untitled finser_30_4_merged_updated_3 afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires december 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in their article, mussa et al., that the short-term financial behavior that the majority of respondents selected “yes” to was . a. always paid credit card in full b. spent less or equal to income c. not overdraft checking account d. none of the above 2. the following variables are associated with good long-term financial behavior in the mussa et al., article. a. all individual tech savvy variables b. no individual tech savvy variables c. mobile use for transfer and mobile use in person d. web app for personal use and web app for work 3. in dilellio and simon’s, mathematical model for taxation, which of the following income sources are taxed at the preferential capital gains tax rate? a. social security b. qualified dividends in the taxable account c. bond interest d. pension benefits 4. in the sensitivity analysis found in seeking tax alpha in retirement income, which variables had the smallest effect on tax alpha? a. percent increase of tax rates after tcja expiration b. glide path transition from stocks to bonds, per year c. inflation rate d. stock and bond rate of return 5. in dr. starr’s article, if we are to ignore variables related to a person’s family of origin, the following kinds of giving are significantly and positively correlated to the amount of retirement savings accumulated, except for: a. donations to religious organizations b. donations to organizations that support youth c. donations to organizations that promote environmental protection d. donations to organizations that help the needy ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning from the editor this issue contains issue 4 of volume 24 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “anchoring, affect and efficiency of sports gaming markets around playoff positioning” is coauthored by kevin krieger at the university of west florida, r. daniel pace at the university of west florida, nicholas clarke at the university of west florida, and clay girdner at keybank corporation. the authors investigate the wagering market of nfl and nba games when participating teams have secured playoff positions. they use both the opening and closing lines (analogous to asset prices) of spread bets to examine if potential “letdown” effects, either psychologically or strategically, are priced. their results demonstrate that the initial opening line consistently provides a profitable strategy for those betting against teams that have clinched positions in the post-season. they show that by the close of the betting cycle, closing lines move in the expected direction as the market partially prices the letdown and that many closing lines tighten to the extent that, after paying commissions, the naı̈ve strategy of betting against clinched teams is less profitable. however, certain wagers, for example betting against nfl teams that have clinched top seeds, are statistically significantly economically profitable after paying commissions. their results appear to support the behavioral finance concept of anchoring. the second article “the perfect withdrawal amount: a methodology for creating retirement account distribution strategies” is coauthored by e. dante suarez at trinity university, antonio suarez (independent advisor), and daniel t. walz at trinity university. the authors present a new way to develop withdrawal strategies from retirement portfolios. it is derived analytically, instead of from empirical testing, and iterates always in the same manner. based on a new measure they develop, the perfect withdrawal amount, for which they discuss how to construct a probability distribution and how to apply it sequentially. they also derive a new measure of sequencing risk and present new strategies built with this framework. the third article, “a new strategy to guarantee retirement income using tips and longevity insurance: a second look” is coauthored by paul j. haensly and k. prakash pai, both at the university of texas of the permian basin. prior research proposes a new investment strategy for retirees that bundles treasury inflation protected securities with a deferred annuity to guarantee real annual withdrawal rates of 5% or more with no risk of financial ruin. the strategy addresses three problems that retirees face: longevity risk, inflation risk, and liquidity risk inherent in the purchase of an immediate annuity. in this research, the authors evaluate the performance of this proposed strategy under realistic assumptions about costs, security design, and markets. in addition, we evaluate how the bequest motive might affect the choice between the proposed strategy and an immediate annuity. the fourth article, “the time perspective of financial advisors and its effect on their decision-making” is authored by kenneth ryack at quinnipiac university.psychological research suggests individuals often display past, present, or future time perspective (tp) biases that impact decision making. the author examines the tp biases of financial advisors from different backgrounds and whether or not the biases impact client recommendations. he financial services review 24 (2015) v–vi 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. finds that consistent with literature that suggests a link between tp and career choice, advisors are future oriented as a group, regardless of their professional background, but contrary to prior tp research, he finds that the bias does not appear to impact their professional decisions. instead, his findings are consistent with research that demonstrates psychological biases are mitigated when professional decision makers perform job related tasks. the final article, “a further examination of equity indexed annuities” is coauthored by andy terry and erick elder, both at university of arkansas at little rock. equity indexed annuities (eias) are deferred annuities that credit interest according to a formula tied to the performance of an underlying equity index. the authors expands previous research, particularly that of reichenstein (2009, 2011), by examining the distribution of returns that could have been created on a rolling monthly basis since 1928 for 11 through 15-year investment horizons. they also examine investment alternatives that include the options imbedded in eias. finally, rather than assuming constant cap rates they allow cap rates to vary with interest rates. they find that for long time horizons the opportunity costs of investing in eias is high. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. in this last issue of 2015, i would like to specifically thank our anonymous reviewers: please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 24 (2015) v–vi manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 from the editor this issue contains issue 4 of volume 23 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “boomers’ life insurance adequacy pre & post the 2008 financial crisis” is coauthored by janine k. scott at massey university and john gilliam at texas tech university. the authors use the survey of consumer finance to examine life insurance adequacy among boomers before and after the financial crisis of 2008. they find a significant difference in 2010 between the baby boomers and the senior generation in life insurance adequacy. variables related to net worth, such as income, marital status, and self-insurability, were also significant predictors of life insurance adequacy. they conclude that given greater life insurance adequacy among those with higher incomes, it may be beneficial for those with mid to low incomes to increase the level of group term insurance. the second article “financial adviser background checks’ is coauthored by bhanu balasubramnian, eric r brisker, suzanne gradisher, all at the university of akron. using the 2009 national financial capability survey, the authors identify demographic characteristics associated with financial adviser users who conduct adviser background checks and/or consider more than one adviser before making a choice in advisors. they find that very few financial adviser users check backgrounds, but find a positive relationship between adviser background checks and trust levels. they conclude that having a reliable, easy, and efficient background check process in place will help improve trust in financial advisers. the third article, “wealth and credit compliance: does economic literacy matter?” is coauthored by celeste varum and alla kolyban both at university of aveiro, campus universitário de santiago. the authors examine the influence of economic literacy upon individuals’ over-indebtedness and households’ wealth. the authors propose that a lack of economic-financial knowledge may have detrimental consequences, in particular reflected in higher exposure to credit and financial risk. the authors provide evidence of the importance of financial literacy in portugal, for both individuals’ over-indebtedness and household wealth. the fourth article, “choosing between value and growth in mutual fund investing’ is coauthored by glenn pettengill at grand valley state university, george chang at grand valley state university, and c. james hueng, at western michigan university. the authors investigate investor’s choices between value and growth mutual funds. they indicate that the value premium demonstrates that value securities outperform growth securities, suggesting that an investor should choose value funds. alternatively, existing studies suggest that growth funds outperform value funds. the authors show that value funds indeed outperform growth funds especially in terms of lower realized risk and higher realized terminal wealth. the authors conclude that previous findings may result from a bias against value in some multifactor models. the final article, “low-beta investing with mutual funds” is by david nanigian at ph.d. financial services review 23 (2014) v–vi 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. the american college and penn state university. the author states that empirical research has shown that investing in low-beta stocks can improve the mean-variance efficiency of an investor’s portfolio. dr. nanigian forms portfolios of mutual funds based on beta and examines whether mutual fund investors can capitalize on this puzzle. he finds that investing in a portfolio of funds in the bottom quintile of beta can improve alpha by a statistically significant 2.9% to 4.9% a year over the top quintile of beta funds. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. i would like to recognize in the table below, those that reviewed for issues 23(1) through 23(4) during the past year. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review thank you to all of the fsr referees for issues 23(1) through 23(4) alex wang university of connec cut jose linares-zegarra university of granada andrei shynkevich kent state university karen benson university of queensland andrew ching university of toronto karen varcoe university of california andy terry university of arkansas at li le rock kevin bracker pi sburg state university annamaria lusardi dartmouth college khaled elkhal university of southern indiana bernhard zwergel university of augsburg larry prather se oklahoma state university chris browning texas tech university larry rose california state cli robb kansas state university ma hew spiegel yale university coleen clark ryerson university michael finke texas tech university dale domian york university michael naylor massey university darren lee university of queensland ning tang san diego state university david nanigian the american college pablo ruiz-verdu universidad carlos iii de madrid david smith university of albany parvez ahmed university of north florida diane docking northern illinois patryk babiarz university of alabama diane schooley boise state peter lichtenberg wayne state university emery trahan northeastern university phillip tew arkansas state university frederick p. schadler east carolina university rachel smith university of mississippi giovanni fernandez stetson university randy gardner university of missouri kansas city j michael collins university of wisconsin robert comment analysis group jaakko aspara aalto universitry robert dubil university of utah jacquelyn humphrey university of queensland robert kunkel university of wisconsin oshkosh jim gilkeson university of central florida sharon devaney purdue university john adams university of texas sherman hanna ohio state university john clinebell university of northern colorodo swarnankur cha erjee university of georgia john grable university of georgia tim query new mexico state university john haslem university of maryland walter woerhide the american college john salter texas tech university william jennins u.s. air force academy john simon australian reserve bank wookjae heo university of georgia john watson monash university yasser alhenawi univeristy of evansville vi editorial / financial services review 23 (2014) v–vi academy of financial services officers president robert moreschi virginia military institute president-elect swarn chatterjee university of georgia executive vice president-program janine scott university shepherd vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations philip gibson winthrop university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president thomas coe quinnipiac university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university terrance martin university of texas–rio grande valley martin seay kansas state university past presidents thomas coe, 2015-2016 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial.org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2018 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 27, no. 1, 2018 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2018 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. call for papers the academy of financial services 32nd annual meeting october, 2018, chicago, il the academy of financial services will hold its annual conference in chicago, il on tuesday and wednesday, october 2-3, 2018. afs will be meeting next year in conjunction with the financial planning association (fpa be). the afs conference will feature speakers, symposia, several special sessions, posters, and a reception. with the generous support of our sponsors, the academy has awarded several best paper awards during past meetings and we anticipate continuing best paper awards in 2018. submission information: research papers and abstracts covering all aspects of individual financial management and education are sought for inclusion in the program. papers in the areas of estate planning, insurance, tax accounting aspects of financial planning, investments, and retirement planning are encouraged. proposals for panel discussions and tutorials devoted to current issues in individual financial management or the practice of financial planning will also be considered for inclusion in the program. several sessions will be registered for continuing education (ce) credit with the cfp® board. for further information: further information will be available soon on the afs website at academyfinancial.org for submission, content questions contact program chair, dr. janine scott at jscott@shepherd.edu. for other details, please contact support@academyfinancial.org. manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 academy of financial services officers president robert moreschi virginia military institute president-elect duncan williams western carolina university executive vice president-program swarn chatterjee university of georgia vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president william chittenden texas state university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama halil kiymaz rollins college frances lawrence louisiana state university tom potts baylor university janine scott massey university martin seay kansas state university past presidents thomas coe, 2015-2016 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2017 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 26, no. 2, 2017 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2017 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president william chittenden texas state university president-elect thomas coe quinnipiac university executive vice president-program robert moreschi virginia military institute vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2016 swarn chatterjee university of georgia vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president lance palmer university of georgia editor, financial services review stuart michelson stetson university directors sherman hanna ohio state university halil kiymaz rollins college frank laatsch univ. of southern mississippi david nanigian the american college tom potts baylor university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake past presidents lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2016 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 25, no. 1, 2016 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2016 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president william chittenden texas state university president-elect thomas coe quinnipiac university executive vice president-program robert moreschi virginia military institute vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2016 swarn chatterjee university of georgia vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president lance palmer university of georgia editor, financial services review stuart michelson stetson university directors sherman hanna ohio state university halil kiymaz rollins college frank laatsch univ. of southern mississippi david nanigian the american college tom potts baylor university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake past presidents lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2014 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 24, no. 1, 2014 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2015 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. from the editor this issue contains issue 1 of volume 25 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “investor’s images of the stock market: antecedents and consequences” is coauthored by dawn m. dobni and marie d. racine both at university of saskatchewan. their research studies stock market image, defined as the sum of impressions about the stock market. the authors seek to understand how several personality-oriented, cognition-based, and demographic variables influence investors’ images of the stock market and how these images impact investing behaviors and outcomes. their findings suggest an individual’s financial literacy, propensity to trust, and sociability are important antecedents of his or her perceptions about the stock market. the data also show that stock market image affects investing motives, risk reduction efforts, emotional responses, and degree of satisfaction associated with investing. the second article “the effects of fund commonality in mutual fund families on fund operating expenses and return correlations: evidence from u.s. equity mutual funds” is authored by youngkyun park at university of idaho. the author examines the effects of the commonality of mutual funds within a fund family, measured by common stock holdings and multi-fund management, on fund operating expenses and return correlations. for u.s. equity funds during the period of 2001�2006, he finds that common stock holdings and multi-fund management are negatively related to fund operating expenses, but positively related to the correlation of fund return residuals, which increases the correlation of fund returns. additionally, he finds that the fund commonalities can have negative net effects on risk-adjusted returns of a portfolio with equity funds that have different investment objectives. the third article, “the overlooked momentum traders in 401(k) plans” is authored by ning tang at san diego state university. using a dataset with over one million 401(k) traders, she investigates momentum trading in 401(k) plans. she identifies momentum traders in each quarter and evaluates how these traders perform. her results indicate the existence of momentum traders. however, there is no evidence that they successfully improve their portfolio performance. instead, momentum sellers sell the outperformed funds. overall, financial services review 25 (2016) v–vi 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. momentum traders could lose up to 2.14% per year. in seeking to explain such losses, it is observed that 401(k) momentum traders follow a naı̈ve momentum strategy. they don’t have the ability to select funds with momentum investing styles but, instead, simply chase past returns. the fourth article, “is a vix etp an investment in the vix?” is coauthored by r. parker clowers at auriemma consulting group, inc. and travis l. jones at florida gulf coast university. they examines vix-based etps (exchange traded products) and illustrate that both the return and risk of these products are not related to the return and risk of the vix index. the authors note that vix etps do not correlate well to the vix index. in fact, these funds are not even designed to have a high correlation to the vix index. individual investors can often mistake vix etps for an investment in the vix index itself, which is incorrect and may lead to a costly mistake. the final article, “a marginal cash flow analysis of mortgagors’ choices” is authored by jim musumeci at bentley university. prior academic research has focused on determinants of the spread between interest rates on conforming versus jumbo and 15-year versus 30-year mortgages, but few authors have helped the borrower determine which choice is better. the authors examine these issues from the borrower’s frame of reference and find that comparisons of mortgage terms can be facilitated by analyzing the marginal cash flows from one mortgage contract to another. for many borrowers the “conventional wisdom” leads to suboptimal choices; making the better choice can easily produce low-risk double-digit returns. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 25 (2016) v–vi in memory of jean louis heck (february 6, 1944 february 5, 2016) you may already be aware of jean’s passing recently on february 5, 2016, but i thought it was important that everyone that knew and loved jean had access to information about jean’s life and final arrangements. jean was the 7th president of academy of financial services during 1991–1992. he helped lay the foundation for our great organization and he stayed actively involved for over 25 years. never one to slow down, jean went on to found the financial education associate and the academy of business education. he launched three academic journals; the journal of financial education, the journal of the academy of business education, and advances in financial education, serving as the journal’s editors for many years. he also served as president and executive director of these organizations until fairly recently, as he prepared for his transition. both academic organizations held annual national conferences, which jean founded and organized every year until this year. in jean’s memory, these conferences, organizations, and journals will continue long into the future. for more than 20 years jean has been close friend and mentor to me and many others. i know he has touched many many lives both personally and professionally. as i’ve personally reflected about the great times and experiences with jean, his memory has brought a warmth to my heart. i hope your memories of jean also bring a sense of serenity as we celebrate his life. financial services review 25 (2016) vii–viii 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. i know jean will be genuinely missed by all of us. our sincere sympathies go out to his family and loved ones for this loss of our great friend. kindest regards, stuart michelson jean louis heck (february 6, 1944 february 5, 2016) u.s. veteran jean louis heck, 71, of wayne, pa, husband of nancy j. (nee nagele) heck, passed away at his home on february 5, 2016, one day short of his 72nd birthday. born on february 6, 1944 in st. louis, missouri, he was the son of the late louis j. and martha j. (nee caples) heck. he was a 1961 graduate of northwest high school in house springs, mo, and a united states navy veteran of the vietnam war. after honorably and proudly serving his country, jean then went on to attend florida junior college in jacksonville, fl where he obtained his associate’s degree, and then onto the university of north florida in jacksonville, fl where he received a bachelor’s degree with majors in economics and science. jean then went on to attend the university of south carolina doctoral course program in finance and statistics where he was awarded his ph.d. in 1980 jean began teaching health care finance in the master’s in health administration program at the medical college of virginia, and then in 1983 he accepted a faculty position teaching finance at villanova university. after 23 years at villanova, jean took early retirement and went to teach at st. joseph’s university in 2006, where he continued to serve on the faculty until his death. at st. joseph’s he was awarded an endowed chair in risk management and insurance. jean had diverse interests outside of academia. he had his pilot’s license, and he also trained to be an emt and volunteered with the narberth ambulance squad. he also served for 16 years as treasurer for the birth center, a free-standing birthing clinic in bryn mawr where his son stephen was born. jean also enjoyed watching the phillies and was thrilled to attend one of the world series games in 2008 when the phillies went all the way to become world champions. jean loved the jersey shore, especially ocean city. he also loved corvettes, owning six over the years, while always searching for the next one. jean was also a big fan of women’s college basketball, attending as many villanova and st. joe’s games as possible. he and nancy also shared a love of cats; rarely was their home without one or two cats during their 35 plus years of their marriage. jean is also survived by his sons, alan heck, of seattle, wa, and stephen heck, of wayne, pa, and his sister, janet davis, of anderson, sc, and many friends and colleagues in the world of academics. he was also preceded in death by his brothers, jimmy white and jerry heck. viii memory / financial services review 25 (2016) vii–viii manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 from the editor this issue contains issue 4 of volume 25 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “exploring the demand for retirement planning advice: the role of financial literacy” is coauthored by martin c. seay at kansas state university, kyoung tae kim at university of alabama, and stuart j. heckman at kansas state university. the authors extend previous research on the relationship between financial literacy and financial advice seeking in three ways: (1) examine financial planner use specifically within the context of retirement planning, (2) incorporate huston’s (2010) framework of financial literacy, and (3) use longitudinal data to investigate the initiation, maintenance, and termination of financial planner use. they find that their results (from the 2010 and 2012 national longitudinal survey of youth 1979) show a positive association between the components of financial literacy and financial planner use for retirement planning. the second article “college student interest in personal finance education” is coauthored by christine harrington and walter smith, both at auburn university at montgomery. the authors investigate demand for investing in financial literacy while in college using survey responses from a cross-section of students. their results indicate that student interest in personal finance education is largely a function of perceived return, time cost, financial independence, and gender where female students have relatively more interest. income, patience in consumption, credit experience, and numerical ability. their results support offering learning opportunities for individual personal finance topics in addition to a personal finance course. the third article, “procedure to determine the optimal roth ira versus deductible ira allocation” is coauthored by robert m. hull at washburn university and john b. hull at american century investment services, inc. the authors develop a procedure to guide the roth ira versus deductible ira (rvd) allocation decision. they find that a modest earning couple can achieve a lifetime wealth gain amounting to about $180,000 in today’s dollars. their models allows changes in key variables such as salary match, adjusted gross income, portfolio returns, and withdrawal years. the fourth article, “determining the return-maximizing portfolio leverage and its limitations” is coauthored by robert a. ott and timothy e. zimmer, both at university of indianapolis. financial services review 25 (2016) v–vii 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. the authors hypothesize that leverage in the risk allocation of an investment portfolio can be an effective strategy in achieving overall portfolio goals. their research focuses on the limitations by explicitly including the volatility drag from leveraging the expected portfolio returns. they show that maximizing the expected portfolio returns with respect to leverage results in a returnmaximizing condition that balances the gains from leverage with the losses in the volatility drag. they graphically illustrate the return-maximizing condition over a range of investment returns to produce a return-maximizing leverage curve. the final article, “household use of financial planners: measurement considerations for researchers” is coauthored by stuart j. heckman at kansas state university, martin c. seay at kansas state university, kyoung tae kim at university of alabama, and jodi c. letkiewicz at york university. using the certified financial planner (cfp) board’s definition of financial planning, the authors evaluate the validity of the measures of financial planner use in publicly available datasets. the author’s review of financial services review, journal of personal finance, journal of financial planning, journal of family and economic issues, journal of consumer affairs, and journal of financial counseling and planning identified seven datasets that were commonly used to investigate financial planner use. they find that of these, the two most promising measures were found in the survey of consumer finances and the national longitudinal study of youth (1979). this article critically evaluates these measures and provides insights into the development of better measures of financial planner use for the future. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. yasser alhenawi university of evansville phil baird duquesne university anup basu australia lew coopersmith rider university brenda cude university of georgia gio fernandez stetson university greg filbeck penn state university philip gharghori monash university angelica gonzalez united kingdom john grable university of georgia suzanne gradisher akron university vickie hampton texas tech university andrea hershatter emory dieter hess university of colgne matthew hurst stetson university david hunter university of hawaii richard kish lehigh university van son lai universite laval kc ma stetson university charles larkin trinity college dublin camilla mazzoli università politecnica david michayluk australia vi editorial / financial services review 25 (2016) v–vii please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review david nanigian csu fullerton david north university of richmond barbera o’neill rutgers wade pfau the american college kenneth ryack quinnipiac university kathyrn simms old dominion university sandeep singh brockport university jerry stevens university of richmond gene stout central michigan dante suarez trinity college ning tang san diego state university barton waring bartonwaring viieditorial / financial services review 25 (2016) v–vii from the editor this issue contains issue 3 of volume 26 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “life insurance as a retirement income tool” is coauthored by russell delibero and wade d. pfau both at the american college of financial services. the authors provide a study to determine whether life insurance can play an important role in an overall retirement portfolio given its tax-preferential treatment. this study develops hypothetical scenarios for different types of individuals with varying ages and distribution periods, while using a historical outlook to determine the proper structure of a variable universal life insurance policy. the authors compare a variable universal life policy to different investment vehicles (both in qualified and non-qualified accounts) on an after-tax basis to better understand the potential tradeoff for tax-deferral and insurance fees within life insurance. the second article “does the source of money determine retirement investment choices?” is coauthored by andrea anthony at golden gate university, kristine beck at california state university northridge, and inga chira at california state university northridge. using a dataset of actual investment choices of oregon state university employees, the authors investigate how investment choices differ among (1) the optional retirement plan (orp) funded by the employer and (2) the investments in 403(b) accounts funded by employees themselves using voluntary salary reduction. they find that the level of risk associated with voluntary, salary reduction investments in 403(b) accounts is lower than the risk these same employees are currently taking in their employer funded 401(a) accounts. they also find that participant investment choices in fidelity are riskier than the choices made by those in tiaa-cref. the third article, “marital status, health and retirement wealth for middle aged and older women” is coauthored by serah shin and hyungsoo kim both at university of kentucky. the authors investigate the different impacts of health problems on wealth among unmarried and married women in their later years. they employ a new health measure of a 22 year sequence of chronic diseases. using data from 2,476 women age 50 or older from the 1992–2014 health and retirement study, they find that unmarried women have more complicated and costly health sequences compared to married women, including multifinancial services review 26 (2017) v–vi 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. morbidity or co-morbidity. they also find that over the 22 year period, retirement wealth for unmarried women with costly health sequences is reduced by approximately $3,600 to $5,400 annually. the fourth article, “relation between financial advisory designations and finra misconduct” is authored by jeffrey m. camarda of camarda wealth advisory group. this study examines misconduct disclosures of undesignated vs. designated florida securities salespeople, and finds adverse disclosure materially decreases for designees. the author finds misconduct increases with males, dual investment advisor/registered representative status, and life insurance sales licensure. the final article, “does the source of cash flow affect spending versus saving?” is coauthored by valrie chambers at stetson university, eugene bland at texas a&m university – corpus christi, and marilyn spencer at texas a&m university – corpus christi. the authors study whether people use different mental accounts for different types of hypothetical revenue windfalls rather than viewing them as fungible in their use consistent with neoclassical economics. they find that the income source sometimes influenced the amount spent/saved and a respondent’s general default as a spender or saver was highly significant. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 26 (2017) v–vi manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 determining the return-maximizing portfolio leverage and its limitations robert a. ott, cfaa,*, timothy e. zimmera aassistant professor, school of business, 1400 east hanna avenue, indianapolis, in 46227, usa abstract leverage in the risk allocation of an investment portfolio can be an effective strategy in achieving overall portfolio goals. while the literature on portfolio leverage is robust, quantifying the amount and discussion of its limitations are often minimized. this article focuses on the limitations by explicitly including the volatility drag from leveraging the expected portfolio returns. maximizing the expected portfolio returns with respect to leverage results in a return-maximizing condition that balances the gains from leverage with the losses in the volatility drag. the return-maximizing condition is graphically illustrated over a range of investment returns to produce a return-maximizing leverage curve. © 2016 academy of financial services. all rights reserved. jel classification: d14 microeconomics: personal finance; g00 financial economics: general keywords: investment; leverage; investment returns; return-maximizing investment leverage 1. introduction behavioral finance recognizes the limitations of individual decision making when the process is complicated by an overwhelming amount of information. rather than facilitating better decisions, the profusion of information can potentially lead to suboptimal outcomes by intimidating individual investors and acting as an impediment to the decision-making process. in this manner, information availability can be its own moral hazard that restricts individual access. to compensate and simplify the decision-making process, individuals * corresponding author. tel.: �1-317-788-3317; fax: �1-317-788-3586. e-mail address: ottr@uindy.edu (r. a. ott) financial services review 25 (2016) 415–425 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. frequently implement rules-of-thumb decision devices that approximate value-maximizing decisions. individuals increasingly find themselves in the role of investment decision makers, especially regarding retirement funds. as defined contribution retirement vehicles grow in number, there is an ever increasing need for individuals to gain the required knowledge to appropriately manage these accounts. according to the employment benefit research institute (2011), the use of 401(k) and individual retirement accounts (iras) among working individuals 21 to 64 has increased from 1996 through 2009 in the united states. in 2009, about 33% of workers utilized a 401(k) and over 20% used iras. providing investment tools and research in portfolio management are essential if the public are to become better stewards of their retirement accounts. a key concern for individuals managing retirement accounts is the preservation of retirement funds while also providing for sufficient growth to maintain future purchasing power. while people wish to be conservative with retirement funds, there is a real risk of capital exhaustion during retirement. excess caution with retirement funds management can be as disastrous to a retirement account as excessive weighting towards risk. an investment portfolio should appropriately balance risk and risk-free allocations to ensure current income and asset growth. using leveraged investments within the risk allocation of an investment portfolio can help provide the desired asset growth. current research assesses the validity of adding leverage to the risk allocation of an investment portfolio. this article adds to current research by deriving an algebraic equation for the return-maximizing level of leverage. 2. review as individuals are increasingly responsible for management of investment portfolios, particularly retirement portfolios, the volume of research examining the methods of management has increased. given limited interest and time availability in respect to the volume of research, simplified approaches which approximate optimized strategies are desired and well received by the public. one of the most cited of these approaches is the 4% rule provided by william bengen (1994). according to this rule, from a well-constructed portfolio, 4% annually can be removed for consumption. it is argued that this approach will provide the portfolio owner with sustainable spending and limit the possibility of portfolio exhaustion. in the years since publication, numerous studies have attempted to enhance the approach (bengen 1997, 2001; cooley, hubbard, and walz, 2003; guyton, 2004; guyton and klinger, 2006; stout, 2006, 2008) in response to critiques. the strategy still maintains significant interest in academic literature and public use. examinations of utility models are some of the earliest forms of investment portfolio research. seminal works on utility approaches include merton (1969) and samuelson (1969). these approaches advocate constant risk portfolios with a mix of fixed return assets combined with risk assets. the risk assets such as stocks allow for growth, while fixed return assets such as bonds ensure a positive income stream even in the event of investment volatility. however, requiring a fixed ratio between bonds and stocks can be problematic, especially during periods of market volatility. in periods of a severe stock market correction, 416 r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 the income generated from the fixed asset portion of the portfolio may be unsustainable as bonds are sold to purchase stocks to maintain the fixed asset ratio. attempts to modify this approach and create a sustainable structure during stock market declines are introduced by perold and sharpe (1988). they propose selling stocks and buying bonds in market downturns to reduce risk and increase sustainability. the approach is called constant proportion portfolio insurance (cppi). dybvig (1995, 1999) uses an expected utility maximization model to determine the optimal investment strategy that includes the preference for sustainable income. however, adoption of this type of strategy is often rigorous, complex, and time-consuming. scott et al. (2009) successfully extend this line of research by simplifying the application. the authors propose a rule-of-thumb for managing retirement funds while maintaining a given level of spending which is dubbed the floor-leverage rule. under this rule, 85% of the retirement assets is allocated to risk-free assets such as bonds to maintain current spending levels. the remaining 15% is allocated to a risk portfolio such as stocks that is designed to build wealth to increase future consumption. the authors demonstrate that this method is at least as efficient as other compensation strategies in maximizing utility. to increase the growth potential of the risk allocation scott and watson (2013) recommend a stock portfolio which is leveraged. the authors suggest investing in stocks leveraged three times through exchanged-traded funds (etfs), despite the limitations of volatility drag as noted by jarrow (2010) and sullivan (2009). the choice of leverage in the scott and watson model is exogenously provided by the authors. this article builds on this work by endogenously deriving an equation to determine the return-maximizing level of leverage unique to each investment scenario. the equation integrates the presence of volatility drag, which had been previously noted but not incorporated. the proposed method is generalized and available to multiple applications. 3. methodology the approach to managing retirement assets suggested by this paper follows that of scott and watson (2013) by first setting aside a risk-free portfolio to guarantee a minimum standard of living before using leverage to optimize the residual portfolio. the advantage of this approach is that the level of risk-taking in the residual portfolio will have no impact on the risk-free portfolio. the 85% risk-free allocation they suggest is a useful rule-of-thumb but may not be appropriate for all portfolio holders. the specific allocation should account for the size of the portfolio and the consumption patterns of the portfolio owner. the larger the portfolio and/or the smaller the minimal consumption needs, the smaller the optimal risk-free portfolio requirement. an assumption of the split-account portfolio model is that sufficient income can be generated from the risk-free assets to sufficiently cover consumption needs. in instances where this assumption does not hold, either alternate strategies or reassessment of consumption needs may be required. in establishing an appropriate allocation mix to achieve both the current income needs and growth, a model should minimally consider the consumption needs of the portfolio owner and the size of the portfolio. if the goal of the risk-free portfolio is to maintain a minimal 417r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 level of real purchasing power, a conservative approach would simply capitalize the minimum annual consumption needs (s) by the real risk-free rate (rf). dividing the resulting value by the size of the portfolio (p) provides a risk-free allocation (z) of the portfolio as shown in the following equation. z � �s/rf�/p (1) this simple formulation illustrates that the lower the minimum consumption needs or the larger the portfolio, the smaller the allocation to a risk-free portfolio. the risk-free allocation of the portfolio would be invested in real risk-free assets, such as tips, to maintain its real purchasing power. the remaining percentage of the portfolio (1 – z) could be invested in riskier assets that focus on maximizing expected returns. as an example, an individual with an investment portfolio of $3 million, minimal annual spending needs of $50,000, and a real risk-free rate of 2% will have a risk-free allocation of approximately 83%.1 the remaining 17% of the portfolio would be available to be invested in riskier assets that can be leveraged to maximize expected return. eq. (1) provides a simple formula for deriving a risk-free allocation of a portfolio. however, the assumption of an infinite time horizon, thus guaranteeing the preservation of the portfolio’s principle in perpetuity, is overly restrictive. a less stringent assumption would allow for a limited time horizon and permitting the principle to be spent over time. eq. (1) can be modified as a fixed annuity adjusted for inflation. the adjustment allows for principle exhaustion at the end of the time horizon (t), as shown in the following equation. z � �s � �1 � �1 � rf� �t�/rf�/p.2 (2) for portfolio planning over a finite number of years, eq. (2) provides a reasonable allocation mix. as an example, by incorporating a 40-year time horizon into eq. (2), the portfolio allocation of risk-free assets needed is reduced from 83% to 46%.3 as 46% will now generate sufficient income to minimally cover the consumption needs, the larger proportion of the remaining assets can be invested in riskier assets to produce maximum future purchasing power. clearly, the decision to allow for possible principle exhaustion has a significant impact on the allocation decision between risk and risk-free assets. given the long-term application of this method, the allocation mix can be examined annually using eq. (2) and the portfolio modified when appropriate. annual adjustment ensures long term sustainability of the portfolio in line with spending rules adjustments as proposed by waring and siegel (2015).4 the second step in this approach to managing retirement assets is to maximize the expected return by leveraging the risk portion of the portfolio. using leverage in the risk allocation of the portfolio can better achieve the long term capital preservation aim of the portfolio holder. a leveraged investment provides benefit to the holder if the investment return (ri) exceeds the cost of borrowing (rb). the leverage and spread between return and borrowing costs can add expected incremental returns to the portfolio that is expressed in the following equation: rp � ri � �ri � rb� �n � 1�, (3) 418 r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 where rp is the expected return on the portfolio using a leverage multiple of n, represented as the ratio of the portfolio to the equity investment. the expression n-1 in this equation is the debt-to-equity ratio. as an example, an investment yielding 8% that is leveraged three times, with borrowings costing 2%, yields an expected portfolio return of 20% [8% � (8%–2%) (3–1)]. as long as the expected investment return exceeds the cost of borrowing, eq. (2) suggests that investors could benefit from ever-increasing levels of leverage. in a world of constant returns, increasing levels of leverage generate infinite returns on the leveraged portfolio.5 however, investment returns are not constant and portfolio returns from ever-increasing leverage are bounded. a limiting factor on investment leverage is volatility drag, as implied in the scott and watson (2013) analysis. the application of leverage within a portfolio increases the potential of higher returns. however, the downside of increased portfolio leverage is cost of volatility drag as noted by many including booth and fama (1992), messmore (1995), sullivan (2009), and jarrow (2010). volatility drag is the difference between the arithmetic and the geometric average returns. as an illustration, scott and watson (pp. 55–56) used the example of an investment that over the course of 250 days randomly earns 1% half the days while losing 1% on the other half. while this investment yields an arithmetic return of zero, the geometric return is �1.24%. in this example, the volatility drag is the 1.24%. for normally distributed returns, the relationship between the arithmetic and geometric returns can be expressed more generally as6 rn � rp � 1⁄2��n 2�, (4) where rp represents the arithmetic return, rn the geometric return, and �n 2 the variance of the returns. in the previous example, the arthritic return is zero and the variance is 2.50%.7 if the returns were normally distributed, the geometric return on the investment would be a �1.25%, instead of the actual �1.24%. the volatility drag is one-half the variance of the returns. the aim of a leveraged investment is to increase the expected portfolio return. however, a leveraged investment also increases the volatility drag because the increase in volatility of leveraged assets carries more risk. the general expression for the variance of an investment leveraged n times is the product of the variance of the investment and the squared multiple as shown below. �n 2 � �2n2. (5) as more leverage is applied, the return volatility of the asset increases as well as the resulting volatility drag. for example, if the investment example of scott and watson illustrated above is leveraged three times, the arithmetic return remains at zero percentage, while the geometric return is estimated to lose 11.25%.8 the increasing volatility drag functions as a binding constraint to increases in the portfolio return from the deployment of portfolio leverage. the return on an investment with leverage n can be derived by combining eqs. (3) through (5) as shown in the following expression. 419r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 rn � ri � �ri � rb� �n-1� � 1⁄2��2n2�,9 (6) where one-half the variance of the leveraged investment of eq. (5) represents the volatility drag. when no borrowings are used, n equals one and eq. (6) simplifies to eq. (4). using the assumptions ri � 8%, rb � 2%, and � � 18%, the leveraged return of eq. (6) can be expressed as:10 rn � 8% � 6% �n-1)�1.62% n2, (7) and depicted in fig. 1 as the leveraged return curve. the straight line represents the expected return from leverage with no volatility drag. this leveraged return curve shows that the expected gain from leveraging eventually is dominated by the volatility drag. in this case the maximum return occurs at a leverage multiple of 1.85 times.11 when estimating the volatility drag in eq. (6), the time horizon of the investor becomes an important factor. blume (1974) showed that the geometric average return r(t) for a particular time horizon t can be approximated as a weighted average of the arithmetic average return and a geometric average return estimated over n years as follows: r�t�n � rp �n-t�/�n-1� � rn �t-1�/�n-1). (8) if t equals one, the average return is the arithmetic return and the leveraged return curve is the straight line in fig. 1. if t equals n, the average return is the estimated geometric return and the leveraged return curve is the curved line in fig. 1. depending on the time horizon of the investor, the average return will lie between the arithmetic and geometric returns, with the leverage return curve lying between the two curves in fig. 1. more important, the volatility drag will lessen as the investor’s time horizon shortens, influencing the leverage that achieves the maximum return. the return on an investment with leverage n shown in eq. (8) can be modified for the investor’s time horizon t by substituting eqs. (3) and (6) into eq. (8). r�t)n � ri � �ri � rb� �n-1� � 1⁄2��2n2��t-1�/�n-1�. (9) fig. 1. leveraged return curve. 420 r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 the investor’s time horizon must lie between 1 and n. if t is equal to one, the volatility drag is zero and the portfolio return is the arithmetic return. as t approaches n, the volatility drag becomes one-half the variance of the leveraged investment expressed in eq. (6). the volatility drag of leverage is a constraint on the portfolio’s performance, which is now realized in the equation for the return on a leveraged investment. the maximum expected return with respect to the level of leverage is determined by taking the first order partial derivative of eq. (9) with respect to the leverage multiple, n, and setting the equation equal to zero. solving for n provides the following expression for the leverage multiple that maximizes the expected return.12 n � �ri � rb�/��2�t-1�/�n-1)). (10) an interpretation of eq. (10) can be better understood by rearranging the equation as the following expression. �ri � rb� n � �2n2 �t-1)/(n-1). (11) the left hand side of the equation represents the return on the leveraged portfolio (rp) less the borrowing cost while the right hand side is the variance of the leveraged return for a time horizon of t. an investor will continue to benefit from leveraging as long as the gain from leveraging is greater than the variance of the leveraged return. substituting eq. (10) into eq. (9) gives the maximum return with leverage as the following expression. r�t�n � rb � 1⁄2 ���2��n2���t-1)/(n-1))].13 (12) the maximum return equals the borrowing cost plus one-half the variance of the leveraged returns. this equation is the straight line depicted in fig. 2 as the leveraged curve. in examining the application of leverage within etf’s, cooper (2010) found a consistent relationship between leverage and returns over multiple time periods and indices as that observed in fig. 2. 4. leveraging the s&p 500 to test the return maximizing leverage of eq. (10), we examined the monthly returns to the s&p 500 index, including dividends, over 65 years from november 1950 through fig. 2. the leveraged curve (risk measured with variance). 421r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 october 2015.14 the average monthly return was 0.688%, representing an annual percentage rate (apr) of 8.25%. the variance of these monthly returns was 0.173% or 2.08% annualized. using eq. (4), the estimated geometric return is 0.601%, representing an apr of 7.21%. the actual geometric return over the 65 years was 0.600% or an apr of 7.20%. in this data series, eq. (4) slightly overestimates the actual geometric return.15 to calculate the return maximizing leverage, a borrowing rate must be specified. we chose to simply use the 3-month treasury bill rate as a proxy for the borrowing rate. over the 65 years, the 3-month treasury bill rate averaged 0.366% monthly or 4.40% apr. using the s&p 500 average return and the variance, the return-maximizing leverage over the 65 years implied by eq. (10) was 1.85 times. to examine the accuracy of this measure, we calculated the leverage that maximizes the actual return. calculating the actual monthly return on a leveraged portfolio using eq. (3), we solved for the leverage multiple that would have generated the greatest return over the 65 years. this optimal leverage multiple was 1.77 times, 4.3% percent less than our estimated value. 5. discussion the derived return-maximizing leverage in eq. (10) allows for easy application within the risk allocation of an investment portfolio. using the assumptions from eq. (7)16 in addition to n � 85, and t � 40, the leverage multiple that maximizes the portfolio return is 3.99 times. as the investor’s time horizon lengthens, the leverage multiple is reduced toward 1.85 times, the leverage multiple if t � n. scott and watson suggest a leverage multiple of 3 implies a 53-year time horizon. rather than an exogenous selection of leverage, eq. (10) provides an investor the means to tailor the leverage to the specific circumstances. the introduction of the volatility drag explicitly in the portfolio return differentiates this study from previous studies. without volatility drag, ever-increasing levels of leverage increase portfolio returns. volatility drag limits the portfolio returns from leveraging, eventually outweighing the gains. continuing the same example, the influence of volatility drag on portfolio returns is displayed in fig. 3 which illustrates the expected leveraged return across the variances of the leveraged portfolio.17 as the expected investment return increases, the fig. 3. risk/return with leverage (risk measured with variance). 422 r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 leverage multiple that maximizes the portfolio return also increases, as expressed in eq. (12) and depicted in fig. 2. once the return-maximizing level of leverage is determined, scott and watson (2013) suggest using leveraged exchange-traded funds (etfs) to achieve the desired level of leverage. an investor can simply choose the etf that uses leverage closest to the returnmaximizing derived leverage. if the exact leverage required is not available, or more precision is sought, it is possible to invest in multiple etfs with different leverages, adjusting the investment weight between them to achieve the desired aggregate leverage. the approach assumes a normal long-term distribution of market returns in an attempt to simplify the model. acknowledging the non-normal long-term market returns, and incorporating terms to account for distribution skewness and kurtosis within market returns would add to the basic model. this area would be open to further research and certainly expand the application of the approach presented in this article. 6. conclusion this article provides endogenously derived equations designed to assist in managing a retirement investment portfolio. the approach is to allocate retirement funds between a risk-free portfolio to support minimum annual consumption spending and a risk portfolio leveraged to provide long-term purchasing power. while leveraging the risk portfolio can enhance expected portfolio returns, this article emphasizes the limitations by explicitly including the volatility drag of leveraging. if the spread of the leveraged portfolio return over the borrowing costs is greater than the variance of the leveraged return, then the investor can enhance expected portfolio returns by increasing leverage. otherwise, the volatility drag of leveraging diminishes the expected portfolio returns. this limitation occurs at higher leverage multiples if an investor has a shorter time horizon. while this article focuses on using leverage in a retirement portfolio, the approach is applicable to other issues involving the use of leverage to enhance portfolio returns. in particular, hedge fund managers and managers of financial institutions typically operate with high levels of leverage and may experience the limitations from volatility drag. this topic is also relevant in addressing the optimal capital structure for a firm in corporate finance. notes 1 this allocation is close to the 85% suggested by scott and watson (2013). 2 as t approaches infinity, the influence of the annuity diminishes back to 1/rf and eq. (2) reverts back to eq. (1). in such a case, the principle never depletes and the portfolio will exist in perpetuity. 3 scott and watson (2013) used 40 years in their example. 4 waring and siegel (2015) propose a spending rule that constantly adjusts the annuity through time so as to never deplete the portfolio. 5 this is modigliani-miller’s (1958) proposition ii. 423r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 6 bodie, kane, and marcus (2011) p. 132. 7 �n 2 � 250 � 0.01% � 2.5%. 8 ri � 0% and �2 � 22.5% [or 250 � 0.01% � 9]. in the scott and watson example, the actual loss from 3 times leverage is �10.64%, p. 56. 9 this equation can also be expressed as: rn � rb � (ri � rb) n � 1⁄2(�2n2). 10 these are the assumptions used by scott and watson (2013, p. 49). 11 when n � 0, the portfolio is leveraged �1 times. this is equivalent to selling the investment portfolio and investing it at the borrowing rate. 12 rn � ri � �ri � rb� �n-1)�(1⁄2�2)(n2���t-1)/(n-1)), �rn/�n � �ri � rb� � �2n��1⁄2�2���t-1)/(n-1)), 0 � ri � rb � �n���2���t-1)/(n-1)), n � �ri � rb�/��2� ��t-1)/(n-1)). 13 eq. (12) can also be expressed as a function of the spread between the investment yield and the borrowing cost as r(t)n � rb � 1⁄2 (ri � rb)2/�2 (t-1)/(n-1)). 14 data series is from yahoo finance. 15 the precise relationship is derived from continuously compounded rates that are lognormally distributed. see jacquier, kane, and marcus (2003). 16 ri � 8%, rb � 2%, and � � 18%. 17 figure 2 is the same curve as figure 1, but with the variance as the dependent variable instead of leverage. references bengen, w. t. (1994). determining withdrawal rates using historical data. journal of financial planning, 7, 171–180. bengen, w. t. (1997). conserving client portfolios during retirement, part iii. journal of financial planning, 10, 84–97. bengen, w. t. (2001). conserving client portfolios during retirement, part iv. journal of financial planning, 14, 110–119. blume, m. e. (1974). unbiased estimators of long-run expected rates of return. journal of the american statistical association, 69, 634–638. bodie, z., kane, a., & marcus, a. j. (2011). investments (9th ed.). new york, ny: mcgraw-hill. booth, d. g., & fama e. f. (1992). diversification returns and asset contributions. financial analysts journal, 48, 26–32. cooley, p. l., hubbard, c. m., & walz, d. t. (2003). a comparative analysis of retirement portfolio success rate: simulation versus overlapping periods. financial services review, 12, 115–128. cooper, t. (2010). alpha generation and the risk smoothing using managed volatility (available at ssrn 1664823). dybvig, p. h. (1995). dusenberry’s ratcheting of consumption: optimal dynamic consumption and investment given intolerance for any decline in standard of living. review of economic studies, 62, 287–313. dybvig, p. h. (1999). using asset allocation to protect spending. financial analysts journal, 55, 49–62. employee benefit research institute. (2011). ebri notes (vol. 32, october). guyton, j. t. (2004). decision rules and portfolio management for retirees: is the ‘safe’ initial withdrawal rate too safe? journal of financial planning, 17, 54–62. 424 r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 guyton, j. t., & klinger, w. j. (2006). decision rules and maximum initial withdrawal rates. journal of financial planning, 19, 48–58. jacquier, e., kane, a., & marcus, a. j. (2003). geometric or arithmetic mean: a reconsideration. financial analysts journal, 59, 46–53. jarrow, r. a. (2010). understanding the risk of leverage etfs. finance research letters, 7, 135–139. merton, r. c. (1969). lifetime portfolio selection under uncertainty: the continuous-time case. review of economics and statistics, 51, 247–257. messmore, t. e. (1995). variance drain. journal of portfolio management, 21, 104–110. modigliani, f., & miller, m. (1958). the cost of capital, corporate finance and the theory of investment. american economic review, 48, 3:261–297. perold, a. f., & sharpe, w. f. (1988). dynamic strategies for asset allocation. financial analysts journal, 44, 16–27. samuelson, p. a. (1969). lifetime portfolio selection by dynamic stochastic programming. review of economics and statistics, 51, 239–246. scott, j. s., sharpe, w. f., & watson, j. g. (2009). the 4% rule: at what price? journal of investment management, 7, 1–18. scott, j. s., & watson, j. g. (2013). the floor-leverage rule for retirement. financial analysts journal, 69, 45–60. stout, r. g. (2008). stochastic optimization of retirement portfolio asset allocations and withdrawals. financial services review, 17, 1–15. stout, r. g. (2006). dynamic retirement withdrawal planning. financial services review, 15, 117–131. sullivan, r. (2009). the pitfalls of leveraged and inverse etfs. cfa magazine, 20, 12. waring, m. b., & siegel, l. b. (2015). the only spending rule article you will ever need. financial analysts journal, 71, 91–107. 425r.a. ott, t.e. zimmer / financial services review 25 (2016) 415–425 from the editor this issue contains volume 27 issue 1 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “credit usage, payment behavior and the accuracy of consumer credit files” is authored by l. douglas smith at university of missouri-st. louis, michael staten at university of arizona, thomas eyssell at university of missouri-st. louis, maureen karig at university of missouri-st. louis, jeffrey feinstein at lexisnexis risk solutions, and cathleen johnson at university of arizona. in this research, the authors conduct interviews, examine credit reports, and rescore corrected credit files, to consider household characteristics, major life events, financial resources, and payment habits in order to study the integrity of credit-bureau data, vulnerability to error, and results of disputes filed with the major credit bureaus. they find that credit usage and management vary widely within demographic groups. they conclude that consumers with moderate credit scores are more likely than those with very high or low scores to see significant improvement in their records when errors are corrected. the second article “age when first employed and retirement wealth of baby boomers” is coauthored by hyungsoo kim at university of kentucky, serah shin at university of kentucky, qun zhang at university of kentucky, and martie gillen at university of florida. the authors examine how age when first employed is related to retirement savings in later years. using data from the health and retirement study, they investigate two specific questions: has age when first employed affected the retirement wealth of baby boomers? if so, to what extent? their results show that age when first employed is negatively associated with accumulated retirement wealth in later years. for college graduates delaying the start of employment cost approximately $35,103 per year in retirement savings after controlling for demographic characteristics, number of working years, and occupation types. the third article, “conflicted advice about portfolio diversification” is coauthored by sally shen at global risk institute in financial services and john a. turner at pension policy center. the authors investigate the validity of the argument by the financial services industry to “roll over your ‘old’ 401(k) plan” because 401(k)-type plans have a limited number of investment options, while iras have a virtually unlimited number of options. financial services review 27 (2018) v–vi 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. they analyze the diversification of a large 401(k)-type plan with only five basic investment options. they find that financial advisers with a conflict of interest may use strategic complexity to encourage rollovers, recommend complex portfolios to impress naı̈ve clients, while not weighing the cost of the complex portfolios against any added benefits of diversification. the fourth article, “the role of perceived quality of personal service in influencing trust and satisfaction with banks” is coauthored by anders carlander, amelie gamble, tommy gärling, jeanette carlsson hauff, lars-olof johansson, and martin holmen, all at university of gothenburg. the authors investigate if trust in banks increases with repeated personal contacts with respect to the customer-employee relationship. they utilize data from an on-line survey of customers of swedish retail banks and find that trust in the bank is influenced by perceived quality of personal service through employees’ perceived competence, perceived benevolence, and perceived transparency. they also find that satisfaction with the bank is influenced by perceived quality of personal service through perceived competence, perceived benevolence, and perceived transparency. the final article, “expense ratios and net alphas of large cap funds: do expenses add value?” is coauthored by abhay kaushik and raymond boisvert, both at radford university. the authors analyze the performance of large cap equity funds over the period january 2000 to december 2103 (a particularly tumultuous period), with the objective to assess the performance as reflected in the alpha of funds conditioned on expenses. the results are mixed for large cap funds across individual categories and when performance is measured conditioned on expenses. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 27 (2018) v–vi what determines risk tolerance? michael guillemette, ph.d., cfp�a,*, david nanigian, ph.d.b adepartment of personal financial planning, college of human environmental sciences, university of missouri, 239b stanley hall, columbia, mo 65211, usa bthe richard d. irwin graduate school, the american college, 270 south bryn mawr avenue, bryn mawr, pa 19010, usa abstract it is important for financial planners to understand what drives risk tolerance as it directly influences the portfolio allocation preference of clients. we hypothesize that habit formation, loss aversion and investor sentiment account for significant variation in risk tolerance. we analyze average monthly scores from a widely used risk tolerance questionnaire. we find that the habit formation, loss aversion, and sentiment proxies account for �1.06%, 38.51%, and 13.21% of the variation in average monthly risk tolerance, respectively. habit formation did not account for additional variation in average monthly risk tolerance when controlling for loss aversion and sentiment. © 2014 academy of financial services. all rights reserved. jel classification: d81 keywords: risk tolerance; loss aversion; habit formation; sentiment 1. introduction it is important for financial planners to understand what factors account for variation in risk tolerance. knowledge of these factors will help planners identify what types of economic situations might affect clients’ preferences for risky assets. according to modern portfolio theory assets with a higher variance should have a higher expected return (markowitz, 1952). this corresponds to the concave form of a typical investor’s utility function. the greater the concavity of someone’s utility function, the less willing they are to accept * corresponding author. tel.: �1-573-884-9188; fax: �1-573-884-8389. e-mail address: guillemettem@missouri.edu (m. guillemette) financial services review 23 (2014) 207–218 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. variation in consumption over time. more risk averse individuals must be compensated with a higher expected return, compared with those who are less risk averse, to accept greater consumption variation. the degree of risk aversion determines the optimal mix of risky and risk-free assets within an investor’s portfolio. the three primary theories that help explain willingness to take investment risk include habit formation (abel, 1990), loss aversion (kahneman and tversky, 1979), and investor sentiment (baker and wurgler, 2007). habit formation might explain why client risk tolerance shifts when a job loss or a sudden windfall is experienced. many financial planners are aware that clients become more willing to take investment risk after experiencing prior investment gains and less willing to take risk after successive market declines. loss aversion and the house money effect have been used to explain this behavior, but to what extent does it drive client risk tolerance? many financial planners recall the euphoria clients experienced in the late 1990s when technology stocks were climbing to record highs. retail investors piled into tech stocks during this period of extraordinarily high sentiment, but many who did lost significant wealth when the bubble popped. the extent to which sentiment drives risk tolerance will help financial planners understand how client preferences for risky assets might change in relation to the economic outlook. the next section of this article provides an overview of the literature on factors that have been used to explain variation in risk tolerance. the literature review also helps to explain why different proxies were selected for habit formation, loss aversion, and investor sentiment. section 3 outlines the conceptual framework and hypothesis. section 4 describes the methodologies used to construct measures for habit formation, loss aversion, and investor sentiment. the article concludes with the results and practical implications for financial planners. 2. literature review consumption-based asset pricing models are based on the covariance between asset returns and consumption growth. individuals prefer a smooth consumption path over their lifecycle to maximize expected lifetime utility, which results in a low covariance between asset returns and consumption growth (campbell, 2003). in this case it is difficult to explain the equity premium without an unrealistically high coefficient of relative risk aversion (rra) (campbell, 2003). variation in risk tolerance has been attempted to be explained in the literature through models that incorporate habit formation. models that incorporate habit formation have been introduced to provide a possible explanation for one of the factors that drives risk tolerance. the relative income hypothesis states that individuals evaluate their consumption levels in relation to those of other people, rather than on an absolute basis (duesenberry, 1949). individuals assess their current consumption levels based on a weighted average of their relative recent past consumption under a theory known as habit formation. habit preferences can be either internal or external. abel (1990) proposes an external habit formation model, which is similar to duesenberry’s (1949) “catching up with joneses” hypothesis. habit formation helps explain why the disutility experienced during recessions is so severe, even though the consumption shock is 208 m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 relatively small given the time horizon of the lifecycle (campbell and cochrane, 1999). habit formation implies that risk aversion is time-varying, which means that the optimal allocation of a household’s portfolio to risky assets also varies over time (heaton and lucas, 2000). models of habit formation imply that risk aversion varies with short-term changes in consumption (campbell and cochrane, 1999). constantinides (1990) claims that the equity premium can be explained in a rational expectations model using habit preferences. mehra and prescott (2003) state that habit preferences cannot resolve the equity premium because it results in extreme aversion to consumption risk. they also question whether individuals actually have significant timevarying countercyclical changes in relative risk aversion (rra) that is implied by habit formation models, such as the one developed by campbell and cochrane (1999). habit preferences can explain the difference between the historically low real returns of treasury securities compared with stocks because increased risk aversion increases the quantity demanded for treasury securities, which drives down the risk free rate (weil, 1989). studies in behavioral finance attempt to explain variation in risk tolerance using prospect theory. prospect theory states that individuals evaluate gains and losses from a reference point and describes the utility function as being steeper in the loss domain compared with the gain domain (kahneman and tversky, 1979). benartzi and thaler (1995) find that the historical equity premium can be explained if investors are loss averse and myopic. thaler and johnson (1990) find that individuals experience less disutility from losses after a prior gain and greater disutility after a prior loss. therefore, models that incorporate loss aversion should decrease an individual’s coefficient of loss aversion, �, after prior losses and increase it after prior gains. it is important to note that the more negative � is, the more someone overweighs losses compared with equivalent gains. the empirical finding of thaler and johnson (1990) implies that risk aversion is timevarying. after experiencing prior financial gains, individuals should become less risk averse because prior gains will protect them from subsequent losses. after experiencing prior losses, current losses should make individuals more risk averse. barberis, huang, and santos (2001) study asset prices by incorporating the findings of kahneman and tversky (1979) and thaler and johnson (1990). they find that individuals are loss averse from fluctuations in consumption and that � is dependent on previous investment returns. their framework helps explain the high historical equity premium, the low correlation between stock returns and consumption growth and the excess volatility and predictability of equity returns. investor sentiment is another factor that may help explain variation in risk tolerance. the closed-end fund discount is one proxy for investor sentiment (lee, shleifer, and thaler, 1991; baker and wurgler, 2006; baker and wurgler, 2007). when closed-end funds are less discounted or are priced above net asset value (nav) investors may be optimistic about future returns (lee, shleifer, and thaler, 1991). during periods of high sentiment equity prices mean revert, resulting in lower future returns. poterba and summers (1988) find evidence of mean reversion in stock returns and state that one of the possible explanations is “price fads” that cause equity prices to deviate from fundamental values. the findings of thaler and johnson (1990) imply that when closed-end funds trade at a significant premium to nav investors have become less risk averse. investor sentiment helps explain why risk aversion decreases during high sentiment periods. 209m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 other proxies for investor sentiment include average stock turnover, trading volume, number of ipos, first-day ipo closing prices, the demand for dividend paying stocks, and the equity-to-debt-issue ratio. when noise traders are optimistic, there is greater stock turnover, which increases liquidity. trading volume is a signal that investors have heterogeneous beliefs and differ in their evaluations of equity prices (hong, scheinkman, and xiong, 2006).1 lowry and schwert (2002) state that ipos tend to be held when investors are optimistic and are, therefore, willing to pay an inflated price. cornelli, goldreich, and ljungqvist (2006) find that high gray market prices (a signal that investors are optimistic) are a good predictor of first-day ipo closing prices. the demand for dividend-paying stocks should rise when investors’ marginal propensity to consume is high and they are pessimistic about future returns. baker and wurgler (2002) find that companies issue more equity relative to debt before periods of low stock market returns. 3. conceptual framework and hypothesis the conceptual framework is displayed in fig. 1. habit formation, loss aversion and sentiment should account for significant variation in risk tolerance. risk tolerance directly influences portfolio allocation preference. habit formation assumes the curvature and slope of the utility function are the same in the gain and loss domains. rational agents should derive the same utility and disutility from equivalent gains and losses. however, observed levels of habit formation over small stakes have translated into unrealistically high levels of risk tolerance over larger stakes (rabin, 2000; rabin and thaler, 2001). prospect theory modified consumption models by changing the slope of the utility function in the loss domain (kahneman and tversky, 1979). given the empirical evidence that people are loss averse (tversky and kahneman, 1992; schmidt and traub, 2002; pennings and smidts, 2003; booij and van de kuilen, 2009) we hypothesis that the loss aversion model will account for greater variation in risk tolerance than the habit formation model. 4. methods risk tolerance is measured using a questionnaire that has been developed by finametrica, a leading provider of risk profiling tools. the questionnaire has been psychometrically tested for validity and reliability (moreschi, 2011; van de venter, michayluk, and davey, 2012) and used to profile more than 500,000 people worldwide. the questionnaire includes 25 risk fig. 1. conceptual framework. 210 m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 tolerance questions that can be found at http://goo.gl/18dkl5. scores range from 0 to 100 with zero being most risk averse and 100 being most risk tolerant. the monthly mean risk tolerance scores (mrts) of individuals surveyed in the united states and canada was provided to us by finametrica. the repeated cross sectional data were collected between january 2003 and december 2010. in total, 357,677 different individuals were surveyed. table 1 provides descriptive statistics on the number of people surveyed per month. the minimum number of people surveyed in any given month was 1,640 so we believe that the sample is representative of the broader population of investors. no demographic or socioeconomic data were provided. we use a model developed by illmanen (1995) in this analysis as a proxy for external habit-based preferences. a proxy for external habit formation is derived by taking the exponentially weighted ratio of past real consumption to current real consumption, �. the ilmanen (1995) model is similar to the habit formation model developed by constantinides (1990), as the subsistence level of consumption is the exponentially weighted mean of past consumption. as the gap between the exponentially weighted ratio of past real consumption to current real consumption rises, rra increases. indexed and seasonally adjusted real monthly personal consumption expenditures are obtained from the federal reserve bank of saint louis.2 ilmanen (1995) assigns smaller weights to consumption levels that are further out in time. a smoothing coefficient of 0.90 is used to capture business cycle effects and the weights for the cumulative last 12 months and cumulative last 36 months are 70% and 95%, respectively (illmanen, 1995). eq. (1) displays the derivation of the habit formation proxy. invct � ���ct�1� � �0.9 � ct�2� � �0.92 � ct�3� � . . . � � 0.1� ct (1) the proxy for loss averse preferences is developed by kahneman and tversky (1992) and barberis, huang, and santos (2001). kahneman and tversky (1992) find that individuals weigh losses 2.25 times more than equivalent gains when they are offered isolated gambles. they estimate that the marginally decreasing aspect of the value function, �, is 0.88. in the barberis et al. (2001) model losses are not evenly weighted as there is evidence that sensitivity differs depending on whether a prior gain or loss preceded the current loss. � increases after a prior gain and decreases after a prior loss because of the house money effect (thaler and johnson, 1990). barberis, huang, and santos (2001) create a parameter, k, to table 1 distribution of sample descriptive n mean 3,726 � 1,356 75th percentile 4,476 median 3,683 25th percentile 2,683 minimum 1,640 maximum 8,047 211m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 determine how much more painful losses are after a prior loss and how much less painful they are after a prior gain. they find that k � 3 results in a mean � that is approximately �2.25. for example, if the stock market falls 10% in a given month k is multiplied by �0.10 and then added to �2.25, which results in a loss aversion weight, w, of �2.55. if the stock market rises five percentage in a given month k is multiplied by 0.05 and then added to �2.25 which results in w � �2.10. the return on fama and french’s value-weighted portfolio of u.s. stocks3 is used to proxy for the market return, mkt. the one-month treasury bill rate, rf, is subtracted from mkt to account for the opportunity cost of investing in the equity market. the derivation of the loss aversion proxy is displayed in eq. (2). �mkt � rf �0.88 if �mkt � rf � � 0, else w���1��mkt � rf ��0.88 (2) shumway (1997) develops an asset pricing model based on loss averse investors. the model explains annual returns better than competing models, but it does not explain monthly, quarterly, or half-year returns. this is consistent with the finding that a one-year evaluation period is utility maximizing assuming that investors are myopic and loss averse (benartzi and thaler, 1995). a one-year moving average is used for the loss aversion proxy. baker and wurgler (2007) develop an index to measure investor sentiment that includes the monthly change in the closed-end fund discount, cefd, detrended log turnover, turn, the number of ipos, nipo, the first day return on ipos, ripo, the dividend premium, pdnd, and the equity share in new issues, s, as factors. the index is standardized to have a mean of zero and a variance of one (baker and wurgler, 2007). eq. (3) displays the baker and wurgler (2007) sentiment index formula.4 �sentiment � �0.17�cefd � 0.32�turn � 0.17�nipo � 0.41�ripo � 0.49pdnd � 0.28�s (3) 5. results descriptive statistics on all of the regression variables are reported in table 2. the average coefficient of loss aversion was �2.37, which is consistent with the prior literature (tversky and kahneman, 1992; schmidt and traub, 2002; pennings and smidts, 2003; booij and van de kuilen, 2009). a correlation matrix is displayed in table 3. the highest correlation in the table 2 descriptive statistics mrts loss aversion habit formation sentiment mean 53.2520 �2.3652 0.7016 �0.1124 � 0.9274 4.6431 0.0275 0.3324 75th percentile 53.9462 0.3154 0.7186 0.1200 median 53.2681 �0.7262 0.7069 �0.0530 25th percentile 52.5293 �4.511 0.6945 �0.4070 minimum 51.1398 �15.0205 0.6055 �0.8070 maximum 55.2460 4.3936 0.7360 0.5380 212 m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 matrix is between the loss aversion proxy and mrts. however, this correlation is only 0.63, assuaging concerns of collinearity problems. a shapiro-wilk normality test was run on mrts. the null hypothesis of a normal distribution was not rejected at conventional confidence levels. mrts for 2,327 individuals were analyzed immediately following the recent global financial crisis (gfc) and lower mrts was found among respondents who perceived the stock market to be riskier than it was two years ago (gibson, michayluk, and van de venter, 2013). a positive relation between mrts and positive stock market expectations were also reported during the gfc (gibson, michayluk, and van de venter, 2013). mrts was found to be highly correlated (0.90) with the s&p 500 during the stock market crash of 2008–2009 (guillemette and finke, 2014). however, although risk tolerance was highly correlated with equity market returns it only declined 5% during the global financial crisis, compared with a much greater decline in dutch stock market returns (hoffmann, post, and pennings, 2013). if the finametrica score is used in a linear manner to determine an equity allocation for a client, the average equity shift during the gfc would have been �4%. such a large shift in one’s asset allocation will have a meaningful impact on wealth outcomes, especially over longer time horizons. table 4 displays the results from ordinary least squares (ols) regressions of mrts on the hypothesized factors that account for variation in risk tolerance. the signs of the parameter estimates are consistent with theory for loss aversion and sentiment. the habit formation proxy was not statistically significant in any model. for a prospect theory utility function that incorporates the house money effect, as � increases, mrts increases. the sentiment index is positively associated with mrts. when each of the three hypothesized determinants of mrts are examined in separate univariate regressions (columns 1–3), the loss aversion proxy explains the greatest amount of the variation in mrts. this is evidenced by the largest adjusted r2 value (0.3851) among the three univariate regressions. overall, loss aversion and sentiment contribute meaningfully to explaining variation in mrts. this is evidenced by the adjusted r2 value of our regression model improving from 0.3851 (column 2) to 0.4107 (column 7) when the loss aversion and sentiment variables are added to the plain vanilla model with only the loss aversion proxy. the values of our dependent variable fall within a finite range. therefore, we also examine the results from a � regression model with a logit link specification. a � regression model is a generalized linear model for dependent variables that are marginally distributed following a � distribution. the model was originally designed for percentage data that range from 0 to 100%. the mrts variable is scaled by a factor of 1/100 to conform to the parameters of a � distribution. the � regression model results, which are displayed in table 5, are consistent with the ols results. table 3 correlation matrix invc sentiment loss aversion mrts invc 1.0000 0.3396** �0.1607 0.0014 sentiment 0.3396** 1.0000 0.3332** 0.3758** loss aversion �0.1607 0.3332** 1.0000 0.6258** mrts 0.0014 0.3758** 0.6258** 1.0000 *p � 0.05. **p � 0.01. 213m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 t ab le 4 v ar ia ti on in m r t s –o l s m od el m od el (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) (7 ) in v c t 0. 04 72 (3 .4 8) 3. 52 91 (2 .7 4) � 4. 80 95 (3 .4 1) 1. 17 90 (2 .9 8) l os s av er si on 0. 12 50 ** (0 .0 2) 0. 12 84 ** (0 .0 2) 0. 11 46 ** (0 .0 2) 0. 11 25 ** (0 .0 2) s en ti m en t 1. 04 83 ** (0 .2 7) 1. 18 34 ** (0 .2 8) 0. 48 19 (0 .2 6) 0. 52 50 * (0 .2 3) c on st an t 53 .2 18 9* * (2 .4 4) 53 .5 47 6* * (0 .0 8) 53 .3 69 8* * (0 .0 9) 51 .0 79 7* * (1 .9 2) 56 .7 59 2* * (2 .4 0) 52 .7 50 1* * (2 .0 9) 53 .5 77 0* * (0 .0 8) a dj us te d r 2 � 0. 01 06 0. 38 51 0. 13 21 0. 38 94 0. 14 11 0. 40 53 0. 41 07 o bs er va ti on s 96 96 96 96 96 96 96 *p � 0. 05 . ** p � 0. 01 . t ab le 5 v ar ia ti on in m r t s � re gr es si on m od el m od el (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) (7 ) in v c t 0. 00 19 ( 0. 01 ) 0. 14 18 (1 .3 1) � 0. 19 33 (� 1. 43 ) 0. 04 75 (0 .4 1) l os s av er si on 0. 50 18 ** (7 .8 6) 0. 51 53 ** (8 .0 4) 0. 46 01 ** (6 .6 4) 0. 45 14 ** (6 .8 5) s en ti m en t 0. 04 21 ** (3 .9 7) 0. 04 75 ** (4 .2 6) 0. 01 93 (1 .9 0) 0. 02 11 * (2 .2 9) c on st an t 0. 12 90 (1 .3 3) 0. 14 22 ** (4 2. 89 ) 0. 13 50 ** (3 6. 47 ) 0. 04 30 (0 .5 7) 0. 27 12 ** (2 .8 5) 0. 11 00 (1 .3 4) 0. 14 33 ** (4 3. 85 ) ln l 31 4 33 7 32 1 33 8 32 2 34 0 34 0 a ic � 62 1 � 66 9 � 63 6 � 66 9 � 63 6 � 67 0 � 67 2 o bs er va ti on s 96 96 96 96 96 96 96 *p � 0. 05 . ** p � 0. 01 . 214 m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 6. conclusions loss aversion and sentiment accounted for significant variation in mrts from 2003 to 2010. loss aversion and sentiment accounted for 41.07% of the variation in mrts. when the habit formation proxy was added to the model with loss aversion and sentiment it did not account for additional variation in mrts. this time period, while relatively short, is important because it encompassed the greatest financial panic since the great depression. analysis over a longer time period, if and when a longer time-series of data becomes available, would be an interesting extension for future research. it is in time periods such as these where the assessment of how a client will react to a severe market downturn will be critical in determining whether they continue to follow their financial planner’s investment recommendations. this article provides evidence that more of the variation in mrts is explained by loss aversion than by sentiment. it is essential for risk tolerance surveys to include questions that measure a client’s level of loss aversion. the marginal effect of the sentiment proxy observed in the ols models also supports the inclusion of questions that measure investor sentiment. we find no evidence that measuring the weighted ratio between current and past consumption improves risk tolerance assessment when loss aversion and sentiment are already being measured. 7. practical implications it is imperative that a financial planner assess risk tolerance correctly; otherwise the client may be in a portfolio that is excessively risky. if a client is in a portfolio that is too risky, it may increase the likelihood that he or she will sell out of stocks after a sharp decline in equity prices. this would result in lower future returns, which could possibly preclude or delay a client’s attainment of his or her goals. this is because periods of low equity valuations are usually followed by higher than average stock returns (basu, 1977; campbell and shiller, 1988; fama and french, 1988). helping clients understand their willingness to take risk before a portfolio allocation is constructed will reduce the likelihood that they will sell stocks during a severe market downturn. from 1991 to 2004 investors lost 1.56% annually in dollar-weighted returns because of market timing (friesen and sapp, 2007). there is evidence that myopic behavior may play a role in the reluctance to invest in stocks. thaler, tversky, kahneman, and schwartz (1997) provide experimental evidence that the frequency at which investment performance is presented can affect a client’s propensity to invest in stocks. in the experiment, different groups of investors were compared. one group was shown return data and participants were asked to allocate their portfolios between stock and bond funds on a monthly basis. another group was shown the same data as the first group but allocated their portfolios between stock and bond funds on an annual basis. participants in the first group allocated 59.1% of their portfolios to bond funds, yet participants in the second group allocated only 30.4% to bond funds. helping clients identify myopic behavior is important because evidence has indicated the risk of equities is decreasing in the length of one’s holding period (blanchett, finke and pfau, 2013). myopic behavior could include checking stock prices or viewing investment state215m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 ments on a daily, monthly or even quarterly basis. since stocks should only be used to meet long-term goals, short-term fluctuations are irrelevant. oftentimes the financial media sells the misperception that short-term fluctuations matter, but evidence suggests that even professional fund managers cannot successfully time the market (carhart, 1997; detzel and weigland, 1998; cremers and petajisto, 2009). behavioral strategies that help clients take a long-term view and stay invested in stocks will help increase the likelihood that they will accomplish their goals. a certified financial planner designation increased investor certainty during periods of underperformance, improving a client’s ability to maintain a consistent investment approach during market downturns (james, 2013). winchester, huston, and finke (2011) found that those who had a financial planner, and particularly those who had a written plan (that included an investment policy statement), were far less likely to shift their wealth into cash during the 2008 recession. simple techniques such as waiting until the end of client meetings to discuss returns, and emphasizing longer run performance when they are discussed, are other ways that may help keep clients in portfolios that are aligned with their preferences. notes 1 it should also be noted that when an index changes its constituents; then index funds need to conduct trades simply to continue to replicate their index. 2 real personal consumption expenditure data can be found at https://research.stlouis fed.org/fred2/series/pcec96/. 3 the return data can be found on kenneth french’s web site at http://mba.tuck.dart mouth.edu/pages/faculty/ken.french/data_library.html#research. we are grateful to kenneth french for providing this data. 4 the sentiment data can be found on jeffrey wurgler’s web site at http://people.stern. nyu.edu/jwurgler/. we are grateful to jeffrey wurgler for providing this data. acknowledgments the authors would like to thank geoff davey and finametrica for supporting the advancement of risk tolerance research. david nanigian is grateful to the new york life insurance company for their financial support. references abel, a. b. (1990). asset prices under habit formation and catching up with the joneses. the american economic review, 80, 38–42. baker, m., & wurgler, j. (2002). the equity share in new issues and aggregate stock returns. the journal of finance, 55, 2219–2257. baker, m., & wurgler, j. (2006). investor sentiment and the cross-section of stock returns. the journal of finance, 61, 1645–1680. 216 m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 baker, m., & wurgler, j. (2007). investor sentiment in the stock market. journal of economic perspectives, 21, 129–151. barberis, n., huang, m., & santos, t. (2001). prospect theory and asset prices. the quarterly journal of economics, 116, 1–53. barberis, n., shleifer, a., & vishny, r. (1998). a model of investor sentiment. journal of financial economics, 49, 307–343. basu, s. (1977). investment performance of common stocks in relation to their price-earnings ratios: a test of the efficient market hypothesis. the journal of finance, 32, 663–682. benartzi, s., & thaler, r. (1995). myopic loss aversion and the equity premium puzzle. quarterly journal of economics, 110, 73–92. blanchett, d., finke, m. s., & pfau, w. (2013). optimal portfolios for the long run. available at ssrn: http://papers.ssrn.com/sol3/papers.cfm?abstract_id�2320828. booij, a. s., & van de kuilen, g. (2009). a parameter-free analysis of the utility of money for the general population under prospect theory. journal of economic psychology, 30, 651–666. campbell, j. y., & cochrane j. h. (1999). by force of habit: a consumption-based explanation of aggregate stock market behavior. journal of political economy, 107, 205–251. campbell, j. y. (2003). consumption-based asset pricing. handbook of the economics of finance, 1, 803–887. campbell, j. y., & shiller, r. j. (1988). stock prices, earnings, and expected dividends. the journal of finance, 43, 661–676. carhart, m. m. (1997). on persistence in mutual fund performance. the journal of finance, 52, 57–82. constantinides, g. m. (1990). habit formation: a resolution of the equity premium puzzle. journal of political economy, 98, 519–543. cornelli, f., goldreich, d., & ljungqvist, a. (2006). investor sentiment and pre-ipo markets. the journal of finance, 61, 1187–1216. cremers, m., & petajisto, a. (2009). how active is your fund manager? a new measure that predicts performance. review of financial studies, 22, 3329–3365. detzel, f. l., & weigand, r. a. (1998). explaining persistence in mutual fund performance. financial services review, 7, 45–55. duesenberry, j. s. (1949). income, savings, and the theory of consumer behaviour. cambridge, ma: harvard university press. fama, e. f., & french, k. r. (1988). dividend yields and expected stock returns. journal of financial economics, 22, 3–25. friesen, g. c., & sapp, t. r. (2007). mutual fund flows and investor returns: an empirical examination of fund investor timing ability. journal of banking & finance, 31, 2796–2816. gibson, r., michayluk, d., & van de venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. guillemette, m. a., & finke, m. s. (in press). changing risk tolerance. journal of financial planning. heaton, j., & lucas, d. (2001). portfolio choice in the presence of background risk. the economic journal, 110, 1–26. hoffmann, a. o., post, t., & pennings, j. m. (2013). individual investor perceptions and behavior during the financial crisis. journal of banking & finance, 37, 60–74. hong, h., scheinkman, j., & xiong, w. (2006). asset float and speculative bubbles. the journal of finance, 61, 1073–1117. ilmanen, a. (1995). time varying expected returns in international bond markets. the journal of finance, 50, 481–506. james, r. (2013). brain activity suggests planning designation helps calm investors. the journal of financial planning, 26, 52–59. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–291. lee, c. m., shleifer a., & thaler r. h. (1991). investor sentiment and the close-end fund puzzle. journal of finance, 46, 75–109. 217m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 lowry, m., & schwert, g. w. (2002). ipo market cycles: bubbles or sequential learning? the journal of finance, 57, 1171–1200. markowitz, h. (1952). portfolio selection. journal of finance, 7, 77–91. mehra, r. (2003). the equity premium: why is it a puzzle? financial analysts journal, 1, 54–69. moreschi, r. w. (2011). an analysis of the ability of individuals to predict their own risk tolerance. journal of business & economics research, 3, 39–48. pennings, j. m., & smidts, a. (2003). the shape of utility functions and organizational behavior. management science, 49, 1251–1263. poterba, j. m., & summers, l.h. (1988). mean aversion in stock prices: evidence and implications. journal of financial economics, 22, 27–59. rabin, m. (2000). risk aversion and expected-utility theory: a calibration theorem. econometrica, 68, 1281– 1292. rabin, m., & thaler, r. (2001). anomalies: risk aversion. journal of economic perspectives, 15, 219–232. schmidt, u., & traub, s. (2002). an experimental test of loss aversion. journal of risk and uncertainty, 25, 233–249. shunway, t. (1997). explaining returns with loss aversion. available at ssrn: http://ssm.com/abstract�58442. thaler, r. h., & johnson, e. j. (1990). gambling with the house money and trying to break even: the effects of prior outcomes on risky choice. management science, 36, 643–660. thaler, r. h., tversky, a., kahneman, d., & schwartz, a. (1997). the effect of myopia and loss aversion on risk taking: an experimental test. the quarterly journal of economics, 112, 647–661. tversky, a., & kahneman, d. (1992). advances in prospect-theory– cumulative representation of uncertainty. journal of risk and uncertainty, 5, 297–323. van de venter, g., michayluk, d., & davey, g. (2012). a longitudinal study of financial risk tolerance. journal of economic psychology, 33, 794–800. weil, p. (1989). the equity premium puzzle and the risk-free rate puzzle. journal of monetary economics, 24, 401–421. winchester, d. d., huston, s. j., & finke, m. s. (2011). investor prudence and the role of financial advice. journal of financial service professionals, 65, 43–51. 218 m. guillemette, d. nanigian / financial services review 23 (2014) 207–218 untitled does the source of money determine retirement investment choices? andrea anthonya, kristine beckb, inga chirac,* agolden gate university, 536 mission st, san francisco, ca 94105 bcalifornia state university northridge, 18111 nordhoff st, northridge, ca 91330 ccalifornia state university northridge, 18111 nordhoff st, northridge, ca 91330 abstract using a unique dataset of actual investment choices of oregon state university employees, we investigate how investment choices differ among (1) the optional retirement plan (orp) funded by the employer and (2) the investments in 403(b) accounts funded by employees themselves using voluntary salary reduction. we find that the level of risk associated with voluntary, salary reduction investments in 403(b) accounts is lower than the risk these same employees are currently taking in their employer funded 401(a) accounts. we also investigate whether the choice of the provider has a significant impact on the asset allocation chosen by the employees and find that participant investment choices in fidelity are riskier than the choices made by those in tiaa-cref. © 2017 academy of financial services. all rights reserved. jel classification: d14; g41 keywords: 403(b); investment choices; highly educated individuals; investment risk 1. introduction according to starr (2010), 403(b) plans, on which the non-profit and public sectors rely, have not been studied as thoroughly as 401(k) plans. in this article we explore the choices and asset allocation decisions within the public sector and focus on the defined contribution investment choices available to university employees. we consider employee decisions for two types of accounts: (1) the optional retirement plan (orp), which is an alternative to the * corresponding author. tel.: �1-818-677-4615; fax: �1-818-677-6079. e-mail address: inga.chira@csun.edu (i. chira) financial services review 26 (2017) 241–254 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. traditional defined benefit pension, and (2) voluntary investments in 403(b) accounts. unlike many of the studies that analyze the 401(k) space, the employees in our sample are highly educated and, in aggregate, wealthier than the national average. these individuals have been shown to take an active role in their retirement planning.1 our goal in this study is to explore the investment decisions public employees choose to make and whether those investment choices differ based on the source of money being invested and the provider chosen. although 403(b) accounts are studied less frequently than 401(k) accounts, there are a number of studies that explore how university employees make investment and savings decisions. unlike the previous studies that focus on the characteristics of the employees and the impact of those characteristics on the saving decision, our goal is to investigate the differences between investment behaviors based on the source of the money being invested. we examine whether individuals treat “free money” differently from “earned” money in terms of risk and vendor choice. angus, brown, kihom smith, and smith (2008) explore the investment options universities offer their employees though tiaa-cref. they conclude that the limited investment options result in reducing the risk-taking of the employees but also decreasing the diversification and as a result, reducing growth. given university employees’ increased reliance on defined contribution accounts and the responsibility that the employees must take for their retirement choices, this study is significant in a number of ways. first, we examine whether university employees treat their “free money” differently from their “earned money.” in other words, we analyze how the investment choices in the optional pension plan funded by the employer (the state) differ from the investment choices funded by employees themselves in the voluntary salary reduction 403(b) accounts. because of the design of the state’s retirement system, our sample presents a natural experiment in which we can observe and compare employees’ distinct investment choices associated with employee contributions versus employer contributions. second, we explore whether the choice of the provider has a significant impact on the asset allocation chosen by the employees. before making the individual investment decisions, university employees choose between fidelity and tiaa-cref.2 although the funds available are roughly equivalent between the two providers, the default investment elections and costs are not identical. additionally, tiaa-cref contains options not offered by fidelity (mainly annuities) that have historically been associated with the concept of investment safety. 2. literature review 2.1. investment choices and portfolio selection retirement asset allocation is a decision that could potentially cost employees hundreds of thousands of dollars over their lifetime. however, how good are individuals at making this decision? so far, the data points to a number of flaws in individual decision making regarding retirement planning. for example, benartzi (2001) shows that for well-performing stocks about 40% of discretionary contributions in 401(k) accounts are funneled into an individuals’ own company stock. based on survey findings, choi, laibson, madrian, and metrick (2005) 242 a. anthony et al. / financial services review 26 (2017) 241–254 find that people do not make 401(k) investment allocation decisions based on their riskreward profiles, the management costs of the investments, or even their own unique circumstances. on the contrary, they make their selection by looking at the past performance (especially when it comes to company stock), and in many cases they opt for the default investment option rather than a deliberate choice. as a result, the authors call for a restriction of the options available in a 401(k) account to contain highly diversified mutual funds. 2.2. concerns with self-directed retirement investment there are a number of reasons that could partially contribute to the observed suboptimal retirement participation and allocation. first, the options available to employees have increased dramatically in the last 15–20 years. the increased choices also contribute to a change in investment behavior. choi et al. (2005) show that employees will distribute their investment allocations based on the choices offered by the employer. for example, the more equity funds that are available, the higher the percentage of employee contributions flows into equity funds. second, the trend towards automatic enrollment comes with the burden on the employer to choose the default investment selection. madrian and shea (2001) argue that in the short term, a significant percentage of employees do not change the default asset allocation. the employer choice for the default investment option then becomes imperative in the retirement well-being of the employees. beshears, choi, laibson, madrian, and milkman (2015) find that those who default to non-enrollment in their 401(k) plan have a counterintuitive negative reaction to information about peer retirement savings, while those who opted into a contribution tend to increase savings in response to information about their peer’s savings rates. third, there is a strong reliance on past performance affecting investment allocation. although modern portfolio theory argues that expected returns should drive the investment decision, in reality, participants often look at the immediate past performance to decide future asset allocation. mitchell and utkus (2003) point to a number of reasons that drive the strong influence of past performance on the investment decision. for example, defined contribution providers make such information easily available, which leads to availability bias. additionally, past performance is easily accessible in the media while expected returns are difficult to find and assess. 2.3. factors that affect investment decisions despite evidence that points to suboptimal investment allocation, not everyone makes suboptimal decisions at the same rate. agnew (2006) shows that although many follow naïve diversification rules when allocating their retirement account, participants making more than $100,000 per year hold less company stock, are less likely to follow behavioral biases such as the framing heuristic, and are more likely to participate in a 401(k) compared with lower paid employees. agnew, andreson, gerlach, and szykman (2008) suggest that partially because of the differential in risk aversion and financial literacy, gender has an impact on investment allocation. as a result, their study finds that women are more likely to choose an annuity compared with men. 243a. anthony et al. / financial services review 26 (2017) 241–254 2.4. current 403(b) research most of the current research in the 403(b) space focuses on the characteristics of university employees that are more prone to contribute to retirement funds. duflo and saez (2002) study the impact of group influence on the choice of mutual fund vendor. the authors show that the peer group influences both the participation in a retirement account and also the provider or vendor that is chosen. deaves, veit, bhandari, and cheney (2007) explore the characteristics of college employees associated with the propensity to plan, utilizing a survey. the authors report that the depth of participation (pension contributions as a percentage of salary) is positively correlated to a planning mindset. they also find that demographic characteristics such as gender, marital status, age, and salary were significantly correlated as well. the propensity to plan is reported to be positively correlated with risk tolerance. the authors argue that this might be “because planners are more financially sophisticated and understanding of the fact that some risk-taking is appropriate in a preretirement portfolio.” kim and hanna (2015) find that those with both defined benefit and defined contribution plans are unrealistic about their future retirement income, and their objective inadequacy increases with age. they also find that those without plans are less realistic than those with retirement savings. households willing to take on above-average investment risk are overly optimistic about retirement adequacy. those with higher education and financial experience are more likely to have realistic estimates of retirement income, but the use of a financial planner does not have the same benefit. studies examining investment choices indicate that while rationally the investment allocation for retirement should be driven by the investment horizon and, thus, tilted towards equity for young employees, the reality is different. similar to findings by choi et al. (2004), the hypothetical allocation experiment organized by benartzi and thaler (2001) showed that university of california employees make the asset allocation choice based on the funds available in the plan menu. an earlier tiaa-cref study finds that participants also do not adjust their asset allocation with age, which may result in an overly risky portfolio. at the plan design level, angus et al. (2008) argue that the set-up of 403(b) plans may also have a negative impact on the participants’ decisions. the authors study the efficiency of tiaa-cref, the largest provider of education institutions’ 403(b) plans, to show that participants could be about 40% better off in terms of wealth when investment choices are not restricted only to tiaa-cref retirement annuities. a small portion of u.s. employees have a generous employer contribution that does not call for employee participation. our sample gives us the unique opportunity to study a plan in which the employer contributes to a separate employee retirement plan and that contribution is significant. this separate plan can be invested differently than the same employee’s 403(b) account. a house money effect has been documented by a number of studies, such as thaler and johnson (1990). the employer contribution may be viewed as the house money. employees may be more willing to take risk with the employer money than their own salary reduction contribution to the 403(b) account when there is an easy way to separate the two accounts and invest them independently of each other. our study focuses on the differences 244 a. anthony et al. / financial services review 26 (2017) 241–254 in investment decisions between the two types of accounts and the sources of funds: the “free money” and the “earned money.” we examine investment choices of university employees within the context of lopes’ (1987) sp/a theory. a psychological framework that examines decision making under uncertainty, the framework attempts to balance the feeling of security (s) with aspirational levels (a), or in our case, the goal of potential retirement. the theory predicts that investors will try to match their savings motives, such as a desire to retire, with their aspirational level. thus, the investment behavior of households will relate to their goal of a successful retirement. as such, we expect to see employees maximizing their retirement accounts when the employer is contributing while minimizing the risk taken with their voluntary, salary reduction investments. 3. methodology 3.1. the framework of the retirement system oregon state university is part of the oregon public universities’ retirement system. upon hire, academic and administrative unclassified employees choose between pers, a defined benefits plan and orp,3 an optional retirement 401(a) plan that can be chosen in lieu of pers. this is a onetime irrevocable decision. the employees are split into four tiers based on date of hire. employer contribution rules are driven by the tier. this paper is focused on the orp population only. there are two accounts associated with the orp: the employer’s and the employee’s. both accounts are funded by the employer. the employer account vests 100% after five years of service and the employee account is vested immediately. for tier i–iii employees, the employee account consists of a 6% automatic contribution by the employer into the employee account. this is not an elective employee salary reduction. in other words, regardless of employee participation, the employer contributes 6% on the behalf of the employee and that amount is vested immediately. this is what we refer to as “free money.” for tier iv employees (hired after july 1, 2014), the employee account is called employer match account and consists of 1% to 4%, based on employee contribution into the tdi 403(b) account. unlike tiers i–iii, there is no free, immediately vested money available to employees unless the employee also contributes to the 403(b). once selected into the orp, the employees have a choice between fidelity and tiaa-cref for their provider.4 in addition to the orp (or pers), the participants may also elect one of the two voluntary retirement plans, the 403(b) or the 457. the default investment options are a money market account for tiaa-cref and the age-appropriate, target date fund for fidelity. 3.2. data description we obtain investment account data from oregon state university as of december 2015. the data includes the type of account: either tdi, voluntary 403(b), or the 401(a), orp. the 245a. anthony et al. / financial services review 26 (2017) 241–254 data provides investment fund and percentage allocation by employee. the data are deidentified, but includes the date of hire, allowing us to identify the hiring tiers, and the current age of the employee. we supplement the data by looking at the individual funds and adding the asset class, expense ratio, and, where available, sharpe ratio, beta, and standard deviation of each fund in the portfolio. our main goal is to compare the risk taken with one’s own money versus the employer money. we use a number of measures of risk. first, we measure risk using beta and standard deviation. then, given data limitations of beta and standard deviation within our sample, we also measure risk using an identifier called risk tier, which we build based on the perceived risk of the asset class. specifically, we assign a risk tier from 1 to 6, where 1 is equivalent to the least risky investment, the money market fund, and 6 is associated with the riskiest investment. tier vi includes the sector specific fund for tiaa-cref and the self-directed brokerage account for fidelity, as detailed in exhibit 1. we develop risk classifications based on the weighted average of the standard deviations for the funds in the given category. we understand that these risk tier classifications are not perfectly linear but feel it is a close estimation given the data limitations.5 given the limited investment information obtained from the retirement plan providers, in addition to the data obtained from the two providers, we also conduct a survey of tier iii and iv employees regarding their allocation, perceptions about retirement accounts, and interest and comfort with making investment choices. the full survey is available upon request. our goal is to better understand how and why participants make their investment choices. the survey was emailed to all university employees; 312 participants finished the survey. we draw on some of the answers to compare self-reported data to the actual data received from the vendors. 3.3. analysis for the multivariate analysis, we use ordinary least squares regressions and apply the newey-west correction for heteroscedasticity in the residuals of each linear regression model. our model is: risk � � � �1fidelitydummy � �2tdi 403(b) � �3age� �4tier� ei (1) exhibit 1: risk tier classification risk tier description i money market accounta ii bond fund iii balanced fund (lifecycle or target date fund) iv domestic equity fund v international equity fund vi real estate (sector) fund or the self-directed brokerage accountb atraditional (minimum guarantee) annuities are assigned a 1 and variable annuities are assigned to the risk tier that corresponds to the annuity’s underlying asset class. bthe self-directed brokerage account is only available for fidelity. 246 a. anthony et al. / financial services review 26 (2017) 241–254 the dependent variable, risk, is measured as the weighted-average portfolio risk tier as specified in exhibit 1. the independent variables in the baseline models are as follows: fidelity dummy equal to 1 if the participant is enrolled in a fidelity fund and 0 otherwise, tdi 403(b) dummy equal to 1 if the participant is enrolled in the 403(b) plan and 0 otherwise, age equal to the current age of the participant as provided by the plan sponsor (fidelity or tiaa-cref), and tier equal to the current tier of the employees based on the date of hire. 4. results 4.1. summary statistics the sample consists of 3,436 fidelity observations and 5,851 tiaa-cref observations. of these, 1,689 are fidelity orp 401(a), 3,734 are tiaa-cref orp 403(b), and an additional 1,747 and 2,117 observations are from the 403(b) accounts, respectively. a total of 1,175 (or 56.93%) unique participants belong to fidelity and 889 (43.07%) to tiaacref. table 1 provides the breakdown by hiring tier. the average age of participants in the full sample is 49.11 years old, with those with fidelity at 50.36 and those with tiaa-cref at 48.37 years. at the vendor level, 56.93% participants are enrolled in fidelity plans versus 43.04% in tiaa-cref. the majority of participants in our sample are tier iii employees hired between 2003 and 2013. the recent hires (tier iv) appear to prefer tiaa-cref. at the plan level, 54.96% of observations are in the orp plan (employer contribution) and 45.04% are in the voluntary 403(b) plan. this is an intriguing number given the voluntary nature of the 403(b). given recent changes in employer contribution for tier iv employees, we expect to see more employees participating in their 403(b) plan. we find that out of 116 participants in tier iv, 60 chose both the 401(a) and the 403(b) plans, 23 participants chose only the 401(a) plan, and 33 participants chose only the 403(b) plan. although it appears that only 19% chose not to participate in both plans and thus did not maximize their employer table 1 sample description all vendors fidelity tiaa orp 401(a) tdi 403(b) tier i 407 (19.72%) 291 (24.77%) 116 (13.05%) 139 (10.37%) 309 (28.14%) tier ii 438 (21.22%) 249 (21.19%) 189 (21.26%) 307 (22.91%) 198 (18.03%) tier iii 1,102 (53.37%) 588 (50.04%) 514 (57.82%) 811 (60.52%) 497 (45.26%) tier iv 116 (5.62%) 47 (4.00%) 69 (7.76%) 83 (6.19%) 93 (8.47%) total 2,064 (100%) 1,175 (100%) 889 (100%) 1,340 (100%) 1,098 (100%) results reported are number of observations in the sample and portion of sample in parentheses. orp � optional retirement plan. 247a. anthony et al. / financial services review 26 (2017) 241–254 contribution, it is important to mention the data availability constraints that may underrepresent the true number of dual plan enrollments.6 next, we examine employee investment choices. given that the two vendors have different default investments (target data fund for fidelity and money market account for tiaa-cref), we present the data separately by vendor. table 2 displays these results. on average, participants distribute their investments across 3.90 funds. the number of funds in tiaa-cref accounts is significantly higher, at 5.49, versus the number of funds in fidelity accounts at 2.69. to gain insight into the differences we look at the distribution by provider. we find that 198 of the 889 tiaa-cref participants (22.27%) allocate their retirement funds to 10 or more funds (maximum is 25). by comparison, only 27 fidelity participants (2.29%) allocated their funds to 10 or more funds (maximum is 17). the difference in the number of chosen funds is significant, especially given that the way the funds are presented to participants on the enrollment form is similar between the two providers. all elections are done online and each participant assigns a percentage of the account to different funds. to understand this difference better, we looked at the survey that several participants completed. an interesting find was that participants who are less knowledgeable and confident in their investment skills tend to choose tiaa-cref over fidelity. some participants explained that they associated tiaa-cref with a “safer option” and “less risk.” this is an interesting marginal finding that shows the difference in employee perception about the choices available. while the gross expenses vary from 0.05% to 1.23% for fidelity and from 0.06% to 0.87% for tiaa-cref (excluding traditional annuity), the mean expense ratios are statistically significantly different at 0.48% and 0.40%, respectively. standard deviation, beta, and risk tier are all significantly different between the groups. mean standard deviations are 10.08 and 11.59, betas are 0.93 and 1.05, and average risk tier investments are 3.58 and 3.82, respectively, for fidelity and tiaa-cref. the average age of our sample is 49.11. fidelity participants’ age averages 50.36 and tiaa-cref is 48.37, a statistically significant difference of two years. in this section we investigate whether the risk associated with tdi(a) investments differs from the risk in the orp 403(b) investment choices. we proxy for risk using the weighted average of the risk tiers of each fund in the individual’s portfolio. alternatively, we also explore the weighted average of the standard deviations, provided by the vendor, of each fund in the individual’s portfolio as a secondary risk measure. table 2 investments choices by vendor full sample mean (n) fidelity sample mean (n) tiaa-cref sample mean (n) number of funds 3.90 (2,064) 2.69 (1,175) 5.49 (889) gross expense 0.43% (9,286) 0.48% (3,435) 0.40% (5,851) beta .9831 (5,407) 0.9297 (2,997) 1.0495 (2,410) sd 10.72 (5,627) 10.08 (3,217) 11.59 (2,410) sharpe ratio 0.9359 (5,627) 0.9246 (3,217) 0.9510 (2,410) risk tier 3.7 (9,287) 3.6 (3,436) 3.8 (5,851) results reported are the mean value for each variable with number of observations in parentheses. orp � optional retirement plan. 248 a. anthony et al. / financial services review 26 (2017) 241–254 table 3 presents the multivariate regression results based on investment choice and the type of account. we use the actual portfolio holdings in each fund to calculate the appropriate weighted average. individuals who have missing observations for our proxies of risk are removed from the sample. if an individual has both an orp account and a tdi account they will appear in the sample twice. thus, the sample number of observations for the multivariate regressions is slightly larger than the number of unique individuals. there were 583 unique participants who are enrolled in the 401(a) plan and 567 in the 403(b) plan. the average age in the full sample is 49.9, with the age in the 403(b) slightly higher at 52.71 years than the 401(a) plan at 47.16 years. in support of the theory, we find that employees treat their 401(a) accounts differently from the 403(b) accounts. the tdi 403(b) dummy is negative and statistically significant. the level of risk associated with voluntary, salary reduction investments in 403(b) accounts is lower than the risk employees are currently taking in their employer funded 401(a) accounts. the results hold irrespective of the measure of risk (using risk tier and standard deviation). these findings are consistent with the notion of a house money effect. employees appear be more willing to take risk with the employer money, or “free money,” than their own salary reduction contribution to the 403(b) account. we also examine a special subpopulation consisting of the 374 individuals in the sample who have both an orp account and a tdi account. for those people, the average risk tier for the orp is 3.44 with a 95% confidence interval between 3.34 and 3.53. the average risk tier for the tdi account is 3.26 with a 95% confidence interval between 3.14 and 3.37. we find a statistically significant difference in risk of 0.177 between the groups. forty-six percent of the individuals have no difference in risk tiers between orp and tdi, 34% have a difference of less than 1 between the two types, 15% of individuals have a difference between 1 and 3, and 5% have a difference of greater than 3. we also observe that age is statistically and negatively related to risk. older employees are more conservative in their investment choices, which is expected given their proximity to retirement. to further test the house money effect, we add a dummy in table 1 for the non-vested employer accounts. our hypothesis that employees take more risk with other people’s money table 3 the relationship between type of account and risk model 1 model 2 model 3 variables combined orp 401(a) tdi 403(b) fidelity dummy 0.236*** (4.77) 0.182*** (2.70) 0.315*** (4.28) tdi 403(b) dummy �0.233*** (�4.64) age �0.00504* (�1.94) �0.00309 (�0.87) �0.00782** (�2.07) tier �0.122*** (�3.60) �0.153*** (�2.97) �0.103** (�2.29) constant 3.874*** (20.05) 3.891*** (14.33) 3.697*** (13.25) observations 2,435 1,340 1,095 r2 0.023 0.014 0.024 the dependent variable is the weighted-average portfolio risk tier. the independent variables are a fidelity dummy equal to 1 if the participant is enrolled in a fidelity fund and 0 otherwise, a tdi 403(b) dummy equal to 1 if the participant is enrolled in the 403(b) plan and 0 otherwise, age equal to the current age of the participant as provided by the plan sponsor, and tier equal to the current tier of the employees based on the date of hire. t-statistics are reported in parentheses. ***, **, * indicate statistical significance at the 1%, 5%, and 10% level, respectively. 249a. anthony et al. / financial services review 26 (2017) 241–254 (free employer money) can be further tested by treating the unvested portion of the account as money belonging to someone else. however, we do not find this to be the case. the difference in the average risk of the account preand post-vesting does not vary significantly. the second objective is to examine investment choices and risk associated with the two vendors. the difference in investment choices between the two providers is shown in table 4. the coefficient for the fidelity dummy is positively and significantly related to the risk tier. it appears that participant investment choices in fidelity are riskier than the choices in tiaa-cref. when looking at the data by provider, we find that the riskiness of investment choices is consistently lower for the 403(b) self-directed account as compared with the employer funded 401(a) account. one of the main limitations of our study is the lack of additional control variables in the multivariate analysis. because of limited data availability, we were not able to control for additional demographic characteristics. to help to alleviate this concern and to get a better understanding of the differences in choices between the two providers, we expand our analysis and look at the characteristics of the two vendor investment choices. table 5 presents the summary results at the vendor level. we first present a general overview of investment choices in the fidelity account in panel a, followed by the choices in the tiaa-cref account in panel b. there were 11.97% of fidelity participants who are enrolled in the target date fund, a balanced fund. another 584 people, or 18.10%, have some participation in the target date fund. a very small percentage, at 0.52%, transferred their funds to fidelity’s self-service brokerage account (that allows enrollment in a large number of mutual funds). there are 1.60% of participants who have their entire investment in a money market account, an active choice made by the participant. the average age of the people in this category is 54.56. finally, 5.52% invested their funds to a certain degree in the s&p 500 etf, the cheapest fund on the menu at a gross expense of 0.05%, while 7.37% invested their money in one of the five funds with a gross expense ratio above 1.00%. it is important to add that the fidelity table 4 the relationship between provider and risk model 1 model 2 model 3 variables combined (risk tier) fidelity (risk tier) tiaa-cref (risk tier) fidelity dummy 0.236*** (4.77) tdi 403(b) dummy �0.233*** (�4.64) �0.304*** (�4.16) �0.161** (�2.33) age �0.00504* (�1.94) �0.00642 (�1.64) �0.00347 (�1.00) tier �0.122*** (�3.60) �0.224*** (�4.42) �0.0411 (�0.90) constant 3.874*** (20.05) 4.239*** (14.99) 3.804*** (14.87) observations 2,435 1,121 1,314 r2 0.023 0.031 0.006 the dependent variable is the weighted-average portfolio risk tier in models 4–6. the independent variables are a fidelity dummy equal to 1 if the participant is enrolled in a fidelity fund and 0 otherwise, a tdi 403(b) dummy equal to 1 if the participant is enrolled in the 403(b) plan and 0 otherwise, age equal to the current age of the participant as provided by the plan sponsor, and tier equal to the current tier of the employees based on the date of hire. t-statistics are reported in parentheses. ***, **, * indicate statistical significance at the 1%, 5%, and 10% level, respectively. 250 a. anthony et al. / financial services review 26 (2017) 241–254 choices include comparable vanguard midcap and small-cap funds, both available for a gross expense of 0.08%. unlike fidelity, the number of participants enrolled in tiaa-cref’s target date fund (lifecycle fund) is much lower. only 1.20% enrolled 100% into this option and a total of 2.73% have the option as part of their portfolio. the difference is not surprising given that fidelity’s default investment option is the target date fund, while the tiaa-cref default investment option is not. participation in the traditional annuity option is only 0.67% for the full enrollment and 8.08% for partial enrollment. by comparison, 3.24% of employees have chosen a variable annuity option and 44.97% chose the partial variable annuity option. cost-wise, 6.81% are enrolled in the cheapest fund, which is the cref international equity index with a gross expense ratio of 0.06%. by comparison, 8.83% are enrolled in the tiaa real estate sector fund, which is the most expensive. in cooperation with administrators at oregon state university, we conducted a web-based survey of oregon state university employees during the summer of 2015. we sent a survey link via email in june 2015 to all oregon state university employees that made a retirement plan choice. we removed respondents that were not eligible for the 403(b) plan and removed those enrolled before tier iii started. there were 339 respondents who completed the survey. the survey asked about a variety of issues related to retirement plan saving and investment, including questions about the participant’s plans for retirement, methods of preparation for retirement, expected job tenure, reasons for enrolling in specific retirement plans, knowledge of how the osu retirement plans worked, financial literacy, confidence in financial ability, risk attitudes, and basic demographic information. the survey questions are available upon table 5 investment choice participation by provider panel a: investment choice participation by: fidelity fidelity n number of participants 100% invested in target date fund 388 (11.97%) number of participants 100% invested in brokerage account 17 (0.52%) number of participants 100% invested mm account (average age 54.56) 52 (1.60%) number of participants invested to some degree in cheapest funda 179 (5.52%) number of participants invested to some degree in most expensive fundsb 239 (7.37%) as&p 500 etf. bbaron growth 1.23%, vi small cap 1.20%, jpm mid cap 1.141%, amg mid cap 1.04% jpm. panel b: investment choice participation by: tiaa-cref tiaa-cref n number of participants 100% invested in target date fund 64 (1.20%) number of participants 100% invested in traditional annuity 36 (0.67%) number of participants 100% invested in variable annuity 172 (3.24%) number of participants 100% mm account (average age ) 17 in mm; 183 vamm (3.76%) number of participants invested to some degree in cheapest funda 362 (6.81%) number of participants invested to some degree in most expensive fundb 469 (8.83%) acref intl eq idx-inst-0.06%. btiaa real estate-0.87%. 251a. anthony et al. / financial services review 26 (2017) 241–254 request. this was a blind survey in that we could not merge the survey responses with the actual investment data found in the earlier part of the paper. the average age of the participants in the sample is 42. forty-two percent of respondents enrolled in the orp plan, 48% enrolled in the defined benefit (pension) plan, and 10% of the respondents are not sure which plan they selected. there were 31% of the respondents who enrolled in the optional 403(b) plan, while 22% are not sure if they enrolled in an optional plan; 53% of respondents use fidelity, 33% used tiaa-cref, 2% use valic, and 12% are not sure which vendor they used for their orp accounts. for the 403(b), 61% of respondents use fidelity, 32% used tiaa-cref, and 7% are not sure which vendor they chose. participants are asked to rate the level of confidence in their investment skills on a scale from 0 to 10. the mean confidence level for the sample is 4.19. the mean confidence level for those selecting a 403(b) plan is 5.21 whereas those without the optional plan selected an average mean confidence level of 4.01. the investment funds selected in their 403(b) accounts are a result of an active choice made by the participant for 72% of the participants, while 28% of participants simply used the default investment settings for each vendor. using multivariate regressions, we find that the choice to enroll in the defined contribution plan (orp) rather than the pension plan is associated with respondents who: (1) figured out how much they need for retirement, (2) are younger, (3) have a higher household income, (4) are the primary decision makers, and (5) are more confident about their investment skills. enrollment into the optional 403(b) plan is associated with respondents who: (1) figured out how much they need for retirement, (2) have a higher household income, (3) are tier iv employees, (4) are more financially knowledgeable, and (5) see investing as an exciting activity. 5. conclusion and implications using a unique dataset of actual investment choices of oregon state university employees, we investigate how investment choices differ between (1) the orp funded by the employer and (2) the investments in 403(b) accounts funded by employees themselves using voluntary salary reduction. we find that the level of risk associated with voluntary, salary reduction investments in 403(b) accounts is lower than the risk employees are currently taking in their employer funded 401(a) accounts. these findings are consistent with the notion of a house money effect, with the employer contribution as the house money. employees appear be more willing to take risk with the employer money, or “free money,” than their own salary reduction contribution to the 403(b) account. we also examine whether the choice of the provider, fidelity or tiaa-cref, is related to the asset allocation choices made by the employees. we find that participant investment choices in fidelity are riskier than the choices made by those who selected tiaa-cref. one possible interpretation and implication is that participants who enroll into fidelity are more sophisticated investors and are better able to assess risk. this result is further reinforced by the answers to our survey. respondents who choose fidelity also tend to be more confident in their investment ability and knowledgeable about the stock market. 252 a. anthony et al. / financial services review 26 (2017) 241–254 lastly, using a survey of 354 oregon state university employees we explore how individuals choose their plan, the level of risk, and the plan provider. there were 28% of those in the survey who chose the default investment option by the provider. enrollment into the optional plan is associated with respondents who are more likely to have figured out how much they need for retirement, have a higher household income, are recently hired (tier iv) employees, are more financially knowledgeable, and see investing as an exciting activity. the study and findings have implications for both the design and choices employers make in their retirement plans and the choices employees make when selecting investments. first, the employer’s choice of a default option has a considerable impact on the investment choice made by the employee. as such, employers should select the default option that has the best potential to be an appropriate option for employees in the event of no involvement by the employee. second, employers should add educational information around the employee and employer accounts during employee orientation when the retirement system is presented. this will assure that employees are clear on the types of investment choices available and the implication of an active or passive choice. lastly, given the financial implication of the decision, in the event that employees do not feel comfortable with making the investment choices, employees should employ professional services rather than spreading funds equally between similar accounts or investing in high fee funds that have low fee comparable options. notes 1 as education and income have been shown to have a strong relationship with retirement planning, we aim to eliminate that impact rather than control for it. for example, joo and grable (2005) show a link between higher education/higher income and the existence of workplace retirement savings. 2 employees are informed about the two options at a human resources-led orientation but they are not advised to choose one over the other. 3 for a complete guide of eligibility and rules, see the state of oregon public employees’ website. 4 valic is also an option but it is a closed provider since 2007 and no valic data is included in this study. 5 one concern is that the tiaa real estate fund is less risky than equity funds. to account for this possibility we specify an alternative ranking, where real estate was assigned a 4, domestic equity a 5, and international equity a 6, with the brokerage account becoming a 7. this change does not have a significant impact on the results. 6 it is possible that some employees chose fidelity for one plan and tiaa-cref for the other. given the lack of employee identifiers, we are unable to crosscheck the two plans and identify such participants. we treat each unique participant from either fidelity or tiaa-cref as non-overlapping, when in reality this may not be so. additionally, participants could choose the 403(b) account and not be enrolled in the tdi(a). rather, they may decide to choose the defined benefit pension alternative. 253a. anthony et al. / financial services review 26 (2017) 241–254 given the data availability, we expect the real number of tier iv employees who do not maximize their benefits to be lower. references agnew, j. (2006). do behavioral biases vary across individuals? evidence from individual level 401(k) data. journal of financial and quantitative analysis, 41, 939–962. agnew, j., andreson, l., gerlach, j., & szykman, l. (2008). why choose annuities? an experimental investigation of the role of gender, faming and defaults. the american economics review, 98, 418–422. angus, j., brown, w., smith, k., smith, j. (2008). what’s in your 403(b)? academic retirement plans and the costs of underdiversification. financial management, 36, 1–38. benartzi, s. (2001). excessive extrapolation and the allocation of 401(k) accounts to company stock. the journal of finance, 56, 1747–1764. benartzi, s., & thaler, r. (2001). naive diversification strategies in defined contribution saving plans. the american economic review, 91, 79–98. beshears, j., choi, j. j., laibson, d., madrian, b. c., & milkman, k. l. (2015). the effect of providing peer information on retirement savings decisions. the journal of finance, 70, 1161–1201. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2005). saving for retirement on the path of least resistance. rodney l white center for financial research – working paper 9. deaves, r., veit, t., bhandari, g., & cheney, j. (2007). the savings and investment decision of planners: a cross-sectional study of college employees. financial services review, 16, 117–133. duflo, e., & saez, e. (2002). participation and investment decisions in a retirement plan: the influence of colleagues’ choices. journal of public economics, 85, 121–148. joo, s., & grable, j. (2005). employee education and the likelihood of having a retirement savings program. journal of financial counseling and planning, 16, 37–49. kim, k. t., & hanna, s. d. (2015). do us households perceive their retirement preparedness realistically? financial services review, 24, 139–155. lopes, l. l. (1987). between hope and fear: the psychology of risk, advances in experimental social psychology, 20, 255–295. madrian, b., & shea, d. (2001). the power of suggestion: inertia in 401(k) participation and savings behavior. the quarterly journal of economics, 116, 1149–1187. mitchell, o. s., & utkus, s. p. (2003). lessons from behavioral finance for retirement plan design. prc working paper no. 2003–6. starr, b. (2010). willingness to pay: understanding 403(b) fees. financial services review, 19, 129–144. thaler, r., & johnson, e. (1990). gambling with the house money and trying to break even: the effects of prior outcomes on risky choice. management science, 36, 643–660. 254 a. anthony et al. / financial services review 26 (2017) 241–254 the grable and lytton risk-tolerance scale: a 15-year retrospective stephen kuzniaka,*, abed rabbania, wookjae heoa, jorge ruiz-menjivara, john e. grablea adepartment of financial planning, housing and consumer economics, university of georgia, athens, ga 30605, usa abstract over a decade ago, grable and lytton (1999) developed, tested, and published a financial risk-tolerance scale in financial services review that has since been widely used by consumers, financial advisers, and researchers to evaluate a person’s willingness to engage in a risky financial behavior. analysis of data (n � 160,279) spanning the timeframe 2007 to 2013 provides evidence that the risk-tolerance scale’s reliability and validity have remained robust since the scale was first developed. the scale’s estimated cronbach’s � was 0.77 during this time period. consistent with the literature, high scale scores (representing a greater willingness to take risks) were found to be associated with equity ownership and negatively related to cash and bond holdings. © 2015 academy of financial services. all rights reserved. jel classification: d. microeconomics: d81 criteria for decision-making under risk & uncertainty keywords: risk; risk tolerance; risk scale; risk scoring; risk assessment 1. introduction in 1999, grable and lytton published an article in financial services review that presented a 13-item financial risk-tolerance scale that has, according to google scholar analytics (google, 2014), been referenced in hundreds of research publications. the scale, which is available online through rutgers new jersey agricultural experiment station * corresponding author. tel.: �1-706-425-2964; fax: �1-706-542-4397. e-mail address: kuzniak@uga.edu (s. kuzniak). financial services review 24 (2015) 177–192 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. (https://njaes.rutgers.edu/money/riskquiz/), has been used by over 200,000 consumers, educators, and researchers. when the grable and lytton (g&l) risk scale was first published, there were few publically available measures of financial risk tolerance. those that did exist tended to be based on income gambles, choice dilemmas, or demographically driven heuristics. the grable and lytton article was among the first to provide published risk scale reliability and validity estimates. since first being published, the scale has been adopted as a client data intake instrument by a variety of firms operating in the financial and investment planning domain. financial advisers often use the measure when providing comprehensive planning services as a way to measure and understand their clients’ risk attitudes before allocating client assets. for individuals, the risk scale is often used to understand their own willingness to take financial risk and analyze investment preferences. the purpose of this article is to provide a 15-year anniversary review of the g&l scale. specifically, this article provides readers with data regarding historical scale response patterns and reliability and validity estimates. as will be shown below, the scale has held up relatively well as a consumer tool and research instrument since first being introduced to the public. this article adds additional evidence that the scale, whereas certainly not perfect, does provide financial planning practitioners and researchers with an acceptable, valid, and reliable assessment of a person’s willingness to take financial risk. 2. background review 2.1. development of the scale when grable and lytton (1999) set out to measure financial risk tolerance they were originally faced with the challenge of finding questions that (1) were germane to the concept of risk, (2) would allow anyone to combine question answers into a risk scale, (3) were relevant to situations faced by typical consumers making financial decisions, (4) were easy to administer, and (5) offered both validity and reliability when combined into a scale. grable and lytton used guidance provided by maccrimmon and wehrung (1986) to help identify and develop appropriate questions. these requirements included ensuring that (1) the multidimensionality of risk tolerance was assessed through the inclusion of simple and complex situational items, (2) the items were consistent and not redundant, (3) the items were interesting to answer, and (4) completion times would be reasonably short. their efforts at building a financial risk-tolerance assessment tool were based primarily on scale development theory and propositions found in modern portfolio theory (mpt). in markowitz’s 1952 article describing the basis of mpt, the theoretical relationship between risk and investment returns was clearly outlined. markowitz noted that risk and return are positively related, and as such, investors who demand a higher return must be willing to accept a higher level of risk (i.e., volatility) in their portfolios. this insight has since been used as a key benchmark of validity whenever a risk-assessment tool has been created. grable and lytton (1999) noted that, as such, any new and useful risk-tolerance measure must align with the prediction that high scores will correspond with a general willingness to take more financial risk. in the context of a financial risk-tolerance scale, risk scores should 178 s. kuzniak et al. / financial services review 24 (2015) 177–192 be positively associated with, say, equity ownership. in addition to this baseline measure of validity, a scale ought to exhibit strong psychometric characteristics. grable and lytton’s (1999) efforts at establishing the reliability and validity of a new scale started by selecting over 100 risk-assessment items from the literature. based on pilot study data, they were able to identify 50 items that matched all of the screening criteria. grable and lytton (1999) used these 50 items to begin the development of a risk-tolerance questionnaire. using traditional item-response procedures, grable and lytton culled the list of items to 20 risk questions. they then grouped items into one of eight categories: (1) guaranteed versus probable gambles, (2) general risk choice, (3) choice between sure loss and sure gain, (4) risk as experience and knowledge, (5) risk as a level of comfort, (6) speculative risk, (7) prospect theory, and (8) investment risk. these efforts were taken to ensure that, at a minimum, the new scale would provide high face validity for practitioners and researchers. that is, their review of the literature indicated that a person’s risk attitude was most closely associated with these eight domains. factor analysis procedures were then used to evaluate data from a convenience sample to obtain a more parsimonious number of items. grable and lytton (1999) were able to reduce the number of items to 13. the final version of the scale was found to represent three factors: (1) investment risk, (2) risk comfort and experience, and (3) speculative risk. scale reliability was measured using cronbach’s �. grable and lytton reported an initial � � 0.75. as noted by cortina (1993) and peterson (1994), this level of reliability matched what is typically found in psychological and marketing studies.1 grable and lytton (1999) took additional steps to measure the scale’s construct validity, which is defined as the extent to which a measure actually assesses its intended purpose. they were able to correlate scores on the 13-item scale to responses to the well-known survey of consumer finances (scf) risk assessment item. the scf item has been used extensively in the literature as a proxy measure of consumer risk attitudes (yao, hanna, and lindamood, 2004). the scf asks: which of the following statements on this page comes closest to the amount of financial risk that you are willing to take when you save or make investments? 1. take substantial financial risk expecting to earn substantial returns 2. take above average financial risks expecting to earn above average returns 3. take average financial risks expecting to earn average returns 4. not willing to take any financial risks the two items were found to be positively correlated (r � 0.54). this was the first reported validity estimate of the new scale. grable and schumm (2010), using a different sample, also conducted a construct validity test using the scf item. they noted, similar to grable and lytton (1999), that the scale was positively and statistically significantly related to the scf item. they were also able to estimate the relative reliability of both the 13-item scale and the scf item. while the cronbach’s � for the scale remained relatively constant, grable and schumm noted that the estimated reliability of the scf item was most likely between � � 0.52 and � � 0.59. as such, they concluded that practitioners and researchers who were interested in obtaining a more robust measure of someone’s willingness to engage 179s. kuzniak et al. / financial services review 24 (2015) 177–192 in a risky financial behavior, and had the space constraints to do so, would be better served using the larger 13-item scale. grable and lytton (1999) concluded their original article by encouraging other researchers to continue to test the scale with diverse audiences. they asserted that their hope was that further research using the scale would lead to a better understanding of risk tolerance, which they defined as a person’s willingness to engage in financial behavior when the outcomes are not known. they noted that with further tests, users of the scale would obtain more confidence in the validity and reliability of the scale. 2.2. further tests of the scale four years later, grable and lytton (2003) revisited the instrument to test the scale’s concurrent validity. concurrent validity refers to how well a scale corresponds with actual behavior. in theory, a financial risk-tolerance scale should exhibit a statistically significant correlation with financial behavior, such as investing. they were able to document that scale scores were positively associated with equity ownership and negatively related to fixedincome and cash ownership. this finding held true in both bivariate and multivariate analyses, controlling for age, gender, marital status, education, income, and other factors. their work helped to support the validity of the original scale. yang (2004) conducted a reliability and validity test of the scale using a college student and adult sample. as expected, she noted that younger respondents scored differently than older respondents, but the differences were not consistent. younger respondents were less averse to investing in hard assets, whereas older respondents were risk seeking in relation to stocks and bonds. yang did note, however, that overall scale scores were not significantly different based on the age of respondents. additionally, both the younger and older samples generated cronbach’s � scores greater than � � 0.70. although yang provided suggestions for new items and refinement to existing questions, her overall conclusion was that the scale worked reasonably well with both younger and older respondents. gilliam, chatterjee, and grable (2010) also conducted a concurrent validity test of the g&l scale. they correlated the scale against responses to the scf risk item. similar to grable and schumm (2010), they reported a statistically significant correlation (r � 0.60). additionally, gilliam and his associates noted that the g&l scale was positively associated with the ownership of risky investment assets. overall, they concluded the scale provides an acceptable indication of a person’s willingness to take on investment risks and that the scale does a better job of assessing financial risk tolerance than a single item measure such as the scf item. 2.3. summary currently, there are a limited number of peer-reviewed risk-tolerance assessments available in the public domain. some instruments and scales are new and lacking historical reliability and validity data (e.g., carr, 2014). other scales were developed primarily for research interests (e.g., grable, 2004; grable and joo, 2001). still other items, instruments, and scales tend to measure financial risk tolerance indirectly through income gambles (e.g., 180 s. kuzniak et al. / financial services review 24 (2015) 177–192 barsky, juster, kimball, and shapiro, 1997; hanna and lindamood, 2004) or other forms of risk taking (e.g., weber, blais, and betz, 2002). the g&l scale is one of the only peer-reviewed public—no cost—assessment tools available to consumers, practitioners, and researchers. since its introduction in 1999, more than 200,000 consumers have used the scale to evaluate their tolerance for financial risk. a question of interest for those who use the scale is whether the instrument’s original psychometric properties have changed since the scale was first published. the remainder of this article provides information to help answer this question. evidence of the scale’s validity and reliability, based on a multiyear data collection process, is presented below. 3. methodology 3.1. sample data for this project were obtained from a multiyear proprietary data collection project sponsored by rutgers new jersey agricultural experiment station. for nearly 10 years, rutgers university has hosted a free web-based site that allows anyone with internet access to answer the g&l risk scale items. the system provides a risk score and a basic review of how the score can be used in practice by consumers. response data from over 160,000 individuals, beginning in late 2007 and ending in 2013, were incorporated into this study. basic demographic data regarding the sample are provided in table 1. in general, the sample was diverse, and in many ways, unique in its coverage of different gender, marital status, education, income, and age cohorts. 3.2. the survey appendix a shows the 13 questions asked online. the survey instrument can be accessed at: njaes.rutgers.edu/money/riskquiz/. scores on the scale can range from 13 to 47. higher scores are descriptive of increased financial risk tolerance. the mean score, among the 160,279 respondents, was 27.53 (sd � 5.48). the reliability of the scale, as measured with cronbach’s �, was � � 0.77. fig. 1 shows the distribution of risk scores across the sample. 3.3. statistical approach as discussed above, the purpose of this article was multifaceted. the first purpose was to present descriptive response data for the g&l risk scale. the second purpose was to evaluate the scale’s overall reliability. as reported earlier, the scale’s reliability (i.e., cronbach’s �), using the full sample, was � � 0.77. table 1 provides more nuanced reliability estimates based on demographic categories. the final purpose was to provide additional evidence of the scale’s validity. correlation and regression procedures were used to help support previous assertions regarding the scale’s criterion-related validity. results from these tests are reported below. 181s. kuzniak et al. / financial services review 24 (2015) 177–192 4. results 4.1. sample characteristics as shown in table 1, the sample was over-represented by male respondents; however, this was not surprising given the general tendency of men to exhibit more intense investing table 1 descriptive statistics for respondents by characteristic variable respondent characteristic scale data frequency percent mean sd cronbach’s � risk-tolerance score 27.53 5.48 .77 gender female 66,996 41.8% 25.94 4.95 .73 male 91,383 57.7% 28.70 4.54 .77 age under 25 85,380 53.9% 27.35 5.53 .77 25 to 34 38,398 17.9% 27.94 5.38 .77 35 to 44 14,300 9.0% 28.25 5.39 .78 45 to 54 13,691 8.6% 27.75 5.27 .78 55 to 64 11,654 7.4% 27.02 5.09 .77 65 to 74 3,818 2.4% 26.59 5.20 .78 75 and older 1,190 0.8% 27.56 8.38 .90 marital status never married 36,545 23.2% 27.49 5.51 .77 living with significant other 21,734 13.8% 27.54 5.42 .77 married 26,954 17.1% 27.69 5.30 .78 widowed 36,133 22.9% 26.94 6.62 .84 shared living arrangement 25,357 16.1% 27.82 6.34 .82 education some high school or less 36,545 23.2% 27.28 5.86 .78 high school diploma 21,734 13.8% 27.15 5.51 .76 some college 26,954 17.1% 26.88 5.21 .76 associate’s degree 10,751 6.8% 26.80 5.27 .77 bachelor’s degree 36,133 22.9% 28.10 5.20 .77 graduate or professional degree 25,357 16.1% 28.43 5.39 .78 household income less than $25,000 35,531 22.9% 27.08 5.58 .77 $25,000 to $49,999 30,441 19.6% 26.59 5.35 .76 $50,000 to $74,999 30,135 19.4% 27.28 5.32 .76 $75,000 to $99,999 20,644 13.3% 27.71 5.27 .76 $100,000 or more 38,597 24.8% 28.65 5.53 .78 decision making make own investment decisions 92,803 57.9% 27.87 5.53 .78 rely on the advice of professional 18,387 11.5% 27.77 5.12 .75 do not have investment assets 46,157 28.8% 26.75 5.40 .76 seasonal effects summer 19,239 12.0% 27.72 5.39 .77 fall 47,656 29.7% 27.61 5.50 .77 winter 49,449 30.9% 27.26 5.39 .76 spring 43,935 27.4% 27.67 5.53 .77 182 s. kuzniak et al. / financial services review 24 (2015) 177–192 behavior. the age profile of respondents was skewed towards those under age 25. even so, other age groups were also widely represented. the dataset included a diverse representation of marital status. in terms of educational profile, the sample was fairly representative of the population, with a slight tilt towards those who had completed some form of college education. given who is typically interested in financial planning and investing topics, this educational characteristic was not unexpected. household income patterns showed that respondents tended to cluster into low and high income categories. table 1 also provides data related to financial decision making as reported by respondents. the majority of respondents (57.9%) indicated making their own investment decisions. nearly 30% of those responding indicated having no investment assets at the current time. the remainder reported that they relied on the advice of another person, such as a stock broker or financial planner, when making investment decisions. these data are important in helping establish a profile of the type of person who may be seeking information about their tolerance for risk. finally, table 1 shows seasonal patterns of response. as expected, data collection was lowest during the summer months. this was likely because of fewer college age people completing the survey and the tendency among investors to postpone investment decision making during the summer. seasonally, use of the online survey in fall, winter, and spring was similar. 4.2. response patterns based on t and analysis of variance (anova) tests, nearly all of the risk scores were statistically significantly different across characteristic categories. it is important to note, however, that much of the statistical significance was likely because of the large sample size. as such, within-sample random sampling procedures were used to confirm results. the random sample was chosen using a sampling protocol in spss 22.0. the spss random fig. 1. distribution of risk-tolerance scores across the sample. 183s. kuzniak et al. / financial services review 24 (2015) 177–192 sampling algorithm was based on equal probability estimates. in general, the demographic profile of those in the random sample matched that of the full dataset. some interesting significant differences were noted when random samples were used. for instance, the gender difference was meaningfully significant, t(15,767) � 31.57, p � 0.001. men, as has been reported in the literature (e.g., arano, parker, and terry, 2010; grable, 2008; neelakantan, 2010), were more risk tolerant than women. a difference in risk tolerance scores across income categories was also noted. as shown in fig. 2, a curvilinear (u-shaped) relationship was observed, � � 0.12, t(15,574) � 15.17, p � 0.001. a similar curvilinear effect was found for education, � � 0.09, t(15,687) � 11.87, p � 0.001 (fig. 3). 4.3. reliability estimates table 1 also provides information about the reliability of the g&l risk scale. overall, the scale’s cronbach’s �, as a measure of scale reliability, was � � 0.77. this estimate is higher than that reported by yang (2004) but in line with what grable and lytton (1999) originally reported. additionally, the scale’s reliability estimate falls squarely in the mean average for similar psychologically-based measures (peterson, 1994). the sixth column of table 1 provides reliability estimates for each respondent characteristic. for example, when the analysis was delimited to include only women, the scale’s � was � � 0.73, whereas for men � was � � 0.77. the highest reliability estimate was noted for those age 75 or older (� � 0.90). overall, the scale appears to be most reliable for: (1) males, (2) older respondents, (3) those who are married or have been previously married, and (4) those who make their own financial and investment decisions. when viewed holistically, this profile matches the description of many investors today. it is important to note, as well, that the variation in reliability estimates was quite small across the respondent characteristics. this supports the notion that the scale provides users with a relatively consistent level fig. 2. g&l risk scores by household income. 184 s. kuzniak et al. / financial services review 24 (2015) 177–192 of response measurement across gender, age, marital status, education, income, and decision making characteristics. 4.4. validity estimates the scale’s validity was evaluated using a combination of correlation and regression tests. the first criterion-related validity test (i.e., an evaluation of the relationship between scale scores and an anticipated outcome or behavior) results are shown in table 2. respondents were asked to think about their current financial situation and to indicate, “approximately what percentage of your personal and retirement savings and investments are in the following categories: (1) cash, such as savings accounts, cds, or money market mutual funds; (2) fixed income investments, such as corporate bonds, government bonds, or bond mutual funds; (3) equities, such as stocks, stock mutual funds, direct business ownership or investment real estate (not your personal residence); and (4) other, such as gold or collectibles. responses to these four categorical assessments were summed. scores ranged from zero to 100%. as shown in table 2, risk scores were negatively associated with cash holdings and positively related to equity ownership. this matched the relationship between risk tolerance and portfolio composition predicted in mpt and the capital asset pricing model (hariharan, chapman, and domian, 2000). the relationship between risk scores and bonds was almost fig. 3. g&l risk scores by education. table 2 correlation of risk score with investment allocation (n � 160,279) risk score % cash % bonds % equities % other risk score 1.00 % cash �0.26 1.00 % bonds �0.01 �0.53 1.00 % equities 0.27 �0.77 0.03 1.00 % other 0.13 �0.33 0.05 �0.10 1.00 all coefficients significant at p � 0.001. 185s. kuzniak et al. / financial services review 24 (2015) 177–192 zero; however, when cash and fixed-income holdings were summed and correlated to risk scores, the relationship was negative (r � �0.31, p � 0.001). holding other assets, such as gold or collectibles, was also found to be positively associated with risk scores. these results mirrored those from grable and lytton (2003). these findings add support to the relative power of the g&l risk scale to explain investment asset holdings at the household level. although the effect size of the associations reported in table 2 were not large, the relationships were as expected. further, the strength of associations reflects the notion that financial risk tolerance is only one input into investment allocation decisions. other factors, including financial capacity and a person’s general socioeconomic profile, also play an important role in shaping investment decisions. based on this concept, a second criterionrelated validity test was undertaken. in this case, it was hypothesized that g&l risk scores should be positively associated with equity ownership, holding gender, age, marital status, education, household income, and investment decision making constant. an ordinary least squares regression model was developed to test this possibility. for the purposes of the test, only those respondents who indicated owning investable assets were included in the analysis. this reduced the sample to approximately 105,000 respondents. given the size of the sample and the possibility of obtaining highly significant results with very small effect sizes, a random sample equal to approximately 10% of the delimited sample was used in the analysis. the demographic profile of this sample matched the characteristics of the larger delimited dataset. within the regression, females were coded 1, otherwise 0. age was coded (1) under 25, (2) 25 to 34, (3) 35 to 44, (4) 45 to 54, (5) 55 to 64, (6) 65 to 74, and (7) 75 and older. the under age 25 category was the reference category. marital status was coded categorically using the following groups: (1) single, (2) living with significant other, (3) married, (4) separated or divorced, (5) widowed, and (6) shared living arrangement. the single group was the reference category. education was coded as follows: (1) some high school or less, (2) high school diploma, (3) some college, (4) associate’s degree, (5) bachelor’s degree, and (6) graduate or professional degree. the graduate and professional degree category was the reference group. household income was coded (1) less than $25,000, (2) $25,000 to $49,999, (3) $50,000 to $74,999, (4) $75,000 to $99,999, and (5) $100,000 or more. the less than $25,000 group was the reference category. financial decision making was recoded so that those who made their own investment decisions, rather than relying on the advice of another person, were coded 1, otherwise 0. it is worth noting that those in this group were more likely to report holding cash, although many also reported holding some fixed-income or other assets, such as gold or collectibles. as such, it was conjectured that financial decision making ought to be negatively associated with equity ownership. results from the regression analysis are shown in table 3. the model was statistically significant, f(23,10898) � 1976.06, p � 0.001. the model explained �31% of the variance in total equity ownership (r2 � 0.31). as shown, g&l risk scores were positively associated with equity ownership at a p � 0.001 level. although not of primary importance in this study, it is worth noting that women were less likely to hold equities. this finding matched that of hallahan, faff, and mckenzie (2004). the association between age and equity ownership was positively concave. equity ownership increased by age category up until age 55 to 64. even so, those in the oldest age group still held a higher 186 s. kuzniak et al. / financial services review 24 (2015) 177–192 proportion of investable wealth in equities compared with those in the lowest age category. the relationship between education and equity ownership was as expected, with those exhibiting low levels of attained education holding fewer equities. overall, income was positively associated with equity ownership; however, respondents in the $25,000 to $49,999 were not significantly different from those whose income was $25,000 or less. respondents who were married and separated/divorced were significantly more likely to hold equities compared to those who were single. no differences were noted among singles, those living with a significant other, respondents who were widowed, and those who were living in a shared arrangement. as hypothesized, respondents who made their own investment decisions were less likely to report holding equities. for confirmation purposes, a similar model was developed (not shown) using the combination of cash and fixed-income asset ownership as the outcome variable. the coefficient for the risk score changed from positive to negative at the p � 0.001 level. this result confirmed that g&l risk-tolerance scores were associated with objective risk taking within the sample, holding other factors constant. table 3 regression results of equity ownership variable b se � gender female �3.78 0.17 �0.06*** age 25 to 34 9.64 0.27 0.13*** 35 to 44 18.15 0.34 0.19*** 45 to 54 21.55 0.34 0.22*** 55 to 64 20.61 0.36 0.20*** 65 to 74 18.79 0.51 0.11*** 75 and older 12.26 0.85 0.04*** marital status living with significant other �0.21 0.34 �0.00 married 2.15 0.25 0.03*** separated or divorced 1.60 0.41 0.01*** widowed �1.18 0.75 �0.00 shared living arrangement �0.24 0.72 �0.00 education some high school or less �9.43 0.31 �0.12*** high school diploma �7.34 0.32 �0.08*** some college �5.26 0.27 �0.06*** associate’s degree �5.38 0.34 �0.05*** bachelor’s degree 1.07 0.23 0.02*** household income $25,000 to $49,999 0.50 0.26 0.00 $50,000 to $74,999 3.63 0.26 0.05*** $75,000 to $99,999 5.80 0.29 0.07*** $100,000 or more 7.99 0.26 0.12*** decision making make own investment decisions �3.50 0.21 �0.04*** financial risk tolerance 1.38 0.02 0.25*** constant �12.22 0.57 **p � 0.01 ***p � 0.001. 187s. kuzniak et al. / financial services review 24 (2015) 177–192 5. conclusion after conducting an extensive review of the literature, grable and joo (2004) reported that the term risk tolerance should be used as a description of a person’s willingness to take part in a behavior in which one or more outcomes are both uncertain and potentially negative. individuals, households, cultures, and societies engage in risky behavior on an hour-by-hour basis. some risks are taken on as a normal part of daily life. other risks are reluctantly taken. in the domain of financial and investment planning, the concept of financial risk tolerance has come to be seen as an important element in shaping the development of strategies designed to help households meet their financial goals. risk tolerance serves as an input into nearly all consumer and household finance decisions. while there have been attempts over the past 50 years to both describe and measure financial risk tolerance, the number and types of assessment instruments available publically has been limited. in 1999, grable and lytton published what was, at the time, a unique scale that they argued offered consumers, financial professionals, and researchers a reasonable level of reliability and validity. since 1999, numerous researchers have taken steps to test the reliability and validity of the g&l risk scale. most studies, however, were based on small convenience samples. this article extends these tests by using data (n � 160,279) collected from late 2007 through early 2014 to better describe scale response patterns, as well as provide an update on reliability and validity estimates for the scale. when evaluating findings reported in this article, it is worth remembering that data were collected over periods that included significant market volatility, and that a degree of self-selection bias was likely present in the data collection process. it is possible that the reported results might have been different had data from the great recession been excluded from the analyses and had others without internet access been asked to complete the assessment. overall, using data from more than 160,000 scale users, and subsequent random samples taken from this sample frame, the findings from this study provide additional evidence that the g&l risk scale has performed reasonably well over its 15 years of public use. based on the full sample, a cronbach’s � of � � 0.77 was estimated. further reliability estimates were made using respondent characteristics. the majority of reliability estimate fell within a range of 0.73 to 0.90, with � � 0.77 being the most typical estimate. validity tests showed that scores on the g&l scale were positively associated with equity ownership and negatively related to cash and fixed-income ownership. these results provide evidence that the scale continues to offer users an economical way to differentiate between individuals who are more or less likely to take financial risk. notes 1. reliability refers to “the extent to which [assessments] are repeatable and that any random influence which tends to make measurements different from occasion to occasion is a source of measurement error” (nunnally, 1967, p. 206). reliability provides an indication of how consistent responses are or will be over time. cron188 s. kuzniak et al. / financial services review 24 (2015) 177–192 bach’s � represents the lower bound of reliability (cortina, 1993). peterson (1994) noted that the average reported cronbach’s � in the psychological and marketing literature ranges from .76 to .77. generally, scores below � � .70 are considered to be useful only in exploratory studies. scores greater than � � .90 are also considered problematic because of item redundancy (boyle, 1991). appendix a. 13-item risk tolerance scale 1. in general, how would your best friend describe you as a risk taker? a. a real gambler b. willing to take risks after completing adequate research c. cautious d. a real risk avoider 2. you are on a tv game show and can choose one of the following, which would you take? a. $1,000 in cash b. a 50% chance at winning $5,000 c. a 25% chance at winning $10,000 d. a 5% chance at winning $100,000 3. you have just finished saving for a “once-in-a-lifetime” vacation. three weeks before you plan to leave, you lose your job. you would: a. cancel the vacation b. take a much more modest vacation c. go as scheduled, reasoning that you need the time to prepare for a job search d. extend your vacation, because this might be your last chance to go first-class 4. if you unexpectedly received $20,000 to invest, what would you do? a. deposit it in a bank account, money market account, or an insured cd b. invest it in safe high quality bonds or bond mutual funds c. invest it in stocks or stock mutual funds 5. in terms of experience, how comfortable are you investing in stocks or stock mutual funds? a. not at all comfortable b. somewhat comfortable c. very comfortable 6. when you think of the word “risk,” which of the following words comes to mind first? a. loss b. uncertainty c. opportunity d. thrill 7. some experts are predicting prices of assets such as gold, jewels, collectibles, and real estate (hard assets) to increase in value; bond prices may fall, however, experts tend to agree that government bonds are relatively safe. most of your investment assets are now in high interest government bonds. what would you do? a. hold the bonds 189s. kuzniak et al. / financial services review 24 (2015) 177–192 b. sell the bonds, put half the proceeds into money market accounts, and the other half into hard assets c. sell the bonds and put the total proceeds into hard assets d. sell the bonds, put all the money into hard assets, and borrow additional money to buy more 8. given the best and worst case returns of the four investment choices below, which would you prefer? a. $200 gain best case; $0 gain/loss worst case b. $800 gain best case; $200 loss worst case c. $2,600 gain best case; $800 loss worst case d. $4,800 gain best case; $2,400 loss worst case 9. in addition to whatever you own, you have been given $1,000. you are now asked to choose between: a. a sure gain of $500 b. a 50% chance to gain $1,000 and a 50% chance to gain nothing 10. in addition to whatever you own, you have been given $2,000. you are now asked to choose between: a. a sure loss of $500 b. a 50% chance to lose $1,000 and a 50% chance to lose nothing 11. suppose a relative left you an inheritance of $100,000, stipulating in the will that you invest all the money in one of the following choices. which one would you select? a. a savings account or money market mutual fund b. a mutual fund that owns stocks and bonds c. a portfolio of 15 common stocks d. commodities like gold, silver, and oil 12. if you had to invest $20,000, which of the following investment choices would you find most appealing? a. 60% in low-risk investments, 30% in medium-risk investments, 10% in high-risk investments b. 30% in low-risk investments, 40% in medium-risk investments, 30% in high-risk investments c. 10% in low-risk investments, 40% in medium-risk investments, 50% in high-risk investments 13. your trusted friend and neighbor, an experienced geologist, is putting together a group of investors to fund an exploratory gold mining venture. the venture could pay back 50 to 100 times the investment if successful. if the mine is a bust, the entire investment is worthless. your friend estimates the chance of success is only 20%. if you had the money, how much would you invest? a. nothing b. one month’s salary c. three month’s salary d. six month’s salary 190 s. kuzniak et al. / financial services review 24 (2015) 177–192 scoring 1. a � 4; b � 3; c � 2; d � 1 2. a � 1; b � 2; c � 3; d � 4 3. a � 1; b � 2; c � 3; d � 4 4. a � 1; b � 2; c � 3 5. a � 1; b � 2; c � 3 6. a � 1; b � 2; c � 3; d � 4 7. a � 1; b � 2; c � 3; d � 4 8. a � 1; b � 2; c � 3; d � 4 9. a � 1; b � 3a 10. a � 1; b � 3 11. a � 1; b � 2; c � 3; d � 4 12. a � 1; b � 2; c � 3 13. a � 1; b � 2; c � 3; d � 4 source: grable, j., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. a answers to questions 9 and 10 can be averaged to obtain a combined score. references arano, k., parker, c., & terry, r. (2010). gender-based risk aversion and retirement asset allocation. economic inquiry, 48, 147–155. barsky, r. b., juster, f. t., kimball, m. s., & shapiro, m. d. (1997). preference parameters and behavioral heterogeneity: an experimental approach in the health and retirement study. quarterly journal of economics, 112, 537–579. boyle, g. j. (1991). does item homogeneity indicate internal consistency or item redundancy in psychometric scales? personality and individual differences, 12, 291–294. carr, n. c. (2014). reassessing the assessment: exploring the factors that contribute to comprehensive financial risk evaluation. unpublished doctoral dissertation. manhattan, ks, kansas state university. (available at http://krex.k-state.edu/dspace/bitstream/handle/2097/17283/nickcarr2014.pdf?sequence�5). cortina, j. m. (1993). what is coefficient alpha? an examination of theory and applications. journal of applied psychology, 78, 98–104. gilliam, j., chatterjee, s., & grable, j. (2010). measuring the perception of financial risk tolerance: a tale of two measures. journal of financial counseling and planning, 21, 30–43. google. (2014). john e grable. (available at http://scholar.google.com/citations?user�6ddmmukaaaaj&hl� en&oi�ao). grable, j. e. (2004). assessing financial risk-tolerance attitudes using semantic differential scales. journal of personal finance, 3, 68–88. grable, j. e. (2008). risk tolerance. in j. j. xiao (ed.), advances in consumer financial behavior research (pp. 1–20). new york, ny: springer. grable, j. e., & joo, s. (2001). a further examination of financial help-seeking behavior. financial counseling and planning, 12, 55–66. grable, j. e., & joo, s-h. (2004). environmental and biopsychosocial factors associated with financial risk tolerance. journal of financial counseling and planning, 15, 73–82. grable, j., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. 191s. kuzniak et al. / financial services review 24 (2015) 177–192 grable, j. e., & lytton, r. h. (2003). the development of a risk assessment instrument: a follow-up study. financial services review, 12, 257–274. grable, j. e., & schumm, w. (2010). an estimate of the reliability of the survey of consumer finances risk-tolerance question. journal of personal finance, 9, 117–131. hanna, s. d., & lindamood, s. (2004). an improved measure of risk aversion. journal of financial counseling and planning, 15, 27–45. hallahan, t., faff, r. w., & mckenzie, m. d. (2004). an empirical investigation of personal financial risk tolerance. financial services review, 13, 57–78. hariharan, g., chapman, k. s., & domian, d. l. (2000). risk tolerance and asset allocation for investors nearing retirement. financial services review, 9, 159–170. maccrimmon, k. r., & wehrung, d. a. (1986). risk management. new york, ny: the free press. markowitz, h. m. (1952). portfolio selection. journal of finance, 7, 77–91. neelakantan, u. (2010). estimation and impact of gender differences in risk tolerance. economic inquiry, 48, 228–233. nunnally, j. c. (1967). psychometric theory. new york, ny: mcgraw-hill. peterson, r. a. (1994). a meta-analysis of cronbach’s coefficient alpha. journal of consumer research, 21, 381–391. weber, e. u., blais, a-r., & betz, n. e. (2002). a domain-specific risk-attitude scale: measuring risk perceptions and risk behaviors. journal of behavioral decision making, 15, 263–290. yang, y. (2004). characteristics of risk preferences: revelations from grable and lytton’s 13-item questionnaire. journal of personal finance, 3, 20–40. yao, r., hanna, s. d., & lindamood, s. (2004). change in financial risk tolerance, 1983–2001. financial services review, 13, 249–266. 192 s. kuzniak et al. / financial services review 24 (2015) 177–192 financial services review, 31(4) i volume 31 issue 4 from the editor welcome to the latest issue of financial services review (fsr). if you are a member of the academy of financial services or have been a devoted reader of the journal, you know that fsr has faced several challenges over the past few years. to begin with, the long-time editor of the journal, dr. stuart michelson, passed away unexpectedly in march 2022. dr. michelson was the model journal editor. dr. michelson was quick to work with authors to shepherd papers through the review process. he was firm with his decisions but kind in his interrelations with authors. everyone associated with fsr quickly realized that dr. michelson, as an editor, professional, and colleague, could never be replaced. yet, fsr needed to move forward, even if moving forward was at a slow speed. it was at this same time that the board of directors of the academy of financial services learned that the financial planning association would no longer provide financial or editorial support to the journal. at this moment of crisis, fsr might have gone away, but it did not. at that critical moment in time, dr. terrance martin—who was an assistant professor at the time—stepped forward, without compensation or fanfare, to pick up the pieces of what remained of fsr. without dr. martin's work, i am doubtful that fsr could have survived. this is where my fsr editorial story begins. i have been a member of the academy of financial services since the mid-1990s when i was in graduate school. my first published “paper” (it was actually a description of something new called the worldwide web) was in fsr, and since then, i have been a devoted reader of fsr. it was clear to me in 2023 that someone needed to step up and let dr. martin serve the academy of financial services in other ways (the daunting task of managing a journal is not something someone in the tenure and promotion process ought to be doing). i agreed to do so, but as i told the academy of financial services board at the time, i see my role with fsr as a bridge between the past and the future. a big part of building this bridge to the future involved creating a new journal management and submission system. fsr is now a diamond open access journal. fsr is housed in the university of georgia library system, which means that every issue of the journal can be accessed freely on the internet. this also ensures that authors will gain maximum exposure to their published work. making the transition to diamond open access could not have occurred without the dedicated efforts of shawn brayman, dr. inga timmerman, dr. wookjae heo, dr. jamie lynn byram, and the support of the academy of financial services board of directors. everyone associated with fsr is still learning about the new system, but i am happy to report that fsr is back on solid footing and moving forward with purpose. i have some compelling evidence to back up that last statement. take a look at the papers in this issue. the lead article was written by drs. keith campbell, jim exley, and patrick doyle. dr. campbell and his associates are, without question, the world’s leading experts when it comes to personality assessment as it relates to financial planning. i am certain that this article will quickly become the seminal paper of reference for anyone who incorporates measures of personality into models designed to describe household and individual financial attitudes and behaviors. the second article was authored by shawn brayman, nicki potts, kira brayman, and yegor komissarov. again, i predict this paper will garner international attention. these authors illustrate ways to map investor risk profiles into suitable portfolios, products, and solutions. this is the type financial services review, 31(4) ii of work regulators and financial planning professionals have been clamoring for. the third article was written by drs. michelle kruger, lance palmer, and joe goetz. this promises to be a widely cited paper because it is among the first articles to explicitly describe how shared financial decision-making behavior among married and cohabiting partners is related to financial satisfaction. this issue of fsr concludes with an article written by drs. andreas oehler and matthias horn. not only is their research methodologically eloquent, but their results provide clear evidence that financial literacy and risk aversion play an important role in describing capital market participation. i very much hope you enjoy this issue of fsr. the journal is on solid footing. the future of fsr looks, from my perspective, very positive. let me end by encouraging you to submit a paper. i promise the review process will be honest, quick, and transparent. in the meantime, let your colleagues know about the “new” fsr. you can help spread the word by sharing “financialservicesreview.org” on social media sites. all the best, john e. grable, ph.d., cfp® structured certificates of deposit: introduction and valuation geng deng, ph.d., cfa, frma, tim dulaney, ph.d., frma, tim husson, ph.d., frma, craig mccann, ph.d., cfaa,* asecurities litigation and consulting group, inc., 3998 fair ridge drive, suite 250, fairfax, va 22033, usa abstract this article examines the properties and valuation of market-linked certificates of deposit (structured cds). structured cds are similar to structured products—debt securities with payoffs linked to market indexes—but while structured products have garnered significant interest in both the financial media and in the academic literature, structured cds have received relatively little attention. we review the market for structured cds in the united states and provide valuations for several common product types. using our methodology, we find significant mispricing of several common types of structured cds across multiple issuers, which is similar in magnitude to the well-documented mispricing in the structured products market. in particular, we estimate that structured cds are typically worth �93% of the value of a contemporaneously issued fixed-rate cd. these results suggest that unsophisticated investors may not understand the value, risks, and subtleties of these ostensibly conservative investments. © 2014 academy of financial services. all rights reserved. keywords: certificate of deposit; structured product; cd 1. introduction structured certificates of deposit (structured cds), also referred to as “market-linked cds,” “equity-linked cds,” and “contingent interest cds,” have existed since the late 1980s. structured cds are fdic insured deposits with interest payments that are contingent upon changes in the levels of indexes, individual equities, and interest rates, or combinations of indexes, individual equities, and interest rates. * corresponding author. tel.: �1-703-246-9381; fax: �1-703-246-9387. e-mail address: craigmccann@slcg.com (c. mccann) financial services review 23 (2014) 219–237 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. because structured cds are not securities registered with the sec, the size of the structured cd market is not clear, but has been estimated to be in the tens of billions of dollars annually and growing.1 investors in structured cds have been described as “riskaverse” and “conservative,” but “dissatisfied with the rates available in the traditional fixed-income space.”2 a recent study conducted by the international organization of securities commissions found that, at least in the united kingdom, structured cd investors were typically over 55 or under 35 with high household income.3 market observers suggest that banks are issuing increasing quantities of structured cds to compete with the low yield on traditional cds and treasury securities. these reports suggest that structured cds are primarily being offered to individual investors. offering documents and marketing materials for structured cds also suggest that they are being sold through financial advisors to individual investors. although they are often marketed alongside structured notes and other retail investments, they can also be sold at or through local banks, potentially bypassing regulatory requirements of brokers and financial advisors. there remains a degree of regulatory uncertainty regarding structured cds, which could lead to unique risks and inappropriate sales to investors. the academic and practitioner literature on structured cds is sparse compared with the considerable body of theoretical and empirical work for structured products.4 king and remolona (1987) showed that banks can typically hedge structured cds efficiently using exchange-traded or synthetic options matching the payout of the structured cd.5 chance and broughton (1988) provided additional analysis and interpretation of these products. more recently, edwards and swidler (2005) argue that equity-linked structured cds do not have equity-like returns. unfortunately, these studies analyzed only a few structured cd types and did not value a significant number of actual products. given these substantial risks and the lack of an established literature for this type of investment, investors and financial advisors may not fully understand these highly complex products. as with structured products, the customizability of structured cds allows issuers to charge additional gross margin into the product without explicitly stating their impact as commissions and fees. fees on structured cds are often between 3% and 4%, and we have observed all-in fees as high as 8%, but could effectively be higher because of mispricing of the embedded options. financial advisors should recognize the potential for this form of mispricing and the features that can exacerbate it. in addition, investors and advisors should recognize that structured cds typically carry fdic insurance, but are not risk-free. the sec has outlined a variety of risks in structured cds, including liquidity risk, market risk, call risk, as well as special tax considerations.6 finra is reportedly investigating how structured cds are sold to investors given their increasing complexity and market growth.7 the fdic has also issued a short investor alert on structured cds,8 and the nyse has provided guidance to issuers regarding sales practices and disclosures.9 however, structured cds remain relatively unexplored by analysts and underreported by the financial media. we extend this literature by (1) describing the market for structured cds using a sample of over 2,000 cds; (2) providing valuations for four of the most common crediting formulas in our sample; and (3) valuing over 300 products issued by a variety of banks and documenting significant discounts to face value. our findings suggest that structured cds 220 g. deng et al. / financial services review 23 (2014) 219–237 tend to be worth less than face value or contemporaneously issued traditional cds, have a high probability of crediting the minimum return, and have death provisions that are often of negative value to the investor. these properties would not be apparent to retail investors. 2. the market for structured cds 2.1. our sample of structured cd we have collected a sample of 2,072 structured cds issued in the past eight years.10 we searched for cds issued in the united states with more than $1,000,000 issued that reported an underlying security to bloomberg. the latter requirement confines our sample to structured cds linked to equities, commodities, and other tangible assets, and therefore, excludes structured cds linked to interest rates, such as steepener and range accrual cds. although we collected products from as early as 2005, our sample is heavily concentrated in 2009– 2013. both the weighted average term—weighted by amount outstanding—and simple average for the structured cds in our sample is 5.8 years. table 1 summarizes our sample by issuing bank.11 although other banks’ average issue size ranges from $1.3 million to $16 million, bank of america has the largest average issue size ($81.6 million). bank of america’s large average issue size is because of several large deals including a billion dollar structured cd (cusip: 06051acf7) issued in february 2008. the great majority of the structured cds in our sample are linked to baskets of commodities, equities, currencies, indexes, or some combination thereof. table 2 summarizes the major underlying assets within our sample. two of the top underlying assets are jp morgan proprietary indexes.12 all of the structured cds linked to these indexes were issued by jp morgan. in fact, more than half of the jp morgan structured cds in our sample (by aggregate amount outstanding) were linked to a jp morgan proprietary index. jp morgan’s use of its proprietary index in its cds adds to the information advantage issuers have over investors. table 1 structured cd market sample by issuer issuer number total face value (mm) hsbc 906 $ 5,827.6 jp morgan chase 499 $ 2,607.3 bank of america 27 $ 2,202.1 barclays 213 $ 1,444.6 suntrust 134 $ 896.3 wells fargo 165 $ 615.6 union bank 23 $ 370.8 citigroup 23 $ 283.5 bmo harris bank 30 $ 161.3 other issuers 52 $ 263.0 total 2,072 $14,672.0 221g. deng et al. / financial services review 23 (2014) 219–237 2.2. how structured cds are marketed and sold in the united states we have also collected a number of marketing brochures and other sales material related to structured cds from a variety of issuers. in general, issuers claim that structured cds combine the safety of cds with additional market-linked upside. wells fargo, for example, claims that structured cds “can provide a creative solution for investors looking to gain access to the markets while reducing their exposure to market risk when held to maturity.”13 our results in section 4 suggest that the potential market-related gains in structured cds are minimal. the offering documents for structured cds and those for structured products are similar.14 both disclose the terms and relevant pricing parameters used in the offering, and tend to follow a similar format for each offering from a given issuer. other marketing materials, such as sales brochures and lists of available products, are also similar to those that exist for structured products (and can often be found on the internet from brokerage firm websites). structured cd materials, of course, tend to prominently note fdic insurance. however, brokers who sell structured cds, just like traditional cds, do not have to be registered nor licensed by any state or federal agency, though many may be associated with banks or licensed broker-dealer institutions.15 according to the nyse rules 401 (“business conduct”) and 405 (“diligence as to the accounts”), structured cds should be “priced at market value on customer account statements, not at the purchase amount or at par unless that is the actual market price.”16 although structured products are now required to prominently disclose a fair market value on offering materials, there is no such disclosure requirement for structured cds. as we will demonstrate below, the fair market value of structured cds is often much less than face value, reflecting a premium to the issuer that is similar in magnitude to those seen in structured products. in fact, not having to disclose fair market values might be a reason for banks to issue structured cds over structured products or traditional cds. fdic insured banks are limited to offering cd rates that are no more than 75 basis points above the average of yields for cds of comparable terms within their local market without a waiver from the fdic.17 fdic insured banks and deposit brokers may issue structured cds to circumvent this limit. for example, instead of issuing a traditional cd at a low fixed rate, a bank could market a structured cd with a higher comparable yield than the fdic mandated cap. an investor may table 2 structured cd market sample by underlying asset underlying asset number total face value (mm) baskets 1,293 $ 9,346.5 s&p 500 index 210 $ 1,924.1 jp morgan etf efficiente 5 index 111 $ 942.8 dow jones industrial average 136 $ 913.3 jp morgan optimax market-neutral index 34 $ 260.7 dow jones-ubs commodity index 26 $ 204.2 russell 2000 index 37 $ 135.6 other underlying assets 225 $ 944.7 total 2,072 $14,672.0 222 g. deng et al. / financial services review 23 (2014) 219–237 believe that this product offers higher yield than a traditional cd, even though its fair value might in fact be lower. 3. how to determine fair value we present a monte carlo method for determining the fair value of some common structured cds. we assume throughout that the underlying asset returns follow geometric brownian motion (black and scholes, 1973; merton, 1973), and simulate asset values with a monte carlo framework in accordance with procedures outlined in (glasserman, 2003). the variance of that geometric brownian motion is set to be the longest-term implied volatility of the underlying asset available from bloomberg. investors in structured cds are not paid distributions and the underlying returns reflect only capital appreciation. we compare our structured cd values to the present value of contemporaneously issued traditional cds of a similar maturity. fig. 1 graphs the simple average of traditional cd rates as reported by the fdic between may 2009 and march 2013.18 for our analysis, we used the “nonjumbo” rates (for deposit amounts under $100,000) as these are most likely to be fdic insured; however, the average difference between jumbo and nonjumbo rates over this period was only two basis points according to the fdic data. the secondary market rates for short-term cds as quoted by the federal reserve board are universally higher than those quoted for cds in an initial offering by the fdic, often by 50 basis points or more. we include the reported national average rates for five-year traditional cds from bankrate.com in fig. 1 to show the magnitude of the discrepancy.19 we use the rates reported by the fdic as a conservative estimate for the relative value of a structured cd. fig. 1. national average rates on traditional five year cds. 223g. deng et al. / financial services review 23 (2014) 219–237 3.1. credit risk investors who purchase structured cds bear credit risk. the fdic made clear to investors in the spring of 2012 that its insurance covers only principal repayment at maturity and accrued interest, not any market-linked or contingent interest payments.20 therefore, as reflected in our valuation equation in section 4, structured cds are subject to the credit risk of the issuer insofar as they are exposed to market fluctuations. as a result, a structured cd that is issued by a less credit-worthy bank is worth less than an identical structured cd issued by a more credit-worthy bank. in addition, there are limits to the amount covered by fdic insurance.21 the limit can be made less restrictive by depositing funds at more than one fdic insured institution, across several ownership categories or through joint ownership of deposited funds. investors should also realize fdic insurance does not protect against losses on early withdrawals, which usually come with heavy penalties or forfeiture provisions.22 any investment in a structured cd over the fdic limit would further expose investors to credit risk. 3.2. liquidity and call risk there are no public sources for market prices of structured cds or any independent market makers for structured cds. secondary markets for structured cds, like those for structured products, may be limited to the broker-dealers or banks who issued them, which may in turn limit the price at which an investor can sell his or her investment before maturity.23 we have not included in our valuations any discount to reflect the illiquidity of structured cds. therefore, our valuations are conservative. some structured cds include a call feature whereby the issuer may redeem the structured cd at par at certain times before maturity. if a structured cds’ underlying asset has increased in value and payment at maturity is likely large, an issuing bank could in theory call back the note at par, depriving the investor of the market-linked return despite having tied up his or her money in the structured cd. call risk is, therefore, a potentially significant factor in structured cd valuation, though none of the products in our valuation sample are callable. 4. valuing four common structured cd types the following four subsections introduce common product structures found in our structured cd sample: contingent coupon, single-observation, ratchet, and average-value cds. contingent coupon cds pay periodic coupons based on the return of an underlying asset. single-observation cds pay back principal at maturity plus an interest component that reflects the return of the underlying asset over the entire term of the cd. ratchet cds also payout only at maturity, but their interest component is based on returns calculated over several periods. the interest component of average value cds is based on the average of the underlying asset level at several times before maturity. we present valuation methods for each type, along with example valuations. unless otherwise stated, we report the simple average of the valuations for investors aged 25 to 85. 224 g. deng et al. / financial services review 23 (2014) 219–237 4.1. contingent coupon cd the most common type of structured cd in our sample is called a contingent coupon cd. contingent coupon cds are typically linked to baskets of assets and pay periodic interest payments contingent upon the return of the basket. before each coupon payment date, the prices of the basket elements are observed. returns are calculated, capped and floored to produce the component returns. if the weighted average of these component returns is larger than the minimum coupon (usually zero), then a market-contingent coupon payment is paid to the investor. otherwise, the minimum coupon is paid to the investor, which could be zero. component returns for this type of cd are calculated in one of two ways: unbuffered or buffered. an unbuffered component return is a continuous function of the basket element’s return. with a buffered component return cd, if the basket element’s return is larger than the buffer level, then the investor is credited with the maximum return (the return cap). the payoff of unbuffered structured cds includes payoffs to vanilla european options and the payoff of buffered structured cds includes payoffs to both vanilla and binary options. for the same level of cap and floor, a buffered cd would be worth more than an unbuffered cd (assuming the buffer is less than the cap). as a result, for a buffered cd and an unbuffered cd to be similarly profitable, either the return cap or return floor must be lower on the unbuffered cd, all else equal. for an example valuation, we take the contingent coupon cd issued by barclays bank delaware on february 29, 2012, maturing on february 28, 2019. the cusip for this cd is 06740arx9 and the aggregate issue size is $2,915,000. the cd was linked to an equally weighted basket of 10 equities. barclays paid coupons annually at a rate determined by a floor of �25%, a cap of between 6.5% and 9.5% and a buffer of 0%. we assumed a cap of 8% when valuing this buffered contingent coupon cd and found that the cd was worth approximately $933.79 per $1,000 face value cd; �91% of the value of a contemporaneously issued fixed-rate cd.24 we found that in two-thirds of the simulations, investors received only the minimum return. as an example of an unbuffered contingent coupon cd, we use the jp morgan cds issued on february 29, 2012 with an aggregate face value of $4,241,000 (cusip: 48123y6g6). the products included a return cap of at least 6% and paid coupons annually with a minimum coupon rate is 0.25% per annum. the return floor for this product is �30%. the cds mature in seven years on february 28, 2019. assuming a 6% cap, we value the jp morgan unbuffered contingent coupon cd at $933.88 per $1,000 face value cd; �91% of the value of a contemporaneously issued fixed-rate cd. in this case, investors received the minimum return in �80% of the simulations. the increase in this statistic relative to the buffered example can be traced to the fact that this cd includes a nonzero minimum coupon. fig. 2 illustrates the component return results for a given basket element return as of each coupon date for both the unbuffered example and buffered examples discussed above. 4.2. ratchet cd a second common type of structured cd is the “ratchet cd,” also known as a “cliquet cd.” the issuer of a ratchet cd observes the underlying asset’s price or level at several 225g. deng et al. / financial services review 23 (2014) 219–237 points during the term of the cd and calculates returns between each observation date. a local cap, and sometimes a local floor, is applied to each of the observed returns. at maturity, the observed returns are summed, and a global floor is then applied; usually a minimum return or 0% (corresponding to a return of principal). products with a local cap but without a local floor are exposed to the risk that a large decline in the underlying asset during one period could wipe out many periods of capped positive returns. this structure is similar to the “simple ratchet equity indexed annuity” of hardy (2003). hsieh and chiu (2007) show that for an initial investment p in an equity-indexed annuity (eia) with guarantee ratio �, minimum guaranteed rate g and term t that the value (veia) of the contract is given by veia � e�t��pe�rt max�1 � r, ��1 � g�t�� (1) where r is the (assumed) constant and continuously compounded riskless rate and r is the arithmetic sum of returns given by r � �i�1 nt min �max ��ri, rf�, rc� (2) here n is the number of times per year that the underlying asset is observed, ri � s(ti)/s(ti–1) � 1, rf is the local return floor, rc is the local return cap and � is the participation rate.25 for structured cds, principal protection sets � � 1 and typically there is no local floor (f � �1) or leverage (� � 1). closed-form solutions for this type of eia, in the absence of the maturity guarantee, were derived by hsieh and chiu (2007). although this formula is general and fits nicely into our context, it does not appropriately apply credit risk or fdic insurance. including the cds rate in the overall discount factor would be inappropriate given the guarantees provided by fdic insurance (� � 1). we alter eq. (1) to include the credit risk related to components that are sensitive to market returns and obtain the value of the structured cd (vscd) fig. 2. example unbuffered contingent coupon cd with cusip 48123y6g6 (a) and example buffered contingent coupon cds with cusip 06740arx9 (b). 226 g. deng et al. / financial services review 23 (2014) 219–237 vscd � e�t��pe�rt � pe��r�c�t max�r,gt�� � pe�rt � e�t��pe��r�c�t max�r,gt�� ❘ market/credit risk (3) where c is the continuously compounded cds rate of the issuer. an example of a ratchet cd is the barclays “certificates of deposit linked to the performance of the s&p 500 index due october 27, 2015” issued in october 2010 (cusip: 06740amd8). at maturity, barclays pays investors a return equal to the sum of the s&p 500 quarterly returns, each capped at between 3% and 5%, subject to a minimum return between 0.25% and 1.00% per annum. the product was priced on october 22, 2010 and will mature on october 27, 2015. although this barclays structured cd has a local cap, it has no local floor. that means that a large negative quarterly return during the term of the note could wipe out the several accumulated capped returns. for example, if the s&p 500 increases at a rate of 10% per quarter for two years, then the cd would have accumulated 3% capped returns for eight quarters. if the s&p 500 subsequently decreases by 24% or more in the next quarter, the total of the capped quarterly returns for this nine quarter period would be zero or less, even though the s&p 500 would have increased over 60% during this period. using the volatilities and market rates as of the pricing date and the midpoint for the terms of the cd (4% local cap and 0.625% minimum return per annum), we obtain a valuation of $976.46 per $1,000 face-value cd. we find that there is �97.2% chance of investors earning the minimum return and the average return above the minimum return is �2% per annum. using the contemporaneous national average rates on traditional cds from the fdic, extrapolated to the appropriate term for this cd (1.66%), we find that this cd is worth �95% of a traditional fixed rate cd. we varied the applied minimum coupon to the structure within the range specified in the preliminary offering document (between 0.25% and 1.00%). for each minimum coupon rate, we determined the local cap that will result in an estimated value equal to the value of the structured cd assuming a 25 year old investor and plotted the results in fig. 3. for this particular structured cd, we find that for each basis point increase in the minimum coupon, barclays would likely require a basis point decrease in the local cap level to achieve the same level of profit. 4.3. average return (asian) cd a third common type of structured cd in our sample resembles an average value option, also known as an asian option. the issuer of an “average return” cd observes the underlying asset at several (n) points during the term of the cd and then calculates the average of those levels (li) to determine the underlying return. a floor (rf), cap (rc), and/or participation rate (�) may then be applied to the observed average return to yield the return at maturity r � min �max �� n �i 1 n li � l0 l0 , rf� , rc� (4) 227g. deng et al. / financial services review 23 (2014) 219–237 where l0 is the initial level of the underlying asset. as with the other types of structured cds, we can write the value of the structured cd (vscd) as a component free of credit risk and a component exposed to market/credit risk: vscd � e�t��pe�rt � pe��r�c�tr� � pe�rt � e�t��pe��r�c�tr� ❘ market/credit risk (5) where p is the face-value of the cd. as an example, on december 23, 2010 suntrust issued $13,235,000 worth of indexlinked certificates of deposit linked to the dow jones industrial average (cusip: 86789vly1) maturing on december 20, 2016. the returns were subject to a minimum return of 8% and a cap of 30%. although the structured cds only pay out at maturity, suntrust observes the dow jones industrial average quarterly and the resulting 24 observed levels are averaged. based on our valuation methodology, we find that this cd was worth approximately $932.46 when it was issued in december 2010, �96% of the value of a contemporaneously issued fixed-rate cd. averaging periodic levels reduces the value of asian options compared with traditional european options. the type of crediting formula used in the structured cds we study here is known as an “averaging-in” procedure, as it uses periodically observed levels, as opposed to an “averaging-out” procedure that would average several observation made closer to maturity (bouzoubaa and osseiran, 2010). “averaging-in” includes levels observed throughfig. 3. contour plot for barclays ratchet structured cd (cusip: 06740amd8). 228 g. deng et al. / financial services review 23 (2014) 219–237 out the term of the note, meaning that their average will incorporate returns with widely differing terms. if the expected return on the index is positive, than on average the “averaging-in” procedure will reduce the value of the product relative to an “averaging-out” style product, all else constant. 4.4. single-observation cd the simplest structured cds do not pay coupons but pay a contingent return at maturity if the index increases during the term of the note.26 the contingent payout can be modeled as payoffs to long-term call options on the linked asset. the return on these cds is capped at a maximum return, which is equivalent to including the payoffs to a short out-of-themoney call option. the participation rate effectively changes the number of call options the investor is long and short. let the participation rate be given by � and the maximum (or capped) return be given by rc and minimum (floored) return rf. the value of the structured certificate of deposit vscd with face-value p and term t is given by vscd � pe�rt � e�t��pe��r�c�t min�max��r,rf�,rc�� ❘ risky component (6) in eq. (6), r is the risk-free rate and c is the cds rate of the issuer. because the market-contingent component is not paid until maturity and these cds are typically longdated, a significant liability could build up with respect to such a cd (especially if it is not callable). eq. (6) can be written in terms of european call options. let c(k,t) represent the value of a european call option with strike price k, expiring at time t. the value of the structured cd is then given by vscd � pe�rt � p s0 e�ct��c�s0,t� � c�s0�1 � rc � � ,t�� � prfe��r�c�t ❘ market/credit risk (7) where s0 is the initial price/level of the linked asset. as an example, consider the citibank, n.a. structured cd due january 29, 2016 linked to the s&p 500 (cusip: 172986bf7). the cd was priced on january 26, 2010 and issued on january 29, 2010. it pays no interest during the six-year term and, at maturity, pays the return of the s&p 500 subject to a maximum return of 41% (5.89% annual percentage yield) and a minimum return of 0%. the issue size was $1,952,000. citibank reported a comparable yield of 3.18% when the average rate for nonjumbo deposits at this time was 2.4% and for jumbo deposits was 2.44%.27 the difference between the comparable yield on this cd and the average yield for fixed-rate deposits is close to the 75 basis point fdic cap. using the terms of the contract and observable market variables, we obtain a valuation of 229g. deng et al. / financial services review 23 (2014) 219–237 $938.67 per $1,000 investment in the citibank structured cd. we find that the structured cd was worth roughly 95% of the value of a contemporaneously issued traditional cd at the national average rate.28 we find that as of the issue date there was a 58% chance that the citibank cd would return no additional payment beyond return of principal investment and a 20% chance that the maximum return would be paid to the investor. to provide intuition for the sensitivity of the structured cds value to the issuer’s discretionary parameters, we show in fig. 4 the likelihood that an investor would realize the maximum return and minimum return as a function of the minimum return. there is a very high probability of obtaining the minimum return when holding the cd to maturity. there is only a small probability of returns that would lead to direct market exposure (i.e., neither capped nor floored), even for a minimum return of zero. 4.5. mortality risk and structured cds structured cds, like traditional cds, have defined payouts if the owner of the cd dies before maturity, which may differ from the current market value of the cd. the products in our sample simply pay the original principal amount to the investor’s beneficiary; essentially, the issuer has sold a binary put option to “buy back” the cd at par upon the investor’s death, known in the literature as the “death put.” a return of principal years after the investor makes the cd will be worth less than the initial principal amount because of the time value of money. therefore, the death put can be of negative value to the investor. however, the effect of the death put may differ between types of structured cds. fig. 4. probability of realizing the minimum return and maximum return as a function of the return floor (cusip: 172986bf7). 230 g. deng et al. / financial services review 23 (2014) 219–237 we have modeled the death put for each of our product types using the unisex annuity 2000 mortality table from the transactions of the society of actuaries. in our monte carlo simulations, for each year of the product’s term we randomly select a number of simulated paths to credit with the death put payment based on the conditional death probabilities for an investor of an assumed age at purchase. we then rerun our simulations for ages between 25 and 85. fig. 5 shows the average value of each of the four types of structured cds described above for investors of varying ages. to clarify fig. 5, an average single-observation structured cd is worth approximately $96.55 to a 25 year old investor. to an 85 year old investor, the average single-observation structured cd is worth approximately $95.47, �110 basis points less. therefore, the curve for the single-observation structured cd in fig. 5 slopes downward from the 100% value to the terminal value of 98.9%. for the single-observation structured cds, the value decreases as a function of age because the payout at death is worth less than the crediting procedure would otherwise return. however, for the ratchet, average and contingent coupon types, the value actually increases with age. effectively, crediting formulas in these products reduce investors’ value more strongly than the time value of money up to death. therefore, a return of principal before maturity would be a more favorable outcome from the investor’s point of view. for all product types, however, the size of the effect is small.29 our results suggest that the death benefit is of little value to investors, and can be either positive or negative depending on product type. fig. 5. average structured cd value normalized by the average value of the structured cd to a 25 year old investor. 231g. deng et al. / financial services review 23 (2014) 219–237 4.6. summary of valuation results we have valued a total of 303 structured cds with aggregate face value of $1.8 billion. this valuation sample accounts for roughly 15% of the structured cds in our sample (12% by aggregate face value). our results indicate that, on average, investors receive �93.8 cents in value for each dollar invested in structured cds. table 3 summarizes our valuation results across product types. the size of the valuation sample is limited both by quality of the data and the idiosyncrasies of individual products. for example, many products reference proprietary indexes for which there is little or no market information, such as those noted above from jp morgan. table 3 also summarizes the structured cd valuation as a percentage of contemporaneously issued fixed-rate cds. the rates assumed for fixed-rate cds is the national average rate given by the fdic for the week that the structured cds were priced. in the event that the term of the structured cd did not match that of a traditional cd reported by the fdic, linear interpolation was used to determine the implied cd rate. with the exception of the average cds, the other structured cds were priced at significant discounts to contemporaneously issued fixed-rate cds according to our valuation results.30 table 4 summarizes the valuation results for our sample of structured cds broken down by issuing bank.31 average initial valuation is written as a percentage of face value. in the table, we also include the probability that an investor will realize the minimum return from investing in the products.32 as discussed in the contingent coupon cd section, contingent coupon cds can come in two varieties: buffered and unbuffered. table 5 decomposes the valuation results across these two types. our analysis suggests that these two types of contingent coupon cds are comparably priced and that issuers likely lower the return cap to compensate for the presence of a buffer in the buffered version of the cds. in summary, these results show that structured cds are priced at significant discounts to face value. as noted in the structured products literature, the difference between the fair value and the issue price is effectively an undisclosed additional charge by the issuer. although the amount of this charge varies by product type, it is approximately as large as the discounts observed in the structured product market.33 although structured cds are often marketed as safe investments with additional market exposure, our results show that their table 3 issue date weighted average fair value by product type product type number aggregate face value (mm) average initial valuation percentage of traditional cd contingent coupon cd 230 $1,028.0 93.18%a 92.44% ratchet cd 31 $ 308.2 96.23%a 95.12% average cd 28 $ 338.2 92.54%a 93.28% single-observation cd 14 $ 122.1 96.35%a 97.03% total 303 $1,796.4 93.80%a 93.37% aindicates that the mean ratio of structured cd values to traditional cd values is statistically different from 1 at the 95% confidence level. 232 g. deng et al. / financial services review 23 (2014) 219–237 unfavorable crediting formulas lead to both little market exposure and a high probability of below-market returns. 5. discussion in this article, we review the market for and common features of structured cds. we use a sample of products to provide aggregate data on this relatively obscure, yet very large table 4 issuer structured cd valuation summary issuer number aggregate face value (mm) initial valuation probability of minimum return contingent coupon cds jp morgan chase bank 194 $ 787.6 93.28%a 69% barclays bank delaware 31 $ 175.3 94.10%a 65% citibank 3 $ 42.2 90.77%a 54% hsbc 1 $ 21.5 86.86% 66% american national bank 1 $ 1.3 93.45% 66% total 230 $1,028.0 93.18%a 68% ratchet cds barclays bank delaware 12 $ 64.7 94.43%a 96% suntrust bank 11 $ 114.9 97.33%a 97% union bank 5 $ 123.2 96.14%a 95% jp morgan chase bank 3 $ 5.4 96.35%a 98% total 31 $ 308.2 96.23%a 96% average return cds suntrust bank 17 $ 230.6 93.03%a 94% jp morgan chase bank 9 $ 26.6 89.81%a 70% wells fargo bank 1 $ 73.2 92.83% 99% citibank 1 $ 7.8 84.81% 53% total 28 $ 338.2 92.54%a 93% single-observation cds jp morgan chase bank 6 $ 54.9 95.48%a 61% suntrust bank 3 $ 25.6 98.65% 76% barclays bank delaware 2 $ 16.3 96.83% 62% citibank 2 $ 24.3 95.67% 60% wells fargo bank 1 $ 1.0 93.86% 98% total 14 $ 122.1 96.35%a 64% aindicates that the mean ratio of structured cd values to traditional cd values is statistically different from 1 at the 95% confidence level. table 5 valuation results for our sample of contingent coupon structured cds with and without buffers return type number aggregate face value (mm) initial valuation probability of minimum return buffered 196 $ 805.1 93.02%a 68% unbuffered 34 $ 222.9 93.78%a 68% total 230 $1,028.0 93.18%a 68% aindicates that the mean ratio of structured cd values to traditional cd values is statistically different from 1 at the 95% confidence level. 233g. deng et al. / financial services review 23 (2014) 219–237 financial market. we provide valuation procedures for several structured cd types included in our sample and value a variety of products based upon these procedures. our primary findings are that (1) structured cds include payoffs to complex derivative positions; (2) structured cds are worth significantly less than their issue price; (3) this mispricing is present across product types and issuers; (4) the crediting formulas in structured cds have a high probability of crediting the minimum return; and (5) the death put can be of positive or negative value to the investor depending on product features and type. we find that most of the apparent benefits of structured cds are illusory and overwhelmed by their inherent risks and embedded fees. each of these results has important implications for individual investors and financial advisors. to our knowledge, our study is the first to examine the structured cds market in detail, but is limited by the few publicly available sources for offering documents. in addition, there are a significant number of products (even in our limited sample) that reference proprietary indexes for which there is little or no market data. this could represent an information asymmetry or conflict of interest between investors and issuers of these cds. we have also noticed a variety of “exotic” structured cds that have extremely complex payout formulas and terms as long as 15 years. these may be fruitful topics for further research. because the structured cd market is entirely over-the-counter, customers may not be able to compare different offerings and make informed decisions about the relative value of a particular structured cd, even with information such as “comparable yields” or original issuer discounts. structured cds can be extremely complex investments and, in many ways, are just as complex as structured products, which are regulated much more stringently and do not enjoy fdic. we think it worth careful consideration whether such products should be sold to retail investors, especially in an unregulated setting. until the regulatory framework for structured cds becomes clear, financial advisors can help fill this information gap between issuers and individual investors. in particular, investors must appreciate the often limited nature of the market-linked exposure, the significance of fdic insurance, and the potential for significant mispricing. these issues are likely to be the subject of continued debate as the market for structured cds continues to expand and evolve. notes 1 see bloomberg structured notes brief (2011a), bloomberg structured notes brief (2011c), bloomberg structured notes brief (2011d), and bloomberg structured notes brief (2011e). 2 see bloomberg structured notes brief (2011b). 3 regulation of retail structured products, international organization of securities commissions, april 2013. 4 see for example henderson and pearson (2010), deng et al. (2010), deng et al. (2011b), deng et al. (2011a), and deng et al. (2012). 5 the authors advocate the use of the underlying asset and futures contracts to “dynamically hedge” the position because of high liquidity and low transaction costs. 6 retrieved from http://www.sec.gov/answers/equitylinkedcds.html. 234 g. deng et al. / financial services review 23 (2014) 219–237 7 see robinson (2012). 8 retrieved from http://www.fdic.gov/consumers/consumer/news/cnspr12/market linkedcds.html. 9 nyse information memos no. 06-12, march 17, 2006. 10 this sample was collected mainly from bloomberg using the srch function, performed on march 13, 2013. the sample was augmented with additional structured cds found from other publicly available sources. 11 we include wachovia structured cds in the statistics of structured cds issued by wells fargo. 12 according to a july 2012 strategy guide, the jpmorgan etf efficiente 5 index represents the returns to a basket of twelve exchange traded funds (etfs) that are rebalanced each month to reflect the allocation that maximizes return for a given level of volatility based on the previous six months of historical data. according to a september 2009 strategy guide, the jpmorgan optimax market-neutral index represents the returns to a basket of 18 to 24 commodity indexes (sub-indices of the s&p gsci). the allocation to each commodity is selected by a proprietary algorithm that is “based on modern portfolio theory and momentum theory.” 13 wells fargo securities, market linked certificates of deposit, accessed july 2, 2013. 14 structured product offering documents are publically available from the secs edgar database, filed as form 424b2. 15 retrieved from http://www.investor.gov/investing-basics/investment-products/ certificates-deposit-cds. 16 nyse information memos no. 06-12, march 17, 2006. 17 fdic rules and regulations §337.6. 18 retrieved from http://www.fdic.gov/regulations/resources/rates/index.html. 19 the difference between the rates for jumbo and non-jumbo five year cds is closer to eight basis points according to bankrate.com data. 20 market-linked cds: don’t let the possibility of higher returns cloud your view of the potential risks, fdic consumer news, spring 2012. 21 the maximum deposit insurance amount was increased from $100,000 to $250,000 by the emergency economic stabilization act of 2008; see public law 110-343 §136(a)(1). subsequently, the helping families save their homes act of 2009 extended this increase through december 31, 2013; see public law 111-22 §204(a) (1) (a). in 2010, the wall street reform and consumer protection act made this increase permanent; see public law 111-203 §335(a). 22 the fdic has taken the position in the past that a potential decrease in principal caused by the imposition of an early withdrawal penalty does not prevent a product from qualifying as a ‘deposit’ for insurance purposes. fdic letter to kevin p. murray, february 27, 2002. 23 issuers have a disincentive to maintain such markets, as developed secondary markets may trigger more stringent regulatory requirements. 24 we note that for some periods of time, cd rates were above treasury rates of the same term. because we discount both structured and traditional cds using treasury 235g. deng et al. / financial services review 23 (2014) 219–237 rates, the value of traditional cds may exceed face value. on february 29, 2012, implied seven year cd rates were higher than seven year treasury rates. 25 eq. 2 is a trivial generalization of the formula in hsieh and chiu (2007) to include more frequent observations. 26 a cd of this type was described by edwards and swidler (2005). 27 the fdic currently reports rates on fixed-rate cds out to a maturity of 60 months on a weekly basis. we are using the rates observed by the fdic on january 25, 2010 and using linear extrapolation to determine the implied rate on a 72 month cd. 28 retrieved from http://www.fdic.gov/regulations/resources/rates/index.html. 29 the value for a 25 year old investor is only 1.1% higher for the single-observation products, 0.4% lower for the ratchet products, 0.5% lower for the average products, and 1.5% lower for contingent coupon cds when compared with the value for an 85 year old investor. 30 retrieved from http://www.fdic.gov/regulations/resources/rates/index.html. 31 for some smaller banks (e.g., union bank), cds data is unavailable as of this writing. to include the effects of credit risk for these banks, we averaged the cds rates for all banks that issued structured cds within our sample. 32 for these calculations, we exclude investors who have died during the term of the notes to estimate the probability of obtaining the minimum return (since these investors do not hold the cds to maturity). 33 bloomberg structured notes brief (2011f). references black, f., & scholes, m. (1973). the pricing of options and corporate liabilities. journal of political economy, 81, 637–654. bloomberg structured notes brief. (2011a). bonacci of cd funding securities on popular kinds of structured cds, growth in market. bloomberg structured notes brief, page 6. bloomberg structured notes brief. (2011b). david fitzsimmons on choosing which structured cds to sell and working with issuers. bloomberg structured notes brief, page 4. bloomberg structured notes brief. (2011c). wells fargo’s ginsburg on selling structured cds to crisis-weary investors. bloomberg structured notes brief, page 4. bloomberg structured notes brief. (2011d). stoddart of capital investment marketing on selling structured cds through bank branches. bloomberg structured notes brief, page 5. bloomberg structured notes brief. (2011e). barclays’s wu and gordon on commodities, cds and selling in latin america. bloomberg structured notes brief, page 5. bloomberg structured notes brief. (2013f). banks are inconsistent on new sec disclosure rules. bloomberg structured notes brief, pages 3–5. bouzoubaa, m., & osseiran, a. (2010). exotic options and hybrids: a guide to structuring, pricing and trading. wiley finance. chance, d., & broughton, j. (1988). market index depository liabilities: analysis, interpretation, and performance. journal of financial services research, 1, 335–352. deng, g., mccann, c., & o’neal, e. (2010). what tivo and jp morgan teach us about reverse convertibles. slcg working paper (available at http://www.slcg.com/). deng, g., guedj, i., mallett, j., & mccann, c. (2011a). the anatomy of absolute return barrier notes. the journal of derivatives, 19, 61–70. 236 g. deng et al. / financial services review 23 (2014) 219–237 deng, g., mallett, j., & mccann, c. (2011b). modeling autocallable structured products. journal of derivatives 4 hedge funds, 17, 326–340. deng, g., dulaney, t., & mccann, c. (2012). valuation of reverse convertibles in the variance gamma economy. slcg working paper (available at http://www.slcg.com/). edwards, m., & swidler, s. (2005). do equity-linked certificates of deposit have equity-like returns? financial services review, 14, 305–318. glasserman, p. (2003). monte carlo methods in financial engineering. berlin: springer. hardy, m. (2003). ratchet equity indexed annuities. university of waterloo working paper. henderson, b., & pearson, n. (2010). the dark side of financial innovation: a case study of the pricing of a retail financial product. the journal of financial economics, 100, 227–247. hsieh, m., & chiu, y. (2007). monte carlo methods for valuation of ratch equity indexed annuities. proceedings of the 2007 winter simulation conference. king, s., & remolona, e. (1987). the pricing and hedging of market index deposits. federal reserve bank new york quarterly review, 11, 9–20. merton, r. (1973). theory of rational option pricing. bell journal of economics and management science, 4, 141–183. robinson, m. (2012). finra examines cds tied to derivatives as sales surge to record. bloomberg news, february 7, 2012. 237g. deng et al. / financial services review 23 (2014) 219–237 financial services review, 31(4) 266 financial satisfaction: the role of shared financial responsibilities and shared financial values among couples michelle kruger,1 lance palmer,2 and joseph goetz3 abstract managing shared finances is an important aspect of a romantic relationship, and satisfaction with one’s financial situation depends on a complex host of issues, including decision-making dynamics and resource sharing. this paper provides insight into this relationship by reporting the results from a study designed to provide evidence of an association between couple’s shared financial decision-making behavior and their financial satisfaction. using a sample of 602 individuals in a committed romantic relationship, this project evaluated how couples’ division of financial responsibilities and agreement on spending and saving behavior affected their perceived financial satisfaction. results of the analysis indicated that the way household finances were shared was associated with perceptions of financial satisfaction. specifically, those who reported combining their finances with their partner were more financially satisfied. couples who reported higher levels of agreement on spending were more likely to be satisfied with their current financial situation. finally, couples with higher levels of agreement on saving were more satisfied financially. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation kruger, m., palmer, l., & goetz, j. (2023). financial satisfaction: the role of shared financial responsibilities and share financial values among couples. financial services review, 31(4), 266-282. introduction financial satisfaction is the subjective perception of one’s financial adequacy and resources (hira & mugenda, 1998). it is an important topic for policymakers, researchers, and educators because it contributes to household outcome measures such as well-being (campbell, 1981; cfpb, 2015; easterlin, 2006; joo, 2008; robb & 1 corresponding author (mkruger@uga.edu). department of financial planning, housing and consumer economics, university of georgia, athens, usa 2 department of financial planning, housing and consumer economics, university of georgia, athens, usa 3 department of financial planning, housing and consumer economics, university of georgia, athens, usa woodyard, 2011) and overall life satisfaction (xiao et al., 2009). many studies on financial satisfaction focus on identifying characteristics of household financial decision-makers (e.g., joo & grable, 2004; tharp, 2017; woodyard & robb, 2016). in this regard, financial satisfaction has been used as both an outcome and explanatory variable associated with objective and subjective demographic and socioeconomic characteristics, https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 31(4) 267 financial attitudes, financial behaviors, and financial knowledge. other research has evaluated how perceptions of a partner’s spending and saving behavior impact measures of relationship satisfaction (archuleta, 2013; mao et al., 2017). this literature shows a positive connection between financial and marital satisfaction (koochel et al., 2020; ross et al., 2021). yet, little is known about how financial mutuality, such as sharing financial goals and money management responsibilities, is associated with couples’ financial satisfaction (archuleta et al., 2013; grable et al., 2021). moreover, though the relation between partner spending and relationship satisfaction is well established (britt et al., 2008; kelley et al., 2022; li et al., 2023), the association between financial satisfaction and agreement on saving and spending behaviors has yet to be fully evaluated. this paper therefore aims to address this issue by examining how couples’ division of financial responsibilities and spending and saving compatibility are associated with their financial satisfaction. background george (1992) defines financial satisfaction as the subjective evaluation of a person’s resource adequacy. joo and grable (2004) define it as one’s overall satisfaction with their current financial situation. given this evolving definition, it is perhaps unsurprising that there is no consensus on the best way to measure financial satisfaction (godwin, 1994). some researchers use scales and indexes, whereas others consider single-item measures (aboagye & jung, 2018; hira & mugenda, 1998; joo & grable, 2004). further, numerous sub-facets of financial satisfaction have been identified (garrett & james, 2013; joo & grable, 2004; tharp, 2017; woodyard & robb, 2016). these include financial strain, attitudes, behaviors, financial knowledge, personal characteristics, and couple-level characteristics. although researchers have studied financial satisfaction for over a century, the modeling of financial satisfaction is a more recent phenomenon. for example, joo and grable (2004) proposed a framework to describe the abovementioned personal and household characteristics associated with one’s assessment of their current financial situation, operationalized as follows: 𝐹𝑆 = 𝑓(𝑆, 𝐴, 𝐵, 𝐾, 𝑃), (1) where fs, financial satisfaction, is a function of five factors represented by s, a vector of financial strain variables; a, a vector of attitudinal variables; b, a vector of behavioral variables; k, financial knowledge; and p, a vector of personal characteristics. following garrett and james (2013), tharp (2017), and woodyard and robb (2016), this study builds on joo and grable’s (2004) model to evaluate the functional domains and variables evaluated in this study: 𝐹𝑆 = 𝑓(𝑆, 𝐴, 𝐵, 𝐾, 𝑃, 𝐶𝐿), (2) where cl is added to capture a vector of variables representing couple-level characteristics. the following sections discuss these six domains in detail. financial strain financial strain results from major life-cycle events that influence the family system and generally cost significant sums of money to solve (joo & grable, 2004), such as unemployment (plagnol, 2011; vera-toscano et al., 2006). these difficulties are associated with lower levels of financial satisfaction (archuleta et al., 2011; spuhler & dew, 2019), and reducing them offers a pathway to increasing financial satisfaction (xiao et al., 2006). attitudes an attitudinal variable of importance in nearly all financial satisfaction frameworks is financial risk tolerance, or a person’s willingness to pursue uncertain and potentially negative outcomes. risk tolerance is positively associated with financial satisfaction (aboagye & jung, 2018; joo & grable, 2004). the prevailing thought is that risk tolerance is a trait-like factor (van de venter et al., 2012), and as such, it helps describe the degree to which someone expresses financial satisfaction. beyond causality, it is reasonable to hypothesize that financial risk tolerance is positively associated with financial decisionmaking and subsequent outcomes which, in turn, relate to feelings of financial satisfaction. kruger et al. 268 behaviors financial satisfaction can be negatively affected by mishandling of household financial management tasks (joo, 2008: porter, 1990; xiao et al., 2014). those who perform routine financial management tasks well (e.g., handling cash and credit accounts) are more likely to feel satisfied with their financial situation. positive financial behavior, such as managing cashflows and credit accounts to fully pay off monthly credit card balances (joo, 1998), is positively associated with financial satisfaction (aboagye & jung, 2018; joo & grable, 2004; xiao et al., 2009), whereas risky financial behaviors, such as spending more than one’s income (aboagye & jung, 2018), is associated with lower levels of financial satisfaction (xiao et al., 2014). financial solvency—a behavioral outcome associated with the ability to pay off all debt and still have assets remaining—is also positively associated with financial satisfaction (joo & grable, 2004; garrett & james, 2013; mugenda et al., 1990; sumarwan & hira, 1993). financial knowledge financial knowledge can be objective or subjective (fan & babiarz, 2019; xiao et al., 2009). objective financial knowledge is sometimes referred to as financial literacy (atlas et al., 2019), which can be measured using a quiz or a gradable survey testing the subject’s knowledge of financial topics. subjective financial knowledge can be an indicator of a decision-maker’s confidence (atlas et al., 2019) and is typically measured by asking the person to rate their own level of financial knowledge. robb and woodyard (2011) found that subjective financial knowledge offers insight into one’s financial behavior, which is corroborated by evidence showing a positive relationship between subjective financial knowledge and financial satisfaction (fan & babiarz, 2019; joo & grable, 2004; xiao et al., 2014). personal characteristics of particular importance when describing financial satisfaction is the demographic profile of a financial decision-maker. fan and babiarz (2019) noted that divorced individuals, particularly women, generally exhibit lower financial satisfaction than those who are married. fan and babiarz found that single women are less financially satisfied than married women. men are more likely than women to report being satisfied with their current financial situation, after controlling for socioeconomic status (hira & mugenda, 1998; xiao et al., 2014), although gender differences in financial satisfaction generally decrease with age (hansen et al., 2008). the relationship between attained education level and financial satisfaction is less clear (fan & babiarz, 2019; joo & grable, 2004; hsieh, 2004). in terms of racial or ethnic background, zurlo (2009) reported that non-whites are significantly less financially satisfied than whites. others found no significant relationship between race or ethnic background and financial satisfaction (e.g., hsieh, 2001, 2004; joo & grable, 2004). age and income are positively associated with financial satisfaction (archuleta, 2013; hansen et al., 2008; sumarwan & hira, 1993), perhaps because older adults tend to have higher net worth and higher incomes than younger adults. however, hansen et al. (2008) found that older adults appear to be more satisfied with their financial situation than younger adults, even when wealth and income levels are the same, suggesting they may have learned to be more content with their circumstances. couple-level characteristics joint financial decision-making relates to perceptions of shared financial values among romantic partners (totenhagen et al., 2019). discrepancies in income-earning between romantic partners can contribute to financial conflict and reduce financial satisfaction (eirich & robinson, 2017). for example, those who report having equal levels of economic power in a relationship (i.e., equal earnings) are more likely to report lower levels of financial conflict (dew & stewart, 2012). hypotheses how a person assesses their financial satisfaction is a function of different personal and household characteristics. of particular interest to this study is how financial satisfaction among those in a committed romantic relationship relates to couple-level characteristics, such as the degree to which they share goals and values and the financial services review, 31(4) 269 division of financial responsibilities. understanding how these characteristics affect financial satisfaction can offer insight and tools to improve one’s financial wellbeing. the present study thus extends the financial counseling literature by investigating whether the relationship between financial satisfaction and shared financial goals and values extends to a couple’s level of agreement on specific financial behaviors. extending joo and grable’s (2004) financial satisfaction framework, the following hypotheses examine five financial characteristics of couples (i.e., financial integration style, financial decision-making style, income-earning style, agreement on spending, and agreement on saving) and how they relate to financial satisfaction: h1: partners who report combining their finances are more likely to be satisfied with their current financial situation. h2: partners who report being jointly responsible for financial decisions and management are more likely to be satisfied with their current financial situation. h3: partners who report being jointly responsible for earning income are more likely to be satisfied with their current financial situation. h4: couples with higher levels of agreement on spending are more likely to be satisfied with their current financial situation. h5: couples with higher levels of agreement on saving are more likely to be satisfied with their current financial situation. methodology data data for this study were collected by a private firm between december 2013 and january 2014 through amazon’s mechanical turk platform. data were analyzed as a secondary dataset. the target population for the sample included people in the united states who were currently living with a romantic partner in which at least one partner makes financial decisions for the couple and at least one partner is responsible for earning income. participants who were not currently living with a significant other were excluded. little’s (1988) t and chi-square tests were used to determine if missing data related to values of other variables in the dataset (sheskin, 2020). in rare cases of missing data, the omissions were determined to be random. the “linear trend at point” method of missing data replacement, a procedure in spss that regresses the existing series on an index variable scaled from 1 to n, was used to replace the missing values. the resulting sample comprised 602 individuals living with a romantic partner. outcome variable the outcome variable of interest in this study was financial satisfaction, measured using a singleitem likert agreement measure that was converted into a dummy variable. participants were asked to rate their level of agreement with the statement, “i am satisfied with my current financial situation.” responses indicating agreement or strong agreement were coded as 1, and 0 otherwise. the choice to dichotomize the variable was made to delineate those who were financially satisfied from those who were less satisfied. based on the coding, approximately 49% of participants reported agreeing or strongly agreeing with the statement. couple-level characteristics financial decision-making style was measured using the following question: "who is responsible for the majority of the financial decisions and management made in your household?" responses were converted into dummy variables as follows: “i am responsible” was coded as 1, and 0 otherwise; “my spouse is responsible” was coded as 1, and 0 otherwise; “my spouse/significant other and i are jointly responsible” (reference item) was coded as 1, and 0 otherwise. income pooling style was assessed by asking study participants to complete this statement: “my spouse/significant other and i ______.” combining all finances with a spouse or partner was coded as 1, and 0 otherwise. keeping some finances separate was coded as 1, and 0 otherwise. keeping finances entirely separate (reference category) was coded as 1, and 0 otherwise. responsibility for earning income was measured by asking, "who is responsible for working kruger et al. 270 (generating income)?" the variable was converted into dummy variables as follows: a participant who was responsible for working was coded as 1, and 0 otherwise; a spouse responsible for working was coded as 1, and 0 otherwise; being jointly responsible (reference category) was coded as 1, and 0 otherwise 0. each couple’s level of agreement on spending was measured by agreement with the following statement: "my spouse/significant other and i agree on issues related to spending money." the variable was converted into two binary variables. a response of agree or strongly agree was coded as 1, and 0 otherwise. a response of neither agree nor disagree, disagree, or strongly disagree (reference category) was coded as 1, and 0 otherwise. each couple’s level of agreement on saving was measured based on agreement with the following statement: "my spouse/significant other and i agree on issues related to saving money." responses were converted into two binary variables: agree or strongly agree was coded as 1, and 0 otherwise; neither agree nor disagree, disagree, or strongly disagree (reference category) was coded as 1, and 0 otherwise. control variables to measure financial strain, two proxies were used. first, unemployment was assessed and coded as 1 if the participant was currently unemployed, and 0 otherwise. second, financial stress was indicated by answers to the following question: “i/we often take money out of savings to pay bills.” responses were measured on a fivepoint likert scale where 1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree. to measure attitudes, risk tolerance was measured using the following question: “how often have others described you as a risk-taker?” respondents answered using a five-point likerttype scale where 1 = never, 2 = rarely, 3 = sometimes, 4 = often, and 5 = very often/always. for financial behavior, spending capacity was measured using the following question: “in the past, how often have you had difficulty spending less than your household earns?” answers were coded as 1 = never, 2 = several times per year, 3 = once a month, 4 = several times a month, 5 = once a week, 6 = several times a week, and 7 = almost every day. a dummy variable was created for which those who reported never having difficulty spending less than their household earns were coded as 1, and 0 otherwise. holding credit card debt was measured using the following question: “how much credit card debt do you/does your household currently have? please round to the nearest dollar amount.” the variable was recoded into a dummy variable where those who reported having $0 in credit card debt were coded as 1, and 0 otherwise. lastly, net worth served as a proxy for solvency. a positive net worth where household assets (cash, investments) exceeded liabilities (debt) was coded as 1, and 0 otherwise. subjective financial knowledge was measured with the following items using a five-point agreement scale: (a) i have explained financial concepts to others in the past, (b) i consider myself a novice or beginner when it comes to managing household finances, (c) i often have to ask others to explain financial terminology, (d) i am very comfortable explaining financial terminology, (e) i enjoy learning about financial concepts and terminology, (f) others seek my advice regarding personal financial matters, (g) i consider myself an expert at managing household finances, and (h) i enjoy thinking about/talking about financial matters. two of the questions were reverse coded so that higher scores on the scale indicated an elevated level of financial knowledge. the reliability of the scale was measured at a cronbach's alpha level of .91. the vector of personal characteristics included several variables. marital status was measured categorically using six classifications: (a) single, never married; (b) married, never divorced; (c) remarried; (d) widowed; (e) divorced; and (f) separated. four dummy coded variables were created for the analysis: (a) single, never married (reference), (b) married, never divorced, (c) remarried, and (d) widowed, divorced, or separated were coded as 1, and 0 otherwise. gender was measured categorically with selfdescribed females coded as 1, and 0 otherwise. education level was measured by asking participants to indicate the highest level of education they had obtained. a dummy variable, financial services review, 31(4) 271 bachelor’s degree or higher, was coded as 1 for respondents with a bachelor’s degree or higher, and 0 otherwise. race/ethnicity was measured categorically using only three dummies due to low variability: (a) white (reference), (b) black or african american, and (c) other. age and income each were measured as a continuous variable. study participants also were asked, "please estimate the approximate total income of your household before taxes last year. include income from earnings (e.g., wages, business profits) and unearned income (passive income from investments such as stocks, bonds, and mutual funds)." income was log-transformed for the analysis. data analyses data were analyzed using spss 28.0. the research hypotheses were evaluated using a variety of parametric and non-parametric statistical tests. mean and frequency descriptive statistics were calculated for the continuous and categorical variables, respectively. tests of the research hypotheses were first made using a chisquare analysis to identify significant differences in financial satisfaction across couple-level characteristics. two logit models were then estimated with financial satisfaction as the outcome variable. multivariate logistic regression was chosen for the analysis because of the binary nature of the dependent variable, the capacity of the model to handle a variety of control and independent variable types, and the ability to interpret results using the odds ratio. to address multicollinearity concerns, the level of agreement on spending and level of agreement on saving variables were analyzed in separate models. the models included the remaining couple-level characteristic variables and all control variables. the logit models were empirically modeled as follows: ln [ 𝑃(𝑌) 1−𝑃(𝑌) ] = 𝛽0 + ′𝛽1𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙𝑆𝑡𝑟𝑎𝑖𝑛1 + 𝛽2𝑅𝑖𝑠𝑘𝑇𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒2 + ′𝛽3𝐵𝑒ℎ𝑎𝑣𝑖𝑜𝑟𝑎𝑙3 + 𝛽4𝐾𝑛𝑜𝑤𝑙𝑒𝑑𝑔𝑒4 + ′𝛽5𝑃𝑒𝑟𝑠𝑜𝑛𝑎𝑙5 + ′𝛽6𝐶𝑙𝑖𝑒𝑛𝑡𝑙𝑒𝑣𝑒𝑙6 + 𝑒𝑖, (3) where ln [ 𝑃(𝑌) 1−𝑃(𝑌) ] is the odds of financial satisfaction, y is the binary outcome, ′𝛽1𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙𝑆𝑡𝑟𝑎𝑖𝑛1 is the vector of variables representing financial strain, 𝛽2𝑅𝑖𝑠𝑘𝑇𝑜𝑙𝑒𝑟𝑎𝑛𝑐𝑒2 is financial risk tolerance, ′𝛽3𝐵𝑒ℎ𝑎𝑣𝑖𝑜𝑟𝑎𝑙3 denotes the vector of financial behavior variables, 𝛽4𝐾𝑛𝑜𝑤𝑙𝑒𝑑𝑔𝑒4 is financial knowledge, ′𝛽5𝑃𝑒𝑟𝑠𝑜𝑛𝑎𝑙5 is the vector of personal characteristic variables, and ′𝛽6𝐶𝑙𝑖𝑒𝑛𝑡𝑙𝑒𝑣𝑒𝑙6 represents the vector of variables representing couple-level characteristics. results table 1 shows the descriptive statistics of the categorial variables used in the study. a small proportion of the sample had never been married but were in a committed relationship and living with their partner. approximately 60% of study participants were male, had completed at least a bachelor’s degree, and had a positive net worth. most were employed at the time of the survey. kruger et al. 272 table 1. descriptive statistics for the categorical variables (n = 602) variable freq. participant is satisfied with their current financial situation strongly disagree 9.5% disagree 27.5% neither agree nor disagree 13.8% agree 35.2% strongly agree 14.0% independent variables financial decision-making responsibilities both partners make financial decisions jointly 57.8% participant makes financial decisions 38.2% partner makes financial decisions 4.0% income earning responsibilities both partners earn income 62.1% participant earns income 25.2% partner earns income 12.7% household financial integration style couple combines finances 65.3% couple keeps some finances separate 25.9% couple keeps all finances separate 8.8% spouse and participant agree on issues related to spending money strongly disagree 2.2% disagree 13.5% neither agree nor disagree 17.4% agree 50.5% strongly agree 16.4% spouse and participant agree on issues related to saving money strongly disagree 2.2% disagree 8.8% neither agree nor disagree 11.6% agree 56.5% strongly agree 20.9% control variables (personal characteristics, financial stress, attitudes, and behavior marital status single, never married 15.3% other marital status 84.7% gender male 60.0% female 40.0% financial services review, 31(4) 273 table 1 (continued). descriptive statistics for the categorical variables (n = 602) variable freq. education level completed an associate's degree or lower 36.7% completed a bachelor's degree or higher 63.3% race white 90.8% black or african american 4.5% other race 4.7% employment status unemployed 4.2% employed 95.8% financial stress strongly disagree that they often take money out of savings to pay bills 33.7% disagree that they often take money out of savings to pay bills 40.3% neither agree nor disagree that they often take money out of savings to pay bills 11.6% agree that they often take money out of savings to pay bills 13.1% strongly agree that they often take money out of savings to pay bills 1.3% risk tolerance others would never describe as a risk-taker 29.9% others would rarely describe as a risk-taker 39.6% others would sometimes describe as a risk-taker 22.9% others would often describe as a risk-taker 6.3% others would very often/always describe as a risk-taker 1.3% credit card behavior do not carry a credit card balance 44.9% carry a credit card balance 55.1% spending behavior never has difficulty spending less than income 36.0% has difficulty spending less than income 64.0% net worth positive net worth 62.1% negative net worth 23.4% zero net worth 14.5% table 2 presents the descriptive statistics for the continuous variables. the average age of study participants was slightly under 38 years with an average annual income of approximately $104,000. scores measuring subjective financial knowledge ranged from the lowest possible score of eight to the highest possible score of 40, with an average score of about 28. kruger et al. 274 table 2. descriptive statistics for the continuous variables (n = 602) variable mean standard deviation minimum maximum age 37.6 10.3 20 74 pre-tax household income $103,899.48 $100,710.29 $800.00 $1,000,000 financial acumen scale score 27.9 7.0 8 40 the first research hypothesis was assessed using a series of chi-square tests. the chi-square analysis revealed significant differences in financial satisfaction in groups with divergent couple-level characteristics. table 3 shows the results of the chi-square tests of independence (*p < .05, **p < .01, ***p < .001). each of the five couple-level characteristics of interest (i.e., financial integration style, financial decisionmaking style, income-earning style, agreement on spending, and agreement on saving) was significantly associated with financial satisfaction. those who were financially satisfied were more likely to report combining their finances with their partner, having the participant be primarily responsible for making financial decisions, being jointly responsible or having the participant be primarily responsible for earning income, agreeing on spending, and agreeing on saving. those who were less financially satisfied were more likely to have separate finances or some separate finances, be jointly responsible or have their partner be responsible for financial decisions, have their partner be primarily responsible for earning income, not agree on spending, and not agree on saving. two logit models were estimated to evaluate the remaining research hypotheses, both controlled for the same variables. model 1 (agree on spending) included all couple-level characteristic variables except the “agree on saving” variable. model 2 (agree on saving) included all couplelevel characteristic variables except the “agree on spending” variable. table 4 shows the results. except for the financial stress variable, the coefficient directions and levels of significance for each model were the same. both models had relatively high explained variance, although model 1 had a slightly higher pseudo-r-squared compared to model 2 (nagelkerke r2 = .534 versus nagelkerke r2 = .521). income was positively associated with financial satisfaction in both models. subjective financial knowledge was also positively associated with reports of financial satisfaction (p < .001). for each additional point scored on the subjective financial knowledge measure, the odds of being financially satisfied increased in model 1 by 8% and in model 2 by 9%. financial stress was significantly associated with financial satisfaction in the agree-on-saving model (p < .05), in which those who were financially stressed were 23% less likely to be financially satisfied; however, financial stress was not significant in the agree-on-spending model (p = .06). paying with credit card balances in full each month (p < .05) and net worth (p < .01) were significantly associated with being financially satisfied in both models. in comparison to those who did not have a positive net worth, those who did were 127% more likely to be financially satisfied in the agree-on-spending model and 110% more likely in the agree-on-saving model. financial services review, 31(4) 275 table 3. differences in financial satisfaction by couple-level financial characteristics not satisfied satisfied χ2 separate finances count 40 13 21.56*** expected count 26.9 26.1 std. residual 2.5 -2.6 some separate finances count 90 66 expected count 79.3 76.7 std. residual 1.2 -1.2 combined finances count 176 217 expected count 199.8 193.2 std. residual -1.7 1.7 participant responsible for financial decisions count 103 127 13.94*** expected count 116.9 113.1 std. residual -1.3 1.3 jointly responsible for financial decisions count 183 165 expected count 176.9 171.1 std. residual 0.5 -0.5 participant's partner responsible for financial decisions count 20 4 expected count 12.2 11.8 std. residual 2.2 -2.3 participant responsible for earning income count 64 88 17.11*** expected count 77.3 74.7 std. residual -1.5 1.5 jointly responsible for earning income count 188 186 expected count 190.1 183.9 std. residual -0.2 0.2 participant's partner responsible for earning income count 54 22 expected count 38.6 37.4 std. residual 2.5 -2.5 do not agree on spending count 138 62 39.56*** expected count 101.7 98.3 std. residual 3.6 -3.7 agree on spending count 168 234 expected count 204.3 197.7 std. residual -2.5 2.6 do not agree on saving count 96 41 26.28*** expected count 69.6 67.4 std. residual 3.2 -3.2 agree on saving count 210 255 expected count 236.4 228.6 std. residual -1.7 1.7 financial services review, 31(4) 276 of the couple-level characteristics, only financial integration style was associated with financial satisfaction (p < .05). neither financial decisionmaking responsibility allocation nor incomeearning responsibility allocation was significantly associated with financial satisfaction. in comparison with those who had completely separate finances, those who had completely combined finances were 1.60 times more likely to be financially satisfied, accounting for agreement on spending, and 1.79 times more likely to be financially satisfied, accounting for agreement on saving. agreement on spending money was associated with financial satisfaction (p < .001). those who reported that they and their partner agree on spending were 61% more likely to be financially satisfied, compared to those who strongly disagreed. agreement on saving money also was associated with financial satisfaction (p < .01). those who reported that they and their partner agree on issues related to saving money were 33% more likely to be financially satisfied than those who disagreed. discussion the results from the statistical tests provide support for the first (i.e., partners who report combining their finances are more likely to be satisfied with their current financial situation), fourth (i.e., couples with higher levels of agreement on spending are more likely to be satisfied with their current financial situation), and fifth (i.e., couples with higher levels of agreement on saving are more likely to be satisfied with their current financial situation) hypotheses. the two regressions were similar, with one main difference. in model 1, where agreement on spending was included as an explanatory variable, financial stress was not significant; however, in model 2, where agreement on saving was included as an explanatory variable, financial stress was significant (p < .05). it is important to note that although the estimations were statistically different for each model, the p-values and association with financial satisfaction for each model were similar (i.e., agree-on-spending model, p = .06, and 21% less likely to be financially satisfied; agree-on-saving model, p = .04, and 23% less likely to be financially satisfied). however, as the threshold for a type-i error was set at p < .05, financial stress was deemed significant in the second model only. more empirical work is needed to understand how financial stress may be associated with financial satisfaction when controlling for couple’s financial characteristics. an important finding from this study is that a couple’s level of agreement on spending is important in describing financial satisfaction. participants who strongly agreed that they or their partner agree on issues related to spending money were much more likely to report being financially satisfied, compared to similar participants who reported disagreeing with this statement. partners who strongly agreed on issues related to saving money also were more likely to report being financially satisfied, compared to those who disagreed with this notion. agreement on spending exhibited a larger effect size than agreement on saving, perhaps because the consequences of disagreeing on spending are felt more immediately than are disagreements about saving. study participants who completely combined finances with their partner were more likely to report being financially satisfied, compared to those who kept their finances completely separate. no differences in financial satisfaction based on responsibility for income earning or responsibility for financial decisions and management were observed. it appears that viewing money as a household good—as a combined asset—is more important in terms of financial satisfaction than how money management responsibilities and tasks are divided. this insight provides support for conclusions made by skogrand et al. (2011) and ward and lynch (2018). skogrand et al. (2011) noted how important trust and communication are in the financial management process. when interpreting their findings and the results from the current study, it is important to remember that conclusions are applicable only in the context of healthy and non-abusive relationships where finances are not used to control or manipulate a partner. financial services review, 31(4) 277 table 4. logistic regression analyses (n = 602) variable b se odds ratio b se odds ratio agreement on spending model agreement on saving model independent variables (couple-level characteristics) participant responsible for financial decisions (ref: joint decisions) -0.135 0.25 0.874 -0.210 0.25 .811 participant's partner responsible for financial decisions (ref: joint decisions) -1.219 0.71 .296 -1.200 0.70 .301 combined finances (ref: completely separate finances) 0.956* 0.43 2.601 1.026* 0.43 2.789 some separate finances (ref: completely separate finances) 0.713 0.44 2.041 0.746 0.44 2.108 participant responsible for earning income (ref: joint income) 0.203 0.28 1.224 0.221 0.28 1.247 participant's partner responsible for earning income (ref: joint income) -0.045 0.39 .956 -0.088 0.39 .916 agree on spending 0.476*** 0.13 1.610 agree on saving 0.286* 0.14 1.331 control variables (personal characteristics, knowledge, financial strain, attitudes, and behavior) single, never married 0.093 0.35 1.097 0.053 0.35 1.055 female -0.150 0.27 .861 -0.131 0.27 .877 bachelor's degree or higher -0.420 0.25 .657 -0.394 0.25 .674 black or african american (ref: white) 0.365 0.54 1.441 0.370 0.53 1.447 other race (ref: white) 0.001 0.53 1.001 0.008 0.52 1.008 age -0.008 0.01 .992 -0.006 0.01 .994 income (log transformed) 0.878*** 0.20 2.407 0.849*** 0.20 2.336 subjective financial knowledge 0.081*** 0.02 1.084 0.082*** 0.02 1.086 unemployed -0.428 0.61 0.652 -0.322 0.60 .725 financial stress -0.240 0.13 .786 -0.266* 0.13 .766 risk tolerance 0.075 0.13 1.078 0.101 0.13 1.106 paying credit card balance in full 0.559* 0.23 1.749 0.552* 0.23 1.737 never overspend 0.876*** 0.26 2.401 0.913*** 0.26 2.492 positive net worth 0.820** 0.26 2.272 0.742** 0.26 2.099 constant -14.781 2.36 -13.930 2.32 notes: *<.05; **<.01; ***<.001; model 1: nagelkerke pseudo r-squared = .534; model 2: nagelkerke pseudo r-squared = .521. financial services review, 31(4) 278 implications the financial literature indicates that higher levels of financial satisfaction are associated with lower levels of marital discord (e.g., betcher & macauley, 1990; dew, 2007) and divorce (e.g., amato & rogers, 1997; grable et al., 2007; hill et al., 2017; zagorsky, 2003). findings from this study provide a strategic pathway for financial counselors looking to mitigate the negative impacts of financial dissatisfaction for their clients. one way to do this is to align a couple’s shared values and behaviors. as noted by britt et al. (2008), agreement about spending and saving is related to financial satisfaction. in this regard, financial counselors should be purposeful when assisting their clients in maximizing life satisfaction and happiness (britt et al., 2017). for example, they should evaluate the couple’s money-related habits and behaviors to determine the root causes of any conflict. when building and delivering financial action plans to these clients, financial counselors can recommend ways to align spending and saving goals. results from this study also indicate that it may be appropriate to introduce strategies designed to increase goal congruence related to spending and saving as a way to improve overall financial satisfaction. doing so should increase confidence in each partner so that, if desired, the couple can begin to combine finances to a greater extent, further enhancing financial satisfaction. limitations and future directions although the findings from this study advance the literature on the relationship between financial satisfaction and shared financial goals and values, certain limitations need to be acknowledged. first, the analyses utilized a secondary dataset, which means that some variables had to be approximated or excluded from the models due to a lack of available proxies (e.g., self-reported relationship status, relationship satisfaction). additionally, data were collected online and thus could be vulnerable to a response bias. data were cross-sectional, so the potential for dual causality cannot be ruled out. future studies using longitudinal data would be useful in helping researchers determine the direction of certain relationships and in making causal inferences. in addition, the data were collected at the individual level, which means that responses signified only one person's opinions. responses from both partners would have allowed comparisons between partners, as well as comparisons between each study participant's perception of their own and their partner's perceptions. additionally, it was not known how many times a participant had been married or involved in a committed romantic relationship or whether the participant was currently in a blended relationship. studies designed to replicate this research should take steps to account for these situations. finally, several studies (e.g., archuleta, 2013; archuleta et al., 2011, 2013) have shown that financial satisfaction is related to relational and marital satisfaction. although it was not possible to account for marital or relational satisfaction in this study, future studies should control for and investigate the possibility that relational and marital satisfaction are associated with financial satisfaction. references aboagye, j., & jung, j. y. (2018). debt holding, financial behavior, and financial satisfaction. financial counseling and planning, 29(2), 208-218. doi.10.1891/1052-3073.29.2.208 amato, p. r., & rogers, s. j. (1997). a longitudinal study of marital problems and subsequent divorce. journal of marriage and the family, 612-624. archuleta, k. l. (2013). couples, money, and expectations: negotiating financial management roles to increase relationship satisfaction. marriage & family review, 49(5), 391-411. doi.org/10.1080/01494929.2013.766296 archuleta, k. l., britt, s. l., tonn, t. j., & grable, j. e. (2011). financial satisfaction and financial stressors in marital satisfaction. psychological reports, 108(2), 563-576. doi.org/10.2466/07.21.pr0.108.2.563576 archuleta, k. l., grable, j. e., & britt, s. l. (2013). financial and relationship satisfaction as a function of harsh startup and shared goals and values. https://psycnet.apa.org/doi/10.1080/01494929.2013.766296 https://psycnet.apa.org/doi/10.2466/07.21.pr0.108.2.563-576 https://psycnet.apa.org/doi/10.2466/07.21.pr0.108.2.563-576 financial services review, 31(4) 279 financial counseling and planning, 24(1), 3-14. atlas, s. a., lu, j., micu, p. d., & porto, n. (2019). financial knowledge, confidence, credit use, and financial satisfaction. journal of financial counseling and planning, 30(2), 175190. doi.10.1891/1052-3073.30.2.175 betcher, w., & macauley, r. (1990). the seven basic quarrels of marriage: recognize, defuse, negotiate, and resolve your conflicts. villard books. britt, s., grable, j. e., goff, b. s. n., & white, m. (2008). the influence of perceived spending behaviors on relationship satisfaction. financial counseling and planning, 19(1), 31-43. britt, s. l., hill, e. j., lebaron, a., lawson, d. r., & bean, r. a. (2017). tightwads and spenders: predicting financial conflict in couple relationships. journal of financial planning, 30(5), 36-42. campbell, a. (1981). the sense of well-being in america: recent patterns and trends. mcgraw-hill. cfpb. (2015). financial well-being: the goal of financial education. consumer financial protection bureau, washington, dc. https://files.consumerfinance.gov/f/2015 01_cfpb_report_financial-well-being.pdf dew, j. (2007). two sides of the same coin? the differing roles of assets and consumer debt in marriage. journal of family and economic issues, 28(1), 89–104. doi.10.1007/s10834-006-9051-6 dew, j. p., & stewart, r. (2012). a financial issue, a relationship issue, or both? examining the predictors of marital financial conflict. journal of financial therapy, 3(1), 43-61. https://newprairiepress.org/jft/vol3/iss1/ 4/ easterlin, r. a. (2006). life cycle happiness and its sources: intersections of psychology, economics, and demography. journal of economic psychology, 27(4), 463-482. doi.org/10.1016/j.joep.2006.05.002 eirich, g. m., & robinson, j. h. (2017). does earning more than your spouse increase your financial satisfaction? a comparison of men and women in the united states, 1982 to 2012. journal of family issues, 38(17), 2371–2399. doi.org/10.1177/0192513x16638384 fan, l., & babiarz, p. (2019). the determinants of subjective financial satisfaction and the moderating roles of gender and marital status. family and consumer sciences research journal, 47(3), 237259. doi.org/10.1111/fcsr.12297 garrett, s., & james, r. n. (2013). financial ratios and perceived household financial satisfaction. journal of financial therapy, 4(1), 39-62. doi.org/10.4148/jft.v4i1.1839 george, l. k. (1992). economic status and subjective well-being: a review of the literature and an agenda for future research. in n. e. cutler, d. w. gregg, & m. p. lawton (eds.), aging, money, and life satisfaction: aspects of financial gerontology (pp. 69-99). springer. godwin, d. d. (1994). antecedents and consequences of newlyweds' cash flow management. financial counseling and planning, 5, 161-190. grable, j. e., britt, s., & cantrell, j. (2007). an exploratory study of the role financial satisfaction has on the thought of subsequent divorce. family and consumer sciences research journal, 36(2), 130-150. doi.org/10.1177/1077727x07309284 grable, j., kruger, m., byram, j-l., & kwak, e. j. (2021). perceptions of a partner’s spending and saving behavior and financial satisfaction. journal of financial therapy, 12(1). https://newprairiepress.org/jft/vol12/iss1 /3/ hansen, t., slagsvold, b., & moum, t. (2008). financial satisfaction in old age: a http://dx.doi.org/10.1891/1052-3073.30.2.175 https://psycnet.apa.org/doi/10.1016/j.joep.2006.05.002 https://psycnet.apa.org/doi/10.1177/0192513x16638384 https://doi.org/10.1111/fcsr.12297 https://doi.org/10.4148/jft.v4i1.1839 https://psycnet.apa.org/doi/10.1177/1077727x07309284 kruger et al. 280 satisfaction paradox or a result of accumulated wealth? social indicators research, 89(2), 323-347. doi.org/10.1007/s11205-007-9234-z hill, e. j., allsop, d. b., lebaron, a. b., & bean, r. a. (2017). how do money, sex, and stress influence marital instability? journal of financial therapy, 8(1), 2142. doi.org/10.4148/1944-9771.1135 hira, t. k., & mugenda, o. m. (1998). predictors of financial satisfaction: differences between retirees and non-retirees. financial counseling and planning, 9(2), 75-84. https://www.afcpe.org/news-andpublications/journal-of-financialcounseling-and-planning/volume-92/predictors-of-financial-satisfactiondifferences-between-retirees-and-nonretirees/ hsieh, c. m. (2001). correlates of financial satisfaction. the international journal of aging and human development, 52(2), 135-153. doi.org/10.2190/9yde-46pamv9c-2jrb hsieh, c. m. (2004). income and financial satisfaction among older adults in the united states. social indicators research, 66(3), 249-266. http://www.jstor.org/stable/27522072 joo, s. (1998). personal financial wellness and worker job productivity. [doctoral dissertation, virginia polytechnic institute and state university]. joo, s. (2008). personal financial wellness. in j. j. xiao (ed.), handbook of consumer finance research (pp. 21-33). springer science + business media. joo, s. h., & grable, j. e. (2004). an exploratory framework of the determinants of financial satisfaction. journal of family and economic issues, 25(1), 25-50. doi.org/10.1023/b:jeei.0000016722.37 994.9f kelley, h. h., chandler, a. b., li, x., curran, m. a., & yorgason, j. b. (2022). spenders and tightwads among newlyweds: perceptions of partner financial behaviors and relational well-being. journal of financial therapy, 13(1), 3. doi.org/10.4148/1944-9771.1288 koochel, e. e., markham, m. s., crawford, d. w., & archuleta, k. l. (2020). financial transparency scale: its development and potential uses. journal of financial counseling and planning, 31(1), 14-27. https://connect.springerpub.com/content /sgrjfcp li, x., curran, m. a., lebaron-black, a. b., jorgensen, b., yorgason, j., & wilmarth, m. j. (2023). couple-level attachment styles, finances, and marital satisfaction: mediational analyses among young adult newlywed couples. journal of family and economic issues, 44(1), 125-142. doi.org/10.1007/s10834-021-09808-x little, r. j. (1988). a test of missing completely at random for multivariate data with missing values. journal of the american statistical association, 83(404), 11981202. mao, d. m., danes, s. m., serido, j., & shim, s. (2017). financial influences impacting young adults’ relationship satisfaction: personal management quality, perceived partner behavior, and perceived financial mutuality. journal of financial therapy, 8(2), 24-41. https://newprairiepress.org/jft/vol8/iss2/ 3/ mugenda, o. m., hira, t. k., & fanslow, a. m. (1990). assessing the causal relationship among communication, money management practices, satisfaction with financial status, and satisfaction with quality of life. lifestyles: family and economic issues, 11(4), 343-360. doi.org/10.1007/bf00987345 plagnol, a. c. (2011). financial satisfaction over the life course: the influence of assets and liabilities. journal of economic psychology, 32(1), 45-64. doi.org/10.1016/j.joep.2010.10.006 https://doi.org/10.4148/1944-9771.1135 https://doi.org/10.2190/9yde-46pa-mv9c-2jrb https://doi.org/10.2190/9yde-46pa-mv9c-2jrb https://doi.org/10.4148/1944-9771.1288 https://psycnet.apa.org/doi/10.1016/j.joep.2010.10.006 financial services review, 31(4) 281 porter, n. m. (1990). testing a model of financial well-being. [doctoral dissertation, virginia polytechnic institute and state university]. robb, c. a., & woodyard, a. s. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning, 22(1), 60-70. http://afcpe.org ross, d. b., gale, j., wickrama, k., goetz, j., vowels, m. j., & tang, y. (2021). couple perceptions as mediators between family economic strain and marital quality: evidence from longitudinal dyadic data. journal of financial counseling and planning, 32(1), 158172. doi. 10.1891/jfcp-18-00065 sheskin, d. j. (2020). handbook of parametric and nonparametric statistical procedures (4th ed.). chapman & hall/crc. skogrand, l., johnson, a. c., horrocks, a. m., & defrain, j. (2011). financial management practices of couples with great marriages. journal of family and economic issues, 32(1), 27-35. doi. 10.1007/s10834-010-9195-2 spuhler, b. k., & dew, j. (2019). sound financial management and happiness: economic pressure and relationship satisfaction as mediators. journal of financial counseling and planning, 30(2), 157174. https://connect.springerpub.com/content /sgrjfcp sumarwan, u., & hira, t. k. (1993). the effects of perceived locus of control and perceived income adequacy on satisfaction with financial status of rural households. journal of family and economic issues, 14(4), 343-364. doi.org/10.1007/bf01013984 tharp, d. t. (2017). three essays on personality characteristics and financial satisfaction. [doctoral dissertation, kansas state university]. totenhagen, c. j., wilmarth, m. j., serido, j., curran, m. a., & shim, s. (2019). pathways from financial knowledge to relationship satisfaction: the roles of financial behaviors, perceived shared financial values with the romantic partner, and debt. journal of family and economic issues, 40(3), 423-437. doi.org/10.1007/s10834-019-09611-9 van de venter, g., michayluk, d., & davey, g. (2012). a longitudinal study of financial risk tolerance. journal of economic psychology, 33(4), 794-800. doi.org/10.1016/j.joep.2012.03.001 vera-toscano, e., ateca-amestoy, v., & serrano-del-rosal, r. (2006). building financial satisfaction. social indicators research, 77(2), 211-243. doi.org/10.1007/s11205-005-2614-3 ward, a. f., & lynch jr, j. g. (2018). on a needto-know basis: how the distribution of responsibility between couples shapes financial literacy and financial outcomes. journal of consumer research, 45(5), 1013-1036. doi.org/10.1093/jcr/ucy037 woodyard, a. s., & robb, c. a. (2016). consideration of financial satisfaction: what consumers know, feel and do from a financial perspective. journal of financial therapy, 7(2), 41-61. https://newprairiepress.org/jft/vol7/iss2/ 4/ xiao, j. j., chen, c., & chen, f. (2014). consumer financial capability and financial satisfaction. social indicators research, 118(1), 415-432. doi.org/10.1007/s11205-013-0414-8 xiao, j. j., sorhaindo, b., & garman, e. t. (2006). financial behaviours of consumers in credit counselling. international journal of consumer studies, 30(2), 108-121. doi.org/10.1111/j.14706431.2005.00455.x xiao, j. j., tang, c., & shim, s. (2009). acting for happiness: financial behavior and life satisfaction of college students. social https://econpapers.repec.org/scripts/redir.pf?u=https%3a%2f%2fdoi.org%2f10.1007%252fs10834-010-9195-2;h=repec:kap:jfamec:v:32:y:2011:i:1:p:27-35 https://psycnet.apa.org/doi/10.1007/bf01013984 https://psycnet.apa.org/doi/10.1007/s10834-019-09611-9 https://doi.org/10.1016/j.joep.2012.03.001 https://psycnet.apa.org/doi/10.1007/s11205-005-2614-3 https://psycnet.apa.org/doi/10.1093/jcr/ucy037 https://doi.org/10.1111/j.1470-6431.2005.00455.x https://doi.org/10.1111/j.1470-6431.2005.00455.x kruger et al. 282 indicators research, 92(1), 53-68. doi.org/10.1007/s11205-008-9288-6 zagorsky, j. l. (2003). husbands’ and wives’ view of the family finances. the journal of socio-economics, 32(2), 127-146. zurlo, k. a. (2009). personal attributes and the financial well-being of older adults: the effects of control beliefs. parc working paper series, wps 09-03. https://repository.upenn.edu/parc_worki ng_papers/27/ choosing between value and growth in mutual fund investing glenn pettengilla,*, george changa, c. james huengb adepartment of finance, grand valley state university, 1 campus drive, allendale, mi 49401-9403, usa bdepartment of economics, western michigan university, 1903 w. michigan avenue, kalamazoo, mi 49008-5330, usa abstract this article informs investors on the choice between value and growth mutual funds. the wellestablished value premium demonstrates that, on average, value securities outperform growth securities, suggesting that an investor may be wise to choose value funds. extant studies, however, suggest that growth funds outperform value funds. we show that value funds indeed outperform growth funds especially in terms of lower realized risk and higher realized terminal wealth, leading us to recommend value funds over growth funds. we argue that previous findings result from a bias against value in some multifactor models. © 2014 academy of financial services. all rights reserved. jel classification: g00; g10; g12; g19 keywords: value premium; mutual fund performance; value funds; growth funds; mutual fund �s 1. introduction the well-known value premium argues that value securities, securities with high bookto-market ratios or low price-to-earnings ratios, outperform other securities when raw returns or returns adjusted only for market risk are considered. there is a substantial literature to support this finding, beginning with basu (1983) studying the relationship between priceto-earnings ratios and returns, and with rosenburg, reid, and lanstein (1985) studying the relationship between book-to-market ratios and returns. these results have been confirmed * corresponding author. tel.: �1-616-331-7430; fax: �1-616-331-7445. e-mail address: pettengg@gvsu.edu (g. pettengill) pettengill gratefully acknowledges support from the e. seidman chair. financial services review 23 (2014) 341–359 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. in numerous other studies including the discipline-altering study of fama and french (1992). not only has the value premium been persistently reported, it is also large. from july 1926 through december 2012 this premium has averaged an annualized value of 6% based on data from kenneth french’s web site. because of the strong evidence of outperformance by value securities, individual investors and financial planners may wish to consider a preference for value mutual funds relative to growth mutual funds. indeed, writing in forbes, clash (1998) argues that investors interested in small-company stocks should invest in small-company value mutual funds on the basis of the superior performance of the russell 2000 value index versus the russell 2000 growth index. clash suggests that small-firm investors might use index funds to operationalize his recommendation. two cautions may be appropriate for financial planners and individual investors considering clash’s recommendations. first, the component stocks in the index funds do not perfectly match the component stocks of the portfolios formed in empirical studies that provide evidence of a value premium. second, superior performance by an index of value stocks relative to an index of growth stocks does not necessarily indicate superior performance by managed value funds relative to managed growth funds, and many individual investors and planners prefer to invest in managed funds. empirical studies heighten the concern that managed value funds might not display the superior performance suggested by the value premium. piotroski (2000) reports that the value premium results from a subset of value securities and that most value securities underperform the market. piotroski suggests that with the use of readily available accounting information one can identify those value securities that will outperform.1 however, are managers of value funds successful in identifying those value securities that will outperform? extant studies comparing the performance of value and growth mutual funds estimate � using the fama-french three-factor model and conclude that growth funds outperform value funds. this finding suggests the inability of value fund managers to identify those value securities that will outperform. in this article, however, we present new empirical evidence supporting the presumption that investors may exploit the value premium through mutual fund purchases especially for small-firm funds and index funds. our conclusion is, in large part, driven by the lower realized risk of value fund portfolios. we argue that previous finding of superior performance of growth funds results from a bias in estimating mutual fund �s using the three-factor model. in the next section we discuss the findings of previous studies on the relative performance of value and growth mutual funds. in section 3, we describe our sample and methodology. we present empirical results in section 4 in three parts: first, we compare traditional risk and return measures; second, we examine the implications of these results for an investor’s end wealth, finally, we reconcile our results with previous studies. in the final section, we offer our conclusion. 2. previous studies a considerable literature seeks to inform investors on choices relative to the wide array of mutual fund investment opportunities. most of the literature concentrates on measuring the 342 g. pettengill et al. / financial services review 23 (2014) 341–359 ability of mutual fund managers to “beat the market,” to earn a return greater than justified by the mutual fund’s risk. from jensen (1968) to fama and french (2010) these studies have generally concluded that fund managers do not possess market-beating stock selection ability. these conclusions are generally based on asset pricing models that find, on average, negative �s for mutual fund portfolios using the single-factor capital asset pricing model in earlier studies and a multifactor model such as the fama-french three-factor model in later studies. writing in this journal, betker and sheehan (2013) suggest that practitioners still tend to use single factor models despite the dominant use of multifactor models in academic studies. in this article we compare the performance of value and growth mutual funds rather than examine the overall market performance of mutual funds. our contribution is to suggest that investors and financial planners can benefit from the use of value funds relative to growth funds even though recent studies, calculating �s using multifactor models, suggest the opposite. before presenting our results we summarize the relatively few previous studies that have specifically compared the performance of value and growth mutual funds.2 shi and seiler (2002), using the morningstar classification matrix, build portfolios of value or growth funds that are subdivided into size classification of large, medium, and small firms for a total of six portfolios each containing 30 randomly selected mutual funds. returns are calculated for each portfolio by equally weighting the returns of the funds over the 10-year period, august 31, 1989 to august 31, 1999. shi and seiler find that the average return is higher for the growth mutual fund in each of the three categories and is significantly higher for the large-firm classification. they find that in all three size categories, risk is significantly higher for the growth mutual fund sample relative to the value mutual fund sample. they calculate the sharpe ratio (using semivariance) for each of the size categories.3 they find that the reward-to-risk ratio is higher for the growth firms in the large funds, but is higher for the value funds in the mid and small cap firms. they conclude that their evidence allows no definitive recommendation for investing in growth versus value mutual funds, and indicate that results may vary over other time periods. it is worthwhile to note that their sample covers a significant portion of the dot.com bubble, where one might expect growth mutual funds to do particularly well. davis (2001) compares the performance of growth and value mutual funds using data from the crsp file for the period 1962 through 1998. applying the fama-french three-factor model he identifies funds as value (small) or growth (large) funds by their respective factor loading on the hml (smb) factor. ten mutual fund portfolios are created from the crsp data set with a univariate sort on the loadings on the hml factor. performance is measured for each portfolio by finding the equally weighted average of the �s of each fund within the portfolio across the sample period. portfolios formed from funds with a low loading on the hml factor are considered growth portfolios whereas portfolios formed from funds with a high loading on the hml factor are considered value portfolios. in general, �s are positive for the portfolios of growth mutual funds and negative for the portfolios containing value mutual funds. the portfolio containing the mutual funds with the highest loading on the hml factor has a significantly negative �. in a bivariate sort, davis creates nine portfolios based on the value and growth classification, and based on the size classification of large, medium, and small firms. each of the 343g. pettengill et al. / financial services review 23 (2014) 341–359 classifications is based on the factor loadings on the hml and smb factors. as in the univariate sort, �s are determined for each of the nine portfolios. for all size classifications the value portfolio has a negative � and the growth portfolio has a positive �. davis (2001, p. 25) concludes that: “perhaps the biggest disappointment in the past three decades is the inability (or unwillingness) of funds to capture the value premium that has been observed in common stock returns during the period.” we suggest and argue more fully below that the value premium may exist but is obscured by a bias against value inherent in the fama-french three-factor model. chan, chen, and lakonishok (2002) calculate average returns for a sample of mutual funds using morningstar data from january 1979 to december 1997. although their article explores a wide variety of issues, they do compare the performance of value and growth mutual funds. consistent with davis they identify value and growth funds on the basis of the funds loading on the hml factor. they also compare value and growth funds for three classifications of size: large cap, mid cap, and small cap. they calculate � by regressing mutual fund returns against the three-factor model and find the average � across fund classifications. consistent with davis’ findings, for each of the size categories the � for the growth funds is greater than the � for the value funds. thus, in contrast to the expectations that an investor might hold given the existence of the value premium, chan et al. find that growth mutual funds outperform value mutual funds. in this study, in contrast to previous findings, we provide new evidence that supports the proposition that investors may benefit from selecting value mutual funds over growth mutual funds. as described in the next section, our results are based on a larger sample size and a longer sample period than those used in previous studies. 3. sample and methodology the purpose of our article is to provide new information for investors choosing between value and growth mutual funds. to do so we make empirical comparisons between the performance of growth and value funds for five separate classifications. first, we gather total index values for the russell 2000 value and growth indexes, and for the russell 1000 growth and value indexes, directly from russell investments. data are collected over the period january 1979 (the inception of the russell indices) through december 2012. we use the total index values, which include dividend payments to calculate monthly returns. on the basis that value and growth index funds, including etfs, are widely available and are able to reasonably track index returns, we treat the returns to the russell indices as proxies for returns to index funds. we recognize that such funds were not available from the start of the index data, but such funds are surely available to current investors.4 to compare the performance of managed growth and value funds, we acquire monthly returns for value and growth mutual funds from the morningstar database. we obtain data from the inception of this database through december 2012. morningstar classifies general purpose equity funds as value (growth) if the fund has low (high) valuations (i.e., low [high] price to earnings ratios and high [low] dividend yields) and slow (fast) growth (i.e., low [high] growth rates for earnings, sales, book value, and cash flow). general purpose funds not 344 g. pettengill et al. / financial services review 23 (2014) 341–359 meeting these categories are classified as blended funds. morningstar also characterizes funds by size as small, mid, and large-firm funds. we compare value and growth funds within each of these three size categories. to provide consistency over time, we include in our sample only those funds that have the same morningstar classification throughout the sample period. for each of the three size classifications, we create a value superfund containing all of the value mutual funds available in a given month and a growth superfund containing all of the growth mutual funds available in a given month. we determine the return to each superfund as the equally weighted return of all funds within the superfund. to ensure statistical accuracy, we require that there are at least 30 individual mutual funds for both value and growth funds before we include either superfund return in our sample. across all three size groups there are more growth funds than value funds. thus, for each size grouping our sample begins when there are a sufficient number of value funds and continues through december 2012. subject to this restriction, the small-firm sample begins in january 1993; the mid-firm sample begins in july 1993; and the large-firm sample begins in august 1984. there are 118,539 (120,375) monthly observations for small-firm (midsize) growth funds and 44,673 (47,921) monthly observations for small-firm (midsize) value funds. the sample size is much higher for the large firm funds both because of the larger sample period and the larger number of observations each month. there are 185,026 monthly observations for the large-firm value funds; and 245,994 monthly observations for the large-firm growth funds. to summarize, our sample data allow comparisons between value and growth index funds using the russell 2000 for small firms and russell 1000 for large firms as proxies for index funds. for these comparisons we have data from january 1979 through 2012. to compare the performance of managed value and growth funds we collect data from the morningstar data base to build portfolios of value and growth mutual funds. these portfolios are subdivided into small-firm, midsized firm, and large-firm portfolios. our sample period for these comparisons all begin later than the beginning of the sample period for the index fund portfolios. all sample periods end at december 2012. as shown in the next section these comparisons strongly favor the selection of value funds. 4. empirical results in this section we report on the relative performance of value and growth mutual funds across five matching data sets. in the first set of comparisons, we examine risk and return summary measures. we compare average monthly returns, total risk based on realized monthly return variations, and the risk-return tradeoff using the sharpe ratio. because mutual fund investors are ultimately interested in wealth creation, and because of potential bias with the arithmetic mean in estimating end wealth, for each of the 10 portfolios we calculate end wealth from an initial investment of $10,000 at the start of the sample period. based on the calculated end wealth, we determine the geometric mean return for all 10 portfolios. results from these tests support the supposition that investors can benefit from the selection of value funds relative to growth funds. because these results are at variance with previous studies, in the final segment of this section we rationalize the difference between our findings and previous findings. 345g. pettengill et al. / financial services review 23 (2014) 341–359 4.1. comparisons of monthly returns: arithmetic mean, return variance, sharpe ratio comparisons of monthly mean returns, variance of monthly mean returns and sharpe ratios across the five matched portfolios are shown in table 1. in four of the five comparisons the average return is greater for the value portfolio. none of the comparisons, however, provide a statistically significant difference in mean returns between the value and growth mutual fund portfolio. consistent with clash’s (1998) recommendation, the largest return difference between value and growth is found in the small-firm index funds as proxied by the russell 2000 index returns. the small-firm annualized fund provides a difference in monthly return that annualizes to �2.5% over the period january 1979 through december 2012. this annualized difference is certainly a considerable amount if compounded over a long-term investment period. however, this difference is less than the amount of the value premium reported on the ken french web site,5 suggesting that the composition of value and growth index funds do not match the composition of portfolios used to measure the value premium in academic studies. thus, although it appears that investors are able to benefit from the value premium by choosing value index funds over growth index funds, the difference in the returns from these funds is not as great as suggested by the value premium reported in academic studies. the difference between value and growth in the russell 1000 (large-firm) index is less table 1 monthly percentage returns: value vs. growth mean variance sharpe ratio (a) index funds russell 2000 value 1.169 25.99 22.93% russell 2000 growth 0.966 45.09 14.39% test-statistic 1.245† 1.729‡ 3.153§ (p-value) (0.106) (0.000)* (0.002) russell 1000 value 1.040 18.61 24.11% russell 1000 growth 0.970 25.85 19.08% test-statistic 0.511† 1.388‡ 1.654§ (p-value) (0.305) (0.000) (0.098) (b) morningstar funds managed small value 0.899 23.649 18.09% managed small growth 0.834 41.552 12.54% test-statistic 0.276† 1.757‡ 1.438§ (p-value) (0.391) (0.000) (0.075) managed mid value 0.815 20.541 17.58% managed mid growth 0.814 36.001 13.16% test-statistic 0.005† 1.753‡ 1.157§ (p-value) (0.498) (0.000) (0.161) managed large value 0.811 16.122 20.19% managed large growth 0.838 24.894 16.80% test-statistic �0.204† 1.544‡ 1.157§ (p-value) ( � 0.5) (0.000) (0.124) * a p-value of 0.000 indicates that the p-value is nonzero, but smaller than 0.0005. † h0: �value � �growth; ha: �value � �growth. ‡ h0: �value 2 � �growth 2 ; ha: �value 2 � �growth 2 . § h0: srvalue � srgrowth; ha: srvalue � srgrowth. the statistical test is based on opdyke (2007). 346 g. pettengill et al. / financial services review 23 (2014) 341–359 than the difference found in the russell 2000 (small-firm) index, suggesting that an investor seeking to exploit the value premium may wish to do so by investing in small firm securities. this difference across firm size is also evident in the managed mutual funds. the excess return of value versus growth is greatest for the small-firm managed mutual funds. for midsize managed funds the returns are almost identical for value and growth; and for the large-firm managed funds the mean return is actually larger for growth rather than value. the value investor does better concentrating on small firms. in addition to the difference in performance of value and growth across firm size, the advantage of value over growth varies greatly between index funds and managed funds. for small-firm index funds the annualized value premium is 2.46% whereas that for the managed funds is less than 0.78%. the annualized difference for the large-firm index funds is 0.84% in favor of value while the large-firm managed growth actually has a slightly higher average return than does the large-firm managed value funds.6 thus, our earlier warning concerning piotroski’s finding that most value securities underperform the market seems to have merit. value fund managers are not able to outperform growth funds by the same margin as do the index funds.7 of course, this finding could result from the relative skills of growth fund managers. still, overall value funds on the whole have higher reported mean returns than growth funds. investors are concerned about both risk and return. the value premium emphasizes the benefit of higher average return from investing in value, but value has a much stronger advantage relative to growth in terms of lower risk within our sample. as shown in table 1, in every comparison the risk, as measured by variance in monthly returns, is higher for the growth portfolios. as with returns, the advantage with regard to risk for value investing is strongest for small-firm securities. for both the managed portfolios and the index portfolios the variance in monthly returns in the small-firm growth portfolio is almost twice the variance for the value portfolio. for the other managed portfolios the monthly return variance is more than 50% higher than that for the growth portfolios. in all cases, the monthly return variance is significantly lower for the value portfolio with a p-value � 0.000. because risk is always significantly lower for value and because return is generally higher, in meanvariance criterion value beats growth! consistent with this observation, as shown in table 1, the sharpe ratio measuring the risk-reward tradeoff is always higher for the value portfolio. the advantage of value over growth is primarily because of lower risk. therefore, one might argue that investors tend to be more concerned with return than risk as measured by variance in return. however, we assert that financial planners guiding investors are not as concerned with monthly average returns as they are concerned with the end wealth of the investors. in that regard, as we show in the next section, the lower variance of value portfolios provides a significant advantage in terms of achieving end wealth goals. 4.2. end wealth comparisons in the previous section we have made five sets of comparison between value and growth mutual funds finding generally higher means and significantly lower variance in returns for the value funds. these results lead us to argue that investors ought to prefer value funds relative to growth funds. we recognize, however, that investors may be more concerned with 347g. pettengill et al. / financial services review 23 (2014) 341–359 return than risk. we further argue that the ultimate concern for investors and financial planners is end wealth. investors are concerned about the amount that they will have in their retirement account or any other account when they plan to use it. financial planners know that determination of this amount relies on calculation with geometric means rather than arithmetic means. financial calculators use the geometric mean to calculate end wealth of a lump sum investment or periodic investments as in an annuity. therefore, the crucial consideration should not be the arithmetic mean as is generally reported by mutual funds and as reported in table 1. because end wealth is the ultimate concern for an investor, the geometric mean is more relevant in comparing the times series returns of two investment alternatives than the arithmetic mean. the difference between the spread of the arithmetic means and the spread of the geometric means between two return time series provides important implications for an investor. as shown below, there is a direct connection between return variability and the difference in the spreads of arithmetic and geometric means of return time series. a lower variability in returns results in less variability in the base that determines returns used to calculate arithmetic mean returns. thus, a time series return with low variability will have a smaller difference between arithmetic mean and geometric mean than will a series with a higher variability in returns. given the direct association between the geometric mean and end wealth, we assert that the lower variation in returns of the value funds provide these funds with a critical advantage in providing higher end wealth to the investor. the advantage in end wealth for the value funds is greater than one would predict using the arithmetic means. to illustrate the impact of return variation on end wealth, in table 2 we present a hypothetical example of return streams for investment a and investment b. assume that investment a has a return series that consists of returns of �10%, 10%, and 3%; and investment b has a return series that consists of returns of �50%, 50%, and 21%. the return series for a has an arithmetic mean return of 1% and a variance of 0.69% squared. the return series for b has an arithmetic mean return of 7% and a sd of 17.65% squared. the higher arithmetic mean return for series b inaccurately suggests a greater return for investing in b at the cost of a greater risk. in fact, if $10,000 was invested in asset a, the end wealth would be $10,197. a similar investment in asset b would result in an end wealth of $9,075. the comparative geometric means of 0.65% for asset a and �3.18% for asset b more correctly reflect the wealth creation experienced by the investor. note, the end wealth is not affected by the ordering of the return series. the reason investment b has a lower end wealth despite table 2 bias in comparative arithmetic mean from differences in return variation year investment a investment b return end wealth return end wealth 0 na $10,000 na $10,000 1 �10% $9000 �50% $5,000 2 10% $9,900 50% $7,500 3 3% $10,197 21% $9,075 arithmetic mean 1% 7% variance 0.69% sq. 17.65% sq. geometric mean 0.65% �3.18% 348 g. pettengill et al. / financial services review 23 (2014) 341–359 having a greater arithmetic mean is because of the larger return variation in investment b. the greater variation in the returns for investment b causes greater fluctuations in the base used to calculate returns. as a result, the calculated arithmetic mean return for investment b will be upwardly biased, which results in a misleading indication concerning the relative wealth creation between investment a and investment b. the comparative geometric means correctly predict the difference in end wealth.8 the same relationship illustrated by the hypothetical investment a and investment b exists in the very real comparisons between value and growth mutual funds. the higher variation in returns for the growth funds causes a greater fluctuation in the base value used to calculate returns for the growth funds. this results in an upward bias in any end wealth prediction for growth funds versus value funds using the arithmetic mean. thus, we calculate end wealth for each of the 10 funds in our sample and then calculate geometric means to provide comparisons between value and growth investing. for each of the five value funds and each of the five growth funds we invest $10,000 at the first of the month for which data becomes available. thus, for the index funds the investment begins january 1979, and for the managed funds the investment begins august 1984 for the large-firm funds, january 1993 for the small-firm funds, and july 1993 for the midsize firm funds. in each case we keep the proceeds fully invested through december 2012. for all five comparisons the end wealth in the value fund is greater than the end wealth in the growth fund. not surprisingly the biggest difference in end wealth occurs for the small-firm funds and the index funds. as shown in table 3, the investment of $10,000 in the russell 2000 value fund accrues to $668,991 as compared with an end wealth of $198,328 for the same investment in the russell 2000 growth fund. any investor who heeded clash’s advice to invest in a small-firm value index fund at the beginning of our sample period would have surely have been pleased with the choice of value over growth. the value investor has 3.37 times as much end wealth as the growth investor. surely, any investor would consider this difference significant. unfortunately, we know of no statistical test to confirm a significant difference in end wealth or corresponding geometric means. the end wealth for the russell 1000 value fund is more than 50% larger than for the russell 1000 growth fund. furthermore, the end wealth for the table 3 end wealth, geometric means, and arithmetic means: value vs. growth fund type end wealth value end wealth growth geometric value geometric growth arithmetic value arithmetic growth russell 2000 $668,991 $198,328 1.04% 0.73% 1.17% 0.92% russell 1000 $464,694 $302,601 0.95% 0.84% 1.04% 0.97% small managed $61,772 $42,073 0.76% 0.60% 0.90% 0.83% mid managed $50,965 $41,381 0.69% 0.61% 0.82% 0.81% large managed $118,992 $112,095 0.73% 0.71% 0.81% 0.84% the geometric and arithmetic means are calculated on a monthly basis. end wealth results from the investment of $10,000 in each fund beginning with the first month of data availability and ending december 2012. for the russell funds the sample period begins january 1979. the beginning of the data for the managed firms is: august 1984 for the large-firm portfolio, january 1993 for the small-firm portfolio and july 1993 for the mid-sized portfolio. 349g. pettengill et al. / financial services review 23 (2014) 341–359 small-firm managed value fund is nearly 50% higher than the end wealth for the small-firm managed growth fund. figs. 1 and 2 show the growth in wealth for value funds relative to growth funds for the small-firm index funds and the small-firm managed funds, respectively. the relative growth $10,000 $110,000 $210,000 $310,000 $410,000 $510,000 $610,000 $710,000 ja n79 ja n81 ja n83 ja n85 ja n87 ja n89 ja n91 ja n93 ja n95 ja n97 ja n99 ja n01 ja n03 ja n05 ja n07 ja n09 ja n11 ja n13 growth value fig. 1. increase in invested wealth: russell 2000 growth index versus russell 2000 value index (january 1979 through december 2012). $10,000 $20,000 $30,000 $40,000 $50,000 $60,000 $70,000 d ec -9 2 d ec -9 3 d ec -9 4 d ec -9 5 d ec -9 6 d ec -9 7 d ec -9 8 d ec -9 9 d ec -0 0 d ec -0 1 d ec -0 2 d ec -0 3 d ec -0 4 d ec -0 5 d ec -0 6 d ec -0 7 d ec -0 8 d ec -0 9 d ec -1 0 d ec -1 1 d ec -1 2 growth value fig. 2. increase in invested wealth: portfolio of small-firm managed growth funds versus portfolio of small-firm managed value funds (january 1993 through december 2012) 350 g. pettengill et al. / financial services review 23 (2014) 341–359 in wealth varies over time. for example, growth funds did very well in the 1990s during the dot.com bubble. indeed, the invested wealth for the small-firm managed growth funds exceeds the invested wealth of the small-firm managed value funds throughout the 1990s. however, that advantage was more than offset during the following decade. the small-firm index fund comparison experiences a similar pattern, but the longer investment history for this comparison results in the invested amount for the value fund always being greater than the invested amount held in the growth fund. in both comparisons we include a decade where growth funds experienced unusually high returns and value funds continue to outperform growth funds in terms of accumulated wealth. for all five comparisons, the higher end wealth in the value funds corresponds to a higher geometric mean. the difference in the geometric mean, which accurately measures end wealth considerations, between value and growth funds is always larger than the difference found in the arithmetic means. the geometric mean reflects the lower return variance for the value fund and the arithmetic mean does not. we make two observations with regard to this difference. for the small-firm index funds, using actual returns shows the end wealth of the value portfolio to be 3.37 times the end wealth of the growth portfolio. applying the geometric mean duplicates this difference. if, however, one would use the monthly arithmetic mean to compound wealth, the resulting end wealth of the value portfolio is only 2.26 times that of the growth portfolio. comparisons of arithmetic mean returns do not fully represent the difference in end wealth, which should be the main concern for the mutual fund investor. a second observation underlying the bias found in comparisons of arithmetic means is provided by the large-firm managed funds. the monthly arithmetic mean is larger for the managed large-firm growth fund than for the managed large-firm value fund. consequently, one might expect end wealth for an investment in managed large-firm growth funds to be greater than for a similar investment in managed large-firm value funds. as shown in table 3, this is not the case. the end wealth for the value fund is greater than the end wealth for the growth fund, albeit by a modest amount. because the arithmetic mean is higher for the growth fund, predictions of end wealth using the arithmetic mean would suggest a greater end wealth for the growth fund. however, forecasts using the geometric mean would correctly predict a higher end wealth for the value fund as the geometric mean of the managed large-firm fund portfolio is greater for the value fund than for the growth fund. crucially, end wealth comparisons suggest that investors ought to choose value over growth. this conclusion is at variance with studies reviewed earlier in the article and in the next section we rationalize this difference. 4.3. reconciling current and previous results we have made five sets of comparisons between value and growth funds with comparisons divided between index and managed funds and by firm size. in four of the five comparisons the average return is greater for value, but in none of the five comparisons is the difference statistically significant. in all cases value funds have significantly lower variance in monthly returns. thus, on the basis of mean-variance comparisons value funds must be judged superior to growth funds. we also show that the lower variance is associated with higher end wealth for value funds relative to growth funds. thus, the superior performance 351g. pettengill et al. / financial services review 23 (2014) 341–359 for value funds is driven in large part by the significantly lower variation in returns, that is, lower total realized risk. as cited above, chan et al. (2002) and davis (2001) reach the conclusion that growth beats value by examining risk-adjusted returns (�s) calculated using the fama-french three-factor model. their conclusions depend on the three-factor model’s measurement of systematic risk. this places our argument in the seemingly dubious position of preferring measures of total risk rather than systematic risk. we defend this position with two arguments. first, we argue that on a realized basis, total risk rather than systematic risk is the appropriate measure of risk for any investor.9 second, we argue that risk-adjusted returns calculated using the fama-french three-factor model are biased against value portfolios and biased in favor of growth portfolios. our position preferring total risk versus systematic risk seems at odds with modern portfolio theory. a basic insight of this theory is that only systematic risk, measured by factor loadings on systematic risk factors, should be considered in measuring risk of securities considered for inclusion into a well-diversified portfolio. total risk is deemed an inappropriate measure because the idiosyncratic portion of total risk will be eliminated in a well-diversified portfolio. however, how do investors judge the performance of a portfolio held over time? are investors concerned about the systematic risk in a portfolio as measured by its factor loadings? or, are investors concerned about the amount of return variation realized? this question is paramount to asking whether investors are concerned about returns predicted by a model or realized returns. obviously, investors care about realized returns. we submit that likewise investors are concerned about realized risk rather than expected risk. realized risk is appropriately measured by variation in returns. because realized risk is the ultimate concern of investors, portfolio systematic risk measurements are useful in determining expected returns and risk-adjusted returns (�) only if these measures correlate with variations in return.10 a corollary to this relationship is that an estimate of risk-adjusted return must be based on the absolute value of a factor loading because it is the absolute value of the factor loading that determines variation in return in response to a systematic factor. for ease of illustration we provide an example using the well-known capital asset pricing model (capm), a single factor asset-pricing model where the only systematic risk is the market risk. consider a portfolio long in the market index. this portfolio would have a � of 1. now, consider a portfolio short in the market index. this portfolio would have a � of �1. these two portfolios would have exactly the same realized risk. the variation in returns, thus the risk, would be the same for these two portfolios. only the sign of the returns would be different between the two portfolios. the relative realized returns between the two portfolios would depend on whether market return is on average positive or negative. although a short position is generally required to have a negative market �, assets may have positive or negative �s for their loading on the hml factor of the three-factor model. in fact, value portfolios load positively and growth portfolios load negatively on the hml factor. the positive loading by value portfolios on the hml factor will increase expected return for these portfolios and consequently decrease the estimated �s. in contrast, the negative loading by growth portfolios on the hml factor will decrease expected return for these portfolios and consequently increase the estimated �s. indeed this relationship is why previous studies measuring �s using the fama-french three-factor model find growth funds 352 g. pettengill et al. / financial services review 23 (2014) 341–359 outperforming value funds while our results clearly show that the opposite is true. in the appendix we discuss in detail this bias against value that creates the erroneous conclusion that growth beats value. 5. conclusion in this article we provide investors with information concerning the relative performance of value and growth funds. the well-known value premium suggests that fund investors ought to favor value over growth. we make five sets of comparisons. we use data from the russell 2000 and russell 1000 indexes to proxy the performance of value and growth index funds. additionally, we gather data from the morningstar database to compare the performance of managed mutual funds across three size classification. in four of the five comparisons, value funds have higher average returns, but in none of the comparisons is the difference statistically significant. comparisons of realized risk, however, are always statistically significant and always favor value fund. thus, on the basis of superior performance in terms of mean-variance analysis, our study suggests that fund investors ought to indeed favor value. in addition, we show that the end wealth for investing using historic returns is always higher for value funds than for growth funds and that this difference is greater than which would be computed using arithmetic means. these results are because of the significantly lower variance in returns of the value funds relative to growth funds. although we argue that mutual fund investors ought to favor value funds, consistent with the well-documented value premium, we find that the benefit to investing in value funds is less than would be expected by the size of the value premium reported in empirical studies. further, we find that the benefit of value investing is greater for index funds than for managed funds and that the benefit of value investing is greater for small-firm funds than for large-firm funds. indeed, there is very little difference observed in our sample between the end wealth of managed large-firm value funds and managed large-firm growth funds. finally, we show that the findings of previous studies that suggest investors would find better performance in growth funds results from an intrinsic bias in multifactor models against value portfolios. notes 1 piotroski finds that the use of nine accounting data points allows the selection of outperforming value securities over the period 1976 through 1996. woodley, jones, and reburn (2011), however, find that piotroski’s methodology is not successful over the period 1997 through 2008. 2 classification of funds between value and growth funds is a relatively new phenomenon. early studies such as carhart (1997) and gruber (1996), when comparing across types of mutual funds, use various categories of growth funds. for example, carhart compares aggressive growth, long-term growth, and growth and income funds. 3 shi and seiler indicate that, unfortunately, the sharpe ratio allows no statistical comparison. since the publication of their article, opdyke (2007) has developed a test for comparing sharpe ratios, which we use in this article. 353g. pettengill et al. / financial services review 23 (2014) 341–359 4 for example vanguard offers index mutual funds tracking the russell 1000 and 2000 value and growth indices. among the etf funds tracking the russell indices are the ishares russell 2000 value etf and the ishares russell 2000 growth etf. 5 for the period january 1979 through december 2012, the value premium is an annualized 5.25% according to the ken french web site. 6 as noted in our sample description the sample time period differs for the index funds and managed funds. 7 our results suggest that value fund managers are inefficient in selecting stocks within the value sector. these results are consistent with the finding of brooks and porter (2012) that from 1994 through 2005, mutual fund managers lost potential gain from sector selection because of poor stock selection. 8 this relationship is also illustrated by the historic comparisons between large-firm equity returns and small-firm equity returns as reported in the ibbotson ssbi classic yearbook (2013). ibbotson reports that the annual arithmetic mean is 16.5% for smallfirm equities and 11.8% for large firm equities. the higher return for the small-firm equities is associated with a much higher sd in annual returns: 32.3% for small-firm equities and 20.2% for large-firm equities. because of this disparity in variability of returns, the difference in arithmetic means provides a poor gauge for the difference in wealth creation from investing in the two assets. this difference is appropriately measured by the geometric means that are reported to be much closer by ibbotson. the geometric mean for small-firm equities is 11.9% and the geometric mean for large-firm equities is 9.8%. 9 in support of this position, we note that reports of historic risk-return tradeoffs by assets classes, such as found in ibbotson’s yearbooks, use measure of realized risk such as sd of annual returns. 10 we conclude that value portfolios are less risky than growth portfolios based on realized risk. although the portfolios used in our sample are sufficiently large to eliminate firm specific risk the question may arise as to how value or growth would covary with more general portfolios. however, to measure the general impact of covariance with other securities, surely the most appropriate measure is covariance with the market measured by market �. as reported in the appendix, in all five comparisons that we make between value and growth portfolios, the loading on the market factor is higher for growth than for value. on this basis, covariance considerations would reinforce the conclusion that value is less risky than growth. 11 we do not assert that this bias is the sole reason for the difference in relative total risk and �s between the value and growth portfolios. we note that in all five comparisons, the growth portfolios have a higher market risk, but this should be reflected in both total risk and �s. part of the difference in total risk may be because of a systematic risk factor not captured by the three-factor model that affects growth portfolios more than value portfolios. of course, there are a number of extensions of the three-factor model that seek to include other risk factors. one should also consider the possibility that growth securities include more idiosyncratic risk. consistent with recent studies (see, e.g., fu [2009]), idiosyncratic risk may not be completely eliminated in the portfolios that we investigate. 354 g. pettengill et al. / financial services review 23 (2014) 341–359 12 a similar bias exists against small-firm portfolios, but small-firms securities have more total risk than large-firm securities. indeed, betkar and sheehan (2013) find the largest difference between the �s computed by single factor and multi-factor models to occur for small-firm value portfolios. 13 complete results for all regressions are available from the authors upon request. appendix fama and french (1992) observe the value premium and assume that value firms have higher returns as compensation for higher risk. subsequently, fama and french (1993, 1996) develop the widely used fama-french three-factor model, shown below in eq. (1), which adjusts returns for this presumed risk. applying this model, chan et al., (2002) and davis (2001) conclude that growth outperforms value. our findings, however, dispute the assumption within the three-factor model that value is riskier than growth and hence dispute the conclusion reached by these studies. we will show that this bias against value when using the three-factor model creates the erroneous conclusion that growth beats value.11 according to the fama-french three-factor model, the exposure of security i to systematic risk is measured by the security’s loadings on: market excess return (rm � rf); a portfolio long in value securities and short in growth securities (hml); and a portfolio long in smallfirm securities and short in large-firm securities (smb): �ri � rf�t � �i � �mi�rm � rf�t � �hi�hml�t � �si�smb�t � �t . (1) by construction, this model assigns greater systematic risk to value portfolios because value portfolios tend to load positively on the hml factor. therefore, despite the lower realized risk that we document for value funds, when � is calculated for value funds using the three-factor model, value funds are penalized for presumed greater risk by the model’s construction. we do not deny that a security’s responsiveness to the hml factor creates return variation. however, this responsiveness should be, as argued in the body of this article, measured by the absolute value of the factor loading. posit a growth portfolio with a negative loading that has the same absolute value as a positive loading of a value portfolio. these two portfolios have exactly the same measured absolute responsiveness to the hml factor. thus, these two portfolios will experience exactly the same impact on return variation resulting from the hml factor. however, the impact from the hml factor on expected (required) return will be positive for the value portfolio and negative for the growth portfolio. thus, the impact from the hml factor on risk-adjusted return (�) will be positive for the growth portfolio and negative for the value portfolio. this differential impact provides a strong bias toward assigning superior performance to the growth portfolio. we illustrate this bias by calculating �s for our 10 sample portfolios using the systematic risk factors included in the three-factor model. we first use only the market and smb factors to calculate �s for the growth and value portfolios. we then recalculate �s using the hml factor as well. in table 4 we report the results using the russell 1000 value and growth index portfolios. 355g. pettengill et al. / financial services review 23 (2014) 341–359 as shown in panel a of table 4, if � is calculated using the market and smb factors alone, the estimated � is positive for the value index, albeit insignificantly different from zero. the � for the growth index is negative with p-value � 0.096. thus, based upon the results from applying the market and the smb factors, the comparative measures are consistent with the very real advantage of the value index. panel b of table 4 shows the impact of including the hml factor. there is little change in the factor loadings on the market and smb factors, but there is a dramatic change in the � for the value portfolio. as expected, the growth and value portfolios have opposite signs for the factor loadings on the hml factor. value loads positively on hml and growth loads negatively on hml. thus, the value portfolio, which had a positive � without the hml factor, now has a negative �. the growth portfolio, which had a negative � without the hml factor, now has a positive � with a p-value � 0.059. the switch from a positive � to a negative � for the value portfolio is directly related to the loading on the hml factor. the positive loading on the hml factor causes an increase in the expected return for the value portfolio, which decreases the risk-adjusted return (�). the switch from a negative � to a positive � for the growth portfolio is also directly related to the loading on the hml factor. the negative loading on the hml factor causes a decrease in the expected return for the growth portfolio, which increases the risk-adjusted return (�). in the case of the russell 1000, the loadings on the hml factor for the value and growth portfolios have approximately the same absolute values but with different signs. because of the difference in sign, risk-adjusted return is increased for growth but decreased for value. thus, the exposure to the hml factor would have the same impact on total risk for both portfolios, but would reduce the � for the value portfolio relative to the growth portfolio. application of the three-factor model biases results against value and in favor of growth.12 table 4 factor loadings and risk-adjusted returns: russell 1000 value and growth indices (january 1979 through december 2012) portfolio risk-adjusted return—�s (p-value) market � (p-value) smb � (p-value) hml � (p-value) (a) excess market and smb loadings growth �0.099% 1.193 �0.090 — (0.096) (0.000)* (0.000) value 0.092% 0.888 �0.053 — (0.235) (0.000) (0.037) (b) excess market, smb, and hml loadings growth 0.074% 1.036 �0.171 �0.325 (0.059) (0.000) (0.000) (0.000) value �0.079% 0.925 0.027 0.320 (0.215) (0.000) (0.209) (0.000) factor loadings result from regressing portfolio returns on the fama-french three-factor model, as shown below. the coefficients and p-values reported in panel a result from regressing returns only against the market and smb factor. the coefficients and p-values reported in panel b result from regressing returns against all three factors. �ri � rf�t � �i � �mi�rm � rf�t � �hi�hml�t � �si�smb�t � �t . * a p-value of 0.000 indicates that the p-value is nonzero, but smaller than 0.0005. 356 g. pettengill et al. / financial services review 23 (2014) 341–359 because our focus is on the impact of the hml factor, and in the interest of space, we limit our discussion on the other four groupings of funds to a summary of the changes in risk-adjusted returns from adding the hml factor to the market and the smb factors. table 5 reports these changes in �s and shows the loadings of the portfolios on the hml factor.13 as reported in table 5, when using only the market and the smb factors, � is negative for the growth portfolios in all five cases and is significantly negative at the 1% level for the large-firm morningstar sample. when the three-factor model is applied, the loading on the hml factor is negative in all but one case for the growth portfolio. in the cases where the loading on the hml factor is negative for the growth portfolios, the estimated � increases and becomes positive in two cases. the inclusion of the hml factors biases results in favor of growth portfolios. in all cases, when only the market and the smb factors are applied, the value portfolios have higher estimated �s than the growth portfolios. in four of the five cases, the estimated � for the value portfolios is positive, albeit in only one of the cases is � significantly positive. inclusion of the hml factor causes a radical shift in these comparisons. in all five cases the loading of the value portfolio on the hml factor is positive, reducing the � for the value portfolio. when the three-factor model is applied, in all five cases the estimated � for the table 5 influence of hml factor on comparative risk-adjusted returns: value and growth portfolios portfolio two-factor model three-factor model risk-adjusted return—�s (p-value) risk-adjusted return—�s (p-value) hml � (p-value) (a) russell 1000 growth �0.099% 0.074% �0.325 (0.096) (0.059) (0.000) value 0.092% �0.079% 0.320 (0.235) (0.215) (0.000) (b) russell 2000 growth �0.179% �0.138% �0.078 (0.140) (0.264) (0.071) value 0.022% �0.099% 0.598 (0.039) (0.140) (0.000) (c) managed small firms growth �0.020% �0.030% 0.004 (0.903) (0.895) (0.944) value 0.210% �0.090% 0.581 (0.129) (0.258) (0.000) (d) managed mid firms growth �0.060% 0.010% �0.151 (0.722) (0.942) (0.005) value 0.110% �0.090% 0.434 (0.382) (0.346) (0.000) (e) managed large firms growth �0.160% �0.060% �0.215 (0.009) (0.228) (0.000) value �0.040% �0.150% 0.243 (0.573) (0.020) (0.000) see notes on table 4. 357g. pettengill et al. / financial services review 23 (2014) 341–359 value portfolio is negative and less than the estimated � for the growth portfolio. the inclusion of the hml factors biases results against value portfolios. to the extent that systematic risk associates with variability in any factor, measurement of that factor risk ought to accurately predict that factor’s impact on realized risk. thus, if variability in a portfolio long in value stocks and short in growth stocks (hml portfolio) impacts portfolio risk, measurement of this impact ought to correctly predict the impact on realized risk. the construction of the three-factor model implies that exposure to the hml factor increases risk for portfolios with value securities and decreases risk for growth securities. however, we have shown that value portfolios whether indexed or managed have significantly less realized risk than growth portfolios. the application of the three-factor model produces this result because in this model the sign of the factor loading matters and it should not. it is the absolute value of the factor loading that impacts risk regardless of the sign. one should not use the three-factor model to make comparisons between value and growth portfolios because of the inherent bias against value. references basu, s. (1983). the relationship between earnings yield, market value and return for nyse common stocks: further evidence. journal of financial economics, 12, 129–156. betker, b., & sheehan, j. (2013). a comparison of single factor and multiple factor alphas used in measuring mutual fund performance. financial services review, 22, 349–365. brooks, l., & porter, g. (2012). mutual fund performance attribution 1994–2005. financial services review, 21, 259–273. carhart, m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. chan, k., chen, h., & lakonishok, j. (2002). on mutual fund investment styles. review of financial studies, 15, 1407–1437. clash, j. (1998). the wallflower strategy. forbes, february 9, 113–117. davis, j. (2001). mutual fund performance and manager style. financial analysts journal, 57, 19–27. fama, e., & french, k. (1992). the cross section of expected stock returns. the journal of finance, 47, 427–465. fama, e., & french, k. (1993). common risk factors in the returns on stocks and bonds. journal of financial economics, 33, 3–56. fama, e., & french, k. (1996). multifactor explanations for asset pricing anomalies. the journal of finance, 51, 55–84. fama, e., & french, k. (2010). luck versus skill in the cross-section of mutual fund returns. the journal of finance, 65, 1915–1947. fu, f. (2009). idiosyncratic risk and the cross-section of expected stock returns. journal of financial economics, 91, 24–37. gruber, m. (1996). another puzzle: the growth of actively managed mutual funds. journal of finance, 51, 783–810. haslem, j., baker, k., & smith, d. (2008). performance and characteristics of actively managed retail equity mutual funds with diverse expense ratios. financial services review, 17, 49–68. ibbotson sbbi classic yearbook. (2013). market results for stocks, bonds, bills, and inflation 1926–2012. chicago, il: morningstar, inc. jensen, m. (1968). the performance of mutual funds in the period 1945–1964. the journal of finance, 23, 389–416. opdyke, j. (2007). comparing sharpe ratios: so where are the p-values? journal of asset management, 8, 308–336. 358 g. pettengill et al. / financial services review 23 (2014) 341–359 piotroski, j. (2000). value investing: the use of historical financial statement information to separate winners from losers. journal of accounting research, 38, 1–41. rosenberg, b., reid, k., & lanstein r. (1985). persuasive evidence of market inefficiency. journal of portfolio management, 11, 9–17. shi, w., & seiler, m. (2002). growth and value style comparison of u.s. stock mutual funds. american business review, 20, 25–32. woodley, m., jones, s., & reburn, t. (2011). value stocks and accounting screens: has a good rule gone bad? journal of accounting and finance, 11, 87–104. 359g. pettengill et al. / financial services review 23 (2014) 341–359 a further examination of equity indexed annuities andy terrya,*, erick eldera acollege of business, university of arkansas at little rock, 2801 s. university, little rock, ar 72204, usa abstract equity indexed annuities (eias) are deferred annuities that credit interest according to a formula tied to the performance of an underlying equity index. this research expands previous research, particularly that of reichenstein (2009, 2011), by examining the distribution of returns that could have been created on a rolling monthly basis since 1928 for 11 through 15-year investment horizons. second, we examine investment alternatives that include the options imbedded in eias. third, rather than assuming constant cap rates we allow cap rates to vary with interest rates. we find that for long time horizons the opportunity costs of investing in eias is high. © 2015 academy of financial services. all rights reserved. jel classifications: g2; equity indexed annuities; investments keywords: indexed annuities; equity indexed annuities; investments 1. introduction equity indexed annuities, hereafter eias, are deferred annuities that credit interest according to a formula tied to the performance of an underlying equity index. academic research has been limited to explaining the complex financial product and calculating and comparing returns on eias with those of alternative investments. this research expands previous research, and particularly that of reichenstein (2009, 2011), in three ways. first, rather than examining limited sub-periods of returns data, we examine the distribution of returns that could have been created since 1928. second, unlike previous research, we also * corresponding author. tel.: �1-501-569-8872; fax: �1-501-683-7021. e-mail address: haterry@ualr.edu (a. terry) financial services review 24 (2015) 411–428 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. examine investment alternatives that include the options imbedded in eias. third, unlike prior research, we allow cap rates to vary with interest rates. we find that for long time horizons alternative strategies dominate eias as an investment vehicle. the article is organized as follows: section 2 discusses eias, their features, and the underlying economics from the issuing insurance company perspective. section 3 presents a summary literature review. section 4 expands the work of reichenstein (2009, 2011), whereas section 5 furthers prior research by examining investment alternatives with options. section 6 concludes the article. 2. eia’s background, contract features, and economics within the insurance industry deferred annuities are typically classified as fixed or variable. variable annuities have investment returns that vary depending on the performance of the underlying subaccount investments. the choices of subaccounts are very similar to the choices available for mutual funds. variable annuities also contain a “death benefit” where the beneficiary is guaranteed to receive at least the initial investment less any withdrawals. variable annuities are treated as securities and are regulated by the sec. fixed annuities are contracts wherein the insurance company guarantees fixed returns and payouts. during the accumulation period the contract earns interest at rates pre-specified by the insurance company or spelled out in the contract formula. they are very similar to guaranteed investment contracts (gics). unlike variable annuities, fixed annuities do not have subaccounts with underlying investments. fixed annuities have been around for quite some time, and most fixed annuities are not considered securities and are not regulated by the sec. within the fixed annuity classification are traditional fixed rate annuities and equityindexed annuities (eias).1 fixed rate annuities have a guaranteed minimum interest rate and a current rate. as the names imply, the guaranteed minimum is the lowest rate that will be credited to the contract value, whereas the current rate is usually adjusted annually based on market conditions and cannot be less than the guaranteed minimum. eias were originally introduced in 1995. eias differ from traditional fixed annuities in how interest is credited. for eias, the interest credited usually varies between zero and an upside amount that is determined by the contract formula, but that is linked to the performance, variously measured, of an underlying equity index, for example, s&p 500. eias are marketed as a way to participate in the market upside while being protected against the downside and while receiving a guaranteed minimum return.2 eias are complex financial products that typically bundle a deferred annuity with death benefits and annuity payout options. the number and variability of the features within a contract make them exceedingly complex. these features, including investment, regulatory, insurance, and tax attributes, have already been discussed at length in mccann (2008) and reichenstein (2009). consequently, only those aspects most pertinent to this research are repeated below.3 412 a. terry, e. elder / financial services review 24 (2015) 411–428 2.1. interest crediting formulas and limits in general interest is credited based on the percentage change in an underlying equity index. the percentage change reflects price appreciation only and does not include dividends. the two most common methods are point-to-point and monthly averaging. in the former, the percentage change between two points is calculated, typically on an annual basis, and the ending index value one period becomes the beginning index value for the next. the latter measures the percentage change in the index level from a starting date, usually the contract purchase date or anniversary date, to an average value over the subsequent period. typically the average is calculated on a monthly basis.4 this, and most of the prior research, examines annual reset, point-to-point interest calculations. contract owners generally do not receive an interest rate equal to the total percentage change in the index value, rather the rate credited is usually limited in one or more of several ways:5 most eias have a cap rate that limits the percentage change to be applied to the eia. for example, a 3% cap means that the interest credit will be limited to 3% of the eia contract value, regardless of how much the underlying equity index actually increased. for contracts with annual resets, insurance companies usually have the option to change the cap rates. contracts may also specify a minimum cap rate for the contract term. in addition to a cap rate, a participation rate indicates what percentage of the index increase will be used to calculate the interest credit. for example, if the participation rate is 80% and the percentage change in the index is 6%, then 4.8% (80% � 6%) would be the rate used to determine the interest credit. like cap rates, the initial participation rate is usually guaranteed and after the first year the insurance company has the option of changing the rate. with a spread or margin, the interest rate used to calculate the interest credit is determined by subtracting a “spread” or “margin” from the percentage change in the index. for example, if the spread is 2.5% and the percentage change in the underlying equity index value, however calculated, is 7%, then 4.5% (7% to 2.5%) would be used to determine the interest credit. some contracts specify a maximum spread and some contracts will combine a spread with a cap. finally, eias have a minimum crediting rate of 0%, which is one of the selling points. 2.2. contract values most contracts have three different values: contract (accumulation) value, guaranteed minimum value, and cash surrender value. the contract value is a notional amount equal to the original premium amount plus any vested bonus plus interest credited, less withdrawals. the guaranteed minimum value is calculated as a percentage of the initial premium, and increases each year by the guaranteed minimum interest rate specified in the contract. this table of values is calculated at the time of the contract and does not change. the cash surrender value is usually the greater of the contract value less surrender charges, or the guaranteed minimum value. surrender charges can start out as high as 20% and last 15 or more years. the actual cash surrender value over time will be higher than that included in the contract illustration page on contract date if the interest rate calculated from the index formula is greater than the guaranteed minimum interest rate. 413a. terry, e. elder / financial services review 24 (2015) 411–428 there are some very important things to note. first, in many cases a contract will have a combination of the above limits. for example, a contract could have both a 6% cap and a 1.6% margin in the first year. the length of the surrender period and surrender percentages will also vary with these interest credit limits, leading to a myriad of possible contract combinations. second, the caps, margins, and participation rates are typically guaranteed for only one year and then may be changed at the insurance company’s discretion (see the economics of eias below). third and really important is that there is a significant difference in the early years of a contract between the contract value, from which yields are calculated and reported to customers on annual statements, discussed in marketing efforts, and reported by vanderpal et al. (2011) and the cash surrender value, from which a true cash on cash yield should be calculated. the implication of this difference is that return comparisons should be made based on surrender values or should examine investment horizons greater than the surrender period. 2.3. insurance features insurance features have already been discussed at length in reichenstein (2009) and mccann (2008) and are not further discussed here. 2.4. the economics of eias below is an excerpt taken from american equity’s 2012 10-k report to shareholders that explains their business model: we specialize in the sale of individual annuities (primarily deferred annuities) and, to a lesser extent we also sell life insurance policies. under accounting principles generally accepted in the united states, or gaap, premium collections for deferred annuities are reported as deposit liabilities instead of as revenues. sources of revenues for products accounted for as deposit liabilities are net investment income, surrender charges deducted from the account balances of policyholders in connection with withdrawals, realized gains and losses on investments and changes in fair value of derivatives. components of expenses for products accounted for as deposit liabilities are interest credited to account balances, changes in fair value of embedded derivatives, amortization of deferred policy acquisition costs, other operating costs and expenses and income taxes. earnings from products accounted for as deposit liabilities are primarily generated from the excess of net investment income earned over the interest credited to the policyholder, or the “investment spread.” in the case of index annuities, the investment spread consists of net investment income in excess of the cost of the options purchased to fund the index-based component of the policyholder’s return and amounts credited as a result of minimum guarantees. (american equity 2012 10-k, pages 18–19) further, according to american equity’s 12/31/12 10-k, note 5 to the financial statements, “on the respective anniversary dates of the index policies, the index used to compute the annual index credit is reset and we purchase new one-year call options to fund the next annual index credit. we manage the cost of these purchases through the terms of our fixed index annuities, which permit us to change caps, participation rates, and/or asset fees, 414 a. terry, e. elder / financial services review 24 (2015) 411–428 subject to guaranteed minimums on each policy’s anniversary date. by adjusting caps, participation rates, or asset fees, we can generally manage option costs except in cases where the contractual features would prevent further modifications.” (italics added) american equity and other companies selling eias, collect premiums from customers, which they treat as deposit liabilities, as would a bank, and then invest the money. their investment portfolio consists mostly of highly rated corporate bonds, but also includes u.s. treasury and government agency securities, state and municipal obligations, and mortgagebacked securities. because equity securities are less than 0.5% of their portfolio, they buy one-year call options on the underlying indexes used in their eia crediting formulas. they make profit by earning more on their investments than they have to pay their customers. this difference is called an “investment spread.” as discussed above, the companies manage this investment spread by changing the caps, participation rates, and margin used to calculate the customers’ interest credit. if yields on the underlying investment portfolio fall, or if call option premiums increase due to an increase in index volatility, the insurance company will change the limits to reduce the amounts credited to the customers’ account values and maintain its spread between its “cost of money”—what it credits to the customers—and what it earns on their deposited funds.6 because of a change in accounting rules, companies like american equity now include the effect of derivatives’ gains and losses with the underlying costs being hedged. consequently, american equity since 2001 explicitly discloses their “yield earned,” their “cost of money,” and their “investment spread.” the “cost of money” represents what they have actually credited to their customers. there are a couple of significant points to note. first, these returns are lower than those reported in the vanderpal et al. (2011) studies discussed below because vanderpal’s returns are based on contract values credited over successive five-year periods, whereas the returns actually credited to customers in the aggregate reflect reductions because of surrender charges or not receiving a full period interest credit. second, it is important to note that customer returns depend more on bond yields than on underlying equity index values. as indicated in table 1, “cost of money” is below 3.5% in both 2010 and 2012 despite s&p 500 returns in excess of 10% both years. table 1 above indicates that the population of american equity’s customers could have performed about as well over the last 11 years by simply purchasing 10-year treasury securities. it is important to note that the credited rates in table 1 are for the customers as a group and do not reflect the expected or actual rates for a specific customer. rates for a specific customer will depend on that customer’s contract specifications and on whether the customer incurs any surrender charges. table 1 spreads of american equity annuities 2012 2011 2010 2009 2008 2007 2006 2005 2004 2003 2002 yield earned 5.28% 5.8% 6.06% 6.30% 6.20% 6.11% 6.14% 6.18% 6.28% 6.43% 6.91% cost of money 2.58% 2.77% 2.91% 3.26% 3.43% 3.51% 3.28% 3.38% 3.37% 3.46% 4.19% investment spread 2.70% 3.03% 3.15% 3.04% 2.77% 2.60% 2.86% 2.80% 2.91% 2.97% 2.72% 10-yr treasury yield 1.97% 3.36% 3.85% 2.46% 3.91% 4.68% 4.37% 4.23% 4.38% 4.07% 5.2% source: american equity 10-k’s; u.s. department of treasury, beginning of year 10-year yield. 415a. terry, e. elder / financial services review 24 (2015) 411–428 reichenstein (2009) says “since, by design, indexed annuities cannot add value through security selection, all eias must produce risk-adjusted returns that trail those offered by readily available marketable securities by their spread, that is by their expenses including transaction costs … we do not need empirical tests to state definitively that indexed annuities do not offer competitive risk-adjusted returns. their structure ensures this is the outcome.” his essential point is that indexed annuities are simply repackaging returns that are already available to investors in the market place without adding any potential security selection or market timing value. the cost of this repackaging is the “spread.” in summary, the simple economics of eias is that investors are paying 2–3% annually in investor spreads to receive returns similar to those already available in the market, trivial insurance benefits, and to receive a no loss guarantee. this spread depends on yields on corporate and government bonds, policy acquisition costs (that include sales commissions paid and administrative costs), and option costs and is actively managed by the selling companies by changing caps and participation rate limits in the contracts after the first year. in higher yield environments, caps and participation rates will be higher for example. currently, because bond yields are so low, caps and participation rates are lower. for example, most annual cap rates are currently in the 3–4% range. 3. summary literature review reichenstein (2009) contains an in-depth review of most of the relevant and relatively little prior academic research into eias as investment products. most of the previous empirical research on eias generally follows one of two methods typically used to analyze optimal withdrawal rates for retirement planning or to analyze pension plan shortfalls. one method uses monte carlo or similar simulations to create a probability distribution of possible outcomes, where the inputs for the simulations come from the historical return distribution over some time period.7 the second method creates a distribution by examining all the returns that could have been generated from the actual history of returns. in contrast, vanderpal, marrion, and babbel (2011) in their series of articles titled “real world index annuity returns” examine five-year returns actually credited to contract values for a sample of annual point-to-point with cap contracts. as discussed in section 2 above, there is a significant difference between the contract value and realizable cash value, primarily because of surrender charges. consequently, assuming their small sample selectively provided by one insurance company is representative, their returns are “real” only if surrender charges are ignored. kuhlemeyer (2000) uses one, five, and nine-year point to point eias and also ignores the impact of surrender charges. huebscher (2011) provides an interesting critique of vanderpal, marrion, and babbel (2011) and further discusses reichenstein (2009) and previous authors, indicating a real dichotomy in the research conclusions of the two groups of authors. table 2 below contains a summary of much of the prior research, indicating the time periods, methodology, and results/conclusions. table 2 excludes research by edwards and swidler (2005) because, though they have similar features, equity linked certificates of deposit are different. 416 a. terry, e. elder / financial services review 24 (2015) 411–428 unfortunately, a more direct comparison of the results of prior research is made more difficult not only by the differing time periods of study, but also by the fact that each study used eia contracts with different characteristics. in reality, eias have several different contract variables that are not constant through time. the option to change contract values is valuable to the insurance companies and consequently ignoring this option likely biases the results in favor of the eias. table 2 summary of prior research author(s) time period/return assumptions methodology results/conclusion collins, lam, and stampfli (2009) 1996–2008 (actual; 228 months of data) compare rolling seven year cumulative point-to-point and annual reset eia payoffs with actual s&p 500 cumulative return “eia contract holder incurred significant opportunity cost during period . . . ” “results highly sensitive to beginning and ending dates” collins, lam, and stampfli (2009) 1973–2008 (5,000 simulations using distribution statistics based on this time period) simulated payoffs on representative seven-year point-to-point and annual reset eias with various t-bill and s&p500 combination portfolios eia’s guaranteed cumulative return � 50% 50% stocks-bills portfolio 16.1% of the time; maximum return � 50% 50% portfolio 42% of the time kuhlemeyer (2000) 1925–1998 random sample of 100,000 monthly returns with replacement to estimate eia returns for one, five, and nine year point to point contracts finds the returns from eias to be less than exciting when compared with conventional alternatives mccann and luo (2006) simulations assuming mean s&p 500 return of 10% and sd of 20% compare simulated returns on a 10-year point-topoint eia with 60% 10yr treasury and 40% s&p 500 “96.9% of the time the investor is better off with treasuries and stocks than with the eia” mccann (2008) simulations assume 12.5% expected annual s&p 500 return, 2.5% dividend yield compares simulated returns on a 14-year annual point-to-point, 4% monthly cap eia with a 70% 14-yr treasury strip and 30% s&p 500 “99.8% of the time the investor would be better off with the treasury securities and stock than with the eias” reichenstein (2009) 1957–2008 actual annual returns compares capm risk adjusted returns on various annual reset eia contract types using an annual reset eia with 7% annual cap, estimates an � of �1.92% and sharpe ratio of �0.22 vanderpal, marrion, and babbel (2011) 1996–2010 compute rolling five-year geometric average annual rates of return based on rates credited to a sample of annuity owners 5-year total returns based on credited rates outperform s&p 500 67% of time and 50% tbills/50% s&p 500 79% of the time 417a. terry, e. elder / financial services review 24 (2015) 411–428 4. extension of reichenstein reichenstein (2009) regressed hypothetical annual returns from a variety of annual reset eia contracts on the s&p 500 over the 1957–2008 period to estimate capm �s. his � estimates ranged from 0.10 for a 4% cap to 0.17 for a 7% cap.8 he then shows that, on a capm risk-adjusted basis, eias underperform the s&p 500 and treasury bills by 2–3% from 1957 to 2008. the magnitude of this underperformance is completely consistent with the investment spreads discussed in section 2 above. reichenstein’s analysis used annual january through december returns, assuming someone invested in january 1957 and remained invested through december 2008, to estimate capm �s, �s, and sharpe ratios. it thus captures only one of many possible return sequences that would have actually been available to investors. we extend his original analysis in three ways. first, we use monthly data going back to 1928 and re-estimate his regression equations for the 7%, 6%, and 5% eia with annual reset.9 these results are reported in table 3. as indicated in table 3, extending the time period to include the volatile and depressed 1930s increases the standard deviation and sharpe ratio of the s&p 500. in addition, including the prior and most recent years when treasury yields were significantly lower drives down the compound annual return on bills from 5.27% to 3.55%. though the geometric average return over the extended period is slightly lower for the 7% cap eias (4.11% v. 4.29%), it nevertheless exceeds that on bills leading to a positive sharpe ratio. the estimated �s for eias over the extended period were less negative and the estimated �s were essentially the same as in reichenstein (2009). these changes make intuitive sense given the inclusion of several periods with negative equity returns. table 3 replication and extension of reichenstein (2009) asset geometric average annual return ending wealth standard deviation sharpe ratio � � 1957–2008 originally reported by reichenstein (2009) and replicated s&p 500 9.33% $103.43 17.74% 0.31 0.0% 1.00 treasury bills 5.27% $14.45 0 0 5-year treasury notes 7.00% $33.79 6.02% 0.31 annual reset 7% cap 4.29% $8.88 4.41% �0.22 �1.92%* 0.17 annual reset 6% cap 3.72% $6.70 4.05% �0.38 �2.37%* 0.15 annual reset 5% cap 3.16% $5.04 3.73% �0.57 �2.82%* 0.13 1928–2012 (extended) s&p 500 9.48% $2,210.14 20.50% 0.39 0.0% 1.00 treasury bills 3.55% $19.32 0 0 5-year treasury notes 5.37% $85.35 5.13% 0.38 annual reset 7% cap 4.11% $30.78 4.58% 0.13 �0.01%** 0.16 annual reset 6% cap 3.56% $19.61 4.23% 0.00 �1.0%* 0.14 annual reset 5% cap 3.01% $12.44 3.92% �0.14 �1.5%* 0.12 *significant at 1% level. **significant at 10% level. 418 a. terry, e. elder / financial services review 24 (2015) 411–428 second, we create portfolios consisting of the s&p 500 and treasury bills and then calculate every possible 11 through 15-year sequence of returns that could have been generated on a rolling monthly basis during this extended time period. we choose 11–15 year horizons because, as discussed previously in section 2 above, eias have surrender periods lasting 10 years and longer. calculation of shorter horizon returns would have to account for early surrender penalties. we create equity/bill (eb) portfolios having the same systematic risk as the eias. therefore, for example, the 7% cap eia is compared with an eb portfolio consisting of 16% equity and 84% bills and having a � of 0.16 (see bottom of table 3). similarly, the 5% cap eia is compared with an eb portfolio consisting of 12% equity and 88% bills and having a � of 0.12. an annual cost ratio of 0.3% is deducted from the s&p 500 raw annual return to convert it to a realistic mutual fund return, though in reality annual expenses and trading costs are currently closer to 0.05% for vanguard’s s&p 500 etf and 0.17% for vanguard’s s&p 500 index mutual fund.10 we understand that neither eias nor etfs existed before the 1990s, but nevertheless can calculate hypothetical returns for those products using the returns generated by history. third, unlike prior eia research where caps are assumed to remain constant over the time period examined, we allow cap rates to vary with interest rates. specifically, we assume the cap rate is equal to the treasury bond yield plus 2.5%. as explained in section 2 above, insurance companies have the option of changing many of the eia contract features after the first year. in particular they change the participation rates, spreads, and cap rates to maintain their investment spreads. because our benchmark eia assumes 100% participation and no spread, we focus on changes in the cap rate. insurance companies change cap rates in response to changes in interest rates. in particular, when interest rates are higher and their bond portfolios are earning more, they will raise cap rates. similarly, as in recent history, when interest rates are low, cap rates will be lower.11 using actual total monthly returns on the s&p 500 dating back to 1928 and assuming an annual expense ratio of 0.30%, every possible sequence of rolling 11-year, 12-year, and so on through 15-year holding period return, beginning at the end of every single month, was calculated.12 for example, there were 889 sequences of rolling 11-year returns and 841 sequences of rolling 15-year returns. the ending value of a $1 invested in the hypothetical eb portfolio is then compared with the ending value obtained from investing $1 in a hypothetical eia. the benchmark eia assumes an s&p 500 index annual point-to-point calculation with caps of 7%, 5%, and t-bond yield � 2.5%, respectively. the first two eia calculations assume that the same cap remains in effect during the entire holding period. for every possible 11 through 15-year holding period the hypothetical eb portfolio ending investment value is divided by the eia ending value. a ratio of 1 indicates the two values are equal (breakeven), while a value greater than one indicates the hypothetical eb portfolio has a higher value than the eia, and a ratio of less than 1 indicates the eb portfolio has a lower value. the result is a distribution of every possible value comparison (ratio) that could have been created since 1928 using actual historical returns. from this distribution percentiles are calculated and presented. the percentiles represent the percentage of times an observed ratio was less than (greater than) the reported ratio for percentiles below (above) the mean. the breakeven column represents the percentage of the time the eb portfolio fails to outperform the eia.13 419a. terry, e. elder / financial services review 24 (2015) 411–428 as table 4 indicates, the (eb) portfolio had a median value that ranged from 12% higher than a 7% cap eia over an 11-year horizon to over 34% higher than a 5% cap eia over a 15-year horizon. at the shortest horizon (11 year) and highest cap (7%) the eb portfolio was outperformed 27.4% of the time by the eia. at this same time horizon and cap level, the eb portfolio returned at least 27% more than the eia approximately the same percentage of the time. at the longest horizon (15 years) and lowest cap (5%) the eb portfolio was outperformed by the eia only 15.7% of the time. at this same time horizon and cap level the eb portfolio returned at least 102% more than the eia over 15% of the time. for the variable cap eia the mean cap over the time period was 7.2%, with a minimum value of 2.6% and a maximum of 18.9%. comparing the same eb portfolio against the constant 7% cap eia versus the variable cap eia we find that the median wealth ratios are similar, but the wealth ratio using the variable cap eia has a lower breakeven of about 5% for all holding periods, indicating it outperforms 5% fewer times. thus, the constant cap assumption used in previous research biases the results in favor of eias. because one of the selling features of eias is the downside protection, table 4 also indicates the percentage of times the eb holding period return was negative (i.e., ending value was less than $1), the minimum value of a $1 investment, and on the upside, the value of a $1 investment at the 95th percentile. as indicated in table 4, for 14 and 15-year holding periods, eb portfolio returns were never less than 0. for the 15-year horizon the minimum return was 5% over all three portfolios. the far right column illustrates that 5% of the time table 4 ending wealth ratio of s&p index fund equity/bill (eb) portfolio to eia years percentile � 0.16 � portfolio/7% cap eia eb portfolio value 1 5 10 25 50 75 90 95 99 n be � $1 min 95th% 11 0.73 0.79 0.83 0.98 1.12 1.29 1.64 1.76 1.84 889 27.4% 0.6% $0.98 $2.86 12 0.72 0.79 0.84 0.99 1.14 1.34 1.72 1.83 1.92 877 25.8% 1.0% $0.97 $3.15 13 0.73 0.78 0.83 1.01 1.17 1.42 1.76 1.89 1.99 865 24.2% 0.2% $0.99 $3.36 14 0.71 0.78 0.84 1.03 1.20 1.51 1.81 1.95 2.05 853 23.7% 0.0% $1.03 $3.61 15 0.70 0.78 0.86 1.05 1.23 1.60 1.87 2.00 2.10 841 22.2% 0.0% $1.05 $3.96 percentile � 0.12 � portfolio/5% cap eia eb portfolio value 11 0.81 0.86 0.90 1.05 1.19 1.42 1.84 1.93 2.01 889 21.5% 0.2% $1.00 $2.78 12 0.81 0.86 0.90 1.08 1.22 1.50 1.93 2.03 2.11 877 20.6% 0.3% $0.99 $3.03 13 0.81 0.86 0.91 1.11 1.26 1.59 2.01 2.11 2.20 865 18.3% 0.0% $1.01 $3.24 14 0.80 0.85 0.91 1.14 1.30 1.70 2.07 2.18 2.27 853 16.5% 0.0% $1.04 $3.47 15 0.78 0.86 0.93 1.16 1.34 1.83 2.15 2.26 2.37 841 15.7% 0.0% $1.06 $3.76 percentile � 0.16 � portfolio/(t-bond � 2.5%) cap eia eb portfolio value 11 0.82 0.87 0.90 1.01 1.11 1.24 1.36 1.41 1.47 889 22.6% 0.6% $0.98 $2.86 12 0.82 0.86 0.90 1.03 1.14 1.27 1.41 1.46 1.52 877 20.3% 1.0% $0.97 $3.15 13 0.83 0.86 0.90 1.04 1.17 1.31 1.45 1.50 1.63 865 20.3% 0.2% $0.99 $3.36 14 0.83 0.86 0.91 1.05 1.19 1.36 1.49 1.56 1.76 853 19.0% 0.0% $1.03 $3.61 15 0.82 0.87 0.92 1.07 1.21 1.39 1.55 1.62 1.76 841 17.4% 0.0% $1.05 $3.96 be, breakeven percentile (percentage of times wealth ratio �1). 420 a. terry, e. elder / financial services review 24 (2015) 411–428 investors would have earned approximately 3.5 or more times their initial investment over the 14 and 15-year holding periods. fig. 1 plots the data from table 4 above for the 15-year time horizon. for long horizon investors there appears to be a significant opportunity cost at stake for the down-side protection provided by the eia. recall that ratios less than one indicate the portfolio was outperformed by the eia, not that the portfolio lost money. there were no negative 15-year holding period returns for any of the portfolios. the worst horizon-period performance occurred for the 12-year horizon, 0.16 � eb portfolio where 9 of 877 portfolios (1.03%) had negative returns, with the worst return being a loss of 2.7% of the initial investment. an equivalent 1% of the time this portfolio would have more than tripled an investor’s money over a 12-year horizon. while previous studies simply compared eias with an investment in the s&p 500 or mixture of the s&p 500 and treasury securities, reichenstein (2009) was the first to analyze eias on a capm systematic risk-adjusted basis. however, his capm risk-adjusted approach does not explicitly take into account one of the major features of an eia which is the downside protection against negative returns. this is a significant selling point and may be very valuable to a set of extremely risk averse investors. however, as just discussed above, the most severe loss for an investor in the � adjusted eb portfolio was 2.3% over a 12-year horizon, and there were no negative horizon period returns for any of the � adjusted eb portfolios over 14 or 15-year horizons. therefore, while negative returns were still possible in any given year(s), they occurred very rarely over long horizons. fig. 1. ratio of � adjusted equity/bill (eb) portfolio to equity indexed annuities (eia) over 15 year horizon. 421a. terry, e. elder / financial services review 24 (2015) 411–428 5. further extension: option strategies to further extend some of the previous research, we also modeled the optionality aspects of an eia. the downside protection feature of eias motivates comparing the performance of eias not with portfolios having the same systematic risk, but rather with investment alternatives that include both downside protection and upside potential. thus, a simple first strategy consists of buying the s&p 500 and buying an at-the-money put option, that is, portfolio insurance. second, the payoff for an eia can be replicated with call options. for example, an annual reset eia with a 7% cap will have a return in any one year equal to the maximum of the price appreciation on the s&p 500 up to 7%, and 0. thus, the exact return from an eia can be replicated by buying an at-the-money call option and writing a call option that is (1 � cap %) out of the money. this second strategy consists of buying an at-the-money call option on the s&p 500, writing a call option on the s&p 500 that is (1 � cap%) out of the money, and investing the remainder at the 5-year treasury yield. we estimate option values using the black-scholes option pricing model (see bodie et al., 2013). we estimate volatility from the previous 60-months standard deviation of returns, annualized. we use the actual dividend yield and treasury-bill return over the option year (i.e., we assume the expected dividend yield is the actual yield over the period). we assume transaction costs equal 5% of the option values. table 5 contains the ratios of the equity/put (ep) portfolio ending wealth values to the eia ending wealth for 11 through 15-year holding periods. in addition, table 5 indicates the table 5 ending wealth ratio of s&p index fund and equity/put (ep) portfolio to eia years percentile ep portfolio/7% cap eia portfolio value 1 5 10 25 50 75 90 95 99 n be � $1 min 95th% 11 0.55 0.69 0.85 1.12 1.49 1.81 2.06 2.28 2.59 829 15.8% 4.9% $0.60 $3.93 12 0.55 0.72 0.88 1.20 1.55 1.88 2.18 2.38 2.71 817 14.0% 4.2% $0.66 $4.31 13 0.54 0.75 0.92 1.22 1.64 1.97 2.31 2.51 2.83 805 13.8% 3.6% $0.63 $4.86 14 0.56 0.75 0.94 1.27 1.69 2.04 2.43 2.69 2.94 793 12.1% 2.9% $0.72 $5.35 15 0.57 0.75 1.00 1.36 1.72 2.13 2.57 2.77 3.38 781 9.9% 2.0% $0.73 $6.10 percentile ep portfolio/5% cap eia portfolio value 11 0.60 0.77 0.95 1.24 1.68 2.08 2.36 2.60 2.97 829 11.5% 4.9% $0.60 $3.93 12 0.60 0.81 0.99 1.34 1.78 2.17 2.54 2.76 3.16 817 10.3% 4.2% $0.66 $4.31 13 0.60 0.85 1.04 1.39 1.90 2.31 2.75 2.98 3.30 805 9.3% 3.6% $0.63 $4.86 14 0.61 0.84 1.07 1.46 1.97 2.42 2.89 3.21 3.61 793 7.6% 2.9% $0.72 $5.35 15 0.63 0.87 1.14 1.58 2.05 2.55 3.09 3.36 4.24 781 7.3% 2.0% $0.73 $6.10 percentile ep portfolio/(t-bond 2.5%) cap eia portfolio value 11 0.63 0.79 0.91 1.10 1.38 1.71 1.96 2.12 2.26 829 15.7% 4.9% $0.60 $3.93 12 0.64 0.84 0.95 1.14 1.44 1.77 2.05 2.18 2.41 817 14.3% 4.2% $0.66 $4.31 13 0.64 0.86 0.97 1.20 1.49 1.86 2.14 2.26 2.42 805 12.2% 3.6% $0.63 $4.86 14 0.65 0.87 0.99 1.24 1.55 1.91 2.24 2.38 2.70 793 10.8% 2.9% $0.72 $5.35 15 0.67 0.90 1.02 1.30 1.58 2.00 2.32 2.51 2.90 781 9.1% 2.0% $0.73 $6.10 be, breakeven percentile (percent of times wealth ratio �1). 422 a. terry, e. elder / financial services review 24 (2015) 411–428 percentage of times the ep portfolio holding period returns were negative, the worst return, and on the upside the value of $1 invested in the portfolio at the 95th percentile. the ep portfolios had median values that ranged from 49% to 105% higher than the eia. eias outperformed the ep portfolios from only 7.3% of the time for the 5% cap, 15-year horizon to 15.8% of the time for the 7% cap, 11-year horizon holding period. although holding period returns for the ep portfolios were negative less than 5% of the time for all caps and holding periods, their worst downside ranged from losing 40% for the 11-year horizon to losing 27% for the 15-year horizon. on the upside, the 95th percentile returns ranged from 293% at the 11-year horizon to 510% for the 15 year horizon, or, alternatively, at the 95th percentile the ep portfolio resulted in ending wealth ranging from 2.28 to 3.36 times as high as the eia. relative to the � adjusted eb portfolio, the ep portfolio has both more downside and significantly more upside. this is also illustrated in fig. 2 where the ratios of these portfolio values to eias for 15-year holding periods are shown.14 comparing table 5 to table 4, the ep portfolios are outperformed by eias only half as often for the fixed cap eias to two-thirds as often for the variable cap eias as the eb portfolios are outperformed, and in addition the ep portfolios have significantly more upside. however, the potential for loss is significantly higher for the ep portfolios. this result is because of consecutive down market years where investors earn nothing on the market but still have to pay expensive insurance (put) premiums. the final strategy consists of replicating the eia return by purchasing an at the money call and writing a call that is (1 � cap%) out of the money and investing the remainder at the five-year treasury yield. we assume a 5% transaction cost on both the purchased and written call. in a given year the total return from this strategy is the yield on the treasury bond, less fig. 2. ratio of equity/bill (eb) portfolio and equity/put (ep) portfolio to equity indexed annuities (eia) over 15 year horizon. 423a. terry, e. elder / financial services review 24 (2015) 411–428 the net option cost, plus the eia return. table 6 contains the ratios of the call/bond (cb) portfolio ending wealth values to the ending wealth from the eia for 11 through 15-year holding periods. in addition, the table contains the breakeven, or percentage of the time the eia portfolio outperformed the cb portfolio, as well as the mean, minimum, and maximum values of a $1 investment in the cb portfolio. it is important to note that the cb portfolio never lost money, and the minimum value over all caps and horizon periods was $1.07. the median value of the cb portfolio ranged from 17% to 28% higher than the eia portfolio. these higher median returns translate into between 1.4% and 2% per year. the magnitude of these differences correlates well with the underlying economics discussed in section 2 and illustrated in table 1 if one were to invest in corporate as opposed to treasury bonds. eias outperformed the cb portfolios approximately 20% of the time for the 7% constant cap and less than 15% of the time for the variable cap eia. we performed two additional robustness tests. first, option premiums are sensitive to the volatility measures assumed, and it is possible our estimate of volatility based on the prior 60 month volatility is not what is reflected in market option premiums. as a robustness check we also calculated option prices using two alternative measures of volatility and compared to the variable cap eia. the first measure follows a similar approach to edwards and swidler (2005) and estimates an annual standard deviation using the previous 12 months and future 12 months standard deviations. there was no significant impact on the cb portfolio results and the ep portfolio actually performed just slightly better using the weighted historical and future standard deviations. as a second check, we used the implied volatility contained in the table 6 ending wealth ratio of calls and bond (cb) portfolio to eia years percentile cb portfolio/7% cap eia cb portfolio value 1 5 10 25 50 75 90 95 99 n be mean min max 11 0.83 0.84 0.87 1.03 1.17 1.43 1.79 1.84 1.88 829 21.5% $2.01 $1.07 $3.41 12 0.82 0.83 0.87 1.04 1.20 1.49 1.87 1.92 1.95 817 21.4% $2.16 $1.09 $3.80 13 0.81 0.83 0.87 1.04 1.23 1.56 1.94 1.99 2.04 805 20.5% $2.32 $1.10 $4.22 14 0.80 0.82 0.87 1.06 1.25 1.65 2.01 2.06 2.12 793 20.1% $2.50 $1.15 $4.68 15 0.80 0.82 0.87 1.06 1.28 1.74 2.08 2.14 2.20 781 19.2% $2.69 $1.20 $5.17 percentile cb portfolio/5% cap eia cb portfolio value 11 0.83 0.84 0.87 1.03 1.17 1.43 1.79 1.84 1.88 829 17.5% $1.91 $1.07 $3.19 12 0.82 0.83 0.87 1.04 1.20 1.49 1.87 1.92 1.95 817 17.1% $2.04 $1.11 $3.51 13 0.81 0.83 0.87 1.04 1.23 1.56 1.94 1.99 2.04 805 16.4% $2.19 $1.12 $3.86 14 0.80 0.82 0.87 1.06 1.25 1.65 2.01 2.06 2.12 793 15.6% $2.34 $1.16 $4.23 15 0.80 0.82 0.87 1.06 1.28 1.74 2.08 2.14 2.20 781 15.2% $2.50 $1.20 $4.61 percentile cb portfolio/(t-bond � 2.5%) cap eia cb portfolio value 11 0.91 0.92 0.95 1.07 1.16 1.31 1.48 1.51 1.54 829 14.5% $2.05 $1.07 $4.03 12 0.90 0.92 0.95 1.08 1.18 1.35 1.52 1.56 1.58 817 14.4% $2.21 $1.10 $4.58 13 0.90 0.92 0.96 1.09 1.20 1.38 1.57 1.60 1.63 805 14.3% $2.38 $1.12 $5.13 14 0.90 0.92 0.96 1.10 1.23 1.44 1.62 1.64 1.68 793 13.5% $2.57 $1.14 $5.76 15 0.90 0.92 0.97 1.11 1.25 1.49 1.66 1.69 1.72 781 13.6% $2.77 $1.16 $6.35 be, breakeven percentile. 424 a. terry, e. elder / financial services review 24 (2015) 411–428 cboe volatility index, or vix, which is based on real-time prices of options on the s&p 500. the vix data are available beginning in 1990, and so this sample contains only about one-sixth as many rolling investment periods as our original sample. the median wealth ratios to the variable cap eia using the vix volatility measures were approximately 25% and 3% lower than the ratios obtained using our historical volatility estimates for the ep and cb portfolios, respectively. these results make intuitive sense because over this more recent time period the vix volatility measures were approximately 4% higher than our estimates of volatility based on historical standard deviations, and thus the market was pricing options more expensively than our estimated volatility implied. for the ep strategy, portfolio insurance was more expensive and had a larger impact. for the cb strategy, both the purchased and written options were more expensive, and so the net effect was small. as a second robustness test we examined the impact of using the 1957–2012 time period rather than our original 1928–2012 time period. there are a couple of reasons for doing so. even though ibbotson associates (2013) reports equity returns linked back to 1928, as noted in reichenstein (2009) the s&p 500 composite contained only 90 stocks before 1957. in addition, our original extended time period of 84 years includes 12 years, or 14%, from the great depression. if this is considered an anomaly or highly unlikely event, then our sample may be unduly influenced by it. we re-estimated the models for the 1957–2012 period using the variable cap eia. table 7 reports the results for the 13-year holding period, as these results were representative qualitatively of all the holding periods and quantitatively fell between the 11-year and 15-year holding periods. the table reports the percentage difference in estimated wealth ratios between the more recent period, 1957–2012, and our original sample period, 1928–2012. a positive percentage indicates that the estimated 1957–2012 period wealth ratios are higher and thus strengthen our original results. for the call/bond portfolio this is in fact the case; had we used the more recent time periods our results would table 7 comparison of ending wealth ratios: whole period 1928–2012 vs. 1957–2012 equity/bill (eb) portfolio/(t-bond �2.5%) cap eia 13 year horizon 1 5 10 25 50 75 90 95 99 percentile 1957–2012 0.83 0.87 0.91 1.04 1.14 1.28 1.40 1.46 1.55 1928–2012 0.83 0.86 0.90 1.04 1.17 1.31 1.45 1.50 1.63 % change 0.9% 1.2% 1.2% 0.3% �2.0% �2.9% �3.3% �2.4% �5.0% equity/put (ep) portfolio/(t-bond �2.5%) cap eia 13 year horizon 1957–2012 0.86 0.94 1.00 1.21 1.47 1.71 1.96 2.10 2.31 1928–2012 0.64 0.86 0.97 1.20 1.49 1.86 2.14 2.26 2.42 % change 34.5% 10.0% 3.8% 0.8% �1.1% �8.0% �8.6% �6.8% �4.5% call/bond (cb) portfolio/(t-bond �2.5%) cap eia 13 year horizon 1957–2012 1.07 1.10 1.14 1.24 1.35 1.54 1.60 1.61 1.63 1928–2012 0.90 0.92 0.96 1.09 1.20 1.38 1.57 1.60 1.63 % change 18.3% 19.2% 19.1% 14.1% 12.4% 11.2% 1.8% 0.5% 0.2% 425a. terry, e. elder / financial services review 24 (2015) 411–428 be stronger. for the eb and ep portfolios using the more recent period that excludes the depression years strengthens our results for wealth ratios below the median and weakens them for those above. intuitively, this result makes sense because the downside protection of the eia would be more valuable in the severe negative equity returns of the depression era. finally, table 8 summarizes the various alternative portfolio strategies in a slightly different format by showing the mean, minimum, and maximum compound annual return of each strategy for 11, 13, and 15-year holding periods. as indicated in the previous tables, the put strategy has the highest risk, with both the highest possible compound annual return and largest possible loss, and the highest mean return. over a 15-year holding period the put strategy has a mean compound annual return that is nearly 4% higher than the variable cap eia. the call/bond portfolio has higher means, minimums, and maximums at all caps and holding periods and dominates the reichenstein � adjusted portfolio strategy. the call/bond strategy has lower minimum and higher maximum compound annual returns than the eias, but also has mean compound annual returns that are approximately 2.3% to 3.1% higher than the eias, consistent with the economics discussed previously. 6. conclusion this research has extended previous research on equity indexed annuities by first extending the historical time horizon and return time-paths used by reichenstein (2009), to examine portfolios of equity and treasury bills having the same systematic risk as eias. second, and perhaps more importantly, because eias are marketed as a product having downside protection and upside potential, we examine two approaches involving options. the first approach uses a portfolio consisting of equity and put options as portfolio insurance. the second approach creates a bond/call option portfolio where option positions are designed to mimic the payoff to eias. a third contribution of this research is to allow cap rates to vary table 8 compound annual returns by holding period 11 year holding period 13 year holding period 15 year holding period mean min max mean min max mean min max 7% cap eia 4.3% 1.9% 6.6% 4.3% 1.6% 6.1% 4.3% 2.2% 6.2% equity/bills �0.16 � 5.8% �0.1% 10.8% 6.1% �0.1% 10.8% 6.3% 0.3% 10.3% calls/bond 6.6% 0.6% 11.8% 6.7% 0.7% 11.7% 6.8% 1.2% 11.6% 5% cap eia 3.2% 1.4% 4.9% 3.2% 1.2% 4.6% 3.2% 1.7% 4.6% equity/bills �0.12 � 5.5% 0.0% 10.4% 5.7% 0.1% 10.3% 5.9% 0.4% 9.8% calls/bond 6.1% 0.7% 11.1% 6.2% 0.9% 10.9% 6.3% 1.2% 10.7% t-bond yield � 2.5% cap eia 4.6% 1.2% 9.5% 4.7% 1.0% 9.4% 4.8% 1.4% 9.3% equity/bills �0.16 � 5.8% �0.1% 10.8% 6.1% �0.1% 10.8% 6.3% 0.3% 10.3% calls/bond 6.7% 0.6% 13.5% 6.9% 0.9% 13.4% 7.0% 1.0% 13.1% equity/put 8.3% �4.5% 15.4% 8.5% �3.5% 14.6% 8.6% �2.1% 15.5% 426 a. terry, e. elder / financial services review 24 (2015) 411–428 with the level of interest rates, rather than to assume they are constant as in prior research. our primary conclusion is summarized in part of the title to collins et al., (2009), “downside protection, but at what cost?” the opportunity costs of investing in eias over long horizons compared with reasonable and implementable alternative strategies are quite high. there may be risk averse enough investors for whom the possible eia returns make sense. however, at a minimum, these opportunity costs should be disclosed to potential investors at time of purchase. notes 1 the sec treats equity index annuities as a separate class of annuity, along with variable and fixed. furthermore, because performance does not depend on investments in a subaccount or separate account, these annuities are “exempt securities” under section 989j of the dodd frank act if they are issued in states that adopt the naic suitability in annuity transactions model. because of the low interest rate environment some insurance companies are creating registered securities that merge the benefits of fixed and variable annuities. 2 equity index annuity sales now comprise over half of the fixed annuity market, with $34 billion sold in 2012. insured retirement institute 2013 fact book. 3 in addition to the investment aspect, there are tax and insurance aspects associated with eias. the potential tax and insurance benefits of eias are beyond the scope of this article. 4 empirically, mccann (2008) constructed 241 10-year, month-by-month rolling time periods from 1975–2004 and found that the monthly averaging method resulted in a final wealth amount that was 44% lower than the annual point-to-point method. 5 see www.indexannuity.org/rates_by_carrier.htm for current rates and to get a sense for the number of products available. surrender periods are not listed, but most eias have surrender periods between 10 and 15 years. 6 because caps, participation rates, spreads, surrender periods, and other contract features differ from contract to contract, and because some features can be changed by the insurance company after the first year, modeling all the possible iterations of eias is beyond the scope of this article. 7 see also carver (2013) for an application of monte carlo simulation to eias as a teaching assignment. 8 in the first year of reichenstein’s time period equity returns were negative so that the remaining years’ cap is effectively the beginning cap as well. 9 we also re-estimate his original regressions for the 1957–2008 time period, to insure the integrity of our data and obtain the same results he reported (see table 3). it should also be noted that the distribution of returns assumption, or alternatively, required utility function assumptions, necessary to support a capm analysis have been widely criticized in the literature. in a later section we attempt to account for the truncated returns of eias by modeling the option characteristics. 427a. terry, e. elder / financial services review 24 (2015) 411–428 10 see either yahoo finance or morningstar.com for expense ratios. for vanguard s&p 500 etf, see symbol voo. for s&p 500 index mutual fund, see symbol vfinx. 11 the variable cap had a mean of 7.2% over the time period, and consequently the variable cap eia is compared with the same .16 � equity/bill (eb) portfolio that was compared with the 7% cap eia. 12 the last beginning month would have been june 2003 because 10-year and greater holding period returns are not available after then. thus, the history will include someone who bought in march 1999 and sold at the low in march 2009, but would not include someone who bought at the low in march 2009 and sold in march 2019. thus the collapse in 2009 is getting captured in the history, but someone who bought in june 2003 and sold in june 2013 and doubled her money would not. furthermore, note that reichenstein (2009) calculated annual returns assuming an investor bought in january and sold in december of every year, and thus, captures only one of several possible sequences of annual returns represented by history. 13 our holding periods contain overlapping periods and thus are not statistically independent. consequently, our statistics are descriptive of returns that occurred for all possible sequences and are not probabilistic statements about likely future returns. our tables report percentiles and not confidence intervals. 14 only the 7% and variable cap are shown because the 5% cap adds little to the illustration. references bodie, z., kane, a., & marcus, a. (2013). essentials of investments, 9th ed. (pp. 532–537). new york, ny: mcgraw-hill/irwin. carver, a. b. (2013). the equity indexed annuity: a monte carlo forensic investigation into a controversial financial product. decision sciences journal of innovative education, 11, 23–28. collins, p. j., lam, h., & stampfli, j. (2009). equity indexed annuities: downside protection, but at what cost? journal of financial planning, 22, 48–57. edwards, m. & swidler, s. (2005). do equity-linked certificates of deposit have equity-like returns? financial services review, 14, 305–318. mccann, c. (2008). an economic analysis of equity-indexed annuities. working paper submitted to north american securities administrators association, september 10. mccann, c. & luo, d. (2006). an overview of equity-indexed annuities. working paper, securities litigation and consulting group. huebscher, r. (2011). fantasy-world returns for equity indexed annuities. advisor perspectives newsletter, cfa institute, may 31. ibbotson associates. (2013). ibbotson stocks, bonds, bills, and inflation 2013 valuation yearbook. chicago, il: morningstar inc. insured retirement institute. (2014). iri 2013 fact book, 12th ed. (p. 163). washington, dc: insured retirement institute. kuhlemeyer, g. (2000). the equity index annuity: an examination of performance and regulatory concerns. financial services review, 9, 327–342. reichenstein, w. (2009). financial analysis of equity-indexed annuities. financial services review, 18, 291–311. reichenstein, w. (2011). can annuities offer competitive returns? journal of financial planning, 24, 36–39. vanderpal, g., marrion, j., & babbel, d. f. (2011). real world index annuity returns. journal of financial planning, 24, 50–59. 428 a. terry, e. elder / financial services review 24 (2015) 411–428 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 finser_23_1 academy of financial services officers president lance palmer university of georgia president-elect william chittenden texas state university executive vice president-program thomas coe quinnipiac university vice president-communications christine mcclatchey university of northern colorado vice president-finance diane docking northern illinois university thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2014 duncan williams william patterson university vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president frank laatsch univ. of southern mississippi editor, financial services review stuart michelson stetson university directors robert moreschi virginia military institute dale domian york university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake jacob sybrowsky utah valley university past presidents frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms can be downloaded from the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2014 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 23, no. 1, 2014 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate james larsen wright state university karen eilers lahey the university of akron international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2014 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. from the editor this issue contains issue 1 of volume 23 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “does a relationship with a financial service professional overcome a client’s sense of not being in control of achieving their goals?” is by danielle d. winchester at north carolina a&t state university. in this article the author examines the effects of individual control beliefs on financial goal progress under the theory of planned behavior. her findings suggest that low control beliefs are significantly associated with less financial-goal progress; although she shows that the receipt of expert financial advice can reduce this negative effect and results in similar, and in some goal areas, in higher levels of goal progress than that of individuals with high control beliefs. the second article “low-income employees: the relationship between information from formal advisors and financial behaviors ,” is coauthored by crystal r. hudson at clark atlanta university and lance palmer at the university of georgia.in this article, the authors investigate the financial literacy of low-income employees, by examining their financial behaviors. using data from the 2010 survey of consumer finances, the authors find a significant and positive relationship between the use of information from formal advisors and low-income employees’ positive financial behaviors. their research demonstrates that low-income employees who use information from formal advisors exhibit better financial behaviors than those who do not use advisors.. the third article, “downside risk – what the consumer sentiment index reveals ,” is coauthored by mark a. johnson at loyola university maryland and atsuyuki naka at university of new orleans. the authors examine the ability of consumer sentiment using different age groups to forecast short-term and long-term equity returns. using a long-horizon asymmetric response regression format, they show that negative changes in sentiment have a greater influence on stock returns than positive changes in sentiment. they observe that younger individuals appear to be less risk-averse than older individuals. this article provides evidence that risk is an important consideration when investing, and that demographic characteristics are a consideration when determining appropriate investing approaches. the fourth article, “investor preference for skewness and the incubation of mutual funds” is coauthored by philip gibson at the university of the incarnate word and michael finke at texas tech university. the authors investigate the performance of funds created through incubation. incubation creates an incentive for fund families to select highly skewed securities since extreme performance during incubation will increase the likelihood that some funds will outperform financial services review 23 (2014) v–vi 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. before they are sold to the public. although incubation is as an innovative fund promotion technique, it may harm investors by creating the perception that random prior returns are a signal of fund quality. the authors find that net new money flow increases with an incubated fund’s skewness. after incubated funds are sold to the public, skewed funds attract more investor dollars and their average performance declines. their results suggest that the use of skewed securities during incubation is an effective method for increasing demand, but may be a poor quality signal of future performance. the final article, “performance and persistence of performance of healthcare mutual funds,” is coauthored by abhay kaushik and lynn k. saubert both at radford university. in this article the authors analyze 10.55 actively managed domestic healthcare mutual funds over a 12-year period. they show that healthcare mutual funds outperform the passive index by roughly 2.97 percent per year after controlling for the market risk premium, growth and size premiums, and momentum effects. additionally they demonstrate that the abnormal performance over and under does not persist over subsequent periods, indicating that under and over performances are mean reverting. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. i’m pleased to announce the recent partnership between academy of financial services and the financial planning association. the joint april announcement stated “the financial planning association® (fpa®) and the academy of financial services (afs) have entered into a partnership that makes the two organizations co-publishers of financial services review. “as the membership organization for cfp® professionals, fpa is committed to bringing together all key stakeholders that advance the profession of financial planning. this partnership enables fpa to connect academicians, practitioners and students in support of its members and the growth of the profession,” said lauren m. schadle, cae, fpa executive director and ceo. “an electronic version of financial services review will be available to all fpa members, which will complement the award winning journal of financial planning and substantiate fpa’s role in expanding the body of knowledge for financial planning.” “fpa and afs share the desire to encourage basic and applied research in personal financial planning, as well as the need for interaction between financial services professionals and academicians. working together, we can bring these groups into greater alignment for the benefit of the profession and the people it serves,” added lance palmer, ph.d., cfp®, president of afs and associate professor at the university of georgia. afs will be invited to co-sponsor academic research presentations and hold a joint session to discuss the needs and priorities for practitioner-focused academic research at fpa be: seattle 2014, the annual conference of the financial planning community sept. 20-22.” best regards, stuart michelson editor financial services review vi editorial / financial services review 23 (2014) v–vi from the editor this issue contains issue 3 of volume 25 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article strategic complexity in investment management fee disclosures” is coauthored by leslie a. muller at grand valley state university and john turner at pension policy center. this paper develops a measure of complexity in fee disclosures. their data validates the measure of complexity and indicates that people who are more financially literate are better able to understand complex fee disclosures, but that even people with the presumption of a relatively high degree of financial literacy are not all able to decipher complex fee disclosures. the second article “the evidence on target-date mutual funds” is authored by sandeep singh at college at brockport, suny. the author conducts a survey of the theory and recommendations on tdf glide paths which confirms a trend towards focusing on meeting retirement liabilities, rather than optimizing asset only portfolios. after a review of performance evaluation metrics for tdfs, he shows that none of the available indexes possess all seven characteristics of an ideal benchmark. he recommends that plan sponsors can provide better outcomes by offering multiple risk profile tdfs, while researchers can focus on improving glide path and benchmark design. the third article, “household ratio guidelines for the amount of investments” is coauthored by sherman d. hanna at ohio state university and kyoung tae kim at university of alabama. the authors investigate three investment ratios typically utilized in textbooks: investments to net worth, investments to annual income, and investments to total assets. using the 2013 survey of consumer finances they estimate regressions on respondent evaluation of the adequacy of retirement income, among households with a non-retired head in the 2013 survey of consumer finances. they find that the investments to total assets ratio has the strongest relationship to adequacy, controlling for selected household characteristics and the investments to net worth ratio (capital accumulation ratio) is inferior to the other two ratios. the fourth article, the behavior heuristics responsible for formation and liquidation of tax holding accounts” is coauthored by matt hurst and monica mendoza both at stetson financial services review 25 (2016) v–vi 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. university. in this paper, the authors propose the tax liquidation hypothesis, a predictable pattern of behavior regarding individuals’ decisions to create and subsequently to liquidate ‘cash holding’ accounts when facing tax liabilities. previous research on tax related trading has focused on minimizing the individual tax burden by holding winners and selling losers. this behavior, described as “optimal tax trading” suggests that individuals should sell stocks that have lost value in the short-term while holding onto stocks that have gained value until the stocks can be sold at the preferential long-term capital gains rate. the authors propose and test the tax liquidation hypothesis based on investor behavioral biases and the current tax environment. they show that individual investors will hold “cash” accounts that are consistent with their preferences for risk and return. they propose that having cash on hand and liquidating for tax purposes are not mutually exclusive events and that these events can coexist harmoniously. the final article, “cat bonds: risk offsets with diversification and high returns” is authored by richard j. kish at lehigh university. the author investigates catastrophe bonds, a relatively new entry into the bond market, which are a form of reinsurance in which insurance firms are able to offset the financial risks from both natural and man-made catastrophes. the author’s analysis finds that cat bonds have generated high returns, but with the advantage of diversification when compared with similarly rated corporate debt. thus, he finds that cat bonds are a viable investment option within a diversified portfolio. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards stuart michelson editor financial services review vi editorial / financial services review 25 (2016) v–vi spread options and risk management: lognormal versus normal distribution approach robert brooksa, brandon n. clineb,* adepartment of economics, finance and legal studies, the university of alabama, 200 alston hall, box 870224, tuscaloosa, al 35487, usa bdepartment of finance and economics, mississippi state university, 312 mccool hall, mississippi state, ms 39762, usa abstract we provide better tools for managing the downside risk related to the spread between the asset portfolio and corresponding liabilities. these tools are particularly applicable for individual investors. we investigate the spread option valuation model where both underlying instruments follow geometric brownian motion, and one where both underlying instruments are assumed to follow arithmetic brownian motion. we show that the risk parameters are often materially different. these results are important in practical applications of risk management for individual investors as well as financial institutions. for most personal financial planning applications, one can safely use the simpler arithmetic brownian motion model. © 2015 academy of financial services. all rights reserved. jel classification: g13 keywords: spread options; basis options; lognormal; normal 1. introduction wealth management firms claim to provide tailor-made solutions to a client’s unique financial situations. the assertion is that custom-made financial advice is provided within the context of a client’s unique preferences. these assertions and claims, when actual practices are examined, often fall woefully short. as evidence, consider that neither the client, nor the wealth management firm can * corresponding author. tel.: �1-662-325-7477; fax: �1-662-325-1977. e-mail address: brandon.cline@msstate.edu (b.n. cline) financial services review 24 (2015) 15–35 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. articulate the relationship between the behavior of the deployed investment portfolio and the behavior of the client’s liabilities. we use the term liabilities here to include both a family’s contractual liabilities as well as intended uses for funds (e.g., college tuition, charitable giving, and retirement). we advocate wealth management firms take an asset-liability management approach, an approach successfully used by financial institutions and corporations for decades. one motivation of this study is to provide better tools for managing the downside risk related to the spread between the asset portfolio and liability portfolio. we seek tools that serve the client better and afford the wealth management firm confidence in their value-added proposition, regardless of financial market behavior. specifically, we explore managing the spread between assets and liabilities with a particular interest in spread options. spread options are ideally suited for measuring and managing spread risk exposure. the value of spread options and related risk measures provide useful information regarding the current cost of insuring adverse moves in the spread. many other authors have sought to connect modern quantitative finance tools and practices to improving individual investor performance. see, for example, dubil (2004, 2007) and johnston, hatem, and scott (2013). kyrychenko (2008) sought to incorporate nonfinancial assets into the optimal asset allocation process. we follow a similar strategy, but also include liabilities with a focus on downside risk. based on prior academic research, closed form solutions for valuing european-style spread options do not exist when the underlying instruments are lognormally distributed.1 consequently, numerical techniques and approximations must be used for these option pricing models. industry practice is to model spread options assuming the spread is normally distributed even when the underlying distributions are known to be non-normal. this assumption is often made because the spread can be and often is negative. for example, when preferred retirement living standards are considered liabilities, then in the context of financial planning, spreads are often negative. although assuming the normal distribution is pragmatic, this internal inconsistency creates significant integrity concerns for the risk management systems of many financial institutions. these integrity concerns are especially manifest for risk management of large portfolios and may result in risk measurement errors for individual investors. for options on underlying instruments other than spreads, current industry practice is to model the underlying instrument options assuming they are lognormally distributed. this assumption is often justified by the fact that the financial instrument prices are non-negative because of limited liability. continuous time models based on geometric brownian motion (gbm) imply the terminal distribution of the underlying instrument is lognormal. we refer to the spread option model that assumes both underlying instruments follow the lognormal distribution as the base model for comparison purposes. industry practice, however, is to model spread options assuming arithmetic brownian motion (abm) because the spreads are often negative. these continuous time models based on abm imply the terminal distribution of the underlying instrument is normal. we refer to spread option pricing models using the normal distribution as the alternate models (to contrast them from the base model). the alternate model would be easier to deploy for individual investors and their financial planners, but the base model is common with broader industry practices. 16 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 these modeling procedures are internally inconsistent because the difference between two variables that are lognormally distributed does not follow a normal distribution.2 if it can be shown, however, that the abm assumption on the spread (alternate models) provides essentially the same risk-measure results as the gbm (base models), the internal inconsistency can safely be disregarded for risk management purposes. consequently, an alternate model using abm can be deployed that has a much larger degree of tractability; thus, minimizing the integrity concern. in this study, we examine whether the associated risk management costs of this practical departure are significant. specifically, the question is raised: does implementing a more parsimonious model and erroneously assuming the spread is normally distributed significantly bias risk measures commonly used by financial institutions? can individual investors and their financial planners use the simpler alternate model? this article contributes to the spread option literature in three ways: first, we report a new simplified computational method for computing the value of spread options under gbm. our option pricing model involves estimating three single integral expressions of the standard normal cumulative distribution function (cdf). this procedure involves no iterative search routines or approximations outside of estimating a single integral and the standard normal cdf. second, we provide a practical methodology for evaluating whether two option pricing models yield different risk parameters. option pricing models are considered different if the measure of the difference in the risk parameters is larger than the measure of the difference in the risk parameters based on the bid-offer spreads. finally, we illustrate model comparisons by varying moneyness, time to maturity, correlation, and the strike price. we demonstrate that when the only concern is the risk parameter �, the abm assumption does not cause a material error when the time to maturity is short or the correlation is high (spread volatility is low). thus, the normal assumption can be safely made and may even be preferred because it provides a greater degree of tractability. for longer maturities and lower correlations (or higher volatilities), however, the two models are shown to be materially different and poses significant problems with respect to risk management. it is important to note that we do not address the empirical question of which spread option pricing model is most suitable for particular spread options during specific periods of history. instead, we address a risk measurement issue. assuming both underlying instruments are known to be lognormally distributed, by erroneously assuming the spread is normally distributed, are various risk measures significantly biased? in practice, parsimonious models are strongly favored because of the practical difficulties in estimating input parameters. in this light, the alternate model would be much easier to deploy by individual investors and financial advisory firms. the examination of actual market data in an effort to identify the best model is not within the scope of this article for a variety of reasons. first, it is very difficult to acquire clean market data as exchange-traded spread options are thinly traded, and over-the-counter spread option data are not accessible.3 second, our focus here is on risk management and not spread option valuation. we assume the market option price is known and seek to better understand the behavior of various risk measures. third, it is well known that even the best valuation model for a particular product can change over time as market participants’ perspectives change. hence, even if we provided detailed empirical results it would not necessarily be 17r. brooks, b.n. cline / financial services review 24 (2015) 15–35 applicable for other spread options or other periods of time. finally, for individual investors, their liability portfolios will be unique to their family. thus, by definition, the measured spread will be unique, and hence the resultant option values and risk measures will be custom-made. the absence of tradeable options is not a major concern here as the goal is improved risk measures and better understanding of the economic costs of spread risk. the article will proceed as follows: section 2 presents the two models under consideration: the lognormal spread option pricing model (lnsopm) and the normal spread option pricing model (nsopm). section 3 provides the methodology for comparing two option models for risk management purposes (numerical examples used to illustrate the differences in the models from a risk management perspective are also included). section 4 contains the conclusion. 2. models in this section, we provide two spread option valuation models, one assuming both underlying instruments follow gbm, the second assuming both underlying instruments follows abm. the general framework of black and scholes (1973), merton (1973), and black (1976) is followed. before delving into the particular spread option models, a brief review of the literature will be helpful. margrabe (1978) provides a closed-form equation for exchange options that are zero strike spread options. poitras (1998) advances a pricing formula for european-style spread options by extending a special case of the bachelier (1964) option-pricing model. the bachelier model for pricing options on futures spreads provides a methodology for pricing europeanstyle spread options, assuming changes in the underlying futures prices follow unrestricted arithmetic brownian motion (uabm). the assumption of uabm proves very useful in that it allows negative sample paths to exist, resulting in call options that are priced higher than under plain arithmetic brownian motion. poitras points out that because the differences of lognormal variables are not lognormal, a simplification is not possible. furthermore, he states that if prices are lognormally distributed, it is only possible to have a closed form solution in the special case of an exchange option. otherwise, some double integral approximation must be used. while it is true that the normality assumption should be questioned, schaefer (2002) demonstrates that option values from the bachelier model are nearly identical to values found using monte carlo simulation assuming gbm. however, this is only accurate provided the spread volatility and the time to maturity are both low. wilcox (1990) assumes abm to derive a closed form solution for pricing spread options. it is shown by poitras (1998), however, that the formula is not consistent with the noarbitrage argument; thus, it is not a valid option pricing formula. despite the limitations of the wilcox formula, many have used it to develop analytic approximations for spread option pricing, including shimko (1994). shimko (1994) provides an analytic approximation based on the goldman/wilcox model and a model developed by rubinstein (1991). the rubinstein model values options on futures spreads using a double-integral solution where both underlying futures contracts 18 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 follow gbm. like rubinstein, shimko assumes underlying prices follow gbm. by applying the jarrow and rudd (1982) approximations to the wilcox model, shimko approximates a single-integral solution for pricing options on futures spreads. at the same time, he assumes a stochastic convenience yield, thus overcoming the limitations of the wilcox model and the complexity of the rubinstein model. as a result, shimko (1994) provides the formula to approximate the “true” lognormal solution. schaefer (2002), motivated by the fact that there were no closed form solutions for pricing options on futures spreads under the assumption that the assets follow correlated geometric brownian motion, provides an analytic approximation for such options. schaefer compares the bachelier model to monte carlo simulation and the binomial methods that assume the spreads follow correlated geometric brownian motion. for options on futures spreads with two or three underlying assets, his results indicate that both the bachelier model and the analytic approximation provide solutions consistent with monte carlo simulation and binomial methods. he notes that as volatility and time to maturity increase the disparities between the models become much more pronounced, a result consistent with those reported here. other analytic approximations have been published, including alexander and scourse (2004), alexander and venkatramanan (2007, 2012), benth and saltyte-benth (2006), li, deng, and zhou (2008), carmona and durrleman (2003, 2006), and dempster and hong (2002). brooks (1995) provides a quadrinomial lattice approach to valuing spread options. heenk, kemna, and vorst (1990) explore asian options on oil spreads. pearson (1995) presents an efficient approximation to pricing spread options assuming the two underlying instruments are lognormally distributed. pearson shows that the double integral can be reduced to a single integral (not counting the traditional n(d) integral). he proceeds to offer an efficient approximation to the resulting expression. the advantage of the lognormal model presented below is that it is not an approximation. unfortunately, it does require integration of functions of the standard normal distribution. because very accurate approximations exist for the n(d) integral, however, standard single dimensional integration can be applied. borovkova, permana, and weide (2007) offer a spread option model based on a shifted lognormal distribution and then derive approximation formulas based on moments matching. carmona and durrleman (2003) provide a detailed overview of spread option pricing models and their uses as well as an approximation formula. the two option pricing models for spread options are reviewed before we turn to appraising model differences from a risk management perspective. 2.1. lognormal spread option pricing model (lnsopm) we briefly review the assumptions and our notation for spread options. the general payoff at expiration, t, of a spread option can be expressed as: csot � max�0,�1i1,t � �2i2,t � x�, (1) psot � max�0,x � �1i1,t � �2i2,t�, (2) 19r. brooks, b.n. cline / financial services review 24 (2015) 15–35 where csot denotes the call option value at time t, psot denotes put option value at time t, �1 � 0 denotes positive constant (index 1 coefficient), �2 � 0 denotes negative constant (index 2 coefficient), �� � x � � denotes strike price, i1,t denotes the value of index 1 at time t (stochastic), and i2,t denotes the value of index 2 at time t (stochastic). if we assume indexes follow gbm with geometric drift, then dij � ��̂j � �j�ijdt � �̂jijdzj; j � 1,2, (3) where �� � �̂j � � denotes the mean growth rate of index j, �� � �j � denotes the carry costs related to index j, �j � � denotes the sd of index j, and �� � dzj � � denotes the standard wiener process associated with index j. the value of spread option today can be expressed generically as, so0 � pv�e0�sot� , (4) where the expectation is taken under the equivalent martingale measure (standard finance assumptions are made; see appendix a). the value of call and put options, using the lognormal distribution can be expressed as (base models): cso1�i1,0i2,0,x,t,�1,�2, 1,2,r� � exp �rt��� 0 � � 0 � max�0,�1i1,t � �2i2,t � x f1�i1,i2�di2di1� , (5) pso1�i1,0i2,0,x,t,�1,�2, 1,2,r� � exp �rt��� 0 � � 0 � max�0,x � �1i1,t 20 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 � �2i2,t� f1�i1,i2�di2di1� , (6) where l � (subscript) denotes the lnsopm, r � risk-free interest rate; annualized with continuous compounding, and f1(i1,i2) � bivariate lognormal density function. although a closed form solution to the lnsopm does not exist, there are several single integral representations. again, we assume the standard finance assumptions that afford using the risk-free rate as the mean for both indexes (see appendix a). for example, the following single integral version of lnsopm is used with a standard numerical integration methodology.4 cso1,0�i1,i2� � i1e��1t� �� � n�d1,1� z��n� z�dz � �2i2 �1 e��2t� �� � n�d1,2� z��n� z�dz � x �1 e�rt� �� � n�d2� z��n� z�dz, (7) pso1,0�i1,i2� � cso1,0�i1,i2� � �1i1e��1t � �2i2e��2t � xe�rt, (8) where n� z� � e �z2 2 �2� , (9) n�di� z�� � � �� di� z� e �x2 2 �2� dx, (10) d1,1� z� � ln� �1i1e�r��1�t��2 1 2t 2 �� 1�tz x � �2i2e�r��2�t� 2 2t 2 �� 1 2t� 2�tz� � �1 � �2� 1 2t 2 1�t�1 � �2� (11) 21r. brooks, b.n. cline / financial services review 24 (2015) 15–35 d1,2� z� � ln��1i1e�r��1�t��2 �1 2t 2 ���1�2t���1�tz x � �2i2e�r��2�t� �2 2t 2 ��2�tz � � �1 � �2� �1 2t 2 �1�t�1 � �2� , and (12) d2� z� � ln� �1i1e�r��1�t��2 �1 2t 2 ���1�tz x � �2i2e�r��2�t� �2 2t 2 ��2�tz� � �1 � �2� �1 2t 2 �1�t�1 � �2� . (13) it is important to emphasize that this solution is not an approximation like pearson (1995), carmona and durrleman (2003, 2006), li, deng, and zhou (2008) and others, rather it is an exact result. we do not, however, claim it is closed-form in the usual finance sense. it is still technically a double integral (recall the standard n(d) function is an integral). however, practically it is a single integral because of the existence of very accurate numerical approximations available to compute the standard n(d) function. even the standard black, scholes, merton option pricing model requires some sort of numerical approximation to n(d). 2.2. normal spread option pricing model (nsopm) if we assume that the spread follows abm with geometric drift, then ds � �ssdt �sdzs , (14) where �� � �s � � denotes the mean growth rate of the spread, �s � � denotes the sd of the spread (same units of measure as s), and �� � dzs � � denotes the standard wiener process associated with the spread. the value of call and put spread options, based on the normal distribution can be expressed as (alternate model): cson�i1,0,i2,0,x,t,�1,�2,�1,2,r� � exp��rt �� 0 � � 0 � max 0,�1i1,t �2i2,t � x� fn�i1,i2�di2di1� , (15) 22 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 pson�i1,0,i2,0,x,t,�1,�2,�1,2,r� � exp��rt��� 0 � � 0 � max �0,x � �1i1,t � �2i2,t fn�i1,i2�di2di1� , (16) where n denotes the nsopm, and fn(i1,i2) bivariate normal density function. the advantage of the normal distribution is that the difference between normally distributed random variables is also normally distributed. hence, there does exist a closed form solution to nsopm. note that the spread is normally distributed and is denoted as: st � �1i1,t � �2i2,t . (17) so that the expected terminal spread is: e�st � �1i1,0e��̂1� 1�t � �2i2,0e��̂2� 2�t � se�st . (18) the variance of the spread is: v�st � �s 2 e2�st � 1 2�s , (19) and �s is the sd of changes in the spread. we assume the usual finance conditions that afford using the risk-free rate (see appendix a). therefore, we have the following version of nsopm (alternate model): cson,0�i1,i2� � e�rt��e�st � x�n�dn� � v�st 1/ 2n�dn� , (20) pson,0�i1,i2� � cson,0�i1,i2� � �1i1e� 1t � �2i2e� 2t � xe�rt , (21) where n�dn� � e �dn 2 2 �2 , (22) 23r. brooks, b.n. cline / financial services review 24 (2015) 15–35 n�dn� � � �� d e �dn 2 2 �2� dx, and (23) dn � e�st� � x v�st�1/ 2 . (24) 3. analysis of model differences as previously discussed, this article does not seek to address a particular spread option, rather spread options in general. therefore, rather than start with a set of option prices and calibrate both models, we generate option prices with the lnsopm and calibrate the nsopm so it generates the same original option prices. the advantage of this approach is that the nsopm has only one spread option volatility parameter (under the normal distribution, volatility is in units not percentage). when we compare appropriately calibrated option models, we are not assuming one model is “correct.” in practice, option prices are observable in the marketplace, and an option pricing model is “calibrated” to the market option price. for example, one could compute the implied volatility. our objective, however, is not to conduct an empirical test of a particular option contract. our objective is solely to explore the implications for risk management by comparing the lognormal and normal models. therefore, we begin this analysis by calibrating the alternate model option price using the base model price. calibration is a common practice for risk management applications. traders take prices as given and then use models to infer risk parameters. we examine the resulting differences in risk parameters from the base gbm and the alternate abm for spread options. by varying the parameters of the option models, we examine under what circumstances the risk parameters are significantly different. recall the objective here is to examine whether assuming the spread is normally distributed, when the two underlying instruments are lognormally distributed, results in materially different risk parameters. the lnsopm requires three volatility parameters, the percentage sd of each asset and the correlation between assets. the nsopm only requires one volatility parameter, the per unit sd of the spread. hence, we assume the base model option prices (lnsopm) are market prices and then calibrate the alternate model (nsopm). for example, using the parameters specified later in table 1, the spread volatility is $20.8. however, the implied volatility from the nsopm is $19.4. the $19.4 volatility generates the call price of $7.56 as do the lnsopm parameters. if the distribution of the underlying instruments is normal, then the implied volatility of the spread option would match perfectly with the analytic spread volatility. the underlying instruments, however, are assumed to follow a lognormal distribution, and hence the difference of lognormal distributions is not lognormal. thus, the implied volatility would not be expected to equal the analytic spread volatility. the spread option value computed with 24 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 the lnsopm will reflect the non-normal distribution. when the normal distribution is imposed, the nsopm implied volatility will adjust the assumed normal distribution to reflect the lnsopm price. once we have both the base and alternate models generating the same option prices, we estimate the risk parameters. it is important to emphasize that risk management is not directly an exercise in option valuation. rather, option valuation models are used to estimate risk parameters, such as �, �, �, and vega. finally, we assess the magnitude of the risk parameter table 1 varying moneyness panel a: analysis of � moneyness call � put � error factor significance factor error factor significance factor �30 0.2167 0.2340 0.0075 0.0098 �20 0.1572 0.1718 0.0234 0.0292 �10 0.1097 0.1212 0.0483 0.0588 0 0.0728 0.0813 0.0785 0.0942 10 0.0455 0.0513 0.1108 0.1318 20 0.0266 0.0302 0.1429 0.1696 30 0.0144 0.0165 0.1741 0.2064 panel b: analysis of � and � moneyness call and put � call and put � error factor significance factor error factor significance factor �30 0.3574* 0.1722 0.0474 0.1920 �20 0.2204* 0.0984 0.0474 0.1182 �10 0.1002* 0.0999 0.0474 0.0529 0 0.0060 0.0255 0.0474* 0.0056 10 0.1005* 0.0784 0.0474 0.0585 20 0.1850* 0.1266 0.0474 0.1068 30 0.2611* 0.1709 0.0474 0.1511 panel c: analysis of vega and correlation moneyness call and put vega correlation � error factor significance factor error factor significance factor �30 333.2600* 0.1920 0.0785 0.1920 �20 3.1001* 0.1182 0.1020 0.1182 �10 0.7135* 0.0529 0.1136* 0.0529 0 0.0819* 0.0056 0.1169* 0.0056 10 0.4830* 0.0585 0.1142* 0.0585 20 0.7265* 0.1068 0.1070* 0.1068 30 0.8911* 0.1511 0.0963 0.1511 this table reports the error and significance factor for selected risk parameters for the lnsopm and nsopm when the option prices are calibrated together. moneyness of the option is varied from �30 to 30. time to maturity is set equal to one year, volatility is 30% for both indexes, the risk-free rate is 5%, no dividends are assumed, and the correlation coefficient is 0.8. index 2 is set to 100.0 and the strike price is set to zero. index 1 is varied in increments of 10.0, allowing moneyness to vary from �30.0 to 30.0. trading costs are assumed to be one percentage for both assets; hence, buying the spread would result in a one percentage decline in index 1 and a one percentage increase in index 2. *denotes the error factor being greater than the significance factor. 25r. brooks, b.n. cline / financial services review 24 (2015) 15–35 difference. by varying the initial inputs, we demonstrate that the risk parameter differences between the two models are often not materially different. however, the question arises: what is materially different? specifically, from a risk management perspective, what dictates a significant difference in risk parameters of the models being compared? are �s of 0.53 and 0.54 materially different? our measure of materiality is based on the impact of trading costs on the value of risk parameters. 3.1. significance measure the bid-ask spread represents an approximation of the marginal cost to the market maker for rebalancing the portfolio. there are clearly other costs, such as fixed costs, market impact costs, and other variable costs. to apply the methodology below with these other costs, one would have to appropriately adjust the estimated bid-ask spread. our goal here is not to decompose the total spread option price change into the greek components, rather merely to acknowledge that trading costs impact the ability to actually implement a hedging strategy. the higher the bid-ask spread, the less frequently the portfolio would be rebalanced, and therefore, the risk parameters need not be as precise. when applied by individual investors, the goal is merely to estimate the current market price of mitigating the risk of a decline in the spread between assets and liabilities. our significance measure between the risk parameters of the base and alternate models, therefore, is a function of the bid-ask spread. consider the following measure of significance. significance measure: assuming the market maker has the ability to rebalance the portfolio, an insignificant difference in risk parameter (rp) of the base and alternate models being compared exists if the risk parameter error factor between the models is less than the significance factor. the significance factor of the base model is: ŝ � | rpbase, ask�t� � rpbase,bid�t� �rpbase, ask�t� � rpbase,bid�t� 2 � | , (25) and the risk parameter error factor between the base and alternate model is: �rp � | rpbase�t� � rpalt�t� �rpbase�t� � rpalt�t� 2 � | . (26) thus, �rp � ŝ implies rpbase � rpalt, 26 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 where rpbase, ask � risk parameter of the option at the “ask” price, base model, rpbase,bid � risk parameter of the option at the “bid” price, base model, rpbase � risk parameter of base model, rpalt � risk parameter of alternate model, t � assumed transaction cost, and � � materially indistinguishable. the significance factor is a practical measure that seeks to incorporate how sensitive risk parameters are to the value of the underlying assets. if small deviations from the current market price of the underlying asset result in dramatic changes in the risk measures, then it will be difficult to hedge the position when trading costs are high. therefore, the significance factor measures how sensitive the risk parameter is to changes in the underlying asset price because of the bid-offer spread. the error factor measures the deviation of the risk parameter between the two option models. thus, an error factor that is lower than the significance factor would be deemed nonmaterial, whereas an error factor in excess of the significance factor would be deemed material. 3.2. illustrations we illustrate �, �, �, vega, and correlation � by varying four parameters: (1) moneyness, (2) time to maturity, (3) correlation, and (4) strike price. before reviewing these results, a few words regarding these particular derivatives are in order. first, we numerically estimate these derivatives using a highly accurate procedure.5 the �, �, and vegas reported are in terms of the first index value. we also compute the �, �, and vegas of the second term, but they are not reported here. second, correlation � measures the change in the spread option with respect to a change in correlation. third, with the nsopm model, volatility is the sd of the spread. using an iterative search method, we estimate the implied correlation coefficient that yields the implied sd of the nsopm model. therefore, we can estimate vega and correlation � for the nsopm model. 3.2.1. moneyness for the first illustration, we vary the moneyness of the option while holding all the other parameters fixed to examine the difference between the risk parameters of the base and alternate spread option pricing models. moneyness of the option is varied from �30 to 30 units (e.g., dollars), by varying the value of index 1. time to maturity is fixed at one year, percentage volatility is 30% for each index, correlation is 0.8, strike price is 0.0, and the risk-free rate is 5%. a one percentage trading cost is assumed on both indexes, according to the method discussed in the previous section. by adjusting volatility, the option price of the nsopm is calibrated to the price of the lnsopm. the risk parameters are then calculated, and the error and significance factor are reported. table 1 reports the results of five risk parameters: �, �, �, vega, and correlation �. we see that for the parameters selected the �s are not materially different. that is, the error in � 27r. brooks, b.n. cline / financial services review 24 (2015) 15–35 between the two models is not greater than the difference in � of the lognormal model, given a one percentage transaction cost. note that both the error factor and the significance factor are monotonically decreasing in moneyness for calls and monotonically increasing in moneyness for puts. because call (put) �s increase with moneyness, the percentage difference declines (increases). the value of � is maximized around at-the-money spread options. hence, we observe the error factor and the significance factor are both minimized when the options are at-themoney. gamma is materially different, however, when the spread option is either inor out-of-the-money. very small changes in � are significant when the value of � is very small. both in-the-money and out-of-the-money spread options result in small �s and hence, the error factor is larger than the significance factor. the error factor for � is not influenced by moneyness, but the significance factor is influenced. therefore, � is materially different at-the-money because the significance factor is minimized. vega error factors are significant in all cases in table 1. again, the significance factor is minimized at-the-money.6 vega significance is the same for both calls and puts based on put-call parity for spread options. the risk parameter estimates are dramatically different between these two models. time decay behavior is different between the lognormal and normal models, in part, because of its influence on volatility. volatility risk is difficult to manage, particularly for thinly traded options. we see here that the nsopm measure of vega is materially different from the lnsopm. recall that the nsopm has only one input for volatility, the volatility of the spread. the lnsopm, however, has three inputs, the volatilities of each underlying asset and the correlation. therefore, there are several different measures of volatility risk that can be computed for the lnsopm. we assume the volatility change was driven by a change in the volatility of the first asset. the goal here is to identify that volatility risk is an important consideration and can be assessed using the significance measure. correlation � is also significant when moneyness is around zero. 3.2.2. time to maturity next we vary the time to maturity of the option while holding the other model parameters fixed to examine the difference between the risk parameters of the base and alternate spread option pricing models. time to maturity of the spread option is varied from 0.25 to 2 years. the remaining parameters are the same as in table 1. index 1 and index 2 are each assumed to have values of 100.0 and the strike price is 0, resulting in an at-the-money spread option. table 2 reports the results of this case. we observe for the parameters selected the �s are materially different for time horizons greater than one year. that is, the error in � between the two models is not greater than the difference in � of the lognormal model, given a one percentage transaction cost, for maturities less than a year. note that the error factor is monotonically increasing with time to maturity, whereas the significance factor is monotonically decreasing with time to maturity. hence, past some inflection point the models are materially different. for longer maturities, the two model �s deviate further, whereas the impact of trading costs on �s declines. the error factors for � in this case are very small and the significant factors are much larger. because �s are relatively large when the options are at-the-money, and trading costs 28 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 do not considerably influence option value, �s are not materially different. for longer maturities, the �s mildly drift further apart. thetas and vegas are materially different for maturities of 0.5 or greater, and correlation �s are all materially different for all maturities. clearly, the longer the time to maturity for at-the-money options, the greater volatility impacts the difference between these two models. table 2 varying time to maturity panel a: analysis of � time to maturity call � put � error factor significance factor error factor significance factor 0.25 0.0371 0.1757 0.0386 0.1888 0.50 0.0521 0.1195 0.0550 0.1325 0.75 0.0634 0.0954 0.0677 0.1084 1.00 0.0728 0.0813 0.0785 0.0942 1.25 0.0810* 0.0788 0.0882* 0.0846 1.50 0.0884* 0.0647 0.0970* 0.0076 1.75 0.0951* 0.0593 0.1051* 0.0722 2.00 0.1013* 0.0549 0.1127* 0.0678 panel b: analysis of � and � time to maturity call and put � call and put � error factor significance factor error factor significance factor 0.25 0.0015 0.0422 0.0118 0.0223 0.50 0.0030 0.0311 0.0236* 0.0112 0.75 0.0045 0.0274 0.0355* 0.0075 1.00 0.0060 0.0255 0.0474* 0.0056 1.25 0.0075 0.0244 0.0594* 0.0045 1.50 0.0090 0.0237 0.0714* 0.0038 1.75 0.0105 0.0231 0.0835* 0.0032 2.00 0.0120 0.0227 0.0956* 0.0028 panel c: analysis of vega and correlation time to maturity call and put vega correlation � error factor significance factor error factor significance factor 0.25 0.0144 0.0223 0.9990* 0.0223 0.50 0.0374* 0.0112 0.5338* 0.0112 0.75 0.0599* 0.0075 0.1719* 0.0075 1.00 0.0819* 0.0056 0.1169* 0.0056 1.25 0.1035* 0.0045 0.3518* 0.0045 1.50 0.1245* 0.0038 0.5462* 0.0038 1.75 0.1452* 0.0032 0.7091* 0.0032 2.00 0.1654* 0.0028 0.8472* 0.0028 this table reports the error and significance factor for selected risk parameters for the lnsopm and nsopm when the option prices are calibrated together. volatility is 30% for both indexes, the risk-free rate is 5%, no dividends are assumed, and the correlation coefficient is 0.8. index 1 and index 2 are set to 100.0 and the strike price is set to zero. time to maturity is varied in increments 0.25, allowing time to maturity to vary from 0.25 to 2.00. trading costs are assumed to be one percentage for both assets; hence, buying the spread would result in a one percentage decline in index 1 and a one percentage increase in index 2. *denotes the error factor being greater than the significance factor. 29r. brooks, b.n. cline / financial services review 24 (2015) 15–35 3.2.3. correlation we now vary the correlation of the option while holding the other parameters fixed to examine the difference between the risk parameters of the base and alternate spread option pricing models. correlation of the spread option is varied from �0.99 to 0.99. the remaining parameters are again the same as in table 1. we assume the rest of the parameters as previously reported. varying the correlation is similar to varying the volatility because a spread option becomes more volatile as the correlation between the two indexes declines (holding the other parameters constant). table 3 reports the results of this case. we see that for the parameters selected the �s are materially different when correlation declines (or volatility is increased). that is, the error in � between the two models is greater than the difference in � of the lognormal model, given a one percentage transaction cost. note that the error factor is monotonically decreasing with increasing correlation, and the significance factor is monotonically increasing. hence, past some inflection point the models are no longer materially different. therefore, as correlation declines (or volatility increases), the two models are materially different with respect to �. gamma is materially different for correlations of 0.25 or below, whereas � and vega are materially different for correlations of 0.75 or below. finally, correlation � is materially different in all cases. recall that the lower the correlation the greater the spread volatility. 3.2.4. strike price we now vary the strike price of the option while holding the other parameters fixed to examine the difference between the risk parameters of the base and alternate spread option pricing models. again, the market is assumed to provide current market values and these values are used to calibrate the alternate model so that both the base and alternate models yield the same price. varying the strike price is different than varying the index values as provided in table 1 because of the influence of positive and negative strike prices. the strike price is varied from �30 to 30 by units of 10. the remaining parameters are the same as in table 1. table 4 reports the results of this case. we see that for the parameters selected the �s are not materially different, similar to table 1. that is, the error in � between the two models is less than the difference in �s of the lognormal model (significance factor), given a one percentage transaction cost. note that the error factor is monotonically increasing with the strike price for calls and monotonically decreasing for puts. the significance factors for the remaining risk factors in this case are similar to those reported in table 1. for high and low strike prices, the �s are not materially different, whereas �s are materially different. vegas are always materially different, and correlation �s are materially different for strike prices at 20 or below. 4. conclusion individual investors and their financial planners should consider modeling both client assets as well as client liabilities. this study offers better tools for measuring and managing the downside risk related to the spread between the asset portfolio and liability portfolio. we 30 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 provide tools that serve the client better and afford the wealth management firm confidence in their value-added proposition, regardless of financial market behavior. specifically, spread options and related risk measures form a useful framework for articulating client portfolio performance. table 3 varying correlation panel a: analysis of correlation correlation call � put � error factor significance factor error factor significance factor �0.99 0.2105* 0.0210 0.2666* 0.0337 �0.75 0.1990* 0.0228 0.2485* 0.0346 �0.50 0.1859* 0.0251 0.2284* 0.0379 �0.25 0.1714* 0.0281 0.2069* 0.0409 0.00 0.1550* 0.0321 0.1834* 0.0434 0.25 0.1359* 0.0382 0.1572* 0.0510 0.50 0.1125* 0.0483 0.1268* 0.0612 0.75 0.0810* 0.0717 0.0882* 0.0846 0.99 0.0168 0.4365 0.0171 0.4495 panel b: analysis of � and � correlation call and put � call and put � error factor significance factor error factor significance factor �0.99 0.0596* 0.0205 0.0205* 0.0006 �0.75 0.0524* 0.0206 0.0241* 0.0007 �0.50 0.0449* 0.0207 0.0279* 0.0008 �0.25 0.0375* 0.0208 0.0316* 0.0009 0.00 0.0300* 0.0211 0.0354* 0.0012 0.25 0.0225* 0.0214 0.0319* 0.0015 0.50 0.0150 0.0222 0.0429* 0.0023 0.75 0.0075 0.0224 0.0467* 0.0045 0.99 0.0003 0.1309 0.0503 0.1111 panel c: analysis of vega and correlation correlation call and put vega correlation � error factor significance factor error factor significance factor �0.99 0.0402* 0.0006 0.0114* 0.0006 �0.75 0.0463* 0.0007 0.0018* 0.0007 �0.50 0.0524* 0.0008 0.0163* 0.0008 �0.25 0.0584* 0.0009 0.0317* 0.0009 0.00 0.0642* 0.0012 0.0484* 0.0012 0.25 0.0699* 0.0015 0.0669* 0.0015 0.50 0.0754* 0.0023 0.0876* 0.0023 0.75 0.0809* 0.0045 0.1116* 0.0045 0.99 0.0859 0.1111 0.1389* 0.1111 this table reports the error and significance factor for selected risk parameters for the lnsopm and nsopm when the option prices are calibrated together. the initial parameters are the same as the previous tables. correlation is varied in increments of 0.25, allowing correlation to vary from �0.99 to 0.99. *denotes the error factor being greater than the significance factor. 31r. brooks, b.n. cline / financial services review 24 (2015) 15–35 to meet practical demands many financial institutions model spread options assuming the spread is normally distributed even when the underlying distributions are known to be non-normal. we investigate this internally inconsistent practice and test the difference between a spread option pricing model where both underlying instruments are lognormally distributed, and one where both underlying instruments are normally distributed. table 4 varying strike price panel a: analysis of � strike price call � put � error factor significance factor error factor significance factor �30 0.0131 0.0154 0.1601 0.1938 �20 0.0259 0.0290 0.1402 0.1640 �10 0.0462 0.0510 0.1113 0.1301 0 0.0728 0.0813 0.0785 0.0942 10 0.1026 0.1156 0.0497 0.0614 20 0.1339 0.1490 0.0292 0.0364 30 0.1665 0.1787 0.0163 0.0201 panel b: analysis of � and � strike price call and put � call � put � error factor significance factor error factor significance factor error factor significance factor �30 0.2377* 0.1589 0.3120 2.1101 0.0365 0.1463 �20 0.1801* 0.1217 0.0641 0.1742 0.0402 0.1072 �10 0.1000 0.0769 0.0513 0.0700 0.0442 0.0602 0 0.0060 0.0255 0.0474* 0.0056 0.0474* 0.0056 10 0.0866* 0.0268 0.0442 0.0495 0.0513 0.0570 20 0.1711* 0.0738 0.0402 0.0969 0.0641 0.1485 30 0.2500* 0.1133 0.0355 0.1365 0.3026 07110 panel c: analysis of vega and correlation strike price call and put vega correlation � error factor significance factor error factor significance factor �30 2.2728* 0.2451 0.1197 0.1411 �20 1.4233* 0.0781 0.1229* 0.1030 �10 0.6649* 0.0543 0.1197* 0.0574 0 0.0819* 0.0056 0.1169* 0.0056 10 0.2932* 0.0478 0.1197* 0.0463 20 0.5073* 0.0963 0.1229* 0.0924 30 0.6279* 0.1367 0.1200 0.1309 this table reports the error and significance factor for selected risk parameters for the lnsopm and nsopm when the option prices are calibrated together. the strike price is varied from �30 to 30 in increments of 10. volatility is 30% for both indexes, the risk-free rate is 5%, no dividends are assumed, time to maturity is one year, and the correlation coefficient is 0.8. index 1 and index 2 are set to 100.0 and the strike price is set to zero. trading costs are assumed to be one percentage for both assets; hence, buying the spread would result in a one percentage decline in index 1 and a one percentage increase in index 2. *denotes the error factor being greater than the significance factor. 32 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 through four sets of option values we demonstrate that the differences between these two spread option pricing models’ risk parameters are often not significantly different. as a result, the normal assumption on the underlying can be made in those cases, because it provides a greater degree of tractability. they are, however, significantly different in many other cases. for longer maturities and lower correlations (or higher volatilities), the two models are shown to be materially different. the particular model resulting in the most accurate risk parameters is an empirical issue and likely period and product specific. for financial planners, the goal is to structure asset portfolios that are highly correlated with the known liability portfolios, thus, raising the correlation. furthermore, downside risk management tends to be focused on the short term. with these two insights, personal financial planners can safely use the simpler abm model. therefore, the results presented here provide the foundation for improved client portfolio management by financial planners. notes 1 one exception is special cases such as exchange options (see margrabe (1978)). when referring to “closed form,” we use the finance definition of an expression whose numerical complexity is no greater than an “easy to compute” function of cumulative normal distribution functions. 2 poitras (1998) points out that because the difference of lognormal variables is not lognormal, a simplification of the bachelier model is not possible. 3 a review of 16 articles related to spread option valuation revealed only one article used actual spread option data. alexander and venkatramanan (2007) used less than one year of crack spread option data. 4 integral solving routines, such as mathcad, can be used to find reduced form results such as this one. although complex in appearance, n(d) is easily approximated and standard univariate integration routines can be used. because bivariate integration is often unstable and therefore unreliable, this single integral solution is very useful. 5 see eberly (2008). we use a centered difference approach with order of accuracy 4. order of accuracy 1 assumes �h (increment) as well as no increment, order of accuracy 2 also includes �2h, and so forth. hence, order of accuracy 4 includes �4h, �3h, �2h, �h, and no increment. 6 note that the significance factor is the same for �, vega, and correlation � because of the options having a zero strike price. appendix a: standard finance assumptions consider the standard set up for modeling prices (see, e.g., harrison and kreps (1979) and harrison and pliska (1981)): 1. �0,�̂�, for fixed �̂ � t � 0, finite time horizon, a finite horizon economy, 0 � t � �̂. 2. (�, ℑ, p), uncertainty is characterized by a complete probability space, where the state space � is the set of all possible realizations of the stochastic economy between time 33r. brooks, b.n. cline / financial services review 24 (2015) 15–35 0 and time �̂ and has a typical element � representing a sample path, ℑ is the sigma field of distinguishable events at time �̂, and p is a probability measure defined on the elements of ℑ. 3. f�{ℑ�t�:t��0, �̂�} the augmented, right continuous, complete filtration generated by the appropriate stochastic processes in the economy, and assume that ℑ��̂��ℑ. the augmented filtration, ℑ(t), is generated by z. ℑ(0) contains only � and the null sets of p. 4. f is generated by a k-dimensional brownian motion, z�t���z1�t�,⎣ ,zk�t��,t� �0,�̂� is defined on {�,ℑ,p}, where {ℑ�t�},t��0,�̂� is the augmentation of the filtration {ℑz�t�}t��0,�̂� generated by z(t), and satisfies the usual conditions. 5. ep(�) denotes the expectation with respect to the probability measure p. 6. all stated equalities or inequalities involving random variables hold p-almost surely. 7. p is common for all agents implying uniqueness of the nature of the stochastic processes. 8. conventional perfect market conditions are also assumed, such as no transaction costs, no taxes, unrestricted short selling, and no regulatory or institutional constraints. references alexander, c., & scourse, a. (2004). bivariate normal mixture spread option valuation. quantitative finance, 4, 637–648. alexander, c., & venkatramanan, a. (2007). analytic approximations for spread options. icma centre discussion papers in finance dp2007-11, 15th august 2007, 1–23. alexander, c., & venkatramanan, a. (2012). analytic approximations for multi-asset option pricing. mathematical finance, 22, 667–689. bachelier, l. (1964). theory of speculation (j. boness, trans.). in p. cootner (ed.), the random character of stock market prices (pp. 17–78). cambridge, ma: mit press. (original work published 1900). benth, f. e., & saltyte-benth, j. (2006). analytical approximation for the price dynamics of spark spread options. studies in nonlinear dynamics & econometrics, 10, 1334–1355. black, f. (1976). the pricing of commodity contracts. journal of financial economics, 3, 167–179. black, f., & scholes, m. (1973). the pricing of options and corporate liabilities. journal of political economy, 81, 637–659. borovkova, s., permana f. j., & weide, h. v. d. (2007). a closed form approach to the valuation and hedging of basket and spread options. the journal of derivatives, 14, 8–24. brooks, r. (1995). a lattice approach to interest rate spread options. journal of financial engineering, 4, 281–296. carmona, r., & durrleman, v. (2003). pricing and hedging spread options. siam review, 45, 627–685. carmona, r., & durrleman, v. (2006). generalizing the black-scholes formula to multivariate contingent claims. journal of computational finance, 9, 43–67. dempster, m. a. h., & hong, s. s. g. (2002). spread option valuation and the fast fourier transform. in h. geman, d. madan, s. r. pliska, & t. vorst (eds.), mathematical finance–bachelier congress 2000 (pp. 203–220). berlin: springer. dubil, r. (2004). the risk and return of investment averaging: an option-theoretic approach. financial services review, 13, 267–283. dubil, r. (2007). economic derivatives markets–new opportunities for individual investors: a research agenda. financial services review, 16, 89–104. 34 r. brooks, b.n. cline / financial services review 24 (2015) 15–35 eberly, d. (2008, march 2). derivative approximation by finite differences. retrieved from http://www. geometrictools.com/ documentation/finitedifferences.pdf. harrison, j. m., & kreps, d. (1979). martingales and arbitrage in multiperiod securities markets. journal of economic theory, 20, 381–408. harrison, j. m., & pliska, s. r. (1981). martingales and stochastic integrals in the theory of continuous trading. stochastic process and their applications, 11, 215–260. heenk, b. a., kemna, a. g. z., & vorst, a. c. f. (1990). asian options on oil spreads. review of futures markets, 9, 510–528. jarrow, r. a., & rudd, a. (1982). approximate option valuation for arbitrary stochastic processes. journal of financial economics, 10, 347–369. johnston, k., hatem, j., & scott, e. (2013). individual investors: asset allocation vs. portfolio insurance (puts or calls). financial services review, 22, 291–310. kyrychenko, v. (2008). optimal asset allocation in the presence of nonfinancial assets. financial services review, 17, 69–86. li, m., deng, s., & zhou, j. (2008). closed-form approximations for spread option prices and greeks. journal of derivatives, 15, 58–80. margrabe, w. (1978). the value of an option to exchange one asset for another. journal of finance, 33, 177–186. merton, r. c. (1973). theory of rational option pricing. bell journal of economics, the rand corporation, 4, 141–183. pearson, n. d. (1995). an efficient approach for pricing spread options. the journal of derivatives, 3, 76–91. poitras, g. (1998). spread options, exchange options, and arithmetic brownian motion. journal of futures markets, 5, 487–517. rubinstein, m. (1991). somewhere over the rainbow. risk, 4, 63–66. schaefer, m. (2002, april). pricing and hedging european options on futures spreads using the bachelier spread option model. paper presented at ncr-134 conference on applied commodity price analysis, forecasting, and market risk management, st. louis, mo. retrieved from http://purl.umn.edu/19055. shimko, d. (1994). options on futures spreads: hedging, speculation, and valuation. journal of futures markets, 14, 183–213. wilcox, d. (1990). energy futures and options: spread options in energy markets. new york, ny: goldman sachs & co. 35r. brooks, b.n. cline / financial services review 24 (2015) 15–35 financial services review, 31(4) 246 profile to portfolio: where is the missing link? shawn brayman1, nicki potts2, kira brayman3, and yegor komissarov4 abstract this paper focuses on comparing reproducible methodologies to map an investor risk profile into portfolios, products, and solutions in a suitable manner. this study is premised on the assumption that financial advisors have access to valid measures of an individual’s tolerance to take investment risk or aggregate investor risk profile, and measures of the riskiness of products and portfolios of products. we compared three methodologies from the academic literature or regulators against investment alternatives we constructed. the alternatives were a range of 14 efficient portfolios using long-term indices in the united states, canada, the united kingdom, and australia. seven were based on an equal distribution of risk (i.e., the standard deviation increased equally between the seven portfolios), and seven portfolios where the percentage return of each portfolio increased by the same amount between each portfolio. the portfolios distributed by risk were discarded in favour of those distributed by return, and these were then mapped to determine the risk level of the investor they were considered suitable for based on the three methodologies. it was determined that (a) behavioural expectation and exposure to equities is a valid heuristic but insufficient to scale to the wide variety of portfolios and products, use of leverage, and other factors in the marketplace; (b) rolling standard deviation measures can lead to significantly understated assessments of risk in some periods; and (c) the var calculation is recognized in multiple sources as the preferred methodology to align investor concerns of drop in the value of their portfolio to the actual products, but like standard deviation, it is highly impacted by the period utilized. after altering two methodologies (i.e., mifid-ii and riskcat) based on altered duration of data and scaling, respectively, we found that the four methodologies tested agreed with less than one risk band variance and an average correlation of 0.95 to 0.97. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation brayman, s., potts, n., brayman, k., & komissarov, y. (2023). profile to portfolio: where is the missing link? financial services review, 31(4), 246-265. introduction suitability issues are the primary area of complaint by investors to regulators or 1 corresponding author (shawn.brayman@gmail.com). smb research consulting, toronto, canada 2 morningstar, sydney, australia 3 jisc, toronto, canada 4 morningstar, toronto, canada ombudsman services in most developed markets (brayman et al, 2015). this paper was conceptualized to gain a better understanding of why there is so much difficulty in mapping https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ brayman et al. 247 investors to suitable products. there are challenges in “knowing your client” but scientifically validated tools exist (although they are not always used by the advisor marketplace). portfolio managers have a wide variety of analytical tools that they can measure variants of the risk of products and portfolios, even if they are difficult to explain to investors. but by what mechanism do many advisors think an “average investor” should have a 50% fixed income and 50% equity portfolio when it is known that only 61% of american’s even own equities? in fairness, we know that there are multiple constraints and frameworks that advisors must work within and there are systemic and demographic factors at play. income, education, age, marital status, and race all have strong correlations to whether someone owns equities. of households with under $40,000 of income only 29% own equities while 84% of americans with family incomes greater than $100,000 own equities. lack of equities is less about risk suitability than fundability. based on licencing, advisors may be constrained to measure investment product risk, product by product whereas other investors can determine suitability at a portfolio level. in practice this means a balanced mutual fund might contain some amount of large cap, small cap, international and emerging markets and be deemed suitable for a client, but if the advisor tried to create the identical allocation using separate etfs for each asset class, they might be precluded from selling the small cap or emerging markets as being unsuitable/too risky. in some cases, regulators might allow measuring suitability at a portfolio level if the advisor/firm has the technology systems and processes to do so – which is far more complicated than monitoring the risk of individual products. as well clients may not have their entire portfolio with a single advisor – so certificates of deposit at a bank and equity investments with an investment specialist. can firms monitor assets “held away” and balance to the overall portfolio suitable for the client or do they need to ensure the client has sufficient defensive assets held with them? clearly advisors have a formidable task and financial advisors who, by definition, need to look at the holistic position of the client and ensure that solutions are in the investors best interest may need to be very creative to do this and ensure they comply with all the requirements of their compliance department and the regulator. in this paper we are addressing the problem from the financial planning perspective. significant research has explored how to measure an investor’s tolerance for risk or willingness to take on risk. limited research has evaluated how advisors combine various investor behavioural and planning factors to arrive at a "risk profile" for an investor (i.e., how do financial advisors adjust for time horizon, risk capacity, etc.). assuming a financial advisor does so, they are then expected to be able to map risktolerance/profile scores to suitable product solutions or portfolios that they intend to recommend. this “mapping” stage requires that there is an acceptable methodology for measuring investment risk related to products as well as relating this investment risk measure back to the profile of the investor. a search of the literature and common practices in the financial advisory field illustrates that there are many possible approaches but little consensus on best practice. this diversity in methodologies exists in part because of the wide variety of considerations— from investor expectations to the nature of the returns in a specific market to the unique features of individual products (e.g., leverage, downside protection, currency), which can be even more obscured with current forms of engineered products. the purpose of this paper is to review approaches to mapping an investor’s risk profile to an investment solution, with the objective of identifying best practices that can be utilized by financial advisors. literature review while there is no universal consensus in the literature, there appear to be four key components of a risk profile: risk capacity, risk need, time horizon, and behavioural risk tolerance (hubble et al., 2020). because short-term volatility does not necessarily equate with long-term underperformance, investment horizon should be more prioritised in mapping risk tolerance to investment portfolios (hanna & chen, 1998). droms and strauss (2003) advocated for a more financial services review, 31(4) 248 qualitative, personalised, and intuitive approach to portfolio selection, still based on a risk profile and investment horizon, but mainly based on the rough characteristics of asset classes and their ‘appropriateness’ in terms of trade-offs. this is still a commonly practised approach, but relatively ad-hoc in its justifications. investors perceive risk as negative, in terms of the possibility of underperformance, financial loss, and/or inability to meet financial goals. swisher and kasten (2005) asked what the minimum level of return that is ‘acceptable’ to the investor is and use minimum acceptable return (mar) as the boundary at which to measure downside risk, then optimise a portfolio based on similar principles of mvo, but in what they call downside risk optimisation (dro). grable (2008) proposed an alternative based on a multiplicative model of risk profile (from risk profile, risk capacity and time horizon). grable called this riskcat, with the result being a rps (i.e., score). with this model, an investment risk index is generated with a beta index of 1 for a 75% u.s. large cap and 25% u.s. small cap portfolio. a var (value at risk) calculation is then used to map to corresponding rps. grable also noted that it may be possible to map the rps to the efficient frontier. davey (2015) outlined a detailed relationship between investor expectations for the percentage of growth assets in a portfolio and also largest drops in value they expect as they relate to risk tolerance. davey based observations on data from 80,000 respondents to the finametrica risktolerance assessment. davey back tested the model against portfolios and historical data to confirm the alignment with the largest historical market declines. regulators have approached the mapping issue differently. in the european union, regulations went into effect in 2018, in which the european securities and markets authority prescribed a method to calculate the risk on investment products and portfolios, which were then rated at one of seven risk levels. the method at its simplest is a standard deviation based on the fiveyear monthly volatility of a fund annualized, or in the absence of a history, the expectation based on representative asset classes. additionally, financial advisors were provided guidelines on how to consider risk outside normal market behaviour based on guarantees, counter-party risk, embedded leverage, currency, and other factors. some have argued that a risk parity approach should be used to build an optimal portfolio using the risk factors of the investments with no consideration for their associated returns. haesen et al. (2017) attempted to balance the risk parity approach with the mean-variance model using black-litterman in a multi-step process. other ways to map investor risk to portfolio risk include (a) shortfall analysis, (b) expected utility (i.e., how much value the investor expects the investment will provide, which is different from the statistically logical choice), and (c) relative risk aversion (hanna & chen, 1998). several models take a risk number and use this as a parameter in a risk model calculation that produces a given return, then map it on to the efficient frontier, or ad-hoc match it to a set of portfolios. models that use risk in calculations do so in different ways. for example, some researchers have used a function that weights risk according to a subjective aversion to produce a spectral risk measure, such as the risk aversion parameter in the black-litterman model and related models such as in haesen et al. (2017). others have equated risk scores with the beta value in a var calculation for the maximum level of risk. with this approach, investor preferences for return and risk described in a spectral utility function are then mapped mathematically onto the efficient frontier. finally, other approaches to selecting optimal portfolios based on risk may not consider an investor’s risk profile at all. methods the objective of this study was to evaluate multiple replicable methodologies for mapping or linking from an investor’s risk profile to product solutions and discover if there is any commonality in the outcomes that would indicate a consensus or ‘best practice’. in the same way financial advisors would or should question tools for measuring tolerance for risk or a risk profile to determine if they give materially different measures for the same investor, or question brayman et al. 249 statistical measures of risk of a product if one measure says a product is low risk and another say the same product is high risk, there should be some degree of consistent outcome when financial advisors map the risk of a product to the profile of an investor. if there is no consensus, can this be reasonably explained and resolved? in this study, we used four existing mapping methodologies (or refinements thereof) to test if there was any material difference in the results. for each methodology, we mapped a series of efficient portfolios using long term historical data, looking at four markets/countries: australia, canada, the united kingdom, and the united states. we explored two methods for distributing portfolios along the efficient frontier, based on even increments of the standard deviation or even increments of expected returns. in total, we examined 14 portfolios for each of the four countries using four methodologies. after consideration, we discarded seven portfolios based on an equal distribution of risk and utilized the seven using an equal distribution of return. (see figure 1.) figure 1. seven efficient frontiers with even distribution based on expected return the remainder of this paper is focused on answering the following question: do the four methodologies map these seven portfolios into the same or similar risk bands for investors? although finametrica/morningstar have a proprietary psychometric stated-preference risk tolerance test, and the grable-lytton test (grable, lytton, 1998) is a well-documented and cited stated-preference psychometric test mentioned in the riskcat paper, the respective mapping methodologies are independent, simply assuming the use of a valid risk tolerance assessment. in the mifid-ii final report (esma, 2018), although guidance is provided by the regulator on accessing an investor risk profile, there is no prescribed risk-tolerance methodology. we assume valid and reliable tests would generally categorize an investor similarly (i.e., a risk averse investor should be discernible in any valid and reliable methodology and a high-risk taking investor should be equally as discernible), hence the individual risk tolerance assessment is outside financial services review, 31(4) 250 of the scope of this paper. instead, we use the mapping methodologies to categorize the seven portfolios/products to determine if they are considered suitable for the same risk level of investors. mapping methodologies the following mapping methodologies were tested: a mapping system based on davey (2015) using available finametrica data of “behavioural expectation” of how much equity/growth investments a consumer “expects to hold” in their portfolio. a secondary mapping methodology based on davey (2015) data of “behavioural expectation” of the largest drop in value a consumer would expect in their portfolio (i.e., downside risk/var). a mapping approach based on the standard deviations of portfolio/products based on fiveyear historical data, which is more in line with stated requirements by some regulators (mifidii in europe in particular) when measuring product risk. a mapping methodology based on riskcat, outlined by grable (2008) that relates an overall profile score to a product risk index pegged against u.s. large and small cap equities using var. to allow an effective comparison between approaches we defined a common framing or distribution of risk profiles. both finametrica and mifid-ii utilize seven risk bands, so we used this as the basis. when evaluating the results, it is important to note the following scoring methodologies: finametrica uses a 0 to 100 risk-tolerance score, which is mapped into seven risk profile groupings based on the standard deviation of score distribution. mifid-ii uses seven risk bands defined by prescribed thresholds of standard deviations of the products. riskcat uses a 0 to 2.5 scale which we mapped into seven evenly distributed bands for consistency. risk tolerance in the population is accepted as being normally distribution (like i.q.). in most profiling approaches this attribute may then be constrained or reduced by other factors when arriving at a final risk profile. as an example, short time horizons or reduced risk capacity might indicate that a high tolerance investor should still take lower levels of risk when investing, as they do not have the time to recover or other resources to rely on in the event of bad outcomes. for this study, we considered the investor as “unconstrained”. for each methodology we took the seven benchmark portfolios defined for each market and compared which risk band each methodology assigns them to. we acknowledge that in doing this that although each methodology may have seven bands, the breakpoints for each band may vary since the finametrica approach is distributed by population, mifid-ii by a product risk range, and riskcat scores evenly. analysis of portfolios by risk or return we used long-term asset allocation data from four countries (i.e., australia, canada, the united kingdom, and the united states). data varied from 44 to 73 years ending in 2022. we used five asset classes for each country: (a) cash, (b) domestic fixed income, (c) domestic large cap, (d) international equity, and (e) emerging markets. due to short data collection histories, domestic small cap equities were not utilized (see appendix 1 for details) other than in the united states for the calculation of the riskcat index. using these data, we used two approaches to generate a series of seven portfolios along the efficient frontier, based on even risk distribution (figure 2) and even return distributions (figure 1 above). for portfolios with an even risk distribution, we defined the risk range for each country as difference between the standard deviation (sd) of the most volatile asset class (emerging markets) and the sd of the lowest risk asset class (cash). similarly, the return range was difference between the highest return asset class (emerging markets) and lowest return asset class (cash). target risks levels for the seven efficient portfolios distributed evenly by levels of risk brayman et al. 251 were generated by taking standard deviation of cash plus risk range/14 for portfolio 1, then adding risk range/7 for portfolios 2 to 7. as an example, for the u.s. (appendix 1) cash has standard deviation of +/3.82 while emerging markets have a standard deviation of 33.11, so: sd portfolio 1 = 3.82 + (33.11-3.82)/14 = 5.91 sd portfolio 2 = 5.91 + (33.11-3.82)/7 = 10.1, etc. for each of the target risks we solved for the efficient portfolio and resulting asset allocation, expected return, standard deviation, and value at risk. we repeated the process but with evenly distributed returns by calculating the return range as return of the highest performing asset class (emerging markets) and the lowest return asset class (cash). we then evaluated the efficacy of each distribution to ensure reasonableness in application. behavioral expectations of equity exposure and maximum decline in value the finametrica risk tolerance questionnaire has question data that can map risk-tolerance scores to both an expected percentage of growth assets (methodology #1) and a largest potential decline in value of investments before investors become uncomfortable (methodology #2) (davey, 2015). in the case of equity/growth exposure, this study used a question about the expectation of high risk, medium risk, and low risk investments that investors expected in their portfolio. although davey’s estimate of 100%, 50%, and 0% equities/growth for the three risk categories respectively is reasonable, expectation is not advice and a different assumption (e.g., 60/40 is medium risk) would skew the results to higher equity content across expectation categories. davey found little statistical variation by country and therefore generalized across countries. in this study, we compared the percentage of growth assets for each of the seven portfolios (distributed by return) for the four countries by combining the proportions of recommended domestic equity, international equity, and emerging markets and mapping them to the risk band defined by davey (methodology #1). we then compared the var of each of the seven portfolios above for the four countries against the expected downside expectation for the seven risk bands (methodology #2). davey used an aggressive 3.5x standard deviation factor which was replicated in this study. five-year monthly standard deviations and mifid-ii mapping mifid-ii regulation outlined a range of standard deviation intervals and their respective mapping into seven risk bands. these are shown in table 1. as previously outlined, in this study, volatility was calculated as the five-year monthly standard deviation, annualized as follows (methodology #3): using asset allocations of the seven efficient portfolios distributed by even returns for each country, and the representative five-year monthly index data, we calculated the standard deviation as prescribed by mifid-ii regulations in europe. mifid-ii also prescribes a method for mapping products based on the standard deviation into one of seven risk bands (see table 1). using the mifid-ii mapping, we tested to see if our efficient portfolios map into the seven prescribed risk bands or are otherwise distributed. because of the inconsistent results of rolling fiveyear standard deviations, we also used the same long-term standard deviations as were used elsewhere and the prescribed bands outlined by mifid-ii. financial services review, 31(4) 252 table 1. mifid-ii mapping rules based on standard deviation intervals risk class volatility intervals equal or above less than 1 0% 0.5% 2 0.5% 2% 3 2% 5% 4 5% 10% 5 10% 15% 6 15% 25% 7 25% riskcat methodology and var riskcat was designed as a methodology by grable (2008) that was intended to provide a robust approach for calculating a risk profile for an investor/portfolio. the method is based on a multiplicative algorithm that combines risk tolerance, risk capacity, and time horizon to arrive at a profile score. grable proposed a method to map from a riskcat score into a multiple of an index based on 75% u.s. large cap and 25% u.s. small cap equities, which was linked to a riskcat score of 1.0 of a potential 2.5. because the scale assumes a 2.5 return multiple of a 100% equity portfolio, it seems “dated” and linked to beliefs predating our planning understanding of the difficulty for financial advisors to consistently and materially outperform index/etf returns, we also ran riskcat rescaled to a 1.25x market as the upper threshold. results and analysis efficient frontier portfolios using long term history for each of the four primary markets considered, we created the longest series of historical data possible for five asset classes: (a) cash, (b) domestic fixed income, (c) domestic large cap, (d) international equity, and (e) emerging markets. we did not include small cap equities in the portfolio construction as there was insufficient data history in countries outside the united states. data from u.s. small caps is shown in appendix 1 as it remains material to the beta calculation for the riskcat methodology. observations on the efficient frontier when the portfolios were distributed by even levels of risk, we ended up with portfolios that were much more heavily equity biased. as seen in figure 3, the portfolios were allocated 100% in equities by portfolio 4 for canada and the united states (1 being the least risky and 7 the riskiest) and portfolio 5 for australia. the efficient frontier (figure 2) of portfolios distributed evenly by risk is displayed based on the arithmetic mean which was used in the efficient frontier calc. brayman et al. 253 figure 2. seven efficient portfolios evenly distributed by increments of risk figure 3. defensive vs growth distributions based on the seven portfolios distributed by increments of risk financial services review, 31(4) 254 when the portfolios were distributed by equal increments of return (figure 4), we ended up with a more balanced distribution of fixed income and equity positions for the seven portfolios for each country. in australia and the united kingdom there were some fixed income assets up to and including portfolio 6 of the seven portfolios and into portfolio 5 for canada and the united states. figure 4. defensive vs growth distributions based on even return distribution for the balance of the analysis, we compared the seven portfolios constructed by the return distribution to the four methodologies and did not utilize the risk distributed portfolios. the return distributed portfolios more closely reflect actual behaviour in the marketplace and expectations of regulators. comparing expected equity and the efficient portfolios davey (2015) used 80,000 responses to the finametrica risk-tolerance assessment. one of the questions asks individuals how much they expect in high risk, medium risk, or low risk investments. davey proposed that high risk was 100% equities (growth assets), low risk was 100% defensive (i.e., cash & bonds), and medium risk was a 50/50 split between equities and defensive assets. davey then mapped the expected growth assets for each risk score. figure 5 shows the level of expected equity exposure on the y-axis compared to the risk tolerance of investors. brayman et al. 255 figure 5. consumer expected growth asset exposure based on risk tolerance5 for the analysis, we compared the expected equity assets using davey’s (2015) methodology based on the mid-range of the seven risk-profile bands, to the efficient portfolios that were constructed based on distribution by returns (figure 1). note that an updated version of davey’s methodology based on 2022 data is used. table 2 shows the expected maximum equity exposure and the downside drop in the value of investments at which point an investor was deemed to start to become uncomfortable. table 2. ranges of equity or decline in value by seven risk bands table 3 illustrates the seven efficient portfolios constructed for each country by the equity exposure. the distribution of the portfolio mappings is very consistent across all countries, matching the davey (2015) methodology, which was based on expected equity within the portfolio. not surprisingly, portfolio 6 and portfolio 7 both mapped to risk band seven as they were 93% equity or higher. 5 https://riskprofiling.com/downloads/asset_allocation_mappings_guide_v3.pdf expected equity exposure or decline portfolio max equity maximum decline 1 8% 3% 2 20% 8% 3 35% 16% 4 52% 24% 5 69% 34% 6 83% 45% 7 100% 72% https://urldefense.com/v3/__https:/riskprofiling.com/downloads/asset_allocation_mappings_guide_v3.pdf__;!!d8dunmsj4idr!94nip9xi02vceko1mx9kkz6n6dserrmnhoaq-ajvi244wxxcgbogxd11s5ofjdsuna7fcthktaj3jbn25susjapxefw$ financial services review, 31(4) 256 table 3. comparison of expected equity vs. efficient portfolios equity mapped to bands equity of 7 portfolios mapped risk band portfolio australia canada united kingdom u.s. australia canada united kingdom u.s. 1 7% 7% 7% 5% 1 1 1 1 2 24% 24% 21% 21% 3 3 2 2 3 41% 42% 34% 35% 4 4 3 3 4 57% 60% 49% 57% 5 5 4 5 5 75% 81% 71% 82% 6 6 6 6 6 94% 100% 93% 100% 7 7 7 7 7 100% 100% 100% 100% 7 7 7 7 comparing downside risk expectation and efficient portfolios on downside risk davey (2015) outlined the expectation for investors to downside risk or largest falls (methodology 2) and concluded that it is largely consistent with expectation of equity exposure. specifically, davey noted that, “… predicted performance is the mean minus 3.5 standard deviations for estimated returns” (p. 35). the use of 3.5 times standard deviation is atypical and reflects a 99.98% certainty or 1 event in 2,149 years if annualized data are assumed. most var analyses use 90%, 95%, or 99% certainty. keeping in mind that a var is only concerned about the downside tail, we used the following multipliers: 1.28, 1.65, and 2.33 times the standard deviation for 90%, 95%, and 99%, respectively, to calculate the var, depending on the level of certainty desired. when the seven portfolios for the four countries were mapped to the same seven risk bands a relatively clear distribution emerged. the united kingdom seems to require additional risk to achieve incremental increases in the return. the downside expectations shown in table 4 are based on an updated version of davey’s (2015) methodology (i.e., updated with 2022 data). table 4. comparison of downside expectation vs efficient portfolios at 3.5x sd var @ 3.5x mapping to 7 risk bands port largest downside aus can u.k. u.s. aus can u.k. u.s. 1 3.0% 7.8% 6.9% 7.2% 6.5% 2 2 2 2 2 8.0% 11.5% 9.1% 11.0% 10.3% 3 3 3 3 3 16.0% 20.0% 16.5% 19.9% 17.3% 4 4 4 4 4 24.0% 30.3% 24.8% 30.9% 27.5% 5 5 5 5 5 34.0% 41.6% 33.9% 45.9% 40.8% 6 6 7 6 6 45.0% 52.8% 44.7% 64.0% 56.2% 7 6 7 7 7 72.0% 73.4% 68.4% 84.2% 84.3% 7 7 7 7 brayman et al. 257 applying the mifid-ii var calculations the seven model portfolios were analysed based on five-years of monthly data ending in 2022. these data were used to calculate the monthly standard deviation and then annualized. results are shown in table 5. table 5: mapping seven portfolios across four countries based on five-year standard deviation and mifid-ii methodology australia canada united kingdom u.s. portfolio sd risk class sd risk class sd risk class sd risk class 1 1.7% 2 1.2% 2 4.2% 3 1.1% 2 2 3.7% 3 3.8% 3 6.4% 4 3.4% 3 3 5.6% 4 6.6% 4 8.6% 4 5.2% 4 4 7.4% 4 9.0% 4 10.3% 5 7.0% 4 5 9.1% 4 11.7% 5 12.2% 5 9.4% 4 6 10.6% 5 13.6% 5 12.7% 5 11.9% 5 7 10.6% 5 13.0% 5 13.4% 5 12.9% 5 using the mifid-ii mapping ranges and five years of monthly data shows that all the portfolios are more tightly clustered and would be classed in risk bands 2 to 5. it is important to observe that even the 100% cash portfolio has a standard deviation that would be mapped into risk band 3 or 4 in all counties. with recent interest rate hikes even the five-year cash portfolio would display significant volatility. it is difficult to conceive what products would be considered appropriate for mifid-ii’s first two risk bands. similarly, even the most aggressive of all equity portfolios are mapped to risk bands 5. it appears as if the regulator considers any well-constructed, diversified portfolio – even when 100% in equity investments – not the highest risk level for two bands of investors. notice that for canada that portfolio 7’s risk is less than portfolio 6. the portfolios utilized are based on optimized long-term data. in the short five-year period, this can lead to unexpected risk outcomes. we ran the mifid-ii methodology against five-year data from 2013 to 2017 and found “inverted outcomes” compared to the most recent five years. in the period 2013 to 2017, u.s. markets materially outperformed the rest of the world with u.s. portfolios exhibiting the lowest risk. in the last five years a different pattern has emerged. the conclusion is that using rolling five-year periods will not result in consistent outcomes of risk expectation for investor’s portfolios. we completed the same exercise using the historical long-term indices for these portfolios. results are shown in table 6. although there is a slightly broader distribution of the portfolios across the seven mifid-ii risk classes, no products, including cash, fall into risk class 1 or 2 for standard deviation below 2% per year. there is, however, a more realistic mapping of the 100% equity portfolios (at least to risk class 6 from 5). financial services review, 31(4) 258 table 6: mapping seven portfolios times four countries based on long-term indices and mifidii thresholds australia canada united kingdom u.s. portfolio sd risk class sd risk class sd risk class sd risk class 1 4.5% 3 3.6% 3 4.0% 3 3.4% 3 2 5.8% 4 4.6% 3 5.5% 4 4.9% 3 3 8.5% 4 7.0% 4 8.4% 4 7.3% 3 4 11.7% 5 9.7% 4 11.9% 5 10.6% 5 5 15.1% 6 12.6% 5 16.6% 6 14.8% 5 6 18.6% 6 16.0% 6 22.1% 6 19.6% 6 7 24.7% 6 23.1% 6 28.3% 6 28.0% 6 riskcat results the riskcat model introduced by grable (2008) proposed a multiplicative “profiler score” from 0.0 to 2.5 and a mapping into an index of 75% u.s. large cap and 25% u.s. small cap equities. grable classified this index as a beta = 1 and then mapped it to a score of 1.0 on the scale. using this approach, an investor with a profile score of 0.5 would be mapped to an index of 50% equities and 50% cash. an investor with a score of 2.0 would map to 2x the index, which means leverage of 50% (i.e., doubling the level of risk). in the original model, grable referenced an “index” with a standard deviation is 23.38% and a var of 12.43% after a 10.95% return. we recalculated the index var based on this study’s indices (morningstar us market tr usd and the russell 2000 total return index a shorter history than used by grable) using the same 75%/25% split. assuming the portfolio is unconstrained by time horizon and risk capacity then the mapping simplifies to a simple risk scale from 0 to 2.5, with seven risk bands as shown in table 7. table 7. var based on riskcat portfolio risk class var @ 68% var @ 90% var @ 95% var @ 99% 1 0.36 2.1% 4.4% 7.5% 13.2% 2 0.71 7.8% 12.5% 18.7% 30.0% 3 1.07 13.6% 20.6% 29.9% 46.9% 4 1.43 19.3% 28.7% 41.1% 63.8% 5 1.79 25.1% 36.8% 52.2% 80.6% 6 2.14 30.9% 44.9% 63.4% 97.5% 7 2.50 36.6% 53.0% 74.6% 114.4% brayman et al. 259 table 8 illustrates what occurs when the seven portfolios for each of the four countries is mapped to the riskcat using 7 equal bands. table 8. var mapping at 90% certainty based on riskcat portfolio 90% australia canada united kingdom u.s. var var risk class var risk class var risk class var risk class 1 4.4% 2.2% 1 1.1% 1 1.6% 1 1.2% 1 2 12.5% 1.4% 1 1.0% 1 1.1% 1 0.6% 1 3 20.6% 1.2% 1 1.0% 1 1.3% 1 1.1% 1 4 28.7% 4.4% 2 3.3% 1 4.4% 2 3.9% 1 5 36.8% 8.0% 2 5.9% 2 9.1% 2 7.9% 2 6 44.9% 11.6% 2 9.1% 2 14.9% 3 12.7% 3 7 53.0% 18.5% 3 17.1% 3 21.4% 4 22.1% 4 a variation in the certainty level should result in a change in the breakpoints from riskcat and in the calculated var from the seven portfolios. we considered recalibrating the investment index for each country, although this became more problematic as countries outside the united states are unlikely to have as strong a concentration on their own domestic equities, have as developed a small cap market and definitely not be as concentrated in u.s. equities. in reviewing the riskcat results, one can see that “by design” the system was conceived to support up to 2.5 times the risk and return of a 100% equity portfolio, so portfolio 3 aligns with this 100% equity portfolio. it is worth noting that when riskcat was originally published, the model was developed using a group of advisors in a session. at that time, the belief in a “secret sauce” for investing was more dominant than today when most financial advisors are focused on marginal improvements in returns or reductions in risk from relevant indices. as such, we utilized the same methodology and rescaled riskcat using a more contemporary assumption of 1.25x market as the upper bound for an investor that has high tolerance for risk, capacity for loss, and time horizon (called riskcat 2 here). the result of the rescaling is shown in tables 9 and 10. financial services review, 31(4) 260 table 9. thresholds based on riskcat 2 (rescaled beta) portfolio risk class var @ 68% var @ 90% var @ 95% var @ 99% 1 0.18 0.8% 0.3% 1.9% 4.7% 2 0.36 2.1% 4.4% 7.5% 13.2% 3 0.54 4.9% 8.4% 13.1% 21.6% 4 0.71 7.8% 12.5% 18.7% 30.0% 5 0.89 10.7% 16.5% 24.3% 38.5% 6 1.07 13.6% 20.6% 29.9% 46.9% 7 1.25 16.5% 24.6% 35.5% 55.3% table 10. mapping based on riskcat 2 (rescaled beta) portfolio 90% australia canada united kingdom u.s. var var risk class var risk class var risk class var risk class 1 0.3% 2.2% 2 1.1% 2 1.6% 2 1.2% 2 2 4.4% 1.4% 2 1.0% 2 1.1% 2 0.6% 2 3 8.4% 1.2% 2 1.0% 2 1.3% 2 1.1% 2 4 12.5% 4.4% 3 3.3% 2 4.4% 3 3.9% 2 5 16.5% 8.0% 3 5.9% 3 9.1% 4 7.9% 3 6 20.6% 11.6% 4 9.1% 4 14.9% 5 12.7% 5 7 24.6% 18.5% 6 17.1% 6 21.4% 7 22.1% 7 comparing the six methodologies figures 6 illustrates for each of the four markets, how each of the six methodologies mapped each portfolio. using the assumption that the seven portfolios that are evenly distributed by the level of return in the portfolio is an approach for seven bands of risk (not necessarily the case, but a baseline assumption), one can observe: many models struggle with what a financial advisor would consider an “ultra conservative” band. the risk of cash in isolation or the most conservative efficient portfolio (which has a lower sd than cash) often falls in risk band 2 or 3. the davey (2015) equity exposure methodology (#1) has been used in production with many countries and for several years. overall, when the davey approach is compared to a simple linear mapping of the seven portfolios into seven risk bands, it is the closest overall, followed by the riskcat 2 methodology. davey’s downside expectation (#2) (var), even though based on an exceptionally high certainty requirement, places third overall. brayman et al. 261 figure 6. mapping seven portfolios using six methodologies into seven risk-bands the two approaches outlined by davey based on finametrica data were the most evenly distribution with seven portfolio mapping into seven client risk bands, but from a product perspective the methodology using expected equity exposure can only be utilized at a diversified portfolio level, not at a product level (i.e., every common stock on every stock market is considered the same). complex solutions involving guarantees or leverage are also outside of the ability to analyse easily. the mifid-ii methodology (#3) using five-year data was the most challenged models at the tails, mapping all portfolios into the middle three or four risk bands. although applying the long-term indices marginally improved the distribution, all portfolios were still clustered in 4 of the 7 risk bands. this methodology appears to be constructed to bucket the universe of all products at extremes of risk on both ends (lower than cash and higher than a 100% equity aggressive portfolio) where the products in isolation might make sense individually for a client unless part of a broader portfolio or investment strategy. this would appear to indicate a “risk profile test” should be scaled to only map into the middle 4 bands. financial services review, 31(4) 262 figure 7. six methodologies mapping into seven risk-bands the riskcat methodology (#4) was the most removed from the consensus for the simple distribution with 3 to 5 portfolios mapped into risk band 2 for all countries. as expected, this was largely caused by the scaling of 2.5x equity markets as the high point of the range. when rescaled to 1.25x market (i.e., a riskcat score of 2 was set to a beta of 1), it was a closer mapping to the seven evenly distributed bands, but still was less adaptive on the tails (risk bands 1 and 7). the challenge appears to be a simple linear mapping is not reflective of the non-linear shape of the risk/return curve where significant increase in return on the conservative end of the spectrum can be achieved with relatively little increase in risk whereas on the risky investing extreme the same increase in return may require exposure to much riskier asset classes. we also compared the mapping results with a simple mean and standard deviation and a correlation analysis between each of the methodologies (table 11). it was determined that there was an average standard deviation of +/1.1 bands between the methodologies, but it could be as high as 1.5 bands (e.g., the u.s. market). if we remove the mifid-ii 5yr and the originally 2.5x scaled riskcat, the equity exposure, maximum brayman et al. 263 expected decline (var), mifid-ii long-term and riskcat2 average between 0.95 and 0.97 with the other 3 methodologies and agree on the bands +/0.8 risk bands. table 11. correlation between mapping results of methodologies equity exposure var mifid 5yr mifid lt riskcat riskcat2 avg corr 1.00 0.98 0.84 0.96 0.81 0.92 equity exposure 0.92 0.98 1.00 0.88 0.96 0.80 0.94 var 0.93 0.84 0.88 1.00 0.86 0.65 0.78 mifid 5yr 0.83 0.96 0.96 0.86 1.00 0.77 0.94 mifid lt 0.91 0.81 0.80 0.65 0.77 1.00 0.88 riskcat 0.82 0.92 0.94 0.78 0.94 0.88 1.00 riskcat 2 0.91 overall, the results from this study show a strong correlation in the mapping results across all countries. furthermore, all methodologies “scale up” as the portfolios become riskier. conclusions although use of equity asset exposure is an interesting heuristic, it may not be appropriate as the best metric for determining risk in nonefficient portfolios. in most countries, there is between a 50% and 100% difference in the standard deviation of primary equity asset classes. using a metric of total growth assets does not recognize this properly. finametrica historically was very clear to provide illustrative portfolios with comprehensive “risk return guides” for each country. they also stated that their methodology can only be reasonably applied to diversified/well-constructed portfolios. additionally, although alignment with downside risk or var appears to be the best approach, as outlined in this study, basing it on a simple five or 10-year standard deviation can vastly understate the risk for an investor when the volatility is low for a period of time. any mapping methodology that relies on a measure of standard deviation or var needs to be properly calibrated for each market. as illustrated in this study, financial advisors are likely to see significant variation in the overall volatility of markets, the efficient reliance on equities to reduce risk, and differences in the risk/return curve and market efficiency in general. it is also worth noting that using a linear mapping is problematic where the risk/return curve is far from linear and significant returns on the low end might be achieved with little incremental risk compared to the increased risk on the high end for smaller returns. we considered taking the average of all six methodologies as “better consensus mapping” but the traditional riskcat variances skewed the results. we might consider this in the future removing this model from the analysis. many compliance departments and regulators are looking for simple definitions and easy to explain systems to determine suitability – put a risk rating on every product, rate the risk level suitable to an investor and ensure that everything the investor owns is at or below the approved risk level. simplistic solutions, although easier to manage, do not reflect that people and products are not the same and product manufacturers can invent investment products that in isolation could be unsuitable for any client, but in combination with other solutions may form a portfolio that is optimal for a client. simplistic use of a single timeframe, whether short, medium or long term can reflect one aspect of the product or portfolio risk but obscure other aspects. longer time horizons can make downturns like experienced in 2020 with covid “disappear”, whereas short financial services review, 31(4) 264 timeframe can significantly understate risk in quiet periods. implications evolving compliance solutions is challenging and often only occurs when something is considered “broken” by complaints. none-the-less there are some considerations financial planners should consider to ensure best practices for clients: there is merit in having a different scale for product risk than people risk so that solutions that in isolation are outside what would be considered suitable bands are differentiated. implying that every product regardless of risk is suitable for a very high risk-taking investors and only those investors is too simplistic. ensure that the mapping methodology utilized is designed to map into the holistic position of the investor. using the mifid-ii model as an example, it appears clients should be mapped into the middle four bands – not all seven product risk bands. consider multiple timeframes for measuring risk of products, a long-term horizon to capture risk on the same timescale as the financial plan, the short-term risk as usually defined by regulators, and the highest-risk, short-term experience, over the longer history of the product. if the objective is to ensure clients are not taken by surprise, ensure the measure of risk being used is not losing this perspective. references brayman, s., finke, m., bessner, e., grable, j., griffin, p., & clement, r. (2015). current practices for risk profiling in canada and review of global best practices. investor advisory panel of the ontario securities commission. https://www.osc.ca/sites/default/files/20 21-02/iap_20151112_risk-profilingreport.pdf davey, g. (2015, march/april). getting risk right. investment management consultant association inc., 33-39. https://investmentsandwealth.org/getatta chment/010bd2bb-7911-4870-92b12ba48b1a8aa9/iwm15marapr-get tingriskright.pdf droms, w. g., & strauss, s. n. (2003). assessing risk tolerance for asset allocation. journal of financial planning, 16(3), 72-77. esma. (2018, may 28). final report: guidelines on certain aspects of the mifid ii suitability requirements, esma. european securities and markets authority. https://www.esma.europa.eu/sites/defaul t/files/library/esma35-43-869_fr_on_guidelines_on_suitability.pdf grable, j. e. (2008). riskcat: a framework for identifying maximum risk thresholds. in personal portfolios. journal of financial planning, 21(10), 52-62. https://fpperformancelab.org/wpcontent/uploads/riskca t-aframework-for-identifying-maximumrisk-threshold s-in-personalportfolios.pdf grable, j., & lytton, r. (1998). investor risk tolerance: testing the efficacy of demographics as differentiating and classifying factors. journal of financial counseling and planning, 9(1), 61-74. https://www.researchgate.net/publicatio n/228225789_investor_risk_tolerance_ testing_the_efficacy_of_demographi cs_as_differentiating_and_ classifying_factors haesen, d., hallerbach, w. g., markwat, t. d., & molenaar, r. (2014). enhancing risk parity by including views. journal of investing, 26(4), 53-68. hanna, s. d., & chen, p. (1998). subjective and objective risk tolerance: implications for optimal portfolios. journal of financial counseling and planning 8(2), 17-26. hubble, a., grable, j. e., & dannhauser, r. w. (2020). investment risk profiling: a guide for financial advisors, cfa institute. https://www.cfainstitute.org/en/researc h/industry-research/investment-riskprofiling http://www.osc.ca/sites/default/files/2021 http://www.osc.ca/sites/default/files/2021 http://www.esma.europa.eu/sites/default/files/library/e http://www.esma.europa.eu/sites/default/files/library/e http://www.researchgate.net/p http://www.cfainstitute.org/en/research/indu http://www.cfainstitute.org/en/research/indu brayman et al. 265 swisher, p., & kasten, g. w. (2005). postmodern portfolio theory. journal of financial planning, 18(9), 74-82. appendix a table 12. asset class risk/return by country market asset class standard deviation arithmetic average return australia cash 5.07% 7.51% 51 years fixed income 7.47% 8.05% australian equity 23.80% 11.58% global equity 19.66% 12.33% emerging 30.82% 13.51% canada cash 4.00% 5.07% 73 years fixed income 8.02% 6.36% canadian equities 16.53% 11.19% international equities 16.50% 11.14% emerging markets 27.76% 13.11% united kingdom cash 4.52% 6.06% 66 years fixed income 10.25% 8.48% united kingdom equity 27.05% 13.93% global equity 18.75% 12.36% emerging markets 32.97% 15.51% u.s. cash 3.82% 4.87% 44 years fixed income 7.10% 6.65% u.s. equity 17.63% 11.92% u.s. small cap equity 18.97% 12.50% international equity 21.50% 11.25% emerging markets 33.11% 14.48% academy of financial services officers president william chittenden texas state university president-elect thomas coe quinnipiac university executive vice president-program robert moreschi virginia military institute vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2016 swarn chatterjee university of georgia vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president lance palmer university of georgia editor, financial services review stuart michelson stetson university directors sherman hanna ohio state university halil kiymaz rollins college frank laatsch univ. of southern mississippi david nanigian the american college tom potts baylor university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake past presidents lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2015 academy of financial services. all rights of reproduction in any form reserved. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2015 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. gender differences in saving behaviors among lowto moderate-income households patti j. fisher, ph.d.a,*, celia r. hayhoe, ph.d.a, jean m. lown, ph.d.b adepartment of apparel, housing, and resource management, virginia tech, 240 wallace hall, 295 w. campus drive, blacksburg, va 24061, usa bfamily, consumer, and human development department, utah state university, 2905 old main hill, logan, ut 84322, usa abstract in this study we explore gender differences in saving behaviors among lowto moderate-income households using data collected online from a national sample of lowto moderate-income households (nc1172) and data on similar income single households from the 2010 survey of consumer finances (scf). results show that saving behaviors differ by gender. with the nc1172 sample, we find gender differences in the effects of high-risk tolerance and being non-white on the likelihood of being a saver. in the scf, the presence of other household members affects savings differently for women and men. educators and counselors can encourage savings among men and women in lowto moderateincome households as a way to reduce financial risk and ensure financial security. © 2015 academy of financial services. all rights reserved. jel classification: d12; j16 keywords: savings behavior; gender; low to moderate income 1. introduction research shows that the financial behaviors of men and women differ significantly. on average, women earn less and hold lower levels of wealth than men. researchers and financial practitioners report that women are, in general, more risk averse than men and are more conservative in their investment choices (bajtelsmit, bernasek, & jianakoplos, 1999; * corresponding author. tel.: �1-540-231-7218; fax: �1-540-231-1697. e-mail address: pafisher@vt.edu (p.j. fisher) financial services review 24 (2015) 1–13 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. bajtelsmit & vanderhei, 1997; embrey & fox, 1997; faff, mulino, & chai, 2008; grable, 2000; hallahan, faff, & mckenzie, 2004; hinz, mccarthy, & turner, 1997; neelakantan, 2010; yuh & hanna, 1997), participate in retirement plans at lower rates (sung, 1997), and are more likely to live in poverty during retirement (pearce, 1989). however, embrey and fox (1997) conclude that although gender affects some investment decisions of single women and men, gender is not the defining determinant of investment choice. less than two women in 10 feel “very prepared” to make wise financial decisions and only one-third have a detailed financial plan (prudential research, 2010). in addition, more than 92% of women retirees have not adequately prepared for retirement (society of actuaries, 2010). the issue of women’s financial well-being is important for women themselves, as well as financial educators, financial professionals, and policymakers. although much research has been conducted on gender differences in income, risk aversion, investment behaviors, and level of wealth, relatively few studies focus on a gender difference in general saving behaviors (sunden & surrette, 1998). fisher (2010b) investigated gender differences in personal saving behaviors using the 2007 survey of consumer finances (scf), a large national data set that oversamples wealthy households. the study explored gender differences in saving behaviors to better understand whether simply being male or female leads to differences in saving behavior or whether factors related to saving differ for men and women. in the current study, we examine gender differences in saving among lowto moderate-income households by examining results from the 2010 survey of consumer finances and the nc1172 data set. we include both sets of data because the nc1172 data has the advantage of being focused specifically on lowto moderate-income respondents (hayhoe & gutter, 2012), whereas the scf is known as the best source of financial data at the household level, but oversamples wealthy households. in addition, the nc1172 is a newer, less utilized data set and we hope to gain insight on how it compares with a more established data set. according to the annual america saves/american savings education council survey (america saves, 2010), households with incomes above and below $50,000 (approximately the median u.s. income) differ in their financial behaviors and preparedness. the majority of households with incomes above $50,000 have sufficient emergency and retirement savings (85% and 73%, respectively), whereas only 52% and 36% of households with income below $50,000 have sufficient emergency and retirement savings. about 70% of households with income above $50,000 save for retirement at work, as compared with only 26% of households with income below $50,000. similarly, about 68% of households with income above $50,000 have a savings plan with specific goals, as compared with only 38% of households with income below $50,000 (america saves, 2010). lerman and steuerle (2012) argue that personal finance for lowand middle-income families differs significantly from that of upper-income families but that upper-income families are typically the focus in the literature. the assets of lowand middle-income families are more associated with human capital, social insurance programs, and homeownership as compared with upper-income families, and research on the financial behaviors of lowand moderate-income families is needed. the present study focuses on differences in saving behaviors between lowto moderateincome women and men who are not married or living with a partner, eliminating the effect of a spouse or partner. we include households where children or other non-partner household 2 p.j. fisher et al. / financial services review 24 (2015) 1–13 members are present. on average, women have lower incomes and wealth, and are much more likely to be living in poverty during retirement. therefore, it is important to better understand the saving behaviors of single women, as well as how these behaviors may differ from single men, particularly among lowto moderate-income groups. 2. literature review women in the united states have historically been dependent on men for financial security (schmidt & sevak, 2006), and large gender differences exist in financial perceptions, behaviors, and satisfaction (hira & mugenda, 2000). twenty-eight percentage of single female-headed households were living in poverty in 2003, as compared with 13.5% of single male-headed households and 5.8% of married couple households (schmidt & sevak, 2006). about half of women at age 65 are likely to live beyond age 85, yet 89% of female pre-retirees and 92% of female retirees do not plan far enough in the future to cover this 20-year period (society of actuaries, 2010). lupton and smith (2003) report that median assets in white female-headed households are two-thirds of those in similarly situated male-headed families but do not find a significant gender difference among minority households. similarly, divorced women own only 55% of the wealth of divorced men, and never married women have slightly less than half of the wealth of never married men (chang, 2004). in contrast, schmidt and sevak (2006) find no significant difference between the wealth holdings of single maleand female-headed households, despite the fact that households headed by women are more likely to include children. several possible reasons for a gender gap in wealth are suggested by researchers (blau & kahn, 1997; o’neill, 2003). one area in which women and men differ is risk tolerance, with gender differences in investment behaviors such as portfolio diversification (hira & loibl, 2005) and women less willing to invest in risky assets as compared with men (jianakoplos & bernasek, 1998; ricciardi, 2008). gender is receiving considerable attention in regards to risk taking (xiao, collins, ford, keller, kim, & robles, 2010). fisher (2010a) found that differences in saving behaviors were due in part to differences in risk tolerance. for both black and white households, those with low risk tolerance were less likely to save. with women typically scoring lower than men on risk tolerance measures, it is possible that risk tolerance contributes to a gender difference in savings. in addition, women historically completed fewer years of education than men and spend less time in the work force, affecting earnings (ryan & siebens, 2012). attachment to the labor force differs between women and men, which could contribute to gender differences in financial behaviors (sierminska, frick, & grabka, 2010). any difference in wealth may partly result from lower female labor force participation (warren, rowlingson, & whyley, 2001), as women are more likely than men to work part-time, have more diversified work histories because of child bearing and child rearing, and more frequent job changes (berger & denton, 2004). a persistent gender gap in earnings is associated with lower wealth holdings among women, even when holding saving rates constant (blau & kahn, 1997, 2000; o’neill, 2003). gender differences in information processing and information sources 3p.j. fisher et al. / financial services review 24 (2015) 1–13 may also play a role in the financial strategies of men and women (graham, stendardi, myers, & graham, 2002; loibl & hira, 2006). women are also less knowledgeable about finances and are less confident and enthusiastic about managing money than men (chen & volpe, 2002; loibl & hira, 2006). fisher (2010b) provides a comprehensive review of the literature on gender differences in the economic well-being and financial behaviors of women. using the 2007 wave of the survey of consumer finances, fisher (2010b) reports evidence of gender differences in personal saving behaviors. poor health and low risk tolerance negatively affect the likelihood of women saving in the short term and saving regularly; education is not significant in explaining women’s saving behaviors. in contrast, with increased years of education, men are more likely to save in the short term and to save regularly. 3. framework as in fisher (2010b), we use the simple model of wealth accumulation of sierminska et al. (2010), with assets in period t � 1 (at � 1) expressed through the following equation: at � 1 � (1 � r)(at � yt – ct) where (r) is the gross rate of return on investments, (yt) denotes income in period t, and (ct) is consumption in period t. it is possible that the gap in asset value held by men and women results from differences in saving behaviors (yt – ct). in this framework, household saving behaviors are affected by level of income, age, and risk aversion in addition to an individual’s preferences and consumption needs in the presence of liquidity constraints (fisher, 2010b; sierminska et al., 2010). 4. method 4.1. sample two data sets are used to examine gender differences in saving among lowto moderateincome households. the nc1172 data were collected by survey sampling international in december 2010 using a national sample of households with incomes up to $80,000. for details on how the sample was collected, see hayhoe and gutter (2012). the second data set is the 2010 scf, which is sponsored by the federal reserve board. for the present analysis, the scf sample is restricted to households with incomes up to $80,000. to deal with missing responses in the scf, five complete datasets called “implicates” are produced through multiple imputation techniques (board of governors of the federal reserve system, 2012; kennickell, 1997). we use all five implicates. the repeated imputation inference procedure (rii) of montalto and sung (1996) and rubin (1987) is used to estimate the descriptive statistics as well as to conduct the logistic regression analyses 4 p.j. fisher et al. / financial services review 24 (2015) 1–13 because the coefficients and estimates of variance derived by rii techniques provide more valid inference and tests of significance. we weight the scf data because the sample does not follow an equal-probability design (board of governors of the federal reserve system, 2012). following the recommendations of kennickell and mcmanus (1993) and montalto (1998), we weight the descriptive statistics. the logistic regression models estimating the effects of the independent variables on the likelihood of saving do not include weights, and pooled data (that does not account for variability in the data from missing values) are used for the likelihood ratio tests. only households with a non-married/partnered respondent with an income of $80,000 or less are included, for a total sample size from the nc1172 data set of 510 (244 women and 266 men) and 2,492 from the scf data set (1,604 women and 888 men). 4.2. empirical model saving is the act of regularly setting aside funds for a goal, and is based on a decision making process (lewis, webley, & furnham, 1995; wärneryd, 1999). we classify savers as those who: (1) describe their spending over the past year as being less than income, giving them the potential to save; and (2) report that they save regularly. for the first part of this, a variable in each dataset comparing spending to income is used. with the nc1172 data, the saving regularly variable is based on the following question: “if you do not save regularly with a bank, credit union, or other financial institution, why not?” the first possible response for this question was “i do save regularly.” with the scf data, the saving regularly variable is based on whether respondents report saving regularly by setting money aside each month. the independent variables are based on a model of wealth accumulation (sierminska et al., 2010) where saving is affected by income, age, risk tolerance, preferences, and consumption needs. socioeconomic control variables are also included in the model. income is included as a continuous variable. the nc1172 data set includes eight income categories (see table 1), so the midpoint of each category is used as the income variable. in the scf data, income is inflation-adjusted to 2010 dollars (data were collected in 2009) to be consistent with the nc1172 data, which were collected in 2010. age is also included as a continuous variable. the risk tolerance measure in the nc1172 data is based on the 5-item financial risk tolerance scale (grable & joo, 2004). we create three categories of risk tolerance: low, average (reference group), and high. saving horizon is included as a proxy for preferences. respondents indicate which time period is most important to them in their saving and spending decisions (next few months, next year, next 1–4 years, next 5–10 years, and longer than 10 years). if respondents indicate that the period most important to them is the next few months or next year, we code them as having a short saving horizon (reference group), whereas the next 1–4 years and next 5–10 years are coded as medium, with respondents indicating longer than 10 years coded as having a long saving horizon. the scf asks the same question but the response options differ slightly. short saving horizon (reference group) includes those selecting the next few months or next year as their most important period, whereas the medium saving horizon category includes the next few years to 10 years. long 5p.j. fisher et al. / financial services review 24 (2015) 1–13 saving horizon includes those whose most important period for saving and spending is longer than 10 years. for consumption needs, we include dummy variables for income uncertainty, health status, presence of other (non-spouse) household members, and retirement status. for the nc1172 data set, income uncertainty is based on the respondent’s job security, which we code as 1 if job security is insecure or very insecure, and 0 otherwise. for the scf data, income uncertainty is based on whether the respondent has a good idea of income in the next year, coded as 1 if the household does not have a good idea of income in the next year and 0 otherwise. we create dummy variables for health status, with categories for fair to excellent health (reference group) and poor health. the retirement status variable is coded as 1 if the respondent is retired and 0 otherwise. socioeconomic variables include the respondent’s race and education. for race, we include only two categories because of the nc1172 sample size: table 1 characteristics of single person households by gender variables nc1172 2010 scf women (n � 244) men (n � 266) women (n � 1,604) men (n � 888) mean/ frequency mean/ frequency mean/ frequency mean/ frequency saver 25.82 22.18 18.21 24.13 income $31,399 $32,035 $28,537 $31,855 age 46.4 years 46.3 years 53.2 years 48.1 years low risk tolerance 33.20 15.79 65.56 49.39 average to high risk tolerance 59.43 74.06 31.67 45.16 above average to high-risk tolerance 7.38 10.15 2.77 5.45 preferences short saving horizon (next few months to next year) 59.02 66.92 52.43 48.04 medium saving horizon (longer than one year but less than 10 years) 31.97 21.05 41.68 44.98 long saving horizon (10 years or longer) 9.01 12.03 5.89 6.98 consumption needs income uncertainty 34.02 36.09 38.73 38.76 health fair to excellent health 97.95 93.23 91.24 93.21 poor health 2.05 6.77 8.76 6.79 other (non-spouse) household members 62.70 57.89 47.04 28.81 retired 28.28 27.07 25.10 19.10 socioeconomic characteristics race white 75.00 76.69 60.82 67.43 non-white 25.00 23.31 39.18 32.57 education less than high school 14.75 24.44 13.51 12.88 high school graduate/ged 74.18 61.66 56.49 54.75 college degree 8.20 9.02 30.00 32.37 notes. bold coefficients represent a difference between men and women at a 0.05 significance level. �2 test used for categorical variables and t-test used for continuous variables. some variables do not sum to 100% as a result of missing values. 6 p.j. fisher et al. / financial services review 24 (2015) 1–13 white (reference category) and non-white. education is divided into three categories: less than high school (reference category), high school graduate, and college graduate. 5. empirical results 5.1. descriptive statistics table 1 shows the descriptive statistics for all non-married/partnered women and men in the nc1172 and 2010 scf data sets. the following variables are significantly different for men and women based on univariate tests in the nc1172 sample (�2 for categorical variables and t test for continuous variables): low risk tolerance, average risk tolerance, medium saving horizon, poor health, and education (less than high school and high school graduate/ged). for the scf data set, the majority of variables differ significantly for men and women based on univariate tests, with the exception of income uncertainty and fair to excellent health. about 26% of women and 22% of men in the nc1172 sample are classified as savers, as compared with about 18% of women and 24% of men in the scf sample. the mean incomes of men ($32,035) and women ($31,399) in the nc1172 sample are similar. in the scf sample, the mean income of women ($28,537) is significantly lower than that of men ($31,855). the mean age of the nc1172 sample is 46 years for both women and men, with a mean age of 53 years for women and 48 years for men in the scf. in the nc1172 sample, a smaller proportion of women than men report a short saving horizon. a greater proportion of women in the scf report a low risk tolerance as compared with men. a greater proportion table 2 logit parameter estimates for determinants of being a saver variable nc1172 2010 scf women men women men income (in $100,000) 0.026*** 0.025*** 1.934*** 1.516** age 0.007 0.008 �0.012 �0.007 low risk tolerance 0.145 0.229 �0.228 �0.373 high risk tolerance �0.329 1.216* �0.234 �0.343 preferences medium saving horizon 0.922** 0.130 0.510** 0.337 long saving horizon �0.207 0.677 0.981*** 0.824* consumption needs income uncertainty 0.331 �0.864* �0.532** �0.719*** poor health �12.855 0.032 �0.092 �0.511 other (non-spouse) household members 0.116 0.257 �0.671*** 0.075 retired �0.076 0.416 �0.000 �0.500 socioeconomic characteristics non-white �0.106 �1.516** 0.191 0.305 education high school graduate/ged 0.014 �0.221 �0.006 0.653 college graduate 0.137 �1.140 0.446 0.594 notes. significant individual coefficients indicated by * p � 0.05. ** p � 0.01. *** p � 0.001. bold coefficients represent a difference between men and women at a 0.10 significance level. 7p.j. fisher et al. / financial services review 24 (2015) 1–13 of women in the nc1172 sample report a long saving horizon whereas a greater proportion of men in the scf sample report a long saving horizon, although the gender difference in long saving horizon is only significant in the scf sample. in the nc1172 sample, a significantly greater proportion of men report poor health, whereas in the scf sample, for which the mean age of women is highest, the reverse is true. the racial distribution for men and women in the nc1172 sample is similar. in the scf sample, a significantly higher proportion of men are white. in the nc1172 sample, a similar proportion of men and women report other (non-spouse/partner) household members, whereas in the scf sample women have significantly higher proportions of additional household members (47% of women vs. 29% of men). the proportion of men and women who are retired is similar in the nc1172 sample, but a significantly higher proportion of women than men are retired in the scf sample, which is consistent with the higher age of women in the scf. a significantly greater proportion of men than women have less than a high school education in the nc1172 sample. in the scf sample, a significantly greater proportion of men than women completed college. 5.2. logistic regression results we first estimate the models separately for men and women for each sample. to investigate gender differences in being a saver, we estimate interaction models using the total sample of men and women (separately for the nc1172 and scf samples) with each independent variable in the model interacting with gender along with each of the noninteracting independent variables and the gender dummy variable. by examining the interacted variables in the interaction model that are significant, we can find significant differences between men and women in individual parameters. in addition, a likelihood ratio test is used to investigate whether including the gender dummy and interaction terms improves the model fit as compared with the non-interacted model. for more information on these procedures, see maddala (1992). for the nc1172 sample, the results of the interaction model indicate a significant difference in saving behaviors between men and women (p � 0.01 in the likelihood ratio test). two variables differ significantly for men and women in the model: high-risk tolerance and being non-white. for men, high risk tolerance is associated with a significantly higher likelihood of saving, whereas no significant effect exists for women. being non-white is associated with a significantly lower likelihood of being a saver for men, but is not significant for women. though not statistically different for men and women, medium saving horizon has a positive effect on the likelihood of saving for women but not men, whereas income uncertainty is associated with a lower likelihood of being a saver for men but not women. we also find evidence of a significant gender difference in being a saver in the scf sample (p � 0.001 in the likelihood ratio test). the effect of having other household members present differs significantly for men and women. having other household members present is associated with a significantly lower likelihood of saving for women, but the variable is not significant for men. income and a long saving horizon are significantly and positively related with the likelihood of being a saver for both women and men. medium saving horizon is significantly and positively related with being a saver for women but not 8 p.j. fisher et al. / financial services review 24 (2015) 1–13 men. income uncertainty is associated with a significantly lower likelihood of being a saver for both women and men. 6. discussion and implications we investigated the relationship between gender and saving behaviors among lowto moderate-income households utilizing two data sets. the scf oversamples wealthy households, whereas the nc1172 sample is limited to lowto moderate-income households. thus, we limited the scf sample to incomes of $80,000 or less to be consistent with the nc1172 income limits and our focus on lowto moderate-income singles. note, however, that households can have high levels of net worth but low income and vice versa. we find significant gender differences in being a saver among lowto moderate-income households with both the nc1172 and scf samples. income is significantly and positively related to the likelihood of being a saver for men and women in both data sets, demonstrating the importance of this variable in the saving behaviors of lowto moderate-income households who do not have the disposable income of those with higher incomes. nc1172 men with high risk tolerance were significantly more likely to be savers. income uncertainty is associated with a significantly lower likelihood of being a saver for men in both samples, with a significantly negative effect on being a saver for women in the scf sample but not the nc1172 sample. this finding may be the result of the small sample size in the nc1172 data set, and warrants further investigation. persons with uncertain or irregular income have a greater need to save than workers with a predictable income and, thus, should be targeted by educators. these results suggest that to encourage saving, educators and counselors should target lowto moderate-income men working in jobs that are less stable, such as manufacturing or labor positions. in addition, programs and policies that encourage and facilitate saving among the lowest income groups are needed. although not a direct outcome from the present study with a focus on low to moderate incomes, the question arises whether educators need to analyze their messages and examples to determine if they may be too middle-income oriented and easily dismissed by learners who feel that they do not have the capacity to save. perhaps too many examples assume a regular, predictable income that is beyond the reach of a growing number of american workers. with mean incomes of about $30,000, these singles are having a tough time making ends meet. educators and counselors need to tailor their examples to the audience; this is not the steak and champagne crowd. in the scf sample, the presence of other household members differs significantly for men and women with respect to being a saver, with the presence of other household members leading to a significantly negative effect on being a saver for women, but no significant effect for men. policymakers and financial service providers should continue to seek ways to help female household heads who have more difficulties saving and becoming financially stable as compared with male household heads. for example, the doorways to dreams fund (d2d) has supported a creative endeavor to encourage saving called prize-linked savings (pls), and michigan credit unions used “save to win” to generate 25,000 new accounts and $40 million in savings by encouraging new accounts and deposits by entering savers into a lottery. “prize-linked savings (pls) engages consumers to save by changing the savings 9p.j. fisher et al. / financial services review 24 (2015) 1–13 experience. savers experience immediate rewards of prizes and incentives. by doing so, pls reframes the act of saving as fun—a game with real rewards, rules, suspense, and possibility” (abbi, hahnel, maynard, & smith-ramani, 2012). because only about one-fourth (18–26%) of these lowto moderate-income singles save, this should be a major target group for educators. the results of the present study also provide support for financial education programs that target men and women separately. although a variety of financial education programs target women, such as financial planning for women, wi$eup: financial planning for generation x & y women (wiseupwomen. tamu.edu),women’sfinancialliteracyproject(http://www.financiallit.org/programs/partners/ wflp.aspx, http://www.financialwoman.com/), and women’s institute for a secure retirement (http://www.wiserwomen.org/), we have found no comparable programs targeting men. as with all research, there are limitations to our study. the survey questions for some of the variables differ in the two data sets. the questions used to classify respondents as savers differ somewhat, as the question regarding saving over the previous year specifically excludes investments in the scf whereas the nc1172 question does not exclude investments. this difference may account for the higher propensity to save for the nc1172 men who reported a high-risk tolerance; they may have been investing as well as saving. although these responses seem contradictory it could be that, because of the recession and perhaps a job loss or cut in income, some respondents spent more than they earned in the previous year but still consider themselves to be a saver because in more typical times they do save regularly. it is also possible that some consumers save through an automatic deposit from each paycheck but then periodically spend the money in the account. should they be considered savers if their account is essentially a short-term holding pen for cash? they may save regularly throughout the year to pay for expenses that occur once or twice a year, such as insurance premiums, property taxes, christmas gifts, or vacations. further research on gender differences in the saving and investing behaviors of lowto moderate-income households is needed to identify different motivations for saving. research on financial well-being using a lowto moderate-income stratified random sample is also needed with a greater representation of non-whites. future studies should examine both income and assets because these variables are not always linked. someone may temporarily have a low income but maintain a high net worth. elder and rudolph (2000) found that pension benefits may be a substitute for saving, so future researchers should consider including retirement benefits in studies on saving. to investigate the impact of current volatile economic conditions (that are atypical but may become the new norm), a longitudinal study is needed. in addition to the different wording of questions in the scf and the nc1172 surveys, it is essential to understand the economic circumstances leading up to and during the two data collection periods. although wealthier americans (oversampled in the scf data) were affected by the 2008 financial crisis, the impact was far more severe for lowand middle-income americans, many of whom lost jobs, are underwater or in foreclosure on their homes, and have fewer assets to cushion the economic shock (mishel & shierholz, 2011). single person households affected by the recession are at an even greater disadvantage than individuals with partners who provide the potential to earn income to help cushion economic blows. thus, singles are even more in need of emergency savings than partnered or married couple households. 10 p.j. fisher et al. / financial services review 24 (2015) 1–13 policy incentives to save regularly and consumer education to accompany those incentives are needed. with interest rates on savings currently below one percentage, today’s interest rate environment does not encourage savings, nor does the irs policy of taxing interest at the same rate as earned income while assessing taxes on capital gains and dividends at lower rates. savings are essential for the financial cushion needed by every household and to provide the capital for loans, but current government tax policies do not provide adequate incentives for lowto moderate-income households to save. one difference between the two data sets is that both income and wealth are low to moderate in the nc1172, whereas in the scf, we limited the sample to low to moderate income but these households have a relatively high net worth for individuals with incomes up to $80,000. although the measures of net worth were vastly different for the two data sets and the values provided by some nc1172 respondents were deemed to be unreliable, the mean net worth for the nc1172 sample was about $35,000 whereas that of the scf sample was about $150,000. thus, it is possible that households are not basing their saving and spending decisions only on income, and that wealth also plays a role in such decisions. we do not include wealth in our study because the measures of net worth in the two data sets are very different, so future studies that include net worth are needed. although many high income and high net worth individuals have access to financial planning services, financial educators and counselors are likely serving lowto moderateincome audiences. the most significant finding for counselors and financial educators serving middle-income clients is that less than one-third of lowto moderate-income singles save regularly. unemployment is a greater threat for single earners than for workers with a partner, so having an adequate emergency fund is critical. although both women and men need encouragement to save, in addition to education on effective strategies, women have different learning styles than men and benefit from education geared toward women and provided with other female learners (hays & flannery, 2000; lown, 2004). thus, educators and advisors should use examples that will appeal to women and men so that both genders feel that the information applies to them or alternatively provide gender-focused education. references abbi, s., hahnel, a., maynard, n., & smith-ramani, j. (2012). playing the savings game: a prize-linked savings report. (available at http://www.d2dfund.org/files/publications/playingtsg_report_2012_singlepgs.pdf). america saves. (2010). as america saves week begins, groups release survey suggesting recession has eroded savings practices and discuss their efforts to stop this erosion. annual america saves/american savings education council survey. (available at http://americasavesweek.org/media-room). bajtelsmit, v. l., bernasek, a., & jianakoplos, n. a. (1999). gender differences in defined contribution pension decisions. financial services review, 8, 1–10. bajtelsmit, v. l., & vanderhei, j. a. (1997). risk aversion and retirement income adequacy. in m. s. gordon, o. s. mitchell, & m. m. twinney (eds.), positioning pensions for the twenty-first century. philadelphia, pa: university of pennsylvania press. berger, e., & denton, m. (2004). the interplay between women’s work histories and financial planning for later life. canadian journal on aging, 23, 99–113. blau, f. d., & kahn, l. m. (1997). swimming upstream: trends in the gender wage differential in the 1980s. journal of labor economics, 15, 1–42. 11p.j. fisher et al. / financial services review 24 (2015) 1–13 blau, f. d., & kahn, l. m. (2000). gender differences in pay. journal of economic perspectives, 14, 75–99. board of governors of the federal reserve system. (2012). codebook for 2012 survey of consumer finances. division of research and statistics. (available from http://www.federalreserve.gov/econresdata/scf/files/ codebk2010.txt). chang, m. (2004). his and hers: explaining the gender wealth gap. annual meeting of the american sociological association. (available from http://www.allacademic.com/meta/p109260_index.html). chen, h., & volpe, r. p. (2002). gender difference in personal financial literacy among college students. financial services review, 11, 289–307. elder, h. w., & rudolph, p. m. (2000). beliefs and actions: expectations and savings decisions by older americans. financial services review, 9, 33–45. embrey, l., & fox, j. (1997). gender differences in the investment decision-making process. financial counseling and planning, 8, 33–40. faff, r., mulino, d., & chai, d. (2008). on the linkage between financial risk tolerance and risk aversion. journal of financial research, 31, 1–23. fisher, p. j. (2010a). black-white differences in saving behaviors. financial services review, 19, 1–16. fisher, p. j. (2010b). gender differences in personal saving behaviors. journal of financial counseling and planning, 21, 14–24. grable, j. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625–630. grable, j. e., & joo, s.-h. (2004). environmental and biopsychosocial factors associated with financial risk tolerance. financial counseling and planning, 15, 73–82. graham, j. f., stendardi, e. j., myers, j. k., & graham, m. j. (2002). gender differences in investment strategies: an information processing perspective. the international journal of bank marketing, 20, 17–26. hallahan, t. a., faff, r. w., & mckenzie, m. d. (2004). an empirical investigation of personal financial risk tolerance. financial services review, 13, 57–78. hays, e., & flannery, d. d. (2000). women as learners. san francisco, ca: jossey-bass. hayhoe, c. r., & gutter, m. (2012). reliability of the scales in the nc1172 complex nature of saving data set. family and consumer sciences research journal, 40, 284–294. hinz, r. p., mccarthy, d. d., & turner, j. a. (1997). are women more conservative investors? gender differences in participant-directed pension investments. in m. s. gordon, o. s. mitchell, & m. m. twinney (eds.), positioning pensions for the twenty-first century (pp. 91–106). philadelphia, pa: university of pennsylvania press. hira, t. m., & loibl, c. (2005). understanding the impact of employer-provided financial education on workplace satisfaction. journal of consumer affairs, 39, 173–194. hira, t. m., & mugenda, t. k. (2000). gender differences in financial perceptions, behaviors, and satisfaction. journal of financial planning, 13, 86–92. jianakoplos, n. a., & bernasek, a. (1998). are women more risk averse? economic inquiry, 36, 620–630. kennickell, a. (1997). multiple imputation and disclosure protection: the case of the 1995 survey of consumer finances. washington, dc: board of governors of the federal reserve system. kennickell, a., & mcmanus, d. a. (1993, august). sampling for household financial characteristics using frame information on past income. federal reserve board of governors. (available from http://www. federalreserve.gov/pubs/oss/oss2/papers/asa93.pdf). lerman, r. i., & steuerle, e. (2012, november). the two worlds of personal finance: implications for promoting the economic well-being of lowand moderate-income families. life-cycle investing: financial education and consumer protection, 85–96. doi: 10.2470/rf.v2012.n3.7. lewis, a., webley, p., & furnham, a. (1995). the new economic mind. the social psychology of economic behaviour. hemel hempstead: harvester wheatsheaf. loibl, c., & hira, t. k. (2006). a workplace and gender-related perspective on financial planning information sources and knowledge outcomes. financial services review, 15, 21–42. lown, j. m. (2004). women’s preferences for learning about financial planning. journal of personal finance, 3, 49–58. 12 p.j. fisher et al. / financial services review 24 (2015) 1–13 lupton, j., & smith, j. p. (2003). marriage, assets, and savings. in s. grossbard-shechtman (ed.), marriage and the economy: theory and evidence from industrialized societies (pp. 129–152). cambridge, england: cambridge university press. maddala, g. s. (1992). introduction to econometrics (2nd ed.). englewood cliffs, nj: prentice hall. montalto, c. p. (1998). the surveys of consumer finances: practical issues and challenges from the users’ perspective. consumer interests annual, 44, 232–233. montalto, c. p., & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances. financial counseling and planning, 7, 133–146. mishel, l., & shierholz, h. (2011, august 31). sustained, high joblessness causes lasting damage to wages, benefits, income, and wealth. washington, dc: economic policy institute briefing paper #324. (available from http://w3.epi-data.org/temp2011/briefingpaper324_final%20%283%29.pdf). neelakantan, u. (2010). estimation and impact of gender differences in risk tolerance. economic inquiry, 48, 228–233. o’neill, j. (2003). the gender gap in wages, circa 2000. american economic review, 93, 309–314. pearce, d. (1989). the feminization of poverty: women, work, and welfare. urban and social change review, 11, 28–36. prudential research. (2010). financial experience and behaviors among women. prudential research report 0182849–00001-00. (available from http://www.prudential.com/media/managed/womens_study_final.pdf). ricciardi, v. (2008). the financial psychology of worry and women. work in progress. (available from http://papers.ssrn.com/sol3/papers.cfm?abstract_id�1093351). rubin, d. b. (1987). multiple imputation for nonresponse in surveys. new york, ny: wiley. ryan, c. l., & siebens, j. (2012, february). educational attainment in the united states: 2009. current population reports, p20–566. (available from http://www.census.gov/prod/2012pubs/p20–566.pdf). schmidt, l., & sevak, p. (2006). gender, marriage, and asset accumulation in the united states. feminist economics, 12, 139–166. sierminska, e. m., frick, j. r., & grabka, m. m. (2010). examining the gender wealth gap. oxford economic papers, 62, 669–690. society of actuaries. (2010, december). the impact of retirement risk on women. risks and process of retirement survey report. (available from http://www.soa.org/files/pdf/research-key-finding-impact-retirewomen.pdf). sunden, a. e., & surrette, b. j. (1998). gender differences in the allocation of assets in retirement savings plans. american economic review, 88, 207–211. sung, j. (1997). a structural analysis of retirement funds in a family context: participation and investment in stocks. unpublished dissertation. the ohio state university, columbus, oh. wärneryd, k. e. (1999). the psychology of saving. a study of economic psychology. cheltenham: edward elgar publishing. warren, t., rowlingson, k., & whyley, c. (2001). female finances: gender wage gaps and gender asset gaps. employment and society, 15, 465–488. xiao, j., collins, m., ford, m., keller, p., kim, j., & robles, b. (2010, august). a review of financial behavior research: implications for financial education. consortium on implications of research in personal finance. national endowment for financial education. denver, co. yuh, y., & hanna, s. (1997). the demand for risky assets in retirement portfolios. proceedings of the academy of financial services. honolulu, hi. 13p.j. fisher et al. / financial services review 24 (2015) 1–13 the time perspective of financial advisors and its effect on their decision-making kenneth n. ryacka,* adepartment of accounting, quinnipiac university, hamden, ct, 06518, usa abstract psychological research suggests individuals often display past, present, or future time perspective (tp) biases that impact decision-making. this article examines the tp biases of financial advisors from different backgrounds and whether or not the biases impact client recommendations. consistent with literature that suggests a link between tp and career choice, advisors are future oriented as a group, regardless of their professional background. however, contrary to prior tp research, the bias does not appear to impact their professional decisions. instead, the findings are consistent with research that demonstrates psychological biases are mitigated when professional decision makers perform job related tasks. © 2015 academy of financial services. all rights reserved. jel classification: d8, d9, d14 keywords: time perspective; time orientation; financial advisors; financial decision making; personal financial planning, career choice researchers have noted that today’s complex financial and economic environment fuels an increasing need for personal financial services and those in need of such services commonly turn to a personal financial advisor for advice (trahan, gitman, and trevino, 2012). this is consistent with research that finds 21.8 million u.s households have sought some type of advice from professional financial advisors (elmerick, montalto, and fox, 2002). growth in the financial planning profession and the increased use of financial advisors in the united states has been linked to increasing life expectancies, earlier retirements, increases in the eligibility age for social security benefits, and rampant financial illiteracy in the general * corresponding author. tel.: �1-203-582-6550; fax: �1-203-582-8664. e-mail address: kenneth.ryack@quinnipiac.edu (k. ryack) financial services review 24 (2015) 387–410 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. population (warschauer, 2002). although the profession has seen significant growth, research on financial planning and financial planning professionals has been lacking, and academics have called for more research in this area (black, ciccotello, and skipper, 2002). i examine the time perspective of professional financial planners and its impact on the advice they give to clients. psychological research has found that individuals often display a bias toward a particular temporal frame of reference when making decisions. in other words, decision makers exhibit time perspective (tp) biases by focusing on the past, present, or future features of a particular decision (d’alessio, guarino, de pascalis, and zimbardo, 2003; nuttin, 1985; zimbardo and boyd, 1999). individuals with a strong present tp are more likely to focus on short-term goals and the immediate benefits of current actions. in contrast, people with a strong future tp are more likely to focus on long-term goals and the future consequences of current actions. the tp literature suggests that there may be a link between an individual’s tp and their career choice whereby individuals working in a profession that is more future (present) oriented may exhibit a future (present) tp bias (gonzalez and zimbardo, 1985). personal financial planning is a very forward oriented profession because it focuses on helping clients financially plan for future events in their lives. thus, professional financial advisors may be future oriented as a group. however, the profession is made up of a diverse group of advisors working in different types of firms (e.g., accounting, insurance, investment, etc.), including some career changers originally from nonfinancial backgrounds. if these advisors are not uniformly future oriented as a group, then it is possible that their tp may vary in accordance with their professional background. in either case, their tp may affect the recommendations they make to clients. an abundance of psychology research demonstrates that tp can affect decision-making in a variety of personal decision-making contexts, including some limited research related to financial choices. personal financial planning (pfp) provides an appropriate setting to study tp among financial professionals because pfp decisions naturally involve trade-offs between more future oriented and more present oriented options. the results of prior research suggest tp will affect pfp decisions. however, there is a lack of research examining the impact of tp on professional decision-making and other psychological research suggests its impact may be mitigated in a professional decision-making context. according to that research, individual characteristics may be less important than task characteristics in complex decision scenarios (e.g., beach and mitchell, 1978; chervany and dickson, 1978; huber, 1983; payne, 1982; payne, bettman, and johnson, 1992), and the impact of psychological biases is lessened when professional decision makers perform job-related tasks (see smith and kida, 1991, for a review). thus, tp may not be as influential in professional decision-making contexts as in personal contexts. however, if individual tp biases do impact professional decision-making, then dysfunctional consequences may arise. for example, if a financial planner is biased toward recommending a more future oriented strategy because the advisor has a strong future tp, then he or she may not be effective in helping clients achieve important shorter term financial goals. i investigate the tp biases that exist among professional financial planners, whether there is a relationship between their professional background (e.g., accounting, insurance, securities, etc.) and their tp, and the extent to which tp biases impact the recommendations they 388 k.n. ryack / financial services review 24 (2015) 387–410 make to clients. while there are some limited exceptions, the results generally show that the financial advisors are future oriented as a group, there is no significant relationship between their professional background and their tp, and there is no significant relationship between their tp and the recommendations they make to clients. the next section presents a literature review and hypotheses followed by the method section, a discussion of the results, and concluding remarks. 2. literature review and hypotheses 2.1. impact of tp on behavior psychological theory suggests individuals develop time perspectives from cognitive processes that partition experiences into past, present, and future temporal frames (d’alessio et al., 2003; nuttin, 1985; zimbardo and boyd, 1999). when a strong orientation toward one of these time frames exists, it serves as a cognitive bias or an individual-differences variable that affects judgment and decision-making (zimbardo & boyd, 1999). this bias is generally referred to in the literature as “time perspective” or “time orientation.” an individual with a strong present orientation will tend to focus on more current features of decision alternatives, such as their convenience and the immediate benefits they provide (strathman, gleicher, boninger, and edwards, 1994; zimbardo and boyd, 1999). in contrast, a person with a strong future tp focuses more on potential long-term outcomes and is willing to sacrifice current rewards in order to obtain some desirable future state (boniwell and zimbardo, 2004; strathman et al., 1994; trommsdorff, 1983; zimbardo and boyd, 1999). numerous studies have found significant relationships between tp and behavior in a variety of situations. for example, research has demonstrated that higher levels of alcohol, drug, and cigarette consumption are predicted by a stronger present orientation, while lower levels of substance usage are predicted by a stronger future orientation (e.g., henson, carey, carey, and maisto, 2006; keough, zimbardo, and boyd, 1999). studies have also found a similar type of relationship between tp and other socially unacceptable behaviors, such as youth delinquency (e.g., cauffman, steinberg, and piquero, 2005; modecki, 2008), risky driving (e.g., zimbardo, keough, and boyd, 1997), risky sexual practices (e.g., dorr, krueckeberg, strathman, and wood, 1999; rothspan and read, 1996), and pathological gambling (e.g., mackillop, anderson, castelda, mattson, and donovick, 2006; toplak, liu, macpherson, toneatto, and stanovich, 2007). in addition, tp has been linked to academic achievement (e.g., adelabu, 2007; joireman, 1999), environmental attitudes and behaviors (e.g., corral-verdugo and pinheiro, 2006; joireman, van lange, and van vugt, 2004), and individual health practices (e.g., orbell and hagger, 2006; ouellette, hessling, gibbons, reis-bergan, and gerrard, 2005). research has generally found that individuals with a strong future tp focus on the achievement of future goals and abstract features such as the desirability of future states resulting from their decisions (boniwell and zimbardo, 2004; d’alessio et al., 2003; strathman et al., 1994; zimbardo and boyd, 1999). in contrast, individuals with a more current orientation tend to be influenced by concrete features of decision alternatives, 389k.n. ryack / financial services review 24 (2015) 387–410 such as convenience and other more immediate benefits (strathman et al., 1994; zimbardo and boyd, 1999). strathman, gleicher, boninger, and edwards (1994) demonstrate this behavior in an oil drilling decision case that involves trade-offs between present and future benefits. their results indicate that participants with a strong future tp are more likely to favor offshore drilling when the disadvantages are presented as immediate and the advantages are presented as distant. however, those with a more present orientation are more convinced of the benefits of drilling when the advantages are immediate and the disadvantages are distant. in a personal finance context, joireman, sprott, and spangenberg (2005) demonstrate similar behavior when marketing students are asked how they would prefer to invest money received from a hypothetical windfall. more future oriented respondents tend to choose the options with longer-term benefits (i.e., paying down credit card debt or putting it into savings to cover college expenses), while the more present oriented generally prefer the options with short-term benefits (i.e., purchasing an item online that is temporarily on sale or going on a trip with friends). there is also research finding that future tp is associated with higher levels of perceived financial planning knowledge, retirement involvement, willingness to invest in a 401(k) plan, and a more aggressive retirement savings profile (hershey and mowen, 2000; howlett, kees, and kemp, 2008; jacobs-lawson and hershey, 2005). thus, there is a large amount of literature demonstrating the significant impact of tp on behavior in a variety of different settings, including personal finance. 2.2. relationship between tp and occupation i am aware of only one study to date that has explored the relationship between tp and occupation. gonzalez and zimbardo (1985) collected data from a sample of over 11,000 psychology today readers that completed the stanford time perspective inventory (stpi), an earlier version of the zimbardo time perspective inventory (ztpi). the results indicate that different types of jobs are associated with different tps. for example, students and a group of workers classified as “semiskilled /unskilled” tend to display stronger present orientations and weaker future orientations. in contrast, teachers, professionals, managers, and white collar workers are more likely to exhibit a strong orientation toward at least one dimension of future tp. based on these findings, gonazlez and zimbardo (1985, p. 26) conclude, “it seems likely that two processes are at work here. individuals select certain occupations because they already have the time orientation called for. once in the job, success and satisfaction depend on intensifying the orientation further.” since personal financial planning is a very forward looking profession, it seems plausible that such a career would attract and retain more future oriented individuals. therefore, i propose the following hypothesis: hypothesis 1: professional financial planners are generally future oriented as a group. 2.3. relationship between professional background and tp while financial planners may be future oriented as a group, they come from a number of different professional backgrounds (e.g., accounting, investment, insurance, etc.) and it is 390 k.n. ryack / financial services review 24 (2015) 387–410 possible that they display different time perspective biases related to their different professional backgrounds. for example, a tax accountant or an enrolled agent who spends most of their time working on the preparation of historical tax returns and planning for the short-term (i.e., upcoming months), may be less future oriented than other planners. this notion appears to be supported by practitioner literature which suggests that while cpas have a number of strengths to draw upon when entering the field of pfp (e.g., technical knowledge, quantitative skills, and analytical abilities), they may also have difficulty with helping individuals plan for the long-term (aicpa, 2005; o’reilly, 2000; wolosky, 2005). in contrast, a seasoned estate planning attorney or life insurance advisor, primarily working with clients on long-term issues that span a lifetime, may be more future oriented. if different tp biases are associated with professional experience, then there are important implications for financial professionals of various backgrounds practicing within the same field. therefore, my second hypothesis is as follows: hypothesis 2: financial planning practitioners from different backgrounds (e.g., accounting, insurance, investment) will exhibit different levels of future and present tp. 2.4. impact of tp on financial planning decisions an important question is whether tp biases affect professional financial planners’ judgment and decision-making, and ultimately, their recommendations to clients. as previously noted, there is a large body of research that demonstrates the impact of tp on behavior in a wide variety of personal decision-making contexts. in financial planning, a choice must often be made between a planning strategy that provides more current rewards to the client and a strategy that results in more long-term benefits. the tp research suggests that planners with a more present (future) orientation would be more likely to recommend the strategy with more current (long-term) benefits. however, that research has not examined professional decision makers performing job related tasks. other psychology research indicates that task characteristics can mitigate the importance of individual characteristics (e.g., beach and mitchell, 1978; chervany and dickson, 1978; huber, 1983; payne, 1982; payne et al., 1992), and that judgment biases may be reduced in settings where experienced decision makers perform tasks for which they have significant knowledge and familiarity (berkeley and humphreys, 1982; ebbesen and konecni, 1980; edwards, 1983; einhorn, 1976; fischoff, 1982, 1987; funder, 1987; hogarth, 1981; kida, moreno, and smith, 2010; smith and kida, 1991). for example, smith and kida (1991) review the research on professional auditor judgments and find that a number of heuristics and biases shown to affect judgment in prior research are often mitigated or modified in contexts where expert decision makers perform familiar tasks. thus, it seems unlikely that tp biases would significantly impact the decisionmaking of financial advisors performing realistic job related tasks. therefore, my third hypothesis is as follows: hypothesis 3: individual tp biases will not affect the financial planning recommendations made by professional financial advisors. 391k.n. ryack / financial services review 24 (2015) 387–410 3. method 3.1. participants to ensure that the participants were experienced professionals qualified to provide comprehensive personal financial planning services, only certified financial planners (cfp®) were used in the study. in order to earn and maintain the cfp® credential, a financial advisor must meet minimum education requirements, pass the comprehensive cfp® certification examination, meet a minimum experience requirement, abide by the cfp® board’s code of ethics and professional responsibility and rules of conduct, and comply with the financial planning practice standards (certified financial planner board of standards, 2015). i used data previously collected by ryack (2012). the sample includes 127 cfp® certificants that completed an online instrument. names of cfp® certificants were obtained through internet searches and through referrals from officers of various financial planning association branches. contact was initially made with approximately 210 cfp® certificants. the link to the online instrument was emailed to 180 qualified advisors who agreed to participate after 22 declined to participate and an additional nine were determined to be unqualified because they were not actively engaged in providing financial planning services. of that group, 136 planners actually logged on to complete the instrument and nine of them were excluded because their responses were only partially complete. the participants came from 17 states where, according to cfp® board statistics, roughly 75% of all cfp® certificants were registered at the time the study commenced (certified financial planner board of standards, 2006). as shown in table 1, the participants included 98 (77.2%) men and 29 (22.8%) women. this is consistent with the population of men (76.5%) and women (23.4%) holding cfp® certifications at the time the data were collected (certified financial planner board of standards, 2006). their mean age was approximately 46 years old and they had an average of approximately 11 years of experience as a financial planner. the sample included 19 planners from accounting firms, 18 from insurance firms, 32 from securities firms, and 58 from general financial planning firms. the number of respondents from the various backgrounds was not equal because it was often difficult to determine the type of firm a person worked for before making contact and there was no way to predict actual response rates. 3.2. instrument participants completed a three part online instrument. the first part included four short pfp cases that were created with the assistance of a former chair of the (cfp®) board examinations committee and were reviewed by three expert planners. the cases were also revised after reviewing the results of a pretest with planners from different backgrounds who were solicited in the same manner as the study’s participants. each case contained background information on a client and two alternative planning strategies. the three expert planners that reviewed the instrument before pretesting agreed that for all the cases, one strategy provided more current benefits whereas the other strategy had more long-term 392 k.n. ryack / financial services review 24 (2015) 387–410 benefits. the experts further agreed that these differences were easily identifiable as current versus long-term in each of the cases. after reading each case, the study participants were asked to respond to a six-point scale that indicated his or her preference for one of the alternatives as well as their degree of preference. scale responses one through three indicated a preference for the first alternative (1 � strong preference, 2 � moderate preference, 3 � slight preference), while four through six indicated a preference for the second alternative (4 � slight preference, 5 � moderate preference, 6 � strong preference). the cases were used specifically to test hypothesis 3 and they assessed the financial advisors’ decisions across a number of contexts often encountered in financial planning. various scenarios were presented that reflected multiple decision table 1 demographic statistics of the financial advisors variable n % age (in years) range 26–67 mean 46.2 median 47.0 gender male 98 77.2% female 29 22.8% highest degree held associate 2 1.6% bachelor 61 48.0% masters 47 37.0% jd 14 11.0% no response 3 2.4% type of firm employed at accounting 19 15.0% insurance 18 14.2% securities 32 25.2% general pfp 58 45.6% experience as a financial planner (in years) range 1–29 mean 10.9 median 9.0 certifications and licenses held cfa 1 0.8% cfp 127 100.0% cfs 2 1.6% chfc 26 20.5% cima 5 3.9% clu 22 17.3% cpa license 37 29.1% lutfc 10 7.9% pfs 19 15.0% real estate license 5 3.9% ria/ria rep 52 40.9% series 6 license 39 30.7% series 7 license 80 63.0% series 22,23 license 10 7.9% series 63,64 license 70 55.1% 393k.n. ryack / financial services review 24 (2015) 387–410 contexts and integrated several pfp subject areas, including cash flow analysis/planning, investment planning, income tax planning, education planning, retirement planning, and estate planning. the cases are presented in appendix a. the second part of the instrument consisted of 30 items used to determine the tp of the financial advisors. for each item, the planner was asked to rate how characteristic the statement was of him or her on an eight point scale ranging from extremely typical to slightly typical at one end of the scale and slightly atypical to extremely atypical on the other end of the scale. twelve items were a reproduction of all the items used in the consideration of future consequences (cfc) scale (strathman et al., 1994) and eighteen items were selected from the present hedonistic scale and the future scale contained in the zimbardo time perspective inventory (ztpi) (zimbardo and boyd, 1999). both scales were designed based on extensive factor analyses. the cfc scale is a one factor scale containing 12 items and it essentially measures tp on a continuum with present tp on one end and future tp on the other end. the ztpi contains 56 items that were found to load on five distinct tp factors; past-negative, past-positive, present-fatalistic, present-hedonistic, and future. in other words, each factor makes up its own scale that measures a unique and distinct aspect of tp. it has been common for researchers to incorporate only the relevant ztpi scales that measure the tp dimension of interest for their particular study. consistent with ryack (2012), i exclude items from the ztpi that measure present-fatalistic, past-positive and past-negative orientations because these factors have been shown to be associated with attitudes and behaviors that are either not relevant to this study, or unlikely to be found among professional financial planners.1 the future factor is characterized by planning for and achievement of future goals, while the present-hedonistic factor reflects an orientation toward present enjoyment and pleasure with a lack of consideration of future consequences. the scale items incorporated in this study and the ryack (2012) study include nine of the original ztpi future scale items that had at least a 0.40 loading in the original factor analyses conducted by zimbardo and boyd (1999) and nine of the original ztpi present-hedonistic scale items that had a loading 0.40 or above. three of the original present-hedonistic scale items and four of the original future scale items were excluded because they either had a negative loading or a loading below 0.40 in the original study by zimbardo and boyd (1999). three additional present-hedonistic items that loaded above 0.40 were excluded because they were either inappropriate for professional financial advisors or were redundant with another item. ryack (2012) tested this configuration of the ztpi present and ztpi future scales in a confirmatory factor analysis with a college student sample and found that it did not significantly alter scale integrity. most studies incorporating the cfc and ztpi items have used the original scales to measure tp. however, the scales were developed primarily with college student samples and some researchers have suggested the factor structures may vary across different populations (e.g., mitina and blinnikova, 2008; ryack, 2012; worrell and mello, 2007). in fact, there have been a number of studies that have tested and found different factor structures for the cfc and ztpi scale items (joireman, balliet, sprott, spangenberg, and schultz, 2008; joireman, shaffer, balliet, and strathman, 2012; petrocelli, 2003; ryack, 2012; toepoel, 2010). in the ryack (2012) study, financial planners exhibited multiple present and future factors beyond those measured by the original cfc and ztpi scales. an exploratory analysis 394 k.n. ryack / financial services review 24 (2015) 387–410 conducted with the 12 items in the cfc scale resulted in two present factors and two future factors from which four subscales were created (cfc_p1, cfc_p2, cfc_f1, and cfc_f2). the analysis of the nine items from the ztpi present-hedonistic scale resulted in two present factors from which two subscales were created (zpres_1 and zpres_2), while items from the ztpi future scale yielded three future factors (zfut_1, zfut_2, and zfut_3). in this study, i measure tp using the 12 scale measures from the ryack (2012) study. this includes the original one factor cfc scale, the adjusted one factor ztpi present-hedonistic scale, the adjusted one factor ztpi future scale, and all the subscales. the subscales are replicated in appendixes b and c. in the last part of the instrument, the financial planners completed a demographic questionnaire. the questions solicited background information such as the participant’s age, gender, professional background, pfp experience, education, and licenses and designations held. 4. results and discussion 4.1. tp scoring and scale descriptives as previously noted, the tp biases of the financial advisors were measured using the original cfc scale, modified versions of the original ztpi present-hedonistic and future scales, and the nine subscales from the ryack (2012) study. a planner’s score on each scale was calculated as the average of the planner’s ratings of each item within the scale, with seven of the items on the original cfc scale requiring reverse scoring. descriptive statistics for each scale are presented in table 2. a score of one (extremely typical) to four (slightly typical) indicates that the planner exhibits the tp measured by that scale, while a score of five (slightly atypical) to eight (extremely atypical) implies that the planner is not exhibiting that tp. in general, the results support hypothesis 1 and indicate that planners tend to be very future oriented. the mean and median scores for each future scale (ztpi future, cfc, zfut_1, zfut_2, zfut_3, cfc_f1, and cfc_f2) are between 1.9 and 2.9, which falls into the very typical to moderately typical range on an eight point scale ranging from very typical to very atypical. further analysis reveals that approximately 99% of the planners exhibit at least a slight future bias as measured by both the ztpi future scale and the cfc scale. on the other future subscales, 90–97% of the planners display at least some future bias, depending on the scale. except for a slight present orientation on the zpres_2 scale, the planners do not generally appear to be present oriented. the mean and median scores for the other present scales (ztpi present, zpres_1, cfc_p1, and cfc_p2) range from 4.4 to 6.5, meaning a present orientation is slightly to moderately atypical. thus, the results taken as a whole indicate that planners tend to be more future oriented than present oriented. this conclusion is supported by a paired samples t-test that shows a significant difference between the advisors’ mean scores on the ztpi present scale versus the ztpi future scale (means: future � 2.47 and present � 4.42, t � 17.961, p � 0.001). a similar result is found with a paired t-test comparing a combined cfc present scale 395k.n. ryack / financial services review 24 (2015) 387–410 composed of all the cfc present items from the cfc_p1 and cfc_p2 scales to a combined cfc future scale composed of all the cfc future items from the cfc_f1 and cfc_f2 scales (means: future � 2.84 and present � 6.06, t � 28.150, p � 0.001). 4.2. testing the relationship between background and tp the second hypothesis examines the relationship between tp and the professional background of financial planners. to investigate this hypothesis, i conducted a series of univariate analysis of variance (anova) tests for each tp scale using tp score as the dependent variable and firm type as the independent variable (i.e., accounting, insurance, securities, and general financial planning). only eight planners indicated they worked at some other type of firm. these planners were reclassified based on their answer to a question that asked them to indicate their primary job function. for example, if a planner indicated his or her primary job function was investment advising, then that planner was reclassified as working for a securities firm. as can be seen in table 3, out of the 12 tests conducted, only the ztpi present scale (f � 3.195, p � 0.026) and the zpres_2 scale (f � 4.540, p � 0.005) appear to be significantly affected by firm type. while a lower score on the ztpi present scale indicates a stronger overall present orientation, a lower score on the zpres_2 scale reflects a stronger desire to maintain the quality of one’s present lifestyle and to keep everyday life interesting. results for the zpres_1 scale (a measure of spontaneity and impulsiveness) are also marginally significant (f � 2.317, p � 0.079). tukey-kramer post hoc pairwise comparisons conducted for these three scales show that the differences are driven by advisors from the insurance table 2 time perspective scores of financial planners scale mean median min max sd ztpi present 4.42 4.44 2.11 6.67 0.95 ztpi future 2.47 2.33 1.22 5.00 0.67 cfc 2.90 2.92 1.17 4.58 0.66 zpres_1 5.10 5.00 2.00 7.25 1.07 zpres_2 3.78 3.75 1.50 6.75 1.13 zfut_1 2.68 2.50 1.00 6.50 0.84 zfut_2 1.93 2.00 1.00 6.00 0.93 zfut_3 2.54 2.33 1.00 5.33 0.96 cfc_p1 5.97 6.00 2.40 8.00 0.98 cfc_p2 6.35 6.50 3.50 8.00 0.85 cfc_f1 2.81 2.67 1.00 6.67 0.84 cfc_f2 2.88 2.50 1.00 7.00 1.21 the scale measures in this table include the original cfc scale, the modified versions of the original ztpi present-hedonistic and future scales, and the nine subscales from ryack (2012). the planners rated each item on an eight point scale (extremely typical to extremely atypical) and their scores for each measure were calculated as the average of the ratings for each item included in that measure. the present items in the original cfc measure are reverse scored. a scale score in the range of 1– 4 means the planner is exhibiting the tp measured by that scale (1 � extremely typical, 2 � very typical, 3 � moderately typical, 4 � slightly typical). a score in the range of 5– 8 means the planner is not exhibiting the tp measured by that scale (5 � slightly atypical, 6 � moderately atypical, 7 � very atypical, 8 � extremely atypical). 396 k.n. ryack / financial services review 24 (2015) 387–410 firms. the contrasts further indicate a significant difference on the ztpi present scale between planners working at insurance firms compared to accounting firms (mean difference � �0.939, p � 0.013). comparisons for the zpres_2 scale further reveal a significant difference between planners working in insurance versus accounting (mean difference � �1.263, p � 0.003), insurance versus securities (mean difference � �0.906, p � 0.026), and insurance versus general financial planning firms (mean difference � �0.797, p � 0.036). these results suggest that the advisors from the insurance firms are generally more present oriented than the advisors from the accounting firms. the advisors from the insurance firms also appear to have a stronger desire to maintain the quality of one’s present lifestyle and to keep everyday life interesting when compared to the advisors from all the other types of firms. however, there are no differences evident in any of the ztpi future scale measures or any of the cfc scale measures. in other words, except for the differences found with the advisors from the insurance firms on the ztpi present scale and the zpres_2 scale, there are no significant differences among the planners on any of the other 12 measures. 4.3. the impact of tp on financial planning decisions the third hypothesis investigates whether or not tp impacts the financial planning recommendations made by the financial advisors. four different personal financial planning cases were used to test this question (see appendix a). for each case, the planner reviewed two alternative planning strategies and then indicated his or her preference on a six-point scale, where one through three indicated a strong to slight preference for option a and four through six indicated a slight to strong preference for option b. i analyzed the impact of tp on the advisors’ recommendations using separate regressions for each case and each tp scale table 3 tp scale means (sd) and analysis of variance by firm type accounting (n � 19) insurance (n � 18) securities (n � 32) general financial planning (n � 58) f ratio p-value ztpi present mean (sd) 4.87 (.997) 3.93 (.771) 4.41 (.941) 4.44 (.932) 3.195 .026 ztpi future mean (sd) 2.44 (.665) 2.43 (.579) 2.48 (.808) 2.48 (.623) .035 .991 cfc mean (sd) 3.03 (.551) 2.81 (.601) 2.96 (.653) 2.86 (.715) .491 .689 zpres_1 mean (sd) 5.53 (1.03) 4.76 (.933) 4.88 (1.06) 5.16 (1.08) 2.317 .079 zpres_2 mean (sd) 4.26 (1.25) 3.00 (.879) 3.91 (1.19) 3.80 (1.01) 4.540 .005 zfut_1 mean (sd) 2.55 (.784) 2.53 (.630) 2.68 (1.04) 2.76 (.796) .507 .678 zfut_2 mean (sd) 1.92 (.917) 1.67 (.767) 2.03 (.772) 2.00 (1.06) .619 .604 zfut_3 mean (sd) 2.63 (.999) 2.82 (1.16) 2.51 (1.10) 2.45 (.791) .732 .535 cfc_p1 mean (sd) 6.00 (.833) 6.18 (.776) 5.78 (1.00) 5.99 (1.07) .674 .570 cfc_p2 mean (sd) 6.38 (.663) 6.47 (.835) 6.19 (.922) 6.38 (.869) .550 .649 cfc_f1 mean (sd) 3.00 (.667) 3.17 (1.24) 2.67 (.799) 2.72 (.734) 1.964 .123 cfc_f2 mean (sd) 3.24 (.903) 2.47 (.795) 2.86 (.891) 2.91 (1.21) 1.260 .291 this table shows the results of individual anovas for each tp measure. the advisors’ scores on the tp scale served as the dependent variable and their professional background as measured by the type of firm they worked for (i.e., accounting, insurance, securities, or general financial planning) was the independent variable. 397k.n. ryack / financial services review 24 (2015) 387–410 measure. in each regression, the preference ratings for a particular case were regressed on the scale scores for a particular tp measure. the results are presented in tables 4 through 7. for all 12 of the tp measures across the four scenarios (i.e., 48 analyses), there is only one significant result at the 0.05 level (see table 7). in the fourth case, the advisors are more likely to choose option a if they have a stronger present orientation as measured by the zpres_2 subscale. items in the zpres_2 subscale reflect a desire to maintain the quality of one’s present lifestyle and to keep everyday life interesting and that is consistent with a preference for option a because it maximizes current cash flow. with the exception of that one significant result, the third table 4 linear regressions of tp measure on case 1 score (n � 127, mean case score � 3.11) dependent variable standardized � coefficient t-stat p-value r2 ztpi present �0.074 �0.827 0.410 0.005 ztpi future �0.073 �0.819 0.414 0.005 cfc 0.051 0.567 0.572 0.003 zpres_1 �0.086 �0.964 0.337 0.007 zpres_2 �0.047 �0.529 0.597 0.002 zfut_1 �0.112 �1.265 0.208 0.013 zfut_2 0.036 0.399 0.690 0.001 zfut_3 �0.044 �0.498 0.620 0.002 cfc_p1 �0.062 �0.689 0.492 0.004 cfc_p2 �0.151 �1.710 0.090 0.023 cfc_f1 �0.025 �0.280 0.780 0.001 cfc_f2 �0.060 �0.669 0.505 0.004 each line of this table presents the results of a single linear regression of the advisors’ tp scores on their case preference scores. a lower score on the tp measure indicates a stronger orientation toward that tp. a lower score on the choice preference scale is indicative of a preference for option a, while a higher score is indicative of a preference for option b. in this case, option a had more future benefits, and option b had more present benefits. table 5 linear regressions of tp measure on case 2 score (n � 127, mean case score � 2.59) dependent variable standardized � coefficient t-stat p-value r2 ztpi present 0.021 0.238 0.812 0.000 ztpi future 0.084 0.939 0.350 0.007 cfc 0.006 0.069 0.945 0.000 zpres_1 �0.048 �0.535 0.594 0.002 zpres_2 0.070 0.779 0.437 0.005 zfut_1 0.048 0.541 0.590 0.002 zfut_2 �0.019 �0.209 0.835 0.000 zfut_3 0.130 1.470 0.144 0.17 cfc_p1 �0.042 �0.473 0.637 0.002 cfc_p2 �0.110 �1.233 0.220 0.012 cfc_f1 �0.026 �0.294 0.769 0.001 cfc_f2 �0.081 �0.903 0.368 0.006 each line of this table presents the results of a single linear regression of the advisors’ tp scores on their case preference scores. a lower score on the tp measure indicates a stronger orientation toward that tp. a lower score on the choice preference scale is indicative of a preference for option a, while a higher score is indicative of a preference for option b. in this case, option a had more present benefits, and option b had more future benefits. 398 k.n. ryack / financial services review 24 (2015) 387–410 hypothesis is supported because the financial planners’ tp generally does not appear to have a direct effect their planning decisions. thus, a planner’s background generally does not appear to impact their tp, and their tp does not appear to directly affect their decisionmaking. while it was not a formal research question, i also analyzed whether or not professional background impacted the planners’ decisions by performing an anova for each case with firm type as the independent variable and the preference rating as the dependent variable. again, no significant relationships were found. if the advisors’ tp affects their decisions, then one might expect the financial advisors would tend to select the future option in each case because they are future oriented. that, table 6 linear regressions of tp measure on case 3 score (n � 127, mean case score � 4.70) dependent variable standardized � coefficient t-stat p-value r2 ztpi present �0.094 �1.054 0.294 0.009 ztpi future �0.069 �0.778 0.438 0.005 cfc �0.131 �1.479 0.142 0.017 zpres_1 �0.073 �0.819 0.414 0.005 zpres_2 �0.088 �0.989 0.324 0.008 zfut_1 �0.055 �0.621 0.536 0.003 zfut_2 �0.060 �0.671 0.503 0.004 zfut_3 �0.042 �0.465 0.642 0.002 cfc_p1 0.121 1.367 0.174 0.015 cfc_p2 0.008 0.093 0.926 0.000 cfc_f1 �0.129 �1.453 0.149 0.017 cfc_f2 �0.082 �0.925 0.357 0.007 each line of this table presents the results of a single linear regression of the advisors’ tp scores on their case preference scores. a lower score on the tp measure indicates a stronger orientation toward that tp. a lower score on the choice preference scale is indicative of a preference for option a, while a higher score is indicative of a preference for option b. in this case, option a had more future benefits, and option b had more present benefits. table 7 linear regressions of tp measure on case 4 score (n � 126, mean case score � 3.70) dependent variable standardized � coefficient t-stat p-value r2 ztpi present 0.151 1.704 0.091 0.023 ztpi future 0.054 0.606 0.546 0.003 cfc �0.070 �0.785 0.434 0.005 zpres_1 0.072 0.798 0.426 0.005 zpres_2 0.177 2.008 0.047 0.031 zfut_1 �0.014 �0.153 0.879 0.000 zfut_2 0.076 0.849 0.397 0.006 zfut_3 0.081 0.907 0.366 0.007 cfc_p1 0.090 1.006 0.317 0.008 cfc_p2 0.112 1.252 0.213 0.012 cfc_f1 0.011 0.122 0.903 0.000 cfc_f2 �0.069 �0.772 0.441 0.005 each line of this table presents the results of a single linear regression of the advisors’ tp scores on their case preference scores. a lower score on the tp measure indicates a stronger orientation toward that tp. a lower score on the choice preference scale is indicative of a preference for option a, while a higher score is indicative of a preference for option b. in this case, option a had more present benefits, and option b had more future benefits. 399k.n. ryack / financial services review 24 (2015) 387–410 however, is not what happened. more advisors chose the future options in cases 1 and 4, while more advisors chose the present options in cases 2 and 3. thus, the results appear to support the hypothesis that time perspective does not impact the advisors’ decisions. an alternate argument can be made that time perspective is a factor, but something else is also at play. although the cases were designed with one option that yields more present consequences and one option that yields more future consequences, a planner may still be focused on the client’s future but select the option that has more present consequences because they feel it will be better for their particular client over the long-term. this certainly could have occurred, but it still points to some other factor at play that ultimately drives the planners’ choices. an examination of the results across all four cases suggests task characteristics are an overriding factor. cases 1 and 2 both present scenarios where the advisor is asked to choose an appropriate strategy given a single client goal. each case offers the client two alternatives, one with more favorable shorter-term implications and one with more favorable longer-term implications. in case 1, option a has more favorable future benefits and less favorable current benefits. drawing on the line of credit means the balance in the 401(k) will continue to earn interest at tax deferred rate of 12% and ultimately result in higher long-term wealth accumulation. however, the client will have a reduced short-term cash flow because they will have to pay 8% a year on the line of credit as opposed to only 6% on the 401(k) loan. in contrast, option b has more present benefits and less favorable future benefits. the short-term interest payments will be lower, but the growth in the 401(k) balance will also be lower. in case 1, approximately 60% of the planners expressed at least a slight preference for the more future oriented option (mean � 3.11). this appears to lend support to the idea that tp impacts the choice made since the planners tend to be future oriented as a group. however, the findings from the linear regressions indicate no significant relationships between any of the tp measures and the preference ratings for case 1. in case 2, the majority of planners selected the option with more present benefits. the investments in both options each produce an annual overall return of 12%. however, option a yields more present benefits because the increasing dividend payment each year produces a higher current cash flow. in contrast, option b produces a lower annual cash flow from dividends, but results in more capital growth and the tax on that growth is deferred until the investment is sold. even though the planners were generally future oriented, 72% expressed at least a slight preference for the more present oriented option (mean � 2.59). these data suggest that the characteristics of the task may be an overriding factor. in this case, for example, financial experts are likely to see income producing stocks with increasing dividends as less risky than growth stocks. in fact, anecdotal evidence from discussions with expert financial advisors who reviewed the case indicates that many planners may focus on the safety of the dividend stock even though the case states that both stocks have equal risk. in effect, the specifics of the decision context likely impact planners’ decisions to a greater extent than their general future orientation. case 3 presents a scenario where the planning options require a trade-off between a primary long-term goal of maximizing the transfer of wealth to beneficiaries and a primary 400 k.n. ryack / financial services review 24 (2015) 387–410 short-term goal of maximizing annual cash flow. similar to case 2, the planners indicated a preference for the option with more present benefits. option a results in significantly larger cash outflows while the grandchildren are attending college because mrs. johnson is paying their college expenses while she is also contributing to the trust. thus, it has more negative short-term consequences. however, it has more positive long-term consequences. the total wealth transferred to the grandchildren is a larger amount because mrs. johnson is in the highest marginal estate tax rate and this option will reduce the estate tax paid upon death. option b has more present benefits because of the lower cash outflow while the grandchildren are in college, but fewer future benefits because a smaller amount of total wealth is transferred to the grandchildren in the long-term. in this case, 70% of the planners indicated at least a slight preference for option b (mean � 4.70). once again, the primarily future oriented advisors overwhelmingly seem to favor the shorter-term goal. why might this occur? section 529 plans are a very popular college funding tool with many benefits, and it is likely that the advisors’ opinions are generally biased toward using them to fund college, regardless of the client’s other goals. in fact, one financial planner noted that he was “blinded” by the benefits of the 529 plan. thus, it appears that the task characteristics, once again, have more of an impact than do personal characteristics like tp. in case 4, advisors were asked to choose among strategies that required a trade-off between maximizing a primary long-term goal of transferring wealth in a tax efficient manner versus achievement of that primary long-term goal to a lesser extent in addition to a secondary short-term goal of maximizing current cash flow. option a achieves the primary goal by utilizing the unified tax credit to transfer the money and avoid estate and gift taxes paid by the client. it achieves the secondary goal because the trust now pays income taxes on the earnings instead of the client, thus increasing the client’s cash flow. option b results in a reduced cash flow for the client because an intentionally defective trust requires the client pay income taxes on the trust’s earnings. however, the tax rate the client pays is smaller and the amount of long-term wealth ultimately transferred is much larger. thus, option b maximizes the long-term benefits, while option a has more current benefits. approximately 44% of the planners chose option a versus 56% that chose option b (mean � 3.70). �2 tests indicate no significant difference between the number of planners choosing each option (�2 � 2.03, p � 0.154), and only one of the twelve regressions shows any significant relationship between tp and the option chosen. in summary, the future oriented planners sometimes chose the option with more present benefits and other times chose the option with more future benefits. it appears that the characteristics of the task override any affect that individual characteristics, such as tp, may have on the decision behavior of professional financial planners. this result supports hypothesis 3 and is consistent with prior research in behavioral decision-making which indicates that task characteristics have a major impact on decision-making (e.g., beach and mitchell, 1978; payne, 1982; payne et al., 1992). it is also consistent with research demonstrating that judgment biases are often mitigated when experienced decision makers perform tasks for which they possess significant knowledge or expertise (berkeley and humphreys, 1982; ebbesen and konecni, 1980; edwards, 1983; einhorn, 1976; fischoff, 1982, 1987; funder, 1987; hogarth, 1981; kida et al., 2010; smith and kida, 1991). 401k.n. ryack / financial services review 24 (2015) 387–410 5. conclusion the first hypothesis examines whether professional financial planners tend to display a particular time orientation as a group. gonzalez and zimbardo (1985) present preliminary results which indicate that tp may be associated with different occupations (e.g., teachers, students, managers, professionals, homemakers, semiskilled or unskilled workers, etc.). the researchers conclude that individuals most likely select a particular career because they have the tp called for by that career. gonzalez and zimbardo also suggest that success in the occupation further intensifies that tp bias. my results are consistent with those conclusions because i find that planners tend to be future oriented as a group. personal financial planning involves work that is a very forward looking and it is logical that such a career would be appropriate for future oriented individuals. the findings combined with the results of the gonzalez and zimbardo (1985) study suggest that tp could be used as a tool in helping individuals plan their career or in assisting companies screening candidates for specific types of positions. the second hypothesis explores whether an advisor’s professional background (e.g., accounting, insurance, investment, etc.) impacts their tp. while it was not a formal research question, i also examined whether professional background influenced decision-making directly. there is some concern raised in the practitioner literature that a planner’s professional background might affect their decisions and recommendations to clients. for example, the former head of the aicpa personal financial planning division indicated that cpas often have trouble with long-term planning (o’reilly, 2000). this might lead one to conclude that a planner’s background may affect their tp and their decisions. however, i do not find much difference between the planners’ tp based on their background, nor do i find that their background directly affects any of their decisions. thus, concerns that a planner’s professional background influences their decisions are mitigated. finally, the third hypothesis investigates the impact of tp on the decision-making of professional financial advisors. prior research has often found a link between tp and behavior in a variety of personal decision-making contexts. however, there is a lack of research examining the effect of tp in professional decision-making contexts. the finding that financial planners tend to be future oriented as a group suggests that their future orientation may have been a factor in their decision to select a career in personal financial planning. however, selection of an occupation is a personal decision context. i test the impact of tp on professional decision-making by examining the job-related decisions (i.e., client recommendations) that the financial advisors make. while the planners are generally future oriented, they do not consistently choose the more future oriented planning option. instead, their choices appear to be driven by the characteristics of the task. these findings are positive given that financial planning strategies need to be customized to an individual client’s goals, where the most future oriented strategy may not always be appropriate. in general, the results suggest that task characteristics may be more important than individual tp biases in the financial decisions made by professional advisors. these findings are consistent with the literature that points to the importance of task characteristics on decision-making (e.g., beach and mitchell, 1978; chervany and dickson, 1978; huber, 1983; payne, 1982; payne et al., 1992). the results are also 402 k.n. ryack / financial services review 24 (2015) 387–410 in line with prior research which demonstrates that judgment biases may be mitigated when experienced decision makers perform job-related tasks (berkeley and humphreys, 1982; ebbesen and konecni, 1980; edwards, 1983; einhorn, 1976; fischoff, 1982, 1987; funder, 1987; hogarth, 1981; kida et al., 2010; smith and kida, 1991). some limitations should be considered when evaluating the results of this study. first, the ability to generalize these results to the entire population of financial planners is limited because i used a convenience sample and because all of the participants held the cfp® certification. there are many other professionals in the practice of financial planning who do not hold the cfp® certification. in order to obtain certification, a candidate must meet specific educational requirements, pass a standardized exam, meet experience requirements, and agree to abide by a set of ethical standards and practice standards. these requirements may act to mitigate any differences between planners due to their different professional backgrounds. thus, future research might examine the tp and decision-making of financial advisors not holding the cfp® certification. the results do provide support for the theory that there is a link between an individual’s tp and their career choice because the financial planners tended to be future oriented as a group. however, i was not able to examine whether an individual’s tp bias is intensified with experience in the profession as suggested by gonzalez and zimbardo (1985). future research could examine this possibility. finally, this study only examines the role of the advisor in financial planning. however, financial planning is an interactive process between the advisor and the client. while i do not find that the planners’ tp affect their professional decisions (i.e., recommendations to clients), it is possible that the client’s tp may have a significant impact on their personal financial decisions. for example, tp could affect a client’s financial risk tolerance, credit attitudes, goals, priorities, and ability to implement their personal financial plan. thus, future research could investigate the impact of the client’s tp. acknowledgments i thank tom kida, tom langdon, tom brashear-alejandro, jim smith, caren rotello, pat o’neil, larry post, john napolitano, steve gill and officers of various financial planning association chapters for their assistance with this research. i also thank two anonymous reviewers for their helpful comments and suggestions. notes 1 for example, the present-fatalistic dimension is characterized by a helpless and hopeless attitude toward the future, the past-positive factor reflects a glowing or nostalgic construction of the past, and the past-negative orientation is associated with depression, anxiety and low self-esteem (zimbardo and boyd, 1999). 403k.n. ryack / financial services review 24 (2015) 387–410 appendix a: financial planning cases the cases below were designed with the assistance of a former chair of the (cfp®) board of examinations committee and reviewed by three expert planners. they all agreed that the cases were realistic and that one option clearly contained more present benefits while the other option contained more future benefits. after reading each case, the study participants were asked to respond to a six point scale that indicated his or her preference for one of the alternatives as well as their degree of preference. scale responses one through three indicated a preference for option a (1 � strong preference, 2 � moderate preference, 3 � slight preference), while four through six indicated a preference for option b (4 � slight preference, 5 � moderate preference, 6 � strong preference). case 1 john and mary parker are in their early forties and have been financial planning clients for a number of years. their daughter jessica is finishing her sophomore year in college and has recently decided to transfer to another college for her junior and senior years. tuition and fees at the new school are significantly higher and the college funds set aside by john and mary are not enough to cover the difference. since jessica is not eligible for any additional financial aid, john and mary have come to you for advice on the best way to fund the additional $14,000 need. you are considering the following two options, neither of which will have a significant impact on the comprehensive financial plan already in place for the parkers. option a john and mary have an available balance of $25,000 on an unsecured bank line of credit that they can draw on to pay the $14,000 tuition increase. since the loan is unsecured, interest will not be deductible. they currently have no plans to use the line of credit for another purpose. any amounts drawn on the line must be repaid at an interest rate of 8%, with principal and interest amortized over a five-year period. option b the $14,000 can be borrowed from john’s 401(k) plan, which has been performing better than your initial projections. the plan has been earning an annual average return of 12% and that return is expected to continue going forward. any loans borrowed from the plan will be paid back to john’s account at a rate of 6% interest with principal and interest amortized over a five-year period. assume that john will continue to work for his current employer during the loan period. case 2 recently, you helped create and implement a comprehensive financial plan for a wealthy couple, carol and mike jones. then, they unexpectedly inherited $50,000 in cash from a distant relative. since their existing financial plan adequately covers all their goals, they decide they would like to invest the money and come to you for advice. carol and mike are considering two different stocks. both stocks have an expected total annual return of 12% and the same beta (i.e., the same level of market risk). assume the tax rate on dividends remains the same as the capital gains tax rate during the investment holding period. option a invest the $50,000 in company a stock. the dividend yield on company a stock was approximately 3% last year. the company has increased its dividend by 10% each year, a policy it will continue into the future. the expected total return each year is 12%. option b invest the $50,000 in company b stock. the dividend yield on company b stock was approximately 2% last year. company b has historically paid the same annual dividend, a policy it plans to continue in the future. the expected total return each year is 12%. (continued) 404 k.n. ryack / financial services review 24 (2015) 387–410 appendix a continued case 3 mrs. johnson is a high net worth client with a taxable estate. she is in the highest marginal estate tax bracket (45%) and the highest marginal income tax bracket (35%). recently, mrs. johnson promised to pay for the college education of her grandchildren (ages 4–6). she would also like to provide them with an inheritance. she is fifty years old, in excellent health, and expects to be around long after her grandchildren complete college. an important goal for mrs. johnson is the maximization of her annual cash flow because she likes to travel and enjoy life’s luxuries. she would also like to maximize the transfer of wealth out of her estate, a goal she considers equally important. assume the tax rates do not change. you are considering the following two planning strategies. option a mrs. johnson will pay for her grandchildren’s educational costs as they arise in the future without any gift tax consequences since such transfers are not considered gifts. in addition, she will immediately begin making annual contributions to an irrevocable trust for the grandchildren to take advantage of the annual gift tax exclusion. taxes on any income generated by the trust assets will be paid by the trust. any assets remaining in her estate when she dies will be transferred by will to her grandchildren. option b mrs. johnson will immediately begin annual contributions to section 529 college tuition plans so she can take advantage of the annual gift tax exclusion and fund the education of her grandchildren. when the section 529 plans are adequately funded, she will use the annual gift tax exclusion to fund an irrevocable trust for the grandchildren. upon her death, all of mrs. johnson’s remaining assets will be transferred to her grandchildren through her will. case 4 you are helping mr. amherst, a high net worth client, with the creation and implementation of a comprehensive financial plan. mr. amherst enjoys his wealth and an important goal is the maximization of his cash flow to support his lavish lifestyle. his primary goal is the transfer of his wealth in a tax efficient manner. he would like to give $1,000,000 in unneeded income producing assets to his nephew. you are considering the following two options. option a transfer the $1,000,000 of income producing property to an irrevocable trust, avoiding gift tax liability through the use of the unified credit. all growth will be transferred to the beneficiary, but the trust will pay income tax at the highest marginal rate of 35%. option b transfer the $1,000,000 of income producing property to an intentionally defective grantor trust (idgt). since the trust is irrevocable, gift tax liability will be avoided through use of the unified credit and all growth is transferred to the beneficiary. because the trust is intentionally defective, the grantor (mr. amherst) will pay income tax at his marginal tax rate of 30%. 405k.n. ryack / financial services review 24 (2015) 387–410 appendix b: consideration of future consequences (cfc) sub-scale items cfc_p1 scale (� � 0.76) my convenience is a big factor in the decisions i make or the actions i take. my behavior is primarily influenced by the immediate (i.e., a matter of days or weeks) outcomes of my actions. since my day to day work has specific outcomes, it is more important to me than behavior that has distant outcomes. i primarily act to satisfy immediate concerns, figuring that i will take care of future problems that may occur at a later date. i primarily act to satisfy immediate concerns, figuring the future will take care of itself. cfc_p2 scale (� � 0.69) i think that sacrificing now is usually unnecessary since future outcomes can be dealt with at a later time. i primarily act to satisfy immediate concerns, figuring that i will take care of future problems that may occur at a later date. i generally ignore warnings about possible future problems because i think the problems will be resolved before they reach crisis level. i primarily act to satisfy immediate concerns, figuring the future will take care of itself. cfc_f1 scale (� � 0.58) i think it is more important to perform a behavior with important distant consequences than a behavior with less-important immediate consequences. i think it is important to take warnings about negative outcomes seriously even if the negative outcome will not occur for many years. i am willing to sacrifice my immediate happiness or well-being in order to achieve future outcomes. cfc_f2 scale (� � 0.71) often i engage in a particular behavior in order to achieve outcomes that may not result for many years. i consider how things might be in the future, and try to influence those things with my day to day behavior. the cronbach’s � for each sub-scale is reported in parentheses after each sub-scale name. the cfc sub-scales were created by ryack (2012) in a factor analysis of professional financial advisors’ responses to all of the original 12 cfc scale items (strathman et al., 1994). note that the following two items loaded on both the cfc_p1 and cfc_p2 factors and, therefore, are included in both sub-scales: “i primarily act to satisfy immediate concerns, figuring that i will take care of future problems that may occur at a later date” and “i primarily act to satisfy immediate concerns, figuring the future will take care of itself.” 406 k.n. ryack / financial services review 24 (2015) 387–410 appendix c: zimbardo time perspective inventory (ztpi) sub-scale items zpres_1 scale (� � 0.77) i do things impulsively. i find myself getting swept up in the excitement of the moment. i make decisions on the spur of the moment. i prefer friends who are spontaneous rather than predictable. zpres_2 scale (� � 0.72) i try to live my life as fully as possible, one day at a time. taking risks keeps my life from becoming boring. it is important to put excitement in my life. it is more important for me to enjoy life’s journey than to focus only on the destination. zfut_1 scale (� � 0.67) i am able to resist temptations when i know there is work to be done. meeting tomorrow’s deadlines and doing other necessary work comes before tonight’s play. i keep working at a difficult, uninteresting task if it will help me get ahead. i complete projects on time by making steady progress. zfut_2 scale (� � 0.79) i meet my obligations to friends and authorities on time. it upsets me to be late for appointments. zfut_3 scale (� � 0.62) i believe a person’s day should be planned ahead each morning. when i want to achieve something, i set goals and consider specific means for reaching those goals. i make lists of things i must do. the cronbach’s � for each sub-scale is reported in parentheses after each sub-scale name. the ztpi sub-scales were created by ryack (2012) in factor analyses of professional financial advisors’ responses to nine of the 15 original ztpi present-hedonistic scale items and nine of the 13 original ztpi future scale items (zimbardo and boyd, 1999). in conducting the factor analyses, the ryack (2012) study excluded three present-hedonistic scale items and four future scale items that the original zimbardo and boyd (1999) study found to have a negative factor loading or a loading below 0.40. three items from the original ztpi present-hedonistic scale were also excluded because two were felt to be inappropriate for professional financial advisors and one was redundant with another item that had a high factor loading. the factor analyses of the 18 items in the ryack (2012) study yielded two distinct present sub-scales (eight total items) and three distinct future scales (nine total items). one of the present items did not make it into the sub-scales because of its low factor loading. 407k.n. ryack / financial services review 24 (2015) 387–410 references adelabu, d. h. (2007). time perspective and school membership as correlates to academic achievement among african american adolescents. adolescence, 42, 525–538. aicpa. (2005). what are personal financial planning services and why should i get involved? retrieved from www.aicpa.org/innovation/fp/pfp/homepage.htm beach, l. r., & mitchell, t. r. (1978). a contingency model for the selection of decision strategies. academy of management review, 3, 349–449. berkeley, d., & humphreys, p. (1982). structuring decision problems and the ‘bias heuristic.’ acta psychologica, 50, 201–252. black, k., jr., ciccotello, c. s., & skipper, h. d., jr. (2002). issues in comprehensive personal financial planning. financial services review, 11, 1–9. boniwell, i., & zimbardo, p. g. (2004). balancing one’s time perspective in pursuit of optimal functioning. in p. a. linley & s. joseph (eds.), positive psychology in practice. hoboken, nj: wiley. cauffman, e., steinberg, l., & piquero, a. r. (2005). psychological, neuropsychological and physiological correlates of serious antisocial behavior in adolescence: the role of self-control. criminology, 43, 133–175. certified financial planner board of standards. (2006). cfp® certificant profile. retrieved from www.cfpboard. org/media/profile.asp certified financial planner board of standards. (2015). become a cfp® professional. retrieved from http:// www.cfp.net/become-a-cfp-professional/cfp-certification-requirements chervany, n. l., & dickson, g. w. (1978). on the validity of the analytic-heuristic instrument utilized in “the minnesota experiments”: a reply. management science, 24, 1091–1092. corral-verdugo, v., & pinheiro, j. q. (2006). sustainability, future orientation and water conservation. european review of applied psychology-revue europeenne de psychologie appliquee, 56, 191–198. d’alessio, m., guarino, a., de pascalis, v., & zimbardo, p. g. (2003). testing zimbardo’s stanford time perspective inventory (stpi) short form an italian study. time & society, 12, 333–347. dorr, n., krueckeberg, s., strathman, a., & wood, m. d. (1999). psychosocial correlates of voluntary hiv antibody testing in college students. aids education and prevention, 11, 14–27. ebbesen, e., & konecni, v. (1980). on the external validity of decision-making research: what do we know about decisions in the real world. in t. s. wallsten (ed.), cognitive processes in choice and decision behavior (pp. 21–45). hillsdale, nj: lawrence erlbaum associates. edwards, w. (1983). human cognitive capabilities, representativeness, and ground rules for research. in p. c. humphries, o. svenson, & a. vari (eds.), advances in psychology: analyzing and aiding decision processes (pp. 507–513). amsterdam: north holland. einhorn, h. (1976). a synthesis: accounting and behavioral science. journal of accounting research, 14 (suppl.), 196–206. elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217–231. fischoff, b. (1982). debiasing. in d. kahneman, p. slovic, & a. tversky (eds.), judgment under uncertainty: heuristics and biases (pp. 422–444). new york, ny: cambridge university press. fischoff, b. (1987). judgment and decision-making. in r. sternberg & e. smith (eds.), the psychology of human thought. new york, ny: cambridge university press. funder, d. (1987). errors and mistakes: evaluating the accuracy of social judgment. psychological bulletin, 101, 75–90. gonzalez, a., & zimbardo, p. g. (1985). time in perspective. psychology today, 19, 20–26. henson, j. m., carey, m. p., carey, k. b., & maisto, s. a. (2006). associations among health behaviors and time perspective in young adults: model testing with boot-strapping replication. journal of behavioral medicine, 29, 127–137. hershey, d. a., & mowen, j. c. (2000). psychological determinants of financial preparedness for retirement. gerontologist, 40, 687–697. 408 k.n. ryack / financial services review 24 (2015) 387–410 hogarth, r. (1981). beyond discrete biases: functional and dysfunctional aspects of judgmental heuristics. psychological bulletin, 90, 197–217. howlett, e., kees, j., & kemp, e. (2008). the role of self-regulation, future orientation, and financial knowledge in long-term financial decisions. journal of consumer affairs, 42, 223–242. huber, g. (1983). cognitive style as a basis for mis and dss designs: much ado about nothing? management science, 29, 567–579. jacobs-lawson, j. m., & hershey, d. a. (2005). influence of future time perspective, financial knowledge, and financial risk tolerance on retirement savings behaviors. financial services review, 14, 331–334. joireman, j. (1999). additional evidence for validity of the consideration of future consequences scale in an academic setting. psychological reports, 84, 1171–1172. joireman, j., balliet, d., sprott, d., spangenberg, e., & schultz, j. (2008). consideration of future consequences, ego-depletion, and self-control: support for distinguishing between cfc-immediate and cfc-future subscales. personality and individual differences, 45, 15–21. joireman, j., shaffer, m. j., balliet, d., & strathman, a. (2012). promotion orientation explains why futureoriented people exercise and eat healthy: evidence from the two-factor consideration of future consequences14 scale. personality and social psychology bulletin, 38, 1272–1287. joireman, j., sprott, d. e., & spangenberg, e. r. (2005). fiscal responsibility and the consideration of future consequences. personality and individual differences, 39, 1159–1168. joireman, j., van lange, p. m., & van vugt, m. (2004). who cares about the environmental impact of cars? those with an eye toward the future. environment and behavior, 36, 187–206. keough, k. a., zimbardo, p. g., & boyd, j. n. (1999). who’s smoking, drinking, and using drugs? time perspective as a predictor of substance use. basic and applied social psychology, 221, 149–164. kida, t., moreno, k., & smith, j. (2010). investment decision-making: do experienced decision makers fall prey to the paradox of choice? journal of behavioral finance, 11, 21–30. mackillop, j., anderson, e. j., castelda, b. a., mattson, r. e., & donovick, p. j. (2006). convergent validity of measures of cognitive distortions, impulsivity, and time perspective with pathological gambling. psychology of addictive behaviors, 20, 75–79. mitina, o., & blinnikova, s. (2008). adjusting the structure of the zimbardo time perspective inventory scales using russian data. international journal of psychology, 43, 42–42. modecki, k. l. (2008). addressing gaps in the maturity of judgment literature: age differences and delinquency. law and human behavior, 32, 78–91. nuttin, j. (1985). future time perspective and motivation: theory and research method. hillsdale, nj: erlbaum. o’reilly, m. (2000). cpas have entered the financial services arena: but will they score? ticker, 5, 36–44. orbell, s., & hagger, m. (2006). temporal framing and the decision to take part in type 2 diabetes screening: effects of individual differences in consideration of future consequences on persuasion. health psychology, 25, 537–548. ouellette, j. a., hessling, r., gibbons, f. x., reis-bergan, m., & gerrard, m. (2005). using images to increase exercise behavior: prototypes versus possible selves. personality and social psychology bulletin, 31, 610– 620. payne, j. w. (1982). contingent decision behavior. psychological bulletin, 92, 382–402. payne, j. w., bettman, j. r., & johnson, e. j. (1992). behavioral decision research: a constructive processing perspective. annual review of psychology, 43, 87–131. petrocelli, j. v. (2003). factor validation of the consideration of future consequences scale: evidence for a short version. journal of social psychology, 143, 405–413. rothspan, s., & read, s. j. (1996). present versus future time perspective and hiv risk among heterosexual college students. health psychology, 15, 131–134. ryack, k. (2012). evidence that time perspective factors depend on the group: factor analyses of the cfc and ztpi scales with professional financial advisors. personality and individual differences, 52, 723–727. smith, j., & kida, t. (1991). heuristics and biases: expertise and task realism in auditing. psychological bulletin, 109, 472–489. strathman, a., gleicher, f., boninger, d. s., & edwards, c. s. (1994). the consideration of future consequences: 409k.n. ryack / financial services review 24 (2015) 387–410 weighing immediate and distant outcomes of behavior. journal of personality and social psychology, 66, 742–752. toepoel, v. (2010). is consideration of future consequences a changeable construct? personality and individual differences, 48, 951–956. doi: 10.1016/j.paid.2010.02.029. toplak, m. e., liu, e., macpherson, r., toneatto, t., & stanovich, k. e. (2007). the reasoning skills and thinking dispositions of problem gamblers: a dual-process taxonomy. journal of behavioral decision-making, 20, 103–124. trahan, e. a., gitman, l. j., & trevino, m. d. (2012). virtual financial planners: product, market, and challenges. financial services review, 21, 95–129. trommsdorff, g. (1983). future orientation and socialization. international journal of psychology, 18, 381–406. warschauer, t. (2002). the role of universities in the development of the personal financial planning profession. financial services review, 11, 201–216. wolosky, h. (2005). rating cpas as financial planners. retrieved www.cpa2biz.com worrell, f. c., & mello, z. r. (2007). the reliability and validity of zimbardo time perspective inventory scores in academically talented adolescents. educational and psychological measurement, 67, 487–504. zimbardo, p. g., & boyd, j. n. (1999). putting time in perspective: a valid, reliable individual-differences metric. journal of personality and social psychology, 77, 1271–1288. zimbardo, p. g., keough, k. a., & boyd, j. n. (1997). present time perspective as a predictor of risky driving. personality and individual differences, 23, 1007–1023. 410 k.n. ryack / financial services review 24 (2015) 387–410 academy of financial services officers president robert moreschi virginia military institute president-elect duncan williams western carolina university executive vice president-program swarn chatterjee university of georgia vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president william chittenden texas state university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama halil kiymaz rollins college frances lawrence louisiana state university tom potts baylor university janine scott massey university martin seay kansas state university past presidents thomas coe quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2016 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 25, no. 4, 2016 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2016 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. finser_23_1 performance and persistence of performance of healthcare mutual funds abhay kaushika,*, lynn k. saubertb, r. wayne saubertc a,b,cdepartment of accounting, finance, and business law, college of business and economics, radford university, radford, va 24142, usa. abstract this study analyzes 115 actively managed domestic healthcare mutual funds over the period 1/2000–12/2011. findings of this study show that, on average, healthcare mutual funds outperform the passive index by roughly 2.97% per year after controlling for the market risk premium, growth and size premiums, and momentum effects. further, this study documents that the abnormal overand under-performance does not persist over subsequent periods. in other words, underand overperformances are mean reverting. © 2014 academy of financial services. all rights reserved. jel classifications: g23; g11; g12 keywords: healthcare mutual funds; performance; persistence of performance 1. introduction performance evaluation of mutual funds has generated a great deal of interest among academic researchers, practitioners, retail investors, and financial services professionals. according to the investment company institute (ici) factbook 2013, !53.8 million american households and roughly 92.4 million individuals in the united states own mutual funds. these towering figures demonstrate the significance of mutual funds in the lives of common investors in the united states. it is particularly interesting to evaluate the performance of actively managed mutual funds to determine whether the value of active management outweighs the cost of the expert management. general findings on equity funds have shown that funds do outperform * corresponding author. tel.: "-540-831-6426; fax: "1-540-831-6209. e-mail address: akaushik@radford.edu (a. kaushik) financial services review 23 (2014) 77–91 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. corresponding benchmarks on a gross return1 basis, but there seems to be a consensus that actively managed funds, in general, perform poorly compared to comparative benchmarks after considering expenses.2 although performance evaluation of active funds has been a center piece of attraction for retail investors, practitioners, and academics, persistence of performance has garnered momentum over the last decade. more specifically, both investors and practitioners are interested to know whether funds are able to repeat their performance or whether performance is a random outcome and eventually abnormal performance (positive or negative) follows the mean reversion theory. in other words, market efficiency holds and any over or under performance is mere “luck.” though a number of studies are available to explain the abnormal performance and persistence of performance for well-diversified funds, only a handful studies exist to evaluate the same phenomenon for sector-specific funds. this research extends the existing literature and analyzes 115 actively managed domestic equity healthcare mutual funds that existed at some point in time over the period 2000–2011. healthcare mutual funds have experienced an astounding growth over the last 10 years. for example, only 38 actively managed healthcare mutual funds existed in the year 2000 and that number increased to 56 by the end of 2011, a combined growth of 47% over the last 12 years.3 though the number of funds varies from year to year, overall this study includes 115 healthcare funds that existed at some point in time over the entire period of january 1, 2000 to december 31, 2011. on the contrary, the growth rate was a paltry 2.09% for the combined mutual fund industry over the same period.4 the extraordinary growth in healthcare funds warrants a thorough investigation behind the popularity of these types of funds. the main objectives of this research are to assess (1) the abnormal performance of actively managed mutual funds that mainly invest in health-care stocks, (2) the persistence of performance of these funds, that is, whether funds are able to repeat their abnormal performance or not, and (3) which cross-sectional attributes may explain the abnormal performance. the article is timely and offers a good insight into a very important sector of the u.s. economy. for example, the u.s. economy has been in slowdown mode since the early 2000s and some analysts argue that defensive stocks, like healthcare, should perform better in these conditions. portfolio managers like to invest in health care stocks because they believe this industry not only weathers business cycles but also includes stocks that create secular demand.5 according to zacks investment research “the health care is one of the most desirable avenues for parking investments when markets are headed south. the demand for such services usually remains unchanged even during an economic downturn and investments in the sector provide sufficient protection to the capital invested.” on the other hand, this sector has been on the “hot-seat” politically. although lawmakers are homogenous in terms of their support to change the healthcare system in the country, they differ significantly in their approach to modify the healthcare system. another important aspect of healthcare stocks is the cost of developing new drugs and loss of revenue after the expiration of patents. according to an article published in the forbes magazine “the average cost of bringing a new drug to market is $1.3 billion whereas the average drug developed by a major pharmaceutical company costs at least $4 billion, and it can be as much as $11 billion.”6 pharmaceutical firms invest billions of dollars to invent new drugs, but they lose a large portion of their profits at the time of expiration of their patents. for example, the new york times article dated march 6, 2011 reported an estimated $10 billion a-year-loss in 78 a. kaushik et al. / financial services review 23 (2014) 77–91 revenue for pfizer, a well-known pharmaceutical firm, at the expiration of patents for its famous cholesterol drug lipitor. “the loss poses a daunting challenge for pfizer, one shared by nearly every pharmaceutical company.”7 given the extreme volatility in this sector, some analysts argue that this sector should underperform compared to the broad market. finally, the number of funds in this sector fluctuated quite vigorously over the last decade. for example, the number of healthcare funds increased rapidly and reached to a record level of 88 funds in 2007 since then healthcare funds experienced a sudden decline and the number reduced to 56 funds by the end of 2011. given the extreme volatility in this sector, difference of opinions among analysts regarding this sector’s performance, and continued political debate over the healthcare reforms, the performance evaluation of healthcare funds become more interesting and informative. this research argues that the healthcare stocks offer volatility, but they also offer opportunities. individual investors may not have resources and abilities to select the right mix, but fund managers, especially those who only invest in this sector, should be able to use their information to their advantage. in this research, we analyze whether expert managers have abilities to select the right mix of healthcare stocks? if so, are they able to repeat their performance? in other words, are these managers simply lucky or are they really able to pick good stocks? what fund specific attribute(s) may have explanatory power behind any such abnormal performance? the remainder of this article is structured as follows: section 2 reviews the existing literature on sector and equity mutual funds; section 3 describes the data; section 4 describes methodology; section 5 summarizes the empirical results; and section 6 concludes the article. 2. literature review sector funds are mutual funds with a narrow focus.8 in essence sector specific funds only invest in a particular sector/industry of the economy; therefore, these funds are not as diversified as other equity funds. because of the homogenous risk in their holdings, sector funds tend to be more volatile; however, some may argue that their niche focus should give fund managers more information about the industry in which they invest and in turn, these funds earn higher return despite taking more risk. in other words, sector fund managers may possess better selectivity skills than their well-diversified counterparts. another argument that might support the diversification of sector funds stems from the fact that although these funds are focused on a specific sector, they do invest in different geographic locations, thus they are not exposed to locational risk exposure. finally, the argument that these funds are small is not necessarily true as some of these funds only invest in large-cap firms within the given sector. in other words, it is no surprise that some sector funds are larger in terms of assets under management compared to some well-diversified funds. a number of studies have examined the performance of sector specific funds. the existing research is limited on its findings on sector funds performance and most of these findings are mixed at best or inconclusive at worst. for example, khorana and nelling (1997) examined a sample of 147 sector funds to find that these funds do not significantly outperform well-diversified funds or at best perform as good as other general equity funds. moreover, 79a. kaushik et al. / financial services review 23 (2014) 77–91 they did not find persistence in performance of their sample of sector funds. their sample also suggests higher risk in sector funds compared to general equity funds. however, in another study on sector funds, burlacu and fontaine (2003) analyzed 102 sector funds and find that sector funds outperform well-diversified funds. their results are robust across different benchmarks. thus, their results are very different from those reported by khorana and nelling (1997) who not only reported no over or under performance by sector funds but also suggested that benchmark selection plays a significant role in estimating abnormal performance of sector funds. tiwari and vijh (2004) analyzed more than 600 sector funds and find no over or under performance of these funds. their abnormal performance models include multifactor models, both conditional and unconditional. brooks and porter (2012) analyze equity funds over the period 1994–2005 and find that fund managers were able to allocate funds adequately across sectors but failed to pick superior stocks. however, they also find that fund managers were able to do both during the bear market cycles. on the other hand, kacperczyk, sialm, and zheng (2005) found superior abnormal performance by funds that concentrated on a few industries. nan and yan (2011) analyze fund performance from another perspective. they analyze performance of chinese equity funds. they analyze funds based on two different models-the sharpe ratio and asset pricing model. their results show that actively managed chinese funds can be a better alternative if investors are looking at the total risk-adjusted returns, but they should invest in indexed funds if they are looking at market risk-adjusted returns. special sector funds like real-estate sector funds also have mixed results of underor over-performance. for example, lafever and canizo (2005) found persistence in the performance of real estate mutual funds on gross return basis that was also true when net returns were used though the performance was less persistent compared to gross returns basis. on the other hand, lin and yung (2004) found no positive performance by real estate funds and they also documented fund performance persisted in short run. dellva, demaskey, and smith (2001) found that sector funds outperform passive markets when the s&p 500 was used as the benchmark; however, results improved dramatically when the benchmark was replaced by dow jones industry and subindustry indices. more recently, kaushik, barnhart, and pennathur (2010) analyzed roughly 1,500 sector funds over the period 1990–2005 including healthcare funds and found that healthcare sector funds outperform the market and they did show similar results in recession periods. their study, however, ignored the most recent recession known as the longest recession since the end of the great depression. larry and weigand (1998) analyze the persistence effect by using the characteristics of holdings rather than using the factor mimicking portfolios model. their results indicate that investors should pay more attention to the trends in the overall market rather than chasing the past performance. fan and addams (2012) analyzed the united states based international mutual funds over the period 2005–2009 and found that these funds outperform passive benchmarks. their findings did not find strong persistence of performance effect and fund specific attributes also were not so relevant for their sample of funds. philpot (2000) evaluates persistence effect for nonconventional bond funds. findings of his study also show that persistence effect is limited only to high yield bond funds. manakyan and liano (1997) also failed to find evidence of persistence effect for funds that were closed to new investors. given the wide asymmetry in the findings of sector funds and other 80 a. kaushik et al. / financial services review 23 (2014) 77–91 specialized funds like real-estate funds, it is relevant to study narrow focused healthcare sector funds especially when this sector of the economy has a very different level of risk and returns and also experienced a big swing in terms of growth over the last decade or so. 3. data and descriptive statistics the majority of the data are taken from the morningstar direct database. only those domestic equity funds that invest in health-care stocks are selected from this dataset. because this research is examining the performance of actively managed domestic equity funds; therefore, funds that are classified as index funds, international funds, hybrid funds, bond funds, fund of funds, global funds, and quant funds are screened out from the selection process. monthly returns of fama-french (ff) (fama and french, 1993) factors such as smb (difference in returns between small and large capitalization stocks), hml (difference in returns between high and low book-to-market stocks), carhart momentum factor, mom (difference in returns between stocks with high and low past returns), and monthly risk free returns and monthly crsp value weighted returns are taken from the web site of kenneth french. monthly returns of sample funds, fund and manager specific variables such as turnover ratio, expense ratio, manager tenure, average-market capitalization of holdings, total net assets, fund’s investment in its top 10% holdings, cash holdings, investment in common stocks and number of holdings are taken from the morningstar direct database. most of the fund specific variables are reported on monthly basis except a fund’s expense ratio and turnover ratio; therefore, funds’ expense ratios and turnover ratios are divided by 12 to find their monthly equivalents. initial screening gives a sample of 124 funds; however, based on the existing literature, we selected only those funds that have at least 36 monthly observations. after removing funds with less than 36 monthly observations, 115 funds are available for empirical purposes. table 1 reports descriptive statistics of the sample funds. the average size (net assets under management) over the 12 year period is $168.50 million that suggests that most of the sample funds are small-cap funds. the investment in cash and common stocks is pretty constant over the period of this study. on average, healthcare funds invested roughly 96% of funds in common stocks and 4% in cash. the turnover ratio varies significantly over the 12 year period. the highest turnover ratio of 584% is observed in year 2000 whereas the lowest of 125% is observed in year 2006. the change in turnover ratio was dramatic; it dropped by almost 50% from an unusually high turnover in year 2000 to year 2001 and then increased for the next two years followed by a sharp decline in the next few years. even though most of these funds are small-cap funds, the average mean market cap of holdings of these funds is roughly $15 billion dollars that suggests that these funds tend to invest in a few large-cap stocks and a large proportion of funds is invested in small-cap stocks. the highest number of funds is found in year 2007 and the lowest in year 2000. on average, 101 funds existed per year over the 12 year period. 81a. kaushik et al. / financial services review 23 (2014) 77–91 4. methodology the sharpe (1964) capital asset pricing model (capm) is the commonly used model to price assets. the capm states that in equilibrium, expected returns are linearly related to their level of risk, more specifically, their ! or systematic risk. many tests and models have been developed over the years to measure performance of the mutual funds/funds’ managers. jensen’s (1968) " is perhaps the best known primary model. rit # rft $ "i % !i ! rmrft % &i,t (1) where: rit # rft is the excess return on fund i over the monthly treasury bill rate, "i is the measure of the portfolio’s performance (jensen’s "), rmrft $ rmt # rft is the excess return on the market (crsp value weighted index), and !i $ is the unconditional measure of risk. existing research indicates that returns on equities are heavily influenced by size, growth factor, and past returns besides market risk premium and these factors are commonly known as the ff factors and the carhart momentum factor. to control any biases that may inflate ! loading of the market factor, we estimated the four-factor model of carhart (1997), which table 1 descriptive statistics year annual turnover ratio (%) cash (%) equity (%) top (%) expense ratio (%) tna (in $ million) average market cap (in $ million) manager tenure (in years) n 2000 583.40 5.35 94.30 43.20 1.75 358.00 13,160.42 6.09 79 2001 271.30 4.18 95.28 44.87 1.72 188.70 16,453.96 6.10 95 2002 284.95 4.14 93.49 46.41 1.91 141.55 13,178.22 6.14 106 2003 361.08 4.10 95.61 41.22 1.98 123.05 10,738.15 6.07 105 2004 195.02 3.03 96.85 41.44 1.93 139.98 13,617.64 6.00 108 2005 166.90 3.42 96.45 44.64 1.87 154.41 16,720.19 5.96 110 2006 124.59 3.39 96.37 44.58 1.83 159.33 16,598.92 6.08 111 2007 128.78 2.95 96.45 45.48 1.82 150.67 17,675.78 6.07 112 2008 137.50 4.58 94.83 47.60 1.81 138.69 16,874.75 6.12 110 2009 207.71 3.61 95.44 46.23 1.88 121.41 14,182.13 6.11 103 2010 171.37 3.17 95.95 44.56 1.81 148.96 12,288.87 6.07 88 2011 167.61 4.37 94.71 42.16 1.73 197.30 13,321.17 6.36 87 average 233.35 3.86 95.48 44.37 1.84 168.50 14,567.52 6.10 101 note. table 1 shows the mean values of fund specific variables per year over the period 01/2000 to 12/2011. turnover ratio is the minimum of aggregated sales or aggregated purchases of securities divided by the average 12-month total net assets of the fund, cash is the average percentage of investment held as cash, equity is the average percentage of investment in common stocks, top is the fund’s percentage of investment in top 10% holdings, expense ratio is the average expense ratio charged by the fund, tna is the average annual net assets under management, average market cap is average market value of a fund’s holdings, manager tenure is the average manager tenure of the fund manager and n is the number of funds per year. 82 a. kaushik et al. / financial services review 23 (2014) 77–91 adjusts fund excess return for the ff factors smb, hml, and the carhart’s momentum factor. rit # rft $ "i % !1i ! rmrft % !2i ! smbt % !3i ! hmlt % !4i ! momt % &i,t (2) where: rmrft is the excess monthly return (market return net of one month t-bill return) on the crsp value weighted index smbt is the difference in returns between small and large capitalization stocks hmlt is the difference in returns between high and low book-to-market stocks momt is the difference in returns between stocks with high and low past returns the use of ff factors is significant for this research as table 1 shows that most of the sample funds are small in size and because of the narrow focus inbuilt in our sample funds, book-to-market and momentum should play an important role in explaining excess returns and abnormal performance of these funds. monthly smbt, hmlt, and momt factors are taken from the kenneth french web site. 4.1. persistence of performance in this section, we evaluate whether past winners can continue to outperform in the subsequent period(s) and past losers continue to poorly perform in the subsequent period(s). if this is true then investors and fund managers can definitely take advantage of a continuous trend to their favor. similar to existing research (e.g., berk and green, 2004; carhart, 1997; saap and tiwari, 2004), at the beginning of each year starting from 2001, we divided funds into quintiles based on their previous year’s returns. at the beginning of each year, funds are ranked by their annual returns in the previous year and quintiles are formed. funds in the bottom 20th percentile returns (q1 also labeled as “losers”) are clustered together as one group whereas funds in the top 20th percentile returns (q5 also labeled as “winners”) are grouped together. because funds’ performance may change every year, therefore, q1 and q5 are rebalanced every year. every year, beginning of the year, same step is repeated and two series are formed over the 2000–2011 period where 2000 is the first formation period and 2011 is the last estimation year. we use eq. (2) to estimate " of q1 and q5 to evaluate the persistence effect of past losers and winners. 4.2. cross-section analysis existing research (e.g., carhart, 1997; chan et al., 2004; wermers, 2000, 2003 among others) has shown that a fund’s abnormal performance is heavily affected by its attributes such as expense ratio, turnover ratio, size, investment in best ideas, cash holdings, to name a few. moreover, it is also a known fact that funds differ from each other in terms of their attributes; therefore, cross-sectional analysis plays a significant role in explaining the average abnormal performance of a portfolio of different funds. finally, funds may have uneven number of observations, that is, some funds might have existed the entire period of the study 83a. kaushik et al. / financial services review 23 (2014) 77–91 whereas some others may have shorter life span. to avoid any bias from the funds with more observations than others, we follow the existing literature and use fama and macbeth (1973) methodology to estimate the cross-sectional effects of fund specific attributes on funds’ ". under fama and macbeth methodology, the entire data are divided by time period (months in our study) and " and !s of each period are estimated. the average of all the monthly "s and !s is the portfolio’s cross-sectional " and !s as opposed to the average of "s and !s of individual funds. we follow existing research (e.g., brown, harlow, and starks, 1996; carhart, 1997; wermers, 2000 among others) to estimate dependent variable " by using the following model. "it $ rit # rft # !̃1it ! rmrft % !̃2it ! smbt % !̃3it ! hmlt % !̃4it ! momt (3) !̃1it ! . . . !̃4it are the ! loading estimated using eq. (2). once "s are estimated by using eq. (3) then we use these "s as the dependent variable and apply fama and macbeth (1973) to estimate the cross-sectional effects of fund specific attributes on funds’ abnormal performance. we use the following model to estimate the impact of cross-sectional effects. "it $ !0 % !1 expense ratioit % !2 turnoverit % !3 sizeit % !4 topit % !5 tenureit % !6 cashit % &it where: expense ratio is the average expense ratio charged by the fund turnover ratio is the minimum of aggregated sales or aggregated purchases of securities divided by the average 12-month total net assets of the fund size is the log value of a fund’s monthly tna top is the fund’s percentage of investment in top 10% holdings, tenure is the average manager tenure of the fund manager and it is log value of manager tenure cash is the average percentage of investment held as cash 5. empirical results 5.1. abnormal performance to estimate abnormal performance, we use both the single-factor market model and the carhart (1997) four-factor model. results in table 2 show that abnormal performance (") is positive and statistically highly significant for both models. on average, healthcare mutual funds outperformed the passive index by 0.2847% per month (3.42% per year) when the single-factor model is used. the abnormal performance is 0.2473% per month (2.97% per year) when the four-factor model is used. as mentioned above, both "s are statistically highly significant. in both models, the coefficients for the market are also positive and highly significant, thus suggesting a strong positive correlation between funds and market returns. furthermore, the four-factor model shows positive and statistically significant ! loadings for 84 a. kaushik et al. / financial services review 23 (2014) 77–91 smb and mom factors. these results indicate that abnormal performance of sample funds is heavily influenced by size and past returns of their holdings. of interest to the author, descriptive statistics in table 1 shows that majority of funds in our sample are small cap funds with average size of $168.50 million over the time period of this study. another interesting observation is the coefficient of hml that is statistically highly significant and negative. the negative coefficient of hml indicates positive value premium for healthcare mutual funds’ abnormal performance. results in table 2 suggest that healthcare mutual funds have potential to outperform the market after controlling for the market risk premium, premiums for small and value stocks, and net of momentum effects. these findings are particularly interesting and worth noting as majority of the existing literature suggests that active funds, on average, underperform the market index especially after incorporating the effects of market and other biases. thus, the findings of this study reverse the general notion and support the selectivity skills of active management. 5.2.1. persistence of performance next, we estimate the persistence effect, that is, whether the performance is a random event or funds are able to repeat their underand over-performances to the subsequent periods. table 3 documents results for the past losers (q1) and winners (q5). panel a reports " estimate of a series of funds that were ranked in the bottom 20% based on their previous table 2 abnormal performance of healthcare funds parameter estimate t value p value panel a: single factor model " 0.2848*** 7.58 0.0000 rmrf 0.7089*** 81.80 0.0000 adj. r2 0.3444 n 14,072 panel b: four-factor model " 0.2473*** 7.58 0.0000 rmrf 0.6947*** 54.92 0.0000 smb 0.2900*** 7.80 0.0000 hml #0.2094*** #10.60 0.0000 mom 0.1198*** 10.38 0.0000 adj. r2 0.4196 n 14,072 note. the above table shows the abnormal performance of the sample funds over the period 1/2000 to 12/2011. the abnormal performance (") is based on the single factor and four-factor models. rit is the excess monthly return of fund i over one month u. s. t-bill return. rmrf is the excess monthly return of the value weighted crsp index over the one month u.s. t-bill return. smb, hml, and mom are monthly returns of size (the difference in returns between small and large cap stocks), book to market (the difference in returns between high and low book-to-market stocks), and momentum (the difference in returns between stocks with high and low past returns) portfolios, respectively. the dependent variable is the individual fund’s monthly excess return over the corresponding one month t-bill rate. " is expressed in percentage per month. results are based on newey-west heteroscedasticity and autocorrelation adjusted standard errors. n is the number of fund month observations. ***, **, and * show the significance at 1%, 5%, and 10% level, respectively. model a: rit # rft $ "i " !1i ! rmrft " &i,t ; model b: rit # rft $ "i " !1i ! rmrft " !2i ! smbt " !3i ! hmlt " !4i ! momt " &i,t. 85a. kaushik et al. / financial services review 23 (2014) 77–91 table 3 persistence of performance parameter estimate t value p value panel a " 0.0215 0.28 0.7824 rmrf 0.8275*** 29.34 0.0000 smb #0.2205*** #4.53 0.0000 hml #0.1841*** #4.17 0.0000 mom 0.0443 1.78 0.0746 adj. r2 0.4207 n 2,375 panel b " 0.0547 0.90 0.3689 rmrf 0.7900*** 39.38 0.0000 smb 0.1569*** 4.18 0.0000 hml #0.3076*** #9.64 0.0000 mom 0.1156*** 5.30 0.0000 adj. r2 0.5482 n 2,628 year positive " negative " insignificant " panel c: short term persistence of performance of q5 funds 2001 25 0 75 2002 0 55.6 44.4 2003 23.8 0 76.2 2004 0 0 100 2005 4.8 0 95.2 2006 0 18.2 81.8 2008 0 0 100 2010 0 0 100 2011 0 0 100 panel d: short term persistence of performance of q1 funds 2001 20 0 80 2002 0 89.4 10.6 2003 0 0 100 2004 0 10 90 2005 9 0 91 2006 0 0 100 2008 0 0 100 2009 0 0 100 2011 0 0 100 note. the above table shows the persistence of performance of the sample funds over the period 1/2000 to 12/2011. every year, at the beginning of year, sample funds are sorted based on previous year’s returns and quintiles are formed. q1 is the group that consists of funds that are ranked as bottom 20% funds based on previous year’s annual returns and q5 is the group that consists of funds that are ranked as top 20% funds based on previous year’s annual returns. this step is repeated every year. q1 is a series of worst performing funds and q5 is a series of best performing funds. a positive " of q5 and a negative " of q1 based on the four-factor carhart model indicate persistence of performance. panel a reposts persistence of performance for q1 and panel b reports the same for q5. panels c and d show short term persistence of performance for q5 and q1 portfolios, respectively. results are based on newey-west heteroscedasticity and autocorrelation adjusted standard errors. n is the number of fund month observations. note. model: rit # rft $ "i " !1i ! rmrft " !2i ! smbt " !3i ! hmlt " !4i ! momt " &i,t. *** and ** show the significance at 1% and 5% level, respectively. 86 a. kaushik et al. / financial services review 23 (2014) 77–91 year’s annual returns whereas panel b reports " value of those funds that were ranked in the top 20% based on their previous year’s annual returns. a negative " for q1 and a positive " for q5 suggest that performance persists. results in panel a show a positive value of 0.0215% value of " (0.258% per year), however, it is statistically insignificant. results in panel b show a positive " value of 0.0547% per month (0.656% per year), but it is also statistically insignificant. the results in table 3 show that performance of both losers and winners is mean reverting and does not persist in the subsequent periods. in other words, a theoretical long position in past winners and a short position in past losers will not deliver a net positive superior " to fund investors. the results are consistent with the findings of saap and tiwari (2004) and berk and green (2004) who documented that non-persistence of performance for well-diversified mutual funds. results also go against the common notion that fund managers of specialized funds such as healthcare funds are “"” managers because they have more insight on the stocks of these sectors and therefore they, especially the winners, should be able to outperform passive benchmarks over the subsequent periods. results of this study show that investors in healthcare funds do not enjoy superior "s consistently. the most important coefficient of interest is "; however, we are also interested to see how ! loadings affect excess returns earned by q1 and q5 portfolios. a closer inspection of results in panels a and b shows very similar pattern for rmrf, hml, and mom factors; however, the coefficient of smb is negative and highly significant for q1 whereas it is positive and statistically highly significant for q5. these diametrically opposite coefficients of smb for q1 and q5 suggest that q1 excess returns are negatively affected by size whereas q5 excess returns increase with small size bias. 5.2.2. short term persistence of performance because persistence of performance was not visible in long term, therefore, we also evaluated whether performance persists in short term or not by using the four-factor model. results in panels c and d (table 3) show that short term persistence is only thinly visible. only 25% of funds that earned positive " in year 2000 (the first year of this study) continued to earn positive " in the next period, year 2001, whereas 75% of those positive " funds returned insignificant "s in the next period, year 2001. the short term persistence results vary from year to year. for example, in year 2002, 55.6% of those funds that earned positive " in year 2001 earned negative " in year 2002 whereas 44.4% of those funds earned insignificant " in year 2002. no fund was able to repeat its positive performance in year 2002. similar results were obtained for those funds that earned lowest "s in the previous period. results of panels c and d of table 3 confirm that healthcare funds were not able to repeat their overor under-performance and prove that performance is a random event. 5.3. cross-sectional analysis numerous studies on mutual funds have shown that performance of actively managed funds is heavily dependent on their size, expense ratio and other fund specific attributes. moreover, it is crucial to analyze the impact of these factors on performance of a portfolio of different funds especially when a portfolio includes funds that are of different sizes, have different fee structures, and the portfolio sample is free of survivorship bias. for example, 87a. kaushik et al. / financial services review 23 (2014) 77–91 carhart (1997) shows that expense ratios and turnover ratios negatively affect a fund’s abnormal performance. in his study of 1,892 well-diversified funds, carhart documented, on average, diminishing abnormal performance by 154 basis points for every 100 basis points increase in a fund’s expense ratio and 95 basis points decrease in a fund’s abnormal performance for every 100 basis points increase in its turnover ratio. in a related study, dahlquist et al. (2000) analyzed swedish mutual funds and found that larger size funds, both equity and bond funds, beat their smaller counterparts. they also showed that funds that charge higher fees perform poorly compared to funds that charge low fee. in a similar note, kaushik and pennathur (2012) found a significant positive relationship between the performance of real-estate sector funds and size. haslem et al. (2008) showed a strong relationship between fund specific attributes and funds’ abnormal performance. similar to this study, they analyzed roughly 1,779 actively managed domestic only equity funds and documented that funds with low expense ratios and larger size perform better than funds with higher expense ratios and smaller sizes. they also found a negative relationship between turnover ratio and abnormal performance. in this study, we use cross-sectional analysis to study the impact of fund specific attributes on the performance of a portfolio of different healthcare mutual funds. specifically, we evaluate the " of a portfolio of healthcare funds as a function of funds’ size, expense ratio, turnover ratio, investment in top 10% assets, cash holdings, and managerial tenure. " $ f : (expense ratio, turnover ratio, size, top, tenure, and cash). cross-sectional results reported in table 4 show that abnormal performance is negatively affected by expense ratio. the coefficient of expense ratio is #2.006 and it is marginally statistically significant. this finding suggests that for every 100 basis points increase in expenses, abnormal performance decreases by 200 basis points. although coefficients of turnover ratio, investment in top 10% holdings, and managerial tenure are all positive, but table 4 cross-sectional analysis variable estimate t value intercept #0.009 (#0.85) expense ratio #2.006 (#1.94) turnover ratio 0.020 (0.32) size #0.000 (#0.63) top 0.006 (0.8) tenure 0.004 (1.11) cash 0.110 (1.63) adj. r2 0.381*** n 5,253 note. the above table shows cross-sectional analysis over the period 1/2000 to 12/2011. the dependent variable is monthly " that is estimated by using "it $ rit # rft # !̃1it#1 ! rmrft " !̃2it#1 ! smbt " !̃3it#1 ! hmlt " !̃4it#1 ! momt where !̃1it#1 ! . . . !̃4it#1 are the beta loading estimated using equation rit # rft $ "i " !1i ! rmrft " !2i ! smbt " !3i ! hmlt " !4i ! momt " &i,t . monthly expense ratio, turnover ratio, size, a fund’s investment in its top 10% assets, managerial tenure, and cash holdings are the explanatory variables. model: "it $ !0 " !1 expense ratioit " !2 turnoverit " !3 sizeit " !4 topit " !5 tenureit " !6 cashit " &it. *** and ** show the significance at 1% and 5% level, respectively. 88 a. kaushik et al. / financial services review 23 (2014) 77–91 they are statistically not significant. our results are similar to those found in the existing literature for both well-diversified and other specialized funds (e.g., carhart, 1997; haslem et al., 2008 among others). 6. conclusion investments in mutual funds have been growing consistently for the past five decades. various studies have examined the performance phenomenon of actively managed mutual funds. though results pertaining to the abnormal performance have been mixed, but a majority of studies indicated underperformance of these funds especially after incorporating funds fees and expenses. in this study, we examined the abnormal performance of 115 healthcare mutual funds that existed at some point in time over the period 2000–2012. our motivation stems from the fact that healthcare funds have grown at a rapid rate over the last 10 years and yet not many studies are available to explain their abnormal performance. moreover, this sector of the economy has been a center piece of attention for both the politicians and policy makers. most of the last decade has witnessed global economic slowdown and many analysts may argue that a defensive stock like healthcare should perform better under these conditions. our results indicate that healthcare funds outperform passive market index by as much as 0.2847% per month (3.42% per year) when the single-factor model is used. the abnormal performance outshines passive market index by 0.2473% per month (2.97% per year) when the four-factor model is used. in other words, after controlling for known biases such as the market risk premium, growth and size premiums, and momentum effects, healthcare funds outperform the market by roughly 3% per year. results of this study suggest that retail investors can add value to their overall portfolio by including a portion of their investment in healthcare funds. this will not only give them necessary diversification, but also improves their chances of earning higher return on invested capital. the other point of enquiry was to evaluate whether this performance persists over subsequent periods? we follow carhart (1997) and berk and green (2004) methodologies and created quintiles based on previous year’s annual returns and formed “winner” and “loser” portfolios that were rebalanced every year. we use portfolio based approach to evaluate persistence effect for both short and long terms. the results of this study show that both winner and loser portfolios are unable to repeat their over and under performances. results show that performance is mean reverting for both winner and loser portfolios. notes 1 grinblatt and titman (1989, 1994), wermers (1997) findings show that mangers’ are able to beat corresponding benchmarks before any expenses are considered. 2 analyzing the performance of equity funds, wermers (2000) finds that funds outperform corresponding benchmark by 1.3%; however, the same funds on net return basis underperform corresponding benchmark by 1%. 89a. kaushik et al. / financial services review 23 (2014) 77–91 3 source: morningstar direct database. 4 source: investment company institute factbook 2012. 5 why might an investor consider the health care sector? fidelity investments. 6 “the truly staggering cost of inventing new drugs,” forbes, february 10, 2012. 7 “drugs firms face billions in losses in ’11 as patents end,” the new york times, march 6, 2011. 8 many articles (dellva, demaskey, and smith, 2001; khorana and nelling, 1997); the investment company institute among others define sector funds as funds that seek capital appreciation through their own specialized (narrow) objectives. references berk, j. b., & green, r. c. (2004). mutual fund flows and performance in rational markets. journal of political economy, 112, 1269–1295. brooks, l. d., & porter, g. e. (2012). mutual fund performance attribution: 1994–2005. financial services review, 21, 259–273. brown, k. c., harlow, w. v., & starks, l. t. (1996). of tournaments and temptations: an analysis of managerial incentives in the mutual fund industry. journal of finance, 51, 85–110. burlacu, r., & fontaine, p. (2003). sector funds in europe: performance, selectivity and market timing. markets and bank, 65, 1–27. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. chen, j., hong, h., huang, m., & kubik, j. d. (2004). does fund size erode mutual fund performance? the role of liquidity and organization. american economic review, 94, 1276–1302. dahlquist, m., engström, s., & söderlind, p. (2000). performance and characteristics of swedish mutual funds. journal of financial and quantitative analysis, 35, 409–423. dellva, l. w., demaskey, a. l., & smith, c. a. (2001). selectivity and market timing performance of fidelity sector mutual funds. the financial review, 36, 39–54. fama, e. f., & french, k. r. (1993). common risk factors in the return on bonds and stocks. journal of financial economics, 33, 3–53. fama, f. e., & macbeth, j. (1973). risk, return, and equilibrium: empirical tests. journal of political economy, 81, 607–636. fan, y., & addams, h. l. (2012). united states based international mutual funds: performance and persistence. financial services review, 21, 51–61. grinblatt, m., & titman, s. (1989). mutual fund performance: an analysis of quarterly portfolio holdings. journal of business, 62, 393–416. grinblatt m., & titman, s. (1994). a study of monthly fund returns and performance evaluation techniques. the journal of financial and quantitative analysis, 29, 419–442. haslem, j. a., baker, k. h., & smith, d. m. (2008). performance and characteristics of actively managed mutual funds with diverse expense ratios. financial services review, 17, 49–68. jensen, m. (1968). the performance of mutual funds in the period 1945–1964. journal of finance, 23, 389–416. kacperczyk, m., sialm, c., & zheng, l. (2005). on the industry concentration of actively managed equity mutual funds. journal of finance, 60, 1983–2012. kaushik, a., & pennathur, a. (2012). an empirical examination of the performance of real estate mutual funds 1990–2008. financial services review, 21, 343–358. kaushik, a., barnhart, s. w., & pennathur, a. (2010). market timing and the determinants of performance of sector funds over the business cycle. managerial finance, 36, 583–602. khorana, a., & nelling, e. (1997). the performance, risk and diversification of sector funds. financial analysts journal, may-june, 62–74. 90 a. kaushik et al. / financial services review 23 (2014) 77–91 lafever, r., & canizo, l. (2005). revisiting performance persistence in real estate funds (pp. 1–47). master’s degree thesis, massachusetts institute of technology. larry, f. d., & weigand, r. a. (1998). explaining persistence in mutual fund performance. financial services review, 7, 45–55. lin, c. y., & yung, k. (2004). real estate mutual funds: performance and persistence. journal of real estate research, 26, 69–93. manakyan, h., & liano, k. (1997). performance of mutual funds before and after closing to new investors. financial services review, 6, 257–269. nan, l., & yan, c. l. (2011). understanding emerging market equity mutual funds: the case of china. financial services review, 20, 1–19. philpot, j. (2000). performance persistence and management skill in nonconventional bond mutual funds. financial services review, 9, 247–258. sapp, t., & tiwari, a. (2004). does stock return momentum explain the “smart money” effect? journal of finance, 59, 2605–2622. sharpe, w. f. (1964). capital asset prices: a theory of market equilibrium under conditions of risk.” journal of finance, 19, 425–442. tiwari, a., & vijh, a. m. (2004). sector fund performance: analysis of cash flow volatility and returns. working paper. wermers, r. (1997). momentum investment strategies of mutual funds, performance persistence, and survivorship bias. working paper, university of colorado. wermers, r. (2000). mutual fund performance: an empirical decomposition into stock-picking talent, style, transactions costs, and expenses. journal of finance, 55, 1655–1695. wermers, r. (2003). is money really “smart”? new evidence on the relation between mutual fund flows, manager behavior, and performance persistence. working paper, university of maryland. 91a. kaushik et al. / financial services review 23 (2014) 77–91 untitled cognitive ability impact on life insurance lapsation barry s. mulhollanda,*, michael s. finkeb adepartment of finance, university of akron, 259 south broadway street, akron, oh 44325-4803, usa bdepartment of wealth management, the american college of financial planning, 630 allendale road, king of prussia, pa 19406, usa abstract life insurance is an important household risk management and financial tool. policy lapsation has economic effects on life insurance companies, policyholders, and beneficiaries that may be detrimental when these lapses are unexpected. prior literature examined several hypotheses of life insurance lapse focusing mainly on macroeconomic factors using aggregate data and household microeconomic factors using household-level data. we introduce and test individual cognitive ability variables in a model of the life insurance voluntary lapse decision by individual policyholders using household-level data from the health and retirement study. we find that one measure of cognitive ability, in particular, numeracy, is related to the voluntary lapse decision. while controlling for numeracy, we find evidence that those individuals with higher levels of net worth are less likely to voluntarily lapse a policy which is consistent with the emergency fund hypothesis. we introduce a new measure of liquidity shock, kids moving home, into the model and find it has a strong positive relationship with the decision to voluntarily lapse a policy. consistent with life insurance demand theory, we find that those who have recently entered retirement are more likely to lapse their policy. © 2023 academy of financial services. all rights reserved. keywords: health and retirement study; life insurance; cognitive ability; lapsation 1. introduction permanent life insurance is an important household risk management and financial tool that generally requires an ongoing premium payment to remain in force. life insurance acts as a hedge against the uncertainty of the labor income flows of household members *corresponding author: tel.: +1-330-972-4524, fax: +1-330-972-5970. e-mail address: bmulholland@uakron.edu 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 73–96 (campbell 1980; fischer, 1973; yaari, 1965), it helps beneficiaries meet their need for smoothing consumption over the life cycle (lewis, 1989), and it helps policyholders meet their bequest motives with a potentially guaranteed source of funds (bernheim, 1991). it can also be used as a tax-advantaged savings tool (rankin, 1987) and as a tool to reduce estate tax erosion (milevsky, 2006; mulholland et al., 2016). life insurance policies are designed to build cash value early in the life of the policy that pays for the increasing cost of providing a death benefit as the insured ages. the failure to make premium payments results in the termination of the policy, also known as lapsation. each year, about 7% of all individual life insurance policies lapse (fang & kung, 2012). lapsation can be costly for consumers because accumulated cash value is a tax-deferred asset that can be borrowed against or efficiently transferred at death, and the secondary market value of the policy can be far greater than the cash value among those whose health is worse than average (daily et al., 2008). consumer lapsation of older policies results in socalled lapsation profits that insurance companies build into the pricing of life insurance products (fang & kung, 2012). gottlieb and smetters (2021) find that policyholders who lapse a policy before death or policy maturity subsidize policyholders who hold their policies until death or policy maturity. gottlieb and smetters (2021) note that consumers are subject to two types of risk: mortality risk that life insurance is intended to mitigate, and other non-mortality background risks such as unemployment, medical expense shocks, or unexpected needs of dependents, that result in a need for liquidity. while consumers can more easily estimate mortality risk when they purchase a policy, they may underweight the risk of experiencing background shocks that affect their ability to make regular premium payments. lapsation often occurs because an insured experiences an income shock; however, changes in income affect lapsation for younger households but not for older households (fier & liebenberg, 2013). another potential background risk is a reduced ability to recognize the financial consequences of failing to make required premium payments. gottlieb and smetters (2021) find evidence in a national survey of insurance purchasers that 37.8% of policy lapses result from consumers forgetting to pay their premiums, and 15.4% are due to unexpected liquidity shocks. banks and oldfield (2007) find that a large portion of older adults are unable to perform simple interest calculations with accuracy. gerardi et al. (2013) find that borrowers with lower levels of numerical ability are more likely to default on their mortgages—another behavior that requires a consistent payment to avoid a contractual default and loss of wealth. christelis et al. (2010) find a positive relationship between the cognitive ability of older individuals as measured by numeracy, verbal fluency, and recall (memory), and complex investments such as stocks, while korniotis and kumar (2011) find that stock investors in their 70s and 80s significantly underperform younger investors. financial literacy related to insurance declines consistently in old age despite higher rates of life insurance ownership among older cohorts (finke et al., 2017). policy lapsation decisions can have both positive and negative effects on the household. intended lapsation of policies that are no longer needed to meet the basic life cycle purposes of life insurance—labor income protection (campbell 1980; fischer, 1973; yaari, 1965), beneficiary needs (lewis, 1989), and bequest motives (bernheim, 1991)—may improve household utility after being lapsed. lapsation decisions can positively affect the liquidity of 74 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 the household through the elimination of the premium payments that release those funds for use in other areas of consumption, such as improving lifetime consumption for current household members. the subsequent removal of a sophisticated financial tool from the household portfolio may make management of the portfolio less burdensome. on the other hand, unintended lapsation may be problematic for policyholders and their beneficiaries. for married couples, the loss of the policy may remove guaranteed funds intended to provide for the final expenses associated with death, such as medical out-ofpocket expenses, health insurance copays, and funeral expenses. it may also reduce the assets available to fund the surviving spouse’s lifestyle since funds from the lapsation will provide fewer survivor proceeds than the death benefit would have paid at the death. in addition, if the lapsed policy has a cash surrender value, that cash is brought into the estate of the policyholder which then makes it subject to ongoing income taxation and future probate distribution, both of which are avoided in an active life insurance policy. in this study, we explore whether the observed decline in cognitive ability in later life is related to a potential wealth loss from insurance lapsation. this article adds to the literature by providing insight into unintentional lapsation among older consumers that represents a potential source of welfare loss. awareness of the risk of cognitive aging on life insurance lapsation represents an opportunity for financial planners to avoid this risk by automating payment, considering a life settlement for seniors who do not value the bequest, or fully paying up the policy to avoid the risk of late life lapsation. 2. objective from the life insurance demand theory, we use several key theories as guides for our work. yaari (1965), fischer (1973), and campbell (1980) collectively indicate that life insurance is a hedge against the uncertainty of labor income flows. lewis (1989) suggests that the beneficiaries’ desire to smooth their expected lifetime consumption will influence life insurance demand, and bernheim (1991) suggests that policyholder bequest motives influence the demand for life insurance. several hypotheses have been suggested for the reason policyholders lapse their life insurance policies: the emergency fund hypothesis (efh), the interest rate hypothesis (irh), and the policy replacement hypothesis (prh). linton (1932) proposed the efh, suggesting households regard the cash value in their life insurance policies as a source of emergency funds that can be accessed in times of financial need. schott (1971) describes the foundations for the irh which suggests that when interest rate arbitrage potential exists between high market rates and low policy interest rates, policyholders may be encouraged to take policy loans or lapse the policies to make the funds available to earn the higher rates. russell et al. (2013), building upon empirical research by outreville (1990), articulate the prh which suggests that policyholders will surrender their policies to replace them with a policy with better pricing or more favorable terms. cole and fier (2021) find that policyholders tend to surrender their policies for the cash values when confronted with longer-term financial issues and tend to take policy loans when b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 75 dealing with shorter-term financial crises. their findings support the emergency fund, alternative funds (interest rate), and the policy replacement hypotheses. nolte and schneider (2017) find mixed results of policyholders surrendering policies before the optimal time to surrender the policy under the efh, which is when there are no other liquidity options available. their analysis suggests behavioral issues have a meaningful impact on surrender decisions, with more financially literate policyholders being less likely to make suboptimal policy surrender decisions. our analysis will be limited to the use of the efh for several reasons. first, the data we use is limited in the cross-section of the policyholders who indicate they lapse a policy with the intention of replacing it with another policy, a practice fang and kung (2012) refer to as insurance optimization. fier and liebenberg (2013) use longitudinal hrs data from the 1996 through 2008 waves in their analysis of the prh. second, liebenberg et al. (2010) suggest that different types of data allow for better analysis when exploring the irh and the efh. they suggest that the irh is best explored with aggregate-level data while the efh is best examined with household-level data. the data used in our analysis is household-level in scope, thus we limit our analysis to the efh. cognitive ability and its impact on financial decision-making have been an important area of current research (banks & oldfield, 2007; christelis et al., 2010; lusardi & tufano, 2009; mcardle et al., 2011; smith et al., 2010). consideration has been given to whether fluid intelligence or crystallized intelligence is the most important aspect of cognition. as defined by smith et al. (2010), fluid intelligence is deliberate processing or the ability to think about a problem in a clear and quick manner, while crystallized intelligence is the accumulation of relevant knowledge about various problems through education and lifetime experience. smith et al. (2010) create a shorthand division of the two components of intelligence by defining fluid intelligence as the thinking part involving memory, abstract reasoning, and executive function or decision-making, and defining crystallized intelligence as the knowing part consisting of education and lifetime experience. several measures exist within these areas of intelligence and have been used by researchers as they explore the impact of cognitive ability on financial decisions. smith et al. (2010) indicate that the three measures are: (1) episodic memory, which is used as a general measure of an important aspect of fluid intelligence because memory access is important to any cognitive ability; (2) numeracy, which is used to measure the actual ability to perform numerical skills learned in schools and is used to measure crystallized intelligence; and (3) mental status scores, which measure elements of both fluid and crystallized intelligence, non-specific cognitive skills needed for everything. both numeracy and episodic memory are related to household total wealth and financial wealth holdings. we model the impact that a policyholder’s cognitive ability, whether measured by episodic memory or numeracy, has on the probability of voluntarily lapse of a life insurance policy when controlling for other factors that have been shown to predict lapsation. 76 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 3. data and methods 3.1. data we use data from the health and retirement study (hrs) from the 2008 and 2010 waves (waves 9 and 10) for our analysis. the hrs is a longitudinal survey capturing the health and economic circumstances of a nationally representative sample of u.s. citizens over the age of 50. conducted every two years by the institute for social research at the university of michigan since its first wave in 1992, the original cohort has been interviewed in each wave. with attrition over the time frame of the survey, additional cohorts have been added in the ensuing years, bringing the complete sample size to nearly 37,000 individuals who participated in the survey to some extent throughout the life span of the survey. the hrs is supported by the national institute on aging (nia) and the social security administration (ssa). liebenberg et al. (2012) indicate that household-level panel data are well suited for showing the impact of events at the household-level on the financial decisions made by these households. we use the hrs survey because it has extensive information on life insurance ownership choices made by household members as well as respondent and household information on health, income, wealth, and family structure. our initial interest in modeling lapse behavior in the presence of cognitive ability measures is on the identified respondents for which information is included in both the 2008 and 2010 waves; our sample is first reduced to 17,217 respondents. because the hrs follows participants until they are deceased, we further limit our sample to those who were alive in 2010 and who indicated their life insurance ownership status in that wave. this further reduces our sample to 14,659 respondents. our dependent variable is those respondents who self-lapse a life insurance policy during the prior two years. consistent with prior research using the hrs (fang & kung, 2012; fier & liebenberg, 2013), we use two questions in combination from the hrs to develop our dependent variable. the first question allows us to identify respondents who allowed a policy to lapse in the prior two years.1 we use the response to a follow-up question to specifically identify self-lapsers.2 combining those who indicated a policy lapse in the first question with an affirmative response that the lapse was their choice creates our dependent variable. because it is not possible to lapse a life insurance policy unless you first own one, we next limit our sample to those who indicated they owned life insurance in 2008. this criterion reduced our sample size to 9,359 individuals. finally, we narrow our sample to those who answered the questions related to our main independent variable of interest, cognitive ability. fisher et al. (2013) indicate that the design of the hrs study creates some methodological issues for measuring cognitive functioning which requires appropriate adaption of the standard tests. in particular, the hrs does allow for proxy respondents for individuals who may have reached a point of incapacity that interferes with their ability to respond, so we limit our sample to self-respondents. our analysis will explore both episodic memory and numeracy to determine if either predicts lapse behavior and if one measure is a more powerful predictor of lapse behavior than the other. b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 77 by limiting our sample to those who performed both the episodic memory (total word recall) test components and answered the series of three numeracy questions, our sample size was reduced to 8,795 respondents. the resulting sample size following the application of each step of our selection criteria can be seen in table 1. 3.2. model our dependent variable is the choice to voluntarily lapse a life insurance policy in the prior two years. our independent variable of interest is cognitive ability measured using numeracy and cognitive ability. mcardle et al. (2011) find that within person correlations are moderate for men and women between episodic memory and numeracy. the decision to lapse a life insurance policy is at its roots a decision about retention of a life insurance policy that is already owned by the policyholder. demographic and economic predictors for life insurance ownership include age, bequest motive, education, employment, children, marital status, homeownership, income, net worth, and race (zietz, 2003). gender is also shown to be an important factor in determining demand for life insurance (gandolfi & miners, 1996), with men more likely to be insured. eling and kiesenbauer (2014) find that women are less likely to lapse life insurance policies than men. we control for these differences by including gender in our model. mulholland et al. (2016) find that the financial sophistication of the policyholder plays an important part in the decision to own life insurance. therefore, our model includes a proxy for financial knowledge. the stated bequest motive is the ideal variable to identify specific desires of the individual to leave a bequest. in the absence of a bequest motive variable in the hrs, bernheim (1991) uses marital status and the existence of children of the insured as proxies for the bequest motive. we do the same in our model. an important factor in life insurance ownership is homeownership, or more importantly the mortgage debt that is associated with homeownership. prior research suggests a positive relationship between the amount of total household debt and the amount of life insurance it holds (frees & sun, 2010; lin & grace, 2007). yet, fier and liebenberg (2013) find increased debt is positively related to the decision to lapse a policy. we include total debt as a control variable, allowing it to proxy for homeownership. table 1 sample selection criteria selection criteria sample size all individuals tracked in the hrs from 1992 to 2010 . . . 36,986 . . . respondents included in the 2008 and 2010 waves 17,217 . . . those who were alive in 2010 and indicated their life insurance ownership status 14,659 . . . those who owned life insurance in 2008 9,359 . . . those who answered the total word recall (twr) and numeracy questions in 2010 8,795 note. the selection criteria are cumulative. hrs = health retirement study. 78 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 fang and kung (2012) indicate that the health condition of the insured is an important factor due to reclassification risk from the increasing cost of life insurance on the spot market due to declines in overall health and the resulting increase in mortality risk of the insured. as the insured’s health declines, they will find it increasingly expensive to replace a policy they currently own. accordingly, we include the health condition of the respondent in our model. life insurance is used as both a tax-sheltered form of savings (brown & poterba, 2006) and for non-human-capital-replacement issues like accumulating cash to pay estate taxes for those households vulnerable to estate taxes (milevsky, 2006). mulholland et al. (2016) find that households vulnerable to u.s. estate taxes are more likely to own cash value life insurance. for these reasons, we control for estate tax vulnerability under the 2010 estate tax laws. prior life insurance loan and lapse research suggest loans and lapses are driven by various shocks affecting the household. liebenberg et al. (2010) find evidence that policy loans increase as a result of household income shocks while kuo et al. (2003) find evidence that the unemployment rate is positively related to the lapse rate. fang and kung (2012) find some evidence that health shocks positively impact policy lapsation. bernheim (1991) finds that bequest shocks reduce the demand for life insurance. liebenberg et al. (2012) find a positive relationship between those who recently retired and those who lapse life insurance policies. we control for health shocks, bequest shocks, and those respondents who have retired in the past two years. bernheim (1991) indicates that households with bequest motives, often as intergenerational transfers, hold life insurance to enhance their bequest utility. but just as income shocks and unemployment with the resulting loss of income have been found to increase the likelihood of lapse, we question if unexpected increases in household expenses cause similar behaviors. in particular, does the sudden appearance of children as members of the household put a drain on household expenses that may cause similar results as an income shock? therefore, we include the increase of children living in the home since the prior wave as a control variable in our model. we use the following logistic regression model to model the life insurance lapse decision: ln pi 1� pi � � ¼ b 0 þ b 1�3ðcognitive ability quartiles > 25%þ þ b 4�6ðincome quartiles > 25%þ þ b 7�9ðnet worth quartiles > 25%þ þ b 10ðfinancial knowledgeþ þ b 11ðmarriedþ þ b 12ðchildrenþ þ b 13�17ðage groups < 60 or > 64þ þ b 18ðgenderþ þ b 19ðlog of total debtþ þ b 20�21ðnon-white race groupsþ þ b 22�24ðeducation > no� hs degreeþ þ b 25�28ðhealth problems > 0þ þ b 29ðestate tax vulnerableþ þ b 30ðnewly retired since 2008þ þ b 31ðincome shockþ þ b 32ðadditional kid at home shockþ þ b 33ðmarriage shockþ þ b 34ðhealth shockþ where pi is the probability of the individual lapsing a life insurance policy. b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 79 3.3. variables in our model, the dependent variable, lapse, is a binary variable set to 1.0 when the respondent has answered affirmatively to both hrs questions about life insurance lapse: first that they did lapse a policy in the two years before responding to the survey in 2010, and second that they chose to proceed with the lapse or cancellation. for all other combined responses, the variable is set to zero. we measure episodic memory using two measures of word recall found in the hrs, immediate word recall and delayed word recall (ofstedal et al., 2005). to test word recall, the respondent is read a list of 10 nouns. the first time they receive the word list, it is drawn randomly from four sets of words, of which no words overlap. in each subsequent wave, the respondent is presented with a different word list from the four lists, resulting in the respondent only receiving the same word list every fourth wave. respondents and their spouses are given different word lists in each wave. for immediate word recall, the respondents are read the list of nouns and then asked to immediately recall as many of the words as they can. the respondent receives a score of the number correctly repeated. after testing for immediate word recall and after approximately five minutes of asking other survey questions, the delayed word recall test is administered. again, the respondent is asked to recall as many words as they can from the list. the respondent again receives a score for the number of correctly recalled words. while smith et al. (2010) create a combined word recall measure for each respondent by averaging her immediate and delayed word recall scores, we follow the common practice of adding the two scores to create a total word recall (twr) score for each respondent, leading to a score range of 0 to 20 (see browning, 2014). following smith et al. (2010), we measure numeracy by using three questions that were first included in the 2002 core survey questions and repeated every second wave since. the first question asks the respondent to calculate the number of people given a known percentage while the second question asks respondents to perform a division problem.3,4 if the respondent gets either of the first two questions correct, they are then asked the third question where they are asked to solve a two-year compounding interest problem.5 they receive one point for each correct answer. we create a numeracy score by adding these results for each respondent, developing a score range of 0 to 3. to create comparable measures for episodic memory and numeracy, we quartile each measure. we separate income and net worth into quartiles. following mulholland et al. (2016), we create a dummy variable to identify those households who may be vulnerable to estate taxes in 2010 because they have net worth, including a second home, of greater than $5 million, the then-current maximum estate tax exemption. following prior research (brown & poterba, 2006; fier & liebenberg, 2013), we separate age into bands to capture the non-linear predicted effect of the life cycle stage and life insurance demand. to proxy for unknown bequest factors, we create dummy variables for the respondent being married and having any living children, setting the affirmative to 1.0. we also create a dummy variable for gender, setting it to 1.0 for males. 80 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 we create a total household-level debt variable by first adding the total mortgage debt to the total other household debt. following fier and liebenberg (2013), we control for household-level debt by using the natural logarithm of the total household-level debt. the categories for race, education, and health problems are separated into dichotomous variables. the categories of race are separated into the binary variables white, black, and other race. education is separated into four variables consisting of less than high school, high school, some college, and college degree. eight separate health problems are identified in the hrs, including high blood pressure, diabetes, cancer, lung disease, heart disease, stroke, psychological problems, and arthritis. respondents are asked in each wave if they have been diagnosed with any of these health problems. similar to fang and kung (2012), we total the number of health problems the respondents indicate, creating a health problem score of 0 to 8. we then create dummy variables for each of 0 through 3 problems and create a dummy variable capturing 4 or more health problems. the five shocks or major changes we control for include newly retired, income shock, additional kids at home shock, marriage shock, and health shock, creating dichotomous variables for each. just as fier and liebenberg (2013) do, we set newly retired equal to 1.0 if the respondent identifies herself in 2008 as not retired and then identified herself as retired in 2010. fier and liebenberg explore the efh over the entire range of income decline and find that lapse is related to income shock for those households with the most extreme levels of income decline, roughly the worst 24% of their sample. since our focus is to explore the impact of cognitive ability and not specifically find additional validation for the efh, we use a more conservative 10% income decline in real dollars from the prior wave to indicate an income shock. we set the dichotomous variable measuring additional kids in the home to 1.0 when there is an increase of one or more kids at home since the prior wave. the marriage shock variable is set to 1 if the marital status has changed to unmarried, whether because of divorce or the death of the spouse, since the 2008 wave. finally, since the number of health conditions usually increases over time (fang & kung, 2012), an increase of one health condition from the prior wave may not capture a true health shock. therefore, we specify health shock to be an increase of two health problems since the prior wave. all variables are listed with the summary statistics in table 2. 4. results 4.1. descriptive analysis approximately 4% of our sample lapsed their policies in the two years prior to 2010. this appears to be in line with the typical 5 to 5.5% annual individual policy lapse rates as reported by acli in their annual industry report (acli, 2012). non-voluntary policy lapsation is occurring as evidenced by the additional 10% decline in life insurance ownership indicated by our sample. b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 81 table 2 variable summary statistics wave 10–2010 n = 8,795 384 respondents reported lapse mean sd insurance ownership variables % of sample own life insurance in 2010 0.86 0.35 life insurance status lapse a policy due to their choice in prior two years 0.04 0.20 explanatory variables cognitive ability episodic memory (total word recall) (no. correct) 9.75 3.48 memory quartiles (no. correct) < 25th percentile 5.99 1.91 25th to 50th percentile 9.50 0.50 50th to 75th percentile 11.48 0.50 > 75th percentile 14.50 1.60 numeracy (no. correct) 1.24 0.89 numeracy quartiles (# correct) < 25th percentile 0.00 0.00 25th to 50th percentile 1.00 0.00 50th to 75th percentile 2.00 0.00 >75th percentile 3.00 0.00 economic factors income in 2010 $68,057 $84,585 income quartiles ($) <25th percentile $14,666 $5,961 25th to 50th percentile $33,672 $5,804 50th to 75th percentile $60,287 $10,703 >75th percentile $163,600 $123,344 net worth in 2010 $480,293 $1,002,237 net worth quartiles ($) <25th percentile $11,517 $48,998 25th to 50th percentile $128,233 $41,899 50th to 75th percentile $340,062 $91,184 >75th percentile $1,442,672 $1,649,010 log total debt 5.27 5.28 financial knowledge financial respondent 0.70 0.46 bequest factors married in current wave 0.63 0.48 has living children 0.93 0.26 demographic factors age in 2010 69.56 9.76 less than 60 0.17 0.37 60 to 64 0.17 0.38 65 to 69 0.15 0.36 70 to 74 0.20 0.40 75 to 79 0.15 0.36 80 and higher 0.16 0.37 (continued on next page) 82 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 the sample averaged slightly fewer than 10 total words recalled and approximately 1.24 numeracy questions answered correctly. both numbers are very similar to results from the analyses by mcardle et al. (2011) and smith et al. (2010). means and standard deviations for both income and wealth are reported for the entire sample as well as the quartiles for both variables. we observe apparent non-linearity of the increase in means across quartiles in both measures. approximately 70% of our sample appears to have financial knowledge since they are identified as the financial respondent for the households they represent. over 60% of the respondents are married while over 90% have living children. demographically, our sample is on average nearly 70 years old, contains a majority of females, racially is predominantly white, has nearly half of its respondents with education beyond high school, and indicates that that a little over half have fewer than three of the eight identified health problems. while our statistical summary suggests a positive level of total household debt for our sample, in an unreported analysis, we find that 48.3% of the sample had no household debt. table 2 (continued) wave 10–2010 n = 8,795 384 respondents reported lapse mean sd insurance ownership variables male 0.44 0.50 race white 0.82 0.39 black 0.16 0.37 other 0.02 0.15 education less than high school degree 0.15 0.36 high school 0.37 0.48 some college 0.24 0.43 college 0.24 0.43 health problems zero problems 0.10 0.30 one problem 0.21 0.41 two problems 0.26 0.44 three problems 0.23 0.42 four or more problems 0.20 0.40 estate tax vulnerable net worth over current exemptiona 0.01 0.09 shocks or major changes newly retired since 2008 0.12 0.32 income shock (decline) of > 10% compared with 2008 0.40 0.49 additional kids residing at home since 2008 0.05 0.22 marriage ended since last wave 0.04 0.20 health shock (2+ additional health issues) since last wave 0.03 0.17 note. a2010 exemption was $0 but was retroactively changed to allow an election of $5,000,000. this variable set at $5,000,000. b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 83 in our final control variable, we see that only about one percent of the sample was vulnerable to u.s. estate taxes due to household net worth in excess of $5 million, the 2010 estate tax exemption amount. we note that the u.s. estate taxes were in legislative flux that year due to the elimination in 2010 of estate taxes as a result of the economic growth and tax relief reconciliation act of 2001 (egtrra-2001). the taxpayer relief act of 2010 (tra-2010) reinstated the estate, gift, and generation-skipping transfer taxes with different exclusion amounts and top tax rates for taxable estates. tra-2010 also presented a choice to the estates of those who died in 2010 on how they wanted to be taxed. they could either avoid any estate taxes while accepting the loss of the stepped-up basis for assets; thus, making the assets susceptible to capital gains taxes, or accept the new transfer tax laws implemented by the tra-2010 that allowed estates to retain the stepped-up basis along with a $5 million estate tax exemption. looking finally to the shock and major change variables we include in our analysis, we see that about 12% of the sample retired between their survey interviews in 2008 and 2010. approximately 40% of the sample experienced real income loss in excess of 10% over that period. additionally, five percent of the sample indicated they had children join them in the home, four percent of their marriages end through divorce or the death of their spouse, and three percent experienced health shocks. 4.2. univariate analysis it is important to know how individual and household factors impact the respondent’s decision to voluntarily lapse a life insurance policy. we first conduct a univariate comparison of the independent variables between those households choosing to lapse a policy and those that did not lapse a policy in 2010. our results are reported in table 3. both measures of cognitive ability appear to play a part in the lapse decision, though the level of cognitive ability appears to be important as to which decision is made. both measures indicate a similar pattern across the levels of cognitive ability. a greater share of those within the lowest quartile in each cognitive measure do not lapse their policies while a larger proportion of those in the highest quartile lapse a policy. more quartiles of the numeracy measure showed a univariate relationship between cognitive ability and the decision lapse to lapse a policy. the household economic factors indicate some level of univariate relationship. those respondents in households with the lowest income indicate a higher proportion that do not lapse a policy while those with the highest income have a higher proportion that lapse. respondents in households falling in the lowest and highest quartiles of net worth have greater percentages that lapse policies while those in the inner quartiles have larger percentages that do not lapse. those who lapsed policies also had higher debt than those who did not. from a demographic perspective, there is a positive univariate relationship between lapsing a policy and being male, or being black, or having a college degree. from a life cycle perspective, we see a relationship among respondents who are age 75 and older and not lapsing a policy. of particular interest is the highly significant difference among respondents in the age 60 to age 64 group, suggesting a univariate relationship between this age group that is often associated with the beginning of retirement (brown, 2013) and the decision to lapse. 84 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 table 3 univariate difference for lapses of life insurance between 2008 and 2010 differences in means of demand determinants for individuals lapsing life insurance explanatory variables (1) lapse = 0 (2) lapse = 1 diff (1) 2 (2) no. of individuals 8411 384 cognitive ability episodic memory quartiles (no. correct) <25th percentile 0.3481 0.2943 0.0538** 25th to 50th percentile 0.2337 0.2318 0.0020 50th to 75th percentile 0.2133 0.2318 �0.0185 >75th percentile 0.2049 0.2422 �0.0373* numeracy quartiles (no. correct) <25th percentile 0.2318 0.1536 0.0782*** 25th to 50th percentile 0.3770 0.3854 �0.0084 50th to 75th percentile 0.3189 0.3594 �0.0405* >75th percentile 0.0723 0.1016 �0.0293** economic factors income quartiles ($) <25th percentile 0.2518 0.2109 0.0409* 25th to 50th percentile 0.2499 0.2526 �0.0027 50th to 75th percentile 0.2500 0.2474 0.0026 >75th percentile 0.2482 0.2891 �0.0408* net worth quartiles ($) <25th percentile 0.2473 0.3281 �0.0808*** 25th to 50th percentile 0.2519 0.1901 0.0618*** 50th to 75th percentile 0.2526 0.1953 0.0573** >75th percentile 0.2481 0.2865 �0.0383* log total debt 5.2279 6.2421 �1.0142*** financial knowledge financial respondent 0.7029 0.7057 �0.0028 bequest factors married in current wave 0.6270 0.6432 �0.0162 has living children 0.9289 0.9271 0.0018 demographic factors age in 2010 less than 60 0.1682 0.1797 �0.0115 60 to 64 0.1686 0.2240 �0.0554*** 65 to 69 0.1549 0.1484 0.0065 70 to 74 0.1952 0.2109 �0.0157 75 to 79 0.1504 0.1198 0.0306* 80 and higher 0.1626 0.1172 0.0455** male 0.4336 0.4818 �0.0482* race white 0.8189 0.7917 0.0273 black 0.1571 0.1901 �0.0330* other 0.0240 0.0182 0.0058 education less than high school degree 0.1531 0.1224 0.0307* high school 0.3667 0.3307 0.0359 some college 0.2385 0.2318 0.0067 college 0.2416 0.3151 �0.0735*** (continued on next page) b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 85 we also observe univariate differences in two of our shock variables, those who are newly retired and those with more kids living with them. this suggests that both of these shocks are related to the decision to voluntarily lapse a policy. we now turn our attention to the multivariate analysis using logistic regression to better understand the life insurance voluntary lapse decision by older individuals. 4.3. logistic regression model analysis we present in two tables our logistic regression analyses of the decision to voluntarily lapse a life insurance policy, using separately the two measures of cognitive ability. the regression incorporating episodic memory is presented in table 4a and the regression incorporating numeracy is presented in table 4b. in each table we present the results of three regression analyses from the same model to explore the relationships that various independent variables have on the decision to voluntarily lapse a policy. the first two regressions are shortened forms of the model. regression 1 in each table is the simplest form of each model where we examine the relationship of only our independent variable of interest, cognitive ability, with the decision to voluntarily lapse a policy. in regression 2, we add variables representing household economic factors and bequest motives from life insurance theory, and financial knowledge from prior literature (mulholland et al., 2016). regression 3 includes our full model with all specified control variables. for our variable of interest, cognitive ability, we see clear differences between the use of episodic memory and numeracy as the measure for this important respondent trait. both variables in the regression 1 analyses show a positive relation between the level of cognitive table 3 (continued) differences in means of demand determinants for individuals lapsing life insurance explanatory variables (1) lapse = 0 (2) lapse = 1 diff (1) 2 (2) health problems zero problems 0.1028 0.1068 �0.0040 one problem 0.2098 0.1979 0.0119 two problems 0.2636 0.2682 �0.0046 three problems 0.2257 0.2318 �0.0061 four or more problems 0.1981 0.1953 0.0028 estate tax vulnerable net worth over current exemptiona 0.0072 0.0182 �0.0110** shocks or major changes newly retired since 2008 0.1172 0.1589 �0.0416** income shock (decline) of >10% compared with 2008 0.4033 0.4271 �0.0238 additional kids residing at home since 2008 0.0487 0.0781 �0.0294*** marriage ended since last wave 0.0414 0.0521 �0.0107 health shock (2+ additional health issues) since last wave 0.0314 0.0313 0.0001 note. data from the 2010 health and retirement study. a t test is used for difference of means of the variables. statistical significance at 0.10, 0.05, and 0.01 levels is denoted by *, **, and ***, respectively. a2010 exemption was $0 but was retroactively changed to allow an election of $5,000,000. this variable set at $5,000,000. 86 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 t ab le 4 a l o g is ti c re g re ss io n s— co g n it iv e ab il it y m ea su re d w it h ep is o d ic m em o ry w av e 1 0 – 2 0 1 0 n = 8 ,7 9 5 3 8 4 re sp o n d en ts re p o rt ed la p se d ep en d en t v ar ia b le : la p se = 1 r eg 1 r eg 2 r eg 3 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 in te rc ep t �3 .2 5 5 < .0 0 0 1 * * * �3 .2 6 6 < .0 0 0 1 * * * �3 .4 4 6 < .0 0 0 1 * * * e xp la na to ry va ri ab le s c o g n it iv e a b il it y: e p is o d ic m em o ry (r ef = t w r q u a rt il e w it h lo w es t a b il it y) t w r q u ar ti le w it h 2 n d lo w es t ab il it y 1 .1 7 3 0 .2 7 0 3 1 .1 3 0 0 .4 0 6 3 1 .1 0 5 0 .5 0 7 3 t w r q u ar ti le w it h 2 n d h ig h es t ab il it y 1 .2 8 5 0 .0 8 3 1 * 1 .2 0 2 0 .2 1 7 1 1 .1 6 3 0 .3 3 7 9 t w r q u ar ti le w it h h ig h es t ab il it y 1 .3 9 9 0 .0 1 9 2 * * 1 .2 6 9 0 .1 1 8 4 1 .2 2 9 0 .2 1 4 5 e co n o m ic fa ct o rs in co m e q u ar ti le s (r ef = lo w es t q u ar ti le ) 2 n d lo w es t q u ar ti le 1 .2 9 5 0 .1 1 3 3 1 .2 8 8 0 .1 2 9 4 2 n d h ig h es t q u ar ti le 1 .1 9 4 0 .3 1 2 9 1 .1 3 9 0 .4 8 1 2 h ig h es t q u ar ti le 1 .2 9 8 0 .1 7 3 0 1 .1 7 3 0 .4 5 1 7 n et w o rt h q u ar ti le s (r ef = lo w es t q u ar ti le ) 2 n d lo w es t q u ar ti le 0 .5 2 5 < .0 0 0 1 * * * 0 .5 4 0 < .0 0 0 1 * * * 2 n d h ig h es t q u ar ti le 0 .5 3 0 < .0 0 0 1 * * * 0 .5 3 3 0 .0 0 0 2 * * * h ig h es t q u ar ti le 0 .7 8 2 0 .1 1 9 9 0 .7 3 7 0 .0 8 0 1 * l o g to ta l d eb t 1 .0 3 1 0 .0 0 5 3 * * * 1 .0 2 3 0 .0 4 1 2 * * f in a n ci a l kn o w le d g e f in an ci al re sp o n d en t 1 .0 6 3 0 .6 2 5 0 0 .9 8 5 0 .9 0 7 9 b eq u es t fa ct o rs m ar ri ed 1 .0 4 2 0 .7 6 2 5 1 .0 1 8 0 .9 0 3 6 h as li v in g ch il d re n 0 .9 8 4 0 .9 3 6 5 0 .9 5 9 0 .8 3 8 0 d em o g ra p h ic fa ct o rs a g e in 2 0 1 0 (r ef = ag es 6 0 to 6 4 ) l es s th an 6 0 0 .8 5 1 0 .3 4 5 1 6 5 to 6 9 0 .7 8 1 0 .1 6 5 1 7 0 to 7 4 0 .9 2 0 0 .6 1 7 4 (c o n ti n u ed o n n ex t p a g e) b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 87 t ab le 4 a (c o n ti n u ed ) w av e 1 0 – 2 0 1 0 n = 8 ,7 9 5 3 8 4 re sp o n d en ts re p o rt ed la p se d ep en d en t v ar ia b le : la p se = 1 r eg 1 r eg 2 r eg 3 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 7 5 to 7 9 0 .7 3 0 0 .1 1 4 8 8 0 an d h ig h er 0 .7 2 0 0 .1 1 9 3 m al e (r ef = fe m al e) 1 .2 4 8 0 .0 5 3 3 * r ac e (r ef = w h it e) b la ck 1 .2 6 1 0 .1 0 8 3 o th er 0 .7 5 8 0 .4 8 0 1 e d u ca ti o n (r ef = le ss th an h ig h sc h o o l d eg re e) h ig h sc h o o l 1 .1 7 7 0 .3 7 4 1 s o m e co ll eg e 1 .1 9 2 0 .3 7 9 0 c o ll eg e 1 .6 0 4 0 .0 2 2 9 * * h ea lt h p ro b le m s (r ef = ze ro p ro b le m s) o n e p ro b le m 0 .9 4 2 0 .7 6 4 2 t w o p ro b le m s 1 .1 0 0 0 .6 2 6 2 t h re e p ro b le m s 1 .1 3 6 0 .5 3 0 4 f o u r o r m o re p ro b le m s 1 .1 1 8 0 .6 0 2 0 e st a te ta x vu ln er a b le n et w o rt h o v er cu rr en t ex em p ti o n a 2 .0 7 8 0 .0 7 9 9 * s h o ck s o r m a jo r ch a n g es n ew ly re ti re d si n ce 2 0 0 8 1 .4 1 0 0 .0 1 9 7 * * in co m e d ec li n e > 1 0 % fr o m 2 0 0 8 (r ea l $ ) 1 .0 8 2 0 .4 8 6 3 a d d it io n al k id s re si d in g at h o m e si n ce 2 0 0 8 1 .5 3 3 0 .0 3 3 4 * * m ar ri ag e en d ed si n ce la st w av e 1 .3 5 4 0 .2 3 6 1 h ea lt h sh o ck si n ce la st w av e 0 .8 9 5 0 .7 1 7 1 p se u d o r 2 = 0 .0 0 2 3 0 .0 1 6 7 0 .0 2 9 8 n o te . d at a fr o m th e 2 0 1 0 h ea lt h an d r et ir em en t s tu d y . s ta ti st ic al si g n ifi ca n ce at 0 .1 0 , 0 .0 5 , an d 0 .0 1 le v el s is d en o te d b y * , * * , an d * * * , re sp ec ti v el y . a 2 0 1 0 e x em p ti o n w as $ 0 b u t w as re tr o ac ti v el y ch an g ed to al lo w an el ec ti o n o f $ 5 ,0 0 0 ,0 0 0 . t h is v ar ia b le se t at $ 5 ,0 0 0 ,0 0 0 . 88 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 t ab le 4 b l o g is ti c re g re ss io n s— co g n it iv e ab il it y m ea su re d w it h n u m er ac y w av e 1 0 – 2 0 1 0 n = 8 ,7 9 5 3 8 4 re sp o n d en ts re p o rt ed la p se d ep en d en t v ar ia b le : la p se = 1 r eg 1 r eg 2 r eg 3 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 in te rc ep t �3 .4 9 8 < .0 0 0 1 * * * �3 .4 1 4 < .0 0 0 1 * * * �3 .5 4 7 < .0 0 0 1 * * * e xp la na to ry va ri ab le s c o g n it iv e a b il it y: n u m er a cy (r ef = n u m er a cy q u a rt il e w it h lo w es t a b il it y) n u m er ac y q u ar ti le w it h 2 n d lo w es t ab il it y 1 .5 4 3 0 .0 0 5 7 * * * 1 .5 8 5 0 .0 0 4 1 * * * 1 .6 0 0 0 .0 0 5 0 * * * n u m er ac y q u ar ti le w it h 2 n d h ig h es t ab il it y 1 .7 0 1 0 .0 0 0 8 * * * 1 .7 1 9 0 .0 0 1 4 * * * 1 .6 8 4 0 .0 0 4 5 * * * n u m er ac y q u ar ti le w it h h ig h es t ab il it y 2 .1 2 0 0 .0 0 0 4 * * * 2 .0 2 4 0 .0 0 2 0 * * * 1 .8 6 7 0 .0 1 1 3 * * e co n o m ic fa ct o rs in co m e q u ar ti le s (r ef = lo w es t q u ar ti le ) 2 n d lo w es t q u ar ti le 1 .2 3 4 0 .1 9 9 0 1 .2 6 5 0 .1 5 9 7 2 n d h ig h es t q u ar ti le 1 .1 1 3 0 .5 4 3 3 1 .1 0 8 0 .5 7 9 0 h ig h es t q u ar ti le 1 .2 0 6 0 .3 2 8 5 1 .1 4 6 0 .5 2 0 5 n et w o rt h q u ar ti le s (r ef = lo w es t q u ar ti le ) 2 n d lo w es t q u ar ti le 0 .5 1 1 < .0 0 0 1 * * * 0 .5 3 3 < .0 0 0 1 * * * 2 n d h ig h es t q u ar ti le 0 .5 0 2 < .0 0 0 1 * * * 0 .5 2 3 < .0 0 0 1 * * * h ig h es t q u ar ti le 0 .7 2 4 0 .0 4 3 7 * * 0 .7 1 8 0 .0 5 7 6 * * l o g to ta l d eb t 1 .0 2 9 0 .0 0 8 2 * * * 1 .0 2 2 0 .0 5 0 6 * f in a n ci a l kn o w le d g e f in an ci al re sp o n d en t 1 .0 0 7 0 .9 5 8 6 0 .9 6 1 0 .7 5 8 6 b eq u es t fa ct o rs m ar ri ed 1 .0 0 8 0 .9 5 5 0 1 .0 1 3 0 .9 3 1 4 h as li v in g ch il d re n 0 .9 9 2 0 .9 6 8 7 0 .9 5 1 0 .8 0 8 5 d em o g ra p h ic fa ct o rs a g e (r ef = ag es 6 6 to 7 5 ) l es s th an 6 0 0 .8 5 0 0 .3 4 0 9 6 5 to 6 9 0 .7 8 4 0 .1 7 1 9 7 0 to 7 4 0 .9 1 8 0 .6 0 7 1 (c o n ti n u ed o n n ex t p a g e) b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 89 t ab le 4 b (c o n ti n u ed ) w av e 1 0 – 2 0 1 0 n = 8 ,7 9 5 3 8 4 re sp o n d en ts re p o rt ed la p se d ep en d en t v ar ia b le : la p se = 1 r eg 1 r eg 2 r eg 3 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 o d d s ra ti o p r > v 2 7 5 to 7 9 0 .7 2 7 0 .1 0 9 1 8 0 an d h ig h er 0 .7 1 3 0 .1 0 0 8 m al e (r ef = fe m al e) 1 .1 7 1 0 .1 6 3 2 r ac e (r ef = w h it e) b la ck 1 .3 6 5 0 .0 3 4 8 * * o th er 0 .7 9 6 0 .5 6 1 8 e d u ca ti o n (r ef = le ss th an h ig h sc h o o l d eg re e) h ig h sc h o o l 1 .0 6 2 0 .7 4 7 2 s o m e co ll eg e 1 .0 4 4 0 .8 3 0 8 c o ll eg e 1 .3 8 3 0 .1 2 8 1 h ea lt h p ro b le m s (r ef = ze ro p ro b le m s) o n e p ro b le m 0 .9 2 7 0 .7 0 4 3 t w o p ro b le m s 1 .0 8 2 0 .6 8 8 5 t h re e p ro b le m s 1 .1 1 8 0 .5 8 4 9 f o u r o r m o re p ro b le m s 1 .0 9 5 0 .6 7 2 3 e st a te ta x vu ln er a b le n et w o rt h o v er cu rr en t ex em p ti o n a 2 .0 8 5 0 .0 7 8 1 * s h o ck s o r m a jo r ch a n g es n ew ly re ti re d si n ce 2 0 0 8 1 .4 3 0 0 .0 1 5 3 * * in co m e d ec li n e > 1 0 % fr o m 2 0 0 8 (r ea l $ ) 1 .0 7 4 0 .5 2 9 3 a d d it io n al k id s re si d in g at h o m e si n ce 2 0 0 8 1 .5 1 8 0 .0 3 8 1 * * m ar ri ag e en d ed si n ce la st w av e 1 .3 7 2 0 .2 1 7 4 h ea lt h sh o ck si n ce la st w av e 0 .8 9 5 0 .7 1 7 6 p se u d o r 2 = 0 .0 0 6 3 0 .0 2 0 8 0 .0 3 3 1 n o te . d at a fr o m th e 2 0 1 0 h ea lt h an d r et ir em en t s tu d y . s ta ti st ic al si g n ifi ca n ce at 0 .1 0 , 0 .0 5 , an d 0 .0 1 le v el s is d en o te d b y * , * * , an d * * * , re sp ec ti v el y . a 2 0 1 0 e x em p ti o n w as $ 0 b u t w as re tr o ac ti v el y ch an g ed to al lo w an el ec ti o n o f $ 5 ,0 0 0 ,0 0 0 . t h is v ar ia b le se t at $ 5 ,0 0 0 ,0 0 0 . 90 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 ability and the probability of lapse. numeracy indicates higher levels of significance and probability. while numeracy maintains its high significance in all three forms of the model, episodic memory quickly loses significance as a predictor of voluntary lapse with the introduction of other independent variables. numeracy maintains consistent direction and magnitude of effect in each form of the model. in addition, using the pseudo r2 values as a measure of relative model strength, we see that the models using numeracy consistently display higher strength of association. based upon our statistically significant numeracy quartiles in our model, we conclude that cognitive ability is an important factor in the ownership decisions of existing life insurance policies. we also conclude that numeracy is a better measure of cognitive ability when modeling lapse decisions. our findings are consistent with the prior research that explores the appropriate measure of cognitive ability in relation to financial decisions (christelis et al., 2010; smith et al., 2010). we observe a consistent pattern across both analyses using the different measures of cognitive ability. higher levels of income, when compared with the lowest quartile of income, is not predictive of voluntary lapse while greater net worth is a significant predictor of a reduced likelihood to voluntary lapse a policy when compared with those respondents in the lowest quartile of net worth. total household debt is also a significant predictor of voluntary lapse, indicating that increasing debt increases the likelihood of policy lapse. for both net worth and total debt, the magnitude and direction of the effect are consistent between the reduced model in regression 2 and the full model in regression 3. we find no difference in likelihood to lapse a policy based upon the respondent’s household financial knowledge as proxied by whether or not they were the identified financial respondent for answering household financial questions in the survey. demographic control variables display similar effect patterns and magnitudes but with shifting significance with the change in cognitive ability measure. men, blacks, and those with college degrees appear more likely to lapse. as suggested in the univariate analysis, we use those in the typical preretirement or early retirement years, ages 60 to 64, as our reference group for the multivariate analyses. while falling just outside the traditional levels of significance, policyholders 75 and older appear to be nearly 30% less likely to lapse their existing policies than the reference age group. the presence of health problems was not indicative of a change in likelihood to voluntarily lapse a policy. when considering estate tax vulnerability of the household, we find that those respondents with household net worth in excess of the $5 million estate tax exemption level are significantly more likely to lapse a policy than those with less than this level of net worth. those from the vulnerable households are about twice as likely to lapse a policy. given prior research that suggests life insurance is a tool to improve the efficient transfer of the estate upon death (milevsky, 2006) and is used as such (mulholland et al., 2016), we initially find this result puzzling. we will address this puzzling result in the next section. finally, two important shock variables indicate a significant increase in likelihood to lapse a policy in the presence of these shocks. newly retired respondents are over 40% more likely to lapse a policy. in a new finding, we see that the addition of a child to the household of these older respondents increases the likelihood of lapsing a policy by over 50%. b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 91 overall, our results suggest numeracy as a measure cognitive ability is a significant predictor of life insurance policy lapse decisions. 5. discussion and conclusions the purpose of this study is to further explore the microeconomic determinants of life insurance lapse by introducing an additional variable into a model of lapse behavior. we find evidence that cognitive ability, when using numeracy as its measure, is a determinant in the decision to voluntarily lapse a life insurance policy. in addition, we find that specification of the correct measure of cognitive ability is important as seen in the lack of significant relationship when using episodic memory as the measure of cognitive ability, but a highly significant relationship when using numeracy as the cognitive ability measure. we conclude that numeracy is the appropriate measure of cognitive ability to include when modeling lapse behavior. our results are statistically significant. numeracy may be related to the ability of the policyholder to perceive the value of lapsing a policy at the appropriate time, such as when a term policy is no longer needed to hedge against the loss of labor income flows at retirement or a permanent policy is no longer needed after a loss of a bequest motive. it is important to note here that we are not able to comment on the specific rationality of the individual respondent’s decision to lapse a life insurance policy. with limitation in the data as to the specific reason the lapse decision is being made, such as loss of bequest motive, income shock, or health shock (fang & kung, 2012), we can only comment in general on the rationality of the decisions made by older policyholders. our results support prior literature in several ways. in the area of life cycle theory, we find that a greater proportion of policyholders are making the decision to lapse life insurance in the time frame surrounding when we see u.s. workers transition from the labor force to retirement, typically in the age 60 to age 64 timeframe (brown, 2013). our results suggest, though just beyond the typical level of significance, that in comparison, policyholders who are well into their retirement years are less likely to lapse their policies, possibly to meet bequest or end-of-life expenses. while it is not the focus of our study, we find some evidence supporting the efh. while controlling for cognitive ability, we find policyholders have an increased likelihood of lapsing a policy in the face of rising household debt. conversely, we find that those policyholders with higher levels of net worth are less likely to lapse a policy. income level is not significant in our model. we do not find evidence that income shock at the level we specify impacts the decision to lapse a policy, which is likely a result of the level of shock chosen. it is in the presence of very large income shocks that prior literature finds income shock to be a significant factor in life insurance lapse (fier & liebenberg, 2013; liebenberg et al., 2012). our findings in this model indicate that those individuals with household net worth above the 2010 tax exemption limit of $5 million are twice as likely to lapse their policy as those with net worth below the exemption limit. this result is counter to prior research that finds estate tax-vulnerable households are more likely to own cash value life insurance (mulholland et al., 2016). while on the surface this may be a puzzling result, we suggest 92 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 this result may be an artifact of the data. in the prior study, mulholland et al. use multiple years of cross-sectional data incorporating multiple waves of the survey of consumer finances (scf) to explore the relationship of life insurance demand to estate tax vulnerability. the 2010 wave of the scf was one of seven waves used. in contrast, our study uses only the 2010 cross section of the hrs. this is important because 2010 was a unique year for estate tax laws in the united states. the egtrra-2001 eliminated the estate taxes in 2010 while also making all assets subject to capital gains taxes. this changed the tax rates for estate transfers. for those households that were vulnerable to estate taxation, this may have encouraged them to reoptimize their life insurance coverage, an area of life insurance lapse theory covered by the prh. this is an area that should be considered for future study. we also find new evidence that policyholders who had one or more children recently move home to live with them are much more likely to lapse a policy. certainly, there may be an added expense component in the household and we suggest this is very similar to an income shock. however, this raises some interesting questions due to of the nature of the “expense shock.” are parents foregoing a typical bequest desire in exchange for current joint utility associated with helping their child? if the child knew about the exchange of utility by the parent and the impact it is having on the child’s own present and future utility, would the saliency of that exchange affect the child’s decision to move home? finally, if there are other children that would have benefited from the life insurance proceeds, the loss of their share of the death benefits is effectively a 100% tax on their proceeds by their sibling that moved home. understood in this manner, would the parent make the same choice or would they have sought policy support from the other siblings? the interplay of policy stakeholders in the lapse decision is an area of possible future research that may reveal some interesting results. our results are important to the various life insurance actors we discuss earlier in this article. for insurance companies, making tools available to policyholders and advisors that are targeted to the different levels of numerical ability may improve policyholder decision-making as they evaluate their current policies. for those policyholders with lower numerical ability, the current in-force ledger illustrations may confuse them to the point that they see lapse as their best, or only, option. it is a similar issue for regulators who are charged with protecting consumers. mandating education and protection tools that are sensitive to the policyholder’s cognitive ability should provide better protection for consumers. for advisors, a better understanding of the cognitive ability of their clients or prospects should allow the advisor to tailor the advice to better meet the consumer’s ability to understand it. if consumers are better informed, the advisors are less likely to the face legal issues like those that glenn neasham was forced to endure. we caution that while understanding that numerical ability is important for financial decisions like the ownership of life insurance, the questions used in the hrs should not be viewed as the best way to measure numeracy among policyholders. we leave that research to those studying financial literacy, which huston (2010) conceptualizes as having two dimensions, understanding and use. these are similar to the dimensions smith et al. (2010) use in their definition of crystallized intelligence—knowledge and experience. because b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 93 numeracy is a form of crystallized intelligence, there may be improved life insurance decisions with improved financial literacy about life insurance. notes 1 hrs question mt036: “in the last two years, have you allowed any life insurance policies to lapse or have any been cancelled?” 2 hrs question mt041: “was this lapse or cancellation something you chose to do, or was it done by the provider, your employer, or someone else?” 3 hrs question md178: “if the chance of getting a disease is 10 percent, how many people out of 1,000 would be expected to get the disease?” 4 hrs question md179: “if 5 people all have the winning numbers in the lottery and the prize is two million dollars, how much will each of them get?” 5 hrs question md180: “let’s say you have $200 in a savings account. the account earns 10 percent interest per year. how much would you have in the account at the end of two years?” references american council of life insurers (acli) (2012). acli life insurers fact book 2012, chapter 7. retrieved from https://www.acli.com/-/media/acli/files/fact-books-public/_factbook2012_entirety_020813.pdf banks, j., & oldfield, z. (2007). understanding pensions: cognitive function, numerical ability and retirement saving. fiscal studies, 28, 143–170. https://doi.org/10.1111/j.1475-5890.2007.00052.x bernheim, b. d. (1991). how strong are bequest motives? evidence based on estimates of the demand for life insurance and annuities. journal of political economy, 99, 899–927. https://doi.org/10.1086/261783 brown, a. (2013). in u.s., average retirement age up to 61: younger nonretirees most likely to expect to retire at a younger age (gallup poll 162560). available at http://www.gallup.com/poll/162560/averageretirement-age.aspx brown, j., & poterba, j. (2006). household ownership of variable annuities. nber working paper no. w11964. retrieved from http://ssrn.com/abstract=877469 browning, c. (2014). cognitive status, self-control, and asset decumulation: evidence from the hrs. retrieved from http://ssrn.com/abstract=2390880 campbell, r. (1980). the demand for life insurance: an application of the economics of uncertainty. the journal of finance, 35, 1155–1172. https://doi.org/10.1111/j.1540-6261.1980.tb02201.x christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54, 18–38. https://doi.org/10.1016/j.euroecorev.2009.04.001 cole, c., & fier, s. (2021). an examination of life insurance policy surrender and loan activity. journal of risk and insurance, 88, 483–516. https://doi.org/10.1111/jori.12329 daily, g., hendel, i., & lizzeri, a. (2008). does the secondary life insurance market threaten dynamic insurance? american economic review, 98, 151–156. https://doi.org/10.1257/aer.98.2.151 eling, m., & kiesenbauer, d. (2014). what policy features determine life insurance lapse? an analysis of the german market. journal of risk and insurance, 81, 241–269. https://doi.org/10.1111/j.1539-6975.2012.01504.x fang, h., & kung, e. (2012). why do life insurance policyholders lapse? the roles of income, health and bequest motive shocks. working paper 17899. available at http://www.nber.org/papers/w17899 fier, s. g., & liebenberg, a. p. (2013). life insurance lapse behavior. north american actuarial journal, 17, 153–167. https://doi.org/10.1080/10920277.2013.803438 94 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 finke, m. s., huston, s. j., & howe, j. s. (2017). old age and the decline in financial literacy. management science, 63, 213–230. https://doi.org/10.1287/mnsc.2015.2293 fischer, s. (1973). a life cycle model of life insurance purchases. international economic review, 14, 132– 152. https://doi.org/10.2307/2526049 fisher, g. g., hassan, h., rodgers, w. l., & weir, d. r. (2013). health and retirement study imputation of cognitive functioning measures: 1992–2010 (final release). hrs user guides. ann arbor, mi: survey research center. available at http://hrsonline.isr.umich.edu/modules/meta/xyear/cogimp/desc/cogimpdd. pdf frees, e. w., & sun, y. (2010). household life insurance demand. north american actuarial journal, 14, 338– 354. https://doi.org/10.1080/10920277.2010.10597595 gandolfi, a. s., & miners, l. (1996). gender-based differences in life insurance ownership. the journal of risk and insurance, 63, 683–693. https://doi.org/10.2307/253478 gerardi, k., goette, l., & meier, s. (2013). numerical ability predicts mortgage default. proceedings of the national academy of sciences of the united states of america, 110, 11267–11271. https://doi.org/10.1073/ pnas.1220568110 gottlieb, d., & smetters, k. (2021). lapse-based insurance. american economic review, 111, 2377–2416. https://doi.org/10.1257/aer.20160868 huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. https://doi.org/ 10.1111/j.1745-6606.2010.01170.x korniotis, g. m., & kumar, a. (2011). do older investors make better investment decisions? the review of economics and statistics, 93(1), 244–265. kuo, w., tsai, c., & chen, w. k. (2003). an empirical study on the lapse rate: the cointegration approach. journal of risk and insurance, 70, 489–508. https://doi.org/10.1111/1539-6975.t01-1-00061 lewis, f. (1989). dependents and the demand for life insurance. the american economic review, 79, 452– 467. liebenberg, a. p., carson, j. m., & dumm, r. e. (2012). a dynamic analysis of the demand for life insurance. journal of risk and insurance, 79, 619–644. https://doi.org/10.1111/j.1539-6975.2011.01454.x liebenberg, a. p., carson, j. m., & hoyt, r. e. (2010). the demand for life insurance policy loans. journal of risk and insurance, 77, 651–666. https://doi.org/10.1111/j.1539-6975.2010.01359.x lin, y., & grace, m. (2007). household life cycle protection: life insurance holdings, financial vulnerability, and portfolio implications. journal of risk and insurance, 74(1), 141–173. linton, m. a. (1932). panics and cash values. transactions of the actuarial society of america, 33, 265–394. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and overindebtedness (no. w14808). cambridge, ma: national bureau of economic research. https://www.nber.org/papers/w14808 mcardle, j. j., smith, j. p., & willis, r. (2011). cognition and economic outcomes in the health and retirement survey. in d. a. wise (ed.), explorations in the economics of aging (pp. 209–233). chicago, il: university of chicago press. milevsky, m. a. (2006). the calculus of retirement income: financial models for pension annuities and life insurance (pp 138–162). new york, ny: cambridge university press. mulholland, b. s., finke, m. s., & huston, s. j. (2016). understanding the shift in demand for cash value life insurance. risk management and insurance review, 19, 7–36. https://doi.org/10.1111/rmir.12031 nolte, s., & schneider, j. (2017). don’t lapse into temptation: a behavioral explanation for policy surrender. journal of banking & finance, 79, 12–27. https://doi.org/10.1016/j.jbankfin.2017.02.011 ofstedal, m. b., fisher, g. g., & herzog, a. r. (2005). documentation of cognitive functioning measures in the health and retirement study. hrs documentation report dr-006. institute for social research. ann arbor, mi: university of michigan. outreville, j. f. (1990). whole-life insurance lapse rates and the emergency fund hypothesis. insurance: mathematics and economics, 9, 249–255. https://doi.org/10.1016/0167-6687(90)90002-u rankin, d. (1987). personal finance: using life insurance as a tax shelter. new york times. available at http://www.nytimes.com/1987/05/10/business/personal-finance-using-life-insurance-as-a-tax-shelter.html? pagewanted=print&src=pm b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 95 russell, d. t., fier, s. g., carson, j. m., & dumm, r. e. (2013). an empirical analysis of life insurance policy surrender activity. journal of insurance issues, 36, 35–57. schott, f. h. (1971). disintermediation through policy loans at life insurance companies. the journal of finance, 26, 719–729. https://doi.org/10.1111/j.1540-6261.1971.tb01725.x smith, j. p., mcardle, j. j., & willis, r. (2010). financial decision making and cognition in a family context. economic journal (london, england), 120, f363–f380. https://doi.org/10.1111/j.1468-0297.2010.02394.x yaari, m. (1965). uncertain lifetime, life insurance, and the theory of the consumer. the review of economic studies, 32, 137–150. https://doi.org/10.2307/2296058 zietz, e. (2003). an examination of the demand for life insurance. risk management and insurance review, 6, 159–191. https://doi.org/10.1046/j.1098-1616.2003.030.x 96 b. s. mulholland and m. s. finke / financial services review 31 (2023) 73–96 the impact of the capitalization of operating leases: a guide for individual investors jack trifts, ph.d.a,*, gary e. porter, ph.d.b acollege of business, bryant university, smithfield, ri 02917, usa bd’amore-mckim school of business, northeastern university, boston, ma, 02115, usa abstract we provide a brief explanation of the new financial accounting standards board (fasb) standard requiring firms to move their off-balance sheet operating leases onto the balance sheet beginning in 2019, and then discuss how the new rule might affect the stock and bond values in the largest 1,000 listed firms. in short, despite dramatic increases in on-balance sheet liabilities in several industries, we caution investors not to anticipate changes in their stock or bond valuations resulting from this change. because asset values change in response to new information, and the information we present in this article regarding changes in total assets and debt ratios is currently available in the notes to the financial statements and from data providers such as bloomberg, it is already being used by professionals to forecast asset values. © 2017 academy of financial services. all rights reserved. jel classification: g11 keywords: operating leases; off-balance sheet financing; asset pricing 1. introduction in february 2016, the financial accounting standards board (fasb) released a revised standard (fasb 2016) on the accounting for leases. the most notable impact of this new standard will be the required capitalization of almost all leases that are currently categorized as operating leases. that is, under current standards operating leases have been not been reported as liabilities on the sheet, but the new standards will bring these obligations onto the * corresponding author. tel.: �1-401-447-2201; fax: �1-401-232-6319. e-mail address: jtrifts@bryant.edu (j. trifts) financial services review 26 (2017) 205–220 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. balance sheet. as we will show and discuss in this article, the impact on corporate balance sheets will range from minor to dramatic for u.s. companies. these changes will affect public companies in the united states beginning with their fiscal 2019 statements.1 however, since u.s. generally accepted accounting principles (u.s. gaap) and the sec require the presentation of the prior year’s historical balance sheet and two prior years of income statements and statements of cash flow, the impact of this new standard will be felt by u.s. corporations almost immediately.2 the purpose of this article is to detail the impact of this new standard on the top 1,000 largest firms and to provide insights for individual investors on the impact of this change. we focus on individual investors because they are most likely to be unfamiliar with the accounting for operating leases and with the impact of the pending change. individuals who have been relying on corporate balance sheets to measure the debt levels of the companies in which they invest may be surprised by the sometimes dramatic changes in the apparent debt load of companies most affected by the new standards. in contrast, most professional investment researchers, starting with graham and dodd, 1934, have long been adjusting for the impact of off-balance sheet operating leases. for example, the curriculum of the chartered financial analyst (cfa) program includes coverage of how to capitalize operating leases and restate the balance sheet for firm’s utilizing off-balance sheet financing (see, e.g., cfa, 2012). this article is organized as follows. we next provide a brief explanation of the change and how it will affect financial statements. we then discuss the literature regarding off-balance sheet accounting. in the following section, we discuss the data and methodology. we then discuss the impact of the change on all firms in the sample and then on specific subsets by industry. we then address key questions that may be on the mind of individual investors. we conclude with a brief summary and conclusion. 2. changing standards for lease accounting under the existing standards, u.s. corporations have long accounted for leases by categorizing them as either financial leases (sometimes referred to as capital leases) or operating leases. the distinction between the two has been driven by four criteria set out under u.s. gaap. essentially, leases could be treated as operating as long as their lease term was (1) less than 75% of the assets life, there was (2) no free transfer of the asset to the lessee at lease end, no (3) bargain price transfer of the asset to the lessee at the end of the lease period, and (4) the present value of the contractual lease payments did not exceed 90% of the value of the assets at the inception of the lease. firms are required to report operating leases in footnotes,3 but by carefully structuring the terms of the lease contracts, companies have been able to acquire the use of a wide range of assets without the need to record any liability on their balance sheets. operating leases, which are typically long-term and non-cancellable, are liabilities that are the equivalent of debt, and the existing standard has systematically understated those liabilities on the balance sheet of companies utilizing these leases. the recognition of this problem is not new. for example, in a november 2003 speech by former sec chairman 206 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 arthur levitt, jr. (levitt, 2004) to the partners of the accounting firm kpmg, he noted that “billions of dollars of lease financing fail to show up on balance sheets.” more recently, katz (2016) noted that fasb technicians found approximately “north of a trillion [dollars] in undiscounted lease obligations that are reported in the footnotes.” under the new standards, the distinction between financial and operating leases will be maintained, but both types of leases must be capitalized, eliminating most opportunities for off-balance sheet financing. the balance sheet treatment of both types of leases will be identical with the present value of lease payments being recorded as debt and the corresponding asset value recorded as an asset. the difference between the two types of leases will show up on the income statement. finance leases, as they will now be called, will be expensed through a combination of depreciation of the asset and the amortization of the interest portion of the debt obligation with the total expense depending on both the implicit interest rate of the lease and the depreciation rate of the asset. in contrast, the periodic expenses of operating leases will equal the lease payments. this treatment of operating leases will result in an expense stream essentially identical to the expenses for current operating leases.4 a major driver of this rule change has been the ongoing attempt of fasb and its international counterpart, the international accounting standards board (iasb) to bring the two sets of standards closer together, commonly referred to as “convergence.” this change in leasing standards will accomplish this but the two sets of standards will still not be identical. the iasb does require the capitalization of all operating leases except those that are truly short-term, defined as less than one year and without renewal options. however, the iasb does not differentiate between finance and operating leases like the fasb standards. while this will result in somewhat different expense recognition patterns for operating leases in the u.s., the balance sheet treatments will be essentially identical. 3. literature there is a voluminous literature on leases and lease accounting (e.g., see wheeler spencer & webb, 2015) but we concentrate on two questions: how has the use of operating leases changed over time and why have companies chosen to use them? as we show in this study, operating leases are currently an important form of financing for many firms. the usage of these leases has grown over time. for example, cornaggia, franzen, and simin (2013) tracked the use of leases for all firms in the merged crspcompustat database (essentially all u.s.-listed public firms) excluding financial and utilities from 1980 to 2007. they document a dramatic increase in the use of operating leases (off-balance-sheet) compared with capital (on-balance-sheet) leases. specifically, they compared the proportion of total debt represented by each lease type and document a 745% increase in the use of operating leases compared to (on-balance sheet) capital leases. why is there such a significant increase in the use of operating leases? while some companies are able to capture tax benefits through leasing versus buying, the tax benefits are not affected by whether a particular lease is classified as capital or operating. beatty, liao, and weber (2010) find that companies with lower accounting quality tend to make higher use 207j. trifts, g.e. porter / financial services review 26 (2017) 205–220 of leases. this may reflect lenders’ desires to maintain formal title to the assets when dealing with firms with lower accounting quality or a preference on the part of the companies to keep debt off-balance sheet with operating leases. perhaps the use of operating leases is driven by the belief that since these obligations do not appear on the balance sheets they are “free debt” and not recognized by lenders and investors. bryan, lilien, and martin (2010) address this motivation, concluding that firms using operating leases count on “functional fixation,” a term for market participants who blindly use only balance sheet and income statements, ignoring reporting in footnotes. this belief, if true, is misguided at least for lenders and professional investors. beatty et al., (2010) note that bank monitoring, and access to private information from the firms to which they lend, substitute for the lower quality of accounting in these firms. also, as noted earlier professional investors are well aware of the impact of operating leases and routinely make adjustments for this off-balance sheet debt when evaluating firms as potential investors. certainly, neither of these groups are ignoring information reported in the notes to financial statements. it may be possible that individual investors have a higher propensity to display “functional fixation.” indeed, it is the purpose of this article to reduce this propensity by educating investors to the significance of off-balance sheet debt as well as the minor impact of the coming rules changes on securities prices. 4. data and methodology data for the largest 1,000 firms by market capitalization (number of outstanding shares times price per share) was taken from bloomberg in early may 2016. this time selection was based on the desire to have all firms with fiscal year ends in both december 2015 and january 2016 included with their most recent annual information. under security and exchange commission (sec) reporting requirements, firms have 60 to 75 days, depending on size, from the end of their fiscal year to file their annual report (10k). most retailers end their fiscal years at or near the end of january so all should have completed their filing by mid-april. we allowed about three weeks to ensure that our data sources would be complete and up to date. to estimate the amount of off-balance sheet debt represented by operating leases, we calculated the estimated present value of the minimum rental obligations as provided by bloomberg and taken from each firm’s 10k. a detailed example using urban outfitters is included in the appendix. current reporting requirement for operating leases require firms to report the minimum operating lease payments for each of the next five years and lump all subsequent payments into one sum, labeled “thereafter” in the notes to the financial statements. to estimate the present value of these payments, two assumptions are required. first, one must make some assumption about how payments beyond year five will occur. we assumed that payments beyond year five would occur at the same rate as those in year five. second, a discount rate is required to calculate the present value of the series. bloomberg provides an estimate of each firms after-tax cost of debt from its overall weighted average cost of capital (wacc). of the 1,000 firms in the sample, 938 had values for this after-tax cost of debt provided. because the discount rate is approximately the pretax cost of debt, we 208 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 estimated this value for each firm by dividing the after-tax cost by 0.65, or (1–0.35), assuming a marginal tax rate of 35%, the top marginal u.s. federal corporate tax rate. sixty two firms did not have cost of debt estimates provided by bloomberg, requiring us to make other estimates. seven of these firms had standard and poor’s debt ratings provided by bloomberg and, in these cases we used the average rates of other firms in the sample with the same rating as proxies for the cost of debt. two additional firms had moody’s debt ratings, one provided by bloomberg and the other reported on factset. in these cases, we matched the moody’s rating to its equivalent standard and poor’s rating and followed the same process as above.5 we could find no direct indicator of credit risk for six of the firms. these ranged from the 238th largest firm by market capitalization, sigma aldrich, to the 990th firm, kite pharma. for these firms, we made the assumption that their credit rating would be the same as other firms with the same four-digit standard industrial classification (sic) code. although this approach does not consider differences in leverage and profitability, it does take into consideration similarities in the business and industry. for industry classifications, we used two digit standard industrial classification (sic) codes. we experimented with using three-digit codes, which provide more specific definitions but result in very small industry groups. for example, food stores fall under sic codes 54 and this group is further subdivided into codes 541 through 546 plus a miscellaneous category that distinguish between grocery stores (541), fish markets (542) and fruit and vegetable stores (543), and so forth. for purposes of examining the usage of operating leases, the two digit codes provide sufficient detail. for each company, we estimated the dollar value of off-balance sheet operating leases. we also looked at the impact of capitalizing this off-balance sheet debt on three metrics. first, we examined the change in total assets that results from the recognition of the assets acquired through operating leases. second, since operating leases represent hidden debt, we examined the impact on two debt-related metrics, book value debt to total assets and book value debt to total capital. this last metric is calculated as the book value of short and long term debt to the sum of the book value of short and long term debt plus the market value of the equity. this is not a perfect measure of firms’ market value debt ratios but substituting the market value of equity results in a measure that more closely approximates true effective leverage than the book value measure. 5. impact of the new leasing standards to examine the potential impact of the new standards on financial reporting, we focused on how the largest 1,000 firms’ financial statements would have been different if the standards had been in place for fiscal 2015. the total amount of off-balance sheet financing utilized by the 1,000 largest companies is estimated to be $742 billion, almost three quarters of a trillion dollars. the largest amount of off-balance sheet operating leases were held by walgreens boots alliance with $30.8 billion, followed by at&t ($25.7 billion), cvs health ($23.8 billion), wal-mart ($18.9 billion), and united continental holdings ($16.4 209j. trifts, g.e. porter / financial services review 26 (2017) 205–220 billion). of the 1,000 firms included in our analysis, only 18 reported no material operating leases. table 1 shows the value of off balance sheet financing for the 25 industries most affected. the number one industry by dollar value of operating leases is air transportation. most airlines use operating leases to acquire at least a portion of their fleets. however, even within this industry, the usage of operating leases varies significantly. united continental, as previously noted, is in the top five largest users of operating leases while its smaller rival allegiant has only $33.1 million. part of this difference is certainly the difference in size of the two airlines but may also reflect differences in business practices. for example, allegiant’s business model relies heavily on acquiring older aircraft that may be more difficult to structure as operating leases. retailers are also heavily represented in this listing. from drug stores to restaurants to hardware discounters to grocery stores, these companies tend to be heavy users of operating leases. for most of these companies, the bulk of their operating leases are on their stores. while the above numbers are large, their relative importance of operating leases for each company must be gauged relative to its size. to do this, we examined the percentage change in total assets that would result from the inclusion of off-balance sheet operating leases. for the 1,000 firms in total, the median change in total assets is only 2.6% and the average is only 7.3%. taken alone, these values might suggest that the pending change to the new standards is only a minor adjustment best left for the accountants to worry about. however, the median and average reflect the wide differences that occur between companies and across industries. firms in some industries tend to use operating leases much more aggressively and the degree of operating lease usage differs even within specific industries. for example, four firms had off-balance sheet assets and corresponding debt that exceeded the total of assets shown in their fiscal 2015 balance sheets. whole foods’ total assets are 125.3% greater (i.e., 2.25 times) when the impact of operating leases are included. the other three exceeding the 100% level are chipotle mexican grill (110.8%), jack in the box (104.0%), and regal entertainment group (101.3%). this suggests that investors must look at specific industries and companies when considering the impact of operating leases, because average market statistics can be misleading. table 2 shows the percentage change in total assets that would result from the capitalization of operating leases. the top three industries and five of the top eight industries are in retail. however, even within this industry, there are wide variations in the use of operating leases as a percentage of total assets. the group with the largest overall change is retail trade–apparel and accessory stores with an average of 63.3% overall. however, there is still a wide range with urban outfitters with 97.1% of additional assets off-balance sheet while under armour shows only a 22.5% increase. one of the most important results of the new standard will be to make the balance sheet more reflective of the true debt usage of the firm. to examine the impact of including operating leases as debt, we estimated the change in each firm’s book value debt ratio, defined as total liabilities divided by total assets. while this measure may not be as important as the market value measures of leverage that will be discussed shortly, it is the measure that can be directly computed from a firm’s balance sheet. for the entire 1,000 firms, the average change in debt ratio was 3.9% and the median change was 1.6%. as with the change in total 210 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 t ab le 1 v al ue of of fba la nc e sh ee t op er at in g le as es l ar ge st 25 di vi si on s, by tw odi gi t si c co de s 20 15 fis ca l ye ar d iv is io n si c fi rm s d iv is io n m ea n ($ m il) fi rm s w ith la rg es t le as t ob lig at io ns fi rm s w ith sm al le st le as t ob lig at io ns t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n by a ir 45 10 $6 ,2 41 .2 8 $1 6, 40 6. 40 u ni te d c on tin en ta l h ol di ng s $3 3. 10 5 a lle gi an t t ra ve l c o r et ai l t ra de –m is ce lla ne ou s r et ai l 59 19 $4 ,4 71 .0 2 $3 0, 78 8. 64 w al gr ee ns b oo ts a lli an ce in c $3 5. 77 8 v is ta o ut do or in c r et ai l t ra de –g en er al m er ch an di se st or es 53 12 $3 ,9 65 .6 4 $1 8, 89 3. 05 w al -m ar t st or es in c $3 .2 35 c as ey ’s g en er al st or es in c r et ai l t ra de –f oo d st or es 54 5 $3 ,3 80 .1 2 $7 ,4 02 .3 4 k ro ge r c o $5 36 .3 98 g n c h ol di ng s in cc la ss a r et ai l t ra de –a pp ar el & a cc es so ry st or es 56 11 $3 ,0 16 .9 4 $7 ,3 76 .9 7 t jx c om pa ni es in c $6 45 .6 46 u nd er a rm ou r in c– c la ss a r et ai l t ra de –b ui ld in g m at er ia ls , h ar dw ar e, g ar de n su pp ly an d m ob ile h om e d ea le rs 52 5 $2 ,9 88 .1 3 $6 ,5 34 .3 8 h om e d ep ot in c $2 87 .9 13 fa st en al c o t ra ns po rt & pu bl ic u til ity –c om m un ic at io ns 48 30 $2 ,8 73 .2 1 $2 5, 72 2. 36 a t & t in c $0 .0 00 c b s c or p– c la ss a vo tin g r et ai l t ra de –e at in g & d ri nk in g pl ac es 58 14 $2 ,2 45 .1 9 $1 0, 67 8. 50 m cd on al ds c or p $4 40 .3 52 b ri nk er in te rn at io na l in c r et ai l tr ad e– h om e fu rn itu re , fu rn is hi ng , & e qu ip m en t st or es 57 4 $2 ,1 78 .2 3 $3 ,1 43 .9 9 b es t b uy c o in c $9 95 .2 51 g am es to p c or pc la ss a se rv ic es –m ot io n pi ct ur es 78 8 $1 ,5 61 .1 0 $3 ,3 51 .8 9 a m c e nt er ta in m en t h ld sc la ss a $9 9. 21 2 l io ns g at e e nt er ta in m en t c or p m an uf ac tu ri ng –p et ro le um r efi ni ng an d r el at ed in du st ri es 29 14 $1 ,4 02 .2 1 $4 ,6 44 .5 1 e xx on m ob il c or p $9 1. 79 2 m ur ph y u sa in c r et ai l t ra de –a ut om ot iv e d ea le rs an d g as ol in e se rv ic e st at io ns 55 10 $1 ,3 10 .4 4 $4 ,1 55 .9 2 pe ns ke a ut om ot iv e g ro up $1 38 .8 42 c op ar t in c fi n in s an d r ea l e st at e– in su ra nc e a ge nt s, b ro ke rs an d se rv ic e 64 5 $1 ,1 73 .4 9 $2 ,1 10 .9 8 m ar sh & m cl en na n c os $1 80 .7 74 b ro w n & b ro w n in c m an uf ac tu ri ng –a pp ar el , an d ot he r fi ni sh ed pr od uc ts m ad e fr om fa br ic an d si m ila r m at er ia ls 23 8 $1 ,1 20 .3 8 $2 ,0 21 .8 5 b ur lin gt on st or es in c $3 11 .8 78 c ol um bi a sp or ts w ea r c o m an uf ac tu ri ng –l ea th er an d l ea th er pr od uc ts 31 2 $1 ,0 88 .8 7 $1 ,2 20 .7 0 c oa ch in c $9 57 .0 44 sk ec he rs u sa in c– c la ss a a gr ic ul tu re –a gr ic ul tu ra l se rv ic es 07 1 $9 96 .1 8 $9 96 .1 8 v c a in c $9 96 .1 77 v c a in c t ra ns po rt & pu bl ic u til ity –r ai lr oa d t ra ns po rt at io n 40 6 $8 96 .7 5 $3 ,0 89 .8 1 u ni on pa ci fic c or p $2 83 .7 78 k an sa s ci ty so ut he rn se rv ic es –s oc ia l se rv ic es 83 1 $7 17 .0 7 $7 17 .0 7 b ri gh t h or iz on s fa m ily so lu tio ns $7 17 .0 70 b ri gh t h or iz on s fa m ily so lu tio ns se rv ic es –h ea lth se rv ic es 80 16 $6 98 .8 1 $2 ,6 21 .4 6 d av ita h ea lth ca re pa rt ne rs in c $7 3. 94 8 c he m ed c or p m an uf ac tu ri ng –r ub be r an d m is ce lla ne ou s pl as tic pr od uc ts 30 7 $6 06 .9 1 $2 ,6 42 .6 5 n ik e in cc la ss b $2 2. 86 7 a rm st ro ng w or ld in du st ri es t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n se rv ic es 47 6 $5 64 .0 5 $1 ,8 98 .5 2 x po l og is tic s in c $1 25 .2 32 e xp ed ito rs in tl w as h in c se rv ic es –a m us em en t an d r ec re at io n se rv ic es 79 4 $5 49 .2 4 $1 ,8 01 .0 6 l iv e n at io n e nt er ta in m en t in c $3 5. 13 7 c hu rc hh ill d ow ns in c c on st ru ct io n– h ea vy c on st ru ct io n ot he r th an b ui ld in g c on st ru ct io n c on tr ac to rs 16 2 $5 38 .4 2 $8 19 .7 4 ja co bs e ng in ee ri ng g ro up in c $2 57 .1 04 fl uo r c or p m an uf ac tu ri ng –i nd us tr ia l an d c om m er ci al m ac hi ne ry an d c om pu te r e qu ip m en t 35 53 $5 02 .1 3 $5 ,7 14 .8 0 a pp le in c $9 .6 45 b w x t ec hn ol og ie s in c se rv ic es –h ot el s, r oo m in g h ou se s, c am ps , an d o th er l od gi ng pl ac es 70 12 $4 98 .1 2 $1 ,9 49 .1 7 h ilt on w or ld w id e h ol di ng s in c $5 7. 17 6 e xt en de d st ay a m er ic a in c 211j. trifts, g.e. porter / financial services review 26 (2017) 205–220 t ab le 2 c ha ng e in to ta l as se ts if pr op os ed ch an ge s w er e ap pl ie d to 20 15 fis ca l ye ar la rg es t 25 di vi si on s, by tw odi gi t si c co de s d iv is io n si c fi rm s d iv is io n av er ag e ch an ge l ar ge st ch an ge in di vi si on sm al le st ch an ge in di vi si on r et ai l t ra de –a pp ar el & a cc es so ry st or es 56 11 63 .3 % 97 .1 % u rb an o ut fit te rs in c 22 .5 % u nd er a rm ou r in cc la ss a r et ai l t ra de –f oo d st or es 54 5 53 .8 % 12 5. 3% w ho le fo od s m ar ke t in c 18 .7 % d un ki n’ b ra nd s g ro up in c r et ai l t ra de –e at in g & d ri nk in g pl ac es 58 14 51 .5 % 11 0. 8% c hi po tle m ex ic an g ri ll in c 4. 8% a ra m ar k a gr ic ul tu re –a gr ic ul tu ra l se rv ic es 07 1 39 .7 % 39 .7 % v c a in c 39 .7 % v c a in c r et ai l t ra de –h om e fu rn itu re , fu rn is hi ng & e qu ip m en t st or es 57 4 39 .5 % 66 .1 % w ill ia m sso no m a in c 23 .0 % g am es to p c or pc la ss a m an uf ac tu ri ng –l ea th er an d l ea th er pr od uc ts 31 2 36 .5 % 46 .7 % sk ec he rs u sa in cc la ss a 26 .2 % c oa ch in c se rv ic es –s oc ia l se rv ic es 83 1 33 .3 % 33 .3 % b ri gh t h or iz on s fa m ily so lu tio ns 33 .3 % b ri gh t h or iz on s fa m ily so lu tio ns r et ai l t ra de –m is ce lla ne ou s r et ai l 59 19 31 .8 % 95 .8 % d ic k’ s sp or tin g g oo ds in c 0. 5% e xp re ss sc ri pt s h ol di ng c o m an uf ac tu ri ng –a pp ar el , an d ot he r fi ni sh ed pr od uc ts m ad e fr om fa br ic an d si m ila r m at er ia ls 23 8 30 .5 % 78 .4 % b ur lin gt on st or es in c 6. 6% h an es br an ds in c b ui ld in g m at er ia ls , h ar dw ar e, g ar de n su pp ly an d m ob ile h om e d ea le rs 52 5 30 .1 % 86 .0 % t ra ct or su pp ly c om pa ny 11 .4 % fa st en al c o se rv ic es –m ot io n pi ct ur es 78 8 28 .6 % 10 1. 3% r eg al e nt er ta in m en t g ro up -a 1. 8% t im e w ar ne r in c t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n by a ir 45 10 26 .3 % 55 .8 % sp ir it a ir lin es in c 2. 4% a lle gi an t t ra ve l c o r et ai l t ra de –g en er al m er ch an di se st or es 53 12 21 .3 % 56 .6 % d ol la r g en er al c or p 0. 1% c as ey ’s g en er al st or es in c r et ai l t ra de � a ut om ot iv e d ea le rs an d g as ol in e se rv ic e st at io ns 55 10 18 .3 % 51 .8 % pe ns ke a ut om ot iv e g ro up in c 2. 7% c ar m ax in c se rv ic es –a m us em en t an d r ec re at io n se rv ic es 79 4 11 .3 % 29 .3 % l iv e n at io n e nt er ta in m en t in 1. 5% c hu rc hi ll d ow ns in c se rv ic es –e du ca tio na l se rv ic es 82 2 10 .8 % 12 .8 % g ra ha m h ol di ng s c oc la ss b 8. 9% h ou gh to n m if fli n h ar co ur t c o t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n se rv ic es 47 6 7. 2% 15 .0 % x po l og is tic s in c 2. 4% pr ic el in e g ro up in c/ t he c on st ru ct io n– h ea vy c on st ru ct io n ot he r th an b ui ld in g c on st ru ct io n c on tr ac to rs 16 2 6. 9% 10 .5 % ja co bs e ng in ee ri ng g ro up in c 3. 4% fl uo r c or p fi n in s an d r ea l e st at e– in su ra nc e a ge nt s, b ro ke rs an d se rv ic e 64 5 6. 7% 11 .6 % m ar sh & m cl en na n c os 3. 6% b ro w n & b ro w n in c se rv ic es –h ea lth se rv ic es 80 16 6. 7% 26 .0 % b ro ok da le se ni or l iv in g in c 2. 1% m ed na x in c w ho le sa le t ra de –d ur ab le g oo ds 50 20 6. 6% 17 .6 % a ir ga s in c 1. 7% a rr ow e le ct ro ni cs in c se rv ic es –b us in es s se rv ic es 73 12 0 6. 5% 36 .6 % t ab le au so ft w ar e in cc la ss a 0. 0% v er is ig n in c se rv ic es –e ng in ee ri ng , a cc ou nt in g, r es ea rc h, m an ag em en t an d r el at ed se rv ic es 87 13 6. 4% 13 .6 % pa re xe l in te rn at io na l c or p 0. 5% se rv ic em as te r g lo ba l h ol di ng w ho le sa le t ra de –n on -d ur ab le g oo ds 51 14 6. 1% 31 .9 % d om in o’ s pi zz a in c 0. 5% pi nn ac le fo od s in c se rv ic es –h ot el , r oo m in g h ou se s, c am ps , an d ot he r l od gi ng pl ac es 70 12 5. 8% 13 .8 % m ar ri ot t in te rn at io na lc la ss a 0. 5% l as v eg as sa nd s c or p 212 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 assets, these relatively small values might suggest that the capitalization of operating leases is a relatively unimportant issue as the change in leverage is small. however, as before, these overall statistics mask the wide range of values across the entire sample. for example, the firm showing the largest change in their debt ratio is whole foods that shows a debt ratio of 1.1% without operating leases and 56.1% with the inclusion of the off-balance sheet debt. second is chipotle (0% before, 52.6% after inclusion), dick’s sporting goods (0.2% to 49.0%), and american eagle outfitters (0% to 47.8%). the magnitude of these differences illustrates an important point about analyzing financial statements. investors who relied on the unadjusted balance sheets to assess financial risk would logically assess the above firms as having no financial risk because they each carry effectively no debt. however, in reality, these firms have debt equal to about half of the book value of their total assets. table 3 shows the changes in ratio of book value of debt to total assets by industry group. the retail industry is highly represented among those companies with the greatest change. however, as with other metrics, a key result is that there is substantial variation even within industry segments. for example, the segment with the largest change in debt ratio is retail trade–apparel and accessory stores. on average, retailers in this segment show debt ratios that are 30.9 percentage points higher when operating leases are capitalized. however, american eagle outfitters debt ratio rises 47.8 percentage points while l brands ratio increases only by 10.8 percentage points. while potential investors would be wise to be suspicious of the impact of operating leases on the debt loads of all retailers, they also need to carefully look at differences across firms, not just rely on industry averages. even industries that tend to use relatively few operating leases can have wide ranges of metrics across firms. consider the services–business services segment near the bottom of table 3. across the 120 firms in this segment, the average change in book value debt ratio is only 4.4%. this relatively low number might lead some investors to not worry about off-balance sheet debt for firms in this segment. however, tableau software does make substantial use of operating leases and their debt ratio is 26.8 percentage points higher when those leases are capitalized. the capitalization of leases adds debt to the balance sheet and thus increases the debt ratio for almost all firms. however, it is possible that book value debt ratios could actually improve with the inclusion of capitalized operating leases. for example, in the retail trade–miscellaneous retail segment, note that michaels company shows a change in the debt ratio of �17.2 percentage points. this results because the company has total debt that exceeds the book value of its total assets (i.e., negative equity). in this unusual case, adding the present value of the operating leases to both the firm’s debt and total assets actually decreases the debt ratio. this anomaly was seen in four additional firms: choice hotels, sba communications, cablevision systems ny, and domino’s pizza. while book value measures of leverage are frequently reported, market value measures more accurately portray the true leverage position of firms. to examine this, we calculated each firm’s debt to total capital ratio, defined as (book value) total short and long-term debt divided by the sum of total short and long-term debt plus the market value of equity. ideally the market value of debt should be used in this ratio, but this value is unobtainable for most firms. furthermore, unless a firm’s default rate has changed dramatically or interest rates have moved dramatically, the book value of debt will closely approximate its market value. 213j. trifts, g.e. porter / financial services review 26 (2017) 205–220 t ab le 3 c ha ng e in bo ok va lu e of de bt to to ta l as se ts if pr op os ed ch an ge s w er e ap pl ie d to 20 15 fis ca l ye ar la rg es t 25 di vi si on s, by tw odi gi t si c co de s d iv is io n si c fi rm s d iv is io n av er ag e ch an ge l ar ge st ch an ge in ea ch di vi si on sm al le st ch an ge in ea ch di vi si on r et ai l t ra de –a pp ar el & a cc es so ry st or es 56 11 30 .9 % 47 .8 % a m er ic an e ag le o ut fit te rs 10 .8 % l b ra nd s in c r et ai l t ra de –h om e fu rn itu re , fu rn is hi ng , & e qu ip m en t st or es 57 4 24 .4 % 39 .8 % w ill ia m sso no m a in c 16 .4 % b es t b uy c o in c m an uf ac tu ri ng –l ea th er an d l ea th er pr od uc ts 31 2 23 .7 % 30 .5 % sk ec he rs u sa -c la ss a 16 .8 % c oa ch in c r et ai l t ra de –f oo d st or es 54 5 22 .7 % 55 .0 % w ho le fo od s m ar ke t in c 3. 7% d un ki n’ b ra nd s g ro up in c r et ai l t ra de –e at in g & d ri nk in g pl ac es 58 14 22 .7 % 52 .6 % c hi po tle m ex ic an g ri ll in c 2. 2% a ra m ar k a gr ic ul tu re –a gr ic ul tu ra l se rv ic es 07 1 18 .5 % 18 .5 % v c a in c 18 .5 % v c a in c m an uf ac tu ri ng –a pp ar el , an d ot he r fi ni sh ed pr od uc ts m ad e fr om fa br ic an d si m ila r m at er ia ls 23 8 15 .8 % 26 .9 % l ul ul em on a th le tic a in c 3. 3% h an es br an ds in c b ui ld in g m at er ia ls , h ar dw ar e, g ar de n su pp ly an d m ob ile h om e d ea le rs 52 5 15 .7 % 43 .0 % t ra ct or su pp ly c om pa ny 6. 6% h om e d ep ot in c t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n by a ir 45 10 14 .9 % 26 .7 % sp ir it a ir lin es in c 1. 3% a lle gi an t t ra ve l c o se rv ic es –s oc ia l se rv ic es 83 1 14 .1 % 14 .1 % b ri gh t h or iz on s fa m ily so lu tio ns 14 .1 % b ri gh t h or iz on s fa m ily so lu tio ns r et ai l t ra de –m is ce lla ne ou s r et ai l 59 19 13 .1 % 48 .8 % d ic k’ s sp or tin g g oo ds in c � 17 .2 % m ic ha el s c os in c/ t he r et ai l t ra de –g en er al m er ch an di se st or es 53 12 11 .3 % 33 .5 % b ig l ot s in c 0. 1% c as ey ’s g en er al st or es r et ai l t ra de –a ut om ot iv e d ea le rs an d g as ol in e se rv ic e st at io ns 55 10 8. 5% 22 .4 % a dv an ce a ut o pa rt s in c 0. 7% c ar m ax in c se rv ic es –e du ca tio na l se rv ic es 82 2 8. 2% 10 .3 % g ra hm h ol di ng s c oc la ss b 6. 1% h ou gh to n m if fli n h ar co ur t c o se rv ic es –m ot io n pi ct ur es 78 8 7. 7% 23 .8 % a m c e nt er ta in m en t h ld sc l a 1. 1% t im e w ar ne r in c se rv ic es –a m us em en t an d r ec re at io n se rv ic es 79 4 6. 0% 15 .1 % l iv e n at io n e nt er ta in m en t in c 1. 0% c hu rc hi ll d ow ns in c c on st ru ct io n– h ea vy c on st ru ct io n ot he r th an b ui ld in g c on st ru ct io n c on tr ac to rs 16 2 5. 8% 8. 8% ja co bs e ng in ee ri ng g ro up in c 2. 8% fl uo r c or p t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n se rv ic es 47 6 4. 9% 8. 3% t ri pa dv is or in c 1. 5% pr ic el in e g ro up in c/ t he fi n in s an d r ea l e st at e– in su ra nc e a ge nt s, b ro ke rs an d se rv ic e 64 5 4. 9% 7. 9% m ar sh & m cl en na n c os 2. 7% b ro w n & b ro w n in c se rv ic es –b us in es s se rv ic es 73 12 0 4. 4% 26 .8 % t ab le au so ft w ar e in cc l a 0. 0% v er is ig n in c w ho le sa le t ra de –d ur ab le g oo ds 50 20 4. 4% 9. 4% po ol c or p 1. 3% a rr ow e le ct ro ni cs in c c on st ru ct io n– sp ec ia l t ra de c on tr ac to rs 17 3 4. 0% 5. 4% e m c o r g ro up in c 2. 8% c hi ca go b ri dg e & ir on c o n v se rv ic es –p er so na l se rv ic es 72 3 3. 8% 9. 3% h & r b lo ck in c 0. 6% se rv ic e c or p in te rn at io na l se rv ic es –e ng in ee ri ng , a cc ou nt in g, r es ea rc h, m an ag em en t an d r el at ed se rv ic es 87 13 3. 7% 9. 7% pa re xe l in te rn at io na l c or p 0. 2% se rv ic em as te r g lo ba l h ol di ng m an uf ac tu ri ng –r ub be r an d m is ce lla ne ou s pl as tic pr od uc ts 30 7 3. 4% 10 .3 % n ik e in cc la ss b 0. 5% a rm st ro ng w or ld in du st ri es 214 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 in contrast, the market value of equity typically will exceed a firm’s book value of equity by a substantial margin. table 4 shows the changes in firms’ debt to total capital ratio. overall, the results are very similar to those shown in table 3. eight of the top ten industries in table 4 are in the top 10 of table 3, the other two ranking in the top 15. in most cases, the impact of including the capitalized operating leases is less dramatic in table 4. consider, for example, the top segment in table 4, retail trade–home furniture, furnishings and equipment stores. the firm with the largest change was williams-sonoma that showed an increase in its debt to capital ratio of 22.7 percentage points, compared with a 39.8 percentage point change in its debt to (book value) total assets, as shown in table 3. for 80% of the firms in top 1,000, their leverage as measured by the book value debt to total asset ratio is higher than when measured by the debt to total capital ratio that uses the market value of equity. however, for 200 of the firms the result is opposite with their debt to total capital ratio falling below their book value debt ratio. this occurs because of differences in how the metrics are calculated and differences in the use of operating liabilities across firms. for example, without including operating leases best buy’s debt ratio is 12.83 and its debt to total capital ratio is 14.20. best buy’s market value of equity, used in the debt to total capital ratio is $10.476 billion, $6.098 larger than its book value of equity. the firm’s total capital includes $1.734 billion of debt, making its total capital $12.210 billion. however, the firm’s total book value of assets is $13.519 billion resulting in a lower book value debt ratio. this result occurs because total assets include those financed both with capital (equity plus interest bearing debt) and with operating liabilities, $6.530 billion of working capital liabilities in best buy’s case. because the firm’s operating liabilities exceed the difference between its book and market values of equities, the debt to total capital ratio is less than its book value debt ratio. this result is also reflected in the values after including the firm’s operating leases, valued at $3.144 billion. its book value debt ratio is then 29.27 compared with a higher debt to capital ratio of 31.77. 6. implications for investors for individual investors, two things should be clear from the discussion so far. first, firms have been allowed to keep massive amounts of debt off-balance sheet through the use of operating leases. second, the changes that have been announced by fasb will correct this situation and make unadjusted leverage metrics based on the balance sheet much more representative of the true leverage of the firm. however, as noted, these changes will not appear until 2019 so, in the interim, individual investors should either learn how to adjust leverage measures for operating leases or rely on a data source, such as bloomberg, that does that for them. however, this pending change is likely to raise other questions. for example, will stock and bond prices be affected when this large amount of previously hidden debt comes onto balance sheets? will firms suddenly violate their debt covenants when high amounts of off-balance sheet debt appear on their balance sheet, pushing them into default? will estimates of future cash flows increase when income statements reveal an increase in 215j. trifts, g.e. porter / financial services review 26 (2017) 205–220 t ab le 4 c ha ng e in va lu e of de bt to to ta l ca pi ta l if pr op os ed ch an ge s w er e ap pl ie d to 20 15 fis ca l ye ar la rg es t 25 di vi si on s, by tw odi gi t si c co de s d iv is io n si c fi rm s d iv is io n av er ag e ch an ge l ar ge st ch an ge by di vi si on sm al le st ch an ge r et ai l t ra de –h om e fu rn itu re , fu rn is hi ng , & e qu ip m en t st or es 57 4 20 .0 % 22 .7 % w ill ia m sso no m a in c 17 .6 % b es t b uy c o in c r et ai l t ra de –a pp ar el & a cc es so ry st or es 56 11 18 .1 % 34 .0 % a m er ic an e ag le o ut fit te rs 3. 1% u nd er a rm ou r in cc la ss a r et ai l t ra de –f oo d st or es 54 5 16 .7 % 41 .3 % w ho le fo od s m ar ke t in c 5. 1% d un ki n’ b ra nd s g ro up in c t ra ns po rt & pu bl ic u til ity –t ra ns po rt at io n by a ir 45 10 13 .2 % 21 .4 % u ni te d c on tin en ta l h ol di ng s 0. 7% a lle gi an t t ra ve l c o m an uf ac tu ri ng –l ea th er an d l ea th er pr od uc ts 31 2 12 .8 % 16 .9 % sk ec he rs u sa in cc la ss a 8. 7% c oa ch in c r et ai l t ra de –e at in g & d ri nk in g pl ac es 58 14 12 .6 % 26 .3 % c he es ec ak e fa ct or y in c/ t he 2. 1% a ra m ar k a gr ic ul tu re –a gr ic ul tu ra l se rv ic es 07 1 12 .4 % 12 .4 % v c a in c 12 .4 % v c a in c r et ai l t ra de –m is ce lla ne ou s r et ai l 59 19 10 .7 % 39 .0 % d ic k’ s sp or tin g g oo ds in c 0. 3% e xp re ss sc ri pt s h ol di ng c o se rv ic es –s oc ia l se rv ic es 83 1 10 .4 % 10 .4 % b ri gh t h or iz on s fa m ily so lu tio ns 10 .4 % b ri gh t h or iz on s fa m ily so lu tio n se rv ic es –e du ca tio na l se rv ic es 82 2 9. 6% 13 .3 % g ra ha m h ol di ng s c oc la ss b 5. 9% h ou gh to n m if fli n h ar co ur t c o m an uf ac tu ri ng –a pp ar el , an d ot he r fi ni sh ed pr od uc ts m ad e fr om fa br ic an d si m ila r m at er ia ls 23 8 9. 5% 20 .6 % b ur lin gt on st or es in c 2. 2% h an es br an ds in c r et ai l t ra de –g en er al m er ch an di se st or es 53 12 8. 9% 26 .3 % b ig l ot s in c 0. 1% c as ey ’s g en er al st or es in c se rv ic es –m ot io n pi ct ur es 78 8 8. 5% 23 .6 % a m c e nt er ta in m en t h ld sc la ss a 0. 8% n et fli x in c c on st ru ct io n– h ea vy c on st ru ct io n ot he r th an b ui ld in g c on st ru ct io n c on tr ac to rs 16 2 6. 9% 11 .2 % ja co bs e ng in ee ri ng g ro up in c 2. 6% fl uo r c or p r et ai l t ra de –a ut om ot iv e d ea le rs an d g as ol in e se rv ic e st at io ns 55 10 6. 2% 16 .5 % a dv an ce a ut o pa rt s in c 0. 9% c ar m ax in c b ui ld in g m at er ia ls , h ar dw ar e, g ar de n su pp ly an d m ob ile h om e d ea le rs 52 5 5. 6% 14 .2 % t ra ct or su pp ly c om pa ny 2. 0% fa st en al c o se rv ic es –a m us em en t an d r ec re at io n se rv ic es 79 4 5. 1% 14 .8 % l iv e n at io n e nt er ta in m en t in 0. 8% c hu rc hi ll d ow ns in c c on st ru ct io n– sp ec ia l t ra de c on tr ac to rs 17 3 4. 5% 5. 8% e m co r g ro up in c 3. 3% c hi ca go b ri dg e & ir on c o n v fi n in s an d r ea l e st at e– in su ra nc e a ge nt s, b ro ke rs an d se rv ic e 64 5 4. 0% 4. 8% m ar sh & m cl en na n c os 2. 3% b ro w n & b ro w n in c w ho le sa le t ra de –d ur ab le g oo ds 50 20 3. 3% 6. 3% a ir ga s in c 0. 9% w w g ra in ge r in c m an uf ac tu ri ng –p et ro le um r efi ni ng an d r el at ed in du st ri es 29 14 3. 2% 7. 4% t es or o c or p 0. 5% m ar at ho n o il c or p se rv ic es –p er so na l se rv ic es 72 3 3. 0% 7. 6% h & r b lo ck in c 0. 7% se rv ic e c or p in te rn at io na l m an uf ac tu ri ng –p ri nt in g, pu bl is hi ng an d a lli ed in du st ri es 27 9 2. 9% 12 .3 % n ew s c or pc la ss a 0. 0% c im pr es s n v t ra ns po rt & pu bl ic u til ity –c om m un ic at io ns 48 30 2. 5% 11 .1 % u s c el lu la r c or p 0. 0% c b s c or pc la ss a v ot in g se rv ic es –h ea lth se rv ic es 80 16 2. 5% 7. 2% b ro ok da le se ni or l iv in g in c 1. 0% m ed na x in c 216 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 expenses related to these leases? fortunately, for investors, the answer to each of these questions is “probably not.” the primary reason for this optimism is that significant revaluations in stock prices occur in response to new information and while the impact of off-balance sheet leases may be news to some individual investors, it is not news to the professional investment community. furthermore, none of the information we have reported is “hidden.” as discussed earlier, that data are reported in footnotes and analysts have long been estimating the value of off-balance sheet debt and revising reported balance sheets with this information. as a result, metrics used to provide valuation estimates for the stocks of these firms already reflect this information. debt covenants may be affected by this new standard and in some cases companies and their lenders will have to renegotiate existing covenants to adjust for the newly reported on-balance sheet debt. in a 2011 survey by deloitte of 178 executives of firms with operating leases, 44% of respondents reported that the new standards would likely affect their companies’ existing debt covenants. they note that “this may lead to renegotiation of outstanding debt instruments, which could provide a potential opportunity to exact more concessions from lenders or borrowers, depending on the condition” (deloitte 2011). however, the impact of this is also likely to be small as paik et al., (2015) notes that lenders to companies with significant operating leases already tend to focus on income statement based coverage ratios rather than balance sheet ratios in their covenants. because the income statements of these companies will be essentially unchanged, the impact on these covenants should be small. overall, it is likely that the impact of this change on security prices will be small and insignificant. the market value of a company can be estimated in a number of ways. discounted cash flow models estimate the present value of future free cash flows at a discount rate that is appropriate for the firm’s business and financial risk. values can also be estimated based on various multiples, including sales and ebitda. the capitalization of operating leases will not affect the cash flows to the firm and professional investors have already been including the impact of operating leases in their estimations of leverage and cost of capital. while companies will now provide estimates of the present value of their operating leases that may be somewhat more accurate than those previously done by analysts based on footnote disclosures, there is no reason to think that these values will be consistently higher or lower and therefore any impact is likely to be small and random.6 7. conclusion this article discusses the new fasb standard that will require firms to capitalize their operating leases beginning in 2019, and the implications to investors. we provide a brief explanation of the new standard and how it will affect financial statements, and we reveal which of the largest 1,000 firms and industries will be impacted significantly by the change. we then ask, “what impact will this have on the stock and bond value of these firms?” in short, while firms that are taking advantage of the current standard, those in the retail trade being the heaviest users for example, will show dramatic increases in liabilities and expenses 217j. trifts, g.e. porter / financial services review 26 (2017) 205–220 as a result of moving their operating leases onto their balance sheets, we caution investors not to anticipate changes in their stock or bond valuations resulting from this change. asset values change in response to new information, and the information we present in this article regarding changes in total assets and debt ratios is currently available from data providers such as bloomberg, and is already being used by professionals to forecast asset values and future cash flows. notes 1 specifically, public companies will be required to meet the new standards beginning with financial statements for periods that begin after december 15, 2018. 2 lessees will be required “to apply a modified retrospective transition approach to each lease that existed at the beginning of the earliest comparative period presented in the financial statements, as well as leases entered into after that date.” (pricewwaterhousecoopers, 2016) consistent with accounting standard asc 250, companies will report the effect of the change on prior periods and the cumulative effect on balance sheet accounts. 3 see the appendix for an example from urban outfitters. 4 the criteria for determining a lease’s status is more subjective under the new standards. for example, a lease may be considered a finance lease if its term is for the “major part” of the assets remaining economic live or its present value exceeds “substantially all” of the fair value of the asset. (spiceland et al, 2017). 5 for the remaining 53 firms, we next searched value line for each firm’s “financial strength” rating and used this as a proxy for the firm’s debt rating. value line (2008) notes that their “financial strength ratings take into account a lot of the same information used by the major credit rating agencies. our analysis focuses on net income, cash flow, the amount of debt outstanding, and the outlook for profits, and the stability of the industry and the individual company returns. other factors also enter into the equation.” using this indicator as a proxy for debt credit ratings, we estimated the cost of debt for an additional 47 firms. 6 boatsman and dong (2011) and altamuro et al. (2014) provide evidence that offbalance sheet leases are priced by investors. appendix: the reporting and valuation of operating leases shown below is note 13 from urban outfitters annual report for the 12 months ending january 31, 2016. to estimate the present value, we assumed that after 2021, the remaining $772,226 will be paid at the same annual rate as in 2021. that is, the estimated lease payments in 2022, 2023 and 2024 will be $199,685 per year leaving a remainder of $173,171 in 2025. discounted at the firm’s estimated pretax cost of debt of 2.39%, the present value of the firm’s operating leases is estimated to be $1,780,113. 218 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 13. commitments and contingencies leases the company leases its stores, certain fulfillment and distribution facilities, and offices under non-cancelable operating leases. the following is a schedule by year of the future minimum lease payments for operating leases with original terms in excess of one year: fiscal year 2017 $272,255 2018 263,876 2019 246,043 2020 228,091 2021 199,685 thereafter 772,226 total minimum lease payments $1,982,176 amounts noted above include commitments for 37 executed leases for stores not opened as of january 31, 2016. the majority of our leases allow for renewal options between five and ten years upon expiration of the initial lease term. the store leases generally provide for payment of direct operating costs including real estate taxes. certain store leases provide for contingent rentals when sales exceed specified levels, in lieu of a fixed minimum rent, that are not reflected in the above table. additionally, the company has entered into store leases that require a percentage of total sales to be paid to landlords in lieu of minimum rent. rent expense consisted of the following: fiscal year ended january 31 2016 2015 2014 minimum and percentage rentals $245,474 $234,982 $205,759 contingent rentals 2,704 3,901 5,542 total $248,178 $238,883 $211,301 the company also has commitments for unfulfilled purchase orders for merchandise ordered from our vendors in the normal course of business, which are satisfied within twelve months, of $407,833. the majority of the company’s merchandise commitments are cancellable with no or limited recourse available to the vendor until the merchandise shipping date. the company also has commitments related to contracts with construction contractors, fully satisfied upon the completion of construction, which is typically within twelve months, of $1,535. references altamuro, j., johnston, r., pandit, s., & zhang, h. (2014). operating leases and credit assesments, contemporary accounting research, 31, 551–580. boatsman, j., & dong, x. (2010). equity value implications of lease accounting, accounting horizons, 25, 1–16. 219j. trifts, g.e. porter / financial services review 26 (2017) 205–220 beatty, a., liao, s., & weber, j. (2010). financial reporting quality, private information, monitoring, and the lease-versus-buy decision. the accounting review, 85, 1215–1238. bryan, s. h., lilien, s., & martin, d. r. (2010). the financial statement effects of capitalizing operating leases. the cpa journal, new york 80.8, pp. 36–41. chartered financial analysts institute. (2012). financial reporting and analysis. cfa program curriculum, 3, 652–659. cornaggia, k., franzen, l., & simin, t. (2013). bringing leased assets onto the balance sheet. journal of corporate finance, 22, 345–360. deloitte. (2011). a deloitte survey on the fasb’s proposed changes to lease changes to lease accounting standards. (available at http://www.iasplus.com/en/binary/usa/1102usleasessurvey.pdf). financial accounting standards board. (2016). leases (topic 842), fasb accounting standards update, no. 2016-02, february. graham, b., & dodd, d. l. (1934). security analysis: principles and techniques, mcgraw hill, new york, ny. katz, d. m. (2016). new fasb lease standard could inflate balance sheets. (available at cfo.com, http://ww2.cfo.com/accounting-tax/2016/02/new-fasb-lease-standard-inflate-balance-sheets/). lease accelerator. (2016). who is most impacted by the new lease accounting standards: an analysis of the fortune 500’s leasing obligations. (available at http://explore.leaseaccelerator.com/wp-content/uploads/ 2016/01/who-is-most-impacted-by-the-new-lease-accounting-standards-d.pdf). levitt, a., jr. (february 2004). reclaiming the profession’s heritage. transcript in the cpa journal, new york 74.2, pp. 22–27. pricewaterhousecoopers. (2016). leases. (available at http://www.pwc.com/us/en/cfodirect/assets/pdf/account ing-guides/pwc-lease-accounting-guide.pdf). paik, d. g. h., van der laan smith, j. a., lee, b. b., & yoon, s. w. (2015). the relation between accounting information in debt covenants and operating leases. accounting horizons, 29, 969. spiceland, j. d., nelson, m., & thomas, w. (2017). intermediate accounting (9th ed.). new york, ny: mcgraw-hill professional publishing. value line. (2008). a complete overview. new york, ny: the value line investment survey. wheeler spencer, a., & webb, t. z. (2015). leases: a review of contemporary academic literature relating to lessees. accounting horizons, 29, 997–1023. 220 j. trifts, g.e. porter / financial services review 26 (2017) 205–220 from the editor this issue contains issue 4 of volume 26 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “does financial risk tolerance change over time? a test of the role macroeconomic, biopsychosocial and environmental, and social support factors play in shaping changes in risk attitudes” is coauthored by stephen kuzniak and john e. grable, both at university of georgia. in this paper, the authors address the need that financial planners, as well as regulators, require evidence documenting to what extent risk tolerance changes over time, and if changes do occur, the variables associated with variability. based on a model that included macroeconomic indicators, biopsychosocial and environmental factors, and measures of social support, they find that risk-tolerance attitudes are remain generally stable over time. additionally, there are groups of test takers that exhibit significant shifts in risk tolerance. they also describe some of the variables associated with these score changes, as well as provide financial planning professionals with guidance on how to identify clients who may be prone to shifting their tolerance for financial risk. the second article “which measures predict risk taking in a multi-stage controlled investment decision process?” is coauthored by kremena bachmann, thorsten hens, and remo stössel, all at the university of zurich. the authors assess the ability of different risk profiling measures to predict risk taking along a multi-stage process that reflects individuals’ willingness to take risks. they find that the individual willingness to take risks varies along the process, but its level is always related to a composite measure of the individual risk tolerance. assessment of the risk tolerance cannot be substituted by a simulated experience, although the latter can improve the perception of the risk and reward potential of the investment and motivate higher risk taking. the risk tolerance measure addresses different notions of risk, but they found that individual loss aversion is the most powerful predictor of risk taking at all stages of the discovery process. by contrast, they found that neither the self-assessed risk tolerance measures financial services review 26 (2017) v–vii 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. nor investment experience are suitable for consistently predicting risk taking at different stages of the process. the third article, “evaluating the relationship between ifa remuneration and advice quality: an empirical study” is coauthored by jiřı́ sindelar and petr budinsky, both at the university of finance and administration prague. the authors investigate the interaction between commission remuneration of independent financial advisers and selected sales factors, including the quality of advice. utilizing data on investment transactions and a linear model with mixed effects, they found that the link between commission and quality of the subsequent recommendation is not homogeneous, and advice-bias potential is present only in a limited range of organizational environments, connected mainly to the flat-structure business model. alternatively, they found that arbitrage between different product classes creates a biasing potential across almost all types of firms, creating potential for market systemic risk. the fourth article, “portfolio insurance using leveraged etfs” is coauthored by jeffrey george and william j. trainor jr., both at east tennessee state university. the authors examine the use of leveraged exchange traded funds (letfs) within a constant proportional portfolio insurance (cppi) strategy. they state that the advantage of using letfs in such a strategy is that it allows a greater percentage of the portfolio to be invested in the risk-free rate relative to a traditional cppi. they indicate that where a standard cppi strategy may require 50% of the portfolio to be invested in equities, using a 2x letf only requires 25%, and a 3x letf only requires 16.7% to attain the same effective exposure to equities. their results show that when the risk-free asset is yielding at least 3% or the 1 year minus 90-day treasury exceeds 1%, the use of letfs within a cppi framework results in annual returns approximately 1–2% higher with better sharpe, sortino, omega, and cumulative prospect values, while reducing value at risk (var) and excess shortfall (es) below var. the final article, “who seeks financial advice?” is coauthored by maher h. alyousif and charlene m. kalenkoski, both at texas tech university. the authors examine the determinants of seeking five types of financial advice and find consistency across different types of advice. additionally, they observe no significant differences among subsamples defined by gender, age, and financial literacy. they show that income and risk tolerance are related positively to the demand for financial advice and affect the probability of seeking advice more than other variables. they also indicate that a low perception of financial knowledge, which can be a proxy for self-confidence, and financial fragility decrease the probability of seeking financial advice. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while vi editorial / financial services review 26 (2017) v–vii fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review aamer sheikh quinnipiac university kc ma stetson university abigail b. sussman university of chicago kenneth n. ryack quinnipiac university andy terry university of arkansas at little rock kenneth white university of georgia angela wheeler spencer oklahoma state university kevin davis us air force academy ann marie hibbert west virginia university kyoung tae kim university of alabama anthony loviscek shu larry r. frank better financial education aron gottesman pace university leonard lundstrum northern illinois university brian betker saint louis university manoj athavale ball state university brian starr lubbrock christian university marc oliver rieger university of trier catherine montalto ohio state university marianna brunetti university of roma tor vergata chris robinson york university marie lachance south dakota state university christine harrington auburn university mark egan university of minnesota carlsom christine mcclatchey university of northern colorado mark k. pyles college of charleston christopher ma stetson university martin c. seay kansas state university chungwen hsu university of vermont mei wang whu-otto beisheim school of mgmt colleen asaad baldwin wallace university melinda morrill nc state cris delatorre university of northern colorado michael fuerst university of miami dale domain york university michael s. gutter university of florida darrol stanley pepperdine university nancy mohan university of dayton david hunter hawaii nilton porto university of rhode island david m blanchett morningstar investments mgt ning tang sd state university david nanigian csu fullerton olivia s. mitchell wharton school demissew diro ejara university of new haven patti fisher virginia tech derek r. lawson kansas state university patti fisher virginia tech diane reyniers london school of economcics paul j. haensly university of texas drew peabody university of north texas peter miu mcmaster university george pennacchi university of illinois reinhold lamb university of north florida giovanni fernandez stetson university riccardo calcagno emlyon business school gowri shankar university of washington richard evans university of virginia grady perdue university of houston clear lake richard toolson washington state university hanna lim oklahoma state university sally mckechnie university of nottingham huy lam shultz collins inc scott j. boylan washington and lee university ivo claev university college of london sean grover bam advisors j. michael collins university of wisconsin shawn brayman planplus jaclyn j. beierlein east carolina university sherman hanna ohio state university james dilellio pepperdine university sonya britt kansas state university jane terpstra-tong monash university stephan whitaker federal reserve board cleveland jerry stevens university of richmond steve todd loyola jessica west stetson university swarn chatterjee university of georgia jim musumeci bentley university tahira hira iowa state university jing j. xiao university of rhode island terrance martin jr university of texas pan am john a. haslem university of maryland tomas dvorak union college john salter texas tech university tracey west griffith university john salter texas tech university travis jones florida gulf coast university kathleen m. rehl rehl wealth mgt wade pfau the american college william jennings us air force academy viieditorial / financial services review 26 (2017) v–vii the performance and market timing ability of chinese mutual funds wei hea, bolong caob, h. kent bakerc,* amississippi state university, mississippi state, department of finance and economics, 310 mccool hall, mississippi state, ms 39762, usa bohio university, ohio university, department of economics, 331 bentley hall annex, athens, oh 45701, usa camerican university, kogod school of business, department of finance and real estate, 4400 massachusetts avenue, nw, washington, dc 20016, usa abstract we examine the performance and market timing ability of actively managed chinese stock mutual funds and investigate how fund characteristics and fund flows relate to performance and market timing ability. based on daily return data and several four-factor models, only about 7.5% of these funds have statistically significant risk-adjusted abnormal returns and even fewer demonstrate market timing ability. after controlling for fund size, management fees, average amount, and volatility of fund flows, older funds show higher sharpe ratios. our evidence also reveals the volatility of fund flows has an inverted-u shape relationship with fund performance. © 2015 academy of financial services. all rights reserved. jel classification: g11; g23 keywords: chinese mutual funds; sharpe ratio; treynor ratio; alpha; market timing 1. introduction although the chinese mutual fund industry started in 1991 with the establishment of the first closed-end fund, it became stagnant because of lack of product diversity and investor interest. the lack of professional management and insufficient supervision from the regu* corresponding author. tel.: �1-202-885-1949; fax: �1-202-885-1949. e-mail address: kbaker@american.edu (h. k. baker) financial services review 24 (2015) 289–311 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. lators also contributed to the sluggish performance. in 1997, china introduced the interim regulations on the securities investment funds, which revitalized the industry’s growth by providing a new framework for developing the investment management industry. after china joined the world trade organization in 2001, open-end mutual funds started to thrive as foreign investment management firms brought their practices and expertise into the chinese capital market and setup joint ventures with chinese firms. since then, the chinese mutual fund industry has become the biggest institutional investor in one of the world’s largest emerging markets. currently, china’s mutual fund industry is only about one-tenth the size of the u.s. market but is likely to grow because of its potential to attract foreign and domestic investors. many chinese mutual funds, such as the china index fund (fxi), provide a way to invest in china and potentially the ability to ride china’s growth wave. the possible appreciation of the chinese yuan against the u.s. dollar makes investments in chinese mutual funds even more attractive to international investors. meanwhile, if the chinese middle class continues to expand and increasingly invest in stock through mutual funds, the chinese mutual fund market has the potential to become one of the biggest in the world. much of the previous research on chinese mutual fund performance and market timing skills is written in chinese. for example, shen and huang (2001) and zhang and du (2002) consider only the market factor and a very limited number of closed-end funds. li and ma (2004) and zou and lin (2004) expand the fama-french three-factor model but their sample size remains small because of the short history in the early stage of the chinese mutual fund industry. guo (2010) has a much larger sample than the previous studies but the regression model is limited to the market factor. extending the sample period to the first half of 2010, tang et al. (2012) find that fund size and performance exhibits an inverted-u shape relationship. however, their performance measures are limited to the capital asset pricing model (capm) alpha, fama-french three-factor alpha, and style benchmark-adjusted return, but the last measure does not show the inverted-u shape relationship between size and performance. understanding the behavior and characteristics of chinese mutual funds can supply crucial knowledge of the financial development dynamics in emerging markets. to facilitate our understanding of this industry, we provide baseline empirical facts on the performance of actively managed chinese stock mutual funds. in this study, we investigate the performance and market timing ability of these funds using an updated sample extracted from the csmar china funds market research database–open-end funds. we also examine how some basic fund characteristics and fund flows influence the performance of these mutual funds. we use a standard four-factor model to evaluate the daily performance of chinese mutual funds. we also provide and analyze the sharpe ratio, treynor ratio, alphas, and market timing skill coefficients in our market timing models between 2001 and 2011. we further examine how some fund characteristics, such as fund size, management fees, fund age, together with the average fund flows and volatility of fund flows, relate to the performance of the actively managed chinese stock mutual funds. we then split our full sample into two sub-periods to check the robustness of our results on these relationships. we apply the robust standard error formula hc2 in mackinnon and white (1985) in all of our regressions to 290 w. he et al. / financial services review 24 (2015) 289–311 control for heteroskedasticity. using daily mutual fund return data between 2001 and 2011, we find that only about 7.5% of active chinese mutual funds have statistically significant positive �s. this finding is similar to the results of eling and faust (2010) on emerging market mutual funds. furthermore, older funds tend to have higher sharpe ratios. our results also confirm a positive correlation between average fund flows and fund performance and the volatility of fund flows shows an inverted-u shape relationship with performance. our study differs from previous studies of chinese mutual funds in several ways. first, to our knowledge, we are the first to use a four-factor model to analyze the performance of actively managed chinese stock mutual funds. our model leads to a more accurate calculation of mutual fund abnormal returns. second, our sample is larger, covers a longer time period, and includes more recent data (until the end of 2011) than previous studies of chinese mutual funds. third, we focus on daily returns whereas previous studies use weekly returns. bollen and busse (2001) show that using daily returns to examine the market timing skill is more appropriate. these enhancements result in a more comprehensive and refined study with more robust findings. we also pay particular attention to fund flows by incorporating both their mean and standard deviation when examining their relationships to performance jointly with other fund characteristics. our study contributes to the literature on mutual funds in several ways. first, we establish a performance profile of the chinese mutual fund industry in its early years, which can help researchers interested in examining the evolution of this industry together with the development of china’s capital market. on average, the more than 300 actively managed stock mutual funds in our sample period fail to beat the market after the fees. the average risk-adjusted return (� in the four-factor model) is very close to zero. our results show that only about 7.5% of the funds produce positive and statistically significant �s. this inability to generate superior net-of-fees returns or risk-adjusted returns is consistent with the literature on mutual fund performance in both developed and emerging markets (eling and faust, 2010; fama and french, 2010; french, 2008). furthermore, we find that less than 5% of our funds demonstrate market timing ability with statistical significance, which is similar to the 6.6% in cao and jayasuriya (2012) for emerging market hedge funds and the 2% in fung et al. (2002) for global hedge funds. this result, however, is in stark contrast with the findings on managed portfolios in the united states. for example, bollen and busse (2001) show that more than 40% of mutual funds of their actively managed stock mutual funds in the united states (u.s.) demonstrate statistically significant market timing skill. these findings indicate that actively managed chinese stock mutual funds have a long way to go to catch up with their developed world counterparts. secondly, the relationship among fund characteristics, fund flows, and performance contributes to our understanding of how the mutual fund industry behaves both in general and in a large emerging economy in particular. in our sample, chinese funds established earlier produce higher sharpe ratios and four-factor �s, which is consistent with the concept of the learning effect in bauer et al. (2005). the positive relationship between average fund flows and performance is consistent with performance chasing behavior considered in berk 291w. he et al. / financial services review 24 (2015) 289–311 and green (2004) and the empirical findings in rakowski and wang (2009) and rakowski (2010). besides smart money chasing good performance, we also believe that the redemptions from poor performing funds after the collapse of the chinese stock market in 2008 contribute to this result. the inverted-u relationship between volatility of fund flows and performance indicates that both very high and very low volatility in fund flows are related to poor performance. for high flow volatility, we find the same performance dragging effect as in rakowski (2010). as we discuss in more details later, the low volatility of fund flows in our sample may result from steady outflows from the poor performing funds, which is another consequence of the 2008 collapse. rakowski’s sample ends in 2006. in contrast, our sample ends in 2011 so our results add to our understanding in this issue by looking at post financial crisis fund behavior. third, our findings have important implications for the investment decisions made by both individual and institutional investors. our evidence on the current performance profile of the actively managed chinese stock mutual funds indicates that investors should focus on index funds in china unless they have solid evidence that some active managers have the ability to consistently generate excess risk-adjusted returns. investors interested in these actively managed funds need to learn how to identify the different sources of returns, such as pure luck, rewards for various risk exposures, and genuine skills. on the other hand, chinese active fund managers could improve their skills by learning from their international and domestic predecessors when developing their investment strategies. finally, the regulators of the chinese mutual fund industry can set higher professional conduct standards to ensure quality work and establish better disclosure requirements to educate the investing public about the sources of returns. the article proceeds as follows. section 2 discusses the development of the chinese mutual fund industry and provides a broad context for our study. section 3 examines performance and market timing measures. section 4 describes the data and sample selection. section 5 reports the empirical findings and section 6 offers a summary and conclusions. 2. the chinese mutual fund market the development of the chinese mutual fund industry consists of three stages: (1) the exploration stage (1991–1997), (2) the experimental and learning stage (1998–2004), and (3) the growth stage (post 2004). during the first stage, the zhuxin fund emerged as the first closed-end fund in august 1991 shortly after the establishment of chinese stock market. zhuhai international trust and investment corporation sponsored the zhuxin fund. the wuhan securities investment fund and the shenzhen nanshan venture capital fund started in october 1991. by 1992, china had 37 closed-end funds. several factors hindered further development of the closed-end fund market including the lack of professional management, illiquidity, fund product homogeneity, and the lack of systematic supervision of the fund managers’ behavior. a breakthrough occurred in 1997 when the china securities regulatory commission issued interim regulations on the securities investment funds. this document provided a framework for promoting the growth of investment funds. the end of the closed-end fund era 292 w. he et al. / financial services review 24 (2015) 289–311 occurred in 1997 with a transition toward open-end mutual funds. over the last decade, open-end mutual funds gradually replaced closed-end funds and became the primary and largest type of fund investment. the china securities regulatory commission has offered extensive support to the industry by intensively monitoring and implementing regulations that are similar to those of the u.s. securities and exchange commission (sec). the next stage in the development of the chinese mutual fund industry occurred between 1998 and 2004. after china joined the world trade organization in 2001, the chinese government gradually opened up access to china’s capital markets to foreign financial firms. since then chinese open-end mutual funds have actively sought to collaborate with successful investment management firms in the developed markets to develop innovative fund products and improve management skills. the forms of collaboration range from technical support to joint ventures. the management teams of these new funds include many chinese portfolio managers who returned to china with their experiences from the world’s leading investment management firms. with the support from jpmorgan fleming asset management, the first open-end mutual fund called huaan innovations fund started in september 2001. this event is a milestone in the chinese mutual fund history. by november 2002, china had 17 mutual funds with an asset value approaching 56.4 billion yuan. meanwhile, the regulatory authority in china realized the importance of regulating the mutual fund industry. china’s national legislature, the national people’s congress, issued the law of the people’s republic of china’s securities investment funds on october 28, 2003 and it went into effect on june 1, 2004. this law established a formal legal framework for china’s investment fund industry. as a result, nanfang progressive allocation fund became the first listed mutual fund in october 2004. china’s first exchange-traded fund (etf), the sse 50 etf, started in late 2004. as chinese investors recognized the benefits of etfs as a form of low cost indexing alternative, the etf sector continued to grow. several events characterize the third development phase of the chinese mutual fund industry, which started in 2005. to better monitor the excessive risk-taking behaviors of fund managers and to motivate the funds to diversify their portfolio in the international market, the china securities regulatory commission, people’s bank of china, and state administration of foreign exchange jointly issued measures on admission of domestic securities investments of qualified domestic institutional investor (qdii) in 2006. in 2007, the asset value of the investment fund industry reached 3.28 trillion yuan. starting in 2008, the chinese mutual fund industry has undergone dramatic structural changes as evidenced by the emergence of different investment philosophies, fund structures, and behaviors. fund managers started seeking both returns and fund flows and incorporating international bonds and stocks into their portfolio. consequently, these funds began to attract more institutional investors. in 2008, china and the united states agreed to allow chinese citizens to invest in the u.s. stock market through mutual fund organizations or other asset fund companies in china. for the first time, domestic chinese citizens could invest in the u.s. stock market (rodier, 2009). like most major economies, the global financial crisis of 2007–2009 adversely affected the chinese economy. for example, the asset value of the chinese mutual fund industry decreased to 2.21 trillion yuan (equivalent to $351 billion) on march 31, 2012 from its peak 293w. he et al. / financial services review 24 (2015) 289–311 of around 3.28 trillion yuan (equivalent to $448 billion) on december 13, 2007. by the end of 2012, 1241 mutual funds had a total of 2.865 trillion chinese yuan (around $460 billion) of assets under management (aum) in the chinese mutual fund industry. a diverse group of both individual and institutional investors contributed to the rapid growth of the chinese mutual fund industry through active participation in various mutual fund families. by the end of 2006, the number of mutual fund shareholders reached 14.27 million with a growth rate of 166% compared with 2005. over the next several years, the desire to meet investor needs led to creating many fund forms such as balanced funds, index funds, and social security funds. these new varieties are mainly responsible for attracting the new investors. by june 2011, the chinese mutual fund industry had 91.6 million shareholders. chinese mutual funds may become attractive for international investors as they seek opportunities in the emerging markets and international diversification if the following events occur. first, the chinese economy must continue to grow. second, maturing chinese capital markets must bring higher efficiency, more transparency, and better regulation. finally, the chinese yuan needs to maintain a steady exchange rate. another factor triggering the growth of mutual funds in china is its unique investor clientele. chinese investors, unlike investors in many western countries, have a long history of saving for future uncertainties. the chinese are accustomed to saving 30% to 40% of their disposable income, which is much higher than the savings rate in many western countries. the chinese are responsible for making their retirement plans. given the low yields on bank accounts, mutual funds appear to be attractive for small investors. the chinese government is promoting private retirement planning, leading to potential growth in the mutual fund industry to accommodate the needs of investors for pension funds and special purpose funds. if the chinese middle class continues to expand and increasingly invest in mutual funds, this creates an opportunity for mutual funds to launch more innovative products to meet investor demand. however, the chinese mutual fund industry is relatively young. our results show that the mutual fund managers in china need to improve their portfolio management skills to draw closer to their peers. 3. issues about fund performance and market timing 3.1. performance and market timing measures we use conventional measures of portfolio performance and market timing to study chinese mutual funds. then we examine how fund characteristics and fund flows might influence a fund’s performance and market timing ability. our study has three major objectives. the first objective is to determine whether active chinese stock mutual funds can generate superior risk-adjusted returns. three widely used risk-adjusted performance appraisal measures are: (1) the sharpe ratio (also known as rewardto-variability), (2) the treynor ratio (also known as reward-to-volatility or excess return to non-diversifiable risk), and (3) ex post � (also known as jensen’s �) (maginn et al., 2007). the sharpe ratio (sharpe, 1966) has become an industry standard in measuring risk294 w. he et al. / financial services review 24 (2015) 289–311 adjusted performance. the sharpe ratio compares excess returns to the fund’s total risk as measured by its standard deviation of excess returns. traditionally, the ex post sharpe ratio is given by: sp � r� p � r� f �̂p (1) where r� p is the average fund return; r�f is the average risk-free rate; and �̂p is the standard deviation of the excess return �rp,t � rf,t�. in the case of the sharpe ratio, the benchmark is based on the ex post capital market line (cml). a skillful fund manager will produce returns that place the fund above the cml. in contrast, the treynor measure (treynor, 1965) relates a fund’s excess returns to the fund’s systematic risk (beta or �). the calculation of the treynor ratio is given as: tp � r� p � r� f �̂p (2) where r� p and r�f are the average values of each variable over the evaluation period and �̂p is the fund’s �. the treynor ratio of the market portfolio is the slope of the security market line (sml). thus, a skillful manager will produce a treynor ratio greater than that of the ex post sml. researchers use various risk factors to analyze the performance of mutual funds. building upon the work of fama and french (1993) where size (smb for small minus big) and value or book-to-market equity (hml for high minus low) factors are first introduced, carhart (1997) uses an additional momentum factor (mom) besides the market, smb, and hml factors to study the performance persistence of u.s. mutual funds. researchers widely follow this methodology. our standard four-factor model takes the following form: rpt � rft � �p � �p �rmt � rft� � �ssmbt � �vhmlt � �mmomt � �pt (3) where for period t, rpt is the fund’s return; rft is the risk-free rate; rmt is the return on the market index; and smbt, hmlt, and momt are returns on value-weighted, zero-investment, factor-mimicking portfolios for size, book-to-market equity, and momentum, respectively. the term �p is the intercept of the regression; the �s are the fund’s sensitivity to different risk factors; and �pt is the random error term. the sign and value of the ex post � (�p) indicates the ability of the manager to generate abnormal risk-adjusted returns. our second objective is to investigate the market timing abilities of fund managers. in terms of total returns instead of relative returns, a fund’s performance comes from three sources: (1) decisions involving the strategic allocation, (2) market timing (i.e., returns attributable to shorter-term tactical deviations from the strategic asset allocation), and (3) security selection (i.e., skill in selecting individual securities within an asset class) (maginn et al., 2007). if fund managers have market timing ability, they should increase portfolio exposure to the market before the market advances and reduce market exposure before the market declines. earlier literature on the market timing skills finds little evidence of fund managers possessing this skill (elton et al., 1993; henriksson, 1984; jensen, 1969; treynor 295w. he et al. / financial services review 24 (2015) 289–311 and mazuy, 1966). more recently, using daily data, bollen and busse (2001) and chance and hemler (2001) find evidence of market timing for a substantial number of funds in the united states. based on the standard four-factor model, we extend the treynor and mazuy (1966) (hereafter referred to as tm) approach in detecting market timing ability using the following regression equation. rp,t � �p � �prm,t � �prm,t 2 � f�smbt, hmlt, momt� � �pt (4) where rp,t is the excess return on a portfolio at time t; rm,t is the excess return on the market; f�smbt, hmlt, momt� is the linear combination of these risk factors as expressed in eq. (3); and �p measures market timing ability. the coefficient on the quadratic term of fund returns, �p should be positive if mutual fund managers exhibit the market timing ability by adjusting their portfolio’s market exposures before the market swings to capture the upside and avoid the downside. henriksson and merton (1981) (hereafter referred to as hm) propose another approach of modeling market timing. we also examine their timing coefficient in our four-factor model as eq. (5) shows: rp,t � �p � �prm,t � �pitrm,t � f�smbt, hmlt, momt� � �pt (5) where it equals one if the market’s excess return and rm,t is positive and zero otherwise. the hm regression allows for the � risk to be different in ex post up and down markets. the term �p measures fund managers’ ability to time the market by altering the portfolio’s �. researchers often consider the �s in the tm and hm models as evidence of security selection skill if the estimated value is positive and statistically significant. 3.2. fund characteristics, fund flows, and fund performance the third objective of this study is to examine how the characteristics and flows of actively managed chinese stock mutual funds explain fund performance and market timing ability. many studies attempt to explain how fund characteristics such as fund size, age, expense ratios or management fees, and investment style help to explain fund performance. with some exceptions, the previous literature generally supports an inverse relationship between scale and fund returns. according to the “liquidity hypothesis,” fund size erodes performance because of the higher trading costs associated with illiquid stocks. large funds are not as flexible as small funds in divesting illiquid stocks (perold and salomon, 1991). chen et al. (2004) reinforce the inverse relationship between fund size and returns for various performance benchmarks and attribute the adverse scale effects to lack of liquidity and organizational diseconomies. in contrast, grinblatt and titman (1989) find mixed evidence that fund returns decline with fund size. otten and bams (2002) also report a positive relationship between size and fund abnormal performance for european mutual funds. because the chinese mutual fund industry only has a short history, their sizes are potentially relatively small to generate a price impact on illiquid stocks. furthermore, unlike manufacturing corporations, at the fund level, the size of the fund management team does not 296 w. he et al. / financial services review 24 (2015) 289–311 have to increase together with the size of aum, especially when the funds are relatively small. the fund management company can also provide basic infrastructure and services to its mutual funds so as to gain economies of scale that can offset the organizational diseconomies. consequently, we do not expect that fund size has much influence on the mutual funds in our sample. another fund characteristic that draws researchers’ attention is the fund’s age. fund performance may improve over time as the fund managers accumulate more experience in managing their portfolios and operating their funds. bauer et al. (2005) hypothesize the presence of a learning effect for fund managers. they also recognize the high startup costs associated with newly launched funds, which potentially offsets the advantages of organizational simplicity. thus, we expect older funds to outperform younger funds. researchers such as sharpe (1966), golec (1996), droms and walker (1996), carhart (1997), and jan and hung (2003) find that expense ratios or management fees are the largest component of expenses and reduce fund performance. haslem et al. (2008) also find that u.s. mutual funds with low expense ratios outperform those with higher expense ratios. however, elton et al. (1996) find that expense ratios are virtually the same for all mutual funds in different deciles and are only slightly responsible for the differences in performance between high ranked and low ranked funds in general. researchers also recognize that fund flows can influence the performance of mutual funds. berk and green (2004) provide a theory suggesting that fund inflows can erode the performance of the mutual funds. furthermore, some empirical evidence supports this negative relationship (frazzini and lamont, 2008; friesen and sapp, 2007). in a study on daily mutual fund flows, rakowski and wang (2009) find that past flows have a positive impact on future returns and an information effect drives this relationship. furthermore, rakowski (2010) shows that volatility of fund flows can hurt the performance of the mutual funds as dramatic changes in the flow pattern can force the fund manager to engage in costly trading. 4. data and methodology we obtain the sample from the csmar china funds market research database–openend funds. for the time period ending on december 31, 2011, this database contains 1,005 open-end funds including 529 stock mutual funds, which consist of 433 contractual mutual funds, 55 listed open-end funds, and 41 etfs. we focus on stock funds because no chinese bond indices have a sufficiently long history to be useful for our analysis. furthermore, to select the active funds, we drop index funds (164 including 40 etfs), an etf following an aggressive growth index (1), balanced funds (8), and new funds (9) that started in 2012 with insufficient performance records in the csmar database. thus, of the 529 stock mutual funds, we have 347 actively managed stock mutual funds in our sample between 2001 and 2011. to ensure the reliability of the performance measures estimated, we require a fund to have at least 100 daily returns on record to be included in our sample. this filter results in excluding 39 of the 347 stock funds. we first summarize some basic information on our 347 297w. he et al. / financial services review 24 (2015) 289–311 funds in tables 1, 2, and 3. then we analyze the performances of the 308 funds with sufficient return history and examine how the fund characteristics and flows are related to their performances and skill coefficients. table 1 contains the self-reported investment styles and the organizational forms of the 347 active chinese stock mutual funds. most of these funds (317 of 347) are contractual open-end funds, which are open-end investment trusts. a contractual fund is set up as an agreement among fund managers, the fund trustee, and investors specifying the rights and obligations of the three parties. investors can purchase or redeem the shares of these funds at most chinese commercial banks, which provide over-the-counter (otc) transaction services, at their net asset value (nav) plus some fees. to some extent, such funds are closer in nature to u.s. closed-end funds except that these funds continuously offer shares to table 1 actively managed chinese stock mutual funds: investment style and fund type investment style fund type total contractual open-end funds listed open-end funds active 1 0 1 aggressive growth 27 8 35 appreciation 35 4 39 enhanced index 0 1 1 growth 68 6 74 income 19 0 19 stable appreciation 3 0 3 stable growth 126 8 134 stable value-added 2 0 2 value 36 3 39 total 317 30 347 this table reports the number of actively managed chinese stock mutual funds classified by investment styles and fund type between 2001 and 2011. table 2 number of new actively managed chinese stock mutual funds between 2001 to 2011 year n % cumulative % total assets (billions of rmb) gta a-share index return (%) 2001 2 0.58 0.58 5.06 �24.66 2002 2 0.58 1.15 10.52 �19.80 2003 12 3.46 4.61 20.71 �4.02 2004 12 3.46 8.07 46.73 �17.61 2005 21 6.05 14.12 51.66 �11.75 2006 39 11.24 25.36 211.64 133.63 2007 39 11.24 36.60 1708.76 182.03 2008 35 10.09 46.69 751.14 �65.23 2009 47 13.54 60.23 1174.14 106.36 2010 66 19.02 79.25 1103.10 �8.87 2011 72 20.75 100.00 851.55 �23.82 total 347 100.00 this table is a summary of the number of new actively managed chinese stock mutual funds between 2001 and 2011, total fund size, and annual market returns. not all funds have total assets recorded in the gta china funds database. 298 w. he et al. / financial services review 24 (2015) 289–311 investors upon demand. a listed open-end fund (lof), which is a special form of chinese open-end fund, can be traded in a stock exchange like a closed-end fund. investors can either invest in the funds through the otc market provided by chinese commercial banks or trade their shares on an exchange. through a transfer process, investors can convert their nontradable shares of a mutual fund into tradable shares on an exchange. table 1 also shows that most funds are concentrated in six major investment styles: aggressive growth, appreciation, growth, income, stable growth, and value. each of these styles has more than 10 funds under its respective category. table 2 shows the number of funds established each year, aum for the 347 funds in our sample, and annual market index return between 2001 and 2011. only 49 funds or 14% of our sample funds started before 2006. by january 2006, the chinese mutual funds market had entered into a growth stage for more than a year and was ready to take off under an established operating environment. as table 2 shows, the aum for the sample funds increased from 51.66 billion of rmb in 2005 to 211.64 billion in 2006, together with the substantial increase in the market price level as seen from the 133.63% gta a-share index return in 2006. this boom ended in 2008 as the global financial crisis spread from the developed markets to emerging markets. shortly after a limited rebound in 2009, the chinese stock market entered into a sideways mode with a moderate downward trend. thus, the period between 2006 and 2011 provides a rich market environment to study how fund characteristics and flows are related to fund performance. furthermore, the 2006–2008 and 2009–2011 sub-periods can provide a contrast of fund behavior in different market cycles. for the above reasons, we choose the period between 2006 and 2011 to study the relationship between fund characteristics, fund flows, and fund performance. despite lackluster stock table 3 descriptive statistics for daily returns from october 19, 2001 to december 31, 2011 strategy n mean (%) median (%) min (%) 1% 99% max (%) sd (%) panel a: mutual fund returns aggressive growth 35,525 0.039 0.057 �14.401 �3.584 3.312 13.280 1.225 appreciation 34,604 0.025 0.057 �13.163 �3.922 3.562 15.112 1.376 growth 59,253 0.031 0.060 �8.987 �3.667 3.223 8.705 1.271 stable growth 70,777 0.005 0.030 �11.647 �3.836 3.368 12.603 1.349 value 44,243 0.057 0.060 �9.131 �3.452 3.188 8.481 1.170 others 20,211 0.031 0.045 �7.761 �3.385 3.049 7.468 1.170 all strategies 264,613 0.029 0.052 �14.401 �3.695 3.306 15.112 1.277 panel b: market return and the risk-free rate gta a-share index 2379 0.032 0.091 �9.125 �5.611 4.945 9.917 1.894 risk-free rate 2379 0.007 0.006 0.005 0.005 0.011 0.011 0.002 panel a of this table reports the summary statistics of daily returns of actively managed chinese stock mutual funds by investment style: aggressive growth, appreciation, growth, stable growth, value, and others. panel b shows the market return and the risk-free rate. the gta a-share index is a value-weighted index using the market capitalization calculated with the outstanding negotiable a shares and the closing price. the proxy for the chinese market risk-free rate is a one-year large denomination time deposit rate. all returns are in percentages. 299w. he et al. / financial services review 24 (2015) 289–311 market performance after 2008, the number of actively managed chinese stock mutual funds keeps growing and their aum remains much higher than before 2006. to guard against possible data errors on mutual fund returns in the csmar database, we clean the mutual fund return data in the following ways. first, we drop fund returns showing no activity, (i.e., the fund nav remains unchanged at 1 yuan within the first 20 days of the fund performance record). second, we identify erroneous “reverse” values in nav records (i.e., consecutive extreme values (i.e., � 99.9 percentile or � 0.1 percentile) with opposite signs in either unit nav returns or accumulated nav returns). we eliminate returns related to this kind of error from our sample. we use accumulated nav returns with share splits and cash dividends considered as our mutual fund returns as long as they are not extreme (i.e., outside of the 0.002% to 99.998% range). the erroneous or extreme values are set as missing instead of being winsorized to ensure the reliability of our results. this cleaning procedure gives missing values to about 0.037% of all daily unit nav returns in the csmar database regardless of fund types. table 3 provides a summary of descriptive statistics for the cleaned daily returns for different investment styles, the gta a-share index, and the risk-free rate in china. these daily returns are net of fee returns, which are the returns on the nav per share of the funds fully adjusted for share split/consolidation and cash dividends. as panel a shows, the value and aggressive growth styles have the highest mean daily returns at 0.057% and 0.039%, respectively, whereas the stable growth style has the lowest mean daily return at 0.005%. from both the standard deviation and the range between the 1st percentile and 99th percentile, we observe the appreciation followed by the aggressive growth styles have the most volatile return profiles. as panel b of table 3 shows, we use the market index for all negotiable gta a-shares provided by the csmar database as the market benchmark. gta a-share stocks are stocks listed on the chinese exchanges that are denominated and traded in chinese yuan. the negotiable shares of a company are the shares that can be traded in the secondary market. for some chinese firms, especially those state-owned or subsidiaries of state-owned companies, part of their common equities cannot be traded in the secondary market without the approval of governmental authorities. the state usually owns these non-negotiable shares. we use the current-value-weighted daily aggregated market returns with cash dividends reinvested of all the negotiable gta a-shares as our benchmark index. the current value refers to the market capitalization calculated as the product of the negotiable shares outstanding and the firm’s closing stock price. this index includes the stocks on both the shanghai stock exchange and shenzhen stock exchange. our proxy for the chinese market risk-free rate is a one-year large denomination time deposit rate. the returns reported in table 3 represent daily rates. as table 3 shows, the average daily return on the gta a-share index is 3.2 basis points, yet the net-of-fees daily return on actively managed chinese stock mutual funds is 2.9 basis points. we calculate the average daily total expense ratio from our database, which is about 0.7 basis points. therefore, the average gross daily return for these mutual funds is 3.6 basis points (i.e., 0.4 basis points above the daily average index return). this calculation shows that these funds on average perform slightly better than the market before the fees but not after. in the spirit of french (2008), chinese investors pay about 0.3 basis points (3.2 minus 2.9) per day for price discovery in the chinese stock market. 300 w. he et al. / financial services review 24 (2015) 289–311 besides the market factor, we also include popular risk factors for mutual fund performance analysis in our study. following tang et al. (2012), we use the size (small minus big, smb) and value (high minus low, hml) factors provided by the tianxiang investment analysis system. using the data provided in the csmar china stock market trading database, we also calculate the zero investment portfolio returns on the momentum factor and the cash flows factor. to construct the momentum factor returns, we follow the fama-french approach described on kenneth french’s website.1 we form six valueweighted portfolios based on size and prior (2–12) monthly returns, with the median size as the break point for size and the 30th percentile and 70th percentile as the break points for prior (2–12) monthly returns. we then calculate the zero-investment momentum factor return as the average return on the two high prior return portfolios (small high and big high) minus the average return on the two low prior return portfolios (small low and big low). to study the relationship between fund characteristics and fund performance, we use 2006 to 2011 as the observation period for performance (hereafter called the full period). to test the robustness of our results, we further split our sample into the 2006–2008 and 2009–2011 sub-periods. because the mutual funds in china report their fund information on a quarterly basis, most fund characteristics summarized in table 4 are quarterly data. the exception is management fees, which are an annualized percentage rate charged as a percentage of daily nav. that is, the rate is an annual rate but when the fee is charged to fund investors, it typically accrues on a daily basis. we report the average management fee rate over the period in which the fund performance is measured because some funds adjust their management fees during these periods. as this expense item provides compensation for the fund managers, it should have some implications for fund performance. chinese mutual funds incur other expenses table 4 descriptive statistics for actively managed chinese stock mutual fund characteristics and flows fund characteristics n mean min max sd panel a: basic statistics management fees (%) (2006–2011) 308 1.527 0.500 8.250 0.403 median fund size (10 billions of yuan) 308 0.362 0.006 2.303 0.441 fund age (years) 308 3.592 0.471 10.285 2.288 mean net flows (%) 308 �0.057 �0.473 0.179 0.114 standard deviation of net flows 308 0.274 0.017 0.692 0.151 management fees median fund size fund age mean net flows panel b: correlation coefficients median fund size (10 billions of yuan) �0.022 fund age (years) �0.082 0.494* mean net flows (%) �0.163* 0.468* 0.632* standard deviation of mean net flows �0.023 0.035 0.268* 0.192* this table summarizes the characteristics of actively managed chinese stock mutual funds including management fees, fund size, the mean, and the standard deviation (sd) of net flows, and fund age for the full sample. management fees and net flows are in percentages and fund size is measured in 10 billion yuan. fund age is the length of time since fund inception. *significant at the 0.01 level. 301w. he et al. / financial services review 24 (2015) 289–311 such as sales service fees and custody fees but they are not directly related to the incentives provided to fund managers. as table 4 shows, the management fees have an average of about 1.527%. despite the wide range shown in table 4, funds typically set the management fee at 1.5% for most of the sample period, which may contribute to our result that this item has no significant relationship with performance. we measure fund size as the median aum over the sample period or sub-periods to avoid the influence of small initial sizes because the older funds tend to experience fast growth during the market boom in 2006 and 2007. however, the initial sizes of the younger funds may be too large because they may have dwindled since establishment in the sluggish markets after 2009. we measure fund age as the number of years between a fund’s establishment date and january 1, 2012. for these 308 funds, the average size is 3.62 billion rmb and the average age is 3.59 years. the csmar china open-end funds database provides quarterly data on fund share flows. we use the difference between shares subscribed (sold to investors) and shares redeemed as the net fund flows in a quarter and normalize the net fund flows using the average of shares outstanding at the beginning and end of that quarter. we then calculate mean net flow for a fund over the sample period. as table 4 shows, on average, the 308 funds lost 0.057% of their shares outstanding each quarter, which may result from a loss of investor interest during the market crash in 2008 and the prolonged downward trend since 2009. we measure the volatility of the net flows with their standard deviations, which have an average of 0.274%, almost five times larger than the average mean net flows. as panel b in table 4 shows, fund size is positively correlated to fund age and mean net flows, whereas fund age and mean net flows are also statistically positively correlated. in other words, older funds tend to be larger and attract higher net flows. this is not surprising considering that younger funds may have been operating in a depressed market since they started after the market crash, which increased the difficulty of attracting positive net flows. by contrast, management fees are slightly negatively related to the mean net flows and statistically significant. the standard deviation of net flows is positively related to fund age and the mean net flows with statistical significance. thus, older funds attract higher net flows and the volatility of their net flows is also higher. the positive correlation between mean net flows and the standard deviation of net flows also indicates funds with a low standard deviation in net flows have mean negative net flows, which we observe from our data. the low volatility results from the steady negative net flows. as mentioned earlier, we believe this is one of the underlying reasons for the inverted-u shape relationship between fund flow volatility and fund performance. the csmar database does not provide a data field to identify whether a fund is “dead” or “alive” (i.e., no longer exists or still operates). thus, we check to determine whether a fund’s daily returns are reported until the last trading day of 2011 (december 30, 2011) to infer the fund’s status. our analysis reveals that the database contains four dead blend funds. thus, all stock funds, active or indexed, were still operating at year-end 2011. because our sample is a subset of stock funds, it contains no dead funds. this result is not surprising because all funds in our sample are relatively young compared with those in developed markets. the csmar open-end funds database, which is now part of wharton research data services (wrds), maintains the records for all funds whether they are dead or alive. therefore, having no dead funds in our sample is neither a consequence of our sample 302 w. he et al. / financial services review 24 (2015) 289–311 selection process nor a database issue. consequently, we do not believe that our sample suffers from survivorship bias. 5. empirical results in this section, we first examine the performance of actively managed chinese stock mutual funds for the full sample period and the two sub-periods. next, we analyze the results of cross-sectional regressions on how fund characteristics and fund flows are related to fund performance. 5.1. chinese stock mutual fund performance: 2001–2011 we first present the performance of actively managed chinese open-end stock mutual funds between october 19, 2001 and december 31, 2011. panel a of table 5a presents the results for the sharpe ratio, treynor ratio, four-factor �, �s, and adjusted-r2 in the fourfactor model regressions. panel b of table 5a reports the � and hm timing coefficients for the four-factor hm model. panel c presents the � and the timing coefficients for the table 5 a. risk-adjusted performance appraisal methods applied to 308 actively managed chinese stock mutual funds between 2006 and 2011 performance measures and coefficients n mean median min 5% 95% max sd panel a: four-factor model sharpe ratio (%) 308 �1.86 �0.92 �19.35 �12.89 4.52 37.49 6.04 treynor ratio (%) 308 �0.27 �0.02 �72.45 �0.21 0.10 0.15 4.13 four-factor � (%) 308 0.00 0.00 �0.15 �0.06 0.04 0.07 0.03 gta a-share index 308 0.69 0.71 0.00 0.44 0.89 1.25 0.14 smb 308 0.07 0.06 �0.51 �0.22 0.35 0.52 0.18 hml 308 �0.27 �0.24 �1.16 �0.60 �0.03 0.42 0.19 mom 308 0.02 0.02 �0.08 �0.02 0.05 0.06 0.02 adjusted r2 308 0.78 0.84 �0.01 0.28 0.92 0.96 0.18 panel b: four-factor hm model hm � (%) 308 �0.01 �0.01 �0.26 �0.07 0.08 0.17 0.05 hm timing 308 0.01 0.02 �0.39 �0.16 0.11 0.25 0.08 adjusted r2 308 0.78 0.84 �0.02 0.29 0.92 0.96 0.18 panel c: four-factor tm model tm � (%) 308 �0.001 0.001 �0.22 �0.06 0.05 0.12 0.04 tm timing 308 0.002 0.06 �5.25 �2.68 2.34 5.42 1.40 adjusted r2 308 0.78 0.84 �0.02 0.29 0.92 0.96 0.18 this table reports the risk-adjusted performance of actively managed chinese open-end stock mutual funds using four-factor models between 2006 and 2011. panel a shows the performance measures are the sharpe ratio, treynor ratio, and jensen’s � in the four-factor model. the coefficients on the market (excess return on the gta a-share index), size (smb), book-to-market equity (hml), and momentum (mom) factors are also shown for the four-factor model. panel b lists the � and the hm timing coefficients for the four-factor henriksson and merton (1981) model. panel c presents the � and the tm timing coefficients for the four-factor treynor and mazuy (1966) model. n is the number of funds. mean, median, min, 5%, 95%, max, and sd are the mean, median, minimum, 5th percentile, 95th percentile, maximum, and standard deviation. 303w. he et al. / financial services review 24 (2015) 289–311 four-factor tm model. to ensure the validity of our results, we use the robust standard error formula hc2 in mackinnon and white (1985) for all regressions. in table 5a, the average adjusted r2 is 78% in all three panels, which indicates that our standard four-factor model has strong explanatory power of the performance of these mutual funds. table 5a also shows that the average daily � is basically zero. among the four factors, the gta a-share index has the largest average coefficient value, which is not surprising because these funds are actively managed stock mutual funds that should have substantial market exposure. the coefficients of size (smb) tend to center more around zero compared with those of value (hml), which are overwhelmingly negative. according to this evidence, these funds place more weight on growth stocks, which is reasonable in an emerging economy. the momentum (mom) factor has coefficients that are distributed around zero in a relatively small range indicating that managers of chinese mutual funds do not place much emphasis on this factor when making portfolio allocations. in panels b and c, the average �s for both market timing models are slightly negative and the average timing coefficients are slightly above zero. table 5b shows the number of statistically significant coefficients in the four-factor models for our mutual fund sample based on the 0.05 level. as panel a of table 5b shows, 23 of 308 active funds have significantly positive (n�*) �s in our baseline model but this number decreases in the market timing regressions. almost all funds have statistically significant and positive exposure to the market factor. for the size factor (smb), 165 funds have statistically significant and positive exposure, which is almost twice the number of funds having statistically significant and negative exposure. however, 96% of the funds (297 of 308) have statistically significant and negative exposure to the value table 5 b. risk-adjusted performance appraisal methods applied to 170 actively managed chinese stock mutual funds between october 19, 2001 and december 30, 2011 performance measures and coefficients n� n�* n� n�* panel a: four-factor model four-factor � (%) 175 23 133 11 gta a-share index 308 306 0 0 smb 183 165 125 92 hml 11 1 297 252 mom 259 124 49 3 panel b: four-factor hm model hm � (%) 128 7 180 17 hm timing 204 14 104 10 panel c: four-factor tm model tm � (%) 159 18 149 14 tm timing 172 8 136 7 this table summarizes the signs and statistical significance of the coefficients in three four-factor models of actively managed chinese open-end stock mutual funds between october 19, 2001 and december 30, 2011. panel a shows the �, coefficients on the market (excess return on the gta a-share index), size (smb), book-to-market equity (hml), and momentum (mom) factors for the four-factor model. panel b lists the � and the hm timing coefficients for the four-factor henriksson and merton (1981) model. panel c presents the � and the tm timing coefficients for the four-factor treynor and mazuy (1966) model. n� and n� are the numbers of funds with positive and negative coefficients, respectively. n�* and n�* are the number of funds that report significantly positive or significantly negative coefficients, respectively. 304 w. he et al. / financial services review 24 (2015) 289–311 factor (hml). although the momentum (mom) factor may not be economically significant for these mutual funds, compared with the size (smb) and value (hml) factors, more than a third (124 out of 308) of the funds demonstrate positive and significant exposures to this factor. panels b and c of table 5b indicate both the market timing and security selection skills of these active funds. the four-factor hm model shows more significantly positive market timing coefficients (14 or 4.5% of 308) than the tm model (8 or 2.6% of 308). although not shown in table 5, our further investigation of the results shows that only five funds (1.6% of 308) appear to have market timing ability in both models. we observe that seven funds in the hm model and 18 funds in the tm model demonstrate positive and statistically significant �s, which indicate security selection skills. after further analysis, we find that among these funds, only six (2% of 308) show selection skills in both models. however, no fund shows both security selection and market timing skills. overall, only a very small fraction of these fund managers appear to have either market timing skills or security selection skills. 5.2. fund characteristics, fund flows, and fund performance: 2006–2011 tables 6 through 8 show the relationship among characteristics, fund flows, and the performance measures observed between 2006 and 2011. table 6 presents the results for the whole period while tables 7 and 8 present evidence for the first and second sub-periods, respectively. the adjusted r2s are positive for all regressions in tables 6 through 8. the fund table 6 regressions on actively managed chinese stock mutual fund characteristics and performance measures between 2006 and 2011 fund characteristics sharpe ratio treynor ratio four-factor � four-factor hm � four-factor hm timing four-factor tm � four-factor tm timing management fees �0.120 0.280 0.000 0.001 �0.002 0.238 (1.222) (0.414) (0.002) (0.060) (0.017) (0.802) fund size �0.627 �0.286 0.000 0.005 �0.004 0.223 (0.427) (0.284) (0.003) (0.010) (0.004) (0.135) age 1.517*** �0.057 0.002** 0.002* �0.001 0.003*** �0.069** (0.133) (0.083) (0.001) (0.001) (0.002) (0.001) (0.034) mean net flows 11.530** 6.707 0.095*** 0.139* 0.076** 1.338 (4.952) (6.117) (0.023) (0.077) (0.035) (1.501) sd of net flows 12.310 10.190 0.097** 0.038 0.056 0.109** �1.661 (7.982) (9.716) (0.046) (0.080) (0.139) (0.054) (2.398) (sd of net flows)2 �18.06 �14.37 �0.186** �0.071 �0.127 �0.190** 1.845 (11.640) (13.600) (0.076) (0.133) (0.230) (0.092) (3.825) constant �7.852*** �1.389 �0.011 �0.018* 0.0132 �0.013 0.157 (2.149) (1.341) (0.009) (0.011) (0.090) (0.026) (1.256) n 308 308 308 308 308 308 308 adjusted r2 0.528 0.018 0.201 0.002 0.014 0.12 0.004 this table shows the regression results of actively managed chinese open-end stock mutual fund performance on fund characteristics and investment styles between 2006 and 2011. the dependent variable is fund performance as measured by the sharpe ratio, treynor ratio, four-factor �, four-factor hm �, four-factor hm timing, four-factor tm �, and four-factor tm timing. the independent variables are various fund characteristics. management fees and mean net flows are in percentages and fund size is measured in 10 billion yuan. fund age is the length of time since fund inception. standard errors are in parentheses. ***, **, and * indicate significance at the 0.01, 0.05, and 0.10 levels, respectively. 305w. he et al. / financial services review 24 (2015) 289–311 performance measures as defined in section 4 are the dependent variables. to avoid data mining, we keep all fund characteristics and flow variables as explanatory variables when possible. table 6 shows the one exception, which is the regression for the hm � in table 6. in this situation, we drop management fees, fund size, and mean net flows because the values of their t-statistic are almost zero. thus, including them would result in a negative or lower adjusted r2. the adjusted r2s for the sharpe ratio, four-factor �, and tm � are comparable with those reported in fung et al. (2002). however, the adjusted r2s are not very high for other risk-adjusted performance measures. one may expect to see this kind of result from a market equilibrium perspective. if a fund characteristic indicates better risk-adjusted performance, investment inflows may chase the characteristic until it disappears (berk and green, 2004). in table 6, fund age is positively related to the sharpe ratio, four-factor �s in both the standard and tm models with a statistical significance above the 0.05 level. the adjusted r2 of these regressions ranges from 12% to 53%. furthermore, tables 7 and 8 show this significantly positive effect for the sharpe ratio for both sub-periods at the 0.01 level. for the four-factor �, this positive effect appears in the first sub-period in table 7 at the 0.10 level. these findings support the idea that funds with a longer operating history can generate better risk-adjusted returns. by contrast, fund age is negatively related to the tm timing coefficients and table 8 shows a similar result for the second sub-period. because the adjusted r2 values are quite low in both regressions, we hesitate to interpret too much out of this negative impact of age on the tm timing coefficient. mean net flows are positively related to the sharpe ratio, as well as to the �s in both the table 7 regressions on actively managed chinese stock mutual fund characteristics and performance measures between 2006 and 2008 fund characteristics sharpe ratio treynor ratio four-factor � four-factor hm � four-factor hm timing four-factor tm � four-factor tm timing management fees �3.874 �0.539 �0.087 0.125 �0.176 0.013 �1.014 (8.169) (0.659) (0.157) (0.092) (0.195) (0.064) (1.039) fund size 0.783 0.031 0.001 0.034*** �0.027** 0.021** �0.195*** (0.732) (0.033) (0.009) (0.012) (0.011) (0.010) (0.072) age 1.981*** 0.059*** 0.005* 0.005 0.001 0.005 �0.025 (0.310) (0.011) (0.003) (0.004) (0.003) (0.004) (0.023) mean net flows �5.098** �0.185* �0.074 �0.291* 0.172* �0.208* 1.238* (2.520) (0.096) (0.051) (0.163) (0.098) (0.113) (0.640) sd of net flows 44.36*** 1.814*** 0.096 0.246** �0.123 0.257** �1.659** (6.895) (0.341) (0.119) (0.121) (0.115) (0.114) (0.762) (sd of net flows)2 �49.65*** �2.012*** �0.097 �0.153 0.050 �0.213* 1.271 (7.885) (0.382) (0.129) (0.122) (0.123) (0.118) (0.801) constant �12.60 0.139 0.103 �0.283** 0.318 �0.093 2.066 (12.680) (1.020) (0.243) (0.140) (0.301) (0.097) (1.609) n 147 147 147 147 147 147 147 adjusted r2 0.577 0.544 0.062 0.302 0.253 0.241 0.280 this table shows the regression results of actively managed chinese open-end stock mutual fund performance on fund characteristics and investment styles between 2006 and 2008. the dependent variable is fund performance as measured by the sharpe ratio, treynor ratio, four-factor �, four-factor hm �, four-factor hm timing, four-factor tm �, and four-factor tm timing. the independent variables are various fund characteristics. management fees and mean net flows are in percentages and fund size is measured in 10 billion yuan. fund age is the length of time since fund inception. standard errors are in parentheses. ***, **, and * indicate significance at the 0.01, 0.05, and 0.10 levels, respectively. 306 w. he et al. / financial services review 24 (2015) 289–311 four-factor standard and tm models with high levels of statistical significance. table 8 shows similar findings for mean net flows on these performance measures in the second sub-period. furthermore, the mean net flows have a positive and statistically significant relationship with the hm timing coefficients at the 0.10 level, which also occurs in both sub-periods. even for the tm timing coefficient, this positive relationship appears in the first sub-period at the 0.10 level and for the whole period. although the sign is positive in the second sub-period, it is not statistically significance at normal levels. this evidence is consistent with the performance chasing behavior of investors. although fund flows often chase good performance, the negative fund flows from underperforming funds in the sluggish sideways market after the global financial crisis of 2007–2009 also merit attention. as table 6 shows, the positive coefficient on the standard deviation of net flows and the negative coefficients on its quadratic term indicate an inverted-u shape relationship. this is significant at the 0.05 level for the �s in both the four-factor standard and tm models. the signs are consistent for this pair of coefficients across all regressions except the tm timing coefficient, which shows no statistical significance. furthermore, both sub-periods have similar results for the four-factor tm �. the second sub-period provides a similar conclusion for the � in the four-factor standard model. table 6 also shows that management fees and fund size do not have any statistically significance coefficients for any of the regressions. moreover, the signs of these coefficients are in different directions for the various performance measures. as table 7 shows, the first sub-period yields similar effects for the volatility of fund flows on the sharpe and treynor ratios. as previously discussed, this inverted-u relationship is table 8 regressions on actively managed chinese stock mutual fund characteristics and performance measures between 2009 and 2011 fund characteristics sharpe ratio treynor ratio four-factor � four-factor hm � four-factor hm timing four-factor tm � four-factor tm timing management fees �0.155 0.225 �0.000 �0.002 0.003 �0.004 0.278 (1.067) (0.239) (0.002) (0.027) (0.051) (0.012) (0.669) fund size 1.562*** 0.360 0.012*** 0.003 0.017 0.006 0.258 (0.520) (0.346) (0.004) (0.007) (0.010) (0.005) (0.165) age 1.357*** 0.112 �0.002*** �0.002* �0.001 �0.001 �0.098** (0.158) (0.097) (0.001) (0.0014) (0.002) (0.001) (0.046) mean net flows 15.500*** 5.895 0.087*** 0.019 0.136* 0.059* 1.698 (4.378) (5.292) (0.021) (0.042) (0.073) (0.031) (1.401) sd of net flows 11.90 8.508 0.131*** 0.102 0.029 0.139*** �1.311 (7.377) (8.175) (0.043) (0.067) (0.122) (0.047) (2.183) (sd of net flows)2 �14.69 �10.04 �0.232*** �0.174 �0.075 �0.236*** 1.295 (9.604) (9.544) (0.072) (0.121) (0.218) (0.083) (3.566) constant �7.081*** �1.699 �0.012 �0.015 0.010 �0.013 0.168 (2.180) (1.601) (0.009) (0.040) (0.077) (0.020) (1.081) n 308 308 308 308 308 308 308 adjusted r2 0.491 0.018 0.118 0.004 0.021 0.045 0.010 this table shows the regression results of actively managed chinese open-end stock mutual fund performance on fund characteristics and investment styles between 2009 and 2011. the dependent variable is fund performance as measured by the sharpe ratio, treynor ratio, four-factor �, four-factor hm �, four-factor hm timing, four-factor tm �, and four-factor tm timing. the independent variables are various fund characteristics. management fees and mean net flows are in percentages and fund size is measured in 10 billion yuan. fund age is the length of time since fund inception. standard errors are in parentheses. ***, **, and * indicate significance at the 0.01, 0.05, and 0.10 levels, respectively. 307w. he et al. / financial services review 24 (2015) 289–311 consistent with the idea that high flow volatility hurts fund performance by disrupting the portfolio managers’ operations with costly trading. by contrast, low flow volatility is correlated with poor performance because the net negative flows from the underperforming funds can be relatively steady. as previously noted, the two sub-periods (2006–2008 and 2009–2011) generally provide evidence supporting the main findings in table 6 for the full sample period. however, some results differ among the sub-periods. for example, fund size in table 7 is positively related to the �s in the two timing models and its coefficients are statistically significant. because some of these results cannot be cross-validated or supported by theory, we refrain from over speculating on their meaning. nonetheless, some results still merit worth discussion. first, the regression for the treynor ratio in table 7 contains highly statistically significant coefficients, which support similar conclusions on fund age and fund flow volatility for other performance measures in table 6. furthermore, in table 7, fund size and fund flow volatility can negatively affect the timing coefficients in both the hm and tm models. this is consistent with our expectation that the price impact from larger fund sizes or disruptive fund flows can negatively influence timing ability. table 7 also shows that the mean net flows are negatively and significantly related to the sharpe ratio, treynor ratio, and �s in the two timing models. a plausible explanation is that heavy cash flows poured into the new funds before the peak in 2007, then the market crashed (post october 2007 to december 2008) and these funds performed poorly. thus, the high mean net flows may be correlated with inferior risk-adjusted performance. for the majority of the statistically significant coefficients in table 8, the results are consistent with table 6. 5.3. robustness checks we run similar regressions using the fund characteristics and fund flows with other dummy variables on fund investment styles or fund types (lof or contractual) as robustness checks. the results for the fund characteristics and fund flows are similar. however, the regression results show very few statistically significant and consistent coefficients for various investment styles. the fund type dummy in the performance or skill regressions has no statistically significant coefficients. therefore, we omit reporting these results, which are available on request. 6. summary and conclusions we examine the performance and market timing ability of actively managed chinese open-end stock mutual funds using daily return data. we also study the relationship between fund characteristics, fund flows, and various risk-adjusted performance measures and market timing skills. to our knowledge, we are the first to examine the performance of such mutual funds using the csmar database based on a four-factor model. the background information that we provide on the development of the chinese mutual fund industry and its unique features may help future researchers design new studies involving this expanding sector of the global capital market. 308 w. he et al. / financial services review 24 (2015) 289–311 based on our results, only about 7.5% of the actively managed chinese stock mutual funds in our sample have positive and statistically significant risk-adjusted returns based on a four-factor model. less than 5% of the funds show statistically significant market timing skills or stock selection skills. our investigation of the relationship of fund characteristics, fund flows, and fund performance indicates that older funds tend to perform better, especially using the sharpe ratio. this finding may reflect the presence of a learning effect. net flows are positively related to performance, which relates not only to investors chasing good performance but also to redemptions from underperforming funds. we find an inverted-u shape relationship between fund flow volatility and performance. this relationship suggests that poor performance is related both to disruptive high fund flow volatility and relatively steady net outflows as investors withdraw their money from underperforming funds. overall, our results provide a broad look at the performance and market timing skills of actively managed chinese stock mutual funds. our study may serve as a catalyst for others who address such questions as: have managers of actively managed chinese stock mutual funds improved their portfolio management skills over time? what is the dynamic relationship among the fund performance, manager skills, and fund flows? as the chinese mutual fund industry continues to grow, our study may help academic researchers, policy makers, and investors better understand this potentially important sector in the world capital markets. notes 1 this website is http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library/ det_mom_factor.html. references bauer r., koedijk k., & otten, r. (2005). international evidence on ethical mutual fund performance and investment style. journal of banking & finance, 29, 1751–1761. berk, j., & green, r. (2004). mutual fund flows and performance in rational markets. journal of political economy, 112, 1269–1295. bollen, n., & busse, j. (2001). on the timing ability of mutual fund managers. journal of finance, 56, 1075–1094. cao, b., & jayasuriya, s. (2012). volatility and the cross-sectional performance of emerging market hedge funds. journal of alternative investments, 14, 40–50. carhart, m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. chance, d., & hemler, m. (2001). the performance of professional market timers: daily evidence from executed strategies. journal of financial economics, 62, 377–411. chen, j., hong, h., huang, m., & kubik, j. (2004). does fund size erode mutual fund performance? the role of liquidity and organization. american economic review, 94, 1276–1302. droms, w., & walker, d. (1996). mutual fund investment performance. quarterly review of economics and finance, 36, 347–364. eling, m., & faust, r. (2010). the performance of hedge funds and mutual funds in emerging markets. journal of banking and finance, 34, 1993–2009. 309w. he et al. / financial services review 24 (2015) 289–311 elton, e. j., gruber, m. j., & blake, c. r. (1996). the persistence of risk-adjusted mutual fund performance. journal of business, 69, 133–157. elton, e. j., gruber, m. j., das, s., & hlavka, m. (1993). efficiency with costly information: a reinterpretation of the evidence for managed portfolios. review of financial studies, 6, 1–22. fama, e. f., & french, k. r. (1993). common risk factors in the returns on bonds and stocks. journal of financial economics, 33, 3–53. fama, e. f., & french, k. r. (2010). luck versus skill in the cross-section of mutual fund returns. journal of finance, 65, 1915–1947. frazzini, a., & lamont, o. (2008). dumb money: mutual fund flows and the cross-section of stock returns. journal of financial economics, 88, 299–322. french, k. r. (2008). presidential address: the cost of active investing. journal of finance, 63, 1537–1573. friesen, g., & sapp, t. (2007). mutual fund flows and investor returns: an empirical examination of fund investor timing ability. journal of banking and finance, 31, 2796–2816. fung, h. g., xu. x. e., & yau, j. (2002). global hedge funds: risk, return, and market timing. financial analysts journal, 58, 19–30. golec, j. h. (1996). the effect of mutual fund managers’ characteristics on their portfolio performance, risk and fees. financial services review, 5, 133–148. grinblatt, m., & titman, s. (1989). mutual fund performance: an analysis of quarterly portfolio holdings. journal of business, 62, 393–416. guo, f. (2010). an empirical study on the performance and market timing ability of chinese mutual funds. journal of harbin university of commerce (social science edition), 6, 22–27. haslem, j. a., baker, h. k., & smith, d. m. (2008). performance and characteristics of actively managed retail equity mutual funds with diverse expense ratios. financial services review, 17, 49–68. henriksson, r. a. (1984). market timing and mutual fund performance: an empirical investigation. journal of business, 57, 73–97. henriksson, r. a., & merton, r. c. (1981). on market timing and investment performance ii. statistical procedures for evaluating forecasting skills. journal of business, 54, 513–533. jan, y., & hung, m. (2003). mutual fund attributes and performance. financial services review, 12, 165–178. jensen, m. (1969). risk, the pricing of capital assets, and the evaluation of investment portfolios. journal of business, 42, 167–185. li, h. q., & ma, c. q. (2004). a theoretical and empirical study on the performance evaluation of chinese mutual funds. journal of finance and economics, 7, 56–65. (in chinese). mackinnon, j. g., & white, h. (1985). some heteroskedasticity-consistent covariance matrix estimators with improved finite sample properties. journal of econometrics, 29, 305–325. maginn, j. l., tuttle, d. l., pinto, j. e., & mcleavey, d. w. (2007). managing investment portfolios–a dynamic process. hoboken, nj: john wiley & sons, inc. otten, r., & bams, d. (2002). european mutual fund performance. european financial management, 8, 75–101. perold, a., & salomon, r. s. (1991). the right amount of assets under management. financial analysts journal, 47, 31–39. rakowski, d. (2010). fund flows volatility and performance. journal of financial & quantitative analysis, 45, 223–237. rakowski, d., & wang, x. (2009). the dynamics of short-term mutual fund flows and returns: a time-series and cross-sectional investigation. journal of banking & finance, 33, 2102–2109. rodier, m. (2009). china’s mutual fund industry goes international. (available at http://www.wallstreetandtech.com/articles/217701673). sharpe, w. f. (1966). mutual fund performance. journal of business, 39, 119–138. shen, w. t., & huang, x. r. (2001). an empirical study and appraisal on the performance of chinese mutual funds. economic research journal, 9, 22–30. tang, k., wang, w., & xu, r. (2012). size and performance of chinese mutual funds: the role of economy of scale and liquidity. pacific-basin finance journal, 20, 228–246. treynor, j. l. (1965). how to rate management of investment funds. harvard business review, 43, 63–75. 310 w. he et al. / financial services review 24 (2015) 289–311 treynor, j. l., & mazuy, k. k. (1966). can mutual funds outguess the market? harvard business review, 44, 131–136. zhang, x., & du, s. m. (2002). can chinese mutual funds beat the market? journal of financial research, 1, 1–22. zou, j. f., & lin, y. r. (2004). an empirical analysis of market timing and stock choosing ability of chinese mutual funds. journal of central university of finance and economics, 11, 26–30. 311w. he et al. / financial services review 24 (2015) 289–311 financial services review, 32(1) 47 how are you doing? financial well-being during covid-19 danah jeong1, benjamin hampton2, and kristy l. archuleta3 abstract optimal financial well-being is a goal for both financial professionals and consumers. the covid-19 pandemic raised concerns about consumers’ financial well-being. this study sought to explore the factors related to financial well-being using the personal financial wellness framework (joo, 2008). data was collected from a diverse sample during the covid-19 pandemic. results indicated that objective financial status (e.g., income), positive pre-pandemic financial behaviors, financial satisfaction, and being older and single mattered in one’s increased level of financial well-being during the covid-19 pandemic. subsequently, financial satisfaction was found to be a mediating factor between subjective financial knowledge and financial wellbeing. implications for financial professionals, researchers, and policymakers are provided. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation jeong, d., hampton, b., & archuleta, k. l. (2024). how are you doing? financial well-being during covid-19. financial services review, 32(1), 47-62. introduction financial professionals’ primary goal is to help clients have successful financial lives by enhancing their financial well-being, ultimately increasing their overall well-being. the covid19 pandemic had a significant impact on people across the world, including the u.s., but how did it affect financial well-being? in response to the pandemic, the u.s. government passed legislation and implemented policies aimed at providing financial support to consumers. the families first coronavirus response act became law on march 18, 2020 (moss et al., 2020). the 1 university of georgia, athens, usa 2 university of georgia, athens, usa 3 corresponding author (karchuleta@uga.edu). university of georgia, athens, usa key provisions of this legislation included nutrition benefits, 12 weeks of paid leave for employees needing to provide childcare, additional funding to states for unemployment benefits, paid sick leave, and insurance coverage for covid-19 testing. on the same day, the u.s. department of housing and urban development and the federal housing finance agency declared an eviction and foreclosure moratorium for mortgage loans backed by fannie mae, freddie mac, and the fha for 60 days (federal housing finance agency, 2020; u.s. department https://creativecommons.org/licenses/by-nc/4.0/ mailto:karchuleta@uga.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review 32(1) 48 of housing and urban development, 2020). the coronavirus aid, relief, and economic security (cares) act became law on march 27, 2020. the cares act provisions included direct payments of $1,200 per eligible individual and $500 for each child that qualified, a maximum of $100,000 of penalty-free withdrawals from retirement accounts for distributions associated with the coronavirus, qualified retirement plan loans up to $100,000, waived required minimum distributions for 2020, the allowance of a $300 charitable donation deduction for taxpayers that chose the standard deduction, and a $600 increase to weekly unemployment benefit payments through the end of july, 2020 (national association of tax professionals, 2020). these government interventions likely had an impact on consumers’ financial well-being. researchers are in the process of learning how individual financial well-being was impacted during and post-pandemic. learning from the crisis can help us better prepare clients for what may lie ahead. in this study, we aimed to explore the factors associated with enhanced financial well-being during the covid-19 pandemic. theoretical framework the personal financial wellness (pfw) framework (joo, 2008) was used to guide this study. in the personal finance wellness framework, financial wellness is defined as an active state of financial health that includes both subjective and objective evaluative terms. financial wellness is seen as one component of overall personal well-being and comprised four sub-constructs: (a) objective status, (b) financial satisfaction, (c) financial behavior, and (d) subjective perceptions. objective status includes financial factors of income and other objective measures of financial health such as financial ratios. according to joo (2008), financial satisfaction can be seen as a global measure of how one feels about their financial situation or comprised of multiple factors that may impact one’s perception of their own financial situation. financial behaviors capture what one is actively doing in various areas of personal finance, such as cash flow management, debt management, savings, etc. subjective perceptions encompass attitudes and subjective financial knowledge. literature has shown that attitudes and subjective financial knowledge (i.e., what one thinks they know about personal finance) can be the largest contributing factor to engaging in positive financial behaviors and having increased financial well-being (robb & woodyard, 2011). in this study, we use the four sub-constructs of joo’s (2008) personal wellness framework to understand financial well-being and will refer to financial wellness as financial well-being in this paper. literature review financial well-being financial well-being has been defined by the consumer financial protection bureau (cfpb) as "a state of being wherein a person can fully meet current and ongoing financial obligations, can feel secure in their financial future, and is able to make choices that allow them to enjoy life” (consumer financial protection bureau, 2017, p. 6). in cfpb’s definition, those with higher financial well-being can absorb financial shocks and have the freedom to make choices to enjoy life. similarly, netemeyer et al. (2018) conceptualized perceived financial well-being as two distinct constructs: (a) the distress related with managing one's money in the present and (b) feeling secure in one's financial future. they found that perceived financial well-being predicts overall well-being and the size of the effect is similar to that of the other life domains (e.g., relationship support satisfaction, physical health assessment, and job satisfaction) combined. vlaev and elliott (2014) contributed to the understanding of financial well-being by surveying young workers and families to determine the components of financial wellbeing. they found the top factor of influence on both groups of participants to be based on the amount of control they had over their finances. individuals with greater control over their finances were likely to indicate a higher state of financial well-being. this is important because during the covid-19 pandemic many people were worried about losing their jobs for pandemic-related reasons outside of their control. young adults’ financial well-being has been linked to overall life satisfaction, health, academic achievement, and psychological wellbeing (shim et al., 2009). arber et al. (2014) jeong et al. 49 examined the association between subjective financial well-being, income, and health during the middle and late stages of life for individuals in britain. they found that subjective financial well-being and income were both independently linked with health in the middle stage of life. during the late stage of life, subjective financial well-being was also linked with health, and it mediated the impact of income on health. objective status objective financial status is central to shaping financial well-being. one measure useful in objectively determining financial status is income. incorporating income enables an effective comprehensive financial evaluation as indicated by the many studies that have included income to determine financial well-being. for example, shim et al. (2009) included students’ income and their parents’ income as one of the objective measures in their models to estimate financial well-being. gerrans et al. (2014) found that objective financial status, including the level of income, assets, and debt, was strongly associated with financial satisfaction. other research has indicated that income was one of the strongest influences of financial well-being along with financial capability, financial inclusion and social capital (muir et al., 2017). using data from estonia, riitsalu and murakas (2019) found a positive relationship between income and financial well-being. other research suggests that the relationship between income and financial well-being may be more nuanced. zyphur et al. (2015) observed that only men had greater levels of subjective financial well-being when their incomes were higher. in contrast, the findings of malone et al. (2010) indicated that the financial well-being of american women elevates with income, education, age, and their contribution to household earnings. financial satisfaction another key component of financial well-being is financial satisfaction. financial satisfaction can be defined as one’s level of satisfaction with their financial circumstances (hira & mugenda, 1998). joo and grable (2004) created a framework for financial satisfaction and its determinants. their findings indicated that financial satisfaction is directly and indirectly associated with financial behavior, financial stress, financial knowledge, income, financial solvency, risk tolerance and education. are financial well-being and financial satisfaction related? according to the literature, the answer is yes. in one study, financial satisfaction was even used to operationalize subjective financial well-being (xiao & o’neill, 2018). however, many other studies view them as distinct concepts (fan & henager, 2022; prawitz et al., 2006; tenney & kalenkoski, 2019). research suggests that financial satisfaction is directly and positively related to financial well-being (fan & henager, 2022; west & cull, 2020). building on joo’s (2008) pfw framework, fan and henager (2022) conceptualized financial satisfaction as being one component of financial well-being. west and cull (2020) defined financial satisfaction as the level of satisfaction with one’s financial circumstances and financial well-being as consisting of personal characteristics, current financial management, and one’s expectations about the future of their personal finances. tenney and kalenkoski (2019) examined participants' objective measures of financial wellbeing using financial ratios and financial satisfaction to look for correlations. they found that the probability of being fully satisfied with one’s present financial circumstances rose by 0.19 with a 1% increase in the participants’ investment ratio. financial behavior joo (2008) identified financial behavior as being one of four sub-constructs of financial wellbeing. financial behavior is “any human behavior that is relevant to money management” (xiao, 2008, p. 70). financial behavior has been shown to affect financial satisfaction and overall financial well-being. the work of castrogonzález et al. (2020) indicated that an individual’s financial behaviors predict their financial well-being. gutter and copur (2011) examined the relationship between financial behavior and financial well-being for 15,797 college students via an online survey. they found differences in the magnitude of financial wellbeing by different financial behaviors and socioeconomic characteristics. specific financial financial services review 32(1) 50 behaviors (e.g., saving, budgeting, compulsive buying, and risky credit card use) were shown to have a significant relationship with financial well-being. of the variables tested by joo and grable (2004), financial behavior was found to have the greatest impact on financial satisfaction. conversely, the findings of robb and woodyard (2011) suggested that financial satisfaction influences financial behavior. these results imply that financial behavior and financial satisfaction are interrelated. woodyard and robb (2016) sought to conduct a study similar to the work of joo and grable (2004) but with a larger sample. their findings indicate that behavior and feelings may contribute to financial satisfaction more than knowledge. subjective perceptions the final sub-construct of financial well-being according to joo’s (2008) framework is subjective perceptions. according to her framework, subjective perceptions are comprised of one’s financial attitudes and financial knowledge (i.e., what one thinks they know about personal finance). research tells us that both constructs are salient. utilizing the cfpb’s financial well-being scale, lee et al. (2020) found that having greater subjective financial knowledge increased financial well-being and the propensity to plan amplified this positive relationship. another study showed that one’s attitude to money influenced financial behaviors, and financial behaviors predicted financial wellbeing (castro-gonzález et al., 2020). subjective financial knowledge was also found to impact financial behaviors (robb & woodyard, 2011). based upon joo’s (2008) personal financial wellness framework and the literature review, the following hypotheses were developed. h1: objective financial status (i.e., income, employment) will be positively associated with increased levels of financial well-being during covid-19. h2: pre-covid-19 pandemic financial behaviors will be positively associated with increased levels of financial wellbeing during covid-19. h3: subjective perceptions (i.e., financial knowledge) will be positively associated with increased levels of financial wellbeing during covid-19. h4: financial satisfaction will be positively associated with increased levels of financial well-being during covid-19. h5: subjective perceptions (financial knowledge) will have an indirect effect on financial well-being through financial satisfaction during covid-19. methodology this study utilized qualtrics panels to recruit the sample. panel members were sent an email invitation or prompted on the respective survey platform to proceed with the survey. the invitation provided a hyperlink and a nominal incentive in which the panel member responded by clicking the link. the incentives offered were not standard but rather unique to the individual. the researchers did not know the incentives that were offered. the only criteria to participate is that respondents needed to be 18 years or older. data were gathered as part of an experimental study about goal setting and overall well-being that included a pre-test, online exercise, and posttest. only the initial survey data were included in the current study as the focus of this study was not to test the outcome of the online exercise rather to examine the relevant variables crosssectional data in the pre-test survey. the data from this survey were collected in april 2020. a total of 145 respondents completed the initial survey. outcome variable the outcome variable was financial well-being. in this study, we used the financial well-being measure developed by the consumer financial protection bureau. according to the cfpb report (2017), the scale is associated with four elements: (a) “control over daily and monthly finances,” (b) “capacity to absorb a financial shock,” (c) “being on track to meet financial goals,” and (d) “the financial freedom to make choices that allow enjoyment of life.” due to those elements, the financial well-being concept is reflected as subjective and perceived. financial well-being is measured by ten questions on a likert scale ranging from 1 to 5. the scores reported were jeong et al. 51 processed according to cfpb guidelines to create a single score that ranged from 0 to 100. independent variables financial behavior financial behavior was measured by eight items adapted from grable & joo (2004) and dew & xiao (2011). the eight items assessed how participants normally handled their money prior to the covid-19 pandemic (before march 6, 2020). the items included: (a) “ i set money aside for savings or an emergency fund;” (b) “i set money aside for retirement;” (c) “i had a plan to reach my financial goals;” (d) “i had a weekly or monthly budget that i followed;” (e) “i paid credit card bills in full and avoided finance charges;” (f) “i reached the maximum limit on a credit card;” (g) “i spent more money than i earned;” and (h) “i paid my bills on time.” each variable response ranged from 1 (never) to 4 (always) except for the questions asking about reaching the maximum limit on a credit card and spending more money than earned. these two items were reverse coded, ranging from 1 (always) to 4 (never). the eight items were summated with scores ranging from eight to a maximum of 32. financial knowledge both subjective and objective financial knowledge were measured. a single selfreflected question assessed respondents’ perceived financial knowledge using likert-type scale. scores on the item could range from 1 (very little financial knowledge) to 10 (very high financial knowledge). objective financial knowledge was assessed through five questions focused on key financial concepts: compound interest (question 1), inflation (question 2), diversification (question 3), retirement planning (question 4), and time value of money (tvm) (question 5). these questions were either multiple-choice or true/false. correct responses were assigned a score of 1, while incorrect ones received 0. the total number of correct answers was then summed, with higher scores indicating a greater extent of objective financial knowledge. the average score for correct answers was 2.4, with a standard deviation of 1.47. summation techniques have been employed across various studies, including those by dew & xiao (2011), grable et al. (2020), lind et al. (2020), mountain et al. (2020), and robb et al. (2012). table 1 displays the questions used to assess the objective financial knowledge. table 1. objective financial knowledge questions q.1  suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow?  q.2  imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, would you be able to buy more than, exactly the same as, or less than today with the money in this account?  q.3  do you think that the following statement is true or false? “buying a single company stock usually provides a safer return than a stock mutual fund.”  q.4  true or false: there are annual contribution limits on the amount you can save in a 401(k) plan or ira that depend on your income.  q.5  assume a friend inherits $10,000 today and his sibling inherits $10,000 three years from now. who is richer because of the inheritance?  financial services review 32(1) 52 financial satisfaction a commonly used single-item financial satisfaction question was used in this study. the question was, “how satisfied are you with your overall current financial situation?” using a 10point likert-type scale, scores could range from 1 (very dissatisfied) to 10 (very satisfied). control variables demographic variables used as control variables included gender, income, education level, race/ethnicity, employment status. in terms of gender, those who identified themselves as females were coded 1, otherwise 0. due to a small sample size, some demographic variables were dichotomized. average income ranged between $40,000 and $50,000 and was used to create a dichotomous variable where “above average income” ($50,000 or above) was coded 1, otherwise 0. educational attainment was also coded dichotomously with those respondents who have a college degree or higher level of education coded 1, otherwise 0. if the race of a respondent was other than white, then race was coded 1, otherwise 0. those who reported that they were single were coded 1, otherwise 0. employment status (fully employed and partially employed, relative to non-employed) was included in our analysis. statistical analysis we conducted descriptive analyses and correlations between explanatory variables. next, we tested the hierarchical ordinary least square (ols) regressions models to estimate the relationship among the core variables, controlling for the socio-economic characteristics. finally, additional analyses were conducted to test for a possible mediating relationship. baron and kenny’s (1986) causal-steps test was employed in the mediation analyses and included the following assumptions: 1. the effect of x on y is significant. 2. the effect of x on m is significant. 3. the effect of m on y controlled for x is significant. 4. the effect of x on y controlled for m is smaller than the total effect of x on y. the sample size of this study (n = 145) is relatively small for the sem analysis. thus, we used the regular causal-steps test without using sem software to test the mediational relationship. fritz and mackinnon (2007) showed that the casual steps test without using sem software was the most frequently used methodologies by psychologists and studies with smaller median sample sizes (n < 159.5) than methods that used sem software. along with the causal steps analysis, bootstrapping and the sobel test were performed to test robustness. results table 2 summarizes the descriptive statistics of this study. the average age of the respondents was 41.5 (sd = 13.8). the average household gross income fell between $40,000 and $50,000. nearly half (49%) of respondents reported that their household income was above $50,000. twenty-six percent of respondents of the sample reported attaining a college degree or higher level of education. about 63% of the sample reported being white, while the other races comprised 37%. thirty-five percent of respondents reported that they were single, more than half (53%) of respondents were female, and 47% were male. slightly less than half of the respondents (49%) were employed full-time at the time of the interview, while 14% were employed part-time, and 37% were non-employed. the average financial well-being score was 49.4 (sd = 13.4). the average pre-pandemic financial behavior scale score was 21.2 out of 32. perceived financial knowledge (m = 6.5, sd = 2.2) appeared higher than actual financial knowledge (m = 2.4; sd = 1.5) for respondents in this study. specifically, the average perceived financial knowledge was 68%, while respondents scored 46% on a five-item test of objective financial knowledge.  jeong et al. 53 table 2. descriptive statistics (n = 145)  variables frequency (%) mean sd age  41.5  13.79  household income (categories)       (1) 0-$20,000  28.3%  0.45  (2) $20,001-$40,000  22.8%  0.42  (3) $40,001-$70,000  22.1%  0.42  (4) $70,001-$100,000  16.6%  0.37  (5) $100,001+  10.3%  0.31  education attainment        high school  31.7%  0.47  some college  42.8%  0.50  college   18.6%  0.39  graduate  6.9%  0.25  race        white  62.8%  0.49  african american  17.9%  0.38  asian  8.3%  0.28  hispanic  9.0%  0.29  native american  2.1%  0.14  marital status        single  35.2%  0.48  married  42.8%  0.50  divorced  12.4%  0.33  separated  1.4%  0.12  others  8.3%  0.28  female  53.8%  0.50  employment        full-time employed 49.0%  0.50  part-time employed 13.8%  0.35  non-employed  36.6%  0.48  financial behavior  21.2  5.80  financial well-being 49.4  13.39  subjective financial knowledge  6.5  2.23  objective financial knowledge  2.4  1.47  financial satisfaction  5.9  2.24  table 3 shows the correlation coefficients among the variables of interest in this study (i.e., income, subjective knowledge, pre-pandemic financial behavior, financial satisfaction, and financial well-being). all the variables were significantly correlated with each other at the p < .05 level or lower, except for subjective financial knowledge and income, which was slightly above the p < .05 standard (p < .059). none of the correlations were above the r < .70 benchmark to indicate potential multicollinearity issues. the results of this correlation analysis showed positive relationships among the variables. the financial services review 32(1) 54 findings suggested that more appropriate financial behavior was correlated to greater confidence in self-reflective financial knowledge; the higher the household income level, the higher the financial satisfaction and the cfpb wellbeing score. the size of the correlation between financial behavior and the cfpb well-being score was 0.61. the results suggest that individuals with better financial behavior before covid-19 are more likely to have a higher cfpb well-being score. higher self-reflective financial knowledge was related to a higher household income level, financial satisfaction, and the cfpb well-being score. also, household income level was positively related to financial satisfaction and cfpb well-being score. financial satisfaction was strongly correlated with the cfpb wellbeing score (0.57). we tested these relationships in multivariate regression models (reported in table 4) that control for the effects of other individual difference variables in multiple steps. table 3. correlation results between major variables   financial behavior score  subjective financial knowledge  household income  financial satisfaction  cfpb well-being score  financial behavior 1.0000              score                subjective 0.2539  1.0000           financial knowledge 0.0021              household 0.3814  0.1571  1.0000        income <.0001  0.0592           financial 0.3242  0.4646  0.2173  1.0000     satisfaction <.0001  <.0001  0.0086        cfpb well-being 0.6146  0.2716  0.3495  0.5650  1.000  score <.0001 0.001  <.0001  <.0001  hierarchical regression the results of the hierarchical ols regressions for the cfpb well-being score are shown in table 4. our regression findings support hypotheses 1 through 4. in model 1, age, attaining a higher education degree, being a race other than white, being single, gender, and fulland part-time employment were examined as predictors of financial well-being. the results indicated that age was positively related to financial well-being and holding a college degree or above was marginally significantly associated with financial well-being. other variables were not statistically significantly related to financial well-being. the r2 for this model was 12%. in the second model, household income was added. the results indicated that age was still statistically significantly associated with financial well-being, but higher education was no longer significantly associated with financial well-being. females became marginally significantly and negatively associated with financial well-being. interestingly, with controlling objective financial status (income), employment status (both full-and part-time, relative to non-employed) was negatively associated with financial well-being at a marginally significant level. income was significantly and positively associated with the financial well-being score. if an individual’s income level was above-average, the financial well-being score was 11 points higher than that of jeong et al. 55 his counterpart. the r2 for this model was 24%. aligning with h1, objective financial status (income) positively correlated with financial well-being. the findings support the hypothesis by suggesting that individuals with higher income levels obtain greater financial well-being. in the third model, the control variables from model 2 were examined in addition to financial behavior. the results showed that age, being single, above-average income level, and higher financial behavior scores were positively related to financial well-being. full-time employment was negatively associated with financial wellbeing at a marginally significant level. the r2 for this model was improved to 50%. h2 proposed that a higher level of financial behavior positively predicted levels of financial well-being. the coefficient on financial behavior was significant and positive. thus, h2 is supported. in the fourth model of this analysis, we included the perceived financial knowledge and objective financial knowledge variables in the model. the results showed that age, being single, higher financial behavior scores, above-average household income, and subjective financial knowledge were positively associated with financial well-being. being a full-time employee (relative to nonemployed) was negatively associated with financial well-being at a statistically significant level. the r2 for this model was 53%. in support of h3, subjective perceptions (i.e., financial knowledge) were positively related to levels of financial well-being. the findings indicated that individuals who regard themselves as financially knowledgeable showed approximately a 1-point higher financial well-being score. in our final model, we included a financial satisfaction variable along with the previous variables. the results indicated that age, being single, above-average household income, and positive financial behavior were significantly and positively associated with the financial wellbeing score while being a full-time employee was negatively associated with financial well-being. however, with the inclusion of financial satisfaction, subjective financial knowledge was no longer significantly associated with the wellbeing score. the final model was significant, accounting for 64% of the total variance in the model. in h4, a positive relationship between financial satisfaction and levels of financial wellbeing was hypothesized. our results support the hypothesis, finding that one point increase in financial satisfaction was related to an increase of 2.3 points in the financial well-being score. mediation analyses in our final regression model, we found that adding financial satisfaction completely removed the effect of subjective perception (subjective financial knowledge) on levels of financial wellbeing. we hypothesized that there would be a mediation effect of financial satisfaction on the association between subjective perception and financial well-being. table 5 and figure 1 present the findings from the mediation test. in path a, the direct association between subjective financial knowledge (i.e., subjective perception) and financial satisfaction was positively significant, indicating that an individual who perceived themselves as financially knowledgeable was more financially satisfied. in the second path, the direct association between financial satisfaction and financial well-being was significant and positive. third, the direct association between subjective financial knowledge and financial well-being was also positively significant. however, in the fourth path, the association between subjective financial knowledge and financial well-being score was found to be statistically insignificant when controlling for financial satisfaction. financial services review 32(1) 56 table 4. ols regression results   model 1 model 2 model 3 model 4 model 5   b   se b   se b   se b   se b   se age  0.27 *** 0.09 0.30 *** 0.08 0.29 *** 0.07 0.28 *** 0.07 0.20 *** 0.06 college or above 4.48 * 2.67 0.57   2.62 0.80   2.13 0.10   2.17 0.35   1.93 other races (ref: white) 2.03   2.34 3.52   2.20 1.44   1.81 1.35   1.78 0.94   1.58 single 1.35   2.38 3.73   2.27 4.02 ** 1.85 3.58 * 1.81 2.94 * 1.61 female -3.77   2.51 -4.04 * 2.33 -2.31   1.91 -1.69   1.94 -2.18   1.72 full-time employed -2.24   2.83 -4.91 * 2.69 -4.20 * 2.19 -5.69 ** 2.20 -6.88 *** 1.96 part-time employed -3.55   3.48 -6.00 * 3.28 -1.38   2.73 -1.97   2.68 -2.92   2.38 above average income       10.90 *** 2.30 5.10 ** 2.00 5.22 *** 1.96 4.16 ** 1.75 financial behavior score             1.31 *** 0.16 1.20 *** 0.16 1.04 *** 0.14 subjective financial knowledge                   1.07 *** 0.39 0.15   0.38 objective financial knowledge                   0.53   0.68 0.49   0.61 financial satisfaction                         2.30 *** 0.38 r2 0.12     0.25     0.50     0.53     0.64     *p<.05, **p<.01, ***p<.001 the results indicated that financial satisfaction fully mediated the association between subjective financial knowledge and financial well-being, making the relationship insignificant. in other words, subjective knowledge influences financial well-being only through the level of financial satisfaction. we ran bootstrap and sobel tests to confirm the mediation effect. the bootstrap results indicated that the indirect effect of financial knowledge on financial well-being was significant. the estimated direct effect of financial knowledge was 0.07, and the indirect effect mediated by financial satisfaction was 1.6. that is, 96% of the total effect of subjective knowledge was mediated by financial satisfaction. we also conducted the sobel test. we calculated the z-value following the sobel test equation suggested by baron and kenny (1986): z value =ab/sqrt (b2 sa 2 + a2 sb 2) in this equation, a = raw (unstandardized) regression coefficient for the association between the independent variable and the mediator; sa = the standard error of a; b = raw coefficient for the association between the mediator and the dependent variable; and sb = the standard error of b. aligning with bootstrap results, the sobel test results suggest that the indirect effect of subjective knowledge on cfpb well-being score via financial satisfaction is significantly different from zero (4.71, p < .0001). the findings support h5 that subjective perceptions (financial knowledge) indirectly affect financial well-being through financial satisfaction. jeong et al. 57 figure 1. mediation results of financial satisfaction between subjective financial knowledge and financial well-being (n = 145) table 5. mediation test results      explanatory variables  outcome variables  b path a  subjective financial knowledge  financial satisfaction  0.47  ***  path b  financial satisfaction  financial well-being 3.38  ***  path c  subjective financial knowledge  financial well-being  1.63  **  path c'  subjective financial knowledge  financial well-being  0.07     *p<.05, **p<.01, ***p<.001 discussion in alignment with joo’s (2008) personal financial wellness framework, objective financial status (e.g., income), positive prepandemic financial behaviors, and financial satisfaction mattered in one’s increased level of financial well-being during the covid-19 pandemic. in addition, age and being single were significant demographic variables that were associated with financial well-being in the final model. of no surprise, displaying positive financial behaviors prior to the covid-19 pandemic, having above-average household income, and being satisfied with one’s personal finances are major factors in raising financial well-being. however, a few interesting observations were made from the analyses that provide implications for practitioners, researchers, and policy makers. first and most notably, full-time employment was negatively and statistically significantly associated with increased financial well-being. this is an interesting insight that lends itself to potential policy implications during a pandemic, such as covid-19. a number of factors could have affected this association, such as difficulty juggling work and personal life, caring for loved ones or oneself being sick, or having reduced or reallocated resources. one factor that could have affected the negative relationship between both full-time and part-time employed respondents (although the relationship between part-time employment and financial well-being was not significant) in this sample is that those who were unemployed would have received additional financial benefits from the federal government under the cares act. with this in mind, our financial services review 32(1) 58 study also indicates that above-average income was positively and significantly associated with financial well-being. another factor that may have impacted the relationship between being employed and financial well-being is that those employed full-time or less may have been concerned about the potential of pandemicrelated job loss or reduction in pay whereas the unemployed did not have employment or wages to lose. support for this idea exists in the literature, like choi et al. (2020) who found that feeling insecure about one’s job status was negatively related to financial well-being. similar to previous research (robb & woodyard, 2011), objective financial knowledge did not seem to matter in any of the models. subjective financial knowledge was important; however, once financial satisfaction was included, subjective financial knowledge was no longer significant. further analyses found that financial satisfaction fully mediated the relationship between subjective financial knowledge and financial well-being. when one feels they know about personal finance and are satisfied with their financial situation, then financial well-being rises. one could raise the question that subjective financial knowledge may be measuring financial self-confidence as the measure specifically asks respondents about how much they think they know. in this case, having confidence about what you think you know plays an important role in financial well-being through financial satisfaction. limitations as with any study, limitations are present. first, financial behavior was a lookback measure. to answer the relevant questions, the respondent had to look back to their financial behaviors before covid-19 (prior to march 2020), which was at least three months prior to the interview date. personal experience or emotional judgement amid the turmoil of covid-19 could be reflected in their responses. thus, their behaviors before covid-19 may be distorted or embellished according to their current financial or health status. second, as previously mentioned, the data were collected online during an unprecedented time in not only u.s. history but also global history. individuals and families faced complex issues of navigating health and safety concerns, psychological issues, and financial distress. while it is beyond the scope of this study to take into account these factors, we capture a glimpse of potential outcomes that americans faced. third, unemployment was found to be a significant factor in understanding financial wellbeing during the covid-19 pandemic. however, we do not know when respondents if respondents were unemployed prior to the pandemic or as a result of the pandemic. we only know that they were unemployed at the time of the survey. knowing when respondents became unemployed may contribute to further understanding of why it was an important factor for increased financial well-being. finally, although the sample represents diverse demographics in terms of age, gender, and education attainment, the sample may not generalize the u.s. population. for example, in regard to race/ethnicity, this sample included fewer people who identified as white (62.8%) and more black/african americans (17.9%) compared to the u.s. population (76.3%, 13.4%, respectively). on the other hand, gender was near the national average in this sample including 53% female compared to the national average of 50.5% (u.s. census bureau, 2019). implications multiple implications can be drawn from this study regarding practice, research, and policy, all in the pursuit of improved client and, in general, consumer financial well-being. first, financial professionals (e.g., financial counselors, financial planners, and financial therapists) should continue striving to not only increase their clients’ objective financial status, but also clients’ positive financial behaviors, confidence surrounding financial matters, and financial satisfaction. findings from this study indicate that clients would benefit from crisis preparation. financial professionals would be prudent to help clients develop positive financial behaviors prior to a crisis as well as equip themselves with knowledge about unemployment benefits and other government benefits should their clients find themselves unemployed or in an unprecedented crisis similar to covid-19. for jeong et al. 59 some financial professionals (e.g., financial planners/advisors), they may not normally work with clients who are unemployed or who may have lower financial well-being when they are employed full-time during a crisis. a stigma against accessing government benefits may exist for some clients and financial professionals. however, this study’s findings demonstrate that government benefits when unemployed, like those offered through the cares act, may in fact ease the financial burden and help clients move to state of enhanced financial well-being. while it was beyond the scope of this study, financial professionals may find it beneficial to understand government benefits, prepare for a crisis response, and help clients overcome the challenges posed during and after a crisis. second, future research should further examine the relationship between unemployment and financial well-being as well as above average income during the covid-19 pandemic. as already noted, unemployment and above average income were significantly related to increased levels of financial well-being. while above average income intuitively makes sense, unemployment does not. this relationship leads to the next implication for policy that indicates that it is possible that government benefits do, at least in the short-term, improve individuals’ financial well-being. longitudinal research would help examine how people who have higher income and who are employed full-time versus unemployed during a crisis fair long after the crisis is over. finally, and conversely to the previous point, longitudinal research should explore how stimulus policies—such as enhanced unemployment benefits— may have negative effects, such as extending dependency on government benefits. conclusion using joo’s (2008) personal financial wellness framework, this study set out to discover how individuals’ financial wellbeing was impacted by the covid-19 pandemic. according to joo’s framework, objective status, subjective perspectives, financial behaviors, and financial satisfaction play a role in financial well-being. however, this study found that subjective financial knowledge was not a significant predicting variable directly related to financial wellbeing when accounting for all the other factors. however, financial satisfaction served as a key mediator between subjective financial knowledge and financial wellbeing. in addition, full-time employment was negatively related to financial well-being while above average income was positively related. considering the timing of the survey—during the covid-19 pandemic—a number of insights were gleaned to help practitioners, researchers, and policy makers further understand factors important to enhancing financial well-being, begging the question, how are you doing? references arber, s., fenn, k., & meadows, r. (2014). subjective financial well-being, income and health inequalities in mid and later life in britain. social science & medicine, 100, 12-20. https://doi.org/10.1016/j.socscimed.201 3.10.016 baron, r. m., & kenny, d. a. (1986). the moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. journal of personality and social psychology, 51(6), 1173. castro-gonzález, s., fernández-lópez, s., reyares, l., & rodeiro-pazos, d. (2020). the influence of attitude to money on individuals’ financial well-being. social indicators research, 148(3), 747–764. https://doi.org/10.1007/s11205-01902219-4 choi, s. l., heo, w., cho, s. h., & lee, p. (2020). the links between job insecurity, financial well‐being and financial stress: a moderated mediation model. international journal of consumer studies, 44(4), 353-360. https://doi.org/10.1111/ijcs.12571 financial services review 32(1) 60 consumer financial protection bureau. (2017). cfpb financial well-being scale: scale development technical report. https://files.consumerfinance.gov/f/docu ments/201705_cfpb_financial-wellbeing-scale-technical-report.pdf dew, j. p., & xiao, j. j. (2011). the financial management behavior scale: development and validation. journal of financial counseling and planning, 22(1), 43-59. fan, l., & henager, r. (2022). a structural determinants framework for financial well-being. journal of family and economic issues, 43(2), 415-428. https://doi.org/10.1007/s10834-02109798-w federal housing finance agency. (2020). fhfa suspends foreclosures and evictions for enterprise-backed mortgages. https://www.fhfa.gov/media/publicaffai rs/pages/fhfa-suspends-foreclosuresand-evictions-for-enterprise-backedmortgages.aspx fritz, m. s., & mackinnon, d. p. (2007). required sample size to detect the mediated effect. psychological science, 18(3), 233-239. doi:10.1111/j.14679280.2007 gerrans, p., speelman, c., & campitelli, g. (2014). the relationship between personal financial wellness and financial wellbeing: a structural equation modelling approach. journal of family and economic issues, 35(2), 145-160. https://doi.org/10.1007/s10834-0139358-z gutter, m., & copur, z. (2011). financial behaviors and financial well-being of college students: evidence from a national survey. journal of family and economic issues, 32(4), 699-714. http://dx.doi.org/10.1007/s10834-0119255-2 grable, j., archuleta, k. l., ford, m. r., kruger, m., gale, j., & goetz, j. (2020). the moderating effect of generalized anxiety and financial knowledge on financial management behavior. contemporary family therapy, 42(1), 15–24. hira, t. k., & mugenda, o. m. (1998). predictors of financial satisfaction: differences between retirees and non-retirees. journal of financial counseling and planning, 9(2), 75. joo, s. (2008). personal financial wellness. in handbook of consumer finance research (pp. 21-33). springer, new york, ny. joo, s., & grable, j. e. (2004). an exploratory framework of the determinants of financial satisfaction. journal of family and economic issues, 25(1), 25–50. https://doi.org/10.1023/b:jeei.00000167 22.37994.9f lee, j. m., lee, j., & kim, k. t. (2020). consumer financial well-being: knowledge is not enough. journal of family and economic issues, 41(2), 218228. https://doi.org/10.1007/s10834019-09649-9 lind, t., ahmed, a., skagerlund, k., strömbäck, c., västfjäll, d., & tinghög, g. (2020). competence, confidence, and gender: the role of objective and subjective financial knowledge in household finance. journal of family and economic issues, 41(4), 626-638. http://dx.doi.org/10.1007/s10834-02009678-9 malone, k., stewart, s. d., wilson, j., & korsching, p. f. (2010). perceptions of financial well-being among american women in diverse families. journal of family and economic issues, 31(1), 6381. https://doi.org/10.1007/s10834-0099176-5 moss, k., dawson, l., long, m., kates, j., musumeci, m., cubanski, j., & pollitz, k. (2020, march 23). the families first coronavirus response act: summary of key provisions. kff. https://www.kff.org/global-healthpolicy/issue-brief/the-families-firstjeong et al. 61 coronavirus-response-act-summary-ofkey-provisions/ mountain, t. p., kim, n., serido, j., & shim, s. (2020). does type of financial learning matter for young adults’ objective financial knowledge and financial behaviors? a longitudinal and mediation analysis. journal of family and economic issues, 42(1), 1–20. muir, k., hamilton, m., noone, j. h., marjolin, a., salignac, f., & saunders, p. (2017). exploring financial wellbeing in the australian context. sydney, australia: centre for social impact and social policy research centre, unsw sydney. national association of tax professionals. (2020). coronavirus aid, relief, and economic security act (cares) summary. https://www.natptax.com/taxknowledg ecenter/covid19resources/pages/care s-act-summary.aspx netemeyer, r. g., warmath, d., fernandes, d., & lynch jr, j. g. (2018). how am i doing? perceived financial well-being, its potential antecedents, and its relation to overall well-being. journal of consumer research, 45(1), 68-89. https://doi.org/10.1093/jcr/ucx109 prawitz, a., garman, e. t., sorhaindo, b., o’neill, b., kim, j., & drentea, p. (2006). incharge financial distress/financial well-being scale: development, administration, and score interpretation. journal of financial counseling and planning, 17(1), 34-50. riitsalu, l., & murakas, r. (2019). subjective financial knowledge, prudent behaviour and income: the predictors of financial well-being in estonia. international journal of bank marketing, 37(4), 934950. https://doi.org/10.1108/ijbm-032018-0071 robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals' advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21(4), 291-306. robb, c. a., & woodyard, a. s. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning, 22(1), 60-70. shim, s., xiao, j. j., barber, b. l., & lyons, a. c. (2009). pathways to life success: a conceptual model of financial well-being for young adults. journal of applied developmental psychology, 30(6), 708723. https://doi.org/10.1016/j.appdev.2009.02 .003 tenney, j. a., & kalenkoski, c. m. (2019). financial ratios and financial satisfaction: exploring associations between objective and subjective measures of financial well-being among older americans. journal of financial counseling and planning, 30(2), 231243. doi:10.1891/1052-3073.30.2.231 u.s. department of housing and urban development. (2020). mortgagee letter 2020-04. https://www.hud.gov/sites/dfiles/ochc o/documents/20-04hsgml.pdf vlaev, i., & elliott, a. (2014). financial wellbeing components. social indicators research, 118(3), 1103–1123. https://doi.org/10.1007/s11205-0130462-0 west, t., & cull, m. (2020). future expectations and financial satisfaction. economic papers: a journal of applied economics and policy, 39(4), 318-335. https://doi.org/10.1111/17593441.12292 woodyard, a. s., & robb, c. a. (2016). consideration of financial satisfaction: what consumers know, feel and do from a financial perspective. journal of financial therapy, 7(2), article 4. https://doi.org/10.4148/1944-9771.1102 xiao, j. j. (2008). applying behavior theories to financial behavior. in j.j. xiao (ed.), financial services review 32(1) 62 handbook of consumer finance research (p. 70). springer. xiao, j. j., & o’neill, b. (2018). propensity to plan, financial capability, and financial satisfaction. international journal of consumer studies, 42(5), 501-512. https://doi.org/10.1111/ijcs.12461 zyphur, m. j., li, w. d., zhang, z., arvey, r. d., & barsky, a. p. (2015). income, personality, and subjective financial well-being: the role of gender in their genetic and environmental relationships. frontiers in psychology, 6, 1493. https://doi.org/10.3389/fpsyg.2015.0149 3 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 financial services review, 32(4) 27 racial/ethnic disparities in financial advice seeking: a decomposition analysis di qing1 and miranda reiter2 abstract financial advice seeking is associated with many positive benefits for consumers. yet, most u.s. households are not working with a financial planner and research has pointed to the lack of racial and ethnic diversity among those who do. this study examines racial/ethnic disparities in using a financial planner. logistic regression analyses show that black and white consumers are more likely than asian and hispanic consumers to use financial planners for saving and investment decisions. a fairlie decomposition analysis shows racial/ethnic differences among the determinants that are associated with financial advice seeking. the differences in the determinants were large between white and hispanic consumers and much narrower between black and hispanic consumers. risk tolerance, objective financial knowledge, and income were the most important determinants to explain racial/ethnic differences in financial planner use. this study provides insight into possible barriers to working with financial planners for a diverse group of consumers. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation qing, d., & reiter, m. (2024). racial/ethnic disparities in financial advice seeking: a decomposition analysis. financial services review, 32(4), 27-50. introduction there is a growing demand for financial planners (u.s. bureau of labor statistics, 2023) as financial advice is becoming more important to ensure the success of long-term financial goals for consumers in the united states (harlow et al., 2020). financial advice is particularly important given the change in the types of employersponsored retirement plans offered to workers. over the past few decades, the industry has gradually shifted from offering employees defined benefit plans to defined contribution 1 corresponding author (dzq0007@auburn.edu). auburn university, auburn, usa 2 texas tech university, lubbock, texas, usa plans (estreicher & gold, 2007; myers & topoleski, 2021). in 1975, 27.2 million private sector employees reported having defined benefit plans, whereas 11.2 million reported having defined contribution plans. by 2021, less than 15 million employees reported having defined benefit plans, and over 85 million reported defined contribution plans (myers & topoleski, 2021). as a result, responsibility for funding employee retirement plans has shifted from the employer to the employee. one unintended consequence of this shift is lower levels of retirement preparedness. munnell and suden https://creativecommons.org/licenses/by-nc/4.0/ mailto:dzq0007@auburn.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(4) 28 (2006) note that employees with 401(k) contribution plans will likely have a shortfall in retirement assets. the authors note that this shortfall may be possibly due to the fact that employees are now faced with multiple investment decisions, such as diversification and rebalancing (munnell & suden, 2006). many of these decisions, which traditionally fall on employers, can leave employees feeling unequipped to make the right choices. financial advice can help consumers improve savings behavior (liu et al., 2019), stock market expectations (reiter & seay, 2022), investment outcomes (gaudecker, 2015), asset allocation (marsden et al., 2011), and the value of their assets (goetz et al., 2020; montmarquette & viennot-briot, 2015). specifically for long-term goals, financial planners, when compared to other financial professionals, play a more important role in helping clients (kim et al., 2018). as such, a financial planner may serve as an invaluable resource for increasing the overall financial wellbeing of consumers as well as preparing a diverse group of clients for long-term financial success. non-white consumers may face unique financial challenges that make financial planning even more relevant to addressing their needs. for black consumers, given historical events and systemic challenges, there are considerable financial barriers and constraints compared to their white peers (craemer et al., 2020). for example, black consumers report the lowest median household wealth when compared to other racial/ethnic groups (aladangady et al., 2023) and have lower ownership than white consumers in high-return investments such as stocks, real estate, and business assets (hanna et al., 2010). black households are also more susceptible to economic shocks, such as the great recession than white households, which may decrease retirement preparedness (wolff, 2023). in addition, black consumers have historically had lower retirement plan participation than whites (porto, 2016). similarly, research on hispanic households has found that they have lower retirement preparedness than white households. blanco et al. (2017) posit that this is partially attributed to hispanic consumers’ plans to continue working rather than retire, which could be attributed to lower savings rates during working years or a lack of retirement savings. moreover, hispanic consumers may feel greater financial responsibility towards their families. for example, hispanic individuals report a greater tendency than non-hispanic individuals to financially support family abroad through remittances (lópez-anuarbe et al., 2016). similarly, some asian consumers face familyoriented financial pressures, such as supporting aging parents (merril, 2023a). while the literature is scant on the experience of asian americans in financial services (hanna et al., 2015), asian consumers generally have lower risk tolerances, higher levels of aversion towards debt, and a greater sense of self-reliance compared to consumers from other racial/ethnic backgrounds (merril, 2023a). however, there is evidence that asian consumers save for retirement on par with white consumers, which is higher than that of black and hispanic consumers (yao, 2016). research investigating racial/ethnic differences in financial planner use has uncovered that the majority of financial planning clients are white individuals. despite this, there is also demand from individuals of color for financial advice. some research has shown that black consumers are more likely to hire a financial planner when factors such as income and education are controlled for in regression models (elmerick et al., 2002; hanna, 2011; reiter & qing, 2023; white & heckman, 2016). conversely, hispanic consumers are less likely to work with financial planners (hanna, 2011; white & heckman, 2016). current literature on asian consumers shows some mixed results. while some sources show that asian consumers are less likely to seek financial advice (hanna et al., 2015), other sources show that asian consumers are more likely to seek financial advice (white & heckman, 2016). ultimately, it may depend on the type of financial advice sought (white & heckman, 2016). while researchers have investigated racial/ethnic differences in seeking financial advice, it is important to understand what contributes to these differences. reiter and qing (2023) conducted a study on both gender and black-white racial qing & reiter 29 differences in seeking financial advice using 2012 data from the national financial capability study (nfcs) dataset. they employed a decomposition technique from jackson and lindley (1989). this technique allowed them to examine the effects of independent variables, which varied by race, and to isolate the effects, specifically, of being black or white on seeking financial advice. in other words, reiter and qing (2023) interacted the race variable with all other predictors and utilized the likelihood ratio test to compare the intermediate model (model without interaction) and the interacted model (model with interaction). the results showed a statistically significant difference between the two models and suggested that the interacted model was more appropriate. the results indicated that asking for financial advice is not determined by race or gender in and of itself but by other variables. however, the decomposition technique from jackson and lindley (1989) could not capture specific differences in predictors when comparing racial groups; specifically, the results could not show which predictor contributes the most difference in seeking financial advice between the two groups. as such, fairlie's (2005) decomposition technique is required. the fairlie (2005) decomposition method can estimate the significance of various observed characteristics that account for differences in financial planner use between groups. this study aims to investigate racial/ethnic differences in using financial planners to make investment and saving decisions. we use the fairlie (2005) decomposition technique to examine if racial/ethnic differences are attributed to various consumer and household characteristics or unobserved variables. specifically, we examine which determinants help explain racial/ethnic differences in financial planner use and which determinants contribute to the gap in financial planner use between racial/ethnic groups. in this study, six pairwise comparisons were conducted: (1) blacks versus whites, (3) blacks versus hispanics, (3) blacks versus asians/others, (4) whites versus hispanics, (5) whites versus asians/others, and (6) asians/others versus hispanics. we calculate the part of the observed differences in the utilization of financial planners attributable to variations in household and economic characteristics. this paper contributes to the literature in a few important ways. first, we investigate the racial/ethnic differences in seeking financial advice (i.e., using a financial planner) for saving and investment decisions using the more recent 2016 and 2019 waves of the survey of consumer finances (scf). the data from the scf allow us to include net worth as a variable, unlike the work from reiter and qing (2023), which does not include this important variable, due to a limitation with the nfcs dataset. also, the financial advice seeking question from the nfcs asks about financial planner use within the past five years, whereas the scf uses a broader question. second, we employ the fairlie decomposition method to understand the significance of the racial/ethnic differences in financial advice seeking characteristics for six pairwise groups, which, to the best of our knowledge, has not been examined before. as such, this paper provides additional insights beyond the findings of reiter and qing (2023), who examined disparities in financial planner use between black and white consumers. the paper will be organized as follows. first, we introduce theoretical considerations, and then, we discuss our methodology, including the dataset, sample, dependent variable, independent variables, and the empirical model specification. next, we provide an analysis of the results, which include descriptive statistics, logistic regression analyses, and decomposition analyses. finally, we discuss the results with implications, suggestions for future research, and limitations. theoretical considerations according to the life-cycle hypothesis by ando and modigliani (1963), seeking help from financial professionals assists individuals with making financial decisions that maximize utility over the life cycle. the pattern of wealth accumulation follows a “hump shape”; in other words, individuals accumulate wealth when they are young and distribute wealth when they are old. to allocate resources optimally, asking for help from financial professionals helps smooth utility during a lifetime. according to economic theory, there should be no racial/ethnic financial services review, 32(4) 30 differences when seeking help from financial professionals. however, economic theories may not explain why those who belong to non-white groups behave differently from whites, even if they have the same characteristics, such as income, net worth, or financial knowledge (shin & hanna, 2015). grable and joo (1999) introduced the financial help-seeking framework as an expansion of suchman’s (1966) help-seeking framework. the framework posits that seeking financial advice happens in five steps: exhibiting financial behaviors, analyzing one's financial behaviors, identifying the causes of the behaviors, deciding to seek help, and finally choosing among help options. while the framework does not explain racial/ethnic differences in seeking financial help, it establishes the process that individuals take to arrive at the action of seeking advice, and it highlights some of the characteristics attributed to advice seekers. contrary to some other studies, grable and joo (1999) found that those most likely to seek a third party for financial assistance had undesirable financial behaviors and financial stressors. grable and joo (1999) explained that seeking financial help could be a coping response. using the 2012 national financial capability study, fan (2021) used stress-coping theories to understand more about individuals’ propensity to engage in professional financial advice. like grable and joo (1999), fan found that having a recent experience with financial stress is positively associated with seeking assistance from a financial advisor. this association could explain the greater propensity for black consumers to seek financial advice, as they have been found to experience greater financial challenges when compared to others in general (lim et al., 2014). however, it does not explain the lack of financial professional use among hispanic consumers, as they also experience financial issues similar to those of black consumers (martin & dwyer, 2021). it is imperative to understand more about these disparities in seeking financial advice. based on prior literature and theoretical considerations, we propose the following hypotheses: h1: white consumers will be less likely to seek financial advice for savings and investing decisions when compared to black consumers. h2a: hispanic consumers will be less likely to seek financial advice for savings and investing decisions when compared to black consumers. h2b: hispanic consumers will be less likely to seek financial advice for savings and investing decisions when compared to white consumers. h3a: asian consumers will be less likely to seek financial advice for savings and investing decisions when compared to black consumers. h3b: asian consumers will be less likely to seek financial advice for savings and investing decisions when compared to white consumers. h4: the factors associated with seeking financial advice for savings and investing decisions will differ across racial/ethnic groups. h5: the determinants that explain the racial/ethnic disparities in seeking financial advice for savings and investing decisions will differ between racial/ethnic groups. methods dataset and sample this study utilizes the 2016 and 2019 waves of the survey of consumer finances (scf), a nationally representative triennial cross-sectional survey of families in the united states. the scf, sponsored by the federal reserve board, collects consumer data on various topics. for the 2016 wave of the survey, 6,248 households were interviewed, and 5,777 households were interviewed for the 2019 survey. multiple imputation is used within the scf to provide respondent privacy and avoid missing data. multiple imputation produces five sets of data, called implicates, for each respondent, representing a range of likely responses (lindamood et al., 2007). previous literature (lindamood et al., 2007; rubin, 1987) recommends using all five implicates via repeated imputation inference (rii). as such, the current study uses rii to apply all five implicates (montalto & sung, 1996). according to pence (2015), the “scfcombo” command in stata qing & reiter 31 software combines imputation uncertainty and bootstrapped standard errors; in other words, this command applies rii to improve the accuracy of the estimation. previous studies using the survey of consumer finances (chang, 2005; lei & kordes, 2020; white & heckman, 2016) applied weights for descriptive statistics. however, applying weights to logistic regression models has yielded more conservative results (lindamood et al., 2007). therefore, we applied weights to our descriptive analyses but not to the regression models, consistent with previous research (e.g., shin & hanna (2015)). dependent variable there were two questions asked in the 2016 and 2019 waves of the scf related to financial advice seeking: (1) “what sources of information do you (and your husband/wife/partner) use to make decisions about borrowing or credit?” and (2) what sources of information do you (and your husband/wife/partner) use to make decisions about saving and investments?”. for the current study, the authors are mostly concerned with consumers’ decisions related to using a financial planner in a more traditional sense, which would include seeking help for investments and savings rather than debt or credit. the scf provides many answer choices for information sources, including service professionals such as lawyers, accountants, bankers, brokers, and financial planners. however, the “financial planner” answer choice was the only one used for the current study. the responses were coded as 1 if respondents chose “financial planner” and 0 if they did not. if respondents chose “other” or “inappropriate,” the responses were coded as missing values. independent variables the factors associated with financial planner use have been well-investigated. independent variables included race/ethnicity, gender, age, marital status, income, net worth, risk tolerance, investment horizon, subjective knowledge, objective knowledge, household size, educational attainment, homeownership, employment status, and emergency account access. race/ethnicity, as categorized in the survey of consumer finances, includes black/african american, hispanic/latino, asian, american indian/alaska native, native hawaiian/pacific islander, and white. for the current study, race/ethnicity was grouped into four categories: white, black, hispanic, and asian/other. the asian/other category is made up of mostly asian respondents (hanna & lindamood, 2015) but also includes american indian/alaska native, native hawaiian/pacific islander, and anyone who chose “another race.” combining these groups into one was necessary due to the low sample sizes of each group alone. gender has been identified as a predictor for financial advice. for example, women are more likely to pay for financial advice compared to men and are more likely to seek advice for retirement planning (finke et al., 2011; joo & grable, 2001). if respondents were women (or female, as described in the survey), the value was coded as 1, and 0 if respondents identified as men (male). age is associated with financial advice seeking (robb et al., 2012). there is evidence that older consumers are more likely to seek specific types of financial advice, such as investment and savings advice (hackethal et al., 2012; lachance & tang, 2012). for the current study, age was treated as a continuous variable. in addition, some previous research has indicated that the relationship between age and financial planner use may be nonlinear (white & heckman, 2016), and therefore, age-squared was also included as an independent variable. marriage has been long understood as a financially-advantaged status compared to being single. as such, married individuals tend to do better financially than those who are not married and are more likely to work with financial planners than those who are single (lachance & tang, 2012; robb et al., 2012). marital status was used as a dummy variable; those who were married were coded as 1, and all others were coded as 0. other household characteristics, such as household size (elmerick et al., 2002), educational attainment (chatterjee & zahirovicherbert, 2010), and employment status (elmerick et al., 2002), have been associated with using a financial services review, 32(4) 32 financial planner. household size was used as a continuous variable that indicated the number of people within the same household. educational attainment was organized into the following categories: (a) lower than high school, (b) a high school diploma, (c) some college, (d) a bachelor’s degree, and (d) a graduate degree. employment status was categorized into four groups for the current study: (a) unemployed, (b) employed (e.g., working for someone else), (c) selfemployed, and (d) retired. financial variables such as wealth (harlow et al., 2022) and income (cummings & james, 2014; finke et al., 2011; joo & grable, 2001) are positively associated with seeking financial advice. income was made into a categorical variable and comprised three groups: (1) less than $50,000; (2) between $50,000 and $99,999; and (3) $100,000 or higher. net worth was used as a continuous variable. if net worth was greater than 0, the log-value was used. if net worth was smaller than 0, the log (0.01) was utilized. homeownership (hanna, 2011) and having an emergency fund (white & heckman, 2016) are predictors associated with using a financial planner (hanna, 2011). homeownership and emergency access were used as binary dummy variables. emergency access is determined by answering the following question: “in an emergency could you (or your husband/wife/partner) get financial assistance of $3,000 or more from any friends or relatives who do not live with you?” if the respondent answered “yes,” the value was coded as 1, and 0 if the respondent answered “no.” risk tolerance has been correlated with a higher likelihood of seeking financial advice in numerous studies (chang, 2005; joo & grable, 2001; moreland, 2018; white & heckman, 2016). risk tolerance was categorized into four groups: (1) not willing to take any financial risks; (2) willing to take average financial risks expecting to earn average returns; (3) willing to take above average financial risks expecting to earn above average returns; and (4) willing take substantial financial risks expecting to earn substantial returns. investment time horizons may also impact financial help-seeking behavior, as households with long-term investment horizons are likelier to work with a financial planner than households with intermediateor short-term investment horizons (white & heckman, 2016). investment time horizon was used as a categorical variable and included the following choices: (a) next few months; (b) next year; (c) next few years; (d) next 5-10 years; and (e) longer than 10 years. objective and subjective financial knowledge predict the likelihood of working with a financial planner, although results are mixed depending on whether the association is positive or negative. there tends to be more support in the literature for a positive association between objective financial knowledge and financial advice seeking (alyousif & kalenkoski, 2017; calcagno & monticone, 2015; seay et al., 2016), but some research has found the opposite (hsu, 2022; sommer & lim, 2022). kramer (2016) found that subjective financial knowledge is associated with a lower likelihood of seeking financial advice, but others report a positive association (fan, 2021). the current study measured objective financial knowledge using the big three scale (lusardi & mitchell, 2011), which evaluates respondents’ knowledge of stocks, interest rates, and inflation. if respondents answered a question correctly, one point was allotted; therefore, a summation scale ranged from 0 to 3, where 0 indicated that respondents had low objective financial knowledge and 3 indicated that respondents had high objective financial knowledge. subjective financial knowledge was measured by the question: “on a scale from 0 to 10, where zero is not at all knowledgeable about personal finance and ten is very knowledgeable about personal finance, what number would you be on the scale?” subjective financial knowledge was treated as a continuous variable. empirical model specification a binomial logistic regression model was used among a pooled sample to investigate the determinants of financial advice seeking when making decisions about saving and investments. four additional logistic regression models were used to examine racial/ethnic groups separately. the main assumption of the regression or logit model indicates that all racial/ethnic groups have the same slope and intercept because it is an identical independent variable. however, the qing & reiter 33 slope and intercept can change depending on the specific group due to different characteristics based on the group. this study utilizes fairlie decomposition techniques (2005), developed by blinder-oaxaca (1994), which capture intercept and slope differences in racial/ethnic groups. fairlie’s decomposition technique is an ideal research method to identify inter-group differences. it allows us to address the explained and unexplained segments and quantify the significance levels of different variables. the explained segments indicate how well the observed variables explain financial advice seeking when making decisions about savings and investments. in contrast, the unexplained segments demonstrate how the unobserved variables, which are not included in this model, explain financial advice seeking. this research method has been previously used to explore racial/ethnic differences in financial behaviors. for example, shin and hanna (2015) utilized a decomposition analysis to examine the racial/ethnic differences in high-return investment ownership after the great recession. lee and kim (2022) investigated racial/ethnic differences in financial knowledge using decomposition techniques. however, to the best of our knowledge, this technique has not yet been applied to racial/ethnic differences in seeking financial advice. for the current study, the decomposition of the racial/ethnic groups was estimated as follows: 𝐻 = 𝐹(𝑋�̂�) (1) �̄�1 − �̄�2 = [ 1 𝑁1 ∑ 𝐹(𝑋𝑖 1�̂�1) −𝑁1 𝑖=1 1 𝑁2 ∑ 𝐹(𝑋𝑖 2�̂�1)𝑁2 𝑖=1 ] + [ 1 𝑁2 ∑ 𝐹(𝑋𝑖 2�̂�1) −𝑁2 𝑖=1 1 𝑁2 ∑ 𝐹(𝑋𝑖 2�̂�2)𝑁2 𝑖=1 ] (2) the first equation indicates general logistic regression. the second equation addresses the differences between two groups. more specifically, this paper examined differences in financial planner use between blacks and whites, hispanics and whites, asians/others and whites, blacks and hispanics, blacks and asians/others, and hispanics and asians/others. the second equation �̄�𝑗(�̄�1, �̄�2) indicates the average probability of using a financial planner for different racial/ethnic groups (j); 𝑁𝑗(𝑁1, 𝑁2) represents the sample size for group 1 and group 2; �̂�𝑗(�̂�1, �̂�2) shows the vector of coefficient estimates for group 1 and group 2; 𝑋𝑖 𝑗 (𝑋𝑖 1, 𝑋𝑖 2) is a row vector of average values of the independent variables. [ 1 𝑁1 ∑ 𝐹(𝑋𝑖 1�̂�1) − 1 𝑁2 ∑ 𝐹(𝑋𝑖 2�̂�1)𝑁2 𝑖=1 𝑁1 𝑖=1 ] (3) [ 1 𝑁2 ∑ 𝐹(𝑋𝑖 2�̂�1) − 1 𝑁2 ∑ 𝐹(𝑋𝑖 2�̂�2)𝑁2 𝑖=1 𝑁2 𝑖=1 ] (4) following fairlie’s techniques (2005), equation (3) is the first part of equation (2) and measures the racial/ethnic gap due to the different distribution of groups 𝑋𝑖 𝑗 (𝑋𝑖 1, 𝑋𝑖 2). equation (4) is the second part of equation (2) and measures the “unexplained” portion of the racial/ethnic gap due to unobservable or unmeasurable endowments, such as data limitations. this study focuses on explaining the “explained” part from the first part of equation (2). since most of the respondents in the scf are white, fairlie’s measurement provides for an adjustment in sample selection bias due to an uneven sample size. in this study, we specified 100 decomposition replications and utilized the mean of 100 estimations to conduct the differences across racial/ethnic groups. the degree to which a variable contributes to the racial/ethnic gap in financial planner use could vary slightly based on the order in which an independent variable is placed in the model (fairlie, 2005). for the decomposition analyses in this paper, the independent variables will be ordered as follows: gender, age, marital status, income, net worth, risk tolerance, investment horizon, subjective financial knowledge, objective financial knowledge, household size, educational attainment, homeownership, employment status, and emergency account access. results descriptive statistics table 1 shows the percentage of populations using a financial planner for saving and investment decisions by race and ethnicity. for the 2016 wave of the scf, about 24% of black, 17% of hispanic, 30% of asian/other, and 38% of white households used financial planners for financial services review, 32(4) 34 saving and investing decisions. usage was similar for the 2019 wave at 23%, 21%, 32% and 37%, respectively. when combining the 2016 and 2019 waves, about 24% of black, 19% of hispanic, 31% of asian/other, and 38% of white consumers used financial planners for saving and investing decisions. table 1. mean levels of financial planner use for savings and investments decisions by race/ethnicity racial/ethnic category black hispanic asian/other white total survey year 2016 percent of group using 24% 17% 30% 38% 34% n 204 106 97 1,723 2,130 survey year 2019 percent of group using 23% 21% 32% 37% 33% n 174 115 103 1,537 1,929 pooled years (2016 and 2019) percent of group using 24% 19% 31% 38% 34% n 378 221 200 3,260 4,059 note: weighted proportion. table 2 shows the characteristics of respondents who used financial planners for savings and investment decisions in the combined 2016 and 2019 waves of the scf. about 80% of the respondents were white, while the rest were 9% black, 5% hispanic and 5% asian/other. about 58% of black respondents were women, compared to 39% white, 49% hispanic, and 35% asian/other respondents. white respondents had the oldest mean age of 56, while hispanic respondents were the youngest, with a mean age of 45. white, hispanic, and asian/other respondents were more likely to be married than black respondents. white and asian/other respondents tended to have higher levels of income. as for risk tolerance, black and hispanic respondents had higher proportions who reported taking no risk or taking substantial risk compared to white and asian/other respondents. in terms of investment horizon, a higher percentage of white (32%), hispanic (27%), and asian/other (30%) fell into the “5-10 years investment horizon" category. in comparison, more black respondents (28%) chose “next few years” as their investment time horizon. in addition, white and asian/other respondents had higher subjective and objective financial knowledge scores than black and hispanic respondents. white and asian/other respondents had similar educational attainment rates at the bachelor’s degree or higher level, with over 70% of their samples represented in this category. black and hispanic respondents were also similar to one another in that 44% and 47% of these groups had attained a bachelor’s degree or higher. about 80% of the sample owned homes, and about 75% were employed. about 78% had access to emergency funds, with black and hispanic respondents having the least access compared to the other groups. qing & reiter 35 table 2. descriptive statistics of respondents using financial planners by race/ethnicity pooled sample n=4,059 black n= 378 hispanic n=221 asian/other n=200 white n= 3,260 variables mean s.d. mean s.d. mean s.d. mean s.d. mean s.d. race/ethnicity white 0.8031 0.3977 black 0.0931 0.2906 hispanic 0.0544 0.2269 asian/other 0.0494 0.2167 female 0.4102 0.4919 0.5788 0.4944 0.4932 0.5011 0.3533 0.4792 0.3885 0.4875 age 54.9076 14.7980 49.6646 14.1296 45.3765 12.9053 49.9232 14.2646 56.4682 14.5806 married 0.6738 0.4689 0.4132 0.4931 0.5656 0.4968 0.6836 0.4662 0.7108 0.4535 income less than $50k 0.1954 0.3966 0.4434 0.4974 0.3339 0.4727 0.1916 0.3946 0.1575 0.3643 $50k-$99,999 0.2069 0.4051 0.2741 0.4466 0.3050 0.4614 0.1786 0.3840 0.1942 0.3956 $ 100k above 0.5977 0.4904 0.2825 0.4508 0.3611 0.4814 0.6297 0.4841 0.6483 0.4776 net worth (log value) 12.8774 4.9852 8.5356 6.6113 10.0247 5.6837 12.6897 5.2382 13.5858 4.3346 risk tolerance no risk 0.1620 0.3685 0.2942 0.4563 0.2941 0.4567 0.1507 0.3587 0.1384 0.3454 average risk 0.4948 0.5000 0.4307 0.4958 0.3937 0.4897 0.4860 0.5011 0.5096 0.5000 above average risk 0.2898 0.4537 0.1931 0.3953 0.2262 0.4193 0.3134 0.4650 0.3038 0.4600 substantial risk 0.0535 0.2250 0.0820 0.2747 0.0860 0.2810 0.0499 0.2183 0.0482 0.2141 investment horizon next few months 0.1101 0.3130 0.2196 0.4145 0.2380 0.4268 0.1277 0.3346 0.0876 0.2828 next year 0.0929 0.2903 0.1492 0.3568 0.1457 0.3536 0.0998 0.3005 0.0823 0.2749 next few years 0.2513 0.4338 0.2783 0.4488 0.2344 0.4246 0.2605 0.4400 0.2488 0.4324 next 5-10 years 0.3093 0.4623 0.2317 0.4225 0.2688 0.4443 0.2954 0.4574 0.3219 0.4673 longer than 10 years 0.2364 0.4249 0.1212 0.3268 0.1131 0.3175 0.2166 0.4129 0.2593 0.4383 subjective knowledge 7.8257 1.7832 7.4884 1.9814 7.3955 2.0136 7.5798 1.9066 7.9091 1.7243 objective knowledge 2.5424 0.7204 2.0466 0.8980 2.1520 0.8814 2.5788 0.6600 2.6241 0.6544 household size 2.5599 1.3205 2.3799 1.3330 3.1548 1.5813 2.9371 1.5826 2.5172 1.2670 educational attainment less than high school 0.0184 0.1345 0.0667 0.2498 0.0814 0.2741 0.0120 0.1091 0.0090 0.0942 high school 0.1021 0.3029 0.1497 0.3573 0.1864 0.3903 0.0539 0.2264 0.0939 0.2917 financial services review, 32(4) 36 some college 0.1970 0.3978 0.3127 0.4642 0.2896 0.4546 0.1667 0.3736 0.1792 0.3835 bachelor 0.2996 0.4581 0.2429 0.4294 0.2489 0.4333 0.2345 0.4248 0.3136 0.4640 graduate 0.3828 0.4861 0.2280 0.4201 0.1937 0.3961 0.5329 0.5002 0.4044 0.4908 homeownership 0.8087 0.3934 0.5397 0.4991 0.6054 0.4899 0.7595 0.4285 0.8567 0.3505 employment unemployed 0.0251 0.1565 0.0556 0.2294 0.0181 0.1336 0.0549 0.2283 0.0202 0.1409 employee 0.4716 0.4993 0.5799 0.4942 0.6561 0.4761 0.5160 0.5010 0.4438 0.4969 self-employed 0.2768 0.4474 0.1413 0.3488 0.1810 0.3859 0.2854 0.4527 0.2984 0.4576 retired 0.2265 0.4186 0.2233 0.4170 0.1448 0.3527 0.1437 0.3517 0.2376 0.4257 access to emergency funds 0.7775 0.4160 0.5624 0.4967 0.6389 0.4814 0.7754 0.4183 0.8120 0.3908 note: weighted; 2016 and 2019 survey of consumer finances logistic regression analyses table 3 shows the results of five logistic regression models. the first model represents a binomial regression for the pooled sample, including all racial/ethnic groups, with the black respondent category as the reference group. we also ran a binomial regression for the pooled sample with white as the reference group (see appendix). the other four regression models represent the findings related to the factors significant for financial planner use when making decisions about saving and investments within the black, hispanic, asian/other, and white groups. in the pooled sample, hispanic and asian/other respondents were significantly less likely to work with a financial planner when making decisions about saving and investments than black respondents. no significant difference was found when comparing black consumers to white consumers. gender, age, income, net worth, risk tolerance, investment horizon, financial knowledge, educational attainment, homeownership, and access to emergency funds were all positively associated with seeking financial advice. the authors conducted a joint hypothesis test and likelihood ratio test (lrt) to test the nonlinear effect of age. the coefficient of the squared term is significant and indicates that there is evidence to suggest a nonlinear association. also, when comparing the full model (age-squared included) and nested model (agesquared excluded), the result of the likelihood ratio test was statistically significant (p = 0.0000), indicating that age-squared is an important predictor when determining adviceseeking behavior. qing & reiter 37 table 3. binomial logistic regressions of financial planner use pooled sample black hispanic asian/other white variables coef. s.e. coef. s.e. coef. s.e. coef. s.e. coef. s.e. race/ ethnicity (ref.= black) white -0.0435 0.0588 hispanic -0.1897** 0.0708 asian/other -0.2794** 0.1011 female (ref.= male) 0.3284*** 0.0324 0.2601** 0.0977 0.0738 0.1126 0.1776 0.1749 0.3784*** 0.0538 age 0.0343*** 0.0072 0.0795*** 0.0188 0.0673* 0.0312 0.0308 0.0337 0.0292** 0.0099 age squared 0.0003*** 0.0001 0.0007*** 0.0001 -0.0008* 0.0003 -0.0002 0.0003 -0.0002** 0.0001 married (ref.= not married) -0.0706 0.0472 0.1885 0.0976 -0.3143* 0.1520 -0.5766* 0.1891 -0.0427 0.0600 income (ref.= less than $50k) $50k-$99,999 0.2189*** 0.0487 0.1173 0.1244 0.3837* 0.1831 0.1271 0.2409 0.2003** 0.0667 $100k and above 0.5294*** 0.0506 0.4951** 0.1801 0.6671** 0.2116 0.4884 0.2949 0.4893*** 0.0687 net worth 0.0225*** 0.0052 0.0001 0.0087 0.0070 0.0164 0.0097 0.0225 0.0350*** 0.0067 risk tolerance (ref.=not willing) average risk 0.7900*** 0.0480 0.7158*** 0.1468 0.6853*** 0.1368 0.7487*** 0.2118 0.7993*** 0.0553 above average risk 0.7823*** 0.0518 0.6966*** 0.1469 0.7672*** 0.1741 0.6821** 0.2420 0.7932*** 0.0618 substantial risk 0.5862*** 0.0919 0.8746*** 0.1896 0.8479** 0.2677 0.2795 0.4098 0.4911*** 0.1114 invest horizon (ref.=next few months) next year 0.0105 0.0591 -0.0909 0.1763 -0.2527 0.2072 -0.0690 0.3401 0.0938 0.0774 next few years 0.1714** 0.0560 0.0397 0.1243 -0.1782 0.1894 0.0353 0.2884 0.2694*** 0.0683 next 5-10 years 0.3205*** 0.0605 0.0219 0.1509 0.2480 0.2148 -0.0876 0.3232 0.4359*** 0.0692 longer than 10 years 0.3778*** 0.0618 0.3793* 0.1673 0.1078 0.2381 -0.0147 0.3919 0.4751*** 0.0686 subjective knowledge 0.0172* 0.0080 0.0608** 0.0212 0.0690** 0.0246 0.0560 0.0468 -0.0041 0.0103 objective knowledge 0.1770*** 0.0254 0.0142 0.0577 0.0741 0.0752 0.3352** 0.1131 0.2147*** 0.0315 household size -0.0353* 0.0145 -0.0921* 0.0403 0.0134 0.0573 0.0270 0.0644 -0.0331 0.0186 educational attainment (ref. = less than high school) high school 0.5133*** 0.1091 -0.1838 0.2087 0.6932* 0.3344 0.7375 3.9594 0.6930*** 0.1500 some college 0.6238*** 0.1011 0.0575 0.1873 0.7778*** 0.2385 1.1054 3.9382 0.7575*** 0.1489 bachelor 0.8328*** 0.1031 0.4592* 0.1954 1.1322*** 0.2261 0.7006 3.9447 0.9628*** 0.1606 graduate 0.9187*** 0.1025 0.4118 0.2159 1.1356*** 0.3134 0.8684 3.9347 1.0648*** 0.1579 homeownership (ref.=no) 0.1552*** 0.0486 0.1994 0.1201 0.2511 0.1397 0.3785 0.2478 0.1252* 0.0623 employed (ref.=unemployed) employee 0.0093 0.0990 -0.2892 0.2378 0.8489 1.2875 -0.0384 0.2724 0.0139 0.1186 self-employed 0.0397 0.1013 0.1145 0.2411 0.9846 1.2813 0.2671 0.2994 -0.0236 0.1155 retired 0.1318 0.1046 -0.2158 0.2830 1.4831 1.2767 0.2627 0.3974 0.0895 0.1191 financial services review, 32(4) 38 have emergency funds (ref.= no) 0.1014* 0.0433 0.3116** 0.0986 -0.0153 0.1414 0.0617 0.1769 0.0725 0.0530 intercept 4.4140*** 0.2608 4.3920*** 0.6276 5.7947*** 1.5528 -4.7735 4.0514 4.6304*** 0.2751 sample size 4,059 378 221 200 3,260 r-squared 0.1172 0.0966 0.1298 0.0889 0.1085 note: unweighted analysis, 2016 & 2019 scf. *p<0.05; **p<0.01; ***p<0.001 black respondents. among black respondents who use a financial planner, being a woman, age, being married, having an income that was $100,000 or more, being willing to take at least some risk, having a time horizon greater than 10 years, subjective financial knowledge, holding at least a bachelor’s degree, and having access to emergency funds were all positively associated with using a financial planner when making decisions about saving and investing. household size and age squared were negatively associated with using a financial planner for saving and investment decisions. hispanic respondents. for hispanic respondents, age, having an income of $50,000 or more, having at least some risk tolerance, subjective financial knowledge, and having at least a high school diploma were associated positively with financial planner use. age squared and marriage were negatively associated with using a financial planner. asian/other respondents. for the asian/other group, having average or above average risk compared to no risk was associated with seeking advice. in addition, objective financial knowledge was positively associated, while marriage was negatively associated. white respondents. among white respondents, being a woman, age, having income greater than $50,000, having at least some risk tolerance, having an investment time horizon of the next few years or more, and objective financial knowledge were associated positively with financial advice seeking. net worth was also positively associated with using a financial planner. as the findings show, there were differences in the characteristics of seeking financial advice among racial/ethnic groups. as such, decomposition estimation was used to investigate further and measure the intergroup differences among these variables. decomposition analyses according to fairlie (2005), when conducting decomposition analyses, the reference group should typically be the group with the lower rate of financial planner use at the descriptive level. for example, the hispanic group served as the reference group compared with other groups since only 19% of hispanic respondents from the pooled sample reported using a financial planner when making decisions about saving and investments. black respondents had lower rates (24%) than white respondents (38%), and as such, they were the reference group. asian/other respondents had a rate of 31% and, therefore, were used as the reference group compared to white respondents. table 4 represents the decomposition analyses utilizing fairlie’s estimation (2005) across three pairwise comparisons between racial/ethnic groups (e.g., black respondents versus other groups). specification 1 shows the comparison between black and white respondents. the total difference in using financial planners when making decisions about saving and investments between these two groups was 0.1399 based on the mean predictions of each group. the explained difference to the total difference was 110%, which indicates that the observed differences in respondents’ characteristics explained approximately 110% of the difference. if white respondents had the same characteristics as black respondents, their probability of seeking financial advice for saving and investment decisions would be lower than that of black respondents. because the total difference is greater than 100%, this indicates that the qing & reiter 39 unexplained difference is having a negative effect on the propensity to use financial planners. income accounted for 23.7% of the racial/ethnic gap, followed by 20.2% for net worth and 19.3% for risk tolerance. negative percentages indicate a narrowing effect of that factor on the racial/ethnic gap in financial planner use. gender contributed to narrowing the racial gap by 10.86% between black and white respondents. as such, this means that gender narrows the gap in financial planner use between black and white consumers. table 4. decomposition analysis of financial planner use (1) black vs. white (2) black vs. hispanic (3) black vs. asian/other variables contribution to difference percent of explained difference contribution to difference percent of explained difference contribution to difference percent of explained difference female -0.0167*** -10.86% 0.0018*** 13.21% -0.0082 -11.46% age 0.0077*** 5.01% -0.0016 -11.85% -0.0018 -2.50% marital status -0.0029 -1.92% -0.0069* -51.12% -0.0441*** -61.73% income 0.0365*** 23.71% -0.0046*** -34.09% 0.0359*** 50.29% net worth 0.0310*** 20.15% -0.00001 -0.09% 0.0102 14.27% risk tolerance 0.0297*** 19.29% 0.0075*** 55.49% 0.0251*** 35.13% investment horizon 0.0152*** 9.88% 0.0006 4.79% -0.0002 -0.21% subjective financial knowledge -0.0005 -0.35% 0.0034*** 25.60% 0.0017** 2.41% objective financial knowledge 0.0218*** 14.17% 0.00001 0.05% 0.0290*** 40.63% household size -0.0004*** -0.27% 0.0108*** 80.06% 0.0030 4.17% education 0.0226*** 14.70% 0.0061** 44.93% 0.0023 3.20% homeownership 0.0072*** 4.70% -0.0010** -7.45% 0.0173** 24.27% employment -0.0002** -0.14% 0.0010 7.31% -0.0013 -1.76% emergency access 0.0034** 2.23% -0.0036*** -27.06% 0.0022 3.13% total difference 0.1399 0.0493 0.0713 explained difference 0.1538 0.0135 0.0714 unexplained difference -0.0138 0.0359 -0.00004 percent of explained difference to total difference 110% 27% 100% note: *p<0.05; **p<0.01; ***p<0.001; decomposition analysis using 2016 & 2019 scf. the reference group is the one that has the lower rate of financial planner usage. financial services review, 32(4) 40 specification 2 in table 4 shows the decomposition analysis between black and hispanic respondents. the probability of using a financial planner between black and hispanic respondents was much smaller than that of the other racial/ethnic groups versus black respondents. the total difference in the predicted mean likelihood of seeking financial advice was estimated at 0.0493. however, the characteristics of respondents accounted for 27% of the gap in the rate of using financial planners between these two groups. if hispanic respondents had the same characteristics as black respondents, the probability of using a financial planner when making decisions about saving and investments would be lower than that of blacks. household size (80.1%), risk tolerance (55.5%), and education (44.9%) were the most important factors. the largest contributors to narrowing the gap were marital status (51.1%) and income (34.1%). the results imply that, for instance, because hispanic respondents had higher income levels than blacks, if they had the same level of income as black respondents, they would have a lower probability of using a financial planner. specification 3 in table 4 shows the contribution of each characteristic of households to the racial/ethnic difference financial advisor use between black and asian/other respondents; the total difference was 0.0713 based on the mean predictions of each group. the explained difference to the total difference was 100%. that is, if black respondents had the same respondent characteristics as asian/other respondents, the probability of using financial planners to make saving and investment decisions would be similar. the most important factor was income, which contributed to the racial/ethnic gap by 50.29%. objective financial knowledge contributed 40.63% to the gap. in addition, marital status contributed to narrowing the financial advisor use gap by 61.73%, which indicates that if black respondents had the same marital status as asian/other respondents, they would have a lower probability of using a financial planner. table 5 represents the decomposition analyses comparing three pairwise groups: (1) whitehispanic, (2) white-asian/other, and (3) asian/other-hispanic. specification 1 of table 5 shows the decomposition estimation between white and hispanic respondents. the total difference in using financial planners when making decisions about saving and investments between these two groups was 0.1892 based on the mean predictions of each group. the explained difference to total difference was 92%, which indicates that respondents’ characteristics accounted for 92% of the gap in using financial advisors. if hispanic respondents had the same characteristics as white respondents, the probability of using financial planners would be lower. in addition, the results suggest that most of the racial/ethnic gap in having financial planners when making decisions about saving and investments between white and hispanic respondents was determined by the difference in characteristics of respondents in racial/ethnic groups, not by racial/ethnicity itself. risk tolerance (22.30%) and educational attainment (19.20%) contributed the most among the predictors. specification 2 of table 5 shows the decomposition of white-asian/other differences. however, the characteristics of respondents accounted for 32% of the gap in the rate of using financial planners between these two groups. if asian/other respondents had the same characteristics as white respondents, the probability of using financial planners would be lower than white respondents. age was the most important factor in explaining differences and contributed 30.73%, followed by objective financial knowledge at 20.53%. in addition, because asian/other respondents had more education than white respondents, if they had the same level of education as white respondents, they would have a lower probability of using financial planners. when comparing asians/others to hispanics (specification 3), the explained difference to total difference was 87%. the result indicates that if hispanic respondents had the same characteristics as asian/other respondents, the probability of using financial planners would be lower than asian/other respondents. risk tolerance (30.31%), income (28.68%), and objective financial knowledge qing & reiter 41 (25.20%) were the biggest contributors to the gap, while marital status and being female narrowed the gap. that is, if hispanics had the same marital status as asian/other respondents, they would have a lower probability of using financial planners. table 5. decomposition analysis of financial planner use (1) white vs. hispanic (2) white vs. asian/other (3) asian/other vs. hispanic variables contribution to difference percent of explained difference contribution to difference percent of explained difference contribution to difference percent of explained difference female -0.0112*** -6.42% 0.0020*** 9.04% -0.0058 -5.47% age 0.0114*** 6.56% 0.0067*** 30.73% 0.0045*** 4.22% marital status -0.0011 -0.64% 0.0003 1.20% -0.0208*** -19.70% income 0.0312*** 17.93% -0.0003*** -1.41% 0.0303*** 28.68% net worth 0.0224*** 12.85% 0.0029*** 13.02% 0.0067 6.31% risk tolerance 0.0388*** 22.30% 0.0025*** 11.51% 0.0320*** 30.31% investment horizon 0.0174*** 10.00% 0.0023*** 10.43% -0.0002 -0.21% subjective financial knowledge -0.0007 -0.42% -0.0004 -1.77% 0.0045** 4.31% objective financial knowledge 0.0204*** 11.71% 0.0045*** 20.53% 0.0266*** 25.20% household size 0.0037*** 2.10% 0.0032*** 14.37% -0.0005 -0.49% education 0.0334*** 19.20% -0.0065*** -29.77% 0.0144 13.69% homeownership 0.0056*** 3.20% 0.0032*** 14.36% 0.0125** 11.84% employment 0.0009** 0.50% 0.0011** 5.12% 0.0001 0.06% emergency access 0.0021** 1.18% 0.0003* 1.42% 0.0012 1.11% total difference 0.1892 0.0686 0.1207 explained difference 0.1740 0.0220 0.1055 unexplained difference 0.0153 0.0466 0.0151 percent of explained difference to total difference 92% 32% 87% note: decomposition analysis using 2016 & 2016 scf. the reference group is the group that has the lower rate of financial planner usage. *p<0.05; **p<0.01; ***p<0.001 financial services review, 32(4) 42 discussion and implications this study examined the racial/ethnic differences in financial advice seeking for saving and investment decisions using the 2016 and 2019 waves of the survey of consumer finances. logistic regression was used to determine if there were significant racial/ethnic differences using a financial planner. the results showed that compared to black and white respondents (see appendix), hispanic and asian/other respondents were less likely to use a financial planner when making decisions about savings and investments, which supported hypotheses 2 and 3. these findings have been corroborated in other studies (hanna, 2011; white & heckman, 2016) and may result from cultural differences among racial/ethnic groups. in particular, hispanic clients may be less likely to seek financial advice if they do not intend to retire (blanco et al., 2017). this would partially explain why nearly three-quarters of hispanic consumers indicate they are not actively engaging in retirement planning (hasler et al., 2023). additionally, asian clients may prefer to selfmanage accounts instead of seeking financial advice (merrill lynch, 2023a). one commonality between hispanic americans and asian americans is that these groups make up a higher percentage of first-generation immigrants compared to white and black americans (budiman et al., 2020; budiman & ruiz, 2021). it may be possible that a sense of isolation and language barriers dissuade these individuals from seeking financial advice (pisnanont et al., 2015). one final point is that hispanic and asian individuals may be more likely to support aging parents and other family members (merrill lynch, 2023b; merrill lynch, 2023a). arguably, the burden of supporting family members might warrant financial advice seeking; however, it may also reduce one's level of disposable income. with less disposable income, these consumers may feel that financial advice is too expensive, or they do not perceive themselves as having enough wealth to warrant seeking advice. the logistic regression analyses also showed that factors associated with seeking financial advice for savings and investing decisions were different across racial/ethnic groups, which supported hypothesis 4. for example, net worth and homeownership were significant, but only among white consumers. having an education beyond high school and age were factors in seeking advice for all consumers except for asian/other consumers. subjective financial knowledge was a positive and significant factor among black and hispanic consumers, while objective financial knowledge was a positive and significant factor among white and asian/other consumers. marriage was significant only for hispanics and asians/others. for black consumers, emergency funds were positively associated with seeking advice, while household size was negatively associated. differences in family dynamics, traditions, customs, and culture and their influence on financial advice seeking warrant further research. no significant difference was found between black and white respondents seeking financial advice for saving and investment decisions, and as such, hypothesis 1 was not supported. this contradicts some previous studies that found differences in blackwhite financial planner use (chang, 2005; elmerick et al., 2002; hanna, 2011; reiter & qing, 2023). however, the discrepancy may be related to the fact that some previous studies investigated financial professional usage by including financial advice seeking behavior around debt, which this study excluded, or by defining financial planner differently than the current study. fairlie’s decomposition estimation (2005) was employed to investigate the most important determinants that explain the racial/ethnic gaps in financial planner use. results indicated that the determinants used to explain the racial/ethnic disparities in seeking financial advice for savings and investing decisions differed between racial/ethnic groups, supporting hypothesis 5. the findings show that white and hispanic households had the most significant total difference between groups (0.1892), and black and hispanic households had the least significant total difference between groups (0.0493). previous research indicates that black and hispanic consumers track similarly in the personal finance domain (white et al., 2021). at the same time, cultural, language, and migration differences may explain why there are differences in financial planner use when these two groups qing & reiter 43 are compared. this difference deserves further investigation. the decomposition analyses found that risk tolerance was a key variable in explaining racial/ethnic differences in financial advice seeking. in four out of six pairwise decomposition analyses, risk tolerance was attributed to differences, and it explained most of the differences in using a financial planner between white and hispanic households and asian/other and hispanic households. this finding is consistent with previous research in that risk tolerance has been positively associated with financial advice seeking (chang, 2005; joo & grable, 2001; moreland, 2018; white & heckman, 2016). in addition, research has pointed to racial/ethnic differences in risk tolerance between whites and hispanics, with hispanic consumers generally taking less risk (fisher, 2020). previous studies have shown that objective financial knowledge is positively related to seeking financial advice (alyousif & kalenkoski, 2017; calcagno & monticone, 2015; seay et al., 2016). our decomposition estimations showed that in three out of six specifications, objective financial knowledge was a key indicator for racial/ethnic differences in financial planner use. there has been an ongoing discussion on how to improve racial/ethnic differences in objective financial knowledge, and our findings serve as further evidence that this is a worthy cause. findings indicate that black and hispanic consumers often have lower levels of financial literacy than white consumers (anong, 2016; porto, 2016) and face barriers to accessing the resources and services necessary for improving financial knowledge. while it is easy to suggest that these groups need more education, it might be more effective to consider the policy and institutional changes that can be made to improve their realities. some may argue that individual or group characteristics are to blame for low financial literacy in black and hispanic groups. however, history suggests that systemic socioeconomic and political barriers, as well as discrimination, have roles to play as well (hamilton & darity, 2017). as such, policymakers should seek effective solutions to increase the financial literacy levels of marginalized groups. however, lifting financial literacy levels alone is not enough. financial literacy is also associated with other factors, such as higher educational levels, so a holistic approach to financial well-being is necessary. income has been positively associated with financial planner use in prior research (elmerick et al., 2002; hanna, 2011), and the decomposition analyses revealed that it is also a factor in racial/ethnic differences in seeking financial advice for savings and investing. it was a factor when looking at three pairwise groups: blackwhite, black-asian/other, and asian/otherhispanic differences. it has long been recognized that there are persistent racial/ethnic differences in income, with white and asian/other consumers making more, on average, than black and hispanic consumers (wilson, 2020). net worth also explained some of the racial/ethnic differences between white and black households, white and hispanic households, and white and asian/other households, albeit at lower percentage rates. net worth has often been touted as a key variable in one’s ability to engage the services of a financial planner (west, 2012). in summary, our findings show racial/ethnic gaps in financial advice seeking. hispanic and asian consumers were significantly less likely to seek financial advice for saving and investments than black and white consumers. these results indicate that there is more to understand regarding attracting hispanic and asian/other clients to financial planning and minimizing barriers. the decomposition analyses revealed more information about the disparities in seeking advice. the findings show that income and risk tolerance explained the gaps in several of the pairwise analyses. income was either the largest or second largest contributor to the gap in seeking financial advice among four pairwise groups: (a) black-white, (b) black-asian/other, (c) whitehispanic, and (d) asian/other-hispanic comparisons. risk tolerance was the largest or second largest factor contributing to the financial advice seeking gap when comparing (a) blackhispanic, (b) white-hispanic, and (c) asian/other-hispanic pairwise groups. these results show that if the racial/ethnic income gap and risk tolerance gaps were remedied, there would be less disparity in financial planner use. black and hispanic consumers lag behind white financial services review, 32(4) 44 and asian consumers in income (greig & eckerd, 2022). evidence supports the idea that racial differences in income help explain the racial wealth gap (ashman & neumeuller, 2020). while there is no quick solution to narrowing either the racial income or wealth gaps, our findings are further evidence that these gaps have broad implications that reach beyond mere finances. this is an area in which policy can be implemented to assist consumers with gaining access to financial planning. the u.s. government has developed such solutions to assist consumers with financial issues. for example, an initiative was established in 2021 by the presidential administration to help build wealth in marginalized communities in an attempt to ameliorate the long-standing and persistent racial wealth gap (the white house, 2021). closing the racial wealth gap is a goal that could help increase the number of marginalized consumers seeking financial advice. similarly, policy could provide governmental financial support to minimize the gap in accessing financial planners. in the current study, the descriptive statistics show that while only 15% and 14% of whites and asians were willing to take no investment risk, nearly 30% of black and hispanics were willing to take no risk. on the other hand, nearly 8% of black and hispanic consumers stated they would be willing to take substantial risk, while only 4% of white and asian/other consumers reported the same. these results align with findings from yao et al. (2005), which found that black and hispanic consumers were more likely to take very high risks but were less likely to take some risks. the results seem to indicate that if black and hispanic consumers had more moderate risk tolerance profiles, this would help close the financial planner use gap. while risk tolerance may not be static, it may be a stretch to suggest that one’s risk tolerance should change. many precipitating factors influence one’s risk tolerance, including experience, knowledge, skills, cognition, and financial satisfaction (grable, 2016). as such, it is insufficient to suggest that education alone can help improve risk tolerance gaps. however, a multi-prong approach to increase awareness and financial literacy play a meaningful role. ultimately, we recommend conducting further research on financial planner use among different racial/ethnic groups. specifically, future research should investigate cultural, immigration, language, and other factors that may play a role in seeking financial advice. qualitative research would be particularly useful in this process. it would allow researchers to learn more about information-seeking among diverse racial/ethnic groups in their own words. limitations some limitations in this study should be noted for subsequent studies. first, as stated by fairlie (2005), unlike in the case of linear models, the matrix of independent predictors from two groups (x1 and x2) depends on the value of other variables; as a result, the order of switching the distribution may result in small differences in the outcome. second, this study mainly examined the disparities from an economic behavior perspective; however, hiring a financial planner is a complex decision-making process. for example, future research should consider respondents’ psychological perspectives and cognitive abilities. third, there are likely unobservable factors, such as cultural differences, at play in addition to those included in this study’s model. for example, it must be noted that certain consumers may face barriers in seeking financial advice due to the lack of access to professionals who share the same language and cultural background. also, differences in results may be related to differences between u.s.-born respondents and respondents who migrated to the u.s. it is likely these two groups may have differing perspectives on financial advice seeking. fourth, asians and consumers from racial/ethnic groups with a comparatively small representation in the sample (i.e., native americans) were combined to make one category, given the limitations in the data. interpretations of the results for this group should be read with an understanding that respondents with diverse racial, ethnic, and cultural identities and languages were combined. finally, this study used cross-sectional data. it would be worthwhile to estimate differences over time utilizing panel data. qing & reiter 45 references aladangady, a., chang c. a., & krimmel j. (2023). "greater wealth, greater uncertainty: changes in racial inequality in the survey of consumer finances," feds notes. washington: board of governors of the federal reserve system https://doi.org/10.17016/23807172.3405 alyousif, m., & kalenkoski, c. (2017). who seeks financial advice? financial services review 26(4), 405-432. ando, a., & modigliani, f. (1963). the “life cycle” hypothesis of saving: aggregate implications and tests. the american economic review, 53(1), 55-84. anong, s.t. (2016). financial issues of african americans. in j. j. xiao (ed.), handbook of consumer finance research (pp. 215– 224). springer. https://doi.org/10.1007/978-3-31928887-1_18 ashman, h., & neumuller, s. (2020). can income differences explain the racial wealth gap? a quantitative analysis. review of economic dynamics, 35, 220– 239. blanco, l. r., aguila, e., gongora, a., & duru, o. k. (2017). retirement planning among hispanics: in god’s hands? journal of aging & social policy, 29(4), 311–331. https://doi.org/10.1080/08959420.2016. 1272161 budiman, a., tamir, c., mora, l., & noebustamante, l., (2020, august 20). facts on u.s. immigrants, 2018. pew research center. https://www.pewresearch.org/hispanic/2 020/08/20/facts-on-u-s-immigrants/ budiman, a., & ruiz, n. g. (2021, april 29). key facts about asian americans, a diverse and growing population. pew research center. https://www.pewresearch.org/shortreads/2021/04/29/key-facts-about-asianamericans/ calcagno, r., & monticone, c. (2015). financial literacy and the demand for financial advice. journal of banking & finance, 50, 363-380. chang, m. l. (2005). with a little help from my friends (and my financial planner). social forces, 83(4), 1469–1498. https://doi.org/10.1353/sof.2005.0061 chatterjee, s., & zahirovic-herbert, v. (2010). retirement planning of younger babyboomers: who wants financial advice? financial decisions, 22(2), 1–12. craemer, t., smith, t., harrison, b., logan, t., bellamy, w., & darity jr, w. (2020). wealth implications of slavery and racial discrimination for african american descendants of the enslaved. the review of black political economy, 47(3), 218– 254. cummings, b. f., & james iii, r. n. (2014). factors associated with getting and dropping financial advisors among older adults: evidence from longitudinal data. journal of financial counseling and planning, 25(2), 129–147. elmerick, s., montalto, c., & fox, j. (2002). use of financial planners by u.s. households. financial services review, 11(3), 217231. estreicher, s., & gold, l. (2007). the shift from defined benefit plans to defined contribution plans. lewis & clark l. rev., 11, 331. fairlie, w, r, (2005). an extension of the blinder-oaxaca decomposition technique to logit and probit models. journal of economic and social measurement, 30(4), 305–316. https://doi.org/10.3233/jem-2005-0259 fan, l. (2021). a conceptual framework of financial advice-seeking and short-and long-term financial behaviors: an age comparison. journal of family and economic issues, 42(1), 90-112. https://doi.org/10.1007/s10834-02009727-3 https://doi.org/10.17016/2380-7172.3405 https://doi.org/10.17016/2380-7172.3405 https://doi.org/10.1007/978-3-319-28887-1_18 https://doi.org/10.1007/978-3-319-28887-1_18 https://doi.org/10.1080/08959420.2016.1272161 https://doi.org/10.1080/08959420.2016.1272161 https://www.pewresearch.org/hispanic/2020/08/20/facts-on-u-s-immigrants/ https://www.pewresearch.org/hispanic/2020/08/20/facts-on-u-s-immigrants/ https://www.pewresearch.org/short-reads/2021/04/29/key-facts-about-asian-americans/ https://www.pewresearch.org/short-reads/2021/04/29/key-facts-about-asian-americans/ https://www.pewresearch.org/short-reads/2021/04/29/key-facts-about-asian-americans/ https://doi.org/10.1353/sof.2005.0061 https://doi.org/10.3233/jem-2005-0259 https://doi.org/10.1007/s10834-020-09727-3 https://doi.org/10.1007/s10834-020-09727-3 financial services review, 32(4) 46 finke, m. s., huston, s. j., & winchester, d. d. (2011). financial advice: who pays. journal of financial counseling and planning, 22(1), 18–26. fisher, p. j. (2020). financial risk tolerance among white and hispanic households. journal of personal finance, 19(1), 33–47. gaudecker, h. v. (2015). “how does household portfolio diversification vary with financial literacy and financial advice?” the journal of finance, 70(2), 489–507. goetz, j. w., palmer, l., zhang, l., & chatterjee, s. (2020). changes in household net financial assets after the great recession: did financial planners make a difference? journal of personal finance, 19(1), 9-17. grable, j. e. (2016). financial risk tolerance. in j. j. xiao (ed.), handbook of consumer finance research (pp. 19-26). springer. https://doi.org/10.1007/978-3-31928887-1_2 grable, j. e., & joo, s. h. (1999). financial helpseeking behavior: theory and implications. journal of financial counseling and planning, 10(1), 14-25. greig, f., & eckerd, g. (2022). racial income inequality dynamics: big data insights from 2013 through covid-19. jpmorgan chase institute. https://www.jpmorganchase.com/institut e/research/household-incomespending/racial-income-inequalitydynamics hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking & finance, 36(2), 509–524. https://doi.org/10.1016/j.jbankfin.2011.0 8.008 hamilton, d. & darity, w.a. (2017). the political economy of education, financial literacy, and the racial wealth gap. federal reserve bank of st. louis review, 99(1), 59–76. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10(1), 36–62. hanna, s. d., wang, c., & yuh, y. (2010). racial/ethnic differences in high return investment ownership: a decomposition analysis. journal of financial counseling and planning, 21(2), 44–59. hanna, s., lee, j. & lindamood, s. (2015). financial behavior and attitudes of asians compared to other racial/ethnic groups in the united states. journal of family and economic issues, 36(3), 309–318. https://doi.org/10.1007/s10834-0159447-2 harlow, w. v., brown, k. c., & jenks, s. e. (2020). the use and value of financial advice for retirement planning. the journal of retirement, 7(3), 46–79. hasler, a., lusardi, a., mitchell, o. s., & sconti, a. (2023). understanding financial vulnerability among asians, blacks, and hispanics in the united states. teachers insurance and annuity association of america, 200, 1–34. https://doi.org/10.2139/ssrn.4425740 hsu, y. l. (2022). financial advice seeking and behavioral bias. finance research letters, 46, 102505. https://doi.org/10.1016/j.frl.2021.10250 5 jackson, j. d., & lindley, j. t. (1989). measuring the extent of wage discrimination: a statistical test and a caveat. applied economics, 21(4), 515–540. joo, s. h., & grable, j. e. (2001). factors associated with seeking and using professional retirement‐planning help. family and consumer sciences research journal, 30(1), 37-63. https://doi.org/10.1007/978-3-319-28887-1_2 https://doi.org/10.1007/978-3-319-28887-1_2 https://www.jpmorganchase.com/institute/research/household-income-spending/racial-income-inequality-dynamics https://www.jpmorganchase.com/institute/research/household-income-spending/racial-income-inequality-dynamics https://www.jpmorganchase.com/institute/research/household-income-spending/racial-income-inequality-dynamics https://www.jpmorganchase.com/institute/research/household-income-spending/racial-income-inequality-dynamics https://doi.org/10.1016/j.jbankfin.2011.08.008 https://doi.org/10.1016/j.jbankfin.2011.08.008 https://doi.org/10.1007/s10834-015-9447-2 https://doi.org/10.1007/s10834-015-9447-2 https://doi.org/10.2139/ssrn.4425740 https://doi.org/10.1016/j.frl.2021.102505 https://doi.org/10.1016/j.frl.2021.102505 qing & reiter 47 kim, k. t., pak, t. y., shin, s. h., & hanna, s. d. (2018). the relationship between financial planner use and holding a retirement saving goal: a propensity score matching analysis. financial planning review, 1(1-2), e1008. https://doi/10.1002/cfp2.1008 kramer, m. m. (2016). financial literacy, confidence, and financial advice seeking. journal of economic behavior & organization, 131, 198-217. https://doi.org/10.1016/j.jebo.2016.08.0 16 lachance, m. e., & tang, n. (2012). financial advice and trust. financial services review, 21(3), 209–226. lee, s. t., kim, k. t. (2022). a decomposition analysis of racial/ethnic differences in financial knowledge and overconfidence. journal of family and economics, 43(4), 815–831. https://doi.org/10.1007/s10834-02109795-z lei, s., & kordes, m. h. (2020). women, wealth, and demand for financial planning services. journal of financial planning, 33(8), 48-55. lim, h., heckman, s., montalto, c. p., & letkiewicz, j. (2014). financial stress, self-efficacy, and financial help-seeking behavior of college students. journal of financial counseling and planning, 25(2), 148-160. lindamood d, s., hanna, s. d. and bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41(2), 195–222. https://doi.org/10.1111/j.1745 6606.2007.00075_1.x liu, f., yilmazer, t., loibl, c., & montalto, c. (2019). professional financial advice, self‐control, and saving behavior. international journal of consumer studies, 43(1), 23-34. lusardi, a., & mitchell, o. s. (2011). financial literacy around the world: an overview. journal of pension economics & finance, 10(4), 497-508. lópez-anuarbe, m., cruz-saco, m. a., & park, y. (2016). more than altruism: cultural norms and remittances among hispanics in the usa. journal of international migration and integration, 17, 539-567. marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32(4), 625–643. https://doi.org/10.1007/s10834-0119258-z martin, e. c., & dwyer, r. e. (2021). financial stress, race, and student debt during the great recession. social currents, 8(5), 424–445. merrill lynch. (2023a). diverse viewpoints: exploring wealth in the aapi community. merrill lynch. https://mlaem.fs.ml.com/content/dam/m l/articles/pdf/aapi-ada.pdf merrill lynch. (2023b). diverse viewpoints: exploring wealth in the hispanic-latino community. merrill lynch. https://mlaem.fs.ml.com/content/dam/m l/articles/pdf/ml-hl-wp.pdf montalto, c. p. & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances, journal of financial counseling and planning, 7, 133-141. montmarquette, c., & viennot-briot, n. (2015). the value of financial advice. annals of economics and finance, 16(1), 69-94. moreland, k. a. (2018). seeking financial advice and other desirable financial behaviors. journal of financial counseling and planning, 29(2), 198– 207. munnell, a., & sundén, a. (2006). 401(k) plans are still coming up short. center for retirement research at boston college. https://doi/10.1002/cfp2.1008 https://doi.org/10.1016/j.jebo.2016.08.016 https://doi.org/10.1016/j.jebo.2016.08.016 https://doi.org/10.1007/s10834-021-09795-z https://doi.org/10.1007/s10834-021-09795-z https://doi.org/10.1111/j.1745-%206606.2007.00075_1.x https://doi.org/10.1111/j.1745-%206606.2007.00075_1.x https://doi.org/10.1007/s10834-011-9258-z https://doi.org/10.1007/s10834-011-9258-z https://mlaem.fs.ml.com/content/dam/ml/articles/pdf/aapi-ada.pdf https://mlaem.fs.ml.com/content/dam/ml/articles/pdf/aapi-ada.pdf https://mlaem.fs.ml.com/content/dam/ml/articles/pdf/ml-hl-wp.pdf https://mlaem.fs.ml.com/content/dam/ml/articles/pdf/ml-hl-wp.pdf financial services review, 32(4) 48 myers, e., & topoleski, j. (2021). a visual depiction of the shift from defined benefit (db) to defined contribution (dc) pension plans in the private sector. (report no. if12007). congressional research service. https://crsreports.congress.gov/product/p df/if/if12007 oaxaca, r. l., & ransom, m. r. (1994). on discrimination and the decomposition of wage differentials. journal of econometrics, 61(1), 5-21. pence, k. (2015). scfcombo: stata module to estimate errors using the survey of consumer finances. https://econpapers.repec.org/software/bo cbocode/s458017.htm pisnanont, j., duong, j., condon, a., cruzviesca, m.., pech, c., & ong, p. (2015). scrimping and saving: a report on financial access, attitudes, and behaviors of lowand moderate-income asian americans and pacific islanders. aapi nexus: policy, practice and community, 13(1–2), 25-44. https://doi.org/10.17953/15450317.13.1.25 porto, n. (2016). financial issues of hispanic americans. in j. j. xiao (ed.), handbook of consumer finance research, (pp, 205– 214). springer. https://doi.org/10.1007/978-3-31928887-1 reiter, m., & qing, d. (2023). racial and gender differences in financial advice seeking: evidence from the national financial capability study. financial planning review, e1169. https://doi.org/10.1002/cfp2.1169 reiter, m., & seay, m. (2022). investment advisor use and stock market return expectations. journal of personal finance, 22(1), 1-18. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21(4), 291-305. rubin, d. b. (1987). multiple imputation for nonresponse in surveys. john wiley & sons. seay, m., kim, k. t., & heckman, s. (2016). exploring the demand for retirement planning advice: the role of financial literacy. financial services review, 25(4), 331–350. shin, s. h., & hanna, s. d. (2015). decomposition analyses of racial/ethnic differences in high-return investment ownership after the great recession. journal of financial counseling and planning, 26(1), 43-62. https://doi.org/10.1891/10523073.26.1.43 sommer, m., & lim, h. (2022). financial knowledge and the relationship with using professional investment advice and granting trading discretion. journal of personal finance, 21(1), 62-77. suchman, e. a. (1966). health orientation and medical care. american journal of public health, 56, 97-105. the white house. (2021, june 1). fact sheet: biden-harris administration announces new actions to build black wealth and narrow the racial wealth gap. https://www.whitehouse.gov/briefingroom/statementsreleases/2021/06/01/fact-sheet-bidenharris-administration-announces-newactions-to-build-black-wealth-andnarrow-the-racial-wealth-gap/ west, j. (2012). financial advisor participation rates and low net worth investors. journal of financial services marketing, 17(1), 50–66. white, k., & heckman, s. (2016). financial planner use among black and hispanic households. journal of financial planning, 29(9), 40–49. https://crsreports.congress.gov/product/pdf/if/if12007 https://crsreports.congress.gov/product/pdf/if/if12007 https://econpapers.repec.org/software/bocbocode/s458017.htm https://econpapers.repec.org/software/bocbocode/s458017.htm https://doi.org/10.17953/1545-0317.13.1.25 https://doi.org/10.17953/1545-0317.13.1.25 https://doi.org/10.1007/978-3-319-28887-1 https://doi.org/10.1007/978-3-319-28887-1 https://doi.org/10.1007/978-3-319-28887-1 https://doi.org/10.1007/978-3-319-28887-1 https://doi.org/10.1002/cfp2.1169 https://doi.org/10.1111/fcsr.12397 https://doi.org/10.1111/fcsr.12397 https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/01/fact-sheet-biden-harris-administration-announces-new-actions-to-build-black-wealth-and-narrow-the-racial-wealth-gap/ https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/01/fact-sheet-biden-harris-administration-announces-new-actions-to-build-black-wealth-and-narrow-the-racial-wealth-gap/ https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/01/fact-sheet-biden-harris-administration-announces-new-actions-to-build-black-wealth-and-narrow-the-racial-wealth-gap/ https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/01/fact-sheet-biden-harris-administration-announces-new-actions-to-build-black-wealth-and-narrow-the-racial-wealth-gap/ https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/01/fact-sheet-biden-harris-administration-announces-new-actions-to-build-black-wealth-and-narrow-the-racial-wealth-gap/ https://www.whitehouse.gov/briefing-room/statements-releases/2021/06/01/fact-sheet-biden-harris-administration-announces-new-actions-to-build-black-wealth-and-narrow-the-racial-wealth-gap/ qing & reiter 49 white, k. j., mccoy, m., watkins, k., chen, x., koposko, j., & mizuta, m. (2021). “we don’t talk about that”: exploring money conversations of black, hispanic, and white households. family and consumer sciences research journal, 49(4), 328-343. https://doi.org/10.1111/fcsr.12397 wilson, v. (2020, september 16). racial disparities in income and poverty remain largely unchanged amid strong income growth in 2019. working economics blog. https://www.epi.org/blog/racialdisparities-in-income-and-povertyremain-largely-unchanged-amid-strongincome-growth-in-2019/. wolff, e. (2023). trends in the retirement preparedness of black and hispanic households in the us (nber working papers 31532). https://ideas.repec.org/p/nbr/nberwo/315 32.html yao, r. (2016). financial wellbeing of asian americans. in j. j. xiao (ed.), handbook of consumer finance research (pp. 225– 238). springer. https://doi.org/10.1007/978-3-31928887-1_19 yao, r., gutter, m. s., & hanna, s. d. (2005). the financial risk tolerance of blacks, hispanics and whites. journal of financial counseling and planning, 16(1), 51–62. https://doi.org/10.1111/fcsr.12397 https://www.epi.org/blog/racial-disparities-in-income-and-poverty-remain-largely-unchanged-amid-strong-income-growth-in-2019/ https://www.epi.org/blog/racial-disparities-in-income-and-poverty-remain-largely-unchanged-amid-strong-income-growth-in-2019/ https://www.epi.org/blog/racial-disparities-in-income-and-poverty-remain-largely-unchanged-amid-strong-income-growth-in-2019/ https://www.epi.org/blog/racial-disparities-in-income-and-poverty-remain-largely-unchanged-amid-strong-income-growth-in-2019/ https://ideas.repec.org/p/nbr/nberwo/31532.html https://ideas.repec.org/p/nbr/nberwo/31532.html https://doi.org/10.1007/978-3-319-28887-1_19 https://doi.org/10.1007/978-3-319-28887-1_19 financial services review, 32(4) 50 appendix binomial logistic regression of financial planner use (white as reference) pooled sample variables coef. s.e. race/ ethnicity (ref.=white) black 0.0435 0.0588 hispanic -0.1463** 0.0563 asian/other -0.2359** 0.0811 female (ref.= male) 0.3284*** 0.0324 age 0.0343*** 0.0072 age squared -0.0003*** 0.0001 married (ref.= not married) -0.0706 0.0472 income (ref.= less than $50k) $50k-$99,999 0.2189*** 0.0487 $ 100k above 0.5294*** 0.0506 net worth 0.0225*** 0.0052 risk tolerance (ref.=not willing) average risk 0.7900*** 0.0480 above average risk 0.7823*** 0.0518 substantial risk 0.5862*** 0.0919 invest horizon (ref.=next few months) next year 0.0105 0.0591 next few years 0.1714** 0.0560 next 5-10 years 0.3205*** 0.0605 longer than 10 years 0.3778*** 0.0618 subjective knowledge 0.0172* 0.0080 objective knowledge 0.1770*** 0.0254 household size -0.0353* 0.0145 educational attainment (ref. = less than high school) high school 0.5133*** 0.1091 some college 0.6238*** 0.1011 bachelor 0.8328*** 0.1031 graduate 0.9187*** 0.1025 homeownership (ref.=no) 0.1552*** 0.0486 employed (ref.=unemployed) employee 0.0093 0.0990 self-employed 0.0397 0.1013 retired 0.1318 0.1046 have emergency funds (ref.= no) 0.1014* 0.0433 intercept -4.4575*** 0.2543 sample size 4,059 r-squared 0.1172 note: unweighted analysis, 2016 & 2019 scf. *p<0.05; **p<0.01; ***p<0.001 relation between financial advisory designations and finra misconduct fsr201714 jeffrey m. camarda, phd, cfa eaa,* a4371 us 17, fleming island, fl 32003, usa abstract registered representatives have no general fiduciary duty. cfp, chfc, and cfa designees have higher ethical duties and education requirements. registrants’ criminal, regulatory, complaint, and other misconduct history is public. this study examines misconduct disclosures of undesignated versus designated florida securities salespeople, and finds adverse disclosure materially decreases for designees; it incidentally finds misconduct increases with maleness, dual investment advisor/registered representative status, and life insurance sales licensure. this appears to be the first such study of adverse disclosure association with financial designations, adding to the emerging misconduct and advisors’ ethics literature. these findings offer important policy and consumer choice insight. © 2017 academy of financial services. all rights reserved. jel classification: g020; g180; g220; g280; g290 keywords: advisor; fiduciary; designation; best interest; misconduct 1. introduction financial industry regulatory authority (finra) registered representatives (rrs) have come to be known as financial “advisors.” rrs, as agents of broker dealer (bds) organizations, are typically commission salespeople generally not required to put customers’ interests ahead of their own compensation, in contrast to investment advisor representatives (iars), who are agents of fiduciary registered investment advisor organizations (rias). many rrs are also iars, with conflicting rules and duties, often to the same clients, * corresponding author. tel.: �1-904-278-1177; fax: �1-278-1070. e-mail address: j@camarda.com financial services review 26 (2017) 271–290 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. depending on the hat worn. the respective codes for certified financial planner (cfp), chartered financial consultant (chfc), and chartered financial analyst (cfa) designees require greater financial expertise and ethical duties to clients than do rr rules. many rrs are designees. rr criminal, regulatory, complaint, and other misconduct history is public record. this study examines the comparative misconduct of undesignated versus designated rrs in florida. it appears to be the first such study of adverse disclosure association with financial designations, so adding to the emerging finra misconduct literature (egan, matvos, & seru, 2016). as the financial advisory industry evolves and regulators and other stakeholders seek to enhance practice quality and duties to investors, these findings may offer important contributions to policy, and help consumers to make better advisor choices. 2. background more than 650,000 financial advisors help manage over $30 trillion of investable assets in the united states (egan, matvos, & seru, 2016), with over one-half of all households and nearly 90% of consumers with investable assets over $100,000 seeking such help (smith, vibhakar, & terry, 2008). advisory wrongdoing has not undergone rigorous investigation and analysis (zingales, 2015). rr sales licensing is relatively simple, and even exam cheaters may become licensed (finra, 2016). even sophisticated consumers can easily confuse rr sales reps for fiduciary advisors (government accountability office [gao], 2011). rr misconduct is widespread, and tends to concentrate at firms that may enable it (egan et al., 2016) or even depend on it as a profit center (woolley, 2016). consumers seem generally ignorant of these factors (mccann, qin, & yan, 2016), and social cost is not yet known (ritholtz, 2016). ethically challenged advisor behavior is estimated to waste some $17 billion per year of consumer wealth (the white house, 2016). while rr misconduct data are public, limited accessibility may damage consumers (piaba foundation, 2015). designations sponsors’ (cfa institute, 2016, cfp board, 2016, the american college, 2016), send consumer quality signals (terry & vibhakar, 2011) implying lower misconduct for designees. findings of valid signals could have important consumer welfare and policy implications, provide justification to adopt broad fiduciary standards across the multiple advisory channels, and nurture standardized professionalism. unlike established professions such as law, medicine, and accounting, advisors have wide variations in ethical duties and required education. regulatory gaps are significant and persistent (gao, 2011). ria companies and their iar agents are required to protect clients’ best interests by being held to a fiduciary standard, but rrs— even when dully licensed as iars—are not. rrs may appear to consumers as fiduciaries, but are generally held only to the lower suitability standard, allowing them to put their own interests first. the data here precede implementation of the emerging dol rule that would apply a limited fiduciary rr duty to retirement accounts only. very many study rrs have the cfp, chfc, and cfa designations (the “study designations”). these require meaningful study, examination, and allegiance to ethical codes with much stronger duties than rr rules. designated and undesignated rrs are hired to sell commission products and operate under the same regulations; neither are legally generally 272 j.m. camarda / financial services review 26 (2017) 271–290 required to act as fiduciaries. many rrs offer advice under marketing flags that inaccurately imply fiduciary duty (raymond james mission statement, 2016), flown by some bds with the very highest misconduct rates (egan, matvos, & seru, 2016). such misleading claims seem common. (vystar ad, 2016, p. b-8). hauptman and roper (2017) find such claims endemic to the bd industry. finra advisors exhibit a high level of misconduct, and an anomalous concentration of repeat offenders, compared with physicians (egan et al., 2016). 3. the question of financial advisor professionalism there are no uniform professional standards to which those calling themselves financial advisors must adhere, and on which the public can rely. the occupation, as a whole, does not rise to established standards of professionalism. while some, such as study designees, do or purport to, it is important to note that these represent an occupational subset, and that the public may not be aware of the difference. this study distills six theoretical attributes of a profession from the literature (dean, 1997; flexner, 1976; flexner & metzger, 1976; khurana, nohria, & penrice, 2005, khurana & nohria, 2008; ragatz & duska, 2010; schaefer, 1984) and tests advisors against them: 1. specialized, arcane, deep, socially useful expert knowledge requiring hard and constant study, that is codified, evolves and is perpetuated: while a rigorous body of academic knowledge is developing (kitces, 2014b), there is no requirement that a practitioner obtain it (kitces, 2015b; moisand, 2008). 2. rigorous testing, vetting, and certification of expert knowledge and ability: no practitioners’ requirement (gao, 2011; kitces, 2015b). 3. the placing of explored client interests before the professional’s: there is no general requirement of fiduciary duty (cummings & finke, 2010; gao, 2011). 4. formal system for professional standards promulgation and oversight: while a number of detailed financial designation ethical codes have emerged as cataloged by ragatz and duska (2010), these do not apply to all financial advisory practitioners. (gao, 2011; ragatz & duska, 2010). 5. monopoly power derived from social contract: this aspect is completely nonexistent in the current environment (cummings & finke, 2010; gao, 2011). 6. commitment to excellence, service, collegiality, and dignified conduct: for the reasons discussed under 4, above, this test is not met. moisand (2012) analyzes financial planning and also concludes it is not a profession. others (financial planning coalition, 2014; frumento & korenman, 2013) concur and suggest the lack of real governmental regulation as a profession promotes unethical and fraudulent activity. while a profession is clearly emerging, a still-nascent commitment to formalized academic development is critical to its success (warschauer, 2002). undefined terms like financial advisor are freely used without regulation (gao, 2011). consumers may not understand the difference between fiduciary iars and rrs who are free to enrich themselves at clients’ expense (cummings & finke, 2010; finke & langdon, 2012; gao, 2011) but market and sell as if they put clients first (hauptman & roper, 2017). 273j.m. camarda / financial services review 26 (2017) 271–290 established monopolistic professions like law and medicine use strong signals to establish professionalism. in nonmonopolistic markets, signal theory purports that agents signal skills with educational credentials; acquiring a credential is essentially a reputation-for-quality purchase (spence, 1973; spence, 2002). in cases like financial advisory where uniform professional standards are lacking, practitioners can distinguish themselves from the merely licensed by associating with certifying bodies (mauldin, wilder, & stocks, 2000). to consumers, signals like md or cpa offer clear indications of baseline expertise, and study designations act similarly to these with the key difference that they are not required by regulation. mauldin, wilder, and stocks (2000) and brockman and brooks (1998) find designations signal objectivity, expertise, and ethics to consumers. smith, vibhakar, and terry (2008) find advisors obtain designations to establish professional expertise and credibility. designations serve as an “umbrella brand for the . . . cfp or cfa . . . (that’s) worked successfully for cpas for decades (smith et al., 2008, p. 308). the effect of the study designations on rr misconduct is this study’s primary research question. the cfp mark has become the preeminent financial planning designation (kitces, 2015), largely because of expensive and effective cfp board marketing efforts (kitces, 2015c). it is now most demanded by consumers, and more recognized by consumers by a factor of over two compared to cfa and over four compared to chfc (cfp board, 2015). it is perceived to be more appropriate for financial planning than the more specialized investments-expert cfa (terry & vibhakar, 2011). non-cfa cfps’ confidence in their investment skills may be overstated as compared to those who have acquired both and perhaps better recognize the limits of their knowledge (cordell, smith, & terry, 2011). it is worth noting that this study deals with investment sales agents, and that the cfa investment knowledge set is much greater than those for cfp and chfc. the cfa designation requires three successive eight hour exams on investments management. the six-hour cfp exam tests nine subject areas, including investments, insurance and risk management, income taxes, estate planning, retirement planning, employee benefits, professional conduct, financial planning principles, and financial plan development. the chfc designation requires the completion of nine courses in financial planning, including income taxation, insurance, retirement planning, investments, estate planning, and case studies (the american college chartered financial consultant, 2016b). the cfp board standards of professional conduct requires certificates to “at all times place the interest of the client ahead of his or her own,” (cfp standards of professional conduct, 2008/2014), but it is worth noting that cfp board’s public position has been that a fiduciary duty is only owed for financial planning engagements as opposed to product sales. the cfa code of ethics requires charter holders to place the interests of clients above their own, to act with integrity, competence, and respect, and to maintain and develop professional expertise (cfa institute, 2015). “chfc advisors are required to do the same for clients that they would do for themselves in similar circumstances, the standard of ethical behavior most beneficial for their clients” (the american college the highest standard, 2016a, p. 1). from a fiduciary perspective, the cfa, chfc, and cfp ethical requirements are functionally equivalent, and well beyond the rr regulatory standard. the designation codes do not carry the force of law, and compliance—and ethics code understandings—may vary considerably among designees. 274 j.m. camarda / financial services review 26 (2017) 271–290 conferring institutions’ claims of designations’ signal validity are compelling and uncompromising. chfcs are said to have “the most extensive educational program . . . strict ethical standards, and . . . serve you with the highest level of professionalism” (the american college the highest standard, 2016, p. 1; the american college code of ethics, 2016c). cfp board notes that “. . . cfps . . . have . . . extensive training and experience . . . and . . . held to rigorous ethical standards . . . will make recommendations in your best interest” (cfp board, 2016, p. 1). cfas are promoted as setting “. . . the global standard for . . . integrity, dedication, and advanced skills . . . no credential is as widely respected . . . (to) serve the best interests of investors and society” (cfa institute value of the cfa charter, 2016, p. 1). finra misconduct studies are nascent. barry and eaglesham (2014) find that data restrictions confound study, and that industry-controlled finra does this intentionally. mccann, qin, and yan (2016) agree, finding this done to promote the illusion of transparency, and conceal that misconduct is recurring and predictable. eagan, matvos, and seru (2016) build a regressable database by accessing individual rrs’ brokercheck records, one at a time, to facilitate statistical analysis, finding that 12% of rrs have misconduct disclosures and 7% have been disciplined for misconduct or fraud, and that those with misconduct are five times more likely to repeat it. they find such misconduct elevates in firms targeting customers in areas with concentrations of people who are elderly, and/or have high incomes. of note to this study, florida has a high concentration of such populations. 4. research design and methodology this study uses secondary finra data. rrs must keep current form u4, which requires answers to 57 misconduct questions, involving criminal, regulatory, civil, complaint, and other adverse items. finra maintains records of disclosure events that may be indicative of unethical practice, which this study refers to as disclosure items, or misconduct. misconduct measures comprise the dependent variables using a unique model developed for this study that compiles adverse yes disclosure answers from rrs’ u4s. this study’s independent variable (iv) is the presence of at least one of the study designations. control ivs include age and gender, years of rr registration, employee verses independent contractor status, the presence or absence of a life insurance sales license and the presence of an active fiduciary investment advisor’s iar license in addition to the rr investments products sales license. given their additional education and ethical requirements, designees should be more effective in subordinating their interests and dispensing better advice than undesignated reps. this might reasonably be expected to be associated with lower disclosure. because this study restricts analysis to rr sales agents, it controls for the difference between suitability sales agents and legally dedicated fiduciary non-rr iars, because these latter do not also sell securities products under suitability. it should be emphasized that some study rrs wear both hats, enhancing conflicts for these. including only those licensed as rrs is intended to highlight any pure designation effect differences associated with those bound by designation fiduciary ethical codes who operate in a suitability standard product sales environment. there is potential endogeneity, but it is unclear what, if any, bias this has on results. would-be designees with misconduct may be prevented, via prescreening, suspensions, or revocations, 275j.m. camarda / financial services review 26 (2017) 271–290 from using a designation, resulting in a mechanical suppression of designee misconduct per se. this could induce a survivorship bias and a cross section study flaw where misconduct is found to decline for designees. it is possible than any misconduct or revocation effect may be somewhat self-correcting because designees disclosing misconduct may find their designations revoked after a lag, who would subsequently be measured as undesignated rrs, and that these fluctuations would damp out in large samples over time. if endogeneity is a significant factor—because designees are closely scrutinized for disclosure events, and because observations of disclosed misconduct are correlated with revocations—one would expect lower scores for surviving designees, which should affirm the desired signaling mechanism. it is also quite possible that a number of errant designees slip through the designators’ enforcement nets. it should be noted that professional bodies such as the cfp board and the cfa institute may only on third party complaints and the self reporting of ethical violations for this information; perhaps routine screening via the brokercheck algorithm would be prudent. it should be also be noted that while the ria/iar disclosure form adv discloses designation suspensions or revocations, the rr form u4 does not. there are no reasonably clear endogeneity effects affecting this study’s results. the regression models equate various measures of misconduct against the study and control independent variables available in the data. the primary dependent variable (dv) is a construct referred to as disclosure incidence score (dis); all other dvs are derived from dis. the dis model assigns a point value subscore to each u4 yes answer depending on nature and severity. dis results shows a censored range from 0 to 56. most registrants’ misconduct scores are at or near zero, producing a very non-normal distribution. the target population are those licensed as retail rrs in florida. those modeled as nonretail rrs, such as analysts, are removed. from 35,361 rrs in the raw data, 26,667 are modeled as retail. of these, 116 (0.4%) have the cfa, 2,534 (9.50%) the cfp, and 970 (3.64%) the chfc. florida is one of the three highest misconduct states, has one of the highest concentrations of rrs (egan et al., 2016), and is one of only 12 states imposing a limited fiduciary duty on rrs (finke & langdon, 2012). given this, the greater florida misconduct incidence noted by egan et al., (2016) may underscore any designation effect found here. u4 data are obtained from a vendor serving industry recruiting interests, and are generally consistent with those on brokercheck and referenced in other studies (barry & eaglesham, 2014; egan et al., 2016; mccann et al., 2016). given this study’s data comprise virtually the entire population, traditional reliability concerns are not an issue. data are such that misconduct scores are cumulative and not transitory. note they allow for only one response to each question, regardless of the number of incidents to which it might apply. for example, a rr whose license has been revoked six times would show the same score for the applicable question as one revoked only once; this will likely produce under-measurement distortions. 5. model specification, dependent and independent variables dis is comprised of the sum of the subscores for all u4 disclosure questions, advisorrelated or not, and whether indicative of mere allegation or clear wrongdoing. dis is thus the broadest misconduct measure in the study. dis scoring methodology is described below. 276 j.m. camarda / financial services review 26 (2017) 271–290 adis (advisor dis), is a dis subset restricted to the sum of all dis subscores that relate only to advisor functions; for instance, financial fraud or regulatory suspension would be included, but domestic abuse or personal bankruptcy would not. adis includes both allegations and indications of culpability. cad (culpable advisor disclosure) sums only the subscores that are both advisor-related and indicate findings of wrongdoing or other clear indications of culpability. as such it is this study’s most specific and important misconduct measure. the b-dvs are binary transformations of the corresponding dvs. for each, if the correspondent is greater than zero, the bvalue would be one; if the correspondent is zero, the b-value is zero. note scalar information is lost; dis � 56 or dis � 1 both become b-dis � 1. the binary transformations are needed for logit and facilitate informative descriptive statistics. b-dis – the binary transformation of dis. b-adis – the binary transformation of adis. b-cad – the binary transformation of cad. anydes is an indicator iv for rrs holding at least any one of the three study designations. this is the study iv. this study also uses control ivs of age, gender, years finra registered, life insurance sales licensure, independent contractor versus employee status, and dual rr/iar registration status. employment status seeks to control for possible higher employer supervision effects verses perhaps reduced compliance oversight for nonemployee contractors. note dual registration means the subjects can do business as both rrs and iars, a conflicted situation given respective suitability versus fiduciary duties. rrs with insurance licenses have the ability to also sell fixed and variable life insurance products such as annuities, often marketed as investments, which have a robust history of complaints. while the control variables are not hypothesized in this study as explanatory, some of their results seem quite interesting and worthy of further study. 6. construction of dependent variables there are 57 individual disclosure categories queried on the form u4. the excerpt in fig. 1 offers an example of these questions. some of these—such as “have you ever been charged with any felony?”—disclose allegation, but not culpability, of a non-advisor–specific disclosure, while others, such as “have you ever been convicted of . . . a misdemeanor involving investments . . . or any fraud . . .” connote a finding of clear advisor-related wrongdoing. to assess the potential impact on advisory quality, the study quantifies each yes disclosure based on the applicable two of four possible conditions: is the disclosure of mere allegation, or does it reasonably indicate culpability? does the disclosure specifically relate to theft or consumer investment harm, or instead relate to nonadvisor-specific unethical behavior? each condition is assigned a dis factor value of 1–4, and the product of the factors corresponding to a specific combination of conditions determines the misconduct subscore for a particular disclosure question. for each registrant, each question’s subscores are summed to determine a regis277j.m. camarda / financial services review 26 (2017) 271–290 trant’s total disclosure score, which is used as the basis for the dependent variables in this study. to reiterate, this study uses four conditions and four dis factors to determine the dis and other study dvs: 1. does the item generally relate to non-investments–specific amoral or unethical behavior? dis factor � 1 2. does the item directly relate to registrant’s actions with respect to investments or theft? dis factor � 2 3. does the item relate to allegations against registrant without other implication of culpability? dis factor � 1 4. does the item relate to allegations against registrant with reasonable implication of culpability? dis factor � 2 both a sum and a product version of the dis calculation were considered for this study. the product version is used as it offers a more appropriate severity score gradient. for instance, as seen in table 1, conditions one and three—mere allegation of a non-advisory item—only produces a subscore of 1 using the product method, one-fourth as much as a finding of an advisory item like financial fraud that would yields a subscore of 4. the sum method would produce a subscore of 2 for a disclosure of a mere allegation of non-advisory fig. 1. typical finra form u4 questions. 278 j.m. camarda / financial services review 26 (2017) 271–290 issue, twice the value produced by the product form. this seems intuitively disproportionate. these subscores are ascribed to specific u4 questions as illustrated in fig. 2. 7. descriptive statistics as seen in table 2, only about 12% of the florida rr population holds any study designation at all. of those that do, cfps are by far the most numerous, with less than half as many chfcs, and with the cfa barely represented. the study population is overwhelmingly male, mostly employee, with over half holding insurance licenses and iar registrations. table 1 dis subscores as function of condition combinations and product form dis subscore condition combination condition 1 or 2 condition 3 or 4 dis question subscore as product of factors 1,3 not investment/theft specific no clear indication culpability 1 1,4 not investment/theft specific clear indication culpability 2 2,3 investment/theft specific no clear indication culpability 2 2,4 investment/theft specific clear indication culpability 4 dis � disclosure incidence score. fig. 2. dis subscoring of u4 questions. 279j.m. camarda / financial services review 26 (2017) 271–290 of note in table 3, non-cfa designees tend to be older and longer-registered than other non-designees. this is also true for insurance licensees and somewhat true for independent contractors. study designees are predominantly male as seen in table 4, with designation holders over 80% male. given the sample gender distribution from table 2, it is not surprising that all iv categories are mostly populated by males. we also see in table 4 that most cfps and cfas are employees, but that most chfcs are independent contractors. it is perhaps not surprising that most cfps and chfcs, and hence most designees, are insurance licensed given these financial planning designations include strong insurance components. it was unexpected that over half of cfas are licensed to sell life insurance, but this does underscore the study model’s utility in identifying cfas who play retail advisory roles, and who are perhaps required to be licensed to sell these products. table 4 also reveals that most independent table 2 independent variables counts and sample weights independent variable count in 26,667 sample percent of sample holds at least one of any designation 3,197 12.00% holds cfp 2,534 9.50% holds chfc 970 3.60% holds cfa 116 0.40% male 19,815 74.30% female 6,852 25.70% employee 16,871 63.30% independent contractor 9,796 36.70% dually licensed as rr and iar 14,964 56.10% holds life insurances/annuities license 15,324 57.50% cfp � certified financial planner; chfc � chartered financial consultant; cfa � chartered financial analyst; rr � registered representatives; iar � investment advisor representatives. table 3 mean ages and years registered by independent variable and individual designations independent variable � age � years registered any designation 51 23 no designation 42 10 cfp 53 22 non-cfp 47 15 chfc 57 26 non-chfc 47 15 cfa 48 18 non-cfa 48 15 male 48 16 female 47 13 independent contractor 52 17 employee 45 14 iar 48 17 non-iar 48 14 insurance license 52 19 no insurance license 42 10 cfp � certified financial planner; chfc � chartered financial consultant; cfa � chartered financial analyst; iar � investment advisor representatives. 280 j.m. camarda / financial services review 26 (2017) 271–290 contractors are insurance licensed, and that a slight majority of all rrs are also registered as fiduciary iars, with this tendency substantially increased for designees, particularly cfps and cfas. rr/iars tend to be employees. finally, table 4 also shows there’s a tendency for rrs to also be licensed as fiduciary investment advisors and to sell commissionable life insurance as well as commissionable securities as rrs. average misconduct scores for the sample and by independent variable are reviewed in table 5. it should be noted that the heavy skewness of the discrete dvs toward the zero bound of the distributions limits the interpretative value of these means. an alternative perspective is offered by the means of the binary dvs that report the percentage of subsample disclosing misconduct by dv, without magnitude. it seems noteworthy that scores and percentage misconduct uniformly rise for the cfp and chfc designations, but that the opposite is observed for cfas. in other words, cfps and chfcs, respectively show higher misconduct than non-cfps and non-chfcs, and the percentages of sample showing any misconduct at all for each measure is higher for these designees than for those without them. while these observations are mitigated somewhat by regression against control ivs, it is interesting that the three designations share similar age and gender profiles, and that gender is clearly the single most powerful confounding factor appearing in these descriptive statistics, with males showing consistently sharply higher misconduct. besides maleness, iar and insurance licensee status also show strong associations with higher misconduct. 8. test hypotheses the null hypothesis is that the study misconduct measures are not affected by having any one of the study designations, so that that there is no disclosure difference associated with table 4 gender, employment, insurance license, and rr/iar status distribution by independent variable independent variable male female iar not iar independent contractor employee insurance license no insurance any designation 82% 18% 70% 30% 49% 51% 89% 11% no designation 73% 27% 54% 46% 35% 65% 53% 47% cfp 81% 19% 74% 26% 45% 55% 91% 9% non-cfp 74% 26% 54% 46% 36% 64% 54% 46% chfc 86% 14% 59% 41% 65% 35% 92% 8% non-chfc 74% 26% 56% 44% 36% 64% 56% 44% cfa 84% 16% 76% 24% 25% 75% 55% 45% non-cfa 74% 26% 56% 44% 37% 63% 57% 43% male 100% 0% 59% 41% 39% 61% 60% 40% female 0% 100% 47% 53% 30% 70% 50% 50% independent contractor 79% 21% 44% 56% 100% 0% 74% 26% employee 72% 37% 63% 37% 0% 100% 48% 52% iar 79% 21% 100% 0% 29% 71% 64% 36% non-iar 69% 31% 0% 100% 47% 53% 49% 51% insurance license 78% 22% 62% 38% 47% 53% 100% 0% no insurance license 70% 30% 48% 52% 23% 77% 0% 100% cfp � certified financial planner; chfc � chartered financial consultant; cfa � chartered financial analyst; rr � registered representatives; iar � investment advisor representatives. 281j.m. camarda / financial services review 26 (2017) 271–290 having any designation compared with having none. the alternate is that scores will be different for holders of at least one of these designations. it is expected that designated rrs should show lower misconduct. h0: �anydes � 0 h1: �anydes � 0 the primary regression method is tobit; data are censored, with results heavily skewed, with most at or close to zero. logit is the primary, and ols a secondary, robustness check. the regressions explore if having any of the study designations is associated with lower misconduct after controlling for other factors available in the dataset, including those identified in the descriptive statistics to be associated with higher levels of misconduct, such as insurance licensure, iar registration, and maleness. as noted, higher misconduct scores are generally associated with designations, the exception being cfas, which are quite scarce in the sample. the regressions seek to unbundle the effects. the tobit and ols regression models specifications are: dis � �1 � �yr � �a � �g � �ic � �rr&iar � �ins � �anydes � � (1) adis � �1 � �yr � �a � �g � �ic � �rr&iar � �ins � �anydes � � (2) cad � �1 � �yr � �a � �g � �ic � �rr&iar � �ins � �anydes � � (3) table 5 dis, adis, cad, and binary analog misconduct by designation, gender, employment status, and insurance and iar licensure independent variable � b-cad � b-adis � b-dis � cad � adis � dis sample mean 0.12 0.15 0.24 0.75 0.86 1.10 any designation 0.18 0.22 0.28 1.10 1.23 1.41 no designation 0.11 0.13 0.23 0.71 0.81 1.06 cfp 0.19 0.22 0.28 1.11 1.25 1.42 non-cfp 0.11 0.14 0.23 0.72 0.82 1.07 chfc 0.18 0.21 0.27 1.10 1.23 1.43 non-chfc 0.12 0.14 0.24 0.74 0.85 1.09 cfa 0.09 0.13 0.17 0.52 0.62 0.70 non-cfa 0.12 0.15 0.24 0.75 0.86 1.11 male 0.14 0.17 0.26 0.91 1.04 1.28 female 0.05 0.07 0.17 0.29 0.34 0.57 independent contractor 0.13 0.15 0.27 0.82 0.93 1.24 employee 0.11 0.14 0.22 0.71 0.82 1.02 iar 0.14 0.18 0.26 0.90 1.04 1.27 non-iar 0.08 0.10 0.20 0.56 0.64 0.90 insurance license 0.16 0.20 0.29 1.02 1.17 1.41 no insurance license 0.06 0.07 0.17 0.39 0.44 0.68 cfp � certified financial planner; chfc � chartered financial consultant; cfa � chartered financial analyst; iar � investment advisor representatives. cad is for culpable advisory disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. adis is for advisory disclosure incidence score and measures allegations and findings of advisory-related misconduct. dis is for disclosure incidence score and measures allegations and findings of advisoryand non-advisory–related misconduct. b-versions of dis, adis, and cad are binary transformations of the corresponding continuous variables. 282 j.m. camarda / financial services review 26 (2017) 271–290 the logit regression models specifications are: b-dis � �1 � �yr � �a � �g � �ic � �rr&iar � �ins � �anydes � � (4) b-adis � �1 � �yr � �a � �g � �ic � �rr&iar � �ins � �anydes � � (5) b-cad � �1 � �yr � �a � �g � �ic � �rr&iar � �ins � �anydes � � (6) where yr � years registered a � age g � gender ic � independent contractor rr&iar � dual registration as rr and iar ins � rr is life insurance sales licensed anydes � rr holds any of cfa, cfp, or chfc designations 9. regressions results as seen in table 6, having at least one of the study designations is associated with lower misconduct by all three measures with high significance in the tobit. having an insurance table 6 tobit dis, adis, and cad variable coefficients, significance, and (standard errors) by years registered, age, gender, employment status, iar and insurance licensure status, and presence of at least one study designation independent variable dis adis cad years registered 0.213 *** 0.449 *** 0.522 *** (0.009) (0.014) (0.017) age 0.008 — �0.039 *** �0.038 *** (0.007) (0.012) (0.014) gender 1.818 *** 3.542 *** 3.893 *** (0.155) (0.263) (0.317) independent contractor 0.460 *** �0.343 — 0.030 — (0.137) (0.210) (0.246) iar 0.841 *** 1.940 *** 1.921 *** (0.13) (0.21) (0.25) insurance license 0.937 *** 2.805 *** 2.840 *** (0.146) (0.235) (0.279) any designation �1.155 *** �0.984 *** �1.020 *** (0.191) (0.268) (0.311) iar � investment advisor representatives. *, **, *** left-adjacent coefficient significant at the less-than 10%, 5%, or 1% level, respectively. associated standard error appears beneath the corresponding coefficient in parentheses. cad is for culpable advisory disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. adis is for advisory disclosure incidence score and measures allegations and findings of advisory-related misconduct. dis is for disclosure incidence score and measures allegations and findings of advisoryand non-advisory–related misconduct. b-versions of dis, adis, and cad are binary transformations of the corresponding continuous variables. 283j.m. camarda / financial services review 26 (2017) 271–290 license, being male, and being dually registered as an iar is associated with higher misconduct, also all with high significance. please see the results section for more detailed discussion. the ols results are presented in table 7. having a designation is associated with lower misconduct by all three measures with high significance in the ols. please see the results section for detailed discussion. readers are reminded that because the data are bounded at zero, there is a non-normal error distribution, rendering ols inappropriate as a primary regression technique because of possible gauss-markov violations. ols is still a reasonable estimator for these data, and presented as a supplemental robustness check on the tobit results. logit regression results are presented in table 8. logit interpretations are different from typical regressions such as ols, and generally focus on a measure called the odds ratio instead of regression coefficients. essentially, the odds ratio gives the probability that a specific study iv is associated with a yes or no value for the misconduct dv. if an iv odds ratio is 0.25 for a misconduct measure of 1 or yes, then that misconduct is not likely to be present, and this should be interpreted as a good or 75% chance that a particular subject does not have this misconduct disclosed compared with a subject that lacks the categorical iv. odds ratios lower than one are proportionately associated with less likelihood of the condition being present, and those higher than one with proportionately greater likelihoods; those close to one are analogous to zero coefficients in typical regressions. the odds ratios results here show a reduced probability of all three measures of misconduct associated with having at least one study designation, all with high significance. probabilities of higher misconduct for all measures are associated with maleness, dual iar registration, and table 7 ols dis, adis, and cad variable coefficients, significance, and (standard errors) by years registered, age, gender, employment status, iar and insurance licensure status, and presence of at least one study designation independent variable dis adis cad years registered 0.077 *** 0.080 *** 0.076 *** (0.002) (0.002) (0.002) age �0.003 * �0.005 *** �0.005 *** (0.002) (0.002) (0.002) gender 0.456 *** 0.424 *** 0.372 *** (0.040) (0.037) (0.035) independent contractor 0.046 — �0.077 ** �0.053 — (0.038) (0.036) (0.034) iar 0.141 *** 0.134 *** 0.102 *** (0.036) (0.034) (0.032) insurance license 0.050 — 0.053 — �0.001 — (0.040) (0.037) (0.035) any designation �0.375 *** �0.292 *** �0.261 *** (0.055) (0.052) (0.049) iar � investment advisor representatives. *, **, *** left-adjacent coefficient significant at the less-than 10%, 5%, or 1% level, respectively. associated standard error appears beneath the corresponding coefficient in parentheses. cad is for culpable advisory disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. adis is for advisory disclosure incidence score and measures allegations and findings of advisory-related misconduct. dis is for disclosure incidence score and measures allegations and findings of advisory-and non-advisory–related misconduct. b-versions of dis, adis, and cad are binary transformations of the corresponding continuous variables. 284 j.m. camarda / financial services review 26 (2017) 271–290 insurance licensure. please see the results section for detailed discussion. logit is the primary robustness check as unlike for ols, the data can be made appropriate to the logit technique via binary transformations. logit odds ratios (�) as the probability of misconduct are discussed, and � plots are provided in figs. 3 (b-dis), 4 (b-adis) and 5 (b-cad). as seen in fig. 3 for b-dis, the odds of misconduct go down with a study designation, but meaningfully up for dual rr/iars, insurance licensure, and for maleness, with the latter quite pronounced. fig. 4, the odds ratios for b-adis, shows the probability of misconduct goes down with a study designation, but substantially up for dual rr/iars, insurance licensure, and for maleness, with the latter two quite pronounced. in fig. 5, for b-cad, the odds of misconduct go down with a study designation, but substantially up for dual rr/iars, insurance licensure, and for maleness, with the latter two quite pronounced. as seen in these three figures, logit results are very consistent for each misconduct measure; please see results section for detailed discussion. 10. results discussion, summary, and conclusions dis is this study’s broadest measure of misconduct disclosure, including all items both advisory and non-advisory–related and including allegations as well as findings of misconduct. having any study designation is associated with lower dis in all of the controlled table 8 logit b-dis, b-adis, and b-cad variable coefficients, significance, odds ratios (�), and (standard errors) by years registered, age, gender, employment status, iar and insurance licensure status, and presence of at least one study designation independent variable b-dis psi b-adis psi b-cad psi years registered �0.047 *** 0.954 �0.083 *** 0.920 �0.090 *** 0.914 (0.002) (0.003) (0.003) age �0.001 — 0.999 0.008 *** 1.008 0.007 *** 1.007 (0.002) (0.002) (0.003) gender 0.213 *** 1.532 0.351 *** 2.018 0.363 *** 2.065 (0.019) (0.027) (0.030) independent contractor 0.041 ** 1.084 �0.067 *** 0.874 �0.033 — 0.935 (0.016) (0.020) (0.022) iar 0.121 *** 1.274 0.217 *** 1.544 0.203 *** 1.502 (0.016) (0.021) (0.023) insurance license 0.137 *** 1.315 0.314 *** 1.875 0.293 *** 1.796 (0.017) (0.023) (0.026) any designation �0.129 *** 0.773 �0.084 *** 0.846 �0.079 *** 0.854 (0.022) (0.025) (0.027) iar � investment advisor representatives. *, **, *** left-adjacent coefficient significant at the less-than 10%, 5%, or 1% level, respectively. associated standard error appears beneath the corresponding coefficient in parentheses. cad is for culpable advisory disclosure and measures regulatory or judicial findings or other strong indications of clear advisory-related misconduct. adis is for advisory disclosure incidence score and measures allegations and findings of advisory-related misconduct. dis is for disclosure incidence score and measures allegations and findings of advisory-and non-advisory–related misconduct. b-versions of dis, adis, and cad are binary transformations of the corresponding continuous variables. 285j.m. camarda / financial services review 26 (2017) 271–290 regressions with high significance. it should be noted, however, that rrs who are male, have an insurance license, or are also registered as iars are associated with higher dis scores. the negative influence of the non-designation factors appears to exert a strong influence on misconduct scores, even for those with designations. it seems noteworthy that being male or fig. 3. b-dis and anydes logit odds ratio graph. fig. 4. b-adis and anydes logit odds ratio graph. 286 j.m. camarda / financial services review 26 (2017) 271–290 also an iar is associated with higher dis misconduct to p � 0.0001 in all three tests, and that having an insurance license also is with the same p value in both tobit and logit. exploration of non-designation effects is left for subsequent research. for dis and b-dis, the findings suggest that the null hypothesis h0: �anydes � 0 can be rejected with a high degree of confidence, with associated p values less than 0.0001 in all three regression tests. consequently, lower dis misconduct seems associated with having at any least one of the study designations. the anydes adis and b-adis findings are similar: the null that associated misconduct does not change when any study designation is present can be rejected with high confidence, in this case with p � 0.001 for tobit and logit, and ols p � 0.0001. the anydes cad and b-cad findings are generally similar to those for dis and adis. recall that cad is the study’s most serious misconduct measure, limited to advisory items only where misconduct is not merely alleged but nearly certain. for the cad dvs, the reduction for anydes is quite robust, with p � 0.001 for tobit, �0.004 for logit, and �0.0001 for ols. once again, the null that misconduct does not change when any study designation is present can be rejected with high confidence. as for dis and adis, similar strongly higher misconduct associations are seen for the cads for gender (all ps � 0.0001), iar (highest p is from ols at �0.002) and with insurance licensure from tobit and logit (both ps � 0.0001; but note that ols finds no significance for insurance). this study’s results support a strong and robust finding that having at least any one of the study designations is associated with lower misconduct. this is found with high statistical significance using the three different regression techniques. these findings are uniform across the severity spectrum from dis to cad, when controlling factors are considered. both allegations and findings for advisory and non-advisory misconduct matters fall fig. 5. b-cad and anydes logit odds ratio graph. 287j.m. camarda / financial services review 26 (2017) 271–290 uniformly for those in the anydes category. for florida rrs at least, holding at least one of the study designations seems to be a strong and reliable quality signal, consistent with the marketing claims of the various conferring institutions, in which we can have confidence. it should be noted that this finding is not consistent with those from the descriptive statistics section, where higher misconduct for cfps and chfcs was found. this may be because of the confounding influence of misconduct-increasing factors such as insurance licensure, iar registration, and maleness. the fact that all three of these characteristics are associated with higher misconduct affirms the descriptive statistics findings, and they all appear more influential than the designation effect. it is possible that these compensation conflicts somewhat outweigh the good that designations may confer. the temptation of high and poorly disclosed insurance or securities commissions, perhaps misleadingly obtained under a false iar or designation fiduciary flag where consumers are allowed to believe their interests are put first, may influence or even enable the designee to act less ethically than implied by the signal, inducing complaints and other misconduct markers. this may be one explanation for the descriptive statistics finding that misconduct rises for chfcs and cfps over non-designees, but the question is left for further research, including a study of misconduct by specific designation now underway by this author. the discovered gender effect has now been confirmed by subsequent research (egan, matvos, & seru, 2017), but an exploration of its likely complex causes is deferred to later study. references barry, r., & eaglesham, j. (2014). regulatory breakdowns in oversight of us stockbrokers. wall street journal [whitepaper]. brockman, c., & brooks, r. (1998). the cfa charter: adding value to the market. financial analysts journal, 54, 81–85. cfa institute. (2015). code of ethics and standards of professional conduct. (available at http://www. cfainstitute.org/ethics/codes/ethics/pages/index.aspx) cfa institute. (2016). value of the cfa charter. (available at https://www.cfainstitute.org/programs/cfaprogram/ charter/value/pages/index.aspx) cfp board. (2015). 2015 annual brand tracking report: executive summary. (available at http://www. letsmakeaplan.org/docs/default-source/default-document-library/2015-annual-brand-tracking-report-executivesummary.pdf) cfp board. (2016). cfp certification: the standard of excellence. (available at http://www.cfp.net/about-cfpboard/cfp-certification-the-standard-of-excellence) cfp board standards of professional conduct. (2008/2014). (available at cfp.net/docs/ethics-enforcement/ standards-of-professional-conduct-final.pdf?sfvrsn�2) cordell, d., smith, r., & terry, a. (2011). overconfidence in financial planners. financial services review, 20, 253–263. cummings, b., & finke, m. (2010). the economics of fiduciary investment advice (ssrn working paper 1701181). (available at ssrn.com/abstract�1701181) dean, p. (1997). examining the profession and practice of business ethics. journal of business ethics, 16, 1637–1649. egan, m., matvos, g., & seru, a. (2016). the market for financial adviser misconduct. (available at http:// www.nber.org/papers/w22050) finra. (2016). registered representatives brochure. (available at http://www.finra.org/sites/default/files/ registered-representatives-brochure.pdf) 288 j.m. camarda / financial services review 26 (2017) 271–290 financial planning coalition. (2014). consumers are confused and harmed: the case for regulation of financial planners. (available at http://financialplanningcoalition.com/wp-content/uploads/2014/06/financialplanning-coalition-regulatory-standards-white-paper-final.pdf) finke, m., & langdon, t. (2012). the impact of the broker-dealer fiduciary standard on financial advice. journal of financial planning, 25, 28–37. flexner, a. (1976). is social work a profession? proceedings of the national conference of charities and corrections. (available at http://scholar.google.com/scholar?hl�en&q�is�social�work�a�profession� flexner�and�metzger�college�and�university�1976&btng�&as_sdt�1%2c10&as_sdtp�) flexner, a., & metzger, w. (1976). is social work a profession? what is a profession? college and university, 52, 42–55. frumento, a., & korenman, s. (2013). professionalism and investment advisors. journal of investment compliance, 14, 32–41. government accountability office. (2011). consumer finance: regulatory coverage generally exists for financial planners, but consumer protection issues remain (gao-11–235). washington, dc: government printing office. hauptman, m., & roper, b. (2017). financial advisor or investment salesperson? washington, dc: consumer federation of america. khurana, r., & nohria, n. (2008). it’s time to make management a true profession. harvard business review, 86, 70–77. khurana, r., nohria, n., & penrice, d. (2005). is business management a profession? harvard business school website. (available at hbswk.hbs.udu/item/4650.html) kitces, m. (2014b). the rise of the financial planning academic. (available at kitces.com/blog/the-rise-of-thefinancial-planning-academic/) kitces, m. (2015). cfp board’s reduced experience requirement quietly takes effect with fpa and napfa silence as tacit support? (available at https://www.kitces.com/blog/cfp-boards-reduced-experiencerequirement-quietly-takes-effect-with-fpa-and-napfa-silence-as-tacit-support/) kitces, m. (2015b). are the licensing and other requirements to become a financial advisor too easy? (available at https://www.kitces.com/blog/are-the-licensing-and-other-requirements-to-become-a-financialadvisor-too-easy/) mauldin, d., wilder, w., & stocks, m. (2000). does aicpa accreditation of nonaudit services add value? the case of personal financial planning. accounting horizons, 14, 49–67. mccann, c., qin, c., & yan, m. (2016). how widespread and predictable is stock broker misconduct? (available at securities litigation and consulting group: http://papers.ssrn.com/sol3/papers.cfm?abstract_id�2768942) moisand, d. (2008). the financial planning act of 2008. journal of financial planning, 21, 78–90. moisand, d. (2012). so close and yet so far. journal of financial planning, 25, 26–27. piaba foundation. (2015). piaba brokercheck letter to sec re file 265–28. (available at https://www. sec.gov/comments/265-28/26528-98.pdf) ragatz, j., & duska, r. (2010). financial codes of ethics. in j. boatright (ed.), finance ethics: critical issues in theory and practice. new york, ny: j. wiley and sons. raymond james mission statement. (2016). mission statement. (available at http://www.raymondjames.com/ about/mission.htm) ritholtz, b. (2016). brokers behaving badly. (available at https://www.bloomberg.com/view/articles/2016-0302/broker-misconduct-is-worse-than-we-thought) schaefer, t. (1984). professionalism: foundation for business ethics. journal of business ethics, 3, 269. smith, r., vibhakar, a., & terry, a. (2008). demarcating designations: chartered financial analyst and certified financial planner. journal of financial services marketing, 12, 299–310. spence, m. (1973). job market signaling. the quarterly journal of economics, 87, 355–374. spence, m. (2002). signaling in retrospect and the informational structure of markets. the american economic review, 92, 434–459. terry, a., & vibhakar, a. (2011). a comparative analysis of the cfa and cfp designations. financial services review, fall, 66–81. 289j.m. camarda / financial services review 26 (2017) 271–290 the american college. (2016b). chfc: chartered financial consultant. (available at https://www.theameri cancollege.edu/designations-degrees/chfc) the american college. (2016a). chfc: the highest standard. (available at http://www.chfchigheststandard. com/) the american college. (2016c). code of ethics. (available at https://www.theamericancollege.edu/about-thecollege/community-responsibility/code-of-ethics) the white house. (2016). the effects of conflicted investment advice on retirement savings. washington, dc: government printing office. vystar ad. (2016). florida times-union, b-8. (http://jacksonville.com/news/metro/2017-08-09/morris-publishingsells-florida-times-union-gatehouse-media) warschauer, t. (2002). the role of universities in the development of the personal financial planning profession. financial services review, 11, 201–216. woolley, s. (2016). it just got harder to trust financial advisors. (available at https://www.bloomberg.com/ news/articles/2016-03-01/it-just-got-even-harder-to-trust-financial-advisers) zingales, l. (2015). does finance benefit society? the journal of finance, 70, 1327–1363. 290 j.m. camarda / financial services review 26 (2017) 271–290 financial services review, 32(3) 32 impact of the financial advisor on clients’ financial outcomes: an integrative model pierre-etienne pilote,1 emilio boulianne,2 and michel magnan3 abstract a financial advisor may either act as a consultant or may be delegated the entire financial advising process. in both cases, researchers tend to conclude that advisors have an impact on their clients’ financial outcomes. however, there is no agreement on the nature and extent of this impact. we argue that such discordance in the results being reported in prior research arises from the different theoretical lens used to observe the phenomenon: agency theory, trust theory, and the concept of knowledge. in our view, the complexity of advisors’ contribution to their clients’ outcomes requires a novel approach that extends beyond a single theory. relying upon a literature review, we propose an integrative multi-theory model that reconciles prior findings and illustrates how financial advisers impact clients’ outcomes at each step within the financial advising process. practice-grounded, the model also provides a causal mechanism clarifying the opaque and complex services provided by the financial advisor. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation pilote, p-e., boulianne, e., & magnan, m. (2024). impact of the financial advisor on clients’ financial outcomes: an integrative model. financial services review, 32(3), 32-67. introduction there is extensive evidence indicating that financial advisors impact their clients’ financial outcomes (angelova & regner, 2013; cici et al., 1 corresponding author (pilote.pierre-etienne@uqam.ca). école des sciences de la gestion, université du québec à montréal, montréal, canada and john molson school of business, concordia university, montréal, canada. for valuable comments, we thank participants and discussants at the 2024 research workshop on management control (jason moschella, hec montréal), 2023 montreal business schools phd symposium (olivier greusard, esguqam) and 2023 esg-uqam workshop. 2 john molson school of business, concordia university, montréal, canada. emilio boulianne acknowledges the financial support from the kpmg entrepreneurial research studies and the autorité des marchés financiers (amf). 3 john molson school of business, concordia university, montréal, canada and cirano, montréal, canada. michel magnan acknowledges the financial support from the s.a. jarislowsky chair of corporate governance as well as from the institute for the governance of private and public organizations. 2017; martin & finke, 2014). however, the scope and nature of such impacts diverge. this situation leads us to put forward the following research question: “what is the impact of the financial advisor on their clients’ financial outcomes?” to https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(3) 33 achieve this aim, this paper reviews the literature and analyzes such research to propose an integrated view on the relation between financial advisors and clients’ financial outcomes. financial advisors are the professionals to whom clients turn when they do not want to deal with their personal finances alone. when an individual considers their personal finances, three options are available: (1) acting alone, unassisted, (2) consulting a financial advisor while keeping decision-making autonomy, or (3) delegating all steps to a financial advisor (calcagno et al., 2017; tang & hu, 2019). many factors come into play in deciding whether to act alone or to seek a financial advisor (balasubramnian & brisker, 2016; finke et al., 2011; west, 2012). the evolving complexity of personal finance resulting from the advent of new regulations, novel financial products, and wealth accumulation increasingly leads individuals to solicit the services of financial advisors.4 according to the financial industry regulatory authority (finra), the term “financial planners” refers to investment advisors, brokers, insurance agents, and accountants.5 the financial consumer agency of canada also states that “a financial advisor is a general term that can apply to anybody who helps you manage your money. this could include an employee of your financial institution, a stockbroker or an insurance agent.”6,7 in this study, a financial advisor is a professional working in one of the fields related to personal finance and offering services to individuals. in using a broad definition of the term “financial advisor,” we take the same approach as prior research that counts financial planners, financial advisor, accountants, lawyers, notaries, brokers, bankers, insurance agents, and related financial credentials as advisors (see 4 in the united states, the cfp board (2023) reports that 53% of americans have a financial planner. in canada, 49% of canadians obtain advice from a professional financial advisor/planner (financial consumer agency of canada, 2019). 5 retrieved from: https://www.finra.org/investors/investing/workingwith-investment-professional/financial-planners (accessed on january 14, 2024). gennaioli et al., 2015; hudson & palmer, 2014; kim et al., 2018; lei, 2019). financial advice covers a broad range of activities. for instance, personal financial planning (pfp) associations and accreditation bodies generally identify six or seven core areas of financial advising, namely: (1) financial management, (2) insurance and risk management, (3) investment planning, (4) retirement planning, (5) tax planning, and (6/7) estate planning and legal aspects (finke et al., 2009; fp canada, 2019). an individual consulting with an advisor may seek advice on a specific topic or they may wish to implement a comprehensive plan (winchester & huston, 2015), but regardless, they expect added value (sweeney et al., 2018). financial advisors aim to fill a knowledge gap or provide a sense of security to the client (bae & sandager, 1997). the outcomes associated with the advising process depend not only on the steps in which the advisor is involved, but also on how the advisor acts. to answer our research question, we use a theoretical review (paré et al., 2015) with the goal of developing a model of the financial advice process that integrates three theoretical perspectives underlying prior research on financial advising. first, agency theory places the financial advisor’s self-interest at the forefront, ahead of the client’s interests, resulting in a negative impact for the client (angelova & regner, 2013; beyer et al., 2013; mullainathan et al., 2012). second, trust theory implies that the advisor inspires trust and makes recommendations in the client’s best interest (barnett white, 2005; lachance & tang, 2012). how trust in the advisor may impact the outcomes of the advising process remains an open question. third, the concept of knowledge emphasizes the role played by the advisor’s proven knowledge 6 retrieved from: https://www.canada.ca/en/financialconsumer-agency/services/savingsinvestments/choose-financial-advisor.html (accessed on january 14, 2024). 7 in the united states, the public often misunderstand the term “financial advisor” (lach et al., 2019; tharp, 2019). https://www.finra.org/investors/investing/working-with-investment-professional/financial-planners https://www.finra.org/investors/investing/working-with-investment-professional/financial-planners https://www.canada.ca/en/financial-consumer-agency/services/savings-investments/choose-financial-advisor.html https://www.canada.ca/en/financial-consumer-agency/services/savings-investments/choose-financial-advisor.html https://www.canada.ca/en/financial-consumer-agency/services/savings-investments/choose-financial-advisor.html pilote et al. 34 and expertise compared to the client’s limited knowledge, suggesting a positive impact for the client (hershey et al., 1990; hershey & walsh, 2000). the theoretical perspectives we adopt to analyze the literature on the financial advising process embed the concept that services rendered by the advisor qualify as a “credence good” as the client may have difficulty evaluating their ultimate effectiveness because of a lack of knowledge (bruhn & miller, 2014; gennaioli et al., 2015; winchester & huston, 2017). a credence good consists of a complex and opaque service that requires specialized knowledge (darby & karni 1973; dulleck & kerschbamer, 2006; dulleck et al., 2011; fong, 2005). as the client may only see the outcomes of financial advice without knowing or understanding how the advisor gets there, agency problems may arise, explaining why agency theory is widely used in financial services research. the information gap between client and advisor also explains the use of trust theory as clients are taking a risk by entrusting advisors with their financial affairs since they are unable to know the details of their advisor’s actions and decisions. finally, the specialized nature of the services rendered by the advisor explains the use of the concept of knowledge in our analysis. from a practical standpoint, our review relies on the six steps8 of the pfp process developed by the certified financial planner (cfp) board of standards, inc. (2017) to proxy for the process of providing financial advice: (1) agreeing on how to work together, (2) gathering information, (3) analyzing the client's situation, (4) providing recommendations, (5) implementing the recommendations, and (6) monitoring progress (cfp board, 2017). put forward to frame the work of cfp professionals, we consider that these steps represent a golden standard in financial services and remain general enough to fit most financial advisory services. in this regard, we 8 we constructed our model based on these six steps. we are aware that the cfp board recently introduced a seven-step model, which is very similar to the six steps. our rationale to keep the six-step model is that prior research used for the model has been published expect most financial advisors to follow similar steps as those proposed by the cfp board. although some articles specify which step they address (e.g., angelova & regner, 2013; barnett white, 2005; hershey et al., 1990), the usage of steps to analyze the articles is mainly ours. in our view, using these steps allows researchers to bridge the gap between research and practice. our main research contributions are as follows. first, we propose a theory-driven review of the financial advising literature, relying on three theoretical lenses to explain the financial advisor’s impact. second, we propose an integrative model of the financial advisor’s impact, linking the three theoretical lenses and explaining the divergent results found in prior research. this model allows one to integrate the reality of the financial advising environment by linking the effects of the advisor’s level of involvement to the steps performed in practice, thus enhancing the model’s relevance for both academics and practitioners. lastly, based on four financial advising core questions developed in this paper, we contribute to better capturing the key concept of credence good, as provided by a financial advisor’s services, and provide explanations on how an advisor’s involvement will impact their clients’ financial outcomes compared to when an individual decides to act in an unassisted manner. theorization of financial advising agency theory agency theory refers to an arrangement whereby the principal delegates some of their decisionmaking power to an agent. as each party seeks to maximize their personal interests, the goals pursued by the principal are in opposition with the goals pursued by the agent (jensen & meckling, 1976). this conflict is embedded in an during the tenure of the six-step model, and the new step is a split of existing ones. after analysis, we did not find significant differences when we looked at our data with an additional step. financial services review, 32(3) 35 asymmetry of information favoring the agent (ross, 1973; wright et al., 2001). financial advising meets the characteristics associated with agency theory. there exists a contractual relationship between a principal (the client) and an agent (the advisor), a clash of goals wherein the client wants their assets to grow, while the advisor seeks to derive personal benefit from the actions taken on the client’s behalf. to overcome this conflict of interest, the client can either set up control mechanisms to monitor the advisor’s actions or motivate the advisor to act in the best interests of both parties by offering incentives (jensen & meckling, 1976; wright et al., 2001). unfortunately, the impact on the outcomes of these controls is uncertain, and controls and incentives add costs to the client. for instance, research indicates that individuals are unable to answer basic financial literacy questions (boisclair et al., 2017; lusardi & mitchell, 2011, 2014).9 thus, clients lack the knowledge to assess the complex steps and actions carried out by an advisor, challenging their ability to implement effective controls. further, clients are in a difficult position to incentivize advisors, since they do not have full control over all the incentives offered to the advisor. for example, the advisor’s employer may compensate the advisor to sell certain products, which influences the advisor’s recommendations and actions (inderst & ottaviani, 2012a, 2012b; kingston & weng, 2014). in such a case, it becomes difficult and costly for the client to implement incentives that can really drive behavior, such as prioritizing the client’s interest. moreover, most clients cannot estimate how much their advisor’s services cost 9 research around the world indicates that between 30% and 55% of respondents correctly answer the three core questions on financial literacy. in canada, the rate is 42% (boisclair et al., 2017), while in the united states the rate is 30% (lusardi & mitchell, 2011). the three core questions are: question 1, on interest rates: “suppose you had $100 in a savings account and the interest rate was 2% per year. after five years, how much do you think you would have in the account if you left the money to grow? more than them, or do not know how their advisor is compensated (cheng & kalenkoski, 2018). trust theory several definitions of trust exist, but the one posited by mayer et al. (1995) is relevant to our study: “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party” (p. 712). common to most definitions is that a trust relationship cannot exist without the presence of a person who trusts (trustor), a trusted party (trustee), and a context involving risk-taking by the trustor. it is remarkable how financial advising meets the characteristics of trust theory. first, it involves a relationship between a person who trusts (the client) and a person the client trusts (the advisor). second, by delegating all or some steps to the advisor, the client is de facto taking a large risk and finds themselves in a vulnerable position. for instance, at the gathering information, analysis, and recommendation steps of the financial advising process, the client must let the advisor take actions that not only influence their financial outcomes, but whose accuracy cannot be fully assessed. the concept of knowledge three types of knowledge are widely recognized (paris et al., 1983). first, declarative knowledge (“know what”) is the accumulation of concepts, facts, rules, laws, and principles (anderson, 1982, 1983; gupta & cohen, 2002; paris et al., 1983). tardif (1992) considers declarative knowledge as fundamentally static, rather than dynamic, and $102; exactly $102; less than $102; don’t know”; question 2, on inflation: “imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after one year, how much would you be able to buy with the money in this account? more than today; exactly the same; less than today; don’t know;” and question 3, on diversification: “is the following statement true or false? buying a single company’s stock usually provides a safer return than a stock mutual fund. true; false; don’t know.” pilote et al. 36 thus must be translated into procedural or conditional knowledge to enable action. second, procedural knowledge (“know how”) relates to the steps and procedure followed to carry out an action (tardif, 1992). but, as anderson (1982) points out, declarative knowledge is necessary to implement procedural knowledge, as it would not be useful to know “how to do” if one does not know “what to do.” third, conditional knowledge (“know when and why”) refers to the conditions surrounding the action (schunk, 2012; tardif, 1992). for instance, when and in what context is it appropriate to use a specific strategy, to favor this or that approach, or to take this or that action? gaining an understanding of “what, how, when, and why,” based on declarative and procedural knowledge, positions conditional knowledge at the top of the knowledge pyramid. knowledge may also be viewed as domainspecific or general (perkins & salomon, 1989; schunk, 2012). specific knowledge applies to explicit domains and is of more limited use, while general knowledge refers to skills like reading and writing, transcending disciplinary fields and extending to many domains and situations (perkins & salomon, 1989; schunk, 2012). researchers agree that solving a given problem requires both specific and general knowledge. further, domain-specific and general knowledge do not add to the three categories presented above, but rather complement them (tardif, 1992). when it comes to financial advising, individuals who deal with an advisor do so on the premise that, due to their cumulative general and specific knowledge, the advisor will add value to their outcomes (sweeney et al., 2018). financial advisors possess higher levels of financial literacy knowledge than individuals (azamian et al., 2022), are more analytical (nofsinger & varma, 2007), and rely on an established problem-solving approach (hershey et al., 1990). furthermore, the understanding acquired by financial advisors encompasses all kinds of knowledge (declarative, procedural, and conditional), namely, the “what, how, when, and 10 at this point, an article can be counted more than once if it appears in more than one database. why.” a large majority of individuals do not have such an advanced level of personal finance knowledge. in other words, due to their lack of knowledge, individuals are very unlikely to be able to adequately manage their personal finances, and more specifically when compared to financial advisors. method based upon a literature review, we developed a theory-grounded integrative model showing the impact of the financial advisor on their clients’ financial outcomes. our choice of a theoretical review rests on the following premises. first, from such a review may emerge theoretical conceptualizations of a phenomenon, or advancement at a theoretical level, otherwise known as a theoretical review (paré et al., 2015), a theory development review (templier & paré, 2018), or an integrative review (snyder, 2019). second, a theoretical literature review proves especially appropriate when studying a phenomenon that covers different fields of expertise drawing on various research perspectives (torraco, 2005). in addition, advisors, coming from different fields such as financial planning, accounting, law, notarial law, securities brokerage, insurance, and banking, to name a few, may be involved at different steps of the financial advising process. data collection the review comprised four phases as described below (figure 1). first, we searched, using three databases, the possible combinations of financial advisor and the three theoretical lenses. the search terms (presented in figure 1) needed to appear either in the title, the keywords, or the abstract of articles. based on these criteria, we identified 2,160 results.10 second, we sought to narrow down the results. we analyzed the title and abstract of each article based on three preestablished criteria: (1) the article must focus on the presence of an advisor in the financial advising process, (2) the article must explicitly or implicitly be linked to one of the three theoretical financial services review, 32(3) 37 lenses, and (3) the article must be from an academic journal.11 based on these criteria, we retained 131 articles.12 third, focusing on these 131 articles and their references, we identified relevant articles that may not have been identified in the initial search (see webster & watson, 2002). following this review, we added 106 articles, thus bringing the total number of articles to 237. finally, we read the 237 articles in their entirety and evaluated them using the same criteria used in the second phase, thus bringing our final dataset to include 88 articles. appendix a summarizes the 88 articles. for each article, the table provides the authors, the type of financial advisor, the method used, the measurement, the outcome, and the study findings. for a better analysis, and in line with the integrative model, we present the articles categorized through the three theoretical lenses. the dataset reflects the diversity of expertise and fields associated with financial advising, as it includes articles published in marketing, finance, economics, psychology, management, accounting, and financial planning. as an aside, even if the pfp field is of prime importance for individuals and the economy, it is surprising to see the limited number of journals dedicated to pfp. figure 1. literature review search steps 11 this criterion mostly explains the sharp reduction of articles since there is an important number of professional publications in financial planning. 12 from this point on, the results from the three databases were combined, eliminating the possibility of the same article being counted more than once. pilote et al. 38 data analysis the data analysis builds upon the three theoretical perspectives and relies on the six steps of the pfp process developed by the cfp board (2017) as a proxy for the financial advising process (figure 2). based on the combination of the theoretical perspectives and the six steps, we pinpoint where in the process and how the advisor affects their clients’ financial outcomes. figure 2. six steps of the financial advising process the first step, agreeing on how to work together, sets the parameters of the relationship between the financial advisor and the client. it consists of a deal specifying the services provided by the advisor, such as the next steps where the advisor will be involved. the second step, gathering information, covers the collection of information and documents from the client for analysis. at the third step, analyzing the situation, the advisor assesses the client’s financial situation to identify relevant strategies and options, weighing the benefits of each, to develop recommendations. at the fourth step, providing recommendations, the advisor provides recommendations to the client on how to best optimize their financial outcomes. the fifth step, implementing the recommendations, involves ensuring the implementation, either by the client or the advisor, according to the terms of the deal established in step 1. the sixth step, monitoring progress, closes the loop. it consists of assessing the evolution of the situation and adjusting when the conditions require it. the sixth step brings a circularity to the process as it leads to a relaunch of the process when needed. winchester and huston (2015) report that financial advice may fall into either a modular approach, when happening on an ad hoc basis regarding a specific topic, or an integrated approach, when advice follows a comprehensive process staggered over time. only the level of involvement of the actors (client or advisor) changes according to the six steps. for instance, figure 3 shows the six steps when a client delegates the process (from steps 2 through 6) to an advisor, while, when clients opt for consultation, they retain the right to implement or not the advisor’s recommendations (thus taking control of step 5). depending on the arrangement, step 6 (monitoring progress) may either be the responsibility of the client or part of the mandate given to the advisor. financial services review, 32(3) 39 figure 3. possible arrangements between the client and the advisor we consider that these six steps, even if developed to frame the financial planning process, adequately capture the financial advising process as they are quite generic and focus on a client-advisor relationship. findings three main findings arise from our analysis. first, the advisor’s impact on their clients’ financial outcomes depends upon the theoretical lenses used and the executed steps. second, the review brings forward the need for an integrative model, since the articles report divergent results. third, the review highlights other factors, such as the type of advisors involved, which refers to the holding of different professional designations. financial advising and agency theory prior research reports on the clash of interests at an aggregate level (dvorak, 2015; hackethal et al., 2012; mullainathan et al., 2012), as well as at specific steps (angelova & regner, 2013, 2018; chalmers & reuter, 2020; foerster et al., 2017). advisors can maximize their own interest at the expense of their clients, first by taking financial advantage of third-party offerings when issuing a recommendation and monitoring (anagol et al., 2017; christoffersen et al., 2013; hackethal et al., 2012), and second by providing minimal effort when analyzing and monitoring (dvorak, 2015; liu, 2005). the financial advisor’s effort is mainly observed through the customization of their clients’ portfolios. in a comparative analysis of advisory firms’ retirement plans and their clients’ retirement plans, dvorak (2015) shows that both retirement plans are very similar, with the exception that plans offered to clients favor investments with higher fees. advisors minimize the effort and maximize the fees charged to clients by limiting the customization of clients’ retirement plans. similarly, foerster et al. (2017) find that advisors offer an imperfect copy of their personal portfolio to their clients. by recommending a comparable asset allocation base to all clients, advisors limit their efforts while receiving compensation for the actions taken. foerster et al. (2017) estimate that clients annually pay fees of around 2.5% of their assets to obtain the same advice given to all other clients. bolton et al. (2007) find that if higher fees can be earned, the more the recommendations will be detrimental to the client. other tactics used by advisors include pressing clients to make more trades, which increases advisors’ fees (hackethal et al., 2012; hoechle et al., 2017), or recommending a suboptimal or wrong investment option to clients while favoring advisors’ fees (angelova & regner, 2013, 2018; kingston & weng, 2014). in short, consistent with an agency theory perspective, some advisors prioritize their own interests over their clients’ interests (burke et al., 2015). pilote et al. 40 however, a long-term horizon relationship (jarratt et al., 2007), a competitive market (bolton et al., 2007; stoughton et al., 2011), and the setting up of control mechanisms (calcagno et al., 2017) do mitigate the conflict of interests in financial advising. considering the precarious position of the clients, chalmers and reuter (2020) even suggest that an individual would receive a better service from well-constructed default financial advice than from a conflicted financial advisor. for instance, advanced software makes it possible nowadays to obtain such default financial advice at a low cost. financial advising and trust theory clients may either fully trust their advisor and thereby consent to the advisor’s power on their financial outcomes, or they may not fully trust the advisor, for a variety of reasons, and will prefer to limit the advisor’s impact. related to financial advising, two steps prove critical regarding the trust relationship: the agreeing step, when the client chooses to either delegate or consult the advisor on an ad hoc basis, and the implementing step, where the client relies on the advisor’s recommendations. our review reveals two main findings related to trust theory. first, the more the client trusts the advisor, the more likely it is that the client will completely delegate to the advisor (calcagno et al., 2017; gennaioli et al., 2015; monti et al., 2014). second, the more the client trusts the advisor, the more likely is it that the client follows the advisor’s recommendations (barnett white, 2005; deng & liu, 2017; eriksson & hermansson, 2019; georgarakos & inderst, 2014; johnson & grayson, 2005; lachance & tang, 2012; pauls et al., 2016). moreover, trust does allow clients to relieve themselves of the mental burden of managing their finances. for instance, in the context of a laboratory experiment measuring brain activity, engelmann et al. (2009) find that clients prefer to delegate the mental burden of making investment decisions by following advisors’ recommendations. financial advising and the concept of knowledge the advisor’s knowledge plays a key role during the financial advising process and ultimately impacts its effectiveness (hershey et al., 1990; hershey & walsh, 2000). translated into the sixstep framework, advisors seem particularly adept at : (1) gathering information, (2) analyzing the situation, and (3) monitoring progress. in short, equipped with superior knowledge, advisors may directly influence their client’s financial outcomes. for instance, regarding their personal financial situation, azamian et al. (2022) report that financial advisors are better prepared for retirement, have less debt, are better insured, and are more likely to have an estate plan than the public. for clients with an advisor, studies report greater value of assets held (liu et al., 2019), better portfolio diversification (pan et al., 2020), better use of tax instruments (cici et al., 2017), and reduced wealth volatility (grable & chatterjee, 2014). as positive influences on their clients, we may cite the clients’ saving practices (chatterjee & lu, 2023; fan, 2021; mountain et al., 2021) and the setting and committing to financial goals (marsden et al., 2011; martin & finke, 2014). some studies on advisors’ knowledge credit their positive impact on their clients’ financial outcomes to knowledge that can be viewed either as declarative (i.e., an accumulation of facts, concepts, or rules) (cloyd, 1995, 1997; hershey & walsh, 2000; spilker, 1995) or as procedural (i.e., systematic steps and procedures) (barrick & spilker, 2003; fischer & gerhardt, 2007; hershey et al., 1990). research indicates that advisors with specialized knowledge obtain better results in information search strategies, the quantity of results, and the relevance of the information found (barrick & spilker, 2003; cloyd, 1995, 1997; hershey & walsh, 2000; spilker, 1995). hence, an advisor’s knowledge positively affects the quality of the analysis performed (hershey & walsh, 2000), with its declarative and procedural forms being the most salient. such knowledge permits them to perform better problem-solving processes, reducing errors, and thereby providing higher quality solutions to clients (hershey et al., 1990). financial professional designations financial services review, 32(3) 41 holding a professional designation (e.g., cfp®) may modulate the impact of the advisor on their clients’ financial outcomes as it positively influences the client’s trust in the financial advisor (agnew et al., 2018; dean, 2017; guillemette & jurgenson, 2017; james, 2013) as well as the level of knowledge of financial advisors (arman & shackman, 2012; blanchett, 2019; kim et al., 2018; lei, 2019). underlying these findings is the evidence that holding a professional designation that includes a code of ethics induces the advisor to put the interests of their clients first (finke et al., 2009). for example, both in canada and the united states, cfps must follow the code of ethics’ principles of their designations. those codes include notably fiduciary duty, duty of loyalty to the client, integrity, competence, diligence, professionalism, and confidentiality (cfp board, 2018; fp canada, 2022). however, we leave the analysis of the impact of designations to future research. toward an integrative model of financial advising we present in this section an integrative model showing the impact of the financial advisor on their clients’ financial outcomes. we elaborate the model in two phases. first, in figure 4, we discuss and illustrate the impacts associated with each of the three theories for each step where they manifest. second, in figure 5, we present the integrative model of financial advising across the three theories, when the process is performed alone (unassisted) versus when the process is performed with a financial advisor. the integrative model shows how the interactions between the three theoretical lenses impact clients’ financial outcomes. impacts on clients’ financial outcomes associated with each theoretical lens figure 4 summarizes the findings where the advisor’s impact on their clients’ financial outcome may be positive or negative. figure 4 includes the case when clients choose to delegate the entire process to an advisor, as well as when clients opt for ad hoc consultations at certain steps. figure 4. summary of the impacts (positive or negative) of the financial advisor on their clients’ financial outcome pilote et al. 42 agency theory. using an agency theory perspective, incentives outside the client’s reach and the lack of control over the advisor’s actions creates a clash of interests that leads advisors to maximize their own interests at the expense of the client. more specifically, the lack of client control over the advisor’s actions is seen at the analyzing (foerster et al., 2017), recommendations (angelova & regner, 2013), and monitoring steps, while the negative impact of incentives plays out at the recommendations (angelova & regner, 2013) and monitoring steps (hoechle et al., 2017). figure 4 illustrates this direct negative impact where advisors are thinking of themselves first when they are presented with the opportunity to do so. trust theory. with a client-advisor trust relationship, it is when agreeing on how to work together and on implementing the recommendations that trust has its greatest impact. the more clients place trust in their advisors, the more they delegate power (monti et al., 2014) and the more they follow their advisors’ recommendations (barnett white, 2005). accordingly, the impact on the clients’ financial outcome can be positive or negative, depending on the level of trust placed in the advisor. but what may be the impact of trust on the other steps? one may say that the importance of trust is rather evident when gathering information, since an advisor’s analysis of a client’s situation depends upon the information available (hershey et al., 1990). in other words, clients may ask themselves if they should provide all relevant personal information to their advisor? fischer and gerhardt (2007) and foerster et al. (2017) argue that when the client provides private information, this allows the advisor to personalize recommendations. accordingly, the client should provide all available information to the advisor. on this point, omarzu (2000) mentions that clients must consider the risk associated with the disclosure of personal information, where such risk brings trust theory into play. the level of trust a client has in the advisor will determine the level of information sharing. here, trust in the advisor can directly and positively impact clients’ financial outcomes. figure 4 illustrates this essential element to reflect the impact of trust theory more accurately at the gathering information step. to our knowledge, this information is often overlooked in the data we analyze (one exception is alsemgeest (2022)). this element appears in figure 4 as a blue plus symbol surrounded by a circle. concept of knowledge. based on prior research, it is when gathering information, analyzing the situation, and monitoring progress that the concept of knowledge plays an important role and where it has a positive impact. equipped with proven knowledge, the financial advisor can: (1) better identify the information needed for analysis, (2) better assess the client-specific situation and possible options, and (3) better determine whether to restart the process based on the results obtained (hershey et al., 1990; hershey & walsh, 2000). figure 4 illustrates the direct and positive impact of the advisor’s level of knowledge. synergistic effect of the three theoretical lenses the interplay between agency theory, trust theory, and the concept of knowledge complexifies an advisor’s contribution to their clients’ financial outcomes due to a synergistic effect. depending on the theoretical perspectives taken, the advisor may have a positive impact or a negative impact. figure 5 proposes an integrative model of this synergistic effect based on four core questions associated with the six steps, namely: 1. does the client trust the advisor? 2. does the advisor prioritize the client’s interest? 3. does the advisor mobilize their financial knowledge? 4. does the client agree to disclose all relevant personal information? the integrative model resembles a decision tree, where the individual answers the four questions by “yes” or “no.” the model helps to understand what is inside the “black box;” that is, the services financial services review, 32(3) 43 provided by the advisor, a credence good. three elements underlie the integrative model: (1) the four questions provide a priority order, (2) the step in the process determines which theoretical lens will dominate, and (3) a theory applies or not depending on whether the individual performs the process alone or is assisted to various degrees by a financial advisor. figure 5 illustrates the advisor’s impact on outcomes by contrasting when the process is carried out alone (unassisted), with the outcome set to zero, or with an advisor, even if factors such as individual financial knowledge may affect the financial outcome. we aim to conceptualize an anticipated positive outcome when the financial advisor is involved, as well as situations where the advisor’s involvement may result in an outcome that may be equal to or worse than when an individual acts alone. we will now turn to an analysis of findings around the four core questions. figure 5. integrative model of the impact of financial advisors on their clients’ financial outcomes does the client trust the advisor? the answer to that question underlies agreeing on how to work together and implementing the recommendations. for these two steps, trust theory plays a primary role over the two other theoretical lenses since the client delegates to the advisor. when the client trusts the advisor, then the effects of agency theory and the concept of knowledge may come into play to influence the outcome of the process. at the implementation step, if the client does not accept the advisor’s recommendations, the negative impact related to an advisor’s selfinterest or the positive impact related to an advisor’s knowledge will not manifest pilote et al. 44 themselves in the financial outcome. stated otherwise, using a trust theory lens, a decision by a client to turn down an advisor’s recommendation results in a positive impact if the recommendation is harmful, and in a negative impact if the recommendation is beneficial (monti et al., 2014). however, if the client accepts the advisor’s recommendations, it places the individual in a vulnerable position where they cannot mitigate the effects, positive or negative, associated with the other steps of the process. as a reminder, trusting can be beneficial, but overtrusting can lead to abuse by the financial advisor (gennaioli et al., 2015; schwartz et al., 2011). does the advisor prioritize the client’s interests? this question puts forward agency theory, with emphasis on analyzing the situation, providing recommendations, and monitoring progress. here, the client trusts the advisor, opening the door to potential conflicts of interest, depending on whether the advisor prioritizes their interests or the clients’ interests. if an advisor maximizes their own interests, not only does it negatively impact their clients’ financial outcomes, but it also cancels the effects of the concept of knowledge. in this regard, blanchett (2019) and bergstresser et al. (2009) report that the advisor’s conflict of interest can negate the benefits associated with the advisor’s knowledge and expertise. alternatively, if advisors prioritize clients’ interests, then they can make use of their knowledge. when the client and advisor’s interests coincide, the advisor can have a strong positive influence on a client’s financial outcomes (finke, 2013). does the advisor mobilize their financial knowledge? this question comes into play when the client trusts the advisor and when the advisor prioritizes clients’ interests. the advisor’s knowledge will impact a client’s financial outcome when gathering information, analyzing the situation, and monitoring progress. when the advisor has specific knowledge helpful for dealing with a given situation, being able to leverage their skills and expertise to maximize the client’s interests, this may have a significant positive impact. alternatively, when the advisor does not have the necessary knowledge or is unable to leverage the required knowledge, this will not only negate the advisor’s influence but will also prevent trust theory from manifesting itself at the gathering information step. does the client agree to disclose all relevant personal information? the last question involves trust theory when gathering information. full disclosure of clients’ personal information to the advisor will impact the clients’ financial outcomes under the following conditions: (1) the client follows the advisor’s recommendations or has delegated the entire process to the advisor, (2) the advisor prioritizes the client’s best interests, and (3) the advisor is competent and mobilizes their knowledge. when these three conditions are present and clients disclose all relevant personal information, advisors find themselves in a position to personalize their advice and achieve “optimal” results (fischer & gerhardt, 2007; foerster et al., 2017). in addition, sharing of private information by the client allows the advisor to achieve a greater positive impact on their clients’ financial outcome. conclusion, limitations, and future research this paper presents the current state of knowledge regarding the impact of financial advisors on their clients’ financial outcomes. based on a review and analysis of prior research and using the six steps of the pfp process as a template for financial advising, we document in detail where the advisor may play a positive or negative role for their clients. we then propose an integrative model, linking three theoretical lenses, namely agency theory, trust theory, and the concept of knowledge, when the process is performed alone (unassisted) or with a financial advisor. the integrated model helps to answer our research question, which is, “what is the impact of the financial advisor on their clients’ financial outcome?” in summary, agency theory underlies clients’ financial outcomes via advisor actions when analyzing, recommending, and monitoring. an advisor’s conflict of interests will manifest itself directly and negatively on clients’ financial outcomes. trust theory provides a useful conceptual lens at various steps of the financial advising process. when gathering information, financial services review, 32(3) 45 the client’s trust with respect to the advisor directly and positively influences the outcome. however, when agreeing on how to work together and when implementing advisor recommendations, trust in the advisor may manifest indirectly, either positively or negatively. as for the concept of knowledge, it positively and directly influences the outcome of the process via its impact on the advisor’s gathering information, analyzing, and monitoring. these three theoretical lenses interact in a specific sequence, as shown in figure 5, providing a better understanding of the impact of the financial advisor. our main research contributions are as follows. first, we propose a theory-driven review of the financial advising literature, relying on three theoretical lenses to explain the financial advisor’s impact. second, we propose an integrative model of the financial advisor’s impact, linking the three theoretical lenses and explaining the divergent results found in prior research. this model allows researchers and practitioners to integrate the reality of the financial advising environment by linking the effects of the advisor’s level of involvement to the steps performed in practice, thus enhancing the model’s relevance for both academics and practitioners. last, based on four financial advising core questions developed in this paper, we seek to better capture the key concept of credence good, as provided by a financial advisor’s services, and provide explanations on how the advisor’s involvement will impact their clients’ financial outcome compared to when an individual decides to act in an unassisted manner. this study contributes to the growing interest of academics, financial advisors, professional associations, regulators, financial institutions, and society in examining the role played by financial advisors for their clients. a healthy and prosperous economy depends notably on financially literate individuals, but the latter are facing an increasingly complex financial marketplace. how can individuals navigate financial decisions in challenging times? should financial decisions rest with the individual alone or be a function of both individual decisions and advisory assistance? in the united states, canada, and abroad, the financial advising ecosystem should enable individuals to access, understand, and use financial products and services to their benefit. it should be easy for everyone to manage their revenues, expenses, debts, and savings. accordingly, the current study also adds to the financial literacy domain. this research has limitations. the role played by financial advisors is very important to investigate but difficult to research, given the complexity of financial advisors’ service as a credence good and its impact on their clients’ financial outcome. even if personal finance is critical for individuals and in the economy, surprisingly, research in the area is rather limited and scattered. for instance, quality academic journals specialized in pfp are very rare. that said, our review of prior research, and our analysis through theoretical frameworks, has still permitted us to identify relevant articles to develop an integrative model. we agree that our model provides a limited picture of a complex construct and requires further investigations. for future research, it would be relevant to validate empirically the integrative model developed. case studies and interviews should be conducted with stakeholders, such as advisors, banks, and clients, to know more about the financial advising process and, more importantly, the role played by financial advisors. professional associations such as fp canada or the us cfp board should be involved. collecting data using a diversity of methods among various stakeholders may only help to better capture the phenomena. financial professional designations also seem to modulate the impact of the advisor on their clients’ outcome through various training, certifications, and diplomas. how designations impact clients’ perceptions in the process requires investigation. future research may also look at how advisors use technology, such as software, to access specialized knowledge. references agnew, j. r., bateman, h., eckert, c., iskhakov, f., louviere, j., & thorp, s. (2018). first impressions matter: an experimental investigation of online pilote et al. 46 financial advice. management science, 64(1), 288-307. alsemgeest, l. (2022). the truth will set you free: the perception of south african financial advisors regarding clients' truthfulness when sharing information. journal of financial services marketing, 27(3), 221-231. anagol, s., cole, s., & sarkar, s. (2017). understanding the advice of commissions-motivated agents: evidence from the indian life insurance market. review of economics and statistics, 99(1), 1-15. anderson, j. r. (1982). acquisition of cognitive skill. psychological review, 89(4), 369. anderson, j. r. (1983). the architecture of cognition. harvard university press. angelova, v., & regner, t. (2013). do voluntary payments to advisors improve the quality of financial advice? an experimental deception game. journal of economic behavior & organization, 93, 205-218. angelova, v., & regner, t. (2018). can a bonus overcome moral hazard? experimental evidence from markets for expert services. journal of economic behavior & organization, 154, 362-378. arman, j., & shackman, j. (2012). the impact of financial planning designations on financial planner income. the service industries journal, 32(8), 1393-1409. azamian, n., beck, k., chiu, h.-h., & timmerman, i. (2022). do as i tell you, not as i do: financial advisors and personal financial decision-making. financial services review, 30(1), 57-68. bae, s. c., & sandager, j. p. (1997). what consumers look for in financial planners. journal of financial counseling and planning, 8(2), 9. balasubramnian, b., & brisker, e. r. (2016). financial adviser users and financial literacy. financial services review, 25(2), 127. barnett white, t. (2005). consumer trust and advice acceptance: the moderating roles of benevolence, expertise, and negative emotions. journal of consumer psychology, 15(2), 141-148. barrick, j. a., & spilker, b. c. (2003). the relations between knowledge, search strategy, and performance in unaided and aided information search. organizational behavior and human decision processes, 90(1), 1-18. bergstresser, d., chalmers, j. m. r., & tufano, p. (2009). assessing the costs and benefits of brokers in the mutual fund industry. the review of financial studies, 22(10), 4129-4156. beyer, m., de meza, d., & reyniers, d. (2013). do financial advisor commissions distort client choice? economics letters, 119(2), 117-119. bhattacharya, u., hackethal, a., kaesler, s., loos, b., & meyer, s. (2012). is unbiased financial advice to retail investors sufficient? answers from a large field study. the review of financial studies, 25(4), 975-1032. bigel, k. s. (2000). the ethical orientation of financial planners who are engaged in investment activities: a comparison of united states practitioners based on professionalization and compensation sources. journal of business ethics, 28(4), 323-337. blanchett, d. m. (2019). financially sound households use financial planners, not transactional advisers. journal of financial planning, 32(4), 30-40. bluethgen, r., gintschel, a., hackethal, a., & mueller, a. (2008). financial advice and individual investors' portfolios. available at ssrn 968197. bluethgen, r., meyer, s., & hackethal, a. (2008). high-quality financial advice wanted!. available at ssrn 1102445. bodnaruk, a., & simonov, a. (2015). do financial experts make better investment decisions?. journal of financial intermediation, 24(4), 514-536. boisclair, d., lusardi, a., & michaud, p.-c. (2017). financial literacy and retirement planning in canada. journal of pension economics & finance, 16(3), 277-296. bolton, p., freixas, x., & shapiro, j. (2007). conflicts of interest, information provision, and competition in the financial services review, 32(3) 47 financial services industry. journal of financial economics, 85(2), 297-330. bruhn, a., & miller, m. (2014). lessons about best interests duty. australasian accounting business & finance journal, 8(4), 23-43. burke, j., hung, a., clift, j., garber, s., & yoong, j. (2015). impacts of conflicts of interest in the financial services industry. working paper series no. wr1076, rand labor and population, santa monica, ca. burke, j., & hung, a. a. (2021). trust and financial advice. journal of pension economics & finance, 20(1), 9-26. calcagno, r., giofré, m., & urzì-brancati, m. c. (2017). to trust is good, but to control is better: how investors discipline financial advisors’ activity. journal of economic behavior & organization, 140, 287-316. certified financial planner board of standards. (2017). consumer guide to financial planning. available at: https://www.cfp.net/docs/defaultsource/consumeroutreach/cfpboard_consumer_guide_to_ financial_planning.pdf?sfvrsn=14. accessed on september 20, 2023. certified financial planner board of standards. (2018). code of ethics and standards of conduct. available at: https://www.cfp.net/-/media/files/cfpboard/standards-and-ethics/cfp-codeandstandards.pdf?la=en&hash=afc0087b5 3ede26b2a478bd84df7dda3. accessed on september 20, 2023. certified financial planner board of standards. (2023). 2023 cfp board consumer sentiment survey – cost of living. available at: https://www.cfp.net/knowledge/reportsand-statistics/consumer-surveys/2023cfp-board-consumer-sentiment-surveycost-of-living. accessed on september 20, 2023. chalmers, j., & reuter, j. (2020). is conflicted investment advice better than no advice? journal of financial economics, 138(2), 366-387. chatterjee, s., & lu, f. (2023). surviving in financial advice deserts: limited access to financial advice and retirement planning behavior. the international journal of bank marketing, 41(1), 70106. https://doi.org/https://doi.org/10.1108/ij bm-01-2022-0022 chen, p., & richardson, m. (2018). conflicted financial advice: disclosure revisited. applied economics letters, 25(12), 826-829. chen, p., & richardson, m. (2019). conflict of interest, disclosure and vertical relationships: an experimental analysis. economic papers: a journal of applied economics and policy, 38(3), 167-181.cheng, y., & kalenkoski, c. m. (2018). lost in fees: an analysis of financial planning compensation. the journal of wealth management, 20(4), 46-54. christoffersen, s. e., evans, r., & musto, d. k. (2013). what do consumers’ fund flows maximize? evidence from their brokers’ incentives. the journal of finance, 68(1), 201-235. cici, g., kempf, a., & sorhage, c. (2017). do financial advisors provide tangible benefits for investors? evidence from tax-motivated mutual fund flows. review of finance, 21(2), 637-665. cloyd, c. b. (1995). prior knowledge, information search behaviors, and performance in tax research tasks. the journal of the american taxation association, 17, 82. cloyd, c. b. (1997). performance in tax research tasks: the joint effects of knowledge and accountability. accounting review, 111-131. cummings, b. f., finke, m. s., & james, r. n. (2013). bounded rationality strikes again: the impact of cognitive ability and financial planners on roth ira adoption and ownership. available at ssrn 1968984. cupach, w. r., & carson, j. m. (2002). the influence of compensation on product recommendations made by insurance https://www.cfp.net/docs/default-source/consumer-outreach/cfpboard_consumer_guide_to_financial_planning.pdf?sfvrsn=14 https://www.cfp.net/docs/default-source/consumer-outreach/cfpboard_consumer_guide_to_financial_planning.pdf?sfvrsn=14 https://www.cfp.net/docs/default-source/consumer-outreach/cfpboard_consumer_guide_to_financial_planning.pdf?sfvrsn=14 https://www.cfp.net/docs/default-source/consumer-outreach/cfpboard_consumer_guide_to_financial_planning.pdf?sfvrsn=14 https://www.cfp.net/-/media/files/cfp-board/standards-and-ethics/cfp-code-and-standards.pdf?la=en&hash=afc0087b53ede26b2a478bd84df7dda3 https://www.cfp.net/-/media/files/cfp-board/standards-and-ethics/cfp-code-and-standards.pdf?la=en&hash=afc0087b53ede26b2a478bd84df7dda3 https://www.cfp.net/-/media/files/cfp-board/standards-and-ethics/cfp-code-and-standards.pdf?la=en&hash=afc0087b53ede26b2a478bd84df7dda3 https://www.cfp.net/-/media/files/cfp-board/standards-and-ethics/cfp-code-and-standards.pdf?la=en&hash=afc0087b53ede26b2a478bd84df7dda3 https://www.cfp.net/-/media/files/cfp-board/standards-and-ethics/cfp-code-and-standards.pdf?la=en&hash=afc0087b53ede26b2a478bd84df7dda3 https://www.cfp.net/knowledge/reports-and-statistics/consumer-surveys/2023-cfp-board-consumer-sentiment-survey-cost-of-living https://www.cfp.net/knowledge/reports-and-statistics/consumer-surveys/2023-cfp-board-consumer-sentiment-survey-cost-of-living https://www.cfp.net/knowledge/reports-and-statistics/consumer-surveys/2023-cfp-board-consumer-sentiment-survey-cost-of-living https://www.cfp.net/knowledge/reports-and-statistics/consumer-surveys/2023-cfp-board-consumer-sentiment-survey-cost-of-living pilote et al. 48 agents. journal of business ethics, 40(2), 167-176. danilov, a., biemann, t., kring, t., & sliwka, d. (2013). the dark side of team incentives: experimental evidence on advice quality from financial service professionals. journal of economic behavior & organization, 93, 266-272. darby, m. r., & karni, e. (1973). free competition and the optimal amount of fraud. the journal of law and economics, 16(1), 67-88. dean, d. h. (2017). the benefit of a trustworthy face to a financial services provider. journal of services marketing, 31(7), 771-783. deng, l., & liu, z. (2017). one-period pricing strategy of ‘money doctors’ under cumulative prospect theory. portuguese economic journal, 16(2), 113-144. direr a. a., & visser b. m. (2013). portfolio choice and financial advice. finance, 34(2), 35-64. dulleck, u., & kerschbamer, r. (2006). on doctors, mechanics, and computer specialists: the economics of credence goods. journal of economic literature, 44(1), 5-42. dulleck, u., kerschbamer, r., & sutter, m. (2011). the economics of credence goods: an experiment on the role of liability, verifiability, reputation, and competition. american economic review, 101(2), 526-555. dvorak, t. (2015). do 401k plan advisors take their own advice? journal of pension economics & finance, 14(1), 55-75. engelmann, j. b., capra, c. m., noussair, c., & berns, g. s. (2009). expert financial advice neurobiologically “offloads” financial decision-making under risk. plos one, 4(3). eriksson, k., & hermansson, c. (2019). how relationship attributes affect bank customers’ saving. international journal of bank marketing. 37(1), 156-170. fan, l. (2021). information search, financial advice use, and consumer financial behavior. journal of financial counseling and planning. 32(1), 21-34. financial consumer agency of canada. (2019). canadians and their money: key findings from the 2019 canadian financial capability survey. available at: https://www.canada.ca/content/dam/fcac -acfc/documents/programs/researchsurveys-studies-reports/canadianfinancial-capability-survey-2019.pdf. accessed on september 20, 2023. financial planning standards council and institut québécois de planification financière. (2020). canadian financial planning: definitions, standards & competencies. available at: https://www.fpcanada.ca/docs/defaultsource/standards/definitions_standards_ 2020_edition_eng_web.pdf. accessed on september 20, 2023. finke, m. (2013). financial advice: does it make a difference? the market for retirement financial advice, oxford academic, oxford. 229-248. finke, m., huston, s., & waller, w. (2009). do contracts influence comprehensive financial advice? financial services review, 18(2), 177-193. finke, m. s., huston, s. j., & winchester, d. d. (2011). financial advice: who pays. journal of financial counseling and planning, 22(1), 18. fischer, r., & gerhardt, r. (2007). investment mistakes of individual investors and the impact of financial advice. paper presented at the 20th australasian finance & banking conference (afbc). foerster, s., linnainmaa, j. t., melzer, b. t., & previtero, a. (2017). retail financial advice: does one size fit all? the journal of finance, 72(4), 1441-1482. fong, y.-f. (2005). when do experts cheat and whom do they target? rand journal of economics, 113-130. fp canada. (2019). cfp® professional competency profile. available at: https://www.fpcanada.ca/docs/defaultsource/standards/cfp-professionalcompetency-profile.pdf. accessed on september 20, 2023. https://www.canada.ca/content/dam/fcac-acfc/documents/programs/research-surveys-studies-reports/canadian-financial-capability-survey-2019.pdf https://www.canada.ca/content/dam/fcac-acfc/documents/programs/research-surveys-studies-reports/canadian-financial-capability-survey-2019.pdf https://www.canada.ca/content/dam/fcac-acfc/documents/programs/research-surveys-studies-reports/canadian-financial-capability-survey-2019.pdf https://www.canada.ca/content/dam/fcac-acfc/documents/programs/research-surveys-studies-reports/canadian-financial-capability-survey-2019.pdf https://www.fpcanada.ca/docs/default-source/standards/definitions_standards_2020_edition_eng_web.pdf https://www.fpcanada.ca/docs/default-source/standards/definitions_standards_2020_edition_eng_web.pdf https://www.fpcanada.ca/docs/default-source/standards/definitions_standards_2020_edition_eng_web.pdf https://www.fpcanada.ca/docs/default-source/standards/cfp-professional-competency-profile.pdf https://www.fpcanada.ca/docs/default-source/standards/cfp-professional-competency-profile.pdf https://www.fpcanada.ca/docs/default-source/standards/cfp-professional-competency-profile.pdf financial services review, 32(3) 49 fp canada. (2022). standards of professional responsibility. available at: https://fpcanada.ca/docs/defaultsource/standards/standards-ofprofessional-responsibility.pdf. accessed on september 20, 2023. gennaioli, n., shleifer, a., & vishny, r. (2015). money doctors. the journal of finance, 70(1), 91-114. georgarakos, d., & inderst, r. (2014). financial advice and stock market participation. working paper no. 1296, european central bank, frankfurt, germany. golec, j. h. (1992). empirical tests of a principal-agent model of the investorinvestment advisor relationship. journal of financial and quantitative analysis, 27(1), 81-95. grable, j. e., & chatterjee, s. (2014). reducing wealth volatility: the value of financial advice as measured by zeta. journal of financial planning, 27(8), 45-51. guillemette, m. a., & jurgenson, j. b. (2017). the impact of financial advice certification on investment choices. journal of financial counseling and planning, 28(1), 129-139. gupta, p., & cohen, n. j. (2002). theoretical and computational analysis of skill learning, repetition priming, and procedural memory. psychological review, 109(2), 401. gurun, u. g., stoffman, n., & yonker, s. e. (2018). trust busting: the effect of fraud on investor behavior. the review of financial studies, 31(4), 1341-1376. hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking & finance, 36(2), 509-524. hackethal, a., inderst, r., & meyer, s. (2010). trading on advice. cepr discussion paper no. dp8091, available at ssrn: https://ssrn.com/abstract=171486 3 hanna, s. d., & lindamood, s. (2010). quantifying the economic benefits of personal financial planning. financial services review, 19(2), 1-21. hershey, d. a., & walsh, d. a. (2000). knowledge versus experience in financial problem solving performance. current psychology, 19(4), 261. hershey, d. a., walsh, d. a., read, s. j., & chulef, a. s. (1990). the effects of expertise on financial problem solving: evidence for goal-directed, problemsolving scripts. organizational behavior and human decision processes, 46(1), 77-101. hoechle, d., ruenzi, s., schaub, n., & schmid, m. (2017). the impact of financial advice on trade performance and behavioral biases. review of finance, 21(2), 871-910. horn, l., meyer, s., & hackethal, a. (2009). smart investing and the role of financial advice-evidence from a natural experiment using data around a tax law change. available at ssrn 1343623. hudson, c. r., & palmer, l. (2014). lowincome employees: the relationship between information from formal advisors and financial behaviors. financial services review, 23(1). hung, a., & yoong, j. (2013). asking for help: survey and experimental evidence on financial advice and behavior change. the market for retirement financial advice, 182, 182-212. inderst, r., & ottaviani, m. (2012a). competition through commissions and kickbacks. american economic review, 102(2), 780-809. inderst, r., & ottaviani, m. (2012b). how (not) to pay for advice: a framework for consumer financial protection. journal of financial economics, 105(2), 393411. james iii, r. n. (2013). brain activity suggests planning designation helps calm investors. journal of financial planning, 26(2), 52. jarratt, d., bossomaier, t., & thompson, j. (2007). trust as an emergent phenomenon in wealth management relationships. global business and economics review, 9(4), 335-352. jensen, m. c., & meckling, w. h. (1976). theory of the firm: managerial behavior, agency costs and ownership https://fpcanada.ca/docs/default-source/standards/standards-of-professional-responsibility.pdf https://fpcanada.ca/docs/default-source/standards/standards-of-professional-responsibility.pdf https://fpcanada.ca/docs/default-source/standards/standards-of-professional-responsibility.pdf https://ssrn.com/abstract=1714863 https://ssrn.com/abstract=1714863 pilote et al. 50 structure. journal of financial economics, 3(4), 305-360. johnson, d., & grayson, k. (2005). cognitive and affective trust in service relationships. journal of business research, 58(4), 500-507. kim, k. t., pak, t. y., shin, s. h., & hanna, s. d. (2018). the relationship between financial planner use and holding a retirement saving goal: a propensity score matching analysis. financial planning review, 1(1-2), e1008. kingston, g., & weng, h. (2014). agency theory and financial planning practice. australian economic review, 47(3), 290-303. kramer, m. m. (2012). financial advice and individual investor portfolio performance. financial management, 41(2), 395-428. krausz, m., & paroush, j. (2002). financial advising in the presence of conflict of interests. journal of economics and business, 54(1), 55-71. lach, p. a., flynn, l. r., & kelly, g. w. (2019). brokers or investment advisers? the us public perception. financial analysts journal, 75(3), 125-131. lachance, m.-e., & tang, n. (2012). financial advice and trust. financial services review, 21(3). lai, k. p. (2016). financial advisors, financial ecologies and the variegated financialisation of everyday investors. transactions of the institute of british geographers, 41(1), 27-40. lei, s. (2019). financial advice and individual investors’ investment decisions. applied economics letters, 26(13), 1129-1132. linnainmaa, j. t., melzer, b., previtero, a., & foerster, s. (2018). financial advisors and risk-taking. unpublished manuscript. https://www.aleprevitero.com/wpcontent/uploads/2019/10/029_previtero_ wp_financialadvicerisktaking.pdf liu, f., yilmazer, t., loibl, c., & montalto, c. (2019). professional financial advice, self‐control and saving behavior. international journal of consumer studies, 43(1), 23-34. liu, w. l. (2005). motivating and compensating investment advisors. the journal of business, 78(6), 2317-2350. lusardi, a., & mitchell, o. s. (2011). financial literacy and retirement planning in the united states. journal of pension economics & finance, 10(4), 509-525. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52(1), 5-44. mackinger, b., jonas, e., & mühlberger, c. (2017). when advisors’ true intentions are in question. how do bank customers cope with uncertainty in financial consultancies?. frontiers in psychology, 8, 219401. marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32(4), 625-643. martin, t. k. j., & finke, m. (2014). a comparison of retirement strategies and financial planner value. journal of financial planning, 27(11), 46. mayer, r. c., davis, j. h., & schoorman, f. d. (1995). integrative model of organizational trust. academy of management review, 20(3), 709-734. mietzner, m., & molterer, m. (2018). you might not get what you need: the discrepancy between financial advice and commissions in germany. economics letters, 162, 167-170. monti, m., pelligra, v., martignon, l., & berg, n. (2014). retail investors and financial advisors: new evidence on trust and advice taking heuristics. journal of business research, 67(8), 1749-1757. montmarquette, c., & viennot-briot, n. (2015). the value of financial advice. annals of economics and finance, 16(1), 69-94. montmarquette, c., & viennot-briot, n. (2019). the gamma factors and the value of financial advice. annals of economics and finance, 20(1), 387-411. moreland, k. a. (2018). seeking financial advice and other desirable financial behaviors. journal of financial counseling and planning, 29(2), 198207. financial services review, 32(3) 51 mountain, t. p., kim, n., serido, j., & shim, s. (2021). does type of financial learning matter for young adults’ objective financial knowledge and financial behaviors? a longitudinal and mediation analysis. journal of family and economic issues, 42(1), 113-132. mullainathan, s., noeth, m., & schoar, a. (2012). the market for financial advice: an audit study. working paper no. w17929, national bureau of economic research, cambridge, ma. nofsinger, j. r., & varma, a. (2007). how analytical is your financial advisor? financial services review, 16(4), 245. oehler, a., & kohlert, d. (2009). financial advice giving and taking—where are the market’s self-healing powers and a functioning legal framework when we need them? journal of consumer policy, 32, 91-116. omarzu, j. (2000). a disclosure decision model: determining how and when individuals will self-disclose. personality and social psychology review, 4(2), 174-185. pan, x., wu, w., & zhang, x. (2020). is financial advice a cure-all or the icing on the cake for financial literacy? evidence from financial market participation in china. international review of financial analysis, 69, 101473. paré, g., trudel, m.-c., jaana, m., & kitsiou, s. (2015). synthesizing information systems knowledge: a typology of literature reviews. information & management, 52(2), 183-199. paris, s. g., lipson, m. y., & wixson, k. k. (1983). becoming a strategic reader. contemporary educational psychology, 8(3), 293-316. park, e., & yao, r. (2016). financial risk attitude and behavior: do planners help increase consistency? journal of family and economic issues, 37, 624638.pauls, t., stolper, o. a., & walter, a. (2016). broad-scope trust and financial advice. working paper, goethe university, frankfurt, germany. perkins, d. n., & salomon, g. (1989). are cognitive skills context-bound? educational researcher, 18(1), 16-25. ross, s. a. (1973). the economic theory of agency: the principal's problem. the american economic review, 63(2), 134139. schunk, d. h. (2012). learning theories: an educational perspective sixth edition. pearson. london, uk. schwartz, j., luce, m. f., & ariely, d. (2011). are consumers too trusting? the effects of relationships with expert advisers. journal of marketing research, 48(spl), s163-s174. shapira, z., & venezia, i. (2001). patterns of behavior of professionally managed and independent investors. journal of banking & finance, 25(8), 1573-1587. smith, h., finke, m., & huston, s. (2012). the influence of financial sophistication and financial planners on roth ira ownership. journal of financial service professionals, 66(6), 69. snyder, h. (2019). literature review as a research methodology: an overview and guidelines. journal of business research, 104, 333-339. spilker, b. c. (1995). the effects of time pressure and knowledge on key word selection behavior in tax research. accounting review, 49-70. stolper, o. (2018). it takes two to tango: households’ response to financial advice and the role of financial literacy. journal of banking & finance, 92, 295-310. stoughton, n. m., wu, y., & zechner, j. (2011). intermediated investment management. the journal of finance, 66(3), 947-980. sweeney, j. c., plewa, c., & zurbruegg, r. (2018). examining positive and negative value-in-use in a complex service setting. european journal of marketing. 52(5-6), 1084-1106. tang, n., & hu, x. (2019). financial literacy and the use of financial advice—a nonmonotonic relationship. journal of trends and issues, 2(157), 12-32. tardif, j. (1992). pour un enseignement stratégique: l'apport de la psychologie cognitive. éditions logiques. montréal, qc. templier, m., & paré, g. (2018). transparency in literature reviews: an assessment of pilote et al. 52 reporting practices across review types and genres in top is journals. european journal of information systems, 27(5), 503-550. tharp, d. (2019). consumer perceptions of financial advisory titles and implications for title regulation. working paper, mercatus center at george mason university, arlington, va. torraco, r. j. (2005). writing integrative literature reviews: guidelines and examples. human resource development review, 4(3), 356-367. van dijk, m., bijlsma, m., & pomp, m. (2008). the price of free advice. applied economics, 40(14), 1889-1903. von gaudecker, h. m. v. (2015). how does household portfolio diversification vary with financial literacy and financial advice?. the journal of finance, 70(2), 489-507. webster, j., & watson, r. t. (2002). analyzing the past to prepare for the future: writing a literature review. mis quarterly, xiii-xxiii. west, j. (2012). financial advisor participation rates and low net worth investors. journal of financial services marketing, 17(1), 50-66. winchester, d., & huston, s. (2017). trust reduces costs associated with consumer-financial planner relationship. journal of financial service professionals, 71(4), 80-91. winchester, d. d., & huston, s. j. (2014). does a relationship with a financial service professional overcome a client’s sense of not being in control of achieving their goals? financial services review, 23(1), 1-23. winchester, d. d., & huston, s. j. (2015). all financial advice for the middle class is not equal. journal of consumer policy, 38(3), 247-264. winchester, d. d., huston, s. j., & finke, m. s. (2011). investor prudence and the role of financial advice. journal of financial service professionals, 65(4), 43. wright, p., mukherji, a., & kroll, m. j. (2001). a reexamination of agency theory assumptions: extensions and extrapolations. the journal of socioeconomics, 30(5), 413-429. zhang, a. c. (2014). financial advice and asset allocation of individual investors. pacific accounting review, 26(3), 226247. financial services review, 32(3) 53 appendix a table of articles authors type of financial advisor methodology measurement of agency, trust, or knowledge outcome measurement findings agency theory anagol et al. (2017) insurance agents field experiment (audit study) choice of self-serving recommendations different life insurance options advisors are more likely to recommend the option that maximizes their commission angelova & regner (2013) participants acting as financial advisors lab experiment choice of self-serving recommendations different investment options advisors are more likely to recommend the option that maximizes their commission lump sums lead the advisor to recommend more the favorable option for the client angelova & regner (2018) participants acting as financial advisors lab experiment choice of self-serving recommendations different investment options advisors are more likely to recommend the option that maximizes their commission competition and reputation can lead the advisor to recommend more the favorable option for the client beyer et al. (2013) participants acting as insurance advisors lab experiment choice of self-serving recommendations different extended coverage options advisors are more likely to recommend the option that maximizes their commission bigel (2000) certified financial planners (cfp) and non-certified financial advisors mail survey no explicit measure, but a measure of ethics comparison of the results on the ethical questions holding a professional title is linked to a higher level of ethics the form of compensation does not affect the level of ethics of planners pilote et al. 54 bluethgen et al. (2008) independent financial advisors, title not specified survey choice of self-serving assets and the choice of actively managing portfolios assets allocation compensation structure and advisor’s rationality are the two most important predictors of the quality of advice provided to clients bolton et al. (2007) not applicable analytical model choice of self-serving recommendations recommendation choice the more profit the financial institution can make by directing clients to a particular type of investment, the more tempted it will be to do so, even if this affects its credibility. burke et al. (2015) financial advisors, title not specified literature review not applicable not applicable financial advisors maximize their own interests at the expense of their clients' interests chalmers & reuter (2020) brokers, title not specified archival data with field experiment and online survey the conflict of interest of brokers is assumed annual return and volatility of assets if a default option for a retirement plan exists and it is well designed, individuals are better off with it than with financial advice with a conflict of interest advisors recommend assets with high commissions chen & richardson (2018) participants acting as financial advisors lab experiment choice of self-serving recommendations client’s monetary results minus the amount paid advisors are more likely to recommend the option that maximizes their commission chen & richardson (2019) participants acting as financial advisors lab experiment choice of self-serving recommendations client’s monetary results advisors are more likely to recommend the option that maximizes their commission christoffersen et al. (2013) mutual funds broker archival data choice of mutual funds with higher incentive the positive or negative flow to a mutual fund commissions from mutual funds affect broker recommendations cupach & carson (2002) insurance agents mail survey choice of self-serving recommendations amount and type of insurance recommended the form of compensation for insurance agents does not financial services review, 32(3) 55 affect their product recommendations danilov et al. (2013) bank employees lab experiment choice of self-serving recommendations different financial product options advisors are more likely to recommend the option that maximizes their commission group incentives amplified the effect dvorak (2015) financial advisory firms archival data choice of mutual funds with higher fees the variation in fees between the funds held in the retirement plan for their own employees and their clients the retirement plans for their clients and their employees are very similar, except that the clients have higher fees finke (2013) financial advisor, title not specified literature review not applicable not applicable the positive contribution of the financial advisor could be negated by the presence of a conflict of interest between the advisor and the client foerster et al. (2017) financial advisor, title not specified archival data no explicit measure investors' portfolio of assets clients' portfolio is an imperfect copy of their advisor's clients pay 2.5% of the value of the portfolio annually to receive the same advice as all of the advisor's other clients golec (1992) financial advisor, title not specified archival data incentives provided by the mutual funds to the advisor performance of the mutual funds mutual funds with base fees and advisor incentives attract more investment and have better financial results hackethal et al. (2012) independent financial advisor and bank-related financial advisor, title not specified archival data incentive to do more trade performance of the investment portfolio clients with either type of advisor perform worse than those investing alone the portfolios of advised clients have more trades than those of clients investing alone pilote et al. 56 hackethal et al. (2010) financial advisor, title not specified archival data and phone survey advisor’s willingness to initiate contact with clients transaction volume and type of product acquired based on financial advice, clients trade more and purchase more products, which benefits the bank hoechle et al. (2017) financial advisor, title not specified archival data no explicit measure comparison of the performance of advisory and nonadvisory trades with benchmarks trades following the advisor's recommendations perform worse than independent trades and benchmarks the effect is worse when it is the advisor who initiates the contact for the client to trade inderst & ottaviani (2012b) financial advisor, title not specified analytical model choice of self-serving recommendations advice on which option to choose commission-based compensation promotes the maximization of the advisor’s interests with naïve clients, but not with wary clients jarratt et al. (2007) wealth management advisor analytical model choice of self-serving recommendations investment choices advisors have a vested interest in finding a balance between maximizing their own shortterm interest and retaining the trust of clients so that they remain their clients kingston & weng (2014) financial advisor, title not specified analytical model choice of self-serving assets allocation assets allocation advisors favorize their own interest by selecting growthmaximizing assets when they receive a percentage of assets under management krausz & paroush (2002) financial advisor, title not specified analytical model choice of self-serving recommendations advice on which option to choose advisor pushes the risky option, which is associated with a better return for himself, even if the client is risk-averse lai (2016) insurance agents, wealth managers, and independent financial advisors mixed methodology presence of incentives not applicable recommendations made by financial advisors are strongly affected by commissions, bonuses, and sales quotas financial services review, 32(3) 57 liu (2005) financial advisor, title not specified analytical model choice of self-serving recommendations and efforts clients’ return on investment simply sharing the return on the investment fails to maximize the advisor's effort mietzner & molterer (2018) bankers, title not specified archival data choice of self-serving recommendations investment in a specific type of asset commissions create a gap between the recommendations made by the advisor and those that would optimize the client's situation mullainathan et al. (2012) financial advisor, title not specified field experiment (audit study) choice of self-serving modifications to the initial portfolio variation on the initial portfolio based on advice advisors encourage the acquisition of funds with high fees and the multiplication of transactions advisors are willing to accommodate client wishes when they coincide with their own interests, but try to reverse the situation when the client's vision does not favor their own interests oehler & kohlert (2009) financial advisor working in a bank, title not specified field experiment (audit study) choice of self-serving recommendations variation of advisors’ recommendations advice provided by the advisor is of higher quality when the client is more financially literate the information provided by the advisor to the client is generally of low quality stoughton et al. (2011) financial advisor, title not specified analytical model presence of kickbacks clients’ wealth kickbacks offered by portfolio managers to financial advisors negatively affect clients’ wealth van dijk et al. (2008) insurance brokers archival data and online survey choice of self-serving recommendations the return on investment of the life insurance policy insurance brokers do not generally recommend the best product to their clients trust theory pilote et al. 58 barnett white (2005) participants acting as financial advisors lab experiment a 7-point likert scale question the decision to follow the recommendation clients’ decision to follow the recommendation is based on the advisor’s ability in risky and non-emotional situations and on the advisor’s benevolence in risky and emotional situations bhattacharya et al. (2012) softwaregenerated advice archival data with field experiment no explicit measure the decision to implement the recommendation very few clients (about 5%) implement unsolicited recommendations the presence of unbiased advice is not enough to ensure that clients follow it burke & hung (2021) financial advisors, different denominations archival data a combination of multiple questions using likert scale the underdiversification of assets and the holding of risky assets individuals who trust are more likely to get financial advice individuals who trust hold more publicly traded stocks unsolicited financial advice has no real impact on clients’ behavior calcagno et al. (2017) financial advisor, title not specified archival data a 5-point likert scale question the decision to delegate decision-making to the advisor clients who trust more their advisor are more likely to completely delegate financial decision-making deng & liu (2017) financial advisor in a broad sense analytical model represented by a variable in the model the portfolio at the end of the period the more trust the client has in the advisor, the more inclined the client is to invest in the risky asset engelmann et al. (2009) the expert providing advice is presented as a professor of economics lab experiment neural activity visible by fmri during the decision-making the decision to follow the expert’s recommendation access to expert advice leads participants to rely on the expert's opinion and to offload the mental burden of decisionmaking eriksson & hermansson (2019) financial advisor working in a bank, title not specified online survey a combination of two questions financial assets purchased from the bank and monthly flow the length of the relationship with the advisor, the environment and the trust financial services review, 32(3) 59 invested in mutual funds placed in the advisor positively affect the total assets held by the client with the bank as well as the amounts invested monthly in mutual funds gennaioli et al. (2015) financial advisor in a broad sense analytical model represented by a variable in the model the portfolio at the end of the period the more trust the client has in the advisor, the more likely they are to completely delegate the process georgarakos & inderst (2014) financial advisor, title not specified archival data and analytical model a dichotomous question the decision to invest in the stock market the more clients trust their advisor, the more likely they are to follow their advisor's recommendations for participating in the stock market gurun et al. (2018) registered investment advisor archival data comparison of the decision to move their assets to a checking account the decision to continue the relationship with the advisor clients who have been affected by financial fraud stop doing business with the advisor who suggested the investment johnson & grayson (2005) financial advisor, title not specified mail survey cognitive and affective trust are both based on five questions the value of sales per account and the number of different products owns by the client clients who trust more their advisor are more likely to want to return to them in the future advisors who have more cognitive trust from their clients have better sales effectiveness lachance & tang (2012) five types of financial advisor archival data a 7-point likert scale question soliciting the services of a financial advisor individuals who trust financial advisors more seek their services in all five areas linnainmaa et al. (2018) financial advisor, title not specified archival data the length of clientadvisor relationship participation in the stock market, i.e., holding risky assets the more clients trust their advisor, the more likely they are to follow their advisor's recommendations to participate in the stock market clients who seek out an advisor participate more in the pilote et al. 60 stock market and take more risk (30% more) in their investments mackinger et al. (2017) financial advisor, title not specified survey a combination of five questions willingness to cooperate with the advisor trust in the advisor affects clients’ intention to cooperate under uncertainty monti et al. (2014) financial advisor, title not specified interviews and survey a combination of four questions clients’ indication on whether they would completely delegate clients report that they delegate primarily because they trust their advisor and lack financial knowledge (77% say they delegate and trust their advisor) pauls et al. (2016) financial advisor, title not specified archival data a dichotomous question the decision to follow the advisors’ recommendation the more trust the client has in the advisor, the more likely they are to follow the advisor's recommendations stolper (2018) softwaregenerated advice archival data no explicit measure the decision to implement the recommendation very few clients (8.8%) implement the recommendations within a year of receiving the advice the presence of unbiased advice is not enough to ensure that clients follow it winchester & huston (2017) financial services professional, title not specified survey clients’ willingness to recommend the advisor to someone else the value of financial advice after cost trust in the advisor increases the value of consulting an advisor by reducing the costs associated with the client's control of the advisor. when the client has more trust, they control less concept of knowledge azamian et al. (2022) financial advisor, 63% of the respondents are online survey no explicit measure, but advisors are considered as professional in debt, retirement plan, investment, estate plan, and cash flow the financial advisor group is better prepared for retirement, has less debt, has more liquidity, is better insured and financial services review, 32(3) 61 certified financial planners (cfp) comparison with the general population is more likely to have an estate plan barrick & spilker (2003) tax specialists lab experiment a combination of eight questions on the topic the information collected by the participants are compared with those of an experts’ panel problem-specific knowledge positively and directly (indirectly, via the search strategy) affects information collection performance in the presence (absence) of collecting assistance bergstresser et al. (2009) mutual funds broker archival data no explicit measure, but brokers are considered as professional in comparison to their clients the performance of mutual funds sold or acquired through a broker or directly by the client mutual funds purchased through a broker perform worse than those purchased directly by the client results could be explained by the conflict of interest between the broker and the client bluethgen et al. (2008) financial advisor, title not specified archival data no explicit measure, but advisors are considered as professional in comparison to their clients diversification of assets in the clients’ portfolio advisory-assisted clients have better portfolio diversification clients who receive advice trade more and therefore have higher transaction costs bodnaruk & simonov (2015) mutual fund managers archival data no explicit measure, but managers are considered as professional in comparison to their clients comparison of the results of mutual fund managers with those of individuals mutual fund managers are no better than individuals when it comes to their personal investments chatterjee & fan (2023) financial advisor and cfp archival data no explicit measure, but the presence of a professional is considered retirement savings, 6 components people living in financial advice deserts are less likely to have retirement accounts and to contribute regularly to their retirement accounts cici et al. (2017) financial advisor, title not specified archival data no explicit measure, but advisors are considered as the change in equity ownership in the funds financial advisors provide tangible benefits to investors, pilote et al. 62 professional in comparison to their clients before and after a tax distribution particularly with respect to taxation cloyd (1995) tax specialists lab experiment a combination of eighteen questions on the topic the information collected by the participants are compared with those of an experts’ panel prior knowledge positively affects the number of relevant items found, the ability to differentiate between relevant and irrelevant items, and reduces the amount of time needed to complete the information search cloyd (1997) tax specialists lab experiment a combination of eighteen questions on the topic the information collected by the participants are compared with those of an experts’ panel prior knowledge positively affects the number of relevant items found cummings et al. (2013) financial advisor, title not specified archival data no explicit measure, but the presence of an advisor is considered a positive influx of knowledge holding an individual tax-free account (roth ira) solicitation of a financial planner positively influences ownership of an individual taxfree account direr & visser (2013) financial advisor, title not specified archival data the holding of a university degree the percentage of assets invested in the stock market in relation to total assets individuals who consult with financial advisors with higher levels of education are more likely to invest in the stock market than clients of those with lower levels of education fan (2021) financial advisor, title not specified archival data no explicit measure, but the presence of an advisor is considered a positive influx of knowledge saving behaviors and credit usage individual's search for information and the external search for information (financial advisor) are positively associated with saving and good credit usage fischer & gerhardt (2007) financial advisor, title not specified analytical model no explicit measure, but advisors are considered as the extent of investment errors made by individuals financial advice helps individuals reduce their investment mistakes financial services review, 32(3) 63 professional in comparison to their clients grable & chatterjee (2014) financial advisor, title not specified archival data no explicit measure, but the presence of an advisor is considered an added value wealth change individuals who sought a financial advisor experienced a 6.25% lower loss of wealth as a result of a recession compared to those who did not seek an advisor hanna & lindamood (2010) unspecified analytical model no explicit measure no explicit measure the value of advice increases the more of the client's wealth may be lost, i.e., wealth volatility hershey et al. (1990) financial planning professional, title not specified lab experiment a combination of a ten-page test and professional functions the type of information requested and the spoken process experts solved the problem in less time and in fewer steps than novices experts have a more structured resolution process than novices experts select more relevant information than novices hershey & walsh (2000) accountants and financial planners lab experiment a combination of thirty-two questions on the topic the deviation between the participants’ solutions and those of an experts’ panel individuals with specialized knowledge select more relevant information than those with general knowledge, who still select more relevant information than naïve individuals individuals with specialized knowledge deviate less from the experts’ solution than those with general knowledge, who still perform better than naïve individuals horn et al. (2009) financial advisor, title not specified archival data no explicit measure, but advisors are considered as the composition of the asset portfolio investors who seek out a financial advisor make better investment choices pilote et al. 64 professional in comparison to their clients hudson & palmer (2014) financial advisor in a broad sense archival data no explicit measure, but advisors are considered as professional in comparison to their clients a combination of ten questions on financial behaviors low-income individuals who seek out a financial advisor have better savings behaviors and better budget management hung & yoong (2013) financial advisor, title not specified lab experiment and archival data no explicit measure, but the presence of an advisor is considered an added value the choice of asset allocation respondents who seek financial advice perform better than those who do not seek advice unsolicited financial advice is not followed by participants in the experiment kramer (2012) financial advisor, title not specified archival data no explicit measure, but advisors are considered as professional in comparison to their clients the performance of the asset portfolio there is no difference in riskadjusted performance between individuals investing alone or with an advisor the asset portfolio of clients seeking an advisor is better diversified liu et al. (2019) financial advisor in a broad sense archival data no explicit measure, but advisors are considered as professional in comparison to their clients annual investment, financial asset holdings and total assets individuals who receive financial advice hold more financial assets marsden et al. (2011) financial advisor, title not specified online survey no explicit measure, but advisors are considered as professional in comparison to their clients having established long-term financial goals, calculating the amount needed for retirement or having an emergency fund respondents who consult a financial advisor have better financial habits financial services review, 32(3) 65 martin & finke (2014) financial advisor, title not specified archival data no explicit measure, but advisors are considered as professional in comparison to their clients wealth accumulated for retirement individuals who seek the help of a financial planner and follow a comprehensive retirement plan are building greater wealth for retirement montmarquette & viennot-briot (2015) financial advisor, title not specified online survey no explicit measure, but advisors are considered as professional in comparison to their clients assets held individuals with a financial advisor for a minimum of four years have more financial assets than those without an advisor the influence of the financial advisor is primarily through improved client savings practices montmarquette & viennot-briot (2019) financial advisor, title not specified online survey no explicit measure, but advisors are considered as professional in comparison to their clients assets held for the period from 2009 to 2013, individuals who retained their financial advisor saw the value of their assets grow by 16.4%, while the value of assets for those who left their advisor grew by 1.7% for the same period. moreland (2018) financial advisor in a broad sense archival data no explicit measure, but the presence of an advisor is considered an added value financial behaviors individuals who consult a financial advisor have better financial behaviors, such as holding a savings account or paying their credit card on time and in full mountain et al. (2021) financial advisor, title not specified online survey no explicit measure, but the presence of an advisor is considered as a source of learning financial behaviors individuals who consult a financial advisor have better financial behaviors pan et al. (2020) financial advisor, title not specified archival data no explicit measure, but the presence of an participation in the stock market financial advice affects stock market participation for individuals with high financial pilote et al. 66 advisor is considered an added value literacy and a preference for diversification park & yao (2016) financial advisor in a broad sense archival data no explicit measure, but advisors are considered as professional in comparison to their clients attitude to risk and asset ownership individuals who use a financial planner are more consistent in their risk taking and financial behaviors shapira & venezia (2001) financial advisor, title not specified archival data no explicit measure, but the presence of an advisor is considered a positive influx of knowledge timing of transactions, transaction volume and yield clients investing with an advisor make more trades than those investing alone clients investing with an advisor earn a higher return before commissions than those investing alone smith et al. (2012) financial advisor, title not specified archival data no explicit measure, but the presence of an advisor is considered a positive influx of knowledge holding an individual tax-free account (roth ira) individuals with a financial advisor or higher level of financial literacy are more likely to hold an individual taxfree account spilker (1995) tax specialists lab experiment a combination of ten questions on the topic the keywords selected by the participants are compared with those of an experts’ panel knowledge positively affects the number of relevant keywords selected the procedural group selects the most relevant keywords, followed by the declarative group and finally the naive group von gaudecker (2015) financial advisor, title not specified archival data no explicit measure, but advisors are considered as professional in comparison to their clients the cost of underdiversification of assets individuals who do not consult a financial advisor or who do not have a high level of financial literacy are at the greatest risk of experiencing losses related to asset underdiversification financial services review, 32(3) 67 winchester & huston (2014) financial advisor, title not specified survey no explicit measure, but advice comes from experts the progress individuals felt they had made in achieving their financial goal an ongoing relationship with a financial advisor improves the financial goal attainment of individuals who perceive themselves to be less in control winchester & huston (2015) financial advisor, title not specified survey no explicit measure, but advice comes from experts individuals rate their performance in accumulating wealth, preventing losses, and smoothing consumption middle-class individuals who receive integrated financial advice are 3 times more ready for retirement, 2 times more able to use their benefits, and nearly 2 times more likely to have an emergency fund winchester et al. (2011) financial advisor, title not specified survey no explicit measure, but advice comes from experts long-term commitment to a financial goal individuals who consult a financial advisor have a greater long-term commitment to their financial goals, i.e., they rebalance their portfolio more zhang (2014) authorized financial adviser (afa) archival data no explicit measure, but advice comes from experts annual return on investment and asset allocation the return on investment of individuals receiving financial advice is not significantly higher than that of individuals investing alone individuals receiving financial advice hold more risky assets (stocks and property) than individuals investing alone evaluating the relationship between ifa remuneration and advice quality: an empirical study jiří šindelářa,*, petr budinskýb auniversity of finance and administration prague, estonská 500, 101 00, prague 10, czech republic buniversity of finance and administration prague, estonská 500, 101 00, prague 10, czech republic abstract this article deals with the interaction between commission remuneration of independent financial advisers and selected sales factors, including the quality of advice. utilizing data on investment transactions and a linear model with mixed effects, we have found that the link between commission and quality of the subsequent recommendation is not homogeneous, and advice-bias potential is present only in a limited range of organizational environments, connected mainly to the flat-structure business model. on the other hand, arbitrage between different product classes was found to create a biasing potential across almost all types of firms, creating potential for market systemic risk. finally, the effect of information provided was proved to be significant only to a very limited extent. © 2017 academy of financial services. all rights reserved. jel classification: g22; g23; g28; d14; d18 keywords: financial advice; conflict of interests; agent principal problem; life insurance; investments funds; systemic distribution risk 1. introduction commission based sales represent the principal distribution channel for financial products in many oecd countries. according to the insurance europe (2014) survey, financial agents (intermediaries, advisers etc.) accounted for nearly half (47.1%) of the new life insurance business in germany, with other central european countries showing a similar situation.1 * corresponding author. tel.: �420-731-537-207; fax: �420-221-628-509. e-mail address: sindelar@mail.vsfs.cz (j. šindelář) financial services review 26 (2017) 367–386 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. one of the most important areas in which advice is provided on a commission basis is pension planning, which in most cases leads to the purchase of a unit-linked life insurance, investment fund or personal pension product. as the oecd (2015) stated in its recent pension outlook, 24% of the member states’ pension-linked assets are in those product classes, with a large portion of them being allocated on the basis of commission-remunerated advice. while commission (third party inducement) remains the principal remuneration mechanism for agents, it is coming under increasing pressure chiefly on european soil. the main argument, as stated in the european insurance and occupational pensions authority (eiopa, 2016, p. 41) advice on the pan-european pension product (pepp), is that “commissions which are often paid by product manufacturers potentially lead to a conflict of interest between the interest of the distributor to gain the commission and the interest of the customers to obtain nonbiased services from the distributor.” similar statements can be found in proposals linked to investment products (markets in financial instruments directive mifid ii) and insurance distribution (insurance distribution directive idd). conflict of interest and its potentially detrimental effect on advice has even led to remuneration restrictions being applied, particularly in the area of unit-linked life insurance. from a theoretical perspective, potential bias created by commission based financial advice is grounded in the general agency theory, as the moral hazard and adverse selection problems (ross, 1995). both result in an inefficient contract for the primary principal (customer), whose bias is amplified by the introduction of a secondary principal (distribution firm). while there are abundant articles pointing to the biased service produced by agents operating on commission (e.g., chalmers and reuter, 2015; gravelle, 1994; inderst and ottaviani, 2011; palazzo and rethel, 2008; schwarz and siegelman, 2015), many of them offer limited empirical background or are based on a less-conclusive (statistical) methodology. some articles, on the other hand, did not find the commission-based remuneration to bear significantly negative consumer consequences (gerhardt and hackethal, 2009) or offered mixed results (glazer, 2007; tseng, 2011). this article seeks to investigate the relationship between paid-out commission, complimentary sales factors and quality of advice provided by intermediaries (agents, financial advisors) in the area of investment products (investment funds, unit-linked insurance) in the czech republic, as the central-eastern europe transit market. the article is divided into three parts: (1) an overview of current empirical findings is provided and research hypotheses constituted, (2) a statistical examination of the relationship of selected factors is carried out, and finally (3) resulting conclusions are summarized and discussed with reference to relevant literature. 2. literature overview as evinced by numerous studies (e.g., lopez et al., 2006; pullins, 2001), a reward scheme plays a crucial role in salesforce motivation. however, its interaction with the quality of advice provided to customers is the subject of scrutiny because of the central role such advice often plays in personal finance. in particular, the effect of a commission-based remuneration 368 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 scheme is a well-covered theme of scientific literature. table 1 summarizes the principal studies in this field. from the factual perspective, the outcome of recent empirical studies underlines the schism outlined in the introduction. although recent literature offers numerous articles on the topic, including an abundant group based on theoretical proofing (e.g., gravelle, 1994; inderst and ottaviani, 2009), no unequivocally dominant pattern is evident. while many articles do point to a compromising effect of commission remuneration, there is a substantial body of research that fails to confirm this link, or even points to the opposite, in terms of customer benefit (a more detailed meta-analysis, with outcomes, can be found, e.g., in burke et al., 2015). as a theoretical assumption for this article, taking a cautious approach, we shall presume that commission remuneration does have a negative effect on subsequent advice quality. yet in reality, this is not a resolute hypothesis, but more of an open question. remuneration scheme, although deemed crucial, is not the only factor potentially influencing the quality of the advice and sales process. in this article, four additional variables were introduced to the model, with the following theoretical background. 2.1. product type although to a large degree unit-linked insurance and investment funds share a common market and are often sold interchangeably, both product classes exhibit differences with regard to fee structure, product features as well as legal framework (for details see e.g., ruprecht, 2007). these have been reported to affect advice quality in some markets, particularly in relation to the insurance business (halan et al., 2014; sane et al., 2013). taking this experience into account, our expectation is that unit-linked life insurance will be more prone to poor advice. 2.2. sales firm structure different internal structures of agent companies have been reported to provide different effects on quality of advice, especially in relation to multilevel marketing systems (reifner et al., 2012). looser structures with lower emphasis on group-incentivizing, on the other hand, have been found to be more supportive of advice quality (danilov and biermann, 2013). we expect to find a similar pattern, with structural networks generally more susceptible to biased advice than flatter “branch like” entities. 2.3. sales firm size there is a conflicting view of how the size of a distribution firm can potentially affect the quality of its service. while some studies suggest that increasing size leads to higher adviser misconduct (egan et al., 2016), others have found quite the opposite, either praising advice provided by medium-large chains (australian securities and investments 369j. šindelář, p. budinský / financial services review 26 (2017) 367–386 t ab le 1 m et aan al ys is of re ce nt em pi ri ca l st ud ie s st ud y m et ho d pr od uc t/r eg io na l fo cu s su rv ey ed sa m pl e r es ul ts a na go l et al . (2 01 2) m ys te ry sh op pi ng (a ud its ), un iv ar ia te re gr es si on l if e in su ra nc e 20 11 (i nd ia ) 30 4 in su ra nc e sa le sag en ts (5 57 au di ts ) c or e fin di ng s of th e qu al ity of ad vi ce ex pe ri m en t: b et w ee n 60 an d 80 % of au di ts en de d w ith a re co m m en da tio n of le ss su ita bl e in su ra nc e po lic y (w ho le in su ra nc e) w ith hi gh er co m m is si on ** * e ve n w he n au di to rs si gn al ed th at th ey ar e m os t in te re st ed in te rm in su ra nc e an d ne ed ri sk co ve ra ge , m or e th an 60 % of au di ts re su lt in w ho le in su ra nc e (u ni tlin k) be in g re co m m en de d* ** a ge nt s pr im ar ily ca te r to cu st om er s (e ith er th ei r be lie fs or ne ed s) by re co m m en di ng th at th ey pu rc ha se te rm in su ra nc e in ad di tio n to w ho le in su ra nc e, as op po se d to re co m m en di ng te rm in su ra nc e al on e. po po va (2 01 0) b eh av io ra l ex pe ri m en t (s en de rre ce iv er ga m e) , w al d te st , f te st in su ra nc e du m m y 20 09 –2 01 0 (g er m an y) 31 4 un de rg ra du at e st ud en ts m aj or fin di ng s of th e be ha vi or al ex pe ri m en t: in al l tr ea tm en ts bu t on e, th e fr eq ue nc y of tr ut hf ul ad vi ce is hi gh er w ith di re ct pa ym en t th an w ith co m m is si on pa ym en t* ** t he ob lig at or y di re ct pa ym en ts by cl ie nt s ar e no t ap pr op ri at e fo r re du ci ng th e co nfl ic t of in te re st of ad vi so rs ** * t he la rg e vo lu nt ar y di re ct pa ym en t by cl ie nt s is th e m os t su cc es sf ul m ec ha ni sm fo r re du ci ng th e co nfl ic t of in te re st of ad vi so rs ** * c ha lm er s an d r eu te r (2 01 5) a nn ua l re tu rn , a nn ua l vo la til ity , o l s re gr es si on r et ir em en t po rt fo lio s (f un ds ) 19 99 –2 00 9 (u sa ) 5 80 7 pa rt ic ip an ts of op tio na l re tir em en t pl an (o r p) m aj or di ff er en ce s of ad vi se d po rt fo lio s in co m pa ri so n w ith ta rg et -d at e fu nd pe rf or m an ce : l ow er af te rfe e an nu al re tu rn s (� � � 2. 98 % )* ** h ig he r vo la til ity of re tu rn s (� � 0. 43 % ) l ow er sh ar pe ra tio ** h ig he r av er ag e fe es (� � 0. 90 % ) c up ac h an d c ar so n (2 00 2) q ue st io nn ai re su rv ey , f te st , � 2 te st l if e in su ra nc e 20 02 (u sa ) 33 6 in su ra nc e sa le sag en ts r es ul ts in di ca te th at : n ei th er am ou nt of co ve ra ge no r ty pe of co ve ra ge re co m m en de d va ri ed ac ro ss th e fiv e al te rn at iv e co m pe ns at io n co nd iti on s (n o st at is tic al ly si gn ifi ca nt lin k) n ei th er co m m is si on le ve l no r fe e fo r se rv ic e le ve l in flu en ce d th e lik el ih oo d of pr od uc t re co m m en da tio n (n o st at is tic al ly si gn ifi ca nt lin k) (c on ti nu ed on ne xt pa ge ) 370 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 t ab le 1 (c on tin ue d) st ud y m et ho d pr od uc t/r eg io na l fo cu s su rv ey ed sa m pl e r es ul ts g er ha rd t an d h ac ke th al (2 00 9) po rt fo lio ch ar ac te ri st ic s (e qu ity sh ar e, sh ar pe ra tio s, an d so fo rt h) , tte st st at is tic s in ve st m en t fu nd s 02 /2 00 6– 07 /2 00 7 (g er m an y) 59 7 in ve st or s w ho sw itc he d fr om no nad vi se d to ad vi se d du ri ng th e sa m pl e pe ri od (s ub sa m pl e) m aj or ef fe ct s of in ve st m en t ad vi ce in co m pa ri so n w ith no na dv is ed in ve st or s: h ig he r tr ad in g ac tiv ity ** * l es s ri sk y an d sp ec ul at iv e tr ad in g r is in g di ve rs ifi ca tio n* */ ** * n o ri si ng ra tio of ex pe ns iv e pr od uc ts so ld (k ic kb ac k pa ym en ts ) l i (2 01 5) in ve st m en t ch ar ac te ri st ic s (e xc es s re tu rn s, ne t flo w , fe es , fr on t lo ad ), o l s re gr es si on in ve st m en t fu nd s 10 /1 99 9– 6/ 20 12 (u sa ) 42 4, 11 5 to ta l ob se rv at io ns (a ct iv el y m an ag ed eq ui ty fu nd s) fu nd flo w s in di ca te th at : r et ur n ch as in g is st ro ng er am on g fu nd s so ld w ith hi gh (u pfr on t) co m m is si on s* ** a m on g m ul tip le as se t cl as se s, re tu rn ch as in g in cr ea se s w ith br ok er co m m is si on s* ** m os t of th e im pa ct is on th e pu rc ha se of pa st w in ne rs ** *, w ith no de te ct ab le ef fe ct on th e re de m pt io n of pa st lo se rs in ve st or s in in st itu tio na l sh ar es ex hi bi t al m os t as m uc h re tu rn ch as in g as re ta il in ve st or s* ** l in ai nm aa et al . (2 01 5) in ve st m en t ch ar ac te ri st ic s (n et re tu rn s, ex ce ss re tu rn s, fr on ten d lo ad s, tr ai lin g co m m is si on s) , tte st , f te st in ve st m en t fu nd s 1/ 19 99 –6 / 20 12 (c an ad a) 58 1, 04 4 in ve st or s, 5, 83 8 ad vi so rs m aj or fin di ng s re ga rd in g ad vi se d po rt fo lio s: if th e ad vi so r be ne fit s fr om th e tr ad e, th e cl ie nt al so be ne fit s fr om th e tr ad e 57 % of th e tim e (t ra de s th at ar e bo th co st ly to th e cl ie nt an d w ith ou t ap pa re nt be ne fit s, bu t be ne fit tin g th e ad vi se r ac co un t fo r 5. 4% ) a di sp ro po rt io na te nu m be r of th e tr ad es id en tifi ed as se lf -s er vi ng (c os tly , on ly ad vi so r be ne fit s) ar e co nc en tr at ed am on g a sm al l nu m be r of ad vi so rs (3 .3 % ), ne t re tu rn s al ph as de cr ea se sh ar pl y in th os e cl ie nt s’ po rt fo lio s t se ng (2 01 1) q ue st io nn ai re su rv ey , � 2 te st l if e in su ra nc e 20 10 (t ai w an ) 36 1 fu lltim e lif e in su ra nc e sa le sp eo pl e m aj or ou tc om es in re la tio n to th e te st ed sc en ar io s: 78 .4 % of th e re sp on de nt s w ou ld se ll a po lic y w ith an in te re st ra te be ne fic ia l to th e cu st om er in st ea d of th e on e be ne fic ia l to th ei r co m pa ny ** * 73 .6 % of th e re sp on de nt s w ou ld se ll a po lic y in lin e w ith cu st om er ne ed s in st ea d of th e on e be ne fic ia l to th ei r co m pa ny ** * 68 .8 % of th e re sp on de nt s w ou ld se ll a po lic y bo th to a he al th y an d un he al th y cu st om er , no tw ith st an di ng th e ef fe ct on hi s co m pa ny *p va lu e � 0. 1; ** p va lu e � 0. 5; ** *p va lu e � 0. 01 . o l s � o rd in ar y l ea st sq ua re s re gr es si on . 371j. šindelář, p. budinský / financial services review 26 (2017) 367–386 commission, 2003), or implying that smaller firms in fact offer limited services and restricted advice (eckardt and räthke-döppner, 2010). based on knowledge of the surveyed market, we presume that larger companies will incline to lower quality of advice, that is, increasing size of the company will have a negative effect on the excellence of its service. 2.4. information available to the salesforce there is little doubt that salesforce competence and professionalism represents a strong stimulus to customer satisfaction and trust (ali et al., 2015; johnson and grayson, 2005; tsoukatos and mastrojianni, 2010). furthermore, a direct link between specialized information provided to individual agents and the subsequent quality of their service has also been proven (eckardt and räthke-döppner, 2010). accordingly, a positive effect of information granted to the salesforce is also expected within our sample. 2.5. research hypotheses based on the previous theoretical overview and prospected model composition, we set a total of five research hypotheses: h1: the amount of commission paid out for insurance products differs significantly from investment funds. h2: there is a significant correlation between the amount of commission paid out and the number of product trainings provided to the salesforce. h3: the there is a significant correlation between the amount of commission paid out and the advice quality. h4: there is a significant difference between the amount of commission paid out for insurance products among diverse sales firm structures. h5: there is a significant difference between the amount of commission paid out for insurance products among diverse sales firm sizes. 2.6. data the data for the empirical part of our survey was provided by eight independent advisory companies (no exclusive ties or direct ownership by financial institutions), who were asked to provide a full listing of the intermediated sales for a random month of the year.2 their overall sales performance is outlined in table 2. by combining the individual listings from the above participants, data on a total of 10,105 transactions performed in the years 2013–2015 on the basis of advice provided by financial agents was gathered. only investment products (ucits3 vehicles) and investment-insurance products (unit-linked4) were concerned. overall, the transactions recorded, encompass 55 372 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 t ab le 2 o ve rv ie w of co m pa ni es pa rt ic ip at in g in th e re se ar ch c om pa ny st ru ct ur e n o. of in di vi du al ad vi se rs (2 01 5) n o. of ne w lif e in su ra nc e co nt ra ct sa so ld (2 01 5) m ar ke t sh ar e in lif e in su ra nc ea –i fa m ar ke t (2 01 5) n o. of ne w in ve st m en t fu nd s co nt ra ct s so ld (2 01 5) m ar ke t sh ar e in in ve st m en t fu nd s– if a m ar ke t (2 01 5) a m l m 4 69 2 66 74 4 27 .1 53 % 37 80 8 19 .1 05 % c po ol 1 78 4 29 67 2 12 .0 71 % 17 78 0 8. 98 4% b m l m 88 6 30 23 2 12 .2 99 % 20 47 2 10 .3 45 % d fl at 37 0 7 29 2 2. 96 7% 14 84 4 7. 50 1% e m l m 95 3 48 8 1. 41 9% 91 6 0. 46 3% h fl at 28 29 2 0. 11 9% 32 4 0. 16 4% g fl at 13 46 8 0. 19 0% 16 0 0. 08 1% f fl at 5 47 0. 01 9% 21 0. 01 1% t ot al — 7 87 3 13 8 23 5 56 .2 37 % 92 32 5 46 .6 52 % if a � in de pe nd en t fi na nc ia l a dv is er s. a r eg ul ar ly pa id co nt ra ct s. 373j. šindelář, p. budinský / financial services review 26 (2017) 367–386 unique insurance/investment products and were advised on by a total of 2,658 individual agents. their basic overview is stated in appendix 1, stipulating that the majority of the recommended investments were following dynamic strategy with a minimum of five years maturity, which is consistent with a longer-term horizon of most financial (pension) plans. furthermore, the survey only incorporated regularly (monthly) paid instruments, which form the backbone of pension planning.5 within the sample, each transaction was described by a set of variables linked to the factors described in the theory chapter. the linkage between general factors and research variables is outlined in table 3. from the structural perspective, the survey sample represents a very diverse portfolio. summary statistics of all variables are outlined in appendix 2. 2.7. quality assessment as mentioned in the theoretical part, the indicator of advice quality (qual) is one of the volatile parts of recent research. in this study, the indication of advice (recommendation) quality is based on the evaluation carried out by panel of independent experts.6 the advantage of this approach is that it can capture additional information above the purely financial/quantitative metrices, as demonstrated by relationships indicated in appendix 1. the panel rated every product that was recommended inside our sample in three basic dimensions: (1) price – economical attributes of the product (fees, potential yield through the life-cycle of the product), (2) quality – availability, accessibility of the product and related customer care, and (3) sustainability – transparency and sustainability of the product (as it is being offered or promoted). from a methodological perspective, all three dimensions of quality were defined in a way that is positively associated with customer utility (i.e., higher value always brings higher benefit) and not mutually contradictory (e.g., better price rating not interfering with the sustainability one), similarly to tseng (2011) and anagol et al. (2012) studies. our aim was not to assess the individual suitability of given products, but rather to evaluate, whether advisers might be stipulated to offer lower quality products with a higher reward on a global scale. each of the experts had to provide his individual multicriterial assessment not only regarding the three quality dimensions, by ordinally sequencing products in given categories (if, uil), but also by setting weights for their relative importance to customer decisionmaking in a given year. every product was then awarded a number of points based on individual weights assigned and their relative placing, normalized between 1 (best rating) and 5 (worst rating), with the points corrected for different numbers of products between categories. the expert body itself was proportionally composed of 355 members: academicians, independent experts, senior bank specialists, and senior financial advisors; with every member being approved by the governing board composed of respected industry figures the internal validity of the framework was further tested on samples of five random products from each category through the governing board ex-post examination. by this procedure, two 374 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 t ab le 3 in de pe nd en t m od el va ri ab le s– ex pl an at io n t he or et ic al fa ct or v ar ia bl e in di ca to r t yp e d en om in at io n c om m is si on re m un er at io n c o m m a m ou nt of fr on t co m m is si on pa id to th e fin al (i nd iv id ua l) ag en ta c on tin uo us c ze ch c ro w n (c z k )b pr od uc t ty pe pr o d pr od uc t cl as si fic at io n n om in al in ve st m en t fu nd , u ni tlin ke d in su ra nc e sa le s fir m st ru ct ur e st r u c fi rm cl as si fic at io n in to th re e gr ou ps n om in al m l m , br ok er -p oo l, fla t st ru ct ur e sa le s fir m si ze si z e fi rm cl as si fic at io n in to th re e gr ou ps o rd in al b ig (� 50 0 if a s) , m ed iu m (5 0– 50 0 if a s) , sm al l (� 50 if a s) in fo rm at io n av ai la bl e to th e sa le sf or ce in fo n um be r of tr ai ni ng s pr ov id ed in re la tio n to gi ve n pr od uc t du ri ng th e la st 12 m on th sc o rd in al m uc h hi gh er th an av er ag e, hi gh er th an av er ag e, si m ila r to av er ag e nu m be r (i n re la tio n to th e pr od uc t ty pe ), le ss th an av er ag e, m uc h le ss th an av er ag e if a � in de pe nd en t fi na nc ia l a dv is er s. a w e ta ke in to ac co un to nl y th e in iti al co m m is si on pa id ou tf or th e sa le (u pfr on t) ,w hi ch is th e va st ly pr ef er re d m et ho d of re m un er at io n in th e ta rg et m ar ke t. t ra ile r co m m is si on s ar e ne gl ig ib le . b a dv ic e co m pa ni es in ou r sa m pl e ut ili ze on ly va ri ab le re m un er at io n w ith no fix ed co m po ne nt (fi xe dco m m is si on m od el is ne gl ig ib le on ta rg et m ar ke t) . b ec au se of co m m is si on be in g de ri ve d fr om si ze of th e tr an sa ct io n, al l of th e co m m is si on am ou nt s w er e tr an sf or m ed to a co m m on co m pa ra tiv e ba si s, re pr es en tin g 1, 00 0 c z k pa ym en t (t he m os t co m m on le ve l of co nt ri bu tio n on ta rg et m ar ke t) . c in tr od uc to ry tr ai ni ng s fo r ne w co m er s w er e ex cl ud ed , on ly pr od uc t tr ai ni ng s w er e ta ke n in to ac co un t. 375j. šindelář, p. budinský / financial services review 26 (2017) 367–386 different measurements were obtained, gaining material for the construction of a monotraitheteromethod (mthm) matrix (campbell and fiske, 1959; crocker and algina, 2008). after correlating the two data lines with goodman and kruskal’s � (p � 0.000), we achieved the following results (table 4). the level of correlation achieved shows strong correspondence with both methods of measurement (crocker and algina, 2008 recommend 0.50 to be the minimum), providing proof of the (convergent) construct validity of the panel evaluation carried out. 3. method as mentioned above, this article deals with the evaluation of the link between selected sales factors and the quality of financial advice, in terms of a client’s subsequent purchase. from the given set of variables, our basic research model is constituted as follows: log(comm � 1) � (size � struc) � (prod � info � qual) � (1�id_comp/id_ifa). (1) for the data analysis, the linear mixed effects models were used. in a classical linear model, with only fixed effects considered, it is assumed that all observations are independent. since this does not hold true for the analyzed data (transactions done by one sales person could not be independent since they depend on the sales person’s knowledge, experience etc., and, moreover, also transactions done under a given company are not independent for similar reasons), the random effects were introduced. two nested random effects appear in our model: an effect of the sales person nested in the random effect of the company. in the model equation is such a setup written as 1id_comp/id_ifa. the fixed effects appear in the model in interactions which is denoted in the model equation by an asterisk. the baseline model of the form (size � struc) * (prod � info � qual) in fact means that we assume that the commission depends on prod, info and qual in a priori different ways in different kinds of companies (according to their size and structure). such differences are further tested and interpreted. the purpose of breaking the whole sample to partial subsamples defined by size and struc is to capture the effect of these factors described in background literature, such as reifner’s et al. (2012) comprehensive study. a p-values less than 0.05 was considered statistically significant. analysis was conducted using r statistical package, version 3.2.3 (r core team, 2015). variance analysis outcomes for the model are summarized in table 5. table 4 mthm matrix m1 m2 m1: main measurement 0.83 0.76 m2: control measurement 0.76 1.0 376 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 going through the p-values of the model, we observe that while two of the sales factors (info, qual) do not have a significant effect on commission on average, all of the factors have a significant relationship with a dependent variable when grouping variables (struc, size) are taken into account. in other words, all of the surveyed sales factors interacted with the amount of commission paid out in each of the company contexts (delimited by the size and sales structure) in a significantly different manner. detailed results in this regard are presented next. 4. results consequently, our results are divided into nine different combinations of company size and sales structure, summarized by table 6. let us use sales structure as our primary differentiator, summarizing mlm, pool, and flat companies of different sizes into three distinct chapters. 4.1. mlm companies the model estimates indicate three principal findings. first, in all of the mlms, irrespective of their size, the difference between the two surveyed product classes (if, uli) has a significant effect on commission paid out, with the unit-linked insurance always providing significantly higher commission. secondly, the information provided to the ifa-force, in terms of training frequency, affects commission level significantly only in a single type of firm—small mlm (positively). in the medium and large sized networks, its effect was not found to be significant on the given p level. finally, our last factor (quality of purchased product) provides a significant outcome only in one environment—large mlm firms. a positive value of the estimate indicates that increasing advice quality provides lower commissions and vice versa; thus, implying that the inducement paid out to the sales force can distort the quality of ifa service in terms of the recommended purchase. intensity of the effect, however, seems rather negligible. table 5 variance analysis outcome sum sq mean sq numdf dendf f value p-value struc 6.9091 3.4546 2 3 11.3158 0.0348 size 1.3570 1.3570 2 183 4.4450 0.0364 prod 42.2420 42.2420 1 2527 138.3682 0.0000 info 0.5126 0.5126 1 8822 1.6792 0.1951 qual 0.0731 0.0731 1 7267 0.2395 0.6246 struc:prod 7.9718 3.9859 2 8847 13.0563 0.0000 struc:info 31.0775 15.5388 2 8771 50.8989 0.0000 struc:qual 19.7648 9.8824 2 9320 32.3708 0.0000 size:prod 3.3587 1.6794 2 8267 5.5009 0.0041 size:info 3.1736 1.5868 2 9375 5.1977 0.0055 size:qual 2.6662 1.3331 2 9061 4.3668 0.0127 377j. šindelář, p. budinský / financial services review 26 (2017) 367–386 4.2. pool companies according to our results, ifas gathered under pool structures also receive significantly different commissions for both product classes, in favor of the uli. contrary to mlms, however, the information provided to the ifa-force does significantly affect the amount of commission in quite an opposite case: with the large companies and in a negative manner. in other words, the more training the salespeople go through, the lower commission they are achieving.7 the most dramatic, however, is the relationship between the amount of commission and the quality of the client�s purchase. found significant in two environments (large, small), this factor exhibited a consistently table 6 results overview estimate standard error z value p-value hypotheses mlm, large sized prod. difference effect �0.504 0.022 �22.666 0.000 h1 accepted info effect �0.003 0.009 �0.341 0.992 h2 not accepted qual effect 0.086 0.034 2.539 0.040 h3 accepted mlm, medium sized prod. difference effect �0.793 0.141 �5.603 0.000 h1 accepted info effect 0.090 0.069 1.294 0.500 h2 not accepted qual effect 0.426 0.426 1.000 0.702 h3 not accepted mlm, small sized prod. difference effect �1.925 0.453 �4.253 0.000 h1 accepted info effect 0.669 0.209 3.201 0.005 h2 accepted qual effect �1.742 0.872 �1.997 0.146 h3 not accepted firm pool, large prod. difference effect �0.685 0.030 �23.172 0.000 h1 accepted info effect �0.162 0.013 �12.301 0.000 h2 accepted qual effect �0.447 0.062 �7.268 0.000 h3 accepted firm pool, medium prod. difference effect �0.973 0.146 �6.656 0.000 h1 accepted info effect �0.069 0.071 �0.973 0.740 h2 not accepted qual effect �0.108 0.431 �0.249 0.997 h3 not accepted firm pool, small prod. difference effect �2.106 0.454 �4.637 0.000 h1 accepted info effect 0.510 0.210 2.432 0.052 h2 not accepted qual effect �2.275 0.875 �2.599 0.033 h3 accepted firm flat, large prod. difference effect 0.117 0.413 0.283 0.993 h1 not accepted info effect �0.443 0.185 �2.396 0.055 h2 not accepted qual effect 2.141 0.769 2.782 0.019 h3 accepted firm flat, medium prod. difference effect �0.172 0.387 �0.444 0.955 h1 not accepted info effect �0.351 0.171 �2.046 0.119 h2 not accepted qual effect 2.480 0.640 3.874 0.000 h3 accepted firm flat, small prod. difference effect �1.304 0.187 �6.969 0.000 h1 accepted info effect 0.228 0.097 2.347 0.057 h2 not accepted qual effect 0.313 0.412 0.759 0.820 h3 not accepted 378 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 negative direction of effect. in other words, advisers operating under a pool umbrella gain significantly higher reward when recommending products with higher quality. in these settings, therefore, the amount of commission does not exhibit a negative potential in terms of advice distortion. 4.3. flat companies the model estimates and p-values indicate that the medium and large sized flat companies represent the most neutral advisory model in our sample. none of the two product classes and or their difference had a significant effect on final ifa remuneration, the same being true for the amount of information provided. the only significant factor was the quality of the recommended product, which interacted with commission in a positive manner. this implies that the rewarding scheme had distortive potential on the final recommendation. the situation with small organizations of flat structure is rather different and resembles previous types. different product classes earn significantly different commissions (in favor of uli). the number of trainings was found (just) to have no significant effect, and product quality is clearly insignificant. such results draw a sharp distinction with medium and large sized flat organizations. reviewing the results through our five research hypotheses, we have obtained rather diverse outcomes. the first hypothesis, based on product class effect on commission, was found effective on a wide scale and was confirmed (h1 accepted) in two-thirds of the organizational types. regarding the hypothesized effect of information provided to the salesforce through product trainings, these significantly affected commission only in two cases (h2 accepted) of diverse structure and size, with no apparent connecting pattern. our third and crucial assumption, depicting a statistically significant link between commission and quality of advice, was found to hold in five out of nine surveyed organizational environments (h3 accepted). finally, the remaining hypotheses (h4 and h5) were both related to the grouping variables (sales firm structure and size) and as such were identified as accepted during the initial variance analysis. all in all, variables included in our model were found significant in most cases, retrospectively validating the model composition. 5. discussion compared with the theoretical basis, our survey for the most part indicates more favorable results than expected by other articles. it was confirmed that in the majority of sales organizations there are significant incentive differences between investment fund and unitlinked life insurance, creating a potential for advice bias and client detriment as described by sane et al. (2013) or halan et al. (2014). despite this outcome, there are organizations that hold limited market share, but prove resistant to commission divergences, operating with flat business structure. regarding the effect of information provided to the sales force through product trainings, observations conducted by eckardt and räthke-döppner (2010) were not 379j. šindelář, p. budinský / financial services review 26 (2017) 367–386 confirmed. significant effects produced by this factor were detected only in a very limited range, indicating that the popular thesis of more education leading to higher earnings is not valid in our ifa sample. sales firm structure and size were identified as crucial elements of the advice process, in accordance with indirect implications published by reifner et al. (2012), danilov and biermann (2013), and egan et al. (2016). confirmation of those two factors shows that judging the whole ifa segment as an internally homogeneous sum of individuals, as exhibited in articles cupach and garson (2002), anagol et al. (2012), and popova (2010) is fundamentally inappropriate, as there are statistically significant functional differences between diverse organizational entities. a “one size fits all” approach, as embodied in many eu regulations (e.g., mifid, idd) and envisaged by part of the academia (reifner et al., 2012), leads to redundant business costs and dubious consumer effect, given our empirical results. principal outcomes of the article are related to the remuneration–advice linkage. theoretical expectations here were more in favor of a negative impact of commission remuneration on quality of subsequent advice. these expectations were largely disproved by our model. only in three organizational environments did the data indicate a negative relationship between quality of a client’s purchase and commission paid out to the ifa, creating a potential discord that could bias the advice. in only two environments of the same business structure (flat organizations) did the model estimate reach major value and these represent a minor part of the ifa market.8 in other words, a remuneration scheme induced potential for recommending products with lower overall quality, as reported by beyer et al. (2013) and chalmers and reuter (2015), or for mis-selling a totally inappropriate product as detected by anagol et al. (2012) is not overly present in the target market. the results related to mlm systems mostly contrast with observations collected in other countries, notably by professor reifner et al. (2012) and his team. reifner’s conclusion that “financial interest in the advice is much more biased” within the structured mlm networks (p. 78) cannot be considered confirmed. 6. conclusions the relationship between ifa remuneration and quality of subsequent advice is a frequent point of current research and policy making. most of the previous studies found that a commission remuneration scheme has a biasing effect on ifa recommendations and subsequent client purchase. in this article, we found that the negative potential created by higher earnings for recommending less quality products is present only in a minority of the ifa organizations, particularly in the flat structures. pool businesses, on the other hand, were diagnosed as more resistant in this regard, not exhibiting undesirable remuneration-based conflict of interest potential. our findings are bounded by three main limitations. we dealt just with the independent advisory part of the market, evading captive (dependent) bank and insurance company networks. although similar results can be foreseen according to some articles 380 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 (reifner et al., 2012), expanding the analysis on captive channels is vital as substantial sales production is realized through them on a (dependent) advice basis. the second limitation is related to the evaluation method utilized with regards to the quality indicator. using a panel of experts’ assessment brings an important new perspective on the topic, yet despite controlled validity, wider back testing of value-added by our alternative approach is vital. the final limitation is related to the macro level of the analysis. as such, it did not attempt to identify mis-selling in relation to individual transactions or clients, but aimed at uncovering main trends on the whole population, delimited by the survey sample. all these differences need to be taken into account, when interpreting study results and they also represent the main directions for following distribution research. notes 1 slightly lower, yet proportionate numbers are true for investment funds (kalus et al., 2015). 2 excluding july, august, and december periods. 3 collective investments as defined by the eu undertakings for the collective investment in transferable securities (ucits) directive. 4 insurance-based investment products as defined by eu directive on insurance distribution (idd). 5 third pillar pension savings product was omitted, because it already has a legal cap on commissions in force, preventing a meaningful analysis at this point. second pillar and occupational pensions are not implemented in the target market. 6 for this purpose, we utilized the financial academy of the golden crown (zlatá koruna, 2016) institute. golden crown provides an independent, arguably most renowned and prestigious high-level financial product rating in the czech republic. as of 2016, it evaluated a total of 191 products in 15 product categories. 7 in the case of small pools, the effect was nearly significant, in a positive direction. 8 according to analysis created by independent group (experti na finance, 2016), out of the top 10 ifa companies in the czech republic, which account for about two-thirds of the independent advice market, mlm represent 78.64%, while pool structures remaining 21.36% (in terms of sales force size). acknowledgment this article was created with the contribution of institutional support for long-term conceptual development of the research organization university of finance and administration. 381j. šindelář, p. budinský / financial services review 26 (2017) 367–386 a pp en di x 1 pr od uc ts pa rt ic ip at in g in th e su rv ey –a n ov er vi ew pr od uc t pr od uc t ty pe pr ofi le r ec om m en de d in ve st m en t ho ri zo n (y ea rs ) r et ur n 1 ye ar (% )a r et ur n 3 ye ar s (c um m ul at iv e, % )a c os ts (s yn th et ic t e r , % )a q ua lit y ra tin g 20 13 q ua lit y ra tin g 20 14 q ua lit y ra tin g 20 15 pr od uc t 1 if l if e cy cl e pr og ra m 25 10 ,2 15 ,9 2 2, 3 2, 48 02 25 2, 48 53 6 pr od uc t 10 if l if e cy cl e pr og ra m 15 1, 37 3, 43 2 2, 27 32 84 2, 31 44 52 2, 13 85 68 pr od uc t 11 if c on se rv at iv e pr og ra m 3 � 0, 49 1, 38 1, 73 2, 49 73 72 pr od uc t 12 if b al an ce d pr og ra m 5 4, 31 6, 28 2, 32 2, 74 30 97 pr od uc t 13 if d yn am ic pr og ra m 5 7, 03 28 ,6 1 0, 63 2, 64 26 56 pr od uc t 14 u l i b al an ce d pr og ra m 5 7, 31 15 ,9 4, 62 2, 88 55 15 pr od uc t 15 u l i d yn am ic pr og ra m 5 5, 43 7, 43 1, 84 2, 39 26 72 pr od uc t 16 u l i d yn am ic pr og ra m 5 14 ,1 2 17 ,9 9 2, 8 2, 20 18 2 1, 97 54 59 2, 18 03 37 pr od uc t 17 u l i l if e cy cl e pr og ra m 25 1, 73 3, 83 4, 85 2, 77 89 98 2, 60 76 32 2, 71 27 75 pr od uc t 18 u l i d yn am ic pr og ra m 5 � 0, 57 11 ,3 4 1, 9 2, 33 98 39 2, 36 10 79 pr od uc t 19 u l i d yn am ic pr og ra m 5 8, 18 5, 27 3, 86 2, 62 75 52 2, 57 19 15 2, 86 48 71 pr od uc t 2 if d yn am ic pr og ra m 5 6, 76 18 ,4 3 2, 3 2, 49 36 47 2, 30 98 5 2, 33 98 45 pr od uc t 20 u l i d yn am ic pr og ra m 5 � 0, 04 3, 98 2, 5 2, 67 16 58 pr od uc t 21 u l i d yn am ic pr og ra m 5 5, 72 12 ,9 2 2, 22 2, 88 93 28 pr od uc t 22 u l i d yn am ic pr og ra m 5 11 ,4 1 16 ,9 6 3, 06 1, 97 11 23 1, 93 38 84 2, 02 43 71 pr od uc t 23 u l i d yn am ic pr og ra m 5 13 ,5 9 � 1, 1 1, 98 2, 77 22 56 2, 76 66 22 2, 72 33 82 pr od uc t 24 u l i b al an ce d pr og ra m 5 0, 55 9, 23 2, 29 2, 21 71 17 2, 34 47 64 pr od uc t 25 u l i b al an ce d pr og ra m 5 6, 43 7, 87 2, 64 2, 83 82 16 pr od uc t 26 u l i d yn am ic pr og ra m 5 21 ,0 8 30 ,3 1, 92 2, 41 47 62 2, 56 02 45 pr od uc t 27 u l i d yn am ic pr og ra m 8 7, 74 10 ,6 4 2, 21 2, 55 51 23 2, 68 41 26 pr od uc t 28 u l i d yn am ic pr og ra m 5 15 ,4 4 29 ,5 4 2, 58 2, 74 26 54 pr od uc t 29 u l i c on se rv at iv e pr og ra m 6 1, 08 6, 91 2, 3 2, 85 69 97 2, 95 11 76 pr od uc t 3 if d yn am ic pr og ra m 5 23 ,2 8 24 ,2 7 2, 4 2, 46 18 65 pr od uc t 30 u l i d yn am ic pr og ra m 5 9, 02 14 ,6 8 3, 54 2, 86 84 03 3, 01 51 77 3, 00 76 38 pr od uc t 31 u l i d yn am ic pr og ra m 6 5, 72 8, 88 3, 32 2, 90 88 75 2, 86 30 58 2, 78 71 48 pr od uc t 32 u l i d yn am ic pr og ra m 5 7, 8 15 ,7 3 3, 85 2, 99 16 77 pr od uc t 33 u l i d yn am ic pr og ra m 8 8, 46 12 ,7 7 2, 29 3, 01 95 17 pr od uc t 34 if d yn am ic pr og ra m 8 7, 5 1, 5 2, 18 2, 53 27 91 pr od uc t 35 if d yn am ic pr og ra m 7 20 33 ,7 1, 78 2, 57 37 75 pr od uc t 36 if l if e cy cl e pr og ra m 25 9, 3 7, 6 2, 2 2, 65 87 01 pr od uc t 37 if d yn am ic pr og ra m 5 20 ,9 21 ,1 2, 16 2, 74 43 53 pr od uc t 38 if c on se rv at iv e pr og ra m 3 2, 7 5, 2 1, 37 2, 75 86 96 pr od uc t 39 if c on se rv at iv e pr og ra m 3 � 0, 4 � 0, 7 0, 75 2, 76 71 68 pr od uc t 4 if c on se rv at iv e pr og ra m 3 4, 88 1, 79 2, 05 2, 27 40 7 pr od uc t 40 if c on se rv at iv e pr og ra m 3 � 3, 4 � 1, 3 1, 39 2, 78 65 88 2, 69 61 1 pr od uc t 41 if d yn am ic pr og ra m 5 1, 9 19 ,2 0, 81 2, 81 25 32 pr od uc t 42 if d yn am ic pr og ra m 5 9, 14 16 ,9 9 2, 44 2, 86 06 14 2, 65 22 45 pr od uc t 43 if d yn am ic pr og ra m 5 2, 97 11 ,5 4 1, 99 3, 08 28 3 3, 00 99 51 pr od uc t 44 u l i d yn am ic pr og ra m 5 21 ,4 7 25 ,3 6 2, 64 2, 90 30 99 pr od uc t 45 u l i d yn am ic pr og ra m 6 3, 25 9, 5 2, 67 2, 94 03 81 2, 84 96 49 pr od uc t 46 if d yn am ic pr og ra m 5 4, 56 6, 78 1, 3 2, 42 72 88 (c on ti nu ed on ne xt pa ge ) 382 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 a pp en di x 1 (c on tin ue d) pr od uc t pr od uc t ty pe pr ofi le r ec om m en de d in ve st m en t ho ri zo n (y ea rs ) r et ur n 1 ye ar (% )a r et ur n 3 ye ar s (c um m ul at iv e, % )a c os ts (s yn th et ic t e r , % )a q ua lit y ra tin g 20 13 q ua lit y ra tin g 20 14 q ua lit y ra tin g 20 15 pr od uc t 47 if d yn am ic pr og ra m 5 16 ,5 21 ,1 1, 71 2, 67 08 91 2, 53 83 73 pr od uc t 48 if b al an ce d pr og ra m 4 1, 2 � 3, 1 1, 54 2, 85 80 9 pr od uc t 49 u l i d yn am ic pr og ra m 5 � 1, 46 12 ,3 2, 11 2, 58 69 47 pr od uc t 5 if d yn am ic pr og ra m 8 1, 11 4, 94 2, 44 2, 29 06 44 pr od uc t 50 if d yn am ic pr og ra m 7 12 ,3 7 16 1, 92 2, 54 94 12 pr od uc t 51 if l if e cy cl e pr og ra m 3 � 0, 05 2, 68 2, 51 2, 66 53 69 pr od uc t 52 if b al an ce d pr og ra m 3 3, 1 0, 8 1, 31 2, 69 29 09 pr od uc t 53 if b al an ce d pr og ra m 4 4, 7 11 ,3 2, 27 2, 72 85 09 pr od uc t 54 if c on se rv at iv e pr og ra m 3 � 0, 6 1, 3 0, 6 2, 76 68 96 pr od uc t 55 if c on se rv at iv e pr og ra m 5 � 0, 4 1, 4 1, 14 2, 89 13 1 pr od uc t 6 if l if e cy cl e pr og ra m 20 9, 91 8, 11 2, 38 2, 37 65 98 pr od uc t 7 if d yn am ic pr og ra m 5 6, 42 11 ,6 1, 77 2, 23 88 84 2, 22 98 73 pr od uc t 8 if d yn am ic pr og ra m 5 15 ,4 21 ,9 1 2, 6 2, 19 83 87 2, 14 07 63 pr od uc t 9 if b al an ce d pr og ra m 3 2, 75 3, 7 1, 22 2, 62 46 33 a a s of 20 16 . t e r � t ot al e xp en se r at io . 383j. šindelář, p. budinský / financial services review 26 (2017) 367–386 a pp en di x 2 in de pe nd en t m od el va ri ab le s– ov er vi ew y ea r 20 13 20 14 20 15 (n � 25 85 ) (n � 33 82 ) (n � 43 61 ) l ow er qu ar til e m ea n/ m ed ia n u pp er qu ar til e l ow er qu ar til e m ea n/ m ed ia n u pp er qu ar til e l ow er qu ar til e m ea n/ m ed ia n u pp er qu ar til e c o m m 30 33 .5 78 3. 8 10 13 9. 6 33 17 .8 87 33 .2 11 36 7. 3 29 13 .8 92 60 .9 12 64 4. 9 58 92 .4 65 53 .3 65 50 .2 c o m m if 13 87 .0 53 00 .6 67 03 .9 13 20 .0 52 29 .3 68 39 .8 13 46 .2 46 91 .6 57 20 .0 33 12 .0 31 78 .0 28 56 .3 c o m m u l i 38 10 .2 86 66 .1 11 20 9. 7 43 20 .0 98 13 .5 12 72 0. 1 46 53 .7 11 32 0. 5 14 99 5. 2 67 49 .7 75 68 .8 89 18 .9 pr o d if � 24 .6 % ; u l i � 75 .4 % if � 23 .5 % ; u l i � 76 .5 % if � 31 .1 % ; u l i � 68 .9 % st r u c m l m � 65 .1 % ; po ol � 27 .0 % m l m � 59 .0 % ; po ol � 33 .3 % m l m � 54 .5 % ; po ol � 38 .6 % fl at � 7. 9% fl at � 7. 7% fl at � 6. 9% si z e b ig � 92 .1 % ; m ed iu m � 7. 7% b ig � 92 .4 % ; m ed iu m � 5. 5% b ig � 93 .1 % ; m ed iu m � 0. 8% sm al l � 0. 2% sm al l � 2. 1% sm al l � 6. 1% in fo m uc h le ss th an av er ag e (� 2) � 7. 93 % m uc h le ss th an av er ag e (� 2) � 11 .9 7% m uc h le ss th an av er ag e (� 2) � 24 .0 7% l es s th an av er ag e (� 1) � 14 .7 8% l es s th an av er ag e (� 1) � 18 .9 7% l es s th an av er ag e (� 1) � 2. 20 % si m ila r to av er ag e nu m be r (0 ) � 12 .0 7% si m ila r to av er ag e nu m be r (0 ) � 20 .2 7% si m ila r to av er ag e nu m be r (0 ) � 25 .8 5% h ig he r th an av er ag e (1 ) � 55 .0 5% h ig he r th an av er ag e (1 ) � 48 .5 5% h ig he r th an av er ag e (1 ) � 47 .7 2% m uc h hi gh er th an av er ag e (2 ) � 10 .1 7% m uc h hi gh er th an av er ag e (2 ) � 0. 18 % m uc h hi gh er th an av er ag e (2 ) � 0. 11 % in d iv 1, 01 4 1, 30 0 1, 39 9 384 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 references ali, a., peranginangin, y., & walsh, m. (2015). the determinants of customer satisfaction in the financial planning industry. jassa, 19, 19–26. anagol, s., cole, s. a., & sarkar, s. (2012). understanding the incentives of commissions motivated agents: theory and evidence from the indian life insurance market. working paper. cambridge, ma: harvard business school. australian securities and investments commission. (2003). survey on the quality of financial planning advice. sydney, australia: asic australian securities and investments commission. beyer, m., de meza, d., & reyniers, d. (2013). do financial advisor commissions distort client choice? economics letters, 119, 117–119. burke, j., hung, a., clift, j., garber, s., & yoong, j. (2015). impacts of conflicts of interest in the financial services industry. working paper, no. wr-1076. santa monica, ca: rand corporation. campbell, d. t., & fiske, d. w. (1959). convergent and discriminant validation by the multitrait-multimethod matrix. psychological bulletin, 56, 81–105. chalmers, j., & reuter, j. (2015). is conflicted investment advice better than no advice? working paper, no. w18158. cambridge, ma: national bureau of economic research. cupach, w. r., & carson, j. m. (2002). the influence of compensation on product recommendations made by insurance agents. journal of business ethics, 40, 167–176, 2002. crocker, l., & algina, j. (2008). reliability and the classical true score model. introduction to classical & modern theory, 2, 114–130. danilov, a., biermann, t., kring, t., & sliwka, d. (2013). the dark side of team incentives: experimental evidence on advice quality from financial service professionals. journal of economic behavior & organization, 93, 266–272. eckardt, m., & räthke-döppner, s. (2010). the quality of insurance intermediary services: empirical evidence for germany. journal of risk and insurance, 77, 667–701. egan, m., matvos, g., & seru, a. (2016). the market for financial adviser misconduct. working paper, no. w22050. cambridge, ma: nber national bureau of economic research european insurance and occupational pensions authority. (2016). eiopa’s advice on the development of an eu single market for personal pension products (ppp). frankfurt am main, germany: eiopa publications. experti na finance. (2016). růst poradenských společností za 3q 2016 v počtu ppz. (available at http:// www.opojisteni.cz/pojistovaci-zprostredkovatele/rust-poradenskych-spolecnosti-za-3q-2016-v-poctu-ppz/) gerhardt, r., & hackethal, a. (2009). the influence of financial advisors on household portfolios: a study on private investors switching to financial advice. working paper, no. ssrn 1343607. frankfurt am main, germany: goethe university frankfurt am main. glazer, d. a. (2007). the effect of commission versus non-commission benefits on customer value: the case of life insurance policy performance. financial services review, 16, 135. gravelle, h. (1994). remunerating information providers: commissions versus fees in life insurance. journal of risk and insurance, 61, 425–457. halan, m., sane, r., & thomas, s. (2014). the case of the missing billions: estimating losses to customers due to mis-sold life insurance policies. journal of economic policy reform, 17, 285–302. inderst, r., & ottaviani, m. (2009). misselling through agents. the american economic review, 99, 883–908. insurance europe. (2014). european insurance in figures: 2014 data. brussels, belgium: insurance europe data. johnson, d., & grayson, k. (2005). cognitive and affective trust in service relationships. journal of business research, 58, 500–507. kalus, p., candela, i., & velie, ch. (2015). fund distribution in germany 2015, focus report. valencia, spain: accelerando associates international sl. li, s. (2015). the influence of financial advisers on return chasing. working paper, no. 86549. berkeley, ca: university of california – berkeley. 385j. šindelář, p. budinský / financial services review 26 (2017) 367–386 linnainmaa, j. t., melzer, b. t., previtero, a., & grace, c. (2015). costly financial advice: conflicts of interest or misguided beliefs? working paper. london: western university. lopez, t. b., hopkins, c. d., & raymond, m. a. (2006). reward preferences of salespeople: how do commissions rate? journal of personal selling & sales management, 26, 381–390. oecd. (2015). pensions at a glance 2015: ocde and g20 indicators. paris, france: oecd publishing. palazzo, g., & rethel, l. (2008). conflicts of interest in financial intermediation. journal of business ethics, 81, 193–207. popova, v. (2010). what renders financial advisors less treacherous? on commissions and reciprocity. economic research papers, no. 2010, 036. jena, germany: friedrich schiller university jena. pullins, e. b. (2001). an exploratory investigation of the relationship of sales force compensation and intrinsic motivation. industrial marketing management, 30, 403–413. r core team. (2015). r: a language and environment for statistical computing. vienna, austria: r foundation for statistical computing. (available at https://www.r-project.org/) reifner, u., neuberger, d., rissi, r., riefa, ch., knobloch, m., clerc-renaud, s., & finger, ch. (2012). study on remuneration structures of financial services intermediaries and conflicts of interest. working paper, no. markt/2012/026/h. hamburg, germany: iff institut für finanzdienstleistungen. ross jr., w. t., anderson, e., & weitz, b. (1997). performance in principal-agent dyads: the causes and consequences of perceived asymmetry of commitment to the relationship. management science, 43, 680–704. ruprecht, w. (2007). as to the difference between (unit linked) life insurance and ucits. berlin, germany: gdv gesamtverband der deutschen versicherungswirtschaft. sane, r., thomas, s., & halan, m. (2013). estimating losses to customers on account of mis-selling life insurance policies in india. working paper, no. wp-2013–007. mumbai, india: indira gandhi institute of development research. schwarcz, d., & siegelman, p. (2015). insurance agents in the 21st century: the problem of biased advice. in d. schwarcz & p. siegelman (eds.), handbook on the economics of insurance law. cheltenham, uk: edward elgar publishing. tseng, l. m. (2011). intra-role conflict and the selling decision: the case of taiwan’s life insurance industry. international journal of conflict management, 22, 373–393. tsoukatos, e., & mastrojianni, e. (2010). key determinants of service quality in retail banking. euromed journal of business, 5, 85–100. zlatá koruna. (2016). statut souteze zlata koruna 2016. (available at http://www.zlatakoruna.info/sites/default/ files/souteze/statut_souteze_zk_2016v2.pdf) 386 j. šindelář, p. budinský / financial services review 26 (2017) 367–386 is a vix etp an investment in the vix? r. parker clowersa, travis l. jonesb,* aauriemma consulting group, inc., 120 broadway, suite 3401, new york, ny 10271, usa blutgert college of business, florida gulf coast university, 10501 fgcu blvd. south, fort myers, fl 33965-6565, usa abstract this article examines vix-based etps (exchange traded products) and illustrates that both the return and risk of these products are not related to the return and risk of the vix index. the authors note that vix etps do not correlate well to the vix index. in fact, these funds are not even designed to have a high correlation to the vix index. individual investors can often mistake vix etps for an investment in the vix index itself, which is incorrect and may lead to a costly mistake. © 2016 academy of financial services. all rights reserved. jel classifications: g11; g12 keywords: vix index; etfs; etps 1. introduction the vix index measures volatility in the equity market and is a good measure of investors’ overall sentiment and level of fear in the stock market. the vix index is able to gauge market expectations of equity performance by tracking the demand for put and call options, through extracting the “price” of implied volatility in prices of s&p 500 index options. copeland and copeland (1999) even note that the vix index can be used as an indicator to rotate between large cap and small cap portfolios as well as between value and growth portfolios. there are a number of vix-related terms in the financial press that individual investors may encounter. the authors of this article use care to distinguish between these terms, which * corresponding author. tel.: �1-239-590-7167; fax: �1-239-590-7367. e-mail address: tljones@fgcu.edu financial services review 25 (2016) 73–85 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. may include: the spot “vix index,” vix-based exchange traded products (“etps” or “funds”), vix futures or options, and the s&p 500 vix short-term futures index (“s&p st index”), among others. a common mistake that is often made by individual investors looking to gain exposure to the vix index (or equity volatility in general) is to assume that vix exchange traded products (etps) accurately track the vix index. vix-based etps are not structured to track the spot vix index, but are designed to track the s&p st index or other related indices that themselves are not well correlated to the spot vix index. the purpose of this article is to examine the differences between vix-based etps and the vix index and to serve as a guide and a possible warning to individual investors about how these products trade. it is important to note that the vix index is not a tradable asset. investment products that are tradable include vix futures, vix options, and vix-based etps (see dzekounoff, 2010; jones, 2011; jones & allen, 2015; moran & dash, 2007; and others for a discussion of vix futures and options). the products that have been introduced over the years are designed to offer investors exposure to the volatility but to not mimic the performance of the vix index. the typical way that these products gain exposure to volatility is to invest in vix futures (and possibly vix options, which most etps do not use). these vix-based etps are created from complex strategies and investors must be aware of the risk involved in trading these particular assets. in an effort to diversify their portfolio, individual investors may seek investment vehicles to hedge against downside moves in the equity market. because volatility, and the vix index specifically, tends to rise when investors become more pessimistic, an investment in volatility can be a good thing, especially as a short-term hedge. thus, vix-based etps can act as a good hedge against a drop in equities, as noted by lydon and chen (2014). jones (2011) also notes that allocating a small portion of a portfolio to volatility could help investors better withstand a bear market. this research notes that a 10% allocation to nearby vix futures (used in many vix etps) can improve portfolio returns more than 15%, on average, when implemented after an increase in the vix index that most often accompanies a decline in the s&p 500. vix-based etps have become more popular, and thus new etps are being created and issued. before these products, individual investors and certain institutions were unable to gain exposure to volatility, because they may have been unable to invest in vix futures. this could be because of restrictions placed on certain institutions and individual investment accounts. these same individuals and institutions can now invest in etps that have vix futures held in trust. most vix-based etps are designed to replicate the one-day return of the s&p 500 vix short-term futures index (“s&p st index”). this underlying index is structured to provide a constant 30-day maturity futures position by rolling a long position in the first and second next to mature monthly vix futures contracts. thus, each day to provide this constant 30-day maturity, the s&p st index adds exposure in the second month to mature vix futures contract and decreases exposure in the next to mature (or nearby) contract. 2. literature review more and more academic research is being published examining vix-based products, including vix futures, options, and etps. empirical evidence presented in whaley (2013), 74 r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 who is known as the “father of the vix (see whaley, 1993),” and others show that investing in vix-based etps is not prudent for a buy-and-hold strategy. the vix futures market is frequently in contango, meaning that the price curve for futures contracts slopes upward, resulting in a decay of the price of the vix futures contracts as they near maturity. this reduction in vix futures prices, thereby reduces the prices of the etps that hold these futures. whaley (2013) found that the price curve for futures sloped upward “80% of all trading days for futures with 30-day maturities.” and that, “this erosion happens because the price of longer-dated futures contracts is almost always higher than the price of shorter-dated futures contracts.” many companies that sponsor vix-based etps note in the prospectus that these funds do not guarantee return of principle, especially if a buy-and-hold strategy is used over the long term. in fact, the prospectus from the barclays (2015) ipath etn (vxx, the largest of these etps) notes, “you (the investor) may lose some or all of your principal if you invest in the etns.” in addition, this prospectus notes, “your etn is not linked to the vix index.” whaley (2013) notes that the prospectus of velocityshares notes that, “the long term expected value of your etns is zero.” however, many individual investors may overlook these facts. jones (2011) discusses investing in the nearby vix futures contracts and shows the damaging effects on a portfolio of a buy-and-hold strategy along with how poorly vix futures track the vix index. lu, wang, and zhang (2012) examine leveraged and inverse etfs and note how, over time, these products do not accurately track the leveraged or inverse return of the benchmark they are deigned to mimic. copeland and copeland (1999) review how changes in the vix index are leading indicators of market performance. they show that, after the vix index increases, portfolios of large cap stocks outperform small caps and portfolios of value stocks outperform growth, while the reverse occurs, after the vix index decreases. following copeland and copeland, boscaljon, filbeck, and zhao (2011) examine a longer timeframe and note that the vix index can be used to time changes in rebalancing between value and growth stocks for holding periods of longer than 30 days. efremidze, dilellio, and stanley (2014) further the examination of this style rotation strategy and note that style rotations between value and growth using entropy measures appear to yield significant risk-adjusted returns. many studies also examine the performance of funds and whether these vehicles add value to investors’ portfolios. chang and krueger (2010) look at enhanced index funds, which include inverse and leveraged funds, and find that these funds, as a whole, perform worse than the pure index, with lower returns and higher risks. these authors note that investors should be wary of enhanced index funds. along these lines, dilellio, hesse, and stanley (2014) note that inverse and leveraged etfs can provide diversification benefits. annual rebalancing with a 10% allocation to an inverse stock fund reduced the coefficient of variation of terminal wealth under flat and rising returns but without an increase in the corresponding sharpe ratio. leveraged etfs were noted to provide an increased sharpe ratio but without a risk-reward benefit in terminal wealth. prather et al., (2009) examine s&p 500 mutual funds against s&p 500 etfs and note that when total annual costs of each are considered, using their model, that etfs dominate the mutual funds for inclusion in individual investors’ portfolios. dilellio and stanley (2011) examine whether etf-only strategies can outperform the s&p 500 and a buy-and-hold benchmark. they conclude that 75r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 these strategies may allow investors to capture inefficiencies in equity markets and that etfs cannot be ignored as potential instruments to enhance portfolio returns. 3. specifics of vix-based etps etps include exchange traded funds (etfs) and exchange traded notes (etns). this section discusses the similarities and differences of etfs and etns and also notes which vix-based etps are classified as etfs and which are classified as etns. as shown in the next section, none of these funds precisely track the vix index, since the assets underlying the etps are vix futures contracts. most vix-based etps are futures based, and the index they are replicating is usually the s&p st index, which is not the vix index but an index created using vix futures contracts. the vix-based etps (to replicate the s&p st index) must then constantly sell the current month futures and buy the subsequent month futures, to keep the constant 30-day maturity. because the vix futures price curve tends to be in contango (i.e., slope upward and becoming more expensive with maturity), the value of the etps will deteriorate over time, because of the purchasing of higher priced second month out contracts and selling of the current month. thus, there is usually a built in “buy high, sell low” trade embedded in these funds. one etp in our sample (vxz) replicates the return of the s&p 500 vix futures mid-term index, which is constructed in a similar manner to the s&p st index, though uses a 5-month constant maturity, rather than a constant 30-day maturity. while vix-based products are not prudent for buy-and-hold strategies, jones (2011) shows that investing in vix futures in a tactical manner could lead to positive returns, if executed correctly. vix-based investment vehicles, however, do have to be carefully monitored, and investors should determine individually if it is wise to invest in these products. thus, since vix-based etps are futures based, the investor in these products should understand, at a minimum, the relation between the etps and the vix index, upon which these investments eventually settle upon. with this said, the intent of this study is not to suggest that individual investors should avoid investing in vix-based etps entirely but to point out the characteristics of these products and caution investors and financial planners of the inherent risks, especially in the longer-term. this study examines the question of how closely vix etps really track the vix index. the next section presents an examination of eight individual etps and notes the differences between each. what follows then compares the returns of these etps against the returns of the vix index to determine if the etp returns do in fact track the returns of the vix index and to what extent. 4. vix-based etps used in this study there are a variety of etps that have been introduced over the past decade that provide investors with an opportunity to invest in volatility. this study examines the vix index returns compared to individual vix-based etp returns from october 4, 2011 through 76 r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 december 31, 2014. this time period was selected because some of the etps included in the study were not listed for trading until october 4, 2011. the eight etps we choose to examine are examined in whaley (2013). four of these etps seek to yield the daily return of a vix-based index. vxx, viix, and vixy replicate the s&p st index, and vxz replicates the s&p mid-term index. two funds (xiv and svxy) seek to yield the inverse daily return of the short-term index, and two etps (tvix and uvxy) seek to provide twice (2�) the daily return of the s&p st index. the specifics of each etp are discussed below. 4.1. vxx – ipath s&p 500 vix short-term futures etn the first vix-based etp introduced and the most popular vix etn is vxx. vxx is an exchange traded note that attempts to replicate the s&p 500 vix short-term futures index total return. the return of this etn follows the mechanics of the s&p 500 vix short-term futures index, as noted above, by rolling long position in the first and second month vix futures contracts each day. the vxx is an unsecured debt obligation of barclays bank, with tax treatment based on the capital gain for the holding period (short or long term) equal to the difference in the amount the investor receives at the time of sale and the amount paid (see barclays, 2015). 4.2. vxz – ipath s&p 500 vix mid-term futures etn vxz is much like vxx but tracks the s&p 500 vix medium-term futures total return index, which has an average settlement date of five months. the return of the ipath s&p 500 vix mid-term futures index continuously rolls long positions in the 4th, 5th, 6th and 7th next months to settle vix futures contracts. the vxz is also an unsecured debt obligation of barclays bank, with tax treatment based on the capital gain for the holding period (short or long term) equal to the difference in the amount the investor receives at the time of sale and the amount paid (see barclays, 2015). 4.3. viix – velocityshares vix short-term etn like the vxx, the viix seeks to replicate the return of the daily performance of the s&p 500 vix short-term futures index. the index was designed to provide investors with and unleveraged exposure in short-term futures contracts. the viix is an unsecured debt obligation of credit suisse ag, with tax treatment based on the capital gain (short or long term) equal to the difference in the amount the investor receives at the time of sale and the amount paid (see credit suisse, 2015). 4.4. xiv – velocityshares daily inverse vix short-term etn the xiv attempts to replicate the inverse performance of the s&p 500 vix short-term futures index and provides investors with (�1�) exposure to the short-term index. when the s&p 500 vix short-term futures index goes up, the xiv returns go down and vice versa. the xiv is also an unsecured debt obligation of credit suisse ag, with tax treatment based 77r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 on the capital gain (short or long term) equal to the difference in the amount the investor receives at the time of sale and the amount paid (see credit suisse, 2015). 4.5. tvix – velocityshares daily 2� vix short-term etn the tvix is a leveraged etn that is linked to a multiple (2�) of the daily return of the s&p 500 vix short-term futures index. the tvix tries to replicate the performance of two times the short-term index. the leveraged component makes the tvix a riskier investment than the vxx. the tvix is an unsecured debt obligation of credit suisse ag, with tax treatment based on the capital gain (short or long term) equal to the difference in the amount the investor receives at the time of sale and the amount paid (see credit suisse, 2015). 4.6. vixy – proshares vix short-term futures etf the shares comprised in the vixy etf make up what is called the “matching fund.” the matching fund attempts to match, before fees and expenses, the performance of the s&p 500 vix short-term futures index. the vixy holds vix futures in the fund and is taxed as a partnership, generating a k-1 (see proshares, 2015a). 4.7. uvxy – proshares ultra vix short-term futures etf the shares comprised in the uvxy etf make up what is called the “ultra fund.” the ultra fund attempts to achieve results, before fees and expenses, that are two times (2�) the performance of the s&p 500 vix short-term futures index each day. the uvxy holds vix futures in the fund and is taxed as a partnership, generating a k-1 (see proshares, 2015b). 4.8. svxy – proshares short vix short-term futures etf the shares comprised in the svxy etf make up what is called the “short fund.” the short fund seeks results, before fees and expenses, that correspond to the inverse (�1�) of the performance of the s&p 500 vix short-term futures index on a daily basis. the svxy holds vix futures in the fund and is taxed as a partnership, generating a k-1 (see proshares, 2015c). five of these etps are exchange traded notes (etns) and three are exchange traded funds (etfs). there are slight differences between the structures of vix etns and vix etfs that individual investors, in particular, must be aware of. etns, in general, promise to match the return of the underlying index. because they are notes, similar to bonds, the creditworthiness of etn issuers is important. etns are structured as debt instruments with a maturity date. etfs, on the other hand, pool together funds from investors and then use these pooled funds to invest in securities in an attempt to match the performance of an index. shareholders do not directly own the underlying investments in the fund, but they own shares of the fund, so indirectly own the underlying assets. 78 r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 5. is a vix etp an investment in the vix? to determine if vix etps truly replicate an investment in the vix index, and if so how much, this section compares the daily returns of each etp with the returns of the vix index. because the vix index is not tradable, it is important to distinguish whether or not an investment in a vix-based etp is a true substitute for an investment in the index. many investors may assume they are gaining highly correlated exposure to the vix index when purchasing a vix etp, when they may not. as mentioned previously, our study analyzes data from october 4, 2011 through december 31, 2014. fig. 1 shows the vix index, over the period of analysis, compared with the unlevered etps (vxx, vxz, viix, and vixy). one can see that none of these etps track the vix index very well over time. all three etps that follow the s&p st index (vxx, viix, and vixy) are very highly correlated with each other. in fact, the three series are not distinguishable from each other. because of this, the return path of vxx, viix, and vixy are combined for clarity. these funds begin at a base of 1.00 in october 2011 and end at a value of 0.04 in december 2014, while the vix index goes from 1.00 to 0.47, during the period. thus, a buy-and-hold strategy in these funds would be devastating to an investor. vxz, the fund that follows the midterm index, fares a bit better than the short-term etps, going from 1.00 to 0.18 over the period, but still ended no where near the vix index. in addition, one can see that the volatility of none of these funds matches that of the vix index. it is important to point out that there are periods where these etps had positive returns. as shown in fig. 1, a buy-and-hold strategy is not what an individual investor would want to pursue. however, the periods of late-2011, april through may 2012, december 2012, and other short-term periods where the vix index spikes would have proved very profitable if an investor chose to purchase one of these etps and sell it before its decline. thus, for the short-term, these products can enhance portfolio returns. examination of fig. 2, shows the two inverse etps (xiv and svxy) versus the vix index. one can see that these two funds performed very well over the period, going from 1.00 to 5.70 (xiv) and 5.81 (svxy), while the vix index goes from 1.00 to 0.47. this performance is in large part because of the contango effect of vix futures, mentioned previously, and the fact that these funds are inverse, effectively shorting the s&p st index. fig. 1. vix index versus short-term and midterm etps from october 2011 through december 2014. 79r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 even though the returns of these etps were very high (around 71% compounded annual return), individual investors should use extreme caution when looking to invest in inverse vix etps, because of their high volatility. as shown in fig. 2, the vix index did decline over this period, contributing to the increase in these funds. however, the vix index does increase at times, which causes these inverse funds to decline significantly. thus, a buyand-hold strategy, would have proved very profitable for an investor, over the period examined. however, there were periods where buying these etps would not have worked in an investor’s favor. for example, if an investor purchased either of these inverse funds in april 2014, they would still have a loss as of december 2014 and would have experienced tremendous volatility in between. thus, individual investors should know this before investing in these products. fig. 3 presents the two twice-levered (2�) etps (tvix and uvxy) against the vix index. as one would expect, given the performance of the 1� funds, these levered etps performed even worse, going from a base of 1.00 to 0.0003 over the period. as is also evident from fig. 3, these funds do not follow the vix index either. thus, an individual investor may choose to invest in one of these twice-levered funds but should definitely not do so with the intention of holding the fund over a long period of time. these funds can be fig. 2. vix index versus inverse etps from october 2011 through december 2014. fig. 3. vix index versus twice-levered etps from october 2011 through december 2014. 80 r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 bought and sold over a short period to take advantage of an increase in volatility, but investors should know the risk—that these funds do not track the vix index and have a strong downward tendency, because of the contango effect of vix futures that is magnified with their leverage. table 1 provides statistics of the vix index versus the eight etps over the examination period. panel a considers all daily data; panel b looks at only the days the vix index goes up, and panel c only the days when the index goes down. table 1, panel a, shows that the vix index had a mean daily return of 0.15%. even though the vix index had a positive mean return, all of the vix etps, except the inverse funds, experienced negative daily mean returns. the mean daily returns of vxx, vxz, viix, and vixy are not statistically significantly different from that of the vix index, while the mean daily returns of the twice-leveraged etps (tvix and uvxy) are significantly different at the 1% level. most interestingly is that the mean daily returns of the inverse etps (xiv and svxy) are not statistically different from that of the vix index. these two etps are the only ones with positive mean daily returns, like that of the vix index. the differences in the index returns and the returns of the etps demonstrate the effect of contango that that these etps experience and further illustrate why these etps are not suitable for a buy-and-hold strategy. in addition to the differences in mean daily returns, panel a also shows the median returns of the index and etps. the vix index had a -0.35% median daily return over the period, while the three short-term etps (vxx, viix, and vixy) had median returns of �0.52%, �0.54%, and �0.52%, respectively. the returns of these etps are correlated, but only about 88% correlated to the vix index. the two inverse etps (xiv and svxy), while replicating the inverse returns of the vix short-term index had median returns of 0.51% and 0.54%, table 1 daily return statistics of vix index vs. etps vix index vxx vxz viix xiv tvix vixy uvxy svxy panel a: all data mean 0.15% �0.33% �0.19% �0.33% 0.29% �0.74%** �0.33% �0.70%** 0.29% median �0.35% �0.52% �0.23% �0.54% 0.51% �1.08% �0.52% �1.08% 0.54% sd 7.04% 3.81% 1.85% 3.83% 3.80% 7.00% 3.82% 7.60% 3.82% skewness 1.02 0.50 0.24 0.49 �0.51 0.35 0.50 0.48 �0.51 correlation n/a 88.41% 76.99% 88.56% �87.89% 85.28% 88.55% 88.42% �87.93% sharpe ratio n/a �0.88 �1.17 �0.88 0.97 �0.69 �0.88 �0.63 0.98 tracking error n/a 4.11% 5.75% 4.09% 10.54% 3.91% 4.09% 3.67% 10.56% panel b: days when vix was up mean 5.60% 2.35% 0.97% 2.36% �2.39% 3.95% 2.36% 4.63% �2.39% median 3.85% 1.52% 0.72% 1.52% �1.65% 2.54% 1.57% 3.05% �1.62% sd 5.91% 3.25% 1.59% 3.27% 3.25% 6.15% 3.26% 6.49% 3.27% % both up n/a 78.99% 72.61% 77.66% 22.07% 78.99% 78.99% 77.39% 22.61% panel c: days when vix was down mean �4.60% �2.65% �1.19% �2.66% 2.60% �4.81% �2.66% �5.32% 2.61% median �3.59% �2.16% �1.07% �2.13% 2.11% �3.80% �2.15% �4.25% 2.11% sd 3.82% 2.56% 1.43% 2.58% 2.55% 4.86% 2.56% 5.10% 2.56% % both down n/a 90.28% 81.48% 90.05% 10.88% 90.05% 90.28% 90.05% 11.11% notes. data from october 4, 2011 through december 31, 2014. all statistics are daily, except sharpe ratio, which uses annualized data. vxx, viix, and vixy replicate the s&p 500 vix short-term futures index. vxz replicates the s&p 500 vix mid-term futures index. xiv and svxy replicate the inverse, while tvix and uvxy replicate 2� of the s&p 500 vix short-term futures index. double asterisks indicate statistical significance in mean from vix index at the 1% level. 81r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 respectively. the two twice-levered etps (tvix and uvxy) both posted median returns of �1.08% (or median returns of �0.54% if they were unlevered), while the midterm etn (vxz) had a median return of �0.23%. individual investors must be aware that over time, the value of these vix etps tends to significantly decline or decay. as one might imagine, from examining the differences in the mean and median daily returns for the vix index and each etp, all of the non-inverse return series are positively skewed, with the inverse etps having negatively skewed return distribution. for comparison, the daily return distribution of the s&p 500, over the period examined, has a skewness of 0.04. the vix index has a skewness of 1.02, with the etps that follow the s&p st index (vxx, viix, and vixy) having a skewness of about 0.50, with the inverse etps having about the same amount of negative skewness. the midterm etp has the lowest skewness of 0.24, while the twice-levered products show a high degree of skewness, as well. table 1, panel a also presents the sharpe ratios of each of the eight etps, as described in sharpe (1994). as one might guess, the etps that track the s&p st index (vxx, viix, vixy, tvix, uvxy), along with vxz, have negative sharpe ratios. it is interesting to note that the vxz, the s&p 500 vix medium-term index etp, has a larger negative sharpe ratio than the s&p st index etps. this is because of the fact that this index has more of the negative returns with less of the volatility when compared to the s&p st index. the sharpe ratios of the two inverse etps, xiv and svxy, are positive at 0.97 and 0.98, respectively. these are both less than the sharpe ratio of the s&p 500, over this period, of 1.25. the tracking error of each vix etp is also noted in table 1. this statistic measures which etps does a better job of tracking the vix index. as one might expect, the two inverse etps, xiv and svxy, have the highest tracking errors, 10.54% and 10.56%, respectively, indicating that these etps demonstrate more volatility around the vix index. even though these tracking errors are highest, these etps have a mean daily return that is not statistically different from that of the vix index. the leveraged etps, tvix and uvxy, resulted in the lowest tracking errors, 3.91% and 3.67%, respectively, signifying that these etps do a slightly better job of tracking the vix index, relative to the other etps that were observed. the remaining s&p st index etps (vxx, viix, and vixy) have tracking errors that range from 4.09% to 4.11%, with the vxz having a tracking error of 5.75%. the high tracking errors exemplify how the vix etps deviate from vix index significantly. table 1, panel b presents the statistics of the vix etps relative to the vix index on the days in which the vix index had a positive return. when looking at the days in which the vix index experienced a positive return, it averaged 5.60% daily, with a 3.85% median return. on the days the vix index moved up, many of the vix etps moved in the same direction, however none had average daily returns or median returns as high as the vix index. for example, vxx experienced an average return of 2.35%, compared to the 5.60% for the vix index. in addition, vxz, the midterm etn, only generated an average return of 0.97% on the days in which the vix index had a positive return. the non-inverse etps that track the s&p st index (vxx, viix, tvix, vixy, and uvxy) only had positive returns 77% to 79% of days in which the vix index had a positive return. table 1, panel c reports the statistics the vix etps in relation to the vix index on days in which the vix index experienced a negative return. these findings are telling and exemplify the issues related to investing in vix etps. on the days in which the vix index 82 r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 had a negative return, it averaged a daily return of �4.60%. for the days in which the vix index had a negative return, only the twice-levered etps (tvix and uvxy) experienced a greater downward average daily return than the vix index of �4.81% and �5.32%, respectively. not surprisingly, during these days, the inverse etps (xiv and svxy), on average, moved in a positive direction. however, there were still trading days in which the inverse vix etps moved in the same negative direction as the vix. the non-inverse funds that follow the s&p st index (vxx, viix, tvix, vixy, and uvxy) were more highly likely to follow the vix index when it was down; all were down over 90% of the days when the vix index was down. thus, panel c illustrates that while, on average, many of these funds did not post as large of a negative return as the vix index, when coupled with the negative average daily returns of these etps when the index went up, the overall performance was much worse than that of the index. this result is also shown in figs. 1 and 3 (excluding the inverse funds). analyzing the returns in the vix-based etps versus the returns in the vix index is noteworthy. one might expect these vix-based etps to act as fair substitutes for investing in the vix index, but, as noted above, these funds do not accurately track the vix index over time. even though these investments do not track the vix index well over time, they still generally move with volatility (but are more correlated with the appropriate s&p 500 vix futures index). thus, investors can use these etps to benefit from changes in volatility but should only invest for a short period. because these funds are not prudent for a buy-and-hold strategy, timing in these investments is critical. investors must consider the benefits, costs, and challenges of investing in vix etps and must also realize that the vix etps do not always move in the same direction as the vix. 6. conclusions this article examines eight vix-based etps that include exchange traded notes etns and etfs and compares the performance and returns of these investment vehicles to that of the vix index. many investors may expect these etps to track the vix index, which this article shows that they do not. in addition, many investors may expect that these funds are designed to replicate the returns of the vix index, which they are not. these funds are structured to replicate the performance of any number of indexes comprised of vix futures. this article looks at etps designed to replicate the performance of two different indexes, the s&p 500 vix short-term futures index and the s&p 500 vix midterm futures index. in addition, etps that track the inverse and twice the performance of the s&p 500 vix short-term futures index are examined. the results of this analysis are crucial for individual investors looking to invest in vix-based etps (or any other vix-based products). investors should realize that the returns of etps, while providing exposure to volatility, do not track the vix index and are exposed to a decline in value, because of the large degree of contango priced into the vix futures contracts used by these funds. thus, vix etps are not generally suitable for a buy-and-hold investment, but may provide return enhancement, if used judiciously and carefully monitored in a portfolio. as noted by dilellio et al. (2014), using inverse and leveraged etfs must be 83r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 considered under the appropriate risk-return tradeoff. using the same assessment, financial planners and individual investors may chose to use vix-based etfs for return enhancement, but each must assess the risk and return benefits of these products. vix-based investment products are still relatively in their infancy, especially etps, and research into these products is just beginning. there are a number of questions that can be examined within the vix-based investments space. this article adds to that literature and provides a general overview and caveat to individual investors of how vix etps behave and some of the factors that must be considered before investing in these products. there is much more research to be done in this area and many more ways to examine using these types of products within an investment strategy, just not from a buy-and-hold perspective. references barclays. (2015). pricing supplement to the prospectus and the prospectus supplement for ipath s&p 500 short-term and mid-term futures etn. barclays ipath (available from http://www.ipathetn.com/us/16/en/ contentstore.app?id�5149530). boscaljon, b., filbeck, g., & zhao, xin. (2011). market timing using the vix for style rotation. financial services review, 20, 35–44. chang, c. e., & krueger, t. (2010). do enhanced index funds live up to their name? financial services review, 19, 145–162. copeland, m. m., & copeland, t. e. (1999). market timing: style and size rotation using the vix. financial analysts journal, 55, 73–81. dilellio, j. a., hesse, r., & stanley, d. j. (2014). portfolio performance with inverse and leveraged etfs. financial services review, 23, 123–149. dilellio, j. a., & stanley, d. j. (2011). etf trading strategies to enhance client wealth maximization. financial services review, 20, 145–163. dzekounoff, d. (2010). understanding vix futures and options. futures magazine, september. efremidze, l., dilellio, j. a., & stanley, d. j. (2014). using vix entropy indicators for style rotation timing. journal of investing, 23, 130–143. jones, t. l. (2011). a look at the use of vix futures in investment portfolios: buy-and-hold versus tactical allocations. journal of trading, 6, 22–29. jones, t. l., & allen, m. t. (2015). a note on the premiums and discounts embedded in vix futures prices. journal of investing, 24, 69–73. lu, lei, wang, j., & zhang, g. (2012). long term performance of leveraged etfs. financial services review, 21, 63–80. lydon, t., & chen, m. (2014). use vix etfs to hedge against market swings. etf trends (available from http://www.etftrends.com/2014/06/use-vix-etfs-to-hedge-against-market-swings/). moran, m. t., & dash, s. (2007). vix futures and options: pricing and using volatility products to manage downside risk and improve efficiency in equity portfolios. journal of trading, 2, 96–105. prather, l. j., chu, t., maxumder, i., & topuz, j. c. (2009). index funds or etfs: the case of the s&p 500 for individual investors. financial services review, 18, 213–230. proshares. (2105a). fact sheet – proshares vix short-term futures eft. proshares (available from http:// www.proshares.com/media/fact_sheet/prosharesfactsheetvixy.pdf). proshares. (2105b). fact sheet – proshares ultra vix short-term futures eft. proshares (available from http://www.proshares.com/media/fact_sheet/prosharesfactsheetuvxy.pdf). proshares. (2105c). fact sheet – proshares short vix short-term futures eft. proshares (available from http://www.proshares.com/media/fact_sheet/prosharesfactsheetsvxy.pdf). 84 r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 sharpe, w. f. (1994). the sharpe ratio. journal of portfolio management, 21, 49–59. (available from http:// app.velocitysharesetns.com/files/prospectus/pricing_supplement_no__vls_etn-1_a36_long_form_ 1.pdf). whaley, r. e. (1993). derivatives on market volatility: hedging tools long overdue. journal of derivatives, 1, 71–84. whaley, r. e. (2013). trading volatility: at what cost? journal of portfolio management, 40, 95–108. 85r. parker clowers, t.l. jones / financial services review 25 (2016) 73–85 financial services review, 32(2) 1 unveiling the winning contribution patterns for enhanced financial health chuck grace,1 adam metzler,2 yang miao,3 longlong feng,4 and alireza fazeli5 abstract by middle age many investors have accumulated nest eggs of comparable size to their annual salary. in this case an additional percentage point of returns has the same mathematical impact on their wealth as a percentage point of savings rate. however, while savings rate falls under the direct control of the investors, investment returns only very weakly so. what is the experience of actual investing canadians in facing this circumstance? studying this topic is important because, despite extensive attention to the topic, savings rates in canada have been slowly eroding for over 20-years. we examine how accumulation patterns impact investor outcomes by examining their investment transactions, over a three-year period ending in august 2022, using advanced data analytics in the form of machine learning to explore previously unknown patterns. this paper gives the resounding answer that investors are overwhelmingly likely to be better served by a focus on savings rather than on returns. we conclude that a consistent pattern of saving is a ‘winning’ strategy for wealth accumulation. saving patterns were by far the most powerful determinant of lifetime utility. the simple act of opening an account and automated regular contributions is the most powerful technique that investors, policy makers, asset managers and advisors can deploy in the pursuit of wealth accumulation. despite this conclusion we observed that most of the investors in this study did not appear to follow a strong saving strategy. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation grace, c., metzler, a., miao, y., feng, l., & fazeli, a. (2024). unveiling the winning contribution patterns for enhanced financial health. financial services review, 32(2), 1-28. introduction by middle age many investors have accumulated nest eggs of comparable size to their annual salary, and many turn their attention to maximizing returns. at this point in their lives, an additional percentage point of returns has the same mathematical impact on their wealth as a 1 corresponding author (fwl@uwo.ca). western university, london, canada 2 wilfrid laurier university, waterloo, canada 3 western university, london, canada 4 western university, london, canada 5 western university, london, canada percentage point of savings rate. what is the experience of actual investing canadians in facing this circumstance? while savings rates fall under the direct control of the investors, investment returns remain at the mercy of unpredictable markets. this paper gives the resounding answer that investors are https://creativecommons.org/licenses/by-nc/4.0/ mailto:fwl@uwo.ca https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(2) 2 overwhelmingly likely to be better served by a focus on savings than on returns. the math behind savings and its economic impact have been well studied. the topic has also been explored in depth by well-known behaviouralists such as kahneman and thaler (2006). it has been extensively mentioned as a contributing factor in the context of financial wellness. and enormous resources have been dedicated to growing savings through investment strategies. but there is little empirical study of real-world savings transactions and how those transaction patterns impact wealth accumulation – especially in a canadian context. we examine unique dataset(s) of investor transactions to examine the relationship between investor behaviours, investment strategies, household savings, and investment outcomes. ultimately our goal was to determine whether investment returns or savings rates drove wealth accumulation for the investors in our datasets. we examine these real-world observed behaviours through advanced data analytics in the form of unsupervised machine learning. we examine trading over a 3-year period ending august 2022, providing us with the opportunity to observe patterns during rising markets, declining markets and the turbulent phases during transitions. the data encompasses hundreds of unique investment strategies including advised and ‘do-it-yourself’ portfolios. the algorithms determined that investors could be clustered into one of three groups (for each dataset) that were determined by saving’s behavior, investment returns and portfolio outcomes. a brief description of the resulting clusters is noted below in table 1. table 1. summary of the clusters derived by the unsupervised machine learning algorithms. dataset 1 dataset 2 cluster 1: a group of investors (13.8% of accounts) whose wealth followed a downward trajectory starting with the market correction in february 2020, and continuing the same trajectory thereafter. cluster 4: a group of investors (5%) whose wealth followed a downward trajectory following major withdrawals from their accounts cluster 2: a group of investors (44.1%) whose wealth generally followed the markets – rising and falling in sync with general market trends. cluster 5: a group of investors (34% of the dataset) whose wealth generally followed the markets – rising and falling with the markets. cluster 3: a group of investors (42.1% of the dataset) whose wealth had an upward trajectory (growing) throughout the period. cluster 6: a group of investors (61%) whose wealth had an upward trajectory (growing) throughout the period. it should be noted that the observed period encompassed much of the covid-19 pandemic. in canada, the federal government introduced the canada emergency response benefit (cerb) in march 2020 before transitioning to the employment insurance program and ending in may 2023. the cerb program provided financial support to employed and self-employed canadians directly affected by covid-19. the $500/week payments, in addition to the savings derived from working-at-home, coincided with a significant increase in household savings (see figure 4). as a result, our study provided a unique opportunity to observe the impact of unusual cash inflows on household resiliency – as measured by wealth. it is also worth noting that the examined period encompasses unusual market conditions marked by negative returns on fixed income investments and mixed returns on equities (see appendix 1). we conclude that a consistent pattern of saving – even in turbulent markets is a ‘winning’ strategy for wealth accumulation. saving patterns were by far the most powerful determinant of wealth accumulation. investment performance played a role in wealth accumulation, but it was muted when compared to savings behaviour. grace et al. 3 we note that systematic saving on a regular or automated schedule enhanced the outcomes. as did saving more frequently – for example, biweekly as opposed to quarterly or irregularly. the observed results were consistent with stochastic simulations implying that ‘the math’ works. we note that ‘keep it simple’ by automating savings can be an effective strategy for wealth accumulation that cuts through the noise and confusion to create a tangible impact on financial wellness. across all the clusters, demographics such as age, gender, geography, risk tolerance and income were statistically immaterial in predicting outcomes with respect to wealth accumulation. we observe that advised investors had higher savings rates and lower withdrawal rates than the diy investors although the size of the diy dataset is significantly smaller and the investors significantly younger. we observed that 56% of the observed investors in this study did not appear to optimize their savings behaviour over the period studied as compared to the active savers. despite the extensive study by multiple disciplines, savings rates in canada are not improving. canadian savings rates have been slowly eroding for over 20 years. they trended up during the 2019 pandemic but have subsequently reverted to pre-pandemic levels levels described by some policy makers as dangerously low. canadian savings rates are currently “middle of the pack” among g20 countries and forecasted to be the lowest amongst the g20 by 2025. in 2023, canadian households are preoccupied with inflation and the impact of rising interest rates, putting pressure on household budgets and the potential for savings rates to decline further. on average, spending is outstripping incomes, household debt levels are rising, and a significant percentage of canadians are worried about having sufficient retirement savings. perhaps the trends noted above could be reversed if more canadians could be encouraged to follow simple, automated savings plans? literature review and background context the topic of ‘saving’ has been explored by several disciplines. over the years, the topic has proven to be important to policy makers, economists, portfolio managers, actuaries, and behavioural scientists. the topic is also important to the financial services industry who look to ‘household saving’ to fuel a plethora of investment products and services. more recently it has also been viewed as intrinsic to the definition of financial wellness. quantitative modelling: savings or returns both empirical evidence and the quantitative modelling sketched below, and expanded in appendix 2, suggest that by middle age investors are likely to accumulate a large enough capital base that their investment decisions are as impacted by return considerations as by considerations of savings rate. it is at this stage, when investors are in their early 40s, that many many people decide to move their assets to a more full-service investment manager, like investors in our dataset. however, savings rates fall under the direct control of these investors while investment returns only weakly so. according to a standard and simple discrete time model in which an investor begins at t = 0 with assets v0, invests a constant fraction f of a constant income x each time period, and invests at a constant return rate μ reinvesting all proceeds, it can be shown that to close approximation the growth in a portfolio over a small k(= 2 or 3) years is given by: vn+k – vn = fkx + μkvn + ½ μ2k(k-1)vn + ½ fμk(k-1)x the impact of μ will be small compared to the impact of f until such time as the portfolio grows to be about twice the income of the investor. this occurs at about the time ln(2)/μ which is 14 for μ = 5%. however, this mathematical sensitivity analysis does not consider the fact that the savings rates are more completely under the control of an investor than the return rate. the reality is that it is hard to move μ very much, and (in contrast to our simple model here), μ is random. the goal of this paper is to see if there are natural groupings of investors – some of whom appear to be working on enhancing μ through market timing, and others concentrating more on savings, financial services review, 32(2) 4 and to compare them over a real time-period to see which group can generate more retirement savings. life cycle hypothesis our research touches on the theoretical life cycle hypothesis framework proposed by modigliani et al. in 1954. the scholars proposed that an optimizing behaviour implies a smooth consumption and lifetime utility where individuals accumulate wealth through savings during their working years. the hypothesis posits that individuals will transition to retirement at a lifetime peak in income and wealth. our research contributes to the research in this area by examining actual individual portfolios as they approach what should be the peak in their consumption curve. financial wellness and savings recently, savings have been linked to the concept of financial wellness (vlaev et al., 2014, kempson et al., 2017, suh 2021, metzler 2021). in their ground-breaking research, kempson et al. (2017) identified three key behaviours that define financial wellness – spending restraint, active saving and borrowing for daily expenses. in previous research, the authors of this paper (metzler et al., 2021) concluded that savings, spending, and debt play uniquely powerful roles in financial resilience. the authors determined that of the 200+ variables used in the clustering, three of the top nine were related to savings. a financially resilient individual can withstand financial setbacks such as sudden loss of income or unanticipated expenses. low levels of financial resilience are a strong predictor of financial stress and can lead to more serious health problems6 . the events of 2020 and the economic impact of covid-19 gave increased urgency to the topic of financial resilience. the financial cost on all levels of government for financial countermeasures to covid-19 are becoming apparent and a stronger understanding of the prevalence and type of financial fragility in canadian households will allow more targeted policy interventions. there is clear value in helping financially stressed individuals understand the root of their financial challenges and then providing them with advice (tailored to their specific circumstances) on the steps they may be able to take to change their circumstances and alleviate their financial stress. savings interventions numerous incentives have been explored by governments and industry to enhance savings. they include policy interventions. policy makers and government agencies have all explored household savings rates as a driver of ‘healthy’ economies and ‘healthy’ households (fcac, 2021; gale et al., 2005, justera et al., 1999, baldwin, 2022). governments around the world regularly incentivize households to ‘save more’ – often with a focus on pensions and retirement. in canada, retirement savings plans, tax free savings accounts and registered education savings plans are popular examples of government sponsored programs with assets under administration measured in the trillions of dollars7. policy makers will also point to savings rates when exploring topics such as poverty and interventions for disadvantaged or vulnerable groups (cruz et al., 2016, hall, 2021). demographic drivers in our analysis we include specific demographic features (see table 2 below) in order to observe wealth accumulation vis a vis potential demographic drivers. researchers have linked savings rates and resiliency to several demographic factors including age (maynard et al., 2008, baldwin, 2022), income (dynan et al., 2004, turner & luea, 2009, macgee 2022, cruz 2016), household composition (cobb-clark et al., 2016), and gender (fisher, 2010). these factors are often combined under ‘life cycle model’ (feiveson et al., 2019). however, metzler et al. (2021) noted that while these demographic traits can be linked to financial resilience, the data does not support a causal relationship. 6 manulife, 2016 financial wellness index 7 statistics canada, www150.statcan.gc.ca/, table 1110-0016-01, released 2020-12-22, sourced june 2023 grace et al. 5 table 2. select features used in the clustering algorithms description features (examples) demographic features general demographic information age, income, gender, marital status, residency know your client information as prescribed by regulations investment knowledge, net worth, risk tolerance, investment horizon behavioural features derived features that can be used as proxies for investor behaviour risk tolerance, automatic versus jit trades (habit), portfolio churn, trading frequency financial and transactional features account detail: the account holding a portfolio on investments – for example rsps or tfsa account type including rsp, tfsa, resp etc. portfolio detail: a basket of securities or holdings security id, units, book value, market value holdings: an individual security security id, risk type, trading exchange transactions: a transaction that changes the book or market value of the portfolio type of transaction, date, units, security, gross, net, currency any fees or commissions derived from a transaction type of fee, date, units, dollar amount, currency bookkeeping: the dealer’s accounting or administrative view of the elements above engineered features/ratios weekly market values, min/max scaling, trading sequences, contributions/withdrawals as % opening balance, contributions/withdrawals as % on income, trades per account, trades per month, internal rate of return behavioural interventions in our analysis we include behavioural features (table 2) in order to observe wealth accumulation vis a vis potential behavioural drivers. kahneman (2012) and thaler et al. (2004) are widely known for theorizing that behavioural attributes drive savings success and that the concept of ‘nudging’ can be used to influence savings decisions. goal setting (soman et al., 2011), mental accounting (shefrin et al., 2004), future self (hershfield et al., 2011, cheema et al., 2011), and risk aversion (cagetti, 2003) have all been linked to savings behaviour. dholakia (2016), in turn, noted that it may be more useful to focus on habits or traits, rather than behaviour, when attempting to predict sustainable saving activities. further research (macinnis et al., 2009, hall, 2021) has noted that interventions meant to nudge decision-makers 8 russell investments (www.russellinvestments.com), sourced june 29, 2023 should be tailored to individual differences and the social forces that impact particular social groups. newmeyer (2020) notes that the benefits of automated savings accrue at a higher rate for individuals with lower incomes and that this benefit depends on the presence of a personal savings orientation (dholakia et al., 2016). the impact of advice in this paper, we examine both advised and ‘doit-yourself’ investors to observe wealth accumulation vis a vis the influence of advice. the role of a financial advisor with respect to household consumption and utility has not been widely researched but there is emerging research that advisors enhance savings behaviour. industry studies (russell8, vanguard9) estimate advisors add 150 to 200 basis points (bps) to portfolio 9 vanguard (www.vanguard.com/pdf/isgqvaa.pdf), sourced june 29, 2023 https://www.tandfonline.com/author/cagetti%2c+marco http://www.russellinvestments.com/ http://www.vanguard.com/pdf/isgqvaa.pdf financial services review, 32(2) 6 growth through coaching and investor discipline. foerster et al. (2017) and linnainmaa (2020) measured advisor value and identified a significantly positive relationship when adding an automatic savings plan and that non-advised investors did not take advantage of automated savings plans. researchers at cirano (montmarquette et al., 2016) determined that investors who work with advisors benefited from greater savings. investment strategies and wealth accumulation in our analysis we include several financial features (table 2) to observe wealth accumulation vis a vis potential investment strategy drivers. in finance and actuarial sciences, researchers have tended to focus on investment risk and return as the primary drivers for wealth accumulation. established techniques such as diversification (markowitz, 1991), asset pricing (merton, 1973), lifetime ruin (bayraktar, 2010), portfolio optimization (markowitz, 2010) and target driven portfolios (blake et al., 2013) are all focused on maximizing returns while minimizing risk – once a basket of savings has been accumulated. canadian investors generally follow one of two trading strategies with their registered retirement savings plans (the focus on this paper) – either active or passive. active trading refers to the periodic trading in specific securities, typically to deliver alpha (unusual returns) or minimize risk (volatility). the antithesis of an active trading strategy would be a passive trading strategy where investors largely ‘buy and hold’ investments for the duration of their investment horizon. do-it-yourself (diy) investors can trade as often as they wish while advised investors are presumably influenced by their advisor’s recommendations and availability. we have not included annual portfolio rebalancing under the definition of active trading as it represents a realignment to the investor’s risk tolerance rather than an attempt to ‘time the market’ or generate alpha. the bottom line despite the interventions noted above, savings trajectories in canada appear to be moving in the wrong direction. in canada, savings rates trended up during the 2019 pandemic but have subsequently reverted to pre-pandemic levels levels described by some policy makers as dangerously low (baldwin, 2022, macgee, 2022) (see figure 1 and 2). as well, while incomes in canada appear to peak in the 45 to 54 age group, savings rates peak in the 35 to 44 age group – well before the peak predicted in the life cycle hypothesis (see figure 3). an empirical study could provide valuable insights into the impact of savings rates, as well as elucidate strategies individuals could adopt to improve savings. grace et al. 7 figure 1. household savings rates as percentage of disposable income10 figure 2. canadian household savings rates as percentage of disposable income11 10 oecd (2023), saving rate (indicator). doi: 10.1787/ff2e64d4-en (accessed on 29 june 2023) 11 https://doi.org/10.25318/3610011201-eng, sourced june 29, 2023 -2.0% 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 canadian household savings rates  https:/doi.org/10.25318/3610011201-eng financial services review, 32(2) 8 figure 3. canadian household savings rates as a percentage of disposable income, by age group (2018/22)12 data description and analytical methodology data/sample the datasets include anonymized data on investors under two distinct circumstances. • dataset 1 encompasses 7,400 investors who work closely with financial advisors. • dataset 2 encompasses 477 investors who do not work with financial advisors – sometimes referred to as diy investors. we examined savings behaviours for the period august 2019 to august 2022, providing us with the opportunity to observe behaviour during rising markets, declining markets and the turbulent phases that transition the two. both datasets encompassed registered retirement savings plans (rsps) to help control for risk tolerance, time horizon and structural constraints (see appendix 3). the datasets included data points down to the daily transaction level. dataset 1 was provided by a registered investment dealer that has provided investment products to canadian retail investors for over 30 12 statistics canada table: 36-10-0587-01 (formerly cansim 378-0152), sourced january 2024 years. the dealer hitherto has approximately 300 advisors who work with approximately 23,000 clients across canada, with over $10 billion canadian dollars (cad) in assets under administration. clients typically have multiple accounts each with different purposes. for example, a client may have accounts for: (i) retirement savings; (ii) children’s education savings; and (iii) other savings. the data are comprised of 7,400 rsp accounts with associated know your client (kyc) information, trade, and transaction details from 2 august 2019 to 5 august 2022 (see table 2). the dealer provides a variety of financial products and services designed to support independent advisors. furthermore, the dealer’s focus is to provide positive outcomes to clients and advisors, and not to push certain financial products. dataset 1 investors work with a registered investment representative or financial advisor. more specifically, the advisors work under an investment dealer governed by the investment industry regulatory organization of canada -30% -20% -10% 0% 10% 20% 30% -$40,000 -$20,000 $0 $20,000 $40,000 $60,000 $80,000 $100,000 $120,000 $140,000 under 35 35 to 44 45 to 54 55 to 64 65 and over cdn household income and saving (2018-22) income savings saving rate https://doi.org/10.25318/3610011201-eng grace et al. 9 (iiroc) 13 . under the iiroc regime, advisors provide a broad range of services and can recommend investment solutions from thousands of investment choices14. investment dealers are obligated to assess their client’s risk tolerance when onboarding. the assessments generally take the form of questionnaire that gathers information on the client (know your client or kyc) and scores the risk tolerance. an effective kyc protocol collects two types of information: (1) objective demographic data (e.g., identity), and (2) subjective information on the client’s investment needs, financial objectives, investment knowledge, appetite for risk and other financial circumstances. in previous research, researchers (e.g., thompson et al., 2021) noted that advisors are diligent at ensuring recommended portfolios match the client’s stated risk tolerance. this determination allowed us to control for risk tolerance in our analysis. dataset 2 was provided by a registered investment dealer that has provided investment products to canadian retail investors for over 9 years. the dealer operates under what is known as a “robo advisor” model where investors open and trade on an account online with minimal advice or service from the dealer. the dealer has approximately 12,000 clients across canada, with over $.79 billion canadian dollars (cad) in assets under administration. clients typically have multiple accounts each with different purposes. for example, a client may have accounts for: (i) retirement savings; (ii) children’s education savings; and (iii) other savings. the data are comprised of 477 rsp accounts with associated kyc information, trade, and transaction details from 2 august 2019 to 5 august 2022. dataset 2 represented younger investors with significantly smaller opening portfolio balances. both originating datasets were edited by the data donors prior to our receipt to ensure all client identifiers were anonymized consistent with canada’s personal information protection and electronic documents act (pipeda) and standard research ethics protocols. even after anonymization practices, there is the possibility that clients could be identified using machine learning algorithms (rocher et al., 2019). therefore, no individuals will be identified or referenced in this paper and any subset of the data cannot be shared with readers. machine learning algorithms in finance machine learning algorithms have been widely used in financial applications. in this paper we are particularly interested in the use of clustering methods for financial trades and transactions. in our study, we deployed machine learning to uncover patterns that are otherwise difficult to discern given the complexity of the data. our datasets encompassed over 200 discrete variables, some of which changed daily over the 36 months of observation. we deployed two machine learning techniques: dynamic time warping and k-means clustering using pycharm, tslearn and dtaidistance in python. our approach allowed us to systematically organize portfolios into clusters that demonstrate distinct investment behaviors. the figures below (figures 4 and 5) represent the trajectories of three distinct client groups within each of dataset 1 and dataset 2. this visual representation helped validate our data-driven grouping and helps to highlight the unique trends within each cluster. further detailed explanations of our methodologies are included in appendix 4. 13 on january 1, 2023, iiroc merged with the mutual fund dealers association and the combined regulatory was renamed ciro or the canadian investment regulatory of canada. 14 product choices for iiroc licensed representatives can include, for example, bonds, debentures, mortgage-backed securities, stocks, warrants, options, futures, mutual funds, exchange traded funds, labour sponsored funds, commodities, trusts, and hedge funds. financial services review, 32(2) 10 figure 4. dataset 1 cluster visual, portfolio values scaled to 1.0, aug 2019 to aug 2022 figure 5. dataset 2 cluster visual, portfolio values scaled to 1.0, aug 2019 to aug 2022 variables variables or features covered a broad range of data elements but can be summarized as demographic, behavioural, financial, and engineered. in our clustering, we focused on 74 discrete data elements and several derived ratios or engineered features as described in appendix 4 and summarized in table 2. results our clustering identified three unique groups per dataset (table 1). within each dataset, the clusters were very similar in terms of kyc data (age, income, gender, risk tolerance etc. but the clusters were differentiated by their saving behaviour (table 4). in particular, the clustering was driven by four dominant features. 1. net contributions, grace et al. 11 2. net contributions as a % of opening balance, 3. net contributions as a % of income and 4. contribution frequency. median net contribution rates (cr) ranged from a low of -59% to a high of 235%, compared to their opening balance. our analysis demonstrated that an active savings strategy was the most effective strategy for building wealth (utility) over the period examined. an active savings strategy was, on average, 5x more powerful at building wealth than relying on investment returns (see table 3, appendix 5 and appendix 6). table 3. portfolio growth attribution15, internal rates of returns (irr) versus contribution rates (cr), by cluster irr (annualized)16 cr (annualized) overall growth (closing balance/opening balance) dataset 1, cluster 1 0.7% -11.9% -11.6% dataset 1, cluster 2 2.3% 5.3% 14.5% dataset 1, cluster 3 2.6% 28.9% 51.6% dataset 2, cluster 4 1.1% -121.2% -50.5% dataset 2, cluster 5 2.5% 3.3% 15.8% dataset 2, cluster 6 1.7% 58.8% 158.7% the six clusters experienced dramatically different outcomes with respect to wealth accumulation. in dataset 1, cluster 1 had (on average) a 12% decrease in wealth over the period while cluster 2 had a 15% increase and cluster 3 a 52% increase. in dataset 2, cluster 1 had (on average) a 51% decrease in wealth over the period while cluster 2 had a 16% increase and cluster 3 a 159% increase. clusters 3 and 6 demonstrated savings patterns that could be considered consistent with the life cycle model predictions for this stage in their lives. i.e. cr was high and consistent with their peak income years. however, clusters 1, 2, 4 and 5 exhibited patterns that were inconsistent with the life cycle model. this observation is important because 56% of the investors observed in this study were from those four clusters. i.e. a majority of investors do not appear to be optimizing their savings rates during their peak income years perhaps contributing to the noted erosion in national savings rates. it is worth noting that in canada, withdrawals from an rsp before retirement are subject to onerous tax implications. the negative savings exhibited in clusters 1 and 4 may therefore suggest that these investors were under unusual financial stress over this period. we found little evidence to suggest active trading (portfolio churn) resulted in superior returns (neither for the advised or the diy investors) or unusual growth in wealth. demographic features, risk tolerance, trading strategies, portfolio mix, and behaviour features 15 anova testing of the irr and cr calculation yielded f-stat values of 109.6 and 393.709 respectively and p-values of 0.000 indicating significant differences in the average value of the clusters. 16 by way of comparison, over the period examined, canadian equity markets were up 5.7% and canadian bond markets were down 4.2%. a balanced portfolio (60eq/40fi) portfolio would have had a return of approximately 3.0%. (see appendix 1) financial services review, 32(2) 12 had minor to insignificant roles in driving the clustering. we noted that saving is a universal strategy the results were the same regardless of age groups, genders, risk tolerances and income levels. table 4. median points for select features by cluster all six clusters had investment returns that were consistent with their risk tolerance and the general market conditions at the time (see appendix 1). median irrs ranged from a low of 0.7% to a high of 2.6%. by way of comparison, over the same period, medium term canadian government bonds had a return of approximately -4.2%, canadian equities 5.7% and u.s. equities 10.8%. a balanced portfolio (60% fixed income, 40% equities) would have had a return of approximately 3.0% before fees. we found no evidence of investment returns driving significant wealth accumulation. instead returns followed a normal distribution curve with random deviations from the mean (see appendix 6 for a description of our investment return methodology). figure 6. dataset 1—change in wealth over time (normalized to $1 on day 1) note. canada emergency response benefit (cerb) payments began in march 2020 and ended in may 2022. grace et al. 13 figure 7. dataset 2—change in wealth over time (normalized to $1 on day 1) in figures 6 and 7 above, it should be noted that there is no overlap between cluster 2 and 3 or cluster 4, 5 and 6 – indicating statistically unique patterns. we noted that savings frequency had a marginal impact on investment returns as measured by irr but a significant impact on savings rates (cr) (appendix 7). investors who saved more frequently (biweekly vs quarterly for example) had a higher cr. and investors who saved systematically and regularly, had significantly higher savings rates than investors who saved periodically. given the administrative burden and the consistency of our observed trading behaviour, we have assumed that weekly, biweekly, and monthly trades were automated. our conclusions led to a discussion of the parsimony principle – to keep it simple. we concluded that when searching for wealth strategies with a powerful impact on financial resilience, keeping it simple – saving and saving often is not only easy to prescribe but effective. finally, we observed that the advised investors (dataset 1) had higher savings rates and lower withdrawal rates than the diy investors (dataset 2) although the size of dataset 2 is significantly smaller and the investors significantly younger than dataset 1. discussion limitations our conclusions are constrained by the datasets provided and the timeframe they cover. it is possible that additional data could influence the feature engineering deployed during our clustering. for example, we were not able to examine savings or trading behaviour in the context of fees or taxes. we observed that the data was not ‘perfect’. it included cases where the data was erroneous. it is not unusual with ‘real world’ data to encounter incorrect values or administrative challenges. these values would eventually be corrected over time, but our dataset was a point in time snapshot. we made efforts to curate the data and account for these outliers. subsequent testing and modelling determined that our curation did not materially impact our final calculated values. likewise, our conclusions are constrained by the unique time-period they cover and its relatively short (3-year) duration. the time-period (2019 to 2022) represents a particularly unique period given the pandemic. it is probable that investment returns would play a stronger role over a longer time-period. over the last 25 years, canadian fixed income yields have averaged closer to 3.6% annually compared to the -4.2% observed in our dataset. we would note however that historical investment returns would still pale in comparison to our strongest observed savings patterns. our conclusions are also limited to a specific view of an investors saving patterns – registered retirement saving plans. we did not have access to an investor’s savings at other institutions such as an employer sponsored pension plan or a savings account at their bank. it is possible that some investors would seek to optimize their savings across multiple accounts and our observations will not reflect those tendencies. nor have we attempted to answer the question ‘how much is enough’. it could be argued that some of our strongest observed savings' patterns financial services review, 32(2) 14 would be difficult to maintain over the long run. but we have left that question for the ‘further research’ section. implications our analysis concluded that: 1. an active savings strategy was more effective at building wealth than relying on investment returns or complex trading strategies alone. 2. saving is a simple, reliable, and powerful technique to build wealth. 3. frequent and disciplined saving is more effective than irregular or just-intime saving. 4. saving is a universal strategy the observed results were the same regardless of age groups, genders, risk tolerances and income levels. our conclusions offer participants a simple tool to cut through the complexity of the background noise and focus on actions that have a tangible impact on wealth accumulation and, by extension, financial resilience. the key would appear to be a focus on participation in savings plans, automated if possible, and incented to run as long as possible. there is no need to make this complicated. the simple act of opening an account and automating regular contributions is the most powerful technique that investors, policy makers, asset managers and advisors can deploy in the pursuit of wealth accumulation. for policy makers, we would encourage their continued sponsorship of savings plans such as retirement savings plans, tax free savings accounts, education saving plans and homeownership saving plans. we would also encourage careful consideration for raising the annual contribution limits in canada for rsps and tfsas, in particular. globally, policy makers have become strong advocates for financial literacy. we would encourage them to make saving a cornerstone strategy within their financial literacy plans. in canada, the financial consumer agency of canada (fcac) has embarked on a project to measure financial resilience. we would strongly advocate to integrate savings into such a measure. in our research we determined that the impact of strong savings behaviour crossed demographic lines such as age, gender, income and location. we would therefore advocate for the sponsorship of saving across the widest breadth of society and not narrowly focused on any one segment. for regulators, we would encourage a balanced perspective that combines transparency and a fiduciary perspective with incentives to save and processes that create simplicity for investors. transparency is a noble objective but when taken too far, it can create complexity and noise for decision makers (investors). complexity has been shown to be a barrier for the saving behaviours for which we advocate. for asset managers, we would encourage more balanced market facing activities that spend more time encouraging saving in general and less time overwhelming investors with investment and economic jargon and communications that create confusion rather than a tangible impact on client outcomes. we encourage a specific focus on the use of systematic saving routines (preauthorized chequing or pacs) as a tangible, simple mechanism for capital accumulation. for advisors we encourage a specific focus on systematic saving routines (preauthorized chequing or pacs). in addition to a strong impact on customer outcomes, automated savings routines can help streamline an advisor's operation and represent a low cost means to increase assets under administration. advisors could also consider a goal-based approach that helps clients keep their investing activities in perspective – i.e., encouraging activities, such as saving, that will have the strongest impact on their end goals. employers are in a unique position to encourage systematic saving through payroll deduction. we would encourage employers to strongly support saving plans for such things as retirement but to also include plans for children’s education and emergency accounts. robust sponsorship and participation in these plans have been shown to improve employee wellness with downstream benefits to the employer in terms of reduced absence, higher productivity and stronger employee loyalty. grace et al. 15 for consumers, the overwhelming conclusion from our research is to embrace saving and simplicity. a simple, automated savings plan into a diversified portfolio is the strongest way to achieve financial resilience. factors such as fees, taxes, rebalancing, asset mix etc. can also be important for some investors, but it starts with saving and the accumulation of investable capital. and the first step need not be intimidating. further research our research points towards ‘what to do’ but not necessarily ‘how to do it’. we would support future research into how policymakers, regulators, industry, and advisors can ensure strong savings behaviours over the long run. in our datasets, we observed a 55% to 60% participation rate in systematic trades. we would ask ‘what would it take’ to move participation rates to 80% or 90%? or to drive the behaviours observed in clusters 3 and 6 from 44% of investors to 60% or 70%? in canada, industry sponsored research has explored the concept of advisor alpha – a measure of the value derived from advice. our research hinted at higher savings rates in our advised dataset, but our diy dataset was too small to be definitive. we would support future research, in collaboration with industry partners, into the role advisors play in encouraging strong savings behaviour. our research was specific to retirement accounts. our conclusions would be strengthened by an examination of other forms of savings such as saving accounts or tax-free accounts. we plan to explore those areas next. our research deployed several methodological approaches to clustering financial data. while we are confident in the robustness of our analysis, further research would be beneficial in helping to identify best practices when researching with financial data. finally, our research has not addressed the question of ‘how much is enough’. is there a recommended minimum saving amount? does that amount change depending on the goal or the time horizon? is it better to save or eliminate debt? we have left all these questions for future papers and collaborations. disclosures acknowledgements: the authors wish to thank and acknowledge matt davison and nicole carleton for their expertise and advice during the final edits on the paper. the authors also thank mark reesor (wilfrid laurier university), kristina sendova (western university) and the many members of our data donor team for their invaluable insights that improved the content and conclusions of this document. author contributions: conceptualization, c.g. and a.m; data curation, l.f., y.m and a.f.; methodology, c.g, a.m., y.m., af; software, l.f.; validation, a.m., y.m.; formal analysis, a.m., y.m. l.f., a.f.; resources, c.g. and a.m.; writing original draft preparation, c.g.; writing—review and editing, c.g, a.m., l.f. y.m. and a.f.; visualization, a.m. and l.f.; supervision, c.g., and a.m.; project administration, c.g. all authors have read and agreed to the published version of the manuscript. funding: this research was supported, in part, by funding from the financial wellness lab at western university, natural sciences and engineering research council of canada, and our anonymous industry partners. institutional review board statement: the study was conducted according to the guidelines of the government of canada’s tri-council policy statement: ethical conduct for research involving humans (tcps 2) and approved by the research ethics board of western university (reb #118582, approved april 2022). informed consent statement: the data source is secondary and provided to us from the private data donor. additionally, the data has been anonymized so individuals cannot be identified by their accounts. this was approved by the ethics board above. data availability statement: the data used in this paper contains personal information for a number of canadians and cannot be shared due to a nondisclosure agreement with our private data donor. conflicts of interest: the dataset discussed herein was designed and collected by a private industry financial dealership that, in part, funded this research. the funders had no role in the analyses financial services review, 32(2) 16 or interpretation of the data, in the writing of the manuscript, or in the decision to publish the results. abbreviations the following abbreviations are used in this manuscript: anova analysis of variance aua assets under administration cad canadian dollars cr contribution rate (savings or deposits) diy do it yourself drip dividend reinvestment program etf exchange traded fund finra financial industry regulatory authority iiroc investment industry regulatory organization of canada irr internal rate of return kyc know your client kyp know your product pac preauthorized contribution rsp registered savings plan rrsp registered retirement savings plan tfsa tax free savings account appendices appendix 1: industry investment return benchmarks (august 2, 2019 to august 5, 2022) asset mix (equity/fi) asset class investment proxy returns 80/20 60/40 50/50 fixed income ishares core canadian universe bond index etf (xbb.to) -4.2% 20% 40% 50% cdn equity ishares core s&p/tsx capped composite index etf (xic.to) 5.7% 50% 35% 30% us equity ishares core s&p 500 index etf (xus.to) 10.8% 30% 25% 20% portfolio returns 5.2% 3.0% 1.8% source: https://ca.finance.yahoo.com/quote/ sourced july 11, 2023, returns annualized (cagr) https://ca.finance.yahoo.com/quote/ financial services review, 32(2) 17 appendix 2: quantitative modelling this paper considers the importance of contributions (as opposed to investment returns) during the accumulation phase of the retirement savings cycle. specifically, we present theoretical and empirical evidence that maintaining a high contribution rate early in the lifetime of the retirement account is crucial to building wealth. this observation, in and of itself, is hardly a new insight. the primary contribution of the paper is not the observation, rather it is the use of unique transaction-level data from thousands of retirement savings accounts in canada, to support both the conventional wisdom and theoretical model. our hope is that this paper provides a reference point for both financial advisors and policy makers, when providing advice and designing incentive programs. theoretical framework perhaps the simplest model for portfolio growth is to assume a deterministic per period investment growth rate of μ, a constant per period salary of x of which a constant fraction f is saved, and that all investment income is reinvested. the original wealth is v0. this model yields the difference equations: discrete vk+1 = (1+μ)vk + fx v0 = given where vk is the value of the investment portfolio at the kth time step. this has a solution vk = v0(1+μ)k + ( fx/μ)[(1+μ)k – 1] expand for small mu. using the binomial theorem (1+μ)k = 1 + kμ + k(k-1)/2 μ2 + higher order terms and [(1+μ)k – 1]/μ = k + kμ(k-1)/2 so… to linear order vk = v0 + μkv0 + fkx, or vk – v0 = fkx+μkv0 to quadratic order vk – v0 = fkx + μkv0 + ½ μ2k(k-1)v0 + ½ fμk(k-1)x this will also be true for beginning at time n and moving forward k steps: vn+k – vn = fkx + μkvn + ½ μ2k(k-1)vn + ½ fμk(k-1)x if k is about 3, mu 5% and f 10%, the 2nd order terms are going to be fairly neglible and so linear order is fine. note that when fx = μvn the first term, which is the savings term, and the second term, which is the growth term, are about equal in dollar value. fx = μvn = μ{v0(1+μ)k + ( fx/μ)[(1+μ)k – 1]} which, if v0 = 0 as is quite reasonable, occurs when fx = fx [(1+μ)k – 1] or 1 = (1+μ)k – 1 or (1 + μ)k = 2. this is when k ln ( 1+ μ) = ln(2) or when k* = ln(2)/ln(1+μ). using the linearization, good for small μ, that ln(1 + μ) = μ, this is approximately k* = ln(2)/μ. for μ = 5% this occurs when k is about 14 years. financial services review, 32(2) 18 appendix 3: retirement savings plans in canada (rsps) both of our datasets encompass registered retirements savings plans. a registered retirement savings plan (rrsp or rsp) is a savings plan, registered with the canadian federal government. investors who contribute funds to an rsp, gain a "tax-advantage" in that the contribution is exempt from income taxes in the year they make the contribution. any investment income earned from investments held within the rsp also grows tax-deferred until it's withdrawn. according to statistics canada 17 , in 2020, over 6.2 million canadians made contributions to a registered retirement savings plan (totalling $50.1 billion). twenty two percent of canadian tax filers made rrsp contributions in 2020 with a median contribution of $3,600. canadians can open an rsp at their financial institution either by • working through a licensed investment representative (advisor) • opening a diy account or • through their employer. saving in the context of rsps generally takes the form of either periodic lump sum payments or automated deposits referred to by the industry as pacs (pre-authorized contributions). in canada, lump sum payments are frequently made in late february each year, just before the rsp contribution deadline for the previous year. pacs are generally set on a monthly or quarterly frequency and are electronically withdrawn from the investors bank account. deposits or savings into an rsp are traditionally referred to as ‘contributions. participants can also transfer funds from other rsps they may own to consolidate their investments. for the purposes of this paper, we did not classify transfers as a saving activity since the savings behaviour was exhibited in a separate account prior to our research. it could be argued that re-invested dividends (drips) or the roll-over of interest payments are 17 statistic canada, https://www150.statcan.gc.ca/n1/dailyquotidien/220401/dq220401a-eng.htm a form of saving but also represent a return on the original capital. for this reason, we include drips and re-invested interest payments in both our internal rate of return (irr) calculations and our contribution rates (cr) calculations. appendix 4: detailed clustering methodology machine learning algorithms have been widely used in financial applications, such as risk modelling, return forecasting, and portfolio construction (emerson et al., 2019), quantitative finance (rundo et al., 2019), financial distress prediction (huang et al., 2019), banking risk management (leo et al., 2019), credit-scoring models and financial crisis prediction (lin et al., 2011), automation through artificial intelligence (donepudi, 2019), market prediction (henrique et al., 2019), and credit risk modeling, detection of credit card fraud and money laundering, and surveillance of conduct breaches at financial institutions (van liebergen, 2017). popular algorithms used in these applications are support vector machines (kim, 2003), neural networks (west et al., 2005), and random forests (patel et al., 2015). in this paper we are particularly interested in clustering methods for financial trades and transactions. recent work in this area includes agglomerative hierarchical clustering for asset allocation (raffinot, 2017) and aggregating stocks using dynamic time-series warping as a distance measure (lim et al., 2020), selforganizing maps and k-means clustering methods in combination with classifier techniques to predict financial distress (tsai, 2014), fuzzy cmedoids clustering method for classifying financial time series (d’urso et al., 2013), and clustering algorithms for financial risk analysis using multiple criteria decision-making methods (kou et al., 2014). absent from this body of work is the use of this broad class of techniques to analyze trading behaviours, the focus of this paper. in our study, we deployed machine learning to help uncover previously unknown patterns in the data. machine learning – and in particular, grace et al. 19 clustering – has proven to be invaluable when examining large, complex datasets. our datasets encompassed over 200 discrete variables, some of which changed daily over the 36 months of observation. we deployed two machine learning techniques: dynamic time warping and kmeans clustering using python, pycharm, tslearn and dtaidistance software. we used dynamic time warping (dtw) to quantify the degree of similarity among various portfolios' weekly average market values. we trained our dtw models using the rrsp portfolio's weekly average market values (the sole variable utilized in our time series analysis). we then conducted a deeper within-cluster analysis on kyc variables such as income and retirement indicators, however these variables weren't included in our model training. by capturing the temporal dynamics of these portfolios, we were able to identify patterns in clients’ trading behaviors. afterward, we applied k-means clustering algorithms to categorize portfolios exhibiting similar trajectories. this approach allowed us to systematically organize portfolios into clusters that demonstrate distinct investment behaviors. the figures below (figures 3 and 4) represent the trajectories of three distinct client groups within each of dataset 1 and dataset 2. this visual representation helps validate our data-driven grouping and helps to highlight the unique trends within each cluster. classification of investor accounts by contribution frequency each of the datasets under study included transaction level detail at a daily level. to examine contribution patterns in terms of frequency, we first curated the data to eliminate transactions that did not affect portfolio market values. for example, administrative adjustments, corrective transactions, and some fees collected directly from the client. we then aggregated the remaining transactions at a daily level for each client to see whether the net sum of these activities was positive for a given day. if the total number of days with positive net sum was greater than 90% of the total number of business days in the interval, the account was classified as a daily contributor, otherwise, the aggregation was repeated, respectively, on weekly, biweekly, monthly, and quarterly levels to classify every account according to its contribution pattern. this bottom-up approach used the same threshold for pattern similarity (i.e., 90%) at every level of aggregation. if an account failed to be classified as one of the predefined periodic contributors, it was categorized as an irregular contributor. results are included in appendix 7. time series data clustering has become an important part of financial data analysis due to its ability to reveal hidden patterns and correlations in time-series data. by grouping similar time series together, it allows for a more efficient and targeted analysis, enabling analysts to draw conclusions about collective behaviour or attributes. studies have corroborated the efficiency of using time series clustering for financial data analysis, highlighting its validity as an approach (dose et al., 2005). min-max scaling before proceeding with the clustering process, it's essential to scale the time series data to ensure that the variance in scale of different features does not distort the distances between data points, which in turn would impact the performance of the clustering algorithm. min-max scaling is an effective method in this regard, as it brings all values within a predetermined range, typically between 0 and 1. this prevents features with larger scales from dominating the calculation of distances. when applying min-max scaling to portfolio weekly market values, it's important to consider the structure of the input. in our case, each time series is associated with a unique account id and scaling must be performed on an account-byaccount basis. suppose we have a time series associated with a particular account id, x =  [x1,  x2,   … ,  xn]. the min-max scaler operation for each account id can be expressed as follows: xscaled  =   x − xmin xmax  −  xmin where xmin and xmax are the minimum and maximum values of the time series x associated financial services review, 32(2) 20 with that account id. this scales the time series xscaled such that all values lie between 0 and 1. this transformation ensures that we're comparing the shape of the time series, rather than being influenced by their magnitude when performing the subsequent clustering with dtw and kmeans. dynamic time warping (dtw) algorithm the dynamic time warping algorithm is a technique used to measure similarity between two sequences which may vary in time or speed. the algorithm considers all possible alignments between the sequences and identifies the optimal alignment that minimizes the total distance between them. for two time series x  =  (x1,  x2,   … ,  xn) and y  =  (y1,  y2,   … ,  ym), which are represented as arrays of respective shapes (n, 1) and (m, 1), the steps involved in the dtw algorithm are as follows: 1. initialization: create an n-by-m matrix where the (i , j )-th element of the matrix contains the distance d(xi , yj ) between the points xi and yj. the distance can be computed using a selected distance metric, commonly the euclidean distance. the calculation formula for euclidean distance is: d(xi,  yj) = √ ∑(xi − yj) 2    . create a second n-by-m matrix d for storing the accumulated distances, where d(i,  j) represents the sum of d(xi,  yj) and the minimum among d(i − 1,  j) , d(i,  j − 1), d(i − 1,  j − 1). 2. matrix filling: iterate over the matrix d, starting from d(1,1) , and compute the accumulated distance for each cell using: d(i,  j)  =  d(xi,  yj)  + min[d(i − 1,  j),  d(i,  j − 1),  d(i − 1,  j − 1)] according to this equation, the accumulated distance is the sum of the distance at that point and the minimum accumulated distance among its neighboring points. 3. path identification: starting from d(n,m), move backwards to d(1,1) by choosing at each step the cell (i − 1,  j), (i,  j − 1) , or (i − 1,  j − 1) that has the smallest accumulated distance. the path that is formed, known as the warping path, represents the optimal alignment between the two-time series. the dtw distance between the two time-series is then given by the value at d(n,m), which represents the minimum sum of distances for aligning the two sequences. the whole process considers the temporal dynamics and can provide a more accurate measure of similarity between time series data compared to traditional euclidean distance, especially when dealing with sequences of different lengths or speeds. the flexibility of the dtw algorithm makes it particularly suited for financial time series analysis, where data can exhibit significant temporal variations. k-means clustering for our research, we used k-means clustering, an iterative technique widely used in machine learning and data mining. the fundamental idea behind k-means clustering is to classify dataset into k different clusters in such a manner that the within-cluster variations are minimized. the iterative process of the k-means algorithm involves partitioning the portfolios into k clusters, computing the centroid of each cluster, and reassigning the portfolio to the cluster whose centroid is closest. the process continues until the positions of the centroids stabilize, indicating the optimal clustering of the data. since the nature of time-series data and the flexibility of dtw in aligning sequences, the centroid calculation can't be as straightforward as simply taking the arithmetic mean of the points in each cluster (petitjean et al., 2011). we use a variant of k-means known as time series kmeans that utilizes the dtw distance as the dissimilarity measure. in this context, the 'centroid' of a cluster is defined using the dtw barycenter averaging (dba) method, which provides an averaged sequence that minimizes the distances to the sequences of the cluster. in each iteration, dba performs three main steps: grace et al. 21 1. computing dtw alignments: in this step, we calculate the dtw between the temporary average sequence (also known as the centroid) and every individual sequence within our set of sequences, denoted as s  =  {s1, … , sn}. this dtw computation allows us to establish links between the coordinates of the average sequence and the coordinates of the individual sequences. 2. updating centroid coordinates: each coordinate of the average sequence is updated as the barycenter (or geometric center) of coordinates linked to it in the previous step. the average sequence at iteration i is represented as c  =  c1, … , ct , and we aim to update its coordinates for the next iteration (i+1), represented as c′ =  c1 ′ , … , ct ′ . now, we use a function 'assoc' that associates each coordinate of the average sequence with one or more coordinates of the sequences in s. this function is computed during the dtw calculation between c and each sequence in s. 3. we can then define the t-th coordinate of the average sequence ct as: ct  =  barycenter(assoc(ct)) 4. the barycenter is the arithmetic mean of a set of points {x1,...,xα} in the vector space: barycenter{x1, … , xn} = (x1 + … + xn) n after computing the new centroid, we then repeat the dtw computation between this updated average sequence and all sequences in s. the associations created by the dtw may change as a result, which is why we iteratively perform this process until the average sequence converges to a stable configuration. together, the combination of min-max scaling, dtw and k-means clustering forms an effective methodology for time series data clustering in our research. appendix 5: investment returns versus contribution rates figure 8. dataset 1—investment returns (irr) and contribution rates (cr) by cluster financial services review, 32(2) 22 figure 9. dataset 2—investment returns (irr) and contribution rates (cr) by cluster appendix 6: investment return calculations (irr) to compare investment performance across different portfolios, we needed a comparison criterion that considers the cash flows into and out of the portfolio, as well as the timing of such cash flows. one simple and commonly used criterion is the internal rate of return (irr). the irr is defined as the discount rate at which the present value of all cash flows in a given period of time equals to 0. consider a portfolio that is invested from time 0 to time t. assume that the market value of the portfolio at time 0 is s0, and that the market value at the conclusion of the investment is st. assume that the investor makes n transactions before time t, where the ith transaction happens at time ti and has amount ci. we further assume that ci > 0 if the ith transaction is an additional contribution to the portfolio, and that ci < 0 otherwise. the irr of this investment is the root to the equation s0 +∑ci n i=1 e−rti − ste −rt = 0, where ∑ ci n i=1 e−tir = 0 if n = 0. notice more than one root may exist if one or more ci is negative, i.e., the investor withdraws at least once from the investment. some approaches have been proposed in the existing literature to select the most useful irr in this case, see, for example, hartman and schafrick (2004). for the rrsp accounts that are analyzed in this project, there is a strong incentive for investors to refrain from withdrawing prematurely. consequently, large withdrawals from the portfolio are less commonly observed compared to other account types. to calculate the irr of different portfolios, we first need to obtain the amount and the timing of all the cash flows. however, as the trading records of the raw data sets contain errors that are challenging, if not impossible, to distinguish from correct records, we need to resort to approximations. to this end, we use the following procedure: 1. calculate the daily change of number of shares for all the securities in a portfolio. determine the reason for such changes and keep only those caused by a trading decision from the investor. for example, reinvested dividends would cause the number of shares to change, but they are not counted as cash flows. 2. calculate the average trading price for each security on each day. calculate the amount of the changes in step 1). 3. add other cash flows that does not cause changes in the number of shares. for example, dividends paid out as cash do not cause the number of shares to change, but they are counted as cash flows since they are returns from the investment. there are two main sources of error in the approximation procedure: the rounding error of the number of shares that are exchanged, and the difference between average trading price and actual trading price. the errors are not material and do not affect the results significantly. grace et al. 23 a histogram of the observed irrs is given in figure 10. the table below summarizes the quantiles of the irr. figure 10. histogram of the realized irr quantile 0.01 0.25 0.50 0.75 0.99 irr 0.433257 0.049519 0.011676 0.054847 .653425 appendix 7: savings frequency and wealth accumulation figures figure 11. dataset 1—savings frequency and wealth accumulation financial services review, 32(2) 24 figure 12. dataset 1—savings frequency, irr versus cr figure 13. dataset 2—savings frequency and wealth accumulation grace et al. 25 figure 14. dataset 2—savings frequency and irr versus cr references avanzi, b., taylor, g., wong, b., & xian, a. (2021). modelling and understanding count processes through a markovmodulated non-homogeneous poisson process framework. european journal of operational research, 290(1), 177195. baldwin, r. d. (2022). the evolving wealth of canadians: who is better fixed for retirement? who is not?. who is not. bayraktar, e., & young, v. r. (2010). optimal investment strategy to minimize occupation time. annals of operations research, 176(1), 389-408. blake, d., wright, d., & zhang, y. (2013). target-driven investing: optimal investment strategies in defined contribution pension plans under loss aversion. journal of economic dynamics and control, 37(1), 195-209. cagetti, m. (2003). wealth accumulation over the life cycle and precautionary savings. journal of business & economic statistics, 21(3), 339-353. ülkümen, g., & cheema, a. (2011). framing goals to influence personal savings: the role of specificity and construal level. journal of marketing research, 48(6), 958-969. cobb-clark, d. a., kassenboehmer, s. c., & sinning, m. g. (2016). locus of control and savings. journal of banking & finance, 73(c), 113-130. cruz, m., & ahmed, s. (2016). on the impact of demographic change on growth, savings, and poverty. world bank policy research working paper, (7805). d’urso, p., cappelli, c., di lallo, d., & massari, r. (2013). clustering of financial time series. physica a: statistical mechanics and its applications, 392(9), 2114-2129. dholakia, u., tam, l., yoon, s., & wong, n. (2016). the ant and the grasshopper: understanding personal saving orientation of consumers. journal of consumer research, 43(1), 134-155. donepudi, p. k. (2019). automation and machine learning in transforming the financial industry. asian business review, 9(3), 129-138. dose, c., & cincotti, s. (2005). clustering of financial time series with application to index and enhanced index tracking portfolio. physica a: statistical financial services review, 32(2) 26 mechanics and its applications, 355(1), 145-151. dynan, k. e., skinner, j., & zeldes, s. p. (2004). do the rich save more?. journal of political economy, 112(2), 397-444. emerson, s., kennedy, r., o'shea, l., & o'brien, j. (2019, may). trends and applications of machine learning in quantitative finance. in 8th international conference on economics and finance research (icefr 2019). fcac. (2021). make change that counts: national financial literacy strategy 2021-2026. https://www.canada.ca/en/financialconsumer-agency/programs/financialliteracy/financial-literacy-strategy-20212026.html feiveson, l., & sabelhaus, j. (2019). lifecycle patterns of saving and wealth accumulation. fisher, p. (2010). gender differences in personal saving behaviors. journal of financial counseling and planning, 21(1). foerster, s., linnainmaa, j. t., melzer, b. t., & previtero, a. (2017). retail financial advice: does one size fit all?. the journal of finance, 72(4), 1441-1482. gale, w. g., iwry, j. m., & orszag, p. r. (2005). the automatic 401 (k): a simple way to strengthen retirement savings (pp. 2006-1). retirement security project. hartman, j. c., & schafrick, i. c. (2004). the relevant internal rate of return. the engineering economist, 49(2), 139-158. hall, c. c. (2021). promoting savings for financial resilience: expanding the psychological perspective. current directions in psychological science, 30(1), 49-54. henrique, b. m., sobreiro, v. a., & kimura, h. (2019). literature review: machine learning techniques applied to financial market prediction. expert systems with applications, 124, 226-251. hershfield, h. e., goldstein, d. g., sharpe, w. f., fox, j., yeykelis, l., carstensen, l. l., & bailenson, j. n. (2011). increasing saving behavior through age-progressed renderings of the future self. journal of marketing research, 48(spl), s23-s37. huang, y. p., & yen, m. f. (2019). a new perspective of performance comparison among machine learning algorithms for financial distress prediction. applied soft computing, 83, 105663. juster, f. t., smith, j. p., & stafford, f. (1999). the measurement and structure of household wealth. labour economics, 6(2), 253-275. kahneman, d. (2012). the human side of decision making. journal of investment consulting, 13, 9-14. kahneman, d., & thaler, r. h. (2006). anomalies: utility maximization and experienced utility. journal of economic perspectives, 20(1), 221-234. kempson, e., finney, a., & poppe, c. (2017). financial well-being: a conceptual model and preliminary analysis. working paper. https://www.researchgate.net/publicatio n/318852257_financial_wellbeing_a_conceptual_model_and_preli minary_analysis?channel=doi&linkid= 59819dd14585150575c00b54&showful ltext=true kim, k. j. (2003). financial time series forecasting using support vector machines. neurocomputing, 55(1-2), 307-319. kou, g., peng, y., & wang, g. (2014). evaluation of clustering algorithms for financial risk analysis using mcdm methods. information sciences, 275, 112. leo, m., sharma, s., & maddulety, k. (2019). machine learning in banking risk management: a literature review. risks, 7(1), 29. grace et al. 27 lim, t., & ong, c. s. (2020). portfolio diversification using shape-based clustering. the journal of financial data science. lin, w. y., hu, y. h., & tsai, c. f. (2011). machine learning in financial crisis prediction: a survey. ieee transactions on systems, man, and cybernetics, part c (applications and reviews), 42(4), 421-436. linnainmaa, j. t., melzer, b. t., & previtero, a. (2021). the misguided beliefs of financial advisors. the journal of finance, 76(2), 587-621. macgee, j., pugh, t. m., & see, k. (2022). the heterogeneous effects of covid‐19 on canadian household consumption, debt and savings. canadian journal of economics/revue canadienne d'économique, 55, 54-87. nenkov, g. y., macinnis, d. j., & morrin, m. (2009). how do emotions influence saving behavior. center for retirement research, 9(8), 1-11. markowitz, h. m. (1991). foundations of portfolio theory. the journal of finance, 46(2), 469-477. markowitz, h. m. (2010). portfolio theory: as i still see it. annu. rev. financ. econ., 2(1), 1-23. maynard, n., de neve, j. e., & tufano, p. (2008). consumer demand for prizelinked savings: a preliminary analysis. harvard business school finance working paper, (08-061). metzler, z., & grace, c. (2021). learning about financial health in canada. quantitative finance & economics, 5(3): 542-570. merton, r. c. (1973). an intertemporal capital asset pricing model. econometrica: journal of the econometric society, 867-887. modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. franco modigliani, 1(1), 388-436. montmarquette, c., & viennot-briot, n. (2015). the value of financial advice. annals of economics and finance, 16(1), 69-94. newmeyer, c., warmath, d., o’connor, g. e., & wong, n. (2021). is savings automation helpful to liquid savings? it depends on whether you have a savings habit. journal of public policy & marketing, 40(2), 285-297. patel, j., shah, s., thakkar, p., & kotecha, k. (2015). predicting stock and stock price index movement using trend deterministic data preparation and machine learning techniques. expert systems with applications, 42(1), 259268. petitjean, f., ketterlin, a., & gançarski, p. (2011). a global averaging method for dynamic time warping, with applications to clustering. pattern recognition, 44(3), 678-693. raffinot, t. (2017). hierarchical clusteringbased asset allocation. the journal of portfolio management, 44(2), 89-99. rocher, l., hendrickx, j. m., & de montjoye, y. a. (2019). estimating the success of re-identifications in incomplete datasets using generative models. nature communications, 10(1), 1-9. rundo, f., trenta, f., di stallo, a. l., & battiato, s. (2019). machine learning for quantitative finance applications: a survey. applied sciences, 9(24), 5574. shefrin, h. m., & thaler, r. h. (2004), mental accounting matters (chapter 14). advances in behavioural economics. princeton university press soman, d., & cheema, a. (2011). earmarking and partitioning: increasing saving by low-income households. journal of marketing research, 48(spl), s14s22. suh, e. (2022). can't save or won't save: financial resilience and discretionary retirement saving among british adults in their thirties and forties. financial services review, 32(2) 28 ageing & society, 42(12), 29402967. thaler, r. h., & benartzi, s. (2004). save more tomorrow™: using behavioral economics to increase employee saving. journal of political economy, 112(s1), s164-s187. thompson, j. r., feng, l., reesor, r. m., & grace, c. (2021). know your clients’ behaviours: a cluster analysis of financial transactions. journal of risk and financial management, 14(2), 50. tsai, c. f. (2014). combining cluster analysis with classifier ensembles to predict financial distress. information fusion, 16, 46-58. turner, t. m., & luea, h. (2009). homeownership, wealth accumulation and income status. journal of housing economics, 18(2), 104-114. west, d., dellana, s., & qian, j. (2005). neural network ensemble strategies for financial decision applications. computers & operations research, 32(10), 2543-2559. van liebergen, b. (2017). machine learning: a revolution in risk management and compliance?. journal of financial transformation, 45, 60-67. vlaev, i., & elliott, a. (2014). financial well-being components. social indicators research, 118, 1103-1123. manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 life insurance as a retirement income tool russell delibero, ph.d., cfpa, wade d. pfau, ph.d., cfab,* athe american college of financial services, 270 s bryn mawr avenue, bryn mawr, pa 19010, usa bthe american college of financial services, 270 s bryn mawr avenue, bryn mawr, pa 19010, usa abstract given its tax-preferential treatment, careful study is warranted to determine whether life insurance can play an important role in an overall retirement portfolio. this study develops hypothetical scenarios for different types of individuals with varying ages and distribution periods, while using a historical outlook to determine the proper structure of a variable universal life insurance policy. we compare a variable universal life policy to different investment vehicles (both in qualified and non-qualified accounts) on an after-tax basis to better understand the potential tradeoff for tax-deferral and insurance fees within life insurance. © 2017 academy of financial services. all rights reserved. jel classification: d14; d81; g11; g23 keywords: life insurance; retirement planning; asset allocation; retirement spending 1. introduction with the largest generation in the united states entering retirement, retirement planning is growing in prevalence and relevance as baby-boomers worry about how to fund their retirements. studies such as bengen (2004) and steiner (2014) have focused on how to spend from savings during retirement. baseline assumptions in such studies rarely consider the impact of taxation. however, there remains widespread concern about taxes and their impact on retirees (mccarthy, 2011; silver, 2013). different types of investment vehicles offer varying tax treatment. life insurance is a tax preferential vehicle that one can use as a piece of an overall asset allocation to help satisfy both retirement income needs, and concerns * corresponding author. tel.: �1-610-526-1569; fax: �1-610-526-1569. e-mail address: wade.pfau@theamericancollege.edu (w.d. pfau) financial services review 26 (2017) 221–240 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. regarding taxation. the strategy of using life insurance as a supplemental retirement income vehicle raises the issue of whether it is appropriate and what opportunity costs may exist. prior research (katt, 2009; parish, 2014; resnick & resnick, 2009) has shown there are pitfalls to using life insurance in this manner. however, there are problems with past research approaches, and a goal of this research is to demonstrate a proper structure to analyze the use of life insurance as a retirement income vehicle. after determining the proper structure and framework, this research will then compare how life insurance as an asset class compares to the performance of equities on both a preand post-tax basis. this comparison will provide transparency to the embedded costs and suitability of insurance for individuals. research in this area often does not factor in the tax consequences of accumulation and distribution strategies. one reason is tax situations can vary so dramatically across individuals. this study will make assumptions regarding tax rates to provide a clearer picture of the outcomes net of taxes. the analysis of taxation is crucial as tax savings provide one of the largest benefits to utilizing life insurance as a supplemental income vehicle. this study develops a framework for properly structuring variable universal life insurance for supplemental cash flow during retirement. two of the most important numbers for this type of study are the rate of return during the accumulation phase and the safe withdrawal rate during the distribution phase. establishing a potential proper rate of return and an appropriate withdrawal rate will help avoid some of the pitfalls from previous studies using variable universal life insurance and provide an alternative to the growing problem of taxable income during retirement. after finding sustainable assumptions for accumulation and distribution rates in a variable universal life policy, the life insurance vehicle cash flow generation provides comparison data to an investments-only strategy determining the probability of success that a qualified or non-qualified investment account would be able to generate for the same after-tax income. the variable universal life policy was chosen to be the main source of analysis because it most closely represents a comparison to an investments-only strategy, providing the owner with flexibility in choosing the underlying investments (from a menu of options) and risk of the policy. the comparison will demonstrate any advantages of the tax preferential treatment along with whether life insurance is a viable replacement for, alternative to, or supplement of retirement income. the key variables in this study include the following: age at initial plan start, gender, health classification, length of accumulation, length of income, amount of death benefit, amount of annual premium paid, historical rates of return, costs of insurance, accumulation rate, and distribution rate. this study will include multiple scenarios for different ages, length of distribution periods, and health classification, but the base case will use the following scenario: y age at initial plan start (45), y health classification (preferred non-smoker), y length of accumulation (19 years), y length of distributions (15 years), y amount of annual premium paid ($50k), y male mortality tables. 222 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 six key research questions reflect the objectives for this study include: (1) what is a safe combined hypothetical accumulation rate and distribution rate to assume when life insurance is intended to serve as a distribution vehicle? (2) what are the disadvantages for using life insurance as a distribution vehicle? (3) what characteristics determine appropriate candidates for using life insurance effectively as a retirement cash flow tool? (4) what is the opportunity cost of using life insurance as an asset class compared with a brokerage account? (5) what are the tax advantages to using life insurance as an asset class? (6) is life insurance a reasonable choice as a supplemental cash flow vehicle for retirement savings? many financial advisors within the industry have shied away from using life insurance as an asset class. katt (2009), resnick and resnick (2009), and parish (2014) have all studied life insurance as a distribution vehicle indicating that both academic and industry studies have shown poor results. often this was because of being too aggressive with the hypothetical accumulation or distribution rates. with these poor experiences, this potentially useful and beneficial strategy may become less common or utilized improperly, which can lead to suboptimal outcomes for clients. this study aims to provide an appropriate framework for utilizing life insurance as a supplemental income vehicle and then compare it to other investment options. this comparison will identify the potential positive attributes for the supplemental income strategy and will demonstrate the proper way to structure the policy to obtain more beneficial results. the benefit for at least some individuals will be another source of retirement funds that offers tax advantages. the more sources that an individual can access for retirement income, the more prepared that individual will be for different future economic environments and changing tax regimes. this study will look at the ability to use life insurance as a distribution vehicle and determine safe hypothetical accumulation and distribution rates to use in conjunction. this study will also compare using life insurance as a distribution vehicle with other investment vehicles on both a preand post-tax basis. several issues arise when looking at using life insurance for this purpose. life insurance will have different costs for different individuals based on age, gender, health classification, type of policy, and carrier chosen. there are also limitations based on the length of the distribution period for the policy. there is less risk and more income potential for a shorter distribution period. a significant advantage for using life insurance as a distribution vehicle is the tax preferential treatment. the comparisons made to other distribution vehicles with different tax treatment will depend on an individual’s tax rates for ordinary income and investment income. while this study utilizes a specific set of assumptions, results will vary in practice based on different individual circumstances. with retirement planning and subsequently tax planning becoming such prevalent topics, research on efficiencies and different potential approaches can be beneficial. while there are many different approaches to retirement income planning with a variety of investment vehicles, tax planning takes these approaches one-step further. retirement planning and tax issues can vary widely for everyone, but it is important to look at a large scope of options to tailor fit these approaches and ensure the vetting and availability of different options. life insurance is a vehicle not originally designed with retirement planning in mind, but because of its tax preferential treatment, it is a potential option. prior research discusses the accumulation and distribution phases of retirement planning, along with different potential tax planning strategies. additional research has discussed life insurance as a retirement planning 223r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 vehicle, but often it has focused on the pitfalls and potential negative ramifications, and concluded that it is not a viable option. this study will review this prior research and look to build and expand upon the potential framework and usefulness of utilizing life insurance as a distribution vehicle for retirement income. 2. literature review retirement planning is more prevalent now than ever before. in past years, u.s. workers often received a defined benefit pension from their employer and could rely on social security. the shift to defined contribution retirement plans in the workforce has changed the scope of retirement planning. in addition, the largest generation in the united states is now entering its retirement years. retirement planning consists of two phases, the accumulation phase, which entails gathering assets, and the distribution phase, which refers to using the accumulated assets to support spending during non-income earning years. bengen (2004) studied what a safe withdrawal rate would be during the distribution phase to avoid outliving assets in an investment portfolio. this research concluded that 4% would be a safe initial withdrawal rate to use based on the worst-case historical outcome from varying 30-year periods in the united states. one critical aspect missing from his original study is a consideration of taxes, which he did discuss in his book in 2006. the safe initial withdrawal of 4% is a gross number before the payment of any taxes and will be lower than 4% on an after-tax basis. his assumption was that the investments would be in a qualified account and, thus, benefit from tax deferral. however, depending on ordinary income tax rates, especially if higher in the future, individuals may not obtain the highest standard of living possible if more tax efficient options are available. not all retirement assets will be inside a qualified account and taxable investments could be subject to taxation on an ongoing basis, thus reducing the net amount of the 4% safe initial withdrawal rate. sumutka, sumutka, and coopersmith (2012) use a comprehensive tax model to evaluate different withdrawal strategies, discussing the three different types of investment accounts: tax-deferred, taxable, and tax-free. sumutka et al. state that the withdrawal sequence from the various accounts will affect overall taxation on the portfolio and find that the optimal tax-efficient strategy produces withdrawal stability utilizing low withdrawal rates during early retirement years. income stability helps to avoid the loss of itemized deductions, the loss of tax favored long-term capital gains treatment, and the imposition of the amt. tax efficiency comes from a withdrawal sequence of taxable assets, tax-deferred assets, and then tax-free assets, keeping in mind offsetting tax deductions and tax-bracket management to avoid higher taxation. other withdrawal strategies may provide for smaller tax payments, but this results in lower wealth creation. optimal withdrawal strategies demonstrate the importance of tax planning during retirement and the ability to have multiple asset vehicles to withdraw from during retirement, including tax diversification and asset location. mccarthy (2011) concludes that taxes will be increasing in the future based on federal spending and the looming federal deficit. tax increases would decrease the standard of living for many retirees. during retirement, this may mean an increased withdrawal rate, which can shorten the time horizon that assets will last, or force a lower standard of living. mccarthy 224 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 states that increasing tax burdens creates opportunities for tax planning. the benefits to minimizing taxes may have a compounding effect and mean the difference between running out of money or a lower lifestyle, and living as planned. with tax planning being such a crucial aspect to retirement income planning, the differential taxation of investment vehicles needs consideration to determine a net distribution to an individual. silver (2013) concludes that existing retirement planning concepts may have served many retirees well, but with changing laws and historically low tax rates, people are looking for alternative solutions. there are several different types of vehicles for retirement planning, and they fall into the three broad categories of tax deferred, taxable, and tax exempt. silver references a 2010 survey conducted by lincoln financial group that found taxes constitute 31% of retiree expenses and the amount of taxes paid surprised 11% of retirees. up to 85% of social security benefits can be taxable, and this is a reason planners are looking to minimize taxable income during retirement. because of the potential to join the next tax bracket and the negative impact on social security income, taxable assets are no longer as desired, even if they offer slightly higher rates of returns, if alternative tax advantaged options are available. this is especially true when people believe taxes will be higher in the future. many people look at permanent life insurance solely for the death benefit it provides, but life insurance can serve multiple purposes. a major benefit to permanent life insurance, with implications for retirement planning, is the ability to accumulate equity within the contract. what truly makes this an advantage is that the equity accumulates on a tax-deferred basis. unlike annuities, life insurance distributions are calculated on a first in, first out basis (fifo). one can withdraw the funds placed into the contract without paying any taxes. the next step is the ability to take loans against the policy. the reason loan provisions are important is that a loan is not a taxable event, even though there is an on-going interest expense. because of this, it is possible to extract a significant amount of cash from within a life insurance policy without paying taxes. this strategy seems simple and beneficial: accumulate equity, withdraw up to the basis, borrow gains, and never pay taxes because the death benefit will also be income tax free. this would certainly help solve problems retirees face with income tax planning during retirement. however, there are many issues to consider with this basic strategy. the tax preferential treatment afforded to life insurance is applicable in general to policies that can accrue cash, and not to any specific type of policy. however, there are several different ways to accumulate the cash value, depending on the type of policy. varying types of permanent life insurance policy have different features and benefits with potentially different accrual methods. often, the type of policy that a person selects depends on the individual’s risk tolerance, desire for control, flexibility, upside potential compared with guarantees, and general beliefs about financial markets. an individual with a more aggressive risk tolerance that would rather choose from a menu of investments has a better option with a variable universal life policy. variable universal life policies offer different subaccounts in which one can make investments; thus, allowing an individual to have equity market exposure. this potential equity exposure allows for greater upside potential but also carries risk if the investments do not perform as expected. the 225r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 owner of a variable universal life policy can purchase additional guarantees to ensure payment of the death benefit regardless of market performance. katt (2015) discusses the history of the different types of cash value life insurance policies and the industry’s evolution over time. in the 1990s when interest rates began to decline, participating whole life (pwl) and universal life (ul) policies became less attractive. this led to the introduction of variable universal life (vul) policies. vuls benefited from investment performance based on stock and bond subaccounts instead of relying on dividends or an interest-crediting rate. however, early on there were abuses by agents in illustrating too high of hypothetical rates of return. most likely this resulted from negligence, but potentially there was an intent to deceive consumers. regulation capped maximum illustration assumptions. vuls, like other equity assets, can be dangerous if policy owners sell the underlying investments at the wrong time. traditional whole life policies work well to provide supplemental income on a taxadvantaged basis. however, non-guaranteed dividends affect cash accumulation, which minimizes investor control and flexibility. kriesel (2010) concludes that many investors were not happy with the conservative gains found in traditional whole life insurance policies. this led to the creation of ul to take advantage of the higher interest rates of the 1980s. however, the marketing of ul policies frequently included maximizing the death benefit with minimal funding. unfortunately, when interest rates fell, these policies suffered. variable universal life policies take advantage of the ability to invest within the policy via mutual funds. because of the tax advantages afforded to life insurance, this can create an opportunity to use variable universal life policies as a roth ira alternative without the income restrictions and contribution limits. katt (2013) further discusses how cash value within life insurance policies is different from a bank account. a client example discusses a situation where an upset client results from the charging of interest to access the cash value within the policy and feels that there should not be a penalty to do so. this is one of the large misunderstandings when it comes to cash value life insurance. katt indicates that it should not be thought about as a bank account, but rather as more like an asset that can serve as collateral to be borrowed against. katt further explains a phenomenon known as phantom income, which occurs with the generation of taxable income from the lapsing of a policy that contains no cash value. repaying loans can often be the only option for a client that has an economic benefit since a lapse or surrender of a policy may result in a substantial taxable event. another potential issue can be in the form of surrender charges, which most forms of universal life policies have. while a surrender charge is also known as an early exit penalty, the calculation on an illustration is the difference between the accumulation value and the surrender value. where additional misconceptions and confusion can occur is if a client reduces the death benefit during the surrender period. even without a cash value withdrawal, the surrender charge still takes place for the reduced portion of the death benefit. there are primarily two different death benefit options for structuring a life insurance policy and the cash value accumulation plays an important role in both approaches. policies most commonly consist of a level death benefit in which the cash value accumulation does not increase the death benefit, but borrowing against or withdrawing the cash value decreases the death benefit. 226 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 the level death benefit approach helps mitigate the increasing cost of insurance structure by lowering the net at-risk amount to the insurance company. the net at-risk amount to an insurance company is the difference between the death benefit and cash value within the policy. cost of insurance increases as life expectancy decreases. the payment of death benefit proceeds exists when there is at least $1 of cash value and does not change if there is a much higher cash value amount. for this reason, in theory it would be beneficial not to accrue more than necessary. this would require knowing at what age a person will pass away, which is why typical policies project to age 100 or beyond. however, if there is a reduction in life expectancy because of major health concerns, a reduction in premiums can prevent unnecessary cash value accumulation. managing a life insurance policy around cash value accumulation and current life expectancy can provide maximum efficiency. the type of life insurance policy selected dictates the method and capacity for accumulation of cash value. each contains its own unique risks for accumulating cash. however, they all also contain another risk that could nullify the previously mentioned tax advantages of life insurance. if a policy is deemed to be a modified endowment contract (mec), the fifo account method is no longer available and will contain additional penalties for accessing the cash value before age 59 1⁄2. the determination comes from a “7 pay test” and looks to avoid excessive overfunding of a life policy up-front. this needs to be a consideration when cash accumulation is a goal. an additional risk exists even when there is the proper structure of a life insurance policy to avoid mec rules. commito (2012) concludes that a tax court and subsequently an appeals court both ruled that a case involving the lapse of a life insurance policy that contained significant policy loans resulted in tax due because of the significant difference in the loan balance and cost basis. the client argued that discharge of indebtedness should apply because of the loan outstanding being greater than the total net worth at the time of the taxable event, rendering the client insolvent. the tax court ruled differently, stating that this did not apply because the client was solvent at the time and loan repayment occurred from the policy values. this ruling now means that in all situations a taxable event will occur if a policy lapses with a loan balance greater than cost basis. further discussion continues about how managing risk is also crucial during retirement. it is more involved than simply accumulating enough and providing retirement income. an entire retirement plan can be at risk if additional unexpected medical expenses or an extended long-term care need occurs. unexpected market downturns, inflation, and increasing taxes all pose additional risks. the design of life insurance helps with transferring multiple retirement risks by creating the availability of additional resources and riders to provide further flexibility for policy owners. policy reprojections, also known as in-force illustrations, provide an understanding of policy longevity incorporating any market returns or interest crediting received up to the date of the reprojection. in-force illustrations have a similar layout to the originally provided sales illustration during policy purchase, but will project policy longevity based on the current interest crediting assumptions, which are usually different from the original. in-force illustrations are readily available from the insurance companies and help monitor the status of a policy to provide a more accurate estimate of income potential during retirement compared with the original projection. overloan protection riders help prevent policies from lapsing because of too many distributions, automati227r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 cally placing them in a paid-up status and avoiding the unexpected taxable event from occurring. other riders are also available for additional risk management against long-term care, chronic illness, and death benefit acceleration upon terminal illness. with a supplemental income objective, variable universal life is the type of policy that many investors gravitate toward because they can experience market exposure and upside potential within the tax efficient wrapper of life insurance. 3. methodology life insurance is a tax-advantaged vehicle that can generate supplemental income. discussions around the need for a tax advantaged vehicle during retirement such as life insurance, which offers tax deferred accumulation, along with the potential for tax-free distributions, warrants research to the viability of such a vehicle. however, discussions have also identified that when using life insurance with the intended purpose of supplemental income, agents often used too aggressive of a hypothetical accumulation rate as well as too aggressive of a distribution rate. this has resulted in large negative consequences for individuals with the misunderstanding of how the policy works and the taxable consequences should the strategy not work as originally designed. this research will first determine an appropriate hypothetical rate of return to avoid the overestimation issues that have occurred in the past. it will then provide a safe withdrawal rate in conjunction with the determined accumulation rate, once premiums have stopped, to prevent the withdrawal of too many distributions that causes the policy to lapse with a taxable event. utilizing these two determined rates jointly will alleviate the issues that have caused trouble for this strategy as noted in previous discussions. overall, this research will show the proper structure for these vehicles based on historical data and rolling period simulations. thus, this will enhance financial literacy on life insurance strategies by providing clarification on previous misuses and solutions on the correct structure for this strategy to be successful. this study will use a quantitative research approach using historical data, combined with other select variables to develop a new quantitative approach when assessing life insurance as a retirement asset. the goal is to determine a quantitative value for the accumulation rate and its associated distribution rate for a set of client circumstances. historical performance will determine the accumulation rate, which is the first part of the equation, as it does not depend on the distribution rate. however, the distribution rate does depend on the accumulation rate when generating the same amount of cash flow (higher accumulation rate, lower distribution rate, and vice-versa). cost of insurance is dependent upon several variables, including gender, age, health classification, and the amount of the death benefit. for this study, these arbitrarily selected variables are necessary to determine the cost of insurance component of the policy. the distribution rate will be dependent upon the duration of the distribution period, cost of insurance, and accumulation rate. another arbitrarily specified variable is the length of distributions. three different life insurance companies will provide a sample of costs of insurance. for the baseline, the study considers a 45-year old male client. starting ages of 35 and 55 will be used for additional scenarios. 228 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 for this research, we will also assume preferred non-smoker health for the base case, with an additional scenario utilizing a standard non-smoker health classification. preferred nonsmoker is often the second-best health rating given by insurance companies. typically, the best classification is uncommon and the goal of this research is to be able to have applicable findings for the widest possible audience. meanwhile, length of accumulation determines how long the assets will be able to accumulate before providing distributions, with longer periods allowing for greater potential asset growth. for this research, the base case will assume a 19-year accumulation period, which for the assumed 45-year-old individual, would put retirement age at 65. additional scenarios will include a 9-year accumulation period for a 55-year-old individual and a 29-year accumulation period for a 35-year-old individual. length of distributions determines how long assets will be able to support cash flow. this duration directly affects the amount of cash flow generation, along with a correlation of risk for the policy to lapse. for the base case, this research will assume a 15-year distribution period, from ages 65 through 79. the intention is for supplemental cash flow during the early part of retirement. this approach is not intended to provide the core source of retirement income over the individual’s lifetime. testing the combination of accumulation and distribution rates also includes monitoring the policy once distributions stop at age 79. the purpose of testing after distributions have ceased will be to make sure the policy remains in force and avoids lapsing for an additional 21 years until age 100, to avoid triggering a taxable event. additional scenarios will include distribution periods of 20 and 30 years. the amount of death benefit directly affects the cost of insurance. a higher death benefit leads to higher costs of insurance. for this research, the goal is to provide supplemental cash flow and because this is the focus, the study will solve for the minimum death benefit that does not trigger a mec. a minimum non-mec death benefit will minimize the cost of insurance, while also preserving the tax advantages of the life insurance policy. the minimum non-mec death benefit allows for overfunding of the policy, which helps alleviate concerns about policies becoming underfunded. the type of death benefit payout directly affects the cost of insurance based on the net at-risk amount to the insurance carrier. this study will assume an increasing (option 2) death benefit payout during the accumulation period, switching to a level (option 1) death benefit payout upon the start of the distribution period. an increasing death benefit is the total of a base death benefit amount plus the cash value, while a level death benefit does not include the cash value. the amount of premium paid directly affects with a positive correlation the amount of insurance, cost of insurance, and potential distributions. this research will assume annual premiums of $50,000. additional scenarios of $10,000 and $25,000 annual premiums will be tested to confirm the linear relationship between premiums and cash flow generation. historical rates of return in the equity markets over an extended period will help determine an accumulation rate suitable to use for variable universal life insurance contracts. the assumed portfolio will be 100% equity investments, using large-capitalization u.s. stocks. the cost of insurance has a negative correlation with the total potential for accumulation and distributions. age, health classification, and mortality tables determine the cost of insurance for each individual insurance company. this study will compare the cost of 229r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 insurance and fees from three separate insurance companies with the goal of providing an accurate representation of the market. it is important to note that the rates for cost of insurance are subject to increase up to a maximum allowable amount. an increase to the maximum allowable amount is historically unlikely, but increases and decreases have occurred. actuarial pricing and experience dictates cost of insurance, which may cause higher or lower costs for different ages. policy fees have a negative correlation to the total potential for accumulation and distributions. policy fees include premium loads, administrative charges, and death benefit charges. this study will compare the fees and cost of insurance from three separate insurance companies with the goal of showing an accurate representation of the market. this study looks to determine a suitable accumulation rate for variable universal life insurance contracts. this is the hypothetical rate of return to use for illustration purposes when someone is buying a new life insurance contract. a proper accumulation rate will avoid setting an unrealistic or too aggressive of a return expectation that results in the policy becoming underfunded and lapsing, if no corrective action occurs. this study also looks to determine a suitable decumulation rate for variable universal life insurance contracts. this is the hypothetical withdrawal rate to use for illustration purposes when someone is buying a new life insurance contract. a proper decumulation rate will avoid taking too many distributions from the policy and causing it to lapse. this study will assume that the distributions from the variable universal life policy will consist of both withdrawals and loans. withdrawals occur until the full recovery of costbasis, after which loans against the policy occur for remaining distributions. all loans will assume a contractual fixed loan interest rate of 3%. current pricing dictates the loan rate used, but has been higher historically and can be higher or lower in the future. while life insurance in general receives the tax preferential treatment as previously discussed, this research will focus on utilizing a 100% equity based variable universal life policy. the 100% equity based policy provides an equity alternative and minimizes fund expenses and variations within the allocation options between policies. the design of the policy for the hypothetical participants will be with the intended goal of providing supplemental retirement cash flow. with this goal, the vul policy can take on more risk as it is only a piece of the overall portfolio, rather than being the main source of retirement income. a key aspect of the policy design is increasing flexibility during retirement with the intention of being a complement to the rest of the portfolio and not a portfolio income replacement. this design will lead to providing the minimum death benefit allowed by irs standards that avoids becoming a mec and losing the previously discussed tax advantages upon distribution. the minimum death benefit design minimizes the insurance costs and looks to meet the stated objective of maximum supplemental cash flow from policy distributions. policy design is crucial to the success of utilizing life insurance as a supplemental cash flow vehicle. the minimum non-mec death benefit minimizes insurance costs and overfunds the policy as much as allowable by irs code while keeping the tax preferential treatment. the overfunding of the policy creates a larger margin of error before poor performance and underfunding jeopardize the policy. underfunded life insurance policies are the primary concern and pitfall experienced when trying to generate supple230 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 mental cash flow. in addition to the minimum non-mec death benefit, the type of death benefit payout is also crucial to the design. the structure of the death benefit payout will also help maximize the income generation and minimize the cost of insurance throughout the life of the policy. with the goal of showing an accurate representation of the marketplace, quotes from three separate insurance companies are used to determine the cost of insurance. volume of variable universal life insurance policies sold and financial strength ratings determine the companies selected. this will provide a good example of carriers likely to implement this strategy as described in the research. while variable universal life policies typically allow for the selection of many different subaccounts with varying equity and bond allocations, this study will assume an all equity allocation maximizing the potential upside accumulation and distributions. with the all equity allocation assumptions, this research will be using an aggressive strategy. for equity returns, this study uses robert shiller’s data set of historical returns from 1871 through 2015 for large-capitalization u.s. stocks (shiller, 2016). rolling historical periods created from the data determine the combined accumulation rate and distribution rates. statistical analysis will stress test various accumulation and distribution rates to determine the historical probability of success. this research will define successful combined accumulation and distribution rates as those that would have had a 95% or greater probability of success based on historical equity performance. testing the combination of accumulation and distribution rates will also include monitoring the policy after discontinuing distributions beyond age 79. the purpose of testing after distributions have ceased will be to make sure the policy remains in force an additional 21 years until age 100 and avoids lapsing, which would trigger a taxable event. the expectation of the statistical analysis are results showing that people implementing life insurance as a supplemental cash flow vehicle may have been too aggressive with the rate of return assumptions and the rates should be more conservative. to conduct this research, the recreation of a variable universal life insurance policy with different assumed starting points is necessary. first, multiple life insurance carrier illustrations with the same parameters show the different costs and charges within the policies. the hypothetical variable universal life policy assumes a 100% investment into a largecapitalization stock index fund. the aggressive allocation will facilitate equity portfolio comparisons while minimizing variations of fund costs. for the base case in this study, we assume the life of the policy will last from age 45 through age 100, with accumulation occurring during the first nineteen years and distributions occurring during the next fifteen years. beyond age 79, the policy contains no inflow or outflow of cash, but requires making sure the policy remains in-force and does not lapse because of insufficient funds potentially causing a taxable event. this means that policy simulations require 55 years of investment returns. the simulations for this research assume a starting point for each year from 1871 through 1961. the study creates 91 simulated life insurance policies, each with a different 55-year historical return scenario. compiling this data into a sample life policy allowed for the calculation of the maximum cash flow distribution rate to keep the policy in-force through age 100. 231r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 4. results table 1 demonstrates the “base” illustration using a specific carrier default 9.41% distribution rate with a selected 6.3% accumulation rate. the illustration shows the end of policy year age of the individual, amount of premium payments each year, the cash account balance, the administrative fees and charges within the policy, cost of insurance (coi), income, cost table 1 base variable universal life example with a 6.3% accumulation rate and insurance carrier software default 9.41% distribution rate age premiums cash value starting balance fees coi income cost basis loan balance cash value ending balance death benefit 46 $50,000 $ �$6,712 �$1,380 $ $50,000 $44,548 $1,088,117 47 $50,000 $44,548 �$4,712 �$1,510 $ $100,000 $93,891 $1,137,460 48 $50,000 $93,891 �$4,712 �$1,653 $ $150,000 $146,190 $1,189,759 49 $50,000 $146,190 �$4,712 �$931 $ $200,000 $202,551 $1,246,120 50 $50,000 $202,551 �$4,712 �$1,059 $ $250,000 $262,327 $1,305,896 51 $50,000 $262,327 �$4,712 �$1,174 $ $300,000 $325,747 $1,369,316 52 $50,000 $325,747 �$4,712 �$1,282 $ $350,000 $393,048 $1,436,617 53 $50,000 $393,048 �$4,712 �$1,400 $ $400,000 $464,463 $1,508,032 54 $50,000 $464,463 �$4,712 �$1,543 $ $450,000 $540,225 $1,583,794 55 $50,000 $540,225 �$4,712 �$1,719 $ $500,000 $620,573 $1,664,142 56 $50,000 $620,573 �$2,120 �$1,863 $ $550,000 $708,585 $1,752,154 57 $50,000 $708,585 �$2,120 �$2,139 $ $600,000 $801,849 $1,845,418 58 $50,000 $801,849 �$2,120 �$2,482 $ $650,000 $900,623 $1,944,192 59 $50,000 $900,623 �$2,120 �$2,841 $ $700,000 $1,005,239 $2,048,808 60 $50,000 $1,005,239 �$2,120 �$3,217 $ $750,000 $1,116,046 $2,159,615 61 $50,000 $1,116,046 �$2,120 �$3,773 $ $800,000 $1,233,242 $2,276,811 62 $50,000 $1,233,242 �$2,120 �$4,131 $ $850,000 $1,357,442 $2,401,011 63 $50,000 $1,357,442 �$2,120 �$4,516 $ $900,000 $1,489,056 $2,532,625 64 $50,000 $1,489,056 �$2,120 �$4,478 $ $950,000 $1,629,003 $2,532,625 65 $ $1,629,003 �$120 �$4,390 �$153,289 $796,711 $1,573,547 $2,388,120 66 $ $1,573,547 �$120 �$4,381 �$153,289 $643,422 $1,514,607 $2,234,876 67 $ $1,514,607 �$120 �$4,211 �$153,289 $490,132 $1,452,134 $2,081,632 68 $ $1,452,134 �$120 �$4,004 �$153,289 $336,843 $1,385,945 $1,928,388 69 $ $1,385,945 �$120 �$3,758 �$153,289 $183,554 $1,315,848 $1,775,144 70 $ $1,315,848 �$120 �$3,473 �$153,289 $30,265 $1,241,638 $1,621,900 71 $ $1,241,638 �$120 �$3,142 �$153,289 �$123,024 �$ 126,715 $1,159,414 $1,467,035 72 $ $1,159,414 �$120 �$2,734 �$153,289 �$ 288,405 $1,071,536 $1,307,582 73 $ $1,071,536 �$120 �$2,233 �$153,289 �$ 454,945 $978,653 $1,143,346 74 $ $978,653 �$120 �$1,664 �$153,289 �$ 626,481 $880,524 $974,182 75 $ $880,524 �$120 �$1,269 �$153,289 �$ 803,163 $776,633 $832,473 76 $ $776,633 �$120 �$1,397 �$153,289 �$ 985,146 $666,060 $723,789 77 $ $666,060 �$120 �$1,606 �$153,289 �$1,172,588 $548,299 $607,674 78 $ $548,299 �$120 �$1,842 �$153,289 �$1,365,654 $422,869 $483,615 79 $ $422,869 �$120 �$2,106 �$153,289 �$1,564,511 $289,255 $351,068 80 $ $289,255 �$120 �$2,401 $ �$1,611,446 $304,798 $368,050 — — — — — — — — — — 100 $ $800,873 �$120 $ $ �$2,910,451 $851,200 $851,200 this table utilizes the life insurance carrier default distribution rate of 9.41% with a user selected 6.3% accumulation rate showing the cash flow each year. displayed are the end of year age, premiums paid, starting cash value balance, deductions for policy fees and cost of insurance, income received, cost basis, loan balance upon recovery of the cost basis, ending cash value balance, and death benefit. the starting base death benefit amount for the policy is $1,043,569. not displayed but factored into the calculations are the annual investment returns. base case for a 45-year old preferred non-smoker male. 232 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 basis, loan balance, ending balance for the cash account, and death benefit. the age of the individual for the base case illustrations will remain between 45 and 100. the beginning and ending cash account balances are crucial to demonstrate the fluctuation of cash value and impact that the fees, insurance costs, investment returns, and income all have. if the policy has a positive cash value, the policy will remain in-force and not lapse, preventing a potential taxable event. if the insured dies, the proceeds of the death benefit pay off the loan and the beneficiary receives the net death benefit. if a death benefit payment occurs, regardless of amount, there is no taxable event. in this example, after distributions and loan balance growth, a death benefit of $851,200 remains at age 100. with the primary focus of this study being supplemental retirement cash flow, the death benefit aspects of the policy receive little focus. with the base example being setup utilizing a carrier default distribution rate of 9.41% and a selected 6.3% accumulation rate, further analysis and testing provided the impact of different accumulation and distribution rates. with the creation of and ability to manipulate a base illustration, the historical performance for each potential starting year determined the maximum distribution rate while keeping the policy from lapsing. with a goal of providing a safe assumed rate of return for both the overall accumulation and distribution periods of the life insurance policy based on historical data, this study will look at a 95% success rate for the combined accumulation and distribution rates. table 2 summarizes the success rate for each combination of an accumulation and distribution rate for the hypothetical variable universal life policy with accumulation rates from a selected range of 6% through 10% and distribution rates from a selected range of 8% through 11%, table 2 historical probability of success for joint assumptions with a 19-year accumulation period, 15-year distribution period, and non-lapse for an additional 21 years distribution rate 8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10 10.2 10.4 10.6 10.8 11 accumulate rate 6 100 100 100 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 92.3 6.2 100 100 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 92.3 90.1 6.4 100 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 90.1 85.7 6.6 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 87.9 84.6 79.1 6.8 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 87.9 83.5 78 65.9 7 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 87.9 83.5 76.9 63.7 53.8 7.2 100 100 98.9 95.6 95.6 94.5 94.5 93.4 92.3 91.2 87.9 82.4 74.7 61.5 52.7 50.5 7.4 100 100 95.6 95.6 94.5 94.5 93.4 92.3 91.2 87.9 82.4 72.5 60.4 52.7 50.5 48.4 7.6 100 95.6 95.6 94.5 94.5 93.4 92.3 90.1 86.8 80.2 67 59.3 50.5 49.5 48.4 46.2 7.8 96.7 95.6 94.5 94.5 93.4 92.3 90.1 86.8 79.1 65.9 57.1 50.5 49.5 48.4 46.2 44 8 95.6 94.5 94.5 93.4 92.3 90.1 85.7 79.1 65.9 56 50.5 48.4 48.4 46.2 41.8 40.7 8.2 94.5 94.5 93.4 92.3 90.1 85.7 79.1 65.9 53.8 50.5 48.4 47.3 45.1 41.8 39.6 37.4 8.4 94.5 93.4 92.3 90.1 85.7 79.1 64.8 53.8 50.5 48.4 47.3 45.1 41.8 38.5 37.4 34.1 8.6 93.4 92.3 90.1 85.7 79.1 63.7 53.8 50.5 48.4 47.3 45.1 41.8 37.4 36.3 34.1 34.1 8.8 93.4 91.2 85.7 79.1 63.7 53.8 50.5 48.4 47.3 45.1 41.8 37.4 35.2 34.1 34.1 33 9 91.2 86.8 79.1 63.7 53.8 50.5 48.4 46.2 44 40.7 37.4 35.2 34.1 34.1 31.9 30.8 9.2 87.9 79.1 64.8 53.8 50.5 48.4 46.2 44 40.7 37.4 34.1 34.1 33 30.8 30.8 30.8 9.4 79.1 65.9 53.8 50.5 48.4 46.2 44 39.6 37.4 34.1 34.1 33 30.8 30.8 30.8 28.6 9.6 65.9 53.8 50.5 48.4 46.2 42.9 39.6 37.4 34.1 34.1 33 30.8 30.8 30.8 28.6 27.5 9.8 53.8 50.5 48.4 46.2 41.8 39.6 37.4 34.1 34.1 33 30.8 30.8 30.8 27.5 27.5 27.5 10 50.5 48.4 46.2 41.8 39.6 37.4 34.1 34.1 31.9 30.8 30.8 30.8 27.5 27.5 27.5 25.3 base case for 45-year old preferred non-smoker male. 233r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 all in 0.2% increments. the combined accumulation and distribution rates are cash flow generation focused and would look different if the intended goal was to provide more consideration to the death benefit feature. across the range of accumulation and distribution rates, multiple combinations allowed for a 95% or greater historical success. the table shows that within the specified 6% to 10% accumulation rate range and 8% to 11% distribution rate range, the upper limit for the accumulation rate is 8%, with anything greater resulting in less than 95% success for the lowest distribution rate shown of 8%. an 8% accumulation rate combined with an 8% distribution rate results in a 95.6% probability of success. in addition, the table shows that within the specified ranges the upper limit for a distribution rate is 10%, with anything greater having a less than 95% success rate with the lowest accumulation rate shown of 6%. a 10% distribution rate results from the accumulation rate being 6%. it is also important to note that a combination of a 7.6% accumulation rate with an 8% distribution rate had a 100% historical probability of success, which represents the highest accumulation rate to do so. with a 6% accumulation rate, a 9.4% distribution rate also results in a 100% historical probability of success. upon further testing of the combined accumulation and distribution rates providing at least a 95% probability of success, the combination that generated the highest supplemental cash flow from the hypothetical variable universal life policy was a 7% accumulation rate with a 9% distribution rate. this combination supported distributions of $158,453. the vul rates have been determined with an approach like bengen’s (2004) safemax 4% rule approach, utilizing rolling historical equity returns to determine the maximum feasible numbers for a given historical probability of success. the base case scenario for this research utilizes $50,000 annual premiums, however, additional premium scenarios were tested for $10,000 and $25,000 (20% and 50% of the base case premiums, respectively). these scenarios resulted in 19.2% and 49.5% of the base case cash flow, respectively. the majority of the life insurance policy expenses are linear with only the monthly policy administrative expense being a fixed cost. this results in a fairly linear relationship between premiums and cash flow generation. in addition to the base scenario of this research study, involving a 45-year old male in preferred non-smoker health supplementing cash flow for 15 years, further analyses provide scenarios with the assumption of a standard non-smoker health classification as well as for varying accumulation and distribution periods. table 3 summarizes the success rate for each combination of an accumulation and distribution rate for the hypothetical variable universal life policy assuming a standard non-smoker health classification with accumulation rates from a selected range of 5% through 10% and distribution rates from a selected range of 7% through 10.75%, all in 0.25% increments. the change in health classification from preferred health to standard health resulted in a 0.50% decrease, 7% to 6.5%, in the accumulation rate with the same 9% distribution rate. this combination supports annual distributions of $145,210, a decrease from the $158,453 annual distributions with the assumed preferred health. for individuals in poor health, this may be a less viable strategy based on the higher costs of insurance. a standard health rating results in a �8.4% cash flow reduction within the given parameters, with further reductions expected for substandard health. 234 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 table 4 summarizes the corresponding accumulation and distribution rates for a 35, 45, and 55-year-old male in preferred non-smoker health, with 15, 20, and 30-year distribution periods. it is important to note that retirement stays at age 65, which is the first year a distribution occurs. shorter accumulation periods and longer income periods both work to reduce the combined feasible accumulation and distribution rates. to compare the life insurance design to a more traditional investment portfolio, examples will use the same specified parameters, including amount of investment, and length of accumulation and distribution periods. utilizing the same historical equity returns, a probability of success for generating the same after-tax distributions inside both a qualified and non-qualified account determines whether the tax advantages overcome the life insurance policy expenses. there are several different types of investment accounts, because they vary table 3 historical probability of success for joint assumptions with a 19-year accumulation period, 15-year distribution period, and 21 years of additional non-lapse for a 45-year old nonsmoker with a standard health classification distribution rate 7 7.25 7.5 7.75 8 8.25 8.5 8.75 9 9.25 9.5 9.75 10 10.25 10.5 10.75 accumulate rate 6 100 100 100 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 6.25 100 100 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 87.9 6.5 100 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 87.9 80.2 6.75 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 85.7 79.1 67 7 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 84.6 79.1 63.7 52.7 7.25 100 100 100 100 100 95.6 94.5 94.5 93.4 93.4 91.2 83.5 75.8 63.7 51.6 49.5 7.5 100 100 100 100 95.6 94.5 94.5 93.4 93.4 90.1 83.5 75.8 61.5 51.6 49.5 47.3 7.75 100 100 100 97.8 94.5 94.5 93.4 93.4 90.1 83.5 73.6 58.2 51.6 49.5 47.3 45.1 8 100 100 97.8 95.6 94.5 93.4 93.4 90.1 83.5 71.4 57.1 51.6 49.5 47.3 44 40.7 8.25 100 97.8 95.6 94.5 93.4 93.4 90.1 83.5 71.4 57.1 50.5 49.5 47.3 42.9 38.5 37.4 8.5 98.9 95.6 94.5 93.4 93.4 91.2 83.5 71.4 57.1 50.5 48.4 47.3 42.9 37.4 37.4 34.1 8.75 95.6 94.5 93.4 93.4 91.2 83.5 71.4 57.1 50.5 48.4 47.3 41.8 37.4 37.4 34.1 34.1 9 94.5 94.5 93.4 91.2 83.5 72.5 57.1 50.5 48.4 46.2 41.8 37.4 35.2 34.1 33 30.8 9.25 94.5 93.4 91.2 84.6 75.8 57.1 50.5 48.4 46.2 40.7 37.4 35.2 34.1 33 30.8 30.8 9.5 93.4 91.2 85.7 75.8 58.2 50.5 48.4 46.2 40.7 37.4 35.2 34.1 33 30.8 30.8 28.6 9.75 92.3 87.9 79.1 60.4 50.5 48.4 46.2 40.7 37.4 34.1 34.1 31.9 30.8 30.8 27.5 27.5 10 89 79.1 63.7 51.6 48.4 46.2 40.7 37.4 34.1 34.1 30.8 30.8 30.8 27.5 27.5 26.4 table 4 joint assumptions summary for accumulation and distribution rates for different ages and distribution lengths based on retirement at age 65 age 35 (29-year accumulation) 45 (19-year accumulation) 55 (9-year accumulation) income duration 15 7.50% 8.75% 7.00% 9.00% 6.75% 7.75% 20 7.50% 7.75% 7.00% 7.75% 7.00% 6.75% 30 7.50% 6.75% 7.00% 6.50% 6.75% 6.00% accumulation/distribution rates. each scenario assumes retirement at age 65 which is the first distribution year. all scenarios continue to monitor and avoid a policy lapse through age 100. 235r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 in the application of taxation. non-qualified (taxable) accounts can often benefit from reduced taxation on dividends and long-term capital gains. qualified accounts benefit from deferring all taxes until distribution. however, upon distribution all income is subject to taxation as ordinary income. tax-free investment accounts, such as a roth-ira, receive the same potential tax-preferential treatment as life insurance with tax-deferred growth and then tax-free distributions. the roth ira is the best investment account from a taxation standpoint and therefore does not require further analysis. however, it may not be available for all individuals based on its income and contribution limits. this study compares life insurance to both a qualified account and a non-qualified account. the non-qualified account will assume on-going taxation of historical annual dividends, with the tax-deferral of price appreciation until withdrawals begin. during withdrawals, taxation occurs upon recovery of the cost basis. while it is not entirely possible to simply withdraw the cost basis first within a non-qualified account, based on the average cost basis, after adjusting for dividends, the first eight distributions are tax-free, and then the remaining seven distributions are taxable. a common approach is to use the annual taxation method for non-qualified accounts, however, it is very unlikely that a non-qualified account will have turnover every year such that there would be annual taxation on the price appreciation. multiple scenarios of varying taxation rates on the annual dividend and distributions demonstrate the relationship between higher taxation and probability of success. taxation rates will range from 15% to 45%, even though the current maximum taxation on dividends and long-term capital gains is 23.8%. the higher taxation scenarios demonstrate potential increases in current taxation levels. table 5 summarizes the probability of success that a non-qualified account with various dividend and long-term capital gains taxation rates can generate the same $158,453 after-tax cash flow as the variable universal life insurance policy for a preferred non-smoker. as taxation increases, the probability of success decreases. while the variable universal life insurance policy has a 95.6% probability of success, the non-qualified account has a lower probability of success with any assumed dividend and long-term capital gains tax rate greater than 15%. it should be noted that one bias in using historical data ranging from 1871 to 2015, is that the historical average dividend yield on stocks was higher in the earlier years than in more recent times. therefore, this procedure will overstate the gains associated with a non-qualified account. table 5 also summarizes the probability of success that a nontable 5 probability of success for non-qualified account to generate matching after-tax income with various tax rates dividend and ltcg tax rates probability of success: matching $158,453 spending for preferred non-smoker probability of success: matching $145,210 spending for standard non-smoker 15% 95.6% 100.0% 20% 90.0% 95.6% 23% 84.4% 93.3% 30% 44.4% 77.8% 35% 40.0% 46.7% 45% 36.7% 41.1% 236 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 qualified account can generate the same $145,210 after-tax cash flow as the variable universal life insurance policy when assuming standard non-smoker health. with a reduced stress on the portfolio needing to generate less cash flow, the probability of success increases. however, any assumed tax rate greater than 20% results in a lower probability of success than the variable universal life insurance policy. depending on the assumed tax rate, there is a lower probability of success for generating the same after-tax cash flow as the vul policy. however, that does not mean the vul policy is a replacement for taxable investment accounts, as there are significant differences. most notable is the cash value within each vehicle after the specified cash flow period has ended. the life insurance policy will have cash value that may be accessible for additional distributions, but it is important not to overdraw the policy and cause it to lapse, triggering a taxable event. therefore, most of the cash value remaining is not accessible. there is a death benefit remaining within the vul policy, however, that does not provide additional resources to the insured, only the beneficiary, which will have varying levels of importance depending on the individual. this differs with the non-qualified account because any additional remaining value has full accessibility should the individual choose to use it, which provides additional economic benefit. the qualified account scenarios reflect that all price appreciation and dividends are tax deferred until distribution, at which time the full distribution receives ordinary income taxation treatment. because of the tax deductibility of qualified accounts, the annual investments are grossed-up based on an assumed preretirement tax bracket. nine total scenarios demonstrate varying preand post-retirement tax brackets ranging from 35% to 55%. table 6 summarizes the probability of success that a qualified account with various preand post-retirement ordinary income taxation rates can generate the same $158,453 after-tax cash flow as the variable universal life insurance policy assuming preferred non-smoker health. within the nine scenarios, only two have a lower probability of success than the vul policy, both of which have a higher ordinary income taxation rate during retirement compared with the preretirement taxation rate. one of the compelling arguments for the life insurance policy is protection against a rising tax environment. however, if tax rates do not rise substantially the qualified plan is clearly more beneficial. with six of the nine scenarios table 6 probability of success for qualified account to generate matching after-tax income with various tax rates pre-retirement tax bracket grossed-up investment retirement tax bracket probability of success: matching $158,453 spending for preferred non-smoker probability of success: matching $145,210 spending for standard non-smoker 35% $76,923 35% 100.0% 100.0% 35% $76,923 45% 95.6% 100.0% 35% $76,923 55% 55.6% 90.0% 45% $90,909 35% 100.0% 100.0% 45% $90,909 45% 100.0% 100.0% 45% $90,909 55% 94.4% 100.0% 55% $111,111 35% 100.0% 100.0% 55% $111,111 45% 100.0% 100.0% 55% $111,111 55% 100.0% 100.0% 237r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 showing a higher probability of success within the qualified plan compared with the vul, individuals should maximize their available qualified plans first before utilizing a life insurance policy as a retirement vehicle. even in the scenario of a modest increase in taxation from 35% to 45%, the qualified plan has the same probability of success as the vul. with the vul also incurring more risk than the qualified plan, the vul should not be viewed as a replacement, but rather a supplement when additional tax deferral is desired. table 6 also summarizes the probability of success that a qualified account with various preand post-retirement ordinary income taxation rates can generate the same $145,210 after-tax cash flow as the variable universal life insurance policy assuming standard nonsmoker health. with a reduced stress on the portfolio needing to generate less cash flow, the probability of success increases, resulting in only one scenario of the nine that has a lower probability of success than the vul policy. the lone scenario with a lower probability of success results from a 20% increased tax rate during retirement. as previously noted with the non-qualified account, the qualified account would also provide additional flexibility to withdraw and deplete any remaining values if an individual chooses to do so. however, any additional distributions are taxed at ordinary income taxation rates. 5. conclusions this study has determined that for the given parameters of the base case, the safe combined accumulation and distributions rates are 7% and 9%, respectively. these rates assume a 100% equity allocation towards u.s. large capitalization stocks. although other combinations of rates provided a 95% or higher probability of success based on historical returns, this combination yielded the highest supplemental cash flow with a minimum of 95% probability of success. these parameters demonstrate success based on historical performance, but in practice it is crucial to monitor the policy performance on an on-going basis. one of the largest benefits of a variable universal life policy is its flexible nature, which allows for adjustments to be made. this will allow higher or lower distributions to occur based on actual performance. this study has also determined that if an individual does not qualify for preferred health, but standard health, the safe accumulation rate based on historical returns reduces from 7% to 6.5% with the same 9% distribution rate, which results in a �8.4% decrease in income with the given parameters. this research shows that while life insurance can generate cash flow, there are additional risks that could result in a policy lapsing and triggering a taxable event. if an individual is in a lower tax bracket during retirement, the tax-preferential treatment would be less advantageous. the higher the taxation assumptions, the more competitive the life insurance options become on an after-tax basis because of increased efficiency. however, even in the scenarios where the life insurance policy provides a higher probability of success compared with the traditional investment accounts, the residual value after the distribution period has limited use. this is a result of the need to maintain the policy in-force to avoid taxation. the investment accounts do not impose such limitations and would provide full access to any residual values available beyond the distribution period. while the insurance policies have a tax-free death benefit, it is for a beneficiary and not the insured. 238 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 life insurance when used as a supplemental cash flow vehicle needs continuous monitoring. using life insurance as an asset class should be a supplement to an existing retirement income portfolio and, therefore, is not suitable for every individual. life insurance can add benefits for individuals in higher income tax brackets, but for individuals in lower income tax brackets or individuals who do not have other income solutions, life insurance may not be appropriate. the reason it is not appropriate for individuals without other income solutions is that life insurance should not be the primary source of retirement income. in addition, while the vul policy is flexible in premiums, concerns could grow if the ability to make premium payments is no longer possible. while corrective action may be available, such as reducing the death benefit or surrendering the policy, the payment for costs of insurance may unnecessarily occur. while this study focused on utilizing life insurance as a retirement vehicle, it is not appropriate for everyone. life insurance is illiquid for generating cash flow and is a long-term time horizon vehicle. many policies have a fee or penalty associated with liquidating the contract during the early years. in addition, because of embedded fees, it often takes 10 years or more before gains appear within the contract. for these reasons, life insurance should be a supplement to the overall portfolio for individuals with a long-term time horizon before requiring income. individuals considering this strategy should also be healthy to help balance the costs of insurance associated with the death benefit of the policy. individuals who have other retirement income vehicles, are in a higher tax bracket, or believe that they will be in the future, and can monitor the policy on an on-going basis, may want to consider utilizing life insurance as a supplemental cash flow vehicle. the purpose of this research is to show that life insurance is usable as a retirement vehicle when properly structured. the tax preferential treatment provided to life insurance allows a consumer to have greater flexibility over which dollars to use during retirement. the cash accumulation within the policy grows on a tax-deferred basis and then upon retirement, access to the cash value can occur on a tax-free basis. the life insurance policy provides an additional option, like the current roth ira, but without funding and income limitations. this research focused on a specific set of parameters to provide supplemental cash flow to a retirement portfolio. the tax advantages of life insurance as a cash flow vehicle increase for individuals who are in a higher income tax bracket or individuals who are looking to protect against future increases in ordinary income tax rates. tax deferral has been shown to be beneficial, but some individuals may have limited access to qualified investment accounts. life insurance removes limitations to contribution limits. uncertainty around future taxation or availability for long-term capital gains treatment may provide reason to look at other vehicles to supplement a portfolio. references bengen, w. (2004). determining withdrawal rates using historical data. journal of financial planning, 17, 64–74. bengen, w. (2006). conserving client portfolios during retirement. denver, co: fpa press. commito, t. (2012). court of appeals discusses “discharge of indebtedness” income when a policy lapses. journal of financial service professionals, 66, 14–15. 239r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 katt, p. (2009). managing life insurance: loans or withdrawals. journal of financial planning, 22, 38–39. katt, p. (2013). a practical understanding of cash value. journal of financial planning, 26, 40–41. katt, p. (2015). history of cash value life insurance and implications for existing policies. journal of financial planning, 28, 33–34. kriesel, w. (2010). creative uses of insurance. cpa journal, 80, 50–56. mccarthy, e. (2011). with higher taxes looming, are your clients ready? journal of financial planning, 24, 22–26. parish, s. (2014). the use of life insurance in retirement income risk planning. journal of financial service professionals, 68, 33–35. resnick, j., & resnick, b. (2009). understanding the policy details of no lapse guarantee and vul. journal of practical estate planning, 11, 23–30. shiller, r. (2016). online date robert shiller (available at http://www.econ.yale.edu/�shiller/data.htm). silver, g. (2013). tax diversification in retirement planning. journal of financial services professionals, 67, 49–55. steiner, k. (2014). a better systematic withdrawal strategy: the actuarial approach. journal of personal finance, 13, 51–56. sumutka, a., sumutka, a., & coopersmith, l. (2012). tax-efficient retirement withdrawal planning using a comprehensive tax model. journal of financial planning, 25, 41–52. 240 r. delibero, w.d. pfau / financial services review 26 (2017) 221–240 academy of financial services officers president william chittenden texas state university president-elect thomas coe quinnipiac university executive vice president-program robert moreschi virginia military institute vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2016 swarn chatterjee university of georgia vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president lance palmer university of georgia editor, financial services review stuart michelson stetson university directors sherman hanna ohio state university halil kiymaz rollins college frank laatsch univ. of southern mississippi david nanigian the american college tom potts baylor university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake past presidents lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2015 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 24, no. 3, 2015 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2015 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. finser_30_4_merged_updated_3 officers president inga timmerman academy of financial services halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 california state university, northridge president-elect executive vice president-program terrance k. martin winston-salem state university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review terrance k. martin winston-salem state university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. institutional membership to academy of financial services is $250 and includes the four annual issues of fsr. how to submit: there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academyfinancial.org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor terrance k. martin at martintk@wssu.edu. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2022 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 30, no. 4, 2022 editor terrance k. martin, winston-salem state university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $250. academic subscription price is $125 and is available by joining the academy of financial services. academic subscription includes four emailed online issues of financial services review each year. hardcopies of the journal are available at additional cost. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. editorial office: terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. advertising information. those interested in advertising in the journal should contact terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu printed in the usa © 2022 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president robert moreschi virginia military institute president-elect duncan williams western carolina university executive vice president-program swarn chatterjee university of georgia vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president william chittenden texas state university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama halil kiymaz rollins college frances lawrence louisiana state university tom potts baylor university janine scott massey university martin seay kansas state university past presidents thomas coe, 2015-2016 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2017 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 26, no. 3, 2017 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2017 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. financial literacy and financial behavior: assessing knowledge and confidence colleen tokar asaada,* aschool of business, baldwin wallace university, berea, oh 44017, usa abstract this article explores how financial literacy, comprised of both actual financial knowledge and perceived financial confidence, affect financial decisions. using national survey data from the united states, results indicate that financial confidence is a critical component of financial literacy and is important across all knowledge levels. however, overconfident individuals, or those with high confidence (or self-assessed) knowledge but low actual knowledge, have a higher propensity to engage in risky (costly) financial behaviors. together, results suggest that financial literacy initiatives should focus not only on factual knowledge, but on helping individuals achieve a healthy dose of confidence. © 2015 academy of financial services. all rights reserved. jel classification: d03; d14; d80 keywords: financial literacy; confidence; overconfidence; financial behaviors; risk 1. introduction financial literacy is a measure of the degree to which one understands key financial concepts and possesses the ability and confidence to manage personal finances through appropriate short-term decision making and sound, long-range financial planning, while mindful of life events and changing economic conditions. (remund, 2010, p. 284) financial decision making is an essential component of day-to-day life, from minor decisions such as deciding whether or not to purchase a latte to major decisions such as taking on a home mortgage. several definitions of financial literacy highlight that to make sound financial decisions, individuals must not only possess the necessary knowledge, but * corresponding author. tel.: �1-440-826-2392; fax: �1-440-826-3868. e-mail address: colleen.tokar@gmail.com financial services review 24 (2015) 101–117 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. must also have “the ability and confidence” to apply their knowledge. this article explores how financial literacy influences financial behaviors. by examining two components of financial literacy, financial knowledge, and financial confidence (or perceived knowledge), this article demonstrates that both components are critically important to sound decision-making. using survey data from finras 2012 national financial capability study (nfcs), financial knowledge is measured by the number of correct answers to multiple-choice and true or false questions. financial confidence reflects a self-assessed level of financial knowledge, which may or may not coincide with measured financial knowledge. this article demonstrates that both knowledge and confidence influence financial behaviors, and surprisingly, the effect of financial confidence on behaviors is just as important as the effect of financial knowledge. furthermore, confidence is an important predictor of financial behavior across all actual financial knowledge level groups. additionally, by examining the interaction of financial knowledge and confidence, this study expands on the literature related to overconfidence, or the tendency to overestimate one’s accuracy and to underestimate risk. in instances where confidence exceeds actual knowledge (i.e., overconfidence), an individual has a greater likelihood of engaging in risky (costly) financial behaviors, such as taking out a title-loan. a key contribution, therefore, is a better understanding of how confidence influences financial behaviors: confidence is good, but not if it greatly exceeds actual knowledge. prior research associating perceived knowledge with individual financial behaviors has failed to reconcile instances in which inaccurate self-assessments can be harmful. this article shows that, overall, positive illusions are good. however, this article also illuminates the particular risky situations in which overconfidence is self-injurious. these findings are relevant across a multitude of disciplines and are pertinent to individuals, practitioners, and institutions alike. there are clear implications for financial literacy initiatives, initiatives that are of utmost importance given the pervasiveness of financial decisions in every individual’s daily life. the article is organized as follows. section 2 provides an overview of the literature, touching on financial literacy, perceived knowledge, and overconfidence. section 3 details the hypotheses, as well as an overview of the data, measures, and methods. section 4 presents the results and section 5 concludes. 2. literature review 2.1. financial literacy mandell (2008, p. 257) describes financial literacy as “the ability of consumers to make financial decisions in their own best shortand long-term interests.” at its most basic level, “financial literacy relates to a person’s competency for managing money” and “is typically measured at the individual level and then aggregated by groups” (remund, 2010, p. 279). because of the changing economic environment (e.g., see organisation for economic c-operation and development [oecd], 2005), financial literacy initiatives have received much attention. 102 c.t. asaad / financial services review 24 (2015) 101–117 research suggests that financial education has a positive effect on financial behaviors: education programs and seminars affect savings and total financial wealth (lusardi, 2004), and individuals who studied economics or business in high school are less likely to be unbanked (bernheim, garrett, and maki, 2001; grimes, rogers, and smith, 2010). however, other research questions the effectiveness of financial literacy initiatives: educating employees about the risks of employer stock does not significantly affect 401k holdings (choi, laibson, madrian, and metrick, 2005) and high school students who complete a semester of a financial literacy course are no more financially literate than high school students who have not completed the course (mandell and klein, 2009). u.s. households with higher levels of knowledge engage in more financial planning, although this positive relationship is weak (alhenawi and elkhal, 2013). if financial knowledge is not enough, what other factors influence the financial-decision making process? this article explores how a specific cognitive element, perceived knowledge (or financial confidence), shapes financial behaviors. 2.2. perceived knowledge (confidence) researchers often emphasize what people actually know at a given time, yet understanding perceptions is also important. park, gardner, and thukral (1988) emphasize that perceived knowledge is related to cognitive functioning, including recognition (schachter, 1983), identification (nelson, gerler, and narens, 1984), and problem solving (metcalfe, 1986). both an individual’s actual financial knowledge and perceived financial knowledge influence investments (kyrychenko and shumb, 2009), retirement planning (parker, bruin, yoong, and willis, 2011), and credit card behaviors (allgood and walstad, 2013). furthermore, carpena, cole, shapiro, and zia (2011) emphasize that aside from numeracy based knowledge, financial literacy may also affect decisions through an individual’s increased awareness and initiative. thus, financial confidence is a critical component of financial decision making. often there is a discrepancy between an individual’s actual knowledge and an individual’s self-perception, or confidence. correlations between actual and perceived financial knowledge vary considerably on an individual basis (agnew and szykman, 2005). it is interesting to look at the interactions and differences between these two measures of knowledge, specifically in situations where confidence exceeds actual knowledge. 2.2.1. overconfidence overconfidence refers to an individual’s propensity to overestimate the accuracy of his or her estimates, meaning that there is “a positive difference between assessed confidence and observed achievement” (campbell, goodie, and foster, 2004, p. 299). such overestimation, is more likely to occur “after unexpectedly difficult tasks” (healy and moore, 2007, p. 4). for example, less skilled financial planners are more confident than the more skilled (cordell, smith, and terry, 2011). individuals who are overconfident have narrow confidence intervals and, therefore, tend to overestimate precision and underestimate risk (goel and thakor, 2008). often, those who take more risk are not necessarily risk-seeking, but are less aware 103c.t. asaad / financial services review 24 (2015) 101–117 of the risk (simon, houghton, and aquino, 2000). even when given high incentives for accuracy, individuals still exhibit overconfidence (williams and gilovich, 2008). in a theoretical model, goel and thakor (2008) posit that ceo overconfidence effects firm value nonmonotonically, meaning that overconfidence is good up to a point (to overcome initial risk aversion) but then is harmful (leading to excessive risk-taking). likely, these findings will hold for the individual financial decisions considered in this analysis: that overconfidence is “good” for most financial decisions, but overconfidence is harmful for risky financial decisions. as johnson and fowler (2011, p. 320) warn, “it seems that we are likely to become overconfident in precisely the most dangerous of situations.” 3. method 3.1. hypotheses and research approach hypothesis 1 (h1): financial confidence predicts financial behavior. because cognitive functioning is a critical component of the decision making process (see section 2.2.), financial confidence likely influences all types of financial decisions. in fact, carpena et al., (2011) find that financial education initiatives do not equip individuals with the numeracy knowledge needed to make complex financial decisions; however, the initiatives greatly affect awareness and familiarity with financial services and products. thus, confidence is likely an important component of financial decision making. logistic regressions explore whether financial confidence affects financial behaviors above and beyond the influence of actual (measured) financial knowledge. hypothesis 2 (h2): overconfident individuals (high confidence, low knowledge) are most likely to engage in risky financial behaviors. overconfidence, or “that upward gap between what we know and what we think we know” (cordell, smith, and terry, 2011, p. 255), results in an overestimation of accuracy and underestimation of risk (see section 2.2.1.). based on goel and thakor’s (2008) theoretical model, overconfidence may be beneficial to overcome initial risk aversion, but also leads to excessive risk-taking. those individuals who self-assess their financial knowledge as higher than their actual knowledge may improperly assess risk levels, resulting in risky financial behaviors. it is hypothesized, then, that in most circumstances higher levels of confidence leads to “better” financial behaviors. however, too much confidence may be harmful in riskier circumstances because the risk level is not properly assessed. logistic regressions explore whether overconfidence leads to an increased propensity to engage in risky (or costly) financial behaviors. 3.2. data data are obtained from the 2012 national financial capability study (nfcs) commissioned by the financial regulatory authority’s (finra) investor education foundation. the study, with support from the u.s. department of the treasury and the president’s 104 c.t. asaad / financial services review 24 (2015) 101–117 advisory council on financial literacy, aims to measure american’s money skills. the state-by-state survey collected data from 25,509 respondents via an online survey. all analyses are weighted based on national distributions within age/gender, ethnicity, education, and census division. table 1 provides descriptive statistics for the survey sample. the sample is �51% female, 67% white, 53% married, and over 62% has an education beyond high school. 3.3. measuring financial literacy 3.3.1. financial knowledge five survey questions are used to measure basic financial knowledge. table 2 documents the five questions and the survey results. the dummy variables interest, inflation, bond, mortgage, and risk are created whereby a 1 represents a correct response and a 0 represents an incorrect response, a “don’t know,” or a refusal to answer.1 approximately 14% of the respondents answered all five financial literacy questions correctly. consistent with previous findings the results point to differences in knowledge across gender and race (fisher, 2010; lusardi, 2008; mandell, 2006).2 these gender differtable 1 characteristics of the sample n % total sample 25,509 100.0 gender/sex male 12,392 48.6 female 13,117 51.4 age 18–24 3,139 12.3 25–34 4,669 18.3 35–44 4,171 16.3 45–54 5,005 19.6 55–64 4,569 17.9 65� 3,956 15.5 ethnicity/race white 16,956 66.5 non-white 8,553 33.5 marital status married 13,782 53.4 single 7,469 28.2 separated or divorced 10,899 14.0 widowed 985 4.4 education not complete high school 2,210 8.7 high school graduate 5,695 22.3 ged 1,818 7.1 some college 9,160 35.9 college graduate 4,105 16.1 post-graduate education 2,519 9.9 data was obtained from 2012 finras national financial capability study and is weighted based on national distributions within age/gender, ethnicity, education, and census division. 105c.t. asaad / financial services review 24 (2015) 101–117 ences hold across age groups (lusardi and mitchell, 2007; lusardi, mitchell, and curto, 2010) and cross-nationally (lusardi and mitchell, 2011). also consistent with the literature, there is an inverted u-shape relation between age and financial literacy and a positive relation between education levels and literacy levels (lusardi and mitchell, 2011). 3.3.2. financial confidence (or perceived knowledge) in the nfcs survey, participants rated their own financial knowledge on a 7-point likert item scale whereby a “1” reflects low self-assessed levels of financial knowledge and a “7” reflects high self-assessed levels of financial knowledge. the question, as presented in the survey, reads, “on a scale from 1 to 7, where 1 means very low and 7 means very high, how table 2 financial knowledge question responses (n) percentage interest: suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow? more than $102 19,112 74.9 exactly $102 1,906 7.5 less than $102 1,407 5.5 don’t know 2,818 11.0 prefer not to say 266 1.0 inflation: imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account? more than today 2,201 8.6 exactly the same 2,176 8.5 less than today 15,630 61.3 don’t know 5,164 20.2 prefer not to say 334 1.3 bond: if interest rates rise, what will typically happen to bond prices? they will rise 5,014 19.7 they will fall 7,168 28.1 they will stay the same 1,290 5.1 there is no relationship� 2,186 8.6 don’t know 9,545 37.4 prefer not to say 306 38.6 mortgage: a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. true 19,142 75.0 false 2,303 9.0 don’t know 3,882 15.2 prefer not to say 182 0.7 risk: buying a single company’s stock usually provides a safer return than a stock mutual fund. true 2,209 8.7 false 12,366 48.5 don’t know 10,715 42.0 prefer not to say 219 0.9 the five financial literacy questions as they appear in finras 2012 national financial capability study. correct answers are italicized. binary variables were created for each of the five literacy topics (interest, inflation, bond, mortgage, and risk) whereby a correct response is coded as 1. if the respondent incorrectly answered the question, responded “don’t know” or “prefer not to say” the answer was coded as not correct (or 0). 106 c.t. asaad / financial services review 24 (2015) 101–117 would you assess your overall financial knowledge?” on average, individuals rated their financial knowledge (overall) as 5.15 on a 7-point scale. table 3 shows that only 9.4% of individuals self-assessed their knowledge level as below average, whereas 15.3% of individuals rated their knowledge as average and 75.2% rated their knowledge as above average. two additional questions, measured on a 7-point scale, assess confidence levels: how strongly do you agree or disagree with the following statements? y i am good at dealing with day-to-day financial matters, such as checking accounts, credit and debit cards, and tracking expenses. y i am pretty good at math. as seen in table 3, the majority of individuals rate themselves as better than average: the mean and median responses are all above four. robb, babiarz, and woodyard (2012) measure financial confidence as an average of the confidence responses. following their method, an “average” financial confidence measure is created (average) as the mean of the three responses (overall, day-to-day, math).3 3.3.3. knowledge and confidence allgood and walstad (2013) develop a measure that accounts for both an individual’s actual financial knowledge and perceived financial knowledge, arguing that the combination provides “more robust and nuanced insights” about how financial literacy affects financial table 3 financial confidence overall day-to-day math n % n % n % 1–very low/strongly disagree 500 2.0 947 3.8 1,177 4.7 2 506 2.0 574 2.3 704 2.8 3 1,329 5.4 865 3.4 991 3.9 4–average/neither agree or disagree 3,792 15.3 3,207 12.8 2,924 11.6 5 8,461 34.2 3,287 13.1 3,403 13.5 6 6,652 26.9 5,698 22.7 5,896 23.4 7–very high/strongly agree 3,479 14.1 10,518 41.9 10,099 40.1 mean 5.15 5.65 5.57 median 5.00 6.00 6.00 standard deviation 1.30 1.60 1.70 three questions in the nfcs survey touch on financial confidence: (1) “on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall financial knowledge?” (overall); “how strongly do you agree or disagree with the following statements?” (2) “i am good at dealing with day-to-day financial matters, such as checking accounts, credit and debit cards, and tracking expenses.” (day-to-day); (3) “i am pretty good at math.” (math) respondent’s answering “don’t know” for overall financial knowledge (528, or 2.1%); day-to-day matters (204, or 0.8%); and math (136, or 0.5%). respondent’s that “prefer not to say” for overall financial knowledge (790, or 3.1%); day-to-day matters (210, or 0.8%); and math (179, or 0.7%). correlations between the confidence measures of 58.2% (day-to-day and math), 41.8% (day-to-day and overall), and 36.0% (overall and math). all correlations significant at the 1% level. an average financial confidence measure (average) is also created and represents the average of responses to the three confidence questions (overall, day-to-day, and math). 107c.t. asaad / financial services review 24 (2015) 101–117 behavior. following their methodology, a composite knowledge measure is created. first, “high” and “low” groups are established for both financial knowledge and confidence, where those individuals with above average scores are categorized as “high” and those individuals with below average scores are categorized as “low.”4 then four additional variables are created to represent the four types of combined (knowledge-confidence) financial literacy (high-high, high-low, low-high, and low-low). about 28% and 20% of the sample are classified as high-high and low-low literacy, respectively. approximately one-third of the sample has high-low literacy and about 10% of the sample is overconfident (low-high knowledge). 3.4. measuring financial behaviors the nfcs surveys numerous financial topics. the behaviors considered in this analysis are detailed in the appendix. although it is not appropriate to label financial behaviors with normative values (e.g., “good” and “bad” behaviors ultimately depend on individual preferences and circumstances), it is possible to discern whether or not an individual is engaging in a “good” or “recommended” financial practice. for example, checking your credit rating is a “good” financial practice while being involved in a foreclosure process is a “bad” financial practice. applying goel and thakor’s (2008) theory, higher levels confidence will lead to “better” financial decisions, except in the riskiest of circumstances. financial behaviors classified as “risky” are costly behaviors that are riskier and pricier than other short-term loan alternatives, including: taking out an auto title loan; taking out a short-term payday loan; receiving a tax advance on a refund; using a pawn shop; and using a rent-to-own facility. if overconfident individuals underestimate risk, they will be more likely than other groups to engage in these costly behaviors that jeopardize their resources. 4. results logistic regressions explore how the two components of financial literacy influence financial behaviors. in logistic regressions, the dependent variable is a binary variable and the models attempt to predict whether or not an individual engages in a specific type of financial behavior. the � coefficients are difficult to interpret; therefore, the odds ratio, a more useful measure of effect size, is reported. the measure is the ratio of the likelihood of an event occurring in one group to the likelihood of the same event occurring in another group. in addition to the four financial knowledge groups, several other demographic factors are considered (almost all of which are dummy variables), including: gender (female � 1); age (18–24, 24–34, 35–44, 45–54, 55–64, or 65�); race (non-white � 1); education (�high school, ged, high school graduate only, some college, college graduate only, or postgraduate); employment status (self-employed, full-time, part-time, homemaker, student, unable, unemployed, or retired); marital status (single, married, divorced or separated, or widowed or widower); dependent children (no children or dependent children, one child, two children, or three or more children); annual income (less than $15k, $15–25, $25–35, $35–50, $50–75, $75–100, $100–150, or $150k or more); income-drop (� 1 if “experienced a large 108 c.t. asaad / financial services review 24 (2015) 101–117 drop in income which you did not expect” in the past 12 months); and risk tolerance (scale from 1 to 10 whereby 1 means “not at all willing” to take risks with financial investments and 10 means “very willing”). the use of these demographic controls is established in the literature (e.g., see allgood and walstad, 2013 or lusardi and mitchell, 2011). unless otherwise noted, the omitted variables are: the low-low group, 18–24 age group, college graduate, full-time employment, married, no dependent children, and annual income of at least $50,000 but less than $75,000. all variables are interpreted in reference to these groups. 4.1. predicting financial behaviors in table 4, financial knowledge and confidence are both statistically significant predictors of financial behavior. because the omitted variable is the low-low group, interpreting the odds ratio of the low knowledge and high confidence group reflects a difference in confidence, and interpreting the odds ratio of the high knowledge and low confidence group reflects a difference in actual knowledge. for example, those with high confidence (holding actual knowledge constant at a low level) are 1.131 times more likely to have obtained a credit report whereas those with high actual knowledge (holding confidence constant at a low level) are 1.294 times more likely to have obtained a credit report. as a robustness check to ensure that confidence affects financial behaviors regardless of the level of actual financial knowledge, table 5 reconsiders the regressions in table 4. instead of using the four knowledge groups as predictors, continuous variables are used to measure knowledge (ranging from 0 of 5 questions correct to 5 of 5 correct) and confidence table 4 financial literacy and financial behavior high knowledge, high confidence high knowledge, low confidence low knowledge, high confidence savings and borrowing behaviors calculate 1.459*** 1.350*** 1.609*** compare credit cards 1.205*** 1.456*** 1.650*** credit report 1.329*** 1.131** 1.294*** credit score 1.247*** 1.218*** 1.380*** health insurance 1.055* 1.116* 1.133** life insurance 1.102*** 1.051 1.215*** investments 1.429*** 1.569*** 1.170*** loan from retirement account 0.801*** 0.724** 1.017 home equity loan 0.943 1.176 1.393*** foreclosure 0.892** 0.846 1.471*** logistic regressions predicting financial behavior whereby the predictor variables control for gender, age, education, employment status, marital status, income, income-drop, and risk tolerance. the dependent variables are listed vertically and represent the financial behaviors (0 � did not engage in behavior; 1 � engaged in behavior), which are detailed in the appendix. the odds ratio for three financial knowledge groups is reported, a ratio of the likelihood of an event occurring in one group to the likelihood of the same event occurring in another group. the omitted variable is the low actual knowledge and low confidence group. the odd ratios of the three reported knowledge groups are interpreted in reference to this omitted group. the results for the control variables are not reported. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively. 109c.t. asaad / financial services review 24 (2015) 101–117 (self-assessed rating from 1 to 7). confidence is measured using overall financial confidence (overall) and an average financial confidence (average). for interpretation purposes, standardized variables are used in the regression so that the odds ratio reflects how a one standard deviation above or below the average affects financial behaviors. echoing the results of table 4, table 5 also demonstrates that both knowledge and confidence are important components of financial behaviors. for example, with a one standard deviation increase above the mean in knowledge and confidence an individual is 1.236 and 1.372 times more likely, respectively, to have calculated how much he or she needs to save for retirement. higher levels of financial literacy lead to “better” financial decision making. however, learning from the consequences of past financial decisions, particularly learning from mistakes, may also lead to higher levels of financial literacy. for example, financial knowledge is positively related to seeking financial advice (collins, 2012; robb, babiarz, and woodyard, 2012). herein lies a potential endogeneity problem: financial literacy levels predict financial behaviors and financial behaviors (experience) may predict financial literacy levels. to address the potential simultaneity issue, correlations examine the inter-relation between different financial behaviors. then, two-stage least squares regressions use high school financial education to instrument for financial literacy. similarly, van rooij, lusardi, and alessie (2011) use economic education in high school as an instrumental variable, arguing that it is correlated with financial literacy (the independent variable) but not correlated with stock market participation (the dependent variable). first, there is not a clear statistical table 5 robustness checks for confidence measure overall average knowledge confidence knowledge confidence savings and borrowing behaviors calculate 1.236*** 1.372*** 1.227*** 1.226*** compare credit cards 1.115*** 1.276*** 1.102*** 1.206*** credit report 1.118*** 1.297*** 1.111*** 1.201*** credit score 1.099*** 1.241*** 1.088*** 1.190*** health insurance 1.089*** 1.033* 1.082*** 1.021 life insurance 0.990 1.157*** 0.981 1.129*** investments 1.324*** 1.298*** 1.320*** 1.169*** loan from retirement account 0.776*** 1.060 0.706*** 0.857*** home equity loan 0.925*** 1.120*** 0.935** 0.991 foreclosure 0.703*** 1.161*** 0.725*** 0.920** the logistic regressions from table 4 are reconsidered. instead of using the four knowledge groups, continuous variables are used for financial knowledge and financial confidence. the financial knowledge measure reflects the number of correctly answered questions, and thus ranges from 0 to 5. several confidence measures are considered: overall, day-to-day, math, and average. in all regressions, both the financial knowledge variable and the respective financial confidence variable are standardized for ease of interpretation of the odds ratios. the control variables gender, age, education, employment status, marital status, income, income-drop, and risk tolerance are not reported. the financial behaviors (0�did not engage in behavior; 1�engaged in behavior) are listed vertically along the left panel and are detailed in the appendix. the odds ratio represents reflects how a one standard deviation above or below the average affects financial behaviors. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively. 110 c.t. asaad / financial services review 24 (2015) 101–117 relationship across financial behaviors, suggesting that financial behaviors are not distinct predictors of financial literacy (see also allgood and walstad, 2013). second, in all specifications, the financial confidence variable is still statistically and economically significant. moreover, previous research has also not found reverse causality to be an issue (allgood and walstad, 2013; van rooij, lusardi, and alessie, 2011). the results in tables 4 and 5 provide support for h1, that financial confidence predicts financial behavior. in addition, these findings support remund’s (2010, p. 284) definition of financial literacy as “a measure of the degree to which one understands key financial concepts and possesses the ability and confidence to manage personal finances” and the work of courchane, gailey, and zorn (2008, p. 137) who find that “optimistic self-assessments, not accurate ones, lead to better financial outcomes.” 4.2. risky behaviors h2 proposes that overconfident individuals are most likely to engage in risky financial behaviors. table 4 hints that this may be the case: the high-high knowledge group is less likely than the low-low knowledge group to have experienced a foreclosure process while the low-high (overconfident) knowledge group is more likely than the low-low group. because of an underestimation of risk, overconfident individuals likely have an increased propensity to engage in risky behaviors. the omitted knowledge group in this next series of regressions is the overconfident group, or those with low knowledge and high confidence (low-high). if the overconfident are more likely to engage in risky behaviors, then the odds ratios of the other knowledge groups should be below one, indicating that these groups are less likely to engage in risky behaviors than the low-high group. after controlling for many other factors, including risk tolerance, table 6 shows that “overconfident” individuals are more likely than other knowledge groups to take these financial risks. compared with “overconfident” individuals, those with low perceived table 6 financial literacy and risky behavior high knowledge, high confidence high knowledge, low confidence low knowledge, low confidence risky, high-cost behaviors title-loan 0.484*** 0.656*** 0.575*** pay-day loan 0.595*** 0.719*** 0.745*** tax advance 0.496*** 0.602*** 0.588*** pawn shop 0.656*** 0.804*** 0.780*** rent-to-own 0.630*** 0.586*** 0.769*** logistic regressions predicting financial behavior whereby the predictor variables control for gender, age, education, employment status, marital status, income, income-drop, and risk tolerance. the dependent variables are listed vertically and represent the financial behaviors (0 � did not engage in behavior; 1 � engaged in behavior), which are detailed in the appendix. the odds ratio for three financial knowledge groups is reported, a ratio of the likelihood of an event occurring in one group to the likelihood of the same event occurring in another group. the omitted variable is the low actual knowledge and high confidence group. the odd ratios of the three reported knowledge groups are interpreted in reference to this omitted group. the results for the control variables are not reported. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively. 111c.t. asaad / financial services review 24 (2015) 101–117 and low actual knowledge are about 25% less likely to have obtained a payday loan and 22% less likely to have taken a tax advance. because actual knowledge is low across these two groups, the increased propensity to engage in these risky behaviors is because of the high-perceived knowledge (confidence), that is, overconfidence. as a robustness check, table 7 reconsiders the regressions in table 6, but instead of using the four knowledge groups as predictor variables, standardized continuous variables are used to measure knowledge and confidence. an interaction term, knowledge*confidence, considers the relation between the two financial literacy components. the statistically significant interaction terms in table 7 indicate that confidence influences the relationship between knowledge and financial behavior, affecting the strength and/or direction of the relationship. fig. 1 illustrates one of the interactions from table 7, how confidence affects the use of payday loans. the slope of the high confidence line is steeper than the low confidence line indicating that the affect of knowledge on payday loan behavior is different for different confidence levels. when knowledge is low, those with high confidence are more likely to engage in the risky behavior and when knowledge is high, those with high confidence are less likely to engage in the risky behavior. this illustrates that confidence is good, but too much confidence is harmful. this link between overconfidence and risky behaviors helps elucidate the discrepancy that parker, bruin, yoong, and willis (2012) could not explain: they found that confidence is positively associated with “good” financial decisions, yet prior research demonstrates a negative association between overconfidence and trading behaviors (e.g., barber and odean, 2000; grinblatt and keloharju, 2009). people tend to have unrealistic self-perceptions, but these positive-illusions can be advantageous (sedikides, 1993). this self-efficacy gives individuals the confidence to act (bandura, 1997). thus, as parker, bruin, yoong, and willis (2012, p. 387) suggest, “confidence may play a role in reducing hesitation and increasing risk taking.” however, in some financial circumstances, inaccurately assessing the level of risk table 7 robustness checks overconfidence measure knowledge confidence knowledge*confidence risky, high-cost behaviors title-loan 0.673*** 1.060** 0.897*** pay-day loan 0.713*** 0.966 0.923*** tax advance 0.652*** 1.066** 0.886*** pawn shop 0.770*** 0.949** 0.926*** rent-to-own 0.667*** 1.074*** 0.884*** the logistic regressions from table 6 are reconsidered. instead of using the four knowledge groups, continuous variables are used for financial knowledge and financial confidence. the financial knowledge measure reflects the number of correctly answered questions, and thus ranges from 0 to 5. the overall confidence measure (1–7) is considered. in all regressions, both the financial knowledge variable and the financial confidence variable are standardized. an interaction term, knowledge*confidence, considers the relation between the two financial literacy components, knowledge, and confidence. the control variables gender, age, education, employment status, marital status, income, income-drop, and risk tolerance are not reported. the financial behaviors (0 � did not engage in behavior; 1 � engaged in behavior) are listed vertically along the left panel and are detailed in the appendix. the odds ratio represents reflects how a one standard deviation above or below the average affects financial behaviors. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively. 112 c.t. asaad / financial services review 24 (2015) 101–117 can lead to risky decisions. the results in table 6, table 7, and fig. 1 support h2, that overconfident individuals are most likely to engage in risky financial behaviors. together, these findings align with goel and thakor’s (2008) theoretical model: some overconfidence is good, but too much is bad. 5. conclusion this analysis examines two components of financial literacy, knowledge, and confidence. not surprisingly, individuals with both high knowledge and confidence are more likely to make “good” financial decisions than individuals with both low knowledge and confidence. somewhat surprising, however, is how influential perceived knowledge is on financial behavior. additionally, when confidence is high and actual knowledge is low, individuals are more likely to take financial risks by engaging in costly behaviors. these findings are robust to measurement changes for the four knowledge groups, confidence, and overconfidence. this article makes two noteworthy contributions. first, confidence, or self-perceived knowledge, is an important component of financial literacy. prior studies found a positive association between an individual’s self-assessed level of confidence with investing, retirement planning, and credit card behaviors. this analysis extends these findings by demonstrating, with a large nationally representative sample, that confidence affects additional savings and borrowing behaviors. in most financial circumstances, higher confidence is beneficial. however, the second contribution of this article is the clarification of when confidence can be detrimental. overconfident individuals tend to overestimate the precision of their knowledge and underestimate risk; thus, perhaps even unbeknownst to them, overconfident individuals are more likely to engage in risky and costly financial behaviors. fig. 1. confidence and the probability of obtaining a payday loan by knowledge level. this figure illustrates the interaction between knowledge and confidence and the effect on the probability of obtaining a payday loan. graph created using the regression coefficients, including control variables, from table 7 and the website: http://www.jeremydawson.co.uk/slopes.htm. 113c.t. asaad / financial services review 24 (2015) 101–117 these findings are based on survey data. as such, the results are only as good as the survey design and participant veracity. survey responses are “vulnerable to social desirability,” meaning that if financial behaviors are viewed as having “a normative valence” then respondents are apt to overstate “good” financial behaviors and understate ‘bad’ financial behaviors (willis, 2008). additionally, financial decisions are not strictly about money but involve balancing life’s tradeoffs, and it is possible that some of these omitted factors, such as personality characteristics, may play a role in shaping financial behaviors. as predicted, confidence is an important part of financial decision making. the strength of the findings may relate to survey design issues: the confidence measure may pick up on a more “general” financial knowledge (or perhaps financial resourcefulness or experience) that the actual knowledge measure misses or may also reflect other confidence issues such as trust or general life outlook. future experimental research may help explicate these measures and their respective affects on financial behaviors. what is clear from the analysis, however, is that financial literacy encapsulates more than just numeracy knowledge. understanding that confidence is just as important as knowledge is of paramount value to educators and policy makers, helping to structure more effective financial literacy initiatives and improve individuals’ everyday financial decision making. the risky financial behaviors addressed in this article are complex and require numeracy knowledge; however, based on carpena et al.’s (2011) findings, more exposure to these financial topics will increase awareness and hopefully positively affect decision making. although numeracy and math-based skills are necessary for specific, concrete calculations, exposing individuals to financial topics and products may help individuals make more informed financial decisions. the challenge then is to create education initiatives that help individuals find a healthy dose of confidence. notes 1 the three questions (risk, interest, and inflation), developed by annamaria lusardi and olivia mitchell, first appear in the 2004 cross-section of the health and retirement study. lusardi and mitchell (2011) describe their rationale for using these three questions as a measure of financial literacy, emphasizing that the measures were chosen keeping four principles in mind: (1) simplicity, that is, basic financial concepts; (2) relevance, that is, pertinent to daily financial decision-making; (3) brevity, that is, small number of questions for widespread adoption; and (4) capacity to differentiate, that is, distinguish different levels of knowledge. 2 lusardi and mitchell (2011) also report that not only are women less likely than men to correctly answer the financial literacy questions, but women are also less likely to respond, that is, women are more likely to answer that they “do not know.” similarly, beierlein and neverett (2013) find that women are less likely to enroll in an elective personal finance course. 114 c.t. asaad / financial services review 24 (2015) 101–117 3 using finras 2009 survey, robb, babiarz, and woodyard’s (2012) average financial confidence measure is computed as an average of four confidence variables. “i regularly keep up with economic and financial news” is not included in the 2012 survey, and thus the average confidence variable is computed as an average of three confidence variables. 4 individuals are classified as having high confidence if they self-assessed their overall financial knowledge as a 6 or 7 on a 7-point scale and classified as having low confidence if they self-assessed their financial knowledge as 5 or lower on a 7-point scale (mean � 5.15; median � 5.00). individual are classified as having high actual financial knowledge if they answered 3, 4, or 5 of 5 questions correctly and as low actual financial knowledge if they answered 2 questions or less correctly (mean � 2.88; median � 3.00). appendix: finras 2012 national financial capability study: selected survey topics and questions savings and loan behaviors calculate have you ever tried to figure out how much you need to save for retirement? compare ccs thinking about when you obtained your most recent credit card, did you collect in formation about different cards from more than one company in order to compare them? credit report in the past 12 months, have you obtained a copy of your credit report? credit score in the past 12 months, have you checked your credit score? health insurance are you covered by health insurance? life insurance do you have a life insurance policy? investments not including retirement accounts, do you have any investments in stocks, bonds, mutual funds, or other securities? retirement loan in the last 12 months, have you or your spouse/partner taken a loan from your retirement account(s)? home equity loan do you have any home equity loans? foreclosure have you been involved in a foreclosure process on your home in the last 2 years? risky, high-cost behaviors auto title loan in the last two years, have you taken out an auto title loan? payday loan in the last two years, have you taken out a short term “payday” loan? tax advance in the last two years, have you gotten an advance on your tax refund? used pawn shop in the last two years, have you used a pawn shop? used rent-to-own in the last two years, have you used a rent-to-own store? references agnew, j. r., & szykman, l. r. (2005). asset allocation and information overload: the influence of information display, asset choice, and investor experience. journal of behavioral finance, 6, 57–70. alhenawi, y., & elkhal, k. (2013). financial literacy of u.s. households: knowledge vs. long-term financial planning. financial services review, 22, 211–244. allgood, s., & walstad, w. (2013). financial literacy and credit card behaviors: a cross-sectional analysis by age. numeracy, 6, 1–26. bandura, a. (1997). self-efficacy mechanism in human agency. american psychologist, 37, 122–147. 115c.t. asaad / financial services review 24 (2015) 101–117 barber, b. m., & odean, t. (2000). trading is hazardous to your wealth: the common stock investment performance of individual investors. journal of finance, 55, 773–806. beierlein, j. j., & neverett, m. (2013). who takes personal finance? financial services review, 22, 151–171. bernheim, b. d., garrett, d. m., & maki, d. m. (2011). education and saving: the long-term effects of high school financial curriculum mandates. journal of public economics, 80, 435–465. campbell, w. k., goodie, a. s., & foster, j. d. (2004). narcissism, confidence, and risk attitude. journal of behavioral decision making, 17, 297–311. carpena, f., cole, s., shapiro, j., & zia, b. (2011). unpacking the causal chain of financial literacy. world bank policy research working paper series, vol. (2011). choi, j. j., laibson, d., & madrian, b. c. (2005). are empowerment and education enough? underdiversification in 401(k) plans. brookings papers on economic activity, 2005, 151–198. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. cordell, d. m., smith, r., & terry, a. (2011). overconfidence in financial planners. financial services review, 20, 253–263. courchane, m., gailey, a., & zorn, p. (2008). consumer credit literacy: what price perception?. journal of economics and business, 60, 125–138. fisher, p. j. (2010). black-white differences in saving behaviors. financial services review, 19, 1–16. goel, a. m., & thakor, a. v. (2008). overconfidence, ceo selection, and corporate governance. journal of finance, 63, 2737–2784. grimes, p. w., rogers, k. e., & smith, r. c. (2010). high school economic education and access to financial services. journal of consumer affairs, 44, 317–335. grinblatt, m., & keloharju, m. (2009). sensation seeking, overconfidence, and trading activity. journal of finance, 64, 549–578. healy, p., & moore, d. (2007). bayesian overconfidence. available at ssrn 1001820. johnson, d. d., & fowler, j. h. (2011). the evolution of overconfidence. nature, 477, 317–320. kyrychenko, v., & shumb, p. (2009). who holds foreign stocks and bonds? characteristics of active investors in foreign securities. financial services review, 18, 1–21. lusardi, a. (2004). saving and the effectiveness of financial education. in o. mitchell & w. utkus (ed.), pension design and structure: new lessons from behavioral finance (pp. 157–184). cambridge: oxford university press. lusardi, a. (2008). financial literacy: an essential tool for informed consumer choice? (available at http:// www.dartmouth.edu/�alusardi/pub.html). lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. (2011). financial literacy around the world: an overview. journal of pension economics and finance, 10, 497–508. lusardi, a., mitchell, o. & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. mandell, l. (2006). teaching young dogs old tricks: the effectiveness of financial literacy intervention in pre-high school grades. presented at the academy of financial services 2006 annual conference, salt lake city, october 11, 2006. mandell, l. (2008). financial education in high school. in a. lusardi (ed.), overcoming the saving slump (pp. 257–279). chicago, il: the university of chicago press. mandell, l., & klein, l. s. (2009). the impact of financial literacy education on subsequent financial behavior. journal of financial counseling and planning, 20, 15–24. metcalfe, j. (1986). feeling of knowing in memory and problem solving. journal of experimental psychology, 12, 288–294. nelson, t. o., gerler, d., & narens, l. (1984). accuracy of feeling-of-knowing judgments for predicting perceptual identification and relearning. journal of experimental psychology: general, 113, 282–300. 116 c.t. asaad / financial services review 24 (2015) 101–117 oecd (organisation for economic co-operation and development) (2005). improving financial literacy. analysis of issues and policies. paris, france: oecd publishing. park, c. w., gardner, m. p., & thukral, v. k. (1988). self-perceived knowledge: some effects on information processing for a choice task. american journal of psychology, 101, 401–424. parker, a. m., bruin, w. b., yoong, j., & willis, r. (2012). inappropriate confidence and retirement planning: four studies with a national sample. journal of behavioral decision making, 25, 382–389. remund, d. l. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. schachter, d. l. (1983). feeling of knowing in episodic memory. journal of experimental psychology: learning, memory, and cognition, 9, 39–54. sedikides, c. (1993). assessment, enhancement, and verification determinants of the self-evaluation process. journal of personality and social psychology, 65, 317–338. simon, m., houghton, s. m., & aquino, k. (2000). cognitive biases, risk perception, and venture formation: how individuals decide to start companies. journal of business venturing, 15, 113–134. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. williams, e. f., & gilovich, t. (2008). do people really believe they are above average? journal of experimental social psychology, 44, 1121–1128. willis, l. (2008). evidence and ideology in assessing the effectiveness of financial literacy education. iowa law review, 94. 117c.t. asaad / financial services review 24 (2015) 101–117 financial teaching by parents and financial education at school or workplace: evidence from japan tsung-ming yeha,* afaculty of economics, kyushu university, 744 motooka nishi-ku, fukuoka 819-0395, japan abstract this study investigates the differential roles of financial socialization within the family versus financial education at school or the workplace, using data from a representative sample of 25,000 japanese individuals. the results indicate that different platforms may play different roles. while the adults’ short-term financial behaviors, which involve regular feedback and immediate consequences for deviation, are primarily related to their parent’s financial advising in childhood, long-term financial behaviors, which require complex planning and decision-making, are primarily related to financial education received at school or the workplace. the results also suggest the benefits of accumulating financial experiences and education in different stages. © 2022 academy of financial services. all rights reserved. jel classifications: g51; g53; p36; p46 keywords: financial literacy; financial education; financial behavior 1. introduction the literature has, in general, revealed an insufficient level or amount of financial literacy around the world. for instance, atkinson and messy (2011) reported that few people across countries could correctly answer basic financial literacy questions across 14 countries, and lusardi and mitchell (2014) compiled a similar pattern in 12 countries. questions arise as to whether financial education programs can effectively improve financial literacy and skills that are instrumental in one’s personal finances. empirical studies emerged, using different *corresponding author: tel.: +81-928025457, fax: +81-928025457. e-mail address: yeh@econ.kyushu-u.ac.jp 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 297–320 designs and methods to address this question. as will be elaborated in the next section, more recent empirical research, even studies applying the most rigorous research design, randomized controlled trials (rct), which can better identify causal effects (e.g., kaiser & menkhoff, 2017, 2020; kaiser et al., 2021), has generally documented evidence of a positive effect of financial education offered in schools, communities, or workplaces. one strand of the research investigated the effect of state-mandated financial education, also finding a positive effect. for instance, stoddard and urban (2020) showed that high school financial education policies reduced nonstudent debt and loan delinquency rates among 19to 29-yearolds. in addition, family is another important platform in which youth can acquire the foundations of financial capability into their adulthood (consumer financial protection bureau, 2016). financial socialization theory postulates that youth can grasp financial knowledge and behavioral values within the family through interactions with family members on monetary issues (gudmunson & danes, 2011; hanson & olson, 2018; jorgensen et al., 2019). empirical research on parental financial teaching usually had to rely on data acquired from survey questionnaires, as it can be difficult to conduct an rct or quasi-experiments on parental teaching within a family. while subject to possible endogeneity and other methodological issues, existing research based on survey data has also revealed a positive association between parental financial advising in childhood or adolescence and financial behaviors in adulthood (e.g., bucciol & veronesi, 2014; grinstein-weiss et al., 2011; sansone et al., 2019). most existing studies investigated the financial experiences or financial education taking place in a specific platform, being family, school, community, or workplace. however, individuals can receive financial education throughout stages of life—as children, students, and employees. the cumulative effect of financial education can play out, as suggested by serido and shim (2014). they reported that financial education, starting in high school and continuing in college, can contribute to more responsible financial behaviors during and after college. wagner and walstad (2019) attempted to investigate the distinct influences of financial education received at high schools, colleges, and workplaces. this study provides additional evidence by further investigating the influence of parental financial advising in childhood vis-à-vis financial education at school or workplace on financial behaviors in adulthood. such investigation is possible by using data from the financial literacy survey (fls) 2016—an online survey conducted by japan’s central council for financial services information in 2015, which contains self-reported experiences of being taught by parents, the financial education experiences at school or workplace, as well as a set of questions relating to financial knowledge, attitudes, and behaviors. furthermore, following wagner and walstad (2019), this study examined two types of financial behaviors— short-term ones defined as “involving a money or credit management task that gives regular and timely feedback to remind people about what they need to do to change their financial behavior to avoid financial penalties and consequences,” and long-term ones as “involving more planning for the future and are less influenced by regular feedback or learning by doing.” the results indicate that those who were only taught by parents how to manage money primarily manifest desirable short-term behaviors measured by “careful consideration before 298 t.-m. yeh / financial services review 30 (2022) 297–320 purchase,” “paying bills on time,” “watching financial affairs closely,” and “having an emergency fund.” those who received only financial education at school or the workplace primarily displayed desirable long-term behaviors measured by investing in stocks and saving for retirement, while also displaying desirable short-term behaviors to a significantly lesser extent. the results are robust to tests using instrumental-variable and propensity score matching methods. it is plausible that short-term financial behavior is shaped within the family as children develop sound financial values and attitudes from direct and indirect socialization with family members, particularly parents, leading to behaviors such as paying bills on time. such socialization and experiences within the family can also strengthen executive functions, resulting in self-control and careful purchase behaviors. parents can provide timely negative feedback or punishment when children deviate from these practices. on the other hand, financial education at school or the workplace aims at improving knowledge of personal finances and budgeting skills, preparing students or employees for long-term financial behaviors such as retirement planning, saving, or home buying (fox et al., 2005). caution is required when interpreting the results of this study, which has some limitations. primary potential concerns include recall error, endogeneity, and omitted variables problem, which, together with other issues, will be discussed in the final section. despite the limitations, this study enriches the financial education literature by providing evidence distinguishing financial socialization in the family from financial education at school or the workplace. the results of differential influences associated with different platforms may provide implications for educators, employers, financial education trade bodies, and policymakers. in addition, evidence from japanese data can enrich the literature, as the influence of financial education on broader population-level financial knowledge and self-efficacy were understudied outside the united states (rothwell & wu 2019). 2. literature review financial well-being is associated with financial capability, the capacity to manage financial resources effectively based on knowledge, skill, and access (consumer financial protection bureau, 2016). the building blocks of financial capability can be acquired through financial education in the early stage of life. the consumer financial protection bureau (cfpb) suggests a developmental framework, based on extensive research, for understanding when, where, and how young people learn and develop the following three building blocks (consumer financial protection bureau, 2016). (1) executive functions— the cognitive processes used to make plans, focus attention, remember information, and perform multitasks, which are essential in saving, setting financial goals, and managing money. (2) financial habits and norms—the values, standards, and heuristics used in financial matters, such as making a point of paying bills on time. (3) financial knowledge—familiarity with financial facts and concepts, which helps efficient money management and effective comparison of financial products. t.-m. yeh / financial services review 30 (2022) 297–320 299 children and youth acquire these building blocks from their family members or school education. financial socialization theory postulates that youth grasp financial knowledge and behavioral values within the family through interactions with family members on monetary issues (gudmunson & danes, 2011; hanson & olson, 2018; jorgensen et al., 2019). schools are also crucial for financial socialization by providing structured curricula or activities such as reality fairs or savings-promotion programs. empirical studies use different research designs to investigate the effect of financial education on financial literacy, attitudes, or behaviors. randomized controlled trials (rct) are ideal for identifying the causal effects. for instance, frisancho (2018) evaluated the impact of a large-scale school-based financial education and found that students improved financial knowledge, self-control, and consumption habits, as of 6;24 months postintervention. batty et al. (2020) reported that engaging in an experiential economic program (called my classroom economy) improved students’ financial knowledge of elementary students in the united states. a few studies have emerged applying meta-analysis on the effects of financial education. fernandes et al. (2014) covered 15 previous rct studies (up to 2013), which showed no significant effect, which is smaller than that among correlational studies. another meta-analysis by miller (2015) covered studies based on rct (up to 2013) and found similar results—the impact appears limited at best in the outcomes such as savings, credit performance, or financial knowledge. however, more recent meta-analyses covering a more extensive set of studies found an overall positive result. kaiser and menkhoff (2020) metaanalyzed 18 previous rct studies (up to 2019) on the effect of school-based education intervention, reporting significant and positive effect size among students in terms of financial literacy and some financial behaviors. another meta-analysis by kaiser et al. (2021) covered a larger set of rct previous studies (=76) up to 2019, which also include those education programs outside schools, reported positive and positive effects size in terms of financial literacy and some financial behaviors, particularly budgeting, saving and investing. the latter two studies compiled a more significant effect on financial literacy than financial behaviors. they also found that financial education is less effective for some specific behaviors, such as handling debt. in addition to rct, meta-analyses by kaiser and menkhoff (2017, 2020) also investigated studies on nonrandom participants of some sort of educational program, finding positive effects on financial literacy and behaviors. there also exist empirical studies based on observational data, using natural or quasiexperiments and surveying. for instance, among the 37 previous studies covered by kaiser and menkhoff (2020) for their meta-analyses, 19 used a nonrandom design method (with the remaining 18 rcts). such studies have more flexibility in the investigated outcomes and more prolonged effects of financial education while also being subject to endogeneity. it requires caution when interpreting the results from observational data. one strand of research using quasi-experiments assessed the effect of financial education by using the variation in u.s. high school mandates across different states. for instance, 29 states mandated some form of consumer education in secondary schools between 1957 and 1985 to prepare students with practical and useful decision-making skills in financial matters. bernheim et al. (2001) found that these mandates had a high positive impact on saving rates and wealth accumulation during adulthood as of 1995. interestingly, cole et al. (2016) showed that these programs did not improve savings, using a much larger sample and a 300 t.-m. yeh / financial services review 30 (2022) 297–320 more flexible specification. they suggested a possible endogeneity explanation—those states had imposed mandates during rapid economic growth periods, which might have explained the higher savings behavior of concurrent graduates. however, more recent studies did find that financial education mandates reduced defaults and higher credit scores among young adults (urban et al., 2020), reduced the likelihood and frequency of payday borrowing among young adults (harvey, 2019a), and increased bank account ownership among young adults with lower educational credentials, while, overall, having no effect no bank account ownership and propensity to save (harvey, 2019b). burke et al. (2020) also reported that state-mandated financial education improves financial well-being, primarily accruing to men and those with college degrees. aside from quasi-experimental studies, empirical studies rely on large-scale survey data, particularly in the literature on the association between financial literacy and financial behaviors. these surveys contain financial knowledge-related questions used to measure one’s financial literacy level. some large-scale survey contains questions about one’s financial experiences and financial education during childhood or adolescence. such surveys complement the literature on the roles of financial socialization within the family, as it is harder to implement rct or quasi-experiments on activities within the family. results based on the survey have, in general, reported a positive relationship between financial socialization during childhood or adolescence and financial behaviors at a later stage. for instance, ashby et al. (2011) investigated data from a national survey in the united kingdom, finding that adults are more likely to save when they had saved at age 16. in addition, saving in adulthood is not related to their receiving pocket money from parents or relatives during adolescence. however, sansone et al. (2019) reported that dutch adults displayed greater inflation-related knowledge and self-assessed financial literacy if they reported receiving pocket money between 8 and 12 years of age or being taught how to run a budget to save between 12 and 16. furthermore, grinstein-weiss et al. (2011) found that those american adults who reported receiving money-management teaching from their parents are associated with higher credit scores and lower credit card debt in adulthood. bucciol and veronesi (2014) also showed that those dutch individuals who reported parental teaching during childhood are more likely to save in adulthood, particularly when they were given pocket money with advice on saving and budgeting. however, caution is required because these studies may be subject to endogeneity issues inherent in nonrandom data as well as recall error, as people may not correctly recall experiences occurring many years ago. despite the limitation, this study aims to present evidence based on a large-scale japanese survey that includes information on the respondent’s parental teaching in childhood and financial education experience received at school or the workplace. the survey data make it possible to separate the roles of parental teaching distinctly from financial education at school or the workplace, while most previous studies can only focus on financial experiences on a certain platform.1 2.1. the financial education system in japan the japanese school curriculum enacted in 1951 recommended the introduction of savings promotion activities in schools through the so-called “children’s banks,” in which t.-m. yeh / financial services review 30 (2022) 297–320 301 students can deposit and withdraw money in financial institutions through schools (messy & monticone, 2016). in 2006, the ministry of education overhauled the basic education act, which stipulated the objectives of education, including “fostering a spirit of autonomy and independence, emphasizing connections to a career and practical life and developing a mindset of active contribution to the building and development of society” (oecd, 2013). consequently, japan revised and renewed the school course guidelines to strengthen financial education, which was implemented at elementary, junior high, and senior high schools in 2011, 2012, and 2013, respectively. however, financial education was not mandatory in the curriculum (until april 2022). in addition, financial knowledge is only provided patchily and sporadically in specific subjects, such as “civics” in junior high school and “social studies” or “politics and economics” in high school, and with limited teaching hours. a survey of japanese school teachers conducted from 2013 to 2014 revealed that more than a majority of teachers answered insufficient teaching hours set aside for financial knowledge, and 40% answered insufficient content (financial education promotion study group, 2014). in addition, financial trade bodies offer a variety of opportunities for financial education, via seminars and visiting lectures, for working adults and citizens. such programs are more practical in nature, covering explanations of financial products, asset management, investment knowledge, family budget management, and life planning (oecd, 2013). 3. data and variables this study investigated whether and how financial behaviors in adulthood are associated with financial socialization in the family versus financial education at school or the workplace. data were drawn from fls 2016—an online survey conducted by japan’s central council for financial services information in 2015 to shed light on japanese individuals’ financial knowledge, attitudes, and behaviors. the sample comprises 25,000 individuals, distributed in proportion to japan’s demographic structure. table 1 summarizes the results of this empirical study. the average age was 48.7 years, ranging from 18 to 79 years. females accounted for half of the sample. by occupation, 32.2% were employed by a company, 20.9% were house workers, 15.6% were not employed, 14% were part-time workers, 7% were self-employed, 4.9% were students, and 3.5% were civil servants. by education degree, 38.6% had a college degree, followed by those with high school education (32.4%), 2-year college degree (11.29%), and vocational education (10.5%), while only 4.2% received graduate school education. regarding household income, the largest cohort is 2.5–5 million yen annually (28.9%), followed by 5–7.5 million yen (16.6%), and up to 2.5 million yen (15.7%). only 6.7% of the survey respondents reported a household income of more than 10 million yen (equivalent to approximately $91,116 as of january 2020). 3.1. variables on financial socialization or financial education two fls questions in the survey are used to construct the variables. one question asks if “your parents or guardians taught you how to manage your finances,” with 19.8% of 302 t.-m. yeh / financial services review 30 (2022) 297–320 respondents replying “yes,” 60.4% “no,” and 19.8% “don’t know.” i define those with a “yes” response as those receiving financial teaching from parents (guardians) at home in childhood. the other fls question asks if “financial education was offered by a school or college you attended or a workplace where you were employed,” with 6.6% of respondents replying “yes,” 75.7% “no,” and 17.7% “don’t know.” i define those with a “yes” response as those receiving financial education at school or in the workplace. the relatively smaller table 1 descriptive statistics variables no. mean or proportion (%) age 25,000 48.71 % female 25,000 50.66 % with occupation = company workers 25,000 32.24 civil servant 25,000 3.51 self-employed 25,000 6.99 part-timers 25,000 14.03 house-work 25,000 20.90 student 25,000 4.85 not employed 25,000 15.64 others 25,000 1.84 % with degree = mandatory education 25,000 2.82 high school 25,000 32.42 vocation school 25,000 10.51 2-year college 25,000 11.29 4-year college 25,000 38.60 graduate 25,000 4.20 other 25,000 0.15 % with household income = 0 mil. yen 25,000 3.60 >0 and <2.5 mil. yen 25,000 15.70 >2.5 and <5 mil. yen 25,000 28.90 >5 and <7.5 mil. yen 25,000 16.64 >7.5 and <10 mil. yen 25,000 9.68 >10 and <15 mil. yen 25,000 5.12 >15 mil. yen 25,000 1.62 don’t know 25,000 18.75 % receiving fin. teaching by parents only 25,000 17.04 % receiving fin. education at school/work only 25,000 3.82 % receiving both at home & school/work 25,000 2.76 % receiving non at home or school/work 25,000 76.38 carefully consider before buying (from 1 to 5) 25,000 3.94 pay bills on time (from 1 to 5) 25,000 4.42 keep a close eye on financial affairs (from 1 to 5) 25,000 3.65 % have a 3-month emergency fund 25,000 54.85 stock investment 25,000 32.64 % estimate post-retirement expenses 25,000 49.39 % have a plan for post-retirement expenses 14,185 35.59 % set aside post-retirement expenses 14,185 26.04 % correctly answered 0 question 25,000 15.22 1 question 25,000 16.19 2 questions 25,000 18.58 3 questions 25,000 20.44 4 questions 25,000 19.59 5 questions 25,000 9.99 t.-m. yeh / financial services review 30 (2022) 297–320 303 number epitomizes the fact that financial education has not been emphasized in the japanese education system. i constructed four dummy variables regarding the financial education experience. “fin. teaching by parents only” dummy indicated those who received financial advice from parents but did not receive financial education at school/workplace. “fin. education at school/work only” dummy indicated those who received financial education at school/workplace but not financial advising by parents. “both at home and school/work” was defined as those receiving financial advice from parents as well as financial education at school/work. finally, the “no fin. education” dummy was defined for the remaining respondents. the middle of table 1 reports that 76.4% received no financial teaching at home or school or workplace, while 17% received financial teaching from parents only, 3.8% at school or work only, and 2.76% both. 3.2. variables on financial behaviors following wagner and walstad (2019), i constructed variables for an individual’s shortterm and long-term financial behaviors. these variables are based on replies to the following fls questions. descriptive statistics are reported in the lower part of table 1. 3.3. short-term financial behavior variables 1. “before i buy something, i carefully consider whether i can afford it.” on a scale of 1 to 5, 33.8% indicated “5 = strongly agree,” 36.7% “4 = agree,” 22% “3 = neutral,” 5.3% “2 = disagree,” and 2.3% “1 = strongly disagree.” the average score was 3.9. a category variable was defined for this behavior, taking values from 1 to 5. 2. “i pay my bills on time.” on a scale of 1 to 5, 63.5% indicated “5 = strongly agree,” 21% “4 = agree,” 11.3% “3 = neutral,” 2.7% “2 = disagree,” and 1.5% “1 = strongly disagree.” the average score was 4.42. a category variable was defined by taking values from 1 to 5. 3. “i watch my financial affairs closely.” on a scale of 1 to 5, 22.5% indicated “5 = strongly agree,” 34.9% “4 = agree,” 30.4% “3 = neutral,” 8.8% “2 = disagree,” and 3.3% “1 = strongly disagree.” the average score was 3.65. a category variable was defined by taking values from 1 to 5. 4. “have you set aside emergency funds that would cover your expenses for three months in case of sickness, job loss, economic downturn, or other emergencies?” here, 54.9% indicated “yes,” 29.7% “no,” and 15.4% “don’t know.” a dummy variable was defined for those who indicated “yes.” as suggested by the categorical variables, most japanese respondents seemingly displayed a prudent financial attitude. most respondents were also prepared for short-term financial needs. 3.4. long-term financial behavior variables 5. “have you ever purchased stocks?” 32.6% indicate “yes” and the remaining “no.” a dummy variable was defined for those who indicated “yes.” 304 t.-m. yeh / financial services review 30 (2022) 297–320 6. “are you aware of the amounts that will be required for your living expenses for retirement?” here, 49.4% indicated “yes” and 50.6% “no.” a dummy variable was defined for those who indicated “yes.” 7. “do you have a financial plan for the living expenses you think you will have to cover in the future?” here, 35.6% indicated “yes” and 64.4% “no.” a dummy variable was defined for those who indicated “yes.” 8. “have you set aside funds for the living expenses you think you will have to cover in the future?” here, 26.0% indicated “yes” and 74% “no.” a dummy variable was defined for those who indicated “yes.” contrary to short-term financial behaviors, the results suggest that most japanese respondents inadequately plan and prepare for their long-term financial needs. 3.5. control variables i also constructed a set of control variables that may influence one’s financial behaviors, such as age, gender, occupation, education attainment, household income, and residence area.2 another control variable is financial literacy, which has been documented to have a bearing on financial behaviors (behrman et al., 2012; disney & gathergood, 2013; klapper et al., 2013; lusardi & mitchell, 2007a, 2007b; rooij et al., 2011; yeh, 2022; yeh & ling, 2022). i constructed a financial literacy variable based on one’s answers to the “big-five” questions commonly used in previous studies (e.g., despard et al., 2020; gathergood & weber, 2017; ooijen & van rooij, 2016). 1. “suppose you put 1 million yen into a savings account with a guaranteed interest rate of 2% per year. how much would there be in the account after five years, disregarding tax deductions?” 2. “imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after one year, how much would you be able to buy with the money in this account?” 3. “true or false? buying a single company’s stock usually provides a safer return than a stock mutual fund.” 4. “if interest rates rise, what will typically happen to bond prices?” 5. “true or false?” “when compared, a 15-year mortgage typically requires higher monthly payments than a 30-year loan, but the total interest paid over the life of the loan will be less.” the respondents who correctly answered each of these questions were 65.7%, 55.6%, 45.8%, 24%, and 68.4%, respectively. the question on inflation and bond prices had the lowest correct rate (24%). alternatively, as reported at the bottom of table 2, out of these five questions, 10% of respondents correctly answered five questions, 19.6% four questions, 20.4% three questions, 18.6% two questions, and 16.2% one question. further, 15.2% failed to answer any question correctly. i constructed a financial literacy variable that indicated the number of correct answers, ranging from zero to five. t.-m. yeh / financial services review 30 (2022) 297–320 305 4. empirical tests and results 4.1. univariate tests the univariate tests compare the financial behaviors among the four types of respondents stratified by their financial education experience. the results are shown in table 2. the results of the one-way analysis of variance (anova) suggest that at least two groups were significantly different in terms of financial behaviors (p < .01). for almost all the financial behavior variables, the “no fin. education” group performed the worst, while “both at home and school/work” performed the best. however, for short-term financial behaviors, the “fin. table 2 financial behaviors stratified by financial education experiences no fin. teaching fin., teaching by parents only fin., education at school/ work only receiving both all panel a: short-term financial behaviors variables mean mean mean mean mean carefully consider before buying 3.90 4.09 3.96 4.13 3.94 anova f value 52.36 — — — — p-value 0.000 — — — — pay bills on time 4.39 4.56 4.36 4.51 4.42 anova f value 41.81 — — — — p-value 0.000 — — — — keep a close eye on financial affairs 3.58 3.87 3.76 4.01 3.65 anova f value 128.94 — — — — p-value 0.000 — — — — variables percent percent percent percent percent have an emergency fund 0.52 0.65 0.62 0.69 0.55 anova f value 115 — — — — p-value 0.000 — — — — panel b: long-term financial behaviors variables percent percent percent percent percent stock investment 0.30 0.33 0.55 0.49 0.32 anova f value 125.84 — — — — p-value 0.000 — — — — estimate post-retirement expenses 0.48 0.51 0.65 0.61 0.49 anova f value 32.77 — — — — p-value 0.000 — — — — have a plan for postretirement 0.33 0.41 0.49 0.52 0.36 anova f value 54 — — — — p-value 0.000 — — — — set aside post-retirement expenses 0.24 0.29 0.35 0.37 0.26 anova f value 23.67 — — — — p-value 0.000 — — — — note. anova stands for analysis of variance. 306 t.-m. yeh / financial services review 30 (2022) 297–320 teaching by parents only” group outperformed the “fin. education at school/work only” group, while the opposite was true for long-term financial behaviors. for instance, 65% in the “fin. teaching by parents only” group have emergency (for three-month) funds compared with 62% in the “fin. education in the school/work only” group. however, regarding “stock investment” (having retirement planning), 33% (41%) in the “fin. teaching by parents only” group were prepared compared with 55% (49%) in the “fin. education in the school/work only group.” the results suggest that education at home or school/work may have different implications for short-term and long-term financial behaviors. as one-way anova could not determine which specific groups were statistically significantly different from each other and neither accounted for other possible factors, i performed further multivariate analyses. 4.2. multivariate tests in this section, i performed (ordered) probit regressions of the short-term financial behavior variables. table 3 summarizes the (ordered) probit regression results for the four variables on short-term financial behaviors. for categorical variables (on a scale of 1 to 5), the columns report the marginal effects on the predicted probability of “5 = strongly agree” due to space limitations. explanatory variables include the dummies for financial education experiences, using the “no fin. education” group as the reference group. table 3 shows that the “fin. teaching by parents only” and “both at home and school/ work” groups are more likely to display the four short-term financial behaviors. for instance, these two groups are 7–8.5% more likely to strongly agree with the statement “before i buy something, i carefully consider whether i can afford it,” at a significant level (p < .01), relative to the “no fin. education” benchmark group. on the other hand, the “fin. education at school/work only” group shows no significant coefficient for “carefully consider whether i can afford it” and “pay bills in time,” respectively. in the other two short-term variables, “fin. education at school/work only” group indicates significant coefficients but with a smaller magnitude compared with the other two groups. in fact, the coefficient equality tests, reported at the bottom of table 3, show that “fin. teaching by parents only” is more likely than “fin. education at school/work only” to display the short-term behaviors at a significant level. table 4 reports the probit result on the long-term financial behaviors. all three groups receiving financial education were more likely than the group receiving no financial education to manifest the four long-term financial behaviors at a significant level (p < .01). individuals receiving education either at school or at work have a higher likelihood than those “taught by parents only.” for instance, compared with the no-financial-education group, the “financial education at school/work only” group has a 16% higher likelihood, but the “taught by parents only” group has 2%, respectively, to have stock investment experience. the coefficient equality tests also show that the “fin. education at school/work only” group is more likely than the “taught by parents only” group to display the short-term behaviors at a significant level. t.-m. yeh / financial services review 30 (2022) 297–320 307 t ab le 3 a v er ag e m ar g in al ef fe ct s (a m e ) o f o rd er ed p ro b it re g re ss io n o f th e sh o rt -t er m fi n an ci al b eh av io rs p re d ic te d p ro b ab il it y o f st ro n g ly ag re ei n g : c ar ef u ll y co n si d er b ef o re b u y in g p ay b il ls o n ti m e k ee p a cl o se ey e o n fi n an ci al af fa ir s h av e an em er g en cy fu n d a m e p a m e p a m e p a m e p f in . te ac h in g b y p ar en ts o n ly 0 .0 7 1 .0 0 0 0 .0 6 9 .0 0 0 0 .0 8 3 .0 0 0 0 .1 0 5 .0 0 0 f in . ed u ca ti o n at sc h o o l/ w o rk o n ly 0 .0 1 6 .2 1 2 �0 .0 2 3 .0 9 0 .0 4 7 .0 0 0 0 .0 6 2 .0 0 0 b o th at h o m e an d sc h o o l/ w o rk 0 .0 8 5 .0 0 0 0 .0 5 5 .0 0 2 0 .1 2 4 .0 0 0 0 .1 5 9 .0 0 0 f in . li te ra cy 0 .0 2 1 .0 0 0 0 .0 4 4 .0 0 0 0 .0 2 3 .0 0 0 0 .0 6 9 .0 0 0 f em al e 0 .0 1 4 .0 2 5 0 .0 8 3 .0 0 0 0 .0 1 3 .0 0 9 0 .1 1 2 .0 0 0 a g e �0 .0 0 2 .0 0 0 0 .0 0 5 .0 0 0 0 .0 0 2 .0 0 0 0 .0 0 6 .0 0 0 h av e lo an s 0 .0 1 4 .0 1 5 �0 .0 6 9 .0 0 0 �0 .0 5 9 .0 0 0 �0 .1 5 5 .0 0 0 c iv il se rv an t �0 .0 2 2 .0 7 3 �0 .0 3 2 .0 3 1 0 .0 0 9 .3 8 1 0 .0 4 7 .0 0 3 s el fem p lo y ed 0 .0 4 0 .0 0 0 �0 .0 3 6 .0 0 1 0 .0 1 0 .2 2 3 �0 .0 4 8 .0 0 0 p ar tti m er s 0 .0 3 4 .0 0 0 �0 .0 2 4 .0 1 0 0 .0 0 8 .2 1 1 �0 .0 5 9 .0 0 0 h o u se -w o rk 0 .0 3 1 .0 0 0 0 .0 2 3 .0 1 8 0 .0 4 1 .0 0 0 0 .0 1 4 .1 6 1 s tu d en t �0 .0 1 3 .3 3 1 0 .0 5 8 .0 0 0 0 .0 5 3 .0 0 0 �0 .1 6 3 .0 0 0 n o t em p lo y ed 0 .0 5 6 .0 0 0 0 .0 2 0 .0 4 3 0 .0 3 4 .0 0 0 �0 .0 2 1 .0 3 0 > 0 an d < 2 .5 m il . y en 0 .0 1 8 .2 5 6 0 .0 2 2 .1 6 1 0 .0 3 4 .0 0 1 0 .0 7 4 .0 0 0 > 2 .5 an d < 5 m il . y en 0 .0 0 2 .8 7 9 0 .0 6 7 .0 0 0 0 .0 5 0 .0 0 0 0 .1 5 6 .0 0 0 > 5 an d < 7 .5 m il . y en �0 .0 2 7 .1 0 1 0 .1 1 1 .0 0 0 0 .0 5 8 .0 0 0 0 .2 2 4 .0 0 0 > 7 .5 an d < 1 0 m il . y en �0 .0 5 2 .0 0 3 0 .0 9 8 .0 0 0 0 .0 6 7 .0 0 0 0 .2 5 6 .0 0 0 > 1 0 an d < 1 5 m il . y en �0 .0 7 3 .0 0 0 0 .1 2 5 .0 0 0 0 .0 8 0 .0 0 0 0 .2 9 6 .0 0 0 > 1 5 m il . y en �0 .1 4 0 .0 0 0 0 .1 2 3 .0 0 0 0 .0 9 5 .0 0 0 0 .3 4 0 .0 0 0 4 -y ea r co ll eg e 0 .0 0 7 .2 2 9 0 .0 0 9 .1 2 4 0 .0 1 5 .0 0 1 0 .0 5 3 .0 0 0 g ra d u at e 0 .0 1 4 .2 7 5 �0 .0 0 5 .7 2 0 0 .0 4 3 .0 0 0 0 .0 9 2 .0 0 0 r es id en ce ar ea s y es — y es — y es y es — w al d x 2 5 5 8 .1 .0 0 0 2 4 0 3 .0 0 0 1 6 5 3 .0 0 0 5 1 7 8 .5 .0 0 0 p se u d o r 2 0 .0 0 9 — 0 .0 5 0 — 0 .0 2 4 — 0 .1 9 0 — c o ef fi ci en t eq u al it y te st x 2 p x 2 p x 2 p b y p ar en ts o n ly = at sc h o o l/ w o rk 1 5 .3 .0 0 0 3 6 .9 .0 0 0 1 0 .2 .0 0 1 7 .0 .0 0 8 a t h o m e an d sc h o o l/ w o rk = b y p ar en ts o n ly 0 .7 .4 0 4 0 .6 .4 3 4 1 0 .0 .0 0 2 7 .9 .0 0 5 a t h o m e an d sc h o o l/ w o rk = at sc h o o l/ w o rk 1 2 .4 .0 0 0 1 2 .4 .0 0 0 2 4 .1 .0 0 0 1 7 .7 .0 0 0 308 t.-m. yeh / financial services review 30 (2022) 297–320 t ab le 4 a v er ag e m ar g in al ef fe ct s (a m e ) o f o rd er ed p ro b it re g re ss io n o f lo n g -t er m fi n an ci al b eh av io rs s to ck in v es tm en t a w ar e o f re ti re m en t ex p en se s h av e a p la n s et as id e p o st -r et ir em en t fu n d s a m e p a m e p a m e p a m e p f in . te ac h in g b y p ar en ts o n ly 0 .0 2 0 .0 0 5 0 .0 6 4 .0 0 0 0 .0 9 3 .0 0 0 0 .0 5 8 .0 0 0 f in . ed u ca ti o n at sc h o o l/ w o rk o n ly 0 .1 6 3 .0 0 0 0 .1 7 0 .0 0 0 0 .1 4 1 .0 0 0 0 .0 9 6 .0 0 0 b o th at h o m e an d sc h o o l/ w o rk 0 .1 2 8 .0 0 0 0 .1 6 1 .0 0 0 0 .1 9 4 .0 0 0 0 .1 3 9 .0 0 0 f in an ci al li te ra cy 0 .0 7 5 .0 0 0 0 .0 2 8 .0 0 0 0 .0 2 9 .0 0 0 0 .0 1 2 .0 0 0 f em al e �0 .0 6 9 .0 0 0 �0 .0 1 3 .1 8 8 0 .0 1 2 .2 4 2 0 .0 0 2 .8 3 0 a g e 0 .0 0 5 .0 0 0 0 .0 1 1 .0 0 0 0 .0 0 9 .0 0 0 0 .0 1 0 .0 0 0 d u m m y fo r h av in g lo an s �0 .0 2 9 .0 0 0 �0 .0 4 4 .0 0 0 �0 .0 7 5 .0 0 0 �0 .0 9 0 .0 0 0 c iv il se rv an t �0 .0 3 9 .0 0 6 0 .0 2 5 .1 8 4 0 .0 3 2 .0 9 4 0 .0 6 5 .0 0 0 s el f� em p lo y ed �0 .0 0 8 .4 9 5 �0 .0 1 6 .3 3 3 �0 .0 0 2 .9 1 0 0 .0 1 1 .4 1 5 p ar tti m er s �0 .0 4 2 .0 0 0 �0 .0 1 5 .2 5 7 �0 .0 0 4 .7 2 6 0 .0 1 9 .0 8 2 h o u se -w o rk �0 .0 0 6 .5 3 0 0 .0 2 9 .0 2 4 0 .0 3 7 .0 0 4 0 .0 9 6 .0 0 0 s tu d en t �0 .1 0 4 .0 0 0 �0 .0 0 2 .9 5 8 �0 .0 7 2 .0 3 3 0 .0 5 9 .1 6 2 n o t em p lo y ed �0 .0 1 7 .0 7 1 0 .0 5 8 .0 0 0 0 .0 6 9 .0 0 0 0 .1 1 4 .0 0 0 h o u se h o ld in co m e < 2 .5 an d > 0 m il . y en 0 .0 0 9 .6 0 4 �0 .0 3 2 .2 7 3 �0 .0 0 3 .9 2 6 �0 .0 3 9 .1 0 5 > 2 .5 an d < 5 0 .0 4 1 .0 2 3 �0 .0 2 4 .3 9 9 0 .0 1 8 .5 3 4 0 .0 2 4 .3 0 5 > 5 an d < 7 .5 0 .0 7 3 .0 0 0 �0 .0 1 5 .6 0 5 0 .0 4 4 .1 3 6 0 .0 4 8 .0 5 1 > 7 .5 an d < 1 0 0 .0 8 6 .0 0 0 0 .0 2 4 .4 3 1 0 .0 9 1 .0 0 3 0 .1 0 5 .0 0 0 > 1 0 an d < 1 5 0 .1 2 3 .0 0 0 0 .0 4 4 .1 7 9 0 .1 1 5 .0 0 0 0 .1 6 6 .0 0 0 > 1 5 0 .1 5 0 .0 0 0 0 .1 0 6 .0 0 8 0 .1 9 6 .0 0 0 0 .2 8 7 .0 0 0 4 -y ea r co ll eg e 0 .0 4 7 .0 0 0 0 .0 1 1 .2 0 2 0 .0 0 8 .3 3 8 0 .0 2 6 .0 0 0 g ra d u at e 0 .0 7 7 .0 0 0 0 .0 5 6 .0 0 3 0 .0 7 3 .0 0 0 0 .0 7 2 .0 0 0 r es id en ce ar ea s y es y es — y es — y es — w al d x 2 4 5 9 2 2 2 8 5 .7 .0 0 0 1 8 0 4 .1 .0 0 0 2 5 1 6 .4 .0 0 0 p se u d o r 2 0 .1 8 6 0 .1 3 5 — 0 .1 1 6 — 0 .2 2 0 — c o ef fi ci en t eq u al it y te st b y p ar en ts o n ly = at sc h o o l/ w o rk 9 5 .5 8 0 .0 0 0 2 2 .9 .0 0 0 5 .6 .0 1 8 4 .7 .0 3 0 a t h o m e an d sc h o o l/ w o rk = b y p ar en ts o n ly 4 2 .3 3 0 .0 0 0 1 5 .6 .0 0 0 1 9 .2 .0 0 0 1 7 .3 .0 0 0 a t h o m e an d sc h o o l/ w o rk = at sc h o o l/ w o rk 3 .0 9 0 .0 7 9 0 .1 .7 7 9 3 .6 .0 6 0 3 .3 .0 7 0 t.-m. yeh / financial services review 30 (2022) 297–320 309 regarding the control variables, financial literacy is positively and significantly associated with all financial behaviors, which is consistent with previous studies. females are more prudent in short-term behaviors and less likely to invest in stocks. however, females are not statistically different from males in retirement savings, probably because they may share the financial resources with their male spouses in the same household. high-earning individuals are also more financially behaved, except “carefully consider before purchase,” which makes sense as they can afford not to do so. senior people have the same tendency as higher-earning people. seniors may not be able to “carefully consider before purchase” as younger ones, probably due to weakening cognitive capability. in contrast, having a loan makes one more careful in buying but is adversely associated with all other financial behaviors. those with advanced education degrees perform better than those without in most financial behaviors. finally, occupations appear to matter, but the coefficients are generally difficult to interpret. in summary, the results suggest that those receiving financial teaching both at home and school/work are the best financially prepared long-term. however, unlike the case for shortterm behaviors, the association is primarily attributed to financial education at school or the workplace, which is more influential than that at home. 4.3. instrumental variables estimation results one concern about including financial literacy as a control variable in studies of financial behavior is that it may be endogenous (lusardi & mitchell, 2014). therefore, i use the instrumental variables (iv) method to address this concern. in existing research, instruments used for financial literacy include mathematical ability during teens (gathergood & weber, 2017; jappelli & padula, 2013), the experience of family members (behrman et al., 2012; rooij et al., 2011), and the number of universities or newspapers circulating in the neighborhood (klapper et al., 2013). in this study, due to data availability, i used the number of nikkei newspapers per household circulating in the respondent’s residing prefecture as the iv variable. furthermore, following bannier and schwarz (2018) and yeh and ling (2022), in addition to this external instrument, i also used instruments constructed by heteroscedasticity, an estimation method developed by lewbel (2012), when no or insufficient external instruments were available. as the lewbel estimator is based on linear regression models, i only apply iv estimation for binary dependent variables. table 5 reports the lewbel estimates for the regressions. the first-stage regression results indicate that the external instrument is positively and significantly associated with financial literacy (p < .01), satisfying the exclusion restriction. the lewbel (2012) method assumes heteroscedasticity in the errors of the first-stage regression. the white test and breusch and pagan test for heteroscedasticity show that the assumption of heterogeneous error terms is met. the weak instruments test results, the cragg and donald statistic (=15.9 and 6.4, respectively), imply that the null hypothesis of weak instruments is rejected if we are willing to tolerate a 10–20% relative bias based on critical values provided by stock and yogo (2005). 310 t.-m. yeh / financial services review 30 (2022) 297–320 column 1 of table 5 reports the iv estimates for the short-term behavior “have an emergency fund.” the results are similar to those in table 3, but the coefficient for “fin. education at school/work only” is no longer significant (p = .751). the conclusion remains unchanged that short-term behaviors are primarily associated with parenting teaching at home. table 5 results for instrumental variables estimation method . have an emergency fund have a plan set aside retirement funds . coef. p coef. p coef. p fin. teaching by parents only 0.065 0.000 0.097 0.000 0.064 0.000 fin. education at school/work only 0.006 0.751 0.154 0.000 0.109 0.000 both at home and school/work 0.079 0.000 0.208 0.000 0.160 0.000 financial literacy 0.159 0.000 0.023 0.310 �0.007 0.737 female 0.156 0.000 0.010 0.528 �0.005 0.702 age 0.004 0.000 0.009 0.000 0.012 0.000 dummy for having loans �0.166 0.000 �0.082 0.000 �0.105 0.000 civil servant 0.040 0.015 0.033 0.100 0.066 0.000 self-employed �0.048 0.000 �0.003 0.877 0.001 0.963 part-timers �0.049 0.000 �0.005 0.692 0.021 0.038 house-work 0.016 0.104 0.041 0.001 0.105 0.000 student �0.139 0.000 �0.012 0.549 0.145 0.000 not employed �0.023 0.022 0.084 0.000 0.160 0.000 household income <2.5 and >0 mil. yen 0.028 0.102 0.000 0.995 �0.015 0.539 >2.5 and <5 0.097 0.000 0.023 0.413 0.059 0.019 >5 and <7.5 0.156 0.000 0.045 0.136 0.070 0.009 >7.5 and <10 0.177 0.000 0.096 0.003 0.132 0.000 >10 and <15 0.214 0.000 0.124 0.000 0.198 0.000 >15 0.232 0.000 0.214 0.000 0.351 0.000 4-year college 0.011 0.295 0.011 0.453 0.036 0.004 graduate 0.025 0.199 0.079 0.004 0.082 0.000 residence areas yes — yes — yes — constant �0.169 0.000 �0.216 0.000 �0.438 0.000 first-stage regression of financial literacy coef. p coef. p coef. p iv: nikkei newspaper circulation 0.034 0.002 0.048 0.001 0.048 0.001 no. 25000 — 14185 — 14185 — underidentification test kleibergen-paap rk lm statistic 207.6 0.000 88.1 0.000 88.1 0.000 weak identification test cragg-donald wald f statistic 15.943 — 6.433 — 6.433 — heteroskedasticity tests white/koenker nr2 test statistic 356.3 0.000 133.3 0.000 133.3 0.000 breusch-pagan/godfrey/cook-weisberg 248.8 0.000 96.9 0.000 96.9 0.000 coefficient equality test x2 p by parents only = at school/work 13.4 0.000 6.3 0.012 5.0 0.026 at home and school/work = by parents only 0.5 0.461 19.1 0.000 17.7 0.000 at home and school/work = at school/work 11.5 0.001 3.3 0.070 3.7 0.056 note. the results are based on lewbel’s iv estimation method, using the dummy for receiving education at home as the external instrument. regressions are estimated using the heteroskedasticity-robust standard errors. all regressions include residence dummies (not reported). *stock-yogo weak id test critical values for 10% maximal iv relative bias 11.3; 20% maximal iv relative bias 6.08; 30% maximal iv relative bias 4.28. t.-m. yeh / financial services review 30 (2022) 297–320 311 in table 5, columns 2 and 3 report the iv estimates for the long-term behavior “having a financial plan for retirement” and “setting aside funds,” respectively. the results are similar to those in table 4. in both columns, the coefficient for “fin. education at school/work only” is significantly larger than “fin. teaching by parents only” (p < .05), suggesting that long-term financial behaviors are primarily associated with financial education at school/ work. the iv method results for the control variables remain similar to the preceding analyses, except for financial literacy—in columns 2 and 3, financial literacy variables are no longer significant. probably, the effect of financial literacy is now captured by financial education variables in the iv estimation. however, the results regarding financial education remain unchanged. 4.4. matching method results whether one receives financial teaching from parents or school can also be related to socioeconomic factors of the individual’s family background, such as wealth or parents’ educational attainment. the preceding regression specifications address this possibility by including various control variables available from the survey data. this section provides additional tests by matching each individual who received financial teaching with a “control” individual with similar characteristics but without financial teaching by parents or school/workplace. the matching is based on having the same gender, financial literacy scores, age cohort, occupation cohorts, household income cohorts, and education degrees without placement.3 subsequently, i rerun the tests, as in tables 3–4, by using a sample of the treatment and control groups. table 6 reports the marginal effect results for regressions using individuals receiving financial teaching by parents only (treatment group) and their control peers receiving no financial teaching. in the treatment group, 3,399 individuals (95.9%) were matched with a control peer, while 146 (4.1%) failed to find a match. in all columns, the marginal effects of financial teaching by parents are statistically significant (p < .01) and are close in magnitude to those reported in tables 3–4. table 7 reports the marginal effect results for regressions using individuals receiving financial education at school/work only (treatment group) and their control peers. in the treatment group, 820 individuals (97.9%) were matched with a control peer, and only 18 (2.1%) failed to find a match. for short-term financial behaviors (columns 1–4), the marginal effect of financial education is only significant for one variable—“watching financial affairs closely.” however, the marginal effects of financial education are significant for long-term financial behaviors with a greater magnitude. for instance, for the “stock investment” experience variable, financial education at home/workplace has a marginal effect of 15.5%, more significant than the corresponding 2.7% reported in table 6, suggesting a more critical role of financial education at school/work in the long-term financial behaviors. 312 t.-m. yeh / financial services review 30 (2022) 297–320 t ab le 6 p ro p en si ty sc o re m at ch in g an al y si s fo r th e tr ea tm en t g ro u p th at re ce iv ed fi n . te ac h in g b y p ar en ts o n ly v er su s co n tr o l g ro u p th at re ce iv ed n o n e d ep en d en t v ar ia b le : c ar ef u ll y co n si d er b ef o re b u y in g p ay b il ls o n ti m e k ee p a cl o se ey e o n fi n an ci al af fa ir s h av e an em er g en cy fu n d s to ck in v es tm en t h av e a p la n s et as id e re ti re m en t fu n d s . m ea n . m ea n . m ea n . m ea n . m ea n m ea n . m ea n . m at ch ed co n tr o l g ro u p 3 .9 0 6 4 .4 4 8 3 .6 0 9 0 .5 7 5 0 .3 2 4 0 .3 3 0 0 .2 3 2 f in . te ac h in g b y p ar en ts o n ly 4 .0 9 8 4 .5 4 0 3 .8 6 6 0 .6 6 6 0 .3 5 0 0 .4 1 0 0 .2 8 1 p -v al u e fo r t te st o f d if fe re n ce 0 .0 0 0 0 .0 0 0 0 .0 0 0 0 .0 0 0 0 .0 2 4 0 .0 0 0 0 .0 0 0 o rd er ed p ro b it o rd er ed p ro b it o rd er ed p ro b it p ro b it p ro b it p ro b it p ro b it a m e p a m e p a m e p a m e p a m e p a m e p a m e p p re d ic te d p ro b ab il it y o f st ro n g ly ag re e o n th e st at em en t 0 .0 8 1 .0 0 0 0 .0 5 3 .0 0 0 0 .0 8 7 .0 0 0 f in . te ac h in g b y p ar en ts 0 .0 9 2 .0 0 0 0 .0 2 7 .0 0 8 0 .0 8 8 .0 0 0 0 .0 5 9 0 0 0 c o n tr o ls y es y es y es y es y es y es y es n o . 6 ,7 9 8 — 6 ,7 9 8 — 6 ,7 9 8 — 6 ,7 9 8 — 6 ,7 9 8 4 ,1 2 9 — 4 ,1 2 9 — w al d x 2 1 8 1 .0 0 0 5 4 7 .5 .0 0 0 4 0 7 .5 .0 0 0 1 2 8 7 .0 0 0 1 2 0 9 .0 0 0 5 2 1 .0 0 0 7 1 4 0 .0 0 0 p se u d o r 2 0 .0 1 0 — 0 .0 4 4 — 0 .0 2 2 — 0 .1 7 9 — 0 .1 7 9 0 .1 1 8 — 0 .2 3 2 — n o te . a m e = av er ag e m ar g in al ef fe ct s. t.-m. yeh / financial services review 30 (2022) 297–320 313 t ab le 7 p ro p en si ty sc o re m at ch in g an al y si s fo r tr ea tm en t g ro u p th at re ce iv ed fi n . ed u ca ti o n at sc h o o l/ w o rk o n ly v er su s co n tr o l g ro u p th at re ce iv ed n o n e d ep en d en t v ar ia b le : c ar ef u ll y co n si d er b ef o re b u y in g p ay b il ls o n ti m e k ee p a cl o se ey e o n fi n an ci al af fa ir s h av e an em er g en cy fu n d s to ck in v es tm en t h av e a p la n s et as id e re ti re m en t fu n d s . m ea n . m ea n . m ea n . m ea n . m ea n . m ea n . m ea n . m at ch ed co n tr o l g ro u p 3 .8 9 0 4 .4 3 8 3 .6 4 0 0 .5 9 8 0 .4 1 3 4 0 .3 6 7 0 .2 6 3 f in . ed u ca ti o n g ro u p 3 .9 6 3 4 .3 3 5 3 .7 6 8 0 .6 2 0 0 .5 6 9 5 0 .5 0 3 0 .3 4 9 p -v al u e fo r t te st o f d if fe re n ce 0 .1 4 1 0 .0 2 3 0 .0 1 1 0 .3 6 3 0 .0 0 0 0 .0 0 0 0 .0 0 4 o rd er ed p ro b it o rd er ed p ro b it o rd er ed p ro b it p ro b it p ro b it p ro b it a m e p a m e p a m e p a m e p a m e p a m e p a m e p p re d ic te d p ro b ab il it y s tr o n g ly ag re e 0 .0 3 2 .0 9 7 �0 .0 2 9 .1 6 0 0 .0 4 5 .0 0 5 f in . ed u ca ti o n d u m m y 0 .0 2 8 .1 9 1 0 .1 5 5 .0 0 0 0 .1 3 2 .0 0 0 0 .0 9 2 .0 0 0 c o n tr o ls y es y es y es y es y es y es y es n o . 1 6 4 0 — 1 6 4 0 — 1 6 4 0 — 1 6 4 0 — 1 6 4 0 9 5 2 — 9 5 2 — w al d x 2 4 7 .0 .0 0 1 1 9 8 .7 .0 0 0 1 0 4 .2 .0 0 0 3 5 1 .0 .0 0 0 2 6 9 .5 .0 0 0 1 4 5 .1 .0 0 0 2 2 0 .3 .0 0 0 p se u d o r 2 0 .0 1 0 — 0 .0 5 6 — 0 .0 2 4 — 0 .1 8 1 — 0 .1 5 4 0 .1 2 9 — 0 .2 5 3 — n o te . a m e = av er ag e m ar g in al ef fe ct s. 314 t.-m. yeh / financial services review 30 (2022) 297–320 4.5. subsample test results the effect of financial education may vary according to an individual’s socioeconomic background. following wagner and walstad (2019), the last robustness check divides the sample into high and low groups for two control variables to investigate whether the results found until now still hold for the subsample group. the sample is divided by household income (below 750 million yen vs. above) and household financial wealth (below 750 million yen vs. above). for each subsample group, the (ordered) probit regressions were performed, as shown in tables 3 and 4. the untabulated results remain qualitatively similar to the full sample. 5. discussion analyzing a representative sample of 25,000 japanese individuals, the empirical studies revealed that short-term financial behaviors in adulthood are primarily related to financial education received within the family. in contrast, long-term financial behaviors are related to financial teaching received at school or the workplace and, to a lesser extent, at home. the results can be explained using the developmental framework suggested by cfpb. it is plausible that short-term financial behavior is shaped within the family as children develop sound financial values and attitudes from direct and indirect socialization with family members, particularly parents, leading to behaviors such as paying bills on time. such socialization and experiences within the family can also strengthen executive functions, resulting in behaviors such as careful consideration when making a purchase, delaying gratification, and closely watching one’s financial affairs. financial teaching in the family is effective because parents can provide timely negative feedback or punishment when children deviate from these practices. such experiences can influence one’s attitude or behavior when turning into adulthood, as suggested by studies by lusardi et al. (2010) and van campenhout (2015). in contrast, financial education delivered at school or the workplace may be, in general, aimed at improving knowledge of personal finances and budgeting skills, which are an instrument in long-term financial behaviors such as retirement planning and saving, and home buying and homeownership (fox et al., 2005). even if education at school or workplace covers short-term financial behaviors, schoolteachers and instructors may not be able to provide regular and immediate feedback, particularly if good practices are not followed, as parents or family members do. the results are consistent with those of wagner and walstad (2019) and bayer et al. (2009). financial teaching by parents can still contribute to long-term financial behaviors but to a lesser extent than that at schools or workplaces. the weaker association may be due to the more complex tasks and advanced financial knowledge required for long-term planning, which is better served by later-stage financial education in schools or workplaces. a comparison of japan and other advanced countries can further shed insights. in fact, japan reported lower self-reported parental and school/workplace financial education responses than the united states and dutch, where many previous related studies are t.-m. yeh / financial services review 30 (2022) 297–320 315 available. for example, the proportion of dutch who were taught by their parents how to manage a budget is 77.4% (out of 2,676 individuals), as reported by sansone et al. (2019). it is higher than the 17.8% in japan (this study). the proportion of the u.s. sample reported by grinstein-weiss et al. (2011) is about 70% (out of 2,389 individuals). in addition, wagner and walstad (2019) reported that 23% (out of 24,729 u.s. sample) receive financial education either at school or workplace, while the proportion in japan (this study) is no more than 7%. not surprisingly, japanese individuals also have lower financial literacy than other developed countries. for instance, for a set of comparable financial literacy questions, japanese respondents reported a correct rate of 58%, lower than germany of 67% and the united states of 65% (central council for financial services information, 2016). furthermore, compared with other developed countries, the japanese hold a lower proportion of stocks in their financial assets. according to oecd’s household financial assets data, in 2015, among the g7 countries, japanese households held the lowest percentage, at 8.9%, of shares and other equity in their financial assets, against 30.3% of the united states, in terms of 20052015 average.4 the situation of low exposure to financial education in japan hardly changed. the more recent japanese financial literacy survey results done in the year 2019, published recently, compiled that 20.3% received parental teaching, as compared with 19.8% in the 2016 survey of this study, and 8.9% received financial education at the workplace/school, as compared with 8.4% in 2016 (central council for financial services information, 2019). one possible reason for low financial literacy and education is the life-long employment, which may be starting to change somewhat, but has remained a norm in the past decades. until recently, employees have been automatically enrolled in a defined-benefit pension scheme, under which they do not need to work out financial matters by themselves. only in the recent decade have some japanese companies started to shift to a defined contribution scheme gradually. another reason may be due to the limited coverage and teaching hours in the japanese school curriculum, as described in the literature review section. given the circumstances, it is imperative to strengthen financial education programs targeted at parents, school students, and employees in japan. 5.1. limitations this study analyses self-reported data from fls questionnaires. since the respondents’ financial education experiences were not strictly assigned in a random manner, it is not possible to claim a rigorously causal relationship. i used the iv and matching method to address this concern, but they may not completely solve the endogeneity issue. however, even if endogeneity might explain the positive association between financial education and financial behaviors, it seems less plausible as an explanation for the largely insignificant effects of financial education at school/workplace on short-term financial behaviors. nonetheless, i used more cautionary wording when describing the results. omitted variable problems may also exist. although parental advising is out of the control of the children, whether parents provide financial teaching may be related to factors such as 316 t.-m. yeh / financial services review 30 (2022) 297–320 family background (tang et al., 2015) that might be associated with the children’s financial behaviors in their adulthood. empirical tests of this study cannot account for such family background factors as the survey do not contain such information. the survey data used in this study may entail recall error, as respondents were asked to recall their financial education experiences which can date back a long time ago.5 although not being able to address the concern completely, i conducted a robust check by running the primary tests on a subset of samples aged below 30, who may be subject to recall errors to a lesser extent. the conclusion remains unaltered. nonetheless, caution is necessary when interpreting the results as recall errors cannot be totally ruled out. another limitation is that the fls did not reveal what kind of financial education was offered, although such knowledge can better contribute to understanding the relationship between financial education and behavior. future research may address this issue by obtaining detailed information on the types and intensities of the educational activities. 5.2. implications despite the limitations, the results of this study have several implications. the results illuminate the importance of financial education within the family, which appears to have a strong association with financial behavior into adulthood. it is useful for public or private educational or financial institutions to provide advice or training programs targeted at parents regarding how to communicate financial advice effectively so that children can develop proper financial attitudes and habits during daily life opportunities. another insight from this study is the importance of cumulative financial education. this suggests that financial education has an accumulative impact. to promote financial education, educational policymakers may adopt a more integrative design that incorporates programs designed at different stages of life. educational programs should be designed with distinct objectives, with earlier stage programs focusing more on short-term financial behaviors (e.g., executive functions and proper financial attitudes and habits) and later-stage programs emphasizing long-term behaviors (e.g., knowledge and skills for asset management and retirement saving). individuals should be encouraged and provided opportunities to receive accumulative financial education at different stages of life, as it is the most effective way to influence financial behavior. notes 1 one exception is wagner and walstad (2019), which isolated financial education received at high schools, colleges, and workplaces. 2 the residence area dummies are kyushu (prefectures of fukuoka, kagoshima, kumamoto, miyazaki, nagasaki, oita, okinawa, and saga), shikoku (ehime, kagawa, kochi, and tokushima), chugoku (hiroshima, okayama, shimane, tottori, and yamaguchi), keihan (kyoto, osaka), kinki (hyogo, nara, shiga, and wakayama), chubu (aichi, fukui, gifu, ishii, mie, nagano, niigata, shizuoka, and t.-m. yeh / financial services review 30 (2022) 297–320 317 toyama), tokyo, kanto (chiba, gunma, ibaragi, kanagawa, saitama, tochigi, and yamanashi), and tohoku (akita, aomori, fukushima, hokkaido, iwate, miyagi, and yamagata). 3 the results remain qualitatively similar when multiple matches are allowed, with replacements, using the nearest-neighbor criteria. 4 https://data.oecd.org/hha/household-financial-assets.htm 5 i am grateful to the referees for suggesting some of the limitations. acknowledgments the author is grateful to the anonymous referees and the editor-in-chief for their comments. the author also acknowledges financial support from jsps kakenhi (jp17k03807), and education and research center for mathematical and data science of kyushu university, and transcosmos foundation. references ashby, j. s., ingrid, s., & paul, w. (2011). save now, save later? linkages between saving behavior in adolescence and adulthood. european psychologist, 16, 227-237. atkinson, a., & messy, f. a. (2011). assessing financial literacy in 12 countries: an oecd/infe international pilot exercise. journal of pension economics and finance, 10, 657-665. bannier, c. e., & schwarz, m. (2018). genderand education-related effects of financial literacy and confidence on financial wealth. journal of economic psychology, 67, 66-86. batty, m., collins, j. m., o’rourke, c., & odders-white, e. (2020). experiential financial education: a field study of my classroom economy in elementary schools. economics of education review, 78, 102014. bayer, p. j., bernheim, b. d., & scholz, j. k. (2009). the effects of financial education in the workplace: evidence from a survey of employers. economic inquiry, 47, 605-624. behrman, j. r., mitchell, o. s., soo, c. k., & bravo, d. (2012). how financial literacy affects household wealth accumulation. the american economic review, 102, 300-304. bernheim, d. b., garrett, d. m., & maki, d. m. (2001). education and saving: the long-term effect of high school financial curriculum mandates. journal of public economics, 80, 435-465. bucciol, a., & veronesi, m. (2014). teaching children to save: what is the best strategy for lifetime savings? journal of economic psychology, 45, 1-17. burke, j., collins, j. m., & urban, c. (2020). does state-mandated financial education affect financial wellbeing? center for financial security working paper. available at https://centerforfinancialsecurity.files. wordpress.com/2020/09/state-fin-ed-well-being_burke_collins_urban_final.pdf central council for financial services information. (2016). financial literacy survey: 2016 results. available at https://www.shiruporuto.jp/e/survey/kinyulite/pdf/16kinyulite.pdf central council for financial services information. (2019). financial literacy survey: 2016 results. available at https://www.shiruporuto.jp/e/survey/kinyulite/pdf/19kinyulite.pdf cole, s., paulson, a., & shastry, g. k. (2016). high school curriculum and financial outcomes the impact of mandated personal finance and mathematics courses. journal of human resources, 51, 656-698. consumer financial protection bureau. (2016). building blocks to telp youth achieve financial capability. washington, dc: consumer financial protection bureau. despard, m. r., friedline, t., & martin-west, s. (2020). why households lack emergency savings? the role of financial capability. journal of family and economic issues, 41, 542-557. disney, r., & gathergood, j. (2013). financial literacy and consumer credit portfolios. journal of banking & finance, 37, 2246-2254. fernandes, d., lynch, j. g., jr., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors.management science, 60, 1861-1883. 318 t.-m. yeh / financial services review 30 (2022) 297–320 financial education promotion study group. (2014). report on financial education in junior and senior high schools (in japanese). available at https://www.jsda.or.jp/about/kaigi/chousa/kenkyukai/content/jittai_rep. pdf fox, j., bartholomae, s., & lee, j. (2005). building the case for financial education. journal of consumer affairs, 39, 195-214. frisancho, v. (2018). the impact of school-based financial education on high school students and their teachers: experimental evidence from peru. idb working paper idb-wp-871. washington, dc: interamerican development bank. gathergood, j., & weber, j. (2017). financial literacy, present bias and alternative mortgage products. journal of banking & finance, 78, 58-83. grinstein-weiss, m., spader, j., yeo, y. h., taylor, a., & books freeze, e. (2011). parental transfer of financial knowledge and later credit outcomes among lowand moderate-income homeowners. children and youth services review, 33, 78-85. gudmunson, c. g., & danes, s. m. (2011). family financial socialization: theory and critical review. journal of family and economic issues, 32, 644-667. hanson, t. a., & olson, p. m. (2018). financial literacy and family communication patterns. journal of behavioral and experimental finance, 19, 64-71. harvey, m. (2019a). impact of financial education mandates on younger consumers’ use of alternative financial services. journal of consumer affairs, 53, 731-769. harvey, m. (2019b). does state-mandated high school financial education affect savings by low-income households? working paper. available at https://static1.squarespace.com/static/5c4d314bb27e3999d515a9e4/t/ 5e0a1b2841180e2960023175/1577720633380/harvey_fined_savings_working+paper_v20191230.pdf jappelli, t., & padula, m. (2013). investing in financial literacy and saving decisions. journal of banking & finance, 37, 2779-2792. jorgensen, b. l., allsop, d. b., runyan, s. d., wheeler, b. e., evans, d. a., & marks, l. d. (2019). forming financial vision: how parents prepare young adults for financial success. journal of family and economic issues, 40, 553-563. kaiser, t., & menkhoff, l. (2017). does financial education impact financial literacy and financial behavior, and if so, when? the world bank economic review, 31, 611-630. kaiser, t., & menkhoff, l. (2020). financial education in schools: a meta-analysis of experimental studies. economics of education review, 78, 101930. kaiser, t., lusardi, a., menkhoff, l., & urban, c. (2021). financial education affects financial knowledge and downstream behaviors. in print. klapper, l., lusardi, a., & panos, g. a. (2013). financial literacy and its consequences: evidence from russia during the financial crisis. journal of banking & finance, 37, 3904-3923. lewbel, a. (2012). using heteroscedasticity to identify and estimate mismeasured and endogenous and endogenous regressor models. journal of business & economic statistics, 30, 67-80. lusardi, a., & mitchell, o. s. (2007a). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205-224. lusardi, a., & mitchell, o. s. (2007b). financial literacy and retirement preparedness: evidence and implications for financial education. business economics, 42, 35-44. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5-44. lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358-380. messy, f. a., & monticone, c. (2016). financial education policies in asia and the pacific. oecd working papers on finance, insurance and private pensions no, 40, 1-66. miller, m., reichelstein, j., salas, c., & zia, b. (2015). can you help someone become financially capable? a meta-analysis of the literature. the world bank research observer, 30, 220-246. oecd. (2013). advancing national strategies for financial education. available at https://www.oecd.org/finance/ financial-education/g20_oecd_nsfinancialeducation.pdf t.-m. yeh / financial services review 30 (2022) 297–320 319 ooijen, r., & van rooij, m. c. j. (2016). mortgage risks, debt literacy and financial advice. journal of banking & finance, 72, 201-217. rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449-472. rothwell, d. w., & wu, s. (2019). exploring the relationship between financial education and financial knowledge and efficacy: evidence from the canadian financial capability survey. journal of consumer affairs, 53, 1725-1747. sansone, d., rossi, m., & fornero, e. (2019). four bright coins shining at me: financial education in childhood, financial confidence in adulthood. journal of consumer affairs, 53, 630-651. serido, j., & shim, s. (2014). life after college: drivers of young adults success. aplus. arizona pathways to life success. wave 3. the university of arizona. available at https://content.fortune.com/wp-content/ uploads/2014/06/ua-aplus-3-0-final.pdf stock, j. h., & yogo, m. (2005). testing for weak instruments in linear iv regression. in j. h. stock & d. w. k. andrews (eds.), identification and inference for econometric models: essays in honor of thomas j. rothenberg (pp.). cambridge university press. stoddard, c., & urban, c. (2020). the effects of state-mandated financial education on college financing behaviors. journal of money, credit and banking, 52, 747-776. tang, n., baker, a., & peter, p. c. (2015). investigating the disconnect between financial knowledge and behavior: the role of parental influence and psychological characteristics in responsible financial behaviors among young adults. journal of consumer affairs, 49, 376-406. urban, c., schmeiser, m., collins, j. m., & brown, a. (2020). the effects of high school personal financial education policies on financial behavior. economics of education review, 78, 101786. van campenhout, g. (2015). revaluing the role of parents as financial socialization agents in youth financial literacy programs. journal of consumer affairs, 49, 186-222. wagner, j., & walstad, w. b. (2019). the effects of financial education on short-term and long-term financial behaviors. journal of consumer affairs, 53, 234-259. yeh, t.-m. (2022). an empirical study on how financial literacy contributes to preparation for retirement. journal of pension economics and finance, 21, 237-259. yeh, t.-m., & ling, y. (2022). confidence in financial literacy, stock market participation, and retirement planning. journal of family and economic issues, 43, 169-186. 320 t.-m. yeh / financial services review 30 (2022) 297–320 untitled from the editor this issue contains issue 2 of volume 24 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “financial literacy and financial behavior: assessing knowledge and confidence,” is authored by colleen tokar asaad at st. bonaventure university. in this paper the author explores how financial literacy, comprised of both actual financial knowledge and perceived financial confidence, affect financial decisions. her results indicate that financial confidence is a critical component of financial literacy and is important across all knowledge levels. she also shows that overconfident individuals or those with high selfassessed knowledge, but low actual knowledge, have a higher propensity to engage in risky financial behaviors. the second article “financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience,” is coauthored by ning tang and paula c. peter, both at san diego state university. the authors explore how financial education, financial experience, and parents’ financial experience influence young adults’ financial knowledge. using data on 3,597 young adults from a national longitudinal survey, they find that financial education, financial experience, and parents’ financial experience all exert a positive impact on young adults’ financial knowledge. they also find that these determinants work interactively and that both individual and parents’ financial experience help narrow the gap in financial knowledge caused by lack of financial education. the third article, do u.s. households perceive their retirement preparedness realistically?,” is coauthored by kyoung tae kim at university of alabama and sherman d. hanna at the ohio state university. in this paper, the authors examine the divergence between objective and subjective assessment of retirement adequacy, analyzing u.s. households with a full-time worker age 35 to 60 in the 2010 survey of consumer finances. they find that of those households, 58% have objective inadequacy, and 54% have subjective inadequacy, but only 52% have objective/subjective consistency. using a logistic regression, they show that households with defined benefit plans and with defined contribution plans are less realistic than those without plans, and, as age increases, realism decreases. financial services review 24 (2015) v–vi 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. the fourth article, “investment performance of aaii stock screens over diverse markets,” is coauthored by david s. north and jerry l. stevens, both at university of richmond. in this paper, the authors extend prior research on aaii screening performance by including more recent investment periods, employing more rigorous factor models, and examining both median and mean returns to address possible skewness. they show that over a period of tumultuous markets, with as little as $50,000 to invest, over 30% of the available screens achieved statistically significant excess rates of return unrelated to transaction costs and multi-factor risks proposed by efficient market theorists. these results should assist individual investors who often rely on information services and products when making their investment decisions. the final article, “the grable and lytton risk-tolerance scale: a 15-year retrospective,” is coauthored by stephen kuzniak, abed rabbani, wookjae heo, jorge ruiz-menjivar, and john e. grable, all at university of georgia. the authors follow up on the financial risk tolerance scale developed by grable and lytton which was originally published in financial services review in 1999. over that last 15 years this scale has been widely used by consumers, financial advisers, and researchers to evaluate a person’s willingness to engage in a risky financial behavior. their data analysis (n � 160,279) spanning the timeframe 2007 to 2013 provide supporting evidence that the risk-tolerance scale’s reliability and validity have remained robust since the scale was first developed. they found that consistent with the literature, high scale scores, representing a greater willingness to take risks, were found to be associated with equity ownership and negatively related to cash and bond holdings. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 24 (2015) v–vi gratitude, finance, and financial gratitude reminders in charitable giving: a repeated experiment over time yi liua, russell n. james iiib,* adepartment of finance, st. john fisher college, 3690 east avenue, rochester, ny 14618, usa bschool of financial planning, texas tech university, lubbock, tx 79409–1210, usa abstract an initial reminder of three good things (tgt) increases charitable giving intentions, while reminders of three good financial things (tgft) or three financial things (tft) reduce them. repeating these reminders daily during the following seven days results in even higher donation intentions for tgt, but shows no consistent additional effects for tgft or tft. donation intentions measured one or thirty days after stopping these reminders fall significantly faster for tgt. no such effects arise for tgft or tft. gratitude reminders without financial references increase donation intentions, especially when repeated over time. however, this gratitude effect fades after the reminders stop. © 2023 academy of financial services. all rights reserved. jel classifications: d64; d91 keywords: charitable giving; financial gratitude; gratitude; positive psychology; behavioral finance 1. introduction according to giving u.s.a.’s annual report on philanthropy, in 2021, americans contributed about $471.44 billion to charitable organizations. total charitable giving grew by 5.1% in 2021, and approximately 70% of charitable donations came from individuals in 2021 (giving usa, 2021). philanthropic motivations for individual charitable giving have received considerable attention in recent years (krishna, 2011; zlatev & miller, 2016). bekkers and wiepking (2011) review the literature and propose that eight mechanisms determine charitable giving: *corresponding author: tel.: +1-806-787-5931, fax: +1-806-742-1849. e-mail address: russell.james@ttu.edu 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 23–34 awareness of need, solicitation, costs and benefits, altruism, reputation, psychological benefits, values, and efficacy. research on charitable giving has often focused on a socioeconomic perspective (bekkers & de graaf, 2006; brown & ferris, 2007; james, 2011; james & sharpe, 2007; james & wiepking, 2008; mesch et al., 2011, 2006). however, there are other lenses from which to analyze the influences of charitable giving. for example, the influence of positive psychology interventions on charitable giving may be fruitful but has received limited attention in the literature (asebedo & seay, 2015). this study fills the gap in the literature and uses a randomized control-group pretest–posttest experimental longitudinal survey to determine if positive psychology interventions, including gratitude and financial/money reminders, effectively influence charitable giving intentions. 2. literature review gratitude is defined as the acknowledgment and understanding that one has benefited from the kindness or altruism of another (mccullough & emmons, 2003; mccullough et al., 2002, 2001). it is an interpersonal emotion and personal virtue similar to appreciation, which keeps it from being focused on itself. as a positive emotion that has emerged from positive psychology, gratitude increases positive affect, subjective happiness, and life satisfaction (cunha et al., 2019; mccullough & emmons, 2003; mccullough et al., 2002) and is also associated with a great range of social and psychological benefits (watkins, 2014). 3. gratitude and generosity there is a growing body of research showing a significant positive relationship between gratitude and generosity—feeling grateful promotes altruism. gratitude acts as a moral incentive that motivates people to participate in prosocial conduct, either toward the benefactor, toward others, or both (mccullough et al., 2001). specifically, gratitude promotes giving to others, generosity, and social responsibility, while a lack of experiencing gratitude may lead to low donations (isen, 1987). people who demonstrate gratitude feel more inner wealth, which makes them capable of sharing gifts with others (isen, 1987). based on the moral motivation theory, grateful people tend to care about others and participate in various prosocial activities because they are inspired to support others (mccullough & emmons, 2003; mccullough et al., 2001; romani et al., 2013). bock et al. (2018) asserted that the moral trait of gratitude influences charitable giving intentions. they found that gratitude is associated with charitable giving behavior and other positive behaviors like helping others, returning favors, and supporting nonprofits than are emotions like happiness or satisfaction (bock et al., 2018). a fair amount of past research has shown that gratitude is a strong and reliable spur to altruistic action (bartlett & desteno, 2006; desteno et al., 2010; karns et al., 2017; komter, 2004; walker et al., 2016). for example, chaplin et al. (2019) conducted a national survey of adolescents and found that encouraging gratitude in teenagers had the advantage of increasing generosity towards others and reducing their materialism. specifically, the study found that those who kept a gratitude journal, defined by the researchers as a notebook in 24 y. liu and r. n. james / financial services review 31 (2023) 23–34 which participants jotted down things they were grateful for, donated 60% more of their earnings than those in the control group, who did not keep a gratitude journal. in addition to this study by chaplin et al. (2019), other studies have demonstrated a similar correlation between gratitude and generosity. walker et al. (2016) conducted six experiments, each showing that gratitude encourages people’s altruistic behaviors. specifically, participants felt more grateful through experiential purchases than material purchases. participants in the experiential purchases group were more inclined to act altruistically and wrote down more money to be given to the recipients and less money to keep for themselves. another study conducted by liu and hao (2017) on social status and charitable giving found that while reciprocity was important in promoting charitable giving in high-status individuals, feelings of gratitude were the most dominant motivator for low-status individuals. 3.1. financial reminders and generosity research shows that money-related concepts produce robust changes in people’s thoughts, motivations, cognitive states, and behaviors toward others (vohs, 2015; vohs et al., 2006; zhou et al., 2009). financial/money reminders weaken sociomoral and prosocial responses (gasiorowska et al., 2016; mok & de cremer, 2016; savani et al., 2016), encourage people to act independently (vohs et al., 2008), reduce people’s tendency to help others (gasiorowska et al., 2012; tang et al., 2008; vohs, 2015), and decrease altruistic behavior (devoe & pfeffer, 2007; pfeffer & devoe, 2009). after being reminded of money, people are less likely to be interested in volunteering their time to an organization (devoe & pfeffer, 2007; pfeffer & devoe, 2009) and less willing to engage in charitable giving behaviors (roberts & roberts, 2012; vohs et al., 2006, 2008). financial/money reminders may strongly influence donation behaviors (vohs et al., 2006, 2008). for example, vohs et al. (2006) found that when participants were primed with money concepts, they became less prone to donating. the authors suggested that money creates a self-sufficient orientation in which people prefer to be free from dependency and dependents. consequently, money reminders caused people to hold more tightly to their resources, which reduced requests for help and also reduced assistance to others. further research by vohs et al. (2008) found that money-reminded participants donated less money to a university student fund than those who were neutral participants in the control group. specifically, participants in the money-reminded group contributed 39% of their endowment compared with those in the control group, who donated 67%. thus, reminders of money were associated negatively with charitable giving. building upon the works of vohs et al. (2006, 2008), ekici and shiri (2018) focused on the effect of “exposure” of money on charitable giving. the authors conducted an experiment in which participants were shown a donation box containing money, which was either wooden (opaque condition) or transparent. the authors found that participants in the transparent box treatment group were less likely to donate money and donate less money to charities than participants in the opaque box treatment group. this study highlighted the effect of the degree of money exposure on charitable giving. additionally, money reminders on charitable giving can be observed not only in adults but also in adolescents (roberts & roberts, 2012) and young children (gasiorowska et al., y. liu and r. n. james / financial services review 31 (2023) 23–34 25 2012, 2016). roberts and roberts (2012) conducted an experiment involving 114 adolescents aged 13 to 14 who were randomly assigned to either money-reminder or control groups. the authors found that adolescents in the money-reminder group gave less money to the food bank than those in the control group. additionally, children who had been reminded about money showed a lower level of generosity as they preferred not to share their stickers with their partners than children in the control group (gasiorowska et al., 2012). therefore, these studies show a negative relationship between money reminders and the amount of money donated to charities. this paper contributes to previous research in two ways. first, the paper connects research findings on gratitude reminders and financial reminders by simultaneously examining the relationship between gratitude reminders, financial reminders, and financial gratitude reminders with charitable giving intentions. specifically, this study identifies how reminders impact charitable giving intentions for three good things (tgt), three good financial things (tgft), and three financial things (tft) interventions. second, the paper contributes to the literature by exploring how the effects of each reminder on charitable giving intentions change over time, either when repeated or not. 3.2. hypotheses research shows that financial reminders could affect one’s behavior and attitudes (roberts & roberts, 2012; vohs et al., 2006, 2008) as they reduce people’s tendency to help others (tang et al., 2008) and result in less charitable giving (roberts & roberts, 2012; vohs et al., 2006, 2008). at the same time, feelings of gratitude foster charitable giving (mccullough & emmons, 2003; mccullough et al., 2002). the effect of positive psychology and financial reminders on charitable giving decisions is investigated through three hypotheses: hypothesis 1: the “three good things” intervention will result in an increased likelihood of charitable giving compared to the control group. hypothesis 2: the “three financial things” intervention will result in a decreased likelihood of charitable giving compared to the control group. hypothesis 3: these effects will increase over time if repeated, but diminish over time if not repeated. because the “three good financial things” intervention combines both positively associated (gratitude) and negatively associated (financial reminders) interventions, no prediction is made as to the effects on charitable giving intentions. 4. sample and methodology 4.1. sample a total of 993 people participated in the experiments. each participated in a randomized control-group pretest–posttest experimental survey administered on the qualtrics platform. given the use of human subjects, this study was reviewed and approved by the human 26 y. liu and r. n. james / financial services review 31 (2023) 23–34 research protection program (irb2018-582) of the second author’s affiliated university. participants were recruited using the amazon mechanical turk (mturk) recruitment service, an online web service that connects researchers to individuals willing to complete tasks for compensation. survey responses from such participants produce results similar to those generated from traditional nonprobability samples (hauser & schwarz, 2016). researchers in a variety of disciplines have found that this source of participants produces reasonable and consistent results similar to other methods of participant recruitment (buhrmester et al., 2011; goodman et al., 2013). participant characteristics are reported in table 1. random assignment to four groups (tgt, tft, tgft, and no intervention) resulted in similar sample sizes for the three intervention groups (n = 246, 243, and 241), with the control group having a slightly larger sample size (n = 263). socio-demographic and economic characteristics were relatively similarly distributed across groups as reflected by table 1. 4.2. experimental methodology on day 1, all participants first responded to an initial baseline question regarding charitable giving intentions. they were asked, “if you were asked in the next 3 months, what is the likelihood you might give money to each of the following organizations? please rate the likelihood from 0% to 100%” about eight nonprofit organizations. each subject was randomly assigned to one of four groups: tgt, tgft, tft, and no intervention (control group). those assigned to an intervention then read the following, “in this exercise, you will remember and list three [good/good financial/financial] things that have happened in your day and reflect on what caused them. these things can vary from relatively small to relatively large in importance to you; these things can be related to any area of your [life/ financial life/financial life] such as [tgt: relationships, work, school, leisure, physical and mental health, spirituality, money, daily living, transportation, and so forth/tgft or tft: spending, saving, budgeting, planning, giving, investing, daily financial transactions, thoughts/feelings about money, conversations with others about money, earning money, and so forth] by completing this table 1 descriptive statistics all three good things three good financial things three financial things no intervention n 993 246 243 241 263 female 48.9% 46.8% 48.2% 49.0% 45.6% male 51.1% 53.3% 51.9% 51.0% 54.4% married 63.5% 66.7% 60.5% 63.9% 63.1% not married 36.5% 33.3% 39.5% 36.1% 36.9% white 77.6% 76.4% 77.0% 80.5% 76.8% other 22.4% 23.6% 23.1% 19.5% 23.2% high school or less 24.0% 25.6% 29.2% 21.2% 20.2% college degree 76.0% 74.4% 70.8% 78.8% 79.9% income < $40k 34.8% 31.7% 36.6% 34.0% 36.9% income $40k–$80k 40.7% 39.4% 42.4% 43.6% 37.6% income > $80k 24.5% 28.9% 21.0% 22.4% 25.5% age (mean) 37.5 36.6 38.1 37.8 37.6 y. liu and r. n. james / financial services review 31 (2023) 23–34 27 exercise, you will intentionally focus on the [good/good financial/financial] things in your life, allowing you to remember the [good/good financial/financial] things that might otherwise have been overlooked.” this was followed by three sets of open-text responses for the following: “think about event [#1/#2/#3] from your day. please record the following: title of event please describe what happened how did this event make you feel at the time? how did this event make you feel later (including now, as you remember it)? explain what you think caused this event — why it came to pass” all participants in all groups were then asked, “taking all things together, how happy would you say you are?” and “how satisfied are you with your current financial situation?” the same procedure (excluding the initial baseline question) using the same reminder tasks was performed on days 2 (one day after the initial survey), 3, 4, 5, 6, and 7. however, on these subsequent days, the task was followed by questions measuring charitable giving intentions. again, this used a 0% to 100% scale and asked, “if you were asked in the next 3 months, what is the likelihood you might give money to each of the following organizations?” about the eight nonprofit organizations referenced initially. no additional reminders occurred after day 7. however, charitable giving intentions were collected once again one day later (on day 8) and 30 days later (on day 38). 4.3. variables the model used these control variables: charitable organization referenced, married status, age, gender, race, education, and income. the charitable organizations were the american cancer society, the nature conservancy, the american humane association, the american red cross, the breast cancer research foundation, ducks unlimited, a local animal shelter, and the salvation army. education was a dichotomous variable equal to 1 if the respondent had a bachelor’s degree or higher and 0 otherwise. income was recorded as three categories: less than $40,000, $40,000 to $80,000, and greater than $80,000. 4.4. model this study estimates the following linear regression model via ordinary least squares: chg� i ¼ b 0 þ b 1tgt þ b 1tgft þ b 1tft þ b 0 ichoþ b 0 jdem þ « where chg� i was a continuous dependent variable that represents the probability of charitable giving. tgt represented the indicator variable of three good things; tgft represents the indicator variable of three good financial things; tft represented the indicator variable of three financial things; cho represented a matrix of eight organization groups; dem was a matrix of demographic variables that comprise married status, age, gender, race, education, and income. 28 y. liu and r. n. james / financial services review 31 (2023) 23–34 5. results the study used multivariate analysis, ordinary least squares linear regression, to investigate the impact of interventions on charitable giving intentions. column 1 of table 2 reports the immediate effect of the intervention on the probability of charitable giving intentions. in line with our hypothesis, the result provides evidence that these initial reminders result in increased giving intentions for tgt and decreased giving intentions for tgft and tft. the results also show that married, white, and male are associated negatively with giving intentions while age is associated positively with giving intentions. column 2 of table 2 presents the results exploring whether repeating these interventions over time has any additional effects. the intervention variable coefficients (tgt, tgft, and tft) reflect the overall propensity for charitable intentions to be different from the control group, controlling for preintervention intentions, across each category. (this is the overall group effect.) the day variable coefficients (day 3, day 4, day 5, day 6, and day 7) reflect the overall propensity for giving intentions to be higher or lower on any particular subsequent day across all groups, relative to the first postintervention measurement. these separately identify overall effects from time and experimental repetition. the interaction variable coefficients are key. they reflect the difference in the effects of each repetition day on the overall propensity for charitable intentions for each intervention group relative to the control group. the tgt intervention results in significantly greater giving intentions on days 4, 5, 6, and 7, relative to the initial postintervention intentions on day 2. although the tgft and tft interventions result in overall lower intentions, these relationships appear not to be consistently higher or lower with repeated interventions, relative to the initial postintervention intentions on day 2. table 3 reports how giving intentions change over time after interventions stop. again, the intervention variable coefficients (tgt, tgft, and tft) reflect the overall propensity for charitable intentions to be different from the control group, controlling for preintervention intentions, across each category. (this is the overall group effect.) the day variable coefficients (day 8 in column 1, day 38 in column 2) reflect the overall propensity for giving intentions to be higher or lower on any either subsequent day without continued interventions across all groups, relative to the first postintervention measurement. (this is the overall time effect.) again, the key results are the interaction variable coefficients. these reflect the difference in the effects of time passage without continued interventions on the overall propensity for charitable intentions for each intervention group relative to the control group. (this is how the effect of time passage varies across the groups.) these results indicate that charitable giving intentions decline significantly for the tgt group relative to the control group following one day without repeating the intervention, controlling for overall time effects. a similar result occurs for the tgt group, with roughly similar magnitude, following 30 days without repeating the intervention. no such significant differences arise for either the tft or tgft groups. y. liu and r. n. james / financial services review 31 (2023) 23–34 29 6. conclusion, discussion, and implications following results from previous single-intervention gratitude experiments, the current results confirm that an initial reminder of tgt increases charitable giving intentions. this matches the prediction of hypothesis 1. following results from previous single-intervention financial reminder experiments, the current results confirm that an initial financial reminder table 2 likelihood of giving the following initial and repeated reminders (ols) (1) (2) giving likelihood after initial reminder giving likelihood change when repeating reminders pre-intervention giving likelihood 0.8250 (0.0080)*** 0.8676 (0.0034)*** three good things 1.2856 (0.6341)** 0.1944 (0.3951) three good financial things �1.4211 (0.6444)** �0.7743 (0.3976)* three financial things �1.2655 (0.6211)** �0.7348 (0.3907)* day (reference: day 2) day 3 0.2518 (0.6551) day 4 �0.277 (0.4933) day 5 0.4025 (0.6597) day 6 �0.2872 (0.6449) day 7 1.4372 (0.6449) day3*three good things 1.4372 (0.9938) day4*three good things 2.0165 (0.7627)*** day5*three good things 2.2466 (1.0418)** day6*three good things 2.3890 (0.8180)*** day7*three good things 2.2662 (1.0104)** day3*three good financial things �0.156 (1.0071) day4*three good financial things 0.2561 (0.7683) day5*three good financial things �0.1426 (1.0509) day6*three good financial things 2.1306 (0.8177)*** day7*three good financial things 1.1254 (1.0532) day3*three financial things �0.8422 (0.9725) day4*three financial things �0.8904 (0.7554) day5*three financial things �0.437 (1.0208) day6*three financial things �0.1872 (0.8163) day7*three financial things 0.1217 (0.9978) charitable organization (reference: american cancer society) the nature conservancy 0.0580 (0.9190) 0.0372 (0.3472) the american humane association 1.3324 (0.9178) 1.8311 (0.3467)*** the american red cross 0.0074 (0.9177) 0.0131 (0.3466) breast cancer research foundation 0.2274 (0.9181) 0.2469 (0.4093) ducks unlimited �0.5413 (0.9280) 0.1925 (0.4132) a local animal shelter 1.5303 (0.9186) 1.1602 (0.4095)** the salvation army 0.9466 (0.9178) 0.9960 (0.4092)** married �1.3847 (0.5031)*** �0.7408 (0.2124)*** age 0.0754 (0.0220)*** 0.0228 (0.0091)** male �1.0259 (0.4735)** �1.2185 (0.1982)*** white �3.1780 (0.5928)*** �1.9368 (0.2491)*** college 0.7090 (0.4836) 0.7972 (0.2033)*** income (reference: 40k) $40k–$80k 2.5625 (0.5430)*** 1.3094 (0.2299)*** $>80k 2.5879 (0.6533)*** 1.1147 (0.2762)*** significance levels: *p < .05, **p < .01, ***p < .001. 30 y. liu and r. n. james / financial services review 31 (2023) 23–34 (tft) reduces charitable giving intentions. this matches the prediction of hypothesis 2. in a new result, a combination of gratitude and financial reminders (tgft) is found to initially reduce charitable giving intentions. hypothesis 3 predicted that these effects will increase over time if repeated, but diminish over time if not repeated. this was confirmed for the gratitude reminder (tgt). repeating these reminders daily during the following seven days results in even higher donation intentions for tgt, controlling for overall time effects. however, results from repeating the interventions over time were not consistent for the financial or financial gratitude reminders (tft, tgft). donation intentions measured one or thirty days after stopping these reminders fell significantly faster for tgt. no such effects arose for tgft or tft. thus, hypothesis 3 is confirmed for gratitude reminders, but not for financial reminders or financial gratitude reminders. charitable decision-making is a topic of interest to social science researchers for a variety of reasons. however, it is also an important topic for practice among fundraisers and financial advisors. financial counselors and financial planners are likely to work with charitably inclined clients. fundraisers will work almost exclusively with those who are charitably inclined. table 3 likelihood of giving after reminders stop (ols) giving likelihood change when reminders stop for 1 day giving likelihood change when reminders stop for 30 days pre-intervention giving likelihood 0.7987 (0.0066)*** 0.6339 (0.0079)*** three good things 2.3623 (0.6919)*** 2.0752 (0.6919)** three good financial things 1.3103 (0.7032)*** 1.3306 (0.7032) three financial things �0.755 (0.6863) �0.5109 (0.6863) day (reference day 7) day 8 (1 day after reminders stop) 0.9258 (0.6456) day 8*three good things �3.5280 (1.0256)*** day 8*three good financial things �1.7271 (1.0045) day 8*three financial things �1.6588 (1.0091) day 38 (30 days after reminders stop) �0.3126 (0.6456) day 38*three good things �3.4133 (1.0256)*** day 38*three good financial things �0.1877 (1.0045) day 38*three financial things �0.6507 (1.0091) charitable organization (reference: american cancer society) the nature conservancy �0.0358 (0.6750) �1.3301 (0.6750) the american humane association 2.0601 (0.6739)*** 1.206 (0.6739) the american red cross �0.4325 (0.6739) �0.3501 (0.6739) breast cancer research foundation �0.4962 (0.7546) �0.5615 (0.7546) ducks unlimited �0.0395 (0.7628) �3.2147 (0.7628)*** a local animal shelter 1.1487 (0.7550) 1.2199 (0.7550) the salvation army 0.9442 (0.7544) �0.6316 (0.7544) married �0.2749 (0.4084) �1.4006 (0.4084)*** age �0.0200 (0.0174) 0.0001 (0.0174) male �1.8485 (0.3779)*** �1.1589 (0.3779)** white �1.5147 (0.4806)*** �1.6529 (0.4806)*** college 1.6071 (0.3898)*** 0.5016 (0.3898) income (reference: 40k) $40k–$80k 1.6700 (0.4424)*** 1.6730 (0.8443)** $>80k 0.69312 (0.5303) 2.4702 (1.0273)** significance levels: *p < .05, **p < .01, ***p < .001. y. liu and r. n. james / financial services review 31 (2023) 23–34 31 these results confirm the importance of gratitude, and gratitude reminders, as a motivator for charitable giving decisions. additionally, they show that the impact of these reminders is stronger when they are repeated over time and fades when that repetition stops. this is important for fundraisers and financial advisors to understand. the power of gratitude references is not fully realized from a one-time reference. instead, they can be more powerful if incorporated into regular, repeated references or conversations. additionally, these results confirm past research showing that money reminders tend to reduce interest in charitable giving. first focusing on the client’s desired philanthropic impact, rather than financial spreadsheets, may lead to more interest in charitable giving. also, these results suggest that whereas general gratitude reminders lead to increased charitable giving, financial gratitude reminders do not. references to gratitude for all things lead to increased charitable giving intentions, whereas references to gratitude strictly for financial things do not. again, this matches with the idea of starting philanthropic conversations with broader, nonfinancial motivations, rather than financial references. the present findings have practical for fundraising professionals and financial services professionals in understanding the behavioral intentions of donors. as a practical matter, it may be better for fundraising professionals not to begin by drawing attention to financial reminders, which increases the money salience for potential donors, but instead, emphasize areas of gratitude and focus on the “good things” or the impact the donations that could bring in. thus, gratitude reminders should be acknowledged to promote fundraising. financial advisors who wish to discuss philanthropic planning with clients may have better conversations when emphasizing broad concepts such as gratitude while reducing money reminders. 7. limitations some limitations must be acknowledged when evaluating these findings. first, this study is focused on the measure of charitable giving intentions, not the estimations of actual giving. as a result, extending our findings to actual altruistic behaviors should be done with caution. given the substantive importance of donation magnitude, more study is needed to see how the individuals’ characteristics and motivations examined here affect the charitable giving intentions related to the absolute amount of charitable giving, rather than a simple yes/no choice. second, participants were asked to write about financial behaviors during a stock market downturn in the united states (the s&p 500 index fell 15% during the month of this study’s data collection in december 2018), as well as during the peak christmas shopping season. it may be insightful for future experimental research to determine how gratitude and financial interventions impact charitable giving intentions in different months and economic cycles. references asebedo, s. d., & seay, m. c. (2015). from functioning to flourishing: applying positive psychology to financial planning. journal of financial planning, 28, 50–58. 32 y. liu and r. n. james / financial services review 31 (2023) 23–34 bartlett, m. y., & desteno, d. (2006). gratitude and prosocial behavior: helping when it costs you. psychological science, 17, 319–325. https://doi.org/10.1111/j.1467-9280.2006.01705.x bekkers, r., & de graaf, n. d. (2006). education and prosocial behavior. unpublished typescript. bekkers, r., & wiepking, p. (2011). who gives? a literature review of predictors of charitable giving part one: religion, education, age and socialisation. voluntary sector review, 2, 337–365. https://doi.org/10.1332/ 204080511x6087712 bock, d. e., eastman, j. k., & eastman, k. l. (2018). encouraging consumer charitable behavior: the impact of charitable motivations, gratitude, and materialism. journal of business ethics, 150, 1213–1228. https://doi. org/10.1007/s10551-016-3203-x brown, e., & ferris, j. m. (2007). social capital and philanthropy: an analysis of the impact of social capital on individual giving and volunteering. nonprofit and voluntary sector quarterly, 36, 85–99. https://doi.org/ 10.1177/0899764006293178 buhrmester, m., kwang, t., & gosling, s. d. (2011). amazon’s mechanical turk: a new source of inexpensive, yet high-quality, data? perspectives on psychological science: a journal of the association for psychological science, 6, 3–5. https://doi.org/10.1177/1745691610393980 cunha, l. f., pellanda, l. c., & reppold, c. t. (2019). positive psychology and gratitude interventions: a randomized clinical trial. frontiers in psychology, 10, 584. https://doi.org/10.3389/fpsyg.2019.00584 chaplin, l. n., john, d. r., rindfleisch, a., & froh, j. j. (2019). the impact of gratitude on adolescent materialism and generosity. the journal of positive psychology, 14, 502–511. https://doi.org/10.1080/17439760.2018.1497688 desteno, d., bartlett, m. y., baumann, j., williams, l. a., & dickens, l. (2010). gratitude as moral sentiment: emotion-guided cooperation in economic exchange. emotion, 10, 289–293. https://doi.org/10.1037/a0017883 devoe, s. e., & pfeffer, j. (2007). hourly payment and volunteering: the effect of organizational practices on decisions about time use. academy of management journal, 50, 783–798. https://doi.org/10.5465/ amj.2007.26279171 ekici, a., & shiri, a. (2018). the message in the box: how exposure to money affects charitable giving. marketing letters, 29, 137–149. https://doi.org/10.1007/s11002-018-9452-y gasiorowska, a., chaplin, l. n., zaleskiewicz, t., wygrab, s., & vohs, k. d. (2016). money cues increase agency and decrease prosociality among children: early signs of market-mode behaviors. psychological science, 27, 331–344. https://doi.org/10.1177/0956797615620378 gasiorowska, a., zaleskiewicz, t., & wygrab, s. (2012). would you do something for me? the effects of money activation on social preferences and social behavior in young children. journal of economic psychology, 33, 603–608. https://doi.org/10.1016/j.joep.2011.11.007 giving usa. (2021). the annual report on philanthropy for the year 2020. chicago: giving usa foundation. goodman, j. k., cryder, c. e., & cheema, a. (2013). data collection in a flat world: the strengths and weaknesses of mechanical turk samples. journal of behavioral decision making, 26, 213–224. https://doi.org/ 10.1002/bdm.1753 hauser, d. j., & schwarz, n. (2016). attentive turkers: mturk participants perform better on online attention checks than do subject pool participants. behavior research methods, 48, 400–407. https://doi.org/10.3758/ s13428-015-0578-z isen, a. m. (1987). positive affect, cognitive processes, and social behavior. in l. berkowitz (ed.), advances in experimental social psychology (vol. 20, pp. 203–253). cambridge, ma: academic press. james, r. n., iii., & sharpe, d. l. (2007). the nature and causes of the u-shaped charitable giving profile. nonprofit and voluntary sector quarterly, 36, 218–238. https://doi.org/10.1177/0899764006295993 james, r. n., iii., & wiepking, p. (2008). a comparative analysis of educational donors in the netherlands. international journal of educational advancement, 8, 71–78. https://doi.org/10.1057/ijea.2008.9 james, r. n., iii., (2011). charitable giving and cognitive ability. international journal of nonprofit and voluntary sector marketing, 16, 70–83. https://doi.org/10.1002/nvsm.402 karns, c. m., moore, w. e., iii., & mayr, u. (2017). the cultivation of pure altruism via gratitude: a functional mri study of change with gratitude practice. frontiers in human neuroscience, 11. https://doi.org/10.3389/ fnhum.2017.00599 y. liu and r. n. james / financial services review 31 (2023) 23–34 33 komter, a. e. (2004). gratitude and gift exchange. in r. a. emmons & m. e. mccullough (eds.). the psychology of gratitude (pp. 195–212). oxford: oxford university press. krishna, a. (2011). can supporting a cause decrease donations and happiness? the cause marketing paradox. journal of consumer psychology, 21, 338–345. https://doi.org/10.1016/j.jcps.2011.02.001 liu, c. j., & hao, f. (2017). reciprocity belief and gratitude as moderators of the association between social status and charitable giving. personality and individual differences, 111, 46–50. https://doi.org/10.1016/j. paid.2017.02.003 mccullough, m. e., & emmons, r. a. (2003). counting blessings versus burdens: an experimental investigation of gratitude and subjective well-being in daily life. journal of personality and social psychology, 84, 377–389. https://doi.org/10.1037//0022-3514.84.2.377 mccullough, m. e., emmons, r. a., & tsang, j. a. (2002). the grateful disposition: a conceptual and empirical topography. journal of personality and social psychology, 82, 112–127. https://doi.org/10.1037//0022-3514.82.1.112 mccullough, m. e., kilpatrick, s. d., emmons, r. a., & larson, d. b. (2001). is gratitude a moral affect? psychological bulletin, 127, 249–266. https://doi.org/10.1037/0033-2909.127.2.249 mesch, d. j., brown, m. s., moore, z. i., & hayat, a. d. (2011). gender differences in charitable giving. international journal of nonprofit and voluntary sector marketing, 16, 342–355. https://doi.org/10.1002/nvsm.432 mesch, d. j., rooney, p. m., steinberg, k. s., & denton, b. (2006). the effects of race, gender, and marital status on giving and volunteering in indiana. nonprofit and voluntary sector quarterly, 35, 565–587. https:// doi.org/10.1177/0899764006288288 mok, a., & de cremer, d. (2016). the bonding effect of money in the workplace: priming money weakens the negative relationship between ostracism and prosocial behaviour. european journal of work and organizational psychology, 25, 272–286. https://doi.org/10.1080/1359432x.2015.1051038 pfeffer, j., & devoe, s. e. (2009). economic evaluation: the effect of money and economics on attitudes about volunteering. journal of economic psychology, 30, 500–508. https://doi.org/10.1016/j.joep.2008.08.006 roberts, j. a., & roberts, c. r. (2012). money matters: does the symbolic presence of money affect charitable giving and attitudes among adolescents? young consumers, 13, 329–336. https://doi.org/10.1016/j.joep.2008.08.006 romani, s., grappi, s., & bagozzi, r. (2013). explaining consumer reactions to corporate social responsibility: the role of gratitude and altruistic values. journal of business ethics, 114, 193–206. https://doi.org/10.1007/ s10551-012-1337-z savani, k., mead, n. l., stillman, t., & vohs, k. d. (2016). no match for money: even in intimate relationships and collectivistic cultures, reminders of money weaken sociomoral responses. self and identity, 15, 342–355. https://doi.org/10.1080/15298868.2015.1133451 tang, t. l. p., sutarso, t., davis, g. m. t. w., dolinski, d., ibrahim, a. h. s., & wagner, s. l. (2008). to help or not to help? the good samaritan effect and the love of money on helping behavior. journal of business ethics, 82, 865–887. https://doi.org/10.1007/s10551-007-9598-7 vohs, k. d. (2015). money priming can change people’s thoughts, feelings, motivations, and behaviors: an update on 10 years of experiments. journal of experimental psychology. general, 144, e86–e93. https://doi. org/10.1037/xge0000091 vohs, k. d., mead, n. l., & goode, m. r. (2006). the psychological consequences of money. science, 314, 1154–1156. https://doi.org/10.1126/science.1132491 vohs, k. d., mead, n. l., & goode, m. r. (2008). merely activating the concept of money changes personal and interpersonal behavior. current directions in psychological science, 17, 208–212. walker, j., kumar, a., & gilovich, t. (2016). cultivating gratitude and giving through experiential consumption. emotion, 16, 1126–1136. https://doi.org/10.1037/emo0000242 watkins, p. c. (2014). gratitude and the good life: toward a psychology of appreciation. dordrecht, the netherlands: springer. zhou, x., vohs, k. d., & baumeister, r. f. (2009). the symbolic power of money. psychological science, 20, 700–706. https://doi.org/10.1111/j.1467-9280.2009.02353.x zlatev, j. j., & miller, d. t. (2016). selfishly benevolent or benevolently selfish: when self-interest undermines versus promotes prosocial behavior. organizational behavior and human decision processes, 137, 112–122. https://doi.org/10.1016/j.obhdp.2016.08.004 34 y. liu and r. n. james / financial services review 31 (2023) 23–34 the perfect withdrawal amount: a methodology for creating retirement account distribution strategies e. dante suareza,*, antonio suarezb, daniel t. walzc atrinity university, department of finance and decision sciences, san antonio, tx 78212, usa bindependent financial advisor, monterrey, mexico ctrinity university, department of finance and decision sciences, san antonio, tx 78212, usa abstract we present a new way to develop withdrawal strategies from retirement portfolios. it is derived analytically, instead of from empirical testing, and it iterates always in the same manner. it is based on a new measure we develop, the perfect withdrawal amount, for which we discuss how to construct a probability distribution and how to apply it sequentially. we also derive a new measure of sequencing risk. we present new strategies built with this framework. © 2015 academy of financial services. all rights reserved. jel classification: j26; d81; d14 keywords: retirement; optimal withdrawal; safe withdrawal rate; withdrawal rule; sequencing risk 1. introduction the question of how much to withdraw from a savings account once an investor has entered retirement is now more relevant than ever, as current demographic trends make it a matter of vital importance for a growing number of people every year. the problem itself arises because retirees generally wish to use their retirement funds to support a standard of living that is as high as possible, but without depleting their account so quickly that the years still ahead become difficult to finance—the so-called failure risk. on the other hand, withdrawing “too little” money might simply translate into excessively high balances in their * corresponding author. tel.: �1-210-999-7860; fax: �1-210-999-8134. e-mail address: esuarez@trinity.edu (e. d. suarez) financial services review 24 (2015) 331–357 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. accounts at the end of the life span horizon—the so-called surplus risk—and the “golden years” lifestyle would have been restricted unnecessarily. the goal for the researcher is then to develop formulas or “rules” that dictate withdrawal amounts in each period based on the retiree’s age, the assets held in the savings account, and the retiree’s willingness to limit consumption in exchange for future safety. there are, of course, other factors that affect the optimality of these rules (inflation, tax implications, covering for emergency withdrawals, etc.), but the three factors mentioned usually have the largest impact on any recommendation. the research thus far undertaken has produced a considerable body of knowledge. the problem is well understood, and many of the pitfalls of the original, simpler rules have been identified and addressed. the optimal strategies currently available are sophisticated and address a wide range of different situations and scenarios. still, it is our opinion that many of these results are, in a sense, heuristic. these approaches typically start with an idea that “makes sense” intuitively and then test it to see if improvement was indeed attained. in contrast, we have tried to develop an entirely analytical treatment of the problem, starting from the functional relationship between the relevant variables and building up from there before producing any rules or strategies. a central concept in this effort is a new measure that we have called the perfect withdrawal amount (pwa). instead of developing a particular rule for withdrawing, we introduce a fundamentally different way of addressing the problem. 2. previous research the origin of the research program on optimal withdrawals in retirement can be traced back to bengen’s (1994) pioneering article, in which he presents the basis for what would come to be known as “the 4% rule.” specifically, he demonstrates that a 4% withdrawal rate from a retirement fund, adjusted for inflation, is generally sustainable for normal retirement periods. a series of studies by cooley, hubbard, and walz (1998, 1999, 2003, and 2011) then strengthened this conclusion, as they report similar findings using overlapping samples of historical stock and bond returns. from a methodological standpoint, the distinguishing feature of these “first generation” articles is that they rely on a constant withdrawal amount, established from the outset, which is only adjusted to replenish its purchasing power. this has motivated attempts to develop “adaptive” rules, aimed at improving the results by applying midcourse corrections. guyton and klinger (2006) develop performance-based rules that decrease or even cancel the inflationary adjustment when return rates are too low, and that modify the withdrawal amount when the implied withdrawal rate falls outside ranges they prescribe. frank, mitchell, and blanchett (2011) use adjustment rules that depend on how much the rate of return deviates from the historical averages. zolt (2013) proposes curtailing the inflationary adjustment to the withdrawal amount to increase the portfolio’s survival rate, and produces different “rules” by varying the degree to which purchasing power is restored. another take on the adaptive theme has been to reassess the situation periodically, taking into account the shortening of the horizon period. spitzer (2008) considers this effect and 332 e.d. suarez et al. / financial services review 24 (2015) 331–357 resets the withdrawal amount every five years. blanchett and frank (2009) annually recalculates the probability of depleting the retirement funds too soon. if it becomes higher (lower) than a set of critical values they posit, the withdrawal amount is decreased (increased) by 3%; otherwise it remains constant. an additional refinement has been to interpret the planning horizon length as a stochastic variable instead of a parameter. under this view, the goal for the planner is to ensure that the funds in the retirement account “outlive” the retiree (instead of the other way around), no matter the number of years involved. stout and mitchell (2006) use mortality tables to make sure that a retirement period of uncertain length can be covered. stout (2008) decreases the withdrawal amount whenever the account balance falls below a measure of the present value of the withdrawals yet to be made, and increases it when the balance is above this measure plus an additional value reserve. mitchell (2011) uses different thresholds to trigger adjustments and also addresses the risk of superannuation. a still more recent approach has treated the selection of withdrawal amounts as a lifetime-utility maximization problem. milevsky and huang (2011) posit as the objective function the total discounted value of the utility derived across the entire retirement period, in a setting where this length is a stochastic variable and the subjective discount rate is a measure of “personal impatience.” williams and finke (2011) use a similar model with more realistic portfolio allocations and also consider other sources of income. blanchett, kowara, and chen (2012) measure the relative efficiency of different withdrawal strategies by comparing the actual cash flows provided by each strategy to the flows that would have been feasible under perfect foresight. the methodology that we present here can also be used to derive adaptive rules and revisiting schemes, and to perform longevity risk and utility maximization analyses. it is also well-suited for trying out alternative distributions for the rates of return, such as those described in blanchett and blanchett (2008), pfau (2012), and blanchett, finke, and pfau (2014). all of these, however, would now be enhanced by understanding the trajectories that the resulting strategies trace out in the pwa dimension. 3. the perfect withdrawal amount we begin the development of our methodology by assuming a scenario where annual withdrawals are made on the first day of the year and annual returns accrue on the last day.1 there’s no inflation (or, alternatively, the rates of return used are in real terms) and no taxes. this simplifies the setup without removing any crucial element in the relationship between the main variables and, as will be shown below, any additional pieces needed to represent a real-world situation are easily added on top of this skeleton framework. we now posit that for any given series of annual returns there is one and only one constant withdrawal amount that will leave the desired final balance on the account after n years (the planning horizon). this can be verified by solving a problem that is formally equivalent to that of finding the fixed-amount payment that will fully pay off a variable-rate loan after n years. in other words, we rederive the traditional pmt() formula found in financial calcu333e.d. suarez et al. / financial services review 24 (2015) 331–357 lators, but with three amendments: (1) interest rates are not fixed but change in every period, (2) the desired ending value is not necessarily zero, and (3) we are dealing with drawdowns from an asset instead of payments to a liability.2 the basic relationship between account balances in consecutive periods is: ki�1 � (ki � w) (1 � ri) (1) where ki is the balance at the beginning of year i, w is the yearly withdrawal amount, and ri is the rate of return in year i in annual percentage. applying eq. (1) chain-wise over the entire planning horizon (n years), we obtain the relation between the starting balance ks (or k1) and the ending balance ke (or kn): ke � ({[(ks � w) (1 � r1) � w] (1 � r2) � w} (1 � r3) . . . � w) (1 � rn) (2) and we solve eq. 2 for w to get: w � �ks � i�1 n (1 � ri) � ke� � �i�1 n �j�i n (1 � rj) (3) eq. (3) provides the constant amount that will draw the account down to the desired final balance if the investment account provides, for example, a 5% return in the first year, 3% in the second year, minus 6% in the third year, and so forth, or any other particular sequence of annual returns. this figure we call the pwa. if one were to know in advance the sequence of returns that will come up in the planning horizon, one would compute the pwa, withdraw that amount, and reach the desired final balance exactly and just in time.3 to provide a concrete example of the relationship between a sequence of returns and its corresponding pwa, we will use one of the estimation runs involved in the exercise presented in the next section to provide a numerical illustration. suppose that the retirement period is just starting out and that we know in advance that in the next 30 years our investment account will yield the returns presented in fig. 1. fig. 1. sequence of returns produced by monte carlo engine through random drawing from s&p 500 historical data. 334 e.d. suarez et al. / financial services review 24 (2015) 331–357 then, assuming that we want to exhaust our account in full (no inheritance) in 30 years, eq. (3) indicates that each year we should withdraw exactly 7.2556 cents per dollar of initial balance. fig. 2 confirms. therefore, with $1 million as starting balance and a final balance goal of zero, $72,556 is the perfect withdrawal amount for this sequence of returns. we withdraw the same amount every year and reach our desired final balance with no ups and downs in the income stream, no portfolio failure, and no surplus. in this manner we are characterizing every sequence of returns using one particular figure: the corresponding pwa. therefore, from here we argue that the retirement withdrawal question is, at its core, a matter of “guessing” what the pwa will turn out to be (eventually) fig. 2. arithmetical confirmation that the pwa for the sequence in fig. 1, when operating on $1 million starting balance and aiming for $0 ending balance, is $72,556. 335e.d. suarez et al. / financial services review 24 (2015) 331–357 for each retiree’s portfolio and objectives. the methodology presented here can then be understood as a way to go about this guessing in a statistically sound way. specifically, our problem now becomes how to estimate the probability distribution for pwas from the probability distribution for the returns on the assets held in the retirement account. before moving on to explore how to derive the pwas distribution, we should mention several results provided by our approach even at this early stage. first, eq. (3) can be restated in a particularly useful way. the term �i�1 n (1 � ri) in the numerator is simply the cumulative return over the entire retirement period, so we’ll now call it rn. we keep the subscript n to underscore the fact that this is the total return, not the average, and thus it depends on the length of the planning horizon. the denominator, in turn, can be interpreted as a measure of sequencing risk: �i�1 n � j�i n (1 � rj) � (1 � r1) (1 � r2) (1 � r3) . . . (1 � rn) � (1 � r2) (1 � r3) . . . (1 � rn) � (1 � r3) (1 � r4) . . . (1 � rn) � . . . � (1 � rn�1) (1 � rn) � (1 � rn) (4) we note that, for any given (unordered) set of rates, this expression decreases if “big” rates show up at the beginning of the retirement period and “small” rates show up at the end, because the last rates appear more times than the first rates in the summation. to make the interpretation more natural, we’ll say that it is the reciprocal �1/�i�1 n �j�i n (1 � rj)� that captures the sequencing effect, because this item goes up when the sequence is favorable.4 this we call sn, and we note that two r-vectors can have the same rn and yet different sn’s—a given cumulative return can come up in different ways, order-wise. now eq. (3) simplifies to: w � (rnks � ke) sn (5) this provides us with an expression that captures the impact of sequencing explicitly, through a factor (sn) that “hits” the withdrawal amount in a perfectly sensible way. this reformulation is important because the sequencing issue is precisely what makes the optimal withdrawal problem unique. in most financial analysis discussions, an understanding of the total accrued return will suffice, but here it is of the essence to know not just how much but when this much. in retirement, it makes a huge difference if “good” financial results come first and “bad” ones later, instead of the other way around, because in this stage of the life cycle the rates of return apply sequentially to an ever-dwindling capital base—the retiree is constantly drawing down her savings to support herself. this is also why previous studies have sought for a way to account for this “sequencing risk” (frank and blanchett, 2010; frank, mitchell, and blanchett, 2011). indeed, even very recent discussions such as pfau (2014) continue to address this problem, developing proxy variables to measure the correction required because of the sequencing issue. although it is reasonable to expect that some of these constructs will capture the general features of the adjustment factor, our eq. (5) comes directly from the simplest, most natural interpretation of the problem—so sn is not a proxy. rather, we think sn is an expression that should be investigated further because it is a measure of orientation (return rates going up, going down, 336 e.d. suarez et al. / financial services review 24 (2015) 331–357 up a little then down a lot, etc.), and this is the crucial element that the adjustment factor should capture. second, we modify eq. (5) to express the withdrawal rate, instead of the withdrawal amount: w/ks � rnsn � sn (ke/ks) (6) additionally, we note how it is dependent on the ratio ke/ks. this means that to derive an optimal withdrawal rate we need to know what fraction of the initial account balance is to be bequeathed. although this might seem like a trivial adjustment to the existing strategies, which may focus on the zero-bequest case only for the sake of clarity, the specific functional form by which inheritance goals affect the optimal recommendation has not been established before5 (bengen, 2006; bernard, 2011; spitzer, 2008). it is the proportion of the desired final balance to the starting balance that matters and not, as might be assumed, the amount by which the initial balance exceeds the target. for example, a natural way to deal with the complication of a positive bequest goal would be to “put aside” the inheritance money—effectively removing it from the analysis—and work with the excess balance as if it were a zero-bequest case (cf. “two-bucket” strategy in bernard, 2011). however, the separated funds will accumulate returns too, and there is no reason for these returns not to be available for consumption (at least the real part). should the planner separate the funds to be inherited but then plow back the real returns into the “consumable” balance? our approach renders these efforts unnecessary by including a bequest term in the equations, making the no-bequest scenario just a specific instance of the general case. 4. construction of a pwa probability distribution we now proceed to construct an actual probability distribution for pwas, assuming a 100% equity portfolio. the decision to use an all-equity scenario is not crucial because we are only illustrating how the methodology works. this allocation was chosen because it maximizes the variance of the returns on the account’s assets, and thus produces the pwa distribution with the largest dispersion possible. by examining this “worst case scenario” we can get a clearer idea of what pwa distributions look like in general, keeping in mind that for “normal” portfolio allocations they would be more clustered than what is shown here. to apply our model we calculated the monthly returns on the s&p500 for the period january 1957 through april 2013 and then used a monte carlo engine to draw 360 values at random from this set (one at a time, so it’s “with replacement”). these were then interpreted as the monthly returns in a 30-year planning horizon, in the same order as they were drawn, so the first 12 values were compounded to get the annual return for year 1, the next 12 values became year 2, and so on.6 the engine repeated this process 20,000 times and computed the cumulative return (rn) and sequencing factor (sn) for each series of returns obtained. this provided us with 20,000 (rn,sn) pairs, each pair standing in for a realistic vector of return rates with 30 entries.7 fig. 337e.d. suarez et al. / financial services review 24 (2015) 331–357 3 presents the frequency distribution of the pwa formula (eq. (5)) evaluated at each of these 20,000 (rn,sn) pairs, using $1 million as starting balance and with $0 as desired ending balance. fig. 4 is the same but with cutoff points for deciles indicated. some of the advantages provided by our model can be seen in these charts. for example, the tabulations used to produce the charts can be used to locate any withdrawal amount, not just milestone values and not only amounts in the lower end of the range. so we can readily advise on the consequences of withdrawing, say, $62,000 every year; from the underlying tabulation we read that 64% of the monte carlo runs produced pwas higher than $62,000, so failure risk for that withdrawal amount is 36%. this distribution was calculated using $0 as desired ending balance, so the failure risk figure estimates the probability of total ruin. however, if the calculations had included a bequest target, the cumulative areas under the pwa curve would represent probabilities of not reaching this target instead of total depletion. the retiree might be interested in separating these two numbers, and the framework provides a way to do this: just compare the failure risk levels, with and without bequest, at that withdrawal amount. fig. 5 shows this graphically. surplus risk, in turn, can be estimated by inverting the roles of the withdrawal amount and the ending balance as dependent/independent variables in the analysis. for the example above, one would use the same set of (rn,sn) pairs to evaluate ke � rnks � 62,000/sn, and then tabulate the resulting figures.8 we used this feature to estimate surplus risk in this same scenario using $43,000 as withdrawal amount (which is approximately the cutoff point for the first decile of the pwa distribution). as expected, close to 90% of the runs left a positive balance behind—90.9% “didn’t fail”—but other points along the distribution of ending balances deserve mentioning. fig. 3. frequency distribution of annual pwas in 20,000 monte carlo runs, with the specific parameters shown in the header. this distribution is then interpreted as a probability distribution for the optimal withdrawal amount. 338 e.d. suarez et al. / financial services review 24 (2015) 331–357 fig. 4. same as fig. 3, but now indicating the figures that delimit 10% areas under the curve. these would be the “milestone” annual withdrawal amounts for the parameter values given in the header. fig. 5. the continuous line is fig. 3. the broken line is the pwa distribution when the desired ending balance is $500k. failure risk at $62k rises from 36% in the no-bequest graph, to 46% in the $500k bequest-goal graph. this means that the probability of a final balance between $0 and $500k is 10%. 339e.d. suarez et al. / financial services review 24 (2015) 331–357 for example, 74% of the runs ended up with more money than they began with, that is, the final balance was larger than the starting balance (and the desired ending balance was zero). the final balance is two times or more the starting balance in 58% of the runs. if one were to take this “4.3% rule” at face value, a 100% equity portfolio would have a 12% probability of ending up with 10 times or more what it had at the beginning! similar results have been found before and have been fully acknowledged (cooley, hubbard, and walz, 1998), but even so we feel that they call for a reassessment of the “safe withdrawal rate” approach because of the large size of the ending balances associated with it. we think this approach could be used as a conservative guideline to start the series of withdrawals, but understanding that it is very likely that we will be able to increase the withdrawal amount significantly and yet stay inside the desired failure-risk range. precisely, we now move on to the mechanics of how is the withdrawal amount adjusted each year using this methodology. 5. sequential application of the pwa formula per our assumptions, we make a withdrawal at the beginning of retirement year 1 and live off this money for the entire period until, at the very end, the yield accrued during the previous 12 months is actually credited to our account. now our pwa distribution needs to be recalculated using eq. (5), with the balance actually showing as the new starting balance and shortening the time horizon by one period. the shortening of the horizon is attained by substituting in a new set of 20,000 (rn,sn) pairs, obtained with the same procedure described in the previous section, but this time drawing return rate sequences that are shorter than before by 12 data (they represent horizons with one year less to go). from this new distribution we choose the withdrawal amount for year 2, presumably—but not necessarily—using the same risk-tolerance profile as in year 1. furthermore, this process is simply repeated every year. we must stress that in pwa the process by which the withdrawal amount is selected is always the same, but this does not mean that the withdrawal amount itself will not change. actually, it is rare for a withdrawal strategy built with this framework to recommend constant withdrawals beyond a small number of periods. pwa incorporates all new information into the set of data available for the next analyses, and this updating will create adjustment pressures that, in all likelihood, will end up modifying the withdrawal amount at some point. also, although a year has gone by, the planning horizon may be adjusted differently. the framework allows the retiree to judge whether the expected number of years still ahead for her has indeed decreased by one. if, for example, the retiree has stayed particularly healthy, she may decide to use the same horizon length once more. or she may even use a longer horizon, if she has just overcome serious illness or made beneficial lifestyle changes. and, sadly, there will be opposite cases where the appropriate length reduction is larger than just one period. however, other than these discretionary length adjustments, which can be considered part of the information update process, the pwa approach is consistent. it calls for the application of the same procedure to produce the withdrawal menu every year, without triggering different schemes if certain conditions are met. 340 e.d. suarez et al. / financial services review 24 (2015) 331–357 this consistency allows us to compute the optimal withdrawal rates in the no-bequest case for the entire length of the planning horizon, and for all confidence ranges, at the same time. if no bequest is being sought, eq. (6) becomes: w/ks � rnsn (7) fig. 6 presents the results of evaluating this equation with the sets of (rn,sn) pairs corresponding to different horizon lengths (obtained as described above) for the 100% equity case considered here. for example, let us consider a retiree that starts out with $1 million, has no desire to leave behind any money, and thinks that 30 years is a reasonable length for her planning horizon. she has assets indexed to the s&p 500 and wants to be 90% certain that she will not have to lower her withdrawal amount later on. our subject only needs to read from the chart to obtain the corresponding rate. she does this and, following its advice, pulls out $43,000 (4.3%). her first year turns out badly; the stock market goes down 10%, so her remaining balance of $957,000 (after withdrawing the $43k) shrinks to $861,000. should she withdraw $43,000 again in the second year? well, $43k represents 5.0% of her now-current balance and the chart tells us that, with 29 years remaining, that would take her out of the 90% confidence range and into the 80–90% region. those are the factors that need to be considered in the decision—nothing else matters. fig. 6. milestone values in the optimal withdrawal rate distribution for the all-equity, no-bequest case. the rates are with respect to the actual balance faced each year, not the balance at the beginning of the retirement period. if, for example, the investor wishes to be 90% safe, each year she should run down the last row and take out the corresponding fraction of her balance at that point, whatever it may be. 341e.d. suarez et al. / financial services review 24 (2015) 331–357 we make this point to highlight the simplicity of our approach, which cuts through the gordian knot of some of the adaptive rules found in the previous literature (bernard, 2011; blanchett and frank, 2009; guyton, 2004; mitchell, 2011; pye, 2000; robinson, 2007; stout and mitchell, 2006). for example, in guyton and klinger (2006) we find that “withdrawals are to increase from year to year to make up for inflation, except that there is no increase after a year where the portfolio’s total return is negative and when that year’s withdrawal rate would be greater than the initial withdrawal rate” (p. 5), or “when a current year’s withdrawal rate has risen more than 20% above the initial withdrawal rate, the current year’s withdrawal is reduced by 10%; this rule expires 15 years before the maximum age to which the retiree wishes to plan” (p. 7). one can find more rules like these in other studies. the pwa approach would call this type of adjustments into question, claiming instead that the statistically appropriate thing to do is to calculate the probability distribution of pwas for the current situation and choose from there—every year. all the relevant information currently available is embedded in that distribution and all new information is captured by the way the distribution changes as time elapses. an adaptive rule is simply a movement into different confidence ranges of the corresponding distribution and this, unless properly understood and consciously chosen, is an arbitrary call. it must be stressed that these simplified results apply only if no bequest is being sought. when the retiree wishes to leave a certain amount behind, the withdrawal rate is given again by eq. (6), which is the general-case expression: w/ks � rnsn � sn (ke/ks) (6) this can be read as “the withdrawal rate for the no-bequest case, reduced by the sequencing factor times the fraction of the current balance that is to be bequeathed.” the performance of the investment portfolio in previous years then comes back into play, but it is fully captured by the balance actually showing in each period, which is relevant only insofar as it affects the bequest-to-balance ratio. 6. what confidence ranges really tell us at this point it becomes important to clarify what is meant by “confidence range” in this framework. because the framework is adaptive, now an expression such as “90% safe” no longer means that there is a 90% chance that the portfolio will not run out of funds. what it means is that there is a 90% chance that you won’t have to lower your withdrawal amount in the future, to achieve your bequest target. we think this interpretation is more informative for the investor because normally in a real life situation a strategy will not be allowed to fail—the withdrawal amount will be reduced before that happens. so the truly pressing question is then how much would the withdrawal amount need to be lowered if we go into the red zone. the model can produce these quantitative assessments by computing and analyzing alternate pwa distributions, which are relatively easy to obtain because they can be derived from the original distribution. for example, the distribution of the pwa after withdrawing an arbitrary amount w* this year can be obtained using a modified version of eq. (3):9 342 e.d. suarez et al. / financial services review 24 (2015) 331–357 w� � �rn (ks � w*) � ke� � �i�2 n �j�i n (1 � rj) (8) where w’ is the new (modified) pwa for the subsequent years, w* is the withdrawal amount being considered for this year, and we leave the expression in the denominator in its explicit form to note that the summation now starts at period 2. we see that the denominator is now “missing” the product of all the returns from year 1 to year n (the summation starts with the product of returns from year 2 to year n), which is rn, so we can express it as 1 sn � rn. and we now rewrite eq. (8) as: w� � �rn (ks � w*) � ke� sn/(1 � rnsn) (9) finally, we can identify some of these terms as our original pwa, w � (rnks � ke) sn, to see that the modified pwa is a linear transformation of the original: w� � (1 � rnsn)�1 w � rnsn (1 � rnsn)�1 w* (10) we can use this expression to compute different pwa distributions using the set of (rn,sn) pairs that we already have. fig. 7 presents two of these alternate distributions, using w*�$20,000 and w*�$100,000. these alternate distributions are then the answer that the framework provides to our pressing question above. where most other approaches simply provide the change in the failure-risk figure, here we can compute the withdrawal amounts corresponding to different fig. 7. pwa distributions immediately after making the first withdrawal, but before the first yield accrues. the thicker line is fig. 3 (the original distribution). these charts answer the question “what happens if i withdraw x dollars this year,” and are the pwa alternative to failure-risk figures. 343e.d. suarez et al. / financial services review 24 (2015) 331–357 safety levels, with one year less to go, after making a withdrawal of a given amount in the current period. from the point of view of the retiree, this is the probability that she will have to settle for a future withdrawal of x dollars or less, after making a withdrawal of y dollars in the current period. fig. 8, which is also made with eq. (10) but using a table form instead of a chart, shows these results.10 these are examples of a more general feature made available by the pwa framework: the possibility of optimizing withdrawal paths meeting a specific profile request, and not just constant streams.11 this can be achieved by introducing the “shape” of the request into the optimization process. consider an investor that is already in retirement but has not yet reached her social security full retirement age. she might prefer to avoid reductions in her benefits by delaying the time of claiming and use her investment account as her sole source of income for a number of years. this would then require a two-stage withdrawal plan, taking out relatively high amounts at the beginning and then smaller amounts after the benefits start to come in. fig. 8. probability that the pwa will drop below a certain amount after a withdrawal of a given size is made in year 1. assumptions as in fig. 7 header. for example, the highlighted cell shows that if a withdrawal amount of $40,000 (4%) is acceptable for the retiree, then she can take out $110,000 in year 1 and still be “90% safe” with respect to staying clear of this self-defined “red zone.” 344 e.d. suarez et al. / financial services review 24 (2015) 331–357 a plan like this can be optimized by rederiving eq. (3), with the withdrawals after the switch year t equal to the first years’ withdrawals minus the benefits b, so that w1 � w2 � … � wt�1 � wt � b � wt�1 � b � … � wn � b. this scheme produces withdrawal recommendations aimed at providing constant total income throughout; the path of withdrawals from the investment account would have to be “kinked,” but this can be handled by the framework without much difficulty. we close this section by discussing another feature related to confidence levels which enhances the model’s flexibility even more. when the retirement plan includes a bequest, the model provides the investor with an additional adjustment lever: the value of that bequest. by eq. (6), the safety level of a given withdrawal amount depends in part on the relation between the current balance and the desired ending balance. this means that by modifying the bequest goal we can attain different confidence levels without changing the withdrawal amount. for example, with $1 million starting balance and a 30-year planning horizon, if the retiree intends to bequeath $500,000 the 90% confidence level is at $34,600. but if she decided to lower that goal to $400,000 the 90%-safe withdrawal amount would now be $36,157—a 4.5% increase. settling for an inheritance of just $200,000 increases this amount an additional 10.4%, to $39,921. doing away with the bequest goal altogether takes us to fig. 4, where we see that the 90%-safe withdrawal amount is then $43,316. fig. 9 shows these bequest-adjustment profiles for several confidence levels. of course, modifying the bequest goal is also an option for midcourse corrections, and it can be combined with other adjustment levers. for example, if the account gets high rates of return for a number of years, the retiree may decide to split the benefits of the good run by increasing the withdrawal amount, the confidence level, and the bequest goal. the pwa framework will provide the relevant terms-of-trade. fig. 9. withdrawal amounts corresponding to different confidence levels for an all-equity portfolio with $1 m starting balance and 30-year horizon. the amounts depend on the investor’s intended bequest. the highlighted cells show that an investor who wants an annual income close to $52k and wishes to bequest $500k will have to accept 70% safety. however, if she does away with her bequest intentions her safety level rises to 80%. 345e.d. suarez et al. / financial services review 24 (2015) 331–357 7. development and testing of a withdrawal rule inside the pwa framework once more, the purpose of this article is not to derive a particular withdrawal strategy but to present new tools for constructing and assessing strategies in general. still, by way of illustration we now present strategies built with this toolkit. suppose that a major concern for a strategy under consideration is whether the withdrawal paths obtained from it will be too “jumpy”—that is, we are worried that the recommended withdrawal amount might turn out to be very different between one year and the next. this may arise if, for example, the strategy demands us to stay inside a confidence range that is very narrow, so that the withdrawal amount has to keep moving up and down to fall back into the mandated confidence range. we may address this concern by proposing a two-tiered strategy that starts by selecting a certain amount of failure risk, but then “tolerates” some degree of variation around this value for the sake of steadiness in the income flow. suppose we choose 50% as the acceptable initial value of failure risk. anything “safer” than that is not agreeable to the retiree because she feels it creates too much surplus risk—she has no heirs and wants to consume her savings down to the last penny. we begin by withdrawing the median amount in the 30-year pwa distribution for an all-equity portfolio (because that is what she happens to have). after making this first withdrawal and getting the return in our account at the end of the year, the withdrawal amount for year 2 is chosen as follows. if the amount withdrawn in the previous year is still inside the 15%-to-85% confidence range (a rather wide 70-point interval) once the pwa distribution is recalculated starting from the now-current balance and with 1 year less remaining in the planning horizon, that same amount is withdrawn once more. otherwise the strategy is “recentered” by taking the median value of the recalculated distribution. this process is then repeated every year. after setting up our monte carlo engine with these parameters, we have run simulations of the withdrawal paths resulting from this strategy using the now customary $1 million starting balance and $0 desired ending balance. figs. 10, 11, and 12 present three of these simulated paths, each one corresponding to return sequences that give rise to (somewhat) flat, rising, and decreasing withdrawal profiles.12 barring perfect foresight, nothing strictly better can be provided for a retiree with these demands and holding these assets. if we want to improve the results of the strategy along either one of the three relevant dimensions (safety, income level, stability), we would have to change either the base safety level, the recentering point or the width of the tolerance range. this last point is an important realization about the nature of the problem at hand. by the structural relationship between the variables, when the parameters of a strategy are modified to enhance one specific property, it is inevitable for one or more of the other properties to suffer. this in turn means that the design of optimal withdrawal strategies is not a search for the dominant scheme; it is a search for the strategy that best suits the needs of the investor being served. the next section closes the discussion by looking more closely at this “planner-as-tailor” aspect of the framework. 346 e.d. suarez et al. / financial services review 24 (2015) 331–357 8. a tale of three strategies we will now use the sequence of returns shown in fig. 1 to present three different strategies, once again starting from $1 million and ending at $0. as can be seen in fig. 1, the dispersion in that sequence is significant (standard deviation 14.5 points) and it is quite jumpy, with the values ranging from �12.0% in year 29 to 44.2% in year 8. therefore, this particular sequence may be considered a worthy challenge for any strategy to handle, even though it’s rather “normal” (fig. 2 informs us that its pwa in this case is $72,556; figs. 3 and 4 shows this amount is quite close to the median value, and in the peak region of the chart). the purpose of this exercise is not to establish which strategy is better. the results shown represent not just the strategy followed, but the specific return sequence used as well. to compare strategies one would have to average their respective performance statistics over a large number of randomized runs to determine the frequency with which they produce “good” or “bad” outcomes, where good and bad are defined by the retiree. one can compare the situation faced by the retiree to that of a shopper in a clothing store right after a big sale. the items in the tables and bins are still misplaced because a horde of customers sorted through them haphazardly the day before. but the most efficient way to find an item of her size and liking is, still, to look in the places where it would normally be; there is a higher probability of finding it there than elsewhere. the fig. 10. this strategy kicks off by withdrawing the median value in the 30-year pwa distribution. then it becomes “sticky,” withdrawing the same amount unless it is “too safe” (more than 85% confidence) or “too risky” (less than 15% confidence). when this happens, the strategy again takes the median value of the pwa distribution applicable at that point in time. the pwa for this particular sequence of returns was $84,792 which, by fig. 4, is somewhat on the high side. 347e.d. suarez et al. / financial services review 24 (2015) 331–357 pwa framework would then be the knowledgeable clerk who will ask you what you want and then will tell you where it is more likely that you may find it. following this metaphor, this section is a sampler of some of the “items carried” in an all-equity portfolio facing a 30-year planning horizon. the first strategy considered is the 4% rule, which withdraws $40,000 every year and is presented in fig. 13. we already know that the pwa for this sequence is $72,556, and therefore that this strategy will be too cautious and end up with surplus balance. the balance reaches $3,013,737 in year 30, which shows in the figure as the convergence point of the tolerance range boundaries (only reference points now). these boundaries always come together at the end because in this framework there is only one possible recommendation at that point: take out all you have, because this is the last year. after continuing to withdraw $40,000 (one last time) and accruing the final return of 9.3%, the account bequeaths $3,250,295. we see that the overaccumulation trend is evident early on, as the confidence level reaches 98% in year 3 after starting out at 93%—perhaps overly high even from the outset. by year 15 the confidence level goes “off the charts,” as there is not a single run in our simulation for that horizon length that produces a pwa of $40,000 or lower starting from $1,832,363 (the balance in year 15). these confidence levels “above 100%” persist until the end of the retirement period. the second strategy is the one we used in the previous section, which we now show in fig. 14 as “sticky median with 70-point tolerance and recentering.” as discussed above, this strategy uses the central value (the median) in the 30-year pwa distribution as the initial fig. 11. same as fig. 10, but this time the sequence of returns to which the strategy was applied had a pwa of $118,498. fig. 4 shows this is unusually high, so the strategy is “surprised” and reacts by increasing the withdrawal amount markedly and frequently. 348 e.d. suarez et al. / financial services review 24 (2015) 331–357 withdrawal amount. then it focuses on stability and withdraws the same amount every year, but only if it is “reasonable” to do so. when taking out the same amount implies falling outside the 70-point confidence interval centered at 50% (between 85% and 15%), we adjust the withdrawal and go back to the center of the current pwa distribution. therefore, the strategy effectively starts over or “recenters” when it deems that it is no longer prudent to favor steadiness over safety (failure-wise or surplus-wise). the median of the 30-year pwa chart is $70,465 (see fig. 4), so the strategy kicks off with that, and then it keeps the amount steady for a number of years. we see how the upper bound gets close to the withdrawal amount in year 7, but the strategy stays the course because the confidence level is 25% (low, but still allowed). the opposite happens in year 20, when the lower bound closes in, but again the reins are held tight because the confidence level only reaches 75%. the tolerance range is not “pierced” until year 23, when holding on to the previous withdrawal amount would put the account at 91% confidence—an unacceptably high risk of overaccumulation. we use the balance at that point ($615,189) to compute the pwa distribution with 8 years to go (year 23 is only starting out), and take out the median value: $94,648. this new withdrawal amount is almost decreased only two years later, when the confidence level drops to 16%, but the tolerance range is only grazed and the strategy holds steady. the balance at the beginning of year 30 is $71,908, which is taken out in full as final withdrawal and creates the steep drop at the end of the chart (�24%). the drop is of course caused by the terrible return rate in year 29 (�12.0%), but in a sense it can also be attributed to the high volatility of this portfolio allocation. still, this final amount is higher than what fig. 12. same as figs. 10 and 11, but here the sequence of returns had a pwa of $53,691. per fig. 4, this value is in the lower half of the distribution, so the strategy is unable to support the median withdrawal amount and adjusts to a lower level. 349e.d. suarez et al. / financial services review 24 (2015) 331–357 the retiree got during the first two decades and is also higher than what would have been provided by the 4% rule. the third strategy is a variation of the second one. it is somewhat exotic, as we modify the three main parameters of sticky median to illustrate the flexibility in strategy design provided by the pwa framework. this version kicks off by withdrawing the amount at the 75% safety level in the pwa distribution, so it is more conservative than sticky median, which starts at 50%. then we narrow the tolerance range around the base safety level to 30 points, so it now spans the interval from 60% to 90%. this means it stays closer to the level of risk with which the retiree feels most comfortable (compared with sticky median), but it demands more frequent adjustments to the withdrawal amount because it allows less “straying.” therefore, the third change is in how are the adjustments applied; instead of going back to the original safety level (75%), this time the strategy “pushes” the withdrawal amount (up or down) by the minimum amount necessary to get back into the tolerance range. this adjustment procedure is more subtle than recentering, but it asks for the retiree to live at the very edge of his risk preference in some years. these changes produce a peculiar but very reasonable strategy, which we have called “sticky first quartile with 30 point tolerance and nudging” and is shown in fig. 15.13 the most noticeable change in the resulting withdrawal path is its upward trend, which is a result of the higher safety level used. the initial withdrawal amount is $55,025 (75% safe) fig. 13. the 4% rule is used with the sequence of returns in fig. 1. as the planning horizon shortens and the account balance rises, it becomes less likely that $40,000 will end up being the pwa. this increases the confidence level, even though the withdrawal amount is always the same. after 30 years, the account bequeaths more than $3 million. 350 e.d. suarez et al. / financial services review 24 (2015) 331–357 and from there the lower bound of the tolerance range keeps nudging the amount higher. the only downward adjustment comes at the very end, and is very slight. after withdrawing $199,620 in year 29 and taking the decrease of �12.0%, the balance reaches $199,197; this is the last withdrawal. these three sets of results shed some light on the consequences of pursuing different goals. an all-out desire to provide withdrawal stability would follow something akin to the 4% rule, even if it means bequeathing an amount much larger than what was intended. an approach that attempts to maximize income without going out on a limb would prefer something like the sticky median strategy, even if the withdrawal amount can suffer major jolts in either direction. if the primary concern is safety, and cash flow steadiness is not essential, a strategy similar to sticky first quartile may be in order even though it is likely to result in a markedly rising withdrawal path. fig. 16 superimposes the three withdrawal paths for a more direct comparison. again, this exercise must not be taken as evidence in favor or against either strategy. if instead of the sequence in fig. 1 we had used the one underlying fig. 12 (the “decreasing” path), the interpretation might have been different. fig. 17 shows the values in that “bad” sequence, which has a pwa of just $53,691 starting from $1 million and aiming for $0. fig. 18 is the chart of the resulting withdrawal paths, which may be compared with fig. 16 for contrast. fig. 14. same as fig. 13, but now the strategy followed is “sticky median with 70-point tolerance and re-centering.” the withdrawal amount is much higher than with the 4% rule and yet retains some degree of stability. here the entire remaining balance is withdrawn in year 30 (no bequest is desired), but this final withdrawal is still “inside the ballpark” set by the previous amounts. 351e.d. suarez et al. / financial services review 24 (2015) 331–357 9. conclusions and next steps a sequence of return rates, together with a starting balance and a desired ending balance, determines a “perfect” withdrawal amount. if the retiree’s goal is to obtain a constant stream of income from her investment account, this is the amount she should take out in every period. if she withdraws more than the pwa, she will run out of funds (or leave behind less money than she intended). if she withdraws less, she will leave money on the table. the problem, of course, is that we do not know in advance what the pwa will turn out to be in each case. to make inferences and take decisions one must construct probability distributions, and the methodology presented here is a way to do this soundly. here we use the historic distribution of return rates as the probability distribution of future returns, assuming it to be independent and identical in all periods. however, any other assumption may be used, as our framework only discusses how to “process” the assumptions (or expectations) held by the researcher. at this stage, we can only make preliminary recommendations. however, one that seems to take shape intuitively is a major departure from conventional wisdom: the best level of failure risk is 50%. to be more precise, we should use the mode of the pwa distribution— because that is the most likely value that the pwa will end up taking in our particular case. for symmetric or moderately skewed distributions, failure risk at the mode will be close to 50%. if the assumptions used for the distribution of return rates fig. 15. same as figs. 13 and 14, but now the strategy followed is “sticky first quartile with 30-point tolerance and nudging.” compared with fig, 14, this strategy is more cautious (it drifts around 75% safety instead of 50%), more strict (tolerance 30 points instead of 70), and more subtle (upon piercing the tolerance the amount is nudged back in instead of recentered). 352 e.d. suarez et al. / financial services review 24 (2015) 331–357 are sound, then close to half of the retirees who follow this policy will be able to keep withdrawing this amount (or more!) for the entire retirement period. for the other half, the procedure outlined here will steer them clear of funds’ depletion or bequest collapse through timely warnings that the amount needs to be decreased—and probably not by fig. 16. strategy comparison. the withdrawal paths in figs. 13, 14, and 15 are superimposed, with the tolerance range lines removed for clarity. the 4% rule provides a low cash flow and leaves a sizable bequest, so the standard of living in retirement is overly restricted. sticky median provides more money, but failure risk reaches very high levels along the way. sticky first quartile plays it safer, but at the expense of frequent adjustments to the amount withdrawn. fig. 17. another random sequence of returns produced by monte carlo engine. this is the sequence that gives rise to fig. 12. 353e.d. suarez et al. / financial services review 24 (2015) 331–357 much. if ours is indeed the case when the sequence of returns awaiting in our future is a terrible one, then the procedure will keep signaling for downward adjustments, taking the withdrawal path asymptotically to a very low value, but probably not too far below what it would have been under perfect foresight. the obvious next step for the framework’s future development is to use it to derive new “standard” rules—à la bengen’s 4%—for different portfolio compositions and retirement objectives, and to look at the rules currently in favor through the lens of pwa. also, it would be useful to add longevity risk into the framework by making the planning horizon a stochastic variable instead of a parameter. and it would seem possible to adapt the model to the accumulation phase of the life-cycle, deriving a perfect contribution amount concept along the lines of pwa. once we know more about the pwa for different portfolio compositions, we can proceed to the derivation of optimal adaptive strategies that produce specific results along any of the relevant dimensions. here we would be interested in exploring “gliding” strategies that change the asset-mix over time, as well. finally, we see an opportunity to improve the computation of the pwa probability distributions by using more advanced methods to construct the probability profiles for the financial assets’ returns. the bootstrapping procedure used here is very practical, but we are aware of its numerous drawbacks. fig. 18. comparison of the same three strategies as in fig. 16, but now the withdrawal paths are the ones produced by the sequence in fig. 17 instead of fig. 1. the 4% rule is closer to the other two in this case, but it still results in a significant bequest of $735,059 when none was sought. 354 e.d. suarez et al. / financial services review 24 (2015) 331–357 notes 1 although not entirely realistic, the assumption of annual withdrawals has been the norm in withdrawal rate studies since the “founding” of the field by bengen (1994). 2 the fact that this is a mirror image of the paying-off-a-loan situation means that the variable “r” in the typical formula is now a return rate instead of an interest rate, the ending value is available capital instead of unpaid balance, and the result of the formula is a payment from the account instead of a payment into the loan. 3 blanchett, kowara, and chen (2012) presents a measure similar to pwa called sustainable spending rate (ssr). our pwa was developed independently, but it turns out to be a generalization of ssr, ssr being the pwa when the starting balance is $1 and the desired ending balance is zero. 4 the sequencing factor formula is the same as the sinking fund factor under variable rates of return, but the interpretation and application is thoroughly different in each case. 5 actually, considering only formal, mathematically explicit models, the authors were unable to find a single study where the desired bequest is an input in the calculations. 6 similar “bootstrapping” procedures are used, for example, in blanchett (2007); blanchett and frank (2009); spitzer (2008); spitzer, strieter, and singh (2007); and zolt (2013). 7 in the previous section we said that we characterize each sequence of returns by its corresponding pwa, but this is only an intuitive description. the formally correct statement is that we characterize each sequence by its corresponding (rn,sn) pair. to obtain an actual pwa for a specific sequence of returns, the current balance and the bequest target must be provided. 8 for the (rn,sn) pairs that produce a negative ending balance, the numerical value of this formula would be wrong because the yields cease to apply once the balance moves into the red. but the tabulation can show the size distribution of the positive ending balances, with the rest marked as “failures.” 9 eq. (8) is obtained simply by changing the standpoint of the evaluation: from before the current year’s withdrawal is made, to after it has been made but before the current year’s return hits the balance. in other words, we perform the evaluation assuming we are in the middle of the year, instead of at the very beginning. this changes the starting balance from ks to ks-w*, but leaves the sequence of yield accruals unchanged. 10 we thank an anonymous reviewer for comments that led to the development of the tables in figs. 8 and 9. 11 approaching the problem from a different angle, robinson and tahani (2010) also produce non-constant withdrawal profiles. 12 the three return sequences involved in the figures were produced randomly by the monte carlo engine, but they are not equally likely. we selected them intentionally, out of the set of 20,000 runs, to illustrate different possibilities for the resulting withdrawal path. 13 this strategy developed from suggestions by an anonymous reviewer, which we thankfully acknowledge. 355e.d. suarez et al. / financial services review 24 (2015) 331–357 acknowledgment we thank david blanchett, wade pfau, and conference participants at the 2014 annual meeting of the academy of financial services at nashville, for helpful comments to an earlier version of this article. references bengen, w. p. (1994). determining withdrawal rates using historical data. journal of financial planning, 7, 171–180. bengen, w. p. (2006). baking a withdrawal plan ‘layer cake’ for your retirement clients. journal of financial planning, 19, 44–51. bernard, g. a. (2011). systematic withdrawal programs: unsafe at any speed. journal of financial service professionals, 65, 44–61. blanchett, d. m. (2007). dynamic allocation strategies for distribution portfolios: determining the optimal distribution glide path. journal of financial planning, 20, 68–81. blanchett, d. m., & blanchett, b. c. (2008). joint life expectancy and the retirement distribution period. journal of financial planning, 21, 54–60. blanchett, d., finke, m., & pfau, w. d. (2014). asset valuations and safe portfolio withdrawal rates. the retirement management journal, 4, 23–36. blanchett, d. m., & frank, l. r. (2009). a dynamic and adaptive approach to distribution planning and monitoring. journal of financial planning, 22, 52–66. blanchett, d. m., kowara, m., & chen, p. (2012). optimal withdrawal strategy for retirement income portfolios. the retirement management journal, 2, 7–20. cooley, p. l., hubbard, c. m., & walz, d. t. (1998). retirement savings: choosing a withdrawal rate that is sustainable. journal of the american association of individual investors, 20, 16–21. cooley, p. l., hubbard, c. m., & walz, d. t. (1999). sustainable withdrawal rates from your retirement portfolio. financial counseling and planning, 10, 39–47. cooley, p. l., hubbard, c. m., & walz, d. t. (2003). a comparative analysis of retirement portfolio success rates: simulation versus overlapping periods. financial services review, 12, 115–129. cooley, p. l., hubbard, c. m., & walz, d. t. (2011). portfolio success rates: where to draw the line. journal of financial planning, 24, 48–60. frank sr., l. r., & blanchett, d. m. (2010). the dynamic implications of sequence risk on a distribution portfolio. journal of financial planning, 23, 52–61. frank sr., l. r., mitchell, j. b., & blanchett, d. m. (2011). probability of failure-based decision rules to manage sequence risk in retirement. journal of financial planning, 24, 46–55. guyton, j. t. (2004). decision rules and portfolio management for retirees: is the ‘safe’ initial withdrawal rate too safe? journal of financial planning, 17, 54–62. guyton, j. t., & klinger, w. j. (2006). decision rules and maximum initial withdrawal rates. journal of financial planning, 19, 49–57. milevsky, m., & huang, h. (2011). spending retirement on planet vulcan: the impact of longevity risk aversion on optimal withdrawal rates. risk management, 23, 24–38. mitchell, j. b. (2011). retirement withdrawals: preventive reductions and risk management. financial services review, 20, 45–59. pfau, w. d. (2012). capital market expectations, asset allocation and safe withdrawal rates. journal of financial planning, 25, 36–43. pfau, w. d. (2014). new research on how to choose portfolio return assumptions. advisor perspectives, 8, 1–7. pye, g. b. (2000). sustainable investment withdrawals. journal of portfolio management, 26, 73–83. 356 e.d. suarez et al. / financial services review 24 (2015) 331–357 robinson, c. d. (2007). a phased-income approach to retirement withdrawals: a new paradigm for a more affluent retirement. journal of financial planning, 20, 44–56. robinson, c., & tahani, n. (2010). sustainable retirement income for the socialite, the gardener, and the uninsured. financial services review, 19, 187–202. spitzer, j. j. (2008). retirement withdrawals: an analysis of the benefits of periodic ‘midcourse’ adjustments. financial services review, 17, 17–29. spitzer, j. j., strieter, j. c., & singh, s. (2007). guidelines for withdrawal rates and portfolio safety during retirement. journal of financial planning, 20, 52–59. stout, r. g. (2008). stochastic optimization of retirement portfolio asset allocations and withdrawals. financial services review, 17, 1–15. stout, r. g., & mitchell, j. b. (2006). dynamic retirement withdrawal planning. financial services review, 15, 117–131. williams, d., & finke, m. s. (2011). determining optimal withdrawal rates: an economic approach. the retirement management journal, 1, 35–46. zolt, d. m. (2013). achieving a higher safe withdrawal rate with the target percentage adjustment. journal of financial planning, 26, 51–59. 357e.d. suarez et al. / financial services review 24 (2015) 331–357 academy of financial services officers president thomas coe quinnipiac university president-elect robert moreschi virginia military institute executive vice president-program duncan williams western carolina university vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president william chittenden texas state university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama halil kiymaz rollins college frances lawrence louisiana state university david nanigian california state university, fullerton tom potts baylor university janine scott massey university past presidents william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2016 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 25, no. 2, 2016 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2016 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. exploring the demand for retirement planning advice: the role of financial literacy martin c. seay, ph.d., cfpa,*, kyoung tae kim, ph.d.b, stuart j. heckman, ph.d., cfpc aassistant professor of personal financial planning, kansas state university, 318 justin hall, manhattan, ks 66505, usa bassistant professor, department of consumer sciences, university of alabama, 312 adams hall, tuscaloosa, al 35487, usa cassistant professor of personal financial planning, kansas state university, 319 justin hall, manhattan, ks 66505, usa abstract this research extends previous literature on the relationship between financial literacy and financial advice seeking in three ways: (1) we examine financial planner use specifically within the context of retirement planning, (2) we incorporate huston’s (2010) framework of financial literacy, and (3) we use longitudinal data to investigate the initiation, maintenance, and termination of financial planner use. results from the 2010 and 2012 national longitudinal survey of youth 1979 (nlsy79) show a positive association between the components of financial literacy and financial planner use for retirement planning. © 2016 academy of financial services. all rights reserved. jel classification: d14; g20 keywords: retirement planning; financial planner use; financial advice 1. introduction u.s. workers face significant difficulty in adequately planning for retirement. this difficulty is reinforced by the transition from defined benefit (db) to defined contribution (dc) plans, which places more responsibility and risk on individuals for their saving and investing * corresponding author. tel.: �1-785-532-1486; fax: �1-785-532-5505. e-mail address: mseay@ksu.edu (m. c. seay) financial services review 25 (2016) 331–350 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. decisions. planning for retirement requires individuals to make complex financial decisions, with financial literacy becoming critical (lusardi and mitchell, 2007; van rooij, lusardi, and alessie, 2011). the shift to self-funding retirement (i.e., dc plans) has helped spur the growth in demand for financial advice and the financial planning profession. while some research has investigated the relationship between individual financial literacy and general financial advice seeking behavior (calcagno and monticone, 2015; collins, 2012; moulton, loibl, samak, and collins, 2013; robb, babiarz, and woodyard, 2012), little work has focused on advice related to retirement planning. further, there are notable limitations in the measures used in previous research, either because of temporal inconsistencies (e.g., the national financial capability survey has a five-year look back) or lack of focus on retirement planning. consequently, this study uses data from the 2010 and 2012 administrations of the national longitudinal survey of youth 1979 (nlsy79) to investigate the relationship between financial literacy and household demands for retirement planning advice. recent retirement adequacy studies have found that more than half of u.s. households are not adequately prepared for retirement. using data from the 2010 survey of consumer finances (scf), kim and hanna (2015) find only 42% of working households aged 35 to 60 are adequately prepared for retirement, while 46% report that they expect to receive adequate income in retirement. munnell, webb, and golub-sass (2012) note an increase in the proportion of working households who are at risk of being unable to maintain their preretirement standard of living in retirement between 2007 and 2010 from 44% to 53%. this increase is attributed to the combined effect of poor investment returns, lower interest rates, and the increase in social security’s full retirement age. despite positive signs of economic recovery, munnell, hou, and webb (2014) find that 53% of households remain at risk of lowered standards of living in retirement using data from the 2013 scf. a growing body of literature indicates that financial planners provide significant benefits, both economic and psychological, in helping individuals prepare for retirement. two key studies investigating the economic benefit of financial advice are blanchett and kaplan (2013) and grable and chatterjee (2014). blanchett and kaplan (2013) quantify the benefit of retirement planning advice as gamma, a measure of the increased potential retirement income an individual receives from working with an advisor. their work suggests that, through managing investments, taxes, and retirement withdrawals, an individual’s retirement income can be increased by 22.6% by working with an advisor. similarly, grable and chatterjee (2014) introduce zeta, a measure of how a financial advice can limit wealth volatility and loss in times of economic turmoil. they find that individuals who met with a financial advisor experienced significantly less wealth volatility over the great recession. in terms of psychological benefits, individuals who meet with a financial advisor are more likely to establish long-term goals and be confident in their retirement plan (marsden, zick, and mayer, 2011). further, households who receive financial planning advice exhibit greater consistency between risk attitudes and financial behaviors (park and yao, 2015). given the important role that financial literary and financial planners play in retirement planning, the current study extends previous literature in three ways. first, the nlsy79 provides a specific measure of financial planner use for retirement planning. second, previous work has not been able to incorporate huston’s (2010) financial literacy framework 332 m.c. seay et al. / financial services review 25 (2016) 331–350 by simultaneously exploring financial knowledge, financial confidence, and financial capability. previous work has also used summated measures of financial knowledge, which may have limited the ability of researchers to detect the types of knowledge associated with help-seeking activity. lastly, the use of longitudinal data allows us to better explore how financial literacy is related to the initiation, maintenance, and termination of financial planner use. 2. literature review 2.1. defining financial literacy the terms financial knowledge and financial literacy have been used when referring to an individual’s ability to make financial decisions. however, these terms have often been used interchangeably and with inconsistent definitions. given this confusion, huston (2010) has provided a clear definitional and theoretical framework for financial literacy. according to huston (2010), financially literate individuals must not only be knowledgeable, but also have the ability to apply that knowledge to specific circumstances. financial knowledge is defined as a measure of an individual’s objective understanding of financial concepts and is typically assessed by asking individuals a series of factual financial questions. an individual’s knowledge is then rated based on the number or difficulty of questions they are able to answer correctly. a review of literature indicates that, in many cases, the term financial literacy is used to convey what huston (2010) defines as financial knowledge. however, to be financially literate individuals must be able to apply this knowledge. huston (2010) indicates that an individual must have confidence in his or her knowledge and be capable of applying that knowledge to a financial scenario. simply put, without confidence in one’s ability and the innate capability to translate knowledge into action, financial knowledge alone may be insufficient to spur positive financial behavior. this article’s approach is similar to huston (2010) as we seek to clearly define and distinguish between financial knowledge and financial literacy. 2.2. financial literacy and financial behavior the majority of research into financial literacy has focused on financial knowledge. financially knowledgeable households are consistently found to be more likely to exhibit beneficial financial behaviors, while less financially knowledgeable households tend to exhibit more troubling behaviors. financial knowledge is negatively associated with high cost debt borrowing instruments (lusardi and scheresberg, 2013; robb et al., 2015) and positively associated with more responsible credit card practices (allgood and walstad, 2013; xiao et al., 2011) and “best practice”1 financial behavior (robb & woodyard, 2011). financial knowledge is also associated with increased stock ownership (calvert et al., 2007), the use of lower cost mortgages (moore, 2003), and retirement planning behavior (lusardi and mitchell, 2009). additionally, moulton et al., (2013) finds that financially knowledgeable individuals are less likely to underestimate their total household debt. 333m.c. seay et al. / financial services review 25 (2016) 331–350 a more complicated relationship has been found between financial confidence and financial behavior. while financial confidence is positively related to “best practice” financial behaviors (robb and woodyard, 2011) and responsible credit card behavior (allgood and walstad, 2013), it is also positively associated with high cost borrowing behavior (robb et al., 2015). this disparity may be somewhat explained by situations in which consumers’ financial confidence is misaligned with their actual knowledge and ability. allgood and walstad (2013) and robb et al. (2015) both find that individuals that exhibit high financial confidence and low financial knowledge are more likely to exhibit poor financial decisions. similarly, moulton et al. (2013) finds that financially overconfident individuals are more likely to engage in suboptimal mortgage borrowing behaviors. financial capability has most often been proxied through cognitive ability or financial sophistication, a measure that blends financial capability, financial behavior, and financial knowledge (huston, finke, and smith, 2012). individuals with higher levels of cognitive ability are more likely to participate in the stock market (christelis, tullio, and padula, 2010), less likely to overreact to market changes (browning and finke, 2015), exhibit fewer behavioral biases (grinblatt, keloharju, and linnainmaa, 2012), and demonstrate more patience when making financial decisions (benjamin, sebastian, and shapiro, 2013). similarly, financially sophisticated households are more likely to understand and take advantage of roth iras (smith, finke, and huston, 2012), take advantage of mortgage leverage strategies (kim, seay, and smith, 2016), and make more appropriate mortgage decisions (smith, finke, and huston, 2011). given data availability in the nlsy, this research uses a measure of cognitive ability as a proxy for financial capability. 2.3. who seeks financial planning advice? according to a recent project sponsored by the certified financial planner board of standards and the consumer federation of america, close to nine in 10 american households engage in some type of financial planning, ranging from very informal (i.e., mental budgeting) to very formal (i.e., building a comprehensive financial plan with a professional) with most households falling somewhere in between (princeton survey research associates international, 2013). the use of professional financial planners in the united states, although not widespread, does seem to be on the rise. an analysis of the scf shows that that 25% of households reported financial planner use in 2007, up from 21% in 1998 (hanna, 2011). many researchers have explored factors that lead a household to seek professional financial help of some kind. in terms of demographics, wealth and income are the leading indicators followed closely by educational attainment and age (hanna, 2011). people with more financial knowledge (collins, 2012; robb et al., 2012), greater risk tolerance (hanna, 2011; robb et al., 2012), and a sense of self-efficacy (lim, heckman, letkiewicz, and montalto, 2014) are more likely to utilize financial help. cummings and james (2014) find that people seeking help for emotional problems will also seek help for financial matters and that experiencing the death of a spouse increases the likelihood of seeking help. finke, huston, and winchester (2011) find those who pay for financial advice are more likely to be older, wealthier, college educated, and female. 334 m.c. seay et al. / financial services review 25 (2016) 331–350 recent literature has also identified trust as being an important predictor of financial help-seeking. gennaioli, shleifer, and vishny (2015) develop a theoretical model in which consumer decisions to hire professionals to manage (i.e., invest) their money is mediated by trust. recent empirical results reinforce the theoretical conclusion that trust plays an important role in financial help-seeking. lachance and tang (2012) find that, “controlling for financial exposure,2 trust and cost are the two most important determinants of financial advice-seeking behavior” (p. 220). they also find that trust is relatively more important in determining saving and investment advice seeking compared to other types of advice, for example, debt counseling. martin, finke, and gibson (2014) explore the relationship between race, trust, and seeking retirement advice. they find lower levels of trust among black and hispanic households and that trust is positively associated with seeking retirement advice from financial planner. some barriers to seeking professional financial help include low financial risk tolerance (grable and joo, 2001), shame and embarrassment, and lack of knowledge about professional sources (du plessis, lawton, and corney 2010). grable and joo (2001) also find that individuals with low satisfaction with their financial situation are more likely to seek advice from family, friends, and work colleagues, rather than professional sources. 2.4. the link between financial literacy and help-seeking past studies have addressed the relationship between the components of financial literacy and help-seeking behavior with some promising findings. lusardi and mitchell (2007) find that greater knowledge increases one’s awareness of the need for assistance and perry and morris (2005) find that potential costs of poor decisions emboldens individuals to make their own financial decisions. in an analysis of college students, lim et al. (2014) find that college students who took financial education courses in either high school or college are more likely to seek financial help. both collins (2012) and robb et al. (2012) analyze the 2009 national financial capability study (nfcs) dataset and find a positive correlation between financial knowledge, financial confidence, and the use of a financial planner. in an investigation of an italian sample, calcagno and monticone (2015) find that financially knowledgeable individuals are more likely to seek advice, but no relationship is found between financial confidence and help seeking behavior. conversely, in a study of first time homebuyers, moulton et al. (2013) find financial confidence to be positively associated with advice seeking behavior, but found no relationship between financial knowledge and the use of a financial coach. finke et al. (2011) find a more complicated relationship between financial confidence and financial advice. overall, those who pay for financial advice have a low level of self-reported knowledge about financial issues. however, among those who pay, those who choose comprehensive management have high self-reported knowledge about financial issues (finke et al., 2011). while a variety of studies have sought to investigate the link between financial literacy and advice seeking behavior, most research has been limited in its inclusion of all three components of financial literacy and focus on financial planner use. using rich data from the nlsy79, this research is able to better measure each component of financial literacy in 335m.c. seay et al. / financial services review 25 (2016) 331–350 investigating its link to financial planner use while controlling for other known predictors of financial advice seeking. 3. method 3.1. dataset and sample selection the nlsy79 is a nationally representative sample of 12,686 young men and women who were between 14 and 22 years old when they were first surveyed in 1979. these individuals were interviewed annually through 1994 and are currently interviewed on a biennial basis. this dataset is particularly appropriate to address the research question because it is longitudinal, has specific questions on the use of a financial planner as well as questions to measure financial knowledge, financial confidence, and financial capability. of the 7,301 respondents who remained in the survey in 2012, we limit our sample to nonretired individuals that responded to both the 2010 and 2012 administrations of the nlsy79. this provided a final sample size of 5,127. 3.2. dependent variable the dependent variables are constructed based on whether or not the respondent “consulted a financial planner about how to plan [your] finances after retirement” in 2010 and 2012. this study uses two different dependent variables. first, a binary dependent variable indicates whether respondents reported using a financial planner for retirement planning in 2012 for a baseline analysis. further, we define four categories of financial planner use between the two survey waves; those who had a financial planner in both 2010 and 2012; those who did not have a planner in 2010, but adopted one in 2012; those who had a financial planner in 2010, but dropped them in 2012; and those who did not have a planner in either 2010 or 2012. 3.3. financial literacy variables 3.3.1. financial knowledge objective financial knowledge is measured using five personal finance questions. the financial knowledge questions, administered in the nlsy79 in 2012, asses an individual’s understanding of diversification, compound interest, inflation, bond pricing, and mortgages. more important, a “don’t know” response option is included to limit the occurrence of random guessing on each question. researchers have used these items individually (lusardi and scheresberg, 2013; seay et al., 2015), to create a summative scale (collins, 2012; robb and woodyard, 2011; robb et al., 2012), and to differentiate individuals with high and low objective knowledge (allgood and walstad, 2013; robb et al., 2015). a careful analysis of the questions leads us to conclude that each question is measuring a different aspect of financial knowledge and should not be used in a manner that counts them as one measure. 336 m.c. seay et al. / financial services review 25 (2016) 331–350 using factor analysis, we find the individuals items have low reliability (� � 0.37), supporting the notion that these questions should be used as separate measures. 3.3.2. financial confidence three different measures are used to measure confidence: subjective financial knowledge, confidence in ability to manage day-to-day financial matters, and rotter locus of control. subjective financial knowledge is measured based on a question asking respondents to rate their overall financial knowledge on a scale from 1 to 7. similarly, individuals are asked to identify, on a scale from 1 to 7, how much they agreed with the statement “i am good at dealing with day-to-day financial matters, such as checking accounts, credit and debit cards, and tracking expenses.” for both of these questions, which are measured in 2012, higher scores are associated with increased confidence levels in financial knowledge and ability to manage finances. lastly, the rotter locus of control scale (rotter, 1966) is used to measure the extent to which an individual believes they are in control of their lives. scores range from 4 to 16 and have been coded such that higher scores signify a high internal locus of control, indicative of higher self-determination in accomplishing tasks. individuals with a high internal locus of control may believe in their ability to change their situation and make them more confident to seek information that will help them in their situation (rotter, 1990). 3.3.3. financial capability an individual’s capability to apply knowledge is proxied using the armed forces qualification test (afqt). the afqt is commonly used as a general measure of individual’s cognitive ability. originally assessed in 1980, raw scores were converted to percentile scores and normed in 2006 to reflect updated standards. 3.4. control variables in addition to financial literacy variables, control variables include age, race (white, black, or hispanic), gender (male/female), married (yes/no), education (less than high school, high school education, some college, or college degree), urban area (yes/no), employment status (unemployed, employed, unable to work, or work/other), health insurance (yes/no), chronic health issue in household (yes/no), log of income, log of net worth, log of retirement account balance, participation in a defined benefit retirement plan, stock ownership (yes/no), home ownership (yes/no), risk tolerance, and trust. risk tolerance is measured on scale from one to 10, with higher scores being associated with an increased willingness to take risks in financial matters. trust is measured on a scale from 1 to 5, with higher scores indicating that an individual is more trusting of other people. a full table of measures can be found in the appendix. 3.5. research hypothesis based on previous research indicating that seeking financial advice is a complement for financial literacy (collins, 2012; robb et al., 2012), three research hypotheses are proposed: 337m.c. seay et al. / financial services review 25 (2016) 331–350 hypothesis 1: the components of financial literacy are positively associated with the use of a financial planner. hypothesis 2: the components of financial literacy are positively associated with adopting a financial planner when compared to those who never had a financial planner. hypothesis 3: the components of financial literacy are negatively associated with dropping a financial planner when compared to those who had a financial planner throughout. 3.6. empirical specification two regression models are employed to test these hypotheses. to test hypothesis one, a binomial logistic regression is conducted to establish a baseline relationship between the financial literacy components and the use of a financial planner. given that financial knowledge is measured in 2012, the dependent variable for this analysis is financial planner use in 2012. logit� p� � log� p 1 � p� � �0 � x1�1 � x2�2 � � � � � xk�k � x� where p � probability of using a financial planner in 2012 x � a vector of a household’s financial literacy variables and characteristics � � a vector of coefficients to be estimated to investigate hypotheses two and three, a multinomial logit regression is utilized to compare four groups based on financial planner use across two time periods: (1) those who had a financial planner in both 2010 and 2012 (throughout); (2) those who did not have a planner in 2010, but adopted one in 2012 (adopted); (3) those who had a financial planner in 2010, but dropped them in 2012 (dropped); and (4) those who did not have a planner in either period (never). we are interested in two specific comparisons. the first is the difference between those that adopted a planner in 2012 (adopted) and those who did not have a planner in either period (never). we hypothesize those who decide to adopt a planner to be more financially literate. the second comparison is between those who dropped a planner 2012 (dropped) and those who had a planner throughout (throughout). we hypothesize those who dropped a planner in 2012 to have lower financial literacy than those who have a planner throughout. the multinomial logit is specified as follows. the probability that the ith household would choose the jth group is described by: pij � pr�rij � rik�, for k � j, j � 0, 1, 2, 3 with rij is the maximum utility attainable for household i if the household holds jth group, and, rij � x�ij �ij � �ij 338 m.c. seay et al. / financial services review 25 (2016) 331–350 where �ij is a vector of coefficients of each of the independent variables. assuming that the stochastic term, �ij, is distributed identically and independently across alternatives, the multinomial logit model is expressed by: pij � exp�x�ij�ij�/��x�ij�ij� the nlsy79 provides weighting information that researchers can use to make the sample representative of the larger u.s. population. consequently, normalized sampling weights from 2012 are used in all analyses, providing more representative and generalizable results (deaton 1997). unfortunately, complex sampling design information is not included in the publically available version of the nlsy79. 4. results 4.1. descriptive results table 1 provides descriptive statistics for the sample, as well as for each of the four different groups of financial planner use. respondent ages range from 47 to 56, an ideal age group in which to investigate retirement planning decisions. the majority of the sample is white (81.5%), male (50.2%), married (67.6%), employed (80.4%), and homeowners (74.6%). overall, respondents are financially knowledgeable, have high levels of financial confidence, and have relatively internal locus of controls. when comparing financial literacy between groups, reported levels of financial knowledge, confidence and capability are highest for those who had a financial planner in both 2010 and 2012 and lowest for those who did not have a planner in either period. 4.2. baseline model: binomial logit analysis results from the binomial logistic regression predicting use of a planner in 2012 are presented in table 2. variance inflation factors were checked to test for any potential multicollinearity issues, but were found to be within the acceptable range (less than 2.5). this baseline analysis provides evidence of the link between financial literacy and seeking retirement planning advice. an understanding of diversification (knowledge), an understanding of mortgages (knowledge), having higher subjective knowledge (confidence), and having a more internal locus of control (confidence) are all associated with planner use. more specifically, correctly answering the diversification and mortgage questions increases the odds that an individual received retirement advice from a financial planner by 45.5% and 42.8%, respectively. similarly, unit increases in subjective knowledge and locus of control increases the odds of financial planner use by 6.4% and 5.3%, respectively. however, no statistically significant relationship is found between cognitive ability (capability) and advice seeking. results also indicate that the likelihood of using a financial planner for retirement purposes is positively correlated with education, health insurance coverage, net worth, retirement assets, stock ownership, homeownership, risk tolerance, and trust. by contrast, income, having a chronic health issue in the household, being male, and living in an urban area are negatively related to the likelihood of using a financial planner. 339m.c. seay et al. / financial services review 25 (2016) 331–350 4.3. multinomial logit analyses results from the multinomial logit most relevant to our hypotheses are presented in tables 3 and 4. table 3 presents the comparison between those who never had a planner and those who adopted a planner in 2012, as this isolates the decision to adopt a planner in 2012. in terms of financial literacy, individuals who are more knowledgeable about diversification and have higher subjective knowledge are more likely to adopt a planner for retirement planning table 1 descriptive statistics of selected variables by changes in financial planner use variable all sample n � 5,127 planner in 2010 and 2012 n � 657 adopted a planner in 2012 n � 369 dropped a planner in 2012 n � 500 no planner n � 3,601 financial literacy measuresa k: diversification 0.68 0.84 0.77 0.75 0.63 k: compound interest 0.74 0.85 0.80 0.78 0.71 k: inflation 0.82 0.87 0.86 0.87 0.80 k: bonds 0.31 0.42 0.37 0.36 0.27 k: mortgage 0.87 0.96 0.92 0.94 0.84 c: subjective knowledge 4.90 5.19 5.21 5.11 4.77 c: day-to-day finances 5.73 6.16 5.83 5.91 5.59 c: rotter locus of control 11.5 12.28 11.62 11.71 11.27 a: afqt (intelligence) 52.76 68.14 60.13 58.18 47.58 control variables mean age 51.4 51.6 51.4 51.6 51.4 white 81.5% 90.3% 84.1% 82.2% 79.1% black 12.6% 6.3% 10.1% 12.4% 14.4% hispanic 5.9% 3.4% 5.8% 5.4% 6.5% male 50.2% 45.7% 56.1% 50.7% 50.5% female 49.8% 54.3% 43.9% 49.3% 49.5% married 67.6% 78.9% 74.1% 71.8% 63.6% less than high school 5.6% 0.4% 2.6% 2.3% 7.7% high school education 39.3% 21.4% 30.5% 33.6% 45.3% some college 24.4% 19.7% 27.5% 26.1% 24.9% college degree 30.7% 58.6% 39.4% 38.0% 22.2% urban 74.4% 75.4% 71.0% 77.8% 74.0% unemployed 16.7% 10.4% 12.5% 9.7% 19.6% employed 80.4% 87.3% 85.3% 87.7% 77.2% unable to work 0.8% 0.7% 0.5% 0.7% 0.9% work/other 2.1% 1.6% 1.7% 1.9% 2.3% has health insurance 86.8% 97.5% 92.5% 91.9% 82.9% chronic health issue in household 10.3% 5.0% 7.7% 7.6% 12.3% mean income $398,534 $83,837 $65,814 $69,968 $42,570 mean net worth $53,425 $893,340 $604,925 $514,538 $244,280 mean retirement account $26,827 $80,237 $37,013 $32,274 $12,608 defined benefit plan ownership 17.8% 22.4% 22.6% 20.3% 15.8% stock ownership 16.3% 33.2% 21.8% 19.3% 11.4% homeowners 74.6% 90.8% 85.2% 83.6% 68.3% mean score of risk tolerance 3.7 4.4 4.1 4.1 3.4 trust 2.2 2.5 2.3 2.2 2.1 source: restricted sample of the nsly79, 2010 and 2012 waves. percentages are weighted proportions. ak � knowledge; c � confidence; a � capability. 340 m.c. seay et al. / financial services review 25 (2016) 331–350 advice than otherwise similar households. in particular, correctly answering the diversification question increases the odds of adopting a planner by 33.4%, while a one unit increase in subjective knowledge increases the odds of adopting a planner by 16.5%. adopting a planner is also found to be positively associated with homeownership, net worth, retirement assets, and risk tolerance. table 4 presents the comparison between those had a planner in each time period and those who dropped a planner in 2012. this comparison is important as it isolates the decision table 2 baseline model: binomial logistic regression of financial planner use, 2012 nlsy79 variable coeff. se odds ratio financial literacy measuresa k: diversification 0.3747*** 0.0912 1.455 k: compound interest 0.1219 0.0953 1.130 k: inflation �0.0438 0.1051 0.957 k: bonds 0.0550 0.0788 1.057 k: mortgage 0.3563* 0.1514 1.428 c: subjective knowledge 0.0622* 0.0312 1.064 c: day-to-day finances �0.0049 0.0255 0.995 c: rotter locus of control 0.0519** 0.0163 1.053 a: afqt (intelligence) 0.00137 0.0019 1.001 control variables age 0.0092 0.0159 1.009 male (ref.: female) �0.2147** 0.0777 0.807 married (ref.: unmarried 0.1160 0.0877 1.123 racial/ethnicity (ref.: white) black 0.0974 0.1400 1.102 hispanic 0.1363 0.1743 1.146 education (ref.: less than high school) high school education 0.6382* 0.2946 1.893 some college 0.9136** 0.3008 2.493 college degree 1.2329*** 0.3064 3.431 employment status (ref.: employed) unemployed �0.2642 0.1624 0.768 unable to work 0.3742 0.4451 1.454 work/other �0.2002 0.2864 0.819 urban (ref.: no) �0.1848* 0.0848 0.831 has health insurance (ref.: no) 0.4955** 0.1620 1.641 chronic health issue in household (ref.: no) �0.2875* 0.1451 0.750 income (ln) �0.0217* 0.0101 0.979 net worth (ln) 0.0447*** 0.0090 1.046 retirement assets (ln) 0.0707*** 0.0075 1.073 defined benefit pension ownership 0.0777 0.0895 1.081 stock ownership 0.3253*** 0.0897 1.384 homeowners 0.2734* 0.1159 1.314 risk tolerance 0.0780*** 0.0161 1.081 trust 0.1022* 0.0444 1.108 intercept �5.8555 0.8912 concordance (mean) 77.1% source: restricted sample of the nsly79, 2012 wave. ak � knowledge; c � confidence; a � capability. *p � .05, **p � .01, ***p � .001. 341m.c. seay et al. / financial services review 25 (2016) 331–350 to drop a planner in 2012. dropping a planner is negatively associated with an understanding of diversification (knowledge) and having an internal locus of control (confidence). specifically, correctly answering the diversification question decreases the odds of dropping a planner by 26.0%, while a one unit increase in the locus of control decreases the odds of dropping a planner by 6.3%. dropping a planner is also negatively associated with health insurance coverage, education, net worth, retirement assets, stock ownership, and trust but positively associated with income. table 3 multinomial logistic regression of financial planner use (reference category: no planner) adopted a planner in 2012 coeff. se odds ratio k: diversificationa 0.2882* 0.1335 1.334 k: compound interest 0.0587 0.1391 1.060 k: inflation 0.0409 0.1579 1.042 k: bonds 0.0823 0.1185 1.086 k: mortgage 0.2970 0.2106 1.346 c: subjective knowledge 0.1529*** 0.0461 1.165 c: day-to-day finances �0.0621 0.0353 0.940 c: rotter locus of control 0.00127 0.0241 1.001 a: afqt (intelligence) 0.00354 0.00289 1.004 control variables age 0.0005 0.0237 1.001 male (ref.: female) 0.0481 0.1167 1.049 married (ref.: unmarried 0.0666 0.1299 1.069 racial/ethnicity (ref.: white) black 0.2244 0.1977 1.252 hispanic 0.3217 0.2379 1.380 education (ref.: less than high school) high school education 0.2403 0.3387 1.272 some college 0.5929 0.3505 1.809 college degree 0.5961 0.3656 1.815 employment status (ref.: employed) unemployed 0.0285 0.2310 1.029 unable to work �0.1170 0.7842 0.890 work/other 0.0251 0.4199 1.025 urban (ref.: no) �0.2493* 0.1242 0.779 has health insurance (ref.: no) 0.2358 0.2095 1.266 chronic health issue in household (ref.: no) �0.1334 0.2016 0.875 income (ln) 0.0066 0.0154 1.007 net worth (ln) 0.0376** 0.0124 1.038 retirement assets (ln) 0.0608*** 0.0117 1.063 defined benefit pension ownership 0.1150 0.1344 1.122 stock ownership 0.1593 0.1424 1.173 homeowners 0.3313* 0.1674 1.393 risk tolerance 0.0628** 0.0235 1.065 trust 0.0127 0.0647 1.013 intercept �5.1226 1.2942 source: restricted sample of the nsly79, 2010 and 2012 waves. reference category is no planner in 2010 and 2012. a k � knowledge; c � confidence; a � capability. *p � .05, **p � .01, ***p � .001. 342 m.c. seay et al. / financial services review 25 (2016) 331–350 5. discussion the purpose of this article is to expand the body of knowledge related to the relationship between financial literacy and a financial planner use for retirement planning advice. this is accomplished by incorporating huston’s (2010) framework for financial literacy, using a retirement specific measure of financial planner, and using longitudinal data that allows exploration of the initiation, maintenance, and termination of financial planner use. table 4 multinomial logistic regression of financial planner use (reference category: planner in 2010 and 2012) dropped a planner in 2012 coeff. se odds ratio k: diversificationa �0.3007* 0.1487 0.740 k: compound interest �0.1235 0.1549 0.884 k: inflation 0.2542 0.1747 1.289 k: bonds 0.0272 0.1246 1.028 k: mortgage 0.0299 0.2657 1.030 c: subjective knowledge 0.0693 0.0507 1.072 c: day-to-day finances �0.0468 0.0422 0.954 c: rotter locus of control �0.0655* 0.0261 0.937 a: afqt (intelligence) 0.0011 0.0031 1.001 control variables age 0.0102 0.0254 1.010 male (ref.: female) 0.2006 0.1242 1.222 married (ref.: unmarried �0.0961 0.1401 0.908 racial/ethnicity (ref.: white) black 0.1771 0.2230 1.194 hispanic 0.0999 0.2863 1.105 education (ref.: less than high school) high school education �1.0003 0.6584 0.368 some college �1.0990* 0.6653 0.333 college degree �1.5323 0.6712 0.216 employment status (ref.: employed) unemployed 0.2152 0.2725 1.240 unable to work �0.5110 0.6859 0.600 work/other 0.3456 0.4579 1.413 urban (ref.: no) 0.2268 0.1398 1.255 has health insurance (ref.: no) �0.5687* 0.2884 0.566 chronic health issue in household (ref.: no) 0.2553 0.2390 1.291 income (ln) 0.0453** 0.0167 1.046 net worth (ln) �0.0468** 0.0145 0.954 retirement assets (ln) �0.0370** 0.0119 0.964 defined benefit pension ownership �0.0851 0.1430 0.918 stock ownership �0.3316* 0.1434 0.718 homeowners 0.0961 0.1927 1.101 risk tolerance �0.0280 0.0258 0.972 trust �0.2285** 0.0709 0.796 intercept 2.3093 1.5134 source: restricted sample of the nsly79, 2010 and 2012 waves. reference category is planner in 2010 and 2012. a k � knowledge; c � confidence; a � capability. *p � .05, **p � .01, ***p � .001. 343m.c. seay et al. / financial services review 25 (2016) 331–350 evidence is found to support hypothesis one, as elements of financial knowledge and financial confidence are associated with seeking retirement planning advice from a financial planner. this analysis is conceptually similar to collins (2012) and robb et al. (2012), and builds upon their work by using a measure of receiving retirement planning advice in the current year and by controlling for trust, a variable that was unavailable in the data on which their analyses were based. our results indicate a more nuanced relationship between financial knowledge and advice seeking behavior than previously understood. collins (2012), calcagno and monticone (2015), and robb et al. (2012) each use composite measures of financial knowledge, which does not allow exploration of the specific elements of financial knowledge that contribute to advice seeking behavior. results of this study indicate that an understanding of higher level concepts (i.e., diversification and mortgages) are key contributors to advice seeking behavior, while no relationship is found for understanding of compound interest, inflation, and bonds. the positive relationship between subjective financial knowledge and seeking advice is similar to previous results in collins (2012), calcagno and monticone (2015), and robb et al., (2012). the relationship between confidence and behavior is reinforced, as individuals with a more internal locus of control are found to be more likely to seek advice from a financial planner. no relationship is found between cognitive ability (capability) and financial planner use. this is surprising, but may be because of the use of a general measure of capability as opposed to one specifically related to finances. supporting evidence is also found for hypotheses two and three. among those who did not have a planner in 2010, individuals who are more knowledgeable about diversification (knowledge) and had higher subjective knowledge (confidence) are more likely to adopt a financial planner for retirement planning advice. similarly, among those who had a planner in 2010, discontinuing planner use in 2012 is negatively associated with an understanding of diversification (knowledge) and an internal locus of control (confidence). these results reinforce the importance of higher level financial knowledge in the decision to seek retirement planning advice, as well as highlighting that different components of financial knowledge may be more or less important in different behaviors. evidence is also provided related to the importance of financial confidence, although depending on the analysis the specific measure of confidence that impacted behavior differed. once again, no relationship is found between cognitive ability (i.e., our proxy for capability) and advice seeking. while the current analysis provides more information about the relationship between financial literacy and financial planner use within the context of retirement planning than in previous literature, care should still be taken in interpreting the current results. data availability limited the measurement of financial knowledge to 2012, while ideally knowledge in 2010 would be used to predict behavior in 2012. this measurement issue severely limits the ability to determine the causal relationship between literacy and planner use is limited. further, given that financial planners often explain financial concepts to clients (i.e., they educate their clients), there may be reverse causality in our model as seeking financial help may improve financial capability. this issue can be addressed upon the release of future waves of the nlsy79. lastly, there are limitations to the measure of financial planner use itself. the term financial planner is not clearly defined in the survey and, consequently, respondents may consider a variety of different individuals (e.g., financial advisor, stockbroker, agents, etc.) to be financial planners. similarly, the financial planner question does 344 m.c. seay et al. / financial services review 25 (2016) 331–350 not clearly indicate a time boundary, which may lead to some inconsistency in the temporal proximity of the planner visit to the question response. 6. conclusions this study reinforces the important role financial advice plays as a compliment to financial knowledge; higher (lower) levels of financial knowledge are associated with initiating and maintaining (dropping) use of a financial planner for retirement planning. results point specifically to the importance of diversification knowledge as a predictor of financial planner use. historically, financial planning services have emphasized investment management and return on investment, only recently expanding value propositions to include multiple aspects of an individual’s financial life (kitces, 2015). as the profession evolves, it will be interesting to see if the relevance of other areas of financial knowledge become more or less important relative to diversification knowledge. however, current clients that are better equipped to understand the value of investment advice are more likely to adopt and use a financial planner, while also being less likely to stop using a financial planner. this suggests that planners should continue to educate clients on the value of diversification and asset allocation. this article also highlights the importance of incorporating huston’s (2010) framework for financial literacy in future research. the inclusion of the three elements of financial literacy provides a better conceptual understanding of one’s ability to evaluate financial scenarios and implement financial planning decisions. similarly, results highlight the importance of carefully evaluating the use of scales to measure financial knowledge. the most prominent studies investigating financial help-seeking behavior have employed a summated scale (collins, 2012; calcagno and monticone, 2015; robb et al., 2012). by including items individually, this research was able to identify the aspects of financial knowledge that were most critical to seeking retirement planning advice from a financial planner. notably, the summated scale used in previous literature was found to have extremely poor reliability (� � 0.37) within the sample of interest. given this result, researchers should be cautious in constructing measures of financial knowledge and be more inclusive of the other components of financial literacy to permit a more complete understanding of phenomenon. omitting the capability and confidence aspects of financial literacy may lead to invalid conclusions. notes 1 robb and woodyard identify best practice financial behaviors as having an emergency fund, obtaining a personal credit report, not overdrafting checking accounts, paying off credit cards in full, having a retirement plan, and owning appropriate insurance. 2 lachance and tang distinguished between five areas of financial advice: saving or investments, tax planning, insurance, mortgage or loan, and debt counseling. their use of the term “financial exposure” is meant to capture how the relevance of each type of advice varies among consumers based on their financial position. for example, debt counseling is most relevant to someone who has debt. 345m.c. seay et al. / financial services review 25 (2016) 331–350 appendix variable coding and descriptive statistics variable name description year collected dependent variables use of financial planner � 1 if respondent answered yes to “consulted a financial planner about how to plan [your] finances after retirement.” 2010, 2012 controls age age of respondent at interview date. continuous variable ranging from 40 to 56. 2012 gender male � 1 if respondent’s reported sex was male. 1979 marital status married � 1 if respondent reported being married. 2012 race/ethnicity white � 1 if respondent’s reported race/ethnicity was white only. 2012 black � 1 if respondent’s reported race/ethnicity was black only. 2012 hispanic � 1 if respondent’s reported race/ethnicity was hispanic. 2012 education less than high school � 1 if highest education level of respondent was less than a high school diploma. 2012 high school � 1 if highest education level of respondent was a high school diploma or equivalent. 2012 some college � 1 if highest education level of respondent was less than four years of college. 2012 college degree � 1 if highest education level of respondent was four years of college or more. 2012 employment status unemployed � 1 if respondent reported being temporarily laid off or unemployed and looking for work. 2012 employed � 1 if respondent reported working now. 2012 unable to work � 1 if respondent reported being disabled and unable to look for work. 2012 work/other � 1 if respondent reported being retired, a homemaker, or other. 2012 other control variables urban area � 1 if respondent reported that residence was located in an urban area. 2012 health insurance � 1 if respondent reported being covered by health insurance/health plan. 2012 chronic health issue in household � 1 if respondent reported that at least one member of the household was disabled or chronically ill. 2012 346 m.c. seay et al. / financial services review 25 (2016) 331–350 appendix (continued) variable coding and descriptive statistics variable name description year collected family income (log) log of total family income. 2012 family net worth (log) log of total family net worth. 2012 retirement assets (log) log of total family retirement assets. 2012 defined benefit pension plan participation � 1 if respondent reported that benefits from any pension/retirement plans were based on a formula. 2012 stock ownership � 1 if respondent reported self or spouse/ partner owning any shares of stock. 2012 home ownership � 1 if respondent reported that residence was owned or being bought by self or spouse/partner. 2012 risk tolerance measured as a continuous variable, “rate yourself from 0 to 10, where 0 means ‘unwilling to take any risks’ and 10 means ‘fully prepared to take risks.’” 2012 trust measured as a continuous variable on a scale of 1 to 5, “generally speaking, how often can you trust other people.” 2008 key predictors financial knowledge diversification � 1 if respondent correctly answered the question, “buying a single company stock usually provides a safer return than a stock mutual fund.” 2012 compound interest � 1 if respondent correctly answered the question, “suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow: more than $102, exactly $102, or less than $102?” 2012 inflation � 1 if respondent correctly answered the question, “imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, would you be able to buy more than, exactly the same as, or less than today with the money in this account?” 2012 bonds � 1 if respondent correctly answered the question, “if interest rates rise, what will typically happen to bond prices? they will rise, they will fall, they will stay the same, there is no relationship between bond prices and the interest rate”; � 0 if answered incorrectly, answered “don’t know,” or refused to answer. 2012 347m.c. seay et al. / financial services review 25 (2016) 331–350 references allgood, s., & walstad, w. (2013). financial literacy and credit card behaviors: a cross-sectional analysis by age. numeracy, 6, 1–26. benjamin, d. j., sebastian, a. b., & shapiro, j. m. (2013). who is ‘behavioral’? cognitive ability and anomalous preferences. journal of the european economic association, 11, 1231–1255. blanchett, d., & kaplan, p. (2013). alpha, beta, and now . . . gamma. the journal of retirement, 1, 29–45. browning, c., & finke, m. s. (2015). cognitive ability and the stock reallocations of retirees during the great recession. journal of consumer affairs, 49, 356–375. doi:10.1111/joca.12065. calcagno, r., & monticone, c. (2015). financial literacy and the demand for financial advice. journal of banking & finance, 50, 363–380. doi:10.1016/j.jbankfin.2014.03.013. calvert, l., campbell, j., & sodini, p. (2007). down or out: assessing the welfare costs of household investment mistakes. journal of political economy, 115, 707–747. christelis, d., tullio, j., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54, 18–38. appendix (continued) variable coding and descriptive statistics variable name description year collected mortgage � 1 if respondent correctly answered the question, “do you think that the following statement is true or false? a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less.” 2012 financial confidence subjective knowledge measured as a continuous variable from the question “on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall financial knowledge?” 2012 day-to-day finances measured as a continuous variable from the question “how strongly do you agree or disagree with the following statements? please give your answer on a scale of 1 to 7, where 1 means “strongly disagree” 7 means “strongly agree,” and 4 means “neither agree nor disagree.” “i am good at dealing with day-to-day financial matters, such as checking accounts, credit and debit cards, and tracking expenses.” 2012 rotter locus of control measured as a continuous variable with scores ranging from 4 to 16. 1979 financial capability armed services vocational aptitude battery percentile score created by the nls, measured as a continuous variable ranging from 0% to 100%. assessed in 1980 and converted to percentile scores in 2006 to reflect updated standards 348 m.c. seay et al. / financial services review 25 (2016) 331–350 collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. cummings, b. f., & james, r. n. (2014). factors associated with getting and dropping financial advisors among older adults: evidence from longitudinal data. journal of financial counseling and planning, 25, 129–147. deaton, a. (1997). the analysis of household surveys: a microeconometric approach to development policy. baltimore, md: johns hopkins university press. du plessis, k., lawton, j., & corney, t. (2010). unemployed and struggling: the lack of financial help seeking behavior exhibited by male construction workers. the australian journal of financial planning, 6, 1–8. finke, m. s., huston, s. j., & winchester, d. (2011). financial advice: who pays? journal of financial counseling and planning, 22, 18–26. gennaioli, n., shleifer, a., & vishny, r. (2015). money doctors. the journal of finance, 70, 91–114. doi:10.1111/jofi.12188. grable, j. e., & chatterjee, s. (2014). reducing wealth volatility: the value of financial advice as measured by zeta. journal of financial planning, 27, 45–51. grable, j. e., & joo, s.-h. (2001). a further examination of financial help-seeking behavior. journal of financial counseling and planning, 12, 55–74. grinblatt, m., keloharju, m., & linnainmaa, j. t. (2012). iq, trading behavior, and performance. journal of financial economics, 104, 339–362. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. huston, s. j., finke, m. s., & smith, h. l. (2012). a financial sophistication proxy for the survey of consumer finances. applied economics letters, 19, 1275–1278. doi:10.1080/13504851.2011.619485. kim, k., & hanna, s. d. (2015). do u.s. households perceive their retirement preparedness realistically? financial services review, 24, 139–155. kim, k., seay, m. c., & smith, h. (2016). after the great recession: financial sophistication and housing leverage. applied economics letters, 23, 1285–1288. doi:10.1080/13504851.2016.1150944. kitces, m. (2015). 6 key value propositions a good financial planner can provide for clients seeking a better “return on life” (available at: https://www.kitces.com/blog/6-key-value-propositions-a-good-financialplanner-can-provide-for-clients-seeking-a-better-return-on-life/). lachance, m.-e., & tang, n. (2012). financial advice and trust. financial services review, 21, 209–226. lim, h., heckman, s. j., letkiewicz, j., & montalto, c. p. (2014). financial stress, self-efficacy, and financial help-seeking behavior of college students. journal of financial counseling and planning, 25, 148–160. lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the role of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. s. (2009). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness (no. w15350). national bureau of economic research (available at: http://www.nber.org/papers/w15350). lusardi, a., & scheresberg, c. b. (2013). financial literacy and high-cost borrowing in the united states (no. w18969). national bureau of economic research (available at: http://www.nber.org/papers/w18969). marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32, 625–643. martin jr., t. k., finke, m., & gibson, p. (2014). race, trust, and retirement decisions. journal of personal finance, 13, 62–71. moore, d. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences (tech. rep. no. 03–39). pullman, wa: washington state university, social and economic sciences research center (available at: http://www.sesrc.wsu.edu/sesrcsite/papers/files/dfi-techreportfinal2-16-04.pdf). moulton, s., loibl, c., samak, a., & collins, m. j. (2013). borrowing capacity and financial decisions of low-tomoderate income first-time homebuyers. journal of consumer affairs, 47, 375–403. doi:10.1111/joca.12021. munnell, a. h., hou, w., & webb, a. (2014). nrri update shows half still falling short. center for retirement research at boston college issue in brief 2014–20. 349m.c. seay et al. / financial services review 25 (2016) 331–350 munnell, a. h., webb, a., & golub-sass, f. (2012). the national retirement risk index: an update. center for retirement research at boston college issue in brief 2012–20. park, e., & yao, r. (2015). financial risk attitude and behavior: do planners help increase consistency? journal of family and economic issues, 37, 624–638. doi:10.1007/s10834-015-9469-9. perry, v. g., & morris, m. d. (2005). who is in control? the role of self-perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39, 299–313. princeton survey research associates international. (2013). financial planning profiles of american huseholds: the 2013 household financial planning survey and index (available at: http://www.cfp.net/docs/publicpolicy/2013-fin-planning-profiles-of-amer-households.pdf). robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. robb, c., babiarz, p., woodyard, a., & seay, m. c. (2015). bounded rationality and the use of alternative financial services. journal of consumer affairs, 49, 407–435. doi:10.1111/joca.12071. robb, c. a., & woodyard, a. s. (2011). financial knowledge and “best practice” behavior. journal of financial counseling and planning, 22, 36–46. rotter, j. b. (1966). generalized expectancies for internal versus external control of reinforcement. psychological monographs: general and applied, 80, 1–28. rotter j.b. (1990). internal versus external control of reinforcement: a case history of a variable. american psychologist, 45, 490–493. doi:10.1037/0003-066x.45.4.489. seay, m. c., asebedo, s., thompson, c., stueve, c., & russi, r. (2015). mortgage holding and financial satisfaction in retirement. journal of financial counseling and planning, 26, 200–216. smith, h., finke, m. s., & huston, s. j. (2011). the impact of financial sophistication on adjustable rate mortgage ownership. journal of financial counseling and planning, 22, 3–15. smith, h. l., finke, m. s., & huston, s. (2012). financial sophistication and housing leverage among older households. journal of family and economic issues, 33, 315–327. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and retirement planning in the netherlands. journal of economic psychology, 32, 593–608. xiao, j., tang, c., serido, j., & shim, s. (2011). antecedents and consequences of risky credit behavior among college students: application and extension of the theory of planned behavior. journal of public policy & marketing, 30, 239–245. 350 m.c. seay et al. / financial services review 25 (2016) 331–350 procedure to determine the optimal roth ira versus deductible ira allocation robert m. hulla,*, john b. hullb aclarence e. king endowed chair of finance, school of business, washburn university, 1621 oxford road, lawrence, ks 66044, usa binternal wholesaler, american century investment services, inc., 4511 w. 70th street, prairie village, ks 66208, usa abstract we offer a procedure to guide the roth ira versus deductible ira (rvd) allocation decision. we require users to input 10 values to generate outputs that include contribution and withdrawal tax rates; maximum gain; and, optimal amount to allocate between the two major ira types. by being at their optimal rvd allocation, we find that a modest earning couple can achieve a lifetime wealth gain amounting to about $180,000 in today’s dollars. we provide figures and tables illustrating rvd outcomes when there are changes in key variables such as salary match, adjusted gross income, portfolio returns, and withdrawal years. © 2016 academy of financial services. all rights reserved. jel classification: a20; c00; d14; g11; h21; k34 keywords: retirement planning; ira; tax rates 1. introduction in this article, we address broad goals in personal financial planning of promoting financial literacy and allocating lifetime income. alhenawi and elkhal (2013) indicate the need for public policies to encourage financial literacy and education, while collins, lam, and stampfli (2015) note the sustainability of adequate lifetime income is a critical portfolio objective. promotion of financial literacy and optimal lifetime investment cannot be realized unless educators and financial advisors have procedures to solve the most crucial financial * corresponding author. tel.: �1-785-670-1600; fax: �1-785-393-5630. e-mail address: rob.hull@washburn.edu (r.m. hull) financial services review 25 (2016) 373–414 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. decisions. in this article, we offer a procedure that provides a definitive solution to a critical financial decision involving the allocation of retirement contributions between the two major ira types: a deductible ira and a roth ira. a deductible ira is synonymous to a traditional ira. we develop the first systematic procedure that renders a defined solution to the quest to discover an investor’s optimal roth ira versus deductible ira (rvd) allocation. we present new formulas and algorithms applicable to income levels for all marginal tax rates. to illustrate this applicability, we perform a detailed rvd illustration for a couple with an adjusted gross income (agi) of $110,000 growing at 3% until retirement. the outcome of our illustration is an optimal range of percentages that should be allocated between a deductible ira and a roth ira. by knowing this range, we demonstrate that our couple can achieve a maximum marginal gain of $28,496 a year during a 20-year retirement period. in today’s dollars, the total lifetime gain in retirement wealth is around $180,000. however, there are variables where modifications in their values can cause the maximum gain to change while also shifting the optimal percentage allocated between a deductible ira and a roth ira. for example, we show that greater employer matching funds will decrease the gain and increase the optimal percentage allocated to a roth ira. we also provide results from scenario analysis illustrating how the optimal rvd allocation changes when modification are made to agi, nominal rate of stock return, and number of withdrawal years. our results are consistent with the research (horan and zaman, 2009; shynkevich, 2013) that suggests investors should contribute more to a roth ira when larger retirement withdrawals are expected. in particular, our scenario tests illustrate that more should be put in a roth ira for situations where investors have greater matching funds, more aggressive investment strategies leading to greater returns, and shorter retirement periods. we cannot find evidence that greater increases in agi during one’s contribution years necessarily indicate that more should always be put in a roth ira. applied mathematical models to cover optimal retirement behavior (ragsdale, seila, and little, 1994; welch, 2008) discuss salient points involved in our article’s rvd procedure. they also examine concerns beyond our rvd emphasis such as that by welch (2015). welch states that comparing the optimal retirement planner (orp) model to other models that do not include progressive income taxes is impractical. because of our focus on the critical rvd decision, our procedure is more simplified. for example, the orp model can require dozens of inputs, while our procedure requires only 10 values to be inputted by the user. we keep our inputs down by supplying values for eight other variables (such as the expected inflation rate) although individual advisors can modify our supplied values. our focus on the rvd allocation choice is justified because of its central importance in retirement planning as the roth ira and deductible ira are the two best choices for retirement investment in that they offer huge tax advantages compared with other alternatives. thus, the main task for retirement planning is deciding what percentages of one’s retirement funds should be allocated to these two ira types. in attempting this task, we extend the rvd research in a pioneering fashion by giving a new procedure that precisely determines the optimal rvd allocation through innovative algorithms and new formulas that include the marginal cost-benefit notion. in addition, our procedure incorporates future tax tables that determine contribution and withdrawal tax rates as well as future and present 374 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 value equations that generate the required future value lump sums and annuity withdrawal cash flows based on expected rates of return and cost of living increases. for simplicity, our withdrawals are annuities. suarez, suarez, and walz (2015) develop withdrawal strategies for retirement portfolios. the use of such strategies could be used as a substitute to our annuity withdrawal choice to enable greater maximum gains. two consideration make our rvd procedure workable in achieving optimal lifetime income allocation. first, we can update inputted values periodically to ensure the investor is allocating their ira contributions correctly. second, we generate an optimal rvd allocation outcome that covers a flat range of optimal percentages determined by one’s contribution tax rate (tc) and withdrawal tax rate (tw). the flat range that occurs around one’s optimal rvd allocation allows for a margin of error for any values inputted in our procedure. we offer three main outcomes when using our procedure to compute an investor’s optimal rvd allocation. first, we determine the tax rate differential defined as �t � tc � tw and use both average and marginal tc and tw values. second, we detect the deductible ira withdrawal (diw) that maximizes the gain from attaining the optimal rvd allocation. the gain is a function of �t and diw. third, we provide the percentage of ira retirement funds that should be allocated between the deductible and roth ira types. the optimal deductible ira withdrawal percentage (odi%) is the optimal diw as a percentage of the maximum diw. while we focus on the role of tc and tw in making the rvd decision, there are other factors that could be applicable giving an advantage to either a roth ira (such as less restrictions on minimum distributions during retirement) or a deductible ira (lowering one’s taxable income to be eligible for educational tax credits). while these factors can be important, we believe they are minor for most investors when compared with the substantial economic value that results from the interplay of tc and tw. we organize the remainder of our article as follows. in section 2, we provide background for understanding the rvd allocation decision. section 3 reviews retirement decisionmaking models and their equations while section 4 introduces and explains the new variables needed to develop an innovative formula for computing the maximum marginal gain needed to pinpoint the optimal rvd allocation. in section 5, we overview the key variables, provide tax bracket information, compute tc using our contribution algorithm, and produce ira cash flow information. section 6 contains our withdrawal algorithm that generates outcomes such as tw, maximum gain, and the optimal rvd allocation. in section 7, we plot the optimal rvd withdrawal range and illustrate how the maximum gain changes based on various assumptions about the match and other income withdrawn during retirement. in section 8, we perform scenario analysis that show how changes in key variables influence the optimal rvd allocation. section 9 offers concluding remarks and a disclaimer. 2. relevant literature on roth ira versus deductible ira choice a deductible ira is an individual retirement account, established in the united states by the employee retirement income security act of 1974. at the time of the introduction of the roth ira brought about by the passage of the taxpayer relief act of 1997, researchers used mathematical frameworks, such as the scholes and wolfson or “sw” (1992), to choose 375r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 among retirement savings alternatives available at that time. these alternatives included a deductible ira, a nondeductible ira, and a non-ira investment. while the deductible ira is viewed as superior because of its tax savings, there are complexities to consider before knowing if a nondeductible ira is superior to a non-ira investment. major complexities involve unknown investment return rates and personal tax rates as well as different tax rates that are applicable to the various types of investments. with the introduction of the roth ira in 1997, the complexities increased and the rvd research formally began (adelman and cross, 2010; anderson and hulse, 2013; crain and austin, 1997; horan, peterson, and mcleod, 1997; horan and peterson, 2001; hrung, 2007; hulse, 2003; krishnan and lawrence, 2001; shynkevich, 2013; sibley, 2002). the rvd research has evolved over time with the changes in legislation that have seen tax rates lowered, income limits to ira contributions raised, and income caps suspended. these changes have not only made past research obsolete in terms of its specifics but also caused ongoing complications in dealing with how an individual determines their optimal rvd allocation. crain and austin (1997) utilize the sw framework to analyze the rvd complexities including the choice between a non-ira investment and a nondeductible ira investment. they state the latter two choices are inferior to a deductible ira and a roth ira. they note the rvd allocation choice favors a deductible ira when the tw is less than the tc. horan, peterson, and mcleod (1997) point out the situations that favor a conversion of a deductible ira into a roth ira. regarding conversion, the decision is similar to crain and austin in that conversion to a roth ira should not be done when tc�tw holds. horan and peterson (2001) analyze the rvd decision when the tax savings from the deductible ira is invested in a mutual fund with some inherent tax-deferral characteristics. they compare a roth ira with an employer-sponsored 401(k) plans that match some or all of an employee’s contributions to a deductible ira.1 like others, they reiterate that the rvd decision favors the deductible ira if tc�tw. horan and peterson wrote in a time of greater restrictions governing employers and roth ira contributions but that would change over time, as restrictions would be loosened. while investors have been able to choose a roth 401(k) for their own contribution since 2006, the employer’s matching plan does not allow the employer’s match to be put into a roth ira. hulse (2003) suggests the conversion option is valuable and investors should consider it in the rvd allocation choice. because the roth ira does not have a conversion option, a deductible ira is ceteris paribus favored over a roth ira. hrung (2007) empirically examines the influence of tax and nontax factors on the rvd choice. he discovers taxpayer liquidity is often more important in choosing between ira types than tax factors. for example, investors with children are more likely to contribute to a deductible ira with the tax savings used for current expenditures or to increase one’s current ira contribution. adelman and cross (2010) and anderson and hulse (2013) revisit the rvd dilemma by comparing the deductible and roth ira types. their results indicate the rvd choice can be influenced by theoretical or practical assumptions an investor makes related to a tax bracket effect, required minimum distributions (rmds), and impact of withdrawals on the amount of social security benefits taxed. anderson and hulse note that tax law changes (such as the american taxpayer relief act of 2012) have relaxed restrictions making the roth rollover 376 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 option available to many more participants. they develop a framework for rvd decisionmaking to help an investor decide if a roth rollover is advantageous. as can be seen from our overview of general background information on the rvd choice, the research has been frequently directed towards decisions other than a strict roth ira versus deductible ira choice. it regularly focuses on conversion from a deductible ira to a roth ira reflecting financial advisors main task, which is dealing with that segment of the population over 50 years of age who hold most of the wealth. in this article, we have a different orientation as we focus on a younger age group who are beginning their retirement investing and have modest earnings. for these future retirees, they should be focusing on their optimal rvd choice. this article addresses this focus and demonstrates that sizeable gains in lifetime wealth can be achieved even for modest income earners. 3. retirement decision alternatives and models in this section, we first discuss the proper comparison between the roth and deductible ira types. we then overview four alternatives investments that can be used for retirement withdrawals. these four investments include two inferior alternatives (non-ira and nondeductible ira) and the two superior alternatives (deductible ira and roth ira). finally, we describe the need to create a procedure to determine a precise allocation between the two superior ira alternatives. while we focus on a comparison between the two superior ira types, other retirement income should not be ignored. as we will illustrate later in figs. 1 and 2, withdrawals from an employer’s match and other investments (oi) during retirement cause the optimal rvd decision to allocate more to a roth ira. 3.1. retirement investment alternatives crain and austin (1997) begin modern day rvd modeling. they point out the superiority of a deductible ira and a roth ira to a nondeductible ira or any non-ira investment. horan, peterson, and mcleod (1997) borrow from the analysis of crain and austin and permit investors to fall into lower tax brackets upon withdrawal of retirement assets. this not only has a significant effect on the rvd choice but also the investor’s choice between nondeductible ira contributions and taxable mutual fund investments. they state that comparing future values of taxable investments and nondeductible ira investments require establishing the formulas governing their after-tax accumulations.2 seida and stern (1998) use the sw framework to analyze the rvd dilemma. they argue that a roth ira investment made after taxes are paid cannot be directly compared with a before-tax investment in a deductible ira because the latter costs less because of savings from a tax deduction. a roth ira investment also cannot be compared with an equal dollar amount of after-tax deductible ira investment if the deductible ira’s initial balance would exceed statutory limitations. seida and stern argue that a proper comparison of a deductible ira to a roth ira requires that a deductible ira be coupled with a supplemental outlay equal to the tax savings. however, they were writing in an era with limited maximum contributions and restrictions on ira contributions caused by ceilings on income that made 377r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 the proper comparison hard to attain. in today’s investment environment, most investors should have little problem in investing the tax savings in a deductible ira without exceeding maximum ira contributions. for example, irs numbers as of 2016 indicate it is possible that a person can invest up to $76,500 (or $89,500 if over 50 years of age) in iras. thus, an investor today should have little trouble in investing the tax savings in a deductible ira. in this article, we focus on the proper comparison that involves contrasting the dollar amount of a roth ira with that of a deductible ira plus its tax savings. except for unusual situations, it should be rare for this principle to be violated. however, if the proper comparison principle is violated, our procedure breakdowns because it assumes an investor’s deductible ira contribution includes all tax savings from the deductible ira. violation of this principle means that one could actually invest more retirement dollars by choosing a roth ira. 3.2. nondeductible ira and a non-ira comparison while we consider the nondeductible ira and non-ira alternatives inferior, they are still important, as some investors will be in a position to save beyond the two superior ira choices. funds withdrawn from the two inferior choices can influence the optimal rvd choice. this is because the roth ira becomes a more favorable choice over a deductible ira when investors have other sources of income that increase their withdrawal tax rate. in addition to deductible and roth iras, the horan, peterson and mcleod, hpm, (1997) formulas (grounded in seida and stern) consider non-iras and nondeductible ira. given that earnings of a nondeductible ira are taxed at the ordinary rate upon withdrawal while its actual principal contributed is not taxed on withdrawal, hpm state the after-tax future value of a nondeductible ira dollar (fvatn) is fvatn � 1 � ([1 � rc]yc � 1)(1 � tw) (1) where tw � ordinary marginal tax rate upon withdrawal, rc � expected rate of return on contribution, and yc � number of years until withdrawal. for a non-ira taxable mutual fund investment (fvbtm), hpm state that a dollar invested has, at withdrawal, a before-tax future value of fvbtm � (1 � rc � rcpoitc � rcpcgtcg)yc (2) where poi � percent of annual return distributed as ordinary income, tc � ordinary marginal tax rate,3 pcg � percent of annual return distributed as capital gains, and tcg � intermediate marginal tax rate on capital gains. hpm add that a capital gain tax is recognized is based on the before-tax future value at withdrawal of fvbtm as given in (2) less the adjusted basis, composed of the initial investment and distributions less the income tax on distributions. at withdrawal, the after-tax future value of a taxable mutual fund investment (fvatm) is 378 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 fvatm � (1 � rc � rcp0itc � rpcgtcg)yc � tcg� (1 � rc � rcp0itc � rcpcgtcg)yc � 1 � rcp0i(1 � tc) (1 � rc � rcp0itc � rcpcgtcg)yc � 1 rc � rcp0itc � rcpcgtcg � rcpcg(1 � tcg) (1 � rc � rcp0itc � rcpcgtcg)yc � 1 rc � rcp0itc � rcpcgtcg � (3) the term inside the brackets of (3) represents the before-tax accumulation at withdrawal less the adjusted basis. hpm state one may want to forgo the ira tax deferral in exchange for paying lower capital gains tax. when comparing a nondeductible ira and a non-ira, the objective is to determine the percentage of capital gains distribution that makes one indifferent between a taxable mutual fund investment and the same investment in a nondeductible ira. hpm set (1) equal to (3) to get 1 � [(1 � rc)yc � 1](1 � tw) � 1 � rc � rcp0itc � rcpcgtcg)yc � tcg� (1 � rc � rcp0itc � rcpcgtcg)yc � 1 �rcp0i(1 � tc) (1 � rc � rcp0itc � rcpcgtcg)yc � 1 rc � rcp0itc � rcpcgtcg � rcp0i(1 � tc) (1 � rc � rcp0itc � rcpcgtcg)yc � 1 rc � rcp0itc � rcpcgtcg � (4) they use this expression to solve for the percentage of return distributed as capital gain (pcg) subject to 0 � pcg � 1, 0 � poi � 1, and 0 � pcg � poi � 1. 3.3. roth ira and deductible ira comparison while the roth ira allows for all future earnings and withdrawals to be free from tax, it is taxed at the ordinary rate when contributed. as such, an after-tax future value roth ira dollar is fvroth � $1(1 � tc)(1 � rc)yc (5) where tc � ordinary marginal tax rate upon contribution. for a conversion into a roth ira, hpm assume all assets being converted are deductible contributions and earnings subject to tax. the after-tax future value of a dollar in an existing deductible ira account is fvatd � $1(1 � rc)yc(1 � tw) (6) where tw � ordinary marginal tax rate upon withdrawal. at the time they were writing, hpm stated that investors with agis of no more than $100,000 might convert existing deductible iras to roth iras. if the tax liability is paid 379r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 from the ira assets in the year of conversion then investors should convert when fvroth � fvatd or when we have fvroth fvatd � (1 � tc)(1 � rc)yc (1 � tw)(1 � rc)yc � 1. (7) equation (7), in essence, states to convert if tw�tc. hpm argue that paying the tax liability from the ira assets is suboptimal if the tax liability paid out of the assets being converted decreases the principal in the new roth ira. in this case, the future value of a converted roth ira is simply (l � tc)(l � rc)yc. alternatively, the tax liability can be paid from assets that would not qualify for tax-deferred status, leaving the principal in the new roth ira unchanged from the deductible ira. in this case, the future value of a converted ira dollar equals the future value of the new roth ira dollar less the after-tax future value of conversion tax. we have fvroth�(1 � rc)yc � tc(fvatm) (8) where fvatm is the after-tax future value of a taxable mutual fund investment from equation (3). the first term in (8) represents the future value of a dollar in the new roth ira. the second term represents the lost future value of the tc dollars used to pay the conversion tax. substituting equation (3) into equation (8) gives fvroth � (1 � rc)yc � tc�(1 � rc � rcp0itc�rcpcgtcg)yc � tcg� (1 � rc � rcp0itc � rcpcgtcg)yc � 1 � rcp0i(1 � tc) (1 � rc � rcp0itc � rcpcgtcg)yc � 1 rc � rcp0itc � rcpcgtcg �rcp0i(1 � tc) (1 � rc � rcp0itc � rcpcgtcg)yc � 1 rc � rcp0itc � rcpcgtcg �� (9) given the above, hpm adds that conversion should be made when fvroth fvatd � (1 � rc)yc � tc(fvatm) (1 � tw)(1 � rc)yc � 1. (10) 3.4. extending prior rvd research the prior rvd research points out the interplay of tw and tc be it converting from a deductible ira to a roth ira or trying to determine which ira type should be put in one’s retirement funds. while financial advisors assumedly know this interplay and can attempt to guide their clients towards an optimal ira allocation, they do not have a rigorous procedure to guide them with scientific precision. we extend the prior rvd research by 380 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 addressing this issue through our rvd procedure that provides definitive guidance to financial advisors. keying on the prior research finding that shows a roth is favorable when tw�tc, our rvd procedure takes this finding to its logical conclusion by computing both marginal and average tc and tw values for use in achieving maximum gain in retirement wealth. with our algorithms and formulas in hand, we determine the maximum gain from all possible deductible ira withdrawals (diws). the point where the gain is maximized establishes the percentage of one’s total ira contribution that should be allocated to a deductible ira with the remainder allocated to a roth ira. 4. approaches and alternatives to computing optimal rvd allocation in this section, we discuss two approaches to determine the optimal rvd allocation: the “year-by-year” approach and the “mean” approach. we then clarify “marginal” versus “average” use of tax rates. we also present our two gain formulas. the average gain formula uses average tax rates and the marginal gain formula uses marginal tax rates. the key to a marginal gain formula lies in identifying discovery points (dps) as they signify the diw dollar that causes the taxable income to jump to a higher marginal tax rate. dps are uniquely determined by different retirement withdrawals that cause different taxable income. besides a diw, a retiree’s taxable income is influenced by tax deductions, social security benefits (ssb), employer salary match, and other investments (oi). when used together, we refer to the latter two items as match/oi. 4.1. the two rvd allocation approaches we develop two approaches to determine an optimal rvd allocation. we first develop a “year-by-year” rvd allocation approach that looks at each year separately. with this approach, we consider each year’s contribution and tax savings. we then, if necessary, separate out tax savings made at different marginal tax rates. we then compute the future value of each individual tax savings. this allows earlier contributions to grow more and have a greater weight. we next partition future values for all years based on savings gotten from their specific marginal tax rates. the partitions form the annual withdrawals that are designed to pay taxes below the tax savings. each withdrawal year has its own gain calculation and other rvd outcomes. while this approach should be more accurate, it has the disadvantage of requiring one to generate a large number of yearly values causing complications including cumbersome complexities when presenting its rvd procedure. thus, we reserve the yearby-year approach for future research until the basic principles of this article’s rvd procedure are better known making its complex presentation more workable. second, we develop a “mean” rvd allocation approach that creates factors generating mean contribution and withdrawal values to replace annual values. this approach has the advantage of quickly creating a plethora of withdrawals from which the maximum gain can be identified while also showing the cost of straying from the optimal. by using the mean approach, we can more easily generate results including those from scenario analysis. for these reasons, this article uses the mean approach as the beginning point in creating research interest on discovering the optimal rvd allocation. 381r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 4.2. the two tax rate alternatives within any optimal rvd allocation approach are two alternative uses of tax rates, namely, the use of either the “average” tax rate or the “marginal” tax rate. we define our average contribution tax rate as average tc � tax savings from deductible ira contributions / deductible ira contributions. our contribution algorithm in section 5.3 computes the average tc. we define our average withdrawal tax rate as average tw � taxes paid on all retirement withdrawals / taxable income on all retirement withdrawals. our withdrawal algorithm in section 6 computes the average tw. while the marginal tc involves the tax rate(s) corresponding to the highest taxable income bracket(s) for which the last dollars earned are taxed, the marginal tw covers tax rate(s) corresponding to the highest taxable income bracket(s) for which the last dollars withdrawn are taxed. we are able to avoid the average versus marginal usage decision for tc because the average tc and marginal tc are similar for our illustration and most other situations. however, similarity in the average tw and marginal tw rarely occurs.4 in this article, we will focus on the use of marginal tax rates. this usage is consistent with marginal analysis that focuses on the incremental benefits and costs of an activity and considers the opportunity cost of the next best alternative. when looking at the incremental benefit of using a dollar of deductible ira, the next best ira alternative would be putting the dollar in a roth ira. when an investor contributes to a deductible ira account, the incremental benefits of tax savings from dollars contributed are viewed as based on the last dollars contributed. these last dollars will be in the highest tax bracket(s). similarly, when an investor withdraws deductible ira contributions, we should view the incremental costs of the last dollars withdrawn and assign them the highest tax bracket(s). thus, if investors have the match/oi included in their retirement withdrawals, these withdrawals are assumed to absorb the lowest marginal tax bracket(s) with the diw partitions absorbing the highest bracket(s). given the above discussion, we will focus on the mean approach and the marginal use of tax rates. nonetheless, we think there is merit in knowing average tax rates. because the average and marginal tc values are the same for this article’s rvd illustration, we focus on what extent, if any, the use of marginal tw disagrees with the use of the average tw. thus, we will report results using both marginal and average tw values. in the process, we can identify situations for which the use of average tax rates render similar results to the use of marginal tax rates. 4.3. discovery points and eight tax rate differentials when using average tax rates, we use an average gain formula. this formula is average gain � (tc � tw)(diw) (11) 382 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 where tc and tw are average tax rates as defined above and diw is the before-tax deductible ira withdrawal. determining a marginal gain formula is more difficult because we require partitioning of diw based on applicable tax brackets. we use these eight brackets: 0, 0.12, 0.18, 0.3, 0.336, 0.396, 0.42, and 0.471.5 the general expression for the marginal gain is �j�1 8 �tj (partitionj) where �tj � (tc � tw)j and each �t is aligned with its corresponding partition. identifying taxable income levels for each �t is a prerequisite for developing a gain formula using marginal tax rates. to perform this task, we first ascertain what we call discovery points (dps). we define a dp as: discovery point � the taxable income dollar that leads to a jump to a higher marginal tax rate. dps enable us to determine the partitions that correspond to unique �t values. dps for each investor’s diw has to be independently found. this is because an investor’s taxable income is influenced by the factors that can be different, namely, diw, tax deduction, ssb, and match/oi. to illustrate dps, we assume tax deduction � $50,000; taxable proportion of ssb � $40,000; match/oi � $0; diw � $15,000; and, tc � 0.18. because tax deduction � ssb � match/oi, we have $50,000 � ($40,000 � $0) � $10,000 for which tw � 0 as it is not taxed. we refer to this $10,000 as the slack and define it as: slack � diw partition that is not taxed where slack � 0 if tax deduction � ssb�match/oi. if there is no slack then the lowest marginal tw that can occur is 0.12. in our illustration, the $10,001st deductible ira dollar will get the marginal tax rate of tw � 0.12. the $10,001st dollar is dp1. before this, tw � 0 and so marginal gain1 � �t1(diw1) � (tc � tw)$10,000 � (0.18 � 0)$10,000 � $1,800. there remains $15,000 � $10,000 � $5,000 in a deductible ira, and so we have: marginal gain2 � �t2(diw2) � (tc � tw)$5,000 � (0.18 � 0.12)$5,000 � $300. the total marginal gain is $2,100. if dp2 is $39,001 because we jump to a higher tax bracket with the $39,001st dollar, then the partition between dp1 and dp2 would be $39,001 � $10,001 � $29,000 and a diw greater than $15,000 could have a marginal gain2 as large as 0.06($29,000) � $1,740. in table 1, we give dps, changes in dps, marginal tw values, and �t values. because there are seven federal tax brackets including the possibility of slack where part of diw can be taxed at zero percentage, we have eight dps. they are important because they reveal where tax rates jump to a higher marginal level. in panel a, we label these eight dps from dp1 to dp8. the eight dps are identified by our withdrawal algorithm presented in section 6. they correspond to eight �t values that cover the changes in tw from 0 to 0.471. when using a marginal analysis, identifying dps are vital in computing the marginal gain from the ira allocation between the two major ira types. in panel b, we report the eight �t ranges with the ranges created by subtracting dps. these differences signify maximum partitions for which a corresponding tw is applicable. in panel c, we give eight tw values corresponding to the eight dp partitions. the beginning of each dp partition represents a jump to a higher tw value. for example, the beginning of the partition of dp3-dp2 is dp2 � $69,111 and the $69,111th diw dollar achieves a jump 383r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 in tw from 0.12 to 0.18. in panel d, we list the eight tax rate differentials (�ts). each �t is computed using the marginal tc (i.e., 0.30 as determined by our contribution algorithm in section 5) and one of the eight tw values. of interest, the changes in larger adjoining dps become the tax bracket ranges that can be gathered from information in the last row of table 2. for example, beginning with dp4dp3 � $151,367, we see that this is the tax bracket range ending with $300,166 for 0.30 and $457,363 for 0.336 that covers $457,363 � $300,166 � $151,367. if there were no ssb, no match/oi, and no tax deductions, dps would be determined solely by tax brackets. 4.4. five-step procedure to compute the marginal gain given the above preliminary information, we can now create our marginal gain formula. it is marginal gain � �t1dp1 � �t2(dp2 � dp1)� . . . � �t8(dp8 � dp7). (12) given (12), we describe our five-step procedure to compute the marginal gain. first, we identify as many dps as needed to form partitions to cover diw. second, we associate each partition with its corresponding marginal tw. using our marginal tc � 0.30, we compute all applicable �t values so that each �t is now associated with its corresponding partition. third, we compute the average and marginal gains given in (11) and (12). to illustrate (12) when the match/oi � 0, consider a diw of $69,111, which is dp2. using the information in table 1, we have: marginal gain � �t1dp1 � �t2(dp2-dp1) � (0.30 � 0)$46,397 � (0.30 � 0.12)($69,111 � $46,397) � 0.30($46,397) � 0.18($22,714) � $13,919 � $4,089 � $18,008. table 1 discovery points, �t ranges, marginal tax rates, and tax rate differentials panel a. eight discovery points determining maximum partitions applicable for a marginal tw dp1 dp2 dp3 dp4 dp5 dp6 dp7 dp8 $46,397 $69,111 $156,512 $307,879 $465,076 $824,525 $890,921 unlimited panel b. eight tax rate differentials (�ts) ranges created by subtracting discovery points (dps) dp1-0 dp2-dp1 dp3-dp2 dp4-dp3 dp5-dp4 dp6-dp5 dp7-dp6 dp8-dp7 $46,397 $22,714 $87,401 $151,367 $157,197 $359,449 $66,396 unlimited panel c. eight marginal tw values corresponding to the eight �t ranges 0.000 0.120 0.180 0.300 0.336 0.396 0.420 0.471 panel d. eight tax rate differentials (�t) when tc is 0.30 �t1 �t2 �t3 �t4 �t5 �t6 �t7 �t8 0.300 0.180 0.120 0.000 �0.036 �0.096 �0.120 �0.171 note: panel a contains the eight discovery points (dps) applicable to this article’s rvd illustration. dp is defined as the taxable income dollar that leads to a jump to a higher marginal tw. dps are identified by our withdrawal algorithm presented in section 6. panel b covers the eight ranges determined by adjoining dps with larger adjoining dps often representing the tax rate bracket ranges given in table 3. panel c provides the eight marginal tw values that correspond to the eight ranges in panel b. panel d gives the eight tax rate differentials (�t) corresponding to the eight ranges in panel b where �t � tc � tw with the marginal tc � 0.30 (as computed in table 4) used for all dps. 384 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 2 values used in rvd procedure panel a. ten values supplied by a couple filing jointly birth year 1986 age 30 retirement year 2050 portfolio mix during contribution years (proportion in stock) 0.95 portfolio mix during withdrawal year (proportion in stock) 0.60 adjusted gross income (agi) $110,000 salary $106,000 ira contribution (without the match/oi and without the reinvested tax savings) $14,661.61 ira employee salary match (3% of combined salary): 0.03($106,000) � $3,180 $3,180 other investments (oi): funds invested that are source of retirement withdrawals $3,180 note: together the match and oi (called match/oi) � $3,180 � $3,180 � $6,360 panel b. eight values supplied by our procedure that can be modified yw (years of withdrawal from ira based on life expectancy) 20 annual inflation rate for remainder of lifetime 1.5% nominal annual rate of return for stocks 8.3743% premium of stocks over non-stock investments 3.5% annual growth rate in agi/salary/match/oi 3% current tax deduction (does not include the tax deduction from the deductible ira) $25,600 social security benefits (ssb) worksheet, line 8 (current value, married joint filer) $32,000 social security benefits (ssb) worksheet, line 10 (current value, married joint filer) $12,000 panel c. computed rates of return rc (nominal annual contribution rate) � 0.95(8.3743%) � (1�0.95)(8.3743%�3.5%) 8.1993% rw (nominal annual withdrawal rate) � 0.60(8.3743%) � (1�0.60)(8.3743%�3.5%) 6.9743% panel d. four factors contribution factor for years 1–35 (based on 1.5% growth) 1.302631083 agi/salary/match/oi factor for years 1–35 (based on 3% growth) 1.727488052 withdrawal factor for years 36–55 (based on 1.5% growth) 1.976078686 ira factor for years 1–35 used with annual rates of return (based on 3% growth) 1.383340574 panel e. mean values agi (years 1–35) � agi/salary/match/oi factor(current agi) � 1.727488052($110,000) $190,024 ira contribution (years 1–35) � ira factor(maximum ira contribution) � 1.383340574($14,661.61) $20,282 match/oi (years 1–35) � ira factor(current match/oi) � 1.383340574($6,360) $8,798 tax deduction (years 1–35) � contribution factor(current tax deduction) � 1.302631083($25,600) $33,347 tax deduction (years 36–55) � withdrawal factor(current tax deduction) � 1.976078686($25,600) $50,588 ssb worksheet, line 8 (years 36–55) � withdrawal factor(current value) � 1.976078686($32,000) $63,235 ssb worksheet, line 10 (years 36–55) � withdrawal factor(current value) � 1.976078686($12,000) $23,713 future estimated annual ssb during retirement (estimated based on agi) $50,441 note: panel a contains 10 current values inputted by our couple who are assumed to have the same age and birth retirement years. an 11th value is the sum of the salary match and other investments (oi). oi can include nondeductible ira, inheritance, cash value life insurance policies, non-ira mutual fund investments, or rental income used for withdrawal during retirement but does not include social security benefits, employer salary match, or deductible and roth iras. the dollar values are annual values. panel b contains eight values supplied by our procedure. panel c computes annual returns for contribution and withdrawal periods. panel d provides the four factors used to simplify analysis and create mean values used to simplify our rvd procedure. panel e gives mean values over time for current values based on the factors in panel d. 385r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 fourth, if the last diw dollar is between two adjacent dps, we substitute it for the larger of the two dps in (12). to illustrate, if the last diw dollar is $50,000 instead of $69,111, we substitute $50,000 for $69,111 and get marginal gain � �t1dp1 � �t2($50,000-dp1) � 0.30($46,397) � 0.18($50,000 � $46,397) � $13,919 � $649 � $14,568. if the withdrawal includes both diw and match/oi, we replace diw with this withdrawal and use its last dollar. fifth, if we have a match and/or oi withdrawal that lies between two adjacent dps, then we use this withdrawal for the smaller of the two dps and any component where all dps are less than the match and/or oi exits (1). to illustrate, if dp2 � $69,111, dp3 � $156,513, and match � $74,655, then we use $74,655 for dp2 and get marginal gain � �t3(dp3match) � (0.30–0.18)(dp3-match) � 0.12($156,513 � $74,655) � $9,823 where the first two components of (1), �t1dp1 � �t2(dp2-dp1), exit the equation. diw for this illustration is dp3 � match � $156,513 � $74,655 � $81,858. thus, whereas the annuity withdrawal is $156,513, only $81,858 is from a deductible ira. 5. inputs, tax brackets, tc, and tw in this section, we introduce the variables and values for our couple. we next provide tax brackets applicable to our couple over their lifetime followed by our contribution algorithm that computes tc. lastly, we generate future values for contributions to deductible and roth iras with these future values supplying retirement withdrawals. 5.1. inputs and values table 3 contains values used in our rvd procedure. in panel a, we display 10 variables and their values supplied by our couple. we assume the same values for birth year, current table 3 tax information from 2016 to 2070 marginal tax rate 0.120 0.180 0.300 0.336 0.396 0.420 0.471 current year (2016) $0-$18,550 $18,551-$75,300 $75,301-$151,900 $151,901-$231,450 $231,451-$413,350 $413,351-$466,950 $466,951 and greater contribution years (2016–2050) $0-$24,164 $24,165-$98,088 $98,089-$197,870 $197,871-$301,494 $301,495-$538,443 $538,444-$608,264 $608,265 and greater retirement year (2050) $0-$30,774 $30,775-$124,992 $124,993-$252,002 $252,003-$383,975 $383,976-$685,746 $685,747-$741,488 $741,489 and greater withdrawal years (2051–2070) $0-$36,656 $36,657-$148,799 $148,800-$300,166 $300,167-$457,363 $457,364-$816,812 $816,813$922,731 $922,730 and greater note: the first row reports marginal tax rates for periods used in our rvd illustration. the marginal tax rate combines the 2016 federal income marginal tax rate information and 2016 state income marginal tax information that is estimated by averaging the mean statutory state tax rates for all fifty states and matching them as best as possible with the federal taxable income brackets. the seven federal tax rates are 0.10, 0.15, 0.25, 0.28, 0.33, 0.35, and 0.396. we only report information relevant to our illustration, which is the “married joint filers” taxable income bracket ranges but information for single filers, married separate filers, and head of household filers could be similarly gotten. 386 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 age and retirement year for our couple. the portfolio mixes are consistent with investors with a long-term, aggressive strategy. the agi, salary, employer’s salary match, other investments (oi), and ira contribution are values that occur at the end of the current year. the ira contribution of $14,661.61 is the maximum amount our couple would invest in an ira if all of their investment went into a roth ira so that there are no tax savings from a deductible ira that could be used for investing in an ira. the match and oi are each $3,180 and together the match/oi is $6,360. to the extent oi consists of dividends and long-term capital gains, oi does not influence tw but can still be a factor in determining one’s optimal rvd allocation. this is because larger diws can increase the long-term capital gains and dividend tax rates by bumping them from 0% in the two lowest tax brackets to either 15% in the intermediate tax brackets or 20% in the two highest tax brackets. in panel b, we provide our supplied values used in our rvd procedure. consistent with the social security life expectancy calculator, we assume a 20-year retirement period. the inflation rate of 1.5% is based on recent periods such as given by http://www.multpl.com/ inflation/table. the latter reports annual inflation rates based on the 12-month change in the consumer price index (data courtesy of the u.s. bureau of labor statistics and robert shiller). we supply a nominal rate of stock return of 8.3743% based on the average of the s&p 500 and nasdaq composite returns since 1971. nonstock investments can be construed as fixed-income investments assumed to earn 3.5% less than stocks. the 3% annual growth rate is consistent with the historical growth in the u.s. gdp over the past 35 years as given by the world bank. we expect the 3% growth rate for the salary match to hold for our couple as this growth will prevent them from reaching the ira contribution limits. this is because they are currently well under the current limits of $47,000 with these limits increasing at age 50 to $61,000. the current tax deduction of $25,600 includes personal exemptions, standard deductions and factors in itemized deductions and tax credits. it does not include tax deductions from a deductible ira as that is determined later once the optimal diw is known. the two social security values of $32,000 and $12,000 are current dollars that grow over time at our annual inflation rate of 1.5%. panel c computes the nominal annual rates of return during contribution and withdrawal years as 8.1993% and 6.9743%, respectively. as can be seen from their calculations, these returns are determined by their portfolio mixes. panel d gives four factors used to create mean values for use in our rvd procedure that uses our mean approach. the use of these factors enable us to avoid computing 35 values for variables used during our contribution period and 20 values for variables used during our withdrawal period. below we explain these factors. the general formula used to compute our first three factors is: factor � (1 � x)t�1 (1 � x)0 � (1 � x)1 � . . . � (1 � x)n�2 � (1 � x)n�1) n (13) where x is the assumed rate of increase or growth per year; t is the year the first cash flow begins; and, n is the number of years in the period being considered. using (13) with x � 1.5%, t � 1 and n � 35, we get our contribution factor. we have: 387r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 contribution factor � (1.0151�1) (1.0150 � 1.0151 � . . . � 1.01533 � 1.01534) 35 � (1.0150) 45.59208789 35 � (1)1.302631083 � 1.302631083. a current value is multiplied by this factor to get the mean value for the 35-year contribution period. our 35-year agi/salary/match/oi factor of 1.727488052 is computed in the same manner, except we use 3% in (13) instead of 1.5%. for our withdrawal factor where t � 35 and n � 20, we have to modify (13) by increasing all exponents by plus one. after doing this, we have: withdrawal factor � (1.015)35 (1.0151 � 1.0152 � . . . �1.01518 � 1.01520) 20 � (1.01535) 23.47052211 20 � (1.68388132)1.17352611 � 1.976078686. to illustrate, a value of $18,550 in the current tax bracket for a married jointly filer becomes 1.976078686($18,550) � $36,656 for the 20-year withdrawal period of 2051–2070. the value of $36,656 is the same value wrought from computing the 20 future values for $18,550 from 2051 through 2070 and then averaging them.6 the process to compute our ira factor in panel d is as follows. first, we calculate the future value of the initial investment of $14,661.61 if it is growing annually at 3% and achieves an annual nominal rate of 8.1993%. we have: $14,661.61�1.03�0�1.081993�34 � . . . � $14,661.61�1.03�34�1.081993�0 � $3,653,762. second, we divide $3,653,762 by the future value annuity factor of (1 � 0.081993)35 � 1 0.081993 � 180.14799987 to generate the equivalent annual annuity cash flow of $20,282. dividing this cash flow by $14,661.61 gives our ira factor of 1.383340574. creating this factor has the advantage of only requiring the advisee to identify their current ira contribution of $14,661.61. because this factor is dependent on the number of contribution years, growth in agi and nominal annual contribution rate, it has to be generated separately when any of these values change. in a spreadsheet format, our factors are recomputed automatically when the values changed. panel e gives mean values for contribution and withdrawal periods by multiplying a current value by its applicable factor. to illustrate, we use the agi/salary/match/oi factor to calculate a mean agi of 1.727488052($110,000) � $190,024 for the contribution years 1–35 (that for our illustration are years 2016–2050). in regards to the $20,282 discussed previously, we can show that this ira contribution is achievable for our couple. first, for investors working for companies with a 401(k) plan, the current allowable ira contribution is $18,000. second, besides the 401(k), their income levels allow them to each qualify for a 388 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 non-401(k) ira of $5,500 that can be invested in an ira. under current law, they can invest $18,000 � $5,500 � $23,500 in either a deductible ira or a roth ira per person or $47,000 per couple. given our contribution factor, the mean for the next 35 years can be represented by 1.302631083($47,000) � $61,224. this amount is not only much greater than the couple’s maximum roth ira contribution of $20,282 but will also more than cover the maximum they can put in a deductible ira where the mean is $20,282/(1-tc) for our couple’s contribution years. the final value in panel e is our couple’s future retirement estimate for ssb. estimating ssb is difficult because of uncertainties surrounding social security’s future solvency. we consulted the social security online benefit calculator to derive our estimate of $50,441.7 our estimate takes into account the current payout for retiring at today’s normal age of 66 and adjusting it downward because our couple will retire at the age of 64. we then use our withdrawal factor and further adjust the number downwards based on current projections about social security payouts. 5.2. four tax bracket tables in table 2, we present tax information for seven taxable income brackets. we only report information relevant to our illustration, which is the married joint filers taxable income bracket ranges. information for single filers, married separate filers, and head of household filers could be similarly gotten. the total marginal tax rate is the federal marginal tax rate plus the marginal state tax rate. we estimate the state tax rate by averaging the mean statutory state tax rates for all fifty states and match them as best as possible to the federal taxable income brackets when computing the total marginal tax rate. row one has tax information for 2016. row two has projected tax information for the 35 contribution years from 2016 to 2050 computed by multiplying each taxable income value in row one times our contribution factor of 1.302631083 given in table 3. row three provides tax information for the retirement year of 2050 by multiplying 1.01534 � 1.658996373 times each taxable income value in row one. the last row has projected tax information for the 20 withdrawal years from 2051 to 2070 computed by multiplying each taxable income value in row one by our withdrawal factor of 1.976078686 given in table 3. 5.3. computing the contribution tax rate (tc) given the information from tables 2 and 3, we compute tc using our contribution algorithm in table 4. if there is an overlap between taxable income ranges, then tc is computed using two tax rates. if there is no overlap, the average tc is the marginal tc. in table 4, we input our couple’s maximum roth ira of $20,282 (line 1), agi of $190,024 (line 2), and tax deduction of $33,347 (line 3). these values are given in panel e of table 3. the tax deduction does not include the deduction from the optimal deductible ira contribution, which is determined later after we introduce our withdrawal algorithm. while not inputting the total tax deduction can overestimate tc such is not the case for our couple because we still get tc � 0.30 (line 100) even if we were to input $33,347 plus the 389r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 4 contribution algorithm to determine the contribution tax rate (tc) 1. enter maximum amount available to invest in a roth ira. $20,282 2. enter adjusted gross income (agi). $190,024 3. enter tax deduction (for now we exclude the tax deduction when investing in a deductible ira because we have not determined the optimal rvd allocation). $33,347 4. line 2 minus line 3. this is the taxable income. $156,677 5. enter taxable income that is first taxed at 0.471. $582,212 6. enter larger between (line 4 minus line 5) and zero. this is the taxable income for those who qualify for the 0.471 tax bracket. $0 7. enter 0 if line 6 is $0; else 1. equals 1 if qualifies at least in part for 0.471 tax savings. 0 8. enter 0.471. this is the marginal tax rate for those with taxable income given on line 5 and greater. 0.471 9. divide line 1 by (one minus line 8). this is maximum deductible ira for 0.471. $38,340 10. enter smaller of line 7 and (line 6 divided by line 9). this is the percentage of the maximum tax savings for those getting 0.471. 0.00% 11. one minus line 10. this is the percentage qualifying for 0.42 overlap if between 0% and 100%. 100.00% 12. multiply lines 9 and 10. this amount qualifies to lower the taxable income using 0.471 tax rate. $0 13. multiply lines 7, 11, and 21. this overlap amount qualifies to lower the taxable income using 0.42 tax rate. $0 14. enter 0 if line 13 is $0; else 1. eliminate all but the overlap at 0.42 tax rate. 0 15. enter 0 if line 11 is 0%; else 1. eliminate those who qualified for tax savings at 0.471. 1 16. multiply lines 8, 9 and 10. dollar tax savings at 0.471. this is also total tax savings to date. $0 17. enter taxable income that is first taxed at 0.42. $538,444 18. multiply line 15 by larger of (line 4 minus line 17) and zero. the taxable income for those who qualify in 0.42 bracket. $0 19. enter zero if line 18 equals $0; else 1. equals 1 if qualifies at least in part for 0.42 tax savings. 0 20. enter 0.42. this is marginal tax rate for those with taxable income falling in the range given by lines 5 and 17. 0.42 21. divide line 1 by (one minus line 20). this is maximum deductible ira for 0.42. $34,969 22. enter smaller of line 19 and (line 18 divided by line 21). this is the percentage of the maximum tax savings for those getting 0.42. 0.00% 23. enter one minus line 22. this is the percentage qualifying for 0.396 overlap if between 0% and 100%. 100.00% 24. enter 1 if line 14 equals 0; else 0. 1 25. multiply lines 21, 22, and 24. this amount qualifies to lower the taxable income using 0.42 tax rate. $0 26. multiply lines 19, 23, and 35. this overlap amount qualifies to lower the taxable income using 0.396 tax rate. $0 27. enter 0 if line 23 equals 0%; else 1. eliminate those that maxed out tax savings at 0.42. 1 28. enter smaller of lines 27 and 15. we have now taken out those who have maxed out to date. 1 29. multiply lines 20, 21, and 22. this is dollar tax savings if getting 0.42. $0 30. multiply lines 11 and 29. this adjusts for overlap and gives new tax savings to add in. $0 31. enter taxable income that is first taxed at 0.396. $301,495 32. multiply line 28 by the larger of (line 4 minus line 31) and 0. the taxable income for those who qualify in 0.396 bracket. $0 390 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 4 (continued) 33. enter 0 if line 32 equals $0; else 1. equals 1 if qualifies at least in part for 0.396 tax savings. 0 34. enter 0.396. this is the marginal tax rate for those with taxable income falling in the range given by lines 17 and 31. 0.396 35. divide line 1 by (one minus line 34). this is maximum deductible ira for 0.396. $33,579 36. enter smaller of line 33 and (line 32 divided by line 35). this is the percentage of the maximum tax savings for those getting 0.396. 0.00% 37. enter one minus line 36. this is the percentage qualifying for 0.336 overlap if between 0% and 100%. 100.00% 38. enter 1 if line 29 equals 0; else 0. 1 39. multiply lines 35, 36, and 38. this amount qualifies to lower the taxable income using 0.396 tax rate. $0 40. multiply lines 33, 37, and 49. this overlap amount qualifies to lower the taxable income using 0.336 tax rate. $0 41. enter 0 if line 37 equals 0%; else 1. eliminate those that maxed out tax savings at 0.396. 1 42. enter smaller of lines 41 and 28. we have now taken out those who have maxed out to date. 1 43. multiply lines 34, 35, and 36. this is dollar tax savings if getting 0.396. $0 44. multiply lines 23 and 43. this adjusts for overlap and gives new tax savings to add in. $0 45. enter taxable income that is first taxed at 0.336. $197,871 46. multiply line 42 by the larger of (line 4 minus line 45) and 0. the taxable income for those who qualify in 0.336 bracket. $0 47. enter 0 if line 46 equals $0; else 1. equals 1 if qualifies at least in part for 0.336 tax savings. 0 48. enter 0.336. this is the marginal tax rate for those with taxable income falling in the range given by lines 31 and 48. 0.336 49. divide line 1 by (one minus line 48). this is maximum deductible ira for 0.336. $30,545 50. enter smaller of line 47 and (line 46 divided by line 49). this is the percentage of the maximum tax savings for those getting 0.336. 0.00% 51. enter one minus line 50. this is the percentage qualifying for 0.30 overlap if between 0% and 100%. 100.00% 52. enter 1 if line 43 equals 0; else 0. 1 53. multiply lines 49, 50, and 52. this amount qualifies to lower the taxable income using 0.336 tax rate. $0 54. multiply lines 47, 51, and 63. this overlap amount qualifies to lower the taxable income using 0.30 tax rate. $0 55. enter 0 if line 51 equals 0%; else 1. eliminate those that maxed out tax savings at 0.336. 1 56. enter smaller of line 55 and line 42. we have now taken out those who have maxed out to date. 1 57. multiply lines 48, 49, and 50. this is dollar tax savings if getting 0.336. $0 58. multiply lines 37 and 57. this adjusts for overlap and gives new tax savings to add in. $0 59. enter taxable income that is first taxed at 0.30. $98,089 60. multiply line 56 by the larger of (line 4 minus line 59) and 0. the taxable income for those who qualify in 0.30 bracket. $58,588 61. enter 0 if line 60 equals $0; else 1. equals 1 if qualifies at least in part for 0.30 tax savings. 1 62. enter 0.30. this is the marginal tax rate for those with taxable income falling in the range given by lines 48 and 59. 0.30 391r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 4 (continued) 63. divide line 1 by (one minus line 62). this is maximum deductible ira for 0.30. $28,974 64. enter smaller of line 61 and (line 60 divided by line 63). this is the percentage of the maximum tax savings for those getting 0.30. 100.00% 65. enter one minus line 64. this is the percentage qualifying for 0.18 overlap if between 0% and 100%. 0.00% 66. enter 1 if line 57 equals 0; else 0. 1 67. multiply lines 63, 64, and 66. this amount qualifies to lower the taxable income using 0.30 tax rate. $28,974 68. multiply lines 61, 65, and 77. this overlap amount qualifies to lower the taxable income using 0.18 tax rate. $0 69. enter 0 if line 65 equals 0%; else 1. eliminate those that maxed out tax savings at 0.30. 0 70. enter smaller of line 69 and line 56. we have now taken out those who have maxed out to date. 0 71. multiply lines 62, 63, and 64. this is dollar tax savings if getting 0.30. $8,692 72. multiply lines 51 and 71. this adjusts for overlap and gives new tax savings to add in. $8,692 73. enter taxable income that is first taxed at 0.18. $24,165 74. multiply line 70 by the larger of (line 4 minus line 73) and 0. the taxable income for those who qualify in 0.18 bracket. $0 75. enter 0 if line 74 equals $0; else 1. equals 1 if qualifies at least in part for 0.18 tax savings. 0 76. enter 0.18. this is the marginal tax rate for those with taxable income falling in the range given by lines 59 and 73. 0.18 77. divide line 1 by (one minus line 76). this is maximum deductible ira for 0.18. $24,734 78. enter smaller of line 75 and (line 74 divided by line 77). this is the percentage of the maximum tax savings for those getting 0.18. 0.00% 79. enter one minus line 78. this is the percentage qualifying for 0.12 overlap if between 0% and 100%. 100.00% 80. enter 1 if line 71 equals 1; else 0. 0 81. multiply lines 77, 78, and 80. this amount qualifies to lower the taxable income using 0.18 tax rate. $0 82. multiply lines 75, 79, and 91. this overlap amount qualifies to lower the taxable income using 0.12 tax rate. $0 83. enter 0 if line 79 equals 0%; else 1. eliminate those that maxed out tax savings at 0.18. 1 84. enter smaller of line 83 and line 70. we have now taken out those who have maxed out to date. 0 85. multiply lines 76, 77, and 78. this is dollar tax savings if getting 0.18. $0 86. multiply lines 65 and 85. this adjusts for overlap and gives new tax savings to add in. $0 87. enter taxable income that is first taxed at 0.12. $0 88. multiply line 84 by the larger of (line 4 minus line 87) and 0. the taxable income for those who qualify in 0.12 bracket. $0 89. enter 0 if line 88 equals $0; else 1. equals 1 if qualifies at least in part for 0.12 tax savings. 0 90. enter 0.12. this is the marginal tax rate for those with taxable income falling in the range given by lines 73 and 90. 0.12 91. divide line 1 by (one minus line 90). this is maximum deductible ira for 0.12. $23,048 92. enter smaller of line 89 and (line 88 divided by line 91). this is the percentage of the maximum tax savings for those getting 0.12. 0.00% 93. enter one minus line 92. this is the percentage qualifying for 0% overlap if between 0% and 100%. 100.00% 392 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 maximum ira deduction of $20,282/(1–0.3) � $28,974 so that the total tax deduction is $62,321. while tc � 0.30 is our best guess, we still need further scrutiny. because our couple has a 3% annual increase in earnings compared to only a 1.5% annual increase in the tax brackets, they will be paying taxes at a lower marginal tax rate earlier in their working lives. thus, their deductible ira contribution over time can have an increasingly higher tc. when we test all 35 contribution years after adjusting each year for increases, we find that by the 2nd year, our couple gets 57% of its maximum contributions at the 0.30 tax savings level and by the 9th year, they get 100% at the 0.30 level. by the 34th year, they are getting all of their tax savings at 0.336. thus, it is highly unlikely that they would ever have to contribute to an ira and get only 0.18 savings on the dollar especially given that their optimal contribution is almost certainly less than their maximum contribution. we also know they are likely to get withdrawals with tax savings greater than 0.30. thus, there is evidence that a marginal tc of 0.30 is a minimum estimation of their true tax savings. finally, for situations where an optimal deductible ira is lower and the growth in earnings is greater than inflation, our couple should wait to contribute to a deductible ira to get the highest possible marginal tc. 5.4. computing future withdrawal cash flow values for ira having computed tc, the next step in the development of our rvd procedure is to supply future withdrawal cash flow values for iras. in table 5, we supply this information. we begin in panel a by computing the maximum deductible ira contribution for the current year of $28,974. maximum refers to the fact our couple would have to put all of their maximum roth ira investment of $20,282 in a deductible ira and then use the tax savings of 0.3($28,974) � $8,692. thus, they could invest a maximum of $20,282 � $8,692 � $28,974 in a deductible ira. as seen in panel a, if we add in the match/oi of $8,798, then we get $37,772. in panel b, we calculate the maximum future value for the roth ira (mfvroth) as $3,653,762 and it is not taxable. the future value for the deductible ira is taxable and so has a maximum before-tax future value deductible ira (mfvbtded) computed as table 4 (continued) 94. enter 1 if line 85 equals 0; else 0. 1 95. multiply lines 91, 92, and 94. this amount qualifies to lower the taxable income using 0.12 tax rate. $0 96. multiply lines 90, 91, and 92. this is dollar tax savings if getting 0.12. $0 97. multiply lines 79 and 96. this adjusts for overlap and gives new tax savings to add in. $0 98. add lines 16, 30, 44, 58, 72, 86, and 97. this is the total tax saving if maximum deductible used. $8,692 99. enter larger of 0.01 and addition of lines 12, 13, 25, 26, 39, 40, 53, 54, 67, 68, 81, 82, and 95. amount taxable deduction if maximum deductible ira taken. this amount times the optimal rvd allocation once determined will be added to line 3. $28,974 100. divide line 98 by line 99. tc � (tax savings from deductible ira contributions)/ (deductible ira contributions). 0.3000 393r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 5 definitions and equations to compute maximum withdrawal cash flows panel a. maximum annual contributions during working years: means for 2016–2050 macroth (maximum annual contribution roth ira) � $20,282 (computed in panel e of table 2) macbfded (maximum before-tax annual contribution deductible ira) � aacroth/(1�tc) � $20,282/(1�0.3) � $28,974 acbfmoi (before-tax annual contribution match/oi) � $8,798 (computed in panel e of table 2) macbfdedbfmoi (maximum before-tax annual contribution deductible ira and match/oi) � macbfded � acbfmoi � $28,974 � $8,798 � $37,772 panel b. maximum before-tax lump sum future values at retirement: means for end of 2050 mfvroth (maximum future value roth ira at retirement for yc � 35 and rc � 8.1993%) � macroth(fvafrc ,yc ) � macroth�(1 � rc)yc � 1 rc � � $20,282�(1 � 0.0811993)35 � 1 0.081993 � � $20,282(180.14799987) � $3,653,762 mfvbtded (maximum before-tax future value deductible ira) � macbfded(fvafrc ,yc ) � $28,974(180.14799987) � $5,219,608 fvbtmoi (before-tax future value match/oi) � acbfmoi(fvafrc ,yc ) � $8,798(180.14799987) � $1,584,942 mfvbtdedbtmoi (maximum before-tax future value deductible ira and match/oi) � mfvbtded � fvbtmoi � $5,219,608 � $1,584,942 � $6,804,550 panel c. maximum after-tax lump sum future values if tw � 0.30 at retirement: means for end of 2050 mfvatded (maximum after-tax future value deductible ira) � (1�tw)(mfvbtded) � (1�0.30)($5,219,608) � $3,653,762 fvatmoi (after-tax future value match/oi) � (1�tw)fvbtmoi � (1�0.30)$1,584,942 � $1,109,459 mfvatdedmoi (maximum after-tax future values for deductible ira and match/oi) � mfvatded � fvatmoi � $3,653,762 � $1,109,459 � $4,763,221 panel d. maximum before-tax annual withdrawals: means for 2051–2070 mawroth (max annual withdrawal from roth ira for yw � 20 and rw � 6.9743%) � mfvroth(1/pvafrw ,yw ) � fvroth�rw/�1 � 1 (1 � rw)yc��� $3,653,762�0.069743/�1 � 1 (1 � 0.069743)20�� � $3,653,762(0.094204462) � $344,201 mawbtded (maximum before-tax annual withdrawal deductible ira) � mfvbtded(1/pvafrw ,yw ) � $5,219,608(0.094204462) � $491,710 awbtmoi (before-tax annual withdrawal match/oi) � fvbtmoi(1/pvafrw ,yw ) � $1,584,942(0.094204462) � $149,309 mawbtdedmoi (maximum before-tax annual withdrawals for deductible ira and match/oi) � mawbtded � awbtmoi � $491,710 � $149,309 � $641,019 panel e. maximum after-tax annual withdrawals if tw � 0.30: means for 2051–2070 mawatded (maximum after-tax annual withdrawal deductible ira) � (1�0.3)mawbtded � (1�0.3)$491,710 � $344,201 awatmoi (after-tax annual withdrawal match/oi) � (1�0.3)awbtmoi � (1�0.3)$149,309 � $104,516 mawatdedmoi (maximum after-tax annual withdrawals for deductible ira and match/oi) � mawatded � awatmoi � $344,201 � $104,516 � $448,717 394 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 $5,219,608 in panel b and a maximum after-tax future value deductible ira (mfvatded) computed as $3,653,762 in panel c if tw � 0.30. thus, when tc � tw � 0.30, we see that mfvroth � mfvatded � $3,653,762. there are two conditions for this equality to hold. first, the tax savings from the deductible ira must be invested in the deductible ira along with the same amount invested in the roth ira. second, tc must equal tw. we provide a formal proof in appendix 1. with a roth ira, the irs collects its taxes upfront. with a deductible ira, the irs collect its taxes on the principal and earnings when withdrawn. in regards to the earnings on a deductible ira, the irs holds the right to get a proportion of the earnings and that proportion is tw. the future value of the annuity tax savings from the deductible ira explains the difference of $1,565,846 between the before-tax lump sum future values of the deductible ira of $5,219,608 and the roth ira of $3,653,762. to illustrate using our future value annuity factor of 180.14799987 in panel b, the lump sum future value of this annuity tax savings is 0.3($28,974)(180.14799987) � $1,565,846. this is the same value if the deductible ira is withdrawn with taxes paid at tw � 0.30 as the value of the taxes paid is 0.30($5,219,608) � $1,565,846.8 when tc � tw, the value of the tax savings is the same as the value of the taxes paid. panel d of table 5 provides the before-tax annual withdrawals for 20 years generated from the lump sum future values in panel c. the value of $149,309 for the before-tax annual withdrawal from the match/oi is important because it raises the taxable income increasing tw. when using a marginal analysis, the match/oi is assumed to have been withdrawn before diw. finally, panel e gives after-tax withdrawal values where we see that the maximum roth annual withdrawal of $344,201 in panel d is equal to the after-tax value of the maximum annual diw of $344,201 in panel e. once again, the equality results because of the two conditions described above. 6. procedure to determine the optimal rvd solution in this section, we introduce our withdrawal algorithm that computes the optimal rvd allocation. we present introductory illustrations in table 6 that use values given earlier such as our couple’s social security benefits (ssb) and tax deductions. in table 7, we determine our couple’s optimal rvd range. among the outcomes of our withdrawal algorithm are the average tw and the marginal tw. the latter is determined from lines that contain them, for example, line 27 has tw � 0.12, line 31 has tw � 0.18, and so forth, for every subsequent fourth line until we reach line 51 where tw � 0.471. 6.1. first discovery point in table 6, we begin filling in our withdrawal algorithm by inputting our couple’s ssb of $50,441 (line 1). for example 1, we input a diw of $0 (line 3) indicating all contributions were put in a roth ira. the percentage of ssb subject to taxes is 0% (line 21). while the zero taxes on ssb is one benefit of a roth ira, the marginal gain of zero (line 58) reflects a lost opportunity by not investing in a deductible ira. 395r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 6 withdrawal algorithm: determines first discovery point (dp1) example 1 example 2 example 3 1. enter social security benefit (found on form 1040, line 20a). $50,441 $50,441 $50,441 2. enter one-half of line 1. $25,221 $25,221 $25,221 3. enter total income on form 1040, line 22 minus ssb line 20a. (all withdrawals are diw; match/oi not yet considered.) $0.00 $46,396.83 $46,396.84 4. enter total of any exclusions/adjustments (typically not applicable so enter $0). $0 $0 $0 5. add lines 2, 3, and 4. $25,221 $71,617 $71,617 6. add lines 23 through 35 from form 1040 (these can lower your agi for most years; enter $0). $0 $0 $0 7. subtract line 6 from line 5. $25,221 $71,617 $71,617 8. enter $63,235 since married filing jointly. $63,235 $63,235 $63,235 9. subtract line 8 from line 7. if zero or less, enter $0. if line 9 is more than zero, go to line 10. $0 $8,382 $8,382 10. enter $23,713 since married filing jointly. $23,713 $23,713 $23,713 11. subtract line 10 from line 9. if zero or less, enter $0. $0 $0 $0 12. enter smaller of line 9 or line 10. $0 $8,382 $8,382 13. enter one-half of line 12. $0 $4,191 $4,191 14. enter smaller of line 2 or line 13. $0 $4,191 $4,191 15. multiply line 11 by 85% (.85). if line 11 is zero, enter $0. $0 $0 $0 16. add lines 14 and 15. $0 $4,191 $4,191 17. multiply line 1 by 85% (0.85). $42,875 $42,875 $42,875 18. enter smaller of line 16 or line 17. $0 $4,191 $4,191 19. enter amount from line 20 of the lump sum ss worksheet. not applicable so enter $0. $0 $0 $0 20. enter smaller of line 18 or line 19. this is the taxable ssb. $0 $4,191 $4,191 21. percent of taxable ssb subject to taxes. divide line 20 by line 1. put in percentage form. put in 0% if no ssb. 0% 8.3090% 8.3091% 22. multiply line 20 by line 56. (have to compute tw later in line 56.) this is taxes paid on ss if we use tw and if ssb are actually taxed. $0 $0 $503 23. subtract line 1 from line 20. this is the nontaxable ssb. $50,441 $46,250 $46,250 24. adjusted gross income (agi). add line 3 and line 20. $0 $50,588 $50,588 25. enter $50,588. this is the total tax deduction. $50,588 $50,588 $50,588 26. subtract line 25 from line 24. this is the annual taxable income during retirement withdrawal. $0.00 $0.00 $0.01 27. enter 0.12. this is the marginal tax rate for $36,656 of taxable income. 0.12 0.12 0.12 28. enter $36,656. this is maximum taxable income for 0.12 marginal tax rate. $36,656 $36,656 $36,656 29. enter smaller of line 26 or line 28. if zero or less, enter $0. this is applicable taxable income for 0.12 marginal tax rate. $0.00 $0.00 $0.01 30. multiply line 27 by line 29. taxes paid at 0.12 marginal tax rate. $0 $0 $0 31. enter 0.18. this is the marginal tax rate for the next $112,143 of taxable income. 0.18 0.18 0.18 32. enter $112,143. this is maximum taxable income for 0.18 marginal tax rate. $112,143 $112,143 $112,143 33. enter smaller of line 32 or (line 26 minus line 28). if zero or less, enter $0. this is applicable taxable income for 0.18 marginal tax rate. $0 $0 $0 396 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 6 (continued) example 1 example 2 example 3 34. multiply line 31 by line 33. taxes paid at 0.18 marginal tax rate. $0 $0 $0 35. enter 0.30. this is the marginal tax rate for $151,367 of taxable income. 0.30 0.30 0.30 36. enter $151,367. this is maximum taxable income for 0.30 marginal tax rate. $151,367 $151,367 $151,367 37. enter smaller of line 36 or (line 26 minus lines 28 and 32). if zero or less, enter $0. this is applicable taxable income for 0.30 marginal tax rate. $0 $0 $0 38. multiply line 35 by line 37. taxes paid at 0.30 marginal tax rate. $0 $0 $0 39. enter 0.336. this is the marginal tax rate for $157,197 of taxable income. 0.336 0.336 0.336 40. enter $157,197. this is maximum taxable income for 0.336 marginal tax rate. $157,197 $157,197 $157,197 41. enter smaller of line 40 or (line 26 minus lines 28, 32, and 36). if zero or less, enter $0. this is applicable taxable income for 0.336 marginal tax rate. $0 $0 $0 42. multiply line 39 by line 41. taxes paid at 0.336 marginal tax rate. $0 $0 $0 43. enter 0.396. this is the marginal tax rate for $359,449 of taxable income. 0.396 0.396 0.396 44. enter $359,449. this is maximum taxable income for 0.396 marginal tax rate. $359,449 $359,449 $359,449 45. enter smaller of line 44 or (line 26 minus lines 28, 32, 36, and 40). if zero or less, enter $0. this is applicable taxable income for 0.396 marginal tax rate. $0 $0 $0 46. multiply line 43 by line 45. taxes paid at 0.396 marginal tax rate. $0 $0 $0 47. enter 0.42. this is the marginal tax rate for $66,396 of taxable income. 0.42 0.42 0.42 48. enter $66,396. this is maximum taxable income for 0.42 marginal tax rate. $66,396 $66,396 $66,396 49. enter smaller of line 48 or (line 26 minus lines 28, 32, 36, 40, and 44). if zero or less, enter $0. this is applicable taxable income for 0.42 marginal tax rate. $0 $0 $0 50. multiply line 47 by line 49. taxes paid at 0.42 marginal tax rate. $0 $0 $0 51. enter 0.471. this is marginal tax rate for unlimited amount of taxable income. 0.471 0.471 0.471 52. enter $9,999,999 as proxy for unlimited for the maximum taxable income for 0.471 marginal tax rate. $9,999,999 $9,999,999 $9,999,999 53. enter smaller of line 52 or (line 26 minus lines 28, 32, 36, 40, 44, and 48). if zero or less, enter $0. this is applicable taxable income for 0.471 marginal tax rate. $0 $0 $0 54. multiply line 51 by line 53. taxes paid at 0.471 marginal tax rate. $0 $0 $0 55. add lines 30, 34, 38, 42, 46, 50, and 54. total taxes paid (on retirement withdrawal). $0 $0 $0 397r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 we input a diw of $46,396.83 for example 2. if there is taxable income, ssb will be taxed at a rate of 8.309% (line 21) implying the taxable ssb is 0.08309($50,441) � $4,191 (line 20). however, the tax deduction of $50,588 (line 25) is greater than $4,191 causing the taxable income to be $0 (line 26). thus, tw is zero. the latter is verified in line 29, where we see taxes are not paid at a 0.12 marginal tax rate. given a contribution tax rate of 0.30, the first tax rate differential is �t1 � (tc � tw) � (0.30 � 0) � 0.30. to the nearest dollar, we saw in table 1 that $46,397 was dp1. using (12) with $46,396.83 for dp1, we have: marginal gain � �t1dp1 � (0.30 � 0)$46,396.83 � $13,919.049 (line 58). in example 3, we see that to the nearest penny dp1 is $46,396.84 (line 3) because, at this point, the marginal tw goes from 0 to 0.12 as seen by $0.01 (line 26) that indicates taxable income is no longer zero. at the precise point of $46,396.84, the taxable income jumps to a marginal tw of 0.12 so that the next $36,656 (line 28) of taxable income would be taxed at tw � 0.12 with �t2 � (tc � tw) � (0.30 � 0.12) � 0.18. thus, the extra $0.01 in taxable income produces 0.18($0.01) � $0.02 in extra gain and we have a gain of $13,919.051 (line 58). the value of 9.44% (line 60) tells us what percentage of $46,396.84 (line 3) is of $491,710 (line 59) where the latter was given in table 5 as the maximum future value of the before-tax annual annuity withdrawal from a deductible ira. the percentage put in a roth ira would be 100% to 9.44% � 90.56%. multiplying 90.56% times the maximum roth ira of $344,201 (line 61) tells us that $311,723 (line 62) will be withdrawn from a roth ira. 6.2. subsequent discovery points in table 7, we do not repeat the instructions given in table 6 in the first column but only provide the line numbers of prior instructions given in table 6. in line 33 of table 7, we find that the taxable income for second marginal tax rate is $0.00 for example 1 and $0.01 for example 2. this indicates that the diw of $69,111.29 (line 3) in example 2 is dp2 to the nearest penny. this also tells us that the marginal tw of 0.18 (line 31) kicks in during the table 6 (continued) example 1 example 2 example 3 56. divide line 55 by line 26. this is average tw � taxes paid/taxable income. 0.00 0.00 0.12 57. enter 0.30. this is tc computed in table 4. 0.30 0.30 0.30 58. compute the marginal gain (using marginal tax rates) � �t1dp1 � �t2(dp2�dp1) � � � �t8(dp8�dp7) replacing dps and dropping components as prescribed by the five-step procedure. $0.000 $13,919.049 $13,919.051 59. enter $491,710. this is the maximum future value of the before-tax annual annuity withdrawal from deductible ira. $491,710 $491,710 $491,710 60. divide line 3 by line 59 and put in percentage form. percent withdrawn from maximum possible. 0.00% 9.44% 9.44% 61. enter $344,201. this is the maximum future value of the before-tax annual annuity withdrawal from roth ira. $344,201 $344,201 $344,201 62. subtract line 60 from 100% and multiply by line 61 to get the dollar amount of the roth ira. $344,201 $311,723 $311,723 398 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 7 withdrawal algorithm to determine tw, optimal rvd, and dp2, dp3 and dp4 line no. example 1 example 2 example 3 example 4 example 5 example 6 example 7 line 1 $50,441 $50,441 $50,441 $50,441 $50,441 $50,441 $50,441 line 2 $25,221 $25,221 $25,221 $25,221 $25,221 $25,221 $25,221 line 3 $69,111.28 $69,111.29 $98,220 $156,512.00 $156,512.16 $307,879.00 $307,879.16 line 4 $0 $0 $0 $0 $0 $0 $0 line 5 $94,332 $94,332 $123,441 $181,733 $181,733 $333,100 $333,100 line 6 $0 $0 $0 $0 $0 $0 $0 line 7 $94,332 $94,332 $123,441 $181,733 $181,733 $333,100 $333,100 line 8 $63,235 $63,235 $63,235 $63,235 $63,235 $63,235 $63,235 line 9 $31,097 $31,097 $60,206 $118,498 $118,498 $269,865 $269,865 line 10 $23,713 $23,713 $23,713 $23,713 $23,713 $23,713 $23,713 line 11 $7,384 $7,384 $36,493 $94,785 $94,785 $246,152 $246,152 line 12 $23,713 $23,713 $23,713 $23,713 $23,713 $23,713 $23,713 line 13 $11,857 $11,857 $11,857 $11,857 $11,857 $11,857 $11,857 line 14 $11,857 $11,857 $11,857 $11,857 $11,857 $11,857 $11,857 line 15 $6,276 $6,276 $31,019 $80,567 $80,567 $209,229 $209,229 line 16 $18,133 $18,133 $42,875 $92,423 $92,423 $221,085 $221,085 line 17 $42,875 $42,875 $42,875 $42,875 $42,875 $42,875 $42,875 line 18 $18,133 $18,133 $42,875 $42,875 $42,875 $42,875 $42,875 line 19 $0 $0 $0 $0 $0 $0 $0 line 20 $18,133 $18,133 $42,875 $42,875 $42,875 $42,875 $42,875 line 21 35.95% 35.95% 85.00% 85.000% 85.00% 85.000% 85.00% line 22 $2,176 $2,176 $6,676 $7,084 $7,084 $9,998 $9,998 line 23 $32,308 $32,308 $7,566 $7,566 $7,566 $7,566 $7,566 line 24 $87,244 $87,244 $141,095 $199,387 $199,387 $350,754 $350,754 line 25 $50,588 $50,588 $50,588 $50,588 $50,588 $50,588 $50,588 line 26 $36,656 $36,656 $90,507 $148,799 $148,799 $300,166 $300,166 line 27 0.12 0.12 0.12 0.12 0.12 0.12 0.12 line 28 $36,656 $36,656 $36,656 $36,656 $36,656 $36,656 $36,656 line 29 $36,656 $36,656 $36,656 $36,656 $36,656 $36,656 $36,656 line 30 $4,399 $4,399 $4,399 $4,399 $4,399 $4,399 $4,399 line 31 0.18 0.18 0.18 0.18 0.18 0.18 0.18 line 32 $112,143 $112,143 $112,143 $112,143 $112,143 $112,143 $112,143 line 33 $0.00 $0.01 $53,851 $112,143 $112,143 $112,143 $112,143 line 34 $0 $0 $9,693 $20,186 $20,186 $20,186 $20,186 line 35 0.30 0.30 0.30 0.30 0.30 0.30 0.30 line 36 $151,367 $151,367 $151,367 $151,367 $151,367 $151,367 $151,367 line 37 $0 $0 $0 $0.00 $0.01 $151,367 $151,367 line 38 $0 $0 $0 $0 $0 $45,410 $45,410 line 39 0.336 0.336 0.336 0.336 0.336 0.336 0.336 line 40 $157,197 $157,197 $157,197 $157,197 $157,197 $157,197 $157,197 line 41 $0 $0 $0 $0 $0 $0.00 $0.01 line 42 $0 $0 $0 $0 $0 $0.00 $0.00 line 43 0.396 0.396 0.396 0.396 0.396 0.396 0.396 line 44 $359,449 $359,449 $359,449 $359,449 $359,449 $359,449 $359,449 line 45 $0 $0 $0 $0 $0 $0 $0 line 46 $0 $0 $0 $0 $0 $0 $0 line 47 0.42 0.42 0.42 0.42 0.42 0.42 0.42 line 48 $66,396 $66,396 $66,396 $66,396 $66,396 $66,396 $66,396 line 49 $0 $0 $0 $0 $0 $0 $0 line 50 $0 $0 $0 $0 $0 $0 $0 line 51 0.471 0.471 0.471 0.471 0.471 0.471 0.471 line 52 $9,999,999 $9,999,999 $9,999,999 $9,999,999 $9,999,999 $9,999,999 $9,999,999 line 53 $0 $0 $0 $0 $0 $0 $0 399r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 69,111th withdrawal dollar causing �t3 � (tc � tw) � 0.30 � 0.18 � 0.12 to also kick in. using (12), we have: marginal gain � �t1dp1 � �t2(dp2-dp1) � (0.30 � 0)$46,397 � (0.30 � 0.12)($69,111 � $46,397) � $18,008 (line 58). in example 3 of table 7, we input diw � $98,220 (line 3). doing this gives 85% (line 21). if we were to input $98,219 (line 3), we would get 84.999% (line 21), while withdrawals greater than $98,220 would still get 85% as this is the maximum percentage at which ssb can be taxed. for a withdrawal of $98,220, taxes paid on ssb is 0.85($50,441) � $42,875 (line 20). the first $36,656 of taxable income (line 28) is taxed at 0.12 and creates 0.12($36,656) � $4,399 in taxes (line 30). the next $53,851 (line 33) is taxed at 0.18 (line 31) and creates 0.18($53,851) � $9,693 in taxes (line 34). because the taxable income range for 0.18 is $112,143 (line 32), we see that there will be no taxable income beyond the marginal tax rate of 0.18 for a withdrawal of $98,220. the total taxes paid are $4,399 � $9,693 � $14,092 (line 55). dividing the total taxes paid by the taxable income of $90,507 (line 26) yields an average tw of $14,092/$90,507 � 0.1557 (line 56). using (12) and noting that dp2�$98,220�dp3, we substitute $98,220 for dp3 (as described in section 4.4 in our five-step procedure), to get marginal gain � �t1dp1 � �t2(dp2-dp1) � �t3(withdrawal-dp2) � (0.30 � 0)$46,397 � (0.30 � 0.12)($69,111 � $46,397) � (0.30 � 0.18)($98,220 � $69,111) � $21,501 (line 58). the percentage allocated to a deductible ira is 19.98% (line 60). this means our couple puts 81.02% in a roth generating an annuity withdrawal from a roth ira of $275,446 (line 62). suppose ssb is zero. for this situation, the taxable ssb falls from $42,875 to $0 and the taxable income falls to $90,507 � $42,875 � $47,632. our new dp1 is $98,220 � $47,632 � $50,588 (that is the slack) and gets a marginal tw of 0 yielding marginal gain1 � (0.30 � 0)($50,588) � $15,176.40. the remaining withdrawal is $98,220 � $50,588 � $42,875. the first $36,656 is taxed at 0.12. because a zero ssb has eliminated the peculiarities of the way withdrawals are taxed, our new dp2 is simply dp1 plus the marginal tax bracket range for 0.12. the latter tax bracket range is $36,656 as can be seen from row four of table 2. thus, dp2 � $50,588 � $36,656 � $87,244. our second incremental gain is marginal gain2 � (0.30 � 0.12)$36,656 � $6,598.08. for the remaining $42,875 � $36,656 � $6,219 that is taxed at 0.18, we get marginal gain3 � (0.30 � 0.18)$6,219 � $746.28. the sum of these marginal gains is $22,521.9 a gain of $22,521 is greater than $21,501 in example 3 showing that the effect of ssb imposes a cost of $22,521 � $21,501 � $1,020. table 7 (continued) line no. example 1 example 2 example 3 example 4 example 5 example 6 example 7 line 54 $0 $0 $0 $0 $0 $0 $0 line 55 $4,399 $4,399 $14,092 $24,584 $24,584 $69,995 $69,995 line 56 0.1200 0.1200 0.1557 0.1652 0.1652 0.2332 0.2332 line 57 0.30 0.30 0.30 0.30 0.30 0.30 0.30 line 58 $18,008 $18,008 $21,501 $28,495.62 $28,495.62 $28,495.62 $28,495.62 line 59 $491,710 $491,710 $491,710 $491,710 $491,710 $491,710 $491,710 line 60 14.055% 14.055% 19.98% 31.83% 31.83% 62.61% 62.61% line 61 $344,201 $344,201 $344,201 $344,201 $344,201 $344,201 $344,201 line 62 $295,823 $295,823 $275,446 $234,641 $234,641 $128,683 $128,683 400 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 suppose ssb and tax deduction are both zero. without ssb and tax deductions, and keeping match/oi at zero, all dps would be easily identifiable because dp1 � 0 and each subsequent dp would be the amount of a taxable income bracket range. for example, dp2 � $36,656, dp3 � $112,143, and so forth. to illustrate, if we input $0 for ssb (line 1) and $0 for the tax deduction (line 25) while keeping our current assumption of match/oi � $0, we get: marginal gain � �t1(dp1) � �t2(dp2-dp1) � �t3(withdrawal-dp2) � (0.30 � 0.0)$0 � (0.30 � 0.12)$36,656 � (0.30 � 0.18)($98,220 � $36,656) � $13,986. without the positive effect of a tax deduction, we see that the gain is significantly lowered. as seen in example 3, the ssb’s taxable income is $42,875. on a marginal basis where ssb is taxed first and diw is large enough to cover the tax deduction, the impact is 0.12($42,875) � $5,145 in taxes given that ssb would not be taxed if the withdrawal was zero because all ira funds were put in a roth ira. thus, on a marginal comparative basis, one could argue that the marginal gain is not $21,501 but $21,501 � $5,145 � $16,356. by not being taxable, the roth ira does not impose this opportunity cost of $5,145. however, the roth ira also does not have the gross advantage of $21,501 given by the deductible ira when using our marginal gain formula. 6.3. identifying the maximum marginal gain range in example 4, we input diw � $156,512 (line 3) and taxable income � $0 (line 37). dp3 is achieved during the withdrawal of the $156,512th dollar as seen in example 5 when diw is increased by $0.16 to $156,512.16 (line 3) because at this point the taxable income is no longer $0 but $0.01 (line 37). using (12), we have: marginal gain � �t1(dp1) � �t2(dp2dp1) � �t3(dp3-dp2) � (0.30–0.0)$46,397 � (0.30 � 0.12)($69,111 � $46,397) � (0.30 � 0.18)($156,512 � $69,111) � $28,495.62 (line 58). this is the same as the marginal gain for a $156,512 because $156,512 is a discovery point where the next �t is zero as we have �t4 � (tc � tw) � (0.30 � 0.30) � 0. we can also see that the applicable taxable income for 0.30 marginal tax rate (line 35) is the same as tc. to illustrate to the nearest penny, we would add the following component to our computation: �t4(withdrawal-dp3) � (0.30 � 0.30)($156,512.16 � $156,512) � 0($0.16) � $0. thus, the marginal gain remains at $28,495.62 until we reach dp4 at which point the gain will fall because �t5 � 0.336 � 0.300 � �0.036. the lower bound optimal rvd allocation involves a diw of $156,512. the optimal diw as a percentage of the maximum diw (called odi%) is 31.83% (line 60). it is our lower bound odi% because our maximum marginal gain first occurs at 31.83%. noting that our maximum marginal gain is a 20-year annuity, we can discount its total value to get this value in today’s dollar. in doing this, we get a lifetime wealth gain amounting to $179,637 or about $180,000. adjusting for our inflation rate of 1.5%, we get $302,487 in future value dollars at the time of retirement. what happens if an investor withdraws enough to jump to the tax rate of 0.336? this is illustrated in examples 6 and 7 of table 7 where there begins a decrease in the marginal gain as the withdrawal increases past our next discovery point of dp4 � $307,879 to the nearest dollar. the applicable taxable income for the marginal tax rate of 0.336 (line 39) is $0 (line 41) in example 6 but becomes positive at $0.01 (line 41) in example 7 when diw � $307,879.16 indicating we have reached dp4 to the nearest penny. if we were to increase 401r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 diw to $307,880, we would find a fall in the gain on line 58 from $28,495.62 to $28,495.59. the three cents fall is explain by the additional component in our gain formula of �t5(withdrawal-dp4) � (0.30 � 0.336)($307,880 � $307,879.16) � �$0.03. the fall continues with greater withdrawals since our couple’s contribution to a deductible ira gets only tc � 0.30 and the marginal tw beginning with dp4 is 0.336 and covers the dp5-dp4 range. finally, as seen in table 7, the odi% is 62.61% (line 60) and this is our upper bound odi%. 6.4. our couple’s optimal rvd outcomes as we just saw, marginal gains are maximized between dp3 and dp4. the incremental increase in gain is zero from the 156,512th withdrawal dollar to the 307,879th withdrawal dollar. this is because tc � tw � 0.30 for this range. during the withdrawal of 307,879th dollar, our couple jumps to a higher tax bracket and begin to be penalized since the marginal tw is greater than tc. the midpoint of the optimal withdrawal range is ($156,512 � $307,879)/2 � $232,196. on a marginal gain basis, there is a broad odi% range from 31.83% to 62.61%. unless our couple has factors other than the tax rate differential to consider that would favor either a deductible ira or a roth ira, an argument can be made that the safest rvd allocation would be odi% � (31.83% � 62.61%)/2 � 47.22%. by choosing this midpoint percentage, we allow leeway on both sides if any inputs are less than precise. thus, a recommendation would be 47.22% of the annual maximum deductible ira contribution and 52.78% of the annual maximum roth ira contribution. given the maximum mean annual contribution of $28,974 for a deductible ira as computed in table 5, we have: 0.4722214($28,974) � $13,682 allocated to a deductible ira. similarly, given the maximum roth ira contribution of $20,282, we have 0.5277786($20,282) � $10,704 allocated to a roth ira. the corresponding optimal annual withdrawals generated at retirement would be 0.4722214($491,710) � $232,196 from a deductible ira and 0.5277786($344,201) � $181,662 from a roth ira. 7. plotting the optimal rvd range in this section, we plot the annual gains and annual withdrawals. the extent to which the annual match/oi influence the gain depends on assumptions about when the match/oi is taxed and how much of it is taxed.10 fig. 1 assumes diw is withdrawn first and the match/oi is withdrawn last. fig. 2 assumes the match is withdrawn first while oi is withdrawn last. both figures show average and marginal gains. 7.1. an optimal rvd allocation without the match/oi for fig. 1, we choose the following 18 annual withdrawals: $0 (all roth), $46,397 (dp1), $69,111 (dp2), $112,812, $156,512 (dp3), $194,354, $232,196 (midpoint of optimal range), $270,037, $307,879 (dp4), $347,178, $386,478, $425,777, $465,076 (dp5), $491,710 (maximum diw), $529,037, $566,365 (maximum diw plus match), $603,692, and $641,019 (maximum diw plus match/oi). fig. 1 has a flat optimal range covering diws from 402 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 $156,512 (dp3) to $307,879 (dp4) that maximize ira retirement wealth by providing a marginal gain of $28,496. the midpoint of this optimal range is $232,196. as seen in fig. 1, the average gain for this range is from $21,095 to $20,571 with the average gain falling throughout the range. the maximum average gain of $21,095 occurs at dp3. after the maximum diw of $491,710 is reached, fig. 1 shows further gain does not occur because the assumption is that the match/oi is withdrawn last with no impact on the gain. fig. 1 reflects this latter assumption by having a flat range at the end where the marginal gain is fixed at $20,280 once the maximum diw of $491,710 is reached. in comparing the average and marginal gains, we see they are equivalent up to dp1 � $46,397. this is because the taxable income is zero before reaching dp1 so the average and marginal tw are both zero. we can also notice the average gain reaches a relative maximum around dp1 with a steady decline after this point until dp2 � $69,111 is reached. we fig. 1. marginal gain (dotted line) and average gain as a function of withdrawals (in units of $1,000)–without match and other investments (oi). 403r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 attribute this relative maximum to the interplay of an increasing tax on ssb in combination with the progressive nature of the u.s. tax system. using the average tw renders an average gain equal to or proportional to the marginal gain for all points except the short withdrawal span where the relative maximum occurs. 7.2. the optimal rvd allocation with only the match considered as seen in table 5, the match/oi of $8,798 added to the maximum deductible ira contribution of $28,974 increases the total annual ira contribution to $37,772. as reported there, the maximum before-tax annuity withdrawal increases from $491,710 without the match/oi to $641,019 with the match/oi, which is an increase of $149,309. because the match and oi are equal, the match is half of the $149,309 or $75,654.50. in fig. 2, the match of $75,654.50 is withdrawn first and absorbs lower tw values. in fig. 2, the withdrawals are somewhat different from fig. 1 because the first withdrawal of $74,654.50 is the match where dp1 and dp2 are not possible as they involve amounts below the match. other than dp1 and dp2, we keep all other key withdrawals the same for both figures. by the time the match is withdrawn, one is already at the 0.18 marginal tax rate and it will stay this way until there is a jump to the 0.30 tax bracket at which point the marginal gain cannot increase. for the match withdrawal of $74,654.50, the taxable income is $46,911. thus, the match covers the $36,656 tax range for 0.12 and causes us to be in the 0.18 tax rate range for $46,911 � $36,656 � $10,255 withdrawal dollars. thus, by the time the match is withdrawn, we have �t3 � 0.12. using (12) and the beginning of the optimal range at dp3, we have marginal gain � �t3(dp3-withdrawal) � (0.30 � 0.18)($156,512 � $74,655) � (0.12)($81,857) � $9,823. the lower marginal gains in fig. 2 compared with fig. 1 is because the match is now assumed to absorb the lower tw values so that the first $74,655 withdrawn produces no gain because it does not get a tax savings during the contribution years. however, the match is valuable for two reasons. first, it was given free by the employer and no taxes were paid by the employee. second, by assuming the match absorbs the lower tax rate brackets, the match is more valuable on an after-tax basis than other subsequent taxable withdrawals because it causes less to be paid in taxes. thus, the value of the match is enhanced by having a greater after-tax value. if this enhancement is a marginal benefit, then it is possible that $9,823 underestimates the gain and we need to analyze a situation where neither the match nor diw absorbs the lowest tax rates. this particular situation requires investigating the average tw where absorption for any withdrawal does not occur first but all withdrawals are treated as equal. in our investigation, we find total taxes paid on the match are $6,246 and the taxable income is $46,911 by the time the first dollar of diw kicks in. thus, we get $6,246/ $46,911 � 0.1331 for the average tw. at the midpoint of our optimal gain (that is the same for both figures), tw is 0.2701 but the match has not been taxed at any tw above 0.1331. in conclusion, the maximum marginal gain of $9,823 only holds if we disregard any positive marginal effect the match usurps from the marginal gain by absorbing the lower tw values. unlike fig. 1, fig. 2 reveals that the average gain values are now greater than the marginal gain values. by considering the match, there is still a defined optimal marginal range in fig. 2 as was found in fig. 1. the same optimal marginal range of withdrawal dollars are found 404 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 except they now include the match so that in terms of the deductible ira dollars, the range is $81,858 � $233,225. for both figures, we find the marginal gain values tapering off rapidly before the optimal range. after the range ends, values gradually taper off. using the average tw values, we can also find flat ranges that contain the maximum average gain. the peak of $15,583 for the average gain in fig. 2 occurs where diw is $307,879, which is dp4. of importance, even when using the average tw, we still find there can be a margin of error because of a somewhat flat range around its optimal average gain of $15,583. in summary, by considering the match and focusing on a marginal analysis, the optimal midpoint withdrawal of $232,196 in fig. 1 did not change in fig. 2, while odi% fell from $232,196/$491,710 � 0.4722 or 47.22% to ($232,196 � $74,655)/$491,710 � 0.3204 or 32.04%. thus, the match causes a fall that is about one-third of fig. 1’s odi%. when the fig. 2. marginal gain (dotted line) and average gain as a function of withdrawals (in units of $1,000)–with match and without other investments (oi). 405r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 match is withdrawn before the deductible ira, we find a decrease in the marginal gain of $28,496 � $9,823 � $18,673. this represents a large fall of about two-thirds from that found in fig. 1. although not shown, we repeated the figures when assuming both the match and oi are withdrawn first and taxed at the ordinary rate. as expected, we still found the same flatness with the gains and odi% both lowered. 8. scenario analysis results when agis, returns, withdrawal years, and match change in the prior section, we found the optimal rvd allocation using the average gain was within the optimal range given by the marginal gain. furthermore, the two gains have correlation coefficients of 0.95 and 0.865 for figs. 1 and 2, respectively. thus, the average gain parallels the marginal gain. given this knowledge and the ease in computing the average gain, we use it for our scenario analysis that requires a large number of computations to make general conclusions about how key variables influence the optimal rvd allocation. whereas many scenario and outcome variables could be chosen, for brevity’s sake, we limit the number of variables. regardless, these illustrations demonstrate how our rvd procedure can help financial advisors provide rvd guidance to their clientele. for our scenario analysis, we will change three variables that we call scenario variables. they are adjusted gross income (agi), nominal stock returns (returns), and number of withdrawal years during retirement (years). agi changes in increments of $30,000, returns change in increments of 2% (nonstock premium remains at 3.5%), and years change in increments of five years. we investigate 10 other variables for outcome changes when our scenario variables change. we call these variables by the name of outcome variables and their abbreviations and definitions are: mdibt � maximum annual before-tax diw mri � maximum annual roth ira withdrawal match � employer’s salary match tc � average contribution tax rate tw � average withdrawal tax rate odibt � optimal annual before-tax diw gain � (tc � tw)odibt (gain is the maximum gain) odi% is odibt/mdibt and put in percentage form odiat � optimal annual after-tax diw � (1 � tw)odibt ori � optimal annual roth ira withdrawal each time a scenario variable changes, we identify the maximum gain from all withdrawals. we report the values for all other outcome variables that occur at the maximum point and report them in table 8. in panels a, b, and c of table 8, we assume no match/oi. in panels d, e, and f, we allow a 3% salary match and assume no oi. the only changes in variables from those used previously are the following. agi is $140,000 (unless used as a scenario 406 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 8 scenario analysis panel a. without match/oi: agis range from $50,000–$260,000 with incremental changes of $30,000 agi mdibt mri match tc tw odibt gain odi% odiat ori $50 $143 $117 $0 0.1800 0.0000 $48 $8.59 33.33% $48 $78 $80 $240 $188 $0 0.2174 0.0000 $40 $8.70 16.67% $40 $157 $110 $369 $258 $0 0.3000 0.1875 $184 $20.76 50.00% $150 $129 $140 $470 $329 $0 0.3000 0.2273 $274 $19.92 58.33% $212 $137 $170 $601 $399 $0 0.3360 0.2607 $401 $30.16 66.67% $296 $133 $200 $707 $470 $0 0.3360 0.2643 $412 $29.58 58.33% $303 $196 $230 $870 $540 $0 0.3792 0.2701 $435 $47.40 50.00% $317 $270 $260 $1,011 $610 $0 0.3960 0.3221 $758 $56.02 75.00% $514 $153 $155 $551 $364 $0 0.3056 0.1915 $319 $27.64 51.04% $235 $157 $73 $303 $173 $0 0.0747 0.1242 $237 $17.05 18.60% $158 $56 n.a. 1.00 1.00 n.a. 0.96 0.91 0.95 0.96 0.77 0.96 0.66 panel b. without match/oi: nominal rates of return from 4%–18% with incremental changes of 2% returns mdibt mri match tc tw odibt gain odi% odiat ori 4% $142 $100 $0 0.3000 0.1647 $142 $19.24 100.0% $119 $0 6% $244 $171 $0 0.3000 0.2184 $244 $19.90 100.0% $191 $0 8% $423 $296 $0 0.3000 0.2294 $282 $19.92 66.67% $217 $99 10% $744 $519 $0 0.3023 0.2197 $248 $20.46 33.33% $193 $346 12% $1,320 $915 $0 0.3067 0.2321 $293 $21.89 22.22% $225 $712 14% $2,351 $1,620 $0 0.3108 0.2274 $274 $22.89 11.67% $212 $1,431 16% $4,176 $2,869 $0 0.3130 0.2285 $278 $23.53 6.67% $215 $2,678 18% $7,396 $5,073 $0 0.3142 0.2365 $308 $23.95 4.17% $235 $4,861 11% $2,100 $1,445 $0 0.3059 0.2196 $259 $21.47 43.09% $201 $1,266 4.9% $2,534 $1,737 $0 0.0061 0.0229 $52 $1.83 40.34% $36 $1,718 n.a. 0.88 0.88 n.a. 0.96 0.71 0.77 0.98 �0.95 0.76 0.88 panel c. without match/oi: withdrawal years from 5 to 40 years with incremental changes of 5 years years mdibt mri match tc tw odibt gain odi% odiat ori 5 $1,228 $850 $0 0.3077 0.2134 $205 $19.30 16.67% $161 $709 10 $711 $496 $0 0.3018 0.2222 $237 $18.85 33.33% $184 $331 15 $546 $382 $0 0.3000 0.2299 $273 $19.16 50.00% $210 $191 20 $470 $329 $0 0.3000 0.2273 $274 $19.92 58.33% $212 $137 25 $427 $299 $0 0.3000 0.2272 $284 $20.71 66.67% $220 $100 30 $401 $280 $0 0.3000 0.2283 $300 $21.56 75.00% $232 $70 35 $384 $269 $0 0.3000 0.2298 $320 $22.44 83.33% $246 $45 40 $373 $261 $0 0.3000 0.2316 $342 $23.38 91.67% $263 $22 22.5 $567 $396 $0 0.3012 0.2262 $279 $20.67 59.38% $216 $200 12.2 $289 $199 $0 0.0027 0.0059 $44 $1.66 25.37% $33 $228 n.a. �0.82 �0.82 n.a. �0.68 0.80 0.98 0.96 0.99 0.98 �0.85 panel d. with match and without oi: agis range from $50,000–$260,000 with incremental changes of $30,000 agi mdibt mri match tc tw odibt gain odi% odiat ori $50 $143 $117 $35 0.1800 0.1200 $36 $2.15 25.00% $31 $88 $80 $232 $188 $56 0.1894 0.1647 $97 $2.39 41.67% $81 $110 $110 $369 $258 $77 0.3000 0.2387 $246 $15.08 66.67% $187 $86 407r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 table 8 (continued) agi mdibt mri match tc tw odibt gain odi% odiat ori $140 $470 $329 $99 0.3000 0.2323 $196 $13.25 41.67% $150 $192 $170 $601 $399 $120 0.3360 0.2642 $301 $21.59 50.00% $221 $200 $200 $707 $470 $141 0.3360 0.2679 $295 $20.05 41.67% $216 $274 $230 $863 $540 $162 0.3747 0.3120 $504 $31.57 58.33% $347 $225 $260 $1,011 $610 $183 0.3960 0.3325 $674 $42.79 66.67% $450 $203 $155 $549 $364 $109 0.3015 0.2415 $293 $18.61 48.96% $210 $172 $73 $303 $173 $52 0.0793 0.0709 $209 $13.83 14.39% $136 $69 n.a. 1.00 1.00 1.00 0.95 0.96 0.94 0.96 0.62 0.95 0.82 panel e. with match and without oi: nominal rates of return from 4%–0.18 with incremental changes of 2% returns mdibt mri match tc tw odibt gain odi% odiat ori 4% $142 $100 $30 0.3000 0.1848 $142 $16.39 100.0% $116 $0 6% $244 $171 $51 0.3000 0.2324 $244 $16.47 100.0% $187 $0 8% $423 $296 $89 0.3000 0.2249 $176 $13.24 41.67% $137 $173 10% $744 $519 $156 0.3023 0.2599 $248 $10.50 33.33% $183 $346 12% $1,318 $915 $275 0.3057 0.2785 $221 $5.98 16.67% $159 $763 14% $2,360 $1,620 $486 0.3136 0.2958 $94 $1.67 3.98% $66 $1,555 16% $4,190 $2,869 $861 0.3153 0.3297 $0 $0.00 0.00% $0 $2,869 18% $7,412 $5,073 $1,522 0.3156 0.3903 $0 $0.00 0.00% $0 $5,073 11% $2,104 $1,445 $434 0.3066 0.2745 $141 $8.03 36.96% $106 $1,347 4.9% $2,541 $1,737 $521 0.0071 0.0648 $101 $7.05 41.77% $76 $1,796 n.a. 0.88 0.88 1.00 0.94 0.97 �0.72 �0.98 �0.93 �0.76 0.88 panel f. with match and without oi: withdrawal years from 5 to 40 years with incremental changes of 5 years years mdibt mri match tc tw odibt gain odi% odiat ori 5 $1,228 $850 $255 0.3077 0.2834 $205 $4.97 16.67% $147 $709 10 $714 $496 $149 0.3047 0.2492 $178 $9.91 25.00% $134 $372 15 $548 $382 $115 0.3018 0.2368 $183 $11.87 33.33% $139 $255 20 $470 $329 $99 0.3000 0.2323 $196 $13.25 41.67% $150 $192 25 $427 $299 $90 0.3000 0.2316 $213 $14.59 50.00% $164 $149 30 $401 $280 $84 0.3000 0.2321 $234 $15.86 58.33% $179 $117 35 $384 $269 $81 0.3000 0.2333 $256 $17.06 66.67% $196 $90 40 $373 $261 $78 0.3000 0.2354 $280 $18.05 75.00% $214 $65 22.5 $568 $396 $119 0.3018 0.2418 $218 $13.20 45.83% $165 $244 12.2 $290 $199 $60 0.0029 0.0178 $36 $4.27 20.41% $29 $212 n.a. �0.82 �0.82 1.00 �0.83 �0.70 0.88 0.96 1.00 0.93 �0.87 note: the first three panels provide scenario results when the match/oi has no impact. the last three panels give results when oi has no impact. the scenario variables are in the first columns for each panel. the outcome variables (with dollars values expressed in units of 1,000) are in the last 10 columns. the values reported for “gain” are maximum value and so the values reported for the other outcome variables occur at that maximum point. panels a and d vary the adjusted gross income (agi); panels b and e modify the nominal stock return rate (returns); and, panels c and f alter the number of withdrawal years (years). the following abbreviation are used: mdibt � maximum annual before-tax diw; mri � maximum annual roth ira withdrawal; odibt � optimal annual before-tax diw; gain � (tc�tw)odibt; odi% � odibt/ mdibt and put in percentage form; odiat � optimal annual after-tax diw � (1�tw)odibt; ori � optimal annual roth ira withdrawal. the last three rows of each panel, respectively, report averages for the prior eight rows, standard deviations for the prior eight rows, and pearson correlation coefficients with the panel’s scenario variable. 408 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 variable), the salary is now the same as the agi, and ssb is adjusted in the systematic fashion described in section 5.1 each time agi is changed. when approximating optimal rvd values for our scenario analysis in panels a, b and c, we assume withdrawals from zero to the maximum before-tax diw. the eleven other withdrawals are chosen so that all withdrawals are equidistant from one another. to test if the maximum gain is properly identified, we allow (if necessary) another eleven smaller withdrawals around the initial maximum gain with shorter equidistances between withdrawals. this process is repeated as many times as needed until the maximum gain and optimal rvd can be suitably estimated. for panels d, e, and f, the first withdrawal is the match and the last withdrawal is the maximum diw plus the match withdrawal. the 10 other withdrawals are chosen so they are equidistant from one another. as described above, we repeat the process (if necessary) by using smaller equidistance withdrawals until the maximum gain can be correctly estimated. while our scenario analysis procedure uses approximations, we believe they are accurate and serve our purpose, which is to find general relations between scenario and outcome variables. finally, the last three lines for all panels for table 8 provide additional statistics. the first two rows include the average and standard deviation for the scenario and outcome variables based on the eight prior rows. the last row in each panel gives pearson correlation coefficients between the panel’s scenario variable and each of the 10 outcome variables. 8.1. scenario analysis without the match/oi considered from the information in the first three panels in table 8, we offer the following conclusions. first, panel a reveals outcome variables increase because agi increases as all 10 correlation coefficient in the last row are positive. the two variables that have lower positive correlations are the optimal percentage of the maximum diw (odi%) and the optimal contribution to a roth ira (ori). odi% takes a nosedive from 33.33% to 16.67% when agi goes from $50,000 to $80,000. odi% then climbs until it reaches an agi of $170,000 and then it falls for two consecutive agis of $200,000 and $230,000 before rising to its highest percentage. we can attempt to explain this roller coaster pattern. some investors will manage to barely get their tax savings in a higher tax rate bracket and barely attain their highest marginal tc for dollars they contribute to a deductible ira even though their agis are low compared with other investors who also get the same tc. because they have lower agis, they can contribute less and possibly withdraw less. with a lower tw values because of less withdrawals, they achieve larger tax rate differentials (�ts) thereby obtaining larger gains even though they have lower odi% values. this is seen in the fact the correlation between �t and odi% is �0.95 for this panel. we can see in places that when a higher odi% results, then a lower ori occurs. examples of the latter are the rows where odi% are 50.00%, 66.67% and 75.00% and the ori takes a dip even though the ori trend is increasing as agi increases. second, as seen in panel b, we find aggressive investors, who invest high proportions in equity (95% in the portfolio mix we use), will want less of their ira contribution to go to a deductible ira as odi% falls from 100% to 4.17%. this means the percentage invested in a roth ira goes from 0% to 95.83% as returns increase. consider the returns from 8% 409r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 to 10%. odi% falls in half from 66.37% (for an 8% return) to 33.33% (for a 10% return). thus, just a 2% increase in returns from 8% to 10% means one’s investment in a roth ira would double. unlike panel a where the gain increases from about $9,000 to over $56,000 as the agi increases, we find that the gain in panel b only increases from about $19,000 to $24,000 when returns increase. we conclude agi (and the amount contributed to an ira) is a more important factor than returns in getting larger maximum gains. in addition, unlike panel a where the ori only went from $78,000 to $153,000, panel b shows that ori went from $0 to about $4,861,000 (with $712,000 if the returns are 12%). thus, returns have a much greater impact on what is put in a roth ira compared with the agi. third, from panel c, we illustrate how investors with smaller years should be putting more of their ira contribution in a roth ira. those who retire early and/or expect to live long should be placing relatively more of their ira contributions in a deductible ira. the reason is obvious as spreading out withdrawals over a longer period can create lower tw values and larger �t values leading to greater maximum gains. as can be seen from comparing the tc and tw values in panel c, �t decreases from 9.43% to 6.84% as years increase from five to forty. finally, we see high positive correlation coefficients between years with gain (0.96) and odi% (0.99). 8.2. scenario analysis with match but without oi from the information in the last three panels of table 8, we offer the following general conclusions first, panel d is like panel a, where we find that when agi increases then the gain increases. like panel a, odi% is not consistently related to agi. for example, amid the upward climb for odi% as agi increases, odi% is 41.67% for agis of $80,000, $140,000 and $200,000 and 66.67% for agis of $110,000 and $240,000. in looking at the third to last row, we find that the average gain for all scenarios is around $18,600 and the average odi% is near 50%.11 finally, an analysis comparing panels a and d reveals that the gain and odi% both decline with a match with the odi% decline small. second, like panel b, we find that greater increases in returns require greater ira contributions to a roth ira. there are some differences in the correlation coefficients when comparing panels b and e. the positive coefficients for odibt and odiat of 0.77 and 0.76, respectively, have been reverse as they are now negative at �0.71 and �0.76. thus, both the before-tax and after-tax optimal diws are no longer positively related to the increases in returns but are negatively related to them. this is consistent with the notion that the match absorbs lower tw values causing an investor to place less in a deductible ira and more in a roth ira. perhaps, most noteworthy, we find a complete reversal in the correlation coefficients between returns and the gain (0.98 to �0.98) when comparing panels b and e. this shows the devastating effect of the match on the gain. finally, an analysis comparing panels b and e indicates that the gain and odi% both fall when the match is introduced and the fall is greater than when comparing panels a and d. third, like panel c, we show in panel f that investors with short retirement periods should be putting more in a roth ira. in comparing panels c and f, we find the same outcomes found in the previous panel comparisons. for example, with a match present, outcome variables tend to change in the same direction as found previously. a noticeable exception 410 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 is for tw as its correlation coefficient changes sign going from 0.80 in panel c to �0.70 in panel f. finally, an analysis comparing panel c and f indicates once again that the gain and odi% both fall with a match. in conclusion, table 8 illustrate how changes in key variables influence the rvd outcomes. financial advisors can gather an in depth understanding of just how much a change in a variable can influence a retiree’s optimal deductible ira allocation. table 8 serves to remind advisors what to expect if any estimation is inaccurate. for example, suppose a future retiree plans to be aggressive by investing in a portfolio heavy in stocks thereby contributing more to a roth ira. if something happens and stocks underperform by as little as 2%, then serious consequences can result in terms of one’s ira allocation choice. similarly, if the agi unexpectedly changes over time, there are consequences in terms of how the ira allocation should change as we saw that greater agis indicate a general increase in a deductible ira contribution is warranted. our scenario analysis also shows that, despite an upward trend in contributions to a deductible ira as agi increases, there can be deviations and so an advisor needs to monitor the ira allocation on a regular basis. in brief, financial advisors should be warned that updating the ira allocation should be ongoing. 9. conclusions and disclaimer this article is motivated by the desire to give financial advisors a concrete tool to help clients fulfill their retirement goal of properly choosing between the two main ira types: a traditional deductible ira and a roth ira. a survey of the literature suggests this tool is missing and financial advisors state such a tool is needed. in response, we set out to create a roth ira versus deductible ira (rvd) procedure to fill in this missing gap in the personal finance planning area. for this purpose, we developed new formulas and used them within a well-defined computational procedure that includes using an algorithmic method, which is a method that dates back to euclid in 300 bc. for our rvd procedure, we only require advisors to input 10 values from clients to produce their optimal rvd allocation. we began our rvd procedure, by introducing the concept of discovery points that are needed to develop a marginal gain formula. discovery points determine when an additional deductible ira withdrawal dollar will cause a jump to a higher marginal tax rate. we next provided values for key variables for our couple. we then projected future tax brackets covering the life span of our couple and computed their contribution tax rate using our contribution algorithm. for the next task, we introduced definitions and equations that enabled us to perform standard lump sum and annuity computations based on values from key variables. we then used a variety of different diws and placed them within our withdrawal algorithm to illustrate the maximum gain from optimally allocating retirement funds between a deductible ira and a roth ira. we then performed various illustrations showing how the optimal rvd outcomes change when key input variables change. by using the procedure in this article, financial advisors will have a tool to help facilitate any behavioral change needed in clients who are not allocating their ira funds optimally among ira types. by better understanding the retirement decision-making process through 411r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 knowing the correct rvd outcome, financial advisors should have more confidence in helping clients plan their retirements. finally, we offer a disclaimer in terms of our rvd application. because this article and its rvd procedure is new, anyone trying to duplicate this procedure to make an estimate of a proper rvd allocation should proceed with caution. subsequent research may cast light on any shortcomings found in our procedure that uses annuities (where rmds are not a factor) and mean values for the contribution and withdrawal periods. future research can improve on our procedure as needed providing more accurate estimation of the optimal rvd allocation. thus, it remains to be seen how accurate the method given in this article will hold up over time. notes 1 for educational and nonprofit employees, a 403(b) plan would be used and would be similar to the 401(k) plan in allowing both a deductible ira contribution and a roth ira contribution. thus, the use of 401(k) can also refer to any similar employee retirement plan. 2 a nondeductible ira is like an employer’s match in two respects. first, it creates taxable income during retirement that can raise the withdrawal tax rate. second, it also does not create a tax deduction. however, unlike a match, investors must pay for the nondeductible ira out of their own pockets. 3 we call it tc because it would be the same tax rate used in (5) in that marginal context. 4 as seen later in the tc columns of table 8, over 60% of the 48 values for tc cover just one marginal tax rate. the tw columns reveals just the opposite as rarely does tw cover one marginal tax rate and then it is either 0% or 12%. 5 we include zero because, as seen later, it is possible to have a diw that is not taxed. 6 for the most part, we use the “round” function in excel to get the values to the nearest dollar. tax bracket values, like $36,656, would be found in the nearest dollar in tax brackets. thus, for reasons such as this, rounding off errors can occur when reporting and comparing values. 7 the link to the ssb calculator is https://www.ssa.gov/oact/quickcalc/. 8 we get $1,565,846 by adjusting the actual value of 0.3($5,219,608) � $1,565,882 for a rounding off error of 0.002% caused from earlier computations. as mentioned previously, we use the “round” function in excel. 9 in formula form using (12), we have: marginal gain � �t1dp1 � �t2(dp2-dp1) � �t3(withdrawal-dp2) � (0.30–0)$50,588 � (0.30–0.12)($87,244–$50,588) � (0.30–0.18)($98,220–$87,244) � $15,176.40 � $6,598.08 � $746.26 � $22,521. 10 for example, the match (like the deductible ira) for a 403(b) can avoid the payment of state taxes on withdrawals during retirement and so not all of it may be taxed. however, more noteworthy, oi does not necessarily lower tw because its withdrawal can create income not taxed at the ordinary tax rate but at lower long-term capital gains and dividends rates. if oi consists of nontaxable investments like municipal bonds, then that is another argument that oi is not a factor influencing tw. 412 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 11 because we use average tw and gain values, we can extrapolate based on numbers from figure 2 by roughly estimating that a marginal gain would likely hover around $10,000 with odi% around 35%. appendix 1 proof that fvroth � fvatded when tax savings from deductible ira invested in the deductible ira and tc � tw. rc is the expected rate of return on ira contribution. step 1: get two components for the after-tax future value roth ira. 1. amount of earnings available to invest in a roth ira � $x. 2. amount of after-tax future value roth ira upon retirement � fvroth � $x(1 � rc)yc. step 2: get after-tax future value of deductible ira. 1. amount of earnings available to invest in a deductible ira � $x (1 � tc). 2. amount of after-tax future value deductible ira upon retirement � fvatded � (1 � tw)$x(1 � rc)yc (1 � tc) . step 3: set tc � tw. the amount of the future value roth ira is not taxed and so fvroth remains at $x(1 � rc)yc. if tc � tw, the amount of the future value deductible ira is: fvatded � (1 � tw)$x(1 � rc)yc (1 � tc) � (1 � tc)$x(1 � rc)yc (1 � tc) � $x(1 � rc])yc. thus, when tc � tw, we get fvroth � fvatded � $x(1 � rc)yc. q.e.d. acknowledgments we would like to thank all financial advisors and planners who helped on this manuscript and supplied the motivation for pursuing this project based on the need to solve the ira allocation dilemma. in particular, we would like to acknowledge the support and input from sam swift, cfa, cfp, tci wealth advisors, inc., tucson, az. references adelman, s. w., & cross, m. l. (2010). comparing a traditional ira and a roth ira: theory versus practice. risk management and insurance review, 13, 265–277. alhenawi, y., & elkhal, k. (2013). financial literacy of u.s. households: knowledge vs. long-term financial planning. financial services review, 22, 211–244. 413r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 anderson, k. e., & hulse, d. w. (2013). in-plan rollovers from traditional 401(k), 403(b), and 457(b) plans to roth accounts: a decision framework. journal of financial service professionals, 67, 89–98. collins, p. j., lam, h., & stampfli, j. (2015). how risky is your retirement income risk model. financial services review, 24, 193–216. crain, t. l., & austin, j. r. (1997). an analysis of the tradeoff between tax deferred earnings in iras and preferential capital gains. financial services review, 6, 227–242. horan, s. m., peterson, j. h., & mcleod, r. (1997). an analysis of nondeductible ira contributions and roth ira conversions. financial services review, 6, 243–256. horan, s. m., & peterson, j. h. (2001). a reexamination of tax-deductible iras, roth iras, and 401(k) investments. financial services review, 10, 87–100. horan, s. m., & zaman, a. a. (2009). iras under progressive tax regimes and income growth. financial services review, 18, 195–211. hrung, w. b. (2007). determinants of the choice between roth and deductible iras. journal of the american taxation association, 29, 27–42. hulse, d. w. (2003). embedded options and tax decisions: a reconsideration of the traditional vs. roth ira decision. journal of the american taxation association, 25, 39–52. krishnan, v. s., & lawrence, s. (2001). analysis of investment choices for retirement: a new approach and perspective. financial services review, 10, 75–86. ragsdale, c. t., seila, a. f., & little, p. l. (1994). an optimization model for scheduling withdrawals from tax deferred retirement accounts. financial services review, 3, 93–108. scholes, m. s., & wolfson, m. a. (1992). taxes and business strategy: a planning approach (1st ed.). englewood cliff, nj: prentice hall, inc. seida, j. a., & stern, j. j. (1998). extending scholes/wolfson for post-1997 pension investments: application to the roth ira contribution and rollover decisions. journal of the american taxation association, 20, 100–110. shynkevich, a. (2013). optimal contribution strategy as a function of the optimal withdrawal decision making: case of deductible ira versus roth ira. financial services review, 22, 51–75. sibley, m. (2002). on the valuation of tax-advantaged retirement accounts. financial services review, 11, 233–251. suarez, e. d., suarez, a., & walz, d.t. (2015). the perfect withdrawal amount: a methodology for creating retirement account distribution strategies. financial services review, 24, 331–357. welch, j. s. (2008). optimal distributions from tax-advantaged retirement accounts (available at: http:// www.i-orp.com/modeldescription/modeldescriptionk.pdf). welch, j. s. (2015). mitigating the impact of personal income taxes on retirement savings distributions. journal of personal finance, 14, 17–25. 414 r.m. hull, j.b. hull / financial services review 25 (2016) 373–414 household use of financial planners: measurement considerations for researchers stuart j. heckman, ph.d., cfpa,*, martin c. seay, ph.d., cfpb, kyoung tae kim, ph.d.c, jodi c. letkiewicz, ph.d.d aassistant professor of personal financial planning, kansas state university, 1324 lovers lane, 319 justin hall, manhattan, ks 66505, usa bassistant professor of personal financial planning, kansas state university, 1324 lovers lane, 318 justin hall, manhattan, ks 66505, usa cassistant professor, department of consumer sciences, university of alabama, 312 adams hall, tuscaloosa, al 35487, usa dassistant professor, school of administrative studies, 4700 keele street, 282 atkinson, york university, toronto, on m3j 1p3, canada abstract using the certified financial planner (cfp) board’s definition of financial planning, this article evaluates the validity of the measures of financial planner use in publicly available datasets. a review of financial services review, journal of personal finance, journal of financial planning, journal of family and economic issues, journal of consumer affairs, and journal of financial counseling and planning identified seven datasets that were commonly used to investigate financial planner use. of these, the two most promising measures were found in the survey of consumer finances and the national longitudinal study of youth (1979). however, an evaluation of these measures raises significant concerns related to their validity. this article critically evaluates these measures and provides insights into the development of better measures of financial planner use for the future. © 2016 academy of financial services. all rights reserved. jel classification: c81; d14; g20 keywords: financial planner use; financial advice; measurement * corresponding author. tel.: �1-785-532-1371; fax: �1-785-532-5505. e-mail address: sheckman@ksu.edu (s. j. heckman) financial services review 25 (2016) 427–446 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. 1. introduction as an academic discipline, personal financial planning is still relatively new. the first doctoral program in personal financial planning was established at texas tech university in 2000 (brandon and welch, 2009)—a mere 16 years ago. although researchers have been investigating issues related to personal financial planning for some time, the development of a dedicated doctoral program is indicative of the developmental stage of the field. since the first program at texas tech, three additional universities (kansas state university, university of georgia, and university of missouri) have established certified financial planner (cfp) board registered doctoral programs that have led to a growth in the number of financial planning researchers and a corresponding increase in the scientific knowledge base. however, as in any field of knowledge, scientific advancement is dependent upon proper measurement of the relevant objects or concepts. in personal financial planning, an important research objective is to determine the effect of financial planner use on household well-being.1 to properly investigate this relationship, it is necessary to determine whether or not a household receives financial planning advice. at first glance, this may seem simple. however, a careful assessment of publically available datasets indicates that this measurement is not straightforward. the objective of this study is to promote increased rigor in the field of personal financial planning research by examining the important issue of measurement as it relates to household use of financial planners. specifically, we analyze the validity of available measures of household financial planner use in publicly available datasets and provide recommendations for the development of new measures of financial planner use. 2. measurement although measurement has been defined in a variety of ways, two definitions are especially important to the current study. stevens (1951) defines measurement as “the assignment of numbers to objects or events according to rules” (p. 22). while this definition is quite good for the natural sciences, it may be limited in social science applications in which abstract concepts are more often the subject of investigation (rather than objects) (carmines and zeller, 1979). zeller and carmines (1980) define measurement as “a process of linking abstract concepts to empirical indicants” (p. 2). carmines and zeller (1979) emphasize the importance of measurement as it allows scientists to test theoretical propositions—if the empirical indicant (i.e., a variable that can be observed) is weakly related to the underlying phenomenon of interest (i.e., the unobserved concept), any analysis of the data may lead to incorrect inferences. therefore, we consider measurement to be a process in which objects, events, or concepts are systemically classified and represented to advance knowledge. with this understanding of measurement, we can return to the objective of the current analysis. to properly classify households according to whether or not they use a financial planner, several challenges become apparent. even among professionals who refer to themselves as financial planners, there is great diversity in the scope of services provided, ranging from primarily investment advice or another specialized area to comprehensive financial 428 s.j. heckman et al. / financial services review 25 (2016) 427–446 planning. consequently, self-reported data from households will be noisy as there is likely substantial variation between households in the types of financial service professionals (e.g., banker, stock broker, insurance agent, etc.) that come to mind when they hear the term, “financial planner.” therefore, an important starting point for research regarding financial planner use is to clearly define the intended event to be measured. although a number of studies examine the effect of financial planner use, little attention is given to clearly defining what is meant by use of a financial planner. in measuring whether an individual received financial planning services, a clear definitional framework is needed. the cfp board (2013) defines financial planning as “the process of determining whether and how an individual can meet life goals through the proper management of financial resources” (p. 9). the subject areas of the financial planning process include, but are not limited to: “financial statement preparation and analysis (including cash flow analysis/planning and budgeting), insurance planning and risk management, employee benefits planning, investment planning, tax planning, retirement planning, and estate planning” (cfp board, 2013, p. 9). considerations in evaluating whether financial planning occurred include “the client’s understanding and intent in engaging in financial planning, the degree to which multiple financial planning subject areas are involved, the comprehensiveness of data gathering, and the depth and breadth of planning recommendations” (cfp board, 2013, p. 9). 2.1. measurement evaluation: reliability and validity once the event (or other phenomenon) is clearly defined and a possible measure (or set of measures) is identified, the next step is to determine how well the measure performs. for evaluation purposes, carmines and zeller (1979) identify two basic properties of a measure: reliability and validity. reliability refers to a measure’s ability to yield consistent results over repeated trials. although there will always be random variation in any measurement, the goal is to have a measure that yields consistent results. validity refers to the extent to which a measure “. . . does what is intended to do” (carmines and zeller, 1979, p. 12). there are three commonly accepted components of validity: (1) content validity, (2) criterion-related validity, and (3) construct validity. content validity refers to the extent to which an empirical measurement reflects a specific domain of context (carmines and zeller, 1979). for example, a measure of investment knowledge among u.s. adults that assessed understanding of asset return but not asset risk would lack content validity. devellis (2012) points out that content validity is closely related to the definition of the phenomenon of interest. to be valid, a measure must capture the aspects of the occurrence identified in its conceptual definition. criterion-related validity refers to the extent to which a measure is related to an empirical behavior that is external to the measure (carmines and zeller, 1979; devellis, 2012). for example, a measure of risk tolerance may have evidence of criterion-related validity if it is strongly related to amount of risk held in a portfolio. lastly, construct validity refers to the degree to which a certain measure relates to other measures in line with theory-based hypotheses concerning the constructs being measured (carmines and zeller, 1979). for example, if a theoretical framework suggests that increased financial stress should predict increased likelihood of 429s.j. heckman et al. / financial services review 25 (2016) 427–446 financial planner use, a measure of financial stress would have evidence of construct validity if there was a strong, positive correlation between the measure and financial planner use. for the purpose of this article, we focus on evaluating the content validity of publically available measures of financial planner use. 2.2. measures in publicly available datasets to identify publicly available, nationally representative datasets in the united states that contain information regarding household use of financial planners, articles published between 2013 and 2015 in financial services review, journal of personal finance, journal of financial planning, journal of family and economic issues, journal of consumer affairs, and journal of financial counseling and planning were reviewed. from this list, seven datasets were identified that contain information about seeking financial help from professionals: asset and health dynamics among the oldest old (ahead), american life panel (alp), health and retirement study (hrs), national financial capability survey (nfcs), national longitudinal survey of youth 1979 (nlsy79), national longitudinal survey of youth 1997 (nlsy97), and survey of consumer finances (scf). the following discussion highlights the available survey questions and the ways in which the literature has used the questions to measure financial planner use. table 1 summarizes each dataset’s available measures. the ahead survey, which was integrated with the hrs in 1998, contains a question in 1993 and 1995 that ask respondents “do you have a financial advisor who helps make decisions?” only one study published in the journals reviewed used this question; cummings and james (2014) analyze factors associated with getting and dropping a financial advisor. the alp has administered at least four surveys (surveys 5, 13, 21, and 332) that collect information regarding the use of financial professionals. surveys 5 and 21 asked whether the respondent consulted “a financial planner or advisor or an accountant” for retirement planning. survey 13 of the alp is especially note-worthy as it includes detailed financial service use questions, including whether the respondent uses a financial professional, how the professional(s) is compensated, how long they have been doing business, and how satisfied they are with the services. survey 33 asks whether respondents relied on a broker of financial advisor for retirement planning. our analysis indicates that only one study in the journals reviewed has used a financial professional use measure from the alp3: knoll and houts (2012) use a concatenated sample of the alp, hrs, and nfcs to investigate financial literacy. they assess the validity of their financial literacy measure by correlating it with “financial planner use,” which was measured using the questions from the 2004 hrs and alp surveys 5 and 21.4 note that the 2004 hrs questions and alp questions from surveys 5 and 21 are the same question asked with the same lead-in questions about retirement planning. the hrs has several different measures of financial planner use. a topical module in 2000 asks preretirees if they consulted a financial planner for retirement savings and asks retirees to (retrospectively) indicate if they consulted a financial planner in their preretirement years. a different 2004 topical module asks respondents if, in the context of retirement planning, they had “. . . consulted a financial planner or advisor or an accountant.” the 2014 hrs asks 430 s.j. heckman et al. / financial services review 25 (2016) 427–446 table 1 review of public, nationally representative datasets dataset question text of available measuresa variable numbers notes ahead do you have a financial advisor that helps make decisions? v1921 (1993) merged with the hrs dataset in 1998; question has not used since. d5318 (1995) alp survey 5 and 21 (lead questions: have you ever tried to figure out how much your household would need to save for retirement? tell me about the ways you tried to figure out how much your household would need.) same questions as hrs retirement planning questions. questions in survey 13 are asked for up to 5 different individual professionals and firms. did you consult a financial planner or advisor or an accountant? r003_5 survey 13 do you currently use any professional financial service providers—include individual professionals and/or firms—for: conducting stock market and/or mutual fund transactions: (e.g.,, purchases and sales of stocks, shares in mutual funds, options contracts, short selling, and so forth). please exclude transactions that involve an employer sponsored retirement account. advising, management, and/or planning: (e.g., financial advising, investment advising, financial planning, money management, retirement planning, estate planning, and so forth). check all that apply. fs1 first we would like to ask you about [conducting stock market and/or mutual fund transactions/ advising, management, and/or planning]. is there an individual professional with whom you personally interact regarding [conducting stock market and/or mutual fund transactions/ advising, management, and/or planning] services? fs2 how do you pay this [individual professional/ firm] for [conducting stock market and/or mutual fund transactions/advising, management, and/or planning] services? please check all that apply [commission (e.g., per transaction), hourly, monthly, or annual rate, flat fee, percentage fee (e.g., % of my account balance), other] fs5 what is the rate that you pay for [conducting stock market and/or mutual fund transactions/ advising, management, and/or planning] services? fs5b about how long have you been doing business with this [individual professional/firm]? fs7 survey 33 for your retirement planning, do you rely on financial software, a website with a financial calculator, or a broker or financial advisor? you may check several answers. fsftexp hrsa (lead question: have you ever made a plan and calculated what you would need at retirement?) g6802 is asked of preretirees. did you consult a financial planner? (lead question: before you retired did you make a plan and calculate what you would need at retirement?) g6802 (2000) g6789 is asked of retirees to retrospectively report whether they consulted a planner. 431s.j. heckman et al. / financial services review 25 (2016) 427–446 table 1 (continued) dataset question text of available measuresa variable numbers notes did you consult a financial planner? g6789 (2000) (lead questions: have you ever tried to figure out how much your household would need to save for retirement? tell me about the ways you tried to figure out how much your household would need.) did you consult a financial planner or advisor or an accountant? jv356 (2004) (lead question: do you have someone such as a friend or relative, or bank officer, lawyer or financial consultant who regularly helps you with handling your money or property or other financial matters such as signing checks, paying bills, dealing with banks and making investments?) who helps you [and your [partner/husband/wife]] with your finances? choose all that apply. [one response option is “financial consultant, accountant, or other professional investment counselor”] ov502m1 (2014) (lead question: have you given permission to a bank, lawyer, broker or other financial advisor to be able to share your information with family members, friends, or others?) with whom can your financial information be shared? [one response option is “financial consultant, accountant, or other professional investment counselor”] ov531 (2014) who is designated (as power of attorney)? [one response option is financial consultant, accountant or other professional investment counselor] ov509 (2014) nfcs in the last 5 years, have you asked for any advice from a financial professional about any of the following? [debt counseling, savings or investments, taking out a mortgage or a loan, insurance of any type, tax planning] k_1–k_5 available in 2009 and 2012 surveys. nlsy79 people begin learning about and preparing for retirement at different ages and in different ways. have you consulted a financial planner about how to plan your finances after retirement? t09628.01 (2006) available 2006–2012. t21836.01 (2008) t30959.01 (2010) t40951.01 (2012) nlsy97 in the past 12 months, who have you talked with about money issues most often? [one response option is “someone with professional expertise in the field’] s84959.00 (2006) t08892.00 (2007) t30024.00 (2008) t44054.00 (2009) available in the 2006–2013, asked of respondents who talked to someone about finances in past 12 months. t60549.00 (2010) t75450.00 (2011) t89760.00 (2013) scf what sources of information do you use to make decisions about borrowing or credit? (do you call around, read newspapers, magazines, material you get in the mail, use information from television, radio, the internet or advertisements? do you get advice from a friend, relative, lawyer, accountant, banker, broker, or financial planner? or do you do something else?) borrowing/credit x7101-x7110, x68479, x6861-x6864 available in all survey waves; however, the financial planner response option was not added until 1998. reponses are recorded for up to 15 responses. 432 s.j. heckman et al. / financial services review 25 (2016) 427–446 “who helps you with your finances?” and one response category is “financial consultant, accountant, or other professional investment counselor.” as shown in table 1, the 2014 hrs also contains a few other questions that include a response option that identifies a financial consultant. only one study (knoll and houts, 2012) published in the journals reviewed used the (2004) hrs measure of financial professional use. the nfcs asks respondents if, in the last five years, they sought advice from a financial professional about the following categories (allowing for unique responses for each category): debt counseling, savings or investments, taking out a mortgage or loan, insurance of any type, or tax planning. a number of researchers have used the nfcs to explore financial advice in different capacities (balasubramnian, brisker, and gradisher, 2014; collins, 2012; lachance and tang, 2012; robb, babiarz, and woodyard, 2012; sass, belbase, cooperrider, and ramos-mercado, 2015; simms, 2014; tang and lu, 2014). balasubramnian et al. (2014) analyze households who reported using a financial adviser for any subject area and examine which households conduct regulatory searches when choosing an adviser. collins (2012) and robb et al. (2012) both examine financial advice use measured as each of the five categories independently plus a category for any advice. lachance and tang (2012) investigate the relationship between trust and financial advice and measure financial advice using each of the five categories independently. sass et al. (2015) examine financial well-being and include financial advice use as a covariate, measured as consulting a financial professional on any subject area. tang and lu (2014) analyze loan decisions to see whether consulting a financial professional, measured using only the debt counseling and taking out a mortgage or loan responses, influenced the use of 401(k) loans. simms (2014) analyzes women’s use of investment advice by using only the saving or investment response on the financial professional question. table 1 (continued) dataset question text of available measuresa variable numbers notes what sources of information do you use to make decisions about saving and investments? (do you call around, read newspapers, magazines, material you get in the mail, use information from television, radio, the internet or advertisements? do you get advice from a friend, relative, lawyer, accountant, banker, broker, or financial planner? or do you do something else?) saving/investment x7112-x7121, x6865-x6869 notes: question text allowing the interviewer to ask the question appropriately for couples or other grammatical adjustments have been omitted for simplicity. ahead � asset and health dynamics among the oldest old; alp � american life panel & hrs � health and retirement study; alp13 � american life panel survey 13; nfcs � national financial capability survey; nlsy79 � national longitudinal survey of youth 1979; nlsy97 � national longitudinal survey of youth 1997; scf � survey of consumer finances. aquestions were identified using the hrs concordance, http://hrsonline.isr.umich.edu/index.php?p�concord, and searching the following phrases: “financial planner,” “financial advisor,” “financial professional,” and “financial consultant.” the hrs may contain other questions that include some information about financial professional use. for example, there are questions about who serves as trustee (e.g., ov517m1) that contain a response code identifying an “investment counselor” or “consultant.” those questions are not included here because of the narrow role the advisor plays, but these questions may be useful in other applications. 433s.j. heckman et al. / financial services review 25 (2016) 427–446 both the nlsy79 and nlsy97 cohort surveys contain questions about financial advice. the nlsy79 contains one question that states the following: “people begin learning about and preparing for retirement at different ages and in different ways. have you [or] [spouse/partner’s name] consulted a financial planner about how to plan your finances after retirement?” several studies have analyzed financial planner use in the 2008 wave of the nlsy79 by using this question (martin, finke, and gibson, 2014; martin and finke, 2014). martin and finke (2014) create a category they refer to as “comprehensive financial planner” that was measured by two components. the first component was based on the financial planner question and the second was based on a question that asked, “have you [or] [spouse/ partner’s name] ever calculated how much retirement income you would need at retirement?” as shown in table 1, these nlsy79 questions are available in survey waves from 2006 to 2012. the nsly97 asks, “in the past twelve months, who have you talked with about money issues most often?” of the possible responses, one category is “someone with professional expertise in the field.” this question has not been used to assess the use of a financial planner in the publications reviewed. the scf contains two questions about the source(s) of information used by the respondent and spouse/partner (if applicable) for (1) saving and investment decisions and (2) borrowing and credit decisions. respondents who were interviewed in person were shown a list of information sources and interviewers read the same list to respondents for telephone interviews. the list includes the following: call around, read newspapers or magazines, information received in the mail, information from television, radio, internet, advertisements, or advice from a friend, relative, lawyer, accountant, banker, broker, financial planner, or other. responses to these questions are coded in the order that they are listed by the respondent for up to 15 responses. the scf has been used to investigate financial planner use as an outcome (elmerick, montalto, and fox, 2002; hanna, 2011) and as a predictor of perception of retirement preparedness (kim and hanna, 2015), life insurance adequacy (scott and gilliam, 2014), disability insurance ownership (scott and finke, 2013), and consistency of risk attitudes and risky behavior (park and yao, 2015). hanna (2011) identifies a household as using a financial planner if “financial planner” was selected on either the saving/investment or borrowing/ credit question. other scf researchers have used only responses on the saving question to indicate financial planner use, for example, kim and hanna (2015). park and yao (2015) measure financial planner use by utilizing only the first response to the saving/investment question and including the following categories: lawyer, accountant, and financial planner.5 scott and finke (2013) include accountant, banker, and broker in their measure of financial planner while scott and gilliam (2014) measure financial planner and nonfinancial planner use, although it is not clear if the saving or borrowing question was used in either study. researchers have also used the scf to examine “comprehensive” financial planner use, defined as households reporting the use of a financial planner on both the saving/investment and borrowing/credit questions (elmerick et al., 2002). lastly, researchers have classified financial planner use as a more general financial professional measure to examine lowincome household saving behavior (heckman and hanna, 2015) and low-income household 434 s.j. heckman et al. / financial services review 25 (2016) 427–446 financial behaviors (hudson and palmer, 2014). heckman and hanna (2015) use responses on either the saving or borrowing question, and hudson and palmer (2014) use only the saving/investment question. 2.3. measures in other datasets the review of literature also revealed a number of proprietary and primary datasets that contained information about financial planner use. although these datasets may not be accessible to other researchers, understanding the measures in these studies is helpful in terms of developing recommendations for future measures, discussed later in this article. winchester and huston (2014) analyze a proprietary dataset, cosponsored by a large independent financial services company and a financial planning professional association, in which financial planner use was based on responses to two questions. the first question asked if the respondent had a written financial plan and the second asked how that plan was developed. respondents were identified as using a financial planner if they had a written plan and indicated that the plan had been tailored to their financial goals after a meeting with a financial planner. a report from the society of actuaries (soa) and a joint report from the cfp board and consumer federation of american (cfa) also examine financial planner use among u.s. households. one question in the soa survey asked respondents the following: “about how often do you (and your spouse/partner) consult with a financial planner or adviser who helps you make decisions about your retirement/financial planning and is paid through fees or commissions?” (society of actuaries, 2013, p. 88). the cfp board and cfa survey includes several items related to financial plans, including type (e.g., written) and recency,6 and financial planner use in preparing those plans (see princeton survey research associates international, 2013). among respondents who reported having a financial plan, two follow-up questions were asked: “did a financial professional help you to prepare this plan? for example, a financial planner, banker, stock broker, accountant, insurance agent, or investment advisor” (princeton survey research associates international, 2013, p. 53). “some financial professionals who help people with their plans, such as certified financial planners and registered investment advisors, have a fiduciary duty. this means they are required to act in the best interest of their clients, when providing financial planning or investment advice. as far as you know, is the financial professional who helped you with your most recent plan a certified financial planner, a registered investment advisor, or other professional with a fiduciary duty to act in your best interest?” (princeton survey research associates international, 2013, p. 53). several studies have used primary data and included measures of professional financial advice in general (i.e., not specific to financial planners). survey questions include whether respondents relied on someone else’s advice when making investment decisions (kuzniak, rabbani, heo, ruiz-menjivar, and grable, 2015), whether respondents met with a financial advisor in the last 12 months (zick, mayer, and kara, 2012), whether respondents took advantage of meeting with a financial coach7 (moulton, loibl, samak, and collins, 2013), 435s.j. heckman et al. / financial services review 25 (2016) 427–446 and whether respondents consulted a financial professional or advisor (eccles, ward, goldsmith, and arsal, 2013; gibson, michayluk, and van de venter, 2013). gibson et al. (2013) also included whether the respondent used an advisor two years ago and whether or not the current and previous advisor (i.e., from two years ago) were the same person. warschauer and sciglimpaglia (2012) obtained the most detailed information about household financial planner use to date. they examine consumer perceptions about the value of financial planning services and their survey included questions about previous experience with financial planners and the perceived qualifications of the respondents’ financial planners. experience questions included whether the respondent had (1) an up-to-date comprehensive written plan, (2) a written plan focused on one or two issues, (3) received professional advice orally but did not have a written plan, and (4) had a plan but it is out-of-date (warschauer and sciglimpaglia, 2012, p. 199). among respondents who reported experience with a financial planner, the survey asked for the planner’s qualifications with the following response categories: (1) “cfp licenses,” (2) “cpa, enrolled agents, or licensed tax preparers,” (3) “licensed attorney,” (4) “stock broker or insurance agent,” (5) “private or personal banker,” (6) “fee-only planner,” and (7) “don’t know” (warschauer and sciglimpaglia, 2012, p. 200). 2.4. summary and gap to summarize, researchers have measured financial planner use among u.s. households in a variety of ways. although publicly available datasets provide survey items regarding u.s. household use of financial planners the literature to date has not carefully evaluated the validity of such data. to the authors’ knowledge, there are no studies that focus on the measurement of financial planner use. therefore, this study contributes to the literature by providing an evaluation of current publicly available measures of financial planner use and by concluding with measurement recommendations for researchers. 3. method 3.1. sample and analysis we evaluate the validity of the financial planner use measures in seven national datasets8 by providing a careful examination of the content validity of the individual survey questions in two ways. first, we evaluate the content validity of each measure using the cfp board’s definition of financial planning. the cfp board’s definitional framework provides nine distinct content domains that are critical to determining whether financial planning has occurred. consequently, each measure is evaluated based on the extent it addresses each of these domains. second, longitudinal datasets are used to evaluate the extent that each measure validly tracks a household’s use of a financial planner over time. specifically, we evaluate what these measures imply about respondents’ changes in financial planner use between two time periods. although the alp, hrs, nlsy97 are all longitudinal, none are suitable for this 436 s.j. heckman et al. / financial services review 25 (2016) 427–446 type of analysis. only five individuals participate in both survey 5 and survey 21 of the alp; the hrs does not use consistent questions in different survey years; and the nlsy97 question is too vague for inference. the remaining datasets (ahead, nlsy79, and scf) are all good candidates for this analysis. we utilize the financial planner use rates reported by cummings and james (2014) in their analysis of the ahead data and analyze financial planner use in the 2007–2009 scf panel and the 2010–2012 waves of the nlsy79 to test whether observed usage rates are consistent with what we might expect based on industry reports. 4. results 4.1. content validity recall that we adopt the cfp board’s definition of financial planning that includes essentially three components: (1) receipt of financial planning services with an emphasis on life goals, (2) subject areas addressed, and (3) the depth and breadth of the relationship and recommendations. therefore, a valid measure of financial planning should identify that a process has occurred and include questions that ask life goals and financial management decisions, allow for the identification of one more content areas covered by financial planning services, and include questions that address the various aspects of planning engagement. as shown in table 2, the nlsy79 specifically identifies “financial planner” as the professional being consulted and the scf and alp survey 13 allow clear identification for a variety of professionals, including financial planner; the question phrasing in the alp/ hrs, nfcs, and nlsy97 questions do not allow for a clear identification of the type of professional used. none of the questions refer to a process or life goals. the alp/hrs, nlsy79, and scf each have clear references to using advice for management decisions. except for the nlsy97, all measures contain information regarding the subject area covered, however, the alp/hrs and nlsy79 only ask about retirement. the scf includes two different areas (i.e., saving and borrowing) and the nfcs and alp survey 13 are the most comprehensive with five or more areas covered. the identification of multiple subject areas may explain why the nfcs and scf have been so widely used in the literature. none of the measures used contain information regarding client intent, comprehensiveness of data gathering, or depth and breadth of recommendations. overall, the questions from the analyzed nationally representative datasets were found to perform poorly on tests of content validity. this suggests that the measures fail to address the components of financial planning, providing sufficient doubt as to whether they effectively measure financial planner use. further, they fail to allow a researcher to distinguish between the type of services received and the extent of the planning engagement. while the questions have utility in narrow applications, they do not withstand any rigorous evaluation of their validity in measuring financial planner use as outlined by the cfp board (2013). 437s.j. heckman et al. / financial services review 25 (2016) 427–446 t ab le 2 e xi st in g m ea su re s an d co m po ne nt s of co nt en t va lid ity fo r fin an ci al pl an ne r us e c om po ne nt a ss et an d h ea lth d yn am ic s am on g th e o ld es t o ld (a h e a d ) n at io na l da ta se ts a m er ic an l if e pa ne l (a l p) & h ea lth an d r et ir em en t st ud y (h r s) a m er ic an l if e pa ne l su rv ey 13 (a l p1 3) n at io na l fi na nc ia l c ap ab ili ty su rv ey (n fc s) n at io na l l on gi tu di na l su rv ey of y ou th 19 79 (n l sy 79 ) n at io na l l on gi tu di na l su rv ey of y ou th 19 97 (n l sy 97 ) su rv ey of c on su m er fi na nc es (s c f) c le ar id en tifi ca tio n of pr of es si on al n o n o y es –m ul tip le n o y es –f in an ci al pl an ne r n o y es –m ul tip le pr oc es s n o n o n o n o n o n o n o l if e g oa ls n o n o n o n o n o n o n o m an ag em en t of r es ou rc es y es : h el ps m ak e de ci si on s y es : h el pe d de te rm in e ne ed fo r re tir em en t sa vi ng s n ot di re ct ly , bu t m ay be im pl ie d n o y es : pl an ni ng fin an ce s af te r re tir em en t n o y es : in fo rm at io n us ed to m ak e de ci si on s m ul tip le a re as c ov er ed n o n o y es y es n o y es y es sp ec ifi c a re as id en tifi ed n o r et ir em en t fi na nc ia l ad vi si ng in ve st m en ts fi na nc ia l pl an ni ng m on ey m an ag em en t r et ir em en t pl an ni ng , e st at e pl an ni ng an d so fo rt h (o th er ) d eb t co un se lin g sa vi ng s/ in ve st m en ts m or tg ag es in su ra nc e t ax pl an ni ng r et ir em en t n o sa vi ng /i nv es tm en t b or ro w in g/ c re di t c lie nt ’s in te nt to e ng ag e a fi na nc ia l pl an ne r n o, bu t m ay be im pl ie d n o n o, bu t m ay be im pl ie d n o n o, bu t m ay be im pl ie d n o n o, bu t m ay be im pl ie d c om pr eh en si ve ne ss of d at ac ol le ct io n n o n o n o n o n o n o n o d ep th an d b re ad th of r ec om m en da tio n n o n o n o n o n o n o n o n ot e: b ec au se th e a l p an d h r s (a l p � a m er ic an l if e pa ne l & h r s � h ea lth an d r et ir em en t st ud y) co nt ai n a va ri et y of qu es tio ns , w e fo cu s th is co nt en tv al id ity an al ys is on th e m ea su re s th at ha ve be en us ed by pr ev io us lit er at ur e, w hi ch in cl ud e qu es tio ns fr om a l p su rv ey 5 an d 21 an d th e h r s 20 04 . w e co m bi ne th e a l p an d h r s an al ys is be ca us e th e qu es tio ns ar e th e sa m e in th os e su rv ey ye ar s. a lth ou gh a l p su rv ey 13 ha s no t be en us ed pr ev io us ly to ou r kn ow le dg e, w e al so in cl ud e it be ca us e of th e un iq ue m ea su re s. 438 s.j. heckman et al. / financial services review 25 (2016) 427–446 4.2. evidence of validity from financial planner use over time both the nlsy79 and the scf panel have longitudinal data enabling us to assess whether the measures yield results consistent with expected behavior patterns over time. use of a financial planner in the scf is classified in four ways: (1) whether a respondent consulted a financial planner in both domains (comprehensive planner), (2) whether a respondent consulted a financial planner in making saving/investment decisions, (3) whether a respondent consulted a financial planner in making credit/borrowing decision, and (4) whether a respondent consulted a financial planner in either domain (any planner). further, patterns in financial planner use were measured with four dummy variables indicating use of a financial planner in both periods, dropped a planner in 2009, adopted a planner in 2009, and no financial planning service in both survey waves. descriptive statistics related to the use of financial planners in the 2007–2009 scf and 2010–2012 nlsy79 can be found in table 3 and are depicted in fig. 1. both descriptive analyses are weighted to be representative of the u.s. population. the 2007–2009 scf includes 3,857 households and the 2010–2012 waves of the nsly79 include 5,584 respondents. results from the 2007–2009 scf indicated that 9.2% of all households used a comprehensive planner in 2007, 22.3% consulted a financial planner for savings decisions, 12.5% used a financial planner for borrowing decisions, and 25.5% reported consulting a financial planner when making either saving or credit decisions. minor decreases in planner use is noted across the board by 2009, with the percentage of the population consulting a planner declining between 1.1 percentage points and 2.3 percentage points depending on the measure. despite these modest changes, great volatility in planner use was noted. almost two-thirds (63.6%, n � 321) of respondents who engaged in comprehensive planning in 2007 reported that they did not engage again in 2009. on the aggregate, this was largely offset by the 6.4% (n � 246) of the sample that adopted a comprehensive planner in 2009. similarly, exit rates of 46.7%, 64.6%, and 45.3% were found for respondents using investment planners, credit planners, and any planner, respectively. a similar pattern was noted in financial planner use in the 2010–2012 nlsy79. consulting a financial planner for retirement was measured as follows: use of a financial planner in both survey waves, dropped a planner in 2012, adopted a planner in 2012, and no financial planning service in either period. overall, the proportion of households who consulted a financial planner for retirement decreased between 2010 and 2012 (from 24.7% to 23.1%). as in the scf, similar volatility in planner use was noted; 43.1% of respondents who consulted a planner in 2010 dropped their service by 2012. our results are consistent with the ahead data results reported by cummings and james (2014)—they found that 52.7% of respondents who reported using a financial advisor in 1993 no longer reported using a financial advisor in 1995. information on client retention rates for financial planners is limited. the best estimate is provided by pricemetrix (2013), which used aggregated data from 7 million retail investors to investigate advisor retention rates between 2009 and 2013. the report found that the median advisor retained roughly 94% percent of clients each year between 2009 and 2013. poor performing advisors, those in the 10th percentile, were still found to annually retain between 81% and 87% of clients over this same time period. while client retention rates were 439s.j. heckman et al. / financial services review 25 (2016) 427–446 t ab le 3 fi na nc ia l pl an ne r us e in th e 20 07 –2 00 9 sc f pa ne l an d 20 10 –2 01 2 n l sy 79 20 07 –2 00 9 sc f pa ne l 20 10 –2 01 2 n sl y 79 c om pr eh en si ve pl an ne r sa vi ng /in ve st m en t pl an ne r on ly c re di t/b or ro w in g pl an ne r on ly a ny pl an ne r sa m pl e (n ) pr op or tio n (% ) sa m pl e (n ) pr op or tio n (% ) sa m pl e (n ) pr op or tio n (% ) sa m pl e (n ) pr op or tio n (% ) sa m pl e (n ) pr op or tio n (% ) pl an ne r in 20 07 50 5 9. 2 1, 14 0 22 .3 63 0 12 .5 1, 26 6 25 .5 pl an ne r in 20 10 1, 42 1 24 .7 pl an ne r in 20 09 43 0 8. 1 1, 05 9 20 .7 53 2 10 .6 1, 16 0 23 .2 pl an ne r in 20 12 1, 27 2 23 .1 pl an ne r in 20 07 an d 20 09 18 4 2. 9 60 8 10 .4 22 3 3. 7 69 2 12 .1 pl an ne r in 20 10 an d 20 12 80 8 15 .3 d ro pp ed a pl an ne r in 20 09 32 1 6. 3 53 2 11 .9 40 7 8. 8 57 3 13 .4 d ro pp ed a pl an ne r in 20 12 61 3 9. 4 a do pt ed a pl an ne r in 20 09 24 6 5. 2 45 1 10 .3 30 9 6. 9 46 8 11 .1 a do pt ed a pl an ne r in 20 12 46 4 7. 8 n ei th er 20 07 no r 20 09 3, 10 6 85 .6 2, 26 6 67 .4 2, 91 8 80 .6 2, 12 4 63 .4 n ei th er 20 10 no r 20 12 4, 83 2 67 .5 t ot al 3, 85 7 10 0 3, 85 7 10 0 3, 85 7 10 0 3, 85 7 10 0 t ot al 6, 71 7 10 0 n ot es : sa m pl e si ze re fle ct s th e ac tu al (t ha t is , un w ei gh te d) nu m be r of ob se rv at io ns an d pr op or tio ns ar e w ei gh te d to be na tio na lly re pr es en ta tiv e. 440 s.j. heckman et al. / financial services review 25 (2016) 427–446 found to vary by advisor, this report would indicate the tremendous volatility in households reporting the use of a financial planner, and more specifically the large exit from the financial advisory market, observed in the ahead, nlsy79, and scf datasets exceeds reasonable expectations. it is important to note that none of the measures would detect individuals that changed planners, transitions that would serve to decrease retention rates reported by pricemetrix (2013), but rather those that no longer consulted any planner. consequently, it would appear these measures may not be validly representing planner use over time. this measurement error may be because of inconsistent definitions of financial planning among the general populace and/or the vagueness of the questions in addressing the extent of a financial planning engagement (e.g., regular or sporadic one-on-one consulting, attending a seminar, etc.). another challenging aspect of the nlsy79 question is that it asks about historical use of a financial planner and not whether the respondent is currently engaged with a planner. notably, the inconsistency in the scf’s measure of planner use was only detectable because of the limited availability of panel data in the scf; usage rates from period-to-period look consistent when examining the cross sectional scf data but the follow up survey on the same respondents reveals poor representation of planner use. 5. discussion and implications this article evaluates the content validity of measures of financial planner use in publically available datasets employed by the literature. a three-year review of financial services review, journal of personal finance, the journal of financial planning, journal of family and economic issues, journal of consumer affairs, and journal of financial counseling and planning identified seven datasets that warranted further investigation. of these datasets, the scf and nlsy79 were found to have the most promising measures. when evaluated within the cfp board’s definition of financial planning, significant validity concerns are noted in fig. 1. percentage of respondents who report adopting or dropping financial planning services between periods. 441s.j. heckman et al. / financial services review 25 (2016) 427–446 both measures. specifically, each fails to assess the comprehensiveness of data collection, the breadth and depth of planning recommendations, the focus on the achievement of life goals, and the comprehensiveness of financial planning subject areas addressed. further, an investigation of household responses over time raises additional concerns. significant variation in household’s responses that exceed behavioral expectations suggests that sizable measurement error is present. the results indicate that existing financial planner use measures are insufficient and do not allow researchers to capture the diversity and complexity of financial planning engagements. given these results, better measures of financial planner use are needed. a single question is likely insufficient to adequately measure financial planner use given the variety of forms that financial planning may undertake. the cfp board’s definition of financial planning provides significant insight into the type of questions that would be needed. specifically, we suggest that survey questions be designed to address each aspect of the definitional framework. recall that the cfp board defines financial planning as “the process of determining whether and how an individual can meet life goals through the proper management of financial resources” (p. 9). additionally, there are other factors that should be considered in determining whether professional financial planning has occurred, such as client perceptions of the financial planning engagement and the breadth and depth of financial planning subject areas, data collection, and recommendations (cfp board, 2013). although primary data were not directly analyzed in the current study, warschauer and sciglimpaglia (2012) collected the most detailed information about financial planner use to date – their study is a useful source of items that may be adopted in the future. additionally, table 4 provides example survey items that may be used to create a more informative measure of financial planner use. the questions are intended as a guide and have not been tested for rigor or validity. future research is needed to develop these measures further. when measuring financial planner use, an important consideration for researchers is that perceptions regarding the financial planning process (i.e., what is financial planning?) and financial planners (i.e., who is a financial planner?) may vary widely among both professionals and households. for example, some professionals may use the title “financial planner” loosely and households may see “financial planner” as a synonym for investment adviser, stock broker, or banker. the cfp board clearly indicates that not all client engagements should be considered financial planning, even if a client works with a professional who often operates as a financial planner. while there is some merit to this, we believe it is important to gather information about all financial planning activities to properly measure and understand the types of help consumers demand. while someone may not receive comprehensive financial planning help, there is some utility in understanding the different types of services people are getting. future research might also consider surveying professionals to understand these service offerings. this knowledge will help financial services professionals better serve consumers and help policy makers understand the areas of consumer finance that may be too complex for the ordinary american. nonetheless, researchers must be careful to distinguish between consulting a professional who holds the title of financial planner and engaging in financial planning as these are not 442 s.j. heckman et al. / financial services review 25 (2016) 427–446 necessarily one in the same. we suggest either clearly defining what is meant by a financial planner in the survey or perhaps avoid using the term to limit measurement error. using a set of questions to determine whether a financial planning engagement has occurred may be more informative than simply asking about financial planner use as it will capture some of the heterogeneity in financial planning engagements (e.g., subjects covered, depth of data collection, breadth of recommendations, etc.). table 4 example survey items for measuring professional financial planning engagements aspect of financial planning definition example survey question process, goals, management of resources, intent to engage do you consult a financial professional when managing your finances to achieve your financial goals? more than one subject area which personal finance subject areas do you discuss with a financial professional? select all that apply. financial statement preparation and analysis insurance planning and risk management education planning employee benefits planning investment planning income tax planning retirement planning estate planning comprehensiveness of data collection how much of your financial information does your financial professional collect before making recommendations? 1 (minimal) . . . 10 (everything) breadth of recommendations how thoroughly does your financial professional address the following areas? (same list as above) 1 (not addressed) . . . 10 (thoroughly addressed) other sources of heterogeneity diversity of professional training and professional expertise what kind of financial professional do you consult? select all that apply. accountant investment advisor attorney insurance agent investment broker financial planner financial counselor frequency and duration of professional engagement how long have you been working a financial professional? how often do you meet with a financial professional to discuss topics related to your financial goals? payment type of financial professional service how do you pay this financial planning service for? please check all that apply [commission (e.g., per transaction), hourly, monthly, or annual rate, flat fee, percentage fee (e.g., % of my account balance)] 443s.j. heckman et al. / financial services review 25 (2016) 427–446 this analysis demonstrates that existing measures of households’ financial planner use, especially in public, nationally representative datasets, lack content validity. as the field continues to grow, definitions of financial planning, measures of financial planner use, and the effects of engaging in professional financial planning on consumer well-being is of utmost importance. financial planning researchers, however, will be significantly limited until better measures of financial planner use are available. future research should focus on developing and testing questions and measures to address this gap in the field. additionally, researchers should be careful to consider other aspects of measurement, including reliability, criterion-related validity, and construct validity. notes 1 the purpose of this study not to answer this question, but rather to clearly articulate the measurement issues involved in pursuing research questions related to the influence of financial planner use. 2 the alp contains over 400 surveys that have been administered so this list of surveys may not be exhaustive. 3 as an example of work published elsewhere that has used the financial professional use question in the alp, parker, bruine de bruin, yoong, and willis (2012) investigated the relationship between confidence and financial planning. the researchers used three questions (1) “have you or your partner ever tried to figure out how your household would need to save for retirement?”; (2) “have you consulted a financial planner or advisor or an accountant?”; and (3) “have you or your partner developed a plan for retirement saving?” to create a single mean score reflecting what they refer to as a “retirement planning index.” 4 this was confirmed via personal correspondence with the authors. 5 park and yao created five categories of information sources: self and social network, financial planner, financial institutions, media, and other. the rationale for including lawyer and accountant with financial planner is that these professionals “often work as a team to assist financial planners in helping clients make saving and investment decisions” (p. 6). 6 fifty-four percent reported that the plan was prepared or updated in the last 12 months. 7 use of a financial coach was part of a field-experiment involving first-time homebuyers. 8 in the case of the alp and hrs, the evaluation focuses on measures that have been used by previous literature. additionally, we analyze alp survey 13 because of its unique measures. references balasubramnian, b., brisker, e. r., & gradisher, s. (2014). financial adviser background checks. financial services review, 23, 305–324. 444 s.j. heckman et al. / financial services review 25 (2016) 427–446 brandon jr., e. d., & welch, h. o. (2009). the history of financial planning: the transformation of financial services. hoboken, nj: john wiley & sons. carmines, e. g., & zeller, r. a. (1979). reliability and validity assessment. thousand oaks, ca: sage publications, inc. cfp board. (2013). standard of professional conduct (available at: https://www.cfp.net/docs/ethics-enforcement/ standards-of-professional-conduct-final.pdf?sfvrsn�2). collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. cummings, b. f., & james iii, r. n. (2014). factors associated with getting and dropping financial advisors among older adults: evidence from longitudinal data. journal of financial counseling and planning, 25, 129–147. devellis, r. f. (2012). scale development: theory and applications (3rd ed.). thousand oaks, ca: sage publications. eccles, d. w., ward, p., goldsmith, e., & arsal, g. (2013). the relationship between retirement wealth and householders’ lifetime personal financial and investing behaviors. journal of consumer affairs, 47, 432–464. doi:10.1111/joca.12022. elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217–231. gibson, r., michayluk, d., & van de venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. heckman, s. j., & hanna, s. d. (2015). individual and institutional factors related to low-income household saving behavior. journal of financial counseling & planning, 26, 187–199. hudson, c. r., & palmer, l. (2014). low-income employees: the relationship between information from formal advisors and financial behaviors. financial services review, 23, 25–43. kim, k., & hanna, s. d. (2015). do u.s. households perceive their retirement preparedness realistically? financial services review, 24, 139–155. knoll, m. a. z., & houts, c. r. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. journal of consumer affairs, 46, 381–410. doi:10.1111/j.17456606.2012.01241.x. kuzniak, s., rabbani, a., heo, w., ruiz-menjivar, j., & grable, j. e. (2015). the grable and lytton risktolerance scale: a 15-year retrospective. financial services review, 24, 177–192. lachance, m.-e., & tang, n. (2012). financial advice and trust. financial services review, 21, 209–226. martin jr., t. k., finke, m., & gibson, p. (2014). race, trust, and retirement decisions. journal of personal finance, 13, 62–71. martin jr., t. k., & finke, m. s. (2014). a comparison of retirement strategies and financial planner value. journal of financial planning, 27, 46–53. moulton, s., loibl, c., samak, a., & collins, j. m. (2013). borrowing capacity and financial decisions of low-to-moderate income first-time homebuyers. journal of consumer affairs, 47, 375–403. doi:10.1111/ joca.12021. park, e., & yao, r. (2015). financial risk attitude and behavior: do planners help increase consistency? journal of family and economic issues. advance online publication. doi:10.1007/s10834-015-9469-9. parker, a. m., bruine de bruin, w., yoong, j., & willis, r. (2012). inappropriate confidence and retirement planning: four studies with a national sample. journal of behavioral decision making, 25, 382–389. doi:10.1002/bdm.745. pricemetrix. (2013). stay or stray: putting some number behind client retention (available at: http://www. pricemetrix.com/cms/wp-content/uploads/pricemetrix-insights_stay-or-stray_english.pdf?t�1393101724). princeton survey research associates international. (2013). financial planning profiles of american households: the 2013 household financial planning survey and index (available at: http://www.cfp.net/docs/ public-policy/2013-fin-planning-profiles-of-amer-households.pdf). 445s.j. heckman et al. / financial services review 25 (2016) 427–446 robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. sass, s. a., belbase, a., cooperrider, t., & ramos-mercado, j. d. (2015). what do subjective assessments of financial well-being reflect? journal of personal finance, 14, 21–33. scott, j. k., & finke, m. s. (2013). the demand for disability insurance. financial services review, 22, 1–12. scott, j. k., & gilliam, j. (2014). boomers’ life insurance adequacy pre & post the 2008 financial crisis. financial services review, 23, 287–304. simms, k. (2014). investor profiles: meaningful differences in women’s use of investment advice? financial services review, 23, 273–286. society of actuaries. (2013). 2013 risks and process of retirement survey (available at: https://www.soa.org/ files/research/projects/research-2013-retirement-survey.pdf). stevens, s. s. (1951). mathematics, measurement, and psychophysics. in s. s. stevens (ed.), handbook of experimental psychology (pp. 1–49). new york, ny: wiley. tang, n., & lu, t. (2014). are your clients making the right loan choice? journal of financial planning, 27, 39–47. warschauer, t., & sciglimpaglia, d. (2012). the economic benefits of personal financial planning: an empirical analysis. financial services review, 21, 195–208. winchester, d. d., & huston, s. j. (2014). does a relationship with a financial service professional overcome a client’s sense of not being in control of achieving their goals? financial services review, 23, 1–23. zeller, r. a., & carmines, e. g. (1980). measurement in the social sciences: the link between theory and data. cambridge, ma: cambridge university press. zick, c. d., mayer, r. n., & kara, g. (2012). the kids are all right: generational differences in responses to the great recession. journal of financial counseling and planning, 23, 3–16. 446 s.j. heckman et al. / financial services review 25 (2016) 427–446 personality and borrowing behavior: an examination of the role of need for material resources and need for arousal traits on household’s borrowing decisions atefeh yazdanparasta, yasser alhenawib,* aassistant professor of marketing, mead johnson endowed chair in business, university of evansville, schroeder family school of business administration, evansville, in 47722, usa bassociate professor of finance, director of the institute of banking and finance, university of evansville, schroeder family school of business administration, evansville, in 47722, usa abstract this research is an empirical examination of the role of psychological characteristics of household decision makers in their borrowing decisions. using a unique household survey data that ties together relevant concepts from the survey of consumer finances (scf) and psychological and attitudinal literature, we obtain direct measures of each surveyed household’s personality scores, relevant attitudes, and financial profiles. following regression analysis, we examine the relationship between attitude towards borrowing and intentions to apply for specific borrowing options and inspect the role of personality traits in such decisions. specifically, we focus on tow personality traits, the need for material resources and the need for arousal, which have been largely ignored in extant household finance literature. our findings indicate that the attitude toward borrowing and the intention to borrow are not always consistent and, more interestingly, the discrepancies between the two vary across personalities, highlighting the role of personality traits in borrowing decisions. specifically, while the positive relationship between attitude towards borrowing and intention to borrow is intuitive, we show that this relationship is trivial for individuals who score low on need for material resources and individuals with low degrees of need for arousal. in contrast, for individuals with higher levels of need for material resources and need for arousal, the positive association between attitude towards borrowing and the intention to borrow is significantly intensified. further, our results suggest that borrowing options are not homogenous and are motivated differently. consistent with the view that individuals with greater need for material resources consider quantity and quality of possessions as the criteria to judge personal success, we find that these individuals have stronger intentions for mortgages, home improvement loans, business loans, personal loans, and payday loans. similarly, con* corresponding author. tel.: �1-504-208-7749; fax: �1-812-488-2872. e-mail address: ya22@evansville.edu (a. yazdanparast) financial services review 26 (2017) 55–85 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. sistent with the view that individuals with greater need for arousal favor stimulation and excitement that is often associated with more spending, we find that the borrowing decisions of these individual are marked with stronger intentions for home improvement loans, business loans, student loans, personal loans, and payday loans. the intentions for credits cards and automotive loans seem to be independent from the borrower’s personality. finally, we report strong evidence that personal attitudinal biases towards money, risk, financial planning, and borrowing as well as certain demographic characteristics influence household’s borrowing behavior. our work contributes to the literature on household finance in several ways. first, we focus on two under-researched personality traits to examine their roles in household borrowing decisions. second, this research recognizes the distinction between the attitude towards borrowing and the intention to borrow and examines the inconsistencies between them. third, we consider a wide spectrum of borrowing options that differ in terms of risk, motivation, and loan maturity. © 2017 academy of financial services. all rights reserved. jel classification: d03; d14 keywords: personality; household finance; financial decision; borrowing attitude; borrowing intention 1. introduction household finance is an emerging field of research (brown & taylor, 2011; campbell, 2006) that has attracted scholars from finance, economics, management, communication, and marketing (e.g., besharat et al., 2014, 2015; duclos, 2015; lynch, 2011; and shefrin & nicols, 2014). this growing and diverse interest is perhaps attributed to the far-reaching impacts of household’s financial decisions on consumer welfare, marketing practices, financial service providers’ strategies, and the stability and reliability of the financial system (lynch, 2011; priog & roberts, 2007). moreover, extant research has recognized the fact that financial decisions are influenced not only by household’s financial profile but also by decision makers’ personality, habits, and other individual characteristics (lynch, 2011). not surprisingly, a growing area of household finance research examines psychological factors impacting household financial decisions (see guiso et al., 2002 for a comprehensive review). our work belongs to this area of literature and focuses on how individual characteristics impact household’s borrowing choices. since borrowing behavior presents a specific context of consumer decision making, it is important to study the factors that impact borrowing decisions. theory of planned behavior, a well-known theory in psychology (e.g., ajzen & fishbein 1980) maintains that behaviors are influenced by attitudes and intentions. in fact, attitudes can influence behavior directly, as well as indirectly through intentions (e.g., bentler & speckart, 1981, hurst & mendoza, 2016). attitudes are generally defined as the lasting (favorable or unfavorable) evaluations toward an object, action, and so forth, while an intention is a particular type of volition that transforms the psychological state into guided bodily responses (bagozzi, baumgartner, & yi, 1989; fishbein & ajzen, 1975, 1980). therefore, the more favorable the attitudes, the higher the intentions for the behavior. however, extant research provides evidence for inconsistency between attitude and intention (see jonas, diehl, & bromer, 1997). after all, we may have a positive attitude toward performing some act but fail to form an intention or 56 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 intend to refrain from doing the action because of some non-attitudinal reason (bagozzi et al., 1989). the extant literature on household finance does not make a clear a distinction between the attitude towards borrowing and the intention to borrow and possible inconsistencies between them as related to decision making. we argue that individuals’ attitudes may not be congruent with the actual intentions to take certain borrowing actions. specifically, we theorize that the relationship between the attitude toward borrowing and the intention to borrow, and the mechanism thereof, is influenced by personality characteristics of the decision maker. the relationship between household finances and personality traits is in fact an understudied area despite the fact that personality traits can influence financial decision-making at the individual and household level (brown & taylor, 2011). previous research has mainly used the well-known big five model of personality (mccrae & costa, 1987) and examined the influence of one, or a combination, of the five personality traits (i.e., openness to experience, conscientiousness, extroversion, agreeableness, and neuroticism) on financial decisions. we extend the extant research by examining the role of two additional personality traits, the need for material resources and the need for arousal, that have been largely ignored by extant literature on household finance. recent research indicates that individuals with greater need for material resources are relatively more concerned with their social image (christopher, marek, & carroll, 2004) and view possessions as means of achieving utilitarian and social status (richins & dawson, 1992). also, individuals who score higher on the need for arousal trait have been shown to have a chronic need for impulse purchases (d’astous, maltais, & roberge, 1990; rook, 1987). therefore, we theorize that these two traits are related to individuals’ financial preferences and, consequently, borrowing choices.1 finally, extant literature often defines borrowing either too broadly or too narrowly. the former approach overlooks the possibility that borrowing options are not homogenous (as in brandstatter, 1996; davey & george, 2011; and nyhus & webley, 2001). the latter approach focuses on a single borrowing option (e.g., credit card usage as in norvilitis et al., 2006 and lee & kwon, 2002) and, therefore, does not offer a side-by-side comparison of different categories of borrowing in a unified framework. unlike extant literature, we consider a wide spectrum of borrowing options that differ in purpose and weigh differently on the risk and maturity scales. this research uses a representative sample of primary decision makers in u.s. households (the final sample includes 849 responses corresponding to 85% response rate) and measures respondents’ attitudes toward financial issues as well as their intentions to apply for eight different categories of loans including credit cards, mortgage, home improvement, business loans, car loans, student loans, personal loans, and payday loans. respondents’ financial knowledge, risk tolerance, and personality characteristics are also measured.2 our findings suggest that the attitude towards borrowing and the intention to borrow do not necessarily accord, highlighting the role of personality characteristics in such behaviors. we also report strong evidence that different borrowing options are motivated differently. individuals with different personalities demonstrate borrowing patterns that reflect their psychological needs. individuals with greater need for material resources have stronger intention to borrow in certain categories that lead to possessing materials. similarly, individuals with greater need for arousal have a higher tendency to use borrowing options 57a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 associated with spending. finally, the results show that attitudinal biases and predispositions (e.g., attitude toward planning and borrowing) as well as past borrowing behaviors play a significant role in household’s borrowing decisions. our work contributes to the extant research in several ways. first, our results suggest that the attitude toward borrowing and the intention to borrow should not be used interchangeably. second, our work indicates that the need for material resources and the need for arousal significantly influence the way in which attitudinal biases translate into intentions to borrow. third, our findings suggest that the aggregation of borrowing behaviors may lead to distorted conclusions, because borrowing options are not homogenous. theoretically, our work provides general support to the behavioral approach to explaining financial decision making, suggesting that psychological factors and individual biases can swerve financial decisions in a systematic manner. 2. literature review and hypotheses development 2.1. personality traits according to lin (2010), personality traits are the dynamic organization of psycho physiological systems that make up a person’s characteristic behavior, thoughts, and feelings. one of the recent approaches in understanding the impact of personality traits on behavior is the meta-theoretic model of motivation and personality (i.e., the 3m model).3 this model accounts for how personality traits interact with the situation to influence consumer attitudes and actions (mowen, 2000). the model incorporates a hierarchical theory of personality and stipulates that personality traits are at one of four levels (i.e., elemental, compound, situational, and surface). elemental traits are the focus of the present study because of their fundamental nature in individual differences (mowen & carlson, 2003). these traits are basic predispositions that arise from genetic endowment and early learning. at the elemental level, the 3m model contains five traits from the big five personality model (i.e., openness to experience, conscientiousness, extroversion, agreeableness, and neuroticism5) as well as three additional traits namely, need for material resources, need for arousal, and body resource needs (mowen et al., 2007). these traits are used as screening tools, because they can explain the variance in performance, are stable over time, and generalize across groups and settings (mowen & carlson, 2003). 2.2. personality traits and borrowing behavior the extant literature on the relationship between borrowing behavior and personality traits defines borrowing broadly as any behavior that constitutes using others’ wealth to satisfy current consumption needs. for instance, davey and george (2011) found that individuals who score higher on openness have a tendency to try different or new products and services, which implies stronger intentions to borrow money to satisfy this desire. moreover, they might be less hesitant to take on a loan as they see it as a new experience. harley and wilhelm (1992) showed that highly conscientious individuals are self-disciplined and, thus, 58 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 are more likely to regularly save compared to carefree people. as a result, these individuals make planned decisions and are less likely in need of borrowing. the work of sadi, ghalibaf, rostami, gholipour, and gholipour (2011) showed that extroverts focus on external elements, decide easily, and live in present time. further, extroversion is related to the need for stimulation and social contact which affects spending behavior and consequently, increases the dependency to borrow (davey & george, 2011). similarly, extraversion is negatively related to the percentage of income that is saved and positively with likelihood of being in debt and relying on credit cards (davey & george, 2011). nyhus and webley (2001) argued that highly agreeable people have fewer investments and need to borrow more. finally, and since shopping and spending are considered as means of temporary mood repair, lack of emotional stability (i.e., high neuroticism) may lead individuals to engage in behaviors such as impulse buying that could bring them short-term gratification (youn & faber, 2000). lack of self-control has been associated with a variety of personal and social problems, including overspending (mansfield, pinto, & parente, 2003), impulsive spending (baumeister & exline 2000; strayhorn, 2002), and compulsive buying (mowen, 2000). consequently, individuals scoring high on neuroticism are less likely to exhibit self-control and save less (brandstatter, 1996; davey & george, 2011). moreover, brown and taylor’s (2011) used individual level data drawn from the british household panel survey and analyzed the influence of big five personality traits on financial decisions regarding unsecured debt acquisition and financial assets and found that personality traits have different effects across the various types of debt and assets held. even though, these findings provide invaluable insights regarding the impact of personality traits on financial behaviors of individuals, there is a gap in extant literature focusing on the impact of personality traits on specific borrowing behaviors. as such, a more recent stream of research, to which this study belongs, has emerged which recognizes that borrowing actions are not similarly motivated and provides more specificity with regards to various borrowing behaviors. 2.3. hypothesis development overall, the review of extant literature identifies three major research gaps. first, the extant research overlooks the fact that borrowing options differ vastly in terms of borrowing horizon and purpose of borrowing and, thus, they may be motivated differently. the works of besharat et al. (2014, 2015) and brown and taylor’s (2011) present a significant attempt to overcome this shortcoming but lack comprehensiveness, as they focus on a single aspect of borrowing such as credit cards and unsecured debt acquisition. in contrast, the present research analyzes a wide spectrum of borrowing options and considers long-term borrowing (i.e., borrowing for big-ticket items that often require financing such as mortgage), intermediate-term borrowing (e.g., automobile and student loans), and short-term repeated borrowing decisions associated with smaller purchases (e.g., the use of payday loans, credit cards, and store cards). the study also encompasses borrowing behaviors induced by asset acquisition (e.g., home and car), consumptions (e.g., credit cards), educational needs (e.g., student loans), and investment (e.g., business loans). we also examine contrasting borrowing options that imply varying levels of sophistication in financial behavior and knowledge (e.g., 59a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 the process of granting business loans is more rigorous than the ad-hoc process of granting payday loans). the work of duclos (2015) indicates that the significance of a financial decision is attributed not only to the size of the deal (e.g., mortgage) but also to the accumulated effects of repeated decisions (e.g., credit card usage). accordingly, we explore a wide range of borrowing options that are inherently different in risk, commitment level, term, and purpose. second, the extant literature does not consider the additional traits that have been recently added to the 3m model (i.e., the need for material resources and the need for arousal).5 the need for material resources is the general need for possessing material and accumulating wealth. this trait stems from the importance one attaches to worldly possessions (belk, 1984). for individuals with greater need for material resources, happiness is associated with possessions. therefore, they have a more positive attitude toward debt (pinto, parente, & palmer, 2000; pirog & roberts, 2007), and they are more willing to take on greater debts (ponchio & aranha, 2008). this personal urge could make a materialistic individual more willing to carry the burden of a loan to satisfy his/her psychological needs. therefore, controlling for income and wealth, we anticipate to find that those who score higher on this trait exhibit a stronger intention to borrow compared to those who score lower. furthermore, the need for arousal is the general need or desire for stimulation and excitement and countering fear (licata, mowen, harris, & brown, 2003; mehrabian & russell, 1974). individuals who score higher on this trait have a chronic need for stimulation and instant gratification. because purchase is frequently cited as a stimulating behavior because of its power to satisfy urges for goods, services, experiences, and status (d’astous et al., 1990; rook, 1987), we anticipate that the need for arousal trait is associated with stronger intention to borrow (after controlling for income and wealth). as such, it is hypothesized that: hypothesis 1a: greater need for material resources is associated with a stronger intention to borrow. hypothesis 1b: greater need for arousal is associated with a stronger intention to borrow. third, the extant literature does not make a clear distinction between attitudes and intentions. this may lead to dubious conclusions because a person’s attitude toward an action may not be consistent with his intention to take an action (ajzen & fishbein, 1977). more specifically, attitude toward a behavior is defined as the lasting evaluations of the action, but the ultimate behavior is best captured by behavioral intentions which are in turn influenced by many factors including the attitude toward the behavior (fishbein & ajzen, 1975, 1980). further, attitudes are individual factors mainly associated with the amount of affect or feeling for or against something while intentions are influenced by situational, individual, and elements of marketing stimuli communicated with consumers. therefore, the distinction between the two is theoretically and empirically important when investigating behaviors (fishbein & ajzen, 1975; sheppard, hartwich, & warshaw, 1988). in the context of this research, we hypothesize that: hypothesis 2: the intention to borrow is not always consistent with the borrower’s attitude towards borrowing. 60 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 furthermore, research has indicated that personality characteristics predict enduring tendencies to engage in general classes of behavior (mowen, 2000). therefore, we anticipate that the relationship between attitude towards borrowing and the intention to borrow is moderated by personality traits. accordingly, we test the following hypotheses: hypothesis 3a: the relationship between attitude towards borrowing and the intention to borrow is stronger for individuals with greater need for material resources. hypothesis 3b: the relationship between attitude towards borrowing and the intention to borrow is stronger for individuals with greater need for arousal. the following graph illustrates the connections between the hypotheses: 3. data collection and methods 3.1. procedure data were collected through an online survey administrated by qualtrics on amazon’s mechanical turk (mturk; buhrmester et al., 2011; rand, 2012). mturk provides behavioral researchers with representative consumer samples because mturk samples tap more diverse populations and yield greater generalizability (buhrmester et al., 2011; goodman et al., 2013; mason & suri 2012). participation was restricted to u.s. citizens or permanent residents who were at least 18 years old and declared to be one of the primary financial decision makers in their household to ensure reliability of responses (see campbell, 2006 and alhenawi & elkhal, 2013). upon qualification, participants responded to four questions that measured enduring involvement in personal finance matters (bloch, sherrell, & ridgway, 1986) followed by questions about attitude toward money christopher et al. (2004) and financial knowledge (knoll & houts, 2012). next, risk tolerance was measured using items from the well-known investment risk tolerance quiz (grable & lytton, 1999). this quiz is shown to be a reliable and valid measure of risk tolerance (e.g., gilliam, chatterjee, & grable, 2010; larkin, lucey, & mulholland, 2013) to demonstrate the maximum amount of uncertainty that individuals are willing to accept when making financial decisions (grable, 2000). personality traits were measured using measures from licata, mowen, harris, and brown (2003) that were first developed by mowen (2000). respondents were presented with short 61a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 phrases and asked “how often do you feel/act this way” (responses were taken on 5-point scales anchored by 1 � never to 5 � always). attitude towards planning was also measured to control for respondents’ perceptions of financial planning (yamauchi & templer, 1982). borrowing behavior was captured using questions from the survey of consumer finances (scf) administered by the federal reserve—a reliable resource that is largely overlooked in academic research (campbell, 2006). this instrument does not only gauge overall borrowing intentions, but also allows for differential analyses of several types of borrowing options. the scf identifies eight categories of borrowing avenues: credit cards, mortgage, home improvement, business loans, car loans, student loans, personal loans, and payday loans. attitude toward borrowing was measured using seven items adapted from the financial consumer agency of canada. these items measure individuals’ overall views about borrowing using the 5-point likert scale (1 � strongly disagree; 5 � strongly agree). the last sections of the questionnaire contained questions related to financial profile. previous research has shown that financial profile measures are important in the study of household financial behavior. for instance, johnson and li (2010) found that a household with a high debt service ratio is significantly more likely to be turned down for credit than other households. similarly, the work of alhenawi and elkhal (2013) indicated that selfassessment of one’s financial aptitude greatly influences financial behavior. accordingly, financial profile questions included questions on major events with dramatic changes on household financial behavior over that past two years,6 home ownership status, credit score category, having or not having a professional financial planner, confidence in making the best choices in managing money, and respondents’ description of their financial situation (i.e., living comfortably, meeting basic expenses with a little left over for extras, etc.). demographic questions included standard controls such as gender, age, income level, marital status, zip code of the residence, household size, ethnicity, education, and so forth. 3.2. sample overall, we collected 1,000 responses, out of which 849 were used in the analyses after eliminating respondents who did not have the required qualifications, had incomplete responses, or did not pass the instructional manipulation check question included to gauge whether participants paid sufficient attention to the instructions (goodman et al., 2013). table 1 provides characteristics of the sample. panel a of table 1 shows respondents’ scores on personality dimensions. panel b shows responses to attitudinal tests, self-assessments, and other personal factors quizzes. respondents were highly involved in personal finance matters, had relatively favorable attitudes toward money and financial planning (79%), and had an average attitude toward borrowing. the average percentage score on the risk tolerance was 49.40% which is comparable to other u.s.-based studies (e.g., ryack, 2011).8 the mean percentage score on the financial knowledge quiz was 59.57% indicating an overall deficiency in financial knowledge at the household level which is consistent with the major belief in the literature (see, e.g., campbell, 2006). similarly, the sample mean on the financial planning quiz was 60.13%, indicating poor financial planning. this is in line with previous research findings (see, e.g., lusardi & mitchell, 2008). 62 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 table 1 sample characteristics cronbach’s � min max mean dev. obs. % panel a: personality traits introversion .89 1.00 5.00 3.00 0.94 849 materialism .90 1.00 5.00 2.53 0.93 849 openness .87 1.00 5.00 3.41 0.80 849 neuroticism .92 1.00 5.00 2.47 0.96 849 agreeableness .90 1.00 5.00 3.72 0.71 849 conscientiousness .87 1.25 5.00 3.61 0.78 849 need for arousal .88 1.00 5.00 2.52 0.86 849 panel b: attitudes, self-evaluations, and quiz scores involvement in personal finance .79 1.00 5.00 4.09 0.76 849 attitude towards money .60 3.00 7.00 4.91 0.63 849 attitude towards financial planning .87 0.20 1.00 0.79 0.13 849 attitude towards borrowing .73 0.20 0.89 0.45 0.14 849 risk tolerance quiz % 26.53% 81.63% 49.40% 9.99% 849 financial knowledge quiz % 0.00% 100.00% 59.57% 21.69% 849 financial planning quiz % 20.00% 100.00% 60.13% 17.74% 849 self-assessment of financial situation don’t know 6 0.71% comfortable 142 16.73% make little more than needed 366 43.11% make just enough 249 29.33% make less than needed 86 10.13% confidence in making financial decisions don’t know 14 1.65% very 111 13.07% somewhat 379 44.64% not too much 226 26.62% not confident at all 119 14.02% panel c: borrowing behavior have credit card and pay on time 361 42.52% have credit card and carry balance 299 35.22% have mortgage account 250 29.45% refinanced home 97 11.43% obtained home-equity loan 41 4.83% obtained home-improvement loan 33 3.89% have business loan 14 1.65% have car loan 261 30.74% have student loan 342 40.28% have personal loan 139 16.37% have pay-day loan 21 2.47% applied for credit in past five years 454 53.47% filed for bankruptcy 67 7.89% make payments on time 631 74.32% likely to default in the future 56 6.60% likely to file bankruptcy in the future 35 4.12% panel d: borrowing intentions credit card 1.00 5.00 2.18 1.30 849 mortgage 1.00 5.00 2.05 1.29 849 home improvement loan 1.00 5.00 1.60 0.94 849 business loan 1.00 5.00 1.58 0.95 849 car loan 1.00 5.00 2.45 1.32 849 student loan 1.00 5.00 1.81 1.25 849 personal loan 1.00 5.00 1.59 0.96 849 payday loan 1.00 5.00 1.30 0.75 849 panel e: household financial profile dramatic event last 2 years? yes/no 326 38.40% own house 412 48.53% income income category 1 10 1.18% income category 2 56 6.60% income category 3 197 23.20% income category 4 228 26.86% income category 5 135 15.90% income category 6 108 12.72% (continued on next page) 63a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 the last two parts of panel b list respondents’ self-evaluations of their financial situation and ability to make financial decisions (adapted from a survey used by the consumer federation of america (2013). an interestingly find was that 57.73% (13.07% � 44.64%) of participants had average to high confidence in their ability to make financial decisions while table 1 (continued) cronbach’s � min max mean dev. obs. % income category 7 60 7.07% income category 8 38 4.48% income category 9 6 0.71% income category 10 11 1.30% credit score don’t know 139 16.37% poor 130 15.31% fair 261 30.74% good 319 37.57% panel f: demographics age 18.00 82.00 34.04 11.51 849 100.00% gender 0.00 1.00 0.45 0.50 849 100.00% family size 1.00 6.00 2.55 1.27 849 100.00% financial education 157 18.49% working in the financial sector 85 10.01% marital status married 280 32.98% single 350 41.22% divorced 56 6.60% widowed 6 0.71% live with partner 157 18.49% ethnicity white 652 76.80% african american 65 7.66% hispanic 49 5.77% american indian 5 0.59% asian 58 6.83% other 9 1.06% prefer not to answer 11 1.30% education less than high school 4 0.47% high school 105 12.37% some college 239 28.15% associate degree 81 9.54% bachelor degree 314 36.98% masters 86 10.13% doctorate 20 2.36% notes: data were collected through an online survey administrated by qualtrics on amazon’s mechanical turk (mturk). participation is restricted to u. s. citizens or permanent residents who are at least 18 and who reported that they were one of the primary financial decision makers in their households. qualified participants responded to quizzes on enduring financial involvement (bloch, sherrell, & ridgway, 1986); attitude toward money (christopher, marek, & carroll, 2004); financial knowledge (knoll & houts, 2012); attitude towards financial planning (yamauchi & templer, 1982); attitude towards borrowing (financial consumer agency of canada) and risk tolerance (grable & lytton, 1999). each participant was assigned a percentage score based on each of the quizzes. this score represents the ratio of his/her total score in the quiz to the highest possible score. personality traits were measured using short quizzes adapted from licata, mowen, harris, and brown (2003). borrowing behavior questions were adapted from the survey of consumer finances (scf) administered by the federal reserve. the scf identifies eight categories of borrowing avenues: credit cards, mortgage, home improvement, business loans, car loans, student loans, personal loans, and payday loans. this classification is appropriate because it captures a great deal of the variations within the numerous borrowing options available to households (see instrument section for more explanation). the questionnaire also includes sections that explored the financial profile and demographic characteristics of the participants. the final sample includes 849 cases. 64 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 the scores in financial knowledge and planning quizzes were below average. this indicates that raising financial literacy may require not only education, but also changing financial perceptions and attitudes (see alhenawi & elkhal, 2013 for similar argument). panel c captures actual borrowing behaviors. as indicated in the table, 42.52% of respondents owned a credit card and paid on time (i.e., they did not have credit card debt). in contrast, 35.22% were indebted to credit card companies (i.e., they carried a monthly balance). the percentages in the table indicate that the most common types of borrowing (excluding credit cards) were student loans (40.28%), car loans (30.74%), mortgage loans (29.45%), and personal loans (16.37%). other types of loans such as home equity loans, home improvement loans, payday loans, and business loans were much less popular (less than 5%). to control for borrowing ability, we also included additional five questions on credit applications and bankruptcy (adopted from the scf). we found that 53.47% of respondents had applied for credit in the past five years, 74.32% paid their loan payments on time, and only 6.6% anticipated defaulting in the future. also, 7.89% of respondents had filed for bankruptcy in the past, and 4.12% anticipated filing for bankruptcy in the future. panel d demonstrates participants’ intentions to use different types of borrowing options (on a scale of 1 to 5, where 1 � very unlikely and 5� very likely). the demand for car loans was the highest (2.45), followed by credit cards (2.18), and homes (2.05). personal loans (1.59), business loans (1.58), and payday loans (1.30) had lower scores. panels e and f demonstrate the financial and non-financial characteristics of the sample. about 38.40% of respondents reported that they endured a significant event in the past two years that had changed their financial behavior and attitudes. the sample was reasonably balanced across various income categories, credit scores classes, marital status, ethnicity, and education levels. average respondents’ age was 34.04 years with a minimum of 18 and a maximum of 82 years. about 45% of respondents were males. average family size was 2.55 with a minimum of one (single) and a maximum of six. about 18.49% of respondents had formal financial education, and about 10% indicated that they worked in the financial sector. overall, the sample was balanced and consistent with previous studies (e.g., zharvalhenawi & elkhal, 2013; hilgert, hogarth, & beverly, 2003; and johnson & li, 2010). 3.3. regression analysis we used multivariate regression models where the regressors are explanatory variables— including personality traits—theorized to influence the regressand. table 2 provides the results of the regression analyses. the regressand in regression one is the attitude toward borrowing while the regressands in regressions two to nine are the intentions to use eight different types of borrowing. in each borrowing intention regression, we included binary variables that indicate historical borrowing behavior in the same category (e.g., in the mortgage regression model, we include three variables that indicate relevant historical borrowing behaviors such as having an existing mortgage loan, refinanced mortgage, or home equity loan). the models also include standard controls such as financial profile and demographics. we first estimated full models with all possible regressors, then we ran concise models with significant coefficients only for better parsimony. we found negligible 65a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 t ab le 2 r eg re ss io n an al ys is –b or ro w in g in te nt io n (1 ) a tti tu de (2 ) c re di t ca rd (3 ) m or tg ag e (4 ) h om e im pr ov em en t lo an (5 ) b us in es s lo an (6 ) a ut om ob ile (7 ) st ud en t lo an (8 ) pe rs on al lo an (9 ) pa yd ay lo an c on st an t 57 .7 39 ** * 0. 50 1 0. 25 1 1. 99 1* ** 2. 01 0* ** 1. 14 0* 2. 43 4* ** 1. 57 5* ** 2. 59 2* ** in tr ov er si on 0. 11 2 0. 09 9* * 0. 07 0 � 0. 02 4 � 0. 00 7 0. 00 1 0. 05 2 0. 01 2 0. 00 8 m at er ia lis m 2. 21 3* ** 0. 13 6* ** 0. 16 2* ** 0. 12 3* ** 0. 10 5* ** 0. 10 3* 0. 16 2* ** 0. 05 3 0. 03 4 o pe nn es s � 0. 38 5 � 0. 00 5 � 0. 08 8 0. 00 7 0. 03 6 0. 00 7 � 0. 06 5 0. 02 9 0. 00 7 n eu ro tic is m 0. 55 8 � 0. 02 6 � 0. 02 4 0. 02 8 0. 01 9 0. 07 0 0. 15 4* ** 0. 06 3* 0. 03 8 a gr ee ab le ne ss � 1. 31 9* 0. 04 8 0. 15 1* * � 0. 01 4 0. 03 2 0. 05 7 0. 02 4 0. 02 7 � 0. 10 2* ** c on sc ie nt io us ne ss � 1. 13 7* � 0. 03 5 0. 05 3 � 0. 06 5 � 0. 06 5 0. 02 5 � 0. 08 9 � 0. 01 2 0. 00 0 n ee d fo r ar ou sa l 1. 30 0* * 0. 17 5* ** 0. 17 8* ** 0. 08 2* * 0. 12 7* ** 0. 10 5* 0. 19 3* ** 0. 08 5* * 0. 05 4* b or r s ta t a pl d 3. 52 0* ** 0. 15 1 0. 23 6* * � 0. 02 8 � 0. 12 3* � 0. 02 8 � 0. 05 0 � 0. 06 2 � 0. 12 6* ** b or r s ta t b nk r pt 0. 90 6 � 0. 05 9 0. 41 1* * 0. 03 3 � 0. 01 7 � 0. 18 6 � 0. 20 4 0. 13 1 0. 12 2 b or r s ta t o nt im e � 2. 67 0* * � 0. 22 2* � 0. 05 9 � 0. 21 5* * � 0. 30 7* ** � 0. 08 8 � 0. 18 4 � 0. 32 8* ** � 0. 30 4* ** b or r s ta t d efl t 0. 72 0 � 0. 13 8 0. 07 4 � 0. 22 5 � 0. 01 1 � 0. 21 7 � 0. 21 7 � 0. 05 4 0. 04 2 b or r s ta t b nk r pt f ut ur e � 3. 96 6 0. 00 3 0. 07 1 � 0. 29 8* � 0. 05 4 � 0. 02 9 � 0. 18 5 � 0. 06 4 0. 09 8 a tt p ln g � 0. 19 9* ** 0. 00 6 0. 00 4 0. 00 0 � 0. 00 2 0. 00 8* * � 0. 00 1 � 0. 00 2 � 0. 00 3* f in q ui z 0. 01 6 � 0. 00 2 � 0. 00 2 � 0. 00 5* ** � 0. 00 5* ** 0. 00 1 � 0. 00 7* ** � 0. 00 4* * � 0. 00 7* ** o w nh ou se 0. 10 6 � 0. 11 4 � 0. 57 2* ** 0. 00 5 � 0. 13 1* * � 0. 14 7 � 0. 22 7* * � 0. 15 5* * � 0. 03 1 in co m e 0. 46 9 � 0. 04 9 0. 04 0 0. 00 7 � 0. 00 1 � 0. 04 1 0. 00 4 � 0. 02 1 � 0. 01 9 c rd t s co re 0. 02 1 � 0. 04 1 � 0. 02 7 0. 00 7 � 0. 03 0 � 0. 09 0* � 0. 02 1 � 0. 04 3 � 0. 02 9 f in a dv is or � 2. 01 7 0. 03 2 � 0. 11 3 � 0. 23 9* ** � 0. 22 1* * � 0. 00 1 � 0. 14 8 � 0. 22 3* * � 0. 25 1* ** se lf e va l � 1. 69 4* ** � 0. 03 9 � 0. 03 7 � 0. 03 7 � 0. 04 3 � 0. 07 1 � 0. 00 5 � 0. 04 6 � 0. 04 7 c on f 0. 36 8 � 0. 12 5* * � 0. 03 7 � 0. 07 6* � 0. 07 7* � 0. 03 0 � 0. 13 9* ** � 0. 03 6 � 0. 08 6* ** e du 1. 03 8* ** � 0. 07 3* * 0. 05 7* � 0. 02 3 � 0. 01 9 � 0. 06 6* � 0. 10 3* ** � 0. 04 4* � 0. 01 8 a tt b o rr o w 0. 02 9* ** 0. 00 6* 0. 00 7* ** 0. 00 9* ** 0. 01 3* ** 0. 00 1 0. 01 6* ** 0. 00 9* ** b or r h is t 0. 34 3* ** b or r h is t 0. 02 9 b or r h is t 0. 01 7 b or r h is t 0. 06 9 b or r h is t � 0. 16 1 b or r h is t 0. 38 6* * b or r h is t 0. 52 1* * b or r h is t 0. 26 5* ** b or r h is t 0. 31 6* ** b or r h is t 0. 33 4* ** b or r h is t 1. 32 8* ** r 2 0. 17 9 0. 19 1 0. 12 7 0. 12 1 0. 14 7 0. 07 0 0. 15 4 0. 19 2 0. 30 1 a dj . r 2 0. 15 8 0. 16 8 0. 10 0 0. 09 6 0. 12 3 0. 04 4 0. 13 0 0. 16 9 0. 28 2 df . r eg re ss io n 84 8 84 8 84 8 84 8 84 8 84 8 84 8 84 8 84 8 (c on ti nu ed on ne xt pa ge ) 66 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 t ab le 2 (c on tin ue d) (1 ) a tti tu de (2 ) c re di t ca rd (3 ) m or tg ag e (4 ) h om e im pr ov em en t lo an (5 ) b us in es s lo an (6 ) a ut om ob ile (7 ) st ud en t lo an (8 ) pe rs on al lo an (9 ) pa yd ay lo an df . r es id ua l 82 7 82 4 82 3 82 5 82 5 82 5 82 5 82 5 82 5 fst at e 8. 58 0 8. 11 8 4. 78 1 4. 92 9 6. 17 4 2. 70 4 6. 51 0 8. 50 9 15 .4 81 si g. 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 n ot es : t he ta bl e sh ow s th e re su lts of ei gh t re gr es si on s th at co rr es po nd to th e ei gh t di ff er en t bo rr ow in g op tio ns em pl oy ed in th is st ud y (r eg re ss io ns 2 th ro ug h 9) .i ta ls o sh ow s a re gr es si on w he re th e de pe nd en tv ar ia bl e is th e at tit ud e to w ar d bo rr ow in g (r eg re ss io n 1) .w e fir st es tim at e fu ll m od el s th at in cl ud e al lp os si bl e ex pl an at or y va ri ab le s. t he n, w e sh or te n th e m od el s to in cl ud e si gn ifi ca nt va ri ab le s on ly .w e fo un d th e di ff er en ce s be tw ee n th e ex pa nd ed m od el s an d th e co nc is e m od el s to be tr iv ia l. t he re fo re , w e on ly re po rt th e fin di ng of th e co nc is e m od el s w hi ch ha ve th e fo llo w in g ge ne ra l fo rm : b o rr in tn � c � in tr v � m a tr � o p n s � n eu ro � a g re e � c o n s � a ro u s � b o rr s ta t a p ld � b o rr s ta t b n kr p t � b o rr s ta t o n ti m e � b o rr s ta t d efl t � b o rr s ta t b n kr p tf u tu re � a tt p ln g � f in q u iz � o w n h o u se � in co m e � c rd t s co re � f in a d vi so r � s el fe va l� c o n f� e d u � a tt b o rr o w � b o rr h is t b or ri nt n is th e bo rr ow in g in te nt io n va ri ab le th at ca pt ur es th e re sp on de nt s’ in te nt io n to us e a ce rt ai n ty pe of bo rr ow in g. in tr v, m at r, o pn s, n eu ro ,a gr ee ,c on s, a ro us re pr es en ts ev en pe rs on al ity tr ai ts .b or r s ta t a pl d,b or r s ta t b nk r pt ,b or r s ta t o nt im e,b or r s ta t d efl t,a nd b or r s ta t b nk r pt f ut ur e ca pt ur e w he th er or no tt he re sp on de nt ap pl ie d fo r cr ed it in th e pa st fiv e ye ar s, w he th er or no tt he y fil ed fo r ba nk ru pt cy in th e pa st ,w he th er or no tt he y m ak e pa ym en ts on tim e, w he th er or no tt he y ar e lik el y to de fa ul to n a lo an in th e fu tu re ,a nd w he th er or no tt he y ar e lik el y to fil e fo r ba nk ru pt cy in th e fu tu re .a tt p ln g ca pt ur es re sp on de nt s’ at tit ud e to w ar ds pl an ni ng .f in q ui z re pr es en ts th e sc or e on th e kn ow le dg e qu iz .o w nh ou se is a di ch ot om ou s th at eq ua ls 1 if th e re sp on de nt ow ns a ho m e; 0 ot he rw is e. in co m e is a hi er at ic al va ri ab le re pr es en tin g ni ne le ve ls of ho us eh ol d in co m e (r ef er en ce gr ou p is “n o in co m e” ). c rd t s co re re pr es en ts th re e le ve ls of cr ed it sc or e qu al ity (g oo d, fa ir , an d po or ). f in a dv is or is a di ch ot om ou s th at eq ua ls 1 if th e ho us eh ol d us es th e se rv ic e of a pr of es si on al fin an ci al ad vi so r; 0 ot he rw is e. se lfe va lr ep re se nt s re sp on de nt s’ se lf -e va lu at io n of th ei r fin an ci al si tu at io n (fi na nc ia lly co m fo rt ab le ,m ak e a lit tle m or e th an ne ed s, m ee tb as ic ne ed s, an d m ak e a lit tle le ss th an ne ed s) w he re th e re fe re nc e ca te go ry is “p oo r.” c on f re pr es en ts re sp on de nt s’ co nfi de nc e in m ak in g fin an ci al de ci si on s (v er y co nfi de nt , so m eh ow co nfi de nt , a lit tle co nfi de nt an d th e re fe re nc e gr ou p is “n ot co nfi de nt ”) .e du ,c ap tu re s ed uc at io n le ve l( do ct or at e, m as te r, ba ch el or ,a ss oc ia te de gr ee ,s om e co lle ge ,h ig h sc ho ol ,a nd th e re fe re nc e gr ou p is “n on e” ). a tt b o rr o w m ea su re s re sp on de nt s’ at tit ud e to w ar ds bo rr ow in g. b or r h is t co nt ro ls fo r pa st be ha vi or in th e sa m e bo rr ow in g ca te go ry as th e de pe nd en tv ar ia bl e (e .g ., in th e au to m ob ile re gr es si on , b or r h is t in di ca te s th at th e re sp on de nt ha ve ha d a ca r lo an in th e pa st ). in th e cr ed it ca rd re gr es si on ,t he fir st b or r h is t in di ca te s th at th e re sp on de nt ha d ha d cr ed it ca rd (s ) in th e pa st bu tt he y pa y on tim e. t he se co nd b or r h is t in di ca te d ca rr yi ng m on th ly ba la nc e. in th e m or tg ag e re gr es si on ,t he fir st b or r h is t in di ca te s ha vi ng a m or tg ag e ac co un t, th e se co nd b or r h is t in di ca te s ha vi ng re fin an ce d a ho m e in th e pa st , an d th e th ird b or r h is t in di ca te s ha vi ng a ho m e eq ui ty lo an in th e pa st . a st er is ks de no te st at is tic al si gn ifi ca nc e at th e 1% (* ** ), 5% (* *) ,o r 10 % (* ) le ve l. fo r ea ch re gr es si on w e re po rt ra w an d ad ju st ed r 2 a ,d eg re es of fr ee do m an d fst at b . a o ur st ud y de si gn ne ce ss ita te s th e us e of nu m er ou s in te rco rr el at ed va ria bl es (n um er ic al re sp on se s to su rv ey qu es tio ns ) to ch ar ac te riz e an un de rly in g la te nt fa ct or (p er so na lit y tra it) .e xp lo ra to ry fa ct or a na ly si s( e fa )h as be en in ve nt ed sp ec ifi ca lly fo rt hi sk in d of co m pl ic at ed st ud ie s. e fa re du ce st he di m en si on al ity of th e or ig in al sp ac e an d ac co un ts fo rp ot en tia lc ova ria nc e in th e ob se rv ed va ria bl es (s te ve ns ,1 99 2; w oo ld rid ge ,2 00 6] .w e ra n an e xp lo ra to ry fa ct or a na ly si s (e fa )w ith pr in ci pa lc om po ne nt ex tra ct io n an d v ar im ax ro ta tio n an d a co rr el at io n an al ys is .t he e fa an al ys is pr ov id ed fu rth er su pp or tf or un id im en si on al ity of pe rs on al ity di m en si on s as al lm ea su re m en t ite m s sh ow ed hi gh lo ad in gs un de rt he ir co rr es po nd in g ac to r. t he co rr el at io n an al ys is re ve al s th at th e pa irw is e co rr el at io n be tw ee n pe rs on al ity tra its ra ng es be tw ee n � 0. 25 0 (b et w ee n co ns ci en tio us ne ss an d ne ur ot ic is m ) an d 0. 25 8 (b et w ee n op en ne ss an d ag re ea bl en es s) .w e co nc lu de th at co rr el at io n am on g in de pe nd en tv ar ia bl es is no tt oo hi gh . b a lo w ad ju st ed r 2 is ve ry co m m on in m ul tiv ar ia te an al ys is an d a se em in gl y lo w r 2 do es no tn ec es sa ri ly m ea n th at an o l s re gr es si on eq ua tio n is us el es s. w he th er or no t a re gr es si on m od el is a go od es tim at e of th e ce te ri s pa ri bu s re la tio ns hi p be tw ee n th e de pe nd en t th e in de pe nd en t va ri ab le s do es no t de pe nd di re ct ly on th e si ze of r 2 (w oo ld ri dg e, 20 06 , pp . 43 – 44 ). 67a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 differences between the full models and the concise models and, consequently, we only present the results with the concise models. the results for the full models are available upon request from the authors. the findings in table 2 generally support our first set of hypotheses. those individuals with higher levels of need for material resources had stronger intentions to borrow in all categories except for personal loans and payday loans. this is consistent with hypothesis 1a. similarly, higher levels of need for arousal was associated with a stronger intention to borrow in all categories that lends support to hypothesis 1b. as expected, the coefficients on the attitude towards borrowing (attborrow) were positive and significant across all borrowing types (except for student loans). this indicates that a positive attitude towards borrowing is positively associated with intentions to borrow. however, when the coefficients of attitudes and intentions across personalities are considered, the results provided evidence supporting hypothesis 2 (i.e., intention to borrow and the attitude towards borrowing may not always accord). specifically, the results indicate that despite high levels of intention to use certain types of borrowing (student loans and personal loans), neurotic individuals have a neutral attitude toward borrowing (i.e., their attitudes towards borrowing is not favorable in general). moreover, individuals who score high on introversion exhibit stronger intentions to open new credit card accounts although despite having neutral attitudes towards borrowing. also, those with higher levels of agreeableness and conscientiousness show negative attitudes towards borrowing, but this neither strengthens nor weakens their intentions to borrow in almost all categories (except for mortgage and payday loans). the coefficients of borrhist are positive, significant, and statistically distinctive from zero within each borrowing category (except mortgage). therefore, previous borrowing behavior is positively related to future borrowing behavior. this implies that borrowing could be a repetitive or an addictive behavior that makes future borrowing more probable. in other words, it could be argued that borrowers somehow create a level of comfort zone in borrowing, since those who have borrowed in a certain manner in the past have relatively stronger intentions to borrow in a similar manner in the future. the finding that attitude towards borrowing is relatively more favorable for those who have applied for credit in the past 5 years (see borrstatapld )) also supports this view. the intention to borrow is relatively weaker for those who pay their bills on time (borrstatontime ). this could be related to the fact that for responsible borrowers who make timely payments, more loans would result in inability to make timely payments in the future. in addition, table 2 unveils several attributes of the relationship between financial behavior and personalities. more specifically, the results in table 2 do not provide support for the findings of davey and george (2011) who provided evidence indicating that openness to experience is associated with higher intentions to borrow. our results do not support the findings of harley and wilhelm (1992) reporting that highly conscientious individuals are less likely in need of borrowing. our findings indicate that openness to experience and conscientiousness are personality traits that do not weaken, nor do they intensify borrowing intentions. moreover, our results indicate that neuroticism is positively associated with certain types of borrowings such as student loans and personal loans. this is consistent with the view that 68 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 neuroticism is associated with overspending (mansfield, pinto, & parente, 2003), impulsive spending (baumeister and exline 2000; strayhorn, 2002; youn & faber, 2000), compulsive buying (mowen, 2000), and poor financial planning (brandstatter, 1996; davey and george, 2011). it is clear that these types of this financial behavior would eventually lead to immediate need for money that is often satisfied with borrowing in the form of personal loans. poor financial planning (another behavior related to this personality trait) could result in higher needs for student loans as well (please note that we have controlled for the effect of income). our results also indicate that introversion is not associated with any change in the intention to borrow across seven different borrowing options except for credit cards. the positive association between this trait and intentions to credit card borrowing is indeed interesting and in line with an earlier finding by davey and george (2011), supplying additional support for our argument that borrowing options are not homogenous. however, our findings do not support those of sadi et al. (2011) who found that extroversion (i.e., the opposite of introversion) increases the tendency to borrow. finally, we found that agreeableness is positively related to the intention to use mortgages and negatively to the intention to use payday loans. nyhus and webley (2001) showed that highly agreeable people have higher needs to borrow more. our results indicates that this conclusions extends only to mortgages. contemplating the differences between our findings and the findings of earlier research, we argue that these differences are attributed to several factors. first, we take a more meticulous approach to borrowing by considering a wide range of borrowing options and adding more specificity to understanding borrowing decisions. second, we use a comprehensive personality profiling approach by examining the two understudied personality traits in addition to the traditionally studied big five traits while specifically focusing on effects of the need for material resources and the need for arousal traits. overall, our results support the main argument that borrowing options are not homogenous and they vary across personalities. 3.4. interaction plots interaction plots are widely used in the literature to display interaction effects. an interaction effect is witnessed when the joint effects of two variables is the focus of attention so that the outcome variable would show significant differences for different levels of the independent variable at different levels of the moderator variable (hair et al., 2006). in the context of this article, we are interested in finding how a certain personality trait affects the relationship between the attitude towards borrowing and the intention to borrow. while the connection between the attitude towards borrowing and the intention to borrow is intuitive, hypotheses hypothesis 3a and hypothesis 3b postulate that the strength of this connection is a function of the borrower’s personality. specifically, the two hypotheses suggest that the impact of attitude toward borrowing on intention to borrow is intensified for individuals with greater need for material resources (h3a) and/or greater need for arousal (h3b). to test these hypotheses, we use the mode probe analysis approach using process model 1 (hayes, 2012) where personality trait is the moderator, the attitude toward borrowing is the independent variable and the borrowing intention is the dependent variable. figure 1 69a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 figure 1 interaction plots–the need for material resources 70 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 is devoted for the need for material resources trait and figure 2 is devoted for the need for arousal trait. the top two charts in figure 1 indicate that the need for material resources trait significantly moderates the relationship between attitude towards borrowing and the intention on to borrow. in other words, the relationship between attitude towards borrowing and the intention to apply for mortgages and home improvement loans shows significantly different patterns at different levels of this personality trait. the results in panel a indicate that at lower levels of the need for material resources (i.e., for those individuals who do not score high on this trait as indicated by one standard deviation below the mean in our analyses), high or low levels of attitudes toward borrowing (i.e., more or less favorable attitudes toward borrowing) do not significantly impact intentions to apply for a mortgage. the average score on the intention to open a new mortgage account for those who have a more favorable attitude towards borrowing is 1.87. the corresponding average score for those who have a less favorable attitude towards borrowing is 1.80. the difference, 0.07, is trivial and is not figure 1 (continued) notes: the figure shows the results of running interaction plots to discern how the need for material resources trait moderates the relationship between the attitude towards borrowing and the intentions to borrow. to run the test, we classify participants into two classes using the mean of scores and the standard deviation of the scores on the need for material resources test. specifically, respondents with a need for material resources scores higher (lower) than one standard deviation above (below) the mean are classified as individuals with greater (lesser) need for material resources. analogically, respondents with an attitude towards borrowing scores higher (lower) than one standard deviation above (below) the mean are classified as individuals with more (less) favorable attitude towards borrowing. we run the model with all eight borrowing options but we report only the findings with significant interaction effects. under each chart, we report the averages scores on the corresponding intention to borrow. we also report the difference and statistical significance (p-value) for the difference in average scores between the more favorable attitude and the less favorable attitude subgroups. asterisks denote statistical significance at the 1% (***), 5% (**), or 10% (*) level. 71a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 figure 2 interaction plots–the need for arousal 72 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 statistically different from zero (p � 0.814). in contrast, when the need for material resources is higher (i.e., in case of individuals who score high on this trait as indicated by one standard deviation above the mean in our analyses), intention to apply for a mortgage is significantly different at different levels of attitude towards borrowing. more specifically, for those with greater need for material resources, when attitude toward borrowing is more favorable, intention to apply for mortgage loans in significantly higher compared to those with less favorable attitudes towards borrowing. the average score on the intention to open a new mortgage account for those who have a more favorable attitude towards borrowing is 2.83 that is 0.92 point higher than the corresponding score for those who have a less favorable attitude (1.91). it is noteworthy that the difference in intentions is almost one scale unit (on a scale of 1 to 5) indicating 18.4% increase in intentions to apply for mortgage that is statistically significant at the 5% level (p � 0.018). the same observations extend to panel b. when the need for material resources is low, the average score on the intention to obtain a home improvement loan for those who have a more favorable attitude towards borrowing is not significantly different from those who figure 2 (continued) notes: the figure shows the results of running interaction plots to discern how the need for arousal trait moderates the relationship between the attitude towards borrowing and the intentions to borrow. to run the test, we classify participants into two classes using the mean of scores and the standard deviation of the scores on the need for arousal test. specifically, respondents with a need for arousal scores higher (lower) than one standard deviation above (below) the mean are classified as individuals with greater (lesser) need for arousal. analogically, respondents with an attitude towards borrowing scores higher (lower) than one standard deviation above (below) the mean are classified as individuals with more (less) favorable attitude towards borrowing. we run the model with all eight borrowing options but we report only the findings with significant interaction effects. under each chart, we report the averages scores on the corresponding intention to borrow. we also report the difference and statistical significance (p-value) for the difference in average scores between the more favorable attitude and the less favorable attitude subgroups. asterisks denote statistical significance at the 1% (***), 5% (**), or 10% (*) level. 73a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 have a less favorable attitude (p � 0.281). in contrast, when the need for material resources is greater, a more favorable attitude towards borrowing is associated with a significantly higher intention to draw a home improvement loan (mean difference is 1.05 points and is statistically significant at the 1% level; p � 0.001). panel c reveals similar findings with the intention to draw a business loan. when the need for material resources is low, the average score for those who have a more favorable attitude towards borrowing is 1.56 while it is 1.31 for those with a less favorable attitude towards borrowing. the difference is 0.25 scale points which is statistically insignificant (p � 0.239). when the need for material resources is higher, the average score for those who have a more favorable attitude towards borrowing is 2.45 while the corresponding score for those who have a less favorable attitude towards borrowing is 1.30. the difference is an impressive 1.15 points and is statistically significant (p-value is less than 0.001). the findings also extend to panel d with one exception. when the need for material resources is low, the attitude towards borrowing there is no significant difference in individuals’ intention to take a payday loan regardless of their attitudes towards borrowing. however, for those who score higher on the need for material resources, the role of the attitude is much stronger in relation to intention to take payday loans. more specifically, when the need for material resources is low, the average score on the intention to take a payday loan for those who have a more favorable attitude towards borrowing is 1.41. the corresponding score for those who have a less favorable attitude towards borrowing is 1.08. the difference of 0.33 is only marginally significant at the 10% (p � 0.058). in contrast, when the need for material resources is greater, the average score on the intention to take a payday loan for those who have a more favorable attitude towards borrowing is 2.11 that is 1.21 point higher than the average score for those who have a less favorable attitude (0.90). the difference is 1.05 points and is statistically significant at the 1% level (p � 0.001). finally, the plots in panel e show that the need for material resources affects how the attitude towards borrowing translate into intentions to obtain a personal loan. the plots show that a more favorable attitude towards borrowing leads to a stronger intention for personal loans, but the relationship is much stronger when the need for material resources is greater. overall, the findings in figure 1 support hypothesis 3a. specifically, they present evidence consistent with the supposition that while the connection between the attitude towards borrowing and the intention to borrow is intuitive, it is meaningfully impacted by one’s need for material resources. in panels a, b, and c, when the need for material resources is low, the attitude towards borrowing seems to have insignificant role in determining one’s intention to borrow. in panels d and e, when the need for material resources is low the attitude towards borrowing has a weaker impact on the intention to borrow. in contrast, when the need for material resources is higher, a more favorable attitude towards borrowing significantly intensifies one’s intention to borrow. the results in figure 2 provide support of the hypothesis regarding the moderating role of the need for arousal in the relationship between attitude towards borrowing and specific borrowing intentions. the plots in panel a indicate that at lower levels of the need for arousal (i.e., those individuals who score low on this trait), the intention to obtain a home improve74 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 ment loan does not significantly change at different levels of attitudes towards borrowing. more specifically, when need for arousal is low, home improvement loan intention for those who have a more favorable attitude towards borrowing is 1.62, and the corresponding score for those who have a less favorable attitude towards borrowing is 1.36. the difference, 0.26, is small and insignificant (p � 0.220). in contrast, when the need for arousal is higher, the average score on the intention for home improvement loan for those who have a more favorable attitude towards borrowing is 2.49 which is 1.21 points higher than the average score for those who have a less favorable attitude (m � 1.28). the difference is statistically significant at the 1% level (p � 0.001). we conclude that when the need for arousal is high, the attitude towards borrowing plays a significant role in enticing one to apply for home improvement loan. when the need for arousal is low, the attitude towards borrowing does not impact the intention to take a home improvement loan. the same observations extend to panels b and c; that is, when the need for arousal is low, the average score on the intention to draw a business loan (panel b) or to take a payday loan (panel c) does not improve significantly when the attitude towards borrowing is more favorable. in contrast, when the need for arousal is greater, the average score on the intention to draw a business loan (panel b) or to take a payday loan (panel c) is systematically higher for those who demonstrate a more favorable attitude towards borrowing. the observations also extend to panel d but with a remarkably lower significance. we still observe a non-significant impact of attitude on the intention to take a student loan for those who have a lesser need for arousal (the difference is a small �0.18 points and is statistically insignificant form zero). when the need for arousal is greater, there is a more positive impact of attitude on the intention to take a student loan. the difference is a reasonable 0.54 point and is statistically significant at 10% only (p-value is 0.066). finally, the plots in panel e show that the need for arousal impacts how the attitude towards borrowing translate into intentions to obtain a personal loan. the plots show that a more favorable attitude towards borrowing leads to a stronger intention for personal loans but the relationship is much stronger when the need for arousal is greater. overall, the findings in figure 2 support hypothesis 3b. they indicate that the relationship between attitude towards borrowing and the intention to borrow extends only to individuals with greater need for arousal. in all five panels, except panel e, when the need for arousal is low, the attitude towards borrowing exhibits an insignificant role in one’s intention to borrow. in panel e, when the need for arousal is low, the attitude towards borrowing exhibits a much weaker impact on one’s intention to borrow. in contrast, when the need for arousal is greater, the attitude towards borrowing significantly strengthens one’s intention to borrow. 3.5. regressions with interaction terms in this section, we focus on the interaction effects between the attitude towards borrowing, the intention to borrow, and personality traits. hypotheses 3a and 3b postulate that the attitude towards borrowing might have a different effect on the intention to borrow for individuals with different personalities. we test these hypotheses by estimating the incre75a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 mental effect of personality on the relationship between the attitude towards borrowing and the intention to borrow. empirically, we estimate the following regression models borrintn � c � �1matr � �2attborrow � �3�matr � attborrow� (1) borrintn � c � �1arous � �2attborrow � �3� arous � attborrow� (2) the first model includes the need for material resources score (matr), the attitude towards borrowing (attborrow), and an interaction between the two (matr � attborrow). the dependent variable is the intention to borrow (borrintn). the null hypothesis, therefore, is that the change in the intention to borrow that is induced by the attitude towards borrowing is the same across individuals with different need for material resources. if interpreted carefully, the coefficients on matr, attborrow, and matr � attborrow are sufficient to reveal the intention to borrow differentials across all possible combinations of need for material resources and the attitude towards borrowing relative to hypothetical base group defined by construction as [matr � 0, attborrow � 0]. see legend of table 3 for more explanation. the second model with the need for arousal trait is a replica of the first model and, thus, is interpreted in an analogous manner. the parameters of second model are reported in table 4. the results in table 3 indicate that the interaction between the attitude towards borrowing and the need for material resources is not consistent across all borrowing options; which supports our core argument that borrowing options are not homogenous and are motivated differently. the coefficients on the interaction term, �2, are significantly positive in regressions 2, 3, 4, 7, and 8 that correspond to the intentions to apply for a mortgage, a home improvement loan, a business loan, a personal loan, and a payday loan, respectively. this indicates that for individuals with greater need for material resources, the attitude towards borrowing shows a stronger association on the intention to use these borrowing options. this is consistent with hypothesis 3a. an interesting find was that in regressions 2 through 8, the coefficients on the attitude towards borrowing, �2, and the need for material resources, �1, are statistically indistinguishable from zero. this implies that neither the attitude towards borrowing nor the need for material resources can, alone, influence one’s intentions to borrow. regression 1 corresponds to the intention to open a credit card account and seems to be structurally different from other regressions. in particular, we observe that both �2 and �1 are positive and statistically significant. this indicates that credit card borrowing is motivated differently from other borrowing options. a more favorable attitude towards borrowing as well as higher values of need for material resources induce stronger intentions for opening a new credit card account. at the same time, we cannot reject the null that the interaction between the two is insignificant. as in table 3, the results in table 4 indicate that the interaction between the attitude towards borrowing and the need for arousal is not consistent across all borrowing options; which supports the argument that borrowing options are not homogenous and are, therefore, motivated differently. 76 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 t ab le 3 r eg re ss io n an al ys es w ith in te ra ct io n te rm s: t he ne ed fo r m at er ia l re so ur ce s (1 ) c re di t ca rd (2 ) m or tg ag e (3 ) h om e im pr ov em en t lo an (4 ) b us in es s lo an (5 ) a ut om ob ile (6 ) st ud en t lo an (7 ) pe rs on al lo an (8 ) pa yd ay lo an c on st an t [c ] 0. 16 78 1. 90 42 ** * 1. 48 30 ** * 1. 46 60 ** * 2. 04 58 ** * 1. 24 01 ** * 1. 03 23 ** * 1. 31 47 ** * a tt b o rr o w [� 2 ] 0. 03 50 ** * � .0 09 8 � 0. 00 76 � 0. 00 77 � 0. 00 04 � 0. 00 21 0. 00 59 � 0. 00 59 m at er ia li sm [� 1 ] 0. 28 48 ** � .0 76 5 � 0. 13 13 � 0. 14 48 � 0. 03 64 0. 19 61 � 0. 09 41 � 0. 02 05 in te ra ct io n [ � 3 ] � 0. 00 24 .0 06 7* * 0. 00 68 ** * 0. 00 71 ** * 0. 00 44 0. 00 15 0. 00 46 ** 0. 00 66 ** * r 2 0. 12 96 0. 04 15 0. 68 1 0. 07 05 0. 03 24 0. 03 93 0. 09 00 0. 07 60 df . r eg re ss io n 84 5 84 5 84 5 84 5 84 5 84 5 84 5 84 5 fst at e 41 .9 6 12 .1 9 20 .6 0 21 .3 5 9. 43 6 11 .5 36 27 .8 54 23 .1 7 si g. 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 n ot es : t he ta bl e sh ow s th e re su lts of ei gh ti nt er ac tio n re gr es si on s th at co rr es po nd to th e ei gh td if fe re nt bo rr ow in g in te nt io ns ex pl or ed in th is st ud y. in ea ch m od el , w e in te ra ct th e at tit ud e to w ar ds bo rr ow in g sc or e (a tt b o rr o w ) w ith th e ne ed fo r m at er ia l re so ur ce s (m at r sc or e. sp ec ifi ca lly , w e ru n th e fo llo w in g m od el s: b or ri nt n � c � � 1 m at r � � 2 a tt b o rr o w � � 3 �m at r � a tt b o rr o w � b or ri nt n is th e bo rr ow in g in te nt io n va ri ab le th at ca pt ur es th e re sp on de nt s’ in te nt io n to us e a ce rt ai n ty pe of bo rr ow in g. t he in tr od uc tio n of an in te ra ct io n te rm in a re gr es si on ca lls fo r a ca re fu li nt er pr et at io n of th e co ef fic ie nt s on th e ot he r va ri ab le s. t he in te ra ct io n te rm m at r � a tt b o rr o w ca pt ur es th e in cr em en ta l ef fe ct of th e at tit ud e to w ar ds bo rr ow in g on th e in te nt io n to bo rr ow fo r in di vi du al s w ith gr ea te r ne ed fo r m at er ia lr es ou rc es .a si gn ifi ca nt ly po si tiv e co ef fic ie nt � 3 im pl ie s th at th e ne ed fo r m at er ia l re so ur ce s af fe ct ho w th e at tit ud e to w ar ds bo rr ow in g ch an ge s on e’ s in te nt io n to bo rr ow ,t ha t is ,a si gn ifi ca nt ly po si tiv e � 3 le ad s to th e co nc lu si on th at th e at tit ud e to w ar ds bo rr ow in g yi el ds a st ro ng er in te nt io n to bo rr ow fo r in di vi du al s w ith gr ea te r ne ed fo r m at er ia l re so ur ce s. � 2 m ea su re s th e im pa ct of th e at tit ud e to w ar ds bo rr ow in g on th e in te nt io n to bo rr ow in th e ab se nc e of a ne ed fo r m at er ia lr es ou rc es .� 1 m ea su re s th e im pa ct of th e ne ed fo r m at er ia lr es ou rc es on th e in te nt io n to bo rr ow w he n th e at tit ud e to w ar ds bo rr ow in g is ze ro .t he re fo re ,t he � 3 co ef fic ie nt su pp lie s a di re ct te st fo r h yp ot he si s 3a an d th e � 1 co ef fic ie nt s su pp ly ad di tio na l te st in g of h yp ot he si s 1a . a st er is ks de no te st at is tic al si gn ifi ca nc e at th e 1% (* ** ), 5% (* *) , or 10 % (* ) le ve l. 77a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 t ab le 4 r eg re ss io n an al ys es w ith in te ra ct io n te rm s– t he ne ed fo r ar ou sa l (1 ) c re di t ca rd (2 ) m or tg ag e (3 ) h om e im pr ov em en t lo an (4 ) b us in es s lo an (5 ) a ut om ob ile (6 ) st ud en t lo an (7 ) pe rs on al lo an (8 ) pa yd ay lo an c on st an t [c ] 0. 68 60 * 1. 66 16 ** * 1. 63 23 ** * 1. 25 23 ** * 1. 52 19 ** * 1. 67 50 ** * 1. 34 18 ** * 1. 41 70 ** * a tt b o rr o w [� 2 ] 0. 02 22 ** � 0. 00 38 � 0. 00 95 � 0. 00 48 0. 01 23 � 0. 01 25 � 0. 00 29 � 0. 00 86 n ee d fo r ar ou sa l [� 1 ] 0. 06 61 0. 00 38 � 0. 20 92 * � 0. 06 69 0. 16 28 0. 00 30 � 0. 22 45 * � 0. 25 36 ** * in te ra ct io n [� 3 ] 0. 00 29 0. 00 48 0. 00 81 ** * 0. 00 62 ** � 0. 00 03 0. 00 60 * 0. 00 82 ** * 0. 00 79 ** * r 2 0. 13 05 0. 03 49 0. 06 08 0. 07 55 0. 02 79 0. 04 03 0. 10 19 0. 07 96 df . r eg re ss io n 84 5 84 5 84 5 84 5 84 5 84 5 84 5 84 5 fst at e 42 .2 6 10 .1 8 18 .2 3 23 .0 0 8. 10 11 .8 4 31 .9 5 24 .3 5 si g. 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 n ot es : t he ta bl e sh ow s th e re su lts of ei gh ti nt er ac tio n re gr es si on s th at co rr es po nd to th e ei gh td if fe re nt bo rr ow in g op tio ns co ns id er ed in th is st ud y. in ea ch m od el , w e in te ra ct th e at tit ud e to w ar ds bo rr ow in g sc or e (a tt b o rr o w ) w ith th e ne ed fo r ar ou sa l sc or e (a ro us ). sp ec ifi ca lly , w e ru n th e fo llo w in g m od el s: b or ri nt n � c � � 1 a ro us � � 2 a tt b o rr o w � � 3 �a ro us � a tt b o rr o w � b or ri nt n is th e bo rr ow in g in te nt io n va ri ab le th at ca pt ur es th e re sp on de nt s’ in te nt io n to us e a ce rt ai n ty pe of bo rr ow in g. a s ex pl ai ne d in t ab le 3 ab ov e, a si gn ifi ca nt ly po si tiv e � 3 im pl ie s th at th e at tit ud e to w ar ds bo rr ow in g yi el ds a st ro ng er in te nt io n to bo rr ow fo r in di vi du al s w ith gr ea te r ne ed fo r ar ou sa l. � 2 m ea su re s th e im pa ct of th e at tit ud e to w ar ds bo rr ow in g on th e in te nt io n to bo rr ow in th e ab se nc e of a ne ed fo r ar ou sa l. � 1 m ea su re s th e im pa ct of th e ne ed fo r ar ou sa l on th e in te nt io n to bo rr ow w he n th e at tit ud e to w ar ds bo rr ow in g is ze ro .t he re fo re ,t he � 3 co ef fic ie nt s su pp ly a di re ct te st fo r h yp ot he si s 3b an d th e � 1 co ef fic ie nt s su pp ly ad di tio na l te st in g of h yp ot he si s 1b . a st er is ks de no te st at is tic al si gn ifi ca nc e at th e 1% (* ** ), 5% (* *) , or 10 % (* ) le ve l. 78 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 the coefficients on the interaction term, �3, are significant and positive in regressions 3, 4, 6, 7, and 8 that correspond to the intentions to apply for a home improvement loan, a business loan, a student loan, a personal loan, and a payday loan, respectively. this indicates that the attitude towards borrowing has a stronger relationship with the intention to borrow for individuals with higher levels of need for arousal. this is consistent with hypothesis 3b. in regressions 2 through 8, the coefficients on the attitude towards borrowing score, �2, and are statistically indistinguishable from zero. this implies that the attitude towards borrowing cannot, alone, influence one’s intentions to borrow. in regression 1, only �2 is positive and statistically significant. this indicates that a more favorable attitude towards borrowing induces a stronger intention for opening a new credit card account. it also indicates that we cannot reject the null that the interaction between the attitude towards borrowing and the need for arousal is insignificant. 3.6. discussion of findings consistent with previous research and in line with behavioral finance theories, the results suggest that psychological characteristics impact individuals’ financial choices. in addition, our findings are generally in line with the behavioral influence perspective of consumer financial decision making. nevertheless, this research adds several new perspectives to the literature on financial behavior and personal characteristics. we document evidence that greater values of need for material resources and need for arousal are associated with stronger intentions to borrow. this lends support to hypotheses 1a and h1b. we also show that for individuals who score higher on need for material resources, the positive association between attitude towards borrowing and the intention to borrow is significantly stronger relative to individuals with a lesser need for materials resources. this supports hypothesis 3a. the same finding extends to the need for arousal trait that is consistent with hypothesis 3b. furthermore, the overall results support a few overarching suppositions tested in this article. first, the attitude towards borrowing and the intention to borrow are not interchangeable. in fact, the intention to borrow is sometimes inconsistent with the borrower’s attitude towards borrowing. this is consistent with hypothesis 2. second, the fact that we were able to document an interaction effect in only a subset of borrowing options indicates that borrowing options are not homogenous and are motivated differently. a further discussion of this issue is supplied in the conclusion section. finally, our analyses take into account conventional controls used in extant literature. we did not find fundamental deviations from previous findings (i.e., we found that attitudinal biases towards money, wealth, indebtedness, planning as well as financial knowledge and other demographics impact borrowing behavior in a predictable manner). 4. conclusions and implications the present research is an interdisciplinary approach to investigate household’s financial borrowing decisions in the light of psychological characteristics of individuals. the review 79a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 of extant literature reveals a three-fold gap in household borrowing research. first, the role of personality traits in household finance remains an area that should be investigated. the extant research has mainly focused on the big five personality traits and has not examined the extensions by the more comprehensive 3m model of motivation and personality. specifically, the need for material resources and the need for arousal traits have not been fully explored. second, current literature has often applied attitude toward a behavior and intention for the behavior interchangeably. third, borrowing behavior has been defined too broadly with minimal attention to the fact that the ultimate purpose of the borrowing might influence the decision to borrow. the present research is an effort to address the aforementioned gaps. first, we use an extended version of the big five model of personality that includes, in addition to the conventional five traits, the two additional personality traits mentioned above. second, we make a clear distinction between the attitude towards borrowing and the intention to borrow. finally, we consider a wide spectrum of borrowing behaviors defined by the scf administered by the federal reserve. the findings are insightful and present several practical and academic implications. first, our analyses indicate that the two additional personality traits, the need for material resources and the need for arousal, play a significant role in shaping borrowing behavior. second, our findings indicate the importance of making a clear distinction between the attitude towards borrowing and the intentions to borrow. the attitude towards borrowing and the intention to borrow do not always accord, and the inconsistency between the two varies across personalities. in other words, our results provide evidence for the interaction between personality and attitudes towards borrowing on intentions for specific borrowing options. third, we demonstrate that different borrowing options are not homogenous and more importantly, are motivated differently across personalities. intuitively, the attitude towards borrowing influences the intention to borrow. our finding, however, shows that this intuitive relationship is not consistent across different borrowing options and is greatly influenced by the borrower’s personality characteristics. overall, individuals with greater need for material resources find happiness in owning material wealth. they view possessions as means of achieving utility and social status as opposed to comfort and pleasure (richins & dawson, 1992). therefore, they are more willing to carry the burden of a loan in exchange for psychological satisfactions. their borrowing behavior, therefore, is marked by stronger urge to possess material wealth. we find that for individuals with greater need for material resources, the impact of the attitude towards borrowing on the intention to borrow is significantly intensified in mortgages, home improvement loans, business loans, personal loans, and payday loans. the same is not true for automobile loans, credit cards, and student loans. we find this divergence wellexplained by the psychological characteristics of individuals with greater need for material resources. they are more concerned about their social image compared to those who score lower on this trait (christopher, marek, & carroll, 2004) and, therefore, they have a stronger urge to own a home, improve their home, and own a business. these individuals are more likely to spend money on themselves than on others and they contribute less money to charities (richins & dawson, 1992) that explains why they have stronger intention to obtain personal loans. finally, they are more likely to get engaged in compulsive buying (mowen, 80 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 2000; mowen & spears, 1999) that explains the stronger intention for payday loans. arguably, they are willing to carry the burden of the debt to satisfy their psychological need to spend and own personal items immediately (as opposed to wait and save up for them). in contrast, individuals who have greater need for arousal derive happiness from the purchasing experience itself and, thus, they enjoy the frequency of purchasing (licata, mowen, harris, & brown, 2003; mehrabian & russell, 1974). therefore, they spend more on frequent purchases including impulse purchases and less on big ticket items. we find that their borrowing pattern is consistent with their psychological characteristics. for individuals with greater need for arousal, the association between attitude towards borrowing and the intention to borrow is significantly stronger in the case of home improvement loans, business loans, student loans, personal loans, and payday loans. the same is not true for mortgages, automobile loans, and credit cards. in fact, the intention for credit cards and automotive loans seems to be independent of the borrower’s personality. this finding is consistent with the anecdotal evidence of the wide prevalence of these types of borrowing options in the society. despite the aforementioned contributions, the findings of this research are limited in terms of generalizability because of our sample size. our findings, however, pave the road for future research in this area that could span beyond academia. for marketers of financial services, our findings suggest that personality traits of individual and their associated inherent tendencies should be considered when communicating with borrowers. more specifically, since personality traits are associated with different borrowing decisions, it is expected that different groups of individuals with varying personality traits should react differently to marketing efforts. as such, borrowers can be categorized based on their personality traits and attitudes towards certain borrowing options and approached differently. we encourage future research to follow experimental designs and examine such differences. moreover, in consumer profiling and target market selections fields, more research may explore the possibility that the attitude towards borrowing and the intention to borrow are not always consistent. our results identify the moderating role of personality traits as a boundary condition that provides insights on these inconsistencies. future research is encouraged to identify additional boundary conditions. our results also highlight the importance of psychological characteristics in financial decisions. according, this research calls for more research by social planners, legislators, personal financial planners, and educators. we specifically recommend more research on how personal differences shape household’s financial behaviors. notes 1 our analysis also incorporates individuals’ attitude towards financial planning and money, borrowing history and self-evaluation of one’s finances and financial decisionmaking abilities. these factors are largely ignored by previous studies. 2 standard controls for demographics, knowledge, and financial endowment are also considered. see for example ryack and sheikh (2016), balasubramnian and brisker (2016), and docking, fortin, and michelson (2013). 3 there are two distinct approaches for studying personality, namely the psychoanalytic theory and the trait theory. the trait theory focuses on various personality traits and 81a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 their impact on consumer decisions. this theory has been used in two approaches. one approach argues that people have the same set of traits, but these traits are manifested dissimilarly. the second approach argues that individual variances are because of different combinations of each trait varying from one individual to another (lin, 2010). in fact, the theoretical mechanism through which individual differences influence decisions is rooted in the social cognition research perspective (witte and morrison, 2000). according to this perspective, individual differences influence individuals’ perceptions of the world and environmental stimuli (fiske & taylor, 1991) and, thus, can influence their behavior. 4 see davey and george (2011) for a detailed discussion of these traits. openness to experience is the general interest in finding novel solutions and original ideas, and the tendency to use the imagination in performing activities. this trait is associated with being open to ideas and values and trying new things that could potentially be costly. conscientiousness is the need to be organized and efficient in tasks and is defined as the extent to which individuals are careful and self-disciplined. conscientious individuals are organized, hard-working, and reliable as opposed to carefree individuals who score lower on this trait. extroversion is the tendency to be warm blooded, friendly, and sociable. agreeableness is the need to express kindness and sympathy to others. scoring low on this trait shows higher levels of self-interest. as a result, low-agreeable individuals have higher tendencies to hoard their money and engage in frugal practices rather than extravagance. neuroticism is the tendency to be emotional, moody and temperamental. neurotic individuals are more likely to experience negative emotions and are more susceptible to stress. scoring high on neuroticism indicates lack of emotional stability and self-control, which results in taking more emotional/ impulsive decisions and less planned/controlled ones. 5 physical/body needs are related to devoting more time to improve one’s body on a daily basis. those who score higher on this trait work hard to keep their bodies healthy and in good shape (mowen, 2000) and have active lifestyles. this trait is, therefore, not related to borrowing behavior or attitude toward borrowing (pirog & roberst, 2007) and is not included in our analyses. 6 this is a simple yes/no question with a text box that allows respondents, if they wish, to specify the nature of the event. about a third of respondents cited events related to work (e. g. promotion) or personal life (e. g. marriage, divorce). the corresponding binary variable, however, was found to have insignificant impact on borrowing behavior. 7 however, it is higher than non-u. s.-based studies. for instance, the averages reported by larkin, lucey, and mulholland, 2013 who used the same risk tolerance quiz on an irish sample fall in the 20 to 30% range. references ajzen, i., & fishbein, m. (1977). the influence of attitudes on behavior. in d. albarracin, b. t. johnson, & m. p. zana (eds.), the handbook of attitudes (pp. 173–221). mahwah, nj: lawrence erlbaum. 82 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 ajzen, i., & fishbein, m. (1980). understanding attitudes and predicting social behavior. englewood cliffs, nj: prentice-hall. alhenawi, y., & elkhal, k. (2013). financial literacy of u.s. households: knowledge vs. long-term financial planning. financial services review, 22, 211–244. bagozzi, r. p., baumgartner, j., & yi, y. (1989). an investigation into the role of intentions as mediators of the attitude-behavior relationship. journal of economic psychology, 10, 35–62. balasubramnian, b., & brisker, e. r. (2016). financial adviser users and financial literacy. financial services review, 25, 127–155. baumeister, r. f., & exline, j. j. (2000). self-control, morality, and human strength. journal of social and clinical psychology, 19, 29–42. belk, r. w. (1984). three scales to measure constructs related to materialism: reliability, validity, and relationships to other measures of happiness. in t. kinnear (ed.), advances in consumer research (vol. 11, pp. 291–297). provo, ut: association for consumer research. bentler, p. m., & speckart, g. (1981). attitudes ‘cause’ behaviors: a structural equation analysis. journal of personality and social psychology, 40, 226–238. besharat, a., carrillat, f. a., & ladik, d. m. (2014). when motivation is against debtors’ best interest: the illusion of goal progress in credit card debt repayment. journal of public policy & marketing, 33, 143–158. besharat, a., varki, s., & craig, a. w. (2015). keeping consumers in the red: hedonic debt prioritization within multiple debt accounts. journal of consumer psychology, 25, 311–316. bloch, p. h., sherrell, d. l., & ridgway, n. m. (1986). consumer search: an extended framework. journal of consumer research, 13, 119–126. brandstatter, h. (1996). saving, income, and emotional climate of households related to personality structure. progress report 38. vsb-center savings project, tilburg university, the netherlands. brown, s., & taylor, k. (2011). household finances and the ‘big five’ personality traits. journal of economic psychology, 45, 197–212. buhrmester, m., kwang, t., & gosling, s. d. (2011). amazon’s mechanical turk a new source of inexpensive, yet high-quality, data? perspectives on psychological science, 6, 3–5. campbell, j. y. (2006). household finance. journal of finance, lxi, 1553–1604. christopher, a. n., marek, p., & carroll, s. m. (2004). materialism and attitude toward money: an exploratory investigation. individual differences research, 2, 109–117. consumer federation of america. (2013). financial planning profiles of american households: the 2013 household financial planning survey and index. certified financial planner board of standards, september 18, 2013. d’astous, a., maltais j., & roberge, c. (1990). compulsive buying tendencies of adolescent consumers. advances in consumer research, 17, 306–312. davey, j., & george, c. (2011). personality and finance: the effects of personality on financial attitudes and behavior. the international journal of interdisciplinary social sciences, 5, 275–294. docking, d., fortin, r., & michelson, s. (2013). the influence of race on the social security early retirement decision for married couples. financial services review, 22, 133–150. duclos, r. (2015). the psychology of investment behavior: (de)biasing financial decision-making one paragraph at a time. journal of consumer psychology, 25, 317–325. financial consumer agency of canada. your attitude toward credit and debt. https://www.canada.ca/en/financial-consumer-agency/services/financial-toolkit/credit/credit-1/2.html accessed on 4-16-2014. fishbein, m., & ajzen, i. (1975). belief, attitude, intention and behavior: an introduction to theory and research. reading, ma: addison-wesley. fishbein, m., & ajzen, i. (1980). predicting and understanding consumer behavior: attitude-behavior correspondent. in i. ajzen & m. fishbein (eds.), understanding attitude and predicting social behavior (pp. 148–172). englewood cliffs, nj: prentice-hall. fiske, s., & taylor, s. (1991). social cognition (2nd ed.). new york, ny: mcgraw-hill. gilliam, j. e., chatterjee, s., & grable, j. (2010). measuring the perception of financial risk tolerance: a tale of two measures. journal of financial counseling and planning, 21, 30–43. 83a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 goodman, j. k., cryder, c. e., & cheema a. (2013). data collection in a flat world: the strengths and weaknesses of mechanical turk samples. journal of behavioral decision making, 26, 213–224. grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625–630. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. guiso, l., haliassos, m., & jappelli, t. (2002). household portfolios. cambridge, ma: mit press. hair, j. f., black, w. c., babin, b. j., anderson, r. e., & tatham, r. l. (2006). multivariate data analysis (6 ed.). upper saddle river, nj: pearson prentice hall. harley, a., & wilhelm, m. (1992). compulsive buying and exploration into self-esteem and money attitudes. journal of economics psychology, 13, 5–17. hayes, a. f. (2012). process: a versatile computational tool for observed variable mediation, moderation, and conditional process modeling [white paper] (available at http://www.afhayes.com/public/process 2012.pdf). hilgert, m., hogarth, j., & beverly, s. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309–332. hurst, m., & mendoza, m. (2016). the behavior heuristics responsible for formation and liquidation of tax holding accounts. financial services review, 25, 279–301. johnson, k., & li, g. (2010). the debt-payment-to-income ratio as an indicator of borrowing constraints: evidence from two household surveys. journal of money, credit and banking, 42, 1373–1390. jonas, k., diehl, m., & bromer, p. (1997). effects of attitudinal ambivalence on information processing and attitude-intention consistency. journal of experimental social psychology, 33, 190–210. knoll, m. a., & houts, c. r. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. the journal of consumer affairs, 46, 381–410. larkin, c., lucey, b., & mulholland, m. (2013). risk tolerance and demographic characteristics: preliminary irish evidence. financial services review, 22, 77. lee, j., & kwon, k.-n. (2002). consumers’ use of credit cards: store credit card usage as an alternative payment and financing medium. the journal of consumer affairs, 36, 239–262. licata, j. w., mowen, j. c., harris, e. g., & brown, t. (2003). on the trait antecedents and outcomes of service worker job resourcefulness: a hierarchical model approach. journal of the academy of marketing science, 31, 256–271. lin, l. (2010). the relationship of the consumer personality trait, brand personality and brand loyalty: an empirical study of toys and video games buyers. journal of product & brand management, 19, 4–17. lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review, 98, 413–417. lynch j. r. (2011). introduction to the journal of marketing research special interdisciplinary issue on consumer financial decision making. journal of marketing research,48, siv–sviii. mansfield, p. m., pinto, m. b., & parente, d. h. (2003). self-control and credit-card use among college students. psychological reports, 92, 1067–1078. mason, w., & suri. s. (2012). conducting behavioral research on amazon’s mechanical turk. behavior research methods, 44, 1–23. mccrae, r. r., & cost, p. t. jr. (1987). validation of the five-factor model of personality across instruments and observers. journal of personality and social psychology, 52, 81–90. mehrabian, a., & russell, j. a. (1974). an approach to environmental psychology. cambridge, ma. mowen, j. c., & carlson, b. (2003). exploring the antecedents and consumer behavior consequences of the trait of superstition. psychology & marketing, 20, 1045–1065. mowen, j. c., & spears, n. (1999). a hierarchical model approach to understanding compulsive buying among college students. journal of consumer psychology, 8, 407–430. mowen, j. c. (2000). the 3-m model of motivation and personality: theory and empirical applications to consumer behavior. boston, ma: kluwer academic press. 84 a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 mowen, j. c., park, s., & zablah, a. (2007). toward a theory of motivation and personality with application to word-of-mouth communications. journal of business research, 60, 590–596. norvilitis, j. m., merwin, m. m., osberg, t. m., roehling, p. v., young, p., & kamas, m. (2006). personality factors, money attitudes, financial knowledge, and credit card debt in college students. journal of applied social psychology, 36, 1395–1413. nyhus, e. k., & webley, p. (2001). the role of personality in household savings and borrowing behavior. european journal of personality, 15 (special issue: personality and economic behavior), 85–103. pinto, m. b., parente, d. h., & palmer, t. s. (2000). materialism and credit card use by college students. psychological reports, 86, 643–652. pirog, s. f. iii, & roberts, j. a. (2007). personality and credit card misuse among college students: the mediating role of impulsiveness. journal of marketing theory and practice, 15, 65–77. ponchio, m. c., & aranha, f. (2008). materialism as a predictor variable of low income consumer behavior when entering into installment plan agreements. journal of consumer behaviour, 7, 21–34. rand, d. g. (2012). the promise of mechanical turk: how online labor markets can help theorists run behavioral experiments. journal of theoretical biology, 299, 172–179. richins, m. l., & dawson, s. (1992). a consumer values orientation for materialism and its measurement: scale development and validation. journal of consumer research, 19, 303–316. rook, d. w. (1987). the buying impulse. journal of consumer research, 14, 189–199. ryack, k. n., & sheikh, a. (2016). the relationship between time perspective and financial risk tolerance in young adults. financial services review, 25, 157–180. ryack, k. (2011). the impact of family relationships and financial education on financial risk tolerance. financial services review, 20, 181–193. sadi, r., ghalibaf a., h., rostami, m. r., gholipour a., & gholipour, f. (2011). behavioral finance: the explanation of investors’ personality and perceptual biases effects on financial decisions. international journal of economics and finance, 3, 234–241. shefrin, h., & nicols, c. m. (2014). credit card behavior, financial styles, and heuristics. journal of business research, 67, 1679–1687. sheppard, b. h., hartwick, j., & warshaw, p. r. (1988). the theory of reasoned action: a meta-analysis of past research with recommendations for modifications and future research. journal of consumer research, 15, 325–343. stango, v., & zinman, j. (2009). exponential growth bias and household finance. journal of finance, 64, 2807–2849. stevens, j. p. (1992). applied multivariate statistics for the social sciences (2nd ed.). hillsdale, nj: erlbaum. strayhorn, j. m. (2002). self-control: theory and research. journal of the american academy of child and adolescent psychiatry, 41, 7–16. witte, k., & morrison, k. (2000). examining the influence of trait anxiety/repression-sensitization on individuals’ reactions to fear appeals. western journal of communication, 64, 1–27. wooldridge, j. m. (2006). introductory econometrics: a modern approach (3ed edition). thomson learning, ct: stamford. yamauchi, k., & templer, d. (1982). the development of a money attitude scale. journal of personality assessment, 46, 522–528. youn, s., & faber, r. j. (2000). impulse buying: its relation to personality traits and cues. advances in consumer research, 27, 179–185. 85a. yazdanparast, y. alhenawi / financial services review 26 (2017) 55–85 finser_23_2 academy of financial services officers president lance palmer university of georgia president-elect william chittenden texas state university executive vice president-program thomas coe quinnipiac university vice president-communications christine mcclatchey university of northern colorado vice president-finance diane docking northern illinois university thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2014 duncan williams william patterson university vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president frank laatsch univ. of southern mississippi editor, financial services review stuart michelson stetson university directors robert moreschi virginia military institute dale domian york university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake jacob sybrowsky utah valley university past presidents frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms can be downloaded from the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2014 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 23, no. 2, 2014 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2014 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. from the editor this issue contains issue 2 of volume 23 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “performance of alternative mutual funds: the average investors hedge fund,” is coauthored by srinidhi kanuri and robert w. mcleod both at the university of alabama. alternative mutual funds (amfs) provide the individual investor with an opportunity to invest in funds that follow strategies similar to those of hedge funds and seek returns uncorrelated with the market. in this article the authors analyze the performance of amfs for a 14-year period using the carhart four-factor model and the fung-hsieh seven-factor model. their results show that most amfs were not able to create any value for their investors over the study period and these funds perfomed even worse during the recent financial crisis. the second article “portfolio performance with inverse and leveraged etfs,” is coauthored by james a. dilellio, rick hesse, and darrol j. stanley all at pepperdine university. in this article the authors examine passive investment strategies that employ inverse or leveraged equity etfs in their asset allocation and quantify the long-term impact on portfolio performance. monte carlo simulations are employed, drawing samples from distributions created by two distinct time periods of historical daily market returns. their findings show that, while these products are generally not recommended within long-term passive investment strategies, potential diversification benefits may exist, dependent on the behavior of equity and debt markets. the third article, “the performance of the faith and ethical investment products: a comparison before and after the 2008 meltdown,” is coauthored by francisca m. beer, james p. estes, and charlotte deshayes all at california state university san bernardino. in this article the authors study the risk and return characteristics of socially responsible investment and faith-based mutual funds prior to and following the market crisis of 2008. their results show a high level of correlation between the indices examined along with a higher volatility than the s&p 500. they also find a significant shift in the mix of performance and volatility of these funds prior to and after the crash of 2008. this is an important consideration for planners and investors in when considering social or faith based investments. financial services review 23 (2014) v–vi 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. the fourth article, “saving for retirement while having more nights with peaceful sleep: comparison of lifecycle and lifestyle strategies from expected utility perspective,” is coauthored by rosita p. chang at university of hawaii at manoa, david hunter at university of hawaii at manoa, qianqiu liu at university of hawaii at manoa, and helen saar at dixie state university. using bootstrapping simulations the authors examine the fit of target-date funds (tdfs) as the main retirement savings instrument for the utilitymaximizing investor who becomes more risk averse as she gets older. the authors also demonstrate that tdfs can provide higher expected utility than the alternative lifestyle strategies. with loss aversion incorporated in the model, they find that the optimal lifecycle strategy over time leads to higher expected utility than the best lifestyle strategy. they conclude that tdfs are preferable to utility-maximizing investors, although an investor’s risk tolerance should be considered when selecting tdf funds. the final article, “the perpetual growth model and the cost of computational efficiency: rounding errors or wild distortions?,” is coauthored by morris g. danielson and jean l. heck both at saint joseph’s university. in this applied article, the authors investigate the constant dividend growth model (gordon, 1962). the model is primarily used because of its computational simplicity, although, value estimates from the model can be highly dependent on future expected cash flows. the authors results suggest that price estimates from the constant growth model can overstate a stock’s intrinsic value by a sizeable amount, in some cases two or three times the underlying intrinsic value. the potential overstatement increases as the firm’s dividend yield decreases, shifting a greater portion of the expected cash flow into later years. because the model is less likely to misstate value for low-growth, highpayout firms, the interesting implication is that the model is most useful when its ability to value growth is needed least. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 23 (2014) v–vi investors’ images of the stock market: antecedents and consequences dawn m. dobnia, marie d. racineb,* adepartment of management and marketing, bdepartment of finance and management science, edwards school of business, university of saskatchewan, 25 campus drive, saskatoon, saskatchewan, canada s7n 5a7 abstract this research studies antecedents and consequences of stock market image, defined as the sum of impressions about the stock market. it seeks to understand how several personality-oriented, cognition-based, and demographic variables influence investors’ images of the stock market and how these images impact investing behaviors and outcomes. the findings from cross-sectional survey data suggest an individual’s financial literacy, propensity to trust, and sociability are important antecedents of his or her perceptions about the stock market. the data also show that stock market image affects investing motives, risk reduction efforts, emotional responses, and degree of satisfaction associated with investing. © 2016 academy of financial services. all rights reserved. jel classification: g19; g10; g02; d10 keywords: stock market image; investor perceptions; stock market investing; stock market participation; retail investor 1. introduction in recent years there has been an interest in how subjective impressions of the stock market affect participation in it. for example, guiso, sapienza, and zingales (2008) studied investors’ perceived trustworthiness of the stock market and found they assess the risk of * corresponding author. tel.: �1-306-966-8406; fax: �1-306-966-2515. e-mail address: racine@edwards.usask.ca (m.d. racine) financial services review 25 (2016) 1–28 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. being cheated when deciding whether to buy in. keller and siegrist (2006) observed that perceptions of stock profit-making and taking as unethical, or of the stock market as a casino, predict behaviors related to trading securities. more generally, macgregor (2002) suggested that the investing behavior of retail investors will be strongly impacted by their images of the stock market, primarily because many of them lack the skills to process the available objective data. an effort to build on this line of inquiry introduced the concept of stock market image and developed a scale for measuring it (dobni and racine, 2015). formally, that work conceptualized and defined stock market image as a multidimensional construct that captures the sum of the total perceptions or impressions an individual holds about the stock market, separate and apart from those that may be held about specific stock or company listings. practically speaking, understanding this “whole picture” view of the stock market was considered to be important because of its potential influence on investors’ behaviors, expectations and experiences, and the potential value of image as a strategic positioning tool (bravo, montaner, and pina, 2012). to some degree, perceptions and impressions about the stock market may be controllable by institutions that facilitate and regulate participation in it (e.g., stock market owners, securities regulators, and investment advisory providers). however, if strategies are to be devised to form, change, downplay, confirm, or otherwise compensate for them, there is a need to better understand their origins and implications (o’shaughnessy, 1987). the main purpose and contribution of this study was to test several hypotheses related to the antecedents and consequences of stock market image. more specifically, it sought to understand how several key personal factors might influence street-level images of the stock market, and how these images might in turn influence investor behaviors and outcomes. as more becomes known about these relationships, strategies to manage or remediate investors’ images of the stock market can be better specified. to accomplish these goals, this article first provides a brief review of the literature on the form and content of stock market image, and develops and advances several hypotheses pertaining to the construct’s antecedents and consequences. next, the field study undertaken to empirically investigate these hypotheses is described, followed by an analysis of the research results and a discussion of robustness. the article concludes by examining the relevance of the findings for managers and identifying research implications and limitations. to guide the discussion that follows, a framework that depicts the key variables and relationships included in the study is presented in fig. 1 below. note that while the concept flow depicted in this figure truly is left to right, and there is a natural implied direction of causality, it is recognized that it is unlikely that the image an investor has in her first investment will be exactly the same image for her tenth investment. thus, while there is in our opinion an inferred direction of causality, the real world investor will likely follow the path depicted in fig. 1 in an iterative process before “settling” on an image that will inform that investor’s decisions. nonetheless, it is rigorously correct to interrogate the impact of stock market image using the process described below, with its theoretically motivated causality direction. 2 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 2. nature of stock market image image research is concerned with the mental impressions or gestalts that are formed about target objects (stern, zinkhan, and jaju, 2001). regarding the stock market, individuals may possess a set of ideas, beliefs, and feelings about it that reflect either a positive or negative evaluation. this might include perceptions of its potential for gain or loss, the safety of its investing environment, and the fairness of its playing field, to name a few. these impressions may or may not be correct, and may be based on information that is less than complete or factual (dichter, 1985). as they fuse into an “image” of the stock market, they are likely to influence an investor’s choices and behaviors. the notion of stock market image-making has deep historical roots. in 1913 the then self-regulated new york stock exchange (nyse) began a public relations program to counteract the threat of external oversight, hoping to transform the stereotype of the nyse from an immoral and engineered “gambling hell” to that of a “free and open” “people’s market” (ott, 2004). decades later, its “own your share of american business” campaign was likewise an effort to reshape public perceptions, this time to encourage mass stockownership (traflet, 2004). marketing tools were used to position stockholding as an act of patriotism and the stock market as a reputable place for ordinary families to benefit from wealth redistribution (traflet, 2004). over the years, the study of image in the financial literature has been quite limited. clark, thrift, and tickell (2004) traced the image of the international finance/investment management industry, observing that it has evolved from a serious, rational entity to an entertainmentbased activity, due primarily to its mediatized environment. image issues around financial intermediaries have also been investigated, such as strader and ramaswami’s (2004) study of investor perceptions of traditional versus online channels for trading securities. the majority of image-related financial studies has focused on the influence of industry, corporate, and brand images or reputations on investor choices and preferences (e.g., aspara and tikkanen, 2008, 2010, 2011; lucey and dowling, 2005; macgregor, slovic, dreman, and berry, 2000; statman, fig. 1. antecedents and consequences of stock market image. 3d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 fisher, and aninger, 2008). the logic underlying this work is that image is a simplifying heuristic that may curtail rational decision making; for better or worse, mental shortcuts have been shown to routinely trump processing effort for investment choices (baker and nofsinger, 2002; da, engelberg, and gao, 2011). more recently, perceptions about the stock market itself have been recognized as having relevance for investors. inquiries of this ilk typically examine one or two image features, like the stock market’s trustworthiness or gambling-likeness (e.g., guiso, sapienza, and zingales, 2008; keller and siegrist, 2006), and correlate them with outcomes like participation or portfolio design. in the marketing and consumer behavior literatures, the study of image relative to objects like product, brand, store, retail, company, and destination has been vast. dominant among this work has been research that focuses on defining and conceptualizing the image construct, and developing and refining both context-specific and general image measurement scales. while there has been wide variation in how these issues are approached, common precedents are to view image as a transactional process between the stimulus object and the perceiver, and the content of image as a collection of perceptions on multiple dimensions (stern, zinkhan, and jaju, 2001). in the consumer behavior tradition, dobni and racine (2015) introduced and examined the concept of “stock market image,” identifying several attributes on which it is based and developing a scale for measuring them. these attributes represented six underlying dimensions: perceptions about the stock market’s immorality (e.g., gambling-likeness, prevalence of insider trading), the effectiveness of its facilitators and regulators (e.g., regulatory credibility, information transparency), its role as an economic bellwether (e.g., utility to economy, measure of economic health), the extent to which it supports wealth creation (e.g., soundness and safety, propensity for long-run payoffs), the extent to which it supports making fast money (e.g., short-term investment value, emphasis on “hot tips”), and the degree to which it is considered to be tilted (e.g., dominance of large investors, bias against small investors). the specific items that comprise this scale are displayed in appendix 1. 3. antecedents to stock market image according to image theory, there are three categories of impression formation agents (burmann, schaefer, and maloney, 2008): the personal and subjective characteristics of the perceivers, determinants related to image management efforts, like the marketing practices of stock market supply chain members, and external uncontrollable factors, such as the media hype surrounding the stock market (clark, thrift, and tickell, 2004). dichter (1985) argues that variables related to the perceiver’s disposition, attitudes, values, and experiences are the most important shapers of image. as such, the area of concern in this study was on the characteristics of individual investors. the first antecedent pertains to trust. trust is thought to be particularly relevant in the finance context because the investing process is complex and indeterminate and has high risk and stakes. olsen (2008) observed that trust lubricates the financial marketplace, encourages most investment, and underpins many stock market pricing anomalies. especially for non-experts, the decision to invest in the stock market is to trust implicitly in the competence and reliability of its expert systems (unerman and o’dwyer, 2004).guiso, sapienza, and zingales (2008) studied 4 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 trust in the stock market, which they described as the subjective probability of being cheated, and found a connection between low levels of trust and stock market avoidance. the propensity to trust has been generally defined as “a stable disposition to believe in the goodness of others” (bernerth and walker, 2009, p. 218), and in the context of stock market investing trust has been defined as “the subjective probability individuals attribute to the possibility of being cheated” (guiso et al., 2008, p. 2557). trust is relevant to stock market image because there is reason to believe it will affect perceptions in a self-fulfilling way through an illusions of control process (bernerth and walker, 2009; mcknight, cummings, and chervany, 1998). that is, to assure themselves that things are under their personal control, individuals will look for and process information, behavior, or other cues in the environment in a way that confirms their level of trust. as such, those who have a more trusting nature are more likely to focus on positive rather than negative information and to interpret activities and events more favorably, and the opposite is likely to hold for those who are less trusting. as mentioned above, increased levels of trust are also associated with perceptions of decreased levels of risk. there is reason to believe that an individual’s sociability will affect his or her perceptions of the stock market. social interaction has been noted to promote stock market participation through such mechanisms as word of mouth information sharing and the enjoyment of discussing the market with peers (hong, kubik, and stein, 2004). talking about the stock market with others is a way not only to gain information about it but also to observe their emotional reactions to it, both of which inform one’s own opinions on the subject (baker and nofsinger, 2002). bravo, montaner, and pina (2012) confirmed the extension of this effect to brand associations. they found word of mouth communications to be a main determinant of the corporate image of financial institutions, corroborating earlier research findings that the opinions of friends, family, and other people mattered more in image development than does organization-sponsored messaging. hence, it is hypothesized as follows: hypothesis 1: investor trust will be a significant predictor of stock market image. hypothesis 2: investor sociability will be a significant predictor of stock market image. images can evolve from different processes, and are dynamic and subject to change. in the destination image literature, for example, images are identified as ranging from organic to induced to complex, a function of the information sources on which they are based and the level of the perceiver’s direct experience with the destination (gunn, 1988). while an organic image arises from non-tourism information, an induced image develops from destination marketing efforts, and a complex image is a product of first-hand experience (byon and zhang, 2010). it is similarly proposed that stock market images can vary as basic financial competences and experiences of investors vary. research shows that stock ownership increases considerably with financial literacy, presumably because low literacy individuals have little knowledge of how the markets work (van rooij, lusardi, and alessie, 2011). as knowledge and experience accumulate, new insights are gained that can fundamentally change beliefs, behaviors, and attitudes (o’shaughnessy, 1987). guiso et al., (2008) observe that new knowledge can help break down the barriers caused by a lack of trust in the stock market. 5d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 further, the literature on consumer expertise suggests that novices and experts process information differently. novices weight negative information more heavily when making evaluations and tend to be influenced more by stereotypical data and less by individuating information (maheswaran, 1994). these differences may be critical to stock market image, as press coverage of the stock market tends toward negativity and sensationalism. experts also have more fully developed schemas of target objects, and better abilities to analyze and elaborate on given information (alba and hutchinson, 1987). as such, investing aficionados can more effectively sort stock market fact from fiction, and more objectively assess the risks and rewards associated with it. therefore, it is hypothesized that: hypothesis 3: investor financial knowledge will be a significant predictor of stock market image. hypothesis 4: investor experience will be a significant predictor of stock market image. the third set of factors hypothesized to affect stock market image relates to demographics. two variables—gender and age—are investigated. research to date provides abundant evidence that men and women differ in financial investment behavior, including their risk tolerance, their interest and confidence in investing, and the composition of their portfolios (hira and loibl, 2008). ford and kent (2010) observed that female and male college students responded differently to perceptions of threat emanating from the stock market. men took a more opportunistic view, seeing its potential for reward, and women were more apt to be intimidated by it and focus on the potential for loss. women were also more pessimistic about future stock market performance, perceiving the outcomes to be more uncertain and uncontrollable. barber and odean (2001) similarly observed gender differences, noting that the overconfidence of men compared with women regarding financial matters leads to unrealistic beliefs about their ability to predict stock market performance. age-related differences in stock market image are likewise expected. popular press reports suggest that most generation y investors will never find the stock market to be in their comfort zones (brown, 2012). for those that might, optimism, cockiness, and long investment payoffs may produce a more opportunistic view (ford, devoto, and kent, 2007). malmendier and nagel (2011) offered evidence of generational effects on beliefs about the stock market because of exposure to different macroeconomic experiences, such as the postwar boom years and stock market peaks and downturns. these memories and influences were found to diminish over time, but particularly for younger people recent experiences were shown to have stronger effects. vissing-jorgensen (2003) similarly observed that different age groups had time-varying differences in expectations about the stock market. thus, it is hypothesized that: hypothesis 5: investor gender will be a significant predictor of stock market image. hypothesis 6: investor age will be a significant predictor of stock market image. 4. consequences of stock market image image is posited to be an important predictor of several consumer responses. in the consumption context, perceptions and beliefs have been credited with influencing all 6 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 aspects of the consumer decision making process, including information search, purchase preferences and choices, post-purchase processes, and future behavioral intentions (o’shaughnessy, 1987). the power of image is said to reside in its use as a heuristic or shortcut to simplify decision making (tversky and khaneman, 1974), particularly in consumption situations that are marked by uncertainty, complexity, or intangibility (stern, zinkhan, and jaju, 2001). the primary mechanisms through which image is thought to exercise influence are by functioning as an information cue and as an expectation (andreassen and lindestad, 1998). expectations are noted to be crucial in economic decisions such as investing, as these decisions are often made under conditions of uncertainty and expectations have a forecast component that will help consumers gauge the likelihood of future outcomes (van raaij, 1991). expectations are adaptive and formed over time, and may or may not be based on rational or solid evidence (aspara and tikkanen, 2008). in addition to shaping the experiences that are sought and expected from it, how a target object is perceived will shape how the perceiver reacts to it (ittelson, 1974). the social psychology research on social stereotypes demonstrates that a perceiver’s actions with respect to a target will be based on stereotype-generated attributions about the target (e.g., such as stereotypes around race, gender, physically attractive vs. unattractive persons; snyder, tanke, and berscheid, 1977). further, these stereotypes were shown to have a self-fulfilling nature: even when the impressions held by the perceivers about the targets were erroneous, they initiated behaviors on the part of the perceiver that behaviorally confirmed the stereotypes. in the same way, it is suggested that stock market image will influence investor behaviors and experiences. this might include the decision to avoid or participate in equities markets, the motives pursued in stock market investing, the ways in which investors mitigate risk, and the specific investment choices they make. regarding investing motives, for example, it could be expected that an investor who sees the stock market as casino-like will be driven by aims related to entertainment or speculation, using the stock market to satisfy needs around recreation, sensation seeking, or wealth aspirations (dorn and sengmueller, 2009). by contrast, impressions of the stock market as a safe and sound place to invest for patient investors could lead to motives for participation such as saving for retirement or accumulating long-term wealth. the proposition that an investor’s perceptions of the stock market will be manifested in his or her responses to it is also supported by theories on attitudebehavior consistency (ajzen and fishbein, 1980; aspara and tikkanen, 2008b; bravo, montaner, and pina, 2012). hence, it is hypothesized that: hypothesis 7: stock market image will be a significant predictor of investing motives. the second consequence of stock market image examined in this study is customer satisfaction. customer satisfaction is a post-purchase construct that is calculated as a comparison of actual with expected outcomes, and in a cumulative sense is defined as an overall evaluation based on consumption experience with a good or service over time (anderson, fornell, and lehmann, 1994). the argument is that an investor develops expectations about the stock market based on the image he or she holds of it, and to the extent these expectations are confirmed or not the satisfaction or dissatisfaction evaluation will be formed. studies in the brand, corporate, and store image literatures offer specific empirical support for the 7d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 relationship posited between image and customer satisfaction (bloemer and deruyter, 1998; martenson, 2007; ryu, han, and kim, 2008). therefore, it is hypothesized that: hypothesis 8: stock market image will be a significant predictor of investor satisfaction. the next set of consequences posited to be influenced by stock market image involves strategies for risk resolution. observing that it is perceived rather than objective risk that directs the decision maker’s behavioral patterns, cho and lee (2006) studied risk-reducing strategies in the context of the perceived level of risk of the stock market. they defined these strategies as behavioral responses adopted by investors to reduce or avoid risk, and noted them to possibly include such actions as engaging in extensive information search, relying on market-provided sources of information, and limiting the proportion of financial assets invested in stocks. their findings showed a significant link between perceived risk associated with the stock market and the use of several of these vulnerability-lowering behaviors. research has shown that investors use schematas and stereotypes like brand image to judge the risk associated with investment assets (ganzach, 2000; jordan and kaas, 2002). aspara and tikkanen (2008b) similarly observe that an investor’s general evaluation of a company can direct various judgments about it, including judgments or expectations of the risk associated with its stocks. in like manner, it is posited that investors’ images of the stock market are evaluative and that inherent in them are expectations and cues about its risk. for example, an investor who views the stock market as being inadequately regulated or corrupt is apt to attribute more risk to it than does an investor who views the regulation as effectively safeguarding investor interests and the playing field as level. correspondingly, different perceptions are likely to cue different use of risk reduction strategies by investors. it is hypothesized that: hypothesis 9: stock market image will be a significant predictor of the use of risk reducing strategies. finally, there are also likely to be emotional or experiential reactions to stock market image. the literature recognizes that emotions are an important part of consumer response (richins, 1997); when consumers shop for, purchase, or use goods and services, they have subjective, internal reactions to them that can vary in strength, intensity, and valence (brakus, schmitt, and zarantonello, 2009). in consumer research there has been wide variation in the content of emotional responses studied, these categorized hierarchically by laros and steenkamp (2005) from the superordinate level of negative and positive affect to a subordinate level of specific emotions that tap anger, fear, shame, contentment, and happiness. just as emotions can impact how a situation is perceived, emotional experiences can be elicited by perceptions. applying the stimulus-organism-response framework to a hospitality setting, jang and namkung (2009) found that perceptions about restaurant offerings helped to provoke emotions in patrons. the negative social connotations associated with gambling have been noted to generate feelings of guilt in gamblers (mageau, vallerand, rousseau, ratelle, and provencher, 2005). similarly, it is argued that the affective states of investors may be evoked by their perceptions of the stock market. for example, some may associate investing with the experiences of helplessness, loss of control, titillation, or excitement, 8 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 based on the extent to which they see the stock market as a casino or its perils as daunting. levels of anxiety, fear, confidence, or worry may be a function of how such stock market features as integrity, expected payoffs, and utility are processed. the stock market is widely perceived to be an emotional place, and it is hypothesized here that impressions about it can in turn affect an investor’s emotional space: hypothesis 10: stock market image will be a significant predictor of affective states of investors. 5. research methodology this study was conducted as part of a larger research project that examined several issues around stock market image. to investigate the hypotheses of interest, a quantitative survey research design was used. the sample for the study was comprised of individuals 18 years and older living in canada who were then currently invested in the stock market or who had previous stock market investing experience, leading to a total of 573 eligible respondents. sample members were recruited from the canada talk now online research panel, and the data were collected online over the period of april 18–25, 2012 by the social sciences research laboratory at the university of saskatchewan. the canada talk now panel has over 60,000 active members and based on total invites broadcasted to total starts, drop-outs, terminates, and quota fills experiences an overall cooperation rate of 15%. relevant statistics on the sample including age, gender, and education are provided in table 1. the sample is split with 49% male and 51% female. on average the respondents are between 56 and 65 years of age, have a technical diploma or certificate, 11.7 years of investing experience and an average net wealth of $383,140. 5.1. survey instrument the survey instrument included all variables identified in fig. 1. the measures for them are briefly described below and are reproduced in appendixes 1 and 2. to operationalize stock market image, the scale developed by dobni and racine (2015) was used. this scale is comprised of 32 statements regarding attributes of the stock market that are scored on a seven-point likert-type scale. as described earlier, these statements represent six underlying image dimensions: eight pertain to the “immorality” dimension, eight to “facilitators and regulators,” four to “economic bellwether,” four to “wealth creating capacity,” three to “fast money,” and three to the “tilted playing field” dimension. the personality and cognition-based antecedent variables were measured as follows: disposition to trust by a three-item scale that asked about trust in and perceived reliability of other people; sociability by a six-item scale in which respondents reported the frequency of their interactions with others in various contexts; financial literacy by a set of ten quiz-like multiple choice questions that tested knowledge of the stock market and investing concepts; and investing experience by asking respondents the number of years they had been investing in the stock market. 9d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 regarding the consequences variables, four separate investing motives—wealth accumulation, hobby and entertainment, saving for retirement, and speculation—were measured by asking respondents the extent to which each described their reasons for participating in the stock market. investing satisfaction tapped the extent to which respondents were satisfied with their stock market returns. two variables pertaining to risk-reducing strategies were assessed: portfolio diversification by the number of stocks respondents typically hold in their portfolios and information search by the number of sources they typically consult for investment decisions. three separate affective states concerning control, excitement, and panic were modeled by asking respondents to record on semantic differential scales how they typically feel about being invested in the stock market. sociodemographic data collected included age, gender, education, and wealth. gender is coded as one for male and two for female. net wealth is calculated as the difference between total value of assets and total value of debts. based on the existing literature, risk proneness, optimism, education, and net wealth were used as control variables (cavezzali and rigoni, 2012; de bondt, 1998; riley and chow, 1992). risk proneness was measured by asking table 1 descriptive statistics for variables used in the antecedent and consequence regressions variable meana standard deviation antecedents trust 3.82** 1.07 sociability 3.45** 0.93 financial literacy 5.07** 2.58 investing experience 11.71** 7.51 age 56–65 1.33 gender (1 � male, 2 � female) 49% male, 51% female 0.50 stock market image dimensions immoral 3.64** 0.95 facilitators and regulators 4.13** 0.98 economic bellwether 4.95** 0.94 wealth creation 4.08** 0.95 fast money 3.67** 1.00 tilted playing field 4.54** 1.02 consequences wealth accumulation 4.91** 1.53 hobby 2.55** 1.67 speculation 2.86** 1.75 satisfaction 4.09** 1.65 information search 2.35** 1.37 diversification 7.43** 8.38 out of control-in control 4.43** 1.55 panicky-at-ease 4.42** 1.30 excited-bored 3.47** 1.21 controls education technical diploma 1.35 net wealth $383,140** 60,0961 risk proneness 3.57** 1.39 optimism 4.73** 1.04 a **,* indicate statistical significance at 1% and 5%, respectively. 10 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 individuals about their willingness to take risks in financial decisions. an individual’s level of optimism was assessed by averaging levels of agreement with six statements related to positivity about his or her future. 5.2. data analysis the potential antecedents of stock market image tested were trust, sociability, financial literacy, investing experience, age, and gender. the relationships between the antecedents and each of the dimensions of stock market image were examined using a linear regression of the form: yj � � � � i�1 6 �ixi � �j (1) where yj � stock market image dimension j � 1 to 6, [y1 � immorality, y2 � facilitators and regulators, y3 � economic bellwether, y4 � wealth creation, y5 � fast money, y6 � tilted playing field], xi � one of the six antecedents, and �j is the error term. regressions of eq. (1) test our understanding of the factors that may influence an investor’s image of the stock market. similarly, linear regression analysis was used to study the influence that an individual’s image of the stock market may have on her or his investment motives, risk reduction strategies, emotional responses to, and satisfaction from, investing. these relationships were investigated using regression equations in the following format: cj � � � � i�1 6 �iyi � � i�1 n iwi � j (2) where cj represents consequences (risk reducing strategies [information search, diversification], affective states [excited-bored, out of control-in control, and panicky-at-ease], investing motives [wealth creation, hobby, and speculation] and satisfaction). yi � stock market image dimension i � 1 to 6 (as presented above), wi � one of the control variables (risk proneness, optimism, education, and net wealth), and j is the error term. the regression implied by eq. (2) allows an analytical assessment of the impact of stock market image on investment decisions. before analyzing the results of the multivariate analyses in eqs. (1) and (2), univariate descriptive statistics were studied. all statistical tests were done at two levels of significance (0.05 and 0.01) but analysis within the text is based on a 5% level of significance unless otherwise noted. an f test on the goodness of fit is used for each regression. the individual hypotheses were tested using t tests and joint hypotheses were based on f tests. tables 2 and 3 provide the univariate statistics (pairwise correlations). table 4 amalgamates the regression results for each of the six stock market image dimensions on the six antecedents. similarly, table 5 summarizes the consequence and stock market image regressions. to avoid problems with units of measurement, the regression coefficients were standardized by multiplying the parameter estimate by the ratio of the standard deviation of the regressor to the standard deviation of the dependent 11d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 variable. this allowed a comparison of the magnitude of the coefficients both within and across regressions. 6. results and discussion 6.1. antecedents the correlations of the antecedents with the stock market image dimensions (table 2) show support for hypothesis 1. trust has significant relationships with five of the six dimensions considered: positive correlations with facilitators and regulators (0.17) and economic bellwether (0.22), and negative relationships with immorality (�0.43), fast money (�0.15), and tilted playing field (�0.25). both the magnitude of the correlations and their signs are what might be expected; overall, a less trusting personality is associated with more unpalatable views of the stock market. according to hypothesis 2, sociability and the stock market image dimensions should be related and the correlations support this for facilitators and regulators (0.15), wealth creation (0.14), and fast money (0.11). under hypothesis 3, we expect that financial literacy will impact one’s image of the stock market. the correlations show that the financially literate investor tends to see the stock market more as an economic bellwether (0.26) or a tool for wealth creation (0.18) and less as an immoral (�0.31) instrument for fast money (�0.29) in a tilted playing field (�0.17). it is interesting that less virtuous images of the stock market are associated with lower degrees of financial literacy. similarly, investing experience should have a relationship with stock market image (hypothesis 4), and this is consistent with the correlations between investing experience and the immorality (�0.22), economic bellwether (0.19), wealth creation (0.10), and fast money (�0.21) dimensions. gender differences only have an impact on the wealth creation perception of the market (�0.16): being a female tends to reduce the view of the market as a vehicle for wealth creation. this provides limited support for hypothesis 5 (that gender should be a significant predictor of stock market image). finally, under hypothesis 6 we expect and find a relationship between age and stock market image. the older an investor the less she/he views the market as immoral (�0.15) and a place for fast money (�0.17) and more as an economic bellwether (0.13). overall, the pairwise correlations support hypotheses 1–6 with trust and table 2 correlations of the antecedents and the stock market image dimensions used in eq. (1) variable trust social financial literacy investing experience gender (2 � female, 1 � male) age immoral �0.4248** 0.0496 �0.3073** �0.2166** �0.0810 �0.1487** facilitators and regulators 0.1747** 0.1517** �0.0628 0.0101 0.0712 �0.0376 economic bellwether 0.2222** 0.0687 0.2556** 0.1891** 0.0059 0.1236** wealth creation 0.0652 0.1404** 0.1752** 0.0918* �0.1557** �0.0681 fast money �0.1508** 0.1097** �0.2864** �0.2056** �0.0391 �0.1725** tilted playing field �0.2510** 0.0328 �0.1669** �0.0484 0.0129 0.0458 **,* indicate statistical significance at 1% and 5%, respectively. 12 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 t ab le 3 c or re la tio ns of co ns eq ue nc es w ith st oc k m ar ke t im ag e di m en si on s w ea lth h ob by sp ec ul at io n sa tis fa ct io n in fo rm at io n se ar ch d iv er si fic at io n o ut -o fin -c on tr ol pa ni ck yat -e as e e xc ite dbo re d im m or al � 0. 17 80 ** 0. 07 34 0. 15 94 ** � 0. 28 63 ** � 0. 08 76 � 0. 16 97 ** � 0. 33 43 ** � 0. 34 77 ** 0. 08 75 fa ci lit at or s an d re gu la to rs 0. 17 21 ** 0. 21 75 ** 0. 17 49 ** 0. 40 53 ** 0. 01 15 0. 05 26 0. 32 70 ** 0. 23 22 ** � 0. 20 24 ** e co no m ic be llw et he r 0. 24 93 ** 0. 10 57 * 0. 01 27 0. 35 13 ** 0. 07 85 0. 18 38 ** 0. 32 04 ** 0. 23 55 ** � 0. 21 11 ** w ea lth cr ea tio n 0. 29 19 ** 0. 33 75 ** 0. 18 18 ** 0. 45 13 ** 0. 14 41 ** 0. 17 01 ** 0. 38 65 ** 0. 25 05 ** � 0. 29 98 ** fa st m on ey � 0. 01 95 0. 16 87 ** 0. 32 43 ** 0. 06 09 � 0. 00 44 � 0. 10 94 * 0. 00 69 0. 00 25 � 0. 11 31 * t ilt ed pl ay in g fie ld � 0. 07 65 � 0. 04 42 0. 06 66 � 0. 27 06 ** � 0. 04 32 � 0. 10 57 * � 0. 25 99 ** � 0. 30 04 ** 0. 14 66 ** * * ,* in di ca te st at is tic al si gn ifi ca nc e at 1% , an d 5% , re sp ec tiv el y. 13d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 financial literacy appearing to be the most important antecedents both in terms of (absolute) magnitude of correlation as well as frequency of significant correlations. multivariate regression analysis was used to assess the unique relationships between stock market image and each antecedent while controlling for the influence of all the remaining right hand side variables. tests for multicollinearity (discussed in the robustness section) show that the coefficient estimates are reliable and the f-tests indicate that each model is statistically relevant so we precede with presentation of the individual regression results. as illustrated in table 4, the antecedents as a group have the highest goodness of fit (r2 � 0.2771) when explaining immorality and the next best fit is for the fast money dimension (r2 � 0.1427). overall the most influential antecedents of stock market image, both in terms of frequency of significance and coefficient magnitude, are financial literacy, trust, and sociability. however, as the discussion below illustrates, the key antecedents vary across image dimensions. the data continue to support the intuition that a person’s ability to trust will impact his or her views of the stock market. similar to the simple correlation results, trust plays a significant role in shaping the image of the stock market on five of the six image dimensions. the more trust investors have, the less likely they think the stock market is immoral (b � �0.38), a venue to make fast money (b � �0.11), or a tilted playing field (b � �0.26), and the more likely they view it as being effectively facilitated and regulated (b � 0.16) and an economic bellwether (b � 0.16). only wealth creating capacity is not significantly related to trust, perhaps suggesting that investors rely on something more tangible than trust when assessing the stock market as a reliable tool of wealth creation. thus, in general, the results support hypothesis 1: investor trust will be a significant predictor of stock market image. does social interaction impact one’s view of the stock market? hypothesis 2 is supported by a significant positive relationship between sociability and facilitators and regulators (b � 0.12), and wealth creating capacity (b � 0.15). furthermore, while a more social person sees the stock market as a place to make fast money (b � 0.10), there is no significant connection between sociability and the tilted playing field dimension. the nature of the sociability may table 4 antecedents of stock market image: standardized regression coefficients, eq. (1) independent variables dependent variables immorality facilitators and regulators economic bellwether wealth creation fast money tilted playing field trust �.3783** .1614** .1623** .0582 �.1123** �.2637** sociability .0725 .1224** .0736 .1530** .1028* .0609 financial literacy �.2738** �.0880 .2315** .1333** �.2684** �.1591** investing experience �.0632 .0317 .0862 .0894 �.0854 �.0023 gender (male � 1, female � 2) �.1170** .0106 .0477 �.1487** �.1186** .0065 age �.0349 �.0506 .0488 �.1141* �.0771 .1252** r2 .2771** .0554** .1233** .0871** .1427** .1040** n 527 534 538 538 546 550 **,* indicate statistical significance at 1% and 5%, respectively. 14 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 t ab le 5 c on se qu en ce s of st oc k m ar ke t im ag e: st an da rd iz ed re gr es si on co ef fic ie nt s, e q. (2 ) d ep en de nt va ri ab le s in ve st in g go al s an d ex pe ct at io ns r is k re du ct io n st ra te gi es a ff ec tiv e ex pe ri en ce s in de pe nd en t va ri ab le s w ea lth h ob by sp ec ul at io n sa tis fa ct io n in fo rm at io n se ar ch po rt fo lio di ve rs ifi ca tio n o ut -o fin -c on tr ol pa ni ck yat -e as e e xc ite dbo re d in te rc ep t 0. 00 00 0. 00 00 * 0. 00 00 * 0. 00 00 0. 00 00 0. 00 00 0. 00 00 * 0. 00 00 ** 0. 00 00 ** im m or al ity � 0. 01 66 0. 26 45 ** 0. 19 74 ** 0. 01 53 � 0. 07 63 � 0. 05 33 � 0. 12 70 * � 0. 21 51 ** � 0. 11 40 fa ci lit at or s an d re gu la to rs � 0. 04 77 0. 13 26 * 0. 14 98 * 0. 21 04 ** � 0. 11 56 � 0. 13 63 0. 06 61 0. 03 19 � 0. 04 25 e co no m ic be llw et he r 0. 11 99 � 0. 01 08 � 0. 00 81 0. 01 69 0. 02 17 0. 10 08 0. 02 53 � 0. 04 35 � 0. 12 17 w ea lth cr ea tio n 0. 21 71 ** 0. 24 05 ** 0. 00 96 0. 23 37 ** 0. 13 38 0. 19 22 ** 0. 26 62 ** 0. 08 73 � 0. 15 03 * fa st m on ey � 0. 07 30 � 0. 02 16 0. 23 68 ** � 0. 07 12 0. 02 75 � 0. 09 55 � 0. 05 26 0. 03 18 � 0. 02 36 t ilt pl ay in g fie ld 0. 08 42 � 0. 05 01 0. 00 82 � 0. 10 45 * 0. 01 68 0. 01 73 � 0. 05 94 � 0. 11 22 * 0. 09 92 r is k pr on en es s 0. 16 79 ** 0. 23 90 ** 0. 24 97 ** 0. 12 25 ** 0. 14 30 ** 0. 04 59 0. 11 95 ** 0. 19 23 ** � 0. 20 65 ** o pt im is m 0. 12 71 ** � 0. 01 24 0. 02 41 0. 09 93 * � 0. 04 05 0. 09 75 0. 14 63 ** 0. 13 50 ** � 0. 00 34 e du ca tio n 0. 04 43 0. 05 06 0. 02 58 0. 08 43 0. 18 97 ** 0. 01 09 � 0. 03 50 � 0. 00 68 � 0. 12 23 ** n et w ea lth 0. 07 42 0. 02 12 � 0. 01 24 0. 04 88 0. 08 00 0. 13 56 ** � 0. 00 94 0. 01 33 0. 06 21 r 2 0. 15 92 ** 0. 19 44 ** 0. 19 85 ** 0. 27 58 ** 0. 10 39 ** 0. 10 89 ** 0. 26 01 ** 0. 21 33 ** 0. 16 08 ** n 42 9 42 7 42 8 42 2 42 5 42 1 43 0 42 9 42 8 * * ,* in di ca te st at is tic al si gn ifi ca nc e at 1% an d 5% , re sp ec tiv el y. 15d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 play a role in this regard, for example, individuals who more regularly interact with a gaming crowd may absorb different notions about stock market biases than those who travel in more conservative investing circles. of the two cognition-based antecedents, financial literacy appears to be more dominant than investing experience in explaining stock market image. financial literacy is expected to lead to the ability to properly digest media coverage of the stock market and to appropriately judge economic events. the findings suggest that the financially savvy individual tends to view the stock market as a sounder investing environment. the more financially literate investor sees the stock market as an economic bellwether (b � 0.23) and a venue for wealth creation (b � 0.13), and is less likely to see it as immoral (b � �0.27), a place to make fast money (b � �0.27), or a tilted playing field (b � �0.16). of interest to the authors, financial literacy is negatively but not significantly related to perceptions of the efficacy of stock market facilitators and regulators, possibly reflecting beliefs that these participants are outmanned, under-resourced, or lack expertise, or are simply not relevant to the financially savvy investor. the second cognition variable, years of investing experience, mathematically strengthens the image of the stock market on every dimension but is not statistically significant. overall, analysis of the cognition antecedents suggests that financial knowledge improves investor perceptions of the stock market, which is consistent with hypotheses 3, and that it may be the substance of an investor’s experience rather than the tenure that is relevant to stock market image formation. each of the demographic variables leads to similar implications. the gender variable indicates that being a female is consistent with a perception of the stock market not only as immoral (b � �0.12), but also as a place to make fast money (b � �0.12) and to create long-term wealth (b � �0.15). these findings parallel in part other research which suggests that men are more optimistic than women about stock market performance (ford and kent, 2010). age significantly impacts two stock market image components: the older the investor is the more likely the stock market is seen as representing a tilted playing field (b � 0.13) and less likely as providing a venue for wealth creation (b � �0.11). being female or growing older suggests that one tends to see the earnings potential of the stock market in a more pessimistic light. thus, both the bivariate and multivariate analysis supports the hypotheses to varying degrees. in particular, trust and financial literacy appear to be the most important antecedents for the six image dimensions. 6.2. consequences stock market image is also an important predictor of investor consequences. of the six image dimensions studied, a perception of the stock market as a place for wealth creation is the most influential and as a venue for fast money is the least influential across the nine consequences. overall, both the univariate and multivariate results demonstrate strong support for the posited relationships between the dimensions of stock market image and investing motives (hypothesis 7), investor satisfaction (hypothesis 8), the use of risk reducing strategies (hypothesis 9), and investors’ affective states (hypothesis 10). to begin the investigation, the correlations between each image dimension and consequence were calculated and reported in table 3. tabulating the frequency of significant 16 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 correlations of the image dimensions across the consequences indicates that wealth creation is significantly correlated with all nine consequences, facilitators and regulators, and economic bellwether are significantly correlated with seven, immorality with six, tilted playing field with five, and fast money is correlated with four consequences. the highest significant correlations, in absolute value, are between the wealth creation image dimension and satisfaction (0.45), between facilitators and regulators and satisfaction (0.41), and between wealth creation and the affective state out-of-control-in-control (0.39). thus, the pairwise correlations support relationships between stock market image dimensions and consequences with different dimensions being more important for different consequences. multivariate regression analysis was then performed to investigate the relationship between each consequence and the six stock market image dimensions while controlling for risk proneness, optimism, education and net wealth. each of the nine regressions has an r-squared that is significantly different from zero at 1% (table 5). in the robustness section, we also investigate potential issues caused by collinear explanatory variables and endogeneity between the image dimensions and the consequences and conclude that we can proceed with the results in table 5. in general, the evidence supports significant relationships between the consequences and the stock market image dimensions. the individual results and the regression diagnostics will be discussed in more detail in the following sections. the most important contributor, both in terms of magnitude and statistical significance, to pursuing a wealth accumulation motive is the view of the market as a place for wealth creation (� � 0.22). none of the other stock market image dimensions is significantly involved. hobby and speculation have similar responses to the stock market image dimensions: they both have positive and significant associations with the immorality and the facilitators and regulators dimensions. the more immoral the stock market is perceived to be, the more likely that the investing motive is recreational (hobby � � 0.26, speculation � � 0.20). hobby and speculation investors also need to believe that there is an efficacious facilitator and regulator presence in the market (hobby � � 0.13 and speculation � � 0.15). in addition, a perception that markets are effective tools of wealth creation contributes to the pursuit of investing as a hobby (� � 0.24), while a speculation motive and views of the markets as a place to make fast money are significantly related (� � 0.24). neither of these two investing motives is related to the tilted playing field or economic bellwether dimensions. thus, in general, even in the presence of controls and allowing for multivariate relationships, the evidence supports a relationship between stock market image dimensions and the investing motives (hypothesis 7). consistent with hypothesis 8, an investor’s satisfaction with her or his stock market experience is positively related to views that the stock market has regulatory credibility (� � 0.21) and creates sound, long-run payoffs (� � 0.23) and negatively related to the perception that it is biased against small investors (� � �0.10). thus, a stock market that investors perceive to be replete with ethical professionals and effective regulation and an equitable playing field for serious wealth creation potential will contribute to investor satisfaction. it appears that investors who perceive the stock market as a lucrative vehicle for building wealth will more actively pursue risk reduction strategies (information search and portfolio diversification efforts increase by 0.13 and 0.19 , respectively). on the other hand, diversification efforts will be curtailed when the stock market is viewed as a venue for making fast money. this 17d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 is likely an effort to leverage the perceived upside of this investing venue. if investors see the markets as primarily for short term goals (fast money), they either do not bother or they diminish risk reduction strategies as they see no value to their efforts or are possibly guided by an attitude of recklessness. consequently, the evidence suggests that the wealth creation image dimension provides the strongest support for risk reduction efforts (hypothesis 9). the affective states tend to be related to wealth creation or to a perception of stock market corruption or to both. an investor will tend towards the excited end of the excited-bored spectrum and experience an increased feeling of being in control as the wealth creation dimension strengthens (� � �0.15 and 0.27, respectively) but it does not affect the panicky-at-ease state. two of the three affective states are significantly negatively affiliated with the immorality dimension (out-of-in-control � � �0.13 and panicky-at-ease � � �0.22). the panicky experiences are also enhanced with the tilted stock market image (panicky-at-ease � � �0.11). these results suggest that hypothesis 10 (affective states) is supported particularly by the wealth creation and immoral image dimensions. in all of the consequence regressions, the control variables were risk proneness, optimism, education and net wealth. these were significant in 89%, 56%, 33%, and 22% of the regressions respectively. further, the standardized regression coefficients for risk proneness were larger, in absolute value, than those for any of the other control variables (� ranges from 0.12 to 0.25), except in the information search regression where education had the strongest impact (� � 0.19). because the stock market is universally accepted as a risky environment, it is reasonable to see that risk proneness played a key role in all but one consequence relationship. as a robustness check, all regressions were rerun without the control variables. the significance of the estimated regression coefficients were virtually unchanged and the overall conclusions that wealth creation and an immoral stock market image dimension are the two most important dimensions for the consequences remains. 6.3. robustness this section addresses several empirical issues that could affect the regression results starting with an investigation into potential multicollinearity in the estimation of eqs. (1) and (2). we also consider endogeneity between the consequences and the stock market image dimensions in the estimation of eq. (2). 6.3.1. multicollinearity collinearity among the right hand side variables in any regression can lead to large variances for the coefficient estimates and tends to manifest itself in rejection of statistical tests and coefficients with incorrect signs. to assess the possibility of multicollinearity among the explanatory variables of the antecedent regressions (eq. 1), we examine the pairwise correlations and the variance inflation factors (vif) as part of the regression diagnostics. the pairwise correlations between the right hand side variables (table 6) indicate that the strongest significant relationship is between investing experience and age (0.43) and the second highest is between financial literacy and investing experience (0.29) and between gender and financial literacy ( �0.29 ). a pairwise correlation of at least 0.8 (in absolute value) is considered a high correlation (kennedy, 2008). because none of our 18 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 correlations meet or exceed this cutoff, we do not expect issues because of multicollinearity. in addition, to deal with possible linear combinations involving more than two variables, the vif is calculated for each explanatory variable in each regression. although the vif is not a statistical test, a commonly applied rule of thumb is that a vif �10 indicates problems because of multicollinearity (kennedy, 2008). in each regression the vif for each independent variable is less than 1.34, which is well below the rule of thumb cutoff of 10, suggesting that multicollinearity is not an issue in estimating eq. (1). thus, based on the correlations and the vif, multicollinearity is not expected to interfere with the regression output for eq. (1). the explanatory variables in the consequence regression are the six stock market image dimensions and four controls. again, we use correlations and the vif to explore possible multicollinearity. the strongest correlations among the image dimensions are for facilitators and regulators with wealth creation (0.64), immoral with economic bellwether (�0.57), facilitators and regulators and economic bellwether (0.57), wealth creation and economic bellwether (.52) (table 7). the lowest significant correlation is 0.10 and 50% of the significant correlations are between 0.2 and 0.3 . among the control variables, the most notable significant correlations are between optimism and net wealth (0.19) and next biggest is between education and net wealth (0.18). because none of the correlations exceeds 0.8, we do not expect issues because of multicollinearity. in addition, the vif is calculated for each explanatory variable in each consequence regression. the calculated vifs ranged in value from 1.14 to 2.43. thus, according to this rule of thumb, multicollinearity is not interfering with the coefficient estimates nor their interpretation. 6.3.2. endogeneity no empirical method can prove causation. our theory and our model (eq. 2) suggest that the stock market image dimensions may predict the consequences. it is possible that the consequence variables and the stock market image dimensions are endogenous and thus our ordinary least squares (ols) regression results would be biased and inconsistent. if this is the case, then an instrumental variables (iv) estimator rather than ols should be used to estimate the equation. we used the hausman test to determine if ols or two stage least squares (2sls) should be used to estimate the equation. to do the hausman test and 2sls, instruments for the potentially endogenous variables must be specified. the instruments must be (1) highly correlated with the endogenous variable and (2) not correlated with the error table 6 correlations of the right hand side variables in the antecedent regressions variable trust social financial literacy investing experience gender (2 � female, 1 � male) age trust 1 social 0.1202** 1 financial literacy 0.0954* �0.1024* 1 investing experience 0.1402** �0.0149 0.2919** 1 gender (2 � female, 1 � male) 0.1476** 0.0972* �0.2897** �0.0442 1 age 0.1482** �0.0914* 0.1140** 0.4296** �0.0559 1 **,* indicate statistical significance at 1% and 5%, respectively. 19d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 t ab le 7 c or re la tio ns of th e ri gh t ha nd si de va ri ab le s in th e co ns eq ue nc e re gr es si on s im m or al fa c an d re g e co n be ll w ea lth cr ea te fa st m on ey t ilt pl ay fie ld r is kp o pt e du c n et w im m or al 1 fa c an d re g � 0. 40 16 ** 1 e co n be ll � 0. 57 25 ** 0. 56 63 ** 1 w ea lth cr ea te � 0. 27 26 ** 0. 63 84 ** 0. 52 00 ** 1 fa st m on ey 0. 22 91 ** 0. 33 95 ** 0. 07 79 0. 32 24 ** 1 t ilt pl ay fie ld 0. 52 80 ** � 0. 23 2* * � 0. 34 14 ** � 0. 22 18 ** 0. 10 76 * 1 r is kp � 0. 14 25 ** 0. 11 60 * 0. 12 78 ** 0. 26 51 ** 0. 00 48 � 0. 16 79 ** 1 o pt � 0. 31 01 ** 0. 16 77 ** 0. 22 66 ** 0. 14 39 ** � 0. 08 28 � 0. 22 64 ** 0. 09 80 * 1 e du c � 0. 06 86 � 0. 08 43 � 0. 01 24 0. 08 05 � 0. 06 78 0. 03 56 0. 04 78 0. 13 96 ** 1 n et w � 0. 08 37 0. 01 16 0. 07 12 0. 07 74 � 0. 07 54 � 0. 09 20 � 0. 01 03 0. 19 23 ** 0. 18 16 ** 1 * * ,* in di ca te st at is tic al si gn ifi ca nc e at 1% an d 5% . re sp ec tiv el y. im m or al � im m or al ity ; fa c an d re g � fa ci lit at or s an d re gu la to rs ; e co n be ll � ec on om ic be llw et he r; w ea lth c re at e � w ea lth cr ea tio n; t ilt pl ay fie ld � til te d pl ay in g fie ld ; r is kp � ri sk pr on en es s; o pt � op tim is m ; e du c � ed uc at io n; n et w � ne t w ea lth . 20 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 term. finding suitable instruments is always an issue. the three most common approaches are: (1) finding a suitable exogenous variable, (2) using the rank of the endogenous variable, and (3) using the lag of the endogenous variable. we used the second approach because the correlations of the stock market image dimensions were strong (0.95 to 0.98), but, as previously noted, the controls and the stock market image dimensions had weak correlations. furthermore, we did not use the third approach because it is meant for time series data. the correlation of each stock market image dimension and its rank ranged from 0.95 to 98. the hausman test applied to eq. (2) using the rank of the stock market image dimension variables as their instruments indicates that ols is preferred to 2sls. this does not mean that we have proved that the image dimensions cause the consequences but rather it suggests that the best way to estimate the specified relationship is to use ols rather than 2sls. thus, our results are robust to potential endogeneity and multicollinearity issues. 7. implications, limitations, and future research directions as it is well recognized that perception is more important than reality, a focus on subjective impressions is promising for stock market research. this study builds on the recent introduction of stock market image to the finance literature (dobni and racine, 2015), proposing and testing several antecedents and consequences of the construct. it emphasizes the need to understand perceptions of the stock market in a comprehensive and integrated way, and helps to fill research gaps in the processes by which images are formed and how these images in turn impact behavior (bravo et al., 2012). insights from this study will assist efforts to manage stock market image as a resource. the findings of this study suggest that different dimensions of stock market image are affected by different subsets of antecedents. supply chain members should contemplate their roles in shaping stock market image and how these findings might relate to that task. for example, the financial literacy antecedent significantly affects five of the six image dimensions, adding to the growing call for promoting financial literacy initiatives among the public (van rooij et al., 2011). the trust and sociability variables significantly affect five and three image dimensions, respectively, suggesting that strategies for stimulating positive word of mouth about and bolstering public trust in the stock market may be useful. these results confirm links previously observed between these variables and investing behavior (guiso et al., 2008; hong et al., 2004), but indicate an alternative mechanism through which they might operate. the results of this research were consistent with the cognitive-affective-behavioral model advanced by ajzen and fishbein (1980), as stock market image dimensions had a significant effect on several behavioral variables. for instance, investors who view the stock market as a wealth creating vehicle seemed to take it more seriously and enjoy more rewarding experiences. perceiving the stock market as immoral was shown to lead to stronger emotional responses in the form of feelings of panic and loss of control and to encourage investing for hobby or speculation purposes. as they help to inform and shape investing responses and outcomes, it is important to understand perceptions about the stock market 21d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 (including distortions, biases, and misperceptions) and how they are generated so that strategies for improving the investor experience can be better designed (o’shaughnessy, 1987). future research could explore other demographic, socioeconomic, and psychographic antecedents of stock market image. for example, this study did not account for other agents that may impact image, such as marketer-controlled variables and external stimuli like the media, or how the effect of antecedents and stock market image might change in response to external shocks such as a financial crisis. more comprehensive models that include these agents need to be developed and tested. it is also arguable that some of the variables specified as consequences of stock market image could be treated as its antecedents. these include satisfaction with stock market returns and emotional reactions to stock market investing. future studies should more definitively assess the size and direction of these relationships. trust was shown to be an influential determinant of stock market image, and it needs to be more fully understood. this could entail developing a context-specific definition and measure of trust and identifying the attributes and actions that engender it. other important questions include: what mental calculations occur when investors make the decision to trust? what are the drivers and root causes of trust and distrust, and are they heterogeneous across investor segments? how are the hurdles created by low levels of trust best overcome? there may be value in probing individual image dimensions, such as the tilted playing field component. because perceived fairness of markets may be a hygiene issue for most lay investors, the factors that shape this dimension, investor understanding of its implications, and ways to adjust for it should be studied. the fast money dimension is linked to behaviors consistent with a financial playground; as some investors are driven by recreational motives, possible research topics include a deeper understanding of their expectations and outcomes, and ways to increase the entertainment value attached to their investing activity. future research could also address the efficacy of executing image management strategies one dimension at a time, including how doing so disrupts overall image profiles. to what extent do the dimensions evolve in tandem, and how does performance on one affect performance on the others? this research was cross-sectional and thus was not designed to examine the processes involved in changing perceptions of the stock market. it would be useful to study organizations engaged in rehabilitation strategies, such as the china securities regulatory commission (see philip, 2012), to better illuminate wrong and right ways to repair or enhance image. this work could track how the images held by investors change over time as the strategies are executed, including how they incorporate new information into existing impressions, which communicative sources and rehabilitative efforts hold the most sway, and how resistant to change image dimensions are. perceptions about the stock market are likely to elicit psychological and behavioral responses beyond the consequences studied here, and future research should identify and investigate them. possible subjects include investor trading practices, decision making, and risk and return management strategies. further, the impact of stock market image relative to other factors that influence investment decision making should be gauged. such insight could go a long way to revealing how much stock market image matters and how much effort should be devoted to monitoring, measuring, and managing it. 22 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 appendix 1 stock market image scale descriptive title and image statements cronbach’s � factor loading immorality .851 the stock market is corrupt. .702 the stock market is rigged. .701 the stock market is under-regulated. .669 the stock market is harmful to society as a whole. .643 investing in the stock market is for suckers. .585 in their ongoing publicity efforts, publicly traded corporations commonly mislead investors. .574 use of insider information is common in the stock market. .527 losses and gains in the stock market are just a matter of chance. .496 facilitators and regulators .876 in general, financial services professionals (for example, financial planners, stock brokers) are trustworthy and honest. .725 in general, financial services professionals (for example, financial planners, stock brokers) have the best interests of investors in mind. .725 in general, financial services professionals (for example, financial planners, stock brokers) provide good information to help make stock market investment decisions. .669 stock market regulators do a good job of safeguarding investor interests. .584 the financial information that publicly traded companies disclose is straightforward and honest. .572 regulation of insider trading is effective. .553 the stock market is fair for all investors. .538 stock market investors are adequately protected by antifraud and mandatory disclosure rules. .557 economic bellwether .790 the stock market plays an important role in supporting the growth of the economy. .721 the stock market is a measuring stick of the health of the economy. .662 the stock market has little relevance to real economic activity.* �.603 there are enough good quality investment opportunities in the stock market. .543 wealth creating capacity .806 the greater financial risk is being out of the stock market rather than being in it. .697 the benefits of investing in the stock market outweigh the costs. .659 investing in the stock market is one of the safest investments an investor can make. .637 the stock market is sound. .566 the odds are in favor of the individual investor making money in the stock market. .565 if one is serious about building wealth, the stock market as an investment vehicle cannot be ignored. .442 fast money .557 the key to successful stock market investing is hot tips. .661 investing in the stock market is a way to make money easily and quickly. .651 if you are smart, it is easy to pick individual stocks that will have better than average returns. .638 tilted playing field .568 the stock market is controlled by large (institutional) investors. .660 only highly skilled investors can consistently make money in the stock market. .562 it is difficult for small investors to make money in the stock market. .422 * indicates items that are reverse coded. 23d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 appendix 2 description of variables disposition to trust (adapted from naef and schupp, 2009) please indicate how much you agree or disagree with each of the following statements: 1. in general, people can’t be trusted. 2. when dealing with strangers, it is better to be cautious before trusting them. 3. nowadays you can’t rely on anybody. [1 � strongly disagree, 2 � disagree, 3 � somewhat disagree, 4 � neither agree nor disagree, 5 � somewhat agree, 6 � agree, 7 � strongly agree] investor sociability (adapted from hong, kubik, and stern, 2004) during the past 12 months how often have you have engaged in: 1. giving or attending a dinner party 2. entertaining people in your home 3. visiting with friends 4. attending a church or other house of worship 5. doing volunteer work 6. talking with or visiting your neighbors [1 � never, 2 � once a year or less, 3 � several times a year, 4 � once a month, 5 � several times a year, 6 � several times a week, 7 � almost every day] investing experience how many years have you been investing in the stock market, either directly or in a self-directed retirement plan? investing motives to what extent does each of the following describe your objectives for investing in the stock market? 1. wealth accumulation: acquiring a higher expected return than on a savings account 2. hobby and entertainment: interest in the stock market 3. saving for retirement: being able to stop working at an earlier age 4. speculation: trying to profit from short-term developments in the stock market [1 � not at all, 2 � slightly, 3 � somewhat, 4 � moderately, 5 � quite a bit, 6 � extremely] affective states when you are/were invested in the stock market, how do/did you typically feel about it? out of control 1 2 3 4 5 6 7 in control excited 1 2 3 4 5 6 7 bored panicky 1 2 3 4 5 6 7 at ease investor satisfaction how satisfied are you with your past returns from your stock market investments? [1 � very dissatisfied . . . 7 � very satisfied] portfolio diversification on average, how many different stocks do you typically hold in your portfolio? net wealth calculated as the difference between responses to the following questions: what is the estimated total value of your assets (please include the value of your house and other real estate holdings, investments in stocks, bonds, term deposits, gics, employer-sponsored pension plans and cash holdings in savings or checking accounts)? and what is the estimated total value of your debt (please include the value of your real estate mortgages, student, personal and payday loans, outstanding credit card balances, outstanding balances on lines of credit)? risk proneness how would you classify your willingness to take risks in financial decision? [1 � very risk averse . . . 7 � very risk seeking] level of optimism (adapted from scheier, carver, and bridges, 1994) (continued on next page) 24 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 appendix 2 (continued) please indicate how much you agree or disagree with each of the following statements: 1. i rarely count on good things happening to me. 2. i’m always optimistic about my future. 3. in uncertain times, i usually expect the best. 4. if something can go wrong for me, it will. 5. overall, i expect more good things to happen to me than bad things. 6. i hardly ever expect things to go my way. [1 � strongly disagree, 2 � disagree, 3 � somewhat disagree, 4 � neither agree nor disagree, 5 � somewhat agree, 6 � agree, 7 � strongly agree] information search which of the following sources of information influence your decisions about the stock market investments you make? (check all that apply) 1. information from the company 2. advice and forecasts from brokers 3. discussions with friends and family members 4. information from magazines and newspapers 5. information from television 6. online research 7. social networking sites 8. other (please specify) financial literacy (pleis, 2007) next, we are going to ask some questions which will help gauge your knowledge of financial matters. if you are not sure about the answer to a question, please do not guess; instead, select the “do not know” option. 1. all of the following are advantages to the purchase of common stock except: (a) it can provide a good income stream (b) there is good liquidity (c) there can be tax advantages for long-term purchases (d) the return is guaranteed by the issuing company (e) do not know 2. if you buy a company’s stock: (a) you own a part of the company (b) you have lent money to the company (c) you are liable for the company’s debts (d) the company will return your original investment to you with interest (e) do not know 3. to minimize investment risks, your best strategy is: (a) only invest in well-established companies that pay annual dividends (b) invest in bonds or real estate (c) keep your money in the bank or invested in a certificate of deposit (d) diversify your portfolio to include stocks, bonds, mutual funds, real estate, and cash (e) do not know 4. over the last 70 years, the type of investment that has earned the most money and the highest rate of return for investors has been: (a) stocks (b) corporate bonds (c) savings accounts (d) do not know 5. when a company wants to issue stock in itself, it issues an ipo. what is an ipo? (a) investment profit organization (b) a dividend (c) a meeting with the securities regulation commission (d) initial public offering (e) do not know 6. which of the following increases the value of your money in stocks? (a) increase in price per share (b) dividends (c) stock splits (d) all of the above (e) do not know 7. in terms of investing, bonds are: (a) loans that investors make to companies (b) investments that provide monthly income to investors (c) a business promise with your broker (d) the stock of a new company (e) do not know 8. a mutual fund is: (a) an investment instrument for married couples (b) any tax-exempt investment (c) a paycheck deduction that goes to your retirement savings (d) a pooled group of investment instruments (e) do not know 9. which of the following organizations insures you against losses in the stock market? (a) fdic (federal deposit insurance company (b) sec (securities and exchange commission) (c) nbc (national bank of canada) (d) none of the above (e) do not know 10. what is short selling? (a) buying shares in the morning and selling them before the close of business on the same day (b) borrowing shares, selling them, and buying them back for less money after the price has fallen (c) buying a small quantity of shares (d) continuously selling shares after holding them for short periods of time (e) do not know 25d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 references ajzen, i. and fishbein, m. (1980). understanding attitudes and predicting social behaviour. new jersey, nj; prentice hall. alba, j. w., & hutchinson, j. w. (1987). dimensions of consumer expertise. journal of consumer research, 13, 411–454. anderson, e. w., fornell, c., & lehmann, d. r. (1994). customer satisfaction, market share, and profitability: findings from sweden. the journal of marketing, 58, 53–66. andreassen, t. w., & lindestad, b. (1998). the effect of corporate image in the formation of customer loyalty. journal of service research, 1, 82–92. aspara, j., & tikkanen, h. (2008). significance of corporate brand for business-to-business companies. the marketing review, 8, 43–60. aspara, j., & tikkanen, h. (2008). interactions of individuals’ company-related attitudes and their buying of companies’ stocks and products. the journal of behavioral finance, 9, 85–94. aspara, j., & tikkanen, h. (2010). consumers’ stock preferences beyond expected financial returns: the influence of product and brand evaluations. international journal of bank marketing, 28, 193–221. aspara, j., & tikkanen, h. (2011). individuals’ affect-based motivations to invest in stocks: beyond expected financial returns and risks. journal of behavioral finance, 12, 78–79. baker, h. k., & nofsinger, j. r. (2002). psychological biases of investors. financial services review, 11, 97–116. barber, b. m., & odean, t. (2001). boys will be boys: gender, overconfidence, and common stock investment. quarterly journal of economics, 116, 261–292. bernerth, j. b., & walker, h. j. (2009). propensity to trust and the impact on social exchange: an empirical investigation. journal of leadership and organizational studies, 15, 217–266. bloemer, j., & de ruyter, k. (1998). on the relationship between store image, store satisfaction and store loyalty. european journal of marketing, 32, 499–513. brakus, j. j., schmitt, b. h., & zarantonello, l. (2009). brand experience: what is it? how is it measured? does it affect loyalty? journal of marketing, 73, 52–68. bravo, r., montaner, t., & pina, j. m. (2012). corporate brand image of financial institutions: a consumer approach. journal of product & brand management, 21, 232–245. brown, j. m. (2012). gen y: post traumatic stock syndrome (available at http://www.thereformedbroker.com/ 2012/01/10/gen-y-post-traumatic-stock-syndrome/). burmann, c., schaefer, k., & maloney, p. (2008). industry image: its impact on the brand image of potential employees. journal of brand management, 15, 157–176. byon, k. k., & zhang, j. j. (2010). development of a scale measuring destination image. marketing intelligence & planning, 28, 508–532. cavezzali, e., & rigoni, u. (2012). know your client! investor profile and tailor-made asset allocation recommendations. journal of financial research, 48, 32–37. cho, j., & lee, j. (2006). an integrated model of risk and risk-reducing strategies. journal of business research, 59, 112–120. clark, g., thrift, n., & tickell, a. (2004). performing finance: the industry, the media and its image. review of international political economy, 11, 289–310. da, z., engelberg, j., & gao, p.(2011). in search of attention. the journal of finance, 66, 1461–1499. de bondt, w. f. m. (1998). a portrait of the individual investor. european economic review, 42, 831–844. dichter, e. (1985). what’s in an image? the journal of consumer marketing, 2, 75–81. dobni, d. m., & d. racine, m. (2015). stock market image: the good, the bad, and the ugly. journal of behavioral finance, 16, 130–139. dorn, d., & sengmueller, p. (2009). trading as entertainment? management science, 55, 591–603. ford, m. w., devoto, s., & kent, d.w. (2007). threat, intimidation, and student financial knowledge: an empirical study. journal of education for business, 82, 131–139. 26 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 ford, m. w., & kent, d. w. (2010). gender differences in student financial market attitudes and awareness: an exploratory study. journal of education for business, 85, 7–12. ganzach, y. (2000). judging risk and return of financial assets. organizational behavior and human decision processes, 83, 353–370. guiso, l., sapienza, p., & zingales, l. (2008). trusting the stock market. the journal of finance, 63, 2557–2599. gunn, c. a. (1988). vacationscape: designing tourist regions. new york, ny: van nostrand reinhold. hira, f., & loibl, c. (2008). gender differences in investment behavior. in j. j. xiao (ed.), handbook of consumer finance research (pp. 253–270). new york, ny: springer. hong, h., kubik, j. d., & steen, j. c. (2004). social interaction and stock market participation. the journal of finance, 59, 137–163. ittelson, w. h. (1974). an introduction to environmental psychology. new york, ny: holt, rinehart & winston. jang, s., & namkung, y. (2009). perceived quality, emotions, and behavioral intentions: application of an extended mehrabian–russell model to restaurants. journal of business research, 62, 451–460. jordan, j., & kaas, k. p. (2002). advertising in the mutual fund business: the role of judgmental heuristics in private investors’ evaluation of risk and return. journal of financial services marketing, 7, 129–140. keller, k. (2003). strategic brand management. englewood cliffs, nj: prentice hall. keller, c., & siegrist, m. (2006). money attitude typology and stock investment. the journal of behavioral finance, 7, 88–96. kennedy, p. e. k. (2008). a guide to econometrics (6th ed.). malden, ma: blackwell publishers. laros, f. j., & steenkamp, j. b. e. (2005). emotions in consumer behavior: a hierarchical approach. journal of business research, 58, 1437–1445. lucey, b.m., & dowling, m. (2005). the role of feelings in investor-decision making. journal of economic surveys, 19, 211–237. luo, x. (2007). consumer negative voice and firm-idiosyncratic stock returns. journal of marketing, 71, 75–88. luo, x. (2009). quantifying the long-term impact of negative word of mouth on cash flows and stock prices. marketing science, 28, 148–165. macgregor, d. g., slovic, p., dreman, d., & berry, m. (2000). imagery, affect and financial judgment. the journal of psychology and financial markets, 1, 104–110. macgregor, d. g. (2002). imagery and financial judgment. the journal of psychology and financial markets, 3, 15–22. mageau, g. a., vallerand, r. j., rousseau, f. l., ratelle, c. f., & provencher, j. (2005). passion and gambling: investigating the divergent affective and cognitive consequences of gambling. journal of applied social psychology, 35, 100–118. maheswaran, d. (1994). country of origin as a stereotype: effects of consumer expertise and attribute strength on product evaluations. the journal of consumer research, 21, 354–365. malmendier, u., & nagel, s. (2011). depression babies: do macroeconomic experiences affect risk taking? quarterly journal of economics, 126, 373–416. martenson, r. (2007). corporate brand image, satisfaction and store loyalty: a study of the store as a brand, store brands and manufacturer brands. international journal of retail & distribution management, 35, 544–555. mcknight, d. h., cummings, l. l., & chervany, n. l. (1998). initial trust formation in new organizational relationships. academy of management review, 23, 473–490. naef, m., & schupp, j. (2009). measuring trust: experiments and surveys in contrast and combination. diw berlin: sop papers multidisciplinary panel data research. olsen, r. a. (2008). trust as risk and the foundation of investment value. journal of socio-economics, 37, 2189–2200. o’shaughnessy, j. (1987). why people buy. new york, ny: oxford university press. ott, j. a. (2004). the “free and open” “people’s market”: public relations at the new york stock exchange, 1913–1929. business and economic history, 2, 1–43. philips, m. (2012). meet china’s master of the markets. bloomberg businessweek, 2012, 43–44. pleis, l. m. (2007). investment decisions: influence of an internet stock message board. ann arbor, mi: proquest. 27d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 richins, m. l. (1997). measuring emotions in the consumption experience. journal of consumer research, 24, 127–146. riley, w. b., jr., & chow, k. v. (1992). asset allocation and individual risk aversion. financial analysts journal, 48, 32–37. ryu, k., han, h., & kim, t. h. (2008). the relationships among overall quick-casual restaurant image, perceived value, customer satisfaction, and behavioral intentions. international journal of hospitality management, 27, 459–469. scheier, m. f., carver, c. s., & bridges, m. w. (1994). distinguishing optimism from neuroticism (and trait anxiety, self-mastery, and self-esteem): a reevaluation of the life orientation test. journal of personality and social psychology, 67, 1063–1078. snyder, m., tanke, e. d., & berscheid, e. (1977). social perception and interpersonal behavior: on the self-fulfilling nature of social stereotypes. journal of personality and social psychology, 35, 656–666. statman, m., fisher, k. l., & anginer, d. (2008). affect in a behavioral asset-pricing model. financial analysts journal, 64, 20–29. stern, b., zinkhan, g., & jaju, a. (2001). marketing images: construct development, measurement issues, and theory development. marketing theory, 1, 201–224. strader, t. j., & ramaswani, s. w. (2004). investor perceptions of traditional and online channels. communications of the acm, 47, 73–76. traflet, j. (2004). spreading the ideal of mass shareownership: public relations and the nyse. essays in economic and business history, 22, 257–273. tversky, a., & kahneman, d. (1974). judgment under uncertainty: heuristics and biases. science, 185, 1124–1131. unerman, j., & o’dwyer, b. (2004). enron, worldcom, andersen, et al.: a challenge to modernity. critical perspectives on accounting, 15, 971–993. van raaij, w. f. (1991). the formation and use of expectations in consumer decision making. in t. s. robertson & h. h. kassarjian (eds.), handbook of consumer behavior (pp.401–418). englewood cliffs, nj: prentice-hall. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. vissing-jorgensen, a. (2003). perspectives on behavioral finance: does ‘irrationality’ disappear with wealth? evidence from expectations and action. in m. gertler & k. rogoff (eds.), nber macroeconomics annual. cambridge, ma: mit press. 28 d.m. dobni, m.d. racine / financial services review 25 (2016) 1–28 investment performance of aaii stock screens over diverse markets david s. northa, jerry l. stevensa,* adepartment of finance, e. c. robins school of business, university of richmond, richmond, va 23173, usa abstract individual investors often rely on information services and products to compete with professional investors. to assist in this effort, the american association of individual investors (aaii) offers a variety of screening tools designed to help individuals construct stock portfolios. we extend prior research on aaii screening performance by including more recent investment periods, using more rigorous factor models, and examining both median and mean returns to allow for skewing. over a period of tumultuous markets with as little as $50,000 to invest, over 30% of the available screens achieved statistically significant excess rates of return unrelated to transaction costs and multifactor risks proposed by efficient market theorists. © 2015 academy of financial services. all rights reserved. jel classification: g11; g14 keywords:personal investing; investment performance; factor models 1. introduction as part of an investment process, investors consider the extent to which they use passive and active approaches to the equity market. individuals often use a passive approach by investing in broadly diversified index funds rather than spend time and energy trying to pick stocks. the passive approach is supported by efficient market arguments that active trading of stocks will not consistently beat the market index on a risk-adjusted basis. the efficient * corresponding author. tel.: �1-804-289-8597; fax: �1-804-289-8878. e-mail address: jstevens@richmond.edu (j.l. stevens). financial services review 24 (2015) 157–176 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. market hypothesis is based on the view that a large numbers of rational investors with access to the same information will generate security prices that are unbiased estimates of security values. as a result, there should be no consistent abnormal returns from active trading.1 active investors seek opportunities for excess rates of return that are generally linked to behavioral explanations for departures from investor rationality. behavioral finance theories maintain that predictable irrationality in security pricing, linked to human biases and heuristics, causes predictable deviations from efficient market pricing. these predictably irrational behaviors of investors have been tested and validated in behavioral finance studies reviewed by nofsinger (2013) and have been translated into investment strategies by pompian (2012). ultimately, the ability to use information and metrics about stocks to earn an excess return is an empirical issue. even with market inefficiencies, small investors may remain passive because they feel that professionals with large research budgets have size, diversification, technology, transaction cost, and information advantages. empirical evidence suggests that individual investors generally do not make good stock selection decisions. barber and odean (2000) found that returns on stocks sold by individuals subsequently turned out to be higher than returns on stocks they bought. in the same study barber and odean found that the average household with an account at a large discount brokerage firm underperformed by an average 15 basis points per month even without transaction cost adjustments. korniotis and kumara (2013) analyzed investment accounts of over 60,000 customers of a major discount broker over the period from 1991 to 1996. they found that retail investors (less informed) underperformed wholesale investors (well informed) by over 200 basis points per year on average. a wide range of professional services are available to help individuals overcome disadvantages relative to well informed investors. one such service available to individual investors is provided by the american association of individual investors (aaii). aaii is a nonprofit investment education organization offering individual investors a set of low cost services supporting active trading strategies. investors have access to stock screens based on what aaii believes well known and highly regarded investors use to identify attractive stocks. aaii claims that 91% of their screened portfolios beat a passive s&p 500 index but there is no information given on performance after transactions costs, risk adjustments, and tests of statistical significance.2 in this article we present a range of performance measures addressing both practical and theoretical issues to test the investment performance of aaii strategies. specifically, prior tests of investment performance from using mechanical screens are extended as follows: 1. we extend aaii performance analysis in prior studies to more recent years that include the period since the 2008 market crash. our results help establish whether or not prior findings are robust over more recent holdout periods. 2. we introduce an analysis of both the mean and median monthly returns to identify potential skewing of returns. 3. we conduct a more rigorous analysis of screened portfolio performance by using factor models. in this way we test for screened portfolios that offered significant excess returns even after more extreme risk factor hypotheses are considered. 158 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 the article offers several layers of investment analysis that may also be instructional for individual investors who have only focused on cumulative wealth, average returns, or sharpe ratios to measure performance. 2. literature review we address the basic question of whether or not a mechanical screening process offered to individual investors by a third party can achieve statistically significant excess riskadjusted returns over a long investment horizon. the relevant background literature includes both studies testing whether information is efficiently priced in general and studies testing whether there is excess return potential from following professional recommendations in particular. even “well informed” professionals do not earn excess risk-adjusted returns net of transaction costs if security prices are efficiently valued, according to the efficient market hypothesis (emh). aaii stock screens would be of no help to individual investors if the stock market is efficient. the aaii screens combine types of weak-form and semi-strong form information to identify stocks. weak-form efficiency maintains that stock prices fully reflect all historical information. early tests of the weak-form hypothesis used correlations in short run holding period returns defined in days or weeks. fama (1970) and fama and macbeth (1973) found that stock prices are not highly correlated, supporting weak form efficiency. conrad and kaul (1988) and lo and mackinlay (1988) both analyzed weekly returns of the new york stock exchange (nyse) stocks and found significant positive serial correlations that were likely to be too small to allow trading opportunities. campbell, grossman, and wang (1993) found momentum in short horizon data for individual transactions of stocks but the trading turnover was likely to be too high to earn excess returns. jegadeesh (1990) constructed portfolios from one-step-ahead forecasts using a serial correlation model and found a difference in abnormal returns of 2.49% between the top and bottom portfolios. campbell, lo, and mackinlay (1997) found momentum in short run holding periods but only about 12% of the daily price movement could be linked to the prior day return. overall, momentum in stock returns tends to be found in short run holding periods but the potential improvement in returns from trading may be too small to exploit. jegadeesh and titman (1993) discovered momentum in stock returns for holding periods longer than days or weeks. they demonstrated that simple strategies based on stocks ranked by cumulative returns over the past 3 to 12 months predicted relative performance over the next 3 to 12 months. they also showed that their results could not be explained by risk factors. jegadeesh and titman broke ranks with prior researchers at the time by suggesting that weak-form imperfections were large enough to derive excess returns. in a follow up study, jagadeesh and titman (2001) used out-of-sample data to show that their findings continued to hold in the decade after their original study. for long holding period returns, debondt and thaler (1985, 1987) demonstrated mean reversion where winners became losers, suggesting inefficiencies from overreaction. shiller (1981) supported the overreaction hypothesis by finding excess volatility beyond what could be explained by material stock information. both metrick (1999) and hirchleifer (2001) 159d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 recognized that tension between rational (well informed) and irrational (uninformed) money movements result in long periods before full information “efficient” prices are in equilibrium, potentially allowing excess returns from trading counter to the market. alternative tests of weak-form efficiency used mechanical filter rules based on past data to construct portfolios. pinches (1970) found that portfolios built on past data patterns do not outperform buy-and-hold portfolios. on the other hand, brush (1986) found abnormal returns with rules from combinations of past data and pruitt and white (1988) found excess returns for rules based on past data adjusted for january effects. bessembinder and chan (1998) used trading rules based on past price movements and found that active trading did not beat passive investing when cost of active trading were considered. overall, the investment potential from screens based on historical data are not clearly proven or disproven. frequency of trading and trading costs play a large role in the performance of these portfolios. tests of the semi-strong version of the emh are based on performance of portfolios constructed from public information about stocks. in an efficient market, new information should be priced rapidly and there should be no excess return. ball and brown (1968) and brennan (1995) demonstrated that the market prices new accounting information efficiently if the information has a real impact on the stock price. other findings of efficient pricing in event studies occur in fama, fisher, jensen, and roll (1969) for stock splits, jensen and ruback (1983) for takeover information, mcconnell and muscarella (1985) for capital expenditure information, pierce and roley (1985) for stock market reaction to macroeconomic data announcements, and klein (1986) for divestiture information. in contrast, ball and brown (1968) found that earnings announcement information was not priced efficiently. both pre and post announcement drifts occurred for earnings surprises. ball (1978) attributed this drift to problems in measuring abnormal returns but rendleman, jones, and latane (1982) conclude that the price drift is because of behavioral under-reaction to new information. exceptions to efficient pricing of public information can be found in a long list of studies. basu (1977) found excess risk-adjusted returns for portfolios of low price-to-earnings stocks. peavy and goodman (1983) confirmed basu’s findings after controlling for size, liquidity, and industry effects. campbell and shiller (1988) found that low price-to-earnings ratios can predict future abnormal returns. dreman and berry (1995) demonstrated that excess returns from low price-to-earnings stock portfolios were not because of inadequate adjustment for risk. banz (1981) demonstrated a size effect where small stocks earned excess returns whereas arbel and strebel (1983) reached a similar conclusion for firms receiving less research attention. peters (1991) looked at the joint effect of low price-to-earnings and growth and found excess returns for firms where the peg (price-to-earnings ratio divided by growth) was low. rosenberg, reid, and lanstein (1985) found that low market to book stocks tend to earn excess rates of return. fama and french (1996) confirm the joint effects of market risk, size, price-to-earnings, market-to-book, and leverage on portfolio returns. however, they choose to define these effects as risk factors rather than inefficiencies. the fama and french three factor model emerges from their study. it is impossible to tell whether the added factors are risk premiums or unpriced sources of information. studies of investment performance from following recommendations of professional investors have had mixed results. oppenheimer (1981) back-tested portfolios constructed 160 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 from suggestions of benjamin graham in periodic editions of “intelligent investor.” the constructed portfolios beat the relevant benchmarks beyond what graham predicted. in more recent studies, palman, sun, and tang (1994) found abnormal returns in an event study based on stock recommendations from business week analysts. anderson and loviscek (2005) found that portfolios built from the top five picks in the book the best stocks to own in america outperformed the market. desai and jain (1995) tested the performance of stocks recommended from a roundtable of “superstar” analysts published in barrons. abnormal returns were found from the date of the roundtable to the publication date but there were no abnormal returns once the recommendations were made public. metrick (1999) tested recommendations from 153 different investment newsletters and found that neither the group of newsletters nor the 153 individual recommendations generated abnormal returns. several studies tested the investment performance from following recommendations of analysts appearing on television. pari (1987); griffin, james, and zmijewski (1995); and beltz and jennings (1997) found abnormal returns for the first trading day after “wall street week with louis rukherser” aired. however, pari (1987) and beltz and jennings (1997) found negative abnormal returns measured over the longer term whereas griffin, james, and zmijewski (1995) found significant positive abnormal returns over the year following the wall street week recommendations. ferreira and smith (2003) found significant positive abnormal returns for wall street week recommendations in both short run and long run periods when abnormal return measures were adjusted. bolsher, trahan, and venkateswarren (2012) followed the suggestions on kramer’s “mad man” and found positive gains for the first day after the show aired but losses occurred for the following 29 days. tests of public recommendations from professionals offer mixed results and it is difficult to sort unpriced information from the behavioral biases and emotions of analysts making recommendations. as an alternative, several studies test the performance from more mechanical professional recommendations. olson, nelson, witt, and mossman (1998) found significantly positive abnormal returns from trading based on investor’s business daily proprietary stock rankings. choi (2000) found positive abnormal returns from investing in stocks with the highest value line timeliness rankings even after controlling for size, momentum, and earnings surprises. nevertheless, choi’s findings were not likely to overcome transaction costs of trading. loviscek and jordan (2000) created portfolios from the top holdings of five-star mutual funds and found that the performance was too weak to recommend. kacperczyk, van nieuwerburgh, and veldkamp (2013) applied a measure of skill to 3,477 mutual funds over the period 1980 through 2005. they found that the top quintile of funds outperform by 300 to 600 basis points per year. overall, there is no clear consensus from these studies and the possibility remains that some sources of information may offer excess returns whereas others do not. schadler and cotton (henceforth noted as s&c) provided the most relevant study for the work we present here. they tested the investment performance of portfolios constructed with the aaii screens from january of 1998 to december 2005. after accounting for transaction costs, statistical significance, and the appropriate “market” index for comparison, they conclude that about 20% of the constructed portfolios and 25% of the low transaction cost portfolios beat the best fit indexes. we extend the s&c study by considering a longer investment period, measuring performance with statistical significance of excess returns 161d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 from multifactor models, and considering favorable skewing of returns by using both mean and median return measures. our approach offers a rigorous test of market efficiency in general and the performance of the aaii screened portfolios in particular. 3. aaii data and services we first follow the basic approach used by s&c as a starting point and then extend the analysis. we use both the traditional single-factor market model and the fama and french three-factor model to measure and test excess risk-adjusted returns (�s) after transaction costs. this approach offers more extensive tests of investment performance beyond the sharpe ratio comparisons and statistical tests of difference in returns. our study spans the period from the initiation of the aaii screens in january 1998 through december of 2011, whereas s&c ended their study in december of 2005. the extended period of analysis includes both investment performance leading up to the great recession and several years of the ongoing sluggish recovery period. these added observations help illustrate how the aaii portfolio screens tend to perform over rapidly changing markets. the investment service provided by aaii allows investors to follow stock screens constructed with what aaii believes specific well known investors use to select stocks. aaii reports monthly returns on a total of 82 portfolios beginning with january 1998. the aaii screens are divided into the following investment styles: passive indexes, growth, value/ growth, specialty/sector, and unclassified. the screen characteristics used by aaii to categorize a screen into a style represent potential market inefficiencies that might identify stocks with excess return potential. of the 82 portfolios, 16 are stock market indices and one is the treasury bill index. of the remaining 65 portfolios, seven do not have complete data, leaving 56 screens for our testing purposes. table 1 provides a listing of the screened portfolios and corresponding investment styles used in this study along with an annotation of how our list of screens is different from the list used by s&c in their study.3 the aaii stock investor pro database is the source of data used to run through the aaii screens and create portfolios. the database includes around 8,500 actively traded stocks and adrs from the u.s. markets. the portfolios are the result of a mechanical adherence to an investment style without emotions entering into the decisions. aaii sends cds to members each month and current financial data on firms in the database may be downloaded each week from the aaii website. the data do not include dividend distributions. the weekly updates are from reuters as of the end of the business day on friday and are available to investors on monday of the following week. at the end of each month, aaii runs the investment screens and then constructs an equally weighted portfolio of stocks passing the screen. the holding period return is calculated from the change between the price of each stock on the last trading day of the month to the price on the last trading day of the following month. aaiis rebalancing approach sells all securities at the end of the month and buys all securities passing the screen for the next month. the turnover approach assures that each portfolio is equally weighted. this 100% turnover of the portfolio each month is used by s&c and in our extension. the equal weighting assumption clearly results in high turnover and transaction costs. an individual investor would incur lower costs with less frequent 162 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 rebalancing or with an alternative weighting scheme. a key point in the analysis is that we use the same approach to transaction costs as s&c for comparison purposes. 4. theoretical and practical issues for performance tests to test claims made by aaii about the value of their screening services, s&c use the same portfolio performance measures that aaii provides. the measures include comparitable 1 names and style classification of the 56 aaii portfolio screens in this study value screensa growth/value screens cash rich firms buffett-hangstrom dividend (high relative yield) buffettology-eps growth dogs of the dow buffettology-sustainable growth dogs of the dow low priced 5 fisher (philip dreman lynch dividend screen–drps muhlenkamp dividend screen–non-drps o’shaughnessy-growth price-to-free-cash-flow oshaughnessy small cap gr.& value price changeb fundamental rule of thumb oberweis octagon graham defensive investor non-utility value on the move-peg with est growth graham enterprising investor ibd stable 70 price to sales lakonishok t. rowe price neff templeton o’shaughnessy-value wanger (revised) p/e rrelative stock market winners piotroski zweig weiss blue chip dividend yield magnet simple price changeb speciality/sector piotroski high f-score price changeb adrs schloss price changeb dual cash flow magic formula price changeb estimated revisions down rule #1 investing price changeb estimated revisions up oshaughnessy tiny titans price changeb estimated revisions up 5% graham defensive investor (utility) murphy technology growth screens o’neil’s canslim screens used in the s&c study but not this studyc o’neils’s canslim revised 3rd ed. dreman revised (no longer in aaii data) foolish small cap 8 all drps (not classified) foolish small cap 8 revised estimated revisions down lowest 30 (not classified) ibd stable 70 estimated revisions up top 30 (not classified) inve$tware quality growth low price to book ratio (no longer in aaii data) return on equity dreman with est. revisions (not classified) magnet complex price changeb estimated revisions down (no longer in aaii data) oshaughnessy growth mkt leaders price changeb a screen names and classifications are determined by aaii. b there are nine portfolio screens in this study that were not available in the c&s study. these portfolio screen names appear in italics. aaii added complete data for these portfolios after the s&c study was conducted. c there are seven portfolio screens in the c&s study that are not used in this study. we deleted a screen if it had complete data over the entire study period or if the screen style could not be classified. 163d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 sons with the s&p index returns, comparisons with the best-fit index returns, and sharpe ratios. these comparisons are made with and without transaction costs. tests for statistical differences in average portfolio returns and the benchmark indexes are also conducted by s&c. we expand the analysis of performance measures to also include a comparison of median returns of portfolios with their benchmarks to account for skewing, a single factor � with significance tests, and a three factor � with significance tests for the screened portfolios. following s&c, we also look at the measures with and without transaction costs. 4.1. index comparisons aaii provides 16 different equity indexes in its database. following s&c we identify a best-fit index for each of the aaii screened portfolios. the best-fit index for a given aaii portfolio is defined as the index with the highest correlation of monthly returns with the aaii portfolio. for each aaii portfolio we use the relevant best-fit index as the benchmark for performance measurement. 4.2. geometric mean monthly return long run buy-and-hold investors who remain fully invested over the measurement period may be satisfied by beating a passive market index without risk adjustment. risk adjustment of returns is not part of the analysis in this case since the investor will ride out the ups and downs of the market with no intention of buying or selling in response to market moves. a typical first step to analyzing performance is to simply measure the extent to which a strategy beats a benchmark over a measurement period. s&c take this approach by measuring the differences in cumulative total rates of return of aaii strategies and the relevant benchmarks. the cumulative total rate of return comparison is equivalent to measuring differences between the geometric mean monthly return (gmmr) of aaii screens and the relevant benchmark. 4.3. mean monthly return the arithmetic mean monthly return (ammr) is an alternative to the gmmr measure. while the geometric mean reflects the compounded return outcome from an investment, the arithmetic mean is the expected value of a random return series. when there is skewing of returns from outliers the median monthly return (mdmr) is a better measure of the “typical” portfolio performance than the ammr. tests for statistical significance of the difference between the portfolio mean (or median) and the benchmark offer additional insight about performance, because a portfolio’s performance may beat a benchmark because of random chance. 4.4. the sharpe reward for risk ratio the sharpe index is a basic reward-for-risk measure used to evaluate portfolio performance. we use the sharpe ratio defined as the excess return (rp � rf) divided by the portfolio standard deviation of returns (�p) over the measurement period. 164 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 sharpe ratio � (rp � rf)/��p) (1) the sharpe index in eq. (1) measures portfolio efficiency in the sense that a higher sharpe index reflects a better tradeoff between portfolio returns and total portfolio risk. one limitation of the sharpe ratio is that there are no tests of statistically significant differences between sharpe ratios. 4.5. alpha measures of portfolio performance and factor models in theory, good performance is measured as a statistically significant excess risk-adjusted return commonly called the portfolio’s �. to measure the portfolio � there must first be a model that specifies the risk factors and risk premiums leading to an expected return (rp). the single-factor model introduced by sharpe (1963) is consistent with the capital asset pricing model where the portfolio is exposed to two sources of uncertainty known as systematic risk and unsystematic risk. the portfolio’s systematic risk measured by � is because of portfolio exposure to movements of the overall market that affect all portfolios to some extent. the equity risk premium is the reward for taking systematic risk and is measured by the difference between the broad market index (rm) and the risk free rate (rf). unsystematic risk (�) is the second source of uncertainty for a portfolio and represents a type of residual risk that can be eliminated with diversification. the single-factor model is specified below: (rp � rf)t � �1 � �(rm � rf)t � �t (2) alpha represents an excess risk-adjusted return and is expected to be zero if the market prices risk efficiently. eq. (2) is an empirical model allowing estimation of a portfolio’s � (intercept) and � (slope) by regressing (rp – rf) on (rm – rf) using the monthly data for the screened aaii portfolio returns, risk free rates, and market index returns. the intercept is often called jensen’s � in this formulation and the t-statistic, calculated as the estimated intercept coefficient divided by the standard error, provide tests for statistical significance of �. statistically significant positive �s not only measure superior portfolio performance for comparison purposes but they also allow for tests of efficiency in market pricing. we would not expect to find more portfolios with statistically significant positive �s than would be generated by chance. this approach is how most of the “anomalies” to efficient market pricing have been identified. fama and french (1996) argue that statistically significant positive �s from the singlefactor model of eq. (2) are likely to be the result of left-out risk factors rather than excess returns. the fama-french three-factor model augments the single-factor model with size (smb) and value (hml) factor variables. the � in the three-factor model (�3) represents the difference between the portfolio’s return and the expected portfolio return based on the portfolio’s sensitivities to the three factors. alpha is expected to be zero if markets are pricing securities efficiently. the three-factor model is specified below: (rp � rf)t � �3 � �1(rm � rf)t � �2 smbt � �3 hmlt � �t (3) 165d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 eq. (3) is simply eq. (2) with the two added factors smb and hml and their �s. smb is measured as the premium in period t for small cap minus large cap stocks and hml is the premium for high book/price stocks minus low book/price stocks at time t. if both �2 and �3 are not statistically different from zero we have the single-factor model of eq. (2). overall, the three-factor model provides a better fit with stock returns than the singlefactor model (higher adjusted r2). we cannot say whether the additional returns due to statistically significant and positive values for �2 and �3 are the result of added risk premiums or are returns to inefficiencies in pricing size and value. we can say that the three factor model represents a more stringent test of portfolio performance. if the � in the three-factor model is statistically significant and positive the portfolio earned excess returns beyond what would be expected from exploiting size and valuation characteristics of stocks. the importance of including the three-factor model � in an analysis of investment performance is supported by bodie, kane, and marcus (2013, p 0.605) as follows: the fama-french (ff) three-factor model…has almost completely replaced the single-index model in academic performance evaluation, and has been gaining ‘market share’ in the investment services industry. 5. transaction costs for the aaii portfolios the aaii screens are performed each month and stocks that pass the screen are used by aaii to create equally weighted portfolios. transaction costs are computed using the same online brokerage fee of $7 per trade used by s&c. we do not have data on the number of trades each month so we follow s& c by using the average number of stocks in the portfolio as a proxy. with these assumptions, the aaii rebalancing approach results in a monthly roundtrip transaction cost calculated as follows: ($7) (2) (average number of stocks in the portfolio) � $ transaction cost. (4) given the flat cost per trade, the transaction costs is higher for those screens that result in a larger number of stocks in a portfolio. the amount of the initial investment is also a relevant factor since the fixed dollar transaction costs is a higher percentage of a smaller total investment amount. this is an important consideration because the aaii screens are designed for individual investors who may not have large amounts to invest. for the january 1998 through december 2011 investment period, table 2 shows the table 2 relationship between initial investments and the mean monthly average returns after transaction costs for aaii portfolios over the investment horizon from january 1998 through december 2011a initial investment $10k $20k $30k $40k $50k $60k $70k $80k $90k $100k number of funds with positive average returns 8 21 31 42 49 52 53 54 55 55 % of funds with positive average returns 14% 38% 55% 75% 88% 93% 95% 96% 98% 98% a the analysis used all 56 aaii funds. transaction cost calculation assumes 100% turnover each month and $7 per trade. the dollar cost each month is $7 (2) (no. stocks in the portfolio). 166 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 number and percentage of portfolios with positive average monthly rates of return after transaction costs. for example, less than 38% of the 56 aaii portfolios in the study generated positive after-transaction cost returns if an investor put less than $20,000 into the strategy. as the amount invested goes up to $100,000 or more virtually all the aaii screens generate positive after-transaction cost monthly returns. in the performance evaluation of the aaii screens that follows we consider $50,000, $100,000, and zero initial investments with corresponding transaction cost scenarios. 6. overall analysis of aaii portfolio screens we analyze performance of the 56 aaii portfolios over the period from january 1998 through december 2011 by style, by transaction cost, and by different measures of performance. the total output from this analysis is too extensive to put in print but is available upon request from the authors. table 3 provides a summary of the key findings. for each performance measure we compare findings with zero transaction costs to findings with after-cost returns for initial investments of $50,000 and $100,000. the columns in table 3 represent six different performance measures for the aaii portfolios. the first four columns are basic performance measures much like those used by aaii and s&c plus a median monthly return (mdmr). the last two columns represent findings for single-factor (1fmm) �s and three-factor (3fmm) �s, respectively. for gmmr relative to the best-fit index (bfi) without transaction costs we find that 79% of the portfolios beat the bfi returns with an average spread over the benchmark of 38 basis points per month. s&c had almost identical findings with 79.6% of the aaii screened portfolios beating the bfi on a raw cumulative return basis without transaction costs. value (v) screens offered the best performance with 83% of the value portfolios beating the bfi with an average gmmr difference of 46 basis points. gmmr performance is much worse when transaction costs are introduced. when the initial investment is $100,000 only 61% of the portfolios beat the benchmark with an average gmmr margin of only 17 basis points. for an investor with $50,000 to invest only 38% of the portfolios beat the gmmr benchmark and the overall average spread is a negative eight basis points. the value style offers the best performance when investors have $100,000 invested but the growth style is the best style for the smaller $50,000 investment. a plausible explanation for this finding is linked to the difference in the number of stocks in growth and value portfolios. of our 56 portfolios, growth strategy portfolios have a lower number of stocks to trade on average (16.6) than any other portfolio style. however, as the amount invested increases transaction costs as a percentage of the investment decrease to eliminate the advantage of trading small numbers of stocks in growth strategies. the ammr without transaction costs for 86% of the portfolios beat the bfi by an average of 51 basis points per month. however, only 39% of the portfolios are statistically significantly higher than the best fit benchmark. when transaction costs are introduced statistically significant higher ammrs fall to 21% for $50,000 and 27% for $100,000 investments. for the statistically significant higher ammrs the best strategy is v. when we use median returns rather than mean returns for the aaii portfolios, perfor167d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 t ab le 3 su m m ar y pe rf or m an ce of a a ii po rt fo lio s fr om ja nu ar y 19 98 th ro ug h d ec em be r 20 11 by pe rf or m an ce m ea su re , st yl e, an d tr an sa ct io n co st lin ke d to si ze a (1 ) g m m r � b fi (2 ) a m m r � b fi (3 ) m d m r � b fi (4 ) sh ar pe ra tio s re la tiv e to b fi sh ar pe ra tio (5 ) 1 fm m � (6 ) 3 fm m � in iti al in ve st m en t (0 00 s) n a 50 10 0 n a 50 10 0 n a 50 10 0 n a 50 10 0 n a 50 10 0 n a 50 10 0 t ra ns ac tio n co st s in cl ud ed n o y es y es n o y es y es n o y es y es n o y es y es n o y es y es n o y es y es pa ne l a . b fi co m pa ri so ns % al l fu nd s � b fi 79 38 61 86 45 63 41 16 32 82 38 52 b p di ff er en ce (a ll fu nd s) 38 � 8 17 51 6 30 0 � 49 � 21 b es t st yl e* v g r v v g r v v g r v g r v v v v % of fu nd s fo r be st st yl e 83 44 71 92 56 75 50 24 41 92 46 67 b es t st yl e b p di ff er en ce 46 7 29 63 2 43 3 � 52 � 15 pa ne l b . si gn ifi ca nt d if fe re nc e % fu nd s st at . si gn .* n a n a n a 39 21 27 29 13 16 n a n a n a b es t st yl e v v v g r v g r v g r v n a n a n a pa ne l c . sh ar pe ra tio s sh ar pe di ff er en ce .0 49 � .0 21 .0 17 b es t st yl e sh ar pe d if fe re nc e .0 6 � .0 03 .0 29 pa ne l d . fa ct or m od el s % of fu nd s si g. � � 80 32 59 61 29 39 a ve . si gn ifi ca nt � b p 11 9 12 1 14 7 16 3 11 5 12 6 b es t st yl e v g r g r v g r g r g r % si g. by be st st yl e 88 44 65 78 44 56 a v ar ia bl es ar e de fin ed as fo llo w s: g m m r is ge om et ri c m ea n m on th ly re tu rn ; b fi is th e be st fit in de x fo r th e gi ve n po rt fo lio ; a m m r is th e ar ith m et ic m ea n m on th ly re tu rn ; b fi sh ar pe ra tio is th e sh ar pe ra tio co m pu te d w ith th e b fi re tu rn as th e po rt fo lio re tu rn ; 1f m m � is th e on e fa ct or m ar ke t m od el � ;3 fm m � is th e th re e fa ct or m ar ke tm od el � .n a is no ta pp lic ab le an d b p re pr es en ts ba si s po in ts .t he as te ri sk (* ) de no te s a 5% le ve lo f si gn ifi ca nc e us in g a on e ta il tdi st ri bu tio n. w e w ou ld ex pe ct 5% of th e fu nd s to be si gn ifi ca nt be ca us e of pu re ch an ce . t he fa ct or m od el s fo r al l po rt fo lio s us e th e s& p 50 0 in de x as th e m ar ke t be nc hm ar k. 168 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 mance is much worse because positive skewing leads to a median lower than the mean. for no transaction costs, only 41% of aaii portfolios had mdmr higher than the bfi and only 29% had a statistically significant higher mdmr than the bfi. only 16% of the portfolios had a mdmr higher than the bfi for a $50,000 initial investment. the growth/value style had the highest percentage of portfolios with mdmr statistically significantly higher than the bfi. once transaction costs are introduced the basis point spread between the mdmr for aaii portfolios and the bfi are negative overall by 49 basis points for a $50,000 investment and 21 basis points for a $100,000 investment. even for the best style (growth/value) the basis point difference between the mdmr and bfi is negative. with positive skewing of returns, an investor must stick with the strategy or risk missing out on a positive “outlier” return. portfolio strategies that look good on a mean return comparison but not a median return comparison require investors to stay with the strategy over a long period of time. without transaction costs, 82% of the portfolios in table 3 have a sharpe index higher than the bfi sharpe index.4 value portfolios offer the best overall performance with 92% of the funds beating the bfi sharpe ratio. however, differences in sharpe ratios tend to be small and there are no tests for statically significant differences. for smaller investors with an initial investment of $50,000, only 38% of the aaii portfolios beat their benchmark sharpe ratios. value style portfolios tend to have more favorable sharpe ratio comparisons, but the spread is again negative for the $50,000 investment. for investors with $100,000 to invest the sharpe ratio comparisons are more favorable but barely half (52%) of the aaii portfolios beat their bfi sharpe ratios. the value strategy is the best style across all levels of investing based on sharpe ratio comparisons. the results from regression analysis measuring the 1fmm �1 and the fama and french 3fmm �3 appear under headings (5) and (6) in table 3.5 the 1fmm �s are statistically significant and positive for 80% of the aaii portfolios without transaction costs. with transaction costs, statistically significant 1fmm �s occur for 32% of the portfolios for investors of $50,000 and 59% for investors of $100,000. since the one-tail 5% level of significance is used in the tests, we would expect 5% of the portfolios to be significant by pure chance. even with transaction costs, the aaii screens offer more significant excess return portfolios than chance would predict. the best style varies with the size of the investment. value dominates the zero transaction cost comparisons whereas growth dominates for the $50,000 initial investment and a growth/value blend performs best for the $100,000 initial investment. the magnitude of the statistically significant �s is very impressive with over 100 basis point excess return for every initial investment in table 3. when transaction costs are zero, the 3fmm � is statistically significant at the 5% level for 61% of the aaii portfolios with zero transaction costs. with transaction costs, the proportion of aaii portfolios with statistically significant 3fmm �s shrinks to 29% for a $50,000 initial investment and to 39% for an initial investment of $100,000. statistically significant 3fmm �s average from 163 basis points for zero transaction costs to 115 basis points for the $50,000 investment. the growth style has the best performing portfolios for the 3fmm. the findings for the 3fmm are interesting because superior performance is found for a large number of the aaii portfolios even after taking out market exposure, size, and valuation factors that should reduce excess returns to zero according to fama and french. the aaii 169d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 t ab le 4 sc re en ed po rt fo lio ra nk in gs –w ith tr an sa ct io ns co st s an d $5 0k in iti al in ve st m en ta ,b po rt fo lio na m e g ro up g m m r � b fi a m m r � b fi m d m r � b fi sh ar pe b fi sh ar pe 1f m m � 3f m m � a ve ra ge st oc ks he ld pi ot ro sk i 9 pr ic e ch g v 1. 68 % (1 ) 2. 01 % * (2 ) � 1. 02 % (4 0) 0. 13 0 (1 ) 2. 71 % ** (1 ) 2. 39 % ** (1 ) 4 (5 3) m a g n e t si m pl e pr ic e ch g v 1. 29 % (2 ) 2. 06 % * (1 ) � 1. 03 % (4 1) 0. 09 7 (7 ) 2. 45 % ** (2 ) 1. 97 % * (2 ) 3 (5 5) o ’n ei l’ s c a n sl im pr ic e ch g g 1. 22 % (3 ) 1. 36 % * (3 ) � 0. 23 % (1 4) 0. 12 9 (2 ) 1. 94 % ** (3 ) 1. 78 % ** (3 ) 7 (5 0) o ’s ha ug hn es sy : t in y t ita ns pr ic e ch g v 1. 14 % (5 ) 1. 14 % ** (6 ) 0. 87 % * (3 ) 0. 09 4 (8 ) 1. 80 % ** (4 ) 1. 33 % ** (6 ) 25 (2 5) e st r ev : u p 5% pr ic e ch g s 1. 08 % (6 ) 1. 23 % ** (4 ) 1. 00 % ** (1 ) 0. 12 1 (3 ) 1. 77 % ** (5 ) 1. 60 % ** (4 ) 45 (9 ) pi ot ro sk i: h ig h fsc or e pr ic e ch g v 1. 20 % (4 ) 0. 98 % * (8 ) � 0. 25 % (1 6) 0. 07 0 (1 1) 1. 66 % ** (6 ) 1. 24 % * (8 ) 24 (2 9) m a g n e t c om pl ex pr ic e ch g g 0. 28 % (1 6) 0. 86 % (1 0) � 1. 28 % (4 4) � 0. 00 3 (2 2) 1. 66 % * (7 ) 1. 55 % * (5 ) 2 (5 6) z w ei g pr ic e ch g g /v 0. 76 % (8 ) 1. 17 % ** (5 ) 0. 97 % ** (2 ) 0. 10 3 (5 ) 1. 57 % ** (8 ) 1. 24 % ** (9 ) 12 (4 3) g ra ha m -e nt er pr is in g in ve st or pr ic e ch g v 0. 93 % (7 ) 1. 11 % * (7 ) � 1. 02 % (3 9) 0. 10 6 (4 ) 1. 53 % ** (9 ) 1. 24 % * (7 ) 4 (5 4) st oc k m ar ke t w in ne rs pr ic e ch g g /v 0. 74 % (9 ) 0. 76 % (1 2) � 0. 15 % (1 2) 0. 10 1 (6 ) 1. 46 % ** (1 0) 1. 19 % ** (1 0) 12 (4 4) fo ol is h sm al l c ap 8 r ev is ed pr ic e ch g g 0. 47 % (1 1) 0. 79 % (1 1) � 0. 32 % (2 1) 0. 04 1 (1 5) 1. 34 % * (1 1) 1. 03 % * (1 1) 6 (5 1) n ef f pr ic e ch g v 0. 54 % (1 0) 0. 89 % ** (9 ) 0. 25 % * (7 ) 0. 08 7 (1 0) 1. 28 % ** (1 2) 0. 86 % ** (1 3) 23 (3 1) o ’n ei l’ s c a n sl im r ev is ed 3r d e di tio n pr ic e ch g g 0. 26 % (1 7) 0. 43 % (1 7) � 0. 53 % (3 1) 0. 01 0 (1 9) 1. 23 % * (1 3) 0. 99 % * (1 2) 8 (4 9) o ’s ha ug hn es sy : sm al l c ap g ro w th & v al ue pr ic e ch g g /v 0. 35 % (1 3) 0. 61 % * (1 4) 0. 67 % * (4 ) 0. 06 9 (1 3) 1. 06 % ** (1 4) 0. 66 % * (1 5) 25 (2 7) pr ic eto -f re ec as hfl ow pr ic e ch g v 0. 34 % (1 4) 0. 64 % (1 3) � 0. 29 % (1 8) 0. 04 5 (1 4) 1. 03 % * (1 5) 0. 51 % (1 8) 30 (1 9) g ra ha m -d ef en si ve in ve st or (n on u til ity ) pr ic e ch g v 0. 30 % (1 5) 0. 52 % * (1 5) 0. 35 % * (6 ) 0. 06 9 (1 2) 0. 97 % ** (1 6) 0. 59 % * (1 6) 20 (3 5) p/ e r el at iv e pr ic e ch g v 0. 08 % (1 8) 0. 14 % (2 3) � 0. 15 % (1 3) 0. 02 4 (1 6) 0. 54 % * (2 3) 0. 37 % (2 2) 33 (1 3) b uf fe tt: h ag st ro m pr ic e ch g g /v 0. 46 % (1 2) 0. 48 % ** (1 6) 0. 04 % ** (8 ) 0. 09 1 (9 ) 0. 49 % ** (2 5) 0. 41 % * (2 1) 30 (2 1) a r an ki ng s ar e fo r 18 sc re en ed po rt fo lio s w ith st at is tic al ly si gn ifi ca nt 1f m m � th at a a ii fo llo w s ov er th e tim e pe ri od 19 98 –2 01 1. e ac h va ri ab le is gi ve n al on g w ith ra nk in g w ith in ea ch va ri ab le in pa re nt he se s. r an ki ng s ar e ba se d on 56 sc re en ed po rt fo lio s. t ab le 4 da ta ar e so rt ed ba se d on ra nk in g of 1f m m � .s ta tis tic al si gn ifi ca nc e of on eta ile d te st is de no te d as * fo r 5% an d ** fo r 1% .a ll st at is tic s in ta bl e in cl ud e tr an sa ct io n co st s ba se d on a $5 0, 00 0 in iti al in ve st m en t. b v ar ia bl es ar e de fin ed as : g m m r is ge om et ri c m ea n m on th ly re tu rn , b fi is be st fit in th e ta bl e in de x fo r gi ve n fu nd , a m m r is m ea n m on th ly re tu rn , m d m r is m ed ia n m on th ly re tu rn ,s ha rp e is sh ar pe ra tio ,1 fm m � is th e 1fa ct or m ar ke tm od el m on th ly � ,3 fm m � is th e 3fa ct or m ar ke tm od el m on th ly � , an d av er ag e st oc ks he ld is av er ag e nu m be r of st oc ks he ld ov er en tir e sa m pl e tim e pe ri od . fo r g ro up de si gn at io n g is fo r g ro w th , v fo r v al ue , s fo r sp ec ia l. 170 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 portfolios demonstrate excess risk-adjusted returns beyond what chance would predict, supporting the argument that there are aaii screens that exploit pricing inefficiencies. 7. top aaii portfolio strategies the analysis up to this point focused on the overall performance of aaii portfolio strategies. we now analyze the performance of specific strategies with higher performance rankings. we first ranked all portfolios from best (1) to worst (56) for each of the six different performance measures and for the average number of stocks held in the portfolio over our period of study. table 4 presents the ranking comparisons when returns are adjusted for transaction costs given a $50,000 initial investment. portfolios are presented in the rank order of statistically significant 1fmm �s using a 5% significance cutoff. we chose the 1fmm � as the base for ranking comparisons since it represents the starting point for our extension of performance measures beyond s&c’s study. the portfolio’s rank order for each performance measure is in parentheses, allowing a comparison of rankings across different performance measures for each portfolio strategy. the 18 screened portfolios in table 4 represent the aaii strategies that had statistically significant 1fmm �s over the 1998 through 2011 period. the 18 portfolios represent about 32% of the portfolios in this study, which is well above the 5% that we would expect to be significant by chance given the 5% level of significance. a few of the top 18 portfolios ranked by the 1fmm � were not statistically significant and were replaced by the next highest ranked portfolio that was statistically significant. only the last two portfolios had statistically significant �s without being in the top 18 based on the size of �. the 1fmm �s in table 4 are not only statistically significant but are materially attractive ranging from 271 basis points to 54 basis points. an important first observation from table 4 is that the (3fmm) �s and (1fmm) �s have very similar rankings. there is also consistency in statistical significance with all but two of the 3fmm portfolio �s achieving at least 5% statistical significance. even after accounting for the added risk factors advocated by efficient market theorists, over 28% (16/56) of the aaii portfolios achieved statistically significant �s given transaction costs with a $50,000 initial investment. statistically significant �s from the 3fmm are also relatively large, ranging from 239 basis points to 41 basis points. these findings run counter to the efficient market hypothesis. the results in table 4 are based on an investment of only $50,000. as the investment is increased transaction costs become a smaller drag on returns and even higher proportions of the aaii portfolios should achieve statistically significant �s. for example, while we do not show the results here to conserve space, we replicated table 4 with an initial investment of $100,000 and found similar ranking patterns with statistically significant 1fmm �s for almost 60% of the aaii portfolios and a little over 39% of the aaii portfolios with statistically significant 3fmm �s. full output with the higher initial investment is available from the authors upon request. the sharpe ratio ranking for portfolios in table 4 are highly correlated with the rankings of 1fmm �s and 3fmm �s. in general, a good portfolio is a good portfolio whether the sharpe ratio, 1fmm �, or 3fmm � is used for performance measurement. nevertheless, 171d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 there are two obvious discrepancies. the magnet complex price change portfolio is a growth style portfolio and is highly concentrated with only two stocks on average. this portfolio has a sharpe ratio rank of 22 but a much better rank of 7 for the 1fmm � and 5 for the 3fmm �. this portfolio has some attractive performance features but does not do well on a reward for total risk basis. on the other hand, buffett: hagstrom price change is a value/growth portfolio with an average holding of 30 stocks. the sharpe ratio rank of nine out of 56 looks good but a much lower rank of 25 for the 1fmm � and 21 for the 3fmm does not look attractive. even though many performance measures are highly correlated it is a good idea to consider all the risk adjusted return measures in table 4. rankings of portfolios in table 4 based on simple gmmr and ammr measures relative to the best fit benchmark indexes are also highly correlated with all but the mdmr measure. skewing of returns explains why a portfolio may rank high in all the other measures but low in the comparison of mdmr with the best fit index. positive skewing because of large positive outliers makes the average higher than the median, resulting in better performance measured by the difference in the average and the bfi than the performance measured by the difference in the median and bfi. investors may prefer positive skewing as a measure of upside potential but it takes a long term investment horizon to capture the outlier returns that play a large role in performance. the first two portfolios in table 4 present an example of how an individual investor should use all the measures to include skewing. the petroski 9 price change and the magnet simple price change portfolios are both highly concentrated (4 and 3 stocks on average, respectively) value portfolios that top the rankings in performance measures. both portfolios also demonstrate positive skewing of returns since average returns are much higher than the median returns relative to the benchmarks. this combination is good if the investor is looking for a long run commitment to a strategy where positive skewing is rewarded. however, the investor must also be comfortable with a highly concentrated portfolio. 8. conclusions as technology advances individual investor gain access to many of the same tools and strategies available to professional investors. aaii has been at the forefront of these advancements. our study extended work by s&c to test the investment performance of aaii screens. we extended the period of analysis used by s&c (1998–2005) to include the period around the great recession of 2009 to see if the aaii strategies weathered the storm. our results with performance measures used by s&c were similar to theirs. we also expanded the analysis to include the single-factor market model and the three-factor f&f model measures of excess returns (�s). even when accounting for the view that size and market/ book characteristics are really risk premiums rather than unpriced information, we find evidence of statistically significant excess returns well beyond what chance would explain. these findings do not support the efficient market hypothesis and suggest that individuals can achieve excess risk adjusted returns from following mechanical trading recommendations. we found evidence that a good aaii portfolio based on one measure tends to be a good portfolio for other performance measures. the exception is when median returns are used and 172 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 the portfolio has skewing. we found positive skewing in some aaii portfolios that long run investors might find attractive when combined with high ranking performance in other measures. for the top performing portfolios, significant �s were not because of reliance on small cap stocks or undervalued stocks based on low market to book ratios. the finding of statistically significant three-factor model �s from mechanical trading screens with relatively small amounts invested over this tumultuous financial market period is impressive. larger investors would find even higher percentages of the 56 screens to have attractive performance since trading costs would be less of a drag on returns. we could not control several features of this study that might make the performance of aaii screened portfolios more attractive. for example, the aaii data do not include dividends in either the portfolio return calculation or index return. furthermore, transaction costs could be lowered if rebalancing periods could be extended from monthly to quarterly, semiannually, or annually. rebalancing methods other than equal weighting might also affect performance for many of the aaii strategies depending on whether the strategy uses momentum or longer run value screens. in general, given the wide variety of different strategies available in the aaii service, it is likely that most strategies will do better with different combinations of holding periods and rebalancing methods. there would seem to be potential for a better service allowing individual investors to match strategies with the most appropriate holding periods and rebalancing methods. additional work is needed on implementation of mechanical screens such as those provided by the aaii service to help individual investors gain more equal footing with professionals. notes 1 fama (1970) is the generally accepted father of the efficient market hypothesis (emh) and passive investment philosophy. 2 see www.aaii.com for more information about aaii products and services. 3 we analyzed differences between the 54 strategies used by s&c and the 56 strategies in our study. we were able to add nine new screens because aaii added complete data for these portfolios after the s&c study was conducted. we lost seven screens used in the c&s study either because there was no longer complete data over the extended study period or because the screen style could not be classified. to test for the implications of the different set of screens we first ran our analysis over the s&c period using the s&c strategies and then using our strategies. the differences were not statistically significant. we do not provide this output here to conserve space but the data are available upon request. 4 differences in performance results in table 3 and findings in the s&c study are small in general, but one exception occurs for the sharpe ratio without transaction costs. we find 82% of our portfolios with no transaction costs beat the bfi sharpe ratio while s&c find only 72% of their portfolios beat the bfi sharpe ratio. when transaction costs are included, our findings for the sharpe ratio are again very similar to the s&c results. differences in our sample portfolios appear to be responsible for this outcome. 173d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 table 1 illustrates the specific differences in our sample and the s&c sample but more specific analysis of differences in samples for the two studies is available upon request. 5 data for the values of smb and hml in the three factor model are available in the following website (hgttp://mba.tuck.dartmough.edu/pages/faculty/ken.french/data_ library.html). references anderson, r. i., & loviscek, a. l. (2005). in search of information content: portfolio performance of the 100 best stocks to own in america. financial services review, 14, 97–109. arbel, a., & strebel, p. (1983). pay attention to the neglected firm. journal of portfolio management, 9, 37–42. banz, r. w. (1981). the relationship between returns and market value of common stock. journal of financial economics, 9, 3–18. ball, r. (1978). anomalies in relationships between securities’ yields and yield surrogates. journal of financial economics, 6, 103–126. ball, r., & brown, p. (1968). an empirical evaluation of accounting income numbers. journal of accounting research, 6, 159–178. barber, t., & odean, b. (2000). trading is hazardous to your wealth: the adverse stock investment performance of individual investors. journal of finance, 55, 773–806. basu, s. (1977). investment performance of common stocks in relation to their price-earnings ratios: a test of the efficient market hypothesis. journal of finance, 32, 663–682. beltz, j., & jennings, r. (1997). wall street week with louis rukheyser recommendations: trading activity and performance. review of financial economics, 6, 15–27. bessembinder, h., & chan, k. (1998). market efficiency and returns to technical analysis. financial management, 27, 5–17. bodie, z., kane, a., & marcus, a. j. (2013). essentials of investments. new york, ny: mcgraw-hill irwin. bolsher, p., trahan, e., & venkateswarren, a. (1998). how mad is mad money? jim cramer as a stock picker and portfolio manager. journal of investing, 21, 27–39. brennan, m. j. (1995). a perspective on accounting and stock prices. journal of applied corporate finance, 8, 43–52. brush, j. s. (1986). eight relative strength models compared. journal of portfolio management, 13, 21–28. campbell, j., grossman, s., & wang, j. (1993). trading volume and serial correlation in stock returns. quarterly journal of economics, 108, 905–934. campbell, j., & shiller, r. (1998). stock prices, earnings and expected dividends. journal of finance, 43, 661–676. campbell, j. w., lo, a. w., & mackinney, a. c. (1997). the econometrics of financial markets. princeton, nj: princeton university press. choi, j. j. (2000). the value line enigma: the sum of known parts? journal of quantitative analysis, 35, 485–498. conrad, j., & kaul, g. (1988). time variation in expected returns. journal of business, 61, 409–425. dreman, d. n., & berry, m. a. (1995). overreaction, uderreaction, and the low-p/e effect. financial analysts journal, 51, 21–30. debondt, w. f. m., & thaler, r. h. (1985). does the stock market overreact? journal of finance, 40, 793–805. debondt, w. f. m., &thaler, r. (1987). further evidence of stock market overreaction and market seasonality. journal of finance, 42, 557–581. desai, h., & jain, p. c. (1995). an analysis of recommendations of “superstar” money managers at barron’s annual roundtable. journal of finance, 50, 1257–1273. fama, e. (1970). efficient capital markets: a review of theory and empirical work. journal of finance, 25, 383–417. 174 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 fama, e., fisher, l., jensen, m., & roll, r. (1969). the adjustment of stock prices to new information. international economic review, 10, 1–21. fama, e. f., & french, k. r. (1996). multifactor explanations of asset pricing anomalies. journal of finance, 51, 55–84. fama, e., & macbeth, j. (1973). risk return and equilibrium: empirical tests. journal of political economy, 81, 607–636. ferreira, e. j., & smith, s. d. (2003). “wall street week”: information or entertainment? financial analysts journal, 59, 45–53. griffin, p. a., james, j. j., & zmijewski, m. e. (1995). how useful are wall street week stock recommendations? journal of financial statement analysis, 1, 33–52. hirchleifer, d. (2001). investor psychology and asset pricing. journal of finance, 56, 1533–1597. jegadeesh, n. (1990). evidence of predictable behavior of security returns. journal of finance, 45, 881– 898. jegadeesh, n., & titman, s. (1993). returns to buying winners and selling losers: implications for stock market efficiency. journal of finance, 48, 65–91. jegadeesh, n., & titman, s. (2001). profitability of momentum strategies: an evaluation of alternative explanations. journal of finance, 56, 699–720. jensen, m. c. (1968). the performance of mutual funds in the period 1945–1964. journal of finance, 23, 2, 389–416. jensen, m., & ruback, r. s. (1983). the market for corporate control: the scientific evidence. journal of financial economics, 11, 5–50. kacperczyk, g. m., van nieuwerburgh, s., & veldkamp, l. (2013). time varying fund manager skill. journal of finance, 69, 1455–1484. klein, a. (1986). the timing and substance of divestiture announcements: individual, simultaneous and cumulative effects. journal of finance, 41, 685–696. korniotis, g. m., & kumar, a. (2013). do portfolio distortions reflect superior information or psychological biases? journal of financial and quantitative analysis, 48, 1–45. lo, a. w., & mackinlay, a. c. (1990). when are contrarian profits due to market overreaction? review of financial studies, 3, 175–206. loviscek, a. l., & jordan, w. j. (2000). stock selection based on morningstar’s ten-year, five-star general equity mutual funds. financial services review, 9, 145–157. mcconnell, j., & muscarella, c. (1985). corporate capital expenditure decisions and the market value of the firm. journal of financial economics, 14, 399–422. metrick, a. (1999). performance evaluation with transaction data: the stock selection of investment newsletters. journal of finance, 54, 1743–1775. nofsinger, j. r. (2013). the psychology of investing (5th ed.). upper saddle river, nj: prentice hall. olson, d. o., nelson, j., will, c., & mossman, c. (1998). a test of the investors’ daily stock ranking system. the financial review, 33, 161–176. oppenheimer, h. (1981). common stock selection: an analysis of benjamin graham’s “intelligent investor” approach. ann arbor, mi: umi research press. palman, o., sun, h., & tang, a. p. (1994). the impact of publication of analysts’ recommendations on returns and trading volume. the financial review, 29, 395–417. pari, r. a. (1987). wall street week recommendations: yes or no? journal of portfolio management, 14, 74–76. peavy, j. w., & goodman, d. (1983). the significance of p/es for portfolio returns. journal of portfolio management, 9, 43–47. peirce, d., & roley, v. (1985). stock prices and economic news. journal of business, 59, 49–67. pinches, g. (1970). the random walk hypothesis and technical analysis. financial analysts journal, 26, 104–110. pompian, m. (2012). behavioral finance and wealth management: how to build optimal portfolios that account for investor biases (2nd ed.). hoboken, ny: john wiley and sons. 175d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 pruitt, s. w., & white, r. e. (1988). who says technical analysis can’t beat the market? journal of portfolio management, 14, 55–58. rendleman, r. j., jones, c., & latane, h. a. (1982). empirical anomalies based on unexpected earnings and the importance of risk adjustment. journal of financial economics, 10, 269–287. rosenberg, b., reid, k., & lanstein, r. (1985). persuasive evidence of market inefficiency. journal of portfolio management, 13, 9–17. schadler, f. p., & cotton, b. d. (2008). are the aaii stock screens a useful tool for investors? financial services review, 17, 185–201. sharpe, w. f. (1963). a simplified model of portfolio analysis. management science, 9, 277–293. shiller, r. (1981). do stock prices move too much to be justified by subsequent changes in dividends? american economic review, 71, 421–436. 176 d.s. north, j.l. stevens / financial services review 24 (2015) 157–176 the overlooked momentum traders in 401(k) plans ning tanga,* adepartment of finance, college of business administration, san diego state university, 5500 campanile drive, sse 3306, san diego, ca 92182-8236, usa abstract using a unique dataset on over one million 401(k) traders, we investigate momentum trading in 401(k) plans. we identify momentum traders in each quarter and evaluate how these traders perform. results indicate the existence of momentum traders. however, there is no evidence that they successfully improve their portfolio performance. instead, momentum sellers sell the outperformed funds. overall, momentum traders could lose up to 2.14% per year. in seeking to explain such losses, we observe that 401(k) momentum traders follow a naïve momentum strategy. they do not have the ability to select funds with momentum investing styles but, instead, simply chase past returns. © 2016 academy of financial services. all rights reserved. jel classification: g11; g12; g23 keywords: pension fund management; momentum; trading; portfolio performance; 401(k) plan 1. introduction momentum trading is a trading strategy whereby investors buy past winners and sell past losers. numerous studies point out that a hypothetical investor who follows momentum trading strategy can generate significant positive returns in the short term (see, e.g., jegadeesh and tittman, 1993, 2001; rouwenhorst, 1998). performance persistence in mutual funds further confirms the possibility of realizing abnormal return by chasing past winner funds (goetzmann and ibbotson, 1994; grinblatt and titman, 1992; hendricks et al., 1993). acknowledgement: this research is supported by the university grants program at san diego state university. the author also acknowledges vanguard for providing recordkeeping data under restricted access conditions and support from the pension research council at the wharton school. the author thanks olivia s. mitchell, stephen p. utkus, marie-eve lachance, and s. g. badrinath for helpful comments. * corresponding author. tel.: �1-619-594-2082; fax: �1-619-594-3272. e-mail address: ntang@mail.sdsu.edu financial services review 25 (2016) 51–72 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. consequently, many portfolio managers, institutional and individual stock investors, and mutual fund traders subscribe to the view that momentum strategies yield significant profits and follow such strategies (jegadeesh and titman, 2001; grinblatt and keloharju, 2000; grinblatt, titman, and wermers, 1995; solomon, solges, and sosyura, 2014). in particular, a series of articles have documented a strong positive relation between mutual fund past performance and subsequent fund inflows (see, e.g., goetzmann and massa, 2002; sirri and tufano, 1998). however, most of previous studies are based on fund level data with all individual traders aggregated. individual momentum traders were not identified and their performances were not evaluated. little is known whether individual traders could successfully implement momentum strategy to yield profit in real life. here we ask whether there exist momentum traders in 401(k) plans and whether they could achieve portfolio performance improvements. this question fits into the larger question of whether 401(k) traders can improve their 401(k) portfolios by implementing a trading strategy—a question of particular interest to policymakers and plan sponsors who oversee the plans, and to plan participants who will rely on their 401(k) accruals to finance their retirements. if 401(k) participants were able to successfully adopt a trading strategy, such as momentum trading, that would boost their 401(k) balance by the time of retirement, this would not only benefit the economic welfare of individual participants but would also increase soundness and stability of the entire retirement system. however, if momentum trading and other strategies used to improve 401(k) performance do not actually generate gains or even lead to losses, participants need to know this and refrain from adopting such strategies. without information on how such strategies perform, individuals could end up siphoning off their retirement wealth with inappropriate behavior, while the plans themselves also experience management costs. unfortunately, compared with the extensive literature on momentum trading outside retirement accounts (e.g., pettengill, edwards, and schmitt, 2006; pettengill, edeards, and griggs, 2009), momentum traders in 401(k) plans have been largely overlooked. the literature contains little on whether 401(k) traders adopt momentum strategies and how these momentum traders perform. in fact, there are many reasons to suspect that investors may exhibit different behavior when trading with retirement and nonretirement accounts. studies on mental accounting observe that investors tend to split their investment into a safe account, which is designed to maintain wealth level, and a risky account, used for speculation; their choices in separate accounts vary (choi, laibson, and madrian, 2007; rockenbach, 2004). a 401(k) account would be considered a safe account, meant for securing retirement wealth, whereas active nonretirement trading accounts would be considered risky accounts and meant for speculation. because active trading accounts and 401(k) plans serve completely different functions, investors may exhibit different behavior when trading with retirement and nonretirement accounts. in fact, literature has shown that 401(k) participants engage in trading infrequently because of inertia, which is sharply different from the excessive trading in discount brokerage accounts (agnew, balduzzi, and sunden, 2003; tang, mitchell, and utkus, 2012); plan sponsors seem not to expect 401(k) plan participants to excel in profitable trading. several studies have also highlighted the behavioral biases and financial literacy constraints that appear to hinder 401(k) plan participant decision-making. for example, plan participants use naive allocation strategies (agnew, 2002; benartzi and thaler, 2001); exhibit inertia in asset allocation and rebalancing (agnew et al., 2003; ameriks and zeldes, 52 n. tang / financial services review 25 (2016) 51–72 2004); display inconsistence between objective and subjective assessment of retirement adequacy (kim and hanna, 2015), and overinvest in employer stock (benartzi et al., 2007; even and macpherson, 2007; huberman and sengmueller, 2004; liang and weisbenner, 2002). the presence of these factors could contribute to the difference in trading outcomes in and outside retirement accounts. hence, it is necessary to study momentum traders in 401(k) plans and examine their performance. this article adds to the literature by investigating momentum traders and their performance in 401(k) plans. it also explores the causes of inefficiency among 401(k) momentum traders. the unique vanguard record-keeping dataset on over one million individual traders allows us to study the trading behavior in 401(k) plans at the individual account level. we follow the literature to identify momentum traders in each quarter by binomial test. our results indicate the existence of a group of traders who follow momentum strategies. however, we find no evidence of performance improvements. on the contrary, momentum sellers lose significantly by selling outperformed funds in certain cases; overall, momentum traders could lose up to 2.14% per year. in investigating why momentum traders in 401(k) plans do not produce gains, we find that these momentum traders aren’t able to identify funds with momentum investing styles, and thus don’t actually benefit from the momentum effect. instead, they naively follow past fund performances. the remainder of the article is organized as follows. section 2 describes our data and provides descriptive statistics. section 3 introduces our method of identifying momentum traders. section 4 shows the performance of momentum traders. section 5 explains the inefficiency of momentum trading. section 6 offers discussions on how our findings differ from prior studies and section 7 concludes. 2. data the data underlying this study, provided by the vanguard group, is a record-keeping dataset of 5,647,728 eligible employees in dc pension plans, mostly 401(k) plans, from january 2005 to december 2010. to demonstrate the representativeness of vanguard sample, table 1 panel a shows the plan level summary statistics as of december 2010. as of december 2010, the whole dataset contained records on over three million employees participating in 2,144 dc plans with a total of $259 billion of assets under management. the data spans 236 four-digit naics industries. it is noted that the sample is not a balanced panel since not every participant stayed in the plan during the sample period. therefore, we observe fewer participants as of december 2010 (3,281,505) than during the whole sample period (5,647,728). vanderhei et al., (2011) report that in 2010, there were 23.4 million 401(k) plan participants with a total asset value of $1.4 trillion in the u.s. market. these figures are from the ebri/ici dataset, which is a representative sample of the estimated universe of 401(k) plans. it implies that vanguard dataset represents 18% of the asset value and 14% of the participants of the extensive ebri/ici dataset. table 1 panel b summarizes participants’ characteristics in the whole sample in 2010. the median age of participants in 2010 is 46—very close to the national median participant age of 45 (vanderhei et al., 2011). the average and median individual account balances are $67,360 and 53n. tang / financial services review 25 (2016) 51–72 $20,998, compared with the national average and median account balances of $60,329 and $17,686. the average individual account risk exposure is 59.6%, close to the national average of 62% (vanderhei et al., 2011). these similarities lead us to believe that the vanguard data we use here is representative of the overall population of 401(k) participants. table 1 panel c shows trading statistics during sample period january 2005 through december 2010. out of 5,647,728 participants in the whole sample, around 1.5 million (26% of participants) traded during the six year sample period. participants in our sample traded 0.78 times per year with an average annual turnover rate of 36.24%.1 this contrasts sharply with the excessive trading in discount brokerage accounts. for example, barber and odean (2000) report a monthly turnover rate of 6% to 7% in their discount brokerage account sample. it is also important to note that, although overall 401(k) trading is limited, investors who do trade, trade a substantial amount. the average number of trades is 6.16 per year and the average annual turnover rate is 239.51% among traders. that is to say that infrequent trading is an unsound argument for disregarding 401(k) participants’ trading behaviors. among those who trade, their choices significantly impact their portfolios. to study how 401(k) participants reacted to past returns, we select those who traded in equity assets, including equity and balanced funds, between january 2005 and december table 1 descriptive plan, participant, and traders statistics a. plan characteristics (as of december 2010) number of participants 3,281,505 number of plans 2,144 total assets under management $259,090,794,149 number of industries 236 mean median b. participants characteristics (in 2010) male (yes � 1) 58.72% 1 age 45.57 46 online 401(k) account registration 64.15% 1 average individual account balance $67,360 $20,998 average individual risk exposure 59.60% 60.00% c. trading statistics (january 2005 through december 2010) no. of participants 5,647,728 no. of traders 1,458,037 no. of equity traders 1,390,392 average number of trades per year (among all participants) 0.78 average number of trades per year (among traders) 6.16 average annual turnover rate (among all participants) 36.24% average annual turnover rate (among traders) 239.51% d. equity traders characteristics (in 2010) male (yes � 1) 63.75% 1 age 48.55 49 online 401(k) account registration 90.41% 1 average individual account balance $127,778 $64,111 average individual risk exposure 62.51% 66.97% note: the table shows the plan-level summary statistics as of december 2010 in panel a, individual participant characteristics in 2010 in panel b, the trading statistics during the sample period january 2005 through december 2010 in panel c, and individual characteristics in 2010 of those who traded equity assets in panel d. a participant is considered as a trader if he traded at least once during the sample period; a participant is considered as an equity trader if he traded equity assets including equity and balanced funds at least once during the sample period. we count a buy or sell transaction on a daily basis as one trade. to calculate individual annual turnover rate, we sum the absolute values of trading amount (both buy and sell) in one year and divide this sum by two. we then divide the annual trading amount by the account balance at the beginning of the year. 54 n. tang / financial services review 25 (2016) 51–72 2010.2 in total, 1,390,392 participants traded equity assets and they are included in our final trading sample. for each trader, we record the name and amount of individual funds purchased or sold every month together with the historical monthly fund returns. the data also reflects the funds each plan offered to its participants each month. we also incorporate individual month-end portfolio holdings, which we can use to calculate individual account balances each month. table 1 panel d reports the individual attributes of our selected trading sample on equity traders in 2010. compared with the results in panel b, we see that the subsection of participants who trade in their 401(k) plans tend to be affluent older men who use the internet to access their 401(k) accounts and who tend to be more risk-seeking in their investments than their non-trader counterparts. this confirms previous findings (agnew et al., 2003; tang et al., 2012). 3. momentum traders in 401(k) plans we follow goetzmann and massa (2002) and agnew et al., (2003) to use individual account activity to classify investors according to their conditional pattern of fund purchases and sells. momentum buyers (sellers) are defined as those traders who are more likely to buy (sell) funds with past positive (negative) returns (goetzmann and massa, 2002). after identifying momentum buyers and sellers, we then study the performance of momentum traders. this implies that our definition of momentum investing is different from the way jegadeesh and tittman (1993) apply the term in their profitable momentum strategies. the reason is that the focus of the article is to study the performance of individual momentum traders, instead of the outcome of a hypothetical investing strategy. specifically, we follow the literature (agnew et al., 2003; goetzmann and massa, 2002) and adopt binomial test to identify momentum traders. the null hypothesis for an investor to follow momentum buying strategy in a quarter is that the ratio of his purchases with positive returns to total purchases is equal to the percentage of funds with positive returns of all funds available to trade in that quarter. using a one-tailed binomial test, if the null hypothesis is rejected—that is, if the frequency of positive-return fund purchases is significantly higher than expected from a random distribution—the investor is classified as a momentum buyer. in a similar manner, the null hypothesis for momentum seller is defined as follows: buy h0:bp ˜ binomial�b, np n � sell h0:sn ˜ binomial�s, nn n � where each individual in each quarter has n funds to trade, of which np funds have positive returns and nn funds have negative returns. in all, he buys b funds, of which bp have positive returns. he also sells s funds, of which sn have negative returns. test results are based on the 1% significance level. this methodology distinguishes investors who follow certain trading strategies from those who trade randomly. because the methodology requires information on the investment 55n. tang / financial services review 25 (2016) 51–72 opportunities available to each investor, few studies have used it because of limitation of the data. instead, most articles have explored correlations between security flows and their returns or have regressed trades on past asset returns (see, e.g., grinblatt and keloharju, 2001; grinblatt et al., 1995). the problem with such methodologies is that they cannot truly distinguish between feedback and random trading strategies, because positive or negative feedback trading can result from market conditions rather than from any conscious strategic trading. we analyze investors’ trading strategies in buy decisions separately from sell decisions, as people may follow different strategies when they buy versus when they sell (grinblatt et al., 1995; sirri and tufano, 1998). we identify momentum traders quarterly. it is because we found 401(k) traders tend to change their trading strategies through time in our preliminary results; hence it is inappropriate to assume they follow the same strategy in a year or over a longer period and investigate their momentum trading performance during that long period. we do not choose monthly level, as 401(k) plan participants are prone to inertia when rebalancing (agnew et al., 2003). therefore, we identify momentum traders quarterly so that we will have sufficient trading data from each individual. momentum trading is considered as a short-term (less than one year) strategy. the time interval commonly selected to study momentum strategy ranges from one month to one year (gruber, 1996; jegadeesh and tittman, 1993, 2001; warther, 1995; zheng, 1999). in the baseline analysis, we focus on investors’ reactions to the past month’s return. we assume investors react to past returns without delay, as the momentum effect only exists in the short-term. specifically, each quarter, we first calculate the number of funds with positive or negative returns one month ago. we then determine if a buyer or a seller follows momentum strategy in response to last month fund return in that quarter using the binomial test. the fund cash flows by momentum traders identified here will be used to construct testing portfolios in the following sections. in the robustness test, we rerun the analysis based on investors’ reaction to funds’ returns over the past quarter and year (time lag j � 3 and 12 months). fig. 1 panel a shows the cumulative number and percentage of traders who follow momentum strategies with various return lags. specifically, we calculate the total number of momentum buyers and sellers, and the percentage of momentum traders out of all buyers and sellers throughout the testing period (january 2005 through december 2010). fig. 1 panel b reports quarterly momentum trading statistics. first, we establish that each quarter, there does exist a group of alert traders who follow momentum trading in 401(k) plans. for example, as depicted in fig. 1 panel a, in reaction to the past month’s return, a total of 76% of the 1.15 million buyers adopted momentum strategies at least once during the sample period; sixty percentage of the 1.04 million sellers sell the previous month’s losers, making them momentum sellers at least in one quarter during sample period.3 quarterly average statistics in fig. 1 pane b show that each quarter, a portion of 58% of buyers followed momentum strategies when evaluated with past month’s return; forty percentage of sellers are momentum traders. in addition, we confirm that patterns in buy and sell decisions vary. the momentum strategy is more popular in buy decisions than in sell decisions. 56 n. tang / financial services review 25 (2016) 51–72 4. performance of momentum traders 4.1. performance measures in this section, we ask whether momentum traders, as identified in the previous section, improve their portfolio performance. to address this point, we follow zheng (1999) to evaluate trading performances. we first construct three portfolios based on the quarterly momentum identifiers (876,941 momentum buyers and 624,307 momentum sellers) in response to past month fund return obtained in the previous section: fig. 1. momentum traders statistics. (a) cumulative statistics during january 2005 through december 2010, (b) quarterly statistics. note: panel a in this figure shows the cumulative number and percentage of buyers and sellers who followed momentum strategies in response to returns over past 1, 3, and 12 months during the sample period (january 2005 through december 2010). percentage of momentum buyers (sellers) is calculated by dividing the total number of momentum buyers (sellers) by the total number of buyers (sellers) throughout the sample period. panel b in this figure reports the quarterly momentum trading strategy. number of momentum buyers (sellers) and total number of buyers (sellers) in each quarter are first calculated; then the average number and percentage of momentum buyers and sellers are shown in panel b. 57n. tang / financial services review 25 (2016) 51–72 1. momentum buyer portfolio: long funds bought by a momentum buyer each month and weighted by fund’s trading amount. 2. momentum seller portfolio: long funds sold by a momentum seller each month and weighted by fund’s trading amount. 3. long-short portfolio: long the momentum buyer portfolio and short the momentum seller portfolio. portfolio 1 and 2 track what a trader purchases or sells each month in the quarter when he is identified as a momentum trader. for example, in the first quarter of 2008, if an individual is identified as a momentum buyer in response to past month return, a momentum buyer portfolio is constructed based on what he purchased each month in that quarter. however, if the individual stopped following momentum trading in the next quarter, his portfolio will not continue to be included in the momentum buyer portfolio next quarter. portfolio 3 takes a long position in portfolio 1 and a short position in portfolio 2. that is, portfolio 3 evaluates the overall performance of momentum traders by buying what momentum buyers purchased and selling what momentum sellers sold. we also assume that investors hold the portfolio for six months in the baseline analysis. therefore, once the portfolio is constructed, we calculate the average monthly return for the six months after the formation of portfolios. we follow zheng (1999) in using both excess returns and risk-adjusted returns to measure portfolio performance. the excess return is calculated as rp,t e � rp,t � rm,t (1) where rp,t e is the excess return of portfolio p in month t; rp,t indicates the value-weighted portfolio’s monthly raw return; and rm,t is the market return in month t. note that more than one trader can follow the same strategy each month. thus, each month we calculate the cross-sectional average of portfolio excess returns realized by all traders following the same strategy. the time-series mean is presented and t-statistics indicate if the average return is statistically different from zero. we use two methods to calculate risk-adjusted returns. the first is the “portfolio regression” approach, which estimates time-series regressions for returns of the portfolios. both capm model and fama-french three-factor model are used to adjust for risk. the second method is the “fund regression” approach, which estimates capm and fama-french threefactor time-series regression for each fund and averages the � estimates across individual funds in the portfolio (zheng, 1999). in the first “portfolio regression” approach, we run the following ols regressions: rp,t �� rf,t � �p 1 � �p 1�rm,t � rf,t� � �p,t (2) rp,t �� rf,t � �p 3 � �p,rm 3 �rm,t � rf,t� � �p,smb 3 smbt � �p,hml 3 hmlt � �p,t (3) where rp,t � is the cross-sectional average of raw returns realized by individual portfolios following the same strategy in month t; rf,t is the risk-free rate in month t; rm,t is the market return in month t; smbt is the return on the mimicking portfolio for the common size factor in stock returns in month t; and hmlt is the return on the mimicking portfolio for the 58 n. tang / financial services review 25 (2016) 51–72 common book-to-market equity factor in stock returns in month t. in each regression, �p is the risk-adjusted excess return. as zheng (1999) points out, the “portfolio regression” approach does not require each fund to survive for a long period of time; however, it does not take into account time-varying portfolio compositions and risk characteristics. by contrast, the “fund regression” approach considers portfolio variation over time, but it requires each fund to have sufficient past return observations (30 monthly past returns in this study).4 the � in the second approach are obtained by averaging the � estimates across individual funds in the portfolio: ri,t � rf,t � �i 1 � �i 1�rm,t � rf,t� � �i,t (4) ri,t � rf,t � �i 3 � �i,rm 3 �rm,t � rf,t� � �i,smb 3 smbt � �i,hml 3 hmlt � �i,t (5) �p,t � ���i � �i,t�* �i,t��i,t (6) where ri,t indicates the rate of return of fund i in month t; �p,t is the excess return of individual portfolio in month t; �i and �i,t, which are used to calculate �p,t, are from eqs. (4) and (5); and �i,t is the portfolio weight of fund i in month t. after obtaining �p,t of each individual portfolio, we take the average across all traders following the same strategy in each month and show the time-series means and t-statistics on these averages. 4.2. results table 2 panel a shows performance of the three portfolios. column (a) shows monthly excess return over market return as calculated by eq. (1). alphas in columns (b) and (c) are risk-adjusted returns under the “portfolio regression” approach estimated by eqs. (2) and (3). alphas in columns (d) and (e) report risk-adjusted returns under the “fund regression” approach estimated by eqs. (4) through (6), which consider time-varying risk characteristics and portfolio compositions. alpha1 in columns (b) and (d) are risk-adjusted returns estimated by capm model, and alpha3 in columns (c) and (e) are risk-adjusted returns estimated by the fama-french three-factor model. results on momentum buyer portfolio (portfolio 1) are mixed and none of the results is significantly different from zero. on the other hand, the significantly positive excess and risk-adjusted returns achieved by momentum seller portfolio (portfolio 2) indicate that momentum sellers are following a losing strategy. they sell funds that significantly outperform the market in the following six months. for example, as estimated by fama-french model under fund regression approach, momentum sellers lose 0.18% per month, which is equivalent to 2.14% per year. overall, performance of portfolio 3 shows that it is a losing strategy to buy what momentum buyers purchased and sell what momentum sellers sold in certain cases. the loss can be as high as 0.17% each month, which is equivalent to a 2.02% annual loss, when measured by capm model under fund regression approach. 59n. tang / financial services review 25 (2016) 51–72 table 2 performances of momentum traders (value-weighted) portfolios (a) excess return (b) alpha1-portfolio regression approach (c) alpha3-portfolio regression approach (d) alpha1-fund regression approach (e) alpha3-fund regression approach a. six-month holding period 1. momentum buyer 0.02% 0.01% 0.07% �0.02% 0.05% (0.22) (0.20) (1.14) (�0.27) (0.62) 2. momentum seller 0.13%** 0.13%** 0.11%** 0.15%** 0.18%*** (2.62) (2.60) (2.00) (2.59) (3.38) 3. portfolio 1 portfolio 2 �0.11% �0.11% �0.04% �0.17%** �0.13%** (�1.39) (�1.40) (�0.53) (�2.13) (�2.08) b. one-month holding period 1. momentum buyer 0.06% 0.07% 0.08% �0.03% 0.03% (0.35) (0.41) (0.44) (�0.16) (0.23) 2. momentum seller 0.01% 0.003% 0.02% 0.08% 0.12% (0.05) (0.02) (0.13) (0.52) (0.81) 3. portfolio 1 portfolio 2 0.05% 0.07% 0.06% �0.11% �0.09% (0.24) (0.30) (0.25) (�0.56) (�0.56) c. one-year holding period 1. momentum buyer 0.03% 0.03% 0.07% 0.02% 0.04% (0.71) (0.78) (1.92) (0.33) (1.00) 2. momentum seller 0.14%*** 0.14%*** 0.17%*** 0.16%*** 0.17%*** (2.95) (2.93) (3.25) (3.14) (3.27) 3. portfolio 1 portfolio 2 �0.11% �0.11% �0.10% �0.14%** �0.13%*** (�1.85) (�1.84) (�1.72) (�2.53) (�2.75) d. financial turmoil period excluded, six-month holding period 1. momentum buyer 0.04% 0.05% 0.09% �0.01% 0.003% (0.48) (0.71) (1.43) (�0.13) (0.04) 2. momentum seller 0.10% 0.10% 0.10% 0.10% 0.14%** (1.95) (1.93) (1.72) (1.70) (2.55) 3. portfolio 1 portfolio 2 �0.06% �0.05% �0.01% �0.11% �0.14%** (�0.76) (�0.67) (�0.17) (�1.39) (�2.01) e. consistent momentum traders (six-month holding period) 1. momentum buyer 0.03% 0.03% 0.08% �0.01% 0.06% (0.33) (0.35) (1.17) (�0.06) (0.72) 2. momentum seller 0.10% 0.11% 0.09% 0.12% 0.15%*** (1.93) (1.91) (1.50) (1.98) (2.71) 3. portfolio 1 portfolio 2 �0.08% �0.08% �0.01% �0.12% �0.09% (�0.86) (�0.86) (�0.13) (�1.44) (�1.36) note: three portfolios are constructed based on 876,941 momentum buyers and 624,307 momentum sellers identified in response to past month fund return: (1) momentum buyer portfolio: purchase funds bought by a momentum buyer in the quarter he is identified as a momentum buyer, weighted by funds’ trading amount; (2) momentum seller portfolio: purchase funds sold by a momentum seller in the quarter he is identified as a momentum seller, weighted by funds’ trading amount; (3) portfolio 1portfolio 2: long momentum buyer portfolio and short momentum seller portfolio. excess return is calculated as the different between portfolio raw return and market return; cross-sectional average of excess returns among all momentum buyers (sellers) are first calculated each month and time-series mean and t-statistics are shown in column (a); to calculate alpha1 and alpha3 under portfolio regression approach, we first obtain time-series raw returns by averaging portfolio raw returns across all momentum buyers (sellers) each month and risk-adjusted returns from capm and fama-french three-factor models are shown in column (b) and (c), respectively; column (d) and (e) show the risk-adjusted returns under fund regression approach; alphas for individual momentum buyers (sellers) portfolios are first calculated using capm and fama-french three-factor models and we take the average of alphas across all momentum buyers (sellers) each month and show the time-series means and t-statistics. panel a includes the whole sample period (january 2005 through december 2010) and assumes holding period to be six months; panel b and c assume holding periods are one month and one year, respectively; panel d excludes the financial turmoil period (october 2008 through march 2010) from the whole sample period and shows average six-month performance; panel e includes only consistent momentum traders (111,082 momentum buyers and 42,574 momentum sellers) and assumes holding periods to be six months. t-statistics are reported in parenthesis. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. 60 n. tang / financial services review 25 (2016) 51–72 4.3. performance with various holding periods thus far, we have assumed that traders hold the portfolio for six months. however, they may hold the portfolio for different periods. accordingly, we next explore whether portfolio performance changes with various holding periods. table 2 panel b shows the monthly excess returns and risk-adjusted returns for a holding period of one month, and panel c shows the results for a one-year holding period. we find that even as the holding period changes, portfolio performance does not improve significantly. none of the portfolio performances is statistically significant when holding period is one month in panel b. with a holding period of one year, momentum sellers lose significantly under all estimation models. the loss can be as high as 0.17% or 2.02% per year. the overall performance of momentum traders still yields significantly negative returns in certain cases. for example, monthly return of portfolio 3 is 0.14%, or 1.67% annually under capm model (fund regression approach). 4.4. effect of the financial crisis since the sample period includes the 2008–2009 financial crisis, it is necessary to test whether the performance results discussed above change when we exclude the turmoil period. we define the period of financial turmoil to be from october 2008 to march 2009. after excluding this period, we repeat the above analysis assuming holding periods being six months. the results, shown in table 2 panel d, are consistent with those reported earlier. momentum traders do not benefit from momentum trading; they may even lose significantly with certain specifications. for example, as measured by fama-french model under fund regression approach, momentum sellers lose 0.14% every month by selling the outperformed funds; overall, momentum traders lose 0.14% each month, which is equivalent to an annual loss of 1.67%. 4.5. equally weighted portfolio the performance of a portfolio is determined by two decisions: what funds are selected and how much is allocated to each selected fund. the above results based on value-weighted portfolios reflect the outcomes of both decisions. is it possible that momentum traders in 401(k) plans still have the ability to select the right funds, but just fail to allocate the money wisely among these funds? to answer this question, we examine the performance of equally weighted portfolios. the three portfolios under analysis are the same as in table 2, except that they are constructed equally weighted. for example, when a trader was identified as a momentum buyer in the first quarter of 2009, we track what he purchased each month in that quarter and equally allocate money to these purchased funds to construct a momentum buyer portfolio (portfolio 1). then we calculate the average monthly return for various holding periods after portfolio was established. table 3 shows the performance of equally weighted portfolios. the results are consistent with previous findings. momentum buyer portfolio (portfolio 1) may be able to outperform 61n. tang / financial services review 25 (2016) 51–72 table 3 performances of momentum traders (equally-weighted) portfolios (a) excess return (b) alpha1-portfolio regression approach (c) alpha3-portfolio regression approach (d) alpha1-fund regression approach (e) alpha3-fund regression approach a. six-month holding period 1. momentum buyer 0.02% 0.02% 0.07% �0.01% 0.05% (0.30) (0.28) (1.23) (�0.19) (0.67) 2. momentum seller 0.13%*** 0.13%*** 0.12%** 0.15%** 0.18%*** (2.77) (2.75) (2.15) (2.64) (3.38) 3. portfolio 1 portfolio 2 �0.11% �0.11% �0.04% �0.16%** �0.13%** (�1.41) (�1.41) (�0.55) (�2.10) (�2.07) b. one-month holding period 1. momentum buyer 0.06% 0.08% 0.09% �0.01% 0.04% (0.39) (0.44) (0.47) (�0.09) (0.28) 2. momentum seller �0.02% �0.03% �0.01% 0.05% 0.09% (�0.13) (�0.15) (�0.03) (0.34) (0.59) 3. portfolio 1 portfolio 2 0.09% 0.10% 0.09% �0.07% �0.05% (0.40) (0.46) (0.39) (�0.36) (�0.32) c. one-year holding period 1. momentum buyer 0.03% 0.04% 0.07%** 0.02% 0.04% (0.79) (0.85) (2.03) (0.05) (0.14) 2. momentum seller 0.15%*** 0.15%*** 0.17%*** 0.16%*** 0.18%*** (3.07) (3.05) (3.33) (3.19) (3.32) 3. portfolio 1 portfolio 2 �0.11% �0.11% �0.10% �0.14%** �0.13%*** (�1.88) (�1.88) (�1.72) (�2.51) (�2.75) d. financial turmoil period excluded, six-month holding period 1. momentum buyer 0.04% 0.05% 0.09% �0.01% 0.01% (0.54) (0.77) (1.50) (�0.09) (0.07) 2. momentum seller 0.11%** 0.11%** 0.10% 0.10% 0.14%** (2.07) (2.06) (1.83) (1.72) (2.55) 3. portfolio 1 portfolio 2 �0.07% �0.05% �0.01% �0.11% �0.13%** (�0.77) (�0.69) (�0.19) (�1.36) (�1.99) e. consistent momentum traders (six-month holding period) 1. momentum buyer 0.03% 0.03% 0.08% �0.001% 0.06% (0.39) (0.40) (1.23) (�0.01) (0.73) 2. momentum seller 0.11%** 0.11% 0.09% 0.12%** 0.15%*** (2.00) (1.99) (1.55) (2.05) (2.69) 3. portfolio 1 portfolio 2 �0.08% �0.08% �0.01% �0.12% �0.09% (�0.87) (�0.87) (�0.14) (�1.45) (�1.36) note: three portfolios are constructed based on 876,941 momentum buyers and 624,307 momentum sellers identified in response to past month fund return: (1) momentum buyer portfolio: purchase funds bought by a momentum buyer in the quarter he is identified as a momentum buyer, each fund is equally-weighted; (2) momentum seller portfolio: purchase funds sold by a momentum seller in the quarter he is identified as a momentum seller, each fund is equally weighted; (3) portfolio 1portfolio 2: long momentum buyer portfolio and short momentum seller portfolio. excess return is calculated as the different between portfolio raw return and market return; cross-sectional average of excess returns among all momentum buyers (sellers) are first calculated each month and time-series mean and t-statistics are shown in column (a); to calculate alpha1 and alpha3 under portfolio regression approach, we first obtain time-series raw returns by averaging portfolio raw returns across all momentum buyers (sellers) each month and risk-adjusted returns from capm and fama-french three-factor models are shown in column (b) and (c), respectively; column (d) and (e) show the risk-adjusted returns under fund regression approach; alphas for individual momentum buyers (sellers) portfolios are first calculated using capm and fama-french three-factor models and we take the average of alphas across all momentum buyers (sellers) each month and show the time-series means and t-statistics. panel a includes the whole sample period (january 2005 through december 2010) and assumes holding period to be six months; panel b and c assume holding periods are one month and one year, respectively; panel d excludes the financial turmoil period (october 2008 through march 2010) from the whole sample period and shows average six-month performance; panel e includes only consistent momentum traders (111,082 momentum buyers and 42,574 momentum sellers) and assumes holding periods to be six months. t-statistics are reported in parenthesis. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. 62 n. tang / financial services review 25 (2016) 51–72 the market in certain cases. however, momentum sellers lose significantly by selling good performers. overall, portfolio 3 has significantly negative risk-adjusted returns with certain specifications. these findings suggest that momentum traders do not display the ability to select the right funds to outperform the market. 4.6. consistent momentum traders the binomial test we used identifies momentum traders quarterly. it is possible for an investor to follow momentum trading strategy in a quarter and abandon the strategy later. the above analysis showed that momentum traders lose significantly. is it possible for a subgroup of momentum traders who consistently follow momentum trading strategies to perform better? if an investor consistently adopts the same strategy whenever he trades, he is expected to take this strategy more seriously and have a higher chance to benefit from it than other strategy followers. in this subsection, we select a subsample of momentum traders who consistently followed the momentum trading strategy during the sample period. a consistent momentum buyer is a trader who followed momentum buy strategy whenever he made a purchase in any quarter during sample period. the same criterion is used to identify consistent momentum sellers. we excluded those who purchased or sold only in one quarter during the sample period. we obtained 111,082 consistent momentum buyers, who represent 9.7% of buyers or 12.7% of momentum buyers during the sample period; there were 42,574 consistent momentum sellers representing 4.1% of all sellers or 6.8% of momentum sellers. we study consistent momentum traders’ portfolio performance as we did in previous sections. as shown in panel e in table 2 and 3, performance of these consistent or most momentum oriented momentum traders is similar to other momentum traders. they do not benefit from momentum trading; instead they sell outperformed funds. 4.7. identify momentum traders with different returns in this subsection, we test if results will change when we use different criteria to define momentum traders. in our baseline analysis, we identified momentum traders based on their reactions to past month return and found their wealth suffers from momentum trading. now we identify momentum traders in each quarter based on their responses to returns over the past quarter (j � 3) and year (j � 12). for example, to identify momentum traders in a quarter based on return over the past 3 months, we calculate the number of funds with positive and negative returns in the previous quarter and rerun the binomial tests described in section 3. we identified 856,358 (589,243) momentum buyers (sellers) in response to returns over the past 3 months and 577,626 (408,253) momentum buyers (sellers) in response to fund returns over the past 12 months. results in table 4 based on newly identified momentum traders’ portfolios are consistent with previous findings. for example, momentum sellers identified by using returns over past quarter lose 0.16% each month by selling outperformed funds, as estimated by capm model under portfolio regression approach. overall, momentum traders lose 0.18% each month, or 2.14% annually in portfolio 3. 63n. tang / financial services review 25 (2016) 51–72 4.8. investment opportunity constraints last, we explore the possibility that 401(k) momentum traders lose because of investment opportunity constraints in plan offerings, instead of their lack of selection ability. it is known that 401(k) participants are offered a limited set of funds to invest in each plan. their investment opportunity might be constrained if the funds offered are underperformed ones, which could cause investment loss (tang et al., 2010). to exclude such possibility, we construct an average fund portfolio, which invests in all available equity assets in the plan. it measures the performance the average investors realize in the 401(k) accounts. the value-weighted average fund portfolio has an excess return of 0.06% assuming holding period of six months; alphas from capm and fama-french models are 0.06% and 0.07%, respectively, under portfolio regression approach; alphas from capm and fama-french models are both 0.07% under fund regression approach. comparing with results in panel a table 2, these statistics indicate that the average equity fund in the sample 401(k) plans outperformed the market during the testing period; however, momentum traders table 4 performances of momentum traders based on j-month lagged returns portfolios (a) excess return (b) alpha1-portfolio regression approach (c) alpha3-portfolio regression approach (d) alpha1-fund regression approach (e) alpha3-fund regression approach a. momentum trading (j � 3) 1. momentum buyer �0.02% �0.02% 0.04% 0.06% 0.03% (�0.22) (�0.23) (0.65) (0.78) (0.43) 2. momentum seller 0.17%*** 0.16%*** 0.14%*** 0.14%** 0.17%*** (3.20) (3.18) (2.77) (2.35) (2.94) 3. portfolio 1 portfolio 2 �0.18%** �0.18%** �0.10% �0.09% �0.14%** (�2.40) (�2.39) (�1.57) (�1.25) (�2.04) b. momentum trading (j � 12) 1. momentum buyer 0.06% 0.02% 0.09% �0.05% �0.07% (0.62) (0.34) (1.35) (�0.56) (�0.80) 2. momentum seller 0.19%** 0.20%** 0.23%*** 0.08% 0.11% (2.16) (2.12) (2.78) (0.93) (1.57) 3. portfolio 1 portfolio 2 �0.13% �0.17% �0.15% �0.13% �0.18%** (�1.24) (�1.84) (�1.37) (�1.32) (�2.34) note: three portfolios are constructed based on 856,358 (589,243) momentum buyers (sellers) identified in response to fund returns over the past three months in panel a, and 577,626 (408,253) momentum buyers (sellers) identified in response to fund returns over the past 12 months in panel c. three constructed portfolios are: (1) momentum buyer portfolio: purchase funds bought by a momentum buyer in the quarter he is identified as a momentum buyer, weighted by funds’ trading amount; (2) momentum seller portfolio: purchase funds sold by a momentum seller in the quarter he is identified as a momentum seller, weighted by funds’ trading amount; (3) portfolio 1portfolio 2: long momentum buyer portfolio and short momentum seller portfolio. holding period is six months. excess return is calculated as the different between portfolio raw return and market return; cross-sectional average of excess returns among all momentum buyers (sellers) are first calculated each month and time-series mean and t-statistics are shown in column (a); to calculate alpha1 and alpha3 under portfolio regression approach, we first obtain time-series raw returns by averaging portfolio raw returns across all momentum buyers (sellers) each month and risk-adjusted returns from capm and fama-french three-factor models are shown in column (b) and (c), respectively; column (d) and (e) show the risk-adjusted returns under fund regression approach; alphas for individual momentum buyers (sellers) portfolios are first calculated using capm and fama-french three-factor models and we take the average of alphas across all momentum buyers (sellers) each month and show the time-series means and t-statistics. t-statistics are reported in parenthesis. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. 64 n. tang / financial services review 25 (2016) 51–72 buy funds underperforming the average fund and sell funds outperforming the average fund, which leads to a significant loss by momentum traders. their loss is because of their lack of selection ability instead of investment opportunity constraints. we obtained the same conclusion with other holding periods and with equally weighted average fund portfolio. hence, we conclude that momentum traders in 401(k) plans lose significantly. although they move towards good performers in certain cases, they move away from good performers as well. even though 401(k) traders trade infrequently and pay limited transaction costs, they do not benefit from momentum trading. if performance improvement is not linked to feedback trading, then what are the causes of such inefficiency? 5. explaining 401(k) momentum trader inefficiency given the meaningful inefficiencies identified among 401(k) momentum traders, the next step is to locate the causes of the inefficiency. at the participant level, it could be because of traders’ lack of fund selection ability. previous studies have indicated that the stock return momentum phenomenon documented by jegadeesh and titman (1993) explains the mutual fund performance persistence (carhart, 1997; grinblatt et al., 1995). specifically, because stock returns are positively correlated in the short term, funds that consistently invest in recent winner stocks would benefit more than other funds from the effects of return momentum of underlying stocks (sapp and tiwari, 2004). such momentum investing style is the key to persistent mutual fund outperformance. consequently, if investors want to realize excess return by chasing past fund performance, it is important to select funds that take advantage of a momentum investing style. chasing recent winners might unwittingly benefits from the momentum effect if the purchased winning funds happen to invest on momentum; however, a strategy of naively chasing past performance while lacking the ability to identify funds with momentum styles cannot guarantee the success of momentum trading (sapp and tiwari, 2004). sapp and tiwari (2004) have indicated that investors of u.s. mutual funds lack fund selection ability; they do not select funds based on a momentum investing style, but rather simply chase funds that were recent winners. if the same pattern holds true among 401(k) traders, this could be one explanation for the inefficiency among 401(k) momentum traders. in this section, we will test whether momentum traders in 401(k) plans have the ability to identify funds with momentum styles or whether they just naively chase past fund performance. 5.1. determinants of fund cash flows if investors have the ability to identify funds with momentum styles, we would expect fund momentum loadings to significantly affect fund cash flows. however, if investors simply chase past fund performance, we would expect lagged fund returns to be the primary determinant of fund cash flows. we follow sapp and tiwari (2004) to examine the explanatory power of momentum factor loading and lagged fund return for fund cash flows. to do so, we first estimate individual funds’ momentum factor loadings each month by the following four-factor model: 65n. tang / financial services review 25 (2016) 51–72 ri,t � rf,t � �i � �1�rm,t � rf,t� � �2smbt � �3hmlt � �4umdt � �i,t (7) where ri,t is the monthly return of fund i; rf,t is the monthly risk-free asset return; rm,t is the market return in month t. smb, hml, and umd are returns on zero-investment factor-mimicking portfolios for size, book-to-market, and one-year momentum in stock returns. therefore, �4 indicates the estimated momentum factor loading of fund i. we then run the regression to investigate the determinants of fund cash flows. specifically, in the quarter when a trader is identified as a momentum trader, we normalize the monthly cash flows of each fund purchased (sold) by the momentum buyer (seller) by his account balance at the beginning of the month. we then take the monthly average of normalized cash flows of each fund across all traders and use the averages as dependent variables. independent variables in the regression include fund’s umd loading and funds’ past month returns. we choose past month return because momentum traders are identified based on their responses to past month fund return. regressions also control for time fixed effects. model i in table 5 uses a fund’s lagged return and umd loading as explanatory variables. results show that umd loading has no significant impact on momentum traders’ cash flows. at the same time, past fund returns are significantly positively related to cash flows (p�.10 among momentum buyers; p�.01 among momentum sellers). model ii includes additional explanatory variables in the regression, namely logarithm of funds total assets at the beginning of the month, and normalized fund cash flows in the prior month. larger funds presumably have greater visibility and thus are expected to have larger transaction amounts involved. results confirm that larger funds have more cash outflows. the positive coefficient on previous month’s fund cash flow among momentum sellers suggests that cash outflow tends to be persistent, probably because of fund reputation or visibility. results from model table 5 determinants of fund cash flows explanatory variables momentum buyer momentum seller i ii i ii intercept 0.10*** �0.47 �0.14*** �0.08*** (6.25) (�1.36) (�17.60) (�8.43) past month fund return 0.50 0.88 0.06*** 0.11*** (1.80) (1.83) (3.08) (5.92) umd loading 0.01 0.002 0.001 0.0003 (0.55) (0.23) (0.18) (0.12) logarithm of fund size 0.03 �0.003*** (1.71) (�8.30) previous month’s fund cash flow �0.0005 0.08*** (�0.53) (8.03) note: the table shows the coefficients from regressions of the determinants of momentum buyers’ and momentum sellers’ normalized fund cash flows. dependents variables are average monthly cash flows of funds purchased (sold) by momentum buyers (sellers) normalized by individual traders’ account balance at the beginning of the month. independent variables include past month fund return, fund’s umd loading estimated from four-factor model, logarithm of the funds total assets at the beginning of the month, and normalized fund cash flow in the previous month. regressions also control for time fixed effects. t-statistics are reported in parenthesis. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. 66 n. tang / financial services review 25 (2016) 51–72 ii also confirm that cash flows of momentum traders are primarily influenced by past returns rather than by fund momentum exposures (coefficient on past month fund return among momentum buyers in model ii is significant at p�.10). in summary, evidence from table 5 suggests that momentum traders in 401(k) plans lack the ability to identify funds with momentum styles. they just naively chase past fund performance. 5.2. relationship between fund cash flows and fund momentum exposures the above findings confirm that past fund returns, instead of fund momentum exposures, are primary determinants of momentum traders’ cash flows. to further explore this issue, we follow sapp and tiwari (2004) and examine whether 401(k) momentum traders persistently trade funds with high momentum exposures. we rank the funds within a plan into quartiles based on funds’ momentum factor loadings as calculated above. we then calculate for each momentum quartile the proportion of funds purchased by momentum buyers and sold by momentum sellers in the quarter they are identified as momentum traders. as shown in table 6, momentum traders do not appear to deliberately pursue a strategy of investing in momentum-style funds. for example, only 32.01% of the funds purchased by the momentum buyers are the “momentum funds” that have the highest momentum exposures, whereas 22.03% of the funds they purchase belong to the lowest momentum factor quartile. the above results indicate that momentum traders in 401(k) plans naïvely chase funds according to funds’ past returns rather than successfully identifying funds that follow a momentum style. these traders do not follow a deliberate strategy of selectively investing in momentum funds. such a naïve momentum strategy hinders the profitability of momentum investing. 6. discussions empirical results on investor trading performance are mixed in the literature. there is evidence indicating “smart money” effects among investors. gruber (1996) and zheng table 6 percentage of funds traded by momentum traders based on momentum factor rankings (in %) momentum factor loading ranks momentum buyer momentum seller mean sd mean sd 0�1st quartile 22.03 30.03 26.42 31.64 1st-2nd quartile 25.29 32.99 23.64 31.55 2nd-3rd quartile 20.67 30.35 19.40 29.20 3rd-4th quartile 32.01 33.87 30.54 32.84 note: momentum factor loading for each fund is estimated by the four-factor model. funds are then ranked within a plan into quartiles based on funds’ momentum factor loadings. the table reports for each momentum factor loading quartile, the percentage of funds purchased by momentum buyers and sold by momentum seller. 67n. tang / financial services review 25 (2016) 51–72 (1999) show that the short-term performance of funds that experience positive cash flow is significantly better than those experiencing negative cash flow, suggesting that mutual fund investors have selection ability and invest accordingly. subsequent work by sapp and tiwari (2004) confirms the smart money effect using the complete universe of u.s. equity mutual funds for the period 1970 through 2000, and it further shows that stock return momentum phenomenon explains the smart money effect. on the other hand, other studies indicate the existence of “dumb money” effects. for example, frazzini and lamont (2008) find that retail investors direct their money to funds that invest in stocks with high past returns (momentum trading). however, by overweighting growth stocks and selecting securities that on average underperform their growth benchmarks, investors end up with stocks having low future returns. mutual fund investors are “dumb” in the sense that they lose money from their reallocations. they fail to utilize momentum in a systematic way as documented in jegadeesh and tittman (1993). before we draw the conclusion that our results contradict the “smart money” effects, we would like to check if we could replicate the smart money effects using zheng (1999) methodology. in particular, zheng (1999) aggregates trading data across individuals. the rich dataset we adopted allows us to study trading behaviors at the individual investor level. as robustness check, we now aggregate fund flows of momentum buyers and sellers. we follow zheng (1999) to construct portfolios at the beginning of each quarter based on the sign of new money in the preceding quarter. portfolio 1: in all funds with positive new money cash flows by momentum buyers and weighted by funds’ new money. portfolio 2: in all funds with negative new money cash flows by momentum sellers and weighted by funds’ new money. these portfolios correspond to portfolios 5 to 6 in zheng (1999). we then follow the portfolios for three months and calculate their monthly returns as zheng (1999) did. as shown in table 7, there is no sign of smart money effects even after we aggregated the cash flows of momentum buyers and sellers. thus, our results are in line with the literature on the “dumb money” effects. we consider several factors that could contribute to the contrast between our findings and prior literature on “smart money” effects. first, portfolios under analysis here differ from prior studies. this article investigates the performance of momentum traders exclusively; whereas gruber (1996) and zheng (1999) focus on trading strategies by all types of traders. the “dumb money” effects among momentum traders found in this article do not exclude the possibility that investors could make “smart money” through other trading strategies. second, we study a different group of investors. as pointed out in the introduction, investors may exhibit different behavior when trading in and outside retirement accounts. given that 401(k) plan participants may have less resource necessary to implement a mechanical funds selection rule, and they may pay less attention to their account performance because of inertial and other behavioral biases than active traders outside retirement accounts, it is reasonable to expect different performance outcomes (see, e.g., agnew et al., 2003; ameriks and zeldes, 2004). third, our sample period covers 2005 through 2010. although we have examined the impact of financial crisis and investment opportunity constraints on momentum 68 n. tang / financial services review 25 (2016) 51–72 traders’ performances, caution should be taken to imply the results to other years. we invite further study to examine momentum trading performance during other sample periods. 7. conclusions this article explores the existence and performance of momentum traders in 401(k) plans. our unique dataset and our use of the binomial test make it possible to distinguish momentum traders from random traders in each quarter. using various measures of risk-adjusted returns with different holding periods, we find that momentum traders in 401(k) plans do not improve their portfolio performance. instead, they could lose up to 2.14% per year. we further confirmed our conclusions by excluding inconsistent momentum traders from the sample. in seeking to explain such inefficiency, we find that 401(k) traders follow a naïve momentum strategy. that is, they do not have the ability to select funds with momentum investing styles but, instead, simply chase past returns. the substantial growth of 401(k) pensions in the american workplace in recent decades has generated much interest in how well these retirement schemes are managed and how plan table 7 performance of new money portfolios (aggregated fund cash flows) portfolios (a) excess return (b) alpha1portfolio regression approach (c) alpha3portfolio regression approach (d) alpha1-fund regression approach (e) alpha 3fund regression approach 1. positive cash flow �0.003% �0.01% �0.01% �0.002% 0.02% (�0.03) (�0.06) (�0.06) (�0.02) (0.19) 2. negative cash flow 0.12% 0.12% 0.11% 0.18% 0.17% (1.13) (1.09) (1.09) (1.56) (1.60) portfolio 1 portfolio 1 �0.12% �0.13% �0.12% �0.18%** �0.15%** (�1.40) (�1.38) (�1.33) (�2.21) (�2.28) note: three portfolios are constructed based on 876,941 momentum buyers and 624,307 momentum sellers identified in response to past month fund return: (1) positive cash flow portfolio: purchase all funds with positive new money cash flows by momentum buyers in the quarter they are identified as momentum buyers, weighted by funds’ new money; (2) negative cash flow portfolio: purchase all funds with negative new money cash flows by momentum sellers in the quarter they are identified as momentum sellers, weighted by funds’ new money; (3) portfolio 1portfolio 2: long positive cash flow portfolio and short negative cash flow portfolio. the holding period for each portfolio is three months. excess return is calculated as the different between portfolio raw return and market return; cross-sectional average of excess returns among all momentum buyers (sellers) are first calculated each month and time-series mean and t-statistics are shown in column (a); to calculate alpha1 and alpha3 under portfolio regression approach, we first obtain time-series raw returns by averaging portfolio raw returns across all momentum buyers (sellers) each month and risk-adjusted returns from capm and fama-french three-factor models are shown in column (b) and (c) respectively; column (d) and (e) show the risk-adjusted returns under fund regression approach; alphas for individual momentum buyers (sellers) portfolios are first calculated using capm and fama-french three-factor models and we take the average of alphas across all momentum buyers (sellers) each month and show the time-series means and t-statistics. t-statistics are reported in parenthesis. *** and ** indicate statistical significance at the 1% and 5% levels, respectively. 69n. tang / financial services review 25 (2016) 51–72 sponsors can help participants better manage their retirement savings. previous studies indicate that americans are not adequately planning for their retirement futures (willett, 2008). academic researchers have indicated some behavioral biases and financial literacy constraints that hurt 401(k) investment efficiency. while previous studies have investigated 401(k) decisions on contribution and asset allocation, we locate another source of 401(k) management inefficiency—loss caused by momentum trading. pension plan participants seem to lack sufficient trading knowledge and fund selection ability to benefit from momentum trading in 401(k) plans. they adopt momentum strategies irrationally. as a consequence, momentum traders lose from trading in 401(k) plans. in addition, inefficient momentum traders may exert negative impacts on long-term investors in the fund. the withdrawal of momentum traders after experiencing negative returns may force open-end funds to sell assets at depressed values to meet redemption requests, which could potentially cause losses to other investors in the fund. one way to rectify errors produced by counterproductive trading would be to improve financial education and investment literacy levels (dolvin and templeton, 2006). in addition, helpful “nudges” by employers are suggested (thaler and sunstein, 2008). as pointed out by thaler and sunstein (2008), well-designed choice architecture could steer people’s choices in directions that will improve their lives. default options are considered helpful to combat irrational investment behaviors. for example, the 2006 pension protection act authorized a series of “qualified default investment alternatives” (qdias). target-date funds (tdfs), one of the alternatives, are funds diversified across stocks and bonds that automatically rebalance toward lower risk investments as participants approach retirement. we expect that such a design will help avoid losses caused by 401(k) participants’ behavior biases in portfolio choices and trading. in addition, pension designers may want to mitigate the impact of counterproductive momentum trading by muting naïve momentum trading itself, perhaps by making short-term performance less salient to investors, especially to those who experience balance shocks or portfolio losses. notes 1 we count a buy or sell transaction on a daily basis as one trade. to calculate individual turnover, we follow annew et al. (2003). we sum the absolute values of trading amount in one year and divide this sum by two. we then divide the annual trading amount by the account balance at the beginning of the year. 2 the majority of 401(k) investment assets under analysis are mutual funds. 3 the total number of buyers and sellers is more than the total number of traders during the sample period because a trader could both buy and sell. in that case, he will be considered as one observation in buyers sample, one observation in sellers sample, and one, instead of two, observation in traders sample. 4 if the fund has fewer than 30 past return observations in month t, traders who purchased or sold that fund in month t are excluded from the sample. losing observations is the disadvantage of the “fund regression” approach compared with the “portfolio regression” approach. 70 n. tang / financial services review 25 (2016) 51–72 references agnew, j. (2002). inefficient choices in 401(k) plans: evidence from individual level data. presented at the 4th annual joint conference for the retirement research consortium ‘directions for social security reform.’ may 2002, washington, d.c. agnew, j., balduzzi, p., & sunden, a. (2003). portfolio choice and trading in a large 401(k) plan. american economic review, 93, 193–215. ameriks, j., & zeldes, s. p. (2004). how do household portfolio shares vary with age? tiaa-cref working paper. barber, b., & odean, t. (2000). trading is hazardous to your wealth: the common stock investment performance of individual investors. journal of finance, 55, 773–806. benartzi, s., thaler, & r. h. (2001). naïve diversification strategies in defined contribution savings plans. american economic review, 91, 79–98. benartzi, s., thaler, r. h., utkus, s. p., & sunstein, c. r. (2007). the law and economics of company stock in 401(k) plans. journal of law and economics, 50, 45–79. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. choi, c. j., laibson, d., & madrian, b. d. (2007). mental accounting in portfolio choice: evidence from a flypaper effect. nber working paper 13656. dolvin, s., & templeton, w. (2006). financial education and asset allocation. financial services review, 15, 133–149. even, w. e., & macpherson, d. a. (2008). pension investments in employer stock. journal of pension economics and finance, 7, 67–93. frazzini, a., & lamont, o. a. (2008). dumb money: mutual fund flows and the cross-section of stock returns. journal of financial economics, 88, 299–322. goetzmann, w. n., & ibbotson, r. g. (1994). do winners repeat? journal of portfolio management, 20, 9–18. goetzmann, w., & massa, m. (2002). daily momentum and contrarian behavior of index fund investors. journal of financial and quantitative analysis, 37, 375–389. grinblatt, m., & keloharju, m. (2000). the investment behavior and performance of various investor types: a study of finland’s unique data set. journal of financial economics. 55, 43–67. grinblatt, m., & keloharju m. (2001). what makes investors trade? journal of finance, 56, 589–616. grinblatt, m., & titman, s. (1992). the persistence of mutual fund performance. journal of finance, 47, 1977–1984. grinblatt, m., titman, s., & wermers, r. (1995). momentum investment strategies, portfolio performance, and herding: a study of mutual fund behavior. american economic review, 85, 1088–1105. gruber, m. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51, 783–810. hendricks, d., patel, j., & zechhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974–1988. journal of finance, 48, 93–130. huberman, g., & sengmueller, p. (2004). performance and employer stock in 401(k) plans. review of finance, 8, 403–443. jegadeesh, n., & titman, s. (1993). returns to buying winners and selling losers: implications for stock market efficiency. journal of finance, 48, 65–91. jegadeesh, n., & titman, s. (2001). profitability of momentum strategies: an evaluation of alternative explanations. journal of finance, 56, 699–720. kim, k. t., & hanna, s. d. (2015). do u.s. households perceive their retirement preparedness realistically? financial services review, 24, 139–155. liang, n., & weisbenner, s. (2002). investor behavior and the purchase of company stock in 401(k) plans: the importance of plan design. finance and economics discussion series 2002–36. washington, dc: board of governors of the federal reserve system. pettengill, g. n., edwards, s. m., & schmitt, d. e. (2006). is momentum investing a viable strategy for individual investors? financial services review, 15, 181–197. 71n. tang / financial services review 25 (2016) 51–72 pettengill, g. n., edwards, s. m., & griggs, f. t. (2009). can individual investors duplicate professional momentum investing? financial services review, 18, 355–380. rockenbach, b. (2004). the behavioral relevance of mental accounting for the pricing of financial options. journal of economic behavior & organization, 53, 513–527. rouwenhorst, k. g. (1998). international momentum strategies. journal of finance, 53, 267–284. sapp, t., & tiwari, a. (2004). does stock return momentum explain the “smart money” effect? journal of finance, 59, 2605–2622. sirri, e., & tufano, p. (1998). costly search and mutual fund flows. journal of finance, 53, 1589–1622. solomon, d. h., solges, e., & sosyura, d. (2014). winners in the spotlight: media coverage of fund holdings as a driver of flows. journal of financial economics, 113, 53–72. tang, n., mitchell, o. s., & utkus, s. p. (2012). trading in 401(k) plans during the financial crisis. in: r. maurer, o. s. mitchell, & m. j. warshawsky (eds.), reshaping retirement security: lessons from the global financial crisis. oxford: oxford university press. tang, n., mitchell, o. s., mottola, g. r., & utkus, s. p. (2010). the efficiency of sponsor and participant portfolio choices in 401(k) plans. journal of public economics, 94, 1973–1085. thaler, r. h., & sunstein, c. r. (2008). nudge: improving decisions about health, wealth, and happiness. new haven, ct: yale university press. vanderhei, j., holden, s., alonso, l., & bass, s. (2011). 401(k) plan asset allocation, account balances, and loan activity in 2010. ebri issue brief #366. warther, v. (1995). aggregate mutual fund flows and security returns. journal of financial economics, 39, 581–562. willett, m. (2008). a new model for retirement education and counseling. financial services review, 17, 105–130. zheng, l. (1999). is money smart? a study of mutual fund investors’ fund selection ability. journal of finance, 54, 901–933. 72 n. tang / financial services review 25 (2016) 51–72 does it pay to diversify? u.s. vs. international etfs srinidhi kanuria, robert w. mcleodb,* adepartment of finance real estate and business law, college of business, the university of southern mississippi, scianna hall, 118 college drive, #5076, hattiesburg, ms 39406, usa bdepartment of economics finance and legal studies, culverhouse college of commerce, the university of alabama, box 870224, tucaloosa, al 35487-0224, usa abstract individual investors seek diversification in their portfolios using a number of approaches. one approach that is commonly used is to diversify globally. this article evaluates the performance and diversification benefits of international etfs for u.s. investors during and after the recent financial crisis. our results show that u.s. etfs outperform all categories of international etfs for the period of our study (january 2008 – june 2013); they have higher average monthly returns, lower risk (standard deviation of returns), higher risk-adjusted performance (sharpe, sortino, and treynor ratios) and the highest cumulative returns over the entire period. when we form equally weighted portfolios of each etf category and compute their risk-adjusted performance, we again find that u.s. etf portfolios had the best performance for the entire period. we also find that u.s. etfs have the lowest tracking error during the entire period. most of these etfs passively track the benchmark and do not manage for positive �. previous research has questioned the diversification benefits of international investing during times of financial distress. we find that international etfs are highly dependent on major u.s. indices during the period of our analysis, and therefore, offered limited diversification benefits for u.s. investors. © 2015 academy of financial services. all rights reserved. jel classification: g11; g12; g15 keywords: international etfs; diversification; portfolio; risk-adjusted performance * corresponding author. tel.: �1-205-448-8993; fax: �1-205-348-0590. e-mail address: rmcleod@cba.ua.edu. financial services review 24 (2015) 249–270 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. 1. introduction u. s. investors can achieve global diversification in a number of ways. they can purchase individual securities directly in foreign capital markets or in u.s. markets through american depository receipts (adrs). other investors attain international diversification by indirect investments such as mutual funds, closed-end funds, or exchange traded funds (etfs). earlier research (e.g., adler and dumas, 1983; black, 1974; heston and rouwenhorst,1994; levy and sarnat, 1970; stulz, 1981) supports the importance to investors to allocate some of their funds into foreign investments as a means of reducing portfolio risk because of low correlations among markets or market segmentation that results in barriers to international investment. however, more recent findings bring into question the diversification benefits of international investing especially during times of financial distress. eun and shin (1989), king and wadhwani (1990), and koch and koch (1991) show that regional dependencies have increased over time. longin and solnik (1995) and jacquier and marcus (2001) examine correlations in country portfolio returns during turbulent market conditions and conclude that they increase. roll (1987) analyzed the crash of october 1987 and reports that all 23 indexes studied declined in a synchronized fashion. pries, kenett, stanley, helbing, and ben-jacob (2012) report that diversification benefits vanish during times of financial distress. russell (1998) looks at the international diversification benefits of u.s. exchange traded securities such as closed ended funds, adrs and multinational corporation (mncs) to provide diversification benefits similar to investment in foreign equity. the result indicate that u.s. exchange-listed securities behave more like host exchange than their home exchange. this results suggests exchange-listed securities on average, do not perform an international diversification role for u.s. investors. johnson et al. (1999) find that diversification benefits of international mutual funds may be less than what previous studies find. they find that during restrictive u.s. monetary policy periods, international mutual fund indexes provide lower excess returns than domestic counterparts. additionally, the correlations between international mutual funds and domestic mutual funds are higher during restrictive monetary policy periods. this evidence may represent a partial explanation for the home country bias exhibited by united states-based individual and institutional investors. aiello and chieffe (1999) compare the performance of international index funds and the s&p 500 from 1989 to 1997 and find that international index funds do not offer superior performance. ho et al. (1999) find that the united states equity market is a large proportion of the international equity market that is available to individual investors, and united states returns are highly correlated with other markets. hanna et al. (1999) look at ten years of historical data (january 1988 through december 1997) from the stock markets in the g-7 countries. across this 10-year period, they find that a portfolio consisting solely of the s&p 500 dominates any portfolio that can be constructed from the s&p 500 and the major market index of the g-7 countries. the growth in the number of international etfs has been significant, especially immediately before the “great recession.” this growth was in response to investor demand for etfs that provided an opportunity to diversify globally using low cost options. in this article we look at the performance and diversification benefits of international etfs for u.s. investors from 2008 through june 2013.1 using etfs that follow total world, total world 250 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 ex u.s., developed markets, and emerging markets, we compare their performance to u.s. etfs that follow the major indices. according to rompotis (2010), investors choose foreign etfs for a number of reasons: (1) it was difficult to invest in securities listed on foreign exchanges before the emergence of etfs because many u.s. brokers were unable to process orders on non-u.s. exchanges and the few good international mutual funds that existed had very high expenses. (2) apart from difficulties of foreign investing, investors choose international etfs for broad diversification without having to directly purchase stocks in foreign countries. in addition, investors want to take advantage of specific macroeconomic or microeconomic trends, such as rapid growth in a particular economy or region. international etfs give u.s. investors a cheaper and less complicated method to invest in foreign stocks rather than direct investment. our article will test whether international etfs outperform major u.s. etfs and whether u.s. investors get diversification benefits by investing in foreign etfs during the financial crisis and subsequent recovery of the u.s. stock market. 2. data to be included in the analysis (for an equal comparison), the etf should have been created on or before january 2008 and have continuous return and trading history from january 2008 through june 2013. these etfs where created at different points of time with u.s. etfs created before total world, total world ex u.s., developed markets, and emerging markets etfs. table 1a provides a list of each etf used in our study, its inception date, the underlying benchmark index, and the category to which the etf belongs. the complete list of different categories of etfs was obtained from morningstar direct. using data obtained from morningstar direct and bloomberg terminal we compare the performance of u.s. etfs following six major u.s. indices (s&p 500, russell 1000, russell 3000, dow jones industrial average, dow jones u.s. total returns, and nasdaq 100) to total world, total world ex u.s., emerging markets, and developed markets etfs. there are a total of 36 etfs (6 u.s., 6 total world, 4 total world ex u.s., 10 emerging markets, and 10 developed markets) included in this analysis. we also look at the potential benefits of diversification for american investors from owning international etfs. descriptive data are provided in table 1b that include the average annual net expense and turnover ratios of each etf from 2008 to 2013 and assets at the end of june 2013. u.s. etfs have the lowest expense ratios while developed market and emerging market etfs have the highest expense ratios. all these values have also been taken from morningstar direct and bloomberg terminal. 3. performance and risk following rompotis (2009, 2010) and shin and soydemir (2010), etfs are compared based on their average monthly returns for the entire period (january 2008 through june 2013). we rank etfs in descending order based on their average returns. the risk of etfs 251s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 is estimated as the standard deviation of returns. as shown in table 2, on average, u.s. etfs and indices have the highest average monthly returns, whereas emerging (adre, eeb, and bkf) markets etfs and benchmarks have the lowest returns over the entire period. similarly, u.s. etfs and indices (with the exception of qqq and its benchmark nasdaq 100) have the lowest standard deviation of returns over the entire period, whereas emerging markets etfs have the highest standard deviation of returns over the entire period. 3.1. sharpe, sortino, and treynor ratios an etf could have higher returns, but it could have done so by assuming higher risk. to compare risk adjusted returns of etfs over the same period, sharpe, sortino, and treynor ratios are used. these measures have been widely used in the literature (e.g., harper, madura, and schnusenberg, 2006; rompotis, 2009, 2010) to compare etf performance. etfs are ranked in descending order (from best to worst based on these ratios) for the entire period (january 2008 through june 2013). table 1a: this table shows the etf, the benchmark it follows, the inception date and category to which it belongs etf name benchmark inception category ivv ishares core s&p 500 etf s&p 500 tr usd 5/15/2000 u.s. iwb ishares russell 1000 index russell 1000 tr usd 5/15/2000 u.s. iwv ishares russell 3000 index russell 3000 tr usd 5/22/2000 u.s. dia spdr dow jones industrial average dj industrial average tr usd 1/13/1998 u.s. iyy ishares dow jones u.s. index dj us tr usd 6/12/2000 u.s. qqq powershares qqq nasdaq 100 tr usd 3/10/1999 u.s. dew wisdomtree global equity income wisdomtree global equity income tr usd 6/16/2006 total world dgt spdr global dow etf dj global tr usd 9/25/2000 total world ioo ishares s&p global 100 index s&p global 100 tr 12/5/2000 total world fgd first trust dj global select dividend dj global select dividend tr usd 11/21/2007 total world lvl guggenheim s&p global dividend opps idx s&p global dividend opport nr usd 6/25/2007 total world tok ishares msci kokusai msci kokusai tr usd 12/10/2007 total world cwi spdr msci acwi (ex-us) msci acwi ex usa gr usd 1/10/2007 total world ex us dnl wisdomtree global ex-us growth wisdomtree global ex us growth tr usd 6/16/2006 total world ex u.s. gwl spdr s&p world ex-us s&p developed ex us bmi tr usd 4/20/2007 total world ex u.s. veu vanguard ftse all-world ex-us etf ftse aw ex us tr usd 3/2/2007 total world ex u.s. adrd bldrs developed markets 100 adr index bony developed markets 100 adr tr usd 11/13/2002 developed dol wisdomtree international largecap div wisdomtree intl largecap dividend tr usd 6/16/2006 developed doo wisdomtree international div ex-finncls wisdomtree intl dividend ex fincl tr usd 6/16/2006 developed dth wisdomtree defa equity income wisdomtree defa equity income tr usd 6/16/2006 developed dwm wisdomtree defa wisdomtree defa tr usd 6/16/2006 developed efa ishares msci eafe msci eafe nr usd 8/14/2001 developed idv ishares dow jones intl select div idx dj epac select dividend tr usd 6/11/2007 developed piz powershares dwa dev mkts technical ldrs dorsey wright dev mrkt tech ldrs nr usd 12/28/2007 developed pxf powershares ftse rafi dev mkts ex-us ftse rafi dvlp ex us 1000 tr usd 6/25/2007 developed vea vanguard ftse developed markets etf ftse developed ex north america nr usd 7/20/2007 developed adre bldrs emerging markets 50 adr index bony emerging markets 50 adr tr usd 11/13/2002 emerging bik spdr s&p bric 40 s&p bric 40 tr 6/19/2007 emerging bkf ishares msci bric msci bric nr usd 11/12/2007 emerging dem wisdomtree emerging markets equity inc wisdomtree em equity income tr usd 7/13/2007 emerging eeb guggenheim bric bny/mellon bric tr usd 9/21/2006 emerging eem ishares msci emerging markets msci em nr usd 4/7/2003 emerging gmm spdr s&p emerging markets s&p emerging bmi tr usd 3/19/2007 emerging pie powershares dwa em mkts technical ldrs dorsey wright em mrkt tech ldrs nr usd 12/28/2007 emerging pxh powershares ftse rafi emerging markets ftse rafi emerging tr usd 9/27/2007 emerging vwo vanguard ftse emerging markets etf ftse emerging tr usd 3/4/2005 emerging 252 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 sharpe ratio is calculated as: sr � �retf � rf�/�etf (1) where retf denotes the monthly returns on the etf, rf is the monthly risk free rate, �etf is the standard deviation of monthly etf returns. table 1b: shows average annual net expense and turnover ratios from 2008 to 2013 and assets at the end of june 2013 etf average annual net expense ratio 2008–2013 average annual turnover ratio 2008–2013 assets $ (june 2013) category ivv 0.09% 5.33% 42,573,234,908 u.s. iwb 0.15% 6.67% 7,686,737,763 u.s. iwv 0.20% 6.33% 4,287,062,577 u.s. iyy 0.20% 5.67% 715,567,400 u.s. dia 0.17% 6.72% 12,568,816,144 u.s. qqq 0.20% 12.19% 33,645,121,238 u.s. dew 0.53% 42.33% 98,284,717 total world dgt 0.51% 27.17% 88,219,982 total world ioo 0.40% 5.67% 1,256,399,204 total world fgd 0.60% 36.83% 311,450,253 total world lvl 0.80% 82.67% 72,989,199 total world tok 0.25% 5.67% 576,636,913 total world cwi 0.34% 6.17% 416,858,949 total world ex u.s. dnl 0.58% 54.00% 76,210,727 total world ex u.s. gwl 0.35% 5.83% 572,349,639 total world ex u.s. veu 0.19% 7.00% 9,041,860,844 total world ex u.s. adrd 0.28% 10.10% 45,196,595 developed dol 0.48% 22.50% 209,073,220 developed doo 0.58% 46.33% 322,821,722 developed dth 0.58% 31.50% 212,323,318 developed dwm 0.43% 35.67% 438,785,622 developed efa 0.34% 6.33% 40,201,315,518 developed idv 0.50% 42.67% 1,969,620,635 developed piz 0.80% 128.17% 246,498,810 developed pxf 0.71% 21.83% 536,149,000 developed vea 0.12% 8.60% 13,152,279,035 developed adre 0.27% 12.04% 235,859,904 emerging bik 0.49% 12.33% 226,925,818 emerging bkf 0.34% 13.33% 491,577,033 emerging dem 0.63% 38.50% 4,848,791,521 emerging eeb 0.63% 11.00% 224,304,688 emerging eem 0.69% 14.33% 34,620,518,926 emerging gmm 0.59% 10.50% 180,418,744 emerging pie 0.90% 171.00% 354,721,563 emerging pxh 0.80% 36.50% 333,461,873 emerging vwo 0.20% 14.67% 49,355,444,391 emerging 253s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 the sharpe ratio evaluates how well an etf compensates its investor for each unit of risk they incur. the higher the sharpe ratio, the better is the performance of the etf. the second measure of risk-adjusted performance is the sortino ratio expressed as: sortino � �retf � rf�/�d (2) where retf and rf are described as above; �d is the standard deviation of etf’s negative returns. table 2 shows average monthly returns and standard deviation of returns (in %) from january 2008 through june 2013 rank etf no. of obs avg. monthly etf return etf sd avg. monthly index return index sd category 1 qqq 66 months 0.7561% 6.1599% 0.7705% 6.1401% us 2 dia 66 months 0.5148% 4.8122% 0.5279% 4.8237% us 3 iyy 66 months 0.4949% 5.4340% 0.5107% 5.4480% us 4 iwv 66 months 0.4949% 5.4871% 0.5059% 5.5027% us 5 iwb 66 months 0.4810% 5.4020% 0.4896% 5.4150% us 6 ivv 66 months 0.4590% 5.2919% 0.4633% 5.3017% us 7 dem 66 months 0.4407% 6.7076% 0.5276% 6.7266% emerging 8 fgd 66 months 0.2923% 7.1894% 0.2776% 7.0462% total world 9 tok 66 months 0.2521% 5.9550% 0.2301% 5.9868% total world 10 dnl 66 months 0.2345% 5.5904% 0.2910% 5.5820% total world ex us 11 idv 66 months 0.1843% 7.6042% 0.1920% 7.7109% developed 12 piz 66 months 0.1534% 7.3502% 0.2334% 7.3026% developed 13 gmm 66 months 0.1441% 7.8320% 0.1644% 8.0012% emerging 14 ioo 66 months 0.1387% 5.6871% �0.2219% 6.8277% total world 15 vwo 66 months 0.0989% 8.1487% 0.1627% 8.0327% emerging 16 lvl 66 months 0.0824% 7.5689% �0.0142% 7.4818% total world 17 eem 66 months 0.0765% 7.9182% 0.1079% 8.0415% emerging 18 veu 66 months 0.0473% 6.9711% 0.0683% 6.8047% total world ex us 19 vea 66 months 0.0311% 6.7140% 0.0172% 6.6193% developed 20 gwl 66 months 0.0252% 6.4864% 0.0837% 6.6414% total world ex us 21 cwi 66 months 0.0236% 6.6716% 0.0482% 6.7425% total world ex us 22 pxf 66 months 0.0137% 7.4310% 0.1014% 7.3730% developed 23 pxh 66 months 0.0077% 7.9612% 0.1705% 8.0847% emerging 24 dgt 66 months �0.0661% 5.5120% 0.2421% 6.0017% total world 25 dew 66 months �0.0166% 6.5695% 0.0027% 6.5988% total world 26 efa 66 months �0.0112% 6.5123% �0.0064% 6.5437% developed 27 pie 66 months �0.0389% 8.4288% 0.3117% 8.1988% emerging 28 adrd 66 months �0.0451% 6.7941% �0.0625% 6.8151% developed 29 dwm 66 months �0.0617% 6.4940% �0.0130% 6.5679% developed 30 dth 66 months �0.0930% 6.8944% �0.0602% 6.9748% developed 31 dol 66 months �0.0943% 6.4759% �0.0814% 6.5380% developed 32 bik 66 months �0.0999% 8.5527% �0.0542% 8.5939% emerging 33 doo 66 months �0.1272% 6.9165% �0.1335% 6.9606% developed 34 adre 66 months �0.1960% 7.7663% �0.1874% 7.7789% emerging 35 eeb 66 months �0.2289% 8.7484% �0.1896% 8.7991% emerging 36 bkf 66 months �0.2402% 8.9804% �0.2082% 8.9795% emerging etfs are ranked in descending order based on average monthly returns. 254 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 the sortino ratio differentiates between good and bad volatility in the sharpe ratio. the differentiation of upward and downward volatility allows the calculation of the risk-adjusted return to provide a performance measure of an etf without penalizing it for positive returns. a large sortino ratio indicates low risk of large losses occurring. similar to the sharpe ratio, the higher the sortino ratio, the better is the performance of an etf. the third measure we use is the treynor ratio that is expressed as: treynor � �retf � rf�/�etf (3) where retf and rf are defined as above, �etf is the systematic risk of the etf. similarly to sharpe and sortino ratios, the higher the treynor ratio, the better is the performance of the etf. the results shown in table 3a indicate again that u.s. etfs have the highest sharpe and sortino ratios (the first six ranks are occupied by u.s. etfs with qqq and dia having the best performance out of all etfs), whereas emerging (adre, bke, and eeb) and developed (dth, doo, and dol) market etfs have the lowest sharpe and sortino ratios. the treynor ratio again indicates that u.s. etfs had the best performance for the entire period as shown in table 3b. as a robustness test, sharpe, sortino, and treynor ratios were computed using the three month interbank libor rate instead of three month t-bill rate as many of these etfs buy international stocks. the results as shown in tables 3a and b did not change (u.s. etfs again occupied the first six ranks). 4. cumulative returns and cumulative wealth index cumulative returns of etfs for the entire period have been computed. following woolridge (2004) we also compute the cumulative wealth index (cwi) for each etf. the cwi measures the outcome of investing $1,000 in each etf at the beginning of january 2008, presuming reinvestment of dividends. etfs are ranked in descending order based on cumulative returns and cwi. u.s. etfs occupy the top six ranks as shown in table 4. for example, $1,000 invested in qqq and dia in january 2008 would have returned $1,451.14 and $1,300.23 by june 2013, respectively. 5. tracking error it is important to consider tracking error when analyzing etf performance. the greater the tracking error the less closely the etf follows the benchmark. if an investor is considering using an etf for international diversification and the etf has a high tracking 255s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 error, the benefit of diversification relative to the benchmark will be lessened. tracking error is the difference in the performance of etf and its benchmark. ideally, the tracking error of an etf should be zero. however, this is not possible because of expenses; dividends payments arising from stocks of an index; as well as size and timing of index rebalancing (frino and gallagher, 2001). following frino and gallagher (2001), tracking error is measured using three different methods. te1–the first method of estimating tracking error is computed as the average absolute differences between the return on the etf and its benchmark index. the equation is given as: te1 � � t � 1 n abs �return on etf � return on the benchmark index�/n (4) te2–the second method to estimate tracking error is to use standard errors from the regression analysis using monthly returns on each etf and its benchmark index. the model is: table 3a: sharpe and sortino ratios calculated using three month t-bill and three month libor rates rank (t-bill) etf sharpe ratio sortino ratio category rank (libor) etf sharpe ratio sortino ratio category 1 qqq 0.1176 0.1684 u.s. 1 qqq 0.1163 0.1663 us 2 dia 0.1004 0.1403 u.s. 2 dia 0.0989 0.1380 us 3 iyy 0.0853 0.1175 u.s. 3 iyy 0.0841 0.1155 us 4 iwv 0.0845 0.1161 u.s. 4 iwv 0.0833 0.1142 us 5 iwb 0.0832 0.1144 u.s. 5 iwb 0.0820 0.1125 us 6 ivv 0.0808 0.1111 u.s. 6 ivv 0.0796 0.1092 us 7 tok 0.0372 0.0505 total world 7 dem 0.0603 0.0860 emerging 8 dnl 0.0365 0.0516 total world ex u.s. 8 dnl 0.0356 0.0502 total world ex us 9 fgd 0.0364 0.0514 total world 9 fgd 0.0356 0.0501 total world 10 idv 0.0202 0.0278 developed 10 tok 0.0363 0.0492 total world 11 ioo 0.0191 0.0261 total world 11 idv 0.0196 0.0268 developed 12 piz 0.0168 0.0224 developed 12 ioo 0.0182 0.0248 total world 13 gmm 0.0145 0.0201 emerging 13 piz 0.0161 0.0214 developed 14 vwo 0.0084 0.0118 emerging 14 gmm 0.0139 0.0192 emerging 15 lvl 0.0069 0.0094 total world 15 vwo 0.0078 0.0109 emerging 16 dem 0.0611 0.0874 emerging 16 lvl 0.0061 0.0083 total world 17 eem 0.0059 0.0082 emerging 17 eem 0.0053 0.0073 emerging 18 veu 0.0025 0.0034 total world ex u.s. 18 veu 0.0018 0.0024 total world ex us 19 vea 0.0002 0.0002 developed 19 vea �0.0005 �0.0007 developed 20 gwl �0.0007 �0.0010 total world ex u.s. 20 gwl �0.0015 �0.0020 total world ex us 21 cwi �0.0010 �0.0013 total world ex u.s. 21 cwi �0.0017 �0.0023 total world ex us 22 pxf �0.0022 �0.0031 developed 22 pxf �0.0028 �0.0040 developed 23 pxh �0.0028 �0.0040 emerging 23 pxh �0.0034 �0.0048 emerging 24 efa �0.0063 �0.0084 developed 24 efa �0.0070 �0.0093 developed 25 dew �0.0071 �0.0093 total world 25 dew �0.0078 �0.0102 total world 26 pie �0.0082 �0.0103 emerging 26 pie �0.0087 �0.0110 emerging 27 adrd �0.0110 �0.0151 developed 27 adrd �0.0117 �0.0159 developed 28 dwm �0.0141 �0.0187 developed 28 dwm �0.0148 �0.0195 developed 29 bik �0.0152 �0.0207 emerging 29 bik �0.0157 �0.0213 emerging 30 dgt �0.0174 �0.0231 total world 30 dgt �0.0182 �0.0240 total world 31 dth �0.0178 �0.0236 developed 31 dth �0.0184 �0.0244 developed 32 dol �0.0191 �0.0253 developed 32 dol �0.0198 �0.0261 developed 33 doo �0.0227 �0.0298 developed 33 doo �0.0233 �0.0305 developed 34 adre �0.0291 �0.0396 emerging 34 adre �0.0296 �0.0403 emerging 35 eeb �0.0295 �0.0408 emerging 35 eeb �0.0300 �0.0414 emerging 36 bkf �0.0300 �0.0410 emerging 36 bkf �0.0305 �0.0415 emerging etfs are ranked in descending order based on sharpe and sortino ratios. 256 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 etfi.t � �i � �i*bri.t � �i,t (5) where etf i.t and br i.t are monthly etf and benchmark returns, respectively. in this model the standard errors from regressions proxy tracking errors. if the etf perfectly follows its benchmark, then the standard deviation of residuals from the regression must be zero. te3–the third method estimates tracking error as the standard deviation of the return difference between an etf and its benchmark index. this method is the one that is most table 3b: treynor ratios calculated using three month t-bill and three month libor rates rank (t-bill) etf treynor ratio category rank (libor) etf treynor ratio category 1 qqq 0.7246 us 1 qqq 0.7199 us 2 dia 0.4859 us 2 dia 0.4812 us 3 iwv 0.4662 us 3 iwv 0.4615 us 4 iyy 0.4661 us 4 iyy 0.4614 us 5 iwb 0.4521 us 5 iwb 0.4474 us 6 ivv 0.4299 us 6 ivv 0.4252 us 7 dem 0.4120 emerging 7 dem 0.4073 emerging 8 fgd 0.2581 total world 8 fgd 0.2535 total world 9 tok 0.2238 total world 9 tok 0.1616 total world 10 dnl 0.2048 total world ex us 10 idv 0.1520 developed 11 idv 0.1568 developed 11 dnl 0.1430 total world ex us 12 piz 0.1228 developed 12 piz 0.1181 developed 13 gmm 0.1168 emerging 13 gmm 0.1120 emerging 14 ioo 0.1096 total world 14 ioo 0.1048 total world 15 vwo 0.0684 emerging 15 vwo 0.0638 emerging 16 lvl 0.0532 total world 16 eem 0.0426 emerging 17 eem 0.0474 emerging 17 veu 0.0124 total world ex us 18 veu 0.0170 total world ex us 18 vea �0.0036 developed 19 vea 0.0010 developed 19 cwi �0.0113 developed 20 gwl �0.0050 total world ex us 20 lvl �0.0091 total world 21 cwi �0.0066 total world ex us 21 gwl �0.0098 total world ex us 22 pxf �0.0162 developed 22 pxf �0.0209 developed 23 pxh �0.0230 emerging 23 pxh �0.0278 emerging 24 efa �0.0415 developed 24 efa �0.0462 developed 25 dew �0.0452 total world 25 dew �0.0497 total world 26 pie �0.0675 emerging 26 pie �0.0721 emerging 27 adrd �0.0753 developed 27 adrd �0.0800 developed 28 dwm �0.0928 developed 28 dwm �0.0975 developed 29 dgt �0.1080 total world 29 dgt �0.1133 total world 30 dth �0.1247 developed 30 dth �0.1294 developed 31 bik �0.1305 emerging 31 dol �0.1302 developed 32 dol �0.1255 developed 32 bik �0.1352 emerging 33 doo �0.1578 developed 33 doo �0.2201 developed 34 adre �0.2264 emerging 34 adre �0.2311 emerging 35 eeb �0.2599 emerging 35 bkf �0.2755 emerging 36 bkf �0.2708 emerging 36 eeb �0.3221 emerging etfs are ranked in descending order based on treynor ratios. 257s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 commonly used and, according to pope and yadav (1994), produces same estimates as method 1 if � in method 2 is equal to 1. te3 � �1 n � 1 � t�1 n �ri,t � rj,t� 2 (6) where table 4 shows cumulative returns and cumulative wealth index (cwi) over the entire period for each etf where the cwi measures the outcome of investing $1000 in each etf at the beginning of january 2008, presuming reinvestment of dividends rank etf cumulative returns (january 2008 through june 2013) cumulative wealth in june 2013 ($1000 invested in january 2008) category 1 qqq 45.11% $1,451.14 u.s. 2 dia 30.02% $1,300.23 u.s. 3 iyy 25.60% $1,256.05 u.s. 4 iwv 25.36% $1,253.57 u.s. 5 iwb 24.60% $1,246.02 u.s. 6 ivv 23.31% $1,233.13 u.s. 7 dem 15.19% $1,151.89 emerging 8 dnl 5.32% $1,053.16 total world ex u.s. 9 tok 4.92% $1,049.23 total world 10 fgd 2.04% $1,020.42 total world 11 ioo �1.52% $ 984.76 total world 12 idv �7.12% $ 928.77 developed 13 piz �7.73% $ 922.73 developed 14 gmm �10.53% $ 894.72 emerging 15 gwl �11.67% $ 883.34 total world ex u.s. 16 vea �12.19% $ 878.09 developed 17 veu �12.30% $ 876.98 total world ex u.s. 18 cwi �12.45% $ 875.49 total world ex u.s. 19 lvl �13.24% $ 867.60 total world 20 dgt �13.46% $ 865.40 total world 21 efa �13.88% $ 861.15 developed 22 dew �14.50% $ 854.97 total world 23 vwo �14.59% $ 854.13 emerging 24 eem �14.70% $ 853.04 emerging 25 pxf �15.87% $ 841.34 developed 26 dwm �16.64% $ 833.63 developed 27 adrd �16.75% $ 832.47 developed 28 dol �18.34% $ 816.65 developed 29 pxh �18.61% $ 813.95 emerging 30 dth �19.86% $ 801.37 developed 31 doo �21.83% $ 781.67 developed 32 pie �24.14% $ 758.59 emerging 33 bik �26.84% $ 731.56 emerging 34 adre �28.36% $ 716.44 emerging 35 eeb �33.53% $ 664.67 emerging 36 bkf �34.97% $ 650.33 emerging etfs are ranked in descending order based on cumulative returns and cwi. 258 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 r i, t and r j, t are etf and benchmark returns during month t. the total tracking error is computed as the average of te1, te2, and te3. the results shown in tables 5a and b indicate that u.s. etfs (with the exception of qqq) have the lowest te among all etfs. the results show that international etfs have high tracking errors. this result is not surprising as they face restrictions like time delays or exposure to unsafe market environments, which negatively affects their replication ability (rompotis, 2009). in addition, international etfs also have higher expenses compared with u.s. etfs, which also increases their tracking error as there is a positive relationship between expenses and tracking error (rompotis, 2009). table 5a: shows te1, te2, and te3 by etf category and the average of te1, te2, and te3 etf te1 te2 te3 average te category ivv 0.004% 0.016% 0.012% 0.011% us iwb 0.009% 0.023% 0.015% 0.016% us iwv 0.011% 0.023% 0.018% 0.017% us dia 0.013% 0.031% 0.015% 0.020% us qqq 0.014% 0.544% 0.296% 0.285% us iyy 0.016% 0.020% 0.019% 0.018% us ioo 0.361% 1.906% 3.350% 1.872% total world lvl 0.097% 2.118% 1.437% 1.217% total world tok 0.022% 0.126% 0.061% 0.070% total world fgd 0.015% 1.407% 0.678% 0.700% total world dew 0.019% 0.378% 0.199% 0.199% total world dgt 0.308% 4.015% 1.531% 1.952% total world veu 0.021% 1.881% 0.891% 0.931% total world ex us cwi 0.025% 0.880% 0.265% 0.390% total world ex us dnl 0.057% 0.393% 0.186% 0.212% total world ex us gwl 0.059% 1.055% 0.334% 0.482% total world ex us adrd 0.017% 0.222% 0.088% 0.109% developed dol 0.014% 0.465% 0.941% 0.473% developed doo 0.006% 0.388% 0.185% 0.193% developed dth 0.013% 0.507% 0.215% 0.245% developed dwm 0.033% 0.592% 0.235% 0.286% developed efa 0.049% 0.138% 0.231% 0.139% developed idv 0.005% 1.213% 0.067% 0.428% developed piz 0.008% 1.021% 0.577% 0.535% developed pxf 0.080% 1.274% 0.403% 0.586% developed vea 0.088% 2.045% 0.511% 0.881% developed adre 0.009% 0.337% 0.090% 0.145% emerging dem 0.087% 0.405% 0.193% 0.228% emerging gmm 0.020% 1.161% 0.542% 0.575% emerging eem 0.031% 1.580% 0.763% 0.791% emerging bkf 0.032% 1.035% 0.553% 0.540% emerging eeb 0.039% 0.118% 0.105% 0.087% emerging bik 0.046% 0.205% 0.125% 0.125% emerging vwo 0.064% 1.577% 0.992% 0.878% emerging pxh 0.163% 1.816% 1.213% 1.064% emerging pie 0.351% 2.429% 1.018% 1.266% emerging 259s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 6. alpha and beta we also test to see if the selections of securities within the etf provide additional value to investors by computing jensen (1968) � as follows: �retf, t � rf, t� � �i � �i*�rbenchmark, t � rf, t� � �i,t (7) where r etf, t and r benchmark, t are monthly returns on the etf and their benchmark index, respectively. r f, t is the three month t-bill rate. table 5b: shows average of te1, te2, and te3 ranked in ascending order (smaller average te is better) rank etf average te category 1 ivv 0.011% u.s. 2 iwb 0.016% u.s. 3 iwv 0.017% u.s. 4 iyy 0.018% u.s. 5 dia 0.020% u.s. 6 tok 0.070% total world 7 eeb 0.087% emerging 8 adrd 0.109% developed 9 bik 0.125% emerging 10 efa 0.139% developed 11 adre 0.145% emerging 12 doo 0.193% developed 13 dew 0.199% total world 14 dnl 0.212% total world ex u.s. 15 dem 0.228% emerging 16 dth 0.245% developed 17 qqq 0.285% u.s. 18 dwm 0.286% developed 19 cwi 0.390% total world ex u.s. 20 idv 0.428% developed 21 dol 0.473% developed 22 gwl 0.482% total world ex u.s. 23 piz 0.535% developed 24 bkf 0.540% emerging 25 gmm 0.575% emerging 26 pxf 0.586% developed 27 fgd 0.700% total world 28 eem 0.791% emerging 29 vwo 0.878% emerging 30 vea 0.881% developed 31 veu 0.931% total world ex u.s. 32 pxh 1.064% emerging 33 lvl 1.217% total world 34 pie 1.266% emerging 35 ioo 1.872% total world 36 dgt 1.952% total world 260 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 alpha (�i) represents the return the etf can achieve above the return of the benchmark. however, as etfs are passively managed and fully invested in the benchmark index, they are not expected to outperform the benchmark index and generate positive �. on the other hand, etfs are expected to have slightly negative �s, as they are going to underperform the benchmark by the amount of expenses they charge. the beta (�i) coefficient is the measure of systematic risk. if ��1, the etf moves more aggressively than the benchmark index, whereas if ��1, the etf manager is much more conservative than the benchmark index. if � � 1, it indicates that etf is very consistent with the benchmark index movements. table 6 shows � and � for each etfs etf � t � t r2 category qqq �0.0001592 [�0.48] 1.002071‡ [184.67] 0.9977 us dia �0.0001197‡ [�9.30] 0.9976226‡ [3193.62] 1.0000 us iyy �0.000146‡ [�8.58] 0.9974398‡ [4991.90] 1.0000 us iwv �0.0000974‡ [�8.63] 0.9971857‡ [4314.20] 1.0000 us iwb �0.0000745‡ [�6.93] 0.9976105‡ [4320.09] 1.0000 us ivv �0.0000346‡ [�4.37] 0.9981557‡ [6374.79] 1.0000 us ioo 0.0029214 [0.96] 0.7275802‡ [3.82] 0.7614 total world lvl 0.0009631 [0.54] 0.9929058‡ [46.71] 0.9641 total world tok 0.0002312‡ [3.57] 0.994646‡ [793.91] 0.9999 total world fgd 0.0001082 [0.13] 1.01569‡ [72.18] 0.9914 total world dew �0.0001941 [�0.80] 0.9951681‡ [265.80] 0.9991 total world dgt �0.0028468 [�1.65] 0.8893292‡ [22.22] 0.9377 total world veu �0.0002154 [�0.20] 1.015923‡ [54.09] 0.9840 total world ex us cwi �0.0002438 [�0.77] 0.9887347‡ [112.71] 0.9986 total world ex us dnl �0.0005678† [�2.44] 1.000929‡ [255.82] 0.9989 total world ex us gwl �0.0005721 [�1.57] 0.9756828‡ [92.78] 0.9980 total world ex us adrd 0.0001712 [1.63] 0.9968484‡ [450.81] 0.9998 developed dol 0.0001388 [0.12] 1.00407‡ [49.19] 0.9805 developed doo 0.0000518 [0.24] 0.9933894‡ [255.90] 0.9993 developed dth �0.0001406 [0.24] 0.9900525‡ [212.84] 0.9990 developed dwm �0.000339 [�1.25] 0.9880473‡ [195.63] 0.9990 developed efa �0.0004915* [�1.82] 0.9883476‡ [167.06] 0.9989 developed idv �0.0000502 [�0.69] 0.9951643‡ [723.20] 0.9999 developed piz �0.0000505 [�0.07] 0.9835803‡ [81.64] 0.9945 developed pxf �0.0008098 [�1.62] 1.004927‡ [98.55] 0.9970 developed vea �0.0008811 [�1.40] 1.005444‡ [78.96] 0.9953 developed adre �0.00009 [�0.85] 0.9983069‡ [296.77] 0.9999 emerging dem �0.0008532‡ [�3.53] 0.9967554‡ [247.09] 0.9992 emerging gmm �0.0001712 [�0.27] 0.976705‡ [84.49] 0.9958 emerging eem �0.0002985 [�0.32] 0.9801021‡ [62.31] 0.9911 emerging bkf �0.0003249 [�0.47] 0.9981108‡ [96.91] 0.9962 emerging eeb �0.0004049 [�3.58] 0.9941872 [847.60] 0.9999 emerging bik �0.0004605‡ [�3.19] 0.9951161‡ [487.74] 0.9998 emerging vwo �0.0006467 [�0.53] 1.006852‡ [64.07] 0.9853 emerging pxh �0.001591 [�1.08] 0.9734497‡ [53.92] 0.9776 emerging pie �0.0035638‡ [�2.80] 1.020725‡ [42.07] 0.9859 emerging t-stats are heteroskedasticity consistent. * significant at 10%. † significant at 5%. ‡ significant at 1%. 261s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 following rompotis (2009), � is also a measure of etfs replication strategy. a � of 1 reflects full replication strategy, whereby the etf invest all its funds in the benchmark index. on the other hand, � that is significantly different than 1 represents a departure from full replication. in such cases, it is assumed that the manager selected stocks anticipating returns better than the benchmark. as expected and shown in table 6, most of the etfs have very small or insignificantly negative �s. u.s. etfs slightly underperform their benchmark (� is significant, but the magnitude of annualized � is very small with the range being �0.04% to �0.17%), whereas for international etfs, � is insignificant in most cases. the � is significantly positive only for one total world etf (tok). even in this case, the magnitude of outperformance is very small (annualized � of 0.2%). � is positive and significant (at 1%) in all cases. in most cases, � is very close to 1 (�� 0.98), which indicates full replication strategy by the etf manager. this result clearly indicates that most of these etfs use passive replication strategies and do not manage for positive �. 7. diversification we now measure the diversification benefits (if any) of international etfs for u.s. investors by first computing the average correlation between s&p 500 and other major u.s. indices from january 2008 through june 2013. the spearman rank correlation test shown in table 7a indicates that all major u.s. indices are highly correlated with the s&p 500 index. the results are also statistically significant at 1% in all cases. for example, the correlation between s&p 500 and russell 3000 (that measures the performance of the largest 3,000 u.s. companies representing approximately 98% of the investable u.s. equity market) is 0.9981 (statistically significant at 1%). similarly, the significance between s&p 500 and dj u.s. total market index is 0.9987 (again significant at 1%). table 7a: spearman rank correlation tests and their significance between s&p 500 and other major u.s. indices major us indicies s&p 500 russell 1000 djia russell 3000 dj u.s. total market index nasdaq 100 s&p 500 1.0000 russell 1000 0.9991‡ 1.0000 djia 0.9804‡ 0.974‡ 1.0000 russell 3000 0.9981‡ 0.9996‡ 0.9715‡ 1.0000 dj us total market index 0.9987‡ 0.9999‡ 0.9724‡ 0.9998‡ 1.0000 nasdaq 100 0.9232‡ 0.9281‡ 0.8691‡ 0.9283‡ 0.9292‡ 1.0000 * significant at 10%. † significant at 5%. ‡ significant at 1%. 262 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 secondly, we measure the correlation between s&p 500 and international etfs over the same period. our results shown in table 7b indicate that all international etfs are highly correlated with the s&p 500 (statistically significant at 1% in all cases). for example, for world ex u.s. etfs, the correlation varies from 0.7824 to 0.9222 (statistically significant at 1% in all cases). even in the case of emerging market etfs, correlation with s&p 500 varies from 0.7948 to 0.8583 (significant at 1% in all cases).2 table 7b: shows spearman rank correlation between s&p 500 and international etfs total world s&p 500 dew dgt ioo fgd lvl tok s&p 500 1 dew 0.9229‡ 1 dgt 0.9569‡ 0.9438‡ 1 ioo 0.9596‡ 0.9678‡ 0.9867‡ 1 fgd 0.9031‡ 0.9636‡ 0.9202‡ 0.933‡ 1 lvl 0.8937‡ 0.9372‡ 0.8889‡ 0.8998‡ 0.9587‡ 1 tok 0.9779‡ 0.9699‡ 0.9765‡ 0.9851‡ 0.942‡ 0.9212‡ 1 total world ex u.s. s&p 500 cwi dnl gwl veu s&p 500 1 cwi 0.9176‡ 1 dnl 0.7824‡ 0.8782‡ 1 gwl 0.9222‡ 0.9969‡ 0.8827‡ 1 veu 0.918‡ 0.9927‡ 0.8711‡ 0.9899‡ 1 developed markets s&p 500 adrd dol doo dth dwm efa idv piz pxf vea s&p 500 1 adrd 0.9189‡ 1 dol 0.9044‡ 0.9865‡ 1 doo 0.9077‡ 0.9735‡ 0.9866‡ 1 dth 0.9022‡ 0.9796‡ 0.9947‡ 0.9928‡ 1 dwm 0.9071‡ 0.9853‡ 0.998‡ 0.987‡ 0.9934‡ 1 efa 0.915‡ 0.986‡ 0.9911‡ 0.9775‡ 0.9823‡ 0.9957‡ 1 idv 0.8969‡ 0.9528‡ 0.9655‡ 0.9754‡ 0.9729‡ 0.9714‡ 0.9658‡ 1 piz 0.8714‡ 0.9256‡ 0.9258‡ 0.9023‡ 0.9031‡ 0.934‡ 0.9446‡ 0.8982‡ 1 pxf 0.9036‡ 0.9788‡ 0.9768‡ 0.9708‡ 0.9754‡ 0.9817‡ 0.9844‡ 0.9651‡ 0.9187‡ 1 vea 0.9181‡ 0.9915‡ 0.986‡ 0.9772‡ 0.9784‡ 0.9886‡ 0.991‡ 0.9608‡ 0.9338‡ 0.978‡ 1 emerging markets s&p 500 adre bik bkf dem eeb eem gmm pie pxh vwo s&p 500 1 adre 0.8459‡ 1 bik 0.7948‡ 0.962‡ 1 bkf 0.8041‡ 0.971‡ 0.9938‡ 1 dem 0.8394‡ 0.9302‡ 0.9259‡ 0.933‡ 1 eeb 0.8241‡ 0.9911‡ 0.9731‡ 0.9816‡ 0.9257‡ 1 eem 0.8583‡ 0.9767‡ 0.9639‡ 0.9754‡ 0.9665‡ 0.968‡ 1 gmm 0.8465‡ 0.9735‡ 0.9786‡ 0.984‡ 0.9686‡ 0.9718‡ 0.992‡ 1 pie 0.8453‡ 0.9321‡ 0.8994‡ 0.9132‡ 0.9204‡ 0.9164‡ 0.9482‡ 0.943‡ 1 pxh 0.8575‡ 0.9752‡ 0.9568‡ 0.9657‡ 0.9578‡ 0.9649‡ 0.9903‡ 0.9826‡ 0.9362‡ 1 vwo 0.8583‡ 0.978‡ 0.9628‡ 0.973‡ 0.9632‡ 0.9691‡ 0.9945‡ 0.9916‡ 0.9501‡ 0.9897‡ 1 * significant at 10%. † significant at 5%. ‡ significant at 1%. 263s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 8. single factor model following pennathur, delcoure, and anderson (2002), we use the following single factor model to estimate the diversification benefits of international etfs for u.s. investors. they used this model to estimate the diversification of international closed-end country funds relative to the s&p 500. retf,t � �i � �i* rs&p 500, t � ei, t (8) where retf,t and r s&p 500,t are monthly returns for international etfs and the s&p 500 index, respectively. table 8 the regression of monthly international etf returns on monthly s&p 500 returns for the entire period (january 2008 through june 2013) following pennathur et al. (2002) etf � t s&p 500 t r2 category dew �0.005464* [�1.75] 1.143574‡ [17.90] 0.8517 total world dgt �0.0052699† [�2.61] 0.9948454‡ [26.41] 0.9157 total world ioo �0.0033817* [�1.68] 1.029334‡ [26.76] 0.9208 total world fgd �0.0027508 [�0.73] 1.224583‡ [14.22] 0.8155 total world lvl �0.005087 [�1.19] 1.275853‡ [10.91] 0.7987 total world tok �0.0025677 [�1.63] 1.098427‡ [38.61] 0.9563 total world cwi �0.0051134 [�1.54] 1.154643‡ [19.73] 0.8419 total world ex us dnl �0.0014771 [�0.34] 0.8249698‡ [7.41] 0.6121 total world ex us gwl �0.0049754 [�1.59] 1.12824‡ [20.96] 0.8504 total world ex us veu �0.0051187 [�1.49] 1.207013‡ [20.79] 0.8427 total world ex us adrd �0.0059064* [�1.79] 1.177599‡ [21.48] 0.8444 developed dol �0.0060609* [�1.77] 1.10468‡ [20.49] 0.8179 developed doo �0.0067583* [�1.85] 1.18415‡ [16.39] 0.8239 developed dth �0.0063656* [�1.73] 1.173226‡ [18.85] 0.8140 developed dwm �0.0057644* [�1.70] 1.111143‡ [20.12] 0.8229 developed efa �0.0053195 [�1.62] 1.123948‡ [20.20] 0.8373 developed idv �.0041174 [�0.97] 1.286446‡ [12.96] 0.8045 developed piz �0.0040626 [�0.87] 1.208036‡ [12.95] 0.7593 developed pxf �0.0057306 [�1.49] 1.266507‡ [17.28] 0.8165 developed vea �0.0050758 [�1.54] 1.162653‡ [22.44] 0.8429 developed adre �0.0077003 [�1.46] 1.239071‡ [11.31] 0.7155 emerging bik �0.0069385 [�1.05] 1.282092‡ [9.21] 0.6316 emerging bkf �0.0087125 [�1.29] 1.362032‡ [10.18] 0.6466 emerging dem �0.0005135 [�0.11] 1.061991‡ [12.12] 0.7046 emerging eeb �0.0085882 [�1.37] 1.359787‡ [10.79] 0.6791 emerging eem �0.0051737 [�1.01] 1.28185‡ [13.43] 0.7366 emerging gmm �0.0043526 [�0.82] 1.250542‡ [10.95] 0.7166 emerging pie �0.0066144 [�1.12] 1.343829‡ [9.85] 0.7145 emerging pxh �0.005889 [�1.16] 1.287644‡ [13.18] 0.7353 emerging vwo �0.0051226 [�0.97] 1.31925‡ [12.52] 0.7367 emerging t-stats are heteroskedasticity consistent. * significant at 10%. † significant at 5%. ‡ significant at 1%. 264 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 here we regress monthly international etf returns on monthly s&p 500 returns. a � close to or higher than 1 would indicate that international etf return mimics the s&p 500, whereas r2 provides information on tracking effectiveness of the etfs. our results shown in table 8 indicate that the coefficient for the s&p 500 is close to or greater than 1 and statistically significant at 1% in all cases. for example, for the four total world ex u.s. etfs, the coefficient for the s&p 500 varies from 0.83 to 1.21 (statistically significant at 1% in all cases). the r2 is also high and varies from 0.6121 to 0.8504. similarly, for total world, developed, and emerging market etfs, coefficient for the s&p 500 is very close to or much greater than 1 in all cases (results are statistically significant at 1% in all cases). r2 is also high in all cases that indicate that international etfs closely track the s&p 500. the results are similar for other major u.s. indices (not reported but available upon request). pennathur et al. (2002) found similar results for international closed end country funds. these results indicate that international etfs closely follow u.s. indices and there are not many diversification benefits from investing in international etfs for u.s. investors. 9. principal component analysis we also use principal component analysis (pca) analysis to compute diversification benefits of international etfs for u.s. investors. this method groups international etfs and s&p 500 returns into principal components in terms of similarities in their return movement patterns. if international etfs and the s&p 500 have high factor loadings in the same principal component, they are highly correlated, and, hence, there is limited diversification benefit international etfs for u.s. investors. if the s&p 500 has low factor loadings in the same principal loadings (than international etfs), then there are significant benefits of diversification. therefore, investors should invest in etfs that have high factor loadings in different principal components than s&p 500 to get benefits of diversification. in this method, the correlation matrix of monthly returns for international etfs and s&p 500 is used as the input for the entire period. the eigen value reported in table 9 for only the first common factor is greater than 1 and explains more than 90% of the variation in all cases. hence, only the first common factor is important and is reported for this analysis (eigen value 2 and its variation are also shown for comparison purposes. detailed results are available upon request.) results again indicate that international etfs are highly correlated to u.s. markets as the factor loadings of international etfs for component 1 are very close to factor loadings of the s&p 500 for component 1. these results hold for other u.s. indices too. 10. risk adjusted performance and cwi of equally weighted portfolios we form equally weighted portfolios of u.s., total world, total world ex u.s., developed, and emerging market etfs and compute their risk adjusted performance (sharpe and sortino ratios) for the entire period. the results from table 10a indicate that u.s. etf portfolio has the best performance (both absolute and risk-adjusted performance) for the entire period. similarly, u.s. etfs portfolios have the highest cumulative returns and cwi. 265s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 table 10b shows the spearman-rank correlation test between s&p 500 and equally weighted etf portfolios. results again indicate that all international etf portfolios are highly correlated with s&p 500 (all the results are statistically significant at 1%). results table 9 eigen values for component 1 and 2 and the principal factor loadings for component1 for international etfs and s&p 500 total world entire period eigen value proportion cumulative component 1 6.66062 0.9515 0.9515 component 2 0.187901 0.0268 0.9784 variable (entire period) factor loading component 1 tok 0.3844 ioo 0.3821 dew 0.3805 dgt 0.3788 fgd 0.3756 s&p 500 0.3754 lvl 0.3687 total world ex u.s. entire period eigen value proportion cumulative component 1 4.66593 0.9332 0.9332 component 2 0.225322 0.0451 0.9783 variable (entire period) factor loading component 1 gwl 0.4598 cwi 0.4592 veu 0.4579 s&p 500 0.4354 dnl 0.4224 developed markets entire period eigen value proportion cumulative component 1 10.5741 0.9613 0.9613 component 2 0.153925 0.014 0.9753 variable (entire period) factor loading component 1 dwm 0.3064 efa 0.3061 dol 0.3057 vea 0.3053 dth 0.3045 adrd 0.3046 doo 0.3038 pxf 0.3033 idv 0.3002 piz 0.2896 s&p 500 0.2863 (continued on next page) 266 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 (not reported) are similar when we regress equally weighted portfolio returns on s&p 500 as well as the pca. these results again indicate that these international etfs are highly dependent on u.s. indices and there were limited benefits of diversification in these etfs for u.s. investors during the period of our analysis. 11. conclusions our results indicate that u.s. etfs outperform international etfs during the period beginning january 2008 through june 2013. u.s. etfs have higher average returns and lower risk (standard deviation of returns) than international etfs. risk adjusted measures of performances (sharpe, sortino, and treynor ratios) also confirm that u.s. etfs outperform international etfs. table 9 (continued) emerging markets entire period eigen value proportion cumulative component 1 10.4036 0.9458 0.9458 component 2 0.284646 0.0259 0.9717 variable (entire period) factor loading component 1 eem 0.3084 gmm 0.3084 vwo 0.3083 pxh 0.3067 adre 0.3056 bkf 0.3045 eeb 0.3042 bik 0.3021 dem 0.2996 pie 0.2958 s&p 500 0.2711 because the eigen value only for component 1 is greater than 1, only factor loading for component 1 are reported. if factor loadings for component 1 are close to each other, there are limited benefits of diversification. table 10a: shows equally weighted portfolios of u.s., total world, total world ex u.s., developed, and emerging market etfs and their risk adjusted performance (sharpe, sortino and treynor ratios), cumulative returns, and cumulative wealth index (cwi) the portfolios rank equally weighted portfolio time period (january 2008 through june 2013) average monthly return sd of monthly returns sharpe ratio sortino ratio cumulative returns cumulative wealth (initial wealth $1,000 in january 2008) number of etfs 1 u.s. 66 months 0.53% 5.35% 0.0939 0.1307 29.26% $1,292.63 6 2 total world 66 months 0.11% 6.26% 0.0134 0.0182 �5.46% $ 945.40 6 3 total world ex u.s. 66 months 0.08% 6.28% 0.0084 0.0115 �7.36% $ 926.37 4 4 emerging 66 months �0.0036% 7.96% �0.0042 �0.0057 �19.45% $ 805.45 10 5 developed 66 months �0.005% 6.82% �0.0051 �0.0068 �14.68% $ 853.19 10 267s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 jensen’s � indicates that most of these etfs have negative or insignificant �s. these results are expected as these etfs are passively managed and closely follow their benchmark, but underperform the benchmark by the amount of expenses they charge. alpha is positive in only one instance (tok), however, even in cases where etfs have significantly positive or negative �s, the amount of out or under performance is very small. � is positive and significant (at 1%) in all cases. in most cases, � is very close to 1 (�� 0.98), which indicates full replication strategy by the etf manager. this clearly indicates passive replication instead of active management for positive �. diversification benefits of international etfs-results indicate that international etfs are highly correlated with major u.s. indices during the entire period. spearman rank correlation tests find that all international etfs are highly correlated with the s&p 500 during the entire period (results are significant at 1%). results are similar for other major u.s. indices (djia, nasdaq 100, russell 1000, russell 3000, and dow jones u.s. total return index). we find similar results with pca. the second model we use (following pennathur, delcoure, and anderson, 2002), where we regress monthly returns of international etfs against s&p 500 returns indicates that all international etfs are highly dependent on s&p 500. these results are statistically significant at 1% or 5% in all cases. we find similar results between international etfs and other major u.s. indices. in conclusion our results indicate that during the financial crisis and the ensuing recovery, u.s. etfs provided superior performance relative to international etfs on both an absolute and risk-adjusted basis. in addition, during this period, international etfs exhibit high correlation with u.s. markets that eliminates most, if not all, of their global diversification benefits. as such, individual investors should be aware that global diversification using etfs may not provide them with any benefits especially during times of extreme financial distress. table 10b: shows the spreaman-rank correlation between equally weighted etf portfolios and s&p 500 for the period of our study (january 2008-june 2013) s&p 500 u.s. etf portfolio total world etf portfolio total world ex u.s. etf portfolio emerging etf portfolio developed etf portfolio s&p 500 1 u.s. etf portfolio 0.9968‡ 1 total world etf portfolio 0.9408‡ 0.9376‡ 1 total world ex u.s. etf portfolio 0.8770‡ 0.8791‡ 0.9651‡ 1 developed etf portfolio 0.8896‡ 0.8876‡ 0.9790‡ 0.9794‡ 1 emerging etf portfolio 0.8091‡ 0.8183‡ 0.8868‡ 0.9519‡ 0.8999‡ 1 * significant at 10%. † significant at 5%. ‡ significant at 1%. 268 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 notes 1 over one half of international etfs were created in 2007 as shown in table 1a. 2 results not reported in the article indicate that international etfs are also very highly (and significantly) correlated with other major u.s. indices over the entire period, and hence there are limited benefits of diversification for the period of our study. references adler, m., & b. dumas, (1983). international portfolio choice and corporate finance. journal of finance, 38, 925–84. aiello, s., & chieffe, n. (1999). international index funds and the investment portfolio. financial services review, 8, 27–35. black, f. (1974). international capital market equilibrium with investment barriers. journal of financial economics, 1, 337–52. eun, c. s., & shim, s. (1989). international transmission of stock market movements. journal of financial and quantitative analysis, 24(02), 241–256. frino, a., & gallagher, d. r. (2001). tracking s&p 500 index funds. the journal of portfolio management, 28, 44–55. hanna, m. e., mccormack, j. p., & perdue, g. (1999). a nineties perspective on international diversification. financial services review, 8, 37–45. harper, j. t., madura, j., & schnusenberg, o. (2006). performance comparison between exchange-traded funds and closed-end country funds. journal of international financial markets, institutions and money, 16, 104–122. heston, s. l., & rouwenhorst, k. g. (1994). correlations in price changes and volatility across international stock markets. review of financial studies, 7, 281–307. ho, k., milevsky, m. a., & robinson, c. (1999). international equity diversification and shortfall risk. financial services review, 8, 11–25. jacquier, e., & marcus, a. (2001). asset allocation models and market volatility. financial analysts journal, 57, 16–30. johnsona, r. r., buetow, g. w., & jensen, g. r. (1999). international mutual fund returns and federal reserve policy. financial services review, 8, 199–210. jensen, m. c. (1968). the performance of mutual funds in the period 1945–1964. the journal of finance, 23, 389–416. king, m. a., & wadhwani, s. (1990). transmission of volatility between stock markets. review of financial studies, 3(1), 5–33. koch, p. d., & koch, t. w. (1991). evolution in dynamic linkages across daily national stock indexes. journal of international money and finance, 10, 231–251. levy, h., & sarnat, m. (1970). international diversification of investment portfolios. american economic review, 60, 668–675. longin, f., & solnik, b. (1995). is the correlation in international equity returns constant: 1960–1990? journal of international money and finance, 14, 3–26. pennathur, a. k., delcoure, n., & anderson, d. (2002). diversification benefits of ishares and closed-end country funds. journal of financial research, 25, 541–557. pries, t., kenett, d., stanley, h., helbing, d., & ben-jacob, e. (2012). quantifying the behavior of stock correlations under market stress. scientific reports,2, article number: 752. pope, p. f., & yadav, p. k. (1994). discovering errors in tracking error. the journal of portfolio management, 20, 27–32. roll, r. (1988). the international crash of october (1987). financial analysts journal, 1988, 19–35. 269s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 rompotis, g. g. (2009). interfamily competition on index tracking: the case of the vanguard etfs and index funds. journal of asset management, 10, 263–278. rompotis, g. g. (2010). investing overseas from home: the case of asian ishares. journal of asset management, 11, 1–18. russell, j. w. (1998). the international diversification fallacy of exchange-listed securities. financial services review, 7, 95–106. shin, s., & soydemir, g. (2010). exchange-traded funds, persistence in tracking errors and information dissemination. journal of multinational financial management, 20, 214–234. stulz, r. (1981). on the effects of barriers to international investment. journal of finance,36, 923–934. woolridge, r. j. (2004). performance of stocks recommended by brokerages. the journal of investing, 13, 23–34. 270 s. kanuri, r.w. mcleod / financial services review 24 (2015) 249–270 from the editor this issue contains issue 2 of volume 26 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “retirement income strategies designed in an expected utility framework” is authored by mark j. warshawsky at relias llc. in this paper, the author evaluates various classes of retirement income strategies and tests their robustness in an expected utility framework. he finds that fixed percentage systematic withdrawals from an investment portfolio combined with laddered purchases of immediate life annuities stand out as a superior strategy for retired defined contribution plan participants and ira holders, yielding better outcomes than alternatives, including longevity insurance. the second article “the financial literacy of generation y and the influence that personality traits have on financial knowledge: evidence from canada” is authored by robert n. killins at seneca college of applied arts and technology. in this paper, the author examines the financial literacy of generation y and explores how personality traits influence individual’s financial knowledge. he uses a financial literacy survey and multiple areas of financial literacy are measured (investments, budgeting, economics, risk management, and retirement planning) along with the well-known big five personality traits. the author finds that the generation y cohort is more knowledgeable in budgeting and risk management segments of financial literacy, but they lack knowledge in retirement planning. additionally, he finds that extraversion and conscientiousness are both important personality traits when regressed on individuals overall financial literacy levels. the third article, “active asset allocation for retirement funds using the fed model” is coauthored by john m. clinebell at university of northern colorado, douglas r. kahl at university of akron, and jerry l. stevens at university of richmond. active asset allocation, also known as market timing, is controversial but potentially effective for individual investors and financial advisors. this study addresses many of the concerns related to market timing studies. the authors control for transaction costs and tax effects by focusing on funds available for retirement accounts within a vanguard fund family which allows for costless monthly transfers. they use a “risk on” or “risk off” approach rather than experiment with arbitrary cutoff rules for switching funds. they test the time series properties of the fed financial services review 26 (2017) v–vi 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. model to build a prediction model and then apply it over a recent five-year hold-out of period. their findings show that the switching portfolio offers attractive performance compared to either of the vanguard funds, especially with respect to enhancing upside to downside risk ratios. the fourth article, “bond laddering and bond indexing: an empirical comparison” is coauthored by c. sherman cheung and peter miu at mcmaster university. bond laddering and bond indexing have been widely accepted approaches to bond investing among retail investors, although bond laddering has virtually been ignored in the academic literature. both approaches are passive strategies with do not attempt to beat the market. open questions are: which approach should an investor favor, is there any room for both to be used at the same time, and what is the appropriate term to maturity for the ladder. the authors investigate the relative attractiveness of the two approaches and identify conditions that favor one over the other. further, the authors examine conditions under which both instruments should be held within an optimal portfolio and identify situations in which a longer-term ladder is more appropriate than a shorter-term ladder. the final article, “the impact of the capitalization of operating leases: a guide for individual investors” is coauthored by jack trifts at bryant university and gary e. porter at northeastern university. in this paper, the authors provide an explanation of the new fasb standard requiring firms to move their off-balance sheet operating leases onto the balance sheet beginning in 2019 and discuss how the new rule might affect the stock and bond values. despite large increases in on-balance sheet liabilities for some industries, the authors caution investors not to anticipate changes in their stock or bond valuations resulting from this change. because asset values change in response to new information, and information regarding changes in total assets and debt ratios is currently available in the notes to financial statements and from data providers, such as bloomberg, this information is already incorporated in forecast asset values developed by professionals. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 26 (2017) v–vi from the editor this issue contains issue 3 of volume 24 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “how risky is your retirement income risk model?” is coauthored by patrick j. collins at university of san francisco, huy lam at schultz collins, inc., and josh stampfli. this article provides a review of various retirement income modeling approaches including historical back testing, monte carlo simulations, and other more advanced risk modeling techniques. the authors show that implausible assumptions underlying common risk models may mislead investors concerning the risk and return expectations of their retirement investment strategies. the authors demonstrate how an over-simplified model may distort the risks facing retired investors. the second article “the impact of superannuation fund choice legislation and the global financial crisis on australian retail fund flows” is coauthored by rakesh gupta at griffith university and thadavillil jithendranathan at the university of st. thomas. the authors examine the extent to which cash flows into the australian superannuation funds are affected by the past performance of the fund, riskiness of the fund, choice of superannuation fund legislation, and the global financial crisis. they find that both retail and wholesale investors base their investment decision on the past performance of the funds and that there is very little evidence that the riskiness of fund returns has a significant effect on the flow of funds. the authors also show that legislation has resulted in more inflows into managed funds and there are higher inflows into managed funds and equity funds since the period of the global financial crisis. the third article, “does it pay to diversify? u.s. vs. international etfs” is coauthored by srinidhi kanuri at the university of southern mississippi and robert w. mcleod at the university of alabama. this article evaluates the performance and diversification benefits of international etfs for u.s. investors during and after the recent financial crisis. the author’s show that u.s. etfs outperform all categories of international etfs during the period january 2008–june 2013. the etfs have higher average monthly returns, lower risk, higher risk-adjusted performance and the highest cumulative returns over the entire period. the authors also find that u.s. etfs have the lowest tracking error during the entire period. financial services review 24 (2015) v–vi 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. finding that most of these etfs passively track the benchmark and do not manage for positive alpha, the authors indicate that international etfs are highly dependent on major u.s. indices during the period of the analysis, and therefore, offered limited diversification benefits for u.s. investors. the fourth article, “return-enhancing strategies with international etfs: exploiting the turn-of-the-month effect” is coauthored by haiwei chen at university of alaska fairbanks, sang heon shin at alabama state university, and xu sun at university of texas–pan american. in this article, the authors show that the average return over the four-day period surrounding the turn of the month is significantly positive for eight out of nine international etfs studied. the strategy of buying-and-holding an etf during tom period and switching to holding t-bills during non-tom period produces significantly positive monthly average returns. this etf-t-bill switching strategy has the lowest risk, highest sharpe ratio, and highest sortino ratio than the traditional strategy of buying-and-holding either an index fund or an etf. investors that pursue this strategy can expect to generate a terminal value twice as large as the strategy of buying-and-holding an etf. the final article, “the performance and market timing ability of chinese mutual funds” is coauthored by wei he at mississippi state university, bolong cao at ohio university, and h. kent baker at the american university. the authors examine the performance, fund flows, and market timing ability of actively managed chinese stock mutual funds. based on daily return data and several four-factor models, only about 7.5% of these funds have statistically significant risk-adjusted abnormal returns and even fewer demonstrate market timing ability. after controlling for fund size, management fees, average amount and volatility of fund flows, older funds show higher sharpe ratios. their results also shows that the volatility of fund flows has an inverted-u shape relationship with fund performance. the results provide an indication that investors should focus on index funds in china unless they have solid evidence that the active managers under consideration have the ability to consistently generate excess risk-adjusted returns. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 24 (2015) v–vi does financial risk tolerance change over time? a test of the role macroeconomic, biopsychosocial and environmental, and social support factors play in shaping changes in risk attitudes stephen kuzniak, ph.d.a, john e. grable, ph.d.a,* adepartment of financial planning, housing and consumer economics, university of georgia, 300 dawson hall, athens, georgia 30602, usa abstract financial planners work in an environment that requires the documentation of a client’s financial attitudes and preferences. financial risk tolerance is one such attitudinal construct that is generally required by regulators to be evaluated. while there are numerous commercial and academic products used to assess client risk attitudes, questions have been raised over the past several decades regarding the stability of scores from risk-tolerance tools. specifically, financial planners, as well as regulators, require evidence documenting to what extent risk tolerance changes over time, and if changes do occur, the variables associated with variability. the purpose of this study was to address these needs. based on a model that included macroeconomic indicators, biopsychosocial and environmental factors, and measures of social support, it was determined that risk-tolerance attitudes remain generally stable over time. however, there are groups of test takers that exhibit significant shifts in risk tolerance. this article describes some of the variables associated with these score changes, as well as providing financial planning professionals with guidance on how to identify clients who may be prone to shifting their tolerance for financial risk. © 2017 academy of financial services. all rights reserved. keywords: financial risk tolerance; macroeconomic indicators; social support; change in risk tolerance * corresponding author. tel.: �1-706-542-4758; fax: �1-706-542-4856. e-mail address: grable@uga.edu (j.e. grable) financial services review 26 (2017) 315–338 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. 1. introduction understanding appropriate investment options and recommending a suitable allocation of a client’s assets is a key component of a well-drafted comprehensive financial plan. accurate assessment of financial risk tolerance, as an element of the asset allocation process, is generally accepted as an essential condition to developing a suitable and quality financial plan for individuals (cfp board, 2015). for those working as a financial planner, financial risk tolerance (frt) can be defined parsimoniously as an individual’s willingness to take risk (dalton and dalton, 2004). in the information and data gathering stage of client work, a suitable risk assessment is generally required to be used to meet regulatory requirements, as well as to formulate the best plan for an individual (roszkowski and davey, 2010). understanding how a person’s frt influences decision making and behavior is becoming an increasingly important aspect of how financial planners formulate and execute recommendations. for researchers, practitioners, policy makers, economists, and financial professionals, understanding the role of risk and frt is closely linked to better understanding the mechanics that combine to influence an individual’s behavior (xiao, 2008). frt assessment serves as a foundation for nearly all financial planning models, frameworks, and recommendations. a well-designed frt assessment is a tool that can be used to anticipate an individual’s decisions, determine optimal financial choices, and maximize utility under the constraint of imperfect knowledge. one question related to the study of frt is of particular importance, specifically: does frt change over time? the concept of frt “traitedness” is gaining traction as a way to answer the question of how much an individual’s frt deviates over time (roszkowski, delaney, and cordell, 2009). the extent to which people will exhibit a personality trait in behaviors across different situations and contexts defines traitedness (baumeister and tice, 1988). for financial planners, policy makers, and researchers, answering the question of how much an individual’s frt changes (if at all) across time is needed to fully understand how clients will react in a variety of situations and within the context of changing macroeconomic environments. the purpose of this study was to document changes in frt across time. an important aspect of the study was to test whether macroeconomic variables and social support, as indicated by country of residence, were associated with changes in frt at the individual/ household level. results from this study help expand the existing literature on the degree to which frt changes over time. furthermore, results provide an insight into the role macroeconomic and household level variables play in shaping changes in frt. 2. research framework if the assumption that frt is an essential element in the development of an accurate and acceptable comprehensive financial plan is true, it then follows that understanding its malleability over time is an important aspect to consider in the financial planning process. roszkowski and davey (2010) delved deeply into how major events, like the global financial crisis, can affect an individual’s measured frt. they noted that some view frt as a completely stable characteristic (trait), while others view frt as something that varies 316 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 depending on the mood or environment of the test taker (state). however, they concluded, based on a review of the literature and their own experience, that frt is relatively stable over time but somewhat susceptible to situational influences and life circumstances. the implications of this insight are important for financial planners to contemplate, especially considering the unique nature of the field in which multidecade relationships are common. to fully understand the impact different variable relationships have on an individual’s willingness to take risk, a model was developed specifically for this study. this model uniquely includes propositions about the associations between and among macroeconomic variables, demographic factors, and social support and frt. the model is shown in fig. 1. the model was developed using concepts from three frameworks of risk taking: macroeconomic theory, the cushion hypothesis, and a model of the determinants of risk taking developed by irwin (1993). it was hypothesized in this study that the macroeconomic condition of any nation may be associated with changes in frt. macroeconomic conditions are complex, with codependent activities that combine to produce and consume resources. macroeconomic factors may influence the willingness of individuals to take risk in two ways. first, negative events may reduce financial capacity, leading to a negative shift in frt. second, perceptions of conditions, rather than the actual impact of macroeconomic events, could shape someone’s willingness to take financial risk. four variables were used in this study to test the impact of macroeconomic conditions on changes in frt: country level gross domestic product (gdp), national unemployment rates, stock market conditions, and global commodity prices. the second element of the framework was based on irwin’s (1993) model of risk taking. irwin surmised that different predisposing factors affect an individual’s risk-tolerance attitude. biopsychosocial and environmental were two concepts irwin used as classifying factors biopsychosocial & environmental factors age income net worth gender education level marital status change in financial risk tolerance macroeconomic indicators gdp unemployment rate market index global commodities index social support aggregate social safety net united states australia united kingdom fig. 1. the financial risk tolerance (frt) model based on changes in frt. 317s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 that influence frt. biopsychosocial factors include variables such as age and gender while environmental factors include income, net worth, education, and marital status, among other factors. taken together, the combination of these factors and characteristics are expected to have a meaningful influence on an individual’s risk-tolerance attitude. imbedded within irwin’s (1993) model are variables related to cultural experiences and socialization. as shown in fig. 1, social support was also included in the model. the choice of this variable was based on the cushion hypothesis. this hypothesis states that individuals who live in collectivist cultures generally have a greater social support system that “cushions” downside risks when making risky decisions (hsee and weber, 1999; weber and hsee, 1998). in theory, when personal risk is minimized, individuals are allowed to try new things, start small businesses, or invest in potentially riskier opportunities that promise a higher return. in other words, the hypothesis posits that as social support increases, so does the willingness to take financial risk at the household level. it is important to note, however, that it can also be hypothesized that the opposite may be true. it may be that risk is often taken because of the necessity of making progress or achieving financial goals. statman (2008) noted that individuals often pay with risk for a chance to move up in life and that in many countries, individuals are willing to take greater risks for potentially higher rewards, even when familial and national support is low. 3. literature review hallahan, faff, and mckenzie (2004) noted the following: “despite its importance in the financial services industry, there remain some unresolved questions with respect to the ‘determinants of financial risk tolerance’” (p. 58). by determinants, hallahan et al. meant the identification of factors or variables that reveal a systematic association with frt. over the years, varied factors have been proposed and tested but the results have been inconsistent. this review highlights literature that has tested some of these factors. 3.1. macroeconomic factors many individuals who were economically active or invested in the markets during the global financial crisis intuitively know that the overall economy likely has some effect on how individuals make decisions. the extent to which economic forces impact individuals and investment markets has been studied by chen, roll, and ross (1986). the results of their research suggested that from the perspective of efficient market theory, asset prices are influenced, to some degree, by macroeconomic factors. in addition, chen et al. concluded that stock returns are exposed to systematic economic news, and assets are priced in relation to this exposure. their study documented an important link in the relationship between the macro economy and the way individuals make investing decisions involving risk. reinhart and borensztein (1994) took a unique approach to measuring the macroeconomic determinants of commodity prices. in their research, they focused on determining real commodity prices beyond that of looking exclusively at demand factors. their research 318 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 examined international developments across eastern europe and the soviet union to help understand the connection between the macro economy and commodity prices. their results, however, were unable to explain the marked and sustained historical commodity price trends throughout the 1980s and 1990s. popular press articles often discuss the relationships between well-known commodities, such as oil and gold, and the association they have with markets and overall economic conditions. while it is possible to see commodity prices as drivers of the economy, nearly all market pundits address commodity issues by looking at the effect market conditions have on investable commodity markets (motley fool, n.d.). in addition to commodity investment markets, countries around the world have varying levels of structural macroeconomic exposure to commodity prices. for several middle eastern economies, for example, commodity prices (including oil) make up disproportionally large components of total revenue and output (world bank, n.d.). west and worthington (2014) examined the relationship between macroeconomic conditions and financial risk attitudes. based in australia, their study relied on surveys of approximately 6,800 households. they noted, consistent with past literature, that demographic characteristics—especially age—had a strong relationship with changes in frt over time. they also noted that macroeconomic conditions were jointly significant in shaping risk attitudes. several of the variables studied were found to be significantly associated with the risk attitudes of individuals. unemployment rates and domestic stock market returns were discussed by yao, hanna, and lindamood (2004). in their work, yao et al. looked at changes in frt during the period 1983–2001. based on the survey of consumer finances, frt exhibited significant increases from 1995 to 1998 during a period of strong stock growth and large drops in unemployment. yao and her associates also noted that poor global economic conditions in asia and russia had a seemingly negligible effect on domestic frt. market conditions have been hypothesized to influence frt. rabbani, grable, heo, nobre, and kuzniak (2017), for example, noted that daily market volatility exhibited a positive association with frt scores in their study, although the relationship was not strong enough to generally warrant a change in portfolio holdings. a similar finding was reported by zeisberger, vrecko, and langer (2010). santacruz (2009) looked at general economic mood and its influence on frt scores. he concluded that there is limited need to make major adjustments to current models. it was noted, however, that financial planners should recognize the herding behavior that can result in investors’ perceptions of recent salient macroeconomic events. in general, however, there continues to be a paucity of research that deals with this topic, and as such, the relationship is still subject to debate. to address this apparent gap in the literature, the relationship between global macroeconomic variables and an individual’s frt was examined in this study using macroeconomic variables, including unemployment rates, national production (gdp), commodity prices, and market pricing. one of the most difficult aspects of examining macroeconomic variables is the interdependent relationship among economic indicators. therefore, one essential step to evaluating the usefulness of economic variables in future studies will be determining which variables are independently related to frt. 319s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 3.2. biopsychosocial and environmental factors age, income, education, and wealth have all been shown to be significantly associated with an individual’s frt (bajtelsmit, bernasek, and jianakoplos, 1999; grable and lytton, 1999; pålsson, 1996), but the explanatory power and magnitude of their effects have been disputed (gollier and zeckhauser, 2002; hariharan, chapman, and domian, 2000). in general, young men and those with more income and wealth are thought to be more risk tolerant compared with older individuals and those with fewer resources. the role of household size in shaping risk attitudes has also been explored. most often, large households tend to exhibit relative risk aversion. this may result from a lack of risk capacity or a preference to be conservative with household resources. similarly, variables associated with human capital have been found to be positively associated with frt. higher attained education, for example, is generally thought to be associated with elevated levels of frt. baker and haslem (1974) showed that some socioeconomic characteristics have a more profound influence in shaping the risk and return preferences of individual investors. among the most important factors are age, gender, marital status, education, and income. the implications of their findings were that a person’s demographic profile can have a strong influence on perceptions of risk and ultimately frt. in 1997, wang and hanna (1997) studied the association between age and frt. based on data from the survey of consumer finances, they tested the life-cycle investment theory. wang and hanna measured frt as the amount of risky assets held as a percentage of total wealth. they concluded that frt increased with age, controlling for other important variables. dahlbäck (1991) found that the propensity to take risks was influenced by saving decisions. individuals who are willing to save more may have the ability to invest more aggressively. this implies that older investors—typically those with more wealth—may be more willing to take more risk. this relationship, however, is out of step with what financial planners typically assume. nearly all financial planners, and some individual investors, simply use heuristics or rules such “age � percent allocated to bonds” to estimate the appropriate risk level within a given portfolio allocation (benartzi and thaler, 2007). however, the effect may not always be related to biological age but instead age acting as a proxy for an investor’s time horizon or risk capacity. by default, as someone ages they lose time to recoup potential losses. as such, there may be no real age effect. a 1996 study by sung and hanna (1996) investigated several factors that are generally thought to have a positive association with a household’s willingness to take a financial risk. based on data from the 1992 survey of consumer finances, they concluded that education, age, and net worth (including liquidity) were positively correlated with a household’s willingness to take some level of risk. it was also shown that female headed households were less likely to be risk tolerant compared with similar male headed or married households. grable (2000) measured risk taking in everyday money matters and the relationships among demographic, socioeconomic, and attitude characteristics both in individuals and groups. his results showed that a higher frt was associated with being male, older, married, professionally employed with higher income, and more education, among other factors. morin and suarez (1983) examined the empirical evidence of the effects of wealth on relative risk aversion. their work investigated a household’s demand for risky investments using a 320 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 dataset of asset holdings based in canada. the results of their study showed a diverging relative risk aversion when housing was excluded from the definition of wealth (or investments) or treated as a riskless asset. in addition, they noted that an investor’s stage in the life cycle and age were uniformly increasing over time with tolerance for risk. bakshi and chen (1994) tested how changes in demographic variables influence investments in capital markets. the life-cycle investment hypothesis suggests that at an early stage an investor will allocate more wealth to housing and then allocate a higher proportion of resources to financial assets at later life stages. using the euler equation, bakshi and chen provided baseline estimates for determining how risk aversion and investor “consumptionportfolios” can be measured for individuals of all ages and across diverse cultural environments. they noted that when the population ages, aggregate demand for financial investments rise and demand for housing declines. one conclusion from their work was that changes in someone’s demographic profile can bring about fluctuations in asset demand. 3.3. social support and country of origin factors cross-cultural frt has emerged over the last 20 years as a niche area of interest among those who study frt. bontempo, bottom, and weber (1997) observed patterns across four different countries. they concluded that uncertainty avoidance in a country may influence risk perceptions. many other studies using international comparisons have observed differences between the united states (or western europe) and asian countries, notably china (fan and xiao, 2005; hsee and weber, 1999; tan, 2011; wang and fischbeck, 2004). findings from these studies have generally indicated that the chinese are more risk seeking in financial arenas but not necessarily across other domains of risk. kim, chatterjee, and cho (2012) looked at the differences in asset ownership of asian immigrants from many different countries including china. they found a strong relationship between country of origin and the holdings of different asset classes, including homeownership, equities, and business ownership. rieger, wang, and hens (2014) presented a comprehensive evaluation of international risk taking in their article. rieger et al. documented the risk preferences of individuals in 53 countries. they reported that individuals across cultures are, on average, risk averse regarding gains and risk seeking with losses. this finding was in line with the propositions found in prospect theory (kahneman and tversky, 1979). rieger et al. also noted that risk preferences appear to be dependent on economic conditions and cultural factors. it was suggested that their results may serve “as an interesting starting point for further research on cultural differences in behavioral economics” (p. 637). two other large-scale international assessments of frt were conducted by statman (2008) and vieider, chmura, and martinsson (2012). studying 22 and 30 countries, respectively, the findings from these studies showed that those from wealthy countries tend to be more risk averse in financial domains. statman explained that, “people in low income countries have high aspiration relative to their current income” and they “pay with risk for a chance to move up in life” (p. 44). the findings of vieider et al. showed a unique relationship between international socioeconomic variables and risk-seeking behavior. they reported a strong negative correlation between frt and personal income. they explained the 321s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 phenomenon by suggesting that risk attitudes act as a transmission mechanism for growth by encouraging entrepreneurial activities throughout the world. when viewed from the perspective of the cushion hypothesis, country of origin variables become important because each country has a unique social support policy. it is possible that countries with generous social support systems create a ‘cushion’ for risk takers who fail in the markets. if true, this ought to increase the willingness of those in these countries to take risk. on the other hand, a robust social support system may dampen frt based on signals that country residents need not take risk to gain financial stability. at this point, neither hypothesis has been fully explored in the literature. 3.4. stability of frt one of the least discussed notions within the frt literature is the likeliness and degree to which risk attitudes change over time (zeisberger et al., 2010). in this regard, roszkowski and his associates (roszkowski et al., 2009) concluded that intrapersonal consistency was stable over time but greater variability was associated with higher risk-tolerance scores. what remains to be discovered are the unique characteristics of individuals who show inconsistency in their frt scores across multiple assessments. the consistency of individual frt over time can be assessed and split into four distinct categories: (1) stability over time, (2) reactions to market conditions, (3) consistency across different dimensions of frt, and (4) consistency across different types of questionnaires (roszkowski et al., 2009). when looking at frt change over time, yao et al. (2004) surmised that if significant time trends are evident after controlling for biopsychosocial and environmental factors, the changes over time can be interpreted to be related to changes in attitudes toward risk, not changes because of other factors. yook and everett (2003), grable and lytton (2001), and yang (2004) each looked at the consistency of different risk questionnaires across time. in generally, they found that psychometrically valid assessment tools with published reliability estimates tend to, on average, generate repeatable scores, but that even with the most reliable instrument, changes in frt scores do occur among some test takers. the general theme of research regarding the intrapersonal consistency of frt across time is that the construct of frt is relatively stable but does show some fluctuation based on environmental factors. for example, zeisberger et al. (2010) noted that risk parameters appear quite stable for the majority of investors, but that it is possible for one-third of investors to exhibit significant instability over time. 4. methodology in an attempt to test the frt model (fig. 1), this study used a secondary dataset made available by finametrica pty ltd. the risk profiling database included information collected in the united states (us), united kingdom (uk), and australia (aus). the choice to retain data from each country was based on two factors. first, it was thought that the risk tolerance exhibited by citizens of each country might differ based on the macroeconomic conditions present in each locale. second, the use of multicountry data allowed for a test of the cushion 322 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 hypothesis. the data contained biopsychosocial and environmental information, as well as composite frt scores for individuals who completed multiple risk assessments. data collection began in january of 2010 and ended in december of 2014. the mean and median time period between tests was 805 and 763 days, respectively (sd � 388.74) or slightly more than two years. the time span provided a unique perspective on the global trends and distinctive macroeconomic environments that existed in the post global financial crisis period. table 1 shows the demographic profile of the sample based on age, education, income, household size, net worth, and gender. keep in mind that education, income, and net worth were measured using ordinal variables (variable coding is discussed later in this section). the sample size used in the regression (n � 4,983) was reduced because of missing data and modeling delimitations. with an average age of 57, the sample population was older than the mean global population, but this was not surprising based on the fact the sample was drawn from individuals seeking financial or investment guidance. average income fell into the $50,000 to $100,000 range, whereas the average net worth for respondents fell into the $250,000 to $500,000 range. the mean education level was the some college or trade school category. the sample was skewed slightly toward males who made up almost 55% of the sample. a unique feature of the dataset was that all respondents took multiple assessments over the course of several months or years. this unique aspect of the dataset allowed for a comparison of respondents at different points in time, which made possible the identification of unique attributes of respondents who exhibited a notable change in their risk-tolerance score (rts). the sample was delimited to include only those respondents who completed multiple assessments. table 2 shows the distribution of risk scores based on the initial risk-tolerance score (rts_1) and the follow up risk-tolerance score (rts_2) test dates. the variables were also coded by country (aus, uk, us). the finametrica scale was utilized across each of the three countries in the sample to create consistency and comparability across countries. because of a common language, translation and semantic issues represented less of a methodological issue in this study compared with other research projects measuring global risk attitudes where survey tools have been translated into multiple languages. minor adjustments to reflect regional dialects may have been used, but inconsistency across differing country boundaries was expected to be minor. the validity and reliability of the assessment tool has been verified in previous studies that have used the finametrica dataset. for example, when testing the validity of the measure, gilliam, chatterjee, and zhu (2010) reported a cronbach’s alpha of 0.89, suggesting a high degree of reliability for the assessment tool. an example of two of the questions used in the assessment includes: compared with others, how do you rate your willingness to take financial risk? 1. extremely low risk taker 2. very low risk taker 3. low risk taker 4. average risk taker 5. high risk taker 6. very high risk taker 323s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 7. extremely high risk taker how easily do you adapt when things go wrong financially? 1. very uneasily 2. somewhat uneasily 3. somewhat easily table 1 demographic profile of sample variable n percent of sample gender males 5,285 54.6% females 4,392 45.4% age 18–34 1,930 25.0% 35–54 1,930 25.0% 55–65 1,930 25.0% 65� 1,930 25.0% education did not complete high school 832 13.7% completed high school 707 11.6% trade or diploma 1,246 20.5% university degree or higher 3,298 54.2% marital status married (or in a de facto relationship) 5,174 83.2% unmarried 1,046 16.8% income (income from all sources) under $30,000 625 10.2% $30,000-$50,000 1,177 19.2% $50,000-$100,000 2,133 34.7% $100,000-$200,000 1,295 21.1% $200,000-$300,000 672 10.9% over $300,000 241 3.9% household size 0 2,180 36.2% 1 1,957 32.5% 2 859 14.3% 3 661 11.0% 4� 366 6.1% net worth under $10,000 46 0.8% $10,000-$25,000 31 0.5% $25,000-$50,000 55 0.9% $50,000-$100,000 116 1.9% $100,000-$150,000 297 4.9% $150,000-$250,000 966 15.9% $250,000-$500,000 2,055 33.8% $500,000-$1,000,000 1,460 24.0% $1,000,000-$2,500,000 735 12.1% over $2,500,000 322 5.3% country australia 1,762 18.2% united states 6,269 64.7% united kingdom 4,564 17.1% 324 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 4. very easily some of the advantages associated with the use of the finametrica system include the academic and theoretical manner in which the scale was conceptualized, wide professional and individual use, and simple to understand interpretations that help financial planners know how to allocate their client’s investments (finametrica, n.d.). 4.1. dependent variable 4.1.1. change in frt score frt, as defined by each respondent’s rts, was the primary outcome variable of interest. the assessment score was based on a 25-item scale that was aggregated to compute a composite risk score. ranging from 1 to 100, higher scores were indicative of having a higher frt. the mean and standard deviation of the initial test (rts_1) for the sample was 47.40 and 9.51, respectively. in addition to measuring overall composite frt scores, another aspect of the sample were matching data pertaining to changes in frt scores across time by individual respondent. the dataset contained an additional score for each respondent (rts_2). the mean and standard deviation for the rts_2 score was 48.10 and 9.61, respectively. overall, frt scores increased less than one point (0.63; sd � 6.13) from the initial test. with such a large sample, one would expect to see a selection of individuals who exhibited both extreme consistency in frt and others who had major fluctuations in their frt scores. the following mean deviation technique, as outlined by roszkowski and spreat (2010), was used to estimate large fluctuations as a way to isolate those with significant changes in frt: 1. subtract the reliability coefficient from 1.0. a. 1.0–0.89 � 0.11 2. calculate the square root of the estimate. a. sqrt(0.11) � 0.33 3. multiply the square root outcome by the test’s standard deviation to estimate the standard error of measurement (sem). table 2 demographic profile of the sample based on financial risk tolerance (frt) scores frt scores n % mean standard deviation min max rts_1 9,692 100.0% 47.40 9.51 14 93 rts_2 9,692 100.0% 48.00 9.61 15 95 aus_rts_1 1,762 18.1% 48.72 9.63 18 86 aus_rts_2 1,763 18.1% 48.92 9.49 16 87 uk_rts_1 6,269 64.3% 46.61 9.56 14 93 uk_rts_2 6,270 64.3% 47.38 9.69 15 95 us_rts_1 1,661 17.0% 49.18 8.81 18 83 us_rts_2 1,662 17.0% 49.73 9.15 21 84 rts � risk-tolerance score; aus � australia; uk � united kingdom; us � united states. 325s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 a. 0.33 * 9.51 � 3.14 4. estimate the 95% confidence interval by multiplying the sem by 1.96 (this is the approximate z score associated with 95% coverage within a normal distribution). a. 3.14 * 1.96 � 6.15 5. based on the test mean of 47.40, any test taker with a rts_2 score between 41.25 to 53.55 (41 to 54 rounded) was considered rts_stable. this methodological approach, based on the standard error of the mean, provided an estimate of how much variation was needed to confidently conclude that a significant change in a rts had occurred. if the difference in test scores between rts_1 and rts_2 dropped below the defined confidence interval, the respondent was placed into the rts_decrease category. if the difference in test scores between rts_1 and rts_2 rose above the defined confidence interval, the respondent was placed into the rts_increase category. again, by measuring respondents at two separate times, with months and/or years in between, and by combing information about time periods, biopsychosocial and environmental variables, macroeconomic factors, and social support, it was possible to draw conclusions about the unique properties of respondents who exhibited variability in their risk attitude. 4.2. independent variables six biopsychosocial and environmental variables were also recorded at the time of each initial test: age, income, net worth, gender, education level, and marital status. country of origin, time and date of initial response, and the date of the follow up survey were also measured. in addition to the information in the dataset, macroeconomic indicator variables were combined with each sampling unit based on the date of the initial survey. in an effort to understand what, if any, macroeconomic variables might influence an individual’s willingness to take risk, the combined dataset allowed for tests of the significance of global macroeconomic factors. three macroeconomic variables were included for each country: unemployment rate, quarterly gdp, and stock market performance. a fourth macroeconomic variable was included to account for global commodity prices. in addition to country specific macroeconomic variables, all countries were also combined to examine the broad global trends. a set of global variables were then used to measure overall and interaction effects on frt. although survey responses were collected daily, some of the global macroeconomic variables were released monthly or quarterly; therefore, the tests focused on these broader macroeconomic data points by matching data based on the date of the initial assessment. the macroeconomic variables were operationalized as follows: y united states gross domestic product (gdp): reported quarterly, the range of us gdp was measured using data from the bureau of economic analysis. the range of gdp from 2010 to 2015 was $14.7 trillion to $18.1 trillion, with a mean of $16.4 trillion. y australia gdp: measured in millions of us dollars, the total annual gdp ranged from 326 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 $1.34 trillion ($1.43 trillion aud) to $1.55 trillion ($1.65 trillion aud) with a mean of $1.45 trillion ($1.54 trillion aud). y united kingdom gdp: measured in us dollars, the chained volume measures were reported in trillions. the annual range of gdp from 2010 to 2015 was $2.53 trillion (£1.60 trillion) to $2.83 trillion (£1.79 trillion), with a mean of $2.67 trillion (£1.69 trillion). y united states unemployment rate: the us bureau of labor statistics produces a monthly account of individuals defined as the percentage of the labor force that is unemployed but actively seeking and willing to work. the estimate was used in this study. y australia unemployment rate: data from the australian bureau of statistics evaluating the monthly unemployment rate was used. the australian unemployment rate measures the number of people actively looking for a job as a percentage of the labor force. y united kingdom unemployment rate: data from the united kingdom office for national statistics were used based on the monthly unemployment rate (seasonally adjusted for all). the united kingdom unemployment rate is defined as individuals currently unemployed, but have actively been seeking work in the past four weeks and are available to begin a job within the next two weeks. y us stock market index: to obtain an idea of general equity market conditions, the composite standard & poor’s (s&p) 500 was used in this study. the s&p 500 is a market capitalization based index of the 500 largest companies listed on the new york stock exchange (nyse) or nasdaq. y australia stock market index: in april of 2000, the asx 200 became the primary investment benchmark for the australian market. the asx accounts for 70% of the equity market. the index contains the top 200 listed companies by way of floatadjusted market capitalism. the asx 200 index was used to measure the australia equity market (denominated in australian dollars). y united kingdom stock market index: the ftse 350 index is a market capitalization weighted stock market index composed of the largest 350 companies whose primary listing is based on the london stock exchange. the ftse 350 index was used to measure the uk equity market (denominated in british pounds). y global commodities index: although given less attention than equity markets, commodity markets are aggressively traded internationally and many countries (e.g., australia, saudi arabia, russia, and brazil) have commodity intensive domestic markets. the green haven continuous commodity index (cci) fund provides a broad based, diversified commodity basket that can be used as a proxy for commodity performance. the cci uses an index of 17 commodity groups including grains, energy, precious metals, cash, and government treasury securities. the trajectory of the global index was used as an indicator for the general supply, demand, and pricing of global commodity markets. although traded daily, a month average was calculated and matched with test score dates to provide a measure of commodity market activity. y composite gross domestic product: to obtain a global perspective on domestic productions’ relationship to frt, a weighted composite model was developed. using 327s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 weighted averages from the three countries represented in the sample, a global gdp variable was created. the formula below was used for the calculation: gdp � us_gdp us_gdp � uk_gdp � au_gdp � usgdp � ukgdp usgdp � ukgdp � augdp � ukgdp � au_gdp us_gdp � uk_gdp � au_gdp � aus_gdp y composite stock market index: in addition to a composite gdp measure, a global stock market index variable was created using combined market information from australia, united kingdom, and the united states. these data were matched, by date of the initial test, to each respondent’s data profile. these data, rather than a change variable, were used in subsequent analyses. other variables were also included in the analysis. to test the effects of initial outliers, a variable was created that separated individuals into categories based on their rts_1. if someone scored extremely low they were coded as low initial score, and if they scored extremely high they were given a high initial score notation. biopsychosocial and environmental factors were also included in the analysis. it is well known that many professional financial planners use biopsychosocial and environmental variables to predict and assess the frt of their clients (spitzer and singh, 2008). previous research has done a relatively thorough job describing the most popular biopsychosocial and environmental variables used by financial planners (grable, 1997; grable and joo, 1998; sung and hanna, 1996) that appear to be associated with financial risk tolerance. some of the most important of these factors were included in this study. each was measured as follows: y age: age was calculated using year of birth at the initial survey date. y income: income was measured using five categories: (1) under $30,000; (2) $30,000$50,000; (3) $50,000-$100,000; (4) $100,000-$200,000; and (5) over $200,000. y net worth: the data for net worth were coded using 10 distinct categories as follows: (1) under $10,000; (2) $10,000-$25,000; (3) $25,000-$50,000; (4) $50,000-$100,000; (5) $100,000-$150,000; (6) $150,000-$250,000; (7) $250,000-$500,000; (8) $500,000$1,000,000; (9) $1,000,000-$2,500,000; and (10) over $2,500,000. y gender: males were coded 1; females were coded 2. y education level: four levels of education were used to measure attained academic achievement: (1) less than high school; (2) completed high school; (3) trade school or some college; and (4) university degree or higher. y household size: household size was the count of all members (including children) in the household. y marital status: marital status was coded dichotomously. those who were married were coded 1, otherwise 0. 328 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 a measure of social support was included in the study. social support is a broad term that describes the aggregate level of transfers from government to individuals. social support can be measured many ways with differing levels of comparability. simply equating absolute numbers does not make sense globally when production, income, and consumption differ widely across regions. social support can comprise many different concepts or programs, including, but not limited to, socialized healthcare, secondary and/or university education, unemployment insurance, and supplemental retirement income. government transfers, as a percentage of gdp, produces a percentage statistic that allows for comparison across any set of countries worldwide. adding the social support variable in this study was done to provide a test of the cushion hypothesis. for the scope of this study, social support was measured by percentage of gdp based on the us oecds index, as shown in table 3. table 4 provides a descriptive summary of the dependent and independent variables used in this study (data for social support are shown in table 3). a mean value is shown when the data were recorded at the interval level. a median score is shown for categorical variables. 4.3. data analysis methodology the following statistical techniques were used in this study: correlation, probability distribution, and logistic regression analyses. after testing the individual variables for normality and potential multicollinearity, a multinomial logistic regression analysis was used to examine the relationships among the independent variables and changes in frt. specifically, the conceptual model was tested using a multinomial logistic linear regression with the dependent variable separated into three different binary categories: decrease in risk score, stable risk score, and increase in risk score. the model was used to evaluate the change of those whose rts decreased and those whose rts increased across time relative to respondents with stable scores. the results provided clarity to which, if any, variables uniquely influenced a respondent’s change in frt across time. 5. results the first step in the analysis involved testing for possible multicollinearity among the independent variables. this test was conducted using a correlation analysis. the associations between and among the biopsychosocial and environmental factors were not particularly table 3 social support by country country social support (% gdp) united kingdom 21.7% united states 19.2% australia 19.0% gdp � gross domestic product. 329s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 high. on the other hand, the correlations among some of the macroeconomic variables were quite high, as shown in table 5. as shown in table 5, worldwide gdp and investment markets were highly correlated during the period of analysis. the correlation between us gdp and uk and au gdp was 0.98 and 1.00, respectively. given the high correlations among these variables, composite variables based on each country’s data were created. the correlations among these new variables are show in table 6. unemployment and gross domestic product were correlated at almost –1.00. overall, the high degree of correlation, as defined as a coefficient over 0.70 (tabachnick, fidell, and osterlind, 2007) indicated a potential multicollinearity issue. because gdp tends to be the primary indicator of economic activity, this variable was chosen to be included in the model. to build the multinomial logistic model, study participants were split into three unique groups. the first split included respondents who exhibited a significant decrease in their rts (n � 938). the second split was based on respondents who exhibited a significant increase in their rts (n � 1,355). the third group included those with a nonsignificant change in their rts (n � 7,399). after separating out the groups, specific factors were identified to examine the differences associated with changes in frt, using the stable group as the reference category. table 4 descriptive summary of the independent variables variable n mean/median standard deviation min max rts_1 9,692 47.4 9.51 14 93 rts_2 9,692 48.1 9.61 15 95 � in rts 9,692 .63 6.13 �36 48 days between tests 9,692 805.0 388.70 0 1985 education 6,113 3.1 1.09 1 4 income 6,143 3.2 1.26 1 6 household size 6,023 1.2 1.28 1 9 net worth 6,083 7.2 1.49 1 10 age 7,722 57.8 11.30 18 93 gender 9,692 male 5,285 54.6 n.a. 0 1 female 4,407 45.4 n.a. 0 2 marital status 9,692 married 5,174 83.2 n.a. 0 1 us gdp 9,692 $15,741.6 657.23 $14,681 $17,914 aus gdp 9,692 $ 1,392.1 44.96 $ 1,326 $ 1,522 uk gdp 9,692 $ 2,611.8 53.11 $ 2,528 $ 2,823 us commodity 9,692 29.9 3.37 21 36 us market 9,692 $ 1,339.8 203.99 $ 1,031 $ 2,107 uk market 9,692 £3,079.3 258.18 £2,598 £3,862 aus market 9,692 aus$4,596.8 355.29 aus$4,009 aus$5,929 us unemployment 9,692 8.6% 0.89 5% 10% uk unemployment 9,692 7.9% 0.40 6% 9% aus unemployment 9,692 5.2% 0.27 5% 6% rts � risk-tolerance score; gdp � gross domestic product; us � united states; aus � australia; uk � united kingdom. 330 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 t ab le 5 m ac ro ec on om ic va ri ab le s co rr el at io n ta bl e u s g d p u k g d p a u s g d p u s m ar ke t u k m ar ke t a u s m ar ke t c om m od ity in de x u s un em pl oy m en t a u s un em pl oy m en t u k un em pl oy m en t u s g d p 1. 00 0. 98 1. 00 0. 91 0. 75 0. 35 � 0. 12 � 0. 98 0. 65 � 0. 45 u k g d p 1. 00 0. 97 0. 93 0. 80 0. 44 � 0. 1 � 0. 97 0. 70 � 0. 53 a u s g d p 1. 00 0. 89 0. 72 0. 31 � 0. 14 � 0. 98 0. 65 � 0. 41 u s m ar ke t 1. 00 0. 93 0. 66 � 0. 07 � 0. 93 0. 71 � 0. 66 u k m ar ke t 1. 00 0. 81 0. 04 � 0. 77 0. 60 � 0. 65 a u s m ar ke t 1. 00 � 0. 17 � 0. 41 0. 60 � 0. 80 c om m od ity in de x 1. 00 0. 16 � .0 60 0. 33 u s un em pl oy m en t 1. 00 � 0. 69 0. 54 a u s un em pl oy m en t 1. 00 � 0. 65 u k un em pl oy m en t 1. 00 u s � u ni te d st at es ; u k � u ni te d k in gd om ; a u s � a us tr al ia ; g d p � gr os s do m es tic pr od uc t. 331s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 table 7 compares the differences in scores between the respondents from rts_1 to rts_2. the overall distribution of changes in risk scores appeared normal. a correlation estimation was made between change in rts and days between tests. the test was conducted to evaluate if a longer (or shorter) time horizon between tests might have explained the likelihood of a shifting rts. the mean score change was 0.63, whereas the mean period between tests was 805 days. a small positive association was noted between the two variables (r � 0.02); however, the effect size was very small, with much of the association resulting from the large sample size. the result of the test confirmed that test scores generally increased over the period of analysis, but that the time gap between tests was not a particularly important variable in explaining this shift. table 8 show the results of splitting respondents into distinct categories based on a meaningful change between rts_1 and rts_2. respondents that had a significant decrease or increase in score over time, as measured by the standard error of mean technique, were separated from respondents who exhibited stable scores across assessments. almost 25% of respondents had a significant change in their rts. in addition, respondents who exhibited significant decreases consistently scored above the mean on the initial assessment, whereas respondents who had significant increases in frt had initial lower than average scores. the results of the multinomial logistical model are shown in table 9. the second and third columns of table 9 show the model comparing those with a decrease in frt to those whose score was stable. the last two columns in table 9 show the model comparing those with an increase in frt to those whose score remained stable. the results from the test provide insights into the change some individuals exhibited in their frt over time. relative to those whose rts did not change: table 6 simplified macroeconomic variables correlation table avg. gdp avg. mkt avg. commodity avg. unemp avg. gdp 1.00 0.66 �0.12 �0.97 avg. mkt 1.00 �0.09 �0.73 avg. commodity 1.00 0.17 avg. unemp 1.00 gdp � gross domestic product; mkt � market; commodity � commodity index; unemp � unemployment. table 7 comparison of initial and follow-up scores variable mean standard deviation standard error mean upper 95% lower 95% initial average score rts1 47.01 6.29 0.07 47.15 46.87 initial average score rts2 47.61 7.59 0.09 47.77 47.44 initial low score rts1 29.83 3.85 0.14 30.11 29.56 initial low score rts2 34.64 7.38 0.27 35.17 34.11 initial high score rts1 64.82 4.72 0.15 65.11 64.52 initial high score rts2 62.38 7.72 0.25 62.87 61.89 rts � risk-tolerance score. 332 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 y older respondents were more likely to be in the decrease category. y older respondents were less likely to be in the increase category. y those with more education were less likely to be in the decrease category. y those who lived in a country with high social support were less likely to be in the decrease category. y those who lived in a country with high social support were less likely to be in the increase category. y those who lived in a country with a high gdp were less likely to be in the decrease category. y those who lived in a country with a high gdp were more likely to be in the increase category. y when the market was initially high, respondents were more likely to be in the decrease category. y those with a low rts_1 score were more likely to be in the decrease category. y those with a low rts_1 score were less likely to be in the increase category. y those with a high rts_1 score were less likely to be in the decrease category. y those with a high rts_1 score were more likely to be in the increase category. y an interaction between gdp and social support was noted for those in the decrease category. y an interaction between gdp and gender was present for those not in the increase category. y an interaction between market and age was noted for those in the increase category. y an interaction between market and gender was present for those in the increase category. to summarize, the regression results provide insights into the unique attributes of individuals who exhibited a change in their frt across time. the following individuals were more likely to show a decrease in their frt: older respondents with less education, who lived in a country with lower social support and gdp with initially high market values. they were also more likely to have a lower initial rts_1 score. among those showing an increase in frt were younger respondents who lived in a country with lower social support and a higher gdp. they also had a higher initial rts_1 score. although not unexpected, it is noteworthy table 8 description of rts by change across time (n � 9,692) variable mean standard deviation standard error mean upper 95% lower 95% % of sample rts total test 1 47.44 9.51 0.10 47.63 47.25 100.0% test 2 48.06 9.61 0.10 48.26 47.87 100.0% rts stable test 1 47.66 8.98 0.10 47.87 47.46 76.3% test 2 47.81 9.02 0.10 48.01 47.60 76.3% rts increase test 1 42.55 9.93 0.27 43.08 42.02 14.0% test 2 53.33 10.13 0.28 53.87 52.79 14.0% rts decrease test 1 52.73 9.72 0.32 53.35 52.10 9.7% test 2 42.50 9.63 0.31 43.11 41.88 9.7% rts � risk-tolerance score. 333s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 that the direction of the effects for each of the independent variables (excluding social support) between respondents who exhibited a rts decrease and a rts increase showed an almost complete inverse relationship. it is worth noting that tests of those respondents who originally had an extremely low rts_1 score tended to report a higher rts_2 score relative to respondents who had stable scores on both tests. likewise, respondents who originally had an extremely high rts_1 score tended to exhibit a decrease in their rts_2 score relative to respondents who had a stable score on both tests. table 9 multinomial logistic model comparing rts decrease/increase to rts stable variable decrease in score increase in score increase b p-value increase b p-value intercept 6.189 0.000 0.194 0.888 age 0.010 0.043*** �0.018 0.000*** education level �0.090 0.046*** �0.048 0.204 income �0.073 0.128 0.033 0.394 household size �0.062 0.176 0.035 0.300 net worth �0.057 0.117 �0.017 0.573 social support �0.105 0.005*** �0.069 0.031*** commodity index �0.010 0.464 �0.014 0.249 gdp �0.001 0.000*** 0.000 0.061*** market 0.001 0.014*** 0.000 0.143 gender �0.110 0.292 0.082 0.348 married �0.103 0.423 0.025 0.824 low initial score 0.719 0.009*** �1.344 0.000*** high initial score �1.185 0.000*** 0.898 0.000*** gdp � age 0.000 0.961 0.000 0.993 gdp � gender 0.000 0.451 �0.001 0.015*** gdp � education 0.000 0.541 0.000 0.597 gdp � income 0.000 0.378 0.000 0.783 gdp � married 0.000 0.920 0.000 0.655 gdp � household size 0.000 0.969 0.000 0.906 gdp � net worth 0.000 0.380 0.000 0.429 gdp � social support 0.000 0.039*** 0.000 0.290 market � age 0.000 0.166 0.000 0.080*** market � gender 0.001 0.106 0.001 0.021*** market � ed 0.000 0.393 0.000 0.972 market � income 0.000 0.535 0.000 0.582 market � married �0.001 0.338 �0.001 0.366 market � household size 0.000 0.572 0.000 0.460 market � net worth 0.000 0.562 0.000 0.349 market � social support 0.000 0.364 0.000 0.510 commodity � age �0.002 0.229 0.001 0.277 commodity � gender 0.037 0.187 0.044 0.101 commodity � education 0.011 0.377 0.006 0.617 commodity � income �0.007 0.598 0.010 0.411 commodity � married �0.041 0.271 0.013 0.713 commodity � household size �0.003 0.772 0.006 0.587 commodity � net worth �0.004 0.654 �0.011 0.258 commodity � social support 0.002 0.837 0.006 0.516 gdp � gross domestic product. n � 4,983: cox and snell (1989) for first model: 0.07; cox and snell (1989) for second model: 0.07. 334 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 6. discussion the principal purpose of this study was to identify biopsychosocial, environmental, macroeconomic, and social support variables associated with changes in frt across time. several noteworthy findings emerged from the analysis. in general, those who were older at the initial test date were more likely to exhibit a significant decline in their risk score. a similar result was noted for those with less formal education. an interesting find was that living in a country with high social support tended to reduce the migration towards either a decrease or increase on frt scores. living in a country with a high gdp was indicative of exhibiting an increase in frt scores. high market values at the initial assessment was predictive of a decrease in frt. the findings from this study can be incorporated into the practice of financial planning. one of the challenges many financial professionals face is the need to gain an understanding of a client’s feelings and attitudes validly and quickly during the data gathering phase of the financial planning process. rapport is often built over time, which makes it difficult to gain a full picture of an individual after a short introductory meeting or two. trying to assess different personality traits or tendencies is often accomplished through various assessments and, for better or worse, financial planner intuition. risk capacity is often examined once all relevant documents (e.g., cash flow, net worth, and insurance forms) have been reviewed, but accurately assessing personality attitudes and traits in a brief period of time is also necessary and, if accurate, helpful for both the client and the financial planner. to help a client allocate their investments, some form of frt assessment is needed. in addition to a basic risk assessment, financial planners also need to know if the information gathered will be relevant now and in the future. it is customary to have a client complete a frt assessment during the data intake process. other than an initial assessment, there are no rules that require any follow-up evaluations. being able to identify clients who are likely to show a frt change can be helpful for both financial planners and individuals assessing their own allocation decisions. findings from this study help financial planners determine approximately how “traited” frt is and what the characteristics are of individuals who may change over time. as shown here, individuals tend to exhibit generally stable frt scores, but as most financial planners know, household dynamics do change over time, which may cause this financial planning data input to change. in general, frt scores increased across the sample, but not enough to warrant a change in portfolio or other financial recommendations. among some respondents, a marked decrease or increase in frt scores was noted. the age of the test taker was an important predictor of change. older respondents were more likely to exhibit a decrease in their rts, whereas younger respondents were more likely to report a higher rts at a later date. another insight is that initial test scores were predictive of future scores. a rts outside the typical range provides an indication that a client may exhibit a meaningful change in his or her frt at some point in the future. if a client initially scores extremely high or extremely low, it may be useful to monitor that individual closely across time. in addition, any major changes to macroeconomic conditions may be an indicator that frt should be reassessed to ensure that portfolio recommendations still match a client’s needs and willingness to take risk. it should also be noted that any major, or potentially major, changes in social policy 335s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 around social retirement plans or national health insurance may influence the way an individual perceives risk. does frt change over time? that was, and still, remains one of the most important questions asked by financial planners, researchers, and policy makers. overall, frt, in this study, was relatively stable. frt did show some deviation across time, but for the majority of respondents, the initial rts changed very little. however, even if only a small portion of clients exhibit inconsistent frt scores, this can cause a problem in practice. in this study, approximatively 75% of individuals exhibited consistent scores across two assessments. so, hypothetically extrapolated, for a midsized firm with 200 clients over a five-year period, almost 50 clients could have significant changes in their frt scores. macroeconomic variables at the time of initial assessment, initial test scores, and social support all had a significant role to play in describing who was likely to exhibit significant a decrease or increase in their frt across two assessments. when interpreting the results from this study it is important to keep in mind that the macroeconomic, stock market, and commodity index variables were based on values when the first test was taken. a few studies have used change in market conditions or domestic production variables to forecast variations in frt scores, but this study used a baseline metric of the conditions present during the initial test. this methodological approach was applied for two reasons. first, the period in which the study was performed was a relatively stable period with generally favorable market conditions occurring after the global financial crisis. second, the applied nature of the study drove the decision. financial planners, when working with clients in developing investment recommendations within a financial plan, must use data at hand. they do not have access to preand postperiod macroeconomic data. the ability to describe potential variations in client frt requires the use of baseline inputs. even so, comparing the results presented here with future studies that use macroeconomic, biopsychosocial, and social support change data would be useful. it is also worth noting that while the results from this study are valuable in establishing baseline metrics for predicting changes in frt, the overall amount of explained variation in the dependent variable was relatively small. although different than residuals in a traditional linear model, cox and snell (1989) developed a methodology for determining the amount of explanation in a given logistic model. for the model tested in this study, the cox and snell coefficient was 0.071. this means the model explained about 7% of the effect for changes in frt scores over time. although not extremely large, the ability to show significant effects for different unique variables is a starting point to begin the discussion for future research about the exact reasons individuals change their willingness to take risk across time. references bajtelsmit, v. l., bernasek, a., & jianakoplos, n. a. (1999). gender differences in defined contribution pension decisions. financial services review, 8, 1–10. baker, h. k., & haslem, j. a. (1974). toward the development of client-specified valuation models. the journal of finance, 29, 1255–1263. bakshi, g. s., & chen, z. (1994). baby boom, population aging, and capital markets. journal of business, 67, 165–202. 336 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 baumeister, r. f., & tice, d. m. (1988). metatraits. journal of personality, 56, 571–598. benartzi, s., & thaler, r. h. (2007). heuristics and biases in retirement savings behavior. the journal of economic perspectives, 21, 81–104. bontempo, r. n., bottom, w. p., & weber, e. u. (1997). cross-cultural differences in risk perception: a model-based approach. risk analysis, 17, 479–488. cfp board. (n.d.). professional standards & enforcement. (available at http://www.cfp.net/for-cfp-professionals/ professional-standards-enforcement) chen, n., roll, r., & ross, s. a. (1986). economic forces and the stock market. journal of business, 59, 383–403. cox, d. r., & snell, e. j. (1989). analysis of binary data (2nd ed.). new york, ny: crc press., 1989. dahlbäck, o. (1991). saving and risk taking. journal of economic psychology, 12, 479–500. dalton, m., & dalton, j. (2004). fundamentals of financial planning. metairie, la: money education. fan, j. x., & xiao, j. j. (2005). a cross-cultural study in risk tolerance: comparing chinese and americans. ssrn working paper. (available at https://ssrn.com/abstract�939438) finametrica (n.d.). risk tolerance profiling. (available at https://www.riskprofiling.com/us) gilliam, j., chatterjee, s., & zhu, d. (2010). determinants of risk tolerance in the baby boomer cohort. journal of business and economic research, 8, 79–88. gollier, c., & zeckhauser, r. j. (2002). horizon length and portfolio risk. journal of risk and uncertainty, 24, 195–212. grable, j., & joo, s. (1998). does financial education affect knowledge, attitudes, and behavior? an empirical analysis. personal finances and worker productivity, 2, 213–220. grable, j. e. (1997). investor risk tolerance: testing the efficacy of demographics as differentiating and classifying factors. unpublished doctoral dissertation. blacksburg, va: virginia polytechnic institute and state university, 1997. grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business & psychology, 14, 625–631. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. grable, j. e., & lytton, r. h. (2001). assessing the concurrent validity of the scf risk tolerance question. journal of financial counseling and planning, 12, 43. hallahan, t. a., faff, r. w., & mckenzie, m. d. (2004). an empirical investigation of personal financial risk tolerance. financial services review, 13, 57. hariharan, g., chapman, k. s., & domian, d. l. (2000). risk tolerance and asset allocation for investors nearing retirement. financial services review, 9, 159–170. hsee, c. k., & weber, e. u. (1999). cross-national differences in risk preference and lay predictions. journal of behavioral decision making, 12, 165–179. irwin, jr., c. e. (1993). adolescence and risk taking: how are they related? in n. j. bell & r. w. bell (eds.), adolescent risk taking (pp. 1–19). thousand oaks, ca: sage. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica: journal of the econometric society, 47, 263–291. kim, j., chatterjee, s., & cho, s. h. (2012). asset ownership of new asian immigrants in the united states. journal of family and economic issues, 33, 215–226. morin, a. r., & suarez, a. f. (1983). risk aversion revisited. the journal of finance, 38, 1201–1216. motley fool. (n.d.). how do gold prices affect the economy? (available at http://www.fool.com/knowledgecenter/2016/03/05/how-do-gold-prices-affect-the-economy.aspx) pålsson, a.-m. (1996). does the degree of relative risk aversion vary with household characteristics? journal of economic psychology, 17, 771–787. rabbani, a. g., grable, j. e., heo w., nobre, l., & kuzniak, s. (2017). stock market volatility and changes in financial risk tolerance during the great recession. journal of financial counseling and planning, 28, 140–154. 337s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 rieger, m. o., wang, m., & hens, t. (2014). risk preferences around the world. management science, 61, 637–648. reinhart, c., & borensztein, e. (1994). the macroeconomic determinants of commodity prices. staff papersinternational monetary fund, 41, 236–261. roszkowski, m. j., & spreat, s. (2010). issues to consider when evaluating “tests.” in j. e. grable, k. archuleta, & r. r. nazarinia (eds.), financial planning and counseling scales (pp. 13–32). new york, ny: springer. roszkowski, m. j., & davey, g. (2010). risk perception and risk tolerance changes attributable to the 2008 economic crisis: a subtle but critical difference. journal of financial service professionals, 64, 42–53. roszkowski, m. j., delaney, m. m., & cordell, d. m. (2009). intrapersonal consistency in financial risk tolerance assessment: temporal stability, relationship to total score, and effect on criterion-related validity. journal of business and psychology, 24, 455–467. santacruz, l. (2009). effect of general economic mood on investor risk tolerance-implications for financial planning. jassa,1, 35–42. spitzer, j. j., & singh, s. (2008). shortfall risk of target-date funds during retirement. financial services review, 17, 143. statman, m. (2008). countries and culture in behavioral finance. paper presented at the cfa institute conference proceedings quarterly. (available at https://www.cfainstitute.org/learning/products/publications/ cp/pages/cp.v25.n3.6.aspx) sung, j., & hanna, s. (1996). factors related to risk tolerance. journal of financial counseling and planning, 7, 11–19. tabachnick, b. g., fidell, l. s., & osterlind, s. j. (2007). using multivariate statistics (6th ed.). new york: pearson. tan, j. (2011). cross-cultural risk behavior in financial decisions and the cushion hypothesis. unpublished senior theses, paper 168, claremont mckenna college. (available at http://scholarship.claremont.edu/cgi/ viewcontent.cgi?article�1134&context�cmc_theses) vieider, f. m., chmura, t., & martinsson, p. (2012). risk attitudes, development, and growth: macroeconomic evidence from experiments in 30 countries. wzb discussion paper. (available at http://www.ferdinandvieider. com/risk_development_growth.pdf) wang, m., & fischbeck, p. s. (2004). incorporating framing into prospect theory modeling: a mixture-model approach. journal of risk and uncertainty, 29, 181–197. wang, h., & hanna, s. d. (1997). does risk tolerance decrease with age? financial counseling and planning, 8, 27–32. weber, e. u., & hsee, c. (1998). cross-cultural differences in risk perception, but cross-cultural similarities in attitudes towards perceived risk. management science, 44, 1205–1217. west, t., & worthington, a. c. (2014). macroeconomic conditions and australian financial risk attitudes, 2001–2010. journal of family and economic issues, 35, 263–277. world bank. (n.d.). global economic monitor commodities. (available at http://databank.worldbank.org/data/ reports.aspx?source�global-economic-monitor-commodities) xiao, j. j. (2008). applying behavior theories to financial behavior. in j. j. xiao (ed.), handbook of consumer finance research (pp. 69–81). new york, ny: springer. yang, y. (2004). characteristics of risk preferences: revelations from grable & lytton’s 13-items questionnaire. journal of personal finance, 3, 20–40. yao, r., hanna, s. d., & lindamood, s. (2004). changes in financial risk tolerance, 1983–2001. financial services review, 13, 249–266. yook, k. c., & everett, r. (2003). assessing risk tolerance: questioning the questionnaire method. journal of financial planning, 16, 48–55. zeisberger, s., vrecko, d., & langer, t. (2010). measuring the time stability of prospect theory preferences. theory and decision, 72, 359–386. 338 s. kuzniak, j.e. grable / financial services review 26 (2017) 315–338 the behavior heuristics responsible for formation and liquidation of tax holding accounts matt hurst,a,* monica mendozab adepartment of finance, stetson university–school of business, 421 n. woodland boulevard, unit 8398, deland, fl 32723, usa bdepartment of accounting, stetson university–school of business, 421 n. woodland boulevard, unit 8378, deland, fl 32723, usa abstract this article proposes the tax liquidation hypothesis, a predictable pattern of behavior regarding individuals’ decisions to create and subsequently to liquidate “cash holding” accounts when facing tax liabilities. previous research on tax related trading has focused on minimizing the individual tax burden by holding winners and selling losers. this behavior, described as “optimal tax trading” suggests that individuals should sell stocks that have lost value in the short-term while holding onto stocks that have gained value until the stocks can be sold at the preferential long-term capital gains rate. this article proposes the tax liquidation hypothesis based on investor behavioral biases and the current tax environment. individual investors will hold “cash” accounts that are consistent with their preferences for risk and return. the cash account holdings may differ across individuals, but the pattern and hypotheses regarding formation and liquidation of these accounts for tax reasons should be consistent with the model proposed by this article. © 2016 academy of financial services. all rights reserved. 1. introduction this article examines the behavioral biases that lead individuals to create accounts with specific characteristics when facing sizeable tax liabilities and the subsequent behavior to liquidate these accounts at the time of tax filing. we call the pattern of selection and subsequent liquidation because of prevalent behavior heuristics the tax liquidation hypothesis. according to constantinides (1983), an investor who engages in optimal tax trading will hold a “cash account” for the purpose of tracking his or her net liability due at tax filing. every transaction, with taxable implications, will have an effect on the balance of this * corresponding author. tel.: �386-822-7428; fax: �386-822-7491. e-mail address: mhurst@stetson.edu (m. hurst) financial services review 25 (2016) 279–301 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. account, with subsequent gains causing an increase in holdings, while losses, which can be used to offset prior gains, will cause a decrease to net holdings. this behavior, described as “optimal tax trading” by constantinides (1983), focuses on minimizing the individual tax burden by holding winners and selling losers; individuals should sell stocks that have lost value in the short-term while, holding onto stocks that have gained value until the stocks can be sold at the preferential long-term capital gains rate. behavioral research across disciplines has shown that individuals do not behave rationally or act in such a way that always maximizes their own self-interest. shefrin and statman (1985) explain the disposition effect as the suboptimal investor behavior of selling winners too soon and holding on to losers too long. the primary explanations for the disposition effect are prospect theory as proposed by the novel prize-winning research of kahneman and tversky (1979) and mental accounting as proposed by thaler (1980), collectively referred to as ptma. under the ptma framework, individuals exhibit loss aversion, an asymmetric response to losses and gains of equal sizes, and focus too narrowly on individual security results as opposed to portfolio results. these biases influence investors to behave in a manner that violates optimal tax trading as described by constantinides (1983), and creates sizable tax liability due at filing. following the creation of a tax liability, as a result of portfolio turnover, an individual must make a decision on where to hold the proceeds until tax filing. it is likely that the choice of holding account varies according to individual risk preferences, investment horizon, financial sophistication, wealth, income, and relative magnitude of potential tax liability. furthermore, the behavioral biases present will also vary by individual as well as magnitude, but the extant literature suggests we may be able to detect evidence of this behavior in assets with specific characteristics. this article holistically examines behavioral biases as they pertain to asset selection and explores the potential effects these biases may have on security returns. the tax liquidation hypothesis proposes that the behavioral heuristics that have been shown to exist will lead to a consistently identifiable pattern in certain asset classes. confirmation of this hypothesis will require identification of likely holding accounts as well as a statistically significant relationship between return and volume characteristics and the deterministic variables. the intent of this research is to link the rules governing u.s. tax filing, and optimal tax trading with the behavioral characteristics that influence the individual selection of “tax holding accounts.” section 1 identifies the rules governing tax filing. section 2 reviews the theoretical optimal tax minimization strategy. section 3 briefly explains optimal tax trading. section 4 proposes the behavioral biases that cause investors to deviate away from optimal tax trading as well as the heuristics that influence cash account selection and liquidation. section 5 presents the hypotheses, summarizes the data and methodology, identifies real asset classes that closely fit the desired characteristics and a brief empirical examination. section 6 concludes. 2. tax timing the u.s. tax codes governing capital gains and losses provide value to an investor who is tax savvy and engages in optimal tax timing as described by constantinides (1983, 1984). these rules are summarized by brickley, manaster, and schallheim (1991) as follows: 280 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 1. capital gains and losses are recognized for tax purposes when they are realized, not as they occur. 2. the tax rate is higher for short-term capital gains and losses than for long-term capital gains. 3. (offset rule) if there are net short-term losses, they must be used to offset net long-term gains before any tax rate is applied. such offsets reduce the tax advantage associated with short-term losses that otherwise could be deducted directly from taxable income. 4. (wash sale rule) losses for tax purposes are not allowed if an asset is sold for a loss and repurchased within 30 days. 5. ($3,000 limit rule) capital losses can be used to deduct a maximum of $3,000 taxable income in any year. losses greater than $3,000 can be carried forward indefinitely. 6. (safe harbor rule) to avoid penalty a taxpayer must make minimum estimated tax payments equal to either 100% (110% for high income individuals) of the previous year’s tax liability, or 90% of the current year’s tax liability.1 from (2) and (3) it is clear that it is optimal to realize losses in the short run which are subject to the higher rates; from (6) it should be clear that it is optimal to prepay 100%2 in years with unanticipated liability and 90% otherwise. in years when there are large uncovered capital gains (gains not covered by losses) the remainder of the tax liability comes due in april. net short term capital losses offset long term capital gains on a dollar for dollar basis. therefore, an increase in the capital gains tax rate increases the amount of taxes that capital losses can offset (constantinides and scholes, 1980). capital gains tax rates also determine the amount of money to be set aside in a holding account for the purposes of paying the aforementioned liability. it is also worth noting that high net worth individuals must make estimated tax payments on a quarterly (monthly) basis in accordance with the rules for safe harbor. a rational investor expecting to have gains in the current year that are larger than the previous year’s gains will opt to pay quarterly payments equal to 100% of the previous year’s tax liability. alternatively, an investor who expects to have gains that are less than the prior year’s gains will opt to make payments equal to 90% of the current year’s tax liability. no rational investor will opt to pay the full amount of taxes, in essence prepaying taxes not due until the following year. this would violate time value of money as the investor would forego any possibility of earning additional income. furthermore, the interest penalty on untimely payments compels investors to make estimated payments; therefore, foregoing the option to delay all payments until the following tax year of concern.3 thus, the tax liquidation hypothesis proposes that the effect is larger in years when there is unanticipated tax liability. 3. theoretical optimal tax trading consider an investor whose liquid wealth, w, is fully invested in two assets a and b. he purchased both assets at the start of the calendar year and has a one-year investment horizon. he is rational and trades according to the strategy described in constantinides (1983, 1984). 281m. hurst, m. mendoza / financial services review 25 (2016) 279–301 under this scenario, the investor realizes capital losses in the short term, “the loss-realization option” (brickley, manaster, and schallheim 1991), and defers capital gains to the long term, at which time he realizes his capital gains to reestablish the short-term status, “the restart option” (dammon, dunn, and spatt 1989). at the end of the calendar year the asset a has gained $1,000 while asset b has lost $1,000. the investor would realize asset b on the last day of the year and postpone the realization of asset a until the start of the next calendar year. the loss in asset b, commensurate with the investor’s marginal tax rate, can be used to offset ordinary income; the gain, which was not realized in the same calendar year, has created a tax liability for the following year commensurate with the long-term capital gains rate. the investor then determines where and how to hold his realized capital gain in a “cash account” with additional capital gains and losses adjusting the level of cash set aside for tax purposes at any given time.4 without making generalizations about the investor’s risk preferences we can infer two things: first, time value of money forces the investor to reinvest his cash account. second, if he has already engaged in optimal tax selling, there is a positive benefit for him reinvesting in assets where he will be able to use additional tax loss selling. constantinides (1983) estimates the value of the tax-timing option as a fraction of each dollar invested in a security, if the investor fails to take advantage of capital losses. therefore, if an investor were to hold his anticipated tax liability in a money market account (with no chance of capital losses), he foregoes any possibility of tax-timing and, therefore, destroys value by not having that option. 4. the cognitive environment this article seeks to understand the decision-making process of individuals facing a tax liability, the decision to establish a holding account, and the subsequent payment of taxes or liquidation of the holding account. for there to be a tax liquidiation effect three cognitive conditions must be satisfied: 1. loss aversion. investors must exhibit a pattern consistent with the disposition effect. the absolute utility of a loss is greater than an equal gain: v(x) � v(-x). investors are more likely to realize gains than losses, thus creating tax liability. 2. a holding account that maximizes utility under mental accounting. investors choose a vehicle consistent with myopic loss aversion, regret aversion, and problems with self-control. 3. liquidation of the holding account. the ‘labeling’ of accounts reduces the substitutability between accounts, thereby causing an account created to keep track of tax liability to be the first liquidated for tax payment. 4.1. the failure of rational tax trading under prospect theory and mental accounting two fundamental and widely accepted cognitive processes are loss aversion and mental accounting. empirical research has shown that individuals behave differently with gains and 282 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 losses (barberis and xiong, 2009; frazzini, 2006; grinblatt and han, 2005; tversky and kahneman, 1979,1992). “prospect theory,” as proposed by kahneman and tversky (1979), and “mental accounting,” as proposed by thaler (1980), may explain the disposition effect. the disposition effect is the investor behavior of selling winners too soon and holding losers too long (shefrin and statman 1985). in prospect theory, utility is defined by gains and losses relative to a predetermined reference point. the utility function is convex in the area of losses and concave in the area of gains. benartz and thaler (1995) point out that empirical estimation of the ratio of the slopes in the two regions are approximately two, indicating that individuals are twice as sensitive to losses compared with gains. rational analysis suggests that if an investor needs to raise cash then the logical choice is to sell an asset that has declined in value. prospect theory and mental accounting, on the other hand, would suggest that individuals would rather sell an asset that has appreciated in value. these two cognitive biases lead investors to engage in loss aversion and narrow framing, looking at individual assets irrespective of the overall changes in portfolio value. odean (1998), using data that tracked the trades of investors with a large discount brokerage firm, found that individual investors were more likely to sell a stock if it had a gain as opposed to a loss. in a later article, barber and odean (2004) find evidence that even tax savvy investors realize gains more frequently than losses. the odean (1998) and barber and odean (2004) findings support mental accounting over rational analysis. furthermore, this finding is evidence that investors, while concerned about tax implications, behave in a manner that would generate sizeable capital gains liabilities. 4.2. establishing a cash account under mental accounting thaler and johnson (1990) describe the house money effect, a behavioral bias whereby money earned is more valuable than money won. the transformation of an investment from paper profits to realized gains seems to be a convenient time to reset the reference point thereby making a distinction between money earned and money won. the implication, for this article, is the implicit method by which an investor selects a reference point. a simple explanation is that an investor has a dynamic reference point that is evaluated over some interval. we suggest that the transformation for paper to real profits will reset the reference point, thereby reducing the house money effect because the money is now “real.” the selection of a tax holding account should then independent of prior asset performance; selection will be dependent on individual forward looking preferences. benartzi and thaler (1995) explain that the attractiveness of a risky asset depends on the time horizon of the investor. they refer to the combination of loss aversion and a short holding period as myopic loss aversion. they find that loss averse investors choosing between a risky asset (such as stocks) and a less risky asset (such as bonds or treasury bills) are more willing to accept the risky asset as the evaluation period, or horizon, becomes longer. additionally, using a set of parameters consistent with the representative decisionmaker, they find that the evaluation period necessary for an investor to be indifferent between the stocks and bonds is approximately one year. the investor who faces a capital gain liability necessarily has an evaluation period no greater than 16.5 months, and in most cases 283m. hurst, m. mendoza / financial services review 25 (2016) 279–301 less than one year.5 therefore, the selection of fixed income, as opposed to equity as an asset class for the cash account seems plausible for reinvesting the expected tax liability. narrow framing may echo investors’ considerations over non-consumption based utility, such as regret. regret, as explained by kahneman and tversky (1982), “is a special form of frustration in which the event one would change is an action one has either taken or failed to take.” the feeling of regret for errors of commission is stronger than for errors of omission. the regret of taking an action, and the choice being incorrect, is worse than failing to take the right action. therefore, the reinvestment in the holding account is likely to be less risky and more likely to avoid regret in the selection of asset choices. again this points to bonds that historically have less volatility than stocks. under the theoretical behavior of constantinides (1984) the investor sets aside his capital gains tax liability in an account designated for tax purposes and, therefore, may distinguish between this account and his general investment account, consistent with mental accounting. moreover, this may be a rational utility maximizing behavior if the investor has problems with self-control. shefrin and thaler (1981, 1988) propose that mental accounting is consistent with having a set of rules that may help with problems of self-control. there is a hierarchy of accounts, with certain accounts being “off limits” to avoid temptation. dividends also help discourage dissaving when self-control is an issue. shefrin and statman (1984) explain that mental accounting may justify an investor’s preference for dividends. dividends help investors segregate gains and losses and thereby increase utility. barberis and thaler (2003) show that investor’ utility with mental accounting is unambiguously higher with dividends. if we define value derived from income, v, the concavity of the utility function because of risk aversion necessitates v(2) � v(8) � v(10). regular coupons, such as those in bond funds would satisfy the preference for regular income. furthermore, regular dividend payments increase utility, compared to an equivalent lump sum, because of the time value of money. shefrin and statman (1984) argue that paying a dividend also helps investors avoid regret especially when self-control is low, as may be the case with the house money effect. dividends are preferred to capital because it circumvents the investor having to make a decision that they may later come to regret if the decision is “wrong” that is, if the asset they sold subsequently increases in value. the behavioral biases discussed up to this point suggest the cash account selected be low risk, dividend paying bond funds. additionally, the loss realization option discussed earlier implies additional investor utility for selecting closed-end funds were gains can be realized at the investor’s discretion and not automatically passed through by the fund manager. investors who are the most sensitive to taxes, and therefore, do the most to avoid paying them, have a preference for municipal bonds over corporate bonds ceteris paribus. the argument that rational investors should exhibit a tax-related dividend aversion does not apply to these funds because of the tax-free nature of municipal bonds. the relationship between investor sentiment and the holding of municipal bond funds should be evident. municipal bond funds pay a tax-free dividend and are consistent with myopic loss aversion, regret aversion, and reduce problems associated with a lack of self-control. therefore, loss-aversion may be a simple but rational reason for investors to choose municipal bond closed-end funds as their cash account. 284 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 4.3. liquidating the cash account under mental accounting shefrin and statman (1994, 2000) explain “behavioral portfolio theory” as an investor having separate mental accounts for different purposes. this phenomenon is supported by several experimental investigations that show people engage in narrow framing; paying attention to individual gains and losses rather than total changes. thaler (1999) explains the separation of sources and uses of funds into specific accounts as the second component of mental accounting. these mental accounts are considered independently and the covariation between the accounts is ignored. according to this narrow framing expenditures are grouped into categories for specific purposes, and investments are categorized similarly. badrinath and lewellen (1991) show most securities with losses are liquidated around the turn-of-the-year; it may be the case that investors are simply less willing to liquidate a stock holding, as opposed to a predesignated cash account, to cover tax liability early in the year. additionally, if the cash account is simply an accumulation of his gains in individual stocks, his decision of where to hold these gains may be driven by loss aversion or regret aversion. the choice to liquidate the cash account should be inseparable from liquidating other assets for tax related liquidity reasons. why then would any asset exhibit a systematic pattern consistent with tax liquidation? the answer proposed in this article is based on the psychological research that has given rise to behavioral finance. thaler (1999) explains that individuals tend to label accounts according to their usage. expenditures may be grouped to form budgets while wealth is allocated into accounts. there is a hierarchy of accounts such that individuals are more likely to spend current assets before future assets. a checking account is more likely to be used before an investor dips into a 401(k) (thaler, 1999). bracketing expenditures and assets by their uses and sources reduces the substitutability between accounts. people use resources differently depending on the labels that are given to each account. by creating an explicit budget individuals are exerting self-control. however, this means that once an account is established as say, “tax liability,” then it will be difficult to repurpose these funds for other uses. the household balance sheet would have a liability for current taxes as well as an asset, the cash account, and the hierarchy would suggest that this is the first place to look before using other sources. essentially, labels matter. 5. hypotheses the tax-liquidation hypothesis will present as joint hypotheses of lower returns and increased volume in the month of march. tax-loss selling by definition reduces the amount of taxes owed in a given year.6 in years with greater tax-loss selling, it is not necessary to liquidate as much to cover tax liability and therefore higher returns are expected. hypothesis 1: there is a positive relationship between march returns and the previous years’ tax-loss selling as measured by abnormal volume. the capital gains tax rate, �, determines the proportion of capital gains that must be set aside to cover tax liability. since closed-end funds have a finite number of shares, when the 285m. hurst, m. mendoza / financial services review 25 (2016) 279–301 capital gains tax rate is higher we would expect a greater proportion of the total funds to be held for tax purposes. therefore, the magnitude of the tax-liquidation hypothesis will be greater when the capital gains tax rate is higher. hypothesis 2a: march returns will be negatively related to capital gains tax rates. hypothesis 2b: march volume measures will be positively related to capital gains tax rates. taxes on unanticipated capital gains can be deferred until filing in april. it is in the tax payer’s best interest to have this money invested as long as possible. therefore, the tax-liquidation hypothesis will be greater when unanticipated capital gains are larger. hypothesis 3a: march returns will be negatively related to unanticipated liability. hypothesis 3a: march volume measures will be positively related to unanticipated liability. 5.1. data data on security prices, shares outstanding, monthly volume, and monthly returns (including and excluding dividends) comes from the center for research in security prices (crsp) monthly stock files. data on security prices, shares outstanding, daily volume, and daily returns (including and excluding dividends) comes from the crsp daily stock files. the sample consists of 168 municipal bond closed-end funds over the 1987 to 2009 time period.7 this article also looks at a sample of non-municipal closed end funds that trade in the same crsp share code(s). crsp share codes are two digit codes that describe the type of security traded. the securities in this article are limited to share codes 14 and 44. the first digit describes the type of security, 1 indicates ordinary common shares, and 4 indicates sbis (shares of beneficial interest).8 the second digit, 4, provides further information that both of these securities are limited to closed-end funds. approximately 66% of funds have crsp share code 14, the remainder have crsp share code 44. the defining characteristic of the funds in crsp codes 14 and 44 is that they are closed end funds that trade on organized exchanges and have uncharacteristically large dividend yields. the average dividend yield is 9.3% per annum over the sample period. on average 9.8 million shares are traded per month. there is a 61% correlation between the average return for municipals and non-municipals in the sample. the summary statistics, reported in table 1, show that the average size in closed end funds has been decreasing through time as the number of funds increased. the change in total value follows the rapid growth and subsequent decline in the number of funds reported in table 1. moreover, from table 1 and fig. 1, we can see the net asset value of the funds, and the number of existing funds follows a pattern that is consistent with the capital gains tax rate changes. the variable of interest, march_returnit, is defined as the returns in the month of march for fund i in year t. the average monthly march return excluding dividends is -1.4% and in 286 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 table 1 descriptive statistics of fund size by year year 5% 25% median 75% 95% mean 1987 $68,266 $180,000 $251,875 $313,875 $1,481,359 $410,964 1988 $50,000 $94,433 $240,833 $444,375 $1,406,353 $351,008 1989 $34,781 $101,250 $211,606 $395,109 $737,603 $299,479 1990 $33,745 $105,830 $185,084 $409,034 $740,759 $294,847 1991 $42,412 $119,170 $203,199 $422,018 $818,751 $311,860 1992 $46,468 $108,609 $188,891 $338,700 $810,294 $279,736 1993 $39,161 $79,512 $141,678 $285,229 $700,638 $234,365 1994 $29,294 $65,327 $128,305 $246,498 $610,256 $203,789 1995 $26,819 $73,650 $149,245 $265,332 $610,301 $215,759 1996 $32,228 $80,790 $160,369 $281,232 $632,618 $229,490 1997 $34,554 $85,121 $164,304 $288,585 $657,765 $239,926 1998 $37,537 $94,246 $168,768 $300,165 $688,365 $248,664 1999 $34,217 $86,765 $160,008 $269,902 $615,480 $224,753 2000 $29,859 $82,095 $147,874 $246,948 $543,165 $204,512 2001 $33,443 $93,627 $171,670 $275,945 $610,246 $234,330 2002 $35,032 $99,979 $176,518 $302,655 $628,480 $248,285 2003 $34,336 $113,535 $182,016 $307,937 $642,584 $254,163 2004 $34,436 $111,322 $181,125 $306,275 $638,152 $252,013 2005 $38,425 $117,276 $185,543 $316,315 $658,978 $658,978 2006 $39,038 $121,109 $186,940 $327,352 $684,099 $264,883 2007 $40,272 $121,257 $186,687 $313,058 $691,640 $260,130 2008 $37,777 $105,118 $164,066 $281,065 $597,404 $230,771 2009 $31,790 $106,078 $164,320 $278,439 $608,199 $237,400 notes: this shows the descriptive statistics of market cap for end funds that exist in a given year for which data was available. there are 168 funds and 22 years of data available. dollar amounts are in 1,000s. fig. 1. total asset value of muni bond closed-end funds by year. the figure shows the cumulative value of all existing municipal bond closed-end funds in the sample with data available during the 1987–2009 time period. the percentages depict the top marginal tax rate for capital gains. before 1997 capital gains were taxed as ordinary income. 287m. hurst, m. mendoza / financial services review 25 (2016) 279–301 a one-sided t test the variable is statistically different from zero. the dependent variable used for the empirical tests is the excess return in march. this is calculated as the current year’s march return less the prior year’s average monthly return excluding january and march. fig. 2a shows the average monthly return excluding dividends for each calendar month. the return in january is consistent with the tax-timing explanation given by starks et al. (2006). fig. 2b compares the average monthly return for the sample to the average monthly return for all other funds in the same crsp share code and to the monthly return for the crsp value weighted index. fig. 2. (a) average monthly return of the municipal bond funds for the 12 calendar months. the figure shows the average return excluding dividends across all municipal bond closed-end funds with data available for each month during the 1990–2009 time period. (b) average monthly return of the municipal bond funds, nonmunicipal funds, and the center for research in security prices (crsp) equal weighted return. this figure shows the average return across all municipal bond and non-municipal closed-end funds with data available for each month during the 1990–2009 time period. 288 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 the primary covariant is a variable that we have constructed using the data from the internal revenue service’ statistics of income. the variable, unanticipated liability, a measure of net capital gains adjusted for capital losses and income tax due at filing.9 unanticipated liability is calculated as the percentage of gains that would be uncovered by 100% (110%) payments of safe harbor liability under the tax code. because of the nature of tax filing, unanticipated liability will either be positive or nonexistent. it is for this reason that the variable is bound on the lower side by zero. in the event that capital gains are lower in time period (t) than in time period (t-1) the rules governing safe harbor would eliminate the possibility of an additional tax liability due in april. the variable for capital gains tax rates is defined as the prevailing tax rate for long-term capital gains in a given year. the irs also provided penalty rates for underpayment. the federal funds rates come from the federal reserve. the capital gains tax rates are available from several websites and were cross checked for accuracy. over the sample period the average per capita capital gains for all individuals was approximately $4,000 compared with $25,800 for filers in the top 10%. the unanticipated tax year over year for all filers was 5.6% compared with 10.5% for filers in the top 10%. the approximate tax due at time of filing on a per capita basis was $560 for all filers and $4,600 for filers in the top 10%. 5.2. methodology we use panel data and maximum likelihood estimation techniques to show that march abnormal returns are positively related to both the fund’s year-end volume and the average volume across funds. additionally, we propose that the tax liquidation hypothesis is more likely when capital gains tax rates are higher and when there are unanticipated capital gains and dividend income. the tax liquidation hypothesis is a joint hypothesis of lower returns and higher volume in the month of march. for robustness we have two models within each regression. model i uses turnover and model i uses volume ratios. the following regression equations are used to examine the tax liquidation hypothesis: march_returnit � retit�1 2;4�12 � �0 � �1year_end_volume_measureit�1 � �2average_year_end_volume_measureit � �3capital_gains_ratet � �4unanticipated_capital_gainst � �i volume_measureit � �0t � �1treturnit c � �2treturnit p � �3capital_gains_ratet � �4unanticipated_capital_gainst � �it 289m. hurst, m. mendoza / financial services review 25 (2016) 279–301 the hausman specification test rejects the null hypothesis that the efficient estimator is preferred.10 therefore, the remainder of the restricted regression models will be run using fixed effects and will control for clustering by year when appropriate. 5.3. the tax liquidation hypothesis in municipal bond closed-end funds to document the tax liquidation hypothesis in closed-end municipal bond funds the average monthly return is calculated across all funds. fig. 2 presents the average return by month for all funds.11 for all years the average in march (not including dividends)12 for all fund years was (1.4)% compared to an average of (0.1)% for the 10 months of the year not including january. this finding is primarily driven by the years before 2003. in 2003 capital gains tax was limited to a maximum of 15% for long-term gains and 35% for short-term gains. the abnormal return for march for the earlier time period was 1.7% compared with 0.3% in the later time period. the significant underperformance in march is supported by a simple time-series regression of cross-fund average returns on a march dummy variable. the results in table 2 show that march has significantly lower returns than the others months. in the first regression the difference is (1.6)% (significant at 1%). the second regression, which excludes the month of january from the estimation shows that the march returns are lower by (1.3)% (significant at the 5% level). for the negative abnormal returns to be attributable to a price pressure effect the underlying assets cannot be systematically affected by macroeconomic forces. there is an inverse relationship between bond prices and interest rates. fig. 3 graphs the average federal funds rate adjustment by month. it is clear from fig. 3 that changes in the short-term interest rate are not clustered in march. if it were the case that macroeconomic forces were the cause of the persistent negative returns in march we would expect to see clustering of rate hikes in march. the sample funds most likely have a duration which is more sensitive to interest rate changes and therefore a comparison to a vehicle with similar duration is in order. i replicate the fig. 2 using the return excluding dividends (capital appreciation) for long term table 2 panel regression of monthly returns on dummy variables monthly return �0.001(intercept)�0.016 (march) (0.50)*** (�2.64)*** radjusted 2 � 0.01 monthly return � �0.001(intercept)�0.013 (march) (�0.40) (�2.19)*** radjusted 2 � 0.01 notes: this table presents two regressions of monthly returns. model (1) presents the regression of monthly returns on a dummy variable for the month of march. model (2) is similar to the previous model except it drops the month of january. there are 168 funds and 22 years of data. all t statistics are based on the panel corrected standard errors (pcses), which adjust for contemporaneous correlation, autocorrelation, and heteroskedasticity (t statistics in parentheses). *indicates statistically significant at the 10% level. **indicates statistically significant at the 5% level. ***indicates statistically significant at the 1% level. 290 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 government bonds. ibbotson reports the historical monthly capital appreciation for intermediate as well as long term corporate and government bonds. fig. 4 shows the average monthly returns. there is a similar seasonality that exists in long term government bonds that appears to be unrelated to domestic tax filing.13 fig. 3. average federal funds rate adjustment by month. the figure shows the average monthly for the federal funds rate. the years available limit the period for this figure from 1990 to 2008. fig. 4. long term government bond capital appreciation by month. the figure below depicts the average monthly return for a single bond portfolio with a term of 20 years and a reasonably current coupon.17 291m. hurst, m. mendoza / financial services review 25 (2016) 279–301 5.4. march abnormal returns and abnormal year-end volume the tax-liquidation hypothesis assumes that investors trade according to the optimal tax strategy (constantinides, 1984), and sell losers to offset the maximum amount of gains. contrary to the optimal strategy discussed by constantinides (1984), badrinath and lewellen (1991) show that most sales of securities with capital losses take place in november and december. the findings of bharbra, dhillon, and ramirez (1999) give additional support to the year-end tax-loss selling hypothesis. the current article proposes a positive relationship between march returns and the turn of the year volume measure. the intuition for volume mitigating the tax liquidation hypothesis is that in years when an investor is able to offset more capital gains he reduces his tax liability and is thus required to sell less to cover his tax liability. the volume measures of the current article are defined identically to starks et al. (2006): turnoverit � average november and december trading volume of fund i in year t number of shares outstanding for fund i at the beginning of year t vol_ratioit � average november and december trading volume of fund i in year t average february to october trading volume of fund i in year t the first measure of volume, turnoverit, equals fund i’s average volume in november and december normalized by the number of shares outstanding. this measure is used to control for volume traded relative to number of shares across funds. the second measure, vol_ratioit, is used to compare the year-end volume relative to the average monthly volume within a fund. in addition to the two year-end volume measures we calculate the following two march volume measures for each fund: mar_turnoverit � march trading volume of fund i in year t number of shares outstanding for fund i at the beginning of year t mar_vol_ratioit � average november and december trading volume of fund i in year t average february to october trading volume of fund i in year t the march volume measures are defined analogously to the year-end volume measures. the tax-liquidation hypotheses proposes that we should see higher volume in the month of march relative to the other months. as shown in fig. 5, march has the highest monthly volume behind the turn-of-the-year effect months (october through january). this is consistent with the hypothesis of increased selling pressure driving down the returns in march. 292 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 the march volume ratio14 shows that on average, the volume in march is 10% greater than the average volume for february through october (12% excluding october). the abnormal monthly volume for march is significant at the 1% level. this was verified with a simple cross sectional regression of monthly volume on a dummy variable for march.15 the average return for closed-end municipal bond funds in the month of march is negative approximately twice as often as it is positive. as shown in fig. 6, the average return of closed-end funds in march are negative in 13 out of twenty years. the average return across funds is negative in only 5 of thirty years. in table 3a the march return is calculated in three separate ways: first for the calendar month, second from march 15th to april 15th, and finally for the seven days after the ex-dividend date to coincide with the theory that an investor should wait as long as possible before liquidating his holdings. the ex-dividend date for the month of april is on average the 13th. therefore, it is more likely to see the tax liquidation hypothesis in the month of march with liquidation occurring subsequent to the march ex-dividend date. the magnitude of the coefficients for turnover and volume_ratio decrease as expected. however, the test statistics increase dramatically. moreover, the r2 statistic increases substantially when the return period is redefined to coincide with optimal liquidation timing. this suggests that a sizeable proportion, 40% (60%) of the abnormal return in late march (the week after the ex-dividend date) is explained by the model. table 3b includes the municipal bond yield change calculated as: �yieldmarch � yieldfeb� 1 � yieldfeb table 3b shows that the variables for unanticipated liability and capital gains tax rates maintain sign and significance after accounting for the change in yield from the end of february to the end of march. the variable for the yield change, delta, is negative and fig. 5. average monthly volume of the municipal bond fund for the 12 calendar months. the figure shows the average volume across all municipal bond closed-end funds with data available for each month during the 1990–2009 time period. 293m. hurst, m. mendoza / financial services review 25 (2016) 279–301 significant. this is consistent with increases in yield leading to negative capital appreciation. the tax-liquidation hypothesis further implies that funds that have had positive returns in the previous year should experience significantly less liquidation selling in march. investors will delay realizing capital gains indefinitely, choosing to first liquidate funds that have performed poorly to take advantage of offsetting capital losses. table 4 presents the regression of march volume on current and previous year’s fund returns. it is reasonable to assume that both the current and prior year return will have an effect on which fund an investor chooses to liquidate. table 4 reports the results for march volume measures regressed on contemporaneous and prior year returns, the capital gains tax rate, and unanticipated liability.16 in model i, the coefficients for lagged returns are negative and significant at the 1% level. this reveals a fig. 6. a mean march return by year. (a) presents the average return excluding dividends for all funds in the sample for the time period of 1990–2009. (b) average return by year. (b) presents the average return excluding dividends for all funds in the sample for the time period of 1990–2009 excluding the months of january and march. 294 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 negative relationship between march turnover and past fund returns. additionally, the coefficient of determination shows that the models explain a non-trivial amount of the march volume. the relationship between a fund’s return and its liquidating volume is vital to showing that the most likely candidates for liquidation are the poorest performing funds. this is consistent with the optimal tax-timing strategy of constantinides (1983, 1984). the funds that have performed well over the past year will be the least likely to be sold for liquidity reasons, postponing any net capital gains until it is optimal. 5.5. march abnormal returns, volume, and the capital gains tax rate the effective capital gains tax rate on an asset depends on the realization of gains and losses on other assets. constantinides and scholes (1984) note that the rules in the tax code regarding capital gains and losses make the optimal liquidation policy non-separable across different assets. if an investor is able to offset more ordinary income with capital losses taken table 3 a: panel regression of march returns on volume measures, net capital gains rate, and unanticipated gains model (1): turnover coefficients model (2): vol_ratio coefficients coefficient estimates (t statistics in parentheses) march_returnit � retit�1 2;4�12 ��0��1year_end_volume_measureit�1��2average_year_end_volumeit�1 ��3capital_gains_ratet��4unanticipated_liabilityt�1��it panel a1: march monthly returns adjusted for previous feb., apr.–dec. returns on previous year’s volume measures, march volume measure, capital gains rate, and unanticipated liability year-end turnover 0.011 year-end volume ratio 0.006 (3.40)*** (4.78)*** average year-end turnover 0.103 average year-end volume ratio 0.023 (19.74)*** (14.25)*** capital gain rate �0.065 capital gain rate �0.046 (�23.80)*** (�19.49)*** unanticipated liability �0.044 unanticipated liability �0.043 (�12.89)*** (�12.05)*** intercept 0.09 intercept 0.065 r2 0.35 r2 0.36 panel b: march 15 to april 15 returns adjusted for previous feb., apr.–dec. returns on previous year’s volume measures, capital gains rate, and unanticipated liability year-end turnover 0.001 year-end volume ratio 0.008 (1.43) (5.71)*** average year-end turnover 0.020 average year-end volume ratio 0.023 (16.46)*** (113.04)*** capital gain rate �0.013 capital gain rate �0.042 (�4.72)*** (�16.36)*** unanticipated liability �0.019 unanticipated liability �0.034 (�4.64)*** (�8.96)*** intercept 0.105 intercept 0.075 r2 0.29 r2 0.33 295m. hurst, m. mendoza / financial services review 25 (2016) 279–301 in municipal bond funds then we would expect to see more tax-liquidation selling in years with higher capital gains tax rates. the results in table 3 show that the returns in march are negatively related to the capital gains tax rate (significant at the 1% level). this finding is robust to clustering by fund and controlling for prior year’s return and year end volume and average year volume measures across funds. the coefficient estimate of �0.05 predicts, at the margin, a unit increase in the natural log of the capital gains tax rate will result in a 5% percent drop in the returns during the month of march. this is economically significant given the historical rate for capital gains tax is between 15 and 33% with changes of between 8% and 15%. the results, presented in table 5, for abnormal march volume on the capital gains rate are equally revealing. the findings for march turnover (volume ratios) suggest that there is increased volume relative to shares outstanding (average monthly volume) when the capital gains tax rate is higher. the findings for abnormal volume measures are positive and significant at the 1% level. 5.6. march abnormal returns, volume, and unanticipated liability the results, reported in table 3, show that march abnormal returns are decreasing in unanticipated liability. this is paramount in proving that the tax liquidation hypothesis is related to tax liability liquidation. tax liability on unanticipated gains can be delayed until tax filing which is due april 15 or october 15 (if an extension is filed). furthermore, the table 3 (continued) panel c: seven day return following march ex-dividend date, adjusted for previous feb., apr.–dec. returns on previous year’s volume measures, capital gains rate, and unanticipated liability year-end turnover 0.001 year-end volume ratio 0.005 (7.54)*** (13.79)*** average year-end turnover 0.013 average year-end volume ratio 0.013 (40.38)*** (27.58)*** capital gain rate �0.017 capital gain rate �0.009 (�21.86)*** (�13.57)*** unanticipated liability �0.003 unanticipated liability �0.000 (�84.28)*** (�0.35) intercept 0.021 intercept �0.002 r2 0.51 r2 0.53 notes: this table shows the coefficients from regression of march returns, adjusted by the previous february through december (excluding march) average returns, on volume measures for year-end trading, contemporaneous march volume, the capital gains tax rate, and unanticipated liability. model (1) measures volume by turnover and model (2) measures volume by the volume ratio. in panel a the return is calculated using all trading days in march. in panel b the return is calculated using only the last 15 trading days of march and the first 15 of april. in panel c the return is calculated for the week following the ex-dividend date for the fund. there are 168 groups and 22 years of data. all t statistics are based on the panel corrected standard errors (pcses), which adjust for autocorrelation, and heteroskedasticity. *indicates statistically significant at the 10% level. **indicates statistically significant at the 5% level. ***indicates statistically significant at the 1% level. 296 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 coefficient of unanticipated_ncg (�0.011) suggests that a 100% increase in the population wide ncg predicts negative march returns of more than four percentage points. the mean unanticipated liability for the time period in this study was 21%. the results found in table 3 are, therefore, statistically and economically significant. furthermore, the measure of net capital gains from the statistics of income is a crude measure for the most tax sensitive investors who are more likely to hold municipal bonds. the existence of the tax liquidation hypothesis is more likely when the capital gains tax rate is higher and when unanticipated liability are higher. this is because of the rules in the tax code that allow for an investor to pay the minimum of 100% of the prior year’s tax liability or 90% of the current year’s anticipated tax liability. the results, reported in table 5, present the logistic regression of the existence of the tax liquidation hypothesis on the natural log of the capital gains tax rate and the percentage of capital gains in excess of the prior year’s capital gains. the limited dependent variable march_effectit takes a value of one if the return in march is negative and zero otherwise. the coefficients and z-statistics from table 4 show that both the capital gains tax rate and unanticipated liability are positive and significant at all conventional levels. the marginal effects are calculated as the partial derivative of the dependent variable with respect to the independent variable. the results table 3 b: panel regression of march returns on volume measures, net capital gains rate, and unanticipated gains model (1): turnover coefficients model (2): vol_ratio coefficients coefficient estimates (t statistics in parentheses) march_returnit � retit�1 2;4�12 ��0��1year_end_volume_measureit�1��2average_year_end_volumeit�1 ��3capital_gains_ratet�1��4unanticipated_liabilityt�1��it panel a1: march monthly returns adjusted for previous feb., apr.–dec. returns on previous year’s volume measures, march volume measure, capital gains rate, and unanticipated liability year-end turnover 0.012 year-end volume ratio 0.006 (2.56)*** (4.87)*** average year-end turnover 0.055 average year-end volume ratio 0.012 (8.85)*** (6.69)*** capital gain rate �0.001 capital gain rate �0.001 (�5.57)*** (�5.58)*** unanticipated liability �0.012 unanticipated liability �0.014 (�6.28)*** (�5.58)*** march muni yield � �9.22 march muni yield � �9.01 (�22.69) (�20.97) intercept �0.016 intercept �0.017 r2 0.41 r2 0.40 notes: this table shows the coefficients from regression of march returns, adjusted by the previous february through december (excluding march) average returns, on volume measures for year-end trading, contemporaneous march volume, the capital gains tax rate, and unanticipated liability. model (1) measures volume by turnover and model (2) measures volume by the volume ratio. in panel a the return is calculated using all trading days in march. in panel b the return is calculated using only the last 15 trading days of march and the first 15 of april. in panel c the return is calculated for the week following the ex-dividend date for the fund. there are 168 groups and 22 years of data. all t statistics are based on the panel corrected standard errors (pcses), which adjust for autocorrelation, and heteroskedasticity. 297m. hurst, m. mendoza / financial services review 25 (2016) 279–301 from table 4 show that tax-liquidation selling is more likely in years when the capital gains tax rate is high or when there are unanticipated liabilities, all else being equal. the results from the logistic regression are more significant when the average volume measures across funds are used instead of the volume measure of each fund. this indicates that the tax loss selling over all funds has a positive impact on which fund is more likely to be sold in march. table 4 panel regression of march volume measures on current year and previous year’s returns, net capital gains rate, and unanticipated gains model (1): turnover coefficients model (2): vol_ratio coefficients coefficient estimates (t statistics in parentheses) volume_measureit ��0t��1treturnit c ��2treturnit p ��3capital_gains_ratet��4unanticipated_capital_gainst��it panel a: march monthly volume adjusted for share outstanding or average monthly volume on current and previous year’s returns, capital gains rate, and unanticipated liability returnp �0.213 returnp �0.365 (�9.29)*** (�7.51)*** returnc �0.288 returnc �0.785 (�4.24) (�5.45) capital gain rate .001 capital gain rate 0.008 (2.34)** (6.30)*** unanticipated liability 0.047 unanticipated liability 0.079 (4.17)*** (3.30)*** intercept 0.25 intercept 0.668 r2 0.06 r2 0.04 panel b: march 15 to april 15 average daily volume adjusted for share outstanding or average monthly volume on current and previous year’s returns, capital gains rate, and unanticipated liability returnc 0.207 returnc 0.202 (2.21)** (4.18)*** returnp �0.288 returnp 0.009 (�3.23)*** (0.21) capital gain rate 0.038 capital gain rate �0.074 (0.83)) (�3.13)*** unanticipated liability 0.126 unanticipated liability 0.05 (3.35)*** (2.56)*** intercept 1.04 intercept 1.20 r2 0.26 r2 0.09 notes: this table shows the coefficients from regression of march volume measure, against contemporaneous and previous year’s returns, the capital gains tax rate, and unanticipated liability. returnc denotes the current year’s return through february and returnp denotes the prior year’s return of the fund. model (1) measures volume by turnover and model (2) measures volume by the volume ratio. in panel a the volume is calculated using all trading days in march. in panel b the volume is calculated as the average daily volume of the last 15 trading days of march and the first 15 of april. there are 168 groups and 22 years of data. all t statistics are based on the panel corrected standard errors (pcses), which adjust for autocorrelation, and heteroskedasticity. *indicates statistically significant at the 10% level. **indicates statistically significant at the 5% level. ***indicates statistically significant at the 1% level. 298 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 6. conclusion this article proposes the tax liquidation hypothesis based on investor behavioral biases and the current tax environment. individual investors will hold cash accounts that are consistent with their preferences for risk and return. the cash account may be different across individuals but the pattern and hypotheses regarding formation and liquidation of these accounts for tax reasons should be consistent with the findings of this article. we propose that having cash on hand and liquidating for tax purposes are not mutually exclusive events. these events can coexist harmoniously. a rational investor holds a portion of his wealth in cash that maximizes his utility over liquidity; given an out of pocket expenditure, that is, tax liability, he may unavoidably have to liquidate some portion of his invested wealth to rebalance his cash to investments ratio. notes 1 the safe harbor rule was not listed by brickley, manaster, and schallheim (1991) but was added here to explain unanticipated tax liability. table 5 logistic regression of march returns on volume measures, net capital gains rate, and unanticipated gains model (1): turnover coefficients model (2): vol_ratio coefficients coefficient estimates (z-statistics in parentheses) tlhit � �0 � �1year_end_volume_measureit�1 � �2average_year_end_volumeit�1 ��3capital_gains_ratet��4unanticipated_capital_gainst��it tax liquidation hypothesis on previous year’s volume measures year-end turnover �1.07 year-end volume ratio �0.56 (4.60)*** (�7.54)*** march turnover 1.23 march volume ratio 0.25 (2.66)*** (2.08)** capital gain rate 2.24 capital gain rate 2.47 (10.27)*** (11.05)*** unanticipated liability 1.04 unanticipated liability 1.08 (6.18)*** (6.35)*** marginal effects year-end turnover �0.25 year-end volume ratio �0.13 march turnover 0.29 march volume ratio 0.06 capital gain rate 0.53 capital gain rate 0.58 unanticipated liability 0.25 unanticipated liability 0.25 notes: this table presents the logistic regression estimates and marginal effects of the existence of the tlh on capital gains tax rates and unexpected liability. the dependent variable tlh takes a value of one if march returns are negative, and zero otherwise. the independent variable ln_ncg_tax_rate is the natural log of the contemporaneous capital gains tax rate; unanticipated_ncg is the current year’s capital gain as a percentage of the previous year’s capital gains. there are 168 funds and 22 years of data. *indicates statistically significant at the 10% level. **indicates statistically significant at the 5% level. ***indicates statistically significant at the 1% level. 299m. hurst, m. mendoza / financial services review 25 (2016) 279–301 2 high net worth individuals will have to pay 110% of the prior year’s tax. 3 the penalty for an underpayment of over $1,000 is 120 percent of the underpayment rate. the historical penalty rates are available from the irs at http://www.irs.gov/ pub/irs-pdf/n746.pdf 4 “cash account” is a term used by constantinides (1983) but does not mean that the account must be held in cash. 5 the period of 16.5 months comes from an investor selling an asset with a capital gain on january 1 and reinvesting the expected liability until the following april. 6 and subsequent years if there are carried forward losses. 7 we would like to thank laura starks, lwe yong, and lu zheng for providing me with the fund codes. 8 shares of beneficial interest resemble common shares—the primary difference is that an sbi gets issued by a “trust entity” instead of a company. 9 all amounts are in actual dollars. 10 the rejection of the null hypothesis that both the consistent and efficient estimators are acceptable was significant at a 1% level. the primary reason for the rejection was the difference in the firm specific volume measure. 11 fig. 5 graphs the march returns and the average returns for all other months by year to show the consistency of the tax liquidation hypothesis through time. 12 the graph for returns including dividends is similar to the one presented in fig. 2. 13 this article tests the hypothesis using controls for both the long term government bond capital appreciation as well as changes in the municipal bond yields. 14 the march volume ratio is the volume in march divided by the average monthly volume for the months february to october. 15 the months of january, november, and december were not included 16 contemporaneous returns are the returns for january and february of the current year. 17 paraphrased from the ibbotson sbbi 2009 yearbook. references badrinath, s. g., & lewellen, w. (1991). evidence on tax-motivated securities trading behavior. journal of finance, 46(1), 369–382. barber, b. m., & odean, t. (2003). are individual investors tax savvy? evidence from retail and discount brokerage accounts. journal of public economics, 88, 419–442. barberis, n. c., & thaler, r. (2003). a survey of behavioral finance. handbook of the economics of finance, volume 1b, financial markets and asset pricing, ed. george m. constantinides, milton harris and rené m. stulz, 1053–1123. amsterdam; london and new york: elsevier, north holland. barberis, n. c., & xiong, w. (2009). what drives the disposition effect? an analysis of a long-standing preference-based explanation. journal of finance, 64, 751–784. benartzi, s., & richard, h. t. (1995). myopic loss aversion and the equity premium puzzle. quarterly journal of economics, 100, 73–92. bhabra, h., dhillon, u., & ramirez, r. (1999). a november effect? revisiting the tax-loss-selling hypothesis. financial mangement 28(4), 5–15. brickley, j., manaster, s., & schallheim, j. (1991). the tax timing option and the discount on closed-end investment companies. journal of business, 64, 287–312. 300 m. hurst, m. mendoza / financial services review 25 (2016) 279–301 constantinides, g. m. (1983). capital market equilibrium with personal tax. econometrica, 51, 611–636. constantinides, g. m. (1984). optimal stock trading with personal taxes: implications for prices and the abnormal january returns. journal of financial economics, 13, 65–89. constantidides, g. m., & scholes, m. s. (1980). optimal liquidation of assets in the presence of personal taxes: implications for asset pricing. journal of finance, 35, 439–449. dammon, r. m., dunn, k. b., & spatt, c. s. (1989). a reexamination of the value of tax options. review of financial studies, 2, 341–372. frazzini, a. (2006). the disposition effect and underreaction to news. journal of finance, 61, 2017–2046. grinblatt, m., & han, b. (2005). prospect theory, mental accounting, and the disposition effect. journal of financial economics, 78, 311–339. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–291. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–292. odean, t. (1998). are investors reluctant to realize their losses. journal of finance, october, 1775–1798. shefrin, h., & statman, m. (1984). explaining investor preference for cash dividends. journal of financial economics, 13, 253–282. shefrin, h., & statman, m. (1985). the disposition to sell winners too early and ride losers too long: theory and evidence. journal of finance, 40, 777–790. shefrin, h., & statman, m. (1994). behavioral capital asset pricing theory. the journal of financial and quantitative analysis, 29(3), 323–349. shefrin, h., & statman, m. (2000). behavioral portfolio theory. the journal of financial and quantitative analysis, 35, 127–151. shefrin, h., & thaler, r. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609–643. starks, l. t., yong, l., & zheng, l. (2006). tax-loss selling and the january effect: evidence from municipal bond closed-end funds. journal of finance, 61, 3049–3067. thaler, r., & johnson, e. j. (1990). gambling with the house money and trying to break even: the effects of prior outcomes on risky choice. management science, 36, 643–660. thaler, r. (1980). toward a positive theory of consumer choice. journal of economic behavior and organization, 1, 39–60. tversky, a., & kahneman, d. (1992). advances in prospect theory: cumulative representation of uncertainty. journal of risk and uncertainty, 5, 297–323. 301m. hurst, m. mendoza / financial services review 25 (2016) 279–301 academy of financial services officers president thomas coe quinnipiac university president-elect robert moreschi virginia military institute executive vice president-program duncan williams western carolina university vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president william chittenden texas state university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama halil kiymaz rollins college frances lawrence louisiana state university david nanigian california state university, fullerton tom potts baylor university janine scott massey university past presidents william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2016 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 25, no. 3, 2016 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2016 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president william chittenden texas state university president-elect thomas coe quinnipiac university executive vice president-program robert moreschi virginia military institute vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2016 swarn chatterjee university of georgia vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president lance palmer university of georgia editor, financial services review stuart michelson stetson university directors sherman hanna ohio state university halil kiymaz rollins college frank laatsch univ. of southern mississippi david nanigian the american college tom potts baylor university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake past presidents lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2014 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 23, no. 4, 2014 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2015 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. the relationship between time perspective and financial risk tolerance in young adults kenneth n. ryack, ph.d., cpa, cfp�a,*, aamer sheikh, ph.d., cpa, cbm, ccs, dabfa, cfc�, cgma�a adepartment of accounting, school of business, quinnipiac university, hamden, ct 06518, usa abstract this study examines the relationship between time perspective (tp) and financial risk tolerance (frt) in young adults. prior research suggests young adults should invest in riskier portfolios to maximize wealth accumulation for retirement. optimally, they will have a future tp and a high frt. the results of this study indicate that tp accounts for a significant amount of variance in frt. however, the relationship between tp and frt is not optimal. future oriented individuals exhibit lower frt and present oriented individuals exhibit higher frt. the paper concludes with discussion of the implications of these findings and suggestions for future research. © 2016 academy of financial services. all rights reserved. jel classification: d8; d9; d14; g11 keywords: financial risk tolerance; time perspective; time orientation; investment behavior; wealth accumulation; retirement portfolio 1. introduction it is a widely held belief among academics and finance professionals that young individuals with a long time horizon until retirement should maximize the size of their retirement nest egg by creating more risky retirement investment portfolios (embrey and fox, 1997; hanna and chen, 1997; sung and hanna, 1996). this is because more risky investment portfolios consisting primarily of stocks tend to outperform portfolios containing less risky * corresponding author. tel.: �1-203-582-6550; fax: �1-203-582-8664. e-mail address: kenneth.ryack@quinnipiac.edu (k. n. ryack) financial services review 25 (2016) 157–180 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. investments such as bonds and certificates of deposit (hanna and chen, 1997; siegel and thaler, 1997). in accordance with the theory of reasoned action (fishbein and ajzen, 1975), yao, sharpe, and wang (2011) posit that an individual’s financial risk attitudes affect their investment behavior and that behavior, in turn, affects wealth accumulation. the theory is supported by research that has shown investors with a higher risk tolerance are more likely to invest in equities over less risky investments, resulting in more significant wealth accumulation for retirement (e.g., bailey and kinerson, 2005; bajtelsmit, bernasek, and jianakoplos, 1999; hariharan, chapman, and domian, 2000; keller and siegrist, 2006; yuh and devaney, 1996). however, it is not enough for a young individual to simply have a strong financial risk tolerance. they must also be thinking about their future (i.e., have a future time perspective). thus, to build a successful retirement portfolio, a young person should possess both a strong financial risk tolerance and a future time perspective. financial risk tolerance and time perspective are two separate constructs that have received a fair amount of attention in the academic literature. while it seems logical that the two constructs both play an important role in investing behavior, there is a lack of research examining the relationship between them. the goal of the current study is to examine the impact of time perspective on financial risk tolerance in young adults after controlling for other commonly studied variables that have been found to affect financial risk tolerance. 2. literature review and hypotheses 2.1. financial risk tolerance financial risk tolerance (frt) can be defined generally as an individual’s willingness to accept uncertainty in financial settings where there is the possibility of a negative outcome (grable, 2000; grable, lytton, and o’neill, 2004; grable and roszkowski, 2007). in an investing context, frt reflects an investor’s comfort level with investment risk and market volatility (faff, mulino, and chai, 2008; grable and lytton, 1998; hallahan, faff, and mckenzie, 2003). thus, the higher one’s frt, the higher is the degree of uncertainty or investment volatility they are willing to bear. in contrast, financial risk aversion is a term used in the literature to represent the inverse of frt and it reflects a person’s unwillingness to take financial risk (anbar and eker, 2010; hallahan et al., 2003). the higher one’s risk aversion, the lower their comfort level with financial uncertainty and investment volatility. practitioners and scholars have long believed that frt is an important factor in investment behavior and the accumulation of wealth. investors with a higher frt are more likely to construct a portfolio containing a larger percentage of risky investments (i.e., equities), while risk averse investors are more likely to hold less risky assets such as bonds and certificates of deposit (e.g., bailey and kinerson, 2005; finke and huston, 2003; hariharan et al., 2000; keller and siegrist, 2006; weber, blais, and betz, 2002). since a portfolio containing more risky assets yields a higher return over the long run, investors with a higher 158 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 frt will therefore accumulate more wealth upon retirement (e.g., bernasek and shwiff, 2001; embrey and fox, 1997; hariharan et al., 2000; jacobs-lawson and hershey, 2005; jianakoplos and bernasek, 1998; yao et al., 2011). a large body of literature on frt and its determinants has accumulated over the years because of its importance in relation to individual investor behavior, retirement wealth accumulation, and personal financial planning. the discussion that follows focuses on only the three variables included in the current study—gender, income, and financial knowledge/ experience. note that while age and education level have also been found to affect frt in previous studies, they are held constant in the current study because the participants are all college freshmen and are predominantly in the same age group. 2.1.1. gender the research on frt has consistently found that females exhibit a lower frt than males. it includes studies spanning a period of roughly two decades that incorporate a variety of measurement instruments and a number of different types of samples both inside and outside the united states. these samples include professional financial advisors (e.g., hartog, ferrer-i-carbonell, and jonker, 2002; olsen and cox, 2001), college faculty and staff (gilliam and chatterjee, 2011; grable, 2000), college students (anbar and eker, 2010; antonites and wordsworth, 2009; powell and ansic, 1997; ryack, 2011; weber et al., 2002), adults nearing retirement age (hariharan et al., 2000; neelakantan, 2010), and the general public (e.g., faff, hallahan, and mckenzie, 2011; gibson, michayluk, and van de venter, 2013; hallahan, faff, and mckenzie, 2004; larkin, lucey, and mulholland, 2013; wong, 2011; yao et al., 2011). two main theories have been advanced to explain why females are more risk averse. one theory is that biological and evolutionary differences between the genders have resulted in men naturally engaging in more risk taking behavior (anbar and eker, 2010; olsen and cox, 2001; wong, 2011). the second theory suggests gender differences in risk taking behavior are related to cultural influences stemming from differences in the traditional societal roles of males and females (anbar and eker, 2010; faff et al., 2011; olsen and cox, 2001; wong, 2011). 2.1.2. income a popular theory advanced in the literature is that frt will increase with higher levels of income because individuals with higher incomes are more able to absorb any potential losses that may result from more risky investments (anbar and eker, 2010; grable and lytton, 1998; hallahan et al., 2004). this is supported by numerous studies that have found a positive relationship between income and frt. as with gender, the finding is robust across various measurement instruments, countries, and populations that include samples of professional financial advisors (hartog et al., 2002), college faculty and staff (grable, 2000; grable and joo, 2004), college students (anbar and eker, 2010; ryack, 2011), and the general public (e.g., faff et al., 2011; gibson et al., 2013; grable and lytton, 1998; hallahan et al., 2004; wong, 2011; yao et al., 2011). 159k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 2.1.3. financial knowledge and experience the literature suggests that investors with increased financial knowledge and experience should have higher levels of frt because such individuals have a better understanding of the uncertainties associated with taking financial risks, more confidence in their financial decisions, and better coping mechanisms to deal with financial uncertainties (grable and joo, 1997; yao et al., 2011). consistent with this theory, a number of studies find a significant positive relationship between frt and self-reported financial knowledge (antonites and wordsworth, 2009; gibson et al., 2013; grable and joo, 1997). grable and joo (2004) also find that higher financial knowledge scores from a 10-item objective measure are correlated with higher frt. in addition, ryack (2011) finds that college students who played a stock market game as part of a course in high school display higher levels of frt. 2.2. time perspective a variable that has received very little attention in the frt literature is time perspective (tp). also referred to as time orientation, tp is believed to be the result of cognitive processes that divide our experiences into past, present, and future temporal frames of reference (d’alessio, guarino, de pascalis, and zimbardo, 2003; nuttin, 1985; zimbardo and boyd, 1999). a bias toward one of these temporal frames of reference creates an individual differences variable that can affect a person’s every day behavior and decision making. for example, a person primarily exhibiting a present tp will tend to focus on the immediate results of their actions and care little about future consequences (strathman, gleicher, boninger, and edwards, 1994; zimbardo and boyd, 1999). an individual who is primarily future oriented will instead concentrate on the achievement of future goals, focusing on how their current behavior affects future outcomes (boniwell and zimbardo, 2004; d’alessio et al., 2003; strathman et al., 1994; zimbardo and boyd, 1999). present tp and future tp are separate constructs and research has found that measures of the two different constructs are not correlated (boniwell and zimbardo, 2004; zimbardo and boyd, 1999). similar to the frt research, the tp literature has also explored the impact of a number of different demographic variables such as gender and income. for example, prior tp research has found a positive relationship between future tp and income level (e.g., appleby et al., 2005; gonzalez and zimbardo, 1985; holman and silver, 2005). however, the results are mixed regarding gender. some studies show that females are more future oriented than males (e.g., gonzalez and zimbardo, 1985; keough, zimbardo, and boyd, 1999; zimbardo, keough, and boyd, 1997) and others find no relationship between gender and tp (e.g., epel, bandura, and zimbardo, 1999; hamilton, kives, micevski, and grace, 2003; harber, zimbardo, and boyd, 2003). there is an extensive body of literature examining the impact of tp on behavior across a variety of contexts. for example, researchers have found that more future (present) oriented individuals are less (more) likely to gamble (hodgins and engel, 2002; mackillop, anderson, castelda, mattson, and donovick, 2006; petry, 2001; toplak, liu, macpherson, toneatto, and stanovich, 2007) and to consume high levels of alcohol, drugs, and cigarettes (henson, carey, carey, and maisto, 2006; keough et al., 1999; 160 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 mackillop, mattson, mackillop, castelda, and donovick, 2007; petry, bickel, and arnett, 1998; vuchinich and simpson, 1998; wills, sandy, and yaeger, 2001). future (present) tp is also positively (negatively) linked with eating healthier, exercising, and participation in screening for hiv and cancer (dorr, krueckeberg, strathman, and wood, 1999; henson et al., 2006; levy, micco, putt, and armstrong, 2006; luszczynska, gibbons, piko, and tekozel, 2004; ouellette, hessling, gibbons, reis-bergan, and gerrard, 2005; shores and scott, 2007). with the exception of the gambling behavior studies, research examining the role of tp in financial settings has been limited. however, that research does suggest a positive link between future tp and fiscally responsible behavior. for example, joireman, sprott, and spangenberg (2005) find business students with a stronger future orientation are more likely to use a financial windfall to pay down a credit card, contribute to savings, or cover college debt. in contrast, the more present oriented students prefer to purchase a sale item online or go on a trip with friends. results from webley and nyhus (2006) further indicate that future tp in adults is associated with their “economic socialization” (i.e., having been encouraged to have a bank account, having earned or been given money as a teenager, and having discussed financial affairs with their parents). hershey and mowen (2000) also find that future tp is positively associated with perceived financial knowledge, retirement involvement, and perceived financial preparedness. there appears to be only one study that has examined both tp and frt. results from jacobs-lawson and hershey (2005) suggest that higher levels of frt and future tp are predictive of saving for retirement. however, that study does not examine present tp or the direct relationship between tp and frt. 2.3. hypotheses on the predicted relationship between frt and tp despite its potential importance, research examining the relationship between tp and frt is lacking. in general, the time perspective research indicates people who are present oriented tend to be more likely to engage in risky activities, while those who are future oriented tend to avoid such activities. therefore, it seems plausible that future (present) tp will be associated with a lower (higher) frt. this leads to the following hypotheses. hypothesis 1: present oriented individuals are more likely to display higher levels of financial risk tolerance. hypothesis 2: future oriented individuals are more likely to display lower levels of financial risk tolerance. hypothesis 3: present time perspective will account for a large amount of variance in financial risk tolerance above and beyond other variables previously found to affect financial risk tolerance. hypothesis 4: future time perspective will account for a large amount of variance in financial risk tolerance above and beyond other variables previously found to affect financial risk tolerance. 161k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 3. procedure this study uses data previously collected by ryack (2011, 2012). the sample consists of 378 new freshmen that completed a research instrument while attending summer orientation sessions at a public university. descriptive statistics for the sample are presented in table 1 and discussed in the results section. the survey instrument includes various demographic questions as well as a number of items used to measure frt, present tp, and future tp. table 1 simple descriptive statistics variable n mean median frt score 340 26.3968 26.6984 ptp 374 5.7678 5.8889 ftp 378 5.4637 5.6111 sd minimum maximum frt score 4.4203 16.0000 40.0000 ptp 1.0065 2.2222 7.8889 ftp 1.0443 2.3333 8.0000 n % cumulative % gender female 215 56.88% 56.88% male 163 43.12% 100.00% total 378 100.00% income less than $20,000 16 4.49% 4.49% $20,000 to $39,999 35 9.83% 14.33% $40,000 to $59,999 75 21.07% 35.39% $60,000 to $79,999 80 22.47% 57.87% $80,000 to $99,999 58 16.29% 74.16% over $100,000 92 25.84% 100.00% total 356 100.00% stkgame no, did not play the stock market game 272 72.15% 72.15% yes, played the stock market game 105 27.85% 100.00% total 377 100.00% race white 328 88.65% 88.65% other 42 11.35% 100.00% total 370 100.00% financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. race is the self-reported race of the student. 162 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 responses to demographic questions about gender, estimated family income, and financial knowledge/experience are used as control variables since they have been found in previous research to have a strong impact on frt. the dependent variable, frt score, is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale (see appendix a). this scale has been utilized to measure frt in a number of prior studies and it has been demonstrated to be both a reliable and valid measure of frt when compared with other measures (gilliam, chatterjee, and grable, 2010; grable and lytton, 1999, 2001, 2003; kuzniak, rabbani, heo, ruiz-menjivar, and grable, 2015). the present and future tp independent variables are measured with eighteen items from the zimbardo time perspective inventory (ztpi; zimbardo and boyd, 1999). the ztpi consists of 56 items that make up five different tp scales: past-negative, past-positive, present-fatalistic, present-hedonistic, and future. these scales resulted from extensive factor analyses with college student samples and each scale represents a distinct factor that measures a unique type of tp. it has been common for researchers to use only the scales relevant to their particular study, with the future and present-hedonistic scales appearing most often in the research. items in the future scale reflect behavior associated with the planning for and achievement of future goals. the present-hedonistic scale items characterize an orientation toward present enjoyment and pleasure with a lack of consideration of future consequences. given that our hypotheses focus on the impact of present tp and future tp on frt, our instrument only includes items from the present-hedonistic and future ztpi scales. consistent with ryack (2012, 2015), we measure present tp with a subset of nine items from the ztpi present-hedonistic scale and we measure future tp with a subset of nine items from the ztpi future scale (see appendix b). six items contained in the original ztpi present-hedonistic scale and four items contained in the original ztpi future scale are excluded in an effort to create more parsimonious scale measures. most of the items excluded had a negative factor loading, a factor loading below 0.40, or were redundant with other scale items presented in the original study by zimbardo and boyd (1999). ryack (2012) tests the modified nine item scales in confirmatory factor analyses with a college student sample and finds that the revisions do not significantly alter scale integrity. for each scale item, the participants rated how characteristic the statement was of him or her on an eight-point scale ranging from extremely typical to slightly typical on one end (i.e., rating of 1– 4) and slightly atypical to extremely atypical on the other end (i.e., rating of 5– 8). before analyzing the data, these ratings were reverse scored so that a higher number indicates the statement is more characteristic of the participant. in other words, a rating of 1– 4 now represents extremely atypical to slightly atypical and a rating of 5– 8 represents slightly typical to extremely typical. this is done to make analysis of the impact of present tp and future tp (the independent variables) on frt (the dependent variable) easier to interpret because a higher frt score indicates the participant has a stronger financial risk tolerance. each participant’s tp score is calculated as the average of their ratings for the relevant scale. thus, a higher average score for the present tp scale items indicates the participant has a stronger present orientation and a higher average score on the future tp scale items indicates a stronger future orientation. 163k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 4. results 4.1. descriptive statistics and t-tests the descriptive statistics are presented in table 1. as previously noted, frt score is measured using participant scores on the grable and lytton (1999) 13-item financial risk tolerance scale (see appendix a). the lowest possible score is 13 and the highest possible score is 47, with a higher score equating to a higher risk tolerance. the mean frt score for the sample is 26.3968 with a range from 16.000 to 40.000. each participant’s tp score is calculated as the average of their ratings of the nine scale items that make up each scale (see appendix a). the lowest possible value for each scale is one and the highest possible value is eight, with a higher score indicating the participant has a stronger orientation toward the tp measured by the scale. the present time perspective (ptp) scores range from 2.2222 to 7.8889, with a mean of 5.7678. the future time perspective (ftp) scores range from 2.3333 to 8.0000, with a mean of 5.4637. gender is a [0,1] indicator variable with 0 coded as female and 1 coded as male. of the 378 students in the sample, approximately 57% are females and 43% are males. in addition, roughly 89% of the students are white, with the remainder belonging to various ethnic groups. the income variable reflects each student’s estimate of the combined income of their parents from all sources, before taxes. the mean student estimate of their parents’ combined income is in the range of $60,000 to $80,000, with approximately 26% estimating their parents’ combined income to be above $100,000. while ryack (2011) uses several variables to measure financial knowledge/experience among college students, he finds the most significant measure is completion of a stock market game as part of a high school course. therefore, that variable is used as the sole measure of financial knowledge/experience in the current study. stkgame is a [0,1] indicator variable with 0 coded as not having taken a high school course in which a stock market game was played and 1 coded as having taken such a course. roughly 72% of the students did not take a high school course in which a stock market game was played. t-tests of the differences in frt score, ptp, and ftp by gender and by stkgame are presented in table 2. there is a significant difference in frt score between females and males, with a mean frt score for the females of 24.9764 as compared to 28.1905 for the males (p � 0.0001). this result is consistent with a large body of prior research that finds females tend to exhibit lower financial risk tolerance as compared to males. there is no significant difference in ptp between females and males. however, the mean female score on the ftp scale, 5.7564, is significantly higher than the mean male score on the ftp scale, 5.0644 (p � 0.0001). the mean frt score for those individuals who have played a stock market game as part of a high school course is significantly higher than those individuals who have not played a stock market game as part of a high school course (mean of 27.8925 vs. 25.8463, p � 0.0001). this is also consistent with prior research that finds a positive relationship between frt and financial experience/knowledge. there is no significant difference in ptp or ftp between those who have played the stock market game and those who have not. 164 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 4.2. correlation analysis pearson bivariate correlations of all the variables are presented in table 3. as predicted by hypotheses 1 and 2, there is a significant positive correlation between frt score and ptp (r � 0.3519, p � 0.01), and a significant negative correlation between frt score and ftp (r � �0.3265, p � 0.01). consistent with prior research, the findings further show a significant correlation between frt score and the previously studied variables of gender, income, and financial knowledge/experience. however, there does not appear to be a significant correlation between race and frt score. thus, race is excluded from the multivariate analyses conducted below. males tend to display higher risk tolerance, frt score increases with estimated family income, and students that played a stock market game in high school exhibit a higher frt score. additionally, females tend to be more future oriented than males (r � �0.3294, p � 0.0001) and there is a negative relationship between income and future tp (r � �0.1675, p � 0.01). 4.3. ordinary least squares hierarchical regressions prior research has examined the impact of various demographic characteristics on frt scale scores using ordinary least squares (ols) hierarchical regressions (gibson et al., 2013; hallahan et al., 2004; ryack, 2011). consistent with this methodology, we run ols hierarchical regressions of the following form: step 1: frt score � f(gender, income, stkgame) step 2: frt score � f(ptp, gender, income, stkgame) step 3: frt score � f(ftp, ptp, gender, income, stkgame) table 2 differences in frt score, ptp, and ftp by gender and by stkgame frt score ptp ftp panel a: by gender female mean 24.9764 5.7341 5.7564 male mean 28.1905 5.8114 5.0644 difference �3.2140 �0.0773 0.6920 t-statistic �7.1073 �0.7316 6.7478 p-value 0.0000 0.4649 0.0000 panel b: by stkgame did not play stock market game mean 25.8463 5.7915 5.4737 played stock market game mean 27.8925 5.6993 5.4291 difference �2.0461 0.0923 0.0446 t-statistic �3.8680 0.7926 0.3699 p-value 0.0001 0.4285 0.7117 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. 165k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 where the dependent variable, frt score, is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. in the first step, we enter gender as a [0,1] indicator variable with 0 coded as female and 1 coded as male, income as the student estimate their parents’ combined income from all sources before taxes, and stkgame as a proxy for financial knowledge/experience. stkgame is a [0,1] indicator variable where 0 indicates the participant did not complete a high school course where they played a stock market game, and 1 indicates they did complete such a course. the three independent variables entered in step 1 have been examined in other studies and serve as control variables in the current study. the new independent variables of interest, ptp and ftp, are added in step 2 and step 3. ptp is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale, and is added as an independent variable in step 2. ftp is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale, and is added as an independent variable in step 3. table 4 presents the results of the ols hierarchical regressions. we check the ols regression assumptions for each of the step 1, step 2, and step 3 regressions by running the regcheck command in stata (greene, 1997; kennedy, 1998). specifically, this command checks for the assumption of homoscedasticity of the errors using the breuschpagan test (gujarati, 2014), the assumption of no multicollinearity using variance inflation factor (vif) values (studenmund, 2005), the assumption that the residuals are distributed normally using the shapiro-wilk test, the assumption that the model is correctly specified using the linktest (statacorp lp, 2015), and the assumption of table 3 pearson bivariate correlations frt score ptp ftp gender income stkgame race frt score 1 ptp 0.3519** 1 336 374 ftp �0.3265** �0.0241 1 338 374 378 gender 0.3615** 0.038 �0.3294** 1 338 372 376 378 income 0.2356** 0.0913 �0.1675* 0.2011** 1 323 352 354 355 356 stkgame 0.2068** �0.0412 �0.0192 0.0884 0.0152 1 337 371 375 375 354 377 race �0.0232 �0.0011 0.094 0.1032 �0.0919 0.0338 1 332 364 368 369 350 367 370 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. race is the self-reported race of the student. n, the number of observations, is shown under the correlation. * p � 0.05, two-tailed; ** p � 0.01, two-tailed. 166 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 appropriate functional form using ramsey’s regression specification error test (wooldridge, 2008). it also checks for influential observations using cook’s d statistic (pardoe, 2006). all regression assumptions are found to hold true for each of the step 1, step 2, and step 3 regressions (untabulated results). as predicted in hypothesis 1, ptp is a significant predictor of frt score in that individuals with a stronger present orientation are more likely to have a higher level of frt. consistent with hypothesis 3, the ptp variable accounts for a large amount of variance above and beyond the control variables as evidenced by an adjusted r2 increase from 0.172 in step 1 to 0.278 (an increase of 0.106 or 61.63%) in step 2. in support of hypothesis 2, the ols hierarchical regressions further demonstrate a significant negative relationship between ftp and frt score. in other words, individuals with a stronger future orientation are more likely to have a lower frt. the addition of the ftp variable increases the adjusted r2 from 0.278 in step 2 to 0.330 in step 3 (an increase of 0.052 or 18.71%), thus supporting hypothesis 4. a comparison of step 3 to step 1 demonstrates that the ptp and ftp variables together result in an adjusted r2 increase of 0.158 (an increase of 91.86%) over a model that only includes gender, income, and financial table 4 ols hierarchical regressions (standardized beta coefficients) step 1 step 2 step 3 frt score frt score frt score ftp �0.251 p-value 0.000 ptp 0.335 0.341 p-value 0.000 0.000 gender 0.295 0.286 0.202 p-value 0.000 0.000 0.000 income 0.178 0.148 0.110 p-value 0.000 0.000 0.016 stkgame 0.181 0.188 0.197 p-value 0.000 0.000 0.000 n 320 317 317 r2 0.179 0.287 0.340 adjusted r2 0.172 0.278 0.330 f-statistic 21.180 27.690 29.060 prob � f-statistic 0.000 0.000 0.000 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. this table presents results of the following ordinary least squared (ols) regressions: step 1: frt score � f(gender, income, stkgame) step 2: frt score � f(ptp, gender, income, stkgame) step 3: frt score � f(ftp, ptp, gender, income, stkgame). 167k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 knowledge/experience (gender, income, and stkgame). thus, we find evidence in support of all four hypotheses. 4.4. sensitivity analyses we run a variety of sensitivity analyses to ensure that our inferences remain qualitatively unchanged if we run regressions separately for certain groups, or if we change the method of estimation. 4.4.1. ols hierarchical regressions by group we begin our sensitivity analyses by running ols hierarchical regressions by gender (results reported in table 5) and by stkgame (results reported in table 6). in table 5, we run regressions of the following form by gender: step 1: frt score � f(income, stkgame) step 2: frt score � f(ptp, income, stkgame) step 3: frt score � f(ftp, ptp, income, stkgame) where the dependent variable and independent variables are defined in the previous section. for both females and males, ptp is a significant predictor of frt score, supporting hypothesis 1. consistent with the full model previously discussed (where gender is an independent variable), the ptp variable accounts for a large amount of variance above and beyond the control variables. this is evidenced by an adjusted r2 increase from 0.095 in step 1 to 0.204 in step 2 for females and from 0.030 in step 1 to 0.161 for males, thus providing support for hypothesis 3. as found with the full regression models, the ols hierarchical regressions by gender also demonstrate a significant negative relationship between ftp and frt score, supporting hypothesis 2. the addition of the ftp variable increases the adjusted r2 from 0.204 in step 2 to 0.306 in step 3 for females and from 0.161 in step 2 to 0.182 in step 3 for males, thus supporting hypothesis 4. a comparison of step 3 to step 1 demonstrates that the ptp and ftp variables together result in an adjusted r2 increase of 0.211 (an increase of 222.11%) for females and an increase of 0.152 (an increase of 506.67%) for males over a model that only includes only income and financial knowledge/experience (income, stkgame). an interesting finding is that income continues to be a significant explanatory variable for females but not for males. as in the full model, we find evidence to support all four hypotheses when we run hierarchical regressions separately by gender. we also run the ols hierarchical regressions separately for participants that did not complete a high school course with a stock market game and for participants that did complete such a course. in table 6, we run regressions of the following form by stkgame: step 1: frt score � f(gender, income) step 2: frt score � f(ptp, gender, income) step 3: frt score � f(ftp, ptp, gender, income) 168 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 where the dependent variable and independent variables are as previously defined. for both individuals without financial knowledge/experience and individuals with financial knowledge and experience, ptp is a significant predictor of frt score, supporting hypothesis 1. consistent with the full model discussed earlier (where stkgame is an independent variable), the ptp variable accounts for a large amount of variance above and beyond the control variables. this is table 5 ols hierarchical regressions by gender (standardized beta coefficients) step 1 step 2 step 3 frt score frt score frt score female gender � 0 ftp �0.329 p-value 0.000 ptp 0.343 0.321 p-value 0.000 0.000 income 0.253 0.205 0.165 p-value 0.000 0.003 0.011 stkgame 0.208 0.230 0.190 p-value 0.003 0.001 0.003 n 178 176 176 r2 0.105 0.218 0.322 adjusted r2 0.095 0.204 0.306 f-statistic 10.290 15.930 20.310 prob � f-statistic 0.000 0.000 0.000 male gender � 1 ftp �0.169 p-value 0.036 ptp 0.367 0.385 p-value 0.000 0.000 income 0.102 0.092 0.059 p-value 0.223 0.237 0.453 stkgame 0.179 0.166 0.198 p-value 0.032 0.034 0.012 n 142 141 141 r2 0.044 0.179 0.206 adjusted r2 0.030 0.161 0.182 f-statistic 3.200 9.980 8.800 prob � f-statistic 0.044 0.000 0.000 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. this table presents results of the following ordinary least squared (ols) regressions by gender: step 1: frt score � f(income, stkgame) step 2: frt score � f(ptp, income, stkgame) step 3: frt score � f(ftp, ptp, income, stkgame). 169k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 evidenced by an adjusted r2 increase from 0.148 in step 1 to 0.253 in step 2 for individuals without financial knowledge/experience and from 0.100 in step 1 to 0.221 for individuals with financial knowledge and experience, thus providing support for hypothesis 3. as found with the full regression models, the hierarchical regressions by stkgame also demonstrate a significant negative relationship between ftp and frt score, supporting hypothesis 2. the addition of the table 6 ols hierarchical regressions by stkgame (standardized beta coefficients) step 1 step 2 step 3 frt score frt score frt score did not play stkgame � 0 ftp �0.224 p-value 0.001 ptp 0.337 0.324 p-value 0.000 0.000 gender 0.304 0.309 0.214 p-value 0.000 0.000 0.001 income 0.203 0.156 0.125 p-value 0.001 0.009 0.032 n 229 227 227 r2 0.155 0.263 0.302 adjusted r2 0.148 0.253 0.289 f-statistic 18.460 25.750 24.740 prob � f-statistic 0.000 0.000 0.000 played stkgame � 1 ftp �0.341 p-value 0.000 ptp 0.357 0.428 p-value 0.000 0.000 gender 0.294 0.251 0.207 p-value 0.005 0.011 0.025 income 0.128 0.138 0.079 p-value 0.214 0.154 0.387 n 91 90 90 r2 0.120 0.247 0.353 adjusted r2 0.100 0.221 0.322 f-statistic 5.660 5.510 7.400 prob � f-statistic 0.005 0.002 0.000 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. this table presents results of the following ordinary least squared (ols) regressions by stkgame: step 1: frt score � f(gender, income) step 2: frt score � f(ptp, gender, income) step 3: frt score � f(ftp, ptp, gender, income). 170 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 ftp variable increases the adjusted r2 from 0.253 in step 2 to 0.289 in step 3 for individuals without financial knowledge/experience and from 0.221 in step 2 to 0.322 in step 3 for individuals with financial knowledge and experience, thus supporting hypothesis 4. a comparison of step 3 to step 1 demonstrates that the ptp and ftp variables together result in an adjusted r2 increase of 0.141 (an increase of 95.27%) for individuals without financial knowledge/experience and an increase of 0.222 (an increase of 222.0%) for individuals with financial knowledge/experience over a model that only includes only gender and income (gender, income). an interesting find is that income continues to be a significant explanatory variable for individuals without financial knowledge/experience but not for individuals with financial knowledge/experience. as in the full model, we find evidence to support all four hypotheses when we run hierarchical regressions separately by stkgame. 4.4.2. hierarchical truncated regressions given that the dependent variable is a scale measure with a limited possible range of 13–47, it can be argued that a truncated regression model would be more appropriate than an ols regression model (davidson and mackinnon, 1993; maddala, 1983). in table 7, we report the results of hierarchical truncated regressions using maximum likelihood estimation table 7 hierarchical truncated regressions step 1 step 2 step 3 frt score frt score frt score ftp �1.065 p-value 0.000 ptp 1.450 1.498 p-value 0.000 0.000 gender 2.578 2.507 1.779 p-value 0.000 0.000 0.000 income 0.488 0.411 0.306 p-value 0.001 0.005 0.031 stkgame 1.824 1.867 1.952 p-value 0.000 0.000 0.000 n 317 314 314 rough estimate of r2 0.179 0.287 0.340 wald �2 56.760 94.950 117.790 prob � �2 0.000 0.000 0.000 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. this table presents results of the following truncated regressions using maximum likelihood estimation with robust standard errors: step 1: frt score � f(gender, income, stkgame) step 2: frt score � f(ptp, gender, income, stkgame) step 3: frt score � f(ftp, ptp, gender, income, stkgame). 171k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 with robust standard errors (the inferences are qualitatively unchanged if these regressions are run without robust standard errors). a rough estimate of the r2 one would find in an ols regression is computed by correlating frt score with the predicted value and squaring the result. we run hierarchical truncated regressions of the following form where the dependent variable and independent variables are as defined earlier: step 1: frt score � f(gender, income, stkgame) step 2: frt score � f(ptp, gender, income, stkgame) step 3: frt score � f(ftp, ptp, gender, income, stkgame) consistent with hypothesis 1, ptp is a significant predictor of frt score. consistent with hypothesis 3, the ptp variable accounts for a large amount of variance above and beyond the control variables as evidenced by an increase in the rough estimate of r2 from 0.179 in step 1 to 0.287 in step 2 (an increase of 0.108 or 60.34%). in support of hypothesis 2, there is a significant negative relationship between ftp and frt score. the addition of the ftp variable increases the rough estimate of r2 from 0.287 in step 2 to 0.340 in step 3 (an increase of 0.053 or 18.47%), thus supporting hypothesis 4. a comparison of step 3 to step 1 demonstrates that the ptp and ftp variables together result in an increase in the rough estimate of r2 of 0.161 (an increase of 89.94%) over a model that only includes gender, income and financial knowledge/experience (gender, income, and stkgame). thus, we continue to find evidence in support of all four hypotheses when we run hierarchical truncated regressions that use maximum likelihood estimation. 4.4.3. ordinal logistic regressions it can also be argued that the nature of the dependent variable, frt score, is better captured by running an ordinal logistic regression model (davidson and mackinnon, 1993; greene, 1997; liu, 2016; maddala, 1983). in table 8, we report the results of hierarchical ordinal logistic regressions using maximum likelihood estimation with robust standard errors (the inferences are qualitatively unchanged if these regressions are run without robust standard errors). we run hierarchical ordinal logistic regressions of the following form where the dependent variable and independent variables are as defined earlier: step 1: frt score � f(gender, income, stkgame) step 2: frt score � f(ptp, gender, income, stkgame) step 3: frt score � f(ftp, ptp, gender, income, stkgame) consistent with hypothesis 1, ptp is a significant predictor of frt score. consistent with hypothesis 3, the ptp variable accounts for a large amount of variance above and beyond the control variables as evidenced by an increase in the pseudo r2 from 0.032 in step 1 to 0.054 in step 2 (an increase of 0.022 or 68.75%). in support of hypothesis 2, there is a significant negative relationship between ftp and frt score. the addition of the ftp variable increases the pseudo r2 from 0.054 in step 2 to 0.068 in step 3 (an increase of 0.014 or 25.93%), thus supporting hypothesis 4. a comparison of step 3 to step 1 demonstrates that the ptp and ftp variables together result in an increase in the pseudo r2 of 0.036 (an increase of 112.50%) over a model that only includes gender, 172 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 income, and financial knowledge/experience (gender, income, and stkgame). thus, we continue to find evidence in support of all four hypotheses when running hierarchical ordinal logistic models. 5. discussion and conclusion prior research has found that frt is impacted by a number of variables such as gender, income, and financial knowledge/experience. this study finds that tp accounts for a significant amount of the variance in frt above and beyond the amount contributed by those previously studied variables. young adults with a stronger present orientation tend to have higher frt scores and those with a stronger future orientation tend to have lower frt scores. however, prior research demonstrates that individuals with a long time horizon until retirement should be investing in portfolios of more risky securities (i.e., equities as opposed to bonds and money market funds) to maximize their long-term wealth accumulation. thus, the results from this study indicate that young adults are not likely to engage in optimal investment behavior. those that are thinking about their future (i.e., they exhibit a future table 8 hierarchical ordinal logistic regressions step 1 step 2 step 3 frt score frt score frt score ftp �0.561 p-value 0.000 ptp 0.698 0.742 p-value 0.000 0.000 gender 1.108 1.156 0.836 p-value 0.000 0.000 0.000 income 0.227 0.191 0.137 p-value 0.001 0.007 0.048 stkgame 0.768 0.817 0.918 p-value 0.000 0.000 0.000 n 320 317 317 wald �2 49.660 84.710 94.650 prob � �2 0.000 0.000 0.000 psuedo r2 0.032 0.054 0.068 financial risk tolerance (frt) score is measured using student scores on the grable and lytton (1999) 13-item financial risk tolerance scale. future time perspective (ftp) is measured using nine items from the zimbardo time perspective inventory (ztpi) future scale. present time perspective (ptp) is measured using nine items from the zimbardo time perspective inventory (ztpi) present-hedonistic scale. gender is a �0,1� indicator variable with 0 coded as female and 1 coded as male. income is estimate of parents’ income is where students were asked to estimate the combined income of their parents from all sources, before taxes. stkgame is a �0,1� indicator variable with 0 coded as not having taken a course in high school in which a stock market game was played and 1 coded as having taken such a course. this table presents results of the following ordinal logistic regressions using maximum likelihood estimation with robust standard errors: step 1: frt score � f(gender, income, stkgame) step 2: frt score � f(ptp, gender, income, stkgame) step 3: frt score � f(ftp, ptp, gender, income, stkgame). 173k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 orientation) tend to be financially risk averse and less likely to create a portfolio that will maximize their wealth. those that are more risk tolerant tend to be present oriented and are less likely to invest for their future. these findings point to a set of challenges for financial professionals, educators, and researchers: 1. how can we get the young adults with the higher frt focused on investing for their future? 2. how can we increase the financial risk tolerance of the young adults that are future oriented? consistent with prior research, this study also finds that females are more financially risk averse than males. while results from previous studies examining the relationship between gender and tp are mixed, females in the current study tend to display a stronger future orientation than males. contrary to prior research, the results show a negative relationship between income and future tp. this may be explained by the fact that the females in the sample tend to be more future oriented than the males, while at the same time they appear to provide lower estimates of their family income than the males do. additionally, the results generally support prior findings of a significant positive relationship between frt and both income and financial knowledge/experience. however, we find some interesting results for the income variable when we run regressions separately for each group (i.e., by gender or by financial knowledge/experience). when the ols regressions are run separately for males and females, income is a significant explanatory variable for females, but not for males. the reason for this difference is not clear, but as previously noted, the females in our sample do appear to provide lower estimates of their family income than the males. our separate regressions by group also show that the income variable is significant for participants that did not complete a stock market game in high school, but is not significant for those that did complete a stock market game. again, the reason for this difference is not clear. it does raise a question as to whether or not certain types of financial experience and knowledge may possibly mitigate the importance of income as a determinant of frt. however, these results may be an artifact of our particular income measure and further research is clearly needed before any conclusions can be made regarding the differences we find in the separate group regressions. future research should also address the two challenges noted above, exploring possible interventions that move present oriented individuals with a high frt toward a stronger future orientation and increase the frt of future oriented individuals that have a low frt. additionally, more research is needed to determine how generalizable the results are. the current sample is from a population of new college freshmen that are primarily white. future research could examine the relationship between tp and frt among samples of participants from different age groups, different race groups, and different education levels. acknowledgments the authors thank michael kraten and two anonymous reviewers for their helpful comments and suggestions. 174 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 appendix: a grable and lytton (1999) 13-item financial risk tolerance scale 1. in general, how would your best friend describe you as a risk taker? a) a real gambler b) willing to take risks after completing adequate research c) cautious d) a real risk avoider 2. you are on a tv game show and can choose one of the following. which would you take? a) $1,000 in cash b) a 50% chance at winning $5,000 c) a 25% chance at winning $10,000 d) a 5% chance at winning $100,000 3. you have just finished saving for a “once-in-a-lifetime” vacation. three weeks before you plan to leave, you lose your job. you would: a) cancel the vacation b) take a much more modest vacation c) go as scheduled, reasoning that you need the time to prepare for a job search d) extend your vacation, because this might be your last chance to go first-class 4. if you unexpectedly received $20,000 to invest, what would you do? a) deposit it in a bank account, money market account, or an insured cd b) invest it in safe high quality bonds or bond mutual funds c) invest it in stocks or stock mutual funds 5. in terms of experience, how comfortable are you investing in stocks or stock mutual funds? a) not at all comfortable b) somewhat comfortable c) very comfortable 6. when you think of the word “risk” which of the following words comes to mind first? a) loss b) uncertainty c) opportunity d) thrill 7. some experts are predicting prices of hard assets such as gold, jewels, collectibles, and real estate to increase in value. bond prices may fall, however, experts tend to agree that government bonds are relatively safe. most of your investment assets are now in high interest government bonds. what would you do? a) hold the bonds b) sell the bonds, put half the proceeds into money market accounts, and the other half into hard assets c) sell the bonds and put the total proceeds into hard assets d) sell the bonds, put all the money into hard assets, and borrow additional money to buy more. 8. given the best and worst case returns of the four investment choices below, which would you prefer? a) $200 gain best case; $0 gain/loss worst case b) $800 gain best case; $200 loss worst case c) $2,600 gain best case; $800 loss worst case d) $4,800 gain best case; $2,400 loss worst case 9. in addition to whatever you own, you have been given $1,000. you are now asked to choose between: a) a sure gain of $500 b) a 50% chance to gain $1,000 and a 50% chance to gain nothing 10. in addition to whatever you own, you have been given $2,000. you are now asked to choose between: c) a sure loss of $500 d) a 50% chance to lose $1,000 and a 50% chance to lose nothing 11. suppose a relative left you an inheritance of $100,000, stipulating in the will that you invest all the money in one of the following choices. which one would you select? a) a savings account or money market mutual fund b) a mutual fund that owns stocks and bonds (continued on next page) 175k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 appendix a (continued) c) a portfolio of 15 common stocks d) commodities like gold, silver, and oil 12. if you had to invest $20,000, which of the following investment choices would you find most appealing? a) 60% in low-risk investments, 30% in medium-risk investments, and 10% in high-risk investments b) 30% in low-risk investments, 40% in medium-risk investments, and 30% in high-risk investments c) 10% in low-risk investments, 40% in medium-risk investments, and 50% in high-risk investments 13. your trusted friend and neighbor, an experienced geologist, is putting together a group of investors to fund an exploratory gold mining venture. the venture could pay back 50 to 100 times the investment if successful. if the mine is a bust, the entire investment is worthless. your friend estimates the chance of success is only 20%. if you had the money, how much would you invest? a) nothing b) one month’s salary c) three month’s salary d) six month’s salary scoring: 1. a � 4; b � 3; c � 2; d � 1 2. a � 1; b � 2; c � 3; d � 4 3. a � 1; b � 2; c � 3; d � 4 4. a � 1; b � 2; c � 3 5. a � 1; b � 2; c � 3 6. a � 1; b � 2; c � 3; d � 4 7. a � 1; b � 2; c � 3; d � 4 8. a � 1; b � 2; c � 3; d � 4 9. a � 1; b � 3 10. a � 1; b � 3 11. a � 1; b � 2; c � 3; d � 4 12. a � 1; b � 2; c � 3 13. a � 1; b � 2; c � 3; d � 4 source: grable, j., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. appendix b: zimbardo time perspective inventory (ztpi) scale items ztpi present-hedonistic scale items: i believe that getting together with one’s friends to party is one of life’s important pleasures. i do things impulsively. i try to live my life as fully as possible, one day at a time. i make decisions on the spur of the moment. it is important to put excitement in my life. taking risks keeps my life from becoming boring. it is more important for me to enjoy life’s journey than to focus only on the destination. i find myself getting swept up in the excitement of the moment. i prefer friends who are spontaneous rather than predictable. ztpi future scale items: i believe a person’s day should be planned ahead each morning. when i want to achieve something, i set goals and consider specific means for reaching those goals. meeting tomorrow’s deadlines and doing other necessary work comes before tonight’s play. it upsets me to be late for appointments. i meet my obligations to friends and authorities on time. i complete projects on time by making steady progress. i make lists of things i must do. i am able to resist temptations when i know there is work to be done. i keep working at a difficult, uninteresting task if it will help me get ahead. consistent with ryack (2012, 2015), the present scale items used in this study include nine of the 15 items from the original ztpi present-hedonistic scale and the future scale items used include nine of the thirteen items from the original ztpi future scale (zimbardo & boyd, 1999). ryack (2012) conducted confirmatory factor analyses on these scale items using a sample of college students and found the revisions did not significantly alter scale integrity. 176 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 references anbar, a., & eker, m. (2010). an empirical investigation for determining of the relation between personal financial risk tolerance and demographic characteristic. ege akademik bakis, 10, 503–522. antonites, a. j., & wordsworth, r. (2009). risk tolerance: a perspective on entrepreneurship education. southern african business review, 13, 69–85. appleby, p. r., marks, g., ayala, a., miller, l. c., murphy, s., & mansergh, g. (2005). consideration of future consequences and unprotected anal intercourse among men who have sex with men. journal of homosexuality, 50, 119–133. bailey, j., & kinerson, c. (2005). regret avoidance and risk tolerance. financial counseling and planning, 16, 24–28. bajtelsmit, v. l., bernasek, a., & jianakoplos, n. a. (1999). gender differences in defined contribution pension decisions. financial services review, 8, 1–10. bernasek, a., & shwiff, s. (2001). gender, risk and retirement. journal of economic issues, 35, 345–356. boniwell, i., & zimbardo, p. g. (2004). balancing one’s time perspective in pursuit of optimal functioning. in p. a. linley & s. joseph (eds.), positive psychology in practice. hoboken, nj: wiley. d’alessio, m., guarino, a., de pascalis, v., & zimbardo, p. g. (2003). testing zimbardo’s stanford time perspective inventory (stpi) short form an italian study. time & society, 12, 333–347. davidson, r., & mackinnon, j. g. (1993). estimation and inference in econometrics. new york, ny: oxford university press. dorr, n., krueckeberg, s., strathman, a., & wood, m. d. (1999). psychosocial correlates of voluntary hiv antibody testing in college students. aids education and prevention, 11, 14–27. embrey, l. l., & fox, j. j. (1997). gender differences in the investment decision-making process. financial counseling and planning, 8, 33–39. epel, e. s., bandura, a., & zimbardo, p. g. (1999). escaping homelessness: the influences of self-efficacy and time perspective on coping with homelessness. journal of applied social psychology, 29, 575–596. faff, r., hallahan, t., & mckenzie, m. (2011). women and risk tolerance in an aging world. international journal of accounting and information management, 19, 100–117. faff, r., mulino, d., & chai, d. (2008). on the linkage between financial risk tolerance and risk aversion. the journal of financial research, xxxi, 1–23. finke, m. s., & huston, s. j. (2003). the brighter side of financial risk: financial risk tolerance and wealth. journal of family and economic issues, 24, 233–255. fishbein, m., & ajzen, i. (1975). belief, attitude, intention, and behavior: an introduction to theory and research. reading, ma: addison-wesley. gibson, r., michayluk, d., & van de venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. gilliam, j., chatterjee, s., & grable, j. (2010). measuring the perception of financial risk tolerance: a tale of two measures. journal of financial counseling and planning, 21, 30–43. gilliam, j. e., & chatterjee, s. (2011). the influence of birth order on financial risk tolerance. journal of business & economics research, 9, 43–50. gonzalez, a., & zimbardo, p. g. (1985). time in perspective. psychology today, 19, 20–26. grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625–630. grable, j. e., & joo, s.-h. (2004). environmental and biopsychosocial factors associated with financial risk tolerance. journal of financial counseling and planning, 15, 73–82. grable, j. e., & joo, s. (1997). determinants of risk preference: implications for family and consumer science professionals. family economics and resource management biennial, 2, 19–24. grable, j. e., & lytton, r. h. (1998). investor risk tolerance: testing the efficacy of demographics as differentiating and classifying factors. financial counseling and planning, 9, 61–73. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. 177k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 grable, j. e., & lytton, r. h. (2001). assessing the concurrent validity of the scf risk tolerance question. financial counseling and planning, 12, 43–54. grable, j. e., & lytton, r. h. (2003). the development of a risk assessment instrument: a follow-up study. financial services review, 12, 257–274. grable, j. e., lytton, r. h., & o’neill, b. (2004). projection bias and financial risk tolerance. the journal of behavioral finance, 5, 142–147. grable, j. e., & roszkowski, m. j. (2007). self-assessments of risk tolerance by women and men. psychological reports, 100, 795–802. greene, w. h. (1997). econometric analysis, 4th edition. upper saddle river, nj: prentice hall. gujarati, d. (2014). econometrics by example, 2nd edition. new york, ny: palgrave macmillan. hallahan, t., faff, r., & mckenzie, m. (2003). an exploratory investigation of the relation between risk tolerance scores and demographic characteristics. journal of multinational financial management, 13, 483–502. hallahan, t., faff, r., & mckenzie, m. (2004). an empirical investigation of personal financial risk tolerance. financial services review, 13, 57–78. hamilton, j. m., kives, k. d., micevski, v., & grace, s. l. (2003). time perspective and health-promoting behavior in a cardiac rehabilitation population. behavioral medicine, 28, 132–139. hanna, s., & chen, p. (1997). subjective and objective risk tolerance: implications for optimal portfolios. journal of financial counseling and planning, 8, 17–26. harber, k. d., zimbardo, p. g., & boyd, j. n. (2003). participant self-selection biases as a function of individual differences in time perspective. basic and applied social psychology, 25, 255–264. hariharan, g., chapman, k., & domian, d. (2000). risk tolerance and asset allocation for investors nearing retirement. financial services review, 9, 159–170. hartog, j., ferrer-i-carbonell, a., & jonker, n. (2002). linking measured risk aversion to individual characteristics. kyklos, 55, 3–26. henson, j. m., carey, m. p., carey, k. b., & maisto, s. a. (2006). associations among health behaviors and time perspective in young adults: model testing with boot-strapping replication. journal of behavioral medicine, 29, 127–137. hershey, d. a., & mowen, j. c. (2000). psychological determinants of financial preparedness for retirement. gerontologist, 40, 687–697. hodgins, d. c., & engel, a. (2002). future time perspective in pathological gamblers. journal of nervous and mental disease, 190, 775–780. holman, e. a., & silver, r. c. (2005). future-oriented thinking and adjustment in a nationwide longitudinal study following the september 11th terrorist attacks. motivation and emotion, 29, 389–410. jacobs-lawson, j. m., & hershey, d. a. (2005). influence of future time perspective, financial knowledge, and financial risk tolerance on retirement savings behaviors. financial services review, 14, 331–334. jianakoplos, n. a., & bernasek, a. (1998). are women more risk averse? economic inquiry, 36, 620–630. joireman, j., sprott, d. e., & spangenberg, e. r. (2005). fiscal responsibility and the consideration of future consequences. personality and individual differences, 39, 1159–1168. keller, c., & siegrist, m. (2006). investing in stocks: the influence of financial risk attitude and values-related money and stock market attitudes. journal of economic psychology, 27, 285–303. kennedy, p. (1998). a guide to econometrics, 4th edition. cambridge, massachusetts: the mit press. keough, k. a., zimbardo, p. g., & boyd, j. n. (1999). who’s smoking, drinking, and using drugs? time perspective as a predictor of substance use. basic and applied social psychology, 21, 149–164. kuzniak, s., rabbani, a., heo, w., ruiz-menjivar, j., & grable, j. e. (2015). the grable and lytton risktolerance scale: a 15-year retrospective. financial services review, 24, 177–192. larkin, c., lucey, b. m., & mulholland, m. (2013). risk tolerance and demographic characteristics: preliminary irish evidence. financial services review, 22, 77–91. levy, a. g., micco, e., putt, m., & armstrong, k. (2006). value for the future and breast cancer-preventive health behavior. cancer epidemiology biomarkers & prevention, 15, 955–960. 178 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 liu, x. (2016). applied ordinal logistic regression using stata�: from single-level to multilevel modeling. thousand oaks, ca: sage publications. luszczynska, a., gibbons, f. x., piko, b. f., & tekozel, m. (2004). self-regulatory cognitions, social comparison, and perceived peers’ behaviors as predictors of nutrition and physical activity: a comparison among adolescents in hungary, poland, turkey, and usa. psychology & health, 19, 577–593. mackillop, j., anderson, e. j., castelda, b. a., mattson, r. e., & donovick, p. j. (2006). convergent validity of measures of cognitive distortions, impulsivity, and time perspective with pathological gambling. psychology of addictive behaviors, 20, 75–79. mackillop, j., mattson, r. e., mackillop, e. j. a., castelda, b. a., & donovick, p. j. (2007). multidimensional assessment of impulsivity in undergraduate hazardous drinkers and controls. journal of studies on alcohol and drugs, 68, 785–788. maddala, g. s. (1983). limited-dependent and qualitative variables in econometrics. new york, ny: cambridge university press. neelakantan, u. (2010). estimation and impact of gender differences in risk tolerance. economic inquiry, 48, 228–233. nuttin, j. (1985). future time perspective and motivation: theory and research method. hillsdale, nj: erlbaum. olsen, r. a., & cox, c. m. (2001). the influence of gender on the perception and response to investment risk: the case of professional investors. the journal of psychology and financial markets, 2, 29–36. ouellette, j. a., hessling, r., gibbons, f. x., reis-bergan, m., & gerrard, m. (2005). using images to increase exercise behavior: prototypes versus possible selves. personality and social psychology bulletin, 31, 610– 620. pardoe, i. (2006). applied regression modeling: a business approach. hoboken, nj: john wiley and sons, inc. petry, n. m. (2001). substance abuse, pathological gambling, and impulsiveness. drug and alcohol dependence, 63, 29–38. petry, n. m., bickel, w. k., & arnett, m. (1998). shortened time horizons and insensitivity to future consequences in heroin addicts. addiction, 93, 729–738. powell, m., & ansic, d. (1997). gender differences in risk behaviour in financial decision-making: an experimental analysis. journal of economic psychology, 18, 605–628. ryack, k. (2011). the impact of family relationships and financial education on financial risk tolerance. financial services review, 20, 181–193. ryack, k. (2012). evidence that time perspective factors depend on the group: factor analyses of the cfc and ztpi scales with professional financial advisors. personality and individual differences, 52, 723–727. ryack, k. n. (2015). the time perspective of financial advisors and its effect on their decision-making. financial services review, 24, 387–410. doi:10.1016/j.paid.2010.02.029. shores, k., & scott, d. (2007). the relationship of individual time perspective and recreation experience preferences. journal of leisure research, 39, 28–59. siegel, j. j., & thaler, r. h. (1997). anomalies: the equity premium puzzle. journal of economic perspectives, 11, 191–200. statacorp lp. (2015). stata base reference manual release 14. college station, tx. strathman, a., gleicher, f., boninger, d. s., & edwards, c. s. (1994). the consideration of future consequences: weighing immediate and distant outcomes of behavior. journal of personality and social psychology, 66, 742–752. studenmund, a. h. (2005). using econometrics: a practical guide, 5th edition. new york, ny: addisonwesley pearson. sung, j., & hanna, s. (1996). factors related to risk tolerance. financial counseling and planning, 7, 11–20. toplak, m. e., liu, e., macpherson, r., toneatto, t., & stanovich, k. e. (2007). the reasoning skills and thinking dispositions of problem gamblers: a dual-process taxonomy. journal of behavioral decision making, 20, 103–124. vuchinich, r. e., & simpson, c. a. (1998). hyperbolic temporal discounting in social drinkers and problem drinkers. experimental and clinical psychopharmacology, 6, 292–305. 179k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 weber, e. u., blais, a.-r., & betz, n. e. (2002). a domain-specific risk-attitude scale: measuring risk perceptions and risk behaviors. journal of behavioral decision making, 15, 263–290. webley, p., & nyhus, e. k. (2006). parents’ influence on children’s future orientation and saving. journal of economic psychology, 27, 140–164. wills, t. a., sandy, j. m., & yaeger, a. m. (2001). time perspective and early-onset substance use: a model based on stress-coping theory. psychology of addictive behaviors, 15, 118–125. wong, a. (2011). financial risk tolerance and selected demographic factors: a comparative study in 3 countries. global journal of international business research, 5, 1–12. wooldridge, j. m. (2008). introductory econometrics: a modern approach, 4th edition. boston, ma: south western cengage. yao, r., sharpe, d. l., & wang, f. (2011). decomposing the age effect on risk tolerance. journal of socio–economics, 40, 879. yuh, y., & devaney, s. (1996). determinants of couples’ defined contribution retirement funds. financial counseling and planning, 7, 31–38. zimbardo, p. g., & boyd, j. n. (1999). putting time in perspective: a valid, reliable individual-differences metric. journal of personality and social psychology, 77, 1271–1288. zimbardo, p. g., keough, k. a., & boyd, j. n. (1997). present time perspective as a predictor of risky driving. personality and individual differences, 23, 1007–1023. 180 k. n. ryack, a. sheikh / financial services review 25 (2016) 157–180 from the editor this issue contains issue 2 of volume 25 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “behavioral and wealth considerations for seeking professional financial planning help,” is coauthored by jodi letkiewicz, chris robinson, and dale domian, all at york university. the authors use a canadian survey to examine the decisions to seek professional financial planning help. they find that people who use a financial planner have more wealth, lower subjective financial stress and higher financial self-efficacy than people who do not use a financial planner. they also find that people with higher self-efficacy in period t-1 are more likely to seek help in period t, leading to the conclusion that high self-efficacy drives one to seek financial planning help. their results indicate that subjective financial stress leads investors to seek financial planning help. the second article “financial adviser users and financial literacy,” is coauthored by bhanu balasubramnian and eric r brisker both at the university of akron. the authors use the 2012 national financial capability study to determine what demographic characteristics are associated with individuals that use financial advisers and whether financial advisers have any impact on the financial literacy of their clients. they find a significant increase in the use of financial advisers over the past decade. their research shows that savings or investments advisers have the largest positive impact on the financial literacy of their clients, followed by mortgage or loan and insurance advisers, even when controlling for financial education and potential endogeneity issues. the third article, “the relationship between time perspective and financial risk tolerance in young adults,” is coauthored by kenneth n. ryack and aamer sheikh both at quinnipiac university. the authors examine the relationship between time perspective (tp) and financial risk tolerance (frt) in young adults. prior research suggests young adults should invest in riskier portfolios to maximize wealth accumulation for retirement because of a future tp and a high frt. the results of their study indicate that tp accounts for a significant amount of variance in frt, however, the relationship between tp and frt is not financial services review 25 (2016) v–vi 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. optimal. future oriented individuals exhibit lower frt and present oriented individuals exhibit higher frt. the fourth article, “hedged etfs: do they add value?” is authored by srinidhi kanuri at the university of southern mississippi. hedged etfs provide individual investors with the opportunity to invest in etfs that follow strategies similar to those of hedge funds and seek returns uncorrelated with the market. the author investigates the performance for 6 different categories of hedged etfs and hedged mutual funds during the period 2008 to 2014. he found that hedged etfs had much lower risk compared to other index etfs, with the exception of bond market etf agg. although this did not translate into superior absolute or risk-adjusted performance. hedged etfs underperformed all other asset categories (with the exception of commodities etf dbc) and the absoluteand risk-adjusted performance of hedged mutual funds was similar to that of hedged etfs. the author concludes that investors would have been better off with index fund etfs. the final article, “investment strategies when selecting sustainable firms” is coauthored by todd m. shank at university of south florida st. petersburg and benjamin shockey at raymond james financial, inc. the authors investigate whether the emerging emphasis on sustainability is financially rewarded by market participants. this study examines the efficacy of passive versus active investment strategies when selecting sustainable firms for inclusion within an equity portfolio. utilizing two groups of “sustainability-focused” firms, they find financial support for an active selection of sustainable firms on a risk-adjusted basis. the author’s results support identifying global sustainability leaders by industry because they collectively showed greater financial performance over the past decade than both the dow jones sustainability index, as well as the broader market. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards stuart michelson editor financial services review vi editorial / financial services review 25 (2016) v–vi return-enhancing strategies with international etfs: exploiting the turn-of-the-month effect haiwei chena, sang heon shinb, xu sunc,* auniversity of alaska fairbanks, 303 tanana loop, street 201, school of management, uaf, fairbanks, ak 99775, usa balabama state university, department of accounting and finance, 915 s. jackson street, montgomery, al 36104, usa cutah valley university, department of finance and economics, 800 west university parkway, orem, ut 84058, usa abstract we show that the average return over the four-day period surrounding the turn of the month is significantly positive in eight out of the nine international exchange-traded funds (etfs). the strategy of buying-and-holding an etf during turn-of-the-month (tom) period and switching to holding t-bills during non-tom period produces significantly positive monthly average returns. this etft-bills switching strategy also has the lowest risk and highest sharpe ratio and sortino ratio than the traditional strategy of buying-and-holding either an index fund or an etf. investors pursuing this switching strategy generate a terminal value twice larger than the next best strategy of buying-andholding an etf. © 2015 academy of financial services. all rights reserved. jel classification: f3; g1 keywords: etf; international equity; turn-of-the-month; return; risk 1. introduction exchange-traded funds (etfs) become a valuable tool for individual investors and financial advisors in the pursuit of higher returns and more effective diversification. in this article, we study whether investors can take advantage of the turn-of-the-month (tom) effect * corresponding author. tel.: �1-801-863-8807; fax: �1-801-863-7218. e-mail address: sunxufin@gmail.com (x. sun) financial services review 24 (2015) 271–288 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. in international etfs. the tom effect refers to the phenomenon that stock returns are higher surrounding the turn of the month. early studies in ziemba (1991) and cadsby and ratner (1992) document that the tom effect exists in several international stock returns. recently, mcconnell and xu (2008) reconfirm the existence of the tom effect in 34 out of 38 international market indexes. hensel and ziemba (1996) and kunkel and compton (1998) show that the trading strategy of investing in low risk fixed-income account during non-tom period and switching to investing in stocks during tom period produces a higher return than the traditional strategy of buying-and-holding stocks. however, chen and chua (2011) show that the trading strategy of holding a standard & poor’s depositary receipt (spdr), the corresponding etf for the standard & poor’s (s&p) 500 index, during tom period and switching to holding t-bills during non-tom period does not produce a higher return than the strategy of investing in spdr throughout the month. given that etfs have high liquidity and extremely low trading cost, can individual investors exploit the tom effect in ishares? we examine the following markets: australia, brazil, canada, france, germany, japan, hong kong, sweden, and united kingdom. all of them allow foreign investors without much restriction. we first show that the tom effect exists in index returns in all markets except for the japanese market, whereas the tom effect exists in all nine etf returns. when we control for other known confounding factors such as the january effect and the weekend effect, the tom effect still exists in all nine etf returns except for the japanese market and in six index returns. in addition, we find that the risk level is lower during tom period than during non-tom period for both index and etf returns in all nine markets. following chen and chua (2011), we compare the performance of the following three strategies for individual investors. the first strategy is for investors who buy and hold an index fund mimicking a foreign stock market index throughout the month. the second strategy is for investors who buy and hold the corresponding etf throughout the month. the third strategy is for investors who invest in the etf during tom period and switch to holding t-bills during non-tom period. we show that this etf-t-bills switching strategy produces the highest return and has the lowest risk compared to the other two strategies. statistically, the mean monthly return from the switching strategy is significantly positive, whereas those from the other two strategies are not significant. economically, this switching strategy produces a terminal value that is at least 50% higher than the other two strategies. therefore, our results show that investors can exploit the tom effect in international etfs. the following of the article is organized as follows. section 2 reviews previous literatures on tom calendar anomaly and etfs. section 3 describes data and methodology used in this study. empirical results are presented in section 4. section 5 summarizes and concludes. 2. related studies 2.1. calendar anomaly ariel (1987) shows that the cumulative returns during a window of (�9, �9) around the first day of the month are non-negligible even after controlling for the impact of the january 272 h. chen et al. / financial services review 24 (2015) 271–288 effect. examining the monthly dow jones industrial average (djia) returns over the period of 1987 through 1986, lakonishok and smidt (1988) find an average of 0.473% return cumulated during the four-day period at the turn of month, which is higher than the average cumulative return of 0.349% in the whole month. similarly, cadsby and ratner (1992) document a tom effect in australia, canada, hong kong, germany, the united states, and the united kingdom. ziemba (1991) also finds a tom effect in a (�5, �2) window for the japanese stock market during 1949–1988. ogden (1990) hypothesizes that the standardization of payment system in the united states causes the tom effect. at the turn of each month, the concentrated payment of wages, dividends, interests, and other liabilities give rise to a surge of cash flow that is used for subsequent investment, which in turn pushes up the stock returns. more recently, mcconnell and xu (2008) reconfirm the existence of tom effect in a four-day window starting from the last day of the month to the third day of the following month using the crsp value-weighted and equal-weighted indexes for the period of 1987 through 2005. they also show that the tom effect exists in index returns in 34 out of the 38 countries examined. they find that the tom effect is not caused by the influence of stocks with small capitalization or low price, or higher volatility at the end of the month. however, they show that the trading volume and the net funds flows at the turn of the month are not significantly higher than those during the rest of the month, questioning the explanation by odgen (1990). studies have been conducted to examine the performance of various trading strategies designed to exploit the tom effect in stock returns. for example, henzel and ziemba (1996) show that the strategy of investing in a s&p 500 index fund during tom period and switching to bonds during non-tom period outperforms the strategy of investing in the stock index fund throughout the month. similarly, kunkel and compton (1998) show that the strategy of investing in a stock fund during tom period and switching to a money market account during non-tom period within the teachers insurance and annuity association— college retirement equities fund (tiaa-cref) fund family with no transaction costs produces 2.1% higher return than the strategy of buy-and-hold stocks. zwergel (2010) shows that the tom effect also exists in the stock index and the corresponding futures in germany, japan, united kingdom, and united states, and that trading strategies designed to exploit the tom effect are profitable even after adjusting for transaction costs. 2.2. etfs and calendar anomaly etfs are created to mimic a stock index. grossmann and beach (2010) show that ishares in four out of six countries examined are more correlated with a sample of foreign stocks than with the sample of corresponding adrs. for investors in the united states, etfs also have the advantages of low costs and tax efficiency over the traditional index funds, as shown in olienyk, schwebach, and zumwalt (1999) and in poterba and shoven (2002). unlike index funds, etfs are traded as regular stocks in the exchanges. investors who invest with index funds have to wait until the end of the day to purchase or redeem shares at the net asset value, which is determined by the closing prices. in contrast, investors can buy or sell etfs any time during the regular trading hours. 273h. chen et al. / financial services review 24 (2015) 271–288 several studies compare the performance of etfs with the performance of other instruments. for example, pennathur, delcoure, and anderson (2002) find that international ishares replicate the foreign index but also have a high degree of exposure to the u.s. market, which limits the diversification potential. examining the performance of etfs and closed-end funds for 14 countries, harper, madura, and schnusenberg (2006) conclude that etfs give investors better returns with a higher sharpe ratio and a positive jensen’s �. chu, mazumder, miller, and prather (2007) show that a strategy designed to exploit the lead-lag relation between ishares delivers significantly higher returns in 7 out of 12 pairs. chen and chua (2011), using a window of (�1, �3), find that returns are significantly higher during the four-day tom window than during the rest of the month for both spdr and the s&p 500 index. they compare five trading strategies and conclude that the etfbuy-and-hold strategy produces higher returns than the index-fund-buy-and-hold strategy. however, the strategy of holding t-bills during non-tom period and then switching to investing in spdr during tom period produces the highest sharpe ratio. chen and chua, therefore, suggest that investors should choose different strategies based on the consideration of their tax status and risk aversion. there is no study examining the presence of the tom effect in international etf returns. as a result, we aim to fill this gap in the literature. by comparing the performance of trading strategies designed to exploit the tom effect in the returns of etfs and their domestic indexes, we provide guidance for u.s. investors interested in international diversification and return enhancement. 3. data we obtain from datastream the daily data for the ishares for the following markets: australia, brazil, canada, france, germany, hong kong, japan, sweden, and the united kingdom. brazil is an emerging market with an active stock market and relatively less restriction on foreign investors. all others are developed markets. these etfs are actively traded on the american stock exchange (amex). we also obtain the dollar returns for the nine stock market indexes from datastream.1 as shown in table 1, except for the brazilian market, the data range is from march 29, 1996 to august 10, 2012. because of different holiday scheduling by different exchanges, trading is not synchronized between an etf in the u.s. market and its corresponding foreign index in an oversee market. there are more trading days from etfs than from the underlying indexes. in this study, we merge the etf and index time series by excluding missing value on either side of the pair, that is, an etf and its corresponding foreign stock index.2 table 2 presents the summary statistics. for returns, we calculate the simple percentage changes in the underlying index/etf level. among the nine markets, brazil has the largest average daily returns and the highest standard deviation for both market index and etf. the mean daily dollar index return is significantly positive only in the german market. for etfs, the mean daily dollar return is insignificant in all nine markets. 274 h. chen et al. / financial services review 24 (2015) 271–288 table 1 etfs and indexes country period etf/ticker market/index australia 3/29/1996 to 8/10/2012 ewa s&p/asx 200 index brazil 7/31/2000 to 8/10/2012 ewz brazil bovespa index canada 3/29/1996 to 8/10/2012 ewc s&p/tsx composite index france 3/29/1996 to 8/10/2012 ewq france cac 40 index germany 3/29/1996 to 8/10/2012 ewg dax 30 index hong kong 3/29/1996 to 8/10/2012 ewh hang seng index japan 3/29/1996 to 8/10/2012 ewj nikkei 225 index sweden 3/29/1996 to 8/10/2012 ewd omxs index u.k. 3/29/1996 to 8/10/2012 ewu ftse 100 index data source: datastream. table 2 descriptive statistics this table gives the brief statistic summary for etfs daily dollar returns and their underlying indexes in nine countries. for returns, we calculate the simple percentage changes in the underlying index/etf level. difference is the difference between the returns on etf and its corresponding index. except for brazil which starts from july 31, 2000, countries cover a sample period from march 29, 1996 to august 10, 2012. return numbers in the table are in percentage. country variable n mean median minimum maximum sd australia etf 3,975 0.0395 0.0000 �12.3898 20.7495 1.8470 index 3,975 0.0358 0.0769 �14.7872 8.8817 1.5029 difference 0.0037 �0.0843 brazil etf 2,915 0.0884 0.1389 �19.6277 25.5807 2.6171 index 2,915 0.0796 0.1383 �16.4422 18.3611 2.4563 difference 0.0087 �0.0251 canada etf 3,969 0.0348 0.0000 �23.1213 12.3607 1.6084 index 3,969 0.0320 0.1088 �12.8809 10.4345 1.4125 difference 0.0028 0.0041 france etf 3,977 0.0246 0.0000 �10.9457 13.0777 1.7342 index 3,977 0.0218 0.0566 �11.0744 12.9115 1.6497 difference 0.0027 �0.0129 germany etf 3,977 0.0191 0.0000 �11.2864 19.7896 1.8068 index 3,977 0.0274** 0.0271 �3.9172 5.8000 0.6510 difference �0.0083 �0.0154 hong kong etf 3,977 0.0185 0.0000 �12.3762 20.2381 2.0461 index 3,977 0.0313 0.0011 �13.6820 18.8512 1.7493 difference �0.0030 �0.0040 japan etf 3,977 0.0075 0.0000 �10.4077 17.1817 1.6336 index 3,977 0.0051 �0.0088 �10.5827 13.3955 1.6183 difference 0.0024 �0.0021 sweden etf 3,941 0.0303 0.0028 �19.1579 13.2918 2.2152 index 3,941 0.0350 0.0054 �9.6050 13.3513 1.7941 difference �0.0048 �0.0082 uk etf 3,977 0.0215 0.0000 �12.0225 17.0642 1.6133 index 3,977 0.0200 0.0594 �10.0018 12.9967 1.4115 difference 0.0015 �0.0000 ** and * denote for significance levels at 1% and 5%, respectively. 275h. chen et al. / financial services review 24 (2015) 271–288 4. empirical results 4.1. the tom effect following mcconnell and xu (2008) and chen and chua (2011), we use the (�1, �3) four-day tom window at the turn of the month. table 3 presents the mean daily returns. as table 3 return by day of the month the table below presents the average return over the four-day turn of month (tom) window, that is, from the last day of the month day (�1) to the first three days of the following month: day (�1), day (�2), and day (�3). difference is the difference between returns in tom period and non-tom period. �return is the difference between etf returns and the corresponding stock index returns. numbers in the table are in percentage. country day (�1) day (�1) day (�2) day (�3) tom non-tom difference australia etf 0.255* 0.354** 0.168 �0.095 0.171** 0.005 0.166** index 0.218* 0.174 0.231* �0.108 0.129* 0.012 0.117 �return 0.037 0.181 �0.063 0.013 0.027 �0.003 brazil etf 0.600** 0.810** 0.103 �0.063 0.363** �0.023 0.340** index 0.610** 0.676** 0.140 0.065 0.372** 0.007 0.365** �return �0.010 0.134 �0.037 �0.128 �0.026 0.006 canada etf 0.210* 0.295** 0.122 0.063 0.173** 0.001 0.172** index 0.199* 0.312** 0.056 �0.009 0.140** 0.005 0.135* �return 0.011 �0.018 0.066 0.072 0.002 �0.004 france etf 0.312** 0.262* �0.037 0.063 0.150** �0.006 0.156* index 0.324** 0.108 0.036 0.008 0.119* �0.003 0.122 �return �0.013 0.154* �0.073 0.055 0.030 �0.006 germany etf 0.268* 0.253 0.066 �0.031 0.139* �0.012 0.151* index 0.112* �0.004 0.058 0.035 0.005* 0.021 0.029 �return 0.156 0.257 0.008 �0.066 0.140** �0.029 hong kong etf 0.244 0.501** �0.038 0.038 0.186** �0.023 0.209** index 0.231* 0.247 0.322** �0.128 0.168** �0.002 0.170* �return 0.013 0.254 �0.359* 0.166 0.038 �0.012 japan etf 0.122 0.443** 0.090 �0.067 0.102 �0.015 0.117 index 0.047 0.175 0.088 �0.108 0.051 �0.005 0.006 �return 0.074 0.267* �0.178 0.041 0.062 �0.013 sweden etf 0.108 0.318* 0.143 0.057 0.156* 0.000 0.156* index 0.240* 0.215 0.364 0.031 0.162** 0.004 0.159* �return �0.131 0.104 �0.021 0.025 0.032 �0.009 uk etf 0.210 0.384** 0.051 �0.042 0.151** �0.011 0.162** index 0.178* 0.237* 0.130 0.002 0.137** �0.010 0.147** �return 0.033 0.147 �0.079 �0.044 0.036 �0.009 ** and * denote for significance levels at 1% and 5%, respectively. 276 h. chen et al. / financial services review 24 (2015) 271–288 can be seen in the table, the tom effect is most pronounced during the first two days for both etfs and the indexes. more important, as shown in the tom column, the average daily return during tom period is significantly positive for both index and etf in all markets except for the japanese market. our findings are consistent with those in mcconnell and xu (2008) and chen and chua (2011). as the last column in table 3 shows, the return difference between tom period and non-tom period is significantly positive for etfs in all markets except for the japanese market. on the other hand, for index returns, the return difference between tom period returns and non-tom period is significantly positive only in five markets. table 3 also reports the return differences on each of the four days at the turn of each month. the differences are largely insignificant for most of the days. for all nine markets, there is no statistical difference between the index returns and the etf returns, indicating that etfs track the underlying market index very closely even though there is a difference in trading hours.3 to account for known factors such as the january effect and the weekend effect, we conduct a regression analysis to test whether the tom effect still exists in return after controlling for these factors. as shown in eq. (1), we include the two additional control variables. rit � a0 � a1tomit � a2januaryit � a3weekendit � eit where rit denotes the daily returns of etfs or stock indexes. tom is a dummy variable that takes a value of 1 if a trading day belongs to the four-day tom window and otherwise 0. other control variables such as january and weekend are also dummy variables. january takes a value of 1 if a day is in january and 0 otherwise. similarly, weekend takes a value of 1 if it is a monday and 0 otherwise. the null hypothesis is that there is no tom effect, that is, �1 � 0. as shown in table 4, the etf regression shows that the coefficient for tom is significantly positive in all markets except for the japanese market and the swedish market. the index regression shows that the coefficient for tom is significantly positive in six markets but not in the french, germany, and japanese markets. therefore, after controlling for other factors, the results in table 4 are largely consistent with those in table 3, confirming that the tom effect exists in the returns of both index and etfs in most of the markets examined. fig. 1 presents a comparison of the cumulative returns during the four-day tom period and those during the rest of the month for the nine etfs. as can be seen, the average cumulative returns in non-tom period are negative in five markets; in contrast, the average cumulative return over tom period is positive in all markets. the sheer magnitude indicates the importance of the tom effect in determining the overall performance for the whole month. are returns more volatile during tom period than during non-tom period? table 5 presents the return standard deviations on (1) each day of the four-day tom period, (2) during tom period, and (3) during the rest of the month for both etfs and stock indexes. the results show several patterns. first, return standard deviation is higher for etfs than for the indexes in all markets, a finding consistent with those in chen and chua (2011), which 277h. chen et al. / financial services review 24 (2015) 271–288 reports that spdr has a higher standard deviation during both tom period and the rest of the month than its underlying s&p 500 index. second, the standard deviation is no higher during tom period than during the rest of the month for both indexes and etfs. in the australian market, returns are less volatile during tom period than during non-tom period for both the etf and the index. similarly, risk is significantly lower during tom period than during non-tom period for the swedish etf returns and the japanese index returns. 4.2. tom effect: robustness check 4.2.1. tests on subsamples does the tom effect exist in recent data? as a robustness check, we evenly divide our whole sample into two subsamples and retest the tom effect in each subsample. as shown in table 6, in the early sample, the tom effect exists in both etf and index returns only in the brazilian market. in contrast, in the more recent subsample, returns are significantly table 4 regression analysis for each market, the dependent variable is the daily return of the etf and the stock index, respectively. tom, january, and weekend are dummy variables if a trading day is during the turn of the month period, in january, and a monday, respectively. regression coefficients use a newey-west correction of standard errors for heteroscedasticity and autocorrelation. country constant tom january weekend n r2 australia etf 0.0003 0.0016* �0.0008 �0.0006 3,975 0.001 index 0.0002 0.0012* �0.0008 �0.0000 3,975 0.001 brazil etf 0.0008 0.0034** �0.0010 0.0028 2,915 0.004 index 0.0004 0.0036** �0.0001 0.0015 2,915 0.003 canada etf �0.0000 0.0017** �0.0001 0.0000 3,969 0.001 index 0.0002 0.0014* �0.0005 0.0005 3,969 0.001 france etf 0.0003 0.0016* �0.0015 �0.0012 3,977 0.002 index 0.0002 0.0012 �0.0014 �0.0005 3,977 0.001 germany etf 0.0001 0.0015** �0.0015 �0.0006 3,977 0.002 index 0.0004** 0.0003 �0.0011* �0.0005 3,977 0.002 hong kong etf 0.0002 0.0021** �0.0010 �0.0017 3,977 0.004 index 0.0001 0.0018* �0.0018 0.0002 3,977 0.003 japan etf 0.0001 0.0012 �0.0008 �0.0010 3,977 0.002 index 0.0000 0.0006 �0.0001 �0.0014* 3,977 0.002 sweden etf 0.0002 0.0016 �0.0013 0.0005 3,941 0.002 index 0.0002 0.0016* �0.0012 �0.0004 3,941 0.001 u.k. etf 0.0001 0.0016* �0.0011 �0.0005 3,977 0.003 index 0.0000 0.0015** �0.0013 0.0002 3,977 0.002 ** and * denote for significance levels at 1% and 5%, respectively. 278 h. chen et al. / financial services review 24 (2015) 271–288 higher during tom period than during non-tom for etfs in four markets. similarly, index returns are significantly higher during tom period than during non-tom period in four markets. it is an open question whether publicity from the researches on this phenomenon and subsequent trading by investors to exploit such return anomaly contributes to a more recent prevalence of the tom effect in data. 4.2.2. the tom effect in local etfs4 previous results are for etfs traded on the u.s. exchanges. because ishares also issue etfs aboard, we examine if the tom effect also exists in local etfs returns. only three local etfs, that is, germany, hong kong, and united kingdom, have daily prices longer than three years.5 as a result, we examine only these three markets and the results are shown in table 7. returns during tom period are insignificant for the local etf in germany, whereas they are significantly positive in both hong kong and united kingdom. for united kingdom’s local etf, daily returns on the first and the second day of each month are significantly positive. however, only united kingdom’s local etf returns statistically exhibit the tom effect. fig. 2 plots the cumulative returns of both the four-day tom period and the remainder of the same month. similar to the results in fig. 1, the returns generated from the four days at the turn of the month are all positive and account for a large portion of the total returns of the entire month. because these local etfs are most likely traded by domestic investors, the lack of evidence for the tom effect in local etf returns in hong kong and germany could be a result of differences in the payment systems or investor behavioral difference. although it fig. 1. cumulative returns during tom period and non-tom period for etfs. 279h. chen et al. / financial services review 24 (2015) 271–288 may be difficult to document the difference in investor behavior, it will be interesting to examine how the three countries differ in the payment system, for example, the frequency by which wages and salary are distributed to employees. 4.3. the performance of three trading strategies chen and chua (2011) document that the trading strategy of buying-and-holding spdr significantly outperforms the strategy of buying-and-holding a s&p 500 index fund by 0.145%. they also show that the strategy of holding a s&p 500 index fund during non-tom period and then switching to holding spdr during tom period outperforms the traditional strategy of buying-and-holding a s&p 500 index fund by 0.068%. in this article, we examine the performances and risks of the following three trading strategies:6 table 5 risk by day of the month the table below presents the risk as measured by daily return standard deviation at the turn of the month from the last day of the month to the first three days of the following month: day (�1), day (�2), and day (�3). tom and non-tom stand for tom period and the remainder of the month period, respectively. difference is the difference between returns in tom period and non-tom period. numbers in the table are in percentage. country day (�1) day (�1) day (�2) day (�3) tom non-tom difference australia etf 1.601 1.882 1.847 1.683 1.762 1.870 �0.152* index 1.406 1.506 1.426 1.238 1.402 1.526 �0.145** brazil etf 2.212 2.722 2.623 2.478 2.525 2.633 �0.167 index 1.946 2.555 2.766 2.259 2.412 2.462 �0.070 canada etf 1.371 1.554 1.820 1.439 1.555 1.620 �0.087 index 1.208 1.473 1.541 1.278 1.385 1.418 �0.027 france etf 1.591 1.802 1.834 1.774 1.755 1.728 0.011 index 1.509 1.686 1.670 1.679 1.639 1.652 0.006 germany etf 1.606 1.930 1.926 1.838 1.831 1.800 0.000 index 0.673 0.643 0.646 0.659 0.655 0.650 0.007 hong kong etf 1.867 2.187 2.048 1.966 2.027 2.049 �0.096 index 1.513 1.923 1.692 1.609 1.697 1.761 �0.091 japan etf 1.513 1.843 1.674 1.346 1.616 1.638 �0.004 index 1.681 1.553 1.579 1.272 1.529 1.640 �0.100* sweden etf 1.872 2.265 2.144 2.082 2.094 2.244 �0.125* index 1.531 1.877 1.803 1.731 1.739 1.806 �0.057 u.k. etf 1.546 1.738 1.591 1.561 1.612 1.611 �0.003 index 1.117 1.639 1.391 1.348 1.397 1.414 �0.012 ** and * denote for significance levels at 1% and 5%, respectively. 280 h. chen et al. / financial services review 24 (2015) 271–288 table 6 robustness check–tom effect in two subsamples the table presents the average daily return at the turn of the month (tom) defined as from the last day of the month to the first three days of the following month, and the rest of the month (non-tom), respectively. panel a is for the first subsample (before 2006 for brazil and before 2002 for the other markets). numbers in the table are in percentage. country tom non-tom difference panel a australia etf 0.128 0.039 0.089 index 0.122 0.040 0.082 brazil etf 0.372** 0.059 0.313* index 0.425** 0.051 0.374** canada etf 0.127 0.036 0.091 index 0.098 0.032 0.066 france etf 0.143 0.004 0.139 index 0.133 0.004 0.108 germany etf 0.127 0.028 0.009 index 0.039 0.030 0.009 hong kong etf 0.158 0.008 0.149 index 0.143* 0.022 0.121 japan etf 0.072 0.007 0.065 index 0.067 0.024 0.042 sweden etf 0.147 0.042 0.105 index 0.153 0.027 0.125 u.k. etf 0.160* �0.004 0.164 index 0.149* �0.008 0.156 panel b australia etf 0.232* �0.038 0.270** index 0.138* �0.026 0.164* brazil etf 0.323 �0.138 0.461 index 0.171 �0.173 0.344 canada etf 0.238** �0.049 0.288** index 0.199** �0.033 0.233** france etf 0.160 �0.021 0.181 index 0.128 �0.011 0.139 germany etf 0.156 �0.066 0.221* index 0.066* 0.010 0.056 hong kong etf 0.226 0.068 0.295* index 0.204 �0.037 0.240* (continued on next page) 281h. chen et al. / financial services review 24 (2015) 271–288 strategy 1: buy and hold a stock index fund each month. strategy 2: buy and hold a corresponding etf each month. strategy 3: invest in the etf during tom period and then switch to investing in t-bills during non-tom period. some brokerage firms such as fidelity and vanguard have launched their own etfs and allow their customers to trade etfs commission-free (with some restriction, of course). therefore, we do not take transaction costs into consideration.7 table 8 reports the mean returns and the standard deviation for each of the three trading strategies in nine countries. the strategy of buying-and-holding an index fund does produce a positive but statistically insignificant return—except for the german market. similarly, the strategy of buying-and-holding a corresponding etf produces a positive but statistically insignificant average return in all but the brazilian market. such a result reflects the fact that returns are negative during non-tom period as shown in table 6, which offsets the positive table 6 continued country tom non-tom difference japan etf 0.144 �0.048 0.192 index 0.028 �0.049 0.077 sweden etf 0.170 �0.065 0.235 index 0.176* �0.032 0.208* u.k. etf 0.138 �0.019 0.158 index 0.119 �0.011 0.130 ** and * denote for significance levels at 1% and 5%, respectively. table 7 return by day of the month–local etfs the table below presents the average daily return of etfs at the turn of month (tom) from the last day of the month, day (�1), to the first three days of the following month, day (�1), day (�2), and day (�3). only these three local etfs have sufficient data. return stands for the mean daily return of corresponding period, and risk is measured by the standard deviation of daily returns. tom and non-tom stand for the mean return during tom period and the remainder of the month period, respectively. difference in the last column is the difference between returns in tom and non-tom period. numbers in the table are in percentage. country day (�1) day (�1) day (�2) day (�3) tom non-tom difference germany return 0.169 0.037 0.037 0.153 0.118 �0.003 0.122 risk 1.421 1.789 1.789 1.553 1.594 1.525 0.069 hong kong return 0.207 0.037 0.469** �0.026 0.225** 0.024 0.201 risk 1.418 1.789 1.987 1.902 1.759 2.022 �0.263 u.k. return �0.056 0.271* 0.271** �0.028 0.109* �0.011 0.119* risk 1.161 1.303 1.303 1.224 1.261 1.306 �0.045 ** and * denote for significance levels at 1% and 5%, respectively. 282 h. chen et al. / financial services review 24 (2015) 271–288 returns during tom period. in contrast, the etf-t-bills switching produces significantly positive returns in all nine markets.8 equally important, as shown in table 8, returns from the etf-t-bills switching strategy have a much lower standard deviation than those from the other two strategies—again across all nine markets. in australia, canada, german, and sweden, the standard deviation of returns from this switching strategy is more than twice lower than that from the other two strategies. as shown in fig. 3, we use the monthly returns to calculate the sharpe ratio for these three strategies in all markets. the sharpe ratio provides a measure for the risk-adjusted performance of an investment. fig. 3 shows that the etf-t-bills switching strategy has the highest sharpe ratio, easily beating the two buy-and-hold strategies. in france, hong kong, japan, and united kingdom, the sharpe ratio from this switching strategy is almost three times higher than that from the two buy-and-hold strategies. in other markets, the sharpe ratio from this etf-t-bills switching strategy is about twice larger than that from the other two strategies. therefore, on a risk adjusted basis, investors are better off pursuing a strategy of investing in t-bills in non-tom period and switching to holding etf during tom period than simply buying and holding either an index fund or an etf. we also calculate the sortino ratio, which differs from the sharpe ratio by replacing the standard deviation with the downside risk. the downside risk is defined as the standard deviation of returns below the target return. by using the downside risk, the sortino ratio focuses on the risk that an investor may be short of reaching the investment target. we use the s&p 500 index return as the target return in each month in the calculation of the downside risk. fig. 4 exhibits the sortino ratio for the three strategies. consistent with the sharpe ratio in fig. 3, the etf-t-bills switching strategy has the highest sortino ratio, outperforming the other two buy-and-hold strategies on a riskadjusted basis. although the return difference between the three strategies is statistically insignificant, different strategies can still generate returns that are economically significant. to fig. 2. cumulative return between four-day tom and non-tom period for three local etfs. 283h. chen et al. / financial services review 24 (2015) 271–288 assess the economic significance, we calculate the terminal value of one dollar invested following each of the three strategies. as fig. 5 shows, the etf-t-bills switching strategy produces a higher terminal value than the other two traditional buy-and-hold strategies in all markets. this etf-t-bills switching strategy produces a terminal value twice larger than the etf-buy-and-hold strategy, the next best strategy, in canada, france, hong kong, japan, and united kingdom. in the other four markets, this etf-t-bills switching strategy produces a terminal value about 50% larger than the next best strategy. notice that the etf-buy-and-hold strategy underperforms the index-fundbuy-and-hold strategy in german, hong kong, and sweden. in these three markets, the index-fund-buy-and-hold strategy has a higher sharpe ratio and sortino ratio than the etf-buy-and-hold strategy. overall, fig. 5 demonstrates the economic significance of exploiting the tom effect in etfs. table 8 trading strategies comparison this table shows the monthly mean return and standard deviation for three trading strategies: strategy 1 of buy-and-hold a stock index fund throughout each month, strategy 2 of buy-and-hold an etf throughout each month, and strategy 3 of investing in the corresponding etf during tom period and then switching to investing in t-bills during non-tom period. numbers in the table are in percentage. country strategy 1 strategy 2 strategy 3 (index buy-and-hold) (etf buy-and-hold) (etf-t-bills switching) australia mean 0.722 0.798 0.806** sd 6.17 6.40 3.19 brazil mean 1.590 1.764* 1.532** sd 10.67 10.24 4.50 canada mean 0.645 0.702 0.813** sd 6.23 6.49 2.91 france mean 0.438 0.43 0.723** sd 6.322 6.396 3.22 germany mean 0.554* 0.386 0.790** sd 3.64 6.75 3.37 hong kong mean 0.632 0.373 0.868** sd 7.32 7.27 3.62 japan mean 0.103 0.151 0.530* sd 6.32 5.74 2.99 sweden mean 0.701 0.605 0.748** sd 7.10 7.98 3.82 u.k. mean 0.403 0.434 0.727** sd 4.79 4.96 2.89 ** and * denote for significance at 1% and 5%, respectively. 284 h. chen et al. / financial services review 24 (2015) 271–288 5. summary and conclusion this article documents that etf returns are higher during the four-day tom window than the rest of the month in eight out of the nine markets examined. regression analysis shows that such an effect is still present in etf returns after controlling for known factors such as the january effect and the weekend effect. the four-day window accounts for most of the positive returns of the month. the rest of the month typically has either lower returns or negative returns. we also find that there is generally no difference between the etf returns and the underlying index returns across the different days of the month. however, etf returns have a higher standard deviation than the index returns. these findings are consistent with the previous literature. for investors interested in exploiting the tom effect, we compare the strategy of holding t-bills during non-tom period and then switching to etfs during tom period against the two strategies of buy-and-hold either an index fund or an etf, respectively. it is shown that only this etf-t-bills switching strategy produces significantly positive average monthly returns. the other two buy-and-hold strategies generate positive but insignificant monthly average returns. we further show that this switching strategy produces the highest riskadjusted returns as indicated by a higher sharpe ratio and a higher sortino ratio. in terms of economic significance, investors pursuing this etf-t-bills switching strategy on average achieve a terminal value that is at least 50% larger than the traditional buy-andhold strategy, in either an index fund or an etf. because investors can trade etfs on the exchange at extremely low transaction costs and with high liquidity, investing in etfs is an attractive alternative to investing in traditional mutual funds. our results show that u.s. fig. 3. sharpe ratios for the three strategies. note: in strategy 1, investors buy and hold the index fund. in strategy 2, investors buy and hold the etf. in strategy 3, investors hold the etf during tom period and switch to holding the t-bills during non-tom period. in calculating the sharpe ratio, the excess return is defined as the difference between the monthly strategy return and the monthly t-bills return. 285h. chen et al. / financial services review 24 (2015) 271–288 fig. 4. sortino ratio for different strategies. note: in strategy 1, investors buy and hold the index fund. in strategy 2, investors buy and hold the etf. in strategy 3, investors hold the etf during tom period and switch to holding the t-bills during non-tom period. in calculating the sortino ratio, the monthly s&p 500 index return is used as the target return and the excess return is defined as the difference between the monthly strategy return and the monthly t-bills return. fig. 5. terminal values for the etf investment strategies. note: this figure exhibits the ending balance of one dollar invested under the three strategies over the sample period. in strategy 1, investors buy and hold the index fund. in strategy 2, investors buy and hold the etf. in strategy 3, investors hold the etf during tom period and switch to holding the t-bills during non-tom period. 286 h. chen et al. / financial services review 24 (2015) 271–288 investors can benefit from using ishares to exploit the well-known tom effect in their pursuit of higher return and more effective international diversification. notes 1 we check for holidays in those markets to avoid the “holiday return effect” as pointed out in klein, zwergel, and fock (2009). 2 we also conduct analyses separately for an etf and its corresponding index. results remain the same for etfs in that the tom effect exits on all markets except for the japanese market, which is consistent with the results reported in table 3. for index returns, the tom effect exists in all markets except for the japanese market, which shows more a potent tom effect since table 3 shows no tom effect in the german market. there is no significant qualitative changes for results in other tables. as a result, they are not reported for brevity and are available upon request. 3 see http://www.ishares.com/us/products/product-list#categoryid�129&lvl2�overview, for more about the underlying index. 4 we thank stephen spathe for pointing out to us and for suggesting this robustness check. 5 specifically, data is available since january 2001 for germany, november 2001 for hong kong, and may 2000 for united kingdom, respectively. 6 although not reported in the article, the strategy of holding t-bills during non-tom period and switching to holding an index fund underperforms the etf-t-bills switching strategy. 7 there can be other costs such as slippage. we thank an anonymous referee for pointing out this cautionary note. 8 there are some mutations in the results in the subsamples. for the index-t-bill switching strategy, significantly positive returns are found in australian, brazilian, canadian, and german markets in the first subsample, but only in markets in hong kong, sweden, and united kingdom in the second subsample. for the etf-t-bill switching strategy, significantly positive returns are found in markets in australia, brazil, canada, and united kingdom in the first subsample, but in canada, france, germany, hong kong, sweden, and united kingdom in the second subsample. it is the same pattern when the strategies are applied to etfs and indexes separately without merging the data. as a result, it seems that the etf-switching strategy is consistently profitable in the more recent period. acknowledgments we thank dr. stuart michelson (editor) and two anonymous referees for detailed suggestions. we also gratefully acknowledge the assistance of cathy anson in editing the manuscript. helpful comments are also received from stephen spathe and participants of the 2013 academy of financial services annual meeting in chicago. 287h. chen et al. / financial services review 24 (2015) 271–288 references ariel, r. (1987). a monthly effect in stock returns. journal of financial economics, 18, 161–174. cadsby, c., & ranter, m. (1992). turn-of-month and pre-holiday effects on stock returns: some international evidence. journal of banking & finance, 16, 497–509. chen, h., & chua, c. a. (2011). the turn-of-the-month anomaly in the age of etfs: a reexamination of return-enhancement strategies. journal of financial planning, 24, 62–67. chu, t., mazumder, i., miller, e., & prather, l. (2007). exploitable cross autocorrelation among ishares. financial services review, 16, 293–308. grossmann, a., & beach, s. (2010). expanding a u.s. portfolio internationally: adrs their underlying assets, and etfs. financial services review, 19, 163–185. harper j., madura, j., & schnusenberg, o. (2006). performance comparison between exchange-traded funds and closed-end country funds. journal of international financial markets, institutions and money, 16, 104–122. hensel, c., & ziemba, w. (1996). investment results from exploiting turn-of-the-month effects. journal of portfolio management, 22, 17–23. klein, c., zwergel, b., & fock, j. (2009). reconsidering the impact of national soccer results on the ftse 100. applied economics, 41, 3287–3294. kunkel, r. a., & compton, w. (1998). a tax-free exploitation of the turn-of-the-month effect: c.r.e.f. financial services review, 7, 11–23. lakonishok, j., & smidt, s. (1988). are seasonal anomalies real? a ninety-year perspective. review of financial studies, 1, 403–425. mcconnnell, j., & xu, w. (2008). equity returns at the turn of the month. financial analysts journal, 64, 49–64. ogden, j. (1990). turn-of-month evaluations of liquid profits and stock returns: a common explanation for the monthly and january effects. journal of finance, 45, 1259–1272. olienyk, j., schwebach, r., & zumwalt, j. (1999). webs, spdrs, and country funds: an analysis of international cointegration. journal of multinational financial management, 9, 217–232. pennathur, a., delcoure, n., & anderson, d. (2002). diversification benefits of ishares and closed-end country funds. journal of financial research, 25, 541–557. poterba, j., & shoven, j. (2002). exchanges-traded funds: a new investment option for taxable investors. american economic review, 92. papers and proceedings of the one hundred fourteenth annual meeting of the american economic association, 422–427. zwergel, b. (2010). on the exploitability of the turn-of-the-month effect—an international perspective. applied financial economics, 20, 911–922. ziemba, w. (1991). japanese security market regularities: monthly, turn-of-the-month and year, holiday and golden week effects. japan and the world economy, 3, 119–46. 288 h. chen et al. / financial services review 24 (2015) 271–288 a marginal cash flow analysis of mortgagors’ choices jim musumecia,* adepartment of finance, bentley university, waltham, ma 02452, usa abstract a great deal of academic research has focused on determinants of the spread between interest rates on conforming versus jumbo and 15-year versus 30-year mortgages, but much less has been done to help the borrower determine what choice is best for him. we examine these issues from the borrower’s frame of reference and find that comparisons of mortgage terms can be facilitated by analyzing the marginal cash flows from one mortgage contract to another. for many borrowers the “conventional wisdom” leads to suboptimal choices; making the better choice can easily produce low-risk doubledigit returns. © 2016 academy of financial services. all rights reserved. jel classifications: g21; d14 keywords: mortgage maturity; jumbo; conforming 1. introduction a great deal of academic research has focused on the determinants of spreads between various types of mortgages, but many homeowners are guided only by general rules of thumb. this article briefly summarizes the sources of the spreads between different mortgages and then discusses some factors that may help borrowers assess their choices in light of their own personal circumstances. we find that while there is no “one size fits all” optimal selection, the conventional wisdoms often lead to suboptimal choices. before analyzing the mortgagor’s choices, it is useful to recognize why different types of mortgages feature different rates. in general, for any two different mortgage contracts, the one that imposes more risk on the lender will feature the higher rate. because long-term securities are more sensitive to interest rate changes than are short-term securities, they are * corresponding author. tel.: �1-781-891-2235; fax: �1-781-891-2896. e-mail address: jmusumeci@bentley.edu financial services review 25 (2016) 87–104 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. more risky to the lender and typically command a premium (as per the liquidity preference hypothesis, e.g., see copeland, weston, and shastri (2005), pp. 262–264). in addition, borrowers have the option to prepay and the option to default on the mortgage. mortgage interest rates reflect the values of these options (e.g., kau and keenan, 1995); starting immediately after the first payment, shorter-term mortgages at all times have a lower outstanding principal than longer-term mortgages, and therefore, the values of both the option to default and the option to refinance are lower for shorter-term mortgages. most borrowers will refinance when the terms are sufficiently favorable, and lenders have a good idea of how to value this option. likelihood of default, however, may be unique to the borrower. some researchers view default as a cold-blooded decision, with borrowers doing so whenever it is in their best interests (e.g., kau and keenan, 1995). in contrast, others (e.g., brueckner, 2000) suggest that personal characteristics, such as the stability of the borrower’s income stream or her reluctance to default on a mortgage, are also important. without loss of generality, we focus on borrowers of the type brueckner describes, specifically, ones who know they are less likely to default than the average borrower. alternately, our analysis can be thought of as an assessment of what benefit the borrower gains if she is willing to reduce the value of her option to default (or to refinance). how can such a borrower signal this characteristic to the lenders and get a better rate than she otherwise would? certainly just saying “i have a low risk of default” is insufficient because all borrowers can allege that. instead, the borrower must signal through her choices that the option to default is not worth much to her. in the insurance literature, there is a similar phenomenon in which insurers try to separate low-risk and high-risk clients from the rest, and then offer each group different rates that better correspond to their risk profiles. such programs as discounted insurance premia for honors students or nonsmokers are two examples of insurance companies’ inferring risk profiles from other personal characteristics. another method is to offer clients a choice of deductibles. if all rates were actuarially fair conditioned on the population average, then high-risk (low-risk) clients would prefer low (high) deductibles. insurance companies, knowing this, will adjust the population-wide actuarially fair rates to offer each group a rate that better corresponds to its risk characteristics (as revealed by the choice of deductible). puelz and snow (1994) find “strong evidence” that insurance companies offer better rates to drivers who choose high deductibles. the same principle applies to mortgage markets. ceteris paribus, some types of loans (e.g., longer terms or smaller down payments) are inherently more risky for the lender, in part because of an increase in default risk by the borrower. borrowers who know that their likelihood of defaulting is much lower than average will prefer to avoid this risk premium, and should, therefore, choose shorter-term mortgages and larger down payments. dhillon, shilling, and sirmans’ (1990) finding that wealthier borrowers (who have greater resources and therefore, a presumably lower default risk) are more likely to choose mortgages with shorter maturities can be viewed as corroboration of the first part of this conjecture. empirical evidence on the second part is mixed, primarily because we cannot perform true experiments, but must rely on observational studies. lenders can observe part of the borrower’s risk profile and insist that high-risk borrowers make larger down payments. thus, 88 j. musumeci / financial services review 25 (2016) 87–104 studies of the relation between down payment and default likelihood are complicated by endogeneity problems (as discussed in harrison, noordewier, and yavas (2004). many financial transactions generate an economy of scale that translates into better consumer rates for larger amounts. for example, cd rates are typically greater if the customer is willing to commit a larger amount of money. mortgages have the opposite characteristic: conforming loans are only available up to a specified amount that depends on the average house price in the area, and mortgages above that threshold are classified as jumbo mortgages and typically charged a higher rate. the reason is that most loans are pooled into diversified portfolios of mortgages and sold, and a larger loan makes it more difficult to diversify the portfolio. in the remaining sections, we evaluate the borrower’s choice between a jumbo and a conforming mortgage, and that between a 15-year and a 30-year mortgage. without loss of generality, we assume that interest rates are sufficiently low that the mortgagor’s option to refinance has a value sufficiently small that the borrower can ignore it (or, alternatively, that interest rates are very stable, in which case the option to refinance is also small), and similarly we assume that the borrower ignores her option to default (presumably based on personal information that this option will not be, or is at least very unlikely to be, exercised). these assumptions are made for convenience only: even in their absence, our analysis can be viewed as an assessment of the benefit the mortgagor obtains by voluntarily reducing the value of these options. finally, we assume the borrower swaps mortgages (jumbo for conforming and 30-year for 15-year) and then we compare marginal cash flows as the mortgagor moves from one type of mortgage to the other. 2. jumbo versus conforming mortgages jumbo rates generally exceed conforming rates. according to the bloomberg website as of august, 2015, the gap for 30-year mortgages is about 38 basis points (4.24% vs. 3.86%). however, as recently as early march, 2013, the gap was closer to 50 basis points (4.20% vs. 3.70%). these seem typical; cotterman and pearce (1996) report that the spread between jumbo mortgages and conforming loans ranges between 15 and 60 basis points. thus, an issue that many borrowers face is whether it is worthwhile to try to come up with a larger down payment to avoid a jumbo mortgage’s premium. the higher rate on a jumbo mortgage is charged on the entire loan amount, not just on the excess over the threshold. this leads to a very high marginal interest rate when a mortgage exceeds the threshold by only a slight amount. one way to assess this marginal rate is to think of a jumbo loan in the amount of j dollars as a combination of the maximal amount for a conforming loan, c, plus a marginal loan amount, m, at the marginal rate, as in fig. 1. if the rate on the jumbo loan is rj, then rj is approximately equal to a weighted average of the rate on the conforming loan, rc, and the rate on the marginal amount, rm, or rj � c c � m rc � m c � m rm.1 solving for rm gives us rm � rj � c m �rj � rc�. because rj � rc, the marginal interest rate rm will always exceed the jumbo rate. furthermore, because this second term increases without bound when m approaches zero, the 89j. musumeci / financial services review 25 (2016) 87–104 marginal rate rm is very high when the threshold is exceeded by only a small amount. consider, for example, a $500,000 threshold for a jumbo loan, and a rate on a jumbo loan is that is 50 basis points higher than the conforming loan rate of 4.0%. suppose also that a borrower takes out a mortgage for $525,000. the interest rate on the marginal $25,000 in excess of the threshold is about 4.5% � 500,000 25,000 �4.5% � 4.0%� � 14.5%. this is a very high interest rate for a relatively low-risk endeavor, and strongly suggests that the borrower would be well advised to find some alternate way to come up with the extra $25,000. for example, ibbotson and sinquefield (2013) report that the long-run historical nominal geometric mean return on the s&p 500 has been only about 9.8% from 1926 to 2012, and so by selling $25,000 worth of stock and adding the proceeds to the down payment, the borrower effectively swaps a risky 9.8% expected return for a very low-risk 14.5% savings. in this rather extreme case, even raiding an ira (and paying the early withdrawal penalty) would merit serious consideration. when the marginal loan amount m is larger, the marginal rate drops somewhat. consider, for example, the same rates as in the previous paragraph, but a loan totaling $600,000, so that the marginal loan amount m is $100,000. now the marginal interest rate is 4.5% � 500,000 25,000 �4.5% � 4.0%� � 7.0%. given that this 7.0% is a very low-risk return, finding another source for the marginal $100,000 merits strong consideration by most, but it is plausible to expect that 7.0% may be below some borrowers’ reservation prices2 and, therefore, acceptable to them. fig. 1. a jumbo loan depicted as a maximal conforming loan plus a marginal loan. the figure on the left depicts a jumbo loan at an interest rate of rj, while the one on the right portrays a maximal conforming loan in the amount of c and at an interest rate of rc, plus a marginal loan in the amount of m and at an interest rate of rm. m is selected as j – c, so both the jumbo loan and the conforming-marginal combination have the same total value. because rj � c c � m rc � m c � m rm, solving for the marginal rate rm gives us rm � rj � c m �rj � rc�. 90 j. musumeci / financial services review 25 (2016) 87–104 when the marginal amount borrowed is larger, the rate on the marginal amount continues to fall. for example, suppose that the loan is for $1,500,000 total, so that the marginal amount m is $1,000,000. now the interest rate on m is 4.5% � 500,000 25,000 �4.5% � 4.0%� � 4.75%. table 1 shows the marginal rates rm for loans and spreads of several sizes. there is no one-size-fits-all when it comes to an individual’s reservation price for interest rates. however, table 1 allows an investor to at least assess the true cost of the marginal loan m. if it is unusually high, she might consider selling other assets to make a larger down payment and bring the loan down to conforming levels, waiting until she has saved a larger down payment, or settling for a less expensive house. 3. 15-year versus 30-year mortgages (ignoring taxes) some borrowers may be constrained by their income because many lenders follow a 28/36 rule (e.g., thangavelu, 2015). under this rule, monthly housing payments should not exceed 28% of gross income, and total debt commitments should not exceed 36% of gross income. if a lender is following this rule in terms of offering a loan, then some borrowers will not have a choice and will have to get a 30-year mortgage. even for borrowers who are not so constrained, the conventional wisdom is that a 30-year mortgage offers more flexibility than a 15-year mortgage and, therefore, is the wiser choice. for example, goff and cox (1998) demonstrate that if the difference in payments is invested in a tax-deferred account like a 401(k), even after adjusting for taxes, mortgagors who live in the house for 30 years will have a higher expected wealth in 30 years.3 several other researchers reach similar conclusions.4 in addition, qualified borrowers who lean towards a 15-year are often steered away with the advice that, if the borrower has extra cash flow during the lifetime of the mortgage, he can always voluntarily pay more towards the principal if he chooses. certainly borrowers table 1 marginal interest rates rm for loans exceeding a maximum conforming amount of $500,000 when rc � 4% rj total loan $525,000 $550,000 $575,000 $600,000 $650,000 $700,000 $750,000 4.625% 17.125% 10.875% 8.792% 7.750% 6.708% 6.188% 5.875% 4.5% 14.500% 9.500% 7.833% 7.000% 6.167% 5.750% 5.500% 4.375% 11.875% 8.125% 6.875% 6.250% 5.625% 5.313% 5.125% 4.25% 9.250% 6.750% 5.917% 5.500% 5.083% 4.875% 4.750% 4.125% 6.625% 5.375% 4.958% 4.750% 4.542% 4.438% 4.375% note. the table features the marginal rate, rm, which is paid on the amount by which mortgage size exceeds the threshold for a conforming loan. in all cases, the maximum conforming loan size is assumed to be $500,000 and the conforming apr assumed to be 4%. various possibilities for the jumbo rate are shown in the different rows, and various sizes in the different columns. for example, in the first entry (rj � 4.625% and size � $525,000), a borrower who accepts the jumbo rate on a $525,000 mortgage is effectively taking out a conforming mortgage of $500,000 at 4% and paying a rate of 17.125% on the $25,000 by which his loan exceeds the conforming maximum. 91j. musumeci / financial services review 25 (2016) 87–104 should not accept a rate or term that results in payments with which they are uncomfortable (grable and lytton, 1999 present a 20-question survey to assess an individual’s degree of risk aversion, and three of the 20 questions pertain to mortgage choices), and if the rates on the two mortgages are the same, then recommending the 30-year is impeccable advice. however, in practice the rates are rarely the same, and the higher rate on the 30-year mortgage is the price of the greater flexibility. ceteris paribus, flexibility is good, but like any other good, it has its reservation price. borrowers who do not need the flexibility are likely to find that its price is too high. we begin with an (admittedly unrealistic) extreme example to better identify the tradeoffs involved. suppose the annual percentage rate (apr) on a 15-year, $500,000 mortgage is 3.6%, while that on an otherwise identical 30-year mortgage is 7.2% (unless otherwise specified, we assume all annual rates are expressed as aprs, i.e., the actual periodic monthly rate is just apr/12). now monthly payments (throughout the article we consider only interest and principal, and not insurance or property taxes; we consider income taxes and the fact that mortgage interest is tax-deductible in section 4) on the 15-year mortgage will be $3599.02, and those on the 30-year will be $3393.94. it is difficult to imagine that many borrowers will find the flexibility of having the option to keep the monthly difference of $205.08 for the first 15 years is worth paying the extra $3393.94/month for the last 15 years of the 30-year mortgage. one way of visualizing the true cost of that flexibility is to find the monthly payments of an imaginary 30-year mortgage at the 15-year mortgage apr of 3.6%. evaluated at the same interest rate as a 15-year mortgage, this hypothetical 30-year mortgage would give the borrower the flexibility to pay $3599.02 – 2273.23 � $1325.79 less per month when compared with the 15-year borrower. however, in exchange for this flexibility to pay $1325.79 less, the 7.2% mortgage rate essentially charges the borrower an additional $3393.94 – 2273.23 � $1120.71/month, leaving the borrower with a net payment that is only $205.08 less than that of the 15-year mortgagor, as depicted in panel a of fig. 2. this is clearly a poor proposition for most borrowers.5 while this example is too extreme to be observed in practice, it establishes that the greater flexibility of the 30-year mortgage has a price, and that price should not be paid if it exceeds the borrower’s reservation price. even with a more realistic example, the flexibility has a nontrivial cost. as depicted in panel b of fig. 2, a $500,000, 15-year mortgage with an apr of 3.75% requires monthly payments of $3636.11, while a 30-year mortgage at a monthly rate of 4.50%/month requires payments of $2533.43. inserting the intermediate step described above finds that a hypothetical 30-year mortgage at 3.75% requires payments of $2315.58, so that in exchange for the flexibility of paying $3636.11 – 2315.58 � $1320.53 less per month, the borrower is paying back $2533.43 – 2315.58 � $217.85. this may be a desirable trade for some, but not for others. the issue is how can we best measure the price of this flexibility so that the borrower can have a solid basis for judgment? one metric arises from considering the marginal payments of a 30-year mortgagor who swaps that mortgage’s cash flows for those of a 15-year mortgage. as in the discussion of jumbo mortgages, by marginal cash flow we mean any additional cash flows the 15-year borrow pays (or receives) as compared with those of the 30-year borrower. this swap produces a greater cash outflow for the first 15 years in exchange for elimination of all payments after year 15. for example, suppose as above that the apr on a $500,000, 30-year 92 j. musumeci / financial services review 25 (2016) 87–104 mortgage is 4.50%, while that on a 15-year is 3.75%. payments on the 15-year mortgage would be $3636.11, or $1102.69 more than the $2533.43 payments under the 30-year mortgage. what does the 15-year borrower get for this extra $1102.69/month? he is finished paying for the house at the end of 15 years, while the holder of the 30-year mortgage still has monthly payments of $2533.43 to make for an additional 15 years. thus, the 15-year mortgagor’s investment of the additional $1102.69/month for 15 years has produced a gain (savings) of $2533.43/month for the 15 years after that. the internal rate of return (irr) of this investment is the value of r satisfying 1102.69(fvifa180, r%) � 2533.43(pvifa180, r%).6 the solution is r � 0.4632%/month, or an apr of 5.56%. a well-established principle in finance is that, when directly comparing the returns on two investments, it is important to account for risk. for example, it is inappropriate to suggest a cd yielding 3% is a worse investment than a stock index fund producing an expected return of 12% because the index fund has significantly higher risk, and much of that 9% differential is compensation for that risk. similarly, because leases entail cash outflows that are relatively fig. 2. the price of a 30-year mortgage’s greater flexibility. panel a: exaggerated example of the price of flexibility. the figure on the left represents the monthly payment for a 15-year, $500,000 mortgage at a 3.6%/year apr, while the one on the right represents that of a 30-year mortgage at an unrealistically high 7.2%/year. the figure in the center represents payments on a hypothetical 30-year mortgage at the 15-year rate of 3.6%. if the mortgagor could secure such a loan, he would have the flexibility to pay $3599.02 – 2273.23 � $1325.79 less per month. however, this flexibility comes with a cost because the actual 30-year rate requires a payment of $3393.94 – 2273.23 � 1120.71 more than the hypothetical mortgage. thus the 30-year borrower pays 1120.71/ month for the option to pay 1325.79 less, and ends up paying only $205.08 per month less than the 15-year borrower. it is difficult to imagine borrowers for whom this $205.08 savings for the first 15 years is worth paying $3393.94/month more for the last 15 years. panel b: moderate example of the price of flexibility. this panel makes the same point as panel a, but with more reasonable values. here the 30-year borrower is paying back $2533.43 – 2315.88 � $217.85 for the option to pay $3636.11 – 2315.58 � $1320.53 less per month. certainly some borrowers will find this reasonable, but when the problem is framed this way, many will not. 93j. musumeci / financial services review 25 (2016) 87–104 certain, ross et al. (2013) point out that the after-tax cost of debt is a more appropriate discount rate when evaluating leases than is the cost of capital. cheung and miu (2015) make a similar observation that real estate is comparable to bonds because “bonds and real estate share very similar risk and return characteristics,” and reichenstein (1998) similarly views a mortgage as a short position in bonds. our example of the marginal cash flows from trading a 30-year mortgage for a 15-year produces cash flows with very low risk, and thus a fair comparison is other investments with low risk. the 5.56% irr of the previous paragraph is a significantly higher return than other contemporary investments with such low risk (e.g., as of early february, 2014, the yield on 30-year t-bonds was only about 3.67%) and is earned at only the cost of reducing the values of the mortgagor’s options to default or refinance.7 another metric is net present value (npv); using the 15-year apr of 3.75%, for example, we find the npv of the marginal cash flows of 15-year mortgage, when compared with those of the 30-year, to be $47,039.79. given our assumption of no default, the actual value of the house plays no role in the calculations because it is not a marginal effect; whether the borrower takes out a 15-year or a 30-year mortgage affects the principal due at any time, but does not affect the actual value of the house itself. the previous analysis assumed the mortgage was held for 30 years. what if it is liquidated before then? for example, suppose the house is sold after seven years. now the 15-year mortgagor still owes $301,169.18, while the 30-year borrower owes $133,962.24 more, or $435,131.42.8 thus, compared with the 30-year borrower, the 15-year borrower has paid 1102.69/month more for these first seven years, but owes $133.962.24 less at liquidation. because both borrowers are liquidating the mortgage, there are no cash flows at all (and, therefore, no marginal cash flows) after year seven. in this case the irr of the 15-year mortgagor’s marginal $1102.69 monthly investment is the value of r satisfying 1102.69(fvifa 84, r% ) � $133,962.24. the solution is 0.84%/month, or an apr of 10.12%, which is significantly greater than the 5.56% apr of the previous example. however, using the same 3.75% as before to find npv, we now find it to be only $21,721.24, which is lower than the borrower’s $47,039.79 npv in that example. the choice of when to sell the house would probably be exogenous. however, if the mortgagor wanted to rank possible times of sale, the fact that the alternatives have different lifespans would make npv an inappropriate ranking device. in such cases, finding an annuity equivalent to the npv (e.g., see brealey, myers, and allen, 2008, pp. 155–160) or even using irr is a better tool for ranking. the borrower in fact saves more money (npv) the longer he owns the house, but his marginal npv, equivalent annuity, and irr decline when the house is sold at a later date. this can be seen in table 2 and fig. 3, which feature several measures of the benefits for various years in which the house is sold or the mortgage otherwise prepaid. the last entry in table 2, which a 15-year mortgagor who sells the house in one year earns a return of 56.39% on his marginal investment, is so large that it seems a likely error. it is not. as before, the 15-year borrower pays an extra $3636.11 – $2533.43 � $1102.69 per month more than the 30-year borrower. in exchange, in 12 months the 15-year borrower owes $474,684.48, or $17,249.38 less than the 30-year borrower’s $491,933.87. the monthly discount rate that represents the return on the marginal $1102.69 investment is the value of r that solves 1102.69(fvifa12, r%) � $17,249.38. this value of r is 4.70% per month, for an apr of 56.39% per year (because here we have monthly compounding, ear � effective 94 j. musumeci / financial services review 25 (2016) 87–104 annual return is calculated as (1 � periodic monthly return)12 – 1 � 73.52%). of course, because of transactions costs, few homeowners would purchase if they knew they were going to sell in a year. nevertheless, if they did sell then, the transactions costs are presumably identical for the 15-year and 30-year borrowers, and so the marginal effect of transactions costs (i.e., any additional cash flows the 15-year borrow pays [or receives] as compared with those of the 30-year borrower) is zero, and the 15-year mortgage still offers a 56.39% return on marginal cash flows relative to the 30-year. moreover, while the average life of a table 2 pre-tax benefits from the marginal investment of a 15-year mortgage’s larger payments house sold at month npv at 15-year pre-tax rate annual change in npv equivalent monthly annuity irr (apr) irr (ear) 360 $47,039.79 $40.39 $217.85 5.56% 5.70% 348 $46,999.40 $116.53 $221.74 5.56% 5.70% 336 $46,882.87 $195.02 $225.58 5.56% 5.71% 324 $46,687.85 $275.96 $229.36 5.57% 5.72% 312 $46,411.89 $359.43 $233.09 5.59% 5.73% 300 $46,052.46 $445.55 $236.77 5.60% 5.75% 288 $45,606.91 $534.41 $240.40 5.62% 5.77% 276 $45,072.50 $626.13 $243.97 5.65% 5.80% 264 $44,446.37 $720.82 $247.50 5.69% 5.84% 252 $43,725.55 $818.59 $250.97 5.73% 5.88% 240 $42,906.96 $919.57 $254.39 5.79% 5.94% 228 $41,987.39 $1,023.88 $257.76 5.85% 6.01% 216 $40,963.50 $1,131.66 $261.08 5.94% 6.10% 204 $39,831.85 $1,243.03 $264.35 6.05% 6.22% 192 $38,588.81 $1,358.15 $267.58 6.19% 6.37% 180 $37,230.66 $1,477.16 $270.75 6.38% 6.57% 168 $35,753.50 $1,600.20 $273.87 6.61% 6.81% 156 $34,153.30 $1,727.45 $276.95 6.88% 7.10% 144 $32,425.86 $1,859.05 $279.98 7.19% 7.43% 132 $30,566.81 $1,995.19 $282.96 7.56% 7.83% 120 $28,571.62 $2,136.03 $285.89 8.01% 8.31% 108 $26,435.59 $2,281.77 $288.78 8.55% 8.89% 96 $24,153.82 $2,432.58 $291.62 9.23% 9.64% 84 $21,721.24 $2,588.68 $294.41 10.12% 10.60% 72 $19,132.56 $2,750.26 $297.16 11.30% 11.90% 60 $16,382.30 $2,917.54 $299.86 12.96% 13.75% 48 $13,464.76 $3,090.73 $302.52 15.47% 16.61% 36 $10,374.03 $3,270.07 $305.13 19.71% 21.59% 24 $7,103.96 $3,455.80 $307.70 28.42% 32.42% 12 $3,648.16 $3,648.16 $310.22 56.39% 73.52% note. this table shows the financial advantage of the cash flows of a 15-year, $500,000 mortgage at an apr of 3.75% when compared with those of a 30-year mortgage with a 4.5% apr. for example, if the 15-year mortgagor lives in the house for 30 years, then the present value of his marginal cash flows (discounted at the 15-year rate) is $47,039.79, and his rate of return on his marginal payments for the first 180 months is 5.56% apr. if he lives in the house for only, say, 84 months, and then sells, he has a lower but still significant $21,721.24 npv, but a higher apr of 10.12% on his marginal investment. npv increases as the house is kept for a longer time, but at a decreasing rate (because the principal is declining). for purposes of comparison, the 1926–2012 geometric mean return on long-term government bonds is 5.7%, with a standard deviation of 9.7%. irr � internal rate of return; npv � net present value; apr � annual percentage rate; ear � effective annual return. 95j. musumeci / financial services review 25 (2016) 87–104 mortgage varies over time, it is typically well under 10 years. for example, freddie mac’s offering circular supplement of june, 2010, suggests a weighted average life of around six and a half years. in this range, the choice of a 15-year mortgage still offers a low-risk, double-digit return when compared with the 30-year mortgage. for purposes of comparison with other low risk investments, we note that ibbotson and sinquefield (2013) report that long-term government bonds have a geometric mean return of 5.7% and a standard deviation of returns of 9.7%. the result that the advantage of the 15-year mortgage is more pronounced when the time the house is sold is shorter has interesting implications. specifically, younger buyers are more likely to move (not only because they are more likely to have children and have other motivations to trade up to larger houses, but also because they are more mobile and likely to move because of their job). thus, while young buyers may place the greatest value on the fig. 3. npv and irr of the marginal cash flows created by moving from a 30-year mortgage to a 15-year mortgage (ignoring taxes). the figures above depict the npv and irr (expressed as an apr) of the marginal investment (higher payments for 15 years) and marginal benefits (nothing owed after 15 years, and a smaller principal if the house is paid off before 15 years). in both cases the x-axis represents the number of years before the house is sold or both mortgages paid off. whether measured by npv or irr, the financial advantage of the 15-year mortgage is significant. the figures shown are for a $500,000 mortgage, with an apr of 3.75% for a 15-year mortgage and 4.5% for a 30-year mortgage. 96 j. musumeci / financial services review 25 (2016) 87–104 flexibility a 30-year mortgage has to offer, they might also be the ones to stand to gain the most from the marginal investment required of a 15-year mortgage. moreover, as samuelson (1994) and bodie, merton, and samuelson (1992) point out, younger investors may find it optimal to take more risks because, if the outcomes are unfavorable, they still have time to make the required adjustments later in life. if any group were to find the higher costs of a 30-year mortgage acceptable, it is likely to be this group, as they want to invest any additional money in higher-risk ventures as goff and cox (1998) suggest.9 on the other hand, this age group may also earn the highest returns from a 15-year mortgage because of the greater likelihood they will sell the house relatively early, which realizes the highest return on the marginal cash flows. the situation appears clearer for older homebuyers who will want to reduce their risks as they get closer to the end of their working years, and for such individuals a 15-year mortgage is more likely to be the optimal choice. whether the extra costs of the 30-year mortgage are acceptable depends on such characteristics as the borrower’s age, the stability of her cash flows, her degree of risk aversion, her overall portfolio, and her progress towards retirement goals. it is easy to say that the $1102.69 is a good investment, but just where is that money to be found? one possibility is that it can take the place of some alternate investments, for example, any part of an investment portfolio that would otherwise be allocated to low-risk investments. the borrower who sells the house after seven years, for example, will be better placed at that time if he has chosen the 15-year mortgage, as he has effectively invested the $1102.69 at an apr of 10.12%, which is equivalent to an ear of 10.60%. this is considerably higher than the expected return of a low-risk bond portfolio, and even higher than the historical geometric mean return of 9.8% from investing in equity. consider, for example, a younger homebuyer who knew a move was likely after about seven years in his current location. because of its substantially lower risk, the riskless ear of 10.6% on the marginal investment in a 15-year mortgage would seem to dominate the risky 9.8% he would expect to earn if the same marginal cash flows were invested in an all-equity portfolio. 4. 15-year versus 30-year mortgages (considering taxes and different levels of interest rates) in a different setting, fortin et al. (2007) find that rules of thumb also do not work very well in a refinancing context. specifically, they found that, when considering taxes and time value of money, the breakeven point for refinancing was 35% to 40% more distant than otherwise estimated. however, in their framework, refinancing required fees that provided no tax shield, and which were used to reduce interest expense, which reduces the tax benefits. in our framework, interest for either mortgage is tax deductible, so taxes do not affect the result in such a lopsided way, as can be seen in table 3. although all calculations were made with after-tax cash flows and after-tax discount rates, all rates in the tables are reported on a pretax basis to facilitate comparisons, with the pretax rate � after-tax rate/(1-tax rate). because interest constitutes a greater proportion of the payments for the 30-year mortgage than for the 15-year, the 30-year has a comparative advantage in this regard when taxes are considered. however, it is also true that more interest is paid, even on an after-tax basis. the 97j. musumeci / financial services review 25 (2016) 87–104 values in table 3 are less than their analogs from table 2, but the difference is not large. for example, in table 2, the irr of the marginal cash flows of a mortgagor who used a 15-year mortgage, but kept the house for 30 years, was 5.56%, but the corresponding entry in table 3 falls to only 5.38%. one reason the drop between after-tax returns is so small (compared with the large difference in fortin et al.) is that the main benefit of the 15-year mortgage is that the principal is paid down faster, and this is not subject to taxes. it is appropriate to compare the riskless returns we have found so far with contemporary riskless returns of the same type (taxable or tax-exempt). by analogy, whether a starting salary of $50,000/year for an entry-level engineering position is better than average depends table 3 after-tax benefits from the marginal investment of a 15-year mortgage’s larger payments house sold at month npv at 15-year after-tax rate annual change in npv equivalent monthly annuity irr (apr) irr (ear) 360 $35,427.62 $38.89 $140.21 5.38% 5.52% 348 $35,388.73 $110.97 $143.30 5.38% 5.52% 336 $35,277.76 $183.49 $146.33 5.39% 5.52% 324 $35,094.27 $256.48 $149.29 5.40% 5.53% 312 $34,837.79 $329.99 $152.19 5.41% 5.55% 300 $34,507.79 $404.05 $155.03 5.43% 5.56% 288 $34,103.74 $478.71 $157.80 5.45% 5.59% 276 $33,625.03 $554.00 $160.53 5.48% 5.62% 264 $33,071.03 $629.97 $163.19 5.51% 5.65% 252 $32,441.06 $706.65 $165.80 5.56% 5.70% 240 $31,734.42 $784.09 $168.35 5.61% 5.75% 228 $30,950.33 $862.33 $170.86 5.67% 5.82% 216 $30,088.00 $941.41 $173.30 5.76% 5.91% 204 $29,146.59 $1,021.39 $175.70 5.86% 6.02% 192 $28,125.20 $1,102.29 $178.05 6.00% 6.17% 180 $27,022.91 $1,184.18 $180.34 6.18% 6.36% 168 $25,838.73 $1,267.09 $182.59 6.40% 6.60% 156 $24,571.64 $1,351.07 $184.80 6.67% 6.87% 144 $23,220.57 $1,436.17 $186.95 6.97% 7.20% 132 $21,784.41 $1,522.43 $189.06 7.33% 7.58% 120 $20,261.98 $1,609.91 $191.12 7.76% 8.04% 108 $18,652.06 $1,698.66 $193.15 8.29% 8.61% 96 $16,953.40 $1,788.73 $195.12 8.95% 9.33% 84 $15,164.68 $1,880.16 $197.06 9.81% 10.26% 72 $13,284.52 $1,973.02 $198.96 10.95% 11.52% 60 $11,311.50 $2,067.35 $200.81 12.56% 13.31% 48 $9,244.16 $2,163.21 $202.63 14.99% 16.07% 36 $7,080.95 $2,260.66 $204.40 19.10% 20.86% 24 $4,820.29 $2,359.75 $206.14 27.51% 31.26% 12 $2,460.54 $2,460.54 $207.85 54.44% 70.31% note. this table shows the financial advantage of after-tax cash flows of a 15-year, $500,000 mortgage at an apr of 3.75% when compared with those of a 30-year mortgage with a 4.5% apr, assuming a 33% tax rate for the mortgagor. npv is calculated using the after-tax 15-year rate [�(pre-tax rate)(1–.33)]. irrs are calculated on an after-tax basis, but expressed here on a pre-tax basis, that is, as (after-tax irr)/(1-tax rate). for example, if the 15-year mortgagor lives in the house for 84 months, and then sells, he has a $15,164.68 npv and a pre-tax irr (expressed as an apr) of 9.81% on his marginal investment. npv increases as the house is kept for a longer time, but at a decreasing rate (because the principal is declining). irr � internal rate of return; npv � net present value; apr � annual percentage rate; ear � effective annual return. 98 j. musumeci / financial services review 25 (2016) 87–104 on whether we are talking about a position in 1975, 1995, or 2015. in our case, for purposes of comparison, bloomberg reported the 30-yield municipal bond yield to be 2.96%, with a taxable equivalent yield of 2.98%/(1–0.33) � 4.45%, as of early march, 2013. a comparison of the rates earned by swapping a 15-year mortgage for a 30-year and the rates offered by t-bonds and municipal bonds of comparable maturities is shown in table 4. next we examine what happens for different levels of interest rates. clearly the strategy of swapping a 30-year for a 15-year mortgage will be more beneficial when the spread between the two rates is larger, so we do not analyze different spreads. however, the question remains whether the strategy becomes more profitable or less profitable when mortgage rates increase. we find npv declines a little for larger mortgage rates, but this should be no surprise—ceteris paribus, larger discount rates necessarily cause present values to fall. we also find irr rises as the mortgage rates rise. for example, assuming the house is sold in 30 or more years, we find that on a pretax basis the gap between the irr earned on the swap and the apr on the 30-year mortgage ranges from 98 to 125 basis points as the 30-year rate assumed values between 3.50% and 6.50%. in contrast with npv, the irrs are increasing in the interest rates, even when the spread between 15-year and 30-year rates is held constant. table 5 summarizes the results for other rates and years until the house is sold. finally, while the 15-year mortgage offers large benefits relative to the 30-year in today’s interest rate environment, whether this is generally true remains to be seen. table 6 uses the freddie mac database to compare the two mortgages if initiated at various times. because the 15-year database goes back only to september, 1991, and interest rates do not change table 4 a comparison of the irrs earned from a 30-year mortgage to 15-year mortgage swap compared with rates on comparable t-bonds (pre-tax) or municipal bonds (after-tax) ignoring taxes years ear of 15-year to 30-year marginal cash flows ear of comparable t-bond 2 32.42% 0.25% 5 13.75% 0.89% 10 8.31% 2.05% 30 5.70% 3.27% considering taxes years taxable equivalent ear of 15-year to 30 year marginal cash flows ear of comparable municipal bond taxable equivalent ear of comparable municipal bond 1 70.31% 0.24% 0.36% 2 31.26% 0.35% 0.52% 5 13.31% 0.82% 1.23% 10 8.04% 1.85% 2.76% 30 5.52% 2.98% 4.45% note. this table compares returns from the mortgage swap (from tables 2 and 3) with those from t-bonds (in the before-tax case) and municipal bonds (in the after tax case). bond returns were taken from the bloomberg website in early march, 2013. in all cases, annual returns are expressed as effective annual returns (ear) to facilitate comparison. the taxable equivalent ear of the mortgage swap and the municipal bonds is found by taking the specified ear of after-tax cash flows and dividing by (1-t). 99j. musumeci / financial services review 25 (2016) 87–104 dramatically from month to month, we compare the two mortgages every three years, starting in january, 1992 and ending in january, 2013. table 6 features the results. as in table 5, the advantage of a 15-year mortgage over a 30-year is generally greater when the overall rates are greater; here we see this is true, even when the gap between rates is smaller. for example, january, 2013 features the lowest rates, but the largest spread between 15-year and 30-year rates at 71 basis points. assuming the mortgage is liquidated in seven years, it also has the second lowest irr of any of the eight times examined. conversely, the january, 1992 and january, 1995 rates are the largest of the set, and offer the largest irrs for a mortgage liquidated in seven years, even though their spreads between the 15-year and 30-year rates are only about average. because our analysis has ignored the option to refinance, it is very likely that this increase in irr is because of the refinancing table 5 a comparison of npv and irr earned from a 30-year for 15-year swap for various levels of mortgage rates 15 year rate � 2.75%, 30-year rate � 3.5% house sold in no taxes tax rate � 33% npv irr (apr) npv irr (apr) 1 year $3,662.60 53.57% $2,465.93 51.79% 2 years $7,155.87 26.57% $4,839.26 25.75% 5 years $16,659.71 11.64% $11,412.36 11.30% 10 years $29,470.17 6.86% $20,593.15 6.66% 30 years $49,974.76 4.48% $36,556.52 4.35% 15 year rate � 3.75%, 30-year rate � 4.5% 1 year $3,648.16 56.39% $2,460.54 54.44% 2 years $7,103.96 28.42% $4,820.29 27.51% 5 years $16,382.30 12.96% $11,311.50 12.56% 10 years $28,571.62 8.01% $20,261.98 7.76% 30 years $47,039.79 5.56% $35,427.62 5.38% 15 year rate � 4.75%, 30-year rate � 5.5% 1 year $3,633.12 59.67% $2,454.71 57.51% 2 years $7,049.87 30.47% $4,799.58 29.45% 5 years $16,096.28 14.34% $11,201.28 13.88% 10 years $27,664.15 9.18% $19,903.38 8.89% 30 years $44,226.88 6.65% $34,239.17 6.42% 15 year rate � 5.75%, 30-year rate � 6.5% 1 year $3,617.55 63.45% $2,448.47 61.02% 2 years $6,993.91 32.74% $4,777.30 31.59% 5 years $15,803.66 15.81% $11,082.90 15.27% 10 years $26,755.14 10.40% $19,522.02 10.04% 30 years $41,550.23 7.75% $33,009.72 7.47% note. values for npv and irr are calculated in the same fashion as for tables 2 and 3, but for different levels of interest rates. npv declines when rates are higher because higher rates cause present to fall, but irr increases as rates increase. irr � internal rate of return; npv � net present value; apr � annual percentage rate. 100 j. musumeci / financial services review 25 (2016) 87–104 option’s greater value when interest rates are higher. the 15-year mortgage reduces the value of this option relative to the 30-year, and in exchange 15-year borrowers are compensated with a higher irr. 5. conclusions we have shown that jumbo loans and 30-year mortgages can be substantially more expensive than conforming loans and 15-year mortgages. what appears to be only a slight gap in the mortgage rates produces large net present values and internal rates of return of marginal cash flows, that is, cash flows of conforming loans compared with those of jumbo loans, and of 15-year mortgages relative to those of 30-year mortgages. the “conventional wisdom” often steers homeowners towards 30-year mortgages because of their “greater flexibility,” but that extra flexibility comes with a steep cost, particularly for homeowners who will liquidate the mortgage before maturity. we find, for example, that even on an after-tax basis, 15-year borrowers who will liquidate their mortgages at the median term of about seven years will earn a riskless internal rate of return of about 10% on their marginal cash flows relative to their 30-year borrower counterparts. notes 1 this is just an approximation in a multi-period context because the loans are paid down at different rates and so the weights change over time. for example, a 30-year, $100,000 mortgage at a rate of 0.40%/month requires payments of $524.67/month, but at 0.80%/month would require payments of $848.16/month. however, the two loans combined do not have an average rate of exactly 0.60%/month, but rather a monthly rate satisfying (524.67�848.16)(pvifa360,r%)�200,000, which is closer to 0.61%/ month. [throughout the paper we use pvifan,r% to denote the present value of an table 6 a comparison of 15-year and 30-year mortgages initiated in january, 1992 through january, 2013 mortgages initiated in 15-year 30-year irr if mortgage liquidated in apr points apr points 1 year 7 years 15 years 30 years january, 1992 8.01% 1.7 8.43% 1.8 43.38% 12.10% 9.58% 9.06% january, 1995 8.80% 1.8 9.15% 1.8 38.87% 12.26% 10.12% 9.67% january, 1998 6.58% 1.4 6.99% 1.4 37.75% 10.23% 8.00% 7.53% january, 2001 6.64% 0.9 7.03% 0.9 36.36% 10.12% 8.00% 7.54% january, 2004 5.02% 0.7 5.71% 0.7 54.09% 10.83% 7.33% 6.57% january, 2007 5.97% 0.4 6.22% 0.4 24.36% 8.13% 6.82% 6.53% january, 2010 4.44% 0.6 5.03% 0.7 54.33% 9.39% 6.41% 5.76% january, 2013 2.70% 0.7 3.41% 0.7 49.07% 8.27% 4.94% 4.21% note. specified irrs are for the marginal cash flows incurred by a 15-year mortgagor relative to those of a 30-year mortgagor. apr and points are taken from the freddie mac website. the borrower is assumed to have a 33% marginal tax rate. irrs are expressed as aprs and on a pre-tax basis, that is, pre-tax irr � (after-tax irr)/(1-tax rate). irr � internal rate of return; apr � annual percentage rate. 101j. musumeci / financial services review 25 (2016) 87–104 annuity of $1 per period at a discount rate of r% per peiord. this can be calculated as pvifan, r% � 1 r�1 � 1 �1 � r�n�, where r is expressed as a decimal, not a percent.] for a further discussion of this issue, and for conditions under which the relationship holds exactly, see miles and ezzell (1980). 2 a reservation price is generally defined to be the greatest price an individual is willing to pay for something, or the lowest price he will accept if he is the seller. in this context, it refers to the highest rate the mortgagor is willing to pay for a loan. 3 there are several differences in the approach taken here and that taken by goff and cox. first, while their assumptions of a 0.5% gap (7.5% vs. 8.0%) between 15-year and 30-year rates was conservative for its time—the freddie mac survey reports an average gap of only 0.36% (6.78% vs. 7.14%) in april, 1988—it is rather small by today’s standards. for example, the freddie mac survey indicates that as of january, 2014, the average gap was 95 basis points (3.48% vs. 4.43%). second, their framework mixes the fairly riskless cash flows from a mortgage with the relatively riskier cash flows from investing in equity (and while it is true that both their investors ultimately invest some marginal cash flows in equity, the 30-year borrower is doing so for a longer time and, therefore, is subject to more risk), while this article compares the riskless cash flows with each other by focusing on irr of the marginal cash flows. finally, while their analysis centered on a borrower who will keep the house for at least 30 years, we consider a number of possible dates on which the mortgage is prepaid. nevertheless, their article presents a number of conditions that favor the 30-year mortgage. 4 for example, while amromin, huang, and sialm (2007) focus on prepayment of existing mortgages rather than the choice between mortgages, they do classify “taking out a mortgage with a maturity shorter than the standard 30 years” as a form of “mortgage prepayment,” which they conclude is inferior to investing the marginal proceeds in a tax-deferred account. we find that, even when taxes are considered, the lower rate on a 15-year mortgage typically makes it a better choice. 5 more extreme theoretical examples are possible. for example, suppose the apr on the 30-year were 8.4%. now the monthly payments of $3812.30 on the 30-year mortgage would exceed those on the 15-year. clearly this is impossible in any reasonable kind of equilibrium, so that the rate on the 15-year mortgage establishes an upper bound on that of the 30-year. 6 fvifan, r% � 1 r ��1 � r�n � 1� denotes the future value of an annuity of $1 per year, and fvifn, r% � �1 � r�n denotes the future value of a dollar now compounded for n periods at r% per period. because we wanted the irrs for many different sets of cash flows, we used the irr function in ms excel to generate the answer to this (and all the irr entries in tables 2, 3, 5, and 6), but the cash flow registers on almost all financial calculators can also find the value of r � irr. 7 the procedure by which this gain is obtained is very similar to riding the yield curve (e.g., bieri and chincarini, 2005), with a couple of exceptions. riding the yield curve 102 j. musumeci / financial services review 25 (2016) 87–104 involves buying a long-term bond (and sometimes hedging it with sale—or shortsale—of a short-term bond) to earn higher returns. here the mortgagor is a net borrower (or seller of a bond), and so eschews the longer-term mortgage in favor of one with a shorter term, thus earning savings from the difference in rates. however, riding the yield curve, even in its hedged form, incurs some risk, while the process described here is riskless from the borrower’s perspective. 8 the amount the 15-year mortgagor still owes at the end of 84 months can be calculated as 3636.11(pvifa96, 0.3125%) or, equivalently, as 500,000(fvif84, 0.3125%) – 3636.11(fvifa84, 0.3125%). the calculation for the 30-year mortgagor is analogous, but with a periodic monthly rate of 0.375%. 9 in addition, they are probably the age bracket that faces the greatest uncertainty in their future cash-flow stream, as well as least likely to be able to afford the higher payments of a 15-year mortgage. goff and cox (1998) identify other advantages of the 30-year relative to the 15-year mortgage. acknowledgments the author is grateful to claude cicchetti, deborah gregory, atul gupta, colleen moore, mark peterson, len rosenthal, richard sansing, dave simon, meg steere, and participants in the 2013 academy of financial services conference for helpful comments. references amromin, g., huang, j., & sialm, c. (2007). the tradeoff between mortgage prepayments and tax-deferred retirement savings. journal of public economics, 91, 2014–2040. bieri, d., & chincarini, l. (2005). riding the yield curve: a variety of strategies. the journal of fixed income, 15, 6–35. bodie, z., merton, r., & samuelson, w. (1992). labor supply flexibility and portfolio choice in a lifecycle model. journal of economic dynamics and control, 16, 427–450. brealey, r., myers. s., & allen, f. (2008). principles of corporate finance (9th ed.) boston, ma: mcgrawhill/irwin. brueckner, j. (2000). mortgage default with asymmetric information. journal of real estate finance and economics, 20, 251–274. cheung, c., & miu, p. (2015). home ownership decision in personal finance: some empirical evidence. financial services review, 24, 51–76. copeland, t., weston, j. f., & shastri, k. (2005). financial theory and corporate policy (4th ed.). boston, ma: pearson/addison-wesley. cotterman, r., & pearce, j. (1996). the effects of the federal national mortgage association and the federal home loan corporation on conventional fixed-rate mortgage yields, studies on privatizing fannie mae and freddie mac. washington, d.c.: u.s. department of housing and urban development. dhillon, u., shilling, j., & sirmans, c. f. (1990). the mortgage maturity decision: the choice between 15-year and 30-year frms. southern economic journal, 56, 1103–1116. fortin, r., michelson, s., smith, s., & weaver, w. (2007). mortgage refinancing: the interaction of break even period, taxes, npv, and irr. financial services review, 16, 197–209. 103j. musumeci / financial services review 25 (2016) 87–104 freddie mac. (2010). freddie mac primary mortgage market survey (available from http://www.freddiemac. com/pmms/pmms30.htm; http://www.freddiemac.com/pmms/pmms15.htm). goff, d., & cox, d. (1998). 15-year versus 30-year mortgage: which is the better option? journal of financial planning, april, 88–95. grable, j., & lytton, r. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. harrison, d., noordewier, t., & yavas, a. (2004). do riskier borrowers borrow more. real estate economics, 32, 385–411. ibbotson, r., & sinquefield, r. (2013). stocks, bonds, bills, and inflation: 2013 classic yearbook. chicago, il: morningstar. kau, j., & keenan, d. (1995). an overview of the option theoretic pricing of mortgages. journal of housing research, 6, 217–244. miles, j., & ezzell, j. (1980). the weighted average cost of capital, perfect capital markets, and project life: a clarification. the journal of financial and quantitative analysis, 15, 719–730. puelz, r., & snow, a. (1994). evidence on adverse selection: equilibrium signaling and cross-subsidization in the insurance market. journal of political economy, 102, 236–257. reichenstein, w. (1998). calculating a family’s asset mix. financial services review, 7, 195–206. ross, s., westerfield, r., & jaffe, j. (2013). corporate finance (10th ed.) boston, ma: mcgraw-hill/irwin. samuelson, p. (1994). the long term case for equities. journal of portfolio management, 21, 15–24. thangavelu, p. (2015). how much house can you buy? (available at http://www.bankrate.com/finance/ mortgages/how-much-house-can-you-buy–1.aspx). 104 j. musumeci / financial services review 25 (2016) 87–104 financial services review, 31(4) 228 the big five personality traits (ocean) and financial planning: a narrative review and recommendations for advisors w. keith campbell,1 jim exley,2 and patrick c. doyle3 abstract financial planning has moved beyond a purely economic model and now incorporates aspects of behavioral economics and counseling psychology to better serve clients. in this review, we suggest that personality psychology, particularly the big five or ocean model of general personality, might also be useful in financial planning. financial planners understand different clients with different personalities bring different opportunities and challenges into the planning session, but planners might benefit from a more formal understanding of client personality. to this end, we describe the big five traits—openness to experience, conscientiousness, extraversion, agreeableness and neuroticism or ocean—and the basic personality science surrounding them. we next examine how each of the ocean traits is associated with key financial outcomes including: income, net-worth or wealth, financial literacy, financial risk tolerance, and financial happiness. we discuss profiles of the big five traits, including resilient, under controlled, and over controlled profiles. finally, we discuss some potential benefits associated with incorporating personality science into financial planning research and practice. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation campbell, w. k., exley, j., & doyle, p. c. (2023). the big five personality traits (ocean) and financial planning: a narrative review and recommendations for advisors. financial services review, 31(4), 228-245. introduction the field of economics assumes “rationality” in decision making (tucker, 2023). however financial planning professionals regularly observe what they perceive to be “irrationality” in decision making. it is clear that individuals are basing financial decisions on more than optimal economic outcomes. in response, the cfp board 1 corresponding author (wkeithcampbell@gmail.com). university of georgia, athens, georgia 2 wealth science analytics, alpharetta, georgia 3 wealth science analytics, alpharetta, georgia of standards has added psychology to the cfp exam curriculum with topics such as behavioral finance, interpersonal financial conflict, basics of counseling, communication, and coping with crises (cfp board, 2023). these topics can be categorized into two broad psychological fields of study: behavioral finance (also referred to as judgment and decision making or jdm) and counseling psychology. this is a great start for https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ campbell et al. 229 appreciating that individuals do not think about finances in purely rational ways and that discussions between a financial planner and a client often will tap into psychological issues and demand some level of counseling-like skills. noticeably missing from this list, however, is personality psychology. personality psychology defines and measures individual differences in personality traits, dynamics, and narratives, and relates these individual differences to important applied and theoretical outcomes. two recent working papers by the federal reserve (green et al., 2023) and the national bureau of economic research (jiang et al., 2023) suggest that economists are beginning to explore and research personality and financial outcomes, and a growing body of literature in financial planning, psychology, business, and related fields suggest an emergent interest in this topic. given our backgrounds, we have written this review from the perspective of personality psychology, but the movement toward understanding clients through the lens of personality traits is widespread across disciplines. for example, a recent meta-analysis (alderotti et al., 2023) examined over 60 peerreviewed articles (with nearly 900 effects analyzed) published between 2001 and 2020 that explored personality’s relationship to income alone. we imagine a more complete model of financial psychology to include (a) personality, (b) behavioral finance, and (c) counseling and therapy as shown in figure 1. figure 1. an integrated model of financial psychology the goal of this review is to introduce the field of financial planning to this third aspect of financial psychology, personality psychology. to this end, we will review the most widely accepted general model of personality: the big five model of personality, also known as the five factor model or by its acronym ocean, and its relationship with important financial outcomes. we hope to convince researchers and financial planners that an individual’s financial behavior—which may seem irrational or at least sub-optimal economically—may become more rational when factoring in the individual’s unique personality. (note: we can imagine an expanded financial psychology that includes other individual differences like attitudes, or numeracy [peters & bjalkebring, 2015], positive psychology like happiness, relationships science, cultural psychology, and even neuroscience; our argument for the utility of personality is not meant to be exclusive.) we will explore the ocean model in detail below, but briefly the big five ocean traits are: openness to experience, or a creative, aesthetic, and philosophical approach to life; conscientiousness, or a dutiful and organized approach to life; extraversion, or an outgoing or financial services review, 31(4) 230 assertive approach to life; agreeableness, or a kind and cooperative approach to life; and neuroticism, or an emotionally unstable and anxious approach to life. we hope that our arguments surrounding the importance of personality will not come as a surprise to financial planners. when a financial professional looks at a client base, they may notice that there are different kinds of people. some clients lean toward aggressive risk taking. they might work in fields like finance or real estate development. some clients are competent, thoughtful, and serious who have done well across a range of industries and professions. and there will be some anxious and uncertain clients who have surprising amounts of wealth. they may be people who inherited wealth or people who find themselves in unstable life situations. for example, clients who inherited money might show a different personality structure than someone who took chances to create first generation wealth. the vanderbilt family is an interesting example with the young “commodore” starting a dangerous boat business in new york harbor in his early teens as a precursor to his adventures in railroad building. generations that followed continued the family business, but with different personality traits and skills (smith, 2007). in the present article, we argue that these personality traits can be described and measured scientifically. further, these same personality traits predict important financial behaviors, from income to risk tolerance, and thus might be useful for financial professionals to consider. much like the financial industry agrees on general definitions of income, net-worth, cash flow, and even financial literacy, the world of psychology agrees on a general framework of personality. our hope is that this article will spur the financial industrial complex to explore, research, and potentially adopt existing psychological personality models in hopes of preventing a long process of recreating the wheel with less-thanideal measures that are not properly psychometrically vetted. the reward to the financial planning field for adopting big five personality psychology, for example, is over 3,000,000 peer reviewed research publications ready and waiting (google scholar, 2023). a brief overview of personality science dynamics, narratives, and traits personality can be thought of as a process or dynamic, like a person who self-sabotages new opportunities without knowing why, or someone who is constantly seeking status or attention and then feeling empty inside, only to repeat the pattern (baumeister, 2010; deyoung et al., 2014). understanding personality dynamics is central to much of classic psychotherapy and psychoanalysis. personality can also be thought of as a narrative, the story you tell about yourself (mcadams, 1993). the hero’s journey (campbell, 1949) is the most well-known of these narrative structures, but there are many patterns to the stories that we tell about ourselves. most commonly, though, personality can be thought of in terms of traits, or patterns of thoughts, feelings, and behaviors that are consistent across time and across situations (allport & odbert, 1936). trait approaches to personality are sometimes referred to as quantitative approaches because of the sophisticated statistical techniques developed to measure and compare different traits—and explore how traits are related to other outcomes like close relationships, health behaviors or job performance. there are many personality traits that have been studied—both more targeted traits like “type a” and “narcissism” and more general trait models like the eysenck’s (1994) three factor model (pen: psychoticism, extraversion, neuroticism), or lee and ashton’s (2004) six factor model (hexaco: honesty-humility, emotionality, extraversion, agreeableness, conscientiousness, and openness). our focus will be on the most prominent and integrative model in personality psychology, the big five or five factor model of ocean: openness, conscientiousness, extraversion, agreeableness, and neuroticism (goldberg, 1990; costa & mccrae, 1992). importantly, general personality traits are scored as continuous variables from low to high and are neither strictly “good” nor “bad” in terms of their benefits to the individual. instead, personality traits typically offer trade-offs in differing social contexts. for example, a person scoring high in self-control may exercise every day and resist temptations. a person with a lower self-control campbell et al. 231 score may be described as impulsive. a high selfcontrol score may be helpful for a student, while a lower self-control score may be more helpful for an expressive artist. averaged across contexts, though, a pattern of higher scores on all the ocean traits (with a lower score on neuroticism) is considered somewhat more adaptive in modern western cultures. advantages of trait approaches as noted, the key properties of personality traits are consistency across time and situations. what this means is that an individual will have a similar personality today and in six months or a year. also, that an individual will have a similar personality at work, at home, and at play. for example, somebody who is brave should be brave six months from now or in a few years. that person is likely to be brave across situations. they might provide leadership at work, they might fiercely defend their family’s values, and they might be adventurous at play. importantly, personality consistency is not absolute but probabilistic people are not machines but are constantly growing and changing. personality at time 1 correlates about .80 with personality two months later (gnambs, 2014). but the situation matters too. during a strong situation like a funeral or final exam you may see people act in similar ways, but at a less defined social gathering like a neighborhood block party are more likely to see a range of personality traits expressed. it is also important to know that there are some common developmental changes to personality traits. for example, when people get married or start jobs, they on average become more psychologically mature (bleidorn, 2015). another interesting finding is that people become less likely to worry about things in old age outside of major health issues (donnellan & lucas, 2008). this can be thought of as a sort of mellowing process that happens to many people. but even with these developmental changes, you will see that the rank order of traits is relatively stable. a highly impulsive teenager is likely to become more organized and self-disciplined with age and experience, but so are all other teenagers, so the relative self-discipline might stay the same even as the absolute level increases. the big five traits: ocean it turns out there are many different personality traits, which are apparent when one examines language. we have many ways to describe a person’s behavior or personality style such as energetic, cheerful, self-confident, domineering or kind. each of these words, in a sense, reflects a personality trait. of course, it is very hard to build a personality science around a thousand different individual personality traits—each with its own adjective. in response, what researchers have done over the last hundred years is figure out the best way to group these words or traits together into higher order combinations depending on the needs of the practitioners or researchers. this search was aided by the development of quantitative tools like factor analysis and the newfound ability to reduce traits into simpler structures. early personality models varied, with 13 to 5, and sometimes fewer, factors (goldberg, 1993). there is no best or right personality model, but what proved most useful over time in modern, largely western culture, is a five factor structure. the big five are “big” because the traits that they capture are so broad and “five” because it captures five traits—ocean: openness, conscientiousness extraversion, agreeableness, and neuroticism. it can be useful to consider these traits as five big buckets because they are large and hold a great number of adjectives. for example, the big five traits of extraversion include both sociability and desire to be with others but also achievement drive and ambition. extraversion is possibly the most misused as people tend to focus on the sociability portion of extraversion rather than the assertive portion. therefore, strong leaders who enjoy reflective time alone often misclassify themselves as introverts discounting their drive for achievement and leadership qualities. because these big five traits capture such a large amount of the naturally occurring personality space, most personality descriptions can be thought about as refined pieces of big five traits or combination of big five traits. for example, the type a personality has been shown to be primarily driven by neuroticism (bruck & allen, 2003). the trade-off is breadth versus financial services review, 31(4) 232 specificity—the big five is useful for getting a broad personality profile or overview, but if you are interested in specific traits, from risk-taking to creativity, it is typically better to use a purposebuilt measure. simply put, there is a lot of bang for the buck applying the thinking and research encompassing the big five personality model to new fields. theoretically, the big five model captures much of the human personality space and, practically, the big five predicts much of the human behavioral space. but this does not negate the importance of more specific personality tools— in fact, research in any field of social and behavioral sciences will have researchers using global personality measures like the big five, and also specific or niche measures that are highly important to that field. one well-known example is the trait of “grit,” or goal focused determination. grit is an aspect of the big five trait of conscientiousness, but many people simply target grit (duckworth et al., 2007). some readers may be more familiar with other popular models of personality—like the meyersbriggs or the enneagram. while these approaches are also oriented around giving people a vocabulary for discussing personality, scientists often use trait approaches instead of type approaches for methodological reasons. two people who have similar meyers-briggs profiles (two entjs, for example) may be very different in terms of how extraverted they are. despite being in the same group, the person with a high extraversion score may be much more willing to be the center of attention than the other who may have only barely crossed the cusp to an entj from an intj. the big five is not necessarily more right, but it’s certainly more specific. one of the challenges of this social-evolutionary process of personality construction is what is called the founder effect, which asserts that early personality models tend to stick around and shape the later ones. to give an extreme example, the sensing, thinking, feeling, and intuition that make up the very popular meyers-briggs scale came from the work of carl jung, and he borrowed them from classical elemental psychology: earth, air, fire, and water. so, an intuitive-thinking type could be a fire-air type. researchers and practitioners should thus avoid getting too fixated on which personality traits are “right” and instead appreciate that different cultural and psychological systems will produce different maps of personality the same way different cultures mapped similar skies into different constellations. principles of personality assessment how do personality traits get defined, operationalized, and measured? let's start with measurements. personality can be assessed in many ways, including structured interviews, observation, analyzing content from personal websites or twitter feeds, peer, spouse or employer reports, experimental tasks (e.g., gambling tasks), and projective tests (e.g., rorschach inkblots). each of these ways has different strengths and weaknesses and a case can be made for using a range of strategies and assessment in order to triangulate on individual personality. for example, if someone is doing research on a specific personality traits in leadership—or wanted to identify potential high level managers at an organization—they ideally want self-report personality data, 360-degree other-report data (e.g., reports from followers, peers, and supervisors, and even spouse or friends outside the organization), some historical data to see the stability of traits, some behavioral data like past job performance, etc. this type of dense personality measurement allows researchers or practitioners to triangulate on the key personality constructs of interest. unfortunately, this kind of research is highly expensive and timeconsuming, so it is reserved for large grant funded science or executive testing firms. even with all the options available, the simplest strategy for measuring personality is the selfreport questionnaire. self-report methods are remarkably accurate and straightforward. typically, self-report assessment involves individuals answering a series of questions about specific aspects of their thoughts, feelings, or behaviors using a multi-point numerical scale. in a low stakes environment where there is little motivation to lie on the test, individuals are willing and able to report on themselves quite readily. this low stakes environment is typical in academic research settings and in places like financial planning using general personality campbell et al. 233 measures. self-report can be more challenging in high stakes settings like forensic or legal settings and with some personnel selection. many non-academics are surprised at how well self-report works as it seems simplistic compared with lab-based tools like reaction time measures, cognitive response tasks, and neuroimaging. however, decades of research have demonstrated that self-report personality assessment is a reliable and stable scientific tool (paulhus & vazire, 2007). there are at least three good reasons why self-report personality measures work. first, researchers developed these tests so that they are worded with appropriate language and complexity so that most people can respond to them easily. in contrast, some risk tolerance measures ask about equities and bonds, terms many individuals cannot define. additionally, the word “risk” is more heavily associated with “danger” than opportunity and often has a negative connotation (clifton, 2022). second, very sophisticated psychometric techniques are used so that the smallest possible number of items can be used in the making of the self-report assessment. this reduces testing fatigue and test administration scheduling challenges. and third, self-report personality tests have been repeatedly validated and improved in the real world for over a century, often in military and business settings but also in social media (liu & campbell, 2017). we know these tests work because they have worked over and over. personality measurement is relatively straightforward. the trickier question is how researchers established the personality traits for which they would later build assessments. for example, “type a” personality exists in common language today. we can go to work and call somebody type a and there is a pretty good chance they'll know what we mean. this description of a personality trait grew out of the scientific literature—specifically health psychology and cardiology where researchers were attempting to identify personality factors correlating with hypertension (caplan & jones, 1975; rosenman, 1990). calling somebody “anal” or “anal retentive” is another old personality trait that made it into the common language. this term was started by sigmund freud (freud, 1932) based on his theory of libidinal types. the same can be said for the big five traits. extraversion was the term coined by the swiss psychiatrist carl jung, along with its opposite introversion. neuroticism is a classic psychiatric term used to describe emotionality. openness to experience was popularized by mccrae and costa (1997) to capture broad creative and philosophical mindedness (see mccrae, 1987). the other two, agreeableness and conscientiousness, are common words used in a technical way (for a history of the big five see goldberg, 1993). the key point is that traits are not natural types. they are not like teeth or bones. traits are constructed based on trends in human culture, and should be as they change and evolve. the name for the scientific process of developing and refining these personality traits is construct validation (cronbach & meehl, 1955). construct validation is an iterative or cyclical process whereby a concept, in this case a personality trait, gets defined, operationalized, and tested. this process results in a new understanding of the construct and better assessment tools and repeats continuously. to use an example from clinical psychology, melancholia was seen as an untreatable deep despair and that construct evolved into what we now call major depression, which is now seen as a treatable disorder. the point is that researchers in various scientific disciplines are constantly creating and refining psychological constructs and these constructs often overlap or compete with each other—which makes personality science a social and historical process. one of the advantages of trait-based approaches to personality—as opposed to dynamic or narrative approaches—is that the measurement of personality allows researchers to explore how all the different personality traits are related to each other. sometimes referred to as the nomological network, quantitative measurement of traits allows test developers to calculate the correlations between them, much like the lines that connect stars within a constellation. the result is a large web of interconnecting personality traits that relate to each other in meaningful ways. financial services review, 31(4) 234 the organization and structure of the big five traits the big five can be placed into a hierarchy from a more complex personality structure to a simpler personality structure. in the direction of a complex structure, the big five traits can be broken down into smaller units, called facets or aspects. for example, extraversion can be broken into a three-facet structure of assertiveness, energy level, and sociability. when we say that someone is extraverted, we mean that they exhibit high levels of these facets, but there can be some variability between them. for example, an extraverted leader might be assertive and energetic, but not as highly sociable. see table 1 for the big five along with three facets (soto & john, 2017) and figure 2 for a graphic representation of hierarchical personality structure. there is no agreed upon number of facets for the big five; there are models that have more facets, such as the neo that contains six facets per big five trait (costa & mccrae, 1992) and those with less, such as two aspects per big five trait (deyoung et al., 2007). different models are better for different uses, and much is at the discretion of the researcher. table 1. big five personality traits and their facets openness conscientiousness extraversion agreeableness neuroticism aesthetic sensitivity organization assertiveness compassion anxiety creative imagination productiveness energy level respectfulness depression intellectual curiosity responsibility sociability trust emotional volatility traits can also be organized into higher order models, like the big two that can refer to plasticity and stability (deyoung, 2006). extraversion and openness combine to form the meta-trait plasticity. an individual high in extraversion and high in openness might be described as flexible or having a high level of plasticity. thinking practically about this trait combination, an outgoing communicative person who is also open to imaginative new ideas may be more willing to change their mind than someone demonstrating lower levels of extraversion and openness. plasticity could be beneficial in artistic and creative endeavors such as music composition (deyoung, 2006). on the other hand, conscientiousness, agreeableness, and (low) neuroticism combine to form the meta trait of stability. an individual high in conscientiousness and low in neuroticism can be described as having stability and is likely to demonstrate calmness in difficult or stressful situations. stability could be a beneficial personality pattern for managers or modern astronauts (deyoung, 2006). at the highest level combining each of the five ocean traits, psychologists have found what is referred to as the big one personality (musek, 2007). the big one personality is an individual who demonstrates a high level of each of the big five personality traits excluding neuroticism. a big one personality would be highly open, highly conscientious, highly extraverted, highly agreeable, and exhibit low neuroticism. campbell et al. 235 figure 2. big five personality with higher and lower order factor structures there is no right or wrong level of abstraction for the big five traits. the big two of plasticity and stability are probably a little too big for assisting individuals in day-to-day activities but may provide an interesting way to look theoretically at specific behaviors such as investing. a fifteenfacet model of the big five can be useful for getting more nuance around the big five traits when needed. however, the big five model seems to strike the right balance between generality and utility in most cases. measurement scales for the big five as researchers and financial practitioners begin to more thoroughly implement personality into their work, there are many available big five measures in the literature. the most common include the big five inventory (bfi; john et al., 1991), bfi-2 (soto & john, 2017), which contains three facets per trait, and the neo (mccrae & costa, 1991) and neo-ipip (johnson, 2014; maples-keller et al., 2019) which contain six facets for each trait. there are also short measures of the big five. these include the very commonly used ten item personality scale (tipi) (gosling et al., 2003). the tipi is often used in large survey research and also used as an adjunct variable in studies when researchers are just curious about the function of personality. there is also a miniature version of the neo-ipip (donnellan et al., 2006). researchers need to consider the following when selecting a big five measure. first, use a wellestablished (i.e., valid and reliable) measure such as those listed. next, focus on length and availability of facets. longer scales are generally better statistically but cost more in time and fatigue. facets give more analytic flexibility, but that isn’t always theoretically necessary. finally, there is the question of open access. the bfi and neo-ipip are available without charge to researchers, as are the shorter scales. the big five personality is everywhere our focus of this paper is on financial behaviors, but we would be remiss in not noting that big five personality traits make themselves known everywhere. because big five personality is easy financial services review, 31(4) 236 to measure at scale using short self-report measures like the tipi, millions of people have provided personality data in exchange for feedback. these data, along with decades of focused research, has mapped the big five personality onto many aspects of life. the big five traits are related to the state that you live in. californians have been shown to demonstrate a collective higher level of openness while georgian’s have been shown to exhibit a higher level of conscientiousness (rentfrow et al., 2008). in fact, each of the 50 states has a predominant personality trait. the big five traits are related to how your office or home is decorated (gosling et al., 2008). your office— and your clients’—say something about the people who work there. big five traits relate to your language. highly open individuals will use words like “ideas” while highly conscientious individuals will use words like “hard work”. the big five traits are related to the music you prefer—for example, if you are high in openness, you are more likely to be more drawn to jazz or other improvisational and experimental music (rawlings & ciancarelli, 1997; rentfrow & gosling, 2003). the big five traits predict a wide range of occupational job performance and occupational preference (barrick & mount, 1991). big five traits correlate with your facebook profile and the number of friends you have on social media. social media companies are aware of an individual’s ocean traits which reflect what is shown on your screen (liu & campbell, 2017). in fact, ocean personality was seen in the background of a scene of the documentary the social dilemma (2020). and finally, most mental disorders can be described by the big five traits. for example, a prominent model known as the hitop (kotov et al., 2017) describes five “spectra” that correlate in part with the big five. and krueger et al. (2013) developed a measure of the pathological big five, the personality inventory for the dsm–5–which can be used to measure pathological extremes of the big five. in summary, the big five traits are projected throughout a person's life from what they say, to the music they listen to, to the state they live in, to their physical and mental health. it is no surprise, then, that the big five are linked to a range of important financial outcomes. the big five and financial outcomes financial psychology is broadly defined and includes a handful of classic variables that have been well studied and many others that have research interest. well studied variables include income, net-worth or wealth, financial literacy, financial risk tolerance, and financial happiness. there are a few different approaches to reviewing literature. we could either go through the financial process variables—income, net-worth, financial literacy, financial risk tolerance, financial happiness—and then discuss how each is related to the different big five traits. the other approach is to work through the big five traits and then examine how each individual trait is related to the range of financial outcomes. we are choosing the latter approach because we think this framework is most closely aligned with how these observations would likely occur in a financial planning setting. that is, a professional would collect and examine the ocean trait profile of an individual as well as a balance sheet, income and cash flow summaries. our reasoning for this format is that we and other researchers (e.g., bogan et al., 2020) perceive that the future of financial planning is primarily a client centered profession and not just a money centered profession, and focusing on client ocean personality is one way to center practice on the client. the cfp board's recent inclusion of psychology into cfp® training confirms this more client-centered orientation. openness the results for openness are mixed. interestingly, high openness has not shown a consistent correlation with higher income (judge et al., 1999; duckworth et al., 2012; exley et al., 2021) or net-worth (duckworth et al., 2012; nabishima & seay, 2015). however, with very high net worth clients you do see high openness scores (leckelt et al., 2019). this might reflect the openness of creative individuals who work in cutting-edge industries and live in elite urban environments. after all, silicon valley is in california where openness is higher (rentfrow et al., 2008). or it might reflect the association campbell et al. 237 between openness and entrepreneurship (zhao & seibert, 2006). high openness has shown mixed correlations with financial literacy. in general, openness seems to correlate with overall risk taking (nicholson et al., 2005) but more research is needed on financial risk taking specifically. also, openness does seem to be a predictor of overall happiness (furnham & petrides, 2003) but more research is needed on financial happiness. one challenge working with high openness clients is that they might seek novelty and may prefer to invest creatively rather than developing a discipline and refining that discipline. an individual with high openness may know a lot about finance and value creativity over results in the short run but be disappointed in not achieving their financial goals in the long run. trying new things may be a great way to learn but may not be a way to get repeatable results. we like to say that “openness is expensive”. understanding that a client is high in openness from the beginning of the relationship allows for conversations about these varying and often opposing outcomes. conscientiousness conscientiousness is what we conceptualize as the primary wealth trait. high conscientious individuals have consistently shown to have higher incomes (alderetti et al., 2023; fentono’creevy & furnham, 2023; exley et al., 2021; nabishima & seay, 2015; duckworth et al., 2012; judge et al., 1999), higher net worth (duckworth et al., 2012; exley et al., 2021; fenton-o’creevy & furnham, 2023; nabeshima & seay, 2015;) and higher financial literacy (exley et al., 2021; pinjisakikool, 2017; letkiewicz & fox, 2014). however, exley et al. (2021) demonstrated that individuals high in conscientiousness may not take enough risk financially and may simply be “gritting” themselves into higher net worth through hard work, higher income, and spending less. high conscientiousness individuals have been shown to be financially happier over time (joshanloo, 2022). understanding that a potential client is high in conscientiousness could lead a financial practitioner to design a plan that encourages spending and enjoyment, which may seem counterintuitive based on accepted professional practices. but overall, conscientious clients should take well to a consistent and disciplined approach to investing. extraversion extraversion is a bit of a trade-off. individuals high in extraversion have shown to be happier about their life in general (kim et al., 2018). a google scholar search (2023) of extraversion and income yields over 50,000 results with higher extraversion consistently predicting higher income (alderotti et al., 2023) and leadership (campbell et al., 2003). however, it seems that on average, extraverted individuals are better at acquiring income, but struggle with converting that income into wealth (fenton-o’creevy & furnham, 2023; exley et al., 2021). interestingly, individuals high in extraversion consistently demonstrate higher financial risk taking. in some samples, extraversion has correlated with lower financial literacy (e.g., exley et al., 2021; killins, 2017; pinjisakikool, 2017), which may provide clues to the lack of wealth creation. understanding a client’s extraversion is important, and as said above, is a bit of a tradeoff. a person high in extraversion may be willing to have less net worth for the tradeoff of fun and experiences. however, the financial planner and client need to be clear about the goal. agreeableness agreeableness is a conundrum. nice people are happy (deneve & cooper, 1998). agreeable people can make ideal friends and clients because they can be nice and cooperative, but they can suffer when it comes to finances. disagreeableness—a trait associated with narcissism and psychopathy—predicts income positively (alderotti et al., 2023). put another way, nice people make less money (judge et al., 2012), have lower net-worths (exley et al., 2021; fenton-o’creevy & furnham, 2023; weir & duckworth, 2012), and prefer less financial risk (pinjisakikool, 2017; wong & carducci, 2013). while financial literacy results have been mixed, a recent study found that nice people do have higher financial literacy (exley et al., 2021) but are unable to overcome the other confounding factors that their high agreeableness brings them. a practitioner understanding a client’s high financial services review, 31(4) 238 agreeableness might encourage them to take more financial risk if appropriate and possibly even ask for a raise at work. neuroticism neuroticism is associated with some of the larger problems and challenges in an individual’s financial life. people who are neurotic are more psychologically vulnerable to risk. being in an uncertain environment is experienced more painfully. investing demands the acceptance of some level of risk over time and neurotic individuals seem to prefer safer investments (oehler & wedlich, 2018). this may explain why people who are neurotic have challenges making money (alderotti, 2023) and have lower networth (exley et al., 2021; fenton-o’creevy & furnham, 2023; furnham, 2023; weir & duckworth, 2012). in higher interest rate environments where a reasonable return is offered by investments like cds, more neurotic individuals should do reasonably well. however, low interest rate environments that offer little in terms of risk-free returns will be a challenging for a more neurotic, fearful investing style. in addition, one study found individuals high in neuroticism also demonstrated lower financial literacy (exley et al., 2021). understanding a client’s neuroticism level may be one of the most important pieces of information a financial planner can gather based on neuroticism’s negative correlations with income, net worth, and financial risk tolerance. to make matters worse, people high in neuroticism may find it difficult to be financially happy over time (joshanloo, 2022). a financial planner who can provide a buffer to this neuroticism and keep their clients in higher risk assets over the longer term will be helpful. trait profiles the big five traits can also be assessed together in the form of a trait profile. statistical procedures like latent profile analysis and clustering can further be used to assess relatively stable profile patterns of personality across the general population. gerlach et al. (2018) identified four of these big five profiles which included what we call a “muted” profile with the big five traits being average across each of the five traits. beyond this muted profile, three classic profiles emerge: resilient, under controlled, and over controlled. these three are sometimes known as the arc types after the authors of the seminal studies that found three profiles: (asendordorpf et al., 2001; caspi et al., 1995; robins et al., 1996). specific to financial outcomes, a recent latent profile study using financial outcomes confirmed these three personality profiles— resilient, under controlled, and over controlled—supporting the arc model (exley et al., 2022). the pattern with the four more positive big five traits elevated (and neuroticism lowered) is seen in figure 3 and is called a resilient (exley et al., 2022) or role model gerlach et al., 2018). as the name suggests, in many cases this will be the healthiest and highest functioning personality profile. campbell et al. 239 figure 3. big five personality traits associated with big one, resilient, and role model profiles (data from exley et al., 2022) the second pattern identified has higher extraversion and neuroticism and lower agreeableness, conscientiousness, and openness as shown in figure 4. this is known as a under controlled (exley et al., 2022) or as a selfcentered type (gerlach et al., 2018). figure 4. big five personality traits associated with under controlled and self-centered profiles (data from exley et al., 2022) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 openness conscientiousness extraversion agreeableness neuroticism resillient/role model big one/resillient/role model 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 openness conscientiousness extraversion agreeableness neuroticism under controlled/self-centered under controlled/self-centered financial services review, 31(4) 240 the resilient profile is more similar to a high performing manager and the under controlled, a more highly competitive sales position or risktaker. both of these profiles can generate wealth, but both will require a little different approach to wealth management, with the former demanding high competence and professionalism and the latter demanding more external impulse control. a third profile—over controlled (exley et al., 2022) or reserved (gerlach et al., 2018) as shown in figure 5—has muted extraversion coupled with elevated agreeableness and neuroticism. this profile might be described as a “nervous nelly” and may need help taking risks to achieve their desired financial futures. figure 5. big five personality traits associated with over controlled and reserved profiles (data from exley et al., 2022) summary each of these trait profiles, of course, could benefit from different aspects of financial planning. a resilient client will need to have competently produced information and performance the same way a pro athlete needs a top trainer. this client might also need some encouragement to relax their conscientiousness from time to time to take more risk in their finances and enjoy their wealth. an under controlled or risky client, on the other hand, may need to calm their risk taking and benefit from the education, discipline, and conscientiousness that an advisor traditionally has provided. finally, an over controlled or fearful client will benefit from some extraversion and risk taking on the part of the advisor, or else they may invest their money in low-risk assets and have a hard time keeping up with inflation. trait profiles are a potentially useful way to approach personality from the perspective of financial planning as opposed to “one size fits all” financial advice that focuses on reducing risk. trait profiles show how two of the three groups might at times need to take more risk—not less. future research and applications the application of personality science to financial decision making is a relatively new and quite small research field. there is a need for a wide range of research simply to create foundational scientific literature. that said, there are a few areas that stand out for potential study. one idea is tailoring financial advice or advice giving to client personality. the idea is that clients with certain personality traits or profiles will benefit from different advice. for example, highly extraverted and lower conscientiousness 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 openness conscientiousness extraversion agreeableness neuroticism over controlled/reserved over controlled/reserved campbell et al. 241 clients might well benefit from being reined in at times. the job of the financial professional might also benefit by framing the development of an investment plan and portfolio that highlights, for example, its novelty for open clients, its aggressiveness for more antagonistic and extraverted clients, its responsibility and discipline for conscientious clients, and its longterm stability for more neurotic clients. obviously, financial planners do this intuitively all the time, but research might uncover some stable and teachable patterns. another step would be matching financial professionals to clients using client personality as one factor. for example, a more agreeable, neurotic client might benefit from a more therapeutic and supportive financial planner; whereas a more open client might do better with a wider read planner who can tolerate the clients harebrained schemes and novelty seeking. finally, personality can be used for selection, either selecting clients or employees. we are not suggesting using the big five as a selection tool; there are many considerations, including legal, that need to be taken to make any selection based on personality. but what we are suggesting is that any practice could be selecting certain kinds of clients. for example, a firm that attracts many state retirees will have clients with different personality profiles than a firm whose client list is heavy with urban professionals or with athletes and entertainers. it would be very useful for any practice to see if they are limiting themselves or perhaps becoming a niche for certain personality profiles. conclusion: a more complete financial psychology financial planning would be easy if it were just about building financial or accounting models— those models are simple to build, but it can be challenging for an individual to make their behavior conform to the model. the question of why individuals make poor and seemingly selfdestructive financial decisions on a regular basis has puzzled economists for centuries. the growth of the heuristics and biases literature is a testament to this and richard thaler winning the nobel prize in economics for demonstrating many of these heuristics and biases is well deserved. but beyond our shared heuristics, our different personalities predict different ways that we can self-destruct (or excel). there are predictable individual differences in financial performance—some people are risk-taking and aggressive and blow up, but others are fearful and take no risks and never acquire enough wealth to even blow it up, and still other people seem to have an almost supernatural discipline and calm that allows them to invest despite the chaos in the markets. each of these people will need different styles of support and planning. for example, the up and coming “wolf of wall street” will need some conscientiousness; the “nervous nelly” will need some confidence to buy equities for the long-term; and the “warren buffett junior” might need some encouragement to spend a little money having fun with his kids. we hope that in addition to learning about judgment and decision making and some of the basics of therapy, and especially the ability to identify psychological problems that might need referral to a mental health professional, understanding the basics of personality psychology will make financial planners and advisors more effective across a range of clients. and while there are many excellent models of personality available, we suggest that the big five or ocean model would be an excellent starting place for the field of financial psychology. references alderotti, g., rapallini, c., & traverso, s. (2023). the big five personality traits and earnings: a meta-analysis. journal of economic psychology, 94, 102570. allport, g. w., & odbert, h. s. (1936). traitnames: a psycho-lexical study. psychological monographs, 47(1), entire volume. asendorpf, j. b., borkenau, p., ostendorf, f., & van aken, m. a. (2001). carving personality description at its joints: confirmation of three replicable personality prototypes for both children and adults. european journal of personality, 15(3), 169-198. financial services review, 31(4) 242 barrick, m. r., & mount, m. k. (1991). the big five personality dimensions and job performance: a meta‐analysis. personnel psychology, 44(1), 1-26. baumeister, r. f. (2010). the self. oxford university press. bleidorn, w. (2015). what accounts for personality maturation in early adulthood? current directions in psychological science, 24(3), 245-252. bogan, v. l., geczy, c. c., & grable, j. e. (2020). financial planning: a research agenda for the next decade. financial planning review, 3(2), e1094. bruck, c. s., & allen, t. d. (2003). the relationship between big five personality traits, negative affectivity, type a behavior, and work–family conflict. journal of vocational behavior, 63(3), 457-472. campbell, j. (1949). the hero with a thousand faces: a brilliant examination, through ancient hero myths, of man’s eternal struggle for identity. fontana press. campbell, l., simpson, j. a., stewart, m., & manning, j. (2003). putting personality in social context: extraversion, emergent leadership, and the availability of rewards. personality and social psychology bulletin, 29(12), 1547-1559. caplan, r. d., & jones, k. w. (1975). effects of work load, role ambiguity, and type a personality on anxiety, depression, and heart rate. journal of applied psychology, 60(6), 713-719. caspi, a., & silva, p. a. (1995). temperamental qualities at age three predict personality traits in young adulthood: longitudinal evidence from a birth cohort. child development, 66(2), 486-498. cfp board. (2023). cfp® exam: what you'll be tested on. cfp® exam: what you'll be tested on | cfp board. retrieved april 23, 2023, from https://www.cfp.net/getcertified/certification-process/examrequirement/about-the-cfp-exam/whatyoull-be-tested-on clifton, j. (2022). blind spot: the global rise of unhappiness and how leaders missed it. gallup press. costa, p. t., & mccrae, r. r. (1992). normal personality assessment in clinical practice: the neo personality inventory. psychological assessment, 4(1), 5-13. cronbach, l. j., & meehl, p. e. (1955). construct validity in psychological tests. psychological bulletin, 52(4), 281302. deneve, k. m., & cooper, h. (1998). the happy personality: a meta-analysis of 137 personality traits and subjective wellbeing. psychological bulletin, 124(2), 197-229. deyoung, c. g. (2006). higher-order factors of the big five in a multi-informant sample. journal of personality and social psychology, 91(6), 1138-1151. deyoung, c. g., quilty, l. c., & peterson, j. b. (2007). between facets and domains: 10 aspects of the big five. journal of personality and social psychology, 93(5), 880-896. deyoung, c. g., quilty, l. c., peterson, j. b., & gray, j. r. (2014). openness to experience, intellect, and cognitive ability. journal of personality assessment, 96(1), 46-52. digman, j. m. (1997). higher-order factors of the big five. journal of personality and social psychology, 73(6), 1246-1256. donnellan, m. b., & lucas, r. e. (2008). age differences in the big five across the life span: evidence from two national samples. psychology and aging, 23(3), 558-566. donnellan, m. b., oswald, f. l., baird, b. m., & lucas, r. e. (2006). the mini-ipip scales: tiny-yet-effective measures of the big five factors of campbell et al. 243 personality. psychological assessment, 18(2), 192-203. duckworth, a. l., peterson, c., matthews, m. d., & kelly, d. r. (2007). grit: perseverance and passion for long-term goals. journal of personality and social psychology, 92(6), 1087-1101. duckworth, a. l., weir, d., tsukayama, e., & kwok, d. (2012). who does well in life? conscientious adults excel in both objective and subjective success. frontiers in psychology, 3, 356. exley, j., doyle, p., snell, m., & campbell, w. k. o.c.e.a.n.: how does personality predict financial success? journal of financial planning, 34(10), 68-86. exley, j., doyle, p. c., grable, j., & campbell, w. k. (2022). ocean wealth profiles: a latent profile analysis of personality traits and financial outcomes. personality and individual differences, 185, 111300. exposure labs, agent pictures, argent pictures, the space program, netflix. (2020). the social dilemma. extraversion and income. (2023). google scholar search retrieved april 30, 2023, from https://scholar.google.com/ eysenck, h. j. (1994). normality–abnormality and the three-factor model of personality. freud, s. (1932). libidinal types. the psychoanalytic quarterly, 1(1), 3-6. furnham, a., & petrides, k. v. (2003). trait emotional intelligence and happiness. social behavior and personality: an international journal, 31(8), 815-823. gerlach, m., farb, b., revelle, w., & nunes amaral, l. a. (2018). a robust datadriven approach identifies four personality types across four large data sets. nature human behaviour, 2(10), 735-742. gnambs, t. (2014). a meta-analysis of dependability coefficients (test–retest reliabilities) for measures of the big five. journal of research in personality, 52, 20-28. goldberg, l. r. (1993). the structure of phenotypic personality traits. american psychologist, 48(1), 26-34. gosling, s. d., rentfrow, p. j., & swann jr, w. b. (2003). a very brief measure of the big-five personality domains. journal of research in personality, 37(6), 504-528. gosling, s. d., gaddis, s., & vazire, s. (2008). first impressions based on the environments we create and inhabit. in n. ambady & j. j. skowronski (eds.), first impressions (pp. 334–356). guilford publications. greene, c., shy, o., & stavins, j. (2023). personality traits and financial outcomes. frb of boston working paper no. 23-4. hahn, e., gottschling, j., & spinath, f. m. (2012). short measurements of personality–validity and reliability of the gsoep big five inventory (bfis). journal of research in personality, 46(3), 355-359. jiang, z., peng, c., & yan, h. (2023). personality differences and investment decision-making (no. w31041). national bureau of economic research. john, o. p., donahue, e. m., & kentle, r. (1991). the big five inventory: versions 4a and 54 [technical report]. berkeley: university of california, institute of personality and social research. johnson, j. a. (2014). measuring thirty facets of the five factor model with a 120-item public domain inventory: development of the ipip-neo-120. journal of research in personality, 51, 78-89. joshanloo, m. (2022). personality trait level and change predict future financial well-being: a longitudinal study in australia. personality and individual differences, 191, 111575. https://www.proquest.com/indexinglinkhandler/sng/author/exley,+jim,+cfp/$n;jsessionid=bd6e5862eb99415a47d6fcc4825e27e8.i-0103e21a79d84736d https://www.proquest.com/indexinglinkhandler/sng/author/doyle,+patrick,+phd/$n;jsessionid=bd6e5862eb99415a47d6fcc4825e27e8.i-0103e21a79d84736d https://www.proquest.com/indexinglinkhandler/sng/author/snell,+michael/$n;jsessionid=bd6e5862eb99415a47d6fcc4825e27e8.i-0103e21a79d84736d https://www.proquest.com/indexinglinkhandler/sng/author/campbell,+w+keith,+phd/$n;jsessionid=bd6e5862eb99415a47d6fcc4825e27e8.i-0103e21a79d84736d https://www.proquest.com/indexinglinkhandler/sng/author/campbell,+w+keith,+phd/$n;jsessionid=bd6e5862eb99415a47d6fcc4825e27e8.i-0103e21a79d84736d financial services review, 31(4) 244 judge, t. a., livingston, b. a., & hurst, c. (2012). do nice guys—and gals—really finish last? the joint effects of sex and agreeableness on income. journal of personality and social psychology, 102(2), 1-18. judge, t. a., higgins, c. a., thoresen, c. j., & barrick, m. r. (1999). the big five personality traits, general mental ability, and career success across the life span. personnel psychology, 52(3), 621652. killins, r. n. (2017). the financial literacy of generation y and the influence that personality traits have on financial knowledge: evidence from canada. financial services review, 26(2), 143-165. kim, h., schimmack, u., oishi, s., & tsutsui, y. (2018). extraversion and life satisfaction: a cross‐cultural examination of student and nationally representative samples. journal of personality, 86(4), 604-618. kotov, r., krueger, r. f., watson, d., achenbach, t. m., althoff, r. r., bagby, r. m., ... & zimmerman, m. (2017). the hierarchical taxonomy of psychopathology (hitop): a dimensional alternative to traditional nosologies. journal of abnormal psychology, 126(4), 454-477. krueger, r. f., derringer, j., markon, k. e., watson, d., & skodol, a. e. (2013). the personality inventory for dsm–5 (pid–5)–adult (full version). washington, dc: american psychiatric association. leckelt, m., richter, d., schröder, c., küfner, a. c., grabka, m. m., & back, m. d. (2019). the rich are different: unravelling the perceived and self‐reported personality profiles of high‐net‐worth individuals. british journal of psychology, 110(4), 769-789. lee, k., & ashton, m. c. (2004). psychometric properties of the hexaco personality inventory. multivariate behavioral research, 39(2), 329-358. letkiewicz, j. c., & fox, j. j. (2014). conscientiousness, financial literacy, and asset accumulation of young adults. journal of consumer affairs, 48(2), 274-300. liu, d., & campbell, w. k. (2017). the big five personality traits, big two metatraits and social media: a metaanalysis. journal of research in personality, 70, 229-240. maples-keller, j. l., williamson, r. l., sleep, c. e., carter, n. t., campbell, w. k., & miller, j. d. (2019). using item response theory to develop a 60-item representation of the neo pi–r using the international personality item pool: development of the ipip–neo–60. journal of personality assessment, 101(1), 4-15. mcadams, d. p. (1993). the stories we live by: personal myths and the making of the self. guilford press. mccrae, r. r. (1987). creativity, divergent thinking, and openness to experience. journal of personality and social psychology, 52(6), 1258-1265. mccrae, r. r., & costa jr, p. t. (1991). the neo personality inventory: using the five‐factor model in counseling. journal of counseling & development, 69(4), 367-372. mccrae, r. r., & costa jr, p. t. (1997). conceptions and correlates of openness to experience. in handbook of personality psychology (pp. 825-847). academic press. musek, j. (2007). a general factor of personality: evidence for the big one in the five-factor model. journal of research in personality, 41(6), 1213-1233. nabeshima, g., & seay, m. (2015). wealth and personality: can personality traits make your client rich. journal of financial planning, 28(7), 50-57. campbell et al. 245 nicholson, n., soane, e., fenton‐o'creevy, m., & willman, p. (2005). personality and domain‐specific risk taking. journal of risk research, 8(2), 157-176. oehler, a., & wedlich, f. (2018). the relationship of extraversion and neuroticism with risk attitude, risk perception, and return expectations. journal of neuroscience, psychology, and economics, 11(2), 63-92. paulhus, d. l., & vazire, s. (2007). the selfreport method. handbook of research methods in personality psychology, 1(2007), 224-239. peters, e., & bjalkebring, p. (2015). multiple numeric competencies: when a number is not just a number. journal of personality and social psychology, 108(5), 802-822. pinjisakikool, t. (2017). the effect of personality traits on households’ financial literacy. citizenship, social and economics education, 16(1), 39-51. rawlings, d., & ciancarelli, v. (1997). music preference and the five-factor model of the neo personality inventory. psychology of music, 25(2), 120-132. rentfrow, p. j., & gosling, s. d. (2003). the do re mi's of everyday life: the structure and personality correlates of music preferences. journal of personality and social psychology, 84(6), 1236-1256. rentfrow, p. j., gosling, s. d., & potter, j. (2008). a theory of the emergence, persistence, and expression of geographic variation in psychological characteristics. perspectives on psychological science, 3(5), 339-369. robins, r. w., john, o. p., caspi, a., moffitt, t. e., & stouthamer-loeber, m. (1996). resilient, overcontrolled, and undercontrolled boys: three replicable personality types. journal of personality and social psychology, 70(1), 157-171. rosenman, r. h. (1990). type a behaviour pattern: a personal overview. journal of social behavior and personality, 5(1), 124. smith, a. d. h. (2007). commodore vanderbilt: an epic of american achievement. cosimo, inc. soto, c. j., & john, o. p. (2017). the next big five inventory (bfi-2): developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power. journal of personality and social psychology, 113(1), 143. stone, a. a., bachrach, c. a., jobe, j. b., kurtzman, h. s., & cain, v. s. (eds.). (1999). the science of self-report: implications for research and practice. psychology press. tucker, i. b. (2023). macroeconomics for today. cengage. wong, a., & carducci, b. (2013). does personality affect personal financial risk tolerance behavior? iup journal of applied finance, 19(3), 5-18. zhao, h., & seibert, s. e. (2006). the big five personality dimensions and entrepreneurial status: a meta-analytical review. journal of applied psychology, 91(2), 259-271. financial services review, 33(1) 102 retail investors and investment fraud victims: is there a connection? christopher rand,1 melisande mccrae,2 and jason martin3 abstract this study analyzed specific characteristics of investment fraud victims. logistic regressions on a national sample of retail investors revealed that overconfident and financially literate investors shared several characteristics with victims of investment fraud. while overconfident investors were the most comfortable with market regulation and making investment decisions that assumed high amounts of risk relative to investment returns, financially literate investors surpassed them in the frequency of annual trading and portfolio allocation to stocks. surprisingly, overconfident investors favored due diligence via background checks on investment professionals, while financially literate investors did not. overall, males and younger investors tended to share characteristics with investment fraud victims. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation rand, c., mccrae, m., & martin, j. (2025). retail investors and investment fraud victims: is there a connection? financial services review, 33(1), 102-119. introduction according to the federal trade commission (2022), investment scams were among the top five types of fraud reported by 2.4 million consumers. moreover, investment scams exacted the greatest financial cost to consumers: a loss of $3.8 billion––more than double the amount reported lost in 2021 and a staggering $4.6 billion in 2023 to investment scams, the highest fraud category that year (federal trade commission, 2023). with a greater reliance on web-based systems and virtual account statements, consumers are more susceptible to this type of crime (shadel & pak, 2017). 1 corresponding author (crand@berkeley.edu). uc berkeley extension, berkeley, ca, usa. 2 independent consultant-research design, philadelphia, pa, usa. 3 swarthmore college, swarthmore, pa, usa. nevertheless, as a research focus, investment scams pose a challenge, which may be due to how the literature has defined fraud, the many different types of fraud, and the inconsistent reporting of fraud (lee et al., 2019). according to the federal bureau of investigation (fbi), investment fraud is defined as the illegal sale of financial products (federal bureau of investigation [fbi], 2020). this illegal activity has taken many forms, such as market manipulation, pyramid or ponzi schemes, pumpand-dump schemes, and affinity fraud targeting specific groups, to name a few (fbi, 2020). furthermore, although kieffer and mottola (2016) predicted that one in ten investors would https://creativecommons.org/licenses/by-nc/4.0/ mailto:crand@berkeley.edu https://creativecommons.org/licenses/by-nc/4.0/ rand et al. 103 be affected by investment fraud at some point in their lives, the actual percentage of individuals defrauded remains unclear. for example, deliema et al. (2017) found that 16.5% of study participants reported investment fraud, while nearly half of survey respondents reported being victimized by another type of fraud. interestingly, although deliema et al.’s (2017) investment fraud research contradicted more conservative estimates, other studies have suggested that underreporting might account for the confusion surrounding the actual number of investment fraud cases. for example, applied research & consulting llc (2015) found that over two-thirds of victims told friends or family about their victimization, yet only 35% reported it to authorities. almost half of those who did not report it to the authorities stated that it would not have made a difference, while 35% indicated that they wanted to put the occurrence behind them. moreover, 29% reported feeling embarrassed by the experience, while nearly half stated that they felt responsible for being victimized. these factors may play a role in underreporting fraud occurrences (applied research & consulting llc, 2015). despite the lack of precision in defining and tracking instances of investment fraud, extensive research has been conducted on the demographic background of fraud victims. typically, victims are older, married, and male, with many classified as unsophisticated investors (deliema et al., 2020; lokanan, 2014; lokanan & liu, 2021). research has also shown that investment fraud victims’ behaviors and attitudes are common. for example, victims take on more investment risk and are more impulsive than non-victims (knutson & samanez-larkin, 2014; shadel & pak, 2017). victims tend to be more trusting than non-victims when making investment decisions, which might account for a failure to exercise due diligence. additionally, victims of investment fraud tend to regard wealth as a measure of success (deliema et al., 2020; lokanan, 2014; shadel & pak, 2017). hence, further investigation is warranted regarding a possible connection between these characteristics and investment fraud vulnerability. while the aforementioned demographics, behaviors, and attitudes represent factors that may increase vulnerability to investment fraud, financial literacy, which is generally low in the united states, is assumed to safeguard against victimization. specifically, financial literacy is an understanding of financial concepts and behaviors necessary to effectively manage personal finances (lusardi & mitchell, 2014). however, the literature is mixed on the adequacy of financial literacy to protect consumers against investment fraud. for example, engels et al. (2021) found that advanced financial knowledge was more effective than basic money management skills in detecting fraud. therefore, one purpose of this study is to determine the role of financial literacy in mitigating attitudes and behaviors associated with investment fraud victims. in addition to objective knowledge and skills, financial literacy is comprised of individuals’ perceptions of competence in managing personal finances. for example, financially literate consumers and knowledgeable investors have sufficient investment knowledge and skills as well as a level of confidence proportionate to their abilities. in contrast, investors are considered overconfident when their selfperceptions exceed their actual abilities. according to the literature (e.g., knutson & samanez-larkin, 2014; kramer, 2014; nussbaumer et al., 2009), overconfident investors engage in behavior that may render them vulnerable to investment fraud. as a result, another purpose of the present study is to explore the extent to which overconfidence predisposes investors to fraud. finally, investors differ in their experiences with the financial markets and the subsequent judgments they form about their capabilities to achieve their investment goals (bandura, 2001; chiu & klassen, 2009). consequently, this study also explores how gender and age differences in self-efficacy may impact susceptibility to investment fraud (farrell et al., 2016; mcavay et al., 1996). this study builds on the literature to identify characteristics of fraud victims, the purpose of which is twofold: financial services review, 33(1) 104 1. to determine the extent to which financially literate and overconfident investors share characteristics with victims of investment fraud; 2. to explore the extent to which gender and age render investors potentially vulnerable to victimization. literature review investment fraud has continued to rise as fraudsters have learned new and innovative ways to take advantage of individuals (shadel & pak, 2017). the term “fraud” has been broadly defined and is likely partly why much fraud goes unreported in the united states. this study specifically investigates investment fraud and the potential impact of investors’ knowledge and confidence on their susceptibility to victimization. specifically, certain attitudes and behaviors are more associated with investment fraud victims than individuals who have not been defrauded. these characteristics are explored in detail following a discussion of investment fraud victim demographics, the possible role of financial literacy in investment fraud prevention, overconfidence in matters that pertain to money management, and how our experiences shape our beliefs in our abilities as investors. investment fraud victim demographics when shadel and pak (2017) studied the differences between fraud victims and the general public, they confirmed previous research that typical victims tended to be male, married, older, and had a higher incidence among veterans. similarly, deliema et al. (2020) found that victims were three times more likely to be male with a higher likelihood of victimization among older individuals, yet fraud was vastly unacknowledged regardless of the victim’s age. additionally, lokanan (2014) found that many victims were classified as unsophisticated investors, retired or in management positions, lacked wealth, and knew their offender as either a family member, acquaintance, or friend. in the context of investing, fraud victims tend to be more trusting than non-victims and fail to conduct background checks on investment advisors as a result. compared to the general public, fraud victims trade at a higher frequency, assume greater risk in investments they believe will yield greater returns, and participate in nonregulated or novel investment opportunities (deliema et al., 2020; lokanan, 2014; shadel & pak, 2017). additionally, victims of investment fraud have reduced impulse control (knutson & samanez-larkin, 2014). financial literacy lusardi and mitchell (2014) provided a widely accepted definition of financial literacy as “the ability to process economic information and make informed decisions about financial planning, wealth accumulation, pensions, and debt.” (p. 2). other researchers have also contributed to the definition of financial literacy by including concepts such as financial marketplaces, financial products, and related services (e.g., anderson et al.). financial literacy encompasses money management behaviors associated with due diligence, such as checking financial statements for errors. nevertheless, it is unclear whether financial literacy adequately protects against fraud. for example, engels et al. (2021) sampled 5,500 u.s. residents and found that prudent financial behaviors such as budgeting and paying bills on time were insufficient to detect fraud, which the researchers defined as unauthorized access to personal accounts (e.g., bank accounts, insurance accounts, and credit and debit cards). however, engels et al. (2021) also found that the higher the individual’s actual financial knowledge, the greater the ability to detect fraud. moreover, kasim et al. (2024) found evidence of a relationship between financial literacy and awareness of investment scams. nevertheless, the national association of securities dealers investor education foundation (2006), which looked specifically at investment fraud, found that fraud victims had higher financial literacy scores than non-victims. ironically, these findings contradicted their hypothesis going into the study. consequently, the mixed results of prior studies warrant further investigation. overconfidence asaad (2020) defined overconfidence as individuals’ perceptions of their abilities to be better than their performance on objective rand et al. 105 measures. importantly, studies have found that individuals who are overconfident in their abilities related to financial knowledge share many characteristics with victims of investment fraud. for example, asaad (2020) found that overconfident investors were more likely to have riskier behaviors and beliefs than other knowledge groups. specifically, the study found that overconfident participants took more risk in investing and were overoptimistic about market performance. moreover, overconfident individuals were more likely to trust market regulations and more likely to invest in higherrisk strategies such as options or trading on margin. overconfident individuals were also less likely to utilize a financial advisor. consistent with prior research, kramer (2014) found overconfidence in investment decisions to be experienced differently by men versus women. this evidence indicated that women were less likely to reflect this overconfident behavior, which could reduce some of their risk factors for investment fraud. self-efficacy differences according to gender and age participation in financial markets requires a certain comfort level with the complex and dynamic nature of investing. unlike general confidence, an investor’s comfort level is shaped by experiences within the domain. individuals’ assessments of their personal capabilities of success (i.e., self-efficacy) are marked by the degree of assurance they have in their competence to perform at a level to satisfy specific goals (bandura, 2001). significantly, self-efficacy determines individuals’ choices of goals, the degree of effort expended toward achieving them, and how long they persist, especially when a given task becomes challenging (bandura, 1977). gender personal finance, generally, and in the investment industry, specifically, are dominated by male role models. for instance, approximately 80% of chartered financial analysts (cfas) and 70% of certified financial planners (cfp®s) are male (blayney, 2016; fender et al., 2016). as such, considerably more males than females communicate financial information in the context of money management. moreover, schunk and dibenedetto (2021) asserted that role models who share similar characteristics with observers can positively impact the observers’ motivation to succeed in a given task. furthermore, men typically assume the lead role in managing household finances, which undoubtedly bolsters their self-efficacy in handling money matters. d’acunto (2015) sought to determine if gender was a predictor of risk tolerance and the amount of money at risk in investing. focusing on identity stereotypes, d’acunto (2015) found that compared to women, men were more tolerant of risk and invested more often when their identity stereotype was primed, while women experienced no difference when primed compared to men. this outcome was also the case when the men were primed or had their identity threatened while acting as the agent of a principal (d’acunto, 2015). thus, in light of the historic leaning toward men as the primary actors in the world of finance, it stands to reason that males are motivated by a greater sense of agency in achieving financial goals than females. age with the transition from traditional defined benefit pension plans to defined contribution plans like the 401(k), the baby boom generation has become responsible for managing large sums of money (lusardi & mitchell, 2014). however, as an individual’s self-efficacy varies across domains, feelings associated with selfdetermination and mastery of a given area can also wax and wane with age. specifically, mcavay et al. (1996) found that older adults experienced a steady decline in financial selfefficacy throughout a longitudinal study. consistent with this age-based decline in financial self-efficacy, shadel and pak (2017) found that there were more investment fraud victims over the age of 70 compared to the public. deliema et al. (2020) also reported a similar finding where investment fraud victimization was positively associated with age. in contrast, other studies have found that younger age groups are more likely to be victims of fraud (lee et al., 2019). as stated earlier, mixed results in financial fraud research may likely be due to financial services review, 33(1) 106 inconsistent definitions of fraud or a focus on financial products designed for specific age cohorts. nevertheless, this study focused on investment fraud specific to individuals who own investments. moreover, our research was based on the presumption that older investors are more likely to share characteristics with victims of investment fraud than younger investors due to a greater accumulation of financial assets. theoretical framework the opportunity model of predatory victimization deliema et al. (2020) utilized the opportunity model of predatory victimization in their research and posited that certain attitudes and behaviors that are common in fraud victims make them an attractive target for offenders. similarly, this study argues that the characteristics associated with fraud victims are also present in overconfident investors while mostly absent in financially knowledgeable investors. specifically, cohen et al. (1981) argued that the risk of victimization increases with the following five factors: proximity to potential offenders, guardianship, target attractiveness, exposure, and definitional properties of specific crimes. in other words, individuals’ routine activities and lifestyles may place them and their property at risk for victimization without due diligence, professional advice, self-restraint, and wisdom. therefore, the current study explores the extent to which the following factors are associated with financially literate and overconfident investors while noting nuances by gender and age: target attractiveness, guardianship, exposure, and proximity. when the opportunity model of predatory victimization is applied to the current study, an investor with low financial knowledge and high confidence is likely to be an attractive target to fraudsters (deliema et al., 2020). one reason for target attractiveness is that overconfident investors tend to utilize guardians less as their confidence level makes them believe they do not need someone looking out, aiding, or performing due diligence for them (nussbaumer et al., 2009). in addition to target attractiveness, overconfident investors also tend to increase their exposure to fraudsters through behaviors like purchasing nonregulated investments, purchasing investments after attending a free dinner seminar, trading at a high frequency, allocating assets primarily to equities, or buying from a cold caller (deliema et al., 2020; odean, 1998; trinugroho & sembel, 2011). moreover, the overconfident investor is likelier to miss cues that other investors may notice and reduce their proximity to criminals. when investor age is considered in the context of the opportunity model of predatory victimization, both older and younger investors are posited to be at risk of investment fraud but to a lesser extent than overconfident investors. for example, such factors as a greater accumulation of assets coupled with social isolation and/or cognitive decline may increase the target attractiveness of elderly investors to fraudsters (boyle et al., 2012; burnes et al., 2017). in contrast, young investors are believed to have increased exposure to investment fraud opportunities because of higher trading frequency and a tendency to assume greater investment risk independent of experienced investors. concerning gender, females are posited to have lower target attractiveness to perpetrators of investment fraud because of lower perceived selfefficacy in the domain of personal finance and because they are less likely to make investment decisions compared to males. in contrast, males have increased target attractiveness due to their predominance in the financial markets, higher risk tolerance, greater self-efficacy in the personal finance domain, and probable susceptibility to gender priming. when financially knowledgeable investors are added to the framework, figure 1 illustrates that females are likely at the lowest end of the investment fraud risk spectrum, followed by financially knowledgeable men. these knowledgeable men are likely to take on more investment risk than women, but since their confidence level aligns with their abilities, the risk level remains on the lower end of the spectrum (d’acunto, 2015; kramer, 2014). rand et al. 107 figure 1. theoretical framework visualization low risk high risk gender age females theoretical framework risk of investment fraud males younger adults older adults older adults younger adults females males research questions this study explores the extent to which overconfident and financially literate investors possess attitudes and behaviors known to increase investment fraud vulnerability. • does financial literacy protect against investment fraud? • does overconfidence predispose investors to investment fraud? • how do gender and age affect investor vulnerability to investment fraud? these research questions are addressed through eight hypotheses, which examine areas of investor concern and confidence about fraud, investor attitude toward risk, investor risk level, and investors’ information-seeking behaviors. hypotheses this study posits the following hypotheses to address the research questions: h1: worry about personal vulnerability to investment fraud – high-knowledge adults (independent of gender or age), followed by overconfident women and overconfident older adults, are more worried about personal vulnerability to investment fraud than overconfident men and overconfident younger adults. h2: level of confidence in financial market regulation – high-knowledge adults (independent of gender or age), followed by overconfident females and overconfident older adults, have less confidence in financial market regulation than overconfident males and overconfident younger adults. h3: comfort level in making investment decisions – high-knowledge adults (independent of age and gender) and overconfident women are less comfortable making investment decisions relative to overconfident men and overconfident adults of all ages. h4: amount of investment risk assumed relative to expected return – highknowledge investors (independent of age and gender), and overconfident women and overconfident older adults take on less investment risk than overconfident males and overconfident younger adults. financial services review, 33(1) 108 h5: annual trading frequency – highknowledge adults (regardless of gender), highknowledge older adults, overconfident women, and overconfident older adults, report trading less frequently than overconfident men and overconfident or high-knowledge younger adults. h6: allocation to stocks in a portfolio – highknowledge adults (regardless of gender), highknowledge older adults, overconfident women, and overconfident older adults, report having a lower allocation to stocks than overconfident men and younger adults who are either in the overconfident or high-knowledge groups. h7: using a financial advisor for information gathering – both females and older adults with high investment knowledge report greater use of financial advisors for gathering information about investments relative to females and older adults in the overconfident group along with men and younger adults in the high investment knowledge group. overconfident men and overconfident younger adults report using financial advisors the least. h8: performance of a background check on a financial professional – females and older adults with high investment knowledge report conducting higher numbers of background checks on a financial professional relative to overconfident females, overconfident older adults, men with high investment knowledge, and younger adults with high investment knowledge. overconfident men and overconfident younger adults report the lowest background checks on financial professionals. methods the investor survey of the finra foundation’s 2018 national financial capability study was utilized to test the hypotheses. all respondents were investors in the united states. the 2018 national financial capability study was the fourth wave, initially collected in 2009, with over 25,000 adults participating each time it was administered. the study’s objective was to benchmark key indicators of individuals’ financial capabilities and evaluate how their characteristics varied by financial literacy, demographics, attitudes, and behaviors (finra investor education foundation, 2020). since the study focused on factors that could affect retail investors, such as vulnerability to investment fraud, the 2018 national financial capability study investor study was uniquely qualified for the analysis. the participants all owned retail non-retirement investment accounts, with most also owning retirement accounts. the survey provided many questions concerning factors known to increase individuals’ vulnerability to investment fraud with subjective and objective measures of financial knowledge. the objective and subjective questions allow the researcher to create overconfident and knowledgeable investor variables to test the hypotheses (finra investor education foundation, 2020). at the time of the analysis, a more current survey with similar questions was completed. however, we decided to proceed with the 2018 survey since the 2021 survey was collected during the pandemic, and it was unclear what effect the pandemic would have on investors’ survey responses. in july 2018, arc research pulled from three online panels used by the state-by-state survey–– emi online research, survey sampling international, and research now––who sent 3,750 email invitations to participants of the 2018 state-by-state survey who identified as owning investments outside of retirement accounts. notably, 2,763 individuals opened the survey link, 2,003 of whom met the final survey criteria and completed the survey (finra investor education foundation, 2018b; nfcs response statistics, 2018). asaad (2020) used the 2015 national financial capability study investor survey to research investor overconfidence. asaad compared the five knowledge question answers of the national sample to their respective questions in the investor survey and noted that the investor survey participants correctly answered all five questions more often than the larger, more diverse sample. the comparison indicated that the investor survey participants had a higher level of financial literacy than the general population. rand et al. 109 variables the study selected eight variables to analyze as dependent with one independent variable, as described below. the control variables were race, income, education level, and approximate total value of the investors’ non-retirement accounts. details for the variables are provided in tables 1 and 2 below. independent variable. the independent variable was coded into three categories based on knowledge and confidence level: overconfident, high-knowledge, and reference group. the variable coding is similar to asaad’s (2020) methodology, which employed four knowledge and confidence groups based on the methodology of allgood and walstad (2013), who used a broader dataset: the 2009 national financial capability study. the respondents were separated into groups based on their level of objective knowledge and subjective confidence. next, the respondents were assessed based on their overall knowledge about investing on a scale from 1 to 7, with 1 = very low knowledge and 7 = very high knowledge. respondents who did not know or preferred not to answer were excluded from the analysis. their responses were then compared to the group mean. those who scored higher than the mean were placed in the high-confidence group, while those below the mean were placed in the lowconfidence group. table 1 below represents the placement of the investors into one of the three categories. “highknowledge” referenced those with a high level of knowledge regardless of their confidence level. “overconfident” were investors who had a low actual knowledge level but high confidence. the reference group represented investors with low actual knowledge and low investor confidence. table 1. three knowledge and confidence groups dependent variables. the dependent variables were divided into investor attitudes or investor behavior categories to address the research questions. the attitudes category consisted of the degree of worry about investment fraud vulnerability, the confidence level in financial market regulation, the comfort level in making investment decisions, and the amount of investment risk assumed relative to return. the behaviors category consisted of the frequency of trading, the allocation to stocks, using a financial advisor for information, and the risk-taking level. table 2 below shows how dependent variables were used to test each hypothesis. knowledge high high knowledge 1,093 (54.95%) low reference 406 (20.41%) overconfident 490 (24.64%) low high confidence financial services review, 33(1) 110 table 2. hypotheses testing and correlating variables relationship variables. this study investigated the effects of two relationship variables: age and gender. most of the literature shows that older individuals are more likely to become fraud victims than those in younger age groups, with men more likely to be fraud victims than women (shadel & pak, 2017). this study sought a deeper understanding of the effects of age and gender on the dependent variables. direction of effect. the study visualized the intended direction of effect as seen in table 3 below: hypothesis question non-focus group responses focus group responses h1 degree of worry about investment fraud vulnerability 4-jan 7-may h2 level of confidence in financial market regulation 7-jan 10-aug h3 comfort level in making investment decesions 7-jan 10-aug h4 amount of investment risk assumed relative to return average risk and return or no risk willing to take above average risk for above average return h5 trading frequency 1-10 times per year 11 or more times per year h6 allocation to stocks less than half of portfolio more than half of portfolio h7 use of financial advisor no yes h8 performed background check on finacial professional no yes rand et al. 111 table 3. direction of effect visualization results descriptive statistics most individuals in the sample of respondents who graduated from college indicated white as their race, indicated male for gender, were older, and had a higher household income than the population in general. the survey results confirmed low investor knowledge in the united states, with only a little over one-third of the respondents accurately answering more than half of the ten survey questions. investors were confused about the cost of their investments, with almost one-third believing that mutual funds did not have fees or expenses. most investors believed they had access to appropriate information for decision-making, and they were more likely to overestimate than underestimate their investment performance, with men being more confident than women in their abilities. table 4 lists the recoded variables selected from the 2018 investor survey to test the researchers’ hypotheses. financial services review, 33(1) 112 table 4. summary statistics characteristics percent number of observations coded = 1 coded = 0 dependent variables worried about fraud victimization 32% 1,989 634 1,355 confidence in regulation 30% 1,989 587 1,402 comfort making investment decisions 42% 1,989 842 1,147 risk willing to take 36% 1,960 699 1,261 investment trading frequency 15% 1,989 287 1,702 allocation to stocks 59% 1,777 1,041 736 use of advisor for information gathering 60% 1,989 1,196 793 background check on advisor 20% 1,989 393 1,596 variables of interest overconfident investor 25% 1,989 490 1,499 high knowledge investor 55% 1,989 1,093 896 reference group 20% 1,989 406 1,583 overconfident male investor 13% 1,989 256 1,733 overconfident female investor 12% 1,989 234 1,755 overconfident age 55 and older investor 12% 1,989 246 1,743 overconfident under 55 investor 12% 1,989 244 1,745 high-knowledge male investor 37% 1,989 730 1,259 high-knowledge female investor 18% 1,989 363 1,626 high-knowledge 55-and-older investor 38% 1,989 755 1,234 high-knowledge under-55 investor 17% 1,989 338 1,651 interaction variables age 55 and older 63% 1,989 1,244 745 males 57% 1,989 1,128 861 control variables investors with $250,000+ investments 34% 1,989 676 1,313 race = white alone 82% 1,989 1,627 362 college graduates 57% 1,989 1,125 864 income $100,000/yr+ 34% 1,989 673 1,316 statistical analysis the study’s empirical model employed logistic regression to examine the vulnerability of overconfident and high-knowledge investors to investment fraud while evaluating the interaction of gender and age and controlling for income, investment account value, race, and education. since the independent variables in the study were categorical, the max re-scaled r2 was utilized. logistic regressions were run for the eight hypotheses, which presented mixed results. the study employed the max re-scaled r2 statistic to measure the models’ explanatory power. the following is a summary of the results for hypotheses 1–8. table 5 below shows the top three groups per model with statistical significance. financial services review, 33(1) 113 table 5. top three groups per model with statistical significance supportive of hypothesis? model 1 model 2 model 3 model 4 model 5 overconfident overconfident men overconfident younger adults nonwhite adults overconfident women nonwhite adluts younger adults nonwhite adults overconfident older adults overconfident overconfident men overconfident younger adults high knowledge overconfident women overconfident older adults large investment account high knowledge high knowledge overconfident overconfident men high knowledge overconfident women men high knowledge younger adults overconfident men overconfident younger adults overconfident younger adults high knowledge high knowledge high knowledge men high knowledge high knowledge high knowledge high knowledge men high knowledge younger younger adults overconfident men overconfident younger high knowledge women high knowledge older large investment account younger adults nonwhite adults younger adults overconfident high knowledge high knowledge high knowledge high knowledge men high knowledge younger men overconfident men large investment account high knowledge women high knowledge older college graduate large investment account college graduate large investment account large investment account large investment account large investment account large investment account large investment account large investment account reference group reference group reference group reference women reference younger white adults reference men white adults overconfident overconfident men overconfident younger overconfident overconfident younger overconfident women younger high knowledge younger younger high knowledge men nonwhite adults h7 use of financial advisor to gather info no support h8 performance of background check on no support h5 number of investment trades made partial support h6 allocation of stocks in portfolio partial support h3 level of comfort in making full support h4 level of risk assumed relative to investment full support h1 worry about investment fraud no support h2 level of confidence in markets full support note: model 1 analyzes the impact of overconfidence, knowledge level, age, gender, and control variables; model 2 subdivides overconfident investors by gender; model 3 subdivides overconfident investors by age; model 4 subdivides high knowledge investors by gender; model 5 subdivides high knowledge investors by age. attitudes of retail investors the results supported three of the four research hypotheses focusing on attitudes associated with investment fraud vulnerability. consistent with one of the predictions, overconfident investors reflected a higher confidence level in the effectiveness of u.s. regulation of financial markets than other groups in the sample. this effect was evident in the overconfident group and when testing the interaction of investors’ gender and age. specifically, men (more than women) and younger (more than older) investors in the overconfident group felt that u.s. financial markets were effectively regulated. this sentiment supported the hypothesis that overconfident investors, especially male or younger investors, are potentially more vulnerable to investment fraud. in other words, compared to their high-knowledge counterparts, overconfident investors appeared to have an unwavering faith in the financial markets, which could blind them to signals of fraudulent activity. concerning investor comfort level in making investment decisions, regardless of age or gender, overconfident adults were more comfortable making investment decisions than other groups in the sample. moreover, when testing gender interaction, overconfident men had the highest level of comfort in making investment decisions. importantly, investors’ confidence in their abilities to interpret the data related to investments without sufficient financial knowledge or outside assistance from an investment professional could place them at a higher risk of victimization. financial services review, 33(1) 114 furthermore, the amount of investment risk investors assume based on the expected return could also render them vulnerable to investment fraud. specifically, the study results indicated that younger investors were willing to take on the highest level of risks for high returns, followed by overconfident investors. this finding was consistent with existing literature, which found that younger investors were willing to assume more investment risk due to a longer time horizon for recovery from investment losses to achieve their goals. additionally, an interaction between gender and age was observed among overconfident investors. overconfident men and overconfident younger adults were willing to take on the highest degrees of risk for return. while this finding was also consistent with predictions, the study results regarding investors’ levels of worry about vulnerability to investment fraud were contrary to expectations. in fact, the opposite of the research hypothesis was found, with overconfident investors having the highest worry about investment fraud vulnerability. moreover, when the interaction of age and gender was tested, overconfident men (more than women) and overconfident younger (more than older) investors had the highest degree of worry about investment fraud vulnerability. behaviors of retail investors when the remaining four hypotheses associated with behaviors that may increase vulnerability to investment fraud were tested, the results indicated partial support for two of the four research hypotheses. contrary to the hypothesis predicting that younger investors in the overconfident and high-knowledge groups and overconfident men would have the highest trading frequencies, high-knowledge investors, regardless of gender, had the highest trading frequencies, followed by overconfident men. when the analysis was modeled to evaluate the interactions of gender and age, the results suggested that men traded more than women, and younger investors traded more than older investors in the high-knowledge and overconfident groups. while the study results provided partial support for the trading frequency of overconfident males, the higher trading frequency of high-knowledge investors was unanticipated and is discussed in the next section. although allocating a significant portion of an investment portfolio to equities is common among overconfident investors, the results for hypothesis 6 lacked statistical significance for most of the models related to the overconfident groups. nevertheless, model 2 had statistical significance for the overconfident group and partially supported hypothesis 6. specifically, overconfident men had higher stock allocations than the reference group. in the remainder of the models, the overconfident group results were higher than the reference group, except for model 2 (overconfident women), but without statistical significance at the 0.05 level. as with the hypothesis 5 results, the highknowledge group had the highest allocations to equities in portfolios, with high-knowledge men more than women and high-knowledge younger more than older investors. while this result was also unanticipated, the behaviors of overconfident men and high-knowledge younger investors were consistent with the hypothesis that the overconfident younger group allocated more investment portfolios to equities than overconfident older investors but without statistical significance. this combined evidence suggests that overconfident and highknowledge male investors who are younger may be at a higher risk of investment fraud than other groups. finally, the study results did not support hypothesis 7 (investor use of a financial advisor for information gathering) or hypothesis 8 (investor performance of background checks). in both instances, high-knowledge investors who were rand et al. 115 either female or older were predicted to use financial advisors and conduct background checks. however, those most likely to use a financial advisor for information gathering were investors with the largest investment accounts and investors who identified as white in two of the models. surprisingly, the overconfident group was most likely to perform a background check on a professional, which was contrary to the hypothesis. in light of these findings, it is important to mention that the r2 was very low for hypotheses 7 and 8 and could only explain approximately 4–5% of the outcomes. conclusion investment fraud is becoming more prevalent, which results in significant financial loss and emotional trauma for consumers. this analysis aimed to examine the extent to which financially knowledgeable and overconfident investors shared characteristics with victims of investment fraud. consistent with study predictions, overconfident investors shared the following three attitudes with victims of investment fraud: confidence in the regulation of markets, comfort with making investment decisions, and willingness to assume high risk for an investment return. in addition, males and younger investors tended to lead these results compared to females and older investors. however, contrary to predictions, overconfident investors were more worried about investment fraud vulnerability than financially knowledgeable investors. moreover, overconfident investors were more likely to conduct background checks on financial professionals. while some results were perplexing, overconfident investors may be beginning to heed messages concerning the potential pitfalls of investor hubris. in other words, although overconfident groups have a high level of confidence in market regulation and are willing to assume more risk due to their higher comfort level with investments, they may retain a level of concern. overconfidence is portrayed negatively in the literature. while most research supports that overconfident individuals’ attitudes put them at a higher risk for investment fraud, this evidence may indicate a change in their attitudes, which has added a level of worry. another factor could be the timing of the data collection. july 2018 was the middle of the first stock market pullback, with the eight previous years yielding positive u.s. market returns, six of which provided double-digit returns. thus, the respondents’ concerns could have been heightened due to the timing of the survey. with many overconfident individuals unaware of their overconfidence, guardians may be best positioned to protect overconfident investors. advisors, healthcare providers, family, and friends who can look out for investors may see fraud signals the overconfident investor would not notice. the evidence provided by hypothesis 8 (i.e., overconfident investors are more likely to perform a background check on a professional) may be a step in the right direction for the overconfident group if the reason for performing the background check is to use a professional more often. the new finra (2018) rule allowing financial institutions to collect a trusted contact on investment accounts may also allow the guardian more options when they notice behavior likely to increase fraud risk. with the research providing evidence of increased fraud risk for overconfident investors, financial advisors and planners should look for signs of overconfidence in the clients they serve. with this new knowledge, advisors should pay closer attention to their overconfident investors’ accounts and actions to act as guardians to protect the interests of their investors. indeed, these investors are likely unaware financial services review, 33(1) 116 that they are at an increased risk of fraud due to their overconfidence. the findings that high-knowledge investors are also at an increased risk for investment fraud vulnerability may call for more research in this area. current financial literacy research indicates that the solution to the lack of overall financial literacy is providing more education to increase literacy. however, additional education does not appear to solve the vulnerability of investment fraud. highknowledge investors have behaviors and attitudes that place them at a higher risk of investment fraud vulnerability. importantly, the likely solution is the type of education investors receive, not the frequency. investors need to be better informed about the risks related to investment fraud to reduce exposure and avoid specific activities. fraud theory notably, this research looked specifically at the extent to which financially literate and overconfident investors shared characteristics with victims of investment fraud, not the actual incidence of or likelihood of being targeted for actual fraud. hence, individuals can be more vulnerable to being investment fraud victims based on their attitudes and/or behaviors common in investment fraud victims while not actually being targeted and victimized. current fraud theory, like the opportunity model of predatory victimization, may not be adequate to fully explain investment fraud (cohen et al., 1981). some of this research’s models had outcomes that could be explained for different reasons. for example, highknowledge investors were found to have the highest vulnerability to investment fraud based on their behavior of trading at a high frequency and having a higher allocation to stocks. research shows that both behaviors are common in fraud victims but could also be interpreted as a way to earn a higher rate of return on portfolios (shadel & pak, 2017). a knowledgeable investor would know that investing a higher percentage of their portfolio in stocks, which could also require more trading throughout the year, historically presents a higher overall rate of return for a portfolio. additional research should consider factors that may have a predictive ability for actual investment fraud incidence. limitations the 2018 national financial capability study investor survey was administered in july 2018 and represented participants’ feelings at a specific point in time, which happened to fall in the middle of a down year for u.s. equities. indeed, financial decisions can be about more than just money. thus, some factors might not have been captured in the responses to the survey (asaad, 2020). the investor survey subsample of the wider national financial capability study stateby-state survey did not represent the u.s. population. hence, individuals who did not own non-retirement accounts or were less wealthy might not have provided the same responses to the narrower data set analyzed (asaad, 2020). therefore, this context must be considered when generalizing this study to the broader u.s. population. future research the literature on fraud research continues to be limited. this study provides evidence that investor confidence influences behaviors and attitudes concerning vulnerability to fraud. however, the evidence is based on a small subset of the population at one point in time. an expansion of the analysis is warranted to continue to test the hypotheses and better evaluate the outcomes of hypotheses 1, 7, and 8. a larger study design that focuses on more common attitudes and behaviors among fraud victims and asks questions about actual fraud incidence is warranted. is there a correlation between an overconfident investor and the various types of fraud? are they at a higher or lower risk of investment fraud than a scam? why is current financial literacy education insufficient to reduce investment fraud vulnerability within the financially literate group of investors? the risk of rand et al. 117 fraud continues to rise, and the cost, especially with the baby boomer generation aging, will likely rise as well. therefore, more research in this area should be a priority for regulators and legislators (metlife study of elder financial abuse, 2011; shadel & pak, 2017). references allgood, s., & walstad, w. (2013). financial literacy and credit card behaviors: a cross-sectional analysis by age. numeracy, 6(2), 1–26. https://doi.org/10.5038/1936-4660.6.2.3 anderson, a., baker, f., & robinson, d. t. (2017). precautionary savings, retirement planning and misperceptions of financial literacy. journal of financial economics, 126(2), 383-398. https://doi.org/10.1016/j.jfineco.2017.07 .008 applied research & consulting llc. (2015, march). non-traditional costs of financial fraud: report of survey findings. finra investor education foundation. https://www.finrafoundation.org/sites/fi nrafoundation/files/nontraditional-costsfinancial-fraud_0_0_0.pdf asaad, c. t. (2020). investor confidence: are you your own worst enemy? financial planning review, 3(2), e1092. https://doi.org/10.1002/cfp2.1092 bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84(2), 191–215. https://doi.org/10.1037/0033295x.84.2.191 bandura, a. (2001). social cognitive theory of mass communication. media psychology, 3(3), 265–299. https://doi.org/10.1207/s1532785xmep 0303_03 blayney, e (2016). what is the future of women in financial planning? journal of financial planning, 29(9), 32–33. https://www.financialplanningassociatio n.org/article/journal/ sep16-whatfuture-women-financial-planning boyle, p. a., yu, l., wilson, r. s., gamble, k., buchman, a. s., & bennett, d. a. (2012). poor decision making is a consequence of cognitive decline among older persons without alzheimer’s disease or mild cognitive impairment. plos one, 7(8), e43647. https://doi.org/10.1371/journal.pone.004 3647 burnes, d., henderson, c. r., jr., sheppard, c., zhao, r., pillemer, k., & lachs, m. s. (2017). prevalence of financial fraud and scams among older adults in the united states: a systematic review and metaanalysis. american journal of public health, 107(8), e13–e21. https://doi.org/10.2105/ajph.2017.30382 1 chiu, m. m., & klassen, r. m. (2009). calibration of reading self-concept and reading achievement among 15-yearolds: cultural differences in 34 countries. learning and individual differences, 19(3), 372–386. https://doi.org/10.1016/j.lindif.2008.10. 004 cohen, l. e., kluegel, j. r., & land, k. c. (1981). social inequality and predatory criminal victimization: an exposition and test of a formal theory. american sociological review, 46(5), 505–524. https://doi.org/10.2307/2094935 d’acunto, f. (2015). identity, overconfidence, and investment decisions. georgetown university. https://doi.org/10.2139/ ssrn.2641182 deliema, m., mottola, g. r., & deevy, m. (2017). findings from a pilot study to measure financial fraud in the united states. stanford center on longevity and the finra investor education foundation. https://doi.org/10.2139/ssrn.2914560 deliema, m., shadel, d., & pak, k. (2020). profiling victims of investment fraud: mindsets and risky behaviors. journal of consumer research, 46(5), 904–914. https://doi.org/10.1093/ jcr/ucz020 financial services review, 33(1) 118 engels, c., kumar, k., & philip, d. (2021). financial literacy and fraud detection. the european journal of finance, 26(4– 5), 420–422. https://doi.org/10.1080/1351847x.2019. 1646666 farrell, l., fry, t. r., & risse, l. (2016). the significance of financial self-efficacy in explaining women’s personal finance behaviour. journal of economic psychology, 54, 85–99. https://doi.org/10.1016/j.joep.2015.07.0 01 federal bureau of investigation (fbi). (2020). scams and safety. www.fbi.gov/scamsand-safety/common-fraudschemes/investment-fraud federal trade commission. (2022). ftc crunches the 2022 numbers. see where the scammers continue to crunch consumers. www.ftc.gov/businessguidance/blog/ 2023/02/ftc-crunches2022-numbers-see-where-scammerscontinue-crunch-consumers federal trade commission. (2023). as nationwide fraud losses top $10 billion in 2023, ftc steps up efforts to protect the public. https://www.ftc.gov/newsevents/news/pressreleases/2024/02/nationwide-fraudlosses-top-10-billion-2023-ftc-stepsefforts-protect-public fender, r., adams, r. b., barber, b. m., & odean, t. (2016). gender diversity in investment management: new research for practitioners on how to close the gender gap. cfa institute research foundation 2016b-5. https://rpc.cfainstitute.org//media/documents/ article/rf-brief/rfbrv2-n5-1.pdf finra. (2018). new finra rules take effect to protect seniors and vulnerable adults from financial exploitation. http://www.finra.org/ newsroom/2018/new-finra-rules-takeeffect-protect-seniors-financialexploitation finra investor education foundation. (2018a). financial capability by state. www.usfinancialcapability.org finra investor education foundation. (2018b). investor survey methodology. https://www.usfinancialcapability.org/d ownloads/ nfcs_2018_inv_meth.pdf finra investor education foundation. (2020). about the national financial capability study. https://www.usfinancialcapability. org/about.php forrest, t. l. (2007). logistic regression models for predicting trip reporting accuracy in gps-enhanced household travel surveys [master’s thesis, texas a&m university]. https://oaktrust.library.tamu.edu/bitstrea m/handle/1969.1/4667/etd-tamu-2005cgeog-forrest.pdf;sequence=1 kasim, e. s., awalludin, n. r., zainal, n., ismail, a., & ahmad shukri, n. h. (2024). the effect of financial literacy, financial behaviour and financial stress on awareness of investment scams among retirees. journal of financial crime, 31(3), 652–666. https://doi.org/ 10.1108/jfc-04-2023-0080 kieffer, c. n., & mottola, g. r. (2016). understanding and combating investment fraud. finra investor education foundation. https://pensionresearchcouncil.wharton. upenn.edu/wpcontent/uploads/2018/08/11.-kiefferand-mottola.pdf knutson, b. & samanez-larkin, g. (2014, april). individual differences in susceptibility to investment fraud. stanford university and yale university. https://www.finrafoundation.org/sites/fi nrafoundation/files/individualdifferences-in-susceptibility-toinvestment-fraud.pdf kramer, m. m. (2014). financial literacy, overconfidence and financial advice seeking. journal of economic behavior rand et al. 119 & organization, 131(a), 198–217. https://doi.org/10.1016/j.jebo.2016.08.0 16 lee, s. j., cummings, b. f., & martin, j. (2019, january 31). victim characteristics of investment fraud. 2019 academic research colloquium for financial planning and related disciplines. https://doi.org/10.2139/ssrn.3258084 lokanan, m. e. (2014). the demographic profile of victims of investment fraud. journal of financial crime, 21(2), 226–242. https://doi.org/10.1108/jfc-02-2013004 lokanan, m. e., & liu, s. (2021). the demographic profile of victims of investment fraud: an update. journal of financial crime, 28(3), 647–658. https://doi.org/10.1108/jfc-09-20200191 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52(1), 5–44. mcavay, g. j., seeman, t. e., & rodin, j. (1996). a longitudinal study of change in domain-specific self-efficacy among older adults. the journals of gerontology series b: psychological sciences and social sciences, 51(5), p243–p253. https://doi.org/ 10.1093/geronb/51b.5.p243 metlife study of elder financial abuse. (2011). metlife mature market institute. https://www.giaging.org/documents/mm i-elder-financial-abuse.pdf national association of securities dealers investor education foundation. (2006, may 12). investor fraud study final report. the consumer fraud research group. http://www.sec.gov/news/press/extra/se niors/nasdfraudstudy051206.pdf nfcs response statistics. (2018). survey samples and response statistics. https://www.usfinancialcapability.org/d ownloads/2018responsestats.pdf nussbaumer, p., slembek, i., lueg, c., mogicato, r., & schwabe, g. (2009, april). understanding information seeking behaviour in financial advisory. 11th international symposium on information science (isi 2009), konstanz, germany, april 1–3, 2009 (session 4, part 2). https://doi.org/10.5281/zenodo.5506547 odean, t. (1998, april). volume, volatility, price, and profit when all traders are above average. the journal of finance, 53(6), 1887–1934. https://faculty.haas.berkeley.edu/ odean/papers/vvpp/vvpp.pdf schunk, d. h. & dibenedetto, m. k. (2021). self-efficacy and human motivation. in a. j. elliot (ed.) advances in motivation science (vol. 8) (pp. 153–179). elsevier. https://doi.org/10.1016/bs.adms.2020.10 .001 shadel, d., & pak, k. (2017). aarp investment fraud vulnerability study. american association of retired persons. https://www.aarp.org/pri/topics/workfinances-retirement/fraud-consumerprotection/investment-fraud-survey.html trinugroho, i., & sembel, r. (2011). overconfidence and excessive trading behavior: an experimental study. international journal of business and management, 6(7), 147. https://doi.org/10.5539/ijbm.v6n7p14 7 about:blank about:blank finser_30_4_merged_updated_3 financial services review 30 (2022) v from the editor dear esteemed readers, in this edition of the financial services review (volume 30 issue 4), we present five impactful papers, each contributing unique insights into various financial issues. james a. dilellio and andreas simon, in their paper “seeking tax alpha in retirement income,” offer a pioneering framework for optimizing tax efficiency in retirement income. by classifying retirees based on their financial circumstances, they uncover a new aspect of retire ment income strategy—tax alpha, which can add a 0.5% annual return benefit. they also sug gest ways for institutions and fintech firms to enhance their financial planning tools. turning our focus to the factors influencing retirement savings, brian t. starr’s “the effects of health, family, and altruism on retirement savings” illuminates the influence of health, children, altruism, and family attributes. this expanded understanding of retirement savings dynamics highlights the impact of children and health issues on savings behaviors. next, “financial capability across generations and technology” by abeba mussa, meeghan rogers, and xu zhang deeply dives into the relationship between technology sav viness and financial behavior. using data from the 2018 national financial capability study, their research identifies a generational disparity in financial behavior influenced by technology. tsung-ming yeh’s study titled “financial teaching by parents and financial education at school or workplace: evidence from japan” delineates the differential roles of financial socialization and formal financial education. his findings underscore the importance of com prehensive financial experience and education at various stages of life. finally, c. w. copeland, john h. young, and crystal r. hudson critically examine african-americans’ financial well-being in their paper, “exploring differences in african americans’ financial well-being based on financial security factors.” through examining homeownership and employment—two crucial variables historically influenced by racism— their research provides a nuanced understanding of the financial challenges faced by middle income african americans. each of these research papers adds significantly to our understanding of various aspects of financial behavior, and we hope they will benefit our readers. yours sincerely, terrance k. martin jr. editor financial services review 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. finser_23_1 low-income employees: the relationship between information from formal advisors and financial behaviors crystal r. hudson, ph.d.a,*, lance palmer, ph.d., cpa, cfpb adepartment of finance, clark atlanta university, 223 james p. brawley drive, atlanta, ga 30314, usa bdepartment of financial planning, housing and consumer economics, university of georgia, 205 dawson hall, athens, ga 30602-2622, usa abstract this study investigates the financial literacy of low-income employees, by examining their financial behaviors. thus, researchers examine the effect that information from formal advisors has on the financial behaviors of low-income employees. in this study, formal advisors include financial planners, bankers, brokers, employers, accountants, insurance agents, and lawyers. using data from the 2010 survey of consumer finances, researchers find a significant and positive relationship between the use of information from formal advisors and low-income employees’ positive financial behaviors. in other words, low-income employees who use information from formal advisors exhibit better financial behaviors than those who do not. © 2014 academy of financial services. all rights reserved. jel classification: d14 keywords: financial behaviors; low-income employees; financial information 1. introduction over the last few decades, american corporations have transitioned from a defined benefit environment to a defined contribution environment, making employees largely responsible for their own financial affairs (garman and kim, 2003; gonyea, 2007; krajnak, burns, and natchek, 2008). in a defined benefit pension plan, employers are responsible for providing retirement income for the employee and, therefore, employers bear the investment risk * corresponding author. tel.: !1-404-880-6413; fax: !1-404-880-6276. e-mail address: crhudson@cau.edu (c.r. hudson) financial services review 23 (2014) 25–43 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. associated with their retirement portfolio (olsen and vanderhei, 1997). conversely, a defined contribution plan shifts the majority of this responsibility and risk associated with retirement savings from the employer to the employee (garman and kim, 2003; gonyea, 2007; olsen and vanderhei, 1997). subsequently, employees are now responsible for their retirement planning and have to decide how much to save, when to save, and how to invest their funds (gonyea, 2007). however, many employees are not prepared for this responsibility, and employers have provided little education to assist employees (garman and kim, 2003). specifically, low-income employees have been the most adversely affected by this transition and face the greatest risk of being unprepared for retirement (kijakazi, 2003; munnell, golub-sass, perun, and webb, 2007). as a result, low-income employees may enter retirement with little or no savings because they typically have difficulties saving for retirement (gonyea, 2007; kijakazi, 2003). further, research finds that only 23% of households in the bottom third of the income distribution participate in employer retirement plans, compared to 66% of households in the top third (munnell et al., 2007). while low-income employees may receive a larger proportionate social security benefit based on their average indexed monthly earnings, still the absence of liquid savings in retirement presents challenges (gonyea, 2007). compounding this dire situation, low-income employees also lack the financial knowledge and skills needed to make sound financial decisions to achieve their financial goals (rand, 2004). for this reason, low-income employees often live from paycheck to paycheck, carry high-cost debt, and are unaware of consumer rights and services that could improve their financial circumstances (rand, 2004). oftentimes, these individuals fall prey to financial scams and predatory lenders partly because they lack the financial savvy to protect themselves (lyons and scherpf, 2004). thus, the combined challenge of low incomes and poor financial behaviors warrants a focus on low-income employees and prompts the following research question: could information from formal advisors positively affect the financial behaviors of low-income employees? the purpose of this study is to examine the relationship between low-income employees’ use of information from formal advisors and their financial behaviors. formal advisors include financial planners, bankers, brokers, accountants, insurance agents, lawyers, and employers. researchers assume that financial information from any one of these formal advisors, specific to a client’s needs, could have a positive impact on financial decisions and on the financial behaviors of low-income employees. the study’s research hypothesis is as follows: h11: a significant and positive relationship exists between the use of financial information from formal advisors and the acceptable savings and acceptable cash-flow management behaviors of low-income employees. h10: a significant and positive relationship does not exist between the use of financial information from formal advisors and the acceptable savings and acceptable cash-flow management behaviors of low-income employees. 26 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 this study is important because it adds to the body of research relating to the financial behaviors of low-income employees as well as the body of research relating to the effectiveness of financial advice or financial information provided by formal advisors. lowincome employees have not participated in the stock market in a significant way, but given improved financial behaviors and skills this segment has the potential to become active, influential investors (roojl, lusardi, and alessie, 2011). this study is also important because researchers take a slightly different approach by defining employees’ income not only by household income but also by household size. this approach should provide a better analysis of an employee’s true financial circumstances and budget constraints. moreover, this study utilizes multiple behavioral indicators to assess an employee’s savings and cash-flow management behaviors. 2. the workplace as an information source the workplace has proven to be an effective source of financial information for all employees (kranjak et al., 2008). after all, individuals spend a large portion of their time at work (csikszentmihalyi and lefevre, 1989) and employers and their financial representatives often connect with employees by offering financial education seminars and retirement planning programs through the workplace (servon and kaestner, 2008). these financial representatives are third party providers typically hired by the employer to provide workplace financial education seminars (clark, d’ambrosio, mcdermed, and sawant, 2003). another reason the workplace serves as an effective source of financial education is because employees often view their employers as authorities or experts, particularly on matters relating to their benefits (jenkins, 2005; krajnak et al., 2008). thus, many employees respect and value the workplace as a source of information (krajnak et al., 2008). moreover, when financial education is provided as an employee benefit, it is free of charge and therefore more accessible to those with lower incomes (edmiston and gillett-fisher, 2006). 3. literature review an individual’s financial behaviors provide insight into their financial literacy (huston, 2010; remund, 2010). according to huston (2010) and remund (2010), financial literacy is the ability to understand and comprehend personal finance information, or financial knowledge, and the ability to apply financial knowledge typically through financial behaviors. for this reason, researchers of this study choose to examine the financial literacy of low-income employees through their financial behaviors. furthermore, the researchers of this study review previous studies that examine the effect of financial information or financial education from formal advisors on financial behaviors of low-income employees (loibl and hira, 2005; rand, 2004) and on the financial behaviors of all employees (bayer, bernheim, and scholtz, 1996; bernheim and garrett, 2003; byrne, 2007; clark, d’ambrosio, mcdermed, and sawant, 2006; dolvin and templeton, 2006). 27c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 the theoretical model that guides the analysis and hypothesis of this study is the “behavioral model of financial services use” (kunovskaya, 2010) an adaptation of the “behavioral model of health services use” (andersen, 1995). this model suggests that after an individual makes the decision to use the services of financial professionals their financial status and/or financial behaviors will be positively affected (kunovskaya, 2010). likewise, this study suggests that financial information from formal advisors can positively affect the financial behaviors of low-income employees. 3.1. impact of financial advisors on low-income employees’ financial behaviors although the primary focus of most workplace financial education studies has been all employees within the workforce, a few studies specifically target low-income employees. these studies provide evidence of a positive and significant relationship between the use of information from formal advisors and acceptable financial behaviors among low-income employees (loibl and hira, 2005; rand, 2004). these previous studies examine the relationship between financial information that has been provided through self-directed learning (loibl and hira, 2005) and the financial behaviors of low-income employees as well as financial information provided through seminars (rand, 2004) and the financial behaviors of low-income employees. loibl and hira (2005) consider the effect of self-directed learning on financial management practices (i.e., financial behaviors), the effect of financial management practices on financial satisfaction, and the effect of financial satisfaction on career satisfaction. loibl and hira (2005) measure self-directed learning by examining the employee’s use of four different sources of financial planning information provided in the workplace. furthermore, researchers measure financial management practices or financial behaviors through surveys that ask whether participants save for goals, evaluate their spending, and make plans about how to use their money (loibl and hira, 2005). subsequently, researchers find that financial planning material that has been provided through self-directed learning has a significant and positive effect on low-income employees’ financial behaviors, which in turn has a significant and positive effect on financial satisfaction, which ultimately has a significant and positive effect on career satisfaction (loibl and hira, 2005). next, rand (2004) examines (1) the relationship between financial information seminars and the financial knowledge of low-income workers and (2) the relationship between financial information seminars and low-income workers’ savings behaviors. in rand’s (2004) study, low-income workers are defined as employees with household income less than 200% of the u.s. poverty level. rand (2004) invites 822 low-income workers to participate in a financial information seminar, as well as complete a pre-post financial knowledge exam. additionally, this study asks participants to save a percentage of their earned income toward a specific goal and agrees to match each $1 participants save with $2. rand (2004) finds that financial information seminars improve low-income workers’ financial knowledge as well as improve their savings behavior. 28 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 3.2. impact of financial information on all employees’ financial behaviors bayer et al. (1996) conduct one of the early workplace financial education studies that examine the effects of financial education on the financial behaviors of all employees within the workplace. researchers use a kpmg peat marwick retirement benefit survey and focus on the employer’s perspective by surveying employers about their company’s retirement plans and whether they provide workplace financial education sessions or financial information (bayer et al., 1996). in a similar study, bernheim and garrett (2003) likewise use data from the kpmg survey but take a different approach by focusing on the employees’ perspective. bernheim and garrett (2003) survey employees about their exposure to workplace financial seminars and financial information and about their retirement and personal savings. both studies find evidence that workplace financial education have a positive effect on the financial behaviors of all employees. bayer et al. (1996) find that employers who offer some form of financial education or information have higher retirement plan participation and contribution rates than those employers who do not. similarly, bernheim and garrett (2003) find that employees that have been exposed to financial education seminars or financial information in the workplace have higher savings accumulation in their retirement plans and higher savings overall. in another previous study whose focus is all employees, clark et al. (2006) conduct an experiment using university employees to examine changes in retirement goals and savings behaviors. these university employees participate in a one-hour seminar on retirement planning and goal setting and afterwards researchers collect data before the seminar, after the seminar, and several months following the seminar (clark et al., 2006). similarly, dolvin and templeton (2006) conduct a clinical study, within a law firm, with the purpose of examining the relationship between participation in retirement planning seminars and employees’ asset allocation decisions. in this study, the law firm offers a 90-minute financial education seminar to its employees and these employees are surveyed about their participation in the seminar and about information related to their portfolios (dolvin and templeton 2006). both studies find evidence that financial information from formal advisors has a positive effect on the employee’s financial behaviors. clark et al. (2006) find that financial information prompts a significant number of participants to change their retirement goals, modify their savings contributions, and reallocate their investment funds. while, dolvin and templeton (2006) find that those who participate in financial education seminars possess more efficient portfolios in their retirement funds than those who do not. 4. methodology 4.1. data and sample this study uses data from the 2010 survey of consumer finances (scf). the scf provides information about u.s. household finances and contains detailed information from household balance sheets and income statements (kennickell, 2007). moreover, the scf 29c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 includes demographic and socioeconomic information as well as information relating to financial behaviors and financial attitudes (kennickell, 2007). the employee subsample within this study is created by segmenting data within the 2010 scf by employment status. among the 6,482 households, 4,280 or 66% are employees. this employee sample includes individuals who reports working at the time of the survey, including those who are working temporarily or seasonally and those who are self-employed. this larger employee sample is further segmented to create three subsamples based on income and household size. low-income employees are defined as employees with household incomes less than or equal to 200% of the 2010 u.s. poverty level. middle-income employees are defined as employees with household incomes greater than 200% and less than and equal to 494% of the 2010 u.s. poverty level. finally, high-income employees are defined as employees with household incomes greater than 494% of the 2010 u.s. poverty level (u.s. department of health and human services, 2010). the employee subsamples’ income levels by household sizes are listed in table 1. 4.2. financial behaviors researchers find that financial information increases financial knowledge, which in turn improves financial behaviors (hogarth, beverly, and hilgert, 2003). for this reason, researchers of this study choose to examine the relationship between the use of financial information from formal advisors and the financial behaviors of low-income employees. thus, the key variables for this study are specific financial behaviors which are operationalized and measured using data from select financial behavior questions within the 2010 scf. researchers assume that any one of these 10 financial behaviors is equally likely to occur and that income does not preclude the practice of any of these behaviors. all variables are binary and all “yes” responses are considered good or acceptable financial behaviors (coded as 1) and all no responses are considered bad or unacceptable financial behaviors (coded as 0). the 10 financial behavior questions are grouped into measures of cash-flow management table 1 income segmentation of the three income subsamples household size 2010 u.s. poverty levela low-income medium-income high-income " or equal to greater than " or equal to greater than 1 $10,800 $21,660 $21,660 $ 53,500 $ 53,500 2 14,570 29,140 29,140 71,976 71,976 3 18,310 36,620 36,620 90,451 90,451 4 22,050 44,100 44,100 108,927 108,927 5 25,790 51,580 51,580 127,403 127,403 6 29,530 57,060 57,060 145,878 145,878 7 33,270 66,540 66,540 164,353 164,353 8 37,010 74,020 74,020 182,829 182,829 note. au.s. 2010 poverty level. adapted from “the 2010 human health services poverty guidelines” by the u.s. department of health and human services (2010) (available at http://aspe.hhs.gov/poverty/10poverty. shtml). 30 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 and measures of savings. in the cash-flow management category, five financial behavior variables represent employee’s cash-flow management behaviors. examples of these variables include, having a checking account (checking), and whether loans are paid on schedule or ahead of schedule (loans on time). conversely, in the savings category, five financial behavior variables represent employees’ savings behaviors. two of these savings variables are, whether employees have a savings account (savings) and whether employees save on a monthly basis (save). next, financial behavior indices are created to rank an employee’s savings behaviors (savings index) and cash-flow management behaviors (cash-flow index) as low, medium, or high. a high ranking represents a response of “1” to 70% or more of the financial behavior variables. a medium ranking represents a response of “1” to more than 25% but less than 70% of the financial behavior variables. a low ranking represents a response of “1” to 25% or less of the financial behavior variables. all financial behavior variables and indexes are listed in table 2. 4.3. information source variables literature suggests that low-income employees obtain their financial information from informal advisors or informal sources that can include family, friends, spouses, or themselves (olsen and whitman, 2007). however, low-income employees have expressed a desire to receive information from formal advisors as long as that information is affordable (garman and kim, 2003). an information source variable is used in this study to assess employees’ source of financial information based on their responses to the following questions: “what sources of information do you use to make decisions about savings and investments?” the formal advisor category consists of responses that include information from a financial planner, a broker, a banker, an accountant, an insurance agent, a lawyer, or material from work or business contact. the informal advisor category consists of responses that include information from friends and family, self, personal experience, other personal research, as well as “do not shop around,” and “do not save or invest.” the 2010 scf question includes two information source options which are difficult to interpret. the “do not shop around” response is coded as an informal advisor because individuals are relying solely on “themselves,” which is also coded as an informal advisor. individuals who indicated they “do not save or invest” are also grouped with the informal advisor category because they did not indicate any information source and implicitly do not seek out information regarding saving or investing, but again rely on “themselves.” the public source category consists of responses that include calling around, magazines/newspapers, material in the mail, online/ internet, advertising and tv (olsen and whitman, 2007). table 2 contains the information source variable and their measurements. 4.4. other variables researchers find evidence of a significant relationship between other variables and good, or acceptable, financial behaviors. therefore, these variables are identified and controlled for in an attempt to examine the relationship between the use of financial information from 31c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 formal advisors and the financial behaviors of low-income employees. one relevant variable for this study is a planning variable (planning). research shows that planners are more financially well off than non-planners (lusardi and mitchell, 2005). furthermore, those who have the propensity to plan tend to display positive financial behaviors (deaves, veit, table 2 list of variables measurements cash-flow management behaviors checking # 1 if reported having a checking account; 0 otherwise loans on time # 1 if reported paying loans ahead of time or on time; 0 otherwise spending # 1 if reported spending was less than or equal to income; 0 otherwise auto deposit # 1 if reported money automatically deposited; 0 otherwise software # 1 if reported using software to manage their money; 0 otherwise cash-flow index # % of acceptable cash-flow management behaviors savings behaviors savings # 1 if reported having a savings account; 0 otherwise save # 1 if reported saving on a monthly basis; 0 otherwise cd # 1 if reported having certificates of deposit; 0 otherwise retire plan # 1 if reported participation in any pension, retirement, or tax-deferred savings plan connected with job; 0 otherwise ira # 1 if reported having money in ira or keogh; 0 otherwise savings index # % of acceptable savings behaviors information source variables formal advisor # 1 if information sources were a financial planner, banker, broker, accountant, insurance agent, lawyer, or material from work/business contact; 0 otherwise informal advisor # 1 if information sources were friends and family, self, personal experience, other personal research, don’t shop around, or do not save or invest; 0 otherwise public source # 1 if information sources were calling around, magazine/ newspaper, material in the mail, online/internet, advertising, or tv; 0 otherwise other variables planning # 1 if plan for next few years, next 5–10 years, or more than 10 years; 0 if plan for next year or next few months age # chronological age of respondent male # 1 if respondent is male; 0 if female education # years of education completed minority # 1 if race is black, hispanic, or other minorities; 0 if race is white self-employed # 1 if self-employed; 0 otherwise financial behavior indices ranking low # response of 1 to 25% or less acceptable behaviors medium # response of 1 between 25% to 70% acceptable behaviors high # response of 1 to 70% or more acceptable behaviors 32 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 bhandari, and cheney, 2007). in this study, the planning variable (planning) indicates to what extent individuals plan for their futures, with longer term planners being coded as “1” and those who plan for a year or less being coded as “0.” this study also includes demographic variables such as age, gender, education, and race because of their correlation with financial behaviors. table 2 lists all of the other variables and their measurements. 4.5. statistical analysis in this study, data from the 2010 scf are analyzed with the sas statistical package (version 9.3). the data are used to investigate the hypothesis (h1) that examines the relationship between the use of financial information from formal advisors and low-income employees’ percentage of acceptable savings behaviors and acceptable cash-flow management behaviors. these percentages rank as low, medium, or high. the first data analysis step generates descriptive statistics to describe the demographic and socioeconomic characteristics of the low-income, middle-income, and high-income employee subsamples. descriptive statistics are weighted to accurately reflect the representation of smaller groups in the population. the next data analysis step consists of ordered logistic regression analyses, which is used to test the significance of the relationships between the independent variables and the dependent variable. two ordered logistic regression models are used to test the relationships between information from formal advisors and the dependent variables (savings index and cash-flow index), while controlling for other variables. the ordered logistic regression model uses the probability odds models to determine the probability that the dependent variable (y) will fall into one category versus another, in particular the higher categories, given the independent variables (snedker, glynn, and wang, 2002). the probability odds models are as follows: logit$p1% ! log p1 1 " p1 ! &1 # bx1 (1) logit$p1 # p2% ! log p1 # p2 1 " p1 " p2 ! &1 # b'x1 (2) logit$p1 # p2 # . . . pk% ! log p1 # p2 # . . . pk 1 " p1 " p2 " . . . pk ! &1 # b'x1 (3) the ordered logistic regression model utilizes the $2 test to determine if an independent variable has a significant correlation with the dependent variable. coefficients which are generated from the ordered logistic regression model for each independent variable can be estimated by maximum likelihood. additionally, an odds ratio estimate is generated for each independent variable (snedker et al., 2002). the 2010 scf contains five implicates, or five duplicate sets of data that uses various methods to account for missing data. to avoid analysis errors related to these five implicates, researchers use the rii technique in this study. two ordered logistic regression models are estimated using the low-income employee sub-sample to determine significant factors affecting the savings behavior and cash-flow 33c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 management behaviors of low-income employees. these two ordered logistic regression models are as follows: savings index # b0 ! b1 formal advisor ! b2 planning ! b3 minority ! b4 self-employed ! b5 education ! b6 male ! b7 age ! e (4) cash-flow index # b0! b1 formal advisor ! b2 planning ! b3 minority ! b4 self-employed ! b5 education ! b6 male ! b7 age ! e (5) for comparison purposes, two ordered logistic regression models also are estimated using data for the middle-income and high-income employee subsamples to assess the relationship between their dependent and independent variables. 5. results 5.1. socioeconomic characteristics of the three income subsamples this section presents socioeconomic characteristics and descriptive results for the three employee income subsamples: (1) low-income employees, (2) middle-income employees, and (3) high-income employees. the demographic and socioeconomic characteristics and descriptive results are reported in tables 3 and 4. the employee sample for this current study is created by segmenting data within the 2010 scf by employment status. this larger employee sample is further segmented to create three income subsamples: (1) the focal group, low-income employees, (2) middle-income employees, and (3) high-income employees, which is based on household income and household size. the focal group contains 1,060 low-income employees and the sample contains 1,432 middle-income employees and 1,787 high-income employees. the majority of respondents who are low-income employees live in households with one to three people (59%), including themselves, whereas the smallest percentage of low-income employees (3%) lives in households with seven to ten people (table 3). the mean household size of low-income employees is 3.2 people, which is significantly larger than the mean household size of middle-income employees that is 2.8 (t (2,492) # 6.12, p " 0.0001) and even larger than the mean household size of high-income employees that is 2.7(t (2,847) # 9.41, p " 0.0001) (table 4). however, like low-income employees, the majority of middleincome employees (67%) and high-income employees (76%) live in households with one to three people, including themselves (table 3). as for race, the majority of low-income employees are white (54%). however, minorities are predominant among low-income employees (46%) in comparison to the percentage of middle-income minority employees (30%) and high-income minority employees (18%). 34 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 hispanics (24%) represent the largest percentage of minorities in the low-income employee subsample. blacks represent the largest percentage of minorities in the middle-income subsample (13%). moreover, the largest percentage of minorities in the high-income subsample is other minorities (7%) (table 3). at the time of the survey, the mean age of low-income employees is 42. this is significantly younger than the mean age of middleincome employees that is 44 (t (2,492) # (3.36, p " 0.0008) and even younger than the mean age of high-income employees, which is 51 (t (2,847) # (17.51, p " 0.0001). the mean household income of low-income employees is $23,078, whereas the mean household table 3 socioeconomic characteristics (percentage of total) characteristics percentage of low-income employees (n # 1,060) percentage of middle-income employees (n # 1,432) percentage of high-income employees (n # 1,787) household size 1–3 59% 67% 76% 4–6 38 32 24 7–10 3 1 0 gender male 65 80 89 female 35 20 11 race white 54 70 82 black 18 13 6 hispanic 24 12 5 other 4 5 7 marital status married 45 58 77 single 55 42 23 education less than high school 21 7 2 graduated high school 39 31 15 some college 24 29 19 college education 11 22 33 graduate degree 5 11 31 age 18–30 24 15 9 31–50 49 54 49 51! 27 31 42 table 4 socioeconomic characteristics (mean and sd) characteristics low-income employees middle-income employees high-income employees mean (sd) mean (sd) mean (sd) household size 3.2 1.6 2.8 1.5 2.7 1.3 income 23,078 11,673 58,754 23,205 184,764 13,282 age 42 14 44 11.7 51 11.8 education (years completed) 12.3 2.9 13.7 2.3 15.5 1.9 35c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 income of middle-income employees is $58,754. finally, the mean household income of high-income employees is $184,764 (table 4). on average, low-income employees complete 12.3 years of education (table 4), with the largest percentage of low-income employees (39%) graduating from high school and the next largest percentage (24%) graduating from high school and completing some college. the largest percentage of middle-income employees likewise graduates from high school (31%), followed by the next largest percentage of middle-income employees who not only graduate from high school but also completes some years of college (29%). finally, the largest percentage of high-income employees obtains a bachelor’s degree (33%), closely followed by those who obtain a graduate degree (31%) (table 3). 5.2. ordered logistic regression results on cash-flow management behaviors the research question asks whether information from formal advisors could have a positive effect on financial behaviors: (1) the cash-flow management behaviors and (2) the savings behaviors of low-income employees. results from the first ordered logistic regression analysis indicate a significant and positive relationship between the use of financial information from formal advisors and the proportion of acceptable cash-flow management behaviors of low-income employees. thus, the cash-flow management behaviors of lowincome employees who use financial information from formal advisors are 1.51 times more likely to rank high or medium, rather than low, than the behaviors of low-income employees who use financial information from informal advisors or public sources. this relationship is not significant for middle-income or high-income employees (table 5). these results confirm the cash-flow management behaviors proportion of the hypothesis (h1) and are reported in table 5. the statistics from this ordered logistic regression model indicate that the model convergence criterion is satisfied and the model is determined to be a good fit for this study, according to the akaike information criterion (aic), the schwarz criterion (sc), and the (2 log l test. additionally, the score test for proportional odds assumption concludes that the ordered logistic coefficients are equal across the three possible outcomes of the dependent variable. 5.3. ordered logistic regression results on other variables (cash-flow management behaviors) as discussed in the methodology section, other variables, specifically planning, gender, race, education, age, and employment status are controlled for and included as independent variables in the ordered logistic regression models. a significant and positive relationship exists between planning and the proportion of acceptable cash-flow management behaviors of low-income, middle-income, and high-income employees. thus, the cash-flow management behaviors of low-income employees who are planners are 1.40 times more likely than those of non-planners to rank high or medium, rather than low. furthermore, the cash-flow management behaviors of middle-income and high-income employees who are planners are 1.29 and 1.75 times more likely than the behaviors of non-planners to rank high or medium, rather than low (table 5). 36 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 on the other hand, results indicate a significant and negative relationship between race and the proportion of acceptable cash-flow management behaviors of low-income employees. in the model, race is represented by a binary variable, minority. thus, the cash-flow management behaviors of low-income minority employees are more likely than those of low-income white employees to rank low, rather than high or medium. this relationship also is significant and negative for middle-income minority employees and high-income minority employees. the relationship between gender, which is represented by a binary variable, male, and the proportion of acceptable cash-flow management behaviors is not significant for low-income employees but is negative and significant for middle-income and highincome employees (table 5). furthermore, results indicate a significant and positive relationship between education and the proportion of acceptable cash-flow management behaviors of low-income, middleincome, and high-income employees. specifically, for each additional year of education, the odds of the cash-flow management behaviors of low-income, middle-income, and highincome employees being ranked as high or medium, rather than low, increases by multiples of 1.18, 1.23, and 1.16, respectively (table 5). moreover, a significant and positive relationship between age and the proportion of acceptable cash-flow management behaviors is found for low-income employees but not for middle-income employees. thus, for each additional year of age, the odds of low-income employees’ cash-flow management behaviors table 5 likelihood of high or medium cash-flow management behaviors variables low-income employees (n # 1,060) middle-income employees (n # 1,432) high-income employees (n # 1,787) coefficients odds ratio coefficients odds ratio coefficients odds ratio formal advisor 0.415* 1.51 0.138 1.15 0.018 1.02 se (0.135) (0.114) (0.111) planning 0.337* 1.40 0.251* 1.29 0.559*** 1.75 se (0.124) (0.110) (0.138) minority (0.738*** 0.48 (0.343* 0.71 (0.456* 0.63 se (0.136) (0.118) (0.160) self-employed 0.019 1.02 (0.214 0.81 (0.515*** 0.60 se (0.162) (0.158) (0.118) education 0.165*** 1.18 0.205*** 1.23 0.147*** 1.16 se (0.023) (0.025) (0.028) male 0.204 1.23 (0.303* 0.74 (0.409* 0.66 se (0.131) (0.136) (0.200) age 0.013* 1.01 0.002 1.00 (0.018*** 0.982 se (0.005) (0.005) (0.005) intercept (4.012 (2.51 (0.136 se (0.415) (0.422) (0.520) intercept (1.083 0.76 5.425 se (0.402) (0.426) (0.780) note. likelihood ratio test, p " .0001; score test, p " .0001; wald test, p " .0001; score test for proportional odds assumption, p ) .05. *p " .05, **p " .01, ***p " .001. 37c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 being ranked high or medium, rather than low, increases by a multiple of 1.01. the relationship is significant and negative for high-income employees (table 5). finally, the relationship between self-employed and the proportion of acceptable cash-flow management behaviors is not significant for low-income or middle-income employees but is significant and negative for high-income employees (table 5). 5.4. ordered logistic regression results on savings behaviors results of the second ordered logistic regression model, which are reported in table 6, indicates a significant and positive relationship between the use of financial information from formal advisors and the proportion of acceptable savings behaviors of low-income employees. this result confirms the savings behavior proportion of the hypothesis (h1). thus, the savings behaviors of low-income employees who use financial information from formal advisors are 1.65 times more likely to rank as high or medium, rather than low, than the savings behaviors of low-income employees who use financial information from informal advisors or public sources (table 6). likewise, the savings behaviors of middle-income employees who use financial information from formal advisors are 1.38 times more likely to rank as high or medium, rather than low, than the savings behaviors of middle-income table 6 likelihood of high or medium savings behaviors low-income employees (n # 1,060) middle-income employees (n # 1,432) high-income employees (n # 1,787) variables coefficients odds ratio coefficients odds ratio coefficients odds ratio formal advisor 0.499*** 1.65 0.319* 1.38 (0.025 0.98 se (0.139) (0.116) (0.100) planning 0.169 1.18 0.504*** 1.66 0.721*** 2.06 se (0.129) (0.112) (0.135) minority (0.483*** 0.62 (0.439*** 0.64 (0.664*** 0.51 se (0.137) (0.119) (0.153) self-employed (0.171 0.84 (0.789*** 0.45 (0.484*** 0.62 se (0.177) (0.160) (0.107) education 0.161*** 1.17 0.218*** 1.24 0.175*** 1.19 se (0.026) (0.025) (0.027) male (0.007 0.99 (0.065 0.94 0.322 1.38 se (0.137) (0.138) (0.186) age 0.021*** 1.02 0.016* 1.02 0.008 1.01 se (0.005) (0.005) (0.004) intercept (6.56 (5.792 (4.545 se (0.489) (0.455) (0.511) intercept (3.02 (2.559 1.028 se (0.438) (0.428) (0.496) note. likelihood ratio test, p " .0001; score test, p " .0001; wald test, p " .0001; score test for proportional odds assumption, p " .05. *p " .05, **p " .01, ***p " .001. 38 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 employees who use financial information from informal advisors and public sources. this relationship is not significant for high-income employees (table 6). the statistics from this ordered logistic regression model indicates that the model’s convergence criterion is satisfied and the model is determined to be a good fit for this study, according to the aic, the sc, and the (2 log l test. the score test for proportional odds assumption concludes that the ordered logistic coefficients may not be equal across the three possible outcomes of the dependent variable (low, medium, and high). this score test result could be because of small sample sizes within one of the three possible outcomes, which can cause the model to fail this test (sas knowledge base, 2012). separate binary models, or models that uses binary dependent variables, are ran to determine whether the results of the ordered logistic model appeared to be biased. the findings from the binary models are consistent with the findings of the ordered logistic models; therefore, the findings from the ordered logistic model are interpreted and used in this study. 5.5. ordered logistic regression results on other variables (savings behaviors) a significant and positive relationship exists between planning and the proportion of acceptable savings behavior of middle-income and high-income employees, but not for the savings behaviors of low-income employees. thus, results indicate that the savings behaviors of middle-income and high-income employees who are planners is 1.66 and 2.06 times more likely than those of non-planners to rank as high or medium rather than low, respectively (table 6). on the contrary, the results indicate a significant and negative relationship between race and the proportion of acceptable savings behaviors of low-income employees. thus, the savings behaviors of low-income minority employees are more likely to rank as low, rather than high or medium, relative to those of low-income white employees (table 6). this relationship is also significant and negative for middle-income and high-income minority employees. the relationship between gender and acceptable savings behavior is not significant for any income subsample (table 6). furthermore, a significant and positive relationship exists between education and the proportion of acceptable savings behaviors of all three income subsamples. specifically, for each additional year of education, the odds of the savings behaviors of low-income employees being ranked as high or medium, rather than low, increases by a multiple of 1.17. likewise, for each additional year of education, the odds of the savings behaviors of middle-income and high-income employees being ranked as high or medium, rather than low, increases by a multiple of 1.24 and 1.19, respectively (table 6). finally, a significant and positive relationship exists between age and the proportion of acceptable savings behaviors of low-income and middle-income employees. thus, for each additional year of age, the odds of the savings behaviors of low-income and middle-income employees being ranked as high or medium, rather than low, increases by a multiple of 1.02, for both groups. this relationship is not significant for high-income employees. the relationship between self-employed and acceptable savings behaviors is not significant for low-income employees but is negative and significant for middle-income and high-income employees (table 6). 39c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 6. discussion 6.1. impact of financial information on low-income employees’ financial behaviors this study finds a significant and positive relationship between the use of information from formal advisors and the proportion of acceptable financial behaviors of low-income employees. although only correlational, this finding suggests that financial information from formal advisors would positively affect the financial behaviors of low-income employees. this is consistent with results from previous studies in which low-income employees are the focal group (rand, 2004; loibl and hira, 2005). loibl and hira (2005) find that financial planning information provided through self-directed learning have a significant effect on low-income employees’ financial behaviors. likewise, rand (2004) finds that financial information seminars improve a low-income worker’s financial knowledge as well as their savings behaviors. although the primary focus of this current study is low-income employees, there are specific findings that relates to middle-income and high-income employees that are worth noting. this study finds that financial information from formal advisors does not have a significant effect on the positive cash-flow management behaviors of middle-income and high-income employees. it is not to say that middle-income and high-income employees do not use information from formal advisors, it is just that information from formal advisors has no significant effect on their positive financial behaviors. in fact, research has shown that high-income employees, in particular, use information from formal advisors more so than informal advisors (olsen and whitman, 2007). additionally, this study finds that the savings behaviors of high-income employees who use information from formal advisors are more likely to be ranked low as opposed to medium or high. however, this finding is not significant and is on the borderline of the low and medium ranking. results from this study are also consistent with previous workplace financial education and financial information studies that focus on all employees in the workforce and not just low-income employees. these previous studies find that information from formal advisors have a positive effect on the financial behaviors of all employees (bayer et al., 1996; bernheim and garrett, 2003; byrne, 2007; clark et al., 2006; dolvin and templeton, 2006). one difference between this study and previous studies by bayer et al. (1996) as well as bernheim and garrett (2003) is that these previous studies focus on the access of financial information rather than the use of financial information from formal advisors. additionally, the focal group of this study is low-income employees whereas these previous studies did not focus on any particular segment within the workplace. nevertheless, the findings of this current study are still consistent with the finding of these previous studies. 6.2. implications of the current study three distinct implications are evident because of this study: (1) formal advisors such as financial planners, employers, brokers, and bankers should target their financial advice and financial education seminars towards low-income employees; (2) employers should make formal advisors available to their employees for financial consultations; and (3) more 40 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 research that focuses on low-income employees or specific segments of the workforce is warranted. based on the findings of this study, it is obvious that the financial advice or financial information that formal advisors provide has a positive impact on the financial behaviors of low-income employees. in fact, based on these findings, that effect would be greater for low-income employees as oppose to middle-income and high-income employees. therefore, if formal advisors target financial information or advice towards low-income employees, there would be a greater probability of improving low-income employees’ financial behaviors and long-term financial status. ultimately, employees would make better financial decisions and experience less financial stress (garman, 1999). as a result, employers would foster less financially stressed, more productive employees (garman, 1999). in addition to tailoring financial information towards low-income employees, employers should use third party financial professionals to administer this financial education or financial information. a likely choice would be the financial professionals who administer the company’s retirement plans. the findings of this study suggest that if third party financial professionals are used, this would be beneficial for low-income employees and employers. in fact, employers would receive a return on this investment in their employees by having more attentive, more productive employees (garman, 1999). finally, low-income employees could become more active investors in the stock market, through their easily accessible retirement plans, thus presenting a major opportunity for the financial planning industry. as previously mentioned, low-income employees do not significantly participate in the stock market (rooij et al., 2011). however, this trend could change for this segment of the population, and more research is needed to understand this segment. specially, research that explores the financial characteristics, financial behaviors and financial decision-making of low-income employees is necessary to design financial advice, education, and financial information that is effective. 7. conclusions these researchers find (1) a significant and positive relationship between the use of information from formal advisors and the acceptable savings behaviors of low-income employees and (2) a significant and positive relationship between the use of information from formal advisors and the cash-flow management behaviors of low-income employees. thus, these findings suggest that low-income employees who use information from formal advisors have better savings behaviors and better cash-flow management behaviors than those who do not. moreover, these improved financial behaviors could result in better financial decisions for low-income. as a result, low-income employees could become more financially literate and more active participants in the stock market. the secondary findings of this study are a significant and positive relationship between those low-income employees who are planners, have more education, and are older and their acceptable cash-flow management behaviors. furthermore, a significant and positive relationship is found between education and age and the acceptable savings behaviors of low-income employees. thus, the financial behaviors of low-income employees who are 41c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 planners, more educated and older are better than the financial behaviors of non-planners, less educated and younger low-income employees. in contrast, a significant and negative relationship is found between race and the acceptable savings and acceptable cash-flow management behaviors of low-income employees. thus, these low-income employees who are minorities have less positive or poorer savings and poorer cash-flow management behaviors than their comparable low-income white counterparts. references andersen, r. (1995). revisiting the behavioral model and access to medical care. journal of health and social behavior, 36, 1–10. bayer, p. j., bernheim, d., & scholz, j. k. (1996). the effects of financial education in the workplace: evidence from a survey of employers. (nber working paper no. 5655). cambridge, ma: national bureau of economic research. bernheim, d., & garrett, d. m. (2003). the effects of financial education in the workplace: evidence from a survey of households. journal of public economics, 87, 1487–1519. byrne, a. (2007). employee saving and investment decisions in defined contribution pension plans; survey evidence from the uk. financial services review, 16, 19–40. clark, l. r., d’ambrosio, b. m., mcdermed, a. a., & sawant, k. (2003). financial education and retirement savings. symposium presentation at the sustainable community development: what works, what doesn’t, and why conference sponsored by the federal research system, march 2003 (available at http://papers. ssrn.com/sol3/papers.cfm?abstract_id#390642). clark, l. r., d’ambrosio, b. m., mcdermed, a. a., & sawant, k. (2006). retirement plans and savings decisions: the role of information and education. journal of pension education and finance, 5, 45–67. csikszentmihalyi, m., & lefevre, j. (1989). optimal experience in work and leisure. journal of personality and social psychology, 56, 815–822. deaves, r., veit, e., bhandari, g., & cheney, j. (2007). the savings and investment decisions of planners: a cross-sectional study of college employees. financial services review, 16, 117–133. dolvin, s., & templeton, w. (2006). financial education and asset allocation. financial services review, 15, 133–149. edmiston, k., & gillett-fisher, m. (2006). financial education at the workplace: evidence from a survey of federal reserve bank employees (community affairs working paper no. 2006–02). kansas city, mo: federal reserve bank of kansas city. garman, t. (1999). employer-sponsored education programs and incentives to improve employees’ financial lifestyle. employment relations today, 25, 63–72. garman, t., & kim, j. (2003). financial education and advice changes in worker attitudes and behaviors. journal of compensation and benefits,1, 7–13. gonyea, j. g. (2007). improving the retirement prospects of low-wage workers in a defined contribution world. the journal of contemporary social services, 88, 453–462. hogarth, j. m., beverly, s. g., & hilgert, m. (2003). patterns of financial behaviors: implications for community educators and policy makers. paper presented at the 2003 federal reserve systems community affairs research conference, march 2003: washington, dc. (available at http://www.federalreserve.gov/ communityaffairs/national/ca_conf_suscommdev/pdf/hogarthjeanne.pdf). huston, s. (2010). measuring financial literacy. the journal of consumer affairs, 44, 296–316. jenkins, v. (2005). educating employers to make an informed choice about financial education. pensions, 10, 331–336. kennickell, a. b. (2007). look and listen; but don’t stop: interviewers and data quality on the 2007 scf. washington, dc: proceedings of the survey research methods section of the federal reserve (pp. 3646– 3656). (available at https://www.amstat.org/sections/srms/proceedings/y2007/files/jsm2007-000648.pdf). 42 c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 kijakazi, k. (2003). low-wage earners: options for improving their retirement income. in p. edelman, d. salisbury, & p. larson (eds.), the future of social insurance: incremental or fundamental reform (pp. 41–81). washington, dc: national academy of social insurance. krajnak, p. a., burns, s. a., & natchek, s. m. (2008). retirement education in the workplace. financial services review, 17, 131–141. kunovskaya, i. (2010). the impact of financial literacy on use of financial services in russia. ph.d. dissertation. university of georgia, athens, georgia. loibl, c., & hira, k. t. (2005). self-directed financial learning and financial satisfaction. financial counseling and planning, 16, 11–22. lusardi, a., & mitchell, o. (2005). financial literacy and planning: implications for retirement wellbeing (wp 2005–108). ann arbor, mi: university of michigan retirement research center. (available at http:// papers.ssrn.com/sol3/papers.cfm?abstract_id#881847). lyons, a., & scherpf, e. (2004). moving from unbanked to banked: evidence from the money smart program. financial services review, 13, 215–231. munnell, a. h., golub-sass, f., perun, p., & webb, a. (2007). households at risk: a closer look at the bottom third (center for retirement working papers). boston, ma: the center for retirement research, boston college. olsen, a., & whitman, k. (2007). effective retirement savings programs: design features and financial education. social security bulletin, 67, 53–72. olsen, k., & vanderhei, j. (1997). defined contribution plans dominance grows across sectors and employer sizes, while mega defined benefit plans remain strong: where we are and where we are going (special report sr-33, ebri issue brief number 190). washington, dc: employee benefit research institution. retrieved from http://www.ebri.org/pdf/briefspdf/1097lb.pdf rand, d. (2004). financial education and asset-building programs for welfare recipients and low-income workers: the illinois experience. washington, dc: the brookings institution center on urban and metropolitan policy. remund, d. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. rooij, m., lusardi, a., & alessie. r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. sas knowledge base. (2012). the proc logistic proportional odds test and how to fit a partial proportional odds model. (available at http://support.sas.com/kb/22/954.html). servon, l., & kaestner, r. (2008). consumer financial literacy and the impact of online banking on the financial behavior of low-income bank customers. the journal of consumer affairs, 42, 271–305. snedker, k., glynn p., & wang, c. (2002). ordered/ordinal logistic regression with sas and stata1. seattle, wa: the university of washington. (available at http://staff.washington.edu/glynn/olr.pdf). u.s. department of health and human services. (2010). the 2010 human health services poverty guidelines. (available at http://aspe.hhs.gov/poverty/10poverty.shtml). 43c.r. hudson, l. palmer / financial services review 23 (2014) 25–43 marital status, health, and retirement wealth for middle aged and older women serah shina,*, hyungsoo kimb adepartment of family sciences, university of kentucky, 315 funkhouser building, university of kentucky, lexington, ky 40506, usa bdepartment of family sciences, university of kentucky, 315 funkhouser building, university of kentucky, lexington, ky 40506, usa abstract we investigate the different impacts of health problems on wealth among unmarried and married women in their later years. we advance the literature by using a new health measure of a 22 year sequence of chronic diseases. using data from 2,476 women age 50 or older from the 1992–2014 health and retirement study, we find that unmarried women have more complicated and costly health sequences compared with married women, including multi-morbidity or comorbidity. over the 22 year period, retirement wealth for unmarried women with costly health sequences is reduced by approximately $3,600 to $5,400 annually. © 2017 academy of financial services. all rights reserved. jel classification: d12; d14; i14 keywords: heath sequences; chronic diseases; unmarried older women; retirement wealth 1. introduction maintaining health and financial security is challenging for near-retirees (laatsch & klein, 2010; yuh, hanna, & montalto, 1998). with limited financial resources in retirement, health problems threaten the financial security of many americans (smith, 1999, 2003). significant differences in health and financial security between married and unmarried women have been suggested (institute for american values, 2011; joyce, 2007; kim, 2006). * corresponding author. tel.: �1-859-539-7136; fax: �1-859-257-3212. e-mail address: serah.shin@uky.edu; serah.shin3@gmail.com (s. shin) financial services review 26 (2017) 255–270 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. of particular concern are the financial and health conditions of middle-aged and older unmarried women relative to their married peers. the financial status of unmarried older women is low and uncertain considering income (68%) and net worth (31%) compared with married women (joyce, 2007; kim, 2006). many unmarried women (36%) over age 65 rely solely on social security benefits for income, whereas only 21% of married women are as financially restricted (national women’s law center, 2015). recent studies have shown that one’s wealth becomes substantially lower when marital status changes from married to single in later years (addo & lichter, 2013; zissimopoulos, 2009). the health status of unmarried older women is also worse than their married peers: the self-rated health of unmarried women (65% of women with a median age of 48) worsen over time (1972–2003) relative to that of married women (liu & umberson, 2008), middle-aged individuals (51–61) who have spent more years divorced or widowed exhibit more chronic conditions and mobility limitations (hughes & waite, 2009), marital disruption leads to a higher prevalence of cardiovascular diseases (maselko, bates, avendano, & glymour, 2009), and higher mortality rates (henretta, 2010). these previous studies have examined differences in either health or financial status separately by marital status; however, studies have rarely investigated how health problems affect financial situations differently according to marital status in the later years. one study articulate different impacts of health events on wealth depletion between married and unmarried older women (kim, 2006). kim (2006) reports severe chronic conditions cause 4% to 10% less wealth accumulation for unmarried women than for married women. considering a 22 year period (1992–2014) from the health and retirement study, we investigate the differences between unmarried and married women in the potential impacts of health problems on wealth in the later years. we expand the previous study in two distinct ways: first, we use a measure of health status as a 22 year term trajectory of chronic health conditions rather than the current health status; second, we use a nationally representative sample of american women aged 50 and over, in contrast to kim (2006), who considers a population of age 70 and older. this period of life is important in that chronic health conditions tend to begin in one’s 50s, and the impacts of such conditions on household finance emerge over the following 20 to 30 years. the findings from this study may provide more comprehensive information on long-term relationships between middle-aged and older women’s health and retirement wealth according to marital status. 1.1. health measures and sequences of chronic conditions chronic diseases are important health measures related to financial security in the later years. the prevalence and incidence of these conditions have been used to measure health: the occurrence of these conditions (adams, hurd, mcfadden, merrill, & ribeiro, 2003), the number of chronic conditions (zajacova & burgard, 2013), and comorbidities (schoenberg, kim, edwards, & fleming, 2007). these health measures document health conditions primarily at a single point in time or over a short period of time; however, they are limited in the ability to track health status over 256 s. shin, h. kim / financial services review 26 (2017) 255–270 the course of a person’s life. to understand the long-term and cumulative impact of health on financial outcomes, we need to consider a sequence of chronic diseases over time. let us assume potential sequences of chronic disease for a 10 year period in the following example: one may have a sequence of no (n) chronic disease for 10 years (symbolized by nnnnnnnnnn); some may have a sequence that begins with high blood pressure (h) for the first 6 years, followed by arthritis (a) in the seventh year that is maintained until the 10th year (symbolized by hhhhhhaaaa). we measure and compare such health sequences in this study. measuring health as a sequence has notable advantages in understanding health problems from a longitudinal perspective. this sequence measure allows us to capture a comprehensive picture of health evolution over time. the prevalence, events, and sequences of health problems are complex and cannot be described and examined as a whole by existing measures (e.g., single diseases or comorbidity). the sequence approach enables us to simultaneously analyze a series of health problems over time as a whole. thus, the sequence measures can directly answer important questions regarding what typical sequences of health problems people have over time, whether patterns of health sequences differ by marital status, and how specific sequences are related to financial outcomes in these two populations. 1.2. marital status impact of health on wealth despite the fact that prior research has been useful in terms of determining the influence of marriage dissolutions on health status or economic status, few studies have investigated the link between marital disruptions, health status, and economic outcomes except kim (2006). kim (2006) summarizes conceptual mechanisms and their empirical supports, through which marital status affects health and wealth. kim presents three channels through which the unmarried state may affect health: losing protective or regulatory effects deriving from spouses, producing stress effects from marital dissolution, or generating selective effects with more healthy people being married. based on these mechanisms and previous empirical findings, we hypothesize as follows: h1: unmarried women have more severe and costly sequences of chronic diseases than their married peers do. the disadvantages that unmarried women face in retirement wealth accumulation are more straightforward than the health disadvantages. in general, unmarried women have lower levels of earning and lower financial resources in their later years. health problems negatively affect retirement wealth regardless of marital status. however, given that unmarried women have more severe and costly health conditions and fewer available financial resources, their medical expense burden is relatively higher than that of married women. this heavier burden can lead to less accumulated wealth. based on these mechanisms and previous empirical findings, we hypothesize as follows: h2: any sequence of chronic diseases has a more negative impact on retirement wealth for unmarried women than for their married peers. 257s. shin, h. kim / financial services review 26 (2017) 255–270 2. method 2.1. data we use data from 12 waves of the health and retirement study (hrs) 1992, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, and 2014. the hrs is a longitudinal study of middle-aged and older americans sponsored by the u.s. national institute on aging (nia). the hrs provides comprehensive information regarding the prevalence of major chronic conditions and financial status. to examine the sequential patterns of individuals’ health conditions, we restrict our sample to female respondents who were aged 51 to 61 in 1992 and who participated in all 12 waves of the survey, with complete information available on the health variables. of the 12,652 respondents interviewed in the 1992 hrs, 6,785 are women, and 5,074 are age-eligible. after excluding data as a result of attritions or missing respondents for a 22 year period, we use a sample of 2,476 women (976 married and 1,500 unmarried) for our analyses. this study focuses on women alive in their later life. currently, the average life expectancy of an american woman is age 86.6 and 25% of women live past age 90. women’s increased longevity can lead women at higher risk of managing health and financial resources. to minimize and better cope with financial risk from living too long with lingering health problems, women need more comprehensive health information in their later years. thus, we focus on age 50 or older women who are still alive in 2014 with our selected sample. this sample selection in our study can lead to selection biases. to address the concerns related to attrition and associated sample construction, we conduct sensitivity tests by applying inverse probability weights to correct for differential sample loss. to create the inverse probability weights, we follow the method used in previous studies (baulch & quisumbing, 2011; watson & wooden, 2009). we use the baseline data with all women respondents who responded in 1992 whether or not they are dropped out after 1992. the dependent variable is a binary variable coded as 1 if they are included in our sample and 0 for otherwise. we include basic demographics, asset, income, and a variety of health status that greatly affect attrition as predictor variables consistent with previous studies. probability to be observed in our sample for each person is estimated by fitting a logistic regression (available upon request). the weighting adjustment is set to the inverse of the probability so that observations that have less probability being included in our sample contribute more to the analysis. the pseudo r2 in our logistic regression is 0.145, which is higher than the previous research (0.133 in baulch & quisumbing, 2011; 0.090 in watson & wooden, 2009). characteristics of sample and statistics applying inverse probability weights are in appendix a, highlighting that the results with adjusting selection bias are also qualitatively consistent and quantitatively similar with our main results. 2.2. measures 2.2.1. health: sequences of chronic diseases the hrs collects data on seven physical chronic diseases that commonly occur later in life: cancer, lung disease, heart condition, stroke, diabetes, high blood pressure, and arthritis. 258 s. shin, h. kim / financial services review 26 (2017) 255–270 to identify common sequences of health conditions over a 22 year period, we adopt a three-step approach that is newly introduced in recent studies (kim, shin, & zurlo, 2015; salmela-aro, kiuru, nurmi, & eerola, 2011). first, we categorize the diseases into three groups based on the degree of threat to life and the related treatment costs (smith, 1999, 2003). smith’s original classification is severe (heart condition, stroke, cancer, or lung disease) and mild (high blood pressure, arthritis, or diabetes). in this study, we include diabetes in severe conditions because it is one of the top ten causes of death (heron, 2015) and has a similar prevalence distribution to the rest of severe conditions unlike high blood pressure or arthritis (available upon request). we subdivide severe conditions into cardiovascular and non-cardiovascular conditions based on the common disease classifications in previous medical studies. cardiovascular disease is one of the most common causes of death in the advanced countries including the united states. previous studies have commonly classified cardiovascular disease as one group and the rest of the diseases as another group for the comparison of mortality (de jager et al., 2009), health care use (christiansen et al., 2013; kim et al., 2011), risk factors (lloyd-jones et al., 2007; stamler et al., 1999), and comorbidities (gerber et al., 2013). in our study, we adopt the following classifications: (1) cardiovascular diseases (heart condition or stroke; hereafter, c), (2) non-cardiovascular diseases (cancer, lung disease, or diabetes; hereafter, o), and (3) mild diseases (high blood pressure or arthritis; hereafter, m). this categorization allows us to determine eight exhaustive disease types that each individual can experience in a given wave, such as c, o, m, combinations of the three (i.e., mc, mo, co, or mco), and n (no diseases). second, each individual’s sequences are identified from eight disease types by chronologically recoding the health states for each wave. some examples of the sequences for 12 waves are n/n/n/n/n/n/n/n/n/n/n/n (e.g., no disease sequence) or m/m/m/m/m/m/m/ mc/mc/mc/mc/mc (e.g., mild and cardiovascular sequence: having mild diseases until the seventh wave, newly experiencing cardiovascular diseases (c) in the eighth wave and maintaining an mc state until the 12th wave). third, many individuals have similar sequences but not identical sequences. an attempt to analyze too many unique sequences makes analyses complicated or infeasible. thus, we group similar sequences into several homogeneous groups with low within-group heterogeneity but high between-group heterogeneity (piccarreta & billari, 2007). to classify groups or patterns with similar sequences, we use an optimal matching algorithm (oma) introduced in social sciences by abbott (1995). oma is a set of techniques that measures sequence resemblance. the basic idea behind oma is to measure the dissimilarity of two sequences by considering how much effort (e.g., insert, delete , or substitute heath status) is required to transform one sequence into the other, whereby the two sequences can belong to the same sequence pattern (barban, 2013). oma do not directly answer questions about sequence pattern; rather, they generate interval-level measures of resemblance between sequences. these measures, taken over a sequence data set, are then input to clustering, scaling, or grouping algorithms, which in turn generate information on typical patterns of sequences (abbott & hrycak, 1990). then, we use ward’s method, which is the best available method for clustering the identified individuals’ sequential patterns (hair, black, babin, & anderson, 2006). ward’s method is a criterion applied in hierarchical cluster analysis, which minimizes the total 259s. shin, h. kim / financial services review 26 (2017) 255–270 within-cluster variance. this method begins with each respondent having a distinct pattern. then, in a step-wise process, two pairs of patterns are merged. the two pairs are selected to minimize any increase in the within-pattern variance. this merging process continues until we select an appropriate number of patterns. unfortunately, there is no standard and objective selection procedure available to determine the number of sequences. one suggested criterion is the extent of change (i.e., a sudden decrease) in explanatory power (e.g., r2) by an additional sequence (hair et al., 2006). we conduct a pseudo-analysis of variance (anova) test to determine the number of sequences (salmela-aro et al., 2011). we select four sequence patterns because after these four patterns, the explanatory power (r2) of an additional sequence decreases by 50% to 0.05 from 0.1 for each of the four patterns (available on request). as a result, we use five common sequence patterns in this study: the four sequences identified by sequence analysis along with a no disease sequence. 2.2.2. marital status we divide our sample into two subgroups: married or unmarried. married is defined as maintaining a marriage for the 22 year period. unmarried is defined as continuously single status or a change from married to single (widowed or divorced) in the 22 year period.1 to examine differences within the unmarried women, we use two additional types of single status. one is a binary variable representing single status: 1 for a change from married to single and 0 otherwise (i.e., single for 22 years). the other is a continuous variable of the single period: the number of waves that a woman remains single. 2.2.3. financial outcome we use the rate of change in household net worth between 1992 and 2014 as a financial outcome variable. the household net worth in each year is equal to the sum of household assets minus household debts. net worth is adjusted using the consumer price index (2014 � 100). because the same dollar amount of wealth changes represents a different magnitude depending on the wealth level, we use the rate of change of net worth rather than the dollar amounts, as in previous studies (hurd & kapteyn, 2003; kim, 2006). 2.2.4. other variables relevant demographic variables are included in the regressions: age (year) at 1992, race (non-white � 1 and white � 0), education (year), employment status (the number of waves at work), and dollar amount of accumulated household income (the sum of household income during the 22 years). we also include net worth in 1992 to control for wealth levels at baseline. health insurance is also included as a variable. the hrs asks respondents if they are currently covered by any health insurance, which falls into one of five categories: medicaid, medicare or other public health insurance programs, insurance provided by current or previous employers including the spouse’s employers, other privately purchased health insurance, or no health insurance. we use the number of waves for which the respondents are covered by one of these health insurance types to represent the individuals’ insurance status and changes over time. 260 s. shin, h. kim / financial services review 26 (2017) 255–270 2.3. analysis we use ols regressions to estimate the impact of health sequence on the change in net worth over a 22 year period. we estimate the impacts for married and unmarried women separately. for this separate estimation, we first test to determine any differences in the coefficients of the health sequence patterns between the two groups. an f test is conducted using an interaction model with a binary variable of marital status, the sequences patterns of chronic diseases, and the interaction terms for marital status and sequence patterns. the null hypothesis is that each coefficient of the corresponding interaction terms between marital status and sequence patterns is not different from zero. based on the result of the f test, this hypothesis is rejected (f � 3.757, p � 0.000). the test results indicate that the impact of health sequence on net worth change differs between unmarried and married women and separate estimations are needed to provide adequate results. we use raw or untransformed data for our analysis. however, financial data tend to have a right-skewed distribution that can lead to biased estimates. thus, we conduct sensitivity tests using natural logarithm to mitigate bias from the right-skewed distribution (see appendix a). the results with transformed data are also qualitatively consistent and quantitatively similar with our main results. 3. results relevant characteristics of the sample based on marital status are presented in table 1. this table emphasizes that unmarried women are socioeconomically worse off than their married peers. unmarried women are older, have lower levels of education attainment, income and wealth, and are more likely to lack health insurance. the circumstances of the unmarried women worsen after 22 years. the median wealth of unmarried women ($116,800) is only 42.6% of that of married women ($274,100) in 1992, and this level further decline to 26.8% ($84,000 vs. $313,100) in 2014. the five sequence patterns used as health measures are present in table 2: multimorbidity, co-morbidity, mild disease, late event, and no disease. five components listed in the common patterns in table 2 indicate the chronic condition states from 1–2 waves, 3–5 waves, 6–8 waves, 9–10 waves, and 11–12 waves, respectively. approximately 21.8% of the respondents exhibit a multi-morbidity pattern. this pattern is a combination of mild (high blood pressure or arthritis), cardiovascular (heart disease or stroke) or non-cardiovascular diseases (cancer, lung disease, or diabetes), and it is the most complicated/costly among the five patterns. co-morbidity (22.7%) is a combination of mild diseases and non-cardiovascular diseases. this pattern features mild diseases at baseline and new non-cardiovascular diseases emerging later. mild disease is the most frequent pattern (36.7%), and it primarily includes mild diseases throughout the entire period. late event (15.6%) shows no disease for most of the time period, with some health events occurring at the end of the period. no disease refers to only 3.3% of the population that have experienced no disease during the entire period. table 3 shows that the distribution of health sequence patterns differ significantly between married and unmarried women. unmarried women are more likely to have poor health status 261s. shin, h. kim / financial services review 26 (2017) 255–270 than married women. in particular, more unmarried than married women have complicate and costly health sequences with either the multi-morbidity (23.1% vs. 19.8%) or comorbidity (24.7% vs. 19.6%) patterns. table 4 shows that married and unmarried women exhibit differences in wealth levels, and the change according to health sequence patterns. in 1992, wealth disparities are substantial across health sequence patterns between these two groups. for example, unmarried women with a no disease or multi-morbidity pattern have wealth levels of $140,443 or $87,302, respectively, constituting only 27.8% or 39.8% of the wealth of married ($506,100 or $219,310, respectively). among unmarried and married groups, those with multi-/co-morbidity patterns typically have lower wealth than women with other health patterns. table 4 also indicates that during the 22 year period, wealth disparities between the married and unmarried widen across the health sequence patterns. unmarried women deplete table 1 characteristics of the sample marrieda unmarriedb statisticsc n 976 1,500 age (mean, sd) at 2014 76.5 (3.0) 77.4 (3.1) �7.0** race (%) white 90.0 74.6 89.4** non-white 10.0 25.4 household net worth ($1,000) at 1992 mean (sd) 461.3 (610.3) 249.8 (431.2) 9.3** median 274.1 116.8 at 2014 mean (sd) 603.5 (818.9) 250.4 (475.4) 12.1** median 313.1 84.0 household income ($1,000)d mean (sd) 1,032.9 (786.3) 535.7 (488.9) 17.5** median 843.7 417.9 education years (mean, sd) 12.6 (2.8) 12.1 (3.0) 4.7** working period (waves, mean, sd)e 3.2 (3.1) 3.4 (3.2) �1.4 health insurance (waves, mean, sd)f medicare 6.5 (1.8) 6.9 (2.0) �5.6** medicaid 0.3 (1.1) 1.2 (2.6) �12.2** employer 7.0 (3.8) 5.3 (4.1) 10.6** others 2.3 (2.9) 2.1 (2.8) 1.3 no insurance 0.7 (1.4) 1.2 (1.8) �8.4** household net worth and income were adjusted by cpi (2014 � 100). amarried for the 22 year period. bsingle for the 22 year period or changed from married to single between 1992 and 2014. cstatistics are t statistics for continuous variables and �2 for categorical variables, indicating the differences between married and unmarried women. dsum of household income for 22 years. ethe number of waves working for pay. fthe number of waves covered by medicare, medicaid, an employer-sponsored insurance plan, any other privately purchased health insurance plan or without health insurance for 22 years. *p � 0.05, **p � 0.01. 262 s. shin, h. kim / financial services review 26 (2017) 255–270 more wealth than married women do. in the last column in table 4, the wealth of unmarried women declines by 28%, whereas the wealth of married women increases by 14% between 1992 and 2014. comparisons across the health sequence patterns reveal clearer differences. for the 22 year period, unmarried (68%) with the no disease pattern show higher rates of wealth accumulation than married (22%); however, all of the sequences with health problems are more detrimental to wealth for the unmarried than for married women. the wealth levels of unmarried with multi-/co-morbidity patterns decline by 41% and 26%, respectively, whereas those of married women decrease only by 7% for multi-morbidity and increase by 21% for co-morbidity. furthermore, the relative wealth level (ratio) gaps between the two patterns and the no disease pattern are wider for unmarried women (22% to 27% from 62%) than for the married women (33% to 46% from 43% to 47%). table 5 presents the results of regression analyses to estimate wealth change rates by marital status between 1992 and 2014. these results show that health sequence patterns have different impacts on wealth changes according to marital status. in particular, the heath sequence patterns lead to a disproportionately weaker increase in wealth for unmarried women. table 2 five common sequence patterns of chronic diseases across a 22 year period (12 waves) sequence patterns n (% of total) representative patternsa multi-morbidity 539 (21.8) m–mc–mco–mco–mcocombination of multiple mild and severe diseases co-morbidity 562 (22.7) m–m–mo–mo–mo combination of mild disease and non-cardiovascular disease mild disease 908 (36.7) m–m–m–m–mmild disease during the period late event 386 (15.6) n–n–n–x–xbmild or severe disease at the end of the period no disease 81 (3.3) n–n–n–n–nno disease during the period total 2,476 (100) m � mild disease; c � cardiovascular disease; o � non-cardiovascular disease; n � no disease. afive components listed in the common patterns indicate the chronic condition states from 1–2 waves, 3–5 waves, 6–8 waves, 9–10 waves, and 11–12 waves, respectively. bx represents any type of chronic condition or their combination. table 3 the distribution (%) of health sequence patterns by marital status married unmarried multi-morbidity 19.8 23.1 co-morbidity 19.6 24.7 mild disease 41.1 33.8 late event 15.2 15.9 no disease 4.4 2.5 all 100 100 �2 25.0 (p � .000) 263s. shin, h. kim / financial services review 26 (2017) 255–270 for unmarried women, the coefficients of the health sequence patterns range from �0.588 to �0.829 in model 1. to better understand these differences, table 6 shows the predicted dollar amount of the wealth gaps between those with the four health sequences and those with the no disease pattern. the unmarried women with health problems save $117,953 (multi-morbidity), $106,657 (co-morbidity), $93,621 (mild disease), and $78,806 (late event) less than those with the no disease pattern. model 2 in table 5 includes additional two variables of single status and periods remaining single to control for the heterogeneous composition of the unmarried sample (i.e., continuously single vs. a change from married to single). the coefficients of the health sequence patterns in model 2 are similar to those in model 1; thus, demonstrating robustness. in contrast to the impact on their unmarried peers, the effect of the health sequence patterns is minor or nonexistent for married women. the coefficients of the health sequence patterns for married women range from �0.048 to 0.207; however, these coefficients are not statistically significant. of the control variables, accumulated household income shows a positive effect on the wealth change, while the amount of net worth in 1992 shows a negative effect on changes in wealth for both married and unmarried women. medicare coverage affect negatively changes in wealth for married women. other control variables such as age, race, and education are not statistically significant. 4. discussion in this study, we advance the literature on marital status difference in the impact of health problems on women’s retirement wealth in two distinct ways. first, we examine the impact table 4 net worth (median) by marital status and health sequence patterns 1992 2014 changeb (b-a)/a $(a) ratioa $(b) ratio married multi-morbidity 219,310 0.43 204,000 0.33 �0.07 co-morbidity 236,180 0.47 285,000 0.46 0.21 mild disease 285,103 0.56 360,500 0.58 0.26 late event 344,785 0.68 330,500 0.53 �0.04 no disease 506,100 1.00 619,100 1.00 0.22 total 274,138 313,142 0.14 unmarried multi-morbidity 87,302 0.62 51,500 0.22 �0.41 co-morbidity 86,881 0.62 64,150 0.27 �0.26 mild disease 146,769 1.05 104,500 0.44 �0.29 late event 155,541 1.11 135,750 0.58 �0.13 no disease 140,443 1.00 235,600 1.00 0.68 total 116,825 84,000 �0.28 net worth values were adjusted by cpi (2014 � 100). athe ratio of each pattern to no disease pattern. b(net worth in 2014 net worth in 1992)/net worth in 1992. 264 s. shin, h. kim / financial services review 26 (2017) 255–270 t ab le 5 r eg re ss io n re su lts fo r th e ne t w or th ch an ge ra te be tw ee n 19 92 an d 20 14 m ar ri ed u nm ar ri ed b (s e ) � p m od el 1 m od el 2 b (s e ) � p b (s e ) � p h ea lth se qu en ce pa tte rn s m ul tim or bi di ty � .0 48 (. 22 3) � .0 14 .8 28 � .8 29 (. 29 2) � .2 12 .0 05 � .8 41 (. 29 4) � .2 13 .0 04 c om or bi di ty .1 31 (. 22 3) .0 37 .5 57 � .7 23 (. 29 0) � .1 89 .0 13 � .7 11 (. 29 3) � .1 85 .0 15 m ild di se as e .2 07 (. 21 0) .0 72 .3 23 � .6 75 (. 28 6) � .1 94 .0 18 � .6 74 (. 28 8) � .1 93 .0 19 l at e ev en t .0 28 (. 22 8) .0 07 .9 02 � .5 88 (. 29 5) � .1 30 .0 47 � .5 88 (. 29 8) � .1 30 .0 48 (n o di se as e) a ge (y ea rs ) in 19 92 � .0 35 (. 02 3) � .0 75 .1 28 .0 22 (. 01 9) .0 42 .2 44 .0 21 (. 01 9) .0 40 .2 66 r ac e n on -w hi te � .2 27 (. 14 4) � .0 48 .1 16 .0 11 (. 10 3) .0 03 .9 17 � .0 19 (. 10 5) � .0 05 .8 55 (w hi te ) e du ca tio n (y ea rs ) .0 25 (. 01 9) .0 49 .1 86 .0 28 (. 01 6) .0 52 .0 84 .0 27 (. 01 7) .0 48 .1 13 w or ki ng pe ri od (w av es ) .0 05 (. 01 5) .0 11 .7 39 .0 24 (. 01 6) .0 46 .1 27 .0 17 (. 01 6) .0 32 .3 01 h ea lth in su ra nc e (w av es ) m ed ic ar e .0 04 (. 04 2) .0 05 .9 24 � .0 69 (. 03 4) � .0 84 .0 43 � .0 71 (. 03 4) � .0 86 .0 40 m ed ic ai d .0 14 (. 04 6) .0 11 .7 66 .0 37 (. 02 2) .0 57 .0 95 .0 28 (. 02 3) .0 42 .2 22 e m pl oy er .0 37 (. 01 9) .0 99 .0 53 .0 20 (. 01 8) .0 50 .2 62 .0 18 (. 01 8) .0 44 .3 27 o th er s .0 16 (. 02 1) .0 34 .4 37 � .0 13 (. 02 0) � .0 22 .5 07 � .0 17 (. 02 0) � .0 29 .3 89 n o in su ra nc e .0 40 (. 04 4) .0 39 .3 60 � .0 40 (. 03 5) � .0 43 .2 62 � .0 44 (. 03 6) � .0 47 .2 20 h ou se ho ld in co m e ($ 1, 00 0) .0 01 (. 00 0) .3 65 .0 00 .0 01 (. 00 0) .3 14 .0 00 .0 01 (. 00 0) .3 24 .0 00 n et w or th in 19 92 ($ 1, 00 0) � .0 01 (. 00 0) � .3 65 .0 00 � .0 01 (. 00 0) � .2 52 .0 00 � .0 01 (. 00 0) � .2 49 .0 00 si ng le st at us m ar ri ed to si ng le a � .1 51 (. 13 4) � .0 45 .2 60 (s in gl e) w av es in si ng le b .0 02 (. 01 5) .0 05 .8 97 in te rc ep t 1. 34 7 (1 .1 05 ) .2 23 � .6 25 (1 .0 05 ) .5 34 � .4 26 (1 .0 37 ) .6 81 a dj us te d r 2 .1 58 .1 32 .1 34 a c ha ng ed fr om m ar ri ed to si ng le . b t he nu m be r of w av es in si ng le . 265s. shin, h. kim / financial services review 26 (2017) 255–270 of long-term chronic health sequences on changes in wealth. this health measure is a new and innovative approach compared with the traditional method of considering current health status. the second contribution stems from our investigation of a population of women aged 50�. this period of life is important in that chronic health conditions tend to begin in one’s 50s, and the impacts of such conditions on household finance emerge over the following 20 to 30 years. we test two hypotheses. first (h1), unmarried women have more severe and costly sequences of chronic diseases than married women. the results shown in table 3 support this hypothesis. these findings are new and informative in that the results indicate potential paths of health problems in women from their 50s to their 80s. second (h2), any sequence of chronic diseases has a more negative impact on the retirement wealth of unmarried women than on married women. the results reported in tables 4 and 5 also support this hypothesis. these findings confirm the disadvantages for unmarried women in terms of their financial security in their later years. the retirement wealth of unmarried women further decreased as a result of specific health sequences that they experienced. decreased retirement wealth ranged from $78,806 to $117,953 over the 22 years of the study period. this amount is equivalent to a less annual accumulation of $3,582 to $5,362 than those with no disease. previous studies estimated $905 to $5,240 of annual wealth depletion as a result of chronic diseases: $905 to $4,211 for adults aged 51–61 with new mild diseases or new severe diseases (1992–1996 hrs data), or $5,240 for adults aged 70 or older with any new chronic diseases (1993–1995 ahead data; smith, 1999); $3,140 to $4,170 for unmarried women aged 70 or older with new severe chronic diseases (1993–2002 ahead data; kim, 2006). all of these estimations for middle-aged and older adults were based on differences in age, marital status, health measures, or estimated periods. nevertheless, our estimates are between the minimum and maximum estimates from other studies. this result could be a good benchmark for understanding the quantitative magnitude of the financial impact of long-term health sequences on unmarried older women. these findings have important implications for health management and retirement savings for middle-aged and older women and practitioners. the finding that middle-aged and older table 6 predicted change in net worth gaps between 2014 and 1992 across health sequence patterns: unmarried women predicted change rate in net worth (a) net worth ($) at 1992 (b) predicted net worth ($) at 2014 (b�(a�b)) difference in difference (dd) of net worth ($)a multi-morbidity �0.003 87,302 87,040 �117,953b co-morbidity 0.127 86,881 97,915 �106,657 mild disease 0.164 146,769 170,839 �93,621 late event 0.250 155,541 194,426 �78,806 no disease 0.838 140,443 258,134 reference add � [(nw2014 hs � 1,2,3 or 4) � (nw1992 hs � 1,2,3 or 4)] � [(nw2014 hs � 5) � (nw1992 hs � 5)], where nw is net worth and hs is health sequences with 1 � multi-morbidity; 2 � co-morbidity; 3 � mild disease; 4 � late event; 5 � no disease. b(87,040–87,302) � (258,134–140,443) � �117,953. 266 s. shin, h. kim / financial services review 26 (2017) 255–270 women have five common health sequences in their 50s to 80s can enable them to predict potential health trajectories they may encounter and prepare for an appropriate health management plan. for example, when middle-aged women have high blood pressure in their early 50s, approximately 27% (24.6% for the married and 28.3% for the unmarried) of them will likely experience heart disease or stroke in their late 50s or early 60s (i.e., multimorbidity sequence). based on these predictable health sequences in the later years, middle-aged or older women can assess their current health issues and potential health status. women can also take preventive actions to minimize the consequences of health problems and to remain healthy. these actions include managing risk factors related to expected diseases, such as smoking, drinking, diet, or exercise factors. financial practitioners may weave these issues into financial planning by actively engaging health conversations with their clients. for example, high blood pressure is very common in older women. to control this disease, some lifestyle changes are recommended such as salt reduction, and limiting alcohol. people can also better control blood pressure by taking medication about the same time everyday. changes in lifestyle and timing of taking medication can lead to a better control of blood pressure. better control results in fewer doctor visits, which reduce the cost of care (mcclanahan, 2014) and improve financial security. in particular, unmarried women are at greater risk for costly and complicated health sequences; hence, they need to pay more attention to these health sequences and the associated preventive actions. health educators and practitioners can use this information to make and provide specific behavioral or dietary guidelines for middleaged or older women. another implication for unmarried women is that individual health management should be incorporated into retirement planning to minimize financial insecurity in their later years. without any health problems, unmarried and married older women alike have similar financial stability in terms of wealth accumulation rates. once health problems occur, however, any health sequences are substantially detrimental to retirement wealth accumulation for unmarried women but not necessarily for married women. this finding can be interpreted as a lack of financial buffer from marital benefits. in particular, for unmarried women, the most influential factor for long-term retirement savings in this study is health sequences whose magnitude of influence exceeds even that of income or baseline wealth. this result strongly indicates that good health management by unmarried women can minimize the need for retirement savings and accordingly represent an alternative to retirement savings. the other major finding can provide practical knowledge for retirement savings plans related to health issues. many previous studies and financial planners have suggested the need for adequate retirement savings to maintain a preretirement living standard (e.g., 6–8 times the last five years annual salary); however, most suggestions do not consider health care costs. our finding suggests that middle-aged and older unmarried women with health problems spend $3,582 to $5,362 more annually than those with no disease. this figure can provide a potential guideline for appropriate health care savings for unmarried women based on their expected health sequences. additionally, this study can add credence for professionals to help their clients assess the impact of lifestyle on future financial projections, including paying attention to health-related factors. for example, financial planners may be 267s. shin, h. kim / financial services review 26 (2017) 255–270 encouraged to do more on the personal side of financial planning, rather than focusing mainly on the technical side. for another example, financial advisors can incorporate longevity planning into this study with their clients. the life expectancy of a woman aged 65 in 2014 is about 87 years old.2 based on the estimation of this study, practitioners may advise that she needs $78,804 to $117,964 (i.e., $3,582 � 22 to $5,362 � 22) for health care costs unless she has no disease during the rest of her life. despite new findings and important implications, this study has some limitations suggesting the need for further studies. our sample includes only women who have lived during the 22 year period and participated in all 12 waves. women who have been excluded because of death or lack of contact may follow different patterns of health sequences and wealth accumulation. although we use the conventional classification of cardiovascular and noncardiovascular diseases within sever chronic conditions, non-cardiovascular group includes three diseases with different etiologies, organ-systems, and probabilities of co-occurrence. more refinement and improvement of disease classification are needed for understating comprehensive sequences of each chronic condition. we did not adjust household net worth for household size because of the methodological issue for women whose marital status changed from married to single. we also did not consider spouses’ health status for the married group because of the data limitations. for married women, the impacts of their spouses’ health status can be an important factor on household net worth changes. a more refined adjustment considering family characteristics are needed in future studies. notes 1 other types of marital status changes, such as changes from single to married, are not included in analysis. 2 https://www.ssa.gov/planners/lifeexpectancy.html appendix a. sensitivity tests to check potential biases (a) characteristics of the sample applying inverse probability weight married unmarried age at 2014 (mean, sd) 76.0 (2.9) 76.9 (3.1) race (%) white 89.0 72.1 non-white 11.1 27.9 education years (mean, sd) 12.3 (2.9) 11.7 (3.2) household incomea ($1,000, mean, sd) 963.2 (741.9) 495.0 (459.5) household net worth ($1,000) at 1992 mean (sd) 395.0 (538.5) 210.3 (387.7) median 227.1 86.6 at 2014 mean (sd) 531.1 (761.8) 212.4 (422.9) median 275.0 64.4 asum of household income for 22 years. 268 s. shin, h. kim / financial services review 26 (2017) 255–270 references abbott, a. (1995). sequence analysis: new methods for old ideas. annual review of sociology, 21, 93–113. abbott, a., & hrycak, a. (1990). measuring resemblance in sequence data: an optimal matching analysis of musicians’ careers. american journal of sociology, 96, 144–185. adams, p., hurd, m. d., mcfadden, d., merrill, a., & ribeiro, t. (2003). healthy, wealthy, and wise? tests for direct causal paths between health and socioeconomic status. journal of econometrics, 112, 3–56. addo, f. r., & lichter, d. t. (2013). marriage, marital history, and black white wealth differentials among older women. journal of marriage and family, 75, 342–362. barban, n. (2013). family trajectories and health: a life course perspective. european journal of population, 29, 357–385. baulch, b., & quisumbing, a. (2011). testing and adjusting for attrition in household panel data. chronic poverty research centre (cprc) toolkit note. (available at http://www.chronicpoverty.org/uploads/assets/ files/toolbox/cprc_toolkit_note_on_attrition-revfinal.pdf) christiansen, c. b., olesen, j. b., gislason, g., lock-hansen, m., & torp-pedersen, c. (2013). cardiovascular and non-cardiovascular hospital admissions associated with atrial fibrillation: a danish nationwide, retrospective cohort study. bmj open, 3, e001800. de jager, d. j., grootendorst, d. c., jager, k. j., van dijk, p. c., tomas, l. m., ansell, d., & wetzels, j. f. (2009). cardiovascular and noncardiovascular mortality among patients starting dialysis. jama, 302, 1782– 1789. gerber, y., melton, l. j., mcnallan, s. m., jiang, r., weston, s. a., & roger, v. l. cardiovascular and noncardiovascular disease associations with hip fractures. the american journal of medicine, 126, 169-e19. hair, j. f., jr., black, w. c., babin, b. j., & anderson, r. e. (2006). multivariate data analysis. upper saddle river, nj: pearson prentice hall. henretta, j. c. (2010). lifetime marital history and mortality after age 50. journal of aging and health, 22, 1198–1212. heron, m. (2015). deaths: leading causes for 2011. national vital statistics reports, 64, 1–96. hurd, m., & kapteyn, a. (2003). health, wealth, and the role of institutions. the journal of human resources, 38, 386–415. hughes, m. e., & waite, l. j. (2009). marital biography and health at mid-life. journal of health and social behavior, 50, 344–358. institute for american values. (2011). why marriage matters: thirty conclusions from the social sciences. washington, dc: institute for american values. joyce, j. a. (2007). women, marriage and wealth: the impact of marital status on the economic well-being of women through the life course. new york, ny: gordon knot books. kim, h. (2006). older women’s health and its impact on wealth. journal of women & aging, 18, 75–91. (b) regression results applying inverse probability weight and ln transformation adopting weight ln (net worth change rate) married unmarried married unmarried b (se) p b (se) p b (se) p b (se) p health sequence patterns multi-morbidity .012 (.239) .959 �.852 (.304) .005 �.006 (.036) .864 �.144 (.047) .002 co-morbidity .147 (.241) .542 �.706 (.303) .020 .020 (.036) .583 �.124 (.047) .008 mild disease .262 (.230) .254 �.705 (.299) .018 .027 (.034) .424 �.131 (.046) .005 late event .143 (.248) .564 �.610 (.309) .048 .005 (.037) .897 �.114 (.048) .017 (no disease) control variables included yes yes yes yes intercept 1.484 (1.103) .179 �.400 (1.080) .712 1.739 (.179) .000 1.808 (.169) .000 adjusted r2 .156 .103 .150 .127 269s. shin, h. kim / financial services review 26 (2017) 255–270 kim, m. h., johnston, s. s., chu, b. c., dalal, m. r., & schulman, k. l. (2011). estimation of total incremental health care costs in patients with atrial fibrillation in the united states. circulation: cardiovascular quality and outcomes, 4, 313–320. kim, h., shin, s., & zurlo, k. a. (2015). sequential patterns of health conditions and financial outcomes in late life: evidence from the health and retirement study. the international journal of aging and human development, 81, 54–82. laatsch, f. e., & klein, d. p. (2010). social security and retirement savings accounts. financial services review, 19, 37–57. liu, h., & umberson, d. j. (2008). the times they are a changin’: marital status and health differentials from 1972 to 2003. journal of health and social behavior, 49, 239–253. lloyd-jones, d. m., dyer, a. r., wang, r., daviglus, m. l., & greenland, p. (2007). risk factor burden in middle age and lifetime risks for cardiovascular and non-cardiovascular death. the american journal of cardiology, 99, 535–540. maselko, j., bates, l. m., avendano, m., & glymour, m. m. (2009). the intersection of sex, marital status, and cardiovascular risk factors in shaping stroke incidence: results from the health and retirement study. journal of the american geriatrics society, 57, 2293–2299. mcclanahan, c. (2014, august 11). how to fix your finances. by fixing your blood pressure. time. (available at http://time.com/money/3086735/how-to-fix-your-finances-by-fixing-your-blood-pressure/) national women’s law center. (2015). women and social security. (available at http://www.nwlc.org/resource/ women-and-social-security) piccarreta, r., & billari, f. c. (2007). clustering work and family trajectories by using a divisive algorithm. journal of the royal statistical society: series a (statistics in society), 170, 1061–1078. salmela-aro, k., kiuru, n., nurmi, j., & eerola, m. (2011). mapping pathways to adulthood among finnish university students: sequences, patterns, variations in family and work related roles. advances in life course research, 16, 25–41. schoenberg, n. e., kim, h., edwards, w., & fleming, s. t. (2007). burden of common multiple-morbidity constellations on out-of-pocket medical expenditures among older adults. the gerontologist, 47, 423–437. smith, j. p. (1999). healthy bodies and thick wallets: the dual relation between health and economic status. the journal of economic perspectives, 13, 145–166. smith, j. p. (2003). consequences and predictors of new health events. in wise, d. a. (ed.), analyses in the economics of aging (pp. 213–237). chicago, il: university of chicago press. stamler, j., stamler, r., neaton, j. d., wentworth, d., daviglus, m. l., garside, d., & greenland, p. (1999). low risk-factor profile and long-term cardiovascular and noncardiovascular mortality and life expectancy: findings for 5 large cohorts of young adult and middle-aged men and women. jama, 282, 2012–2018. watson, n., & wooden, m. (2009). identifying factors affecting longitudinal survey response. in lynn, p. (ed.), methodology of longitudinal survey. chichester, uk: john wiley & sons, ltd. yuh, y., hanna, s., & montalto, c. p. (1998). mean and pessimistic projections of retirement adequacy. financial services review, 7, 175–193. zajacova, a., & burgard, s. a. (2013). healthier, wealthier, and wiser: a demonstration of compositional changes in aging cohorts due to selective mortality. population research and policy review, 32, 311–324. zissimopoulos, j. m. (2009). gain and loss: marriage and wealth changes over time. rand working paper wr724. (available at https://www.rand.org/content/dam/rand/pubs/working_papers/2010/rand_wr724.pdf) 270 s. shin, h. kim / financial services review 26 (2017) 255–270 employees’ financial behaviors following the 2007–2009 financial crisis crystal hudson, ph.d.a,*, wookjae heo, ph.d.b, heejung park, m.s.c, lance palmer, ph.d., cfpd aclark atlanta university, 223 james p. brawley drive, atlanta, ga 30314, usa bsouth dakota state university, box 2275a/wagner hall 151, brookings, sd 57007-0295, usa cuniversity of wyoming, department 3275, 1000 e university avenue, laramie, wy 82071, usa duniversity of georgia, 205 dawson hall, athens, ga 30602-2622, usa abstract lowand middle-income employees make up the bulk of potential participants in employer sponsored retirement plans; however, employers find it difficult to alter their savings behavior. financial crises may have unintended positive effects on low-income employees’ behavior. therefore, this study examined the effect of the 2007–2009 financial crisis on employees’ financial behaviors; through ordered logistic regression analyses of data from the survey of consumer finances. following the crisis, all employees’ and low-income employees’ savings behavior significantly improved. moreover, all employees’ cash flow management behavior improved following the crisis, while it had no effect on low-income employees’ cash flow management behavior. © 2017 academy of financial services. all rights reserved. jel classification: d14 keywords: financial behaviors; low-income employees; financial crisis 1. introduction in 2013, only 40.8% of all american workers between the ages of 25 and 64 participated in an employer-sponsored retirement plan (copeland, 2014). this low participation rate is * corresponding author. tel.: �1-404-880-6413; fax: �1-(404)-880-6276. e-mail address: crhudson@cau.edu (c. hudson) financial services review 26 (2017) 19–36 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. most likely because of employees’ lack of financial literacy, lack of money management skills, and their inability to save for the future (lusardi & mitchell, 2011). employees need a certain level of financial literacy to manage their financial affairs in today’s ever-changing financial environment (braunstein & welch, 2002; hogarth, beverly, & hilgert, 2003; loibl & hira, 2005, 2006; volpe, chen, & liu, 2006). additionally, increased financial sophistication is necessary, as financial brokers sometimes push risky and complex financial products on unsuspecting consumers (asaad, 2014; lyons & scherpf, 2004). in 2013, only 16% of employees making less than $20,000 per year participated in an employer-sponsored retirement plan (copeland, 2014). on average, lower income employees are particularly ill-equipped to navigate today’s financial environment, or to make sound financial decisions to enhance their financial well-being, and they lack the financial knowledge and education needed to improve their financial situation (andersen, zhan, & scott, 2004; hudson & palmer, 2014; lusardi & mitchell, 2005; lyons, chang, & scherpf, 2006). additionally, lower income individuals have always had challenges saving for emergencies and saving for retirement (fisher, hayhoe, & lown, 2015). opportunities to expand savings and improve cash flow management practices of employees could lead to greater retirement plan participation and savings rates. thus, low-income employees can represent growth for this market if more is understood about their financial behavior. the financial crisis from 2007 to 2009 provides an opportunity to examine whether financial behaviors changed among all employees, and specifically among low-income employees, as a result of external influences. in this crisis, real gross domestic product (gdp) fell by 4.3%, while housing prices fell by 30% (rich, 2013). moreover, unemployment nearly doubled, from 5% in december 2007 to 9.5% in june 2009 (rich, 2013). the 2007–2009 financial crisis hit most americans the hardest in the form of unemployment, lost wealth, and housing insecurity (bricker, bucks, kennickell, mach, & moore, 2011; taylor et al., 2010). this dramatic financial event affected every american, either directly or indirectly, through close friends or family, and provided ample opportunities for deep reflection and evaluation of savings and cash flow management practices. while a crisis could represent a threat to one’s present well-being, it could also represent an opportunity for growth, mastery, and gains (slaikeu, 1990). in other words, could an otherwise detrimental financial crisis have a positive effect on employees’ financial behaviors? the purpose of this study was to examine the effects of the 2007–2009 financial crisis on employees’ and low-income employees’ savings and cash flow management behaviors. thus, this research contributes to the literature by further identifying factors that influence organic financial behavior changes, including external factors like a financial crisis, and examining the persistence of such changes years after the external influence. from this, appropriate supports and systems can be put into place to further reinforce the intrinsic motivation to change financial behaviors. furthermore, while most previous studies explored the effects of a financial crisis on financial behaviors of the general population, this study is unique and adds to the body of literature by examining the effects of a financial crisis on employees’ financial behaviors, because of their ready access to retirement plans and also adds to the body of literature by examining the effects of a crisis on low-income employees’ financial behaviors, because of their low retirement plan participation and contribution rates. having access to an employer20 c. hudson et al. / financial services review 26 (2017) 19–36 based retirement plan gives employees a greater ability to save in comparison to the general population. additionally, this research controls for household constraint by using poverty levels, which combine household size and income, rather than simple income measures. multiple behaviors within the savings and cash flow management categories are measured providing a broader picture of behavior change. 2. literature review a review of previous studies found that a financial crisis positively affects financial behaviors (bricker et al., 2011; fidelity investments, 2013; metlife, 2009; o’neill & xiao, 2012; taylor et al., 2010). fidelity investments (2013) conducted a study that examined financial behaviors after the 2007–2009 financial crisis by surveying the attitudes of 1,154 investors between 2008 and 2013. fidelity’s study found that the 2007–2009 financial crisis boosted the financial confidence of most investors (fidelity investments, 2013). in fact, more than half (56%) of investors reported feeling scared and confused as the crisis started, but otherwise reported being prepared and confident in their financial affairs after the financial crisis. additionally, 42% of investors noted that they had increased the amount of their emergency funds and either sought financial advice from a financial advisor (30%) or financial advice from their spouse or family member (26%). likewise, o’neill and xiao (2012) examined the budgeting, savings, and spending behaviors of 10,661 randomly selected individuals who participated in an online survey from january 2005 to december 2010. similar to this study, o’neill and xiao (2012) used a t test and multiple variate regression analysis to examine data before and after the 2007–2009 financial crisis. they found that the financial behavior of the respondents was better after the 2007–2009 financial crisis than before. respondents reported increased budgeting practices, higher savings, and more positive spending behaviors after the financial crisis. similarly, other previous studies also acknowledged the positive effects that an otherwise stressful financial crisis can have on an individual’s financial behaviors (metlife, 2009; taylor et al., 2010). in a metlife (2009) online survey conducted in 2009, 2,243 participants were randomly selected from across the u.s. thus, 75% of participants reported feeling stress about their financial situation or feeling insecure about their job stability during the 2007–2009 financial crisis. however, the majority of participants reported that the crisis was a wake-up call for them and motivated them to change their behaviors for the better (metlife, 2009). in march 2010, pew research randomly selected a national sample of 2,967 adults and interviewed them about how the 2007–2009 financial crisis affected their lives. respondents felt that the 2007–2009 financial crisis led to a new frugality in american spending and borrowing (taylor et al., 2010). not surprisingly, participants reported eating out less or eating at home more often and also reported shopping at big box discount stores with more frequency. additionally, most participants described having a positive attitude about their 21c. hudson et al. / financial services review 26 (2017) 19–36 finances, and most participants felt that their personal finances would improve in the coming year (taylor et al., 2010). 3. theoretical models previous studies have shown that financial behaviors change and are influenced by external and internal forces (bricker et al., 2011; fidelity investments, 2013; metlife, 2009; o’neill & xiao, 2012; taylor et al., 2010). the transtheoretical model of change (ttm) provides a theoretical framework of how people change, both on their own as well as when they work with a professional (prochaska, diclemente, & norcross, 1992). likewise, the crisis theory provides a theoretical framework of how individuals learn, adapt and grow after a crisis disrupts their normal lives (caplan, 1971). several key processes of change within the ttm model contribute to cognitive reevaluation and perspective shifting, which in turn lead to outwardly observable behavior changes. the processes of consciousness-raising, dramatic relief, environmental revaluation, and social liberation can be triggered when an individual experiences or witnesses someone else going through a negative event, such as a financial crisis (prochaska et al., 1992). consciousness-raising is simply becoming more aware of certain behaviors and the effect those behaviors have on the individual and those with whom he or she associates. furthermore, during the financial crisis, personal finance topics, such as savings and cash flow management, were a common topic of discussion in the news, which led to increased awareness of positive practices, such as saving and budgeting. palmer, bliss, goetz, and moorman (2010a, 2010b) found that even simple consciousness raising activities can lead to significant financial behavior changes. dramatic relief simply refers to having an emotional response (i.e., feelings) when certain topics are discussed (o’neill & xiao, 2012). as unemployment, foreclosures, and bankruptcies abounded, and individuals experienced these things firsthand or indirectly, many likely developed strong emotions associated with basic personal finance vocabulary and practices. finally, social liberation and self-reevaluation further accelerate the behavior change process as individuals recognize that society is more supportive and rewarding of positive financial behaviors, such as saving (social liberation; xiao, newman, prochaska, leon, bassett, & johnson, 2004). opportunities for individuals to experience these cognitive processes of change abounded during the 2007– 2009 financial crisis and likely led to widespread changes in consumer financial behavior. the crisis theory is derived from the psychoanalytic theory and ego psychology and proposes that a crisis disrupts and causes chaos in an individual’s life, but provides them with an opportunity to problem solve, adapt, and grow, and perhaps move to a higher state of being (caplan, 1971; woolley, 1990). caplan (1971) defines a crisis as a threat to individuals’ normal state of being or an obstacle to their important life goals. during a crisis, an imbalance exists and confusion as well as disorder ensues (caplan, 1971). as this occurs, an individual makes many attempts to resolve this crisis, using known and new problem solving techniques, to maintain or restore balance (caplan, 1971). thus, immediately following the crisis, as a state of disequilibrium occurs, an individual must find some way of coping with the crisis (woolley, 1990). a new state 22 c. hudson et al. / financial services review 26 (2017) 19–36 of equilibrium occurs, either at a higher level of functioning where growth has occurred, or at a lower level of functioning, where individuals have fallen to a regressed state of functioning (woolley, 1990). as proposed by this study, a crisis presents individuals with an opportunity to problem solve, adapt, grow, and perhaps move to a higher level of being. based on these theories and a review of literature, the four applicable hypotheses for this study are as follows: hypothesis 1: the overall savings behavior of all employees will be significantly improved after the 2007–2009 financial crisis versus before the 2007–2009 financial crisis. hypothesis 2: the overall cash flow management behavior of all employees will be significantly improved after the 2007–2009 financial crisis versus before the 2007–2009 financial crisis. hypothesis 3: the overall savings behavior of low-income employees will be significantly improved after the 2007–2009 financial crisis versus before the 2007–2009 financial crisis. hypothesis 4: the overall cash flow management behavior of low-income employees will be significantly improved after the 2007–2009 financial crisis versus before the 2007–2009 financial crisis. 4. methodology 4.1. data and sample data from the 2004 and 2013 survey of consumer finances (scf) were used for this study. these two years were selected because they were definitely before and after the financial crisis. the scf, sponsored by the federal reserve bank, is a triennial crosssectional survey of u.s. families that is collected by the national opinion research center (norc), a research organization at the university of chicago. the scf includes descriptive information about a family such as location and household size, as well as financial information such as household income, assets, liabilities, expenses, and banking relationships. the scf also provides information on a household’s financial behaviors such as savings, spending and investing behavior. most information within scf represents the overall household. however, information that is pertinent to an individual, such as employment status, refers to the head of household. in a mixed sex marriage, the male represents the head of household; and in a same sex marriage the older individual represents the head of household (federal reserve, 2014). a subsample of respondents was created by segmenting data by employment status. among the 4,519 households within the 2004 scf, 3,259 (72%) were employees. among the 6,482 households within the 2013 scf, 3,987 (62%) were employees. in total, 7,622 employees were included in this study; however, 376 employees failed to report their income, and therefore, only 7,246 employees were utilized in the analysis. this larger employee sample was further segmented to create the low-income, middleincome, and high-income employee subsamples. these segments were based on income 23c. hudson et al. / financial services review 26 (2017) 19–36 and household size. although all employees and low-income employee segments were the focal groups, middle-income and high-income subsamples were created for further comparison. low-income employees are defined as employees with household incomes less than or equal to 200% of the u.s. poverty level. the u.s. poverty level is based on household size and income and the categorization of households in this research follows a similar design (hudson & palmer, 2014). middle-income employees are defined as employees with household incomes between 200 and 400% of the u.s. poverty level, and high-income employees are defined as employees with household incomes above 400% of the u.s. poverty level. the poverty levels of household size and income are listed in tables 1 and 2 for both 2004 and 2013 (u.s. department of health and human services, 2004, 2013). the overall employee sample had more male respondents than female respondents in both 2004 and 2013. when this employee sample was segmented by income, the highest proportion of female respondents was observed among low-income employees in both 2004 and 2013, while the highest proportion of males was observed among high-income employees for both years. furthermore, the employee sample primarily consisted of whites in both 2004 table 1 2013 income subsamples based on poverty levels household size low-income middle-income high-income less than 200% between 200 and 400% greater than 400% 1 $22,980 $22,980 $45,960 $45,960 2 $31,020 $31,020 $62,040 $62,040 3 $39,060 $39,060 $78,120 $78,120 4 $47,100 $47,100 $94,200 $94,200 5 $55,140 $55,140 $110,280 $110,280 6 $63,180 $63,180 $126,360 $126,360 7 $71,220 $71,220 $142,440 $142,440 8 $79,260 $79,260 $158,520 $158,520 note: u.s. 2013 poverty level. adapted from the 2013 human health services poverty guidelines, by the u.s. department of health and human services, 2013 (available at http://aspe.hhs.gov/poverty/10poverty .shtml/). table 2 2004 income subsamples based on poverty levels household size low-income middle-income high-income less than 200% between 200 and 400% greater than 400% 1 $18,620 $18,620 $37,240 $37,240 2 $24,980 $24,980 $49,960 $49,960 3 $31,340 $31,340 $62,680 $62,680 4 $37,700 $37,700 $75,400 $75,400 5 $44,060 $44,060 $88,129 $88,129 6 $50,420 $50,420 $100,840 $100,840 7 $56,780 $56,780 $113,560 $113,560 8 $63,140 $63,140 $126,280 $126,280 note: u.s. 2004 poverty level. adapted from the 2004 human health services poverty guidelines, by the u.s. department of health and human services, 2004 (available at http://aspe.hhs.gov/poverty/04poverty.shtml/). 24 c. hudson et al. / financial services review 26 (2017) 19–36 and 2013. thus, the highest proportion of white respondents was among high-income employees, while the highest proportion of hispanic and black respondents was among low-income employees. the overall employee sample had more married individuals in both 2004 and 2013. when segmented by income, the highest proportion of married individuals was observed among high-income employees for both years. furthermore, the employee sample primarily consisted of individuals who attained a high school degree or less for both 2004 and 2013, while the highest proportion of individuals with advanced levels of educational attainment was found among high-income employees for both 2004 and 2013 (see table 4). 4.2. financial behavior variables previous research has found that financial behaviors provide insight into an individual’s financial knowledge and financial literacy (hogarth et al., 2003). based on hogarth et al. (2003), this study used groupings of similar variables to develop composite financial behavior scores: (a) savings and (b) cash flow management. the savings category included three financial behavior questions: (a) “do you save on a regular basis?” (b) “do you have a savings account?” and (c) “do you have money automatically deducted directly into an account?” responses indicating positive financial behavior were coded as 1, and negative responses were coded as 0. an overall savings index was created from a summation of these three questions. table 3 contains a detailed list and measurements of these variables, as well as all possible responses. in the cash flow management category, three financial behavior questions were utilized in a similar fashion: (a) “do you have a checking account?” (b) “is your spending less than or equal to your income?” and (c) “do you pay your loans on time?” additionally, an overall cash flow management index was created from these three questions. three control variables, race, education and age were included in this study. race, a categorical variable, was included with possible responses of white (reference), black, hispanic, and other race. education, an ordinal variable, and age, a continuous variable, were also included in the regression analysis. again, table 3 contains a detailed list and measurements of these variables, as well as all possible responses. table 3 savings and cash flow management variables variables measurements cash flow management behaviors checking � 1 if reported having a checking account; 0 otherwise loans on time � 1 if reported paying loans ahead of time or on time; 0 otherwise spending � 1 if reported spending was less than or equal to income; 0 otherwise savings behaviors savings � 1 if reported having a savings account; 0 otherwise save � 1 if reported saving on a monthly basis; 0 otherwise autosave � 1 if reported money automatically deposited in an account; 0 otherwise 25c. hudson et al. / financial services review 26 (2017) 19–36 t ab le 4 so ci oe co no m ic ch ar ac te ri st ic s, em pl oy ee s st at is tic s (% of to ta l) , w ei gh te d l ow -i nc om e em pl oy ee s m id dl ein co m e em pl oy ee s h ig hin co m e em pl oy ee s a ll em pl oy ee s 20 04 (n � 62 9) 20 13 (n � 98 7) 20 04 (n � 78 9) 20 13 (n � 1, 00 2) 20 04 (n � 1, 84 1) 20 13 (n � 1, 99 8) 20 04 (n � 3, 25 9) 20 13 (n � 3, 98 7) se x m al e 60 .0 8% 62 .1 8% 75 .0 2% 77 .6 4% 88 .5 6% 87 .5 0% 79 .7 9% 78 .7 5% fe m al e 39 .9 2% 37 .8 2% 24 .9 8% 22 .3 6% 11 .4 4% 12 .5 0% 20 .2 1% 21 .2 5% r ac e/ et hn ic ity w hi te 57 .0 9% 55 .5 8% 72 .9 6% 68 .3 8% 82 .3 8% 80 .6 9% 75 .2 2% 71 .3 8% b la ck 17 .1 5% 20 .0 9% 14 .2 3% 14 .2 8% 8. 22 % 6. 92 % 11 .4 0% 12 .0 3% h is pa ni c 22 .3 3% 21 .4 3% 8. 81 % 13 .5 3% 4. 62 % 5. 07 % 9. 05 % 11 .2 5% o th er 3. 43 % 2. 89 % 4. 00 % 3. 81 % 4. 78 % 7. 32 % 4. 33 % 5. 34 % m ar ita l st at us m ar ri ed 32 .4 9% 35 .9 0% 52 .9 5% 53 .0 6% 69 .6 9% 66 .4 4% 58 .4 6% 55 .5 2% si ng le 67 .5 1% 64 .1 0% 47 .0 5% 46 .9 4% 30 .3 1% 33 .5 6% 41 .5 4% 44 .4 8% e du ca tio n l es s th an hi gh sc ho ol 80 .5 0% 77 .6 2% 68 .6 1% 60 .1 7% 37 .4 8% 32 .8 5% 53 .3 2% 50 .8 0% a ss oc ia te de gr ee 5. 28 % 7. 37 % 7. 34 % 9. 29 % 7. 73 % 6. 60 % 7. 16 % 7. 47 % b ac he lo r de gr ee 9. 79 % 11 .6 7% 17 .5 4% 19 .7 6% 29 .9 8% 33 .0 8% 23 .0 7% 24 .4 3% g ra du at e de gr ee 4. 02 % 3. 28 % 6. 51 % 10 .4 8% 24 .4 1% 27 .1 0% 16 .1 4% 17 .0 3% c er tifi ca te 0. 40 % 0. 07 % 0. 00 % 0. 31 % 0. 39 % 0. 36 % 0. 30 % 0. 28 % a ge 18 –3 0 31 .2 4% 25 .1 7% 20 .9 0% 16 .6 1% 9. 59 % 10 .5 0% 16 .5 1% 15 .6 7% 31 –5 0 44 .9 0% 43 .8 5% 57 .2 3% 53 .8 7% 54 .7 7% 48 .7 6% 53 .4 6% 48 .8 3% 51 � 23 .8 6% 30 .9 8% 21 .8 7% 29 .5 3% 35 .6 4% 40 .7 3% 30 .0 3% 35 .5 0% 26 c. hudson et al. / financial services review 26 (2017) 19–36 4.3. statistical analysis descriptive statistics were generated to describe the demographic and socioeconomic characteristics of all employees, and the low-income, middle-income, and high-income employee subsamples. the distribution of all employees, and the low-income, middleincome, and high-income employees that responded affirmatively or negatively to each financial behavior was determined through a frequency analysis. an overall savings percentage and overall cash flow management percentage was created by combining the total affirmative responses. an ordered logistic regression model was used to analyze data to determine if there was a significant change in savings and cash flow management behavior from 2004 to 2013. data from the 2004 and 2013 scf were combined, and the overall savings index and cash flow management index were calculated for all of the sample respondents. the primary independent variable was a binary variable used to indicate whether the savings and cash flow management indices were being measured before or following the financial crisis. other control variables included race, education, and age. ordered logistic regression models were used to evaluate whether there was a significant difference in all employees’ and low-income employees’ savings behavior and cash flow management behavior. these two samples were the focal groups of this study, and their regression models were used to prove or disprove the study’s hypotheses. moreover, middle-income and high-income employees’ savings and cash flow management behaviors were evaluated in the same manner as a method of comparison to low-income employees. the ordered logistic regression model was utilized to analyze the relationship of a set of independent variables and an ordinal dependent variable. in this study, the ordinal dependent variables were the savings behavior index and cash flow management behavior index, with four possible ordered responses. the ordered logistic regression model is based on the assumption that outcome i corresponds to the probability that the estimated linear function is within the range of the cutoff points estimated by the outcome. the probability odds models are as follows (snedker, glynn, & wang, 2002): logit(p1) � log p1 1 � p1 � � 1 � bx1 (1) logit(p1 � p2) � log p1 � p2 1 � p1 � p2 � � 1 � b�x1 (2) logit(p1 � p2 � � � � pk) � log p1 � p2 � � � � pk 1 � p1 � p2 � � � � pk � �1 � b�x1 (3) coefficients and odds ratios were generated for each independent variable to estimate the maximum likelihood of this independent variable (snedker et al., 2002). if a significant positive relationship exists between the dependent variable of low-income employees’ savings behavior index and the independent binary variable of the behavior occurring before 27c. hudson et al. / financial services review 26 (2017) 19–36 the financial crisis or after the financial crisis, it would be interpreted as “low-income employees were x% more likely to report better savings behavior after the financial crisis than before the financial crisis.” on the other hand, if a significant negative relationship exists between the same dependent and independent variables, it would be interpreted as “low-income employees were less likely to report better savings behavior after the financial crisis than before the crisis.” the rii technique was utilized to avoid analysis errors that may occur when using scf data. the scf contains five implicates, which are essentially five duplicate sets of data with substituted estimates for missing data (montalto & sung, 1996). this study’s analysis accounted for the scf complex sampling design (i.e., dual-frame complex sampling) and multiple imputation methodology (federal reserve, 2014). 5. results 5.1. financial behaviors comparison results preliminary results from the financial behavior frequency analysis found that the mean number of affirmative savings behaviors low-income employees engaged in during 2013 was significantly greater than the mean number of affirmative savings behaviors lowincome employees engaged in during 2004. as shown in table 5, low-income employees’ mean reported affirmative savings behaviors was 1.72 for 2013 and 1.49 for 2004. within the individual savings behaviors, the percentage of low-income employees who had a savings account, as well as those who reported having money automatically deposited into an account, appeared to be greater in 2013 (see table 6). there were also significant differences in the mean number of affirmative savings behaviors between 2004 and 2013 for middle-income employees and high-income employees, with both groups showing improved savings behaviors (see table 5). the proportion of individuals in both groups that had a savings account and had money automatically deposited into a savings account increased from pre-financial crisis to post-financial crisis (see table 6). overall, the mean number of affirmative cash flow management behavior among low-income employees and middle-income employees increased from 2004 to 2013, while over the same table 5 comparison of financial behavior frequency results, weighted behavior variables low-income middle-income high-income 2004 2013 p 2004 2013 p 2004 2013 p mean of cash flow 1.99 2.08 �0.05a 2.37 2.48 �0.001a 2.69 2.70 �0.05a mean of savings 1.49 1.72 �0.001a 2.01 2.25 �0.001a 2.46 2.57 �0.001a note: samples are 629 (low-income in 2004), 987 (low-income in 2013), 789 (middle-income in 2004), 1,002 (middle-income in 2013), 1,841 (high-income in 2004), and 1,998 (high-income in 2013). afor mean comparisons (i.e., t test), alternative hypotheses are: mean value in 2013 is greater than mean value in 2004 (i.e., one-tail comparison). 28 c. hudson et al. / financial services review 26 (2017) 19–36 period, the mean number of affirmative cash flow management behavior among high-income employees was unchanged (see table 5). as shown in table 6, the percentage of low-income and middle-income employees with checking accounts increased from 2004 to 2013, and for middleincome employees, in particular, a higher proportion in 2013 appeared to be spending less than or equal to their total income. among high-income employees, there was relatively little change in the mean number of affirmative cash-flow management behaviors. again, these results were simply preliminary results. 5.1. ordered logistic regression results results from the ordered logistic regression analysis indicated that, following the crisis, employees were 53% more likely to report higher savings behaviors than before the financial table 6 financial behavior frequency results, weighted behavior variables low-income middle-income high-income 2004 (n � 629) 2013 (n � 987) 2004 (n � 789) 2013 (n � 1,002) 2004 (n � 1,841) 2013 (n � 1,998) cash flow management do you have a checking account? yes 76.59% 83.48% 91.68% 96.26% 99.14% 99.77% no 23.41% 16.52% 8.32% 3.74% 0.86% 0.23% is your income equal to or less than your spending? yes 75.32% 76.65% 77.08% 85.56% 88.78% 91.17% no 24.68% 23.35% 22.92% 14.44% 11.22% 8.83% do you pay your loans on tine? yes 47.74% 47.68% 65.36% 66.62% 81.87% 79.19% no 52.26% 52.32% 34.64% 33.38% 18.13% 20.81% total cash flow management none 4.72% 3.38% 2.10% 0.39% 0.15% 0.01% one 23.79% 18.40% 13.87% 8.13% 3.85% 2.66% two 38.59% 45.27% 31.84% 34.14% 22.06% 24.52% three 32.89% 32.96% 52.19% 57.34% 73.94% 72.81% savings do you save on a regular basis? yes 65.40% 67.74% 76.72% 80.26% 93.06% 91.23% no 34.60% 32.26% 23.28% 19.74% 6.94% 8.77% do you have a savings account? yes 35.42% 39.31% 56.81% 59.86% 69.75% 73.15% no 64.58% 60.69% 43.19% 40.14% 30.25% 26.85% do you have money automatically deposited? yes 47.79% 65.13% 66.77% 85.02% 83.67% 92.19% no 52.21% 34.87% 33.23% 14.98% 16.33% 7.81% total savings none 17.46% 9.39% 6.89% 3.64% 0.84% 0.74% one 33.80% 32.13% 20.44% 15.44% 9.99% 7.05% two 31.41% 35.40% 38.15% 33.07% 31.02% 27.11% three 17.33% 23.08% 34.52% 47.85% 58.15% 65.10% 29c. hudson et al. / financial services review 26 (2017) 19–36 crisis, as shown in table 7. higher income households were more likely to exhibit higher overall savings behaviors. if households were to move from the low-income category to the middle-income category, they would be 126% more likely to exhibit higher overall savings behaviors. hispanics were 48% less likely relative to whites to report higher overall savings behaviors. those with more education were also more likely to have higher overall savings behaviors. when the ordered logit analysis was limited to specifically low-income employees, overall savings behavior followed similar patterns to that of the full sample. low-income employees were 55% more likely to report higher savings behavior following the crisis (see table 7). hispanics were 55% less likely to report higher savings behavior relative to whites, and those with higher educational attainment were 44% more likely to have higher savings behavior among low-income employees. the savings behavior of middle-income employees following the financial crisis was significantly improved and indicated a strong positive association with education. middle-income employees were more likely than lowand high-income employees to have higher savings behavior following the crisis. hispanics were 40% less likely to report higher savings behavior than whites among middle-income employees from 2004 to 2013. high-income employees were also more likely to have higher savings behavior following the financial crisis. among high-income employees, a racial gap was observed. blacks were 30% less likely than whites to have higher savings behavior. the racial gap between blacks and whites was not observed among lowand middle-income employees; rather, an ethnicity gap between hispanics and whites was observed among lowand middle-income employees. similar to savings behavior, all employees were 15% more likely to report better cash flow management behavior following the financial crisis (see table 8). increased income was associated with a higher likelihood of increased cash flow management behavior. education and age were also positively associated with increased cash flow management behavior in 2013 when compared with 2004. minorities were less likely to have higher cash flow management behavior in 2013 compared with whites. in contrast to the total sample, low-income employees’ cash flow management behavior did not change following the crisis when other factors were controlled for, as shown in table 8. similar to the model for all employees, low-income employees with higher education and who were older were 35 and 2% more likely to report higher cash flow management behavior for each incremental increase in education and age, respectively. racial and ethnic differences were observed across all income categories. middle-income employees were 31% more likely to have higher cash flow management practices following the financial crisis. middle-income employees were the only subgroup of employees more likely to have higher savings behavior and cash flow management behavior following the financial crisis. among high-income employees, overall cash flow management behavior was not different before or after the crisis. however, among high-income employees, blacks and hispanics were less likely to have high cash flow management behavior relative to white households. 30 c. hudson et al. / financial services review 26 (2017) 19–36 t ab le 7 sa vi ng be ha vi or ch an ge be tw ee n 20 04 an d 20 13 t ot al sa m pl ea h ig hin co m eb m id dl ein co m ec l ow -i nc om ed n � 7, 24 6 n � 3, 83 9 n � 1, 79 1 n � 1, 61 6 b o r b o r b o r b o r po st -c ri si s (a ft er cr is is � 1) 0. 42 ** * 1. 53 0. 31 ** * 1. 36 0. 52 ** * 1. 69 0. 44 ** * 1. 55 in co m e le ve l (l ow � 1, hi gh � 3) 0. 82 ** * 2. 26 — — — — — — r ac e (w hi te � r ef .) — — — — — — — — b la ck � .1 6* 0. 85 � 0. 36 ** 0. 70 � 0. 11 0. 89 � 0. 10 0. 91 h is pa ni c � .6 6* ** 0. 52 � 0. 30 0. 74 � 0. 51 ** 0. 60 � .8 0* ** 0. 45 o th er ra ce � .1 5 0. 86 0. 00 1. 00 � 0. 08 0. 92 � 0. 56 0. 57 e du ca tio n .2 2* ** 1. 24 0. 17 ** * 1. 18 0. 23 ** * 1. 26 .3 6* ** 1. 44 a ge � .0 1 0. 99 � 0. 01 ** * 0. 99 0. 00 1. 00 0. 01 1. 01 in te rc ep t 1 1. 13 ** * 0. 32 � 4. 86 ** * 0. 01 � 2. 84 ** * 0. 06 � 1. 40 ** * 0. 25 in te rc ep t 2 0. 71 ** * 2. 03 � 2. 42 ** * 0. 09 � 1. 07 ** * 0. 34 0. 43 ** 1. 54 in te rc ep t 3 2. 37 ** * 10 .7 2 � 0. 58 ** * 0. 56 0. 53 ** 1. 71 2. 05 ** * 7. 77 n ot e: b � co ef fic ie nt of va ri ab le s; o r � od ds ra tio . a fo r to ta l sa m pl e m od el , � 2 � 15 26 .3 0* ** ; ps eu do r 2 � 0. 09 . b fo r hi gh -i nc om e m od el , � 2 � 93 .1 5* ** ; ps eu do r 2 � 0. 01 . c fo r m id dl ein co m e m od el , � 2 � 91 .0 2* ** ; ps eu do r 2 � 0. 02 . d fo r lo w -i nc om e m od el , � 2 � 16 6. 82 ** *; ps eu do r 2 � 0. 04 . *p � .0 5; ** p � .0 1; ** * p � .0 01 . 31c. hudson et al. / financial services review 26 (2017) 19–36 t ab le 8 c as h flo w m an ag em en t be ha vi or ch an ge be tw ee n 20 04 an d 20 13 t ot al sa m pl ea h ig hin co m eb m id dl ein co m ec l ow -i nc om ed n � 7, 24 6 n � 3, 83 9 n � 1, 79 1 n � 1, 61 6 b o r b o r b o r b o r po st -c ri si s (a ft er cr is is � 1) 0. 14 ** 1. 15 0. 00 1. 00 0. 27 ** 1. 31 0. 14 1. 55 in co m e le ve l (l ow � 1, hi gh � 3) 0. 72 ** * 2. 05 — — — — — — r ac e (w hi te � r ef .) — — — — — — — — b la ck � 0. 82 ** * 0. 44 � 0. 56 ** * 0. 57 � 0. 89 ** * 0. 41 � 0. 87 ** * 0. 42 h is pa ni c � 0. 54 ** * 0. 58 � 0. 49 * 0. 62 � 0. 58 ** * 0. 56 � 0. 48 ** * 0. 62 o th er r ac e � 0. 23 * 0. 79 � 0. 22 0. 81 � 0. 36 * 0. 70 � 0. 13 0. 88 e du ca tio n 0. 16 ** * 1. 17 0. 08 ** 1. 08 0. 22 ** * 1. 25 0. 30 ** * 1. 35 a ge 0. 01 ** * 1. 01 � 0. 01 * 0. 99 0. 01 ** * 1. 01 0. 02 ** * 1. 02 in te rc ep t 1 � 2. 59 ** * 0. 08 � 7. 26 ** * 0. 00 � 4. 06 ** * 0. 02 � 2. 72 ** * 0. 07 in te rc ep t 2 � 0. 30 ** 0. 74 � 3. 66 ** * 0. 03 � 1. 59 ** 0. 20 � 0. 58 ** 0. 56 in te rc ep t 3 1. 66 ** * 1. 66 � 1. 31 ** * 0. 27 0. 28 1. 32 1. 35 ** * 3. 85 n ot e: b � co ef fic ie nt of va ri ab le s; o r � od ds ra tio . a fo r to ta l sa m pl e m od el , � 2 � 12 25 .4 8* ** ; ps eu do r 2 � . 00 9. b fo r hi gh -i nc om e m od el , � 2 � 30 .9 6* ** ; ps eu do r 2 � 0. 01 . c fo r m id dl ein co m e m od el , � 2 � 10 9. 66 ** *; ps eu do r 2 � 0. 03 . d fo r lo w -i nc om e m od el , � 2 � 13 6. 50 ** *; ps eu do r 2 � 0. 04 . *p � .0 5; ** p � .0 1; ** *p � .0 01 . 32 c. hudson et al. / financial services review 26 (2017) 19–36 6. discussion and implications this study found that the overall savings behavior of all employees improved following the 2007–2009 financial crisis relative to pre-financial crisis practices and persisted for several years following the crisis. even among low-income working households, savings practices are flexible and can be altered. while the financial crisis was an extreme external event, the findings of this study suggest that low-income employees’ savings behaviors are malleable despite limited income. the savings behavior findings of all employees in this study were consistent with results from previous studies that also examined the effects of a financial crisis on savings behaviors of the general population (fidelity investments, 2013; o’neill & xiao, 2012). in particular, o’neill and xiao’s (2012) study utilized a similar methodology of this study to examine preand post-financial behaviors. similar to this study, o’neill and xiao (2012) found that the participants’ savings behavior was better after the 2007–2009 financial crisis. likewise, the fidelity (2013) study saw an increase in participants’ emergency funds and retirement savings after the 2007–2009 financial crisis as opposed to before the crisis. moreover, the cash flow management behaviors of employees, in general, improved after the 2007–2009 financial crisis. however, when the sample was segmented by income, only middle-income employees’ cash flow management behavior, as a group, was more likely to improve following the financial crisis, while the cash flow management behaviors of low-income and high-income employees were unchanged after the financial crisis. similarly, in the o’neill and xiao (2012) study, participants’ budgeting and spending behaviors, which is typically considered cash flow management behaviors, improved after the 2007–2009 financial crisis. moreover, in the fidelity (2013) study, participants decreased their debt after the 2007–2009 crisis, which would increase participants’ cash flow. this study has very practical implications for employers and financial planners who advise on retirement plans and who are looking for ways to increase employee participation and contribution rates in their retirement plans. all employees, regardless of income level, can adjust their savings behavior. employers and retirement plan advisers, who conduct new-hire enrollment seminars or ongoing outreach to plan participants, should focus on motivating increased savings since savings behavior appears to be flexible. when thinking about motivating increased savings in an employer-sponsored retirement plan, less emphasis should be placed on cash flow management behaviors such as banking relationships, debt management, and spending behaviors, since these behaviors seem less responsive to change. savings behavior appears to be more responsive to change; therefore, employers and retirement plan sponsors should target it directly. strategies that incorporate processes of the transtheoretical model of change will provide low-income employees greater intrinsic motivation to change their savings behavior. financial planners may find that consciousness raising activities with a focus on fostering an emotional response (dramatic relief) from potential participants are important first steps. given the lasting change observed in this study, one way financial advisors could trigger the desired emotional response is to appropriately rekindle emotions connected with the financial crisis. plan sponsors and advisors should also consider providing opportunities for lowincome employees to self-identify as savers (self-reevaluation) and then also provide strong 33c. hudson et al. / financial services review 26 (2017) 19–36 public reinforcement of positive savings decisions. these strategies coupled with easy enrollment procedures, matching contributions, and simplified investment options may be effective strategies in inducing savings behavior change among employees, particularly among low-income employees. 7. limitations and future research researchers observed several limitations while conducting this study. while most questions used to measure savings and cash flow management behaviors were adopted from hogarth et al. (2003), there could be some overlap in the participants’ interpretation of these questions. an example of this overlap may exist between the two questions; “do you save on a regular basis?” and “do you have money automatically deposited into an account?” furthermore, scf has gradually added financial behaviors over the years, but the number and the variety of financial behavior questions within scf are still limited. additionally, limitations in how cash flow management and savings behavior were conceptualized and modeled in this study, applying variable constructs from previous studies, could have affected observed relationships in the models. future research will again focus on employees’ financial behavior; however, the behavior categories would expand to investment behaviors as well as retirement and retirement planning behaviors. also, segmenting the overall employee sample by classes such as blue collar workers and white collar workers and then examining their financial behaviors might produce interesting results. finally, instead of examining unintended factors, future research would examine the effects of intended factors such as financial education, automatic deposit into an account and automatic 401k enrollment on behavior. references asaad, c. t. (2014). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101–117. andersen, s. g., zhan, m., & scott, j. (2004). targeting financial management training at low-income audiences. journal of consumer affairs, 38, 167–177 (available at http://onlinelibrary.wiley.com/doi/10.1111/j.17456606.2004.tb00470.x/abstract). braunstein, s., & welch, c. (2002). financial literacy: an overview of practice, research and theory. federal reserve bulletin, 88, 445–457. bricker, j., bucks, b., kennickell, a. b., mach, l. t., & moore, k. (2011). surveying the aftermath of the storm: changes in the family finances from 2007–2009. working paper 2011–17. washington, dc: finance and economics discussion series, federal reserve board (available at http://www.federalreserve.gov/pubs/feds/ 2011/201117/201117pap.pdf). caplan, g. (1971). an approach to community mental health. new york, ny: grune & stratton. copeland, a. (2014). employer based retirement plan participation: geographic differences and trends. issue brief # 405. washington, dc: employee benefit research institute (available at https://www.ebri.org/ publications/ib/index.cfm?fa�ibdisp&content_id�5451). federal reserve. (2014). codebook for 2013 survey of consumer finances. washington, dc: board of governors of the federal reserve system (available at https://www.federalreserve.gov/econresdata/scf/files/ codebk2013.txt). 34 c. hudson et al. / financial services review 26 (2017) 19–36 fidelity investments. (2013). fidelity study finds financial crisis was wake-up call for investors, triggering positive and permanent behaviors. smithfield, ri: fidelity investment services (available at https://www.fidelity.com/about-fidelity/ individual-investing/fidelity-study-finds-financial-crisis-was-wake-up-call-for-investors). fisher, p. j., hayhoe, c. r., & lown, j. m. (2015). gender differences in savings behaviors among lowto moderate income households. financial services review, 24, 1–13. hogarth, j. m., beverly, s. g., & hilgert, m. (2003). patterns of financial behaviors: implications for community educators and policy makers. paper presented at the 2003 federal reserve systems community affairs research conference, washington, dc (available at http://www.federalreserve.gov/communityaffairs/ national/ca_conf_suscommdev/pdf/hogarthjeanne.pdf). hudson, c. r., & palmer, l. (2014). low-income employees: the relationship between information from formal advisors and financial behaviors. financial services review, 23, 25–43. loibl, c., & hira, k. t. (2005). self-directed financial learning and financial satisfaction. journal of financial counseling and planning, 16, 11–22. loibl, c., & hira, k. t. (2006). a workplace and gender related perspective on financial planning information sources and knowledge outcomes. financial services review, 15, 21–42. lusardi, a., & mitchell, o. (2005). financial literacy and planning: implications for retirement well-being (wp 2005–108). ann arbor, mi: university of michigan retirement research center (available at http:// www.mrrc.isr.umich.edu/publications/papers/pdf/wp108.pdf). lusardi, a., & mitchell, o. (2011). financial literacy and retirement planning in the u.s. (wp 107/11). ann arbor, mi: center for research and pensions and welfare policies (available at https://core.ac.uk/download/ pdf/7189099.pdf). lyons, a., & scherpf, e. m. (2004). moving from unbanked to banked: evidence from the money smart program. financial services review, 13, 215–231. lyons, a., chang, y., & scherpf, e. m. (2006). translating financial education into behavior change for low-income populations. journal of financial counseling and planning, 17, 27–45. metlife. (2009). the 2009 met life study of the american dream. new york, ny: metropolitan life insurance corporation (available at http://www.metlife.com/assets//cao/gbms/studies/09010229_ 09amdreamstudy_web.pdf). montalto, c. p., & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances. journal of financial counseling and planning, 7, 133–146. o’neill, b., & xiao, j. (2012). financial behavior before and after a financial crisis: evidence from an online survey. journal of financial counseling and planning education, 23, 33–46 (available at http://afcpe.org/ assets/pdf/v23_j3.pdf). palmer, l., bliss, d. l., goetz, j. w., & moorman, d. (2010a). helping undergraduates discover the value of a dollar through self-monitoring. american journal of business education, 3, 103–108. palmer, l., bliss, d. l., goetz, j. w., & moorman, d. (2010b). improving financial awareness among college students: assessment of a financial management project. college student journal, 44, 659–676. prochaska, j. o., diclemente, c. c., & norcross, j. c. (1992). in search of how people change: applications to addictive behaviors. american psychologist, 47, 1102–1114. rich, r. (2013). the great recession of 2007–09. washington, dc: federal reserve system (available at http://www.federalreservehistory.org/events/detailview/58). slaikeu, k. a. (1990). crisis intervention: a handbook for practice and research (2nd ed.). boston, ma: a pearson education company. snedker, k., glynn, p., & wang, c. (2002). ordered/ordinal logistic regression with sas and stata. seattle, wa: university of washington (available at http://staff.washington.edu/glynn/olr.pdf). taylor, p., morin, r., kochhar, r., parker, k., cohn, d., lopez, m. h., fry, r., wang, w., velasco, g., dockterman, d., hinze-pifer, r., & espinoza, s. (2010). a balanced sheet at 30 months: how the recession has changed life in america. washington, dc: the pew research center, social and demographic trends project (available at http://www.pewsocialtrends.org/files/2010/10/759-recession .pdf). 35c. hudson et al. / financial services review 26 (2017) 19–36 u.s. department of health and human services. (2004). the 2004 human health services poverty guidelines. washington, dc: office of the assistant secretary for planning and evaluation (available at https://aspe.hhs.gov/2004-hhs-poverty-guidelines). u.s. department of health and human services. (2013). the 2013 human health services poverty guidelines. washington, dc: office of the assistant secretary for planning and evaluation (available at https://aspe.hhs.gov/2013-poverty-guidelines). volpe, r.p., chen, h., & liu, s. (2006). an analysis of the importance of personal finance topics and the level of knowledge possessed by working adults. financial services review, 15, 81–98. woolley, n. (1990). crisis theory: a paradigm of effective intervention with families of critically ill people. journal of advanced nursing, 15, 1402–1408. xiao, j. j., newman, b. m., prochaska, j. m., leon, b., bassett, r. l., & johnson, j. l. (2004). applying the transtheoretical model of change to consumer debt behavior. financial counseling and planning education, 1, 89–100. 36 c. hudson et al. / financial services review 26 (2017) 19–36 finser_23_1 investor preference for skewness and the incubation of mutual funds philip gibsona,*, michael finkeb adepartment of accounting, finance and economics, winthrop university, rock hill, sc 29733, usa bdepartment of personal financial planning, texas tech, lubbock, tx 79401, usa abstract mutual fund companies market the strong performance of funds created through incubation to gain the attention of investors who value recent returns. this creates an incentive for fund families to select highly skewed securities because extreme performance during incubation will increase the likelihood that some funds will outperform before they are sold to the public. although incubation is as an innovative fund promotion technique, it may harm investors by creating the perception that random prior returns are a signal of fund quality. we find that net new money flow increases with an incubated fund’s skewness. after incubated funds are sold to the public, skewed funds attract more investor dollars and their average performance declines. these results suggest that the use of skewed securities during incubation is an effective method for increasing demand, but may be a poor quality signal of future performance. © 2014 academy of financial services. all rights reserved. jel classification: g23; g11 keywords: mutual funds; incubation; skewness; expense ratio; performance 1. introduction mutual fund companies use a variety of tactics to gain the attention of investors. given that most investors have no formal training in what factors to assess when selecting a fund, they must look for easily understood cues of product quality (barber and odean, 2008). * corresponding author. tel.: !1-803–323-2186; fax: !1-803-323-3960 e-mail address: pgibson@uiwtx.edu financial services review 23 (2014) 63–75 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. faced with high search costs, many investors simply select funds that have high recent performance (sirri and tufano, 1998). mutual fund managers have an incentive to capture investor attention by creating funds that have high recent returns and marketing the performance of these funds to consumers. mutual fund incubation is a tactic that some fund companies use to create new fund offerings. in incubation, families develop numerous new funds often with a limited amount of seed money. after a period of time, funds with a good performance track record are opened to the public, whereas those that underperform are liquidated before investors ever become aware of their existence (evans, 2010). because the highest performing funds are likely to capture the greatest investor attention, there is an incentive for fund managers to select securities with more highly skewed returns that are most likely to achieve well above average performance. both incubation and skewness of mutual funds are relatively unexplored areas within the financial literature. the purpose of this study is to investigate whether fund families have an incentive to initiate highly skewed incubated funds. we estimate net investor cash flow to incubated mutual funds that are launched with a positively skewed return at inception and compare characteristics and performance of incubated to non-incubated funds. the results of this study reveal that incubated mutual funds on average have a higher expense ratio than non-incubated funds. incubated funds have significantly lower performance after they are opened to investors. we find evidence that fund flows are higher among incubated funds that are launched with a more positively skewed return at inception. our results suggest that fund managers have a strong incentive to select highly skewed securities when incubating funds, and that the process of culling low performance before opening funds to the public can create a powerful, but ultimately false, signal of quality. 2. review of literature 2.1. mutual fund incubation incubation is the process where a mutual fund company creates several mutual funds (incubator funds) seeded with their own resources and operated in private for a specified period of time (palmiter and taha, 2009). this process can either be done privately or publicly. before a mutual fund company can market a new fund, it must first register the fund with the securities exchange commission (sec). if a mutual fund company decides to wait, and then register the incubator fund right before it becomes publicly available, it is referred to as “private incubation.” on the other hand, “public incubation” occurs when the mutual fund company registers the incubator fund with the sec when it is created, but does not actively market the performance of this fund until it is known (evans, 2010). those incubator funds that are unsuccessful in terms of realized returns are often eliminated and never publicized. successful funds are then marketed to prospective investors. much of the literature on mutual fund incubation indicates that once an incubated fund becomes publicly available, it underperforms the market. for instance, garavito (2008) shows that incubated domestic equity funds outperform non-incubated domestic equity funds on a risk-adjusted basis for the initial three years of existence. however, after three years the 64 p. gibson, m. finke / financial services review 23 (2014) 63–75 incubated funds no longer outperform. likewise, ackerman and loughran (2006) find that the average incubator fund outperformed the market by 358 basis points during incubation but underperformed the market by 423 basis points once it became publicly available. similarly, evans (2010) finds a negative relation between fund returns during the incubation stage and subsequent returns. deciding when to launch a fund is an important aspect of the incubation process. to attract investors, mutual fund companies must launch their incubated funds once they have achieved returns for a particular period of time. garavito (2008) finds that mutual fund companies tend to launch incubated funds when their returns are above the industry median. the incubation process can last from a few months to several years. ultimately, the goal is to launch the funds when returns are high enough to capitalize on the return chasing behavior of investors. by creating several funds, randomness, luck, or both will lead to one or few funds posting superior returns. by hiding the funds that did not perform well, investors are allowed to believe that the mutual fund manager was indeed able to identify underpriced securities in the market. 2.2. skewness despite the empirical evidence showing that investors are better off investing in passively managed funds (carhart, 1997; jensen, 1968), actively managed funds continue to thrive and attract consumers. the demand investors have for actively managed mutual funds may be attributed to a preference for skewness, or the increased likelihood of extreme returns. investor preference for skewness is well documented within the financial literature. in his seminal article, arditti (1967) demonstrates that investors prefer positive skewness in the return distribution. he maintains that an investor who has a decreasing absolute risk aversion forgoes an expected portfolio return to benefit from skewness. similarly, harvey and siddique (2000) advocate that investors should prefer portfolios that are right-skewed instead of those that are left-skewed. thus, assets that decrease a portfolio’s skewness are less desirable and should require higher expected returns. kraus and litzenberger (1976) discover that investors are averse to variance, however, they have a preference for skewness. likewise, kumar (2009) finds that most individual investors demand lottery type stocks during poor economic periods and invest disproportionately more in stocks that display a higher skewness. barberis and huang (2007) imply that some investors prefer skewness because it enables them to have a more lottery-like wealth distribution. investor preference for skewness can be better understood with the aid of cumulative prospect theory (cpt). under cpt (tversky and kahneman, 1992) investors have a tendency to overweight small probabilities making highly skewed instruments more attractive. by overweighting the tails of a distribution, a mutual fund that exhibits positively skewed performance might appear to be more desirable. as a result, their returns become more lottery-like, gaining investors attention because they offer a small probability of winning with an especially high reward. incubated mutual funds with a track record of high returns create the illusion that the fund may produce high positive returns in the future. this perception of lottery-like characteristics can make a mutual fund more attractive and provide 65p. gibson, m. finke / financial services review 23 (2014) 63–75 an incentive for fund families to select highly skewed securities during the incubation process. 3. hypothesis the incubation process allows a mutual fund company to select high performing funds to promote to consumers. they have an incentive to select securities within these funds that exhibit positively skewed returns to attract investors seeking high recent returns. given the non-linear relationship between net investor cash flow and performance, companies gain more from promoting funds with greater skewness. we hypothesize that at inception incubated funds that are most skewed will receive a higher inflow of cash relative to other funds. 4. data the source of the equity mutual fund data comes directly from morningstar direct and the national association of securities dealer (nasd) ticker creation date data.1 morningstar direct reports historical net asset values, cash flow, expense ratios, and return data for live and defunct mutual funds. a sample of equity funds from the united states is collected. following evans (2010), only funds that have an inception date greater than or equal to january 1, 1996 is included in the sample. this allows funds at the beginning of the sample to be incubated for a minimum of three years since the ticker creation date data begins in january 1999. to determine which mutual funds were incubated we merge the data from morningstar direct with the nasd creation date data by the ticker assigned to each share class. the ticker creation date is the date a ticker was assigned to a particular fund. excluded from the sample are indexed funds, foreign mutual funds, sector funds, closed-end funds, specialized funds, institutional funds, and funds with less than 30 months of return data. this brings the final sample of funds to 1,698. the sample is free of survivorship bias as it includes both funds that are extinct and funds that are currently active. given that different share classes of a fund have claims to the same underlying portfolio and they do not differ in trades or investment holdings, we combine monthly total net assets across all shares for each fund, and the mutual funds returns and expenditure are then weighted accordingly. 5. summary statistics to determine whether a mutual fund was incubated, we observe the difference between the ticker creation date and the inception date of the fund. if the difference is positive it indicates a deferral between the start of the fund and the application for and the approval of a ticker for the fund. to get a clear distinction, if there is a difference of 12 months or more between the ticker creation date and the mutual fund’s inception date the fund is classified as being incubated. comparable to evans (2010), we find that roughly 22.91% of the sample 66 p. gibson, m. finke / financial services review 23 (2014) 63–75 is incubated. presented in tables 1 and 2 are descriptive statistics showing some of the similarities and the differences between incubated and non-incubated funds sorted according to morningstar direct global category. 5.1. expense ratio t tests table 3 contains the result of a two sample t test in which the expense ratio, total assets under management, and manager tenure are compared for incubated and non-incubated mutual funds. according to morningstar direct, a fund’s expense ratio is the percentage of fund assets used to pay for operating expenses and management fees, including 12b-1 fees, administrative fees, and all other asset-based costs incurred by the fund, except brokerage costs. in examining the difference in mutual fund expenditure, as expected incubated funds on average have a statistically higher expense ratio than non-incubated funds. prior studies indicate that mutual fund expense ratios have a negative relation with performance (carhart, 1997; gruber, 1996; jensen, 1968). consequently, mutual fund investors who purchase higher expense mutual funds end up losing as a portion of their returns goes to the mutual fund company to cover their expenses. furthermore, investors are often unaware of the negative effect that high expenses will have on their returns (alexander, jones, and nigro table 2 summary statistics for non-incubated funds sorted by morningstar direct global category morningstar global category no. of observation (freq) expense ratio (%) total net assets ($ millions) manager tenure u.s. large cap blend 316 (24.14%) 1.45 (0.57) 140.25 (368.02) 5.58 (3.40) u.s. large cap growth 320 (24.44%) 1.70 (0.51) 121.98 (287.20) 5.43 (3.78) u.s. large cap value 179 (13.67%) 1.72 (0.42) 82.40 (158.64) 6.21 (3.75) u.s. equity mid cap 269 (19.78%) 1.72 (0.52) 147.63 (371.15) 6.12 (3.73) u.s. equity small cap 225 (17.18%) 1.84 (0.54) 87.62 (220.22) 5.83 (3.64) all 1309 1.67 (0.54) 120.34 (305.49) 5.78 (3.66) note. table 2 presents the aggregate summary statistics for non-incubated mutual funds. means (sd) are presented for expense ratio, total net assets, and manager tenure. additionally, presented is the number of funds along with their relative frequencies. table 1 summary statistics for incubated funds sorted by morningstar direct global category morningstar global category no. of observation (freq) expense ratio (%) total net assets ($ millions) manager tenure u.s. large cap blend 91 (23.39%) 1.53 (0.65) 88.52 (197.62) 6.50 (4.90) u.s. large cap growth 82 (21.10%) 1.73 (0.50) 162.54 (310.75) 5.57 (4.90) u.s. large cap value 56 (14.39%) 1.90 (0.91) 52.17 (108.91) 6.71 (3.96) u.s. equity mid cap 86 (22.11%) 1.72 (0.51) 76.70 (154.58) 5.91 (3.96) u.s. equity small cap 74 (19.02%) 1.99 (0.47) 57.24 (103.87) 5.78 (4.13) all 389 1.76 (0.63) 79.70 (180.57) 6.10 (4.36) note. table 1 presents the aggregate summary statistics for incubated mutual funds. means (sd) are presented for expense ratio, total net assets, and manager tenure. additionally, presented is the number of funds along with their relative frequencies. 67p. gibson, m. finke / financial services review 23 (2014) 63–75 1998). the high expense ratio of incubated mutual funds might be attributed to a variety of factors. because incubated funds are smaller, economies of scale may make them more costly to operate. garavito (2008) finds that incubated mutual funds generally belong to smaller fund families. haslem, baker, and smith (2008) show that funds with low expense ratio generally outperform those with higher expense ratios. also presented in the table is the difference between total net assets (tna) and manager tenure. incubated mutual funds are statistically smaller than non-incubated mutual funds. however, this result should be taken with caution as it appears that the mean of nonincubated mutual funds are driven by some very large funds as indicated by the standard deviation. finally, we find that there is no statistical difference the in the length of time a manager has been with a fund. 6. incubated funds and skewness given the widely acknowledged non-linear relationship between skewness and net investor cash flow, it may be safe to infer that an investor’s preference for skewness is based on the upside potential signaled by funds in the market that experience recent high returns. skewness is estimated mathematically in the equation below: skewness ! !i"ri " ##3 $3 (1) where ri represents the monthly return of a mutual fund, # and $ symbolize the mean and the sd, respectively. one important feature of total skewness is that it is scaled by the variance of returns; this adjusts for any relationship between skewness and variance. skewness not only captures the first two moments (mean of the return distribution and volatility risk) but it also captures asymmetry that is characterized by the third moment. consistent with the idea that mutual fund companies launch their best incubated funds, we anticipate that the returns of incubated mutual funds are more positively skewed while in incubation (fig. 1). similar to consumers who purchase a lottery ticket with the hope of experiencing a windfall, mutual fund investors will purchase incubated mutual funds that signal the ability to generate above average returns. in other words, by marketing the high returns of incubated funds, mutual fund companies are sending a signal to consumers that they have identified fund managers with superior stock picking skills. table 3 two sample t test comparing incubated and non-incubated funds variable incubated nonincubated difference test-statistic expense ratio 1.75 1.67 $0.08** $2.29 tna 79.90 120.3 40.44*** 3.25 manager tenure 6.09 5.0 $0.30 $1.27 note. table 3 presents the results of a two sample t test used to compare the differences in expense ratio, total net assets (tna), and manager tenure of incubated and non-incubated mutual funds. additionally, present in the table is the statistical difference and swatterthwaite test statistics. incubated n % 389, non-incubated n % 1,039. *significant at 5%, **significant at 1%, and ***significant at 0.1%. 68 p. gibson, m. finke / financial services review 23 (2014) 63–75 6.1. results table 4 presents the descriptive characteristic of incubated mutual funds preand postincubation. the results in the table indicate that the average raw return net of fees is higher for incubated funds while in incubation. this result is also statistically significant; however, this can be attributed to the mutual fund families launching their best performing funds ex post returns. arguably this will always be the case, because those funds that perform poorly in incubation are eliminated and not made available to the public. using incubation as a strategy to gain investor attention appears to be a success. the result show that, on a ve ra ge r et ur n months before/after inception fig. 1. average return of incubated funds post-incubation and pre-incubation. table 4 t-test comparing various characteristics of incubated mutual funds post-incubation and during incubation variable postincubation during incubation (9,168) difference t value raw return 0.37 0.55 $0.17*** 3.77 expense ratio 1.77 1.80 0.01 0.93 tna 89.28 18.27 71.00*** 29.04 turn 91.68 96.87 $5.19*** 4.22 number of observation 42,085 9,168 note. this table contains the descriptive statistics of mutual funds, post-incubation and during incubation. the mean and statistical differences with t values are report for skewness, annual returns net of fees, expense ratio, total net assets (tna), and turnover ratio (turn). *significant at 5%, **significant at 1%, and ***significant at 0.1%. 69p. gibson, m. finke / financial services review 23 (2014) 63–75 average, incubated mutual funds quadruple in size shortly after being made publicly available. fig. 2 plots the average raw returns of incubated funds preand post-incubation. the reduction in performance post-incubation can be attributed to a few factors. first, there is clearly a selection bias since funds that perform well during incubation are more likely to be sold. however, this strong performance during incubation can simply be attributed to luck or random chance that that is not sustainable post-incubation. second, mutual fund companies are able to give their mutual funds preferential treatment during the incubation process. as expected, there is a decline in incubated funds net investor cash flow in years following inception as an investor’s decision are driven by performance. fig. 3 conveys the relative fund flows to incubated funds post-inception. the graph shows that there is an increase in investor cash flow during the first year after inception. however, there is a steady decrease from year two going forward. this is consistent with the notion that mutual fund investors are myopic return chasers. therefore, given that incubated funds do not perform well outside of incubation they do not attract the interest of individual investors. the process of incubation is clearly beneficial to mutual fund companies. by posting funds that are positively skewed, they are effectively able to grab the intention of investors who are looking for funds that have superior management that can result in lottery like returns. this in turns increases assets under management generating higher income for the mutual fund investment company. 6.2. the impact of incubation on fund flows in this section, we examine the investor preference for skewness when investing in incubated mutual funds at inception. it is important to point out that our measure of net investor cash flow comes directly from morningstar direct and it is estimated by stripping out two types of activities. one is expected growth of the assets because of capital market movements. the other is reinvestment of the capital gains and dividend distributions that a ve ra ge fu nd n et c as h fl ow s $ years after inception fig. 2. this figure shows net investor cash flow post-incubation. 70 p. gibson, m. finke / financial services review 23 (2014) 63–75 occur during the calculation month. similar to evans (2010), the dependent variable is ranked by year and month. we assign a fractional rank to each fund based on its net dollar flow for that year. we use a fractional rank instead of the direct measure of net investor cash flow for two reasons. first, the hypothesis being tested is whether incubated funds that are most skewed at inception attract a greater net dollar flow of funds. given variation in the size of younger funds, using a percentage rank reduces the probability that the results are driven by outliers. second, because there is variation in the net cash flow to mutual funds on a yearly basis, ranking funds within each time-period controls for the volatility. while observing mutual funds net investor cash flow, it is essential to control for various mutual fund characteristics that could impact the flow of money going into a fund. we control for the size of the fund, the fund’s family, the age of the fund, the fund’s expense ratio, as well as additional fees. controlling for the size of the fund is important for several reasons. first, sirri and tufano (1998) infer that an equal dollar flow will have a larger impact on smaller funds. second, barber, odean, and zheng (2005) advise that it will ensure that the results are not being driven by small funds. barber, odean, and zheng (2005), evans (2010), and sirri and tufano (1998), all document a negative relation between the total net assets of a fund and net investor cash flow, providing evidence that investors have a preference for investing in smaller funds. to control for the effect of outliers, the natural logarithm of total net assets (lntna) is included as a control variable. because larger fund families are more identifiable in the financial market, we control for the effects of the fund family size on net investor cash flow. evans (2010) shows a positive relation between a mutual fund’s family size and net investor cash flow. controlling for fund family size is important since fund families may steer money into new funds. similar to evans (2010), we also control for fund expenses and turnover. the non-linear relationship between mutual fund flows and performance is a years after inception a ve ra ge t n a fig. 3. this figure shows mutual funds total net assets post-incubation. 71p. gibson, m. finke / financial services review 23 (2014) 63–75 well-documented phenomenon in the mutual fund literature. for example, chevalier and ellison (1997) provides evidence that shows a non-linear relationship between mutual fund flows and performance. therefore, we control for performance using the annual market adjusted return over 12 months. summary statistics are presented in table 5. presented are various mutual fund characteristics that are known to impact net investor cash flow. the results are sorted into quintiles based on total skewness that is calculated using eq. (1). funds that with the highest returns appear to have a higher expense ratio (exp). funds with the highest skewness have a higher turnover ratio. those funds that are able to produce returns that are most skewed are larger and also belong to bigger fund families. consistent with the notion that investors chase after funds that have performed well, the results show that funds with a more positively skewed return experience a higher the net investor cash flow. finally, the results show a positive relation between net investor cash flow and skewness. funds with skewness in the highest quintile experience a greater inflow of new money compared to those the lowest quintile. 6.3. panel regression results table 6 conveys the results of a panel regression that was used to explore the investor preference for incubated funds that are skewed at inception. consistent with the findings of evans (2010), results show that there is a positive relation between funds that are incubated and net investor cash flow. the signs of the control variables are also consistent with previous literature. columns 3 and 4 show investors prefer investing in mutual funds that are positively skewed. in column 5, we use an indicator variable to capture those incubated funds that belong to the highest skewness quintile at inception. investors prefer investing in incubated funds that have a positively skewed return at inception. in column 6, we examine funds that are incubated and belong to the lowest skewness quintile. the results here are statistically insignificant. mutual fund managers appear to have an incentive to create funds in incubation and to launch these funds when their returns are highest. mutual funds investors appear to be driven by recent performance. because incubated table 5 descriptive statistics quintile skewness expense ratio turnover ratio tna family size new money flow 1 (low) $1.29 1.67 87.23 119.44 774.17 0.42 2 $0.77 1.70 89.36 139.76 821.14 0.25 3 $0.43 1.70 91.56 129.77 769.63 0.63 4 $0.06 1.72 90.25 134.23 870.54 1.02 5 (high) 0.53 1.74 92.00 144.46 975.49 1.07 note. this table provides the descriptive statistics for various mutual fund characteristics that are known to impact net investor cash flow. the results are sorted into quintiles by total skewness computed using eq. 1. expense ratio represents the mutual funds expense ratio. skewness represents the annual skewness of the fund. turnover ratio denotes the mutual funds turnover ratio. tna and family size represents the total net assets of the mutual fund and the mutual fund’s family, respectively. new money flow denotes net investor cash flow into the fund. 72 p. gibson, m. finke / financial services review 23 (2014) 63–75 funds have that have achieved high recent performance can be selectively promoted, they are appealing to mutual fund investors. generally, investors are not able to differentiate incubated and non-incubated funds; they simply select the fund that they believe will generate the highest possible return. this is problematic because the characteristic they often focus on as a quality signal is recent returns. this characteristic can be manipulated through the incubation process to create the false impression of positively skewed performance. evans (2010) details the light regulation by the sec as it pertains to incubated funds. because mutual funds are such a vital part of investors’ portfolios, especially those who are selecting funds for retirement, the incubation of funds may create a predictable welfare loss. table 6 panel regression: incubated funds at inception and net investor cash flow variable 1 2 3 4 5 6 intercept 0.51*** 0.64*** 0.75*** 0.51*** 0.77*** 0.76*** id new incubated 0.06*** 0.05*** 0.04*** (7.74) (6.22) (5.24) skewness 0.01** (1.96) id incep hiskew 0.04** 0.04** (2.56) (2.51) id incep lowskew 0.020 (0.94) logtnat-1 $0.05*** $0.01*** $0.01*** $0.01*** $0.01*** ($6.22) ($6.61) ($7.00) ($7.05) ($7.00) age $0.01*** $0.01*** $0.01*** $0.01*** $0.01** ($6.29) ($6.06) ($5.88) ($5.675) ($5.76) log famsize 0.01** 0.01 (0.00) 0.01 0.00 (2.27) (1.36) (1.28) (1.21) (1.30) expense ratio $0.04*** $0.03*** $0.04*** $0.04*** (5.72) ($5.93) ($5.80) ($5.79) turnover ratio $0.00 $0.00 $0.00 $0.00 ($0.31) ($0.22) ($0.02) ($0.28) ret1 0.003*** 0.003*** 0.003*** 0.003*** (14.99) (14.99) (14.97) (14.98) number of observations 159,982 159,982 160,222 159,982 159,982 159,982 fixed effects yes yes yes yes yes yes pseudo r2 0.01 0.1 0.2 0.2 0.2 0.2 note. this table shows the coefficients from a regression of investor cash flow on fund characteristics, including whether the fund is incubated and in the highest skewness quintile at inception. the dependent variable is net cash flow ranked by month and year. each fund is assigned a fractional rank between zero (lowest) and one (highest) based on net cash flow for that year. id incubated is represented by one if the fund is incubated, zero otherwise. skewness denotes the annual skewness of the mutual fund, computed using the formula in eq. 1. id incep hiskew and id incep lowskew represents incubated funds that belong to the highest and lowest skewness quintile at inception. lntnat-1 represents the log lag of a mutual fund’s total net assets. age represents the number of years a mutual fund has been in existence since its inception. log famsize represents the natural log of a mutual fund’s family size. expense, denotes the mutual funds expense ratio. turnover represents the mutual funds turnover. ret1 is the market adjusted return for the prior 12 months also reported are robust t statistics for each variable, number of observations, and pseudo r2. the asterisks statistical significance as followed: *significant at 5%, **significant at 1%, and ***significant at 0.1%. 73p. gibson, m. finke / financial services review 23 (2014) 63–75 this loss may be reduced by reducing a fund’s ability to backfill data or increasing awareness among financial advisors or investors of the incubation process. 7. conclusion consumers faced with high search costs when selecting mutual funds look for salient quality signals such as recent prior returns. recent returns can be manipulated by fund families through the process of incubation in which successful funds are marketed and unsuccessful funds are eliminated. we hypothesize that funds have an incentive to select highly skewed securities when incubating funds to increase the likelihood that some funds will achieve significant outperformance before they are sold to the public. to test this hypothesis, we investigate whether new investor dollars flow toward more highly skewed incubated funds after inception. the results of this study reveal that incubated funds carry a higher expense ratio relative to non-incubated funds. given that incubated funds are launched only after they achieve high returns, mutual fund companies are able to charge greater fees because they know that investors respond positively to performance. however, as the results of this study show, incubated funds do not have exceptional performance post-incubation. as a result, investors are made worse off investing in high expense funds that are created through incubation. the results of this study also reveal that when incubated mutual funds are made available to the public, those funds highly skewed before launch receive a larger inflow of funds relative to other funds. a possible explanation is the upside potential demonstrated by these funds during incubation. arguably, investors are not identifying managers with superior stock picking ability, but they are identifying funds whose possibly random outperformance during incubation is marketed as a signal of quality. this practice may be misleading and, given that these funds carry higher fees and do not perform as well once they are publicly available, there is a potential loss of wealth to investors. notes 1 i would like to thank richard evans at the university of virginia for providing me with this data. references ackerman, c., & loughran, t. (2007). mutual fund incubation and the role of the securities exchange commission. journal of business ethics, 70(1), 33–37. alexander, g. j., jones, j. d., & nigro, p. j. (1998) mutual fund shareholders: characteristics, investor knowledge, and sources of information. financial services review, 7, 301–316. arditti, f. (1967). risk and the required return on equity. journal of finance, 22, 19. barber, b. m., & odean t. (2008) all that glitters: the effect of attention and news on the buying behavior of individual and institutional investors. review of financial studies, 21, 785–818. 74 p. gibson, m. finke / financial services review 23 (2014) 63–75 barber, b., odean, t., & zheng, l. (2000). the behavior of mutual fund investors. working paper. barberis, n., & huang, m. (2007). stocks as lotteries: the implications of probability weighting for security prices. working paper. carhart, m. (1997). on the persistence of mutual fund performance. journal of finance, 52, 57–82. chevalier, j., & ellison, g. (1997). risk taking by mutual funds as a response to incentives. journal of political economy, 105, 1167–2000. evans, r. (2010) mutual fund incubation. journal of finance, 65, 1581–1611. garavito, f. (2008). mutual fund incubation: innovation or marketing tool. working paper. gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds. the journal of finance, 51, 783–810. harvey, c., & siddique, a. (2000). conditional skewness in asset pricing test. journal of finance, 55(2), 1263–1295. haslem, j., baker h., & smith d. (2008). performance and characteristics of actively managed retail equity mutual funds with diverse expense ratios. financial services review, 17, 49–68. jensen, m. (1968) the performance of mutual funds in the period 1945–1964. journal of finance, 23, 389–416. kraus, a., & litzenberger, r.(1976). skewness preference and the valuation of risk assets. journal of finance, 31, 1085–1100. kumar, a. (2009). who gambles in the stock market? journal of finance, 64, 0022–0082. palmitter, a., & taha, a. (2009). star creation: the incubation of mutual funds. vanderbilt law review, 62, 1485–1534. sirri, e., & tufano, p. (1998). costly search and mutual funds flow. the journal of finance, liii, 1589–1622. tversky, a., & kahneman, d. (1992). advances in prospect theory: cumulative representation of uncertainty. journal of risk and uncertainty, 5, 298–323. 75p. gibson, m. finke / financial services review 23 (2014) 63–75 boomers’ life insurance adequacy pre & post the 2008 financial crisis janine k. scott, ph.d.a,*, john gilliam, ph.d., mba, cfp, chfc, club aschool of economics & finance, massey university, palmerston north, new zealand bdepartment of personal financial planning, texas tech university, box 41210, lubbock, tx 79409-1210, usa abstract the baby boomers represent a large percentage of the u.s. population and their preparation for retirement, or lack thereof, can affect the economy at large. in light of the 2008 financial crisis, boomer households may be delaying retirement, choosing to work longer. using the 2004 and 2010 survey of consumer finance, logistic regression analyses are used to examine life insurance adequacy among boomers before and after the financial crisis of 2008. we find a significant difference in 2010 between the baby boomers and the senior generation in life insurance adequacy. variables related to net worth, such as income, marital status, and self-insurability, were significant predictors of life insurance adequacy. given greater life insurance adequacy among those with higher income, increasing group term insurance may help mid to low income households. further implications to practitioners, agents, and educators are discussed. © 2014 academy of financial services. all rights reserved. jel classification: g22 keywords: life insurance; baby boomers; financial planner; human capital 1. introduction the oldest baby boomers (born between 1946 and 1964), became eligible to draw on social security during the financial crisis of 2008 (farrell, 2013). as there are a large number of older americans, specifically the baby boomers, examining life insurance adequacy * corresponding author. tel.: �64 6 3569099, ext 83844; fax: �64 6 350 5660. e-mail address: j.k.scott@massey.ac.nz (j.k. scott) financial services review 23 (2014) 287–304 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. among this subgroup of the population is an important recourse. by 2030 the boomer population is projected to be 61 million (knickman and snell, 2002). in the event that baby boomers do not have sufficient amounts of life insurance, they may sustain financial hardships that have significant economic consequences. the impact of a financial crisis can especially reduce the purchasing power of older households (kirkpatrick and tennant, 2002). because these households are closer to retirement, they are more susceptible to income shocks. holding inadequate amounts of life insurance implies that more households may have to rely on social security survivor benefits, placing additional strain on the economy. because of the purchasing and consumption influence boomers possess, depressed spending can weaken an already strained economy (beinhocker, farrell, and greenberg, 2009). older households are interested in insurance not only from a pure death protection standpoint (i.e., term insurance). they are also interested in its saving component, estate tax benefits, in addition to bequest motives (baek and devaney, 2005; ibbotson, milevsky, chen, and zhu, 2007; mittra, 1995). this article includes the face amounts of both term and cash value life insurance. in the absence of sufficient insurance to cover insurable risks (e.g., untimely death), households will place too much emphasis on these risks while neglecting managing uninsurable risks (e.g., investing in the stock market; mitchell, 2011). this article seeks to assess life insurance adequacy pre and post the 2008 financial crisis by investigating if there is a difference in life adequacy between boomers and seniors (those born before 1946); current life insurance holdings among the boomers in comparison with other age cohorts preand post-financial crisis; and whether having a financial planner increases the likelihood of boomers having adequate life insurance post-financial crisis. 2. background historically, baby boomer wealth has exceeded their predecessors by more than 100%. they thrived during a time when saving was not a priority (beinhocker et al., 2009). a burgeoning stock market alongside loose lending practices and low borrowing costs weakened the savings motive. moreover, boomers as a whole did not share the same aversion to credit that had preoccupied the previous generation, who faced severe economic distress periods, such as the depression and world war ii. boomers may be wealthier than their predecessors, but the same characteristics that define their generation are creating unique challenges for their future that will have significant economic implications (gokhale and kotlikoff, 2000). they are living longer than earlier generations and choosing to exist independently from their children. they are spending their wealth at a faster rate than previous cohorts. many choose or need to work longer to save more for retirement. as a result, a larger portion of wealth is annuitized because of the expansion of programs such as medicare and social security; there are increasing healthcare costs, and more wealth is tied up in annuities, which results in less bequests. boomers also tend to be less risk adverse than younger households. this may be because of a decreased expected human capital value and a greater level of accumulated assets. for 288 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 older boomers, there may be no more dependents to provide for and there are fewer years to protect for the surviving spouse (lin and grace, 2007). a perception of decreased financial vulnerability leads some boomers to have inadequate life insurance holdings. the matter of inadequate life insurance among the baby boomers has been under examined. few studies compare how baby boomers utilized life insurance before and after the financial crisis of 2008; however, there is widespread research that examines how the crisis impacted the way baby boomers prepared for retirement (rosnick and baker, 2010; tres, 2010). the analysis of retirement preparation before and after the crisis provides a lens through which we can more broadly understand how baby boomer wealth was affected by the crisis, allowing us to further examine how life insurance utilization was shaped by the crisis. for example, munnell, golub-sass, soto, and webb (2008) find that in the context of the housing crash wealth effect, boomers may not have sufficient savings to sustain a desired standard living in retirement. it may not be surprising to find that boomers may not have adequate life insurance policies either. 2.1. life insurance inadequacy there are various definitions of “inadequacy” when associated with life insurance usage. many financial advisors use an individualized capital needs analysis to calculate the amount of life insurance an individual needs. however, according to the 2010 life insurance study (limra), the average american has enough life insurance to replace less than four years of income, with an average amount of $155,000 (lifejacket, 2011). the 2011 lifejacket study reveals that one-third of life insurance policyholders bought their policy a decade ago, suggesting coverage amounts may not have changed with life circumstances (lifejacket, 2011). inadequate life insurance can have a large impact on retirement preparedness in addition to significantly reducing living standards for widows (auerbach and kotlikoff, 1991; devaney, 1995). past studies demonstrate the severe economic impact of inadequate life insurance among older widows, the consequences of which include falling into poverty (bernheim, forni, gokhale, and kotlikoff, 1999). one of the primary reasons households possess insufficient amounts of insurance coverage is there is little income to extend beyond household necessities. the 2011 insurance barometer study reveals that 22% of respondents cite inadequate life insurance coverage. approximately 28% of respondents who are married would like their spouse or partner to have life insurance or to add more to their existing coverage. the study also cites that respondents with life insurance still express concerns about coverage contrasted with those who own long-term care or medical insurance—perhaps because of insufficient coverage and high financial risk in the event of a loss. almost half the respondents who have insufficient coverage state the price of life insurance as the foremost barrier to purchase, second to having other financial priorities. unawareness of need or risk also results in insufficient amounts of life insurance (johnson, 1970; mitchel, 2003), implying that educating households about the inherent necessity of life 289j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 insurance adequacy, and the risks of insufficiency, should result in adequate life insurance. this produces a simple equation: awareness of need and risk � knowledge � sufficient financial assets (should) � adequate life insurance. this would not only suggest an opportunity for financial planners to educate clients, but insinuate that simply delivering the information, updating clients about the changes in delivery and access to insurance, the wider use of the internet as regards life insurance, the use of cell phones for financial transactions, and so forth, should result in changed behavior. however, we know that the situation is far more complicated because of behavioral biases and other extenuating factors influencing household insurance behavior. for example, many households choose to postpone paying all debts in the event of a spouse’s death. they find alternative ways to fund education for their children. they may be willing to reduce living standards to adjust after losing a spouse or pin hopes on remarrying to justify lower amounts of life insurance. all of these reasons would be sufficient to warrant inadequate coverage of life insurance, but would leave room for copious “what ifs,” which may cost them future goals. additionally, the top 10% of life insurance companies had approximately one-quarter of their assets tied up in mortgagebacked securities by 2006—assets that were a part of many individual household portfolios. when many homeowners stopped paying their mortgages, this decreased the underlying value of these securities, and thus, severely reduced many household portfolios. for many households “sufficient financial assets” in the simple equation above was more than compromised (baranoff and sager, 2009). calculators are commonly used to determine life insurance adequacy by computing insurance needs that are based on income and arbitrary numbers, human capitalized method, capital needs analysis, the multiple income method, and the economic life cycle method (mitchell, 2003). the human capitalized model projects income into the future and then discounts it to the present. the capital needs method takes into consideration the reduction of household income because of the death of a wage earner and the decrease in living standard suffered by the survivors. this information is then incorporated into insurance needs. the multiple income method uses a predetermined figure and multiplies it by earnings to derive total life insurance needs. the economic life cycle is based on the life cycle model—finding an optimal, smooth consumption path over the insured’s lifetime and determining appropriate life insurance needs. because these are different methods, incorporating different household financial information, the amount of life insurance will vary based on inputs. past studies that examine life insurance demand use similar variables to assess adequacy. using the 1992 wave of the health and retirement survey (hrs), and the economic security planner (esplanner) financial planning software to compute life insurance needs, bernheim et al. (1999) finds underinsurance among single households, non-white households, younger households, and other groups. finke, huston, and waller (2009) developed the most recent and extensive model of life insurance adequacy. this model includes current life insurance, household income, estimated taxes, household economies of scale, and other key factors to determine adequacy. finke, huston, and waller’s (2009) model for life insurance adequacy will serve as the framework for this study, discussed later in the article. independent variables used in the analyses are discussed below. 290 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 2.2. independent variables 2.2.1. age the effect of age on insurance demand is varied (chen, wong, and lee, 2001; showers and shotick, 1994; zietz, 2003). as age increases, we would expect life insurance face value amounts to decrease because the need to replace living expenditures would also be decreasing. because the value of human capital decreases and the cost of insurance rises, it follows that there should be a decrease in life insurance coverage (campbell, 1980). alternatively, an increase in mortality risk (and poor health) can necessitate greater insurance holdings, depending on household preferences (finke, huston, and waller, 2009). for example, older households, such as the baby boomers may hold life insurance because of bequest motives or estate planning reasons (baek and devaney, 2005). given that prior studies find that households fail to adjust life insurance coverage after initial purchase, we expect no difference between baby boomers and the senior generation in life insurance adequacy pre and post the 2008 financial crisis. 2.2.2. financial professional the inclusion of a financial planner, accountant, broker, and banker in the descriptive analysis is pertinent because these professionals may at some point dispense life insurance advice (mulholland, finke, and huston, 2012). although the definition of a financial planner in the survey of consumer finances is not particularly clear, as it does not distinguish between financial professionals who present themselves as advisors with or without a cfp designation, it is interesting to see how a financial planner compares with the other previously listed professionals. based on the findings of finke, huston, and waller (2009) and scott and finke (2013), we expect a positive relation between the use of a financial planner (compared with using a non-financial planner professional) on life insurance adequacy. 2.2.3. education education also has varied results from prior studies, demonstrating both positive and negative associations with life insurance demand (zietz, 2003). however, stemming from human capital theory, households with more education should have greater protection because of a steeper earning capacity (bryant and zick, 2006). they should also be better able to make optimal insurance decisions than those with less education or utilize a financial service professional to meet insurance needs. therefore, we expect a positive association between education and life insurance adequacy. 2.2.4. marital status previous literature reports positive and negative as well as non-significant findings between marital status and life insurance demand (zietz, 2003). because married households possess the advantage of pooling resources as well as exhibiting greater needs for insurance than single households, we expect a positive effect on life insurance adequacy. 291j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 2.2.5. race ethnic group differences also play a role in life insurance ownership. table 1 shows differences between blacks, hispanics, and whites when compared with each other. these differences are key in pinpointing specific strategies for each group, considering their preferences and goals for life insurance. for example, blacks have an overall preference for life insurance compared with hispanics and whites. they are more likely to consider burdening family members as a top reason for purchase. blacks and whites will use life insurance to cover funeral expenses more than hispanics, who will use insurance to fund education needs (mitchel, 2011). though race has been significantly related to the findings in past studies, we do not expect any differences among racial classes when examining adequacy. 2.2.6. risk tolerance as risk aversion increases, a household should possess greater amounts of life insurance (ibbotson et al., 2007). households who are substantial risk takers display a preference for risk seeking behavior; the risk tolerance question used in the survey is limited and may not fully reflect risk tolerance. therefore, we do not expect a significant relation between risk tolerance and life insurance adequacy. 2.2.7. self-employment given the volatility of earnings within self-employed households, there should be greater life insurance coverage than for households who are not self-employed. from a human capital standpoint, greater life insurance adequacy may be warranted for self-employed households. 2.2.8. net worth net worth and life insurance demand has been found to be negatively related (baek and devaney, 2005), as those with greater net worth may be able to self-insure or afford an unexpected shock to income. table 1 ethnic group difference in life insurance ownership* blacks hispanics whites positive attitude about li 1 1 1 feel agents/financial professionals are knowledgeable 1 2 1 have a greater concern about placing financial burden on others 1 2 2 seek to understand the product 1 1 1 prefer to purchase at their workplace 2 2 2 research on the internet 1 1 1 prefer to purchase on the internet 2 1 2 prefer to purchase from agent/financial professional 1 2 1 * adopted from the 2011 insurance barometer study. 292 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 2.2.9. ability to self-insure if assets are greater than the present value of human capital, households can afford to self-insure; thus, having a negative impact on life insurance adequacy. although net worth and the ability to self-insure are indeed correlated, the correlation does not interfere with the analysis evident by correlation tests. 2.2.10. income bernheim et al. (1999) finds more underinsurance at moderate levels of income. as household income increases, underinsurance increases also. however, past studies also demonstrate greater affordability of insurance with increased income (browne and kim, 1993; lewis, 1989). as income variation increases, the present value of human capital decreases, leading to less demand for life insurance (finke, huston, and waller, 2009). 2.2.11. expectation of income growth this is one of the variables that aim to capture attitudes related to life insurance adequacy. ibbotson et al. (2007) cite that with higher income, a higher discount rate should be used to value human capital and, thus, would depress life insurance holdings. 2.2.12. spending greater than income this variable accounts for liquidity constraints of the household (bernheim et al., 1999). households who spend greater than income will be less likely to have adequate life insurance. 2.2.13. social security income the receipt of social security benefits may depress life insurance holdings (bernheim, 1991; fitzgerald, 1987). as with other sources of income, social security benefits should be included in life insurance demand as it influences survivor benefits (lewis, 1989). 2.2.14. health related to human capital, greater health should have a positive effect on life insurance coverage (baek and devaney, 2005). better health promotes or facilitates better formation of human capital and it can also translate into more affordable premiums. however, there may be no effect on adequacy. 2.2.15. bequest motive households with a bequest motive should demand life insurance at a greater rate than households without a bequest motive. however, having a bequest motive and executing financial strategies to prepare for the distribution of assets are two separate matters. therefore, although we would expect a positive relation between having a bequest motive and adequate life insurance, it would not be surprising to find insignificant results. 2.2.16. presence of children life insurance demand should be greater for those households with children (lewis, 1989; mulholland, finke, and huston, 2012). 293j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 2.2.17. expectation of having a sizeable estate at death the inclusion of this variable in the model is similar to including a bequest motive variable. again, preferences or expectations may not always lend to financial preparedness. 2.2.18. homeowner based on the findings of other studies, homeownership should have a positive effect on life insurance coverage (gandolfi and miners, 1996). homeownership can also serve as a signal for greater life insurance adequacy because of the accumulation of assets within and beyond the building years of a household’s life cycle stage. 3. theory a household seeks to maximize the expected utility of wealth, that is, pursue a level consumption path over the life cycle. defined as the present value of expected future market wages, human capital is a non-tradable asset in the household portfolio. labor income is used to fund most of household consumption, and so for a majority of households this dominates the financial capital side of the household portfolio (campbell, 1980). life insurance serves as an actuarial hedge against the loss of income whereas other financial assets seek to reduce diversifiable risk (collins and lam, 2011). life cycle hypothesis supports the use of financial capital to supplement an unexpected loss of human capital to mitigate consumption shocks that can jeopardize financial goals. a household’s age in the life cycle influences life insurance needs. in a broad sense, younger households with greater human capital should possess adequate life insurance more than older households with less human capital to protect. all households should rationally prefer a smooth transition from one stage of the life cycle to the next, but this is not always possible. older adults, specifically baby boomers in the period of analysis, are still in the building stage and susceptible to income and consumption shocks that may alter life insurance needs or preferences. in short, their life cycle path can become steeper depending on unexpected variation in income and/or the economic environment. this leads to greater need for life insurance adequacy (not just coverage) compared with elders with greater stability in income and/or less exposure to the stock market. significant predictors of demand life insurance include household size, number, and age of dependent children, and income (baek & devaney, 2005). lewis (1989) developed his life insurance model based on the survivor’s preferences for insurance. demand for life insurance should decline with age as the value of human capital decreases (baek and devaney, 2005). however, demand for life insurance should theoretically terminate at retirement as the value of human capital shrinks to zero (collins and lam, 2011; finke, huston, and waller, 2009). however, many households continue to hold onto policies beyond retirement age, most specifically, cash value life insurance policies for a number of reasons (brown, 1999). 294 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 4. methodology 4.1. model to determine life insurance adequacy, the following method adopted from finke, huston, and waller (2009) is used. first, it is necessary to determine the number of children under age 18 present in the household. then, using the u.s. census bureau household poverty and income information, we determined if a household is above or below the poverty line based on income and dependents. the sample is then censored to households ages 35 to 70 (35 to 76 in 2010) who are working full-time. the 2010 sample is increased to age 76 to account for the senior generation. using bernheim et al. (2001) we adjust for household economies of scale to assess replacement needs in a two adult household. taxes are estimated based on filing status, adjusted gross income, and personal exemptions. unlike finke, huston, and waller (2009), we use the original filing status question in the survey of consumer finances (scf) to be consistent with the survey years. insurance coverage is calculated based on the face value of both term and cash value life insurance. instead of recreating the insurance ratio developed by finke, huston, and waller (2009), it seems more straightforward to view the insurance ratio as current life insurance to life insurance needs. the life insurance adequacy model is presented in eq. (1) below. adq li � � p�1 p vp �it � tt � f� � ��1 � �1 � rrr� �w� rrr � (1) life insurance adequacy is a function of human capital and household financial characteristics. where vp in the numerator represents the face value of cash value and term life insurance policies; w, the difference between the respondent’s age and retirement age; it, the household’s current income; f, household economies of scale ratio (ability to pool resources in a dual-income household); tt, estimated taxes based on filing status, personal exemptions, and standard deduction; rr, expected real interest rate. a household has adequate life insurance when its insurance value (in the numerator) to insurance need (represented by the denominator) is greater than or equal to one. we assume that income growth is equal to the rate of inflation and to determine overall replaceable need, a discount rate of 2.3% is used. the regression model is shown below in eq. (2). adequate life insurance � �0 � �1age � �2financial professional � �3demographic characteristics � �4financial characteristics � �5financial attitudes � � (2) adequate life insurance is a binary variable equal to 1 if the household has an insurance coverage to insurance needs ratio is greater than or equal to 1. age includes boomers, seniors, 295j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 and generations younger than boomers. financial professional includes the use of a financial planner and non-financial planner. demographic characteristics is a vector of independent variables representing race, education, children, and marital status. financial characteristics includes net worth, income, self-insurability, self-employment, and homeownership. financial attitudes contains a bequest preference, a substantial risk tolerance level, the expectation of greater income in the future, and spending more than annual income. 4.2. dataset as we are examining life insurance adequacy among baby boomers pre and post the 2008 financial crisis, the 2004 and 2010 scf datasets are used. this survey contains the necessary household demographic and financial characteristics for this research study. there were 4,519 respondents in 2004 and 6,482 in 2010. for the descriptive analyses, the scf population weights will be used to represent the u.s. population as a whole but not for the multivariate analysis (rubin, 1987). following finke, huston, and waller (2009), the sample in 2004 was censored to married households working full-time, between age 35 and 70 (expanded to age 76 in 2010) and possessing income above the poverty line. 4.3. descriptive statistics descriptive statistics are presented in tables 2 and 4. table 3 shows the coding for each variable in the study. table 2 displays the mean insurance and mean insurance needs for all respondents. of the generational groups, boomers have the highest insurance mean of $345,297 in 2004 but a greater insurance need than seniors in both years. insurance holdings on average are higher among respondents who report seeking advice from a financial planner ($478,396) compared with those who use another financial professional ($284,421). however, insurance needs are greater (by �$81,000) for households with a financial planner. table 4 displays demographic, financial, and other household characteristics. the first column for each year includes all respondents and the second column is censored/restricted based on age, households with more than one member and income above the poverty line. the sample size in 2004 for the censored sample is 1,966 households and 2,612 in 2010. table 2 mean insurance and insurance need 2004 2010 insurance insurance need insurance insurance need boomers $345,297.34 $676,507.76 $419,124.74 $ 700,545.65 seniors $235,741.31 $325,580.83 $295,727.41 $ 286,945.90 genx and below $235,952.48 $808,799.39 $342,860.90 $1,212,846.14 financial planner $478,396.10 $736,178.60 $586,766.69 $ 888,523.82 non-financial planner $284,421.30 $655,043.47 $351,424.69 $ 872,127.81 296 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 approximately 15% of these households possess adequate life insurance in 2004 and 14% have adequate life insurance in 2010. in both survey years, the majority of respondents are baby boomers followed by seniors in 2004 and genx and below in 2010. the use of a financial planner is greater in the restricted sample for respondents versus the use of a planner in the full sample. financial attitudes such as leaving a bequest and expecting to leave a sizeable estate did not change greatly from 2004 to 2010 in the restricted samples. table 3 coding for each variable from the 2004 and 2010 survey of consumer finances data variables coding adequate life insurance (dependent) 1 � adequate life insurance boomers 1 � age between 39 and 58, 0 otherwise in 2004; 1 � age between 45 and 64, 0 otherwise in 2010 seniors 1 � age � 59, 0 otherwise in 2004; 1 � age � 65, 0 otherwise in 2010 generation x and below 1 � age � 39, 0 otherwise in 2004; 1 � age � 45, 0 otherwise in 2010 less than high school 1 � yes, 0 otherwise high school graduate 1 � yes, 0 otherwise some college 1 � yes, 0 otherwise college graduate 1 � yes, 0 otherwise single 1 � single, divorced, or widowed, 0 otherwise non-single 1 � married, 0 otherwise white 1 � yes, 0 otherwise black 1 � yes, 0 otherwise hispanic 1 � yes, 0 otherwise non-financial planner 1 � banker, accountant, or broker, 0 otherwise financial planner 1 � yes, 0 otherwise spends more than income 1 � yes, 0 otherwise expect income growth 1 � yes, 0 otherwise substantial risk taker 1 � yes, 0 otherwise self-employed 1 � yes, 0 otherwise log of net worth continuous self-insurable 1 � assets is greater than the present value of human capital, 0 otherwise social security income 1 � yes, 0 otherwise income �$35,150 1 � yes, 0 otherwise income between $35,150–$90,800 1 � yes, 0 otherwise income between $90,800–$147,050 1 � yes, 0 otherwise income between $147,050–$288,350 1 � yes, 0 otherwise income �$288,350 1 � yes, 0 otherwise poor health 1 � yes, 0 otherwise good or excellent health 1 � yes, 0 otherwise bequest motive 1 � very important/important, or positive, differing among spouses, weak positive, or negative � 0 have children 1 � kids living at home or away from home, 0 otherwise expect to leave sizeable estate 1 � yes or possibly, 0 otherwise homeowner 1 � own ranch/farm/mobile home/house/condo/coop/�, 0 otherwise 297j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 5. results a pearson correlation test was implemented with all variables in the regression models (e.g., between net worth and income) to avoid multi-collinearity issues. we did not detect any evidence of multi-collinearity. two binomial logistic regression models are tested. the first, shown in table 5, represents all respondents in the censored sample, while table 6 shows results for only baby boomers in the censored sample. table 4 weighted frequencies of all respondents and censored sample (used for adequate life insurance; in %) 2004 2010 all respondents (n � 4,519) restricted sample (n � 1,966) all respondents (n � 6,482) restricted sample (n � 2,612) adequate life insurance (dependent) 15.41 14.16 cohorts boomers 69.18 64.36 59.58 70.81 seniors 16.02 28.05 5.70 11.08 genx and below 14.80 7.59 34.72 18.11 education less than high school education 9.16 3.30 9.91 4.59 high school grad 18.97 15.51 20.72 16.76 some college 17.90 17.16 19.82 14.32 college grad 53.97 64.03 49.55 64.32 good or excellent health 88.56 91.42 86.73 88.11 financial professional financial planner 15.51 21.78 15.68 22.16 non-financial planner 84.49 78.22 84.32 77.84 marital status married 81.99 89.77 80.37 98.38 single 18.01 10.23 19.63 1.62 race black 7.22 5.28 8.66 2.97 hispanic 7.27 1.98 9.79 3.78 white 85.50 92.74 81.55 93.24 have children 93.08 95.38 92.82 95.41 attitudes/expectations spends � income 13.12 10.89 15.45 11.35 expect income growth 32.25 35.64 23.06 27.03 substantial risk taker 5.54 6.27 5.03 6.22 bequest motive 57.27 54.46 55.01 54.05 expect to leave sizeable estate 68.41 74.92 66.84 75.41 self-employed 39.01 47.19 32.27 42.43 homeowner 85.10 94.06 81.27 95.14 ability to self-insure 47.91 71.62 45.34 62.16 social security benefits 5.95 9.57 5.93 12.43 income �$35,150 12.77 6.27 12.88 2.43 $35,150–$90,800 32.25 26.40 33.94 24.59 $90,800–$147,050 14.39 14.85 17.13 17.30 $147,050–$288,350 12.46 17.16 13.38 23.51 �$288,350 28.13 35.31 22.67 32.16 298 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 in the first regression (see table 5) generational groups vary significantly according to amounts of adequate life insurance in 2010. the baby boomer generation was 62% more likely than the senior generation to have adequate life insurance. the financial crisis also impacted racial disparities among household life insurance holdings. in 2004, hispanic households were less likely than white households to have adequate life insurance; however, there was not a significant difference in 2010. households who can afford to self-insure, having assets greater than human capital, demonstrated greater life insurance adequacy in both survey years. in 2010, married household as compared with single households were more likely to have adequate life insurance coverage. this is consistent with what we expected, given household economies of scale. table 5 regression results: dv � adequate insurance (insurance ratio �1) variables 2004 (n � 1,966) 2010 (n � 2,563) parameter estimate odds ratio parameter estimate odds ratio adequate life insurance (dependent) generation (senior) boomer �0.06 0.70 0.34** 1.62 genx and below �0.25 0.57 �0.20 0.95 education (�high school) high school grad 0.14 1.99 0.11 1.42 some college 0.22 2.14 �0.09 1.15 college grad 0.18 2.07 0.22 1.58 marital status (single) non-single 0.14 1.32 1.11*** 9.14 race (white) black 0.42 1.17 �0.12 0.66 hispanic �0.68* 0.39 �0.17 0.63 financial professional (non-financial planner) financial planner 0.21** 1.53 0.14 1.33 spends more than income 0.06 1.12 �0.08 0.84 expect income growth 0.05 1.10 0.04 1.08 substantial risk taker 0.03 1.06 0.07 1.15 self-employed �0.05 0.91 �0.02 0.95 log of net worth �0.01 0.99 0.00 1.00 self-insurable 0.52*** 2.84 0.25*** 1.66 social security income 0.11 1.25 0.33* 1.93 income (�$35,150) $35,150–$90,800 0.04 1.33 0.19 3.70 $90,800–$147,050 0.04 1.32 0.20 3.72 $147,050–$288,350 0.27 1.67 0.63*** 5.73 �$288,350 �0.11 1.15 0.09 3.34 health (poor) good or excellent health 0.03 1.06 �0.11 0.80 bequest motive �0.11 0.80 �0.04 0.93 have children 0.16 1.38 0.15 1.35 expect to leave sizeable estate �0.09 0.83 �0.09 0.83 homeowner 0.26 1.70 0.29* 1.79 intercept �2.23 �3.40 data are from the survey of consumer finances. *p � 0.05, **p � 0.01, ***p � 0.001. 299j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 other positive predictors include receiving social security income, being a homeowner and possessing income greater than $147,000. specifically, respondents reporting income between $147,050 and $288,350 represented the highest odds of insurance adequacy, 473% higher compared with the lowest income group (less than $35,150). having a financial planner was a significant and positive predictor of insurance adequacy in 2004, highlighting greater comprehensive planning in risk management than among other financial professionals. however, post the 2008 crisis, having a planner had no significant effect on life insurance adequacy. the second regression analysis includes only baby boomers (see table 6). as with the all respondent censored sample in 2004, the odds of a boomer having adequate life insurance were 156% greater if they can self-insure in comparison with respondents with insufficient table 6 regression results censored by baby boomers: dv � adequate insurance (insurance ratio �1) variables 2004 (n � 1,360) 2010 (n � 1,527) parameter estimate odds ratio parameter estimate odds ratio adequate life insurance (dependent) education (�high school) high school grad 0.16 2.69 0.22 1.69 some college 0.38 3.37 �0.03 1.32 college grad 0.29 3.07 0.12 1.53 marital status (single) non-single �0.04 0.93 1.07*** 8.48 race (white) black 0.64* 1.68 �0.37 0.48 hispanic �0.77* 0.41 0.00 0.70 financial professional (non-financial planner) financial planner 0.17 1.41 0.10 1.23 spends more than income 0.01 1.02 0.02 1.04 expect income growth 0.01 1.03 0.00 1.00 substantial risk taker 0.05 1.10 0.04 1.08 self-employed �0.02 0.96 �0.07 0.87 log of net worth 0.03 1.03 �0.03 0.97 self-insurable 0.47*** 2.56 0.37*** 2.12 social security income �0.25 0.61 0.64*** 3.63 income (�$35,150) $35,150–$90,800 0.14 1.65 0.08 4.95 $90,800–$147,050 �0.04 1.38 0.26 5.89 $147,050–$288,350 0.31 1.96 0.87*** 10.87 �$288,350 �0.05 1.37 0.30 6.15 health (poor) good or excellent health 0.05 1.10 �0.11 0.81 bequest motive �0.05 0.91 0.01 1.01 have children 0.15 1.35 0.27 1.71 expect to leave sizeable estate �0.12 0.79 �0.12 0.79 homeowner 0.30 1.81 0.40 2.22 intercept �3.11 �2.68 data are from the survey of consumer finances. *p � 0.05, **p � 0.01, ***p � 0.001. 300 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 resources to self-insure. if an individual can self-insure, coverage may be unnecessary but adequate. in 2004, hispanic boomers were 59% less likely to be adequately insured than white boomer households. as expected from previous research, black boomers were more likely to be adequately insured than white boomers in 2004; however, in 2010, there was not a remarkable discrepancy. examining over-insurance among black households would be a worthwhile research topic in the future. post 2004 we see similar results in the boomer sample. in 2010, there was a positive relation between life insurance adequacy and boomers who were married, receiving social security income, having assets to self-insure, and being within the second highest income category. variables related to net worth and income had more of an impact post the 2008 financial crisis, whereas racial status, in addition to having substantial assets, influenced insurance adequacy before the crisis. this is not surprising because higher net worth households can draw on assets in difficult economic times more so than those with lower net worth. 6. conclusion little in the literature exists that specifically examines how the financial crisis of 2008 affected life insurance adequacy among the baby boomer generation. this study aims to expand that literature by addressing the unique characteristics and challenges that define the boomer’s generation and economic context, and by examining boomer life insurance adequacy before and after the financial crisis of 2008 as compared with older and younger generations. because life insurance is often sold not bought, we did not expect to see differences in adequacy within the survey years used in the study. however, using binomial logistic regression analyses for the 2004 and 2010 scf years, we find compelling disparities between baby boomers and the senior generation in 2010 but not in 2004, revealing that the boomer generation is 62% more likely than the older cohort to have adequate life insurance post the financial crisis of 2008. these results were surprising, given that boomers display a higher risk tolerance than the senior generation, but positive for boomer households in a postfinancial crisis economic context. when restricting the sample to only boomers, the regression analysis demonstrated similar results to the analysis of the senior generation and the younger generation insofar as race, self-insurability, income, and marital status impacted life insurance adequacy in the survey years. for the 2004 scf survey year, both samples reflected racial discrepancies wherein black households are more likely to be adequately insured than white households, and hispanic households are at a greater risk of being inadequately insured; however, there is little racial difference in 2010. households whose assets exceed human capital, and can afford to self-insure, showed significant life insurance adequacy. in some cases when an individual can self-insure, coverage may not even be necessary, though it is certainly adequate. married households, in 2010, are more likely than single households to be adequately insured. social security income is also a positive predictor of life insurance 301j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 adequacy. consistent with prior studies, social security acts as a substitute for life insurance but is also indicative of household wealth (browne and kim, 1993; lewis, 1989). the use of a financial planner is higher among households with adequate life insurance compared with the all respondents sample. furthermore, having a financial planner compared with households without a planner is a significant predictor in the all respondent regression sample in 2004, suggesting that they are making a positive impact on household insurance behavior. the results indicate that variables related to net worth (specifically assets used to calculate self-insurability) drive life insurance adequacy. those who can afford to fully insure, having a ratio of insurance to insurance need greater than or equal to one, are able to hold suitable amounts of life insurance or none at all. though there is not enough evidence to reject the null hypothesis post 2008, as boomers had better life insurance protection than the senior generation, this topic warrants further investigation and has pressing implications to financial planning practitioners. given longevity and changes in the definition and the delay of retirement for many, it is important to better devise ways to aid households in acquiring and sustaining adequate life insurance. financial planning practitioners and life insurance agents can improve their services to clients by recognizing generational cohort differences and illustrating dynamic life insurance needs analysis in front of clients. this needs analysis may be different for each generation. for example, younger consumers are more interested in the price of insurance, whereas older households want to see product comparisons (lifejacket study, 2011). purchase preferences of baby boomers are different than younger age cohorts. boomers prefer face-to-face interaction, especially those households who have been divorced, are widowed or separated (mitchell, 2011). sales strategies that worked to attract boomers may not work for later generations such as generation y. technology innovation, in addition to the increase in speed of accessing information now than in the past, provides a new ground on which practitioners can capture both younger and older age groups. regarding life insurance purchase via the internet, purchase strategies tailored specifically to men may be worthwhile, because more males than females shop for life insurance online (mitchell, 2011). for example, using language that evokes emotion, for example, “breadwinner,” or “do not leave your family unprotected,” rather than deterring consumers with morbid language is important. social policy implications are evident from the study. auerbach and kotlikoff (1991) suggest that employers may consider increasing group life insurance coverage. other considerations to lessen potential hardships from the lack of adequate life insurance include expanding social security survivor benefits. however, both of these suggestions warrant further research and resources. education is key (thomas, 2006) and insurance agents have cultivated a reputation in the industry—whether good or bad. however, agents must realize that trust precedes education, because it is a vital factor in purchasing decisions, especially among boomers. for example, blacks are more likely than whites to report that agents are trustworthy and females are more likely than males to have a positive view of insurance salespersons (mitchell, 2011). being transparent about fees and sales charges is another integral part of educating clients and building trust among all age cohorts (thomas, 2006). taking generational differences into 302 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 consideration can be of tremendous value to insurance agents, financial service professionals, and educators. we are in an indispensable position to help manage preferences and shape perceptions about the future. references auerbach, a. j., & kotlikoff, l. (1991). life insurance inadequacy-evidence from a sample of older widows (no. w3765). cambridge, ma: national bureau of economic research. baek, e., & devaney, s. a. (2005). human capital, bequest motives, risk, and the purchase of life insurance. journal of personal finance, 4, 62–84. baranoff, e. g., & sager, t. w. (2009). the impact of mortgage-backed securities on capital requirements of life insurers in the financial crisis of 2007–2008. the geneva papers on risk and insurance-issues and practice, 34, 100–118. beinhocker, e. d., farrell, d., & greenberg, e. (2009). why baby boomers will need to work longer. mckinsey quarterly, 1, 118–127. bernheim, b. d. (1991). how strong are bequest motives? evidence based on estimates of the demand for life insurance and annuities. journal of political economy, 99, 899–927. bernheim, b. d., forni, l., gokhale, j., & kotlikoff, l. j. (1999). the adequacy of life insurance: evidence from the health and retirement survey (no. w7372). cambridge, ma: national bureau of economic research. bernheim, b. d., carman, k. g., gokhale, j., & kotlikoff, l. j. (2001). the mismatch between life insurance holdings and financial vulnerabilities: evidence from the survey of consumer finances (no. w8544). cambridge, ma: national bureau of economic research. brown, j. r. (1999). are the elderly really over-annuitized? new evidence on life insurance and bequests (no. w7193). cambridge, ma: national bureau of economic research browne, m. j., & kim, k. (1993). an international analysis of life insurance demand. journal of risk and insurance, 60, 616–634. bryant, w. k., & zick, c. d. (2006). the economic organization of the household. new york, ny: cambridge university press. campbell, r. (1980). the demand for life insurance: an application of the economics of uncertainty. journal of finance, 35(5), 1155–1172. chen, r., wong, k. a. & lee, h. c. (2001). age, period and cohort effects on life insurance purchases in the u.s. the journal of risk and insurance, 68, 303–327. collins, p., & lam, h. d. (2011). asset allocation, human capital, and the demand to hold life insurance in retirement. financial services review, 20, 303. devaney, s. a. (1995). retirement preparation of older and younger baby boomers. financial counseling and planning, 6, 25–33. farrell, c. (2013). don’t weep for boomers close to retirement. businessweek.com, 1. finke, m., huston, s., & waller, w. (2009). do contracts impact comprehensive financial advice. financial services review, 18, 177–193. fitzgerald, j. (1987). the effects of social security on life insurance demand by married couples. journal of risk and insurance, 54, 86–99. fortune, p. (1973). a theory of optimal life insurance: development and tests. journal of finance, 28, 587–600. gandolfi, a. s., & miners, l. (1996). gender-based differences in life insurance ownership. journal of risk and insurance, 63, 683–693. gokhale, j., & kotlikoff, l. j. (2000). the baby boomers’ mega-inheritance: myth or reality? federal reserve bank of cleveland, research department. ibbotson, r. g., milevsky, m. a., chen, p., & zhu, k. x. (2007). lifetime financial advice: human capital, asset allocation, and insurance. charlottesville, va: research foundation of cfa institute. johnson, r. e. (1970). the college professor and the adequacy of his insurance coverage. journal of risk & insurance, 37, 105–114. 303j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 kirkpatrick, c., & tennant, d. (2002). responding to financial crisis: the case of jamaica. world development, 30, 1933–1950. knickman, j. r., & snell, e. k. (2002). the 2030 problem: caring for aging baby boomers. health services research, 37, 849–884. lewis, f. (1989). dependents and the demand for life insurance. the american economic review, 79, 452–467. the lifejacket study. (2011). seven key insights to help close the coverage gap. life insurance. new york, ny: genworth life insurance company. lin, y., & grace, m. f. (2007). household life cycle protection: life insurance holdings, financial vulnerability, and portfolio implications. journal of risk and insurance, 74, 141–173. mitchell, j. o. (2011). 2011 insurance barometer study. usa: life foundation and ll global, inc. mitchel, j. o. (2003). the adequacy of life insurance coverage in u.s. households. journal of financial service professionals, 57, 54–63. mittra, s. (1995). practicing financial planning: a complete guide for professionals. troy, mi: mittra & associates. mulholland, b. s., finke, m. s., & huston, s. (2012). understanding the shift in demand for cash value life insurance. lubbock, tx: texas tech university. munnell, a. h., golub-sass, f., soto, m., & webb, a. (2008). do households have a good sense of their retirement preparedness? issue in brief, 8–11. chestnut hill, ma: center for retirement research at boston college. rubin, d. b. (1987). multiple imputation for nonresponse in surveys. new york, wiley. rosnick, d., & baker, d. (2010). the impact of the housing crash on the wealth of the baby boom cohorts. journal of aging & social policy, 22, 117–128. scott, j. k., & finke, m. j. (2013). the demand for disability insurance. financial services review, 22, 1–12. showers, v. e., & shotick, j. a. (1994). the effects of household characteristics on demand for insurance: a tobit analysis. journal of risk and insurance, 61, 492–502. thomas, t. (2006). limra looks at boomers and opportunities for growth. national underwriter/life & health financial services, 110, 7–41. trese, h. (2010). boomers in crisis. agent’s sales journal, february 2010, 24–26. zietz, e. (2003). an examination of the demand for life insurance. risk management and insurance review, 6(2), 159–191. 304 j.k. scott, j. gilliam / financial services review 23 (2014) 287–304 a new strategy to guarantee retirement income using tips and longevity insurance: a second look paul j. haensly, ph.d.a,*, k. prakash pai, ph.d.a acollege of business and engineering, the university of texas of the permian basin, 4901 e. university blvd., odessa, tx 79762-0001, usa abstract shankar (2009) proposes a new investment strategy for retirees that bundles treasury inflation protected securities with a deferred annuity to guarantee real annual withdrawal rates of 5% or more with no risk of financial ruin. this strategy addresses three problems that retirees face: longevity risk, inflation risk, and liquidity risk inherent in the purchase of an immediate annuity. in our article, we evaluate the performance of this proposed strategy under realistic assumptions about costs, security design, and markets. in addition, we evaluate how the bequest motive might affect the choice between shankar’s strategy and an immediate annuity. © 2015 academy of financial services. all rights reserved. jel classification: d14; g11; g22 keywords: retirement planning; withdrawal rates; tips; longevity risk; financial ruin 1. introduction an inflation-indexed immediate annuity appears, at first glance, to be an ideal solution to two important financial problems that retirees generally face: maintaining a desired level of real income and avoiding the risk of running out of money, that is, financial ruin. however, shankar (2009) observes that investors have been extremely reluctant to commit their life savings to immediate annuities. as an alternative, shankar proposes an investment strategy for retirees that bundles treasury inflation protected securities (tips) with a deferred * corresponding author. tel.: �1-432-552-2198; fax: �1-432-552-2174. e-mail address: haensly_p@utpb.edu (p. j. haensly) financial services review 24 (2015) 359–386 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. annuity. shankar’s strategy calls for construction of a tips ladder for a fixed number of years and purchase of a non-refundable deferred life annuity that begins payments at the end of the life of the ladder. the individual buys the tips and the deferred annuity at retirement. the ladder and deferred annuity can be designed so that the real annual payout is the same in each year of retirement. shankar (2009) makes the case that this strategy guarantees real annual withdrawal rates, eliminates the risk of financial ruin, and avoids much of the liquidity problem inherent in a non-refundable immediate life annuity. moreover, he argues that the strategy guarantees real withdrawal rates substantially higher than the 4% rule of thumb that financial advisors often recommend for retirees who hold a conventional portfolio of stocks and bonds. in this article, we contribute to the literature on personal finance by providing a more nuanced assessment of the performance of shankar’s (2009) proposed strategy. we apply a bootstrap simulation analysis under realistic assumptions about cash management, trading costs, security design, and markets. we construct simulated yield curves for tips from historical treasury yields and simulated expected inflation curves and apply these yield curves to price the tips in the ladder rather than use deterministic weighted average real yields as in shankar. we examine wider ranges of retirement ages (55 to 75 years instead of just 60 or 65 years as in shankar) and tips ladder lengths (10 to 30 years rather than just 15 or 20 years), and we consider deferred annuities that are indexed for inflation (the ideal situation) as well as those that are not (the current situation in practice). we examine the effects of cash management procedures in which payments of tips coupons and par are converted to monthly cash payouts via a money market account that earns a nominal t-bill rate. we adjust for mortality credits when valuing the deferred annuity rather than assuming a fixed number of years based on median life expectancy at the end of the tips phase as in shankar’s modeling. for each simulated investor, we determine the purchasing power of actual monthly cash payouts and apply stochastic mortality risk to determine time of death. this approach enables us to estimate the expected utility of consumption and bequests from the perspective of the investor’s retirement date. an important question is why investors are reluctant to buy an immediate inflationindexed annuity with their entire savings. as shankar (2009) observes, economists offer a variety of explanations, including unattractive design of annuity products, existence of programs such as social security that perform essentially the same function, and behavioral explanations. we apply expected utility analysis to explore how the strength of the bequest motive affects a rational investor’s choice between a non-refundable immediate life annuity and a tips ladder bundled with a non-refundable deferred life annuity. our results are in close agreement with shankar (2009) for the specific scenarios that he presents in detail. we show that, under a wider range of retirement ages and ladder periods, a tips ladder/inflation-indexed deferred annuity strategy provides a real payout rate that is modestly less than that from a corresponding immediate annuity but without the irreversible commitment while still avoiding risk of ruin. we also confirm that the target real payout rate is strongly determined by market real yields at time of retirement, where the target rate is the rate calculated in the construction of the tips ladder/deferred annuity strategy. on the other hand, we identify circumstances under which shankar’s (2009) optimistic conclusions about his proposed strategy do not hold. first, because the target real payout rate 360 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 is determined by market yields and expected inflation at time of retirement, our simulation suggests that investors at retirement may see low target real payout rates when market real yields are low. the same problem arises (and for the same reason) with immediate annuities; shankar’s proposed strategy does not eliminate this drawback. we explore which combinations of retirement age and tips ladder period offer investors the best odds of an attractive real payout rate regardless of market conditions. second, shankar’s (2009) strategy does not guarantee a fixed real payout rate. when the deferred annuity is indexed for inflation, we show that the realized real payout rate can dip modestly. during the tips payout phase, these dips are rare and transient but may be as much as 0.5%. during the deferred annuity payout phase, the retiree may face fluctuations (as much as 0.25%) when the annuity is indexed for inflation with a lag. on the other hand, when the deferred annuity is not indexed for inflation, investors who survive to the deferred annuity phase have a good chance of experiencing significantly lower realized real payout rates late in life when they are least able to redress the situation. this point is important as long as a market for inflation-indexed deferred annuities does not exist. third, we identify circumstances under which the allocation to the deferred annuity premium could be much greater than the 20 to 30% in the scenarios that shankar (2009) discusses. we explore circumstances under which this allocation is likely to be low and circumstances under which it is likely to be high. 2. literature review the literature on sustainable withdrawal rates (i.e., rates that do not lead to financial ruin) has two threads. in the earlier thread, the authors focus on passive strategies in which the asset allocation and withdrawal rate are selected and fixed at retirement. for example, cooley, hubbard, and walz (2003) examine risk of ruin for fixed withdrawal rates from an all equity portfolio. milevsky & robinson (2005) study risk of ruin when mortality risk and rates of return are stochastic. ervin, faulk, and smolira (2009) evaluate the effects of different combinations of asset allocation, lengths of savings and retirement periods, savings rate, and availability of social security income in scenarios where individuals set their withdrawal rates to smooth lifetime income. in the later thread, the authors consider dynamic strategies in which various features are adjusted as circumstances warrant. for example, stout (2008) and spitzer (2008) examine flexible withdrawal rates, while gupta, pavlik, and synn (2012) evaluate semi-passive balanced fund portfolios. neither passive nor dynamic strategies proposed in the literature guarantee a stream of retirement income that is independent of future portfolio returns and is guaranteed for life. shankar’s (2009) proposed strategy potentially can do both. little attention has been given in the literature to bundling bond ladders and deferred annuities. shankar presents the earliest systematic analysis and description of strategies that combine a tips ladder with a deferred life annuity. (a more recent article by sexauer, peskin, and cassidy (2012) proposes the same strategy.) the key to the strategy is that the investor can design a “buy and forget” portfolio by allocating savings at retirement between a suitably designed tips ladder and a deferred annuity in a manner that, in principle, locks in a real payout rate over the investor’s 361p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 remaining life. therefore, it offers the advantages of a non-refundable immediate, inflationindexed life annuity without the associated liquidity problem. bernheim (1991) concludes that bequest motives are strong for a large segment of the population. thus, the bequest motive might play a significant role in explaining why less than two percentage of retirees annuitize their retirement savings, as shankar (2009) notes. we examine the bequest motive in an attempt to ascertain the circumstances under which an investor prefers an immediate inflation-indexed annuity to a strategy composed of a tips ladder and a deferred annuity. 3. strategy design our analytical model is analogous to the one that shankar (2009) presents. however, the details of our approach are sufficiently different that we provide a complete description. 3.1. key similarities and differences as in shankar (2009), we assume that the investor uses all of savings at retirement to purchase the tips for the ladder and to pay a lump-sum premium for the deferred annuity. another similarity is that strategies with an inflation-indexed deferred annuity are designed so that the annual inflation-indexed payout is approximately the same each year, whether the payout is from the tips ladder in the first phase of retirement or from the deferred annuity in the second phase of retirement. in addition, if the deferred annuity is not indexed, then the ladder is designed so that the fixed nominal value of the annual annuity payment equals the level annual payout from the tips ladder expressed in dollars at retirement. one important difference is that we apply pension annuity factors that are discrete-time versions of the formulas developed by milevsky (2006). these formulas correctly account for the mortality risk. furthermore, in our analysis we apply inflation-indexed yields based on tips (rather than real yields) when we calculate the number of tips to use and the cost of the deferred annuity. this approach is more consistent with the pricing of tips and inflation-indexed deferred annuities, especially if the insurance company were to invest the premium in tips to match its future real cash outflows to the promised annuity payments. finally, we work with monthly rather than annual payouts. this approach facilitates the simulation of cash flows from tips as they occur throughout the year. 3.2. pension annuity factors the immediate pension annuity factor (ipaf) is the present value of a one-dollar real immediate annuity where the present value calculation takes mortality credits into account. (see section 10.6 in milevsky (2006) for a good explanation of what mortality credits are and how annuities work.) we apply the ipaf to calculate payments from the immediate, inflation-indexed annuities that we compare to the tips ladder/deferred annuity strategies. please see appendix a for how we adapt milevsky’s ipaf formula for our analysis. 362 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 we also use the ipaf to calculate the deferred pension annuity factor. the deferred pension annuity factor (dpaf*) is the present value of a one-dollar real deferred annuity where the present value calculation takes mortality credits into account. to determine the monthly real dollar annuity payment, we calculate the dpaf*, assuming discrete, real monthly payouts. in this article, we apply a discrete version of milevsky’s (2006) eq. (6.14). please see appendix a for an explanation of how we adapt milevsky’s dpaf* formula for our analysis. 3.3. construction of the tips ladder the investor’s task is to construct a tips ladder and purchase an inflation-indexed deferred annuity whose first payment is one month after the end of the life of the ladder. in our article, as in shankar (2009), we design the ladder and the annuity so that the target real annual payout is the same from each. the monthly inflation-indexed deferred annuity payment (in terms of dollars at retirement) is p*da � �1 � wtl�w�1 � fda� dpaf* , (1) where dpaf* is calculated at time of retirement, w is savings at retirement, wtl is the proportion of savings used to construct the tips ladder, and fda is the percentage of the premium paid as fees to the insurance company. (we assume that this fee is 2%.) let the tips ladder be an n-year ladder designed to support consumption for the first n years in retirement. for the first year, set aside sufficient funds from the allocation to the tips ladder to cover consumption in the first year. for each remaining year, identify a coupon tips in the market that matures before but as close as possible to the start of that year. for example, for year two in the ladder, find a coupon tips that matures as late in the first year as possible; for year three, find a coupon tips that matures as late in the second year as possible; and so forth. choosing the maturities in this fashion minimizes interest rate risk and inflation risk when the ladder has an annual structure as in shankar (2009). the par from a maturing tips is deposited in a cash account that earns the nominal return on one-month treasury bills rolled over month to month. let p*tl be the monthly payout, expressed in retirement month dollars, to be supported by the tips ladder. then p*tl � a 12 , where a is the annual inflation-indexed payout. we design the ladder and annuity so that p*tl � p*da. the ideal annual payout is a � w�1 � fda� �dpaf*/12� � �1 � fda���1 � fb�pt��11 � 1� . (2) please see appendix b for details about the solution, including how we calculate the number of tips of each maturity to buy. shankar (2009, p. 67) observes that commercially available deferred annuities are not indexed for inflation. hence, we also examine strategies where the deferred annuity is not 363p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 indexed. the construction of the tips ladder and its associated deferred annuity follow the same steps as when the annuity is indexed. in particular, the nominal deferred annuity payout is set equal to the tips ladder payout in dollars at retirement. however, because the nominal deferred annuity payout is fixed, its real value tends to decline over time. 3.4. expected utility in addition to evaluating risk of financial ruin, we also estimate expected utility to evaluate investor preferences. the general form of the constant relative risk aversion (crra) utility function in this study is u�c; �, �� � exp� �1 � ��ln�c � �� � 1 1 � � , c � 0, (3) where � is the risk aversion level, � � 1, � � 0 is a scaling factor, c is the cash flow (either consumption in the given period or the bequest at death) and u�c; 1, �� � ln�c � � , c � 0, (4) when the risk aversion level � � 1. in all cases, the absolute risk-aversion is a�c� � � u��c� u��c� � � c . (5) note that the absolute risk-aversion is scale independent. in the simulation, we use a scale factor �c � $1,000 when evaluating utility of consumption and �b � $10,000 when evaluating utility of bequests. this adjustment compresses the numerical range of the utility function and makes the results more numerically manageable. (as a result, the strength of the bequest motive, d, is implicitly conditional on the scale factors. however, we apply the same scaling factors throughout, so ordering based on expected utility is consistent.) when � � 1, a numerical problem arises in the utility of consumption and utility of bequest functions defined in eqs. (3) through (5). as c approaches 0, the utility approaches minus infinity. we mitigate this problem by replacing very low or zero dollar payouts and bequests with de minimus levels. please see appendix c for details. in the simulation, we estimate expected utility for a given strategy by averaging the realized utility for all simulated investors who use the strategy. we calculate the realized utility for a given simulated investor as u�c���, b���; �, d� � �1 � d� � t�1 t��� 1 tu�ct���; �, �c� d t���u�b���; �, �b�, (6) 364 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 where d is the strength of the bequest motive, 0 � d � 1; � is the one-month time discount factor for utility; t(�) is time of death in months after retirement for simulated life �; ct(�) is the consumption in month t after retirement, expressed in dollars at retirement (and c(�) is a vector of these values); and b(�) is the bequest at time t(�) in retirement year dollars. 4. simulation methods 4.1. markets in the simulation treasury auction schedules are determined by the financing needs of the federal government. hence, new maturities are not offered every month. the auction schedule in this simulation is similar to that followed by the u.s. treasury for most of 2010–2013. specifically, 5-year tips are auctioned in april, august, and december; 10-year tips are auctioned in january, march, may, july, september, and november; and 30-year tips are auctioned each february, june, and october. thus, in any given month, an investor at retirement must go to the secondary market to buy tips with suitable maturities for the ladder. small-scale purchases by individuals through treasury direct generally incur no purchase fees. in the simulation, investors incur transactions costs only for tips purchased in the secondary market. we adopt the simplifying assumption that all auctions, purchases, coupon dates, and maturity dates are end-of-month. tips are indexed to the non-seasonally adjusted consumer price index for all urban consumers (cpi-u). the treasury sets the reference price level for tips up for auction in a given calendar month as the cpi-u for the third preceding calendar month. for example, if the auction month is july, then the reference price level for july 1 is the april cpi-u. if the auction date is midmonth, then the official indexation lag is two and a half months. if the auction date is end-of-month, then the official indexation lag is three months. in the simulation, we set the indexation lag at three months, because we assume all auctions are end-of-month. trading in the secondary market exposes individual investors to market friction in the form of lot size restrictions and trading costs. in the secondary market the retail investor pays dealer markups, commissions, and bid-ask spreads that may be as much as 2% (e.g., see aschkenasy (2005), bullock (2005)). we assume that investors incur transactions costs of 2% of the bond price for tips bought or sold in the secondary market, but we simplify by ignoring lot size restrictions and assuming that tips may be traded in any multiple of $100 par (indexed for inflation after the auction date). 4.2. construction of inflation-indexed zero-coupon yield curves inflation-indexed tips have been issued only since 1997. to create simulated histories that are more representative of historical yield and inflation environments in the united states, in each simulated month we construct an inflation-indexed zero-coupon treasury yield curve. before the simulation begins, we construct historical nominal yield curves. our data sources for 1926–2012 are the daily constant maturity rate (cmt) database compiled by the 365p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 board of governors of the federal reserve system (2013), the fred economic data web site at the federal reserve bank of st. louis (2013), and morningstar (2013). the construction proceeds in three stages. in the first stage, we construct a discrete nominal treasury yield curve at the end of each historical month with bond equivalent yields (assuming semiannual interest) at monthly maturity intervals from one month to six months and then at six-month intervals from month six out to 30 years. while spline interpolation is preferred (and is the approach currently applied by the u.s. treasury to construct its yield curves), we apply it only for historical months february 1977 and later. for earlier historical months, the number of knot points determined by available yield data are too small for spline interpolation to work well. hence we apply linear interpolation to construct yield curves for earlier historical months. in the second stage, we convert the historical discrete nominal yield curve to a yield curve for nominal treasury zero rates by applying an iterative method to the coupon bond yields each month. (see, e.g., hull [2011, pp. 86–88] for an illustration.) in the process, we also convert discretely compounded yields to continuously compounded, annualized yields. in the third stage, we apply a natural cubic spline to interpolate for maturities in monthly intervals between the six-month maturities on the nominal zero-coupon yield curve. the end result is a historical nominal zero-coupon treasury yield curve for each historical month with yields at monthly intervals along the curve. we adapt models for expected inflation that we then apply to construct historical yields and prices for inflation-indexed bonds in the simulation. the simulation applies fama’s (1975) model for one-month estimates of expected inflation. for time horizons of one year and longer, the simulation uses a modified version of kothari and shanken’s (2004) model. in their model, the predictor variables are a short-term yield, a yield spread, a realized treasury bill real return, and inflation over the most recent month. a crucial task at the start of each time step in the simulation is construction of an expected inflation curve. at each simulated month, the simulator constructs an expected inflation curve with a maturity spectrum from one month to 30 years in one-month intervals. using the historical data drawn for the simulated month and the model coefficients estimated at the start of the investor’s life, the simulator calculates the expected inflation for one-month, one-year, two-year, and three-year horizons. for horizons less than three years, the simulator applies cubic splines to interpolate between the one-month, one-year, two-year, and three-year estimates of expected inflation. for longer horizons, the simulator sets expected inflation equal to the three-year estimate. the simulator uses the expected inflation curve to determine the zero-coupon inflation-indexed tips yield curve and to carry out bond pricing and indexing of nominal cash flows as needed for the current time step. the inflation-indexed zero-coupon treasury yield curve is determined in the following way. let yt�m� represent the annualized yield at time t on a nominal zero-coupon treasury bond that matures in m months. let y*t �m� represent the annualized yield at time t on an inflation-indexed zero-coupon treasury bond that matures in m months. a model for the relation between these two yields at t is yt�m� � �et�r̃t�m�� � rprt�m�� � �et�h̃t�m�� � rpit�m��, (7a) 366 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 y*t�m� � �et�r̃t�m�� � rprt�m�� � rplt�m�, (7b) where y et�r̃t�m�� � expected real yield at time t on a zero-coupon treasury bond that matures in m months, y rprt�m� � risk premium at time t for uncertainty about future real yield over t to t m, y et�h̃t�m�� � expected inflation at time t for the period from t to t m, y rpit�m� � risk premium at time t for uncertainty about inflation over t to t m, and y rplt�m� � risk premium at time t for liquidity of the inflation-indexed bond relative to liquidity of the nominal treasury bond, where both bonds mature in m months. we simplify by assuming away convexity issues and relative uncertainty of tax burdens on nominal versus indexed bonds. this model is consistent with the decomposition of bond yields in the literature (e.g., barnes, bodie, triest, and wang (2010); bekaert and wang (2010); kothari and shanken (2004)). the real rate of return is the sum of the expected real yield and the real yield risk premium. thus, the inflation-indexed yield in eq. (7b) equals the real rate of return plus a relative liquidity risk premium. solve eq. (7a) for the real rate of return and substitute the result into eq. (7b): y*t�m� � � yt�m� � et�h̃t�m�� � rpit�m�� � rplt�m�. (8) this equation serves as a template for constructing the inflation-indexed yield curve at each time step in the simulation, given yields yt�m� on the nominal zero-coupon treasury yield curve and the expected inflation rates et�h̃t�m��. eq. (8) requires two risk premiums: the liquidity risk premium for inflation-indexed treasury bonds, and the inflation risk premium for a nominal treasury bond with the same maturity as the corresponding indexed bond. evidence in the literature indicates that zero is a reasonable estimate of the liquidity risk premium (e.g., d’amico, kim, and wei (2008) and christensen and gillan (2011)). we model the inflation risk premium based on average risk premiums for 1990–2007 estimated by d’amico, kim, and wei (2008, figs. 2c and fig. 3c). they report an average inflation risk premium of about 0.25% for one-year maturities and about 0.75% for 10-year maturities. we apply linear interpolation to assign risk premiums as a function of maturity from one to 10 years. for maturities less than one year, we interpolate between a zero risk premium at maturity and 0.25% risk premium with one year to maturity. for maturities greater than 10 years, we assign a risk premium of 0.75%. to estimate the price of a tips in the simulation, we use the inflation-indexed zerocoupon treasury yield curve constructed in each simulated month and apply a pricing model based on evans (1998) to expected coupons and par. 4.3. outline of the bootstrap simulation procedures we initialize the simulator for each strategy and scenario with a different seed for the random number generator. hence, simulation results that we report for each strategy and scenario are statistically independent. each time that we run the simulator, it 367p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 performs 10,000 statistically independent replications, where each corresponds to the simulated lifetime of an investor. at the start of each replication, a simulation time clock initializes. in each simulated lifetime, the stationary block bootstrap works as follows. at the initial time step, the simulator draws a historical month at random from 1926 to 2012 along with the corresponding zero-coupon nominal treasury yield curve and inflation rate. each month has an equal probability that it will be drawn. at the next time step, with probability 1/�, the simulator draws the next month at random; otherwise, it draws the next consecutive historical month. (if the previous historical month is december 2012, then the simulator draws january 1926.) the expected length of a sequence of consecutive historical months is �. (the results that we report are for � � 60. test runs for � � 48 and � � 72 indicate the simulation results are robust to choice of �.) the simulation has three phases for each simulated investor: prehistory, market generation, and payout. in the prehistory phase, the simulator generates a time series of yields to begin estimating expected inflation as well as to estimate components of the bond pricing function. the simulator draws a stationary block bootstrap sample of 1,200 historical months with their respective nominal treasury yield curves and inflation rates. then the simulator estimates the coefficients of the model for expected inflation. we make the simplifying assumption that the inflation model coefficients are stationary. hence, these coefficients apply throughout the remainder of the simulated investor’s lifetime. (an alternative approach is to re-estimate the coefficients for the expected inflation curve in each simulated month based on the simulated historical data to date in the simulated history. however, the expected inflation curves are unrealistically volatile from one month to the next. moreover, the end results are generally similar to those when the expected inflation model coefficients are estimated once per lifetime, and the simulation runs an order of magnitude longer when the coefficients are re-estimated each month.) the purpose of the market generation phase is to create a simulated secondary market in tips. at each time step, information about outstanding securities in the secondary market is updated and new coupon tips may be issued, depending on the auction schedule. to determine the coupon rate for a newly issued bond, the simulator first calculates the coupon rate as if the bond were priced at par, where price equals the sum of the discounted real coupon payments and par, where discounting is back to the auction date. (we assume no arbitrage opportunities; hence the auction yield is the same as the market yield.) then, following the treasury custom at auctions for new issues, the simulator rounds the coupon rate down to the nearest one eighth of one percentage (with a floor of 0.125%) and prices the bond accordingly. the market is saturated after 360 months. the payout phase assumes that the investor has accumulated $500,000 in savings and is about to retire. in all cases, on retirement the investor uses all savings either to purchase an immediate, inflation-indexed annuity or to construct a tips ladder and buy a deferred annuity (that may or may not be indexed, depending on the scenario). the simulator uses mortality data to determine when the investor dies. we construct a unisex mortality table based on the society of actuaries rp 2014 mortality tables (2014) for healthy annuitants. 368 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 5. results 5.1. payout rate a key point to keep in mind is that the target real payout rate determined in the construction of the tips ladder/deferred annuity strategy (described earlier in section 3.3) is not necessarily the realized real payout rate. practical matters, such as cash management during the tips payout phase and how (if at all) the deferred annuity is indexed for inflation, can cause the realized rate to deviate from the target rate. we examine three general questions. first, what is the target level of the annual real payout rate, given age at retirement and length of the tips ladder? second, what is the relation between the target real payout rate and market yields at retirement? finally, how much uncertainty is there about the realized real payout rate over retirement, given market yields at retirement and choice of strategy? alternatively, does the strategy assure that a retiree will receive a fixed real payout rate? 5.1.1. target level of the real payout rate the target level of the real payout rate from a tips ladder/inflation-indexed deferred annuity strategy is close to but lower than the level of the target real payout from an inflation-indexed immediate annuity, conditional on the retirement age. please see the first three columns in table 1. the distribution of the payout rates across all simulation histories (hence, across the distribution of market yield curves at retirement) is similar, because both strategies are indexed for inflation. the payout rates on the tips ladder/deferred annuity strategy, however, are lower than for the corresponding immediate annuity, because the immediate annuity investment returns are boosted by the mortality credits whereas tips returns are not. for strategies with 10-year ladders and inflation-indexed deferred annuities, the payout distribution across all simulation histories is the closest to that for an inflation-indexed immediate annuity. for example, among investors who retire at age 65 and opt for a 10-year tips ladder, the median real payout rate is 5.45% versus a median of 5.65% for immediate annuities; the 5th-percentile rate is 4.66% versus 4.80%; and the 95th-percentile is 8.31% versus 8.80%. given the retirement age, the distribution of the real payout rates shift downward as the length of the tips ladder increases, as illustrated by the 5th-percentile, median, and 95th-percentile statistics. the reason is that the longer the tips ladder, the shorter the deferred annuity payout phase, and hence the greater the proportion of the investment that is not benefiting from mortality credits. put another way, the shorter the ladder, the more the tips ladder/deferred annuity strategy resembles an immediate annuity. table 1 also shows that, given the length of the tips ladder, the later in life that the investor retires, the greater the real payout rate. the same pattern holds for immediate annuities for the same reason: a given level of retirement wealth needs to support fewer years of payouts. when the deferred annuity is not indexed for inflation, the target real payout rate from the tips ladder/deferred annuity strategy is substantially higher than the corresponding 369p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 strategy with an indexed deferred annuity and may even exceed that from an inflationindexed immediate annuity. please see the last three columns in table 1. for example, among investors who retire at age 65 and opt for a 10-year tips ladder, the median target real payout rate is 6.70% versus a median of 5.65% for immediate annuities; the 5th-percentile rate is 5.47% versus 4.80%; and the 95th-percentile is 10.18% versus 8.80%. however, the target real payout rate holds only during the tips ladder phase. the nominal value of the deferred annuity payout is set equal (by construction) to the target real dollar payout. however, the purchasing power of these nominal payouts is eroded by inflation over the investor’s retirement. hence, the realized real payout rate during the deferred annuity payout phase can be substantially lower than the target rate for the strategy. the distribution of the target real payout rate for a tips ladder/non-indexed deferred annuity strategy is higher than that for a corresponding strategy with an inflation-indexed deferred annuity, because the non-indexed annuity is less expensive than its indexed counterpart. hence, for a given level of wealth at retirement and a given length ladder, more of that wealth can be invested in the tips ladder. because the realized real payout rate during the tips ladder phase usually is the target rate by design, this rate is higher when a greater proportion of wealth at retirement is dedicated to the ladder. table 1 distribution of target real annual payout rates (%) by strategy across simulated investor histories retirement age (years) strategy deferred (or immediate) annuity indexed for inflation deferred annuity not indexed for inflation 5%-tile median 95%-tile 5%-tile median 95%-tile 55 immediate annuity 3.44 4.29 7.40 55 10-year ladder & da 3.40 4.24 7.20 4.39 5.66 9.43 15-year ladder & da 3.35 4.17 7.04 4.15 5.24 8.41 20-year ladder & da 3.28 4.09 6.85 3.89 4.82 7.54 25-year ladder & da 3.15 3.92 6.70 3.56 4.36 7.03 30-year ladder & da 2.99 3.72 6.38 3.21 3.94 6.55 60 immediate annuity 4.01 4.87 7.91 60 10-year ladder & da 3.95 4.81 7.61 4.86 6.10 9.65 15-year ladder & da 3.86 4.63 7.42 4.55 5.57 8.54 20-year ladder & da 3.69 4.43 6.99 4.16 5.03 7.70 25-year ladder & da 3.47 4.24 6.93 3.73 4.48 7.07 65 immediate annuity 4.80 5.65 8.80 65 10-year ladder & da 4.66 5.45 8.31 5.47 6.70 10.18 15-year ladder & da 4.46 5.20 7.90 5.02 5.98 8.63 20-year ladder & da 4.14 4.86 7.46 4.45 5.28 7.80 70 immediate annuity 5.86 6.72 9.76 70 10-year ladder & da 5.57 6.34 9.03 6.24 7.37 10.47 15-year ladder & da 5.12 5.83 8.36 5.52 6.38 8.98 75 immediate annuity 7.41 8.24 11.43 75 10-year ladder & da 6.68 7.40 10.05 7.17 8.22 11.07 the payout rates are real annual payouts as a percentage of total funds at retirement. for strategies with a deferred annuity (da), the annuity payments begin one month after the tips ladder phase. when the deferred annuity is indexed for inflation, it is indexed in the same manner as tips, that is, indexed to the cpi-u with a three-month lag. 370 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 5.1.2. influence of market yields market yields at retirement determine the target payout rate in a complex fashion by eq. (2). (in the notation of this equation, the annual real payout rate is a/w.) yields at retirement enter into the calculation through the deferred pension annuity factor and the market prices of the tips in the ladder. if the deferred annuity is indexed for inflation, then the inflation-indexed zero-coupon treasury yield curve at retirement determines both dpaf*, the deferred pension annuity factor, and the vector of tips prices, p. both are in the denominator of eq. (2). the higher the yields, the smaller the value of dpaf* (see eq. [a1]), hence the greater the payout rate. furthermore, the higher the yields, the lower the tips prices (all else equal), hence the greater the payout rate. intuitively, if the inflation-indexed zero-coupon treasury yield curve shifts upward (loosely speaking, if real yields rise), then the present value of the inflation-indexed payments from the tips and deferred annuity is smaller. equivalently, a given level of wealth at retirement can purchase a stream of larger inflation-indexed payouts. in table 2, we present results from linear regressions of the target real payout rate on the value-weighted real yield of the tips ladder at retirement, conditional on retirement age and length of the tips ladder; see the first three columns for versions of the strategy that use an inflation-indexed deferred annuity. goodness of fit as measured by r2 is 90% or better. the value-weighted real yield on the tips ladder is a proxy for the yield curve. the longer the tips ladder, the better this average yield represents the entire yield curve, and the tighter table 2 linear regression of target real annual payout rate on the value-weighted real yield of the tips ladder at retirement retirement age (years) length of tips ladder (years) deferred annuity indexed for inflation deferred annuity not indexed for inflation intercept estimate slope estimate r2 intercept estimate slope estimate r2 55 10 3.54% 0.654 89.6% 4.90% 0.755 70.2% 15 3.39% 0.647 94.3% 4.40% 0.680 82.3% 20 3.22% 0.647 97.1% 3.90% 0.644 90.7% 25 3.02% 0.646 98.1% 3.44% 0.634 95.1% 30 2.81% 0.651 98.3% 3.02% 0.640 96.8% 60 10 4.07% 0.644 90.4% 5.32% 0.748 72.9% 15 3.87% 0.636 95.2% 4.74% 0.658 85.5% 20 3.62% 0.631 97.7% 4.14% 0.628 93.5% 25 3.33% 0.631 98.2% 3.60% 0.222 96.7% 65 10 4.74% 0.638 92.0% 5.90% 0.713 76.7% 15 4.44% 0.620 96.3% 5.15% 0.626 88.8% 20 4.05% 0.615 98.2% 4.39% 0.608 95.7% 70 10 5.62% 0.622 98.5% 6.60% 0.678 80.4% 15 5.10% 0.598 97.3% 5.59% 0.607 93.5% 75 10 6.71% 0.606 95.9% 7.46% 0.639 86.2% the payout rates are real annual payouts as a percentage of total funds at retirement. the deferred annuity payments begin one month after the tips ladder phase. when the deferred annuity is indexed for inflation, it is indexed in the same manner as tips, that is, indexed to the cpi-u with a three-month lag. all coefficient estimates are significantly different from zero at the one percentage level. 371p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 the relation between the payout rate and this average yield. the slope estimate is approximately the same (about 0.6) across retirement ages and ladder lengths in the simulation. hence, each increase of 1% in the value-weighted real yield of the tips ladder translates into about a 0.6% increase in the strategy payout rate, on average. when the deferred annuity is not indexed for inflation, the target payout rate is determined by eq. (2) where dpaf* is replaced by dpaf, the deferred pension annuity factor calculated with yields from the nominal zero-coupon treasury yield curve. this curve is related to the inflation-indexed version; please see eq. (7). ignoring the risk premiums, the nominal yield at each point on the yield curve equals the inflation-indexed yield plus expected inflation for that maturity. hence, the effect of the inflation-indexed treasury yields on the ideal payout rate is muddied by the expected inflation curve. nonetheless, the positive relation between the market yields at retirement (represented by the value-weighted real yield on the tips ladder) and the target real payout rate still is strong. see the last column in table 2. given the investor’s wealth at retirement (w), the deferred annuity fee (fda), and the secondary market trading fee for tips (fb), the target annuity payout for the tips ladder/ deferred annuity strategy is completely determined by the inflation-indexed zero-coupon treasury yield curve at retirement (along with the expected inflation curve at retirement, if the annuity is not indexed), conditional on the tips currently available in the market and their real coupon rates. this last factor is captured by � in eq. (2). in the simulation (as in actual markets), the coupon rates and maturities of available tips change over time. hence, to be precise, the distribution of the target real payout rates in our simulation is determined not only by the market yields at retirement but also by the specific characteristics of the tips available at retirement. 5.1.3. comparison with shankar the simulation results concerning target real payout rates are close agreement with the deterministic calculations reported by shankar (2009). shankar carries out this analysis for assumed real returns of 0%, 1%, 2%, and 3% for tips and the calculation of the deferred annuity premium. shankar’s results are for strategies with inflation-indexed deferred annuities. in table 3, we compare the payout rates reported in shankar’s table 2 with our simulation results. the level of the target real payout rates in the simulation is about the same as in shankar (2009) for comparable versions of the tips ladder/deferred annuity strategy and market real yields at retirement. for example, consider an investor who retires at age 65, set up a 20-year tips ladder, and buys the corresponding inflation-indexed deferred annuity. the statistics in the “left tail” column of table 3 correspond to real market yields at retirement of about 0%. the average real payout rate in shankar is 4.47%, while the corresponding rate from the simulation (based on the 5th-percentile statistic) is 4.14%. the statistics in the “middle” column correspond to real yields of about 1%. the average payout rate in shankar is 4.97%, while the corresponding rate from the simulation (based on the median) is 4.86%. shankar (2009, p. 54) claims that his proposed strategy “. . . would allow retirees to enjoy real withdrawal rates substantially higher than the 4% . . .” suggested by the current con372 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 census in the literature. we find that this claim is true for many but not all circumstances. specifically, the 5th-percentile statistic for the target real payout rate (assuming an inflation-indexed deferred annuity) is less than 4% when the retirement age is 55 or 60; the median target real payout rate is less than 4% for early retirement (age 55) with long tips ladders (25 years or longer). please see the first two columns in table 1. our linear regression results in table 2 show that these lower real payout rates occur when market real yields are low. nonetheless, if an investor waits at least until age 65 to retire, then the tips ladder/inflation-indexed deferred annuity strategy is highly likely to have a target real payout rate exceeding 4%, provided that the value-weighted real yield on the tips ladder is at least 0%. table 3 comparison of target real annual payout rates: simulation vs. shankar (2009) retirement age (years) length of tips ladder (years) gender source target real payout rate in retirement (%)* left tail† middle‡ right tail§ 60 20 male shankar 4.23 4.73 5.54 female shankar 3.86 4.37 5.20 unisex simulation 3.69 (5%-tile) 3.62 (e0) 4.43 (m) 4.25 (e1) 5.15 (75%-tile) 5.20 (e2.5) 65 15 male shankar 5.28 5.78 6.57 female shankar 4.73 5.23 6.04 unisex simulation 4.46 (5%-tile) 4.44 (e0) 5.20 (m) 5.06 (e1) 5.93 (75%-tile) 5.99 (e2.5) 65 20 male shankar 4.59 5.09 5.88 female shankar 4.35 4.85 5.66 unisex simulation 4.14 (5%-tile) 4.05 (e0) 4.86 (m) 4.66 (e1) 5.53 (75%-tile) 5.59 (e2.5) all strategies in this table consist of a tips ladder and an inflation-indexed deferred annuity. the deferred annuity payments begin one month after the tips ladder phase. the strategies are designed so that the target real annual payout rate is the same during the tips ladder phase and the deferred annuity phase. * shankar’s annual real payout rates in retirement are based on a deterministic analysis for assumed real returns of 0%, 1%, 2%, and 3% for tips and the calculation of the deferred annuity premium. results are from his table 2. the rates from our simulation are target real payout rates for individual investors, and the statistics are based on 10,000 replications for each strategy. † payout rates from shankar correspond to calculations based an assumed real return of 0%. target payout rates from our simulation either are the 5%-tile statistics or the expected target payout based on linear regression analysis (e0), given a 0% value-weighted average real yield on the tips ladder. (note: the 5%-tile of the value-weighted average real yield on the tips ladder is about 0.09% for 15-year ladders and about 0.14% for 20-year ladders in the simulation.) ‡ payout rates from shankar correspond to calculations based an assumed real return of 1%. target payout rates from our simulation either are the median statistics (m) or the expected target payout based on linear regression analysis (e1), given a 1% value-weighted average real yield on the tips ladder. (note: the median of the value-weighted average real yield on the tips ladder is about 1.15% for 15-year ladders and about 1.32% for 20-year ladders in the simulation.) § payout rates from shankar correspond to the simple average of the results based assumed real returns of 2% and 3%. target payout rates from our simulation either are the 75%-tile statistics or the expected target payout based on linear regression analysis (e2.5), given a 2.5% value-weighted average real yield on the tips ladder. (note: the 75%-tile of the value-weighted average real yield on the tips ladder is about 2.47% for 15-year ladders and about 2.55% for 20-year ladders in the simulation.) 373p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 5.1.4. shortfall risk an important claim in shankar (2009) as well as in sexauer, peskin, and cassidy (2012) is that the tips ladder/deferred annuity strategy guarantees a fixed real payout rate. we analyze this claim by examining shortfall risk over retirement relative to the target real annual payout rate. specifically, we examine the difference between the target real annual payout rate and the lowest realized 12-month real payout rate. 5.1.4.1. shortfall risk when the deferred annuity is indexed. first, consider tips ladder/ deferred annuity strategies in which the annuity is indexed for inflation. we find that the shortfall risk is likely to be small. please see the first three columns in panel a of table 4. the median shortfall in the real payout rate is on the order of about 0.1% across all strategies, and the 95th-percentile shortfall is about 0.75% or less. in real dollar terms, if the ideal payout ratio is 4% (corresponding to annual real payout of $20,000 when wealth at retirement is $500,000), then a shortfall of 0.75% is about 19% of the annual payout (or $3,750 in purchasing power). over the tips ladder phase, however, we find the shortfall is a rare, transient event. depending on the retirement age and length of the ladder, the odds of a shortfall are about 5% to 15%. for almost all simulated investors who experience a shortfall, the shortfall occurs only once (specifically, in only one ladder year). for example, among investors who retire at age 65 and elect a 20-year tips ladder, 86.78% experience no shortfall during the ladder period (i.e., the realized real payout rate equals the target rate); 12.74% experience one month with no payout; and 0.48% experience two months with no payout. these relatively small fluctuations are a consequence of the cash management strategy in the simulation. specifically, during the life of the ladder, the tips for a given annual rung in the ladder are selected to mature shortly before that year. the proceeds are held in a cash account that earns a nominal one-month nominal t-bill return. hence, fluctuations in the t-bill rate relative to the realized inflation rate occasionally cause the retiree to come up short at the end of a year in the tips ladder phase. over the deferred annuity payout period, fluctuations in the real payout rate are much more common but also are small and on the order of 0.25% or less. please see the last three columns of panel a in table 4. in the simulation, when deferred annuities are inflationindexed, they are indexed to the cpi-u with a three-month lag. we calculate the real value of each realized payout in terms of retirement date dollars. as a consequence, the indexation lag induces relatively small fluctuations in the real payout rate during the deferred annuity payout phase. in some cases, the fluctuations may even be in favor of the retiree; see the 5th-percentile column. when this shortfall statistic is negative, it means that the realized real payout rate in all rolling 12-month periods exceeded the target payout rate for at least five percent of investors. these fluctuations also affect inflation-indexed immediate annuities. in our simulations, the median range for immediate annuity payouts (measured as the difference between the highest and lowest 12-month rolling realized real payout rates) is about 0.25% and the 95th-percentile range is about 0.5% or less. in short, any annuity that is indexed with a lag will not perfectly track changes in purchasing power as measured by changes in the cpi-u. 374 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 5.1.4.2. shortfall risk when the deferred annuity is not indexed. the story is much worse when the deferred annuity is not indexed for inflation. the shortfall over retirement can be large. please see panel b in table 4. over the tips ladder phase, the shortfall risk is low, and magnitude of the shortfall is about the same as for corresponding scenarios when the deferred annuity is indexed. however, if the retiree survives to the deferred annuity phase, then the shortfall can be dramatic. the median shortfall is on the order of about half of the median target real payout ratio. this shortfall is not transient. although deflationary periods can occur in the simulation as over the historical period from which our expected inflation curves are estimated, the table 4 shortfall risk: target real annual payout rate minus lowest realized rolling 12-month real payout rate (%) retirement age (years) tips ladder (years) over retirement over life of tips ladder over deferred annuity payout 5%-tile median 95%-tile 5%-tile median 95%-tile 5%-tile median 95%-tile panel a: strategies with inflation-indexed deferred annuity 55 10 0.00 0.11 0.42 0.00 0.00 0.40 �0.03 0.09 0.22 15 0.00 0.10 0.61 0.00 0.00 0.56 �0.04 0.08 0.20 20 0.00 0.08 0.60 0.00 0.00 0.56 �0.05 0.06 0.19 25 0.00 0.05 0.57 0.00 0.00 0.53 �0.07 0.05 0.18 30 0.00 0.00 0.50 0.00 0.00 0.47 �0.08 0.04 0.17 60 10 0.00 0.12 0.47 0.00 0.00 0.44 �0.05 0.09 0.23 15 0.00 0.09 0.67 0.00 0.00 0.61 �0.05 0.07 0.21 20 0.00 0.05 0.63 0.00 0.00 0.58 �0.07 0.06 0.20 25 0.00 0.01 0.57 0.00 0.00 0.54 �0.09 0.04 0.18 65 10 0.00 0.12 0.53 0.00 0.00 0.50 �0.06 0.08 0.25 15 0.00 0.07 0.73 0.00 0.00 0.67 �0.08 0.07 0.23 20 0.00 0.01 0.64 0.00 0.00 0.59 �0.10 0.05 0.20 70 10 0.00 0.11 0.61 0.00 0.00 0.57 �0.09 0.08 0.28 15 0.00 0.02 0.75 0.00 0.00 0.68 �0.12 0.06 0.24 75 10 0.00 0.08 0.67 0.00 0.00 0.64 �0.13 0.07 0.30 panel b: strategies with deferred annuity that is not indexed for inflation 55 10 0.00 3.20 6.15 0.00 0.00 0.48 1.12 3.33 6.22 15 0.00 2.98 5.44 0.00 0.00 0.66 1.18 3.17 5.56 20 0.00 2.65 4.87 0.00 0.00 0.62 1.08 2.98 5.07 25 0.00 2.22 4.46 0.00 0.00 0.56 1.12 2.78 4.82 30 0.00 0.70 3.97 0.00 0.00 0.49 0.97 2.60 4.51 60 10 0.00 3.10 5.99 0.00 0.00 0.52 0.83 3.29 6.08 15 0.00 2.79 5.24 0.00 0.00 0.70 0.94 3.10 5.36 20 0.00 2.36 4.68 0.00 0.00 0.62 0.89 2.95 5.02 25 0.00 1.04 4.16 0.00 0.00 0.55 0.99 2.78 4.68 65 10 0.00 3.01 5.93 0.00 0.00 0.57 0.75 3.33 6.10 15 0.00 2.50 5.05 0.00 0.00 0.72 0.69 3.10 5.29 20 0.00 1.38 4.44 0.00 0.00 0.63 0.66 2.92 4.83 70 10 0.00 2.73 5.72 0.00 0.00 0.63 0.51 3.31 5.94 15 0.00 1.74 4.87 0.00 0.00 0.74 0.50 3.10 5.26 75 10 0.00 2.07 5.60 0.00 0.00 0.69 0.12 3.25 5.96 all strategies consist of a tips ladder and a deferred annuity. annuity payments begin one month after the tips ladder phase. when the annuity is indexed for inflation, it is indexed in the same manner as tips, that is, to the cpi-u with a three-month lag. payout rates are real annual payouts as a percentage of total funds at retirement. the statistics in this table are conditional on the retiree collecting at least 12 payments in the respective period. 375p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 general trend in purchasing power in the united states over the last century has been downward. hence, over the tips ladder phase of 10 to 30 years, purchasing power of the nominal deferred annuity payment is likely to be reduced significantly. this trend generally continues during the annuity payout. hence, the shortfall statistics in panel b of table 4 represent shortfall at or near the end of the retiree’s life. 5.2. initial allocation and reluctance to annuitize the earlier the age when the deferred annuity payout begins and the shorter the life of the tips ladder, the higher the allocation at retirement to the deferred annuity premium. this relation generally holds true whether the deferred annuity is indexed for inflation or not. please see panel a of table 5. these results are consistent with common financial sense. the earlier the age when a deferred annuity begins its payout, the more payments the investor is likely to collect. hence the annuity is more expensive. furthermore, conditional on age of retirement, the shorter the life of the tips ladder, the longer the period in retirement that the investor expects to require payouts from the deferred annuity. hence, the deferred annuity is more expensive. thus, the allocation at retirement to the deferred annuity premium is greater. see panel b in table 5. we also examine the relation between the allocation to the deferred annuity and the valueweighted real yield on the tips ladder at retirement. the latter serves as proxy for market yields at retirement. conditional on age of retirement and length of the tips ladder in years, the allocation to the deferred annuity generally declines the higher the market yields. for example, see fig. 1. again, this relation makes financial sense. the lower the discount rate applied to real dollar payouts from the deferred annuity, the greater the present value of the deferred annuity. the distribution of market yield curves at retirement across simulated retiree histories accounts for the range of initial allocations to the deferred annuity, given retirement age and length of the tips ladder. please see the 5thand 95th-percentile statistics in table 5. an important claim in shankar (2009) is that a tips ladder/deferred annuity strategy significantly reduces the liquidity problem present when buying a non-refundable immediate life annuity. shankar (p. 58) states, “ . . . the ipra strategy allows the bulk of the retirement portfolio to be held in tips and requires only a small fraction of the retirement savings to be used to pay the non-refundable longevity policy premium; this mitigates the ‘large irreversible commitment’ inherent in buying an immediate lifetime annuity with the entire retirement savings.” however, we determine that for shorter ladders and low real market yields at retirement, the investor may need to allocate 50% or more of savings at retirement to the deferred annuity premium. please see table 5. nevertheless, the simulation results show that the investor can keep the allocation to the deferred annuity low regardless of market real yields at retirement if the investor chooses a long tips ladder. the earlier that the investor retires, the longer the tips ladder needs to be. for example, an investor who wants to cap the allocation to a non-indexed deferred annuity to about 25%, regardless of market rates at retirement, needs a 20-year ladder if retiring at age 55 but only a 15-year ladder if retiring at age 65. 376 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 5.3. the bequest motive and preferences for retirement strategies to evaluate the effect of the strength of the bequest motive, d, on reluctance to commit all savings at retirement to an immediate annuity, we analyze expected utility at several table 5 initial asset allocation for strategies with tips ladder and deferred annuity retirement age (yrs.) tips ladder (yrs.) allocation to inflation-indexed deferred annuity in shankar age when annuity payout begins (years) allocation to inflationindexed deferred annuity in the simulation allocation to non-indexed deferred annuity in the simulation 5%-tile median 95%-tile 5%-tile median 95%-tile panel a: scenarios ranked by median allocation to the inflation-indexed deferred annuity in the simulation 55 30 85 0.037 0.084 0.117 0.007 0.034 0.062 60 25 85 0.051 0.104 0.140 0.013 0.048 0.082 65 20 male: 0.073; female: 0.116 85 0.075 0.135 0.172 0.023 0.073 0.112 55 25 80 0.081 0.165 0.220 0.020 0.077 0.131 70 15 85 0.120 0.188 0.228 0.048 0.118 0.165 60 20 male: 0.138; female: 0.203 80 0.119 0.207 0.264 0.035 0.114 0.173 75 10 85 0.207 0.291 0.327 0.115 0.216 0.271 55 20 75 0.153 0.275 0.347 0.046 0.152 0.231 65 15 male: 0.190; female: 0.267 80 0.173 0.276 0.332 0.071 0.177 0.243 60 15 75 0.222 0.354 0.424 0.090 0.229 0.313 70 10 80 0.285 0.393 0.442 0.158 0.295 0.367 55 15 70 0.264 0.419 0.497 0.105 0.274 0.372 65 10 75 0.344 0.477 0.534 0.188 0.358 0.445 60 10 70 0.397 0.540 0.605 0.219 0.414 0.507 55 10 65 0.432 0.592 0.660 0.236 0.456 0.555 panel b: scenarios ranked by age at retirement and length of the tips ladder 55 10 65 0.432 0.592 0.660 0.236 0.456 0.555 15 70 0.264 0.419 0.497 0.105 0.274 0.372 20 75 0.153 0.275 0.347 0.046 0.152 0.231 25 80 0.081 0.165 0.220 0.020 0.077 0.131 30 85 0.037 0.084 0.117 0.007 0.034 0.062 60 10 70 0.397 0.540 0.605 0.219 0.414 0.507 15 75 0.222 0.354 0.424 0.090 0.229 0.313 20 male: 0.138; female: 0.203 80 0.119 0.207 0.264 0.035 0.114 0.173 25 85 0.051 0.104 0.140 0.013 0.048 0.082 65 10 75 0.344 0.477 0.534 0.188 0.358 0.445 15 male: 0.190; female: 0.267 80 0.173 0.276 0.332 0.071 0.177 0.243 20 male: 0.073; female: 0.116 85 0.075 0.135 0.172 0.023 0.073 0.112 70 10 80 0.285 0.393 0.442 0.158 0.295 0.367 15 85 0.120 0.188 0.228 0.048 0.118 0.165 75 10 85 0.207 0.291 0.327 0.115 0.216 0.271 simulation results are based on unisex mortality risk, whereas shankar (2009) evaluates scenarios separately for male and female mortality risk. results from shankar’s table 2 are based on a deterministic analysis for an assumed real return of 1% for tips and the calculation of the deferred annuity premium. by comparison, the median value-weighted average yield on tips ladders in the simulation is about 1.15% for 15-year ladders and about 1.32% for 20-year ladders. 377p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 levels of d ranging from zero (consumption has utility but bequests do not) to one (bequests have utility but consumption does not). when d � 0, we hypothesize that the investor will always prefer the immediate annuity to a strategy of a tips ladder combined with a deferred annuity. at the other extreme, when d � 1, we hypothesize that the investor will always prefer a strategy of a tips ladder combined with a deferred annuity over an immediate annuity. the simulation results are consistent with both hypotheses. at the extremes for d, the expected utilities of the inflation-indexed immediate annuity and the tips ladder/inflationindexed deferred annuity strategy are significantly different and differ in the predicted direction. these results hold for all variations that we examined of the tips ladder/deferred annuity strategy, where annuities (immediate and deferred) are indexed to the cpi-u in the same manner as tips, and for all levels of risk aversion from 0.5 to 5.0 in our utility model. for intermediate values of the bequest motive, d, the general pattern is that the higher the level of risk aversion, the lower the cross-over value of d at which the investor switches from preference for an inflation-indexed immediate annuity to preference for a tips ladder/ inflation-indexed deferred annuity strategy. please see table 6. the expected utility of a bequest contributes more to total expected utility in two ways. first, as d increases, greater weight is given to it; see eq. (6). second, for a given retirement age and length of the tips ladder, the expected utility of a bequest grows exponentially as the risk aversion increases. fig. 1. this figure plots the allocation to the inflation-indexed deferred annuity for each simulated investor against the market value-weighted real yield on the tips ladder; both values are at the time of retirement. for this figure, all investors retire at age 65 and survive at least 12 months in retirement. at retirement, each investor buys the tips for a 20-year ladder and pays a premium on a deferred annuity that begins payouts one month after the 20-year ladder phase ends. 378 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 please see column 6 in table 7. that is, the bequest is relatively more valuable the greater the level of risk aversion under the utility models in our analysis. in addition, for risk aversion levels greater than one, the crossover value of d tends to fall the older the investor is when retiring, all else equal (specifically, for ladders of the same length). for example, compare strategies with 10-year tips ladders in table 6. this result is consistent with the fact than an older retiree has a shorter life expectancy and thus is more likely to be able to leave a bequest in the form of the balance of the tips ladder at death. the fact that most investors in the real world are reluctant to commit all of their savings at retirement to immediate annuities is consistent with bernheim’s (1991) conclusion that bequest motives are strong for a large segment of the population. our results in table 6 illustrate a more subtle point. the higher the level of risk aversion and the older the investor at retirement, the weaker the bequest motive needs to be for the investor to prefer an investment strategy that avoids committing all of the investor’s savings to a non-refundable immediate life annuity. given the current unavailability of inflation-indexed deferred annuities, an interesting question is whether the same observations hold for tips ladder/deferred annuity strategies table 6 preference for immediate annuity vs. tips ladder/deferred annuity strategy in terms of strength of bequest motive (d); cross-over value for d where preference switches from the immediate annuity to the tips ladder/deferred annuity retirement age (years) length of tips ladder (years) level of risk aversion (�) 0.5 1.0 2.0 3.0 4.0 5.0 55 10 0.95 0.95 0.90 0.65 0.65 0.55 55 15 0.95 0.95 0.85 0.65 0.65 0.65 55 20 0.95 0.95 0.85 0.65 0.55 0.55 55 25 0.95 0.95 0.85 0.65 0.45 0.45 55 30 0.95 0.95 0.85 0.55 0.35 0.35 60 10 0.95 0.95 0.75 0.55 0.55 0.55 60 15 0.95 0.95 0.85 0.55 0.55 0.55 60 20 0.95 0.95 0.85 0.55 0.45 0.45 60 25 0.95 0.95 0.85 0.45 0.35 0.35 65 10 0.95 0.95 0.75 0.45 0.45 0.35 65 15 0.95 0.95 0.75 0.45 0.35 0.35 65 20 0.95 0.95 0.85 0.50 0.25 0.25 70 10 0.95 0.95 0.65 0.40 0.25 0.25 70 15 0.95 0.95 0.80 0.40 0.25 0.25 75 10 0.95 0.95 0.70 0.30 0.15 0.15 the comparison is between inflation-indexed immediate annuities and strategies with a fixed-life tips ladder and an inflation-indexed deferred annuity that begins payouts after the ladder phase. all annuities are indexed to the cpi-u with a three-month lag, analogous to tips. we compare the expected utilities for the immediate annuity and the ladder/deferred annuity strategies for d, the value of the strength of the bequest motive, where d � 0.0, 0.1, 0.2, . . ., 1.0. if d � 0, then investors gain utility only from consumption and not from bequests; if d � 1, then investors gain all utility from bequests and none from consumption. the cross-over point is the midpoint between the highest value of d for which the investor prefers the immediate annuity to the ladder/ deferred annuity based on expected utility (i.e., the expected utility is significantly higher for the immediate annuity) and the lowest value of d for which the investor prefers the ladder/deferred annuity. to test for significance, we calculate the version of the wilcoxon statistic for the difference in expected values, where average rank replaces tied ranks. we define the difference as significant if the two-sided p-value is less than 0.10. 379p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 when the deferred annuity is not indexed. not surprisingly, the crossover point for d frequently is higher, because the real payout ratio falls over the course of the deferred annuity phase. however, the general pattern illustrated in table 6 still holds, because the tips ladder is the source of funds for a bequest. while the non-indexed deferred annuity has a lower real payout than either the tips ladder or a corresponding inflation-indexed deferred annuity, this payout is in the late phase of retirement, and the retiree might not even survive to that phase. 6. conclusions the ideal solution for a retiree is a source of income that presents no risk of financial ruin, protects against loss of purchasing power, shields against longevity risk, and provides a target standard of living. in theory, a strategy that combines a fixed-term tips ladder with an inflation-indexed deferred annuity addresses the first three concerns. we show that shankar’s (2009) claims about the advantages of a tips ladder/deferred strategy are valid under many circumstances but not all. the strategy performs best if the deferred annuity is indexed for inflation. because the target real payout rate is determined by the market real yields at retirement, some combinations of retirement age and length of the tips ladder have payout rates less than 4% when yields are very low. to assure a real payout rate significantly above 4%, we show that the investor should wait to retire until age 65 or later. we also find that when we make realistic assumptions about cash management methods during the tips ladder phase and the indexation lag of the deferred annuity, the realized payout rate for an individual investor can fluctuate over retirement. however, these fluctuations usually are small enough not be catastrophic, rarely occur during the tips ladder phase, and may be above as well as below the target payout rate during the deferred annuity phase (and, in any case, are no worse than for a retiree with an inflation-indexed immediate annuity). table 7 expected utilities at different levels of risk aversion; retirement at age 65 risk aversion (�) immediate inflation-indexed annuity 20-year tips ladder with inflation-indexed deferred annuity expected utility of consumption de minimus expected utility of bequest* expected utility of consumption expected utility of bequest absolute expected utility of bequest† 0.5 205.52 �0.76 164.04 1.56 2.32 1 160.57 �1.27 132.93 0.23 1.50 2 104.57 �4.98 91.09 �1.99 2.99 3 73.38 �27.41 61.67 �12.24 15.16 4 54.53 �184.37 13.25 �83.14 101.23 5 40.76 �1384.03 �244.33 �624.42 759.61 * the immediate annuity strategy uses all cash at retirement to purchase an irrevocable immediate life annuity. hence, no cash is available for a bequest. however, when we calculate utilities, we set de minimus values on monthly consumption and the bequest at death to avoid extreme negative values when the consumption or bequest are tiny. thus, the “expected utility of bequest” for an immediate annuity is the average of the utility of the de minimus bequest discounted back to retirement at the inflation rates for each investor’s history. † the absolute expected utility of bequest is the expected utility of bequest minus the de minimus value that corresponds to no bequest. 380 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 a major advantage that a tips ladder/deferred annuity strategy has over a non-refundable immediate life annuity is that only part of the investor’s wealth at retirement is irrevocably allocated to an annuity. because of investors’ reluctance to annuitize, this feature is attractive. however, if market real yields are low at retirement, then the allocation to the deferred annuity premium could be much greater than 30%. for example, if an investor retires at age 65, sets up a 10-year tips ladder, and buys a non-indexed deferred annuity, then the allocation to the deferred annuity might be 19% or lower (if market real yields are high) or 44% or higher (if market real yields are low). nonetheless, the investor has an excellent chance of keeping the deferred annuity premium below 30% of wealth at retirement if the investor builds a sufficiently long ladder. on the other hand, if the deferred annuity is not indexed for inflation, then the realized real payout rate over the deferred annuity phase can decline substantially at a point in life when the retiree is least able to compensate. unfortunately, as shankar concedes, a market does not yet exist for inflation-indexed deferred annuities. moreover, the difficulty in long-term forecasting of inflation seems likely to continue to deter insurance companies from offering such a product. finally, we explore the relation between the strength of the bequest motive and preference for an inflation-indexed immediate annuity versus a tips ladder/deferred annuity strategy. our results are consistent with evidence in the literature that the bequest motive is strong. if investors make choices based on expected utility, then they prefer an immediate annuity when strength of the bequest motive is low, and they prefer the tips ladder/deferred annuity when strength of the bequest motive is high. moreover, the greater the investor’s risk aversion, the more likely the investor prefers the tips ladder/deferred annuity for a given level of strength of the bequest motive. these results suggest that the tips ladder/deferred annuity strategy may have considerable appeal to individual investors who are reluctant to annuitize their life’s savings. appendix a: pension annuity factors to determine the monthly real dollar annuity payment, first calculate the immediate pension annuity factor (ipaf), assuming discrete, real monthly payouts. this article applies a discrete version of milevsky’s (2006) eq. (6.3). in the discrete version, the stochastic present value of a pension annuity is ax�d̃x� � � i�1 d̃x exp�� y*t� x��i� i 12� , (a1) where y*t� x��i� is the annualized inflation-indexed yield at time t(x) on a zero-coupon, default-risk free tips that matures in i months; t(x) is the date on which the investor is age x (in months); and d̃x is the random number of months until death, given that the investor is alive at age x. mortality index tables generally are constructed so that p�d̃x � i�current age x� � mi� x � i�/mi� x�, (a2) where mi(x) is the mortality index value at age x (in months). thus, 381p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 p�d̃x � i�current age x� � 1 � mi� x � i�/mi� x�. (a3) when d̃x is a discrete, integer-valued random variable, p�d̃x � d�current age x� � p�d̃x � d�current age x� � p�d̃x � d � 1�current age x�. (a4) thus, p�d̃x � d�current age x� � �mi� x � d � 1� � mi� x � d��/mi� x�. (a5) the ipaf is defined as a� x � � d�1 d� x� ax�d� p�d̃x � d�current age x�, (a6) where d(x) is the maximum possible months until death in the mortality table, given that the investor is alive at age x. the deferred pension annuity factor (dpaf*) is the present value of a one-dollar real deferred annuity where the present value calculation takes mortality credits into account. to determine the monthly real dollar annuity payment, first calculate the dpaf*, assuming discrete, real monthly payouts. this article applies a discrete version of milevsky’s (2006) eq. (6.14). in the discrete version, the stochastic present value of a deferred pension annuity is a� x,u � a� x up�d̃x � u�current age x�exp� � y*t� x��u� u 12� , (a7) where the first payment of the deferred annuity is u 1 months after the investor turns age x (in months); a�x u is defined by eq. (a6), where x u replaces x; and the probability is defined by eq. (a2). we assume that the deferred annuity is indexed to inflation in the same manner as tips. in particular, the inflation index is the consumer price index for all urban consumers (cpi-u), non-seasonally adjusted, and the annuity is indexed with the same lag as tips. we also analyze the strategy when the deferred annuity is not indexed. in that case, we carry out the calculations for the deferred pension annuity factor, dpaf, with yt�x��i�, the annualized nominal yield at time t(x) on a zero-coupon, default-risk free treasury bond that matures in i months. the fixed nominal value of the deferred annuity payment is set equal to the time of retirement dollar value of the level inflation-indexed payouts from the tips ladder. appendix b: construction of the tips ladder let i,n be the sum of the inflation-indexed cash flows to be paid in year i by the bond that matures in ladder year n. i,n is expressed in retirement date dollars. when i � n, i,n is the sum of the two coupon payments to be paid in ladder year i. when i � n, i,n is the sum of the one or two coupon payments to be paid in ladder year n and the par to be paid at maturity. when i � n, i,n � 0. thus, in year k, the total of the payments from the tips is 382 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 ak � � j�k n�1�k, jnj, k � 1, 2, . . . , n � 1, (b1) where nj is the number of tips that mature in ladder year j, n is number of years in the ladder, and ak is expressed in terms of retirement date dollars. (keep in mind that tips payments in ladder year k finance consumption in ladder year k � 1.) consistent with the financial planning paradigm of consumption smoothing (see, e.g., kotlikoff (2007)), we impose the constraint that ak � a for all years k. for convenience, we express the system of equations in (b1) in matrix notation: �n � a1, (b2) where � is an n-1 by n-1 upper triangular array. by design, this array is non-singular. hence we can solve for the number of bonds of each maturity: n � a��11. (b3) total cost of the ladder is wtlw � a�1 � fb�pt��11 � a, (b4) where wtl is the allocation of wealth w at retirement to the tips ladder; fb is the transactions cost paid to buy tips, expressed as a percentage of price; p is a vector of prices for the tips maturing in years k � 1, 2, . . ., n 1 of the ladder (where price is dollars per $100 par); and the second term, a, on the right-hand side is the cash set aside for consumption in the first year of the ladder. at this stage in the calculation, we simplify by assuming that the transactions cost, fb, is the same rate for all tips. in the simulation, once the investor has determined how many of each tips issue to buy, the investor then pays transactions fees that depend on the venue: 2% of the market price if purchased in the secondary market, and no fees if purchased at an auction. let p*tl be the monthly payout, expressed in retirement month dollars, to be supported by the tips ladder. then p*tl � a 12 . we design the ladder and annuity so that p*tl � p*da, where p*da � �1 � wtl�w�1 � fda� dpaf* , (b5) and fda is the percentage of the total annuity premium paid as a transactions fee to the issuer. hence, a 12 � �1 � wtl�w�1 � fda� dpaf* . (b6) solve for wtl in eq. (b4) and substitute into eq. (b6): 383p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 a 12 � �1 � ��1 � fb�pt��11�1)(a/w))w(1�fda) dpaf* . (b7) after solving for a and rearranging terms, we have a � w�1 � fda� �dpaf*/12� � �1 � fda���1 � fb�pt��11 � 1� . (b8) to solve for the number of tips for each maturity, substitute the solution from eq. (b8) into eq. (b3). in the simulation, we round down to the nearest whole number of tips, then calculate the actual cost of constructing the ladder, assuming that an amount of cash equal to a is set aside for the consumption in the first year. we then recalculate the actual weight on the ladder, wtl, by dividing actual cost by w. substituting the actual weight back into eq. (b5) gives the actual monthly deferred annuity payout, p*da, in retirement year dollars. because the investor cannot purchase fractional tips, the target real payout from the tips ladder and the target real payout from the deferred annuity will differ. however, because we assume that the investor’s life savings is $500,000, this difference is modest and does not materially affect conclusions about the strategy. appendix c: utility functions when � � 1, a numerical problem arises in the utility of consumption and utility of bequest functions defined in eqs. (3) through (5). as c approaches 0, the utility approaches minus infinity. in this study, consumption in a given month might be zero, and the bequest might be small or zero. it might seem plausible simply to omit calculation of the utility for that specific month (or for the bequest). however, doing so implicitly sets the utility equal to zero. at a month when c � �c or b � �b, the utility is negative. in that instance, not calculating utility is equivalent to setting a floor on consumption or a floor on bequests that equals the scale factor. on the other hand, if the analysis allows low or zero values for consumption or the bequest, then the utility calculation may explode or produce an undefined result. our solution is to set a floor lower than the cutoff by an order of magnitude. specifically, the cutoff for consumption is �c � $100, and the cutoff for bequests is �b � $1,000. that is, the utility function for consumption in period t is u�c; �, �� � � exp��1 � ��ln� c �c �� � 1 1 � � , c � �c exp� �1 � ��ln��c �c �� � 1 1 � � , c �c (c1) when risk aversion � � 1, and 384 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 u�c; �, �� � � ln� c �c � , c � �c ln��c �c � , c � �c (c2) when � � 1. the analogous definitions apply to utility of bequests. in effect, we implicitly assume a de minimis level of monthly consumption and bequest at death. a consequence of defining utility in this manner is that realized utility for a given simulated investor is biased upward if any payout in retirement or the bequest falls below the corresponding cutoffs. the magnitude of the bias depends on frequency and magnitude of events where ct � �c and/or the bequest b � �b. when we compare estimates of expected utility for different strategies and scenarios, the range of expected utilities may be compressed because of this bias. thus, tests for significance of differences in expected utility might not be significant when, in fact, they are, that is, the bias is toward the null hypothesis of no difference. hence, the tests are more conservative in rejecting the null hypothesis of no difference in expected utility. references aschkenasy, j. (2005). staggering choices: many clients are using bond ladders to climb beyond interest-rate risk and inflation risk-but not all ladders are created equal. financial planning, 1. barnes, m. l., bodie, z., triest, r. k., & wang, j. c. (2010). a tips scorecard: are they accomplishing their objectives? financial analysts journal, 66, 68–84. bekaert, g., & wang, x. s. (2010). inflation risk and the inflation risk premium. economic policy, 25, 755–806. bernheim, b. d. (1991). how strong are bequest motives? evidence based on estimates of the demand for life insurance and annuities. journal of political economy, 99, 899–927. board of governors of the federal reserve system. (2013). economic research & data: statistical releases and historical data. retrieved from http://www.federalreserve.gov/releases/h15/data.htm#top/ bullock, n. (2005, january 23). smartmoney: bond ladders go online. the wall street journal (eastern edition), p. 3. retrieved http://proquest.umi.com/ christensen, j., & gillan, j. (2011). tips liquidity, breakeven inflation, and inflation expectations. frbsf economic letter, 19, 1–5. cooley, p. l., hubbard, c. m., & walz, d. t. (2003). a comparative analysis of retirement portfolio success rates: simulation versus overlapping periods. financial services review, 12, 115–128. d’amico, s., kim, d. h., & wei, m. (2008). tips from tips: the informational content of treasury inflationprotected security prices. the federal reserve board finance and economics discussion series, 30, 1–64. ervin, d. m., faulk, g. k., & smolira, j. c. (2009). the impact of asset allocation, savings, and retirement horizons, savings rates, and social security income in retirement planning: a monte carlo analysis. financial services review, 18, 313–331. evans, m. d. (1998). real rates, expected inflation, and inflation risk premia. the journal of finance, 53, 187–218. fama, e. f. (1975). short-term interest rates as predictors of inflation. american economic review, 65, 269–282. federal reserve bank of st. louis. (2013). fred economic data. retrieved from http://research.stlouisfed.org/ fred2/categories/116/ gupta, n. j., pavlik, r., & synn, w. (2012). adding “value” to sustainable post-retirement portfolios. financial services review, 21, 19–33. hull, j. c. (2011). fundamentals of futures and options markets, 7th ed. new york, ny: prentice hall. 385p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 kothari, s. p., & shanken, j. (2004). asset allocation with inflation-protected bonds. financial analysts journal, 60, 54–70. kotlikoff, l. j. (2007). is conventional financial planning good for your financial health? in z. bodie, d. mcleavey, & l. b. siegel (eds.), the future of life-cycle saving and investing (pp. 55–71). charlottesville, va: the research foundation of cfa institute. milevsky, m. a., & robinson, c. (2005). a sustainable spending rate without simulation. financial analysts journal, 61, 89–100. milevsky, m. a. (2006). the calculus of retirement income: financial models for pension annuities and life insurance. new york, ny: cambridge university press. morningstar. (2013). ibbotson® sbbi® 2012 classic yearbook: market results for stocks, bonds, bills, and inflation 1926–2012. chicago, il: morningstar inc. sexauer, s. c., peskin, m. w., & cassidy, d. (2012). making retirement income last a lifetime. financial analysts journal, 68, 74–84. shankar, s. g. (2009). a new strategy to guarantee retirement income using tips and longevity insurance. financial services review, 18, 53–68. society of actuaries. (2014). rp-2014 mortality tables (exposure draft). schaumburg, il: society of actuaries. spitzer, j. j. (2008). retirement withdrawals: an analysis of the benefits of periodic “midcourse” adjustments. financial services review, 17, 17–29. stout, r. g. (2008). stochastic optimization of retirement portfolio asset allocations and withdrawals. financial services review, 17, 1–15. 386 p.j. haensly, k.p. pai / financial services review 24 (2015) 359–386 catastrophe (cat) bonds: risk offsets with diversification and high returns richard j. kisha,* aperella department of finance, lehigh university, 621 taylor street, bethlehem, pa 18015, usa abstract catastrophe bonds, a relatively new entry into the bond market, are a form of reinsurance in which insurance firms are able to offset the financial risks from both natural and man-made catastrophes. although the primary offset is within the reinsurance market, starting in the 1990s insurance firms started using the financial markets to offset catastrophe risks. anecdotal evidence shows that the entry of cat bonds made the reinsurance market more efficient and allowed investors an opportunity to participate in what has been a very profitable investment opportunity. our analysis shows that on average, cat bonds have generated high returns but with the advantage of diversification when compared with similarly rated corporate debt. thus, cat bonds are a viable investment option within a diversified portfolio. © 2016 academy of financial services. all rights reserved. jel classification: g11 keywords: cat bonds; reinsurance; catastrophe 1. introduction catastrophes, both natural (hurricanes, earthquakes, floods, and droughts) and man-made (terrorist attacks, fire, aviation, maritime, and oil disasters), abound worldwide. individuals, firms, and governments often hedge against the potential of extreme losses from these types of catastrophes. economic losses from natural disasters alone have averaged $180 billion annually over the last decade.1 however, the financial losses fail to encompass the full impact of these disasters. for instance, the 2015 earthquake overlapping the countries of * corresponding author. tel.: �1-610-758-4205; fax: �1-610-758-6429. e-mail address: rjk7@lehigh.edu financial services review 25 (2016) 303–329 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. afghanistan and pakistan illustrates the magnitude of nonfinancial losses that are regularly associated with these disasters with over 300 deaths and over 17,000 families displaced when their homes were destroyed.2 most of these losses (both financial and human) are not insured, especially within developing countries. on average only 30% of losses worldwide are covered through insurance policies.3 because of the potential magnitude of these losses, most insurance companies as well as several developing countries offset part of their risk through reinsurance and other financial instruments such as bonds, futures, and sidebars. this article focuses on catastrophe (cat) bonds, their role as an insurance risk offset, and as a viable investment option within a diversified portfolio. starting with a review of the literature, we position our article as a key to understanding cat bonds from the investor’s viewpoint. background is a major factor in understanding cat bonds as an investment option, so we define cat bonds, outline their structure, and sketch how various triggers impact their returns. documenting the historical costs to the insurance industry in the next section sheds light on the magnitude of the losses associated with catastrophic events, why insurance firms need to offset part of this risk, and the role the state of florida played in the advancement of cat bonds. this leads to the analysis of the historical returns using various proxies for an investment in cat bonds. the conclusion ties the presentation together highlighting the fact that cat bonds offer similar returns to comparable corporate issues with the added benefit of diversification. 2. literature review the analysis of cat bonds has spanned a wide spectrum of research from descriptive to empirical. for instance, cummins (2008 and 2012) outlines the state of the cat bond market. in these two descriptive articles, cummins demonstrates a tie between the role played by cat bonds and the reinsurance market. both cat bonds and reinsurance offer the underlying insurance firms a means to offset the risk associated with catastrophic disasters, but cat bonds offer insurance firms an alternative avenue of funding that taps into the capital markets. as reported in cummins (2008), starting with the first successful cat bond underwritten in 1994 for $85 million by hannover re, cat bonds have been used primarily to offset the high layers of reinsurance protection. this sector is not typically covered within the reinsurance market because of the high margins taken by the reinsurers and the lack of faith in the creditability of the reinsurers to offer protection for the magnitude of these low probability events. froot (2001) stresses the role cat bonds have on reducing the barriers of entry into the reinsurance marketplace by offering another avenue insurers can use to reduce their exposure to risk. by examining eight theoretical explanations concerning the pattern of hedging against catastrophe risks from the insurance firm’s perspective, froot finds that the decision between reinsurance and cat bonds is a tradeoff between the lower adverse selection costs associated with reinsurance and the costs stemming from the reinsurance markup. the theoretical work of finken and laux (2009) continues this line of research by demonstrating that cat bonds tied to parametric triggers offer an additional layer of protection to the underlying insurer beyond the reinsurance market. because there is an imperfect 304 r.j. kish / financial services review 25 (2016) 303–329 correlation between the bond’s payoff and the insurer’s loss, diversification is an additional benefit offered by cat bonds. cummins et al. (2004), also exploring the effects of various triggers but through simulation, highlights the advantages to the insurer dependent on the trigger mechanism in place. in a related article, subramanian and wang (2015) attempt to explain why the cat bond market is so small, even though its benefits have been highlighted in the literature (e.g., bantwal and kunreuther, 2000; barrieu and louberge, 2009; cummins and trainar, 2009; dieckmann, 2011; durbin, 2001; hagendorff et al., 2014). subramanian and wang’s claim is that insurers only utilize the cat bond market when potential losses increase beyond where it is efficient to be covered within the reinsurance market. the benefits of cat bonds abound in the literature. for instance, bantwal and kunreuther (2000) rely on simulations to illustrate the benefits of cat bonds to potential investors. the claim is that cat bonds offer a unique opportunity to enhance portfolios with an investment option that provides a high-yielding return that is uncorrelated with the market. during the time of their study the spreads on investment linked securities (ils), of which cat bonds are a subset, were considerably higher than the spreads for comparable speculative-grade corporate debt. this wide spread is not supported within the current market. barrieu and louberge (2009) support earlier findings that the cat bond market offers benefits through diversification and high returns, but the main focus of their research, similar to subramanian and wang (2015), is on explaining why the cat bond market remains small. they argue for the introduction of a hybrid cat bond tied to protection against a stock market crash to add appeal and hopefully expand the cat bond market sector. cummins and trainar (2009) deal with the supply side of the cat bond market. their core finding is that the benefits of cat bonds versus reinsurance increase via the magnitude of the potential losses. from the demand side, dieckmann (2011) supports the significant correlation between returns from cat bonds and speculative grade corporate bonds, but warns of the effects of the low probability of a major catastrophe on returns. durbin (2001), one of the earlier works within this field, reports on the advent of cat bonds and their role in reinsurance from catastrophe events. hagendorff et al., (2014) investigate the cat bond market from the utilization side. they conclude that utilizers of cat bonds are less likely to exhibit risky underwriting and thus have easier access to the financial markets (i.e., the cat bond market); have more efficient hedging outcomes; and because of this easy access may lead to more risky behavior in the future. they also show a negative relationship between cat bond issuance and the size of underwriting losses. hagendorff et al. (2013) takes a different track focusing on the lack of wealth effects to shareholders of the issuing firm as a motive for participating in the cat bond market. using data from 1970 through 1994, froot and o’connell (2008) uncover overpricing within the reinsurance market which contributed to the advent of the alternative catastrophe offsets such as cat bonds. lakdawalla and zanjani (2012) through simulation illustrate the advantages of cat bonds for improving the efficiency of offsetting risks for insurance firms within specific risk settings. thus, most of the literature focuses on the advantages of cat bonds from the issuer’s viewpoint. little work has been done on the actual returns to investors within this market sector, which is one of this article’s contributions. 305r.j. kish / financial services review 25 (2016) 303–329 3. what is a cat bond? to understand the role of cat bonds as an investment option, one must first understand what a cat bond is. a catastrophe-link (cat) bond is a debt obligation in which the interest (coupon payments) and the return of principal are tied to the payoff requirements of an insurance company.4 another common name for cat bonds associated with natural disasters are “act of god” bonds because their payoff is reduced or eliminated if an “act of god” occurs. but catastrophes can also be man-made, such as the 9/11 terrorist attract on the world trade towers in new york city, which was the most costly man-made catastrophe in the united states with insured losses of $32.5 billion.5 similar to reinsurance, cat bonds offer protection to the insurance firm against the extremes, that is, covered losses above a certain manageable threshold. for example, an insurance company could issue a series of bonds in which the payoff is tied to the payout of claims occurring from a natural disaster, such as a hurricane. the bonds offer a two-sided bet. the bondholders are betting that if a hurricane occurs, the insurance payout will be below the threshold established in the bond’s covenants. if the payouts are below the thresholds established, the bondholders will continue to receive their periodic coupon payments and the return of their principal at maturity. if on the other hand, the insurance payouts are above the established thresholds, coupon and principal will be reduced or eliminated to pay these claims. thus, the risks of huge losses to the insurance firms are transferred to the bondholders. 3.1. the current and historical state of the cat bond market cat bonds are a relatively new investment vehicle. they were first issued in the mid-1990s and reached record levels in 2015.6 at the end of 2015, bloomberg reported 189 active cat bonds that were issued from insurance companies within three countries: bermuda (132 issues; 69.8%), cayman island (43 issues; 22.8%), and ireland (13 issues; 6.9%). the remaining bond was issued through a supra national (snat).7 the majority of issues were denominated in u.s. dollars (172 issues; 90.29%; $24,007mm), followed by euros (12 issues; 7.72%; 2,053mm€), japanese yen (four issues; 1.72%; 457mm¥), and swiss franc (one issue; 0.07%; 71mm¥). there were only three bonds with fixed coupons (with an average coupon of 3.73%), 54 zero coupon bonds, and 132 bonds with floating rates. the average floating rate coupon was 4.5%. the adjustments for the floating rate debt were constructed on a base three month rate (euribor—nine issues; libor—11 issues; and treasury bills—112 issues) plus a premium. the average premium was 575 basis points (euribor—299 bps; libor—562 bps; and treasury bills—611 bps). the average maturity was 3.5 years with 68 bonds also including an extension option averaging two additional years. the average size of an issue denominated in euros was 135 million; in yen was 13,284 million; and in u.s. dollars was 141 million. the top five managing firms held close to 70% of the securities outstanding.8 several firms were multiple issuers including the following seven firms (# issues: ticker) with six or more issues: sector re (24: sector), residential re (17: resid), kane sac (10: kanesl), market re (seven: markre), sanders re (six: sandre), tradewinds re (six: trdwyn), and vitality re (six: vitali). 306 r.j. kish / financial services review 25 (2016) 303–329 a summary of amounts issued and outstanding by year for cat bond and insurance linked securities (ils) are summarized in the fig. 1. over the period from 1997 through 2015, the yearly range of issues was $786 million (1997) to $9,094 million (2014) with an average yearly issue of $3,858 million. the cumulative bonds outstanding peaked in 2015 at $25.96 billion. there were several years in which the amount of outstanding bonds declined because of a small volume of new bonds being issued and a large volume of bonds maturing. cat bond payoffs are dependent on the lack of occurrence of the insured disasters. but even when they occur, if the underlying loss fails to reach the defined trigger point, then the bondholders suffer no loss. when the trigger point is breached, there will be a reduction in the expected return on the bonds. these triggers are typically based on the cumulative losses experienced by the underlying insurance companies. triggers, which are defined in the bond indenture, can be structured in a variety of ways. for instance, they can be based on a cumulative loss threshold, a sliding scale of actual losses experienced by the issuer, or tied to an index of weather/disaster conditions. cat bonds can be structured on a per-occurrence event (i.e., exposure to a single major loss event); on the aggregate (i.e., exposure to multiple events over the course of each annual risk-period); or on a multiple loss approach (i.e., payoff reductions are triggered by second and subsequent events). all cat bonds within the sample were issued under the sec’s 144a regulation which limits resales to qualified institutional buyers (qibs). thus, institutions, not individuals regardless of wealth or sophistication, are the only investors eligible to buy these securities within the first year.9 thus, access for the individual investors is only possible through participation in mutual or hedge funds that qualify as qibs. one of the attractions of cat bonds is that they typically offer a higher rate of return versus similarly rated corporate debt instruments. many of the bonds offer variable rates based on a premium over some three month index such as euribor, libor, or u.s. treasury bills. this higher expected return could also be adjusted through the price because there are several zero coupon cat bond issues. besides the potential for high yields, cat bonds also offer the buyer some diversification because of the fact that catastrophic events are not correlated with market cycles or fig. 1. source: http://www.artemis.bm/deal_directory/cat_bonds_ils_issued_outstanding.html. 307r.j. kish / financial services review 25 (2016) 303–329 macroeconomic variables.10 even with these listed benefits, cat bonds should only be a minor portion of an investor’s portfolio because of the small possibility of a huge loss. this is what researchers within finance and economic studies have identified as tail risk. there are varying amounts of risk depending upon where the bond’s coverage occurs. cat bonds issued against catastrophic events in the united states have more data for analysis available versus catastrophic events across the rest of the world. however, even with this additional data, there are no guarantees that historical data will capture the probability for future events as was the case with terrorist attack on the twin towers of the world trade center in new york city on september 11, 2001. although, both natural and man-made disasters are rare, when they occur, the losses are often substantial. thus, insurance firms need to protect themselves against this small probability of extreme loses that might cripple their ability to pay out claims and still stay operational. besides the potential of having a loss of coupon and principal to the underlying insurance firm, other risks to investors of cat bonds include a lack of liquidity (weak secondary market); lack of sec oversight (unregistered securities without periodic disclosure issued under rule 144a—thus only offered to qualified institutional buyers—qibs) and counterparty risk (does the special purpose vehicle, spv, have adequate assets to prevent default). one of the 189 issues outstanding was in default (calabash re iii ltd.: calaba). cat bonds can cover large areas or be quite focused. for example a cat bond issued by travelers, an american insurance company in may 2012, through their spv—long point re iii (series 2012–1) covered excess losses against hurricanes in several northeast states (connecticut, delaware, maine, maryland, massachusetts, new hampshire, new jersey, new york, pennsylvania, rhode island, vermont, and virginia) and the district of columbia. the amount issued was $250 million.11 details of this issue are summarized in table 1. reinforcing the risks that cat bond investors face are highlighted within the prospectus include the potential loss of principal or coupon payments because of the underlying trigger event, the option of the spv to extend the maturity of the bond or to redeem the bonds before table 1 series 2012–1 notes class a ceding company travelers indemnity company (and several affiliates) original principal $250,000,000 initial modeled trigger probability 0.97% initial modeled exhaustion probability 0.77% initial modeled expected loss 0.88% modeling firm air worldwide risk period june 7, 2012 through june 7, 2015 trigger indemnity per occurrence covered event hurricane covered area connecticut, delaware, maine, maryland, massachusetts, new hampshire, new jersey, new york, pennsylvania, rhode island, vermont, and virginia) and the district of columbia rating (s&p) bb� collateral treasury money market funds investor spread 6.00% source: swiss re (2012b). 308 r.j. kish / financial services review 25 (2016) 303–329 maturity, limited recourse to assets of the special purpose vehicle (spv) and no recourse to the assets of travelers, the possibility of insolvency of the issuer and the consequential loss of some or all of their investment, potential negative tax consequences, and limited liquidity. two additional examples of cat bonds highlight the differences in the range of area covered by the underlying event. first, pennunion re ltd. (series 2015–1) was issued october 2015 by cedent amtrak to cover u.s. storm surge (new york city and delaware) and wind damage (connecticut, delaware, maryland, massachusetts, new jersey, new york, pennsylvania, and rhode island) caused from named storms or earthquakes. the size of the issue was $275 million. the bond’s rating by s&p was bb-. this noninvestment grade rating is typical within the cat bond market. as with most cat bond issues, this one originates in bermuda, one of the leading reinsurance and cat bond issuing countries. the second example is the compass re ii ltd. (series 2015–1) issued by one of the leading cedent, aig. the primary coverage for this bond is protection from u.s. wind damage. the size of the issue was $300 million. this bond was rated by fitch at b�. the coverage area is extensive ranging from the gulf to the east coast (texas through massachusetts). this is a short term bond lasting from the official start of the hurricane season (june 1 through november 30). this bond was structured as a zero coupon bond. as shown in tables 2 and 3, the extent of damages from major catastrophe events has exposed insurance and reinsurance firms to the potential of huge losses. thus, the cat bond market was created to help further diversify the risks beyond the reinsurance market. 3.2. cat bond structure cat bonds are issued through a spv to separate the legal and financial liabilities of an insurance firm from the liabilities associated with the bonds. similar to the reinsurance market, cat bonds provide insurance firms an opportunity to offset payout risk above a predetermined threshold (trigger event). in addition to insurance firms, reinsurance firms and some government (both foreign and domestic) agencies also issue cat bonds. based on the underlying event being covered, the spv sets the terms for the bonds (i.e., coupon rate, maturity, amount, and most importantly, defining the trigger event). as with all bond issues, the terms of the contract must be spelled out within the bond indenture. after the terms of the bond are set, the spv issues the cat bonds to investors. the principal (selling price) is then deposited into a collateral account. additionally, part of the insurance premiums are transferred on a periodic basis to the spv as additional collateral. both the principal and the partial premiums are held until they are earned by the bondholders. to maximize returns to the underlying firm, these held funds are reinvested in low risk securities typically highly rated money market funds. these combined cash flows are used to pay the cat bond coupons that are commonly floating rate issues (based on three-month rates such as from treasury bills plus a fixed spread). the typical maturity for cat bonds is three years and ranges from one to five years with coupon resets on an annual basis with payments of quarterly coupons. because the investors’ coupon, or interest payments, are made up of interest the spv makes from the collateral and the premiums the sponsor pays, this aspect of the cat bond returns is very safe. the size of a cat bond issue is typically 309r.j. kish / financial services review 25 (2016) 303–329 table 2 the 25 most costly insurance losses from natural disasters (1970–2014) ranka,b insured loss (bn)c victimsd date (start) event country 1*# $78.638 1,836 8/25/2005 hurricane katrina; storm surge, damage to oil rigs united states, gulf of mexico, bahamas 2 $36.828 18,520 3/11/2011 earthquake (mw 9.0)e triggers tsunami japan 3# $36.079 237 10/24/2012 hurricane sandy; massive storm surge united states, caribbean 4*# $26.990 43 8/23/1992 hurricane andrew; floods united states, bahamas 5 $25.104 2982 9/11/2001 terror attack on wtc, pentagon, other buildings united states 6 $22.355 61 1/17/1994 northridge earthquake (mw 6.6) united states 7*# $22.258 136 9/6/2008 hurricane ike united states, caribbean, gulf of mexico 8 $16.836 181 2/22/2011 earthquake (mw 6.3) aftershocks new zealand 9*# $16.157 119 9/2/2004 hurricane ivan; damage to oil rigs united states, caribbean, barbados 10 $15.783 815 7/27/2011 floods caused by monsoon rains thailand 11*# $15.234 35 10/19/2005 hurricane wilma; torrential rain, floods united states, mexico, jamaica, haiti 12# $12.240 34 9/20/2005 hurricane rita; floods, damage to oil rigs united states, gulf of mexico, cuba 13 $11.339 123 7/15/2012 drought in the corn belt united states 14*# $10.087 24 8/11/2004 hurricane charley united states, cuba, jamaica 15 $9.813 51 9/27/1991 typhoon mireille japan 16# $8.730 71 9/15/1989 hurricane hugo united states, puerto rico 17 $8.682 562 2/27/2010 earthquake (mw 8.8) triggers tsunami chile 18 $8.458 95 1/24/1990 winter storm daria france, united kingdom 19 $8.241 110 12/25/1999 winter storm lothar switzerland, united states, france 20 $7.681 321 4/22/2011 major tornado outbreak: 343 tornadoes; hail united states 21 $7.418 177 5/20/2011 major tornado outbreak: 180 tornadoes united states 22 $6.959 54 1/18/2007 winter storm kyrill; floods germany, united kingdom 23 $6.456 22 10/15/1987 storms and floods in europe france, united kingdom 24# $6.449 38 8/26/2004 hurricane frances united states bahamas 25 $6.134 50 8/22/2011 hurricane irene; torrential rainfall, flooding united states, canada, bahamas a(*) also listed under bankrate.com’s “10 costliest natural disasters. other disasters not included in the above table include: 1988 drought and heat wave affecting central and eastern united states ($76.4 billion damages and 5,000 to 10,000 deaths); 1994 northridge earthquake in california ($67 billion in damages and 60 deaths); 1980 drought and heatwave affecting central and eastern united states ($54.8 billion damages and 10,000 deaths); and 1993 midwest floods affecting central united states (32.8 billion damages and 48 deaths. (source: http:// www.bankrate.com/finance/insurance/top-10-costliest-natural-disasters-1.aspx). b(#) also listed under cnbc.com’s “10-most expensive hurricanes in u.s. history (source: http://www.cnbc.com/2013/10/03/the-10-most-expensive-hurricanes-in-the-history-of-the-united-states.html?slide�11). c table values in us$ billion, 2014 prices. d dead or missing. e mw is the earthquate magnitude scale (mw � 2/3 log10 (mo) � 10.7 where mo is the seismic moment). source: swiss re sigma no. 2/2015. 310 r.j. kish / financial services review 25 (2016) 303–329 over $100 million to help compensate for the transaction costs, modeling risks, and marketing fees. although cat bonds have a low probability that the underlying catastrophe will occur and consequential forfeiture of principal and future coupons, they are still classified as high risk bonds. thus, they are usually rated within the bb, b, and ccc categories, which indicates non-investment (high yield or junk) grade securities by the three large rating agencies (fitch, moody’s, and s&p). although the rating agencies, the insurance firms, and reinsurance firms have their own security analysts, a few key firms specializing in risk assessment (i.e., probability modeling, weather forecasts, seismology, and other technical factors associated with the events being analyzed) dominate the analysis of this type of bond from both the buyer’s and seller’s vantage. catastrophe modeling is vital to cat bond transactions to provide analysis and measurement of events which could cause a loss, as well as, to define the exposed geographical region. bond modeling is much more prevalent within the u.s. market because of the amount of historical data available. these analysts rely on the modeling of historical and simulated data to estimate the likelihood of a catastrophic event occurring and the financial impact of the event if it occurs. thus, mutual and hedge fund investors with cat bond holdings are at a disadvantage to the insurance and reinsurance firms when undertaking the risk assessment of cat bond holdings. this is definitely not a level playing field. one other distinct disadvantage to the potential investor, is that these bonds are not subject to sec registration and disclosure requirements. besides the risk of a triggering event, cat bondholders are also exposed to additional risks. as previously mentioned, there is modeling risk. even the best models cannot fully table 3 disasters consequences disasters natural man-made no. victims natural man-made victims 2010 304 167 (55%) 137 (45%) 303,573 297,127 (98%) 6,446 (2%) 2011 325 175 (54%) 150 (46%) 34,729 29,026 (84%) 5,703 (16%) 2012 318 168 (53%) 150 (47%) 13,929 8,948 (64%) 4,981 (36%) 2013 325 166 (51%) 159 (49%) 25,903 20,201 (78%) 5,702 (22%) 2014 336 189 (56%) 147 (44%) 12,777 7,066 (55%) 5,711 (45%) five-year average 322 173 (54%) 149 (46%) 78,182 72,474 (93%) 5,709 (7%) note: no. victims equals dead or missing. total losses insured losses natural man-made financial losses 2010 218,000 43,475 (20%) 39,869 (18%; 92%) 3,606 (2%; 8%) 2011 370,000 115,814 (31%) 110,021 (30%; 95%) 5,794 (1%; 5%) 2012 186,000 77,238 (42%) 71,279 (38%; 92%) 5,960 (4%; 8%) 2013 138,000 44,917 (33%) 37,047 (27%; 82%) 7,870 (6%; 18%) 2014 110,000 34,708 (32%) 27,749 (25%; 80%) 6,958 (7%; 20%) five-year average 204,000 63,230 (31%) 57,193 (28%; 90%) 6,038 (3%; 10%) $us million $us million (insured/total; insured/natural or man-made) note: insured losses account for property and business interruptions, excluding liability and life insurance losses. source: swiss re sigma 2011–2015. 311r.j. kish / financial services review 25 (2016) 303–329 account for the differences between simulated events and actual events. this showed up in the mismatch of forecasted hurricanes between 2001 and 2005 and the actual number of hurricanes that occurred. industry models to forecast extreme meteorological events are weak at best. a second risk to consider is liquidity because of the fact that cat bonds are highly illiquid. typically, they are bought and held with little aftermarket trading. third, is marketability, that is, cat bonds are issued under rule 144a which makes them accessible to qibs only during the first year of the offering. these bonds are not subject to normal sec filings that makes pricing and risk assessment that much harder. finally, these bonds have counterparty risk. if the associated insurance firm runs into financial distress, the bondholders may not receive all that is promised. additionally, cat bonds can be issued as a single class or as a multiple class bond with various tranches. this is comparable to collateralized mortgage obligations (cmos) in which each class has different risks and thus different compensation. cat bonds, like any other bond class, can be bundled into cdos (collateralized debt obligations) further obscuring the relationships between risk and return. the spv must monitor for the trigger event. this could be a single or multiple event clause. if a qualifying event occurs that triggers a payout, the spv liquidates enough collateral required to make the payment and reimburse the counterparty to satisfy the terms of the cat bonds. thus, the promised cash payments (coupons) stop flowing to the bondholders, along with the expected principal that would have occurred at maturity. these funds are transferred back to the insurers to cover the catastrophic event. if no trigger event occurs, then the bondholders receive their expected cash flows and any remaining collateral is liquidated at maturity and firm’s investors are repaid. fig. 2 shows a typical cat bond structure including where the capital flows from one party to another. depending on if and fig. 2. cash flow analysis. notes: (1) insurance/reinsurance/government agency enters into a risk transfer contact (i.e., the creation of cat bonds) with a special purpose company (spv) established specifically for this transaction. (2) the spv issues the catastrophe (cat) bonds to investors in the capital markets in an amount equal to the limit of the risk transfer contract. (3) proceeds from the sale of the bonds and the periodic premiums transferred to the spv are invested in a collateral trust account of safe investments such as u.s. treasury bills or other money market investments. (a) if no covered event occurs, the bonds will be redeemed at face value. (b) however, if a covered event meeting the defined threshold limits occurs, funds will be withdrawn from the collateral account to cover this excess above the threshold. thus, bondholders could lose all of their principal and any remaining coupons. source: swiss re (2012a). 312 r.j. kish / financial services review 25 (2016) 303–329 when the trigger event occurs, the spv may be forced to extend the maturity of the bond (from three months to two years) until the insurance loss is verified. 3.3. classification of triggers events one of the key areas of contention with cat bonds, highlighted by finken and laux (2009), froot (2001), and cummin et al. (2004), is what qualifies as a trigger event. for instance, the financial industry regulatory authority (finra) outlines the five key trigger descriptions: parametric, modeled loss, industry loss index, indemnity, and hybrid.12 a parametric bond is triggered if specific, objective “parameters” are met—for example, wind speed for a hurricane-linked bond or ground acceleration for an earthquake-linked bond. this is the most transparent and easiest to verify of the triggers and presents the least potential for a sponsor to influence the bond’s performance. it typically pays a lower yield than bonds with other trigger types, as it may not cover all of sponsor’s losses. the second category, modeled loss, measures the sponsoring firm’s exposure or expected loss. it is calculated by computer models that use objective data, such as actual wind speeds or ground acceleration. the bond is triggered if the sponsor’s exposure exceeds a specified dollar amount. it allows for faster verification than the industry-loss or indemnity triggers described below, but there is a heavy reliance on computer modeling to determine when the trigger has occurred. if the computer program is good, then the estimates are good. within the industry-loss index, a bond is triggered when the amount of the overall industry loss from an event, usually determined by an independent third party, exceeds a certain amount. in this situation, there is minimal potential for a sponsor to influence the bond’s performance, as the index is based on industry-wide losses for each event. it typically pays a somewhat higher yield than for parametric triggers, but compared with parametric and modeled loss triggers, it takes longer to compute the final amount of industry loss, leading to increased uncertainty for investors. within the indemnity category, a bond is triggered when the sponsor’s actual underwritten loss on specific insurance policies exceeds a predetermined amount. for example, a sponsor’s insurance claims resulting from a florida hurricane may need to exceed $1 billion to the sponsor before investors lose their interest and principal. it typically pays the highest yield of the different trigger types, as it provides the best protection to the sponsor, but it presents the most potential for the sponsor to influence bond performance, as payouts are based on the individual policy claims against the sponsor and the way the sponsor settles those claims. furthermore, a long period of time can be needed to calculate total loss claims, again leading to increased uncertainty for investors. the fifth classification, a hybrid trigger, is created by combining any of the above triggers. it is useful for bonds linked to multiple events and can be structured to cushion investors’ losses and/or enhance yield potential. however, depending on the components of the hybrid trigger, it can be complicated and difficult to understand or verify. according to swiss re (2012a, 2012b), the natural catastrophic bond trigger breakdown by classification were industry index (40%), indemnity (37%), parametric index (12%), modeled losses (6%), hybrid (4%), and other (2%). 313r.j. kish / financial services review 25 (2016) 303–329 4. catastrophic costs and the need for offsetting risks catastrophic costs vary from disaster to disaster and from year to year. for instance, the tenth costliest natural disaster in the united states between 1980 and 2010 reported by bankrate.com was hurricane katrina in 2005.13 the states impacted the most were alabama, florida, louisiana, and mississippi with total damages of $145 billion ($79 billion insured) and 1,836 deaths. the top 10 disasters totaled over $500 billion in damages and more than 20,000 deaths. besides hurricane damage, terroristic attacks (ex. the 2001 9/11 new york city twin tower attack: $32.5 billion of insured losses and 2,976 deaths) and earthquakes (ex. 2011 9.0 mw tohuku japan earthquake: $36.8 billion of insured losses and 18,520 deaths). from a worldwide perspective, 2011 produced the largest economic losses from earthquakes and natural disasters tallying over $365 billion and over 20,000 deaths. several of these disasters, such as the 2011 japanese earthquake and the 9/11 bombing of the twin towers in new york city, illustrate the type of disasters in which claims are not settled for years after the disaster occurred. 4.1. historical loss perspective table 2 tallies the 25 costliest worldwide insurance losses from natural disasters. for instance, hurricane katrina in 2005 is listed as the costliest natural disaster tallying $78.6 billion in insured losses from the united states (primarily alabama, fl, la, and mississippi), mexico, and the bahamas). table 2 also attributes 1,836 deaths (dead or missing) occurring directly because of this hurricane. thus, the key takeaway from the complete listing is the massive amount of losses that have occurred overtime. these losses justify the need for insurance companies to have a mechanism to offset part of their exposure to catastrophic risks. jarzabkiwski, bednarek, and spee (2015) detail that the primary offset in the past has been through the reinsurance market, but that a small and growing segment for this offset is within the financial markets primarily with cat bonds. 4.2. need for offsetting risks swiss re, a leading reinsurance firm, offers additional support for the need of insurance firms to offset catastrophic risks. as shown in table 3, over the past five years the average annual number of disasters was 322; 173 (54%) from natural disasters and the remaining 149 (46%) from manmade disasters. the breakdown between natural and manmade disasters was consistent over the last five years. although the number of victims over this same five year period averaged 78,000, this average was highly skewed by one event in 2010. during that year, the 7.0 earthquake in haiti accounted for 222,570 deaths, which is more than the total from the next four years combined. the haiti earthquake also accounted for over 300,000 injured and 1,200,000 homeless individuals. although the insured losses totaled $100 million, this was a minuscule amount compared with the total damage of over $10 billion. the bottom half of table 3 focuses on the financial losses over this same five year period (2010–2014). the average annual loss was $204,000 million with approximately 30% of the 314 r.j. kish / financial services review 25 (2016) 303–329 total insured losses. the proportion of these insured losses from natural versus man-made disasters was 9 to 1. 4.3. the florida impetus as the financial markets have become more heavily regulated, reserve capital requirements have increased. this is one of the reasons insurance firms have been forced to offset part of their risks onto the financial markets. an additional reason for reinsurance and the assorted financial instruments they have sprouted flow from the desire of insurance firms to stay solvent. an illustration of a key driver within the reinsurance market was the reaction of the state of florida to the financial crisis caused by hurricane andrew in 1992. after hurricane andrew caused unprecedented damage (over $16 billion of insured losses) in florida, the insurance industry was in a crisis. seven florida based companies and one national firm became insolvent as several others become “technically insolvent.” because of the extent of damages from this storm, many of the firms were threatening to withdraw from the florida market. in 1992, the florida department of insurance enacted two emergency rules (4er92–11 and 4er92–15). the first rule limited the ability of insurance firms to abandon the high-risk areas of the florida market. the rule placed a 90 day moratorium on all firms wishing to withdraw. during this time a firm wishing to withdraw from the florida market had to file a written statement of intent including details for the withdrawal and any projected effects it would have on the market. it also limited the firm to withdrawing only 5% of their policies per year. this rule was originally to last only 6 months, but was extended for an additional three years.14 the second rule established the florida property and casualty joint underwriting association (fpcjua) to make sure that insurance coverage would be available to all. this was combined with the florida windstorm underwriting association (fwua), which insured beach-front properties, together to form citizens property insurance corporation (cpic) in 2012. this new entity is the largest insurer for hurricane disasters in the florida market. in 2012, cpic sponsored the secondlargest transaction in the market’s history by floating $750 million in cat bonds.15 florida created the florida hurricane catastrophe fund (fhcf) in 1993 to provide additional reinsurance capacity following hurricane andrew. this is a state government trust fund, exempt from federal taxes, which requires all insurers (residential and commercial) operating within the state to provide a cushion against catastrophic losses caused from future hurricanes. the creation of new laws, the cpic, and the fhcf were just part of the six key changes to the insurance market in the aftermath of hurricane andrew which included: (1) more carefully managed coastal exposure; (2) a larger role of government (both federal and state) in insuring risks; (3) the introduction of hurricane deductibles; (4) greater use of reinsurance capital from worldwide sources including the financial markets; (5) the expansion and refinement of sophisticated catastrophic modeling; and (6) support and enforcement of tougher building codes.16 the fhcf has been a success in ensuring the solvency of insurance firms operating within florida as evidenced by the florida office of insurance regulation terminating the 1.3% assessment on most property insurance policies 18 months ahead of schedule on july 22, 2014.17 further support is offered in the annual “report prepared for the florida hurricane 315r.j. kish / financial services review 25 (2016) 303–329 catastrophe fund (fhcf): claims-paying capacity estimates,” which forecasts florida’s ability to weather future hurricanes very positively.18 key factors leading to this conclusion include: (1) the requirement of all insurers within the state to contribute to fhcf; (2) forecasts of 50, 100, and 250 years to capture one in a lifetime events; (3) strong debt ratings (aa by all three rating agencies—moody’s, standard & poor’s, and fitch) with capacity for future borrowings; (4) the ability to levy emergency assessments on all property and casualty insurance lines; and (5) successful entry into the reinsurance and financial markets. 5. analysis of returns the reinsurance market was very lucrative from the 1990s through the early 2000s. warren buffett in his 50th anniversary letter to berkshire hathaway investors states the above average returns the firm was able to generate was because of the insurance/reinsurance market.19 because of the firm’s strong capital base, they were a “go to” reinsurance firm. insurance firms participating in reinsurance with berkshire hathaway projected little risk of failure to pay claims so it is surprising that the firm is pulling back from their participation in future reinsurance investments. buffett feels that because of the amount of competition from other reinsurance firms, hedge funds, and the financial markets, the expected returns do not justify the risk being undertaken.20 thus, the financial markets have helped make the reinsurance market more competitive. however, the prediction of future catastrophic risks from population migration to areas of potential catastrophic disasters, such as the hurricane belts across florida, dictate the need for insurance firms to continue to offset these risks. an additional complication is the rise in extreme weather patterns associated with climate change. note also that it is rare for cat bondholders to lose all their funds. the insurance journal in 2013 cite that “only eight … deals issued since 1997 have been triggered—four triggered as a result of losses from natural disasters and others by damages as a result of the 2008 financial crisis.”21 however, what type of returns have cat bondholders received? 5.1. cat indices returns from cat bonds are difficult to calculate because they are traded infrequently so a number of proxies are used. our first test relies on two broad proxies, the eurekahedge ils advisors index (ils) and the mercury investible catastrophe risk index (micrix). the eurekahedge ils advisors index tracks the performance of participating insurance linked investment funds. this index allows a comparison between different fund managers in the insurance-linked securities, reinsurance, and cat bond investment field. eurekahedge, the index manager, is the world’s largest compiler of alternative asset fund databases and ils advisers are an independent advisory service for insurance-linked investments. a sample of monthly returns from this index are shown in table 4 over the period 1/2010 through 3/2016. the mean monthly return is 0.41% coupled with a 0.61% standard deviation. the maximum and minimum monthly returns over this period is 1.20% and �3.94%, respectively. although not shown in table 4, since its inception in 2005, the best monthly return is 1.60% versus the worst monthly return of �3.94%. the sharpe ratio, as a measure of reward to volatility, 316 r.j. kish / financial services review 25 (2016) 303–329 at 0.1897 indicates a positive return for the risk undertaken. the percentage of positive returns over the period 1/2010 through 3/2016 is 94.7% (71 out of 75 months). table 4 shows the historical monthly returns over this seven year period and table 5 shows a summary of the statistics over various sample periods. the second index used as a proxy for the cat bond performance is the mercury investible catastrophe risk index (micrix). it was started in 2006 to track the performance of a diversified portfolio of peak peril industry loss warranties (ilw’s). mercury capital ltd. is a bermuda based fund manager. their index data are based on prices collected from a panel of reinsurance brokers. the micrix index tracks the performance of a balanced portfolio of the peak peril exposures from us quake, us regional wind, european wind, japanese quake, and japanese wind. the index experienced only three down months in the 87 months since inception. this index shows a mean monthly return from 1/2010 through 3/2016 of 0.72% matched with a monthly standard deviation of 2.63%. similar to the ils index, there is a positive sharpe ratio (0.1605). summary statistics over the same time periods as the ils index are shown in table 5. monthly returns, shown in the bottom half of table 4, indicate higher returns compared to the ils index, but with greater variability. the three year correlation of monthly returns between these two cat indices (ils and micrix) is 0.74 indicating a strong positive relationship between these two cat performance proxies. to analyze the performance of these two broad cat bond proxies, their returns are evaluated against the returns from the merrill lynch (ml) non-investment grade corporate bond indices.22 monthly returns of the ml indices are shown in table 6. the top half of the table shows the returns from the bb rated bond index; the middle section shows the returns from the b rated corporate bond index; and the bottom section, the returns from the ccc table 4 ils and micrix historical monthly performance jan. feb. mar. april may june july aug. sept. oct. nov. dec. ils 2016 0.21% 0.54% 0.37% n/a n/a n/a n/a n/a n/a n/a n/a n/a 2015 0.39% 0.24% 0.21% 0.08% 0.16% 0.15% 0.40% 0.84% 1.03% 0.27% 0.31% 0.08% 2014 0.50% 0.50% 0.45% 0.32% 0.08% 0.21% 0.41% 0.81% 0.86% 0.61% 0.14% 0.43% 2013 0.68% 0.75% 0.64% 0.85% 0.44% 0.00% 0.40% 0.92% 1.20% 0.64% 0.49% 0.42% 2012 0.18% 0.19% 0.32% 0.43% 0.59% 0.57% 0.62% 0.94% 1.19% �0.51% 0.27% 1.02% 2011 0.70% 0.18% �3.94% 0.06% 0.21% 0.71% 0.67% 0.12% 0.53% 0.73% �0.04% �0.04% 2010 0.92% 0.94% 0.45% 0.49% 0.29% 0.16% 0.52% 0.75% 1.17% 0.90% 0.29% 0.42% mean 0.41% standard deviation 0.61% maximum 1.20% minimum �3.94% micrix 2016 0.54% 0.40% 0.33% n/a n/a n/a n/a n/a n/a n/a n/a n/a 2015 0.42% 0.42% 0.32% 0.23% 0.23% 0.61% 0.81% 2.06% 2.74% 1.44% 0.48% 0.42% 2014 0.54% 0.53% 0.40% 0.28% 0.28% 0.70% 0.92% 2.28% 3.03% 1.64% 0.58% 0.52% 2013 0.63% 0.63% 0.48% 0.35% 0.35% 0.91% 1.19% 2.97% 3.95% 2.11% 0.69% 0.61% 2012 0.74% 0.74% 0.56% 0.39% 0.39% 0.99% 1.29% 3.28% 4.42% �12.39% 0.87% 0.79% 2011 0.58% 0.57% �14.93% 0.22% 0.22% 0.90% 1.18% 3.29% 4.51% 2.39% 0.64% 0.55% 2010 0.61% 0.60% 0.44% 0.30% 0.30% 0.85% 1.07% 2.75% 3.70% 2.00% 0.66% 0.59% mean 0.72% standard deviation 2.63% maximum 4.51% minimum �14.93% source for ils index: http://www.artemis.bm/eurekahedge_ils_advisers_insurance_linked_securities_fund_ index/; source for micrix index: http://www.artemis.bm/mercury_micrix/ ils � investment linked securities; micrix � mercury investible catastrophe risk index. 317r.j. kish / financial services review 25 (2016) 303–329 rated corporate bond index. the highest monthly return during the 1/2010–3/2016 period for the bb, b, and ccc bonds are 4.07%, 4.00%, and 10.03%, respectively. however, this is contracted with the lowest returns over the same period of �2.67%, �2.85%, and �6.54%. the mean monthly returns (standard deviations) for these three indices are 0.54% (1.20%), 0.45% (1.43%), and 0.47% (2.83%), respectively, for the bb, b, and ccc indices, which is similar to the cat proxies (ils and micrix). these three ml indices are highly correlated with each other (bb with b 0.89; bb with ccc 0.77; and b with ccc 0.93) as expected for three noninvestment grade bond indices. the correlations are summarized in table 8. because we wish to compare results (monthly returns from cat proxies and the ml speculative bond indices) by matched time periods, our analysis focuses on the inferences about the mean of the differences between the two populations in a paired difference tests.23 for this test of the mean of the differences, our null hypothesis test forecasts no difference between the monthly returns of the junk bond indices and the cat bond indices as: table 5 summary statistics for catastrophe (cat) indices period statistic ils micrix swiss bb swiss g swiss us whole mean 0.41% 0.72% 0.48% 0.61% 0.63% � 0.61% 2.63% 0.90% 0.80% 0.88% max 1.20% 4.51% 2.55% 2.23% 2.45% min �3.94% �14.93% �4.89% �3.56% �3.92% # positive 71 73 64 68 66 sharpe 0.1897 0.1605 0.1963 0.3894 0.3714 count 75 75 75 75 75 1 year mean 0.35% 0.85% 0.27% 0.35% 0.36% � 0.30% 0.81% 0.48% 0.48% 0.53% sharpe 0.8890 0.9471 0.3886 0.5482 0.5302 # positive 12 12 8 10 9 2 years mean 0.40% 0.91% 0.31% 0.42% 0.44% � 0.27% 0.83% 0.44% 0.44% 0.48% sharpe 0.3976 0.7529 0.0412 0.2926 0.3127 # positive 24 24 17 21 20 3 years mean 0.47% 1.02% 0.41% 0.57% 0.60% � 0.30% 0.95% 0.45% 0.49% 0.53% sharpe 1.2346 0.9682 0.6873 0.9493 0.9223 # positive 36 36 29 33 32 4 years mean 0.47% 0.81% 0.46% 0.63% 0.66% � 0.34% 2.20% 0.58% 0.63% 0.69% sharpe 1.0441 0.3130 0.5864 0.8053 0.7723 # positive 47 47 39 43 42 5 years mean 0.38% 0.65% 0.42% 0.57% 0.58% � 0.66% 2.90% 0.92% 0.81% 0.89% sharpe 0.1993 0.1389 0.1894 0.3941 0.3761 # positive 51 53 45 49 47 6 years mean 0.42% 0.73% 0.48% 0.62% 0.63% � 0.62% 2.69% 0.91% 0.82% 0.90% sharpe 0.2376 0.1731 0.2334 0.4305 0.4085 # positive 68 70 61 65 63 note: 1 year period (1/2015–12/2015); 2 year period (1/2014–12/2015); 3 year period (1/2013–12/2015); 4 year period (1/2012–12/2015); 5 year period (1/2011–12/2015); 6 year period (1/2010–12/2015); and whole period (1/2010–3/2016). ils � investment linked securities. 318 r.j. kish / financial services review 25 (2016) 303–329 t ab le 6 b b an d b ra te d co rp or at e de bt hi st or ic al m on th ly pe rf or m an ce ja n. fe b. m ar . a pr il m ay ju ne ju ly a ug . se pt . o ct . n ov . d ec . b b 20 16 � 0. 90 % 1. 17 % 2. 32 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 0. 57 % 1. 50 % � 0. 27 % 0. 85 % 0. 47 % � 0. 57 % 0. 08 % � 0. 70 % � 1. 31 % 2. 28 % � 1. 03 % � 1. 42 % 20 14 0. 35 % 1. 07 % 0. 18 % 0. 53 % 0. 46 % 0. 36 % � 0. 99 % 1. 12 % � 1. 19 % 1. 26 % � 0. 10 % � 0. 42 % 20 13 0. 83 % 0. 43 % 0. 62 % 1. 17 % � 0. 20 % � 1. 94 % 1. 91 % 0. 14 % 0. 96 % 1. 57 % 0. 55 % 0. 47 % 20 12 2. 16 % 1. 51 % 0. 32 % 0. 85 % � 0. 27 % 1. 22 % 1. 39 % 1. 00 % 1. 11 % 0. 97 % 0. 41 % 1. 02 % 20 11 1. 31 % 0. 87 % 0. 54 % 0. 92 % 0. 58 % � 0. 52 % 1. 05 % � 2. 54 % � 2. 67 % 4. 07 % � 1. 02 % 1. 72 % 20 10 1. 14 % 1. 09 % 2. 50 % 1. 09 % � 1. 85 % 1. 32 % 2. 39 % 0. 64 % 2. 17 % 1. 47 % � 0. 78 % 1. 18 % m ea n 0. 54 % st an da rd de vi at io n 1. 20 % m ax im um 4. 07 % m in im um � 2. 67 % b 20 16 � 2. 18 % � 0. 17 % 3. 63 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 � 0. 17 % 2. 25 % � 0. 37 % 1. 31 % 0. 58 % � 0. 89 % � 1. 64 % � 2. 35 % � 2. 64 % 1. 78 % � 2. 13 % � 2. 31 % 20 14 0. 54 % 1. 06 % 0. 34 % 0. 48 % 0. 51 % 0. 57 % � 0. 84 % 1. 00 % � 1. 19 % 0. 54 % � 0. 45 % � 1. 56 % 20 13 1. 25 % 0. 68 % 0. 95 % 1. 05 % 0. 02 % � 1. 39 % 1. 74 % 0. 21 % 0. 87 % 1. 60 % 0. 66 % 0. 47 % 20 12 2. 11 % 1. 55 % 0. 47 % 0. 90 % � 0. 53 % 1. 37 % 1. 48 % 0. 99 % 1. 32 % 1. 04 % 0. 73 % 1. 09 % 20 11 1. 70 % 1. 09 % 0. 50 % 1. 04 % 0. 53 % � 0. 34 % 0. 84 % � 2. 85 % � 2. 48 % 4. 00 % � 0. 92 % 1. 43 % 20 10 1. 05 % 0. 22 % 2. 51 % 1. 37 % � 2. 05 % 1. 24 % 2. 68 % 0. 34 % 2. 33 % 1. 59 % � 0. 40 % 1. 70 % m ea n 0. 45 % st an da rd de vi at io n 1. 43 % m ax im um 4. 00 % m in im um � 2. 85 % c c c 20 16 � 3. 85 % � 2. 42 % 10 .0 3% n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 � 0. 31 % 2. 86 % � 0. 66 % 1. 13 % 0. 13 % � 1. 77 % � 2. 83 % � 3. 06 % � 2. 72 % 0. 90 % � 4. 07 % � 5. 10 % 20 14 1. 22 % 0. 94 % 0. 52 % 1. 01 % 0. 96 % 0. 37 % � 1. 14 % 0. 14 % � 2. 21 % � 1. 16 % � 0. 89 % � 3. 07 % 20 13 3. 20 % 0. 62 % 1. 49 % 2. 15 % 0. 74 % � 1. 92 % 1. 90 % 0. 45 % 0. 75 % 1. 90 % 1. 18 % 0. 62 % 20 12 4. 14 % 3. 63 % 0. 59 % 0. 98 % � 2. 71 % 2. 43 % 1. 25 % 1. 75 % 2. 48 % 0. 91 % 0. 22 % 2. 90 % 20 11 2. 68 % 1. 29 % 0. 57 % 1. 54 % 0. 05 % � 1. 29 % 0. 49 % � 6. 54 % � 6. 38 % 7. 98 % � 3. 89 % 2. 59 % 20 10 2. 65 % 0. 58 % 4. 85 % 3. 23 % � 4. 45 % 0. 92 % 2. 87 % � 1. 00 % 3. 14 % 3. 58 % � 0. 29 % 3. 40 % m ea n 0. 47 % st an da rd de vi at io n 2. 83 % m ax im um 10 .0 3% m in im um � 6. 54 % so ur ce : b lo om be rg , b an k of a m er ic am er re ll l yn ch u s h ig h y ie ld b b , b , an d c c c d eb t in di ce s w ith m at ur iti es of on e to fiv e ye ar s. 319r.j. kish / financial services review 25 (2016) 303–329 h0: �d � 0 versus ha: �d � 0 the test statistic is a one-sample t, because we are analyzing a single sample of differences: test statistic: t � x�d � 0 sd/�nd where �d � the mean of the difference of the monthly returns of the speculative (junk) bond index and the cat bond index; x�d � sample mean of differences; sd � sample standard deviation of differences; and nd � number of differences. a comparison of the monthly differences of returns using the 3 ml speculative corporate bonds indices (bb, b, and ccc) and our two broad cat indices (ils and micrix) finds no cases in which the speculative corporate sector of bonds outperformed the cat proxies. there are no incidences of positive t-values at any reasonable level of significance (10%, 5%, or 1%). that is, the null hypothesis of mean of the differences of monthly returns show either nonsignificant results favoring the returns from the cat proxies versus the speculative grade bonds or in a few time periods statistically significant results favoring the cat bond indices. see table 7 for a summary of the significance tests over the total time period and the six yearly test periods (1-, 2-, 3-, 4-, 5-, and 6-years). table 7 means of the differences tests for cat indices t-values ils micrix swiss bb cat swiss global cat swiss us cat bb corporate whole 0.853 �0.530 0.377 �0.433 �0.506 1 year �0.835 �1.809 �0.583 �0.741 �0.769 2 year �1.241 �2.759** �0.761 �1.214 �1.281 3 year �1.236 �3.301*** �0.812 �1.667 �1.796 4 year �0.213 �1.041 �0.096 �1.159 �1.282 5 year 0.293 �0.534 0.030 �0.755 �0.807 6 year 0.737 �0.584 0.253 �0.557 �0.623 b corporate whole 0.183 �0.775 �0.151 �0.892 �0.946 1 year �1.635 �2.204* �1.401 �1.508 �1.511 2 year �2.072* �3.111*** �1.678 �2.014 �2.051 3 year �1.826 �3.412*** �1.482 �2.162* �2.257* 4 year �0.871 �1.320 �0.743 �1.613 �1.707 5 year �0.261 �0.749 �0.429 �1.167 �1.202 6 year 0.180 �0.786 �0.182 �0.937 �0.987 ccc corporate whole 0.168 �0.548 �0.018 �0.426 �0.466 1 year �2.303* �2.753** �2.111 �2.188 �2.190 2 year �2.821*** �3.541*** �2.531** �2.777** �2.805** 3 year �1.936 �3.106*** �1.727 �2.212* �2.288** 4 year �0.699 �1.229 �0.641 �1.215 �1.292 5 year �0.534 �0.863 �0.625 �1.073 �1.104 6 year 0.068 �0.655 �0.142 �0.586 �0.626 note: cat � catastrophe; ils � investment linked securities; micrix � mercury investible catastrophe risk index. significance levels: * (0.025% significance); ** (0.010% significance); *** (0.005% significance). 320 r.j. kish / financial services review 25 (2016) 303–329 adding to the diversification effect of the cat bonds is the low level of correlation between these two cat bond indices and the bb-rated, b-rated, and ccc-rated ml corporate bond indices. correlation between micrix and the bb, b, and ccc corporate bond indices are approximately 0 (i.e., �0.016, �0.103, and �0.124, respectively). for the eurekahedge ils index, the correlations are similar but slightly positive (i.e., 0.117, 0.016, and 0.113, respectively).24 further evidence of the diversification effects of an investment in cat bonds is the lack of correlation between these two broad cat bond proxies and the s&p 500 index and a one to three year maturity us treasury index. the correlations are summarized in table 8. a second broad comparison was also undertaken using swiss re cat bond indices. swiss re reports results from three indices: a bb rated cat bond index, a global cat bond index, and a u.s. based cat bond index. the results are similar to the results from the previous proxies. the means, standard deviations, and sharpe ratios are similar: 0.48%/ 0.90%/0.1963, 0.61%/0.80%/0.3894, and 0.63%/0.88%/0.3714, respectively, for the bb, global, and u.s. cat bond indices. see the last three columns of table 5 for a summary of the results. using the means of the differences hypothesis testing, there is no significant results indicating that speculative bond proxies outperformed the cat bond proxies. with the exception of one test, all the results were negative with a few times periods showing significant results favoring the cat proxies. again, the correlations (shown in table 8) reinforce the diversification effects offered by the cat funds. correlations with the three cat fund proxies and the speculative bond indices ranged from 0.008 (swiss re bb and corporate bb) to �0.052 (swiss re bb and corporate b). the null hypothesis test results of the mean of the differences of monthly returns are summarized in the last three columns of table 7. 5.2. cat funds our final set of proxies for the performance of cat bonds are a number of off-shore hedge funds investing primarily in cat bonds. table 9 summarize the monthly returns from a table 8 correlations for cat indices spx busy13 bb b ccc swiss bb swiss g swiss us micrix ils spx 1.000 �0.166 0.743 0.727 0.596 �0.136 �0.152 �0.161 �0.092 �0.067 busy13 1.000 0.172 �0.046 �0.061 0.159 0.195 0.204 0.100 0.241 bb 1.000 0.894 0.772 0.008 0.034 0.029 �0.016 0.117 b 1.000 0.928 �0.052 �0.003 �0.012 �0.103 0.016 ccc 1.000 0.062 0.111 0.100 �0.124 0.113 swiss bb 1.000 0.969 0.971 0.707 0.924 swiss g 1.000 0.998 0.682 0.940 swiss us 1.000 0.697 0.939 micrix 1.000 0.744 ils 1.000 note: cat � catastrophe; ils � investment linked securities; micrix � mercury investible catastrophe risk index. 321r.j. kish / financial services review 25 (2016) 303–329 t ab le 9 c at as tr op he (c a t ) fu nd s hi st or ic al m on th ly pe rf or m an ce ja n. fe b. m ar . a pr . m ay ju ne ju ly a ug . se pt . o ct . n ov . d ec . a z 20 16 � 0. 21 % 0. 27 % 0. 06 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 0. 11 % � 0. 07 % � 0. 19 % � 0. 15 % � 0. 11 % � 0. 11 % � 0. 11 % 0. 21 % 0. 28 % � 1. 03 % 0. 17 % � 0. 28 % 20 14 0. 17 % 0. 08 % 0. 13 % 0. 02 % � 0. 06 % � 0. 06 % 0. 08 % 0. 26 % 0. 06 % 0. 13 % 0. 09 % 0. 08 % 20 13 0. 35 % 0. 37 % 0. 31 % 0. 31 % 0. 11 % � 0. 02 % 0. 04 % 0. 29 % 0. 36 % 0. 19 % 0. 06 % 0. 08 % 20 12 0. 02 % � 0. 16 % 0. 06 % 0. 20 % 0. 48 % 0. 38 % 0. 43 % 0. 31 % 0. 41 % � 0. 08 % 0. 10 % 0. 39 % m ea n 0. 09 % st an da rd de vi at io n 0. 25 % m ax im um 0. 48 % m in im um � 1. 03 % g a m 20 16 0. 02 % 0. 18 % 0. 29 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 0. 23 % 0. 05 % � 0. 03 % � 0. 04 % � 0. 12 % � 0. 16 % 0. 45 % 0. 71 % 1. 05 % 0. 25 % 0. 10 % � 0. 08 % 20 14 0. 57 % 0. 42 % 0. 48 % 0. 34 % � 0. 15 % 0. 15 % 0. 38 % 0. 78 % 1. 11 % 0. 71 % 0. 04 % � 0. 05 % 20 13 0. 77 % 1. 15 % 0. 93 % 1. 01 % 0. 43 % 0. 10 % 0. 48 % 1. 00 % 1. 35 % 0. 90 % 0. 51 % 0. 39 % 20 12 0. 17 % 0. 07 % 0. 29 % 0. 27 % 0. 97 % 1. 82 % 0. 31 % 1. 27 % 1. 55 % � 0. 75 % 0. 93 % 1. 22 % m ea n 0. 49 % st an da rd de vi at io n 0. 51 % m ax im um 1. 82 % m in im um � 0. 75 % l g t 20 16 � 0. 45 % � 0. 54 % 1. 08 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 1. 28 % 0. 34 % � 1. 68 % 1. 71 % � 0. 34 % � 1. 01 % 0. 43 % 1. 51 % 5. 99 % 1. 18 % � 1. 96 % � 4. 10 % 20 14 3. 13 % 3. 06 % 3. 07 % 1. 11 % � 3. 13 % � 2. 37 % 2. 05 % 4. 85 % 8. 07 % 1. 26 % � 0. 83 % � 1. 59 % 20 13 12 .9 1% 15 .9 9% 7. 31 % 9. 81 % 6. 50 % � 0. 14 % 0. 83 % 8. 74 % 8. 16 % 6. 07 % 1. 83 % 4. 77 % m ea n 2. 69 % st an da rd de vi at io n 4. 45 % m ax im um 15 .9 9% m in im um � 4. 10 % pl en um 20 16 0. 63 % 0. 52 % 1. 86 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 1. 23 % 0. 27 % � 0. 27 % � 0. 74 % 0. 54 % � 0. 13 % 0. 81 % 2. 95 % 6. 26 % 2. 64 % 0. 41 % 4. 42 % 20 14 2. 92 % 1. 92 % 2. 94 % 2. 78 % � 1. 01 % � 2. 27 % � 0. 88 % 3. 22 % 8. 36 % 5. 98 % 0. 55 % � 0. 88 % 20 13 5. 40 % 11 .3 9% 9. 77 % 11 .9 3% 3. 75 % 0. 90 % 1. 78 % 4. 73 % 12 .1 7% 3. 93 % 2. 43 % 2. 37 % m ea n 2. 96 % st an da rd de vi at io n 3. 64 % m ax im um 12 .1 7% m in im um � 2. 27 % sc hr od 20 16 � 0. 03 % 1. 72 % 2. 83 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 4. 56 % 1. 08 % � 4. 44 % 2. 79 % � 5. 38 % � 8. 69 % 12 .0 4% 36 .1 6% 19 .6 2% � 1. 97 % 5. 75 % 0. 40 % 20 14 37 .7 6% 20 .9 6% 11 .9 2% � 33 .2 5% 13 .0 8% 25 .6 8% 64 .8 8% 32 .9 8% 32 .2 4% � 8. 96 % � 3. 56 % m ea n 10 .0 1% st an da rd de vi at io n 19 .8 1% m ax im um 64 .8 8% m in im um � 33 .2 5% so lid um 20 16 � 0. 43 % 6. 74 % 1. 89 % n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a n/ a 20 15 1. 25 % 0. 14 % 0. 08 % 0. 28 % � 0. 86 % � 0. 79 % 0. 59 % 1. 83 % 2. 84 % � 4. 73 % 0. 46 % 1. 12 % 20 14 1. 59 % 1. 00 % 0. 91 % 0. 65 % � 0. 88 % 0. 44 % 0. 93 % 2. 81 % 1. 39 % 2. 27 % � 0. 04 % 2. 04 % 20 13 3. 02 % 4. 72 % 4. 13 % 2. 14 % 1. 00 % � 0. 64 % 0. 75 % 4. 42 % 4. 86 % 2. 90 % 1. 56 % 1. 00 % m ea n 1. 37 % st an da rd de vi at io n 1. 99 % m ax im um 6. 74 % m in im um � 4. 73 % 322 r.j. kish / financial services review 25 (2016) 303–329 sample of six funds representing the six fund families: az cat bond (8 funds), gam cat bond (14 funds), lgt cat bond (22 funds), plenum cat bond (15 funds), schroder cat bond (12 funds), and solidum cat bond (4 funds). see the appendix for a fuller description of the funds within lgt cat bond fund family as representative of the range of funds found within each of the cat fund families.25 note that funds within each fund family may differ by currency, fee structure (front and back loads, management fees, and performance fees), inception dates, and country of incorporation. several of the individual funds are relatively new (i.e., only established in 2015). fund returns have been mixed over time. the monthly returns, reported for a represented fund within each fund family, are very variable as shown by the range of monthly returns that vary from extremes in the schroder fund (minimum �33.25%; maximum 64.88%; mean 10.01%; standard deviation 19.81%) to the more compact set of returns fund within the az fund (minimum �1.03%; maximum 0.48%; mean 0.09%; standard deviation 0.25%). monthly return values are reported in table 9. the percentage of months with negative returns ranged from a low of 15.7% (gam cat bond) to a high of 29.3% (lgt cat bond). other relevant factors show low betas (0.3) for all six funds which supports the cat bonds role in diversifying a portfolio and alphas of approximately 0. more summary statistics by time period are reported in table 10. a cautionary factor is that not every fund within the fund family duplicates these results. the summaries are just representative of the results. all table 10 summary statistics for catastrophe (cat) funds period statistic az gam lgt plenum schroder solidum whole mean 0.10% 0.49% 3.25% 3.39% 10.01% 1.48% � 0.24% 0.51% 5.05% 4.48% 19.81% 2.09% max 0.48% 1.82% 17.09% 19.85% 64.88% 6.74% min �1.03% �0.75% �4.10% �2.27% �33.25% �4.73% positives 39 43 29 33 18 33 sharpe �0.8332 0.3655 0.5375 0.7320 0.4901 0.5377 alpha �0.001 0.076 0.057 0.032 0.147 0.040 beta 0.337 0.330 0.372 0.333 0.316 0.325 count 53 51 41 40 26 40 1 year mean �0.11% 0.20% 0.28% 1.53% 5.16% 0.19% � 0.34% 0.37% 2.49% 2.12% 12.46% 1.86% positives 4 7 7 9 8 9 2 years mean �0.01% 0.30% 0.92% 1.75% 11.11% 0.64% � 0.26% 0.37% 2.89% 2.61% 20.85% 1.54% positives 14 17 15 17 16 19 3 years mean 0.06% 0.45% 2.91% 3.13% 1.26% � 0.25% 0.43% 4.56% 3.74% 1.83% positives 25 29 26 29 30 4 years mean 0.10% 0.51% � 0.25% 0.52% positives 35 40 inception: earliest 11/1/2011 10/31/2011 5/1/2001 9/6/2010 10/21/2013 9/30/2009 inception: latest 10/1/2013 8/31/2015 12/16/2013 9/29/2014 3/31/2015 4/5/2012 note: 1 year period (1/2015–12/2015); 2 year period (1/2014–12/2015); 3 year period (1/2013–12/2015); 4 year period (1/2012–12/2015); and whole period varies by fund. 323r.j. kish / financial services review 25 (2016) 303–329 the cat funds have positive correlations among each other ranging from 0.44 (az and schroder) to 0.86 (az with solidum). this offers support that the funds are operating within the same set of investments. the cat funds show low correlations with the ml speculative grade bond funds (range �0.128 to 0.440 with most correlations close to 0). correlations to the s&p500 index are also close to 0. see table 11 for the complete correlation table. similar to the earlier analysis, the null hypothesis of the test of the mean of the differences predicting no advantage of to either the cat fund or the speculative rated corporate funds show in general that the cat funds outperformed the speculative grade corporate bond indices as indicated by their negative values except within the analysis of the outlier fund az cat fund. for example, plenum cat bond fund shows significantly higher returns over all 3 speculative bond funds (bb, b, and ccc) during the whole period and over both the twoand three-year periods. however, the az cat fund shows positive values indicating subpar performance when measured against the speculative grade corporate bond sector. unlike the previous comparisons for the broad market cat proxies, there are many cases of statistically significant results. a summary of the results are shown in table 12. in general, the null hypothesis of equal returns between the speculative bond sector and the cat bond sector could not be supported. 6. conclusion cat bonds were created to give insurance firms, reinsurance firms, and government entities an additional option for hedging catastrophic risks from disasters (both natural and man-made). these risks are documented as extremes further justifying the need for an offset. through a set of proxies (indices and hedge funds), the returns generated more than offset the risks undertaken by investors, as indicated by the results from the mean of the difference tests. the cat bond indices produced similar to superior monthly returns versus those within the speculative grade bond sector. with returns at or above the speculative grade bond sector coupled with a low probability of the bondholder losing both interest and principal if the threshold losses experienced by the underlying insurance firm are met offer an additional investment opportunity for a diversified investment strategy. thus, cat bonds which table 11 correlations spx busy13 bb b ccc az gam lgt plenum schrod solidum spx 1.000 �0.166 0.743 0.727 0.596 �0.242 0.044 0.142 0.088 �0.050 �0.166 busy13 1.000 0.172 �0.046 �0.061 0.207 0.205 �0.004 0.166 0.095 0.231 bb 1.000 0.894 0.772 �0.128 0.170 0.210 0.137 0.021 �0.016 b 1.000 0.928 �0.059 0.136 0.299 0.132 �0.029 0.002 ccc 1.000 0.126 0.229 0.437 0.179 �0.067 0.103 az 1.000 0.601 0.568 0.469 0.440 0.863 gam 1.000 0.838 0.846 0.796 0.751 lgt 1.000 0.789 0.700 0.672 plenum 1.000 0.518 0.652 schroder 1.000 0.593 solidum 1.000 324 r.j. kish / financial services review 25 (2016) 303–329 provide returns comparable to the risk undertaken with the added benefit of diversification could be used to supplement a small portion of a diversified portfolio. notes 1 the un office for disaster risk reduction reports economic losses from natural disasters average between $250 billion and $300 billion annually. this average includes two components: (1) the international insurance industry’s global loss ($180 billion to $200 billion) and (2) $70 billion to $100 billion in losses from developing countries from smaller-scale floods, fires, and storms. http://globalnews.ca/news/1865514/economic-losses-from-natural-disasters-between-250-billion-and-300-billion-un/ 2 for data on the afghanistan and pakistan earthquake see http://www. huffingtonpost.com/entry/afghan-quake-death-toll-mounts-as-survivors-still-waitfor-aid_5630c8a4e4b063179910268d 3 swiss re sigma no. 2/2015 http://media.swissre.com/documents/sigma2_2015_en_ final.pdf 4 see http://www.artemis.bm/library/what-is-a-catastrophe-bond.html for more information. 5 see http://www.claimsjournal.com/news/national/2011/09/09/190969.htm 6 at the end of 2015, there were over $25.9 billion of cat bonds outstanding. http://www.finra.org/investors/alerts/catastrophe-bonds-and-other-event-linkedsecurities 7 a comprehensive database containing the details of these cat bonds can be found at http://www.artemis.bm/deal_directory/ table 12 means of the differences tests for catastrophe (cat) funds t-values az gam lgt plenum schroder solidum bb corp whole 2.581*** �0.126 �3.763*** �4.359*** �2.518** �3.298*** 1 year 0.365 �0.436 �0.325 �1.858 �1.375 �0.182 2 year 0.631 �0.778 �1.282 �2.720** �2.530** �1.175 3 year 1.202 �1.100 �3.560*** �4.594*** �2.838*** 4 year 2.424** �0.430 b corp whole 1.194 �0.873 �4.118*** �4.623*** �2.608*** �3.550*** 1 year �0.790 �1.346 �0.963 �2.143 �1.468 �0.780 2 year �0.724 �1.742 �1.805 �3.021*** �2.607** �1.745 3 year 0.053 �1.754 �3.917*** �4.836*** �3.179*** 4 year 1.202 �1.052 ccc corp whole 0.543 �0.464 �4.228*** �4.551*** �2.682*** �2.937*** 1 year �1.641 �2.077 �1.857 �2.536* �1.651 �1.412 2 year �1.862 �2.592** �2.584** �3.395*** �2.728** �2.433* 3 year �0.673 �1.914 �4.492*** �5.054*** �3.339*** 4 year 0.582 �0.826 note: significance levels: * (0.025% significance); ** (0.010% significance); *** (0.005% significance). 325r.j. kish / financial services review 25 (2016) 303–329 8 the investment managers held 21.32% of the securities outstanding. the five leading investment managers are stone ridge asset management llc (26.78%); unicredit spa (15.63%); jp morgan chase & co. (10.61%); clariden lue ltd (9.15%); and falcon fund management ltd (7.03%). 9 since 2008, the sec has granted exceptions to allow resales after a 6-month window. see http://media.mofo.com/files/uploads/images/faqrule144a.pdf for answers to frequently asked questions about rule 144a. 10 the correlation of a representative number of cat funds, used as proxies for the cat bonds, with the market is approximately 0. see tables 8 and 11. 11 see swiss re (2012b). 12 see http://www.finra.org/investors/alerts/catastrophe-bonds-and-other-event-linkedsecurities 13 see following 3 web links: http://www.bankrate.com/finance/insurance/top-10-costliestnatural-disasters-11.aspx; : http://citywire.co.uk/money/the-world-s-10-biggest-insuranceclaims/a433781#i�10; and http://www.iii.org/fact-statistic/catastrophes-global 14 see mcchristian (2012) and cummins (2007). 15 see http://www.propertycasualty360.com/2013/03/15/top-5-catastrophe-bond-markettrends?page�5 16 see http://www.insurancejournal.com/news/southeast/2012/08/21/2559960.htm 17 see http://www.florir.com/pressreleases/viewmediarelease.aspx?id�2069 18 see http://www.sbafla.com/fhcf/portals/5/advisory%20council/20150514_fhcf_ may2015bondingcapacity.pdf; 19 see http://www.artemis.bm/blog/2015/03/02/reinsurance-the-engine-of-berkshire-hath away-warren-buffett/ and http://www.artemis.bm/blog/2014/03/15/warren-buffett-u-scatastrophe-rates-too-low-for-berkshire-hathaway/ for summaries and http://berkshire hathaway.com/letters/2014ltr.pdf for the actual buffett letter. 20 see http://www.finra.org/investors/alerts/catastrophe-bonds-and-other-event-linkedsecurities 21 see insurance journal article “pension funds looking for higher yields from cat bonds.” http://www.insurancejournal.com/news/international/2013/04/09/287723.htm 22 two other speculative bond indices (bank of america/merrill lynch 1–3 year u.s. cash pay fixed maturity high yield constrained index—j1hc and bank of america/ merrill lynch 1–3 year u.s. cash pay high yield index—j1a0) were used with similar results. 23 the differences between two populations of means is not an appropriate test. the null hypothesis that there is no difference in the mean returns from the cat bonds proxies (mutual funds and indices) and similarly rated corporate debt (bb, b or ccc: speculative grade debt or junk bonds) matched with the alternative hypothesis is that there is a significant difference in their monthly returns (i.e., h0: (�cat � �junk) � 0 versus ha: (�cat � �junk) � 0) fails to consider the timing of the returns. thus, a test that relies on the two-sample t statistic where sp 2 � �ncat � 1�scat 2 � �njunk � 1�sjunk 2 ncat � njunk � 2 and 326 r.j. kish / financial services review 25 (2016) 303–329 t � �x�cat � x�junk� � 0 �sp 2� 1 ncat � 1 njunk � is not appropriate for this analysis because the assumption of independent samples is invalid. 24 the correlation between the ils-eurekahedge index and the bb-rated and b-rated debt was 0.21 and 0.17, respectively. for the micrix index the corresponding correlation values were 0.002 and 0.013. the analysis was also undertaken against the ccc bond index with similar results. 25 a listing of the funds within each of the six cat fund families is available from the author. appendix: fund summaries lgt cat bond fund is a family of funds. the first group is an open-end fund incorporated in luxembourg. the objective of the fund is to achieve a return in the reference currency of the individual share classes in excess of the three-month money market (libor). the fund invests in insurance-linked securities of all kinds that are traded on a stock exchange. this fund family consists 11 of separate funds: luxembourg based: (1) lgt-cat bd—chf c (figi bbg003fwf7n9; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (2) lgt-cat bd—eur b (figi bbg003fwdrx6; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (3) lgt-cat bd—usd c (figi bbg003fwf6t5; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (4) lgt-cat bd—eur b2 (figi bbg003fwf2v1; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (5) lgt-cat bd—b eur (figi bbg003fwf1h9; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class retail); (6) lgt-cat bd—eur b2 (figi bbg003fwf2d1; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (7) lgt-cat bd—chf b2 (figi bbg003fwf544; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (8) lgt-cat bd—eur c (figi bbg003fwf759; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class institutional); (9) lgt-cat bd—chf b (figi bbg003fwf1z9; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 9/28/2012; share class retail); (10) lgt lux i-cat bond fund—imusd (figi bbg003qtsbh8; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 12/27/2012; share class institutional); and (11) lgt-cat bd— chf f (figi bbg007qvl1v7; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 12/22/2014; share class retail). 327r.j. kish / financial services review 25 (2016) 303–329 the second group of funds are based in switzerland. this open-end fund invests in a broadly diversified portfolio of catastrophic bonds. the fund’s objective is a stable return that is above the money-market yield and has only a low correlation to other movements on the financial markets. this fund family consists five of separate funds: switzerland based: (s1) lgt ch-cat bond fund—eur a (figi bbg000lxpp00; front load 3.0% back load 0.0%; management fee 1.75%; performance fee 0.0%; inception date 5/01/2001; share class retail); (s2) lgt ch-cat bond fund—usd a (figi bbg000lxppp3; front load 3.0%; back load 0.0%; management fee 1.75%; performance fee 0.0%; inception date 5/01/2001; share class retail); (s3) lgt ch-cat bond fund—chf a (figi bbg000lxpb47; front load 3.0%; back load 0.0%; management fee 1.75%; performance fee 0.0%; inception date 5/01/2001; share class retail); (s4) lgt ch-cat bond fund—chf ia (figi bbg000vbxyz1; front load 2.5%; back load 0.0%; management fee 1.25%; performance fee 0.0%; inception date 2/29/2008; share class institutional); and (s5) lgt ch-cat bond fund—eur 1a (figi bbg000v1tvf6; front load n.a. back load n.a.; management fee n.a.; performance fee n.a.; inception date 5/31/2010; share class institutional). the third group of funds are based in liechtenstein. the fund invests in a broadly diversified portfolio of catastrophic bonds. this open-end fund’s objective is a stable return above the money-market yield and has a low correlation to other movements on the financial market. this fund family consists six of separate funds: liechtenstein based: (l1) lgt select-cat bond—im (figi bbg005t7cjj6; front load n.a.; back load 0.0%.; management fee n.a.; performance fee 0.0%; inception date 12/16/2013; share class institutional); (l2) lgt lie-cat bond fund—chf b (figi bbg000nfwsz5; front load n.a. back load n.a.; management fee n.a.; performance fee n.a.; inception date 7/31/2009; share class retail); (l3) lgt lie-cat bond fund—usd b (figi bbg0030nfwxd8; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 7/31/2009; share class retail); (l4) lgt lie-cat bond fund—eur d (figi bbg000nfxzy9; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 7/31/2009; share class institutional); (l5) lgt lie-cat bond fund—eur b (figi bbg000nfwvn1; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 7/31/2009; share class retail); and (l6) lgt lie-cat bond fund—chf d (figi bbg000nfwxy5; front load n.a.; back load n.a.; management fee n.a.; performance fee n.a.; inception date 7/31/2009; share class institutional). references bantwal, v, & kunreuther, h. (2000). a cat bond premium puzzle? journal of psychology and financial market, 1, 76–91. barrieu, p., & louberge, h. (2009). hybird cat bonds. journal of risk and insurance, 76, 547–578. cummins, j. d. (2008). cat bonds and other risk-linked securities: state of the market and recent developments. risk management and insurance review, 11, 23–47. cummins, j. d. (2012). cat bonds and other risk-linked securities: product design and evolution of the market. the geneva reports: risk and insurance research, 5, 39–61. 328 r.j. kish / financial services review 25 (2016) 303–329 cummins, j. d., lalonde, d., & phillips, r. d. (2004). the basis risk of catastrophic-loss index securities. journal of financial economics, 71, 77–111. cummins, j. d., & trainar, p. (2009). securitization, insurance, and reinsurance. journal of risk and insurance, 76, 463–492. dieckmann, s. (2011). a consumption-based evaluation of the cat bond market. working paper (darden school of business, university of virginia). durbin, d. (2001). managing natural catastrophe risks: the structure and dynamics of reinsurance. the geneva papers on risk and insurance, 26, 297–309. finken, s., & laux, c. (2009). catastrophe bonds and reinsurance: the competitive effect of informationinsensitive triggers. journal of risk and insurance, 76, 597–605. froot, k. a. (2001). the market for catastrophe risk: a clinical examination. journal of financial economics, 60, 529–571. froot, k. a., & o’connell, p. g. (2008). on the pricing of intermediated risks: theory and application to catastrophe reinsurance. journal of banking & finance, 32, 69–85. hagendorff, b., hagendorff, j., & keasey, k. (2013). the shareholder wealth effects of insurance securitization: preliminary evidence from the catastrophe bond market. journal of financial services research, 44, 281–301. hagendorff, b., hagendorff, j., keasey, k., & gonzalez, a. (2014). the risk implications of insurance securitization: the case of catastrophe bonds. journal of corporate finance, 25, 387–402. jarzabkowski, p., bednarek, r., & spee, p. (2015). making a market for acts of god: the practice of risk-trading in the global reinsurance industry. cambridge: oxford university press. lakdawalla, d., & zanjani, g. (2012). catastrophe bonds, reinsurance, and the optimal collateralization of risk transfer. the journal of risk and insurance, 79, 449–476. mcchristen, l. (2012). hurricane andrew and insurance: the enduring impact of an historic storm. insurance information institute, august 2012, 1–19. subramanian, a. & wang, j. (2015). catastrophe risk transfer. working paper. (available at http://ssm.com/ abstract�2321415). swiss, r. (2012a). what are insurance linked securities (ils), and why should they be considered? presentation to the cane fall meeting, september. 1–12. swiss, r. (2012b). long point re iii case study. presentation to the cane fall meeting, september 13–17. 329r.j. kish / financial services review 25 (2016) 303–329 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 retirement income strategies designed in an expected utility framework mark j. warshawskya,* arelias llc, 903 brentwood lane, silver spring, md 20902, usa abstract various classes of retirement income strategies are evaluated and their robustness tested in an expected utility framework. fixed percentage systematic withdrawals from an investment portfolio combined with laddered purchases of immediate life annuities stands out as a superior strategy for retired defined contribution plan participants and ira holders, yielding better outcomes than alternatives, including longevity insurance. this broad strategy is then applied, step by step, and customized across a wide range of household risk preferences and situations, with due consideration of product costs, taxes, and economic risks and returns. © 2017 academy of financial services. all rights reserved. jel classifications: g11; g22; d14 keywords: retirement income strategies; systematic withdrawal; life annuity; advanced life delayed annuity (alda) 1. introduction products and strategies are emerging in the market to help retirees invest and draw down assets from their 401(k) accounts and iras during retirement. given the competing desires for lifetime income and wealth, for security and flexibility, and for current and future needs, it is appropriate to compare solutions carefully. this article gives a quantitative analysis of different strategies for households, including systematic withdrawals, immediate life annuities, mixed strategies of systematic withdrawals with life annuities, and also the use of * corresponding author. tel.: �1-301-908-2846; fax: �1-301-592-9277. e-mail: markjwarshawsky@gmail.com financial services review 26 (2017) 113–141 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. longevity insurance (advanced-life delayed annuities, alda). the alda product has been featured in other research, and in a recent amendment to the required minimum distribution (rmd) rules for retirement accounts in the united states. as realistically as possible, this articles considers fee charges and pricing for various products, income flows from social security and pensions, taxes, and random realizations of asset returns, interest rates, and mortality (normal and impaired life). results are computed from a model using an expected utility framework, which assigns a different utility value to any small change in consumption or final wealth levels. the robustness of strategies is tested across households with varying preferences and circumstances. a household’s chosen strategy is assumed to remain fixed throughout retirement. this essentially acknowledges the complexity of issues facing older retirees, such as their trust in financial institutions and their cognitive agility to make midcourse changes, as well as the governance capacities of financial providers and advisors. the idea of simplifying and easing the path for investors is also found in the fixed strategies used by the popular target-date funds and managed accounts during the accumulation phase. at the same time, it is possible to reoptimize at any time in response to major changes in the retired household’s situation, such as the death of a spouse. a more dynamic strategy cognizant of the practical complexities of financial and insurance products and taxes would necessitate a much more complex model which, in turn, would be more opaque to retirees. among the many options considered, a new strategy stands out: combining laddered purchases of nominal immediate life annuities (that is, dollar cost averaging) with systematic withdrawals (fixed percentage) from a dynamically changing retirement investment portfolio. this combination strategy outperforms the alternatives using alda and inflation-indexed annuities. 2. evolution of thinking on income strategies there is growing research literature on retirement income solutions. in a simple loss aversion framework, pang and warshawsky (2009) compared products and strategies for producing income and managing wealth in retirement accounts, assuming an initial 50–50 equity-bond allocation. the strategies compared ran the spectrum of liquidity and potential growth.1 at one end, systematic withdrawals as a fixed percentage of the retirement account provides complete liquidity and growth potential but no guarantees, and significant risk. at the other end, full annuitization using a straight immediate (nominal) life annuity provides no liquidity or growth potential but is fully guaranteed for life in nominal terms. midspectrum strategies include a mix of systematic withdrawals and gradual annuitization over ten years, a deferred variable annuity with a guaranteed minimum withdrawal benefit rider (“va�gmwb” with a 5% withdrawal rate), and an immediate variable life annuity invested in bonds and equities. in general, among the strategies analyzed, the highest real income and lowest chance of income shortfalls are achieved by combining systematic withdrawals and gradual but complete annuitization over ten years—that is, the annuity purchase ladder strategy. only the purchase of an immediate fixed annuity at retirement, using the entire account balance, achieved a higher real income, at the median outcome, but that strategy leaves no wealth 114 m.j. warshawsky / financial services review 26 (2017) 113–141 balances ever. the ladder strategy maintains some balances through the first 10 years of retirement. the other strategies (excluding the immediate variable annuity) offer the advantage of significant real balances (i.e., liquidity). the liquidation value in va � gmwb tends to run to zero in later life, owing to withdrawals and fees. a more aggressive portfolio (70–30 equity-bond split) gives a small relative boost to the va�gmwb, because the investor takes greater advantage of the insurer’s guarantee, but the basic results remain. the relative levels of fees are critical to the analysis. with an overall low level of fees, perhaps owing to institutional pricing, as long as the fees are reduced proportionally across products, the advantage of the annuity ladder strategy still holds. warshawsky (2012, chapter 7) and warshawsky (2016) are the main studies that develop an analysis of the laddered annuity purchase strategy. their focus is on optimizing the ladder strategy of immediate annuity purchases, both because it compared well in earlier research with other strategies and because it closely resembles the optimum in the theoretical fully dynamic model of pang and warshawsky (2010) described below. the components of strategy design optimization include the withdrawal rate, and the length and extent of annuitization. pang and warshawsky (2010) developed a formal theoretical, somewhat stylized, dynamic model of expected utility maximization in retirement, where the household optimizes consumption and allocates wealth across equities, bonds, and life annuities, with consideration of the receipt of lifetime income flows such as social security and defined benefit pensions. in addition to stochastic capital market returns and mortality, households are exposed in this model to uninsured health care cost risks, which increase with age and income decile. absent a bequest motive, the results indicate that retired households should optimally start annuitizing their wealth around their mid-70s and annuitize fully in their 80s. retirees continue purchasing annuities throughout their lives as they save some income for sequential purchases of higher yielding life annuities to effectively insure for higher health care expenses later in life. moreover, with the use of guaranteed annuities, the optimal equity exposure in the remaining portfolio increases with age, reaching nearly 100% for highincome households, until nonannuitized wealth is used up. the consumption level is fairly sustained over the retirement life cycle with the support of annuity income. a modest bequest motive tempers these results, in particular cutting the ultimate use of life annuities, but does not overturn them. in the face of uncertain health care expenses, it seems logical to expect, and some studies indeed show, immediate life annuities to be a poor investment, owing to their lack of liquidity. however, uninsured health care spending exposure is highest later in life, especially for long-term care needs, precisely when newly purchased immediate life annuities (if available) generate their highest returns, owing to large and growing-with-age mortality credits. hence, immediate life annuities are a hedge for long-term care and health spending in the absence of complete health and long-term care insurance coverage. note that this point is orthogonal to the issue of the current health status and, therefore, life expectancy of the potential insured and the appropriateness of the current purchase of immediate life annuities. dus, maurer, and mitchell (2005, dmm) analyze various strategies, including an immediate real (that is, inflation-indexed) life annuity, delayed annuitization and systematic withdrawals. the withdrawal rate is determined according to a fixed dollar benefit level. 115m.j. warshawsky / financial services review 26 (2017) 113–141 under a fixed dollar formula, benefits are paid until the plan participant dies or funds are exhausted. alternate variable formulas are based on fixed percentages or relate to remaining life expectancy. their evaluative approach is broadly similar to warshawsky (2012, chapter 7) using shortfall risk of expected payouts and real balances. if no annuities are available, the optimal strategy is a fixed percentage withdrawal of 7.3%, and the optimal equity share is 75%—remarkably close to the results mentioned above. dmm compare these results to a fixed payment systematic withdrawal equal to the immediate real life annuity payout. focusing on the expected present value of benefits paid (a concept similar to money’s worth in the annuity pricing literature—the larger, the better) as the best single evaluative measure, the fixed percentage approach is shown to be superior to the fixed dollar approach.2 dmm find that annuitization is more appealing to older retirees, as compared with systematic withdrawals, another result quite similar to those summarized above. they infer that a strategy of systematic withdrawals followed by full annuitization at age 75 or 85 (delayed annuitization) increases the expected present value of benefits and shrinks the expected present value of shortfall income. this evokes the annuity ladder. finally, they evaluate the immediate purchase of a delayed annuity (to pay at age 75 or 85) at the beginning of the retirement period—the longevity insurance strategy advocated by some analysts and market participants. dmm find that for retirees who desire any bequests or liquidity, these outcomes are generally inferior to delayed annuitization, particularly for longevity insurance paying at age 85. moreover, dmm assumed the load on the delayed annuity is the same as for immediate annuities, whereas empirical evidence finds that immediate life annuities have lower loads (see below). a related article by horneff, maurer, mitchell, and dus (2008) uses a utility-based framework, with stochastic capital markets (but not inflation) and uncertain lifetimes, to evaluate phased withdrawal plans and fixed payout annuities. they find that the fixed benefit rule performs poorly, running out of funds by age 80, is consistent only with very low levels of risk aversion, and is dominated by other payout rules (the fixed percentage in particular) and by the life annuity. horneff et al., also consider combination strategies, in particular the life annuity and the life expectancy withdrawal rule. they find that as risk aversion increases above relatively low levels, the devotion of a significant share of wealth (approaching 90% with no bequest motive) to the immediate life annuity increases welfare significantly. for moderate risk aversion, it is better to delay annuitization until around age 80. economists refer to the actual dearth of voluntary annuitization—despite the evidence of economic theory and simulations showing the high utility value of immediate life annuities as insurance against outliving wealth—as the “annuity puzzle.” various explanations have been suggested. brown and warshawsky (2004), dushi and webb (2004), and inkmann et al., (2011), among others, show that the extent of annuitization is affected by the availability of annuities in retirement plans, the share of wealth represented by social security and defined benefit plans, the load on life annuities arising from adverse selection, extra marketing costs and other factors, levels of financial wealth, life expectancy, education, and bequest motive. behavioral biases are now favored as an explanation of the annuity puzzle, prompting calls for new strategies and public policies to overcome cognitive blocks. one such proposed 116 m.j. warshawsky / financial services review 26 (2017) 113–141 strategy is longevity insurance, a deeply deferred life annuity, also known as an advancedlife delayed annuity (alda). an alda is purchased at retirement but payouts do not begin until the retiree reaches an advanced age, usually 85. it has been promoted by milevsky (2005) and scott (2008), as well as by the prior, obama, administration.3 the alda strategy claims to provide liquidity through most of retirement, that is, partial annuitization with longevity risk coverage late in life for a premium that might be perceived as a cheap price. for example, sexauer et al. (2012) show that in 2010, a 65-year-old retiree would have needed “only” 12% of the portfolio to purchase a deferred (albeit nominal) annuity such that the first payout at age 85 would equal the last payout from a self-amortizing (over 20 years) portfolio of laddered treasury inflation protected securities (tips). haensly and pai (2015) have also examined the tips/alda strategy. it should be noted though that there is no upside investment potential with this strategy and it may not therefore be desirable to many retired investors. gong and webb (2010) compare longevity insurance with complete immediate annuitization, delaying the complete annuitization to a late age, and systematic withdrawals; they do not consider ladders of annuity purchases. they use an expected utility model, but with no risky assets or bequest motive, and assume all annuities are loaded equally. with retirement at age 65 and moderate risk aversion, gong and webb find that complete immediate annuitization wins out over longevity insurance (although gong and webb say they favor the alda). gong and webb also calculate loads on commercial annuity products using the standard money’s worth methodology of mitchell et al. (1999). they find that the load is about five percentage points higher on real annuities than on nominal annuities, and that the load is fairly constant across ages 60 through 75 but increases thereafter, again by about five percentage points, based on the best pricing of immediate annuities among three issuers, and a pricing model using treasury bond yields and general population mortality. a comparison of the loads on longevity insurance (paying at age 85) with immediate annuities, on a nominal basis and issued at age 65, finds the load on the deferred annuity to be about five percentage points higher than for the immediate annuity at one insurance company, and about 20 to 30 percentage points higher at another. although these findings reflect pricing on one day only (in january 2008), finkelstein and poterba (2004) reached broadly similar conclusions based on data from a large united kingdom insurance company over a 17-year issuance period through 1998. they find that the load is from five to 10 percentage points higher on real than on nominal immediate annuities. we will use these results in our empirical simulations below. note that even though there is no real alda being sold in the market, i do model it in a couple of the simulations below, and make an inference on what its load would be from the above literature. this literature review has focused on studies bearing directly on retirement income strategies. other recent studies, however, are also quite relevant to this article because they address directly and empirically parameter values for the bequest motive that are only guessed at in the previous simulation literature. ameriks et al. (2011), de nardi et al. (2010), and lockwood (2012) carefully estimate preference parameters in structurally similar lifecycle expected utility models. they find varying degrees of bequest prevalence and strength of the bequest motive. below, we use the average and range of their parameter estimates in 117m.j. warshawsky / financial services review 26 (2017) 113–141 evaluations of income strategies. similarly, survey-based formal evidence on the parameters for risk aversion is found in kimbal et al. (2008). their risk aversion parameter estimates are measured precisely, based on answers to hypothetical risk situations. by contrast, the time preference parameter is less consistently and more widely measured in a large literature, so we hew more closely here to convention and intuition. 3. a utility-maximizing framework for the evaluation of strategies 3.1. lifetime utility i now specify the algorithm or model by which retired households find their optimal strategy. this stochastic simulation optimization model is in the spirit of the theoretical model by pang and warshawsky (2010) but is more realistic of product and market conditions. the particular functional forms used here are chosen to (1) enable the use of empirical results in the literature estimating various parameter values (which assumed these functional forms), (2) they are more easily manipulated in stochastic simulation work than other forms, and (3) are fairly common and long-standing in both the theoretical and empirical literatures. a household is assumed to seek an income and wealth management strategy that maximizes its retirement lifetime utility, which is defined as a function of consumption flows and the bequest amount upon death. its lifetime utility in retirement is expressed as follows: v� � �t�0 ��1 ��thtu�ct ht �� � ��v�b��, where the realized lifetime utility v� depends on the survival from t � 0 (retirement age 65) through � (stochastic, maximum age 105), ct is household consumption and ct/ht is on a per capita basis with ht being the effective number of adults, b� is monetary wealth as bequest upon death, and � � 0.97 is the discount factor (time preference) initially (later we will vary and increase the time preference parameter). ignoring children, we set ht to 1 for single retirees and 20.5 for couples, taking into account economies of scale in consumption. the period utility function of consumption takes the constant relative risk aversion (crra) form: u�c� � c1�� 1�� , where a higher value of � indicates greater risk aversion. the utility from bequest has a similar functional form, as follows: v�b� � ��� � b � �1�� 1�� , 118 m.j. warshawsky / financial services review 26 (2017) 113–141 where � indicates the strength of the bequest motive and � imposes a threshold of consumption (in thousands of dollars) above which the bequest motive becomes operative. as the value of � rises, bequests are increasingly considered a luxury. using rounded averages of parameter values estimated by de nardi et al., (2010) and lockwood (2012),4 we initially set � � 4, � � 30, and � � 10. these parameters depict a modestly conservative attitude of retirees toward risk and a middling motive for a bequest. bequests are feasible for households, but are not necessarily easy to manage with the consumption threshold of $30,000 a year (� � 30). we consider alternative values later in sensitivity tests. the objective of households is to maximize the expected v� over all possible life outcomes of �. the numerical result of v� is a ranking of utility and does not conveniently measure the magnitude of welfare gain or loss of one strategy versus another. therefore, we calculate the so-called certainty equivalent (ce) consumption that would generate the same level of value v� for a life path �. this constant ce consumption, on a per capita basis, is determined such that: �t�0 � ��thtu�ce�� � v�. further, over n possible life paths (i.e., n series of simulations from t � 0 through �), we calculate the average certainty equivalent consumption (ace) as: ace � �j�1 n cej n . the objective of households is now transformed to search for an income distribution and wealth strategy that achieves the highest ace. when evaluating strategies, both across broad categories and for specific implementations within a category, the highest ace should be chosen, for a particular set of preference and economic condition parameters. 3.2. parameterizations and stochastics households (singles or couples) are initially assumed to start with $250,000 in accounts upon retirement. household consumption ct is equal to income that is the sum of systematic withdrawals, annuity payouts, and social security benefits. the social security benefit is initially set to be 12 � $1150for singles or survivors and 12 � $1150 � 1.5 for couples (a 50% spousal benefit). these amounts are based on the 50th percentile of the new social security awards to retired workers in 2010. the bequest b� is set as the fund balance upon death, if greater than zero, plus one month’s social security benefit. households are initially assumed to have an initial 50 –50 equity-bond portfolio mix. this simple allocation, common in many target-date funds at the point of retirement, is chosen so that we can focus on the implications of various withdrawal and annuitization strategies; we leave the analysis of more complex asset allocations, such as international, developing market, real assets, and so on, to future work. as wealth is annuitized, the equity share increases in the remaining assets (up to 100%) so as to maintain roughly the same overall risk exposure.5 this dynamic asset allocation strategy is an approximation to the optimal one resulting from the theoretical model of pang and warshawsky (2010) 119m.j. warshawsky / financial services review 26 (2017) 113–141 mentioned above. although some retirees might initially balk at this high allocation to equities, its riskiness is only apparent and not real because of the substitution of fixed life annuities for bonds; this logic could be explained to retired households by financial advisors. variable returns on equities and bonds are proxied by the s&p 500 and the united states 10-year government bonds total return indexes, respectively. inflation is measured by the change in the cpi-u index. the stochastic dynamics of asset returns and inflation are modeled as a vector autoregressive (var) process, following campbell and viceira (2005). the var coefficients and variance-covariance matrix, estimated on 1962–2011 quarterly data, are embedded in the simulations to generate a large number of multiyear series of rates and returns. this approach captures the serial correlations among variables and the contemporaneous correlations of market shocks. summary statistics are reported in table 1 and simulation details are in the appendix, section a. investment management fees are subtracted from returns. we also allow for insurer bankruptcies, with small probabilities, and stochastic partial policyholder recoveries. the management fee for investments is initially assumed to be a relatively low 25 basis points, which is consistent with a portfolio composed mainly, but not entirely, of indexed equity and bond funds in employer-sponsored retirement accounts or discount iras. for annuity purchases before age 75, a 10% load is assumed for nominal immediate life annuities, a 15% load for real immediate life annuities, a 15% load for the nominal alda, and a 20% load for the real alda (as mentioned above, this latter product does not exist in the market, so it should be regarded as hypothetical here). for annuity purchases after age 75, loads are further assumed to increase linearly by up to 5% until age 85 and flatten off thereafter. these load differentials are consistent with the empirical evidence of age-varying actuarial fairness of annuities reported in gong and webb (2010).6 single households purchase single life annuities, while married couples purchase joint and survivor (j&s) annuities with 75% survivor benefit. the process of annuitization can be gradual for immediate annuities, taking as long as 30 or more years, although in actual practice, it would be appropriate to limit the process up to age 90, in line with current u.s. market practices for the maximum age for immediate annuity sales. for the alda, it is a one-time purchase upon retirement at age 65, with payment commencing at 85—the approach advocated by some analysts, market makers, and policymakers. table 1 summary statistics of simulated annual rates and returns equity return bond return bond yield inflation real (%) mean 4.9 2.8 2.5 — standard deviation 17.8 9.7 2.4 — nominal (%) mean 8.9 6.9 6.5 4.1 standard deviation 17.3 9.0 2.5 2.7 source: author’s simulations based on 1962–2011 data. 120 m.j. warshawsky / financial services review 26 (2017) 113–141 the underlying assets for life annuities are assumed to be invested in nominal bonds. the calculation of annuity price/factor uses the 10-year government bond yield, which is stochastic through time, and life tables for annuitants rather than those for the general population, which reflects adverse selection in the voluntary annuity market. note that this annuity pricing module, including the loads mentioned above, is quite conservative and biases the optimal strategies somewhat away from annuitization. the survivals of households are simulated in the model based on general population mortality rates. 3.3. implementations and strategies considered i now consider several broad strategies, including systematic withdrawals, immediate annuities, and deeply deferred annuities, nominal and real. i model retired investors searching among the implementations for the best strategies, first simple ones and then more complex combinations. 3.3.1. systematic withdrawals: fixed real dollars or fixed percentage of balances retirees can continue to hold investment funds and take periodic systematic withdrawals, which can be constant in real dollars. that is, investors withdraw a certain amount in the first period and adjust the amount for inflation in the following periods. this strategy provides retirees with the same purchasing power over time, but they risk outliving resources at older ages. this approach is related closely to the so-called bengen rule, discussed below. alternatively, withdrawals can be a fixed percentage of the portfolio balance in each period, which will not exhaust the retiree’s wealth and implicitly assumes some selfdiscipline or flexibility on consumption because withdrawals can be very low in adverse investment climates. this approach provides liquidity to investors and bequest potential to their heirs. it allows investors to consume more when funds perform well, but also exposes them to possibly painful declines in consumption when investments fare poorly.7 3.3.2. immediate life annuity: nominal or real life annuities address longevity risk and offer a steady flow of income. lacking an annuity, retirees’ income flow and consumption hinge on how quickly they draw down wealth, how long they live, and their investment outcomes. retirees who consume too quickly may outlive their financial resources, especially given ever-increasing life expectancy, while the overly cautious may consume well below their means. we consider the most widely available nominal immediate life annuities, whose payouts are constant in nominal terms, as well as real life annuities, whose payouts are indexed to inflation, with extra cost loads, as observed in the market. 3.3.3. advanced-life delayed annuity as a relatively recent innovation to insure against longevity, the alda purchased at retirement begins payouts at an advanced age, such as 85, to surviving investors. the alda premium is typically a fraction of the premium for an immediate annuity with the same 121m.j. warshawsky / financial services review 26 (2017) 113–141 payout. the annuity load is nonetheless higher because alda providers face higher risk in guaranteeing the interest rate during the interval between purchase and payouts and also possibly from a greater extent of adverse selection among purchasers. determining the desired level of alda payouts and smoothly managing income before benefits begin remains practically challenging for investors. our simulations assume that investors purchase the alda at age 65, make systematic withdrawals before 85, and lower withdrawals by the amount of alda payouts at 85. the lifetime income flow can be volatile because of the uncertainty in investment outcomes. 4. results the strategies may use nominal or real (inflation-adjusted) product components in their implementations, but reported outcomes are always adjusted for stochastically realized inflations, that is, reported incomes and balances are in real dollars. to initially speed up the simulation computations, which involve searching across a grid of expected utility values to find the optimum, the fixed percentage withdrawal has a one percentage point increment, and the fixed real dollar (annual) withdrawal has a $5,000 increment. the annuitization process can take up to 30 or more years, with a 5-year time period increment, and the purchase can start and end in the range of 0–100% of expected wealth, with a five percentage point increment. later simulations do use a more refined search grid. i consider both singles and couples as retired households. 4.1. singles with normal life expectancy although the at-least-partial use of annuities is indicated by the literature, and the ace values (for like preference parameters) reported below are higher than found in strategies using systematic withdrawals alone, i start with systematic withdrawals alone to gain some insight into the best form and level of systematic withdrawals. among fixed-dollar inflationindexed withdrawals, the simulations find that a $15,000 withdrawal (that is, 6% of the initial balance, inflation-indexed in subsequent years) generates the highest lifetime utility. real consumption remains constant, as long as the accounts are not exhausted. if the distribution strategy is based on a fixed percentage of asset balances with a varying dollar amount, the optimal withdrawal is 9% of the portfolio balance each period. relative to the fixed real-dollar withdrawals, this strategy tends to generate higher consumption in earlier years but lower consumption in later years, owing to a higher volatility of incomes. overall, the ace is higher here than with the fixed dollars implementation, indicating the superiority of the fixed percentage approach. see table a-1 in the appendix, section b, for the consumption and wealth outcomes for single retirees. financial planners often advise retirees to draw down their assets using a 4% rule, as suggested by bengen (1994). by this rule, households in our model would withdraw $10,000 a year in real dollars, which is lower than the optimal withdrawal amount given above and produces lower expected lifetime utility (by an ace measure of 23.0, not reported, vs. 25.1 in table a-1). one important factor is that my analysis includes a safe and inflation-indexed 122 m.j. warshawsky / financial services review 26 (2017) 113–141 lifetime income flow from social security, which supports a higher withdrawal and reduces the risk to income of wealth running out. scott et al. (2009) argue that the 4% rule is inefficient because it leaves a significant portion of resources unspent. also, most financial advisors deal with clients with above-average wealth; their spending needs and desires may be relatively modest compared with their assets. let individuals now consider adding (laddered) purchases of immediate life annuities (table 2). in scenario 2a, with both fixed dollar withdrawals and annuity payouts indexed to inflation, singles annuitize 40% of wealth initially and 45% over 25 years.8 they also simultaneously withdraw $10,000 a year in real terms. this combination foregoes some liquidity (lower balances) but generates a higher and more sustainable income flow than systematic withdrawals alone, implying a greater lifetime utility (ace of 25.6 vs. 25.1 in table a-1). when fixed percentage withdrawals and nominal life annuities are considered in scenario 2b, it is optimal to not annuitize initially (zero percentage), but purchases of immediate annuities reach 45% of wealth by 20 years. the postponed and laddered purchase of nominal annuities helps maintain the real purchasing power of payouts and reduces the timing risk of purchases. for early years in retirement, income is achieved primarily through withdrawals (8% of balance), which are nontrivial amounts given the size of account balances. the ace is highest here across the strategies/implementations analyzed, at 26.3. with withdrawals fixed in real dollars along with a ladder of nominal annuities considered in scenario 2c, singles withdraw $10,000 a year and also generate income through significant annuity purchases—initially 30% of wealth and reaching 55% by 25 years. the ace is lower than in the above “nominal” strategy, because the rigid fixed dollar withdrawal imposes greater risk of running out of funds. overall, the strategy of fixed percentage withdrawals combined with laddered purchases of nominal immediate annuities wins.9 table 2 search for optimal strategies among systematic withdrawals and laddered purchases of immediate life annuities—singles real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed real dollars � real annuity optimal strategy: withdrawal $10, annuity initial 40% of wealth, ending 45% by 25 years, ace 25.6 income 32.6 29.5 18.6 28.0 4.4 balance 199.8 89.4 0.0 88.9 72.5 b. options: fixed percentage � nominal annuity optimal strategy: withdrawal 8%, annuity initial 0% of wealth, ending 45% by 20 years, ace 26.3 income 35.6 28.2 21.2 28.4 4.6 balance 250.0 105.3 0.0 113.2 87.9 c. options: fixed real dollars � nominal annuity optimal strategy: withdrawal $10, annuity initial 30% of wealth, ending 55% by 25 years, ace 25.8 income 33.3 30.1 18.0 28.4 4.9 balance 219.2 101.3 0.0 99.4 82.1 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 123m.j. warshawsky / financial services review 26 (2017) 113–141 in table 3, an alda is considered instead of laddered purchases of immediate annuities. in scenarios 3a and 3c, the retiree would optimally withdraw $15,000 systematically, inflation-adjusted, and use 5% of initial wealth to purchase an alda, with payouts indexed to inflation or nominal, respectively. the simulated level of demand for the alda seems trivial. recall, however, that the alda payout commences 20 years after purchase. the ages covered by an alda have higher mortality rates than earlier years, which actuarially reduces the cost for any life annuity payout. further, with discounting for time (interest) for the 20-year waiting period, the alda premium is substantially lower. discounting would lower the premium by about 65% for a real alda and 85% for a nominal alda, assuming the average interest and inflation rates reported in table 1. put differently, the age-85 payout from an alda that is purchased with 5% of wealth is equivalent to the payout from an immediate annuity that is purchased with 15–30% of initial wealth, depending on the contract terms. nonetheless, the alda has significant shortcomings. first, achieving a smooth connection between withdrawals and commencement of payout is difficult. consumption may experience cliff changes by the time the alda begins payouts because of declines in wealth. notably in scenario 3b, the alda loses its appeal entirely when a fixed percentage withdrawal is used. and second, households are likely to be better off using immediate annuities rather than an alda, as indicated by higher aces in table 2 versus table 3, in part owing to the lower loads on immediate annuities. more important, the immediate annuities perform better in managing a steady income flow and avoiding extremely low incomes. 4.2. singles with a short life expectancy i examine now how the strategies/implementations would vary for single retirees with an impaired life expectancy. i model impaired mortality as equivalent to those with a spinal cord injury, based on the estimates of strauss et al. (2005). their life expectancy is about seven table 3 search for optimal strategies among systematic withdrawals and age-85 alda—singles real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed real dollars � real alda optimal strategy: withdrawal $15, alda 5% of wealth, ace 25.2 income 28.8 28.8 15.4 27.2 4.3 balance 295.9 161.4 0.0 151.1 100.6 b. options: fixed percentage � nominal alda optimal strategy: withdrawal 9%, alda 0%, ace 25.7 income 36.7 26.7 17.6 27.2 6.4 balance 254.1 140.3 41.0 146.3 71.4 c. options: fixed real dollars � nominal alda optimal strategy: withdrawal $15, alda 5% of wealth, ace 25.2 income 28.8 28.8 15.2 27.2 4.3 balance 293.7 160.0 0.0 150.7 100.2 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 124 m.j. warshawsky / financial services review 26 (2017) 113–141 years shorter than that for the general population. as shown in table 4, these households would make more aggressive withdrawals—$5,000 more a year as a fixed dollar withdrawal strategy or three percentage points higher as a fixed percentage strategy (compared with cases 2a and 2b above, respectively). being less likely to reach advanced ages, these retirees would also generally reduce their purchases of immediate life annuities. here mortality is simulated based on the impaired population life table but assumes that annuity pricing uses the regular annuitant life table, because annuity underwriting is uncommon and expensive in the united states. 4.3. couples tables 5 and 6 report the results for two-person (same age) retired households when both annuities and systematic withdrawals are considered. the optimal strategies for couples are table 4 search for optimal strategies among systematic withdrawals and immediate annuities—singles with an impaired life expectancy real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed real dollars � real annuities optimal strategy: withdrawal $15, annuity, initial and ending, 25% of wealth, ace 27.2 income 35.7 34.0 18.7 32.6 4.6 balance 220.5 138.8 0.0 126.0 72.0 b. options: fixed percentage � nominal annuity optimal strategy: withdrawal 11%, annuity, initial 0% of wealth, ending 30% by 20 years, ace 27.9 income 41.3 32.5 23.3 32.8 6.3 balance 250.0 138.5 6.8 137.7 79.6 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. table 5 search for optimal strategies among systematic withdrawals and laddered purchases of immediate life annuities—couples real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed real dollars � real annuity optimal strategy: withdrawal $10, annuity initial 35% of wealth, ending 45% by 25 years, ace 23.5 income 37.8 33.3 17.4 30.4 6.7 balance 213.6 87.3 0.0 90.5 80.0 b. options: fixed percentage � nominal annuity optimal strategy: withdrawal 8%, annuity initial 0% of wealth, ending 50% by 25 years, ace 24.3 income 41.3 31.3 19.9 31.1 7.0 balance 250.0 88.9 0.0 102.6 87.2 c. options: fixed real dollars � nominal annuity optimal strategy: withdrawal $10, annuity initial 25% of wealth, ending 65% by 25 years, ace 23.8 income 39.3 34.2 17.2 31.0 7.4 balance 223.3 77.3 0.0 88.8 86.1 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 125m.j. warshawsky / financial services review 26 (2017) 113–141 similar to those for singles. using systematic withdrawals alone, it is optimal for couples to withdraw $15,000 in real dollars, or alternatively 9% of balances each year (table a-2 in appendix, section b). when the option of laddered purchases of immediate life annuities is considered, couples should eventually annuitize from 45 to 65% of wealth, varying with the combinations of nominal and real products (table 5). initial annuitization is generally lower for couples than for singles. also note that the ace is lower for couples than for singles because the same resources have to be shared for a couple even when there are some economies of scale in living. for retired couples, the optimal strategy is a program of systematic withdrawals of 8% of balances combined with a laddered purchase of immediate nominal life annuities over 25 years, rising from zero percentage to 50% of wealth. when an alda is considered, a small fraction (5%) of wealth on alda purchase is optimal and improves welfare relative to the systematic withdrawals alone, according to the measure of ace (table 6). overall, households benefit from adding annuities to their retirement portfolios. immediate life annuities again serve better than aldas. 4.4. other household preferences and situations using the combined withdrawal-laddered strategy i now consider alternative preferences and household situations. given the prior results indicating the overall superiority of combinations of fixed percentage withdrawals with laddered purchases of nominal life annuities, we search for optimal strategies just within that class. table 7 shows the specific optimal strategies and range of results in several cases. with greater wealth (here $600,000), there is more room for the operation of the bequest motive (the minimum consumption threshold is easily met), and therefore the withdrawal rate is lower as is the ultimate extent of annuitization (compare table 7a with table 2b). table 6 search for optimal strategies among systematic withdrawals and age-85 alda—couples real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed real dollars � real alda optimal strategy: withdrawal $15, alda 5% of wealth, ace 23.6 income 35.7 35.7 16.2 30.7 6.5 balance 296.5 142.1 0.0 139.4 105.5 b. options: fixed percentage � nominal alda optimal strategy: withdrawal 9%, alda 0%, ace 23.9 income 43.1 30.0 17.5 30.2 8.5 balance 250.3 126.8 37.0 136.8 71.6 c. options: fixed real dollars � nominal alda optimal strategy: withdrawal $15, alda 5% of wealth, ace 23.7 income 35.7 35.7 16.1 30.7 6.6 balance 296.3 142.8 0.0 139.6 104.4 source: authors’ simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 126 m.j. warshawsky / financial services review 26 (2017) 113–141 hence, the balance holdings across possible outcomes are all proportionately higher here. if risk aversion is higher (� � 6 vs. 4 in the baseline), the single-member household would boost eventual annuitization by 10 percentage points across the life cycle (compare table 7b with table 7a) owing to its greater preference for security, in particular about income; indeed income outcomes are uniformly larger and balances lower. in situation 7c, i change a number of preferences and situations simultaneously: the household is wealthier ($1,000,000), more risk averse (� � 7.6), more desirous of leaving a bequest (� � 20 vs. 10 in the baseline), more forward-looking/patient (� � 1.10), while our search grid across solutions to find the optimal strategy is more refined both for withdrawal rates, and the extent and timing of annuitization. these changes drop the fixed percentage withdrawal rate substantially, while the annuitization pattern is largely unchanged. wealth outcomes are uniformly and relatively higher here because the optimal strategy consistent with the preferences and situation gives more emphasis to retaining balances. in the final frame of table 7, i move in the opposite direction—preferences remain the same, but wealth is lower and retirement for the couple is at age 62. the optimal strategy places a higher emphasis on income, as evident in the much greater extent of annuitization, albeit the purchase laddering takes place over a longer horizon because the younger couple has a much higher probability of at least one member surviving to old age. the longer horizon for the household also explains the lower rate of withdrawals. the specific optimal strategies vary widely with household preferences and demographic situations. table 7 search for optimal strategies among fixed percentage systematic withdrawals and laddered purchases of nominal immediate life annuities—alternative situations and preferences real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. situation: initial wealth is $600,000 optimal strategy: withdrawal 7%, annuity initial 0% of wealth, ending 20% by 20 years, ace 41.49 income 61.92 45.39 30.04 45.74 balance 638.12 359.16 75.66 363.56 b. situation: as in a. above, but with higher risk aversion � � 6 optimal strategy: withdrawal 7%, annuity initial 10% of wealth, ending 30% by 20 years, ace 43.40 income 62.78 46.14 30.79 46.51 balance 579.88 319.14 57.62 322.98 c. situation: initial wealth is $1,000,000, high risk aversion � � 7.6 stronger bequest motive � � 20 more forward-looking, patient � � 1.10, refined search grids optimal strategy: withdrawal 4%, annuity initial 7.5% of wealth, ending 20% by 21 years, ace 49.1 income 79.6 55.9 39.5 56.9 12.5 balance 1314.7 806.5 348.2 806.4 301.4 d. situation: as in a. above, except couple both age 62, initial wealth is $150,000, and � � 1.0 optimal strategy: withdrawal 6%, annuity initial 20% of wealth, ending 75% by 30 years, ace 19.76 income 31.69 26.56 17.36 25.31 4.83 balance 127.57 48.77 0.0 54.32 45.98 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 127m.j. warshawsky / financial services review 26 (2017) 113–141 4.5. add consideration of household taxes i now add consideration of personal taxes on the retired household to the model. in particular, i add parameters for the payment of taxes (aggregating federal, state and local government impositions) on income and bequests made at the average effective rates of the household. for most households, estate tax rates are low or zero, but income tax rates, even for middle-class retirees, can be 15% or higher. i also add to the model the minimum distribution requirements on tax-deferred account balances which can force distributions in excess of optimal levels determined, particularly at older ages; for example, beyond the ages of 90 and older, distributions of 10% and more are required. if the legally required distribution in a year is higher than the optimal, then we place the distribution in excess of the optimal in a taxable investment account that will eventually be used for liquidity or bequest purposes, that is, to support consumption if all other retirement wealth has been used up, but otherwise held in reserve. i illustrate in table 8 the impact of taxes on optimal strategies for four cases shown above, in panels 2b, 5b, 7b, and 7c. i assume a 15% effective income tax rate and a zero effective table 8 search for optimal strategies among fixed percentage systematic withdrawals and laddered purchases of nominal immediate life annuities—alternative situations and preferences, with additional consideration of taxes and minimum distribution requirements real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. situation: panel 2b optimal strategy: withdrawal 9%, annuity initial 0% of wealth, ending 55% by 20 years, ace 23.2 after-tax income 31.5 24.8 17.4 24.4 4.6 tax-deferred balance 250.0 80.9 0.0 97.5 88.9 taxable balance at age 95 0.5 0.0 0.0 0.1 0.5 b. situation: panel 5b optimal strategy: withdrawal 8%, annuity initial 0% of wealth, ending 60% by 25 years, ace 21.03 after-tax income 35.2 26.6 16.7 26.5 6.0 tax-deferred balance 250.0 77.8 0.0 95.8 88.2 taxable balance at age 95 0.1 0.0 0.0 0.1 0.9 c. situation: panel 7b optimal strategy: withdrawal 7%, annuity initial 15% of wealth, ending 40% by 20 years, ace 38.3 after-tax income 54.2 40.1 26.9 40.4 8.8 tax-deferred balance 545.9 285.9 25.2 290.1 172.8 taxable balance at age 95 40.7 9.9 0.0 13.8 14.0 d. situation: panel 7c optimal strategy: withdrawal 5%, annuity initial 5% of wealth, ending 20% by 30 years, ace 44.30 after-tax income 70.3 50.6 33.7 50.8 11.4 tax-deferred balance 1195.5 746.6 250.7 734.5 296.8 taxable balance at age 95 331.8 138.0 42.6 154.9 94.5 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 128 m.j. warshawsky / financial services review 26 (2017) 113–141 estate tax rate. in general, because a significant portion of retirement resources are sent to the government, to produce income levels (and utility) roughly comparable (but inevitably a bit lower) to the situation before the consideration of taxes, higher withdrawals and greater annuity purchases must be made. the remaining balances will be lower. in addition, with the addition of minimum distribution requirements, for higher wealth levels or favorable investment outcomes, some tax-favored retirement assets will eventually be placed in taxable investment accounts, particularly at older ages. 4.6. add pensions finally, i add defined benefit pensions to the model. the pension benefits can be either indexed to inflation or not, and can be either for the individual alone or for both members of the couple (assumed as a joint-and-two-thirds-to-survivor annuity). here i also only use a refined search grid to find the optimal strategy and limit the laddering of annuities up to age 90. with the addition of pension income but no subtraction of other retirement resources, we naturally will expect a higher ace, everything else equal. because the pension is paid as a life annuity, one would also expect less need for purchasing immediate life annuities as well as lower withdrawals. these are indeed the model outcomes shown below in table 9 when i add pensions (at the same level as social security, but unindexed and split among both members of the couple), on top of taxation and minimum distribution requirements, to cases 7c (8d) and 7d. 4.7. the full model treatment of three disparate example cases here i show three example cases, with divergent preferences and household demographic and economic situations, their optimal strategies determined by the full model, and the range table 9 search for optimal strategies among fixed percentage systematic withdrawals and laddered purchases of nominal immediate life annuities—alternative situations and preferences, with additional consideration of taxes and minimum distribution requirements, and defined benefit pensions added real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. situation: panel 7c (8d) optimal strategy: withdrawal 4.5%, annuity initial 0% of wealth, ending 15% by 24 years, ace 49.0 after-tax income 76.4 56.8 39.1 56.8 11.5 tax-deferred balance 1301.8 813.4 276.5 798.5 316.9 taxable balance at age 95 527.1 202.2 67.7 238.6 149.9 b. situation: panel 7d optimal strategy: withdrawal 4.0%, annuity initial 0% of wealth, ending 80% by 27 years, ace 21.9 after-tax income 35.3 29.9 18.9 28.4 5.6 tax-deferred balance 167.3 73.6 0.0 75.7 60.2 taxable balance at age 95 31.6 1.3 0.0 4.6 9.6 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 129m.j. warshawsky / financial services review 26 (2017) 113–141 of possible outcomes over the households’ retirement lifetimes. table 10 gives the full specifications of the cases in terms of preference parameters, demographic situation, wealth, and so on for each household. now i show in table 11 the optimal strategies and range of outcomes for the three households. the first household has fairly moderate retirement means, but a large bequest motive, and is somewhat risk averse. their solution is a quite modest withdrawal rate, and significant annuitization over time, which produces fairly steady income and, relative to their means, significant asset holdings, both in the tax-deferred and taxable table 10 description of three different retired households household age(s) wealth equity allocation investment expense tax rate health status social security pension a 68, 63 $400k 70% 77bps 20% good $18k $12k both b 70 $1.5m 75% 47bps 25% poor $18k $12k c 66, 64 $250k 40% 33bps 10% good $15k $6k both, cola household � � � a 1.075 30 7.6 b 1.075 30 5.1 c 1.15 3 10.4 source: author. table 11 search for optimal strategies among fixed percentage systematic withdrawals and laddered purchases of nominal immediate life annuities—three households, full model real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. household: 10a optimal strategy: withdrawal 4%, annuity initial 0% of wealth, ending 35% by 21 years, ace 27.6 after-tax income 48.3 38.8 26.1 37.9 7.2 tax-deferred balance 505.5 257.1 20.4 258.2 157.7 taxable balance at age 95 211.8 64.3 12.1 103.1 94.0 b. household: 10b optimal strategy: withdrawal 6%, no annuity, ace 43.4 after-tax income 107.4 82.7 49.7 80.3 17.9 tax-deferred balance 1927.5 1354.7 658.3 1313.8 393.8 taxable balance at age 95 250.0 124.5 na 154.2 67.6 c. household: 10c optimal strategy: withdrawal 4%, annuity initial 0% of wealth, ending 80% by 24 years, ace 27.4 after-tax income 41.8 34.6 24.8 33.7 5.3 tax-deferred balance 265.5 119.7 0.0 122.5 96.0 taxable balance at age 95 43.8 4.2 0.1 10.9 19.7 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 130 m.j. warshawsky / financial services review 26 (2017) 113–141 accounts. the second household is composed of an older sick person with significant asset holdings. because of his large bequest motive and his poor health, no annuitization is optimal while the moderate withdrawal rate, combined with social security and pensions, gives plenty of income and preserves most of the tax-deferred and taxable assets to grow and to leave as an inheritance. the third household has even more modest retirement means than household a, but is more risk-averse, more desirous of future than current spending, and much less interested in leaving a bequest. therefore, the optimal strategy for household c places a much higher allocation to annuitization, laddered over an extended period, which in turn produces a less volatile income flow over the household’s retirement lifetime. figs. 1 through 4 show some of the outcomes graphically for household c, using the optimal strategy indicated above. fig. 1 shows the mean of stochastic outcomes under the optimal strategy in terms of pretax income, by source and by age. social security and, in this case, pensions are inflationindexed, so the income flow from them is steady, although they do decline, on average, with age, because of the possibility that one member of the couple dies. income from the series of nominal life annuities being purchased over time grows through age 90, reaching about $12,000 annual income, in real terms, but then declines because purchases of the fig. 1. sources of total annual income, mean, by age ($ real). source: author’s simulations, for household c in table 11. 131m.j. warshawsky / financial services review 26 (2017) 113–141 immediate life annuities stop and owing to the force of inflation. even though this household prefers future spending over current spending, which keeps optimal income high and steady over time, eventually income will decline with age because the large force of mortality makes current spending more salient, and, for a couple, one member is likely to have passed away and spending for him/her inevitably drops. withdrawals from the retirement investment portfolio, although fixed in percentage terms, decline and eventually disappear as the retirement investments are withdrawn and also transferred, because of the minimum distribution requirements, to a taxable portfolio, and life annuities are purchased. minimum distributions (shown in fig. 2) will be made and transferred to a taxable portfolio (see fig. 3) only if the withdrawal rate and annuity purchases are less than the minimum legally required (which initially is less than 4% at age 71 but increases with age), and, obviously, if there is value left in the retirement investment portfolio, which in turn depends on return performance. here minimum distributions are made only in the upper percentiles and at upper ages. finally, fig. 4 illustrates the formula for asset allocation given in footnote 5 above, as applied to the optimal strategy for this particular household. here, because annuitization is quite full, so is the allocation to equity in the remaining investment portfolios (both tax-deferred and taxable). fig. 2. stochastic range of required minimum distributions, by age ($ real). source: author’s simulations, for household c in table 11. 132 m.j. warshawsky / financial services review 26 (2017) 113–141 4.8. different views on investment returns and interest and inflation rates thus far, i have used a model (from campbell and viceira, 2005) of investment returns and interest and inflation rates estimated on quarterly data from 1962 through 2011. this is a reasonable approach for, essentially, forecasting the range of possible experience in the future. nevertheless, some might want to put a greater emphasis on current conditions (in 2013), particularly as interest and inflation rates have declined so dramatically recently, as presumably have overall expected investment returns. it is possible to have the simulations adjusted to reflect this presumed different environment, without, however, completely discarding the longer view, based on more distant past experience. going a step further, it is even possible to fix interest and inflation rates for the current year, allowing a smaller dispersion in the second year, slowly spreading thereafter; equity returns would continue to be as random as modeled earlier. this latter approach may indeed be the best way to express uncertainty for a household currently retired and planning to implement the produced strategy immediately. for households that are not yet retired, however, the former approach (whether emphasizing current conditions or completely historical), reflecting a wider range of uncertainty, is more appropriate. in table 12 below, i give the optimal strategies and range of outcomes for cases 10a and 10c. the first simulation, “current conditions,” gives lower expected interest and inflation rates and overall investment returns than the completely historical approach; indeed the real fig. 3. stochastic range of after-tax investment portfolio values, by age ($ real). source: author’s simulations, for household c in table 11. 133m.j. warshawsky / financial services review 26 (2017) 113–141 interest rate is reduced more than 100 basis points here. this may reduce the attractiveness of annuitization somewhat, as it certainly will reduce the ace. the second simulation, “current conditions with fixed initial interest rates,” uses, somewhat arbitrarily, the average interest rate from the simulation as the fixed initial rate. it should be noted that a fixed initial interest rate reduces considerably the uncertainty around purchasing a life annuity in the first couple of years of the plan horizon; this should have the effect of increasing the attractiveness of initial annuitization, compared with either historical approach. the expected changes occurred. of further note, given the poorer investment environment, the model favors income over assets, everything else being equal. 5. policy discussion and conclusions this analysis is based on a model of rational decision-making by retired individuals and couples, using currently available investment and insurance products. yet people are not likely to use such strategies without a more formalized structure from which to obtain them. at the least, plan sponsors, plan record keepers, or financial providers or advisors must make them available to participants. fig. 4. optimal equity allocations of investment portfolios, by age, percent. source: author’s simulations, for household c in table 11. 134 m.j. warshawsky / financial services review 26 (2017) 113–141 to encourage the use of life annuities, as strongly recommended by theoretical and simulation models, efforts must go beyond designing the best strategy. as brown et al. (2008) show, it is critical to frame the issue so that plan participants understand that their account balances—although seemingly a large amount of money in an investment frame— may seem less adequate in a consumption frame as a lifetime annual income flow. indeed, based on a survey of people ages 50 and older, brown et al. find that when the alternatives are presented in a consumption rather than investment frame, the vast majority prefer an annuity. indeed james, martinez and iglesias (2006) make a related convincing argument that the high annuitization rate (more than two-thirds) in the chilean individual account retirement system is because of its regulatory structure. they ascribe the high annuitization rate to a limited range of payout options (effectively only life annuities or systematic withdrawals), and to the absence of a public defined benefit plan except for a minimum pension guarantee. given these results, proposed regulations by the obama administration, also reflected in proposed bipartisan legislation, could have significantly changed retiree behavior. the department of labor proposed guidelines in 2013 to require defined contribution plan table 12 search for optimal strategies among fixed percentage systematic withdrawals and laddered purchases of nominal immediate life annuities—two households, full model, alternative investment views real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. household: 10a, current conditions optimal strategy: withdrawal 4.5%, annuity initial 0% of wealth, ending 35% by 21 years, ace 26.0 after-tax income 47.8 39.9 25.1 38.9 6.8 tax-deferred balance 428.2 193.3 1.6 205.4 143.1 taxable balance at age 95 90.0 23.1 3.7 32.1 30.4 b. household: 10a, current conditions with fixed initial interest rate optimal strategy: withdrawal 3%, annuity initial 21.5% of wealth, ending 52.5% by 21 years, ace 26.0 after-tax income 45.6 37.2 24.5 35.9 7.2 tax-deferred balance 380.0 151.7 0 168.0 130.0 taxable balance at age 95 205.7 54.3 12.3 74.2 66.8 c. household: 10c, current conditions optimal strategy: withdrawal 4%, annuity initial 0% of wealth, ending 67.5% by 24 years, ace 25.2 after-tax income 37.7 32.5 22.6 31.0 5.0 tax-deferred balance 250.0 84.1 0.0 100.6 90.4 taxable balance at age 95 20.7 3.3 0.2 5.7 8.2 d. household: 10c, current conditions with fixed initial interest rate optimal strategy: withdrawal 3.5%, annuity initial 31.5% of wealth, ending 80% by 24 years, ace 25.4 after-tax income 38.0 33.9 22.9 31.6 5.4 tax-deferred balance 175.9 64.4 0.0 73.5 64.2 taxable balance at age 95 34.3 6.7 1.4 10.9 13.1 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 135m.j. warshawsky / financial services review 26 (2017) 113–141 sponsors to give participants annual income illustrations. these are attempts to achieve more realistic framing and to provide some incentives for partial annuitization. given strongly ingrained behavior and market conditions in the u.s. leaning in the opposite direction, the government might have to take an active role to even the playing field with pure asset strategies. certainly the private sector, especially including life insurers, financial companies, and plan sponsors, need to be more aggressive and creative in their product design and marketing activities in the retirement field, to emphasize the need for lifetime retirement income. based on surveys of hypothetical choices, beshears et al. (2014) find that allowing individuals to annuitize a fraction of their wealth increases annuitization relative to an “all or nothing” decision. the empirical simulation analysis here indicates that some life annuities should indeed be part of the portfolios for many retirees. life annuities work to establish minimum necessary consumption and a certain level of hedging against longevity risk. while this insurance is being lost with the decline of defined benefit pension plans, it can be restored. some have proposed advanced life delayed annuities as providing the essence of insurance at lower premiums. maintaining a sustainable income flow before alda payments begin, however, is no easy task, and the alda has higher loads. welfare measures using alda strategies are lower. a cheaper, less risky and more transparent strategy is the combination of systematic withdrawals (fixed percentage) with laddered purchases of nominal immediate life annuities. this strategy is supported by the analytical work shown here and in the literature, and is robust across the spectrum of household preferences and situations. notes 1 ameriks, veres, and warshawsky (2001) also use a shortfall framework, but with historical, not stochastic, simulated data. 2 these are unpublished results, which have been kindly provided by dmm, based on united states financial and mortality data over the 1967–2002 period for a 65-year-old single man. 3 the obama administration enacted a special dispensation, effective in 2014, from the retirement account minimum distribution requirements for longevity insurance with deferred payment up to age 85; the dispensation removes the premium for longevity insurance from the calculation base used in determining required distributions subject to income tax; the delayed payments, however, are taxed when paid. reportedly, some individuals in high tax brackets are using this dispensation to reduce their tax bill rather than to manage risks. 4 to be consistent with our functional form, we have rearranged the specifications of de nardi et al. (2010) and lockwood (2012) and recalculated the parameter values. ameriks et al. (2011) find their benchmark estimates of � � 7.28 and � � 47.6. these parameters would lead to little consumption and substantial bequest for households in 136 m.j. warshawsky / financial services review 26 (2017) 113–141 our model because the consumption threshold $7,280 is easily exceeded ($13,800 from social security alone for a single retiree, see assumptions below) and the bequest motive is particularly strong. nevertheless, we do consider the effect of larger bequest motive parameters than denardi’s but maintain the consumption threshold, so that lower-income retired households will likely depend more on life annuities, ceteris paribus. 5 let e denote the equity share in the remaining non-annuitized wealth and a denote the degree of annuitization (i.e., a � aw/w, annuitized wealth divided by total wealth). for the desired 50–50 risk exposure, e*(w-aw)/w � 0.5. rearranging the equations gives e � 0.5/(1-a). 6 using some six months of pricing data from august 2011 through april 2012 provided to me by a large, highly rated, united states insurance company and our own computations of fair value using government bond yields and general population mortality, we calculated that the load differential between a nominal alda commencing payment at age 85 and a single-premium immediate straight fixedpayout life annuity, both issued at age 65, indeed averaged about five percentage points. 7 there is some rigidity in these withdrawal options. the fixed dollar strategy does not respond to realizations of asset returns and the fixed percentage option does not speed up distribution toward the end of life, as might be desired to avoid leaving too large a bequest. rather, they are easy to implement in old age, easy to explain, and require minimal governance from product issuers. households could re-optimize their portfolios any time, perhaps with the help of financial advisors, but to avoid a lot of extra costs and governance requirements, the rerunning of the algorithm should be done sparingly, perhaps only at major life events such as the death of a spouse. sun and webb (2012) show that households could spend according to the irs table for required minimum distributions (rmd), plus interest and dividends, and get better utility than from alternatives in the class of systematic withdrawals. collins and lam (2011) use a case study approach to discuss how financial advisors can utilize a credible simulation model to help investors make informed retirement planning decisions. 8 the degree of annuitization at any age is expressed as a percentage of the accumulated value of initial wealth, which reflects the increased with interest (at the average rate). one could have alternatively expressed it as a percentage of the account balance, but this would have been misleading. an extreme example: 100% annuitization of the remaining $1 balance appears as a strong preference for the life annuity but may be trivial if $99 has been withdrawn over prior years. 9 while the strategy of fixed dollar withdrawals and a real annuity gives completely steady real income—a goal perhaps desired by the highly risk averse, the combination of fixed percentage and a real annuity is not examined here because the nominal annuity has a lower load than the real annuity and the increase in income to cover inflation can be accommodated through the laddering purchase of smaller amounts of immediate annuities over time. 137m.j. warshawsky / financial services review 26 (2017) 113–141 appendix appendix a: simulations of rates and returns asset returns are simulated as a vector autoregressive process (var). the var coefficients and variance matrix are first empirically estimated and then embedded in the simulations with stochastic shocks. technically, let v be a vector containing the variables. the vector evolves in an autoregressive pattern: vt � �0 � � k�1 k �kvt�k � �t where � �(0, �) denotes a vector of serially uncorrelated normal errors with e�t�s � 0, t � s. the contemporaneous correlations of shocks are incorporated via the variance-covariance matrix � and serial correlations of the variables via the coefficients �. the econometric regression on historical data yields estimates of the coefficients, �̂�s and the variancecovariance matrix �̂. the simulations follow several steps: first, a cholesky factorization table a-1 search for optimal strategies among systematic withdrawals—singles real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed dollars, inflation adjusted optimal strategy: withdrawal $15, ace 25.1 income 28.8 28.8 13.8 27.0 4.8 balance 315.5 174.2 0.0 163.0 106.3 b. options: fixed percentage of balance, nominal optimal strategy: withdrawal 9%, ace 25.7 income 36.7 26.7 17.6 27.2 6.4 balance 254.1 140.3 41.0 146.3 71.4 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. table a-2 search for optimal strategies among systematic withdrawals—couples real $000 95th percentile 50th percentile 5th percentile mean standard deviation a. options: fixed dollars, inflation adjusted optimal strategy: withdrawal $15, ace 23.2 income 35.7 35.7 13.8 30.4 7.2 balance 314.0 158.3 0.0 151.7 109.4 b. options: fixed percentage of balance, nominal optimal strategy: withdrawal 9%, ace 23.9 income 43.1 30.0 17.5 30.2 8.5 balance 250.3 126.8 37.0 136.8 71.6 source: author’s simulations. higher aces (average certainty equivalent consumption) indicate better outcomes for households, for a given set of preference parameters. 138 m.j. warshawsky / financial services review 26 (2017) 113–141 decomposes the variance-covariance matrix to a triangle matrix. that is, the factorization finds a triangle matrix w so that w�w � �̂. second, a vector of random values are generated according to iid n(0, 1). multiplying this vector by the cholesky factor matrix generates correlated shocks to rates and returns. third, multiplying the var coefficients by previous period returns, plus the shocks, gives current period returns. the procedure is repeated forward until the end-of-time horizon under consideration. following the specification in campbell and viceira (2005), asset classes include money market (90-day t-bills), stocks (proxied by the s&p 500 total return index), and bonds (proxied by the 10-year u.s. government bond total return index). these rates and returns in the var estimation are expressed in logarithm real terms (after adjusting for inflation measured by the change in the cpi-u index), using quarterly data. additionally, three forecasting variables (state variables), which help form expectations of future rates and returns, include short-term nominal interest rate (nominal t-bills), equity dividend yield, and the slope of the yield curve (yield spread as the difference between u.s. 10-year t-note zero-coupon yield and the yield on 90-day t-bills). the entire system is estimated on 1962–2011 quarterly data. bankruptcy of insurance companies may occur. it is assumed that an insurance provider fails with a probability of 0.15% per annum (uniform distribution), based on moody’s global analysis of default probability for corporate bonds rated a for 1970–2005. the size of the loss of insurance contract value is simulated, within the empirical range of economic contractions estimated by barro (2006). appendix b: results for additional scenarios table a-1 shows the optimal results for single retirees considering systematic withdrawals alone. for the optimal fixed-dollar inflation-indexed withdrawals, the median level of real consumption is $28,800, including $13,800 from social security. the median real balance is $174,200 among survivors. table a-2 shows the optimal results for married retirees who are considering systematic withdrawals alone, either fixed real dollars or a fixed percentage of balances. acknowledgments i thank professors jing ai, raimond maurer, and anthony webb for helpful comments and access to unpublished empirical findings, the three anonymous reviewers for this journal, and, especially, gaobo pang, for his considerable assistance and advice on this article. references ameriks, j., caplin, a., laufer, s., & van nieuwerburgh, s. (2011). the joy of giving or assisted living? using strategic surveys to separate public care aversion from bequest motives. journal of finance, 66, 519–561. 139m.j. warshawsky / financial services review 26 (2017) 113–141 ameriks, j., veres, r., & warshawsky, m. j. (2001). making retirement income last a lifetime. journal of financial planning, 14, 60–76. barro, r. (2006). rare disasters and asset markets in the twentieth century. quarterly journal of economics, 121, 823–866. bengen, w. p. (1994). determining withdrawal rates using historical data. journal of financial planning, 7, 171–180. beshears, j., choi, j. j., laibson, d., madrian, b. c., & zeldes, s. p. (2014). what makes annuitization more appealing? journal of public economics, 116, 2–16. brown, j. r., kling, j., mullainathan, s., & wrobel, m. (2008). why don’t people insure late life consumption? a framing explanation of the under-annuitization puzzle. american economic review, 98, 304–309. brown, j. r., & warshawsky, m. j. (2004). longevity-insured retirement distributions from pension plans: market and regulatory issues. in g. william, j. shoven, & m. j. warshawsky (ed.), private pensions and public policies (pp. 332–381). washington, dc: brookings institution press. campbell, j., & viceira, l. (2005). the term structure of the risk-return trade-off. financial analysts journal, 61, 34–44. collins, p., & lam, h. d. (2011). asset allocation, human capital, and the demand to hold life insurance in retirement. financial services review, 20, 303–325. de nardi, m., french, e. b., & jones, j. b. (2010). why do the elderly save? the role of medical expenses. journal of political economy, 118, 37–75. dus, i., maurer, r., & mitchell, o. s. (2005). betting on death and capital markets in retirement: a shortfall risk analysis of life annuities versus phased withdrawal plans. financial services review, 14, 169–196. dushi, i., & webb, a. (2004). household annuitization decisions: simulations and empirical analyses. journal of pension economics and finance, 3, 109–143. finkelstein, a., & poterba, j. (2004). adverse selection in insurance markets: policyholder evidence from the u.k. annuity market. journal of political economy, 112, 183–208. gong, g., & webb, a. (2010). evaluating the advanced life deferred annuity: an annuity people might actually buy. insurance: mathematics and economics, 46, 210–221. haensly, p. j., & pai, k. p. (2015). a new strategy to guarantee retirement income using tips and longevity insurance: a second look. financial services research, 24, 359–386. horneff, w., maurer, r., mitchell, o.s., & dus, i. (2008). following the rules: integrating asset allocation and annuitization in retirement portfolios. insurance: mathematics and economics, 42, 396–408. inkmann, j., michaelides, a., & lopes, p. (2011). how deep is the annuity market participation puzzle? review of financial studies, 24, 279–319. james, e., martinez, g., & iglesias, a. (2006). the payout stage in chile: who annuitizes and why? journal of pension economics and finance, 5, 121–154. kimball, m., sahm, c., & shapiro, m. (2008). imputing risk tolerance from survey responses. journal of the american statistical association, 103, 1028–1038. lockwood, l. m. (2012). incidental bequests: bequest motives and the choice to self-insure late-life risks. working paper, department of economics, northwestern university. milevsky, m. a. (2005). real longevity insurance with a deductible: introduction to advanced-life delayed annuities (alda). north american actuarial journal, 9, 109–122. mitchell, o. s., poterba, j., warshawsky, m. j., & brown, j. r. (1999). new evidence on the money’s worth of individual annuities. american economic review, 89, 1299–1318. pang, g., & warshawsky, m. j. (2009). comparing strategies for retirement wealth management: mutual funds and annuities. journal of financial planning, 22, 36–47. pang, g., & warshawsky, m. j. (2010). optimizing the equity-bond-annuity portfolio in retirement: the impact of uncertain health expenses. insurance: mathematics and economics, 46, 198–209. scott, j. s. (2008). the longevity annuity: an annuity for everyone. financial analysts journal, 64, 40–48. scott, j. s., sharpe, w., & watson, j. g. (2009). the 4% rule—at what price? journal of investment management, 7, 31–48. 140 m.j. warshawsky / financial services review 26 (2017) 113–141 sexauer, s. c., peskin, m. w., & cassidy, d. (2012). making retirement income last a lifetime. financial analysts journal, 68, 74–84. strauss, d. j., vachon, p. j., & shavelle, r. m. (2005). estimation of future mortality rates and life expectancy in chronic medical conditions. journal of insurance medicine, 37, 20–34. sun, w., & webb, a. (2012). should households base asset decumulation strategies on required minimum distribution tables? working paper, center for retirement research, boston college. warshawsky, m. j. (2012). retirement income: risks and strategies, cambridge, ma: mit press. warshawsky, m. j. (2016). new approaches to retirement income: an evaluation of combination laddered strategies. journal of financial planning, 29, 52–61. 141m.j. warshawsky / financial services review 26 (2017) 113–141 do u.s. households perceive their retirement preparedness realistically? kyoung tae kima,*, sherman d. hannab adepartment of consumer sciences, university of alabama, tuscaloosa, al 35487, usa bdepartment of human sciences, ohio state university, columbus, oh 43210, usa abstract this study examines the divergence between objective and subjective assessment of retirement adequacy, analyzing u.s. households with a full-time worker age 35 to 60 in the 2010 survey of consumer finances. of those households, 58% have objective inadequacy, and 54% have subjective inadequacy, but only 52% have objective/subjective consistency. our focus is on households with objective inadequacy, and what factors were related to being an optimist despite having objective retirement inadequacy. a logistic regression shows that households with defined benefit plans and with defined contribution plans are less realistic than those without plans, and as age increases, realism decreases. © 2015 academy of financial services. all rights reserved. jel classification: d12; d14; d91; j26 keywords: retirement assessment; retirement adequacy; cognitive ability; survey of consumer finances (scf); financial education 1. introduction retirement adequacy of current workers is an important issue in policy debates about social security reform, and in proposals for the restructuring of public and private defined benefit plans, as well as for income tax incentives for retirement savings, and penalties for early withdrawal of funds from tax-sheltered retirement accounts. however, retirement planning is becoming increasingly challenging because workers face economic uncertainty, * corresponding author. tel.: �-205-348-9167; fax: �1205-348-8721. e-mail address: ktkim@ches.ua.edu (k.t. kim) financial services review 24 (2015) 139–155 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. social security insolvency, and increased life expectancy. munnell (2012) notes that the substantial decline in the wealth-to-income ratio in the 2010 survey of consumer finances (scf) is a signal of even more serious economic problems for future retirees. bricker, bucks, kennickell, mach, and moore (2011) report that over 60% of u.s. households had decreases in their wealth over the two-year period, 2007–2009. according to the 2014 old-age and survivors insurance and federal disability insurance trustees report (social security administration, 2014), if no changes in taxes or benefits are implemented, the combined trust fund would be depleted by 2033, and income would be sufficient in the combined fund to pay only 77% of scheduled benefits. choi, laibson, and madrian (2004) note that defined benefit pension plans have been steadily being replaced with defined-contribution pension plans so that workers are more responsible for their own retirement savings. therefore, it is important that workers have accurate assessments of their financial status in the retirement planning process. comparing objective assessments of projected retirement adequacy to individual assessments of future retirement adequacy would provide insights into potential problems. analysis of factors related to discrepancies between objective and subjective assessments could provide a better focus for financial education. the main purpose of this study is to assess the consistency between objective and subjective projected retirement adequacy. to assess the objective retirement adequacy, we calculate the mean income replacement rate by following the retirement income stage method (chen, 2007; kim, hanna, and chen, 2014). compared to the benchmark ratios for different income categories estimated from the 2010 consumer expenditure survey, the adequacy of retirement resources is determined. the scf variable of the respondent’s perception of the adequacy of retirement income is used as a subjective measure of having an adequate retirement. based on objective/subjective consistency, we identify four groups; unrealistic optimists, pessimists, adequate realists, and inadequate realists (table 1). we focus our analysis in this article on households that are projected to have objective inadequacy (i.e., unrealistic optimists & inadequate realists) for the purpose of public policy interests. by analyzing households aged 35 to 60 with a full-time worker in the 2010 survey table 1 perception vs. objective retirement adequacy of u.s households with head employed full-time and age 35 to 60, 2010 scf objective measurementb total adequate retirement inadequate retirement subjective measurementa perceived adequate retirement 19.6% (adequate realists) 26.1% (unrealistic optimists) 45.7% perceived inadequate retirement 22.3% (pessimists) 32.0% (inadequate realists) 54.3% total 41.9% 58.1% 100% a for the purpose of this study, we recoded the perceived retirement adequacy variable as a binary category. when the response is coded 1 or 2 (totally inadequate and inadequate), it is defined as having a perception of an inadequate retirement. if the response has 3, 4, or 5, perceived adequacy coded as having a perception of an adequate retirement. b see the method section. 140 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 of consumer finances (scf) dataset, we analyze factors related to being an unrealistic optimist among households with objective inadequacy. 2. literature review 2.1. behavioral versus the life cycle model the dominant theory for analyzing retirement saving behaviors is the life cycle saving (lcs) model (modigliani and brumberg, 1954). the lcs model (ando and modigliani, 1963) assumes that attempt to smooth consumption, and therefore, typically savings will be related to an individual’s stage in the life cycle. those who use the lcs model to explain household behavior are assuming that households can make rational decisions and can project future patterns of non-investment income and life expectancy. furthermore, the lcs model has the assumption that consumption smoothing can be achieved by borrowing when earnings are low and saving for wealth accumulation when earnings are high, and dissaving in retirement (browning and crossley, 2001). in the traditional approach, households are assumed to be fully informed about their life-cycle wealth and when they will retire. on the other hand, behavioral economists are interested in how people make decisions in the face of incomplete information, limited cognitive resources, and decision biases (knoll, 2010). the behavioral life-cycle hypothesis integrates a traditional lcs model with a psychological model including saving motives (shefrin and thaler, 1988), the buffer stock saving model (carroll, 1997), and the hyperbolic consumption model (angeletos, laibson, repetto, tobacman, and weinberg, 2001; laibson, 1997). because this study focuses on the divergence between a household’s objective and subjective assessment of retirement resources and its adequacy, we review some empirical studies that have been discussed to analyze the effect of cognitive ability on financial behavior related to households’ retirement saving (or planning). 2.2. cognitive ability and economic decisions previous studies have focused on the relationship between cognitive ability and economic decisions. frederick (2005) and dohmen, falk, huffman, and sunde (2010) discuss cognitive ability and its relation with two important decision-making characteristics: time preference and risk preference. frederick (2005) uses the cognitive reflection test (crt) score as a proxy of the individual’s cognitive ability, and finds that among 3,428 respondents at various universities, the higher crt group was more patient and willing to take more risks than those who scored lower. similarly, dohmen et al., (2010) examines whether an individual’s cognitive ability is related to the key traits of time preference and risk aversion. from the german socio-economic panel (soep) dataset, individuals with a lower cognitive ability are significantly more impatient and more risk averse. moreover, recent studies find that cognitive ability plays an important role in investment portfolio choices and stock ownership (christelis, jappelli, and padula, 2010; grinblatt, keloharju, and linnainmaa, 141k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 2011), wealth accumulation (mcardle, smith, and willis, 2011), and avoiding financial mistakes (agarwal and mazumder, 2013). 2.3. self-assessment and perception of retirement the role of cognitive ability on subjective assessment of retirement preparedness has received little attention in previous retirement research, even though it may affect retirement decisions, including how individuals behave when deciding if, how, and when to save for retirement. with the replacement of traditional defined benefit (db) pension plans with defined contribution (dc) plans, an accurate assessment of retirement resources is becoming more vital in retirement planning. gustman, steinmeier, and tabatabai (2007) report that more than a third of health and retirement study (hrs) respondents cannot identify whether their pension plan is a db or a dc plan. similarly, chan and stevens (2008) find that many respondents in the hrs with db or dc pension plan do not know key components of their pension plan, such as normal or early age for db plans and annual contribution amount for dc plans, and there is substantial discrepancy of pension details between self-reported and employer reported data. lusardi and mitchell (2011) conclude that lack of financial literacy and financial sophistication are critical in retirement planning, and fornero and monticone (2011) find that more financially knowledgeable people are more likely to participate in retirement plans. cummings, finke, and james (2011) report that respondents with higher cognitive ability are more likely to own a roth ira and adopt it early. in addition to assessing how workers are currently preparing for retirement, it is also essential to investigate how they perceive their retirement preparedness. glamser (1976), and kilty and behling (1985) find a positive link between perceptions of retirement and actual retirement planning. joo and grable (2001) report that workers with positive and proactive attitudes toward retirement are more likely to use retirement planning advice from a financial professional. however, relatively little research on the impact of retirement perception on retirement adequacy has been conducted. malroutu and xiao (1995) examine pre-retirees’ (age 65 or younger) perception of having adequate retirement income by using the 1989 scf. about 39% of pre-retirees answer that they would have adequate retirement income. younger respondents, females, whites, households with relatively low incomes and selfemployed are less likely to perceive having adequate retirement income. munnell, golub-sass, soto, and webb (2008) investigate the accuracy of self-assessment about retirement security. the national retirement risk index (nrri) provides an objective measurement of retirement security, which is the percentage of households are at risk of being unable to maintain their standard of living in retirement. in the 2004 survey of consumer finances, about 60% of households have a good sense of their retirement security. munnell et al. (2008) focus on households whose self-assessment is inconsistent with the nrri. households owning a home, having college degree, unwilling to take risk and those with one-earner have higher probabilities of being in the “too worried” group compared with being appropriately worried. renters, those with less than a college education, those lacking a defined benefit plan, those with poor health and those willing to take some risk are more 142 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 likely to be in the “not worried enough” group compared with being appropriately not worried. 2.4. retirement adequacy hanna and chen (2008) report that previous research studies on the projected retirement adequacy of working households have produced a wide range of estimates, from 31% to 80% having an adequate retirement. however, studies using the survey of consumer finances (scf) have a narrower range of estimates. yuh, hanna, and montalto (1998) conclude that about 52% of households in 1995 would have enough assets and income for retirement assuming investment assets earn historical mean returns, but, based on pessimistic projection of investment returns, only 42% would have adequacy. yuh (2011) reports that 56% of pre-retired households in 2004 would be able to maintain 70% of permanent income in retirement. kim and hanna (2013) find that about 42% of working households in the 2010 scf are adequately prepared for retirement, based on mean projection of investment returns. 3. methodology 3.1. data and sample selection we use the 2010 survey of consumer finances (scf) dataset, a cross-sectional dataset sponsored by the federal reserve board. the scf provides comprehensive and detailed information on the financial status of u.s. households (bricker, kennickell, moore, and sabelhaus, 2012). our initial analytic sample is composed of households with a head who is age 35 to 60 and currently working full-time. some previous studies, for example, yuh et al. (1998), montalto, yuh, and hanna (2000), and chen (2007), use a sample of employed household heads age 35 to 70, but because many workers retire between 60 and 70 we use a different age criteria, to reduce possible selection bias. we assume that retirement age is exogenous, as is assumed by scholz, seshadri, and khitatrakun (2006) and brown, fang, and gomes (2012). for calculation of retirement adequacy, for workers who answer a question about the expected age for retirement from full-time work with “never retire,” we assume an expected retirement age of 70. there are 6,482 households in the public release of the 2010 scf dataset, and 2,283 of the households meet our sample criteria. this study mainly focuses on households that are projected to have objective inadequacy, therefore, for our multivariate analysis we exclude households with objective adequacy, and the final sample size is 1,203. 3.2. the dependent variable 3.2.1. objective measurement–projection for retirement adequacy in this study, retirement resources include social security benefits, projected defined benefit (db) pensions, projected part-time wages after retirement, and annuity distributions from projected retirement assets of the household head and any spouse/partner. our calculation of resources during retirement follows the retirement income stage method reported by 143k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 chen (2007) and kim et al. (2014). the projected retirement needs are based on estimates of each household’s expenditures, using mean expenditure to pretax income ratios for different income ranges (palmer, 1992, 1994), so high income households have a lower replacement ratio than low income households because of higher proportions of pretax income going for income taxes and saving. the replacement ratio is equal to the projected retirement income divided by estimated preretirement expenditures. we estimate benchmark replacement ratios derived from the 2010 consumer expenditure survey published by the bureau of labor statistics (u. s. department of labor, 2012). we use the normal household income of the household, and in the corresponding published income category in the bls, we set the benchmark ratio as the ratio of average annual expenditure divided by average pretax income in that bls category. for example, a household with a normal income of $45,000 would be assumed to have expenditures equal to 91% of pretax income, whereas a household with a normal income of $100,000 would be assumed to have expenditures equal to 69% of pretax income. our approach to defining objective retirement expense needs is more sophisticated than many previous studies, but it does not take into account the heterogeneity of households with the same income level, for instance, as discussed from a normative life cycle approach by scholz et al. (2006). the projected retirement income includes annual withdrawals from projected accumulated retirement assets with mean returns (cf., yuh et al., 1998), and other retirement income including social security benefits, defined benefit pensions, and part-time wages. if the projected retirement replacement ratio is equal to or greater than the benchmark replacement ratio, this household would have adequate retirement resources to sustain retirement needs. 3.2.2. subjective measurement–perception of retirement adequacy the scf has a variable for the respondent’s perception of the adequacy of retirement income, with five levels–totally inadequate (coded as 1), inadequate (coded as 2), enough to maintain living standards (coded as 3), satisfactory (coded as 4), and very satisfactory (coded as 5). though few studies have focused on the retirement perception variable in the scf survey, it plausibly reflects the respondent’s perception of having an adequate retirement. for the purpose of this study, the subjective measurement is a dichotomous indicator of households’ perception of having an adequate retirement with value equal to 1 if the value of indicator is 3, 4, or 5 (adequate), otherwise the value is 0 (inadequate). 3.2.3. categories for the dependent variable our initial analysis is based on all households with a head employed full-time and age 35 to 60. based on objective/subjective consistency, we define four categories of objective/ perceived retirement adequacy: realists having adequate resources (adequate realists), realists having inadequate resources (inadequate realists), households having only subjective adequacy (unrealistic optimists), and households having only objective adequacy (pessimists). given the importance of households with projected objective retirement inadequacy, we then focus only on households with projected inadequate retirement resources. we create a binary variable for objectively inadequate households for descriptive and multivariate analysis (unrealistic optimists vs. inadequate realists). 144 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 table 1 indicates the proportion of u.s. households in four categories of objective/ perceived retirement adequacy. only 42% of working households are adequately prepared for retirement based on an objective measure, whereas 46% rate their future retirement income adequate. there are various possible explanations for the low proportion of working households having objective adequacy, including hyperbolic discounting (angeletos et al., 2001; laibson, 1997) and simple mistakes (campbell, 2006). it is also possible that our assumptions about household preferences and expectations are not accurate, though we make assumptions similar to those made by most authors analyzing retirement adequacy. we return to this issue in our conclusions. the proportion of households rating their future retirement income as adequate is similar to the proportion of households with objective inadequacy, but only 52% of households have consistency between subjective and objective adequacy. about 20% of households are adequate realists and have both subjective and objective adequacy. pessimists (objective adequacy but subjective inadequacy) comprise 22% of households, while 26% are unrealistic optimists (subjective adequacy but objective inadequacy). about 32% of households are inadequate realists and have inadequacy consistently between the two retirement adequacy measurements. 3.3. independent variables presumably a worker’s ability to accurately assess retirement adequacy depends on cognitive ability and experience. the scf does not include a direct measure of cognitive ability, but kyrychenko and shum (2009) and stango and zinman (2009) use education as a proxy for financial cognition or sophistication. lusardi and mitchell (2007) find that the level of formal education is related to measures of financial literacy. we measure educational attainment by five dummy variables: less than high school, high school graduate, some college, bachelor degree, and post-bachelor degree. for multivariate analysis, we use a continuous variable, years of education of the head. huston, finke, and smith (2012) propose that the interviewer’s assessment of how well the respondent understands the scf survey questions is a proxy for financial sophistication, because it is related to the respondent’s understanding of personal finance. kim and hanna (2013) use this variable in a retirement adequacy study. the variable has four levels of understanding of scf questions: excellent, good, fair, and poor. for the purpose of this study, we code it as a binary variable because of the distribution of responses; it is coded as good understanding if the assessment is excellent or good, and as poor understanding if the response is fair or poor understanding of the scf survey. we also include use of a financial planner for saving and investment decision as a proxy for the ability to judge retirement adequacy. use of a financial planner is not necessarily a substitute for financial literacy (collins, 2012), and hanna (2011) reports that controlling for other factors, use of a financial planner increases with education. it is possible that a financial planner might give a client a false sense of confidence (e.g., cordell, smith, and terry, 2011). in the scf measure, it is not possible to identify the qualifications of financial planners that respondents report using. however, it is plausible that use of a financial planner should tend to improve a worker’s ability to judge retirement adequacy. some demographic variables may be related to the experience of the household, and 145k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 therefore might be plausibly related to the ability to judge retirement adequacy. presumably as one ages, experience will lead to a better ability to judge retirement adequacy. however, fluid intelligence decreases with age among adults (mcardle et al., 2011, p. 213). agarwal, driscoll, gabaix, and laibson (2009) suggest that the combined effect of decreasing cognitive ability and increasing experience results in the incidence of financial mistakes decreasing until age 53, then increasing. for descriptive analyses, we categorize age of the household head into three categories: 35–44, 45–54, and 55–60. for multivariate analysis, we use age as a continuous variable. of the four racial/ethnic groups identified in the scf, white, black, hispanic, and asian/other, it is plausible that whites have the most experience in dealing with investments, and so forth (hanna and lindamood, 2008; yao, gutter, and hanna, 2005). for the multivariate analysis, to obtain more robust estimates of effects, a binary variable of racial/ethnic group is created: white versus combination of other three race categories. we include a number of other variables as controls. marital status is measured by using four categories: married couple, female or male single household and partner. employment status of household head is measured with binary variables: employed and self-employed. economic status variables include normal income and retirement planning variables. to capture the possible nonlinearity of the relationship, household income is transformed into the natural log of normal income. retirement variables consist of having a defined benefit pension, having a defined contribution pension, and expected retirement age. the expected retirement age includes four categories partly based on social security benefit rules: before 62, between 62 and 65, over 65, and “never retire.” financial attitude variables include spending behavior, saving for retirement, and the respondent’s risk tolerance. the spending behavior includes three categories: reported spending is greater than income (deficit), spending is equal to income, and spending is less than income (surplus). households who report retirement as a savings goal are coded saving for retirement. the level of risk tolerance is measured as four dummy variables for no risk, average, above average, and substantial risk. 3.4. analysis a logistic regression model is used to test the effect of our selected explanatory variables on the likelihood of being unrealistic optimists among households with inadequate resources. we use the repeated-imputation inference (rii) method, which provides variance estimates more closely representing the true variances than estimates obtained by only one implicate (lindamood, hanna, and bi, 2007). means tests are used with the rii technique to examine differences of the projected retirement adequacy between households with different perception of retirement. our multivariate analysis is unweighted (lindamood et al., 2007). 3.5. research hypotheses hypothesis 1: cognitive ability is negatively related to the likelihood of being an unrealistic optimist. 146 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 the highest educational attainment of the household head and understanding of the scf survey question are used as proxies to cognitive ability. use of a financial planner for saving and investment decisions may supplement an individual’s cognitive ability. hypothesis 2: financial experience is negatively related to the likelihood of being an unrealistic optimist. financial experience variables include age of household head, race/ethnic identity, and having a defined benefit/contribution plan. 4. results 4.1. household characteristics of households with objective inadequacy table 2 shows patterns of selected household characteristics by two categories of objective/perceived retirement adequacy, for households with objective inadequacy (i.e., unrealistic optimist and inadequate realist). those who have a college degree (43%) are less likely to be an unrealistic optimist than those with less than a high school degree (48%). those who report using a financial planner for saving and investment decisions (49%) have a higher rate of being unrealistic than those who do not use a financial planner (44%). households with a poor understanding of survey questions (52%) have higher proportion of being unrealistic than those with a good understanding (44%). the rate of being unrealistic increases with the age of household head, with 41% of those age 35 to 44 unrealistically thinking they will have adequate retirement income, but 50% of those age 55 to 60 being unrealistic. whites have a lower rate of being unrealistic than each of the other racial/ethnic groups. because of the low number of blacks, hispanics, and asian/others in the analytic sample and the similar rates of being unrealistic in each of those groups, in the multivariate analysis we combine the three groups. for the descriptive comparison, the combined group has a rate of being unrealistic of 48%, compared with 43% for whites, and that difference is significant. for households having a db plan, the rate of being unrealistic is higher than the rate for those without a db plan (58% vs. 43%.) there is a similar pattern for having a dc plan, with the rate of being unrealistic 55% among those with a dc plan and 38% for those without one. the proportion of being unrealistic is higher for married couples than for other groups. as the household head’s expected retirement age increases, the rate of being unrealistic decreases, from 55% for those who expect to retire before 62 to 42% for those who expect to retire after 65, and the rate among those who expect to never retire is very low, only 26%. households that spend less than or equal to income are more likely to be unrealistic than those who report spending more than income. households that report having a saving objective related to retirement are more likely to be unrealistic than those who do not report having a retirement goal. as risk tolerance increases, the proportion of being unrealistic increases from 40% (no risk) to 52% (substantial risk). for the multivariate analysis, we 147k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 table 2 rate of unrealistic optimists by selected household characteristics, among households with head employed full-time and age 35 to 60, with objective retirement inadequacy 2010 scf variable distribution, among households with objective inadequacy % unrealistic optimists significance level education of household head less than high school 11.1 47.6 reference high school graduate 31.8 44.6 0.2093 some college 18.3 47.6 0.9945 bachelor degree 27.9 43.1 0.0597 post-bachelor degree 10.9 42.8 0.0920 use of financial planner use 24.6 48.6 �.0001 do not use 75.4 43.7 reference understanding of the scf survey question good understanding 89.6 44.1 0.0006 poor understanding 10.4 51.6 reference age of household head 35–44 40.3 40.5 reference 45–54 42.1 47.1 �.0001 55–60 17.6 49.6 �.0001 racial-ethnic category white 64.0 43.4 reference black 14.7 47.9 0.0207 hispanic 16.2 46.2 0.1309 asian or others 5.0 50.5 0.0179 combination of black, hispanic, asian/other 36.0 47.5 0.0029 defined benefit (db) plan have db plan 12.5 58.2 �.001 do not have db plan 87.5 43.0 reference defined contribution (dc) plan have dc plan 43.2 54.6 �.0001 do not have dc plan 56.8 37.5 reference marital status married 58.8 47.6 reference single male 13.5 40.8 0.0006 single female 21.9 42.2 0.0013 partner 5.8 36.9 0.0002 employment status salary worker 91.4 46.5 reference self-employment 8.6 27.6 �.0001 expected retirement age of head retirement age � 62 25.2 55.3 reference 62 � retirement age � 65 37.3 49.3 0.0005 retirement age � 65 18.1 41.6 �.0001 never retire 19.4 26.0 �.0001 spending behavior spending � income 50.2 50.2 �.0001 spending � income 33.1 42.4 �.0001 spending � income 16.7 34.1 reference having retirement purpose have 52.4 48.4 �.0001 do not have 47.6 41.0 reference (continued on next page) 148 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 combine the substantial and above average responses because of the very small number of substantial responses. 4.2. multivariate analyses table 3 presents the logistic regression results of the likelihood of being an unrealistic optimist among those who are projected to have objective inadequacy. as years of education increase, the likelihood of being unrealistic decreases. the other variables assumed to be related to ability, the interviewer’s perception of the respondent’s understanding of the survey, and use of a financial planner for saving or investment decisions, are not significantly related to being unrealistic. most variables related to financial experience have effects contrary to expectations. age of the household head is positively related to the likelihood of being unrealistic. households with a defined benefit pension and those with a defined contribution plan are more likely to be unrealistic than similar households not having a defined benefit or a defined contribution plan. however, white respondents are less likely to be unrealistic than similar households in the other three racial/ethnic categories, so to the extent that whites have more financial experience than those in the other categories, experience may be related to better assessment of retirement adequacy. households with a self-employed head are less likely to be unrealistic than those with a head working for a salary. households who expect to retire after age 65 or to never retire are less likely to be unrealistic than those with an expected retirement age less than 62. households who spend less than or equal to income are more likely to be unrealistic than those with a deficit. lastly, households willing to take above-average or substantial risks are more likely to be unrealistic than those unwilling to take any risk. 5. discussions and implications the primary goal of this study is to analyze the deviation between a household’s objective and subjective assessment of retirement adequacy. four different types of households— adequate realists, inadequate realists, unrealistic optimists, and pessimists—are categotable 2 (continued) variable distribution, among households with objective inadequacy % unrealistic optimists significance level risk tolerance no risk 43.5 40.2 reference average risk 38.3 47.2 �.0001 above average risk 15.5 51.0 �.0001 substantial risk 2.7 52.0 0.0033 combination of above average and substantial risk 18.2 51.1 �.0001 restrictions are described in the method section, and include head being 35 or older, but no more than 60 and being in the labor force. n � 1,203. significance levels based on rii means tests. 149k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 rized by objective/subjective retirement assessments. only 42% have objective retirement adequacy, whereas 46% perceive they will have an adequate retirement. about 52% have a consistency between objective and subjective adequacy. for our multivariate analysis, we focus on the likelihood of being an unrealistic optimist among a sample of households with objective inadequacy. our descriptive results shown in table 2 indicate among households with inadequate retirement resources, the households with a head with a college degree have lower rate of being unrealistic optimists than those with less than a high school degree. those reporting use of a financial planner are significantly more likely to have to be unrealistic than those who do not use a financial planner, raising questions about the accuracy or benefit of financial table 3 logistic regression analysis of likelihood of being unrealistic optimists, of households with head employed full-time and age 35 to 60, and objective retirement inadequacy, 2010 scf variable coefficient two-tail p-valuea standard error odds ratio cognitive ability: education of household head, use of financial planner, good understanding of the scf survey question education of household head (continuous variable) �0.0681 0.0094 0.0263 0.934 use of financial planner (reference category: no) 0.0962 0.5301 0.1525 1.101 good understanding of the scf survey question (reference category: no) �0.2777 0.2045 0.2192 0.758 financial experience: age of household head, racial-ethnic status, having a retirement plan age of household head (continuous variable) 0.0204 0.0307 0.0094 1.021 racial-ethnic category (reference category: white) combination of black, hispanic, and asian/other 0.3648 0.0140 0.1488 1.440 have defined benefit plan (reference category: no) 0.5289 0.0088 0.2025 1.697 have defined contribution plan (reference category: no) 0.4976 0.0006 0.1445 1.645 control variables: marital status, self-employed, log of income, expected retirement age, spending behavior, have retirement purpose, risk tolerance marital status (reference category: married) single male �0.1838 0.3755 0.2072 0.832 single female 0.0199 0.9081 0.1737 1.020 partner �0.1007 0.7171 0.2817 0.904 self-employment (reference: salary worker) �0.5618 0.0102 0.2190 0.570 log of income �0.0164 0.8244 0.0730 0.984 expected retirement age (reference category: under 62) 62 � retirement age � 65 �0.2072 0.2347 0.1744 0.813 retirement age � 65 �0.5484 0.0057 0.1981 0.578 never retire �0.9492 �.0001 0.2061 0.387 spending behavior (reference category: spending � income) spending � income 0.4015 0.0461 0.2014 1.494 spending � income 0.6515 0.0006 0.1910 1.919 have retirement purpose (reference: no) 0.046 0.7495 0.1443 1.047 risk tolerance (reference category: take no risk) average risk 0.252 0.1082 0.1569 1.287 combination of above-average and substantial risk 0.4357 0.0243 0.1938 1.546 concordance (mean) 67.6% a unweighted rii analysis of 2010 scf dataset, analysis of 1,203 households with full-time employed head age 35–60, and with objective retirement inadequacy. 150 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 planning services. those who have good understanding of the scf survey question are less likely to be unrealistic than those who have a poor understanding, suggesting that cognitive ability does play a role in accurate assessment of retirement adequacy. among financial experience variables, age of the household head is positively associated with the likelihood of being unrealistically optimistic. it is possible that cognitive decline may play a role in this pattern, though given that agarwal et al. (2009) suggest that the incidence of financial mistakes might decrease with age until about age 53, it seems unlikely that this could be the only factor related to this pattern. the effect of age on over-optimism might be related to the cognitive dissonance of some households. the theory of cognitive dissonance (festinger, 1957) posits that individuals are distressed by conflicting cognitive elements. morton (1993) suggests that individuals attempt to decrease their dissonance by either: (1) changing their past values, feelings, or opinions, or, (2) attempting to justify or rationalize their choices. in behavioral finance, financial cognitive dissonance is used in that individuals change their investment styles or beliefs to support their financial decisions (ricciardi and simon, 2000). with financial experience, people should be better able to evaluate objective situations, but cognitive dissonance may lead to accepting lower standard of living in retirement. another explanation might be that an individual’s investment knowledge and skills change by age because of experience. korniotis and kumar (2011) find that older investors are more likely to have greater investment experience and knowledge while they may face adverse effects of cognitive aging on investment skills. older investors would have more experience in stock market cycles and have more belief in the possibility of a stock market recovery by retirement, but younger investors, with more limited experience, might be pessimistic. for evaluation of retirement adequacy, experience and salience should increase with age, and cognitive ability probably plays less of a role than it might for investment decisions. whites are less likely to be unrealistic than households categorized as black or hispanic and asian/others, perhaps because of greater financial experience. whites have more financial experiences with stock ownership than other minority households (hanna and lindamood, 2008). in addition, whites are more likely to be financially literate than blacks and hispanics (lusardi and mitchell, 2011). loving, finke, and salter (2012) find that the racial/ethnic difference in stock market participation in 2004 (whites vs. a combined group of blacks and hispanics) is not significant after controlling for measures of cognitive ability and investor experience. we do not have good proxies for either factor, but we are controlling for education, and it seems plausible that the racial/ethnic differences in being realistic about retirement adequacy might be because of differences in investor experience. households having a defined benefit pension are more likely to be unrealistic than similar households not having a defined benefit pension. this is a contrary result to our expectation because households with defined benefit plan are able to assess their guaranteed retirement income. one possible explanation would be that households might not have good numeracy related to retirement income, and may just assume that simply having these plans will achieve adequacy. similarly, households with a defined contribution plan are more likely to be unrealistic than those without a defined contribution plan. it is plausible that many workers with defined contribution plans are unfamiliar with features of their plans, may not 151k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 be able to accurately assess retirement adequacy, or they may assume that just having a plan may lead to retirement adequacy. households who expect that they will retire after 65, or never retire, are less likely to be unrealistic than those with an expected retirement age less than 62. unrealistic optimists may plan to retire at a relatively earlier age because of their optimism for retirement adequacy. as expected, households willing to take above-average or substantial risk have a higher likelihood of being unrealistic than similar households unwilling to take risk. pirinsky (2013) suggests that more confident people are more likely to take risks than those with less confidence, and optimism is found to be one of the strongest determinants of being willing to take risk. the more sophisticated respondents might assume that despite the stock market crash leading to big decreases in their retirement wealth, mean reversion of returns might lead to higher returns than the historical means. many of the descriptive patterns shown in table 2 also are significant after controlling for other factors in the logistic regression (table 3.) the likelihood of being an unrealistic optimist decreases with years of education. however, understanding of the survey questions and use of a financial planner are not significantly related to being unrealistic. the fact that use of a financial planner has a positive though insignificant effect on being unrealistic does not support the idea that financial planners contribute to better understanding of retirement adequacy. the likelihood of being an unrealistic optimist increases with age, so the previous discussion about possible explanations based on the descriptive patterns in table 2 is relevant. it is possible that as workers get older, they lower their assessment of what is acceptable. it is unclear whether they are being unrealistic (compared with an objective standard that assumes a goal of maintaining a similar standard of living in retirement) or are just accepting of the likelihood of a lower standard of living in retirement. it is also possible that older workers, having experienced more economic cycles in financial markets, are more likely than younger workers to expect mean reversion in returns. our objective measure projects balances reported in 2010 forward using historical mean returns. controlling for other factors, whites are less likely than those in other racial/ethnic groups to be unrealistic. this result suggests the need for retirement education targeted at these groups. as with the descriptive results in table 2, controlling for other factors, those with a defined benefit plan are more likely to be unrealistically optimistic than those without a defined benefit plan. there is a similar effect for having a defined contribution plan. assuming that our projection of objective adequacy is valid, these two results suggest the need for better employer education for workers with such plans. it is also possible that workers with these plans have greater financial experience, and perhaps are more likely in 2010 to expect mean reversion in returns than less experienced workers without such plans. our study highlights the substantial divergence between objective and subjective assessment of retirement adequacy. in addition, the findings in this research are partially consistent with our two hypotheses, indicating that households with greater cognitive ability (as proxied by education) and more financially experienced (as proxied by racial/ethnic group) are more likely to have accurate assessment of their retirement preparedness. however, some of our empirical findings appear puzzling, perhaps because we cannot directly measure the degree of cognitive ability and financial experiences from the scf dataset. to obtain more robust 152 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 results for retirement perception research, a different dataset should be used with direct measures of cognitive ability and/or financial literacy related to the retirement context, such as the health and retirement study (hrs). references agarwal, s., driscoll, j., gabaix, x., & laibson. d. (2009). the age of reason: financial decisions over the life-cycle and implications for regulation. brookings papers on economic activity, 2, 51–117. agarwal, s., & mazumder, b. (2013). cognitive abilities and household financial decision-making, american economic journal: applied economics, 5, 193–207. ando, a., & modigliani, f. (1963). the life cycle hypothesis of saving: aggregate implications and tests, american economic review, 53, 55–84. angeletos, g., laibson, d., repetto, a., tobacman, j., & weinberg, s. (2001). the hyperbolic consumption model: calibration, simulation, and empirical evaluation, the journal of economic perspectives, 15, 47–68. bricker, j., kennickell, a. b., moore, k. b., & sabelhaus, j. (2012). changes in u.s. family finances from 2007 to 2010: evidence from the survey of consumer finances. federal reserve bulletin, 98, 1–80. bricker, j., bucks, b., kennickell, a., mach, t., & moore, k. (2011). surveying the aftermath of the storm: changes in family finances from 2007 to 2009. feds working paper, 17. brown, j., fang, c., & gomes, f. (2012). risk and returns to education. social science research network working papers series (available at ssrn: http://papers.ssrn.com/sol3/papers.cfm?abstract_id�2021598). browning, m., & crossley, t. f. (2001). the life-cycle model of consumption and saving. journal of economic perspectives, 15, 3–22. campbell, j. y.(2006). household finance. journal of finance, 61, 1553–1604. carroll, c. d. (1997). buffer-stock savings and the life cycle/permanent income hypothesis. quarterly journal of economics, 112, 1–55. chan, s., & stevens, a. (2008). what you don’t know cannot help you: pension knowledge and retirement decision-making. review of economics and statistics, 90, 253–266. chen, c. c. (2007). changes in retirement adequacy, 1995–2004: accounting for retirement stages. dissertation, the ohio state university. choi, j., laibson, d., & madrian, b. (2004). plan design and 401(k) savings outcomes. national tax journal, 57, 275–298. christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54, 18–38. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. cordell, d. m., smith, r., & terry, a.(2011). overconfidence in financial planners. financial services review, 20, 253–263. cummings, b. f., finke, m. s., & james, r. n. (2011). the impact of cognitive ability on roth ira ownership. social science research network working papers series (available at ssrn: http://ssrn.com/abstract� 1968984). dohmen, t., falk, a., huffman, d., & sunde, u. (2010). are risk aversion and impatience related to cognitive ability? american economic review, 100, 1238–1260. festinger, l. (1957). a theory of cognitive dissonance. stanford, ca: stanford university press. fornero, e., & monticone, c. (2011). financial literacy and pension plan participation in italy. cerp working papers, center for research on pensions and welfare policies, turin, italy. frederick, s. (2005). cognitive reflection and decision making. journal of economic perspectives, 19, 25–42. glamser, f. (1976). determinants of a positive attitude toward retirement. journal of gerontology, 31, 104–107. grinblatt, m., keloharju, m., & linnainmaa, j. (2011). iq and stock market participation. journal of finance, 66, 2119–2164. 153k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 gustman, a. l., steinmeier, t. l., tabatabai, n. (2007). imperfect knowledge of plan type. nber working paper, no. 13379. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. hanna, s. d., & chen, c. c. (2008). retirement savings. in j. xiao (ed.), handbook of consumer finance research (pp. 35–46). new york, ny: springer publishing,. hanna, s. d., & lindamood, s. (2008). the decrease in stock ownership by minority households. journal of financial counseling and planning, 19, 46–58. huston, s. j., finke, m. s., & smith, h. (2012). a financial sophistication proxy for the survey of consumer finances. applied economics letters, 19, 1275–1278. joo, s., & grable, j. e. (2001). factors associated with seeking and using professional retirement-planning help. family and consumer science research journal, 30, 37–63. kilty, k., & behling, j. (1985). predicting the retirement intentions and attitudes of professional workers. the journal of gerontology, 40, 219–227. kim, k. t., & hanna, s. d. (2013). does financial sophistication matter in retirement preparedness? evidence from the 2010 survey of consumer finances, consumer interests annual, 59. kim, k. t., hanna, s. d., & chen, s. c. (2014). consideration of retirement income stages in planning for retirement. journal of personal finance, 13, 52–64. knoll, m. (2010). the role of behavioral economics and behavioral decision making in americans’ retirement savings decisions. social security bulletin, 70, 1–23. korniotis, g., & kumar, a. (2011). do older investors make better investment decisions? review of economics and statistics, 93, 244–265. kyrychenko, v., & shum, p. (2009). who holds foreign stocks and bonds? characteristics of active investors in foreign securities. financial services review, 18, 1–21. laibson, d. (1997). golden eggs and hyperbolic discounting. quarterly journal of economics, 62, 443–478. lindamood, s., hanna, s. d., & bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41, 195–222. loving, a. c., finke, m. s., & salter, j. r. (2012). explaining the 2004 decrease in minority stock market ownership. review of black political economy, 39, 403–425. lusardi, a., & mitchell, o. s. (2007). baby boomers’ retirement security: the role of planning, financial literacy and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. s. (2011). financial literacy and planning: implications for retirement wellbeing. nber working paper, no. 17078. malroutu, y., & xiao, j. (1995). perceived adequacy of retirement income. financial counseling and planning, 6, 17–23. mcardle, j. j., smith, j. p., & willis, r. (2011). cognition and economic outcomes in the health and retirement survey. in d. a. wise (ed.), explorations in the economics of aging (pp. 209–233). chicago, il: university of chicago press. modigliani, f. & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. in k. kurihara (ed.), post-keynesian economics (pp. 388–436). new brunswick, nj: rutgers university press. montalto, c. p., yuh, y., & hanna, s. d. (2000). determinants of planned retirement age. financial services review, 9, 1–15. morton, h. (1993). the story of psychology. new york, ny: bantam double dell publishing group, inc. munnell, a. h. (2012). 2010 scf suggests even greater retirement risks. center for retirement research at boston college issue in brief 12–15. munnell, a. h., golub-sass, f., soto, m., & webb, a. (2008). do households have a good sense of their retirement preparedness? center for retirement research at boston college issue in brief 8–11. palmer, b. a. (1992). establishing retirement income objectives: the 1991 retire project. benefits quarterly, third quarter, 6–15. palmer, b. a. (1994). retirement income replacement ratios: an update. benefits quarterly, second quarter, 59–75. 154 k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 pirinsky, c. (2013). confidence and economic attitudes. journal of economic behavior & organization, 91, 139–158. ricciardi, v., & simon, h. k. (2000). what is behavioral finance? business, education & technology journal, 2, 1–9. scholz, j., seshadri, a., & khitatrakun, s. (2006). are americans saving “optimally” for retirement? journal of political economy, 114, 607–643. shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609–643. social security administration. (2014). the 2014 annual report of the board of trustees of the federal old-age and survivors insurance and federal disability insurance trust funds. (available at http://www.ssa.gov/ oact/tr/2014/tr2014.pdf). stango, v., & zinman, j. (2009). exponential growth bias and household finance. the journal of finance, 64, 2807–2849. u. s. department of labor. (2012). consumer expenditure report: consumer expenditures in 2010. (available at http://www.bls.gov/cex/csxann10.pdf). yao, r., gutter, m. s., & hanna, s. d. (2005). the financial risk tolerance of blacks, hispanics and whites. financial counseling and planning, 16, 51–62. yuh, y. (2011). assessing adequacy of retirement income for us households: a replacement ratio approach. geneva papers on risk and insurance-issues and practice, 36, 304–323. yuh, y., hanna, s. d., & montalto, c. p. (1998). mean and pessimistic projections of retirement adequacy. financial services review, 9, 175–193. 155k.t. kim, s.d. hanna / financial services review 24 (2015) 139–155 do financial networks matter in retirement investment decisions? evidence from generation yers yunhyung chunga, youngkyun parka,* acollege of business and economics, university of idaho, 875 perimeter drive ms 3161, moscow, id 83844-3161, usa abstract using experimental survey data collected from a sample of generation yers, we examine the joint influence of financial literacy and financial networks on individual retirement investment decisions. we find, first, that financial literacy and financial network intensity (the network strength with the financially literate) are positively related to stock allocation. second, the positive relationship between financial literacy and stock allocation, however, is significant only among those having high financial network intensity. this finding suggests that the positive effects of financial literacy documented in the literature can be limited to only those who have strong networks with the financially literate. © 2015 academy of financial services. all rights reserved. jel classification: g11; j26 keywords: retirement investment decisions; financial networks; financial literacy 1. introduction while the financial environment becomes complex and volatile, individual financial decision-making has been considered an important factor of determining an individual’s financial well-being. thus, a large body of research has investigated what factors influence financial decision-making and has found that financial literacy is a key determinant in many areas such as money management, credit, investment, and retirement planning (e.g., campbell, 2006; lusardi & tufano, 2009; moore, 2003; perry & morris, 2005). however, * corresponding author. tel.: �1-208-885-7154; fax: �1-208-885-5347. e-mail address: youngpark@uidaho.edu (y. park) financial services review 24 (2015) 77–99 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. financial literacy alone may not be able to improve financial behaviors substantially (e.g., fernandes, lynch, & netemeyer, 2014; kiviat & morduch, 2012).1 because of the fastchanging financial product offerings and industry environment, acquiring and processing up-to-date financial knowledge and information is quite challenging even to those who are financially literate (willis, 2011). consequently, individual investors often rely on other people to make investment decisions, instead of solely on their own knowledge. relatively recently, research on financial decision-making has begun to recognize the importance of social interactions in individual investment decisions. because individual investors may be able to reduce time and effort to acquire financial knowledge and information through social interactions with others, “social” people are more likely to participate in financial markets than others who do not engage in social interactions (brown, ivković, smith, & weisbenner, 2008; duflo & saez, 2002; hong, kubik, & stein, 2004; ivković & weisbenner, 2007; kaustia & knupfer, 2012; lu, 2011). in line with this argument, research on social networks in management finds that individuals may acquire a high volume of and diverse work-related knowledge and information through frequent communication with co-workers or experts who have task-related knowledge (hansen, mors, & lovas, 2005; mcfadyen & cannella, 2004; reinholt, pedersen, & foss, 2011; wang & noe, 2010). because tacit and confidential knowledge tend to be shared only with whom people trust (nahapiet & goshal, 1998; yli-renko, autio, & sapienza, 2001) and transferred through strong network ties (chung & jackson, 2013; hansen, podolny, & pfeffer, 2001), strong connections with people who are financially literate may facilitate individuals to obtain reliable and opportune information for investments without excessive time and effort. drawing on the literature on financial literacy and social networks, we propose that an individual investor’s decision-making may be affected by not only his or her financial knowledge but also social networks. among various types of social networks (e.g., career advice networks and task information networks), we focus on individuals’ social networks for acquiring financial or investment information, which are dubbed “financial networks.” while prior studies focused on social activities, few studies examined how social networks influence investment decisions.2 thus, this study aims to fill this gap by taking account of financial networks in discussing individuals’ investment decisions. among financial network characteristics, we focus on financial network intensity, defined as communication frequency with the financially literate, because frequent interactions with the financially literate may enable an individual investor to acquire critical and reliable information and knowledge for his or her investments. taken together, we examine how financial knowledge and financial network intensity are associated with individual investment decisions. among various individual investment decisions, we consider the following sequential decision-making process for retirement investment: first, whether an individual investor chooses a default investment option or not and, second, when selecting “no default,” to what extent he/she allocates contributions to stocks. for this examination, we collected experimental survey data from senior business college students at age 20–26 in 2012, who belong to the generation y (ages 18–35 in 2012). most of them will enter an early stage of their working career upon graduation and likely become defined contribution plan participants. we chose this sample because retirement investment decisions in their early career may significantly affect their retirement 78 y. chung, y. park / financial services review 24 (2015) 77–99 wealth. once new employees make a retirement investment decision, they tend to stick with the status quo and avoid changing their retirement investment portfolio (choi, labibson, madrian, & metrick, 2004, 2006; madrian & shea, 2001). in addition, given that the social security trust fund is projected to be exhausted in 2034 and at that time that social security will be able to cover only 75% of scheduled benefits (social security and medicare boards of trustees, 2014), how to construct a retirement portfolio would be more important to generation yers than older generations. with a sample (n � 97) of senior students in a business college, first, a choice of no default in a retirement plan is positively associated with financial literacy (or financial knowledge), but not significantly associated with financial network intensity (or the network strength with the financially literate).3 second, respondents’ contribution allocation to stock funds is positively associated with financial literacy or financial network intensity. third, financial literacy does not interact with financial network intensity on a choice of a default option, but they interact with each other on respondents’ contribution allocation to stock funds in a retirement portfolio. in an additional analysis on the interaction effect, the positive effects of financial literacy on stock allocation in a retirement portfolio are significant only among those who have high financial network intensity.4 the interaction between financial literacy and financial network intensity also affects how much respondents’ retirement portfolios deviate from age-appropriate stock allocations, the deviation that is evaluated with respect to the morningstar lifetime allocation indexes. we find that those with a high level of financial literacy have smaller deviations from the age-appropriate stock allocations only when they have strong networks with the financially literate. this study enhances an understanding of retirement investment decision-making and long-term financial planning. while extant research finds financial literacy as a key determinant in retirement investment decision-making and long-term financial planning (e.g., alhenawi & elkhal, 2013; lusardi & mitchell, 2007a, 2007b), our research finds that financial literacy and social networks interplay in retirement investment decision-making. specifically, the positive effects of financial literacy on retirement investment decisions documented in the literature can be limited to only those who have strong networks with financially literate people. the results of this study may also provide practical implications to individual investors, financial planners, and employer-sponsored financial education programs. to improve investment decisions, individual investors’ own financial knowledge alone may not be enough; when they synthesize their own financial knowledge and the knowledge acquired through interactions with the financially literate, their investment decisions can be significantly enhanced. therefore, our findings shed light on the importance of financial networks with family, co-workers, and professional financial planners when individual investors make investment decisions. finally, employer-sponsored financial education programs may be designed not only to elevate employees’ financial knowledge but also to provide more opportunities to build networks for financial advice (such as more opportunities to communicate with financial planners or advisors). the remainder of the article is organized as follows. section 2 provides a literature review on financial literacy, financial network intensity, and retirement investment decisions and addresses hypotheses. section 3 describes the sample, survey instruments, and model 79y. chung, y. park / financial services review 24 (2015) 77–99 specification to test the hypotheses. section 4 presents regression results indicating that financial literacy and/or financial network intensity can affect retirement investment decisions. section 5 concludes with a summary of findings and suggestions for future research. 2. financial literacy, financial network intensity, and retirement investment decisions managing an investment portfolio requires an individual investor to spend significant time and effort in understanding various finance concepts such as asset returns, volatility, and covariance between asset returns. thus, information costs—for example, costs of acquiring and processing information about risks and returns—represent a significant barrier for individual investors to build their own investment portfolio (christelis, jappelli, & padula, 2010; vissing-jørgensen, 2003). building on the notion of information cost, we argue that individual investors’ financial literacy and financial networks reduce information cost, which consequently influences their retirement investment portfolios. 2.1. financial literacy information costs for investments may be substantially large for people who have low financial knowledge and skills (vissing-jørgensen, 2003). because they have to spend significant time and effort in acquiring and processing financial information, they may give up the costly process for investment decision-making and seek a simple solution (bettman & park, 1980). agnew and szykman (2005) find that individual investors with a low level of financial knowledge are likely to opt for a default investment option more often than those with a high level of financial knowledge. in addition, choi et al. (2004) find from several 401(k) plans that young, female, and low-income participants with short tenure, who are typically associated with a low level of financial literacy (banerjee, 2011; hung, parker, & yoong, 2009; lusardi & mitchell, 2007a), are more likely to choose a default investment option than others. based on the theoretical and empirical findings in the previous literature, we argue that individuals who have little knowledge of finance are more likely to select a default investment option than those who have much knowledge of finance. on the contrary, individuals with a high level of financial literacy are more likely to choose no default and build their own retirement investment portfolio than those with a low level of financial literacy. in this respect, we hypothesize the following: h1a: financial literacy is positively associated with a choice of “no default” in a retirement plan. individual investors who are financially literate may be able to reduce the costs of acquiring and processing financial information and, as a result, increase stockholding.5 because financial literacy helps reduce the fixed costs of acquiring and processing financial 80 y. chung, y. park / financial services review 24 (2015) 77–99 information (e.g., vissing-jørgensen, 2003),6 financially literate investors are likely to hold stocks (van rooij et al., 2011; yoong, 2010).7 in addition, once individual investors acquire financial knowledge or information, they may apply this knowledge or information to a larger volume of the risky asset with an increase in the expected rate of return (delavande, rohwedder, & willis, 2008). in this respect, we expect that individual investors with a high level of financial literacy are likely to allocate more contributions to stocks than those with a low level of financial literacy, all else being equal. h1b: financial literacy is positively associated with stock allocation in a retirement investment portfolio. 2.2. financial network intensity individuals’ investment decisions are affected by not only their own financial literacy but also social interactions such as word-of-mouth or observational learning (banerjee, 1992; bikchandani et al., 1992; ellison & fudenberg, 1993, 1995). for example, individuals’ investment behavior is affected by investment decisions or outcomes of their co-workers (duflo & saez, 2002; lu, 2011) or neighbors (brown et al., 2008; hong et al., 2004; ivković & weisbenner, 2007; kaustia & knupfer, 2012) because social interactions, like financial literacy, may be able to reduce the costs of acquiring and processing financial information. network ties can serve both instrumental/informational and expressive/social purposes (balkundi & harrison, 2006; cascioaro & lobo, 2008). this study focuses on the instrumental role of networks, assuming that instrumental external ties are most relevant to acquiring financial knowledge and information. for instrumental external ties, the network strength (in terms of communication frequency) with the financial literate is of our particular interest; the costs of acquiring and processing information related to investment may be reduced by frequent interactions with people who are financially knowledgeable. social network literature documents that frequent communications generate “strong ties” and, thus, trust between people in the network (e.g., krackhardt, 1992). trustworthy information built on frequent communications in one’s networks, then, can affect his or her investment decisions (e.g., guiso, sapienza, & zingales, 2008).8 because people tend to share critical and confidential knowledge only with whom they trust (nahapiet & goshal, 1998; yli-renko, autio, & sapienza, 2001), frequent interactions with people who have high financial literacy may provide individuals easy access to a high quality of investment and knowledge. therefore, through social networks with the financially literate, individuals may acquire reliable and opportune information for their investments. drawing on the studies on social network, we expect that financial network intensity, indicating the relationship strength with the financially literate, may influence individuals’ investment decisions. specifically, for a given default investment option, we expect that 81y. chung, y. park / financial services review 24 (2015) 77–99 individual investors who have strong networks with the financially literate are likely to build their own retirement portfolio instead of selecting a default option. thus, we hypothesize the following: h2a: financial network intensity is positively associated with a choice of “no default” in a retirement plan. a strong relationship with people who have a high level of financial literacy may also encourage individuals to increase stock allocation in their retirement investment portfolio. because individuals are likely to obtain critical and reliable financial or investment information through frequent interactions with the financially literate, they may significantly be able to reduce time and effort to obtain and process the information necessary for their investments. individuals usually have limited time and effort to spend on constructing a retirement portfolio. thus, strong networks with financially literate people (i.e., high financial network intensity) may enable individual investors to easily obtain investment information with lower costs and, as a result, to increase stockholdings. in this regard, we hypothesize the following: h2b: financial network intensity is positively associated with stock allocation in a retirement investment portfolio. 2.3. interactions of financial literacy and financial network intensity in the previous sections, we developed the hypotheses that financial literacy or financial network intensity affects individuals’ investment decisions for retirement. the two factors may also interact for retirement investment decisions. specifically, we argue that financially knowledgeable individuals are more likely to build their own portfolio and increase stock allocations when they have higher financial network intensity (i.e., stronger networks with the financially literate) than otherwise. individual investors who have high financial knowledge may have greater ability to understand broad and diverse knowledge about financial investment obtained through interactions with the financially literate. in addition, financially knowledgeable investors may easily synthesize their own financial knowledge and the knowledge acquired through social networks with the financially literate. as a result, financial networks may enable individual investors with high financial literacy to reduce their opportunity cost of obtaining and processing financial knowledge, which consequently eases their investment decision-making. hence, we argue that individual investors’ financial network intensity complements their financial literacy: any positive effects of financial literacy on retirement investment decisions would be stronger among those having high financial network intensity than those having low financial network intensity. accordingly, we expect the following: 82 y. chung, y. park / financial services review 24 (2015) 77–99 h3a: financial network intensity interacts with financial literacy on the selection of a default investment option in a retirement plan. h3b: financial network intensity interacts with financial literacy on stock allocation in a retirement investment portfolio. 3. sample, survey instruments, and model specification 3.1. sample and procedure to examine the effects of financial literacy and financial network intensity on investment decisions, we collected data from 111 senior college students in a business school located in the northwestern united states, using a traditional paper-and-pencil survey in the classroom in spring and fall 2012 semesters. to enhance survey participants’ calculation accuracy, calculators were distributed at the beginning of the survey. students who participated in the survey received class engagement credits. four respondents who were 35 or older and 10 respondents who did not provide all the information are excluded from the sample, resulting in a sample of 97 respondents. 3.2. experimental survey design, model specification, and measures 3.1.1. investment decisions for retirement using an experimental survey method, we developed a scenario regarding investment decisions in a context of a 401(k) plan. in the scenario, we asked respondents to assume that they were recently hired and asked to make investment decisions for a 401(k) plan with an array of investment options provided by a company. we assessed two dependent variables from sequential retirement investment decisions. first, survey respondents were asked to make a decision for a default investment option, which is a money market fund. if a respondent did not choose the default option, then he or she was asked to build his or her own retirement portfolio from an array of investment options: 10 different mutual funds including a money market fund.9 the information for each fund was provided in a separate brochure so that survey participants could compare investment objectives, strategies, expenses, risk and return, and the performance of the funds. the fund information was provided using actual funds in the marketplace, but actual fund names including fund family names were not disclosed in the brochure to prevent respondents from selecting a fund only because of their familiarity with fund family names. table 1 summarizes fund information provided for survey participants. respondents were asked to use the fund information and allocate their contributions as a percentage to each fund up to the total contribution of 100%. 3.1.2. model specification to reflect respondents’ sequential investment decisions, we estimate the following heckman two-stage selection model: 83y. chung, y. park / financial services review 24 (2015) 77–99 selection: s*i � �0 � �1fli � �2fni � �3�flifni� � �4zi � ui (1) si � 1 if s*i � 0 and si � 0 if s*i � 0 stock allocation: yi � �0 � �1fli � �2fni � �3�flifni� � �4wi � �i (2) we use the heckman two-stage selection model because the two-stage selection model would overcome a potential selection bias problem that the error term (�i) in eq. (2) can be correlated with the error term (ui) in eq. (1). in the selection equation, the dependent variable (si) is a dummy, which indicates whether individual i chooses the default option—that is, a money market fund—or not. if a respondent selects “no default,” the variable takes a value of 1 and otherwise 0. in the stock allocation equation, the dependent variable (yi) indicates individual i’s contribution allocation to stock funds as a percentage. stock allocation (yi) is calculated by summing respondent i’s contribution allocations to domestic and international stock funds as a percentage. as key independent variables, the model includes fli and fni, which indicate individual i’s financial literacy and financial network intensity, respectively, and their interaction term (flifni). zi and wi indicate a set of control variables for the selection equation and the stock allocation equation, respectively. 3.1.3. financial literacy to measure an individual’s financial literacy, first, we assume that financial literacy can be captured by a level of financial knowledge, following prior studies on financial literacy (e.g., alhenawi and elkhal, 2013; collins, 2012; hilgert et al., 2003; lusardi and mitchell, table 1 summary fund information provided for survey participants fund types 1. money market fund (designated as a default investment option) 2. small-cap growth fund 3. mid-cap blend fund 4. large-cap value fund 5. large-cap growth fund 6. foreign small/mid-cap blend fund 7. foreign large-cap value fund 8. short-term bond fund 9. intermediate-term bond fund 10. inflation-protected bond fund fund information 1. investment objective 2. principal investment strategies 3. principal investment risks 4. annual expense ratio 5. return and risk a. three-year average return, sd, and sharpe ratio b. five-year average return, sd, and sharpe ratio 6. 10-year total annual returns (a bar graph including data labels) notes: the information for ten funds was provided in a separate brochure so that survey participants could compare investment objectives, strategies, expenses, risk and return, and the performance of the funds. the fund information was provided using actual funds in the marketplace, but actual fund names including fund family names were not disclosed in the brochure to prevent respondents from selecting a fund due to family names. 84 y. chung, y. park / financial services review 24 (2015) 77–99 2007a). next, to evaluate respondents’ financial knowledge, we use a battery of five questions from the national financial capability study supported by the finra investor education foundation. these questions cover fundamental concepts of economics and finance, such as calculations about interest rates and inflation, the relationship between interest rates and bond prices, the relationship between interest payments and maturity in mortgages, and risk diversification.10 a higher score indicates a higher level of financial literacy. 3.1.4. financial network intensity following chung and jackson (2013) and reinholt, pedersen, and foss (2011), we use an egocentric network technique to assess respondents’ financial network intensity. respondents list the first and last name initials of up to 10 people (including spouse/partner, parents, siblings, relatives, friends, co-workers, financial planners or advisors, and others) who they believe are most important sources to their financial or investment decisions. limiting the list of possible contacts to 10 people may not allow respondents to describe their entire financial networks, but constraining the number of contacts listed has the benefit of making data collection more feasible (morrison, 2002). for each person in their network, respondents provide scores to answer two questions: communication frequency and a level of financial knowledge (see table 2). using the responses to the two questions, we yield an individual’s financial network intensity measured by communication frequency with the financially literate in his or her financial networks. we define individual i’s financial network intensity (fni) as follows: fni � �j�1 j �cfreqij � nfkij� maxzi for j � 1, 2, ..., j � � 10� (3) table 2 summary of questions for financial networks question 1. during the five years, how often have you talked to each person to acquire financial and/or investment information? 0 � never 1 � once a year or less 2 � several times a year 3 � once a month 4 � 2–3 times a month 5 � once a week or more question 2. indicate to what extent each person has financial knowledge. 1 � very little 2 � below average 3 � average 4 � above average 5 � very high notes: survey respondents were asked to list the first and last name initials of up to 10 people (including spouse/partner, parents, siblings, relatives, friends, coworkers, financial planners or advisors, and others) who they believe are most important to their financial or investment decisions. for each person respondents listed, they were also asked to answer the above two questions about communication frequency and financial literacy. 85y. chung, y. park / financial services review 24 (2015) 77–99 first, individual i identifies persons (j) in his or her financial network (called “alters”), but the number of alters is constrained up to 10 (j �10). j indicates the total number of persons in individual i’s financial networks. second, individual i evaluates communication frequency with a person j in his or her network (cfreqij), using a scale of 0 (never) to 5 (once a week or more) (table 2: question 1). we assume that a large value of communication frequency indicates a strong relationship with a person in one’s network. third, individual i evaluates the financial literacy of person j, using a scale of 0 (very little) to 5 (very high) (table 2: question 2). the responses are converted to a binary variable (nfkij), which indicates whether person j in individual i’s network is financially literate or not. nfkij takes a value of one for a response of “above average” or “very high” and a value of zero for a response of “very little,” “below average,” or “average.” fourth, because nfkij is zero for person j’s average or lower financial literacy in individual i’s network, the numerator in the eq. (3) is computed by summing up individual i’s communication frequencies with only the financially literate in his or her network. finally, individual i’s network strength with the financially literate is evaluated with respect to the maximum possible value (maxzi) that the individual can have from the relationships with the financially literate. maxzi is defined as a product of j, the maximum possible value of cfreqij (� 5), and the maximum possible value of nfkij (� 1) for individual i. for example, when individual i has five persons in his or her financial networks, his or her maximum possible value (maxzi) is 25 (� 5 � 5 � 1). thus, individual i’s financial network intensity (fni) takes a value between 0 and 1; a larger value of fni indicates greater network strength with the financially literate. 3.1.5. control variables respondents’ network size, investment experience, risk tolerance, grade point average (gpa), current mood state, and demographic backgrounds are included as control variables. first, following hansen et al. (2001), we define network size as the total number of persons from whom a respondent acquires financial or investment information. those who have a larger size of network may have more chances to obtain a broader spectrum of information and knowledge (e.g., chung and jackson, 2013; gargiulo, ertug, and galunic, 2009). second, we include respondents’ risk tolerance, following prior studies on financial risk tolerance and investment behavior (e.g., gibson, michayluk, and venter, 2013; guiso et al., 2008; hong et al., 2004; van rooij et al., 2011; yoong, 2010). respondents’ decisions on a choice of a default option and stock allocation may be affected by their risk tolerance. to measure risk tolerance, we use a battery of 13 items developed by grable and lytton (1999).11 third, we control for investment experience. we expect that a respondent’s prior investment experience may influence his or her decisions on a choice of a default option and stock allocation. thus, survey participants were asked to answer whether they have owned any stocks or mutual funds during the past five years. when a respondent answers “yes” in either stocks or mutual funds, the respondent is regarded as one who has investment experience. fourth, we control for the current mood state by including the 20-item positive affect and negative affect scale (panas; watson et al., 1988), relying on the findings of the literature on personality and social psychology that risk perceptions and associated choices are 86 y. chung, y. park / financial services review 24 (2015) 77–99 significantly influenced by the mood state (or feelings) at the moment of decision making (e.g., forgas, 1995; isen, 2000; leith and baumeister, 1996; loewenstein et al., 2001). fifth, we control for respondents’ gpa. individuals who have high academic performance are likely to have a stronger belief in their ability than those who have low academic performance (bandura, 1993, 1997; becker and gable, 2009). thus, we expect that respondents having a higher gpa are less likely to choose a default investment option. respondents’ gpa, however, is excluded from the stock allocation equation because there is no previous evidence on the relationship between gpa and the propensity to invest in stocks. furthermore, the exclusion of gpa from the stock allocation equation avoids any problems for identification with the heckman two-stage selection model.12 thus, we assume that respondents’ gpa has no direct effect on stock allocation. finally, we control for individual demographic characteristics: gender (female � 1), age, and ethnicity (white � 1, otherwise � 0). 4. results 4.1. descriptive statistics table 3 presents descriptive statistics of the sample. the mean score of respondents’ financial literacy is 4.1, which is greater than the 3.4 that lusardi (2011) reported for college graduates or higher in the national survey sample of 1,488 adults in 2009. the higher score of financial literacy may result from our sample characteristics that all respondents are business students. for financial networks, the respondents in the sample identify, on average, 4–5 people from whom they have obtained financial or investment information. the average financial network intensity is 0.28. this indicates that, for example, when a respondent has five persons in his or her financial network who all are financially literate, he or she has contacted one person at the frequency of “once a month” but the other four persons at the frequency of “once a year or less.” for risk tolerance, our respondents have an average score of 26.49, which is close to the average scores of 28.83 and 27.03 reported by grable and lytton (2003) and grable, lytton, and o’neill (2004), respectively. about 39% of our respondents have investment experience. female respondents consist of 41% of the sample. respondents’ age ranges from 20 to 26 with a mean age of 22.1. about 87% of the respondents are white. panel b of the table reports correlation coefficients between respondents’ selection of “no default,” stock allocation, financial literacy, and financial network variables. in particular, the variables of financial literacy, financial network size, and financial network intensity are not significantly correlated with each other. hence, an inclusion of these three variables in the same regression equation would not incur significant multi-collinearity problems among the variables. 4.2. effects of financial literacy and financial network intensity on retirement investments to examine the effects of financial literacy and financial networks on respondents’ investment decisions for retirement, we use the heckman two-stage selection model de87y. chung, y. park / financial services review 24 (2015) 77–99 scribed in eqs. (1) and (2). column 1 of table 4 presents regression results of the two-stage selection model: column 1a presents probit regression results for decisions on the default option (the first stage), whereas column 1b presents ordinary least square regression results for stock allocation (the second stage). about two-thirds of the 97 respondents select “no default,” and they build their own retirement portfolios. since we ask respondents to build their retirement investment portfolios only when they answer “no default” for the default investment option (a money market fund), the two-stage selection model would overcome a potential selection bias problem that the error term (�i) in the stock allocation equation is correlated with the error term (ui) in the selection equation. first, respondents’ financial literacy is significantly positively related to a choice of “no default” (column 1a) and stock allocation (column 1b). for example, if a respondent’s financial literacy score increases from 4 to 5 (the largest score), the probability of selecting “no default” increases by 0.189 (from 0.708 to 0.897) and stock allocation increases by 12.7 percentage points (from 48.4 to 61.0%), with the other variables held at their mean values. the results support hypotheses h1a (a positive relationship between financial literacy and a table 3 sample descriptive statistics and correlation panel a: descriptive statistics variables number of obs. min max median mean sd financial literacy (score) 97 2 5 4 4.09 0.84 financial network: size 97 1 10 4 4.60 2.46 intensity 97 0 0.92 0.25 0.28 0.22 risk tolerance (score) 97 18 37 26 26.49 4.20 investment experience (yes � 1) 97 0 1 0 0.39 0.49 gender (female � 1) 97 0 1 0 0.41 0.49 age 97 20 26 22 22.06 1.20 ethnicity (white � 1) 97 0 1 1 0.87 0.34 gpa 97 1.90 4.00 3.11 3.14 0.41 current mood state: positive (score) 97 17 46 37 37.24 5.22 negative (score) 97 10 33 18 18.56 5.02 selection of a default option (“no default” � 1) 97 0 1 1 0.66 0.48 contribution allocation to stock funds (for only those selecting “no default”) 64 0 1.00 0.50 0.52 0.20 panel b: correlations of selection of no default, stock allocation, financial literacy, and financial networks 1 2 3 4 5 1. selection of “no default” 1.000 2. stock allocation (for only those selecting “no default”) — 1.000 3. financial literacy 0.365** 0.390** 1.000 4. financial network size 0.096 �0.162 �0.113 1.000 5. financial network intensity 0.128 0.399** 0.120 �0.084 1.000 **p � 0.01. 88 y. chung, y. park / financial services review 24 (2015) 77–99 choice of “no default”) and h1b (a positive relationship between financial literacy and stock allocation). second, financial network intensity is not significantly related to a selection of “no default” (column 1a), but it is significantly positively related to stock allocation (column 1b). for example, a respondent’s stock allocation increases by 2.7 percentage points as his or her network intensity increases by 10%. the results do not support hypothesis h2a (a positive relationship between financial network intensity and a choice of “no default”) but do support h2b (a positive relationship between financial network intensity and stock allocation). table 4 regression results: default investment option and contribution allocation to stock funds dependent variables (1) (2) (1a) (1b) (2a) (2b) selection of no default (first stage) contribution allocation to stock funds (second stage) selection of no default (first stage) contribution allocation to stock funds (second stage) financial literacy (score) 0.715** 0.127** 0.768** 0.111** (0.222) (0.035) (0.235) (0.035) financial network size 0.104 0.006 0.109 0.003 (0.069) (0.009) (0.070) (0.008) financial network intensity 0.794 0.268** 0.650 0.184* (0.751) (0.091) (0.744) (0.087) (financial literacy) � (financial network intensity) �0.734 0.269* (0.699) (0.107) risk tolerance (score) �0.033 0.017** �0.034 0.016** (0.042) (0.005) (0.042) (0.005) investment experience (yes � 1) 0.938* 0.088 0.962* 0.085 (0.385) (0.049) (0.387) (0.046) gender (female � 1) �0.051 0.048 �0.027 0.036 (0.345) (0.041) (0.350) (0.039) ethnicity (white � 1) 0.011 0.162** 0.110 0.121* (0.456) (0.061) (0.470) (0.059) current mood state positive (score) 0.011 0.012* 0.007 0.013** (0.032) (0.004) (0.032) (0.004) negative (score) �0.077* �0.001 �0.079* �0.001 (0.037) (0.005) (0.037) (0.005) gpa 1.065** 1.040* (0.408) (0.412) constant �4.839* �1.281** �1.511 �0.675** (3.389) (0.290) (1.990) (0.228) �: mill’s ratio 0.149 0.137 [p-value] [0.065] [0.071] observations 97 64 97 64 pseudo r2 0.305 0.313 adjusted r2 0.514 0.567 notes: when an interaction term of financial literacy and network intensity is included in column 2, the variables of financial literacy and network intensity are centered with respect to the mean. this centering, however, does not affect the coefficients and their significance. standard errors are in parentheses. **p � 0.01; *p � 0.05. 89y. chung, y. park / financial services review 24 (2015) 77–99 accordingly, the results reported in columns 1a and 1b suggest that when making a decision on whether or not to choose a default option, respondents are likely to rely on their financial literacy but not on their financial networks. however, when constructing a retirement portfolio such as determining stock allocations, respondents are likely to draw on both their financial literacy and financial networks. hence, financial literacy and financial network intensity were positively associated with stock allocation. in addition, respondents’ investment experience, current mood state, and gpa are significantly related to their selection on the default investment option (column 1a) or stock allocation (column 1b). first, when a respondent has any investment experience during the past five years, he or she is likely to select “no default” but not likely to significantly increase stock allocation. second, different mood states involve different investment tasks. a negative (or unpleasant) mood state is significantly related to a choice of the default investment option whereas a positive (or pleasant) mood state is not. this may be because a negative mood state inhibits people from making rational decision-making because of excessive stress (leith and baumeister, 1996). in addition, a positive (pleasant) mood state is likely to increase stock allocation in a retirement portfolio, but a negative mood state is not. because people who have a positive mood state tend to underestimate the possibility of loss, they are more willing to take a risk (au, chan, wang, & vertinsky, 2003; seo, goldfarb, & barret, 2010). respondents’ gpa is positively related to a choice of “no default,” as expected. 4.3. interaction effects of financial literacy and financial network intensity in addition to the main effects of financial literacy and financial network intensity on investment decisions for retirement, we hypothesized that the two variables interact with each other. column 2 of table 4 presents regression results of the heckman two-stage selection model that includes an interaction term of financial literacy and financial network intensity. the interaction term is not significantly related to respondents’ selection of the default option (column 2a), not supporting hypothesis h3a. the interaction term, however, is significantly positively related to stock allocation (column 2b), supporting hypothesis h3b. the positive sign of the interaction term indicates a complementary effect between financial literacy and financial network intensity on stock allocation in a retirement portfolio. to further examine the interaction effect, following aiken and west (1991), we depict respondents’ contribution allocation to stock funds with respect to two different levels of financial literacy and financial network intensity. a high or low level is defined by 1 sd above or below the mean, respectively. fig. 1 shows that the relationship between financial literacy and stock allocation is positive and significant in a high level of financial network intensity (with a slope of 0.171, p-value � 0.001), but this relationship is not significant in a low level of financial network intensity (with a slope of 0.050, p-value � 0.300). the results suggest that the positive effects of financial literacy on stock allocation documented in the literature (e.g., van rooij et al., 2011; yoong, 2010) are limited to those who have a high level of financial network intensity (i.e., high network strength with the financially literate). 90 y. chung, y. park / financial services review 24 (2015) 77–99 4.4. effects of financial literacy and financial network intensity on stock allocation: deviation from age-appropriate stock allocations among those who select “no default” and build their own retirement portfolio with the 10 mutual funds offered, some respondents may not construct an appropriate portfolio—for example, too little or too much stock allocation. in this section, we attempt to see whether respondents construct an appropriate retirement investment portfolio based on their age and risk tolerance level, focusing on stock allocation. for doing this, we use as a benchmark the 2012 morningstar lifetime allocation indexes—u.s. investors (morningstar, 2012). the morningstar indexes provide the information of asset allocation over lifetime for three different risk profiles: aggressive, moderate, and conservative. we assume that the asset allocations of the morningstar indexes represent age-appropriate ones for retirement. to calculate any deviation of respondents’ stock allocation from the stock allocations of the morningstar indexes, we use the following formula: devi � | yi � indexk| (4) where the variable yi indicates individual i’s contribution allocation to stock funds as a percentage and where indexk indicates stock allocations of the morningstar indexes for k � aggressive, moderate, or conservative. a k is determined according to a respondent’s risk tolerance level: for example, if a respondent’s risk tolerance level is low, a conservative allocation of the morningstar indexes is regarded as his or her appropriate asset allocation. a respondent’s risk tolerance is categorized into three levels based on his or her risk tolerance score. a low level of risk tolerance is defined with scores of less than 24, a medium fig. 1. interaction effects of financial literacy and network intensity on contribution allocation to stock funds. notes: this figure depicts the interaction effects based on the regression results in column 2b of table 4. 91y. chung, y. park / financial services review 24 (2015) 77–99 level with scores of 24–32, and a high level with scores of more than 32. the threshold scores of 24 and 32 are derived from the findings of grable and lytton (2003) and grable et al. (2004).13 those who selected “no default” have constructed their retirement portfolios with a deviation of 7.2–80.9 percentage points from the age-appropriate stock allocations. the mean deviation is 35.8 percentage points (panel a of table 5). to examine how the deviation is related to respondents’ financial literacy and financial network intensity, we estimate a heckman two-stage selection model that includes the deviation from the age-appropriate stock allocations as a dependent variable in the second stage. the model also includes an interaction term of financial literacy and financial network intensity. panel b of table 5 presents the regression results for the deviation from the ageappropriate stock allocations based on respondents’ age and risk tolerance level. column 1b shows that the deviation is significantly negatively related to financial literacy and financial network intensity. in addition, the interaction term of financial literacy and financial network intensity is significantly negatively related to the deviation. the negative sign of the interaction term indicates a complementary effect between financial literacy and financial network intensity on reducing the deviation from the benchmark stock allocations. to further examine the interaction effects, we depict the deviation from the age-appropriate stock allocation with respect to two different levels of financial literacy and financial network intensity. a high or low level is defined by 1 sd above or below the mean, respectively, as in fig. 1. fig. 2 shows that the effect of financial literacy on the deviation is significant in the high financial network intensity (with a slope of �0.156, p-value � 0.001), but is not significant in the low financial network intensity (with a slope of �0.056, p-value � 0.212). the results indicate that a reduction of the deviation from the ageappropriate stock allocation by enhancing financial literacy would be effective only among those who have a high level of financial network intensity (i.e., strong relationships with the financially literate). 5. conclusion this article examines the role of financial literacy and financial network intensity in retirement investment decisions of generation yers. with a complete sample of 97 senior business college students, aged 20–26 in 2012, we find first that when a money market fund is designated as a default investment option, respondents who have high financial literacy are more likely to select “no default,” but their financial networks are not significantly related to their choice. second, respondents who have high levels of financial literacy or financial network intensity are likely to allocate more contributions to stock. third, financial literacy and financial network intensity significantly interact on stock allocation in a retirement portfolio. specifically, the positive effect of financial literacy on stock allocation is significant in the high financial network intensity, but the financial literacy effect is not significant in the low financial network intensity. last, when respondents construct their retirement portfolios, stock allocations are largely deviated from the age-appropriate stock allocations given their age and risk tolerance level. the deviations, however, are significantly reduced 92 y. chung, y. park / financial services review 24 (2015) 77–99 when financially knowledgeable respondents have strong relationships with the financially literate. our findings contribute to an understanding of how financial literacy and financial networks are jointly associated with individual investors’ decision-making. although most previous studies argue that financial literacy is a key determinant of individual investors’ decision-making, our study shows that both financial literacy and financial network intensity table 5 contribution allocation to stock funds: deviation from age-appropriate allocations panel a: descriptive statistics obs. min max median mean sd deviation from age-appropriate stock allocations 64 0.072 0.809 0.383 0.358 0.174 panel b: regression results dependent variables (1-a) (1-b) selection of no default (first stage) deviation of stock allocation (second stage) financial literacy (score) 0.768** �0.106** (0.235) (0.032) financial network size 0.109 �0.001 (0.070) (0.008) financial network intensity 0.650 �0.169* (0.744) (0.080) (financial literacy) � (financial network intensity) �0.734 �0.223* (0.699) (0.097) risk tolerance (score) �0.034 �0.009 (0.042) (0.005) investment experience (yes � 1) 0.962* �0.077 (0.387) (0.042) gender (female � 1) �0.027 �0.037 (0.350) (0.036) ethnicity (white � 1) 0.110 �0.103 (0.470) (0.054) current mood state: positive (score) 0.007 �0.013** (0.032) (0.004) negative (score) �0.079* �0.001 (0.037) (0.005) gpa 1.040* (0.412) constant �1.511 1.345** (1.990) (0.209) �: mill’s ratio �0.135 [p-value] [0.051] observations 97 64 pseudo r2 0.313 adjusted r2 0.513 notes: the variables of financial literacy and network intensity are centered with respect to the mean. this centering, however, does not affect the coefficients and their significance. standard errors are in parentheses. **p � 0.01; *p � 0.05. 93y. chung, y. park / financial services review 24 (2015) 77–99 are critical factors influencing individuals’ investment decision-making. in particular, the significant interaction between financial literacy and financial network intensity suggests that financial literacy and financial networks are complementary to each other. this finding implies that an individual’s retirement investment decisions can be significantly enhanced when his or her financial knowledge is synthesized with the knowledge acquired through social networks with the financially literate. this result also provides an insight into employer-sponsored financial education. for retirement investments, individual investors may not always use the information provided by employers, but rather rely on their social networks (e.g., duflo & saez, 2002, 2003). thus, for those who are likely to lack social networks with the financially literate, employers need to provide more opportunities to build networks for financial advice with financial planners or advisors because financial planners or advisors may be able to be substituted for their financial networks. although this study contributes to the literature that pertains to financial literacy, social interactions, and investment behavior in retirement plans, it has several limitations. first, this study does not use actual contribution allocations of defined contribution plan participants. thus, the results from our sample may not represent actual investing behavior of defined contribution plan participants. second, because of the small sample size, the results in the study may not represent the investing behavior of generation yers as a whole. third, because the sample consists of college students who tend to have a higher level of risk tolerance (e.g., gilliam, chatterjee, & grable, 2010) and have less investing experience than the general investing population, the results may be biased towards allocating more contributions to risky assets. despite these limitations, our study using an experimental survey design (i.e., fig. 2. interaction effects of financial literacy and network intensity on the deviation from age-appropriate stock allocations. notes: this figure depicts the interaction effects based on the regression results in column 1b, panel b of table 5. 94 y. chung, y. park / financial services review 24 (2015) 77–99 providing respondents a scenario of retirement investments and having them make investment decisions) provides an implication for generation yers’ retirement investments: financial literacy combined with financial networks could increase the effectiveness of any efforts for enhancing their investment decisions. this could also be applied to improving long-term financial planning such as budgeting and saving. therefore, future research is called for to collect field data from those who engage in the first-time retirement investment or long-term financial planning and investigate how their financial literacy and financial networks jointly influence their decision-making quality. notes 1 several recent articles provide excellent literature reviews on the effects of financial literacy and financial education on financial outcomes (e. g., collins & o’rourke, 2010; fernandes, lynch, & netemeyer, 2014; hastings, madrian, & skimmyhorn, 2013). 2 prior studies on social interactions focused on social activities, such as visiting neighbors and attending church (hong et al., 2004), asking for advice about products and brands (brown et al., 2008), and participating in a sport or social club, or a political or community-related organization (christelis, jappelli, & padula, 2010). 3 in this article, respondents’ financial literacy is measured with their financial knowledge, as in most prior studies on financial literacy (e. g., alhenawi & elkhal, 2013; collins, 2012; lusardi & mitchell, 2007a; robb, babiarz, & woodyard, 2012; van rooij et al., 2011; yoong, 2010), while respondents’ financial network intensity is measured with their network strength with financially literate ones. we will discuss financial literacy and financial network intensity in more detail in section 3. 4 the complementary effects of financial literacy and financial network intensity on contribution allocation to stock funds are similar to the finding of collins (2012). he finds that individuals with a higher level of financial literacy are more likely to receive financial advice. 5 several studies on stock-market participation show that information and transaction costs deter people from participating in the stock market (e.g., haliassos & bertaut, 1995; vissing-jorgensen, 2003). 6 for example, vissing-jorgensen (2003) mentions that once people understand diversification, they can apply their insights to a larger portfolio without cost. 7 van rooij et al. (2011) use the information of respondents’ economics education in school as an instrument for financial literacy and find a positive relationship between financial literacy and stock-market participation. 8 guiso et al. (2008) explore a relationship between trust and stock market participation. using dutch and italian survey data, they find that less trusting individuals are less likely to be stockholders. 9 since this survey was conducted in a context of a 401(k) plan, the survey participants were informed before the survey that they could rebalance their retirement portfolio later at no cost. 95y. chung, y. park / financial services review 24 (2015) 77–99 10 for the exact wording of the five questions, see the survey questions posted to the website of the national financial capability study, http://www.usfinancialcapa bility.org/survey_data.html. 11 we use the items of grable and lytton (1999), instead of questions about hypothetical gambles over lifetime income in the health and retirement study (hrs), because most of our survey participants may not be familiar with the questions used in the hrs because of their little or limited full-time work experience. our respondents’ median (mean) full-time work experience is only 0.5 years (1.8 years). 12 we include gpa in the selection equation but not in the stock allocation equation. the variable gpa works as an exclusion restriction. the heckman two-stage selection model should have at least one exclusion restriction; otherwise, the second stage of heckman two-stage model is likely to suffer from a collinearity problem, which provides imprecise estimates as a result (wooldridge, 2002). 13 grable and lytton (2003) report an average score of 28.83 and a sd of 4.49 with a sample of 303 respondents, while grable et al. (2004) report an average score of 27.03 and a sd of 5.18 with a sample of 421 respondents. as a range of a medium level of risk tolerance, we use the common interval of �one sd from the two studies, which is a range of 24–32. acknowledgments the authors would like to thank michele o’neill and anonymous referees for their helpful comments. references agnew, j. r., & szykman, l. r. (2005). asset allocation and information overload: the influence of information display, asset choice, and investor experience. journal of behavioral finance, 6, 57–70. aiken, l. s., & west, s. g. (1991). multiple regression: testing and interpreting interactions. newbury park, ca: sage. alhenawi, y., & elkhal. (2013). financial literacy of u.s. households: knowledge vs. long-term financial planning. financial services review, 22, 211–244. au, k., chan, f., wang, d., & vertinsky, i. (2003). mood in foreign exchange trading: cognitive processes and performance. organizational behavior and human decision processes, 91, 322–338. balkundi, p., & harrison, d. a. (2006). ties, leaders, and time in teams: strong inference about network structure’s effects on team viability and performance. academy of management journal, 49, 49–68. bandura, a. (1993). perceived self-efficacy in cognitive development and functioning. educational psychologist, 28, 117–148. bandura, a. (1997). self-efficacy: the exercise of control. new york, ny: w. freeman. banerjee, a. (1992). a simple model of herd behavior. quarterly journal of economics, 107, 797–817. banerjee, s. (2011). how do financial literacy and financial behavior vary by state? employee benefit research institute note, 32, 9–13. becker, s. p., & gable, r. k. (2009). the relationship of self-efficacy with gpa, attendance, and college student retention. nera conference proceedings, paper 26. 96 y. chung, y. park / financial services review 24 (2015) 77–99 bettman, j. r., & park, c. w. (1980). effects of prior knowledge and experience and phase of the choice process on consumer decision processes: a protocol analysis. journal of consumer research, 7, 234–248. bikchandani, s., hirshleifer, d., & welch, i. (1992). a theory of fads, fashion, custom and cultural change as informational cascades. journal of political economy, 100, 992–1026. brown, j. r., ivković, z., smith, p. a., & weisbenner, s. j. (2008). neighbors matter: causal community effects and stock market participation. journal of finance, 63, 1509–1531. campbell, j. (2006). household finance. journal of finance, 61, 1553–1604. cascioaro, t., & lobo, m. s. (2008). when competence is irrelevant: the role in interpersonal affect in task-related ties. administrative science quarterly, 53, 655–684. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2004). for better or for worse: default effects and 401(k) savings behavior. in: d. a. wise (ed.), perspectives on the economics of aging (pp. 81–121). chicago, il: university of chicago press. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2006). saving for retirement on the path of least resistance. in: e. j. mccaffrey & j. slemrod (eds.), behavioral public finance: toward a new agenda (pp. 304–351). new york, ny: russell sage foundation. christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54, 18–38. chung, y., & jackson, s. e. (2011). coworker trust and knowledge creation: a multilevel analysis. journal of trust research, 1, 65–83. chung, y., & jackson, s. e. (2013). the internal and external networks of knowledge-intensive teams: the role of task routineness. journal of management, 39, 442–468. 4 collins, j. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. delavande, a., rohwedder, s., & willis, r. (2008). preparation for retirement, financial literacy and cognitive resources. michigan retirement research center working paper, wp2008-190. duflo, e., & saez, e. (2002). participation and investment decisions in a retirement plan: the influence of colleagues’ choices. journal of public economics, 85, 121–148. duflo, e., & saez, e. (2003). the role of information and social interactions in retirement plan decisions: evidence from a randomized experiment. quarterly journal of economics, 118, 815–842. ellison, g., & fudenberg, d. (1993). rules of thumb for social learning. journal of political economy, 101, 93–126. ellison, g., & fudenberg, d. (1995). word of mouth communication and social learning. quarterly journal of economics, 110, 93–125. fernandes, d., lynch, j. g., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors. management science, 60, 1861–1883. forgas, j. p. (1995). mood and judgment: the affect infusion model (aim). psychological bulletin, 117, 39–66. gargiulo, m., ertug, g., & galunic, c. (2009). the two faces of control: network closure and individual performance among knowledge workers. administrative science quarterly, 54, 299–333. gibson, r., michayluk, d., & venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. gilliam, j., chatterjee, s., & grable, j. (2010). measuring the perception of financial risk tolerance: a tale of two measures. journal of financial counseling and planning, 21, 30–43. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. grable, j. e., & lytton, r. h. (2003). the development of a risk assessment instrument: a follow-up study. financial services review, 12, 257–274. grable, j., lytton, r., & o’neill, b. (2004). projection bias and financial risk tolerance. journal of behavioral finance, 5, 142–147. guiso, l., sapienza, p., & zingales, l. (2008). trusting the stock market. journal of finance, 63, 2557–2600. haliassos, m., & bertaut, c. (1995). why do so few hold stocks? economic journal, 105, 1110–1129. hansen, m., podolny, j. m., & pfeffer, j. (2001). so many ties, so little time: a task contingency perspective on 97y. chung, y. park / financial services review 24 (2015) 77–99 the value of social capital in organisations. in: s. m. gabbay & r. th. a. j. leenders (eds.), research in the sociology of organizations (vol. 18, pp. 21–57). london: jai press. hansen, m. t., mors, m. l., & lovas, b. (2005). knowledge sharing in organizations: multiple networks, multiple phases. academy of management journal, 48, 776–793. hastings, j. s., madrian, b. c., & skimmyhorn, w. l. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5, 347–373. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 7, 309–322. hong, h., kubik, j. d., & stein, j. c. (2004). social interaction and stock-market participation. journal of finance, 59, 137–163. hung, a., parker, a. m., & yoong, j. (2009). defining and measuring financial literacy. rand labor and population working paper, no. 708. isen, a. m. (2000). positive affect and decision making. in: j. m. haviland (ed.), handbook of emotions (pp. 261–277). london: guilford press. ivković, z., & weisbenner, s. j. (2007). information diffusion effects in individual investors common stock purchases: covet thy neighbors investment choices. review of financial studies, 20, 1327–1357. kaustia, m., & knüpfer, s. (2012). peer performance and stock market entry. journal of financial economics, 104, 321–338. kiviat, r., & morduch, j. (2012). from financial literacy to financial action. working paper, mcgraw-hill research foundation. krackhardt, d. (1992). the strength of strong ties: the importance of philos in organizations. in: n. nohria & r. eccles (eds.), networks and organizations: structure, form, and action (pp. 216–239). boston, ma: harvard business school press. leith, k. p., & baumeister, r. f. (1996). why do bad moods increase self-defeating behavior? emotion, risk-taking, and self-regulation. journal of personality and social psychology, 71, 1250–1267. loewenstein, g., weber, e. u., hsee, c. k., & welch, n. (2001). risk as feelings. psychological bulletin, 127, 267–286. lu, t. (2011). social interaction effects and individual portfolio choice: evidence from 401(k) pension plan investors. pension research council working paper, prc wp2011-19. lusardi, a. (2011). americans’ financial capability. pension research council working paper, prc wp201102. lusardi, a., & mitchell, o. s. (2007a). financial literacy and retirement planning: new evidence from the rand american life panel. michigan retirement research center working paper, wp2007-157. lusardi, a., & mitchell, o. s. (2007b). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and over-indebtedness. cambridge, ma: national bureau of economic research, nber working paper 14808. madrian, b. c., & shea, d. f. (2001). the power of suggestion: inertia in 401(k) participation and saving behavior. quarterly journal of economics, 116, 1149–1187. mcfadyen, m. a., & cannella, a. a. (2004). social capital and knowledge creation: diminishing returns of the number and strength of exchange relationships. academy of management journal, 47, 735–746. moore, d. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences, pullman, wa: washington state university social and economic sciences research center, technical report 03-39. morningstar. (2012). morningstar lifetime allocation indexes—u. s. investors. chicago, il: morningstar, inc. morrison, e. w. (2002). newcomers’ relationship: the role of social network ties during socialization. academy of management journal, 45, 1149–1160. nahapiet, j., & ghoshal, s. (1998). social capital, intellectual capital, and the organizational advantage. academy of management review, 23, 242–266. perry, v. g., & morris, m. d. (2005). who is in control? the role of self-perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39, 299–313. 98 y. chung, y. park / financial services review 24 (2015) 77–99 reinholt, m., pedersen, t., & foss, n. j. (2011). why a central network position isn’t enough: the role of motivation and ability for knowledge sharing in employee networks. academy of management journal, 54, 1277–1297. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. social security and medicare boards of trustees. (2012). the 2012 annual report of the trustees of the federal old-age and survivor’s insurance and disability insurance trust funds. baltimore, md: social security administration. seo, m., goldfrab, b., & barrett, l. f. (2010). affect and the framing effect within individuals over time: risk taking in a dynamic investment simulation. academy of management journal, 53, 411–431. van rooij, lusardi, m., a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. van rooij, lusardi, m., a., & alessie, r. (2012). financial literacy, retirement planning and household wealth. economic journal, 122, 449–478. vissing-jørgensen, a. (2003). perspectives on behavioral finance: does “irrationality” disappear with wealth? evidence from expectations and actions. cambridge, ma: national bureau of economic research macroeconomics annual. wang, s., & noe, r. a. (2010). knowledge sharing: a review and directions for future research. human resource management review, 20, 115–131. watson, d., clark, l. a., & tellegen, a. (1988). development and validation of brief measures of positive and negative affect: the panas scales. journal of personality and social psychology, 54, 1063–1070. willis, l. (2011). the financial education fallacy. american economic review, 101, 429–434. wooldridge, j. m. (2002). econometric analysis of cross section and panel data. cambridge, ma: mit press. yli-renko, h., autio, e., & sapienza, h. j. (2001). social capital, knowledge acquisitions, and knowledge exploitation in young technology-based firms. strategic management journal, 22, 587–613. yoong, j. (2010). financial illiteracy and stock market participation: evidence from the rand american life panel. pension research council working paper, prc wp2010-29. 99y. chung, y. park / financial services review 24 (2015) 77–99 credit usage, payment behavior, and the accuracy of consumer credit files l. douglas smith, ph.d.a,*, michael staten, ph.d.b, thomas eyssell, ph.d.c, maureen karig, mbad, jeffrey feinstein, ph.d.e, cathleen johnson, ph.d.f acenter for business and industrial studies, university of missouri-st. louis, one university blvd, st. louis, mo 63121, usa bcollege of agriculture and life sciences, university of arizona, forbes hall room 211, tucson, az 85721, usa cdepartment of finance, university of missouri-st. louis, one university blvd, st. louis, mo 63121, usa ddepartment of supply chain and analytics, university of missouri-st. louis, one university blvd, st. louis, mo 63121, usa edepartment of global analytic strategy, lexisnexis risk solutions, 1000 alderman drive, alpharetta, ga 30005, usa fdepartment of philosophy, university of arizona, 1145 e. south campus drive, tucson, az 85721, usa abstract through intensive interviews, examination of credit reports, and rescoring of corrected credit files, the researchers consider household characteristics, major life events, financial resources, and payment habits as they study the integrity of credit-bureau data, vulnerability to error, and results of disputes filed with the major credit bureaus. credit usage and management are found to vary widely within demographic groups. vulnerability to error and outcomes of disputes depend primarily on the credit record itself. consumers with moderate credit scores are more likely than those with very high or low scores to see significant improvement in their records when errors are corrected. © 2018 academy of financial services. all rights reserved. jel classification: d1; d3; g2 keywords: credit; credit scores; household behavior; individual financial management * corresponding author. tel.: �1-314-516-6108; fax: �1-314-516-7201. e-mail address: nldsmith@umsl.edu (l.d. smith) financial services review 27 (2018) 1-28 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. 1. introduction the credit-driven consumer economy in the united states is supported by a reporting and regulatory system that involves multiple parties. in addition to the seller and the buyer of property, goods, or services, third parties are typically engaged in financing the transaction, servicing the loan, collecting debt, and disseminating related information. to support this system, the three major credit bureaus in the united states (equifax, experian, and transunion) maintain credit histories on approximately 200 million consumers and process nearly two billion items of information each month (avery, calem, canner, and bostic, 2003) from data furnishers (credit card companies, mortgage servicers, debt collectors, etc.). the result is a set of extremely comprehensive files that reveal the ways that u.s. consumers use credit and meet their financial obligations. credit scores based on this information are used to facilitate decisions in lending and other financial transactions. accurate credit-bureau data can help in properly assessing credit risk and commensurately pricing credit, but individuals with errors in their credit files can suffer unreasonable denials of credit, inflated borrowing costs, higher cost for insurance, inability to rent a home or even denial of employment. in short, credit-bureau records provide what appears to be microscopic historic detail on individuals’ use and management of credit. credit scores determined from the credit-bureau data can greatly affect individuals’ future economic and personal experience. systems for collecting, maintaining, verifying, correcting, and disseminating data are, therefore, subject to constant scrutiny. the fair credit reporting act (fcra 1970) established regulatory guidelines and a legal framework for the credit-reporting industry in the united states. it was amended in 1996 to include processes designed to improve accuracy of data and to designate responsibilities of the credit bureaus and data furnishers when confronted with consumer claims that their data are inaccurate. in 2003, the fair and accurate credit transactions (“fact”; fair and accurate credit transactions act, 2003) act imposed further reinvestigation duties on data furnishers.1 staten and cate (2004) describe the fcra and its subsequent amendments as taking the “remedial approach” to regulation, whereby the regulating authority “. . . designates the consumer as the ‘quality-control’ inspector . . . and places the responsibility for monitoring file accuracy on the party who can determine accuracy at the lowest cost.” unfortunately, it is not evident that most consumers review their credit files and take actions to get errors corrected. lyons, rachlis, and scherpf (2007) report that many consumers “. . . still lack specific knowledge about what information is contained in credit reports, how to dispute errors, and the possible impact of their credit history on such factors as insurance premiums and employment.” in 2012, a national consumer poll for the national foundation for credit counseling (nfcc 2012) found that just 38% of respondents had reviewed their credit reports within the previous 12 months despite the fact that the federal trade commission (ftc) publicizes the annual availability of free credit reports from the web site https://www.annualcreditreport. com. an understanding of consumer use of credit and the effects of industry and regulatory practice on individual consumers is important, as imposition of regulations to protect consumers can be very costly and have disproportionate effects on small financial institutions (elliehausen and lowrey, 2000). 2 l.d. smith et al. / financial services review 27 (2018) 1-28 consumer advocates have raised concerns that some sectors of society may be disadvantaged in the gathering and reporting of consumer credit information. from a survey of 154 adults, the united states public interest research group (uspirg) concluded that “79% of the credit reports surveyed contained either serious errors or other mistakes” and one-fourth of the reports “contained serious errors that could result in the denial of credit” (national association of state pirgs, 2004). in testimony before congress, a consumer advocacy group asserted that inaccuracies in credit reports could cause at least eight million americans to be improperly categorized as subprime risks, and pay tens of thousands of dollars in excess interest payments over the term of a 30-year mortgage loan (brobeck, 2003). unfortunately, these studies were based on very limited and possibly biased samples and were prone to other methodological flaws. avery, brevoort, and canner (2012) matched elements in credit reports with demographic data for over 200,000 consumers and examined the extent to which they could inadvertently serve as surrogates for demographic characteristics (e.g., race and ethnicity) that are deliberately not used in constructing credit scores, for fear of disparate impact of the creditscoring system on segments of society. they concluded that their research provided “little or no evidence that the credit characteristics used in credit history scoring models operate as proxies for race or ethnicity.” they noted, however, an inevitable relationship between average age of accounts and age of the consumer; thus, providing some unavoidable disparate impact for young consumers. their work does not directly address related questions about the propensity of individuals with different demographic characteristics and household situations to utilize credit and manage it well. in the most comprehensive study of credit-bureau accuracy to date, investigators from the university of missouri-st louis and the university of arizona, on behalf of the federal trade commission (ftc), engaged a representative sample of 1,001 u.s. consumers in a detailed review of their credit reports from the three major u.s. credit bureaus (smith et al., 2013). twenty-six percent of participants in this study claimed to find at least one potentially material error in a credit report and, with guidance from the university research associates who helped them review the credit reports, filed formal disputes with the relevant bureau(s). for 78% of the consumers who filed disputes (20% of participants overall), at least one bureau altered their credit report in a manner that addressed the consumers’ concerns. thirty-three percent of disputants (8.7% of all participants) experienced a resulting increase of 10� points in one or more of their fico scores; 21% of disputants (5.5% of study participants) had one or more scores cross a threshold that would typically result in more favorable terms of credit. the investigators interpreted these results as providing evidence that “current regulatory regimes and industry practice are geared to providing data that promote efficiency in our consumer economy while leaving a small percentage of individual consumers vulnerable to significant misrepresentations of their creditworthiness.” in this paper, we extend the research of smith et al. (2013), which concentrated on the incidence of errors in credit-bureau files and the results of disputes filed by consumers to rectify alleged errors. we examine further how different segments of society engage in the credit-driven consumer economy, manage their obligations, are vulnerable to reporting errors, and fare when they file formal disputes of information in their credit files. to be addressed are the following questions: 3l.d. smith et al. / financial services review 27 (2018) 1-28 1. how successfully do different consumer groups manage their credit obligations? 2. are some consumer segments more vulnerable to inaccuracies in their credit reports than others? 3. do some consumer groups improve their credit standing more than others when corrections are imposed on their files in response to formal disputes filed with the credit bureaus? the unique research design used in this study provides information that enables a coherent multivariate analysis of the full credit cycle. it allows the validation of self-reported data through a cross-check of information in each participant’s credit reports from the three major credit bureaus. impacts of errors in the bureaus’ files are assessed by an actual rescoring of frozen credit files by the leading provider of credit scores the analytical process is structured to reveal associations between outcomes of interest and the interrelated factors that explain them. 2. related research utilization of credit and management of personal or household finance have changed dramatically since the 1950s when a consumer might typically have carried a home mortgage, perhaps a car loan and used a couple of branded credit cards for gasoline service stations or major department stores. household debt increased from an average of 55% of personal disposable income in 1960 to 133% in 2007 (glick and lansing, 2009). home equity lines of credit, guaranteed student loans, bank credit cards, and loans from unconventional sources (e.g., payday loans) have entered the credit mix (bricker et al., 2014). in short, prudent management of credit has become a necessary skill in the modern economy. as such, questions about differences in credit usage across demographic categories, proneness to delinquency and default, problems of coping with debt, and psychological traits associated with credit use and financial stress all become increasingly important areas of study. we next summarize findings from such prior research and then present a graphical illustration of the dynamics of the consumer credit cycle to provide context for our research and analysis. 2.1. credit usage a substantial amount of empirical evidence suggests that individuals with different backgrounds and household situations vary significantly in their use of consumer credit. o’neill and xiao (2014) surveyed over 1,000 consumers to assess their financial sophistication and inquired about their attitudes and behavior regarding the use of credit. they found that older individuals, people with higher incomes and higher education levels who are married without children, white, and male managed credit better than individuals with lower incomes, less education, and those who were single parents, minorities, females, and young adults. javine (2013) surveyed 521 students and determined that key demographic factors predict the level of student-loan debt. african americans, first-generation college students, financially independent students, those in a later year in school, those with lower incomes, and students with lower gpas tend to have more student-loan debt. she observed that students 4 l.d. smith et al. / financial services review 27 (2018) 1-28 who were assessed as having greater financial knowledge were more likely to have student loans in excess of $10,000. smith, finke, and huston (2012) studied the increasing tendency of older adults to carry mortgage debt. they concluded that rather than being a prelude to financial disaster for these individuals and a result of a change in attitudes against savings and thrift, the steady increase in mortgage debt and housing leverage among americans close to retirement age is evidence of their greater financial literacy and sophistication in allocating resources and taking advantage of tax incentives. in 7,592 survey responses from the panel study of income dynamics between 1968 and 2003, grafova (2007) reported that noncollateralized debt obligations (from credit cards, student loans, medical or legal bills, loans from relatives, etc.) were more likely to be used by individuals with unhealthy lifestyles or conditions (smoking, obesity). while recognizing this relationship, the author judged that other factors (liquidity constraints, time preference, risk aversion, hyperbolic discounting, and less self-control) may be more directly related to use and management of credit. 2.2. proneness to delinquencies and defaults sullivan, warren, and westbrook (2006) examined bankruptcy filings over three decades and concluded that general financial distress is related to loss of income, more lenient lending practices, and decline in housing values (i.e., personal financial circumstances, macroeconomic factors, and business practices of financial institutions). ratcliffe et al. (2014) examined the nature and geographic distribution of current financial distress in the united states. using transunion credit reports from september 2013, they focused on individuals who were at least 30 days late on a nonmortgage payment or had debt in collections. they estimated that nearly 12 million american adults (5.3% with a credit file), have nonmortgage debt past due with an average past due balance of $2,258. further, 77 million american adults (35% of adults with a credit file), had some debt collection activity reported on their credit reports with an average “debt in collections or charged-off” amount owed of $5,178. both debt reported past due and debt collection activity were more concentrated in the south, where the aftermath of the housing boom and bust was the greatest. 2.3. psychological and behavioral linkages to use and management of credit individual psychology and attitudes are important determinants of credit utilization and management of debt. norvilitis and maclean (2010) link college students’ debt levels to the students’ general reluctance to delay gratification, financial education received from parents, and the students’ beliefs about whether parents would help out if they should become overextended. individual attitudes toward money, determinants of social standing, and disposition to financial risk have been related to personal financial decisions, spending patterns and management of credit (engelberg and sjoberg, 2007; tang, 1992). in surveys of 1,000 service personnel ready for deployment in 2010, bell et al. (2014) found less anxiety and a better sense of well-being among soldiers who had resources available for a financial emergency, higher levels of self-assessed net worth, and higher 5l.d. smith et al. / financial services review 27 (2018) 1-28 perceived financial knowledge. those with high credit-card debt and large amounts due on automobile loans felt more personal stress. archuleta, dale, and spann (2013) surveyed students who sought financial counseling and report that subjective self-assessments of financial status are more important than actual credit-card debt and student loans in predicting individual levels of financial anxiety. 2.4. personal bankruptcy how individuals cope with overwhelming debt is a topic in itself. fay, hurst, and white (2002) concluded that a personal bankruptcy decision is primarily related to the financial benefits versus the costs of the decision and also is more likely in jurisdictions where bankruptcies are more common. consistent with this, staten (1993) found that increased availability of credit following bankruptcy (because of easing of lending practices of financial institutions) appeared to reduce the deterrent for bankruptcy. himmelstein, warren, thorne, and woolhandler, (2005) found that unexpected medical bills (without insurance coverage) are significant precipitators of personal bankruptcy. 3. research framework and methodology the previous research suggests a complex interplay of institutional practice and individual behavior which affects personal and household use of credit, the management of debt, and maintenance of credit records that affect individuals’ financial opportunities. in fig. 1 we depict the main elements. the use of credit is seen as influenced by several forces: (1) family resources and needs, (2) personal attitudes and behavior about consumption, investment and debt, (3) education about costs of credit and maintenance of a good credit record, (4) business practices in the sales and marketing of products and services and competitive promotional efforts of financial institutions, (5) regulations intended to promote fair lending and so forth, fig. 1. dynamics of credit usage, debt management, and maintenance of individuals’ credit records. 6 l.d. smith et al. / financial services review 27 (2018) 1-28 and (6) economic factors including major family events such as unemployment, divorce, or illness. these same factors may influence the ways that individuals meet their obligations and build their credit records. institutional practice and regulatory constraints regarding the collection, matching, storage, and dissemination of relevant information in credit-bureau files affect the depiction of individuals’ creditworthiness and their vulnerability to errors. with each credit line and financial transaction, there is a very small risk of error in processing and reporting. a file with more credit lines and transactions (and especially with late payments and collection activity) is more prone to errors almost axiomatically because each line and transaction presents an opportunity for an error to occur. of course, accounts with indications of late payments or collection activity present highest risks of error because a financial institution may have credited the wrong account or there may have been a mismatch of records or negligence in reporting payments for collection activity. individuals’ vigilance and actions in identifying potential errors, and institutions’ investigations and actions in response to disputes, help to keep the files clean and determine the consumer’s continuing access to credit and the terms under which future credit is granted. to investigate the full credit cycle from the assumption of debt through payment of obligations and maintenance of the credit record, we engaged a representative sample of 1,001 individuals on behalf of the federal trade commission (ftc). recruitment of participants was accomplished in collaboration with senior economists at the ftc and the three major credit bureaus (equifax, experian, and transunion). a meticulous stratified random sampling procedure (smith et al., 2013) was used to achieve equal representation of individuals in credit-score quintile groups while also providing good representation geographically and by standard demographic attributes. overall, there was excellent representation in the final sample of completed interviews according to the primary criterion (credit score), good representation from all age groups, and an excellent mix according to gender. participants were recruited from each of the 50 u.s. states and the number from each state was generally proportional to the size of the adult population. university research associates engaged participants in thorough reviews of their credit reports and administered a telephone survey to obtain information about participants’ habits in managing credit, their family circumstances, and major life events. they inquired whether the study participants had previously obtained copies of their credit reports and attempted to have alleged errors corrected, whether they had used forms of credit not reflected in the credit reports, and whether such obligations had been discharged. the research associates also asked about monthly payment behavior (e.g., whether the consumer generally paid off monthly balances in full). perhaps surprisingly to some readers, this cannot be deduced from monthly balances of accounts in the credit-bureau files. one consumer may pay the minimum balance on an account on time, not incur further charges in the ensuing month, display no delinquency and yet be under considerable financial stress in doing so. another consumer may incur regular monthly charges, pay off the balance completely each month, and have an account that looks very much like the first consumer’s account despite having very large financial reserves. we expect this information (though admittedly self-reported) to be an important indicator of whether the consumer is under financial stress, vulnerable to delinquency and default, and then vulnerable to recording errors from collection activity. 7l.d. smith et al. / financial services review 27 (2018) 1-28 the checklist provided to consumers to prepare for the telephone interview, the interviewing guide for the intensive review of hard copies of the credit reports which had been mailed to them, and the closing survey after the telephone interview are provided in appendix b. note that, for context, we inquired about other financial assets (such as pensions) owned by the consumer, whether their families recently had suffered stressful events, and whether they had financial reserves to cover unexpected expenses or temporary loss of income. finally, we collected standard demographic information including income. we used income ranges to reduce resistance to disclosing income and recognizing that income estimates are always approximate regardless. only 45 of 1001 participants declined to indicate their family income category. implicitly, in our statistical models, these individuals are placed in the middle income category ($50,000 to $75,000 per year). if a participant alleged that there was a “material” error in any of the credit-bureau reports, the research associate helped craft a dispute letter which was signed by the participant and mailed to the bureau from their home address. we defined a material error as one that could affect a credit score at any of the three bureaus, suggest a mismatching of records from another individual, or suggest the possibility of identity theft (as when unknown accounts or inexplicable balances appeared in a credit report). we tracked the results of the disputes and determined exactly how any corrections to the disputed file would have affected the fico credit score. this was accomplished by having fico compute the credit scores for the three credit reports reviewed by a participant, freeze the files that were reviewed, draw new credit report(s) after an appropriate interval to determine the results of dispute(s) filed, impose corrections made in response to disputes on the original frozen file(s)and rescore the corrected file(s) using the same scoring algorithm. it should be noted that our methodology is unique. other published research has tended to use mathematical models which produced pseudo credit scores as an alternative to this labor-intensive procedure. to our knowledge, no other study of credit usage and credit management has involved an in-depth verification of data for each consumer by a thorough review of all three credit reports using a nationally representative sample. no other study has analyzed the entire credit cycle from credit usage through correction of errors in the credit file with complementary survey data giving detailed demographics and related information about family circumstances and life events. fig. 2 shows the flow of participants through the study process. in the sections to follow, we use multivariate analysis to study how credit utilization and payment history varied for all 1,001 participants and also to examine how the outcomes of disputes varied for 263 participants who allegedly discovered potentially material errors in one or more of their credit files. summary statistics for variables used in the analysis are provided in appendix a. 4. analysis we use a hierarchical set of multiple regression and logistic regression models to investigate portions of the dynamics represented in fig. 1. the models were structured with the following general forms: 8 l.d. smith et al. / financial services review 27 (2018) 1-28 fig. 2. flowchart of the study process. 9l.d. smith et al. / financial services review 27 (2018) 1-28 fig. 2. (continued). 10 l.d. smith et al. / financial services review 27 (2018) 1-28 y credit utilization (two measures) � f(demographic group, household characteristics, family resources, credit-management education, stressful events) using regression. y credit management (two measures) � f(credit utilization, demographic group, household characteristics, family resources, credit-management education, stressful events) using logistic regression. y credit record (one measure) � f(credit management, credit utilization, demographic group, household characteristics, family resources, credit-management education, stressful events, industry practice) using regression. y consumer action for remediation (one measure) � f(credit record, credit management, credit utilization, demographic group, household characteristics, family resources, creditmanagement education, stressful events, industry practice) using logistic regression. y remediation outcome (two measures) � f(consumer action for remediation, credit record, credit management, credit utilization, demographic group, household characteristics, family resources, credit-management education, stressful events, industry practice) using logistic regression. specific variables used to represent the various factors are enumerated in table 1. using this methodology, we are able to demonstrate the extent to which the main factors explain variation in the target variables—not just the effects of individual variables which compose the factors. we investigate whether blocks of explanatory variables for the major factors are significantly related to the outcome measures on their own and whether they contribute marginal information after accounting for the effects of the other factors. the statistical significance for blocks of variables in the regression models is determined by performing nested f tests for significant changes in residual squared error for the regression models. statistical significance of blocks of explanatory variables in the logistic models is similarly determined by performing �2 tests on the change in (�2 log(likelihood)) for the fits of the respective models when the block of variables is removed. table 1 variables used in multivariate models characteristic measurement variables credit utilization revolving credit utilization relative to credit limits; revolving credit relative to family income credit management pays more than minimum balance monthly; pays off all balances monthly credit record average credit score across three bureaus consumer action for remediation alleges material error in one or more reports and files dispute(s) with credit bureau(s) remediation outcome change in a credit score � 10 points; a score crosses a lending threshold generally used to set terms for credit demographic group age (whether under 30), gender, race/ethnicity (whether asian or african american); level of education (some college, whether holds graduate degree) household characteristics marital status (married), owns home, family size, number of children under 18, employed fulltime family resources family income (whether under $25k, between $25k and $50k, or over $75k), has contingency funds to cover two months without income; participates in retirement plan, has other retirement savings stressful events unemployment; drop in income; recent birth; divorce, separation or death of spouse; major medical expense 11l.d. smith et al. / financial services review 27 (2018) 1-28 after assessing the impact of the relevant factors on the predictive power of the models, we shall demonstrate the marginal effects of the individual explanatory variables upon the eight target (dependent) variables when all explanatory variables are included and then, recognizing collinearity, eliminate the most statistically insignificant terms, one at a time, until all remaining variables are statistically significant at the 0.05 level for a one-tailed test. the coefficients of the resulting “reduced” models are examined to indicate their (statistically significant) marginal effects on the target variables after the other explanatory variables in the models are considered. the first regression model analyzes credit utilization (total outstanding balance on all revolving accounts) as a percentage of aggregate credit limits on revolving accounts. if the credit limit on an individual revolving account (such as a credit card) is not reported, we substitute the “largest past balance” to impute the credit limit for that account.2 table 2 contains the results of significance testing for factors explaining this measure of credit utilization. in constructing the models for table 2, we excluded 16 extreme cases where the imputed revolving credit utilization exceeded 150%.3 ten different regression models, based on 872 cases with values for all constituent variables, are implicitly compared in table 2.4 the groups of variables for each of the factors (occurrence of an adverse event, demographic characteristics, family financial resources, and household characteristics) are all highly statistically related to credit utilization when considered individually (i.e., with p-values � 0.0001 when regressed against credit utilization as the only predictive factor). when the incremental informational content is tested by the removal of single factors from the full model, we see that household characteristics, family financial resources, and adverse events are all very statistically significant on the margin as well (with extremely small p-values for the blocks of variables). demographic variables are less statistically significant. next we consider revolving credit utilization as a percentage of annual household income. again we eliminated cases where the measure of credit utilization exceeded 150%. the numerator of this ratio is the total of outstanding balances on all revolving accounts. the denominator is the midpoint of the annual household income category that the participant identified as applicable to the household when completing the telephone interview (capped at $250,000). table 3 summarizes the results of the tests for this model. remarkable in table 3 is the much smaller values for the r2 statistic—with the full model explaining only 5% of the variation in credit utilization relative to household income. this is partly attributable to the wide bands for table 2 significance tests on regression models for revolving credit utilization relative to credit limits factors included r2 f value df p-value sample size full model 0.17 7.71 23 �0.0001 872 ex adverse events 0.16 3.58 5 0.003 872 ex demographic 0.16 2.26 6 0.036 872 ex resources 0.13 7.83 6 �0.0001 872 ex household char 0.14 5.86 6 �0.0001 872 only adverse events 0.04 8.11 5 �0.0001 872 only demographic 0.04 5.38 6 �0.0001 872 only resources 0.11 17.74 6 �0.0001 872 only household char 0.08 12.45 6 �0.0001 872 reduced model 0.16 12.81 13 �0.0001 872 12 l.d. smith et al. / financial services review 27 (2018) 1-28 reporting of household income but the models suggest that there is an enormous amount of variation in the extent to which individuals take on revolving debt relative to their total household income that is not explained by the factors we measured. adverse events and demographic variables accounted for the majority of the systematic variation in this measure. habits in managing credit were measured by a pair of binary variables determined by “yes-no” responses to (a) whether the participant always paid at least the minimum balance and (b) whether the participant always paid the entire balance on all credit obligations each month. table 4 provides the results of tests on the logistic model predicting whether a participant always pays at least the minimum amount due on all accounts each month. table 5 provides the results of tests on the logistic model predicting whether a participant always pays all outstanding balances on all accounts each month. the two indicators of individuals’ self-reported behavior in meeting credit obligations are highly related to each of the individual explanatory factors. on the margin, family resources and demographic characteristics contribute most information. next we consider the credit record as represented in the credit-bureau files. for this, we use the average of the three credit scores from the major bureaus as the consolidated indicator of the credit record as reflected in the participant’s credit reports. table 6 contains the results of significance tests for regressions of this measure against the explanatory factors used in previous models and with an additional factor representing the two self-reported habits on payment behavior. from the results in table 6, we see that all factors are highly statistically related to the average credit scores of participants when considered individually and on the margin. together the five factors explain 60% of variation in the average credit scores. table 3 significance tests on regression for revolving credit utilization relative to household income factors included r2 f value df p-value sample size full model 0.05 2.14 23 0.002 942 ex adverse events 0.04 2.04 5 0.071 942 ex demographic 0.03 2.87 6 0.009 942 ex resources 0.04 1.94 6 0.072 942 ex household char 0.04 1.24 6 0.282 942 only adverse events 0.02 2.85 5 0.015 942 only demographic 0.02 2.48 6 0.022 942 only resources 0.01 1.62 6 0.139 942 only household char 0.01 1.32 6 0.244 942 reduced model 0.04 5.96 6 �0.0001 942 table 4 significance tests on logistic models for whether participant pays at least minimum payment on all obligations each month factors included �2 df p-value sample size full model 256.7 23 �0.0001 997 ex adverse events 23.4 5 0.0003 997 ex demographic 36.9 6 �0.0001 997 ex resources 35.2 6 �0.0001 997 ex household char 14.8 6 0.022 997 reduced model 248.2 10 �0.0001 997 13l.d. smith et al. / financial services review 27 (2018) 1-28 we now turn attention to the likelihood that an individual finds a potentially material error in one or more of his or her credit files. continuing in our hierarchical analysis, we add the person’s credit score (strength of the credit record) as an explanatory factor at this stage. table 7 contains the results of statistical tests for this set of models. family resources, the credit record (average credit score) and general payment habits are the prime factors determining the likelihood that the individual will have a potentially material dispute with information in the credit file. this is perhaps not surprising, as individuals are more likely to question items that negatively affect the credit record. payment behavior also (almost axiomatically) affects the strength of the credit record and therefore the likelihood of negative information being present. we acknowledge that disputes in this study involve inaccurate information in the credit file (errors of commission rather than errors of omission) and especially information that may have an adverse impact on the credit score. unreported credit (such as loans from family members or loans from unconventional lenders who do not report to the bureaus) is ignored. finally, we examine the outcomes of formal disputes to see if a credit score increases by more than 10 points or whether a credit score for the individual crosses a lending threshold that would generally dictate the terms of an automobile loan. a summary of the outcomes is presented in table 8. for this analysis, we grouped the disputants according to the breakpoints for fico credit-score quintiles (where the quintiles were determined from the nationwide sample used to calibrate the fico scoring model). the greater representation in lower quintiles is because of the fact that individuals with low credit scores have more derogatory information in their credit files, and are, therefore, more likely to have something to dispute. table 5 significance tests on logistic models for whether participant pays total amounts due on all obligations each month factors included �2 df p-value sample size full model 293.8 23 �0.0001 997 ex adverse events 22.1 5 0.0005 997 ex demographic 50.2 6 �0.0001 997 ex resources 51.3 6 �0.0001 997 ex household char 32.5 6 �0.0001 997 reduced model 286.7 12 �0.0001 997 table 6 significance tests on regression models for the participant’s average credit score factors included r2 f value df p-value sample size full model 0.60 58.43 25 �0.0001 997 ex adverse events 0.58 10.06 5 �0.0001 997 ex demographic 0.58 9.21 6 �0.0001 997 ex resources 0.56 15.97 6 �0.0001 997 ex household char 0.58 8.67 6 �0.0001 997 ex payment habits 0.51 115.16 2 �0.0001 997 only adverse events 0.18 43.14 5 �0.0001 997 only demographic 0.24 51.31 6 �0.0001 997 only resources 0.22 48.73 6 �0.0001 997 only household char 0.23 48.73 6 �0.0001 997 only payment habits 0.42 364.58 2 �0.0001 997 reduced model 0.60 90.98 16 �0.0001 997 14 l.d. smith et al. / financial services review 27 (2018) 1-28 note that the relationship of outcomes with credit score is not linear. the probability of improvement is highest for individuals with moderate credit scores. many individuals with average fico scores in the top two quintiles already have three credit scores that qualify them for most favorable terms of borrowing; so corrections to their files did not significantly affect their credit standing. individuals in the lowest quintile tended to have so much derogatory information in the credit report that correcting errors did not generally improve the scores sufficiently to move them to a lower-risk group. individuals with average credit scores in the low-to-middle group who find errors in their credit reports have the most to gain immediately from seeking corrections to their files. a third of individuals in the lowermiddle quintiles may expect to have one or more of their credit scores cross a threshold that is used to determine the price of an automobile loan.5 significance tests for logistic models pertaining to the likelihood of having an increase of 10 or more points after dispute resolution are presented in table 9. significance tests for logistic models pertaining to the likelihood of crossing a lending threshold are presented in table 10. the logistic models are constructed with (0–1) indicator variables for the creditscore quintile to accommodate the nonlinear relationship. together, these results in tables 9 and 10 suggest that the outcomes of disputes filed with the credit bureaus to address potentially material errors in credit report(s) are not influenced table 7 significance tests on logistics models for the likelihood of having a dispute of potentially material information in one or more credit reports factors included �2 df p-value sample size full model 136.2 26 �0.0001 995 ex adverse events 5.2 5 0.39 995 ex demographic 10.3 6 0.11 995 ex resources 17.5 6 0.01 995 ex household char 2.3 6 0.89 995 ex payment habits 8.0 2 0.02 995 ex credit record 3.5 1 0.06 995 reduced model 121.7 6 �0.0001 995 table 8 numbers and percentages of cases where corrections to file(s) resulted in fico scores(s) crossing a lending threshold (589, 619, 659, 689, or 719 points) score crossed threshold average fico score overall �590 590–679 680–749 750–789 �790 no number 75 53 42 28 11 209 percent 91.5 72.6 62.7 93.3 100.0 79.5 yes number 7 20 25 2 — 54 percent 8.5 27.4 37.3 6.7 — 20.5 total number 82 73 67 30 11 263 percent 100.0 100.0 100.0 100.0 100.0 100.0 15l.d. smith et al. / financial services review 27 (2018) 1-28 by characteristics or circumstances of the individual who files the dispute. the overwhelming factor is the content of the credit record itself. we must acknowledge, however, that this presumes the individual has help in clearly stating the nature of the alleged error and identifying the specific items in the credit report(s) that should be corrected (and how). this was assured in our study by the individuals’ receiving letters prepared by university research associates to be completed by them and mailed to the relevant credit bureaus. in the analysis to this point, we have determined the extent to which various factors account for consumers’ experience with the u.s. system of granting credit and reporting credit information. finally, we present the parameters for individual variables in the regression and logistics models to show their individual statistical significance and magnitudes of their marginal impact. to facilitate the interpretation of the logistic models, instead of giving the logit coefficients, we indicate the impact of unit changes in the respective explanatory variables upon the corresponding odds ratios for the respective binary outcomes.6 table 11 contains parameters (coefficients) for the three “full” regression models and the “odds ratio” effects for variables in the five “full” logistic models. table 12 contains the coefficients for the eight “reduced models” that were produced by stepwise backward elimination of variables (one at a time) that did not meet a 0.1 level of statistical significance (i.e., 0.05 level for a one-tailed test) at each stage. considering the reduced model in the last column of table 12, for example, the odds of a having a credit score cross a lending threshold(the probability of crossing a threshold divided by the probability of not crossing a threshold) are estimated to be table 9 significance tests on logistics models for the likelihood that one or more credit scores will increase 10� points because of correction(s) of potentially material error(s) factors included �2 df p-value sample size full model 21.6 29 0.84 262 ex adverse events 3.56 5 0.61 262 ex demographic 3.34 6 0.77 262 ex resources 4.92 6 0.55 262 ex household char 5.76 6 0.45 262 ex payment habits 2.87 2 0.24 262 ex credit record 3.36 4 0.50 262 reduced model 4.2 1 0.01 262 table 10 significance tests on logistics models for the likelihood that one or more credit scores will cross a lending threshold because of correction(s) of potentially material error(s) factors included �2 df p-value sample size full model 35.7 29 0.18 262 ex adverse events 5.54 5 0.35 262 ex demographic 4.33 6 0.63 262 ex resources 8.73 6 0.19 262 ex household char 7.73 6 0.26 262 ex payment habits 0.11 2 0.94 262 ex credit record 19.53 4 0.0006 262 reduced model 10.9 3 0.01 262 16 l.d. smith et al. / financial services review 27 (2018) 1-28 t ab le 11 r eg re ss io n co ef fic ie nt s an d od ds ra tio ef fe ct s fo r th e “f ul l m od el s” v ar ia bl e c re di t ut ili za tio n/ lim its (t ab le 2) c re di t ut ili za tio n/ in co m e (t ab le 3) pa ys � m in du e od ds ef f (t ab le 4) pa ys al l ba la nc es od ds ef f (t ab le 5) a ve ra ge cr ed it sc or e (t ab le 6) m at er ia l di sp ut e od ds ef f (t ab le 7) m at er ia l sc or e in cr ea se od ds ef f (t ab le 9) sc or e cr os se s th re sh ol d od ds ef f (t ab le 10 ) in te rc ep t 45 .4 2* ** 6. 64 ** nr nr 62 3. 55 ** * nr nr nr u ne m pl oy m en t 6. 88 ** � 1. 24 0. 60 ** 0. 77 � 19 .3 3* ** 1. 07 0. 49 .5 1 in co m e dr op 4. 03 3. 36 ** 0. 72 0. 57 ** * � 10 .5 2* * 1. 37 * 1. 23 1. 83 * r ec en t bi rt h � 4. 29 � 3. 68 * 0. 81 0. 86 2. 64 1. 07 0. 73 0. 26 ** d iv or ce /s ep ar at io n 4. 30 2. 06 0. 97 1. 25 � 17 .0 5* * 1. 20 1. 09 1. 34 m aj or m ed ic al 5. 25 * 0. 55 0. 55 ** * 0. 71 � 15 .4 2* ** 0. 81 0. 88 1. 33 a ge un de r 30 � 5. 04 � 4. 93 ** * 1. 91 ** * 1. 45 * � 7. 20 0. 67 * 0. 65 0. 73 m al e � 3. 59 0. 18 0. 99 1. 38 ** 1. 72 0. 74 * 1. 06 1. 69 * a si an � 11 .0 4* * � 3. 16 1. 67 4. 46 ** * 7. 28 0. 99 1. 60 0. 73 a fr ic an a m er ic an 1. 31 � 4. 44 ** 0. 37 ** * 0. 39 ** * � 39 .8 9* ** 1. 19 1. 13 0. 55 so m e co lle ge � 5. 00 2. 12 1. 19 1. 13 11 .5 1* 0. 69 0. 64 0. 67 g ra du at e de gr ee � 6. 11 1. 81 2. 26 ** 2. 09 ** * 21 .1 9* ** 0. 69 0. 52 0. 74 in co m e ov er $7 5k 1. 03 0. 03 0. 94 1. 08 4. 47 2. 01 ** * 1. 32 2. 89 ** in co m e $2 5– $5 0k 3. 35 3. 08 * 0. 64 * 1. 02 � 3. 13 0. 91 0. 89 3. 21 ** in co m e le ss $2 5k � 11 .2 8* * 3. 66 0. 69 1. 30 4. 43 0. 59 * 0. 68 1. 08 c on tin ge nc y fu nd s � 8. 33 ** * 1. 86 2. 33 ** * 2. 51 ** * 30 .5 8* ** 0. 90 1. 02 0. 97 pe ns io n pl an � 2. 87 � 1. 27 0. 87 1. 13 � 2. 94 1. 00 1. 37 1. 33 o th er re tir em en t � 10 .9 6* ** � 2. 50 * 1. 72 ** 1. 93 ** * 28 .8 3* ** 0. 90 1. 52 2. 31 ** h om eo w ne r � 9. 20 ** * 2. 31 1. 48 * 1. 75 ** * 16 .4 2* ** 1. 01 � 0. 56 * 0. 59 m ar ri ed � 6. 02 ** � 2. 27 1. 54 * 1. 61 ** 8. 93 * 1. 01 1. 24 0. 65 e m pl oy ed fu ll tim e 6. 65 ** 1. 91 1. 06 0. 62 ** � 21 .9 1* ** 1. 16 0. 68 0. 59 y ea rs w ith em pl oy er � 0. 21 � 0. 02 1. 01 1. 01 0. 46 0. 98 0. 99 0. 98 h ou se ho ld si ze 3. 57 ** * � 0. 29 0. 99 0. 77 ** * � 4. 38 * 0. 81 * 1. 18 1. 50 ** k id s un de r 18 � 0. 44 0. 26 0. 75 ** 1. 11 � 2. 96 1. 12 0. 85 0. 83 pa ys � m in ba la nc e 40 .3 0* ** 1. 24 1. 26 1. 15 pa ys al l ba la nc es 50 .5 3* ** 0. 70 * 1. 79 * 0. 89 a ve ra ge cr ed it sc or e 0. 99 ** * na na a ve ra ge sc or e q ui nt ile 1 1. 00 0. 13 ** * a ve ra ge sc or e q ui nt ile 2 0. 94 0. 82 a ve ra ge sc or e q ui nt ile 4 0. 79 0. 08 ** * a ve ra ge sc or e q ui nt ile 5 0. 19 ** nr * si gn ifi ca nt at 0. 1 le ve l, ** si gn ifi ca nt at 0. 05 le ve l, ** * si gn ifi ca nt at 0. 01 le ve l, nr � no t re le va nt , na � no t ap pl ic ab le . 17l.d. smith et al. / financial services review 27 (2018) 1-28 t ab le 12 r eg re ss io n co ef fic ie nt s an d od ds ra tio ef fe ct s fo r “r ed uc ed m od el s” v ar ia bl e c re di t ut ili za tio n/ lim its (t ab le 2) c re di t ut ili za tio n/ in co m e (t ab le 3) pa ys � m in du e od ds ef f (t ab le 4) pa ys al l ba la nc es od ds ef f (t ab le 5) a ve ra ge cr ed it sc or e (t ab le 6) m at er ia l di sp ut e od ds ef f (t ab le 7) m at er ia l sc or e in cr ea se od ds ef f (t ab le 9) sc or e cr os se s th re sh ol d od ds ef f (t ab le 10 ) in te rc ep t 44 .9 4* ** 12 .6 7* ** nr nr 62 2. 79 ** * nr nr nr u ne m pl oy m en t 6. 68 ** 0. 50 ** * � 18 .3 7* ** 0. 54 ** in co m e dr op 5. 30 ** 2. 94 ** 0. 51 ** * � 9. 69 ** r ec en t bi rt h � 3. 55 * 0. 33 ** d iv or ce /s ep ar at io n � 16 .8 1* * m aj or m ed ic al 5. 62 * 0. 55 ** * 0. 69 * � 15 .4 2* ** a ge un de r 30 � 5. 68 ** * 2. 07 ** * 0. 65 * m al e 1. 35 ** 0. 76 * a si an � 11 .6 2* * 4. 37 ** * a fr ic an a m er ic an � 4. 58 ** * 0. 34 ** * 0. 36 ** * � 39 .7 0* ** so m e co lle ge s � 5. 99 * 11 .3 3* g ra du at e de gr ee � 7. 22 * 2. 12 ** * 1. 86 ** * 22 .3 5* ** in co m e ov er $7 5k 1. 97 ** * 2. 45 ** in co m e $2 5– $5 0k 3. 18 ** in co m e le ss $2 5k � 12 .4 8* ** c on tin ge nc y fu nd s � 9. 52 ** * 2. 58 ** * 2. 56 ** * 30 .7 1* ** o th er re tir em en t � 11 .5 0* ** � 2. 38 ** 1. 89 ** * 1. 92 ** * 29 .2 0* ** 1. 96 ** h om eo w ne r � 8. 26 ** * 1. 54 ** 1. 65 ** * 17 .7 6* ** m ar ri ed � 6. 55 ** � 2. 97 ** 1. 61 ** 1. 47 ** 11 .4 2* * e m pl oy ed fu llt im e 4. 05 * 0. 70 ** � 22 .8 9* ** y ea rs w ith em pl oy er 0. 54 * h ou se ho ld si ze 3. 03 ** * 0. 81 ** * � 5. 90 ** * 0. 89 ** k id s un de r 18 0. 73 ** * pa ys � m in ba la nc e 40 .0 1* ** pa ys al l ba la nc es 50 .9 0* ** 0. 71 * a ve ra ge cr ed it sc or e 0. 99 ** * a ve ra ge sc or e q ui nt ile 1 0. 20 ** * a ve ra ge sc or e q ui nt ile 2 a ve ra ge sc or e q ui nt ile 4 0. 10 ** * a ve ra ge sc or e q ui nt ile 5 nr * si gn ifi ca nt at 0. 1 le ve l, ** si gn ifi ca nt at 0. 05 le ve l, ** * si gn ifi ca nt at 0. 01 le ve l, nr � no t re le va nt . 18 l.d. smith et al. / financial services review 27 (2018) 1-28 80% lower (changed by a factor of 0.2) for individuals with credit scores in the lowest quintile relative to those in the middle quintiles after considering family income, whether the family had a recent birth, and whether the individual had other forms of retirement savings. concentrating on the coefficients for the reduced models, with other things considered, we see higher credit utilization relative to credit limits among individuals who suffered recent or extended unemployment, a drop in income, major medical expenses, were employed fulltime, or lived in large households. lower usage of available credit was seen for individuals in the lowest income group, individuals who had contingency funds to cover an unexpected two-month interruption in income, who had retirement savings or pension plans, were homeowners, asians, college-educated individuals, and married persons. recall, however, that 83–84% of the variation of credit utilization relative to credit limits was unexplained by the regression models. individuals who were unemployed, had major medical expenses, were african american, or who had more children at home, ceteris paribus, were less likely to pay at least the minimum amounts due on all accounts each month. conversely, if the person was under 30, had a graduate degree, had contingency funds or retirement plan, was a homeowner or married, he or she was more likely to pay more than the minimum due on all accounts each month. payment habits, of course, are highly influential in determining the individual’s credit score. after accounting for those habits and other variables in the model, average credit scores were still higher for college-educated individuals, for people with contingency funds, homeowners, married individuals and people who are employed longer with their current employer. average scores were lower for individuals who were unemployed, who had recent drops in income, were divorced or separated, had major medical expenses, were african american, employed full-time (vs. part-time)7 and from larger households. material disputes, ceteris paribus, were more likely to arise among individuals with average lower credit scores and also among individuals with family income over $75k. some further variation was explained by age, gender, drops in income and household size. following disputes, improvement in credit standing was less likely for african americans and more likely for individuals in the highest and lowest income categories but this is likely because of associations between average credit scores for individuals (i.e., the nature of the credit record itself) as discussed earlier. race (black) becomes insignificant at the 0.05 level when indicators of credit-score quintiles are included in the reduced model to accommodate the nonlinear effects of the quality of the credit record. the last pair of models suggests that individuals in the lowest quintile are as likely as those in quintile 3 to receive a 10-point increase in their credit scores but that is insufficient to have them cross a lending threshold because of other deficiencies in their records. 5. implications for credit-reporting practice based on qualitative observations from the intensive reviews of credit-bureau records in this study (and decades of personal experience in the use of such data for risk management in financial institutions), we are able to offer suggestions for possible changes to credit19l.d. smith et al. / financial services review 27 (2018) 1-28 reporting practices that may improve the quality and utility of information in credit-bureau files while respecting consumer privacy. setting aside the fact that bureau files represent only one side of the ledger (i.e., not providing information about income or financial assets), there were several recurring themes observed by our research associates as they discussed the credit reports with study participants. they pertained to measures of: (1) credit utilization, (2) alleged applications for new credit, and (3) collection activity. the construction of credit scores is, as acknowledged earlier, hampered by the lack of information on credit limits and payments made on revolving accounts. neither of these variables can be deduced unequivocally from account balances and delinquency status. this makes estimates of utilization and whether individuals pay accounts in full each month prone to considerable error. because of concerns for individual and institutional privacy, we frankly doubt that satisfactory remedies exist for these shortcomings. “hard pulls” of bureau records (caused by applications for new credit) reduce credit scores. they may occur, however, for reasons with dramatically varying implications about risk—from trivial credit-checks for a new cell-phone account to highly leveraged financial gambits. more information about the reason for “hard” inquiries and the amount of the credit line would be helpful (if the borrower were to authorize such disclosure). finally, collection accounts can grossly overstate the amounts of debt involved and often involve disputed obligations. late payments and service charges (sometimes arbitrarily imposed by unscrupulous collectors) can overwhelm the amount of an original debt. it is often impossible to tie a collection account to an original obligation or to nature of product or service involved. a collector may have purchased accounts receivable from landlords, medical clinics, or from a bankrupt enterprises that failed to meet its obligations to the consumer. further documentation regarding the nature of the collection account, the original creditor, current amount owing, cumulative interest and service charges, and magnitude of the original debt or credit line (before default) would improve the record. 6. conclusion and directions for future research with this research, we have confirmed some patterns observed in previous studies on the use of consumer credit, refined analysis of consumers’ habits in managing debt, and extended the inquiry to encompass the roles that individual consumers must play to ensure that their credit records are accurate. a series of regression and logistic models at different stages of the credit cycle reveals the dynamics of the conceptual model from fig. 1. consistently with previous research, we see that credit usage depends somewhat on socio-economic factors but still varies a great deal among individuals with similar educational backgrounds, income, family resources, ethnicity, major family events, and so forth. misuse of credit and vulnerability to adverse events occurs in a broad socio-economic spectrum. addressing our first research question (whether we can account for differences in individuals’ success in managing their credit obligations), we see that the credit record (as revealed in the average fico scores from the three major credit bureaus) is highly related to each of the factors studied (family resources, household characteristics, demographic attributes, and occurrence of adverse events)—and especially related to individuals’ self-reported 20 l.d. smith et al. / financial services review 27 (2018) 1-28 habits in meeting their monthly obligations. we should note that, in our sample, just 46% of respondents claimed to pay off all account balances each month while the triennial survey of consumer finances (bricker et al., 2014) reported that the percentage of households that habitually pay off credit card balances each month increased by 10.5 percentage points (from 57.9 to 64.0) over the period 2001 to 2013. responses in our study were given after an intensive review of the individuals’ credit reports. this review, even though the credit reports do not explicitly indicate whether balances on revolving credit are paid in full each month, may increase the accuracy of consumers’ responses to questions about their payment habits. with credit usage and payment behavior related, however, to household resources, adverse events, and demographic characteristics, some of the difference may be related to composition of the respective samples or to interpretation of the meaning of “normally paying of all balances” in the respective surveys. addressing the second research question (whether individuals in some demographic groups are more vulnerable to inaccuracies in their credit reports), we see that the propensity to identify alleged errors in the credit records is primarily related to the nature of the person’s credit record rather than standard demographic characteristics of the person. some demographic groups, however, are more prone to usage of credit without paying off all balances each month. that makes them more vulnerable to reporting of derogatory information in the credit file and, indirectly, more prone to errors that can affect the credit report. our findings in this regard reinforce those of avery et al., (2012) but with added precision from our directly addressing the issues in our extensive interviews with the consumers and our joint reviews of their credit reports. incidence of adverse impact is much lower than previously reported in studies by various advocacy groups but a significant number of consumers in our sample (5.5%) had a credit score cross a threshold for more favorable credit terms after errors were corrected in their credit files. addressing the final research question (whether some demographic groups are more successful than others in getting alleged errors rectified), we see that the outcomes of disputes are similarly related to the credit record rather than to the demographic characteristics of the disputant (provided, of course, that disputes are registered with necessary clarity). disputes filed with credit bureaus to correct alleged errors are most likely to produce material improvements in the disputant’s credit score if the current score is in the second or third quintile (at or slightly below average)—regardless of the demographic characteristics of the disputant. individuals with very poor or exceptionally good credit records are less likely to see material differences in their credit standing from corrections to the credit files, but for different reasons. several issues call for further research. first, while our results suggest that the dispute resolution process is reasonably effective, it should be remembered that study participants were carefully guided through the dispute process to ensure that our measures of accuracy and adverse impact were as accurate as possible. we did not test whether individuals with alleged errors in their credit reports are able to document them effectively on their own and be as successful in having changes made as did the participants in our study. second, given that fewer than one in six of our study participants had reviewed their credit reports in the two years before joining the study, one is compelled to ask why consumers do not exert the effort necessary to do so. in an era of well-publicized data breaches and identity theft, one 21l.d. smith et al. / financial services review 27 (2018) 1-28 would expect consumers to be more vigilant than they seem to be. finally, given that the regulatory model in place is reliant on the ability and willingness of consumers to monitor the accuracy of the information in their credit files, one must ask if this process could be improved with the implementation of an educational process which describes the system and the importance of monitoring. in the process of reviewing the credit reports with university associates, study participants received a thorough education on the scope and significance of information contained in their credit reports. many expressed gratitude for having a better understanding of factors that affect their credit scores and how good personal financial decisions could help to improve them. future research might indicate whether more efforts at financial literacy would contribute to better personal financial management and measurable improvements in consumer credit scores. notes 1 additional information about the fcra and subsequent amendments is available from the federal trade commission (ftc) web site at http://www.ftc.gov/os/statutes/ 031224fcra.pdf. 2 financial institutions (especially credit-card issuers) are reputedly reluctant to report the credit limits because they tempt competitors to poach business by offering cards with higher limits. the absence of proper credit limits, however, reduces the informative content of credit scores on which they all rely for business decisions. this seems to be a tradeoff that financial firms are willing to bear. 3 this can happen, for example, when credit limits are lowered after delinquency. the patterns of statistical significance of the individual factors were similar when the extreme values were included, but the r2 statistic for the full model dropped from 17% to 10% when the extreme cases were included. 4 if any of the elements of a factor or the dependent variable was missing, the case was not included. 5 crossing a threshold would be expected to change the pricing of a loan only if the potential lender was using that particular score as the determining factor. if the lender had another but lower score in hand, the latter would tend to dominate. 6 the probability that a binary logistic outcome equals one is computed by exp(logit)/ (1�exp(logit)) for our logistic models. the odds-ratio effects are the inverse logarithms of the corresponding logistic parameters. 7 effects of full-time employment may be spurious in this model (because of relationships with other variables such as a spouse’s having a high income); it was not significantly correlated with average credit score over all. acknowledgment this work was performed with support from the united states federal trade commission. it does not necessarily reflect the views of the ftc. 22 l.d. smith et al. / financial services review 27 (2018) 1-28 appendix a: summary statistics for variables used in the models variable nonmissing obs. minimum maximum mean standard deviation max rev. bal. to max credit line (%) 876 0 140.3 30.44 33.93 max. rev. bal. to annual income (%) 943 0 129.6 9.22 17.18 average fico score 1,001 459.7 823.3 698.15 95 always pays more than min. payments 1,001 0 1 0.79 0.41 pays off all revolving balances monthly 1,001 0 1 0.46 0.5 filed material dispute 1,001 0 1 0.26 0.44 score increased � 10 points with corrections 263 0 1 0.33 0.47 score crossed threshold 263 0 1 0.21 0.40 recent extended unemployment 1,001 0 1 0.22 0.41 reduced household income 1,001 0 1 0.38 0.49 recent new birth in family 1,001 0 1 0.1 0.3 divorce separation death of spouse 1,001 0 1 0.07 0.25 major medical bill 1,001 0 1 0.17 0.38 age under 30 1,001 0 1 0.21 0.41 male 1,001 0 1 0.51 0.5 asian 1,001 0 1 0.04 0.19 black 1,001 0 1 0.13 0.34 some college (including undergrad degree) 1,001 0 1 0.62 0.49 graduate degree 1,001 0 1 0.26 0.44 income over $75k 1,001 0 1 0.39 0.49 income $25k–$50k 1,001 0 1 0.23 0.42 income under $25k 1,001 0 1 0.13 0.34 adequate funds for contingencies 1,001 0 1 0.69 0.46 has pension plan 1,001 0 1 0.56 0.5 has nonemployee retired savings account 1,001 0 1 0.38 0.49 owns home 1,001 0 1 0.63 0.48 married 1,001 0 1 0.54 0.5 employed fulltime 1,001 0 1 0.57 0.5 years with current employer 1,001 0 44 5.15 7.67 household size 997 1 10 2.57 1.31 number of children under 18 years 1,001 0 6 0.54 0.94 23l.d. smith et al. / financial services review 27 (2018) 1-28 appendix b: credit-review checklist, interview guide, and closing survey (reformatted to conserve space) credit review checklist it would be most helpful when we call to interview you if you focus on these items: 1. are your name, address and identifying information on the front of the report correct? 2. if you currently have a mortgage on a home is the name of the lender correct? 3. is the outstanding balance correct? (remember that the balance could be from any day of the previous month). 4. is the date the loan was opened correct? 5. are the number of delinquencies, if any, correct? 6. if you have had a car loan in the past seven (7) years, find that in the report and check the date the loan was opened, lender name, loanbalance and payment history. are these items correct? 7. for each credit card you currently have, locate it in the report and check the date issued, lender name, outstanding balance and payment delinquencies, if any. are these items shown correctly? 8. if you currently have other types of monthly installment loans for purchasing goods or services, please locate these in the report. again, pay special attention to the date of the loan, lender name, outstanding balance and payment delinquencies, if any. are all of these items correct? 9. for any closed loans (mortgages, car loans, credit cards, and so forth) carefully examine those that show there may have been delinquent payments. does this information appear correct? 10. does the report show any loans that went to collection? is the information reported correctly? (a collection account is one where you missed a payment and the lender has hired a colllection agency to collect the money from you. it is listed at the end of the credit report). 11. after you have reviewed one credit report in detail, please do the same with the other two reports. do you see any significant differences among the three reports? what are they? please make note of questions you have and to discuss with our research associate during your phone interview. thank you again for participating in this study. interviewing guide good evening/afternoon . . . this is ___________________________________ calling from the university of missouri/university of arizona regarding the federal trade commission study on the accuracy of credit report information. thank you for participating. may we review your reports now? (if not, ask for a time to reschedule, and either confirm or tell him/her you will call them back to confirm a new, convenient time.) first, i’ll answer any questions you may have about the content of your three credit reports. then, we’ll discuss any possible inaccuracies you may have identified during your own review. next, we discuss the findings from our comparison of the information in the three credit reports. and finally, i’ll ask you for some general demographic information to complete our study. 1. using the checklist that we sent you along with your three credit reports, did you find any questionable items? if no, confirm that they have covered every category in the list. mention each item in the list. if yes, what did you find? obtain clarification for each item mentioned or alleged error starting with the disputed credit report and then compare the corresponding item in the other credit reports. when all the participant’s questions have been answered, address any inconsistencies or irregularities identified in the preparatory review: (continued on next page) 24 l.d. smith et al. / financial services review 27 (2018) 1-28 appendix b: (continued) from our own review, we have a few questions we would like to ask . . . if there are any derogatory items review these with the participant to be sure the facts are presented correctly. if the participant finds alleged errors explain the dispute process and explain the difference between material and immaterial errors. if the participant claims an immaterial error suggest he/she call the lender to resolve. if the participant claims a material error, explain how you will help them file a dispute. helping the participant file a dispute: ● be sure you fully understand what the participant feels is in error and why; ● explain that you will prepare a letter (to that particular cra) and send it to the participant; ● he/she will need to review the information, sign, fill in the identifying information (social security number and date of birth) and mail the letter to the cra; ● also explain that you will include a self-addressed stamp postcard for the participant to return to the study office to confirm that the dispute letter(s) were mailed to the cra; ● ask the participant to forward any correspondence they receive from the credit reporting agency in response to the dispute letters; ● tell the participant you will follow up with them by email or phone to find out what the cra told them about the dispute, or what the cra has done about it; ● explain that, in about eight weeks, the research team will draw a new credit report and will contact them to let them know what changes (if any) were made in response to the dispute letters. closing survey (administered to all participants at the completion of the interview) “i think we have covered everything in your credit reports. now i need to record some additional information to help in our interpretation of the study data . . .” credit report accuracy and disputes have you ever previously requested a copy of your credit reports? yes ___ no__ if yes: a) where did you get them? _________________ b) when did you last request a credit report? _____________ have you ever previously disputed any inaccurate information on any of your credit reports or requested that some negative information removed? yes __ no __ if yes: a) when did you last dispute inaccurate information on your credit report? ___________ b) were you successful in getting the item changed or removed? yes no c) if yes, how did you accomplish this? for example did you write to a bureau, telephone the creditor, write to a creditor, use a bureau website, or get help from a credit counselor? d) how long did it take? (hours of effort on the consumer’s part) _______________ e) how were you notified of the result?: e-mail __ mail ___ phone__ other ___ none___ f) have you previously obtained your credit scores? for example, have you ever obtained your fico score, vantage score, or some other credit score? yes __ no__ if yes: a) where did you get your credit score? ___________________ b) when did you last obtain your credit score? ________________________ c) have you obtained credit in the past seven years that is not reflected in your credit reports? yes __ no __ if yes: a) what was the total amount? _________________________ b) was there a formal payment schedule? yes __ no __ if yes: c) were all payments made on time? yes __ no__ d) are the loans completely paid off? yes __ no __ (continued on next page) 25l.d. smith et al. / financial services review 27 (2018) 1-28 appendix b: (continued) do you generally pay more than the minimum balance on all your credit cards every month? yes __ no __ do you generally pay off all your credit card balances from month to month? yes__ no __ in the past 12 months have you had any late payment fees for credit cards, overdraft charges on bank accounts, or other late payment fees? yes __ no __ within the past year have you had any of your credit limits on your credit cards or other accounts reduced? yes __ no __ if yes: a) on how many? ________ b) by how much? ______________________________ in the past 12 months have you used pay day loans, pawn shops, auto title loans or similar forms of credit to meet you immediate needs for cash? yes __ no __ if yes: a) how many times in the past 12 months? ________ b) did you use these types of loans to avoid high late payment fees or interest charges on other forms of credit? __________________________________________ household characteristics “. . . the ftc needs to ensure that it has based its study on a properly representative sample of u.s. consumers. we need, therefore, to provide (anonymously) some information about each participant and their households. we have a few general questions in that regard . . .” 1. in which age bracket would you place yourself? (a) under 30 (b) 31–40 (c) 41–50 2. (d) 51–60 (e) over 60 3. what is your gender? m__ f__ 4. would you classify yourself as: (a) white, (b) african-american, (c) hispanic 5. (d) asian, (e) other 6. are you currently: (a) married, (b) living with a partner, (c) never married, (d) divorced, (e) separated, (f) widowed 7. what is your highest level of education? (a) no high school diploma (b) high school diploma (c) some college (d) associates degree (e) bachelor’s degree (f) graduate degree 8. rating yourself on a scale of 1–5, with one being not knowledgeable at all to 5 being very knowledgeable, how knowledgeable do you consider yourself about credit related matters? 9. using the same rating scale of 1–5 with one being not confident at all to 5 being very confident, how confident do you consider yourself in your ability to manage your own finances? 10. “. . . the ftc is interested in the most effective ways to educate american consumers about credit matters and personal financial management . . .” 11. did you participate in any personal finance courses in high school or (if applicable) in college? yes__ no__ 12. “. . . some states have begun to offer mandatory education about personal finance and credit management to high-school students.” 13. in which state and city did you attend high school? ______________________ 14. are you currently employed? yes __ no __ and, researchers note the employment research category: (a) employed (b) self employed (c) homemaker, (d) retired, (e) unemployed, (f) disabled, (g) other 15. full time or part time 16. (if employed) how many years have you been with your current employer? ___ 17. (if employed) what is your occupation? ___________________________ and, researchers note occupational research categories (a) professional (b) admin/mgr (c) trade and/or technical staff (d) sales (e) clerical (d) retired (e) disabled (continued on next page) 26 l.d. smith et al. / financial services review 27 (2018) 1-28 references archuleta, k. l., dale, a., & spann, s. m. (2013). college students and financial distress: exploring debt, financial satisfaction, and financial anxiety. journal of financial counseling and planning, 24, 50–62. avery, r. b., calem, p. s., canner, g. b., & bostic, r. w. (2003). an overview of consumer data and credit reporting. federal reserve bulletin, 89, 47–73. avery, r. b., breevoort, k. p., & canner, g. (2012). does credit scoring produce a disparate impact? real estate economics, 40, s65–s114. bell, m. m., nelson, j. s., spann, s. m., malloy, c. j. , britt, s. l., nelson, g., & briana, s. (2014). the impact of financial resources on soldiers’ well-being. journal of financial counseling and planning, 25, 41–52. bricker, j., dettling, l. j., henriques, a., hsu, j. w., moore, k. b., sabelhaus, j., thompson, j., windle, r. a., devlin-foltz, s., & krimmel, j. (2014). changes in u.s. family finances from 2010 to 2013: evidence from the survey of consumer finances. federal reserve bulletin, 100, 1–41. elliehausen, g., & lowrey, b. r. (2000). the costs of implementing regulatory changes: the truth in savings act. journal of financial services research, 17, 165–179. engelberg, e., & sjoberg, l. (2007). money obsession, social adjustment, and economic risk perception. the journal of socio-economics, 36, 686–697. fair and accurate credit transactions act of 2003 (fact act or facta). (2003). public law 108–159-dec 4, 2003. (available at http://www.gpo.gov/fdsys/pkg/plaw-108publ159/pdf). fair credit reporting act (fcra). (1970). 15 u.s.c. § 1681 et seq. (available at http://www.ftc.gov/os/statutes/ 031224fcra.pdf). appendix b: (continued) 18. how many people are in your household? ________________ 19. are there any children under the age of 18 living in the household? yes __ no __ if yes, how many? __________________ 20. in which general category does your total household income fall? (a) under $25k (b) 25–49.9k (c) 50–74.9k (d) 75–99.9k (e) 100–149k (f) 150–200k (g) �200k (h) declined 21. do you own or rent your living quarters? own __ rent __ 22. have you, in the past two years, experienced any of the following events that may affect your current financial situation? i’ll read five events: a. an extended period of unemployment (3 months or longer) when you couldn’t find a job? b. a significant reduction in your household income? c. birth of a family member for which you have financial responsibility? d. divorce, separation or death of a spouse? e. major medical bill that was not covered by insurance? 23. without taking out a loan or using a credit card, do you have savings or other financial resources to come up with $2,000 within 30 days if you needed it for an emergency? (yes/no) 24. do you have an employer-provided retirement plan other than social security? (yes/no) if yes, is it: (a) a defined benefit plan where the employer pays a predetermined amount for life beginning at retirement (usually based on years of service and salary (b) a defined contribution plan where both employee and employer may contribute to individual’s accounts and the benefit amount at retirement is based on the amount invested and the investment performance of those accounts, (c) both (a) and (b) or (c) some other type of employer-provided retirement plan? (specify) _________________________________________________ 25. do you have another type of registered retirement account (not provided by an employer) such as a selfdirected ira or roth ira (i.e., a retirement plan where the owner of the account, usually through a trustee, makes all contributions and investment decisions)? the benefit amount at retirement is based on the amount invested and the investment performance of those accounts. (yes/no) 27l.d. smith et al. / financial services review 27 (2018) 1-28 fay, s., hurst, e., & white, m. (2002). the household bankruptcy decision. american economic review, 92, 706–718. glick, r., & lansing, k. j. (2009). u.s. household deleveraging and future consumption growth. federal reserve board of san francisco, economic letter, 16. grafova, i. b. (2007). your money or your life: managing health, managing money. journal of family and economic issues, 28, 285–303. himmelstein, d. u., warren, e., thorne, d., & woolhandler, s. (2005). illness and injury as contributors to bankruptcy. marketwatch. health affairs web exclusive 2. project hope–the people-to-people health foundation, inc. javine, v. (2013). financial knowledge and credit usage in college students. financial services review, 22, 367–87. lyons, a. c., rachlis, m., & scherpf, e. (2007). what’s in a score? differences in consumers’ credit knowledge using ols and quantile regressions. the journal of consumer affairs, 41, 223–249. national association of state pirgs. (2004). mistakes do happen: a look at errors in consumer credit reports. washington, dc: national association of state pirgs. national foundation for credit counseling. (2012). the 2012 consumer financial literacy survey prepared by harris interactive inc. rochester, ny: national foundation for credit counseling. norvilitis, j. m., & maclean, m. g. (2010). the role of parents in college students’ financial behaviors and attitudes. journal of economic psychology, 31, 55–63. o’neill, b., & xiao, j. j. (2014). post-recession, post-legislation credit use: insights from an online survey. journal of personal finance, 13, 65–76. ratcliffe, c., mckernan, s. m., theodos, b., kalish, e., chalekian, j., guo, p., & trepel, c. (2014). delinquent debt in america: an opportunity and ownership initiative brief. washington, dc: urban research institute. smith, h. l., finke, m. s., & huston, s. j. (2012). financial sophistication and housing leverage among older households. journal of family and economic issues, 33, 315–327. smith, l. d., staten, m., eyssell, t., karig, m., freeborn, b. a., & golden, a. (2013). accuracy of information maintained by u.s. credit bureaus: frequency of errors and effects on consumers’ credit scores. journal of consumer affairs, 47, 588–601. staten, m. e. (1993). the impact of post-bankruptcy credit on the number of personal bankruptcies. credit research center, working paper no. 58. krannert graduate school of management, purdue university west lafayette, in. staten, m. e., & cate, f. h. (2004). does the fair credit reporting act promote accurate credit reporting? building assets building credit: a symposium on improving financial services. joint center for housing studies, harvard university, babc 04–14, 1–58. sullivan, t. a., warren, e., & westbrook, j. l. (2006). less stigma or more financial distress: an empirical analysis of the extraordinary increase in bankruptcy filings. stanford law review, 59, 213–256. tang, t. l. (1992). the meaning of money revisited. journal of organizational behavior, 13, 187–202. 28 l.d. smith et al. / financial services review 27 (2018) 1-28 financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience ning tanga,*, paula c. peterb adepartment of finance, san diego state university, 5500 campanile drive, san diego, ca 92182-8239, usa bdepartment of marketing, san diego state university, 5500 campanile drive, san diego, ca 92182-8239, usa abstract this article explores how financial education, financial experience, and parents’ financial experience influence young adults’ financial knowledge. we rely on a general model of learning to hypothesize the determinants of financial knowledge acquisition. using data on 3,597 young adults from a national longitudinal survey, we find that financial education, financial experience, and parents’ financial experience all exert a positive impact on young adults’ financial knowledge. moreover, these determinants work interactively. both individual and parents’ financial experience help narrow the gap in financial knowledge caused by lack of financial education. © 2015 academy of financial services. all rights reserved. jel classification: d14; a2 keywords: financial literacy; financial education; financial experience; parental socialization 1. introduction young adults face unprecedented financial obligations and complexity in today’s demanding financial environment. as they become financially independent from their guardians, they must make choices about student loans, debt, insurance, mortgages, and retirement * corresponding author. tel.: �1-619-594-2082; fax: �1-619-594-3272. e-mail address: ntang@mail.sdsu.edu (n. tang). financial services review 24 (2015) 119–137 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. funds. the financial decisions made early in life can have significant long-term economic and social effects (montoya and scott, 2013). better financial behavior requires financial knowledge, specifically knowledge in essential personal finance concepts and products (chen and volpe, 1998; jacobs-lawson and hershey, 2005; van rooij, lusardi, and alessie, 2011). because choices made at the beginning of financial independence exert such a strong influence later, it is especially urgent that young people develop basic knowledge and skills in economics and finance (shim et al., 2013). alhenawi and elkhal (2013) infer from their findings that to promote financial planning, we should strongly foster financial education at early stages of life. policymakers, the financial service industry and educators have promoted numerous programs and initiatives to combat low levels of financial knowledge among young people. for example, in 2009, 21 states required an economics course for high school graduation whereas 13 states required a personal finance course (council for economic education, 2009). as of 2013, these numbers have been increased to 22 and 17, respectively (council for economic education, 2014). in 2010, under the dodd-frank act (h.r. 4173), the u.s. congress created the consumer financial protection bureau to further promote financial education through its consumer engagement and education group. there also has been a surging interest in financial education by u.s. financial institutions and their associations (worthington, 2006). community banker (2003) shows that in 2003, 98% of u.s. community banks sponsored financial literacy programs and 72% offered their own programs. despite these efforts, lack of financial knowledge among young people is still widespread (see, e.g., chen and volpe, 1998; mandell, 2009). to formulate effective interventions to increase financial knowledge among the young, we need to identify and manage aspects that influence the process through which people acquire financial knowledge. in the past, the main focus has been on the formal education system to disseminate financial knowledge to young people. however, because of poor results, researchers and policymakers have started questioning the roles parents and personal experience play in effective financial knowledge learning. for example, johnson and sherraden (2007) show that students can successfully obtain necessary financial concepts by participating in programs that provide education as well as hands-on investment and management experience, such as “save for america” and “illinois bank-at-school” programs. lusardi, mitchell, and curto (2010) suggest parents’ financial experience is an important variable affecting a young adult’s financial knowledge in addition to financial education. surveying financial literacy among college students, chen and volpe (1998) find that parents, participants’ own mistakes, and school courses are all listed as people’s sources of personal finance education. however, no study to our knowledge has tested for the concurrent roles of financial education, financial experience and parents’ financial experience in reducing the gap in financial knowledge. by taking a general model of learning from the education literature (kolb, 1984) and applying it to financial knowledge learning, our study aims to identify the concurrent roles of financial education, financial experience, and parent’s financial experience as determinants of financial knowledge acquisition. moreover, the interaction among these variables is also explored. results from subgroup analysis and poisson regression based on 3,597 young adults from the 1997 national longitudinal survey of youth indicate that financial educa120 n. tang, p.c. peter / financial services review 24 (2015) 119–137 tion, financial experience, and parents’ financial experience all significantly improve young adults’ financial knowledge. moreover, they work interactively. young adults lacking financial education benefit more from financial experience and parents’ financial experience. that is, both individual and parents’ financial experience can help narrow the gap in financial knowledge caused by lack of financial education. the article offers two main contributions. first, the article adapts a solid model of learning from the field of education to the field of financial literacy to create a financial knowledge acquisition framework. second, the article identifies three determinants of financial knowledge learning—financial education, financial experience, and parents’ financial experience— and provides empirical evidence proving that these determinants significantly affect financial knowledge and operate interactively. taken together, these contributions flesh out the theory behind financial knowledge acquisition and can be used to evaluate and improve financial education programs. better programs mean youth more educated in financial knowledge and more prepared to make the crucial financial decisions faced at the beginning of financial independence. the remainder of the article is organized as follows: section 2 introduces the conceptual framework and hypotheses; section 3 describes our data and measures; section 4 shows the results; and section 5 offers conclusions. 2. conceptual framework and hypotheses in discussing the major gaps in evaluation literature on financial education and counseling, collins and o’rourke (2010) indicate a lack of guiding theories in the literature. they pointed out that because of the lack of a prevailing theoretical framework in the field, most studies failed to cite a specific theory or understand the theoretical underpinnings of their work. we find that the experiential learning theory (elt) by kolb (1984) in education research literature can help us fill the gap and we use it to guide our study. to our knowledge, this is the first study to use this theory to explain the acquisition of financial knowledge. according to this model, an individual acquires knowledge through experiences (e.g., owning a stock or bond), observations (e.g., having parents with financial experience), and conceptualizations (e.g., receiving formal financial education) and then tests that knowledge through active experimentation (e.g., practice over time), which results in new experiences. kolb’s model implies the importance of experience and reflective observation with the phenomena being studied rather than merely conceptualizing it. when taken as a comprehensive theoretical framework, elt can be used to explore learning processes and educational issues in various disciplines such as education, management, computer and information sciences, psychology, medicine, nursing, accounting, and law (see kolb, boyatzis, and mainemelis, 1999 for a review). perplexingly, no study to our knowledge has suggested the possible application of elt to financial knowledge acquisition. this research adopts kolb’s elt model to financial knowledge acquisition and explores the determinants that may have an impact on financial knowledge. conceptualizations, such as financial education acquired in the classroom, as well as experiences and observations, such as the ones acquired through personal and parents’ financial experience, may all 121n. tang, p.c. peter / financial services review 24 (2015) 119–137 influence financial knowledge acquisition. moreover, we extend this model by considering the influence of financial experience and parents’ financial experience on the effects of financial education on financial knowledge acquisition. fig. 1 offers an illustration of kolb’s theoretical learning model adapted to the financial knowledge acquisition process. the focus of our research is to select proxy measures or examples for “financial education,” “financial experience,” and “parents’ financial experience” components, and use empirical data to validate the hypothesized financial knowledge acquisition framework. 2.1. financial education according to the elt model, coursework at school is the major source of conceptualization learning in the current education system. therefore, financial education is expected to exert a positive impact on financial knowledge. a burgeoning collection of literature has assessed how well financial education actually improves financial knowledge, but findings are not conclusive. some research shows financial education has a positive effect on financial knowledge (see, e.g., danes, huddlestoncasas, and boyce, 1999; tennyson and nguyen, 2001). other studies find no significant effect (e.g., mandell, 2009). lyons, rachlis, and scherpf (2007) conclude that the discrepancy in mixed results regarding financial education effectiveness is because of differences in the programs various researchers evaluated and in the methods they used to evaluate those programs, or differences in what was measured and how. for example, self-reported financial education exposure measures might lead different results than college major measures, which seem to be less prone to selection bias. to address this issue, this article uses college major to measure financial education exposure, which is less prone to selection bias; it also considers the concurrent roles of subjects’ financial experience and their parents’ financial experience to evaluate the effect of formal financial education at school, which we hypothesize has a positive effect on financial knowledge. thus: hypothesis 1: young adults should have higher level of financial knowledge if they fig. 1. hypothesized financial knowledge acquisition framework. 122 n. tang, p.c. peter / financial services review 24 (2015) 119–137 possess higher financial education considering the concurrent roles of financial experience and parents’ financial experience. 2.2. financial experience in the 2001 surveys of consumers, respondents reported personal financial experience as the most important way they had learned cash-flow management, credit management, saving, and investment practices (hilgert, hogarth, and beverly, 2003). lyons, rachlis, and scherpf (2007) further confirm that personal experience impacts financial knowledge of credit and debt management. these findings are unsurprising in the context of the elt model, where experience learning emphasizes the roles that direct experience and focused reflection play in increasing knowledge. this view has inspired numerous financial education campaigns that introduce programs focused on hands-on investment and management experience to traditional formal education. for example, the “bank in school” program has students open a saving account at school, make deposits, calculate simple interest, and track their saving balance; the “huntington bank kids’ club” program in columbus, ohio, provides students with an on-site school bank to promote hands-on experience learning at school (consumer bankers association, 2002). the general assumption underlying these programs is that individual financial experience can help narrow the gap in financial knowledge caused by lack in financial education exposures. despite empirical evidence showing that direct personal financial experience affects financial knowledge, no study to our knowledge has systematically incorporated financial experience into a financial knowledge learning framework and tested whether it positively affects financial knowledge when taken with other influencing factors (i.e., financial education and parents’ financial experience). moreover, no study to our knowledge has explored the possibility for financial experience to compensate for lack of financial education. thus: hypothesis 2a: young adults should have higher levels of financial knowledge if they possess higher financial experience considering the concurrent roles of financial education and parents’ financial experience. hypothesis 2b: financial experience should positively impact financial knowledge especially in the absence of financial education. 2.3. parents’ financial experience significant family members, especially parents, present incomparable socialization influences on young adults’ learning processes (xiao et al., 2011). by interacting with parents, children develop consumer skills, knowledge, and attitudes. even as they enter early adulthood, parental influence remains a potentially important socializing force (bowen, 2002; norvilitis and maclean, 2010; shim et al., 2013). in a survey of 924 students enrolled at various universities, 74% of women and 68% of men stated that they obtained their personal finance knowledge from their parents (chen and volpe, 2002). furthermore, findings from the 2001 parents, youth, and money survey suggest that parents who think they do an “excellent” or “good” job managing their money are more likely to provide their 123n. tang, p.c. peter / financial services review 24 (2015) 119–137 children with financial guidance than those parents who think they do a “fair” or “poor” job managing their money (employee benefit research institute, 2001). therefore, we expect financially experienced parents to be more capable of helping their children. we hypothesize that parents’ financial experience positively influences youth financial knowledge and could help to narrow the gap in financial knowledge when financial education at school is not available. this leads us to our final hypotheses: hypothesis 3a: young adults should have higher levels of financial knowledge if their parents possess higher financial experience considering the concurrent roles of financial education and financial experience. hypothesis 3b: parents’ financial experience should positively impact financial knowledge especially in the absence of financial education. 3. data and measures 3.1. data this article uses data from the 1997 national longitudinal survey of youth (nlsy97) provided by the u.s. bureau of labor statistics. the nlsy97 is a nationally representative sample of the u.s. youth population. to reach a total sample of 8,984 respondents, nlsy97 interviewers screened 75,291 households in 147 primary sampling units that did not overlap (a primary sampling unit is a metropolitan area or, in nonmetropolitan areas, a single county, or group of counties). the longitudinal dataset follows the same group of respondents from 1997 (wave 1) to 2010 (wave 14), recording data annually. the survey contains extensive information on respondents’ demographic and socioeconomic characteristics, family backgrounds, and educational experiences. because few surveys simultaneously gather data on an individual’s financial education, family background, and financial experience, few studies have evaluated how these three determinants of financial knowledge acquisition work together. the rich dataset from nlsy97 and its longitudinal feature enables us to fill this gap. our study used 3,597 respondents from the nlsy97 dataset because these records had valid responses for all of our study variables. 3.2. measures 3.2.1. financial knowledge the nlsy97 2007 survey (wave 11) asked respondents the following three financial knowledge questions aimed at testing basic but fundamental financial concepts regarding risk diversification, interest rate, and inflation. 1. do you think that the following statement is true or false? buying a single company stock usually provides a safer return than a stock mutual fund. (true/false) 2. suppose you had $100 in a savings account and the interest rate was 2% per year. 124 n. tang, p.c. peter / financial services review 24 (2015) 119–137 after 5 years, how much do you think you would have in the account if you left the money to grow: more than $102, exactly $102, or less than $102? (a. more than $102; b. exactly $102; c. less than $102) 3. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, would you be able to buy more than, exactly the same as, or less than today with the money in this account? (a. more than today; b. exactly the same as today; c. less than today) these questions have been shown to differentiate well between financially knowledgeable and financially naïve respondents and were also included in the 2004 health and retirement survey (hrs2004), the 2009 american life panel (alp2009), and the 2009 national financial capability study (nfcs2009; lusardi and mitchell, 2008, 2011; lusardi, mitchell, and curto, 2010). using the responses to these three financial knowledge questions, we created a financial knowledge score that sums the number of correct answers across the three questions. we use this financial knowledge score for the rest of our analysis. 3.2.2. financial education in surveys conducted from 1997 to 2007 (waves 1 through 11), respondents were asked about their majors when they were in college. we associated majoring in economics and business management with higher financial education exposure. we consider this measure more reliable compared with self-reported data on financial education collected years after the respondents left school (e.g., peng et al., 2007). it is expected that those who benefited more from financial education are more likely to remember and, therefore, report having such education. therefore, self-reported data could cause selection bias in estimation (collins and o’rourke, 2010). in addition, mandell and klein (2007) point out that the low financial knowledge scores among young adults, even after they have taken a course in personal finance, are related to a lack of motivation to learn or retain these skills. consequently, to mitigate the bias caused by students’ motivation to acquire financial knowledge, we use college majors that are the choice of respondents to measure financial education exposures. it is noted that the selected sample also includes respondents who were not in college before 2007. their value of exposure to college level financial education is set to zero. 3.2.3. financial experience we aggregated responses from the 1998 to 2007 surveys (wave 2 through 11) on whether young adults invested in stocks, mutual funds, cds, bonds, or t-bills. we used responses to evaluate respondents’ financial experience before 2007. 3.2.4. parents’ financial experience the 1997 survey (wave 1) asked respondents’ parents whether they had financial experience in stocks, bonds, or pension funds. we used this as indicator of parents’ financial experience. 3.2.5. covariates covariates include individual demographic and socioeconomic characteristics. we control for individual’s gender (wave 1), age (wave 1), race (wave 1), and income earned in 2006 125n. tang, p.c. peter / financial services review 24 (2015) 119–137 (wave 11). these factors have been shown to significantly impact one’s financial knowledge (see, e.g., lusardi, mitchell, and curto, 2010; mandell, 2009; worthington, 2006). we also control for young adults’ education level in 2007 (wave 11) and high school gpa collected in 1999 (wave 3) as lusardi, mitchell, and curto (2010) show that education attainment and cognitive ability affect financial knowledge. the highest educational attainment by parents is also included in covariates. our final control variable is whether young adults asked parents about financial issues. this is to control for the opportunity children had to learn from their parents. 3.3. summary statistics our study used a sample of 3,597 respondents who had valid responses for all of our study variables listed above. table 1 summarizes data on respondents’ financial knowledge, financial education exposure, personal financial experience, and parents’ financial experience in our selected sample. the average financial knowledge score among young adults in the selected sample is 1.84 out of 3. eighteen percent of the respondents majored in economics or business management in college before answering the financial knowledge questions in 2007. respondents reported limited personal financial experience, with only 24.6% reporting that they had invested in stocks or bonds before 2007. fifty-eight percent of parents reported that they had financial experience. our selected sample consists of young adults with an average age of 24 in 2007. fifty percent of them are males; the sample earned an average annual income of $24,011 in 2006. the breakdown of ethnicities is: white (65.1%); black (23.4%); american indian, eskimo, or aleut (0.8%); asian or pacific islander (1.4%); and other (9.4%). as of 2007, 26% of respondents in the selected sample had obtained a college degree, 3% had obtained a graduate degree, and 71% had not obtained a college degree. the average high school gpa was 2.89. forty-one percent of respondents’ parents had high school or less as their highest education attainment, and 31% had college or more. seventy percent of young adults in the sample had asked their parents about financial issues. 4. results 4.1. subgroup analysis we first adopted one-way analysis of variance (anova) model to investigate whether there are any differences in mean financial knowledge scores in different subgroups. f-statistics from the test indicate whether the difference between the means of the subgroups is significant or not. because the focus of the article is to examine the roles of financial education, financial experience, and parents’ financial experience on financial knowledge, we defined subgroups by these three factors. the results are summarized in table 2. it is found that all three components in our financial knowledge acquisition framework have significant impact on financial knowledge. economics or business management majors scored higher than those who did not major in economics or business management. respondents who had 126 n. tang, p.c. peter / financial services review 24 (2015) 119–137 financial experience had an average financial knowledge score of 2.19, compared with 1.73 by those who without financial experience. the difference is statistically significant. our results also suggest that respondents whose parents had investment experience scored higher than their counterparts. 4.2. poisson regression–main effects in this section, we will test the concurrent roles of financial education, financial experience, and parents’ financial experience on financial knowledge, after controlling for other covariates. specifically, we evaluated the research hypotheses using poisson regression. table 1 summary statistics mean sd min max dependent variables total financial knowledge score 1.84 .92 0 3 independent variables majored in economics or business management before 2007 18.0% 38.5% 0 1 invested in stocks or bonds before 2007 24.6% 43.1% 0 1 parents invested in stocks or bonds or had pension accounts 58.3% 49.3% 0 1 covariates male 50.1% 50.0% 0 1 age in 2007 24.34 1.47 22 28 race white 65.1% 47.7% 0 1 black 23.4% 42.3% 0 1 american indian, eskimo, or aleut 0.8% 9.1% 0 1 asian or pacific islander 1.4% 11.6% 0 1 other 9.4% 29.2% 0 1 education non-high school 11.8% 32.3% 0 1 high school 49.8% 50.0% 0 1 some college 9.4% 29.1% 0 1 college 26.0% 43.9% 0 1 graduate school 3.0% 17.1% 0 1 high school gpa 2.89 .59 .42 4 income from wage and salary in 2006 (in $) $24,011 $18,255 0 $175,000 under $10,000 23.9% 42.6% 0 1 $10,000 to $25,000 37.7% 48.5% 0 1 $25,000 to $50,000 33.0% 47.0% 0 1 $50,000 or more 5.5% 22.8% 0 1 parents’ education non-high school 10.8% 31.1% 0 1 high school 30.1% 45.9% 0 1 some college 28.6% 45.2% 0 1 college 15.6% 36.3% 0 1 graduate school 14.9% 35.6% 0 1 ask parents about finance issues 70.1% 45.8% 0 1 n � 3,597. the table shows the summary statistics of dependent variable, three independent variables, and covariates of our analysis. mean, standard deviation, minimum and maximum values are reported. 127n. tang, p.c. peter / financial services review 24 (2015) 119–137 previous research has used poisson regression to model the count outcome of financial behaviors among young adult populations (worthy, jonkman, and blinn-pike, 2010). poisson regression is more appropriate in this study than standard linear regression for several reasons. first, preliminary analyses of our data demonstrated that using linear regression models resulted in models generating out of bounds predictions. for example, predictions of young adults’ financial knowledge scores may have values higher than three based on ols regression. in addition, poisson regression is more appropriate than linear regression because our dependent variable (financial knowledge score) is a discrete count variable, it is not overdispersed (e.g., financial knowledge scores’ variance does not exceed its mean), and it fits the poisson distribution well (deviance goodness-of-fit test �2 (df) � 1808.14 (3580), p � 0.99; pearson goodness-of-fit test �2 (df) � 1417.84 (3080), p � 0.99). therefore, we ran poisson regression to assess the effects of financial education, financial experience, parent’s financial experience and their interaction effects on financial knowledge. coefficients in poisson regression are the difference between the logs of expected counts, which is difficult to interpret intuitively. to help clarify our results, we report both the regression coefficients (�) and incidence rate ratios (irrs), calculated as (e�). irrs are interpreted as the change in the rate ratio of financial knowledge scores for one unit change in the independent variable. in other words, an irr � 1 (�1) implies a one unit increase in an independent variable will decrease (increase) the predicted rate of financial knowledge scores by a factor of the reported irr for the independent variable. table 3 summarizes the results from poisson regression model on the main effects of financial education, financial experience, and parents’ financial experience on financial knowledge scores, after controlling for other covariates. we find that financial education significantly increases financial knowledge (p � 0.01), as does financial experience (p � 0.01) and parents’ financial experience (p � 0.10). for example, choosing finance related subjects as college majors is associated with a 12% increase in financial knowledge score. having financial experience and parents having financial experience increase the rate of table 2 subgroup analysis no. of obs. financial knowledge score financial education before 2007 a. majored in economics or business management before 2007 649 2.17 b. did not major in economics or business management before 2007 2,948 1.77 f statistics 102.43*** financial experience before 2007 a. invested in stocks or bonds before 2007 886 2.19 b. did not invested in stocks or bonds before 2007 2,711 1.73 f statistics 173.04*** parents’ financial experience a. parents invested in stocks or bonds or had pension accounts 2,096 1.99 b. parents did not invest in stocks or bonds or had pension accounts 1,501 1.64 f statistics 123.15*** mean financial knowledge scores for each subgroup are reported. f statistics from anova test is used to indicate whether financial knowledge scores differ significantly between subgroups. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. 128 n. tang, p.c. peter / financial services review 24 (2015) 119–137 financial knowledge scores by 10% and 4% respectively. these results confirm hypotheses hypothesis 1, hypothesis 2a, and hypothesis 3a. with respect to the covariates in the study, there is a significant difference between men and women (p � 0.01) with respect to the expected number of correct financial knowledge questions answered. the results indicate, all else held constant, men are expected to exhibit 19% higher financial knowledge scores than women. age is not statistically significant. only black and respondents belong to the “other” race group are significantly different from the reference group (white) with respect to financial knowledge level. all else equal, black respondents report only 0.94 times as many correct financial knowledge questions as white respondents (p � 0.01). as expected, both respondents’ and parents’ educational attainment and respondents’ income are highly significant and positively associated with financial knowledge (p � 0.01 for all variables). it is surprising to find that “ask parents about finance issues” predicts lower financial knowledge. one explanation could be that those who turn to parents for financial advice are more likely to be those who lack financial knowledge. table 3 poisson regression results main effects irr coefficient sig. se intercept .95 �.05 .14 factors financial education before 2007 1.12 .11 *** .02 financial experience before 2007 1.10 .10 *** .02 parents’ financial experience 1.04 .04 * .02 covariates male 1.19 .17 *** .02 age in 2007 .99 �.01 .01 race (ref: white) black .94 �.06 *** .02 indian .97 �.03 .08 asian .97 �.03 .05 other .95 �.05 * .03 education 1.07 .07 *** .01 high school gpa 1.14 .13 *** .02 income from wage and salary in 2006 (ref: under $10,000) $10,000 to $25,000 1.06 .06 *** .02 $25,000 to $50,000 1.12 .12 *** .02 $50,000 or more 1.16 .15 *** .03 parents’ education 1.03 .03 *** .01 ask parents about finance issues .96 �.04 ** .02 n 3597 wald �2 (df) 987.74 (16) aic 10276.38 bic 10381.57 the table shows the poisson regression results. dependent variable is the number of correct financial knowledge questions answered (0–3). independent variables are respondents’ financial education before 2007, financial experience before 2007 and parents’ financial experience. control variables include gender, age in 2007, race, education, high school gpa, respondents’ income from wage and salary in 2006, parents’ educational attainment, and whether respondents asked parents about finance issues. regression coefficients (�), incidence rate ratios (irrs) calculated as (e�), significance level and standard error are reported. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. 129n. tang, p.c. peter / financial services review 24 (2015) 119–137 4.3. poisson regression–interaction effects to evaluate hypothesis 2b and hypothesis 3b (the interaction effects), we incorporate interaction terms between financial education and financial experience, and between financial education and parents’ financial experience into our poisson regression model. the model is otherwise the same as the main effects only model. before interpreting the coefficients on interaction terms in table 4, it should be noted that there is a baseline difference in financial knowledge scores between those who had financial education in college and those who did not. after mean-centering, marginal means of financial knowledge scores estimated from table 4 poisson regression results–interaction effects irr coefficient sig. se intercept .95 �.06 .14 factors financial education before 2007 1.20 .18 *** .03 financial experience before 2007 1.12 .12 *** .02 parents’ financial experience 1.05 .05 ** .02 financial education before 2007* .93 �.07 ** .03 financial experience before 2007 financial education before 2007* .93 �.07 * .04 parents’ financial experience covariates male 1.19 .18 *** .02 age in 2007 .99 �.01 .01 race (ref: white) black .94 �.06 *** .02 indian .96 �.04 .08 asian .98 �.02 .05 other .95 �.05 * .03 education 1.07 .07 *** .01 high school gpa 1.14 .13 *** .02 income from wage and salary in 2006 (ref: under $10,000) $10,000 to $25,000 1.06 .05 ** .02 $25,000 to $50,000 1.12 .11 *** .02 $50,000 or more 1.16 .15 *** .03 parents’ education 1.03 .03 *** .01 ask parents about finance issues .96 �.04 ** .02 n 3597 wald �2 (df) 996.38 (18) aic 10277.70 bic 10395.27 the table shows the results of poisson regression with interaction effects. dependent variable is the number of correct financial knowledge questions answered (0–3). independent variables are respondents’ financial education before 2007, financial experience before 2007, parents’ financial experience, interaction between financial education before 2007 and financial experience before 2007, and interaction between financial education before 2007 and parents’ financial experience. control variables include gender, age in 2007, race, education, high school gpa, respondents’ income from wage and salary in 2006, parents’ education attainment, and whether respondents asked parents about finance issues. regression coefficients (�), incidence rate ratios (irrs) calculated as (e�), significance level and standard error are reported. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. 130 n. tang, p.c. peter / financial services review 24 (2015) 119–137 poisson regression for those who received financial education in college are 1.80 (se � 0.07) and 1.50 for those who did not receive finance education in college (se � 0.04). this difference needs special attention because poisson regression coefficients can be interpreted as the percentage predicted change in financial knowledge scores expected by one unit change in a predictor, not the unit predicted change in financial knowledge scores (coxe, west, and aiken, 2009). thus, an equal percentage change for those with and without financial education will ultimately result in an even greater difference in financial knowledge scores because those who had financial education in college have a higher baseline of predicted financial knowledge scores (i.e., as the value of any number becomes larger, a constant percentage change will result in an increasingly larger absolute change in the number). for the aforementioned reasons, one intuitive and recommended approach to better understand the results of poisson regression interactions is to visually plot the marginal predicted financial knowledge scores for those had financial education and those did not across values of each respective independent variable, holding all other variables constant at their mean or reference value (coxe, west, and aiken, 2009). therefore, we depict these marginal mean plots in figs. 2 and 3, along with table 4 to demonstrate the interaction effects. table 4 shows that the positive relationship between financial experience and financial knowledge score is stronger among those who lack financial education than those who have it. the slope difference test is significant (� � �0.07, p � 0.05). these findings support hypothesis 2b. the results indicate that, the financial knowledge gap between those who had finance education and those who did not will narrow as financial experience levels increase. fig. 2 clearly depicts this point. the positive relationship between parents’ financial experience and financial knowledge score is stronger among those who lack financial education than those who have it. the fig. 2. marginal means of financial knowledge score by financial education *financial experience (95% confidence interval). 131n. tang, p.c. peter / financial services review 24 (2015) 119–137 difference is statistically significant (� � �0.07, p � 0.10). as depicted in fig. 3, those who lacked financial education in college are able to narrow the gap with those who had financial education with respect to financial knowledge scores if their parents’ have financial experience. for example, without parents’ financial experience, having financial education is predicted to exhibit 0.30 (1.8 � 1.5) increase in financial knowledge score, but the difference is only 0.19 (1.76 � 1.57) when parents of young adults have financial experience. these findings support hypothesis 3b. 4.4. reverse causality issue our measure on financial education in college (college major) could subject the impact of financial education to a selection bias. a bias would arise if youth who are more financially knowledgeable are more likely to choose finance-related majors. therefore, the positive relationship between college major in finance-related subjects and financial knowledge does not necessarily indicate that financial education in college improves financial knowledge. to mitigate the bias, we used college major data before 2007 when financial knowledge questions were asked in the baseline analysis in previous sections. that is, the students chose their majors before their financial knowledge were evaluated. this way we could reduce the possibility that financial knowledge affects college major choice. in addition, we test for the possibility that it is because financial knowledge affects college major choice that we found a positive relationship between them in tables 3 and 4. specifically, we follow bernheim, garrett, and maki (2001) and create a new variable “financial education after 2007,” which uses college major data after the financial knowledge questions were asked in fig. 3. marginal means of financial knowledge score by financial education *parents’ financial experience (95% confidence interval). 132 n. tang, p.c. peter / financial services review 24 (2015) 119–137 2007. we run poisson regression as in the previous section and replace the variable indicating college major before 2007 with college major after 2007. as shown in column 1 in table 5, the effect of financial education after 2007 is not significant, which indicates that the difference in financial knowledge between individuals who chose finance-related majors versus those who chose non-finance related majors exists only after the major is chosen, not before. therefore, it is exposure to financial education that improves one’s table 5 poisson regression results–interaction effects (college major and financial experience after 2007) 1. financial education after 2007 2. financial experience after 2007 irr coefficient sig. se irr coefficient sig. se intercept .75 �.29 .34 .94 �.06 .17 factors financial education 1.15 .14 .10 1.20 .18 *** .04 financial experience 1.13 .12 *** .04 1.08 .08 .05 parents’ financial experience 1.09 .09 * .05 1.05 .05 ** .02 financial education* financial experience .97 �.03 .12 1.00 .00 .07 financial education* parents’ financial experience .83 �.19 * .11 .92 �.08 * .04 covariates male 1.22 .20 *** .04 1.21 .19 *** .02 age in 2007 1.00 .00 .01 .99 �.01 .01 race (ref: white) black .89 �.11 ** .05 .94 �.06 *** .02 indian .96 �.04 .16 .97 �.03 .08 asian .81 �.21 .14 .98 �.02 .06 other .97 �.03 .06 .93 �.07 ** .04 education 1.05 .05 ** .03 1.07 .07 *** .01 high school gpa 1.17 .16 *** .04 1.14 .13 *** .02 income from wage and salary in 2006 (ref: under $10,000) $10,000 to $25,000 1.04 .04 .05 1.04 .04 * .02 $25,000 to $50,000 1.15 .14 *** .05 1.11 .10 *** .02 $50,000 or more 1.30 .26 *** .09 1.14 .13 *** .04 parents’ education 1.00 .00 .02 1.03 .03 *** .01 ask parents about finance issues .96 �.00 .04 .96 �.04 * .02 n 591 2,885 wald �2 (df) 133.68 (18) 676.80 (18) aic 1717.23 8235.23 bic 1800.49 8348.61 the table shows the results of poisson regression with interaction effects. dependent variable is the number of correct financial knowledge questions answered (0–3). independent variables in model 1 are respondents’ financial education after 2007, financial experience before 2007, parents’ financial experience, interaction between financial education after 2007 and financial experience before 2007, and interaction between financial education after 2007 and parents’ financial experience. model 2 used financial education before 2007 and financial experience after 2007; other variables are the same as in model 1. control variables include gender, age in 2007, race, education, high school gpa, respondents’ income from wage and salary in 2006, parents’ education attainment, and whether respondents asked parents about finance issues. regression coefficients (�), incidence rate ratios (irrs) calculated as (e�), significance level and standard error are reported. ***, **, and * indicate statistical significance at the 1%, 5%, and 10% levels, respectively. 133n. tang, p.c. peter / financial services review 24 (2015) 119–137 financial knowledge score, not financial knowledge that pushes one to become exposed to financial education (i.e., choose one of the finance-related majors). similarly, the positive relationship between financial experience and financial knowledge scores could be subject to reverse-causality bias, the problem commonly suffered by most studies on this topic. given the positive correlation between financial experience and financial knowledge, it is hard to determine the direction of causality. the longitudinal feature of nlsy97 dataset allows us to mitigate and test for such bias. first, we only used financial experience before 2007 when financial knowledge questions were asked in our baseline analysis in the previous section. if the financial knowledge level is evaluated after the financial experience, to some extent, it reduces the possibility that financial knowledge affects financial experience. however, it is possible that financial knowledge is correlated over time. that is, not investing in stocks in the pre-question period may be a reflection of low financial knowledge, which is then explicitly measured by the questions asked later. to exclude such a possibility, we rerun the poisson regression by using financial experience after 2007. the results are shown in column 2 table 5. the effect of financial experience after 2007 is not significant. the difference in financial knowledge between individuals who have financial experience and those who do not disappears if financial experience was measured after the financial knowledge question. therefore, the positive correlation between financial knowledge and financial experience observed in tables 3 and 4 is mainly caused by the effects of financial experience on financial knowledge. our conclusions are robust. 5. conclusions financial knowledge can lead to better decision-making. to formulate effective public policy interventions to increase financial knowledge among the young, we need to identify and manage aspects that influence the process through which individuals acquire financial knowledge. this article adapts kolb’s model of learning process, which relies on the concurrent roles of the educational, observational, and experiential components of knowledge acquisition. with a longitudinal study, we test the effects of three hypothesized factors on financial knowledge: financial education, financial experience, and parents’ financial experience. based on our results, all three factors significantly improve young adults’ financial knowledge. moreover, they work interactively. young adults who do not have financial education benefit more from financial experience and parents’ financial experience. our results indicate the indispensable role of hands-on experience and parents’ influence especially when school education is not available. our findings have important implications for financial literacy program evaluation, design, and implementation. our model indicates that financial education, financial experience and parents’ financial experience work concurrently and interactively on financial knowledge acquisition. therefore, it’s inappropriate to evaluate the effectiveness of one determinant independently from the other factors. for example, to evaluate the effectiveness of a personal finance curriculum, we need to consider students’ previous financial experience and their parents’ financial sophistication. otherwise, the results can be biased. in addition, our results show that school-based education is not the only way young adults learn financial 134 n. tang, p.c. peter / financial services review 24 (2015) 119–137 knowledge. both individual and parents’ financial experience could help narrow the gap in financial knowledge caused by lack of financial education. again, these findings can help in the design of better financial education programs. we encourage more policy support for community and family-based interventions, especially among those who lack formal financial education. we would like to offer an evaluation of the three financial knowledge questions used in the nlsy97 survey. although many studies use them, the questions mainly focus on evaluating how well respondents have grasped certain theoretical concepts. these questions may fail to capture the application-oriented knowledge learned from parents or gained through investment experience. this is part of the larger problem of a lack of rigorous measures of financial literacy for researchers to use in their studies (huston, 2010; schmeiser and seligman, 2013; volpe, chen, and liu, 2006). as measures of financial knowledge become more rigorous, the results of studies such as ours will become more exact and useful. another area of possible improvement is our financial knowledge acquisition framework measures. our study uses college major as a proxy for financial education, and uses investment experience as a proxy for financial experience. future studies can support and expand our theoretical framework by considering an array of variables that might fall under financial education and financial experience. for example, financial education in high school or state financial education curriculum mandates could be alternative measures of young adults’ exposure to financial education. studying the influence of financial experience in areas such as cash-flow management, credit management, savings, or retirement planning would be another productive avenue for future inquiry. it would also be interesting to extend our model by analyzing the difference in financial knowledge between other subgroups such as male versus female (alhenawi and elkhal 2013; chen and volpe, 1998). last, as previous literature has argued, there are many unobservable factors that might influence the relationship between our variables of interest. the type of regressions that are reported in the text, even though we make attempt to address reverse causality, still suffer from omitted variables bias. to address this issue, future studies should use an experimental approach, where respondents are divided into “control” and “treatment” groups and exposed to different doses of financial education and experiential learning (see, e.g., wiedrich et al., 2014). references alhenawi, y., & elkhal, k. (2013). financial literacy of u.s. households: knowledge vs. long-term financial planning. financial services review, 22, 211–244. bernheim, b. d., garrett, d. m., & maki, d. m. (2001). education and saving: the long-term effects of high school financial curriculum mandates. journal of public economics, 80, 435–465. bowen, c. f. (2002). financial knowledge of teens and their parents. financial counseling and planning, 13, 93–101. chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. 135n. tang, p.c. peter / financial services review 24 (2015) 119–137 collins, j. m., & o’rourke, c. m. (2010). financial education and counseling-still holding promise. journal of consumer affairs, 44, 483–498. community banker. (2003). banks stepping up financial literacy efforts. community banker, 12, 71–72. consumer bankers association. (2002). cba 2002 survey of bank-sponsored financial literacy programs. cba survey, april 2002. (available at https://www.faircredit.org/files/57.pdf). council for economic education. (2009). survey of the states – economic, personal finance & entrepreneurship education in our nation’s schools in 2009. (available at http://www.councilforeconed.org/wp/wp-content/ uploads/2011/11/2009-survey-of-the-states.pdf). council for economic education. (2014). survey of the states – economic and personal finance education in our nation’s schools 2014. (available at http://www.councilforeconed.org/wp/wp-content/uploads/2014/02/ 2014-survey-of-the-states.pdf). coxe, s., west, s. g., & aiken, l. s. (2009). the analysis of count data: a gentle introduction to poisson regression and its alternatives. journal of personality assessment, 91, 121–136. danes, s. m., huddleston-casas, c., & boyce, l. (1999). financial planning curriculum for teens: impact evaluation. financial counseling and planning, 10, 26–39. employee benefit research institute. (2001). parents, youth & money 2001. (available at http://www.ebri.org/ surveys/pym-es.pdf). hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 2003, 309–322. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. jacobs-lawson, j. m., & hershey, d. a. (2005). influence of future time perspective, financial knowledge, and financial risk tolerance on retirement saving behaviors. financial services review, 14, 331–344. johnson, e., & sherraden, m. (2007). from financial literacy to financial capability among youth. journal of sociology and social welfare, 34, 119–146. kolb, d. a. (1984). experiential learning: experience as the source of learning and development. englewood cliffs, nj: prentice-hall, inc. kolb, d. a., boyatzis, r. e., & mainemelis, c. (1999). experiential learning theory: previous research and new directions. http://www.d.umn.edu/�kgilbert/educ5165-731/readings/experiential-learningtheory.pdf. lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review: papers and proceedings, 98, 413–417. lusardi, a., & mitchell, o. s. (2011). financial literacy and planning: implications for retirement wellbeing. in o. s. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 17–39). oxford: oxford university press. lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lyons, a. c., rachlis, m., & scherpf, e. (2007). what’s in a score? differences in consumers’ credit knowledge using ols and quintile regressions. journal of consumer affairs, 41, 223–249. mandell, l. (2009). the impact of financial education in high school and college on financial literacy and subsequent financial decision making. (available at http://www.americasaves.org/elements/www. americasaves.org/file/fin%20behavior%20and%20ed%20v2.pdf). mandell, l., & klein, l. s. (2007). motivation and financial literacy. financial services review, 16, 105–116. montoya, d. y., & scott, m. l. (2013). the effect of lifestyle-based depletion on teen consumer behavior. journal of public policy and marketing, 32, 82–96. norvilitis, j. m., & maclean, m. g. (2010). the role of parents in college students’ financial behaviors and attitudes. journal of economic psychology, 31, 55–63. peng, t. m., bartholomae, s., fox, j. j., & cravener, g. (2007). the impact of personal finance education delivered in high school and college courses. journal of family and economic issues, 28, 265–284. schmeiser, m. d., & seligman, j. s. (2013). using the right yardstick: assessing financial literacy measures by way of financial well-being. journal of consumer affairs, 47, 243–262. 136 n. tang, p.c. peter / financial services review 24 (2015) 119–137 shim, s., serido, j., bosch, l., & tang, c. (2013). financial identity-processing styles among young adults: a longitudinal study of socialization factors and consequences for financial capabilities. journal of consumer affairs, 47, 128–152. tennyson, s., & nguyen, c. (2001). state curriculum mandates and student knowledge of personal finance. journal of consumer affairs, 35, 241–262. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. volpe, r. p., chen, h., & liu, s. (2006). an analysis of the importance of personal finance topics and the level of knowledge possessed by working adults. financial services review, 15, 81–98. wiedrich, k., collins, m., rosen, l., & rademacher, i. (2014). financial education and account access among elementary students. findings from the assessing financial capability outcomes (afco) youth pilot. cfed working paper, april 2014. worthington, a. c. (2006). predicting financial literacy in australia. financial services review, 15, 59–79. worthy, s. l., jonkman, j., & blinn-pike, l. (2010). sensation-seeking, risk-taking, and problematic financial behaviors of college students. journal of family and economic issues, 31, 161–170. xiao, j. j., tang, c., serido, j., & shim, s. (2011). antecedents and consequences of risky credit behavior among college students: application and extension of the theory of planned behavior. journal of public policy & marketing, 30, 239–245. 137n. tang, p.c. peter / financial services review 24 (2015) 119–137 the pros and cons of remaining in a 401(k) plan after retirement olivia s. mitchella,*, catherine reillyb, john a. turnerc adepartment of insurance/risk management & business economics/policy, the wharton school of the university of pennsylvania, 3620 locust walk, st 3000 sh-dh, philadelphia, pa 19104–6302, usa bone mifflin place, suite 404, cambridge, ma 02138, usa cpension policy center 3713 chesapeake street, washington, dc 20016, usa abstract this paper examines whether retirees would benefit from staying in their companies’ 401(k) plans after retirement, versus rolling their savings over to individual retirement accounts (iras). our focus is on individuals having low or moderate levels of financial literacy. we conclude that many such retirees would likely find it financially rewarding to retain their assets in their 401(k) plans. while iras currently offer a wider range of advice or distribution options compared with 401(k) plans, we close by pointing to legislative and technological developments that may produce better outcomes, more retirement confidence, and greater security for retirees. © 2023 academy of financial services. all rights reserved. jel classifications: g53; d14; j32 keywords: financial literacy; pension choices; ira rollovers; retirement decision-making 1. introduction many models of household financial behavior have traditionally analyzed decisions made by rational, well-informed individuals. in this paper, we focus instead on the vast majority of americans who enter retirement with low or moderate levels of financial literacy (lusardi & mitchell, 2014; lusardi et al., 2014).1 in particular, we evaluate the choices that these individuals confront at retirement when they have accumulated assets in a 401(k) plan and *corresponding author: tel.: +1-215-746-5706, fax: +1-215-573-3418. e-mail address: mitchelo@wharton.upenn.edu 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 1–21 must either roll their funds over to individual retirement accounts (iras), or leave their assets in their workplace 401(k) plans.2 in what follows, we focus mainly on those two options, while noting in passing that a third option, generally the least desirable of the three, is to cash out the plan and pay taxes (and perhaps penalties) on the assets withdrawn.3 these choices are available to many older workers, depending on the group considered. for example, 64% of private sector employees (excluding agricultural workers, household workers, and the self-employed) have access to defined contribution plans (bureau of labor statistics, 2021). for those participating in these plans, whether to roll over when they leave the labor force or transition from one job to another is a key financial choice they must make. as we show below, several factors can usefully be integrated into this decision, including the levels of expenses charged by different options, the types of funds available in which retirees can invest, the availability and cost of financial advice, and the retiree’s need for various types of decumulation products. in what follows, we first review the investment and administrative costs associated with iras and 401(k) plans. while these costs are likely to vary both between providers and different savers within each category, it is useful to highlight what is, and is not, included in these tallies. next, we examine different ways that financial advice is provided to 401(k) and ira participants, along with the various costs associated with these alternatives. another important factor shaping the decision to remain in an employer plan or roll over to an ira pertains to the size of the participant’s account balance. we also touch on both employer and financial advisor perspectives concerning the rollover decision. we find that, for people in employer plans with high fees, an ira rollover could be a beneficial financial decision. however, most people with low or moderate levels of financial literacy could do better financially by remaining in their employer 401(k) plans, in part because they have better fiduciary protection in those plans than in iras. we close with a summary of the issues facing participants, along with a brief discussion of potential improvements in the decumulation process for 401(k) account balances.4 2. understanding 401(k) and ira fees, and the role of plan size the institutional retirement income council (iric, 2021) quantified the additional retirement income participants can receive by retaining their assets in their employer-sponsored defined contribution plans throughout retirement, versus using iras that they access through advisors or brokers. the iric report proposed that employer-sponsored defined contribution plans can use their institutional bargaining power to provide their participants with better investment and decumulation options than those available from individually managed retirement accounts. in what follows, we investigate this proposition. three principal types of costs that ira and 401(k) participants must pay for their accounts involve investment management, administrative, and advisory fees. we first discuss investment management and administrative fees, including the importance of plan size for plan costs. advisory fees are discussed in section 4. 2 o. s. mitchell et al. / financial services review 31 (2023) 1–21 2.1. investment fees according to the investment company institute (duvall, 2021), 401(k) investors typically pay lower investment fees than do ira investors. that source noted that the asset-weighted expense ratio of 401(k) equity mutual funds was 39 basis points (bps) in 2020, compared with 57 bps for equity mutual funds in ira accounts. moreover, this comparison may understate the fee advantage of 401(k) plans, since the largest employer-sponsored retirement plans typically use collective investment trusts (cits) rather than mutual funds. one reason that 401(k) plans may have lower average fees than iras is that large 401(k) plans can often access low-cost institutional share classes or commingled investment trusts (cits) that are not available to retail investors. for plans with more than $500 million in assets, nearly 45% of their assets were in cits in 2022, a percentage that has grown over time. smaller plans have not shifted to cits, partly because cit minimums may be too high for smaller plans (morningstar, 2022). the cost differential between institutional and retail share classes may be larger in actively managed funds versus index funds. this cost differential also appears to have diminished over time, reducing the cost advantage of cits over mutual funds. for example, while one very large plan sponsor offers its 401(k) savers a passive target date fund (tdf) cit at 6.5 bps, vanguard recently lowered the cost of its retail target date mutual funds (available to ira savers) to 8 bps (szala, 2022). given these small differences, using a 401(k) rather than an ira does not automatically lead to lower investment fees, and financially sophisticated savers could build very low-cost ira portfolios for themselves. nevertheless, most participants are likely to lack the skills to do a good job constructing their own portfolios without advice. for these participants, the ability to stay in an institutionally priced plan with a professionally designed default investment option is likely to lead to a lower-cost outcome. 2.2. administrative fees in addition to investment fees, 401(k) providers also charge fees for asset custody, recordkeeping, and third-party administration. these are charged either on a per-participant basis, or as a percentage of assets under management. additionally, plans may charge for specific services, such as participant loans or withdrawals. while some ira providers do charge participants a small annual administrative (account) fee, many do not, with their income instead generated from investment expense charges levied on assets under management (folger, 2022). one large provider, for instance, has no setup, annual, or maintenance fee. a large recordkeeper offers an ira with a $20 annual fee plus a $20 annual fee on mutual funds for balances less than $10,000. participants can waive both fees by signing up for electronic delivery of account documents (ira reviews, 2021). robo-advisors are also available; these tend to be low-cost automated online platforms that manage participant assets based on employee preferences elicited via online surveys. one large robo-advisor charges the same fees (in basis points) for small and large accounts (betterment, 2021). o. s. mitchell et al. / financial services review 31 (2023) 1–21 3 relatively less information is available on 401(k) plan administrative fees. in a recent pension fees lawsuit, one large company was reported to have charged 86 bps in administrative fees, much higher than the average 44 bps charged by comparable plans having a similar asset range ($250 to $500 million; manganaro, 2019). a different pension fee lawsuit was filed against a company with annual recordkeeping fees of about $80 per participant (manganaro, 2021). schimmer (2021) reported that the median annual recordkeeping fee was $59 per participant in 2017. recordkeeping fees are charged either as per participant flat fees or as a percentage of assets across the whole plan; for low-balance participants, the latter approach is more advantageous. sometimes employers will cover active employees’ administrative fees, so these workers need not incur these costs out of pocket. 2.3. the impact of plan size on 401(k) fees fees for 401(k) plans vary considerably by plan size. the majority of 401(k) participants are in large plans, which usually charge lower fees than do small plans. for instance, 88% of private sector defined contribution participants were covered by plans with 100+ participants in 2018 (employee benefits security administration, 2021). brightscope/investment company institute (2020) found that 401(k) plans with under $1 million in assets had average total costs of 144 bps, versus 48 bps for plans with $100–$250 million in assets, and 28 bps for plans with over $1 billion in 2017. one reason that large 401(k) plans charge lower fees than small plans is that the former tend to be more heavily invested in index funds; accordingly, the higher fees that characterize small plans are in part due to the investment choices selected by plan sponsors. specifically, index funds comprised 23% of assets in plans with under $1 million in plan assets, versus 40% of assets in plans with over $1 billion (brightscope/investment company institute, 2020). over 95% of 401(k) plans with $10+ million in assets offered index funds in their investment options, while 79% of 401(k) plans with under $1 million did so in 2017. in addition to being more likely to offer index funds, participants in larger plans often also have access to lower-cost institutional share classes. for instance, 401(k) plans with assets below $1 million charged an average of 76 bps for target date mutual funds and 18 bps for index mutual funds in 2018; this compares with 37 bps for target date mutual funds and 6 bps points on index mutual funds in plans with assets over $1 billion (brightscope/investment company institute, 2020). another reason that larger plans have lower costs is that they may benefit from economies of scale in performing their administrative tasks. for instance, many administrative functions such as nondiscrimination testing or regulatory reporting can cost the same, regardless of the size of the plan. as a result, administrative fees as a percentage of assets tend to be higher for small plans than for large plans; this is one important reason for why small plans tend to be more expensive than large plans. the 401(k) average fees have also fallen over time. in addition to fee compression for both investments and administrative pricing, an additional explanation for this is the increase in the proportion of participants in low cost, large plans. according to morningstar (2022), 4 o. s. mitchell et al. / financial services review 31 (2023) 1–21 plans with more than $500 million in assets covered 34% of participants in 2011, but by 2019, they had added almost 13.5 million more participants and covered 43% of plan participants. around 15% of plans covered 90% of participants. these large plans have also pushed for lower fees by switching to index funds and investing in collective investment trusts. 3. how participant balances and financial literacy influence rollover decisions participant levels of financial literacy will also influence the choice between remaining in an employer plan or moving to an ira. financially literate participants who wish to manage their own assets will be able to use the wide range of competitively priced ira investment options to construct very low-cost portfolios for themselves. nevertheless, the less financially literate are more likely to need ira advice, as they will no longer have a default option selected by a plan sponsor on which to fall back. for participants who need advice, their level of accumulated assets will shape the type of advice they can access. table 1 provides background information on the distribution of 401(k) account balances for people in their 60s as they move into retirement. the median account balance was $90,385 in 2018 for people with 20–30 years tenure, making it a reasonable approximation of this group’s 401(k) accruals. at the 25th percentile, the account balance was $28,398, and at the 90th percentile, it was $557,589. we use this information to evaluate the availability of advice for individuals having high ($550,000), median ($100,000), and low ($30,000) plan balances at retirement. interestingly, the evidence shows that relatively few—only about one-fifth of low-balance retirees—knew that they pay investment fees on their accrued assets, versus 53% of retirees having balances over $100,000 (general accountability office, 2021). we hypothesize that the lower the accumulated balance, the more likely it is that the individual has low or moderate financial literacy and will need advice to construct a low-cost retirement portfolio. in table 2, we evaluate the cost differential to participants in employer plans of different sizes, comparing remaining in their employer plans versus rolling into iras. we assume that the cost of the 401(k) plan includes investment and administrative fees, whereas the cost of the ira contains only the cost of the investments. these calculations do not include the cost of financial advice. when participants are unsophisticated, we assume that they invest in “average cost” ira mutual fund portfolios; conversely, for financially sophisticated participants, we assume that they invest in low-cost passive target date funds. this may understate the cost of the ira option to unsophisticated participants, because, in addition to table 1 percentile distribution of 401(k) account balances for people in their 60s by tenure in the plan, 2018 job tenure (years) 10th percentile 25th percentile 50th percentile 75th percentile 90th percentile all 7,039 28,398 90,385 257,752 557,589 5–10 7,144 19,983 51,654 140,907 307,093 10–20 4,138 19,229 54,683 152,088 355,609 20–30 8,286 34,638 106,617 283,509 590,145 >30 15,892 64,987 203,793 487,292 911,783 source: ebri database; we thank jack van derhei for providing these calculations. o. s. mitchell et al. / financial services review 31 (2023) 1–21 5 the cost of the investment funds, they may also need to pay separately for advice to build their investment portfolios. results show that some participants in small, high-cost 401(k) plans might find less costly options, even if they are not financially sophisticated, by rolling into iras. in most cases, however, financially unsophisticated participants are likely to encounter lower costs by remaining in their employer plans rather than withdrawing their assets and moving them to iras. participants with low account balances are likely to do better by leaving their money in their employer plans instead of rolling over to iras. as this group is likely to be the least financially literate, it is more likely to need financial advice. nevertheless, due to low balances, these participants will have the least access to affordable advice outside employer plans. while robo-advice is available to participants with very low balances, such participants may be unaware of robo-options or lack the confidence to engage with them. for financially sophisticated participants, or those with larger balances, the situation is less clearcut. participants with large balances will be able to choose whatever form of advice they find most helpful. an attractive option for large balance participants who are in competitively priced employer plans is to leave assets that they do not need to finance current retirement spending in their employer plans. nevertheless, high-balance participants may have more sophisticated investment and advice needs, for which iras can provide a wider range of options. 4. financial advice in addition to the investment and account fees that retirement savers incur, plan participants also face costs if they need financial advice relating to their retirement portfolios. in this section, we discuss costs of financial advice provided by traditional human advisors, robo-advisors, and hybrid advisors, for 401(k) versus ira participants. in general, the table 2 gain or loss (in basis points) from rolling out of the employer plan into an ira: excludes explicit advice cost type of account plan size and fees average ira fees (57 bps) benchmark ira tdf fee (10 bps)a small 401(k) (<$10m) (104 bps) 47 94 medium 401(k) (<$250m) (48 bps) �9 38 large 401(k) (>$1bn) (22 bps) �35 12 megab (8 bps) �49 �2 source: authors’ calculations. notes: small, medium, and large 401(k) fee data from brightscope/investment company institute (2020). 401(k) fees include investment and administration fees for plans using mutual funds. ira investment fees from table 1. abenchmark ira tdf fee based on average of vanguard retail tdf fee (8 bps) and fidelity freedom index tdf fee (12 bps). bvery large “mega” plans using passive cits may have significantly lower costs: for example, one large plan with $4.5 bn in total assets offers tdfs for 6.5 bps + approx. $50 p.a. admin costs, which on our median balance of $200,000 would be the equivalent of 8 bps of total cost. 6 o. s. mitchell et al. / financial services review 31 (2023) 1–21 evidence shows that participants with higher account balances are most likely to use professional financial advisors: 49% of retirement savers with investable assets over $500,000 use advisors, compared with only 16% of those having under $50,000 (cerulli associates, 2021a). 4.1. financial advice options for 401(k) participants while 401(k) sponsors do not generally provide actual investment advice, in practice, participants in 401(k) plans often receive implicit financial advice from their plan providers at no additional cost beyond the investment and administrative charges they pay on their retirement savings. one way in which this occurs is that plan sponsors select the investment menus for the offered plans. these are often interpreted by employees as firm-provided advice (mitchell & utkus, 2022). additionally, most 401(k) plans have a default investment option, typically a target date fund suite, for participants who do not make active investment choices. according to a large retirement plan provider, target date funds are offered by nearly 90% of employer-sponsored defined contribution plans, such as 401(k) plans (finra, 2022). under u.s. law, plan providers have a fiduciary duty to provide investment options in participants’ best interests, in terms of diversification, risk, expected return, fees, and the choice of investment options. this fiduciary oversight provides 401(k) plan participants a level of protection that does not exist in the typical ira setting. sponsor fiduciary duty has been vigorously enforced through a series of recent lawsuits arising when 401(k) participants have questioned the level of fees and the quality of investment options available (turner, 2021). such lawsuits are likely to have been a factor in 401(k) fees declining more rapidly over time, compared with average fees for mutual funds, as is evident in table 3. the 401(k) participants who wish for more customized portfolios than those provided by plan sponsor default options may decide to pay for additional advice regarding their retirement plan investments. one increasingly widespread option is a managed account, where the plan’s available investment options are combined with employee attributes such as risk tolerance, salary, or outside assets, to construct and manage a personalized investment portfolio. fees for managed accounts vary but are often in the range of 25–35 bps, in addition to the cost of the underlying investments (cerulli associates, 2021b). alternatively, participants can hire their own external advisors—either human or robo—to help them choose from the investment options in their 401(k) plans. external advisors can also give advice on participants’ non-401(k) assets. nevertheless, participants within a 401(k) plan who use table 3 asset-weighted equity mutual fund expense ratios: 2010 and 2020 equity mutual fund market segment 2010 expense ratio (bps) 2020 expense ratio (bps) ratio of 2020 to 2010 (percent) overall 83 50 60.2 401(k) plans 70 39 55.7 ira 85 57 67.0 source: ici (2021a, 2021b) and authors’ calculations. o. s. mitchell et al. / financial services review 31 (2023) 1–21 7 external advisors will need to implement these recommendations themselves, and advisors can only help them select from the investment options available on the plan menu (or through a plan brokerage window, if available). 4.2. financial advice options for ira participants if participants wish to roll their assets out of employer plans into iras, they must first select an ira custodian in the form of a bank, brokerage firm, or insurance company; alternatively, participants can go directly to a mutual fund company (laponsie, 2022). because ira managers offer a wide range of investments from which participants can select, with diverse fees and charges, the choice of investment products is an important decision, requiring a level of knowledge not typically demanded of 401(k) participants. ira participants can also avail themselves of other advice strategies to help them make their investment choices. for instance, they can choose a do-it-yourself (diy) approach, going completely on their own when selecting an ira provider, and then selecting from among the investment funds available. to construct a very low-cost portfolio, a financially sophisticated participant could take advantage of the quite inexpensive index funds available on many ira platforms: for instance, s&p 500 index funds have fees starting below 2 bps, and passive target date funds are available starting at 8 bps (dehaan et al., 2021; vanguard, 2022). nonetheless, financially illiterate participants could be less likely to find these products and hence be prone to making suboptimal decisions. ira investors can also choose to pay separately for advice, for instance by hiring an independent financial advisor; nevertheless, for people with low account balances, this option is unlikely to be available. an alternative would be to access a robo-advisor, a relatively inexpensive advice option that is also available to individuals with low account balances. still another option is that ira holders can choose a hybrid approach, where a robo-advisor also offers some contact with a human advisor. such hybrid services tend to be intermediate in cost between pure roboand human advisors.5 if retirement savers roll assets over from 401(k) plans where they did not need advice or were able to access advice through managed account services, to iras, where they do require advice concerning choice of provider and/or investments, the cost difference can be substantial, particularly if the starting point is a low-cost, large employer plan (see table 4). robo-advisors in the united states generally charge an annual advice fee of around 25 bps, in addition to the cost of the underlying investment options. these often require no or very low minimum account balances (fisch et al., 2019). accordingly, robo-advisors can suit clients who otherwise might not have access to affordable financial advice. some robo-advisors provided by mutual fund companies may not even charge separately for advice, though, in turn, they limit investment choices to the fund options offered by the parent companies. in addition to pure robo-advisors, many companies have launched hybrid advice options. these combine digital advice and algorithmic portfolio construction with some access to a human advisor, often in the form of financial planning services offered via a certified financial planner (cfp). hybrid advisors usually require higher minimum account balances of 8 o. s. mitchell et al. / financial services review 31 (2023) 1–21 t ab le 4 c o st co m p ar is o n s fo r d if fe re n t ty p es o f ad v ic e 4 0 1 (k ): m an ag ed ac co u n ta ir a : r o b o -a d v is o rb ir a : h y b ri d ad v is o rc ir a : h u m an ad v is o r a d v ic e fe e 2 0 – 2 5 b p s 1 5 – 2 5 b p s 3 0 – 4 0 b p s ; 1 0 0 b p s a ss et m in im u m $ 5 ,0 0 0 – $ 2 5 ,0 0 0 $ 0 – $ 3 ,0 0 0 $ 5 0 ,0 0 0 – $ 1 0 0 ,0 0 0 u su al ly at le as t $ 1 0 0 ,0 0 0 in v es tm en t fe e s am e co st as em p lo y er p la n in v es tm en t o p ti o n s t y p ic al ly u se s p as si v e in v es tm en ts as b u il d in g b lo ck s, ar o u n d 1 0 b p s t y p ic al ly u se s p as si v e in v es tm en ts , ar o u n d 1 0 b p s in v es tm en ts se le ct ed b y ad v is o r, m ay al so co n ta in ac ti v e fu n d s a cc es s to h u m an ad v is o r s o m e ac ce ss n o s o m e ac ce ss y es n o te s: a m an ag ed ac co u n t b as ed o n f id el it y m an ag ed ac co u n t at 2 5 b p s fo r as se ts u n d er $ 2 0 0 ,0 0 0 an d 2 0 b p s fo r as se ts ex ce ed in g $ 2 0 0 ,0 0 0 . b r o b o ad v is o r b as ed o n av er ag e o f v an g u ar d (2 0 2 2 ) d ig it al ad v is o r (e st im at ed 1 5 b p s n et ad v is o ry fe e) an d b et te rm en t (2 5 b p s ad v is o ry fe e) p lu s 1 0 b p s as su m ed co st o f in v es tm en t fu n d s. c h y b ri d ir a b as ed o n av er ag e o f v an g u ar d p er so n al a d v is o r s er v ic es (2 0 2 2 ) (3 0 b p s) an d b et te rm en t p re m iu m (4 0 b p s) p lu s 1 0 b p s as su m ed fo r in v es tm en ts (b et te rm en t, 2 0 2 2 ). o. s. mitchell et al. / financial services review 31 (2023) 1–21 9 $50,000–$100,000 and charge an annual 30–40 bps for advice, in addition to the cost of the underlying investments. most robo assets are invested with hybrid advisors (ponte, 2022). human financial advisors typically charge around 100 bps per year for advice on assets under management, in addition to the cost of the underlying investments (advisoryhq, 2021). these financial advisors may require clients to have minimum investable assets of $100,000 or more (ludwig, 2017). some advisors have started to offer one-time advice sessions to savers for a flat fee, rather than charging a percentage of assets under management. in such cases, the advisor can help savers create financial plans, but savers will typically need to implement them on their own. 5. employer and financial advisor perspectives for various reasons, some plan sponsors could prefer that retirees remain in their employer-sponsored plans. one is that retaining assets increases the pool of investable funds, which may allow sponsors to negotiate lower fees from retirement plan service providers. to this point, some employers have already expressed increased interest in participant retention postretirement. a recent survey documented that 69% of all plan sponsors and 84% of those with assets over $500 million stated that they favored retaining participant assets in the company’s 401(k) plan after retirement (cerulli associates, 2021a). additionally, asset managers are starting to develop target date suites that convert participants’ accumulated balances into income streams after retirement, and some include embedded annuity products (blackrock, 2021; state street global advisors, 2021). in addition to the advantages of retaining scale, plan sponsors can offer these integrated products to give participants greater confidence that their money will not run out in retirement. nevertheless, employers do continue to bear fiduciary liability for retirees who remain in 401(k) plans. this concern can now be alleviated from 2021, if employers join a pooled employer plan (pep), as the pep can assume most of the fiduciary responsibility for the plans. so far, however, there has been limited uptake of these plans (morningstar, 2022). another potential retention cost relates to firms’ requirement to continue communicating with retirees as plan participants. in most cases, however, communicating with participants is delegated to recordkeepers (erisa advisory council, 2020), so such communication is unlikely to be very costly. financial advisors also have views as to whether 401(k) participants should roll over to iras, depending on whom they work for. if a participant’s advisor is also the plan’s advisor, either directly or through a managed account, the advisor is compensated for managing the account within the 401(k) plan and can allocate the participant’s assets across the plan’s investment options. independent advisors unaffiliated with employers’ plans will not usually receive asset-based fees on assets retained within the 401(k) plans, nor will they be able to manage the assets directly. additionally, advisors may wish to use investment options unavailable on the retiree’s plan menu, or they may wish for more flexible distribution options than the 401(k) platform offers. for these reasons, advisors often prefer that participants roll their assets to iras after retirement. nevertheless, as managed account solutions 10 o. s. mitchell et al. / financial services review 31 (2023) 1–21 become more sophisticated, and more employer plans add investment products and other functionality specifically designed for the retirement phase, advisors may find it increasingly feasible and attractive to manage retirement assets within employer plans. 6. evaluating participants' decisions to roll over next, we outline key features of iras and 401(k) plans that may encourage or discourage employees from rolling over their assets from their 401(k) plans at retirement. 7. why roll over to an ira? several features of iras may help explain why retirees may move their retirement assets out of their 401(k) plans. these include: 7.1. greater awareness of iras as a postretirement solution traditionally, participants have rolled their assets out of their employer plans into iras upon retirement. although surveys show that increasing numbers of employers are expressing an interest in retaining participant assets post-retirement, some employers may not be actively promoting this option. by contrast, ira providers do actively market their services through multiple channels. in some cases, participants may be unaware that they could use their employer plans to pay them a stream of income in retirement. and in other cases, retirees might not wish to have their retirement assets tied to former employers and, therefore, they would prefer to roll the money into their personal iras after leaving their firms. 7.2. additional investment choices some see the chance to access a wider range of investment choices as an advantage of iras, compared with 401(k) plans. in 2017, the average large 401(k) plan offered 28 investment options, of which 13 were equity funds, three were bond funds, and eight were target date funds. moreover, more than four-fifths of plans offered target date funds (brightscope/ investment company institute, 2020). by contrast, iras can provide access to thousands of different investment funds, and they may also offer access to individual stock and bond purchases (porter, 2021). offsetting that argument, substantial research has shown that offering fewer choices may be better for people with low financial literacy, due to mental overload associated with many choices (carosa, 2011; iyengar & lepper, 2000). specifically, offering too many investment options in 401(k) plans can reduce participation rates (iyengar et al., 2004). moreover, shen and turner (2018) found that participants do not need very many investment options to be adequately diversified. this was specifically the case in the thrift savings plan (2021) that offered only five diversified funds to federal employees, the military, and members of o. s. mitchell et al. / financial services review 31 (2023) 1–21 11 congress.6 furthermore, offering a wide range of investment options can increase the chances that participants will make risky investment choices that could endanger their retirement security. nowadays, some iras allow participants to invest in stock options (parker, 2022), cryptocurrency (bitcoinira, 2022), and gold (best, 2022). it is unlikely that most 401(k) participants would benefit from such highly risky assets that require substantial financial sophistication to manage effectively. offsetting the appeal of more funds available in iras versus 401(k) plans, 401(k)s can also offer investments frequently unavailable in iras. specifically, 401(k) plans can offer less expensive institutionally priced mutual funds and collective investment trusts, and they may also include stable value funds and guaranteed investment contracts not usually available outside employer-provided plans. many large 401(k) plans also offer brokerage windows through which participants can access most of the investment options available on providers’ ira platforms.7 7.3. account consolidation another rationale for rolling assets into iras after retirement may be participants’ desire for simplification. although it is usually possible to transfer assets from one 401(k) plan to another if the saver continues to work, the process can be laborious. by contrast, rolling assets from a 401(k) to an ira is relatively easy at retirement. as an example, when a retiree has accumulated assets in several 401(k) plans over a long career, consolidating them all into a single ira could be an attractive option. unsurprisingly, participants who have existing iras are more than twice as likely (41%) to roll their assets into iras when leaving their employers, compared with participants who did not previously have iras (19.5%; the pew charitable trusts, 2021). 7.4. more flexible access to savings while almost all 401(k) plans allow participants to stay in their employer plans after retirement, many still place limitations on how often participants can access their savings. according to one large recordkeeper (vanguard, 2021b), 80% of the participants on its platform are in plans that offer installment payments other than required minimum distributions (rmds), yet only14% are in plans that offer annuities. additionally, 71% of all participants have access to some level of ad hoc partial distributions, although this feature is most prevalent in the very largest plans. plans can also limit the number of partial distributions available to participants, and recordkeepers may charge separately for each distribution. for these reasons, participants who need to use their savings to fund spending needs in early retirement, rather than preserving them as a contingency fund, may find the distribution options offered by 401(k) plans overly restrictive. this is most likely to be the case for low balance participants. indeed, according to one survey (the pew charitable trusts, 2021), only 24% of participants with balances between $5,000–$25,000 left their assets in their 401(k) plans when leaving their employers, compared with 36.5% of those with balances over $100,000. among low balance participants, 33% rolled their assets to iras, versus 12 o. s. mitchell et al. / financial services review 31 (2023) 1–21 28% of participants with assets over $100,000. the biggest difference was in the share of lump sum withdrawals: 30% of low balance participants took lump sum withdrawals, compared with only 8% of participants with assets over $100,000.8 7.5. annuities there is increasing interest among policy analysts and researchers in ways to help protect retirement savers against longevity risk (the risk associated with outliving their assets in old age). a natural way to insure against this risk is to include annuities in 401(k) lineups that promise to pay income checks for life. to date, however, relatively few 401(k) plans have offered annuities, though some providers are beginning to integrate these insurance products into target date funds (dierking, 2017). one large plan recordkeeper reported that 14% of its plans with 5,000+ participants and 12% of all its plans offered annuities (vanguard, 2021b). also, recently enacted and proposed legislation such as the secure act and the secure 2.0 act, include provisions making it easier for 401(k) plan sponsors to offer annuities. the reality is that ira holders can purchase annuities more readily than 401(k) participants can today. nevertheless, the take-up of annuity products in the united states remains low. part of the reason may be that to gain an accurate understanding of how annuities work and appreciate the benefits of owning them requires substantial financial literacy and active education by the advisor, on the benefits of owning an annuity (brown et al., 2017, 2021). 7.6. roth conversion while this paper focuses on differences in fees, investment options, access to advice, fiduciary protection, and disbursement options, another issue for some participants may be the tax treatment of the different accounts. a retiree with a traditional 401(k) plan may decide that he or she would prefer to roll over to a roth ira since there are no required minimum distributions, and the retiree’s assets can continue to grow with preferential tax treatment. moreover, the beneficiary (usually the participant’s heirs) may have no tax liability when the retiree dies. a disadvantage is that taxes, which could be substantial, would be due at the time of the rollover, rather than being postponed and taken with rmds or other disbursements. 8. why remain in an employer plan? several features of 401(k) plans provide incentives for retirees to stay in employer plans, including the ones mentioned below. 8.1. fiduciary protection plan sponsors have a fiduciary duty to participants to manage the plans in participants’ best interests; moreover, sponsors can be held liable through private sector lawsuits or department of labor enforcement if they charge excessive fees or provide poorly o. s. mitchell et al. / financial services review 31 (2023) 1–21 13 performing investment options (turner, 2021). by contrast, department of labor enforcement does not apply to iras, and lawsuits concerning iras are rare because they are conducted on an individual basis, while 401(k) lawsuits are conducted on a class basis. the investment menus and default investment options in 401(k) plans are selected and overseen by erisa fiduciaries charged with the responsibility to act in participants’ best interests. in an investigation of an employer investment fund screening, sialm et al. (2015) reported that 401(k) participants benefited from plan sponsors dropping poorly performing funds and adding stronger performers. the sec plays an important role in investor protection in the retail space, regulating the products that can be sold to retail investors. by contrast, iras do not provide participants with a default investment or an investment menu selected by a fiduciary. this is particularly concerning when persons with low or moderate levels of financial literacy lack the expertise to successfully implement a do-it-yourself approach to constructing portfolios or to seek out the most cost-effective ira options. as a result, they must either pay for somebody to help them construct a portfolio, or they will be likely to construct portfolios with inappropriate risk levels and probably excessive costs. 8.2. less need for advice financially illiterate participants are more likely to need financial advice if they roll over to an ira than if they remain in their employer’s plans. one reason is that most target date funds offered by 401(k) plans convert into balanced funds after retirement, with a risk profile deemed suitable for most retirees. therefore, many 401(k) participants will achieve a diversified investment portfolio without financial advice, relying on the default options preselected by employers. the employer-selected default options are likely to be particularly valuable for low balance holders who lack financial literacy, as these individuals will tend to find it difficult to access affordable advice outside their employer plans. 8.3. institutional pricing on investments and advice many retirees who remain in the larger employer plans will benefit from institutional pricing for investment products, which tends to be lower than average ira pricing. also, many employer plans offer managed account services that enable participants to receive advice to create investment portfolios based on the plan menu that considers a wider range of inputs, compared with purely retirement date-based target date funds. the average fees paid for investments in iras are higher than those in 401(k)s, which may imply that most participants do not seek the least expensive products. pension participants surveyed by turner and korczyk (2004) were asked if they knew how much they paid in fees (in dollars or as a percentage of assets) for the stock mutual funds they held in their 401(k) plans, and that study concluded that three-quarters of respondents could not answer the question. hastings et al. (2011) also showed that people with greater financial knowledge paid lower fees for mutual funds. it may also indicate that ira participants need advice for 14 o. s. mitchell et al. / financial services review 31 (2023) 1–21 portfolio construction, and they are paying indirectly for this advice in the form of higher fees for investment products. 8.4. protection from creditors a participant’s 401(k) plan assets are protected from creditors under u.s. law, while ira money is not exempt when a person files for bankruptcy (there is some variability in state law, as noted in folger, 2021). 8.5. early access people who retire from employers who offer 401(k) plans may access their 401(k) assets at age 55 without paying a 10% penalty, compared with age 59½ in an ira. 9. how new technology and legislation could change the balance employer-sponsored 401(k) plans were originally intended to be supplementary retirement savings vehicles to complement conventional defined benefit plans. nonetheless, with defined benefit plans disappearing from the private sector, fewer americans today have access to guaranteed retirement income from their retirement savings, compared with the past. there has recently been increased interest among plan sponsors, legislators, and asset managers in keeping participants in employer plans after retirement, but so far, these have gained limited traction due to both legal and technological hurdles. recent and pending legislation, as well as technological advances, may change this situation going forward. legislators have introduced several bills with provisions that make it easier for 401(k) plan sponsors to offer products allowing participants to convert their defined contribution savings into retirement income stream. as noted above, the 2019 secure act gave plan sponsors a safe harbor for annuity selection by allowing them to rely on state insurance regulators’ assessments of the creditworthiness of insurance providers. it also included provisions permitting easier portability of annuity products. the proposed life act of 2022 (norcross, 2022) would allow qualified default investment alternatives (qdias) to include allocations to illiquid and potentially variable annuities in 401(k) accounts as well. the securing a strong retirement act (2021) that recently passed the house of representatives includes additional provisions facilitating the inclusion of annuities in 401(k) plans, including raising the percentage and potentially the asset limit that can be used for qualified life annuity contracts (qlacs) and expanding the range of products that can be used. these innovations will have limited effect if the problem is lack of demand for annuities from participants, rather than a lack of easily accessible supply. the sec’s (2019) “regulation best interest: the broker-dealer standard of conduct” requires advisors and brokers to act in the best interest of clients when considering rollovers from employer plans to iras. the department of labor (2021) also signaled its intention to apply more stringent regulation to advice on potential rollovers from 401(k) plans to iras. o. s. mitchell et al. / financial services review 31 (2023) 1–21 15 as noted above, participants in small, high-fee plans might benefit the least by remaining in their employer’s plans after retirement. nevertheless, since early 2021, u.s. employers have gained the opportunity to join pooled employer plans (peps) that permit otherwise unrelated employers to use the same retirement plan while offloading most administrative and fiduciary responsibility to the pooled plan provider (ppp). so far, peps have gained limited traction, but as they gain scale, they could aggregate multiple small or medium-sized employer plans into one larger plan. this could significantly reduce costs and improve the rationale for staying in the employer-sponsored plan after retirement. technology has so far been one of the obstacles to keeping participants in employer plans after retirement. this is because 401(k) recordkeeping systems are designed primarily for the accumulation phase, and they have traditionally not offered participants much flexibility in drawing down assets. the options for receiving advice within employer-sponsored retirement plans have also been limited. this is now starting to change, however, thanks to partnerships between policymakers, asset managers, advisors, insurers, and technology providers. for instance, policymakers have recommended adding annuity products to retirement plans (irs notice 2014-66; iwry & turner, 2009) and asset managers are beginning to embed annuity products into target date funds, the most used default investments in 401(k) plans (blackrock, 2021; state street global advisors, 2021). insurers, recordkeepers, and asset managers are partnering to deliver new retirement income solutions (pechter, 2022). managed account providers are also partnering with insurance hubs to provide retirement planning advice and annuity offerings within employer plans (investment news, 2022). the technology is improving, and managed accounts are expanding to provide additional options for the postretirement phase. this is likely to make it increasingly attractive for advisors to manage their clients’ assets within 401(k) plans. 10. conclusions and implications for the decumulation period this paper has examined the question of how individuals should think about the choice of retaining their retirement assets in their 401(k) plans after retirement versus rolling them over to iras. our particular focus is on people with low or moderate levels of financial literacy, who comprise much of the older workforce. against this backdrop, we analyze the choices for retirees most likely to make mistakes concerning rollovers. we conclude that, for people in plans with high fees, which are mostly small plans and thus represent a relatively small group of people, rolling over to iras could be a sensible financial decision. yet most people with low or moderate levels of financial literacy will do better financially by remaining in their employer 401(k) plans. an important reason is that 401(k) plans, through the use of defaults, are more likely to provide retirees more appropriate investment portfolios at lower cost compared with ira accounts. even though most 401(k) plans are likely to offer lower fees in retirement than iras, many retirees still do roll their assets over to iras. one reason is that the advantages of better access to advice, easier account consolidation, and greater flexibility of withdrawals 16 o. s. mitchell et al. / financial services review 31 (2023) 1–21 offered by iras can be seen to outweigh the cost advantages of 401(k)s. another is that they may not understand that they could leave their assets in their employer plans after retirement, as plan sponsors may not actively promote this option. by contrast, ira providers continue to actively market their services. we suggest that this may be changing, as legislators make it easier to include income products in 401(k) plan menus. moreover, asset managers, managed account providers, insurers, and recordkeepers are beginning to partner with new technology providers to make it easier for retirees and their advisors to continue using their 401(k) plans to manage assets in retirement. both 401(k) plans and iras must evolve if they are to provide better options for participants in the decumulation phase of life. fornia and doonan (2022) argue that a typical defined benefit (db) plan with longevity risk pooling, lower fees, and professional money management can deliver almost twice as much income per dollar invested as a typical dc plan, with four-fifths of this advantage stemming from more efficient management of the post-retirement phase. incorporating longevity risk pooling into the management of dc assets in the retirement phase could enable many dc plans to adopt attractive features of db plans while avoiding employer liability as well as the funding and portability problems that the latter have experienced. one new approach proposed by fullmer and turner (2022) would be to enable robo-advisors to provide tontines. these involve financial risk pools where participants “mutually and irrevocably agree to receive payouts while living and share the proceeds of their accounts upon death” (fullmer & sabin, 2018). longer term, retaining assets in employer plans after retirement could usher in a new era of collectively managed investment solutions for retirees that could help plan sponsors “put the pension back” into dc plans (horneff et al., 2020; iwry et al., 2021). ultimately, and most importantly, this could yield better outcomes, more retirement confidence, and greater security for retirees. notes 1 for instance, only 34% of older americans over age 50 could correctly answer the “big three” questions on interest rates, inflation, and stock risk posed in the 2004 health and retirement study (lusardi & mitchell, 2014). 2 there is also a suboption, where participants who retire with benefits at more than one workplace plan may choose to consolidate all their benefits into one of these workplace plans. 3 generally, the amounts an individual withdraws from an ira or retirement plan before reaching age 59½ are called “early” distributions. individuals must pay an additional 10% early withdrawal tax unless an exception applies. one exception is that employees in 401(k) plans who leave their job after age 55 are not subject to the penalty tax (internal revenue service, 2021). 4 while we refer specifically to 401(k) plans, our analysis applies generally to employer-sponsored defined contribution plans, including money purchase, 403(a), 403(b), and governmental 457(b) plans, and the federal thrift savings plan. 5 some robo-advisors have also started to provide advice concerning the decumulation phase, though agnew and mitchell (2019) and fisch et al. (2019) report that much remains to be done in this regard. o. s. mitchell et al. / financial services review 31 (2023) 1–21 17 6 a similar point was made by tang et al. (2010) for a large sample of private sector 401(k) plans. turner et al. (2016) investigate issues relating to rollovers from the thrift savings plan, and turner and klein (2014) study rollovers from 401(k) plans. 7 nevertheless, one large provider reported that very few (1%) of the 30% of participants provided access to brokerage windows through their plans actually used them (vanguard 2021a). 8 these numbers refer to all distributions on termination of employment, not just termination at retirement. some employers may force participants with balances below $5,000 to make an immediate distribution into an ira at termination. acknowledgments this study received research support from the pension research council/boettner center at the wharton school of the university of pennsylvania. for useful comments, we thank without implicating peter brady and mark iwry; reilly also acknowledges helpful collaboration and input from jodan ledford, ceo of smart usa co. opinions and conclusions expressed herein are solely those of the authors and do not represent the opinions or policy of any institutions with which the authors are affiliated. references advisoryhq. (2021). average advisor fees in 2021: everything you need to know. available at https://www. advisoryhq.com/articles/financial-advisor-fees-wealth-managers-planners-andfee-only-advisors/ agnew, j. & mitchell, o. s. (eds.). (2019). the disruptive impact of fintech on retirement systems. oxford: oxford university press. best, r. (2022). best gold ira companies. investopedia. available at https://www.investopedia.com/best-goldira-companies-5087720 betterment. (2021). explore our pricing. available at https://www.betterment.com/401k/plan-pricing/ betterment. (2022). pricing at betterment. available at https://www.betterment.com/pricing?utm_content= brand&utm_medium=sem&&lid=437000 32812145085&ds_s_kwgid=58700004022642666&ds_a_cid= 405186663&ds_a_caid=92013 89213&ds_a_agid=91383124217&ds_a_lid=kwd-305491188865&utm_ source=bing&utm_campaign=&utm_medium=cpc&utm_keyword=%2bwww.betterment.com&utm_ matchtype=p&utm_creative=&utm_term=%2bwww.betterme nt.com&msclkid=69f9548d192912b9cbfa350d24467855& utm_content=www.betterment.co m&gclid=clc7j83w9_mcfsugiakdzbegmg&gclsrc=ds bitcoinira. (2022). invest in bitcoin tax-free with your ira. available at https://bitcoinira.com/buy-bitcoin/? aid=701&keyword=bitcoin%20ira&network=g&device=c&placement=&adgroup=126567578382&campaign= 15147644605&creative=580350434104&adp=&gclid=cjwkcai a1jgrbhbseiwaxxblwtfzjzbdbd7u uxddcpyk3a1tqccmlsa6klfxm7g0m7tyq6dn vnqcnroc6fyqavd_bwe blackrock. (2021). blackrock’s retirement income solution: a major step forward for american retirement security. available at https://ir.blackrock.com/news-and-events/press-releases/press-releases-details/2021/ blackrocks-retirement-income-solution-a-major-step-forward-for-american-retirement-security/default. aspx brightscope/investment company institute. (2020). the brightscope/ici defined contribution plan profile: a close look at 401(k) plans, 2017. available at https://www.ici.org/system/files/attachments/pdf/20_ppr_ dcplan_profile_401k.pdf brown, j. r., kapteyn, a., luttmer, e., & mitchell, o. s. (2017). cognitive constraints on valuing annuities. journal of the european economic association, 15, 429–462. https://doi.org/10.1093/jeea/jvw009 brown, j. r., kapteyn, a., luttmer, e., mitchell, o. s., & samek, a. (2021). behavioral impediments to valuing annuities: evidence on the effects of complexity and choice bracketing. the review of economics and statistics, 103, 533–546. https://doi.org/10.1162/rest_a_00892 18 o. s. mitchell et al. / financial services review 31 (2023) 1–21 bureau of labor statistics. (2021). 67 percent of private industry workers had access to retirement plans in 2020. the economics daily: u.s. bureau of labor statistics (bls.gov) carosa, c. (2011). how many investment options should 401(k) plan sponsors offer? fiduciary news, october 18. cerulli associates. (2021a). us retirement end investor 2021, solving for the decumulation phase. available at https://www.cerulli.com/research-areas/us-retirement cerulli associates. (2021b). the cerulli edge: us retirement edition 1q 2021, edition #58. available at https:// www.cerulli.com/research-areas/us-retirement dehaan, e., song, y., xie, c., & zhu, c. (2021). obfuscation in mutual funds. journal of accounting and economics, 72, 101429. https://doi.org/10.1016/j.jacceco.2021.101429 dierking, d. (2017). target-date funds investing in annuities may provide guaranteed income. mutualfunds. com, july 4. available at https://mutualfunds.com/retirement-channel/target-date-funds-investing-annuitiesprovide-guaranteed-income/. department of labor. (2021). temporary enforcement policy on prohibited transaction rules applicable to investment advice fiduciaries. field assistance bulletin 2021-02. duvall, j. (2021). ira investors are concentrated in lower-cost mutual funds. investment company institute. available at https://www.ici.org/system/files/file/2021-07/21_view_irafees_print.pdf employee benefits security administration. (2021). 2018 form 5500 annual reports abstract. available at https://www.dol.gov/sites/dolgov/files/ebsa/researchers/statistics/retirement-bulletins/private-pension-planbulletins-abstract-2018.pdf erisa advisory council. (2020). considerations for recognizing and addressing participants with diminished capacity. available at https://www.dol.gov/sites/dolgov/files/ebsa/about-ebsa/about-us/erisaadvisory-council/2020-considerations-for-recognizing-and-addressing-participants-with-diminishedcapacity.pdf finra. (2022). target-date funds—find the right target for you. available at https://www.finra.org/ investors/learn-to-invest/types-investments/retirement/target-date-funds-find-right-target-you#:;:text=one% 20increasingly%20popular%20investment%20option,as%20401(k)%20plans fisch, j. e., labouré, m., & turner, j. a. (2019). the emergence of the robo-advisor. in j. agnew & o. s. mitchell (eds.). the disruptive impact of fintech on retirement systems (pp. 13–37). oxford: oxford university press. folger, j. (2021). which retirement funds are protected from creditors? investopedia. available at https:// www.investopedia.com/articles/personal-finance/040716/which-retirement-funds-are-protected-creditors.asp. folger, j. (2022). what roth ira fees do i pay? investopedia. available at https://www.investopedia.com/ what-types-of-roth-ira-fees-do-i-pay-4771122#:;:text=some%20roth%20ira%20providers%20charge,% 2425%20and%20%245 0%20per%20year fornia, w. b., & doonan, d. (2022). a better bang for the buck 3.0: post-retirement experience drives pension cost advantage.washington, dc: national institute on retirement security. fullmer, r., & sabin, m. (2018). individual tontine accounts – yes, seriously! journal of retirement. available at https://retirementincomejournal.com/article/individual-tontine-accounts-yesseriously/ fullmer, r. k., & turner, j. a. (2022). robo tontines. journal of retirement, 10, 63–73. general accountability office (gao). (2021). many participants do not understand fee information, but dol could take additional steps to help them. available at https://www.gao.gov/assets/gao-21-357.pdf hastings, j. s., mitchell, o. s., & chyn, e. t. (2011). fees, framing, financial literacy in the choice of pension managers. in o. s. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 101–115). oxford: oxford university press. horneff, v., maurer, r., & mitchell, o. s. (2020). putting the pension back in 401(k) plans: optimal retirement plan design with longevity income annuities. journal of banking & finance, 114, 105783. https://doi.org/ 10.1016/j.jbankfin.2020.105783 institutional retirement income council. (2021). delivering additional income through lower-cost products and services available in the defined contribution industry. available at https://tinyurl.com/3fxnatus o. s. mitchell et al. / financial services review 31 (2023) 1–21 19 internal revenue service. (2021). retirement topics—exceptions to tax on early distributions. available at https://www.irs.gov/retirement-plans/plan-participant-employee/retirement-topics-tax-on-early-distributions investment company institute. (2021a). the economics of providing 401(k) plans: services, fees, and expenses. available at https://www.ici.org/system/files/2021-06/per27-06.pdf investment company institute. (2021b). release: quarterly market data. available at https://www.ici.org/ statistical-report/ret_21_q1 investment news. (2022). morningstar partners with hueler on annuities. available at https://www. investmentnews.com/morningstar-partners-with-hueler-on-annuities-217445 ira reviews. (2021). the largest ira account provider companies in 2021. available at https://www.irareviews.com/ira/largest-ira-companies.aspx irs notice 2014-66. lifetime income provided through target date funds in section 401(k) plans and other qualified defined contribution plans. available at: https://www.irs.gov/pub/irs-drop/n-14-66.pdf iwry, j. m., john, d. c., pulliam, c., & gale, w. g. (2021). collective defined contribution plans. brookings institute. available at https://www.brookings.edu/research/collective-defined-contribution-plans/ iwry, m., & turner, j. (2009). automatic annuitization: new behavioral strategies for expanding lifetime income in 401(k)s. the retirement security project. iyengar, s. s., & lepper, m. (2000). when choice is demotivating: can one desire too much of a good thing? journal of personality and social psychology, 79, 995–1006. https://doi.org/10.1037/0022-3514.79.6.995 iyengar, s. s., jiang, w., & huberman, m. g. (2004). how much choice is too much? contributions to 401(k) retirement plans. in o. s. mitchell & s. utkus (eds.), pension design and structure: new lessons from behavioral finance (pp. 83–95). oxford: oxford university press. laponsie, m. (2022). 7 ira accounts to consider for 2022. us news, jan 7. available at https://money.usnews. com/money/retirement/articles/best-ira-account-providers ludwig, l. (2017). the rise of the robo-advisors – should you use one? investorjunkie. available at https:// investorjunkie.com/35919/robo-advisors/ lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. https://doi.org/10.1257/jel.52.1.5 lusardi, a., mitchell, o. s., & curto, v. (2014). financial literacy and financial sophistication in the older population. journal of pension economics & finance, 13, 347–366. https://doi.org/10.1017/s1474747214000031 manganaro, j. (2019). recordkeeping fees under the microscope. planadvisor, november 17. available at https://www.planadviser.com/recordkeeping-fees-microscope/. manganaro, j. (2021). lack of revenue sharing questioned in erisa lawsuit. plansponsor, august 17. https:// www.plansponsor.com/lack-revenue-sharing-questioned-erisa-lawsuit/ mitchell, o. s., & utkus, s. p. (2022). target date funds and portfolio selection in 401(k) plans. journal of pension economics & finance, 21, 519–536. morningstar. (2022). retirement plan landscape report: an in-depth look at the trends and forces reshaping u.s. retirement plans. center for retirement & policy studies. norcross, d. (2022). news releases. available at https://norcross.house.gov/media-center/pressreleases/repsnorcross-walberg-reintroduce-legislation-improve-retirement parker, m. (2022). can you put stock options in an ira account? the nest. available at https://budgeting. thenest.com/ira-rules-selling-stocks-reinvesting-24947.html pechter, k. (2022). ‘income america’ offers ‘5forlife.’ retirement income journal. available at https:// retirementincomejournal.com/article/income-america-offers-5forlife/ the pew charitable trusts. (2021). how employees handle their retirement savings during work transitions. available at https://www.pewtrusts.org/en/research-and-analysis/issue-briefs/2021/05/how-employees-handle-theirretirement-savings-during-work-transitions ponte, b. (2022). vanguard manages two-thirds of robo assets. ignites. available at https://www.ignites.com/ c/3559244/455664 porter, t. j. (2021). ira vs. 401(k) differences – which retirement plan is better? moneycrashers.com. available at https://www.moneycrashers.com/ira-vs-401k-differences-retirementplan-better/ 20 o. s. mitchell et al. / financial services review 31 (2023) 1–21 sec. (2019). regulation best interest: the broker-dealer standard of conduct. available at https://www.finra. org/rules-guidance/key-topicis/regulation-best-interest securing a strong retirement act 2021. available at: https://waysandmeans.house.gov/wp-content/uploads/ 2021/05/secure-2.0-section-by-section-5.3.21-1.pdf schimmer, p. (2021). fees for 401(k) services: what plan sponsors need to know. harris. available at https:// harriscpas.com/fees-for-401k-services-what-plan-sponsors-need-to-know/. shen, s., & turner, j. a. (2018). conflicted advice about portfolio diversification. financial services review, 27, 47–82. sialm, c., starks, l., & zhang, h. (2015). defined contribution pension plans: sticky or discerning money? the journal of finance, 70, 805–838. https://doi.org/10.1111/jofi.12232 state street global advisors. (2021). addressing longevity risk. available at https://www.ssga.com/us/en/ institutional/ic/insights/addressing-longevity-risk szala, g. (2022). vanguard cuts target date fund fees. thinkadvisor, february 14. available at https://www. thinkadvisor.com/2022/02/14/vanguard-cuts-target-date-fund-fees/ tang, n., mitchell, o. s., mottola, g., & utkus, s. (2010). the efficiency of sponsor and participant portfolio choices in 401(k) plans. journal of public economics, 94, 1073–1085. https://doi.org/10.1016/j.jpubeco.2010.06.016 thrift savings plan. (2021). administrative and investment expenses. available at https://www.tsp.gov/tspbasics/administrative-and-investment-expenses/. turner, j. a. (2021). pension litigation as a way of improving participants’ outcomes: investments and fees. chapter 2. in new york university review of employee benefits and executive compensation (eds) nyu school of continuing and professional studies and d. pratt. new york: lexisnexis. pp 2–1 to 2–15. turner, j. a., & klein, b. w. (2014). retirement savings flows and financial advice: should you roll over your 401(k) plan? benefits quarterly, 30, 42–54. turner, j. a., klein, b. w., & stein, n. p. (2016). financial illiteracy meets conflicted advice: the case of thrift savings plan rollovers. the journal of retirement, 3, 47–65. https://doi.org/10.3905/jor.2016.3.4.047 turner, j. a., & korczyk, s. (2004). pension participant knowledge about plan fees. washington, dc: aarp data digest. available at http://assets.aarp.org/rgcenter/post-import/dd105_fees.pdf vanguard. (2021a). the brokerage option in dc plans. available at https://institutional.vanguard.com/iam/pdf/ cirbop_052021.pdf vanguard. (2021b). how america saves 2021. available at https://institutional.vanguard.com/content/dam/inst/ vanguard-has/insightspdfs/21_cir_has21_has_fsreport.pdf vanguard. (2022). vanguard digital advisor. available at https://investor.vanguard.com/advice/digital-advisor vanguard personal advisor. (2022). finally. financial advice powered by relationships, not commissions. available at https://investor.vanguard.com/advice/financial-advisor o. s. mitchell et al. / financial services review 31 (2023) 1–21 21 household ratio guidelines for the amount of investments sherman d. hanna,a,* kyoung tae kimb adepartment of human sciences, ohio state university, 1787 neil avenue, columbus, oh 43210, usa bdepartment of consumer sciences, university of alabama, 312 adams hall, box 870158, tuscaloosa, al 35487, usa abstract some textbooks suggest using financial ratios to provide simple indicators of whether households are making appropriate financial decisions. we investigate three investment ratios mentioned in textbooks: investments to net worth, investments to annual income, and investments to total assets. we conduct regressions on respondent evaluation of the adequacy of retirement income, among households with a non-retired head in the 2013 survey of consumer finances. the investments to total assets ratio has the strongest relationship to adequacy, controlling for selected household characteristics. the investments to net worth ratio (capital accumulation ratio) is inferior to the other two ratios. © 2016 academy of financial services. all rights reserved. jel classification: d14; d91; g11; j26 keywords: capital accumulation ratio; financial ratios; investing; lifecycle theory of savings; survey of consumer finances 1. introduction 1.1. background household financial ratios are used to help provide simple rules for financial decisions, because many households have trouble with more complex analyses (greninger, hampton, kitt, and achacoso, 1996; harness, chatterjee, and finke, 2008). financial ratio guidelines are intended to provide easily understandable rules, without necessarily allowing for indi* corresponding author. tel.: �1-614-292-4584; fax: �1-614-292-4339. e-mail address: hanna.1@osu.edu (s.d. hanna) financial services review 25 (2016) 263–277 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. vidual circumstances. for instance, the guideline of saving 10% of income is widely accepted by financial experts (greninger et al., 1996), and may provide a starting point for discussion between financial advisers and clients, but it is clearly not justified in terms of the normative life cycle model (hanna, fan, and chang, 1995), as most households should not save in years when income is far below normal, and perhaps should save much more than 10% in some situations. however, as a simple educational device, financial ratio guidelines may be useful for educators and advisers. one topic covered by a few proposed ratio guidelines is the appropriate amount of investments that should be held by households. the investment ratios we examine are related to retirement adequacy. a complete assessment of retirement adequacy is complex, and involves many assumptions (greninger, hampton, kitt, and achacoso, 2000). survey analyses of retirement adequacy involve careful identification of the size and composition of retirement investments, expected retirement age, and projection of the levels of social security and any other pensions (kim and hanna, 2015; yuh, hanna, and montalto, 1998). for many households, having multiple retirement income stages can further complicate the analysis of retirement adequacy (kim, hanna, and chen, 2014). investment ratios can potentially provide a quick assessment of whether a household is on track to an adequate retirement. guidelines related to the amount of investments ignore the role of home equity in the overall household portfolio (cheung and miu, 2015) and make many other simplistic assumptions. the most frequently mentioned investment ratio is the capital accumulation ratio, which has been analyzed in a number of empirical analyses (harness, finke, chatterjee, 2009; letkiewicz and hanna, 2013; moon, yuh, and hanna, 2002; yao, hanna, and montalto, 2002, 2003). the capital accumulation ratio, defined as the proportion of net worth held in investment assets, was proposed by lytton, garman, and porter (1991). it has been used in two personal finance textbooks, in editions of the garman and forgue (2003) from about 2000 to 2010, and in devaney (1997). these authors propose that having a capital accumulation ratio of at least 25% is a good indicator of the ability to accumulate capital for future goals, as it shows that net worth is not being devoted to vehicles and one’s personal residence. in the greninger et al. (1996) article reporting a delphi survey of financial planners and educators, a level of 50% for the ratio is suggested. however, as letkiewicz and hanna (2013) note, the capital accumulation ratio guideline advocated by garman and forgue (2003) in previous versions of their textbook before 2012 is not commonly suggested in financial planning textbooks. garman and forgue (2012, p. 78) suggest using the ratio of investments to total assets to answer the question of whether one is investing enough. they state guidelines of having the ratio be at least 10% for those in their 20’s, 11% to 30% for those in their 30’s, and over 30% for those aged 40 and over. yao et al. (2002) mention the alternative of using the investments to assets ratio instead of the capital accumulation ratio because of the problem of households having negative net worth. dalton, dalton, cangelosi, guttery, and wasserman (2005, p. 124) suggest using the ratio of investment assets to annual income, with a goal of having sufficient investments to generate the amount of income needed in retirement. they suggest having a ratio of 10 at retirement, with ratios of 3 to 4 about 10 years before retirement and a ratio of at least 1 264 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 about 20 years before retirement. this ratio is similar to the normative analysis presented in ibbotson (2008, p. 199). fig. 1 shows graphically the pattern implied by the results in ibbotson. as shown, the needed ratio depends not only on age, but also on income, because of the progressive nature of social security pension benefits and taxes on current earnings. while a number of empirical studies have been conducted using the capital accumulation ratio (see review by letkiewicz and hanna, 2013), there are no empirical studies of one of the newer ratios, the ratio of investment assets to annual income (dalton et al., 2005). there is one brief abstract related to the other newer ratio used by garman and forgue (2012), the ratio of investment assets to total assets. yao and hanna (2002) report comparisons how well the 25% capital accumulation ratio guideline is related to retirement adequacy to how well the 25% investment to assets guideline is related to retirement adequacy, and note that neither guideline is a very good indicator of retirement adequacy. many of the previous empirical studies on investment ratios focus on meeting simple thresholds, and therefore, have not provided much insight into what levels might be best for households. there have been no rigorous analyses of whether particular ratio levels are optimal (moon, et al, 2002). 1.2. objectives lytton et al. (1991) suggest that establishment of ratio guidelines “… would be dependent on … empirical research to determine appropriate empirical ranges.” harness et al. (2009) also suggest that ratio guidelines should be research based. they note that “a problem with using ratios as tools is that the extant literature testing their value is limited. for example, there is little evidence that a capital accumulation ratio of 0.7 is better than one of 0.3.” the fig. 1. ratio of needed amount of retirement investments to income, ibbotson analysis of amount to accumulate to replace 80% of net pre-retirement income, by age and income level. calculations by author, based on table in ibbotson (2008), p. 199. 265s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 objective of this article is to provide insights into three household investment ratios, what value of each ratio is best, and whether one ratio is better than the others. we do not directly test whether guidelines proposed by textbook authors make sense, as there is no rigorous analysis presented by any of the authors. we present distributions of the two newer ratios, along with the capital accumulation ratio, based on analyses of the 2013 survey of consumer finances. furthermore, for an exploratory attempt to derive normative results for levels of the three investment ratios, three ordinary least squares (ols) regressions are utilized, with respondent assessment of the adequacy of future retirement income, as a function of age, homeownership status, having a defined benefit plan, and employment status along with linear and quadratic variables for each investment ratio. our conclusion on which ratio is the best is based on the explanatory power of each ratio regression, and also the plausibility of the effects of the ratio on the assessment of retirement adequacy, and the mathematical properties of each ratio. 2. method 2.1. mathematical considerations for financial ratios harness et al. (2009) note that many financial ratios have a non-normal distribution. the component variables of the three ratios related to investments are income, investments, net worth, and total assets. all have skewed distributions, especially the asset and net worth variables, with very long tails of the distribution. further, many households have zero or negative net worth, making interpretation of the capital accumulation ratio problematic, because net worth is in the denominator of the ratio. 2.2. data and sample in this study, the 2013 survey of consumer finances (scf) released by federal reserve board (bricker et al., 2014) is used. the total sample size of the 2013 scf is 6,015 and for our main analyses, households with non-retired heads (n � 5,665) are analyzed. 2.3. dependent variable if an investment ratio is to provide some guidance for financial planning, presumably the level of the ratio should be related to projected retirement adequacy. this is the approach used by yao et al. (2003), although they use arbitrary levels of the capital accumulation ratio and analyze the effect on an objective measure of retirement adequacy. in this exploratory study, we use the household’s subjective assessment of retirement adequacy as a dependent variable. the scf variable (x3023) has five levels: x3023 using any number from one to five, where one equals totally inadequate and five equals very satisfactory, how would you rate the retirement income you receive (or expect to 266 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 receive) from social security and job pensions? include 401(k) accounts and all other types of pensions. 1. *totally inadequate 2. 3. *enough to maintain living standards 4. 5. *very satisfactory kim and hanna (2015) show that this variable is related to an objective estimate of retirement adequacy, although not perfectly. if we want to include households of all ages, and not make arbitrary assumptions about the retirement ages of those who did not list a specific retirement age, the subjective measure of retirement adequacy has advantages over the objective measure. three ols regressions on subjective retirement adequacy are run, controlling for age of head, age squared, homeownership status, having a defined benefit pension, employment status and one of the ratio variables. to allow for nonlinear effects of the ratios on subjective retirement adequacy, a quadratic term for the ratio is also included. we include linear and quadratic terms for both age and for each ratio. appendix a shows how to obtain the optimum level (extreme point) for age and for each ratio using basic principles from calculus. 2.4. independent variables 2.4.1. key variables: three investment ratios each of the ratios has the value of investments in the numerator. investment assets consist of all financial assets other than monetary assets such as checking and saving accounts, plus nonfinancial assets such as art work, antiques, net business assets, and real estate other than the personal residence. monetary (liquid) assets include checking, savings, money market, and call accounts, and are not counted as part of investment assets, although certificates of deposit are included as investment assets. net worth is the sum of monetary assets, investment assets, and non-financial assets minus consumer debt and property debt. as shown in the descriptive results below, there are some extreme values of the ratio, so some adjustments are made for the regression analysis. in our analytic sample, 71% of households have positive amounts of investments. the capital accumulation ratio is defined as investment assets-to-net worth and is calculated from information on investment assets and net worth. if net worth is zero or negative, then the ratio is defined as equal to the value of investments, in other words, the denominator will be assumed to be equal to one (cf., letkiewicz and hanna, 2013). table 1 illustrates the rationale for recoding the ratio for values of net worth less than or equal to zero. assume that the value of investment assets equals $10,000. for a household with net worth of $100,000, the ratio is 0.1, and for a household with $10,000 of net worth, the ratio is 1.0. however, as net worth approaches zero, the ratio increases, for example, it is 10,000 for net worth of $1, and the ratio is not defined for net worth of zero. conceptually, as net worth decreases in the positive range, the increase in the ratio values represents increased leverage. however, as net worth decreases from zero in the negative range, the ratios are 267s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 negative. for net worth of $1, the ratio is 10,000, but for a net worth of �$1, the ratio is �10,000. this switch does not make sense in terms of our objective of relating values of the ratio to retirement adequacy. in both cases the household is highly leveraged. consider somebody who just graduated from college with $110,000 of debt and no assets other than an ira worth $10,000. even though the situation is different from somebody with net worth of $1 and an ira worth $10,000, it seems reasonable to assign a value of the ratio of 10,000 to both cases, rather than a ratio of �0.1 to the person with $110,000 of debt. therefore, we follow the example of yao et al. (2002) and yao et al. (2003) and calculate the ratio as investments divided by 1.0 if net worth is non-positive. in our analytic sample, there are 579 households with zero or negative net worth (11.7% of the sample based on a weighted analysis.). the ratio of investments to annual income (dalton et al., 2005) is calculated as the ratio of investments to annual household income. dalton et al. (2005) do not specify whether income should be measured as current or normal income, but we use respondent estimates of normal income. if the denominator is zero, the ratio is defined at the value of investments. as with the capital accumulation ratio, there are some extreme values of the ratio, so adjustments are made for the regression analysis. dalton, dalton, and oakley (2014) discuss the ratio of investments plus monetary assets to gross earnings, although this ratio does not address the idea of needing investments that will grow. however, as a worker approaches retirement, it is plausible that those that are risk averse will shift some investments to monetary assets such as money market accounts. in our analytic sample, there are eight households reporting zero normal income (0.08% of the sample based on a weighted analysis). the ratio of investments to total assets (garman and forgue, 2012) is calculated as the ratio of investment assets to total assets. total assets include both financial assets and non-financial assets. for the 82 households with zero assets (1.56% of the sample based on table 1 illustration of distribution of investments to net worth ratio, assuming investments � $10,000 net worth ($) ratio 100,000 0.1 10,000 1 5,000 2 1,000 10 100 100 10 1,000 1 10,000 0.1 100,000 0.01 1,000,000 �0.01 �1,000,000 �0.1 �100,000 �1 �10,000 �100 �100 �1,000 �10 �10,000 �1 �100,000 �0.1 note: table created by authors. 268 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 a weighted analysis.), the ratio is defined as the value of investments, which in all cases is equal to zero. 2.4.2. control variables in each regression, in addition to the ratio and the square of the ratio, a few household characteristic variables directly related to retirement adequacy are included as control variables. the age of the household head and the square of the age (divided by 10,000 to scale the estimated coefficients) were included. to test for whether optimal ratio levels depend on age, interaction variables are created between age and the ratio and the square of the ratio. one disadvantage of adding interaction terms is that there is multicollinearity because of the interaction terms, resulting in some effects not being significantly different from zero, so for comparison, we also present regressions without interaction terms (appendix b). dummy variables for whether the household owns its home, and whether the head has a defined benefit pension are included. furthermore, dummy variables for whether the head works part-time, and whether the head is not working but not retired are included (both relative to the head working full-time). 2.5. descriptive and multivariate analyses for descriptive analyses, quantiles, means, and skewness of the ratios are obtained with weighted analyses, both for the entire 2013 scf sample and for the analytic sample of households with a non-retired head. in addition, the means and medians of the ratios for age groups are obtained with weighted analyses. ols regressions are utilized to estimate the effect of ratio levels on respondent assessment of the adequacy of retirement income. the suggestions in lindamood, hanna, and bi (2007) are followed in analyzing the data, including use of unweighted regressions. 3. results 3.1. descriptive results for all 6,015 households in the 2013 scf, the maximum level of net worth in the 2013 scf is $1,324,417,600, with 11.59% of households having negative net worth, and 1.33% of households having zero net worth. (the scf includes vehicles, housing, and financial assets, but does not include the value of personal possessions such as furniture.) the maximum value of investment assets is over one billion dollars, $1,320,813,200. however, 30.57% of households have zero investment assets, and one household is coded as having negative investments, presumably because of debt not recorded elsewhere. the maximum value of assets is $1,324,540,600, and there is also a case with a negative value for total assets. table 2 shows the distribution of the three ratios for all households and for households with non-retired heads. for all households, the maximum value of the investments to net worth ratio is almost 59,000,000, partly because of the almost 13% of households with zero or negative net worth. for instance, a young recent college graduate with $1,000 in a mutual 269s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 fund and $30,000 in student debt would have an investments to net worth ratio of 1,000. the maximum value of the investments to annual income ratio is over 125,000,000, whereas the maximum value of the investments to total assets ratio is only 1.00. for all households, the median level of the investments to net worth ratio is 0.35, the median for the investments to normal income ratio is 0.31, and the median for the investments to total assets ratio is 0.17. for all households, the skewness of the investments to net worth and the investments to normal income ratios are very high (74 and 56, respectively), but the skewness of the investments to total assets ratio is only 0.7. the columns for the ratio with non-retired heads correspond to the sample used for the regression analyses. in addition, for these columns and for the regression, extreme values of the investments to net worth ratio and of the investments to normal income ratio are recoded. for all households, the 95th percentile of the investments to net worth ratio is 130, so for households with values of the ratio above 130, the ratio is recoded to 130 to avoid having overly influential values in the regression analysis. the 95th percentile of the investments to normal income ratio is 11.43 for all households. that level is not extreme in terms of the ibbotson (2008) normative analysis (see fig. 1), so we instead use the approximate level of the 99th percentile in the full sample, 30, as the point of recoding for our analytic sample of non-retired households. values of the investments to normal income ratio above 30 are recoded to 30. the rationale for both of these recoding rules is to limit the effect of a few extreme values on the dependent variable. table 3 shows mean and median values of the three ratios by age ranges. the mean values of the investments to net worth ratio decrease by age, contrary to textbook discussions of the desirable patterns of the ratio by age (devaney, 1997; garman and forgue, 2003). the median value of the investments to net worth ratio increases by age up to the 55 to 64 age range. the mean value of the investments to normal income ratio increases by age, from 0.50 for those under 25 to 3.99 for those 65 and over. the median value of the investments to table 2 distribution of three investment ratios, all households (untrimmed) and households with non-retired head, trimmed ratios distribution investment/net worth investment/normal income investment/total assets all households all non-retired* all households all non-retired** all households all non-retired mean 1026.43 6.98 383.17 2.32 0.28 0.29 maximum 58,580,000 130.00 125,340,000 30.00 1.00 1.00 99th percentile 20,000 130.00 30.35 30.00 0.98 0.98 95th percentile 130.00 100.00 11.43 11.74 0.76 0.88 75th percentile 0.76 0.77 2.12 2.25 0.52 0.53 median 0.35 0.37 0.31 0.38 0.17 0.19 25 percentile 0.00 0.00 0.00 0.00 0.00 0.00 minimum 0.00 0.00 0.00 0.00 0.00 0.00 skewness 74.35 4.57 55.80 3.10 0.70 0.65 note: analysis by authors of 2013 survey of consumer finances, weighted. ratios defined as value of investments if denominator �0. recoded to 0 if investments �0. *values above 130 recoded to 130. **values above 30 recoded to 30. 270 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 normal income ratio increases by age, from 0.00 for those under 25 to 1.02 for those 65 and over. the mean value of the investments to assets ratio increases by age up to the 55 to 64 range, from 0.13 for those under 25 to 0.36 for those in the 55 to 64 range. the median value of the investments to assets ratio increases by age up to the 55 to 64 range, from 0.00 for those under 25 to 0.33 for those in the 55 to 64 range. 3.2. regression results table 4 shows the results for three ols regressions, with the dependent variable being the respondent’s subjective assessment of the adequacy of retirement income, including interaction terms between age of the head and the ratio and the square of the ratio. (appendix b has the same regressions without the interaction terms.) the effects of the ratio of investments to net worth and the square of that ratio, as well as the interactions of age with the ratio and with the ratio squared, are not significant. age and age squared are significant, and the combined effect implies that at a value of the ratio equal to zero, subjective assessment of retirement adequacy decreases with age up to age 40, then increases as age increases above age 40. homeowners have a higher level of perceived retirement adequacy than otherwise similar renters. those with a defined benefit pension have significantly higher assessments of adequacy than those without one. households with a head employed part-time, and those with a head not working but not retired, have higher assessments of adequacy than those with a head employed full-time. the adjusted r2 for the regression is 0.044. for the second ols regression shown in table 4, the effects of the ratio of investments to normal income and the square of that ratio, as well as the interactions of age with the ratio and with the ratio squared, are not significant. in a version of the regression without interaction terms (appendix b), the effects of the ratio and the ratio squared are significant, table 3 mean and median of ratios by age groups age range investments/net worth* investments/normal income** investments/assets mean �25 18.30 0.50 0.13 25–34 16.13 0.64 0.20 35–44 8.92 1.20 0.27 45–54 5.80 2.03 0.32 55–64 4.13 3.15 0.36 65 and over 1.25 3.99 0.32 median �25 0.00 0.00 0.00 25–34 0.27 0.08 0.07 35–44 0.40 0.29 0.18 45–54 0.46 0.63 0.25 55–64 0.49 0.95 0.33 65 and over 0.29 1.02 0.24 note: analysis by authors of households with non-retired heads of 2013 survey of consumer finances, weighted. ratios defined as value of investments if denominator �0. recoded to 0 if investments �0. *values above 130 recoded to 130. **values above 30 recoded to 30. 271s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 and the combined effect of the ratio and ratio squared are positive up to a level of 14. age and age squared are significant, and the combined effect implies that at a value of the ratio equal to zero, subjective assessment of retirement adequacy decreases with age up to age 41, then increases as age increases above age 41. homeowners have a higher level of perceived retirement adequacy than otherwise similar renters. those with a defined benefit pension have significantly higher assessments of adequacy than those without one. households with a head employed part-time, and those with a head not working but not retired, have higher assessments of adequacy than those with a head employed full-time. the adjusted r2 for the regression is 0.050. for the third ols regression shown in table 4, the effects of the ratio of investments to assets and the square of that ratio do not have significant effects on retirement adequacy, but the interaction of age and the ratio is significant and positive. in a version of the regression without interaction terms (appendix b), the effects of the ratio and the ratio squared are significant, and the combined effect of the ratio and the ratio squared are positive up to a level of 0.52. age and age squared are significant, and the combined effect implies that at a value of the ratio equal to zero, subjective assessment of retirement adequacy decreases with age up to age 46, then increases as age increases above age 46. those with a defined benefit pension have significantly higher assessments of adequacy than those without one. households with a head employed part-time, and those with a head not working but not retired, have higher assessments of adequacy than those with a head employed full-time. the adjusted r2 for the regression is 0.058. table 5 shows the levels of each ratio for which there is an extreme point in terms of the combined effects of the ratio and ratio squared and the interaction terms in each regression, based on the formula shown in appendix a. all of the numbers shown in table 5 represent maximum points, for example, for a household head age 25, the level of the investments to assets ratio that maximizes the assessment of retirement income is 0.44. there is no table 4 regressions on subjective assessment of retirement adequacy, among households with non-retired head, by investment ratios, controlling for age, homeownership, having defined benefit plan, working part-time, and not working variable investment/net worth investment/normal income investment/assets coefficient p-value coefficient p-value coefficient p-value ratio �0.0446 0.2071 �0.0227 0.4699 0.2566 0.6800 ratio squared 0.0003 0.2054 �0.0002 0.8594 �0.5280 0.4593 age * ratio 0.0014 0.1295 0.0009 0.0767 0.0266 0.0182 age * ratio squared �0.00001 0.1231 �0.00002 0.3338 �0.0205 0.1064 age of head �0.0261 �.0001 �0.0225 �.0001 �0.0301 �.0001 age of head/10000 3.2870 �.0001 2.7423 �.0001 3.2392 �.0001 homeowner 0.1565 0.0001 0.1399 0.0007 0.0693 0.0930 have defined benefit pension 0.3922 �.0001 0.3893 �.0001 0.3682 �.0001 employment status of head (reference category: full-time) head working part-time 0.1344 .0204 0.1387 0.0172 0.1504 .0095 head not working but not retired 0.1700 .0009 0.1933 0.0002 0.2331 �.0001 intercept 2.6254 �.0001 2.5700 �.0001 2.6884 �.0001 adjusted r2 0.0444 0.0500 0.0577 note: analysis by authors of 2013 scf, households with non-retired heads. extreme values of ratios recoded. unweighted analyses. rii technique used. 272 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 monotonic pattern between age and the level of the investments to net worth ratio. for the investments to normal income ratio, for all ages above 25, the optimal level of the ratio increases with age, from 6.06 at age 35 to 14.80 at age 65. for the investments to assets ratio, the optimal level of the ratio increases slightly with age, from 0.44 at age 25 to 0.53 at age 65. 4. summary and implications 4.1. summary the investments to net worth ratio (also known as the capital accumulation ratio) is the only household financial ratio related to the amounts of investments for which there has been empirical research, and it is discussed in two personal finance textbooks (devaney, 1997; garman and forgue, 2003). however, it has some problematic mathematical properties (harness, et al., 2008), as it cannot be directly calculated for almost 13% of the households in the 2013 scf. the ratio is not related to subjective assessment of retirement adequacy among households with non-retired heads. another financial ratio, the investments to annual income ratio, is also discussed in a textbook (dalton et al., 2005). however, it has extreme values for a small proportion of households. it is not as strongly related to subjective assessment of retirement adequacy as the ratio of investment assets to total assets. the investments to assets ratio is discussed in a textbook (garman and forgue, 2012). the investments to total assets ratio has advantages in terms of some of the mathematical issues discussed by harness et al. (2008). even though it has a skewed distribution, it is not as skewed as the other two ratios. furthermore, it is almost always possible to calculate the investments to total assets ratio, as long as the household does not have zero assets. the regression with the investments to assets ratio has a higher adjusted r2 than the regressions with the other two ratios. further, the investments to total assets ratio has a positive effect on subjective retirement adequacy for most households, even though in the ols regression we might interpret a small decrease in subjective adequacy as the ratio increases above a level of 0.6. by contrast, the investments to net worth ratio does not have a statistically significant effect on subjective retirement adequacy. the investments to normal income ratio does have a significant positive effect on subjective retirement adequacy, though it has some very extreme values, unlike the investments to assets ratio. table 5 optimum values of ratios by age, based on regressions age investments/net worth investments/normal income investments/assets 25 68.28 �0.07 0.44 35 56.01 6.06 0.48 45 61.70 10.01 0.50 55 62.58 12.77 0.52 65 62.94 14.80 0.53 note: based on regressions in table 4. 273s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 4.2. implications as with any household financial ratio above, the investments to asset ratio can at best be a simplistic guide to initial thinking about financial decisions. further, the ols regression with subjective retirement adequacy also has significant effects for homeownership, age, and having a defined benefit pension. therefore, for instance, if two households are similar in terms of age and homeownership status, and have the same value of the investments to asset ratio, but one has a defined benefit pension but the other has nothing but social security, it makes sense that they would have different assessments of retirement adequacy. if an author or educator wants to use a financial ratio guideline related to the amount of investments in terms of two of three criteria stated in section 1.2, the investments to assets ratio is superior to other ratio guidelines that have been proposed. for the regressions of ratios on perceived retirement adequacy, the investments to assets ratio has the highest r2 and has the best mathematical properties. the relationship between ratio levels and retirement adequacy are plausible for both the investments to assets ratio and the investments to normal income ratio, though in the regressions with interaction terms with age (table 4), none of the terms involving the investments to normal income ratio are significantly different from zero, while for the investments to assets ratio, one term is significant. in the regressions with no age interaction terms (appendix b), the ratio and the square of the ratio have highly significant effects for both the investments to normal income ratio and the investments to assets ratio. for the investments to assets ratio, a ratio guideline of 0.5 might be plausible, based on our regression results. as with any financial ratio guideline, one should be cautious in applying such guidelines, as a more complex analysis of a household’s situation is always desirable. even though the investments to assets ratio is best in terms of mathematical properties and explanatory power, the investments to normal income ratio has some plausibility in terms of financial planning models (e.g., fig. 1). both the median and mean levels of both the investments to normal income ratio and the investments to assets ratio by age group (table 3) are far below the optimal levels of these two ratios based on the regressions (table 5). this divergence between the optimal patterns and actual patterns is consistent with most research on retirement adequacy (hanna, kim, and chen, 2016). the investments to net worth ratio, also known as the capital accumulation ratio, has poor mathematical properties, and has limited explanatory power in terms of the regression on perceived retirement adequacy. even though this ratio is the only investment ratio with an extensive amount of empirical research, it should not be used in financial education or advising in the future. appendix: a finding maximum levels for age and for each ratio from regression results. if y is the assessment of retirement adequacy, and x is either age or the ratio, the relevant results from each regression without interaction terms between age and the ratio can be expressed as: 274 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 y � ax � bx2. (1) dy/dx � a � 2bx. (2) at extreme point, slope �0. so, xe � �a/2b. (3) if a � 0, xe is a maximum. if a � 0, xe is a minimum. the calculus is more complicated with the interaction terms. with x � ratio, g � age, coefficients � a, b, c, d y � ax � bx2 � cgx� dgx2. (4) dy/dx � a � 2bx � cg � 2dgx. (5) at slop � 0, x � �(a � cg)/(2b � 2dg). (6) appendix b table b1 shows regressions of the ratios and the square of the ratios, without interaction terms between age and the ratio and ratio squared variables. (the regressions with interaction terms for age, shown in table 4, have very high variance inflation factors for the interaction terms, indicating multicollinearity, which resulted in all but one of the interaction terms not being statistically significant.) for the investments to net worth ratio regression, the ratio and its table b1 regressions on subjective assessment of retirement adequacy, among households with non-retired head, by investment ratios, controlling for age, homeownership, having defined benefit plan, working part-time, and not working variable investment/net worth investment/normal income investment/assets coefficient p-value coefficient p-value coefficient p-value ratio 0.0062 .6627 0.0340 �.0001 1.6088 �.0001 ratio squared �5.4e-05 .6220 �0.0012 �.0001 �1.5457 �.0001 age of head �0.0255 �.0001 �0.0274 �.0001 �0.0301 �.0001 age of head/10,000 3.3069 �.0001 3.4001 �.0001 3.3601 �.0001 homeowner 0.1574 �.0001 0.1307 .0003 0.0776 .0591 have defined benefit pension 0.3915 �.0001 0.3905 �.0001 0.3678 �.0001 employment status of head (reference category: full-time) head working part-time 0.1345 .0204 0.1405 .0151 0.1693 .0034 head not working but not retired 0.1616 .0015 0.1863 .0003 0.2242 �.0001 intercept 2.6063 �.0001 2.6401 �.0001 2.5785 �.0001 adjusted r2 0.0444 0.0503 0.0578 note: analysis by authors of 2013 scf, households with non-retired heads. extreme values of ratios recoded. unweighted analyses. rii technique used. 275s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 square are not significant, and the combined effect implies that perceived retirement adequacy is maximized at a level of the ratio of almost 57, an extreme level of household leverage. for the investments to normal income ratio regression, both the ratio and its square have significant effects, and the combined effect implies that perceived retirement adequacy is maximized at a level of the ratio of almost 14. for the investments to asset ratio regression, both the ratio and its square have significant effects, and the combined effect implies that perceived retirement adequacy is maximized at a level of the ratio of 0.5. as with the regressions with interaction terms between age and ratio and ratio squared (table 4), the adjusted r2 is highest for the investments to assets ratio regression, and lowest for the investments to net worth ratio regression. references bricker, j., dettling, l. j., henriques, a., hsu, j. w., moore, k. b., sabelhaus, j., et al. (2014). changes in u.s. family finances from 2010 to 2013: evidence from the survey of consumer finances. federal reserve bulletin, 100, 1–40. cheung, c. s., & miu, p. (2015). home ownership decision in personal finance: some empirical evidence. financial services review, 24, 51–76. dalton, m. a., dalton, j. f., cangelosi, r. r., guttery, r. s., & wasserman, s. a. (2005). personal financial planning theory and practice. le crosse, wi: kaplan financial. dalton, m. a., dalton, j. f., & oakley, k. f. (2014). cases in financial planning: analysis and presentation (2nd ed.). st. rose, la: money education. devaney, s. a. (1997). using financial ratios. in e. t. garman, j. j. xiao, & b. g. brunson (eds.), the mathematics of personal finance: using calculators and computers (pp. 141–153). houston, tx: dame publications. garman, e. t., & forgue, r. (2012). personal finance (11th ed.). mason, oh: south-western. garman, e. t., & forgue, r. (2003). personal finance (7th ed.). boston, ma: houghton mifflin company. greninger, s. a., hampton, v. l., kitt, k. a., & achacoso, j. a. (1996). ratios and benchmarks for measuring the financial well-being of families and individuals. financial services review, 5, 57–70. greninger, s. a., hampton, v. l., kitt, k. a., & achacoso, j. a. (2000). retirement planning guidelines: a delphi study of financial planners and educators. financial services review, 9, 231–245. hanna, s., fan, j. x., & chang, y. r. (1995). optimal life cycle savings. financial counseling and planning, 6, 1–15. hanna, s. d., kim, k., & chen, s. c.-c. (2016). retirement savings. in j. xiao (ed.), handbook of consumer finance research (2nd ed., pp. 33–43). new york, ny: springer publishing. harness, n. j., chatterjee, s., & finke, m. (2008). household financial ratios: a review of literature. journal of personal finance, 6, 77–97. harness, n. j., finke, m. s., & chatterjee, s. (2009). the effects of the capital accumulation ratio on wealth. journal of financial counseling and planning, 20, 44–57. ibbotson associates. (2008). stocks, bonds, bills, and inflation yearbook. chicago, il: ibbotson associates. kim, k., & hanna, s. d. (2015). do u.s. households perceive their retirement preparedness realistically? financial services review, 24, 139–155. kim, k., hanna, s. d., & chen, s. c.-c. (2014). consideration of retirement income stages in planning for retirement. journal of personal finance, 13, 52–64. lindamood, s., hanna, s. d., & bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41, 195–222. letkiewicz, j., & hanna, s. d. (2013). the impact of stock and real estate returns on the propensity of households to meet the capital accumulation ratio guideline. journal of financial counseling and planning, 24, 34–44. 276 s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 lytton, r. h., garman, e. t., & porter, n. m. (1991). how to use financial ratios when advising clients. journal of financial counseling and planning, 2, 3–23. moon, s. j., yuh, y., & hanna, s. d. (2002). financial ratio analysis of korean households. family and consumer sciences research journal, 30, 496–535. yao, r., & hanna, s. d. (2002). the investment-asset ratio: a plausible replacement for the capital accumulation ratio guideline? proceedings of the association for financial counseling and planning education (p. 66). scottsdale, arizona. yao, r, hanna, s. d., & montalto, c. p. (2002). factors related to meeting the capital accumulation ratio guideline. financial services review, 11, 153–171. yao, r, hanna, s. d., & montalto, c. p. (2003). the capital accumulation ratio as an indicator of retirement adequacy. journal of financial counseling and planning, 14, 1–12. yuh, y., hanna, s., & montalto, c. p. (1998). mean and pessimistic projections of retirement adequacy. financial services review, 7, 175–193. 277s.d. hanna, k.t. kim / financial services review 25 (2016) 263–277 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 from the editor this issue contains issue 1 of volume 24 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “gender differences in saving behaviors among lowto moderateincome households” is coauthored by patti j. fisher at virginia tech, celia r. hayhoe at virginia tech, and jean m. lown at utah state university. in this paper, the authors explore gender differences in saving behaviors among lowto moderate-income households using data collected online from a national sample of lowto moderate-income households and data on similar income single households from the 2010 survey of consumer finances. their results indicate that saving behaviors differ by gender. they find gender differences in the effects of high risk tolerance and being non-white on the likelihood of being a saver. they also find that the presence of other household members affects savings differently for women and men. they recommend that counselors encourage savings among men and women in lowto moderate-income households as a way to reduce financial risk and ensure financial security. the second article “spread options and risk management: lognormal versus normal distribution approach” is coauthored by robert brooks at the university of alabama and brandon n. cline at mississippi state university. the authors provide tools for managing the downside risk related to the spread between the asset portfolio and corresponding liabilities for individual investors. they investigate the spread option valuation model where both underlying instruments follow geometric brownian motion, and one where both underlying instruments are assumed to follow arithmetic brownian motion. they show that the risk parameters are often materially different. for most personal financial planning applications, one can safely use the simpler arithmetic brownian motion model. the third article, ”minority household size and the life insurance purchase decision “ is coauthored by michael a. guillemette at the university of missouri, m. monica hussein at california state university, northridge, and g. michael phillips at california state university, northridge. the authors use the 1992-2010 survey of consumer finances to analyze whether the likelihood of life insurance ownership and the face value amount of life insurance changes for minorities as household size changes. they find that the likelihood of financial services review 24 (2015) v–vi 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. life insurance ownership declines for black and larger hispanic families as household size increases when controlling for socioeconomic and demographic variables. they also find a significant decline in the face value amount of term life insurance purchased by black families as household size rises. they recommend that financial planners recognize potential cultural biases and the possible role of household members when meeting with their clients to discuss life insurance coverage. the fourth article, “home ownership decision in personal finance: some empirical evidence” is coauthored by c. sherman cheung and peter miu both at mcmaster university. in this paper, the authors examine the home ownership question for households with differing risk tolerance and demonstrate the interaction effect between financial assets and home ownership. they also examine whether the economic case for home ownership varies across u.s. regions. for households that decide to rent instead of own, the authors recommend hedging housing consumption risk with investments in real estate investment trusts. the final article, “do financial networks matter in retirement investment decisions? evidence from generation yers” is coauthored by yunhyung chung and youngkyun park both at the university of idaho. the authors use experimental survey data collected from a sample of generation yers to examine the joint influence of financial literacy and financial networks on individual retirement investment decisions. they find that financial literacy and financial network intensity are positively related to stock allocation. they also show that the positive relationship between financial literacy and stock allocation is significant only among those having high financial network intensity. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 24 (2015) v–vi seeking tax alpha in retirement income james a. dilellioa,*, andreas simonb adepartment of decision sciences, information systems and strategy, graziadio business school, pepperdine university, 24255 pacific coast highway, malibu, ca 90263, usa bleventhal school of accounting, marshall school of business, the university of southern california, 3660 trousdale parkway, los angeles, ca 90089-0441, usa abstract we provide a framework to find an optimal decision for tax-efficient retirement income. by developing a model for income and capital gains tax with stock and bond investments in taxdeferred, tax-exempt, and taxable accounts, we identify three categories of retirees based on their income needs and net worth. we propose and evaluate a simple heuristic to determine the optimal retirement income strategy, quantifying a 0.5% annual return benefit. we call this benefit tax alpha and show its robustness to varying model input parameters. we also suggest approaches for large institutions or fintech firms to improve their existing financial planning tools. © 2022 academy of financial services. all rights reserved. jel classifications: g11; h21 keywords: retirement income; longevity risk; tax efficiency; asset location 1. introduction as the baby boomer generation exits the workforce, their needs for tax-efficient use of their retirement assets grow. according to the st. louis fed, approximately 10,000 u.s. baby boomers expect to retire every day from 2019 to 2039.1 similar trends are seen in other western countries, like the united kingdom, germany, australia, and others due to the mid20th century baby boom.2 many of these retirees will have built up substantial assets in taxdeferred accounts, for example, the u.s. 401(k) plans, since their introduction in 1981.3 accounts like traditional or rollover u.s. iras, 403(b)s, and 457 plans offer similar tax*corresponding author. tel.: +1-714-403-0085; fax: +1-949-223-2575. e-mail address: james.dilellio@pepperdine.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 223–249 deferred investing for non-profit and government employees, and in some cases, have been around for even longer.4 consequently, many retirees are now beginning to draw down these assets, either voluntarily or involuntarily, due to required minimum distributions (rmds).5 in addition to tax-deferred accounts, many baby boomers also have tax-exempt accounts, such as the u.s. roth iras, roth 401(k)s, and roth 403(b)s. tax-exempt accounts in the united states have only been around for less time compared with tax-deferred accounts and typically have fewer assets. in the united states, the roth iras also restrict direct contributions for higher wage earners, further limiting the assets in these accounts. brown et al. (2017) reported tax-exempt retirement income as very beneficial for all wage earners entering retirement as these assets can be used to support higher levels of retirement income in a progressive tax system, like in the united states, germany, united kingdom, canada, australia, among others. moreover, many retirees may have direct ownership in other assets residing in accounts that are taxable each year, such as stocks, bonds, mutual funds, exchange traded funds (etfs), cds, and savings accounts. with few exceptions, retirees are also entitled to government benefits, like social security in the united states, or the superannuation guarantee in australia (see gerrans et al., 2009). lastly, some retirees may have access to pension plans. while tax law does vary from country to country, most of the basic concepts of investor taxation (e.g., capital gains, taxes on dividends, and tax-deferred accounts) exist globally in developed countries. therefore, the problem facing our current and future generations of retirees throughout the world is how they may draw down their assets in the most tax-efficient manner. improving tax efficiency will extend portfolio longevity and increase the retiree’s bequest. the wide variation in retirement assets and retirees living longer, as discussed in poterba (2014), further complicates this problem. indeed, nobel laureate bill sharpe described tax-efficient retirement drawdowns as “one of the most difficult financial problems” he has ever attempted (see milesvsky, 2020). the paper rises to this challenge, showing how 0.5% of added investment return may be by following optimal decisions in retirement drawdowns. addressing tax-efficient retirement drawdowns falls into the category of prescriptive analytics and optimization. unfortunately, the optimization model is non-linear, constrained, and stochastic. non-linearities arise from a progressive tax system, and constraints occur from income needs, limited (if any) loan options to fund retirement, and rmds (when applicable). uncertainty appears in future tax law, investment returns, and inflation in such a model. in this article, we provide a comprehensive mathematical representation of this stochastic optimization problem and provide a simple heuristic to inform future retirement planning. it includes the use of the sequential least squares programming (slsqp) approach, and a heuristic to identify when it can (and cannot) be applied. we also conduct a detailed sensitivity analysis to show how the optimal solution is (or is not) impacted by uncertainty. 2. how this work contributes to the current body of knowledge the literature in retirement planning generally falls into two categories of analysis using either objective wealth functions or utility functions. we discuss each category below, and how this work contributes and extends previous research in this area. 224 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 2.1. objective wealth functions and tax law models the first group of research investigates optimization and constraints imposed by federal tax laws. al zaman (2008) was the first to introduce a retiree’s objective that includes tax rates of their heirs in account selection to fund retirement income. we contribute to this work by including it as part of the optimal decision-making process. we show that the optimal decision can be significantly affected by this parameter. similarly, recent work by pfau et al. (2017) analyzed three unique scenarios, including home equity, social security, and other retirement assets in identifying potential enhancements. cook et al. (2015) show the significant benefit of roth conversions in extending portfolio longevity using specific use cases. in the results shown here, we demonstrate a similar increase in longevity by modifying the common rule withdrawal strategy to include taxdeferred account withdrawals dependent on the heir’s tax rate. the common rule withdrawal strategy is often used as the baseline approach used by coopersmith and sumutka (2011), sumutka et al. (2012), and others. the common rule withdrawal strategy is also widely discussed in the popular press by solin (2010), rodgers (2009), and lange (2009), endorsed by large retail investment firms fidelity (see fidelity, 2014) and vanguard (see vanguard, 2016). the common rule withdrawal strategy first applies rmds to tax-deferred accounts, when applicable, each year. any unmet income is then met by taxable account funds until exhausted. next, voluntary tax-deferred account withdrawals are made until these accounts are exhausted. finally, taxexempt accounts are used to satisfy retirement income. when tax-exempt funds are exhausted, the retiree has insufficient assets to support their desired retirement income. however, practitioners like piper (2013) and demuth (2016) provide many insights into when and how such an approach is far from optimal. we contribute to this work by showing that the common rule strategy provides an important heuristic in seeking a global optimal retirement income decision. geisler and hulse (2018) extended work on the common rule by expanding it to include social security benefits and rmds. they confirmed the persistence of outperformance against the common rule withdrawal strategy. the effect of social security on tax-efficient retirement income has also been investigated by meyer and reichenstein (2013) and reichenstein & meyer (2018), who identify that, for certain lower-income retirees, a spike in marginal tax rates occurs that they call the “tax torpedo.” in our work here, we avoid this issue by assuming a retiree’s modified adjusted gross income (magi) is high enough that 85% of social security income is taxable as ordinary income, and the remaining 15% is not.6 coopersmith and sumutka (2011) highlight the benefits of general-purpose optimization routines, as well as the mathematical and implementation challenges it faces due to non-linear constraints imposed by progressive taxes. their work builds on the pioneering work of linear programming for retirement income planning by ragsdale et al. (1994). most recently, welch (2016, 2017) showed how linear programming models can measure tax-exempt conversion effectiveness and be adapted to upward or downward trending markets. we contribute to this work by using the sequential least squares programming (slsqp) routine, which is freely available in the scipy package of the python 3 open-source programming language.7 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 225 coopersmith and sumutka (2017) evaluated the effectiveness of tax-exempt conversions using a linear programming model when different rates of returns are used. their article is the first to highlight how different investment returns in tax-deferred and tax-exempt accounts influence the benefit of tax-exempt conversions. dilellio and ostrov (2017) show that, under progressive income tax rates, there is a simple geometric representation of account consumption that maximizes wealth transferred to an heir. they also note the existence of multiple optimal solutions for income falling within a given tax bracket, which can confound general-purpose optimization algorithms. this article uses this important insight to develop a “modified” common rule that, with insight into the heir’s marginal tax rate, can extend portfolio longevity or increase a bequest. in dilellio and ostrov (2020), the authors extend their previous study by including a taxable account with stock reinvested. they develop a complex algorithm that finds the optimal switching time between tax-exempt and taxable accounts. we extended this work by showing that, under the regime of sufficient but not excessive assets, switching times can also be found using a general-purpose optimization routine, avoiding the complexity of a customized algorithm. horan (2006) wrote one of the pioneering articles that quantified, under a progressive tax system, tax-efficient withdrawal strategies. he showed that tax-deferred accounts could take advantage of deductions and lower tax brackets to significantly increase wealth accumulation over the common rule withdrawal strategy. we build upon this seminal work by generalizing tax efficiency across three categories of a retiree’s net worth, retirement income needs, and retirement horizon. these categories are retirees with insufficient assets, sufficient assets, or excessive assets relative to their income needs and retirement horizon. 2.2. economic utility functions another group of retirement income research uses utility functions. in brown et al. (2017), the authors find that tax-exempt accounts can be important for high net worth retirees subject to uncertain future investment returns and tax rates. dammon et al. (2004) conclude that in the presence of liquidity shocks, there is a strong preference for holding taxable bonds in a tax-deferred account and stocks in a taxable account. fischer and gallmeyer (2017) reinforce these findings in the context of mutual fund tax efficiency. garlappi and huang (2006) predict that the preference to hold bonds in a tax-deferred account can change when tax benefits are more uncertain. most recently, kobor and muralidhar (2020) developed and tested a goal-based, lifetime-income approach that has a better chance of success than a traditional glide-path approach using target-date funds. many authors also used life-cycle models that focus on both the accumulation and decumulation phases. lachance (2013) solves a life-cycle model for both the accumulation and decumulation phases using tax-deferred and tax-exempt accounts. they find that choosing a tax-deferred over a tax-exempt account creates a reduction in wealth, most noticeable for those with higher incomes and pensions, which increases social security taxation. marekwica et al. (2013) also examine a life-cycle model, including renting versus owning a home. they find that a tax arbitrage exists from mortgage interest payments and investing in 226 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 tax-deferred accounts. zhou (2009) developed a life-cycle model and finds that asset location decisions are very sensitive to a progressive tax system. the life-cycle model by zhou (2012) identified the optimal stock market participation between taxable and tax-deferred accounts, finding a preference for higher tax-deferred account participation earlier in life, and higher participation in taxable accounts later. 3. model in this section, we model the various accounts available to retirees, as well as other sources of income. we also describe our tax model and its basis in current tax law in the united states and other developed countries of the world. for details on u.s. rates and income limits, please see details in the appendix. there are two methods for defining this optimization problem. the first approach uses power-law utility functions, which may include a coefficient of relative risk aversion, such as those used in brown et al. (2017), sialm (2006), and others. then, future consumption is discounted at some discount factor. these models often include variables for pre-tax income in retirement, taxable accounts, tax-deferred, and tax-exempt retirement accounts. asset allocation in these models usually consists of a stock portfolio and a riskless bond. progressive tax brackets may be simplified and future tax rates and investment returns found using a bootstrapping method or a stochastic process model. these models provide strong evidence for optimal decision-making for large portions of the population and help inform policy. however, the use of utility functions prevents their direct application by fintech companies interested in building software tools needed by the unique circumstances that current and future retirees face. cook et al. (2015), al zaman (2008), dilellio and ostrov (2017, 2020), coopersmith and sumutka (2011), ragsdale et al. (1994), among others, used an alternative optimization approach. these authors impose an objective function to maximize portfolio longevity and/ or the size of the bequest. to varying degrees, these authors include additional aspects of their models to better reflect tax law. for instance, many include rmds, all current tax brackets, qualified versus non-qualified dividends, and the special treatment of capital gains taxes (see sumutka et al., 2012). we emphasize this latter approach here, as many retirees look to the financial services industry for retirement income planning software. in turn, many in the financial services industry look to academia to advance research that can support the development of algorithms necessary to improve their ongoing practice supporting their clients’ needs. 3.1. decision variables and objective function our model definition begins with our decision variables. for each year over a fixed retirement horizon of n years, a retiree has to decide how much of their retirement income comes from tax-deferred accounts, tax-exempt accounts, and taxable accounts. we define these decision variables mtaxable;t, respectively, as 1 � n vectors representing after-tax income j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 227 sources for each year t = 1 to n.8 in the following section, we provide details on assumptions on how these after-tax income sources are determined. next, we state the optimization problem to maximize our objective wealth function for retirees. max g 1� t heirð þwtax-deferred;n þwtax-exempt;n � �þwtaxable;n � � (1) eq. (1) represents the total wealth transfer to a retiree’s non-spouse heirs. here, their is the marginal tax rate of the retiree’s heir, so represents an estimate of the government’s share in this account, as described in reichenstein et al. (2012). the variables wtax-deferred;n and wtax-exempt;n are the tax-deferred account and tax-exempt account balances at the retirement horizon n, inflated by a factor g ≥ 1. this inflation factor represents the additional value in these retirement accounts if the non-spouse heir were to exercise the full stretch provision of 10 years, established by the secure act, as shown in dilellio and kinsman (2020). for countries that do not permit stretch provisions in these accounts, g ¼ 1: the variable wtaxable account;n is the taxable account balance, which assumes non-spouse heirs receive a full step-up in cost basis. we also neglect the effect of estate taxes. in the united states, and as reported by huang and cho (2017), affects about 0.2% of taxpayers. we assume the retiree may have other sources of income. this income may include social security or some other lifetime annuity annually adjusted at a rate it. this income may also include pension benefits that remain fixed in current-year dollars. we defined these variables as sst;taxable, sst;tax free, and pt, respectively, in retirement year t. 3.2. constraints this optimization problem in eq. (1) is subject to the inequality constraints for rmds and account balances. mtax-deferred;t ≥rmdt for 1≤ t≤n (2) etax-deferreed;t ≥ 0 for 1≤ t≤n (3) etax-exempt;t ≥ 0 for 1≤ t≤n (4) etaxable;t ≥ 0 for 1≤ t≤n (5) in eq. (2), we determine rmds from balances of the previous year’s tax-deferred account balances based on the retiree’s life expectancy let. if the retiree was born before june 30, 1949, then the retiree’s age at the end of the year, at, is compared with 70.5 years. otherwise, the age is compared to 72 years. we obtain life expectancy from the irs tables, form 590-b, appendix b.9 note that this table does not distinguish between men and 228 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 women, so our reference to retiree and spouse in the following calculations can be either gender. rmdt ¼ 0 for at < 70:5 or 72 etax-deferred;t�1 let for at > 70:5 or 72 ( (6) eqs. (3–5) represent account balances at the end of each retirement year and thus restrict negative balances in any account. we determine account balances each year using the real rate of return of the retiree’s investments and deduct any taxes owed. for the tax-deferred account, we find the end-of-the-year account balance in year t, etax�deferred;t, as etax-deferred;t ¼ etax-deferred;t�1 � mtax-deferred;t þ ttax-deferred;tð þ½ � � 1þmtð þ: (7) the term mt is the time-dependent real rate of return of a stock-bond portfolio, and ttax-deferred;t are the income taxes due from the tax-deferred account distribution. because there may also be retiree income tax due to their social security and pension benefits received, ttax�deferred;t is found after accounting for this other income. our model assumes that 85% of social security benefits are taxed as ordinary income, which is an upper bound on the amount of tax that could be due from social security. otherwise, the value would not solely depend on both tax-deferred account distributions, increasing the complexity of the optimization problem. all examples shown in the following sections satisfy this condition, which is an individual with a modified adjusted gross income of at least $34,000 or at least $44,00 for a married couple filing jointly. for the tax-exempt account, we find the end-of-the-year account balance in year t, etax-exempt;t, as etax-exempt;t ¼ etax-exempt;t�1 �mtax-exempt;t½ � 1þmtð þ: (8) note that we assume that the asset allocation in the tax-deferred account is also applied to the tax-exempt account, so the tax-exempt account grows at the same rate mt each year. we shall also apply this asset allocation to the taxable account. therefore, we determine the taxable account end of year balance in year t, etaxable;t, as etaxable;t ¼ etaxable;t�1 � mtaxable;t þ tgains;tð þ½ � 1þmtð þ: (9) we model our capital gains tax calculation of tgains;tin the following ways. first, we assume all gains are long-term, so we tax these gains at the lower capital gains rate, which is very likely for retirees that are not actively trading assets in their taxable account. tax brackets used in this paper appear in the appendix table a1 and a2, for u.s. taxpayers. however, other developed countries also use a progressive tax system, such as the united kingdom, germany, canada, and australia, among others. therefore, while the results that j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 229 follow will implement only u.s. tax law, we expect these results to be generalizable to any country with a progressive tax system and different rates for capital gains. we also assume a single cost basis for the taxable account. in fact, a retiree may have purchased stocks or bonds at multiple basis, and could realize additional tax savings if higher basis investments are liquidated first. however, it is also possible that a retiree may not be aware of this fact, and liquidate lower basis investment first, increasing their tax liability. therefore, the single cost basis represents a weighted average of all cost basis. the framework of the model could be expanded to include multiple cost bases, but the generalization of the findings would be unchanged. consequently, we do not attempt to reinvest stock dividends and bond coupon payments, but rather use them to satisfy current-year retirement income needs. in the event that tax-deferred account withdrawals, along with other interest income, pensions, and social security exceed the income needed, we transfer these excess amounts to a zero-dividend investment. using a zero-dividend investment minimizes annual taxes on dividends while maximizing the step-up in cost basis. this approach is similar to the strategy employed by cook et al. (2015), except we minimize the tax drag of dividends in this separate taxable account, which will be transferred in its entirety to the heirs. next, we assume a “glide path” asset allocation of stock and bonds in the taxable account identical to the tax-deferred and tax-exempt accounts, which is consistent with the practice among financial advisors. additionally, varying asset allocations between accounts can bias optimal drawdowns, delaying account drawdowns with higher stock allocations due to their higher expected returns. the glide path is initially set based on the retiree’s age so that stock allocation is equal to 120-age. thus, a 60-year-old retiree starts with an asset allocation of 60% stock and 40% bonds, with a 1% shift from stocks to bonds each year of retirement. a similar glide path is set for the spouse, in case their age differs. we do not include investor sentiment to market risk, which is likely to affect their asset allocation decisions during retirement. eqs. (7–9) above can accommodate time variation, so our model’s framework is sufficiently flexible to accommodate this effect. lastly, we assume stock dividends are all qualified, so they are taxed as a capital gain, like the ishares core s&p 500 index etf (ticker: ivv). alternatively, we assume bond income is taxed as ordinary income, like the ishares aggregate bond index etf (ticker agg). because return uncertainty is not included in this model, we do not model tax-loss harvesting of either short-term or long-term losses on taxable account investments. however, we do generalize our results using a sensitivity analysis that varies the expected return (among other inputs) to capture the effect of different long-run stock and bond total return performances. we show that the optimization is largely insensitive to stock returns and bond returns. fig. 1 provides a graphical representation of these assumptions. this model does include many aspects of the tax systems throughout the world, such as deductions, income tax, and capital gains taxes stack on top of one another. however, other elements, like the u.s. alternative minimum tax (amt), the united states 3.8% medicare tax, and local province or state taxes are not included. sumutka et al. (2012) developed and tested a tax-efficient model that included these additional elements for u.s. retirees. as these taxes are almost always much smaller than those imposed at a national level, we do not expect our findings to change significantly if these country-specific taxes were imposed. 230 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 finally, the equality constraint for the after-tax retirement income needed, ct, is pt þ sst;taxable þ dt � tð þ þmtda;t þ d�t � tqdiv;t þmta;t � �þmtea;t þ sst;tax free ¼ ct: (10) the first terms represent income taxed as ordinary income, where pt is the pre-tax pension, sst,taxable is pre-tax social security, and dt are pre-tax bond coupon payments from the brokerage account. the ordinary income tax paid on these three income sources is t. the next terms in parentheses are the pre-tax qualified dividends d�t from the taxable brokerage account and the taxes paid on the qualified dividends tqdiv;t. the term sst;tax free is the taxfree portion of social security benefits. note that in eq. (10), time-depended retirement income can be accommodated by this model, consistent with retirees differing views of using retirement assets for generating income, supporting health, and intergenerational planning as shown in chambers et al (2021). 3.3. withdrawal strategies we implement three strategies to solve the optimization problem described in the previous section. we find the optimal strategy depends on the relationship between a retiree’s net worth (including the present value of annuities) and the retiree’s desired retirement income. fig. 2 presents a summary of this finding that will be supported by results in the next section. in the three categories appearing in fig. 2, the common rule strategy, as described in coopersmith and sumutka (2011), provides an important heuristic in selecting an optimal strategy. here, we define the common rule strategy as one that exhausts rmds first, then taxable account funds, then tax-deferred account funds, then lastly taxexempt funds. fig. 1. conceptual model for retirement income sources and taxation. in this illustration, the retiree’s deductions exceed their income from social security, pension, and taxable bond interest. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 231 referring to the upper left of fig. 2, this category corresponds to a situation where the common rule strategy exhausts all funds over the retirement horizon. dilellio and ostrov (2017) show that portfolio longevity can be extended by more tax-efficient planning that includes knowledge about the heir’s tax rate. cook et al. (2015) show that the most effective approach to possibly extending portfolio longevity is careful roth conversions during either preor early retirement years. in the following section, we show that the modified common rule withdrawal strategy provides a similar benefit of three additional years found in cook et al. (2015) by first drawing down from the tax-deferred account up to the heir’s tax rate. the modified common rule strategy is one that initially withdraws tax-deferred account funds up to the heir’s tax bracket, before using taxable account funds. otherwise, the modified common rule strategy is identical to the common rule strategy. we also show that the modified common rule produces 0.35% of additional return over the common rule strategy. the lower left and upper right in fig. 2 represents the most significant retiree category for increasing a bequest. we identify this category by applying the common rule strategy and seeing that there are sufficient, but not excessive funds, in meeting retirement income needs. put another way, the retiree cannot solely fund their retirement income without voluntary withdrawals from their tax-deferred, tax-exempt, or taxable accounts. in the following section, we show that using the sequential least squares programming (slsqp) withdrawal strategy can add the equivalent of 0.54% to the annual return of the common rule strategy. the slsqp withdrawal strategy uses non-linear constrained optimization methods to solve this optimization problem. the last category appears in the lower right of fig. 2 when a retiree has some years when rmds plus other income sources like dividends, pensions, and social security exceed their income needs. in this case, the optimal strategy is again the modified common rule, which takes full advantage of their tax-deferred account up to the tax bracket of their heir. note that the slsqp strategy is often not feasible here because the equality constraint in eq. (10) may not be satisfied in some retirement years. in this last category, we model this so-called fig. 2. the optimal withdrawal strategy depends on the relationship between the retirees net worth + present value of annuities and their retirement income needs * their retirement horizon. 232 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 “excess income” as a direct contribution to another taxable brokerage account that is invested in a zero-dividend stock, to minimize the corrosive effects of dividends, as suggested by demuth (2016). investing a retiree’s excess income in this way maximizes the benefits of the step-up in cost basis realized by a retiree’s non-spouse heirs. 4. results the following subsections demonstrate the results using the three categories that appear in fig. 2. using the common rule withdrawal strategy, three outcomes occur. 1. retiree and spouse exhaust all account assets over their retirement horizon. 2. retiree and spouse have sufficient assets to fund their retirement but some of their account assets must be drawn down. 3. retiree and spouse have an excess of assets, generating more income from dividends, interest, coupon payments, and rmds than is needed over their retirement horizon. 4.1. retiree and spouse run out of funds over their retirement horizon table 1 shows the assumptions that, after following the common rule strategy, exhaust all account funds. we use the average spousal age differential of 3 years.10 these results include the effect of a surviving spouse, who can realize a step-up in cost basis because we assume they reside in a community property state. we also assume these retirees need to begin taking their rmds in the year they turn 70.5.11 the implied life expectancy of 95 and 100 years in table 1 are conservative estimates often used by financial planners. we investigate shorter life expectancies in the following subsections. we determine the cost basis for stocks and bonds using the nominal return found lower in the table, to be consistent with current assumptions about capital markets. using the ishares s&p 500 etf (ticker ivv) as a proxy for stock investing, its lifetime annualized return (from may 15, 2000, to april 27, 2021) was 7.2%. similarly, the ishares aggregate bond index etf (ticker agg) was used as a proxy for bond investing, and its lifetime annual return (from september 22, 2003, to april 27, 2021) was 4.0%. with these returns in mind, a taxable investment in each of these funds 10 years before retirement yields cost bases of 50% for stocks and 68% for bonds. additionally, we assume a 1% transition from stocks to bonds in retirement, which follows conventional wisdom. to provide support for empirical observations by gerrans et al. (2009) and investment theory by samuelson (1969), we include a sensitivity analysis with this input parameter. we set social security to its median value, assume it grows at the rate of inflation appearing lower in table 1, and begins at the full retirement age of 67.12 similarly, we set pension benefits to their median value of $11,000 per year for private and union pensions, but do not adjust its annual payment for inflation.13 we derive the inflation rate using the long-term average from the consumer price index from 1992 to 2020. however, given how much inflation may change, we include this input as part of the sensitivity analysis explored in the later sections. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 233 using the variables in table 1, the retiree’s spouse exhausts all their funds in the 37th year, for portfolio longevity of 36.5 years. the upper pane of fig. 3 shows how the retiree and spouse use each source of after-tax income to meet the annual income needs using the common rule withdrawal strategy. here, all taxable account funds (orange) are exhausted first, then tax-deferred account funds (blue), and tax-exempt account funds last (green). also, note that the rmds (yellow bars overlapping the blue) appear completely over the tax-deferred account bars when representing a binding constraint in years 5 through 17. lastly, income tax brackets are constant each year, because all the results are in the current year, not future dollars.14 however, income brackets drop in year 31, when the surviving spouse must file their taxes as “single.” the 0% tax bracket corresponds to the income below the standard deduction in the united states or income below the so-called exclusion limit for retirees paying taxes outside the united states. the lower pane of fig. 3 shows the modified common rule withdrawal strategy increasing portfolio longevity to 39.7 years or about 3.2 years. put another way, the common rule withdrawal strategy would need to increase its investment returns by 0.35% per year to achieve the same portfolio longevity.15 recall that the modified common rule strategy first uses tax-deferred account assets up to the top of the income bracket associated with greater efficiency relative to the heir. in this case, the heir’s marginal tax rate is 25%, so tax-efficient tax-deferred account withdrawals can occur up to the top of the 24% bracket, but not the bracket above it corresponding to the 32% income tax rate. the most noticeable difference between the upper and lower panes in fig. 3 is the rmds are no longer binding during any table 1 input variables for a retiree and spouse that lead to insufficient funds for retirement age of retiree, spouse 65, 62 after-tax income needed, per year $150k (retiree + spouse), $140k (surviving spouse) retirement time horizon, retiree, and spouse 30 years (retiree), 40 years (spouse) community property state? yes account information retiree and spouse tax-deferred account starting balance $800,000 and $100,000 retiree and spouse tax-exempt account starting balance $400,000 and $50,000 taxable account starting balance $1,000,000 cost basis 50% (stocks), 68% (bonds) initial asset allocation 60% stocks, 40% taxable bonds glide slope transition from stock to bonds 1% per year filing status married filing jointly, then single the deduction, tax year the standard deduction, 2020 retiree’s social security amount and age to start $18,500 at age 67 spouse social security amount and age to start $18,500 at age 67 retiree’s pension and start age $3,667 at age 65 spouse’s pension and start age $3,667 at age 65 stretch tax-deferred account withdrawals for 10 years no heir’s tax rate 25% inflation 2.1% nominal return, stocks 7.2% nominal return, bonds 4% stock dividend rate 2% bond coupon rate 2.5% 234 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 year. there is also a longer duration of stock dividends (gray) and bond interest (light blue) contributing to retirement income needs because the taxable account assets are not used up as quickly in the lower pane of fig. 3. fig. 4 shows how up to four different taxes are paid each year of retirement: two sources of income taxes and two sources of capital gains taxes. in the early retirement years, using the common rule strategy in the upper pane, the retiree pays very little tax, because the income tax from ordinary income barely exceeds the standard deduction, and no tax-deferred account withdrawals occur yet. additionally, the retiree and spouse can realize gains by selling stock and bond shares in their taxable brokerage account that, with an initial fraction of their cost basis shown in table 1, does not trigger any long-term capital gains taxes. this fig. 3. meeting after-tax income using the common rule withdrawal strategy (upper pane) and modified common rule withdrawal strategy (lower pane) using variables in table 1. the retiree and spouse exhaust all their accounts. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 235 small capital gains tax liability is thanks to the fairly large 0% capital gains bracket that is mostly available due to limited ordinary income. in the years 5 to 17, rmds initiate and are a binding constraint. also, long-term capital gains are realized above the 0% capital gain rate, so taxes become more significant. once the retiree exhausts the taxable brokerage account in year 18, retirement income shifts to withdrawals entirely from the tax-deferred account, significantly increasing income taxes paid in years 19–25. the retiree exhausts the tax-deferred account in year 26. in the later years, taxable income barely exceeds the standard deduction, thanks to the funds from the tax-exempt account, so minimal taxes are due. the lower pane of fig. 4 shows how shifting taxes to earlier years increased the portfolio longevity. in years 1–10, the retiree uses tax-deferred funds instead of taxable brokerage account funds. the modified common rule withdrawal strategy identifies their use because fig. 4. taxes paid with the common rule and modified common rule withdrawal strategies using conditions in table 1. the retiree and spouse exhaust all their accounts. 236 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 they are more efficient than the possible transfer to the heir, whose tax rate is assumed at 25% in table 1. then, in the later years, the retiree pays some long-term capital gains taxes, along with a small amount of income tax. after year 31, the taxes paid are identical to the common rule withdrawal strategy. therefore, a key insight here is that portfolio longevity increases if one defies conventional wisdom to always defer taxes until future years. these results demonstrate that withdrawing from the tax-deferred account earlier than the rmds suggest can provide real benefits. in this case, the longevity increased by about 3 years. or, in terms of investment returns, provides a tax alpha of about 0.35%. in the sensitivity section that follows, we show that these results are mostly generalizable across several varying conditions changed from table 1. 4.2. retiree and spouse have sufficient funds for their retirement in this case, the retiree has a sufficient, but not excessive, net worth. recall, sufficient retirement funds mean assets must be drawn down at some point in retirement. excessive means that interest, coupons, and dividends, along with rmds, exceed the retiree’s income needs at some point during retirement. therefore, there is an opportunity to increase their bequest. the retiree and spouse’s assumptions are identical to the values appearing in table 1, except we reduced the retirement horizons to 15 years (retiree) and 20 years (spouse). table 2 summarizes the final account values under three withdrawal strategies. it shows that the increase in wealth transfer to non-spouse heirs is 22% larger with the slsqp withdrawal strategy over common rule. for the common rule strategy to produce the same wealth transfer amount as the slsqp strategy, asset returns would need to increase by 0.54% per year. the drawdown decisions for the common rule and slsqp withdrawal strategies appear in fig. 5 upper and lower panes. there are several unique differences in these decisions. first, for the slsqp withdrawal strategy, tax-deferred account withdrawals (in blue) occur much earlier, so rmds are no longer binding constraints. also, tax-deferred account withdrawals occur throughout retirement, helping to prevent a higher level of income tax in the later years, especially for the surviving spouse filing as single in years 16–20. lastly, the retiree uses the taxable brokerage account much more sparingly so that the heir can realize the step-up in cost basis from this account. table 2 account balances at the end of retirement for common rule, modified common rule, and slsqp withdrawals when the retiree and spouse have sufficient, but not excessive, funds for retirement strategy taxdeferred account, retiree tax-exempt account, retiree taxable brokerage balance tax-deferred account, spouse tax-exempt account balance, spouse wealth transfer to heirs common rule $461,937 $798,355 $0 $57,742 $99,794 $1,287,908 modified common rule $0 $798,355 $632,418 $0 $99,794 $1,530,567 slsqp $20,186 $265,305 $1,255,685 $2,523 $33,163 $1,571,185 note: slsqp = sequential least squares programming. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 237 the upper and lower panes of fig. 6 show the significant difference in taxes paid between the common rule and slsqp withdrawal strategies. the most obvious difference is that the retiree pays taxes much earlier in the slsqp versus the common rule strategy. however, notice that the total taxes paid by the slsqp strategy are in the low $20,000 range for the first seven years, then dropping to between about $5,000 and $10,000 for the remaining retirement horizon. the common rule strategy follows a more “conventional wisdom” recommended by many cpas that often suggest delaying taxes for as long as possible. under the common rule withdrawal strategy, the retiree pays very little tax in years 1–4. like in the previous fig. 5. meeting after-tax income using the common rule and sequential least squares programming (slsqp) withdrawal strategies and conditions in table 1 with a shorter retirement horizon. the retiree and their spouse have sufficient, but not excessive, funds for retirement. 238 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 subsection, low taxes are due to the initial fraction of the taxable account cost basis shown in table 1 not triggering any long-term capital gains taxes. in years 5–15, there is a gradual increase in taxes from both the rmds and long-term capital gains. years 16–17 see a decrease in taxes because the taxable brokerage account offsets the higher tax rates of the surviving spouse. the final five years 18–20 see a significant tax increase from withdrawals from the tax-deferred account and the surviving spouse filing taxes as single. therefore, in these final years, the surviving spouse pays nearly $35,000 of income tax to maintain their desired consumption of $140,000 per year. these later years show how sensitive the surviving spouse is to losing the larger standard deduction by no longer being able to file their taxes as married filing jointly. fig. 6. taxes paid with the common rule withdrawal sequence and sequential least squares programming (slsqp) withdrawal strategies and conditions in table 1 with a shorter retirement horizon. the retiree and their spouses have sufficient, but not excessive, funds for retirement. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 239 comparing the upper and lower pane of fig. 6 indicates that tax minimization may not occur. this finding is consistent with sumutka et al. (2012), who noted that “. . . several common tax minimization and estate planning strategies do not produce optimal results.” these results provide further evidence that simply minimizing taxes paid does not lead to a larger bequest to the retiree’s heir. 4.3. retiree and spouse have a large excess of funds in this last case, the retiree and spouse have a large net worth with more income from dividends, interest, coupon payments, and rmds than is needed at some point over their retirement horizon. here, the modified common rule provides a solution marginally better than the common rule. in this next example, we begin by using the data from table 1. like in the previous subsection, we reduce the retirement horizon to 15 years for the retiree and 20 years for the spouse. we also increase the retiree’s tax-deferred account to start at $2.7m, and consumption while married filing jointly to $180,000 per year. the account values at the end of retirement using the common rule and modified common rule withdrawal strategies appear below in table 3. reviewing the results from table 3, we see that the modified common rule strategy increases common rule wealth transfer by about 7%. increasing the stock and bond returns by 0.24% with the common rule strategy produces the same wealth as the modified common rule strategy. this example is also one that highlights the limitations of general-purpose constrained optimization programs like slsqp. if we apply the slsqp withdrawal strategy here, it does find an optimal solution that satisfies the equality constraint in eq. (10). however, the total wealth transfer to heirs is $3,764,923, which is less than found with the modified common rule. thus, the slsqp strategy appears to have gotten “stuck” in a local maximum, so it did not yield a global maximum. and, if the slsqp withdrawal strategy could not find a solution, which may occur when the equality constraint in eq. (10) is unmet, this general-purpose optimization algorithm could spend 10–15 minutes and report back that no solution could be found. the deployment of any enterprise application that consumes precious computational resources while not providing actionable insights must be avoided. thus, we have table 3 account balances at the end of retirement for common rule and modified common rule withdrawals strategy taxdeferred account, retiree tax-exempt account, retiree taxable brokerage balance taxdeferred account, spouse tax-exempt account balance, spouse excess income account balance wealth transfer to heirs common rule $2,501,404 $798,355 $617,383 $92,645 $99,794 $73,004 $3,534,072 modified common rule $1,594,256 $798,355 $1,629,133 $59,047 $99,794 $0 $3,767,258 note: retiree and spouse have excessive amounts of retirement assets. 240 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 evidence that there is a need for a more customized optimization algorithm for this category of retirees, as we demonstrate with the modified common rule strategy. fig. 7 presents the retirement income sources used to satisfy the retiree and spouse’s income needs, then later the surviving spouse’s income needs. the upper panel shows the results using the common rule withdrawal strategy, where rmds are a binding constraint in years 5–20. the rmds produce excess income as mentioned previously. to minimize the effect of taxes, income over $140,000 in years 14–20 is reinvested in a zero-dividend stock mutual fund or etf. we assume this zero-dividend investment made in these later years has a nominal return equal to the stock investment in the retiree’s other investment accounts.16 fig. 7. meeting after-tax income using the common rule and modified common rule withdrawal strategies and conditions in table 1 with shorter retirement horizon, higher tax-deferred account starting account balances, and higher retiree and spouse income needs. retirees and spouses have excessive amounts of retirement assets. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 241 the lower panel of fig. 7 shows drawdowns with the modified common rule, where all rmds from the tax-deferred account are kept as a non-binding constraint. it also prevents excess income from occurring in years 14–20 for the surviving spouse. fig. 8 shows the taxes paid corresponding to the retirement income decisions shown in fig. 7. similar to what appeared in fig. 6, the optimal decision increases taxes paid in earlier retirement years. however, in later years, taxes paid with the modified common rule fall below those paid with the common rule. thus, we see another example where the benefit of the step-up in cost basis provided by the taxable brokerage account outweighs the additional and earlier taxes paid by the retiree. fig. 8. taxes paid with the common rule withdrawal sequence and modified common rule withdrawal strategies and conditions in table 1 with shorter retirement horizon, higher tax-deferred account starting account balances, and higher retiree and spouse income needs. retirees and spouses have excessive amounts of retirement assets. 242 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 5. sensitivity analysis of the heir's tax rate we begin by revisiting the results from the previous subsection to see how portfolio longevity increases with the modified common rule strategy and changes with the heir’s tax rate. recall that the modified common rule strategy is identical to the common rule, except it draws down the tax-deferred account initially up to the heir’s tax bracket. also, note that the common rule strategy is not affected by changing the heir’s tax rate. the results in table 4 suggest that drawing down the tax-deferred account faster is more efficient for the retiree than the heir to pay income taxes on these distributions. however, because the retiree exhausts all of their accounts, delaying this drawdown when the retiree has an heir at a lower tax rate provides little benefit. this is primarily because rmds for these distributions are less tax-efficient than could be realized by an heir with a lower marginal tax rate. in the extreme case when the heir is a qualifying charitable organization with a 0% tax rate, the modified common rule strategy provides only a fraction of a year of additional longevity. we next examined how the effect changing the heir’s tax rate can have on the results in tables 2 and 3. table 5 summarizes the results that vary the heir’s tax rate away from the 25% rate used to produce table 2 when the retiree and spouse have sufficient retirement assets. we also applied the modified common rule withdrawal strategy to each of these examples, and we found the slsqp strategy always provided a higher wealth transfer to the heir. the primary observation here is that the slsqp withdrawal strategy provides the greatest improvement when the heir’s tax rate is higher. table 6 provides a similar sensitivity analysis, using conditions used to create table 3 when the retiree and spouse have excessive retirement assets. we also applied the slsqp withdrawal strategy but found that the modified common rule strategy was superior in all cases we examined. also, the slsqp had very long run times in each case in table 6. table 4 portfolio longevity sensitivity (years) to heir’s tax rate sheir common rule modified common rule improvement vs. common rule equivalent tax alpha 0% 36.5 36.8 0.3 0.03% 15% 36.5 37.5 1.0 0.11% 25% 36.5 39.7 3.2 0.35% 35% 36.5 39.7 3.2 0.35% note: base case results use inputs from table 1 with t heir = 25%. retiree and spouse have insufficient funds for retirement. table 5 wealth transfer sensitivity to heir’s tax rate sheir common rule slsqp $ improvement vs. common rule equivalent tax alpha 0% $1,417,828 $1,544,800 $126,972 0.21% 15% $1,339,876 $1,524,163 $184,287 0.33% 25% $1,287,908 $1,571,185 $283,277 0.54% 35% $1,235,940 $1,568,913 $332,973 0.69% note: base case results use inputs from table 1, but shortened retiree’s horizon to 15 years and spouse’s horizon to 20 years. retiree and spouse have sufficient retirement assets. slsqp = sequential least squares programming. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 243 similar to table 4, table 6 shows that the modified common rule withdrawal strategy provides the greatest improvement when the heir’s tax rate is higher. tables 4, 5, and 6 indicate that the common rule withdrawal strategy can provide an important diagnostic. recall the results from table 5 were based on the common rule that produced no excess income, or involuntary income above the income need. in all situations where this occurred, the slsqp strategy always produced superior results. similarly, tables 4 and 6 showed that the modified common rule strategy was superior in the presence of insufficient or excess income. and, in the case of table 6, provided significantly shorter computational run times and guaranteed a solution that satisfied all constraints. 5.1. sensitivity analysis to understand the effect of selected parameters, we recalculated the results from the previous sections, using high and low values that appear below in table 7. we changed each value throughout the forecast, except for the values in the second to the last row. in the row labeled “percent increase to tax rates,” we assumed it begins in 2026, which is the year the 2017 tax cuts and jobs act (tcja) expires. we chose the high and low values based on ranges observed from historical averages. for an excellent history of income tax rates from 1916 to 1999 (see sialm (2006). fig. 9 shows how the tax alpha of 0.35% changes using the modified common rule strategy changes. the tax alpha is most sensitive to increasing tax rates, exceeding 0.50% if tax rates increase in the future. increasing the allocation towards bonds also increases tax alpha, as there is a greater opportunity for tax efficiency. interestingly, there appears to be little sensitivity to stock and bond rate of returns, indicating the robustness of tax alpha to longterm market return trends. fig. 10 shows the sensitivity of the 0.54% tax alpha when the retiree and their spouse shorten their retirement horizon to 15 and 20 years. tax alpha is the most sensitive future table 6 wealth transfer sensitivity to heir’s tax rate sheir common rule modified common rule $ improvement vs. common rule equivalent tax alpha 0% $4,182,585 $4,183,990 $1,405 < 0.01% 15% $3,793,477 $3,935,081 $141,604 0.14% 25% $3,534,072 $3,767,259 $233,187 0.23% 35% $3,274,667 $3,601,928 $327,261 0.34% note: base case results use inputs from table 1, but shortened retiree’s horizon to 15 years and spouse’s horizon to 20 years, as well as increasing the retiree’s tax-deferred account to start at $2.7m, and increasing consumption while married filing jointly to $180,000 per year. retiree and spouse have excessive retirement assets. table 7 baseline, low and high values for sensitivity analysis low value baseline high value asset allocation to stocks (bonds) 50% (50%) 60% (40%) 70% (30%) stock rate of return 6% 7.2% 8.4% bond rate of return 3% 4% 5% inflation rate 1.5% 2.1% 4.2% percent change to tax rates starting in 2026 �40% 0% +40% asset allocation glideslope 0% 1% 2% 244 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 tax rate, as well as to glide path. it is also very sensitive to the retiree’s initial stock allocation, and least sensitive to stock and bond returns. finally, fig. 11 shows that tax alpha for retirees with excessive retirement income is most sensitive to inflation, followed by a glide path, and the tax increases after the tcja expires. tax alpha is least sensitive to stock and bond returns. 5.2. implications to practice large financial institutions, like fidelity (2014) and vanguard (2016), currently provide retirement income planning tools that rely solely on the common rule withdrawal strategy. our findings suggest that their enterprise applications should not be entirely discarded due to their heuristic benefits. instead, results from the common rule strategy can be used to guide the next level of optimization that has a substantial benefit to tax-efficient retirement plans. some fintech firms, including personal capital, betterment, and others, are beginning to move in this direction. we hope this article’s insights provide the financial services industry with guidance to enhance and improve outcomes for retirement income plans. our findings also challenge the conventional wisdom advocated by many cpas to defer taxes for as long as possible. we show the value of strategically paying some taxes earlier to avoid large taxes later due to rmds or switching tax filing from married to single. lastly, we show that tax alpha or the additional pre-tax return realized by an optimal strategy is usually most fig. 9. sensitivity analysis for retirees with insufficient retirement funds. fig. 10. sensitivity analysis for retirees with sufficient retirement funds. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 245 sensitive to future tax rates, and least sensitive to the rate of return for the stock market. this latter finding suggests simulation of stock returns is likely unnecessary to quantify the tax alpha. 6. conclusions in this article, we developed a framework for finding the optimal decision in tax-efficient retirement income. we showed that optimizing a retirement income plan that begins with using the common rule withdrawal strategy can offer important insights into the next stage of this optimization problem. we identified three categories of retirement income plans and showed how there is no “one size fits all” solution to the problem. we provide model-driven evidence that general-purpose optimization routines can provide significant improvements over the common rule in one category, while not in all. we provide a simple heuristic to show how alternative optimization methods perform and include a sensitivity analysis to generalize the results. contrary to conventional wisdom, we also show that paying income taxes earlier due to tax-deferred account distributions can be more tax-efficient, contributing about 0.5% of additional annual returns. we also demonstrated how these three categories of tax-efficient withdrawal strategies performed under differing assumptions on investment returns, inflation, and tax rates. we closed on the implications these findings have on large financial institutions and smaller fintech startups in the financial planning space, as well as insights for cpas involved in multiyear tax planning. notes 1 https://www.stlouisfed.org/on-the-economy/2019/may/how-many-people-will-beretiring-in-the-years-to-come 2 https://en.wikipedia.org/wiki/mid-20th_century_baby_boom 3 http://benna401k.com/401k-history.html 4 https://www.newretirement.com/retirement/what-is-a-403b/ fig. 11. sensitivity analysis for retirees with excess retirement funds. 246 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 5 in 2019, the secure act delayed the onset of rmds from 70.5 to 72 for individuals born after june 30, 1949. 6 in the united states, 85% of social security is taxable income if magi exceeds $34,000 (single) or $44,000 (married filing jointly). 7 https://docs.scipy.org/doc/scipy/reference/optimize.minimize-slsqp.html 8 after-tax drawdowns are more helpful in supporting after-tax income needs, so all decision variables are after-tax values. 9 https://www.irs.gov/pub/irs-pdf/p590b.pdf 10 https://www.rgj.com/story/money/business/2018/11/28/retirement-planningchallenges-age-gap-relationships/2142317002/ 11 changing their rmds to the year they turn 72 does not have a meaningful effect on the results. (longevity = 33.5 with common rule and 37.4 with modified common rule) 12 https://www.aarp.org/retirement/social-security/questions-answers/how-muchsocial-security-will-i-get.html 13 http://www.pensionrights.org/publications/statistic/income-pensions/ 14 this approach to keeping income brackets fixed each year means they are constant in real dollars. it is also consistent with the long-standing practice at the irs annually increasing brackets by the rate of inflation. 15 we found the value of 0.35% by re-running this case with the common rule withdrawal strategy and gradually increasing the stock and bond returns until the portfolio longevity was 39.7 years. 16 the largest zero-dividend etf is first trust dow jones internet index etf, ticker: fdn, with over $10b in management. appendix income and capital gains tax rates in the united states and other developed countries table a1 u.s. income tax rates and brackets, 2020 tax year income tax rate income limit, single income limit, married filing jointly 10% $9,875 $19,750 12% $40,125 $80,250 22% $85,525 $171,050 24% $163,300 $326,600 32% $207,350 $414,700 35% $518,400 $622,050 37% none none j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 247 acknowledgments this research was supported by the julian virtue professorship endowment at the graziadio business school, pepperdine university. we are also grateful for our research assistants on this project, airika corley and sophia venegas. references brown, d. c., cederburg, s., & o’doherty, m. s. (2017). tax uncertainty and retirement savings diversification. journal of financial economics, 126, 689-712. chambers, b., walker, r., feng, j., & gu, y. (2021). the silver tsunami: an enquiry into the financial needs, preferences and behaviours of retirees. accounting & finance, 61, 645-687. cook, k. a., meyer, w., & reichenstein, w. (2015). tax-efficient withdrawal strategies. financial analysts journal, 71, 16-29. coopersmith, l. w., & sumutka, a. r. (2011). tax-efficient retirement withdrawal planning using a linear programming model. journal of financial planning, 24, 50-59. coopersmith, l. w., & sumutka, a. r. (2017). the impact of rates of return on roth conversion decisions and retiree savings wealth. journal of personal finance, 16, 29-39. dammon, r. m., spatt, c., & zhang, h. h. (2004). optimal asset location and allocation with taxable and taxdeferred investing. the journal of finance, 59, 999-1037. demuth, p. (2016). the overtaxed investor: slash your tax bill & be a tax alpha dog. los angeles, ca: alpha dog press. dilellio, j. a., & ostrov, d. n. (2017). optimal strategies for traditional versus roth ira/401(k) consumption during retirement. decision sciences, 48, 356-384. dilellio, j. a., & ostrov, d. n. (2020). toward constructing tax efficient withdrawal strategies for retirees with traditional 401(k)/iras, roth 401(k)/iras, and taxable accounts. financial services review, 28, 67-95. dilellio, j. a., & kinsman, m. (2020). the secure act and your retirement objectives. graziadio business review, 23. fidelity. (2014). detailed methodology: fidelity retirement income planner. available at http://personal. fidelity.com/planning/retirement/pdf/rip_methodology.pdf. fischer, m., & gallmeyer, m. (2017). taxable and tax-deferred investing with the limited use of losses. review of finance, 21, 1847-1873. garlappi, l., & huang, j. (2006). are stocks desirable in tax-deferred accounts? journal of public economics, 90, 2257-2283. geisler, g., & hulse, d. (2018). the effects of social security benefits and rmds on tax-efficient withdrawal strategies. journal of financial planning, 31, 36-47. gerrans, p., clark-murphy, m., & speelman, c. (2009). asset allocation and age effects in retirement savings choices. accounting & finance, 50, 301-319. kobor, a., & muralidhar, a. (2020). targeting retirement security with a dynamic asset allocation strategy. financial analysts journal, 76, 38-55. lange, j. (2009). retire secure! hoboken, nj: wiley. lachance, m. (2013). roth versus traditional accounts in a life-cycle model with tax risk. journal of pension economics and finance, 12, 28-61. table a2 u.s. capital gains tax rates and brackets, 2020 tax year capital gains tax rate income limit, single income limit, married filing jointly 0% $40,000 $80,000 15% $441,450 $496,600 20% none none 248 j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 marekwica, m., schaefer, a., & sebastian, s. (2013). life cycle asset allocation in the presence of housing and tax-deferred investing. journal of economic dynamics and control, 37, 1110-1125. horan, s. m. (2006). withdrawal location with progressive tax rates. financial analysts journal, 62, 77-87. huang, c. c., & cho, c. (2017). ten facts you should know about the federal estate tax. center on budget and policy priorities, 1-7. meyer, w., & reichenstein, w. (2013). the tax torpedo: coordinating social security with a withdrawal strategy to minimize taxes. retirement management journal, 3, 25-32. milesvsky, m. (2020). retirement income recipes in r. switzerland: springer. piper, m. (2013). can i retire? st. louis, mo: simple subjects, llc. reichenstein, w., & meyer, w. (2018). understanding the tax torpedo and its implications for various retirees. journal of financial planning, 31, 38-45. reichenstein, w., horan, s. m., & jennings, w. w. (2012). two key concepts for wealth management and beyond. financial analysts journal, 68, 14-14. rodgers, r. (2009). the new three-legged stool. hoboken, nj: wiley. sialm, c. (2006). stochastic taxation and asset pricing in dynamic general equilibrium. journal of economic dynamics and control, 30, 511-540. solin, d. r. (2010). the smartest retirement book you’ll ever read. new york, ny: perigee trade. sumutka, a. r., sumutka, a. m., & coopersmith, l. w. (2012). tax-efficient retirement withdrawal planning using a comprehensive tax model. journal of financial planning, 25, 41-52. poterba, j. m. (2014). retirement security in an aging population. american economic review, 104, 1-30. pfau, w., tomlinson, j., & vernon, s. (2017). optimizing retirement income by integrating retirement plans, iras, and home equity. schaumburg, il: society of actuaries. available at https://www.soa.org/493461/ globalassets/assets/files/resources/research-report/2017/2017-optimize-retire-income-plans-iras-equity.pdf ragsdale, c. t., seila, a. f., & little, p. l. (1994). an optimization model for scheduling withdrawals from tax deferred retirement accounts. financial services review, 3, 93-108. samuelson, p. a. (1969). lifetime portfolio selection by dynamic stochastic programming. the review of economics and statistics, 51, 239-246. vanguard. (2016). how to turn retirement savings into retirement income. valley forge, pa: the vanguard group, inc. available at https://retirementplans.vanguard.com/segment/pdfs/retincshortbro.pdf welch, j. s. (2016). measuring the financial consequences of ira to roth ira conversions. journal of personal finance, 15, 47-55. welch, j. s. (2017). a 3-step procedure for computing sustainable retirement savings withdrawals. journal of financial planning, 30, 45-55. zaman, a. a. (2008). reconsidering ira and roth ira keeping bequests and other options in mind. journal of wealth management, 11, 82-92. zhou, j. (2009). the asset location puzzle: taxes matter. journal of economic dynamics and control, 33, 955-969. zhou, j. (2012). life-cycle stock market participation in taxable and tax-deferred accounts. journal of economic dynamics and control, 36, 1814-1829. j. a. dilellio and a. simon / financial services review 30 (2022) 223–249 249 finser_23_3 art 4 investor profiles: meaningful differences in women’s use of investment advice? kathryn simms, ph.d., cpaa,* aresearch faculty, college of health sciences, room 3121, old dominion university, norfolk, va 23508, usa abstract women in the united states face numerous financial challenges: they typically earn less than men do; they have greater probabilities of living in poverty; and they need substantial retirement funds, given their average longevity. consequently, a comprehensive understanding of how women use investment advice to remedy these challenges is vital. however, the literature is largely mute on this issue. this study helps to fill this gap in the literature by evaluating two profiles of female investors through cluster analysis and logistic regression conducted on a large, nationally representative database collected recently. predictors of seeking investment advice vary considerably across profiles. © 2014 academy of financial services. all rights reserved. jel classification: j16, economics of gender keywords: economics of gender; female investors; investment advice; financial services; financial literacy 1. introduction in the united states, women outnumber men as a function of increasing age to the extent that the ratio of women-to-men is 2-to-1 by the age of 85 (united states census bureau, 2010). consequently, a general absence of retirement planning among women (lusardi & mitchell, 2008) seems particularly problematic. aside from such age-related challenges, women are more likely to experience several other life difficulties associated with financial complications: about 10 million single women are raising at least one child, which repre* corresponding author. tel.: !1-757-683-7130. e-mail address: ksimms@odu.edu (k. simms) financial services review 23 (2014) 273–286 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. sents a 52% increase since 1970 (united states census bureau, 2011). additionally, more women than men live in poverty (united states census bureau, 2012). overall, women earn about 81 cents for each dollar that men earn (united states department of labor, 2013b). overcoming such financial adversities may be particularly challenging for women to do alone because their financial literacy is generally lower than that of men (lusardi & mitchell, 2011). these factors combined suggest that women may stand to benefit considerably from high quality investment advice. such advice may help to offset shortcomings in women’s financial knowledge and the financial hardships that tend to be more specific to their gender. ideally, the associated research literature would be quite advanced and ready to facilitate investment advisors and policymakers in accommodating women’s specific needs for investment advice services. however, relatively little is known about patterns of utilizing financial services independent of gender, much less about women’s specific usage patterns (collins, 2012; robb, babiarz, & woodyard, 2012). the upshot is that there is a gap in the literature in terms of assimilating a comprehensive picture of how women’s investor characteristics and other demographic traits combine to help predict the conditions under which women are the most likely to reach out for outside investment advice. the purpose of this study, then, is to help advance the extant literature by attempting to integrate these characteristics into more comprehensive profiles of female investors. the end-goal of such analysis is to identify more comprehensive patterns associated with how and why women use (or do not use) investment advice. these goals are attained through cluster analysis and logistic regression analysis of data derived from a nationally representative database, the 2012 national financial capability study (nfcs). 2. perspectives a common perspective on why women may have a need for professional investment advice is that, on average, they have lower financial literacy than men do. although there are differences across countries, this phenomenon is not unique to the united states (lusardi & mitchell, 2011). rather, it is documented both in undeveloped and developed countries alike including germany, netherlands, and japan, independent of demographic characteristics such as age. in fact, this gap between male and female financial literacy appears to exist at relatively high levels of educational attainment. for example, chen and volpe (2002) detect strong evidence of this gap among a sample of 924 undergraduate and graduate students who represented diverse majors and who were drawn from multiple institutions of higher education. on chen and volpe’s 36-question assessment, women outscored men on only one question, and underperformed men on 22 other questions—with the differential being more than 10% on 10 of those questions. the female students also did not rank personal finances to be as important as male students did, nor were the female students as confident about their financial knowledge. at even higher education levels, this gender gap may disappear, as suggested by chalmers and reuter’s (2010) analysis of the utilization of investment advice services provided through the oregon university system’s retirement fund. their analysis detected little evidence of a gender gap in financial knowledge. similarly, dolan and stevens (2013) report that having a college 274 k. simms / financial services review 23 (2014) 273–286 education seems to eliminate gender gaps that favor men in terms of the rationality financial decision-making. from another perspective, gender-based differences in financial literacy and associated measures seem unlikely to be a sufficient explanation for why women use outside investment advice. robb et al.’s analysis (2012) showed that women’s gender was a significant predictor of the use of investment advice after controlling not only for financial literacy, but also for self-perceived financial knowledge (or financial confidence), financial satisfaction, risk aversion, and a variety of other demographic variables including age and educational attainment. in fact, after controlling for these factors, their logistic regression analysis indicated that women have 34% higher odds of seeking investment advice than men do. collins (2012) reported a similar finding. gender was a tangent rather than a focal point of these important studies; therefore, these researchers had no particular reason to expound upon this finding. haslem’s discussion (haslem, 2008) of why some investors pay investment advisors to purchase mutual funds when no-load options are readily available may provide some insight: women may be seeking validation of their investment choices, attempting to resolve marital disputes over investment decisions, or benefiting from ancillary financial services such as an evaluation of their overall financial positions. investment advice may benefit women in several additional ways such as better asset allocation strategies. such strategies are particular important—given that women seem to have higher relative risk aversion (bajtelsmit, bernasek, & jianakoplos, 1999), even though they also seem to invest more in retirement funds all else equal relative to male investors (deaves, veit, bhandari, & cheney, 2007). furthermore, women are likely to benefit from advice that promotes proactive financial behaviors such as disciplined spending habits, paying bills on a timely basis, and being financially prepared (schmeiser & hogarth 2013). the remainder of this study attempts to contribute to the existing literature by evaluating whether different profiles of female investors exist and whether these profiles (if any) exhibit different patterns of utilizing investment services. 3. method 3.1. participants and data participants in this study are the 13,117 female participants in the state-by-state 2012 nfcs. this portion of the nfcs is a nationally representative database derived from online surveys conducted from july to october 2012. participants were selected based on a nonprobability quota sample from over a million, paid participants in preexisting online surveys. participants’ identities and self-reported demographics were verified. the finra investor education foundation funded the nfcs. it developed the nfcs in conjunction with organizations such as the u.s. department of the treasury and president obama’s advisory council on financial capability. the main objectives of these organizations were (1) to collect key benchmarks for assessing financial capability in the united 275k. simms / financial services review 23 (2014) 273–286 states and (2) to examine variability in these benchmarks. this current study relies on data from the public-use version of the database. 3.2. measures 3.2.1. use of investment advice and evidence of investing/saving use of investment advice is an indicator variable that provides self-reported data about whether participants asked for advice about investing or savings during the last five years. having an emergency, “rainy day” fund to cover at least three months of expenses in the event of illness, unemployment, or economic downturns serves as evidence of actual investing or saving behaviors. in many respects, this measure is ideal because other measures may have less malleable origins. for example, having sufficient retirement savings is, at least partially, a function of whether participants’ employers offer retirement plans and of the nature of these plans (i.e., defined benefit or defined contribution). 3.2.2. investor characteristics each participant’s financial literacy score is the number of correct responses to five questions about investing, borrowing, and personal financial management included the nfcs. this and similar measures have been well accepted in the literature (e.g., collins, 2012; lusardi, mitchell, & curto, 2010; lusardi & mitchell, 2011). self-perceived overall financial knowledge varies from 1 " very low to 7 " very high. additionally, participants compare their financial knowledge to that of other household members as follows: 1 " the participant knows the most, 2 " someone else knows the most, 3 " the participant and someone else in the household are about equally knowledgeable, 4 " the participant does not know, 5 " the participant prefers not to say, and 6 " there is no other person with whom to compare the participant’s knowledge in the household. participants’ self-reported willingness to take risks when making financial investments ranges from 1 to 10, where 1 indicates the lowest level of self-perceived risk-taking and 10 indicates the highest level of such behavior. current financial satisfaction is measured from 0 " not at all satisfied to 10 " extremely satisfied. approximate annual household incomes range from 1 " less than $15,000 to 8 " $150,000 or more. employment status consists of being self-employed, employed full-time, or employed part-time; or being a homemaker, a full-time student, disabled, unemployed, or retired. partners’ employment status has the same response categories, except that an additional category accounts for participants who have no “significant others.” 3.2.3. demographic variables demographic variables include number of financially dependent children, marital status, educational attainment, age, and race. number of financially dependent children is measured as 1 " 1 to 4 " 4 or more. marital status includes being married, single, separated, divorced, or widowed. educational attainment ranges from 1 " did not complete high school to 5 " completed graduate education. participants’ ages vary from 1 " 18–24 to 6 " over 65. information about race is restricted to whether or not participants are white in the public-use database. 276 k. simms / financial services review 23 (2014) 273–286 3.3. data analysis first, this study uses descriptive statistics to evaluate basic information about the data. next, it relies on cluster analysis (jain, murty, & flynn, 1999; norušis, 2011) to access the existence of female investor “profiles” based on the investor characteristics defined in the prior section. more specifically, the goal of cluster analysis is to determine the number of profiles of female investors (if any) and to assess individual group membership. this study relies on two-step cluster analysis, a method appropriate when theory does not suggest the number of clusters and when sample sizes are large. two-step clustering derives the number of clusters through iterative log likelihood estimation based on a goodness of fit measure (i.e., the bayesian information criterion [bic]). next, profiles are identified via (1) the importance statistics for each investor characteristic, (2) the rank of each characteristic within each cluster, and (3) distributional characteristics of each measure (i.e., frequencies and means) within each cluster. the importance statistic for investor characteristics assesses a variable’s role in assigning participants to a particular cluster. it ranges from 1 (i.e., highly important in distinguishing clusters) to 0 (i.e., not important in distinguishing clusters). demographic variables are entered as evaluation fields, so that these variables do not play a role in classifying investor profiles. hence, the importance statistic for each evaluation field permits an assessment of the connection between each demographic variable and each cluster. finally, logistic regression is used to study the explanatory power of variables related to women’s decisions to seek investment advice. 4. results 4.1. descriptive data analysis only about 27%1 (n " 3,565) of female participants report having used investment advice in the last five years; about 70% (n " 9,227) report that they did not use such services, with a small fraction (2%, n " 325) reporting that either they do not know whether they have used such services or that they prefer not to say whether they have done so (table 1). over a third (37% or n " 4,829) have a rainy day fund that would cover at least three months of expenses; 59% (n " 7,708) do not have such a fund, and 4% (n " 580) either do not know or prefer not to say whether they have a rainy day fund. on average, participants answered fewer than three questions correctly on the financial literacy assessment for an average percentage score of 52%. at the same time, participants’ mean self-perceived financial knowledge is 5 on a 7-point likert-type scale, and 27% believe that their knowledge is equivalent to someone else’s in the household. participants’ mean level of financial satisfaction is about 5 on a 10-point likert-type scale, and their mean level of self-perceived risk-taking is about 4 out of 10 possible levels. table 2 provides additional demographic statistics about the sample. 4.2. cluster analysis two-step cluster analysis indicates the presence of two distinct profiles of female investors (table 3). profile 1 (n " 7,658) can be characterized as females who typically are not 277k. simms / financial services review 23 (2014) 273–286 prepared for financial emergencies (i.e., 89% did not have a rainy day fund), are not satisfied financially (m " 3 on a 10-point scale), and have limited household incomes (m " $25,000 per year to under $35,000 per year). these women are also unlikely to seek out investment advice (i.e., 11% have sought out such advice during the last five years). furthermore, their self-perceived financial knowledge (m " 5 on a 7-point scale) tends to exceed their financial literacy scores (m " 2 or 40% of questions answered correctly). profile 2 (n " 4,975) is more likely to ask for investment advice (m " 53%), is more likely to have a rainy day fund (m " 78%), and tends to be fairly financially satisfied (m " 7 on a 10-point scale) with a mean annual household income of $75,000 to under $100,000. additionally, participants’ self-perceived financial knowledge (m " 6 on a 7-point scale) also appears to exceed their financial literacy scores (m " 3 or 60% of questions answered correctly). their mean financial literacy score exceeds the mean score for the full sample (p # 0.001). although working full-time is the most frequent work status for both profiles, 23% of women2 in profile 1 work full-time compared with 37% of women in profile 2. selfperceived investment risk-taking is the most important factor in distinguishing between clusters. however, it is also among the lowest factors in distinguishing between participants within each cluster (i.e., eighth out of nine categories). about 75% of participants in profile 1 rank their self-perceived level of risk-taking as being 5 or lower, whereas self-perceived table 1 investor characteristics for female participants (n " 13,117) in the 2012 national financial capability study measure mean (m) and sd or % n or n use of financial advice about investing/savings yes 27.2% 3,565 no 70.3% 9,227 do not know 1.4% 189 prefer not to say 1.0% 136 rainy day fund yes 36.8% 4,829 no 58.8% 7,708 do not know 2.8% 370 prefer not to say 1.6% 210 financial literacy score (0 to 5 correct) m " 2.6 13,117 sd " 1.4 self-perceived risk-taking (1 low to 10 high) m " 4.3 13,117 sd " 2.5 perception of financial knowledge (1 low to 7 high) m " 5.0 12,633 sd " 1.3 comparison of financial knowledge participant knows the most in household 24.1 3,160 someone else knows the most 9.0 1,179 equally knowledgeable 26.7 3,507 does not know 2.5 332 prefers not to say .6 84 no comparison 37 4,855 financial satisfaction (0 low to 10 high) m " 4.8 13,117 sd " 2.9 278 k. simms / financial services review 23 (2014) 273–286 risk-taking is distributed relatively evenly throughout the 10-response categories for participants in profile 2. the only investor characteristic that has an importance statistic of less than 1 is how participants rank their financial knowledge compared with that of another member of the household (importance statistic " 0.75). this characteristic also ranks last within each cluster. participants in profile 1 typically do not have a partner with whom to compare their financial knowledge (frequency " 47%), whereas participants in profile 2 are more likely to rank their knowledge as being equal to someone else’s in the household (frequency " 37%). in terms of evaluation fields for the demographic variables (table 4), 16% of participants in profile 1 have completed either their undergraduate or graduate degrees; by comparison, 38% of participants in profile 2 have completed at least one of these degrees. the importance statistic is 1 for this evaluation field. marital status has an importance statistic of 0.77. frequency analysis indicates that 43% of participants in profile 1 are married compared with 71% of participants in profile 2—with “married” being the most frequent marital status for each profile. profile 1 also tends to be younger (i.e., the mean age range for profile 1 is 35–44 compared with 45–54 for profile 2, with age having an importance statistic of 0.71). the importance statistics for the remaining evaluation fields—number of the children and race— are 0.11 and 0.07, respectively. over 50% of participants in each profile report having no financially dependent children, and over 60% of participants in each profile are white. table 2 descriptive statistics for female participants (n " 13,117) in the 2012 national financial capability study measure mean (m) and sd or % n or n number of financially dependent children m " 0.82 13,117 (1 to 4 or more) sd " 1.1 marital status married 54.1% 7,098 single 25.6% 3,361 separated 2.0% 265 divorced 12.7% 1,663 widowed 5.6% 730 educational attainment did not complete high school 9.8% 1,285 completed high school only 32.02% 4,200 some college 34.02% 4,484 undergraduate degree 15.5% 2,035 graduate education 8.66% 1,113 age 18–24 12.3% 1,616 25–34 17.0% 2,233 35–44 16.5% 2,167 45–54 18.8% 2,468 55–64 19.9% 2,487 65! 15.5% 2,146 race white 64.2% 8,425 non-white 35.8% 4,692 279k. simms / financial services review 23 (2014) 273–286 4.3. logistic regression the results of logistic regressions in table 5 provide the odds that women in each profile will seek investment advice for each relevant investor characteristic and demographic variable. to enhance the reader’s interpretations, i summarize results that differ across profiles first, and follow that discussion with a description of the relationships that are more similar across profiles. for profile 1, how the participant ranks her financial knowledge in comparison to someone else’s in the household does not predict the odds of seeking investment advice (panel a). however, for profile 2, as long as the participant assesses the table 3 results of cluster analysis reported in the order of importance statistics for distinguishing between clusters measure profile 1 n " 7,658 profile 2 n " 4,975 self-perceived investment risk importance statistics 1 1 within cluster rank 8 8 mean 4 out of 10 5 out of 10 asked for investment advice in the last five years importance statistic 1 1 within cluster rank 4 1 mean .11 .53 rainy day fund importance statistic 1 1 within cluster rank 1 2 mean .11 .78 overall financial satisfaction importance statistic 1 1 within cluster rank 2 3 mean 3 out of 10 7 out of 10 perception of financial knowledge importance statistic 1 1 within cluster rank 5 4 mean 5 out of 7 6 out of 7 financial literacy score importance statistic 1 1 within cluster rank 6 6 mean 2 out of 5 3 out of 5 household income importance statistic 1 1 within cluster rank 3 5 mean $25,000 to under $35,000 $75,000 to under $100,000 current work status importance statistic 1 1 within cluster rank 7 7 mode full-time (23%) full-time (37%) comparison of financial knowledge importance statistic .75 .75 within cluster rank 9 9 mode no partner (47%) equal (37%) note: modes are provided as the best descriptor of central tendency for nominal variables with more than two categories. 280 k. simms / financial services review 23 (2014) 273–286 differential in knowledge, the odds of seeking investment advice are consistently higher compared with the reference category (i.e., the participant believes that someone else knows more than she does). more specifically, when the participant believes that she knows more than someone else in the household, the odds are 28% higher that the participant sought investment advice; if the participant believes that she is equally as knowledgeable as someone else in the household, the odds are 41% higher; and, if there is no one else in the household with whom to compare her knowledge, then odds of seeking investment advice are 81% higher. for profile 2, a one-unit increase in perceptions of financial knowledge is associated with 10% reduced odds of seeking investment advice (panel b). this predictor is not significant for profile 1. for profile 1, a one-unit increase in the participant’s financial literacy score is associated with a 13% increase in the odds of seeking advice, whereas this predictor is not significant for profile 2. for profile 1, earning more than $15,000 per year is associated with 49% to 84% greater odds of seeking financial advice up to the point that participants earn under $75,000 (panel c). in contrast, for profile 2, three household income ranges are associated with from 46% to 60% reduced odds of seeking investment advice compared with annual household income of less than $15,000 per year: earning $15,000 to less than $25,000; earning $25,000 to less than $35,000; and earning $75,000 to less than $100,000. for profile 1, being a full-time student is associated with 41% greater odds of seeking investment advice compared with working full-time, whereas being retired is associated with 41% lower odds of seeking investment advice compared with working full-time (panel d). for profile 2, being a homemaker, unemployed, and retired are associated with 36%, 120%, and 42% higher odds, respectively, of seeking investment advice compared with working full-time. for profile 1, having one additional child is associated with 9% higher odds of seeking advice (panel e), and being single compared with being married is associated with 34% greater odds of seeking advice (panel f). neither the number of children nor any aspect table 4 evaluation field for cluster analysis measure profile 1 n " 7,658 profile 2 n " 4,975 education importance 1 1 mean high school some college marital status importance .77 .77 mode married (43%) married (71%) age importance 0.71 .71 mode 35–44 45–54 children importance .11 .11 mean .90 .69 race importance .07 .07 mean .61 .69 note: modes are provided as the best descriptor of central tendency for nominal variables with more than two categories. 281k. simms / financial services review 23 (2014) 273–286 table 5 logistic regression: predictors of financial advice cluster 1 odds ratio cluster 2 odds ratio panel a: comparison of financial knowledge participant knows the most 1.024 1.284* someone else knows the most — — equally knowledgeable 1.225 1.411** does not know .655 .642 prefers not to say .666 1.016 no comparison 1.006 1.814** panel b: risk, satisfaction, knowledge, and financial literacy self-perceived risk-taking 1.081*** 1.144*** overall financial satisfaction .967* .949** perception of financial knowledge .962 .899** financial literacy score 1.126*** 1.039 panel c: annual household income less than $15,000 — — $15,000 to less than $25,000 1.523** .426** $25,000 to less than $35,000 1.840*** .402** $35,000 to less than $50,000 1.488** .633 $50,000 to less than $75,000 1.753*** .660 $75,000 to less than $100,000 1.141 .539** $100,000 to less than $150,000 1.007 .635 $150,000 or more .639 .811 panel d: employment status self-employed .975 1.274 employed full-time — — employed part-time 1.326* 1.499*** homemaker .771 1.361** full-time student 1.414* .763 disabled .851 1.445 unemployed 1.001 2.201*** retired .590** 1.416** panel e: children number of financially dependent children 1.093* 1.045 panel f: marital status married — — single 1.337* 1.131 separated 1.293 .635 divorced 1.038 1.001 widowed 1.476 1.013 panel g: age 18 to 24 — — 25 to 34 .681** .672* 35 to 44 .517*** .497*** 45 to 54 .691* .718* 55 to 64 .894 1.100 65! 1.155 1.274 panel h: race white .828* 1.255** non-white — — panel i: educational attainment did not complete high school .380*** .527** high school .529*** .602*** some college .111 .085 .796* .768** undergraduate degree — — graduate degree .983 1.142 constant .166*** 1.398 note: results for spouses’ work status were not significant and have been omitted to reduce table length. *p $ 0.05; **p $ 0.01; ***p $ 0.001. 282 k. simms / financial services review 23 (2014) 273–286 of marital status is a significant predictor for profile 2. for profile 1, being white is associated with 17% reduced odds of seeking advice, but for profile 2, being white is associated with 26% increased odds of seeking advice (panel h). a one unit-increase in self-perceived risk-taking is associated with 8% and 14% higher odds of seeking investment advice for profiles 1 and 2, respectively (panel b). a one-unit increase in being financially satisfied decreases the odds of seeking investment advice by 3% and 5% for profiles 1 and 2, respectively (panel b). being employed part-time compared with working full-time is associated with 33% and 50% greater odds of seeking investment advice for profiles 1 and 2, respectively (panel d). for both groups being older than 18 to 24 is associated with 28% to 50% reduced odds of seeking financial advice until participants surpass the age range of 45–54 (panel g). not having obtained an undergraduate degree is associated with from 20% to 62% reduced odds of seeking advice for both profiles, depending on participants’ final levels of educational attainment (panel i). finally, partner’s work status is not a significant predictor of seeking investment advice for either profile. 5. discussion and conclusions despite women’s risk of having insufficient retirement assets and of living at or below poverty (lusardi & mitchell, 2008; united states census bureau, 2012), only 27% of the women in this nationally representative study have sought investment advice during the last five years. additionally, only a little more than a third have a rainy day fund sufficient to cover at least three months of living expenses. this statistic is particularly disconcerting, given that the u.s. department of labor (2013a) reports that the duration of unemployment is 15 weeks or more for 53% of those who have been unemployed recently. however, these statistics also appear to conceal an important part of the picture. more specifically, the findings suggest that not only are there two groups, or profiles, of female investors, but also that the factors associated with the odds of seeking investment advice often vary across these profiles. put most simply, one profile of female investors might be labeled as “struggling” and the other profile might be identified as “thriving.” the chief factor that distinguishes these two profiles is that the “strugglers” tend to have higher self-assessed risk aversion. other discriminating factors yield the following more detailed profiles: only about 90% of the “strugglers” have a rainy day fund or have sought investment advice in the last five years. they also have more limited household incomes, averaging from $25,000 to under $35,000. over three-quarters of the “thrivers” have a rainy fund and over half have sought financial advice during the last five years. “thrivers” have higher household incomes, averaging from $75,000 to under $100,000 per year. the “thrivers” are more likely to have earned a graduate or undergraduate college degree, but more than 60% of them have not done so. the most common marital status for each group is being married, but 71% of “thrivers” are married compared with 43% of “strugglers.” despite conventional wisdom, race and number of financially dependent children do not particularly characterize these profiles. several different patterns in the factors associated with seeking investment advice emerge for these profiles. the “thrivers” appear to seek investment advice based on their perceived 283k. simms / financial services review 23 (2014) 273–286 financial knowledge, whereas the odds that strugglers will seek advice may be more connected to their actual financial knowledge. more specifically, when “thrivers” believe that there is someone else in the household who knows more than they do about finances, they are less likely to reach out for investment advice. at the same time, being more confident about own their financial knowledge is also associated with reduced odds that “thrivers” will seek outside help. on the other hand, it is more likely that “strugglers” will reach out for investment advice when their objectively measured financial knowledge is higher. additionally, some of the evidence suggests that “strugglers” are more likely to reach out for advice when they earn more (i.e., up to $75,000), so that the price of the advice may be a factor in their decision-making. on the other hand, “thrivers” are more likely to seek advice when they are living below poverty compared with when their earnings fall into one of several higher income brackets. some evidence also aligns with the theory that when “strugglers” work status is probably transitory—such as being a full-time student and working part-time—they are more likely to reach out for investment advice. by contrast, if their work status might be more permanent, such as being retired, they have reduced odds of seeking such advice. on the other hand, some of the evidence suggests that “thrivers” may react self-protectively by seeking investment advice when they work less than full-time, regardless of whether or not this work status is usually considered to be permanent. more specifically, they reach out for advice when they work part-time, are homemakers, are unemployed, or are retired. these findings require further corroboration, particularly in terms of validating the existence of these profiles in other datasets. however, these findings do suggest that— although it may often seem practical and suitable to talk about female investors as a single group—such a classification strategy can obscure the existence of meaningful differences in investor profiles within gender. not accounting for these differences may result in failures to launch investment advice strategies successfully. encouraging one group (i.e., the “strugglers”) to seek investment advice may be more about improving financial education, overcoming concerns about the costs of investment advice, and, potentially, mitigating fixedmindsets (dweck, 2007). encouraging the other group (i.e., “the thrivers”) to seek investment advice may be more about educating investors about the extent of financial advisors’ knowledge compared with their own and about capitalizing on what may be more of “growth” mindsets (dweck, 2007). both groups, though, are certainly urged to seek out financial advice—given women’s overall potentially more precarious financial situations— and to assess the quality of their financial advisors through checking references, determining the veracity of advertised certifications, evaluating whether conflicts of interest might exist, reading financial information in detail, and considering getting second opinions. findings in this study also expound on the extant literature. for example, haslem’s (2008) suggestion that women may be seeking validation of their investment choices when they seek investment advice appears to be on target, but only for women with characteristics similar to the “thrivers.” his suggestion that seeking advice may be an attempt to resolve a marital dispute over an investment decision appears less likely. “thrivers” are likely to seek investment advice even when they do not have a partner within whom to compare their 284 k. simms / financial services review 23 (2014) 273–286 knowledge. they seem willing to accept another member of the household’s authority, if they perceive that that person has more knowledge about finances than they do. additionally, this study suggests a nuance in how financial literacy predicts women’s odds of seeking financial services. nongender focused research suggests that both objective and subjective financial literacy are associated with the odds of using investment advice (collins, 2012; robb et al., 2012). findings in this study suggest that only objectively measured financial literacy is relevant for the profile that uses less financial advice. the subjective measures of financial literacy appear to be relevant for the profile that uses more of these services. however, this second group does tend to have higher financial literacy. further research is needed to overcome the limitations of this study. for example, this research is correlational, rather than causal. hence, for example, we cannot be sure that investment advice leads to greater household incomes and greater financial satisfaction. instead, such advice-seeking behavior may merely co-occur with household income without making a causal contribution. although causal connections are always difficult to establish, longitudinal data may be useful in studying these connections. additionally, the “price” of having a large, nationally representative dataset is usually limited access to details about the nuances of each variable. having additional data about the availability, type, source, price, frequency, and specific outcomes associated with investment advice may illuminate the investor profiles developed in this study further. in the future, either smaller scale studies or other large datasets collected primarily to study patterns in investment advice may provide additional useful insights. notes 1 supplemental analysis indicates that 30% of male participants had used these services, which is significantly higher than women’s usage via a one-tailed t test (p $. 001). 2 including partner’s work status in the analysis produces clusters that are not cohesive or distinct. hence, partner’s work status is excluded from the final analysis. references bajtelsmit, v. l., bernasek, a., & jianakoplos, n. a. (1999). gender differences in defined contribution pension decisions. financial services review, 8, 1–10. chalmers, j., & reuter, j. (2010). what is the impact of financial advisors on retirement portfolio choices and outcomes? the national bureau of economic research (nber working paper no. 18158). (available at http://www.nber.org/programs/ag/rrc/nb10–05%20chalmers,%20reuter%20final-revised.pdf). chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. collins, m. j. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–320. deaves, r., veit, e. t., bhandari, g., & cheney, j.(2007). the savings and investment decisions of planners: a cross-sectional study of college employees. financial services review, 16, 117–133. dolan, r. c., & stevens, j. l.(2013). rationality, irrationality, and “predictable irrationality”: does education, curriculum, or gender matter? financial services review, 22, 115–132. dweck, c. (2007). mindset: the new psychology of success. new york, ny: ballantine books. 285k. simms / financial services review 23 (2014) 273–286 haslem, j. a. (2008). why do mutual fund investors employ financial advisors? journal of investing, 17, 91–94. jain, a. k., murty, m. n., & flynn, p. j. (1999). data clustering: a review. acm computing surveys, 31, 265–323. lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review: papers & proceedings, 98, 413–417. lusardi, a., & mitchell, o. s. (2011). financial literacy around the world: an overview. journal of pension economics and finance, 10, 497–508. lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young. the journal of consumer affairs, 44, 358–380. norušis, m. j. (2011). ibm spss statistics 19 statistical procedures companion. upper saddle river: nj: prentice hall. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. schmeiser, m., & hogarth, j. (2013). good advice, good outcomes? how financial advice-seeking relates to self-perceived financial well-being. working paper series. social science research network. (available at http://ssrn.com/abstract"2261707). united states department of labor. (2013). economic news release: table a-12. unemployed persons by duration of unemployment. (available at http://www.bls.gov/news.release/empsit.t12.htm). united states department of labor. (2013). equal pay. (available at http://www.dol.gov/equalpay/). united states census bureau. (2010). dp-1: profile of general population and housing characteristics: 2010. (available at http://factfinder2.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src"bkmk). united states census bureau. (2011). profile american: facts for features. (available at http://www.census.gov/ newsroom/releases/archives/facts_for_features_special_editions/cb12-ff08.html). united states census bureau. (2012). current population survey (pov01): age and sex of all people, family members and unrelated individuals iterated by income-to-poverty ratio and race: 2012 below 100% of poverty—all races. (available at http://www.census.gov/hhes/www/cpstables/032013/pov/pov01_100.htm). 286 k. simms / financial services review 23 (2014) 273–286 behavioral and wealth considerations for seeking professional financial planning help jodi letkiewicz, ph.d.a,*, chris robinson, ph.d., cfp�a, dale domian, ph.d., cfp�a aschool of administrative studies, york university, toronto, ontario, canada m3j 1p3 abstract this study uses a canadian survey to examine the decisions to seek professional financial planning help. we find that people who use a financial planner have more wealth, lower subjective financial stress, and higher financial self-efficacy than people who do not use a financial planner. using the longitudinal design of the survey we find that people with higher self-efficacy in period t-1 are more likely to seek help in period t, leading to the conclusion that high self-efficacy drives one to seek financial planning help. we do not find that subjective financial stress leads one to seek financial planning help. implications for practitioners, consumers, and policy makers are discussed. © 2016 academy of financial services. all rights reserved. jel classification: d14 (household saving; personal finance); g20 (financial institutions and services: general); p46 (consumer economics) keywords: financial planning; help-seeking; financial stress; self-efficacy 1. introduction whether a financial planner helps increase well-being has become particularly relevant in recent years given the increased responsibilities shouldered by consumers and the complexities of the financial marketplace. we use the transtheoretical model of behavioral change (prochaska, 1984) and theories of stress and coping (lazarus, 1991) to form hypotheses relating self-efficacy, stress, and wealth with a consumer’s decision to seek professional * corresponding author. tel.: �1-416-736-2100; ext. 33630. e-mail address: jodilet@yorku.ca (j. letkiewicz) financial services review 25 (2016) 105–126 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. financial help. a critical concept in our research is self-efficacy, which refers to people’s beliefs in their capabilities to produce a desired outcome (bandura, 1997). we use a large unique survey with longitudinal data across three waves to assess whether self-efficacy and stress are correlated with having a financial planner and if high stress and high self-efficacy lead one to seek a financial planner. we find that while subjective stress is negatively correlated with use of a financial planner, high subjective stress by itself is not significantly related to the decision to seek professional financial help. financial self-efficacy is correlated with both the use of a financial planner as well as the decision to seek help. 2. background and review of literature professional financial planning is an important option for overall financial well-being. financial planners have tools and expertise lacking in the general population and can help families facing complex circumstances decide which financial decisions are in their best interest. collins (2010) distinguishes four different roles that financial planners can play: technical expert, transactional agent, counselor, and coach. given the wide breadth of these roles, the outcome measures from the use of a financial planner can vary substantially. given the variability of advice, zick and mayer (2013) question whether the appropriate outcome measures are best measured in dollars, time, or psychological states. existing studies have quantified financial planner advice in monetary terms (hanna, 2011), while others measure it by evaluating financial milestones such as establishment of long term goals and a retirement plan (marsden, zick, and mayer, 2011). blanchett and kaplan (2013) show how financial planning techniques can increase certainty-equivalent income in retirement. kinniry, jaconetti, dijoseph, and zilbering (2014) show that financial planning strategies could add 3% per year in net returns. while taylor, jenkins, and sacker (2011) do not explicitly consider financial planners, they show that making good financial decisions improves people’s psychological health. 2.1. help-seeking behavior the decision to seek help is relevant across multiple domains, including medicine, mental health, and personal finance. much research focuses on whether people faced with financial or medical problems decide to seek help for those problems. research in health fields has been more prolific than research in the financial field, but financial planning is gaining traction. some of the first researchers to consider help-seeking in a financial context were grable and joo (1999). grable and joo (1999) conceptualize help-seeking behavior as a coping strategy to deal with financial troubles. they develop a framework consisting of five stages including the recognition and evaluation of one’s own financial behaviors in the process of seeking professional help when people recognized problems with which they needed help. grable and joo (2001) examine the choice between obtaining financial advice from professionals versus nonprofessionals. they find that individuals with low financial risk tolerance and low satisfaction with their financial situation are more likely to seek advice 106 j. letkiewicz et al. / financial services review 25 (2016) 105–126 from family, friends, and work colleagues, instead of from professional sources. according to du plessis, lawton, and corney (2010), barriers to financial help-seeking include shame and embarrassment, as well as lack of knowledge about professional sources. dearden, goode, whitfield, and cox (2010) demonstrate gender differences in help-seeking for debt problems. hanna (2011) finds greater usage of financial planners among people with above average risk tolerance, post-bachelor degree education, higher household income, and higher net worth. hanna also finds an increase in the use of financial planners, from 21% of households in 1998 to 25% in 2007. collins (2012) finds that individuals with higher incomes, educational attainment, and levels of financial literacy are more likely to receive financial advice, and suggests that financial advice more often serves as a complement to, rather than a substitute for, financial capability. robb, babiarz, and woodyard (2012) find individual characteristics differentiate which type of financial advice people seek (e.g., debt counseling vs. investment planning). financial knowledge and satisfaction are positively related to using any type of financial advice while knowledge and satisfaction are inversely related to the use of debt counseling. cummings and james (2014) analyze decisions to either get or drop financial advisors. the most significant factors in getting a financial advisor include becoming a new widow(er), asking family members for assistance with financial decisions, seeking help for emotional problems, and positive changes in income and net worth. new widow(er)s and people with increased net worth are less likely to drop their financial advisors. finke, huston, and winchester (2011) find those who pay for financial advice are more likely to be female, relatively older, wealthier, and college educated with a low level of self-reported knowledge about financial issues. of those who pay for help, they find those who choose comprehensive management are more likely to be under 65, wealthy, and have high self-reported knowledge about financial issues. 2.2. financial stress and self-efficacy the cognitive theory of stress and coping defines stress as a certain situation that an individual assesses as taxing or exceeding his or her resources and consequently threatens his or her well-being (folkman, schaefer, and lazarus, 1979; lazarus and folkman, 1984). within a meta-theoretical system approach lazarus (1991) views the complex processes of emotions (i.e., stress) as composed of causal antecedents, mediating processes, and effects. antecedents are individual variables, such as commitments, beliefs, or environmental variables, such as demands or situational constraints. a mediating process is an appraisal of a situation and an assessment of personal coping options. effects can be immediate or long-term and may include areas such as psychological well-being, somatic health and social functioning. in this context, self-efficacy, along with other factors such as locus of control, anxiety or self-esteem, is considered a mediating process that helps an individual to manage situational requirements (bandura, 1992; jerusalem and schwarzer, 1992). given a stressful situation, such as financial distress, self-efficacy can act as a coping mechanism and thus help lessen the stressful situation. financial stress can arise from personal, family, or financial situations (joo, 1998) and the impact of financial stress can be far-reaching. wrosch, heckhausen, and lachman (2000) 107j. letkiewicz et al. / financial services review 25 (2016) 105–126 explore how financial stress affects perceptions of well-being. kim and garman (2003) show that employees with financial stress have greater absenteeism and were less committed to their organizations. britt, gable, goff, and white (2009) find that in couples, the partner’s spending behavior is a key factor influencing relationship satisfaction. self-efficacy refers to people’s beliefs in their capabilities to produce given attainment (bandura, 1997). self-efficacy is domain specific (bandura, 1997, 2006; lown, 2011), meaning that it is not universal across all aspects of one’s life. bandura (2006) differentiates between general self-efficacy and domain-specific self-efficacy. for example, the belief in one’s ability to compete in a triathlon does not necessarily mean the person believes they can manage money with the same sense of confidence. self-efficacy affects every area of human endeavor including health, academic performance, and personal finances (grembowski et al., 1993; lent, brown, and larkin, 1986; lapp, 2010). self-efficacy is particularly important in the context of financial decisions and helpseeking because it influences individuals’ behavioral changes (bandura, 1977; gecas, 1989). research in the health and exercise fields demonstrates that self-efficacy can be boosted to encourage health-promoting behaviors (grembowski et al., 1993). individuals with high levels of self-efficacy are more successful than those with low self-efficacy in coping with adversity (park and folkman, 1997). lapp (2010) finds higher financial self-efficacy is associated with lower debt, fewer financial problems, lower financial stress, higher savings, and greater financial happiness. in studies examining self-efficacy, risk tolerance, age, and education are positively correlated with self-efficacy (lown, 2011). engelberg (2007) finds that respondents with a high sense of self-efficacy are less likely to perceive themselves being at risk for disrupted income, unforeseen expenses, and unsuccessful investments, as compared with those with low self-efficacy. the study finds that those with high self-efficacy report a sense of financial control, less attachment to the importance of money, better economic knowledge, a more optimistic view of their financial situation and less distrust in money matters. lacking a sense of economic self-efficacy is associated with feelings of stress, negative emotion, and in more extreme cases, depression (e.g., burgogne, 1990; ennis, hobfoll, and schroeder, 2000; krause and baker, 1992; mates and allison, 1992). remund (2010) proposes that better consumer financial decision making stems from financial self-efficacy—a belief that one can effectively manage his/her personal financial affairs. along the same lines, lapp (2010) finds that financial self-efficacy is the missing link between knowledge and effective action and that given awareness or knowledge about a situation, self-efficacy will propel someone into action. because seeking the help of a financial planner is a positive consumer decision, then it is reasonable to expect that financial self-efficacy will be positively associated with the decision to seek financial planning help. those with a high sense of financial self-efficacy may believe they have the ability to handle their financial affairs and be able to identify what they can manage and when they need help. those low in financial self-efficacy may be less able to manage their financial affairs and therefore unable to determine when they need help. there is some evidence to suggest that this is the case. a study by parker et al. (2012) finds that confidence, a trait closely related to self-efficacy, is positively correlated with the likelihood of retirement planning and suggests that confidence may be needed to start the overwhelming 108 j. letkiewicz et al. / financial services review 25 (2016) 105–126 process or even to make an appointment with a financial advisor. lim et al. (2014) find that college students with high levels of financial stress are generally less likely to seek financial help, but that effect is somewhat moderated for those high in self-efficacy. 3. transtheoretical model of behavior change we bring together two strands of research—the transtheoretical model of behavior change (ttm; prochaska and velicer, 1997) and the cognitive theory of stress and coping (folkman et al., 1979). we combine these theories to create hypotheses of why some people seek financial planning assistance and others do not. the ttm assesses an individual’s willingness or ability to change a behavior and outlines processes to help guide individuals through the stages. the ttm emerged from a comparative analysis of leading theories of psychotherapy and behavior change in an effort to integrate a field that had fragmented into more than 300 theories of psychotherapy (prochaska, 1984). the ttm has been used in studies to explain how people stop unhealthy behaviors and develop healthy ones (xiao et al., 2004) and extended to explain and predict positive changes in financial behaviors based on participation in financial education programs (shockey and seiling, 2004). the ttm outlines five stages that individuals go through when changing behaviors: pre-contemplation (not ready), contemplation (getting ready), preparation (ready), action, and maintenance. to progress through those stages, individuals need awareness that the advantages of the change outweigh the disadvantages, confidence they can make and maintain the changes (self-efficacy), and strategies to help them maintain their new behaviors. we think of ttm as a process one progresses through when adopting good habits. in the process of making a decision there are antecedents, and in this case we hypothesize that stress is an antecedent and self-efficacy is the coping mechanism that propels an individual into action. if stress is experienced in the pre-contemplation and contemplation stage, then self-efficacy can help move people to the preparation and action phase. separating the stages of positive habit formation may help explain why people with low financial self-efficacy or high financial stress may resist getting financial planning help. there are many ways to change and improve behaviors; this study concentrates on the help-seeking aspect. seeking financial help is a positive decision that can help deal with financial issues. 4. hypotheses this study focuses on the first four stages of the ttm (pre-contemplation, contemplation, preparation, and action) to drive the hypotheses on help-seeking behavior. the dataset does not contain information that would allow us to test the maintenance phase, such as implementation of a financial plan or some other specific set of financial management behaviors. we propose three sets of testable hypotheses, two of them using the ttm and the third relating to the nature of the professional help that is the subject of the financial planning standards council (fpsc) survey, described in the next section. 109j. letkiewicz et al. / financial services review 25 (2016) 105–126 hanna (2011), hanna and lindamood (2011), and robb et al. (2012) find that families with higher income and/or wealth are much more likely to engage financial planners. these families receive more financial value from the advice, or they perceive that they will and they have the means to pay for the planning services. accordingly, we propose the first hypothesis: hypothesis 1: professional help-seeking is positively correlated with the level of assets, income and homeownership. we conceptualize financial stress as being a trigger in the pre-contemplation and contemplation stages of the ttm. as a way to manage that stress, individuals, or households will seek professional financial help. the experience of stress is likely to drive one to address it. given the research on stress and financial well-being, it is likely that people with planners have a lower level of stress. it is important to test for the correlation of stress and use of a planner as well as the help-seeking aspect. therefore, we propose: hypothesis 2a: financial stress is negatively correlated with having a financial planner. hypothesis 2b: financial stress is positively correlated with help-seeking. financial self-efficacy is conceptualized as a coping mechanism for stress, so we model it as a trigger that leads to the preparation stage; then the action stage of the ttm is when financial planning help is sought. we expect that self-efficacy drives one to seek help, but may also be positively correlated with having a financial planner. thus, we propose the third set of hypotheses: hypothesis 3a: self-efficacy is positively correlated with having a financial planner. hypothesis 3b: self-efficacy is positively correlated with help-seeking. 5. data 5.1. the fpsc survey the financial planning standards council (fpsc) conducted a survey on use of financial advisors, with particular attention to those identified as financial planners and certified financial planners among the general english-speaking population of canada. the survey excluded québec because that is the only province which regulates use of the term “financial planner.” the fpsc surveyed three waves of respondents: 1. wave 1: between august 2009 and january 2010, a total of 117,467 individuals were solicited with 7,383 surveys completed, resulting in a response rate of 6.2%. 2. wave 2: between february and july 2011, the population of respondents from wave 1 (7,383) was solicited again and 2,471 surveys were completed resulting in a response rate for wave 2 of 33%. 3. wave 3: the third wave, from april to august 2012, included the 7,383 panelists from wave 1 and 2 along with 88,247 new invites. a total of 1,045 from waves 1 and 2 and 7,510 new surveys were completed for a response rate of 8.9%. 110 j. letkiewicz et al. / financial services review 25 (2016) 105–126 in terms of response rates, a national survey of student engagement (nsse) study (fosnacht, sarraf, howe, and peck, 2013) concluded that even relatively low response rates provided reliable estimates and in other studies the total number of respondents has been shown to be more important in assuring reliable estimates than response rates (e.g., pike, 2012). we believe the final sample size used in this study is sufficient in assuring reliable estimates. there are two steps in the analysis and the samples used in each step are different. for the analysis in model 1, we exclude the panelists in wave 2, and panelists from wave 3 who are resurveys from wave 1, because these respondents do not constitute independent observations. after removing a few more cases because they are missing the respondent identification variable, we combine data from waves 1 (7,275) and 3 (7,502) to create a single dataset containing 14,777 observations. a total of 709 cases had one or more financial variables missing (primarily income, assets, and debt). we report results using the database without missing variables, or 14,068 observations. the panel data are used in the second step of the analysis. we only include data from respondents who completed the survey in all three waves (1,045). after accounting for missing data across all three waves, the final panel data sample size is reduced to 826. the survey targeted households likely to seek help from a financial advisor, with quota minimums based on the type of financial advice received (comprehensive/integrated planning, limited planning); credentials of the advisors (certified planners vs. noncertified advisors); and those using or not using an advisor. therefore, the sample is not nationally representative. while this is certainly a limitation of the dataset, we do not see any bias in it that invalidates its use to investigate the hypotheses we test. the financial planning standards council (2013) published some initial results of the survey in the value of financial planning. findings indicate that on average, people who seek professional financial-planning help will experience greater financial and emotional well-being. the survey offers much more scope than the simple descriptive statistics in the initial report, and our study utilizes the data along with a theoretical foundation to delve deeper into the question of what affects the decision to use a financial planner. 5.2. control variables the demographic variables include age (categorical variables), gender, marital status (married or not currently married, which includes single, widowed and separated), educational attainment (no college degree, college degree or some college, have a university degree), employment status (unemployed, employed, or retired), and whether or not they have children. it should be noted that in canada, “college” refers to institutions similar to community colleges, tech, or trade schools in the united states. these variables likely affect help seeking tendencies, but they are not our primary concern. for example, a household with children will need to determine life insurance needs or college savings, which may require outside assistance. we expect a positive sign for higher levels of education because it is also highly correlated with income and wealth, and a positive relationship for the age group 50–64 because that is the period in the life cycle when people are accumulating investment assets and need retirement planning. these variables must be included to remove their effect before we can determine the significance of our hypotheses. financial variables, which we 111j. letkiewicz et al. / financial services review 25 (2016) 105–126 will refer to collectively as wealth variables, include total assets, income levels (categorical variables), home ownership, and non-housing debt. a binary variable controls for the survey year (wave 1). this ensures that the results presented from model 1 are not time-dependent. the survey poses a series of questions on the respondent’s feelings or opinions about his or her own financial well-being. we use these questions to construct measures of financial stress and financial self-efficacy. finally, there are two variables that we believe are related to financial stress and are more objective than feelings about well-being: amount of nonhousing debt and whether the respondent believes he or she is likely to lose his or her job in the future. 5.3. measuring financial stress both subjective measures of stress, which indicate an internal emotion, and objective measures of stress, which are external or environmental factors, have been used by researchers to measure financial stress (britt et al., 2008; kim and garman, 2003; wrosch et al., 2000). the subjective measures in this study (listed below) are similar to questions used by other researchers. for example, kim and garman (2003) use two measures, “my income is enough for me to meet my monthly living expenses” (reverse coded) and “i worry about how much money i owe.” another study (wrosch et al., 2000) uses two subjective measures of stress, “i have enough money to meet my needs” and “i have difficulty paying monthly bills.” we use the following three questions to construct the measure of subjective financial stress: y i worry a lot about my financial situation y i feel i barely get by every month y my finances are out of control responses for all items were based on a 9-point likert scale (1 � strongly disagree to 9 � strongly agree). we use principal component analysis to create the subjective financial stress variable (stress, henceforth) with a mean of zero, and positive score indicating a relatively higher level of stress. the resulting construct has a cronbach’s � of 0.807, indicating good internal consistency. we distinguish between self-efficacy as the person’s belief that she or he can take action to improve the situation, and stress as the person’s belief or feeling that his or her situation is negatively impacting their well-being. stress has two inter-related components. the stress variable attempts to measure the person’s subjective feeling of stress. where financial matters are involved, a family is more likely to feel stressed when matters go badly or their outcome is uncertain. we identify two questions in the survey that measure objective financial stress: amount of non-housing debt and unemployment risk. joo (1998) creates a list of 24 objective financial stressors that are used in several studies. these stressors include events such as loss of a job, serious illness, divorce, wage garnishment, and ability to pay off debt. while the dataset used in this study is not robust enough to measure all 24 items, the two objective items we include (debt and employment risk) are both represented in joo’s original study (1998). our examination of the evidence on hypothesis 2a and hypothesis 2b as null hypotheses seeks to determine if we can reject the hypotheses that these variables are positively related 112 j. letkiewicz et al. / financial services review 25 (2016) 105–126 to seeking help or negatively related to the use of a planner. the strongest evidence will be rejection or failure to reject for all three variables, stress and the two objective measures. 5.4. measuring financial self-efficacy self-efficacy is a judgment of capability and a state of mind that we cannot measure directly. researchers measure it by constructing scales using survey questions. schwarzer and jerusalem’s (1995) 10-item general self-efficacy scale (gses) has been validated in 30 countries. the gses is a general measure that does not assess specific behavior, so consumer economics researchers develop measures that relate to personal financial behavior. bandura (2006) provides some guidance for the development of a self-efficacy scale. self-efficacy is an indication of perceived capability and, therefore, items measuring selfefficacy should be phrased in terms of “can do” rather than “will do.” multiple scales have been used to measure financial self-efficacy (e.g., dietz et al., 2003; danes and haberman, 2007; lapp, 2010; lown, 2011) and all propose somewhat different scales using somewhat different survey questions. five questions (listed below) were identified in the fpsc survey as measures of financial self-efficacy. a comparison of the fpsc measures and the measures of the four aforementioned studies are listed in table 1. all of their work and the measures in this study exhibit more similarities than differences. the survey questions of these four articles map reasonably well to questions on the fpsc survey. we place them on the same rows when the questions are similar, but the reader can see that all the scales are similar. based on the guidance of bandura (2006) and a comparison of the similar measures in table 1, the five questions used to construct the financial self-efficacy measure are: y i feel that i am prepared to manage through tough economic times y over the last 5 years, i have improved my ability to save y i don’t know what to do to improve my financial situation (reverse coded) y i feel prepared in the event of an unexpected financial emergency y i am on the right track in terms of financial affairs each of these statements is about a perceived capability. for example, reverse coding “i don’t know what to do to improve my financial situation” indicates that one knows how to improve their financial situation. responses for all items were based on a 9-point likert scale (1 � strongly disagree to 9 � strongly agree). we use principal component analysis to create a financial self-efficacy variable (se, henceforth) with a mean of zero, and positive score indicating a relatively higher level of self-efficacy. se has a cronbach’s � of 0.812, which indicates a good level of internal consistency. to determine the measures of stress and se, we started with 12 possible statements in the survey. we determined which statement measured each construct beginning with a theoretical basis and then conducted principal components analysis to confirm the theoretical basis. the stress and se measures loaded into their respective classifications, confirming the use of each question in their respective measures. 113j. letkiewicz et al. / financial services review 25 (2016) 105–126 t ab le 1 m ea su re s of fin an ci al se lf -e ffi ca cy d ie tz et al . (2 00 3) d an es an d h ab er m an (2 00 7) l ap p (2 01 0) l ow n (2 01 1) fp sc su rv ey qu es tio ns i ha ve lit tle co nt ro l ov er fin an ci al th in gs th at ha pp en to m e it is ha rd to st ic k to m y sp en di ng pl an w he n un ex pe ct ed ex pe ns es ar is e i fe el pr ep ar ed in th e ev en t of an un ex pe ct ed fin an ci al em er ge nc y t he re is lit tle i ca n do to ch an ge m an y of th e im po rt an t m on ey is su es in m y lif e i fe el co nfi de nt ab ou t m ak in g de ci si on s th at de al w ith m on ey i la ck co nfi de nc e in m y ab ili ty to m an ag e m y fin an ce s i do n’ t kn ow w ha t to do to im pr ov e m y fin an ci al si tu at io n i be lie ve th e w ay i m an ag e m y m on ey w ill af fe ct m y fu tu re i w as sa tis fie d w ith m y fin an ci al si tu at io n it is ch al le ng in g to m ak e pr og re ss to w ar d m y fin an ci al go al s i am on th e ri gh t tr ac k in te rm s of fin an ci al af fa ir s i w as ab le to sa ve m on ey w he n un ex pe ct ed ex pe ns es oc cu r i us ua lly ha ve to us e cr ed it o ve r th e la st 5 ye ar s, i ha ve im pr ov ed m y ab ili ty to sa ve i of te n fe el he lp le ss in de al in g w ith th e m on ey pr ob le m s of lif e w he n fa ce d w ith a fin an ci al ch al le ng e, i ha ve a ha rd tim e fig ur in g ou t a so lu tio n i fe el th at i am pr ep ar ed to m an ag e th ro ug h to ug h ec on om ic tim es i w as go od at pl an ni ng fo r m y fin an ci al fu tu re i w or ry ab ou t ru nn in g ou t of m on ey in re tir em en t 114 j. letkiewicz et al. / financial services review 25 (2016) 105–126 5.5. the dependent variable the survey question that generates the dependent variable, help-seeking, is: for which, if any, of the following services did you obtain help from professional financial advisor(s) (includes financial planner, life insurance advisor, investment advisor, debt counselor etc.) to assist you in the past five years? please select as many as apply. [the italicized text was underlined in the survey instrument.] the first answer choice is: i have not obtained help from a financial advisor for any of the following financial matters in the past 5 years. if the respondent chose the first answer, the help-seeking variable is coded 0. if the respondent chose instead one or more of the comprehensive list of financial advisory activities listed, the help-seeking variable is coded 1. the details of which particular services a respondent used are beyond the scope of our study. 6. method we use two models to test the hypotheses proposed. model 1 utilizes a logistic regression that has the general form: financial help-seeking � f�objective stress, stress, se, controls� the dependent variable is categorical: it takes the value one if help was sought in the last five years, or zero, if it was not. the relationship between the predictor and response variables is not a linear function; instead a logit transformation is used to arrive at the following relationship: logit�� � x�� � ln� � � x� 1 � � � x�� � � � �1x1 � �2x2 · · · � �kxk � ei (1) the full sample (n � 14,068) is used in this step of the analysis. to further probe the relationship between financial self-efficacy and use of a financial planner, we make use of the longitudinal aspect of the survey in model 2. the panelists in the sample we use completed all three waves of the survey (n � 1,045). cases missing key variables or with inconsistent responses for the financial planner use variables were excluded from the analysis, resulting in a sample size of 826 total observations. a multinomial logit was conducted to compare the changes in self-efficacy between those who reported adopting a planner in wave 2 or wave 3 compared with those who reported not using a planner in the same period (or previous periods in the case of wave 3). this comparison tests whether self-efficacy in a prior period is higher for those who decide to seek help in a subsequent period when compared with those who did not use an advisor in either period. a multinomial logistic regression is a classification method that generalizes logistic regression to multiclass problems, where there are more than two possible discrete outcomes. this allows us to test 115j. letkiewicz et al. / financial services review 25 (2016) 105–126 the directional effects of planner use (i.e., does se in period 1 predict planner use in period 2?). multinomial logistic regression uses a linear predictor function f(k, i) to predict the probability that the observation i has outcome k as modeled below: f�k, i� � �k � xi (2) where �k is the set of regression coefficients associated with outcome k, and xi is the set of explanatory variables associated with observation i. in this analysis we use two time periods, t and t-1, so that the function takes the form: f�kt, it� � �kt�1 � xit�1 (3) where f(kt, it) is a linear function to predict the probability that observation i has outcome k in time period t; �kt�1 is the set of regression coefficients associated with outcome k in the previous wave of data collection; and xit�1 is the set of explanatory variables associated with observation i in the in the previous wave of data collection. 7. results 7.1. descriptive results table 2 displays summary statistics describing the sample. the sample is representative of the canadian population in terms of wealth, income, and homeownership. for example, statistics canada reports the median net worth in 2012 was $243,800 (statistics canada, 2012) and in 2011 69% of canadians owned their homes (statistics canada, 2011). the sample in this study has median net worth of $269,000 and a 70% homeownership rate. almost 62% of the respondents reported seeking financial planning help at some point during the previous five years. out of the total sample studied 58% are female, over 65% are married, and approximately 90% are at least 30 years of age. while approximately 80% of the sample had some form of postsecondary schooling, only 25% of non-help–seekers finished a university degree, in comparison with 40% of help-seekers. over 60% of the overall sample has children, with help-seekers more likely to be parents than non-help– seekers. we can expect that having children might also affect help-seeking tendencies, as it puts additional financial pressure on the respondents. owning a house might be another factor affecting help-seeking as 79% of help-seekers own a house instead of renting it, while only 56% of non-help–seekers do so. the amount of assets and debts differ slightly between help-seekers and non-seekers, with the former having a greater amount of both. assets and debt are right-skewed. most people have assets, particularly since 70% own a home, but a few have large asset holdings, with the maximum being $81.5 million. while data on the smaller panel sample are not shown, descriptive analysis of that sample indicates similar patterns. table 2 reports statistics on the cases that have no missing values, and we use that data set for most of the hypothesis tests. we examined the descriptive statistics for the data set that 116 j. letkiewicz et al. / financial services review 25 (2016) 105–126 includes the 709 cases with missing values, and also the correlation matrix of all the variables. there is no significant difference between the two data sets. table 3 shows the statistics on the components and total scores for the two behavioral variables we create from the survey: stress and financial self-efficacy (se). help seekers have higher levels of financial se but lower levels of stress compared with those who have not sought help and the differences are statistically significant. 7.2. logistic regression results table 4 shows the results of the logistic regressions in model 1 testing correlation with use of a financial planner (hypothesis 1, hypothesis 2a, and hypothesis 2b). the measures of stress and se are highly correlated (�0.705), and while the large sample size can alleviate table 2 descriptive statistics variable full sample help-seeker non-help–seeker n � 14,068 n � 8,669 (62%) n � 5,399 (38%) n percentage n percentage n percentage gender female* 8,191 58.2% 4,949 57.1% 3,242 60.0% male* 5,877 41.8% 3,720 42.9% 2,157 40.0% age 18–29* 1,584 11.3% 795 9.2% 789 14.6% 30–49* 5,641 40.1% 3,329 38.4% 2,312 42.8% 50–64* 5,089 36.2% 3,388 39.1% 1,701 31.5% 65 or older* 1,754 12.5% 1,157 13.3% 597 11.1% marital status married* 9,240 65.7% 5,989 69.1% 3,251 60.2% not married* 4,828 34.3% 2,680 30.9% 2,148 39.8% education did not go to college* 3,017 21.4% 1,449 16.7% 1,568 29.0% college degree/some college* 6,141 43.7% 3,696 42.6% 2,445 45.3% university degree* 4,817 34.2% 3,477 40.1% 1,340 24.8% employment status unemployed* 909 6.5% 406 4.7% 503 9.3% employed* 8,663 61.6% 5,509 63.5% 3,154 58.4% retired* 3,250 23.1% 2,198 25.4% 1,052 19.5% have children* 9,621 68.4% 6,148 70.9% 3,473 64.3% own a house* 9,857 70.1% 6,833 78.8% 3,024 56.0% plan to retire in the next five years* 2,095 14.9% 1,501 17.3% 594 11.0% unemployment risk* 1,715 12.2% 1,070 12.3% 645 11.9% income less than $50,000* 4,589 32.6% 2,156 24.9% 2,433 45.1% between $50k to $100k* 5,026 35.7% 3,346 38.6% 1,683 31.2% above $100,000* 3,655 26.0% 2,759 31.8% 896 16.6% mean median mean median mean median assets* $551,878 $269,000 $634,454 $370,000 $419,304 $100,000 non-housing debt* $37,793 $5,000 $38,831 $6,000 $28,260 $5,000 *indicates significance difference between help-seekers and non-help–seekers (p � .05); t tests for continuous variables and pearson’s �2 for categorical variables. 117j. letkiewicz et al. / financial services review 25 (2016) 105–126 some of the problems associated with multicollinearity, we enter them into the model separately to test the effect of each on the dependent variable. therefore, this step is performed in two stages. model 1a regresses everything but se on the help-seeking variable and model 1b regresses everything but stress on the help-seeking variable. the second column of the table shows the expected sign, if any, and whether the particular variable is one of the controls (c) or part of a hypothesis (hypothesis 1, hypothesis 2a, and hypothesis 3a). as expected, those with more education and those ages 50–64 are more likely to seek financial planning help. both these expectations had significant coefficients and moderately high odds ratios. we noted that having children was associated with help-seeking in the descriptive statistics and that relationship appears also in the regressions. the coefficients are virtually the same in all three models. we had no specific expectations of the other control variables, and most of them are not significant. we hypothesized that help-seeking would be positively related to value of assets, the higher income categories and home ownership. all three are statistically significant with positive signs and odds ratios greater than 1. the interpretation is quite simple and unsurprising. people with more assets and income have the ability to pay for professional financial table 3 behavioral variables 3a: financial self-efficacy factor (se) full sample help-seeker non-help–seeker (n � 14,068) (n � 8,669) (n � 5,399) i feel that i am prepared to manage through tough economic times 5.50 5.81 4.99 over the last five years, i have improved my ability to save 5.52 5.78 5.11 i do not know what to do to improve my financial situation (reverse coded) 5.62 5.90 5.18 i feel prepared in the event of an unexpected financial emergency 5.01 5.44 4.32 i am on the right track in terms of financial affairs 6.02 6.40 5.42 mean self-efficacy principal component scores* �0.016 0.183 �0.336 all items are on a 1 to 9 likert scale (1 � strongly disagree, 9 � strongly agree). cronbach’s � � 0.812. *mean difference is significant at the p � 0.05 level. 3b: financial stress factor (stress) full sample help-seeker non-help–seeker (n � 14,068) (n � 8,669) (n � 5,399) i worry a lot about my financial situation 5.22 5.00 5.57 i feel i barely get by every month 4.57 4.14 5.26 my finances are out of control 3.59 3.29 4.08 mean financial stress principal component scores* 0.018 �0.130 0.255 all items are on a 1 to 9 likert scale (1 � strongly disagree, 9 � strongly agree). cronbach’s � � 0.807. *mean difference is significant at the p � 0.05 level. 118 j. letkiewicz et al. / financial services review 25 (2016) 105–126 advice and are likely to gain a net benefit from it. this hypothesis is not the primary motivation of this paper, but it is essential to include the wealth variables and test the hypothesis to be able to test the stress and self-efficacy hypotheses with these variables also in the regression. the coefficients and the significance are virtually the same in all three models. the findings from model 1a indicate that objective stress (non-housing debt and unemployment risk) is positive and statistically significant, while subjective stress (stress) is negative and significant. we interpret this to mean that stress can be alleviated by the use of a financial planner while objective stress is an external factor that may not be influenced by a financial planner (i.e., one might be worried about losing their job, but a financial planner may not be able to remedy this). thus we find partial support for hypothesis 2a, that subjective stress (stress) is negatively correlated with having a financial planner. let us repeat the definition of financial self-efficacy: the belief in one’s own ability to succeed at financial tasks. the measure of se was to capture this psychological state with table 4 logistic regression tests of the main hypotheses variable name expected sign hypothesis (h) or control (c) model 1a model 1b � odds ratio wald � odds ratio wald wave 1 c 0.059 1.061 2.168 0.086* 1.090 4.587 gender female c 0.019 1.019 0.246 0.020 1.020 0.256 age (reference category: age 30–49) age 18–29 c 0.044 1.045 0.476 �0.010 0.990 0.026 age 50–64 �c 0.126** 1.134 6.104 0.156** 1.169 9.340 age 65� c �0.091 0.913 1.161 �0.067 0.935 0.634 marital status non-married c 0.118** 1.125 6.911 0.134** 1.143 8.779 education (reference category: college degree/some college) less than college �c �0.348*** 0.706 52.016 �0.335* 0.715 47.483 university degree �c 0.294*** 1.342 43.049 0.280 1.323 38.487 employment status (reference category: employed) unemployed c �0.097 0.908 1.525 �0.049 0.952 0.368 retired c 0.214** 1.239 10.873 0.131** 1.141 4.003 plan retire in next five years �c 0.322*** 1.380 28.886 0.271*** 1.311 20.013 have children �c 0.115* 1.122 6.486 0.117* 1.125 6.692 income level (reference category: � $50k) $50,000 to $100,000 �h1 0.387*** 1.473 69.120 0.357*** 1.429 57.983 more than $100,000 �h1 0.587*** 1.799 99.455 0.502*** 1.652 71.383 assets (ln) �h1 0.120*** 1.127 274.201 0.112*** 1.118 237.217 own a house �h1 0.174** 1.190 10.901 0.159** 1.173 9.051 non-housing debt (ln) �h2 0.006 1.006 1.755 0.013** 1.013 8.588 may lose job �h2 0.285*** 1.330 22.949 0.272*** 1.313 20.968 financial stress �h2 �0.152*** 0.859 48.933 financial self-efficacy �h3 0.321*** 1.378 215.405 constant �1.607*** 0.200 275.985 �1.516*** 0.220 242.666 number of observations 14,068 14,068 nagelkerke r2 (-2log likelihood) .176 (16,785.72) .190 (16,614.98) likelihood ratio test: �2 (df) 1,946.59 (19)*** 2,210.33 (19)*** *p � 0.05; **p � 0.01; ***p � 0.001. 119j. letkiewicz et al. / financial services review 25 (2016) 105–126 respect to personal finances. we expect that use of a financial planner is positively correlated with se—that is, someone using a financial planner will have more confidence in their financial affairs. the coefficient is significant and positive in model 1b, and the odds ratio is quite high thus finding support for hypothesis 3a. it should be noted that the wald statistic for se is high and only the constant term and assets have higher wald values. 7.3. multinomial regression results the next step of the analysis is to test the help-seeking hypotheses—whether stress or se lead one to seek financial help (hypothesis 2b and hypothesis 3b). the panel portion of the data, as discussed earlier, was used in this step. to test direction, the sample was split into four groups: (1) those who used a planner in all three waves; (2) those who did not use a planner in w1 or w2 but adopted a planner in w2 or w3, respectively; (3) those who used a planner in w1 or w2, but did not report using one in w2 or w3, respectively; and (4) those who reported not using a planner in all three waves. the groups of most interest are those who adopted a planner compared with those who did not. we can use the measures of se and stress in the initial period to compare these two groups. did those who adopted a planner in a subsequent period have higher se or stress when compared to someone who did not adopt a planner in the subsequent period? a multinomial logit was used to compare the two groups (those who adopted a planner in one period vs. those who did not) over two time periods using the panel portion of the dataset. results from the multinomial analysis are shown in table 5. the findings indicate that individuals who adopt a planner in period t had higher self-efficacy in period t-1 than people who did not adopt a planner. the control variables in the multinomial models are the same variables used in the logistic regression models. the result is consistent for those who adopted a planner in wave 2 and wave 3. this provides some evidence that higher self-efficacy is associated with the decision to seek financial planning help, thus we cannot reject hypothesis 3b and the support for it seems quite strong. the same pattern is not evident for stress which leads us to reject hypothesis 2b. while measures of stress are higher for those who seek help, the effect is not significant, and thus, we do not find evidence that stress is a motivating factor for seeking financial help. limitations in the dataset do not allow for a multinomial analysis of objective stress, so we can only conclude that those using a financial planner are more likely to exhibit signs of objective stress. table 5 multinomial results for likelihood of adopting a planner in t given se measure in t-1 � se odds ratio p value wave 2: adopted a planner in t (n � 41) self-efficacyt-1 .576 .267 1.778 .016 stresst-1 .436 .276 1.546 .057 wave 3: adopted a planner in t�1(n � 48) self-efficacyt .535 .277 1.707 .027 stresst .267 .269 1.306 .161 all variables in the logistic regressions were included as control variables. reference group is no planner. sample size for no planner (comparison group) was 228 in w2 and 284 in w3. 120 j. letkiewicz et al. / financial services review 25 (2016) 105–126 a robustness check was run to further probe the se finding. so far, we know that se is positively correlated with use of a planner and help-seeking, but does using a planner increase se? to test this, a multivariate repeated-measures analysis of variance (anova) was conducted to test within-subject changes in self-efficacy. this will help determine if se changed significantly for each individual across the three time periods based on their planner use behavior. results indicate that mean scores of se were not significantly different across waves for those who adopted a financial planner in w2 (f � 1.42, p � 0.245) or w3 (f � 0.002; p � 0.969). we interpret this to mean that while se is positively correlated with the decision to seek a planner and having a planner, it does not necessarily increase with the adoption of a planner. 7.4. interpretation of results the consistent results for hypothesis 1, and the reduction in nagelkerke r2 values when wealth variables are removed, provide considerable confidence in the validity of hypothesis 1. people with higher incomes and wealth are more likely to seek professional financial advice. stress is lower for individuals who use a financial planner, but high stress does not necessarily motivate someone to seek help. the negative correlation between wealth and assets might explain this finding (�.348, p � 0.001). the strong effect of wealth may be suppressing the effect of stress on help-seeking. the finding that stress is lower for those who use a financial planner supports the notion that assistance helps reduce stress. se is found to be a highly significant and important variable in the decision to seek help as well as the use of a financial planner. we find substantial support for the effect of se on both counts. while self-efficacy does not change with use of a financial planner, higher initial self-efficacy was correlated with their decision to use a planner. looking at results comprehensively, we conclude that people with a stronger sense of financial self-efficacy are more likely to seek professional financial advice when they need it and that use of a financial planner is correlated with lower stress. 8. summary and implications our results, and indeed the increasing literature in behavioral finance, show that financial planners need more than merely technical competence and marketing skills if they are to attract, serve and retain clients. this study highlights the importance of self-efficacy in the decision to seek professional financial planning help. these findings suggest that a society high in self-efficacy may make greater use of financial planners. although it seems like a plausible conjecture that financial stress would be a trigger for help-seeking, we do not find support for that supposition. the findings in this study have implications for financial planners, financial institutions, government agencies, and policy makers. first, we will discuss self-efficacy. the strong and consistent effects of self-efficacy are an important contribution of this article. then we discuss policy implications derived from our findings. 121j. letkiewicz et al. / financial services review 25 (2016) 105–126 8.1. increasing self-efficacy bandura (1977) suggests four ways to increase self-efficacy: performance accomplishments, vicarious experience, verbal encouragement, and physiological states. each of these strategies can be applied to personal finance in various ways to increase financial self-efficacy. accomplishments influence one’s sense of mastery and can lead to a greater sense of self-efficacy. one way to increase self-efficacy using the concept of performance accomplishments is to structure financial decisions ways that allow for small accomplishments while learning new skills. financial institutions and other agencies that deal with consumer finance issues should consider this when structuring consumer interactions. a suggestion is to keep some financial offerings simple and straightforward so consumers can understand how they work and feel confident in their ability to manage them. anderson (2012) suggests individuals set a simple goal; for example, create a plan to reduce spending and pay off a credit card balance. the plan requires some money management skills and discipline and is a good way to inspire confidence and motivation for more advanced tasks. vicarious experiences occur when one observes someone similar to them succeeding at a task. commercials, public service announcements, and other communications can be structured in a way to appeal to a diverse audience and provide valuable information and guidance for helping people get started. verbal encouragement takes place when one is encouraged to take on a task with the belief they can accomplish it. constructive feedback is important to building and maintaining a sense of self-efficacy. this is important for financial advisors to keep in mind when working with clients and for public workers dealing with a financially illiterate population. finally, the way people experience, interpret and evaluate emotional states is important for how they develop self-efficacy beliefs. extremely nervous or anxious people tend to doubt themselves and may therefore have a weak sense of self-efficacy. this is in line with our finding that stress can either paralyze or mobilize someone to seek financial help. household finances can be stressful for families. one way to reduce the stress and anxiety is to establish basic ground rules and commit to a plan with your partner (anderson, 2012) and your planner (if you have one). this can help facilitate an environment with well-established goals and principles and can facilitate positive communication and behaviors around a shared goal, thus reducing stress and anxiety. 8.2. policy implications there are ongoing efforts in both the united states and canada to improve financial literacy. the essential argument is that if citizens are more financially literate, they will make better financial decisions which will improve their well-being. there is much debate about the effectiveness of financial education (willis, 2008) that has led researchers to focus on behavioral interventions. studies by fernandes, lynch, and netemeyer (2014) and parker et al. (2012) find that confidence is key to improving financial well-being and the results of this study support those findings. given the positive 122 j. letkiewicz et al. / financial services review 25 (2016) 105–126 behaviors associated with self-efficacy and confidence, both should be considered when designing and deploying financial literacy programs. the effect of wealth on the decision to seek financial advice will not surprise anyone, and it suggests a direction for public policy. when we think about the services that society deems as important, governments or other agencies provide some of them for little or no direct cost to consumers. for example, in canada, healthcare is universally available without additional fees, and in the united states more than 100 million people have medical expenses paid by medicare and medicaid. legal aid is available in canada and the united states to those who cannot otherwise afford legal counsel. we know that stress from financial issues can cause mental and physical illness (e.g., buckland, 2012; ennis et al., 2000), but this knowledge has not translated into financial help for most citizens. buckland (2012) describes the patchwork nature of financial advising support and access to financial services for lower income families in canada. many families get their basic financial information from banks and credit unions, but 9% of americans, 8% of u.k. residents and 3% of canadians do not have a bank account, and the percentages are much higher for the lowest income decile. some agencies provide financial counseling to low income families, but these services are not universally available. for example, land grant colleges in the united states offer financial counseling to low income families, but these services are not widely available and no such practice exists in canada. some services such as tax clinics for low income families depend almost exclusively on volunteers. debt counseling agencies depend on voluntary funding from financial institutions that is not guaranteed in the long run. in short, access to financial counseling is not well-coordinated or funded and perhaps we need a more comprehensive policy to address this. the exploration of financial clinics modeled on legal aid accessibility would be of some merit and policy options that ensure everyone has some basic access to financial counseling may be a worthy endeavor. acknowledgments the authors thank the fp foundationtm for its generous support and natallia uborceva for research support. references anderson, c. (2012). financial self-efficacy in your clients’ lives. journal of financial planning practice management blog. [online] available at http://practicemanagementblog.onefpa.org/2012/03/20/financialself-efficacy-in-your-clients-lives/ (accessed july 23, 2015). bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84, 191–215. bandura, a. (1992). exercise of personal agency through the self-efficacy mechanism. in r. schwarzer (ed.), self-efficacy: thought control of action (pp. 3–38). washington, d.c.: hemisphere. bandura, a. (1997). self-efficacy: the exercise of control. new york, ny: freeman. 123j. letkiewicz et al. / financial services review 25 (2016) 105–126 bandura, a. (2006). guide for constructing self-efficacy scales. in t. urdan & f. pajares (eds.). self-efficacy beliefs of adolescents. charlotte, nc: information age publishing. blanchett, d., & kaplan, p. (2013). alpha, beta, and now … gamma. the journal of retirement, 1, 29–45. britt, s., grable, j. e., goff, b. s. n., & white, m. (2008). the influence of perceived spending behaviors on relationship satisfaction. journal of financial counseling and planning, 19, 31–43. buckland, j. (2012). hard choices: financial exclusion, fringe banks and poverty in urban canada. toronto: university of toronto press. burgogne, c. (1990). money in marriage: how patterns of allocation both reflect and conceal power. sociological review, 38, 634–665. collins, j. m. (2010). a review of financial advice models and the take-up of financial advice. center for financial security working paper, 10–5. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. cummings, b. f., & james, r. n. iii. (2014). factors associated with getting and dropping financial advisors among older adults. journal of financial counseling and planning, 25, 129–147. danes, s. m., & haberman, h. r. (2007). teen financial knowledge, self-efficacy, and behavior: a gendered view. financial counseling and planning, 18, 48–60. dearden, c., goode, j., whitfield, g., & cox, l. (2010). credit and debt in low-income families. york, uk: joseph rowntree foundation. dietz, b. e., carrozza, m., & ritchey, p. n. (2003). does financial self-efficacy explain gender differences in retirement saving strategies? journal of women & aging, 15, 83–96. du plessis, k., lawton, j., & corney, t. (2010). unemployed and struggling: the lack of financial help-seeking behavior exhibited by male construction workers. the australian journal of financial planning, 6, 1–8. engelberg, e. (2007). the perception of self-efficacy in coping with economic risks among young adults: an application of psychological theory and research. international journal of consumer studies, 31, 95–101. ennis, n. e., hobfoll, s. e., & schroeder, k. e. e. (2000). money doesn’t talk, it swears: how economic stress and resistance resources impact inner-city women’s depressive mood. american journal of community psychology, 28, 149–173. fernandes, d., lynch, j. g., jr., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors. management science, 60, 1861–1883. financial planning standards council. (2013) the value of financial planning. (available at http://www.fpsc.ca/ value-financial-planning). finke, m. s., huston, s. j., & winchester, d. d. (2011). financial advice: who pays. journal of financial counseling and planning, 22, 18–26. folkman, s., scheaefer, c., & lazarus, r. s. (1979). cognitive processes as mediators of stress and coping. in v. hamilton & d. m. warburton (eds.), human stress and cognition: an information-processing approach. london, uk: wiley. fosnacht, k., sarraf, s., howe, e., & peck, l. (2013). how important are high response rates. paper presented at the annual meeting of the association for institutional research, long beach, ca. gecas, v. (1989). the social psychology of self-efficacy. annual review of sociology, 15, 291–316. grable, j. e., & joo, s. (1999). financial help-seeking behavior: theory and implications. financial counseling and planning, 10, 14–25. grable, j. e., & joo, s. (2001). a further examination of financial help-seeking behavior. journal of financial counseling and planning, 12, 55–65. grembowski, d., patrick, d., diehr, p., durham, m., beresford, s., kay, e., & hecht, j. (1993). self-efficacy and health behavior among older adults. journal of health and social behavior, 34, 89–104. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. hanna, s. d., & lindamood, s. (2010). quantifying the economic benefits of personal financial planning. financial services review, 19, 111–127. jerusalem, m., & schwarzer, r. (1992). self-efficacy as a resource factor in stress appraisal processes. in r. schwarzer (ed.), self-efficacy: thought control of action (pp. 195–213). washington, d.c.: hemisphere. 124 j. letkiewicz et al. / financial services review 25 (2016) 105–126 joo, s. (1998). personal financial wellness and worker job productivity (order no. 9949872). proquest dissertations & theses global. (304467171). (retrieved from http://ezproxy.library.yorku.ca/login?url�http:// search.proquest.com/docview/304467171?accountid�15182). kim, j., & garman, e. t. (2003). financial stress and absenteeism: an empirically derived model. journal of financial counseling and planning, 14, 31–42. kinniry, f., jaconetti, c., dijoseph, m., & zilbering, y. (2014). putting a value on your value: quantifying vanguard advisor’s alpha. (available at http://www.vanguard.com/pdf/isgqvaa.pdf). krause, n. & baker, e. (1992). financial strain, economic values, and somatic symptoms in later life. psychology and aging, 7, 4–14. lapp, w. m. (2010). behavior models for prosperity: a statistical assessment of savings and behavioral change. earn research brief. earned assets resource network [online]. (available at https://www.earn. org/wp-content/uploads/2015/03/5_-_behavioral_models_for_prosperity-_a_statistical_assessment_of_savings_ and_behavioral_change-1.pdf). lazarus, r. s. (1991). emotion and adaptation. london: oxford university press. lazarus, r. s., & folkman, s. (1984). stress, appraisal, and coping. new york, ny: springer. lent, r. w., brown, s. d., & larkin, k. c. (1986). self-efficacy in the prediction of academic performance and perceived career options. journal of counseling psychology, 33, 265–269. lim, h., heckman, s. j., montalto, c. p., & letkiewicz, j. (2014). financial stress, self-efficacy, and financial help-seeking behavior of college students. journal of financial counseling & planning, 25, 148–160. lown, j. m. (2011). development and validation of a financial self-efficacy scale. journal of financial counseling and planning, 22, 54–63. marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32, 625–643. mates, d., & allison, k. r. (1992). sources of stress and coping responses of high school students. adolescence, 27, 461–474. park, c. l., & folkman, s. (1997). meaning in the context of stress and coping. review of general psychology, 1, 115–144. parker, a. m., bruin, w. b., yoong, j., & willis, r. (2012). inappropriate confidence and retirement planning: four studies with a national sample. journal of behavioral decision making, 25, 382–389. pike, g. r. (2012). nsse benchmarks and institutional outcomes: a note on the importance of considering the intended uses of a measure in validity studies. research in higher education, 54, 149–170. prochaska, j. o. (1984). systems of psychotherapy: a transtheoretical analysis (2nd ed.) new york, ny: dorsey. prochaska, j. o., & velicer, w. f. (1997). the transtheoretical model of health behavior change. american journal of health promotion, 12, 38–48. remund, d. l. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. schwarzer, r., & jerusalem, m. (1995). generalized self-efficacy scale. measures in health psychology: a user’s portfolio. causal and control beliefs, 1, 35–37. shockey, s. s., & seiling, s. b. (2004). moving into action: application of the transtheoretical model of behavior change to financial education. journal of financial counseling & planning, 15, 41–52. statistics canada. (2011). homeownership and shelter costs in canada. national household survey, 2011. (available at http://www12.statcan.gc.ca/nhs-enm/2011/as-sa/99-014-x/99-014-x2011002-eng.pdf). statistics canada. (2012). changes in wealth across the income distribution, 1999 to 2012. insights on canadian society. (available at http://www.statcan.gc.ca/pub/75-006-x/2015001/article/14194-eng.pdf). taylor, m., jenkins, s., & sacker, a. (2011). financial capability and psychological health. journal of economic psychology, 32, 710–723. wrosch, c., heckhausen, j., & lachman, m. e. (2000). primary and secondary control strategies for managing health and financial stress across adulthood. psychology and aging, 15, 387–399. 125j. letkiewicz et al. / financial services review 25 (2016) 105–126 willis, l. e. (2008). against financial-literacy education. iowa law review, 94, 197–285. xiao, j. j., newman, b., prochaska, j. m., leon, b., bassett, r. l., & johnson, j. l. (2004). applying the transtheoretical model of change to consumer debt behavior. journal of financial counseling & planning, 15, 89–100. zick, c. d., & mayer, r. n. (2013). evaluating the impact of financial planners. in o. s. mitchell & k. smetters (eds.), the market for retirement financial advice (pp. 153–181). oxford: oxford university press. 126 j. letkiewicz et al. / financial services review 25 (2016) 105–126 who seeks financial advice? maher h. alyousifa,*, charlene m. kalenkoskib atexas tech university, 5301 chicago avenue apt 6103, lubbock, tx 79414, usa bprogram, texas tech university, 1301 akron avenue, lubbock, box 41210, lubbock, tx 79409-1210, usa abstract the determinants of seeking five types of financial advice are examined and are found to be consistent across the different types of advice. in addition, no significant differences are found among subsamples defined by gender, age, and financial literacy. income and risk tolerance are related positively to the demand for financial advice and more greatly affect the probability of seeking advice than do other variables. a low perception of financial knowledge, which can be a proxy for self-confidence, and financial fragility decrease the probability of seeking financial advice. © 2017 academy of financial services. all rights reserved. jel classification: d14; g20 keywords: financial advice; risk tolerance; financial knowledge; financial literacy; financial fragility 1. introduction the demand for professional financial advice by the u.s. population is estimated to be within the range of 25–33% (collins, 2012) despite the fact that many american households are experiencing financial difficulty (brooks, wiedrich, sims, and rice, 2015). according to a liquid asset poverty measure by assets and opportunities scorecard,1 for example, 44% of u.s. households have less than three months of savings. moreover, 55% of consumers have credit scores that make reasonably priced loans unattainable (brooks et al., 2015), and only 22% of workers are very confident about having enough money to live comfortably during retirement (vanderhei and copeland, 2015). understanding the correlates of financial* corresponding author. tel.: �966-55-622-6786; fax: �966-13-664-0602. e-mail address: maheryousif@hotmail.com (m.h. alyousif) financial services review 26 (2017) 405–432 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. advice-seeking behavior helps to explain the coexistence of reported low financial satisfaction and measured low demand for financial advice among american households. investors who rely on their own understanding often make poor financial decisions because of a lack of knowledge, information costs, and behavioral biases (fischer and gerhardt, 2007). these challenges warrant the use of professional advisers, who serve different purposes, deal with various products, and can help their clients navigate the high degree of financial uncertainty. using the 2012 national financial capability study (nfcs), a cross-sectional study that was funded by the financial industry regulatory authority’s (finra) investor education foundation, this article investigates the characteristics of financial-advice-seeking behavior for five types of financial advice: debt counseling, savings/investment, mortgages/loans, insurance, and tax planning. a probit regression model is estimated to examine the associations between income, risk tolerance, financial knowledge, financial literacy, financial fragility, and a set of demographic variables and the probability of seeking financial advice. additionally, this article examines the determinants of financial-advice-seeking behavior for subsamples defined by gender, age, and financial literacy. 2. literature review the existing literature on the characteristics of financial-advice-seeking behavior examines this conduct generally and for specific types of advice such as debt counseling, retirement planning, and investment management (collins, 2012; finke, huston, and winchester, 2011; grable and joo, 1999; hackethal, haliassos, and jappelli, 2012; heo, grable, and chatterjee, 2013; inderst and ottaviani, 2012; kramer, 2012; robb, babiarz, and woodyard, 2012; heo, grable, & chatterjee, 2013; salter, harness, and chatterjee, 2010; scott and finke, 2013; seay, kim, and heckman, 2016; simms, 2014). these studies identify age, gender, wealth, income, home ownership, education, financial knowledge, confidence, risk tolerance, and negative life events as factors that influence the demand for financial advice. age is a significant determinant of seeking advice in all areas of personal finance, has been found to be related positively to debt counseling for those aged 25–54, and is related negatively to debt counseling for respondents who are aged 65 or older (robb et al., 2012). grable and joo (1999) find that younger households and those who do not own homes are more likely to seek financial help compared with homeowners and older individuals who may experience self-concealment2 to protect their perceived life achievement. in addition, individuals who demonstrate bad financial behaviors (e.g., overspending, overusing credit, and not saving for retirement) and who experience financial stressors (e.g., death of a family member, divorce, and loss of a job) are more likely to seek financial help. however, hackethal et al. (2012) find that older clients (over 50) are more likely to use a financial adviser compared with younger clients aged 18–30. gender influences the decision to seek financial advice. because of their overconfidence in managing finances, males resist financial counseling and are less likely to seek financial advice compared to females (finke et al., 2011; hackethal et al., 2012; robb et al., 2012). 406 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 in contrast, tang and lachance (2012) find that gender and home ownership do not affect the demand for financial advice. income has been found to be related positively to the demand for financial advice (robb et al., 2012). however, other studies indicate that wealth has more of an impact on the decision to seek financial advice compared to income (finke et al., 2011; hackethal et al., 2012; hanna, 2011). advisers are inclined to provide their services to clients who are self-employed, female, have high wealth, and have more work experience (hackethal et al., 2012). on the other hand, calcagno and monticone (2014) do not find support for the predicted associations between high wealth or high income and the probability of seeking financial advice. although education increases the likelihood of seeking financial advice (finke et al., 2011; hanna, 2011; inderst and ottaviani, 2012), perceived knowledge about managing finances reduces the likelihood of asking for help (finke et al., 2011). however, other studies find that knowledge and confidence are correlated positively with the use of financial advice (calcagno and monticone, 2014; collins, 2012; inderst and ottaviani, 2012; robb et al., 2012). the literature also investigates the determinants of advice-seeking behavior from other angles. studies about the sources of advice examine an individual’s tendency to seek financial advice from nonprofessional versus professional sources (grable and joo, 2001), bank-affiliated versus independent advisers (hackethal et al., 2012), social networks versus paid advisers (chang, 2005; loibla and hira, 2006), and the use of financial planners (hanna, 2011; letkiewicz, robinson, and domian, 2016). studies that examine advice seeking by certain groups focus on less-sophisticated or low-income clients (kramer, 2012; tang and lachance, 2012), older adults (cummings and james, 2014), affluent retirees (salter et al., 2010), and the middle class (winchester and huston, 2015). they also examine the effects of financial literacy on the use of financial advice (calcagno and monticone, 2014; collins, 2012; robb et al., 2012; seay et al., 2016) and the determinants of seeking comprehensive versus partial financial advice (elmerick, montalto, and fox, 2002; finke et al., 2011; tang and lachance, 2012). financial risk tolerance and financial satisfaction have been found to play a role in determining whether people seek financial help from professionals or nonprofessionals such as family members, friends, or work colleagues (grable and joo, 2001; lin and lee, 2004). chang (2005) finds that low socioeconomic status affects people’s decisions to seek information about investment and savings from their social network rather than from paid financial advisers. elmerick et al. (2002) find that the determinants of seeking comprehensive financial advice and seeking advice regarding savings and investment are different from the determinants of seeking advice regarding debt and borrowing. education, income, net worth, and financial assets are related positively to the probability of seeking comprehensive financial advice, while age is related negatively to the use of comprehensive financial planners. hanna (2011) studies the demand for personal financial planners and finds that age increases the likelihood of using a planner until the age of 42 then decreases it. the determinants that increase the likelihood of using a financial planner include education, risk tolerance, being a single-female-headed household, and being black (hanna, 2011). cummings and james (2014) examine the factors that influence the decision to begin or discontinue the use of financial advisers among older adults and find that becoming wid407m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 owed, receiving family help, and experiencing an increase in income or net worth are significant factors in influencing the demand for financial advisers. studying the sentiment of financial-advice-seeking behavior among the middle class, winchester and huston (2015) find that the expected benefit relative to income is a more significant determinant of seeking financial advice than individuals’ attitudes regarding cost. financial literacy increases the probability of seeking financial advice (calcagno and monticone, 2014), and such advice is a complement to rather than a substitute for financial capability (collins, 2012). as income, education, and financial knowledge increase, the likelihood of seeking financial advice increases; however, self-assessment of financial literacy is related negatively to seeking financial advice, while measured financial literacy has no effect on the demand for such advice (kramer, 2016). this article contributes to the literature that examines the determinants of seeking professional financial guidance by focusing on five specific types of financial advice and investigating three subsamples that are defined by gender, age, and financial literacy. because each type of financial advice has a specific purpose, studying the determinants of seeking advice about debt, savings/investment, mortgages/loans, insurance, and tax planning provides valuable insights into advice-seeking behavior. in addition to financial knowledge and risk attitudes that robb et al. (2012) examine in their study, this article constructs two variables, financial fragility and financial literacy, to comprehend the effect of financial difficulty and the grasp of basic financial concepts on seeking financial advice. the focus on females, the young, and the financially illiterate is related to specific characteristics, examined in the empirical literature, that distinguish and influence the financial behavior of these subsamples. females and young respondents are most likely to experience financial stress and difficulties (orc, 2015; simms, 2014), and the financially illiterate are susceptible to suboptimal financial decisions (lusardi, 2008; lusardi and mitchell, 2009; lusardi and tufano, 2009; van rooij, lusardi, and alessie, 2011). the empirical literature about gender differences in financial knowledge finds that females score lower than males in financial literacy tests, are more likely to be dissatisfied with their personal financial situation, and are less confident in their financial skills and their ability to manage financial emergencies (goldsmith and goldsmith, 2006; hira and mugenda, 2000; hung, yoong, and brown, 2012). gender differences in investment knowledge, financial skills, and risk tolerance between females and males might explain and exacerbate the economic status disadvantage of females that manifests in lower lifetime earnings, lower wealth, and lower retirement-plan participation (bajtelsmit and bernasek, 1996; hung et al., 2012). while females are more patient than males in the measurement of rate of time preference, they exhibit more risk aversion and less interest in financial subjects (donkers and van soest, 1999). the gender role differences and division of labor within households provide another explanation for the disparity in the consumption of financial services (burton, 1995; morris and meyer, 1993). the literature on financial competency among young adults shows weak financial literacy and a lack of understanding of basic financial knowledge, which affect the quality of their financial decisions and lead them to commit costly financial mistakes (lusardi, 2008; lusardi and mitchell, 2014; lusardi, mitchell, and curto, 2010). a high level of debt at an early age, for example, impedes the accumulation of wealth and forestalls their contributions to employer408 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 provided retirement plans (lusardi et al., 2010). additionally, weak financial numeracy has negative impacts on critical decisions related to financing an education and making major purchases such as buying a car (lusardi, 2012). laibson, gabaix, driscoll, and agarwal (2007) find that financial sophistication has a hump-shaped pattern, which could explain the high borrowing costs in terms of interest rates and fees by younger and older adults. research indicates that financial literacy influences financial-decision making and that the understanding of basic financial concepts is associated with retirement planning, stock market participation, and individuals’ borrowing behavior (hastings and mitchell, 2011; lusardi, 2008; lusardi and mitchell, 2009; van rooij et al., 2011). individuals who are not financially sophisticated are less likely to own stocks because they do not comprehend the working of financial markets and asset pricing and are more likely to seek financial advice from friends and family members than from financial professionals (van rooij et al., 2011). 3. data the dependent variables in the analysis in this article are indicators for whether or not five different types of financial advice were sought, debt counseling, savings/investment, taking out a mortgage/loan, insurance of any type, and tax planning. each variable takes a value of 1 if the specific type of advice was sought from a financial professional and 0 if it was not. the independent variables are gender, age, race, education, marital status, number of children, income, risk tolerance, perceived financial knowledge, financial literacy, and financial fragility. because the three subsamples are defined by age, gender, and financial literacy, those variables are excluded from their regressions. female is a dichotomous variable that takes a value of 1 if the respondent is female and 0 if the respondent is male. age is categorized into six ranges: 18–24, 25–34, 35–44, 45–54, 55–64, and 65 or more. a categorized dichotomous variable for each age range is defined (the omitted category is 65�). race is a dichotomous variable that takes a value of 1 if the respondent is white and 0 if the respondent is nonwhite.3 education is categorized into three levels: high school or less, some college, and college or more (the omitted category is college or more). marital status is categorized as married, living with a partner, and single (the omitted category is married). the number of financially dependent children is categorized into five choices: not having any children, having one child, having two children, having three children, and having four children or more. the omitted category is not having any children. income is categorized into eight ranges, and for each range a dichotomous variable is defined. the comparison group is less than $15,000. the risk tolerance variable is a subjective answer by respondents to the following question: “when thinking of your financial investment, how willing are you to take risk?” the answers fall on a 10-point scale that ranges from 1 (not at all willing) to 10 (very willing). in this analysis, they are aggregated into three risk tolerance levels4 and the omitted category is low risk tolerance. the financial knowledge variable is a subjective assessment by respondents to the following question: “how would you assess your overall financial knowledge?” the answers fall on a seven-point scale that ranges from 1 (very low) to 7 (very 409m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 high). in this analysis, they are aggregated into three perceived financial knowledge levels5 and the omitted category is high financial knowledge. financial fragility is constructed from seven questions that examine respondents’ tendency to experience overspending, difficulty in covering expenses, the lack of an emergency fund, inability to come up with $2000 in the next month, the absence of a retirement plan, and incurring too much debt. this variable is a sum of these signs of financial fragility. overspending is a dichotomous variable that takes a value of 1 if the respondent’s spending is more than income and 0 otherwise. the difficulty of covering expenses and paying all bills is a dichotomous variable that takes a value of 1 if the respondent indicated it was very difficult or somewhat difficult to cover expenses and 0 otherwise. having no emergency fund that would cover three months of expenses is a dichotomous variable that takes a value of 1 if the respondent answered “no” and 0 otherwise. the confidence to come-up with $2000 is a dichotomous variable that takes a value of 1 if the respondent could probably not or is certain she/he could not come-up with that amount and 0 otherwise. having no retirement plan is a dichotomous variable that takes a value of 1 if the respondent has neither a private plan nor a plan through a current or a previous employer and 0 otherwise. having too much debt is a dichotomous variable that has a value of 1 if the respondent agrees or strongly agrees with that statement and 0 otherwise. financial literacy consists of five questions that measure respondents’ understanding of compound interest, inflation, bond prices, mortgage interest, and risk. this variable is a sum of the correct answers to these questions and has a range of 0–5. table 1 provides the distribution of correct financial literacy answers and shows that respondents who answered 4–5 questions correctly are between 16 and 26%. fig. 1 shows that respondents have difficulty understanding the effect of interest rates on bond prices and the risk-return trade-off in buying a single company’s stock versus purchasing a share of a mutual fund. 4. model the model estimated in this article is a probit model: yij * � b0 � xi �b � �ij (1) yij � � 1 if yij * � 0 0 if yij * � 0 where yij * is a latent variable representing the net benefit an individual i perceives he or she will receive from seeking financial advice related to task j where j is one of the following: debt counseling, savings/investment, a mortgage/a loan, insurance, and tax planning,6 yij is equal to 1 if the respondent reported seeking that type of financial advice and 0 otherwise; xi is a matrix of explanatory variables representing income,7 risk tolerance, perceived financial knowledge, financial literacy, financial fragility, female, white, age, education, marital status, and number of children; and uij is an error term that follows the standard normal distribution. 410 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 table 1 summary statistics mean standard error dependent variables debt counseling 0.0906 0.0022 savings or investment advice 0.2871 0.0033 mortgage or loan advice 0.2020 0.0030 insurance advice 0.3028 0.0034 tax planning 0.1812 0.0029 independent variables gender male 0.4858 0.0037 female 0.5142 0.0037 age (years) 18–24 0.1231 0.0027 25–34 0.1830 0.0030 35–44 0.1635 0.0027 45–54 0.1962 0.0029 55–64 0.1791 0.0028 65� 0.1551 0.0026 race white 0.6647 0.0037 non-white 0.3353 0.0037 education level high school or less 0.3812 0.0037 some college 0.3591 0.0036 college or more 0.2597 0.0030 marital status married 0.5403 0.0037 living with a partner 0.0816 0.0021 single 0.3782 0.0037 number of children no children 0.3181 0.0035 one child 0.1699 0.0028 two children 0.1312 0.0025 three children 0.0567 0.0018 four children or more 0.0337 0.0014 no financial dependent children 0.2905 0.0033 annual income less than $15,000 0.1426 0.0027 $15,000 to less than $25,000 0.1225 0.0025 $25,000 to less than $35,000 0.1155 0.0024 $35,000 to less than $50,000 0.1470 0.0026 $50,000 to less than $75,000 0.1882 0.0029 $75,000 to less than $100,000 0.1153 0.0023 $100,000 to less than $150,000 0.1076 0.0023 $150,000 or more 0.0613 0.0017 risk-tolerance level low 0.3517 0.0035 medium 0.4388 0.0037 high 0.1746 0.0029 (continued on next page) 411m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 5. hypotheses h1: income is expected to be related positively to seeking financial advice about savings/ investment, mortgages/loans, insurance, and tax planning, and to relate negatively with debt counseling for the entire sample and subsamples. previous literature finds a positive relation between income and the demand for financial advice. h2: risk tolerance is expected to be related positively to seeking financial advice for the entire sample and subsamples. research indicates that this factor has been found to increase the likelihood to seek financial help from professionals. h3: perceived financial knowledge is expected to be related negatively to seeking financial advice for the entire sample and subsamples. although some studies find that perceived knowledge reduces the likelihood of asking for advice, others report a positive relation between knowledge and the use of financial advice. h4: financial literacy is expected to be related positively to seeking all types of financial advice except debt counseling for the entire sample and subsamples. the literature finds that financial literacy increases the probability of seeking advice. however, some studies differentiate between the effect of subjective and objective assessment of financial literacy on the demand for financial advice. h5: financial fragility is expected to be related positively to seeking financial advice for the entire sample and subsamples. although respondents who experience financial stressors are more likely to seek advice, those who are financially fragile might not afford the purchase of financial advice. table 1 (continued) mean standard error perceived financial knowledge low 0.0915 0.0022 medium 0.1487 0.0027 high 0.7288 0.0034 financial literacy 2.8781 0.0110 financial fragility 2.3821 0.0133 number of observations 25,509 0% 5% 10% 15% 20% 25% 30% 35% debt counseling savings or investment mortgage or loan insurance tax planning pe rc en ta ge o f a dv ic e u se rs type of financial advice fig. 1. demand for financial advice. source: author’s tabulation of data from the 2012 finra national financial capability study. 412 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 6. descriptive statistics the summary statistics of the dependent and independent variables are provided in table 1. the first important observation to be made is the low demand for financial advice, which is utilized by 9–30% of the population, depending on the type of advice. sixty-six percent of respondents are white and 34% are non-white respondents. seventy-four percent of respondents have some college education or less. married individuals are the majority at 54%, followed by singles at 38%, and individuals who are living with partners at 8%. thirty-two percent of respondents have no children and 29% have children who are financially independent. proportions are distributed evenly among the income categories, except for the $50,000 to $75,000, which represents 19% of the population, and those making $150,000 or more, which represents 6% of the population. only 17% of respondents have a high-risk-tolerance level, while the majority of respondents (44%) have a medium-risk-tolerance level.8 each type of financial advice serves a specific purpose, which explains the advice use distribution in fig. 1 and shows that the two most sought after types of financial advice are insurance and savings/investment. even though 86% of respondents to a cfp stress awareness survey point to debt and daily expenses as the two primary sources of stress (orc, 2015), debt counseling is the least demanded type of advice at 9%. although 73% of respondents rated themselves high when asked to give a subjective assessment of their overall financial knowledge,9 average financial literacy on a scale of 0–5 is only 2.9. financial fragility is measured on a scale of 0–6, and each number represents the cumulative signs of financial difficulty across the seven financial fragility questions. table 2 reveals that only a quarter of respondents do not experience any of the six signs of financial fragility. fig. 2 shows that 56% of respondents report difficulty in covering expenses and paying bills and that 55% have no emergency fund that could cover expenses for 3 months. the comparison between females and males is provided in table 3. the t test results indicate that the significant difference between females and males is related to seeking financial advice about savings/investment, mortgages/loans, and tax planning. as for debt counseling, and insurance, there is no evidence of a statistically significant difference. the comparison between the young (18–44) and the old (45�) is provided in table 4. the t test results indicate that the significant difference between the young (18–44) and the old table 2 distribution of financial fragility measure fragility degree level percentage of respondents 0 23.56% 1 15.88% 2 14.59% 3 14.98% 4 16.63% 5 11.14% 6 3.21% the financial fragility measure consists of seven questions in the 2012 nfcs, which examine a respondent’s tendency to experience overspending, difficulty in covering expenses, lack of an emergency fund, inability to raise $2,000 in the next month, lack of any retirement plan, and having a high level of debt. the table shows the percentage of respondents who experience different degrees of financial fragility. 413m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 (45�) is related to seeking financial advice about debt counseling, savings/investment, and mortgages/loans. as for insurance and tax planning, there is no evidence of a statistically significant difference. the comparison between the financially illiterate and financially literate respondents is provided in table 5. financial illiteracy is defined as answering two questions or less 0% 10% 20% 30% 40% 50% 60% pe rc en ta ge o f r es po nd en ts fig. 2. distribution of financial fragility issues. source: author’s tabulation of data from the 2012 finra national financial capability study. table 3 summary statistics (females vs. males) female male mean standard error mean standard error dependent variables debt counseling 0.0878 0.0028 0.0935 0.0034 savings or investment advice 0.2718 0.0043 0.3033 0.0051 *** mortgage or loan advice 0.1872 0.0037 0.2177 0.0046 *** insurance advice 0.3019 0.0045 0.3037 0.0051 tax planning 0.1660 0.0036 0.1973 0.0045 *** independent variables annual income less than $15,000 0.1517 0.0036 0.1329 0.0040 *** $15,000 to less than $25,000 0.1385 0.0035 0.1055 0.0035 *** $25,000 to less than $35,000 0.1267 0.0033 0.1037 0.0035 *** $35,000 to less than $50,000 0.1490 0.0036 0.1449 0.0039 $50,000 to less than $75,000 0.1774 0.0037 0.1997 0.0045 *** $75,000 to less than $100,000 0.1031 0.0029 0.1281 0.0037 *** $100,000 to less than $150,000 0.0950 0.0029 0.1209 0.0035 *** $150,000 or more 0.0586 0.0023 0.0642 0.0026 risk-tolerance level low 0.4286 0.0049 0.2703 0.0049 *** medium 0.4169 0.0049 0.4620 0.0056 *** high 0.1130 0.0032 0.2397 0.0049 *** perceived financial knowledge low 0.1055 0.0031 0.0768 0.0031 *** medium 0.1647 0.0037 0.1317 0.0039 *** high 0.6930 0.0046 0.7668 0.0049 *** financial literacy 2.6110 0.0141 3.1609 0.0166 *** financial fragility 2.5171 0.0180 2.2391 0.0195 *** *significance at 10% level; **significance at 5% level; ***significance at 1% level. 414 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 correctly out of the five financial literacy questions in the survey. the t test results indicate that the significant difference between the two groups is related to seeking all types of financial advice. 7. results table 6 reports the estimation results for five probit regression models on the entire sample. the dependent variables are indicators for whether or not five different types of financial advice were sought, debt counseling, savings/investment, taking out a mortgage/ loan, insurance of any type, and tax planning. to examine how advice seeking varies by gender, age, and financial illiteracy, three dummy variables representing those subsamples are included in the model. the results of the probit regression models on the entire sample show consistently that income and risk tolerance are related positively to seeking all types of financial advice. these results confirm that the existing findings in the literature extend to these specific applications. the two constructed variables, financial literacy and financial fragility, have an opposite table 4 summary statistics (young vs. old) young (18–44) old (45�) mean standard error mean standard error dependent variables debt counseling 0.1160 0.0037 0.0681 0.0025 *** savings or investment advice 0.2610 0.0050 0.3103 0.0045 *** mortgage or loan advice 0.2272 0.0047 0.1796 0.0037 *** insurance advice 0.3060 0.0052 0.2999 0.0044 tax planning 0.1790 0.0044 0.1831 0.0037 independent variables annual income less than $15,000 0.1876 0.0045 0.1027 0.0031 *** $15,000 to less than $25,000 0.1302 0.0039 0.1157 0.0032 *** $25,000 to less than $35,000 0.1201 0.0037 0.1115 0.0031 * $35,000 to less than $50,000 0.1444 0.0040 0.1493 0.0035 $50,000 to less than $75,000 0.1812 0.0044 0.1944 0.0039 ** $75,000 to less than $100,000 0.1100 0.0035 0.1199 0.0031 ** $100,000 to less than $150,000 0.0820 0.0031 0.1303 0.0033 *** $150,000 or more 0.0444 0.0023 0.0763 0.0025 *** risk-tolerance level low 0.2784 0.0050 0.4166 0.0049 *** medium 0.4508 0.0057 0.4283 0.0048 *** high 0.2288 0.0049 0.1266 0.0033 *** perceived financial knowledge low 0.1091 0.0036 0.0760 0.0027 *** medium 0.1727 0.0043 0.1274 0.0033 *** high 0.6832 0.0053 0.7693 0.0042 *** financial literacy 2.5062 0.0164 3.2074 0.0140 *** financial fragility 2.7394 0.0190 2.0657 0.0180 *** *significance at 10% level; **significance at 5% level; ***significance at 1% level. 415m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 effect on seeking financial advice. while financial literacy is related positively to the demand for all types of financial advice, except for debt counseling, financial fragility decreases the demand for advice about savings/investment, insurance, and tax planning, but increases the demand for debt counseling. financial literacy alerts people to the value of financial advice in improving their well-being because they realize the complexity of financial topics and issues. however, financial literacy might be endogenous to the demand for advice. to test this potential endogeneity and revers causality, the article instruments for financial literacy using scores for the quality of public schools for 50 states and the district of columbia in 2012. the results of a wald test of exogeneity indicate endogeneity of financial literacy. therefore, it cannot be concluded that changes in financial literacy influence the demand for financial advice. on the other hand, financial difficulties such as overspending, lack of an emergency fund, and having a high level of debt discourage people from purchasing financial advice. in addition, a low perception of financial knowledge, which could proxy selfconfidence, has been found to decrease the probability of seeking financial advice. the correlation between a high perception of financial knowledge and financial literacy is found to be 0.26, which reflects a weak positive linear relation between these key variables. this finding reveals a lack of consistency between objective and subjective assessment of financial knowledge. table 5 summary statistics (financially illiterate vs. financially literate) financially illiterate financially literate mean standard error mean standard error dependent variables debt counseling 0.1078 0.0041 0.0799 0.0026 *** savings or investment advice 0.2036 0.0051 0.3387 0.0043 *** mortgage or loan advice 0.1523 0.0046 0.2327 0.0039 *** insurance advice 0.2459 0.0054 0.3379 0.0043 *** tax planning 0.1351 0.0044 0.2097 0.0037 *** independent variables annual income less than $15,000 0.2300 0.0052 0.0886 0.0027 *** $15,000 to less than $25,000 0.1748 0.0048 0.0901 0.0027 *** $25,000 to less than $35,000 0.1436 0.0044 0.0982 0.0028 *** $35,000 to less than $50,000 0.1450 0.0044 0.1482 0.0033 $50,000 to less than $75,000 0.1500 0.0045 0.2118 0.0038 *** $75,000 to less than $100,000 0.0723 0.0032 0.1418 0.0032 *** $100,000 to less than $150,000 0.0545 0.0029 0.1405 0.0032 *** $150,000 or more 0.0298 0.0022 0.0808 0.0024 *** risk-tolerance level low 0.3931 0.0061 0.3261 0.0043 *** medium 0.3700 0.0060 0.4814 0.0046 *** high 0.1741 0.0049 0.1749 0.0036 perceived financial knowledge low 0.1419 0.0044 0.0604 0.0023 *** medium 0.1827 0.0048 0.1276 0.0031 *** high 0.6118 0.0061 0.8012 0.0037 *** financial fragility 2.8809 0.0209 2.0738 0.0166 *** *significance at 10% level; **significance at 5% level; ***significance at 1% level. 416 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 6 fi na nc ia l ad vi ce pr ob it d eb t co un se lin g sa vi ng s/ in ve st m en t m or tg ag e/ lo an in su ra nc e t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) in de pe nd en t va ri ab le s g en de r (m al e) fe m al e � 0. 00 34 0. 00 44 0. 03 02 0. 00 63 ** * 0. 00 40 0. 00 59 0. 03 65 0. 00 69 ** * 0. 00 60 0. 00 57 r ac e (n on -w hi te ) w hi te � 0. 02 05 0. 00 48 ** * � 0. 00 17 0. 00 73 0. 02 04 0. 00 67 ** * � 0. 01 00 0. 00 78 � 0. 00 88 0. 00 64 a ge (6 5� ) 18 –2 4 0. 02 40 0. 01 04 ** 0. 01 55 0. 01 40 0. 13 43 0. 01 41 ** * 0. 04 76 0. 01 58 ** * 0. 03 95 0. 01 28 ** * 25 –3 4 0. 03 98 0. 00 94 ** * � 0. 04 10 0. 01 22 ** * 0. 15 00 0. 01 17 ** * 0. 03 98 0. 01 36 ** * 0. 01 11 0. 01 09 35 –4 4 0. 01 18 0. 00 88 � 0. 11 15 0. 01 19 ** * 0. 07 92 0. 01 15 ** * 0. 00 86 0. 01 34 � 0. 04 76 0. 01 06 ** * 45 –5 4 0. 00 85 0. 00 83 � 0. 09 97 0. 01 05 ** * 0. 03 94 0. 01 04 ** 0. 00 64 0. 01 20 � 0. 04 69 0. 00 95 ** * 55 –6 4 0. 01 80 0. 00 78 ** � 0. 03 91 0. 00 97 ** * 0. 03 06 0. 00 98 ** * 0. 01 44 0. 01 12 � 0. 02 02 0. 00 86 ** e du ca tio n le ve l (c ol le ge or m or e) h ig h sc ho ol or le ss � 0. 03 28 0. 00 59 ** * � 0. 09 35 0. 00 84 ** * � 0. 06 22 0. 00 79 ** * � 0. 07 05 0. 00 91 ** * � 0. 06 05 0. 00 76 ** * so m e co lle ge � 0. 01 58 0. 00 52 ** * � 0. 04 29 0. 00 75 ** * � 0. 01 33 0. 00 69 * � 0. 01 46 0. 00 82 * � 0. 03 22 0. 00 65 ** * m ar ita l st at us (m ar ri ed ) l iv in g w ith a pa rt ne r � 0. 00 43 0. 00 84 0. 00 86 0. 01 24 � 0. 01 01 0. 01 10 � 0. 01 10 0. 01 30 � 0. 01 96 0. 01 11 * si ng le 0. 00 24 0. 00 55 0. 00 90 0. 00 79 � 0. 04 10 0. 00 75 ** * � 0. 01 60 0. 00 85 * � 0. 01 51 0. 00 70 ** n um be r of ch ild re n (n o ch ild re n) o ne ch ild 0. 03 87 0. 00 67 ** * 0. 03 59 0. 00 99 ** * 0. 04 91 0. 00 90 ** * 0. 06 62 0. 01 05 ** * 0. 04 50 0. 00 87 ** * t w o ch ild re n 0. 04 57 0. 00 73 ** * 0. 03 54 0. 01 10 ** * 0. 05 42 0. 00 98 ** * 0. 07 97 0. 01 17 ** * 0. 04 33 0. 00 96 ** * t hr ee ch ild re n 0. 04 03 0. 00 94 ** * 0. 01 57 0. 01 51 0. 05 85 0. 01 32 ** * 0. 04 47 0. 01 55 ** * 0. 02 34 0. 01 34 * fo ur ch ild re n or m or e 0. 05 46 0. 01 13 ** * 0. 06 56 0. 01 90 ** * 0. 06 97 0. 01 64 ** * 0. 08 80 0. 01 96 ** * 0. 04 51 0. 01 64 ** * n o fin an ci al ly de pe nd en t ch ild re n 0. 00 53 0. 00 71 0. 01 73 0. 00 91 * 0. 02 09 0. 00 89 ** 0. 03 09 0. 01 01 ** * 0. 01 39 0. 00 85 a nn ua l in co m e (l es s th an $1 5, 00 0) $1 5, 00 0 to le ss th an $2 5, 00 0 0. 04 54 0. 00 89 ** * 0. 06 08 0. 01 42 ** * 0. 05 66 0. 01 41 ** * 0. 10 21 0. 01 44 ** * 0. 05 16 0. 01 37 ** * $2 5, 00 0 to le ss th an $3 5, 00 0 0. 05 04 0. 00 90 ** * 0. 07 45 0. 01 42 ** * 0. 08 25 0. 01 35 ** * 0. 10 31 0. 01 46 ** * 0. 06 85 0. 01 35 ** * $3 5, 00 0 to le ss th an $5 0, 00 0 0. 05 81 0. 00 86 ** * 0. 08 63 0. 01 35 ** * 0. 08 50 0. 01 30 ** * 0. 11 58 0. 01 40 ** * 0. 08 36 0. 01 27 ** * $5 0, 00 0 to le ss th an $7 5, 00 0 0. 06 16 0. 00 89 ** * 0. 11 63 0. 01 34 ** * 0. 11 51 0. 01 28 ** * 0. 13 11 0. 01 39 ** * 0. 10 68 0. 01 25 ** * $7 5, 00 0 to le ss th an $1 00 ,0 00 0. 06 96 0. 01 00 ** * 0. 13 20 0. 01 47 ** * 0. 13 00 0. 01 39 ** * 0. 11 67 0. 01 55 ** * 0. 12 53 0. 01 36 ** * $1 00 ,0 00 to le ss th an $1 50 ,0 00 0. 06 29 0. 01 12 ** * 0. 15 16 0. 01 52 ** * 0. 14 33 0. 01 46 ** * 0. 13 91 0. 01 63 ** * 0. 14 02 0. 01 40 ** * $1 50 ,0 00 or m or e 0. 05 19 0. 01 27 ** * 0. 17 51 0. 01 73 ** * 0. 16 82 0. 01 61 ** * 0. 15 75 0. 01 85 ** * 0. 18 66 0. 01 53 ** * (c on ti nu ed on ne xt pa ge ) 417m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 6 (c on tin ue d) d eb t co un se lin g sa vi ng s/ in ve st m en t m or tg ag e/ lo an in su ra nc e t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) r is kto le ra nc e le ve l (l ow ) m ed iu m 0. 00 92 0. 00 50 * 0. 11 11 0. 00 70 ** * 0. 03 95 0. 00 66 ** * 0. 07 38 0. 00 76 ** * 0. 05 55 0. 00 64 ** * h ig h 0. 04 99 0. 00 62 ** * 0. 16 85 0. 00 93 ** * 0. 06 55 0. 00 87 ** * 0. 11 87 0. 01 02 ** * 0. 10 46 0. 00 82 ** * pe rc ei ve d fin an ci al kn ow le dg e (h ig h) l ow � 0. 00 89 0. 00 73 � 0. 03 42 0. 01 33 ** * � 0. 05 16 0. 01 18 ** * � 0. 05 15 0. 01 31 ** * � 0. 03 67 0. 01 22 ** * m ed iu m � 0. 01 60 0. 00 62 ** * � 0. 01 90 0. 00 95 ** � 0. 01 39 0. 00 86 � 0. 03 36 0. 00 99 ** * � 0. 01 63 0. 00 85 * fi na nc ia l lit er ac y � 0. 00 34 0. 00 16 ** 0. 02 23 0. 00 24 ** * 0. 01 68 0. 00 22 ** * 0. 02 12 0. 00 26 ** * 0. 01 17 0. 00 21 ** * fi na nc ia l fr ag ili ty 0. 02 09 0. 00 16 ** * � 0. 03 86 0. 00 23 ** * 0. 00 79 0. 00 19 ** * � 0. 00 66 0. 00 25 ** * � 0. 01 46 0. 00 21 ** * b an kr up tc y 0. 18 03 0. 00 80 ** * h om eo w ne rs hi p 0. 12 34 0. 00 72 ** * t he fin an ci al ly ill ite ra te 0. 02 07 0. 00 83 ** 0. 03 28 0. 01 23 ** * 0. 02 37 0. 01 15 ** 0. 01 32 0. 01 33 0. 02 28 0. 01 10 ** t he yo un g 0. 01 57 0. 00 49 ** * 0. 00 99 0. 00 73 0. 06 65 0. 00 68 ** * 0. 02 11 0. 00 79 ** * 0. 02 55 0. 00 65 ** * in te ra ct io n va ri ab le s fe m al e* in co m e $1 5, 00 0 to le ss th an $2 5, 00 0 0. 08 56 0. 12 00 � 0. 20 89 0. 10 18 ** 0. 03 92 0. 11 18 � 0. 16 47 0. 09 07 * � 0. 07 62 0. 11 79 $2 5, 00 0 to le ss th an $3 5, 00 0 0. 06 24 0. 12 28 � 0. 07 53 0. 10 15 0. 17 87 0. 10 85 * � 0. 09 29 0. 09 16 0. 15 11 0. 11 61 $3 5, 00 0 to le ss th an $5 0, 00 0 0. 05 31 0. 11 34 � 0. 05 68 0. 09 48 0. 19 94 0. 10 09 ** � 0. 12 77 0. 08 56 0. 14 03 0. 10 60 $5 0, 00 0 to le ss th an $7 5, 00 0 � 0. 02 72 0. 11 35 � 0. 01 43 0. 09 08 0. 08 80 0. 09 65 � 0. 01 34 0. 08 23 0. 13 91 0. 10 07 $7 5, 00 0 to le ss th an $1 00 ,0 00 � 0. 13 01 0. 12 65 � 0. 19 14 0. 09 78 ** 0. 18 62 0. 10 35 * � 0. 10 85 0. 09 01 0. 13 67 0. 10 76 $1 00 ,0 00 to le ss th an $1 50 ,0 00 � 0. 11 95 0. 13 96 � 0. 11 25 0. 10 01 0. 16 06 0. 10 68 � 0. 17 05 0. 09 33 * 0. 25 60 0. 10 95 ** $1 50 ,0 00 or m or e � 0. 01 74 0. 16 44 � 0. 01 67 0. 11 27 0. 23 36 0. 11 80 ** � 0. 05 85 0. 10 51 0. 27 87 0. 11 96 ** fe m al e* r is kt ol er an ce m ed iu m � 0. 02 26 0. 06 98 � 0. 02 08 0. 05 12 0. 06 97 0. 05 31 � 0. 01 72 0. 04 81 0. 03 99 0. 05 67 h ig h 0. 05 95 0. 08 58 � 0. 10 09 0. 06 54 � 0. 07 47 0. 06 80 � 0. 05 59 0. 06 23 � 0. 05 39 0. 07 04 fe m al e* pe rc ei ve d fin an ci al kn ow le dg e l ow � 0. 20 68 0. 10 63 * � 0. 09 29 0. 09 59 0. 04 71 0. 09 43 � 0. 07 88 0. 08 43 � 0. 09 00 0. 10 71 m ed iu m � 0. 13 14 0. 08 43 � 0. 01 76 0. 06 81 � 0. 01 71 0. 07 01 � 0. 04 99 0. 06 28 0. 07 42 0. 07 61 fe m al e* fi na nc ia l fr ag ili ty le ve l 1 0. 11 67 0. 12 03 0. 09 63 0. 06 43 0. 05 53 0. 07 00 0. 07 60 0. 06 46 0. 08 93 0. 06 87 2 0. 11 46 0. 12 07 � 0. 02 78 0. 06 78 0. 08 64 0. 07 32 0. 01 43 0. 06 73 0. 02 99 0. 07 35 3 0. 25 59 0. 11 75 ** � 0. 08 28 0. 07 30 0. 13 49 0. 07 75 * � 0. 01 02 0. 07 03 0. 01 11 0. 07 93 4 0. 38 61 0. 11 97 ** * � 0. 04 58 0. 07 86 0. 19 59 0. 08 17 ** 0. 07 22 0. 07 34 0. 16 32 0. 08 81 * (c on ti nu ed on ne xt pa ge ) 418 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 6 (c on tin ue d) d eb t co un se lin g sa vi ng s/ in ve st m en t m or tg ag e/ lo an in su ra nc e t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) 5 0. 37 06 0. 12 73 ** * � 0. 02 75 0. 09 59 0. 18 64 0. 09 52 ** 0. 11 04 0. 08 48 0. 12 54 0. 10 21 6 0. 10 15 0. 19 43 0. 06 44 0. 18 53 0. 43 24 0. 16 22 ** * � 0. 10 30 0. 13 99 0. 13 10 0. 18 15 y ou ng *i nc om e $1 5, 00 0 to le ss th an $2 5, 00 0 � 0. 18 42 0. 11 69 � 0. 25 33 0. 10 02 ** 0. 06 15 0. 10 86 0. 17 70 0. 08 75 ** 0. 14 44 0. 11 86 $2 5, 00 0 to le ss th an $3 5, 00 0 � 0. 23 26 0. 11 65 ** � 0. 31 87 0. 10 07 ** * � 0. 03 19 0. 10 60 0. 11 15 0. 08 84 � 0. 07 12 0. 11 68 $3 5, 00 0 to le ss th an $5 0, 00 0 � 0. 10 88 0. 11 04 � 0. 30 20 0. 09 48 ** * � 0. 15 47 0. 09 95 0. 12 33 0. 08 32 � 0. 08 58 0. 10 84 $5 0, 00 0 to le ss th an $7 5, 00 0 � 0. 00 26 0. 11 23 � 0. 30 25 0. 09 14 ** * � 0. 08 34 0. 09 66 0. 17 68 0. 08 08 ** � 0. 21 67 0. 10 45 ** $7 5, 00 0 to le ss th an $1 00 ,0 00 � 0. 09 36 0. 12 57 � 0. 49 64 0. 09 83 ** * � 0. 10 58 0. 10 40 0. 04 76 0. 08 91 � 0. 27 27 0. 11 11 ** $1 00 ,0 00 to le ss th an $1 50 ,0 00 � 0. 13 54 0. 13 76 � 0. 47 52 0. 10 25 ** * � 0. 10 58 0. 10 83 0. 06 60 0. 09 37 � 0. 31 64 0. 11 54 ** * $1 50 ,0 00 or m or e � 0. 31 43 0. 16 05 ** � 0. 61 88 0. 11 86 ** * � 0. 23 46 0. 12 26 * � 0. 02 58 0. 10 91 � 0. 45 51 0. 12 77 ** * y ou ng *r is kt ol er an ce m ed iu m 0. 14 43 0. 06 78 ** � 0. 01 23 0. 05 18 0. 04 71 0. 05 27 0. 03 56 0. 04 79 0. 08 76 0. 05 67 h ig h 0. 14 75 0. 08 60 * 0. 18 07 0. 06 64 ** * 0. 09 66 0. 06 92 0. 14 16 0. 06 30 ** 0. 26 27 0. 07 10 ** * y ou ng *p er ce iv ed fin an ci al kn ow le dg e l ow � 0. 13 70 0. 10 08 � 0. 08 82 0. 09 05 � 0. 05 97 0. 08 99 � 0. 18 49 0. 07 87 ** � 0. 01 82 0. 10 02 m ed iu m � 0. 08 31 0. 08 18 � 0. 15 66 0. 06 63 ** � 0. 10 36 0. 06 82 � 0. 14 96 0. 06 11 ** � 0. 03 94 0. 07 39 y ou ng *f in an ci al fr ag ili ty le ve l 1 � 0. 06 71 0. 11 73 0. 16 53 0. 06 91 ** � 0. 07 93 0. 07 35 0. 11 16 0. 06 91 0. 15 41 0. 07 39 ** 2 � 0. 20 43 0. 11 62 * 0. 31 97 0. 07 21 ** * � 0. 27 43 0. 07 60 ** * 0. 04 65 0. 07 14 0. 21 09 0. 07 85 ** * 3 � 0. 38 15 0. 11 52 ** * 0. 36 56 0. 07 64 ** * � 0. 22 88 0. 08 04 ** * 0. 03 68 0. 07 36 0. 18 20 0. 08 34 ** 4 � 0. 40 57 0. 11 74 ** * 0. 45 13 0. 08 25 ** * � 0. 32 56 0. 08 28 ** * � 0. 07 93 0. 07 56 0. 28 11 0. 08 95 ** * 5 � 0. 55 29 0. 12 50 ** * 0. 37 21 0. 09 64 ** * � 0. 33 78 0. 09 52 ** * � 0. 05 11 0. 08 52 0. 20 28 0. 10 29 ** 6 � 0. 45 97 0. 17 21 ** * 0. 36 57 0. 16 49 ** � 0. 24 03 0. 14 88 � 0. 06 66 0. 13 22 0. 08 82 0. 17 75 il lit er at e* in co m e $1 5, 00 0 to le ss th an $2 5, 00 0 0. 27 35 0. 11 90 ** 0. 09 51 0. 10 11 � 0. 04 90 0. 11 02 0. 15 76 0. 08 93 * � 0. 04 73 0. 11 93 $2 5, 00 0 to le ss th an $3 5, 00 0 0. 15 27 0. 12 03 0. 00 18 0. 10 21 � 0. 13 51 0. 10 77 0. 04 14 0. 09 07 � 0. 03 24 0. 11 84 $3 5, 00 0 to le ss th an $5 0, 00 0 0. 06 81 0. 11 38 � 0. 02 67 0. 09 75 � 0. 15 42 0. 10 30 0. 13 03 0. 08 69 � 0. 05 76 0. 11 01 $5 0, 00 0 to le ss th an $7 5, 00 0 0. 11 31 0. 11 61 0. 00 86 0. 09 45 � 0. 13 86 0. 10 05 0. 10 08 0. 08 52 0. 00 59 0. 10 65 $7 5, 00 0 to le ss th an $1 00 ,0 00 0. 12 77 0. 13 20 0. 12 47 0. 10 71 � 0. 27 74 0. 11 23 ** 0. 06 30 0. 09 83 � 0. 05 41 0. 11 85 $1 00 ,0 00 to le ss th an $1 50 ,0 00 0. 39 37 0. 14 59 ** * 0. 06 97 0. 11 36 � 0. 25 38 0. 12 25 ** 0. 08 47 0. 10 81 � 0. 20 23 0. 12 67 $1 50 ,0 00 or m or e 0. 45 75 0. 16 81 ** * 0. 28 66 0. 13 72 ** � 0. 25 36 0. 13 99 * 0. 29 96 0. 12 61 ** � 0. 11 63 0. 14 22 (c on ti nu ed on ne xt pa ge ) 419m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 6 (c on tin ue d) d eb t co un se lin g sa vi ng s/ in ve st m en t m or tg ag e/ lo an in su ra nc e t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) il lit er at e* r is kto le ra nc e m ed iu m � 0. 02 01 0. 06 98 0. 08 78 0. 05 49 0. 04 57 0. 05 71 0. 01 13 0. 05 03 0. 05 02 0. 06 13 h ig h 0. 22 39 0. 08 38 ** * 0. 23 48 0. 06 86 ** * 0. 30 93 0. 07 22 ** * 0. 19 00 0. 06 54 ** * 0. 25 23 0. 07 35 ** * il lit er at e* pe rc ei ve d fin an ci al kn ow le dg e l ow � 0. 01 87 0. 10 37 0. 08 57 0. 09 40 0. 16 39 0. 09 12 * 0. 14 99 0. 08 13 * 0. 00 97 0. 10 44 m ed iu m � 0. 08 28 0. 08 56 0. 19 43 0. 07 08 ** * 0. 03 04 0. 07 29 0. 04 04 0. 06 45 0. 03 70 0. 07 98 il lit er at e* fi na nc ia l fr ag ili ty le ve l 1 � 0. 10 71 0. 12 74 � 0. 05 88 0. 07 89 � 0. 05 75 0. 08 96 � 0. 02 38 0. 07 98 0. 02 54 0. 08 60 2 � 0. 12 30 0. 12 29 0. 02 92 0. 07 90 0. 01 66 0. 08 77 0. 05 28 0. 07 86 0. 13 16 0. 08 75 3 � 0. 38 04 0. 12 30 ** * � 0. 00 44 0. 08 35 � 0. 18 14 0. 09 17 ** 0. 06 49 0. 08 06 � 0. 04 12 0. 09 20 4 � 0. 44 84 0. 12 44 ** * � 0. 08 20 0. 08 86 � 0. 11 69 0. 09 40 0. 03 61 0. 08 21 � 0. 06 39 0. 09 93 5 � 0. 39 45 0. 13 12 ** * 0. 02 22 0. 10 19 � 0. 17 58 0. 10 42 * 0. 01 09 0. 09 07 � 0. 17 11 0. 11 09 6 � 0. 07 88 0. 18 15 0. 02 66 0. 17 64 � 0. 09 26 0. 15 70 � 0. 15 04 0. 13 80 � 0. 03 80 0. 18 11 n um be r of ob se rv at io ns 25 ,5 09 25 ,5 09 25 ,5 09 25 ,5 09 25 ,5 09 t he yo un g va ri ab le is in cl ud ed in a se pa ra te re gr es si on w ith ou t th e ag e ra ng es to av oi d co lli ne ar ity . *s ig ni fic an ce at 10 % le ve l; ** si gn ifi ca nc e at 5% le ve l; ** *s ig ni fic an ce at 1% le ve l. 420 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 to examine additional reasons that can explain the demand for different types of advice, the article uses as a proxy for liquidity constraints the difficulty to cover expenses and pay one’s bills. the t test results in table 7 indicate that respondents who do not experience liquidity constraints show a higher demand for advice about saving/investment, mortgages/ loans, and insurance compared with those with a liquidity problem. furthermore, homeownership has been used as a proxy for socioeconomic status and financial stability. the t test results in table 8 show significant differences between homeowners and non-homeowners. the percentage of homeowners who seek financial advice is higher than that for nonhomeowners across all types of advice except debt counseling. in addition, table 9 shows that seeking debt counseling by respondents who declared bankruptcy is significantly different from those who did not experience bankruptcy. to test the presence of significant differences in financial advice seeking behavior by females, the young, and the financially illiterate, the regression model in table 6 uses interaction terms between those groups and the factors of interest over the whole sample. the results for females indicate that financial fragility only is related positively to seeking debt counseling and advice about mortgages/loans for respondents who experience at least three signs of financial difficulty. for the young group, the results show a negative relation between income and seeking advice about savings/investment, and tax planning. high risk tolerance is related positively to seeking advice about savings/investment, insurance, and tax planning. however, financial fragility is related negatively to debt counseling and advice about mortgages, and positively to advice about savings/investment, and tax planning. for the financially illiterate group, high risk tolerance is related positively to seeking all types of financial advice, while financial fragility has a negative association to seeking debt counseling. table 7 t-test of means for financial advice (liquidity constraint) type of advice have difficulty no difficulty mean standard error mean standard error debt counseling 0.1186 0.0033 0.0523 0.0027 *** savings/investment 0.2201 0.0041 0.3785 0.0054 *** mortgages/loans 0.1934 0.0039 0.2137 0.0046 *** tax planning 0.1512 0.0036 0.2220 0.0046 *** *significance at 10% level; **significance at 5% level; ***significance at 1% level. table 8 t-test of means for financial advice (home ownership) type of advice yes no mean standard error mean standard error savings/investment 0.3668 0.0046 0.1782 0.0046 *** mortgages/loans 0.2671 0.0042 0.1130 0.0037 *** insurance 0.3533 0.0045 0.2338 0.0050 *** tax planning 0.2386 0.0041 0.1028 0.0036 *** *significance at 10% level; **significance at 5% level; ***significance at 1% level. 421m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 7.1. subsample results: females versus males the female variable is found to be significant for seeking financial advice about savings/ investment, and insurance only. the probit regression results for the female subsample are provided in table 10. income and risk tolerance are related positively, while a low perception of financial knowledge and financial fragility are related negatively to seeking both types of financial advice. these findings for the female subsample are identical to the findings for the male subsample in table 11, except for the effect of financial knowledge. a low perception of financial knowledge has a greater effect on the demand for financial advice for females compared with males. therefore, the characteristics that influence the demand for financial advice for females and males appear to be similar and gender differences do not distinguish the consumption of financial advice between these two groups. 7.2. subsample results: young versus old the young (aged 18–44)10 variable is found to be significant for seeking all types of financial advice, except savings/investment. table 12 reports the estimates of four probit regression models for the young subsample. income and risk tolerance are related positively to seeking all types of financial advice. a low perception of financial knowledge decreases the probability of seeking financial advice about mortgages/loans, insurance, and tax planning, while financial fragility is related positively to seeking debt counseling and negatively to seeking advice about insurance and tax planning. these findings are similar to those for the old group in table 13, except for the low perception of financial knowledge, which does not appear as significant for the old compared with the young subsample. age classification does not explain the consumption of financial advice. 7.3. subsample results: financially illiterate versus financially literate the financially illiterate variable is found to be significant for seeking all types of financial advice except insurance. table 14 provides the estimates of four probit regression models for the financially illiterate subsample. income and risk tolerance are related positively to seeking the four types of financial advice. a low perception of financial knowledge is related negatively to seeking advice about mortgages/loans and tax planning and appears to have no significance on seeking advice about debt and savings/ investment. while financial fragility increases the probability of seeking debt counseling, it decreases the probability of seeking advice about savings/investment and tax table 9 t-test of means for financial advice (bankruptcy) type of advice bankrupt no bankruptcy mean standard error mean standard error debt counseling 0.4873 0.0208 0.0760 0.0021 *** *significance at 10% level; **significance at 5% level; ***significance at 1% level. 422 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 planning. these findings are similar to the results for the financially literate subsample in table 15. therefore, the factors affecting the demand for financial advice show no significant differences based on the financial literacy level only. 8. conclusion this article uses the 2012 nfcs to investigate the correlates of seeking five types of financial advice: debt counseling, savings/investment, mortgages/loans, insurance, and tax table 10 financial advice probit (female) independent variables savings/investment insurance marg. effects (se) marg. effects (se) race (non-white) white �0.0024 0.0096 �0.0101 0.0103 age (65�) 18–24 �0.0375 0.0184 ** 0.0049 0.0210 25–34 �0.1167 0.0170 *** �0.0246 0.0191 35–44 �0.1710 0.0167 *** �0.0269 0.0186 45–54 �0.1207 0.0145 *** �0.0160 0.0166 55–64 �0.0570 0.0134 *** 0.0052 0.0156 education level (college or more) high school or less �0.1095 0.0113 *** �0.0786 0.0125 *** some college �0.0429 0.0105 *** �0.0023 0.0116 marital status (married) living with a partner 0.0224 0.0164 �0.0122 0.0171 single 0.0218 0.0106 ** �0.0002 0.0114 number of children (no children) one child 0.0175 0.0133 0.0570 0.0141 *** two children 0.0232 0.0151 0.0674 0.0157 *** three children �0.0093 0.0204 0.0493 0.0206 ** four children or more 0.0711 0.0240 *** 0.0804 0.0248 *** no financially dependent children �0.0035 0.0126 0.0176 0.0141 annual income (less than $15,000) $15,000 to less than $25,000 0.0302 0.0181 * 0.0819 0.0183 *** $25,000 to less than $35,000 0.0610 0.0186 *** 0.0909 0.0187 *** $35,000 to less than $50,000 0.0758 0.0177 *** 0.0996 0.0185 *** $50,000 to less than $75,000 0.1137 0.0175 *** 0.1318 0.0184 *** $75,000 to less than $100,000 0.1058 0.0196 *** 0.1035 0.0209 *** $100,000 to less than $150,000 0.1300 0.0205 *** 0.1087 0.0223 *** $150,000 or more 0.1728 0.0233 *** 0.1509 0.0251 *** risk-tolerance level (low) medium 0.1154 0.0092 *** 0.0774 0.0099 *** high 0.1762 0.0138 *** 0.1341 0.0151 *** perceived financial knowledge (high) low �0.0418 0.0168 *** �0.0563 0.0167 *** medium �0.0188 0.0124 �0.0403 0.0128 *** financial literacy 0.0311 0.0033 *** 0.0312 0.0035 *** financial fragility �0.0407 0.0031 *** �0.0060 0.0034 * number of observations 14,127 14,127 *significance at 10% level; **significance at 5% level; ***significance at 1% level. 423m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 planning. although only 24% of respondents are satisfied with their personal financial condition, the use of financial advice is within the range of 9–30% of the u.s. population, depending on the type of advice. while 73% of respondents assessed themselves as financially knowledgeable, only 16% were able to answer five basic financial literacy questions correctly. this discrepancy raises the complex question of why individuals are reluctant to seek professional financial advice. the analysis of the multivariate results reveals a consistent effect of key factors on the demand for the five types of financial advice and no significant differences have been found among the subsamples, which are defined by gender, age, and financial literacy. income and table 11 financial advice probit (male) independent variables savings/investment insurance marg. effects (se) marg. effects (se) race (non-white) white �0.0004 0.0108 �0.0073 0.0117 age (65�) 18–24 0.0756 0.0212 *** 0.1025 0.0238 *** 25–34 0.0312 0.0175 * 0.1054 0.0196 *** 35–44 �0.0501 0.0172 *** 0.0502 0.0193 *** 45–54 �0.0691 0.0153 *** 0.0392 0.0173 ** 55–64 �0.0174 0.0141 0.0281 0.0160 * education level (college or more) high school or less �0.0808 0.0124 *** �0.0615 0.0134 *** some college �0.0439 0.0106 *** �0.0251 0.0116 ** marital status (married) living with a partner �0.0021 0.0184 �0.0074 0.0198 single �0.0159 0.0120 �0.0452 0.0130 *** number of children (no children) one child 0.0519 0.0147 *** 0.0703 0.0159 *** two children 0.0476 0.0160 *** 0.0900 0.0174 *** three children 0.0429 0.0221 * 0.0365 0.0236 four children or more 0.0671 0.0311 ** 0.1042 0.0320 *** no financially dependent children 0.0350 0.0133 *** 0.0401 0.0147 *** annual income (less than $15,000) $15,000 to less than $25,000 0.0920 0.0222 *** 0.1259 0.0228 *** $25,000 to less than $35,000 0.0832 0.0220 *** 0.1146 0.0231 *** $35,000 to less than $50,000 0.0916 0.0207 *** 0.1327 0.0215 *** $50,000 to less than $75,000 0.1096 0.0204 *** 0.1275 0.0213 *** $75,000 to less than $100,000 0.1446 0.0218 *** 0.1260 0.0233 *** $100,000 to less than $150,000 0.1582 0.0227 *** 0.1582 0.0241 *** $150,000 or more 0.1624 0.0257 *** 0.1565 0.0272 *** risk-tolerance level (low) medium 0.1066 0.0109 *** 0.0697 0.0119 *** high 0.1590 0.0128 *** 0.1046 0.0142 *** perceived financial knowledge (high) low �0.0234 0.0214 �0.0453 0.0214 ** medium �0.0192 0.0148 �0.0240 0.0157 financial literacy 0.0151 0.0036 *** 0.0124 0.0039 *** financial fragility �0.0373 0.0035 *** �0.0080 0.0037 ** number of observations 11,382 11,382 *significance at 10% level; **significance at 5% level; ***significance at 1% level. 424 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 12 fi na nc ia l ad vi ce pr ob it (t he yo un g 18 – 44 ) in de pe nd en t va ri ab le s d eb t co un se lin g m or tg ag e/ lo an in su ra nc e t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) g en de r (m al e) fe m al e � 0. 01 57 0. 00 69 ** � 0. 01 48 0. 00 98 0. 00 20 0. 01 07 � 0. 02 81 0. 00 88 ** * r ac e (n on -w hi te ) w hi te � 0. 01 49 0. 00 68 ** 0. 00 98 0. 01 00 � 0. 02 20 0. 01 07 ** � 0. 03 32 0. 00 89 ** * e du ca tio n le ve l (c ol le ge or m or e) h ig h sc ho ol or le ss � 0. 03 76 0. 00 91 ** * � 0. 07 87 0. 01 29 ** * � 0. 06 46 0. 01 39 ** * � 0. 04 59 0. 01 17 ** * so m e co lle ge � 0. 01 80 0. 00 80 ** � 0. 02 78 0. 01 15 ** � 0. 02 68 0. 01 27 ** � 0. 03 14 0. 01 04 ** * m ar ita l st at us (m ar ri ed ) l iv in g w ith a pa rt ne r � 0. 00 75 0. 01 13 � 0. 02 46 0. 01 54 � 0. 01 24 0. 01 68 � 0. 01 36 0. 01 41 si ng le � 0. 00 03 0. 00 90 � 0. 08 74 0. 01 26 ** * � 0. 03 62 0. 01 37 ** * � 0. 03 01 0. 01 12 ** * n um be r of ch ild re n (n o ch ild re n) o ne ch ild 0. 05 44 0. 00 96 ** * 0. 03 26 0. 01 33 ** 0. 07 14 0. 01 46 ** * 0. 04 18 0. 01 20 ** * t w o ch ild re n 0. 06 10 0. 01 02 ** * 0. 03 64 0. 01 39 ** * 0. 08 19 0. 01 54 ** * 0. 03 28 0. 01 26 ** * t hr ee ch ild re n 0. 04 74 0. 01 24 ** * 0. 03 53 0. 01 80 ** 0. 03 93 0. 01 95 ** 0. 02 38 0. 01 66 fo ur ch ild re n or m or e 0. 06 08 0. 01 48 ** * 0. 04 52 0. 02 16 ** 0. 06 47 0. 02 37 ** * 0. 02 61 0. 01 95 n o fin an ci al ly de pe nd en t ch ild re n 0. 00 93 0. 01 62 � 0. 00 40 0. 02 36 0. 01 04 0. 02 38 0. 00 27 0. 02 15 a nn ua l in co m e (l es s th an $1 5, 00 0) $1 5, 00 0 to le ss th an $2 5, 00 0 0. 04 30 0. 01 30 ** * 0. 07 01 0. 02 03 ** * 0. 12 07 0. 01 99 ** * 0. 06 19 0. 01 79 ** * $2 5, 00 0 to le ss th an $3 5, 00 0 0. 04 54 0. 01 33 ** * 0. 09 20 0. 01 97 ** * 0. 10 74 0. 02 02 ** * 0. 05 51 0. 01 78 ** * $3 5, 00 0 to le ss th an $5 0, 00 0 0. 06 28 0. 01 26 ** * 0. 08 06 0. 01 88 ** * 0. 12 58 0. 01 95 ** * 0. 07 28 0. 01 69 ** * $5 0, 00 0 to le ss th an $7 5, 00 0 0. 07 20 0. 01 27 ** * 0. 13 05 0. 01 85 ** * 0. 14 51 0. 01 92 ** * 0. 07 69 0. 01 67 ** * $7 5, 00 0 to le ss th an $1 00 ,0 00 0. 07 22 0. 01 42 ** * 0. 14 29 0. 02 07 ** * 0. 10 31 0. 02 17 ** * 0. 08 72 0. 01 86 ** * $1 00 ,0 00 to le ss th an $1 50 ,0 00 0. 06 23 0. 01 67 ** * 0. 15 88 0. 02 26 ** * 0. 12 99 0. 02 41 ** * 0. 09 00 0. 02 03 ** * $1 50 ,0 00 or m or e 0. 05 21 0. 01 87 ** * 0. 16 89 0. 02 58 ** * 0. 14 03 0. 02 82 ** * 0. 11 75 0. 02 27 ** * r is kto le ra nc e le ve l (l ow ) m ed iu m 0. 02 23 0. 00 79 ** * 0. 05 43 0. 01 12 ** * 0. 07 96 0. 01 20 ** * 0. 06 45 0. 01 06 ** * h ig h 0. 05 79 0. 00 89 ** * 0. 09 50 0. 01 35 ** * 0. 13 33 0. 01 46 ** * 0. 12 92 0. 01 20 ** * pe rc ei ve d fin an ci al kn ow le dg e (h ig h) l ow � 0. 01 66 0. 01 16 � 0. 06 61 0. 01 79 ** * � 0. 06 86 0. 01 87 ** * � 0. 03 79 0. 01 68 ** m ed iu m � 0. 01 98 0. 00 92 ** � 0. 03 12 0. 01 32 ** � 0. 05 02 0. 01 42 ** * � 0. 01 84 0. 01 21 fi na nc ia l lit er ac y � 0. 00 73 0. 00 24 ** * 0. 01 63 0. 00 35 ** * 0. 01 87 0. 00 38 ** * 0. 00 23 0. 00 31 fi na nc ia l fr ag ili ty 0. 01 63 0. 00 24 ** * � 0. 00 55 0. 00 35 � 0. 00 81 0. 00 37 ** � 0. 01 25 0. 00 32 ** * n um be r of ob se rv at io ns 11 ,1 35 11 ,1 35 11 ,1 35 11 ,1 35 *s ig ni fic an ce at 10 % le ve l; ** si gn ifi ca nc e at 5% le ve l; ** *s ig ni fic an ce at 1% le ve l. 425m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 13 fi na nc ia l ad vi ce pr ob it (t he ol d 45 � ) in de pe nd en t va ri ab le s d eb t co un se lin g m or tg ag e/ lo an in su ra nc e t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar ge (s e ) g en de r (m al e) fe m al e 0. 00 39 0. 00 59 0. 01 61 0. 00 71 ** 0. 06 67 0. 00 89 ** * 0. 03 30 0. 00 66 ** * r ac e (n on -w hi te ) w hi te � 0. 02 99 0. 00 69 ** * 0. 03 02 0. 00 90 ** * 0. 00 24 0. 01 12 0. 02 09 0. 00 85 ** e du ca tio n le ve l (c ol le ge or m or e) h ig h sc ho ol or le ss � 0. 02 36 0. 00 75 ** * � 0. 04 64 0. 00 95 ** * � 0. 07 00 0. 01 19 ** * � 0. 06 67 0. 00 89 ** * so m e co lle ge � 0. 01 06 0. 00 70 � 0. 00 18 0. 00 82 0. 00 01 0. 01 05 � 0. 02 61 0. 00 75 ** * m ar ita l st at us (m ar ri ed ) l iv in g w ith a pa rt ne r 0. 01 05 0. 01 29 0. 01 01 0. 01 62 � 0. 00 62 0. 02 07 � 0. 02 76 0. 01 65 * si ng le 0. 00 37 0. 00 70 � 0. 00 31 0. 00 89 0. 00 14 0. 01 09 0. 00 86 0. 00 82 n um be r of ch ild re n (n o ch ild re n) o ne ch ild 0. 01 33 0. 00 95 0. 05 39 0. 01 19 ** * 0. 04 71 0. 01 49 ** * 0. 02 03 0. 01 13 * t w o ch ild re n 0. 01 58 0. 01 09 0. 05 40 0. 01 39 ** * 0. 05 42 0. 01 77 ** * 0. 01 38 0. 01 33 t hr ee ch ild re n 0. 03 52 0. 01 52 ** 0. 06 90 0. 02 05 ** * 0. 03 07 0. 02 60 � 0. 04 40 0. 02 00 ** fo ur ch ild re n or m or e 0. 05 38 0. 01 91 ** * 0. 08 28 0. 02 64 ** * 0. 12 25 0. 03 58 ** * 0. 01 53 0. 02 81 n o fin an ci al ly de pe nd en t ch ild re n � 0. 00 56 0. 00 76 0. 02 43 0. 00 95 ** 0. 02 36 0. 01 16 ** 0. 00 60 0. 00 89 a nn ua l in co m e (l es s th an $1 5, 00 0) $1 5, 00 0 to le ss th an $2 5, 00 0 0. 05 92 0. 01 22 ** * 0. 05 61 0. 01 91 ** * 0. 06 16 0. 02 05 ** * 0. 03 26 0. 01 95 * $2 5, 00 0 to le ss th an $3 5, 00 0 0. 07 44 0. 01 22 ** * 0. 10 67 0. 01 87 ** * 0. 07 77 0. 02 09 ** * 0. 07 64 0. 01 90 ** * $3 5, 00 0 to le ss th an $5 0, 00 0 0. 07 70 0. 01 19 ** * 0. 13 20 0. 01 80 ** * 0. 08 74 0. 02 01 ** * 0. 09 31 0. 01 79 ** * $5 0, 00 0 to le ss th an $7 5, 00 0 0. 07 14 0. 01 26 ** * 0. 16 38 0. 01 79 ** * 0. 09 74 0. 02 02 ** * 0. 12 78 0. 01 76 ** * $7 5, 00 0 to le ss th an $1 00 ,0 00 0. 08 41 0. 01 43 ** * 0. 19 09 0. 01 91 ** * 0. 10 50 0. 02 23 ** * 0. 14 95 0. 01 88 ** * $1 00 ,0 00 to le ss th an $1 50 ,0 00 0. 08 09 0. 01 51 ** * 0. 20 79 0. 01 95 ** * 0. 12 43 0. 02 27 ** * 0. 16 61 0. 01 90 ** * $1 50 ,0 00 or m or e 0. 08 65 0. 01 73 ** * 0. 25 01 0. 02 11 ** * 0. 15 06 0. 02 53 ** * 0. 21 59 0. 02 03 ** * r is kto le ra nc e le ve l (l ow ) m ed iu m 0. 00 20 0. 00 66 0. 03 19 0. 00 78 ** * 0. 07 35 0. 00 97 ** * 0. 03 91 0. 00 74 ** * h ig h 0. 03 53 0. 00 94 ** * 0. 04 97 0. 01 14 ** * 0. 09 24 0. 01 43 ** * 0. 06 50 0. 01 04 ** * pe rc ei ve d fin an ci al kn ow le dg e (h ig h) l ow � 0. 00 51 0. 01 04 � 0. 05 16 0. 01 49 ** * � 0. 02 11 0. 01 82 � 0. 02 71 0. 01 52 * m ed iu m � 0. 00 98 0. 00 82 � 0. 00 81 0. 01 11 � 0. 01 11 0. 01 37 � 0. 01 06 0. 01 10 fi na nc ia l lit er ac y � 0. 00 25 0. 00 23 0. 01 88 0. 00 29 ** * 0. 02 52 0. 00 35 ** * 0. 01 92 0. 00 27 ** * fi na nc ia l fr ag ili ty 0. 02 46 0. 00 21 ** * 0. 00 92 0. 00 26 ** * � 0. 00 64 0. 00 33 * � 0. 01 49 0. 00 26 ** * n um be r of ob se rv at io ns 14 ,3 56 14 ,3 56 14 ,3 56 14 ,3 56 *s ig ni fic an ce at 10 % le ve l; ** si gn ifi ca nc e at 5% le ve l; ** *s ig ni fic an ce at 1% le ve l. 426 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 14 fi na nc ia l ad vi ce pr ob it (t he fin an ci al ly ill ite ra te ) in de pe nd en t va ri ab le s d eb t co un se lin g sa vi ng s/ in ve st m en t m or tg ag e/ lo an t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) g en de r (m al e) fe m al e � 0. 01 86 0. 00 85 ** � 0. 01 48 0. 01 14 � 0. 01 94 0. 01 04 * � 0. 02 27 0. 00 99 ** r ac e (n on -w hi te ) w hi te � 0. 03 01 0. 00 86 ** * � 0. 01 84 0. 01 15 0. 00 49 0. 01 06 � 0. 02 10 0. 01 00 ** a ge (6 5� ) 18 –2 4 0. 01 38 0. 01 94 0. 03 17 0. 02 26 0. 10 17 0. 02 34 ** * 0. 06 91 0. 02 14 ** * 25 –3 4 0. 02 37 0. 01 84 � 0. 00 91 0. 02 22 0. 10 95 0. 02 20 ** * 0. 04 54 0. 02 06 ** 35 –4 4 � 0. 00 84 0. 01 92 � 0. 09 44 0. 02 29 ** * 0. 04 53 0. 02 29 ** � 0. 01 49 0. 02 12 45 –5 4 0. 00 12 0. 01 83 � 0. 07 55 0. 02 11 ** * 0. 02 90 0. 02 18 � 0. 02 11 0. 02 00 55 –6 4 0. 00 23 0. 01 83 � 0. 01 53 0. 02 11 0. 02 98 0. 02 21 � 0. 01 24 0. 02 02 e du ca tio n le ve l (c ol le ge or m or e) h ig h sc ho ol or le ss � 0. 05 63 0. 01 07 ** * � 0. 11 03 0. 01 45 ** * � 0. 08 04 0. 01 35 ** * � 0. 07 09 0. 01 30 ** * so m e co lle ge � 0. 04 53 0. 01 06 ** * � 0. 04 89 0. 01 45 ** * � 0. 03 36 0. 01 31 ** � 0. 04 67 0. 01 26 ** * m ar ita l st at us (m ar ri ed ) l iv in g w ith a pa rt ne r � 0. 01 15 0. 01 41 0. 03 09 0. 01 89 � 0. 01 48 0. 01 66 0. 00 13 0. 01 61 si ng le � 0. 00 72 0. 01 03 0. 01 51 0. 01 37 � 0. 03 22 0. 01 22 ** * � 0. 02 16 0. 01 17 * n um be r of ch ild re n (n o ch ild re n) o ne ch ild 0. 04 62 0. 01 19 ** * 0. 05 31 0. 01 61 ** * 0. 05 06 0. 01 41 ** * 0. 04 68 0. 01 38 ** * t w o ch ild re n 0. 04 13 0. 01 36 ** * 0. 04 27 0. 01 85 ** 0. 06 31 0. 01 59 ** * 0. 03 19 0. 01 58 ** t hr ee ch ild re n 0. 05 27 0. 01 67 ** * � 0. 02 13 0. 02 49 0. 04 64 0. 02 05 ** 0. 03 23 0. 02 03 fo ur ch ild re n or m or e 0. 04 23 0. 02 03 ** 0. 03 31 0. 02 84 0. 04 94 0. 02 41 ** 0. 03 06 0. 02 41 n o fin an ci al ly de pe nd en t ch ild re n 0. 00 39 0. 01 39 0. 01 92 0. 01 71 0. 02 68 0. 01 68 0. 01 59 0. 01 69 a nn ua l in co m e (l es s th an $1 5, 00 0) $1 5, 00 0 to le ss th an $2 5, 00 0 0. 07 75 0. 01 40 ** * 0. 06 60 0. 01 89 ** * 0. 06 22 0. 01 82 ** * 0. 04 61 0. 01 78 ** * $2 5, 00 0 to le ss th an $3 5, 00 0 0. 07 26 0. 01 45 ** * 0. 06 71 0. 02 02 ** * 0. 08 42 0. 01 80 ** * 0. 06 27 0. 01 83 ** * $3 5, 00 0 to le ss th an $5 0, 00 0 0. 07 65 0. 01 45 ** * 0. 07 28 0. 01 98 ** * 0. 08 47 0. 01 81 ** * 0. 07 32 0. 01 74 ** * $5 0, 00 0 to le ss th an $7 5, 00 0 0. 07 84 0. 01 53 ** * 0. 11 16 0. 01 99 ** * 0. 12 43 0. 01 82 ** * 0. 09 92 0. 01 74 ** * $7 5, 00 0 to le ss th an $1 00 ,0 00 0. 09 12 0. 01 83 ** * 0. 13 62 0. 02 37 ** * 0. 12 68 0. 02 15 ** * 0. 10 50 0. 02 07 ** * $1 00 ,0 00 to le ss th an $1 50 ,0 00 0. 10 80 0. 02 10 ** * 0. 15 54 0. 02 56 ** * 0. 14 31 0. 02 46 ** * 0. 09 79 0. 02 32 ** * $1 50 ,0 00 or m or e 0. 11 09 0. 02 30 ** * 0. 21 52 0. 03 14 ** * 0. 16 36 0. 02 78 ** * 0. 14 39 0. 02 60 ** * r is kto le ra nc e le ve l (l ow ) m ed iu m 0. 01 35 0. 00 95 0. 10 92 0. 01 22 ** * 0. 04 38 0. 01 13 ** * 0. 05 17 0. 01 13 ** * h ig h 0. 07 04 0. 01 14 ** * 0. 18 07 0. 01 53 ** * 0. 10 00 0. 01 42 ** * 0. 11 24 0. 01 34 ** * pe rc ei ve d fin an ci al kn ow le dg e (h ig h) l ow � 0. 01 57 0. 01 27 � 0. 02 92 0. 01 86 � 0. 03 85 0. 01 58 ** � 0. 03 38 0. 01 70 ** m ed iu m � 0. 02 74 0. 01 13 ** 0. 00 13 0. 01 50 � 0. 01 71 0. 01 36 � 0. 01 00 0. 01 36 fi na nc ia l fr ag ili ty 0. 01 76 0. 00 30 ** * � 0. 03 96 0. 00 39 ** * � 0. 00 34 0. 00 36 � 0. 01 81 0. 00 35 ** * n um be r of ob se rv at io ns 8, 92 1 8, 92 1 8, 92 1 8, 92 1 *s ig ni fic an ce at 10 % le ve l; ** si gn ifi ca nc e at 5% le ve l; ** *s ig ni fic an ce at 1% le ve l. 427m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 t ab le 15 fi na nc ia l ad vi ce pr ob it (t he fin an ci al ly lit er at e) in de pe nd en t va ri ab le s d eb t co un se lin g sa vi ng s/ in ve st m en t m or tg ag e/ lo an t ax pl an ni ng m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) m ar g. ef fe ct s (s e ) g en de r (m al e) fe m al e 0. 00 01 0. 00 53 0. 04 73 0. 00 76 ** * 0. 01 31 0. 00 70 * 0. 01 83 0. 00 66 ** * r ac e (n on -w hi te ) w hi te � 0. 01 40 0. 00 61 ** 0. 01 20 0. 00 95 0. 03 48 0. 00 85 ** * 0. 00 22 0. 00 82 a ge (6 5� ) 18 –2 4 0. 03 28 0. 01 49 ** � 0. 01 54 0. 01 99 0. 05 55 0. 01 97 ** * 0. 00 53 0. 01 84 25 –3 4 0. 03 92 0. 01 13 ** * � 0. 07 06 0. 01 53 ** * 0. 11 11 0. 01 43 ** * � 0. 01 19 0. 01 33 35 –4 4 0. 02 16 0. 01 07 ** � 0. 12 34 0. 01 45 ** * 0. 05 67 0. 01 36 ** * � 0. 05 93 0. 01 27 ** * 45 –5 4 0. 00 88 0. 01 02 � 0. 10 99 0. 01 26 ** * 0. 01 89 0. 01 22 � 0. 05 51 0. 01 11 ** * 55 –6 4 0. 02 95 0. 00 93 ** * � 0. 04 67 0. 01 13 ** * 0. 02 12 0. 01 11 * � 0. 02 19 0. 00 98 ** e du ca tio n le ve l (c ol le ge or m or e) h ig h sc ho ol or le ss � 0. 02 35 0. 00 72 ** * � 0. 09 76 0. 01 03 ** * � 0. 06 32 0. 00 96 ** * � 0. 06 16 0. 00 92 ** * so m e co lle ge � 0. 00 46 0. 00 62 � 0. 05 04 0. 00 89 ** * � 0. 01 08 0. 00 82 � 0. 03 00 0. 00 77 ** * m ar ita l st at us (m ar ri ed ) l iv in g w ith a pa rt ne r 0. 00 59 0. 01 10 � 0. 00 26 0. 01 63 � 0. 00 09 0. 01 46 � 0. 03 33 0. 01 49 ** si ng le 0. 00 80 0. 00 67 0. 00 45 0. 00 99 � 0. 04 47 0. 00 94 ** * � 0. 00 92 0. 00 87 n um be r of ch ild re n (n o ch ild re n) o ne ch ild 0. 03 49 0. 00 85 ** * 0. 02 21 0. 01 25 * 0. 04 39 0. 01 15 ** * 0. 04 06 0. 01 11 ** * t w o ch ild re n 0. 05 06 0. 00 89 ** * 0. 02 96 0. 01 38 ** 0. 04 55 0. 01 23 ** * 0. 04 58 0. 01 21 ** * t hr ee ch ild re n 0. 03 23 0. 01 16 ** * 0. 03 78 0. 01 96 * 0. 06 09 0. 01 72 ** * 0. 00 66 0. 01 70 fo ur ch ild re n or m or e 0. 06 87 0. 01 43 ** * 0. 08 96 0. 02 64 ** * 0. 08 43 0. 02 25 ** * 0. 04 99 0. 02 22 ** n o fin an ci al ly de pe nd en t ch ild re n 0. 00 75 0. 00 84 0. 01 35 0. 01 12 0. 01 72 0. 01 07 0. 01 13 0. 00 99 a nn ua l in co m e (l es s th an $1 5, 00 0) $1 5, 00 0 to le ss th an $2 5, 00 0 0. 02 53 0. 01 24 ** 0. 05 43 0. 02 08 ** * 0. 06 77 0. 02 05 ** * 0. 05 29 0. 02 05 ** * $2 5, 00 0 to le ss th an $3 5, 00 0 0. 04 40 0. 01 24 ** * 0. 08 09 0. 02 05 ** * 0. 11 26 0. 01 97 ** * 0. 07 01 0. 01 97 ** * $3 5, 00 0 to le ss th an $5 0, 00 0 0. 05 74 0. 01 16 ** * 0. 09 76 0. 01 91 ** * 0. 12 79 0. 01 86 ** * 0. 09 16 0. 01 87 ** * $5 0, 00 0 to le ss th an $7 5, 00 0 0. 05 96 0. 01 18 ** * 0. 12 36 0. 01 89 ** * 0. 16 90 0. 01 82 ** * 0. 11 53 0. 01 83 ** * $7 5, 00 0 to le ss th an $1 00 ,0 00 0. 06 61 0. 01 28 ** * 0. 13 75 0. 01 99 ** * 0. 20 05 0. 01 93 ** * 0. 14 14 0. 01 92 ** * $1 00 ,0 00 to le ss th an $1 50 ,0 00 0. 04 96 0. 01 40 ** * 0. 15 92 0. 02 06 ** * 0. 21 76 0. 01 98 ** * 0. 16 40 0. 01 96 ** * $1 50 ,0 00 or m or e 0. 04 51 0. 01 62 ** * 0. 17 44 0. 02 26 ** * 0. 25 32 0. 02 14 ** * 0. 21 24 0. 02 09 ** * r is kto le ra nc e le ve l (l ow ) m ed iu m 0. 00 87 0. 00 61 0. 11 62 0. 00 87 ** * 0. 03 89 0. 00 81 ** * 0. 05 67 0. 00 78 ** * h ig h 0. 03 56 0. 00 78 ** * 0. 16 61 0. 01 19 ** * 0. 05 73 0. 01 10 ** * 0. 09 69 0. 01 04 ** * pe rc ei ve d fin an ci al kn ow le dg e (h ig h) l ow � 0. 00 67 0. 01 04 � 0. 04 47 0. 01 89 ** � 0. 08 90 0. 01 65 ** * � 0. 03 62 0. 01 65 ** m ed iu m � 0. 00 91 0. 00 75 � 0. 03 92 0. 01 22 ** * � 0. 02 44 0. 01 11 ** � 0. 02 19 0. 01 08 ** fi na nc ia l fr ag ili ty 0. 02 22 0. 00 19 ** * � 0. 04 07 0. 00 30 ** * 0. 00 30 0. 00 28 � 0. 01 40 0. 00 26 ** * n um be r of ob se rv at io ns 16 ,5 88 16 ,5 88 16 ,5 88 16 ,5 88 *s ig ni fic an ce at 10 % le ve l; ** si gn ifi ca nc e at 5% le ve l; ** *s ig ni fic an ce at 1% le ve l. 428 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 risk tolerance are related positively to the demand for all types of financial advice and more greatly affect the probability of seeking advice than do other variables. the finding for income is consistent with the article’s expectation that those with relatively high incomes might have a sufficient level of financial sophistication to seek financial advice in the areas of savings/investment, mortgages/loans, insurance, and tax planning. the positive relation between income and debt counseling was not expected, however. this relation could be interpreted as the result of a tendency of those who see increases in income to accumulate debt to fund a lifestyle that exceeds their income level. risk tolerance plays a significant role and demonstrates a strong positive effect on the demand for all types of financial advice. however, the subjective assessment of risk tolerance in the survey raises a question about the accuracy and reliability of this measure in reflecting respondents’ actual risk tolerance and their understanding of its significance for their financial investments. a low perception of financial knowledge decreases the demand for all types of financial advice except debt counseling. this finding does not support the expected negative relation between perceived financial knowledge and the demand for financial advice and contradicts some findings in prior research. this subjective assessment of financial knowledge might become a psychological barrier that decreases the demand for financial advice because respondents are not confident in their ability to assess financial products and monitor agency relationships. on the other hand, financial fragility, has been found to be related negatively to seeking financial advice about savings/investment, insurance, and tax planning, and related positively to seeking debt counseling. people who struggle with their expenses and are not able to save for retirement might not have the luxury to think about investment or tax planning. financial stress would draw their attention away from long-term plans toward immediate short-term concerns. the survey question about seeking the five types of financial advice refers to this behavior in the past five years and does not necessarily indicate that respondents never seek professional financial advice or use alternative sources such as their social network. furthermore, because the survey focuses on individual responses, the household’s behavior may not be observed accurately. if the spouse, for example, seeks financial advice, then the other spouse may not indicate seeking such advice. understanding the demand for professional financial advice requires an examination of the effect that salient and hidden fees have on people’s decisions to contract financial advisers. in addition, future research has to examine the determinants of trust because respondents lack the ability to assess service quality and evaluate outcomes. financial advice is a mosaic of services, and several factors influence the demand for different types of advice. the similarity of payment-reward trade-off (i.e., fee payment for investment return) makes financial advice a unique service arrangement because individuals’ mode of payment is that exact commodity that they aim to preserve and grow to smooth their consumption power over their life cycle. notes 1 the assets & opportunities scorecard is a comprehensive look at americans’ financial security based on 130 outcome and policy measures. the scorecard enables states 429m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 to benchmark their outcomes and policies against other states in five areas: financial assets & income, businesses & jobs, housing & homeownership, health care, and education. http://assetsandopportunity.org/scorecard 2 self-concealment refers to the psychological tendency to keep perceived negative or intimate personal information secret. older homeowners might conceal their financial difficulty to protect their social status and perceived financial competency. 3 the survey’s questionnaire asks for detailed race and ethnicity information but the dataset provides information regarding white and non-white only. 4 the subjective risk tolerance levels as per the 10-point scale are as follows: ● 1–3: low risk tolerance ● 4–7: medium risk tolerance ● 8–10: high risk tolerance 5 the financial knowledge levels as per the seven-point scale are as follows: ● 1–3: low financial knowledge ● 4: medium financial knowledge ● 5–7: high financial knowledge 6 the survey question for the dependent variables is: in the last five years, have you asked for any advice from a financial professional about any of the following? debt counseling—savings or investment—taking out a mortgage or a loan—insurance of any type—tax planning. 7 wealth is a preferable factor in the decision to purchase financial advice, but is not included in the 2012 nfcs. therefore, income has been used to proxy wealth in the model as the level of earning power might indicate a level of financial sophistication and capability to seek professional financial advice. 8 risk tolerance has been aggregated into three levels because the survey question measures it on a scale from 1-very low to 10-very high and the responses are almost evenly distributed across the scale. 9 subjective assessment of financial knowledge has been aggregated into three levels because the survey question measures it on a scale from 1 to 7. 10 the classification of younger respondents as those aged 18–44 follows the methodology in cfp 2015 stress awareness month survey report. references bajtelsmit, v. l., & bernasek, a. (1996). why do women invest differently than men? journal of financial counseling and planning, 7, 1–10. brooks, j., wiedrich, k., sims, l., & rice, s. (2015). excluded from the financial mainstream: how the economic recovery is bypassing millions of americans. (available at http://assetsandopportunity.org/score card) burton, d. (1995). women and financial services: some directions for future research. international journal of bank marketing, 13, 21–28. calcagno, r., & monticone, c. (2014). financial literacy and the demand for financial advice. journal of banking and finance, 50, 363–380. 430 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 chang, m. (2005). with a little help from my friends (and my financial planner). social froces, 83, 1469–1497. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. cummings, b. f., & james, r. n. (2014). determinants of seeking financial advice among older adults. (available at ssrn 243468) donkers, b., & van soest, a. (1999). subjective measures of household preferences and financial decisions. journal of economic psychology, 20, 613–642. elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217. finke, m., huston, s., & winchester, d. (2011). financial advice: who pays. journal of financial counseling and planning, 22, 18–26. fischer, r., & gerhardt, r. (2007). investment mistakes of individual investors and the impact of financial advice. in 20th australasian finance & banking conference, sydney, australia. goldsmith, r. e., & goldsmith, e. b. (2006). the effects of investment education on gender differences in financial knowledge. journal of personal finance, 5, 55–69. grable, j. e., & joo, s. (1999). financial help-seeking behavior: theory and implications. financial counseling and planning, 10, 13–24. grable, j. e., & joo, s. (2001). a further examination of financial help-seeking behavior. financial counseling and planning, 12, 55–74. hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking and finance, 36, 509–524. hanna, s. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. hanna, s., & lindamood, s. (2010). quantifying the economic benefits of personal financial planning. financial services review, 19, 111–127. hastings, j. s., & mitchell, o. s. (2011). how financial literacy and impatience shape retirement wealth and investment behaviors (no. w16740). national bureau of economic research. heo, w., grable, j. e., & chatterjee, s. (2013). life insurance consumption as a function of wealth change. financial service review, 22, 389–404. hira, t. k., & mugenda, o. (2000). gender differences in financial perceptions, behaviors and satisfaction. fpa journal, 200, 1–7. hung, a., yoong, j., & brown, e. (2012). empowering women through financial awareness and education. oecd working papers on finance, insurance and private pensions, (14), 1. inderst, r., & ottaviani, m. (2000). misselling through agents. the american economic review, 99, 883–908. inderst, r., & ottaviani, m. (2012). financial advice. journal of economic literature, 50, 494–512. kramer, m. m. (2012). financial advice and individual investor portfolio performance. financial management, 41, 395–428. kramer, m. m. (2016). financial literacy, confidence and financial advice seeking. journal of economic behavior & organization, 131, 198–217. laibson, d., gabaix, x., driscoll, j., & agarwal, s. (2007). the age of reason: financial decisions over the lifecycle. national bureau of economic research working paper 13191. letkiewicz, j., robinson, c., & domian, d. (2016). behavioral and wealth considerations for seeking professional financial planning help. financial services review, 25, 105–126. levin, j., milgrom, p., & segal, i. (2004). consumer theory. (available at http://web.stanford.edu) loibl, c., & hira, t. k. (2006). a workplace and gender-related perspective on financial planning information sources and knowledge outcomes. financial services review, 15, 21. lin, q., & lee, j. (2004). consumers’ information search when making investment decisions. financial services review, 13, 319–332. lusardi, a. (2008). financial literacy: an essential tool for informed consumer choice? (no. w14084). national bureau of economic research. lusardi, a. (2012). numeracy, financial literacy, and financial decision-making (no. w17821). national bureau of economic research. 431m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 lusardi, a., & mitchell, o. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. lusardi, a., mitchell, o., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lusardi, a., & mitchell, o. s. (2009). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness (no. w15350). national bureau of economic research. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and overindebtedness (no. w14808). national bureau of economic research. morris, g. a., & meyer, r. l. (1993). women and financial services in developing countries: a review of the literature. (available at https://kb.osu.edu) orc. (2015). stress awareness month survey report. (available at http://www.cfp.net) robb, c. a, babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. salter, j. r., harness, n., & chatterjee, s. (2010). utilization of financial advisors by affluent retirees. financial services review, 19, 245–263. scott, j. k., & finke, m. s. (2013). the demand for disability insurance. financial services review, 22, 1–12. seay, m. c., kim, k. t., & heckman, s. j. (2016). exploring the demand for retirement planning advice: the role of financial literacy. financial services review, 25, 331–350. simms, k. (2014). investor profiles: meaningful differences in women’s use of investment advice? financial services review, 23, 273–286. tang, n., & lachance, m.-e. (2012). financial advice: what about low income consumers? journal of personal finance, 11, 121–158. vanderhei, j., & copeland, c. (2015). retirement confidence survey. retreived from http://www.ebri.org. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. winchester, d. d., & huston, s. j. (2015). all financial advice for the middle class is not equal. journal of consumer policy, (april), 247–264. 432 m.h. alyousif, c.m. kalenkoski / financial services review 26 (2017) 405–432 from the editor this issue contains volume 27 issue 2 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “what do financial planning organizations communicate to stakeholders and consumers? an empirical narrative analysis” is coauthored by wookjae at south dakota state university, heo, narang park at university of georgia, robin henager at whitworth university, and john e. grable at university of georgia. the authors examine how the financial organizations present relevant information to stakeholders and consumers and what differences exist between what organizations intend to deliver and what consumers and stakeholders perceive from the communication channels. using text mining techniques, their results show that financial planning organizations were successfully addressing their own position in the financial planning profession; however, they were failing to communicate specifics about their value to consumers. the second article “an examination of the federal employee retirement system (fers) survivor annuity benefit” is coauthored by kevin davis at usaf academy, steve p. fraser at florida gulf coast university, and william w. jennings at usaf academy. the authors present a study on the federal employees retirement system (fers) which provides survivor annuity benefits for employees who forfeit a portion of their annuity as a premium. they develop a monte carlo simulation to describe the distributions and implied internal rates of return for fers annuitants who elect a joint and survivor annuity. their analysis suggests that the survivor benefit program is quite lucrative for most male retirees. in contrast, the program is less rewarding for female retirees, especially if younger than their spouse. for many female retirees, the program actually produces a negative return. the third article, “a framework for analyzing defined benefit pension insurance: the survivor benefit plan for veterans,” is coauthored by william w. jennings at united states air force academy, jeffrey c. merrell at university of colorado, boulder, thomas c. o’malley at united states air force academy, and brian c. payne at university of colorado, colorado springs. in this study, the authors present a framework for making a pension insurance method decision for us military veterans’ survivor benefit plan (sbp). they find that federal government subsidies generate a positive expected net payout for financial services review 27 (2018) v–vi 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. sbp. while insurance outcomes are typically skewed, they show the asymmetry of sbp outcomes. they describe a scenario where 5% of participants receive 60% of benefits. an alternative financial planning approach incorporates private insurance and investments and often bests the sbp, when utilizing actuarially-correct life expectancy, moral hazard, taxes, and individual financial needs. the fourth article, “who uses robo-advisory services, and who does not?” is coauthored by martha fulk, john e. grable, kimberly watkins, and michelle kruger all at university of georgia. the authors compare the demographic, attitudinal, and behavioral characteristics of us consumers in their current and expected use of robo-advisory services, traditional financial planning services, or a combination of the two. their findings show a difference between those who used robo-advisory services and those who used traditional financial planning services. overall, those who used a traditional financial planner were older and reported higher levels of net worth. additionally, they find those who used traditional financial planning services reported a larger percentage of their total net worth from an inheritance, and users of robo-advisory services generally (a) had lower income, (b) had lower net worth, (c) had received no or less inheritance, and (d) were less impulsive financially. the final article, “risk and uncertainty in style rotation” is authored by timothy a. krause at penn state, behrend. the author states the cboe® vix (volatility) index has been established as an indicator of style returns because increases in this “fear index” lead to outperformance of “value” vs “growth” stocks. in this paper, the author introduces the concept of “uncertainty” as an additional indicator of returns to value, as measured by the cboe® vvix (“volatility of volatility”). his research shows that increases in expected volatility lead to short-term positive returns to value, while increases in uncertainty lead to negative short-term returns to value. he also observes that these are especially strong during economic downturns and following decreases in the vix index. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 27 (2018) v–vi the perpetual growth model and the cost of computational efficiency: rounding errors or wild distortions? morris g. danielsona, jean l. hecka,* adepartment of finance, saint joseph’s university, 5600 city avenue, philadelphia, pa 19131, usa abstract the constant growth model (gordon, 1962) plays an important role in the stock selection process for individual investors, in part, because of its computational simplicity. however, value estimates from the model can be highly dependent on cash flows to be received in the distant future. if future events might constrain a firm’s growth or lead to its demise, the unadjusted gordon model can substantially overstate value. because the model is less likely to misstate value for low-growth, high-payout firms, the ironic implication is that the model is most useful when its ability to value growth is needed least. © 2014 academy of financial services. all rights reserved. jel classification: g30 keywords: constant growth model; perpetual growth; stock valuation 1. introduction “nothing lasts forever,” or so the old saying goes. when valuing assets, however, individual investors often ignore this adage and assume cash flows will grow at a constant rate in perpetuity. the gordon (1962) growth model uses this assumption and is a first-cut tool for estimating stock prices and calculating terminal values in two-stage growth models, or discounted cash flow analyses. bradley and jarrell (2008) note that this model “… is taught in all top-tier business schools and used widely throughout the financial community. it is found in virtually all graduate-level corporate finance textbooks and valuation manuals.” * corresponding author. tel.: �1-610-660-3148; fax: �1-610-660-1986. e-mail address: jean.heck@sju.edu (j. heck) financial services review 23 (2014) 189–206 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. recent surveys (e.g., block, 1999; demirakos, strong, and walker, 2004; dukes, peng, and english, 2006; imam, barker, and clubb, 2008) suggest that value estimates from dividend discount models (such as the gordon model) are used to inform the recommendations of many analysts. the constant-growth assumption greatly simplifies the calculation of the present value of an infinite dividend stream. indeed, one of the main reasons why the gordon model plays such a prominent role in valuation theory and practice is because it is so easy to use. however, value estimates from the perpetual-growth-model can rely heavily on cash flows to be received in the distant future. shaffer (2006) notes that when the required return is 4% and the growth rate is 3%, 62% of the value estimate is accounted for by cash flows to be received more than 50 years in the future. danielson, heck, and shaffer (2008) observe that almost 70% of an asset’s value stems from cash flows to be received in years 11 to infinity if the discount rate is 8% and the growth rate is 4%. from a mathematical perspective, a present value (today) of $20 is worth $20, regardless of whether the cash flow supporting this present value will be received in five years or 500 years. as long as cash flow estimates five and 500 years in the future are equally reliable, the timing of the future cash flows is not a cause for concern. from a practical standpoint though, it is decidedly easier to forecast future events (and cash flows) five years out, than to predict what will happen in 500 years. in the short term, consumer preferences are likely to evolve in subtle ways, current patents will remain in force, and technological innovations may already be in the works (e.g., new medications in the clinical trial process), making their implications at least somewhat identifiable. over longer periods of time, market leaders can be replaced (e.g., sears and k-mart by wal-mart) and entire product markets can become obsolete (e.g., buggy whips and vcr machines). in response to this concern, numerous valuation models allow the assumed growth rate to decrease (or change) over time. examples include miller and modigliani (1961), holt, (1962), mao (1966), fielitz and muller (1985), gordon and gordon (1997), danielson (1998), and o’brien (2003). more recently, shaffer (2006) extends the gordon model to include a constant, annual probability of permanent failure. both of these approaches (allowing for decreasing growth or permanent failure) effectively reduce the portion of an estimated present value accounted for by cash flows to be received far in the future. however, the gordon model maintains computational advantages over even the simplest of these alternative models, accounting for its ongoing popularity. nevertheless, if future events might constrain a firm’s growth rate, or lead to its eventual demise, the use of the unadjusted constant growth model can overstate an asset’s value. the goal of this article is to quantify the size of these potential valuation errors. are these errors modest (i.e., rounding differences), in which case the benefits of the model’s computational simplicity would outweigh the potential costs created by its imprecise value estimates? or, are the valuation errors large enough to materially distort investment decisions? the analytical results in this article suggest that price estimates from the constant growth model can overstate a stock’s intrinsic value by a sizeable amount—in some cases the valuation errors can be two or three times the underlying intrinsic value! the potential overstatement increases as the firm’s dividend yield decreases, shifting a greater portion of the expected cash flow into later years. these results do not imply that the constant growth 190 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 model should be abandoned as a valuation tool. instead, the goal of this article is to promote a better understanding of the limitations of the model, and to place guardrails on its use. in particular, the constant growth model is most useful when the firm faces a low default probability, when only a small portion of the growth will be the result of positive net present value investments, and when the firm’s dividend yield is sufficiently large. in all other cases, value estimates obtained from the constant growth model have the potential to significantly overstate value. the ironic implication is that the perpetual growth model is most useful when it is needed least. that is, when the model is used to value low-growth, high-payout firms. 2. the constant-growth model: an overview in the constant-growth model, a firm’s stock price (or an asset’s value) is a function of the firm’s future dividends (i.e., cash dividends or stock repurchases); the key assumption is that the dividends will increase at a constant annual rate, forever. to begin, we define each variable in real terms: d1 is the dividend expected to be paid next year, stated in current dollars; r is the required (real) rate of return; and g is the (real) perpetual-growth rate. using these definitions, the model can be written as eq. (1). p0 � d1 r � g (1) in future years, a firm’s cash flows can grow as the result of new investments or inflation. section 2.1 rewrites eq. (1) to focus on growth from new investments when all inputs are stated in real terms. section 2.2 extends the model further, to allow the firm’s cash flow stream to increase with inflation. 2.1. growth and new investments eq. (1) is often expanded to calculate a stock price as a function of the amount invested in new projects and the return on these investments (see, e.g., brealey and myers, 2003). to do this, the variable e1 is defined as the perpetual earnings stream expected to be generated by the current operations (stated in current dollars), the plowback rate, b, is the portion of the earnings reinvested each year, and rn is the economic (not accounting) real return on new investments. using these additional definitions, eq. (1) becomes eq. (2). p0 � d1 r � g � �1 � b� e1 r � brn (2) when rn exceeds r, each new project is expected to have a positive net present value and p0 will increase with the reinvestment rate b.1 if rn equals r, eq. (2) simplifies to p0 � e1/r and value does not depend on the reinvestment rate b. in this case, the firm’s dividend still increases each year at the rate g � brn. however, this dividend growth simply compensates 191m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 investors for deferring dividends through the reinvestment process; the growth will not increase the estimated value p0. 2.2. growth and inflation to convert eq. (2) into nominal terms, the following definitions are used: h � the annual inflation rate r*n � the nominal return on new investments � (1 � rn)(1 � h) – 1 r* � the discount rate in nominal terms � (1 � r)(1 � h) – 1 e*1 � next year’s nominal expected earnings � (1 � h)e1 g* � the nominal growth rate � (1 – b)h � b r*n the nominal growth rate, g*, has two terms on the right-hand side to allow the firm’s entire cash flow stream (including amounts currently paid as dividends) to increase with inflation. the first term captures the growth (because of inflation) of the portion of the cash flow stream currently paid as a dividend. the second term is the growth created through the reinvestment process; the inflation rate h is embedded in the nominal investment return r*n. bradley and jarrell (2008) also define the nominal growth rate in this way (but with different notation). when these definitions are substituted into the constant growth model, the inflation terms ultimately cancel out and the model simplifies to eq. (2), with r, rn, and g defined as real rates.2 eq. (3) outlines this process. p0 � �1 � b� e*1 r* � g* � �1 � b��1 � h� e1 ��1 � r��1 � h� � 1� � ��1 � b�h � b��1 � rn��1 � h� � 1�� � �1 � b� e1 r � brn (3) bradley and jarrell (2008) show that when the constant growth model is not adjusted properly to incorporate nominal discount and growth rates, the model will produce incorrect value estimates. in particular, if the discount rate is stated in nominal terms (i.e., the discount rate is r*), but the growth rate is defined simply as the product of the reinvestment rate b and r*n, rather than as g* (as defined above), the constant growth model can understate an asset’s value.3 however, this result does not mean that the constant growth model produces “conservative” value estimates. instead, the analysis in bradley and jarrell (2008) simply means that when the model is applied incorrectly (and inputs are not defined consistently in either nominal or real terms) the model’s value estimates will be wrong (i.e., garbage in, garbage out). because the constant growth model produces identical value estimates using properly defined real or nominal inputs, and because the model’s notation is simpler using real interest rates, the analysis in the remainder of this article defines all valuation inputs in real terms. 192 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 3. value creation and the timing of future cash flows as noted by shaffer (2006), danielson, heck, and shaffer (2008), and others, a substantial portion of the value estimates obtained from the constant growth model can be attributed to cash flows that may not be received for decades. to calculate the percentage of a perpetualgrowth value estimate attributed to cash flows after year t, first note that the present value of a finite (t-period), growing dividend stream (i.e., the dividends will grow at a constant rate for t periods before dropping to $0) can be written as follows (welch, 2009): p0 � d1 r � g �1 � �1 � g 1 � r� t� (4) then, subtract eq. (4) from eq. (1) and divide the result by eq. (1). this yields eq. (5), which quantifies the portion of an asset’s value created by cash flows to be received after year t. pv�t � 1 to ��/pv�0 to �� � �1 � g 1 � r� t (5) table 1 uses eq. (5) to calculate the percentage of an asset’s estimated value—using eq. (1)—that is expected to be received after year t, for three discount rates (7, 10, and 13%) and a range of growth rates. for each combination of r and g, the table lists the portion of the value estimate created by cash flows received after years 10, 20, 30, 40, 50, and 100. in addition, the table lists the macaulay (1938) duration for each combination of r and g. although r and g can be (and typically are) estimated individually, payne and finch (1999) point out that the key determinant of p0 in eq. (1) is the difference between r and g. for example, if d1 is estimated to be $1 and the difference between r and g is 5%, the estimated stock price is $20 regardless of whether r � 7% and g � 2%, or r � 13% and g � 8%. in each case, the implied dividend yield (d/p) is 5%.4 table 1 reveals that the portion of a value estimate attributable to cash flows after a specified year is also closely related to the difference r – g (� d/p), but this relation is not exact. for example, when r � 7% and g � 2%, 38.4% (9.1%) of the asset’s value is created after year 20 (year 50); when r � 13% and g � 8%, 40.4% (10.4%) of the value is expected to be realized after year 20 (year 50). as the difference between r and g becomes smaller, a progressively larger portion of the value estimate will depend on cash flows from the out years. for example, when r – g is four percentage points (d/p � 4%), over 20% of the estimated value will be realized after year 40. when r – g is two percentage points (d/p � 2%), more than twice as much (i.e., nearly 50%) of the asset’s value will be received after year 40. as r and g converge (and d/p decreases toward zero), value estimates from the gordon model become highly dependent on cash flows to be received in the distant future. when r – g � 1%, almost 70% of the value will be realized after year 40, and �40% will be received after year 100. the macaulay duration in this case exceeds 100. when r – g � 0.5%, over 60% of the estimated value will be received after year 100 and the macaulay duration exceeds 200! 193m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 only when the dividend yield is large, can the majority of the estimated value be linked to cash flows within a foreseeable time horizon. for example, assume that g � 0 and d/p � r. if r � 13% (r – g � d/p � 0.13), over 90% of the value will be received over the next 20 years; if r � 10% (r – g � d/p � 0.10), over 90% of the value will be received during the next 30 years; and when r � 7% (r – g � d/p � 0.07), over 90% of the value will be received in the next 40 years. 4. by how much can the perpetual growth model overstate value? table 1 demonstrates that perpetual-growth-model value estimates can be highly dependent on the present value of cash flows that will not be received for decades (i.e., 20, 50, 100, table 1 percent of value created after year t panel a (r � 7%) g � 0 0.02 0.03 0.04 0.05 0.06 0.065 t � 0 100% 100% 100% 100% 100% 100% 100% 10 50.8% 62.0% 68.3% 75.2% 82.8% 91.0% 95.4% 20 25.8% 38.4% 46.7% 56.6% 68.6% 82.9% 91.1% 30 13.1% 23.8% 31.9% 42.6% 56.8% 75.5% 86.9% 40 6.7% 14.7% 21.8% 32.1% 47.0% 68.7% 82.9% 50 3.4% 9.1% 14.9% 24.1% 38.9% 62.5% 79.1% 100 0.1% 0.8% 2.2% 5.8% 15.2% 39.1% 62.6% duration 15.3 21.4 26.8 35.7 53.5 107.0 214.0 panel b (r � 10%) g � 0 0.05 0.06 0.07 0.08 0.09 0.095 t � 0 100% 100% 100% 100% 100% 100% 100% 10 38.6% 62.8% 69.0% 75.8% 83.2% 91.3% 95.5% 20 14.9% 39.4% 47.7% 57.5% 69.3% 83.3% 91.3% 30 5.7% 24.8% 32.9% 43.6% 57.7% 76.0% 87.2% 40 2.2% 15.6% 22.7% 33.1% 48.0% 69.4% 83.3% 50 0.9% 9.8% 15.7% 25.1% 40.0% 63.3% 79.6% 100 0.0% 1.0% 2.5% 6.3% 16.0% 40.1% 63.4% duration 11.0 22.0 27.5 36.7 55.0 110.0 220.0 panel c (r � 13%) g � 0 0.03 0.08 0.09 0.11 0.12 0.125 t � 0 100% 100% 100% 100% 100% 100% 100% 10 29.5% 39.6% 63.8% 69.7% 83.6% 91.5% 95.7% 20 8.7% 15.7% 40.4% 48.6% 70.0% 83.7% 91.5% 30 2.6% 6.2% 25.7% 33.9% 58.5% 76.6% 87.5% 40 0.8% 2.5% 16.4% 23.7% 49.0% 70.1% 83.7% 50 0.2% 1.0% 10.4% 16.5% 40.9% 64.1% 80.1% 100 0.0% 0.0% 1.1% 2.7% 16.8% 41.1% 64.2% duration 8.7 11.3 22.6 28.3 56.5 113.0 226.0 notes. this table lists the percentage of a perpetual-growth model value estimate attributed to cash flows received after year t (where t � 0, 10, 20, and so forth) calculated using eq. (5). the calculations use required returns of 7% (panel a), 10% (panel b) and 13% (panel c), and a range of assumed growth rates, g. for each combination or r and g, the table also lists the macaulay duration of the value estimate. 194 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 or more years in the future). this fact would not be a cause for concern if the future cash flows were certain to be paid. however, economic conditions can change dramatically over the course of any 20, 50, or 100 year period; the survival of no firm is guaranteed. indeed, the experience of the past 100 years highlights the challenges individual firms face to remain in business (let alone to maintain a positive growth rate) over an extended period. during the past century, the economy has been transformed by two world wars, a global depression, and innovations in the transportation, communication, and information systems industries. orville and wilbur wright’s first flight was in 1903, the ford model t debuted in 1908, at&t completed the first transcontinental telephone line in 1915, the first electrical, binary, programmable computer was invented during the 1930s, and the construction of the interstate highway system commenced in 1956. for an individual firm to have maintained a constant (or even positive) growth rate across the past century, the firm would have been required to reinvent itself multiple times to adjust to these (and other) events. simply maintaining a competitive advantage (and a constant, positive growth rate) over a 25 year period can be a challenge. as noted by sheth (2009), many of the 62 firms highlighted in the 1982 book “in search of excellence” (including sears, xerox, kodak, dana, and digital computer corp.) experienced financial hardships in the ensuing 25 years. thus, investors should be wary of purchasing any asset at a price that can only be justified by cash flows to be received in the unforeseeable, distant future. previous studies identify two alternative ways to minimize the importance of far distant cash flows in the valuation process. first, shaffer (2006) extends the gordon model to include a constant, annual probability of permanent failure. second, numerous valuation models allow the assumed growth rate to decrease over time (e.g., miller and modigliani, 1961; holt, 1962; mao, 1966; fielitz and muller, 1985; gordon and gordon, 1997; danielson, 1998; o’brien, 2003). this section reports on potential valuation errors embedded in perpetual growth model price estimates if future cash flows will be constrained by default, finite growth, or both. 4.1. valuation models: default and finite growth shaffer (2006) extends the perpetual growth model to allow for the possibility that the firm will irreversibly fail (and the cash flow stream will end) in the future. assuming that the annual failure probability (p) is constant over time, eq. (1) can be rewritten as eq. (6), and eq. (2) can be rewritten as eq. (7). p0 � d1�1 � p� r � p � g�1 � p� (6) p0 � e1�1 � b��1 � p� r � p � brn�1 � p� (7) shaffer (2006) separately modifies the perpetual growth model to allow the growth rate to permanently decrease—from a supernormal level (i.e., rn � r) down to the required return—in any future year. again, shaffer assumes that this event occurs with a constant 195m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 annual probability. however, shaffer does not develop a form of the perpetual growth model that allows for both default (where the cash flow stream ends) and finite growth (where the firm’s growth rate permanently decreases, but the cash flow stream continues). to develop such a model, we simplify the assumptions, and stipulate that the firm’s return on new investments, rn, be equal to the required return, r, starting in the specific future year t � 1. as derived in appendix 1, the adjusted perpetual growth model, allowing for default and finite growth, can be written as eq. (8). p0 � e1�1 � b��1 � p� r � p � brn�1 � p� �1 � � (1 � p)(1 � brn) (1 � r) � t� � � (1 � p)(1 � brn) (1 � r) � t� e1(1 � b)(1 � p) r � p � br(1 � p)� (8) if the default probability is zero (p � 0), eq. (8) can be written as eq. (9). this is the traditional finite growth model from miller and modigliani (1961) and gordon and gordon (1997). p0 � e1�1 � b� r � g �1 � �1 � g 1 � r� t� � e1(1 � g)t r(1 � r)t (9) 4.2. estimation errors with default (but perpetual growth) in this section, we compare value estimates from eq. (7) to those obtained using eq. (2). if the intrinsic value of the firm is defined by eq. (7) (because the firm might fail in the future) by how much will the perpetual growth model overstate firm value? the percentage overstatement embedded in a perpetual growth value estimate in this case (positive default probability; perpetual growth) is eq. (10). appendix 2 derives this equation. % overstatement �p � 0; perpetual growth� � p�1 � r� �r � g��1 � p� (10) table 2 reports on the potential valuation errors—calculated using eq. (10)—that can arise when the perpetual growth model is used to value a firm that might fail in the future. panel a lists the assumptions used in this exercise, and calculates firm value using the unadjusted perpetual growth model. in particular, the firm is expected to have earnings per share of $1 next year, the required return (stated as a real interest rate) is 7%, and the return on new investments (again stated as a real interest rate) is 7%. the unadjusted perpetual growth model, eq. (2), estimates the current stock price as $14.29 regardless of whether the reinvestment rate is 0%, 50%, or 99%. the results in table 2, panel a illustrate this for eight different reinvestment rates, b. although each reinvestment rate produces the same value estimate when plugged into the perpetual growth model, they do not create identical challenges for the firm’s managers, or impose equal risks on the firm’s shareholders. mathematically, value is preserved if retained earnings are reinvested in projects earning a return equal to the cost of capital, and if the default probability is zero. however, 196 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 identifying projects that can earn a return equal to the cost of capital is not a trivial task in a competitive economy. for example, if the economy is growing at a real rate of 2% per year, but the firm must earn 7% (i.e., the cost of capital) on new investments to preserve value, the process of deferring dividends into the future might require the firm to grow faster than the economy as a whole, simply to maintain its value.5 it is possible that the aggressive policies required to achieve this growth could ultimately cause the firm to “grow broke,” contributing to (or magnifying) the firm’s default probability. table 2 estimation errors–default panel a: assumptions and unadjusted value estimates (default probability � 0) scenario 1 2 3 4 5 6 7 8 r � 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 r � 0.07 0.07 0.07 0.07 0.07 0.07 0.07 0.07 b � 0.0000 0.1429 0.2857 0.4286 0.5714 0.7143 0.8571 0.9286 g � br � 0 0.01 0.02 0.03 0.04 0.05 0.06 0.065 r � g � 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.005 ppgm � 14.286 14.286 14.286 14.286 14.286 14.286 14.286 14.286 panel b: default probability � 0.0025 scenario 1 2 3 4 5 6 7 8 padj � 13.759 13.675 13.559 13.388 13.113 12.597 11.265 9.299 %os � 3.8% 4.5% 5.4% 6.7% 8.9% 13.4% 26.8% 53.6% panel c: default probability � 0.005 scenario 1 2 3 4 5 6 7 8 padj � 13.267 13.111 12.899 12.593 12.114 11.259 9.290 6.883 %os � 7.7% 9.0% 10.8% 13.4% 17.9% 26.9% 53.8% 107.5% panel d: default probability � 0.0075 scenario 1 2 3 4 5 6 7 8 padj � 12.806 12.589 12.297 11.884 11.253 10.173 7.899 5.459 %os � 11.6% 13.5% 16.2% 20.2% 27.0% 40.4% 80.9% 161.7% panel e: default probability � 0.01 scenario 1 2 3 4 5 6 7 8 padj � 12.375 12.105 11.747 11.247 10.502 9.274 6.865 4.518 %os � 15.4% 18.0% 21.6% 27.0% 36.0% 54.0% 108.1% 216.2% notes. this table compares price estimates from the unadjusted perpetual growth model, eq. (2), to price estimates from the perpetual growth model with default, eq. (7), and reports the percentage overstatement of value embedded in the unadjusted perpetual growth model price estimate. the firm’s estimated earnings next year, e1, is assumed to be $1, the required return, r, is 7%, and the return on new investments, r, is 7%. panel a lists combinations of reinvestment rate b and growth rate g that produce a stock price of $14.286 using eq. (2). panels b through e list the adjusted price estimates from eq. (7), padj, for each scenario using default probabilities of 0.0025 (panel b), 0.005 (panel c), 0.0075 (panel d), and 0.01 (panel e), and report the % difference between the unadjusted and adjusted value estimates: %os � (ppgm � padj)/padj. the values reported in the %os rows can also be calculated using eq. (10). 197m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 table 2, panels b to e, report potential overstatements embedded in perpetual-growth value estimates when the default probability is positive. the table reports results for failure probabilities ranging from 0.25% to 1%. we use this range of failure probabilities because shaffer (2006) reports that the average annual failure rate for u.s. businesses during the 1955 to 1995 period was 0.6%. the table 2 results reveal that the potential overstatement is directly related to the timing of the implied future cash flows. if r – g is 0.5 percentage points—that is, d/p is 0.5% (scenario 8) and the bulk of the estimated value is created by cash flows from distant years—the perpetual growth model can overstate value by almost 54% (over 216%) if the firm faces a constant, annual, failure probability of 0.25% (1%). in contrast, if r – g is seven percentage points (scenario 1), the potential overstatement implied by the perpetual growth model is only 3.8% (15.4%) when the annual failure probability is 0.25% (1%). if a firm faces a positive default probability, the process of deferring dividends creates additional risk (i.e., default risk) for shareholders. thus, for a firm to maintain value after deferring dividends, the return on investment earned by the new projects must exceed the cost of capital in those outcomes in which the firm does not fail. that is, the project must produce an “ex-post” positive npv to compensate investors for the firm’s default risk. eq. (1), which is derived in appendix 3, calculates the return on new investments required to preserve value for a firm that retains and reinvests earnings, if the firm faces an annual default probability of p. rn�breakeven� � r � p 1 � p (11) thus, if a firm’s annual default probability is 1% and its required return is 7%, new projects must earn an 8.08% return in those outcomes in which the firm is successful for the investments to preserve value. 4.3. estimation errors with finite growth (no default) because the gordon model assumes growth will continue forever, the model can overstate the value of an asset if rn � r, and if competition will limit the period of time positive net present value projects will be available (stigler, 1963). in this section, we assume that the default probability is zero and focus directly on the value overstatement created by the use of a growth phase that is too long (i.e., perpetual vs. finite growth period). if eq. (9) defines an asset’s intrinsic value, the use of the perpetual growth model, eq. (2), will produce a value estimate that is too high. to calculate this valuation error as a percentage of the intrinsic value, subtract eq. (9) from eq. (2), and divide the difference by eq. (9). after much algebra, this yields eq. (12). %overstatement �p � 0; finite growth� � 1 1 � �1 � g 1 � r� t� b 1 � b��rn r � 1� � 1 (12) 198 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 table 3 reports potential valuation errors using eq. (12). the inputs include four combinations of r, g, b, and rn (with p � 0) and competitive advantage periods ranging from 10 to 100 years. each combination of r, g, b, and rn produces a p/e ratio of 30 using the unadjusted perpetual growth model.6 if competition will limit a firm’s growth phase to 10 years (but the default probability is zero) table 3 shows that the unadjusted perpetual growth model will overstate firm value by over 65% when the dividend yield is 3% (scenario 1), and by almost 100% when d/p is 0.5% (scenario 4).7 if the firm’s growth phase can be maintained for 20 years, eq. (2) will still overstate firm value by over 42% when d/p is 3% and by over 91% when d/p is 0.5%.8 even if the true growth period is 100 years, the unadjusted perpetual growth model will overstate value by almost 49% when d/p is 0.5%. in contrast, the unadjusted perpetual growth model only overstates value by 3% when the growth period equals 100 years and the dividend yield is 3%. these results suggest that the perpetual growth model is an appropriate first-cut valuation tool for a fairly limited set of firms: high-dividend firms with very strong competitive advantages (i.e., 50 or more years). table 3 estimation errors–finite growth scenario 1 2 3 4 r � 0.07 0.07 0.07 0.07 r � 0.4 0.125 0.0857 0.0765 b � 0.1 0.4 0.7 0.85 g � br � 0.04 0.05 0.06 0.065 r � g � 0.03 0.02 0.01 0.005 ppgm � 30 30 30 30 t � 10: padj � 18.175 16.988 15.694 15.005 %os � 65.1% 76.6% 91.2% 99.9% t � 20: padj � 21.102 19.225 16.976 15.691 %os � 42.2% 56.0% 76.7% 91.2% t � 50: padj � 26.209 23.883 20.173 17.567 %os � 14.5% 25.6% 48.7% 70.8% t � 100: padj � 29.085 27.619 23.855 20.163 %os � 3.1% 8.6% 25.8% 48.8% notes. this table compares value estimates from the unadjusted perpetual growth model, eq. (2), to value estimates from the extended model allowing finite growth (but no default), eq. (9), and reports the percentage overstatement embedded in the unadjusted perpetual growth model value estimate. the firm’s estimated earnings next year, e1, is $1, the required return, r, is 7%, and the firm’s p/e ratio is 30. the table lists four combinations of reinvestment rate b, return on new investments r, and growth rate g that produce a stock price estimate of $30 (p/e � 30) using eq. (2). for each combination, the table reports the adjusted value estimate from eq. (9), padj, and the % difference between the unadjusted and adjusted value estimates: %os � (ppgm � padj)/padj. the values reported in the %os rows can also be calculated using eq. (12). this information is reported assuming the period in which the firm can invest in projects where r � r will last 10, 20, 50, and 100 years. 199m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 4.4. estimation errors with default and finite growth table 4 compares value estimates from the perpetual growth model, eq. (2), to value estimates from eq. (8), which adjusts the perpetual growth model for both finite growth and potential default. the examples in table 4 calculate potential valuation errors, for competitive advantage periods ranging from 10 to 100 years, using the same four combinations of r, g, b, and rn as in table 3. panel a lists the assumptions used in each scenario. panels b through e list potential value overstatements—embedded in value estimates from eq. (2)—when firms face positive default probabilities ranging from 0.25% to 1%. table 4 estimation errors–default and finite growth panel a: assumptions and perpetual growth model value estimate scenario 1 2 3 4 r � 0.07 0.07 0.07 0.07 r � 0.4 0.125 0.0857 0.0765 b � 0.1 0.4 0.7 0.85 g � br � 0.04 0.05 0.06 0.065 r � g � 0.03 0.02 0.01 0.005 ppgm � 30 30 30 30 panel b: default probability � 0.0025 scenario 1 2 3 4 t � 10: padj � 17.384 15.935 13.900 11.944 %os � 72.6% 88.3% 115.8% 151.2% t � 20: padj � 20.087 17.959 14.994 12.470 %os � 49.4% 67.1% 100.1% 140.6% t � 50: padj � 24.593 21.9791 17.593 13.839 %os � 22.0% 36.5% 70.5% 116.8% t � 100: padj p � 26.911 24.916 20.311 15.556 %os � 11.5% 20.4% 47.7% 92.9% panel c: default probability � 0.005 scenario 1 2 3 4 t � 10: padj � 16.653 14.997 12.465 9.910 %os � 80.1% 100.0% 140.7% 202.7% t � 20: padj � 19.151 16.834 13.410 10.330 %os � 56.6% 78.2% 123.7% 190.4% t � 50: padj � 23.135 20.317 15.550 11.372 %os � 29.7% 47.7% 92.9% 163.8% t � 100: padj � 25.007 22.636 17.583 12.556 %os � 20.0% 32.5% 70.6% 138.9% (continued on next page) 200 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 the results in table 4 show that the introduction of a positive default probability can magnify the potential overstatements reported in table 3 (where the default probability is 0). if the firm’s dividend yield is 3% (scenario 1), table 4, panel b reveals that the percentage overstatement in a perpetual growth model value estimate is 72.6% when p � 0.25% and t � 10 years (vs. 65.1% in table 3, p � 0), and is 11.5% when p � 0.25% and t � 100 (vs. 3.1% in table 3, p � 0). if the firm’s dividend yield is only 0.5% (scenario 4), the differences between the percentage overstatements in table 4, panel b and table 3 become more table 4 (continued) panel d: default probability � 0.0075 scenario 1 2 3 4 t � 10: padj � 15.975 14.157 11.291 8.461 %os � 87.8% 111.9% 165.7% 254.6% t � 20: padj � 18.288 15.828 12.115 8.806 %os � 64.0% 89.5% 147.6% 240.7% t � 50: padj � 21.815 18.856 13.895 9.621 %os � 37.5% 59.1% 115.9% 211.8% t � 100: padj � 23.330 20.693 15.432 10.463 %os � 28.6% 45.0% 94.4% 186.7% panel e: default probability � 0.01 scenario 1 2 3 4 t � 10: padj � 15.344 13.399 10.312 7.375 %os � 95.5% 123.9% 190.9% 306.8% t � 20: padj � 17.488 14.925 11.038 7.665 %os � 71.5% 101.0% 171.8% 291.4% t � 50: padj � 20.615 17.564 12.531 8.316 %os � 45.5% 70.8% 139.4% 260.7% t � 100: padj � 21.844 19.025 13.704 8.928 %os � 37.3% 57.7% 118.9% 236.0% notes. this table compares value estimates from the unadjusted perpetual growth model, eq. (2), to value estimates from the extended model allowing both default and finite growth, eq. (8), and reports the percentage overstatement embedded in the unadjusted perpetual growth model value estimate. the firm’s estimated earnings next year, e1, is $1, the required return, r, is 7%, and the firm’s p/e ratio is 30. panel a lists combinations of reinvestment rate b, return on new investments r, and growth rate g that produce a stock price estimate of $30 (p/e � 30) using eq. (2). panels b through e list the adjusted value estimates from eq. (8), padj, for each scenario using default probabilities of 0.0025 (panel b), 0.005 (panel c), 0.0075 (panel d), and 0.01 (panel e), and report the % difference between the unadjusted and adjusted value estimates: %os � (ppgm � padj)/padj. within each panel, this information is reported assuming the period in which the firm can invest in projects where r � r will last 10, 20, 50, and 100 years. 201m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 pronounced. the percentage overstatement in a perpetual growth model value estimate is 151.2% when p � 0.25% and t � 10 years (vs. 99.9% in table 3, p � 0), and is 92.9% when p � 0.25% and t � 100 (vs. 48.8% in table 3, p � 0). as the estimated default rate increases from panel b to panel e, the potential overstatement embedded in perpetual-growth-model value estimates continues to grow. if the firm’s dividend yield is 3% (scenario 1), table 4, panel e reveals that the percentage overstatement in a perpetual-growth-model value estimate is 95.5% when p � 1% and t � 10 years (vs. 65.1% in table 3, p � 0), and is 37.3% when p � 1% and t � 100 (vs. 3.1% in table 3, p � 0). if the firm’s dividend yield is only 0.5% (scenario 4), the percentage overstatement is 306.8% when p � 1% and t � 10 years (vs. 99.9% in table 3, p � 0), and is 236.0% when p � 1% and t � 100 (vs. 48.8% in table 3, p � 0). the results in table 4 further limit the population of firms for which the perpetual growth model is an appropriate valuation tool. in addition to having a high dividend yield and a strong competitive advantage, the firm should also be financially strong, with a very low probability of default. 5. conclusions the perpetual-growth model provides investors with a simple way to calculate the present value of a perpetual dividend stream. the model is a very useful teaching tool, as it clearly illustrates the links between reinvestment policies, reinvestment returns, and stock prices. in addition, the economic interpretation of valuation ratios (such as the price-to-earnings ratio) is often explained within the framework of the constant growth model (e.g., leibowitz and kogelman, 1990). thus, it is not surprising that the model “… is taught in all top-tier business schools …” as noted by bradley and jarrell (2008). moving from the classroom to the investing world, however, the usefulness of the model can be called into question. to exploit the computational elegance of the model, one must assume that cash flows will grow at a constant rate forever. this assumption is of dubious validity in a competitive economy, where a firm’s growth can be derailed by internal mistakes, innovations by other firms, or macroeconomic shocks. the analytical results in this article suggest that the perpetual-growth model should be applied in a very limited set of circumstances: when the firm faces a low default probability, when only a small portion of the growth will be the result of positive net present value investments, and when the firm’s dividend yield is sufficiently large. along these lines, foerster and sapp (2005) show that the constant-growth model produced reasonable value estimates for the bank of montreal (e.g., value estimates that approximated historical stock prices), a large, mature, dividend-paying company. for firms that do not fit this profile, individual investors, financial analysts, and financial planners should use the perpetualgrowth model with caution. 202 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 notes 1 we acknowledge that this statement is true (for the perpetual growth model) only over the range of reinvestment rates, b, in which the implied growth rate g (� brn) is less than the required return r. 2 danielson and scott (2000) and weston, chung, and siu (1998) use a similar approach to reconcile the constant growth model in nominal terms to the constant growth model in real terms. 3 for example, assume that b � 0.4, e1 � $1, r � 8%, rn � 12%, and h � 2%. thus, next year’s nominal earnings are $1.02, the nominal required return is 10.16%, and the nominal growth rate is 6.896% (� (1 – 0.4)(2%) � 0.4(14.24%)). using real inputs, the stock price is $18.75 (� [$1(1 – 0.4)]/[0.08 – 0.4(0.12)]. the stock price using eq. (3) and nominal inputs is also $18.75 (� [$1.02(1 – 0.4)]/[0.1016 – 0.06896]. in contrast, if the nominal growth rate is defined simply as the retention rate times the nominal return on investments, 5.696% � 0.4(14.24%), the estimated stock price using this (incomplete) nominal growth rate, the nominal discount rate, and the nominal earnings is $13.71 (� [$1.02(1 – 0.4)]/[0.1016 –0.05696]. thus, the incorrect application of the constant growth model will understate value by almost 27%. 4 according to the gordon model, if the expected dividend next year is d1, the required return is r, and the perpetual growth rate is g, then the estimated stock price today is p0 � d1/(r – g). this expression can be rearranged to write the dividend yield as d1/p0 � (r – g). 5 the firm’s growth rate will depend on the reinvestment rate and the return on new investments. for example, if the retention rate b is 90% and the return on new investments is 7%, the firm’s growth rate is 6.3%. for an extreme illustration of the challenge facing firms, assume that a firm currently has a market share of 2%. if the firm’s product market is growing at an annual rate of 2%, and the firm grows at an annual rate of 6.3%, the firm’s revenues will become larger than those of its industry in 95 years. however, the examples in table 1 show that a sizeable portion of firm value must be realized after year 100 when the difference between r and g is small. 6 the p/e ratio can be derived from the constant growth model by dividing each side of eq. (2) by e1. this yields p/e � (1 – b)/(r – brn). 7 empirical evidence in fuller, huberts, and levinson (1993) and lakonishok, shleifer and vishney (1994) reveals that high p/e firms (e.g., p/e � 30) experience higher future growth rates than low-p/e firms. however, the superior growth of these firms typically lasts for less than 10 years. 8 the traditional finite growth model, and eqs. (8) and (9), make the unrealistic simplifying assumption that a firm’s competitive advantage period will end abruptly at a specified future date, and that the return on new investments will at that point fall immediately to the cost of capital. it is perhaps more likely that the return on new investments will decrease gradually over some future period. for example, if the return on new investments remains equal to rn for the next 10 years, and then gradually decreases to r over a 10 year period, the true valuation errors will fall between the overstatements reported in the t � 10 and t � 20 rows of table 3. 203m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 appendix 1 this appendix derives the perpetual growth model adjusted to allow for both permanent failure and finite growth. to begin, note that the perpetual growth model, allowing for permanent failure, is defined by eq. (7). this expression is repeated here as eq. (a1). p0 � e1�1 � b��1 � p� r � p � brn�1 � p� (a1) if the return on new investments remains rn forever, the stock price at the end of year t can be written as eq. (a2). if the return on new investments changes to r* starting in year t � 1, whereas the plowback rate (b) and the failure probability (p) remain constant, the stock price at the end of year t (p*t ) can be written as eq. (a3). pt � et�1�1 � b��1 � p� r � p � brn�1 � p� (a2) p*t � et�1�1 � b��1 � p� r � p � br*�1 � p� (a3) given that the annual failure probability is p, there is a (1 – p)t chance that the firm will survive past year t. in addition, only the discounted values of eqs. (a2) and (a3) impact the stock price today; the future stock prices must be divided by (1 � r)t to calculate their present values. thus, the stock price today, allowing for the shift in the investment return, can be written as eq. (a4). p0 � e1�1 � b��1 � p� r � p � brn�1 � p� � � (1 � p) (1 � r)� t �p*t � pt� (a4) in each of years 1 through t, the firm’s earnings will grow at the annual rate brn. thus, the firm’s earnings in year t � 1 can be written as eq. (a5). et�1 � e1�1 � brn�t (a5) if the firm cannot invest in positive net present value projects after year t (that is the typical assumption in finite growth models), r* equals r. using this assumption, and substituting eq. (a5) into (a4), produces eq. (a6), which is eq. (8) in the text. p0 � e1�1 � b��1 � p� r � p � brn�1 � p� �1 � � (1 � p)(1 � brn) (1 � r) � t� � � (1 � p)(1 � brn) (1 � r) � t� e1(1 � b)(1 � p) r � p � br(1 � p)� (a6) 204 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 appendix 2 this appendix derives an equation that quantifies the percentage overstatement embedded in perpetual growth model value estimates for firms that may fail in the future, with an annual failure probability of p, but will grow at the perpetual annual rate g if they do not fail. the valuation error is stated as a percentage of the true stock price, calculated by eq. (7). to do this, subtract eq. (7) from eq. (2), and divide the difference by eq. (7). this step yields eq. (a7), which simplifies to eq. (a8). % overstatement �p � 0; perpetual growth� � d1 r � g � d1�1 � p� r � p � g�1 � p� d1�1 � p� r � p � g�1 � p� (a7) % overstatement �p � 0; perpetual growth� � r � p � g�1 � p� �r � g��1 � p� � 1 (a8) after rewriting the integer 1 on the right-hand side of eq. (a8) as 1 � [(r – g)(1 – p)]/ [(r – g)(1 – p)], the equation further simplifies to eq. (a9), which is eq. (10) in the text. % overstatement �p � 0; perpetual growth� � p�1 � r� �r � g��1 � p� (a9) appendix 3 this appendix derives the breakeven return on new investments for a firm with a positive probability of failure. if a firm’s default probability is zero, the firm’s breakeven reinvestment return will equal the required return. if a firm faces a positive default probability, the process of deferring dividends through the reinvestment process creates additional default risk for shareholders. for a firm to maintain value after increasing its plowback ratio, the expected return on new investments must exceed the cost of capital. the breakeven return on new investments is the return a firm must earn, in outcomes in which the firm does not default, to maintain value if the firm reduces its payout ratio. to illustrate this, assume that eq. (a1) defines firm value before it changes its plowback rate b. if the firm increases its plowback rate to b � �, firm value can be written as eq. (a10). p*0 � e1�1 � b � ���1 � p� r � p � �b � �� rn�1 � p� (a10) to solve for the breakeven return on new investments, set eq. (a1) equal to (a10), and solve algebraically for rn. this expression simplifies to eq. (a11), which is eq. (11) in the text. rn�breakeven� � r � p 1 � p (a11) 205m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 references block, s. (1999). a study of financial analysts: practice and theory. financial analysts journal, 55, 86–95. bradley, m., & jarrell, g. (2008). expected inflation and the constant-growth valuation model. journal of applied corporate finance, 20, 66–78. brealey, r. a., & myers, s. c. (2003). principles of corporate finance. new york: mcgraw-hill/irwin. danielson, m. (1998). a simple valuation model and growth expectations. financial analysts journal, 54, 50–57. danielson, m., heck, j., & shaffer, d. (2008). shareholder theory: how opponents and proponents both get it wrong. journal of applied finance, 18, 62–66. danielson, m., & scott, j. (2000). expanding the scope of valuation analysis. financial practice and education, 10, 52–61. demirakos, e. g., strong, n. c., & walker, m. (2004). what valuation models do analysts use? accounting horizons, 18, 221–240. dukes, w., peng, z., & english, p. (2006). how do practitioners value common stock? journal of investing, 15, 90–104. fielitz, b., & muller, f. (1985). a simplified approach to common stock valuation. financial analysts journal, 41, 35–41. foerster s., & sapp, s. (2005). the dividend discount model in the long-run: a clinical study. journal of applied finance, 15, 55–75. fuller, r., huberts, l., & levinson, m. (1993). returns to e/p strategies, higgledy-piggledy growth, analysts’ forecast errors, and omitted risk factors. journal of portfolio management, 19, 13–24. gordon, m. (1962). the investment, financing, and valuation of the corporation. homewood, il: irwin. gordon, j., & gordon, m. (1997). the finite horizon expected return model. financial analysts journal, 53, 52–61. holt, c. (1962). the influence of growth duration on share prices. journal of finance, 17, 465–475. imam, s., barker, r., & clubb, c. (2008). the use of valuation models by uk investment analysts. european accounting review, 17, 503–535. lakonishok, j., shleifer, a., & vishny, r. (1994). contrarian investment, extrapolation, and risk. journal of finance, 49, 1541–1578. leibowitz, m., & kogelman, s. (1990). inside the p/e ratio: the franchise factor. financial analysts journal, 46, 17–35. macaulay, f. (1938). some theoretical problems suggested by the movements of interest rates, bond yields, and stock prices in the united states since 1856. new york: columbia university press. mao, j. (1966). the valuation of growth stocks: the investment opportunities approach. journal of finance, 21, 95–102. miller, m., & modigliani, f. (1961). dividend policy, growth, and the valuation of shares. journal of business, 34, 411–433. o’brien, t. (2003). a simple and flexible dcf valuation model. journal of applied finance, 13, 54–62. payne, t., & finch, j. h. (1999). effective teaching and use of the constant growth dividend discount model. financial services review, 8, 283–291. shaffer, s. (2006). corporate failure and equity valuation. financial analysts journal, 62, 71–80. sheth, j. (2009). the self-destructive habits of good companies. philadelphia: wharton school publishing. stigler, g. j. (1963). capital and rates of return in manufacturing industry. princeton, nj: university press. weston, j. f., chang, s. c., & siu, j. (1998). takeovers, restructuring, and corporate governance. upper saddle river, nj: prentice hall. welch, i. (2009). corporate finance. new york: prentice hall. 206 m.g. danielson, j.l. heck / financial services review 23 (2014) 189–206 the evidence on target-date mutual funds sandeep singha,* adepartment of accounting, economics and finance, school of business administration and economics, college at brockport, suny, brockport, ny 14420, usa abstract this paper assimilates the knowledge and evidence on target-date mutual funds (tdfs). it begins with a discussion of the environment that contributes to the tremendous growth of tdfs. next, a survey of the theory and recommendations on glide paths indicates a trend towards focusing on meeting retirement liabilities, rather than optimizing asset only portfolios. a review of performance evaluation metrics for tdfs shows that none of the available indexes possesses all seven characteristics of an ideal benchmark. plan sponsors can provide better outcomes by offering multiple risk profile tdfs while researchers can focus on improving glide path and benchmark design. © 2016 academy of financial services. all rights reserved. jel classification: j26; c00; d14; g11 keywords: target-date funds; life cycle funds; retirement; asset allocation 1. introduction the department of labor’s safe harbor provision of 2007 has provided a significant boost to the popularity of target date mutual funds (tdfs). these funds are now included as a default investment option in defined contribution plans along with managed accounts and balanced funds. according to department of labor, the qualified default investment alternative (qdia) provides a plan sponsor “safe harbor relief from fiduciary liability for investment outcomes ebsa (2008).” a consequence of the pension protection act of 2006, employers are increasingly adopting the automatic enrollment option in 401(k) plans, which further supplements the assets under management (aum) base of tdfs. * corresponding author. tel.: �1-585-395-5519; fax: �1-585-395-2542. e-mail address: ssingh@brockport.edu financial services review 25 (2016) 235–262 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. investors held about $1 trillion in target date and lifestyle funds at the end of 2014, compared to a total of $2.75 billion at the end of 1995 (ici handbook, 2015). this calculates to a compounded annual growth rate (cagr) of about 35% over the 10-year period, an impressive flow of investor money to these funds.1 a tdf is managed from its purchase date through an expected retirement year, and in some cases, the fund provides an additional “during retirement” investment option. on the other hand, a lifestyle fund is directed to a broad age group and tailored to its purported risk tolerance. for example, a lifestyle fund focused on younger investors in their 30s will usually have a high equity exposure, while a lifestyle fund aimed at retirees will likely have a heavy emphasis on fixed income securities. assuming that one’s risk tolerance decreases with age, an investor will move from one lifestyle fund to the next one catering to their “lifestyle” as they age, until they are in retirement. in the case of a tdf, this asset allocation is automatically shifted for the investor. a good way to understand the difference between target date and lifestyle funds is to think of lifestyle funds as building blocks of a tdf. so, which of these two fund types should investors prefer? chang et al. (2014) use a utility maximization framework and use bootstrap simulations to compare welfare benefits of both types of funds. the primary focus of their research is to measure utility derived from fixed and decreasing equity allocation for an individual investor over time. they find that a decreasing equity allocation provides better welfare benefits than a static one. this implies that tdfs are superior investment vehicles from a utility maximization perspective. the authors caution that there is no one-size-fits all tdf, the investor should select such funds based on their risk tolerance. for the purposes of this paper, the term “target date fund” is used to designate a broad class of open-end mutual funds and exchange traded funds (etfs) that include both target date and lifestyle funds. the appeal of a target-date mutual fund lies in the convenience it provides to the investor. she does not have to monitor, and periodically alter the asset allocation because of passage of time. in most cases, the fund mandate automatically provides for that. based on the chosen retirement age, the fund manager allocates the funds to a predetermined allocation schedule; say 85% stocks and 15% fixed income initially when the investor is young. this allocation may eventually reverse to 15% stocks and 85% fixed income closer to the “target date.” this change in asset allocation that occurs over many years is commonly referred to as the fund’s “glide path.” the vanguard target retirement fund 2055 (ticker: vffvx), is provided as an example. the 2055 in the fund name indicates the anticipated retirement year for the investor. in 2015, this fund is recommended for a 30-year old anticipating retirement at age 70 or a 25-year old anticipating retirement at age 65. based on the fund’s glide path, the initial asset allocation is quite aggressive, with up to 90% in equity securities. according to vanguard, this becomes “more conservative over time, meaning that the percentage of assets allocated to stocks will decrease while the percentage of assets allocated to bonds and other fixed income investments will increase.” after the year 2055, the fund will mimic the allocation of vanguard target retirement income fund (ticker: vtinx), another fund in the vanguard funds family. for a comparison of the two funds, the reader is referred to table 1 (vanguard group [2015]). the glide path of the vanguard target retirement funds is shown in fig. 1. as expected, the glide path of the fund becomes more conservative as the target date approaches. based 236 s. singh / financial services review 25 (2016) 235–262 on the fund’s prospectus the manager of the fund has significant discretion in the asset allocation decision. for example, while the allocation to equities is required to diminish from 90% to 30% by the retirement date, the manager has discretion of plus or minus seven years around the target date to accomplish that. such discretion may not be the case for all target date mutual funds. many have a more rigid timetable for asset allocation change. there are multiple approaches to designing tdfs with each fund family emphasizing their approach as superior, perhaps in the pursuit of market share. some families use index funds in the lineup of tdfs while others rely on actively managed ones. some have rigid or semi-rigid asset allocations while others provide their managers with broader latitude in such decisions. the wide range of design approaches can be illustrated by comparing the vantable 1 comparative information on vanguard target retirement fund 2055 and target retirement income funda composition of funds vanguard target retirement 2055 fund vanguard target retirement income fund ticker vvfvx vtinx assets under management $1.67 billion $11.21 billion target equity 90% 30% target fixed income and cash 10% 70% acquired fund fees and expenses 0.18% 0.16% asset allocations as of 9/30/2014 total stock market index 63.0% 21.1% total international stock market index 27.0% 8.8% total bond market ii index 8.0% 39.3% total international bond index fund 2.00% 14.0% short-term inflation-protected securities 0.0% 16.8% note: areported for september 30, 2014. fig. 1. an illustrative glide path for vanguard target date fund. 237s. singh / financial services review 25 (2016) 235–262 guard funds to pimco funds. pimco tdfs utilize a liability driven investing (ldi) approach to asset allocation changes. the design is motivated by the goal of meeting inflation-adjusted withdrawals (liabilities) during retirement (a more in-depth discussion on liability driven investing [ldi] is provided in a later section of this paper). fig. 2 depicts the glide path of pimco tdfs. the wide range of approaches to tdf design is starkly contrasted by figs. 1 and 2. the key difference that can be observed is the exposure to inflation insensitive assets like tips, commodities and, real estate during retirement in the two glide paths. to the best of my knowledge, this is the first paper that provides a comprehensive review and evaluation of the body of knowledge on tdfs. the focus of this paper is four fold. in the first section, a discussion of the current state of the target date mutual fund market is provided. the discussion includes analysis of factors that have aided in the growth of this segment of the investment companies market. in the second section, a discussion of the theoretical basis and a summary of varied suggestions for designing glide paths are presented. in the third section performance and benchmarking of these funds is discussed. in the fourth, recommendations and suggestions for plan sponsors, financial advisors and researchers are made. concluding remarks are provided in the final section. 2. the tdf market over the 10-year period 2005–2014, assets under management in tdfs have grown about five-fold. according to the investment company institute (ici), investors had a choice of 797 tdfs for a total of 3,036 mutual fund share classes at the end of 2014 (ici [2015]). bauer, fig. 2. glide path of pimco target-date funds. 238 s. singh / financial services review 25 (2016) 235–262 phillips, and white (2009) predict that 80% of all new defined contribution plan investments will go to tdfs and the target-date market will be $2.3 trillion by end of 2018. there are three broad factors playing a role in the remarkable growth of this market namely, the regulatory environment, demographics, and investor affinity for passive investments. before tdfs became popular, agnew et al. (2003) studied 7,000 accounts of a large defined contribution plan over the period 1994 to1998. they find significant participant indifference towards the management of their retirement assets. most asset allocations are at extremes, either 100% or 0% equities. they also find “very limited portfolio reshuffling” implying that rebalancing of portfolios is rare. more alarmingly, in a study of over a million retirement accounts from about 1,000 pension plans, tang and mitchell (2010) find that most employers provide an efficient menu of investment choices in pension plans. however, participants undo this by making inefficient portfolio selection decisions. this costs the participants a reduction of about one-fifth in potential retirement wealth. in addition to participant education, the natural corollary of these findings would be to have investment vehicles in retirement plans that do not require significant participant intervention over time, that is “do no harm.” each of the preceding studies shows that fund owners are not likely to manage their retirement funds to their best advantage. this provides a compelling argument in favor of tdfs. funds that absolve the portfolio owner of monitoring and periodic adjustments to reflect altered risk tolerance. the most significant boost to tdfs comes from two regulatory developments: (1) the automatic enrollment (opt-out option) feature under the pension protection act of 2006 and (2) the department of labor directive of 2007 that provides safe harbor relief to plan sponsors if a tdf is offered as a qualified default investment option (qdia). the other two qdia options that emerged from the pension protection act of 2006 (ppa) are managed accounts and balanced funds ebsa (2013). this regulatory protection has greatly increased the popularity of tdfs in qualified defined contribution retirement plans. according to landsberg (2007), ppa provides changes in five broad areas: defined benefit plan funding, hybrid qualified plans, defined contribution plans, executive compensation, and plan distributions. two broad set of key changes promote the growth of assets invested in tdfs. the first relates to enrolment procedures and regulatory relief. this includes permission to employers to offer automatic enrollment in 401(k) plans and safe harbor provisions for default options that include tdfs in automatic enrollment plans. a default investment option becomes the deemed choice of a participant when no selection of investment alternatives is specifically made. the second contributing factor makes permanent the provisions of economic growth and tax relief reconciliation act (egtrra) of 2001 with respect to increased contributions, inflation indexing and “catch up” provisions for contributions to various individual retirement arrangements (iras) and 401(k) plans. balduzzi and ruetter (2015) report increased heterogeneity in risk taking and associated returns of target date mutual funds since the passing of ppa of 2006. they attribute this divergence in performance to risk taking by fund families with low market shares attempting to differentiate themselves to attract plan sponsors. consequently, participants in different funds of the same vintage year might experience returns significantly different from the average. in a study of its participants, tiaa-cref researchers richardson and bissette (2014) find that over the period 2005–2011 tdfs garnered a significant proportion (22%) of all 239s. singh / financial services review 25 (2016) 235–262 contributions. these funds are the preferred choice of younger participants with lower account balances and are utilized as a fund of funds strategy, meaning that the target date fund is the primary investment of the participant. a similar trend is reported for the broader 401(k) market by ici. at the end of 2013, 70% of 401(k) plans offered tdfs and the proportion of recently hired participants in their 20s and 30s holding tdfs is 51% and 53%, respectively. the comparable numbers for 2006 are 29.4% and 28.5%, respectively. equally interestingly, 85% and 80% of the account balance of these participants, respectively, is held in tdfs. there seems to be a demonstrable trend towards increasing use of tdfs by younger workers. if investment choices in retirement plans are assumed to be stable, then the expectations of a significant increase in assets of tdfs with the passage of time seems logical. for the year 2014, ici reports that new fund flows into indexed funds increased by 30% in addition to a 93% increase in 2013. a significant portion of these funds was into equity index funds. remarking on the growth ici concludes, “demand for index mutual funds remained strong in 2014.” index funds reflect a low cost tax efficient passive investment philosophy. tdfs are steeped in a similar belief that discourages frequent readjustment of asset allocation and portfolio holdings. it may not be a leap to assume that an investor embracing index funds is likely to be attracted to tdfs and vice versa. given the trend of increasing assets directed towards tdfs, the design of these retirement savings vehicles becomes important to retirement well-being of a significant proportion of the working population. 3. investment theory and glide paths towers watson (2010) reports “… a 2009 watson wyatt research study found that employees categorized as ‘intermediate/ long-horizon’ investors planning to retire in 15 years or more experienced typical median declines of 27% to 37% in the value of their plans. this magnitude of loss was attributable to the large exposure to equity investments (51% to 95% across 2020–2055 target horizon funds evaluated) considered appropriate for investors focused on wealth accumulation. what is troubling, however, is that ‘short-horizon’ investors planning to retire in two to seven years (2010–2015 target horizon funds) suffered a median loss of 21%, and as high as 36% for some funds. these individuals have limited ability to recover from the losses incurred other than by deferring retirement and significantly increasing their rate of retirement saving (p. 1).” d’antona (2008) aptly highlights a similar plight of tdf owners retiring in the period 2008–2010. she also points to the wide range of losses suffered by such investors because of deficient glide paths. the glide path of tdfs has a significant influence on investment outcomes and consequently on retirement well-being. a fund’s glide path primarily refers to the manner in which the fund asset allocation changes over time until the target date is reached or beyond. the evidence on the efficacy of glide path design is discussed next. gomes, kotlikoff, and viceira (2008), study asset allocation in the context of a variable labor supply. they propose that the ability of individuals who do poorly in financial markets can make up the performance shortfall by working more. they demonstrate that the welfare loss is negligible when using a typical life cycle mutual fund. comparatively, there is a 240 s. singh / financial services review 25 (2016) 235–262 significant loss of welfare in investing in stable value funds. they argue in favor of higher equity exposures during retirement as the presence of bond like cash flows in the form of pension benefits provide better risk taking capabilities. pfau and kitces (2014) further endorse the idea of upward sloping glide paths during retirement, that is, those that increase equity exposure during retirement. using monte carlo simulations, they demonstrate that glide paths that increase equities from 30% to 60% during retirement reduce the probability and magnitude of run outs as compared with glide paths that do the opposite that is, reduce equity exposure from 60% to 30% during retirement. later kitces and pfau (2015) combine a rising equity allocation glide path with a dynamic asset allocation scheme. in this scheme, equity exposures are increased when equity markets are deemed undervalued and reduced when overvalued. using actual return data and overlapping periods, they demonstrate that a rising glide path is most suited in overvalued markets. blanchett (2015) points to the fact that the initial condition of financial markets is often ignored in many asset allocation decisions as one is embarking upon, or is close to retirement. utilizing a range of varying market conditions, he concludes that upward sloping glide paths representing increasing allocations to equities seem to perform well in relatively higher return environments and higher allocations to lower risk assets seem to do well in low return markets. basu and drew (2009) are perhaps the first to question the conventional wisdom of reducing equity allocation in tdf portfolios as one nears the target date. they point to the fact that individuals contribute savings to the retirement portfolio on a regular basis and the sum total of the initial portfolio increases with time even if one assumes a rate of return of zero on initial portfolio and contributions. reducing equity allocation in such portfolios at a point in time close to the withdrawal phase is detrimental to portfolio size at terminal date. the difference between the lifecycle and contrarian strategy (one that increases equity allocation closer to terminal date) is mostly driven by performance close to terminal years. the authors do acknowledge the chance of ruin increase under such a strategy and that the behavioral aspects of such a strategy cannot be ignored. pfau (2010) uses monte carlo simulations to illustrate that investors with modest risk aversion will find traditional glide paths more suitable as compared with a fixed allocation. he points to the failure to consider expected utility and, therefore, comparative risk aversion in contrarian glide paths of some studies. contrary to basu and drew, he shows that higher utility is derived from traditional life cycle strategies when the investor wishes to avoid probability of ruin. capitalizing further on the earlier findings, basu, bryne, and drew (2011) also suggest a dynamic asset allocation strategy predicated on portfolio value rather than the age of the participant. the key to such a strategy would be the setting of an accumulation rate and this accumulation rate guides the asset allocation over time. the authors utilize simulations to demonstrate the benefits of such a strategy. yoon (2010) suggests a risk budget based methodology to designing glide paths. under such a system, periodic adjustments to asset allocation are driven by the available risk budget. the advantage of such a strategy as yoon points out is that the asset allocation stays within the risk tolerance at all times. the downside of such a system is the complexity of the system and the inability to maximize returns. spitzer and singh (2011) study the glide path of 22 fund families and classify glide paths utilized by these fund families in two broad categories, “late descent” or “early descent.” 241s. singh / financial services review 25 (2016) 235–262 in late descent glide paths, the percentage of stocks stays constant at a high proportion for the first part of the life cycle and then begins to fall rapidly over the latter part of the investment period. in “early descent” the asset allocation begins with a high proportion in equities and, the percentage of stocks falls gradually over the entire holding period. both strategies are shown to be inferior to a static allocation. the investment company institute (2014) uses three similar classifications to illustrate glide paths. the first, “allocates 50% of its assets to equity as of the target date (‘ret’) and reaches its most conservative allocation (20% equity) 15 years later.” in the second, a constant proportion is maintained for the first 20 years and then the asset allocation glides to the most conservative allocation five years before the retirement date. in the third set of glide paths, a constant allocation is maintained in the first 20 years and the asset allocation is gradually reduced to more conservative through 30 years into retirement. even with same securities in all portfolios in the preceding strategies, the investment outcomes are likely to be different because of asset allocation and its adjustment. the other factor affecting the investment outcome in these funds is the investment strategy and security selection in the portfolio. the primary distinction is between the uses of actively and passively managed funds in the lineup. findings of the ici study broadly confirm results reported by similar studies related to glide path efficacy. arnott, sherard, and wu (2013) are critical of traditional glide paths. if the goal is to maximize wealth accumulation during earning years and minimize longevity and lifestyle risk during retirement, they argue that traditional glide paths fail on both accounts. the reasons for the failure are suboptimal asset class exposure, inefficient risk and return balance including lack of diversification and an assumption of constant risk premiums. estrada (2014) validates their findings in a global context through analysis of returns of 19 countries covering a period of 110 years. he finds that ten alternate strategies that include five different contrarian and five static equity allocations with varying holding periods, tend to outperform traditional glide paths. the interesting query raised by both sets of authors is what constitutes risk for an individual investor. by traditional measure of volatility as a proxy for risk, tdfs tend to provide greater certainty of outcomes but as shown many times before, other strategies tend to provide higher end of period wealth accumulation and lower run outs (longevity risk). idzorek, stempien, and voris (2013) provide a glide path stability score (gpss) to each fund family offering tdfs. they calculate this score by comparing the consistency in asset allocation of different vintage tdfs from the same family. a lower glide path stability score indicates that the glide paths of different vintages of the same family have more variability in asset allocation for the same time remaining to retirement. they conclude, “while glide path changes are not necessarily bad, we believe that unannounced and unjustified changes in glide paths should be viewed with extreme scrutiny, given that investors and sponsors select these investments based on expectations of risk (p. 81).” optimal glide paths during retirement have also received some attention from researchers. spitzer and singh (2008) use a bootstrap simulation to study the shortfall risk of tdfs during the retirement years. shortfall risk is defined as the probability of running out of money during retirement. they classify tdfs into three types of glide paths: steep, gentle, and fixed 25/75. they show that all three glide path strategies have higher shortfall risk than a constant 50/50 allocation. they urge the designers of target-date mutual funds to “rethink their asset allocation during retirement (p. 151).” kalman (2011) uses a similar bootstrap 242 s. singh / financial services review 25 (2016) 235–262 methodology to confirm the findings of spitzer and singh. he also demonstrates that after accounting for equity risk premium, a bond heavy portfolio in retirement reduces the probability of run outs. the question of whether glide paths should end at retirement or extend through retirement remains unresolved. clark and hood (2009) recognize the need for maintaining a constant real withdrawal rate during retirement. to that end, they focus on designing tdfs that are especially suitable for the withdrawal phase of the life cycle. such tdfs will allocate higher and a constant proportion to riskier assets to accomplish the goal of constant real withdrawals during retirement. this brings up a worthwhile question: how is the choice of various glide paths supported by investment theory? two broad inter-related theoretical constructs are used to examine this question: (1) life-cycle investing incorporating human capital and, (2) liability driven investing. these are often referred to frequently in the context of glide paths utilized by tdfs. a discussion of each theory in the context of glide paths is presented next. both lines of examination assume an underlying adherence to mean-variance optimization and portfolio theory constructs, albeit with constraints in both cases. 3.1. life-cycle investing incorporating human capital the life cycle theory of consumption and portfolio choice posits that at any point in time in a human’s life, one’s endowment consists of financial wealth and human capital. the human capital is the present value of lifetime earnings, which is stochastic. this variability in earnings results from controllable and uncontrollable factors. increasing human capital through skills and education and the ability to work the number of hours chosen along with capacity to retire early can be treated as controllable factors. uncontrollable factors include being forced to work fewer than intended hours or not being able to work at all because of unemployment, sickness or premature death. over a lifetime, an individual converts human capital into financial capital and it is assumed that one has exhausted all human capital at retirement. for a detailed exposition, see mayers (1972), williams (1978) and merton (1969). as a practical application of this theory, financial advisors usually look at a human life as compilation of four consecutive stages: accumulation, consolidation, withdrawal, and gifting. the first stage starts when the individual has accumulated some human capital through training and/or education and is ready to start the conversion of this capital into financial capital. most analysis and research assumes an initial endowment of zero for financial capital at the start of this stage of the life cycle. this is also perhaps the longest of the four stages of the life cycle that starts at around 22 to 25 years of age and continues to about 10 years before retirement. the consolidation phase usually follows the accumulation phase. in this phase, the preparation for retirement begins. the primary activity during this stage is the transfer of financial capital to less risky investments, since there is a diminished ability to recover from a significant loss of portfolio value because of a shortened time horizon. after the consolidation phase, the depletion of the accumulated financial capital starts in the withdrawal phase. the final stage of the life cycle is the gifting phase where one plans for the inevitable. 243s. singh / financial services review 25 (2016) 235–262 while a significant number of tdfs broadly adhere to the preceding life cycle framework, there is a wide disparity in terms of period allocated to each stage of the cycle. while some funds move the asset allocation to bonds quite early in the cycle, others maintain a constant exposure to fixed income assets over extended periods. research themed around human capital and life cycle investing relevant to tdfs is discussed next. viciera (2009) describes a case where an individual has a stable job that implies steady cash inflows. the employment prospects and size of compensation are also not highly correlated with the performance of the stock market. these resemble cash flows from a bond more than common stock. this implies that the investment portfolio of such an individual should hold a high proportion of stocks. as one gets closer to retirement, the bond like component of one’s portfolio (human capital) is depleted. this creates the need to increase actual bonds in this integrated portfolio. viceira also indicates that in cases where human capital is volatile and strongly positively correlated to the stock market, equity exposure should be limited. finally, since expected returns change over long holding periods, occasional adjustments to glide paths may be useful. he concludes that target date investing is consistent with the human capital approach. for long holding periods viciera (2009) finds that having a tdf in a pension plan as compared to a money market fund alone results in higher utility for the participant. another suggestion for better tdf design implies a significant exposure to real assets in the asset allocation. funds with higher allocation to tips, therefore, are superior in design in viciera’s view. with respect to accumulation years, shiller (2005) makes some interesting observations. the optimal asset allocation between human and equity capital is dependent on the correlation between the returns of these two assets. he notes that these correlation estimates vary over a broad range. he also points to recommended equity exposures as a proportion of portfolio value in the literature, of as low as 20% to as high as 300% for young workers depending on the assumptions made regarding the relationship of human capital and equity returns. a comprehensive examination of the influence of labor income on the portfolio decision is provided by cocco et al. (2005). this provides a good basis for a discussion on the optimal design of glide paths as related to tdfs. in the presence of labor income that has low correlation to equity returns they find support for decreasing proportion of equity investments as one ages. an interesting find was that the most significant welfare loss of about 2% per year occurs when one ignores labor income in the portfolio decision. the loss is less significant when one ignores labor income risk only. the most significant loss in welfare arises from “disastrous labor income realization,” meaning unemployment or disability unsupported by safety nets. in light of the findings, glide paths that have declining equity exposures over time, that incorporate nontraded labor income in the asset allocation choices can result in higher welfare outcomes. bodie and treussard (2007) provide additional insight into the role of riskiness of human capital in determining an optimal glide path. they classify certain individuals as “natural tdf holders.” these are individuals with risky human capital, that is, their human capital betas are high. such individuals will experience substantial welfare gains if offered a relatively safe tdf. to reflect the conversion of an individual’s portfolio from human capital 244 s. singh / financial services review 25 (2016) 235–262 to financial capital and the translation of human capital from “stock like” to “bond like,” the glide path should be “humped”; not the traditional linear reduction with age. boscalajon (2011) provides a discussion of “critical wealth,” a point where financial and human capital is equal. he argues that it is from this point on that a systematic transfer to less risky financial assets should start taking place. utilizing monte carlo simulations and utility functions that have a coefficient of risk aversions ranging from 1 to 10, pfau (2011) finds support for glide paths that adhere to traditional strategy of high initial equity allocation and declining risky assets over time as utility maximizing. as mentioned earlier, the value of human capital is the present value of expected future earnings. using bureau of economic analysis (bea) data and yields on investment grade corporate bonds in various industries, blanchett and straehl (2015) attempt to quantify human capital in 12 industries. the authors treat the estimate of the return and risk of human capital across industries as a separate asset class. these then constitute inputs in a portfolio optimization scheme along with other traditional asset class returns. as might be expected, they find that “the optimal equity allocation decreases with age, riskier employment, and riskier homeownership, whereas it increases with guaranteed pension income (p.1).” another and more recent stream of thought affecting glide path design is liability driven investing (ldi). it is derived from defined benefit pension plan design. under such a strategy, the determination of retirement liabilities is first made and then a glide path is designed to optimally meet them. the body of knowledge as related to glide path design motivated by ldi is discussed next. 3.2. liability driven investing ldi derives its inspiration from the management of assets to meet future liabilities, akin to the management of defined benefit plan assets. examples of prominent practitioners of ldi are defined-benefit plan sponsors and insurance companies. traditional portfolio theory assumes that the optimal portfolio is independent of investor’s risk preferences markowitz (1952), sharpe (1964). risk aversion is incorporated through the proportion of risk free holdings of the investor. ldi on the other hand, focuses on designing portfolios to meet future liabilities. it is an individual exercise for retirement portfolio design and management as each individual liability set is different. a defined-contribution plan can be managed similarly if one considers retirement spending needs as a liability. under ldi, for tdfs the goal in building a retirement portfolio, is not a maximization of portfolio expected returns given a level of risk, but rather to ensure that the portfolio has sufficient assets to support retirement withdrawals and that it does not run out of money during the lifetime of the investor. minimizing longevity risk is the priority under such a strategy. the goal is to attain a set of real cash flows that will last with certainty throughout retirement. thus, inflation protection during retirement years is an embedded objective of all glide paths in a ldi based framework. in the design of glide paths, attaining this goal often takes precedence over portfolio optimization. according to meder (2012), “the first order asset allocation decision is no longer focused on the split between equities and core fixed income but focused on deriving the split between a return-seeking asset (rsa) and a liability hedging assets (lha) component. the rsa 245s. singh / financial services review 25 (2016) 235–262 component seeks to generate returns in excess of the expected liability return (growth in the present value of the liability attributable to the passage of time), similar to the discount rate on the liability (p. 117).” while the preceding comment is in the context of a defined benefit pension, it provides a unique manner of looking at glide paths of tdfs. it provides a dual goal in the accumulation phase of the investing horizon; first, to hedge the anticipated liability and the second, to provide growth to the portfolio. idzorek (2008) provides a detailed discussion on designing tdfs that incorporate liability relative portfolio optimization. if an individual’s retirement expenses can be thought of as a set of real cash flow liabilities, then a retiree’s portfolio during the saving years is optimized subject to this liability constraint. this liability constraint is implemented through a short position in a portfolio of tips. idzorek provides a comparative visual for such optimization vis-a-vis an asset only optimization. the significant difference is the heavy relative overweighting in real assets like tips and commodities in such portfolios that are subject to the liability constraint of a short position in tips. another advantage of liability relative investing is the determination of the fund status. this is analogous to an underfunded or over funded status of a defined benefit plan. if a retiree finds early on in their life of the underfunded status, and they are responsible about it, then there may be an opportunity to alter their savings-consumption mix accordingly. there seems to be sufficient support for designing glide paths that embrace meeting retirement liabilities as opposed to optimizing asset only portfolios as the primary goal. it may be possible to jointly accomplish both asset only optimization and ldi, but in the presence of constraints, for example long only portfolios, the task becomes quite challenging. additionally, optimization processes are extremely sensitive to the value of inputs, for example, a small change in volatility causes a significant shift in the optimal portfolio. in light of the preceding, a ldi only based approach, where portfolio optimization is a secondary constraint to designing glide paths might be worth pursuing. whitten and thuerbach (2015) provide an illustration of using ldi driven glide paths. the goal is to provide a level of real income during retirement. using a glide path that has a higher proportion of traditional inflation hedge assets (tips � commodities � real estate) and lower proportion of traditional equities over time, through monte carlo simulations, they demonstrate that such glide paths provide narrower distributions and lower value at risk (var) across all retirement horizons. it is important to point out that the analysis is to retirement and does not extend through retirement. anecdotally, it seems that ldi is garnering increasing interest of practitioners. in addition to pimco, dimensional funds advisors (2016), a mutual fund company, recently introduced tdfs of various vintages premised on ldi. a review of the promotional literature and prospectus of these funds provide a good example of application of the ideas discussed previously in this section. for example, the 2005, 2010, and 2015 vintages at the end of may 31, 2016 held about two-thirds of their assets in tips. this is a good illustration of implementation of the recommendations of whitten and thuerbach (2015). a summary of the research on the human capital aspects and liability driven investing as pertaining to tdfs appears in table 2. for comparative evaluation and justifying changes to existing tdf design, it becomes important to measure outcomes from tdf investing. tdfs have now been in existence for 246 s. singh / financial services review 25 (2016) 235–262 t ab le 2 t ar ge tda te fu nd re se ar ch su m m ar y a ut ho r( s) y ea r t op ic fo cu s pu bl ic at io n k ey fin di ng /c on tr ib ut io n a gn ew et al . 20 03 a na ly si s of 40 1( k ) po rt fo lio s a m er ic an e co no m ic r ev ie w pa rt ic ip an t as se t al lo ca tio n ne gl ec t an d ra re re ba la nc in g. t an g et al . 20 10 jo ur na l of pu bl ic e co n. o pt im al po rt fo lio of fe ri ng s un do ne by po or pa rt ic ip an t ch oi ce . l an ds be rg 20 07 t d f m ar ke t an d re gu la to ry en vi ro nm en t j of d ef er re d c om p. e nr ol m en t pr oc ed ur es an d re g. re lie f pr om ot e as se t gr ow th . l an ds be rg 20 14 j of p. pl an . & c om pl ia nc e fi du ci ar y re sp on si bi lit y re la te d to t d fs fo r pl an sp on so rs . r ic ha rd s et al . 20 14 t ia a -c r e f d ia lo gu e t d fs ga rn er in g m os t re ce nt re tir em en t co nt ri bu tio ns . b al du zz i et al . 20 15 n b e r w or ki ng pa pe r si gn ifi ca nt ri sk ta ki ng di ve rs ity in lo w er m ar ke t sh ar e fu nd s. b as u & d re w 20 09 g lid e pa th de si gn j of po rt fo lio m an ag em en t pe rf or m an ce in la te ye ar s do m in at es re tir em en t ac cu m ul at io n. c la rk & h oo d 20 09 jo ur na l of in ve st in g e qu iti es to in cr ea se in re tir em en t fo r co ns ta nt re al re tu rn s. pf au 20 10 fi na nc ia l se rv ic es r ev ie w l if e cy cl e fu nd s pr ov id e hi gh er ut ili ty if ri sk av er si on is m od es t. pf au 20 11 a pp lie d e co no m ic l et te rs d et er m in e gl id e pa th s ba se d on co ef fic ie nt of ri sk av er si on . y oo n 20 10 j of a ss et m an ag em en t r is k bu dg et ba se d gl id e pa th s in st ea d of de cl in in g ri sk to le ra nc e. b as u et al . 20 11 j of po rt fo lio m an ag em en t d yn am ic al lo ca tio n gu id ed by ac cu m ul at io n ra te in st ea d of tim e. a rn ot t et al . 20 13 jo ur na l of r et ir em en t t ra di tio na l gl id e pa th s fa il to lim it lif es ty le an d lo ng ev ity ri sk . id zo re k et al . 20 13 jo ur na l of in ve st in g g lid ep at h st ab ili ty sc or e (g ss ) to ex am in e gl id e pa th dr if t. pf au & k itc es 20 14 j of fi na nc ia l pl an ni ng r ec om m en d in cr ea si ng eq ui tie s du ri ng re tir em en t ye ar s. b la nc he tt 20 15 j of fi na nc ia l pl an ni ng si gn ifi ca nc e of in iti al m ar ke t co nd iti on s at re tir em en t. k itc es & pf au 20 15 j of fi na nc ia l pl an ni ng c on di tio na l gl id e pa th s ba se d on eq ui ty m ar ke t va lu at io ns . 247s. singh / financial services review 25 (2016) 235–262 t ab le 2 (c on tin ue d) a ut ho r( s) y ea r t op ic fo cu s pu bl ic at io n k ey fin di ng /c on tr ib ut io n n ag en ga st , b uc ci , & c oa ke r 20 06 t d f pe rf or m an ce ev al ua tio n po pp in g th e h oo d t d fs la ck im ag in at io n an d ca nn ot ou tp er fo rm m ar ke t. su rz & is ra ls en 20 07 j of pe rf . m ea su re m en t t d fs fa il to ou tp er fo rm pu re t ar ge tda te in de xe s. sp itz er & si ng h 20 08 fi na nc ia l se rv ic es r ev ie w t ra di tio na l gl id e pa th s la g co ns ta nt m ix po rt fo lio . b la nc he tt et al . 20 10 j of pe ns io n b en efi ts in co rp or at in g ri sk lo w er s di sp er si on an d re tir em en t fa ilu re . d ol vi n et al . 20 10 j of fi na nc ia l pl an ni ng g lid e pa th s ou tc om es ar e eq ui va le nt to st at ic as se t al lo ca tio ns . k al m an 20 11 fi na nc ia l se rv ic es r ev ie w b on d he av y ri sk pr em iu m ad ju st ed po rt fo lio s re du ce ru n ou ts . l ip to n & k is h 20 11 fi na nc ia l se rv ic es r ev ie w l if e cy cl e fu nd s ad d lit tle va lu e on ri sk ad ju st ed ba si s. b ar e & g re ve s 20 13 v ie w po in t r us se ll r es . fo cu s on re tu rn so ur ce s an d de -e m ph as iz e re la tiv e pe rf or m an ce . c as si dy et al . 20 14 j of po rt fo lio m an ag em en t w id el y di sc lo se d be nc hm ar k en ab le s in st itu tio na l be st pr ac tic es . c ha ng et al . 20 14 fi na nc ia l se rv ic es r ev ie w t d fs pr ov id e be tte r w el fa re w he n co m pa re d to lif e cy cl e fu nd s. e st ra da 20 14 j of po rt fo lio m an ag em en t c on fir m s a rn ot t et al ., u s fin di ng s fo r 19 ot he r co un tr ie s. t an g & l in 20 15 j of a ss et m an ag em en t in ap pr op ri at e be nc hm ar k ch oi ce re su lts in w el fa re lo ss of 67 % . w ill ia m s 19 78 in co rp or at in g hu m an ca pi ta l th eo ry in t d fs jo ur na l of b us in es s in iti al pa pe r on in co rp or at in g hu m an ca pi ta l in po rt fo lio ch oi ce . c oc co et al . 20 05 r ev . of fi na nc ia l st ud ie s ig no ri ng hu m an ca pi ta l re su lts in a 2% w el fa re lo ss pe r ye ar . sc hi lle r 20 05 t he e co no m is t’ s v oi ce h um an an d fin an ci al ca pi ta l re tu rn s co rr el at io n ef fe ct . b od ie et al . 20 07 fi na nc ia l a na ly st s jo ur na l r ec om m en d hu m pe d gl id e pa th s to in co rp or at e hu m an ca pi ta l. g om es et al . 20 08 a m er ic an e co n. r ev ie w a ss et al lo ca tio n in th e pr es en ce of va ri ab le la bo r su pp ly . 248 s. singh / financial services review 25 (2016) 235–262 t ab le 2 (c on tin ue d) a ut ho r( s) y ea r t op ic fo cu s pu bl ic at io n k ey fin di ng /c on tr ib ut io n v ic ei ra 20 09 l if e c yc le fu nd s (c ha pt er ) h um an ca pi ta l is bo nd lik e so m or e bo nd s cl os er to re tir em en t. b os ca ljo n 20 11 fi na nc ia l se rv ic es r ev ie w in tr od uc es th e co nc ep t of cr iti ca l w ea lth th at af fe ct s al lo ca tio n. b la nc he tt et al . 20 15 fi na nc ia l a na ly st s jo ur na l q ua nt if y hu m an ca pi ta l re tu rn an d ri sk in 12 in du st ri es . id zo re k 20 08 im m un iz in g re tir em en t lia bi lit ie s u np ub lis he d m an us cr ip t l ia bi lit y co ns tr ai nt m od el ed as a sh or t po si tio n in t ip s. m ed er 20 12 jo ur na l of in ve st in g po rt fo lio sp lit am on g lia bi lit y he dg in g an d re tu rn se ek in g as se ts . w hi tte n et al . 20 15 pi m c o so lu tio ns h ig he r ra tio of in fla tio n he dg e as se ts re su lt in lo w er v a r s. 249s. singh / financial services review 25 (2016) 235–262 sufficient time to provide return data for such a preliminary assessment. a discussion on benchmarking and evaluation of tdf performance is presented next. 4. benchmarking and performance evaluation in this section, a discussion on two aspects of tdf performance is presented. in the first, a framework for benchmarking and performance attribution of tdfs is discussed. in the second, a summary of accumulated empirical evidence on the performance of tdfs is provided. before the discussion on performance evaluation and attribution of tdfs, an important fact regarding tdf strategies need to be pointed out. most tdf strategies assume a long-term investment horizon. many tdfs, for example, those of 2050 or 2055 vintage, assume a holding period of 40 years or more, and this is for just “to” retirement tdfs not “through” retirement. the long horizon becomes even more significant when one takes into account the finding that returns in the last few years of the investment horizon are the most significant, as pointed out by basu et al., therefore, any performance evaluation methodology should take into account the assumed long-term investment horizon of the investor. it seems that the traditional one, three, or five-year returns might not be adequate to evaluate tdf performance.2 a significant portion of dc plan assets is tdfs, and as mentioned in the earlier part of the paper, assets invested in tdfs are expected to grow. monitoring duties of the plan sponsor usually involve performance evaluation of plan offerings. cassidy et al. (2014) emphasize the need for a benchmark based performance evaluation in defined contribution plans. the benchmark enables better communication between plan sponsors and investment managers. it affords the plan participants benefits of institutional best practices. a benchmark also lets the plan sponsor clearly define risk to the investment managers. cassidy et al. recommend separate benchmarks for accumulation and decumulation stages. the subject of benchmarking and attribution of tdf performance is best understood in the attribution framework borrowed from bailey, richards, and tierney (2007). let m be the return that represents a neutral asset allocation and passive return for underlying asset classes of a tdf. as applicable to tdfs the benchmark, m captures the utility function of the investor including an appropriate degree of risk aversion. neutral asset allocation in this context would result in minimum acceptable welfare derived from the tdf investment strategy. for an actively managed tdf, a portfolio manager can add value and improve welfare, through three broad sets of decisions. first, the manager can deviate from neutral asset class weights and contribute value to the portfolio. for example, overweighting fixed income securities when relative returns from this asset class are forecasted to be favorable in the view of the manager. let this be denoted by w. next, they can alter allocation to sectors within an asset class. for example, over weighting mortgage-backed securities relative to treasuries. let this be denoted by s. finally, security selection can be a source of value. let this be denoted by a. a represents over and underweighting of securities relative to the benchmark. assuming interaction affect between s and a to be minimal, the portfolio return p can then be represented as, 250 s. singh / financial services review 25 (2016) 235–262 p � m � w � s � a (1) for a tdf with underlying index funds as investment options, both s and a are required to be 0, so the portfolio return becomes, p � m � w (2) in cases, where no deviation from a neutral allocation is permitted, but underlying funds are actively managed, p � m � s � a (3) in cases, where no deviation from a neutral allocation is permitted, and underlying funds are indexed, p � m (4) in the case of eq. (4) the role of the manager is largely administrative as they can exercise no discretion in asset allocation or security selection decision. a couple of additional interesting points about the above frame work. in most cases, w is restricted to a range of real life funds. second, being that a tdf is essentially a fund of funds, s and a are often beyond the control of the tdf manager and a domain of the manager of the underlying fund. assuming a similar setup, bare and greves (2013) recommend that investment performance and participant success be segregated in the performance evaluation of tdfs. as participant success in accomplishing investment goals is individual and varies across participants, the more objective endeavor would be to study investment performance. the other recommendation made by the authors is to deemphasize peer relative performance evaluation. the long-term asset allocation of the fund as reflected by the fund’s glide path represents the tdf’s strategic allocation mix. if the fund relies on active management, then there is an additional component of security selection and tactical return if the fund mandate permits alteration of asset allocation, perhaps within a range. the focus should be upon identifying the sources of risk and return of the tdf portfolio, according to the authors. bare and greves (2013) suggest a three-pronged approach to performance attribution. first, developing a simple benchmark representing the growth and capital preservation components of the fund. second, determining a composite benchmark representing the returns of the specific target components of the fund. for example, performance of real assets in a neutral allocation. together, this is analogous to m in eq. (1). finally, a return comparison between composite and simple benchmark, and then between the fund return and the simple benchmark to isolate the value added by the manager through security selection and tactical choices. this is equivalent to w�s�a in eq. (4). with this framework in place, it is instructive to examine how the benchmarking tools available stand up to fundamental principles of a good benchmark. a number of index providers have created target date indexes for benchmarking and performance evaluation. according to bailey, richards, and tierney (2007), a valid benchmark should possess the following properties: it is (1) unam251s. singh / financial services review 25 (2016) 235–262 biguous, (2) measurable, (3) appropriate, (4) reflective of current investment opinions, (5) prescribed in advance, (6) investable, and (7) accountable. at present, there are multiple target-date indexes available for benchmarking tdf performance. descriptive literature provided by each index provider highlights the superiority of each index design. a summary of these indexes appears in table 3. please note that despite reasonable efforts, the author of this paper is unable to find up to date information on some of these indexes. in the table, the prominent feature of each benchmark is highlighted. to provide the reader with an academic assessment of index qualities rather than comparative positives and negatives, an aggregate analysis based on bailey, richards, and tierney’s seven properties of a valid benchmark are presented. as can be observed from table 3, a wide array of investment philosophies is found in the design of target date indexes. there are many different glide paths used by tdfs. similarly, there are wide and varying approaches to the underlying glide paths tracked by the indexes. note that the tracking error of any tdf will likely be different by benchmark, given the wide divergence in the underlying glide path of each index. from the perspective of plan sponsors, there seem to be sufficient historical returns to allow back testing and to determine which index best fits their preferred glide path. all indexes shown in table 3 take into account the need for asset protection in their design and some aim to benchmark real returns close to and during retirement. this recognizes the need to incorporate investor risk tolerance and volatility of human capital related cash flows. a couple of index sponsors have started offering more than one index series for the same target date. the choice of a particular series incorporates investor risk preferences to some extent in the glide path decision. a word of caution regarding the calculation of returns, index rebalancing and reconstitution might be appropriate. some indexes report gross returns, others report returns net of fees, while others report both. the influence on tracking error from this difference in reporting returns should be recognized. all index returns are calculated from underlying indices that have their own rebalancing and reconstitution rules. the impact on returns of this fact is not widely understood or has been investigated by researchers. a wide variation in the frequency of reconstitution and rebalancing of the indexes is revealed in table 3. time between reconstitution and rebalancing can range from one month to one year. too frequent rebalancing can be expensive and may not provide sufficient time to benefit from trending markets, while a period too long between rebalancing may lead to significant deviations from target asset allocations for the portfolio. again, this has not been analyzed nor there is a consensus among researchers on an optimal time between rebalancing and reconstitution. with respect to the seven properties of a valid benchmark, none of the indexes meets all seven. the efficacy of target date indexes is undermined by the fact that they are indexes derived from other indexes. all six indexes are measurable and given sufficient information, the manager whose portfolio is benchmarked can agree to be accountable to any of the six benchmarks. indexes based on consensus cannot be specified in advance and, therefore, are not investable. similarly, equal weighted indexes do not represent the underlying market and are technically not investable. benchmarks based on proprietary algorithms and rebalancing methodologies that are not clearly disclosed, or understood, do not meet the unambiguous criteria. arguably, a benchmark that attempts to incorporate investor risk preferences is more 252 s. singh / financial services review 25 (2016) 235–262 t ab le 3 t ar ge tda te fu nd in de xe s fe at ur es in de xe s c ha ra ct er is tic sa c al la n d ow jo ne s r ea l r et ur n d ow jo ne s g lo ba l d ow jo ne s u .s . m or ni ng st ar l if et im e s& p t ar ge t d at e in de x t d a o n t ar ge t in de xe s r is k pr ofi le va ri et y 1 1 1 1 3 1 4 g lid e pa th s c on se ns us dr iv en c os in e cu rv e ba se d c os in e cu rv e ba se d c os in e cu rv e ba se d h um an ca pi ta l an d lia bi lit y dr iv en c on se ns us dr iv en l os s pr ob ab ili ty dr iv en m od el a ss et cl as se sb 4 4 3 3 4 4 4 in fla tio n pr ot ec tio n r e it s, c om m od iti es , an d t ip s r e it s, c om m od iti es , an d t ip s n o sp ec ifi c al lo ca tio n. t hr ou gh un de rl yi ng in de xe s. n o sp ec ifi c al lo ca tio n. t hr ou gh un de rl yi ng in de xe s. t ip s an d c om m od iti es r e it s, c om m od iti es , an d t ip s r e it s, c om m od iti es , an d t ip s r eb al an ci ng a nn ua l se m ian nu al m on th ly m on th ly q ua rt er ly m on th ly m on th ly in de x re co ns tit ut io n a nn ua l se m ian nu al m on th ly m on th ly a nn ua l a nn ua l m on th ly u nd er ly in g in de xe s m et ho do lo gy v ar io us e qu al w ei gh te d e qu al w ei gh te d e qu al w ei gh te d c ap w ei gh te dc c ap w ei gh te d v ar io us u nd er ly in g in de xe s re co ns tit ut io n m et ho d v ar io us pr op ri et ar y pr op ri et ar y pr op ri et ar y pr op ri et ar y c om m itt ee v ar io us u nd er ly in g co m po ne nt s v ar io us d ow jo ne s in de xe s d ow jo ne s in de xe s d ow jo ne s in de xe s m or ni ng st ar in de xe s an d fu tu re s e t fs an d c om m od ity t ru st m ut ua l fu nd s an d e t fs r et ur ns ca lc . st ar t da te ja nu ar y 20 06 d ec em be r 19 99 ja nu ar y 19 83 ja nu ar y 19 83 d ec em be r 20 01 m ay 20 07 ja nu ar y 19 98 n ot e: a t ab le cr ea te d fr om in fo rm at io n pr ov id ed in m or ni ng st ar , s& p, an d c al la n a ss oc ia te s do cu m en ts . b a ss et cl as se s ar e cl as si fie d as e qu ity , fi xe d in co m e, in fla tio n h ed ge a ss et s, an d c as h. in de xe s w ith th re e as se t cl as se s do no t in cl ud e in fla tio n he dg e as se ts se pa ra te ly . c c ap w ei gh te d in de xe s ar e flo at ad ju st ed . c om m od ity ex po su re is pr od uc tio n w ei gh te d. 253s. singh / financial services review 25 (2016) 235–262 appropriate than the one that relies upon the investor holding suitable levels of cash to incorporate risk tolerance. two of the six indexes attempt this, and, therefore, seem better designed on the appropriateness factor. blanchett and kasten (2010) provide initial support for incorporating risk aversion in index design. they demonstrate that using three broad categories of risk and dynamically adjusting allocation provide lower dispersion of returns and more certain outcomes. therefore, a risk appropriate benchmark may capture a neutral welfare maximizing allocation more appropriately. a proportion of tdfs are supported by underlying indexed funds and etfs that are actively managed. sometimes fund managers have latitude regarding asset allocation in many tdfs. the property of the benchmark being reflective of current investment opinion becomes significant. a manager who has discretion in terms of asset allocation and security selection should understand factors that affect the benchmark and have an opinion with regard to the factors. a majority of tdf strategies is executed through indexed portfolios and predetermined glide paths. under such circumstances, the property of the benchmark being reflective of current investment opinion is not so relevant, as manager discretion is limited. based on the evaluation of the seven benchmarks, once the underlying indexes and their proportions are disclosed, all of the indexes meet the investment opinion criterion. at best, tdf benchmark design is in its infancy and plan sponsors can expect more refinements in time to come. once an appropriate benchmark is selected, it becomes relevant to measure the performance in context of the benchmark. to this point, most performance evaluation has relied on benchmark independent performance evaluation and mostly on simulations. the empirical evidence on tdf performance is presented next. nagengast, bucci, and coaker (2006) study performance of tdfs from six major fund families. they utilize a weighted score of six metrics: structure/strategy, expenses, allocation, performance, and two measures of risk. the authors find the returns from these funds to be in line with the performance of the market. they observe, “the asset allocations of most of the fund families lack imagination.” surz and israelsen (2007) determine that tdfs “failed to measure up to the risk-adjusted performance standards established by the pure target-date indexes.” to arrive at this conclusion, the authors first prescribe the characteristics possessed by an ideal tdf. they also build four benchmarks ranging from “defensive” to “aggressive” depending upon the glide path applicable to each category of tdf. they arrive at the preceding conclusion based on the risk classification of a tdf and performance relative to the benchmark. borrowing performance metrics from the annuity industry, lewis (2008) uses actual annuity quotes to calculate the probability of attaining an income replacement ratio. this enables one to evaluate the efficacy of a particular glide path and the shortfall risk associated with each. in the opinion of the author, this makes comparative risk assessment of glide paths easier and, perhaps less dependent on simulated data as in other studies. using bootstrap simulations, dolvin et al., (2010) find that certain dynamic strategies that reduce equity exposure over time are equivalent to some static strategies. they also conclude that most glide paths follow a 120-age allocation scheme, similar to a 100-age strategy. lipton and kish (2011), highlight the opacity in disclosure and the wide variability in glide paths, management fees and outcomes in life cycle investing. they construct simple 254 s. singh / financial services review 25 (2016) 235–262 rolling indexes based on the widely used asset allocation rule of 100-age in equities. an evaluation of a wide range of target dates and tdfs leads them to conclude “life cycle funds add little value on a risk-adjusted basis” and “that individuals with a long time horizon may want to consider whether the apparent convenience of life cycle funds outweighs the difficulties in measuring and attaining return (p. 92).” in a study of 36 fund families, tang and lin (2015) find significant welfare loss from investing arbitrarily in a tdf. they measure loss from two sources: portfolio selection and glide path. the authors conclude that an inappropriate tdf, one whose glide path does not reflect the risk preference of the investor can cost the investor a loss of as high as 17% in welfare. of this loss in welfare, the most significant contributor to welfare loss of about 67% is attributed to inappropriate glide path choice and the remainder could be attributed to portfolio selection. the authors advocate a risk-based tdf selection strategy. a summary of the accumulated research on performance evaluation of tdfs appears in table 2. some initial evidence on the actual performance of tdfs is now available. in a 2016 morningstar study, holt et al. (2016), calculate the historical 10-year annual average total return on tdfs and compare them to various classes of mutual funds. they find that on a total return basis the tdfs lag funds that include only equities, but do better than categories of bond funds. there has been a tremendous flow of funds into tdfs over the past 10 years, a point made on multiple occasions in this study. returns calculated after taking cash flows in and out of the portfolio might be a better gauge of the returns actually experienced by the investor. when comparing dollar-weighted return, holt et al., find that with the exception of sector and diversified funds, tdfs exceeded the performance of all other categories. the dollar-weighted return for tdfs was 5.16% compared with the all funds average of 4.35%. when comparing dollar-weighted returns of various vintages, as might be expected because of their heavier equity weightings, later vintages outperformed the nearer vintages. now that some actual return data are available, it will be interesting to evaluate performance on a risk-adjusted basis with actual returns rather than relying on simulations. a recommendation in this regard is made in the next section. 5. recommendations this section is divided into three parts, each aimed at a different set of tdf stakeholders. the first set of recommendations is directed towards investors and their financial advisors. the second is for plan sponsors and mutual fund companies and finally and the last set is aimed at researchers and presented as ideas for future research. 5.1. investors and financial advisors in light of the wide variety of tdf offerings, deliberated choices by individuals and their advisors specific to the individual situation can improve investment outcomes. as with other mutual fund recommendations, fees are a significant variable affecting long-term results and the investor or their advisor should closely analyze these. lipton and kish (2011) report front-end loads ranging from 4.75% to 5.75%. they also report that management fees and 255s. singh / financial services review 25 (2016) 235–262 expense ratios decline with the approach of the target date. given due diligence and fiduciary duty requirements, it is important that advisors monitor fund fees over time. economies of scale dictate that as assets under management of a fund increase, the expense ratio of the fund should decline. investors should avoid a fund that increases fees under such circumstances. tdfs at a basic level are fund of funds. the underlying funds have their own fees. many fund families do not charge overlay fees at the tdf level and the investor pays only the fees of the underlying funds. such funds should be preferred to ones that tack on another layer of fees. the total fees paid by the investor over time should be the ultimate consideration though. passive funds have attracted significant amounts of investor funds and indexing as an investing strategy is gaining increasing popularity. the empirical evidence against the long-term performance of actively managed funds and their cost disadvantage do not make them a competitive alternative in a tdf vis-à-vis those with indexed underlying funds. there is growing theoretical evidence that a one-size fits all tdf is not an optimal choice. incorporating risk tolerance of the investor in the tdf choice provides higher welfare. for example, an investor whose human capital return has low correlation to equity market return has perhaps higher risk tolerance in their tdf choices. under such circumstances, investors can consider a tdf whose target date is past the intended retirement year. additionally, a choice that incorporates increased longevity and longer life spans will likely result in higher utility. there are two sets of cash flows that are often ignored in the choice and design of tdfs. these are: (1) the investor’s holdings of real estate in the form of primary residence and (2) defined benefit pension cash flows like social security. the latter is quite like a real asset with “bond like” characteristics. according to jennings and reichenstein (2003) “the financial profession ignores the value of the db plan in calculating asset allocations, it places an implicit value of zero on db benefits (p. 197).” incorporating both of these often-ignored assets in the tdf choice might indicate a different level of risk aversion and, thus, suggest a different asset allocation pederson (2015). the importance of including real assets in tdfs is emphasized in many studies. in ldi, many scholars have modeled the retirement liability as a set of real outflows analogous to the payout on a real annuity. therefore, inflation protection appears to be a primary objective of any sound glide path design. investors and their advisors should favor tdfs that have a healthy dosage of real assets like tips, commodities and other such assets, especially through retirement. the task of finding utility maximizing choices is facilitated when plan sponsors and mutual fund companies offer tdfs designed with this goal in mind. 5.2. plan sponsors and mutual fund companies benz (2015) characterizes tdfs as “blunt instruments.” there is widespread agreement among academics and practitioners that tdf efficiency can be improved by broadly tailoring funds to investor risk preferences. here are some simple suggestions for improving efficiency and there are long-term structural changes that can be affected to improve outcomes for plan participants. anecdotally, numerous defined contribution plans offer tdfs from one provider only. this virtually locks the participant into a glide path that may or may not be relevant to their 256 s. singh / financial services review 25 (2016) 235–262 risk tolerance and overall portfolio objectives. plan sponsors can easily increase the menu choices by offering funds from more than one family. in adding another family of tdfs, it would be ideal to have new ones that have significantly different glide paths from the current choice. while this would increase fiduciary responsibility related to monitoring, it would benefit the sponsor with regard to the fiduciary duty related to “risk appropriateness.” it is acknowledged that the question of picking the fund best suited to the risk aversion function of an employee partially defeats the notion of a single qualified default investment alternative (qdia). however, if this selection is made part of an employee’s onboarding process, it will lead to more desirable welfare outcomes. for a good exposition on fiduciary responsibility related to tdfs for pension plan sponsors see landsberg (2014). the risk profile of the tdf is significantly influenced by the correlation of human capital returns to equity returns (bodie and treussard, 2007; idzorek, 2008; shiller, 2005). bernard (2009) points out that “target-date funds with a retirement date of 2010 have stock allocations from 9.15% to 65% of their portfolios.” prospective buyer of tdfs would appear to have a wide range of risk choices with which to tailor their holdings. plan sponsors should strive to consider work force risk characteristics when offering tdfs to employees. one simple suggestion is to take into account the correlation of employer stock returns to the broader stock market. for example, where returns are highly correlated to stock market returns, in line with recommendations of bodie and treussard (2007), a safer tdf might be appropriate. stempien and zoll (2015) report a growing trend towards custom glide paths by plan sponsors with a unique work force. if the workforce has a similar demographic, for example 85% of the workers are between the ages of 25 to 35, then a custom tdf will have a greater fiduciary duty demonstrability than an off the shelf product. the advice in this regard from bauer, phillips, and white (2009) is twofold: custom offerings with indexed funds as underlying options. the benefit of indexed offerings relate to compliance and due diligence issues while the second of custom tdf address transparency and relevance for the work force requirements. with individualized asset allocation through robo-advisors becoming increasingly likely, it might be possible soon to move from custom target date to individualized tdfs. it may be possible to develop individualized glide paths for participants or at least provide a recommendation based on information provided by an employee. the first step in this regard would be to incorporate risk tolerance information as part of personnel information. with respect to mutual fund families, it might be worthwhile to offer three variants of the same target year, or at the very least two. the glide path can be aggressive, balanced, and conservative or just aggressive and conservative. this will enable investors to tailor risk to some extent without borrowing or short selling in other parts of the portfolio. alternatively, as suggested earlier, this would mitigate the need to seek a tdf of a vintage different from the intended retirement year. it is widely acknowledged in the mutual fund industry that vanguard funds have one of the lower expense ratios in the industry. in bell v. anthem the plaintiff alleges excess fees charged by the anthem 401(k) plan that has about $5 billion in assets. all the funds in the lineup are vanguard funds. the litigation highlights the need for establishing due diligence procedures. tramell (2009) points to the need for monitoring and controlling fund fees as these can significantly affect retirement outcomes for participants. 257s. singh / financial services review 25 (2016) 235–262 5.3. researchers given longevity increases and long retirement horizons, an answer to the “to and through retirement” glide path question is quite vital. this is premised on the basic issue of the responsibility of the employer or the mutual fund to the employee. does it end at retirement or at end-of-life? at this time there seems to be no definitive answer. each set of providers claim their glide path to be superior. researchers should try to answer both aspects of this question definitively. the discussion on glide paths earlier in this paper, create a new set of research possibilities. there have been numerous suggestions on an optimal glide path. blanchett and straehl (2015) present a unified single-period framework for incorporating non-tradeable assets like residential home and human capital in the portfolio optimization process. their analysis provides many insights into the design of next generation glide paths and tdfs. the question of the optimality of a glide path given a utility function that incorporates non-tradeable assets needs to be explored in greater detail and definitively answered. two recent developments in the tdf market have created a host of new research questions; one set is regulatory and the other technological. the first is the new treasury rule that permits the holder of an ira or 401(k) to invest up to the larger of 25% or $125,000 of the portfolio in a longevity annuity. more important, through notice 2014–66, the u.s. treasury (2014) further permits tdfs to offer deferred annuities as a part of the asset allocation, without violating discrimination rules, as long as certain conditions are met. one of the questions examined by researchers previously (pfau, 2014; spitzer and singh, 2008), focuses on the longevity risk faced with different glide paths during retirement. the inclusion of longevity and deferred annuities alters the optimal asset allocation profiles of tdfs, and the actuarial probabilities of longevity risk. a host of new questions need to now be answered. as the permission to offer annuities is fairly recent, research needs to be conducted on the optimal pricing of these annuities and their influence on the risk profiles of tdfs. the second development is the mechanization of investment advice. the infiltration of robo-advisors in the investment advice market provides a fertile ground for the implementation of the idea of individualized glide paths. tdf glide paths that incorporate risk tolerance individually may not be a far-fetched idea. the day may not be far off where an individualized glide path is generated for participants based on the information collected from their personnel file or the initial employment interview. tdf evaluation mechanisms are still in their infancy. there are a number of indexes now available to benchmark tdf performance. each index has strengths and weaknesses. a comprehensive comparative index evaluation would provide useful information to defined contribution plan providers and consultants. moreover, the impact of building indexes that has underlying indexes instead of individual securities as components, needs attention and analysis. one of the other growing and popular approaches to designing glide paths is liability driven investing or ldi. currently most ldi related to tdfs assume the retirement liability to be a set of real cash outflows. to the best of this author’s knowledge, no effort to model stochastic shocks to accumulation or decumulation cash flows, like unanticipated health care costs, has been attempted. in the modeling of glide paths incorporating such situations, an 258 s. singh / financial services review 25 (2016) 235–262 interesting line of inquiry would involve a comparative evaluation of risk tolerance of a group of employees as opposed to the risk capacity needed to meet their retirement liabilities. an effort in this regard, would be a valuable addition to the body of knowledge. 6. conclusion aided by a favorable regulatory environment, tdfs have experienced unprecedented flow of funds over the past eight years. there seem to be two complimentary approaches to the design of tdf glide path. the first is in the theory of portfolio choice in the presence of nontraded assets like human capital and the second is rooted in the ultimate goal of meeting retirement liabilities. most of the empirical evidence on tdfs calls for better design of glide paths. given the long-term horizon of tdf strategies, performance measurement and evaluation remain a challenge even though indexes are now available to benchmark performance. given the nature of tdfs, none of these indexes fully meets ideal benchmark qualities. the empirical evidence on, and the design of, tdfs is still in its infancy. much work still remains. notes 1 during the period 1995 to 2006, tdf assets grew by about $300 billion. since the passing of the ppa act, over the 2007 to 2014 period assets in tdfs increased by about $700 billion. bauer, phillips, and white (2009) predict that tdf assets will increase by another $1.2 trillion by end of 2018. 2 the author would like to thank an anonymous referee for pointing this fact and for suggesting an attribution framework. acknowledgment the author would like to thank john j. spitzer for his review and comments on the article. the author would also like to thank the editor and three anonymous referees for their comments and suggestions. remaining errors and omissions are mine. references agnew, j., balduzzi, p., & sundén, a. (2003). portfolio choice and trading in a large 401(k) plan. american economic review, 93, 193–221. arnott, r. d., sherrerd, k. f., & wu, l. (2013). the glidepath illusion …. and potential solutions. journal of retirement, 1, 13–28. bailey, j. v., richards t. m., & tierney, d. e. (2007). evaluating portfolio performance. in john l. maginn, donald l. tuttle, jerry e. pinto, & dennis w. mcleavey (eds.), managing investment portfolios: a dynamic process (pp. 717–782). hoboken, nj: john wiley & sons, inc. 259s. singh / financial services review 25 (2016) 235–262 balduzzi, p., & reuter j. (2012). heterogeneity in target-date funds: optimal risk taking or risk matching? nber working paper 17886, boston, ma. bare, r., & greves, j. (2013). evaluating target date investment performance. veiwpoint russell research, russell investments monograph, russell investments, tacoma, wa. basu, a. k., & drew m. e. (2009). portfolio size effect in retirement accounts: what does it imply for lifecycle asset allocation funds? journal of portfolio management, 35, 61–72. basu, a. k., byrne, a., & drew, m. e. (2011). dynamic lifecycle strategies for target date retirement funds, journal of portfolio management, 37, 83–96. bauer, d. j., phillips, b. f., & white, j. r. (2009). target-date retirement funds: the new defined contribution battleground. casey quirk and associates (available at http://tinyurl.com/p43ppgs). benz, c., & stipp, j. (2015). morningstar inc. video, what the best ‘in-retirement’ mutual funds can teach you (available at http://tinyurl.com/ol647). bernard, t. s. (2009). target-date funds: seven questions to ask before jumping in. new york times, june 29 (available at http://tinyurl.com/m9lqdr). blanchett, d. m., & kasten, g. w. (2010). improving the “target” in target-date investing. journal of pension benefits, 18, 17–27. blanchett, d. m. (2015). initial conditions and optimal retirement glide paths. journal of financial planning, 28, 52–60. blanchett, d. m., & straehl, p. u. (2015). no portfolio is an island. financial analysts journal, 71, 15–23. bodie, z., & treussard, j. (2007). making investment choices as simple as possible, but not simpler. financial analysts journal, 63, 42–47. boscaljon, b. (2011). determining the “glide path” for target date funds. financial services review, 20, 113–128. burtless, g. (2010). do workers prepare rationally for retirement? in a. drolet, n. schwarz, & c. yoon (eds.), the aging consumer: perspectives from psychology and economics (pp. 103–127). new york: taylor & francis. callan associates (2015). target date fund index primer (available at http://tinyurl.com/p6kwpwz). cassidy, d. p., peskin m. w., seigel, l. b., & sexauer, s. (2014). be kind to your retirement plan – give it a benchmark. journal of portfolio management, special issue, 33–36. chang, r. p., hunter, d., qianqiu l., & saar h. (2014). saving for retirement while having more nights with peaceful sleep: comparison of lifecycle and lifestyle strategies from expected utility perspective, financial services review, 23, 169–188. clark, j. m., & hood, m. (2009). policy implications for modeling next generation of target date funds. the journal of investing, 18, 53–61. cocco, j. f., gomes, f. j., & maenhout, p. j. (2005). consumption and portfolio choice over the life cycle. the review of financial studies, 18, 491–533. d’antona, j. jr. (2008). disappointing performance hits 2010 target-date funds. pensions & investments (available at http://tinyurl.com/mz2l2z). dimensional fund advisors. (2016). target date retirement income fund (available at http://tinyurl.com/hgxk8sa). dolvin s. d., templeton, w. k., & rieber, w. j. (2010). asset allocation for retirement: simple heuristics and target-date funds. journal of financial planning, 23, 60–71. employee benefit security administration. (february 2013). u.s. department of labor, washington, dc. target date retirement funds—tips for erisa plan fiduciaries. p. 2–3. employee benefits security administration. (april 2008). u.s. department of labor, washington, dc. regulation relating to qualified default investment alternatives in participant-directed individual account plans. p. 1–2. estrada, j. (2014). the glidepath illusion: an international perspective. journal of portfolio management, 40, 52–64. gomes, f. j., kotlikoff, l. j., & viceira, l. m. (2008). optimal life-cycle investing with flexible labor supply: a welfare analysis of life-cycle funds. american economic review: papers and proceedings, 98, 297–303. holt, j., yang, j., acheson l., & wasserman, s. (2016). 2016 target-date fund landscape. morningstar research, 68. 260 s. singh / financial services review 25 (2016) 235–262 idzorek, t., stempien, j., & voris, n. (2013). bait and switch: glide path instability. journal of investing, 22, 74–82. idzorek, t. (2008). lifetime asset allocation methodologies for target date maturity funds. unpublished manuscript, ibbotson associates, 22–26. investment company institute. (2015). investment company fact book, 2015. p. 45, p. 230–231. investment company institute. (2015). frequently asked questions about target date or lifecycle funds. p. 2–3: investment company institute, washington, dc. investment company institute. (2014). 401(k) plan asset allocation, account balances and loan activity in 2013, ici research perspective, 20, 40–47. jennings w. w., & reichenstein, w. (2003). valuing defined-benefit plans. financial services review, 12, 179–199. kalman, m. b. (2011). the role of the equity risk premium in the shortfall risk of target-date funds. financial services review, 20, 265–284. kitces, m. e., & pfau, w. d. (2015). retirement risk, rising equity glide paths, and valuation-based asset allocation. journal of financial planning, 28, 38–48. landsberg, r. d. (2014). fiduciary risk management and target date funds. journal of deferred compensation, 19, 86–94. landsberg, r. d. (2011). target date funds and the proposed sec rule on disclosure. journal of pension planning & compliance, 36, 86–94. landsberg, r. d. (2007). an overview of the pension protection act of 2006. journal of deferred compensation, 13, 13–33. lewis, n. d. (2008). assessing shortfall risk in life-cycle investment funds. journal of wealth management, 11, 15–19. lipton, a. f., & kish, r. j. (2011). life cycle funds: lack of disclosure and lack of return. financial services review, 20, 75–94. markowitz, h. m. (1952). portfolio selection. journal of finance, 7, 1, 77–91. mayers, d. (1972). non-marketable assets and capital market equilibrium under uncertainty. in m. jensen (ed.), studies in the theory of capital markets (pp. 223–248). new york: praeger. mcgraw hill financial. (2015). dow jones target date indexes: risk attributes and performance (available at http://tinyurl.com/nsqk3py). meder, a. h. (2012). level 1 ldi: selecting an appropriate benchmark. journal of investing, 21, 117–127. merton, r. c. (1969). lifetime portfolio selection under uncertainty: the continuous-time case. review of economics and statistics, 51, 247–257. morningstar inc. (2015). target date indexes: comparison grid (available at http://tinyurl.com/o42pbdf). nagenast, j. c., bucci, j., & coaker, w. j., ii. (2006). popping the hood: an analysis of major life cycle fund families (2nd ed.). turnstone advisory group llc marina del ray, ca. pederson, n. (2015). rethinking retirement risk (available at http://tinyurl.com/ngj227c). pfau, w. d. (2010). lifecycle funds and wealth accumulation for retirement: evidence for a more conservative asset allocation as retirement approaches. financial services review, 19, 59–74. pfau, w. d. (2011). an optimizing framework for the glide paths of life cycle asset allocation funds. applied economic letters, 18, 55–58. pfau, w. d., & kitces, m. e. (2014). reducing retirement risk with a rising equity glide path. journal of financial planning, 27, 38–45. richardson, d. p., & bissette, b. (2014). trends in premium and asset allocation by tiaa-cref participants: 2005–2011. tiaa-cref institute research dialogue, 112, 23–24. sharpe, w. f. (1964). capital asset prices: a theory of market equilibrium. journal of finance, 19, 3, 425–442. shiller, r. j. (2005). life-cycle portfolios as government policy. the economists’ voice, 2, 1–14. spitzer j. j., & singh s. (2011). assessing the effectiveness of lifecycle (target-date) funds during the accumulation phase. financial services review, 20, 327–341. spitzer, j. j., & singh, s. (2008). shortfall risk of target-date funds during retirement. financial services review, 17, 143–153. 261s. singh / financial services review 25 (2016) 235–262 stempien, j., & zoll a. (2015). morningstar inc. video, what’s behind target-date funds’ popularity? (available at http://tinyurl.com/owznmlz). surz, r. j., & israelsen, c. l. (2007). evaluating target date lifecycle funds. the journal of performance measurement, winter, 62–70. tang, l., & lin y.-t. (2015). the efficiency of target-date funds. journal of asset management, 16, 131–148. tang, n., & mitchell, o. (2010). the efficiency of sponsor and participant portfolio choices in 401(k) plans. journal of public economics, 94, 1073–1085. torregrosa, p. t., & koenig, d. a. (2008). target date funds: the case for a more conservative glide path. institutional investor guide to defined contribution and defined benefit services, wells fargo bank. towers watson. (2010). defined contribution: lessons after the crisis (available at http://preview.tinyurl.com/ nqorkdp). tramell, s. (2009). glide path. cfa magazine, jan-feb 2009, 36–43. united states treasury. (2014). treasury issues guidance to encourage annuities in 401(k) plans (available at http://tinyurl.com/jr8nc82). vanguard group. (2015). vanguard target retirement funds prospectus. january 27, 62x. viceira, l. m. (2009). life cycle funds. in a. lusardi (ed.), overcoming the saving slump: how to increase the effectiveness of financial education and saving programs (pp. 140–178). chicago, il: university of chicago press. whitton, b., & thuerbach, k. (2015). pimco’s glide path: an outcome focused design. pimco featured solution, march, 2–8. williams, j. (1978). risk, human capital, and the investor’s portfolio. journal of business, 51, 65–89. yoon, y. (2010). glide path and dynamic asset allocation of target date funds. journal of asset management, 11, 346–360. 262 s. singh / financial services review 25 (2016) 235–262 academy of financial services officers president tom potts baylor university president-elect executive vice president-program terrance k. martin winston-salem state university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations michelle cull western sydney university vice president-mktg & public relations shawn brayman financial planning research consultant immediate past president inga timmerman university of north florida editor, financial services review terrance k. martin winston-salem state university directors jason andreson university of kansas jasmine fang massey university norah feng massey university john garble university of georgia wookjae heo purdue university phillip gibson winthrop university barry mulholland university of akron utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. institutional membership to academy of financial services is $250 and includes the four annual issues of fsr. how to submit: there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academyfinancial.org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor terrance k. martin at martintk@wssu.edu. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2023 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 31, no. 1, 2023 editor terrance k. martin, winston-salem state university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $250. academic subscription price is $125 and is available by joining the academy of financial services. academic subscription includes four emailed online issues of financial services review each year. hardcopies of the journal are available at additional cost. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. editorial office: terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. advertising information. those interested in advertising in the journal should contact terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu printed in the usa © 2023 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president robert moreschi virginia military institute president-elect swarn chatterjee university of georgia executive vice president-program janine scott university shepherd vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations philip gibson winthrop university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president thomas coe quinnipiac university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university terrance martin university of texas–rio grande valley martin seay kansas state university past presidents thomas coe, 2015-2016 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial.org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2017 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 26, no. 4, 2017 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2017 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. strategic complexity in investment management fee disclosures leslie a. muller, ph.d.a,*, john a. turner, ph.d.b aassistant professor of economics, grand valley state university, 50 front street sw, grand rapids, mi 49504, usa bpension policy center, 3713 chesapeake st. nw, washington, dc 20016, usa abstract this article develops a measure of complexity of fee disclosures, based on previous work assessing the grade level of language, and validates that measure through a survey where students are asked to independently rate the complexity of fee disclosures. in addition, the article hypothesizes that high fee providers are more likely to engage in strategic complexity in fee disclosures than are low fee providers. we hypothesize that high fee providers use strategic complexity to take advantage of the lack of financial sophistication of many people, making it difficult for them to compare fees across service providers and to understand the level of fees they are paying. © 2016 academy of financial services. all rights reserved. jel classification: g02; g2 keywords: fees; complexity; financial literacy; disclosures 1. introduction while much attention has been given to financial education for pension participants and their lack of financial sophistication (e.g., lusardi and mitchell, 2007; mccarthy and turner, 2000), less attention has been given to the quality of information they receive from financial service providers. efforts by pension participants and other investors to learn about financial market products and services may be offset by efforts by financial * corresponding author. tel.: �1-616-331-7473; fax: �1-616-331-7421. e-mail address: mullerle@gvsu.edu (l. a. muller) financial services review 25 (2016) 215–234 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. service providers to obfuscate through complexity in fee structures and in financial disclosures concerning fees. while transparency is widely viewed as desirable in financial disclosures and in the features of financial products, to take advantage of the lack of financial sophistication of their clients, some advisers may engage in strategic complexity. with strategic complexity, they structure their fees and fee disclosures in complex ways. for example, they may disclose important information in footnotes or use terminology that is not commonly understood by pension participants. abeler and jäger (2015) comment that there is growing evidence that people often do not respond optimally when they are faced with complex financial information. complexity is a multidimensional strategy that we hypothesize is used by some investment management or advisory companies to make it difficult to compare advisory services on the basis of fees, and to determine what level of fees a prospective investor would be paying. the strategy takes advantage of the lack of financial sophistication of many investors, but in some instances even sophisticated investors may not be able to make comparisons or obtain accurate fee information based on the disclosures provided on the internet. “financial sophistication” can come about through taking specific courses (e.g., finance class in college), life experience in making financial choices, and/or working in a position that requires financial decision-making. strategic complexity presumably reduces competition based on fees in the market for advisory services and permits service providers to charge higher fees. prices and fees are essential information in the functioning of the market for a product or service. strategic complexity in the disclosure of fees is presumably designed to make a market less competitive to increase the income of service providers. strategic complexity raises search costs. it may explain in part the limited search that many people make when shopping for financial services and products because the complexity in fee disclosures reduces the benefits from searching. data from the survey of consumer finances provide information about the shopping practices of americans for financial services. in 2010, based on a self-assessment of the effort made to shop for investments that offered the best terms, 55% of families reported making a moderate effort, 23% reported shopping a great deal for the best terms on investments, and 21% reported not shopping at all (bricker et al., 2013). 2. this article we hypothesize that low fee providers of advisory services use less complexity in their fee disclosures than high fee providers. strategic complexity in fee disclosures makes it difficult for people who are not financially sophisticated to compare fees across service providers. the article develops analytical tools for analyzing complexity in fee disclosure. we do this by using an expanded metric of complexity in language to evaluate the complexity of a sample of disclosures, and then test this metric by asking college business students to compute hypothetical fees and to rate the complexity of the fee disclosures. 216 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 while fees are charged in a wide variety of ways, the traditional and most widely used approach is to charge fees based on a percentage of the assets being managed. this approach accounts for 85% of the fees received by advisory firms (maxey, 2011). we analyze fee disclosures for ongoing investment advisory or management services that are available over the internet. data from the survey of consumer finances indicate that the internet is used for information on investing by 80% of the population age 35 and younger (bricker et al., 2012). a number of major companies do not disclose their fees for financial advisory services over the internet, but instead require that the client contact them to obtain that information. for this reason, this analysis of fee disclosures is for a truncated sample, which is the sample of companies that disclose fees over the internet. the article begins by discussing related literature. it then examines the issue of complexity in fee disclosures by adopting a methodology that is used to assess the grade level of writing and expanding it to measure the complexity beyond simple sentence structure and word choice. it applies this metric to all of the 10 largest advisory companies that provide fee disclosures over the internet, plus a sampling of other companies. because many companies do not disclose their fees over the internet and because of the time consuming nature of the process of finding the fees and then rating the associated disclosures, we have a small sample of fee disclosures (n � 10). to validate this expanded metric, we survey a sample of 618 college business students for an independent rating of the disclosures. finally, the article presents concluding comments. 3. literature review this article on complexity in fee disclosures relates to several areas in the economic literature. in related work in computational linguistics, loughran and mcdonald (2013) examine the readability or complexity of financial disclosures. more closely related to the hypotheses of our article, li (2008) argues that business managers attempt to hide the poor earnings prospects of their firms by increasing the complexity of their written disclosures. he finds that when annual reports are harder to read, good news they contain is more transitory and bad news is more persistent in its effects on future earnings. smith and taffler (1992, 2000) find that when a firm’s financial situation deteriorates, its accounting reports tend to become more opaque. older research finds issues of readability and complexity in material presented in footnotes (healy, 1977; smith and smith, 1971). using consumer search models, ellison and ellison (2009) use the term obfuscation to describe marketing practices where firms make price comparisons complex or confusing. obfuscation can lead to reduced consumer learning about prices. carlin (2009) and wilson (2010) also present models of strategic obfuscation where firms intentionally increase the search costs of consumers. carlin (2009) finds that when some consumers are unsophisticated, firms can use complex pricing and price obfuscation to charge supracompetitive prices. ellison and wolitsky (2011) present a search model where it is rational for firms to increase the search costs of consumers. prices and firm profits increase as search costs increase. price dispersion occurs in markets where consumers are differentially informed 217l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 (ellison and wolitsky 2011). thus, price dispersion may occur in the market for financial advisers because of heterogeneity in the level of financial sophistication of clients. ellison and wolitzky (2012) find in some theoretical models that higher price mark-ups are associated with greater obfuscation. they note that obfuscation is bad for consumers both in that it raises the prices they pay and that it increases their search costs. their model is based on obfuscation increasing the amount of time it takes for consumers to discover prices. the analysis in our article differs in that it focuses more on factors that make it difficult for consumers to understand the fees that they would be charged, given that they have done the search required to find that information. this article, which focuses on advisory fee disclosures over the internet, also relates to literature on use of the internet as a search mechanism. ellison and ellison (2009) argue that the ease of search over the internet has caused some firms to increase the obfuscation concerning the disclosure of price and quality of products. in an analysis of fee disclosures, the government accountability office (u.s. government accountability office [gao], 2013) reviewed fee disclosures online for 10 large providers of individual retirement accounts (iras). it finds that fees often are disclosed in ways that make them difficult to understand. the lack of transparency involves a number of different aspects of the disclosures, which may suggest a strategy of lack of transparency in fees. gao finds that fee information is generally scattered across the providers’ websites in ways that make it difficult to find all the applicable fees. the results of the u.s. gao (2013) non-generalizable survey suggest that ira fees are often disclosed in ways that imply that fees are not important. the fees are sometimes located in difficult-to-find places. one example is fees provided in the last section of a 49-page document. that section uses small font type to disclose the information, another implicit message that the information is not important. fees are often located in footnotes in small font type. footnotes generally are viewed as being reserved for technical information that is not of interest to general readers. further obfuscating the fee disclosures, often the word “fee” is difficult to find. black et al. (2002) argue for comprehensive personal financial planning, but express concern that such planning services may lead to less transparency in fees. hung et al. (2008) conduct a study for the securities and exchange commission (sec). their main purpose is to provide the sec with a factual description of the investment advisory and brokerage industries to assist the sec in its evaluation of the legal and regulatory environment concerning investment professionals. in their survey of investors, they find that many investors find the disclosures provided by financial advisers to be difficult to understand. starr (2010) examines issues relating to fees in 403(b) plans in the public and nonprofit sectors. mazzoli and nicolini (2010) investigate determinants of the financial adviser’s choice relating to the transparency versus opaqueness of pricing policy for financial advice using italian data. they do not include the level of fees as a factor explaining the degree of opaqueness of the fee structure. they measure opaqueness by a binomial variable that is categorized as opaque if the advisory fee is not charged as a separate element but is part of other fees. thus, they do not measure degrees of opaqueness for advisers who charge fees as a percentage of assets, but consider all disclosures of that type of fees as transparent. almost all investment advisers charge fees based on assets under management, and thus their fee 218 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 disclosures are considered to be transparent. lachance and tang (2012) find that trust is a factor in the use of financial advisers. it may also be a factor in the use of financial advisers who have opaque fee disclosures. a study of consumers’ shopping strategies for medicare part d policies finds that insurers profit when consumers fail to shop around for the best priced products (ho, hogan, and morton 2015). we argue in our article that some mutual fund providers encourage that result by making information on fees difficult to find and difficult to understand. 4. complexity in language in 1998, the securities and exchange commission adopted rules on plain english disclosure in financial reports. the underlying argument behind the rules is that (1) firms could use vague language to hide adverse information, and (2) the average investor may not understand complex disclosures, which could reduce the efficiency of financial markets (li, 2008). a possible alternative motivation for the use of complex terminology in disclosures is that it may impress some clients as to the sophistication of the service provider and the difficulty of the subject. the client may view the apparent difficulty of the subject as the barrier to comprehension, rather than viewing the barrier as being because of strategic complexity in choice of language and expression. complexity can take the form of increasing the amount of time it takes to find fee information. it can also take the form of increasing the amount of time it takes to understand fee information, which would reduce the likelihood that fee information will be understood. this section focuses on complexity that reduces the likelihood that fee information will be understood. linguists have studied complexity in language and have developed empirical measures of complexity. these measures rate written documents as to the grade level of education required to understand them. these measures generally take into account sentence length and the difficulty of the vocabulary used. longer sentences are viewed as increasing the complexity of language. use of jargon or specialized terminology is one aspect of the difficulty of the vocabulary used. one commonly used measure for assessing complexity in the english language is the dale-chall readability score (dale and chall, 1948). the formula that produces that score is score � (0.1579 � pdw) � (0.0496 � asl) � 3.6365 (1) where pdw is percentage of difficult words and asl is average sentence length.1 in using the formula, 10% for pdw would be entered as 10. the score is then translated into a grade level (see table 1). any word above the school grade 4 level (age 10) is rated as a difficult word. the percentage of difficult words has a much larger effect on grade level of the writing than sentence length. for example, a sentence can be 27 words long and still be rated at grade level 4 if it has no difficult words, while if it has three difficult words, it would be rated at a grade level 7–8. this formula differs from some other readability formulas in that it assesses word difficulty based on a subjective assessment rather than on counting number of letters or 219l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 syllables in words. the dale-chall approach uses a list of 3,000 words that are considered the vocabulary of non-difficult words. a calculator is available on the internet that checks text against the list of 3,000 words (readability formulas, 2013b).2 a simple example demonstrates the use of the dale-chall approach for evaluating financial disclosures. this analysis focuses on fidelity personalized portfolios (fidelity, 2013). its fee disclosure is the following: “gross annual advisory fee: between 0.55% and 1.5% of eligible assets invested.” using the microsoft word program to count the number of words, this sentence has 12 words. that word counter appears to count the spaces between strings of characters, so that 0.55% is counted as one word. analyzing this disclosure as a sentence, the dale-chall readability score is grade 16 and above. this example points out a weakness of the application of the dale-chall readability formula using their vocabulary list of 3,000 words for the purposes of analyzing financial disclosures. the following words are not included in the list and, thus, are counted as difficult words: annual, advisory, assets, invested. while these would be difficult words for many fourth graders, they would presumably not be difficult words for pension participants reading fee disclosures. thus, just as loughran and mcdonald (2013) conclude that the gunning-fog measure, an alternative measure of language complexity, is not well-suited for analyzing financial disclosures, we conclude that the 3,000 word vocabulary associated with the dale-chall formula is too limited for the purposes of this article. we substitute instead a subjective assessment of which words are difficult words for 401(k) participants. to measure readability, the target audience needs to be taken into account. the approach used in this article focuses on both difficult words and difficult concepts. for example, the words “eligible” and “assets” are not assessed as difficult words for users of financial advisory services, but the concept “eligible assets” is rated as a difficult or unclear concept when it is not otherwise defined or explained. similarly, while the word “fee” is not a difficult concept, the phrase “gross fee” is rated as a difficult concept. in the above example difficult concepts would be “gross fee” and “eligible assets.” thus, for a sentence of 12 words, with four difficult words, applying the dale-chall formula with our assessment of difficult words yields a rating of grades 13–15 (table 1). long average sentence length is not the only aspect of sentence length that contributes to complexity. long total length of a disclosure can add to complexity. furthermore, having table 1 converting raw score to grade level using the dale-chall readability score raw score grade level 0 to 4.9 4 or below 5.0 to 5.9 5 to 6 6.0 to 6.9 7 to 8 7.0 to 7.9 9 to 10 8.0 to 8.9 11 to 12 9.0 to 9.9 13–15 (college) 10.0 and above 16 and above source: dale and chall (1948). 220 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 a single long sentence in a disclosure with multiple short sentences can add to complexity. both of these aspects of complexity are not measured by the dale-chall formula. for our analysis, we stick, however, with the traditional approach of focusing on long average sentence length. because we use a different assessment of difficult words than is used for the dale-chall formula, we use the following score formula in the remainder of the article. this formula uses the same parameter values as the traditional dale-chall formula for percentage difficult words and average sentence length. the formula differs only in that we use a higher constant because we use a more sophisticated vocabulary in measuring percentage difficult words. the higher constant is used to calibrate the formula so that it yields a score and grade level that is roughly comparable to what the dale-chall score would be if it used our vocabulary of difficult words. adjusted score � (0.1579 � pdw) � (0.0496 � asl) � 6 (2) in our formula, the percentage of difficult words is generally lower than for the dale-chall formula because of using an expanded vocabulary of acceptable words. appendix 1 contains a partial list of the words found in disclosures that we rate as difficult. however, because of our calibration of the constant term, table 1 is still used to convert the score to a grade level. the formula is specifically designed to rate complexity in documents that are designed for a more sophisticated reader than one with a low level of education. because of the constant term, it cannot be used for rating documents for fourth graders. in measuring the grade level of fee disclosures, a natural question is what should be the grade level at which documents are written? according to one commentator on health communications, “health communications professionals generally recommend designing adult targeted public education print materials for about a fifthor sixth grade reading level, to accommodate individuals who read at lower levels” (sanner, 2003). the joint forum of financial market regulators (2004) in canada has proposed that disclosures be written at a fifth grade level. 5. other elements of complexity while we focus on complexity in language, the concept of complexity used in this article is broader than just complexity in language. it is based on the idea that anything that increases the difficulty that unsophisticated investors have in understanding how much they pay in fees is part of the complexity in the presentation of fees. complexity in fee disclosures can be achieved in a number of different ways. developing these complex approaches imposes higher search and comprehension costs on users of financial advisory services. this section discusses other ways in which fee disclosure can be made complex. a later section provides examples of actual disclosures demonstrating these issues. one strategy that financial advisers use is to provide an apparent disclosure that is simple to understand, while the actual disclosure, often provided in a footnote, is difficult to 221l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 understand. the apparent disclosure may contain information on the level of fees, but sometimes it does not. the apparent disclosure may be primarily a marketing tool. it differs from the actual disclosure in that either it does not contain information on the level of fees or the actual disclosure contains information that substantially modifies it. the actual disclosure may incorporate the apparent disclosure, or it may be a separate disclosure that contradicts in some respects the apparent disclosure. the fee structure can be complex. it can require multiple calculations. it can involve multiple fees for different aspects of a service. it can involve different ways that fees are charged, with it being unclear as to which way would apply to the investor, or when there is choice, which way would be most advantageous to the investor. some companies make it difficult to understand the asset base against which fees are being charged. they may use terms such as “eligible assets,” “gross assets,” or “net assets.” complexity can be increased by placing disclosures in places that would signal that the information is not important, such as footnotes. an example of complex disclosure is placing the disclosure at the end of a lengthy segment of text, such as at the 28th line of footnotes. complexity can also be increased by disclosing fees in areas where they would not be expected to be disclosed, such as under a different heading from where the apparent disclosure occurs. while most fee disclosures are short, some are lengthy. sometimes length is because of the addition of worked examples, or supplementary information to assist in understanding the fees. we do not include worked examples and supplementary information as part of the basic fee disclosure when we count sentence length. complexity can be increased by indicating a degree of uncertainty or ambiguity as to actual fees, with use of the word “may” to indicate that actual fees may be greater or less. often this involves listing factors that may affect the level of fees, but not disclosing by how much fees would be affected. some disclosures provide information that may mislead unsophisticated investors. for example, edelman (2013) provides the following statement: “you never pay any commissions, brokerage fees or trading costs.” what this statement presumably means is that the client does not pay these fees or costs as part of their advisory fee to edelman. the client does, of course, pay trading costs relating to the underlying investments. an unsophisticated investor is not likely to see this distinction. complexity can be increased by making it more difficult to find fee information. one measure of difficulty is the number of clicks it takes to get from the initial web page describing advisory services to the page providing fee information. with nondisclosure, it is impossible to obtain information as to level of advisory fees because that information is not provided. a number of major financial service providers do not disclose information about their fees on the internet, but instead require that a potential client talk to an adviser. complexity can involve not disclosing information about the fees charged by the investment products that the adviser recommends. in practice, rarely do disclosures provide information as to the range of fees or typical fees of the investment products they recommend. while that is not an aspect of the compensation of the adviser, it is an aspect of the cost to the client. 222 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 6. empirical analysis rating fee disclosures in this section, we examine the fee disclosures of a small number of firms. because we individually code the complexity of the fee disclosures, the sample size is small. this empirical analysis is intended to demonstrate the use of the analytical tools we develop, but also to provide suggestive evidence as to the relationship between level of complexity of fee disclosure and level of fees. we argue that complexity is a multidimensional (or complex) strategy. our small sample of fee disclosures is limited in several ways. first, it is limited to firms that disclose their fees over the internet. second, it is limited to firms that do not also receive compensation through commissions. part of the fee paid by the participant may be hidden in the form of higher expense ratios on the investments to offset the commissions the mutual fund companies pay directly to the adviser. third, the sample is limited to firms with a minimum asset level for advice of no more than $500,000. we pick that figure because it is the minimum used by a number of companies. arguably, investors with higher amounts would tend to be more sophisticated and thus more likely to be able to understand complex fee disclosures. furthermore, a limit above that amount would exclude most people. in rating grade level of the text of the disclosures, we apply the adjusted dale-chall readability formula (eq. 2), but instead of using their vocabulary list, we apply our own subjective assessment as to which words would be considered difficult. computer analyses of texts can have difficulty determining sentence length because not every period denotes the end of a sentence. that problem is not an issue in the analysis in this article, where we visually count sentence length. the 10 largest mutual fund companies in the united states are vanguard, american, fidelity, t. rowe price, pimco, franklin templeton, blackrock, oppenheimer, jp morgan, and columbia (roth, 2012). the top three companies—vanguard, american, and fidelity—controlled $2.6 trillion in assets in 2008 (vohwinkle, 2008). we examine fee disclosures for all of these companies. not all of these companies, however, provide on-going investment advisory services. oppenheimer, for example, does not appear to provide an advisory service for ongoing financial advice. of these 10 mutual fund companies, we were able to find on their internet websites information about fees for ongoing financial management or advice for three: vanguard, fidelity, and t. rowe price. perhaps as important as fee disclosure is fee nondisclosure. while it appears that some of the companies do not provide ongoing advisory services for individual clients, some companies describe those services but do not disclose their fees at their website. for example, a search of the website for blackrock managed accounts did not yield any information concerning fees (blackrock, 2013). because of our inability to find information on the internet about fees for most of the largest companies, we supplement the information for those companies that is available with information from a convenience sample of companies that do provide information about their fees on the internet. because the sample is small, and it is a nonrandom sample, the sample cannot be used to statistically test our hypothesis. the companies in the sample are included based on their providing a fee disclosure on the internet. the companies range in size from 223l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 very large ones to very small ones. they include companies that also sell investment products and companies that only provide advisory services. table 2 provides some descriptive information about fee disclosure. some companies provide information about fees in the apparent fee disclosure, but do not actually disclose the level of fees. we have not included those companies in this analysis. many companies do not provide information about their fees at their websites, and those are not been included. the companies in table 2 are thus a convenience sample of companies that provide actual fee disclosures on their websites. a general pattern can be seen among the small number of advisor firms in table 2. firms charging higher fees are more likely to provide a simplified apparent disclosure along with an actual disclosure. for those companies, the actual disclosure tends have longer sentences and more total words, both indicators of complexity. table 3 rates both the apparent fee disclosure and the actual fee disclosure in terms of grade level. the apparent fee disclosure is at a considerably lower grade level than the actual fee disclosure when both are provided. the main finding from the sample in table 3 is that the grade level of the actual disclosure is considerably higher in high fee companies than in low fee companies. thus, while complexity is presumably a strategy, high fee providers can be identified solely based on the complexity of the language they use. an alternative hypothesis is that high-fee providers have more complex disclosures because they provide a more complex range of services. while that may be true for some table 2 description of fee disclosures adviser company apparent disclosure number of words (mean per sentence) actual disclosure number of length (mean per sentence) fees on first dollar (minimum investment) mean total mean total 1. edelman financial services (2013) 5 5 58 58 2.0% 2. edward jones (2013) – – 28 309 1.8% ($500,000) 3. fidelity (2013) 12 12 13 65 1.5% ($200,000) 4. west financial services 11 11 25 76 1.25% plus $250 an hour for financial planning consulting 5. summit financial services (2013) – – 16 16 1.0% ($500,000) 6. fbb capital partners – – 17 17 1.0% ($500,000) 7. vanguard (2013) – – 13 80 0.9% ($500,000) 8. zero commission portfolio (2013) – – 17 17 $149 a month 9. bloombergblack (2013) – – 11 11 $100 a month 10. nestwise20 (2013) – – 8 8 $250 initial fee and $575 a year source: authors’ compilation. note: the average sentence length is in some cases an approximation because of the use of headings and tables. the total length of disclosures does not include a worked example, if one is provided. providing an example adds to the length, but it would not be an aspect of complexity in presentation. the level of fees is the fee charged on the first dollar of investments. when a range of fees is specified without other information, the highest fee is reported. 224 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 types of financial advisers, that hypothesis is rejected in our sample by examining the services provided. in all cases, the only service provided is advice for managing investments. a further hypothesis is that any person doing comparison shopping for a financial management firm would tend to choose one with low fees and an understandable disclosure. similarly, anyone initially choosing an adviser with high fees, would tend to change to a lower fee adviser over time. those hypotheses assume rational, well-informed behavior, and would argue against the existence of high-fee and low-fee providers providing similar services, which is actually what is observed in the market. 7. survey of college students: validation of our rating tool in the previous section, we note several dimensions of complexity—complexity in language, complexity in computations, complexity in finding relevant information. in the remainder of the article, we focus on the dimension of the complexity of language. we make that choice because we believe that dimension is an important dimension of complexity, and also because we want to narrow the scope of our research so as to be able to focus on one aspect of the issue. to validate our rating tool as a measure of the difficulty of understanding fee disclosures, we surveyed 618 students in the seidman college of business at grand valley state university, a midsize university in allendale, mi. while all of the students were taking business classes, and most of the students in the sample are business majors, 27% of the respondents are not business majors, including 16% who are studying mathematics intensive majors, such as engineering, but are not business majors (table 4). while the sample of students is selective in some respects, it is less selective than the sample used by choi, laibson, and madrian (2010), where the sample consisted largely of students from harvard and wharton, and the sat scores averaged in the 99th and 98th percentiles, respectively. the survey was voluntary and administered online by the school blackboard site in october table 3 rating fee disclosures adviser company grade level fee on first dollar apparent disclosure actual disclosure 1. edelman 9–10 16 2.0% 2. edward jones (2013) – 13–15 1.8% 3. fidelity (2013) 9–10 13–15 1.5% 4. west financial services 7–8 11–12 1.25% plus $250 an hour for financial planning consulting 5. summit financial service (2013) – 7–8 1.0% 6. fbb capital partners – 7–8 1.0% 7. vanguard (2013) – 9–10 0.9% 8. zero commission portfolio (2013) – 7–8 $145 a month 9. bloombergblack (2013) – 7–8 $100 a month 10. nestwise (2013) – 7–8 $250 initial fee, $575 annual fee source: authors’ compilation, based on a portfolio of $425,000. 225l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 2014. the goal was to present the students with two actual fee disclosures that we have rated: one that we rated as “simple” and one that we rated as “complex,” using our adjusted readability formula (eq. 2). we did not identify in the survey which disclosure was simple and which was complex. the students would then compute the fee charged, as well as rate the complexity of the disclosure on a scale from 1 to 10. the full survey is found in appendix 2. summary statistics for the sample are found in table 4. a little over half of respondents are male, and 73% of the students are in some type of business major. nonbusiness majors comprise 27% of the sample, with 11% having a non-math intensive major such as education or sociology, and 16% having a math-intensive major such as engineering or statistics. additionally, 88% are age 25 or under, with the remaining 12% being nontraditional students or individuals pursuing their mba. ninety-six percent of respondents list english as their first language. some characteristics of our sample may bias the results either upwards or downwards with respect to the population of those reading financial disclosures. the literature suggests that age is positively related to financial knowledge, so given that most of the respondents are under age 25, the age of the respondents in our sample would bias the results downward, meaning they would be more likely to rate a disclosure as complex. for example, lusardi and mitchell (2011) find that financial literacy tends to have an inverted-u pattern with respect to age, being lowest among the young and the old. however, the literature also shows that individuals with business degrees make more sophisticated financial decisions (allgood et al., 2011), biasing the results upwards. given that we do not have sufficient information to determine the degree of bias either way, we should keep this in mind when extrapolating the results to the general population. table 4 summary statistics for survey sample characteristic percentage of sample male 54 female 46 age 25 and under 88 age over 25 12 english is first language 96 english is not first language 4 non-business majors 27 non-math intensive 11 math intensive 16 business majors 73 accounting 20 economics 5 finance 12 management 14 marketing 14 general business 6 mba 2 source: authors’ tabulations, based on a survey sample of 618 students. note: non-business, non-math intensive majors would include, for example, education or sociology majors. non-business, math intensive majors would include, for example, engineering and statistics majors. 226 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 in table 5 we present the answers the students gave when asked to compute the fee from both the simple and the complex disclosure (questions 5a and 6a, respectively; see appendix 2). the results clearly show that the students had a more difficult time with the complex disclosure. while half got the correct answer to the simple disclosure, only 15% computed the fee correctly for the complex one. in addition, while only 24% answered “don’t know” for the simple disclosure, 45% gave this answer for the complex one. the results in table 6 confirm the findings in table 5—the disclosure we rate as complex is more difficult than the one we rate as simple. while 40% of students rated the simple disclosure as very easy (1–3 on a scale from 1 to 10), only 24% rated the complex disclosure as such. furthermore, while approximately 10% rated the simple disclosure as a 9 or 10 (most difficult), 18% rated the complex one as such. thus, our findings are consistent with earlier studies finding heterogeneity in the ability of people to deal with complexity in financial issues. 8. logit analysis we also ran logit regressions to examine what characteristics of the student were correlated with a correct fee calculation for the simple fee disclosure and for the complex table 5 student responses to questions calculating the amount of fees, in percentages response simple disclosure (percent) complex disclosure (percent) can’t tell from information provided 14.8 10.2 answer given, correct 49.8 15.0 answer given, incorrect 11.3 29.8 don’t know 24.1 45.0 total 100.0 100.0 number of responses 576 618 source: authors’ tabulations from survey. simple disclosure is question 5a; complex disclosure is question 6a (see appendix 2). table 6 rating the difficulty of financial disclosures student rating of the complexity of the fee disclosure simple disclosure complex disclosure percent cumulative percent percent cumulative percentage 1–3 40.4 40.4 23.5 23.5 4–5 24.1 64.5 20.5 44.0 6–8 25.7 90.2 38.3 82.3 9–10 9.8 100.0 17.7 100.0 total 100.0 – 100.0 – number of responses 618 – 611 – source: authors’ tabulations from survey. ratings range from 1 (easiest) to 10 (most difficult). simple disclosure is question 5b; complex disclosure is question 6b (see appendix 2). 227l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 disclosure. this analysis provides insights as to the effect of complexity on the ability to correctly determine a price, which is the fee charged for financial advice. the marginal effects are reported in table 7 for both regressions. in both models, age and whether the student is a native english speaker are both statistically insignificant. for the simple fee disclosure, men and women are equally likely to provide the correct answer. however, for the complex disclosure, men are 8% more likely to correctly compute the fee than were women. other studies show that women tend to have lower financial literacy than men (lusardi and mitchell, 2008). we are especially interested in whether the student’s major has an effect on whether he or she answered the question correctly, as courses related to financial decision making are one of the ways a person can develop financial sophistication. hence course major can serve as a proxy of financial sophistication. accounting and finance majors are the most likely group to have specifically studied fee disclosure. in addition, mba students are also exposed to higher-level accounting and finance material that may aid them in answering the questions. mba students tend to be older and have had experience in financial decision making, so they may be even more financially sophisticated than the undergraduates. however, since we are not able to disentangle the effects of this aspect of sophistication from the knowledge they gain in the mba program, we cannot assume that the mba variable is a proxy of only coursework. the remaining business majors—marketing, management, and general business—have had exposure to business terminology and may have had accounting or finance classes, but not studied fee-disclosure type problems extensively. economics and nonbusiness majors with math-intensive concentrations (e.g., engineering or statistics) may not have had the exposure to business terms, but may have the mathematical capability to compute the fees. the last group—and omitted category in the regression—is students who have a nonbusiness, non-math intensive major (e.g., elementary education or sociology). these table 7 logit analysis characteristic low fee–correct answer high fee–correct answer age over 25 .04 (.07) .06 (.05) male .01 (.05) .08 (.03)** native english speaker .19 (.10) .03 (.05) major: accounting .30 (.07)** .33 (.15)* economics .14 (.11) .07 (.14) finance .25 (.08)** .23 (.15) management .09 (.09) .11 (.12) marketing .02 (.10) .13 (.13) general business .07 (.10) �.01 (.10) mba .29 (.11)* .51 (.21)** non-business/math .18 (.10) .17 (.14) n 575 617 source: authors’ calculations from survey. note: the estimates above are marginal effects. the category non-business/math refers to non-business majors whose major include a significant amount of math (e.g., engineering). the omitted major category is a non-business major in a concentration with little to no math (e.g., elementary education, sociology). *p � 0.05, **p � 0.01. 228 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 students would be the least likely to be able to correctly compute the fees, as their exposure to business terminology and mathematical computations is most likely minimal. for the most part, the marginal effects with respect to major are consistent with what is learned in the classroom. for the simple disclosure, accounting and finance majors are 30% and 25% more likely, respectively, to compute the correct answer than are students with a nonbusiness, non-math intensive major. mba students are also more likely to compute the correct answer. generally, it might be assumed that most people would understand the simple disclosure, while people who were more sophisticated in terms of financial literacy would be more likely to understand the complex disclosure. however, while there is still a large, statistically significant effect for mba students, for accounting and finance majors the effect for the complex disclosure was less statistically significant or insignificant, respectively. so there is some evidence that financial sophistication helps, but helps to a lesser extent with the more complex fee disclosures. 9. policy analysis clearly, some fee disclosures are written at a high grade level. whether that is done to purposely obscure the disclosure, as argued in this article, or for other reasons, a public policy response could be to have readability requirements for fee disclosures. american plain english statutes in some states require that legal text be written below a specified grade level (stumpff, 2013). 10. conclusions while much attention has been given to financial education for pension participants, less attention has been given to the quality of information they receive from financial service providers. this article hypothesizes that some financial advisers take advantage of the lack of sophistication of their clients by disclosing fee information in complex ways. complexity is a multidimensional strategy that some investment management or advisory companies use to make it difficult, if not impossible, to compare advisory services on the basis of fees, and to determine what level of fees a prospective investor would be paying. we note several dimensions of that strategy—complexity in language, complexity in computations, complexity in finding relevant information—but in the end focus on the dimension of the complexity of language. we make that choice because we believe that dimension is an important dimension of complexity, and also because we wanted to narrow the scope of our research so as to be able to focus on one aspect of the issue. the strategy takes advantage of the lack of financial sophistication of many investors, but in some instances even sophisticated investors may not be able to make comparisons or obtain accurate fee information based on the disclosures provided on the internet. this strategy presumably reduces competition in the market for advisory services and permits service providers to charge higher fees, raising their profits. 229l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 our article develops a hypothesis concerning strategic complexity in fee disclosures, develops a measure of complexity in fee disclosures, collects a small sample of fee disclosures, notes the simple correlation between complexity in fee disclosure according to our measure and level of fees, and collects a substantial sample to validate our measure of complexity of fee disclosures. our nonscientific sample of 10 fee disclosures is only suggestive of the relationship between complexity in fee disclosures and level of fees. our larger sample validates our measure of complexity. it indicates that people who are more financially literate (based on college studies) are better able to understand complex fee disclosures, but that even people with the presumption of a relatively high degree of financial literacy are not all able to decipher complex fee disclosures. this article develops analytical tools for measuring complexity in fee disclosures. it presents suggestive evidence, based on a small sample (10 advisory companies), that higher-fee advisers use more complex disclosures than lower-fee advisers. it also finds that complexity in disclosure in one dimension is generally accompanied by complexity in disclosure in other dimensions, suggesting use of strategic complexity in disclosures. in particular, higher-fee advisers tend to use more complex language and longer sentences, present important information in footnotes, and have more complex fee structures. however, it is possible to identify high fee providers solely on the basis of the complexity of the language they use. the article rejects by the use of counterexamples the argument that complexity is inherent and that simple fee disclosures involve a tradeoff against completeness. while financial education has been recommended by a number of commentators as the solution to dealing with lack of financial sophistication, with strategic complexity in fee disclosure, financial education may be of little use. for some fee disclosures, the complexity of language is beyond that that would reasonably be understood. the issue of strategic complexity in fee disclosures may exist in other aspects of financial services. for example, future studies could explore this issue concerning credit card fees, fees for financial products, fees for mortgages and other loans, fees for insurance products, and banking fees. the structure of products in all these areas has become increasingly complex. further research could explore those areas and test whether in those areas higher-fee providers also tend to use more complex fee disclosures. notes 1 a different version of the formula only includes the constant 3.6365 if the pdw is five percent or greater, which is generally for reading at grade 3 or lower. 2 we examined two other options for rating complexity of language, and found both to be inappropriate for evaluating fee disclosures. the fog index, sometimes called the gunning-fog index, developed by robert gunning (1952), bases readability on sentence length and the proportion of words with more than two syllables. in an analysis of the use of that index for measuring the readability of financial disclosures 230 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 loughran and mcdonald (2013) conclude that it is not well suited for measuring the complexity of financial disclosures because words with two or more syllables, such as “corporation,” are common in financial disclosures. the flesch reading ease test is also based on sentence length and the average number of syllables per word (readability formulas, 2013a). acknowledgment we have received helpful comments from bernard casey, vijay gondhalekar, olivia mitchell, david mccarthy, an anonymous reviewer, and participants in the 21st annual conference of superannuation researchers in sydney, australia; the conference on pension reforms in post-socialist and other countries in poznan, poland, the 2013 conference of the european network for research on supplementary pensions (enrsp) in münster, germany, and the 2015 western economic association international conference in honolulu, hawaii. appendix 1: difficult words and concepts in financial disclosures the following is a list of words or concepts that appear in financial advisory fee disclosures that are subjectively rated as difficult in terms of their effect on understanding the fees that a person would pay. eligible assets, gross assets, net assets, gross fees, net fees, 12b-1 fees, transfer fees, sec fees, wrap fees, net assets, load, may. appendix 2: survey understanding fees for financial products and services please answer the following questions. you may use a calculator. demographic questions 1. please indicate your gender m or f 2. what is your age? 3. is english your native language? y or n 4. what is your major/intended major? fee disclosures for financial advice 5. “start with a free 60-day trial, then pay $100 a month.” 231l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 a. based on that fee structure, how much would you pay the first year for an investment account that averaged $500,000 over the year? check one. can’t determine from the information . the amount would be . i don’t know b. on a scale of 1 to 10, with 1 being easy and 10 being impossible, how difficult was it to answer question 5a? 6. “we feature a single annual fee, calculated and debited quarterly from your account. it appears directly on your statement. the annual fee schedule is shown below.” amount invested fee first $150,000 2.00% next $250,000 1.65% next $350,000 1.25% next $250,000 1.00% next $2 million 0.75% next $7 million 0.60% next $15 million 0.50% above $25 million negotiable a. based on that fee structure, how much would you pay the first year for an investment account that averaged $500,000 over the year? for simplicity, assume the account balance is $500,000 each quarter. check one. can’t determine from the information . the amount would be . don’t know b. on a scale of 1 to 10, with 1 being easy and 10 being impossible, how difficult was it to answer question 6a? references abeler, j., & jäger, s. (2015). complex tax incentives. american economic journal: economic policy, 7, 1–28. allgood, s., bosshardt, w., van der klaauw, w., & watts, m. (2011). economics coursework and long-term behavior and experiences of college graduates in labor markets and personal finance. economic inquiry, 49, 771–794. black, k., ciccotello, c. s., & skipper, h. d. (2002). issues in comprehensive personal financial planning. financial services review, 11, 1–9. blackrock. (2013). what are managed accounts? http://www2.blackrock.com/us/individual-investors/productsperformance/managed-accounts. bricker, j., kennickel, a., moore, k., & sabelhaus, j. (2012). changes in u.s. family finances from 2007 to 2010: evidence from the survey of consumer finances. federal reserve bulletin, 98. carlin, b. (2009). strategic price complexity in retail financial markets. journal of financial economics, 91, 278–87. choi, j.j., laibson, d., & madrian, c. (2010). why does the law of one price fail? an experiment using index mutual funds. journal of financial studies, 23, 1405–1432. dale, e., & chall, j. (1948). a formula for predicting readability. educational research bulletin, 11–20. 232 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 edelman financial services. (2013). invest with edelman financial (available at http://www.ricedelman.com/ cs/invest_with_edelman/investment_management_how_emap_works). ellison, g., & ellison, s. (2009). search, obfuscation, and price elasticities on the internet. econometrica, 77, 427–452. ellison, g., & wolitzky, a. (2012). a search cost model of obfuscation. rand journal of economics, 43, 417–441. fidelity. (2013). fidelity managed accounts (available at https://www.fidelity.com/managed-accounts/services). franklin templeton investments. (2013). separately managed accounts (available at https://www.franklintempleton.com/ retail/pages/generic_content/prog_serv/ftpa/separately_mgd_accts.jsf). gunning, r. (1952). the technique of clear writing. new york: mcgraw-hill. healy, p. (1977). can you understand the footnotes to financial statements? accountants journal, 219–222. ho, k., hogan, j., & morton, f. (2015). the impact of consumer inattention on insurer pricing in the medicare part d program. nber working paper no. 21028. hung, a. a., clancy, n., dominitz, j., talley, e., berrebi, c., & suvankulov, f. (2008). investor and industry perspectives on investment advisors and broker-dealers. lrn-rand center for corporate ethics, law, and governance within the rand institute for civil justice and commissioned by the u.s. securities and exchange commission (sec) (available at https://www.sec.gov/news/press/2008/2008–1_randiabdreport. pdf). joint forum of financial market regulators (canada). (2004). guidelines for capital accumulation plans, (may 28). lachance, m.-e., & tang, n. (2012). financial advice and trust. financial services review, 21, 209–226. li, f. (2008). annual report readability, current earnings, and earnings persistence. journal of accounting and economics, 45, 221–247. loughran, t. & mcdonald, b. (2013). measuring readability in financial disclosures (available at http:// papers.ssrn.com/sol3/papers.cfm?abstract_id�1920411). lusardi, a., & mitchell, o. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. (2008). planning and financial literacy: how do women fare? nber working paper no. 13750, (january). lusardi, a., & mitchell, o. (2011). financial literacy around the world: an overview. nber working paper no. 17107 (available at https://core.ac.uk/download/files/153/6358633.pdf). maxey, d. (2011). how to pay your financial adviser, the wall street journal, december 12 (available at http://online.wsj.com/article/sb10001424052970204554204577024152103830414.html). mazzoli, c., & nicolini, g. (2010). to fee or not to fee: pricing policies in financial counseling. financial services review, 19, 307–322 (available at https://www.researchgate.net/profile/gianni_nicolini/publication/ 46536799_to_fee_or_not_to_fee_pricing_models_in_financial_counselling/links/5441490f0cf2a6a049a 56e3e.pdf). mccarthy, d., & turner, j.(2000). pension education: does it help? does it matter? benefits quarterly, 16, 64–72. readability formulas. (2013a). the flesch grade level readability formula (available at http://www.readabilityformulas.com/flesch-grade-level-readability-formula.php). readability formulas. (2013b). free dale-chall readability formula with word list—original and revised versions (available at http://www.readabilityformulas.com/free-dale-chall-test.php). roth, a. (2012). morningstar grades ten largest mutual fund families. moneywatch, (july 9). sanner, b. (2003). are your written materials missing the mark? journal on active aging, (july-august), 18–24. smith, m., & taffler, r. j.(1992). the chairperson’s statement and financial performance. accounting and finance, 32, 75–90. smith, m., & taffler, r. j. (2000). the chairperson’s statement: a content analysis of discretionary narrative disclosures. accounting, auditing & accountability journal, 13, 624–647. smith, j. e., & smith, n. p.(1971). readability: a measure of the performance of the communication function of financial reporting. accounting review, 46, 552–561. 233l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 starr, b.(2010). willingness to pay: understanding 403(b) fees. financial services review, 19, 125–144. stumpff, a. (2013). synthetic common law: toward simpler legal rules. paper presented at conference at university of michigan. t. rowe price. (2013). advisory planning services (available at http://individual.troweprice.com/public/retail/ products-&-services/advisory-planning-services). u.s. government accountability office (gao). (2013). 401(k) plans: labor and irs could improve the rollover process for participants. gao-13–30. vanguard. (2013). vanguard asset management services fees (available at https://investor.vanguard.com/whatwe-offer/investing-help/partner-with-me-for-ongoing-help?link�facet). vohwinkle, j. (2008). list of 50 largest mutual fund companies ranked by assets for 2008. generation x finance (available at http://genxfinance.com/list-of-50-largest-mutual-fund-companies-ranked-by-assets-for-2008/). wells fargo. (2013). wells fargo advisors’ services and fees (available at https://www.wellsfargo.com/beyond today/tools-checklists/advisorinfo/advisorfees). wilson, c. (2010). ordered search and equilibrium obfuscation. international journal of industrial organization, 28, 496–506. 234 l.a. muller, j.a. turner / financial services review 25 (2016) 215–234 academy of financial services officers president william chittenden texas state university president-elect thomas coe quinnipiac university executive vice president-program robert moreschi virginia military institute vice president-communications martin seay kansas state university vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2016 swarn chatterjee university of georgia vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president lance palmer university of georgia editor, financial services review stuart michelson stetson university directors sherman hanna ohio state university halil kiymaz rollins college frank laatsch univ. of southern mississippi david nanigian the american college tom potts baylor university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake past presidents lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2014 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 24, no. 2, 2015 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2015 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 from the editor dear readers, i am delighted to present the latest edition of the financial services review (volume 31 issue 1), packed with valuable research and insights on various aspects of finance and financial behavior. as the editor of this esteemed academic journal, i am privileged to introduce the thought-provoking articles featured in this issue. one of the significant topics explored in this edition is the debate surrounding the benefits and drawbacks of remaining in a 401(k) plan after retirement. the article authored by olivia s. mitchell, catherine reilly, and john a. turner examines whether retirees would be better off retaining their assets in their companies’ 401(k) plans or rolling their savings over to individual retirement accounts (iras). the study focuses on individuals with low or moderate levels of financial literacy. the findings reveal that many such retirees could find it financially rewarding to continue with their 401(k) plans. while iras offer a more comprehensive range of advice and distribution options, the article highlights the potential for legislative and technological developments to improve retirement outcomes, instill greater retirement confidence, and enhance security for retirees. in another captivating research piece, yi liu and russell n. james iii explore the intriguing relationship between gratitude, finance, and charitable giving. the study presents results from a repeated experiment over time, revealing the impact of different reminders on charitable giving intentions. an initial reminder of “three good things” (tgt) increases such purposes, while reminders focused on “three good financial things” (tgft) or purely “three financial things” (tft) decrease them. the researchers also observe that repeating gratitude reminders with financial references can further heighten donation intentions, though this effect wanes after the reminders stop. these findings offer valuable insights for individuals and organizations seeking to optimize their philanthropic efforts. moreover, melissa j. wilmarth, kyoung tae kim, and robin henager’s research delves into the financial behaviors of military and civilian households in the united states. their study investigates the role of financial knowledge and financial education in shaping shortterm and long-term financial behaviors. the results highlight that military households display higher financial knowledge scores and are more likely to receive financial education, which positively correlates with higher financial behaviors. this research provides valuable insights for policymakers and financial practitioners aiming to enhance financial literacy and behaviors within military and civilian communities. 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) v–vi another article by tracey west, laura de zwaan, and di johnson examines the gender bias in financial literacy measurement. the study investigates why women opt for the nonresponse option in financial literacy questions, and the findings indicate that a sustained lack of confidence in financial information is a significant factor. these results emphasize the importance of addressing gender bias in financial literacy measurement tools and advocate for targeted policies and resources to bridge the gender gap in financial literacy and ability. lastly, barry s. mulholland and michael s. finke contribute to this issue with their research on the impact of cognitive ability on life insurance lapsation. the study introduces individual cognitive ability variables to model the voluntary lapse decision by policyholders. the findings suggest that numeracy, as a measure of cognitive ability, plays a role in the voluntary lapse decision. additionally, individuals witdh higher levels of net worth are less likely to voluntarily lapse a policy, consistent with the emergency fund hypothesis. the article also introduces a new measure, “kids moving home,” which exhibits a strong positive relationship with the decision to lapse a policy voluntarily. the research aligns with life insurance demand theory, indicating that those who recently entered retirement are more likely to lapse their policies. in conclusion, the articles featured in this issue of the financial services review offer valuable contributions to the understanding of various financial phenomena, providing crucial insights for policymakers, practitioners, and researchers in the field of finance. i extend my heartfelt appreciation to all the authors, reviewers, and editorial team members for their dedication and commitment to academic excellence. happy reading and engaging discussions! sincerely, terrance k. martin jr. editor, financial services review 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. vi t. k. martin / financial services review 31 (2023) v–vi performance of alternative mutual funds: the average investor’s hedge fund srinidhi kanuria,*, robert w. mcleoda athe university of alabama, box 870224, tuscaloosa, al, 35487-0224, usa abstract alternative mutual funds (amfs) provide the individual investor with the opportunity to invest in funds that follow strategies similar to those of hedge funds and seek returns uncorrelated with the market. financial planners, advisors, and investors need to be aware of how well amfs deliver absolute or positive returns regardless of market conditions and their relatively high expense ratios. in this article we analyze the performance of amfs for the period january 1998 through december 2011 using the carhart four-factor model and the fung-hsieh seven-factor model. our results indicate that most amfs have not been able to create any value for their investors over the period of our study. furthermore, the performance of these funds was even worse during the recent financial crisis. © 2014 academy of financial services. all rights reserved. jel classification: g11; g14 keywords: alternative investments; hedge funds; mutual fund performance 1. introduction alternative mutual funds (amfs) are relatively new entrants into the mutual fund industry. as characterized by morningstar, they are also known as hedged mutual funds or non-traditional mutual funds. it is important for financial planners, advisors, and investors to be knowledgeable of the performance and costs of these funds when making investment recommendations or decisions. these funds follow investment strategies similar to those of hedge funds and are attractive to individual investors who are often unable to invest in hedge * corresponding author. tel.: �1-205-348-7842; fax: �1-205-348-0590. e-mail address: skanuri@cba.ua.edu (s. kanuri) financial services review 23 (2014) 93–121 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. funds because of high initial investment requirements and longer lock-up periods. amfs normally have a goal of providing individual investors with access to investment strategies that offer non-correlated returns and diversification benefits. this goal of amfs is in contrast with traditional or long only funds that try to beat a benchmark such as s&p 500 or russell 1000. amfs have grown rapidly in the last few years. the growth in assets under management (aum) has been significant. according to goldman sachs asset management (2012), inflows into these funds were $29.7 billion in 2005 (11% of the total mutual fund inflows). by 2009, these funds experienced inflows of $121 billion (25% of total mutual fund flows). further, by december 2011, total assets under management for all the surviving funds were $132.82 billion. amfs have more flexibility than long only funds. they can buy underpriced securities and short overpriced ones. these funds can also use leverage, derivatives, options, and swaps (like hedge funds) to seek higher returns (see appendix a1 for an explanation of nine different types of amfs). even though amfs have more flexibility than traditional mutual funds, they have more constraints than hedge funds. some of the regulations with which amfs must comply include daily liquidity, covering short positions, borrowing less than one-third of total assets, limiting investments in illiquid assets to less than 15% of assets under management.1 amfs’ active management strategy of buying undervalued securities and shorting overvalued ones could increase returns manifold, if managers make good investments; it can also increase risk, if managers make poor choices. therefore, the investment manager’s skill in buying and shorting securities is extremely important. amfs are more actively managed than traditional long-only mutual funds.2 studies by brooks and porter (2012) find that the returns from actively managed funds dominate the returns from passively managed funds and dowell and mann (2004) reached a similar conclusion in regards to fixed income funds. however, previous research by carhart (1997), elton, et al. (1995), and others does not support the existence of skilled or informed mutual fund managers.3 although amfs funds are relatively new, there has been some research in this field. koski and pontiff (1999) and deli and varma (2002) find that the flexibility to use derivatives, sell securities short, and borrow money to create leverage help managers to control expenses, risk, and manage cash flows more efficiently that makes the amfs appear to be an attractive alternative to standard mutual funds and subject to analysis. agarwal, et al. (2009) look at the performance of 52 hedged mutual funds over the period 1994–2004. they find that these hedged mutual funds outperform traditional mutual funds, but underperform similar hedge funds. broussard and neely (2011) study a similar sample of 36 long/short and marketneutral funds and find that managers of these funds do not create any alpha. our article adds to the literature by analyzing the performance of amfs during the recent financial crisis that began in 2007. we also utilize the fabozzi and francis (1979) model to test the monthly performance of these funds in both up and down markets. this analysis is important as many of these funds are sold to individual investors with the promise of absolute returns regardless of market conditions. in addition, our data include a wider variety of amfs (nine different categories), and also a much larger sample (318 funds) of which about 180 were created in the last five years. 94 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 2. hypotheses in our article we are interested in determining whether or not amfs can provide benefits to individual investor through their promise of delivering returns that are uncorrelated with the market. do they provide absolute returns and positive alphas regardless of market conditions? to examine these issues we test the following hypotheses: hypothesis 1: alternative mutual funds have more flexibility than long-only mutual funds. they can take long (short) positions in undervalued (overvalued) securities. additionally, they can use derivatives (including forwards, options, and swaps) to seek absolute returns. therefore, they should have a positive alpha. hypothesis 2: alternative mutual funds (like hedge funds) seek returns uncorrelated with the market. lipper defines them as seeking “positive returns in all market conditions” without measuring themselves against investable indexes. therefore, during bear markets and major financial crisis, they should have a positive alpha. 3. data and descriptive statistics elton, et al. (1996) find that previous mutual fund studies suffered from survivorship bias as funds that merge or die have worse performance than funds that do not and failing to account for survivorship bias will lead to higher risk-adjusted returns for mutual funds. brown, et al. (1992) also find that survivorship bias can give a false impression about persistence in mutual fund performance. to avoid this problem, we include in our analysis all alternative mutual funds that ever existed as found in the morningstar direct data. by including all the dead amfs in the analysis, we control for survivorship bias problem. 3.1. summary statistics table 1a contains some descriptive statistics on alternative mutual funds. the last column contains all the funds that ceased operations before december 2011. there were 256 surviving funds and 62 dead funds at the end of december 2011. all these funds are included in the analysis. the total assets under management for all surviving funds at the end of december 2011 were $132.82 billion. a further description of the funds is found in table 1b that includes information on management fees, net expense ratio, and turnover for different categories of alternative mutual funds. most of these funds have annual expense ratios close to or over 2%. this ratio is much higher than expense ratios for long only equity or bond mutual funds.4 according to morningstar direct, average annual expense ratios for various classifications of funds were as follows: large growth (1.43%), large value (1.43%), mid growth (1.55%), mid value (1.37%%), small growth (1.61%), small value (1.43%), long term bond (1.09%), and multisector bond (1.22%). 95s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 table 1a descriptive statistics alternative no. of living funds aum (december 2011) average (aum) standard deviation (aum) median (aum) dead funds long/short 76 29,667.70 390.36 1,124.08 44.70 33 multialternative 51 10,694.80 205.67 365.17 65.30 2 market neutral 29 23,152.10 798.35 1,181.30 94.40 18 currency 19 9,730.90 512.15 1,319.90 58.80 3 managed futures 13 6,438.50 495.27 474.33 378.50 0 inverse debt 7 515.80 73.69 105.04 29.50 0 inverse commodities 3 11.30 3.77 2.17 3.40 0 bear market 30 3,777.8 118.05 321.07 13.60 6 non-traditional bond 28 48,831.50 1,743.98 3,468.09 271.45 0 all 256 132,820.40 62 notes. aum � assets under management. all figures in millions of dollars (december 2011). the last column contains number of funds that died before december 2011. table 1b expenses and turnover (source: morningstar direct) comparison mean standard deviation median long/short management fee 1.18 0.43 1.20 annual net expense ratio 2.14 0.96 1.98 turnover (%) 423.80 1,238.68 193.25 multialternative management fee 0.93 0.48 0.98 annual net expense ratio 1.60 0.73 1.56 turnover (%) 264.90 545.68 112.00 market neutral management fee 1.29 0.33 1.25 annual net expense ratio 1.95 0.73 1.89 turnover (%) 336.78 944.34 216.00 currency management fee 0.80 0.19 0.85 annual net expense ratio 1.44 0.48 1.30 turnover (%) 129.47 326.99 28.00 managed futures management fee 1.24 0.37 1.06 annual net expense ratio 2.67 1.43 1.95 turnover (%) 312.90 619.88 70.00 inverse debt management fee 0.78 0.06 0.75 annual net expense ratio 1.88 0.41 1.81 turnover (%) 974.26 424.29 1,107.00 inverse commodities management fee 0.88 0.14 0.84 annual net expense ratio 1.65 0.68 1.51 turnover (%) 103.75 38.42 83.00 bear market management fee 0.85 0.15 0.90 annual net expense ratio 1.95 0.54 1.90 turnover (%) 506.96 507.97 473.00 non-traditional bond management fee 0.71 0.35 0.60 annual net expense ratio 1.23 0.50 1.10 turnover (%) 276.53 314.06 141.00 96 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 3.2. data our selection of the beginning period for our analysis was based on regulatory changes. before september 1997, mutual funds managers were limited in their ability to use investment strategies that involved timing the market because of the “short–short” rule that requires that mutual funds not earn more than 30% of their gross income from sales of securities held for less than three months.5 failure to comply with this rule would result in a tax of 35% on the entire gain. in september 1997, the “short–short” rule was eliminated. this change led to a proliferation of amfs who could now use short-term hedging and trading strategies irrespective of the 30% of gross earnings limitation. our analysis of amfs begins after this rule change and extends from january 1998 through december 2011. we begin our data collection first by developing a comprehensive list of all alternative mutual funds (surviving as well as dead) from the morningstar direct data. this list then was used in conjunction with data from the center for research in security prices (crsp) survivorship bias free mutual fund database to gather monthly returns, net asset value (nav), and assets. following bauer, et al. (2005, 2006, 2007), among surviving funds all funds with at least 12 months of return data are included in the analysis.6 the monthly fama-french three factors, the momentum factor (for carhart analysis), and the monthly risk-free rate were all taken from the wharton research data services (wrds) database. information on expense ratios, 12b-1 fees, turnover, inception data (for calculating fund age), and load fees were obtained from morningstar direct. 4. methodology to evaluate the performance of the amfs we compute the � using a mutual fund model and also a hedge fund model.7 the models are as follows: 4.1. carhart four-factor model according to elton, gruber, and blake (2011), the most frequently used multifactor model for measuring portfolio performance is the three factor model developed by fama and french (1993). the three-factor model is used as fama and french provide evidence that the three factors (excess market return, size factor, and value vs. growth factor) explain about 90% of diversified portfolio returns (as they are associated with risk). according to davis (2001), if the three factors do measure risk, then the fund manager should be able to earn returns to compensate for this risk. furthermore, the premiums associated with factors can be earned by a passive strategy of buying a diversified portfolio of stocks with sensitivity similar to the factors. therefore, if active fund management has any economic value, it should be able to outperform these passive strategies. carhart’s (1997) four-factor model is used as a performance benchmark. carhart fourfactor model is similar to fama-french three factor model, but it includes an additional factor for momentum (mom), which is the return difference between a portfolio of past 12-month 97s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 winners and a portfolio of past 12-month losers. the four-factor model is consistent with a model of market equilibrium with four risk factors. the model is as follows: ri,t � rf,t � �i � �i (rm,t � rf,t) � �s smbt � �v hmlt � �m mom � �i,t (1) where: ri,t � the percentage return to fund i in month t. rf,t � us t-bill rate for month t. rm,t � return on crsp value-weighted index for month t. smbt � realization on capitalization factor (small-cap return minus large-cap return) for month t. hmlt � realization on value factor (value return minus growth return) for month t. mom � the momentum factor �i,t � an error term. small company stocks will have a positive loading on smb (positive slope, �s), whereas big-company stocks tend to have a negative loading. similarly, a positive estimate on �v indicates sensitivity to value factor and a negative estimate indicates sensitivity to growth factor. a positive loading on �m would show sensitivity to momentum effects. finally, a positive intercept (�) would indicate superior performance; whereas a negative intercept would indicate underperformance, compared with the four-factor model. 4.2. fung-hsieh seven-factor model all the previous models were mutual fund models. because these funds follow strategies similar to hedge funds, mutual fund models may not give an accurate description of performance. therefore, the widely used fung-hsieh seven factor hedge fund model (fung and hsieh, 2001, 2004) is also used to determine performance. returns for bond, currency, and commodity lookback straddles have been obtained from dr. david a. hsieh’s web site.8 the seven factors are as follows: i. equity market factor – s&p 500 index monthly returns. ii. size factor – russell 2000 monthly index returns – s&p 500 monthly index returns. iii. bond market factor – the monthly change in the 10-year treasury constant maturity yield. iv. credit spread factor – the monthly change in the moody’s baa yield less 10-year treasury constant maturity yield. v. bond trend-following factor – return of ptfs bond lookback straddle. vi. currency trend-following factor – return of ptfs currency lookback straddle. vii. commodity trend-following factor – return of ptfs commodity lookback straddle. ri,t – rf,t � �1 � �1 equity � �2 size spread � �3 bond market � �4 credit spread� �5 bond trend � �6 currency trend � �7 commodity trend � �i,t (2) 98 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 5. empirical results the results from the mutual fund model indicate that most amfs have a significantly negative alpha (all alphas have been annualized) as shown in table 2a. non-traditional bond, currency and managed-futures mutual funds also had positive alphas, but the results were not significant. the carhart four-factor model indicates that the returns of most of these funds were driven by value stocks and past winners. there were some exceptions like multialternative and non-traditional bond funds which either had significant exposure to growth stocks and/or followed contrarian strategies (not buying past winners). multialternative and managedfutures funds also had significant exposure to large cap stocks. inverse debt, inverse commodities, and bear market funds had the worst performance (highly negative and statistically significant alphas) among all the categories. “all funds” is an equally weighted portfolio of all amfs within all investment styles. carhart four-factor model finds that all funds had significantly (at 1%) negative alphas. annualized all funds alphas are �3.07%. 5.1. fung-hsieh model the seven-factor model confirms the results of the mutual fund models that most of the fund categories have significantly negative alphas as shown in table 2b. the only major differences are performance of managed-futures funds. fung-hsieh model shows that managed futures funds have a significantly positive alpha of 4.64% (alpha was insignificant with carhart four-factor model). non-traditional bond mutual funds again have a positive alpha, but the results are not significant. inverse-debt, inverse commodities, and bear market funds again show the worst performance among all categories. all funds have an annualized alpha of �3.101% (statistically significant at 1%). both the models show that that all funds have very low r2 that is what we expect from a well diversified portfolio. 6. gross performance of alternative funds until now only net performance of mutual funds has been considered. this means that expenses have already been deducted from fund’s return. previous literature (jensen, 1968; malkiel, 1995; gruber, 1996; detzler, 1999) indicates that mutual fund performance net of expenses has not been generated excess returns. however, using gross returns, superior performance can be identified (blake, elton, and gruber, 1993; detzler, 1999) with alpha insignificantly different from zero. this finding is consistent with grossman and stiglitz (1980) theory of informationally efficient markets, where informed investors are compensated for their information gathering. to test this hypothesis, fund’s monthly gross return is calculated by adding 1/12 of fund’s annual expense ratio to monthly net returns the results of which are seen in table 3. some of these funds are able to follow the market with alphas insignificantly different than zero. there were exceptions, like non-traditional bond and managed futures funds that have 99s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 t ab le 2a t hi s ta bl e re po rt s th e re su lt s fo r c ar ha rt fo ur -f ac to r m od el fo r th e pe ri od ja nu ar y 19 98 th ro ug h d ec em be r 20 11 c ar ha rt a nn ua li ze d a lp ha (� 10 0) a lp ha s m b h m l k m o m r 2 n um be r of fu nd s l on g/ sh or t [� 1. 86 % ] [� 0. 00 15 77 8] ** * [0 .0 08 87 23 ] [0 .1 02 64 65 ]* ** [0 .5 47 76 69 ]* ** [0 .0 15 61 07 ] 0. 46 65 10 9 m ul ti al te rn at iv e [� 1. 32 % ] [� 0. 00 11 07 ]* ** [� 0. 09 59 00 5] ** * [� 0. 08 36 06 5] ** * [0 .3 87 54 49 ]* ** [� 0. 03 99 09 3] ** * 0. 51 15 53 m ar ke tn eu tr al [� 2. 02 % ] [� 0. 00 17 05 7] ** * [0 .0 56 84 86 ]* * [0 .0 01 33 26 ] [0 .2 50 64 15 ]* ** [0 .0 51 23 67 ]* ** 0. 12 65 47 c ur re nc y [0 .0 76 % ] [0 .0 00 06 29 ] [� 0. 01 30 32 3] [0 .0 20 24 19 ] [0 .1 06 89 79 ]* ** [0 .0 17 11 63 ] 0. 03 21 22 m an ag ed f ut ur es [1 .4 3% ] [0 .0 01 18 56 ] [� 0. 29 11 97 3] ** * [0 .2 27 49 5] ** * [0 .1 53 14 07 ]* ** [0 .0 96 08 93 ] 0. 08 62 13 in ve rs e d eb t [� 10 .3 3% ] [� 0. 00 90 41 8] ** * [0 .0 55 88 54 ] [� 0. 10 39 06 2] [0 .0 14 21 87 ] [� 0. 04 33 40 5] 0. 00 99 7 in ve rs e c om m od it ie s [� 12 .2 7% ] [� 0. 01 08 53 7] [0 .4 68 83 84 ] [0 .4 16 40 45 ] [� 1. 03 05 7] ** * [� 0. 24 33 48 1] 0. 29 74 3 b ea r m ar ke t [� 9. 21 9% ] [� 0. 00 80 27 4] ** * [� 0. 07 05 93 ] [0 .3 22 12 59 ]* ** [� 1. 39 49 72 ]* ** [0 .0 96 69 94 ]* ** 0. 61 62 36 n on -t ra di ti on al b on d [1 .0 19 % ] [0 .0 00 84 54 ] [0 .0 33 14 09 ] [� 0. 07 56 94 9] ** * [0 .1 55 11 44 ]* ** [� 0. 07 42 36 4] ** * 0. 33 78 28 a ll f un ds [� 3. 07 % ] [� 0. 00 25 95 3] ** * [0 .0 06 07 25 ] [0 .1 35 49 29 ]* ** [0 .0 16 15 97 ] [0 .0 39 92 17 ]* ** 0. 05 32 2 31 8 n ot es . r ep or te d ar e th e o l s es ti m at es fo r eq ua ll y w ei gh te d po rt fo li os pe r in ve st m en t st yl e. a ll al ph as ha ve be en an nu al iz ed . a ll f un ds is an eq ua ll y w ei gh te d po rt fo li o of al l al te rn at iv e m ut ua l fu nd s w it hi n sp ec ifi c in ve st m en t st yl e. s ta nd ar d er ro rs ar e he te ro sk ed as ti ci ty co ns is te nt . ** *s ig ni fi ca nt at 1% , ** si gn ifi ca nt at 5% . 100 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 t ab le 2b t hi s ta bl e re po rt s th e re su lt s fo r f un gh si eh se ve nfa ct or m od el fo r th e pe ri od ja nu ar y 19 98 th ro ug h d ec em be r 20 11 f un gh si eh a nn ua li ze d a lp ha (� 10 0) a lp ha e qu it y s iz e s pr ea d b on d m ar ke t c re di t s pr ea d b on d t re nd c ur re nc y t re nd c om m od it y t re nd r 2 ‘ n um be r of fu nd s l on g/ sh or t [� 2. 55 7% ] [� 0. 00 21 55 9] ** * [0 .5 24 ]* ** [0 .1 05 ]* ** [� 0. 37 5] [� 1. 11 8] ** * [� 0. 00 2] [0 .0 05 ] [0 .0 00 4] 0. 43 2 10 9 m ul ti al te rn at iv e [� 1. 37 3% ] [� 0. 00 11 51 2] ** * [0 .3 88 ]* ** [� 0. 04 7] ** [� 0. 48 2] [� 1. 15 9] ** * [0 .0 07 ]* * [0 .0 04 ] [0 .0 03 ] 0. 49 53 53 m ar ke tn eu tr al [� 2. 16 8% ] [� 0. 00 18 25 1] ** * [0 .2 03 ]* ** [0 .1 12 ]* ** [� 0. 48 0] [� 0. 81 8] [� 0. 00 3] [0 .0 07 ] [� 0. 00 06 ] 0. 10 33 47 c ur re nc y [� 0. 39 5% ] [� 0. 00 03 29 9] [0 .0 95 8] ** * [0 .0 22 ] [� 0. 66 8] [� 0. 45 6] [� 0. 00 5] [0 .0 09 ] [0 .0 03 1] 0. 03 17 22 m an ag ed f ut ur es [4 .6 4% ] [0 .0 03 78 8] ** [0 .1 28 ]* * [� 0. 16 8] ** [5 .3 23 ]* ** [0 .9 71 ] [0 .0 21 ]* * [0 .0 12 4] [0 .0 56 ]* ** 0. 22 28 13 in ve rs e d eb t [� 9. 25 % ] [� 0. 00 80 6] ** * [� 0. 04 4] [0 .0 09 ] [5 .1 75 ]* ** [0 .1 38 ] [� 0. 02 8] [0 .0 04 4] [� 0. 00 09 ] 0. 12 02 7 in ve rs e c om m od it ie s [� 14 .7 4% ] [� 0. 01 32 05 1] ** [� 0. 91 2] ** * [0 .0 92 ] [� 0. 50 2] [1 .8 02 ] [� 0. 13 0] ** [0 .0 03 4] [� 0. 03 9] 0. 25 98 3 n on -t ra di ti on al b on d [0 .7 07 4% ] [0 .0 00 58 76 ] [0 .1 30 ]* ** [0 .0 09 ] [� 1. 68 2] ** * [� 2. 54 1] ** * [� 0. 00 8] ** [0 .0 03 ] [� 0. 01 3] ** * 0. 37 67 36 b ea r m ar ke t [� 7. 1% ] [� 0. 00 61 18 ]* ** [� 1. 52 6] ** * [� 0. 31 3] ** * [� 1. 18 1] [� 2. 72 7] ** * [� 0. 04 1] ** * [0 .0 06 ] [0 .0 11 ] 0. 60 48 28 a ll f un ds [� 3. 10 1% ] [� 0. 00 26 21 7] ** * [� 0. 02 6] [0 .0 14 ] [� 0. 28 8] [� 1. 33 6] ** * [� 0. 01 5] ** * [0 .0 06 ]* * [0 .0 02 8] 0. 05 32 7 31 8 n ot es . r ep or te d ar e th e o l s es ti m at es fo r eq ua ll y w ei gh te d po rt fo li os pe r in ve st m en t st yl e. a ll al ph as ha ve be en an nu al iz ed .“ a ll f un ds ” is an eq ua ll y w ei gh te d po rt fo li o of al l al te rn at iv e m ut ua l fu nd s w it hi n sp ec ifi c in ve st m en t st yl e. s ta nd ar d er ro rs ar e he te ro sk ed as ti ci ty co ns is te nt . ** *s ig ni fi ca nt at 1% , ** si gn ifi ca nt at 5% . 101s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 significantly positive alphas, whereas inverse debt funds, inverse commodities funds, and bear market funds still underperform (significantly). overall, results from the mutual fund model indicate that all funds gross alpha is negative, but not statistically significant. 6.1. fung-hsieh model the fung-hsieh model has similar results. managed futures and non-traditional bond mutual funds have significantly positive alphas whereas inverse debt and bear market funds have significantly negative alphas (similar to mutual fund models). therefore, the underperformance of these funds is not because of expenses. all funds has insignificantly negative alpha of �0.752%. these results indicate that amfs are able to follow the indices, but that the fund expenses are too high to be able to keep pace with or outperform the market. table 3 this table shows the carhart four-factor and fung-hsieh seven-factor net and gross alphas for the period january 1998 through december 2011 carhart annualized net alpha (�100) net alpha annualized gross alpha (�100) gross alpha number of funds long/short [�1.86%] [�0.0015778]*** [0.2355%] [0.0001955] 109 multialternative [�1.32%] [�0.001107]*** [0.367%] [0.0003054] 53 market-neutral [�2.02%] [�0.0017057]*** [�0.235%] [�0.0001958] 47 currency [0.076%] [0.0000629] [1.577%] [0.0013045] 22 managed-futures [1.43%] [0.0011856] [5.41%] [0.0044015]** 13 inverse debt [�10.33%] [�0.0090418]*** [�8.56%] [�0.007428]*** 7 inverse commodities [�12.27%] [�0.0108537] [�11.289%] [�0.0099325] 3 bear market [�9.219%] [�0.0080274]*** [�5.867%] [�0.0050259]*** 36 non-traditional bond [1.019%] [0.0008454]** [2.29%] [0.0018846]*** 28 all funds [�3.07%] [�0.0025953]*** [�0.7725] [�0.0006461] 318 fung-hsieh annualized net alpha (�100) net alpha annualized gross alpha (�100) gross alpha number of funds long/short [�2.557%] [�0.0021559]*** [�0.35%] [�0.0002922] 109 multialternative [�1.373%] [�0.0011512]*** [0.31%] [0.0002588] 53 market-neutral [�2.168%] [0.0018251]*** [�0.42%] [�0.0003481] 47 currency [�0.395%] [�0.0003299] [1.10%] [0.0009127] 22 managed futures [4.64%] [0.003788]** [8.7%] [0.0069782]*** 13 inverse debt [�9.25%] [�0.00806]*** [�7.46%] [�0.0064437]*** 7 inverse commodities [�14.74%] [�0.0132051]** [�13.78%] [�0.0122837] 3 bear market [0.7074%] [0.0005876] [�3.657%] [�0.0030994]*** 28 non-traditional bond [�7.1%] [0.006118]*** [1.97%] [0.0016267]*** 36 all funds [�3.101%] [�0.0026217]*** [�0.752] [�0.0006288] 318 notes. reported are the ols estimates for equally weighted portfolios per investment style. all alphas have been annualized. “all funds” is an equally weighted portfolio of all alternative mutual funds within specific investment style. standard errors are heteroskedasticity consistent. ***significant at 1%, **significant at 5%. 102 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 7. performance of alternative funds during bull and bear markets because one of the alleged advantages of amfs is their ability to provide performance uncorrelated with the market, we test to see if performance varies depending on whether or not there is a bull or bear market. we use fabozzi and francis’s (1979) modification of the jensen model. the jensen’s alpha was modified to allow alphas and betas to vary during different market conditions. the modified model is as follows: ri,t – rf,t � �1 � �2 (dt) � �1 (rm,t � rf,t) � �2 (dt) (rm,t � rf,t) � �i,t (3) where: dt � 1 if the period is bull market and zero otherwise. �1 � bear market alpha (�bear). �2 � difference between bull and bear market alphas (�bull – �bear). �1 � �2 � bull market alpha (�bull).9 similarly, �1 is the bear market beta. �2 is the difference between bull and bear market beta and �1 � �2 is bull market beta. only the alpha specification from the article is used to determine whether alternative funds are outperforming in bull or bear markets. following fabozzi and francis (1979), the mutual fund and hedge fund models are modified to determine the performance of these funds during different market conditions. ri,t – rf,t � �1 � �2 (dt) � �i (rm,t � rf,t) � �s smbt � �v hmlt � �m mom � �i,t (modified carhart) (4) where: dt � 1 if (rm,trf,t) �0 and zero otherwise. �1 � bear market alpha (�bear). �2 � difference between bull and bear market alphas (�bull – �bear). �1 � �2 � bull market alpha (�bull). similarly, the modified fung-hsieh model is given by: ri,t – rf,t� �1 � �2 (dt) � �1 equity � �2 size spread � �3 bond market � �4 credit spread � �5 bond trend � �6 currency trend � �7 commodity trend � �i,t (5) where: dt � 1 if s&p 500 monthly return �0 and zero otherwise. �1� bear market alpha (�bear). �2 � difference between bull and bear market alphas (�bull – �bear) �1 � �2 � bull market alpha (�bull). 103s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 results from table 4 indicate that most of the amfs do not outperform during down markets. most of the down market alphas (�1) are negative for all alternative fund categories. this result is true even for bear market funds that are supposed to outperform during down markets. these results are consistent with fabozzi and francis (1979) who find no evidence of mutual funds outperformance during bear markets. results also show that during a bull market, only half (or less than half) of the funds of all categories (with the exception of managed-futures funds) have positive alphas. most of the managed-futures funds (11 out of 13) had positive alphas during bull markets. fung-hsieh seven-factor model shows that 23 out of 28 non-traditional bond mutual funds have positive bull market alphas. table 4 this table shows the fabozzi and francis (1979) modified carhart four-factor and fung-hsieh seven-factor bull and bear market alphas carhart �1 positive �1 significant �2 positive �2 significant �bull positive number of funds long/short 35 24 58 18 46 109 multialternative 13 7 38 7 27 53 market-neutral 20 6 26 6 21 47 currency 11 6 8 3 8 22 managed-futures 1 0 12 0 12 13 inverse debt 0 5 5 0 0 7 inverse commodities 1 1 1 0 1 3 bear market 8 5 6 2 5 36 non-traditional bond 15 5 16 4 16 28 fung-hsieh �1 positive �1 significant �2 positive �2 significant �bull positive number of funds long/short 34 22 61 19 44 109 multialternative 18 8 38 5 31 53 market-neutral 15 11 28 4 23 47 currency 12 4 8 5 9 22 managed-futures 2 0 10 0 11 13 inverse debt 0 5 5 0 0 7 inverse commodities 1 0 0 0 0 3 bear market 13 7 5 7 6 36 non-traditional bond 15 3 17 4 23 28 notes. ri,t � rf,t � �1 � �2 (dt) � �i (rm,t � rf,t) � �s smbt � �v hmlt � �m mom � �i,t (modified carhart) where dt � 1 if (rm,t � rf,t) �0 and zero otherwise. ri,t � rf,t� �1 � �2 (dt) � �1 equity � �2 size spread � �3 bond market � �4 credit spread � �5 bond trend � �6 currency trend � �7 commodity trend � �i,t (modified fung-hsieh model). where dt � 1 if s&p 500 monthly return �0 and zero otherwise. �1 � bear market alpha (�bear). �2 � difference between bull and bear market alphas (�bull � �bear). �1 � �2 � bull market alpha (�bull) is the sum of �1 and �2 (se bhardwaj and brooks, 1993). a t-test is used to test for the significance of coefficient. following fabozzi and francis (1979), a two tail t-test (where �1.96 � t-stat � 1.96) is used since the alternative hypothesis is that the tested coefficient is not equal to zero. 104 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 8. performance during the 2007 financial crisis these amfs strive to maintain a low correlation with the overall market. therefore, these funds should have had decent performance during the financial crisis. to test this hypothesis the performance of amfs is tested during the 2007 financial crisis. according to the wall street journal, the u.s. bear market (2007–2009) was declared in june 2008 when djia fell 20% from its october 11, 2007 high. the djia peaked at 14,198.10 on october 11, 2007 before starting its decline. the bear market reversed course by the end of march, 2009. this analysis is for the period of october, 2007 through march, 2009. the results indicate that these funds have even worse performance (significantly negatively) during the recent financial crisis as shown in table 5a. the only exception was currency funds that had positive alpha, but the results were not significant. the biggest surprise again was performance of bear-market funds. bear market funds lost significant value during the crisis period. carhart four-factor model shows that bear market funds had negative annualized alphas of over �18% (significant at 1%). all funds had significantly negative (significant at 1%) alphas of �7.51% (carhart). 8.1. fung-hsieh model the hedge fund model shows that most of these categories had negative returns but the results were not statistically significant as seen in table 5b. bear market funds have an alpha of �21.43% (significant at 1%) during the crisis. inverse debt funds have an alpha of �10.125%, whereas the three inverse commodities funds lost over half their value during the financial crisis (annualized alpha of �52.78% significant at 1%). all funds had a significantly negative alpha of �6.97% (significant at 1%) during the crisis. these results clearly demonstrate that these funds are not able to deliver absolute returns (returns uncorrelated to the market) during financial crisis. 9. robustness checks 9.1. conditional factor models all the models previously used were unconditional factor models. the assumption is that investors and managers use no information about the state of the economy to form expectations. however, if managers trade on publicly available information, and use dynamic strategies, unconditional models may produce inferior results. to address these concerns, chen and knez (1996) and ferson and schadt (1996) advocate using conditional models. this adjustment is made by using time-varying conditional expected returns and betas instead of unconditional betas. the instruments used are commonly available and proven to be useful for determining stock returns. the instruments are: (1) 1 month t-bill rate; (2) dividend yield on market index; (3); slope of the term structure (treasury constant maturity 10-year rate – treasury constant maturity 3-month rate); (4) quality spread in the corporate 105s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 t ab le 5a t hi s ta bl e re po rt s th e re su lt s fo r c ar ha rt fo ur -f ac to r m od el fo r th e pe ri od o ct ob er 20 07 th ro ug h m ar ch 20 09 c ar ha rt a nn ua li ze d a lp ha (� 10 0) a lp ha s m b h m l k m o m r 2 n um be r of fu nd s l on g/ sh or t [� 3. 06 8% ] [� 0. 00 25 93 7] ** [0 .0 78 63 24 ] [� 0. 04 94 47 4] [0 .6 37 76 5] ** * [0 .0 42 24 94 ] 0. 55 08 10 9 m ul ti al te rn at iv e [� 2. 20 % ] [� 0. 00 18 53 7] [� 0. 03 63 93 6] [� 0. 11 10 57 2] ** * [0 .4 20 92 63 ]* ** [0 .0 35 62 93 ] 0. 55 74 53 m ar ke tn eu tr al [� 5. 6% ] [� 0. 00 47 90 9] ** [0 .1 09 20 6] [� 0. 12 64 68 1] [0 .3 28 98 42 ]* ** [0 .0 71 01 99 ] 0. 18 8 47 c ur re nc y [0 .6 2% ] [0 .0 00 51 97 ] [0 .0 32 34 49 ] [0 .0 02 68 77 ] [0 .0 86 40 39 ] [0 .0 23 21 04 ] 0. 01 92 22 m an ag ed f ut ur es [0 .8 2% ] [0 .0 00 68 82 ] [� 0. 60 51 94 4] [0 .1 44 03 16 ] [� 0. 14 32 93 4] [0 .2 00 97 02 ] 0. 55 89 13 in ve rs e d eb t [� 13 .9 1% ] [� 0. 01 24 02 2] ** [0 .1 68 46 3] [� 0. 30 43 74 9] [� 0. 08 21 79 1] [� 0. 11 94 31 6] 0. 04 03 7 in ve rs e c om m od it ie s [� 36 .5 4% ] [� 0. 03 71 95 5] ** [1 .1 92 13 9] [0 .8 34 44 5] ** [� 1. 25 54 57 ]* ** [� 0. 33 27 84 5] 0. 34 46 3 n on -t ra di ti on al b on d [� 3. 19 % ] [� 0. 00 26 98 6] ** [0 .1 45 99 12 ]* * [� 0. 27 47 78 ]* ** [0 .2 09 16 56 ]* ** [� 0. 08 37 11 3] ** * 0. 46 76 28 b ea r m ar ke t [� 18 .1 06 % ] [� 0. 01 65 07 4] ** * [� 0. 17 34 49 9] [0 .2 57 28 29 ]* ** [� 1. 37 51 52 ]* ** [0 .0 37 13 85 ] 0. 60 64 36 a ll f un ds [� 7. 51 % ] [� 0. 00 64 84 1] ** * [0 .0 73 69 36 ] [� 0. 00 21 54 5] [0 .0 47 92 61 ] [0 .0 19 72 05 ] 0. 06 97 2 31 8 n ot es . r ep or te d ar e th e o l s es ti m at es fo r eq ua ll y w ei gh te d po rt fo li os pe r in ve st m en t st yl e. a ll al ph as ha ve be en an nu al iz ed .‘ a ll f un ds ’ is an eq ua ll y w ei gh te d po rt fo li o of al l al te rn at iv e m ut ua l fu nd s w it hi n sp ec ifi c in ve st m en t st yl e. s ta nd ar d er ro rs ar e he te ro sk ed as ti ci ty co ns is te nt . ** *s ig ni fi ca nt at 1% , ** si gn ifi ca nt at 5% . 106 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 t ab le 5b t hi s ta bl e re po rt s th e re su lt s fo r f un gh si eh se ve nfa ct or m od el fo r th e pe ri od o ct ob er 20 07 th ro ug h m ar ch 20 09 . f un gh si eh a nn ua liz ed a lp ha (� 10 0) a lp ha e qu it y s iz e s pr ea d b on d m ar ke t c re di t s pr ea d b on d t re nd c ur re nc y t re nd c om m od it y t re nd r 2 l on g/ sh or t [� 2. 09 7% ] [� 0. 00 17 64 1] [0 .6 03 46 45 ]* ** [0 .0 40 31 36 ] [� 0. 54 53 61 7] [� 0. 20 99 29 6] [� 0. 02 29 63 ]* [� 0. 02 06 57 2] ** * [0 .0 24 58 81 ]* * 0. 51 5 m ul tia lte rn at iv e [� 0. 07 44 9% ][ � 0. 00 00 62 1] [0 .3 78 45 09 ]* ** [� 0. 12 55 39 8] ** [� 0. 78 07 77 58 ]* [� 1. 22 31 22 ]* ** [� 0. 00 00 84 2] [� 0. 01 63 31 9] ** [0 .0 22 84 87 ]* * 0. 53 85 m ul ti -n eu tr al [� 3. 13 9% ] [� 0. 00 26 53 9] [0 .2 72 69 33 ]* ** [0 .0 17 94 93 ] [0 .5 18 36 95 ] [� 0. 61 98 57 5] [0 .0 04 99 55 ] [� 0. 00 45 66 4] [0 .0 20 70 93 ] 0. 14 32 c ur re nc y [� 1. 11 5% ] [� 0. 00 09 33 7] [0 .0 48 85 54 ] [0 .0 51 19 36 ] [� 1. 08 48 83 ] [0 .0 87 74 79 ] [� 0. 01 85 46 8] [� 0. 00 65 78 4] [0 .0 01 94 04 ] 0. 02 39 m an ag ed f ut ur es [� 0. 83 7% ] [� 0. 00 06 99 8] [� 0. 07 75 63 5] [� 0. 38 92 78 3] [3 .5 51 39 ] [2 .5 51 32 8] [0 .0 03 21 33 ] [� 0. 04 25 91 4] [0 .1 40 88 49 ]* ** 0. 69 89 in ve rs e d eb t [� 10 .1 25 % ] [� 0. 00 88 56 ] [� 0. 18 55 87 4] * [� 0. 00 77 32 8] [0 .0 45 80 92 ] [� 2. 45 56 14 ] [� 0. 14 85 84 8] ** [0 .0 40 18 42 ] [� 0. 04 30 36 3] 0. 21 28 in ve rs e c om m od iti es [� 52 .7 8% ] [� 0. 06 06 3] ** * [� 1. 12 92 33 ]* ** [1 .8 13 74 6] ** [� 1. 86 83 92 ] [5 .9 83 55 5] [� 0. 40 91 80 8] ** [0 .2 73 92 49 ]* ** [� 0. 29 95 75 4] ** 0. 50 13 n on -t ra di tio na l b on d [� 0. 74 9% ] [� 0. 00 06 26 3] [0 .1 25 68 11 ]* ** [� 0. 12 86 47 3] ** [� 3. 70 47 02 ]* ** [� 2. 54 28 29 ]* ** [� 0. 05 29 44 8] ** * [0 .0 01 63 79 ] [� 0. 00 91 88 5] 0. 51 22 b ea r m ar ke t [� 21 .4 3% ] [� 0. 01 99 00 9] ** * [� 1. 50 56 99 ]* ** [� 0. 03 86 93 ] [� 1. 26 53 79 ] [� 3. 78 45 02 ]* ** [� 0. 08 02 63 5] ** [0 .0 41 68 14 ]* ** [0 .0 04 73 57 ] 0. 57 55 a ll f un ds [� 6. 97 % ] [� 0. 00 59 98 7] ** * [0 .0 08 39 69 ] [0 .0 07 00 2] [� 0. 98 77 03 2] * [� 1. 13 76 54 ]* * [� 0. 03 92 15 2] ** [0 .0 03 47 54 ] [0 .0 09 13 44 ] 0. 06 94 8 n ot es . r ep or te d ar e th e o l s es ti m at es fo r eq ua ll y w ei gh te d po rt fo li os pe r in ve st m en t st yl e. a ll al ph as ha ve be en an nu al iz ed .“ a ll f un ds ” is an eq ua ll y w ei gh te d po rt fo li o of al l al te rn at iv e m ut ua l fu nd s w it hi n sp ec ifi c in ve st m en t st yl e. s ta nd ar d er ro rs ar e he te ro sk ed as ti ci ty co ns is te nt . ** *s ig ni fi ca nt at 1% , ** si gn ifi ca nt at 5% . 107s. kanuri, r.w. mcleod / financial services review 0 (0) 93–121 bond market (moody’s baa rated corporate bond yield – moody’s aaa rated corporate bond yield). all the instruments are lagged by one month. for the carhart four-factor model, market beta, smb, hml, and mom are allowed to vary over time. 9.2. conditional carhart four-factor model ri,t � rf,t � �i � �i (rm,t � rf,t) � �s smbt � �v hmlt � �m mom � �1{zt-1*(rm,t � rf,t) } � �2{zt-2* smbt } � �3{zt-3* hmlt } � �4{zt-4*mom} � �i,t (6) for the fung-hsieh seven-factor model, all the seven factors are allowed to vary over time. 9.3. conditional fung-hsieh seven-factor model ri,t – rf,t � �1 � �1 equity � �2 size spread � �3 bond market � �4 credit spread � �5 bond trend � �6 currency trend � �7 commodity trend � �1{zt-1* equity} � �2{zt-2* size spread} � �3{zt-3* bond market} � �4{zt-4* credit spread} � �5{zt-5* bond trend} � �6{zt-6* currency trend} � �7{zt-7* commodity trend} � �i,t (7) 9.4. results of conditional models tests the results are robust with conditional carhart model as shown in table 6. the conditional fung-hsieh model produces similar results with two exceptions. inverse commodity funds have a positive alpha of 3% but the results were not significant (unconditional model shows an alpha of �14.74% significant at 1%) whereas bear market funds have a conditional alpha of �3.75% (significant at 1%) compared with an unconditional alpha of �7.1% (also significant at 1%). the conditional alpha of all funds is �2.627% (significant at 1%) versus an unconditional alpha of �3.1% (also significant at 1%). 9.5. mutual fund market-timing and selectivity previous literature (treynor and mazuy, 1966; kon and jen, 1978; henriksson and merton, 1981, lee and rahman, 1990) studies mutual fund market timing and selectivity. they find that mutual fund managers have limited success in market timing and selectivity. however, amfs use hedge fund strategies. research (chen, 2007; chen and liang, 2007) on hedge funds indicates that hedge fund managers have some success in market timing. many of these amfs are run by a manager with some hedge fund experience. this experience factor is confirmed by agarwal, et al. (2009) who find that managers of at least half the hedged mutual funds have prior hedge fund experience. the following two models are used to test for mutual funds market timing and selectivity. 108 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 t ab le 6 t hi s ta bl e sh ow s th e c on di ti on al vs . u nc on di ti on al al ph a fo r c ar ha rt fo ur -f ac to r m od el s an d f un gh si eh se ve nfa ct or m od el s fo r th e pe ri od ja nu ar y 19 98 th ro ug h d ec em be r 20 11 c ar ha rt u nc on di ti on al a nn ua li ze d a lp ha (� 10 0) u nc on di ti on al a lp ha ( � ) r 2 c on di ti on al a nn ua li ze d a lp ha (� 10 0) c on di ti on al a lp ha ( � ) r 2 n um be r of fu nd s l on g/ sh or t [� 1. 86 % ] [� 0. 00 15 77 8] ** * 0. 46 65 [� 2. 12 % ] [� 0. 00 17 86 5] ** * 0. 47 35 10 9 m ul ti al te rn at iv e [� 1. 32 % ] [� 0. 00 11 07 ]* ** 0. 51 15 [� 1. 74 3] ** * [� 0. 00 14 64 4] ** * 0. 53 04 53 m ar ke tn eu tr al [� 2. 02 % ] [� 0. 00 17 05 7] ** * 0. 12 65 [� 2. 31 % ] [� 0. 00 19 45 ]* ** 0. 14 03 47 c ur re nc y [0 .0 76 % ] [0 .0 00 06 29 ] 0. 03 21 [0 .3 2% ] [0 .0 00 26 99 ] 0. 04 42 22 m an ag ed -f ut ur es [1 .4 3% ] [0 .0 01 18 56 ] 0. 08 62 [� 2. 72 % ] [� 0. 00 22 97 9] 0. 30 15 13 in ve rs e d eb t [� 10 .3 3% ] [� 0. 00 90 41 8] ** * 0. 00 99 [� 9. 76 8% ] [� 0. 00 85 29 4] ** * 0. 08 84 7 in ve rs e c om m od it ie s [� 12 .2 7% ] [� 0. 01 08 53 7] 0. 29 74 [� 6. 22 % ] [� 0. 00 53 39 4] 0. 46 32 3 b ea r m ar ke t [� 9. 21 9% ] [� 0. 00 80 27 4] ** * 0. 61 62 [� 7. 66 % ] [� 0. 00 66 14 8] ** * 0. 62 49 36 n on -t ra di ti on al b on d [1 .0 19 % ] [0 .0 00 84 54 ] 0. 33 78 [0 .8 76 % ] [0 .0 00 72 72 ] 0. 41 37 28 a ll f un ds [� 3. 07 % ] [� 0. 00 25 95 3] ** * 0. 05 32 2 [� 3. 33 8% ] [� 0. 00 28 25 9] ** * 0. 05 30 3 31 8 f un gh si eh u nc on di ti on al a nn ua li ze d a lp ha (� 10 0) u nc on di ti on al a lp ha ( � ) r 2 c on di ti on al a nn ua li ze d a lp ha (� 10 0) c on di ti on al a lp ha (� ) r 2 n um be r of fu nd s l on g/ sh or t [� 2. 55 7% ] [� 0. 00 21 55 9] ** * 0. 43 2 [� 2. 75 7% ] [� 0. 00 23 26 7] ** * 0. 43 79 10 9 m ul ti al te rn at iv e [� 1. 37 3% ] [� 0. 00 11 51 2] ** * 0. 49 53 [� 1. 2% ] [� 0. 00 10 03 1] 0. 51 65 53 m ar ke tn eu tr al [� 2. 16 8% ] [� 0. 00 18 25 1] ** * 0. 10 33 [� 1. 95 1% ] [� 0. 00 16 40 2] ** 0. 11 32 47 c ur re nc y [� 0. 39 5% ] [� 0. 00 03 29 9] 0. 03 17 [� 0. 56 % ] [� 0. 00 04 67 8] 0. 04 84 22 m an ag ed f ut ur es [4 .6 4% ] [0 .0 03 78 8] ** 0. 22 28 [5 .7 5% ] [0 .0 04 67 44 ]* * 0. 40 84 13 in ve rs e d eb t [� 9. 25 % ] [� 0. 00 80 6] ** * 0. 12 02 [� 10 .1 75 % ] [� 0. 00 89 02 4] ** * 0. 18 21 7 in ve rs e c om m od it ie s [� 14 .7 4% ] [� 0. 01 32 05 1] ** 0. 25 98 [3 .0 03 6% ] [0 .0 02 46 92 ] 0. 47 09 3 n on -t ra di ti on al b on d [0 .7 07 4% ] [0 .0 00 58 76 ] 0. 37 67 [0 .8 42 % ] [0 .0 00 69 87 ] 0. 44 84 36 b ea r m ar ke t [� 7. 1% ] [� 0. 00 61 18 ]* ** 0. 60 48 [� 3. 75 % ] [� 0. 00 31 83 8] ** * 0. 62 22 28 a ll f un ds [� 3. 10 1% ] [� 0. 00 26 21 7] ** * 0. 05 32 7 [� 2. 62 7% ] [� 0. 00 22 16 ]* ** 0. 05 30 7 31 8 n ot es . r ep or te d ar e th e o l s es ti m at es fo r eq ua ll y w ei gh te d po rt fo li os pe r in ve st m en t st yl e. a ll al ph as ha ve be en an nu al iz ed .“ a ll f un ds ” is an eq ua ll y w ei gh te d po rt fo li o of al l al te rn at iv e m ut ua l fu nd s w it hi n sp ec ifi c in ve st m en t st yl e. s ta nd ar d er ro rs ar e he te ro sk ed as ti ci ty co ns is te nt . ** *s ig ni fi ca nt at 1% , ** si gn ifi ca nt at 5% . 109s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 the first model was developed by treynor and mazuy (1966). this model adds a quadratic term to capm or the market model to capture market timing and selectivity. the equation is as follows: ri,t � rf,t � �s � �1 (rm,t � rf,t) � �2 (rm,t � rf,t) 2 � ei,t (8) a positive and significant �2 indicates superior market timing ability. however, a negative and significant �2 indicates inferior market timing. if �2 is not different than 0, then the manager has no market timing ability. similarly, �s measures selectivity. the second model, which was developed by henriksson and merton (1981), replaces the quadratic term with a variable max (0, rm). the equation is as follows: ri,t � rf,t � �s � �1 (rm,t � rf,t) � �[dt(rm,t � rf,t)] � ei,t (9) where: dt � 0 if rm,t � rf,t (�1 otherwise). here, � measures market-timing ability, whereas �s measures selectivity. these results, shown in table 7a, confirm previous mutual fund literature that mangers of these funds in general do not have success in market-timing or stock selection. only non-traditional bond mutual fund managers seem to have success in selectivity, but no market-timing ability. some of the other categories do have positive �2 and � (market-timing ability), but the results are not significant. all categories with the exception of nontraditional bond mutual funds had negative (some of them statistically significant) or insignificantly positive �s. both the models find that all funds (that is an equally weighted portfolio of all amfs within specific investment style) do have any success in market-timing or selectivity. following ferson and schadt (1996), the conditional market timing and selectivity of these funds is given by: conditional treynor and mazuy ri,t � rf,t � �s � �1 (rm,t � rf,t) ��1{zt-1*(rm,t � rf,t) } � �2 (rm,t � rf,t)2 � ei,t (10) conditional henriksson and merton ri,t � rf,t � �s � �1 (rm,t � rf,t) ��1{zt-1*(rm,t � rf,t) } � �[dt(rm,t � rf,t)] � ei,t (11) where: dt � 0 if rm,t � rf,t (�1 otherwise). the results remain same. only non-traditional bond mutual funds have some selectivity, but no market-timing ability. other amfs have no market-timing or selectivity. 110 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 9.6. persistence grinblatt and titman (1992), brown, et al. (1992), hendricks, et al. (1993), brown and goetzmann (1995), goetzmann and ibbotson (1994), kahn and rudd (1995), malkiel (1995), elton, et al. (1996), and carhart (1997) have tested the persistence of mutual fund total returns in time. grinblatt and titman (1992) find evidence that differences in performances between funds persists over time and this persistence is consistent with the ability of fund managers to earn abnormal returns. hendricks, et al. (1993) find that relative performance no-load growth funds persist in the near term, with the strongest evidence for a one-year horizon. goetzmann and ibbotson (1994) find strong evidence that past mutual fund performance predicts future mutual fund performance. their data suggests that winner and losers are likely to repeat, even when performance is adjusted for relative risk. kahn and rudd (1995) investigate performance persistence for fixed income and equity mutual funds and found performance persistence only for fixed income funds. however, this persistence edge cannot overcome the average underperformance of fixed-income funds resulting from fees and expenses. elton, et al. (1996) find that risk-adjusted performances of mutual funds persist, that is, funds that did well in the past continue to do well in the future. deztel and table 7a reported are the results from treynor and mazuy (1966) and henriksson and merton (1984) models treynor and mazuy �s �2 r2 number of funds long/short [�0.0010545]** [�0.1481033] 0.4613 109 multialternative [�0.0002566] [�0.2660947] 0.4979 53 market-neutral [�0.0004345] [�0.3265041] 0.1168 47 currency [�0.0003485] [0.1284035] 0.0313 22 managed-futures [0.0009739] [�0.2675342] 0.0285 13 inverse debt [�0.0091004]*** [0.0422014] 0.0017 7 inverse commodities [�0.0125913] [0.5599629] 0.2335 3 non-traditional funds [0.0026413]*** [�0.5331332]*** 0.3057 28 bear market [�0.0034579]*** [�1.550607]*** 0.6073 36 all funds [�0.0018099]*** [�0.2233747] 0.05329 318 henriksson and merton �s � r2 number of funds long/short [�0.0015555]** [�0.0002089] 0.461 109 multialternative [�0.0008719] [0.0004863] 0.496 53 market-neutral [�0.0003722] [0.0021221] 0.1158 47 currency [�0.000494] [�0.0011674] 0.0311 22 managed-futures [�0.0230618]*** [0.0445176] 0.5666 13 inverse debt [�0.0068348]** [0.0048976] 0.0032 7 inverse commodities [�0.016618] [�0.012518] 0.2349 3 non-traditional funds [�0.00004059] [�0.0024351] 0.3021 28 bear market [�0.0100744]*** [�0.0058585] 0.6026 36 all funds [�0.0037009]*** [�0.0027284]** 0.05339 318 notes. for the treynor and mazuy (1996) model, �s measures selectivity whereas �2 measures markettiming. similarly, for the henriksson and merton (1984) model, �s measures selectivity whereas � measures market-timing. “all funds” is an equally weighted portfolio of all alternative mutual funds within specific investment style. standard errors are heteroskedasticity consistent. ***significant at 1%, **significant at 5%. 111s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 weigand (1998) find that adjusting fund returns for the size of the stocks in which funds invest and financial ratios intended to capture fund manager investment styles explains all the persistence in mutual fund returns from 1976 to 1985, the period in which persistence is most prevalent. philpot (2000) looks at the performance persistence of non-conventional bond funds (high-yield bonds, global issues, and convertible bonds) and finds that short-term performance persistence is present, but limited to the high-yield bond subsample. survivorship bias plays a very important role in performance persistence. this impact is because of truncation of the data set because of disappearance of poorly performing funds from the sample. studying only surviving funds will overstate performance. brown, et al. (1992) show that early studies exaggerate the extent of persistence by relying on survivorship-biased data sets. because survivorship bias has been controlled, there will be no such problems. carhart (1997) finds that in his survivorship bias free sample of u.s. equity funds, persistence disappears after accounting for momentum in stock returns. however, recent studies argue that after properly considering fund styles, there is persistence in u.s. equity funds (ibbotson and patel, 2002; wermers, 2000). following kahn and rudd (1995), the following model is used test whether alternative mutual funds have any performance persistence. table 7b reported are the results from conditional treynor and mazuy and henriksson and merton models conditional treynor and mazuy �s �2 r2 number of funds long/short [�0.0010298]** [�0.1524436] 0.4639 109 multialternative [�0.0011016]** [0.2324004] 0.5105 53 market-neutral [�0.0004432] [�0.3883459] 0.1202 47 currency [�0.0003479] [�0.0222527] 0.0371 22 managed-futures [�0.0007113] [0.5235891] 0.1667 13 inverse debt [�0.0100143]*** [0.3859164] 0.0197 7 inverse commodities [�0.0146829] [2.237837] 0.2678 3 non-traditional funds [0.0018968]*** [�0.3312209] 0.3643 28 bear market [�0.0034545]*** [�1.630376]*** 0.6101 36 all funds [�0.0016509]*** [�0.5044269]*** 0.05317 318 conditional henriksson and merton �s � r2 number of funds long/short [�0.0009166] [0.0011572] 0.4638 109 multialternative [�0.0005413] [�0.0001244] 0.5095 53 market-neutral [�0.0002004] [0.0027451] 0.1189 47 currency [�0.0007876] [�0.0008534] 0.0372 22 managed-futures [0.0052106] [0.0098067] 0.1722 13 inverse debt [�0.0082083]*** [0.002002] 0.0186 7 inverse commodities [�0.0114461] [�0.0042874] 0.2586 3 non-traditional funds [�0.0006605] [�0.0036931]** 0.3624 28 bear market [�0.0085499]*** [�0.0019516] 0.6049 36 all funds [�0.0030748]*** [�0.0002986] 0.0532 318 notes. for the conditional treynor and mazuy model, �s measures selectivity whereas �2 measures markettiming. similarly, for the conditional henriksson and merton model, �s measures selectivity whereas � measures market-timing. “all funds” is an equally weighted portfolio of all alternative mutual funds within specific investment style. standard errors are heteroskedasticity consistent. ***significant at 1%, **significant at 5%. 112 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 period 2 performance is regressed against period 1 performance. performance (2) � a � b x performance (1) (12) where “performance” is annual returns. positive estimates of coefficient b with significant t-statistics are evidence of persistence: period 1 performance contains useful information about period 2 performance. the results from table 8 indicate that none of the mutual fund categories display any performance persistence. 9.7. factors affecting performance of amf following otten and bams (2002) and bauer, et al. (2007), the following regression is performed to estimate the marginal effect of expense ratios and other variables on performance of all funds. the results of this analysis are presented in table 9. �i � c0 � c1 ln (assetsi) � c2 (turnoveri) � c3 expense ratioi � c4 ln(agei) � �i (13) table 8 reported are the results of kahn and rudd (1995) performance persistence model a b long/short .0207159** �0.0410406 [2.23] [�0.66] multialternative .0211119** �0.239867*** [2.33] [�3.25] market-neutral 0.0081719 0.0770294 [0.77] {0.89} currency .0204124*** �0.1299019} [2.68] [�1.05] managed-futures �0.0283912 �0.1391153 [�1.06] [�0.33] inverse debt �0.1000085*** �0.2209817 [�5.32] [�1.26] inverse commodities �0.2029987*** �0.2472212 [�3.85] [�0.75] non-traditional bond .0424145*** �0.2435717 [2.65] [�1.59] bear market �0.1160574*** �0.0197846 [�4.39] [�0.26] all funds �0.0148607** 0.0346342 [�2.47] [0.72] notes. “all funds” is an equally weighted portfolio of all alternative mutual funds within specific investment style. standard errors are heteroskedasticity consistent. t-stats are in brackets. period 2 performance is regressed against period 1 performance. performance (2) � a � b � performance (1). where ‘performance’ is annual returns. positive estimates of coefficient b with significant t-statistics are evidence of persistence: period 1 performance contains useful information about period 2 performance. ***significant at 1%, **significant at 5%. 113s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 where: �i � alpha for fund i from both the models (carhart and fung-hsieh). ln (assetsi) � ln of total assets for fund i at end of 2011. ln is used instead of total assets as this variable may be non-linearly related to performance. turnoveri � turnover for fund i at end of 2011. expense ratioi � expense ratio for fund i at end of 2011. ln (agei) � ln of total fund’s age at end of 2011. ln is used instead of total age as this variable may be non-linearly related to performance. ln (assets) is significantly positive in both cases indicating economies of scale for larger funds (similar to otten and bams, 2002). turnover and expenses are negative in both cases, but the results are not statistically significant. these results, however, are consistent with previous literature that turnover and expenses are negatively related to performance (blake, et al., 1993; malkiel, 1993; carhart, 1997; dellva and olson, 1998). finally, the influence of fund age is considered. results indicate that that there is significantly negative relationship between fund age and performance (significantly in the case of fung-hsieh model). these results are consistent with webster (2002) who finds a strong negative relationship between fund age and market adjusted returns. this analysis is only for surviving funds at the end of 2011. we also ran a robustness check (not reported in the article) where we included all the dead funds (based on the final value in their year of death). this did not change the nature of the cross-sectional results. 10. index funds versus alternative mutual funds would investors have been better off with index mutual funds? index funds tracking the s&p 500 are passively managed and have lower expense ratios and management fees. only table 9 the influence of fund characteristics on risk adjusted performance. model constant ln (assets) turnover expense ln (age) carhart �0.0032062** [�1.99] 0.0010181*** [5.37] �0.0000709 [�1.40] �0.0646934 [�1.36] �0.0005993 [�1.73] fung-hsieh �0.0012341*** [�3.19] 0.0005377*** [3.75] �0.0000136 �0.26 �0.0227598 [�0.71] �0.0004795** [�2.03] notes. following otten and bams (2002) and bauer, et al. (2007), the following regression is performed to estimate the marginal effect of expense ratios and other variables on performance of “all funds.” standard errors are heteroskedasticity consistent. t-stats are in brackets. �i � c0 � c1 ln (assetsi) � c2 (turnoveri) � c3 (expense ratioi) � c4 ln(agei) � �i where �i � alpha for fund i from both the models (carhart and fung-hsieh). ln (assetsi) � ln of total assets for fund i at end of 2011. ln is used instead of total assets as this variable may be non-linearly related to performance. turnoveri � turnover for fund i at end of 2011. expense ratioi � expense ratio for fund i at end of 2011. ln (agei) � ln of total fund’s age at end of 2011. ln is used instead of total age as this variable may be non-linearly related to performance. ***significant at 1%, **significant at 5%. 114 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 equity amfs are compared to index funds for this apples to apples comparison. therefore, only amfs which have significant equity exposure are included in this analysis. the performance of all index funds tracking the s&p 500 (living as well as dead) from january 1998 through december 2011 is computed. there are a total of 62 index funds tracking the s&p 500 during this period. table 10a shows the differences in expense ratios and management fees between index funds and alternative funds. index funds have average expense ratio of only 0.65% compared with 2.14% for long/short, 1.60% for multialternative funds, 1.95% for market neutral, and 1.95% for bear market. the performance comparison is made using the carhart four-factor model. fung-hsieh seven-factor model is not used for computing alpha of index funds (tracking s&p 500) as s&p 500 returns (equity factor) is one of the seven factors for fung-hsieh model. table 10b clearly shows that performance of index funds is better than any of the equity amfs over the entire sample period. the annualized alpha of index funds with the carhart table 10a expenses and turnover for index funds tracking s&p 500 and equity amfs comparison number mean standard deviation index mutual funds 62 management fee 0.20 0.16 annual net expense ratio 0.65 0.47 turnover(%) 9.89 24.602 long/short 109 management fee 1.18 0.43 annual net expense ratio 2.14 0.96 turnover(%) 423.80 1,238.68 multialternative 53 management fee 0.93 0.48 annual net expense ratio 1.60 0.73 turnover(%) 264.90 545.68 market neutral 47 management fee 1.29 0.33 annual net expense ratio 1.95 0.73 turnover(%) 336.78 944.34 bear market 36 management fee 0.85 0.15 annual net expense ratio 1.95 0.54 turnover(%) 506.96 507.97 table 10b this table shows difference in performance between passively managed index mutual funds tracking s&p 500 and equity alternative mutual funds using alpha from the carhart four-factor model for the period january 1998 through december 2011 carhart annualized alpha (�100) alpha r2 no. index funds tracking s&p 500 [�0.5598%] [�0.0004677]*** 0.9746 62 long/short [�1.86%] [�0.0015778]*** 0.4665 109 multialternative [�1.32%] [�0.001107]*** 0.5115 53 market-neutral [�2.02%] [�0.0017057]*** 0.1265 47 bear market [�9.219%] [�0.0080274]*** 0.6162 36 notes. reported are the ols estimates for equally weighted portfolios per investment style. all alphas have been annualized. standard errors are heteroskedasticity consistent. fung-hsieh seven-factor model is not used for computing alpha of index funds (tracking s&p 500) as s&p 500 returns (equity factor) is one of the seven-factors for the fung-hsieh model. ***significant at 1%, **significant at 5%. 115s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 four-factor model is �0.56% (significant at 1%). this result is much better than any of the equity amfs. even during the 2007 crisis (table 10c), the s&p 500 index funds lost much less value than equity amfs. 11. conclusion our results indicate that most alternative funds have not delivered on their promise to provide absolute returns regardless of market conditions and have not created any value for investors. using the mutual fund model, we do find that non-traditional bond mutual funds have positive alphas while the fung-hsieh seven-factor model shows similar results for managed-futures funds. however, the conditional carhart model finds that alpha of nontraditional bond mutual funds is positive, but not significant. the performance of amfs was even worse during the 2007 financial crisis that shows that these funds do not deliver absolute returns. on a gross return basis, some of the categories are able to follow the market with alphas insignificantly different than zero. there are again exceptions like inverse-debt, inverse-commodities and bear market funds which still underperform on a gross basis. the managers of these funds do not have any success in market timing or selectivity and none of the fund categories display any performance persistence. based on our findings, investors should be wary of having unrealistic expectations of performance and diversification benefits of these alternative mutual funds. our analysis does not support the hypotheses that amfs will have positive alphas and that they will provide superior performance in bear markets. notes 1 see, for example, agarwal, et al. (2009). 2 see table 1b for turnover ratios for amfs and appendix a-3 for the turnover ratios for traditional long-only funds. table 10c this table shows difference in performance between passively managed index mutual funds tracking s&p 500 and equity alternative mutual funds using alpha from the carhart four-factor model for the period october 2007 through march 2009 (2007 financial crisis) carhart annualized alpha (�100) alpha r2 no. index funds tracking s&p 500 [�1.556%] [�0.0013062]*** 0.9908 62 long/short [�3.068%] [�0.0025937]** 0.5508 109 multialternative [�2.20%] [�0.0018537] 0.5574 53 market-neutral [�5.6%] [�0.0047909]** 0.188 47 bear market [�18.106%] [�0.0165074]*** 0.6064 36 notes. reported are the ols estimates for equally weighted portfolios per investment style. all alphas have been annualized. standard errors are heteroskedasticity consistent. fung-hsieh seven-factor model is not used for computing alpha of index funds (tracking s&p 500) as s&p 500 returns (equity factor) is one of the seven-factors for the fung-hsieh model. ***significant at 1%, **significant at 5%. 116 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 3 additional studies on fund performance show mixed results. according to sewell (2011) for every article that supports market efficiency (most of these articles were published before 1990), there are three articles that reject the efficient markets hypothesis. for example, jensen (1968), ippolito (1989), droms and walker (1996), davis (2001), baras, et al. (2010) find no superior performance of mutual fund managers, whereas grinblatt and titman (1992), hendricks, et al. (1993), goetzman and ibbotson (1994), elton (1996), gruber(1996), wermers (2000), haslem, et al. (2008), and budiono and martens (2010) find evidence of superior performance studies on international funds tend to reject market efficiency. see, for example, otten and bams (2002), huij and post (2011). 4 t-test confirm these results (see also appendix a2 and a3 for summary statistics of these comparison mutual fund categories). 5 internal revenue service code section 851 (b)(3). 6 table 1a shows number of dead and surviving funds for each category. some of the categories such as managed futures, inverse debt, and inverse commodities have 13, 7, and 3 funds, respectively. we have monthly data for these funds for the last six to seven years. we compute alpha for these categories for the available time period. the same approach has been used by bauer, et al. (2005, 2006, 2007) and otten and bams (2002). they compute alpha for mutual fund categories with 1–15 funds. 7 the authors also used the capital asset pricing model and the fama-french threefactor model to evaluate the performance of amfs. although the results are not reported in this article, they are consistent with the carhart four-factor model. 8 https://faculty.fuqua.duke.edu/�dah7/hfrfdata.htm. 9 the bull market alpha (�1 � �2) for a mutual fund is simply the sum of bear market alpha (�1) and difference between bull and bear market alpha (�2) for that fund and as such there is no significance level for this alpha (see bhardwaj and brooks, 1993). acknowledgments we thank tom downs, sherwood clements, jesse ellis, douglas cook, javier vidal garcia, roger otten and our anonymous referees for helpful comments and suggestions. appendix a1 definitions (source-morningstar): multialternative: these funds offer investors exposure to a combination of strategies like long-short equity and debt, managed futures, global macro, and convertible arbitrage, among others. these strategies may change in response to market conditions. short exposure is usually greater than 20%. managed futures: these funds typically take long and short positions in futures or other derivative contracts according to a trend-following or momentum strategy. currency mutual funds: these funds invest in united states and foreign currencies using 117s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 short term money market instruments, derivatives (including forwards, options, swaps), and cash deposits. long/short mutual funds: these funds take long (short) positions in undervalued (overvalued) assets. long/short structure varies from 120/20 to 150/50 with 130/30 being the most popular. because of regulation t, short selling is limited to 50%. market neutral: these funds try to earn income by maintaining low correlation with the market. these funds usually have 50% of net assets in long positions whereas holding 50% of net assets in short positions. their goal is to deliver positive returns regardless of fluctuations in market. inverse debt: these funds seek to generate returns equal to an inverse fixed multiple of short-term returns of a fixed-income index. inverse commodities: these funds seek to generate returns equal to an inverse multiple of short-term returns of a commodity index. non-traditional bonds: many funds in this group describe themselves as “absolute return” portfolios, which seek to avoid losses and produce returns uncorrelated with the overall bond market; they use a variety of methods to achieve those aims. bear market funds: bear-market portfolios invest in short positions and derivatives to profit from stocks that drop in price. because these portfolios often have extensive holdings in shorts or puts, their returns generally move in the opposite direction of the benchmark index appendix a2 total number of funds (including dead funds) and total assets under management (aum) for comparison mutual fund categories at the end of december 2011 category total funds living funds dead funds aum (december 2011) large value 620 352 268 $540.40 large growth 846 430 416 $780.80 mid value 209 115 94 $ 84.10 mid growth 483 235 248 $198.60 small value 202 104 98 $ 57.00 small growth 470 231 239 $120.90 multisector bond 104 59 45 $136.20 long term bond 59 15 44 $ 14.70 note. aum in billions of dollars. 118 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 references agarwal, v., boyson, n. m., & naik, n. y. (2009). hedge funds for retail investors? an examination of hedged mutual funds. journal of financial and quantitative analysis, 44, 273. barras, l., scaillet, o., & wermers, r. (2010). false discoveries in mutual fund performance: measuring luck in estimated alphas. the journal of finance, 65(1), 179–216. bauer, r., koedijk, k., & otten, r. (2005). international evidence on ethical mutual fund performance and investment style. journal of banking & finance, 29, 1751–1767. bauer, r., derwall, j., & otten, r.(2007). the ethical mutual fund performance debate: new evidence from canada. journal of business ethics, 70, 111–124. bauer, r., otten, r., & rad, a. t. (2006). ethical investing in australia: is there a financial penalty? pacific-basin finance journal, 14, 33–48. bhardwaj, r. k., & brooks, l. d. (1993). dual betas from bull and bear markets: reversal of the size effect. journal of financial research, 16, 269–283. blake, c. r., elton, e. j., & gruber, m. j. (1993). the performance of bond mutual funds. journal of business, 66, 371–403. appendix a3 expenses and turnover for comparison mutual fund categories category mean standard deviation median large value management fee 0.71 0.71 0.21 annual net expense ratio 1.43 1.30 1.00 turnover (%) 103.83 68.00 310.11 large growth management fee 0.71 0.71 0.21 annual net expense ratio 1.43 1.30 1.00 turnover (%) 103.83 68.00 310.11 mid value management fee 0.74 0.70 0.18 annual net expense ratio 1.37 1.25 0.55 turnover (%) 61.03 52.00 41.33 mid growth management fee 0.77 0.75 0.20 annual net expense ratio 1.55 1.43 0.66 turnover (%) 122.35 108.00 103.17 small value management fee 0.84 0.85 0.19 annual net expense ratio 1.43 1.35 0.44 turnover (%) 71.36 51.00 88.35 small growth management fee 0.86 0.85 0.23 annual net expense ratio 1.61 1.48 0.67 turnover (%) 127.75 100.00 129.63 multisector bond management fee 0.61 0.60 0.23 annual net expense ratio 1.22 1.15 0.51 turnover (%) 100.44 50.21 105.65 long term bond management fee 0.50 0.50 0.18 annual net expense ratio 1.09 1.00 0.51 turnover (%) 71.51 30.00 109.08 no. model number of factors 1 unconditional carhart 4 2 unconditional fung-hsieh 7 3 conditional carhart 20 4 conditional fung-hsieh 35 119s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 brooks, l. d., & porter, g. e. (2012). mutual fund performance attribution: 1994–2005. financial services review, 21, 259–273. broussard, j., & neely, w. (2011, september). mutual alternatives to hedge funds: market neutral and long/short funds. in midwest finance association 2012 annual meetings paper. brown, s. j., goetzmann, w., ibbotson, r. g., & ross, s. a. (1992). survivorship bias in performance studies. review of financial studies, 5(4), 553–580. brown, s. j., & goetzmann, w. n. (1995). performance persistence. the journal of finance, 50, 679–698. budiono, d. p., & martens, m. (2010). mutual funds selection based on funds characteristics. journal of financial research, 33, 249–265. carhart, m. m. (1997). on persistence in mutual fund performance. the journal of finance, 52, 57–82. chen, y. (2006). timing ability in the focus market of hedge funds. available at ssrn 687230. chen, y., & liang, b. (2007). do market timing hedge funds time the market? journal of financial and quantitative analysis, 42, 827–856. chen, z., & knez, p. j. (1996). portfolio performance measurement: theory and applications. review of financial studies, 9, 511–555. davis, j. l. (2001). mutual fund performance and manager style. financial analysts journal, 57, 19–27. detzel, f. l., & weigand, r. a. (1998). explaining persistence in mutual fund performance. financial services review, 7, 45–55. deli, d. n., & varma, r. (2002). contracting in the investment management industry: evidence from mutual funds. journal of financial economics, 63, 79–98. dellva, w. l., & olson, g. t. (1998). the relationship between mutual fund fees and expenses and their effects on performance. financial review, 33, 85–104. detzler, m. l. (1999). the performance of global bond mutual funds. journal of banking & finance, 23, 1195–1217. dowen, r. j., & mann, t. (2004). mutual fund performance, management behavior, and investor costs. financial services review, 13, 79–91. droms, w. g., & walker, d. a. (1996). mutual fund investment performance. the quarterly review of economics and finance, 36(3), 347–363. elton, e. j., gruber, m. j., & blake, c. r. (1995). fundamental economic variables, expected returns, and bond fund performance. the journal of finance, 50, 1229–1256. elton, e. j., gruber, m. j., & blake, c. r. (1996). survivor bias and mutual fund performance. review of financial studies, 9, 1097–1120. elton, e. j., gruber, m. j., & blake, c. r. (2011). holdings data, security returns, and the selection of superior mutual funds. journal of financial and quantitative analysis, 46, 341. fabozzi, f. j., & francis, j. c. (1979). mutual fund systematic risk for bull and bear markets: an empirical examination. the journal of finance, 34, 1243–1250. fama, e. f., & french, k. r. (1993). common risk factors in the returns on stocks and bonds. journal of financial economics, 33, 3–56. ferson, w. e., & schadt, r. w. (1996). measuring fund strategy and performance in changing economic conditions. the journal of finance, 51, 425–461. fung, w., & hsieh, d. a. (2001). the risk in hedge fund strategies: theory and evidence from trend followers. review of financial studies, 14, 313–341. fung, w., & hsieh, d. a. (2004). hedge fund benchmarks: a risk-based approach. financial analysts journal, 60, 65–80. goetzmann, w. n., & ibbotson, r. g. (1994). do winners repeat? the journal of portfolio management, 20, 9–18. goldman sachs asset management. (2012). breaking with tradition: mutual funds offer an alternative route; implementing non-traditional mutual fund investments within a core and satellite framework. gsam white paper gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds. the journal of finance, 51, 783–810. 120 s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 grinblatt, m., & titman, s. (1992). the persistence of mutual fund performance. the journal of finance, 47, 1977–1984. grossman, s. j., & stiglitz, j. e. (1980). on the impossibility of informationally efficient markets. the american economic review, 70, 393–408. haslem, j. a., baker, h. k., & smith, d. m. (2008). performance and characteristics of actively managed retail equity mutual funds with diverse expense ratios. financial services review, 17, 49–68. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974–1988. the journal of finance, 48(1), 93–130. henriksson, r. d., & merton, r. c. (1981). on market timing and investment performance. ii. statistical procedures for evaluating forecasting skills. journal of business, 54, 513–533. henriksson, r. d. (1984). market timing and mutual fund performance: an empirical investigation. journal of business, 57, 73–96. huij, j., & post, t. (2011). on the performance of emerging market equity mutual funds. emerging markets review, 12, 238–249. ibbotson, r., & patel, a. (2002). do winners repeat with style? (no. ysm 253). yale school of management. ippolito, r. a. (1989). efficiency with costly information: a study of mutual fund performance, 1965–1984. the quarterly journal of economics, 104, 1–23. jensen, m. c. (1968). the performance of mutual funds in the period 1945–1964. the journal of finance, 23, 389–416. kahn, r. n., & rudd, a. (1995). does historical performance predict future performance? financial analysts journal, 61, 43–52. kon, s. j., & jen, f. c. (1978). estimation of time-varying systematic risk and performance for mutual fund portfolios: an application of switching regression. the journal of finance, 33, 457–475. koski, j. l., & pontiff, j. (1999). how are derivatives used? evidence from the mutual fund industry. the journal of finance, 54(2), 791–816. lee, c. f., & rahman, s. (1990). market timing, selectivity, and mutual fund performance: an empirical investigation. journal of business, 63, 261–278. malkiel, b. g. (2003). the efficient market hypothesis and its critics. the journal of economic perspectives, 17, 59–82. otten, r., & bams, d. (2002). european mutual fund performance. european financial management, 8, 75–101. philpot, j. (2000). performance persistence and management skill in nonconventional bond mutual funds. financial services review, 9, 247–258. sewell, m. (2011). fund performance. university college of london, rn, 11, 03. treynor, j., & mazuy, k. (1966). can mutual funds outguess the market. harvard business review, 44, 131–136. webster, d. (2002). mutual fund performance and fund age. available at ssrn 1764543. wermers, r. (2000). mutual fund performance: an empirical decomposition into stock-picking talent, style, transactions costs, and expenses. the journal of finance, 55, 1655–1703. 121s. kanuri, r.w. mcleod / financial services review 23 (2014) 93–121 from the editor this issue contains issue 1 of volume 26 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “individual estimates of life expectancy and consumption patterns,” is coauthored derek r. lawson and stuart j. heckman, both at kansas state university. using the 2013 survey of consumer finances, the authors investigate the relationship between subjective life expectancy and three indicators of consumption: financial assets, credit card debt, and saving behavior. their results indicate that remaining work life expectancy and retirement life expectancy are associated with financial assets and that retirement life expectancy is associated with savings behavior. the second article “employees’ financial behaviors following the 2007–2009 financial crisis,” is coauthored by crystal hudson at clark atlanta university, wookjae heo at south dakota state university, heejung park at university of wyoming, and lance palmer at university of georgia. this paper is based on the foundation that lowand middle-income employees make up the bulk of potential participants in employer-sponsored retirement plans and employers find it difficult to alter the employees savings behavior. noting that financial crises may have unintended positive effects on low-income employees’ behavior, the authors examine the effect of the 2007–2009 financial crisis on employees’ financial behaviors using data from the survey of consumer finances. their results show that following the financial crisis, all employees’ and low-income employees’ savings behavior significantly improved and employees’ cash flow management behavior improved following the crisis, while it had no effect on low-income employees’ cash flow management behavior. the third article, “value line quarterly eps forecast error: analyst credibility or management appeasement?” is authored by philip baird at duquesne university. the author studies value line quarterly earnings forecast errors from 1999 through 2016 and shows that the direction of forecast bias and forecast efficiency with respect to earnings news depend on investment rating. he concludes that patterns of bias and inefficiency indicate that value line analysts are primarily motivated to maintain credibility with investors than to appease company managers. he also shows that for buy-rated stocks, forecast bias is pessimistic, and forecasts are inefficient with respect to good earnings news. when news is bad for buy-rated stocks, forecasts are unbiased and efficient. for sell-rated stocks, forecast bias is optimistic, financial services review 26 (2017) v–vi 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. and forecasts are inefficient with respect to bad earnings news. when news is good for sell-rated stocks, forecasts are unbiased and efficient. the fourth article, “personality and borrowing behavior: an examination of the role of need for material resources and need for arousal traits on household’s borrowing decisions” is coauthored by atefeh yazdanparast and yasser alhenawi, both at university of evansville. the authors examine the role of psychological characteristics of household decision makers in their borrowing decisions using the survey of consumer finances and direct measures of each surveyed household’s personality scores, relevant attitudes, and financial profiles. they examine the relationship between attitude towards borrowing and intentions to apply for specific borrowing options and inspect the role of personality traits in such decisions. their findings indicate that the attitude toward borrowing and the intuition to borrow are not always consistent and the discrepancies between the two vary across personalities, highlighting the role of personality traits in borrowing decisions. the authors report strong evidence that personal attitudinal biases towards money, risk, financial planning, and borrowing as well as certain demographic characteristics influence household’s borrowing behavior. the final article, “the impact of age differences and race on the social security early retirement decision for married couples: an extension with gender role reversals,” is coauthored by diane scott docking at northern illinois university, rich fortin at new mexico state university, and s. e. michelson at stetson university. the authors examine the impact of age differences on the social security early and delayed retirement decision for married couples. they develop an excel model to compute the breakeven internal rate of return for each of the race combinations. they find for couples the greater the age difference, the greater the incentive to retire early, as the hurdle rate is lower to overcome. because women almost always have higher be irrs than men, they find that in marriages where the wife is the breadwinner and the older partner, it is more difficult for the couple to retire early, as compared to marriages where the husband is the breadwinner and the older partner. hispanics have higher hurdle rates; while whites have lower hurdle rates. because of the difference in break-evens, hispanics have a more difficult time retiring early, while whites have a less difficulty retiring early. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 26 (2017) v–vi finser_23_1 does a relationship with a financial service professional overcome a client’s sense of not being in control of achieving their goals? danielle d. winchester, ph.d.a, sandra j. huston, ph.d.b aassistant professor, department of accounting & finance, north carolina a&t state university, school of business and economics, 1601 east market street, greensboro, nc 27411, usa bassociate professor, department of personal financial planning, college of human sciences, broadway & akron, lubbock, tx 79409 abstract although considerable evidence suggests that setting goals increases the odds of behavior changes that lead to goal attainment, less research has been conducted examining what underlying traits allow some to successfully attain their goals when others do not. this study, using the theory of planned behavior, examines the affects of individual control beliefs on financial goal progress. findings suggest that low control beliefs are significantly associated with less financial-goal progress; however, the receipt of expert financial advice can reduce this negative effect and result in higher levels of goal progress than that of individuals with high control beliefs. © 2014 academy of financial services. all rights reserved. keywords: behavioral control; self-efficacy; goal progress; financial advice jel classification: d14 1. introduction the establishment and pursuit of financial goals figures prominently in the financial planning process, and financial service professionals routinely emphasize the importance of developing clear specific, measureable, achievable, realistic, and timed (smart) autonocorresponding author. tel.: !1-336-256-3361; fax: !1-336-256-2055. e-mail address: ddwinche@ncat.edu (d.d. winchester) financial services review 23 (2014) 1–23 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. mous (reflecting personal interest and values) financial goals (koestner, otis, powers, pelletier, and gagnon, 2008; latham and locke, 1991; schunk, 1991). goals serve as motivational units for behavior modification by: directing attention and/or effort toward goal-relevant activities and away from goal irrelevant-activities; serving an energizing function, whereby high priority goals lead to greater effort than low priority goals; increasing persistence; and leading to the arousal, discovery, and/or use of task-related knowledge and strategies (latham and locke, 1991; phillips and gully, 1997). goals also increase the ability to concentrate when necessary, delay gratification, and follow instructions (sheldon and kasser, 1998). as such, the setting and evaluating of goals for financial success is a grounding pillar of the financial planning process as outlined by the certified financial planner board of standards, cfp board of standards.1 the setting of goals plays a critical role in aligning people’s behaviors with the actions necessary for goal attainment (bagozzi and dholakia, 1999; koestner, lekes, powers, and chicoine, 2002; neukam and hershey, 2003). stawski, hershey, and jacobs-lawson (2007) find that goal clarity motivates individuals to plan for retirement. cai and yang (2012) find goal clarity influences risk tolerance and ultimately risk-taking strategies in goal attainment. although researchers agree that certain types of goals are more likely to be initiated by individuals and that certain goal types are more likely than others to lead to behavioral change and ultimate goal attainment, very little is known in the finance domain as to why some people, with similar goals, are able to achieve their goals and others are not. personality traits and personal beliefs associated with goal setting behavior have been found to either facilitate or thwart the attainment of goals. as such, the medical field has embraced the inclusion of personality traits and beliefs into predicting behavior change and goal attainment. anecdotal evidence in the finance domain also supports this finding; o’neill et al. (2000) find that the most frequently reported resource needed for people to make progress toward financial goals is personal qualities; whereby personal qualities include “will power and determination,” “positive thinking,” “self control,” and “belief in self.” however, to date, empirical research investigating personal characteristics and goal progress, financial behaviors, or both, is limited to determining the influence of demographic characteristics on a person’s willingness to set goals or initiate goal-attaining behavior (chatterjee, greenpimentel, and turner, 2010), a person’s readiness to begin goal-based behaviors (o’neill, brennan, and bristow, 2001; xiao et al., 2003), and more recently and related, examining the association between locus of control and personality type (type a or b) on the likelihood of displaying certain financial behaviors and taking risk (britt, cumbie, and bell, 2013; carducci and wong, 1998). studies investigating the influence of personality traits on health-related behaviors are more common. however, there exist fundamental differences between financeand healthrelated goals, for example financial goals (1) take longer to realize, (2) typically are not as conducive to tangible measurements of progress, (3) are not completely under the volition of the goal-setter, (4) have less concrete outcomes, and (5) have higher variability in performance than health-related goals (eisingerich and bell, 2007). as such, this study theoretically articulates and empirically examines the affects of personal qualities and beliefs, more specifically, perceived control beliefs (self-efficacy and controllability), on financial goal progress. more expressly, do personal assessments of self-efficacy and controllability impact 2 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 financial goal progress similarly as they do health-related goals. it also seeks to examine if having an on-going committed relationship with a financial services professional differentially affects the goal attainment progress of individuals with high and low control beliefs. 2. literature review many financial goals are long-term in nature, require the occurrence of a particular event, or both, to know whether or not they have been achieved. for example, the goal of having $2.5 million at retirement may not be achieved for 30 to 40 years. likewise, many financial goals do not lend well to quantitative measurement. however, to achieve a financial goal, a person must display behaviors that are congruent with that goal, and it is these behaviors that culminate in the attainment of a goal. additionally, it is these behaviors, not the actual goal, that are under greater volitional control of the goal setter; for example, the behavior of contributing $2,000 per year to a coverdall education savings account is much more under the control of the goal-setter than amassing $30,000 in a college savings account; as such, most financial goals are stated in terms of implementation behaviors, the linking of goals with goal-directed behaviors (gollwitzer, 1999). therefore, this study uses the terms financial-goals and behaviors interchangeably (ajzen, 2002; bagozzi and dholakia, 1999). this review of literature will include studies that focus on goal setting (stating a goal intention), goal striving (initiating behaviors that will lead to goal attainment), personality traits and goal striving, and personality traits and advice. 3. goal setting, goal striving, and behavior change setting goals is a committing to reaching a desired outcome or to perform a desired behavior (gollwitzer, 1999). much of consumer behavior is goal-directed. personal goals are self-investments that provide individuals with a sense of purpose, structure, and identity (elliot, sheldon, and church, 1997). it is generally accepted that the formation of goals influences consumer behavior through four mechanisms: serving a directive function by directing attention, effort toward goal-relevant activities and away from goal irrelevantactivities, or both; providing an energizing function, whereby high priority goals lead to greater effort than lower priority goals; stimulating persistence; and leading to the use of task-related knowledge and strategies (latham and locke, 1991). conceptually, goal-oriented behavior can be classified into two phases: goal setting (i.e., intention) and goal striving. goal setting involves decision-making processes in which goals are identified (bagozzi and dholakia, 1999), and goal striving is the initiation of actions or behaviors that lead to, as well as the assessment of progress toward, goal attainment (gollwitzer and brandstätter, 1997). there is widespread agreement that achieving and effectively striving toward personal goals leads to heightened states of well-being (brunstein, 1993; christiansen, backman, little, and nguyen, 1999). feelings of competency and goal mastery are essential for one’s vitality and self-esteem and result in a significant decline in discrepancy reduction (i.e., the 3d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 engagement in non-goal related behaviors) and an increase in overall performance (elliot et al., 1997). goal striving is also linked to increases in goal-related thinking, motivation, and overall goal commitment (zhang and huang, 2010). schunk (1995) finds that people who perceive satisfactory goal progress feel capable of improving their skills and goal attainment, and litmanen, hirsto, and lonka (2010) find that students who perceived progress in attaining their education goals, advanced more rapidly in their studies. 4. personality traits and behavior change personality refers to internal factors that explain a person’s unique and relatively stable patterns of behavior (hogan, hogan, and roberts, 1996). researchers agree that personality is best characterized by the big-five personality traits, extraversion, openness to experience, emotional stability, conscientiousness, and agreeableness. zweig and webster (2004) investigate and compare the impact of these personality traits on performance-orientated goals and find that the big-five do not account for all outcome variance in predicating outcomes. as a result, self-efficacy is commonly paired with personality traits to predict academic achievement (caprara, vecchione, alessandri, gerbino, and barbaranelli, 2011), health-behavior changes (strecher, devellis, becker, and rosenstock, 1986), and academic-goal progress (lent et al., 2005). personality traits in addition to the big-five, such as rotter’s (1966) locus of control (schunk, 1990), perfectionism (powers, koestner, and topciu, 2005), and personality strengths, 24 distinct strengths that range from creativity to leadership to humor (linley, nielsen, gillett, and biswas-diener, 2010), have also been used to predict behavior changes. of these studies, only a very limited number have examined the impact of these traits on financial behaviors; of which, the majority focus on locus of control (joo, grable, and bagwell, 2003) or self efficacy (hilgert, hogarth, and beverly, 2003; perry and morris, 2005). for example, joo et al., (2003) find that college students with higher external locus of control have more positive attitudes toward credit card use, and britt et al., (2013) find that college students with an external locus of control exhibit the worst financial behavior. however, only two studies could be identified that investigate the impact of personality traits on financial goal striving, progress, or both. davis and hustvedt (2012) find that perceived behavioral control is the most important predictor in savings behaviors; however, the participants in this study had all received prior former consumer economics, personal finance education, or both (i.e., had higher than average levels of financial knowledge, self-efficacy, or both) and may or may not have used the assistance of a financial services professional in their financial decision making. xiao and wu (2008) use the theory of planned behavior (tpb) framework to investigate the impact of perceived behavioral control on the actual behavior of completing a debt management plan. 5. expert advice, personality traits, and behavior change in goal striving, it oftentimes becomes advisable to consider input from experts (koestner et al., 2001). in the health field it has been found that expert advice proves consistently 4 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 effective in promoting the recommended behaviors for health-goal attainment. expert recommendations can cause a change in exercise belief and behaviors among cancer patients. oncologists’ recommendations to exercise not only help patients develop successful strategies to maintain an exercise program (i.e., goal setting) but also result in greater levels of exercise (i.e., behavioral change) (ingram, wessel, and courneya, 2010; jones, courneya, fairey, and mackey, 2005). in the finance field, a survey reports that nearly 80% of respondents cite the greatest benefit of receiving expert financial advice as the “motivation to do what’s needed to meet retirement goals;” and empirically, persons who have met with a financial advisor are more likely to engage in goal-related retirement planning activities (marsden, zick, and mayer, 2011). however, research suggests that specific personality types may be more prone to seek and comply with experts’ advice than others. burish, carey, wallston, stein, jamison, and lyles (1984) contend that externally oriented patients (i.e., those with an external locus of control) may be more receptive than internally oriented (i.e., internal locus of control) patients to advice from health-care professionals. however, rosenstock (1985) finds no correlation between personality type and advice compliance when examining the behavioral changes of diabetic patients. 6. theoretical framework much of the work done on goal intention and striving has been done under the framework of the theory of reasoned actions. however, the progress made on financial goals is characterized as earned progress. earned progress, according to zhang and huong (2010), is the situation where a person attributes the degree of their advancement toward a goal to themselves (i.e., mostly under their own volition) and interpret that the progress reflects their own active pursuit of the goal. as such, the theory of planned behavior may serve as a better model for financial goal attainment, as this theory differs from reasoned action in its assertion that perceived behavioral control influences the likelihood of intentions and actions (madden, ellen, and ajzen, 1992). 7. theory of planned behavior 7.1. goals and beliefs ajzen and madden (1986) contend that goal intention and striving behaviors are guided by three consideration: (1) beliefs about the likely consequences or other attributes of the behavior (behavioral beliefs), (2) beliefs about the normative expectations of other people (normative beliefs), and (3) beliefs about the presence of factors that may thwart or facilitate the performance of the behavior (control beliefs). behavioral beliefs result from a favorable or unfavorable attitude toward the behavior; normative beliefs result from perceived social pressure concerning the goodness, appropriateness of the behavior, or both; and control 5d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 belief affect the perceived ease or difficulty in performing the behavior (webb, christian, and armitage, 2007). 7.2. beliefs’ impact on intentions and behaviors although all three beliefs play critical roles in goal initiation (i.e., setting) and completion, behavioral belief and normative beliefs occur during the first, “predecisional,” phase of goal striving and result in the formation of goal intentions, and these goal intentions are immediate antecedents of behavior (ajzen, 2002). goal intentions are significantly positively associated with goal striving (koestner et al., 2002); however, considerable evidence suggests that goal intentions do not necessarily translate into action. in fact the intentionbehavior correlation is quite small with up to 82% of the variance in behavior being left unexplained (mccaul, sandgren, o’neill, and hinsz, 1993; webb et al., 2007). additionally, gollwitzer (1999) finds that setting a goal for example, “i intend to do x,” and goal-attaining behavior are modestly correlated; intention accounts for between 20% and 30% of the variation in initiating behaviors that are in congruence with goal realization. the weak intention-behavior relation is hypothesized to be a result of people having good intentions but failing to act on them. on the other hand, control beliefs significantly add to the prediction of health behavioral action and explain between 34% and 43% of the variance in the display of the goal-achieving behavior (mccaul et al., 1993). control beliefs also indirectly impact behavior through increased intentions; a high level of perceived control increases effort and perseverance, strengthen a person’s intention to perform the behavior, and subsequently increase the likelihood of displaying appropriate behavior (ajzen, 2002). fig. 1 diagrams these relations as theorized by ajzen and madden (1986). according to the tpb, goal intention is the immediate antecedent of goal-directed behavior, and goal intentions are influenced by behavioral beliefs, normative beliefs, and control beliefs. this theory differs from the theory of reasoned behavior by its inclusion of control beliefs and theorizing that these beliefs not only impact behavior indirectly through intentions (illustrated by the dashed relation line), but also directly. people with low or no control beliefs believe that they have neither the resources nor the opportunities to perform or exhibit a certain behavior. they are unlikely to have strong behavioral intentions to engage in an action even if they hold a favorable attitude and believe that others would approve of them performing the behavior, that is, have strong behavioral and normative beliefs (ajzen and madden, 1986). on the other hand, given a sufficient degree of actual control over behavior (i.e., high control beliefs), people are expected to change their behavior when the opportunity arises (ajzen, 2002). 7.3. components of control beliefs: self-efficacy and controllability the more resources and opportunities a person believes he or she has regarding the performance of a specific behavior, the greater their perceived behavioral control (pbc), and perceived behavioral control is an independent determinant of behavior (madden et al., 1992). perceived behavioral control can, and usually does, vary across situations and actions 6 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 and is comprised of two components, perceived self-efficacy and controllability. these two components are operationalized by different indicators, but together they comprise the higher-order concept of pbc (see fig. 2). perceived self-efficacy is an internal assessment of whether one has the necessary resources, abilities, or talents to perform a specific behavior. it is based on a person’s perceived ability to execute a course of action required to attain a specific outcome or to perform a behavior (ajzen, 2002; mccaul et al., 1993; phillips and gully, 1997). this construct is a relatively stable personality quality. people who have a low sense of efficacy may avoid specific task, on the other hand, those who are more efficacious tend to work harder and persist longer when they encounter difficulties (schunk, 1991). fig. 1. impact of beliefs on goal initiation behaviors. based on the theory of planned behavior proposed by ajzen and madden (1986). fig. 2. hierarchical model of perceived behavioral control based on ajzen (2002). 7d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 controllability represents the extent to which performance is up to the person, as well as the perceived ease or difficulty of performing the behavior and the existence of opportunities to perform the behavior successfully (ajzen, 2002; mccaul et al., 1993). both controllability and self-efficacy can reflect both internal and external factors, and both have been found to account for significant portions of variance in behavior. 8. conceptual framework 8.1. additive moderation the tpb is used to address the differences in financial goal progress between individuals with varying level of perceived behavioral control. this work builds on the goal setting research of phillips and gully (1997) by examining the impact of both self-efficacy and controllability on the goal setting progress. it has also been suggested that there exists a moderating relation between controllability and self-efficacy on financial goal progress; therefore, this study uses an additive moderation model (perry and morris, 2005). it is suggested that controllability cannot immediately be increased, so this study seeks to investigate if its influence on behavior can be impacted by receiving expert advice. this investigation extends the investigative line of inquiry, as initiated by jones et al. (2005) in the health domain, by investigating the moderating affect of expert advice on perceived controllability. therefore, it tests if having an on-going committed relationship with a financial services professional differentially affects the goal attainment progress of individuals with high and low control beliefs. the conceptual model tested is presented in fig. 3. fig. 3. conceptual model of the impact of perceived controllability and self-efficacy on financial goal progress, including the moderating affect of expert financial advice on the effect of controllability on goal progress. 8 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 9. method 9.1. hypotheses based on the above conceptual model, this study tests the following hypotheses: hypothesis 1: individuals with lower perceived controllability will have lower levels of goal progress than those with higher level of perceived controllability. hypothesis 2: individuals reporting lower levels of self-efficacy will report lower levels of financial goal progress than those with higher levels of self-efficacy. hypothesis 3: the negative relation between lpc and financial goal progress will be further increased by lpse, that is, there is a moderating relation between perceived self-efficacy and controllability. hypothesis 4: the negative impact of lpc on financial goal progress will be lessened by the attainment of expert advice, that is, expert advice counteracts lpc in goal progress. 9.2. data and sample this study uses proprietary data cosponsored by a large independent financial services company and a financial planning professional association. a third party research company conducted the study online and collected the data in the summer of 2008 within the united states. the survey is designed to provide data that describe consumer attitudes and behavior in a changing economy. to be included in the study, respondents had to be adults and meet a threshold of having at least $50,000 in annual income or a minimum of $50,000 of investable assets. the sample contains data for 3,022 respondents. the individual is the unit of measure for nearly all of the questions asked. two variables, income and investable assets, are household level variables. because this study examines the impact of perceived behavior control on financial goal progress, the data are censored to only those respondents who have reported having a particular financial goal. as such, the sample for each financial goal may or may not be identical. 9.3. variable description and measurement 9.3.1. dependent variables bagozzi and dholakia (1999) use three items to measure action initiation in the goal setting and goal pursuit progress: (1) “how well have i enacted my plan,” “am i making progress toward my goal,” and “are there adjustments that need to be made,” likewise this study uses perceived progress in goal attainment as the measurement of behavior initiation. respondents that report having a goal were asked to rate how much progress had been made in achieving their financial goals, that is, to “rate your progress in achieving this goal.” the goals included (1) reducing taxes, (2) planning for retirement, (3) insurance or financial protection needs, (4) saving money and/or accumulating of wealth, (5) credit and/or debt 9d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 management, and (6) estate planning. the level of progress for these goals was measured on a 1 to 4 likert scale where 1 represents “little or no progress” and 4 represents “great deal of progress.” 9.3.2. independent variable controllability is a measure of whether the individual perceives an outcome to be under his or her control or under the control of outside forces. there is an abundance of research whereby beliefs about control of health would have been used to operationalize controllability in predicting health-related behavior (winefield, 1982). as such, this study converts respondent’s responses to “i have control of my finances” from a 1 to 5 likert scale to binary variable where 1 represents lpc (likert scale responses 1–3), and 0 represents hpc (likert scale responses 4–5). 9.3.3. moderating variables self-efficacy directs the choice to actively pursue a goal. self-efficacy is typically an assessment of an individual’s capacity to do what is required to accomplish a specific goal. an individual’s belief that he or she possesses the knowledge, skills, and or abilities required to achieve that goal is thought to instill confidence in goals attainment and to spur goaldirected efforts (affleck et al., 2001). following zimmerman and bandura (1994) who find a strong correlation between self-efficacy and perceived knowledge, self-efficacy is operationalized in this study by the respondents 1–5 likert scale response (1 " strongly disagrees to 5 " strongly agrees) to understanding financial-related issues.2 respondents were further classified into two groups, those with low level of perceived self-efficacy (likert scale responses 1–3) and those with hpse (likert scale responses 4–5). expert financial advice is measured as a binary variable, based on the respondent’s responses to, “do you have a written financial plan?” and “which of the following describes how your financial plan was developed?. . . comprehensive plan personalized for me after meeting and discussing with my financial planner about my goals.” this variable is coded as 1 for those who are and/or have received expert financial advice regarding their financial goals, responded yes, and as 0 for those who responded no. 9.3.4. control variables in addition to the main variable of interest, self-efficacy and controllability, certain sociodemographic variables are controlled for in this study. these variables theoretically may impact a person’s beliefs regarding attitude toward the behavior or perceptions of social norms, or play a critical role in goal progression and completion. these variables include gender, education level, income, and investable assets. by including these factors, this study attempts to control for the impact of one’s financial condition on goal progress as well as better isolate the impact of control beliefs on goal progress. 10 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 10. analysis 10.1. preliminary analyses table 1 presents the descriptive statistics and !2 statistics for the dependent, independent, and moderating variables, as well as the other constraint variables used in the study. approximately 10% and 27% of the sample have low levels of perceived control and self-efficacy, respectively; with lower self-efficacy being reported among persons with lpc. fifty-percent of those with lpc reported low perceptions of self-efficacy compared to 21% of those with hpc. this difference in reporting is significantly different among the two groups, as measured by a !2 statistic of 220.878 (p # 0.001). pearson correlation coefficients (see table 2) add further support to the descriptive statistics by highlighting a statistically significant positive correlation between perceived controllability and perceived self-efficacy; meaning having low controllability is associated with also having lpse (pearson correlation coefficient of 0.27, p # 0.001). table 1 between sample statistics variable total sample (n " 3,022) low controllability (n " 312) high controllability (n " 2,710) !2 statistic low perceived control 10.3% moderating variables low perceived self-efficacy 27.3% 50.3% 21.0% 220.878*** expert advice 24.9% 13.4% 28.1% 58.872*** financial goals reducing taxes 24.8% 16.1% 27.2% 33.317*** planning for retirement 66.5% 63.4% 67.3% 3.481 protecting assets 15.4% 14.6% 15.6% 0.379 saving and wealth accumulation 58.1% 52.1% 59.8% 12.194*** credit and debt management 28.1% 46.1% 23.1% 134.291*** estate planning 26.0% 17.1% 28.5% 34.334*** control variables female 41.5% 46.0% 40.3% 6.878** college education or more 72.4% 68.8% 73.4% 5.367* income less than $50,000 2.1% 1.7% 2.2% 0.625 $50,000–149,999 57.9% 67.2% 55.4% 29.322*** $150,000–249,999 26.2% 18.6% 28.3% 24.489*** $250,000 or more 24.4% 13.7% 27.4% 51.646*** investable assets less than $50,000 24.5% 40.1% 20.2% 109.360*** $50,000–249,999 19.9% 18.6% 20.3% 0.884 $250,000–999,999 20.6% 16.3% 21.7% 9.126** $1,000,000 or more 18.1% 13.7% 19.3% 10.709** note. this table reports the mean values of the control beliefs, report of having each specific financial goal, and other variables hypothesized to impact the perceived financial goal progress. means are compared between persons reporting a low perceived level of controllability and those with high perceived controllability. all data were included in this analysis. source: author’s estimates using proprietary data collected by an independent research company. *p # 0.05, **p # 0.01, ***p # 0.001. 11d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 t ab le 2 pe ar so n co rr el at io n m at ri x of lo w pe rc ei ve d co nt ro lla bi lit y an d se lf -e ffi ca cy ,u se of ex pe rt ad vi ce ,a nd go al pr og re ss m ea n sd 1 2 3 4 5 6 7 8 9 10 1: c on tr ol la bi lit y 0. 22 0. 41 1. 00 2: se lf -e ffi ca cy 0. 27 0. 44 0. 27 ** * 1. 00 3: e xp er t ad vi ce 0. 25 0. 43 $ 0. 14 ** * $ 0. 13 ** * 1. 00 4: sa vi ng fo r ed uc at io n 2. 41 0. 74 $ 0. 18 ** $ 0. 19 ** * 0. 28 ** * 1. 00 5: r ed uc in g ta xe s 2. 58 0. 68 $ 0. 22 ** * $ 0. 16 ** * 0. 11 ** 0. 57 ** * 1. 00 6: pl an ni ng fo r re tir em en t 2. 75 0. 67 $ 0. 34 ** * $ 0. 25 ** * 0. 21 ** * 0. 43 ** * 0. 34 ** * 1. 00 7: pr ot ec tin g as se ts 2. 82 0. 79 $ 0. 30 ** * $ 0. 29 ** * 0. 22 ** * 0. 30 0. 50 ** * 0. 55 ** * 1. 00 8: sa vi ng an d w ea lth ac cu m ul at io n 2. 58 0. 65 $ 0. 37 ** * $ 0. 28 ** * 0. 18 ** * 0. 44 ** * 0. 33 ** * 0. 65 ** * 0. 55 ** * 1. 00 9: c re di ta nd de bt m an ag em en t 2. 47 0. 65 $ 0. 29 ** * $ 0. 21 ** * 0. 11 ** 0. 42 ** * 0. 34 ** * 0. 31 ** * 0. 41 ** * 0. 53 ** * 1. 00 10 : e st at e pl an ni ng 2. 86 0. 83 $ 0. 18 ** * $ 1. 17 ** * 0. 20 ** * 0. 31 ** * 0. 48 ** * 0. 55 ** * 0. 48 ** * 0. 47 ** * 0. 41 ** * 1. 00 n ot e. t hi s ta bl e re po rt s th e pe ar so n co rr el at io n co ef fic ie nt s am on g th e in de pe nd en t, m od er at in g, an d de pe nd en t va ri ab le s. so ur ce : a ut ho r’ s es tim at es us in g pr op ri et ar y da ta co lle ct ed by an in de pe nd en t re se ar ch co m pa ny . *p # 0. 05 ,* *p # 0. 01 ,* ** p # 0. 00 1. 12 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 table 1 suggests that a person’s control beliefs may also impact the type of financial goals that are initiated as well as the goal progress attained. persons with lpc are statistically more likely to report having financial goals related to credit and debt management. nearly half of the persons reporting lpc have goals related to credit and debt management compared to roughly one-fifth of those with high controllability beliefs (!2 statistic of 12.194, p # 0.001). on the other hand, those with hpc more frequently report having goals related to reducing taxes and savings and wealth accumulation than those with lpc (27% compared to 16%, and 60% compared to 52%). there appears to be no difference in the frequency of reporting retirement savings, or asset protection goals between levels of perceived controllability. likewise perceived controllability also appears to impact the level of reported goal progress. table 3 illustrates that in all six goals, high controllability persons report with greater table 3 between sample goal progress statistics variable (likert scale; ascending level of progress) low controllability high controllability !2 statistic reducing taxes (n " 750) 1 14.3% 4.0% 45.402*** 2 55.4% 33.6% 3 25.7% 56.4% 4 4.8% 5.9% planning for retirement (n " 2008) 1 7.8% 1.5% 271.788*** 2 60.9% 24.8% 3 28.6% 61.9% 4 2.7% 11.8% protecting assets (n " 465) 1 19.0% 3.0% 51.485*** 2 35.8% 19.5% 3 36.8% 58.1% 4 8.4% 19.5% saving and wealth accumulation (n " 1,757) 1 13.6% 1.3% 274.943*** 2 65.8% 34.3% 3 19.2% 58.0% 4 1.5% 6.4% credit and debt management (n " 848) 1 6.0% 1.6% 76.124*** 2 67.7% 42.0% 3 24.3% 49.1% 4 2.0% 7.3% estate planning (n " 786) 1 12.6% 5.5% 33.034*** 2 39.6% 20.1% 3 34.2% 50.8% 4 13.5% 23.6% note. this table reports the frequency reporting of goal progress by respondents. means are compared between respondents of varying levels of perceived controllability (i.e., low vs. high). source: author’s estimates using proprietary data collected by an independent research company. percentages may not add to 100% because of rounding. *p # 0.05, **p # 0.01, ***p # 0.001. 13d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 frequency higher level of progress than low-controllability persons. the three goals that receive the highest progress for those with high controllability are those related to asset protection, estate planning, and retirement planning. conversely, those with lpc report higher levels of progress on goals related to estate planning, asset protection, and retirement planning, in descending order of progress. evidence is provided that personality type influences the seeking of expert advice. persons with high controllability more frequently report using expert financial advice than those with low levels of controllability (nearly 30% compared to nearly 15%; !2 statistic of 58.872, p # 0.001, table 1). additionally, both low controllability and low self-efficacy are significantly negatively associated with seeking expert advice (pearson correlation coefficient of $0.14 and $0.13, respectively; p # 0.001, table 2). the preliminary analyses further offers support hypothesis 1 and 2 by showing statistically significant negative relations between low controllability and low self-efficacy and progress on all six reported financial goals. 10.2. central analyses an ordinary least squares regression is used to test hypothesis 1–4. the additive statistical moderation model used is presented in fig. 4. six regression analyses were used in this study; controllability expert financial advice self-efficacy financial goal progress controllability * expert advice controllability * self-efficacy other constraints ey b1 c'1 c'2 c'4 b7 fig. 4. statistical additive moderation model of the impact of perceived controllability, expert advice, and self-efficacy on financial goal progress, including the moderating affect of expert financial advice and selfefficacy on the effect of perceived controllability on goal progress. 14 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 each modeled a unique financial goal’s progress—financial goals related to (1) reducing taxes, (2) retirement planning, (3) asset protection, (4) savings and wealth accumulation, (5) credit and debt management, and (6) estate planning. in fig. 4, c%1 represents the direct effect of perceived low controllability on financial goal progress; this parameter estimate addresses hypothesis 1; perceived low controllability is associated with lower levels of financial goal progress. hypothesis 2, the impact of lpse on goal progress, is assessed by parameter coefficient, b1. the significance of the moderating relationship of lpc and lpse is represented by b7 and that of lpc and expert advice is represented by c%4, representing the determination of hypothesis 3 and hypothesis 4, respectively. therefore, the conditional effect of controllability on goal progress is represented by the equation, c%1 ! b7 (selfefficacy) ! c%4 (expert advice). 10.3. direct effects lpc has a significant influence in goal progress for five of the six studied financial goals, lpc does not have a significant influence on estate planning goal progress. in three of the five goal areas, lpc reduces goal progress by one-half point or more. lpc reduces asset protection goal progress by 0.60 points (c%1 " $0.60, p # 0.001); savings and wealth accumulation goal progress by 0.51 points (c%1 " $0.51, p # 0.001); and tax reduction goals by 0.47 points (c%1 " $0.47, p # 0.001). progress in the two remaining impacted goals, retirement planning and credit and debt management, is reduced by 0.43 and 0.33 points, respectively. these finds lend support to hypothesis 1. see table 4 for complete regression results. lpse is consistently associated with lower financial goal progress; as represented by a significant b1 parameter estimate in all models of financial goals. the analysis suggest that lpse results in the greatest reduction in goal progress related to asset protection; a lpse decreases asset protection goal progress by 0.34 points (b1 " $0.34, p # 0.001). likewise, saving and wealth accumulation progress is reduced by 0.24 points (b1 " $0.24, p # 0.001); credit and debt management by 0.19 points (b1 " $0.19, p # 0.05); and between 0.18 and 0.11 points for the remaining three goal classifications, reducing taxes, retirement planning, and estate planning; lending support to hypothesis 2. expert advice does not consistently influence goal progress. receiving expert advice is associated with increased progress in saving and wealth accumulation (c%2 " 0.0692, p # 0.05) and estate planning (c%2 " 0.2537, p # 0.001). 10.4. moderating effects the conditional effect of lpc because of its interaction, which is, moderating relation, with low self-efficacy is not unanimously supported in this study. of the six financial goals examined, one, estate planning, reports a significant b7 parameter estimate (b7 " $0.4563, p # 0.05). this finding does not lend support the hypothesis 3; lpse does not appear to influence the effect of lpc on goal progress. analyses suggest a significant moderating relation between lpc and expert advice; in four of the six models the parameter estimate c%4 was positively significant. these models were 15d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 t ab le 4 c om pl et e o l s re gr es si on an al ys es r ed uc in g ta xe s r et ir em en t pl an ni ng a ss et pr ot ec tio n sa vi ng an d w ea lth ac cu m ul at io n c re di t an d de bi t m an ag em en t e st at e pl an ni ng pa ra m et er es tim at e se pa ra m et er es tim at e se pa ra m et er es tim at e se pa ra m et er es tim at e se pa ra m et er es tim at e se pa ra m et er es tim at e se l ow pe rc ei ve d co nt ro lla bi lit y (c % 1) $ 0. 17 9* 0. 07 2 $ 0. 43 4* ** 0. 04 8 $ 0. 59 6* ** 0. 12 2 $ 0. 51 1* ** 0. 04 8 $ 0. 33 3* ** 0. 06 3 $ 0. 11 5 0. 13 2 l ow pe rc ei ve d se lf -e ffi ca cy (b 1 ) $ 0. 46 8* ** 0. 10 2 $ 0. 15 2* ** 0. 03 7 $ 0. 34 0* ** 0. 08 9 $ 0. 24 3* ** 0. 03 9 $ 0. 19 3* * 0. 06 1 $ 0. 11 2 0. 08 6 e xp er t ad vi ce (c % 2) 0. 05 0 0. 05 7 0. 06 3 0. 03 3 0. 15 4 0. 08 3 0. 06 9* * 0. 03 4 0. 04 6 0. 06 9 0. 25 4* ** 0. 06 4 l ow co nt ro lla bi lit y ! lo w se lf -e ffi ca cy (b 7 ) 0. 10 1 0. 15 1 $ 0. 06 7 0. 06 9 0. 62 8 0. 22 5 0. 11 9 0. 07 3 0. 02 6 0. 09 3 $ 0. 45 6* 0. 18 2 l ow co nt ro lla bi lit y ! ex pe rt ad vi ce (c % 4) 0. 39 5* 0. 17 1 0. 36 3* ** 0. 09 3 0. 62 8* * 0. 22 5 0. 24 2* 0. 09 6 0. 03 4 0. 13 4 0. 06 3 0. 19 5 fe m al e 0. 01 7 0. 05 1 $ 0. 01 6 0. 02 7 $ 0. 06 8 0. 06 7 $ 0. 05 2 0. 02 8 $ 0. 03 2 0. 04 4 $ 0. 08 2 0. 05 9 c ol le ge ed uc at io n or m or e $ 0. 03 6 0. 05 8 0. 00 3 0. 03 1 0. 19 5* * 0. 07 3 0. 05 9 0. 03 1 0. 06 3 0. 04 6 $ 0. 15 7* 0. 06 8 in co m e 0. 06 5 0. 12 0 $5 0, 00 0– 14 9, 99 9 0. 01 9 0. 12 6 0. 03 3 0. 08 6 $ 0. 02 9 0. 18 5 $ 0. 04 9 0. 07 3 0. 04 4 0. 13 3 0. 09 9 0. 14 1 $1 50 ,0 00 –2 49 ,9 99 0. 04 2 0. 12 6 0. 06 7 0. 08 8 $ 0. 01 8 0. 19 6 $ 0. 03 2 0. 07 6 $ 0. 04 3 0. 14 0 0. 15 4 0. 14 9 $2 50 ,0 00 or m or e 0. 04 6 0. 08 9 0. 09 2 0. 19 8 $ 0. 03 8 0. 07 7 $ 0. 13 7 0. 14 6 0. 12 9 0. 15 0 in ve st ab le as se ts $ 0. 00 1 0. 08 4 $5 0, 00 0– 24 9, 99 9 0. 08 1 0. 07 7 0. 16 2* ** 0. 03 8 0. 02 5 0. 09 7 0. 10 2* 0. 04 0 $ 0. 03 5 0. 05 8 0. 11 9 0. 10 9 $2 50 ,0 00 –9 99 ,9 99 0. 19 5* 0. 08 1 0. 32 5* ** 0. 03 8 0. 22 9* 0. 09 4 0. 32 2* ** 0. 03 9 0. 16 3* * 0. 05 9 0. 10 5 0. 09 6 $1 ,0 00 ,0 00 or m or e 0. 50 1* ** 0. 04 2 0. 20 9* 0. 10 2 0. 49 5* ** 0. 04 2 0. 31 4* * 0. 08 6 0. 30 6* * 0. 09 6 o bs . 75 0 2, 00 8 46 5 1, 75 7 84 8 78 6 a dj us te d r 2 0. 08 61 0. 24 53 0. 22 5 0. 28 21 0. 13 70 0. 11 28 *p # 0. 05 ,* *p # 0. 01 ,* ** p # 0. 00 1. 16 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 t ab le 4 c on tin ue d sa vi ng an d w ea lth ac cu m ul at io n c re di t an d de bt m an ag em en t e st at e pl an ni ng pa ra m et er es tim at e se pv al ue pa ra m et er es tim at e se pv al ue pa ra m et er es tim at e se pv al ue l ow pe rc ei ve d co nt ro lla bi lit y (c % 1) $ 0. 51 1 0. 04 8 0. 00 0 $ 0. 33 3 0. 06 3 0. 00 0 $ 0. 11 5 0. 13 2 0. 38 3 l ow pe rc ei ve d se lf -e ffi ca cy (b 1 ) $ 0. 24 3 0. 03 9 0. 00 0 $ 0. 19 3 0. 06 1 0. 00 2 $ 0. 11 2 0. 08 6 0. 19 4 e xp er t ad vi ce (c % 2) 0. 06 9 0. 03 4 0. 00 2 0. 04 6 0. 06 9 0. 50 7 0. 25 4 0. 06 4 0. 00 0 l ow co nt ro lla bi lit y ! lo w se lf -e ffi ca cy (b 7 ) 0. 11 9 0. 07 3 0. 10 2 0. 02 6 0. 09 3 0. 77 9 $ 0. 45 6 0. 18 2 0. 01 3 l ow co nt ro lla bi lit y ! ex pe rt ad vi ce (c % 4) 0. 24 2 0. 09 6 0. 01 2 0. 03 4 0. 13 4 0. 80 0 0. 06 3 0. 19 5 0. 74 7 fe m al e $ 0. 05 2 0. 02 8 0. 06 2 $ 0. 03 2 0. 04 4 0. 46 9 $ 0. 08 2 0. 05 9 0. 16 0 c ol le ge ed uc at io n or m or e 0. 05 9 0. 03 1 0. 06 2 0. 06 3 0. 04 6 0. 17 3 $ 0. 15 7 0. 06 8 0. 02 1 in co m e $5 0, 00 0– 14 9, 99 9 $ 0. 04 9 0. 07 3 0. 50 6 0. 04 4 0. 13 3 0. 74 2 0. 09 9 0. 14 1 0. 48 3 $1 50 ,0 00 –2 49 ,9 99 $ 0. 03 2 0. 07 6 0. 67 9 $ 0. 04 3 0. 14 0 0. 76 1 0. 15 4 0. 14 9 0. 30 3 $2 50 ,0 00 or m or e $ 0. 03 8 0. 07 7 0. 62 5 $ 0. 13 7 0. 14 6 0. 34 8 0. 12 9 0. 15 0 0. 39 2 in ve st ab le as se ts 0. 01 1 $5 0, 00 0– 24 9, 99 9 0. 10 2 0. 04 0 0. 00 0 $ 0. 03 5 0. 05 8 0. 54 5 0. 11 9 0. 10 9 0. 27 7 $2 50 ,0 00 –9 99 ,9 99 0. 32 2 0. 03 9 0. 00 0 0. 16 3 0. 05 9 0. 00 6 0. 10 5 0. 09 6 0. 27 2 $1 ,0 00 ,0 00 or m or e 0. 49 5 0. 04 2 0. 31 4 0. 08 6 0. 00 3 0. 30 6 0. 09 6 0. 00 2 o bs . 1, 75 7 84 8 78 6 a dj us te d r 2 0. 28 21 0. 13 70 0. 11 28 *p # 0. 05 ,* *p # 0. 01 ,* ** p # 0. 00 1. 17d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 financial goals related to reducing taxes, retirement planning, asset protection, and savings and wealth accumulation. see fig. 5 for a visual representation of this subsuming relation in financial goals related to protecting financial assets. this figure illustrates how a positively significant moderating parameter estimate suggests lpc supplemented by expert financial advice results in greater goal progress than that of both lpc and hpc without expert advice, as well as hpc with expert advice. additionally, the inclusion of the moderating relation increases the overall explanability of the models. the inclusion of the interaction variable (lpc ! expert advice) increases the adjusted r-square of the models by 8%, 2%, 6%, and 1% for tax, retirement, protection, and wealth goals, respectively. this findings offer credence to hypothesis 4. in the remaining two models, credit and debt management and estate planning, there is no evidence to support that the goal progress level for those with lpc and expert advice is statistically different from the expertly advised with both hpc and lpc or those with low controllability and are not expertly advised. 11. discussion and implications it is common practice for individuals to set financial goals, and as individuals become more responsible for their financial futures being able to achieve these goals is more imperative. this study provides a strong support for a model linking personality traits and goal striving. more specifically, this study suggests that the tpb can be used to model goal striving in the finance domain by finding that lpbc is associated with low financial goal progress and/or striving. this knowledge may be helpful in determining why some people fig. 5. to help visualize and interpret the nature of the moderation of perceived controllability’s effect in asset protection goal progress, predicted values of goal progress were generated using various values of perceived controllability and the moderator. the covariates in the model were set to their sample mean when deriving the predicted values. this form of visualization is consistent with preacher and hayes (2004). 18 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 with the same financial goals are able to achieve their goals and others are not, as well as identifying people who may significantly benefit from expert advice in goal attainment. financial service professionals can use these findings to not only better serve their current client base, but also to appeal to lpbc persons who might not have considered seeking expert advice in the past. for the average person financial advice is positively associated with goal progress, and persons who receive expert advice have significantly higher goal progress than those without expert advice in goals related to savings and wealth accumulation and estate planning and marginally significant influence on retirement-planning goals, ceteris paribus. it also suggests an opportunity for advisors to increase their value to clients as related to attainment of retirement, because these goals have been cited as the most important financial goal for many people. in attracting a new clientele, these findings can be used to encourage lpbc persons, who would normally not seek expert advice, to seek advice in attaining their financial goals through marketing campaigns, referral request, and other communications. lpse has a dual negative impact on financial success. first it impedes goal progress as well as erodes goal motivation and second, it creates a self-fulfilling prophecy of failure and erodes overall well being. perceived self-efficacy can be improved through appropriate financial education as well as through positive persuasive feedback and decision anxiety reducing techniques. additionally, tackling moderately difficult, short-term financial goals before moving on to longer-termed more complex goals allows a client to do well and associate this success with his or her abilities; thereby, increasing self-efficacy. although lpc cannot be increased in the short-run, receiving expert advice reduces its negative influence on goal progress. in some financial areas there is no evidence to support that the goal attainment of persons with lpc is any different from that of persons with hpc, and in other goal domains, such as tax reduction, retirement planning, asset protection, and wealth accumulation, the goal progress of lpc persons exceeds that of all others. this is especially important as more individuals report planning for retirement as their main financial goal. these findings give more credence to value of expert financial advice and the role of advice in helping people with varying personality traits reach their financial goals. 11.1. data limitations and future research a possible limitation of this study is that the wealth and/or income requirements to be included in the survey (at least $50,000 in investable assets or income) may single out a portion of the u.s. population that has higher than average control beliefs, that is, those with low control beliefs may largely exist in the lower-income segments of the population. as such, the sample used may not contain an adequate representation of individuals with low control beliefs. however, when comparing the percentage of individuals in this sample that are of low control beliefs to that of a recognized nationally representative sample, the 1979 national longitudinal survey of youth (nlsy), the percentage of low control belief individuals is higher in the sample. therefore, the income and/or asset restrictions to be included in the study does not inhibit the number of low-control individuals and there are adequate members of the low controlled group to ensure reliable estimates to be reported for that group. other data limitations include the small sample sizes associated with individuals 19d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 that utilize professional financial advice. it is well established in the literature that the number of households that use professional financial advice is small. despite the data’s limitations, they offer an opportunity to begin to examine the role of pbc on an individual’s ability and willingness to pursue financial goals. these data afford the opportunity to measure respondent’s perceived financial control as measured by questions specifically related to one’s level of control over his or her finances instead of the rotter-locus of control scale, which assesses general loci of control and has been used in prior studies. they also allow for ample control variables to be used in the analysis that better allow for the isolation of the effect of pbc on goal attainment from the respondent’s financial condition. future studies of pbc on goal attainment should strive to address the limitations of this study as well as contain ample survey questions pertaining to personality traits, in addition to self-efficacy, that might impact financial goal progress. notes 1 the cfp is a regulatory organization that creates and enforces uniform standards of competency, practice and ethics of financial planners. 2 this measure of self-efficacy serves as both a measure of both perceived knowledge and perceived ability. bell and kozlowski (2002). find that high-ability individuals have the capabilities to do well on the difficult aspects of tasks and are therefore are expected to experience higher levels of self-efficacy. the present research, therefore, examines in more detail the interaction between goal achievement and an individual’s level of cognitive ability. acknowledgment the author would like to acknowledge the journal reviewers for feedback and comments in this project. the results, conclusions, and any errors are those of the author. this manuscript has not been published elsewhere and it has not been submitted simultaneously for publication elsewhere. references affleck, g., tennen, h., zautra, a., urrows, s., abeles, m., & karoly, p. (2001). women’s pursuit of personal goals in daily life with fibromyalgia: a value-expectancy analysis. journal of consulting and clinical psychology, 69, 587. ajzen, i. (2002). perceived behavioral control, self-efficacy, locus of control, and the theory of planned behavior. journal of applied social psychology, 32, 665–683. ajzen, i., & madden, t. j. (1986). prediction of goal-directed behavior: attitudes, intentions, and perceived behavioral control. journal of experimental social psychology, 22, 453–474. bagozzi, r., & dholakia, u. (1999). goal setting and goal striving in consumer behavior. the journal of marketing, 63(4), 19–32. 20 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 bell, b. s., & kozlowski, w. (2002). goal orientation and ability: interactive effects on self-efficacy, performance, and knowledge. journal of applied psychology, 87, 497, 2002. britt, s., cumbie, j., & bell, m. (2013). the influence of locus of control on student financial behavior. the college student journal, forthcoming. brunstein, j. (1993). personal goals and subjective well-being: a longitudinal study. journal of personality and social psychology, 65, 1061, 1993. burish, t., carey, m., wallston, k., stein, m., jamison, r., & lyles, j. (1984). health locus of control and chronic disease: an external orientation may be advantageous. journal of social and clinical psychology, 2, 326–332. cai, y., & yang, n. (2012). financial goal clarity and risk tolerance: an experimental investigation. consumer interests annual, 58, 1–4. caprara, g., vecchione, m., alessandri, g., gerbino, m., & barbaranelli, c. (2011). the contribution of personality traits and self-efficacy beliefs to academic achievement: a longitudinal study. british journal of educational psychology, 81, 78–96, 2011. carducci, b., & wong, a. (1998). type a and risk taking in everyday money matters. journal of business and psychology, 12, 355–359. chatterjee, s., green-pimentel, l., & turner, p. (2010). financial education and consumers’ willingness to change behavior. (available at http://businessperspectives.org/journals_free/imfi/2010/imfi_en_2010_03_ chatterjee.pdf) christiansen, c., backman, c., little, b., & nguyen, a. (1999). occupations and well-being: a study of personal projects. the american journal of occupational therapy, 53, 91–100. davis, k., & hustvedt, g. (2012). it’s a matter of control: saving for retirement. international review of social sciences and humanities, 3, 248–261. eisingerich, a., & bell, s. (2007). maintaining customer relationships in high credence services. journal of service marketing, 21, 253–262. elliot, a., sheldon, k., & church, m. (1997). avoidance personal goals and subjective well-being. personality and social psychology bulletin, 23, 915–927. gollwitzer, p. (1999). implementation intentions: strong effects of simple plans. american psychologist, 54, 493. gollwitzer, p., & brandstätter, v. (1997). implementation intentions and effective goal pursuit. journal of personality and social psychology, 73, 186. hilgert, m., hogarth, j., & beverly, s. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, july, 309–322. hogan, r., hogan, j., & roberts, b. (1996). personality measurement and employment decisions: questions and answers. american psychologist, 51, 469. ingram, c., wessel, j., & courneya, k. (2010). women’s perceptions of home-based exercise performed during adjuvant chemotherapy for breast cancer. european journal of oncology nursing, 14, 238–243. jones, l., courneya, k., fairey, a., & mackey, j. (2005). does the theory of planned behavior mediate the effects of an oncologist’s recommendation to exercise in newly diagnosed breast cancer survivors? results from a randomized controlled trial. health psychology, 24, 189. joo, s., grable, j., & bagwell, d. (2003). credit card attitudes and behaviors of college students. college student journal, 37, 405–419. koestner, r., gingras, i., abutaa, r., losier, g., didio, l., & gagné, m. (2001). to follow expert advice when making a decision: an examination of reactive versus reflective autonomy. journal of personality, 67, 851–872. koestner, r., lekes, n., powers, t., & chicoine, e. (2002). attaining personal goals: self-concordance plus implementation intentions equals success. journal of personality and social psychology, 83, 231. koestner, r., otis, n., powers, t. a., pelletier, l., & gagnon, h. (2008). autonomous motivation, controlled motivation, and goal progress. journal of personality, 76, 1201–1230. latham, g., & locke, e. (1991). self-regulation through goal setting. organizational behavior and human decision processes, 50, 212–247. lent, r., singley, d., sheu, h., gainor, k., brenner, b., treistman, d., et al. (2005). social cognitive predictors 21d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 of domain and life satisfaction: exploring the theoretical precursors of subjective well-being. journal of counseling psychology, 52, 429. linley, p., nielsen, k., gillett, r., & biswas-diener, r. (2010). using signature strengths in pursuit of goals: effects on goal progress, need satisfaction, and well-being, and implications for coaching psychologists. international coaching psychology review, 5, 6–15. litmanen, t., hirsto, l., & lonka, k. (2010). personal goals and academic achievement among theology students. studies in higher education, 35, 195–208. madden, t., ellen, p., & ajzen, i. (1992). a comparison of the theory of planned behavior and the theory of reasoned action. personality and social psychology bulletin, 18, 3–9. marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32(4), 625–643. mccaul, k., sandgren, a., o’neill, h., & hinsz, v. (1993). the value of the theory of planned behavior, perceived control, and self-efficacy expectations for predicting health-protective behaviors. basic and applied social psychology, 14, 231–252. neukam, k., & hershey, d. (2003). financial inhibition, financial activation, and saving for retirement. financial services review, 12, 19–38. o’neill, b., brennan, p., & bristow, b. (2001). application of the transtheoretical model of change to financial behavior. consumer interests annual, 47, 1–9. o’neill, b., xiao, j., bristow, b., brennan, p., & kerbel, c. (2000). successful financial goal attainment: perceived resources and obstacles. financial counseling and planning, 11, 1–12. perry, v., & morris, m. (2005). who is in control? the role of self-perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39, 299–313. phillips, j., & gully, s. (1997). role of goal orientation, ability, need for achievement, and locus of control in the self-efficacy and goal–setting process. journal of applied psychology, 82, 792. powers, t., koestner, r., & topciu, r. (2005). implementation intentions, perfectionism, and goal progress: perhaps the road to hell is paved with good intentions. personality and social psychology bulletin, 31, 902–912. preacher, k., & hayes, a. (2004). spss and sas procedures for estimating indirect effects in simple mediation models. behavior research methods, 36, 717–731. rosenstock, i. (1985). understanding and enhancing patient compliance with diabetic regimens. diabetes care, 8, 610–616. rotter, j. (1966). generalized expectancies for internal versus external control of reinforcement. psychological monographs: general and applied, 80, 1. schunk, d. (1990). goal setting and self-efficacy during self-regulated learning. educational psychologist, 25, 71–86. schunk, d. (1991). self-efficacy and academic motivation. educational psychologist, 26, 207–231. schunk, d. (1995). self-efficacy, motivation, and performance. journal of applied sport psychology, 7, 112–137. sheldon, k., & kasser, t. (1998). pursuing personal goals: skills enable progress, but not all progress is beneficial. personality and social psychology bulletin, 24, 1319–1331. stawski, r., hershey, d., & jacobs-lawson, j. (2007). goal clarity and financial planning activities as determinants of retirement savings contributions. the international journal of aging and human development, 64, 13–32. strecher, v., devellis, b., becker, m., & rosenstock, i. (1986). the role of self-efficacy in achieving health behavior change. health education & behavior, 13, 73–92. webb, t. l., christian, j., & armitage, c. j. (2007). helping students turn up for class: does personality moderate the effectiveness of an implementation intention intervention? learning and individual differences, 17, 316–327. winefield, h. (1982). reliability and validity of the health locus of control scale. journal of personality assessment, 46, 614–619. xiao, j. j., o’neill, b., prochaska, j. m., kerbel, c. m., brennan, p., & bristow, b. j. (2003). a consumer 22 d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 education programme based on the transtheoretical model of change. international journal of consumer studies, 28, 55–65. xiao, j. j., & wu, j. (2008). completing debt management plans in credit counseling: an application of the theory of planned behavior. journal of financial counseling and planning volume, 19(2), 29–45. zhang, y., & huang, s. (2010). how endowed versus earned progress affects consumer goal commitment and motivation. journal of consumer research, 37, 641–654. zimmerman, b., & bandura, a. (1994). impact of self-regulatory influences on writing course attainment. american educational research journal, 31, 845. zweig, d., & webster, j. (2004). what are we measuring? an examination of the relationships between the big-five personality traits, goal orientation, and performance intentions. personality and individual differences, 36, 1693–1708. 23d.d. winchester, s.j. huston / financial services review 23 (2014) 1–23 hedged etfs: do they add value? srinidhi kanuria,* adepartment of finance real estate and business law, college of business, the university of southern mississippi, scianna hall, 118 college drive, #5076, hattiesburg, ms 39406, usa abstract hedged exchange traded funds (etfs) provide individual investors with the opportunity to invest in etfs that follow strategies similar to those of hedge funds and seek returns uncorrelated with the market. in this article i analyze the performance of six different categories of 49 hedged etfs and 539 hedged mutual from january 2008 to december 2014, and compared them with five different asset categories of index etfs. hedged etfs and mutual funds had highly negative or low correlation with other index etfs which indicates that they did help investors diversify. hedged etfs also had much lower risk compared with other index etfs with the exception of bond market etf agg. however, this did not translate into superior absolute or risk-adjusted performance, and hedged etfs underperformed all other asset categories (with the exception of commodities etf dbc). the absoluteand risk-adjusted performance of hedged mutual funds was similar to that of hedged etfs. based on these findings investors would have been better off with index fund etfs. © 2016 academy of financial services. all rights reserved. keywords: hedge funds; investments; exchange traded funds (etfs); index fund etfs 1. introduction hedged exchange traded funds (etfs) are relatively new entrants into the etf industry. these etfs follow investment strategies similar to those of hedge funds and are attractive to individual investors who are often unable to invest in hedge funds because of high initial investment requirements and longer lock-up periods. these hedged etfs offer hedge-fundlike strategies for a fraction of the cost, and zero restrictions around getting into or out of these funds. these etfs normally have a goal of providing individual investors with access * corresponding author. tel.: �1-601-266-4809; fax: �1-601-266-6110. e-mail address: srinidhi.kanuri@usm.edu financial services review 25 (2016) 181–198 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. to investment strategies that offer non-correlated returns and diversification benefits. this goal of hedged etfs is in contrast with traditional or long only etfs that try to replicate a benchmark such as s&p 500 or russell 3000. there are a number of ways etf can act or replicate hedge fund returns.1 the different methods are direct approach, hedge fund replication, and copycat (see appendix b for a brief description of these three different methods). 2. motivation this article looks at the merits of holding hedged etfs versus holding different (asset) categories of index etfs. there has been a significant increase in the number of hedged etfs. lot of retail and institutional investors are increasingly drawn to etfs that aim to mimic hedge-fund strategies. as of december 2014, there were 34 live and 15 dead hedged etfs. the assets under management (aum) under the surviving hedged etfs as of december 2014 were $3.42 billion. with the increase in the number of funds and the growth in assets under management it is obvious that investors thought that hedged etfs would provide higher risk adjusted returns or benefits from diversification. 3. literature this is the first article that looks at the characteristics and performance of hedged etfs as an asset class. previous literature has only looked at hedge funds or alternative or hedged mutual funds (amfs). although amfs are relatively new, there has been some research in this field. koski and pontiff (1999) and deli and varma (2002) find that the flexibility to use derivatives, sell securities short, and borrow money to create leverage help managers to control expenses, risk, and manage cash flows more efficiently that makes the amfs appear to be an attractive alternative to standard mutual funds and subject to analysis. agarwal et al., (2009) were the first to look at the performance of 52 hedged mutual funds over the period 1994–2004. they form a single portfolio of six different categories of 52 hedged mutual funds from 1994 to 2004 and compare them to traditional mutual funds and hedge funds. they find that these hedged mutual funds outperform traditional mutual funds, but underperform similar hedge funds. kanuri and mcleod (2014) conduct a similar study of 256 amfs period january 1998 through december 2011using the carhart four-factor model and the fung-hsieh seven-factor model. their results indicate that most amfs have not been able to create any value for their investors over this period. furthermore, the performance of these mutual funds was even worse during the recent financial crisis (october 2007 through march 2009). this article looks at the performance of surviving as well as dead hedged etfs since their inception and compare them to u.s. stock market (ivv), aggregate bond market (agg), total world ex u.s. (veu), real estate market (iyr), and commodities market (dbc). we compare the performance of hedged etfs to different index etfs for an equal comparison or compare performance after expenses. elton, gruber, and blake (1996) find that previous 182 s. kanuri / financial services review 25 (2016) 181–198 mutual fund studies suffered from survivorship bias as funds that merge or die have worse performance than funds that do not and failing to account for survivorship bias will lead to higher risk-adjusted returns for mutual funds. excluding dead etfs can lead to similar problems. therefore, dead etfs were included in the analysis to control for survivorship bias. following agarwal et al., (2009), a single portfolio of all hedged etfs (surviving and dead) is formed from january 2008 through december 2014. 4. hypothesis in this article i am interested in determining whether or not hedged etfs can provide benefits to individual investor through their promise of delivering returns that are uncorrelated with the market. do hypothesis 1: hedged etfs have more flexibility than long-only etfs (which try to which try to mimic a benchmark such as s&p 500 or russell 3000). they can take long (short) positions in undervalued (overvalued) securities. additionally, they can use derivatives (including forwards, options, or swaps) to seek absolute returns that is, positive returns in all market conditions. therefore, they should have lower correlation and better absoluteand risk-adjusted performance compared to traditional benchmark etfs. hypothesis 2: because of the reasons mentioned in hypothesis 1, hedged etfs should help investors diversify, and have low correlation with other asset categories. 5. data and methodology the list of hedged etfs has been taken morningstar direct database. hedged etfs were first created in late 2007. appendix a shows the different categories of hedged etfs. therefore, for an equal comparison, equally weighted portfolios for hedged etfs have been formed monthly from january 2008 through december 2014. there are total of 34 live and 15 dead hedged etfs at the end of december 2014. we include all surviving hedged etfs in the analysis that have at least 12 months of returns as of december 2014. we also compare the performance of hedged etfs to the u.s. stock market that is, s&p 500 etf (ivv), foreign stock market etf that is, ftse all world ex u.s. (veu), barclays aggregate market etf (agg), u.s. real estate etf (iyr), and commodity etf that is, powershares db commodity tracking etf (dbc). we also form two different portfolios: 1. 65% u.s. stock market (ivv)/35% bond market (agg) – following stout and mitchell (2006) and brown et al., (2003), a portfolio of 65% in a broad index of index of equities of u.s. corporations and 35% intermediate term bonds is formed. this is also the allocation for rep. demint’s social security savings act of 2003. 2. 45% u.s. stock market (ivv)/10% foreign stock market (veu)/5% real estate market (veu)/5% commodity market (dbc)/35% bond market (agg) – as a 183s. kanuri / financial services review 25 (2016) 181–198 robustness test, another portfolio comprising of 45% u.s. stocks,10% foreign stocks, 5% real estate, 5% commodities, and 35% bonds is also used for comparison purposes. appendix c shows all the hedged etfs and their inception date. appendix d shows the different index etfs used for comparison. the monthly returns, annual expenses, annual turnover, and assets under management at the end of each year for hedged etfs, ivv, agg, veu, iyr, and dbc have been obtained from morningstar direct database. 6. results table 1 shows descriptive statistics. the total aum for 34 surviving hedged etfs, ivv, agg, veu, iyr, and dbc of december 2014 were $3.42 billion, $69.69 billion, 23.02 billion, $24.09 billion, $12.27 billion, $6.02 billion, and $4.04 billion, respectively. hedged etfs are much more expensive than all other categories of etfs and have average expense ratio of 0.84% and are more expensive than all other categories of etfs (with the exception of dbc). the range of expense ratios for hedged etfs varies from 0.24% to 1.65%. the average turnover for hedged etfs was also very high compared with most other index etfs. the average turnover for hedged etfs was 138.58%. in comparison, the turnover of s&p 500 etf (ivv) was only 5%. expenses and turnover are very important as previous literature finds that expenses and turnover are negatively related to fund performance (blake et al., 1993; carhart, 1997; dellva and olson, 1998; domian and reichenstein, 1998; dowen and mann, 2004; golec, 1997; kanuri and mcleod, 2014). table 2 shows the average, maximum and minimum net asset allocation of hedged etfs. hedged etfs are highly diversified across different asset classes (both foreign and domestic). hedged etfs have highly negative allocations indicate short selling of assets exploit arbitrage conditions. this is consistent with koski and pontiff (1999) and deli and varma (2002). table 1 shows the summary statistics for hedged etf portfolio and other index etfs for the period of our study etf category total etfs number surviving (december 2014) number dead (december 2014) aum for surviving etfs (december 2014) average expense ratio average turnover ratio hedged etfs 49 34 15 $3,415,247,306 0.84% 138.58% ivv (s&p 500) 1 1 0 $69,686,294,171 0.07% 5% agg (total bond) 1 1 0 $24,092,634,308 0.08% 180% veu (total world ex u.s.) 1 1 0 $12,272,804,784 0.14% 4% iyr (real estate) 1 1 0 $6,021,577,713 0.45% 27.00% dbc (commodities) 1 1 0 $4,036,424,446 0.88% 0% 184 s. kanuri / financial services review 25 (2016) 181–198 6.1. correlation table 3 shows the results for spearman rank correlation test between hedged etf portfolio, ivv, agg, veu, iyr, and dbc, 65% u.s. stock market (ivv)/35% bond market (agg) portfolio and 45% u.s. stock market (ivv)/10% foreign stock market (veu)/5% real estate market (veu)/5% commodity market (dbc)/35% bond market (agg) portfolio during the period of our study. results indicate that hedged etf portfolio have highly negative correlation with all other asset categories with the exception of commodities (dbc). correlation between hedged etf portfolio and ivv (s&p 500) is �0.4285, whereas correlation between hedged etf portfolio and agg (total bond market) is �0.0643. hedged etf portfolio is only positively correlated with commodity index etf. however, even correlation between hedged etf portfolio and commodity index etf is very low (0.0247). results were statistically significant in the case of ivv, veu, iyr, and 45% u.s. stock market (ivv)/10% foreign stock market (veu)/5% real estate market (veu)/5% commodity market (dbc)/35% bond market (agg) portfolio. these results indicate that hedged etfs are highly diversified and, therefore, have highly negative or very low correlation with all other asset categories. 6.2. returns and standard deviation table 4 shows average monthly returns, median monthly returns, standard deviation of monthly returns, and cumulative returns over the entire period (january 2008 through december 2014). hedged etf portfolio severely underperformed all asset categories with the exception of powershares db commodity tracking etf and has much lower average monthly returns compared with all other asset categories p value of average returns are significantly different than zero (5% or better) for all portfolios in fact, hedged etf portfolio lost value and had negative average monthly returns. the only other category that had negative average monthly returns was commodity index etf. however, hedged etf portfolio had lower standard deviation of returns compared with all other categories except the total market bond etf (agg). the cumulative returns for hedged etf portfolio for the entire period were �14.76%. only powershares db commodity tracking etf lost more value (�40.56%) over this time period. table 5a shows the annualized returns every year, average annual returns (both arithmetic and geometric) as well as standard deviation of annual returns. results again indicate that hedged etf portfolio underperformed most asset categories (with the exception of powertable 2 shows the average net allocations of hedged etfs hedged etfs asset allocations equity % (net) asset allocations bond % (net) asset allocations cash % (net) asset allocations non-u.s. bond % (net) asset allocations non-u.s. equity % (net) asset allocations other % (net) average 25.33% 9.19% 63.49% 3.95% 4.28% 1.99% max 104.25% 91.33% 192.35% 23.08% 31.00% 22.60% min �97.95% �99.19% �4.25% 0.00% �19.19% �15.77% 185s. kanuri / financial services review 25 (2016) 181–198 t ab le 3 sh ow s th e sp ea rm an r an k c or re la tio n te st be tw ee n eq ua lly w ei gh te d h ed ge d e t f po rt fo lio an d ot he r in de x e t fs fo r th e pe ri od of ou r st ud y (j an ua ry 20 08 th ro ug h d ec em be r 20 14 ) c or re la tio n h ed ge d e t fs iv v (s & p 50 0) a g g (t ot al bo nd ) v e u (t ot al w or ld e x u .s .) iy r (r ea l es ta te ) d b c (c om m od iti es ) 65 % iv v /3 5% a g g 45 % iv v /1 0% v e u /5 % iy r /5 % d b c /3 5% a g g h ed ge d e t fs 1 iv v (s & p 50 0) � 0. 18 35 * 1 a g g (t ot al bo nd ) � 0. 04 01 � 0. 00 54 1 v e u (t ot al w or ld e x u .s .) � 0. 23 54 ** 0. 87 62 ** * 0. 06 72 1 iy r (r ea l es ta te ) � 0. 23 15 ** 0. 76 63 ** * 0. 23 39 ** 0. 74 12 ** * 1 d b c (c om m od iti es ) 0. 05 92 0. 51 93 ** * � 0. 00 82 0. 58 90 ** * 0. 32 68 ** 1 65 % iv v /3 5% a g g � 0. 17 34 0. 99 14 ** * 0. 08 76 0. 89 08 ** * 0. 79 94 ** * 0. 50 69 ** * 1 45 % iv v /1 0% v e u /5 % iy r /5 % d b c /3 5% a g g � 0. 19 85 * 0. 96 64 ** * 0. 12 98 0. 93 45 ** * 0. 83 33 ** * 0. 56 00 ** * 0. 98 50 ** * 1 * si gn ifi ca nt at 10 % ; ** si gn ifi ca nt at 5% ; ** * si gn ifi ca nt at 1% . 186 s. kanuri / financial services review 25 (2016) 181–198 shares db commodity tracking etf) and had negative average annualized returns (both geometric and arithmetic). however, standard deviation of annualized returns for hedged etf portfolio was much lower than all asset categories with the exception of barclays aggregate bond market etf (agg). following woolridge (2004), i also compute the cumulative wealth index (cwi) for each category. the cwi measures the outcome of investing $1,000 in each category at the beginning of january 2008, presuming reinvestment of dividends. investors who would have invested in hedged etfs would have lost the most money compared with other categories (with the exception of commodities index etf which lost 40.56% during this time period). 6.3. risk adjusted performance a portfolio may have higher returns, but it could have achieved them by taking higher risk. therefore, we compute risk adjusted performance to compare the different portfolios. table 4 shows the average monthly returns, median monthly returns, standard deviation of monthly returns, and cumulative returns (january 2008 through december 2014) for equally weighted hedged etf portfolio and other index etfs for the period of our study january 2008 through december 2014 average monthly median returns standard deviation cumulative returns (january 2008 through december p value of average returns hedged etfs �0.12% �0.22% 3.51% �13.94% 0.008*** ivv (s&p 500) 0.70% 1.41% 4.85% 62.91% 0.011** agg (total bond) 0.38% 0.32% 1.00% 37.43% 0.002*** veu (total world ex u.s.) 0.17% 0.50% 6.35% �2.87% 0.014** iyr (real estate) 0.87% 2.02% 7.93% 57.78% 0.017** dbc (commodities) �0.42% �0.30% 6.20% �40.56% 0.013** 65% ivv/35% agg 0.59% 1.17% 3.19% 57.12% 0.007*** 45% ivv/10% veu/5% iyr/5% dbc/35% agg 0.49% 0.94% 3.35% 43.78% 0.007*** * significant at 10%; ** significant at 5%; *** significant at 1%. table 5a shows the annualized returns every year, average annual returns (both arithmetic and geometric), and standard deviation of annual returns for equally weighted hedged etf portfolio and other index etfs for the period of our study (2008–2014) annual returns hedged etfs ivv agg veu iyr dbc 65% ivv/35% agg 45% ivv/10% veu/5% iyr/5% dbc/35% agg 2008 2.68% �36.95% 5.88% �44.02% �40.02% �30.80% �23.85% �24.10% 2009 �4.69% 26.43% 5.14% 38.89% 30.14% 15.08% 19.08% 20.27% 2010 �2.77% 14.96% 6.30% 11.85% 26.36% 11.86% 12.37% 12.53% 2011 �5.27% 2.03% 7.58% �14.25% 5.63% �2.71% 4.24% 2.50% 2012 �4.20% 15.91% 4.04% 18.55% 18.36% 3.31% 11.78% 11.61% 2013 0.83% 32.31% �2.15% 14.50% 1.05% �7.57% 19.23% 13.96% 2014 �1.16% 13.62% 6.04% �4.05% 26.62% �28.18% 10.99% 7.36% arithmetic average �2.08% 9.76% 4.69% 3.07% 9.73% �5.57% 7.69% 6.30% geometric average �2.12% 7.22% 4.65% �0.42% 6.73% �7.16% 6.67% 5.32% standard deviation 3.00% 22.76% 3.20% 26.77% 24.56% 18.12% 14.83% 14.49% 187s. kanuri / financial services review 25 (2016) 181–198 t ab le 5b sh ow s th e c um ul at iv e w ea lth in de x (c w i) y ea r c w ih ed ge d e t fs c w iiv v c w ia g g c w iv e u c w iiy r c w id b c c w i65 % iv v /3 5% a g g c w i45 % iv v /1 0% v e u /5 % iy r /5 % d b c /3 5% a g g 20 08 $1 ,0 26 .7 7 $6 30 .4 5 $1 ,0 58 .8 0 $5 59 .8 0 $5 99 .8 2 $6 92 .0 0 $7 61 .4 9 $7 59 .0 4 20 09 $9 78 .5 9 $7 97 .0 7 $1 ,1 13 .2 1 $7 77 .5 3 $7 80 .6 1 $7 96 .3 5 $9 06 .8 1 $9 12 .8 7 20 10 $9 51 .4 5 $9 16 .3 2 $1 ,1 83 .3 6 $8 69 .6 5 $9 86 .3 7 $8 90 .8 2 $1 ,0 19 .0 1 $1 ,0 27 .2 7 20 11 $9 01 .3 2 $9 34 .9 1 $1 ,2 73 .0 3 $7 45 .7 5 $1 ,0 41 .9 1 $8 66 .7 1 $1 ,0 62 .2 3 $1 ,0 52 .9 8 20 12 $8 63 .5 0 $1 ,0 83 .7 0 $1 ,3 24 .4 9 $8 84 .0 6 $1 ,2 33 .1 6 $8 95 .4 1 $1 ,1 87 .3 5 $1 ,1 75 .2 1 20 13 $8 70 .7 1 $1 ,4 33 .8 1 $1 ,2 96 .0 2 $1 ,0 12 .2 4 $1 ,2 46 .0 9 $8 27 .5 8 $1 ,4 15 .6 6 $1 ,3 39 .3 1 20 14 $8 60 .5 7 $1 ,6 29 .1 3 $1 ,3 74 .2 6 $9 71 .2 7 $1 ,5 77 .7 9 $5 94 .3 6 $1 ,5 71 .2 0 $1 ,4 37 .8 4 c um ul at iv e (2 00 8– 20 14 ) � 13 .9 4% 62 .9 1% 37 .4 3% � 2. 87 % 57 .7 8% � 40 .5 6% 57 .1 2% 43 .7 8% t he c w i m ea su re s th e ou tc om e of in ve st in g $1 ,0 00 in ea ch ca te go ry at th e be gi nn in g of ja nu ar y 20 08 , pr es um in g re in ve st m en t of di vi de nd s. 188 s. kanuri / financial services review 25 (2016) 181–198 we calculate sharpe ratio (1964), sortino ratio (1991), omega ratio (2002), and kappa 3 ratio (2004) for each portfolio from january 2008 through december 2014 to compare their risk-adjusted performance (see appendix e for a brief description of these measures). results indicate that hedged etf portfolio had lower risk-adjusted performance compared to all asset categories with the exception of powershares db commodity tracking etf (table 6). 7. hedged mutual funds (hmfs) versus hedged etfs as a robustness test, hedged etfs are also compared with the same six categories (long/short, market neutral, managed futures, multi-alternative, bear market, and nontraditional bond) of hedged mutual funds (hmfs) for this time period (january 2008 through december 2014). the list of all hmfs (surviving and dead) was obtained from morningstar direct database. following bauer et al. (2005, 2006, and 2007), all funds with at least 12 months of return data were included in the analysis. this means that funds that were created after january 2014 were excluded from the analysis. following baks (2003) and wermers et al. (2012), funds with several share classes were combined and the assetweighted returns were computed. also following agrawal et al. (2004) a single portfolio of all hedged mutual funds was formed to compare with hedged etfs. there were a total of 539 hedged mutual funds with $149.77 billion in assets as of december 2014. table 7a shows the descriptive statistics for both hedged mutual funds and etfs. as expected, hedged mutual funds are much expensive (average expense ratio of 1.82%) and also have much higher turnover (327.08%). table 6 shows risk-adjusted performance measures (sharpe, sortino, kappa 3, and omega ratios) for equally weighted hedged etf portfolio and other index etfs from january 2008 through december 2014 january 2008 through december 2014 sharpe ratio sortino ratio kappa 3 ratio omega ratio hedged etfs �0.041 �0.064 �0.018 0.851 ivv (s&p 500) 0.139 0.196 0.064 1.432 agg (total bond) 0.358 0.696 0.120 2.642 veu (total world ex u.s.) 0.023 0.032 0.011 1.064 iyr (real estate) 0.107 0.148 0.057 1.373 dbc (commodities) �0.072 �0.092 �0.033 0.826 65% ivv/35% agg 0.177 0.253 0.071 1.580 45% ivv/10% veu/5% iyr/5% dbc/35% agg 0.139 0.194 0.056 1.447 table 7a shows the descriptive statistics for hedged etfs and hedged mutual funds etf category total number (surviving and dead) december 2014 aum for surviving funds (december 2014) average expense ratio average turnover ratio hedged etfs 49 $3,415,247,306 0.84% 138.58% hedged mutual funds 539 $149,772,989,903 1.82% 327.08% 189s. kanuri / financial services review 25 (2016) 181–198 t ab le 7b sh ow s th e sp ea rm an r an k c or re la tio n te st be tw ee n h ed ge d e t fs , h ed ge d m ut ua l fu nd s an d di ff er en t in de x e t fs c or re la tio n h ed ge d e t fs h ed ge d m ut ua l fu nd s iv v (s & p 50 0) a g g (t ot al bo nd ) v e u (t ot al w or ld e x u .s .) iy r (r ea l es ta te ) d b c (c om m od iti es ) 65 % iv v /3 5% a g g 45 % iv v /1 0% v e u /5 % iy r /5 % d b c /3 5% a g g h ed ge d e t fs 1 h ed ge d m ut ua l fu nd s 0. 25 72 ** 1 iv v (s & p 50 0) � 0. 18 35 * � 0. 24 71 ** 1 a g g (t ot al bo nd ) � 0. 04 01 � 0. 17 52 � 0. 00 54 1 v e u (t ot al w or ld e x u .s .) � 0. 23 54 ** � 0. 15 90 0. 87 62 ** * 0. 06 72 1 iy r (r ea l es ta te ) � 0. 23 15 ** � 0. 19 41 * 0. 76 63 ** * 0. 23 39 ** 0. 74 12 ** * 1 d b c (c om m od iti es ) 0. 05 92 0. 03 60 0. 51 93 ** * � 0. 00 82 0. 58 90 ** * 0. 32 68 ** 1 65 % iv v /3 5% a g g � 0. 17 34 � 0. 21 45 ** 0. 99 14 ** * 0. 08 76 0. 89 08 ** * 0. 79 94 ** * 0. 50 69 ** * 1 45 % iv v /1 0% v e u /5 % iy r /5 % d b c /3 5% a g g � 0. 19 85 * � 0. 18 98 * 0. 96 64 ** * 0. 12 98 0. 93 45 ** * 0. 83 33 ** * 0. 56 00 ** * 0. 98 50 ** * 1 * si gn ifi ca nt at 10 % ; ** si gn ifi ca nt at 5% ; ** * si gn ifi ca nt at 1% . 190 s. kanuri / financial services review 25 (2016) 181–198 table 7b shows the spearman rank correlation test between hedged etfs, hedged mutual funds, and different index etfs. correlation between hedged etfs and mutual funds was 0.2572. results were statistically significant at 5%. both hedged etfs and mutual funds have highly negative or low correlation with other index etfs. both hedged etfs and mutual funds had positive correlation only with commodity index fund (dbc). however, the correlation was very low even in this case. these results indicate that hedged etfs and mutual funds did help investors diversify. 7.1. risk, returns, and risk-adjusted performance table 8a shows the average monthly returns, median monthly returns, and standard deviation for hedged etfs and mutual funds. the average monthly returns were very similar for hedged etfs and hedged mutual funds (�0.12% and �0.10%, respectively). results were statistically significant at 1% in both the cases. however, hedged etfs had much higher standard deviation (of monthly and annual returns) or risk compared with hedged mutual funds. table 8b shows the annualized returns and cumulative wealth index (cwi) from 2008 to 2014. both hedged etfs and hedged mutual funds lost value, and had negative cumulative returns of �13.94% and �8.34%, respectively. both categories also had negative risk-adjusted performance. table 8a shows the average monthly returns, median monthly returns, and standard deviation of monthly returns for hedged etfs and hedged mutual funds january 2008 through december 2014 average returns median returns standard deviation cumulative returns (january 2008 through december 2014) p value of average returns hedged etfs �0.12% �0.22% 3.51% �13.94% 0.008*** hedged mutual funds �0.10% �0.14% 0.58% �8.34% 0.001*** * significant at 10%; ** significant at 5%; *** significant at 1%. table 8b shows the annualized returns (2008–2014) and cumulative wealth index (cwi) for hedged etfs and hedged mutual funds annual returns hedged etfs hedged mutual funds cwi (hedged etfs) cwi (hedged mutual funds) 2008 2.68% �0.34% $1,026.77 $996.64 2009 �4.69% 0.26% $978.59 $999.27 2010 �2.77% �0.46% $951.45 $994.72 2011 �5.27% �2.01% $901.32 $974.69 2012 �4.20% �2.29% $863.50 $952.35 2013 0.83% �3.33% $870.71 $920.62 2014 �1.16% �0.44% $860.57 $916.56 arithmetic average �2.08% �1.23% geometric average �2.12% �1.24% standard deviation 3.00% 1.32% cumulative returns �13.94% �8.34% the cwi measures the outcome of investing $1,000 in each category at the beginning of january 2008, presuming reinvestment of dividends. 191s. kanuri / financial services review 25 (2016) 181–198 8. conclusions hedged etfs help retail or individual investors invest in etfs that follow strategies similar to hedge funds at a fraction of the cost. they are also very liquid and do not have high initial investment or longer lock-up periods that hedge funds have. in this paper i analyze the performance of six different categories of hedged etfs (long/short, market neutral, managed futures, multi-alternative, bear market, and non-traditional bond) for the time period january 2008 through december 2014. the performance of these hedged etfs was compared to five categories of index fund etfs. our results indicate that hedged etfs have not delivered on their promise of absolute returns that is, positive returns regardless of market conditions. hedged etfs have very high expense ratios and turnover compared with other index etf categories. hedged etfs did help investors diversify and had highly negative or low correlation with all other asset categories. however, this did not translate into superior absolute or risk-adjusted performance, and hedged etfs underperformed all other categories (with the exception of commodities etf dbc). as a robustness test the performance of hedged etfs was also compared with the same six categories of hedged mutual funds during this time period (table 9). there were 539 hedged mutual funds (surviving and dead) at the end of december 2014 with $149 billion in assets. both hedged etfs and mutual funds have highly negative or low correlation with other asset categories. these results again indicate that hedged etfs and mutual funds did help investors diversify during the period of our study. however, this did not help investors as both the categories had negative absoluteand risk-adjusted performance. based on our findings, investors would have been better off with index funds. notes 1 http://www.etf.com/etf-education-center/21043-article-46-alternatives-etfs-can-anetf-replicate-a-hedge-fund.html. acknowledgments i would like to express my sincere gratitude to the editor and two anonymous referees for their helpful suggestions that greatly improved the article. all remaining errors and omissions are my responsibility alone. table 9 shows the risk-adjusted performance (sharpe, sortino, kappa 3, and omega ratios) for hedged etfs and hedged mutual funds january 2008 through december 2014 sharpe ratio sortino ratio kappa 3 ratio omega ratio hedged etfs �0.041 �0.064 �0.018 0.851 hedged mutual funds �0.217 �0.262 �0.044 0.571 192 s. kanuri / financial services review 25 (2016) 181–198 appendix a etf categories (source: morningstar) bear market: bear-market portfolios invest in short positions and derivatives to profit from stocks that drop in price. because these portfolios often have extensive holdings in shorts or puts, their returns generally move in the opposite direction of the benchmark index. long-short: long-short portfolios hold sizable stakes in both long and short positions. some funds that fall into this category are market neutral–dividing their exposure equally between long and short positions in an attempt to earn a modest return that is not tied to the market’s fortunes. other portfolios that are not market neutral will shift their exposure to long and short positions depending upon their macro outlook or the opportunities they uncover through bottom-up research. managed futures: these etfs typically take long and short positions in futures or other derivative contracts according to a trend-following or momentum strategy. market neutral: these etfs try to earn income by maintaining low correlation with the market. these funds usually have 50% of net assets in long positions while holding 50% of net assets in short positions. their goal is to deliver positive returns regardless of fluctuations in market. multi-alternative: these etfs offer investors exposure to a combination of strategies like long-short equity and debt, managed futures, global macro, and convertible arbitrage, among others. these strategies may change in response to market conditions. non-traditional bonds: many etfs in this group describe themselves as “absolute return” portfolios, which seek to avoid losses and produce returns uncorrelated with the overall bond market; they use a variety of methods to achieve those aims. appendix b shows the different replication strategies for hedged etfs 1. direct approach: the easiest way for hedged etfs to replicate hedge fund returns would be to hold the hedge fund themselves. however, hedge funds have much longer lock-up periods and etfs require their assets to be traded daily. therefore, hedged etfs can directly follow hedge funds strategies wherever strategy permits. for example, hedged etfs can buy underpriced securities and short overpriced ones. these etfs can also use leverage, derivatives, options, and swaps (like hedge funds) to seek higher returns. 2. hedge fund replication: lot of these hedged etfs seeks to track a custom index that, in turn, seeks to track the risk-adjusted return characteristics of hedge funds. this process is called hedge fund replication. hedge funds report their returns to a hedge fund indexing firm. the hedge fund indexing firm then creates an index that replicates the returns of the hedge funds either broadly or by a specific strategy. these etfs try to replicate the returns of hedge funds by buying actual assets that hedge funds hold using regressions and other statistical process. (however, according to morningstar direct some of these hedged etfs do not have a benchmark.) 193s. kanuri / financial services review 25 (2016) 181–198 3. copycat: hedge funds by law are required to disclose their holdings. hedged etfs rely on 13f filings reported by the hedge funds. however, hedge funds are secretive and don’t publish their holding unlike etfs. however, the law requires hedge funds to disclose their data on a quarterly, lagged basis. most hedge funds could have already sold their stock by the time the filings are public. appendix c: shows the list of hedged etfs, the category to which it belongs and their inception date number etf ticker category inception 1 rydex inverse 2� s&p midcap 400* rms bear 11/5/2007 2 rydex inverse 2� russell 2000* rry bear 11/5/2007 3 rydex inverse 2� s&p select sector engy* rec bear 6/10/2008 4 rydex inverse 2� s&p select sector fincl* rfn bear 6/10/2008 5 rydex inverse 2� s&p select sector hlth* rho bear 6/10/2008 6 rydex inverse 2� s&p select sector tech* rtw bear 6/10/2008 7 macroshares major metro housing down* dmm bear 6/30/2009 8 direxion daily 2 yr trsy bear 3� shares* twoz bear 2/25/2010 9 advisorshares ranger equity bear etf hdge bear 1/26/2011 10 advisorshares athena intl bear etf* hdgi bear 7/18/2013 11 powershares nasdaq-100 buywrite* pqbw long short 6/12/2008 12 advisorshares accuvest glbl lg sht etf agls long short 7/8/2010 13 advisorshares qam equity hedge etf qeh long short 8/7/2012 14 first trust cboe s&p 500 vixtail h etf vixh long short 8/29/2012 15 powershares s&p 500 downside hedged etf phdg long short 12/5/2012 16 us equity high volatility put write etf hvpw long short 2/27/2013 17 janus velocity tail risk hdgd lg cp etf trsk long short 6/20/2013 18 janus velocity volatility hdgd lg cp etf spxh long short 6/20/2013 19 wisdomtree managed futures strategy etf wdti managed futures 1/5/2011 20 first trust morningstar mgd futsstrt etf fmf managed futures 8/1/2013 21 ishares diversified alternatives trust* alt market neutral 10/6/2009 22 iq merger arbitrage etf mna market neutral 11/17/2009 23 proshares rafi long/short rals market neutral 12/2/2010 24 quantshares u.s. mkt neut anti-momen etf* nomo market neutral 9/7/2011 25 quantshares u.s. market neut quality etf* qlt market neutral 9/7/2011 26 quantshares u.s. market neutral size siz market neutral 9/7/2011 27 quantshares u.s. market neut momentum mom market neutral 9/7/2011 28 quantshares u.s. mrkt neut high beta etf* btal market neutral 9/13/2011 29 quantshares u.s. market neut anti-beta btal market neutral 9/13/2011 30 quantshares u.s. market neutral value chep market neutral 9/13/2011 31 advisorshares rockledge sectorsam etf* ssam market neutral 1/11/2012 32 iq hedge market neutral tracker etf qmn market neutral 10/3/2012 33 proshares merger mrgr market neutral 12/11/2012 34 iq hedge multi-strategy tracker etf qai multialternative 3/25/2009 35 proshares hedge replication hdg multialternative 7/12/2011 36 first trust tactical high yield etf hyls non traditional bond 2/25/2013 37 market vectors trs-hdgd hi-yld bd etf thhy non traditional bond 3/21/2013 38 proshares high yield-interest rate hdgd hyhg non traditional bond 5/21/2013 39 proshares investment grade-intr rt hdgd ighg non traditional bond 11/5/2013 40 wisdomtree bofa mrl lynch hybd ngtdr etf hynd non traditional bond 12/18/2013 41 wisdomtree bofa mrl lynch hybd zrdr etf hyzd non traditional bond 12/18/2013 42 wisdomtree barclays us aggtbd ngtdur etf agnd non traditional bond 12/18/2013 43 wisdomtree barclays us aggtbd zr dur etf agzd non traditional bond 12/18/2013 44 wisdomtree japan interest rate strat etf jgbb non traditional bond 12/18/2013 45 iq hedge macro tracker etf mcro other allocation 6/9/2009 46 iq real return etf cpi other moderate allocation 10/27/2009 47 wisdomtree global real return etf rrf other global fixed income 7/14/2011 48 spdr ssga multi-asset real return etf rly other allocation 4/25/2012 49 alphaclone alternative alpha etf alfa other u.s. equity large cap growth 5/31/2012 etfs marked with a * in front of them have been liquidated, merged, or closed as of december 2014. 194 s. kanuri / financial services review 25 (2016) 181–198 a pp en di x d : sh ow s th e di ff er en t in de x e t fs us ed fo r co m pa ri so n n um be r e t f t ic ke r c at eg or y in ce pt io n b en ch m ar k 1 is ha re s c or e s& p 50 0 iv v u .s . e t f l ar ge b le nd 5/ 15 /2 00 0 s& p 50 0 2 is ha re s c or e u .s . a gg re ga te b on d a g g u .s . e t f in te rm ed ia te -t er m b on d 9/ 22 /2 00 3 b ar cl ay s u .s . a gg b on d t r u sd 3 v an gu ar d ft se a llw or ld ex -u .s . e t f v e u u .s . e t f fo re ig n l ar ge b le nd 3/ 2/ 20 07 ft se a w e x u .s . t r u sd 4 po w er sh ar es d b c om m od ity t ra ck in g e t f d b c u .s . e t f c om m od iti es b ro ad b as ke t 2/ 3/ 20 06 po w er sh ar es d b c om m od ity in de x 5 is ha re s u .s . r ea l e st at e iy r u .s . e t f r ea l e st at e 6/ 12 /2 00 0 d ow jo ne s u .s . r ea l e st at e in de x 195s. kanuri / financial services review 25 (2016) 181–198 appendix e shows the different risk-adjusted measures of performance sharpe ratio: the sharpe ratio (sharpe, 1964) evaluates how well an etf compensates its investor for each unit of risk they incur. the higher the sharpe ratio, the better is the performance of the etf. sharpe ratio � �rp � rf� rf where rp denotes the monthly returns on the portfolio. rf is the monthly risk free rate. �p is the standard deviation of portfolio’s excess returns. sortino ratio: the sortino ratio (sortino and van der meer, 1991) differentiates between good and bad volatility in the sharpe ratio. the differentiation of upward and downward volatility allows the calculation of the risk-adjusted return to provide a performance measure of an investment without penalizing it for positive returns. similar to the sharpe ratio, the higher the sortino ratio, the better is the performance of a portfolio. the sortino ratio is shown as follows: sortino ratio � �rp � rf� �df where rp and rf are described as above and �d is the standard deviation of portfolio’s negative returns. omega ratio: introduced by keating and shadwick (2002), it is a way of measuring the performance of financial assets based on the level of returns they offer in return for the risk of investing in them. it is a ratio of weighted gains to weighted losses. the measure divides expected returns into two parts – gains and losses, or returns above the expected rate (the upside) and those below it (the downside). therefore, in simple terms, consider omega as the ratio of upside returns (good) relative to downside returns (bad). while the sharpe ratio covers only the first two moments of return distribution (means and variance), omega ratio covers all moments of return distribution or the omega ratio is an alternative measure of asset performance that gives the investor the information the sharpe ratio discards. � � � r b �1 � f� x��dx � a r f� x�dx where f(x) is the cumulative probability distribution (i.e., the probability that a return will be less than x), r is a threshold value selected by the investor and a,b are the investment 196 s. kanuri / financial services review 25 (2016) 181–198 intervals. it is effectively equal to the probability weighted gains divided by the probability weighted losses after a threshold. kappa 3 ratio: motivated to find a more generalized risk-adjusted performance measure, kaplan and knowles (2004) developed the kappa-measure. they show that the omega and the sortino ratios are only special cases of kappa, whereby the parameter n of kappa determines whether the sortino ratio, omega, or another risk-adjusted return measure is generated. the general form of kappa is described by the expression below. kn��� � � � � n�lpmn��� choosing n � 1 and n � 2 yields the omega (�k1) and the sortino ratio (�k2), respectively. in general, any number is possible for the parameter n. kappa 3 (�k3) however, seems to be the most frequently used version of the kappa measure in literature (e.g., eling and schumacher (2007), kaplan and knowles (2004)). thus, k3 shall also be applied in the empirical examination of performance measures in this composition. references agarwal, v., boyson, n. m., & naik, n. y. (2009). hedge funds for retail investors? an examination of hedged mutual funds. journal of financial and quantitative analysis, 44, 273. baks, k. p. (2003). on the performance of mutual fund managers. working paper, emory university. bauer, r., koedijk, k., & otten, r. (2005). international evidence on ethical mutual fund performance and investment style. journal of banking & finance, 29, 1751–1767. bauer, r., otten, r., & rad, a. (2006). new zealand mutual funds: measuring performance and persistence in performance. accounting & finance, 46, 347–363. bauer, r., derwall, j., & otten, r. (2007). the ethical mutual fund performance debate: new evidence from canada. journal of business ethics, 70, 111–124. blake, c. r., elton, e. j., & gruber, m. j. (1993). the performance of bond mutual funds. journal of business, 66, 371–403. brown, j.r., liang, n., & weisbenner, s. (2007). individual account investment options and portfolio choice. behavioral lessons from 401(k) plans. journal of public economics, 91(10), 1992–2013. carhart, m. m. (1997). on persistence in mutual fund performance. the journal of finance, 52, 57–82. davidoff, t., brown, j. r., & diamond, p. a. (2003). annuities and individual welfare (no. w9714). national bureau of economic research. deli, d. n., & varma, r. (2002). contracting in the investment management industry: evidence from mutual funds. journal of financial economics, 63, 79–98. dellva, w. l., & olson, g. t. (1998). the relationship between mutual fund fees and expenses and their effects on performance. financial review, 33, 85–104. domian, d. l., & reichenstein, w. (1997). performance and persistence in money market fund returns. financial services review, 6, 169–183. dowen, r. j., & mann, t. (2004). mutual fund performance, management behavior, and investor costs. financial services review, 13, 79–91. eling, m., & schuhmacher, f. (2007). does the choice of performance measure influence the evaluation of hedge funds? journal of banking & finance, 31, 2632–2647. elton, e. j., gruber, m. j., & blake, c. r. (1996). survivor bias and mutual fund performance. review of financial studies, 9, 1097–1120. 197s. kanuri / financial services review 25 (2016) 181–198 golec, j. h. (1996). the effects of mutual fund managers’ characteristics on their portfolio performance, risk and fees. financial services review, 5, 133–147. kanuri, s., & mcleod, r. w. (2014). performance of alternative mutual funds: the average investor’s hedge fund. financial services review, 23, 93–121. kaplan, p. d., & knowles, j. a. (2004). kappa: a generalized downside risk-adjusted performance measure. journal of performance measurement, 8, 42–54. keating, c., & shadwick, w. f. (2002). a universal performance measure. journal of performance measurement, 6, 59–84. koski, j. l., & pontiff, j. (1999). how are derivatives used? evidence from the mutual fund industry. the journal of finance, 54, 791–816. sharpe, w. f. (1964). capital asset prices: a theory of market equilibrium under conditions of risk*. the journal of finance, 19, 425–442. sortino, f. a., & van der meer, r. (1991). downside risk. the journal of portfolio management, 17, 27–31. stout, r., & mitchell, j. b. (2006). dynamic retirement withdrawal planning. financial services review, 15, 117–131. wermers, r., yao, t., & zhao, j. (2012). forecasting stock returns through an efficient aggregation of mutual fund holdings. review of financial studies, 25, 3490–3529. woolridge, r. j. (2004). performance of stocks recommended by brokerages. the journal of investing, 13, 23–34. 198 s. kanuri / financial services review 25 (2016) 181–198 home ownership decision in personal finance: some empirical evidence c. sherman cheunga, and peter miua,* adegroote school of business, mcmaster university, hamilton, ontario l8s 4m4, canada abstract despite being one of the most important decisions a household has to make and extensively covered in personal finance textbooks, there is very little empirical guidance as to whether it pays to own a house. we examine this empirical question for households with different risk tolerance. by including home ownership into the general portfolio analysis of financial assets, we demonstrate clearly the interaction effect between financial assets and home ownership. we also offer a comprehensive analysis of 20 regional housing markets to determine whether the economic case for home ownership varies across regions. for households that decide to rent instead of owning a house, this study offers evidence on the effectiveness of hedging housing consumption risk with investments in real estate investment trusts. © 2015 academy of financial services. all rights reserved. jel classification: g10; g11 keywords: home ownership; portfolio diversification 1. introduction the home ownership decision is a major decision for most households. all textbooks in personal finance cover the topic in details (e.g., see madura, 2014). almost all textbooks examine the home ownership decision as a “buy versus rent” decision. what is lacking in personal finance textbooks is the empirical evidence on whether a typical u.s. household should in fact own a house. this absence of empirical evidence is somewhat unusual given the emphasis on empirical evidence in modern finance. theory is not enough unless it is * corresponding author. tel.: �1-905-525-9140 ext. 23981; fax: �1-905-521-8995. e-mail address: miupete@mcmaster.ca (p. miu) financial services review 24 (2015) 51–76 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. supported by empirical evidence. in the literature, there is not much empirical evidence that buying a house is in fact a wise decision. this lack of evidence is a cause for concern since a house is a major asset for most households and the decision is relatively expensive and inconvenient to be reversed. the question of whether a household should purchase a house takes on more urgency since the collapse of the housing market in 2008. our article examines empirically whether it pays to own a house by estimating the diversification benefit of home ownership in a representative investor’s portfolio. similar to holding any financial assets, home ownership may enhance the wealth of a portfolio investor in the form of capital gain, whereas at the same time help diversifying the risk of the overall investment portfolio. if home ownership offers statistically significant enhancement in the risk-return trade off, then it pays to own a house. if a house does not offer such benefit, the case for home ownership is no longer obvious. perhaps the chapter on “buying versus renting” a house in personal finance textbooks should be approached differently. we advocate an approach where the decisions of home ownership, savings, and investment are made jointly. therefore, this study broadens the scope of services offered by financial advisors and planners. the common practices for most financial advisors and planners tend to focus exclusively on investments, retirement planning, estate planning, and insurance needs. residential ownership, which may represent a great part of a household’s total wealth, tends to be ignored. this study equips financial advisors or planners to offer a more comprehensive guidance to their clients beyond savings and investment decisions. the finance literature does address the home ownership decision with the emphasis on the interactions of housing choices with other financial assets. some researchers examine the ex ante optimality of joint portfolio choices of liquid assets such as stocks, bonds, and the house. yao and zhang (2005), for example, derive a stylized model and provide empirical evidence for the ability of the explanatory variables in their model to predict individuals’ decisions to own versus to rent a house. whenever the investor in their model chooses to own a house, the model predicts the investor will reduce the equity proportion in their net worth, reflecting the substitution effect of the house for risky stocks. flavin and yamashita (2002) examine the effects of an exogenous housing ownership on the financial portfolio over a life cycle. unfortunately, in their model, a house is not endogenously included in a household’s portfolio based on its risk-return characteristics and, therefore, it cannot address the issue of whether an individual should own a house based on the risk-return characteristics of housing ownership. hennessey (2003) investigates the home ownership choice using two case studies. his prime focus is the payoff from the ownership without explicitly taking into consideration the risk dimension. given the importance of risk consideration in modern finance, any analysis without accounting for risk will be incomplete at best. with the exception of goetzmann (1993) and wu and pandey (2012), the existing literature seems to be mostly concerned with the ex ante modeling of the housing choice and the empirical verification of the proposed models. different from the existing literature, our article focuses on the ex-post diversification benefit of owning a house assuming very little about the model governing the behavior of our representative investor other than the standard mean-variance framework. the basic question we address in this study is whether an investor should own a house instead of renting one based on the historical risk and return characteristics of home ownership. 52 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 to capture the historical risk and return characteristics of owning a house, we first use the aggregate u.s. residential property market price index as a proxy for the return on owning a house. we also examine twenty regional residential property markets. home ownership differs from the ownership of financial assets in one very significant way. in the case of stocks, one can obtain exposure to the whole market easily by buying an exchange-traded fund (etf) that covers the overall united states or international markets. common market indices such as the standard and poor’s 500 index are the appropriate proxy for the ownership of stocks. in the case of houses, a home owner usually owns only one house. an aggregate national market index for the overall u.s. housing market becomes meaningless if there are significant variations in regional housing markets. this necessitates the examination of individual regional markets. by examining individual regional markets, we can also shed light on whether the housing markets in the u.s. sun belt are indeed more attractive than the housing markets in the northeast united states? further, the difference in the attractiveness of various regional housing markets is an important issue for the “buy versus rent” decision because of the fact that individuals quite often have to relocate to a different city for career or other reasons. for individuals who choose rental instead of outright ownership, the concern for them is the consumption risk associated with rental because of future rent increases. can these individuals replace the investments in houses with investments in real estate investment trusts (reits)? can they hedge their housing consumption risk by owning reits? is the ownership of reits an alternative way to obtain exposure to the residential housing markets? we attempt to address this issue by verifying the attractiveness of reits in lieu of the outright ownership of residential housing. goetzmann (1993) and wu and pandey (2012) examine the optimal allocation to a residential property using the historical risk and return characteristics of home ownership similar to part of our study. as explained later, they, however, fail to provide the statistical tests that can tell us whether a house belongs to an investor’s portfolio in the first place. they implicitly assume an individual should always own a house and proceeds with estimating what fraction of his or her wealth should be allocated to the house investment. in this study, we provide the necessary statistical tests on the home ownership decision. it is well known that any in-sample portfolio performance analysis ought to result in a risk-adjusted return on the optimal portfolio no worse than that before an extra asset class is added to the portfolio. the issue is whether such improvement is too large to be explained by chance and this can only be settled by statistical tests. goetzmann (1993) and wu and pandey (2012) also fail to examine alternatives to home ownership such as investments in reits. some may question the appropriateness of our choice of a portfolio optimization model involving financial assets and a house. individuals who contemplate home ownership versus rental are likely to be first-time buyers who lack the financial ability to own financial assets. should a portfolio approach involving both financial assets and a house be appropriate when the first-time buyer may have no extra money for financial assets? first, individuals may have financial assets in their pension accounts in which they have entitlement but without accessibility for tax and other institutional reasons. the 2001 survey of consumer finances (scf) shows that about two-thirds of the u.s. households own their primary residences and 53c.s. cheung, p. miu / financial services review 24 (2015) 51–76 the home value accounts for only 55% of a homeowner’s total assets, on average. secondly, home ownership is a long-term decision. although first-time buyers may not have immediate ownership in financial assets, they will eventually accumulate financial assets through their 401(k). our approach can therefore be thought of as a long-term validation of a current decision. finally, the use of a portfolio model is a common approach adopted in the literature (e.g., brueckner, 1997; flavin and yamashita, 2002; goetzmann, 1993). by adopting an approach that is consistent with those commonly used, we ensure our results are readily comparable with those of other studies. it is our focus on the interaction effect between real estate and financial assets that enables us to gain important insights into the impact of home ownership on the investments in other financial assets. our findings suggest that owning a house quite often results in the displacement of bond investments in the portfolio as the two assets exhibit very similar risk-return characteristics. thus, investors and financial planners are likely to arrive at suboptimal overall investment portfolios if they focus on asset allocation involving only traditional financial assets while ignoring the substitution effect between real estate ownership and bond investments. in summary, this study addresses a gap in the literature and is far more comprehensive than the previous studies in that we examine the suitability of home ownership for a variety of potential home owners with different degrees of risk aversion.1 the optimal investment choice of an investor of high risk tolerance could be dramatically different from that of one who hardly wants to bear any risk given the difference in how they trade-off return versus risk. potentially as a key component of the overall investment portfolio, home ownership may be perceived differently by the two kinds of investors in terms of its contribution to the growth of their overall wealth. for example, no matter how much diversification benefit it brings about in constraining the risk of the overall portfolio, home ownership may still be considered as a poor decision by an investor of high risk tolerance if it fails to deliver a sufficiently high return. on the other hand, a modest expected return from the appreciation of real estate price may provide sufficient incentive for a risk averse investor to decide owning the same residential property since the diversification benefit of having the real estate investment as part of his or her overall investment portfolio is highly treasured by this kind of investors. we design our empirical analysis in a way that allows us to dissect the implication of risk aversion on home ownership. we also examine the risk-return characteristics of different regional residential housing markets and how they may influence the home ownership decision. moreover, we examine the case for investing in reits as an alternative to outright home ownership.2 finally, we provide the necessary statistical test results in substantiating the arguments for or against home ownership. by using a comprehensive approach encompassing both home ownership and financial assets, we illustrate how and to what extent the current practice of financial planning involving financial assets alone can be very problematic. our results indicate home ownership pays for the conservative investors not because of the price appreciation of residential properties. despite the potentially false impression created by the recent housing crisis, a house is attractive for its stabilizing influences in times of financial turbulences rather than its potential for price appreciation. there are regional variations in the attractiveness of home ownership with the sun belt lagging far behind the west coast. 54 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 2. optimal portfolio of a representative u.s. homeowner in this article, we consider a representative u.s. investor originally holding u.s. equities, non-north american equities, and u.s. government bonds. for this investor, we examine the effects on the risk-return tradeoff of adding real estate to the existing portfolio. specifically, we gauge the value-adding of real estate investment by examining the extent to which it enhances the risk-adjusted expected return of the overall investment portfolio as measured by the resulting sharpe ratio of the optimal portfolio under a mean-variance optimization framework. therefore, it captures the diversification benefit given the fact that real estate returns are far from perfectly correlated with those of other financial assets. we compare the sharpe ratios of two optimal portfolios: (1) the maximum sharpe ratio of the original portfolio with the three existing financial assets as our baseline, and (2) the maximum sharpe ratio when real estate is added to the investor’s baseline portfolio. any statistically significant improvement in the sharpe ratio will prove that home ownership indeed offers a diversification benefit to the investor and enhances his or her risk-adjusted return. if short selling is allowed, the maximum sharpe ratio can be derived by first finding out the weights w of the tangency portfolio.3 w � ��1� z̄ � rf � 1� b � a � rf , (1) where z̄ is the vector of the mean returns of assets � is the variance-covariance matrix of asset returns 1 is the unit vector rf is the constant risk-free rate a�1' � ��1 � 1 b�1' � ��1 � z̄ the sharpe ratio of the tangency portfolio is then equal to: �p � z̄p � rf �p � z̄' � w � rf �w' � � � w�0.5 , (2) where z̄p and �p are, respectively, the mean and sd of the return of the tangency portfolio. because an individual cannot short sell a house the way a stock is sold short and most investors do not short sell individual asset classes, we consider the portfolio allocations when short selling is disallowed throughout this article. when short selling is disallowed, the weights of the tangency portfolio can be solved with the extra inequality constraint that the vector of weights is non-negative. the question of interest is whether the sharpe ratio of the tangency portfolio when a house is added is indeed statistically significantly higher than the sharpe ratio of that without the house. this will give us confidence that the improvement in the sharpe ratio is not because of sampling errors. the details of the test procedure is developed by glen and jorion (1993) and explained in appendix b. one simple way to test the diversification benefits is to estimate and compare the sharpe ratios of the tangency portfolios with and without real estate using historical average return 55c.s. cheung, p. miu / financial services review 24 (2015) 51–76 of t-bill for rf in eq. (2). this ex post approach is the most common practice and is also used here. in addition to using historical t-bill returns, we also consider the resulting sharpe ratios at different values of rf. by varying rf, we in effect trace out the efficient frontier by solving for different tangency portfolios.4 this approach allows us to solve for different optimal portfolios on the frontier and evaluate the statistical significance of any difference in the respective sharpe ratios for investors with different risk aversion parameters. by allowing for different degrees of risk aversion, we can detect if the diversification benefits are different for investors with different risk preference. goetzmann (1993) and wu and pandey (2012) develop the efficient frontier of the financial assets and the house using historical data. what they fail to do is to examine the baseline case without the house so as to illustrate the enhancement of the risk-adjusted return as a result of investing in the residential property. by ignoring the baseline case without the house, they implicitly assume that the house has to be in the optimal portfolio and thus ruling out the possibility that home ownership may not be a value-adding proposition for (at least) some of the investors in the first place. in this study, we proxy the return on home ownership with the price appreciation measured by the standard and poor’s (s&p)/case-shiller home price indices. it is obvious that the return on home ownership is more than the price appreciation of the residential property. a home is both a consumption good and an investment. the consumption component is the shelter provided by the house. by not accounting for the consumption component in our study, we clearly underestimate the return on home ownership. there are two other potential tax benefits associated with home ownership we ignore in this study. first, interest payments embedded in the mortgage are tax deductible, whereas rents in general are not. second, the realized gain upon selling the house also receives more favorable tax treatment than capital gains from financial asset investments. fortunately, ignoring these advantages of home ownership will only reinforce our conclusions if our test results favor home ownership based on the price appreciation alone. 3. data the proxies for united states and non-north american equities are the monthly total return series of the center for research in security prices’ (crsp’s) value-weighted index and the msci europe, australasia and far east (eafe) index, respectively. we use the total return series of the intermediate government bonds provided by the ibbotson sbbi yearbook to proxy for the monthly returns on u.s. government bonds. several indices are used as proxies for the price appreciation of residential real estate. the s&p/case-shiller home price indices are used to construct our monthly returns on home ownership as they are the most popular and comprehensive measures of u.s. residential real estate prices. we consider both the case-shiller 10-city composite (cs composite 10) and the 20-city composite (cs composite 20) indices as our proxies for the returns on the national market. the s&p/caseshiller index family also includes 20 regional indices for the 20 metropolitan statistical areas (msas). in this study, we also consider the investment in u.s. real estate investment trusts (reits) as an alternative to home ownership. to proxy for the return on reits, we use the 56 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 popular all reits monthly return series published by the national association of real estate investment trusts (nareit). because the s&p/case-shiller indices are only available from january 1987, we consider the sample period from january 1987 to december 2011 in conducting our main empirical analysis. however, data for five of the 20 msas are not available until later in our sample period. specifically, data for the phoenix msa are available from january 1989. data for the seattle msa are available from january 1990. data for both atlanta and detroit msas are only available from january 1991. finally, data for the dallas msa are from january 2000. as a result of the late availability of the dallas msa data, the cs composite 20 index, which includes the dallas msa, also starts from january 2000. although we could have conducted all our analyses using only data since january 2000 for all our time series to maintain uniformity in the time coverage across all msas, we also want to have the longest time series possible to obtain a long-term perspective of home ownership as home ownership tends to be a long-term proposition for most individuals. very few individuals own a house for just a year or two and then switch to renting. as a result, we perform our portfolio optimization exercise for the cs composite 10 index and the indices of 15 msas together with matching financial asset returns data series covering the full sample period since january 1987. for the remaining five msas and the cs composite 20 index of late data availability, we perform our analyses using shorter time series with starting dates dictated by the respective data availability dates mentioned earlier. in conducting the portfolio optimization; therefore, we consider the respective matching subsamples of the returns of other financial assets. 4. results 4.1. descriptive statistics the risk and return characteristics of the financial assets and the house price indices are reported in table 1 panel a. among all the house price and financial asset indices, u.s. equities and reits have the highest average returns of 0.81% and 0.82% per months, respectively. proxied by the national house price indices of cs composite 10 or cs composite 20, the average monthly return on home ownership is lower than those of all financial assets. international equities represented by eafe and reits are the most risky asset classes based on the sds of their monthly returns, whereas residential real estate represented by the cs composite 10 and cs composite 20 indices are the least risky. bonds appear to be moderate in both risk and returns. evidence based on the cs composite 10 and cs composite 20 indices in table 1 appears to indicate that a house should be considered as a consumption good to be enjoyed rather than an investment for future returns. if home ownership is an investment at all, it is the safest investment among the assets under consideration here. as for individual msas, portland stands out as the city with the best mean return at 0.40% per month, which is about one half of the mean return of u.s. equities. detroit, on the other hand, has the worst mean return at 0.08% per month. not one single city reports a negative mean return over the sample period. in other words, the worst scenario for owning a 57c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 1 su m m ar y st at is tic s pa ne l a : r is kre tu rn ch ar ac te ri st ic pa ne l b : c or re la tio n co ef fic ie nt n m ea n sd c as esh ill er c om po si te -1 0 c as esh ill er c om po si te -2 0 c r sp v w in te r. g ov . b on d m sc ie a fe c s c om po si te -1 0 29 9 0. 29 % 0. 91 % c s c om po si te -2 0 14 3 0. 22 % 1. 12 % c s c om po si te -2 0 1. 00 c r sp v w 29 9 0. 81 % 4. 62 % c r sp v w 0. 04 0. 09 in te r. g ov . b on d 29 9 0. 56 % 1. 33 % in te r. g ov . b on d � 0. 04 � 0. 05 � 0. 08 m sc ie a fe 29 9 0. 54 % 5. 16 % m sc ie a fe 0. 09 0. 15 0. 72 � 0. 09 a ll r e it 29 9 0. 82 % 5. 21 % a ll r e it 0. 12 0. 22 0. 60 � 0. 02 0. 48 a z -p ho en ix 27 5 0. 16 % 1. 35 % a z -p ho en ix 0. 77 0. 84 c a -l os a ng el es a 29 9 0. 34 % 1. 27 % c a -l os a ng el es a 0. 92 0. 93 c a -s an d ie go a 29 9 0. 35 % 1. 21 % c a -s an d ie go a 0. 87 0. 89 c a -s an fr an ci sc oa 29 9 0. 35 % 1. 41 % c a -s an fr an ci sc oa 0. 86 0. 87 c o -d en ve ra 29 9 0. 30 % 0. 74 % c o -d en ve ra 0. 52 0. 69 d c -w as hi ng to na 29 9 0. 35 % 1. 02 % d c -w as hi ng to na 0. 92 0. 95 fl -m ia m ia 29 9 0. 24 % 1. 15 % fl -m ia m ia 0. 81 0. 88 fl -t am pa 29 9 0. 17 % 1. 01 % fl -t am pa 0. 79 0. 89 g a -a tla nt a 25 1 0. 09 % 0. 96 % g a -a tla nt a 0. 65 0. 72 il -c hi ca go a 29 9 0. 25 % 1. 03 % il -c hi ca go a 0. 74 0. 85 m a -b os to na 29 9 0. 26 % 0. 92 % m a -b os to na 0. 76 0. 79 m id et ro it 25 1 0. 08 % 1. 20 % m id et ro it 0. 66 0. 75 m n -m in ne ap ol is 27 5 0. 22 % 1. 23 % m n -m in ne ap ol is 0. 74 0. 83 n c -c ha rl ot te 29 9 0. 18 % 0. 60 % n c -c ha rl ot t 0. 54 0. 62 n v -l as v eg as a 29 9 0. 11 % 1. 39 % n v -l as v eg as a 0. 68 0. 80 n y -n ew y or ka 29 9 0. 26 % 0. 81 % n y -n ew y or ka 0. 86 0. 89 o h -c le ve la nd 29 9 0. 21 % 0. 88 % o h -c le ve la nd 0. 55 0. 62 o r -p or tla nd 29 9 0. 40 % 0. 87 % o r -p or tla nd 0. 55 0. 80 t x -d al la s 14 3 0. 09 % 0. 85 % t x -d al la s 0. 55 0. 60 w a -s ea ttl e 26 3 0. 31 % 0. 99 % w a -s ea ttl e 0. 66 0. 80 n ot e: su m m ar y st at is tic s of th e m on th ly re tu rn s on th e s& p/ c as esh ill er in di ce s, c r sp va lu ew ei gh te d (v w ) in de x, in te rm ed ia te go ve rn m en t bo nd s, m sc ie a fe in de x, an d th e a ll r e it in de x fr om ja nu ar y 19 87 to d ec em be r 20 11 .t he sa m pl e pe ri od s of 5 m sa s an d th e c s c om po si te 20 in de x ar e sh or te r (p ho en ix m sa fr om ja nu ar y 19 89 ; se at tle fr om ja nu ar y 19 90 ; a tla nt a an d d et ro it fr om ja nu ar y 19 91 ; an d d al la s an d th e c s c om po si te 20 in de x fr om ja nu ar y 20 00 ). a c s c om po si te -1 0 co m po ne nt s 58 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 residential property in our sample period is that the owner enjoys the shelter without much price appreciation. san francisco has the highest risk and the lowest volatility goes to charlotte. the sds of returns of the former and latter msas are 1.41% and 0.60%, respectively. los angeles, san diego, and san francisco share very similar risk and return characteristics. the individual msa data indicate that there are sufficient variations in the performance. the experience of a home owner in one part of the country may be different from that in other parts. the implications of the variations on the home ownership decision remain to be seen. it may simply mean a somewhat lower return without reversing the conclusion that home ownership still pays when an individual is relocated from a high-return city to a low-return one. whether a house is a worthwhile investment depends not just on its risk-return characteristics, but also on the correlation of its returns with those of other asset classes held by the investor. we report the pair-wise correlation coefficients of the returns of the house price and financial asset indices in table 1 panel b. among all pairs of different asset classes, the u.s. equities and eafe stocks have the highest return correlation at 0.72. reits also have relatively high correlations with u.s. equities and eafe at 0.60 and 0.48, respectively. the cs composite 10 and composite 20 indices tend to have very low correlations with all other asset classes including reits.5 it is quite clear from the descriptive statistics that residential properties are very stable assets unaffected by movements in the prices of other financial assets in the long run. further, residential properties (as proxied by the s&p/case-shiller indices) and reits are very different assets with the latter being a high risk and high return proposition. the correlations between individual housing markets and the cs composite 20 index are relatively high with magnitudes at no lower than 0.60. most of the correlations are greater than the highest correlation among financial assets, which is between u.s. equities and eafe stocks at 0.72. the correlations between individual msas and the cs composite 10 index are somewhat lower since the cs composite 10 index does not include some of the individual msas under consideration.6 being a safe asset does not necessarily imply that it must be included in an optimal portfolio. likewise, a risky asset class such as equities may as well be a worthwhile investment as long as it can deliver the return that can justify the bearing of its risk. further, portfolio choice varies from one individual to another depending on the individual’s risk tolerance. to examine the potential risk-adjusted return enhancement effect brought about by home ownership, we perform the optimal portfolio analysis described in section 2 that allows for the consideration of individual’s risk tolerance. 4.2. national residential market the detailed results by using the cs composite-10 index as our proxy for home ownership are reported in table 2. column 1 lists the various levels of monthly risk-free interest rate, rf , being considered, with the last number (0.31%/month) being the ex post historical average risk-free rate based on short-term t-bills. as previously discussed, we solve for the respective optimal portfolios while disallowing short selling. the optimal allocations to 59c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 2 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p w ith c s c om po si te -1 0 in de x as pr ox y r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s c om po si te -1 0 c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 05 79 0. 45 72 0. 00 00 0. 48 49 0. 58 06 0. 12 08 0. 87 92 0. 00 00 0. 47 52 0. 00 0 0. 00 10 0. 06 79 0. 50 10 0. 00 00 0. 43 10 0. 45 33 0. 12 62 0. 87 38 0. 00 00 0. 39 53 0. 00 1 0. 00 20 0. 08 82 0. 58 94 0. 00 00 0. 32 24 0. 33 43 0. 13 46 0. 86 54 0. 00 00 0. 31 56 0. 03 2 0. 00 30 0. 14 87 0. 85 13 0. 00 00 0. 00 00 0. 23 65 0. 14 87 0. 85 13 0. 00 00 0. 23 65 — 0. 00 40 0. 17 79 0. 82 21 0. 00 00 0. 00 00 0. 15 87 0. 17 79 0. 82 21 0. 00 00 0. 15 87 — 0. 00 50 0. 27 41 0. 72 59 0. 00 00 0. 00 00 0. 08 62 0. 27 40 0. 72 60 0. 00 00 0. 08 62 — 0. 00 60 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 04 54 1. 00 00 0. 00 00 0. 00 00 0. 04 54 — 0. 00 70 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 02 37 1. 00 00 0. 00 00 0. 00 00 0. 02 37 — 0. 00 31 a 0. 15 12 0. 84 88 0. 00 00 0. 00 00 0. 22 64 0. 15 12 0. 84 88 0. 00 00 0. 22 64 — n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t c s c om po si te -1 0 ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or ts el lin g di sa llo w ed an d ba se d on th e fu ll sa m pl e pe ri od fr om fe br ua ry 19 87 to d ec em be r 20 11 .t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 31 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 60 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 various asset classes for the case with and without residential housing are reported in columns 2–5 and columns 7–9, respectively. when residential housing is added, the allocation is nil when the historical average t-bills return of 0.31% is used to solve for the tangency portfolio. the optimal allocation in residential housing is also zero at higher levels of risk-free rate. residential housing enters our investor’s optimal portfolio only at lower levels of risk-free rate. note that lower risk-free rates correspond to investors with a lower risk tolerance. this should come as no surprise as residential housing displays extremely low risk and return characteristics and, therefore, is attractive to conservative investors with low risk tolerance. the sharpe ratios corresponding to the cases with and without residential housing are reported in columns 6 and 10, respectively. when the residential housing enters the optimal portfolios at low levels of rf , the improvement in the sharpe ratio is quite meaningful. in fact, whenever it enters the optimal portfolio, the resulting sharpe ratio is always no worse than the baseline case without the investment in residential property. this confirms the well-known fact stated earlier that whenever the estimation of the input parameters, the optimization, and the performance measure are done over the same sample period, the performance of the optimized portfolios with the extra asset class cannot be inferior by construction. any conclusion based on this observed improvement in performance is meaningless without statistical testing. we conduct the statistical tests based on the simulation method of glen and jorion (1993) described earlier and in appendix b. a p-value of 0.05 or lower will indicate that the sharpe ratio with home ownership is statistically significantly higher than the sharpe ratio without home ownership at the usual five-percent level, leading us to conclude that the benefit from home ownership is real rather than purely because of chance. the resulting p-values at different levels of risk-free rate are reported in the last column of table 2. when proxied by the cs composite 10 index, residential housing offers statistically significant diversification benefit to our representative investor at monthly rf of 0.2% (about 2.4% annually) or lower. given the current low interest rate environment of below 2.4%, home ownership can in fact be an optimal decision. at zero risk-free rate, table 2 indicates the proper allocation to a house is 48.5% of an individual’s total wealth.7 in the absence of home ownership, the optimal bond holding for this same individual is 87.9%. home ownership reduces the optimal bond holding to 45.7%, that is, a reduction of about 42 percentage points. in effect, the investment in a house comes about almost exclusively at the expense of bond holding because of their similar risk characteristics. a financial planner, who focuses on the allocation among financial assets alone, may therefore potentially recommend an overinvestment in bonds because of the failure to capture the interaction effect between the return from home ownership and those from other financial assets. table 2 (and subsequent tables) illustrates the importance of a more comprehensive approach in financial planning. while goetzmann (1993) and wu and pandey (2012) incorporate both home ownership and financial assets in their analysis, by ignoring the baseline case of the optimal portfolio without home ownership, they cannot demonstrate the displacement effect of a house on bonds in an optimal portfolio and perform the statistical tests as we have done here. it is important to note that home ownership is attractive not for its potential price appreciation. as pointed out in table 1, the mean return on the cs composite-10 index at 0.29% per month is no match for the 0.81% return from u.s. equities. 61c.s. cheung, p. miu / financial services review 24 (2015) 51–76 we examine the robustness of our results by repeating the previous analysis with the cs composite 20 index as our proxy for the return of home ownership. the results are reported in table 3. the case for home ownership is found to be weaker than that documented in table 2. we offer several reasons for this finding. first, the cs composite 20 index includes some weak housing markets such as atlanta, detroit, phoenix, and dallas that are not included in the cs composite 10 index (individual housing markets will be examined later). second, the cs composite 20 index is available only since january 2000. with its shorter sample period (only a total of 143 monthly observations), the financial/housing crisis of 2008 exerts a stronger influence on the cs composite 20 index than the cs composite 10 index; thus, weakening the case for owning a residential property. nevertheless, despite the relatively stronger negative effect of the financial/housing crisis, home ownership as represented by the cs composite 20 index still offers statistically significant diversification benefit at a risk-free rate that is close to zero based on the p-value of glen and jorion (1993) test (see last column of table 3). in other words, for a risk-averse investor, it still pays to own a house despite the severity of the recent financial/housing crisis and the fact that, as of the end of our sample period, most of the housing markets are still far from fully recovered from their losses. for longer-term home owners, the cs composite 10 index, which covers a much longer time period of about 25 years, offers a more representative indication of the benefit of owning a house. 4.3. individual metropolitan markets as for individual cities, although there are enough similarities to draw definitive statements about home ownership, there are also interesting differences when we repeat the above optimal portfolio analysis using individual msa indices to proxy for the return from home ownership. among the 20 msas, portland presents the best outcome for home ownership. residential housing enters the optimal portfolio at all reasonable levels of risk-free interest rate including the historical average of 0.31%/month (see table 4). only at monthly rf of 0.4% (i.e., about 4.8% annually) or above do we find residential housing becomes unattractive. these high levels of rf correspond to individuals with relatively high-risk tolerance. the logical move for these relatively aggressive individuals is to stay out of safe assets like residential housing and invest in only stocks as indicated in table 4. our results suggest that, for most residents in portland with reasonable degrees of risk aversion, home ownership definitely makes sense economically even without accounting for the consumption value of the shelter a house will provide. to appreciate the attractiveness of real estate to residents of portland, note that the mean return and sd of the u.s. equities over the same sample period (reported in table 1) are 0.81% and 4.62%, respectively. the ratio of the mean to the sd corresponds to the sharpe ratio at a zero risk-free rate. this will work out to be about 0.175% of mean return for each percentage sd of risk for u.s. equities. according to the results in table 4 for portland, the sharpe ratio at the zero risk-free rate for an investor with home ownership is about 0.680% for each percentage point of risk. the improvement over u.s. equities investment is very dramatic. part of the reason for the improvement is the investment in other financial assets in addition to real estate. comparing the sharpe ratio of 0.680 for an investor with home 62 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 3 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p w ith c s c om po si te -2 0 in de x as pr ox y r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s c om po si te -2 0 c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 06 92 0. 61 06 0. 00 00 0. 32 02 0. 49 28 0. 11 05 0. 88 95 0. 00 00 0. 44 81 0. 01 4 0. 00 10 0. 07 43 0. 67 72 0. 00 00 0. 24 85 0. 37 92 0. 10 50 0. 89 50 0. 00 00 0. 36 11 0. 10 0 0. 00 20 0. 08 57 0. 82 37 0. 00 00 0. 09 05 0. 27 58 0. 09 59 0. 90 41 0. 00 00 0. 27 46 0. 37 8 0. 00 30 0. 07 81 0. 92 19 0. 00 00 0. 00 00 0. 18 90 0. 07 81 0. 92 19 0. 00 00 0. 18 90 — 0. 00 40 0. 02 75 0. 97 25 0. 00 00 0. 00 00 0. 10 68 0. 02 75 0. 97 25 0. 00 00 0. 10 68 — 0. 00 50 0. 00 00 1. 00 00 0. 00 00 0. 00 00 0. 03 28 0. 00 00 1. 00 00 0. 00 00 0. 03 28 — 0. 00 60 — — — — — — — — — — 0. 00 70 — — — — — — — — — — 0. 00 18 a 0. 08 28 0. 78 85 0. 00 00 0. 12 87 0. 29 19 0. 09 78 0. 90 22 0. 00 00 0. 28 90 0. 30 7 n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t c s c om po si te -2 0 ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or t se lli ng di sa llo w ed an d ba se d on th e sa m pl e pe ri od fr om fe br ua ry 20 00 to d ec em be r 20 11 . t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 18 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 63c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 4 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p in po rt la nd r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s po rt la nd c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 04 24 0. 37 95 0. 00 00 0. 57 81 0. 67 96 0. 12 08 0. 87 92 0. 00 00 0. 47 52 0. 00 0 0. 00 10 0. 04 78 0. 39 67 0. 00 00 0. 55 55 0. 53 82 0. 12 62 0. 87 38 0. 00 00 0. 39 53 0. 00 0 0. 00 20 0. 05 73 0. 42 74 0. 00 00 0. 51 53 0. 39 93 0. 13 46 0. 86 54 0. 00 00 0. 31 56 0. 00 0 0. 00 30 0. 07 87 0. 49 64 0. 00 00 0. 42 50 0. 26 71 0. 14 87 0. 85 13 0. 00 00 0. 23 65 0. 02 2 0. 00 40 0. 17 20 0. 79 72 0. 00 00 0. 03 09 0. 15 88 0. 17 79 0. 82 21 0. 00 00 0. 15 87 0. 48 4 0. 00 50 0. 27 40 0. 72 60 0. 00 00 0. 00 00 0. 08 62 0. 27 40 0. 72 60 0. 00 00 0. 08 62 — 0. 00 60 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 04 54 1. 00 00 0. 00 00 0. 00 00 0. 04 54 — 0. 00 70 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 02 37 1. 00 00 0. 00 00 0. 00 00 0. 02 37 — 0. 00 31 a 0. 08 35 0. 51 20 0. 00 00 0. 40 45 0. 25 11 0. 15 12 0. 84 88 0. 00 00 0. 22 64 0. 03 0 n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t c s po rt la nd m sa in de x ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or t se lli ng di sa llo w ed an d ba se d on th e fu ll sa m pl e pe ri od fr om fe br ua ry 19 87 to d ec em be r 20 11 . t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 31 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 64 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 ownership with the sharpe ratio of 0.475 for one with only financial assets and also at zero risk-free rate (see first row of table 4), the difference of about 0.205 in sharpe ratio can be attributed to the benefit of adding real estate to the optimal portfolio. this increase in the sharpe ratio still exceeds the sharp ratio of 0.175 for investing in u.s. equities alone. home ownership in portland clearly offers an improvement in the risk-adjusted payoff over that of investing in u.s. equities alone. similar to the value-adding proposition as demonstrated in the previous subsection by using the national cs composite 10 index, the real estate in portland is an attractive addition to an investor’s portfolio mainly because of its ability in risk reduction rather than the potential for price appreciation. table 1 indicates that portland, being the best real estate market, merely offers a mean return of 0.40% per month over our sample period, which is not even half of the mean return of 0.81% per month for u.s. equities. what is the verdict of home ownership for residents of other major metropolitans? residential real estate in the cities of denver, washington dc, los angeles, san diego, san francisco, and seattle offer slightly higher average monthly returns (ranging from 0.30% to 0.35%) than that of the cs composite 10 index (at 0.29%), but they are also (marginally) riskier than the composite index (see the means and sds of returns reported in table 1 panel a). our optimal portfolio analysis using the individual msa indices of these cities confirms that they share the same desirability in home ownership as we have previously demonstrated using the cs composite-10 index. residential property offers statistically significant diversification benefit at risk-free interest rate of no higher than about 0.2% per month (or about 2.4% per year). to conserve space, we only report the outcome for san diego in table 5 as the representative results for these six cities.8 home ownership is still attractive enough for most individuals except for those with high risk tolerance. note that the optimal weight for real estate in san diego at the zero risk-free rate representing the most conservative investor is 38.82%. in fact, the portfolio weights for real estate at all levels of risk-free rate are consistently lower than the allocations for portland residents. given the significantly higher housing prices for west coast cities in california, home owners in california may have to invest more than the optimal weight of 38.82% of their wealth in real estate because of the fact that fractional ownership is not very common. comparing the asset mixes of the optimal portfolios with and without real estate in san diego at zero risk-free rate, the addition of real estate is mainly at the expense of bond investment. the optimal weight on bonds drops from about 88% to about 55% when real estate is added. it appears that real estate serves as a close substitute for bonds because of the similarity in their (low) risk characteristics. if california residents have to overinvest in real estate because of its lumpiness, this overinvestment probably should be accommodated by a reduction in bonds. again, our results clearly demonstrate the importance of comprehensiveness in financial planning with the investor’s home ownership central to the asset allocation decision. according to the risk-return characteristics as reported in table 1 panel a, residential real estate in the cities of miami, chicago, boston, charlotte, new york, and cleveland offer slightly lower return but also lower risk than the cs composite 10 index. based on our optimal portfolio analysis results, these six cities are similar in terms of the degree of enhancement of risk-adjusted return home ownership can offer our representative investor. results for chicago are reported in table 6 as the representative of this group. residential 65c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 5 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p in sa n d ie go r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s sa n d ie go c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 06 66 0. 54 52 0. 00 00 0. 38 82 0. 56 04 0. 12 08 0. 87 92 0. 00 00 0. 47 52 0. 00 0 0. 00 10 0. 07 49 0. 57 16 0. 00 00 0. 35 35 0. 44 85 0. 12 62 0. 87 38 0. 00 00 0. 39 53 0. 00 1 0. 00 20 0. 08 96 0. 61 80 0. 00 00 0. 29 23 0. 34 01 0. 13 46 0. 86 54 0. 00 00 0. 31 56 0. 02 2 0. 00 30 0. 12 25 0. 72 21 0. 00 00 0. 15 54 0. 24 01 0. 14 87 0. 85 13 0. 00 00 0. 23 65 0. 24 8 0. 00 40 0. 17 79 0. 82 21 0. 00 00 0. 00 00 0. 15 87 0. 17 79 0. 82 21 0. 00 00 0. 15 87 — 0. 00 50 0. 27 41 0. 72 59 0. 00 00 0. 00 00 0. 08 62 0. 27 40 0. 72 60 0. 00 00 0. 08 62 — 0. 00 60 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 04 54 1. 00 00 0. 00 00 0. 00 00 0. 04 54 — 0. 00 70 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 02 37 1. 00 00 0. 00 00 0. 00 00 0. 02 37 — 0. 00 31 a 0. 13 00 0. 74 55 0. 00 00 0. 12 45 0. 22 85 0. 15 12 0. 84 88 0. 00 00 0. 22 64 0. 31 2 n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t c s sa n d ie go m sa in de x ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or t se lli ng di sa llo w ed an d ba se d on th e fu ll sa m pl e pe ri od fr om fe br ua ry 19 87 to d ec em be r 20 11 . t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 31 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 66 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 6 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p in c hi ca go r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s c hi ca go c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 07 42 0. 57 04 0. 00 00 0. 35 54 0. 51 92 0. 12 08 0. 87 92 0. 00 00 0. 47 52 0. 00 1 0. 00 10 0. 08 97 0. 64 20 0. 00 00 0. 26 83 0. 41 21 0. 12 62 0. 87 38 0. 00 00 0. 39 53 0. 02 9 0. 00 20 0. 12 23 0. 79 28 0. 00 00 0. 08 49 0. 31 65 0. 13 46 0. 86 54 0. 00 00 0. 31 56 0. 35 1 0. 00 30 0. 14 87 0. 85 13 0. 00 00 0. 00 00 0. 23 65 0. 14 87 0. 85 13 0. 00 00 0. 23 65 — 0. 00 40 0. 17 79 0. 82 21 0. 00 00 0. 00 00 0. 15 87 0. 17 79 0. 82 21 0. 00 00 0. 15 87 — 0. 00 50 0. 27 40 0. 72 60 0. 00 00 0. 00 00 0. 08 62 0. 27 40 0. 72 60 0. 00 00 0. 08 62 — 0. 00 60 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 04 54 1. 00 00 0. 00 00 0. 00 00 0. 04 54 — 0. 00 70 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 02 37 1. 00 00 0. 00 00 0. 00 00 0. 02 37 — 0. 00 31 a 0. 15 12 0. 84 88 0. 00 00 0. 00 00 0. 22 64 0. 15 12 0. 84 88 0. 00 00 0. 22 64 — n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t c s c hi ca go m sa in de x ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or t se lli ng di sa llo w ed an d ba se d on th e fu ll sa m pl e pe ri od fr om fe br ua ry 19 87 to d ec em be r 20 11 . t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 31 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 67c.s. cheung, p. miu / financial services review 24 (2015) 51–76 housing offers statistically significant diversification benefit only at rf of 0.1% per month or lower. the case for home ownership in these cities is weaker than that of the previous group of cities (i.e., denver, washington dc, los angeles, san diego, san francisco, and seattle) as reported in table 5. it is also weaker than the outcome when we use the cs composite 10 index as our proxy (see table 2). at zero risk-free rate an investor still has meaningful optimal exposure to real estate at 35.54%. phoenix, minneapolis, las vegas, dallas, and tampa are somewhat less attractive comparing with the other markets we have examined so far. table 7 reports the results for dallas as the representative of this group. home ownership in this group of cities offers statistically significant diversification benefit only at risk-free interest rate that is close to zero. in other words, home ownership is only attractive to the somewhat more risk-averse individuals. at zero risk-free rate, optimal bond holdings with and without real estate are 59.96% and 88.95%, respectively. in other words, the addition of 33.36% in real estate displaces 28.99% in bonds. consistent with earlier results, bond and real estate are almost perfect substitutes. as a result, financial planners should not focus on investment decisions involving financial assets alone without accounting for home ownership. however, the results for dallas have to be interpreted with caution as there are only 143 monthly observations as opposed to 299 observations for most other cities and the cs composite 10 index. because the cs dallas msa data are only available since january 2000, the 2008 financial/housing crisis has a relatively greater impact on dallas than other cities and thus potentially resulting in a bias against the case for home ownership. in which of the metropolitans should a potential home buyer really think twice before signing on the dotted line? based on our optimal portfolio analysis results, atlanta and detroit are the only two cities where residential property offers no significant diversification benefit regardless of the investor’s risk preference. it is actually not a surprising result if we take a closer look at the statistics reported in table 1 panel a. among our sample of 20 cities, atlanta and detroit offer the lowest average monthly return of 0.09% and 0.08%, respectively; whereas both cities display sds higher than that of the cs composite 10 index. based on this risk-return characteristics alone, we are quite sure that these two cities have very little to offer. only the portfolio results of atlanta are reported here (see table 8). with the exception of atlanta and detroit, home ownership in all cities can significantly enhance the risk-adjusted return of our representative investor when the risk-free interest rate is close to zero. because we have ignored the consumption benefit of home ownership in our analysis, it is fair to say that home ownership pays in 18 of the 20 markets. it is harder to make a definite statement about owning a house in atlanta and detroit, because the consumption value may or may not be able to compensate for the lack-luster portfolio results documented above. in general, our results indicate that residential housing is an attractive investment despite the recent great recession. the conclusion would have been stronger if the consumption component provided by the residential property has been accounted for. an interesting find was that certain cities in the sun belt regions such as atlanta and to a lesser extent phoenix, dallas, tampa, and las vegas appear to be less attractive. housing properties in the west coast cities such as los angeles, san diego, san francisco, seattle, and portland offer better risk and return tradeoff in addition to the pleasant weather. 68 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 7 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p in d al la s r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s d al la s c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 06 68 0. 59 96 0. 00 00 0. 33 36 0. 47 98 0. 11 05 0. 88 95 0. 00 00 0. 44 81 0. 03 7 0. 00 10 0. 08 83 0. 77 82 0. 00 00 0. 13 35 0. 36 34 0. 10 50 0. 89 50 0. 00 00 0. 36 11 0. 34 9 0. 00 20 0. 09 59 0. 90 41 0. 00 00 0. 00 00 0. 27 46 0. 09 59 0. 90 41 0. 00 00 0. 27 46 — 0. 00 30 0. 07 81 0. 92 19 0. 00 00 0. 00 00 0. 18 90 0. 07 81 0. 92 19 0. 00 00 0. 18 90 — 0. 00 40 0. 02 75 0. 97 25 0. 00 00 0. 00 00 0. 10 68 0. 02 75 0. 97 25 0. 00 00 0. 10 68 — 0. 00 50 0. 00 00 1. 00 00 0. 00 00 0. 00 00 0. 03 28 0. 00 00 1. 00 00 0. 00 00 0. 03 28 — 0. 00 60 — — — — — — — — — — 0. 00 70 — — — — — — — — — — 0. 00 18 a 0. 09 78 0. 90 22 0. 00 00 0. 00 00 0. 28 90 0. 09 78 0. 90 22 0. 00 00 0. 28 90 — n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou tc s d al la s m sa in de x ar e re po rt ed at di ff er en tl ev el s of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or t se lli ng di sa llo w ed an d ba se d on th e sa m pl e pe ri od fr om fe br ua ry 20 00 to d ec em be r 20 11 . t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 18 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 69c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 8 d iv er si fic at io n be ne fit s of ho m e ow ne rs hi p in a tla nt a r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe c s a tla nt a c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 10 77 0. 70 97 0. 00 00 0. 18 27 0. 48 94 0. 13 69 0. 86 31 0. 00 00 0. 48 21 0. 10 8 0. 00 10 0. 14 30 0. 85 70 0. 00 00 0. 00 00 0. 39 98 0. 14 30 0. 85 70 0. 00 00 0. 39 98 — 0. 00 20 0. 15 22 0. 84 78 0. 00 00 0. 00 00 0. 31 79 0. 15 22 0. 84 78 0. 00 00 0. 31 79 — 0. 00 30 0. 16 80 0. 83 20 0. 00 00 0. 00 00 0. 23 68 0. 16 80 0. 83 20 0. 00 00 0. 23 68 — 0. 00 40 0. 20 18 0. 79 82 0. 00 00 0. 00 00 0. 15 74 0. 20 18 0. 79 82 0. 00 00 0. 15 74 — 0. 00 50 0. 32 27 0. 67 73 0. 00 00 0. 00 00 0. 08 51 0. 32 27 0. 67 73 0. 00 00 0. 08 51 — 0. 00 60 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 04 87 1. 00 00 0. 00 00 0. 00 00 0. 04 87 — 0. 00 70 1. 00 00 0. 00 00 0. 00 00 0. 00 00 0. 02 65 1. 00 00 0. 00 00 0. 00 00 0. 02 65 — 0. 00 26 a 0. 16 12 0. 83 88 0. 00 00 0. 00 00 0. 26 57 0. 16 12 0. 83 88 0. 00 00 0. 26 57 — n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t c s a tla nt a m sa in de x ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or ts el lin g di sa llo w ed an d ba se d on th e sa m pl e pe ri od fr om fe br ua ry 19 91 to d ec em be r 20 11 . t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 26 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 70 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 4.4. real estate investment trusts finally, we conduct an optimal portfolio analysis for individuals who intend to use reits as a substitute for owning a residential property. the results (reported in table 9) indicate that reits offer no statistically significant diversification benefit at all levels of rf. 9 the risk and return characteristics of reits as reported in table 1 more or less speak for the futility of reits in our representative investor’s portfolio. the summary statistics in table 1 clearly indicate that reits have a somewhat higher risk (based on the sd of return) than the u.s. equity as proxied by the crsp value-weighted index; whereas they have essentially the same average return. further, reits have the second highest correlation with the u.s. equities among the return series. this renders reits a rather unattractive investment except for the most aggressive investor (i.e., as represented by the cases of high levels of risk-free interest rate reported in table 9). this finding is compatible with that of liu and mei (1992) who find that the returns on reits resemble those of small cap stocks. the improvement in the sharpe ratio in these cases of higher levels of rf is, nevertheless, still statistically insignificant. direct investment in a house pays in the long-run. investment in an indirect securitized asset class like reits is no substitute. this information should be useful to renters who may want to hedge the future housing consumption needs with reits. tables 7, 8, and 9 clearly provide examples of the situation where the inclusion of an extra asset always results in the sharpe ratio no worse than the baseline case without the extra asset. without the backing of any statistical tests, such a finding does not necessarily imply an investor will be benefited by including the extra asset in constructing his or her optimal portfolio. the results of our statistical tests suggest that part of the observed enhancement in the in-sample performance when residential property is added may in fact be the result of a pure random effect. the statistical tests performed in this study are crucial to ascertain the contribution of real estate investment to the investor’s/homeowner’s optimal portfolio. 5. conclusion without any good solid empirical evidence on the payoff of home ownership, it is hard to make sound decision regarding one of the biggest decisions a typical household has to make. to compound the matter, financial advisors tend to ignore the home ownership question and focus on the savings and investment decisions involving financial assets. by providing the empirical evidence for the economic benefit of home ownership in a portfolio context involving financial assets and home ownership, we provide a complete analysis beyond what normally offered by financial advisors. our evidence indicates that the financing of home ownership quite often comes about at the expense of bond investments because bonds and real estate share very similar risk and return characteristics. any financial planner who ignores a client’s home ownership will likely recommend an overinvestment in bonds. while the economics of home ownership for conservative individuals is compelling based on our empirical findings, the “buy versus rent” decision also depends on the subjective lifestyle choice as pointed out in some personal finance texts. 71c.s. cheung, p. miu / financial services review 24 (2015) 51–76 t ab le 9 d iv er si fic at io n be ne fit s of in ve st in g in a ll r e it r is kfr ee ra te po rt fo lio s w ith re al es ta te po rt fo lio s w ith ou t re al es ta te g j pva lu e po rt fo lio w ei gh ts sh ar pe ra tio po rt fo lio w ei gh ts sh ar pe ra tio c r sp v w in te r. g ov . bo nd m sc ie a fe a ll r e it c r sp v w in te r. g ov . bo nd m sc ie a fe 0. 00 00 0. 09 93 0. 87 03 0. 00 00 0. 03 04 0. 47 77 0. 12 08 0. 87 92 0. 00 00 0. 47 52 0. 22 2 0. 00 10 0. 10 29 0. 86 42 0. 00 00 0. 03 30 0. 39 77 0. 12 62 0. 87 38 0. 00 00 0. 39 53 0. 24 5 0. 00 20 0. 10 85 0. 85 48 0. 00 00 0. 03 67 0. 31 80 0. 13 46 0. 86 54 0. 00 00 0. 31 56 0. 26 5 0. 00 30 0. 11 80 0. 83 89 0. 00 00 0. 04 31 0. 23 89 0. 14 87 0. 85 13 0. 00 00 0. 23 65 0. 28 4 0. 00 40 0. 13 74 0. 80 62 0. 00 00 0. 05 64 0. 16 13 0. 17 79 0. 82 21 0. 00 00 0. 15 87 0. 30 2 0. 00 50 0. 20 07 0. 70 00 0. 00 00 0. 09 94 0. 08 94 0. 27 40 0. 72 60 0. 00 00 0. 08 62 0. 30 7 0. 00 60 0. 61 28 0. 00 00 0. 00 00 0. 38 72 0. 04 88 1. 00 00 0. 00 00 0. 00 00 0. 04 54 0. 29 9 0. 00 70 0. 58 12 0. 00 00 0. 00 00 0. 41 88 0. 02 59 1. 00 00 0. 00 00 0. 00 00 0. 02 37 0. 30 4 0. 00 31 a 0. 11 96 0. 83 61 0. 00 00 0. 04 43 0. 22 88 0. 15 12 0. 84 88 0. 00 00 0. 22 64 0. 28 8 n ot e: po rt fo lio w ei gh ts an d sh ar pe ra tio s of ta ng en cy po rt fo lio s w ith an d w ith ou t a ll r e it ar e re po rt ed at di ff er en t le ve ls of m on th ly ri sk -f re e in te re st ra te s. o pt im al po rt fo lio s ar e ob ta in ed w ith sh or ts el lin g di sa llo w ed an d ba se d on th e fu ll sa m pl e pe ri od fr om fe br ua ry 19 87 to d ec em be r 20 11 .t he pva lu es of g le n an d jo ri on (g j) te st s on eq ua l sh ar pe ra tio s ar e al so re po rt ed . a t he la st ro w co rr es po nd s to a le ve l of ri sk -f re e in te re st ra te of 0. 31 % /m on th th at is eq ua l to th e m ea n re tu rn on sh or tte rm t re as ur ie s ov er th e sa m pl e pe ri od . 72 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 by varying the degree of risk aversion, we show that home ownership is attractive to conservative investors because of the extremely low risk in investing in residential property as a result of its relatively stable rate of price appreciation. although the market value of residential real estate may have been affected by the volatility in the financial markets during the recent housing crisis, it is in general insensitive to the movements in the financial markets over our 25-year sample period. the correlations of the returns on residential real estate with those of financial assets are in fact very low. these near-zero correlations suggest an individual should own a house because of its stabilizing influences in times of financial turbulences rather than its potential for price appreciation. the price appreciation of the best residential market, portland, is merely one half that of the u.s. equities. the 2008 housing crisis should not displace our perception of the long-run stability of residential property price. reits simply lack the conservative characteristics of a regular house to be qualified as viable substitutes. another major contribution of this article is the comprehensive examination of the 20 regional residential property markets. the sun belt housing markets are no more attractive than the northeastern part of the united states. the west coast in general and portland in particular tend to stand out as the more attractive markets. notes 1 the degree of risk aversion of an individual is likely to be associated with a number of factors including gender, age, wealth and income levels, and other personal characteristics related to his or her inherent risk preference. 2 an investor can easily obtain exposure to the reit sector by buying an etf that replicates a reit index or by buying an actively managed mutual fund that specializes in reits. for empirical evidence on the performance of actively managed mutual funds that specialize in the reit sector, see kaushik and pennathur (2012). 3 see, for example, ingersoll (1987, ch. 4), for the derivation. appendix a to this article also provides a detailed exposition of our portfolio framework. 4 elton, gruber, brown, and goetzmann (2006, ch. 6) suggest the variation of the risk-free rate as a way to trace out the frontier. 5 perhaps not surprisingly, the cs composite 10 and composite 20 indices are almost perfectly positively correlated with each other. 6 pair-wise correlations between individual msas are computed but not produced here to conserve space. the statistics are available on request. 7 the allocation to home ownership would have been higher if the consumption value of the house has been explicitly accounted for as part of the return. 8 results for all individual cities are available upon request. 9 stevenson (2000) also tests for the statistical significance of diversification benefits of adding reits to an equity portfolio without bonds and finds the sharpe ratio of the combined portfolio to be not significantly greater than that of the pure equity portfolio. 73c.s. cheung, p. miu / financial services review 24 (2015) 51–76 appendix a: sharpe ratio and optimal portfolio weight sharpe ratio � is a commonly used measure of risk-adjusted return. it can be defined as the amount of expected return (z̄) over the risk-free interest rate (rf) per unit of return sd (�). that is, � � z̄ � rf � the higher the sharpe ratio the more (excess) return can be generated by an asset with the same amount of risk. for a portfolio of assets, optimizing the sharpe ratio will lead us to invest only at portfolios that lie on the efficient frontier in a plot of expected returns against sds of all possible portfolio compositions as defined by the proportions of the total investment amount (i.e., weights) on individual assets making up the portfolios (see figure a1). in figure a1, z̄p and �p are, respectively, the expected return and the sd of return of the portfolio. given a risk-free interest rate of rf, the portfolio that delivers the highest sharpe ratio is referred to as the tangency portfolio, which can be represented by the tangency point at the efficient frontier of a straight line having an intercept of rf (see figure a1). to solve for the weights w of the tangency portfolio, we can consider the portfolio optimization problem of an investor having the opportunity to invest $1 in a portfolio making up of n risky assets and a risk-free asset delivering a return of rf . let z̄ denotes the n � 1 vector of the expected returns of risky assets and � the n � n variance-covariance matrix of risky asset returns. let � denotes the n � 1 vector of the amounts of money invested in each of the risky assets. the assumption of fully investing $1 ensures that the amount of money invested in the risk-free asset equals to $1 � �' � 1, where 1 is the n � 1 unit vector. suppose fig. a1. 74 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 the investor targets an expected return of rp to be generated from his portfolio. the optimal portfolio that he or she should hold is, therefore, the portfolio that can deliver an expected return of rp, whereas at the same time being the least risky (i.e., with the smallest sd or variance). the investor should invest in a portfolio with composition � such that the variance of portfolio return (�' � � � �) can be minimized, whereas at the same time conforming to the constraint that the target expected return of rp can be achieved. that is, min � ��' � � � �� subject to �' � z̄ � rf � �$1 � �' � 1� � rp the first order condition of the above optimization problem gives us the optimal value for �: � � � � ��1 � � z̄ � rf � 1� where � is a certain constant. the weights w of the risky assets within the optimal risky portfolio (i.e., the tangency portfolio) is, therefore, the normalized value of �. that is, w � � 1' � � � � � ��1 � � z̄ � rf � 1� � � 1' � ��1 � � z̄ � rf � 1� � ��1 � � z̄ � rf � 1� 1' � ��1 � z̄ � rf � 1' � ��1 � 1 this gives us eq. (1). the sharpe ratio of the tangency portfolio is its expected excess return over the risk-free interest rate divided by its sd of return. that is: �p � z̄p � rf �p � z̄' � w � rf �w' � � � w�0.5 , which is our eq. (2). appendix b let us define t as the number of monthly observations. also, let �1 be the sharpe ratio of the tangency portfolio before the addition of the residential property and �2 be the sharpe ratio after the addition. the null hypothesis to be tested is h0 : �1 � �2. when short selling is allowed, the test statistic presented in gibbons, ross, and shanken (1989) and jobson and korkie (1989) is: f � �t � 4� � �2 2 � �1 2 1 � �1 2 , (3) where f follows a f-distribution with 1 and t-4 degrees of freedom. in this study, we consider the case where short selling is disallowed. under this condition, the f statistic of eq. (3) will not be following the f-distribution and needs to be simulated 75c.s. cheung, p. miu / financial services review 24 (2015) 51–76 before hypothesis tests can be conducted. we conduct the simulations by following the method proposed by glen and jorion (1993). we first estimate the means, variances, and the covariances using historical return data. the expected return of the additional asset class (i.e., real estate in this study) is then modified so that the original tangency portfolio is still mean-variance efficient after this additional asset class is included. this in effect ensures the null hypothesis is satisfied in the subsequent simulation exercise. with these modified parameters, t random samples of joint returns are drawn from a multivariate normal distribution. based on these simulated returns, a new set of means and variance-covariance matrix are estimated. sharpe ratios of the tangency portfolios with and without the additional asset class can then be estimated. finally, the value of the test statistic of eq. (3) is computed and recorded. the empirical distribution of the statistic is generated under the null hypothesis by repeating this process 1,000 times. references brueckner, j. k. (1997). consumption and investment motives and the portfolio choices of homeowners. journal of real estate finance and economics, 15, 159–180. elton, e. j., gruber, m. j., brown, s. j., & goetzmann, w. n. (2006). modern portfolio theory and investment analysis (7th ed.). new york, ny: wiley. flavin, m., & yamashita, t. (2002). owner-occupied housing and the composition of the household portfolio. the american economic review, 92, 345–362. gibbons, m. r., ross, s. a., & shanken, j. (1989). a test of the efficiency of a given portfolio. econometrica, 57, 1121–1152. glen, j., & jorion, p. (1993). currency hedging for international portfolios. journal of finance, 48, 1865–1886. goetzmann, w. (1993). the single family home in the investment portfolio. journal of real estate finance and economics, 6, 201–222. hennessey, s. m. (2003). the impact of housing choice on future household wealth. financial services review, 12, 143–164. ingersoll, jr., j. e. (1987). theory of financial decision making. lanham, md: rowman & littlefield publishers. jobson, d., & korkie, b. (1989). a performance interpretation of multivariate tests of asset set intersection, spanning, and mean variance efficiency. journal of financial and quantitative analysis, 24, 185–204. kaushik a.,& pennathur a. k. (2012). an empirical examination of the performance of real estate mutual funds 1990–2008. financial services review, 21, 343–358. liu, c., & mei, j. -p. (1992). the predictability of returns on equity reits and their co-movement with other assets. journal of real estate finance economics, 5, 401–418. madura, j. (2014). personal finance, 5th edition. upper saddle river, nj: pearson education, inc. stevenson, s. (2000). international real estate diversification: empirical tests using hedged indices. journal of real estate research, 19, 105–131. wu, c. y., & pandey, v. k. (2012). the impact of housing on a homeowner’s investment portfolio. financial services review, 21, 177–194. yao, r., & zhang, h. h. (2005). optimal consumption and portfolio choices with risky housing and borrowing constraints. review of financial studies, 18, 197–239. 76 c.s. cheung, p. miu / financial services review 24 (2015) 51–76 financial adviser users and financial literacy bhanu balasubramniana, eric r. briskera,* acollege of business administration, the university of akron, akron, oh 44325, usa abstract using the 2012 national financial capability study we determine what demographic characteristics are associated with individuals that use financial advisers and whether financial advisers have any impact on the financial literacy of their clients. we consider five types of financial advisers: debt counselors, savings or investment, mortgage or loan, insurance, and tax planning. we find a significant increase in the use of financial advisers over the past decade. we also find that savings or investments advisers have the largest positive impact on the financial literacy of their clients, followed by mortgage or loan and insurance advisers, even when controlling for financial education and potential endogeneity issues. © 2016 academy of financial services. all rights reserved. jel classification: d12; d14; d81 keywords: personal financial planning; financial advisers; financial planners; financial literacy; instrumental variable models; propensity score matching 1. introduction we can argue that financial clients follow the advice from their financial advisers without learning from their interactions or financial experience. on the other hand, it is possible that financial clients do learn from their financial experience and interactions with financial advisers. thus, financial advisers may play a significant role in disseminating financial knowledge to their clients that positively impacts their financial literacy. in this article, using the survey data from the 2012 national financial capability study (nfcs), conducted by the financial industry regulatory authority (finra) investor education foundation, we first identify demographic characteristics that are associated with individuals that use financial * corresponding author. tel.: �1-850-766-8315; fax: �1-330-972-5970. e-mail address: ebrisker@uakron.edu (e. r. brisker) financial services review 25 (2016) 127–155 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. advisers in the united states. we then determine if financial advisers have any impact on the financial literacy of their clients. over the past few decades financial products have become more complex to the point that making financial decisions has become very challenging for most people. in addition, retirement planning decisions have become more difficult for many people because of a shift in employee retirement plans away from defined benefit plans and towards defined contribution plans where individuals must make their own investment allocation decisions. for example, in recent years there has been a large increase in the number of investment products available that are typically used for retirement savings such as thousands of new mutual funds, exchange traded funds, target date (life-cycle) funds, and international equity funds. there has also been an increase in the number of non-investment products such as mortgage loan products (i.e., adjustable rate mortgages, reverse mortgages, and interest-only mortgages), insurance products (i.e., universal life insurance, annuities), and personal health insurance products. as these more complex financial products become available investors are more likely to make non-optimal financial decisions simply because they do not understand the financial products they are consuming or how to manage risks. as financial products have grown in complexity the importance of financial literacy in our society has also grown. several studies have considered how financial literacy impact financial decision making and the well-being of individuals. for example, campbell (2006), in a study of household finance, finds that poorer and less educated households having lower financial literacy levels are more likely to make investment mistakes when making financial decisions, and the presence of these types of households may inhibit complex financial innovation. klapper, lusardi, and panos (2013) find that financial literacy is positively related to participating in financial markets and negatively related to the use of informal sources of borrowing, and that individuals with higher levels of financial literacy and unspent income are better able to deal with macroeconomic shocks. other research has shown that financial literacy leads to higher portfolio returns (calvet, campbell, and sodini, 2009), greater wealth and probability of investing in stocks (van rooij, lusardi, and alessie, 2011), better retirement planning (lusardi and mitchell, 2007a, 2007b), and lower cost of borrowing (huston, 2012). on the other hand, lower financial literacy levels lead to higher mortgage costs (moore, 2003), excessive financial burden because of debt (gathergood, 2012), and other debt related problems (lusardi and tufano, 2009). in summary, all prior literature agree that improving financial literacy can lead to better outcomes for both individuals, households, and the entire economic system. to compensate for lower levels of financial literacy, individuals have the option to seek out expert financial advice from professional advisers who will assist them in making complex financial decisions. for example, winchester and huston (2014), using the theory of planned behavior, find that low control beliefs are significantly associated with lower financial-goal progress, however, receiving expert financial advice can help reduce this negative effect and actually result in higher levels of goal progress compared with individuals having high control beliefs that receive no expert financial advice. however, we cannot assume that everyone using financial advisers have lower levels of financial literacy. it is certainly possible that many individuals that use financial advisers already have high levels 128 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 of education and financial literacy, but are not able to stay informed about new financial products or strategies on their own. black, ciccotello, and skipper (2002) propose two personal financial advising delivery models that are based on portfolio theory. in the first model, known as the “specialist model,” the financial consumer, or household, works directly with multiple types of financial advisers such as a debt counselor, investment adviser, mortgage or loan counselor, insurance agent, and tax planner. in the second model, known as the “planner model,” the financial consumer, or household, works directly with a single financial adviser, and this financial adviser acts as the intermediary between the consumer and the relevant financial adviser specialists. the consumer’s choice between these two models depends on the amount and type of financial advice they are seeking. as both of these models use financial advisers, it is important to understand not only who uses financial advisers, but also what specific types of financial advisers they are using. in 2012, finra investor education foundation conducted a national financial capability study. this national survey collected data on demographic characteristics of each survey participant regarding gender, age, ethnicity, education level, marital status, income level, employment status, and the region they live in within the continental united states. the survey also asks if the individual has ever received a financial education from high school, college, an employer, or from the military, and if the respondent has used a specific type of financial adviser. the five specific types of financial advisers included were debt counseling, savings or investments, mortgage or loan, insurance, and tax planning. the survey also asks five questions designed by annamaria lusardi and olivia s. mitchell to assess the financial literacy of the survey respondents (see table 1 and appendix a for a list of these questions). in determining who uses financial advisers, we find that there has been a significant increase in the number of individuals who use financial advisers over the last 14 years. elmerick, montalto, and fox (2002), using the 1998 survey of consumer finances, found that only 21.20% of those surveyed claimed to have used a financial adviser. using the more recent 2012 finra survey, we find that over 53% of those surveyed claimed they used a financial adviser indicating the usage of financial advisers has more than doubled since the 1998 survey. overall, we find that those who have received a financial education, are female, have higher education and income levels, and those that are self-employed are more likely to use a financial adviser. those in younger age groups and non-married individuals are less likely to use financial advisers. when considering the specific financial adviser types separately, we find those having received a financial education and those self-employed are more likely to use all of the different adviser types. females are more likely to use insurance advisers, black (nonhispanic) individuals are more (less) likely to use debt counseling (mortgage or loan) advisers, and younger individuals are more likely to use debt counseling and mortgage or loan advisers. individuals having higher education levels are more likely to use savings or investments and mortgage or loan advisers, those that are married are more likely to use mortgage or loan, insurance, and tax planning advisers, and those with higher income levels are more likely to use all of the different types of advisers but debt counselors. 129b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 1 financial literacy questions panel a: financial literacy questions q1 time value of money suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow? q2 inflation imaging that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account? q3 interest rates if interest rates rise, what will typically happen to bond prices? q4 interest rates a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. q5 diversification buying a single company’s stock usually provides a safer return than a stock mutual fund. panel b: percentage answering financial literacy questions correctly q1 q2 q3 q4 q5 overall 78.6% 66.1% 30.9% 79.4% 53.5% financial adviser user 82.1 71.4 36.3 84.9 61.7 non-user 74.7 60.1 24.8 73.0 44.1 t stat 13.28*** 17.77*** 18.77*** 21.88*** 26.59*** debt counseling user 73.0 56.3 28.2 80.0 48.0 non-user 79.2 67.1 31.2 79.3 54.0 t stat �5.91*** �9.18*** �2.82*** 0.74 �5.12*** savings or investments user 83.8 74.0 40.5 86.1 67.5 non-user 76.2 62.5 26.5 76.2 47.0 t stat 13.61*** 17.64*** 20.36*** 18.24*** 29.78*** mortgage or loan user 82.6 71.0 36.5 88.6 63.2 non-user 77.5 64.8 29.4 76.7 50.7 t stat 8.15*** 8.36*** 9.25*** 21.45*** 15.87*** insurance user 81.6 70.5 35.4 85.1 60.7 non-user 77.2 64.0 28.8 76.5 50.0 t stat 7.69*** 9.77*** 9.84*** 15.76*** 15.27*** tax planning user 82.8 71.7 41.5 85.9 66.2 non-user 77.6 64.8 28.3 77.7 50.3 t stat 8.04*** 9.07*** 16.28*** 13.35*** 19.68*** panel a identifies the five financial literacy questions asked in the survey. panel b reports the percentage of individuals who answered each question correctly for financial adviser users vs. non-users. results are reported for the full sample (overall), for those financial adviser users who used at least one or more of the five types of financial advisers vs. non-users (financial adviser), and for each financial adviser type user vs. non-users, separately. all values are in percentage. t stats are reported indicating the significance in the difference between users vs. non-users. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 130 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 with a large increase in the use of financial advisers in the united states raises the question if financial clients learn from their interactions with financial advisers. bucherkoenen and lusardi (2011) declare that “those exposed to financially knowledgeable people become more financially knowledgeable themselves.” kolb (1984), in the education literature known as the “experiential learning theory,” includes active experimentation, education, observation, and experience as essential components in the general model of learning. bandura (1977) identifies observational learning as the process of social interaction impacting the knowledge of an individual. therefore, individuals who use the services of a financial adviser are likely to learn from their interactions, discussions, experience, and financial outcomes. there is similar literature that shows that individual’s financial literacy improves when their parents have higher levels of education (lusardi, mitchell, and curto, 2010), when they have financial experience (johnson and sherraden, 2007), and when they live in a zip code with higher average education levels (lachance, 2014). tang and peter (2015) also show that financial education, financial experience, and parent’s financial experience exert positive influence on young adult’s financial knowledge. however, there are no studies that examine whether financial advisers have any impact on the financial literacy of their clients. we find that those who use financial advisers answer more of the financial literacy questions correctly compared to those who do not use financial advisers, even when controlling for demographic characteristics that are expected to impact financial literacy. we also control for previous financial education received that has already been shown to be positively related to financial literacy by grimes, rogers, and smith (2010), and walstad, rebeck, and macdonald (2010). when considering the different financial adviser types separately, we find that savings or investment advisers have the largest positive impact on financial literacy, followed by mortgage or loan and insurance advisers, respectively. similar to results reported by robb, babiarz, and woodyard (2012) using the 2009 nfcs survey, we also find a negative relationship between using debt counseling1 advisers and financial literacy scores. it is possible that people having lower financial literacy get into debt and foreclosure problems and because they are poor learners, they do not learn from their experience or interaction with debt counselors. it is also possible that debt counseling services do not offer opportunities to expand their client’s financial literacy in the same manner that savings and investments advisers can since they advise in multiple areas pertaining to personal financial planning. we also find tax planning advisers do not have a significant impact on the financial literacy of their clients as they specialize in tax advising that is not necessarily related to financial literacy topics covered in the questions asked. overall, our results indicate that financial advisers are serving more financial consumers in the united states in recent years, and a positive relationship exists between using certain types of financial advisers and higher financial literacy levels even when controlling for demographic characteristics, such as financial education and education levels, that are known to impact financial literacy. robb, babiarz, and woodyard, (2012) and allgood and walstad (2016) find financial literacy influences the decision to use a financial adviser. allgood and walstad (2016) acknowledge that this relationship may reflect reverse causality, but they largely rule this possibility out without actually testing for reverse causality. in this study, we examine if 131b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 financial advisers actually do influence the financial literacy of their clients, and up to this point we have attempted to control for demographic characteristics that impact financial literacy, such as age, financial education, and education, and we still find a statistically significant positive relationship between financial literacy and using financial advisers. however, we have not directly dealt with the potential endogeneity issues that may exist between the use of financial advisers and financial literacy. we address this potential endogeneity issue by using instrumental variable (iv) generalized method of moments (gmm) and iv probit models where various instruments are used that are correlated with using a specific type of financial adviser but uncorrelated with financial literacy. to correct for the possibility of a confounding variables problem in our base-line model, we also conduct propensity score matching tests where we match financial adviser users with non-users having similar demographic characteristics and then determine how financial advisers impact financial literacy using this matched sample. we continue to find that savings or investments, mortgage or loan, and insurance advisers have a positive influence on financial literacy scores, but tax planning advisers have no significant impact on financial literacy scores. however, we find that debt counseling advisers have a negative influence on financial literacy scores when using iv gmm estimation, but they have no significant impact on financial literacy scores when using iv probit estimation. overall, our results remain robust even when controlling for endogeneity and confounding variables issues indicating a significant positive relationship between financial literacy and using savings or investments, mortgage or loan, and insurance advisers. we contribute to the individual/household financial literacy literature by providing evidence that specific types of financial advisers have a positive influence on financial literacy. this suggests clients learn from their financial experience and interactions with specific types of financial advisers rather than just take advice from them. to the best of our knowledge, this is the first article to consider the impact financial advisers have on the financial literacy of their clients. the remainder of the article is organized as follows. section 2 provides details about the finra nfcs 2012 survey, and our multivariate analysis empirical results are reported in section 3. section 4 presents our iv models and propensity score matching results, and we conclude in section 5. 2. data and descriptive analysis 2.1. finra 2012 national financial capability study the finra investor education foundation conducted the state-by-state financial capability survey that was developed in consultation with the u.s. department of treasury, the president’s advisory council on financial literacy, and multidisciplinary team including academics and practitioners. the data were collected through an online survey of 25,509 adults (18�) across the united states, with approximately 500 respondents per state plus the district of columbia. the survey was self-administered by respondents on a website from 132 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 july to october of 2012 and provides an unprecedented set of information on financial behaviors across the united states. survey respondents were asked to identify several demographic characteristics such as their gender, age, ethnicity, education level, marital status, income level, employment status, region of the united states they lived in, and if they have ever received a financial education in the past. they were also asked if they have used a debt counseling, savings and investments, mortgage or loan, insurance, or tax planning adviser within the past five years. available responses included yes, no, do not know, and prefer not to say. after limiting our sample to respondents who answered either yes or no for each of the financial adviser type use questions we are left with 22,218 out of the 25,509 survey respondents for our analysis. out of these 22,218 survey respondents, 53.4% used at least one of the five types of financial advisers compared with 46.6% that had not. the survey respondents are also asked five basic finance questions designed by annamaria lusardi and olivia s. mitchell to assess their financial literacy. the five questions asked are designed to test the individuals understanding of basic concepts related to the time value of money, interest rates, inflation, and risk diversification. these questions are widely used as a measure of financial literacy in the literature. we will use the survey respondent’s answers to these five questions to measure their financial literacy level. 2.2. descriptive analysis in table 2 we report the percentage of each demographic characteristic group that used at least one type of financial adviser in the fa user column, and for each specific type of financial adviser in the debt counselor, savings or investments, mortgage of loan, insurance, and tax planning columns. the demographic characteristics considered are financial education, gender, age group, ethnicity, education, marital status, income, employment, and region. these demographic characteristics have been shown to impact personal financial behavior and the probability of using a financial adviser when making financial decisions (barber and odean, 2001; elmerick, montalto, and fox, 2002). we report t-stats indicating the significance in the difference between the two identified groups within each demographic characteristic. we find a large variation in the percentage of individuals using at least one type of financial adviser based on their education and income levels. there is a much larger percentage of individuals using financial advisers who are either a college graduate (62.9%) or have post graduate college (69.7%) compared with those who have either completed high school (44%) or less than high school (30.4%). not surprisingly, we also find that that there is a monotonic increase in the percentage of financial adviser users across higher income levels. for instance, 75.1% of those earning more than $150,000 used a financial adviser compared with only 40.3% of those earning $15,000 to $24,999. when considering the specific finance adviser types, we find that a larger percentage of individuals falling in lower income categories and younger age groups reported using debt counselors. we find a larger percentage of individuals in lower age groups and higher income levels used mortgage or loan advisers. there are also a larger percentage of individuals falling in higher age groups, higher education levels, and higher income levels 133b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 2 percent of individuals that use financial advisers fa user debt counseling savings or investments mortgage or loan insurance tax planning total % using 53.4% 8.8% 31.6% 22.1% 32.9% 19.9% financial education taken 65.6 11.5 42.4 29.5 42.9 27.9 not taken 49.8 8.0 28.4 19.9 29.9 17.5 taken not taken t test 20.70*** 7.23*** 18.15*** 13.58*** 16.70*** 15.01*** gender male 54.3 8.6 33.8 23.3 33.2 21.5 female 52.7 8.9 29.9 21.2 32.7 18.6 male female t test 2.29** �0.72 6.16*** 3.81*** 0.71 5.38*** age group 18–24 43.7 9.0 26.1 17.2 26.5 16.3 25–34 52.6 13.9 29.3 28.5 34.9 20.4 35–44 52.4 11.5 26.1 27.0 35.1 18.2 45–54 50.0 8.1 27.1 20.9 32.7 17.8 55–64 55.4 7.1 34.9 19.8 33.3 20.1 65� 62.3 3.9 43.7 18.2 32.3 25.4 (18–24) (65�) t test �13.90*** 7.28*** �14.10*** �0.99 �4.77*** �8.55*** ethnicity white 54.2 7.5 32.6 22.4 32.6 20.2 black 51.4 12.4 29.0 21.2 33.8 19.0 white black t test 3.59*** �10.41*** 5.15*** 1.96** �1.66* 2.05*** education � high school 30.4 6.6 10.6 8.9 18.9 7.8 high school 44.0 7.2 23.1 16.0 25.2 13.3 ged 38.0 8.3 16.2 14.5 24.6 11.0 some college 54.1 8.7 30.8 22.3 34.3 18.7 college graduate 62.9 11.6 40.6 28.5 39.5 25.3 post-graduate 69.7 8.1 50.1 30.8 41.4 34.4 �hs–post-grad t test �27.33*** �1.92* �32.70*** �19.71*** �16.76*** �24.07*** marital status married 59.6 8.6 36.2 26.5 36.7 24.0 single 43.9 9.2 26.2 16.7 26.7 14.4 separated 39.4 10.6 18.0 11.9 28.4 14.2 divorced 47.3 8.7 23.9 16.8 29.4 13.9 widowed 52.0 7.6 30.3 14.5 30.7 17.3 married widowed t test 4.45*** 1.13 3.70*** 9.72*** 3.73*** 5.10*** income �$15,000 27.4 6.2 11.5 7.9 16.5 6.4 $15k-$24,999 40.3 10.1 16.8 11.8 26.3 9.5 $25k-$34,999 45.7 9.5 22.4 16.7 28.6 13.1 $35k-$49,999 51.8 11.1 26.7 19.7 32.6 16.3 $50k-$74,999 59.2 9.1 35.1 25.6 36.8 21.3 $75k-$99,999 64.2 9.1 43.1 30.4 38.6 26.5 $100k-$149,000 70.3 7.0 50.0 32.8 42.0 32.8 � $150,000 75.1 6.8 57.6 36.8 45.3 43.4 �$15k �$150k t test �34.09*** �0.87 �33.07*** �21.83*** �19.92*** �27.55*** employment self employed 60.1 9.2 36.3 25.7 41.3 28.8 full-time 59.2 11.0 35.0 29.4 37.5 21.6 part-time 51.4 9.6 31.8 20.2 32.3 20.0 homemaker 45.4 8.0 21.3 19.9 28.7 15.9 student 43.3 9.4 27.8 15.5 25.4 13.9 134 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 that use savings and investments advisers. in fact, there is a monotonic increase in the percentage of individuals using a savings and investments, mortgage or loan, insurance, or tax planning adviser across both higher education and income levels. individuals using insurance or tax planning financial advisers are mixed across the other demographic groups, and gender, ethnicity, marital status, employment, and region reveal mixed results for each financial adviser type. in table 1, panel a, we identify the five financial literacy questions used in the survey, and in table 1, panel b, we report the percentage of individuals answering each of the five financial literacy questions correctly. the first row in panel b, overall, reports the percentage of survey respondents answering each question correctly for the entire sample, followed by user versus non-user comparisons for those using one or more of the different financial adviser types (financial advisor), and for each of the financial adviser types, separately. we also report t-stats indicating the significance in the difference between the percentages of financial adviser(s) users versus non-users who answered each question correctly. overall, we find that 78.6% of individuals answered the question regarding the time value of money correctly, but only 30.9% answered the question pertaining to the relationship between interest rates and bond prices correctly. just over 66% answered the question pertaining to inflation correctly, but nearly 80% answered the question regarding mortgage interest expenses correctly. surprisingly, only 53.5% answered the question pertaining to diversification using mutual funds correctly. we find that a greater percentage of those using one or more of the five types of financial advisers, reported under financial adviser, answered all five questions correctly compared with those not using any of the five types of advisers, and these differences are all highly significant at the 1% level. when considering the five types of financial advisers separately, we find that a higher percentage of individuals using table 2 (continued) fa user debt counseling savings or investments mortgage or loan insurance tax planning disabled 37.0 8.2 12.6 11.3 25.6 6.9 unemployed 34.4 8.5 15.6 11.4 20.8 10.7 retired 59.2 4.4 40.8 17.4 31.7 23.6 self-employed retired t test 0.63 6.37*** -3.24*** 6.97*** 7.02*** 4.15*** region northeast 53.6 8.5 34.0 22.1 32.0 21.2 midwest 53.6 7.7 31.6 21.5 33.8 19.8 south 50.4 9.2 29.5 19.9 31.0 18.7 west 57.1 9.4 32.8 25.5 35.2 20.7 northeast west t test �3.33*** -1.50 1.24 �3.79*** �3.28*** 0.62 the table shows the percent of individuals surveyed that used financial advisers. the fa user column report results for people that used at least one or more of the five types of financial advisers. the specific financial adviser columns report the percentage of individuals using that specific type of adviser. all values are in percentages. t stats are reported indicating the significance in the difference between the two identified groups within each demographic characteristic. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 135b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 savings or investments, mortgage of loan, insurance, and tax planning advisers answered all five questions correctly compared to those not using those types of advisers (with the statistical significance of these differences being highly significant at the 1% level). however, we find that a lower percentage of those using debt counselors answered all of the questions correctly compared to those not using debt counselors except for the question regarding mortgage interest expenses (q4). 3. multivariate analysis 3.1. financial adviser users to determine who is using financial advisers, we use multivariate ols regressions to relate each demographics characteristic associated with individuals that used one or more of the five types of financial advisers while controlling for all other demographic characteristics. we also estimate a multinomial logit regression model to estimate the effect each demographic characteristic has on the probability of using at least one of the five types of financial advisers while simultaneously controlling for the effects of all other demographic characteristics. the dependent variable in both models, financial adviser users, is a dummy variable set equal to 1 if the survey respondent used at least one of the five types of financial advisers, or 0 otherwise. the demographic characteristics included are financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region. financial education is a dummy variable set equal to 1 if the respondent had any financial education training, or 0 otherwise. female is a dummy variable set equal to 1 for females and 0 for males. black (non-hispanic) is a dummy variable set equal to 1 if the person is black (non-hispanic), or 0 otherwise. the categories within the age group, education, marital status, income, employment, and region demographic characteristic as identified in table 3 are set equal to 1 if the respondent identifies themselves to be in that categorical group, or 0 otherwise. the first two columns in table 3 report coefficient estimates and t-stats from ols estimation. we find the ols coefficient estimates for financial education and female is positive and highly significant indicating people having received financial education in the past and females tend to use financial advisers. we also find there is a highly statistically significant monotonic increase in the use of financial advisers across higher education and income levels. those in younger age groups and non-married people are less likely to use financial advisers. however, results are mixed for different employment and region categories and not significant for black (non-hispanic) indicating ethnicity does not appear related to using a financial adviser. the last three columns in table 3 report logit coefficient estimates, z-stats, and odds ratios from the logit regression results. positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to use a financial adviser compared with the reference category identified in the table for each demographic characteristic. consistent with the ols results, we find that those having a financial education are 62% more likely, and females 14% more likely, to use a financial adviser. we also find there is a monoton136 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 3 financial adviser user ols and logit regressions financial adviser users ols coefficient t stat logit coefficient z-stat odds ratio constant 0.294*** 12.68 �0.921*** �8.47 0.398 financial education 0.105*** 13.79 0.483*** 13.49 1.621 female 0.030*** 4.52 0.135*** 4.44 1.144 age group 25–34 �0.018 �1.25 �0.082 �1.28 0.921 35–44 �0.054*** �3.70 �0.243*** �3.71 0.784 45–54 �0.069*** �4.85 �0.309*** �4.83 0.734 55–64 �0.026* �1.77 �0.116* �1.74 0.891 65� 0.024 1.41 0.109 1.44 1.116 black (non-hispanic) �0.008 �1.02 �0.035 �1.01 0.965 education (reference category: did not complete high school) high school 0.037*** 2.58 0.177*** 2.63 1.194 ged 0.005 0.26 0.035 0.43 1.036 some college 0.100*** 7.21 0.443*** 6.76 1.558 college graduate 0.147*** 9.80 0.649*** 9.24 1.913 post-graduate 0.159*** 9.77 0.717*** 9.25 2.049 marital status (reference group: married) living-with-partner �0.022* �1.78 �0.099* �1.76 0.906 single �0.043*** �5.34 �0.194*** �5.44 0.823 income (reference category: less than $15,000) $15,000 $24,999 0.102*** 7.80 0.481*** 7.85 1.618 $25,000 $34,999 0.134*** 9.73 0.610*** 9.71 1.840 $35,000 $49,999 0.177*** 13.24 0.787*** 12.87 2.196 $50,000 $74,999 0.230*** 17.35 1.004*** 16.45 2.729 $75,000 $99,999 0.262*** 17.73 1.142*** 16.79 3.134 $100,000 $149,999 0.305*** 20.04 1.349*** 18.76 3.853 $150,000 or more 0.339*** 20.01 1.532*** 18.26 4.628 employment (reference category: self employed) employed full-time �0.038*** �3.05 �0.175*** �3.05 0.839 employed part-time �0.039** �2.50 �0.176** �2.51 0.839 homemaker �0.102*** �6.41 �0.453*** �6.32 0.636 full-time student �0.097*** �4.41 �0.426*** �4.36 0.653 disabled �0.066*** �3.63 �0.296*** �3.53 0.744 unemployed �0.105*** �6.48 �0.481*** �6.38 0.618 retired �0.036** �2.47 �0.169** �2.49 0.845 region (reference category: northeast) midwest 0.021** 2.07 0.093** 2.07 1.098 south �0.002 �0.18 �0.007 �0.17 0.993 west 0.042*** 4.25 0.188*** 4.21 1.207 observations 22,218 22,218 adjusted r2 0.110 this table reports ols and logit regression results where the dependent variable, financial adviser users, is set equal to 1 if the person used at least one of the five types of financial advisers, or 0 otherwise. the independent variables include financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region. positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to use a financial adviser compared to the indicated reference category. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 137b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 ically increasing likelihood of using a financial adviser across higher education and income levels, but age group is significantly related to a reduced probability of using a financial adviser for those in younger age groups. compared with the univariate results presented in table 2, this shows that it is income rather than age that increases the probability of an individual using a financial adviser. we also see that married people are more likely to use financial advisers than either single or “living-with-partner” couples, and black (nonhispanic) does not impact the likelihood of using a financial adviser. when considering employment, those that are self-employed are more likely to use financial advisers when compared with all other employment categories. finally, we find that those from the midwest (west) are 10% (21%) more likely to use a financial adviser compared with those from the south or northeast. in table 4 we estimate the same ols and logit models for each of the five financial adviser types separately. in this case, each financial adviser type to include debt counseling, savings or investments, mortgage or loan, insurance, and tax planning is set equal to 1 if the person used that type of adviser, and zero otherwise. we use the same demographic characteristics and categories that were used in table 3. in discussing the ols estimation results (first two columns for each adviser type), we find that financial education is both positive and highly statistically significant for all financial adviser types indicating people having received financial education in the past are associated with using all of the five types of financial advisers. however, females are only associated with using insurance advisers with all other coefficient estimates for female being not significant for other adviser types. we also find that black (non-hispanic) is positively (negatively) related to using debt counseling (mortgage or loan) advisers, but not significantly related to using the other three financial adviser types. we find a monotonically decreasing relationship between age groups and debt counseling where lower (higher) age groups are positively (negatively) related to using debt counseling. we also find that the 65� age group is positively related to using savings or investment advisers, the 25 to 34 age group is positively related to using mortgage or loan advisers. for education, people having a high school education or higher are associated with using savings or investment and mortgage or loan advisers, while only those having some college or higher are associated with using insurance and tax planning advisers. for marital status, married couples are associated with using mortgage or loan, insurance, and tax planning advisers. for income, we find those in lower income levels are associated with using debt counseling while those in higher income levels are increasingly associated with using savings or investment, mortgage or loan, insurance, and tax planning advisers. for employment, all coefficient estimates are either significantly negative or not significant, indicating the reference category, self employed, is positively related to using all of the financial adviser types compared with the other employment groups. for region, we find those in the west are associated with using mortgage or loan advisers, and those in the midwest and west are associated with using insurance advisers. the logit estimation results (last three columns for each adviser type) are consistent with the ols results. those receiving a financial education are from between 39% to 53% more likely to use all of the financial adviser types compared to those not receiving 138 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 4 specific type of financial adviser user ols and logit regressions debt counseling savings or investments ols coefficient t stat logit coefficient z-stat odds ratio ols coefficient t stat logit coefficient z-stat odds ratio constant 0.050*** 3.58 �2.901*** �15.27 0.0550 0.130*** 6.58 �2.181*** �16.15 0.113 financial education 0.030*** 6.08 0.355*** 6.48 1.426 0.083*** 10.86 0.399*** 11.03 1.490 female 0.004 0.92 0.034 0.67 1.034 0.003 0.47 0.019 0.59 1.019 age group (reference category: 18–24) 25–34 0.038*** 4.11 0.367*** 3.70 1.444 �0.043*** �3.42 �0.279*** �3.82 0.757 35–44 0.016* 1.67 0.171 1.62 1.186 �0.106*** �8.36 �0.622*** �8.26 0.537 45–54 �0.014 �1.62 �0.166 �1.55 0.847 �0.088*** �7.09 �0.520*** �7.11 0.594 55–64 �0.019** �2.05 �0.225** �1.97 0.799 �0.030** �2.30 �0.210*** �2.82 0.811 65� �0.039*** �3.94 �0.675*** �4.52 0.509 0.028* 1.80 0.051 0.62 1.052 black (non-hispanic) 0.030*** 5.94 0.325*** 6.04 1.384 �0.007 �1.04 �0.037 �0.96 0.964 education (reference category: did not complete high school) high school 0.005 0.68 0.080 0.64 1.083 0.038*** 3.65 0.451*** 4.79 1.570 ged 0.009 0.92 0.139 0.95 1.149 �0.000 �0.04 0.154 1.36 1.166 some college 0.015* 1.95 0.218* 1.83 1.244 0.075*** 7.39 0.664*** 7.26 1.943 college graduate 0.032*** 3.59 0.397*** 3.17 1.488 0.142*** 12.10 0.979*** 10.37 2.663 post-graduate 0.011 1.17 0.167 1.20 1.181 0.162*** 11.76 1.030*** 10.41 2.802 marital status (reference category: married) living-with-partner �0.012 �1.53 �0.145 �1.56 0.865 0.004 0.39 0.012 0.19 1.012 single �0.009* �1.78 �0.116* �1.92 0.890 0.001 0.09 0.010 0.26 1.010 income (reference category: less than $15,000) $15,000 $24,999 0.042*** 5.42 0.577*** 5.39 1.780 0.039*** 3.94 0.366*** 4.45 1.443 $25,000 $34,999 0.030*** 3.81 0.432*** 3.86 1.540 0.083*** 7.69 0.665*** 8.27 1.944 $35,000 $49,999 0.045*** 5.64 0.588*** 5.41 1.801 0.114*** 10.56 0.842*** 10.93 2.320 $50,000 $74,999 0.023*** 2.95 0.338*** 3.02 1.402 0.185*** 16.69 1.181*** 15.70 3.257 $75,000 $99,999 0.018** 2.10 0.277** 2.24 1.319 0.252*** 19.31 1.473*** 18.30 4.361 $100,000 $149,999 0.001 0.16 0.046 0.34 1.047 0.304*** 22.01 1.680*** 20.35 5.364 $150,000 or more �0.001 �0.05 0.016 0.11 1.016 0.364*** 22.28 1.923*** 21.13 6.840 employment (reference category: self employed) employed full-time 0.007 0.84 0.064 0.70 1.066 �0.037*** �3.02 �0.183*** �3.06 0.832 employed part-time 0.000 0.05 0.007 0.06 1.007 0.007 0.47 0.045 0.61 1.046 homemaker �0.023** �2.40 �0.280** �2.31 0.756 �0.072*** �5.05 �0.417*** �5.25 0.659 full-time student �0.012 �0.84 �0.121 �0.77 0.886 �0.033 �1.64 �0.143 �1.32 0.867 disabled �0.000 �0.01 �0.001 �0.01 0.999 �0.077*** �5.15 �0.582*** �5.38 0.559 unemployed �0.007 �0.73 �0.077 �0.61 0.926 �0.066*** �4.67 �0.436*** �4.89 0.647 retired �0.018** �2.24 �0.321** �2.56 0.725 0.005 0.38 0.018 0.26 1.018 region (reference category: northeast) midwest �0.011* �1.86 �0.142* �1.81 0.868 0.004 0.44 0.028 0.58 1.028 south 0.000 0.07 0.004 0.06 1.004 �0.010 �1.14 �0.056 �1.22 0.945 west �0.000 �0.01 0.005 0.07 1.005 0.004 0.47 0.028 0.58 1.028 observations 22,218 22,218 22,218 22,218 adjusted r2 0.023 0.127 mortgage or loan insurance ols coefficient t stat logit coefficient z-stat odds ratio ols coefficient t stat logit coefficient z-stat odds ratio constant 0.135*** 7.37 �2.215*** �14.87 0.109 0.207*** 9.53 �1.500*** �12.75 0.223 financial education 0.060*** 8.40 0.336*** 8.69 1.399 0.097*** 12.26 0.431*** 12.52 1.539 female 0.004 0.77 0.025 0.70 1.025 0.023*** 3.58 0.112*** 3.60 1.118 age group (reference category: 18–24) 25–34 0.032*** 2.76 0.117 1.51 1.125 0.011 0.87 0.044 0.66 1.045 35–44 �0.009 �0.80 �0.134* �1.68 0.874 �0.008 �0.57 �0.048 �0.70 0.953 45–54 �0.065*** �5.71 �0.458*** �5.73 0.633 �0.026** �1.97 �0.132* �1.94 0.876 55–64 �0.076*** �6.39 �0.533*** �6.36 0.587 �0.022 �1.58 �0.114 �1.62 0.892 65� �0.092*** �6.62 �0.635*** �6.58 0.530 �0.037** �2.27 �0.184** �2.28 0.832 black (non-hispanic) �0.026*** �3.99 �0.163*** �3.95 0.850 0.011 1.50 0.053 1.49 1.054 education (reference category: did not complete high school) high school 0.030*** 3.18 0.353*** 3.40 1.423 0.006 0.51 0.079 1.02 1.082 ged 0.022* 1.83 0.282** 2.31 1.325 0.010 0.65 0.093 1.01 1.098 some college 0.068*** 7.19 0.616*** 6.17 1.851 0.071*** 5.84 0.406*** 5.48 1.501 139b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 4 (continued) mortgage or loan insurance ols coefficient t stat logit coefficient z-stat odds ratio ols coefficient t stat logit coefficient z-stat odds ratio college graduate 0.086*** 7.89 0.700*** 6.78 2.015 0.088*** 6.49 0.472*** 6.05 1.603 post-graduate 0.087*** 6.81 0.706*** 6.51 2.026 0.087*** 5.67 0.465*** 5.55 1.591 marital status (reference category: married) living-with-partner �0.022** �2.04 �0.123* �1.84 0.885 �0.026** �2.13 �0.120** �2.03 0.887 single �0.054*** �8.20 �0.357*** �8.02 0.700 �0.040*** �5.10 �0.188*** �5.01 0.829 income (reference category: less than $15,000) $15,000 $24,999 0.023*** 2.76 0.336*** 3.50 1.400 0.082*** 7.07 0.517*** 7.32 1.677 $25,000 $34,999 0.051*** 5.34 0.612*** 6.57 1.844 0.089*** 7.28 0.556*** 7.75 1.744 $35,000 $49,999 0.068*** 7.09 0.732*** 8.14 2.080 0.118*** 9.78 0.695*** 10.02 2.004 $50,000 $74,999 0.113*** 11.52 0.997*** 11.40 2.709 0.146*** 12.13 0.819*** 11.98 2.268 $75,000 $99,999 0.146*** 12.40 1.151*** 12.42 3.163 0.150*** 10.89 0.834*** 11.20 2.303 $100,000 $149,999 0.168*** 13.33 1.252*** 13.20 3.499 0.175*** 12.00 0.938*** 12.20 2.556 $150,000 or more 0.202*** 13.33 1.404*** 13.79 4.071 0.201*** 11.82 1.043*** 12.30 2.836 employment (reference category: self-employed) employed full-time 0.005 0.40 �0.020 �0.32 0.980 �0.064*** �4.98 �0.291*** �5.23 0.747 employed part-time �0.025* �1.85 �0.141* �1.75 0.868 �0.062*** �4.01 �0.271*** �3.88 0.763 homemaker �0.052*** �3.76 �0.319*** �3.86 0.727 �0.111*** �7.09 �0.511*** �7.02 0.600 full-time student �0.079*** �4.51 �0.491*** �4.01 0.612 �0.126*** �6.20 �0.581*** �5.65 0.559 disabled �0.040*** �2.80 �0.328*** �2.88 0.720 �0.060*** �3.42 �0.261*** �2.97 0.770 unemployed �0.064*** �4.87 �0.485*** �4.92 0.615 �0.118*** �7.63 �0.592*** �7.40 0.553 retired �0.037*** �2.84 �0.254*** �3.21 0.776 �0.073*** �4.87 �0.331*** �4.91 0.718 region (reference category: northeast) midwest 0.005 0.64 0.029 0.54 1.029 0.027*** 2.82 0.130*** 2.81 1.139 south �0.002 �0.22 �0.012 �0.24 0.988 0.003 0.31 0.013 0.29 1.013 west 0.037*** 4.35 0.223*** 4.32 1.250 0.028*** 2.93 0.133*** 2.92 1.143 observations 22,218 22,218 22,218 22,218 adjusted r2 0.074 0.051 tax planning ols coefficient t stat logit coefficient z-stat odds ratio constant 0.166*** 9.28 �2.110*** �13.34 0.121 financial education 0.063*** 9.11 0.379*** 9.46 1.461 female �0.005 �0.83 �0.028 �0.75 0.972 age group (reference category: 18–24) 25–34 �0.024** �2.15 �0.250*** �2.93 0.779 35–44 �0.073*** �6.44 �0.595*** �6.76 0.551 45–54 �0.074*** �6.64 �0.599*** �6.95 0.549 55–64 �0.067*** �5.76 �0.552*** �6.27 0.576 65� �0.036*** �2.61 �0.371*** �3.84 0.690 black (non-hispanic) 0.005 0.80 0.036 0.83 1.037 education (reference category: did not complete high school) high school 0.007 0.83 0.193* 1.76 1.212 ged �0.002 �0.20 0.073 0.56 1.076 some college 0.035*** 4.05 0.431*** 4.10 1.539 college graduate 0.069*** 6.79 0.637*** 5.87 1.891 post-graduate 0.111*** 8.94 0.815*** 7.24 2.259 marital status (reference category: married) living-with-partner �0.038*** �4.03 �0.300*** �3.89 0.740 single �0.020*** �3.19 �0.139*** �3.02 0.870 income (reference category: less than $15,000) $15,000 $24,999 0.026*** 3.31 0.382*** 3.63 1.465 $25,000 $34,999 0.055*** 6.17 0.688*** 6.74 1.990 $35,000 $49,999 0.080*** 8.84 0.898*** 9.21 2.455 140 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 a financial education. females are 12% more likely to use insurance advisers, but we find no other significant difference related to gender in using the other financial adviser types. those that are black (non-hispanic) are 38% more likely to use debt counseling advisers, but 15% less likely to use mortgage or loan advisers. we find a monotonically decreasing probability of using debt counseling and mortgage or loan advisers across higher age groups, but all other results related to age group are mixed. we find a monotonically increasing probability of using savings or investments and mortgage or loan advisers across higher education levels, while those having some college or higher are far more likely to use insurance and tax planning advisers. for marital status, we find that those that are married are 12% to 30% more likely to use mortgage or loan advisers, 11% to 17% more likely to use insurance advisers, and 13% to 26% more likely to use tax planning advisers compared with those that are single or “living-withpartner.” we find a nearly monotonic decreasing probability of using debt counseling advisers across higher income levels, but a monotonically increasing probability of using savings or investments, mortgage or loan, insurance, and tax planning advisers across higher income levels. for employment, those that are self-employed are more likely to use any of the financial adviser types compared with the other employment groups. for table 4 (continued) tax planning ols coefficient t stat logit coefficient z-stat odds ratio $50,000 $74,999 0.120*** 12.90 1.164*** 12.30 3.202 $75,000 $99,999 0.163*** 14.39 1.401*** 14.04 4.060 $100,000 $149,999 0.214*** 17.35 1.650*** 16.26 5.205 $150,000 or more 0.307*** 20.13 2.041*** 19.10 7.698 employment (reference category: self-employed) employed full-time �0.095*** �8.41 �0.583*** �9.15 0.558 employed part-time �0.049*** �3.65 �0.264*** �3.25 0.768 homemaker �0.084*** �6.28 �0.528*** �6.10 0.590 full-time student �0.121*** �7.00 �0.798*** �5.99 0.450 disabled �0.098*** �7.55 �0.871*** �6.59 0.419 unemployed �0.084*** �6.51 �0.574*** �5.73 0.563 retired �0.049*** �3.70 �0.277*** �3.67 0.758 region (reference category: northeast) midwest 0.007 0.85 0.051 0.93 1.053 south �0.003 �0.39 �0.023 �0.44 0.977 west 0.004 0.47 0.031 0.58 1.032 observations 22,218 22,218 adjusted r2 0.088 this table reports ols and logit regression results where the dependent variable, debt counseling, savings or investments, mortgage or loan, insurance, or tax planning is set equal to 1 if the person used that type of financial adviser, or 0 otherwise. the independent variables include financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region. positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to use a financial adviser compared with the indicated reference category. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 141b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 region, those in the west are 25% more likely to use mortgage or loan advisers, and those in the midwest and west are 14% more likely to use insurance advisers. 3.2. financial literacy we will now determine if financial advisers have an impact on the financial literacy of individuals that use them versus those that do not. we first estimate a linear ols regression model to determine if using at least one of the five types of financial advisers impacts the number of correct answers giving for the five financial literacy questions asked in the survey. we then estimate an ordered logit regression model to determine if using at least one of the five types of financial advisers impacts the probability of answering more of the financial literacy questions correctly. the dependent variable in both models, financial literacy score, is set equal to the number of questions the survey respondent answered correctly ranging from 0 to 5. the independent variable of interest, fa user, is a dummy variable set equal to one if the survey respondent used at least one of the five types of financial advisers in the past five years, or 0 otherwise. we then re-estimate both models where we replace the fa user independent dummy variable with separate dummy variables for each financial adviser type to include debt counseling, savings or investments, mortgage or loan, insurance, and tax planning that are set equal to 1 if the survey respondent used that type of financial adviser, or 0 otherwise. it is important for us to include the demographic characteristics financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region as control variables in our models since many of them are expected to influence financial literacy. the first two columns in the left panel of table 5 report the coefficient and t-stat results from the ols model specification using fa user as the main independent variable of interest. we first notice that financial education is positively related to higher financial literacy scores as expected, but female and black (non-hispanic) are significantly related with lower financial literacy scores. we also find a monotonically increasing positive relationship with higher financial literacy scores across higher age groups, education levels, and income levels. both married and self-employed people scored higher financial literacy scores compared with the other categories under marital status and employment, and those in the northeast scored lower financial literacy scores compared with the other categories under region. our main variable of interest, fa user, is positive, and highly statistically significant, indicating that those using at least one of the five types of financial advisers answered more of the financial literacy questions correctly compared with those that did not use any financial advisers. this is our first evidence that financial advisers positively influence the financial literacy of their clients, even after controlling for other demographic characteristics that are also positively related to higher financial literacy scores. the last three columns of the left panel in table 5 report the ordered logit regression model results. in this model specification, a positive (negative) coefficient estimate indicates that characteristic increases (decreases) the probability of answering more of the financial literacy questions correctly. consistent with the ols results, we find that financial education increases the probability of answering more financial literacy questions correctly by 142 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 5 financial literacy ols and ordered logit regressions financial literacy score: 0 to 5 financial literacy score: 0 to 5 ols coefficient t stat o-logit coefficient z-stat odds ratio ols coefficient t stat o-logit coefficient z-stat odds ratio constant 1.639*** 27.63 3.986*** 42.39 53.833*** 1.667*** 28.16 3.952*** 42.08 52.044*** fa user 0.202*** 11.88 0.289*** 11.26 1.335*** debt counseling �0.277*** �9.38 �0.432*** �9.67 0.649*** savings or investments 0.154*** 7.30 0.247*** 7.73 1.280*** mortgage or loan 0.109*** 5.05 0.150*** 4.59 1.161*** insurance 0.063*** 3.12 0.075** 2.50 1.078** tax planning �0.016 �0.68 �0.023 �0.64 0.978 financial education 0.265*** 13.47 0.414*** 13.76 1.512*** 0.271*** 13.75 0.425*** 14.09 1.529*** female �0.455*** �26.83 �0.728*** �27.89 0.483*** �0.451*** �26.59 �0.721*** �27.62 0.486*** age group (reference category: 18–24) 25–34 0.189*** 5.31 0.244*** 4.56 1.276*** 0.198*** 5.56 0.258*** 4.83 1.294*** 35–44 0.445*** 12.21 0.623*** 11.32 1.865*** 0.455*** 12.48 0.642*** 11.65 1.901*** 45–54 0.637*** 17.87 0.897*** 16.63 2.452*** 0.640*** 17.95 0.906*** 16.79 2.475*** 55–64 0.751*** 20.24 1.079*** 19.18 2.943*** 0.754*** 20.31 1.085*** 19.25 2.958*** 65� 0.881*** 20.67 1.289*** 19.86 3.630*** 0.882*** 20.70 1.291*** 19.86 3.637*** black (non-hispanic) �0.294*** �15.16 �0.440*** �14.99 0.644*** �0.284*** �14.63 �0.427*** �14.53 0.652*** education (reference category: did not complete high school) high school 0.347*** 9.45 0.466*** 8.45 1.594*** 0.346*** 9.44 0.469*** 8.49 1.598*** ged 0.265*** 5.90 0.380*** 5.64 1.462*** 0.265*** 5.91 0.386*** 5.73 1.471*** some college 0.714*** 20.00 0.990*** 18.35 2.691*** 0.716*** 20.06 0.995*** 18.43 2.705*** college graduate 0.963*** 25.05 1.403*** 24.00 4.067*** 0.966*** 25.14 1.410*** 24.09 4.098*** post-graduate 1.111*** 26.32 1.700*** 26.29 5.474*** 1.108*** 26.26 1.696*** 26.18 5.450*** marital status (reference category: married) living-with-partner �0.072** �2.26 �0.113** �2.40 0.893** �0.077** �2.42 �0.122*** �2.60 0.885*** single �0.016 �0.78 �0.005 �0.15 0.995 �0.019 �0.94 �0.011 �0.34 0.989 income (reference category: less than $15,000) $15,000 $24,999 0.135*** 4.00 0.187*** 3.72 1.206*** 0.154*** 4.58 0.214*** 4.26 1.239*** $25,000 $34,999 0.231*** 6.62 0.342*** 6.49 1.407*** 0.243*** 6.98 0.359*** 6.81 1.431*** $35,000 $49,999 0.368*** 10.81 0.513*** 10.01 1.670*** 0.385*** 11.33 0.537*** 10.48 1.711*** $50,000 $74,999 0.531*** 15.60 0.742*** 14.43 2.100*** 0.536*** 15.76 0.748*** 14.55 2.114*** $75,000 $99,999 0.644*** 16.96 0.941*** 16.25 2.561*** 0.640*** 16.84 0.933*** 16.06 2.541*** $100,000 $149,999 0.755*** 19.03 1.146*** 18.88 3.144*** 0.744*** 18.69 1.126*** 18.46 3.083*** $150,000 or more 0.835*** 18.55 1.372*** 19.52 3.942*** 0.818*** 18.02 1.341*** 18.93 3.824*** employment (reference category: self-employed) employed full-time �0.070** �2.20 �0.127*** �2.61 0.881*** �0.069** �2.15 �0.125** �2.57 0.882** employed part-time �0.123*** �3.13 �0.204*** �3.41 0.816*** �0.126*** �3.21 �0.209*** �3.50 0.812*** homemaker �0.165*** �4.10 �0.235*** �3.87 0.790*** �0.170*** �4.21 �0.244*** �4.01 0.783*** full-time student 0.157*** 2.91 0.197** 2.40 1.218** 0.154*** 2.85 0.187** 2.29 1.206** disabled �0.219*** �4.68 �0.344*** �4.88 0.709*** �0.214*** �4.58 �0.337*** �4.77 0.714*** unemployed �0.104** �2.48 �0.177*** �2.81 0.838*** �0.104** �2.49 �0.179*** �2.84 0.836*** retired �0.095** �2.53 �0.146** �2.53 0.864** �0.101*** �2.67 �0.154*** �2.67 0.857*** region (reference category: northeast) midwest 0.064** 2.54 0.098** 2.54 1.103** 0.063** 2.48 0.094** 2.44 1.099** south 0.041* 1.72 0.045 1.22 1.046 0.042* 1.77 0.046 1.26 1.047 west 0.167*** 6.66 0.240*** 6.29 1.272*** 0.169*** 6.75 0.243*** 6.37 1.276*** observations 22,218 22,218 22,218 22,218 adjusted r2 0.296 0.297 this table reports ols and ordered logit regression results where the dependent variable, financial literacy score, is set equal to the number of financial literacy questions answered correctly, ranging from 0 to 5. the independent variables include financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region. the independent variable of interest in the left panel, fa user, is set equal to 1 if the person used one or more of the five types of financial advisers, or 0 otherwise. the independent variable of interest in the right panel, the five specific financial adviser types, are set equal to 1 if the person used that financial adviser type, or 0 otherwise. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 143b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 51%. we also find that female is associated with a 52% decline in the probability of answering more financial literacy questions correctly which is consistent with lusardi and mitchell (2008) that find females have lower levels of financial literacy. black (nonhispanic) is associated with a 36% decline in the probability of answering more financial literacy questions correctly. we find a monotonically increasing probability of higher financial literacy scores across higher age groups, education levels, and income levels. for marital status and employment, those “living-with-partner” are 11% less likely to answer more financial literacy questions correctly compared with those that are married, and self-employed individuals are more likely to answer more questions correctly. for region, those living in the midwest and west are 10% and 27% more likely to answer more financial literacy questions correctly, respectively. our main variable of interest, fa user, indicates that those using financial advisers are 33.5% more likely to answer more financial literacy questions correctly compared with those that did not use financial advisers. this estimate is highly statistically significant, and reinforces the ols results that financial advisers positively influence the financial literacy of their clients, even when controlling for other demographic characteristics that are also positively related to higher financial literacy scores. in the right panel of table 5 we re-estimate both the ols and ordered logit models where we replace the fa user independent dummy variable with dummy variables for each financial adviser type, separately, to determine the impact each adviser type has on financial literacy. the results for the demographic characteristic control variables are similar to those reported for the original model specification above. in the first two columns using the ols specification, we find that only savings and investments, mortgage or loan, and insurance advisers are positively related to higher financial literacy scores, but tax planning advisers do not significantly impact financial literacy scores. the ordered logit model reveals that savings and investments increase the probability of obtaining a higher financial literacy score by 28%, while mortgage or loan and insurance advisers increases this probability by 16% and 8%, respectively, and tax planning advisers have no significant impact on financial literacy. we find that financial literacy scores are negatively related to debt counseling advisers, where debt counselors appear to negatively influence the financial literacy of their clients and reduce the probability of a higher financial literacy score by 65%. robb, babiarz, and woodyard (2012), using the 2009 nfcs survey, also find a negative relationship between using debt counselors and financial literacy scores because those individuals appear to make financial mistakes in the first place and do not learn from their experiences. gathergood (2012) suggests debt counseling services involves offering solutions to existing debt problems, such as debt-repayment plans, usually to clients that lack self-control, have lower levels of financial literacy, and are considered poor learners. it appears that the debt counselor results are consistent with this characterization of individuals that use debt counselors. this is also consistent with our univariate results reported in table 1, panel b, where we find that debt counselor users perform very well on the financial literacy question related to the total interest paid over a 15-year mortgage versus a 30-year mortgage, but underperform on the other financial literacy questions. this suggests that debt counselors do not influence the financial literacy of their clients in the same manner that other adviser types 144 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 do because they focus on topics specifically related to interest rates and interest payments related to debt products. in summary, using both ols and ordered logit models, we find that financial advisers, in general, have a positive influence on the financial literacy of the individuals that use them suggesting they play a role in helping disseminate financial knowledge to their clients rather than just offering them advice about what they should do in making financial decisions. when considering the five different types of financial advisers separately, we find that it is savings or investment advisers that have the largest impact on financial literacy, followed by mortgage or loan and insurance advisers, respectively. this is understandable since the five financial literacy questions asked pertain to the time value of money, interest rates, inflation, and risk diversification which are topics that a savings or investment adviser should cover with their client. tax planning advisers, which we find have no influence on financial literacy scores, and debt counselors, which have a negative influence on financial literacy scores, would be less likely to cover these types of topics with their clients. 4. endogeneity and confounding variables there is the possibility that our model suffers from endogeneity issues in the choice to use a financial adviser. it is possible that those having higher levels of financial literacy are more likely to use financial advisers. even though we control for demographic characteristics that we find to be directly related to financial literacy in all of our regression specifications, such as financial education and education, this may not be enough to control for this endogeneity bias. lacking longitudinal data that would directly address this problem, we must try to control for endogeneity using econometric techniques given our data limitations. we address these concerns in this section using an instrumental variables (iv) model and propensity score matching. 4.1. instrumental variables model to address the potential endogeneity issues between financial literacy and the choice to use a financial adviser, we implement a two-stage iv estimation model. we estimate iv models for each of the financial adviser types to include debt counseling, savings or investments, mortgage or loan, insurance, and tax planning using both the gmm estimator and iv probit estimation.2 in the first stage, we use responses to available survey questions that are expected to predict the choice to use a specific type of financial adviser as instruments, but these instruments must be uncorrelated with the second stage dependent variable that is the financial literacy score. the same demographic control variables from table 5 are also used in both the firstand second-stage model specifications. table 6, panel a, identify the instruments (survey questions) used for each of the five financial adviser types. we also report the first-stage coefficient estimates for the instruments used from the gmm estimator and iv probit estimation, separately.3 all instruments used are positively related to using each specific type of financial adviser, and they are all highly 145b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 6 instrumental variable regressions panel a: first stage instruments for each financial adviser type instruments used instrument survey question: instrumented: debt counseling home foreclosure have you been involved in a foreclosure process on your home? declared bankruptcy have you declared bankruptcy in the last two years? instrumented: savings or investments employer retirement plan do you have any retirement plan through a current or previous employer? other retirement plan do you have any other retirement accounts not through an employer? instrumented: mortgage or loan own your home do you currently own your home? home foreclosure have you been involved in a foreclosure process on your home? instrumented: insurance health insurance are you covered by health insurance? life insurance do you have a life insurance policy? instrumented: tax planning earn salary or wages did you receive any income in the form of salary or wages? social security benefits did you receive any income from social security retirement benefits? first stage instruments coefficient estimates iv gmm t stat iv probit z-stat debt counseling: home foreclosure 0.166*** 9.29 0.166*** 15.38 declared bankruptcy 0.381*** 18.87 0.381*** 34.28 savings or investments: employer retirement plan 0.082*** 11.33 0.082*** 11.41 other retirement plan 0.285*** 34.95 0.285*** 39.75 mortgage or loan: own your home 0.119*** 18.81 0.119*** 17.97 home foreclosure 0.161*** 9.25 0.161*** 10.65 insurance: health insurance 0.026*** 2.98 0.026*** 2.82 life insurance 0.093*** 12.51 0.093*** 12.55 tax planning: earns or wages 0.035*** 5.28 0.035*** 5.18 social security benefits 0.099*** 10.13 0.099*** 11.12 panel b: second stage dependent variable: financial literacy score (0 to 5) financial literacy score: 0 to 5 financial literacy score: 0 to 5 iv gmm t stat iv probit z-stat iv gmm t stat iv probit z-stat constant 1.756*** 27.24 0.677*** 7.02 1.587*** 23.04 0.531*** 4.99 debt counseling �0.870*** �7.57 �0.071 �0.39 savings or investments 1.174*** 16.10 1.248*** 7.69 financial education 0.313*** 16.15 0.366*** 7.86 0.189*** 8.67 0.303*** 5.81 female �0.445*** �25.72 �0.212*** �6.40 �0.456*** �24.67 �0.238*** �6.66 age group (reference category: 18–24) 25–34 0.221*** 5.71 0.203*** 3.76 0.195*** 4.58 0.173*** 2.89 35–44 0.446*** 11.20 0.235*** 4.17 0.508*** 11.56 0.308*** 4.75 45–54 0.603*** 15.64 0.394*** 7.00 0.656*** 15.44 0.466*** 7.26 55–64 0.725*** 18.25 0.505*** 8.16 0.714*** 16.48 0.490*** 7.20 65� 0.850*** 19.10 0.590*** 7.51 0.785*** 16.36 0.523*** 6.17 black (non-hispanic) �0.269*** �13.02 �0.120*** �3.40 �0.278*** �12.69 �0.120*** �3.18 146 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 6 (continued) financial literacy score: 0 to 5 financial literacy score: 0 to 5 iv gmm t stat iv probit z-stat iv gmm t stat iv probit z-stat education (reference category: did not complete high school) high school 0.354*** 8.81 0.318*** 6.32 0.338*** 8.15 0.315*** 5.74 ged 0.273*** 5.57 0.227*** 3.64 0.289*** 5.72 0.283*** 4.16 some college 0.749*** 19.25 0.603*** 11.72 0.670*** 16.53 0.540*** 9.59 college graduate 1.019*** 24.57 0.722*** 11.68 0.850*** 19.22 0.590*** 8.52 post-graduate 1.151*** 26.39 0.687*** 8.86 0.976*** 20.47 0.555*** 6.34 marital status (reference category: married) living-with-partner �0.092*** �2.91 �0.032 �0.58 �0.084** �2.43 �0.036 �0.60 single �0.039* �1.82 �0.119*** �3.20 �0.031 �1.39 �0.126*** �3.15 income (reference category: less than $15,000) $15,000 $24,999 0.185*** 5.04 0.153*** 3.02 0.120*** 3.16 0.118** 2.18 $25,000 $34,999 0.277*** 7.16 0.118** 2.22 0.162*** 4.00 0.034 0.58 $35,000 $49,999 0.435*** 11.73 0.262*** 4.73 0.283*** 7.21 0.129** 2.10 $50,000 $74,999 0.594*** 16.35 0.450*** 7.59 0.372*** 9.10 0.251*** 3.59 $75,000 $99,999 0.705*** 17.59 0.362*** 5.27 0.420*** 8.96 0.107 1.27 $100,000 $149,999 0.810*** 19.98 0.579*** 6.97 0.475*** 9.55 0.245** 2.40 $150,000 or more 0.899*** 19.86 0.580*** 5.60 0.506*** 8.97 0.197 1.56 employment (reference category: self-employed) employed full-time �0.068** �2.12 0.029 0.45 �0.033 �0.96 0.081 1.16 employed part-time �0.129*** �3.18 �0.013 �0.18 �0.132*** �3.06 �0.008 �0.10 homemaker �0.205*** �4.89 �0.043 �0.58 �0.074* �1.66 0.076 0.95 full-time student 0.135** 2.34 0.178* 1.88 0.149** 2.39 0.181* 1.76 disabled �0.232*** �4.79 �0.074 �0.91 �0.140*** �2.76 0.005 0.06 unemployed �0.137*** �3.17 �0.020 �0.27 �0.027 �0.58 0.083 1.04 retired �0.119*** �3.21 �0.092 �1.15 �0.099** �2.52 �0.110 �1.29 region (reference category: northeast) midwest 0.054** 2.07 0.050 1.06 0.067** 2.39 0.089* 1.76 south 0.039 1.59 0.085* 1.91 0.049* 1.85 0.110** 2.31 west 0.172*** 6.71 0.144*** 2.99 0.167*** 6.08 0.167*** 3.24 observations 22,069 22,069 21,095 21,095 adjusted r2 0.275 0.199 financial literacy score: 0 to 5 financial literacy score: 0 to 5 iv gmm t stat iv probit z-stat iv gmm t stat iv probit z-stat constant 1.669*** 24.94 0.552*** 5.31 1.657*** 21.98 0.405*** 3.24 mortgage or loan 0.410*** 2.61 0.938*** 3.59 insurance 0.306 1.58 1.383*** 3.85 financial education 0.260*** 12.28 0.293*** 5.99 0.249*** 9.27 0.206*** 3.45 female �0.456*** �26.50 �0.225*** �6.71 �0.459*** �25.47 �0.249*** �6.94 age group (reference category: 18–24) 25–34 0.183*** 4.72 0.172*** 3.13 0.172*** 4.33 0.183*** 3.18 35–44 0.438*** 11.06 0.252*** 4.40 0.428*** 10.52 0.249*** 4.13 45–54 0.641*** 16.26 0.460*** 7.70 0.616*** 15.42 0.441*** 7.21 55–64 0.768*** 18.72 0.581*** 8.81 0.743*** 18.10 0.545*** 8.20 65� 0.911*** 19.75 0.676*** 8.19 0.884*** 19.25 0.653*** 7.78 black (non-hispanic) �0.277*** �13.37 �0.088** �2.44 �0.298*** �14.49 �0.132*** �3.57 education (reference category: did not complete high school) high school 0.339*** 8.56 0.297*** 5.71 0.356*** 8.91 0.304*** 5.62 ged 0.250*** 5.17 0.204*** 3.20 0.262*** 5.36 0.209*** 3.12 some college 0.708*** 17.92 0.535*** 9.74 0.717*** 17.49 0.502*** 8.26 college graduate 0.959*** 22.47 0.641*** 9.69 0.972*** 21.93 0.619*** 8.50 post-graduate 1.119*** 24.89 0.605*** 7.42 1.121*** 23.96 0.574*** 6.63 marital status (reference category: married) living-with-partner �0.075** �2.38 0.001 0.01 �0.068** �2.15 0.006 0.10 single �0.012 �0.55 �0.063 �1.58 �0.012 �0.55 �0.053 �1.30 income (reference category: less than $15,000) $15,000 $24,999 0.140*** 3.89 0.134*** 2.63 0.131*** 3.28 0.054 0.89 $25,000 $34,999 0.226*** 5.82 0.071 1.28 0.237*** 5.56 0.014 0.22 $35,000 $49,999 0.369*** 9.81 0.204*** 3.50 0.379*** 8.62 0.115 1.58 $50,000 $74,999 0.524*** 13.09 0.346*** 5.24 0.536*** 11.27 0.253*** 3.06 147b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 table 6 (continued) financial literacy score: 0 to 5 financial literacy score: 0 to 5 iv gmm t stat iv probit z-stat iv gmm t stat iv probit z-stat $75,000 $99,999 0.626*** 13.58 0.235*** 2.97 0.659*** 12.89 0.192** 2.09 $100,000 $149,999 0.739*** 15.28 0.427*** 4.51 0.767*** 14.04 0.356*** 3.27 $150,000 or more 0.827*** 14.97 0.401*** 3.41 0.842*** 13.57 0.268** 2.09 employment (reference category: self employed) employed full-time �0.081** �2.53 0.018 0.28 �0.067** �1.97 0.077 1.07 employed part-time �0.120*** �2.99 0.012 0.16 �0.124*** �2.98 0.028 0.35 homemaker �0.168*** �3.99 �0.000 �0.00 �0.152*** �3.26 0.088 1.01 full-time student 0.168*** 2.86 0.224** 2.30 0.160** 2.47 0.265** 2.41 disabled �0.218*** �4.51 �0.042 �0.50 �0.211*** �4.27 �0.007 �0.08 unemployed �0.113*** �2.60 0.032 0.42 �0.093* �1.92 0.134 1.52 retired �0.087** �2.34 �0.062 �0.77 �0.089** �2.27 �0.034 �0.40 region (reference category: northeast) midwest 0.063** 2.44 0.051 1.06 0.056** 2.10 0.026 0.51 south 0.038 1.56 0.081* 1.80 0.035 1.44 0.083* 1.80 west 0.155*** 5.92 0.109** 2.19 0.162*** 6.17 0.111** 2.17 observations 22,020 22,020 21,645 21,645 adjusted r2 0.286 0.288 financial literacy score: 0 to 5 iv gmm t stat iv probit z-stat constant 1.753*** 22.38 0.595*** 4.75 tax planning �0.162 �0.59 0.591 1.24 financial education 0.292*** 11.44 0.319*** 5.63 female �0.455*** �26.52 �0.219*** �6.46 age group (reference category: 18–24) 25–34 0.166*** 4.24 0.194*** 3.45 35–44 0.405*** 9.23 0.250*** 3.75 45–54 0.583*** 13.51 0.429*** 6.43 55–64 0.704*** 16.22 0.524*** 7.48 65� 0.842*** 18.68 0.609*** 7.44 black (non-hispanic) �0.294*** �14.55 �0.122*** �3.42 education (reference category: did not complete high school) high school 0.364*** 9.14 0.321*** 6.23 ged 0.270*** 5.54 0.240*** 3.76 some college 0.742*** 18.74 0.576*** 10.52 college graduate 1.012*** 22.58 0.704*** 9.88 post-graduate 1.174*** 22.39 0.657*** 6.85 marital status (reference category: married) living-with-partner �0.087*** �2.66 0.004 0.06 single �0.027 �1.26 �0.098** �2.52 income (reference category: less than $15,000) $15,000 $24,999 0.161*** 4.37 0.132** 2.52 $25,000 $34,999 0.285*** 6.97 0.092 1.54 $35,000 $49,999 0.437*** 10.29 0.215*** 3.16 $50,000 $74,999 0.611*** 12.60 0.377*** 4.53 $75,000 $99,999 0.751*** 12.52 0.279*** 2.66 $100,000 $149,999 0.868*** 12.09 0.438*** 3.27 $150,000 or more 0.973*** 10.16 0.394** 2.16 employment (reference category: self-employed) employed full-time �0.098** �2.36 0.065 0.81 employed part-time �0.146*** �3.45 �0.002 �0.03 homemaker �0.199*** �4.18 �0.005 �0.06 148 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 statistically significant at the 1% level. this indicates they are legitimate instruments to use in predicting the probability of using each type of financial adviser. table 6, panel b, report results for the second-stage estimates for each model where financial literacy score is the dependent variable, and the predicted financial adviser type user from the first stage model and the demographic characteristics are used as independent variables. using the gmm estimator we find that debt counseling remains negatively related to financial literacy scores, however, the iv probit estimates indicate that debt counseling is not significantly related to financial literacy scores. savings or investments and mortgage or loan remains positively related to higher financial literacy scores using both the gmm estimator and iv probit estimates. insurance advisers remain positively related with higher financial literacy scores only when using the iv probit estimates, and not significant when using the gmm estimator. tax planning remains not significantly related to financial literacy scores using both the gmm estimator and iv probit estimates. overall, our results in table 6 are similar to those reported in table 5 with the exception that debt counseling is not significant when using the iv probit estimate, and insurance is not significant when using the iv gmm estimator. savings or investments and mortgage or loan advisers appear to have the strongest impact on the financial literacy of their clients, even when controlling for demographic characteristics that are shown to impact financial table 6 (continued) financial literacy score: 0 to 5 iv gmm t stat iv probit z-stat full-time student 0.116* 1.76 0.209* 1.86 disabled �0.236*** �4.24 �0.015 �0.15 unemployed �0.131*** �2.69 0.025 0.29 retired �0.107*** �2.70 �0.089 �1.06 region (reference category: northeast) midwest 0.060** 2.33 0.057 1.19 south 0.036 1.49 0.089** 1.98 west 0.173*** 6.77 0.150*** 3.06 observations 21,911 21,911 adjusted r2 0.287 this table reports instrumental variable (iv) model results where the first stage dependent variable, debt counseling, savings or investments, mortgage or loan, insurance, or tax planning is set equal to 1 if the person used that type of adviser, or 0 otherwise. the independent variables used in the first stage model include instruments for each financial adviser type and the demographic control variables financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region. panel a report the instruments used for each financial adviser type and iv gmm and iv probit coefficient estimates and t stats (z-stats) for the instruments from the first stage model results, respectively. the second stage model dependent variable is the financial literacy score that is set equal to the number of financial literacy questions answered correctly, ranging from 0 to 5. the independent variables used in the second stage model include the predicted financial adviser user variable from the first stage model and the same demographic control variables used in the first stage model. panel b report the iv gmm and iv probit coefficient estimates and t stats (z-stats) for the second stage model results for each financial adviser type separately. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively. 149b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 literacy and when controlling for potential endogeneity issues between financial literacy and the choice to use a financial adviser. 4.2. propensity score matching in an attempt to reduce any confounding variable biases in our results we implement a propensity score matching model to estimate the impact financial advisers have on the financial literacy of their clients. we first estimate the same logit regression used in table 3 where fa user, the dependent variable, is regressed on to the different demographic characteristic control variables to include financial education, female, age group, black (non-hispanic), education, marital status, income, employment, and region. using the propensity scores from this logit estimation we perform a nearest neighbor match that ensures that a financial adviser user is paired with a non-user with statistically the same demographic characteristics. we conduct the matching with replacement to achieve the best match for each financial adviser user with a non-user. we separately conduct the matching without replacement to ensure we have a 1:1 match for each financial adviser user with a non-user that maximizes our matched sample size. we then use these matched samples to estimate an ols univariate regression model where the dependent variable is financial literacy score and the independent variable is fa user. we repeat this process using the logit model from table 4 for each of the financial adviser types separately. table 7, panel a, report results using propensity score matching with replacement. consistent with earlier results, we find there is a positive and statistically significant relationship between fa user and financial literacy scores. when considering the financial adviser types separately, we find that debt counseling is negatively related to financial literacy score, while savings or investments, mortgage or loan, insurance, and tax planning are all positively related to financial literacy scores. table 7, panel b, report results using propensity score matching without replacement. we again find that there is a positive and statistically significant relationship between fa user and financial literacy scores. when considering the financial adviser types separately, we continue to find that debt counseling is negatively related to financial literacy score, while savings or investments, mortgage or loan, and insurance are all positively related to financial literacy scores. however, we find that insurance is not significantly related to financial literacy when using matching without replacement. these results continue to indicate that savings or investments advisers have the strongest influence on the financial literacy of their clients, followed by mortgage or loan and insurance advisers. debt counselors is consistently shown to be negatively related to financial literacy, and tax planning advisers are only marginally, or not at all, related to the financial literacy of their clients. 5. conclusions using the 2012 finra investor education foundation national financial capability study, we determine the demographic characteristics associated with people that use finan150 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 cial advisers, and the impact financial advisers have on the financial literacy of their clients. we consider five different types of financial advisers: debt counselors, savings or investments, mortgage or loan, insurance, and tax planning. we first find that there has been a significant increase over the last decade in the number of individuals who use financial advisers. elmerick, montalto, and fox (2002), using the 1998 survey of consumer finances that is a similar survey to the 2012 finra survey we are using, reported that only 21% of those surveyed had used a financial adviser. using the more recent 2012 finra survey, we find that over 53% of those surveyed reported using a financial adviser indicating the usage of financial advisers has more than doubled since the 1998 survey was conducted. when considering the demographic characteristics associated with those using a financial adviser, we find that females are 13% more likely, and those that have received a financial table 7 propensity score matching ols univariate regression: financial literacy score (0 to 5) � adviser user dummy fa user debt counseling savings or investments mortgage or loan insurance tax planning panel a: propensity score matching with replacement number of users 11,872 1,949 7,025 4,914 7,312 4,422 matches 4,494 1,660 3,692 3,151 4,142 2,872 intercept 2.95*** 3.06*** 3.20*** 3.21*** 3.14*** 3.31*** (145.09) (91.52) (146.47) (134.12) (148.88) (132.08) coefficient 0.41*** �0.20*** 0.32*** 0.21*** 0.19*** 0.17*** (17.14) (�4.47) (11.79) (6.95) (7.27) (5.37) panel b: propensity score matching without replacement number of users 10,346 1,949 7,025 4,914 7,312 4,422 matches 10,346 1,949 7,025 4,914 7,312 4,422 intercept 2.77*** 3.06*** 3.31*** 3.29*** 3.25*** 3.45*** (201.72) (99.02) (209.72) (171.76) (204.81) (171.76) coefficient 0.60*** �0.20*** 0.21*** 0.13*** 0.08*** 0.04 (30.69) (�4.65) (9.28) (4.68) (3.75) (1.25) this table reports propensity score matching results where we first estimate the logit regression in table 3 to match each financial adviser user with a non-user having statistically the same financial education, gender, age group, ethnicity, education, marital status, income, employment, and region. we then estimate a univariate regression using this matched sample where the dependent variable is the financial literacy score that is set equal to the number of financial literacy questions answered correctly from 0 to 5, and the independent variable is the fa user dummy that is set equal to 1 if the person used one or more of the five types of financial advisers, or 0 otherwise. we repeat this process for each financial adviser type separately to include debt counseling, savings or investments, mortgage or loan, insurance, and tax planning using the logit regressions from table 4 to determine the propensity score matches. the dummy variables debt counseling, savings or investments, mortgage or loan, insurance, and tax planning are set equal to 1 if the person used that type of financial adviser, or 0 otherwise. panel a report results where propensity score matches were created with replacement in order to determine the best match for each observation. panel b report results using matching without replacement in order to increase the sample size used in the univariate regressions. ***, **, and * denote significance at the 1%, 5%, and 10% levels, respectively, with t stats reported in parenthesis. 151b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 education in the past nearly 50% more likely, to use financial advisers. those that are married, or self-employed, are also more likely to use financial advisers. we find that there is a monotonically increasing probability of using financial advisers across higher education and income levels, but younger age groups are less likely to use financial advisers. surprisingly, ethnicity, after controlling for all other demographic characteristics, does not appear to significantly impact the likelihood of using a financial adviser. finally, we find that respondents from the midwest and west are more likely to use a financial adviser than respondents from the south or northwest. when considering the demographic characteristics of those that used a specific type of financial adviser we find that those receiving a financial education in the past are more likely to use all of the five financial adviser types compared with those that did not. females are more likely to use insurance advisers compared with males, but not more likely to use any of the other four adviser types. black (non-hispanic) people are more likely to use debt counselors but less likely to use mortgage or loan advisers. we find that older people are more likely to use savings and investments advisers but younger people are more likely to use debt counselors and mortgage or loan advisers. those having a high school education or higher are more likely to use savings or investment advisers, but only those having some college or higher are more likely to use insurance or tax planning advisers. people with higher income levels are less likely to use debt counselors, but more likely to use all of the other adviser types. those with lower income levels are more likely to use debt counselors. married couples are more likely to use mortgage or loan, insurance, and tax planning advisers, and those that are self-employed are more likely to use all of the five financial adviser types. finally, we find that those living in the west are more likely to use mortgage or loan advisers, and those in the midwest and west are more likely to use insurance advisers. finally, after controlling for demographic characteristics that are shown to impact financial literacy, we continue to find a positive and statistically significant relationship between financial literacy scores and the use of financial advisers suggesting they have a positive influence on the financial literacy of their clients. when considering the different financial adviser types separately we find that savings or investments, mortgage loan, and insurance advisers have the highest positive influence on the financial literacy of their clients. however, debt counselors appear to have a negative influence on the financial literacy of their clients which is consistent with robb, babiarz, and woodyard (2012) suggesting those individuals appear to make financial mistakes in the first place and do not appear to learn from their experiences. we find that tax planning advisers do not significantly impact the financial literacy scores of their clients. the relationship between the choice to use a financial adviser and financial literacy may suffer from endogeneity issues and confounding variable biases in the models we use, so we attempt to correct these biases using instrument variables and propensity score matching methodology. our results remain robust indicating that savings or investment, mortgage or loan, and insurance advisers have a positive influence on the financial literacy of their clients. even though it is difficult to measure the impact financial advisers have on the financial literacy of their clients with precision, we have established that a positive relationship exists between financial advisers and financial literacy after controlling for demographic characteristics that are shown to impact financial literacy. this relationship also remains robust 152 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 when controlling for the possibility of endogeneity and confounding variable problems that may exist in a model relating financial adviser use with financial literacy scores. the positive relationship we find that exists between financial adviser use and financial literacy scores suggest that clients appear to learn from their interactions with financial advisers. these results show that financial advisers do not just offer financial advice, but also play a significant role in disseminating financial knowledge to their clients that has a positive impact on financial literacy. future research using longitudinal data that measures financial literacy before and after using a financial adviser, or that is able to track the financial adviser-client relationship through time, would be able to measure the impact financial advisers have on the financial literacy of their clients with more precision. notes 1 debt counseling refers to the services provided to individuals who have problems repaying their debt on time. for further details, please see: http://www.consumer finance.gov/askcfpb/1449/whats-difference-between-credit-counselor-and-debtsettlement-company.html. 2 we also estimated the iv model using two-stage least squares (2sls) with results nearly identical to the reported gmm results. 3 coefficient estimates for the demographic control variables are nearly identical to what is reported in table 4. acknowledgment research funded by the university of akron-finance advisory board 2014 economic summit awards program grant. appendix a: financial literacy questions below are the five financial literacy questions asked in the survey that is used to measure the financial literacy of the survey respondent: 1. suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow? a. more than $102 b. exactly $102 c. less than $102 d. don’t know e. refused 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account? 153b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 a. more than today b. exactly the same c. less than today d. don’t know e. refused 3. if interest rates rise, what will typically happen to bond prices? a. they will rise b. they will fall c. they will stay the same d. there is no relationship between bond prices and the interest e. don’t know f. refused 4. a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. a. true b. false c. don’t know d. refused 5. buying a single company’s stock usually provides a safer return than a stock mutual fund. a. true b. false c. don’t know d. refused references allgood, s., & walstad, w. b. (2016). the effects of perceived and actual financial literacy on financial behaviors. economic inquiry, 54, 675–697. bandura, a. (1977). social learning theory. englewood cliffs, nj: prentice-hall. barber, b. m., & odean, t. (2001). boys will be boys: gender, overconfidence, and common stock investment. the quarterly journal of economics, 116, 261–292. black, k. jr., ciccotello, c. s., & skipper, h. d., jr. (2002). issues in comprehensive financial planning. financial services review, 11, 1–9. bucher-koenen, t., & lusardi a. (2011). financial literacy and retirement planning in germany. journal of pension economics and finance, 10, 565–584. calvet, l. e., campbell, j. y., & sodini, p. (2009). measuring the financial sophistication of households. american economic review, 99, 393–398. campbell, j. y. (2006). household finance. journal of finance, 59, 1553–1604. elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u. s. households. financial services review, 11, 217–231. 154 b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 gathergood, j. (2012). self-control, financial literacy, and over-indebtedness. journal of economic psychology, 33, 590–602. grimes, p. w., rogers, k. e., & smith, r. c. (2010). high school economic education and access to financial services. journal of consumer affairs, 44, 317–335. huston, s. j. (2012). financial literacy and the cost of borrowing. international journal of consumer studies, 36, 566–572. johnson, e., & sherraden, m. (2007). from financial literacy to financial capability among youth. journal of social welfare, 34, 119–146. klapper, l., lusardi, a., & panos, g. a. (2013). financial literacy and its consequences: evidence from russia during financial crisis. journal of banking & finance, 37, 3904–3923. kolb, d. a. (1984). experiential learning: experience as the source of learning and development. englewood cliffs, nj: prentice-hall, inc. lachance, m. (2014). financial literacy and neighborhood effects. journal of consumer affairs, 48, 251–273. lusardi, a., & mitchell, o. s. (2007a). baby boomers retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. s. (2007b). financial literacy and retirement preparedness: evidence and implication for financial education. business economics, 42, 35–44. lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review: papers and proceedings, 98, 413–417. lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and overindebtedness. nber working paper 14808, national bureau of economic research, inc. moore, d. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes and experiences. technical report 03–39. social and economic sciences research center, washington state university. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professions’ advice: the role of financial knowledge, satisfaction, and confidence, financial services review, 21, 291–305. tang, n., & peter, p. c. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience. financial services review, 24, 119–137. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. walstad, w. b., rebeck, k., & macdonald, r. a. (2010). the effects of financial education on the financial knowledge of high school students. journal of consumer affairs, 44, 336–357. winchester, d. d., & huston, s. j. (2014). does a relationship with financial service professional overcome a client’s sense of not being in control of achieving their goals? financial services review, 23, 1–23. 155b. balasubramnian, e.r. brisker / financial services review 25 (2016) 127–155 the effects of health, family, and altruism on retirement savings brian t. starra,* adepartment of economics, lubbock christian university, 5601 19th street, lubbock, tx 79407, usa abstract using data from the panel study of income dynamics, this paper broadens the analysis of retirement savings by examining the effects of health, children, altruism, and family of origin attributes on the decision of whether to save for retirement and on how much retirement savings are accumulated. the presence of children in the household generally reduces the probability of saving for retirement. poor personal or parental health diminishes retirement savings outcomes. altruistic behavior generally presents as complementary to retirement savings, and the evidence suggests children of mothers who saved for retirement are more likely to do the same. © 2022 academy of financial services. all rights reserved. jel classifications: c72; g51; g53 keywords: retirement savings; family of origin; health; philanthropy 1. introduction standard economic theory promulgates a (typically) representative agent who emerges into life, introspectively surveys his or her preferences and tolerances, estimates life expectancy and lifetime income patterns, forecasts a real rate of return on savings, sets up a hamiltonian to solve for the optimal time path of consumption and then saves and spends accordingly. of course, these assumptions are necessary to reduce enormously complex decisions into mathematically tractable problems so that economic life can be modeled and studied. but no economist believes that a person enters the world prepared to optimally *corresponding author: tel.: +1-405-425-5000, fax: +1-405-425-5090. e-mail address: brian.starr@oc.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 251–271 make these decisions. rather, consumption smoothing is a learned behavior, formed and shaped by the people and circumstances surrounding us. given the observed lack of savings altogether by some agents, one must wonder if some people ever learn this behavior at all. this paper seeks to examine some of those forces that shape our retirement savings behavior. specifically, we will examine the effects that health, progeny, altruism, and learning from one’s family of origin might have on retirement savings decisions. data for our consideration will be taken from the panel study of income dynamics (psid). while this data set is not as specific to retirement savings as another survey, it allows us to observe in a broader sphere of circumspection the variables surrounding the retirement decision and has been used at times to study the economics of retirement. see, for example, bernheim et al. (2001). 2. literature review a rich body of literature covers a broad range of retirement savings topics, several relevant to our tasks. we start by recalling that the maturation of the defined contribution plan has significantly altered the savings landscape. while our standard consumption smoothing models do not differentiate savings accumulated in a retirement plan from savings accumulated elsewhere, there is little empirical doubt that the increasing replacement of the defined benefit pension plan by the defined contribution plan is affecting the way american laborers save for retirement. wise (2007) noted that 401(k) plans had not existed for a full working career of those then retiring and the full impact of this change could not yet be anticipated. indeed, those who were joining the workforce in their early 20s when congress created section 401(k) through the revenue act of 1978 are just now entering the traditional retirement years. still, studies have emerged that shed some light on where we are and where we seem to be headed. whether the existence of iras and 401(k)’s has increased the overall level of savings has been significantly debated with some concluding they have (poterba et al., 1996) and others finding evidence to the contrary (engen et al., 1996). the question put on a finer point would be how these defined contribution structures anticipated outcomes compare with those expected to obtain in a defined benefit plan. some evidence suggests that the defined contribution outcome is likely to be preferred in all but the lowest decile of earnings. the reason for this outcome lies in the fact that stock market returns, so critical to the defined contribution outcome, are largely uncorrelated to any given worker’s unique circumstances so a worker is likely to experience highs that counterbalance the lows during the working career. conversely, define benefit plan outcomes are more determined by one’s level of income at career end, a metric that is subject to substantial risk. thus, conversely, to some popular notions, the study suggests that the defined contribution outcome is generally preferable to the defined benefit outcome (samwick & skinner, 2004). within the 401(k) realm, we observe the positive correlation of participation and savings with income, age, education, and job tenure (munnell et al., 2000). these outcomes are intuitive as we would expect the first variable to be positively correlated with the others and 252 b. t. starr / financial services review 30 (2022) 251–271 income itself to be positively correlated with participation and savings. a caveat is in order, though. the literature indicates that a significant number of workers at the lowest decile of pay simply do not save at all, running counter to our standard economic predictions. why would a person even on a very meager wage not opt to smooth consumption by saving at least a small portion of their income? anecdotal evidence suggests a variety of reasons, including the notion that they feel too poor to save, they plan to work until they die, or they expect support from other sources. the latter idea will be briefly explored later in the essay. research has also shown that plan design plays an important role in retirement saving and investing outcomes among plan participants, though the nature of the specific conclusions must be carefully unpacked. according to one study, the presence of a matching contribution seems a crucial factor influencing the contribution decision, but the amount of the matching contribution is statistically insignificant (munnell et al., 2000). the same study summarizes other research which demonstrates that a very high rate of the match is negatively correlated to the amount a participant saves in the plan, perhaps suggesting that the income effect dominates the substitution effect. in essence, the participants are acting as if they have pre-determined what overall percentage of income they wish to have contributed to the plan and a generous employer match crowds out their own contributions. the notion that the presence of a matching contribution has a significant effect on plan participation is not without debate. choi et al. find that, after controlling for the liquidity and investment constraints embedded within the plan designs, the presence of a match only enticed 10% of the eligible employees to contribute to the plan. workers effectively forfeit 50% of the matching contribution that is available to them by failing to contribute up to the match rate. the study suggests that a matching contribution to incent participation is a relatively weak instrument. indeed, it sometimes appears as if “at any point in time employees are likely to do whatever requires the least current effort” (choi et al., 2001), deeply implicating plan design decisions (choi et al., 2004). stronger measures such as an automatic enrollment (mitchell et al., 2007) or fresh start nudge (beshears et al., 2021) seem more critical to obtaining optimal savings outcomes. such behavioral inertia toward an intuitive path of least resistance might extend to the investment decision as well (brown et al., 2007). thus, it is no surprise that strong evidence has emerged that automatic enrollment, a levering of behavioral inertia, significantly affects savings outcomes in 401(k) plans. before the advent of this technique, the default election of each participant was deemed to be a deferral of nothing into the plan until the plan-eligible participant signed a deferral agreement electing some positive amount to be saved. automatic enrollment implicitly assumes that such an option is not welfare-maximizing and makes the default savings percentage some positive number (say 3%). thus, the election to defer to the plan becomes essentially a negative election. only by making an affirmative election to the contrary can the participant move to another rate of deferral, including the formerly ubiquitous 0% election. early research demonstrated an escalation in participation of 48 percentage points for new hires under automatic enrollment (madrian & shea, 2000). follow-up research reiterated the drastic increase in participation and revealed that between 65% and 87% of new plan participants “elected” the default rate defined under the plan (choi et al., 2001). even after controlling for matching contributions, the efficacy of automatic enrollment appears robust (beshears et al., 2007). enrollment processes that seek to simplify the contribution and investment election b. t. starr / financial services review 30 (2022) 251–271 253 decision-making process have been shown to increase participation (choi et al., 2006) though not to the same degree as through the mechanism of automatic enrollment (beshears et al., 2007), and the extent to which simplification is effective in improving outcomes is still up for debate (cardella et al., 2021). these findings are telling and somewhat discombobulating to standard economic theory. before this research, economists could simply argue that many non-savers had very large discount rates on their future utilities and affirmatively elected not to save, even in full knowledge of the future implications of such non-savings. a more moderate view would have held that utility discounting is dynamic and that participants would later regret their earlier non-savings decision. however, in either version of the narrative, participants were held as rational thinkers. research on automatic enrollment introduces the notion that participants might not be thinking at all, begging the question of the importance of financial literacy in retirement savings outcomes. a large body of literature examines the impact of financial literacy on financial decision making, with a subset of that literature exploring how we learn to save and invest. not surprisingly, some of that learning emerges from our own experiences (choi et al., 2009) and education, though not necessarily in financial literacy per se, plays a key role in determining financial outcomes (cole et al., 2014). some data suggests that children learn general savings behavior from their parents, particularly when the children have received little financial training elsewhere (tang & peter, 2015), though individual behavioral traits of the children seem to ultimately prove more critical than parental training (barboza et al., 2021). a probit analysis of intergenerational asset holdings finds a significant correlation between bank account and stock holdings between parents and children and that the correlation is more than that which can be explained by intergenerational transfers (chiteji & stafford, 2000). using the same panel study of income dynamics (psid) data, charles and hurst find persistence in intergenerational wealth that also transcends mere income and education measures. parents and children tend to hold similar assets, even after controlling for risk tolerance variables. they suggest that the explanation lies in the idea of children learning about financial market participation from their parents (charles & hurst, 2002). while we tend to perceive the flow of knowledge, behavioral traits, and economic outcomes like income volatility (shore, 2011) as passing directionally from parents to children, the advent of children in a household indubitably changes that household’s economic behavior. for example, love (2010) finds the number of children living in a household reduces the share of risky assets in the household’s portfolio. we will attempt to observe familial effects on retirement savings from both directions—the modeled behavior of the saver’s parents and the presence of children in the saver’s household. we will also add variables representing charitable giving and health status to test the potential effects of altruism and health on retirement savings decisions. 2.1. psid data our dataset is drawn from the university of michigan’s panel study of income dynamics. periodically, this study includes a special module to gather general wealth 254 b. t. starr / financial services review 30 (2022) 251–271 information, including retirement savings data. this paper analyzes data from the 1989 and 2019 waves that have some detailed pension-related data points associated. the purpose of using psid data instead of a dataset more keenly focused on retirement savings is to enable an analysis of factors beyond those typically encompassed in surveys specific to retirement and to analyze intergenerational effects. by using the psid data, we can observe the retirement savings behavior of parents in 1989 and compare it to that of their adult children in 2019. our first set of observations, however, will not yet incorporate intergenerational effects and will simply focus on data from the 2019 wave. the reason for this is straightforward— by doing so we can work with a larger sample size. the analysis throughout will focus on heads of households and will exclude those over the age of 67. once we begin analyzing intergenerational effects, we will only be able to include 2019 observations for those heads of households who are children of a head of household responding to the 1989 wave, reducing the number of observations with which we can work. on a smaller scale, even working within the 2019 wave itself, our sample size will vary from one exercise to the next depending on whether the participant elected to answer specific questions at hand. for example, participants who responded that they were covered by a pension plan at work but answered that they did not know how much money they had in the plan would be counted in our probit analysis (as saving for retirement) but not in our ordinary least squares (ols) regression in which we analyze accumulation. we will examine retirement plan participation in two ways. in the simplest form, we will simply observe whether the household is actively saving for retirement. we will not differentiate between saving by the head of household or by the spouse—saving by either or both are equally valid. the presence of a positive ira or other retirement plan balance, or the contribution to a retirement account within the past five years will result in a household being classified as saving for retirement. otherwise, the household is coded as not saving for retirement. armed with this binary classification, we can analyze the data with a standard probit, according to the equation: prob saving for retirement 1ð þ ¼ u xx b s � � (1) in eq. (1), xx is our vector of independent variables explaining the decision of whether to save for retirement, exclusive of the family of origin effects. later, we will simply modify eq. (1) such that: prob saving for retirement 1ð þ ¼ u xi b s � � (2) in eq. (2), xi is our vector of independent variables explaining the decision to save for retirement, inclusive of the family of origin effects. descriptive statistics of independent variables are found in table 1. b. t. starr / financial services review 30 (2022) 251–271 255 the second way that we will observe retirement plan savings is cardinal in nature. each household is asked what amount it has invested in an ira or other annuity and how much it has in its retirement account. we will add these two together as the measure of a household’s designated retirement savings and use ols regression to analyze it according to the standard form: y ¼ xxb þ y (3) here, y is the dollar amount of the household’s retirement savings and xx is the vector of explanatory variables, exclusive of family of origin effects. as with the probit analysis, we will then estimate: y ¼ xib þ y (4) table 1 descriptive statistics of psid data (author’s calculations) mean n min max sd dependent variables (2019) saving for retirement 0.46 8,404 0 1 0.50 retirement savings $61,130 7,881 0 $4,700,000 $226,971 independent variables (2019) age 42.37 8,404 31 66 10.29 income $80,048 8,404 �$267,900 $2,125,100 $91,002 education 13.43 8,310 0 17 2.55 male 0.68 8,404 0 1 0.47 married 0.51 8,404 0 1 0.50 health status 2.55 8,382 1 5 1.03 children 0.89 8,404 0 10 1.23 minority 0.48 8,328 0 1 0.50 lived with parents 0.65 8,262 0 1 0.48 union member 0.09 8,404 0 1 0.29 health insurance 0.91 8,348 0 1 0.29 religious $ $602 8,344 0 $62,000 $2,521 combo $ $78 8,373 0 $20,000 $526 needy $ $96 8,344 0 $30,000 $636 health $ $32 8,383 0 $15,000 $287 education $ $37 8,391 0 $15,000 $356 youth $ $17 8,388 0 $6,000 $176 cultural $ $15 8,398 0 $30,000 $354 community $ $6 8,395 0 $5,000 $85 environment $ $11 8,395 0 $5,000 $106 peace $ $10 8,390 0 $10,000 $150 other $ $31 8,395 0 $32,675 $519 independent variables (1989 family of origin) head interactive $19,773 3,713 0 $1,412,200 $48,616 spouse interactive $10,053 3,713 0 $465,080 $24,953 health of head 2.49 3,713 1 5 1.12 parent income $35,028 3,713 1 $1,412,200 $48,654 head education 4.70 3,713 1 9 1.83 spouse work 0.76 3,713 0 1 0.43 note. psid = panel study of income dynamics. 256 b. t. starr / financial services review 30 (2022) 251–271 in eq. (4), we are simply adding additional variables to the ols regression so that xi includes all independent variables included in xx plus the family of origin variables at the bottom of table 1. clearly, the data we analyzed in the probit analysis, actively saving for requirement, is a necessary condition for the actual accumulation of retirement savings and our evaluation thereof. the accumulated amount, however, is a function of many more variables, some of which we can analyze directly or indirectly with psid data (like income and age) and others (like risk tolerance) that lie beyond the scope of our analysis. still, our variables of interest will also prove salient in this second type of analysis and will be discussed below. 3. why doesn’t everyone save for retirement? 3.1. the effects of health while much research and the popular press demonstrate a good deal of worry over the fact that many americans simply do not save for retirement, it is useful to rehearse some rational reasons why a worker might choose the path of non-savings. such theoretical reasons will inform our investigation. perhaps the most intuitive reason for non-savings would be a state of health that is deemed insufficient for any significant post-retirement life expectancy. why forego current consumption to fund a part of the life-cycle one does not expect to attain in meaningful measure? of course, the psid survey team is not so gauche as to ask whether the participant expects to live to see retirement. we must rely on a slightly weaker instrument—namely the revealed perception of participant health. the survey question allows the respondent to choose excellent, very good, good, fair, or poor health, with the ordinal level 1 assigned to those describing themselves as having excellent health and level 5 designated to those in poor health. one might also wonder if the observation of parental health during childhood years might impact the retirement savings decision. a rational response to the observation of the poor health of one’s parent would be to conclude that one’s own life expectancy might be below average, thus, reducing the stock of financial capital needed to finance the retirement years. fortuitously, we have in the 1989 psid data the self-reported parental health assessment, on the same scale of 1–5, allowing us to test whether the observation of parental health might have a bearing on retirement savings decisions. 3.2. the effects of progeny the presence of children in the household might also have a bearing on retirement planning decisions. modern seminal work regarding intergenerational transfers was performed by becker (1974) who assumed that altruism motivated these transfers. later literature also explored the alternative motive that any intergenerational transfer proffered was done so with the expectation of getting the favor returned in the future (cox et al., 1998) rather than more altruistic motives (cigno, 1993). in many cultures throughout history, family expectations b. t. starr / financial services review 30 (2022) 251–271 257 included caring for those who could not care for themselves (see i timothy 5:8 for an ancient dictum to provide for one’s immediate family). societal laws and mores are still such that those who will not provide for their own children are viewed dimly. care for one’s elderly parents, however, might generate a broader array of responses. while children have no means by which to care for themselves, aged parents have presumably had a working lifetime to accumulate retirement savings and, in many cases, a retirement benefit stream from a government-funded plan. most savings models simply focus on the rational choice of a purely introspective agent, but some studies invoke the tools of game theory to help systematize behavior that is otherwise difficult to rationally explain (cigno, 1993). appendix briefly presents a set of simple games between parents, their children, and the government. the games are presented in extensive form for visual clarity. in each case the parent can be alternatively viewed in one of two ways. in the more pejorative view, the parent is selfish and cares nothing for the welfare of his or her children. in the more forgiving view, the parent’s expectation is that he or she will provide for both children and geriatric parents and that the children, in turn, might return the favor. for simplicity, we assume a sequential game of full information. a more realistic game might be one of the simultaneous moves with incomplete information on the payoffs of other players, but the game is styled simply to make some basic points. in the initial game, parents move first, deciding whether to save for retirement. children observe their parents’ actions and then decide, if the parent has elected not to save, whether to effectively “bail” out the parent. the subgame perfect equilibrium, of course, depends on the child’s altruism toward parents and the perception of the parents on that altruism. a more interesting version of the game is sketched in game 2, where the government also enters the picture. given recent governmental intervention in the economy, it seems reasonable to postulate that the government might go into deeper deficit spending mode to enhance social security payouts to prevent rampant geriatric poverty among the aged who have failed to save for retirement. the government moves second and whether it decides to bail out the non-saving parent is a function of its own empathy versus the deadweight loss associated with a bailout. it is also a function of its belief about what the child as last mover will do. we reverse the order of child and government in game 3 to demonstrate the advantage of priority when all payoffs are known. the altruistic second mover has the luxury of foregoing a costly bailout of the non-saving parent if it knows that the final mover will fund the mutually desired wealth transfer. and, of course, in all games the rational parent who expects to live to retirement will elect to save if it believes that there are no other players who will bail them out. a reasonable assumption on the coefficient d (<1) ensures that the parent would rather save some now than to suffer or prematurely expire due to insufficient post-retirement resources. suppose parents believe that their children will help care for them in retirement and that these parents fit the mold described for our simple games. in that case, the parent will maximize lifetime utility by minimizing (or even eliminating) retirement savings during the working years. we have a theoretically interesting reason to include in our parameters the variable of how many children live with the head of household. from south korea, we have empirical evidence showing crowding out by expanding pensions of intergenerational transfers to retired parents from their working-age children (jung et al., 2016). in thailand, 258 b. t. starr / financial services review 30 (2022) 251–271 expectation of support from children reduces the probability of retirement savings (witvorapong & yoon, 2021). indeed, cross-country analysis suggests, at least in countries without robust financial markets, children represent a de facto retirement savings (galasso et al., 2009). we will examine whether a correlated phenomenon might be happening in the united states. 4. the effects of altruism the familiar edgeworth box diagram demonstrating an agent’s interior bliss point when said agent values the welfare of the other player is instructive when we consider the potential effect of altruism on the retirement savings decision. if there are agents within the economy who value the immediate welfare of others more than their own future welfare, we might expect altruistic behavior to crowd out savings toward future consumption. for example, consider an agent who wishes to maximize the present value of his lifetime welfare stream w which is based on some function u of that agent’s own consumption c and an altruistic valuing of the consumption of others o all discounted at a homogenous time preference rate r . more formally, the agent wishes to maximize, subject to standard economic constraints: w ¼ ð fu ct, otð þge�r tdt (5) with @w @c > 0 and @w @o > 0 depending on value of r and on the agent’s relative preferences driving the u function to a specific form, some combination of one’s own current surplus and another’s current shortage will reduce (or possibly eliminate) the savings one would otherwise undertake. the psid dataset generously breaks down charitable giving into multiple categories that allow us to break down the effects of giving by type. in our list of variables in the tables at the end of the paper, we denote philanthropic giving with a dollar sign at the end of the variable. thus, philanthropic variables being tested for significance range from religious $ to other $. 4.1. the effects of family of origin we implicitly acknowledge that saving for retirement is a learned behavior and examine the possible effects of one’s family of origin on the retirement savings decision and commensurate outcomes in the level of such savings obtained. even with our largest sample of data, that which is exclusively from the 2019 survey, we have one variable giving insight into these potential effects—a binary variable indicating whether the respondent grew up in a household with both parents. at the cost of reducing sample size, we will more robustly b. t. starr / financial services review 30 (2022) 251–271 259 specify our model to include attributes of the 2019 respondent’s parents from the psid survey responses of those parents in 1989. 4.2. other variables our analysis also includes the standard lineup of demographic variables along with a binary variable indicating whether the respondent lived with both parents during childhood. we add variables for income and education, and dummy variables indicating whether the respondent is a union member and whether the respondent’s family is covered by health insurance. summary statistics of our variables are listed in table 1. we first see descriptive data on the two dependent variables—the binary variable in our probit analysis which indicates whether the respondent is actively saving for retirement and the cardinal variable indicating the level of accumulated retirement savings. we next see descriptive data on the independent variables observed for the respondents in 2019. we finally see descriptive data for the responses of their parents to the psid survey in 1989. the 1989 family of origin data elements include the ordinal measure of the health of the parental head of household, measures of that head’s education (recorded on a slightly different ordinal scale in 1989 than in 2019), parental income, and whether the spouse (worded as “spouse” in the 1989 survey) was working. we also consider an interactive variable for each parent which is family income multiplied by a dummy variable indicating whether that parent is participating in a retirement plan. an affirmative response to survey questions concerning whether the head (and, in turn, spouse) is covered by a retirement plan or contributing to an ira or annuity gives the binary variable a value of one. otherwise, the binary multiplier of the interactive variable is zero. given the wording of the 1989 survey, the spouse would be the spouse of the respondent. we can gain some insight on whether the example of one parent might be more potentially influential than the other. 4.3. analysis of results from 2019 data table 2 (probit) and table 3 (ols) show the results of the analysis of our 2019 data, with a more robust sample size. table 2 can be interpreted as an analysis of the first step of the retirement decision—whether to actively save for retirement at all, while table 3 analyzes the quantifiable result of the decision of, among other parameters, how much to save. in both cases, we have only one variable indicating the potential effect of one’s family of origin, the binary variable telling whether one grew up in a household with both parents. let us first consider the probit results of table 2. much of the result comports with intuition. income, age, education, marital status, union membership and the complementary benefits package of health insurance are all highly significant and take a positive algebraic sign on their coefficients. disconcerting, and beyond the scope of this paper, is the negative coefficient on the minority variable. comporting with intuition is the statistically significant negative coefficient associated with health status— poorer health is associated with a lower probability of saving for retirement. more intriguingly, we observe the highly significant and negative coefficient associated with the number 260 b. t. starr / financial services review 30 (2022) 251–271 of children the respondent has. it seems the presence of children in a household makes it less likely that household will save for retirement, a position bolstered by the fact that of the statistically significant philanthropic variables, only donations toward youth causes take a negative coefficient. this could lend credence to the game-theoretic reasoning we have discussed. parents might be looking to their children to help care for them during their retirement years. conversely, it could be that parents believe they have only enough discretionary income to care for their children and that saving for retirement is something they will undertake once their children have matured. besides charitable contributions to organizations that support youth, all other philanthropic variables of statistical significance take a positive coefficient. this suggests altruistic behavior does not crowd out retirement savings. rather, those who donate philanthropically, at least to organizations serving religious or environmental needs, are more likely to save for retirement. we also see in table 2 the first glimpse of the effect of family of origin. while table 2 probit analysis of 2019 psid data (n = 7,994) exclusive of the family of origin variables philanthropic contributions followed by “$” dependent variable: saving for retirement coefficient se z p-value constant �4.60345 0.245827 �18.73 < .0001*** income 5.83488e-06 3.34007e-07 17.47 < .0001*** age 0.0653389 0.0101352 6.447 < .0001*** age2 �0.000736212 0.000114324 �6.440 < .0001*** education 0.137487 0.00775512 17.73 < .0001*** male �0.0129764 0.0482848 �0.2687 .7881 married 0.329825 0.0488800 6.748 < .0001*** health status �0.0771171 0.0170285 �4.529 < .0001*** children �0.0937214 0.0153356 �6.111 < .0001*** minority �0.282718 0.0345042 �8.194 < .0001*** lived with parents 0.0994510 0.0356185 2.792 .0052*** union member 0.622159 0.0569302 10.93 < .0001*** health insurance 1.04121 0.0783915 13.28 < .0001*** religious $ 4.62806e-05 9.92698e-06 4.662 < .0001*** combo $ 4.58896e-05 4.86573e-05 0.9431 .3456 needy $ 9.18933e-07 3.36444e-05 0.02731 .9782 health $ 0.000280908 0.000143534 1.957 .0503* education $ �3.47601e-05 9.87521e-05 �0.3520 .7248 youth $ �0.000292925 0.000116167 �2.522 .0117** cultural $ 0.000323009 0.000195115 1.655 .0978* community $ �3.39467e-05 0.000353272 �0.09609 .9234 environment $ 0.00152810 0.000494932 3.087 .0020*** peace $ 0.000214843 0.000191108 1.124 .2609 other $ 7.58910e-06 3.94299e-05 0.1925 .8474 mean dependent variable 0.462222 sd dependent variable .498602 mcfadden r2 0.289363 adjusted r2 .285014 log-likelihood �3921.419 akaike criterion 7890.839 note. psid = panel study of income dynamics. the number of cases “correctly predicted” = 6,176 (77.3%). philanthropic donations followed by “$”. *p < .10, **p < .05, ***p < .01. b. t. starr / financial services review 30 (2022) 251–271 261 in this iteration of the analysis we know nothing of parental behavior, we see that those who grew up with both parents in the household are more likely to save for retirement. next, consider in table 3 our cardinal analysis of retirement savings accumulation as a function of the same variables analyzed from the binary analysis. through ols regression, we observe, when known, the total amount of retirement savings accumulation. one immediately notes several points of continuity between what we observed in the first stage of the retirement savings decision (whether to do so) and the amount which has been accumulated. as expected, income, age (or, in this case, its square) and education take positive coefficients in the ols regression. minority status continues to be both economically and statistically significant with a negative coefficient and poor health is associated with lower retirement account balances. similarly, all statistically significant philanthropic behavior (note more of the various types of philanthropy are statistically significant here than in our probit analysis) table 3 ols analysis of 2019 psid data (n = 7,512) exclusive of the family of origin variables philanthropic contributions followed by “$” dependent variable: retirement savings accumulated coefficient se t-ratio p-value constant �122706 18039.4 �6.802 < .0001*** income 0.827321 0.0335647 24.65 < .0001*** age2 35.3199 2.21022 15.98 < .0001*** education 6521.88 1006.48 6.480 < .0001*** male 6477.73 6719.82 0.9640 .3351 married �2189.18 6880.26 �0.3182 .7504 health status �10734.1 2353.69 �4.561 < .0001*** children �2830.25 2086.15 �1.357 .1749 minority �30057.6 4959.56 �6.061 < .0001*** lived with parents 2996.90 5036.12 0.5951 .5518 union member �20101.7 8227.37 �2.443 .0146** health insurance �3422.28 8146.88 �0.4201 .6744 religious $ 2.10752 0.958022 2.200 .0278** combo $ 10.7533 4.88625 2.201 .0278** needy $ 9.79702 4.16270 2.354 .0186** health $ 21.3235 8.58277 2.484 .0130** education $ �7.89183 7.32756 �1.077 .2815 youth $ �27.0763 13.7150 �1.974 .0484** cultural $ 17.2087 6.19687 2.777 .0055*** community $ 49.3573 28.9072 1.707 .0878* environment $ 174.036 21.6076 8.054 < .0001*** peace $ 32.6903 15.6816 2.085 .0371** other $ 11.9140 4.24040 2.810 .0050*** mean dependent variable 62001.83 sd dependent variable 228074.5 sum squared residuals 2.93e + 14 se of regression 197896.9 r2 0.249327 adjusted r2 .247122 f(22, 7489) 113.0630 p-value(f) .000000 log-likelihood �102260.2 akaike criterion 204566.3 note. ols = ordinary least squares, psid = panel study of income dynamics. philanthropic donations followed by “$”. *p < .10, **p < .05, ***p < .01. 262 b. t. starr / financial services review 30 (2022) 251–271 except for charitable contributions for youth are positively correlated with retirement savings accumulation. we also observe some points of departure from the table 2 results. while union membership seems to make it more likely that the member will save for retirement, the amount of accumulated savings is negatively correlated with union membership. we might conjecture that union members are more likely to also enjoy some type of defined benefit pension plan coverage under which an actuarial equivalent of the current account balance is unreported by the survey respondent. consequently, the results in table 3 might tell us little about the likely retirement well-being of union versus non-union members. another interesting point of discontinuity is the sole variable giving insight into the effects of one’s family of origin. while living with both parents while growing up leads to a greater likelihood of saving for retirement, it seems to have no bearing on the actual amount of retirement savings accumulation. just so, while the number of children in the household can have an adverse effect on the decision of whether to save for retirement, the number of children is not a statistically significant indicator of how much those parents who decide to save for retirement actually accumulate. 4.4. intergenerational analysis as noted earlier, the inclusion of parental variables from the 1989 wave reduces our sample size since not all 2019 responding heads of households had a parental head of household who responded to the 1989 survey. herein lies the weakness of this new step, but taking it allows us to specify the model more fully and analyze whether children seem to learn anything affecting their retirement savings behavior from their parents. following our earlier order of operations, we first observe the probit analysis reported in table 4. this analysis follows that described in table 2, but with the family of origin variables added. much remains unchanged among the variables earlier analyzed, but contributions to youth organizations are no longer of statistical significance in predicting whether a respondent saves for retirement. family of origin variables here typically present as statistically insignificant, with the notable and algebraically positive exception of the interactive variable constructed as the product of family income and the mother’s participation in a retirement plan. to control for the possibility that the effect might be a result simply of the mother’s participation in the workforce, a typically necessary condition to also save for retirement, we include a binary variable indicating whether the mother worked outside the home but find that variable to be statistically insignificant. if family-of-origin behavior affects the decision of whether to save for retirement, that behavior seems more likely to be transmitted through the mother. what happens when we analyze the amount of retirement balance accumulation with an ols regression that includes a family of origin variable? table 5 gives the answers. we should not be surprised to learn that income, age, education, union membership, and minority status, continue to display the same properties earlier discussed. however, we note that contributions to youth organizations now display statistical significance and a positive algebraic sign, joining the other statistically significant philanthropic variables as complements to retirement savings accumulation. and while a mother’s retirement savings behavior might influence the initial retirement savings decision, it seems to have no significant bearing on b. t. starr / financial services review 30 (2022) 251–271 263 the ultimate balances obtained within retirement accounts. the most interesting difference we see is the effect of parental health. not only is the health status of the parental head negatively correlated with retirement savings accumulation, but it also presents as highly significant while the health of the saver himself or herself is now displaced from statistical significance. this data suggests that retirement savers might be looking at parental health over their own health as a better predictor of the savers’ longevity and the commensurate magnitude of retirement savings need. table 4 probit analysis of 2019 psid data (n = 3,598) inclusive of family of origin variables philanthropic contributions followed by “$” family of origin variables in italics dependent variable: saving for retirement coefficient se z p-value constant �4.73750 0.435574 �10.88 < .0001*** income 6.69183e-06 5.57630e-07 12.00 < .0001*** age 0.0725972 0.0169832 4.275 < .0001*** age2 �0.000837659 0.000187176 �4.475 < .0001*** education 0.138385 0.0137482 10.07 < .0001*** male �0.0206323 0.0641102 �0.3218 .7476 married 0.287069 0.0713396 4.024 < .0001*** health status �0.0882530 0.0262294 �3.365 .0008*** children �0.0675705 0.0252273 �2.678 .0074*** minority �0.297096 0.0564493 �5.263 < .0001*** lived with parents 0.147792 0.0580016 2.548 .0108** union member 0.672718 0.0880006 7.644 < .0001*** health insurance 0.900498 0.107364 8.387 < .0001*** religious $ 3.38554e-05 1.44712e-05 2.340 .0193** combo $ 2.46305e-05 7.09484e-05 0.3472 .7285 needy $ �5.62748e-06 5.15755e-05 �0.1091 .9131 health $ 0.000536065 0.000289928 1.849 .0645* education $ �0.000109256 0.000156249 �0.6992 .4844 youth $ �0.000182149 0.000187711 �0.9704 .3319 cultural $ 0.000385624 0.000285471 1.351 .1767 community $ 0.00154095 0.000992411 1.553 .1205 environment $ 0.00184249 0.000837875 2.199 .0279** peace $ 0.000478458 0.000522544 0.9156 .3599 other $ 8.44776e-05 9.69104e-05 0.8717 .3834 head interactive �6.77236e-07 1.21394e-06 �0.5579 .5769 spouse interactive 2.43506e-06 1.21300e-06 2.007 .0447** health head 0.00680264 0.0251455 0.2705 .7868 parent income 1.20853e-06 1.33266e-06 0.9069 .3645 parent education 0.0151172 0.0164922 0.9166 .3593 spouse work �0.0300312 0.0608267 �0.4937 .6215 mean dependent variable 0.469983 sd dependent variable .499168 mcfadden r2 0.310560 adjusted r2 .298500 log-likelihood �1714.952 akaike criterion 3489.903 note. ols = ordinary least squares, psid = panel study of income dynamics. number of cases “correctly predicted” = 2,810 (78.1%). philanthropic donations followed by “$”. *p < .10, **p < .05, ***p < .01. 264 b. t. starr / financial services review 30 (2022) 251–271 5. practical implications for financial planners as the role of financial advisors evolves and expands (sommer et al., 2022), client conversations should be handled with concomitantly evolving prudence and tact. health, whether parental or personal, apparently matters to many savers and the perception of bad health can present as a barrier to saving for retirement. the skilled advisor, rather than relying simply on averages or mortality tables, might ask gentle questions about whether those table 5 ols analysis of 2019 psid data (n = 3,411) inclusive of family of origin variables philanthropic contributions followed by “$” family of origin variables in italics dependent variable: retirement savings accumulated coefficient se t-ratio p-value constant �136051 33972.7 �4.005 < .0001*** income 1.06365 0.0566082 18.79 < .0001*** age2 42.9840 3.80438 11.30 < .0001*** education 5708.48 1983.25 2.878 .0040*** male 6982.38 9541.63 0.7318 .4644 married 892.960 10620.2 0.08408 .9330 health status �5610.92 3834.08 �1.463 .1434 children �1362.58 3640.49 �0.3743 .7082 minority �20873.7 8502.90 �2.455 .0141** lived with parents �1444.96 8468.30 �0.1706 .8645 union member �28366.0 13315.1 �2.130 .0332** health insurance �12515.5 12629.5 �0.9910 .3218 religious $ 0.696535 1.58741 0.4388 .6608 combo $ �5.24681 8.08576 �0.6489 .5165 needy $ 29.1985 6.99521 4.174 < .0001*** health $ 77.2639 21.7333 3.555 .0004*** education $ 17.5444 11.1140 1.579 .1145 youth $ 90.1074 25.4434 3.541 .0004*** cultural $ 5.63254 6.82375 0.8254 .4092 community $ 173.802 47.0682 3.693 .0002*** environment $ 214.067 27.4320 7.804 < .0001*** peace $ �18.7385 36.7719 �0.5096 .6104 other $ 38.6322 10.6156 3.639 .0003*** head interactive �0.0702475 0.161830 �0.4341 .6643 spouse interactive �0.131690 0.156531 �0.8413 .4002 health of head �12924.1 3713.55 �3.480 .0005*** parent income 0.0520071 0.172694 0.3012 .7633 head education 2996.05 2416.05 1.240 .2150 mean dependent variable 70064.03 sd dependent variable 256521.9 sum squared residuals 1.51e + 14 se of regression 211289.5 r2 0.326939 adjusted r2 .321567 f(27, 3383) 60.86260 p-value(f) 1.7e-266 log-likelihood �46648.16 akaike criterion 93352.32 note. ols = ordinary least squares, psid = panel study of income dynamics. philanthropic donations followed by “$”. *p < .10, **p < .05, ***p < .01. b. t. starr / financial services review 30 (2022) 251–271 265 national averages seem right for the client and, if not, why not. the answer to such questions might minimize the need to save for retirement. conversely, the planner might appropriately nudge into salubrious financial action a non-saver who otherwise (perhaps falsely) assumes a premature death will fully obviate the need for retirement savings. without a helpful nudge, such a client might experience the ironic joy of serendipitous life accompanied by the regret of having underfinanced it. family can matter as well. the skilled planner might consider asking what their clients learned about saving from their parents, leveraging all helpful knowledge previously gained, and filling in the apparent gaps. more importantly, financial advisors can be keenly alert to the possibility that the presence of children in the household might be delaying the client’s saving for retirement. again, such action could be rational if the family has a sound history of intergenerational care that is inculcated into each successive generation. perhaps more likely, the conversation will provide an opportunity to educate the client about the time value of money and the crucial importance of saving early. on a more relieving note, if an advisor worries that altruistic behavior might crowd out a client’s ability to save, that advisor can take some comfort in evidence that suggests altruistic giving and saving for retirement are typically complementary behaviors. an examination of the client’s portfolio should quickly confirm whether the client fits this pattern or proves to be an exception, with further conversation needed. 6. conclusion while psid data are not as specific to retirement savings as other surveys, its possession of intergenerational data and other elements that retirement-specific surveys might ignore make it worthy of exploration. here is what the data suggests: 1. health status plays a predictable role in the retirement savings decision. those who enjoy good health are more likely to save for retirement. when analyzing the amount of retirement savings accumulated, the good health of the saver drives higher savings until we consider intergenerational variables. once those variables are included, parental health is more significant than one’s health in predicting the retirement savings level. 2. the number of children in the household is inversely correlated with the household’s likelihood of saving for retirement, possibly suggesting some parents plan to draw some retirement support from their children. however, the number of children in the household is statistically insignificant as a predictor of the actual amount of retirement savings accumulation. 3. to the extent that it is statistically significant, most altruistic giving is complementary to retirement plan savings. crowding-out effects are only observed with those contributions made to serve youth, and that effect disappears when adding intergenerational variables. 4. there is evidence that the decision to save for retirement is positively impacted by having lived with both parents growing up, and the mother’s retirement savings behavior might positively influence the retirement savings decision of her children. 266 b. t. starr / financial services review 30 (2022) 251–271 appendix games parents, children, and governments might play for the following games between players parents, children, and government, let the following variables define the payoffs. p = present value of the parent’s lifetime utility resulting from the consumption that obtains from their own productivity. c = present value of utility enjoyed by the government that obtains from knowing its elderly citizens are provided for correctly. x = present value of utility enjoyed by children from knowing their parents are provided for correctly. h = present value of the utility parent receives from a transfer from child. w = present value of the child’s cost of providingh to the parent. g = present value of utility parent enjoys from a government transfer. dwl = deadweight loss government sustains to fund g. d = parental discount factor associated with running out of resources prematurely. (p, χ) (p+θ, χ—ψ) parent child save don’t save baildon’t bail if χ ˃ ψ, subgame perfect equilibrium, or “spe” = (don’t save, bail) if χ ˂ ψ, spe = (save, don’t bail) (δp, 0) parent, child game 1 b. t. starr / financial services review 30 (2022) 251–271 267 (p, γ, χ) (p+g, γdwl, χ) (δp, 0, 0) (p + θ, γ, χ-ψ) parent government child save don’t save baildon’t bail baildon’t bail if χ ˃ ψ, spe = (don’t save, don’t bail, bail) if x < ψ and γ < dwl, spe = (save, don’t bail, don’t bail) if χ ˂ ψ and γ ˃ dwl, spe = (don’t save, bail, don’t bail) parent, child, and government game 2 268 b. t. starr / financial services review 30 (2022) 251–271 references barboza, g., bongini, p., & rossolini, m. (2021). financial (il)literacy vs. individual’s behavior: evidence on credit card repayment patterns. financial services review, 29, 247–276. becker, g. (1974). a theory of social interactions. journal of political economy, 82, 1063–1093.december). bernheim, b. d., skinner, j., & weinberg, s. (2001). what accounts for the variation in retirement wealth among u.s. households. american economic review, 91, 832–857. beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2007). simplification and saving. nber working paper 12659. cambridge, ma: national bureau of economic research. beshears, j., choi, j. l., & madrian, b. (2007). the impact of employer matching on savings plan participation under automatic enrollment. 9th annual joint conference of the retirement research consortium: challenges and solutions for retirement security (pp. 1–24). washington, d.c. (p, χ, γ) (p+θ, χ -ψ, γ) (δp, 0, 0) (p + g, x, γ-dwl) parent child gov’t save don’t save baildon’t bail baildon’t bail if χ ˃ ψ, spe = (don’t save, don’t bail, bail) if x < ψ and γ < dwl, spe = (save, don’t bail, don’t bail) if χ ˂ ψ and γ ˃ dwl, spe = (don’t save, bail, don’t bail) parent, child, and government game 3 b. t. starr / financial services review 30 (2022) 251–271 269 beshears, j., dai, h., milkman, k. l., & benartzi, s. (2021). using fresh starts to nudge increased retirement savings. organizational behavior and human decision processes, 167, 72–87. brown, j. r., liang, n., & weisbenner, s. (2007). individual account investment options and portfolio choice: behavioural lessons from 401(k) plans. journal of public economics, 91, 1992–2013. cardella, e., kalenkoski, c. m., & parent, m. (2021). less is not more: 401(k) plan information and retirement planning choices. journal of pension economics and finance, 1–21. charles, k. k., & hurst, e. (2002). the correlation of wealth across generations. nber working paper 9314. cambridge, ma: national bureau of economic research. chiteji, n., & stafford, f. (2000). asset ownership across generations. the levy economics institute of bard college working paper 314. annandale-on-hudson, ny. choi, j. j., laibson, d., & madrian, b. c. (2004). plan design and 401(k) savings outcomes. national tax journal, 57, 275–298. choi, j. j., laibson, d., & madrian, b. c. (2006). reducing the complexity costs of 401(k) participation through quick enrollment(tm). nber working paper 11979. cambridge, ma: national bureau of economic research. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2009). reinforcement learning and savings behavior. the journal of finance, 64, 2515–2534. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2001). defined contribution pensions: plan rules, participant decisions, and the path of least resistance. tax policy and the economy. cambridge, ma: national bureau of economic research. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2009). reinforcement learning and savings behavior. the journal of finance, 64, 2515–2534. choi, j. j., laibson, d., madrian, b., & metrick, a. (2001). for better or for worse: default effects and 401(k) savings behaviour. nber working paper 8651. cambridge, ma: national bureau of economic research. cigno, a. (1993). intergenrational transfers without altruism. european journal of political economy, 9, 505– 518. cole, s., paulson, a., & gauri, k. s. (2014). smart money? the effect of education on financial literacy. review of financial studies, 27, 2022–2051. cox, d., eser, z., & jiminez, e. (1998).motives for private transfers over the life cycle: an analytical framework and evidence for peru. journal of development economics, 55, 57–80. engen, e., williams, g., & scholz, j. k. (1996). the illusory effects of saving incentives on savings. journal of economic perspectives, 10, 113–138. galasso, v., gatti, r., & profeta, p. (2009). investing for the old age: pensions, children and savings. international tax and public finance, 16, 538–559. jung, h., pirog, m., & lee, s. k. (2016). do public pensions crowd out private transfers to the elderly?: evidence from south korea. journal of pension economics and finance, 15, 455–477. love, d. (2010). the effects of marital status and children on savings and portfolio choice. review of financial studies, 23, 385–432. madrian, b., & shea, d. (2000). the power of suggestion: inertia in 401(k) participation and savings behavior. nber working paper 7682. cambridge, ma: national bureau of economic research. mitchell, o. s., utkus, s. p., & yang, t. (2007). turning workers into savers? incentives, liquidity, and choice in 401(k) plan design. national tax journal, 60, 469–489. munnell, a., sunden, a., & taylor, c. (2000). what determines 401(k) participation and contributions: crr working paper 2000-12. boston college. poterba, j. m., venti, s. f., & wise, d. a. (1996). how retirement savings programs increase savings. journal of economic perspectives, 10, 91–112. samwick, a. a., & skinner, j. (2004). how will 401(k) pension plans affect retirement income? american economic review, 94, 329–343. shore, s. (2011). the intergenerational transmission of income volatility: is riskiness inherited? journal of business & economic statistics, 29, 372–381. 270 b. t. starr / financial services review 30 (2022) 251–271 sommer, m., lim, h., & macdonald, m. (2022). financial advisor use, life events, and the relationship with beneficial intentions. financial services review, 30, 69–88. tang, n., & peter, p. c. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experiences. financial services review, 24, 119–138. wise, d. a. (2007). the economics of aging. nber reporter, 4, 1–10. witvorapong, n., & yoon, y. p. (2021). do expectations for post-retirement family and government support crowd out pre-retirement savings? insights from the working-age population in thailand. journal of pension economics & finance, 21, 218–236. http://dx.doi.org/10.1017/s1474747220000360 b. t. starr / financial services review 30 (2022) 251–271 271 the performance of the faith and ethical investment products: a comparison before and after the 2008 meltdown francisca m. beera,*, james p. estesa, charlotte deshayesa adepartment of finance, california state university san bernardino, 5500 university parkway, san bernardino, ca 92407, usa abstract this article explores the risk and return characteristics of socially responsible investment and faith-based mutual funds before and after the market crisis of 2008. findings show a high level of correlation between the indices studied as well as a higher volatility than the s&p 500. we also find a significant shift in the mix of performance and volatility of these funds before and after the crash of 2008. this is an important consideration for both planners and investors in making an informed decision that is tempered by both the intensity of their social or faith based investment preferences and resultant risk and return on those investments. © 2014 academy of financial services. all rights reserved. jel classification: g11; g15; z12 keywords: mutual funds; downside risk; socially responsible investment; faith based investment 1. introduction in the classical financial theory of markowitz (1952), the choice of an efficient portfolio of assets, is predicated on the maximization of the investor’s maximization of return and minimization of risk, disregarding other rewards of a social nature. increasingly, planners are faced with client discussions regarding socially, ethically, and religious investment preferences and should be aware of the various implications that each type of investment presents * corresponding author. tel.: �1-909-537-5709; fax: �1-909-537-5078. e-mail address: fbeer@csusb.edu (f. beer) financial services review 23 (2014) 151–167 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. in a pioneering article, sen (1999) showed the significance of social, moral, or ecological motives on the selection of investment products and that selection on the growth of market economics. the first study used data between 1998 and 2008 whereas this study updated those findings with data taken between 2008 and 2012; continuing to explore the risk and return characteristics of socially responsible investing and faith-based investing by comparing three indices, the msci kld 400 social index (kld 400), the updated ftse kld catholic values 400 index (cv400), and the dow jones islamic market index (djim). this study also compares the performance of the socially responsible investing and faith-based indices with the s&p 500 before and during the 2008 meltdown and then again in the recovery through 2012. the traditional measures of performance of sharpe, treynor, and jensen are again used to measure performance. however, because these measures are based on the assumptions that the portfolio returns can be totally characterized by the mean and variance of the return distribution (condition 1) and that investors care only about the mean and variance of the return distribution (condition 2), the traditional performance measures are appropriate when returns are normally distributed. reality, however, is complex and returns are rarely normally distributed. returns distribution is often characterized by “fat tails” and “skewness” (maheu and mccurdy, 2009). to account for these features of the data we use the lower partial moment as an alternative measure for risk (baillie and degennaro, 1990; nantell, price, and price, 2009). the authors then calculate the sortino measure of performance and modify the treynor measure and the jensen alpha by using the semi-standard deviation as the measure of risk. findings show that all of these performance measures generally yield similar inferences. results also show that socially responsible investment and faith based funds provide above average profit opportunity. empirical studies about socially responsible investing and faith-based are important and should be pursued even further. sparkes and cowton (2004) reported that the socially responsible investing and faith-based industry has grown and matured significantly since the early 2000. they showed that socially responsible investing and faith-based industries are now mainstream investments. socially responsible investing and faith-based investing have evolved from an activity managed by a small number of specialty retail investment funds into an investment philosophy adopted by a growing proportion of some of the large investment institutions. this may be evolving even further with classes in ethics now required at many colleges and universities. this trend is reflected in student managed investment funds, which may well be a precursor to the investment philosophy of future investors. clinebell (2013) points out that incorporating socially responsible investing into student managed investment funds may prove to be “a useful vehicle for addressing the ethical and social issues faced in today’s investment environment.” this bodes well for increased focus on socially responsible investing and its relevance to advisors and investors alike. only a few studies have compared the performance of socially responsible funds with faith based funds (hakim and rashidiank, 2002; sadeghi, 2008). these studies have different focuses. sadeghi (2008) focus is not on performance but rather on the impact of the introduction of shariah-compliant index on the malasian stock exchange. hakim and 152 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 rashidian (2002) examines the stochastic properties of the islamic index. the authors also investigate the relation between the islamic index and the broader stock market represented by the wilshire 5000 index. increasingly, investors and financial planners are seeking socially responsible or faith based funds for investments in an effort to address concerns based on social or religious values. although this is important, it also introduces a level of violently that may not be anticipated but is relevant to planners and investors. while some trusts or faith based investments are required to remain within the guidelines, alternatives should be discussed with clients in light of a balanced portfolio and its implications for risk and volatility. should these alternatives be rejected by the investors or advisory boards then the advisors have done their job in considering appropriate alternatives in increasing performance, and/or decreasing risk and volatility. in addition to a different focus, there are no studies that use both traditional and less traditional measures of performance in comparing these indices. this article addresses these shortcomings. it consists of 5 sections. section 2 reviews the literature. section 3 introduces the data and the methodology. section 4 presents the findings. section 5 concludes the study. 2. literature review over the past two decades, exchange-traded funds (etfs) and mutual funds have become the preferred investment option for small-scale investors and many independent planners. today, the number of mutual funds exceeds the number of listed securities in the new york stock exchange (nyse) (rouwenhorst, 2004). with the advent of even more etfs this gap continues to widen. there a wide variety of mutual fund categories and objectives. this article will focus on specialty funds. specialty funds are mutual funds that mainly invest in a specific market, a region, an industry, or a somewhat narrow group of assets. socially responsible funds (srf) and faith based funds (fbf) belong to this category. the increasing interest for faith-based funds has grown significantly and inspired recent empirical studies. kreander, mcphail, and molyneaux (2004) examined the motives of socially responsible investors and came to the conclusion that in many cases, the demand for these funds is based on “riba.” there are two types of “riba.” the first type prohibits practices that could lead to one party increasing his or her wealth without providing services to the other party. the second type prohibits exchanges of commodities in unequal quantities. girard and hassan (2005) described investment based on riba as: “… muslims deemed that profit should come as a result of efforts”; this is not the case in interest-dominated investments. this growth in interest for this type of fund would seem to confirm a growing concern by ethically motivated investors about issues such as the environment, women, employees, and communities and the selection of their investments accordingly. some studies also investigated the screening process used by faith-based funds. scatizzi (2010) explained that most faith-based funds avoid companies that produce alcohol, tobacco, and pornography, companies engaged in producing and selling firearms and those in the oil industry. types of etfs include islamic, christian faiths, and other religions. lyn and zychowicz (2010) studied the impact of faith-based screens on investment performance from 153f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 may 2001 to february 2008 in the united states. the comprehensive set of tests show that faith-based funds mostly outperformed the market, including socially responsible investing funds. there are some interesting differences between faith-based mutual funds and etfs (screened) and unscreened funds. geczy, stambaugh, and levin (2005) noticed significant differences in the expense ratios between screened and unscreened funds. screened funds have an average expense ratio of 1.3% versus 1.1% for unscreened ones. screened funds have lower turnover (81.5% average vs. 175.4%) and they tended to be smaller ($150 million average assets vs. $260 million). evidence related to srf performance is mixed. barnett and salomon (2006) showed that as the number of social screens increase, financial returns decline. stone, guerard, gultekin, and adams (2001), supporting the findings of waddock and grave (1977) concluded: no cost in risk-adjusted return means that an organization can affirm its social values without foregoing return. however, if socially responsible investing provides better risk-adjusted returns, then it pays to be socially responsible even if there is no issue of affirming social values.” an important point for advisors in the discussion of this type of funds. bauer, koedijk, and otten (2005) did not find evidence of a significant difference in risk-adjusted returns between ethical and conventional funds for the 1990–2001 periods. renneboog, ter horst, and zhang (2008) concluded that existing studies “hint but do not unequivocally demonstrate” that srf investors are willing to accept suboptimal financial performance to pursue social or ethical objectives. barnett and salomon (2003) concluded that the screened funds underperformed the s&p 500 on a nominal basis. the authors also indicated that the best performers were those with the strongest and weakest social screens. other studies introduced evidence more in favor of the hypothesis that srf outperformed most traditional investments. in similar studies statman (2000, 2005) and sauer (1997) compared the performance of the srf and the s&p 500 index. utilizing jensen’s alpha and sharpe’s ratio, they found that the ds400 raw return and risk-adjusted returns are higher than the s&p 500 index. goodmacher (2006) found that “the mean raw [excess] returns for the group of srf funds are actually superior to those of the group of non-srf funds, although this difference was not statistically significant.” although we believe that socially and ethically conscious investors seek to invest in assets that yield good profit and at the same time give them peace of mind, results from previous studies are contradictory. are srf and fbf outperforming the s&p 500, or is it the other way around? our first goal is thus to revisit previous findings using different types of performance measures. specially, our study uses traditional measures of performance and, as mentioned earlier, to mitigate the fact that returns do not follow a normal distribution, some measures based on downside risk. as stated previously, this important limitation is often ignored in studies. violation of the normality assumption can compromise empirical studies results. beer, estes, and munte (2011), prove that measures of downside risks explain the downturn of worldwide markets in 2008. one purpose of this article is to compare the performance of christian and muslim funds and the impact of the screening process on the risk and return of these funds. a dynamic 154 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 explored in this article is the effect on performance of the inclusion of financial stocks in the s&p 500 and their specific exclusion from the catholic, muslim, and social index funds when explored in the downturn in 2008 and corresponding recovery for the next four years. scholars comparing christianity and muslim faith report both similarities and differences (smith, 2002; hsu et al., 2008). these similarities and differences explain why investors’ religious beliefs are a determinant of their investment decisions. empirical studies attempted to evaluate the impact of economic downturn on sri performance. dania and malhotra (2012) identify linkages between four major islamic indexes and the corresponding “conventional” indexes. in particular they found evidence of a positive and significant spillover from conventional indexes on their corresponding islamic indexes. even if sri and market have common performance patterns, the screening process of sri can also protect the investors, in some cases, from market risks. for example, when the stock market crashed in 2008, one of the most impacted sectors was financial services. funds serving christian principles never bought shares in those companies because of their support for alternative lifestyles such as including same-sex couples in employee-benefit plans. 3. data and methodology 3.1. data data for the indexes used in this article were collected from various databases, that is, the msci database, yahoo finance and crsp.1 (these indices are described below). data were also collected for monthly treasury bills using the u.s. department of the treasury web site. 3.1.1. the msci kld 400 social index (ds400) kld research and analytics (kinder, lydenberg, domini, and company), inc. established the domini 400 social index (ds400) in may 1990 as the first ethical-social index that is concerned with environmental, social and governance factors (esg). kld uses a two-step screening process for selecting companies for the ds400 (www.msci.com). first, companies involved in alcohol, gambling, tobacco, military weapons, civilian firearms, nuclear power, adult entertainment, and genetically modified organisms are excluded from the index: second, using the list mentioned above, kld screens companies based on considerations of esg performance, sector alignment and size representation. 3.1.2. the msci u.s.a. catholic values index (cv 400) “the cv 400 is a free float-adjusted market capitalization index designed to be used as a u.s. equity benchmark for catholic investors who seek equity ownership in alignment with the moral and social teachings of the catholic church. the cv 400 consists of 400 companies selected from the msci u.s.a. investable market index (imi).” (www.msci. com). each company’s catholic values performance is evaluated based on respecting 155f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 human life, promoting human dignity, reducing arms production, pursuing economic justice, protecting the environment, and encouraging the corporate responsibility. 3.1.3. the dow jones islamic market index (djim) the third index used in this study is the djim, the main indicator for the performance of islamic funds and provides a benchmark tracking sharia-compliant, based on the koran, securities. the index selected companies in 34 countries whose activities are consistent with islamic principles. the majority of financial institutions are prevented from been part of the index along with companies involved in the production or distribution of alcohol, pornography, tobacco, gambling, weapons, music, entertainment, and pork meat or non-halal meat, hotels and airlines that serve alcohol on their premises. typical holdings are technology, telecommunications, steel, engineering, transportation, health care, utilities, construction, and real estate. 3.1.4. rate of returns as commonly done in the financial literatures, all index series returns are calculated using the continuously compounded formula (hussein and omran, 2005): rt � ln� pt pt�1 � (1) where pt and pt�1 represent the closing price of an index at time t and t�1, respectively, and ln is natural logarithm. our data include 174 monthly observations over the periods july 1998 to june 2008 and january 2008 to december 2012. the major benchmark for these indices is the s&p 500 for which the data are taken from crsp. in figs. 1 and 2, the monthly returns of cv400, ds400, and djim indices are plotted with the monthly return of s&p 500. figs. 1 and 2 show some monthly patterns between the indices. further as depicted in table 1, for the period 1998–2008, the mean return for the s&p 500 was lower than those of the socially responsible and faith based funds. during the same period, the monthly mean return of the djim is the highest; an investment in the djim index was slightly more rewarding than investments in other indices. notice, however, that the djim is also the most volatile index. after the crisis, it is the cv400 that provides the largest mean return and the ds400 that exhibit the lower volatility. msci for the period 1998–2008 and for the period 2008–2012, the correlation coefficient between the cv400 and the ds400 remain the higher coefficient reaching a value of 99%. the cv400 is similar to the kld’s domini 400 social index with additional layers of screens covering abortion, contraceptive products and embryonic/fetal stem cells. the high correlation coefficient can probably be explained by the fact that both indices are produced by kld research and analytics. for the period 1998–2008 and the period 2008–2012, the smallest coefficient is found between the djim and the cv400. the djim is, however, more restrictive than the cv400 156 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 and the ds400. the djim excludes companies involved with pork products, hotel and leisure industries, and conventional financial services (banking, insurance, etc.). the magnitude of the correlation coefficients, explain why kempf and osthoff (2007) used the terms fig. 1. monthly returns distribution 1998–2008. fig. 2. monthly returns distribution 2008–2012. 157f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 “conventional funds in disguise” when referring to srf and fbf. the correlation between the s&p 500 and the other three indexes rose from all three religious and socially based indexes rose more than 1% in the post 2008 era. 3.2. methodology as mentioned earlier, we rely on six different performance measures: sharpe’s ratio, treynor’s ratio, jensen’s alpha, sortino ratio, treynor semi-standard deviation (ssd), and the jensen ssd. each measure is briefly presented below. 3.2.1. sharpe ratio sharpe ratio � ri � rf �i (2) in eq. (2), ri represents the return of the index; rf is the benchmark asset, that is, the treasury bill and �i is standard deviation of the index. a higher sharpe ratio indicates superior performance, whereas lower sharpe ratio indicates poor performance. table 1 descriptive statistics and correlation cv400 ds400 djim s&p 500 panel a: 1998–2008 mean 0.0029 0.0030 0.0037 0.0010 median 0.0037 0.0039 0.0066 0.0070 maximum 0.1052 0.1052 0.1037 0.0923 minimum �0.1484 �0.1481 �0.1398 �0.1576 standard deviation 0.0456 0.0455 0.0507 0.0436 correlation cv400 ds400 djim s&p 500 cv400 1 0.9998 0.9308 0.9819 ds400 1 0.9314 0.9819 djim 1 0.9316 s&p 500 1 panel b: 2008–2012 cv400 ds400 djim s&p 500 mean 0 �0.0001 �0.0012 �0.0005 median 0.0059 0.0061 0.003 0.0081 maximum 0.1114 0.0989 0.1077 0.1023 minimum �0.1777 �0.1701 �0.1991 �0.1856 standard deviation 0.0570 0.0548 0.0629 0.0561 correlation cv400 ds400 djim s&p 500 cv400 1 0.9991 0.9255 0.9919 ds400 1 0.9258 0.9931 djim 1 0.9464 s&p 500 1 158 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 3.2.2. treynor ratio treynor index � ri � rf �i (3) in eq. (3) ri and rf are as defined previously, and �i is beta of the index. the higher the value of treynor index, the more return is gained per unit of systematic risk. 3.2.3. jensen alpha alpha � � � ri � �rf � �i�rm � rf�� (4) in eq. (4) ri, rf, and �i, are as defined above and rm is the s&p 500 composite index. alpha evaluates the returns that the fund has generated against the returns actually expected out of the fund given the level of its systematic risk, a positive jensen alpha indicates the index profit more than expected. the sharpe ratio, the treynor index, and the jensen alpha have the convenient property of being completely captured by a risk-return frontier, from which one can interpret them as the slope of the efficient line, or return per unit risk (pedersen and rudholm-alfvin, 2003). as a consequence the performance measures of sharpe, treynor, and jensen alpha are appropriate when returns are normally distributed. in addition, the capm assumes that a quadratic utility function adequately apprehended investors’ preferences. this assumption is counterintuitive. a quadratic utility function implies that as people become wealthier their risk aversion increases. in reality, investors are concerned with earning less than expected not more than expected. rational investors should try to avoid “downside” volatility only. when the underlying assumptions are violated, alternative measures of risk can be used. one of the best known alternative measures of risk is the lower partial moments or lpm (t,k) (fishburn and kochenberger, 1979; luce, 1980; sarin (1987); fishburn, 1980, 1981; kahnemann and tversky, 1979; thaler, 1993). 3.2.4. lower partial moment lpmi�ben, k� � �e�ben � ri� k�ben � ri� 1/k (5) in eq. (5), ben is a specific benchmark against which performance is measured (e.g., the risk free rate); k measures the sensitivity to extreme losses (pedersen and rudholm-alfvin, 2003) whereas ri was being defined previously. as shown in (5), lpm uses only the returns that are less than the predetermined benchmark. several subcases of lpm (ben,k) can be used, that is, absolute shortfall (as), the probability of shortfall (ps), and semistandard deviation (ssd). this study relies on the ssd, mainly because it’s well established relationship with the capm 3.2.5. capital asset pricing model ssd � lpmi�rf ,2� � �e�rf � ri� k�ben � ri� 1/k (6) 159f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 hogan and warren (1972), bawa and lindenberg (1977), harlow and rao (1989), and satchell (1996) introduced a ssd derived capm. in the model, “the risk measure, �i ssd, is defined in terms of the “semicovariance” with the market.” pedersen and satchell (2002) demonstrated that when risk is identified by the ssd equation, the risk/return frontier exhibits the same desirable convexity properties as the traditional mean-variance frontier. the ssd has lead to the development of a treynor ssd and an alpha ssd (henriksson and merton (1981), henriksson (1984), pedersen and satchell (2000)). these measures are presented below. 3.2.6. treynor ratio ssd treynor (ssd) � ri � rf �i ssd (7) 3.2.7. jensen alpha ssd jensenalpha(ssd) � �ssd � ri � ⎣rf � �i ssd�rm � rf�⎦ (8) in addition to the treynor (ssd) and the jensen alpha (ssd), the sortino ratio introduced by sortino and price (1994) was also calculated. the sortino ratio is expressed below. 3.2.8. sortino ratio ssd sortino � ri � rf ssd (9) as shown in eq. (9), the sortino ratio is similar to the sharpe measure. the only difference being the denominator, that is, the measure of risk. the performance measures presented above have an important advantage. pedersen and rudholm-alfvin (2003), show that these measures are appropriate (the measures capture essential features of the asset return distribution), derived from solid models (the measures have a solid foundation either in finance theory or are a universally applied market standard) and they provide clarity (the measures are easy to explain to a nontechnical individual). pedersen and satchell (2002) find evidence that the sortino ratio is a reliable measure of performance. barndorff-nielsen, kinnebrock, and shephard (2008) report that realized semivariances have important predictive qualities for future market volatility. 4. findings results in table 2 show the s&p 500 returns depart from normality for both the preand post-crisis periods. results also indicate that for the other indices, the normality hypothesis cannot be rejected. the betas (�i) presented in table 3 show the indices beta coefficients computed as the regression coefficients of the indexes studied on the portfolio returns. as shown in table 3 between 1998 and 2008, ethical-based and faith-based investment betas were somewhat 160 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 higher than one, showing volatility slightly higher than the s&p 500. the highest beta found is for the djim index. when calculating �i ssd, ethical-based and faith-based investment betas also remain slightly above one. �i ssd are also obtained by regression, however, �i ssd uses only observations that fall below the risk free rate. indeed, semivariance estimates the average loss that a portfolio could incur. with the exception of the ds400 �i ssd, betas for the period 2008–2012 are somewhat lower, another indication of a less volatile market. anova f-tests and p-levels presented in table 4 confirm that we cannot reject the hypothesis that the monthly mean returns are different from each other when we used data for the entire period sampled. the hypothesis of monthly mean different returns before 2008 cannot be rejected either. further, the hypothesis of monthly mean different returns after 2008 cannot be rejected. finally, for the ethical-based, faith-based investment and s&p500 cannot be rejected. clearly, although monthly means for the ethical-based and faith-based investment (ds400, cv400, and djim) performed better than the s&p 500 on a risk adjusted basis during the period before 2008 and generally better after with the exception of the djim after 2008 (see table 5). results are not significant. 5. summary and conclusion this article explores the risk and returns characteristics of socially responsible investment and faith-based investment investments and compares them to the s&p 500 index for the table 2 normality test jarque-bera statistic cv400 ds400 djim s&p 500 panel a: 1998–2008 skewness �0.3593 �0.3522 �0.4671 �0.6447 kurtosis 3.4129 3.4201 3.1852 3.8963 jarque-bera 3.4339 3.3636 4.5351 12.3306*** panel b: 2008–2012 cv400 ds400 djim s&p 500 skewness �0.6165 �0.6510 �0.7192 �0.7698 kurtosis 0.5306 0.4934 0.0830 0.8909 jarque-bera 4.5039 4.8466 6.1061 7.9100 note. ***significant at 1%. table 3 beta and beta-ssd index beta beta ssd panel a: 1998–2008 cv400 1.02679 1.00050 ds400 1.02454 1.00046 djim 1.08193 1.00171 s&p 500 1 1.0002 panel b: 2008–2012 cv400 0.9919 0.9931 ds400 0.9536 1.0274 djim 0.9926 0.9003 s&p 500 1 .9833 161f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 period 1998 through 2012. the analyses are carried using the traditional risk-adjusted measures of sharpe, treynor, and jensen. these performance measures are also adjusted to account for downside risk (barberis and huang, 2001; ang, bekaert, and liu, 2005). in contrast to other studies, the analysis concentrates on indices and not investment funds, eliminating issues of transaction costs of funds, the timing activities and the skill of the fund management. findings show a high level of correlation between the indices studied, mainly between cv400 and ds400, probably because of the similarity of the screening methods used. planners and investors should note that diversification between ethical investments is not recommended. additionally, the findings show a statistically different level of performance before and after the 2008 market collapse. a compromise may in fact be possible with socially responsible investing, although less so with faith based investing. finding, also show that when we compare the performance measure before and after 2008, we have reasons to reject the null of means equality. our results are different to those reported by sauer (1977), statman (2000, 2006), and goodmacher (2006) who found that the risk adjusted returns of srf funds are superior to those of other mutual funds. our results are similar to schroder (2006) and kreander (2004) who found sri stock indices do not exhibit a different level of risk-adjusted return than conventional benchmarks. however, many sri indices have a higher risk relative to the benchmarks. the exclusion of the financial sector and health care sectors contributed to both stability table 4 risk-adjusted performance measures index sharpe treynor jensen alpha sortino treynor (ssd) jensen (ssd) panel a: 1998–2008 cv400 0.004469 0.00021 0.00197 0.01132 0.00021 0.88893 ds400 0.00636 0.00028 0.00204 0.01532 0.00029 0.88958 djim 0.01945 0.00091 0.00283 0.04535 0.00098 0.88037 s&p 500 �.03918 �.00171 .000001 �.099010 �.001710 .000000 panel b: 2008–2012 cv400 �0.05475 �0.00315 0.00050 �0.06898 �0.00314 0.00050 ds400 �0.05944 �0.00341 0.00023 �0.07437 �0.00317 0.00049 djim �0.06924 �0.00439 �0.00074 �0.09165 �0.00484 �0.00107 s&p 500 �0.06508 �0.00371 �0.00006 �0.08078 �0.00365 0.00000 table 5 f-test and p level f-test p level whole sample 16.25345 .e�0 sharpe .09875 .96058 treynor .0445 .98748 jensen alpha .0209 .99588 sortino 11.4 .96 treynor (ssd) .03148 .99246 jensen (ssd) .01 .99 162 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 and better investment performance during the collapse in 2008; but seems to have contributed to under performance in the years following the collapse. although it is important for investors and planners to take into account faith and social preferences when investing, it is equally important that both understand the ramifications of making those choices and present them logically to their clients. although some trust funds and managed funds for religious organizations may require compliance with religious or social values, most investors express only a preference for this type of investing. a planner or investor can compensate, should the client agree, by investing in separate etfs that are specific to health care or financials. this effectively eliminates their exclusion from the portfolio but may not solve the investor preferences. however, in a study by beal (1998) return did not appear to be the main consideration of socially conscious investors, but a lower return may not be consistent with the required return that planners have established to reach retirement goals. the difficult decision for the planner is the strength of the preference versus appropriate diversification and exposure to growth sectors and the overall return and performance of the portfolio in light of long term retirement goals. notes 1. crsp (center of research in security prices) the university of chicago. references ang, a., bekaert, g., & liu, j. (2005). why stocks may disappoint. journal of financial economics, 76, 471–508. baillie, r. t., & degennaro, r. p. (1990). stock returns and volatility. journal of financial and quantitative analysis, 25, 203–214. balzer, l. (1994). measuring investment risk: a review. the journal of investing, 3, 47–58. barberis, n. c., & huang, m. (2001). mental accounting, loss aversion, and individual stock returns journal of finance, 56(4), 1247–1292. barndorff-nielsen, o., kinnebrock, s., & shephard, n. (2008). measuring down-side risk-realised semivariance. working paper. oxford-man institute of quantitative finance. barnett, m. l., & salomon, r. m. (2006). beyond dichotomy: the curvilinear relationship between social responsibility and financial performance. strategic management journal, 27, 1101–1122. baron, d. p. (2009). a positive theory of moral management, social pressure, and corporate social performance. journal of economics & management strategy, 18, 1, 7–43. bauer, r., koedijk, k., & otten, r. (2005). international evidence on ethical mutual fund performance and investment style. journal of banking & finance, 29, 1751–1767. bawa, v., & lindenberg, e. (1977). capital market equilibrium in a mean-lower partial moment framework. journal of financial economics, 5, 189–200. beal, d., & guyen, m. (1998). putting your money where your mouth is: a profile of ethical investors. financial services review, 7, 129–143. beer, f. m., estes, j. p., & munte, h. j. (2011). the performance of faith and ethical investment products: an empirical investigation of the last decade. journal of the academy of business and economics, 30, 101–124. bello, z. (2005). socially responsible investing and portfolio diversification. the journal of financial research, 28, 41–57. 163f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 bookstaber, r., & clarke, r. (1984). option portfolio strategies: measurement and evaluation. the journal of business, 57, 469–492. carhart, m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. chieffe, n., & lahey, k. e. (2009). helping clients select sri mutual funds and firms. journal of financial planning, 22, 60–70. clinebill, j. (2013). socially responsible investing and student managed investment funds: expanding investment education. financial services review, 22, 13–22. cowton, c. j. (1994). the development of ethical investment products, ethical conflicts in finance. oxford: blackwell. dybvig, p. h., & ross, s. a. (1985). differential information and performance measurement using a security market line. journal of finance, 40, 383–399. egan, w. j. (2007). the distribution of s&p 500 index returns. available at http://ssrn.com/abstract�955639. eling, m., & schuhmacher, f. (2007). does the choice of performance measure influence the evaluation of hedge funds? journal of banking and finance, 31, 2632–2647. elton, e. j., gruber, m. j., & blake, c. r. (1995). fundamental economic variables, expected returns, and bond fund performance. journal of finance, 40, 1229–1256. elton, e. j., gruber, m. j., & blake, c. r. (1996). the persistence of risk adjusted mutual fund performance. journal of business, 69, 133–157. elton, e. j., gruber, m. j., & goetzmann, w. n. (2006). modern portfolio theory and investment analysis (7th ed.). new york: john wiley and sons. fishburn, p., & kochenberger, g. (1979). two-piece. von neumann – morgernstern utility functions. decision sciences, 10, 503–518. fishburn, p. (1980). foundations of risk measurement: i. risk as a probability of loss. management science, 30, 396–406. fishburn, p. (1981). foundations of risk measurement: ii. effects of gains on risk. journal of mathematical psychology, 25, 226–242. girard, e., & hassan, m. k. (2005). faith-based ethical investing: the case of dow jones islamic indexes. fma papers. available at http://www.fma.org/slc/papers/faith-basedethicalinvesting.pdf. geczy, c., stambaugh, r. f., & levin, d. (2005). investing in socially responsible mutual funds. available at ssrn http://ssrn.com/abstract�416380 or doi: 10.2139/ssrn.416380. goetzmann, w. n., & ibbotson, r. g. (1994). do winners repeat? journal of portfolio management, 20, 9–18. goodmacher, ch. (2006). the effects of ethical screens on socially responsible fund performance. working paper, oxford university. gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51, 783–810. hakim, s., & rashidian, m. (2002). risk & return of islamic stock market indexes. economic research forum annual meetings, sharjah, uae hamilton, s., jo, h., & statman, m. (1993). doing well while doing good? the investment performance of socially responsible mutual funds. financial analysts journal, 49, 62–66. harlow, w., & rao, r. (1989). asset-pricing in a generalized mean-lower partial moment framework: theory and evidence. journal of financial and quantitative analysis, 24, 285–311. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974–1988. journal of finance, 48, 93–130. henriksson, r., & merton, r. (1981). on market timing and investment performance. ii statistical procedures for evaluating forecasting skills. journal of business, 54, 513–533. henriksson, r. (1984). market timing and mutual fund performance: an empirical investigation. journal of business, 57, 73–96. hogan, w., & warren, j. (1972). computation of the efficient boundary in the e-s portfolio selection model. journal of financial and quantitative analysis, 7, 1881–1896. hsu, b., reynolds, a., hackett, c., & gibbon, j. (2008). estimating the religious composition of all nation: an 164 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 empirical assessment of the world christian database. journal for the scientific study of religion, 47(4), 678–693. huij, j. (2007). new insights into mutual funds: performance and family strategies. erim series research in management no. 99. available at ssrn http://ssrn.com/abstract�961819. hull, c. e., & rothenberg, s. (2008). firm performance: the interactions of corporate social performance with innovation and industry differentiation. strategic management journal, 29, 781–789. hussein, m., & omran, m. (2005). ethical investment revisited evidence from dow jones islamic indexes. the journal of investing, 14(3), 105–126. jarque, c. m., & bera, a. k. (1980). efficient tests for normality, hemoscedasticity, and serial independence of regression residuals. economics letters, 6, 255–259. jensen, m. c. (1968). the performance of mutual funds in the period 1945–1964. journal of finance, 23, 389–416. jones, s., van der laan, s., frost, g., & loftus, j. (2008). the investment performance of socially responsible investment funds in australia. journal of business ethics, 80, 181–203. kahnemann, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–291. kempf, a., & osthoff, p. (2007). the effect of socially responsible investing on portfolio performance. european financial management, 13(5), 908–922. kenourgios, d., & petropoulos, i. (2005). the persistence of mutual funds performance: evidence from the u.k. stock market. ekonomia, cyprus economic society and university of cyprus, 7, 121–138. kreander, n., mcphail, k., & molyneaux, d. (2004). god’s fund managers: a critical study of stock market investment practices of the church of england and uk methodist. accounting, auditing, & accountability journal, 17, 408. kurtz, l. (1997). no effect, or no net effect? studies on socially responsible investing. journal of investing, 6, 37. kurtz, l., & dibartolomeo, d. (2005). the kld catholic values 400 index. the journal of investing, 14, 101–104. litner, j. (1965). the valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. review of economics and statistics, 47, 13–37. lyn, e. o., & zychowicz, e. j. (2010). the impact of faith-based screens on investment performance. the journal of investing, 19(3), 136–143. luce, r. (1980). several possible measures of risk. theory and decision, 12, 217–228. mackey, a., mackey, t. b., & barney, j. b. (2007). corporate social responsibility and firm performance: investor preferences and corporate strategies. academy of management review, 32, 817–835. maheu, j. m., & mccurdy, t. h. (2009). how useful are historical data for forecasting the long-run equity return distribution? journal of business and economic statistics, 27, 95–112. malhotra, d. k., & dania, a. (2012 october). an empirical examination of the performance of faith based investing: the case of islamic index funds. published in the proceedings of academy of financial services conference, san antonio, tx. margolis, j. d., & walsh, j. p. (2001). people and profits? the search for a link between a company’s social and financial performance. mahwah, nj: lawrence erlbaum. markowitz, h. (1952). portfolio selection. journal of finance, 7, 77–91. mclachlan, j., & gardner, j. (2004). a comparison of socially responsible and conventional investors. journal of business ethics, 52, 11–25. mossin, j. (1969). security pricing and investment criteria in competitive markets. american economic review, 59, 749–756. nantell, t. j., price, k., & price, b. (1982). mean-lower partial moment asset pricing model: some empirical evidence. journal of financial and quantitative analysis, 17(5), 763–782. doi: 10.2307/2330861. pedersen, c., & satchell, s. (2002). on the foundation of performance measures under asymmetric returns. quantitative finance, 3, 217–223. 165f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 pedersen, c., & rudholm-alfvin, t. (2003). selecting a risk-adjusted shareholder performance measure. journal of asset management, 4, 152–172. price, k., price, b., & nantell, t. j. (1982). variance and lower partial moment measures of systematic risk: some analytical and empirical results. journal of finance, 37, 843–55. renneboog, l., horst, j. t., & zhang, c. (2008). socially responsible investments: institutional aspects, performance, and investor behaviour. journal of banking and finance, 32, 1723–1742. reyes, m. g., & grieb, t. (1998). the external performance of socially-responsible mutual funds. american business review, 16, 1–7. richardson, b. j. (2007). protecting indigenous people through socially responsible investment. indigenous law journal, 6(1), 1–30. rouwenhorst, k. g. (2004). the origins of mutual funds. yale icf working paper no. 04–48. available at ssrn http://ssrn.com/abstract�636146. sadeghi, m. (2008). financial performance of shariah-compliant investment: evidence from malaysian stock market. international research journal of finance and economics, 20, 15–26. sarin, r. (1987). extensions of luce’s measure of risk. theory and decision, 22, 125–141. satchell, s. (1996). lower partial moment capital asset pricing model: a re-examination. ifr birkbeck college discussion paper, 20. sauer, d. a. (1997). the impact of social-responsibility screens on investment performance: evidence from the domini 400 social index and domini equity mutual fund. review of financial economics, 6, 137–149. scatizzi, c. (2010). faith based investing. american association of individual investors. available at http:// www.aaii.com/financial-planning/article/faith-based-investing-article?adv�yes. schroder, m. (2004). the performance of socially responsible investments: investment funds and indices. financial markets and portfolio management, 18, 122–142. sen, a. (1999). development as freedom. oxford: oxford university press. sharpe, w. (1964). capital asset prices: a theory of capital market equilibrium under conditions of risk. journal of finance, 19, 425–442. sharpe, w. (1966). mutual fund performance. journal of business, 39, 119–138. sharpe, w., alexander, g. j., & bailey, j. w. (1998). investments (6th ed.). englewood cliffs, nj: prentice hall. smith, t. (2002). religious diversity in america: the emergence of muslims, buddhists, hindus, and others. journal for the scientific study of religion, 41, 577–585. smith, t. (2002). the muslim population of the united states: the methodology of estimates. public opinion quarterly, 66, 404–417. social investment forum. (2005). 2005 report on socially responsible investing trends in the united states: 10-year review. available at http://www.socialinvest.org. sortino, f., & price, l. (1994). performance measurement in a downside risk framework. journal of investing, 3, 59–65. sparkes, r. (2001). ethical investment: whose ethics, which investment? business ethics: a european review, 10, 194–205. sparkes, r., & cowton, c. j. (2004). the maturing of socially responsible investment: a review of the developing link with corporate social responsibility. journal of business ethics, 52, 45–57. statman, m. (2000). socially responsible mutual funds. financial analysts journal, 56, 30–39. statman, m. (2006). socially responsible indexes: composition, performance and tracking error. the journal of portfolio management, 32, 100–109. stone, b. k., guerard, g. b., gultekin, m. n., & adams, g. (2001). socially responsible investment screening: strong evidence of no significant cost for actively managed portfolios. social investment forum and co-op america. available at http://www.socialinvest.org/pdf/research/moskowitz/2001%20honorable%20mention %20-%20moskowitz.pdf thaler, r. (1993). advances in behavioral finance. new york, ny: russell sage foundation. todd, m. j., & barrett, d. b. (2004). quantifying alternate futures of religion and religious. futures, 36, 947–960. treynor, j. l. (1966). how to rate management of investment funds. harvard business review, 43, 63–75. 166 f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 van liedekerke, l., de moor, l., & vanwalleghem, d. (2007). risk-return of belgian sri funds. available at ssrn http://ssrn.com/abstract�1088979. waddock, s. a., & grovers, s. b. (1997). quality of management and quality of stakeholder relations: are they synonyms? business and society, 36, 250–279, 1997. zheng, l. (1999). is money smart? a study of mutual fund investors’ fund selection ability. the journal of finance, 54, 901–933. 167f.m. beer, j.p. estes, c. deshayes / financial services review 23 (2014) 151–167 academy of financial services officers president swarn chatterjee university of georgia president-elect janine scott shepherd university executive vice president-program inga chira california state university, northridge vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations philip gibson winthrop university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president thomas coe quinnipiac university editor, financial services review stuart michelson stetson university directors shawn brayman planplus global charles chaffin cfp board of standards victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university terrance martin university of texas–rio grande valley past presidents robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial.org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2018 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 27, no. 4, 2018 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international wade d. pfau the american college lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2018 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. factors related to the risk tolerance of households in china and the united states: implications for the future of financial markets in china sherman d. hannaa,*, kyoung tae kimb, lishu zhangc ahuman sciences department, ohio state university, 1787 neil avenue, columbus, oh 43210, usa bdepartment of consumer sciences, 312 adams hall, box 870158, university of alabama, tuscaloosa, al 35487, usa cdepartment of economics, shenzhen university, 3688 nanhai avenue, nanshan, shenzhen, guangdong, china abstract we analyzed factors related to the financial risk tolerance of chinese households, using the 2011 china household finance survey (chfs). the risk tolerance question was similar to one in the u.s. survey of consumer finances (scf), and we found that chfs respondents had slightly higher risk tolerance than scf respondents, but the percentage of households with stock assets was 9%, compared with 49% in the united states. our multivariate analyses found many household characteristics in the chfs had effects on risk tolerance similar to those found in the 2013 scf. we discuss implications for the future of chinese investment markets. © 2018 academy of financial services. all rights reserved. jel classifications: d14; g11 keywords: risk tolerance; individual investing; china household finance survey; survey of consumer finances 1. introduction risk tolerance is an important topic for household financial choices because it affects many types of household decisions, including portfolio selection, insurance choices, and saving decisions. there have been many normative analyses on the relationship between risk tolerance and * corresponding author. tel.: �1-614-292-4584; fax: �1-614-292-4339. e-mail address: hanna.1@osu.edu (s.d. hanna) financial services review 27 (2018) 279-302 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. optimal financial decisions, for example, campbell and viceira (2002). standard advice for households, based on normative analyses by economists, is that they should consider their risk tolerance in making investment decisions. there are extensive studies on the relationships between the risk tolerance and decision-making. grable and joo (2004) suggested that one’s subjective financial risk tolerance is largely determined by social factors such as demographic, socioeconomic, and attitudinal characteristics. households willing to take some risk have been found to be more likely to own high return investment (shin and hanna, 2015) and stocks (wang and hanna, 2007). kim and hanna (2015) found that those willing to take average or above average risk had higher likelihoods of having an adequate retirement than those unwilling to take any risk. kim, wilmarth, and choi (2016) found that as a household’s level of risk tolerance increased, outstanding credit card balance and installment loan balances increased. lastly, households not willing to take any investment risk have been found to be more likely to own life insurance (gutter and hatcher, 2008). there are many studies on the factors related to the financial risk tolerance of u.s. households. however, the risk tolerance of chinese households has been understudied. in one of the few studies that attempted to compare the risk tolerance of u.s. respondents (fan and xiao, 2006), the authors noted that their study was exploratory and their chinese respondents were not representative of all of china. they concluded that there was a need for future research using more comparable data. the only other studies we found that attempted to directly compare the risk tolerance of u.s. and chinese respondents used small student samples, for instance, in terpstra-tong and terpstra (2013), the u.s. sample included only 70 students at one university, and 452 students at universities in hong kong and macao. those authors noted that despite that fact that china had a rapidly growing capital market, little was known about investors in china, and the increasing middle class represented a great opportunity for financial services companies. the primary research objective of our study is to ascertain the factors that affect respondent financial risk tolerance, in a large sample representative of all households in china. a secondary objective is to compare the effects of household characteristics on risk tolerance in china to the results of a similar analysis of u.s. data. lv (2007) stated that risk tolerance is not commonly considered by the wealth management industry in china in making recommendations for clients. in the united states, low proportions of households own stocks directly, but many own through mutual funds, especially inside retirement accounts controlled by workers, the proportion of households with direct or indirect ownership of stocks increased from 32% in 1989 to over 50% by 2001, and has remained at about that level since, with a decrease after the financial crisis of 2008 (fig. 1). as we will demonstrate (table 5), only about 9% of households in china in 2011 had direct or indirect ownership of stocks, which is not surprising given much lower income levels in china compared with the united states, but the high level of economic growth in china makes it likely that china will approach u.s. levels of household income in the coming decades. will future investment patterns of households in china be similar to the patterns in the united states? comparison of factors related to risk tolerance in china and the united states may provide insights into future investment patterns in china. in the united states, government and employer policies related to household financial decisions, including optimal settings for “nudge” policies to encourage households to make better decisions, take 280 s.d. hanna et al. / financial services review 27 (2018) 279-302 household risk tolerance levels into account, and it may be appropriate to consider risk tolerance levels of households in china for such policies. 2. literature review 2.1. why measure risk tolerance? recommendations for investment portfolios often are based on the expected utility model, which despite criticisms that it represents actual household behavior poorly, is the best normative model available (hanna and chen, 1997; schoemaker, 1982). economists have used expected utility analysis to evaluate whether individuals have made optimal financial decisions (e.g., calvet, campbell, and sodini, 2007; cocco, gomes, and maenhout, 2005). a crucial part of normative household finance is a plausible estimation of utility function parameters, as discussed by brown and poterba (2000). the expected utility model, as a normative model for decision making involving risky choices (hanna and chen, 1997), includes risk aversion as an important parameter of individual utility functions. pratt (1964) had one of the earliest presentations of mathematical forms of utility under risk, and arrow (1971) and deaton and muellbauer (1980) are among those with expositions of the model. in particular, arrow (1971, p. 94) presented the mathematical forms for absolute risk aversion and relative risk aversion. hanna and chen (1997) noted that it is reasonable to assume that most households are risk averse, as risk seeking utility functions imply the willingness to accept a chance of zero consumption, and even for households with plausible levels of risk aversion, optimal portfolios include some risky investments. bailey, olson, and wonnacott (1980) concluded that the phenomenon of people engaging in gambling is better explained by direct utility from the experience, rather than from a risk-seeking level of risk aversion. 10 15 20 25 30 35 40 45 50 55 1989 1992 1995 1998 2001 2004 2007 2010 2013 pe rc en t o f h ou se ho ld s ow ni ng year direct or indirect holdings direct holdings fig. 1. percent of u.s. households owning stocks directly and/or indirectly, 1989–2013. created by authors, based on data from federal reserve board, https://www.federalreserve.gov/econres/files/scf2013_tables_ internal_real.xls. 281s.d. hanna et al. / financial services review 27 (2018) 279-302 a relative risk aversion level of zero is consistent with risk neutrality, so the optimal choice is based on maximizing expected value. for household investment decisions, both relative risk aversion and the share of total household wealth that is human wealth are important factors in determining optimal allocations of the financial investment portfolio. campbell and viceira (2002, p. 188) presented analyses showing that for retired households, the optimal allocation to stocks ranges from 80% for relative risk aversion of 2, to 13% for a relative risk aversion of 12. cocco (2005), gomes and michaelides (2005), and horneff, maurer, and stamos (2008) also presented analyses showing how optimal investment choices depend on the household’s level of risk aversion. some normative articles such as hubbard, skinner, and zeldes (1994), scholz, seshardi, and khitatrakun (2006), and gomes and michaelides, (2003) derived optimal savings levels under risk with several different levels of relative risk aversion. hubbard et al. (1994) derived optimal savings levels and examined predictions of a life-cycle simulation model subject to uninsured idiosyncratic risk (risky earnings, medical expense, and life span). campbell and cocco (2003) showed how the optimal choice between a fixed rate and a variable rate mortgage loan depended on the level of the household’s risk aversion. clearly, the prescriptions of normative household finance (campbell, 2006) depend on the level of risk aversion assumed, and analysis of public policies related to household financial decisions also depend on the level of risk aversion assumed. risk tolerance is the inverse of risk aversion (barsky et al., 1997). is there heterogeneity in levels of household risk tolerance among chinese households? while there is considerable research about that question for u.s. households, the evidence is limited for chinese households. as terpstra-tong and terpstra (2013) noted, little is known about investors in china, but comparisons of risk tolerance levels and factors related to risk tolerance between china and the united states may provide insights into future financial markets in china. ownership of stock assets in the united states was somewhat limited before the 1990s (haliassos and bertaut, 1995), which seemed puzzling to economists given the equity premium (siegel and thaler, 1997). siegel and thaler discussed whether the equity premium was a puzzle in countries other than the united states, for instance, in japan and germany. they concluded that even in countries with wartime destruction of the value of stocks, there still was an advantage of investment in stock assets over the long run, compared with bonds and cash equivalents. therefore, the advantage of investments in stocks over alternative financial investments (e.g., hanna and chen, 1997) may apply to china despite the difference in the economic history of the two countries. 2.2. measuring risk tolerance according to hanna, gutter, and fan (2001), there are at least four different methods of measuring risk tolerance: (1) asking about investment choices, (2) asking a combination of investment and subjective questions, (3) assessing actual behavior, and (4) asking hypothetical questions with carefully specified scenarios. below, we briefly review selected articles using each method. an example of the first method is in the survey of consumer finances (scf), a nationally representative survey conducted every three years in the united states (bricker et al., 2014), which has since 1983 included a risk tolerance measure. yao, hanna, and lindamood (2004) 282 s.d. hanna et al. / financial services review 27 (2018) 279-302 was the first study published in a journal that provided the history of the scf risk tolerance measure. the question in the scf simply asks respondents whether they were willing to take substantial risk to make a substantial return, above average return, average risk to make an average return, or no risk when making investments. yao et al. (2004) presented a conceptual model that suggested that the answer to the scf risk tolerance question might be affected not only by the level of respondent risk aversion, but also perceptions influenced by recent stock market changes, and by one’s life cycle stage. the scf risk tolerance measure has been used in many studies both as a dependent variable and an independent variable. grable and lytton (2001) listed six previous studies using it as a dependent variable and 11 studies using it as an independent variable. we estimate that at least 50 additional publications have used the scf risk tolerance measure. we do not discuss all studies that have used the scf risk tolerance measure, focusing instead on studies that used approaches consistent with the approach used in this article. an example of the second method is in grable and lytton (1999), who presented a measure based on 13 questions, based both on investment choices and financial beliefs, feelings, needs, and aspirations. the grable-lytton measure has been used in dozens of studies, but it has not been tested in a nationally representative survey in the united states. some of the components are situational, for instance, would you cancel a “once-in-alifetime” vacation if you lost your job three weeks before the vacation? obviously, a person’s answer would depend on other resources and other household income, the person’s age, and other factors. grable and lytton (2001) reported that their risk tolerance measure has good validity. an example of the third method, assessing actual behavior to measure risk tolerance, is wang and hanna (1997), who used the theoretical framework proposed by friend and blume (1975) to infer risk aversion from the proportion of total wealth invested in risky assets. attanasio, banks, and tanner (2002)’s equity premium estimates were also based on actual investment choices. however, this method could be problematic. as hanna, waller, and finke (2008) noted, investment choices may be based on risk tolerance, but also expectations and risk capacity. they presented a model of factors affecting investment choices, and stated “the scf measure, which asks respondents whether they are willing to take greater risk to achieve greater returns, may be an imperfect measure of risk tolerance, as people may be thinking of all four elements on the left side of their model in stating how much investment risk they would be willing to take.” based on this observation, we have modified the model presented in hanna et al. (2008) in our fig. 2, suggesting that answers to the scf risk risk aversion risk capacity feelings about volatility expectations investment choices fig. 2. conceptual model of investment choices involving risk modified version of model in hanna et al. (2008). 283s.d. hanna et al. / financial services review 27 (2018) 279-302 tolerance question might be affected by all of the factors on the left side, risk aversion, risk capacity, expectations, and feelings about volatility. for instance, for two individuals with identical risk aversion, if one expects the stock market to keep increasing and the other expects a 10 year slump, they might choose different portfolios, or, for those with no current risky investments, different potential portfolios as reflected in the choices in the scf risk tolerance question. the fourth method, using hypothetical questions to infer risk tolerance, may control for other factors that determine actual investment choices and, therefore, may reveal each respondent’s true risk aversion. for example, kimball (1988) presented hypothetical income gambles to infer relative risk aversion. barsky et al. (1997) used the framework proposed by kimball (1988) to include hypothetical income gamble questions in the health and retirement study survey. the hypothetical scenarios presented to respondents attempted to control for other factors such as household resources, and if a constant relative risk aversion utility function is assumed, relative risk aversion levels can be inferred from respondent choices. considerable individual variation is present, for instance, hanna and lindamood (2004) found an interquartile range from 2 to 6, and in a sample of older adults (fang, hanna, and chatterjee, 2013), 39% had levels of relative risk aversion over 7.5 and over 40% had levels under 3.8. hanna and lindamood (2004) and hanna et al. (2001) noted some flaws in the hypothetical income gambles used in the u.s. health and retirement study dataset reported in fang et al. (2013), which might have led to downward bias in estimates of relative risk aversion. even though there are a variety of ways to measure risk tolerance, only one, the scf measure, has been used in nationally representative samples in the united states over a period of more than 30 years. furthermore, the scf measure is based on one simple question, rather than a complex question based on hypothetical income gambles (barsky et al., 1997) or a combination of many questions that include investment experiences and risk capacity (grable and lytton, 1999). most national household surveys have a very limited capacity for additional questions, so it is unlikely that the grable-lytton measure will be used in such surveys. therefore, we focus on the scf question, since a nationally representative survey in china has used a very similar question. 2.3. empirical research on factors that affect respondent risk tolerance there have been many empirical studies with risk tolerance as a dependent variable, with a large variety of measures of risk tolerance. in this section we focus on selected empirical studies using the scf risk tolerance variable as a dependent variable. there have been dozens of publications reporting analyses of risk tolerance in the united states, but we focus on three studies using scf datasets from 1983 to 2004 (table 1). sung and hanna (1996) used the 1992 scf to test the effect of financial variables and demographic variables on risk tolerance for a subsample of households with employed heads. they used a logistic regression with a dependent variable of whether the respondent was willing to take some risk (average, above average, or substantial) versus no risk. yao et al. (2004) analyzed a combination of the 1983 to 2001 scf datasets, and noted that even though the scf risk tolerance measure had four levels, a statistical test indicated that an ordered logit was not 284 s.d. hanna et al. / financial services review 27 (2018) 279-302 t ab le 1 se le ct ed st ud ie s on fa ct or s re la te d to ri sk to le ra nc e: u .s . st ud ie s us in g na tio na lly re pr es en ta tiv e sa m pl es ; c hi ne se st ud ie s us in g va ri ou s sa m pl es st ud y sa m pl e d ep en de nt va ri ab le (a na ly si s) se le ct ed ef fe ct s in m ul tiv ar ia te an al ys is of re sp on de nt ch ar ac te ri st ic s se le ct ed ef fe ct s in m ul tiv ar ia te an al ys is of ho us eh ol d ch ar ac te ri st ic s su ng an d h an na (1 99 6) u . s. sc f, 19 92 , ho us eh ol ds w ith an em pl oy ed re sp on de nt , n � 2, 65 9 sc f ri sk to le ra nc e (l og is tic re gr es si on ) y ea rs to re tir em en t: po si tiv e n on -i nv es tm en t in co m e: po si tiv e l iq ui d as se ts � 3 m on th s in co m e: po si tiv e e du ca tio n: po si tiv e se lf -e m pl oy ed : po si tiv e r ac ia l/e th ni c: w hi te � h is pa ni c, w hi te � a si an /o th er h ou se ho ld si ze : n s h om eo w ne rs hi p: n s si ng le fe m al e: n eg at iv e y ao , h an na , an d l in da m oo d (2 00 4) u . s. sc f, 19 83 –2 00 1, al l ho us eh ol ds , n � 24 ,1 32 sc f ri sk to le ra nc e (l og is tic re gr es si on ) a ge : n eg at iv e c hi ld re n � 18 : n eg at iv e e du ca tio n: po si tiv e in co m e: po si tiv e r ac ia l/e th ni c w hi te : po si tiv e se lf -e m pl oy ed : po si tiv e m al e: po si tiv e h om eo w ne r: n s m ar ri ed : po si tiv e po or he al th : n eg at iv e w an g an d h an na (2 00 7) u . s. sc f, 19 92 –2 00 4, al l ho us eh ol ds , n � 21 ,4 71 sc f ri sk to le ra nc e (l og is tic re gr es si on ) a ge : n eg at iv e c hi ld re n � 18 : n eg at iv e e du ca tio n: po si tiv e in co m e: po si tiv e r ac ia l/e th ni c w hi te : po si tiv e h av e bu si ne ss : po si tiv e m al e: po si tiv e h om eo w ne r: po si tiv e m ar ri ed : n s fa n an d x ia o (2 00 6) w or ke rs in g ua ng zh ou , c hi na , 19 98 , n � 47 0; u .s . sc f, 19 98 , no nfa rm fu lltim e w or ke r ho us eh ol ds , n � 2, 67 1 sc f ri sk to le ra nc e (m ul tino m in al lo gi st ic re gr es si on ) a ge : n eg at iv e in bo th in co m e: po si tiv e in bo th e du ca tio n: po si tiv e in bo th se lf -e m pl oy ed : po si tiv e in c hi ne se r ac ia l/e th ni c w hi te : po si tiv e h ou se ho ld si ze : n s m al e: po si tiv e in u ni te d st at es , n s in c hi ne se h om eo w ne rs hi p: po si tiv e in bo th si ng le fe m al e: n eg at iv e m ar ri ed : po si tiv e in u ni te d st at es , n s in c hi ne se (b ot h � si gn ifi ca nt ef fe ct in bo th c hi ne se an d u .s . an al ys es ) t er pe rs tr at on g an d t er pe rs tr a (2 01 3) st ud en ts in u ni te d st at es , h on g k on g, m ac ao , n � 52 2 g ra bl e m ea su re (o l s re gr es si on ) c hi ne se re sp on de nt s no t si gn ifi ca nt ly di ff er en t fr om u .s . re sp on de nt s (m od el 3) n a a ge : n eg at iv e m al e: po si tiv e py le s, l i, w u, an d d ol vi n (2 01 6) su rv ey of st ud en ts at a u ni te d st at es un iv er si ty (n � 21 5) an d 2 c hi ne se un iv er si tie s (n -6 20 ) sc f m ea su re (l og it) an d g ra bl e m ea su re (o l s re gr es si on ) c hi ne se re sp on de nt s m or e ri sk to le ra nt th an u .s . re sp on de nt s n a ex ce pt fo r pa re nt ed uc at io n an d in co m e l ow pa re nt ed uc at io n: po si tiv e fo r sc f m ea su re , n s fo r g ra bl e m ea su re fe m al e: n eg at iv e fo r bo th m ea su re s h ig h pa re nt in co m e: n eg at iv e fo r sc f m ea su re , n s fo r g ra bl e m ea su re l ow pa re nt in co m e: n s fo r sc f m ea su re , ne ga tiv e fo r g ra bl e m ea su re fo r st ud ie s an al yz in g sc f ri sk to le ra nc e va ri ab le , ef fe ct s fo r so m e ri sk vs . no ri sk re po rt ed un le ss ot he rw is e no te d. u . s. sc f � u .s . su rv ey of c on su m er fi na nc es ; n s � no t si gn ifi ca nt ly di ff er en t fr om 0 at 5% le ve l. 285s.d. hanna et al. / financial services review 27 (2018) 279-302 appropriate, so they used a cumulative logit model, with three separate logistic regressions on some risk, high risk, and substantial risk. wang and hanna (2007) analyzed a combination of the 1992–2004 scf datasets, using the same cumulative logit approach used by yao et al. (2004). wang and hanna obtained results generally similar to those found by yao et al., including a significant negative effect of age squared, with the combined effect of age and age squared implying that at mean values of other variables, the likelihood of being willing to take some risk was about 42 percentage points lower for a respondent aged 80 than for one aged 25. they estimated that a household with a head with a bachelor’s degree would be almost 38 percentage points more likely to be willing to take some risk than an otherwise similar household with less than a high school degree. households with a female respondent had a predicted likelihood of being willing to take some risk 14 percentage points lower than an otherwise similar household with a male respondent. income was positively related to risk tolerance. 2.4. empirical research on risk tolerance in china very few studies have directly addressed financial risk tolerance in china, and even fewer studies have attempted to directly compare risk tolerance in china to risk tolerance in the united states. fan and xiao (2006) compared the risk tolerance of chinese workers and comparable u.s. respondents in 1998 (table 1). this research used a self-collected data of 407 workers from four categories of enterprises in guangzhou, a major city in china. they reported that a higher proportion of u.s. workers were willing to take some risk (72%) than were chinese workers (65%). the multivariate analyses showed cross-sectional patterns of age being negatively related to risk tolerance, and education and income being positively related to risk tolerance, with self-employed workers being more risk tolerant than employees. homeowners had a higher likelihood of being willing to take average risk but a lower likelihood of being willing to take substantial risk than renters. they also found that more risk-tolerant individuals were more likely to own stocks. fan and xiao (2006) noted that their chinese sample had limited generalizability to all of china. terpstra-tong and terpstra (2013) compared the financial risk tolerance of three samples of students at universities in hong kong and macao and at a university in the united states based on a self-collected data, with a total of 522 responses (table 1). in a multivariate model controlling for various respondent characteristics, there were no significant differences in the responses of chinese respondents and u.s. respondents. however, the sample from the united states consisted of only 70 students from one university. the authors attempted to compare residents of china attending a university in macao to students from hong kong and macao. men had higher risk tolerance than women in hong kong but not in the other samples, and age was negatively related to risk tolerance. pyles et al. (2016) surveyed 215 students at a u.s. university and 620 students at universities in hong kong and macao (table 1). their questions included both the scf measure and the grable-lytton measure. the chinese students were more risk tolerant than otherwise similar students in the united states. women were less risk tolerant than otherwise similar male students. we found one other journal article (lv, 2007) reporting analyses of 286 s.d. hanna et al. / financial services review 27 (2018) 279-302 risk tolerance in china, but it was based on a survey with 88 bank customers in china, so we did not include it in table 1. 2.5. summary of consistent empirical effects most research, including research based on u.s. household surveys, has found men have higher risk tolerance than women. research using household surveys has generally found that risk tolerance is positively related to education and income. white respondents in the u.s. have been found to be more likely to be willing to take some risk than otherwise similar respondents with other racial/ethnic identifications. households with a self-employed head or with a business have been found to be more willing to take risk. the effect of being a single head versus being a couple has been found to have inconsistent effects in the united states and in china. homeowners have been found to be more likely to be willing to take some risk. it is difficult to compare the results of household surveys to the results of student surveys, as the age and education levels of student respondents do not vary much, and the most recent analysis of household survey data from china was for a 1998 survey of workers in one city, so analyses of more recent household survey data from both countries should provide valuable insights into factors affecting risk tolerance. in the next section, we present a conceptual model for factors related to risk tolerance. it is plausible that the respondent and household characteristics are related to risk preferences and to risk capacity. selection of the variables for our empirical analyses was based on our conceptual model, but also by variables used in previous empirical research. thus, in our study, we control for respondent and household characteristics, including gender, education level, age, and household income to identify the factors that could have significant effect on chinese respondents’ risk tolerance. the main contribution of this research is to use the china household finance survey (chfs), a nationwide dataset to test the risk tolerance of chinese households. in addition, we included ethnicity in our regression model, han versus one of the other 56 ethnicities recognized in china, because of the consistent result in u.s. research that white respondents are more likely to be willing to take some risk than otherwise similar respondents with other racial/ethnic identities. moreover, we present a comparison of the u.s. respondents and chinese respondents in terms of the distribution of risk tolerance and present multivariate analyses of u.s. risk tolerance parallel to our multivariate analyses of risk tolerance of chinese households. 3. conceptual model and research hypotheses risk tolerance is sometimes assumed to be a preference, and economic theory provides limited insights into the determinants of preferences (yao et al., 2004). previous authors examining factors related to levels of risk tolerance have used psychological and sociological models for conceptual models. for instance, yao et al. (2004) used a psychological theory that people are affected more strongly by recent events to explore patterns of risk tolerance over time. yao et al. also presented a conceptual model proposing that willingness to take investment risk (i. e., the scf risk tolerance measure) 287s.d. hanna et al. / financial services review 27 (2018) 279-302 is influenced by risk aversion, recent stock market changes, and risk capacity as related to life cycle stage, and so forth. a risk tolerance measure that is based on a respondent’s willingness to invest in risky assets may reflect many factors, not just risk tolerance as the inverse of the economist’s concept of the inverse of risk aversion. fig. 2 shows a modified version of the model presented by hanna et al. (2008), who suggested that a household’s actual investment choices may depend on risk tolerance, and other factors. some households do not have funds to make any investments, but the scf and chfs risk tolerance questions could reflect potential investment choices. statistical analyses with controls for many household characteristics may provide some insights into variations of risk tolerance responses. based on the model in fig. 2, and also on discussion in shin and hanna (2015), we created the hypotheses shown in table 2. campbell and viceira (2002) and hanna and chen (1997) presented normative analyses based on historic returns on financial assets, and demonstrated that the optimal proportion of stock assets in the financial investment portfolio depended on risk capacity, which is related to human wealth. somebody with a particular level of risk aversion (as defined by economists) should decrease the proportion of the portfolio in stocks as retirement approaches, even if there is no change in risk aversion. therefore, the willingness to take financial risk may decrease with age, even if risk aversion does not change, because of the decreasing proportion of household wealth in human wealth. the negative relationship between age and risk tolerance has been found by almost all researchers in studies of households with a wide age range, so we expected to find the same pattern. in most empirical studies of households of risk tolerance, men have been found to be more risk tolerant than women. yao and hanna (2005) found this pattern in a combined sample of the 1983–2001 scf datasets, with unmarried men being most willing to take some risk, followed by married men, unmarried women, and married women. they suggested that the differences could be because of genetic factors, socialization, and culture. controlling for other factors, women should be willing to invest somewhat more aggressively than men because of the longer life expectancies of women (ho, milevsky, and robinson, 1994), but most research has found that men have higher risk tolerance than women, and we expect to find a similar pattern in both the chfs and the scf. marital status has been found to have mixed effects in previous research, and as yao and hanna (2005) discuss the contradictory results they obtained for marital status and respondent gender. a couple would have a longer life expectancy than a single person of the same age, so objectively should be more willing to take investment risk, but there may be a selection effect, with more risk tolerant people being more likely to be single. we do not have a specific expectation for the effect of marital status on risk tolerance, so it is listed as a null hypothesis in table 2. education has been found to be related to risk tolerance in most empirical studies, though if income is controlled, the cause of this relationship is not immediately apparent. in terms of the scf and chfs risk tolerance measures, in fig. 2, willingness to take financial risks might be reflected in actual investment choices for some households, but for households without any investments, the answer to the risk tolerance questions might reflect potential investment choices. haliassos and bertaut (1995) posed the question “why do so few hold 288 s.d. hanna et al. / financial services review 27 (2018) 279-302 stocks?” one of their answers they discussed is the information costs of investing in stocks, and education could be related to the cost of understanding information about investing. more educated respondents may also be more future-oriented, having been willing to defer earning to obtain more education. therefore, we expect a positive relationship between education and risk tolerance. employment status could have an effect on risk tolerance, both in terms of risk capacity and also exposure to information about investments, for instance, with employer-based retirement plans in the united states. there might also be a selection effect for selfemployment, as those choosing self-employment over being an employee may tend to be table 2 hypotheses and theoretical justifications variables category expected effect theoretical justification age � as age increases, the investment horizon shortens, and human wealth decreases. gender male no rigorous theoretical justification, though virtually all empirical studies have shown that women have lower risk tolerance scores than men. female � education less than high school education level is related to financial knowledge, and cognitive burdens and information costs are required in investing. even after controlling for income, there might be a positive relationship between education and risk tolerance. high school � some college without degree � bachelor degree � employment status employee without the stability of a salary, respondent would be less willing to invest in a high return asset. however, self-selection might result in different preferences for self-employed. self-employed other than farmer � farmer � retired � other job status � health status poor health if household expects poor health in the future, it might have a greater need for more stable funds to cover medical expenses. fair health � good health � excellent health � race/ethnicity han controlling for other factors including income and age, there should not be any differences.non-han 0 home ownership yes � renters are less likely to be willing to invest in high return assets because they would choose to invest in assets with low volatility to buy a house in the future. no household type couple couple household might have higher expected lifetime income than single households, and this leads to higher likelihood to be willing to hold high return investments. single � income � higher income households have more resources to invest in high return investments. financial assets � those with more financial assets are more able to cover short term expenses, so will be more willing to invest in high return assets (�) is positive effect, (�) is negative effect, 0 is no effect. the reference category is indicated in bold. 289s.d. hanna et al. / financial services review 27 (2018) 279-302 more risk tolerant. empirical studies in the united states have consistently found that self-employed respondents are more risk tolerant than respondents who are employees (e.g., yao et al., 2004), and we expect to find a similar pattern. it is plausible that health would be related to risk tolerance because poor health may shorten the investment horizon, given the possibility of medical expenses. in u.s. research, white respondents have been found to be more willing to take some risk (vs. no risk) than those identifying as black or hispanic. yao et al. (2005) discussed both cultural differences and differing levels of exposure to information about financial markets as possible reasons for the lower risk tolerance of black and hispanic respondents compared with white respondents. in yao et al. (2004) the respondents who identified as “other,” which hanna and lindamood (2008) noted were mostly of asian identification, were less likely than otherwise similar white respondents to be willing to take some risk. yao et al. (2004) found that blacks and hispanics were less willing than white respondents to be willing to take some risk, but more willing to take substantial risk. this apparent inconsistency might be because of the lack of financial sophistication and lower exposure to financial information in these groups. controlling for other characteristics, we do not expect han chinese to have different level of risk tolerance compared with other ethnic groups. household resources, including financial assets, homeownership, and income, have been found to be positively related to risk tolerance, and the part of the reason might be because of the information costs of making investment decisions, as discussed by haliassos and bertaut (1995). a household with low resources might have very limited funds and the information costs of making an investment choice would be higher than the expected gains of choosing a higher return investment. therefore, we expect a positive relationship between resources (financial assets, homeownership, and income) and risk tolerance. 4. method 4.1. dataset and sample selection we used the 2011 china household finance survey (chfs). the survey is sponsored by the southwest university of finance and economics in cooperation with the finance research branch of people’s bank of china, and according to gan et al. (2012), when it was released, it was the first nationally representative survey on household finances in china. this dataset is highly focused on detailed level of economic status as well as various demographic characteristics of households. among the existing datasets in china, the chfs can be considered as the dataset with the highest response rate (gan et al., 2013). the overall response rate of the chfs is around 88%. in the 2011 chfs, the survey respondent was selected as the person who was most financially knowledgeable in the household. the total sample size of survey respondents was 8,438. we selected respondents with a valid response regarding to his or her risk attitude. there were 169 cases where the risk tolerance question was missing or unlisted, and those cases were excluded from our analyses. our analytical sample size was 8,269. we also analyzed the 2013 290 s.d. hanna et al. / financial services review 27 (2018) 279-302 scf (bricker et al., 2014), which had 6,015 households, but after excluding 13 cases with values of the risk tolerance variable imputed because of respondent nonresponse (hanna et al., 2018), our analytic sample was 6,002. 4.2. weighting gan et al. (2012) described the sampling method and creation of the chfs survey weight. as with the u.s. scf (bricker et al., 2014), wealthy households were oversampled, and the survey weight also reflects variations in geographic sampling rates, so the data need to be weighted by the survey weight variable for results to be better representative of chinese households. thus, we reported weighted results for descriptive and multivariate analysis. for comparison with the logistic regressions for the chfs, we ran logistic regressions based on the 2013 scf dataset for cumulative risk tolerance variables, following the methods reported in yao et al. (2004). we followed the recommendations discussed in shin and hanna (2017) for applying population and replicate weights in multivariate analyses of scf data. 4.3. measurement of variables 4.3.1. dependent variable the chfs has a variable for a household’s financial risk tolerance with five different levels, indicating the household’s willingness to take high, above-average, average, below-average, or no financial risk. the english translation of the codebook has the following: [a4003] assume you have some assets to invest, which type of project would you invest in? 1. high risk, high return 2. slightly above-average risk, slightly above-average return 3. average risk, average return projects 4. slightly below-average risk, slightly below-average return 5. unwilling to take any risk this variable is similar to the risk tolerance question in the u.s. survey of consumer finances (scf): [x3014] which of the statements comes closest to the amount of financial risk that you are willing to take when you save or make investments? 1. take substantial financial risks expecting to earn substantial returns 2. take above average financial risks expecting to earn above average returns 3. take average financial risks expecting to earn average returns 4. not willing to take any financial risks based on the empirical specification (see section 4.4.), we created three dichotomous composite variable of one’s financial risk tolerance; substantial risk, high risk, and some risk as follows: 291s.d. hanna et al. / financial services review 27 (2018) 279-302 x1�1 if high risk, and 0 otherwise; x2�1 if high, above-average, and average, and 0 otherwise, x3�1 if high, above-average, average, or below average, and 0 otherwise 4.3.2. independent variables based on previous research on risk tolerance and our research questions, the independent variables for the logistic regressions (table 6) included age of respondent, gender of the respondent (male vs. female), marital status (couple vs. single), education of the respondent (less than high school, high school degree, some college, and bachelor degree), employment status (salary worker, self-employed other than farming, farming, retired, and others), health status (excellent, good, fair, and poor), race/ethnicity of the respondent (han or not), homeowner (homeowner vs. renter), and total family income and the level of financial assets. family income included income from all sources. for comparison, our logistic regressions for the 2013 scf (appendix) included comparable variables: age of respondent, gender of the respondent, marital status (couple vs. single), education of the respondent (less than high school, high school degree, some college, bachelor degree), employment status (salary worker, self-employment other than farming, farming, retired and unemployed), health status (excellent, good, fair and poor), race/ethnicity of the respondent (white or not), homeownership, and total family income and the level of financial assets. 4.4. empirical specification given the characteristics of the scf risk tolerance variable, it would seem reasonable to use an ordered logit to analyze it. the assumption in the ordered logistic regression model is proportional odds (i.e., parallel), with constant effects across response categories. a score test indicated that the proportional odds assumption was not valid, and therefore, ordered logit was not appropriate. clogg and shihadeh (1994) proposed using k-1 separate cumulative logistic regression models. we followed previous scf studies analyzing risk tolerance (e.g., yao et al., 2004; yao & hanna, 2005) in using cumulative logistic regressions. therefore, three separate cumulative logit models were used in this study: logit�p�y � 1�� � log�y � 1 y � 1��1 � �1x (1) logit�p�y � 2�� � log�y � 2 y � 2��2 � �2x (2) logit�p�y � 3�� � log�y � 3 y � 3��3 � �3x (3) the cumulative logit models examine the effect of explanatory variables on the probability for households willing to take substantial risk, high risk and some financial risk tolerance. yao et al. (2004) discussed issues related to analysis of the scf risk tolerance 292 s.d. hanna et al. / financial services review 27 (2018) 279-302 variable, and concluded that it is not appropriate to analyze the variable as one continuous variable. our method is similar to the approach by yao et al. (2004) in defining three binary dependent variables. 5. results 5.1. descriptive results in the 2011 chfs, there were 8,269 valid responses to the risk tolerance question. table 3 shows the mean age of respondents, and the distribution of selected other characteristics. the mean age was about 49, and 54% of the respondents were male. most (68%) respondents did not have a high school degree, and only 8% had a college degree. in the scf, 34% of respondents had a college degree. in the chfs, almost 86% were married or partners, compared with only 57% in the scf. in the chfs, over 40% were in farming or otherwise self-employed, compared with less than 9% in the scf. in the chfs, 56% reported being table 3 distribution of selected characteristics, 2011 chinese household financial survey and 2013 scf variable percentage or mean 2011 chfs 2013 scf age of respondent (mean) 48.8 50.5 gender of respondent male 53.6 47.4 female 46.4 52.6 education of respondent less than high school degree 68.1 9.9 high school degree 12.1 28.6 some college 11.9 27.2 bachelor degree or higher 7.9 34.3 marital status (married or partners) 85.6 56.7 employment status of respondent employee 25.2 54.7 farming 31.6 0.3 self-employed other than farming 9.1 8.4 retired 12.8 10.0 other 21.3 26.5 health status of respondent excellent 12.3 23.9 good 31.8 48.8 fair 40.1 21.0 poor 15.8 6.3 ethnicity (han or white) 95.7 70.1 homeowner 81.8 65.2 mean household income, renminbi. (in u.s. dollars at 6.4 to 1) 55,226 ($8,629) $86,707 mean financial assets, renminbi (in u.s. dollars at 6.4 to 1) 63,968 ($9,995) $251,156 n 8,269 6,002 scf � survey of consumer finances; chfs � china household finance survey. chfs and scf results are weighted. 293s.d. hanna et al. / financial services review 27 (2018) 279-302 in poor or fair health, compared with 27% in the scf. almost 96% were of the han ethnicity, with the others distributed among the 55 other ethnic groups/nationalities listed in the codebook, whereas in the scf, 70% of respondents were white. almost 82% in the chfs were homeowners, compared with 65% in the scf. the mean household income in the chfs was only about 10% of the level of the united states, with an even greater relative difference for financial assets. table 4 presents descriptive results of distribution of the 2011 chfs risk tolerance, and the distribution of the comparable 2013 scf risk tolerance question (two categories in the chfs are combined for comparability). in the 2011 chfs, about 45% said they were not willing to take any investment risk, 42.5% said below average or average risk, 7.2% said above average risk, and 5.6% gave a response of being willing to take substantial risk. the distributions are somewhat similar between chfs and scf, though the proportion of no risk responses in the scf, 46.6%, is significantly higher than the rate in the chfs, using a simple z test for comparing proportions in two independent samples. the risk tolerance questions in the scf and in the chfs measure willingness to take risks, though there is not a direct connection of these questions to the economic concept of risk aversion. hanna and lindamood (2004) presented respondents with both the scf risk tolerance question, and a series of hypothetical income gamble questions with a graphical extension of the questions presented in barsky et al. (1997). hanna and lindamood (2004) estimated a regression of relative risk aversion on the responses to the scf risk tolerance question, to be able to project the distribution of risk aversion levels for households of all ages in the u.s. population. the middle section of table 4 relative risk aversion estimate, are based on the regression in hanna and lindamood (2004). based on our discussion in section 2.2., one should be skeptical of these estimates, but while very crude, the estimates allow for comparison with normative analyses of optimal portfolios by a number of economists, for example, campbell and vicerea (2002) and hanna and chen (1997). based on the hanna and chen analysis, even households with the no risk response should be willing to have 50% or more of their investment portfolio table 4 risk tolerance responses, 2011 chfs and 2013 scf 2011 chfs relative risk aversion estimateb 2013 u.s. scf response distribution response distribution 1. unwilling to take any risk 44.7%c � 6.7 1. no risk 46.6% 2. slightly below average or average riska 42.5%c 5.4 2. average risk 36.3% 3. slightly above average risk 7.2%c 4.0 3. above average risk 14.1% 4. substantial risk 5.6%c �2.7 4. substantial risk 3.0% sample size 8,269 6,002 scf � survey of consumer finances; chfs � china household finance survey. chfs and scf results are weighted. aslightly below average, 15.91%; average, 25.55% bestimates of relative risk aversion based on regression in hanna and lindamood (2004). crate significantly different from corresponding scf rate at p � 0.05. 294 s.d. hanna et al. / financial services review 27 (2018) 279-302 in stocks until they approach retirement. assuming that a no risk response corresponds to a relative risk aversion level of 8, and a substantial risk response corresponds to a relative risk aversion level of 2, the mean relative risk aversion level would be about 6.3 in both surveys, which according to hanna and chen (1997) would imply fairly aggressive retirement portfolios until about 20 years before retirement, and some stock assets even at retirement. table 5 shows direct and indirect ownership of stocks. the scf has very detailed questions that are used to estimate whether stocks are in mutual funds, including retirement accounts under the control of respondents, as well as to estimate the dollar value of stock holdings for each household. the chfs does not have comparable questions about indirect stock holdings. as table 5 shows, almost 9% of households in the chfs owned stocks directly, compared with about 14% of households in the chfs. the chfs has questions about ownership of mutual funds, financial derivatives, and wealth management products, but very few households reported owning these products who did not also own stocks directly. if we assume that all such products contained stock assets, then 9% of households in the chfs directly and/or indirectly owned stock assets, compared with almost 49% of households in the scf. although analysis of factors related to stock ownership in the chfs would be interesting (cf. wang and hanna, 2007), the proportion is small and the definitions in the survey are unclear, so we will not attempt that in this article. 5.2. multivariate results an ordered logistic regression model was tested with the chfs data, but the result of the score test (cf., yao et al., 2004) indicated that ordered logistic regression was not appropriate. a cumulative logistic regression model was utilized for testing the impact of each independent variable on each level of risk tolerance for chinese respondents (table 6), and for comparison, we also present the same logistic regressions estimated with the 2013 scf (appendix). we divided risk tolerance into three levels: substantial risk tolerance, high risk tolerance (willing to take above average or substantial risk vs. no risk or average risk) and some risk tolerance (the three levels vs. unwilling to take any risk). table 5 ownership of stocks and stock assets, 2011 chfs and 2013 scf percent of households type of investment 2011 chfs 2013 scf direct ownership of stocks 8.63 13.75 direct and/or indirect ownership of stocks 9.03 48.85 scf � survey of consumer finances; chfs � china household finance survey. weighted analyses of all households in each dataset. estimate for chfs for direct and/or indirect ownership of stocks is an upper estimate, based on the assumption that all mutual funds, financial derivatives, and wealth management products contain stock assets. 295s.d. hanna et al. / financial services review 27 (2018) 279-302 5.2.1. some risk tolerance the first three columns of table 6 contain the logistic regression results on willingness to take some risk in the chfs. if the respondent was willing to take average, above average, or substantial risk then “some risk” was set to 1, otherwise, “some risk” was set to 0. the likelihood of being willing to take some risk decreased with age in both the chfs and in the scf (appendix). the odds ratios for both survey results indicate that men have odds 1.8 times those of women in being willing to take some risk. for both the chfs logit and the scf logit (appendix), the likelihood of being willing to take some risk increased strongly with education. in both surveys, being self-employed (other than farming) was associated with the likelihood of being willing to take some risk, though farmers in the chfs were not significantly different from others, while those in farming related jobs in the scf were more likely to be willing to take some risk than others. those with fair health status in the chfs were less likely to be willing to take some risk than those with poor health, while in the scf, those with excellent health were more table 6 logistic regressions for some, high, and substantial risk tolerance, 2011 chfs variable some risk high risk substantial risk coefficient p value odds ratio coefficient p value odds ratio coefficient p value odds ratio age of respondent �0.056 �.001 0.945 �0.050 �.001 0.951 �0.030 �.001 0.970 male (ref: female) 0.582 �.001 1.789 0.476 �.001 1.609 0.474 �.001 1.606 single (ref: couple) �0.037 0.653 0.964 �0.034 0.671 0.967 0.279 0.041 1.321 education of respondent (ref: less than a high school degree) high school degree 0.483 �.001 1.621 0.385 �.001 1.469 0.100 0.506 1.106 some college 0.710 �.001 2.035 0.535 �.001 1.708 0.088 0.577 1.092 college degree 1.228 �.001 3.414 0.873 �.001 2.393 0.012 0.948 1.012 employment status of respondent (ref: employee) self-employed (other than farmer) 0.421 �.001 1.523 0.366 �.001 1.442 0.390 0.015 1.477 farmer 0.117 0.133 1.125 0.144 0.061 1.155 0.102 0.500 1.108 retired 0.066 0.505 1.069 0.000 0.997 1.000 �0.117 0.623 0.890 other job status 0.060 0.524 1.062 0.170 0.063 1.186 �0.300 0.101 0.741 health status of respondent (ref: poor health) excellent �0.174 0.085 0.841 0.218 0.028 1.244 0.299 0.083 1.348 good �0.041 0.605 0.960 0.159 0.050 1.172 �0.488 0.003 0.614 fair �0.168 0.024 0.846 �0.005 0.948 0.995 �0.267 0.079 0.766 han (other ethnic) �0.111 0.366 0.895 �0.227 0.058 0.797 �0.277 0.195 0.758 homeowner (ref: renter) 0.141 0.050 1.152 0.206 0.003 1.229 0.038 0.769 1.038 log (income) 0.028 0.002 1.029 0.031 0.001 1.031 0.019 0.274 1.020 log (financial assets) 0.049 �.001 1.050 0.031 �.001 1.032 �0.015 0.354 0.985 intercept �1.736 �.001 0.816 �.001 �1.438 �.001 mean concordance rate 74.6% 72.7% 66.0% chfs � china household finance survey. weighted analyses of 2011 chfs. total sample size is 8,269. for dummy variables or sets of dummy variables, reference category is in parentheses. the p values are based on two-tail tests. 296 s.d. hanna et al. / financial services review 27 (2018) 279-302 likely to be willing to take some risk than those with poor health. in the scf, white respondents were much more likely to be willing to take some risk than those with other racial/ethnic identifications (black, hispanic, and asian/other) but han respondents in the chfs were not significantly different from non-han respondents. homeowners were more likely than renters in the chfs to be willing to take some risk, but there was no difference in the scf. in both surveys, the likelihood of being willing to take some risk increased strongly with income and with financial assets. 5.2.2. high risk tolerance the middle three columns of table 6 present logistic regression results on willingness to take high risk in the chfs. if the respondent was willing to take above average risk or substantial risk then “high risk” was set to 1, otherwise, high risk was set to 0. the likelihood of being willing to take high risk decreased with age in both the chfs and in the scf. men had odds of being willing to take high risk 1.61 times as high as women in the chfs, and the scf had a similar pattern. single respondents were not significantly different from couple respondents in the chfs, but respondents in single households in the scf were more likely to be willing to take high risk than those in married or partner households. education was positively related to having high risk in the chfs, while in the scf, college educated respondents were more likely to willing to take high risk than those without a high school degree. self-employed respondents in both the chfs and the scf were more willing than others to take high risk. ethnic status was not significant in either the chfs or the scf. those with excellent or good health status in the chfs were more likely to be willing to take high risk than those with poor health, while in the scf, health status was not significantly related to being willing to take high risk. in the chfs, homeowners were more willing to take high risk than renters, but the effect was not significant in the scf. in both surveys, income and financial assets were positively associated with the likelihood of being willing to take high risk. 5.2.3. substantial risk tolerance the last three columns of table 6 contain logistic regression results on willingness to take substantial risk. if the respondent was willing to take substantial risk then “substantial risk” was set to 1, otherwise, substantial risk was set to 0. age was negatively related to willingness to take substantial risk, with the odds ratio indicating almost a 3% decrease in the odds of being willing to take substantial risk with every one year increase in age. in the appendix table, the effect of age was similar for the u.s. respondents, with a 2% decrease in the odds for every one year increase in age. male respondents were more likely to be willing to take substantial risk than female respondents, with odds 1.6 times as high in the chinese survey and 1.4 times as high in the u.s. survey. being single was related to substantial risk tolerance in the chfs but not in the scf. education did not have a significant effect in the chfs on substantial risk tolerance, and had a mixed effect in the scf, with those with some college being more likely than those without a high school degree to be willing to take substantial risk. self-employed respondents (other than farmers) had odds of being willing to take substantial risk 1.5 times that of other respondents in the chinese survey and 1.9 times as high in the u.s. survey. one large difference between the 297s.d. hanna et al. / financial services review 27 (2018) 279-302 two surveys was the effect of being a farmer or in a related occupation, while in china, farmers were not significantly different from others in being willing to take substantial risk, while in the united states, those who were farmers or in a related occupation had odds of being willing to take substantial risk over six times as high as those in other occupations. han respondents were not significantly different from non-han respondents in being willing to take substantial risk in the chfs. in the scf (appendix), white respondents were less likely than others to be willing to take substantial risk, opposite to the effect for some risk tolerance, similar to a result reported by yao et al. (2004). these results provide additional support beyond the results of the score test for using a cumulative logit model rather than an ordered logit model. 6. discussion and implications our descriptive results show that despite the much higher income and education levels of u.s. respondents, the risk tolerance levels of chinese respondents were very similar to those of u.s. respondents, with estimated levels of relative risk aversion almost identical in the two surveys. however, about 49% of households in the united states had stock assets, compared with only 9% of households in china. this suggests that as household incomes continue to increase in china, there may be a rapid increase in the ownership of stocks, perhaps through mutual funds. education and appropriate regulation will be important in the future for households in china. some studies have found relationships between education, financial literacy, and the ownership of risky investments. most research has found that respondents with higher education levels and/or more financial knowledge are more risk tolerant (gibson, et. al, 2013; grable, 2000; nobre and grable, 2015). better understanding of risk will enhance the involvement and utilization of appropriate investments by chinese households to better achieve their financial planning goals, which could have a substantial impact on retirement adequacy (kim and hanna, 2015). based on cumulative logistic regressions, many household characteristics affected respondent risk tolerance for chinese respondents and also for u.s. respondents, and for seven types of variables, generally in ways consistent with our research expectations. there was a cross-sectional negative relationship between age and the willingness to take risk, and while this does not necessarily imply that people will become less risk tolerant as they get older, in terms of financial education and policies, age should be considered as an important correlate to risk tolerance. men were more willing to take risk than women in both surveys. the willingness to take some risk and to take high risk increased with education in the chfs and generally in the scf. in both surveys, self-employed respondents were more willing to take investment risk than others. those with excellent health in the scf were more likely to be willing to take some risk than those with poor health, but otherwise health was not related to risk tolerance. health status had a few significant effects in the chfs, but with no consistent pattern. risk tolerance increased with income and financial assets in both the scf and in the chfs. being a homeowner was related to being more willing to take some risk and to take high risk in the chfs but not in the scf. in the scf, white respondents were more willing to take some risk 298 s.d. hanna et al. / financial services review 27 (2018) 279-302 but less willing to take substantial risk than respondents identifying as black, hispanic, or asian, a result previously found by yao et al. (2004). in the chfs, han respondents were not different in risk tolerance from respondents of other ethnic identities, which may reflect the shrinking income gap (gustafsson and shi, 2003). despite some differences between the chfs and the scf in the effects of household characteristics on risk tolerance, the similarity of the distribution of risk tolerance levels and the general similarity of effects in the logistic regressions suggest that risk tolerance is perceived in a somewhat similar way in china and the united states. therefore, future research with chinese household surveys could benefit from careful consideration of research with united states datasets. for instance, analysis of factors related to ownership of stocks (e.g., wang and hanna, 2007) could provide valuable insights for china. campbell and viceira (2002) derived optimal stock percentages for households, based on age, risk aversion, and the correlation of human wealth with financial returns. given their assumptions about patterns of human wealth and age, the optimal percentage of the total portfolio in stocks was very sensitive to the assumed level of risk aversion, for instance, for a worker 35 years away from retirement, with zero correlation between human wealth and investment returns, the optimal stock percentage ranged from 148% for a relative risk aversion level of 3, to 27% for a relative risk aversion level of 12. for any given level of risk aversion, the optimal stock allocation decreased with age, for instance, for a risk aversion level of 5, decreasing from 76% at 35 years from retirement to 42% at 5 years from retirement. however, even at very low risk tolerance levels (relative risk aversion level of 12), households should have some stock assets. the general optimal patterns obtained by campbell and viceira (2002) and others (e.g., hanna and chen, 1997) should be considered in terms of the risk tolerance findings reported in this article. individual advice to households should consider their risk tolerance and also their risk capacity (cordell, 2002; hanna and chen, 1997). thaler and sunstein (2008) argued that trying to educate households to make good decisions about complex issues such as retirement investments is likely to fail, and present choice architecture, or “nudges” as a way to improve household decision-making. however, the key issue for many of the “nudge” suggestions is starting with the appropriate default levels, for instance, the default portfolio composition of worker retirement portfolios. the normative analyses of campbell and viceira (2002) and of poterba, rauh, and venti (2005) show that optimal investment allocations depend crucially on the level of risk aversion of households. therefore, measurement of risk tolerance is important in devising policies that will help households make better decisions. our research on the risk tolerance of households in china represents a first step in that task for china. future household surveys in china should use several measures of risk tolerance and risk aversion, for instance, including an income gamble question to try to isolate preferences from situational factors (hanna et al., 2001). as financial investment choices expand in china, it should be better to obtain improved estimates of the relationship between risk tolerance and investment choices. 299s.d. hanna et al. / financial services review 27 (2018) 279-302 references arrow, k. j. (1971). essays in the theory of risk-bearing. chicago, il: markham publishing co. attanasio, o. p., banks, j., & tanner, s. (2002). asset holding and consumption volatility. journal of political economy, 110, 771–792. bailey, m. j., olson, m., & wonnacott, p. (1980). the marginal utility of income does not increase: vorrowing, lending, and friedman-savage gambles. american economic review, 70, 372–379. barsky, r. b., juster, f. t., kimball, m. s., & shapiro, m. d. (1997). preference parameters and behavioral heterogeneity: an experimental approach in the health and retirement survey. quarterly journal of economics, 112, 537–579. bricker, j., dettling, l. j., henriques, a., hsu, j. w., moore, k. b., sabelhaus, j., thompson, j., & windle, r. a. (2014). changes in u.s family finances from 2010 to 2013: evidence from the survey of consumer finances. federal reserve bulletin, 100, 1–41. appendix logistic regressions for some, high, and substantial risk tolerance, 2013 scf variable some risk high risk substantial risk coefficient p value odds ratio coefficient p value odds ratio coefficient p value odds ratio age of respondent �0.022 �.001 0.978 �0.026 �.001 0.974 �0.020 0.002 0.980 male (ref: female) 0.545 �.001 1.725 0.455 �.001 1.575 0.337 0.001 1.401 single (ref: couple) 0.153 0.006 1.165 0.151 0.011 1.163 0.162 0.386 1.175 education of respondent (ref: less than a high school degree) high school degree 0.422 �.001 1.523 0.092 0.583 1.096 0.242 0.469 1.273 some college 0.763 �.001 2.145 0.442 0.012 1.555 0.715 0.048 2.045 college degree 1.419 �.001 4.131 0.789 �.001 2.202 0.417 0.225 1.517 respondent’s employment status (ref: employee) self-employed (other than farmer) 0.427 �.001 1.532 0.313 0.005 1.368 0.620 0.005 1.858 farmer 1.413 �.001 4.107 0.703 0.084 2.019 1.936 0.001 6.928 retired 0.174 0.058 1.190 �0.114 0.222 0.892 �0.196 0.396 0.822 not working �0.198 0.047 0.820 �0.204 0.061 0.815 0.020 0.926 1.020 health status (ref: poor health) excellent 0.425 0.001 1.530 0.282 0.160 1.325 0.654 0.156 1.924 good 0.220 0.060 1.246 0.250 0.217 1.283 0.617 0.195 1.852 fair 0.173 0.170 1.188 0.165 0.421 1.179 0.842 0.054 2.321 white (ref: black, hispanic, or other) 0.272 �.001 1.313 0.099 0.191 1.105 �0.410 0.006 0.664 homeowner (ref: renter) 0.061 0.419 1.063 0.090 0.338 1.095 �0.053 0.780 0.948 log (income) 0.158 �.001 1.172 0.101 0.030 1.106 0.037 0.650 1.038 log (financial assets) 0.139 �.001 1.149 0.060 �.001 1.062 �0.004 0.826 0.996 intercept �3.277 �.001 �3.111 �.001 �3.968 �.001 mean concordance rate 81.9% 71.2% 64.5% scf � survey of consumer finances rii procedure with population and replicate weights applied. total sample size is 6,002. for dummy variables or sets of dummy variables, reference category is in parentheses. the p value are based on two-tail tests. 300 s.d. hanna et al. / financial services review 27 (2018) 279-302 brown, j. r., & poterba, j. m. (2000). joint life annuities and annuity demand by married couples. the journal of risk and insurance, 67, 527–554. calvet, l., campbell, j. y., & sodini, p. (2007). down or out: assessing the welfare costs of household investment mistakes. journal of political economy, 115, 707–747. campbell, j. y. (2006). household finance. journal of finance, 61, 1553–1604. campbell, j. y., & cocco, j. f. (2003). household risk management and optimal mortgage choice. the quarterly journal of economics, 118, 1449–1494. campbell, j. y., & viceira, l. m. (2002). strategic asset allocation: portfolio choice for longterm investors. new york, ny: oxford university press. clogg, c. c., & shihadeh, e. s. (1994). statistical models for ordinal data. thousand oaks, ca: sage publications. cordell, d. m. (2002). risk tolerance in two dimensions. journal of financial planning, 15, 30–35. cocco, j. f., gomes, f. j., & maenhout, p. j. (2005). consumption and portfolio choice over the life cycle. review of financial studies, 18, 491–533. cocco, j. f. (2005). portfolio choice in the presence of housing. the review of financial studies, 18, 535–567. deaton, a., & muellbauer, j. (1980). economics and consumer behavior. new york, ny: cambridge university press. fan, j. x., & xiao, j. j. (2006). cross-cultural differences in risk tolerance: a comparison between chinese and americans. journal of personal finance, 5, 54–75. fang, m. c., hanna, s. d., & chatterjee, s. (2013). the impact of immigrant status and racial/ethnic group on differences in responses to a risk aversion measure. journal of financial counseling and planning, 23, 64–77. friend, i., & blume, m. e. (1975). the demand for risky assets. american economic review, 65, 900–922. gan, l., yin, z., jia, n., xu, s., & ma, s. (2013). asset allocation of chinese households and housing demand analysis. journal of finance research, 4, 1–14. gan, l., yin, z., jia, n., xu, s., ma, s., & zheng, l. (2012). data you need to know about china: research report of china household finance survey. berlin: springer science & business media. gibson, r., michayluk, d., venter, g. v. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. gomes, f. j., & michaelides, a. (2005). optimal life-cycle asset allocation: understanding the empirical evidence. journal of finance, 60, 869–904. gomes, f. j., & michaelides, a. (2003). portfolio choice with internal habit formation: a life-cycle model with uninsurable labor income risk. review of economic dynamics, 6, 729–766. grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625–630. grable, j. e., & joo, s. h. (2004). environmental and biopsychosocial factors associated with financial risk tolerance. financial counseling and planning, 27, 73–82. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. grable, j. e., & lytton, r. h. (2001). assessing the concurrent validity of the scf risk tolerance question. financial counseling and planning, 12, 43–52. gutter, m. s., & hatcher, c. b. (2008). racial differences in the demand for life insurance. journal of risk and insurance, 75, 677–689. gustafsson, b., & shi, l. (2003). the ethnic minority-majority income gap in rural china during transition. economic development and cultural change, 51, 805–822. haliassos, m., & bertaut, c. c. (1995). why do so few hold stocks? the economic journal, 105, 1110–1129. hanna, s. d., & chen, p. (1997). subjective and objective risk tolerance: implications for optimal portfolios. financial counseling and planning, 8, 17–26. hanna, s. d., gutter, m. s., & fan, j. x. (2001). a measure of risk tolerance based on economic theory. financial counseling and planning, 12, 53–60. hanna, s. d., kim, k. t., & lindamood, s. (2018). behind the numbers: understanding the survey of consumer finances. journal of financial counseling and planning, 29, 410–418. 301s.d. hanna et al. / financial services review 27 (2018) 279-302 hanna, s. d., & lindamood, s. (2008). the decrease in stock ownership by minority households. journal of financial counseling and planning, 19, 46–58. hanna, s. d., waller, w., & finke, m. (2008). the concept of risk tolerance in personal financial planning. journal of personal finance, 7, 96–108. ho, k., milevsky, m. a., & robinson, c. (1994). asset allocation, life expectancy and shortfall. financial services review, 3, 109–126. horneff, w. j., maurer, r. h., & stamos, m. z. (2008). optimal gradual annuitization: quantifying the costs of switching to annuities. the journal of risk and insurance, 75, 1019–1038. hubbard, r. g., skinner, j., & zeldes, s. p. (1994). the importance of precautionary motives in explaining individual and aggregate saving. carnegie-rochester conference series on public policy, 40, 59–125. kim, k., & hanna, s. d. (2015). does financial sophistication matter in retirement preparedness? journal of personal finance, 14, 9–20. kim, k., wilmarth, m., & choi, s. (2016). credit use of u.s. households after the great recession: the role of credit constraints. family & consumer sciences research journal, 44, 280–294. kimball, m. s. (1988). farmers’ cooperatives as behavior toward risk. the american economic review, 78, 224–232. lv, x. l. (2007). analysis of bank customer’s risk tolerance. mathematical statistics and management, 26, 1062–1065. nobre, l. h. n., & grable, j. e. (2015). the role of risk profiles and risk tolerance in shaping client investment decisions. journal of financial service professionals, 69, 18–21. pyles m. k., li, y., wu, s., & dolvin, s. d. (2016). cultural influences on risk tolerance and portfolio creation. journal of behavioral and experimental finance, 9, 43–55. poterba, j., rauh, j., & venti, s. f. (2005). utility evaluation of risk in retirement saving accounts. in d. a. wise (ed.), analyses in the economics of aging (pp. 13–58). chicago, il: university of chicago press. pratt, j. w. (1964). risk aversion in the small and in the large. econometrica, 32, 122–136. schoemaker, p. j. h. (1982). the expected utility model: its variants, purposes, evidence and limitations. journal of economic literature, 20, 529–563. scholz, j. k., seshadri, a., & khitatrakun, s. (2006). are americans saving “optimally” for retirement? journal of political economy, 114, 607–643. shin, s. h., & hanna, s. d. (2015). decomposition analyses of racial/ethnic differences in high return investment ownership after the great recession. journal of financial counseling and planning, 26, 43–62. shin, s. h., & hanna, s. d. (2017). accounting for complex sample design in analyses of the survey of consumer finances. journal of consumer affairs, 51, 433–447. siegel, j., & thaler, r. h. (1997). anomalies: the equity premium puzzle. journal of economic perspectives, 11, 191–200. sung, j., & hanna, s. d. (1996). factors related to risk tolerance. journal of financial counseling and planning, 7, 11–19. terpstra-tong, j. l., & terpstra, r. h. (2013). differences in financial risk tolerance: a study of the u.s., china, hong kong and macao. euro asia journal of management, 23, 43–65. thaler, r. h., & sunstein, c. r. (2008) nudge: improving decisions about health, wealth, and happiness. new haven, ct: yale university press. wang, c., & hanna, s. d. (2007). the risk tolerance and stock ownership of business owning households. journal of financial counseling and planning, 18, 3–18. wang, h., & hanna, s. (1997). does risk tolerance decrease with age? financial counseling and planning, 8, 27–31. yao, r., gutter, m. s., & hanna, s. d. (2005). the financial risk tolerance of blacks, hispanics and whites. financial counseling and planning, 16, 51–62. yao, r., & hanna, s. d. (2005). the effect of gender and marital status on financial risk tolerance. journal of personal finance, 4, 66–85. yao, r., hanna, s. d., & lindamood, s. (2004). changes in financial risk tolerance, 1983–2001. financial services review, 13, 249–266. 302 s.d. hanna et al. / financial services review 27 (2018) 279-302 which measures predict risk taking in a multi-stage controlled investment decision process? kremena bachmanna, thorsten hensb,c,e*, remo stösseld adepartment of banking and finance, university of zurich, plattenstrasse 32, 8032 zürich, switzerland bdepartment of banking and finance, university of zurich, plattenstrasse 32, 8032 zürich, switzerland and cdepartment of finance, norwegian school of economics, nhh, bergen ddepartment of banking and finance, university of zurich, plattenstrasse 32, 8032 zürich, switzerland edepartment of economics, university of lucerne, switzerland abstract we assess the ability of different risk profiling measures to predict risk taking along a multistage process that reflects individuals’ discovery of their willingness to take risks. we find that the individual willingness to take risks varies along the process, but its level is always related to a composite measure of the individual risk tolerance. assessment of the risk tolerance cannot be substituted by a simulated experience, although the latter can improve the perception of the risk and reward potential of the investment and motivate higher risk taking. the risk tolerance measure addresses different notions of risk, but we found that the individual loss aversion is the most powerful predictor of risk taking at all stages of the discovery process. by contrast, we found that neither the self-assessed risk tolerance measures nor the investment experience are suitable for consistently predicting risk taking at different stages of the process. © 2017 academy of financial services. all rights reserved. jel classification: d81; g11 keywords: risk profiling; risk tolerance; risk attitude; risk preferences; risk taking; experience sampling 1. introduction an essential task in investment management is determining the amount of risk an investor should take. in principle, investors can identify their willingness to bear risks through investment in the financial market, but this approach is costly because a considerable amount * corresponding author. tel.: �41 44 634 48 22. e-mail address: kremena.bachmann@bf.uzh.ch (k. bachmann) financial services review 26 (2017) 339–365 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. of wealth can be lost because of inconsistent decisions during the learning process. to assist investors and justify their recommendations as required by regulators, financial professionals use various techniques to determine the level of risk that their clients should take. in this study, we evaluate the suitability of such risk profiling techniques based on their power to explain and predict individual risk-taking behavior. more important, we believe that the relationship between the assessed risk profile and the subsequent risk taking may not be stable if individuals are still in the process of identifying their willingness to take risks. the involvement in such a process is likely because individuals are not always able to correctly anticipate their emotional reactions to possible outcomes (kahneman, 2009). to shed some light on this issue, we conduct an experimental study on whether an individual’s risk taking changes over different stages of a process along which private investors are expected to correct misperceptions and discover their true willingness to take risks. we then analyze how the predictability of risk profiling questions varies over the stages of such a process. the goal of the study is to identify risk profiling measures that consistently explain and predict risk taking at all stages of the discovery process. this consistency is important because investment advisors usually do not know which stages of the process their clients have completed. using a risk profiler that is suitable only if clients have completed certain stages of the discovery process can lead to inappropriate advice being given. to determine the relevant stages of the discovery process, we consider evidence from previous studies reporting that individual risk taking varies with certain characteristics of the decision setting, such as ambiguity, personal experience, and feedback. we use these features to design a multistage discovery process that reflects the investment experience of a typical private investor. for simplicity, investors only decide between one risky asset and cash. at the beginning of this process, it is assumed that investors decide within an ambiguous situation, that is, they know the return of holding cash, but they do not know anything about the return distribution of the risky asset. afterwards, the ambiguity is revealed while investors can choose its presentation format. in the third stage, the investors are asked to answer some risk profiling questions. in the next stage of the process, they experience the risk-return characteristics of different asset allocations based on simulations. in the fifth stage, the investors learn which return they have made, and in the last stage, they are able to reconsider their investment decision using a three-day break. we analyze whether individual risk taking changes over the different stages of the process, that is, whether investors are involved in a process of discovering their willingness to take risks. we then analyze the ability of different risk profiling questions to consistently predict risk taking over the different stages of the process. nobre and grable (2015) suggest that risk profiling questions should consist of questions assessing the risk need (the amount of risk required to meet a particular financial goal), the risk capacity (the client’s ability to absorb a possible financial loss resulting from the financial risk taken) and the financial risk tolerance (an individual’s willingness to accept uncertainty related to the outcome of a financial decision). carr (2014) analyses the optimal weighting of these dimensions. in this study, we focus on the assessment of the investors’ risk tolerance, which is a psychological concept. the assessment of the risk need and the risk capacity are purely financial issues that can be managed with financial planning tools. in our study, we use a broad definition of risk tolerance that 340 k. bachmann et al. / financial services review 26 (2017) 339–365 refers to losses as an additional notion of risk. moreover, we use different formats to state the questions, that is, some questions use lotteries and others use verbal alternatives; we also consider questions based on a self-assessment. additionally, we consider other factors that may affect risk taking, such as investment experience outside of the study and the investors’ risk awareness as reflected in the misperception of the true risk-reward profile of their investments. we also analyze whether simulated experience can substitute for risk profiling based on questions. we find that some aspects of an individual’s risk tolerance explain risk taking at all stages of the decision process, while the risk awareness and the self-stated investment experience cannot. moreover, although simulated experience improves risk awareness and supports risk taking, it cannot be used as a substitute for the assessment of individual risk tolerance when explaining and predicting risk taking. while risk tolerance can be measured in many ways, we find that the individuals’ loss aversion is the most suitable measure because it most accurately predicts the risk-taking behavior of investors involved in a process of discovering their willingness to take risks. of interest, we find that self-assessed risk tolerance measures are not suitable for predicting risk taking at any stage of the decision process. if individuals’ risk tolerance cannot be assessed and one must rely on socioeconomic characteristics, then only gender can be used as a predictor of risk taking. the results of our study have important policy implications. regulators in most developed countries acknowledge the importance of using risk profilers, and professional advisors use various risk profiling methods to justify their recommendations. however, it is not clear whether the risk profilers used in practice are suitable for determining the optimal level of risk taking (brayman, finke, grable, and griffin, 2017). their external validity is sometimes tested based on real asset allocation decisions (corter and chen, 2006; gilliam, chatterjee, and grable, 2010; grable and lytton, 2003; m. guillemette, finke, and gilliam, 2012; wärneryd, 1996). however, it is unclear whether an asset allocation at a certain point of time is a good assessment criterion because clients may still be involved in the process of discovering their willingness to take risks. our analysis explicitly considers the impact of this discovery process on the suitability of different risk profiling measures. we identify measures that consistently predict risk taking at every stage of the process. this is important for advisors because they usually do not know which stages of the discovery process their clients have already passed. using questions that consistently predict risk taking at all stages of the discovering process increases the probability that clients remain satisfied with the recommendations. at the same time, making recommendations based on questions that consistently predict risk taking at all stages of the discovery process should support the advisors’ confidence that these recommendations match the clients’ risk tolerance and do not encourage misperceptions that are corrected over time. 2. literature review and research hypotheses using different measures of individual risk tolerance, previous studies have found that these measures are related to individual investment risk taking. for example, barsky and juster (1997) find that risk tolerance revealed in a hypothetical choice between uncertain 341k. bachmann et al. / financial services review 26 (2017) 339–365 income streams predict stock ownership. yook and everett (2003) find a significant positive correlation between the total score of several risk tolerance measures and the percentage of actual stock holdings in portfolios. corter and chen (2006) propose another risk tolerance measure and show that it is positively correlated with the riskiness of the actual investment portfolios chosen. wärneryd (1996) finds a significant relationship between the individual investment attitude based on risk-return considerations and the risk in portfolios of dutch households. gilliam et al. (2010) find a significant positive association between broadly used risk tolerance measures and equity ownership. while these studies show that the evaluation of the individual risk tolerance is important for explaining investment risk taking, it remains unclear whether the explanatory power remains stable over time because individuals change their risk-taking behavior. for this reason, we designed a controlled laboratory experiment that stays close to the advisory processes found in praxis so that the setting is not too artificial. we also consider informationand experience-driven changes in investment risk taking. at the beginning, investors are expected to make investment decisions under ambiguity, that is, they may not know the exact risk-return characteristics of the alternatives that they consider for investment. frisch and baron (1988) argue that ambiguity arises from the perception of missing information relevant for a probability judgment, which supports the normative status of utility theory. from a theoretical perspective, ambiguity is important because it motivates lower stock market participation compared with the basic expected utility model (see, e.g., epstein and schneider, 2010 among others). antoniou et al. (2015) confirm the prediction of the theoretical ambiguity literature. in particular, they find that an increase in ambiguity is associated with reductions in capital flows into equity mutual funds. hence, providing information that makes probability judgments easier can increase risk taking. based on this literature, we conjecture that our participants take less risk under ambiguity, that is, in the first stage, than in later stages of our experiment. in the second stage of our experiment, the participants can acquire three different descriptions of the returns of the risky asset. previous studies have shown that even if individuals are provided with identical information, the presentation format can influence the utilization of information. in a classic demonstration of this phenomenon, slovic et al. (1978) observe that the presentation of formally equivalent statistics influences risk-taking behavior. similar types of framing effects have been reported in the literature on decision-making (tversky and kahneman, 1981). framing effects have been extensively used to modify risk-relevant behavior, facilitate cooperative conflict resolutions and advance knowledge or attitudes (see rohrmann, 1992 for an overview). we focus on the last aspect and hypothesize that individuals have different abilities to utilize information in different formats, which may influence their risk-taking behavior. in the third stage of our experiment, the participants are asked to answer questions regarding their risk tolerance and investment experience. the effect, wherein individuals change their behavior in response to being monitored, has been widely discussed in health economics (parsons, 1974) and consumer behavior research (fitzsimons and williams, 2000). in our study, we consider the existence of assessment effects in the context of investment risk taking. 342 k. bachmann et al. / financial services review 26 (2017) 339–365 in the fourth stage of our experiment, the participants can experience the return distribution by drawing samples from it before they can decide how to invest. converging findings show that there are systematic differences between decisions based on experience and decisions based on description (hertwig and erev, 2009), particularly in the context of decisions involving rare events (hertwig et al., 2004). kaufmann et al. (2013) show that communicating risk with the help of experience sampling and graphical displays leads to higher risk taking. goldstein et al. (2008) suggest that using interactive methods allowing individuals to explore the probability distributions of potential outcomes can be beneficial for inferring preferences and predicting subsequent risk-taking behavior. in line with this research, we hypothesize that experience sampling influences risk taking. in particular, we analyze whether experience sampling can substitute the assessment of individual risk tolerance in explaining and predicting risk-taking behavior. in the next stage of our experiment, the participants have a break of three days in which they can carefully study the design of the experiment and what they have done so far. previous research suggests that decision-makers switch to simpler strategies if decisions have to be made under time pressure, which can explain preference reversals (ordonez and benson, 1997). in negotiations, for example, individuals appear to reach higher-quality agreement after a break because the latter allows them to assess strategies and behavior (harinck and de dreu, 2008). we hypothesize that giving individuals time to re-evaluate the decision problem may have an impact on their subsequent risk taking. in the last stage of our experiment, the participants learn the outcomes of their previous investments and decide for the last time whether and how to revise them. given that all relevant information is available before a decision is made, the outcome of a decision should not be used to improve subsequent decisions. however, fischhoff (1975) demonstrates the existence of a hindsight bias, an effect of the outcome information on the judged probability for different outcomes. his explanation for observing this bias is that outcome information calls attention to information that would make a decision good or bad. for example, bad outcomes call attention to the risks associated with the decision as an argument against taking the decision. we hypothesize that the information on the outcomes of previous decisions may affect the subsequent risk taking and take the effect into account when assessing the suitability of risk profiling questions. 3. survey design our study consists of six stages, which differ either in the information that individuals receive or in the tasks they have been asked to perform. table 1 provides an overview of all stages. it specifies the information that is also provided at every stage and the tasks that the individuals were asked to perform after receiving the new information. a common task at every stage is an investment decision. at each stage, individuals were given financial wealth expressed in experimental currency units (ecu) and asked to split the wealth between a risky and a riskless asset. the amount in ecu varied between individuals dependent on their true financial situation, which was assessed in advance 343k. bachmann et al. / financial services review 26 (2017) 339–365 together with other demographic and socio-economic characteristics. the monetary value of all ecu endowments was 10 euros. the investment decisions between stages were independent. the individuals were informed that one of their investment decisions would be relevant for their final payment and that the relevant decision would be determined randomly at the end. in the first stage, individuals were asked to make an investment decision under ambiguity, that is, the individuals knew only the return of the riskless asset but did not have any information about the return distribution of the risky asset. the latter was provided in the second stage using different formats. the graphical format used histograms, the verbal format was based on scenarios, and the statistical format used descriptive statistics (see appendix d). the individuals could use the format that they considered most helpful. acquiring information was not mandatory. subsequently, individuals were asked to make an investment decision for a second time. in the next stage, no new information was provided. instead, individuals were asked questions about their risk tolerance, financial knowledge and investment experience. because asking such questions may change the individual risk-taking behavior, we asked individuals to make a third investment decision. afterwards, individuals were asked questions assessing their risk awareness, that is, their understanding of the risks and rewards associated with different investment decisions. in the fourth stage, individuals received the opportunity to experience the risk of investment in the risky asset. our experience sampling tool is based on the same idea as the tool used by kaufmann et al. (2013), that is, individuals draw different scenarios on the realization of the risky asset and observe how the return distribution of different asset allocations table 1 survey structure new information provided tasks after receiving new information stage 1: ambiguity information on the return of the riskless asset make an investment decision stage 2: return information return distribution of the risky asset (described by graphics, scenarios, and statistics) make an investment decision stage 3: profile estimation 1. answer questions assessing risk tolerance, financial knowledge, and experience 2. make an investment decision 3. answer risk awareness questions (1st time) stage 4: simulated experience experience the risk-return profile of different asset allocations through simulations 1. answer risk awareness questions (2nd time) 2. make an investment decision stage 5: time break three days break make an investment decision stage 6: feedback receive report of returns with all previous investment decisions 1. state satisfaction/expectations 2. make an investment decision 344 k. bachmann et al. / financial services review 26 (2017) 339–365 emerge. to make asset allocations comparable, we allowed individuals to simultaneously observe the final outcomes of two different asset allocations side-by-side (see figure a-1 in the appendix). both asset allocations use the same return realization of the risky asset and the same investment horizon of 1 year. the simulations were restarted with every change in the asset allocation. to avoid framing effects, both return distributions were scaled in the same way. after observing the final outcomes of at least two hundred scenarios (this required at least 10 drawings), the individuals were asked to answer our risk awareness questions for the second time and to make an investment decision. the payoff of the participants depended on this investment decision but not on the decisions made while drawing outcomes of different asset allocations. in the fifth stage, the individuals were informed that they would have a three-day break. in reality, the clients received factsheets with investment information. similarly, individuals were given the option to download the description of the assets for further reference. after a three-day break, the individuals were asked to make their fifth investment decision. they were also asked to state which investment decision they consider the best one, that is, which investment decision they would consider relevant for their payment. in the sixth stage, individuals received a report on the realized returns with each of their five investment decisions. for each decision, the individuals were asked to state to what degree they are satisfied and to what degree they are positively or negatively surprised. afterwards, the individuals were asked to make a final investment choice. 3.1. incentives the participants received a base payment of 13.25 euros and a payoff based on one of the five investment decisions. the relevant decision was selected randomly. the payoff in the selected decisions depended on the preferred exposure to the risky asset and the return of the risky asset, which was drawn from the previously communicated distribution of the risky asset. additionally, the participants could gain or lose 2% (20 cents) of their initial endowment with every correct (incorrect) answer to the risk awareness questions. all questions that were relevant for the final payment were marked in red, and the instructions stated that this indicates payoff relevance. the median completion time was 27 min, excluding the three-day break. the total payments varied between 21.75 and 27.65 euros with an average of 26.20 euros. 3.2. participants the survey was conducted online1 in january 2014 with 439 germans aged between 18 and 65. the sample was provided by a professional market research agency and included individuals from a national panel of over 200,000 germans. socioeconomic questions were used to apply a quota sampling procedure for selecting participants from the general population to ensure the representativeness of the sample. we used the time those individuals took to read the instructions and answer the questions to exclude those that are most likely to provide random answers.2 the filtered sample includes 320 individuals. a summary of their socioeconomic profiles is provided in table b-1 in the appendix. most of the individuals have no children, have a high school degree, 345k. bachmann et al. / financial services review 26 (2017) 339–365 work as employees without supervisory responsibilities, have a monthly net income between 1,300 to 2,600 euros and have a financial wealth of between 2,500 to 10,000 euros. 3.3. definitions the risk profiler is a composite of several measures that predicts investment risk taking. nobre and grable (2015) suggest that risk profiling questions should consist of questions assessing the risk need (the amount of risk required to meet a particular financial goal), the risk capacity (the client’s ability to absorb a possible financial loss resulting from the financial risk taken) and the financial risk tolerance (the individual’s willingness to accept uncertainty related to the outcome of a financial decision). we focus on the assessment of the risk tolerance since the first two concepts are purely financial issues that can be managed with financial planning tools. in our setting, the investor’s risk tolerance is a multidimensional construct that reflects an investor’s attitude toward risk. we use different notions of risk that refer to the uncertainty of payoffs and to payoffs below a certain reference point. the attitude toward uncertainty is usually called risk aversion. the attitude toward payoffs below a certain reference point is called loss aversion. we consider the investor’s risk awareness as an additional driver of risk taking. it measures the discrepancy between the perceived and the true risk-reward characteristics of the chosen investment. 3.4. questions design the questions used in our survey assess an individual’s risk tolerance, risk awareness and investment experience, along with socio-economic and demographic characteristics as potential drivers of financial risk taking. the questions are provided in the appendix. the questions assessing an investor’s risk tolerance address different notions of risk. in line with the results of morrison and oxoby (2014), who find that loss aversion influences decisions involving risk beyond the effects of risk aversion, we assess risk aversion and loss aversion as separate descriptions of an individual’s risk tolerance. the estimation of individual’s risk aversion is based on self-assessments. an individual’s loss aversion is estimated with a price table task, which is similar to the one used by holt and laury (2002). in this task, the individuals were asked to make eight binary comparisons. in each comparison, they were asked to select either the safe option or the risky option. a control question describing the individual’s choice asks individuals to confirm or revise their decision. the question assessing individual risk awareness aimed to evaluate an investor’s understanding of the return distribution of the risky asset. we used multiple choice questions with individually randomized answers. in addition to answering the questions, we asked individuals to state their confidence in the correctness of their answers. to compare the different question types, we apply the same seven-point likert scale to all questions.3 for three questions, it was not appropriate to use a likert scale. in these cases, we ensured that the questions had seven answer options with equal psychological distance, that is, we used numbers such as years for the financial experience questions, which precisely defined the steps between the answers. in the empirical analysis, we treated the answers as an interval-based numerical dataset.4 346 k. bachmann et al. / financial services review 26 (2017) 339–365 4. results 4.1. risk taking along the discovery process our experimental design is based on the idea that individuals facing investment decisions are involved in a process of discovery of their willingness to take risks. to test this conjecture, we first consider the individual changes in risk taking between two subsequent stages of the decision process. because participants had to allocate their wealth between a risky and a risk-free asset, we take the percentage allocated in the risky asset as the measure of risk taking. the summary statistics reported in table 2 suggest that at all stages, about half of all individuals change their risk-taking behavior. except in the stage after the experience sampling, where individuals increase their risk taking by 4% on average, risk-taking revisions do not have a clear direction. next, we test whether the risk-taking revisions are associated with individual characteristics observable in the corresponding stages. the relevant characteristics of the stages that differ among individuals are linked to (1) the demand for information on the risky asset, (2) an improvement in the risk awareness after the experience sampling, and (3) the average portfolio return with past investment decisions, expectations and satisfaction with these returns. table 3 reports summary statistics on risk-taking revisions between two subsequent decisions. it also includes the results of independent tests on the association of individual characteristics observed in different stages of the decision process and the risk-taking revisions. we observe that individuals acquiring information on the risky asset are more likely to change their risk taking. additional kruskal-wallis tests, which are not reported here, suggest that the description type (verbal, graphical, and statistical) is not associated with either the risk-taking revisions or with the level of risk taking in the second stage. furthermore, we observe that individuals who improve their awareness of extreme outcomes and extreme positive outcomes after the experience sampling take more risks on average. finally, we observe that individuals change risk taking after receiving information on the outcomes of previous decisions. in particular, individuals who receive a bad (nonpositive) outcome on average reduce their risk taking, while individuals who receive a good (positive) outcome with previous decisions take more risks on average. significantly more individuals change their risk taking after bad outcomes than individuals who change their risk taking after good table 2 risk taking revisions percentage of individuals changing risk taking risk taking revisions mean (in%) sd (in%) min (in%) max (in%) stage2-stage1 (after ambiguity reduction) 55.9% �0.067 14.43 �57 50 stage3-stage2 (after risk profiling questions) 46.3% 0.214 12.39 �55 55 stage4-stage3 (after experience sampling) 61.3% 4.019 16.07 �90 65 stage5-stage4 (after break) 54.1% �1.299 13.09 �60 50 stage6-stage5 (after outcome feedback) 56.2% 0.189 12.87 �50 55 347k. bachmann et al. / financial services review 26 (2017) 339–365 t ab le 3 r is k ta ki ng re vi si on s an d in di vi du al ch ar ac te ri st ic s l ev el of ri sk ta ki ng re vi si on s m ea n (i n% ) sd (i n% ) m in (i n% ) m ax (i n% ) k ru sk al -w al lis t es t (p -v al ue ) in di vi du al s ch an gi ng ri sk ta ki ng pe ar so n � 2 -t es t (p -v al ue ) a cq ui re in fo rm at io n n o � 0. 48 14 .5 6 � 50 50 0. 47 y es 0. 08 14 .4 2 � 57 50 0. 51 0 0. 59 0. 03 6 r is k aw ar en es s q1 (e xt re m e re tu rn s) d et er io ra tio n � 2. 78 17 .0 6 � 48 30 0. 65 n o ch an ge 4. 03 16 .2 2 � 90 65 0. 62 im pr ov em en t 6. 33 14 .7 5 � 30 53 0. 07 0 0. 57 0. 66 5 q2 (l ow re tu rn s) d et er io ra tio n 4. 53 13 .7 2 � 20 30 0. 68 n o ch an ge 4. 10 15 .9 5 � 90 65 0. 61 im pr ov em en t 3. 19 18 .2 8 � 48 65 0. 87 2 0. 62 0. 79 0 q3 (e xt re m e lo w re tu rn s) d et er io ra tio n 4. 05 12 .6 1 � 30 35 0. 59 n o ch an ge 4. 82 15 .5 9 � 40 65 0. 61 im pr ov em en t � 6. 95 24 .4 1 � 90 10 0. 35 1 0. 63 0. 95 9 q4 (e xt re m e hi gh re tu rn s) d et er io ra tio n 6. 60 15 .0 5 � 20 50 0. 76 n o ch an ge 3. 16 15 .6 8 � 90 65 0. 59 im pr ov em en t 9. 26 19 .2 4 � 30 65 0. 06 8 0. 65 0. 24 9 q5 (v ol at ili ty ) d et er io ra tio n 4. 32 14 .6 0 � 48 35 0. 66 n o ch an ge 4. 21 16 .0 8 � 90 65 0. 61 im pr ov em en t 2. 41 17 .9 6 � 40 60 0. 42 0 0. 56 0. 69 0 q6 (a ve ra ge re tu rn ) d et er io ra tio n 8. 27 16 .3 9 � 15 63 0. 63 n o ch an ge 3. 73 16 .3 2 � 90 65 0. 61 im pr ov em en t 2. 57 11 .6 9 � 30 35 0. 46 4 0. 65 0. 87 4 a ve ra ge ou tc om e n on -p os iti ve � 17 .9 4 14 .5 2 � 50 0 0. 89 po si tiv e 1. 27 11 .9 5 � 50 55 0. 00 0 0. 54 0. 00 3 e xp ec ta tio ns c om fo rt ed 1. 61 11 .2 4 � 50 45 0. 52 d is ap po in te d � 2. 68 15 .3 0 � 50 55 0. 00 1 0. 65 0. 01 6 sa tis fa ct io n c om fo rt ed 1. 58 12 .6 0 � 50 55 0. 54 d is ap po in te d � 3. 22 12 .9 4 � 48 30 0. 00 2 0. 61 0. 14 9 t he ta bl e pr es en ts su m m ar y st at is tic s of ri sk ta ki ng re vi si on s as w el l as th e pe rc en ta ge of in di vi du al s ch an gi ng ri sk ta ki ng ov er tw o su bs eq ue nt de ci si on s. it al so re po rt s th e re su lts of in de pe nd en t te st s on th e as so ci at io n be tw ee n ri sk ta ki ng re vi si on s an d in di vi du al s’ ch ar ac te ri st ic s in di ff er en t st ag es . in th e ca se of va ri ab le s w ith tw o ca te go ri es , th e pe ar so n � 2 -t es t is eq ui va le nt to th e on esi de s fi sh er ex ac t te st . 348 k. bachmann et al. / financial services review 26 (2017) 339–365 outcomes. similarly, individuals disappointed by their previous returns tend to reduce their risk taking, while individuals pleased with their previous returns tend to increase their risk taking. so far, we find that the stages of the decision process under consideration are associated with significant changes in individual risk taking. however, do individuals learn something about their willingness to take risks by going through the various stages? to answer this question, we asked individuals to state which investment decision they consider the best one. to avoid outcome bias, we asked this question just before the outcomes of their investment decisions were revealed to them. approximately 30% of the participants who revised their risk taking stated that their best decision was the last one. moreover, the association between risk-taking revisions and choosing the last decision as the best one is statistically significant (fisher exact test, p value: 0.02). we conclude that the provided decision stages were helpful for participants involved in a process of discovering their willingness to take risks. overall, we find that individual risk taking changes significantly after receiving information on the risky asset, while the direction of risk taking depends on individual risk tolerance. moreover, the individual risk taking increases significantly after improving risk awareness in the experience sampling task. although the outcome of previous decisions should not change risk taking because outcomes cannot be accumulated over stages, there are significant differences in the risk-taking revisions of individuals experiencing good or bad outcomes on average with their previous decisions. finally, we find that individuals involved in discovering their willingness to take risks learn successfully over the different stages of the decision process. 4.2. explaining risk taking in this section, we analyze the importance of individual risk tolerance, risk awareness and financial experience as drivers of investment risk taking. the evaluation of these factors is based on a factor analysis. the analysis shows that the answers to the twenty questions evaluating individuals’ risk tolerance, risk awareness and financial experience can be summarized by three different factors, which are not correlated with each other (see appendix c for more details). in the following, we use these factors in ordinary least square regressions to test whether they can explain risk taking as expressed by the amount of wealth that individuals invest in the risky asset at each stage. previous research suggests that demographic and socioeconomic characteristics influence an individual’s risk tolerance and risk taking (see, e.g., grable and lytton, 2003; sundén and surette, 1998; xiao, 1996). to take this into account, we use age, gender, number of children, education, job position, income, and wealth as controls in each regression. as an additional independent variable, we include an indicator variable that captures whether the individual acquires information on the risky asset. in the last decision, we include the average return of the previous investment decisions as a further independent variable. the estimation results are reported in table 4. we observe that among the three factors capturing the individuals’ risk tolerance, risk awareness and financial experience, only the risk tolerance factor explains risk-taking 349k. bachmann et al. / financial services review 26 (2017) 339–365 t ab le 4 r is k ta ki ng dr iv er s in ve st m en t in th e ri sk y as se t (0 –1 00 ) d ec is io n 1 r is k to le ra nc e 8. 62 8* ** 9. 65 1* ** 8. 62 3* ** 9. 62 2* ** (1 .1 98 20 4) (1 .0 95 ) (1 .2 09 ) (1 .0 98 ) r is k aw ar en es s � 1. 16 6 0. 51 4 � 0. 39 6 0. 75 9 (1 .4 92 ) (1 .4 34 ) (1 .3 84 ) (1 .2 89 ) fi na nc ia l ex pe ri en ce � 0. 45 1 1. 55 5 0. 34 7 0. 98 5 (1 .6 04 4) (1 .3 31 ) (1 .4 86 ) (1 .2 01 ) a cq ui re 9. 15 9* ** 10 .2 01 ** * 9. 65 2* ** 9. 76 1* ** 8. 84 9* * 10 .3 49 ** * 9. 48 6* ** 9. 73 8* ** in fo rm at io n (2 .4 90 ) (2 .3 04 ) (2 .8 39 ) (2 .7 96 ) (2 .7 43 ) (2 .5 70 ) (2 .6 52 ) (2 .5 19 ) c on tr ol s y es n o y es n o y es n o y es n o a dj us te d r 2 0. 25 06 0. 22 97 0. 11 66 0. 04 13 3 0. 11 5 0. 04 50 5 0. 24 57 0. 22 74 d ec is io n 2 r is k to le ra nc e 9. 35 3* ** 10 .2 28 ** * 9. 46 4* ** 10 .2 18 ** * (1 .2 33 8) (1 .1 37 ) (1 .2 44 ) (1 .1 39 ) r is k aw ar en es s � 0. 10 56 1. 34 9 0. 75 9 1. 61 2 (1 .5 51 ) (1 .4 94 ) (1 .4 24 4) (1 .3 37 ) fi na nc ia l ex pe ri en ce 0. 04 71 6 1. 65 1 0. 97 1 1. 00 5 (1 .6 65 ) (1 .3 88 ) (1 .5 29 ) (1 .2 45 ) a cq ui re 10 .2 95 ** * 10 .7 61 ** * 10 .1 57 ** * 9. 67 6* * 10 .1 08 ** * 10 .9 19 ** * 10 .1 27 ** * 9. 64 5* ** in fo rm at io n (2 .5 64 ) (2 .3 92 ) (2 .9 51 ) (2 .9 13 ) (2 .8 48 ) (2 .6 80 ) (2 .7 28 ) (2 .6 11 ) c on tr ol s y es n o y es n o y es n o y es n o a dj us te d r 2 0. 27 0. 23 68 0. 12 34 0. 04 44 5 0. 12 34 0. 04 62 5 0. 26 66 0. 23 73 d ec is io n 3 r is k to le ra nc e 9. 39 8* ** 10 .1 72 ** * 9. 40 2* ** 10 .1 96 ** * (1 .2 33 9) (1 .1 40 ) (1 .2 45 ) (1 .1 45 ) r is k aw ar en es s � 1. 13 8 0. 23 1 � 0. 29 8 0. 56 5 (1 .5 51 2) (1 .4 97 ) (1 .4 25 ) (1 .3 44 ) fi na nc ia l ex pe ri en ce � 0. 48 7 0. 34 4 0. 38 9 � 0. 24 7 (1 .6 67 ) (1 .3 92 ) (1 .5 31 ) (1 .2 52 ) a cq ui re 10 .7 67 7* ** 10 .7 39 ** * 11 .2 28 2* ** 10 .5 15 ** * 10 .4 30 ** * 10 .7 35 ** * 11 .0 49 ** * 10 .2 73 ** * in fo rm at io n (2 .5 65 ) (2 .3 98 ) (2 .9 51 ) (2 .9 20 ) (2 .8 50 ) (2 .6 87 ) (2 .7 31 ) (2 .6 26 ) c on tr ol s y es n o y es n o y es n o y es n o a dj us te d r 2 0. 27 1 0. 23 41 0. 12 48 0. 04 17 9 0. 12 34 0. 04 19 0. 26 62 0. 22 97 d ec is io n 4 r is k to le ra nc e 9. 63 5* ** 10 .7 19 ** * 9. 85 7* ** 10 .6 7* ** (1 .3 87 2) (1 .2 48 ) (1 .3 92 6) (1 .2 45 ) (c on ti nu ed on ne xt pa ge ) 350 k. bachmann et al. / financial services review 26 (2017) 339–365 t ab le 4 (c on tin ue d) in ve st m en t in th e ri sk y as se t (0 –1 00 ) r is k aw ar en es s 1. 28 3 2. 74 1 2. 00 13 2. 68 6 (1 .6 80 ) (1 .6 00 ) (1 .5 60 7) (1 .4 43 ) fi na nc ia l ex pe ri en ce 0. 77 89 1. 72 2 1. 90 91 1. 14 6 (1 .8 44 8) (1 .5 04 ) (1 .7 15 3) (1 .3 54 ) a cq ui re 12 .6 46 ** * 13 .1 57 ** * 11 .3 1* ** 10 .7 72 ** * 12 .1 54 ** * 12 .6 77 ** * 12 .1 37 ** * 11 .5 59 ** * in fo rm at io n (2 .8 82 ) (2 .6 28 ) (3 .2 24 ) (3 .0 76 ) (3 .1 48 3) (2 .9 14 ) (3 .0 07 4) (2 .7 78 ) c on tr ol s y es n o y es n o y es n o y es n o a dj us te d r 2 0. 22 79 0. 22 84 0. 09 95 1 0. 05 76 3 0. 09 82 4 0. 05 28 3 0. 22 98 0. 23 39 d ec is io n 5 r is k to le ra nc e 9. 17 6* ** 10 .1 37 ** * 9. 08 7* ** 10 .1 53 ** * (1 .2 62 ) (1 .1 44 ) (1 .2 72 ) (1 .1 47 ) r is k aw ar en es s � 0. 99 3 0. 88 0 � 0. 50 3 0. 87 3 (1 .5 40 4) (1 .4 81 1) (1 .4 26 ) (1 .3 30 ) fi na nc ia l ex pe ri en ce � 1. 83 7 0. 09 3 � 0. 97 4 � 0. 41 (1 .6 87 ) (1 .3 89 ) (1 .5 67 ) (1 .2 48 ) a cq ui re 10 .4 14 ** * 11 .5 4* ** 10 .2 58 ** * 10 .3 71 ** * 9. 29 1* * 10 .9 38 ** * 10 .4 14 ** * 10 .9 47 ** * in fo rm at io n (2 .6 23 ) (2 .4 08 ) (2 .9 55 7) (2 .8 47 ) (2 .8 80 ) (2 .6 92 ) (2 .7 47 ) (2 .5 61 ) c on tr ol s y es n o y es n o y es n o y es n o a dj us te d r 2 0. 24 31 0. 23 35 0. 10 46 0. 04 46 1 0. 10 7 0. 04 35 6 0. 23 91 0. 22 99 d ec is io n 6 r is k to le ra nc e 9. 42 23 ** * 10 .1 43 ** * 9. 39 97 ** * 10 .1 58 6* ** (1 .3 30 2) (1 .1 84 ) (1 .3 41 ) (1 .1 86 ) r is k aw ar en es s � 0. 09 9 1. 43 7 0. 42 8 1. 43 0 (1 .6 17 ) (1 .5 22 ) (1 .5 02 ) (1 .3 75 ) fi na nc ia l ex pe ri en ce � 1. 62 5 0. 07 6 � 0. 66 0 � 0. 44 1 (1 .7 72 ) (1 .4 28 ) (1 .6 51 ) (1 .2 90 5) a cq ui re 9. 99 2* ** 10 .3 46 ** * 9. 36 8* * 8. 82 4* * 8. 90 5* * 9. 74 3* ** 9. 60 7* * 9. 39 7* ** in fo rm at io n (2 .7 63 ) (2 .4 92 ) (3 .1 03 ) (2 .9 25 ) (3 .0 24 ) (2 .7 68 ) (2 .8 96 ) (2 .6 47 ) a ve ra ge re tu rn 2. 86 2* ** 2. 91 1* ** 3. 21 5* ** 3. 37 1* ** 3. 20 4* ** 3. 37 9* ** 2. 86 9* ** 2. 89 9* ** (0 .3 52 ) (0 .3 31 ) (0 .3 74 ) (0 .3 56 ) (0 .3 73 ) (0 .3 55 ) (0 .3 54 ) (0 .3 33 ) c on tr ol s y es n o y es n o y es n o y es n o a dj us te d r 2 0. 19 57 0. 21 37 0. 05 45 0. 03 42 0. 05 73 0. 03 15 0. 19 08 0. 21 16 t he ta bl e re po rt s th e es tim at io n re su lts of or di na ry le as ts qu ar e re gr es si on s w ith th e pe rc en ta ge of w ea lth in ve st ed in th e ri sk y as se t( 0 –1 00 ) as a de pe nd en t va ri ab le in ea ch re gr es si on .s ta nd ar ds er ro rs ar e gi ve n in pa re nt he se s. a ge ,g en de r, nu m be r of ch ild re n, ed uc at io n, jo b po si tio n, in co m e, an d w ea lth ar e us ed as co nt ro ls . ** *, ** , an d *i nd ic at e si gn ifi ca nc e le ve ls of 1% , 5% , an d 10 % , re sp ec tiv el y. 351k. bachmann et al. / financial services review 26 (2017) 339–365 behavior at each stage. its impact on risk taking is stable over different decision modes and is robust to demographic and socio-economic characteristics used as controls. the influence of the factors capturing individuals’ risk awareness and financial experience on risk taking is statistically not significant. interestingly, we observe significant and robust differences in the risk taking associated with the demand for information on the risky asset. individuals who acquire information on the risky assets invest approximately 10% more in the risky asset than individuals who do not acquire information on the risky asset. although individuals cannot accumulate returns of subsequent investment decisions, their risk taking in the last stage changes with the average outcome of their previous investment decisions. 4.3. predicting risk taking in the following, we analyze which combination of single questions has the strongest predictive power for risk-taking behavior. we apply a cross-validation analysis.5 table 5 reports the estimated coefficients of the variables with significant predicting power. the risk awareness assessed before (after) the experience sampling is used to predict the first (last) three investment decisions. the average return on past investment decisions is used only in the prediction of the last decision. we observe that risk taking at all stages is best predicted by individuals’ loss aversion. its assessment is, however, critical. while a general loss aversion formulation is not helpful in predicting risk taking, a verbal question specifying returns and a quantitative version based on a lottery question are able to predict risk taking in all decision modes. by contrast, risk aversion measures based on self-assessment cannot be used to predict risk taking. another important predictor of risk taking is the returns of past decisions. although the odds of the outcomes do not change over time and returns cannot be accumulated, the participants take significantly more (less) risks after observing positive (negative) average returns with their past investment decisions. in the context of the assessed risk tolerance, demographic and socio-economic characteristics have limited predictive power. to shed some light on them, we repeat the crossvalidation analysis while excluding the risk tolerance and the investment experience questions. table 6 reports the estimation results. we observe that among the demographic and socioeconomic characteristics, gender is the most reliable variable in predicting risk taking. females are less willing to take risks. as in the previous analysis, age can be a good predictor of risk taking but only in certain situations, while income loses predicting power. the effect of previous returns on subsequent risktaking remains strong. we conclude that assessed individuals’ loss aversion is the most powerful predictor of risk taking at all stages and in the context of all other questions that we use with a potential impact on risk taking. we find that self-assessed knowledge, experience, and risk aversion are not useful in predicting individual risk taking. finally, recommending less risky investment can be optimal for female individuals if there is no option to assess their risk tolerance. 352 k. bachmann et al. / financial services review 26 (2017) 339–365 t ab le 5 pr ed ic tin g po w er of si ng le qu es tio ns d ec is io n 1 d ec is io n 2 d ec is io n 3 d ec is io n 4 d ec is io n 5 d ec is io n 6 g en er al ri sk ta ki ng 3. 55 32 * (1 .3 72 5) g en er al fin an ci al ri sk ta ki ng c ur re nt fin an ci al ri sk ta ki ng pa st fin an ci al ri sk ta ki ng g en er al fin an ci al lo ss av er si on v er ba l fin an ci al lo ss av er si on 6. 52 5* ** (1 .0 82 ) 6. 43 8* ** (1 .1 15 ) 6. 37 5* ** (1 .1 02 ) 7. 53 7* ** (1 .2 04 ) 8. 33 4* ** (1 .1 25 ) 5. 04 0* ** (1 .2 68 ) q ua nt . fin an ci al lo ss av er si on 5. 63 2* ** (1 .0 96 ) 5. 08 8* ** (1 .1 08 ) 7. 90 5* ** (1 .1 02 ) 8. 04 ** * (1 .2 04 ) 4. 31 7* ** (1 .1 40 ) 3. 94 1* ** (1 .0 86 ) fi na nc ia l in ve st in g fo r th ri ll pr of es si on al ex pe ri en ce in fin an ce c on su m pt io n of fin an ci al ne w s fi na nc ia l kn ow le dg e st at is tic al kn ow le dg e fi na nc ia l tr ad in g ex pe ri en ce t ra di ng fr eq ue nc y 3. 53 9* ** (1 .0 64 ) 3. 57 6* * (1 .1 74 ) r is k aw ar en es s 1 r is k aw ar en es s 2 r is k aw ar en es s 3 r is k aw ar en es s 4 a ge cl as s � 2. 75 4* * (0 .9 98 ) fe m al e n um be r of ch ild re n e du ca tio n pr of es si on al st at us m on th ly in co m e 2. 67 4* * (1 .0 09 ) w ea lth a ve ra ge pa st re tu rn 8. 13 7* ** (0 .9 71 ) a cq ui re in fo rm at io n 3. 95 6* ** (1 .0 11 ) 4. 31 6* ** (1 .0 07 ) 4. 63 7* ** (1 .0 51 ) 2. 74 5* * (1 .0 42 ) a dj us te d r 2 0. 27 89 0. 34 91 0. 29 95 0. 29 78 0. 28 89 0. 43 63 t he ta bl e re po rt s th e es tim at es of cr os sva lid at io n an al ys is w ith th e pe rc en ta ge of w ea lth in ve st ed in th e ri sk y as se t (0 –1 00 ) as a de pe nd en t va ri ab le in ea ch re gr es si on . st an da rd s er ro rs ar e gi ve n in pa re nt he se s. ** *, ** , an d * in di ca te si gn ifi ca nc e le ve ls of 1% , 5% , an d 10 % , re sp ec tiv el y. 353k. bachmann et al. / financial services review 26 (2017) 339–365 t ab le 6 pr ed ic tin g po w er of de m og ra ph ic an d so ci oe co no m ic ch ar ac te ri st ic s d ec is io n 1 d ec is io n 2 d ec is io n 3 d ec is io n 4 d ec is io n 5 d ec is io n 6 a ge cl as s � 3. 41 6* * (1 .1 61 ) � 2. 92 1* (1 .1 62 ) fe m al e � 3. 04 4* * (1 .1 45 ) � 3. 84 3* * (1 .1 60 ) � 3. 77 2* * (1 .1 71 ) � 4. 17 3* ** (1 .1 65 ) � 3. 40 1* * (1 .0 78 ) n um be r of ch ild re n e du ca tio n pr of es si on al st at us m on th ly in co m e w ea lth � 2. 70 3* (1 .2 01 ) a ve ra ge pa st re tu rn 10 .4 63 ** * (1 .0 78 ) a cq ui re in fo rm at io n 4. 73 9* ** (1 .1 60 ) 4. 99 5* ** (1 .1 71 ) 5. 53 5* ** (1 .2 88 ) 4. 86 1* ** (1 .1 70 ) a dj us te d r 2 0. 03 5 0. 09 2 0. 07 2 0. 05 2 0. 25 2 t he ta bl e re po rt s th e es tim at es of cr os sva lid at io n an al ys is w ith th e pe rc en ta ge of w ea lth in ve st ed in th e ri sk y as se t (0 –1 00 ) as a de pe nd en t va ri ab le in ea ch re gr es si on . st an da rd s er ro rs ar e gi ve n in pa re nt he se s. ** *, ** , an d * in di ca te si gn ifi ca nc e le ve ls of 1% , 5% , an d 10 % , re sp ec tiv el y. 354 k. bachmann et al. / financial services review 26 (2017) 339–365 5. discussion and implications we found strong evidence that individuals’ risk tolerance is a more powerful predictor of risk taking than investors’ self-assessed investment experience or risk awareness. more important, we found that the association between risk tolerance and risk-taking remains significant over different decision stages related to reduced ambiguity, extended experience and feedback on previous decisions. with respect to the impact of these decision stages on risk taking, we find that reduced ambiguity influences risk taking, but it does not necessary increase it, as documented by antoniou et al. (2015). however, we find that extending experience with the risky asset through simulations increases risk taking, which is in line with the results of kaufmann et al. (2013) and bradbury et al. (2014). furthermore, we observe that the average return of previous decisions influences the subsequent risk taking, although the odds of the possible outcomes remain the same and returns cannot be accumulated. as suggested by fischhoff (1975), this behavior can be explained with a stronger focus on the risks (returns) after negative (positive) returns. it is also possible that individuals use outcomes to judge the quality of their previous decisions, as suggested by baron and hershey (1988). in this case, positive (negative) outcomes would increase (decrease) the confidence in the decision quality and individuals would increase (decrease) subsequent risk taking, as we observe in our experiment. risk tolerance measures are usually multidimensional, and the components can be correlated (guillemette et al., 2015). we analyzed the predicting power of the components and found that an individual’s loss aversion is the most powerful predictor of risk taking in all decision modes. this supports previous findings that loss aversion measures are more powerful in explaining risk taking than the arrow-pratt measures (guillemette et al., 2012). moreover, we found that self-assessed risk tolerance has no predicting power. among the questions assessing investment experience, we found that only the question related to the trading frequency can predict risk taking in some decision modes. overall, we do not find a positive relationship between investment experience and risk taking, which is in contrast to the results of corter and chen (2006). this can be explained with differences in the measures. while corter and chen (2006) ask individuals to evaluate their investment experience relative to other individual investors, our measures are based on individual trading experience. several studies suggest that risky asset ownership can be explained by demographic and socioeconomic variables (see for example grable and lytton, 2003; sundén and surette, 1998; xiao, 1996). we found that among the assessed demographic and socioeconomic characteristics, only gender can predict risk taking in most decision modes but only if the individual risk tolerance cannot be assessed. if the risk tolerance is assessed, gender loses its predicting power. this observation is in line with the results of wärneryd (1996) and grable and lytton (2003). our results have important implications for the design of risk profilers. to predict risk taking, the latter should include questions assessing the individual risk tolerance, which should include a question on the investor’s loss aversion. gender is a useful predictor of risk taking only if the risk tolerance cannot be assessed. by contrast, self-assessed investment experience is not a reliable predictor of risk taking, but the stated trading frequency can be used as a proxy for investment experience when predicting risk taking. 355k. bachmann et al. / financial services review 26 (2017) 339–365 another important predictor of risk taking is the past investment return. the latter influences the desired risk taking beyond the level based on the assessed risk tolerance. hence, in addition to assessing an individual’s risk tolerance, a risk profiler should either consider an investor’s misperception of risk, or the latter should be corrected through additional measures. otherwise, investors will be willing to revise their risk taking for no good reason. 6. conclusions the optimal amount of risk an investor should take is one of the most important issues in wealth management. since answering this question through real-life investment experience can be costly, several studies suggest risk profiling measures and prove their suitability by showing that they can explain risk taking. this article studied whether and how the suitability of different risk profiling measures varies if individuals are involved in a process of discovering their willingness to take risks. this process included situations with reduced ambiguity, extended experience and feedback on the outcomes of previous decisions, which reflect the experience of private investors. the results show that private investors are often involved in the process of discovering their willingness to take risks. the average risk-taking behavior changes over the different stages of the learning process, but it is always associated with a composite measure of the individual’s risk tolerance. overall, we did not find any significant association between risk taking and investment experience outside of the study, although sometimes the self-reported trading frequency can predict risk taking. letting investors experience the riskiness of different asset allocations through simulations reduces biases in the risk-reward perception of the investors, but the investors’ risk tolerance and loss aversion in particular remains a significant predictor of the individual risk-taking behavior at all stages of the decision process. by contrast, self-assessed risk tolerance measures appear to not be suitable for predicting risk taking at any stage of the decision process. these results suggest that risk profiling measures should be selected carefully. when investors are involved in a process of discovering their willingness to take risks, some measures are more stable predictors of risk taking than others and should not be missed by risk profilers. by contrast, other measures may predict risk taking in only some or none of the stages, which limits their suitability. using measures that predict risk taking at all stages of the discovery process increases the probability that clients remain satisfied with the derived recommendations. at the same time, making recommendations based on questions that consistently predict risk taking at all stages of the discovery process should support the advisors’ confidence that these recommendations match the clients’ risk tolerance and do not create misperceptions that are corrected over time. notes 1 online studies allow effective access to a sample of the general population. moreover, they allow tracking of the time individuals spend on each question. 2 we excluded all individuals who needed less than one and a half minutes to read the instructions and less than 15 minutes to finish the survey. 356 k. bachmann et al. / financial services review 26 (2017) 339–365 3 for the quantitative financial loss aversion question, we presented eight answer options. the last two possibilities were merged because only three individuals used the seventh possibility in their choices. the results of a robustness test with the combined answer possibilities shows that the results remain stable. 4 according to the literature, likert scales can be considered an interval-based measure, that is, parametric analysis is appropriate (carifio and perla, 2007; norman, 2010; pell, 2005). 5 the analysis uses recursive feature elimination that removes the least important predictors of a model step-by-step. first, a model with all predictors is trained on a training set. the model is then used to predict the test set. the least important predictor is then removed from the model, and the whole procedure is repeated for all the subsequent subsets of predictors. to avoid any selection bias (e.g., over-fitting predictors and samples), the train and test data sets are resampled with a 10-fold cross-validation. after the resampling iterations, the most appropriate number of predictors is determined based on the resampling output. the predictors with the best rankings across all the resampling iterations are then used to fit the final model. appendix a experience sampling fig. a-1: illustration of the experience sampling. 357k. bachmann et al. / financial services review 26 (2017) 339–365 b socioeconomic and demographic characteristics table b-1: sample description n percentage variable type age categorical variable 18–24 54 16.88% 0 25–34 44 13.75% 1 35–44 70 21.88% 2 45–54 82 25.63% 3 55–64 70 21.88% 4 gender indicator variable male 147 45.94% 0 female 173 54.06% 1 number of children ordinal variable 0 201 62.81% 0 1 62 19.38% 1 2 43 13.44% 2 3 10 3.13% 3 4 4 1.25% 4 education categorical variable primary school 10 3.13% 0 secondary school 65 20.31% 1 high school 96 30.00% 2 bachelor 39 12.19% 4 master 45 14.06% 5 phd 11 3.44% 6 other education 53 16.56% 7 no education 1 0.31% 8 professional status categorical variable self-employed/in family business 37 11.56% 0 employee in top management 18 5.63% 1 employee with leadership position 65 20.31% 2 employee without leadership position 108 33.75% 3 apprentice 47 14.69% 4 unemployed 45 14.06% 5 monthly income categorical variable 0–1,300 euro 60 18.75% 0 1,300–2,600 euro 94 29.38% 1 2,600–3,600 euro 74 23.13% 2 3,600–5,000 euro 54 16.88% 3 5,000–18,000 euro 11 3.44% 4 � 18,000 euro 1 0.31% 5 no answer 26 8.13% wealth categorical variable 0–500 euro 47 14.69% 0 500–2,500 euro 44 13.75% 1 2,500–10,000 euro 59 18.44% 2 10,000–30,000 euro 46 14.38% 3 30,000–65,000 euro 32 10.00% 4 65,000–175,000 euro 30 9.38% 5 175,000 euro 11 3.44% 6 no answer 51 15.94% 358 k. bachmann et al. / financial services review 26 (2017) 339–365 c factor analysis d questions the following questions are assessed on a seven-point scale ranging from “not true at all” to “absolutely true.” general risk tolerance: in general, i am a risk loving person. general financial risk tolerance: my risk tolerance when i am investing money is generally high. current financial risk tolerance: my current willingness to take risk in financial decisions is low. past financial risk tolerance: my risk tolerance in financial decisions was high in the past. general financial loss aversion: when i am confronted with an important financial decision then i do concern more with the possible losses than with the possible gains. table c-1: factor loadings with a varimax rotation factors (before experience sampling) factors (after experience sampling) risk tolerance financial experience risk awareness risk tolerance financial experience risk awareness general risk taking 0.73 0.18 �0.11 0.73 0.19 �0.11 general financial risk taking 0.87 0.29 �0.09 0.87 0.29 �0.04 current financial risk taking 0.65 0.15 �0.01 0.65 0.15 �0.02 past financial risk taking 0.56 0.34 �0.16 0.56 0.34 �0.17 general loss aversion 0.4 0.16 0.03 0.4 0.16 0.05 verbal loss aversion 0.74 0.11 �0.16 0.75 0.11 �0.09 quantitative loss aversion 0.49 0.11 0.13 0.5 0.11 0.17 financial investing for thrill 0.6 0.49 �0.12 0.61 0.48 �0.08 professional experience in finance 0.07 0.59 �0.14 0.08 0.58 �0.14 consumption of financial news 0.3 0.67 �0.02 0.3 0.66 �0.01 financial knowledge 0.33 0.74 0.01 0.32 0.75 �0.02 statistical knowledge 0.16 0.47 0.27 0.15 0.48 0.18 trading experience 0.15 0.74 0.14 0.14 0.75 0.13 trading frequency 0.44 0.63 0.02 0.43 0.64 0.01 risk awareness 1 0 0.07 0.72 0 0.09 0.77 risk awareness 2 �0.16 0 0.73 �0.1 0.02 0.68 risk awareness 3 �0.08 �0.03 0.62 �0.12 �0.08 0.75 risk awareness 4 0.05 0.02 0.89 0.14 0.04 0.88 ss loadings 3.81 3.05 2.45 3.82 3.08 2.55 proportion variance 0.21 0.17 0.14 0.21 0.17 0.14 cumulative variance 0.21 0.38 0.52 0.21 0.38 0.53 proportion explained 0.41 0.33 0.26 0.4 0.33 0.27 cumulative proportion 0.41 0.74 1 0.4 0.73 1 359k. bachmann et al. / financial services review 26 (2017) 339–365 verbal financial loss aversion: for a 50-percent chance to earn a high amount of money with a financial investment i would be willing to risk an equal amount of money. financial investing for thrill: i already invested very often money because of the thrill if its value will go up or down. professional experience in finance: i collected the big part of my professional experience in the financial sector (investment advisory, insurance, asset management, trustee, tax counseling, auditing, and accounting). consumption of financial news: i am very interest in economic news. financial knowledge: i can explain to a friend very well at which things he or she has to look after in the case of risky financial assets. statistical knowledge: i can explain to a friend very well what a probability distribution is. quantitative financial loss aversion: you have the choice to invest 500 ecu in a risky or in a risk-free asset. the wealth will be invested for one year. with an equal probability (each with 50%) the risky asset will result in a positive return of 50% p.a. (i.e., 250 ecu) or in a negative return. the risk-free asset will result in a positive return of 2% p.a. (i.e., 10 ecu). are you sure? in comparison to the risk-free asset (2%) you prefer the risky asset (50% chance to get a return of 50% p.a. [i.e., 250 ecu]) if the possible negative return is not higher than �8%. p.a; beginning at a possible negative return of �15% p.a. you prefer the risk-free asset. is this really your final decision? financial trading experience since how many years do you trade financial asset by yourself? y i have never traded financial assets by myself. y i buy and sell financial assets since about 1 to 3 years. y i buy and sell financial assets since about 4 to 6 years. y i buy and sell financial assets since about 7 to 9 years. risky asset decision risk-free asset 50% probability to get a return of 50% probability to get a return of i prefer the risky asset i prefer the risk-free asset 100% probability to get a return of 50% p.a. (250 ecu) �8% p.a. (�40 ecu) 2% p.a. (10 ecu) 50% p.a. (250 ecu) �15% p.a. (�75 ecu) 2% p.a. (10 ecu) 50% p.a. (250 ecu) �22% p.a. (�110 ecu) 2% p.a. (10 ecu) 50% p.a. (250 ecu) �29% p.a. (�145 ecu) 2% p.a. (10 ecu) 50% p.a. (250 ecu) �36% p.a. (�180 ecu) 2% p.a. (10 ecu) 50% p.a. (250 ecu) �43% p.a. (�215 ecu) 2% p.a. (10 ecu) 50% p.a. (250 ecu) �50% p.a. (�250 ecu) 2% p.a. (10 ecu) 360 k. bachmann et al. / financial services review 26 (2017) 339–365 y i buy and sell financial assets since about 10 to 12 years. y i buy and sell financial assets since about 13 to 15 years. y i buy and sell financial assets since more than 15 years. trading frequency how many times do you reallocate your financial assets, that is, how often do you buy and sell financial assets? y not at all y about every second year y about once a year y about twice a year y about four times a year y about every month y at least once a week risk awareness 1 the asset allocation with the highest probability for a strong negative and a strong positive return is: y 10% risk-free asset/90% risky asset y 40% risk-free asset/60% risky asset y 80% risk-free asset/20% risky asset y 35% risk-free asset/65% risky asset how confident are you with your answer?: not sure at all 1-2-3-4-5-6-7 absolutely sure. risk awareness 2 which asset allocation does not allow you to get a return higher than 2%? y 5% risk-free asset/95% risky asset y 0% risk-free asset/100% risky asset y 100% risk-free asset/0% risky asset y 75% risk-free asset/25% risky asset how confident are you with your answer?: not sure at all 1-2-3-4-5-6-7 absolutely sure. risk awareness 3 the asset allocation with the greatest risk for negative return in the worst out of 100 cases is: 361k. bachmann et al. / financial services review 26 (2017) 339–365 y 50% risk-free asset/50% risky asset y 40% risk-free asset/60% risky asset y 10% risk-free asset/90% risky asset y 45% risk-free asset/55% risky asset how confident are you with your answer?: not sure at all 1-2-3-4-5-6-7 absolutely sure. risk awareness 4 the asset allocation with the greatest potential for positive returns in the best out of 100 cases is: y 60% risk-free asset/40% risky asset y 20% risk-free asset/80% risky asset y 5% risk-free asset/95% risky asset y 15% risk-free asset/85% risky asset how confident are you with your answer? not sure at all 1-2-3-4-5-6-7 absolutely sure. risk awareness 5 the asset allocation with the smallest variation of returns is: y 20% risk-free asset/80% risky asset y 45% risk-free asset/55% risky asset y 80% risk-free asset/20% risky asset y 30% risk-free asset/70% risky asset how confident are you with your answer? not sure at all 1-2-3-4-5-6-7 absolutely sure. risk awareness 6 the asset allocation with the highest expected return is: y 5% risk-free asset/95% risky asset y 10% risk-free asset/90% risky asset y 40% risk-free asset/60% risky asset y 25% risk-free asset/75% risky asset how confident are you with your answer? not sure at all 1-2-3-4-5-6-7 absolutely sure. descriptions on the risky asset y graphical description 362 k. bachmann et al. / financial services review 26 (2017) 339–365 in the following graphic, you see the realized returns and their frequencies of 280 randomly drawn scenarios for the risky asset. higher bars mean higher frequencies. y verbal description the average return for the risky asset over all possible scenarios is 7% per annum. in 70 out of 100 scenarios one can expect that the return falls between �10% and 24% per annum, and in 30 out of 100 scenarios the return is lower than �10% and higher than 24% per annum. the positive or negative deviation from the average return is the same, and has the same probability. for example, a return of �3% has the same probability as a return of 17%. y statistical description the returns are normally distributed with a mean of 7% and a sd of 16%. the normal distribution has the property that returns close to 7% are more probable than those further away, and that the probability of a return of �3% has the same probability as a return of 17%. acknowledgment we are grateful for the valuable comments of the participants of the 2016 european financial management conference, the 2016 behavioural finance working group conference as well as the seminar participants at the university of basel, university of sussex, university of liechtenstein, university of zürich, university of innsbruck, and the university of muenster. financial support by the swiss national science foundation (grant 100018-149934) and the urpp-finreg of the university of zurich is gratefully acknowledged. references antoniou, c., harris, r. d. f., & zhang, r. (2015). ambiguity aversion and stock market participation: an empirical analysis. journal of banking & finance, 58, 57–70. fig. d-1: example of a return distribution used in the graphical description of the risky asset. 363k. bachmann et al. / financial services review 26 (2017) 339–365 baron, j., & hershey, j. c. (1988). outcome bias in decision evaluation. journal of personality and social psychology, 54, 569–579. barsky, r., & juster, f. (1997). preference parameters and behavioral heterogeneity: an experimental approach in the health and retirement study. the quarterly journal of economics, 112, 537–579. bradbury, m. a. s., hens, t., & zeisberger, s. (2014). improving investment decisions with simulated experience. review of finance, 19, 1–34. brayman, s., finke, m., grable, j., & griffin, p. (2017). current practices for risk profiling in canada and review of global best practices. journal of financial service professionals, 71, 71–81. carifio, j., & perla, r. j. (2007). ten common misunderstandings, misconceptions, persistent myths and urban legends about likert scales and likert response formats and their antidotes. journal of social sciences, 3, 106–116. carr, n. (2014). reassessing the assessment: exploring the factors that contribute to comprehensive financial risk evaluation. unpublished doctoral dissertation. kansas state university, manhattan, kansas. corter, j. e., & chen, y. j. (2006). do investment risk tolerance attitudes predict portfolio risk? journal of business and psychology, 20, 369–381. epstein, l. g., & schneider, m. (2010). ambiguity and asset markets. annual review of financial economics, 2, 315–346. fischhoff, b. (1975). hindsight � foresight: the effect of outcome knowledge on judgment under uncertainty. journal of experimental psychology: human perception and performance, 1, 288–299. fitzsimons, g. j., & williams, p. (2000). asking questions can change choice behavior: does it do so automatically or effortfully? journal of experimental psychology: applied, 6, 195–206. frisch, d., & baron, j. (1988). ambiguity and rationality. journal of behavioural decision making, 1, 149–157. gilliam, j., chatterjee, s., & grable, j. (2010). measuring the perception of financial risk tolerance: a tale of two measures. journal of financial counseling and planning, 21, 30–43. goldstein, d. g., johnson, e. j., & sharpe, w. f. (2008). choosing outcomes versus choosing products: consumer-focused retirement investment advice. journal of consumer research, 35, 440–456. grable, j. e., & lytton, r. h. (2003). the development of a risk assessment instrument: a follow-up study. financial services review, 12, 257–274. guillemette, m. a., yao, r., & iii, r. n. j. (2015). an analysis of risk assessment questions based on lossaverse preferences. journal of financial counseling and planning, 26, 17–29. guillemette, m., finke, m., & gilliam, j. (2012). risk tolerance questions to best determine client portfolio. journal of financial planning, 25, 36–44. harinck, f., & de dreu, c. k. w. (2008). take a break! or not? the impact of mindsets during breaks on negotiation processes and outcomes. journal of experimental social psychology, 44, 397–404. hertwig, r., barron, g., weber, e. u., & erev, i. (2004). decisions from experience and the effect of rare events. psychological science, 15, 534–539. hertwig, r., & erev, i. (2009). the description-experience gap in risky choice. trends in cognitive sciences, 13, 517–523. holt, c., & laury, s. (2002). risk aversion and incentive effects. the american economic review, 92, 1644–1655. kahneman, d. (2009). the myth of risk attitudes. journal of portfolio management, 36, 1. (available at http://search.ebscohost.com/login.aspx?direct�true&db�bth&an�45267051&site�ehost-live) kaufmann, c., weber, m., & haisley, e. (2013). the role of experience sampling and graphical displays on one’s investment risk appetite. management science, 59, 323–340. morrison, w., & oxoby, r. (2014). loss aversion in the laboratory. iza discussion papers, 8309. nobre, l. h. n., & grable, j. e. (2015). the role of risk profiles and risk tolerance in shaping client investment decisions. journal of financial service professionals, 69, 18–21. norman, g. (2010). likert scales, levels of measurement and the “laws” of statistics. advances in health sciences education, 15, 625–632. ordonez, l., & benson, l. i. (1997). decisions under time pressure: how time constraint affects risky decision making. organizational behavior and human decision processes, 71, 121–140. 364 k. bachmann et al. / financial services review 26 (2017) 339–365 parsons, h. (1974). what happened at hawthorne? science, 183, 922–932. pell, g. (2005). use and misuse of likert scales. medical education, 39, 970. rohrmann, b. (1992). the evaluation of risk communication effectiveness. acta psychologica, 81, 169–192. slovic, p., fischhoff, b., & lichtenstein, s. (1978). accident probabilities and seat belt usage: a psychological perspective. accident analysis and prevention, 10, 281–285. sundén, a. e., & surette, b. j. (1998). gender differences in the allocation of assets in retirement savings plans. american economic review, 88, 207–211. tversky, a., & kahneman, d. (1981). the framing of decisions and the psychology of choice. science, 211, 453–458. wärneryd, k.-e. (1996). risk attitudes and risky behavior. journal of economic psychology, 17, 749–770. xiao, j. j. (1996). effects of family income and life cycle stages on financial asset ownership. financial counseling and planning, 7, 21–30. yook, k., & everett, r. (2003). assessing risk tolerance: questioning the questionnaire method. journal of financial planning, 48–55. 365k. bachmann et al. / financial services review 26 (2017) 339–365 saving for retirement while having more nights with peaceful sleep: comparison of lifecycle and lifestyle strategies from expected utility perspective rosita p. changa, david huntera, qianqiu liua,*, helen saarb ashidler college of business, university of hawaii at manoa, 2404 maile way, honolulu, hi 96822, usa budvar-hazy school of business, dixie state university, 225 south 700 east, st. george, ut 84770, usa abstract we evaluate the fit of target-date funds (tdfs) as the main retirement savings instrument for the utility-maximizing investor who becomes more risk averse as she gets older. using bootstrapping simulations, we show that tdfs can provide higher expected utility than the alternative lifestyle strategies. with loss aversion incorporated in the model, we still find that the optimal lifecycle strategy over time leads to higher expected utility than the best lifestyle strategy. therefore, tdfs are preferable to the utility-maximizing investor. however, lifecycle strategies are not one-size-fits-all solution and investor’s risk tolerance has to be considered when selecting tdf funds. © 2014 academy of financial services. all rights reserved. jel classification: g11; g23; d14 keywords: target-date funds; life cycle investing; retirement saving 1. introduction since being endorsed by the department of labor, target-date funds (tdfs) have grown in both size and popularity.1 according to the investment company institute (2013), at the end of 2012 the total assets of tdfs were $481 billion, which is about 9.5% of the total assets of defined contribution plans. this represents a significant increase since 2007, when tdfs only had a 4.2% share of defined contribution plans. with this growing market share, it is * corresponding author. tel.: �1-808-956-8736; fax: �1-808-956-9887. e-mail address: qianqiu@hawaii.edu (q. liu) financial services review 23 (2014) 169–188 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. important to verify the central tenet of tdfs: investors should gradually diminish the importance of risky assets in their retirement portfolio returns over time. moreover, it is essential to establish if the current variety of tdf offerings satisfy the risk preferences of all individual investors. if not, individual investors and financial advisors need to be aware of this, so that they can evaluate how to alter the portfolios properly to match their risk profiles and maximize their utilities. tdfs are distinguishable by both the target retirement date and a glide path, which is defined as the specific strategy that decreases the portfolio’s risky asset allocation throughout its accumulation period.2 instead of tdfs, an investor could alternatively invest in lifestyle funds that hold equity allocations fixed over time. the superiority of decreasing equity allocations over holding them fixed has been questioned for a long time, even before the emergence of tdfs. many financial planners have advocated that the equity weighting in an individual investor’s portfolio should be inversely related to the investor’s age. the most common rule of thumb is that the percentage of wealth invested in stocks should be equal to 100 minus the investor’s age (ameriks and zeldes, 2004). following this rule an investor who is 45 years old should have no more than 55% of her wealth invested in stocks. academic literature refers to this idea as the lifecycle investment hypothesis (bakshi and chen, 1994). although financial planners seem to generally agree with decreasing risky allocations over time, the academic support for this argument has been less straightforward. we do know that stocks are not as risky over a long investment horizon (bali, demirtas, levy, and wolf, 2009; malkiel, 2011), and that stocks are essential to accumulate funds for different financial goals (samuelson, 1989; dynan, skinner, and zeldes, 2002). we also know that younger employees are endowed with future labor income that will help them recover from potential investment losses if they earn poor returns from risky investments, and likewise, that all investors have desire to attain high terminal values at the end of their investment horizon (bodie, merton, and samuelson, 1992; jagannathan and kocherlakota, 1996). our findings add one more reason to this list: as an investor’s risk aversion increases with age, the equity allocation in her portfolio needs to decrease to remain optimal for the utility-maximizing investor. an investor’s risk aversion is directly related to her utility, which represents the satisfaction that the investor gets from her consumption of goods or services. unlike returns on investments or the terminal wealth, utility takes into account the cost of reaching monetary goals; such as an investor’s sleepless nights that are attributable to taking too much risk. keeping too much wealth in risky assets when the investor’s risk tolerance is low might lead to lower expected utility despite the potential large dollar amount that could be earned by doing so. experts in different fields, such as psychology, economics, finance, and management, among others, have conducted research on demographics of individual risk aversion with data from many countries. they have determined an individual’s risk aversion with: analysis of actual holdings of risky assets (see, for instance, morin and suarez, 1983; bakshi and chen, 1994; pålsson, 1996; agnew, balduzzi, and sundén, 2003; and gibson, michayluk, and van de venter, 2013); experimental studies looking at the participant’s choice between risky alternatives, and studies that involve monitoring brain activity during decision-making 170 r.p. chang et al. / financial services review 23 (2014) 169–188 that involves risk (deakin, aitken, robbins, and bahakian, 2004; lee, leung, fox, and gao, 2007). in general, these studies support the lifecycle risk aversion hypothesis. specifically, they provide evidence that the arrow-pratt measure of relative risk aversion increases with age, and older investors hold lower proportion of risky assets in their total wealth.3 in this article we use expected utility over different accumulation periods to compare differences between lifecycle and lifestyle strategies whereas assuming that investors are utility maximizers.4 we run simulations with a range of portfolio weights on stocks and bonds for both lifecycle and lifestyle strategies, with a range of representative investors with different risk aversion characteristics. our investigation includes portfolio strategies that are more common among the funds offered by fund families and the best performing strategies. we find that for the investor whose risk aversion increases as she gets older, the lifecycle portfolio strategies lead to higher expected utility than strategies that keep equity allocation fixed over the whole accumulation period. prior studies in the field of behavioral finance have concluded that investors have different attitude towards downside risk and are more sensitive to the negative changes to their wealth (kahneman and tversky, 1979). we therefore incorporate the concept of loss aversion into the expected utility model. the results find further evidence that an investment strategy that decreases the weight of risky assets in the portfolio as the target retirement date nears, leads to higher expected utility than that of the best lifestyle strategy and is, therefore, preferable to the utility-maximizing investor. it is also important to remember that not all tdfs with the same target date are the same. balduzzi and reuter (2012) find increased heterogeneity among the tdfs offered in the market and conclude that this differentiation can lead to varying levels of performance and risk. idzorek (2009) stresses that it is important to look beyond the target date and evaluate the fund strategy in terms of investors’ risk preference and risk capacity. the morningstar industry survey (2009) reports that in 2008 the equity allocations for 2010 tdfs range from 26% to 72%. because of this wide difference in glide paths, we consider a variety of lifecycle strategies, including different glide path lengths, gentle and steep descent glide paths (also referred to as glide paths with kink), as well as aggressive and conservative glide paths in our analysis. as the popularity of tdfs increases, there is an emerging research on this topic in recent years. studies have been both critical of the lifecycle model as well as finding support for the suitability of tdf strategy. the main focus in most of these studies has been on the accumulated wealth by the target retirement year and the appropriateness of the asset allocation (see, for instance, bodie and treussard, 2007; schleef and eisinger, 2007; basu and drew, 2009; branch and qiu, 2011). spitzer and singh (2008, 2011) analyze the performance of lifecycle strategies both during accumulation period and during retirement, and find that tdf strategies underperform the lifestyle strategies and are not as safe as implied. liu, chang, de jong, and robinson (2011) evaluate the performance of the two lifecycle funds with gentle and steep descent glide paths and seven lifestyle funds throughout accumulation and withdrawal phases examining the total accumulated wealth at the retirement date. they conclude that lifecycle strategies are beneficial to the investor, especially during the withdrawal phase. pang and warshawsky (2011) find that lifecycle strategies are less risky than comparable balanced strategies and are proper for investors who wish to protect their accumulated funds. recent article by lipton and kish (2011) finds that lifecycle 171r.p. chang et al. / financial services review 23 (2014) 169–188 strategies fail to reach the benchmarks, and there is little uniformity in allocation and timing among the tdf families. pfau (2010, 2011) studies the utility from retirement savings by using standard constant relative risk-aversion utility function. he focuses on four different lifecycle strategies and compares them to 11 lifestyle strategies, and finds that investors with moderate risk aversion will reach higher expected utility from investing their retirement savings in tdf strategy than fixed asset allocation strategy. this article extends the existing literature by incorporating investor’s increasing risk aversion and loss aversion into the expected utility framework introduced in pfau (2010, 2011). we also include in our analysis a wider range of real-world lifecycle and lifestyle strategies and examine the lifecycle strategy with the highest expected utility in theory from bootstrapping simulations. our results suggest that for the representative investor there always exists a tdf strategy that generates higher lifetime utility than those from the conventional lifestyle strategies. we also find that tdfs are not one-size-fits-all solution. although many tdf strategies already exist in the market, they may not be optimal to utility-maximizing investors. investors who are less risk tolerant need a tdf strategy that starts at lower levels of equity in the portfolio. financial advisers and plan sponsors should take into account the risk tolerance profiles of individual investors to make sure the chosen tdf strategy is a good fit. 2. data and methodology 2.1. mean-variance utility model the objective of this article is to evaluate the accumulation period utility of investors with similar, but uniquely different perceptions about risk from different portfolio strategies. our first analysis assumes that the investor’s risk aversion increases as she gets older. in the second model we also assume that the investor becomes more risk averse as she experiences losses. based on these assumptions we look at how the change in risk aversion over the accumulation period and the portfolio composition affect the investor’s expected utility. more specifically, do lifecycle strategies yield higher expected utility compared with lifestyle strategies for the utility-maximizing investor who becomes more risk averse as she gets older? we use the expected utility in the accumulation period to compare the differences between lifecycle and lifestyle strategies for a representative investor who becomes less risk tolerant as she gets older. at any month t, we express the utility of the representative investor using the mean-variance model adopted by friend and blume (1975): ut � e�rp,t� � 1 2 at�p,t 2 (1) where rp,t,�p,t are the return and variance of the portfolio that the investor holds, and at is arrow and pratt’s measure of relative risk aversion at month t.5 this mean-variance model can be motivated by assuming quadratic utility for arbitrary distributions, or assuming that the return of the risky portfolio is normally distributed for arbitrary preferences (huang and 172 r.p. chang et al. / financial services review 23 (2014) 169–188 litzenberger, 1988). we calculate monthly utilities for the representative investor with different scales of risk aversion. many articles have studied the value of the risk aversion coefficient using theoretical models and experimental frameworks, resulting in a wide range of results. the majority of the studies conclude that the value of a relative risk aversion coefficient is somewhere between 0 and 5, and 2 to 4 for a typical investor. for example, friend and blume (1975) estimate that a coefficient of relative risk aversion is about 2 assuming stock returns are the only stochastic component of wealth. grossman and shiller (1981) find the coefficient of relative risk aversion has to be at least 4 to explain the variability in stock prices. pålsson (1996) finds the range for the coefficient of relative risk aversion is between 2 and 4. in our analysis we assume that for the representative investor the risk aversion level starts from zero to 2 at the beginning of the accumulation period, and ends between 4 and 6 at the target retirement year.6 the expected utility for all these different risk aversion ranges is calculated for the finite number of portfolio strategies based on equity/bond mix. the expected utility for the whole accumulation period (e.g., 40 years from age 25 to age 65) is defined as the present value of all monthly expected utilities: e�u� � � t�1 t �t�e(rp,t) � 1 2 at�p,t 2 � (2) where � is the discount factor, t is the length of accumulation period in months, and t stands for a specific month. we use a discount factor of 0.99.7 in this expected utility model we assume the investor’s relative risk aversion increases over her lifetime. we assume that the change in risk aversion is linear, but for robustness we also examine cases where the investor’s risk aversion stays constant for a certain period (e.g., first 10 years) and starts increasing closer to the target retirement year. 2.2. mean-variance utility model with loss aversion behavioral finance studies have shown that investors are loss averse, meaning they are more sensitive to the negative changes to their wealth than gains (thaler, tversky, kahneman, and schwartz, 1997). in the case of loss aversion, utility function is steeper for losses than gains. in the second model we incorporate the concept of loss aversion into the original mean-variance utility model. we add to the expected utility function the loss aversion coefficient, �, which increases the investor’s risk aversion when the prior period’s portfolio return was negative. the value of the loss aversion coefficient, �, is determined based on the realized portfolio return in the previous month. following tversky and kahneman (1992), we assume this parameter to be equal to 2.25 when the portfolio return was negative in the previous month. with the loss aversion coefficient the accumulation period expected utility function is expressed as follows, e�u� � � t�1 t �t�e(rp,t) � 1 2 at�t�p,t 2 � (3) 173r.p. chang et al. / financial services review 23 (2014) 169–188 where �t � � 1 when rp,t�1 s � 0 2.25 when rp,t�1 s � 0 (4) rs p,t-1 stands for the portfolio return for the month t � 1. if during the previous month the portfolio return was negative, the investor’s risk aversion will increase 2.25 times next month. in the case of a positive portfolio return, the investor’s risk aversion does not deviate from the regular risk aversion that the investor normally has. 2.3. return data and simulation expected portfolio return is calculated based on the portfolio mix for each strategy at the respective accumulation period. the lifecycle portfolios follow different glide path strategies that are characterized by the beginning equity allocation, ending equity allocation, and the time point when the equity allocation starts to decrease in the portfolio (e.g., for the first 10 years the weight of equity is kept at the maximum creating a kink in the glide path). based on these parameters, the portfolio allocations of lifecycle strategies change each year by decreasing the equity in equal increments but stay the same within the year.8 lifestyle strategies keep the equity allocation constant over the entire accumulation period. the monthly mean return of the portfolio for each strategy is calculated as follows: e�rp,t� � wbond,t � r� bond � weq,t � r� eq (5) where wbond,t and weq,t are the weights on bond and equity in the portfolio at month t, respectively. r� bond is the mean return of the bond and r� eq is the mean return of diversified equity portfolio. the portfolio variance for each strategy is calculated for each period as follows: �t 2�rp,t� � wbond,t 2 �� bond 2 � weq,t 2 �� eq 2 � 2wbond,tweq,t�� bond�� eq bond,eq (6) where is the correlation between the returns of the equity portfolio and the returns of the fixed income portfolio. in the simulations we use a diversified equity portfolio for equity allocation with 45% invested in the s&p 500 index, 30% invested in the russell 2000 index, and 25% invested in the msci eafe index, following liu, chang, de jong, and robinson (2009). for the fixed income holding we use a 10-year u.s. treasury bond. the monthly return data for equity indices and treasury bonds is retrieved from crsp and datastream. the sample period is from january 1970 to december 2010. table 1 shows the descriptive statistics for the three equity indices: the s&p 500 index, the russell 2000 index, and the msci eafe index, the 10 year u.s. treasury bonds, and the diversified equity portfolio. 2.4. tdfs and their glide paths the glide path of the tdf is characterized by two attributes: length of the glide path and change in the asset allocation over the fund lifetime. as balduzzi and reuter (2012) note, the fund families try to differentiate themselves from other tdfs there are many different glide 174 r.p. chang et al. / financial services review 23 (2014) 169–188 paths and asset allocation strategies in the market. table 2 summarizes the characteristics of glide paths offered by fund families collected from the prospectus of individual tdfs as of september 2012. as of september 2012, there were 230 tdfs with 45 fund families sponsoring these funds. most tdfs are founded 40 years before the target year with five year increments. the most common beginning equity level among the tdfs offered currently in the market is 90% equity. the highest level of beginning equity in the glide path is 100%. on average the tdfs reach the level of 40% to 45% equity by the target year. the minimum level of equity at the target year is 20%. the most common combination is the glide path from 90% equity allocation 40 years before the target year and 50% equity at target year. five fund families follow this glide path strategy with their tdfs. the next most common combinations are 90% to 40%, 90% to 30%, and 90% to 20% equity. among the more conservative strategies table 1 descriptive statistics for monthly returns for assets used in simulated strategies for the period from january 1970 to december 2010 10-year t-bond s&p 500 russell 2000 msci eafe diversified equity portfolio mean 0.00689 0.00896 0.00908 0.00755 0.00864 median 0.00595 0.01202 0.01308 0.00978 0.01187 standard deviation 0.02357 0.04555 0.06392 0.05018 0.04417 sample variance 0.00056 0.00207 0.00409 0.00252 0.00195 kurtosis 1.17384 1.97282 4.29705 1.08029 3.55604 skewness 0.35503 �0.48217 �0.03284 �0.34430 �0.60492 minimum �0.06682 �0.21580 �0.30615 �0.20239 �0.22394 maximum 0.09999 0.16811 0.39515 0.17874 0.21885 table 2 summary statistics for the glide paths of tdfs offered by the mutual fund families percentage of equity year of the kink before target year length of the glide path in years 40 years before target year at target year before target year after target year mean 91 42 30 43 19 median 90 45 30 40 20 mode 90 45 25 40 10 min 80 20 10 40 0 max 100 60 35 50 40 market leaders fidelity 90 20 n/a 45 0 vanguard 90 50 25 50 5 t. rowe price 90 45 25 45 40 percentage of tdfs with kink in the glide path 57% continue past target year 66% glide path longer than 40 years before target year 41% number of fund families 45 number of funds 230 175r.p. chang et al. / financial services review 23 (2014) 169–188 the common combination is from 80% to 40%, 80% to 30%, and 80% to 20% equity. among the tdf families, 57% have a kink in their glide path, meaning that they keep the equity at the maximum level for the first five to 30 years. most commonly the glide path is flat for the first 10 years and starts decreasing after 10 years. tdfs that continue the glide path after the target year are called “through” tdfs. on average the glide path continues 19 years past the target year and 66% of tdfs are “through” tdfs. given there is a wide range of glide path strategies used by actual tdfs in the marketplace, we calculate the expected utility over the accumulation period for different glide paths. each glide path can be characterized by its beginning equity level, ending equity level, and the steepness of the glide path or the presence of the kink in the glide path. we analyze glide paths that begin with equity allocation from 100% to 50% and end with the equity allocation from 70% to 0%. in addition we assume the glide path can be flat in the beginning of the accumulation period for 10, 20, or even 30 years, creating a kink into the glide path. given the different combinations between the beginning and ending equity and the kink in the glide path, we analyze well over 900 different portfolio strategies. we simulate 1,000 bootstrap samples for each portfolio strategy. in this article, we present the results for the more common strategies used by the fund families as well as the best performing strategy. for the base case we examine the representative investor who starts saving for retirement in the beginning of her professional career and has at least 40 years to accumulate funds to support her retirement years. in our results we present the expected utility for the following seven common lifecycle strategies: 90% to 50%, 90% to 40%, 90% to 30%, 90% to 20%, 80% to 40%, 80% to 30%, and 80% to 20%. for comparison we also present results from the best performing lifecycle and lifestyle strategies. 3. empirical results 3.1. expected utility for the representative investor with different accumulation periods we first consider the case of a young investor who has at least 40 years to save up for her retirement and calculate the total accumulation period expected utility following eq. (1) for a wide range of different portfolio strategies; both lifestyle and lifecycle. table 3 summarizes the results for the common strategies offered by the fund families as specified in section 2.4. the results are generated from 1,000 simulations in each scenario, with equity allocation 5% increments ranging from 0% to 100% (e.g., 20 times). table 3 reports four levels of beginning-ending risk aversion: {0–4; 0–6; 2–4; 2–6} in the first column. we report the expected utility for the best lifestyle strategy and its corresponding glide path in the second to third columns. for ease of comparison, we use the expected utility for the best lifestyle strategy as a benchmark, and report the expected utilities for the seven lifecycle strategies as a percentage difference relative to the benchmark. in addition, the last three columns in table 3 compare the best lifecycle strategy with the best lifestyle strategy by reporting the percentage difference, the glide path, and the t-statistics. panel a reports an investor with 40 years of investment horizon followed by panels b, c, and d with 30-years, 20-years, 10-years of investment horizons respectively. from table 3, we 176 r.p. chang et al. / financial services review 23 (2014) 169–188 can see that as risk aversion levels increase, the total expected utilities decrease. even though, theoretically, one can always find a lifecycle strategy which has higher expected utility than lifestyle strategy, but for investors with high levels of risk aversion, the current common market lifecycle strategies do not consistently generate higher total expected utilities than lifestyle strategies. in our base case the representative investor is a young person who has just entered the work force and is risk neutral (a�0) when she starts saving up for retirement, and becomes less risk tolerant as she gets older, with risk aversion level increasing in equal increments every year and reaching the level of 4 at target retirement year. on average this implies risk aversion of 2. even though our investor is risk neutral when she starts investing, we find that 100% portfolio does not yield him the highest expected utility among the lifestyle strategies and she should invest in the 55% equity portfolio instead. panel a of table 3 shows that all table 3 total accumulation period expected utility for the investor who becomes more risk averse as she gets older risk aversion best lifestyle strategy lifecycle strategies best theoretical lifecycle strategy e(u) equity 90 to 50 90 to 40 90 to 30 90 to 20 80 to 40 80 to 30 80 to 20 glide path t-stat panel a: 40 year investment horizon 0 to 4 3.284 55% 2.7% 3.5% 3.8% 3.7% 2.8% 3.0% 2.7% 4.5% 100 to 25 57.45** 0 to 6 3.075 40% �1.0% 1.4% 3.1% 4.0% 1.7% 3.1% 3.7% 4.4% 100 to 10 28.66** 2 to 4 3.051 40% �6.8% �4.9% �3.6% �2.8% �2.9% �1.8% �1.4% 0.3% 50 to 30 5.25** 2 to 6 2.903 35% �13.3% �9.7% �6.9% �5.0% �6.6% �4.2% �2.7% 0.6% 45 to 20 3.89** panel b: 30 year investment horizon 80% 80% 75% 70% 70% 70% 65% 0 to 4 2.575 55% 2.2% 2.9% 2.6% 1.7% 2.1% 2.1% 1.1% 4.5% 100 to 25 84.55** 0 to 6 2.411 40% �0.6% 1.6% 2.9% 3.2% 1.6% 2.8% 2.9% 4.3% 100 to 10 29.72** 2 to 4 2.397 40% �4.5% �2.9% �1.1% �0.4% �1.4% �0.6% �0.1% 0.3% 50 to 30 3.54** 2 to 6 2.279 35% �9.9% �6.6% �3.0% �0.9% �4.2% �2.1% �0.3% 0.6% 45 to 20 5.31** panel c: 20 year investment horizon 70% 65% 60% 55% 60% 60% 50% 0 to 4 1.796 55% 1.6% 1.7% 1.1% �0.2% 1.2% 1.1% �0.9% 4.6% 100 to 25 31.46** 0 to 6 1.682 40% �0.5% 1.5% 2.3% 2.0% 1.3% 2.3% 1.4% 4.4% 100 to 10 17.48** 2 to 4 1.675 40% �2.8% �0.8% 0.1% 0.0% �0.4% 0.1% 0.0% 0.3% 50 to 30 6.94** 2 to 6 1.591 35% �7.1% �3.1% �0.6% 0.5% �2.3% �0.6% 0.6% 0.6% 50 to 20 8.98** panel d: 10 year investment horizon 60% 55% 45% 40% 50% 50% 35% 0 to 4 0.941 55% 0.9% 0.7% 0.2% �2.4% 0.2% 0.0% �3.3% 5.0% 100 to 25 35.41** 0 to 6 0.881 40% �0.5% 1.1% 1.1% 0.2% 0.8% 1.7% �0.6% 4.9% 100 to 10 49.28** 2 to 4 0.879 40% �1.6% �0.1% 0.3% �0.6% 0.0% 0.3% �1.0% 0.3% 50 to 30 5.10** 2 to 6 0.835 30% �5.1% �1.7% 0.4% 0.5% �1.2% 0.2% 0.1% 0.6% 45 to 25 8.42** notes. the table reports the best lifestyle strategy as a benchmark for the expected utility (e(u)) of the lifecycle strategies. the results for the lifecycle strategies are reported as a percentage difference from that benchmark. the last three columns compare the best lifecycle strategy with the best lifestyle strategy by reporting the percentage difference, the original 40 year glide path (beginning and ending level of equity), and the t-statistics. **denotes significance at the 1% level. results for each accumulation period are reported in separate panels. first row of panels b through d give the approximate level of equity for each lifecycle strategy at the beginning of accumulation period (e.g., 30 years before target date). 177r.p. chang et al. / financial services review 23 (2014) 169–188 the more common lifecycle strategies included in our investigation yield higher expected utility during accumulation period than the best lifestyle strategy. the highest utility would be reached with a theoretical lifecycle portfolio that starts out at 100% but then smoothly decreases the equity allocation to 25% by the target date. such a strategy would yield to our representative investor 4.5% higher utility and it is significantly higher than the 55% lifestyle strategy at the 1% level. among the more common tdf strategies her best choice would be to start investing to the tdf with the glide path from 90% to 30% equity, which generates a 3.8% higher utility than that of the best lifestyle strategy. if the young investor becomes more risk averse as she gets older, reaching the risk aversion of 6 at the target date, the best lifestyle strategy is 40% equity portfolio; more conservative than above. for this investor all except the 90% to 50% lifecycle strategy outperform the best fixed equity portfolio. among the lifecycle strategies the highest expected utility is reached with the theoretical glide path from 100% equity to 10%, with utility 4.4% higher and is also significantly higher than that of the best lifestyle strategy at the 1% level. our results show that for the investor with higher risk aversion the appeal of more common lifecycle strategies starts to diminish. in fact none of the more common lifecycle strategies reported in table 3 manage to outperform the best lifestyle strategy when the investor’s beginning risk aversion is as large as 2. however, it is still possible to construct a lifecycle strategy that generates higher utility. for example, for the investor with risk aversion from 2 to 4, the theoretical lifecycle strategy with a conservative glide path from 50% equity to 30% equity would outperform the best lifestyle strategy with 40% equity fixed. the percentage difference in expected utility is though only 0.3%, but is still significant at the 1% level.9 the higher the investor’s beginning risk aversion the lower should be the equity allocation in the beginning of the accumulation period. high risk aversion at the target retirement year shifts the preference towards a more conservative glide path with lower beginning and ending equity level. because we run our analysis for a wide range of strategies we find that even with higher levels of risk aversion, there exist lifecycle strategies that outperform lifestyle strategies although none of the currently offered tdfs have sufficiently conservative glide path. to the investor who has above average risk aversion when she is young, we recommend a more conservative tdf that has a target year shorter than the investor’s planned retirement. the tdf with the shorter target date has already decreased the equity allocation to the desired lower level. by picking the tdf with the target date that does not match her desired target retirement year, the investor may find a fund with the glide path that matches her risk aversion and will maximize her expected utility. one of the main selling points of tdfs is that the target year stated in the fund name is an easy way for the investor to find the suitable fund. in addition quite often it is the plan sponsor who picks the fund for the investor as the investor herself has failed to make her choice. therefore, the suggestion to deviate from the tdf with the matching target year may not be very helpful. therefore, we recommend that fund families should include in their menu tdfs with more conservative glide paths that may be more suitable for more risk averse investors. so far, we have assumed that a young person enters the work force, for example, after 178 r.p. chang et al. / financial services review 23 (2014) 169–188 finishing her undergraduate education at around age 25 and immediately starts making contributions to the retirement savings plan. unfortunately, this is quite often not the case as young people do not take advantage of time as their ally in investing for retirement. we analyze whether tdfs would also be a good choice for the person who joins the defined contribution plan later in life, for example 30, 20, or even only 10 years before the desired target retirement date in the remaining panels in table 3. panel b of table 3 assumes that our young investor postpones saving for retirement by 10 years. we assume that our investor picks a tdf that has a target year similar to her planned year of retirement. in such a case the tdf that our investor starts making contributions to has already decreased the equity allocation for the first 10 years, assuming the smooth glide path. for example for the investor who picks a tdf with the matching target date and glide path from 90% equity to 40% equity, the level of equity in the portfolio is at about 80% when she starts making investments into this fund. in our analysis we assume our investor’s risk aversion is at its minimum when she starts making contributions into the retirement fund. if risk aversion of the risk-neutral investor increases from 0 to the level of 4 over her 30-year accumulation period, the lifestyle strategy that yields the highest utility should have 55% equity. from the more common lifecycle strategies listed in panel b in table 3, all of them outperform the best lifestyle strategy in the context of total expected utility. the highest utility is reached with the strategy with the original glide path from 100% equity to 25% equity at the target date. this tdf would have decreased the equity allocation to about 81% when the investor starts to make contributions to her retirement plan, and it yields a utility 4.5% higher than that from the best lifestyle strategy, which is significant at the 1% level. if the investor is more risk averse and her risk aversion increases from 2 at 30 years before the target year to 4 at the target year, her best choice would be a portfolio strategy with original glide path from 50% equity to 30% equity. the increase in expected utility is only 0.3% still significantly higher than the utility from the best lifestyle (40% equity) strategy. if the investor is very risk averse with risk aversion increasing from 2 to 6, we still find that a conservative lifecycle strategy with the glide path from 45% equity to 20% over 30 years yields investor higher utility. however, none of the tdfs in the market have decreased the equity to such a low level 30 years before the target date. therefore, the conservative investor needs a tdf with a far more conservative glide path than any of the fund families currently offer. if the investor postpones making contributions into a retirement account even longer and leaves only 20 years for contributing towards her retirement nest egg, the glide path of lifecycle strategies has brought the equity level in the portfolio even lower. the summary of expected utilities under the same assumptions is brought in panel c of table 3. if our investor is still neutral towards risk when she starts saving for retirement and her risk aversion reaches the level of 4 when she retires, she is still better off with the lifecycle portfolio strategy. among the more common lifecycle strategies the highest expected utility is reached with the original glide path strategy from 90% equity to 40% equity. this strategy would yield our investor a total accumulation period expected utility that is 1.7% higher than the utility from the best lifestyle strategy, which has 55% on equity. for the investor who has high risk tolerance 20 years before the target retirement, the tdfs that have decreased the level of equity to 55% or less may be too conservative. starting out with relatively high equity level 179r.p. chang et al. / financial services review 23 (2014) 169–188 at 20 years before target retirement might help the investor to catch up a little with the lost years of capital accumulation. however, it is important to keep in mind that in case of periods of bear markets investors have less time left to recover the losses. less risk tolerant investors may now find that for a suitable tdf at 20 years before the target year the level of equity has to decrease sufficiently. for example, less risk tolerant investor with risk aversion changing from 2 to 6 over her 20-year accumulation period will benefit from the conservative lifecycle strategy with the glide path from 50% equity to 20% equity. it provides significantly higher expected utility than that of the best performing lifestyle strategy (with 35% equity). this lifecycle strategy starts out at 70% equity 40 years from the target date and reaches around 50% level 20 years before the target date. when the investor postpones making contributions to her retirement account even more and the accumulation period is only 10 years, the glide paths of lifecycle strategies have brought the equity level down very close to the level at the target date. we find, though, from panel d of table 3 that for our representative investor who is risk-neutral when she starts investing towards retirement and reaches a risk aversion level of 4 at the retirement date, the lifecycle strategy with the original glide path from 100% equity to 25% equity yield the highest expected utility that is 5.0% higher than utility from 55% equity lifestyle strategy that yields utility of 0.941 and the difference is significant at the 1% level. it is interesting to note that even the more risk averse investors with risk aversion increasing from 2 to 6 by the target year may now reach higher utility with some of the common tdfs as they have decreased the level of equity sufficiently. for example the tdf with the original glide path from 90% equity to 20% has reached an approximate equity level of 40% 10 years before the target date. over the 10 year accumulation period this strategy would yield our investor 0.5% higher utility than the fixed 30% equity portfolio, which is the best lifestyle strategy. 3.2. expected utility of the loss averse investor with different accumulation periods prior research has concluded that investors have different attitude towards downside risk, thus, they are more sensitive to the negative changes to their wealth. table 4 summarizes the results for the analysis that uses eq. (5) that incorporates the representative investor’s loss aversion in our model. similar to table 3, in table 4 we assume an accumulation period of 40, 30, 20, or 10 years and the total expected utility is calculated as the sum of the present value of monthly utilities over the different accumulation periods. in case the representative investor’s portfolio strategy yields a loss during the given month, the investor is assumed to be more risk averse the next month. we use the loss aversion of 2.25 as suggested by tversky and kahneman (1992). panel a in table 4 shows that for the loss averse investor who is risk-neutral when she starts contributing towards the retirement portfolio and has a 40 year accumulation period, moderate lifecycle strategies yield higher expected lifetime utility than the lifestyle strategies. specifically for the risk-neutral investor who reaches a risk aversion of 4 by the target year, the best portfolio strategy is the lifecycle portfolio with the glide path from 90% equity to 20% equity. such a portfolio strategy will yield her a total accumulation period expected utility that is 1.6% higher than that from the 45% equity lifestyle portfolio, and the difference is statistically significant at the 1% level. the more common tdf strategies, except the 180 r.p. chang et al. / financial services review 23 (2014) 169–188 aggressive glide path from 90% to 50% all outperform the lifestyle strategies in terms of expected utility. for the investor who is more risk-averse when she is young and has a risk aversion of at least 2 when she starts investing towards retirement and reached risk aversion level of 4 by retirement the lifecycle strategy with equity decreasing from 50% to 25% over the accumulation period, yields the highest expected utility. the more common lifecycle strategies fail to outperform the 35% equity lifestyle strategy. the investor needs a more conservative glide path. as suggested before, this can be accomplished if our investor picks a tdf with the target date closer to today than her desired retirement year. as the main target investor of tdfs is the investor who fails to pick the fund herself, it would be more important that fund families should start to offer tdfs with more conservative glide paths. we find that even when the loss averse investor postpones making contributions into the retirement account, conservative lifecycle strategies yield higher expected utility but the table 4 total accumulation period expected utility for the loss averse investor risk aversion best lifestyle strategy lifecycle strategies best theoretical lifecycle strategy e(u) equity 90 to 50 90 to 40 90 to 30 90 to 20 80 to 40 80 to 30 80 to 20 glide path t-stat panel a: 40 year investment horizon 0 to 4 2.679 45% �1.1% 0.2% 1.2% 1.6% 0.5% 1.4% 1.6% 1.6% 90 to 20 7.20** 0 to 6 2.540 35% �6.0% �3.3% �1.1% 0.2% �2.2% �0.2% 0.9% 1.2% 60 to 20 10.77** 2 to 4 2.523 35% �11.6% �9.2% �7.4% �6.1% �6.6% �4.8% �3.9% 0.2% 50 to 25 1.01 2 to 6 2.395 30% �17.9% �13.9% �10.8% �8.5% �10.3% �7.3% �5.4% 0.1% 45 to 25 0.25 panel b: 30 year investment horizon 80% 80% 75% 70% 70% 70% 65% 0 to 4 1.703 40% �1.0% 0.3% 1.1% 1.3% 0.4% 1.2% 1.1% 1.3% 90 to 20 4.59** 0 to 6 1.611 35% �5.3% �2.4% �0.3% 0.8% �1.7% 0.3% 1.0% 1.0% 70 to 20 3.36** 2 to 4 1.610 35% �8.3% �5.6% �3.6% �2.3% �4.0% �2.1% �1.2% 0.2% 50 to 25 1.10 2 to 6 1.525 30% �13.7% �9.3% �5.8% �3.5% �7.0% �3.7% �1.9% 0.2% 45 to 25 0.37 panel c: 20 year investment horizon 70% 65% 60% 55% 60% 60% 50% 0 to 4 1.072 40% �0.9% 0.3% 1.0% 0.7% 0.4% 0.9% 0.4% 1.1% 80 to 35 2.31* 0 to 6 1.011 35% �4.4% �1.5% 0.6% 1.1% �1.0% 0.8% 1.0% 1.2% 75 to 35 2.81** 2 to 4 1.014 35% �5.6% �2.9% �0.9% �0.1% �2.2% �0.3% 0.1% 0.2% 50 to 30 2.33* 2 to 6 0.958 30% �10.1% �5.5% �1.9% �0.2% �4.3% �1.1% 0.2% 0.3% 45 to 25 0.37 panel d: 10 year investment horizon 60% 55% 45% 40% 50% 50% 35% 0 to 4 0.559 40% �0.8% 0.3% 0.4% �0.4% 0.3% 0.4% �0.6% 0.5% 80 to 35 2.68** 0 to 6 0.525 35% �3.8% �0.9% 0.7% 0.4% �0.8% 0.6% 0.2% 0.8% 75 to 30 3.09** 2 to 4 0.531 35% �4.1% �1.2% 0.3% 0.0% �1.0% 0.3% �0.1% 0.4% 50 to 30 3.80** 2 to 6 0.500 30% �8.2% �3.2% �0.2% 0.1% �2.8% 0.0% 0.2% 0.2% 45 to 25 0.45 notes. the table reports the best lifestyle strategy as a benchmark for the expected utility (e(u)) of the lifecycle strategies. the results for the lifecycle strategies are reported as a percentage difference from that benchmark. the last three columns compare the best lifecycle strategy with the best lifestyle strategy by reporting the percentage difference, the original 40 year glide path (beginning and ending level of equity), and the t-statistics. ** and *denote significance at the 1% and 5% levels, respectively. results for each accumulation period are reported in separate panels. first row of panels b through d give the approximate level of equity for each lifecycle strategy at the beginning of accumulation period (e.g., 30 years before target date). 181r.p. chang et al. / financial services review 23 (2014) 169–188 utility difference between the best lifestyle and best lifecycle strategies is smaller than in case of the investor who is not loss averse.10 if the investor postpones making contributions into a retirement account and leaves 30 years for contributing towards her retirement nest egg, the glide path of lifecycle strategies has brought the equity level in the portfolio lower. the summary of expected utilities under the same assumptions is brought in panel b of table 4. if our risk-neutral investor is still risk neutral towards risk when she starts saving for retirement and her risk aversion reaches the level of 4 when she retires, she reaches higher expected utility with most of lifecycle portfolio strategies. among the more common five lifecycle strategies the highest expected utility is reached with the original glide path strategy from 90% equity to 20% equity. this strategy would yield our investor a total accumulation period expected utility that is 1.3% higher than the utility from the 40% equity lifestyle portfolio (e(u)�1.703) and the difference is significant at the 1% level. less risk tolerant investor with risk aversion changing from 2 to 6 over her 30-year accumulation period will benefit from more conservative lifecycle strategy with equity changing from 45% to 25%, which is only 0.2% higher than that of the best lifestyle strategy (30% equity fixed). in case our investor’s accumulation period is only 20 years, the glide paths of lifecycle strategies have brought the equity level closer to the level at the target date. we find, though, from panel c of table 4, that for our representative investor who is risk neutral when she starts investing towards retirement and reaches a risk aversion level of 4 at the retirement date, the lifecycle strategy with the original glide path from 80% equity to 35% equity yields the highest expected utility. for the investor whose risk aversion is 2 when she starts investing and reaches 6 by the target year, a more conservative strategy with original glide path from 45% equity to 25% equity provides the highest utility. for the loss averse investor with the 10 year accumulation period, the glide paths of many common tdfs have reached low enough levels of equity that a conservative investor may reach higher utility with lifecycle strategies compared with best lifestyle strategy. however, the percentage difference in expected utilities is very small. in summary, our results show that for the loss averse investor the portfolio strategies have to be more conservative to yield higher utility. both the best lifestyle and best lifecycle strategies need to have a lower level of equity. we also find that the difference between the utility from reported lifestyle and lifecycle strategies has become smaller. for the more risk tolerant investor the utility from the common lifecycle strategies is less than 1.6% higher than the utility from the best lifestyle strategy and the most aggressive glide path (90% to 50%) does not outperform the best lifestyle strategy. for the more risk averse and loss averse investor the utility from the common lifecycle strategies is even less beneficial compared with the best lifestyle strategy. such investor needs more conservative tdfs than currently offered on the market. 3.3. expected utility from a kinked tdf glide path or a kinked risk aversion so far we have examined so-called smooth glide path strategies in comparison to constant equity strategies. poterba and samwick (1997) study the age and cohort effects on investor portfolio allocation and find that households start decreasing equity in their overall portfolio after age 43. in addition, as shown in table 2, among the tdfs offered in the market, 57% 182 r.p. chang et al. / financial services review 23 (2014) 169–188 have a kink in their glide paths. many tdfs keep the equity allocation at maximum in the beginning of the accumulation period for five to 15 years thus creating a kink into the glide path. most tdfs in the market do not keep equity at the maximum level for that long, but have a kink in the glide path at 35 to 30 years before the target year. to investigate the impact of a kink, we study the expected utility for glide path strategies with the kink at 30 years before the target year and report the results in table 5.11 we examine the lifecycle strategies with the kink at 30 years before the target year in comparison to the best lifestyle strategy as in table 3. in panel a of table 5 we still assume that the investor’s risk aversion changes linearly and starts increasing right after she starts making contributions to her retirement account. this investor, however, has a kinked glide path 30 years before the target year. when the investor’s beginning risk aversion is neutral she can easily select a tdf among the most common lifecycle strategies that outperform the best lifestyle strategy. the higher her ending risk aversion, the more conservative should be her strategy, keeping both the beginning and ending equity allocation at lower levels. for the investor whose risk aversion in the beginning of the accumulation period is 2, although the table 5 expected utility with kinked glide path and/or risk aversion risk aversion best lifestyle strategy lifecycle strategies best theoretical lifecycle strategy t-stat e(u) equity 90 to 50 90 to 40 90 to 30 90 to 20 80 to 40 80 to 30 80 to 20 glide path panel a: glide path kinked at 30 years before target year, risk aversion linear 0 to 4 3.284 55% 2.4% 3.3% 3.9% 4.1% 3.0% 3.4% 3.4% 4.5% 100 to 15 37.59** 0 to 6 3.075 40% �8.9% �0.2% 1.6% 2.7% 0.7% 2.2% 3.1% 3.4% 80 to 10 26.12** 2 to 4 3.051 40% �9.2% �7.6% �6.5% �5.8% �4.6% �3.7% �3.2% 0.3% 45 to 30 28.20** 2 to 6 2.903 35% �17.2% �14.1% �11.7% �10.0% �9.5% �7.4% �6.1% 0.6% 40 to 20 29.76** panel b: glide path linear, risk aversion kinked at 30 years before target retirement 0 to 4 3.445 75% 4.0% 4.2% 4.1% 3.6% 2.8% 2.6% 1.9% 5.6% 100 to 35 125.9** 0 to 6 3.222 50% 4.3% 5.7% 6.6% 6.8% 4.7% 5.3% 5.4% 8.1% 100 to 20 101.9** 2 to 4 3.092 40% �4.6% �3.0% �2.0% �1.5% �1.5% �0.7% �0.5% 0.5% 55 to 30 17.07** 2 to 6 2.974 35% �8.3% �5.3% �3.2% �1.8% �3.3% �1.5% �0.4% 1.0% 55 to 20 30.34** panel c: both glide path and risk aversion kinked at 30 years before target retirement 0 to 4 3.445 75% 4.7% 5.3% 5.5% 5.5% 3.7% 3.9% 3.7% 7.1% 100 to 25 169.8** 0 to 6 3.222 50% 4.4% 6.1% 7.3% 8.0% 5.1% 6.1% 6.6% 9.4% 100 to 10 78.17** 2 to 4 3.092 40% �6.5% �5.1% �4.1% �3.6% �2.8% �2.0% �1.7% 0.5% 55 to 30 8.68** 2 to 6 2.974 35% �11.0% �8.3% �6.3% �3.4% �5.2% �3.5% �2.4% 0.9% 50 to 20 17.18** notes. the table reports the best lifestyle strategy as a benchmark for the expected utility (e(u)) of the lifecycle strategies. the results for the lifecycle strategies are reported as a percentage difference from that benchmark. the last three columns compare the best lifecycle strategy with the best lifestyle strategy by reporting the percentage difference, the original 40 year glide path (beginning and ending level of equity), and the t-statistics. **denotes significance at the 1% level. panel a of the table reports the results for different lifecycle strategies that keep the level of equity at the maximum for the first 10 years of the accumulation period. investor’s risk aversion is assumed to increase smoothly over the accumulation period. panel b reports the results for lifecycle strategies with smooth (linear) glide path, but assumes that the investor’s risk aversion is constant for the first 10 years and starts increasing only 30 years before target date. panel c reports the results for the lifecycle strategies with the kink in the glide path at 30 years before the target year and also assumes that the investor’s risk aversion starts increasing only at 30 years before the target retirement. the full accumulation period is assumed to be 40 years. 183r.p. chang et al. / financial services review 23 (2014) 169–188 more common lifecycle strategies do not yield higher expected utility in comparison with the best lifestyle strategy, a very conservative lifecycle strategy is still able to outperform this constant equity strategy. for example, to the investor with risk aversion increasing from 2 to 6, the best strategy is to invest in the 40% to 20% glide path fund, which has a utility significantly higher than that of the best lifestyle strategy at the 1% level. we next examine the expected utility for the investor whose risk aversion starts increasing many years after she enters the work force and begins making contributions to the chosen retirement plan; a scenario reported by poterba and samwick (1997). in the beginning of the accumulation period the accumulated capital is still small and worries about potential losses are low. certain changes in life can change the investor’s risk tolerance; like for example starting a family. agnew, balduzzi, and sundén (2003) find that investor’s risk aversion is low even up to age 45 to 54 and starts increasing after that. panel b of table 5 shows the expected utilities for the investor whose risk aversion starts increasing 30 years before the target retirement year. the strategies listed in this panel have a smooth glide path or in other words start decreasing the equity allocation in the portfolio right away. as our representative investor’s risk aversion starts increasing only later in life, her average risk aversion over the accumulation period is relatively lower. therefore, in general this investor reaches higher utility with more aggressive strategies compared with the investor whose risk aversion starts increasing right away. more specifically our investor who is risk neutral when she is young and whose risk aversion starts increasing 30 years before retirement reaching the level of 4 by the target date, would reach the highest total expected utility when investing in the tdf that has 100% equity in the beginning and decreases it to the level of 35% equity by the target date. the more risk averse investor whose risk aversion changes from 2 to 6, would benefit from choosing a lifecycle strategy with the glide path from 55% equity to 20% equity. in both cases, the improvements are significant at the 1% level. panel b shows that if investor’s risk aversion is constant for a period and starts increasing later in life, the lifecycle strategies still provide higher expected utility than lifestyle strategies. in panel c of table 5, we present the results for the case where the glide path has a kink 30 years before target date and investor’s risk aversion starts increasing 30 years before the target retirement. the more risk tolerant investor (risk aversion increasing from 0 to 4 and from 0 to 6) will reach higher expected utility from investing in the common lifecycle strategies than from lifestyle strategies. given the flexibility in the kinked glide path, the percentage increase in the expected utilities is higher among the lifecycle strategies. for example the investor with risk aversion from 0 to 4 (with the kink in 30 years before the target year) will reach the highest utility with the lifecycle strategy that has a steep glide path that holds equity at 100% for the first 10 years and decreases it to 25% over the remaining 30 years. such strategy will yield him 7.1% higher expected utility (significant at the 1% level) than the best lifestyle strategy with 75% equity fixed over the 40 years of investment.12 in summary, when the investor has a kinked glide path and/or kinked risk aversion before the target retirement year, our results show that the best lifecycle strategy still outperforms all lifestyle strategies which have the fixed equity allocation. as in the previous analyses, for the more risk averse investor the utility from the more common lifecycle strategies is 184 r.p. chang et al. / financial services review 23 (2014) 169–188 typically less than the best lifestyle strategy, thus needs more conservative tdfs than currently offered on the market. 4. conclusions prior literature has examined several reasons why investors should follow the conventional wisdom of switching their investment portfolio gradually into safer assets when they get older. this article adds investors’ increasing risk aversion to this list and shows that for the investors who become less risk tolerant over their lifetime, the lifecycle strategy used by tdfs is a choice that would yield higher expected utility than a constant allocation strategy. for investors who start saving for retirement early and have 40 years to accumulate wealth to support their golden years, the best choice for the retirement plan depends on their risk aversion. to the investors who are less risk averse when young and become moderately risk averse by the time they plan to finish working, the best fit would be a plan that is relatively aggressive and starts the portfolio glide path at a high level of equity (close to 100%), decreasing it over the accumulation period to a low level (about 25%) at the target date. investors, who are more risk averse and become even less risk tolerant throughout their life, may not find a beneficial lifecycle strategy among the tdfs offered currently by the fund families and would need a lifecycle strategy with a more conservative beginning and ending level of equity. procrastinators who leave saving for retirement until later in life are also better off with investment strategies that have a glide path with decreasing stock allocation over time. in most cases they would still be better off picking the tdf that matches their planned retirement year. if the investor’s risk tolerance starts decreasing later in life, for example 30 years before their planned retirement, lifecycle strategies will still provide a higher expected utility over the accumulation period than the constant allocation strategy. compared to the lifestyle strategies, the investor also enjoys higher utility from a kinked tdf strategy that keeps the equity allocation at the maximum in the beginning years. after incorporating the investor’s loss aversion into the conventional expected utility framework, our results show that conservative lifecycle investment strategies provide higher expected utility than strategies with constant equity allocation, though the tdfs currently offered by fund families are not conservative enough for the loss averse and extremely risk averse investor. we find it important to remember that tdfs are not a one-size-fits-all solution. investors should look at the tdfs strategy a bit closer than just the target date. financial advisors should determine individual investors’ risk aversion characteristics before suggesting the appropriate tdf. though it is hard to predict an investor’s future risk tolerance, her beginning risk aversion can give at least some indication for the starting point of the glide path of the portfolio strategy. as one of the qdia options, lifecycle funds are targeted at individuals who have failed to pick the fund for their retirement contributions. our analyses suggest that the plan sponsors who make the choice for them should become more familiar with the plan 185r.p. chang et al. / financial services review 23 (2014) 169–188 participants’ risk aversion. in particular, fund families should consider offering some tdfs with more conservative glide paths for those investors who are more risk averse. notes 1 with the enactment of the pension protection act (ppa) of 2006, the u.s. department of labor defines qualified default investment alternatives (qdias) for retirement plan participants who fail to choose an allocation for their retirement contributions. one of the four mechanisms described in the ppa says that the qualified default instrument should have a portfolio mix that considers an investor’s age or retirement date (employee benefits security administration, 2009). tdfs, also known as lifecycle funds, change asset allocation on a pre-stated schedule, depending upon an investor’s age relative to retirement, can therefore be used by plan sponsors as a default option for those participants who fail to choose an investment allocation for themselves. 2 in this article we use terms “lifecycle fund” and “target-date fund (tdf)” interchangeably. 3 when wealth is taken into account, some studies have shown that as investors get older and their wealth increases, their (absolute) risk aversion may not increase. however, siegel and hoban (1982) show that there is an increasing relative risk aversion with wealth if a broader based sample and a more comprehensive measurement of wealth are used. 4 in this article we focus on a representative investor who has failed to choose a fund for his contributions towards retirement, or who wants to make his contributions with minimal or no interference during the full investment period. this is an investor who is kept in mind in the ppa 2006 that defined qdias. this is also the investor who without this default choice would have made no or very small contributions toward his retirement welfare. therefore, it is important to remember that an investor who actively manages his retirement portfolio may find other strategies more suitable. 5 hereafter, all risk aversions included in this article mean relative risk aversion. 6 we assume that the relative risk aversion changes annually. within each month during a year the investor’s risk aversion is assumed to stay the same. 7 for robustness, we also test with � of 0.95 and 0.90. the results are available upon request. 8 as in liu, chang, de jong, and robinson (2011), we rebalance the portfolio annually in the beginning of each year. therefore, within the year the level of equity in the portfolio is kept constant. 9 in eq. (2), the lifetime expected utility is based on the mean and variance of the portfolio, which are the same in the 1,000 simulations. this leads to a small standard deviation of the utility difference series and very large t-statistic correspondingly. 10 unlike table 3, the utility difference between the best lifecycle and lifestyle strategies is not statistically significant at any conventional level for many highly risk averse investors. this is because of the much higher standard deviation of the utility differ186 r.p. chang et al. / financial services review 23 (2014) 169–188 ence series in the 1,000 simulations, generated from a path-dependent utility model specified in eq. (3). 11 we also calculate the expected utility for glide path strategies with the kink at 20 and 10 years before the target year and our results are qualitatively similar as in the 30 year case. these results are unreported to save place, and are available upon request. 12 the results are similar when we assume the investor’s risk aversion starts to increase 20 years before the target retirement. they are available upon request. references agnew, j., balduzzi, p., & sundén, a.(2003). portfolio choice and trading in a large 401(k) plan. american economic review, 93, 193–215. ameriks, j., & zeldes, s. p. (2004). how do household portfolio shares vary with age? working paper, columbia business school. bakshi, g. s., & chen, z. (1994). baby boom, population aging, and capital markets. journal of business, 67, 165–202. balduzzi, p., & reuter, j. (2012). heterogeneity in target-date funds and the pension protection act of 2006. working paper, national bureau of economic research. bali, t. g., demirtas, k. o., levy, h., & wolf, a. (2009). bonds versus stocks: investors’ age and risk taking. journal of monetary economics, 56, 817–830. basu, a. k., & drew, m. e. (2009). portfolio size effect in retirement accounts: what does it imply for lifecycle asset allocation funds? journal of portfolio management, 35, 61–72. bodie, z., merton, r. c., & samuelson, w. f.(1992). labor supply flexibility and portfolio choice in a life cycle model. journal of economic dynamics and control, 16, 427–449. bodie, z., & treussard, j.(2007). making investment choices as simple as possible, but not simpler. financial analysts journal, 63, 42–47. branch, b., & qiu, l. (2011). exploring the pros and cons of target date funds. financial services review, 20, 95–111. deakin, j., aitken, m., robbins, t., & bahakian, b. j. (2004). risk taking during decision-making in normal volunteers changes with age. journal of international neuropsychological society, 10, 590–598. dynan, k. e., skinner, j., & zeldes, s. p. (2002). the importance of bequests and lifecycle saving in capital accumulation: a new answer. american economic review, 92, 274–278. employee benefits security administration. (2009). factsheet: regulation relating to qualified default investment alternatives in participant-directed individual account plans. employee benefits security administration (ebsa) newsroom. friend, i., & blume, m. e. (1975). the demand for risky assets. american economic review, 65, 900–922. gibson, r., michayluk, d., & vane de venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. grossman, s. j., & shiller, r. j. (1982). the determinants of the variability of stock market prices. american economic review, 71, 222–227. huang, c., & litzenberger, r. h. (1988). foundations for financial economics. upper saddle river, nj: pearson education company. idzorek, t. (2009). target date solutions: is a target date enough? cfa institute conference proceedings quarterly, 26, 81–89. investment company institute. (2013). 2013 investment company institute fact book: review of trends and activities in the u.s. investment company industry. washington, dc: investment company institute. jagannathan, r., & kocherlakota, n. r. (1996). why should older people invest less in stocks than younger people? federal reserve bank of minneapolis quarterly review, 20, 11–23. 187r.p. chang et al. / financial services review 23 (2014) 169–188 kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–291. lee, t. m. c., leung, a. w. s., fox, p. t., & gao, j. (2007). age-related differences in neural activities during risk taking as revealed by functional mri. social cognitive and affective neuroscience, 3, 7–15. lipton, a. f., & kish, r. j.(2011). life cycle funds: lack of disclosure and lack of return. financial services review, 20, 75–94. liu, q., chang, r. p., de jong, j. c., & robinson, j. h.(2009). reality check: the implications of applying sustainable withdrawal rate analysis to real world portfolios. financial services review, 18, 123–139. liu, q., chang, r. p., de jong, j. c, & robinson, j. h. (2011). are lifecycle finds getting a bum rap? a comprehensive comparison of lifecycle versus lifestyle retirement strategies from accumulation through withdrawal. journal of wealth management, 14, 68–84. malkiel, b. g. (2011). a random walk down wall street: time-tested strategy for successful investing. new york, ny: w.w. norton & company. morin, r. a., & suarez, a. f. (1983). risk aversion revisited. journal of finance, 38, 1201–1216. morningstar fund research. (2009). target-date series research paper: 2009 industry survey. pålsson, a. (1996). does the degree of relative risk aversion vary with household characteristics? journal of economic psychology, 17, 771–787. pang, g., & warshawsky, m. (2011). target-date and balanced funds: latest market offerings and risk-return analysis. financial services review, 20, 21–34. pfau, w. d. (2010). lifecycle funds and wealth accumulation for retirement: evidence for a more conservative asset allocation as retirement approaches. financial services review, 19, 59–74. pfau, w. d. (2011). the portfolio size effect and lifecycle asset allocation funds: a different perspective. journal of portfolio management, 37, 44–53. poterba, j. m., & samwick, a. a. (1997). household portfolio allocation over the life cycle. working paper, national bureau of economic research. samuelson, p. a. (1989). a case at last for age-phased reduction in equity. proceedings of the national academy of sciences of the united states of america, 86, 9048–9051. schleef, h. j., & eisinger, r. m. (2007). hitting or missing the retirement target: comparing contribution and asset allocation schemes of simulated portfolios. financial services review, 16, 229–243. siegel, f. w., & hoban, j. p. jr. (1982). relative risk aversion revisited. review of economics and statistics, 64, 481–487. spitzer, j. j., & singh s. (2008). shortfall risk of target-date funds during retirement. financial services review, 17, 143–153. spitzer, j. j., & singh s. (2011). assessing the effectiveness of lifecycle (target-date) funds during the accumulation phase. financial services review, 20, 327–341. thaler, r. h., tversky, a., kahneman, d., & schwartz, a. (1997). the effect of myopia and loss aversion on risk taking: an experimental test. quarterly journal of economics, 112, 647–661. tversky, a., & kahneman, d. (1992). advances in prospect theory: cumulative representation of uncertainty. journal of risk and uncertainty, 5, 297–323. 188 r.p. chang et al. / financial services review 23 (2014) 169–188 afs annual meeting october 15-16, 2015 orlando, fl call for proposals 2015 due april 13, 2015 submission information: research papers and abstracts covering all aspects of financial planning, individual financial management, and education are sought for inclusion in the program. papers in the areas of estate planning, insurance, tax accounting aspects of financial planning, investments, and retirement planning are encouraged. proposals for panel discussions and tutorials devoted to current issues in individual financial management or the practice of financial planning will also be considered. each submission will be reviewed anonymously by at least two members of the program committee, and authors will be notified of the decisions on or about june 1, 2015. papers already accepted for publication in a refereed journal should not be submitted. there is no submission fee. submissions must be prepared as a word document and submitted as an attachment within proposalspace, this year’s online submission format. new option: this year for the first time, the conference will include an afs poster session similar to last year’s joint conference where fta sponsored a poster session during the reception. in your submission to proposalspace, you will have the option to allow your submission to be considered for a poster. submission process: all submissions will be made online this year, for the first time. no direct submissions to the program chair will be accepted. complete instructions appear below and help is available if you have questions. please submit here: http://proposalspace.com/calls/d/463 best paper awards: at least two awards ($1,000 each) will be conferred: the finametrica behavioral finance award,, is designated for the best paper to demonstrate a specific trait or behavior and the relevance to financial advisors and how they might apply it with their clients; the planplus analytical excellence awardy , is designated for the best paper that outlines a new or improved analytical technique or understanding that will allow financial planners to generate better financial forecasts and strategies for their clients. manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $50 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees can be paid online or mailed to the editor when a manuscript is submitted electronically. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online or mail the application, along with a check for annual dues and submission fee ($125 total; $75 for a one-year membership and $50 submission fee) to the editor. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services northern illinois university department of finance dekalb, il 60115 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 academy of financial services officers president lance palmer university of georgia president-elect william chittenden texas state university executive vice president-program thomas coe quinnipiac university vice president-communications christine mcclatchey university of northern colorado vice president-finance diane docking northern illinois university thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations tom warschauer san diego state university vice president-mktg & public relations a. william gustafson texas tech university vice president-membership larry prather southeastern oklahoma state university vp local arrangements 2014 duncan williams william patterson university vp local arrangements 2015 benjamin cummings saint joseph’s university immediate past president frank laatsch univ. of southern mississippi editor, financial services review stuart michelson stetson university directors robert moreschi virginia military institute dale domian york university charles chaffin cfp board of standards rich fortin new mexico state university grady perdue university of houston clear lake jacob sybrowsky utah valley university past presidents frank laatsch, 2012-13 univ. of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994 -95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms can be downloaded from the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2014 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 23, no. 3, 2014 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jean louis heck saint joseph’s university financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2014 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. from the editor this issue contains issue 3 of volume 23 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “what determines risk tolerance?” is coauthored by michael guillemette at the university of missouri and david nanigian at the american college. in this article, the authors investigate what drives risk tolerance as it directly influences the portfolio allocation preference of clients. analyzing average monthly scores from a widely used risk tolerance questionnaire, they find that habit formation, loss aversion, and sentiment proxies account for -1.06%, 38.51% and 13.21% of the variation in average monthly risk tolerance, respectively. they also find that habit formation did not account for additional variation in average monthly risk tolerance when controlling for loss aversion and sentiment. the second article “structured certificates of deposit: introduction and valuation,” is coauthored by geng deng, tim dulaney, tim husson, and craig mccann, all with the securities litigation and consulting group, in this paper, the authors examine the properties and valuation of market-linked certificates of deposit (structured cds). they review the market for structured cds and provide valuations for several common product types. they find significant mispricing of several common types of structured cds across multiple issuers, which is similar in magnitude to the well-documented mispricing in the structured products market. in particular, they estimate that structured cds are typically worth approximately 93% of the value of a contemporaneously issued fixed-rate cd. these results suggest that un-sophisticated investors may not understand the value, risks, and subtleties of these seemingly conservative investments. the third article, “using the new portability election of deceased spouses: a pedagogical example” is coauthored by de’arno de’armond, darlene pulliam, and robin patterson, all at west texas a&m university. the authors investigate the tax relief, unemployment insurance reauthorization, and job creation act of 2010. the act contains a provision which will allow the unused portion of a decedent’s exclusion (taxable estate protected by the unified credit) to be used upon the subsequent death of the surviving spouse. the portability election is simple for situations where it appears the surviving spouse will not financial services review 23 (2014) v–vi 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. remarry, however, it becomes much more complicated if the surviving spouse should remarry. the authors provide an interesting explanation, along with examples and an excel application model developed for use in the academic environment. . the fourth article, “does active management work? evidence from equity sector funds’” by crystal y. lin at eastern illinois university, in this research, the author presents evidence that equity sector mutual funds, the nine fidelity select portfolios, have provided better after-expense returns against broader market etf, spy, and their peer sector etfs, the nine select sector spdr funds, over the sample period 1999 -2010. not only do they achieve higher nominal returns over the twelve years, some of the funds also generate higher risk-adjusted returns measured by sharpe ratio and alpha from various asset pricing models. none of the sector mutual funds generates a significant negative alpha for the sample period for all asset pricing models. the final article, “investor profiles: meaningful differences in women’s use of investment advice,” by kathryn simms at old dominion university. u.s. women face numerous financial challenges, they typically earn less than men do; they have greater probabilities of living in poverty; and they need substantial retirement funds, given their average longevity. the author investigates how women use investment advice to remedy these challenges. this article fills the gap in the literature by evaluating two profiles of female investors through cluster analysis and logistic regression conducted on a large, nationally representative database. she finds that predictors of female investors seeking investment advice vary considerably across profiles. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. thanks to those who make the journal possible, especially the referees and contributing authors. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 23 (2014) v–vi bond laddering and bond indexing: an empirical comparison c. sherman cheunga, peter miua,* adegroote school of business, mcmaster university, hamilton, ontario l8s 4m4, canada abstract bond laddering and bond indexing have been widely accepted approaches to bond investing among retail investors. however, bond laddering has virtually been ignored in both the academic literature and most of the popular investment textbooks. one thing both approaches have in common is that they are passive strategies with no attempt whatsoever to beat the market. there are many unresolved issues about the two seemingly similar approaches. first, which approach should an investor favor? is there any room for both to be used at the same time? second, if an investor decides to use a ladder, what is the appropriate term to maturity for the ladder? there is hardly any theoretical or empirical guidance as to which is a better approach to use and the right term of a ladder. the relative attractiveness of the above two approaches are empirically examined in this study. we identify conditions that favor one over the other. conditions under which both instruments should be held within an optimal portfolio are also identified. we also identify conditions in which a longer term ladder is more appropriate than a shorter term ladder. © 2017 academy of financial services. all rights reserved. jel classification: g10; g11 keywords: ladder bond portfolio; passive bond investments 1. introduction both laddering and indexing are popular approaches to bond investing. holding a bond portfolio that replicates an index can be entirely consistent with asset pricing theories in which an investor in equilibrium holds a market portfolio of stocks and bonds. bond * corresponding author. tel.: �1-905-525-9140 ext. 23981; fax: �1-905-521-8995. e-mail address: miupete@mcmaster.ca (p. miu). financial services review 26 (2017) 181–203 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. laddering is somewhat of an enigma to the academic profession. bond laddering consists of investing a roughly equal dollar amount in bonds with different maturities and holding the bonds until maturity. when the shortest term bond matures, the proceeds will be reinvested in the longest term bond to maintain the same maturity structure of the bond portfolio over time. the idea is to avoid locking into a low return bond for a long period of time. the ladder approach allows investors to stay fully invested and take advantage of rising yields when the opportunity arises. the approach is widely accepted by practitioners but virtually ignored in investment textbooks.1 except the few studies mentioned below, the academic literature contains virtually no serious discussions of the laddering approach to bond investing. for example, specialized fixed-income books by garbade (1996), kaufman, bierwag, and toevs (1983), and sundaresan (1997) and make no mention of the topic.2 the ladder approach is not considered to be important enough to be mentioned in some of the most popular investment textbooks, such as alexander, sharpe, and bailey (2001), bodie, kane, and marcus (2007), and jones (1996).3 one thing both the approaches of bond laddering and indexing have in common is that they are passive strategies with no attempt whatsoever to add value by betting on interest rate movements or the selectivity of individual bonds. while laddering may be a lower risk approach as maturities are shorter, its return at the same time may be lower. how an investor should choose between these two passive strategies is far from obvious. this study provides an empirical guide. judd, kubler, and schmedders (2011) build an intertemporal model that justifies the use of a ladder bond portfolio in an investor’s investment holding. their model shows the familiar separation theorem in which an investor will hold a risk-free asset and a common risky portfolio. the ladder bond portfolio acts like a risk-free asset delivering a steady stream of income in an intertemporal world. their model assumes no human capital, labor income, and inflation. it is entirely possible any labor income from the human capital will serve the same purpose, thus negating the need for a ladder bond portfolio. further, as pointed out by campbell and viceira (2001), inflation makes nominal bonds used in the ladder portfolio risky in real term. more important, canner, mankin, and weil (1997) have documented empirically that separation theorem judd, kubler, and schmedders rely on to justify the existence of a bond ladder portfolio has no empirical support in the real world. cheung, kwan, and sarkar (2010) justify the existence of bond ladders based on bound rationality. a rationale for the existence of ladders, however, does not help investors to choose between seemingly similar alternatives. the present study on bond indexing complements the growing literature on stock indexing. the performance of stock indexing has been investigated in many studies. prather, chu, mazumder, and topuz (2009) compare alternative passive ways to invest in the standard and poor’s (s&p) 500 index. while the standard and poor’s depository receipts (spdrs) have lower advertised annual expenses, investors in spdrs face bid-ask spreads and commissions. the s&p 500 index mutual funds involve no additional trading costs. the question for them is whether the undisclosed trading costs alter the choice between spdrs and the s&p 500 index mutual funds. chang and krueger (2010) examine whether enhanced index funds (eifs) live up to their name and enhance portfolio performance. they find that eifs have mostly lower returns, much higher risks, and lower risk-adjusted returns. lu, wang, and 182 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 zhang (2012) examine leveraged and inverse exchange trade funds (etfs) and note how, over time, these products do not accurately track the leveraged or inverse return of the benchmark they are designed to mimic. dilellio and stanley (2011) look at actively and passively managed investment strategies utilizing only etfs. they test whether etf-only strategies such as sector rotation strategy can typically outperform the s&p 500 index, and more appropriately a representative benchmark on an absolute and/or risk-adjusted basis and find that these strategies may allow investors to capture inefficiencies in equity markets. 2. optimal portfolio of a representative united states investor to examine the contribution of a bond index or a bond ladder in an investment portfolio, we consider a representative u.s. investor who is interested in holding an optimal portfolio consisting of u.s. equities, non-north american equities, and u.s. bonds. for this investor, we examine the effects of using either a bond index or a bond ladder on the risk-return tradeoff of the portfolio. specifically, we estimate the sharpe ratio of the optimal portfolio under a mean-variance optimization framework when either the bond index or the bond ladder is held with u.s. equities and non-north american equities. therefore, we compare: (1) the maximum sharpe ratio of the portfolio with the two equities and the bond index, and (2) the maximum sharpe ratio of the portfolio with the two equities and the ladder strategy. the bond strategy that offers the higher sharpe ratio is the superior approach to adopt. to ensure that any difference in the attractiveness between the bond ladder and the bond index is not because of sampling errors, we test if the extra diversification benefit offered by either approach is indeed statistically significant. if short selling is allowed, the maximum sharpe ratio can be derived by first finding out the weights w of the tangency portfolio.4 w � ��1 � z� � rf � 1� b � a � rf , (1) where z� is the vector of the mean returns of assets � is the variance-covariance matrix of asset returns 1 is the unit vector rf is the constant risk-free rate a � 1� � ��1 � 1 b � 1� � ��1 � z� the sharpe ratio of the tangency portfolio is then equal to: �p � z�p � rf �p � z�� � w � rf �w� � � � w�0.5 (2) where z�p and �p are, respectively, the mean and standard deviation of the return of the tangency portfolio. because most investors do not short sell an entire asset class even they 183c.s. cheung, p. miu / financial services review 26 (2017) 181–203 may short a stock occasionally, we consider the portfolio allocations where short selling is disallowed throughout this article. when short selling is disallowed, the weights of the tangency portfolio can be solved with the extra inequality constraint that the vector of weights is nonnegative. 3. data the test period covered in this study is from 1980 to 2015 inclusively. unlike other asset classes such as stock and bond indices where returns can be readily obtained, returns on ladder strategies must be constructed from individual bonds. let us take the five-year ladder as an example. the return on the ladder at time t can be calculated as the equally weighted average return of the one-year, two-year, three-year, four-year, and five-year bonds at time t. this ladder portfolio will be updated at time t � 1 by dropping the return of the one-year bond that is maturing while adding the five-year bond return observed at time t � 1. this process will be repeated over time. by adding five-year bonds repeatedly as we roll over the ladder strategy, the portfolio eventually consists of four past five-year bonds and a current five-year bond. thus, the return on the ladder portfolio is simply an equally weighted average of past and current five-year bond returns.5 likewise, the return on a seven-year ladder is the equally weighted average of past and current seven-year bond returns. to the extent that longer term bonds tend to offer a higher coupon, longer term ladders will offer a higher coupon returns. the bond yields used in calculating the time-series of ladder returns of various maturities are obtained from the federal reserve website. specifically, historical annual yield data on three-year bonds, five-year bonds, and seven-year bonds are obtained to construct a three-year ladder, a five-year ladder, and a seven-year ladder, respectively.6 annual return data are used as the maturing bonds in ladders tend to be reinvested annually. by using an equally weighted average of current and past five-year bonds as the return on the five-year ladder, we implicitly ignore any mark-to-the-market price changes before the maturity dates of the bonds. consider a five-year bond issued now at par with the yield-tomaturity equals to its coupon rate. this yield-to-maturity will be used in computing the return on the ladder for the current and also the next four periods. this approach ignores the price gain or loss of the bond in the next four periods as a result of any changes in the market interest rates.7 this approach is deemed to be valid if the investor holds the bond until maturity as in the case of a ladder and is adopted in our main analysis. we will consider the implications of mark-to-market price changes on the performance of the ladder strategy later in our analysis. the bond index used as an alternative to the ladder strategy is the barclays u.s. aggregate bond index, a leading benchmark for fixed-income investors.8,9 we obtain the annual index returns over our sample period of 1980–2015 from bloomberg. we use the annual total return series of the center for research in security prices’ (crsp’s) value-weighted index and the msci europe, australasia and far east (eafe) index to proxy for the united states and non-north american equity returns, respectively. 184 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 4. analytical framework we perform five types of analyses on bond laddering and bond indexing. first we examine the historical risk, return, and the sharpe ratio for each of the two alternatives to study their risk and return characteristics in isolation. this univariate analysis allows us to determine the attractiveness of each asset class. while bond laddering and bond indexing are similar in that they are both passive in style with no attempt to beat the market, they are not identical. the bond index contains bond assets with a higher credit risk and a longer duration and, therefore, it is important to quantify the resulting difference in the risk and return characteristics because of credit risk and duration.10 second, we compare the performance of ladder strategies of different term structures. for investors who pursue the ladder approach, what should be the ideal term structure? should the investor pursue a five-year ladder instead of a seven-year ladder? there is no theoretical guidance in the literature. we intend to provide some empirical guidance on this issue by examining the attractiveness of a three-year ladder, a five-year ladder, and a seven-year ladder versus a passive bond index. third, the attractiveness of each bond strategy is examined in a portfolio context. because most investors own both bonds and stocks, the modern portfolio approach demands that any bond strategy must be judged by its contribution to an overall diversified portfolio. fourth, it is quite possible that neither the ladder nor the bond index dominates each other uniformly for all investors with different risk preferences. the difference in their risk and return characteristics could be in such a way that they are both appropriate depending on the risk aversion of the individual investor. to examine this possibility, we investigate the ladder with different maturities versus the bond index in a portfolio context for investors with different risk aversion parameters. finally, we also perform statistical tests to ensure any difference in the attractiveness of the two passive fixed-income approaches is not because of sampling variations. 5. results table 1 provides the descriptive statistics for the key individual asset classes. here we focus on the five-year ladder without marking to market.11 later in the article we will table 1 summary statistics of annual returns on different assets (36 years from 1980 to 2015) crsp vw msci-eafe five-year ladder barclays bond index mean 0.1252 0.1106 0.0631 0.0814 standard deviation 0.1715 0.2219 0.0307 0.0694 sharpe ratioa 0.4645 0.2929 0.5704 0.5163 correlation crsp vw 1.000 0.683 0.155 0.169 msci-eafe 1.000 0.292 0.048 five-year ladder 1.000 0.575 barclays bond index 1.000 note: asharpe ratio calculated based on a risk-free rate of 4.56%, which is the average yield of the three-month t-bill over the sample period. 185c.s. cheung, p. miu / financial services review 26 (2017) 181–203 consider ladders of other maturities and ladders that are marked to market. among the four asset classes under consideration here, u.s. equities and international equities offer the best average annual returns at 12.52 and 11.06%, respectively. as expected, these two asset classes also have the highest risk based on the standard deviations of their returns. between the two fixed-income assets, the bond index has a higher return but also comes with a higher risk. the five-year ladder is the safest asset class with the worst return. the difference in the risk and return between the two fixed income approaches is probably because of the higher credit risk and longer duration associated with the bond index. the sharpe ratio points to the five-year ladder being the best asset class because of its extremely low risk. as expected, the correlation coefficients between equities and the two fixed-income assets are in general quite low. the next step is to examine the attractiveness of the five-year ladder versus the bond index in a portfolio context. we gauge their attractiveness by comparing the sharpe ratios of optimal portfolios making up of u.s. equities, international equities, and each of the two fixed-income assets. the correlation structure between the ladder and the two equity asset classes and the ladder’s low standard deviation may make the ladder more attractive despite its low return. the most common way to construct the optimal portfolio is the ex post approach. specifically, the tangency portfolio on the efficient frontier is obtained by using the historical average return of t-bill as the risk-free interest rate. we then examine the relative attractiveness of the two fixed-income assets by comparing the sharpe ratios of the tangency portfolios based on a ladder with united states and international equities and the bond index with united states and international equities according to eq. (2). in addition to the ex post approach, we also consider the resulting sharpe ratios at different values of the risk-free rate rf in eq. (2). by varying rf, we in effect trace out the efficient frontier by solving for different tangency portfolios.12 this approach allows us to solve for different optimal portfolios on the frontier for investors with different risk aversion parameters. thus, we can ascertain the diversification benefits to investors with different degrees of risk aversion based on the sharpe ratios and their statistical significance. this should be a useful exercise for dissecting the performance of the ladder strategy since its lower standard deviation of returns (table 1) suggests it is a somewhat less risky way to invest in bonds relative to simple bond indexing. this may make laddering a more suitable bond strategy for conservative investors holding a diversified portfolio of equity and fixed-income assets. by allowing for different degrees of risk aversion, we will be able to examine the suitability of the two bond strategies for investors of different risk preferences. table 2 reports the attractiveness of the ladder approach versus the bond index in a portfolio context. the results are consistent with the univariate results in table 1, which show that both fixed-income assets have the highest sharpe ratios among all asset classes. they continue to play a dominant role in the optimal portfolios. panel a in table 2 pertains to the case where portfolios are formed with the two equity assets and the five-year ladder. when the risk-free rate is zero, the ladder makes up about 96% of the optimal portfolio. note that a zero risk-free rate corresponds to a highly risk averse investor who loads up with bonds at the expense of equities. as the risk-free rate increases, u.s. equities start entering into the optimal portfolios. higher risk-free rates correspond to investors with higher risk tolerance. 186 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 based on the annual ex post risk-free rate of 4.56%, the optimal portfolio consists of about 12% in u.s. equities and 88% in the five-year ladder. international equities do not play any role in the optimal portfolios. an interesting finding is that the home bias, which is deemed to be undesirable in the finance literature, virtually causes no harm here because of the unattractive risk-return characteristics of the international equities during our sample period. panel b in table 2 contains the results for the bond index. similar to the above portfolio results for the five-year ladder, the bond index is a dominant asset class especially at lower levels of risk-free rates. this comes as no surprise as low levels of risk-free rates correspond to investors with high risk aversion. comparing the ladder portfolio results in panel a with those for the bond index in panel b, note that the efficient frontier with the ladder intersects the efficient frontier with the bond table 2 optimal portfolio allocations with short sale disallowed at different levels of risk-free interest rates panel a: with five-year ladder strategy risk-free rate portfolio weights mean return standard deviation of return sharpe ratiocrsp vw msci-eafe five-year ladder 0.0000 0.0366 0.0000 0.9634 0.0654 0.0312 2.096 0.0100 0.0426 0.0000 0.9574 0.0658 0.0314 1.776 0.0200 0.0515 0.0000 0.9485 0.0663 0.0317 1.459 0.0300 0.0656 0.0000 0.9344 0.0672 0.0324 1.147 0.0400 0.0917 0.0000 0.9083 0.0688 0.0341 0.845 0.0500 0.1569 0.0000 0.8431 0.0729 0.0401 0.570 0.0600 0.6074 0.0000 0.3926 0.1008 0.1067 0.383 0.0700 1.0000 0.0000 0.0000 0.1252 0.1715 0.322 0.0456a 0.1191 0.0000 0.8809 0.0705 0.0363 0.686 panel b: with barclays u.s. aggregated bond index risk-free rate portfolio weights mean return standard deviation of return sharpe ratiocrsp vw msci-eafe barclays bond index 0.0000 0.1397 0.0323 0.8281 0.0885 0.0683 1.295 0.0100 0.1527 0.0292 0.8181 0.0890 0.0687 1.149 0.0200 0.1699 0.0252 0.8050 0.0896 0.0693 1.004 0.0300 0.1932 0.0197 0.7871 0.0905 0.0703 0.861 0.0400 0.2270 0.0117 0.7613 0.0917 0.0719 0.720 0.0500 0.2792 0.0002 0.7207 0.0937 0.0749 0.583 0.0600 0.3523 0.0000 0.6477 0.0969 0.0812 0.454 0.0700 0.5193 0.0000 0.4807 0.1042 0.1002 0.341 0.0456a 0.2533 0.0055 0.7412 0.0927 0.0733 0.643 note: optimal portfolios are constructed based on the estimated risk-return parameters of the asset classes during the sample period of 1980–2015 as reported in table 1. athe optimal portfolio weights in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 187c.s. cheung, p. miu / financial services review 26 (2017) 181–203 index at about the ex post risk-free rate. below the ex post level, the frontier with the ladder dominates (higher sharpe ratios); whereas the frontier with the bond index dominates when the risk-free rate is above the ex post level (higher sharpe ratios). the reason for this result is because of the ability of the ladder in risk reduction. as reported in table 1, the risk level of the ladder in isolation is about one half of the bond index (standard deviation of 3.07% vs. 6.94%) and thus becomes very attractive to own for a risk averse investor. hence the frontier with the ladder dominates at lower level of risk-free rates. for example, table 2 indicates that, at the zero risk-free rate, the portfolio risk (i.e., standard deviation of return) with ladder is 3.12% versus that of 6.83% for the bond index portfolio. this meaningful risk reduction leads to a much better sharpe ratio for the ladder portfolio than the bond index portfolio (2.10 vs. 1.30). for an investor with higher risk tolerance, the lower risk associated with the ladder becomes less of an attraction. the ladder has the lowest return among all asset classes under consideration and its significance in the optimal portfolios drops at risk-free rates above the ex post level of 4.56%. in fact, u.s. equities become the dominant assets for investors with somewhat high risk tolerance. because u.s. equites have a lower standalone sharpe ratio than the ladder (see table 1), the introduction of this asset class results in lower portfolio sharpe ratios under the higher risk tolerance cases. it is this dominance of u.s. equities that causes the sharpe ratios of the ladder portfolios to be below those of the bond index portfolios in high risk tolerance cases. the evidence so far points to the superiority of the ladder over the bond index for risk averse investors as our five-year ladder offers more meaningful risk reduction and hence better sharpe ratios at lower levels of risk-free rates. the next question is whether the difference in the two efficient frontiers (or sharpe ratios) is indeed statistically significant. the statistical technique used here builds on the spanning tests used in the finance literature.13 the spanning tests address the question that when an additional asset class is added to a benchmark portfolio of existing asset classes, does the resulting efficient frontier of the combined asset classes indeed display a statistically significant improvement over that of the original benchmark asset classes? if the frontier of the combined asset classes displays a statistically significant shift to the left, the additional asset class should be added to the investor’s benchmark portfolio. the usual spanning tests are not used here as they require the testing of the entire frontier. as mentioned earlier, we are interested in the suitability of the ladder versus bond indexing for investors with different risk aversion. this requires the test of difference in the sharpe ratios given by eq. (2) at different levels of risk-free rates. let �1 defined by eq. (2) be the sharpe ratio of the tangency portfolio based on the benchmark asset classes and �2 be that after the addition of another asset class corresponding to the same risk-free rate. the null hypothesis to be tested is h0: �1 � �2. the benchmark portfolio consists of u.s. equities, international equities (i.e., non-north american equities as proxied by eafe), and the barclays bond index. we estimate the maximum sharpe ratio corresponding to a given level of risk-free rate for this benchmark portfolio. the combined portfolio then contains the benchmark asset classes plus the five-year ladder. if the sharpe ratio for the combined portfolio is statistically no different from that of the benchmark at the same level of risk-free rate, the ladder will be considered as a redundant asset. this test is repeated at different levels of risk-free rates. if the ladder is shown to be redundant at all levels of risk-free rates, portfolio investors should be content with only holding bond index, 188 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 without bothering with bond laddering. in designing the test this way rather than directly testing the statistical significance of the difference in the two sharpe ratios in each row across panels a and b of table 2, we can also entertain the possibility that both the ladder and the bond index may coexist in an investor’s optimal portfolio to enhance performance in a way that either one alone cannot do. the details of the test procedure as developed by glen and jorion (1993) are outlined in the appendix a. table 3 reports the test results. columns 7–9 are the portfolio allocations for the benchmark asset classes at various levels of risk-free rates and columns 2–5 are allocations for the combined asset classes. the results for the combined asset classes in columns 4–5 are most interesting. at very low levels of risk-free rates, the bond index is completely displaced by the ladder and the opposite is true at high levels of risk-free rates. at moderate risk-free rates, both fixed-income assets coexist. the last column reports the p-value of the null hypothesis that the sharpe ratio for the benchmark case is equal to that for the combined case. a p-value of 0.05 means that the equality can be rejected at five-percent level. at both zero and one percentage risk-free rates, the sharpe ratio for the combined portfolio that happens to contain no bond index is indeed statistically better than that of the benchmark portfolio. at six percent and seven percent risk-free rates, the combined portfolio contains no ladder. the combined and benchmark portfolios are in effect the same with the same asset allocations and the same sharpe ratios. the test of equal sharpe ratios is therefore irrelevant at these two levels of risk-free rates. between two percent and four percent risk-free rates, the difference between the sharpe ratios are statistically significant. the two fixed-income assets together in fact do a better job than either one alone. at five percent risk-free rate, the difference between the two sharpe ratios is no longer statistically significant. finally, at the ex post risk-free rate of 4.56%, the difference is marginally significant (p-value equals to 0.057). table 3 diversification benefits of five-year ladder: portfolio weights and sharpe ratios of tangency portfolios with and without ladder are reported at different levels of annual risk-free interest rates risk-free rate portfolios with five-year ladder portfolios without five-year ladder gj p-value portfolio weights sharpe ratio portfolio weights sharpe ratio crsp vw msci-eafe barclays bond index five-year ladder crsp vw msci-eafe barclays bond index 0.0000 0.0366 0.0000 0.0000 0.9634 2.096 0.1397 0.0323 0.8281 1.295 0.000 0.0100 0.0426 0.0000 0.0007 0.9567 1.776 0.1527 0.0292 0.8181 1.149 0.000 0.0200 0.0511 0.0000 0.0214 0.9275 1.460 0.1699 0.0252 0.8050 1.004 0.000 0.0300 0.0648 0.0000 0.0555 0.8797 1.152 0.1932 0.0197 0.7871 0.861 0.001 0.0400 0.0911 0.0000 0.1208 0.7881 0.861 0.2270 0.0117 0.7613 0.720 0.016 0.0500 0.1613 0.0000 0.2952 0.5435 0.610 0.2792 0.0002 0.7207 0.583 0.135 0.0600 0.3523 0.0000 0.6477 0.0000 0.454 0.3523 0.0000 0.6477 0.454 — 0.0700 0.5193 0.0000 0.4807 0.0000 0.341 0.5193 0.0000 0.4807 0.341 — 0.0456a 0.1198 0.0000 0.1920 0.6883 0.712 0.2533 0.0055 0.7412 0.643 0.057 note: optimal portfolios are obtained with short-sale disallowed and based on risk-return parameters estimated over the sample period of 1980–2015. the p-values of glen and jorion (gj) tests on equal sharpe ratios are reported in the last column of the table. athe portfolio results presented in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 189c.s. cheung, p. miu / financial services review 26 (2017) 181–203 in summary, the above empirical evidence based on the five-year ladder conclusively supports the argument for adding the ladder to a diversified equity portfolio for risk averse and moderately risk averse investors. 5.1. optimal term structure of a ladder portfolio the choice of the five-year maturity for our ladder is somewhat arbitrary. the practitioners’ literature is silent on the optimal term. as mentioned earlier, the return on the five-year ladder portfolio is simply an equally weighted average of past and current five-year bond returns. likewise, the return on a seven-year ladder is the equally weighted average of past and current seven-year bond returns. to the extent that longer term bond tends to offer a higher coupon, the longer term ladder will offer a higher coupon return. one would therefore expect a longer term ladder such as a seven-year ladder to deliver a higher return than a three-year ladder. at the same time, the conventional wisdom is that longer term fixedincome assets are deemed to be more volatile than shorter term ones. the question is whether the extra return associated with the longer term ladder is high enough to justify its higher risk. this is essentially an empirical question. table 4 provides the summary statistics for the three-year and seven-year ladders together with other asset classes being considered earlier. again, we assume each bond in the ladder portfolio is held till maturity and make no attempt to mark the value of each bond to its market value as interest rates change. the mark-to-the-market issue will be addressed in the next section. the surprise here is the higher return at lower risk for the longer term ladder. although we expect longer term ladders to earn higher returns, we do not expect at the same time they exhibit lower return volatility. how can we explain this seemingly counterintuitive finding? first, we need to realize that, since all bonds in each ladder are held till maturity without marking to the market, the current interest rate environment has no impact on the returns of the bonds already in the ladder. thus, the only reason why the return on a five-year ladder changes from one period to the next is because of the reinvestment of the proceeds obtained when the oldest five-year bond matures. this reinvestment risk for most part drives the risk of the ladder. the key to understand the absence of the expected result is the amount of reinvestment associated with ladders of various maturities. a three-year ladder intable 4 summary statistics of annual returns on different assets (36 years from 1980 to 2015) crsp vw msci-eafe three-year ladder five-year ladder seven-year ladder barclays bond index mean 0.1252 0.1106 0.0583 0.0631 0.0672 0.0814 standard deviation 0.1715 0.2219 0.0347 0.0307 0.0278 0.0694 sharpe ratioa 0.4645 0.2929 0.3660 0.5704 0.7770 0.5163 correlation crsp vw 1.000 0.683 0.118 0.155 0.158 0.169 msci-eafe 1.000 0.177 0.292 0.306 0.048 three-year ladder 1.000 0.968 0.928 0.590 five-year ladder 1.000 0.983 0.575 seven-year ladder 1.000 0.532 barclays bond index 1.000 note: asharpe ratio calculated based on a risk-free rate of 4.56%, which is the average yield of the three-month t-bill over the sample period. 190 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 volves one third of the portfolio to be reinvested each year; whereas a seven-year ladder churns over only one seventh of its portfolio every year. hence the three-year ladder exposes one third of its portfolio to reinvestment risk, comparing to one seventh for a seven-year ladder. the longer the term of a ladder, the smaller amount of reinvestment is required each year and thus the smaller its exposure to interest rate risk. this explains why, in table 4, we observe a monotonic decreasing standard deviation of return as the term of the ladder increases. analogous to table 3 for the five-year ladder, tables 5 and 6 provide the portfolio results for the three-year and seven-year ladders, respectively. the seven-year ladder in table 6 completely displaces the bond index in the combined portfolio at risk-free rates up to three percent. the sharpe ratio for the combined portfolio without bond index at the zero risk-free rate is almost twice that of the benchmark one with bond index (2.44 vs. 1.30). the dominance of the ladder at these levels of risk-free rates is statistically significant according to the p-values of glen-jorion test (last column in table 6). at the very high risk-free rate of seven percent, the bond index becomes more important again as this aggressive investor pursues return at the expense of risk and the safe seven-year ladder no longer fits her high risk preference profile. at the ex post risk-free rate of 4.56%, both fixed-income assets coexist to provide a statistically better sharpe ratio than the benchmark with the bond index alone. note that, in this case, the presence of the bond index in the combined portfolio is somewhat low at merely 5.4% of the overall combined portfolio. the three-year ladder in table 5 does contribute significantly to the sharpe ratio of the combined portfolio at risk-free rates up to three percent, demonstrating its attractiveness to risk averse portfolio investors. however, this superiority is achieved with the presence of the bond index in the combined portfolio, whereas the seven-year ladder delivers the superiority without any help from the bond index (see table 6). table 5 diversification benefits of three-year ladder: portfolio weights and sharpe ratios of tangency portfolios with and without ladder are reported at different levels of annual risk-free interest rates risk-free rate portfolios with three-year ladder portfolios without three-year ladder gj p-value portfolio weights sharpe ratio portfolio weights sharpe ratio crsp vw msci-eafe barclays bond index three-year ladder crsp vw msci-eafe barclays bond index 0.0000 0.0616 0.0000 0.0584 0.8800 1.771 0.1397 0.0323 0.8281 1.295 0.000 0.0100 0.0703 0.0000 0.0862 0.8435 1.496 0.1527 0.0292 0.8181 1.149 0.001 0.0200 0.0835 0.0000 0.1287 0.7878 1.229 0.1699 0.0252 0.8050 1.004 0.003 0.0300 0.1063 0.0000 0.2016 0.6921 0.975 0.1932 0.0197 0.7871 0.861 0.028 0.0400 0.1545 0.0000 0.3560 0.4895 0.749 0.2270 0.0117 0.7613 0.720 0.135 0.0500 0.2792 0.0001 0.7207 0.0000 0.583 0.2792 0.0002 0.7207 0.583 — 0.0600 0.3523 0.0000 0.6477 0.0000 0.454 0.3523 0.0000 0.6477 0.454 — 0.0700 0.5193 0.0000 0.4807 0.0000 0.341 0.5193 0.0000 0.4807 0.341 — 0.0456a 0.2154 0.0000 0.5514 0.2332 0.646 0.2533 0.0055 0.7412 0.643 0.328 note: optimal portfolios are obtained with short-sale disallowed and based on risk-return parameters estimated over the sample period of 1980–2015. the p-values of glen and jorion (gj) tests on equal sharpe ratios are reported in the last column of the table. athe portfolio results presented in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 191c.s. cheung, p. miu / financial services review 26 (2017) 181–203 the evidence so far points to the dominance of longer term ladders for risk averse investors. for the shorter term ladder, an investor will also need to invest in the bond index for its higher return and yet to retain the stability of the fixed-income asset class. 5.2. mark to the market effects one potential criticism of the above analysis is that we have ignored marked-to-market (mtm) price changes. a falling interest rate will not only lower the reinvestment return for a ladder but also produce price gain if the bond portfolio is to be liquidated. given the longer time to maturity of its constituent bonds, such price gain should be larger for a longer term ladder. in the above analysis, we ignore the pricing impact while focusing only on the reinvestment return. this approach can be justified by the assumption that all bonds in a ladder portfolio are held to maturity, which automatically eliminates any price risk from the ladder portfolio.14 what will be the implications if we also take the pricing effect into account when we measure the risk-return characteristics of our ladders? it is natural to expect return to be more volatile as the price risk also comes into play. more important, the significance of the mtm effect is expected to be different for ladders of different terms. given the longer time to maturity of its constituent bonds, the mtm effect should be stronger for longer term ladders. will the consideration of the mtm effect invalidate our previous conclusions drawn regarding the attractiveness of a ladder to a portfolio investor? to address this question, we replicate the mtm annual returns of our three-year, five-year, and sevenyear ladders over our sample period of 1980–2015 using the yields of one-year, two-year, three-year, five-year, and seven-year bonds obtained from the federal reserve website.15 let us take a three-year ladder as an example. a three-year ladder makes up of three bonds that mature in one, two, and three years, respectively. at the beginning of each year, we find out table 6 diversification benefits of seven-year ladder: portfolio weights and sharpe ratios of tangency portfolios with and without ladder are reported at different levels of annual risk-free interest rates risk-free rate portfolios with seven-year ladder portfolios without seven-year ladder gj p-value portfolio weights sharpe ratio portfolio weights sharpe ratio crsp vw msci-eafe barclays bond index seven-year ladder crsp vw msci-eafe barclays bond index 0.0000 0.0240 0.0000 0.0000 0.9760 2.442 0.1397 0.0323 0.8281 1.295 0.000 0.0100 0.0280 0.0000 0.0000 0.9719 2.086 0.1527 0.0292 0.8181 1.149 0.000 0.0200 0.0338 0.0000 0.0000 0.9662 1.733 0.1699 0.0252 0.8050 1.004 0.000 0.0300 0.0427 0.0000 0.0002 0.9571 1.382 0.1932 0.0197 0.7871 0.861 0.000 0.0400 0.0578 0.0000 0.0274 0.9149 1.039 0.2270 0.0117 0.7613 0.720 0.003 0.0500 0.0911 0.0000 0.0885 0.8204 0.717 0.2792 0.0002 0.7207 0.583 0.015 0.0600 0.2298 0.0000 0.3423 0.4279 0.462 0.3523 0.0000 0.6477 0.454 0.267 0.0700 0.5193 0.0000 0.4807 0.0000 0.341 0.5193 0.0000 0.4807 0.341 — 0.0456a 0.0724 0.0000 0.0543 0.8733 0.855 0.2533 0.0055 0.7412 0.643 0.002 note: optimal portfolios are obtained with short-sale disallowed and based on risk-return parameters estimated over the sample period of 1980–2015. the p-values of glen and jorion (gj) tests on equal sharpe ratios are reported in the last column of the table. athe portfolio results presented in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 192 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 the initial value of the ladder by calculating the prices of the three bonds with the prevailing one-year, two-year, and three-year yields, respectively. without any information regarding the specific coupon structure of the constituent bonds, here we assume they are all zerocoupon bonds and the observed yields are all zero-yields. to find out the year-end value of the ladder, we revalue the originally two-year (three-year) bond with the one-year (two-year) yield observed at the end of that year. note that the one-year bond held at the beginning of the year matures at year end yielding the par value and hence requiring no mtm. with the values of the ladder at the beginning and the end of the year, we can then calculate its mtm annual return for that year. in doing so, we recognize the fact that, at the end of each year, the shortest term bond will mature and the proceeds received will be reinvested in a new three-year bond. the above calculations are repeated for each of the years so as to replicate the historical annual mtm returns of the three-year ladder over our sample period.16 the historical mtm returns for the five-year and seven-year ladders are calculated in a similar fashion. analogous to tables 1 and 4, table 7 presents the summary statistics of our three mtm ladders together with all other asset classes under consideration. the mtm consideration seems to have very little effect on the return. the five-year mtm ladder has roughly the same return as the one without mtm at about 6.5% per annum. likewise, the three-year ladder with or without mtm offers an annual return of about 5.9%. the major difference the mtm makes is on risk. not surprisingly, the risk difference gets more pronounced as the term of the ladder increases, as the price risk being amplified by the longer time to maturity of the constituent bonds. for example, the standard deviation of the seven-year ladder with mtm at 5.96% is about twice that of the one without mtm at 2.78%. in the case of a three-year ladder, the risk for the three-year ladder without mtm is 3.47%, which is not much lower than the standard deviation of 4.39% for the three-year ladder with mtm. note that all ladders with or without mtm still have lower risk but also have lower return than the barclay bond index. the sharpe ratio of ladders with mtm is a monotonic increasing function as the term increases, albeit at a slower rate than ladders without mtm. in fact, the sharpe ratio for the seven-year mtm ladder at 0.436 is much higher than that of the three-year mtm ladder at 0.317. this suggests that the longer term ladder is more attractive than the shorter term one if they are held in isolation of a portfolio. to assess the value-adding ability of mtm ladders in a portfolio setting, we repeat the above portfolio analysis on our three mtm ladders. the results are reported in tables 8 to 10. based on the p-value of the glen-jorion test reported in the last column of table 8, the three-year mtm ladder contributes significantly to the optimal portfolio performance at risk-free rates up to about two percent, whereas the same ladder but without mtm consideration (see table 5) offers diversification benefit at risk-free rates up to three percent. note that the three-year ladder with mtm is completely displaced from the combined portfolio at risk-free rates four percent or higher. with mtm, the three-year ladder is attractive to the more risk averse investors comparing to the case without mtm. while it is obvious that fixed income assets should play a prominent role in the portfolio of a risk averse or moderately risk averse investor, what should be the composition of the fixed income assets and the effect of mtm on the composition is less obvious. tables 5 and 8 address the proper composition of fixed income assets and the mtm effect on the composition. first, both ladders (with or 193c.s. cheung, p. miu / financial services review 26 (2017) 181–203 t ab le 7 su m m ar y st at is tic s of an nu al re tu rn s on di ff er en t as se ts (3 6 ye ar s fr om 19 80 to 20 15 ) c r sp v w m sc ie a fe t hr ee -y ea r la dd er t hr ee -y ea r m t m la dd er fi ve -y ea r la dd er fi ve -y ea r m t m la dd er se ve nye ar la dd er se ve nye ar m t m la dd er b ar cl ay s bo nd in de x m ea n 0. 12 52 0. 11 06 0. 05 83 0. 05 95 0. 06 31 0. 06 58 0. 06 72 0. 07 16 0. 08 14 st an da rd de vi at io n 0. 17 15 0. 22 19 0. 03 47 0. 04 39 0. 03 07 0. 05 07 0. 02 78 0. 05 96 0. 06 94 sh ar pe ra tio a 0. 46 45 0. 29 29 0. 36 60 0. 31 66 0. 57 04 0. 39 84 0. 77 70 0. 43 62 0. 51 63 c or re la tio n c r sp v w 1. 00 0 0. 68 3 0. 11 8 0. 08 2 0. 15 5 0. 02 9 0. 15 8 0. 02 9 0. 16 9 m sc ie a fe 1. 00 0 0. 17 7 0. 02 6 0. 29 2 -0 .0 21 0. 30 6 -0 .0 18 0. 04 8 t hr ee -y ea r la dd er 1. 00 0 0. 88 8 0. 96 8 0. 78 1 0. 92 8 0. 68 5 0. 59 0 t hr ee -y ea r m t m la dd er 1. 00 0 0. 81 8 0. 96 1 0. 77 2 0. 90 0 0. 78 5 fi ve -y ea r la dd er 1. 00 0 0. 72 2 0. 98 3 0. 64 8 0. 57 5 fi ve -y ea r m t m la dd er 1. 00 0 0. 67 5 0. 98 2 0. 89 2 se ve nye ar la dd er 1. 00 0 0. 60 6 0. 53 2 se ve nye ar m t m la dd er 1. 00 0 0. 93 7 b ar cl ay s bo nd in de x 1. 00 0 n ot e: a sh ar pe ra tio ca lc ul at ed ba se d on a ri sk -f re e ra te of 4. 56 % , w hi ch is th e av er ag e yi el d of th e th re em on th t -b ill ov er th e sa m pl e pe ri od . 194 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 without mtm) and the bond index coexist in the optimal portfolio of a risk averse or moderately risk averse investor. secondly, comparing tables 5 and 8, the allocation to the bond index increases at the expense of the ladder when the mtm is introduced in table 8. for example, at the risk-free rate of two percent, the optimal allocation to the bond index increases from 12.9% to 22.4% when we incorporate the mtm effect in the returns of the three-year ladder. the reasons are the increase in the correlation between the three-year ladder and the bond index caused by the mtm from 0.590 to 0.785 and also the increase in the standard deviation of the three-year ladder from 3.47% to 4.39% (see table 7). these increases in correlation and risk dampen the conservative investor’s enthusiasm for the three-year ladder in favor of the bond index. note that mtm also reduces the overall demand for fixed income assets. the total fixed income holding at the zero, one-percent, and two-percent risk-free rates are 94%, 93%, and 92% of the overall portfolio, respectively, without mtm (table 5). table 8 shows the total fixed income holding now drops to 89%, 88%, and 86% at the same respective risk-free rates because of the less attractive risk profile of the mtm ladder. in the case of the five-year ladder results reported in table 9, the optimal portfolio based on the ex post risk-free rate contains no five-year mtm ladder, whereas the five-year ladder without mtm makes up about 69% of the portfolio allocation at the same risk-free rate (see table 3). note that the absence of the five-year mtm ladder in the portfolio at the ex post risk-free rate is replaced by about 74% of the portfolio in the bond index (see last row in table 9). the five-year mtm ladder is clearly an inferior fixed-income product comparing to the bond index. the reason for this outcome is not difficult to understand. the five-year mtm ladder and the bond index has a high correlation at 0.892 (see last column of table 7). in effect, the five-year mtm ladder and the bond index are close substitutes of each other. yet the bond index has a higher sharpe ratio (0.516 vs. 0.398). hence it is no surprise that the five-year mtm ladder fails to make the cut when comparing to the bond index. nevertheless, table 8 diversification benefits of three-year mtm ladder: portfolio weights and sharpe ratios of tangency portfolios with and without ladder are reported at different levels of annual risk-free interest rates risk-free rate portfolios with 3-year mtm ladder portfolios without 3-year mtm ladder gj p-value portfolio weights sharpe ratio portfolio weights sharpe ratio crsp vw msci-eafe barclays bond index three-year mtm ladder crsp vw msci-eafe barclays bond index 0.0000 0.0978 0.0114 0.0560 0.8347 1.494 0.1397 0.0323 0.8281 1.295 0.009 0.0100 0.1098 0.0115 0.1223 0.7564 1.277 0.1527 0.0292 0.8181 1.149 0.022 0.0200 0.1282 0.0116 0.2243 0.6359 1.070 0.1699 0.0252 0.8050 1.004 0.067 0.0300 0.1602 0.0118 0.4015 0.4266 0.880 0.1932 0.0197 0.7871 0.861 0.203 0.0400 0.2270 0.0117 0.7613 0.0000 0.720 0.2270 0.0117 0.7613 0.720 — 0.0500 0.2792 0.0002 0.7207 0.0000 0.583 0.2792 0.0002 0.7207 0.583 — 0.0600 0.3523 0.0000 0.6477 0.0000 0.454 0.3523 0.0000 0.6477 0.454 — 0.0700 0.5193 0.0000 0.4807 0.0000 0.341 0.5193 0.0000 0.4807 0.341 — 0.0456a 0.2533 0.0055 0.7412 0.0000 0.643 0.2533 0.0055 0.7412 0.643 — note: optimal portfolios are obtained with short-sale disallowed and based on risk-return parameters estimated over the sample period of 1980–2015. the p-values of glen and jorion (gj) tests on equal sharpe ratios are reported in the last column of the table. athe portfolio results presented in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 195c.s. cheung, p. miu / financial services review 26 (2017) 181–203 the five-year mtm ladder completely replaces the bond index at risk-free rates up to two percent. thus, the five-year mtm ladder is still very attractive to investors with high risk aversion. as in the case of the three-year ladder, the mtm reduces the percentage allocation to the total fixed income holding and also tilts the composition of the fixed income asset away from the less attractive ladder toward bond index for the moderately risk averse investors. in the only two cases in table 9 where the two fixed income assets coexist at the risk-free rates of three percent and four percent, the p-values indicate that the coexistence of the two fixed income components adds no statistically significant diversification benefit. the results in table 9 favor the five-year mtm ladder for the risk averse investors and the bond index otherwise.17 the results in table 10 provide the case of the seven-year mtm ladder in which the effect of the mtm is strongly felt. at first brush, the seven-year mtm ladder appears to be very dominant as it completely or almost completely displaces the bond index at risk-free rates up to three percent. also recall from the summary statistics reported in table 7 that the seven-year mtm ladder has the best sharpe ratio among the three mtm ladder strategies. it therefore manages to make a presence, albeit small, in the optimal portfolio at the ex post risk-free rate in table 10, as opposed to the cases for the three-year and five-year mtm ladders that are completely absence at the same ex post rate. unfortunately, the p-values indicate the presence of the seven-year mtm ladder adds no statistically significant benefit except at the zero risk-free rate. its dominance at one-percent risk-free rate is weakly statistically significant at the 10% level. for most part, the ladder is not a potent instrument except to the most risk averse investors. as in the cases of the three-year and five-year ladder, the mtm reduces the percentage allocation to the total fixed income holding and also tilts the composition of the fixed income asset away from the less attractive ladder toward bond index for the moderately risk averse investors. among the cases of risk-free rates being considered, the two fixed income assets coexist only at risk-free rates of three percent and table 9 diversification benefits of five-year mtm ladder: portfolio weights and sharpe ratios of tangency portfolios with and without ladder are reported at different levels of annual risk-free interest rates risk-free rate portfolios with five-year mtm ladder portfolios without five-year mtm ladder gj p-value portfolio weights sharpe ratio portfolio weights sharpe ratio crsp vw msci-eafe barclays bond index five-year mtm ladder crsp vw msci-eafe barclays bond index 0.0000 0.1244 0.0149 0.0000 0.8607 1.473 0.1397 0.0323 0.8281 1.295 0.009 0.0100 0.1369 0.0130 0.0000 0.8501 1.274 0.1527 0.0292 0.8181 1.149 0.021 0.0200 0.1545 0.0102 0.0000 0.8353 1.078 0.1699 0.0252 0.8050 1.004 0.061 0.0300 0.1765 0.0095 0.1739 0.6402 0.890 0.1932 0.0197 0.7871 0.861 0.151 0.0400 0.2177 0.0086 0.5333 0.2404 0.722 0.2270 0.0117 0.7613 0.720 0.375 0.0500 0.2792 0.0002 0.7207 0.0000 0.583 0.2792 0.0002 0.7207 0.583 — 0.0600 0.3523 0.0000 0.6477 0.0000 0.454 0.3523 0.0000 0.6477 0.454 — 0.0700 0.5193 0.0000 0.4807 0.0000 0.341 0.5193 0.0000 0.4807 0.341 — 0.0456a 0.2533 0.0055 0.7412 0.0000 0.643 0.2533 0.0055 0.7412 0.643 — note: optimal portfolios are obtained with short-sale disallowed and based on risk-return parameters estimated over the sample period of 1980–2015. the p-values of glen and jorion (gj) tests on equal sharpe ratios are reported in the last column of the table. athe portfolio results presented in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 196 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 four percent. however, the p-values indicate that the coexistence of the two fixed income components adds no statistically significant diversification benefit. the results in table 10 also favor the seven-year mtm ladder for the risk averse investors and the bond index otherwise. what can we say about the implication of the mtm effect across ladders of different terms? first, note that none of the shorter term mtm ladders enters the optimal portfolio at the ex post risk-free rate. moreover, the weights of mtm ladders drop quite dramatically at higher risk-free rates including the ex post risk-free rate comparing to the ladders without mtm. the weights at the ex post risk-free rate for the three-year, five-year, and seven-year ladders without mtm are 23.32%, 68.83%, and 87.33%, respectively. the mtm effect reduces the weights of first two ladders to zero and the seven-year ladder to a mere 3.9%. there seems to be a relationship between the magnitude of the impact caused by the mtm and the term of the ladder with a greater reduction in the weight of the longer term ladder. to understand the above effect of mtm on the allocation of money to ladders, one has to see the allocation behavior with mtm and then the allocation behavior without mtm. first, in the case of allocations with mtm, notice that the correlation between the mtm ladders and the bond index rises as the term increases. the correlation coefficient increases monotonically with the term ranging from 0.785 for the three-year mtm ladder to 0.937 for the seven-year mtm ladder according to table 7, whereas the correlations between the ladders without mtm and the bond index are no higher than 0.6. the seven-year mtm ladder is an almost perfect substitute for the bond index. the high correlations reduce any potential diversification benefit of holding both fixed income assets simultaneously and invalidate the rationale for both fixed income assets to coexist. the decision to own predominantly only one fixed income asset depends on the risk preference of the individual investor. at higher risk-free rates, the bond index offers the investor with higher risk tolerance a better return and table 10 diversification benefits of seven-year mtm ladder: portfolio weights and sharpe ratios of tangency portfolios with and without ladder are reported at different levels of annual risk-free interest rate risk-free rate portfolios with seven-year mtm ladder portfolios without seven-year mtm ladder gj p-value portfolio weights sharpe ratio portfolio weights sharpe ratio crsp vw msci-eafe barclays bond index seven-year mtm ladder crsp vw msci-eafe barclays bond index 0.0000 0.1555 0.0156 0.0000 0.8289 1.389 0.1397 0.0323 0.8281 1.295 0.039 0.0100 0.1680 0.0133 0.0000 0.8187 1.217 0.1527 0.0292 0.8181 1.149 0.064 0.0200 0.1850 0.0101 0.0000 0.8049 1.047 0.1699 0.0252 0.8050 1.004 0.097 0.0300 0.2039 0.0079 0.1187 0.6696 0.880 0.1932 0.0197 0.7871 0.861 0.192 0.0400 0.2299 0.0063 0.4250 0.3388 0.723 0.2270 0.0117 0.7613 0.720 0.351 0.0500 0.2792 0.0002 0.7207 0.0000 0.583 0.2792 0.0002 0.7207 0.583 — 0.0600 0.3523 0.0000 0.6477 0.0000 0.454 0.3523 0.0000 0.6477 0.454 — 0.0700 0.5193 0.0000 0.4807 0.0000 0.341 0.5193 0.0000 0.4807 0.341 — 0.0456a 0.2534 0.0050 0.7026 0.0390 0.643 0.2533 0.0055 0.7412 0.643 0.473 note: optimal portfolios are obtained with short-sale disallowed and based on risk-return parameters estimated over the sample period of 1980–2015. the p-values of glen and jorion (gj) tests on equal sharpe ratios are reported in the last column of the table. athe portfolio results presented in the last row are obtained by assuming a risk-free rate equals to the average three-month t-bill yield of 4.56% over the sample period of 1980–2015. 197c.s. cheung, p. miu / financial services review 26 (2017) 181–203 thus dominates the longer-term ladders with mtm. this explains in part the drop in the weight of the seven-year ladder from 87.33% to a mere 3.9% when mtm is considered. this also explains the earlier result that the seven-year mtm ladder with a lower risk and lower return profile is preferred at lower risk-free rates. the net outcome is to hold either the bond index or the seven-year ladder depending on the levels of the risk-free rates. the second reason for the larger drop in the allocations for the longer term ladders in the presence of mtm can be explained by examining the allocation behavior without the mtm. the longer term ladders without mtm not only have lower risk than the shorter term ladders without mtm because of the lower reinvestment risk mentioned earlier, they also have slightly lower correlations with the bond index. these attractive risk and correlation characteristics translate into a greater diversification benefit as the term increases. in fact, one can argue that it is the lack of mtm consideration that makes ladders unusually attractive. this is especially true in the case of the seven-year ladder without the mtm. the seven-year ladder without mtm has the lowest standard deviation among all fixed income assets at 2.78% and at the same time the lowest correlation with the bond index. the above attractive risk-return characteristics brought about by the absence of the mtm consideration make the seven-year ladder the most dominant asset at 87.33% in the optimal portfolio at the ex post risk-free rate. this unusual advantage disappears with mtm and hence the dramatic reduction in portfolio weight. the above results of higher correlations between longer term mtm ladders and the bond index also lead to another interesting result. longer term mtm ladders do not statistically add significant diversification benefit to the benchmark portfolio with the bond index. none of the p-values supports the coexistence of a seven-year ladder or a five-year ladder with the bond index whenever mtm is introduced. in the absence of mtm, the p-values support far more incidences of statistically significant improvement in the shape ratios when the ladders and the bond index coexist for investors with moderate risk aversion. the above insights can only be gained through our use of tests of statistical significance. it is our statistical tests that pinpoint the effect of mtm of longer term ladders and its investment implications. the previous conclusion that ladder portfolios are more attractive to more risk averse investors still holds when the mtm is introduced. the mtm also increases the correlations and the standard deviations of longer term ladders and reduces its attractiveness in a portfolio context. these two factors lead to binary outcomes for longer term ladders in which an investor will hold either the bond index or the longer term ladders depending on the investor’s risk aversion but not both fixed income assets at the same time. 6. conclusion this article attempts to bridge a major gap in personal finance by addressing a popular approach to bond investing that is rarely mentioned in the academic literature. investors are confronted with a choice between two similar fixed-income investment instruments that are passive in nature with no way to choose either one or both together. further, there is absolutely no objective guidance as to the proper term of the ladder strategy in the existing literature. 198 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 this article offers insights into the above missing pieces. first, by allowing the risk aversion behavior to change, we show that the two instruments are suitable for investors with different risk preference. because of their risk reduction ability, ladders are particularly suitable for conservative investors with somewhat higher degrees of risk aversion. second, longer term ladders appear to offer investors a better risk-return tradeoff than shorter term ladders. this is especially true for investors who believe that marking to market is not necessary as all bonds are held till maturity. third, marking to the market in general raises the correlations between ladders and the bond index, and increases the risk of ladders as the term increases. this dampens the usefulness of longer term ladders except for the most conservative investors. the higher correlations also make the coexistence of longer term ladders and the bond index less likely. fourth, ladders and bond index may coexist within an optimal portfolio to improve the risk-adjusted return especially for investors with moderate risk aversion in the world without mtm. in the case of mtm, the coexistence of both fixed income assets is likely only for shorter term ladders. marking to the market has a profound effect on the composition of the fixed-income assets in an optimal portfolio. first, it makes fixed-income assets as a group less attractive as mtm heightens the correlation structure of fixed-income assets. second, it affects the decision to go with a long term ladder or a short term ladder. if an investor believes mtm is unimportant because the bonds are held till maturity, the longest term ladder offers the best outcome. finally, it reduces the likelihood that the two fixed-income assets will coexist in a portfolio as longer term ladders act like the bond index. notes 1 see, for example, the coverage of the strategy in hirt, block, and basu (2008, chapter 12), bernstein (2003), bohlin and strickland (2004), and scatizzi (2009). 2 fabozzi (2007) mentions the ladder strategy in the context of active yield curve strategies for investment professionals. his focus there is on comparing the return performance of a barbell strategy versus a bullet strategy when the yield curve shifts, whereas retail investors/advisors use laddered portfolios as a passive strategy. laddered portfolios as examined in the academic literature are usually in conjunction with barbells and bullets. again, as in the case of fabozzi, the emphasis is on the risk-return characteristics of various active bond strategies as interest rates changes. 3 fabozzi (1995) is the refreshing exception in which laddering is mentioned. again, its mention is in the context of interest rate risk and the comparison is against active barbell and bullet strategies. 4 see, for example, ingersoll (1987, chapter 4), for the derivation. 5 some may construct a ladder by using bonds maturing every six months or every three months. in the case of a five-year ladder using bonds maturing every six months, the ladder will consist of the six-month bond, one-year bond, 1 1/2-year bond, . . . , and the five-year bond. the six-month bond maturing six month later will be reinvested in a new five-year bond. the initial one-year bond maturing one year later will be reinvested in a five-year bond to maintain the same maturity structure. a new five-year 199c.s. cheung, p. miu / financial services review 26 (2017) 181–203 bond will be bought every 6 months in this case. the return on this ladder portfolio is again simply an equally weighted average of past and current five-year bond returns. while ladders can be built with different “steps” such as six months or three months, our statement that the portfolio return is simply an equally weighted average of past and current five-year bond returns remains true. we use bonds maturing every year in the construction of ladder portfolios in this article. 6 the annual yield data of bonds of other maturities are also obtained to address the “mark-to-the-market” issue as explained later. 7 it would appear that we have ignored the reinvestment of coupons at returns different form the initial yield. this is not an issue if the investor uses certificates of deposit (cd) with compounding of interest at the initial rate. for investors who use treasury notes with regular coupons, the reinvestment returns on the intermediate coupons may be different from the initial yield. we will consider the scenario that the reinvestment returns happen to be the prevailing t-bills rates at the time the coupons are received later in the article. see note 17. 8 for a description of the origin of the barclays u.s. aggregate bond index, see cui (2013). 9 both vanguard and blackrock offer exchange traded funds (etfs) replicating the index. the ishares core us aggregate bond etf offered by blackrock is the largest bond etf among its offerings at about $35 billion of assets under management as of march 31, 2016. it has a management expense ratio of 0.08 percent. 10 we do not expect the bond index to be subject to much credit risk. as of december 31, 2015, about 73 percent of the barclays u.s. aggregate bond index consists of aaa bonds. 11 while the choice of five-year ladder is somewhat arbitrary, it is not highly unusual. us treasury notes, which are ideal for the construction of ladders, are available in terms of 2, 3, 5, 7, and 10 years. most banks offer cds with maturities up to 10 years. the five-year ladder happens to be the medium term of available treasury notes or cds. 12 elton, gruber, brown, and goetzmann (2006, chapter 6) suggest the variation of the risk-free rate as a way to trace out the frontier. 13 see jobson and korkie (1989) for a discussion of the spanning tests. 14 the bond index requires portfolio re-balancing to maintain its term structure and hence the use of market prices for their bond holding can be justified. 15 unfortunately, yield information for four-year and six-year bonds are unavailable from the federal reserves website. we impute the four-year (six-year) yield by interpolating the yields between the three-year (five-year) and five-year (seven-year) bonds. 16 see appendix b for a detailed numerical illustration of the procedure. 17 instead of using zero coupon bonds in constructing the marked-to-the-market portfolio, we repeated the exercise for the five-year ladder using coupon bonds issued at par and the coupons are invested at the prevailing risk-free rates until maturity. the bonds are marked to the market at the end of each year. the return on this five-year ladder over the same sample period is slightly lower than those reported in tables 7 and 9. the conclusion that the ladder and the bond index do not coexist remains unaffected. to conserve space, the detailed results are not produced here and available on request. 200 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 appendix a: statistical test for the difference in portfolio sharpe ratios let us define t as the number of monthly observations. also, let �1 be the sharpe ratio of the tangency portfolio based on the three benchmark asset classes and �2 be the sharpe ratio after the addition of an addition asset class. the null hypothesis to be tested is h0: �1 � �2. when short selling is allowed, the test statistic presented in gibbons, ross, and shanken (1989) and jobson and korkie (1989) is: f � �t � 4� � �2 2 � �1 2 1 � �1 2 , (3) where f follows a f-distribution with 1 and t-4 degrees of freedom. in this study, we consider the case where short selling is disallowed. under this condition, the f statistic of eq. (3) will not be following the f-distribution and needs to be simulated before hypothesis tests can be conducted. we conduct the simulations by following the method proposed by glen and jorion (1993). we first estimate the means, variances, and the covariances using historical return data. the expected return of the additional asset class (i.e., a ladder strategy in this study) is then modified so that the original tangency portfolio is still mean-variance efficient after this additional asset class is included. this in effect ensures the null hypothesis is satisfied in the subsequent simulation exercise. with these modified parameters, t random samples of joint returns are drawn from a multivariate normal distribution. based on these simulated returns, a new set of means and variance-covariance matrix are estimated. sharpe ratios of the tangency portfolios with and without the additional asset class can then be estimated. finally, the value of the test statistic of eq. (3) is computed and recorded. the empirical distribution of the statistic is generated under the null hypothesis by repeating this process 1,000 times. appendix b: a numerical example of calculating the annual mtm ladder return suppose we start implementing a three-year ladder strategy at the end of 1979. by assuming the observed one-year, two-year, and three-year bond yields of 11.98%, 11.39%, and 10.71% as zero yields, the prevailing prices of the one-year, two-year, and three-year zero-coupon bonds with face value of $100 are $89.30, $80.59, and $73.70, respectively. consider a notional starting investment value of $100 for our three-year ladder portfolio. it is equally split among the three bonds (i.e., we are investing $33.33 in each of the three bonds). after one year (i.e., at the end of 1980), the first bond matures and its price is $100. the second bond now has a remaining time to maturity of one year; whereas the third bond will mature in two years. suppose the one-year and two-year bond yields are now 14.88% and 14.08%. based on these yields, the prevailing market prices of the second and third bonds become $87.05 and $76.84. the annual mtm return of the ladder portfolio in 1980 is, therefore, the weighted average return of the three bonds over that year, i.e., 201c.s. cheung, p. miu / financial services review 26 (2017) 181–203 $33.33 $100 � �$100.00 � $89.30� $89.30 � $33.33 $100 � �$87.05 � $80.59� $80.59 � $33.33 $100 � �$76.84 � $73.70� $73.70 � 8.09% at the end of 1980, the market values of the positions on the three bonds are respectively $37.33 (� $33.33 � 100/89.30), $36.01 (� $33.33 � 87.05/80.59), and $34.75 �� $33.33 � 76.84 73.70�. the overall value of the ladder portfolio is, therefore, $108.09. as the first bond has matured, the $37.33 is reinvested in a newly issued three-year bond. with a three-year bond yield of 13.65%, a newly issued three-year zero-coupon bond should be selling at $68.12. now, the ladder portfolio consists of a three-year bond (the first bond), a one-year bond (the second bond), and a two-year bond (the third bond). after another year (i.e., at the end of 1981), the second bond matures and its value is $100. the third bond now has a remaining time to maturity of one year; whereas the first bond will mature in two years. suppose the one-year and two-year bond yields are now 12.85% and 13.29%. based on these yields, the prevailing market prices of the first and third bonds become $77.91 and $88.61. the annual mtm return of the ladder portfolio in 1981 is, therefore, the weighted average return of the three bonds over that year, i.e., $37.33 $108.09 � �$77.91 � $68.12� $68.12 � $36.01 $108.09 � �$100.00 � $87.05� $87.05 � $34.75 $108.09 � �$88.61 � $76.84� $76.84 � 14.84% at the end of 1981, the market values of the positions on the three bonds are respectively $42.69 (� $37.33 � 77.91/68.12), $41.37 (� $36.01 � 100.00/87.05), and $40.07 �� $34.75 � 88.61 76.84�. as the second bond has matured, the $41.37 is reinvested in a newly issued three-year bond. now, the ladder portfolio consists of a two-year bond (the first bond), a three-year bond (the second bond), and a 1-year bond (the third bond). the above calculations are repeated for each of the subsequent years of the sample period to obtain the time-series of annual mtm return of the three-year ladder. references alexander, g. j., sharpe, w. f., & bailey, j. v. (2001). fundamentals of investments. upper saddle river, nj: prentice hall. bernstein, p. j. (2003). investing after 50. journal of accountancy, 195, 20–27. bodie, z., kane a., & marcus, a. j. (2007). essentials of investments. new york, ny: mcgraw hill. bohlin, s., & strickland, g. (2004). climbing the ladder: how to manage risk in your bond portfolio. aaii journal, 5–8. campbell, j. y., & viceira, l. m. (2001). who should buy long-term bonds? the american economic review, 91, 99–127. 202 c.s. cheung, p. miu / financial services review 26 (2017) 181–203 canner, n., mankin, n. g., & weil, d. n. (1997). an asset allocation puzzle. the american economic review, 87, 181–191. chang, c. e., & krueger, t. (2010). do enhanced index funds live up to their name? financial services review, 19, 145–162. cheung, c. s., kwan, c. c. y., & sarkar, s. (2010). bond portfolio laddering: a mean-variance perspective. journal of applied finance, 1, 103–109. cui, c. (2013). barclays agg had modest origin. wall street journal (03 apr 2013: c.2). dilellio, j. a., & stanley, d. j. (2011). etf trading strategies to enhance client wealth maximization. financial services review, 20, 145–163. elton, e. j., gruber, m. j., brown, s. j., & goetzmann, w. n. (2006). modern portfolio theory and investment analysis (7th ed.). new york, ny: wiley. fabozzi, f. j. (2007). bond markets, analysis, and strategies (6th ed.). upper saddle river, nj: pearson. fabozzi, f. j. (1995). investment management. englewood cliffs, nj: prentice hall. garbade, k. d. (1996). fixed income analytics. cambridge, ma: the mit press. gibbons, m. r., ross, s. a., & shanken, j. (1989). a test of the efficiency of a given portfolio. econometrica, 57, 1121–1152. glen, j., & jorion, p. (1993). currency hedging for international portfolios. journal of finance, 48, 1865–1886. hirt, g.a., block, s. b., & basu, s. (2008). investment planning for financial professionals. new york, ny: mcgraw-hill. ingersoll jr., j. e. (1987). theory of financial decision making. lanham, md: rowman & littlefield publishers. jobson, d., & korkie, b. (1989). a performance interpretation of multivariate tests of asset set intersection, spanning, and mean variance efficiency. journal of financial and quantitative analysis, 24, 185–204. jones, c. p. (1996). investments (5th ed.). new york, ny: wiley. judd, k. l., kubler, f., & schmedders, k .(2011). bond ladders and optimal portfolios. the review of financial studies, 24, 4123–4166. kaufman, g., bierwag, g. o., & toevs, a. (1983). innovations in bond portfolio management: duration and immunization. greenwich, ct: jai press. lu, l., wang, j., & zhang, g. (2012). long term performance of leveraged etfs. financial services review, 21, 63–80. prather, l. j., chu, t, mazumder, m. i., & topuz, j. c. (2009). index funds or etfs: the case of the s&p 500 for individual investors. financial services review, 18, 213–230. scatizzi, c. (2009). laddered bond portfolios. aaii journal, june 2009. http://www.aaii.com/journal/article/ laddered-bond-portfolios sundaresan, s. (1997). fixed income markets and their derivatives. cincinnati, oh: southwestern publishing. 203c.s. cheung, p. miu / financial services review 26 (2017) 181–203 portfolio performance with inverse and leveraged etfs james a. dilellioa,*, rick hessea, darrol j. stanleya agraziadio school of business and management, pepperdine university, 24255 pacific coast highway, malibu, ca 90263, usa abstract turbulent economic and financial times require investors and financial planners to investigate new ways to handle the goal of wealth maximization. this article investigates passive investment strategies that use inverse or leveraged equity exchanged-traded funds (etfs) in their asset allocation, and quantifies the long-term impact on portfolio performance for the purpose of improving the risk-reward tradeoff. monte carlo simulations are used, drawing samples from distributions created by two distinct time periods of historical daily market returns. the findings suggest that, whereas these products are generally not recommended within long-term passive investment strategies, potential diversification benefits exist, dependent on the behavior of equity and debt markets. these findings could materially alter long-term passive portfolio construction methods currently in use by financial planners and individual investors seeking potential diversification benefits using etfs. © 2014 academy of financial services. all rights reserved. jel classification: g1; c6 keywords: inverse etfs; leveraged etfs; diversification; portfolio; asset allocation 1. introduction the purpose of this article is to provide individual investors and financial planners with guidance on the possible use of inverse and leverage financial instruments to improve the distribution of terminal wealth. the recent economic downturn has called into question the effectiveness of diversification and led some investors to consider alternative investments as part of their overall portfolio. a recent study by arshanapalli et al. (2010) shows that, * corresponding author. tel.: �1-714-403-0085; fax: �1-949-223-2575. e-mail address: james.dilellio@pepperdine.edu (j. dilellio) financial services review 23 (2014) 123–149 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. although more severe than many of the bear markets before it, the recent downturn did benefit from diversification and alternative investments such as gold. so-called “wide diversification” is not entirely new (mulvey et al., 2007), and suggests that increasing the number of asset classes offers significant benefit. ammermann et al. (2011) suggests the effectiveness of a style-rotation over sector/industry rotation using exchange-traded funds (etfs) as a long-flat or synthetic put equity strategy whereas johnston et al. (2013) suggested portfolio insurance as an alternative. however, the use of inverse and leveraged products as a new asset class in a diversified portfolio has yet to be adequately examined. the results shown here represent a first attempt to quantify their long-term impact for passive investors seeking improved diversification. lack of study at a portfolio level does not suggest little is known about the impact of holding these types of investments long-term. to the contrary, a comprehensive study by cheng and madhaven (2009) shows how inverse etfs need to be rebalanced on a daily basis to maintain a constant leverage, and how this can lead to wealth destruction. this wealth destruction occurs largely because of the resulting path dependence on accumulated wealth that can easily diverge from the underlying index over longer holding periods. additionally, lu, wang, and zhang (2012) show how the longer term performance diverges from the benchmark through periods of up to one year, and caution investors on their use as substitutes for benchmark indices. this wealth destruction is also aggravated by higher volatility, although trainor and baryla (2008) notes an interesting corollary that suggests some of these leveraged etfs can outperform their respective benchmarks in periods of low volatility. giese (2010) aptly summarizes the benefit of holding leveraged funds in bullish markets, as well as their benefit as an investment product remaining positive, unlike a short position, but highlights these benefits are offset by increased performance volatility. this article is organized as follows: section 2 discusses the background and history of inverse and leveraged funds. section 3 states the research hypothesis, and section 4 describes the methods used to test our hypothesis. section 5 presents the terminal wealth distributions for a large number of simulated returns randomly sampled from non-overlapping historical stock and bond time series. section 6 highlights the relevant statistics from the simulated return distributions and also quantifies risk-adjusted returns using the sharpe ratio. section 7 enhances the results from section 6 by considering other investment options for stock and bond funds. the article concludes with section 8. 2. background of inverse and leveraged etfs although the first inverse etf came into existence in 2007, both inverse and leveraged open-ended mutual funds have been in existence far longer. since 1997, profunds has offered an inverse and leveraged version of the s&p 500 index through two mutual funds. by mid-2010, 150 different inverse and leveraged etfs were available with a total of $30b of assets under management (see guedj et al., 2010) although their track record is brief, the history of these products so far suggests that they are meeting the objectives contained within their offering prospectuses. specifically for this proposed research article, we assume that they effectively meet their daily objectives of inverse (�1�) or double (2�) daily returns of 124 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 their underlying index through the use of swaps and other derivative instruments. unfortunately, the daily releveraging of these products results in potential investment management challenges and risks.1 many researchers, including barnhorst and cocozza (2010), cheng and madhaven (2009), guedj et al. (2010), trainor and baryla (2008), and lu et al. (2012) have noted that daily rebalancing has tended to reduce their stated investment effectiveness over longer holding periods. furthermore, the limited history of these products prevents longer term studies of historical returns within a diversified portfolio. 3. research hypothesis this research article attempts to uncover whether there are any potential diversification benefits to a passively managed portfolio that holds a portion of its holdings in either inverse or leveraged equity etfs. to test this hypothesis, three typical long-term investors were considered who seek diversification by holding a broad stock and bond index fund in a tax deferred retirement account that is rebalanced annually. these investors were assumed to be representative of three separate tolerances to risk, and that the risk tolerance can be adjusted based on their portfolio exposure to stocks and bonds. the objective is to determine whether there is any risk-return benefit of these products in a diversified portfolio, or whether, as the previous studies have shown, they should be avoided entirely by long-term passive investors. perhaps, these products should remain clearly in the hand of speculators and short-term traders. greater regulatory protections for retail investors may be warranted. the strategic allocation between stocks and bonds is highly dependent on the time horizon of the investor and their tolerance for risk. this failure to quantify risk in an operational context for the strategic allocation of investment assets has caused the rise of mechanical methods that reflects classical thinking by connecting risk-aversion to one’s age. thus, the “rule of 120” has appeared in the literature. this suggests that for the stock allocation should be 120 minus the age of the individual. the rule has been modified by others, from about 100 to 130, to accommodate different economic and investment cycles. the remainder of the money is then allocated to fixed interest instruments.2 the popularity of this approach has motivated many fund providers to offer single funds that automatically follow this approach, naming them either lifecycle or target date funds (tdfs), although these styles may be too aggressive for risk adverse investors (pfau, 2010). table 1 was formatted to reflect the strategic asset allocations used in this study. additionally, the investor is assumed to annually rebalance their portfolio to these allocations over a 10-year investment period. in contrast, this study is interested in comparing simulated returns of the above investors versus ones who choose to allocate a small portion of their assets at the beginning of each year of a 10-year investment period to either an inverse or leveraged stock fund. this alternative group of investors is represented by allocations that appear in table 2, and includes the same stock to bond ratio as established in table 1. the distinction in table 2 is that 10% of the investor’s retirement account will be allocated towards an inverse or leveraged stock fund. again, this study assumes that the investor rebalances to these allocations at the beginning of each year. 125j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 4. research methods the previous section noted the logic for the strategic asset allocation between stocks and bonds. although many believe that one must at the same time be specific about the types of equity or fixed income asset chosen, this article suggests that this is more of a tactical asset allocation question. for example, equities could be small cap, mid cap, large cap, or reits. for fixed income, this could be short-term bonds; intermediate-term bonds; long-term bonds, high yield bonds; inflation-indexed; or money markets and/or t-bills. further complicating the issue, one could consider for equities international exposure and for bonds the question of corporate versus government bonds further complicated by domestic versus international. thus, the possible pragmatic tactical choices are large. to minimize potential problems that could arise from the tactical allocation question, two well-known and large mutual funds were chosen to begin our analysis, representative of rational investment choices and proxies for stocks and bonds. the vanguard s&p 500 index fund (vfinx) was chosen as the proxy to represent domestic equity. this fund attempts to track the performance of the s&p 500, a widely recognized benchmark of the u.s. stock market. this index represents large capitalization firms almost equally weighted between value and growth. the vanguard total bond market index fund (vbmfx) was chosen as the proxy to represent fixed income. this fund attempts to track the barclays capital us aggregate float adjusted index. this index is classified as an intermediate-term domestic index. both funds are no-load and low-fee as well as being passively administrated. the funds are highly popular with vanguard investors. the s&p fund is second in size whereas the bond fund is third. vanguard is further known to have very low benchmark tracking errors thereby making these funds valid proxies. for this reason, we also consider other table 1 traditional stock and bond allocations risk tolerance high medium low typical age 25 40 55 bond allocation 5% 20% 35% stock allocation 95% 80% 65% stock:bond ratio 19:1 4:1 13:7 (1.86:1) notes. this table shows the traditional stock and bond allocations assumed for each simulated 10-year period, rebalanced annually. table 2 alternative stock, bond, and inverse or leveraged stock allocation risk tolerance high medium low typical age 25 40 55 bond allocation 4.5% 18.0% 31.5% stock allocation 85.5% 72.0% 58.5% alternative stock allocation 10.0% 10.0% 10.0% stock:bond ratio 19:1 4:1 13:7 (1.86:1) notes. this table shows the alternative stock, bond, and inverse or leveraged stock allocations assumed for each simulated 10-year period, rebalanced annually. 126 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 vanguard stock and bond index funds to enhance our initial findings, based on alternative candidates for an investor’s stock and bond allocation. the inverse and leveraged equity returns were modeled from the daily returns, adjusted for any dividends, of the stock returns, multiplied by negative one (�1) for the inverse stock fund, and positive two (2) for the leveraged fund. annual fees were assumed at 1%, and were deducted from monthly returns, to include the combined cost of the expense ratio, trading commissions, and a bid-ask spread potentially resulting from lower volume etfs.3 although the value of 1% may appear to be insufficient to cover all investor expenses, it was found by starting with the expense ratio for an existing inverse or leveraged etf, such as the proshares short and ultra s&p 500 etfs (symbols sh and sso), which are both 0.90%. we then applied the bid-ask model developed by dilellio and stanley (2011) based on the threemonth moving average volume at the end of 2013 for these etfs, producing a bid-ask spread cost of 0.10% and 0.08%. to account for commissions, we assumed one buy and one sell per year on a portfolio of $100,000 to maintain the 10% allocation to these alternative investments, which produces an annual cost of $20/(10% of $100,000) � 0.2%. lastly, because we are “replicating” an inverse or leverage etf using an existing index mutual fund, we wanted to avoid double-counting the expense of managing the fund, assuming that management was sufficiently covered by the 0.90% expense ratio already identified. thus, we subtract from this total the expense ratio of 0.17% for vfinx, and express the “penalty” as: total replication expense (inverse etf) � 0.90% � 0.10% � 0.20% � 0.17% � 1.03% total replication expense (leveraged etf) � 0.90% � 0.08% � 0.20% � 0.17% � 1.01% no attempt to model divergence from the daily benchmark was made, which can occur when such a fund trades at either a discount or premium. a recent study of inverse fund performance suggests that such a premium or discount remain fairly small, and quickly revert back to the performance modeled here. gerasimos (2011) found that in particular, emphasis is given to the ability of these etfs to meet their daily investment target. in this respect, an average deviation from the daily target amounting to �0.034% is computed. applying a classification to the deviation from the daily return goal, they found that for about 62% of the examined trading period’s duration the return of the average short etf abstains from its target a maximum rate of 0.5%, either below or above the target. the inverse and leveraged equity model used herein appears to be a good starting point for this investigation, but does assume certain risks of leveraged and inverse funds are negligible. in fact, investors are exposed to other risks because of the construction of the inverse equity funds. inverse equity funds replace equity shares with futures and swaps to guarantee the applicable multiples of return. futures have the benefit of having a clearing corporation stated as the counterparty, which has a very favorable credit risk advantage. on the other hand, swaps clear through banks; this adds another element of credit risk. this is a major area of concern not fully understood. in addition, futures also require standard amounts and fixed times to expiration as well as being marked to the market. swaps do not. instead, they are more flexible that accounts for their popularity. choi and elston (2009) reported that proshares inverse s&p 500 etf held weightings of 91% in swaps and but 9% 127j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 in futures. further, cheng and madhavan (2009) noted that daily return streams from paired leveraged and inversed leveraged etfs do not net out on a daily basis. because of the recent development of inverse and leveraged etfs, and uncertainty of return distributions of future markets, the monte carlo simulation was developed. by randomly drawing from actual frequency distributions observed within over two separate 10 year periods, we used an approach similar to cheng and estes (2010) and ervin et al. (2009). each frequency distribution was obtained from different, non-overlapping time periods, thereby making them temporally uncorrelated. these time periods were categorized in terms of overall equity market returns as a “flat” from 2001 to 2010, and “rising” from 1991 to 2000. the descriptive statistics over these periods are summarized in table 3. annualized values were determined by assuming an average of 20 trading days in a month. then, annualized returns are found from the average daily returns multiplied by 20*12 and annualized standard deviations were found from daily standard deviations multiplied by (20*12)1/2. daily returns were found as “adjusted closing price at day t” divided by “adjusted closing price at day t-1” minus 1. a review of table 3 suggests several important behaviors were observed in the time periods considered. first, bond returns were always positive in a fairly narrow annualized return range of �5–7%, and mean daily bond returns were not statistically different between the two 10-year sample periods (p-value � 0.28). bond volatility was also nearly constant, as measured by the annualized standard deviation ranging between 4.39% to 4.49%. the bond returns were in stark contrast to the returns exhibited by stocks over these periods. annualized stock returns varied significantly, between 3.51% and 16.27%, and the mean daily stock returns were statistically different at a p-value � 0.11. the period from 2001 to 2010 also had a noticeable increase in volatility of �6% (annualized) versus the previous decade. the inverse stock investment generated a total return that was negative in all cases, which is consistent with the wealth destruction expected analytically from cheng and madhavan (2009), because of the negative expected return over long holding periods conversely, the leveraged stock fund had a large positive total return in the 1991–2000 case, and a negative total return in the 2001–2010 time period. table 3 return statistics from daily returns over historical 10-year periods return (annualized) standard deviation (annualized) total return bonds 5.22% 4.39% 72.5% stock 3.51% 21.3% 13.9% inverse stock �3.51% 21.3% �45.5% leveraged stock 7.03% 42.6% �19.5% “flat” january 2, 2001 to december 31, 2010, n � 2,516 samples bonds 7.30% 4.49% 113.33% stock 16.27% 14.57% 395.5% inverse stock �16.27% 14.57% �83.88% leveraged stock 32.53% 29.15% 1860.2% “rising” january 2, 1991 to december 29, 2000, n � 2,527 samples notes. this table shows the return statistics from the two 10-year periods observed from adjusted closing prices of stock and bond market proxies. inverse and leveraged stock return statistics are generated assuming perfect replication of negative and double daily return from the stock return, respectively. 128 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 an important aspect of table 3 is the distinction between annualized and total returns. the annualized returns for inverse stocks were always “negative” of the stock returns and the leveraged stock was always 2� the stock returns, but the total returns differ. this behavior is because of the path dependency discussed previously, and will be examined through simulating other paths drawn from the daily return distributions that match historical returns. for the historical periods considered here, the time series of cumulative returns are illustrated in fig. 1 for 2001 to 2010 and fig. 2 for 1991 to 2000. note that the extreme difference in scale between fig. 1 and fig. 2 were intentional so that it is clear that two very different scenarios of equity returns would be simulated. to proceed with a simulation of a portfolio of bonds, stocks, and inverse or leveraged stock funds, two alternatives were considered and one selected to simulate future returns. the first was to take a set of historical returns, fit an appropriate distribution to it, and use this distribution to generate random samples. this article chose not to follow this method, because of concern about the behavior at the tails of the distribution not accurately reflecting observed returns. instead, this article chose an alternate approach that considered a 20-day consecutive return that was available from an observed return history. this method is shown in the appendix. the simulation randomly samples over 2,500 monthly returns derived from actual history to create 120 months of returns. the stock, bond and inverse or leveraged monthly returns use the same random number to preserve the historical correlation of returns between each. the same random sample was also used to determine the representative monthly risk-free rate based on t-bills,5 which is used to determine excess returns and compute the sharpe ratio for each trial, as described in sharpe (1994). fig. 1. cumulative return of bonds, stock, inverse, and leveraged stock funds from january 2001 through december 2010. 129j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 we based our sampling method assuming that the stock market follows a markov (e.g., memoryless) process, and is also related to our assertion about efficient markets, that suggests successive price changes are independent (see fama, 1965). our specific approach of sampling from a historical distribution is often termed “bootstrapping,” which was originally called computer-intensive methods. two excellent references for this methodology are davison and hinkley (1993) and efron and tibshirani (1993). the rationale for randomly sampling supports the assumption that there are no discernible patterns in the returns of stocks and bonds. furthermore, the efficient market hypothesis (emh) states that there are no discernible patterns for returns for stocks and bonds, so consequently, the markets are efficient at all times. many will disagree with that statement. it is not novel, and it has been a contentious subject for decades. there are a few strong believers on both sides with the vast majority falling somewhere in between. however, the fact remains that passive index investors have the better argument. it can be noted that a large percentage of managers failed to outperform their benchmarks over a longer-time horizon. jones and wermers (2011) noted that active returns (adjusted for risk) across managers and time probably average close to zero, net of fees and other expenses, above their benchmark. their conclusion confirmed what should be expected in a mostly efficient market. in such a market, one should expect fierce competition among active managers which drives average (net) active risk-adjusted returns towards zero. they further state that there exists so-called superior active mangers (sams) who should be able to develop better sharpe ratios. thus, investors could be rewarded by exposure to such active strategies by sams. a similar trend has been observed with enhanced index funds (eifs), as shown by chang and krueger fig. 2. cumulative return of bonds, stock, inverse and leveraged stock funds from january 1991 through december 2000. 130 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 (2010). therefore, the controversy will likely continue, but the odds today still favor the efficient market hypothesis. 5. return distributions of terminal wealth the distributions of terminal wealth for the three investors classified previously are shown below in figs. 3 and 4, corresponding to drawing samples from the 2001–2010 and 1991–2000 returns, respectively. within each figure, two distributions for each investor are shown, corresponding to the investor with and without a 10% exposure to the inverse stock fund. the distributions are represented by boxplots that indicate first percentile (p01), 25th percentile (q1), 50th percentile (median), 75th percentile (q3), and 99th percentile (p99) values obtained for terminal wealth, assuming an initial wealth of $1,000. referring to the results in figs. 3 and 4, it appears that for all investors, the distribution of terminal wealth becomes more positively skewed with the addition of the inverse stock fund. the dispersions of the distributions also appear to decrease with the addition of the fig. 3. terminal wealth distributions for investors using 2001–2010 return distributions. distribution assumes $1,000 initial wealth, rebalancing each year, and a 1% annual fee incurred by the inverse stock fund. fig. 4. terminal wealth distributions for investors using 1991–2000 return distributions. distribution assumes $1,000 initial wealth, rebalancing each year, and a 1% annual fee for the inverse stock fund. 131j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 inverse stock fund. these two observations translate to what is expected from knowledge of how the inverse stock fund behaves as a hedge against the stock fund. thus, the average return is decreased and the variance of the return is also decreased. the results from figs. 3 and 4 are also encouraging because they show that the inclusion of the inverse stock fund appears to be simultaneously reducing volatility and return. in contrast, figs. 5 and 6 show the distributions of terminal wealth when an investor includes the leveraged stock fund as part of their asset allocation. unlike what was seen in figs. 3 and 4, figs. 5 and 6 demonstrate the increased volatility that the leveraged fund has on portfolio performance. the results here are less encouraging than what was previously shown, because there does not appear to be a commensurate increase in return for taking the additional risk. this is consistent with trainor and baryla’s (2008) finding that over one year, a 2� leveraged fund produces 2� the standard deviation of returns, but only increase the return by a factor of 1.4. thus, the question remains whether the reduction (increase) in volatility was sufficient enough to justify the reduced (enhanced) returns. tables 4 and 5 provide an answer this question more completely. fig. 5. terminal wealth distributions for investors using 2001–2010 return distributions. distribution assumes $1,000 initial wealth, rebalancing each year, and a 1% annual fee for the leveraged stock fund. fig. 6. terminal wealth distributions for investors using 1991–2000 return distributions. distribution assumes $1,000 initial wealth, rebalancing each year, and a 1% annual fee for the leveraged stock fund. 132 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 6. return statistics for terminal wealth and sharpe ratio to compare the effect of the alternative allocations on terminal wealth and sharpe ratio statistics found for each of the 100,000 trials, we determined a coefficient of variation (cv) for each. the cv is found from the standard deviation divided by the corresponding mean. we propose that a rational investor prefers any alterative allocation that reduces, but keeps positive, the cv for their terminal wealth (cv-tw) and sharpe ratio (cv-s). tables 4 and 5 list these statistics for the simulated periods 2001–2010 and 1991–2000, respectively. referring to the upper panel of table 4 in the row labeled “terminal wealth statistics,” smaller positive cv-tw occur in the alternative allocation to the inverse equity fund than the traditional allocation. this reduction is statistically significant at less than a 0.001 level, based on comparing the reduced values (0.458, 0.371, and 0.289) to a 99.9% confidence interval for the cv-tw defined as 99.9% confidence interval for cv-tw � �tw x�tw � z99.9% �tw �n . here �tw and x�tw are the standard deviation and mean of terminal wealth, z99.9% is the z-statistic corresponding to a 99.9% interval, and n � 100,000. this equation follows the form of a coefficient of variation, but the denominator has been changed from the point estimate to an interval estimate, thereby allowing for direct calculation the confidence interval. we propose that cv-tw that fall outside this region are statistically significant at less than a 0.001 level. for the results in table 4, evaluation this equation yields intervals of [0.573, 0.580], [0.472, 0.476], and [0.378, 0.381] for the 25-year old, 40-year old, and 55-year old investors, respectively. thus, we can make the claim that the cv-tw has been significantly reduced when the inverse fund is used. this reduction suggests that although returns are reduced as shown in a reduction in means, risk as measured by the standard deviation, is reduced to a greater degree. whereas the reduction in cv-tw is encouraging, it is also relevant to note that a similar, but less significant, reduction is possible if the alterative 10% allocation applied to the inverse fund was switched to the risk-free asset. for example, the 25-year old investor’s cv-tw reduces from 0.576 to either 0.458 or 0.516 whether their 10% allocation were using inverse funds versus the risk-free assets. risk free assets also do not reduce the cv-tw as significantly as the inverse funds for 40 and 55-year old investors, with the traditional allocation producing a 0.379 cv-tw for the 55-year old, versus 0.289 and 0.342 for the inverse fund and risk free asset used as the alternative investment. therefore, the 55-year old investor’s cv is reduced by 23.7% when inverse funds are used, versus a 9.8% reduction when risk-free assets are used. unfortunately, the benefit of inverse funds discussed above in the context of reducing the cv-tw is not seen in the sharpe ratio statistics. as can be seen in the row labeled “sharpe ratio statistics” in the upper panel of table 4, the cv-s is always larger when the alternative allocation with inverse funds is used. furthermore, and as expected from the capital market 133j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 t ab le 4 s ta ti st ic s us in g 20 01 –2 01 0 re tu rn di st ri bu ti on s fo r si m ul at io n sa m pl in g t ra di ti on al al lo ca ti on a lt er na ti ve al lo ca ti on w it h in ve rs e fu nd a lt er na ti ve al lo ca ti on w it h ri sk -f re e as se t 25 -y ea r ol d 40 -y ea r ol d 55 -y ea r ol d 25 -y ea r ol d 40 -y ea r ol d 55 -y ea r ol d 25 -y ea r ol d 40 -y ea r ol d 55 -y ea r ol d in ve rs e fu nd (� 1� ) t er m in al w ea lt h st at is ti cs s ta nd ar d de vi at io n $ 77 0 $ 65 8 $ 54 7 $ 56 6 $ 47 5 $ 38 3 $ 68 0 $ 58 2 $ 48 4 m ea n $1 ,3 36 $1 ,3 88 $1 ,4 42 $1 ,2 37 $1 ,2 81 $1 ,3 26 $1 ,3 19 $1 ,3 66 $1 ,4 15 c oe ffi ci en t of va ri at io n 0. 57 6 0. 47 4 0. 37 9 0. 45 8 0. 37 1 0. 28 9 0. 51 6 0. 42 6 0. 34 2 s ha rp e ra ti o st at is ti cs s ta nd ar d de vi at io n 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 m ea n 0. 01 7 0. 02 8 0. 04 3 0. 00 3 0. 01 2 0. 02 6 0. 01 6 0. 02 7 0. 04 2 c oe ffi ci en t of va ri at io n 5. 35 2 3. 35 8 2. 17 9 32 .5 36 7. 82 3 3. 60 0 5. 74 9 3. 51 3 2. 24 6 l ev er ag ed fu nd (2 � ) t er m in al w ea lt h st at is ti cs s ta nd ar d de vi at io n $ 77 1 $ 66 0 $ 54 8 $ 87 7 $ 77 4 $ 67 1 m ea n $1 ,3 32 $1 ,3 85 $1 ,4 39 $1 ,3 40 $1 ,3 87 $1 ,4 36 c oe ffi ci en t of va ri at io n 0. 57 9 0. 47 6 0. 38 1 0. 65 5 0. 55 8 0. 46 7 s ha rp e ra ti o st at is ti cs s ta nd ar d de vi at io n 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 0. 09 3 m ea n 0. 01 7 0. 02 7 0. 04 2 0. 01 6 0. 02 4 0. 03 5 c oe ffi ci en t of va ri at io n 5. 49 6 3. 41 8 2. 20 6 5. 77 8 3. 87 2 2. 68 1 n ot es .t hi s ta bl e sh ow s ce nt ra lt en de nc y an d di sp er si on of te rm in al w ea lt h an d s ha rp e ra ti os fo r ea ch of th e th re e in ve st or s id en ti fi ed .r es ul ts w er e ob ta in ed fr om 10 0, 00 0 tr ia ls , an nu al re ba la nc in g, an d a 1% an nu al fe e (c ha rg ed m on th ly ) fo r th e si m ul at ed in ve rs e or le ve re d st oc k fu nd . 134 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 t ab le 5 s ta ti st ic s us in g 19 91 –2 00 0 re tu rn di st ri bu ti on s fo r si m ul at io n sa m pl in g t ra di ti on al al lo ca ti on a lt er na ti ve al lo ca ti on w it h in ve rs e fu nd a lt er na ti ve al lo ca ti on w it h ri sk -f re e as se t 25 -y ea r ol d 40 -y ea r ol d 55 -y ea r ol d 25 -y ea r ol d 40 -y ea r ol d 55 -y ea r ol d 25 -y ea r ol d 40 -y ea r ol d 55 -y ea r ol d in ve rs e fu nd (� 1� ) t er m in al w ea lt h st at is ti cs s ta nd ar d de vi at io n $1 ,9 32 $1 ,4 56 $1 ,0 71 $1 ,2 19 $ 91 5 $ 66 8 $1 ,5 65 $1 ,1 94 $ 88 9 m ea n $4 ,8 98 $4 ,2 93 $3 ,7 56 $3 ,6 61 $3 ,2 40 $2 ,8 63 $4 ,3 99 $3 ,9 00 $3 ,4 53 c oe ffi ci en t of va ri at io n 0. 39 4 0. 33 9 0. 28 5 0. 33 3 0. 28 3 0. 23 3 0. 35 6 0. 30 6 0. 25 7 s ha rp e ra ti o st at is ti cs s ta nd ar d de vi at io n 0. 09 1 0. 09 1 0. 09 2 0. 09 2 0. 09 2 0. 09 2 0. 09 1 0. 09 1 0. 09 1 m ea n 0. 27 9 0. 28 6 0. 29 5 0. 25 1 0. 25 4 0. 25 6 0. 27 9 0. 28 6 0. 29 5 c oe ffi ci en t of va ri at io n 0. 32 8 0. 32 0 0. 31 1 0. 36 7 0. 36 3 0. 36 1 0. 32 7 0. 31 9 0. 31 0 l ev er ag ed fu nd (2 � ) s ta nd ar d de vi at io n $1 ,9 29 $1 ,4 54 $1 ,0 70 $2 ,5 88 $2 ,0 49 $1 ,5 98 m ea n $4 ,8 96 $4 ,2 90 $3 ,7 53 $5 ,7 65 $5 ,1 30 $4 ,5 58 c oe ffi ci en t of va ri at io n 0. 39 4 0. 33 9 0. 28 5 0. 44 9 0. 39 9 0. 35 1 s ha rp e ra ti o st at is ti cs s ta nd ar d de vi at io n 0. 09 1 0. 09 1 0. 09 1 0. 09 1 0. 09 1 0. 09 1 m ea n 0. 27 8 0. 28 6 0. 29 4 0. 28 3 0. 29 0 0. 29 7 c oe ffi ci en t of va ri at io n 0. 32 8 0. 32 0 0. 31 1 0. 32 2 0. 31 4 0. 30 6 n ot es .t hi s ta bl e sh ow s ce nt ra lt en de nc y an d di sp er si on of te rm in al w ea lt h an d s ha rp e ra ti os fo r ea ch of th e th re e in ve st or s id en ti fi ed .r es ul ts w er e ob ta in ed fr om 10 0, 00 0 tr ia ls , an nu al re ba la nc in g, an d a 1% an nu al fe e (c ha rg ed m on th ly ) fo r th e si m ul at ed in ve rs e or le ve re d st oc k fu nd . 135j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 line, the use of the risk-free asset in the alternative allocation has no effect on the sharpe ratio mean or standard deviation.6 as shown in the lower panel of table 4, there is no diversification benefit from using a leveraged stock fund. here, we reran the 100,000 trials in our simulation, which produced a small differences in the fourth significant digit, between the lower and upper panels for the “traditional allocation.” in all cases, the cv-tw and cv-s increase when the leveraged stock fund is included as the alternative asset. this result is not surprising, because this time period sampled (2001–2010) did not have rising equity prices. for completeness, we also tested levels of statistical significance between the top and bottom panels in table 4 using the confidence intervals estimated above. recalling our 99% confidence intervals cited previously for the traditional allocation, we note that the cv-tw and cv-s in the bottom panel labeled “traditional allocation” all fall within these intervals, which is expected since it represents simply another random draw of 100,000 trials in the simulation. the results so far have suggested that the inverse stock fund may be a useful addition to a diversified portfolio of stocks and bonds, because the cv-tw was consistently reduced, to varying degrees. this benefit always exceeded the benefit an investor could realize if a similar allocation was made to a risk-free asset. however, one may argue the contrary if attention is solely given to the sharpe ratio statistics.7 additionally, the use of the leveraged stock in the alternative allocation does not improve diversification in this “flat” equity case from 2001 to 2010. table 5 provides the alternative perspective on inverse and leveraged stock funds as an alternative investment in “rising” equity markets, such as those seen that in the period from 1991 to 2000. intuitively, one may suspect that the use of an inverse stock fund in rising equity markets should be avoided. indeed, the upper panel of table 5 in the row labeled “terminal wealth statistics” shows that the mean terminal wealth is significantly reduced, but so is the standard deviation. in fact, the cv-tw is still significantly reduced, but to less of a degree then what was seen in table 4. these reduced values all fall below the 99.9% confidence intervals for cv-tw in table 5, which are [0.393, 0.396], [0.338, 0.340], and [0.284, 0.286] for the 25-year old, 40-year old, and 55-year old investors, respectively. these findings suggest that the cv-tw can be improved when long-term equity prices are either rising or flat. unfortunately, the benefits of allocations with inverse stock funds still do not reduce the cv-s. referring to the upper panel of table 5 in the row labeled “sharpe ratio statistics,” we see the same relationships observed in table 4. that is, the cv-s does not improve with the inclusion of the inverse fund as an alternative asset. lastly, and somewhat surprisingly over a period of rising equity prices, there is a no diversification benefit in the cv-tw from using a leveraged etf, as shown in the lower panel of table 5. in all cases, the cv-tw increases, suggesting that the leveraged stock fund provides no risk-return benefit. this finding may be counterintuitive, when considering the positive upward trend of equities in the 1991–2000 time period, but is partly reconciled when reviewing the lower panel of table 5 labeled “sharpe ratio statistics.” here, the cv-s shows a marginal risk-adjusted improvement from the use of the leveraged stock fund in a rising equity market, because these values all fall slightly below the 99.9% confidence 136 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 intervals for the cv-s in table 5 of [0.325, 0.327], [0.317, 0.319], and [0.311, 0.313]. thus, we see that there is a benefit to an alternative allocation using leveraged equity funds, provided an investor relies on sharpe ratio, not terminal wealth statistic, in rising equity markets. again for completeness, we also tested levels of statistical significance between the top and bottom panels in table 5 using the confidence intervals estimated above. recalling our 99% confidence intervals cited previously for the traditional allocation, we again note that the cv-tw and cv-s in the bottom panel labeled “traditional allocation” all fall within these intervals. recall that this is expected since it represents another random draw of 100,000 trials in the simulation. 7. enhancing the study by utilizing other stock and bond funds one criticism that could be raised by the results shown so far is that it only incorporated one specific stock index fund (vfinx) and one specific bond index fund (vbmfx). thus, the results shown above may not be generally applicable to other funds utilized by individual investors and financial planners, even if these stock and bond funds are major players in their respective investment categories. therefore, to strengthen the findings from the previous section, we identified additional stock and bond index funds to see if the results shown previously still hold. for stock funds, we identified four potential stock indices to compliment and extend the findings previously shown that were based on the s&p 500 index fund, which is a large capitalization stock fund. the four candidate stock funds represented mid and small capitalization stocks, along with the nasdaq and dow jones industrials indices. although many other stock index funds could be considered, such as international developed and emerging markets, we chose these because of their wide familiarity to individual investors and financial planners who we believe would treat them as likely candidates for their stock investment. we also identified two additional bond funds to extend the midterm maturity high quality corporate and u.s. government bond fund examined previously. these new bond funds provide short and long-term maturities. in terms of an expanded set of bond funds, although we recognized that there are potentially many other bond funds from which to select, we believe these carry risks that may prevent many bond investors from considering them as their primary bond investment. these alternatives included junk, international, and emerging market bonds. lastly, we chose not to include municipal bonds because we are interested in investments in a tax deferred or tax exempt account, such as an ira or 401k. a summary of the original stock and bond funds, which appears on the first row, along with the expanded set of funds to be analyzed, appear in tables 6 and 7. including these additional stock and bond index funds represented a challenge, as our simulation approach required daily returns starting no later than january 1, 1991. for some of these funds, the inception dates noted in tables 6 and 7 are many years after this date. consequently, the updated simulation results include the small and midcap results in both 10-year time periods (1991–2000 and 2001–2010), but our nasdaq, short term, and long term bond index mutual funds could only support simulated results for the 10-year time period of 2001–2010 because of their later inception date. we also did not to include the dow 137j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 industrials index as no appropriate index fund could be found. we did consider the use of the dow 30 etf index fund (spdr dow jones industrial average index etf, symbol dia), which became available on january 13, 1998. however, its high degree of similarity measured by correlation from 2001 to 2010 daily returns against vfinx was 0.96, suggesting it would generate results that would be very similar to those already provided. our modeling approach to develop the inverse and leveraged etf returns followed the approach used in section 4. tables 8 and 9 summarize the total replication expense, which in some cases, differed from our previous assumption of a 1% cost. thus, the results in this section were found by using the total replication expense determined in the last column of these tables. to begin understanding the potential of the expanded set of funds, descriptive statistics are once again found and appear in table 10. the top panel of table 10 shows the return statistics for the period of “flat” equity prices of 2001–2010, while the bottom is for “rising” equity prices of 1991–2000. over the period of 2001–2010, longer term bonds provided a higher return but also a higher volatility. midcaps and small-cap funds also generated higher returns than the s&p 500 large-cap fund (reported in table 3 at 3.51%) with a marginally higher volatility, whereas the nasdaq fund returned nearly the same as the large-cap fund, but at nearly three times the volatility. the results over the period of 1991–2000, a period of rising equity prices, showed less difference between holding a large-cap stock versus holding either small or midcap stocks. annualized returns between large, mid, and small cap stocks are all within 1% of each other from 1991 to 2000, and annualized volatilities are within 2% of each other. when considering the effect of holding the inverse funds over the both 10-year time periods, the annualized returns were all negative and slightly larger in magnitude than their long-index counterparts. similarly, the annualized returns of holding the leveraged funds were all slightly less than double. these results are consistent with the higher expenses table 6 previously assumed stock fund and additional stock index funds utilized to enhance the study index stock fund selected symbol inception date s&p 500 vanguard 500 index mutual fund vfinx august 31, 1976 mid-caps vanguard extended market index fund vexmx december 21, 1987 small-caps vanguard small cap index, investor class naesx october 10, 1960 nasdaq usaa nasdaq-100 index usnqx november 9, 2000 notes. the first row of this table lists the fund used in the previous sections. the later rows show the additional funds utilized in this section. table 7 previously assumed bond fund and additional bond index funds utilized to enhance the study index bond index fund symbol inception date total bond market vanguard total bond market index vbmfx december 10, 1986 short-term bond market vanguard short-term bond index vbisx february 28, 1994 long-term bond market vanguard long-term bond index vbltx february 28, 1994 notes. the first row of this table lists the fund used in the previous sections. the later rows show the additional funds utilized in this section. 138 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 associated with these funds, as well as the effect volatility can have on expected returns of these types of investments over longer periods (see lu et al., 2012). similarly, the leveraged funds over each of these two time periods were also consistent with the longer period expectation of providing slightly less than two-times the annualized return of their longindex counterpart, while doubling the volatility. another notable result from table 10 is that positive annualized returns generally yielded positive total returns and negative annualized returns produced negative total returns. however, because the annualized returns were obtained from arithmetically averaged daily returns, there is a notable exception to this relationship when the average daily returns is very small, as in the case of nasdaq that had a negative cumulative return because of the dot-com crash in the early 2000s. lastly, although there was statistical significance between the s&p 500 returns between 1991 and 2000 and 2001–2010 at the 0.11 level, there was no statistical difference in the mid and small-cap returns (p-values � 0.41 and 0.53, respectively). 7.1. results for midcaps, small caps, and nasdaq index funds tables 11 and 12 provide results for terminal wealth and sharpe ratio statistics obtained from a 100,000 trials using the midcap fund identified in table 6, and simulating the inverse and leveraged version of it. during the period of “flat” equity prices from 2001 to 2010 shown in table 11, the use of the inverse midcap fund, shown in the upper panel, is consistent with the findings in the previous section when using large cap funds were assumed. specifically, the cv-tw is reduced below the 99.9% confidence intervals of [0.725, 0.736], [0.596, 0.603], and [0.476, 0.481] for the 25-year old, 40-year old, and 55-year old investors, when the inverse midcap fund is used, indicating statistical signifitable 8 selected proshares short (�1�) etf products etf symbol fund name expense ratio estimated bid-ask spread commissions long index expense ratio total replication expense sh short s&p 500 0.90% 0.10% 0.2% �0.17% 1.03% myy short mid-cap400 0.95% 0.37% 0.2% �0.24% 1.28% rwm short russell 2000 0.95% 0.13% 0.2% �0.24% 1.04% psq short qqq 0.95% 0.19% 0.2% �0.64% 0.7% table 9 selected proshares leveraged (2�) etf products etf symbol fund name expense ratio estimated bid-ask spread commissions long index expense ratio total replication expense (annual) sso ultra s&p 500 0.90% 0.08% 0.2% �0.17% 1.01% mvv ultra midcap400 0.95% 0.16% 0.2% �0.24% 1.07% uwm ultra russell 2000 0.95% 0.15% 0.2% �0.24% 1.06% qld ultra qqq 0.95% 0.11% 0.2% �0.64% 0.62% 139j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 cance at less than 0.001 level. the use of the risk free asset also produces a cv-tw that is lower than this interval, but is less statistically significant. also like shown previously, cv-s does not corroborate this finding, as the use of leveraged midcap fund in this period of “flat” equity prices does not reduce the cv-s relative to the traditional allocation case. table 12 corresponds to a period of “rising” equity prices from 1991 to 2000, and results are similar to the previous results when the large cap fund. again, cv-tw is significantly reduced when the inverse fund is used instead of the traditional allocation, since each investor’s cv-tw is outside the 99.9% confidence interval obtained for the traditional allocation of [0.565, 0.571], [0.476, 0.481], and [0.390, 0.393]. a similar, but less statistically significant reduction occurs when the risk-free asset is used in the alternative allocation. the cv-s is not reduced when the inverse fund is used, which is consistent with previous results. as seen previously, this period of rising equity prices does not show any benefit of the use of the leveraged fund based on the cv-tw. again, the use of the leveraged fund provides an improvement to the cv-s over the traditional allocation, based on the values of 0.474, 0.459, and 0.441 all lower than the lower bound on the 99.9% confidence interval for cv-s of [0.487, 0.492], [0.469, 0.473], and [0.442, 0.447]. tables 13 and 14 provide results for terminal wealth and sharpe ratio statistics obtained from a 100,000 trials using a small-cap fund. the period of “flat” equity prices from 2001 to 2010 shown in table 13 indicates the previous results hold, where the use of the inverse fund reduces the cv-tw below the 99.9% confidence interval for the traditional allocation of [0.747, 0.759], [0.617, 0.625], and [0.494, 0.499] for the 25-year old, 40-year old, and table 10 return statistics from daily returns over historical 10-year periods return (annualized) standard deviation (annualized) total return bonds (short term) 4.27% 2.88% 55.8% bonds (long term) 7.00% 9.88% 97.9% mid-cap 8.31% 23.5% 78.9% inverse mid-cap �9.59% 23.5% �72.6% leveraged mid-cap 15.6% 47.0% 59.4% small-cap 9.68% 24.6% 100.9% inverse small-cap �10.7% 24.6% �76.3% leveraged small-cap 18.30% 49.1% 90.7% nasdaq 3.55% 30.2% �9.76% inverse nasdaq �4.25% 30.2% �60.4% leveraged nasdaq 6.48% 60.3% �70.4% “flat” january 2, 2001 to december 31, 2010, n � 2515 samples mid-cap 15.6% 16.1% 349.4% inverse mid-cap �16.9% 16.1% �85.2% leveraged mid-cap 30.1% 32.3% 1267.6% small-cap 15.3% 14.6% 346.7% inverse small-cap �16.3% 14.6% �84.0% leveraged small-cap 29.5% 29.2% 1322.0% “rising” january 2, 1991 to december 29, 2000, n � 2527 samples notes. this table shows the return statistics from the two 10-year periods observed from adjusted closing prices of stock and bond market proxies. inverse and leveraged stock return statistics are generated assuming perfect replication of negative and double daily return from the stock return, respectively, along with incurring an average daily expense as shown in tables 8 and 9. 140 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 55-year old, respectively. a similar, but less significant reduction occurs if the risk-free asset is used. furthermore, these results are still not corroborated by looking at the cv-s value. there is also still no evidence that the leveraged fund provides any benefit during “flat” equity prices, as both the cv-tw and cv-s show increases. in the period of “rising” equity prices from 1991 to 2000 shown in table 14, the results for the terminal wealth statistics are similar to before, with the inverse of the small cap fund significantly reducing the cv-tw relative to the traditional allocation’s 99.9% confidence intervals of [0.555, 0.561], [0.467, 0.472], and [0.382, 0.385]. this is also true, but less significant, when the cv-tw under the risk-free asset is evaluated. and again, like the large and mid caps, the cv-s does not show any improvement when the inverse small cap fund is used. the use of the leveraged small cap index fund produces results similar to before, where a significant improvement in cv-s occurs against the traditional 99.9% confidence intervals of [0.482, 0.487], [0.463, 0.467], and [0.439, 0.443]. table 15 provides the last extension to alternative equity funds and the use of inverse and leveraged funds from them in an alternative allocation. here, which covers the period of “flat” equity prices from 2001 to 2010, all the findings shown previously continue to hold. the cv-tw, with the use of the inverse nasdaq index fund, is still reduced when compared to the traditional allocation, based on 99.9% confidence intervals of [0.928, 0.946], [0.745, 0.757], and [0.583, 0.590], and the reduction is more significant than if the risk-free asset table 11 statistics using 2001–2010 return distributions for simulation sampling, mid-cap index fund (vexmx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $1,597 $1,260 $ 967 $1,086 $ 850 $ 639 mean $2,187 $2,102 $2,021 $1,857 $1,791 $1,727 cv-tw 0.730 0.600 0.478 0.585 0.474 0.370 sharpe ratio statistics standard deviation 0.095 0.095 0.096 0.095 0.096 0.096 mean 0.084 0.092 0.104 0.073 0.080 0.091 cv-s 1.129 1.032 0.918 1.310 1.195 1.054 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $1,349 $1,073 $ 829 $1,949 $1,602 $1,296 mean $2,067 $1,994 $1,924 $2,332 $2,252 $2,174 cv-tw 0.653 0.538 0.431 0.836 0.712 0.596 sharpe ratio statistics standard deviation 0.096 0.096 0.096 0.095 0.095 0.095 mean 0.084 0.092 0.104 0.084 0.090 0.099 cv-s 1.140 1.041 0.925 1.131 1.053 0.964 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 141j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 were used. unfortunately, the sharpe ratio statistics do not support a reduction in cv-s when the inverse fund is used. further, the use of the leveraged nasdaq index fund provides no benefit in terms of cv-tw or cv-s, which is similar to the results shown for other equity funds during the period of “flat” equity prices from 2001 to 2010. 7.2. results for short-term and long term bond funds to complete our evaluation of alternative funds, we considered two variations of bond holdings beyond the vtbmi fund. these two funds represented shorter and longer maturity bond indices, and we evaluated each as its own case when the vtsmi fund was used along with an alternative allocation to either inverse of the stock fund, 2� leverage of the stock fund, or a risk-free asset alternative. the results for this variation in bond funds appear in tables 16 and 17, and confirm what was seen previously. once again, use of the inverse stock fund provides a diversification benefit by reducing the cv-tw beyond the 99.9% confidence intervals of [0.578, 0.585], [0.474, 0.479], and [0.377, 0.380] for the short-term bond fund index in table 16, with a similar, but less significant reduction in the cv-tw when the risk-free asset is used. again, there appears to be no benefit in terms of cv for the sharpe ratios. lastly, the use of the leveraged fund does not provide any benefit in terms of either terminal wealth or sharpe ratio, which is again consistent with the 2001–2010 period with “flat” equity prices already reported. table 12 statistics using 1991–2000 return distributions for simulation sampling, mid-cap index fund (vexmx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $2,705 $2,004 $1,441 $1,705 $1,259 $ 896 mean $4,762 $4,191 $3,683 $3,597 $3,196 $2,836 cv-tw 0.568 0.478 0.391 0.474 0.394 0.316 sharpe ratio statistics standard deviation 0.097 0.097 0.096 0.097 0.097 0.096 mean 0.198 0.206 0.216 0.179 0.183 0.190 cv-s 0.488 0.470 0.447 0.541 0.527 0.508 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $2,181 $1,641 $1,198 $3,672 $2,868 $2,203 mean $4,284 $3,814 $3,391 $5,654 $5,051 $4,506 cv-tw 0.509 0.430 0.353 0.649 0.568 0.489 sharpe ratio statistics standard deviation 0.097 0.097 0.097 0.096 0.096 0.096 mean 0.198 0.205 0.216 0.203 0.210 0.218 cv-s 0.490 0.472 0.450 0.474 0.459 0.441 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 142 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 similar findings appear in table 17, where the 99.9% confidence intervals for cv-tw are [0.578, 0.585], [0.479, 0.484], and [0.395, 0.398] for the 25-year old, 40-year old, and 55-year old investors, respectively. there is a significant decrease in the cv-tw when the inverse fund is used, and to a lesser degree when the risk-free asset is included. no benefits could be found in terms of the cv from the sharpe ratio statistics. 8. conclusions this article investigated the risks and possible opportunities of a “120 – age” annual reallocation strategy of stocks and bonds, but also included a small allocation of inverse or leveraged stock fund. the assessment was based on three different risk aversion levels, and simulated inverse and leveraged equity fund returns that included expenses and fees. we also included several variations of alternative stock and bond funds that might be selected by individual investors and financial planners in their asset allocation decision. terminal wealth and sharpe ratio statistics were obtained by monte carlo simulation that selected from return histories using two unique historical periods, representing both “flat” and “rising” cases of long-term equity returns, rather than assuming prices followed standard geometric brownian motion. table 13. statistics using 2001–2010 return distributions for simulation sampling, small-cap index fund (naesx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $1,912 $1,480 $1,112 $1,277 $ 981 $ 725 mean $2,538 $2,384 $2,239 $2,100 $1,984 $1,872 cv-tw 0.753 0.621 0.497 0.608 0.495 0.387 sharpe ratio statistics standard deviation 0.096 0.096 0.096 0.096 0.096 0.096 mean 0.102 0.110 0.121 0.091 0.098 0.108 cv-s 0.939 0.874 0.795 1.056 0.984 0.892 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $1,594 $1,245 $ 944 $2,374 $1,918 $1,524 mean $2,370 $2,239 $2,115 $2,738 $2,592 $2,452 cv-tw 0.673 0.556 0.446 0.867 0.740 0.621 sharpe ratio statistics standard deviation 0.096 0.096 0.097 0.096 0.096 0.096 mean 0.102 0.110 0.121 0.101 0.107 0.115 cv-s 0.941 0.876 0.797 0.945 0.893 0.832 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 143j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 the findings, developed under various stock and bond alternatives, showed that under both “flat” and “rising” return histories, annual rebalancing to include a 10% allocation towards an inverse stock fund provides a diversification benefit by reducing the cv of terminal wealth. the diversification benefit is strengthened when stock returns are “flat.” further, it always exceeded the benefit of simply using a risk-free asset. unfortunately, the benefit observed by measuring the cv of terminal wealth was not corroborated with sharpe ratio statistics. additionally, leveraged etfs were never found to provide a risk-reward benefit preferred by rational investors interested in terminal wealth, but sharpe ratio statistics did show a benefit in rising equity markets. these results, and the potential diversification benefits of inverse and leveraged etfs, call into question the current recommendation that they are always detrimental to long-term investors, and only beneficial for short-term trading. although these findings suggest that additional analytical and empirical studies may be warranted to better assess their impact on asset allocation decisions made by individual investors and financial planners, both must consider the addition of inverse or leverage stock funds to improve wealth maximization under appropriate risk-reward tradeoff. the source of this somewhat unexpected behavior likely lies in a return history that is not normally distributed in the observed distribution’s tails, and thus may influence long-range plans that are important to individual investors and financial planners. future work in this area can cover several areas. first, a more comprehensive set of cases table 14 statistics using 1991–2000 return distributions for simulation sampling, small-cap index fund (naesx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standarddeviation $2,556 $1,902 $1,373 $1,626 $1,205 $ 860 mean $4,579 $4,053 $3,582 $3,494 $3,121 $2,785 cv-tw 0.558 0.469 0.383 0.465 0.386 0.309 sharpe ratio statistics standard deviation 0.093 0.093 0.093 0.094 0.094 0.094 mean 0.192 0.200 0.211 0.174 0.179 0.186 cv-s 0.486 0.466 0.441 0.541 0.524 0.502 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $2,076 $1,565 $1,145 $3,495 $2,741 $2,114 mean $4,125 $3,692 $3,300 $5,424 $4,870 $4,367 cv-tw 0.503 0.424 0.347 0.644 0.563 0.484 sharpe ratio statistics standard deviation 0.094 0.094 0.093 0.094 0.094 0.093 mean 0.192 0.200 0.211 0.198 0.204 0.213 cv-s 0.489 0.469 0.444 0.475 0.458 0.438 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 144 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 could be considered. for example, evaluating the effects of “3�” leveraged and “�2�/ �3�” inverse etfs is suggested to see if the observations found using “�1�” inverse and “2�” leveraged etfs considered here persist. like giese (2010), we also observe that the optimal leverage strongly depends on the prevailing market conditions, which were not exhaustively considered here. further, alternative simulation approaches could be used, such as simulating the price path of stocks and bonds with variations in their stochastic parameters. lastly, other measures of diversification could be examined, such as the diversification effect as studied by hight (2010). acknowledgments the authors wish to thank the participants at the academy of financial services 2011 annual meeting in las vegas, nevada, who provided useful feedback on an earlier version of this work. specifically, we wish to acknowledge brian boscaljon for his insightful suggestion to include risk-free assets as an alternative asset for comparison. lastly, although every attempt was made to eliminate any errors from this article, any remaining issues are solely the responsibility of the authors. table 15 statistics using 2001–2010 return distributions for simulation sampling, nasdaq index fund (usnqx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $1,280 $1,063 $ 859 $ 923 $ 758 $ 598 mean $1,366 $1,415 $1,465 $1,269 $1,310 $1,351 cv-tw 0.937 0.752 0.587 0.728 0.579 0.443 sharpe ratio statistics standard deviation 0.092 0.092 0.092 0.093 0.093 0.092 mean 0.014 0.021 0.031 0.004 0.010 0.019 cv-s 6.572 4.451 3.004 22.837 9.266 4.783 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $1,135 $ 945 $ 764 $1,536 $1,310 $1,101 mean $1,363 $1,404 $1,446 $1,387 $1,429 $1,473 cv-tw 0.833 0.673 0.528 1.107 0.916 0.748 sharpe ratio statistics standard deviation 0.093 0.093 0.092 0.092 0.092 0.092 mean 0.014 0.021 0.031 0.014 0.019 0.026 cv-s 6.558 4.442 3.000 6.827 4.960 3.604 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 145j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 notes 1 fidelity noted that because of rebalancing and other risks, leverage and inverse leveraged etfs are intended as short term trading vehicles for sophisticated investors actively monitoring their portfolios on a daily basis. source: http://personal.fidelity. com/research/etf/content/leveraged_etn_etf.shtml. 2 the u.s. securities and exchange commission (2011), while not stating the rule of 120, implied it indirectly in their suggestion to for investors to consider target date funds (tdf) in ‘beginners’ guide to asset allocation, diversification, and rebalancing.” source: http://www.sec.gov/investor/pubs/assetallocation.htm. 3 for example, see http://www.proshares.com/funds/sh.html. 4 the correlation between stocks and bonds were as follows: (1991–2000, 20012010) � 0.047 and �0.255. 5 the authors wish to thank ken french from making daily risk-free rates available at his web site. source: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data _library.html. 6 the small variation seen in the sharpe ratio means at the third decimal place is because of limiting the number of trials to 100,000. 7 an interesting but problematic issue occurs when simulating time periods that have “falling” equity prices, such as observed over 1999 to 2008. this case was not reported table 16 statistics using 2001–2010 return distributions for simulation sampling, short-term bond index fund (vbisx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $ 770 $ 647 $ 527 $ 565 $ 467 $ 369 mean $1,324 $1,358 $1,392 $1,227 $1,256 $1,284 cv-tw 0.581 0.477 0.379 0.461 0.372 0.287 sharpe ratio statistics standard deviation 0.093 0.093 0.093 0.093 0.093 0.093 mean 0.016 0.023 0.034 0.001 0.007 0.015 cv-s 5.852 4.013 2.737 76.871 13.945 6.084 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $ 676 $ 570 $ 465 $ 873 $ 760 $ 649 mean $1,314 $1,344 $1,375 $1,329 $1,360 $1,391 cv-tw 0.515 0.424 0.338 0.657 0.559 0.466 sharpe ratio statistics standard deviation 0.093 0.093 0.093 0.093 0.093 0.093 mean 0.016 0.023 0.034 0.015 0.020 0.028 cv-s 5.951 4.050 2.750 6.296 4.573 3.329 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 146 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 because it violated our desire to use temporally uncorrelated historical periods. nevertheless, when the simulation drew from this time period, a negative mean sharpe ratio was found, making interpreting a cv of sharpe ratio statistics problematic because both a small standard deviation or large negative mean of the sharpe ratio could lead to a large negative cv of the sharpe ratio. thus, the negative cv value for the sharpe ratio statistics for the “falling” equity price case are not as easily reconcilable to preference for rational investors. 8 this fund was acquired by vanguard. its actual inception date is in 1960, but began operating as a passive low-cost index fund in 1989. source: http://socialize.morning star.com/newsocialize/forums/p/94965/94965.aspx#94965. table 17 statistics using 2001–2010 return distributions for simulation sampling, long-term bond index fund (vbltx) traditional allocation alternative allocation with inverse (�1�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $ 781 $ 690 $ 605 $ 573 $ 497 $ 426 mean $1,343 $1,432 $1,526 $1,243 $1,317 $1,395 cv-tw 0.581 0.482 0.397 0.461 0.377 0.306 sharpe ratio statistics standard deviation 0.093 0.093 0.094 0.093 0.093 0.094 mean 0.018 0.034 0.055 0.004 0.019 0.042 cv-s 5.096 2.745 1.694 23.859 4.827 2.263 alternative allocation with risk-free asset alternative allocation with leveraged (2�) fund 25-year old 40-year old 55-year old 25-year old 40-year old 55-year old terminal wealth statistics standard deviation $ 686 $ 606 $ 530 $ 882 $ 801 $ 722 mean $1,331 $1,410 $1,493 $1,349 $1,429 $1,514 cv-tw 0.515 0.430 0.355 0.654 0.560 0.477 sharpe ratio statistics standard deviation 0.093 0.093 0.094 0.093 0.093 0.093 mean 0.018 0.034 0.056 0.017 0.029 0.045 cv-s 5.084 2.742 1.693 5.420 3.162 2.060 notes. this table shows central tendency and dispersion of terminal wealth and sharpe ratios for each of the three investors identified. results were obtained from 100,000 trials, annual rebalancing, and an annual fee as shown in the last column of tables 8 and 9 for the simulated inverse or leveraged stock fund. 147j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 references ammermann, p., runyon, l., & conceicao, r. (2011). a new quantitative approach for the management of a student-managed investment fund. managerial finance, 37, 624–635. arshanapalli, b. g., & nelson, w. b. (2010). yes virginia, diversification is still a free lunch. journal of wealth management, 13, 34–40. barnhorst, b. c., & cocozza, c. r. (2010). inverse and leveraged etfs: considering the alternatives. journal of financial planning, 24, 44–49. chang, c. e., & krueger, t. (2010). do enhanced index funds live up to their name? financial services review, 19, 145–162. chen, h., & estes, j. (2010). a monte carlo study of the strategies for 401(k) plans: dollar-cost-averaging, value-averaging, and proportional rebalancing. financial services review, 19, 95–109. cheng, m., & madhavan, a. (2009). the dynamics of leveraged and inverse exchange traded funds. journal of investment management, 7, 43–62. choi, d., & elston, f. (2009). inverse etfs. in proceedings of the academy of accounting and financial studies international conference in new orleans, louisiana. allied academies international conference. retrieved at alliedacademies.org/public/proceedings/proceedings24/aafs%20proceedings.pdf davison, a. c., & hinkley, d. v. (1993). bootstrap methods and their application. new york, ny: cambridge series in statistical and probabilistic mathematics. dilellio, j. a., & stanley, d. j. (2011). etf trading strategies to enhance client wealth maximization. financial services review, 20, 145–163. efron, b., & tibshirani, r. (1993). an introduction to the bootstrap. boca raton, fl: chapman and hall/crc. ervin, d., faulk, g., & smolira, j. (2009). the impact of asset allocation, savings, and retirement horizon, saving rates, and social security income in retirement planning: a monte carlo analysis. financial services review, 18, 313–331. fama, e. (1965). random walks in stock-market prices. selected papers no. 16, booth school of economics, university of chicago, p. 7. retrieved at http://www.chicagobooth.edu/�/media/34f68ffd9cc04ef1a 76901f6c61c0a76.pdf gerasimos g. r. (2011). a study on the third-generated exchange traded funds: the case of short etfs. the journal of index investing, 2, 25–43. giese, g. (2010). on the risk-return profile of leveraged and inverse etfs. journal of asset management, 11, 219–228. guedj, i., guohua, l., & mccann, c.. (2010). leveraged and inverse etfs, holding periods, and investment shortfalls. journal of index investing, 1, 45–57. appendix step simulation process 1 assign a numerical index of 1 to n for each of the daily returns observed in the historical 10-year period. 2 for each day, determine the cumulative return up to and including the previous 20 days, so that the accumulated returns from these days are representative of the distribution of monthly returns. 3 generate 120 random numbers ranging from 1 to n. 4 select returns from the distribution of representative monthly returns determined in step 2 using the random numbers found in step 3, and generate 10 years of monthly returns. use the same set of random numbers for each trial to select monthly returns from each of investment category, ensuring that the historical correlation among assets is preserved.4 5 repeat for 100,000 trials, collecting terminal wealth and sharpe ratio for each trial. 6 determine mean, standard deviation, and coefficient of variation in terminal wealth and sharpe ratios found from each of the 100,000 trials generated in step 5. 148 j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 hight, g. n. (2010). diversification effect: isolating the effect of correlation on portfolio risk. journal of financial planning, 23, 54–61. johnston, k., hatem, j., & scott, e. (2013). individual investors: asset allocation vs. portfolio insurance (puts or calls). financial services review, 22, 291–310. jones, r. c., & wermers, r. (2011). active management in mostly efficient markets. financial analysts journal, 67, 29–45. lu, l., wang, j., & zhang, ge. (2012). long term performance of leveraged etfs. financial services review, 21, 63–80. mulvey, j. m., ural, c., & zhang, z. (2007). improving performance for long-term investors: wide diversification, leverage, and overlay strategies. quantitative finance, 7, 175–187. pfau, w. d. (2010). lifecycle funds and wealth accumulation for retirement: evidence for a more conservative asset allocation as retirement approaches. financial services review, 19, 59–74. sharpe, w. f. (1994). the sharpe ratio. journal of portfolio management, 21, 49–59. trainor, w. j., & baryla, e. a. (2008). leveraged etfs: a risky double that doesn’t multiply by two. journal of financial planning, 21, 48–55. 149j.a. dilellio, r. hesse, d.j. stanley / financial services review 23 (2014) 123–149 exploring financial behaviors of military households: do financial knowledge and financial education matter? melissa j. wilmartha, kyoung tae kimb,*, robin henagerc adepartment of consumer sciences, university of alabama, 316-a adams hall, box 870158, tuscaloosa, al 35487, usa bdepartment of consumer sciences, university of alabama, 316-c adams hall, box 870158, tuscaloosa, al 35487, usa cschool of business, whitworth university, 300 w. hawhorne road, spokane, wa 99251, usa abstract this study explores short-term and long-term financial behaviors of military and civilian households in the united states. we investigate the role of financial knowledge and financial education on financial behaviors. using the 2018 national financial capability study (nfcs), results indicated military households had higher financial knowledge scores, greater receipt of financial education, and higher financial behaviors. multivariate analyses show that objective and subjective financial knowledge were associated positively with short-term and long-term financial behaviors of military and civilian households. experiencing financial education was positively associated with the longterm behaviors of military households. this study provides insights for policymakers and financial practitioners. © 2023 academy of financial services. all rights reserved. jel classifications: g510; g530 keywords: financial education; financial knowledge; financial behaviors; military households; national financial capability study 1. introduction despite the attention of both the popular press and the government on the topic of finances in military households, there is still limited empirical research in this area (carlson et al., *corresponding author: tel.: +1-205-348-9167, fax: +1-205-348-8721. e-mail address: ktkim@ches.ua.edu 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 35–54 2015; skimmyhorn, 2016a). developing a better understanding of the financial behaviors of military households is an important topic as it may facilitate ways to help military personnel reduce stress and strain (luther et al., 1998). commonly reported among all types of american households is that money, work, and the economy are very or somewhat significant sources of stress in their lives (american psychological association, 2017). given the great recession, housing market struggles, the complexity of financial products, as well as the economic impact of the covid-19 pandemic, it is important to consider ways to improve the financial well-being of americans, especially our military households, now and into the future. financial behaviors, including financial management, savings behavior, and investments have been analyzed for the general (civilian) population (e.g., henager & cude, 2016; xiao, 2008). additionally, specific sub populations have been investigated including the collegeaged population and millennials (e.g., henry et al., 2001; kim et al., 2019; lee & kim, 2020). unfortunately, less is known about the military population, as the limited research has focused on descriptive analyses rather than multivariate analyses. research from the military has identified that financial management skills and well-being are positively affected by financial knowledge (bell et al., 2014, 2009). further, younger members of the military had greater financial security than their civilian counterparts (finra ief, 2016). carlson and colleagues (2015) found that financial behaviors were positively impacted by high levels of financial knowledge (subjective), high self-mastery, and lower levels of financial anxiety for military members. additionally, they found that soldiers with emergency savings had better financial behaviors than those without emergency savings, while soldiers with credit card debt had worse financial behaviors than those without credit card debt. this study is aimed to explore financial behaviors of military households in the united states in two ways; short-term and long-term financial behaviors. the focus for short-term behaviors is emergency savings, spending within one’s limits, managing a checking account, and budgeting while the focus for long-term behaviors is retirement planning, savings, investments, and having a will. further, we investigate the role of financial knowledge and financial education as salient factors for these financial behaviors. empirical results will provide an important insight into policymakers as well as financial practitioners. while prior work on military households has provided this study a groundwork on the roles of financial knowledge and financial behaviors, little work has investigated the differences among shortterm and long-term financial behaviors and financial education. 2. literature review 2.1. financial behaviors, knowledge, and education understanding the role of financial knowledge and financial education on financial behaviors is important as decisions made throughout the life course can have lasting effects on a households’ present and future financial decision making. having higher levels of financial 36 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 literacy when younger leads to better financial decisions and improved quality of life later in life (james et al., 2012). there is a growing body of literature identifying the relationship between financial knowledge and financial behaviors. behaviors such as having a checking and/or savings account, making payments on time, having an emergency fund, and tracking expenses are associated with higher levels of financial knowledge (hilgert et al., 2003). additionally, higher financial knowledge was associated with ownership of investments and saving for long-term goals (moore, 2003) and retirement planning (delavande et al., 2008). having a will is an important component of estate planning and older adults with more assets were more likely to plan for wealth transference using a will (goetting & martin, 2001). however, planning for a lifetime of wealth transference should begin when an individual is young and be updated over time (garman & forgue, 2012). research has found that respondents with higher levels of financial knowledge were more likely to plan and to succeed in their planning (lusardi & mitchell, 2007, 2011). whereas, having low financial knowledge has been found to negatively impact long-term financial behaviors as well as daily financial management (braunstein & welch, 2002). specifically, more negative financial behaviors (e.g., high interest rate mortgages, limited savings and investments, and over borrowing) are associated with low levels of financial knowledge (lusardi, 2008). when investigating short-term and long-term financial behaviors, both subjective and objective financial knowledge was positively associated with both types of financial behaviors (henager & cude, 2016). in addition to objective financial knowledge, subjective financial knowledge has been linked to greater likelihood to plan (hadar et al., 2013), participate in best financial practices (robb & woodyard, 2011), and better credit card practices (allgood & walstad, 2016). the body of literature has also identified that socio-demographic status is important when examining financial behaviors. for example, gender, age, race, marital status, dependent children, employment status, educational attainment, and income have all been found to impact financial behaviors (fernandes et al., 2014; henager & cude, 2016; lee & kim, 2020; robb & woodyard, 2011; xiao et al., 2015; zick et al., 2012). financial education programs generally include both prevention and intervention strategies to improve financial knowledge, engagement, and communication in an effort to increase financial wellness (borden et al., 2016). the principle idea behind financial education programs is that many individuals and households lack the financial knowledge needed to make appropriate financial decisions (tang & baker, 2016). literature regarding the impact of financial education on financial knowledge shows mixed results; some researchers have found a significant positive association (kaiser et al., 2021; tang & peter, 2015) while others have found no significant association between financial education courses and financial knowledge (mandell, 2008, xiao et al., 2011 despite this, the potential benefit of financial education is often viewed as one of many approaches to increasing financial knowledge (gale & levine, 2010). as a result of concern for financial well-being and low levels of financial knowledge, government, business, and nonprofit entities have started developing programs aimed at improving financial knowledge with the goals of improving financial behaviors (fernandes et al., 2014). such programs are available specifically to military members through programming m. j. wilmarth et al. / financial services review 31 (2023) 35–54 37 developed within each service branch as well as nonprofits targeting military members (borden et al., 2016). 2.2. finances in military households proper management of finances is important for all types of families, as it may reduce financial strain and distress. both civilian and military families experience similar amounts of financial stress (skimmyhorn, 2014); however, there are differences that need to be considered (e.g., griffith, 2015). military families experience stressors that civilian families do not usually experience, including deployment and frequent moves required by the military (hosek & wadsworth, 2013). having confidence in their deployment support (e.g., trust in chain of command and available support options) was associated with decreased financial difficulties while deployed (griffith, 2015). conversely, having seen others wounded or killed in combat or post-deployment experiences (e.g., anger, frustration) was associated with increased financial difficulties (griffith, 2015) and growing literature is identifying the implications of post-traumatic stress disorder (ptsd) on financial stress and behaviors among service members (e.g., harrison et al., 2010; olson et al., 2018; wang & pullman, 2019). in an effort to help military families cope financially, pay increases have led to earnings of military members being higher than civilians with equivalent education levels (hosek & wadsworth, 2013). military families spend less on food, healthcare, personal items, and taxes when compared with civilian families (hosek & wadsworth, 2013). however, military spouses’ are more likely to be unemployed or work fewer hours than they prefer, as compared with their civilian counterparts (hosek & wadsworth, 2013). service members may be experiencing higher rates of debt than in the past, including more incoming members entering service with debt (26% in 1997 to 42% in 2003; hosek & wadsworth, 2013). not all military families experience finances in the same way, as there may be variance based on rank and branch of service. within military families, active duty families tend to be better off financially than those that are reserve service families (london & heflin, 2015), whereas those in the army have lower financial well-being than those in the air force (skimmyhorn, 2014). families of officers have better financial well-being than those of enlisted service members’ families (hosek & wadsworth, 2013). specific to financial behaviors, service members may differ from their civilian counterparts. results published from the national financial capability study (nfcs), identified that credit card holders (both civilian and military) participate in negative financial behaviors such as only paying the minimum payment, incurring late fees, or using a cash advance from a credit card. however, military service members were more likely than civilians to engage in at least one of these negative behaviors (finra ief, 2013a, 2013b). among a marine corp sample, commonly found financial problems included bounced checks and/or suspensions of check-cashing privileges, high credit card debt, overuse of credit, and high phone bills (varcoe et al., 2003). using the nfcs, skimmyhorn (2016b) found that military members had more types of savings accounts (e.g., has an emergency fund, has nonretirement investments) and greater credit card behaviors (e.g., not paying balance in full, paying 38 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 a late fee, and using a cash advance) that were problematic compared with the civilian population. on the other hand, over half of military households (57%; 36% for civilians) did not have difficulty covering their monthly expenses, 41% reported some difficulty, and 10% reported a great deal of difficulty, much related to pay grades (those in higher pay grades had easier times making ends meet) (finra ief, 2013a, 2013b). more of military respondents (51%) reported spending less than their income as compared with civilians (41%; finra ief, 2013a, 2013b). not only has pay grade within the military been found to correlate with financial behaviors; but also financial behaviors vary between the branches (skimmyhorn, 2014). as compared with army counterparts, those in the air force were equally as likely to report spending more than their income, while those in the navy were less likely and marines were more likely (skimmyhorn, 2014). in addition, military members tend to be better at saving than their civilian counterparts, from the 2012 survey 54% of service members had an emergency fund to cover three months of living expenses, while only 40% of civilians did (finra ief, 2013a, 2013b). of those in active duty, over half (57%) reported saving for retirement (defense manpower data center, 2016). focusing on retirement, differences between career military and noncareer military has been identified in terms of total family income, percent of income saved, retirement income sources, and total number of pension plans; as well as differences in financial satisfaction (brunson et al., 1998). according to a recent report from the consumer financial protection bureau (cfpb, 2019), veterans’ financial skills and behaviors (e.g., budgeting, spending within budget) were positively associated with their financial situation and financial well-being. additionally, the analysis identified that financial education increases financial behaviors and financial well-being (cfpb, 2019). 2.3. research questions given the previous studies on financial behavior, knowledge, and education discussed above, this study extends the existing literature by examining the associations between financial knowledge and education with financial behaviors of military households by addressing the following research questions. research question (rq) 1: are financial knowledge and financial education associated with positive short-term financial behaviors of military households? research question (rq) 2: are financial knowledge and financial education associated with positive long-term financial behaviors of military households? 3. method 3.1. dataset and sample selection the data used for this study came from the 2018 national financial capability study (nfcs) state-by-state survey instrument sponsored by the financial industry regulatory m. j. wilmarth et al. / financial services review 31 (2023) 35–54 39 authority (finra). the questionnaire was administered on a state-by-state basis to achieve approximately 500 observations from each state and the district of columbia and was designed to assist in better understanding financial capability in the united states (mottola & kieffer, 2017). the self-reported data were collected from june through october in 2018 (finra ief, 2019). the total sample size of the 2018 nfcs is 27,091 and this study includes 20,796; observations were dropped from the sample if the respondent chose “prefer not to say” for the objective financial knowledge and financial behavior questions and/or answered “prefer not to say” or “don’t know” for the subjective knowledge and financial education questions. the main analytic sample of those with military experience includes 3,045 households, including only households with a head of household who is active duty or previously a member of the u.s. armed service. within our analytic sample, 631 were active duty in the u.s. armed services, while 2,413 were formerly members of the armed services. as a reference group, we conducted the analyses with civilian households (n = 17,751). 3.2. dependent variables two key dependent variables were investigated, long-term and short-term financial behaviors as measured by henager and cude (2016) and kim et al. (2019). the 2018 nfcs collects one new question, “do you currently have a will?” and we adopted this variable as one of long-term financial behaviors. the long-term financial behavior index was created based on responses to four questions asking if the respondent had ever done any planning to evaluate the amount needed for their retirement, owned any retirement plans, owned any investments outside of their retirement accounts, and had a will. the short-term financial behavior index was created based on four questions asking if the respondent had an emergency fund, spent less than or equal to their income, did not overdraw their checking account occasionally, and used a budget. for each index, the four variables were coded as binary variables, one indicating the financial behavior. the responses were summed to create the two indices; each index ranging 0–4. the four items included in each index were equally weighted in the summation of items. 3.3. independent variables key independent variables reflect the level of financial knowledge, financial education, and years since military completion. objective financial knowledge was based on the number of correct answers to the six questions in the survey (ranging 0–6) and subjective financial knowledge was based on a scale of 1 = very low to 7 = very high. whether the respondent reported having received and participated in financial education was coded with 1 = yes, 0 = no. the exact wording of variables in the survey can be found in the appendix. additionally, two variables were included to categorize military groups. one category included years since military completion was categorized as follows; the respondent completed military service in the past year, 1–3 years, 4–10 years, more than 10 years, and currently active duty (reference). the other category for types of military service grouped respondents into branches of service; army, navy, air force, and others. 40 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 following previous studies on financial behaviors (e.g., henager & cude, 2016; kim et al., 2019), this study includes the set of following control variables; age, gender (male, female), race/ethnicity (white, black, hispanic, and asian/others), marital status (married, single, and separated/divorced/widowed), presence of dependent child(ren) (yes/no), employment status (employed, otherwise), education (less than high school, high school diploma, some college, bachelor degree, and post-bachelor degree), household income, substantial income drop (yes/no), banking status (yes/no) and homeownership (yes/no). lastly, we also controlled for state of residence. 3.4. analyses given the ordered nature of the dependent variables, we conducted ordered logistic regression analyses on composite variables of financial behaviors, which provide a general overview of the financial behaviors of military households. also, we conducted similar analyses for civilian households as a reference group. model 1: short-term financial behaviors = f (financial knowledge, financial education, socio-demographic status) model 2: long-term financial behaviors = f (financial knowledge, financial education, socio-demographic status) the nfcs provides a survey weight to be representative of the national population in terms of age, gender, ethnicity, education, and census division (with adjustments for the oversampled states for comparability with previous years), so all of our results are weighted. 4. results 4.1. descriptive results descriptive results for both samples’ characteristics are presented in table 1. in terms of short-term financial behaviors, the percentages of the military sample that participated in the four short-term behaviors were as follows: having emergency funds (66.7%), spending less than income (45.6%), not experiencing an overdraft (65.0%), and keeping a budget (58.2%). the mean composite score of short-term behaviors for military households was 2.35. in terms of long-term financial behaviors for the military sample; 63.3% had figured out the amount of savings they needed for retirement, 79.2% had a retirement plan(s), 52.2% owned investments outside of the retirement account, and 61.5% had a will. the mean composite score for long-term behaviors was 2.56. the civilian household sample had less than half of the sample participating in three of the four short-term behaviors; having emergency funds (48.3%), spending less than income (41.7%), and keeping a budget (41.0%). while 75.2% of the civilian households had not experienced an overdraft. the mean of the short-term behaviors index for civilian households was 2.06. looking at the long-term behaviors for civilian households, 43.7% of civilian households had figured out the amount of savings they needed for retirement and 61.7% m. j. wilmarth et al. / financial services review 31 (2023) 35–54 41 table 1 sample characteristics of military respondent and civilian households, 2018 national financial capability study (nfcs) variables military household (n = 3,045) civilian household (n = 17,751) short-term behaviors (0–4) mean (sd): 2.35 (1.25)*** mean (sd): 2.06 (1.33) emergency funds (ownership) 66.72%** 48.26% spending less than income 45.56% 41.71% no overdrafts 64.98%*** 75.16% budgeting 58.17% 41.00% long-term behaviors (0–4) mean (sd): 2.56 (1.37)*** mean (sd): 1.66 (1.34) retirement planning (amount needed) 63.28% 43.65% retirement account (ownership) 79.18%*** 61.65% investments (ownership) 52.24%*** 31.07% having a will 61.54%*** 29.89% objective financial knowledge (0–6) mean (sd): 3.39 (1.57)*** mean (sd): 3.14 (1.63) objective financial knowledge questions interest 76.63% 75.81% inflation 58.90% 58.35% bond price 33.58%*** 27.25% mortgage 82.38%*** 75.87% portfolio 51.17% 45.59% time value of money 36.58%*** 31.28% subjective financial knowledge (1–7) mean (sd): 5.66 (1.31)*** mean (sd): 5.09 (1.34) financial education 31.9%*** 22.2% mean age mean (sd): 50.9 (17.8)*** mean (sd): 46.3 (16.6) gender male 85.30%*** 43.09% female 14.70%*** 56.91% race/ethnicity white 64.03% 64.00% black 18.50%*** 10.68% hispanic 11.33%*** 16.58% asian/others 6.15%*** 8.75% marital status married 63.35% 49.57% single 22.28%*** 33.69% separated/divorce/widow 14.36%*** 16.74% having dependent children 42.28%*** 35.20% employed 59.04% 57.51% education less than high school 0.85%*** 2.79% high school degree 21.02%*** 28.83% some college 34.42%*** 27.52% bachelor’s degree 17.43% 18.83% post-bachelor’s degree 26.30%*** 22.03% household income less than $15,000 5.06%*** 12.12% $15,000–$24,999 7.63%*** 10.99% $25,000–$34,999 8.18%*** 11.35% $35,000–$49,999 13.35%* 15.01% $50,000–$74,999 18.29% 19.24% $75,000–$99,999 23.97%*** 12.60% $100,000–$149,999 16.89%*** 12.03% $150,000 or more 6.63% 6.66% (continued on next page) 42 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 had a retirement plan(s). further, 31.1% owned investments outside of their retirement account and only 29.9% had a will. the long-term behavior mean composite score for civilian households was 1.66. for both short-term and long-term behaviors, the means of the composite scores were statistically higher for the military households than civilian households. see figs. 1 and 2 for the comparison of short-term and long-term financial behaviors by household type. in terms of financial knowledge and education, military households had significantly higher mean scores for both objective and subjective financial knowledge than civilian households. military households had a mean score of 3.39 for objective financial knowledge as compared with 3.14 for civilian households. for subjective financial knowledge, military households in our sample had a mean score of 5.66 as compared with 5.09 for civilian households. in terms of financial education, 31.9% of the military household sample had received and participated in some form of financial education, significantly higher than the 22.2% of civilian households who had experienced financial education. the distribution of financial knowledge scores by household type is presented in figs. 3 and 4. other socio-demographic characteristics are presented in table 1. fig. 1. short-term financial behaviors by household type, 2018 national financial capability study (nfcs). weighted results. table 1 (continued) variables military household (n = 3,045) civilian household (n = 17,751) substantial income drop 31.11%*** 20.16% banked 94.77%*** 92.51% homeownership 76.08%*** 57.73% note. weighted results. *p < .05, **p < .01, ***p < .001. m. j. wilmarth et al. / financial services review 31 (2023) 35–54 43 specifically investigating the military sample, a comparison of financial behaviors and knowledge scores by the receipt of financial education is presented in table 2. military households that had received financial education had a significantly higher mean subjective financial knowledge score of 5.96 versus 5.52, but no significant difference in objective knowledge was found. a mixed pattern exists with financial behaviors, military households who had received financial education had a lower mean for short-term behaviors (2.29 vs. 2.39), but a higher mean for long-term behaviors (2.93 vs. 2.40). 4.2. multivariate results results for military households (see table 3) showed that the key independent variables were all associated with higher odds of having higher scores on composite indices for both short-term and long-term financial behaviors. in particular, a one unit increase in objective financial knowledge increased the odds of having a higher level of short-term behaviors by 6.8% and 13.3% for long-term behaviors. a one-unit increase in subjective financial knowledge increased the odds of having a higher level of short-term behaviors (51.4%) and longterm behaviors (43.9%). financial education experience increased the odds of having a higher level of long-term behaviors by 56.3%, the association was not held significantly with short-term behaviors. for military households, current military members had higher odds of long-term financial behaviors than retired members. air force members had higher odds of having a higher level of short-term and long-term behaviors than army members. respondents who were female, white, single, and did not have dependent children had higher odds of having a higher short-term behaviors index while age was positively associated with the odds for fig. 2. long-term financial behaviors by household type, 2018 national financial capability study (nfcs). weighted results. 44 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 long-term behaviors. additionally, higher income levels (relative to income less than $25,000) had increased odds of having higher levels of both short-term and long-term behaviors. the odds increase with higher income levels and are higher for long-term behaviors than short-term behaviors. having a substantial income drop was negatively associated with short-term behaviors while positively associated with long-term behaviors. banked households and homeowners had higher odds for both short-term and long-term financial behaviors. as a reference, results for both short-term and long-term behaviors of civilian households are presented in table 3. objective financial knowledge and subjective financial knowledge were associated with higher odds of having higher scores on each of the indices of financial behaviors for civilian households. similar to military households, experiencing financial education was only associated with long-term behaviors. respondents who were male, asian/other, single, did not have dependent children, and had a bachelor’s degree all had odds of having higher short-term behavior index scores. age, being white, being married, having no dependent children, being employed, and education were all associated with increased odds of having higher long-term behavior index scores. a similar pattern in income, substantial income drop, banking status, and homeownership as we observed with military households was found for civilian households. 5. discussion and relevance this study explored financial behaviors of military households in the united states by analyzing short-term and long-term financial behaviors. we mainly investigated the role of fig. 3. distribution of objective financial knowledge by household type, 2018 national financial capability study (nfcs). weighted results. m. j. wilmarth et al. / financial services review 31 (2023) 35–54 45 financial knowledge and financial education on financial behaviors. both objective and subjective financial knowledge show increased odds of having higher scores on both short-term and long-term behaviors, with larger odds for subjective financial knowledge for both indices. over-confidence can be a concern, specifically for younger adults (henager & cude, 2016), where the subjective knowledge and objective knowledge do not always align. robb and woodyard (2011) also reported that financial knowledge and financial confidence had a low correlation but both affect behavior. in this case, subjective knowledge shows that this is an important factor in the behavior of this military sample. in other words, if they think they can they will, as those with higher confidence perform more positive behaviors than those with lower confidence. the results for financial education indicate higher odds for long-term financial behaviors. the relationship was not significant for short-term financial behaviors. this is encouraging as the concern for military families grows and financial education targeted at military families has been in the policy conversation, particularly financial planning and preparation. borden and colleagues (2016) encourage the use of financial education programs to increase protective factors and coping strategies for military families as they manage their finances and potentially deal with financial strain. recommendations include financial education programs including preventative and interventional approaches focusing on financial communication and financial engagement, in addition to financial knowledge (borden et al., 2016). an interesting finding was that the military sample had a higher percentage of those with a will. this is an important part of long-term planning and an important topic for financial education for everyone (kotlikoff, 1988). military service members and their families have fig. 4. distribution of subjective financial knowledge by household type, 2018 national financial capability study (nfcs). weighted results. 46 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 access to legal assistance, which is free, which covers the writing of a will (military.com, 2020). having this free service available likely contributes to the higher number of wills, but also military personnel deployed likely want to make sure their family is taken care of and are encouraged to have a will. it is concerning, the number of civilian households without a will, and is a topic that ought to be included in financial education at all levels. 5.1. limitations while the nfcs offers a wealth of information about financial situations, characteristics, behaviors, and education of civilian and military households, there are still some drawbacks with the data and results should be viewed with them in mind. it is important to note that the nfcs is self-report data and not observed by a third party, so accuracy of the self-reported data are not fully known. the publicly available dataset does not contain information on whether the military member was an enlisted member or commissioned as an officer. this would have an impact on the salary while the member was still active military, as well as career trajectory, and potentially retirement positions. the data did not allow for complete control of the variation of those who have completed military service at different times. while we controlled for length of time since service was completed, there is still a broad range of personnel that are included in the sample. future work would benefit from the ability to have additional specific information to target service branch and career of active duty and retired personnel. additionally, while we have important information on financial education, we do not know specifics on the type of financial education the respondent received (e.g., through the military, through a military-affiliated partner, through a nonmilitary related source, type of programming, length, etc.). with many of the service branches offering increasing opportunities for financial education, in varying forms, it will be important to be able to further table 2 financial behaviors of households with a military head by financial education, 2018 national financial capability study (nfcs) variables military household with financial education (n = 962) military household without financial education (n = 2,057) objective financial knowledge (0-6) mean (sd): 3.31 (1.58) mean (sd): 3.44 (1.57) subjective financial knowledge (1-7) mean (sd): 5.96 (1.23)*** mean (sd): 5.52 (1.32) short-term behaviors (0–4) mean (sd): 2.29 (1.20)* mean (sd): 2.39 (1.28) emergency funds (ownership) 75.90%*** 62.55% spending less than income 38.60% 48.81% no overdrafts 52.36%*** 71.23% budgeting 62.48% 56.30% long-term behaviors (0–4) mean (sd): 2.93 (1.28)*** mean (sd): 2.40 (1.37) retirement planning (amount needed) 75.17%*** 58.04% retirement account (ownership) 84.01%*** 77.29% investments (ownership) 64.80% 46.65% having a will 69.00%* 58.41% note. weighted results. *p < .05, **p < .01, ***p < .001. m. j. wilmarth et al. / financial services review 31 (2023) 35–54 47 t ab le 3 o rd er ed lo g it re g re ss io n re su lt s o f sh o rt -t er m an d lo n g -t er m fi n an ci al b eh av io rs , m il it ar y re sp o n d en t an d ci v il ia n h o u se h o ld s, 2 0 1 8 n at io n al f in an ci al c ap ab il it y s tu d y (n f c s ) v ar ia b le s m il it ar y h o u se h o ld s n = 3 ,2 6 6 c iv il ia n h o u se h o ld s n = 2 0 ,1 8 3 s h o rt -t er m b eh av io rs l o n g -t er m b eh av io rs s h o rt -t er m b eh av io rs l o n g -t er m b eh av io rs o d d s ra ti o x 2 o d d s ra ti o x 2 o d d s ra ti o x 2 o d d s ra ti o x 2 f in an ci al k n o w le d g e an d fi n an ci al ed u ca ti o n o b je ct iv e fi n an ci al k n o w le d g e 1 .0 6 7 6 * 6 .0 6 1 8 1 .1 3 2 7 * * * 2 0 .7 9 4 0 1 .0 9 3 7 * * * 8 2 .1 2 6 5 1 .2 6 3 6 * * * 5 1 3 .7 7 5 6 s u b je ct iv e fi n an ci al k n o w le d g e 1 .5 1 4 1 * * * 1 8 0 .9 3 9 7 1 .4 3 8 5 * * * 1 3 9 .7 9 1 5 1 .3 7 2 7 * * * 7 6 3 .1 6 8 4 1 .3 5 5 1 * * * 6 2 3 .7 5 6 3 f in an ci al ed u ca ti o n 0 .8 7 3 8 3 .1 0 2 0 1 .5 6 3 0 * * * 3 0 .5 9 5 0 1 .0 0 5 0 0 .0 2 0 9 1 .3 6 5 3 * * * 7 8 .0 7 8 8 y ea rs si n ce m il it ar y co m p le ti o n (r ef er en ce : c u rr en tl y ac ti v e d u ty ) w it h in o n e y ea r 1 .5 9 9 5 1 .9 1 3 0 0 .4 9 9 4 * 4 .0 5 1 1 n /a 1 to 3 y ea rs ag o 1 .0 5 6 8 0 .0 7 8 5 0 .5 8 6 3 * * 6 .9 9 3 6 4 to 1 0 y ea rs ag o 0 .6 7 2 3 * 4 .9 0 3 9 0 .3 2 7 5 * * * 3 6 .9 9 6 6 m o re th an 1 0 y ea rs ag o 0 .9 0 0 7 0 .3 8 4 1 0 .2 1 1 8 * * * 7 9 .1 3 0 7 t y p e o f m il it ar y se rv ic e (r ef er en ce : a rm y ) n av y 0 .9 8 0 4 0 .0 3 9 5 1 .1 9 7 6 3 .2 5 4 7 n /a a ir f o rc e 1 .3 0 7 5 * * 6 .8 9 3 1 1 .5 2 6 8 * * * 1 7 .0 2 1 9 o th er s 1 .1 6 3 1 1 .3 5 6 7 1 .1 9 4 9 1 .8 5 8 3 c o n tr o l v ar ia b le s a g e 1 .0 0 3 5 0 .7 7 3 9 1 .0 1 9 3 * * * 2 2 .0 3 1 7 1 .0 0 1 5 1 .6 5 2 0 1 .0 2 7 9 * * * 5 2 3 .6 1 3 6 g en d er (r ef er en ce : f em al e) 0 .7 8 5 8 * 6 .1 0 7 3 1 .0 7 4 1 0 .5 0 7 8 1 .0 7 6 6 * 6 .2 7 4 8 0 .9 8 0 2 0 .4 3 2 5 r ac e/ et h n ic it y (r ef er en ce : w h it e) b la ck 0 .6 6 7 9 * * * 1 6 .6 1 9 2 1 .2 1 6 5 3 .3 4 9 4 0 .9 2 0 9 2 .9 9 0 0 1 .0 0 6 9 0 .0 1 8 7 h is p an ic 1 .0 4 1 3 0 .1 3 1 6 0 .8 8 7 5 1 .0 8 9 2 1 .0 7 3 6 3 .2 1 0 6 0 .8 3 4 7 * * * 1 9 .0 1 9 7 a si an /o th er s 0 .8 9 5 7 0 .5 9 5 6 0 .9 3 5 4 0 .2 1 4 2 1 .1 4 3 4 * * 7 .0 1 7 4 0 .9 7 1 8 0 .3 0 2 7 m ar it al st at u s (r ef er en ce : m ar ri ed ) s in g le 1 .2 2 9 4 * 4 .5 4 3 7 1 .1 2 7 3 1 .3 1 1 8 1 .2 3 3 4 * * * 2 8 .3 9 2 9 0 .9 6 9 3 0 .5 9 4 7 s ep ar at ed /d iv o rc e/ w id o w 1 .2 2 0 5 3 .7 2 2 6 1 .0 0 1 2 0 .0 0 0 1 0 .9 8 2 5 0 .1 6 8 8 0 .8 0 0 8 * * * 2 5 .3 6 8 0 d ep en d en t ch il d re n (r ef er en ce : n o ) 0 .6 6 1 9 * * * 2 2 .8 2 2 9 0 .9 8 1 7 0 .0 4 1 7 0 .5 4 7 3 * * * 3 4 0 .4 1 2 3 0 .8 5 8 6 * * * 2 0 .5 3 8 9 e m p lo y ed (r ef er en ce : n o ) 1 .1 6 6 5 2 .8 9 5 9 1 .1 7 4 9 3 .1 1 5 1 0 .9 8 8 4 0 .1 3 1 7 1 .0 9 5 3 * * 7 .3 7 2 0 e d u ca ti o n (r ef er en ce : l es s th an h ig h sc h o o l) h ig h sc h o o l d eg re e 0 .9 9 0 2 0 .0 0 0 7 0 .6 5 2 3 1 .1 6 6 9 1 .1 1 8 8 1 .5 3 0 2 1 .8 9 0 8 * * * 3 2 .8 1 7 7 s o m e co ll eg e 0 .6 5 6 7 1 .2 0 9 6 0 .5 0 8 2 2 .8 9 8 3 0 .9 6 2 2 0 .1 7 5 6 1 .9 6 5 0 * * * 3 6 .5 1 6 9 b ac h el o r’ s d eg re e 1 .1 3 3 7 0 .1 0 4 9 0 .6 0 3 6 1 .5 7 5 1 1 .2 3 9 6 * 5 .0 5 8 0 2 .8 7 6 0 * * * 8 4 .8 3 5 3 p o st -b ac h el o r’ s d eg re e 0 .9 2 8 3 0 .0 3 7 5 0 .6 7 5 2 0 .9 6 8 7 1 .0 6 6 8 0 .4 6 9 5 2 .8 6 9 4 * * * 8 5 .8 7 1 7 (c o n ti n u ed o n n ex t p a g e) 48 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 t ab le 3 (c o n ti n u ed ) v ar ia b le s m il it ar y h o u se h o ld s n = 3 ,2 6 6 c iv il ia n h o u se h o ld s n = 2 0 ,1 8 3 s h o rt -t er m b eh av io rs l o n g -t er m b eh av io rs s h o rt -t er m b eh av io rs l o n g -t er m b eh av io rs o d d s ra ti o x 2 o d d s ra ti o x 2 o d d s ra ti o x 2 o d d s ra ti o x 2 h o u se h o ld in co m e (r ef er en ce : l es s th an $ 1 5 ,0 0 0 ) $ 1 5 ,0 0 0 – $ 2 4 ,9 9 9 0 .6 9 7 7 3 .3 7 6 2 1 .3 3 0 0 2 .0 0 7 8 0 .9 7 3 2 0 .2 1 0 6 1 .6 6 2 1 * * * 5 7 .8 4 3 2 $ 2 5 ,0 0 0 – $ 3 4 ,9 9 9 0 .8 3 0 3 0 .8 8 9 3 1 .9 8 5 8 * * * 1 1 .5 4 8 5 1 .1 7 5 0 * * 7 .3 5 1 0 2 .4 9 0 0 * * * 1 9 0 .7 5 8 5 $ 3 5 ,0 0 0 – $ 4 9 ,9 9 9 1 .2 1 0 9 1 .0 6 5 7 3 .5 3 7 4 * * * 4 3 .6 0 7 0 1 .5 2 0 7 * * * 5 3 .3 5 1 2 3 .9 8 0 5 * * * 4 7 1 .5 5 0 6 $ 5 0 ,0 0 0 – $ 7 4 ,9 9 9 2 .0 2 9 1 * * * 1 4 .5 6 6 6 4 .9 0 8 2 * * * 6 8 .8 0 0 8 1 .9 3 6 1 * * * 1 3 2 .9 2 7 6 5 .2 1 4 8 * * * 6 7 6 .1 8 5 2 $ 7 5 ,0 0 0 – $ 9 9 ,9 9 9 2 .2 0 3 6 * * * 1 7 .8 3 6 3 7 .3 6 7 7 * * * 1 0 4 .4 2 4 9 2 .6 9 8 5 * * * 2 3 8 .3 2 4 7 7 .5 1 1 2 * * * 8 2 8 .7 9 2 4 $ 1 0 0 ,0 0 0 – $ 1 4 9 ,9 9 9 2 .8 1 0 9 * * * 2 8 .5 3 2 5 6 .6 5 6 5 * * * 8 9 .0 6 0 1 3 .4 3 0 1 * * * 3 3 4 .6 9 2 4 8 .5 6 6 8 * * * 8 7 3 .3 9 7 1 $ 1 5 0 ,0 0 0 o r m o re 4 .3 2 8 0 * * * 4 0 .9 4 1 7 1 0 .2 7 5 9 * * * 9 7 .4 8 1 5 5 .0 7 8 4 * * * 4 1 7 .7 8 5 1 1 2 .9 0 3 5 * * * 9 3 3 .3 8 1 6 s u b st an ti al in co m e d ro p (r ef er en ce : n o ) 0 .3 7 2 0 * * * 1 0 7 .3 2 6 8 1 .9 2 3 4 * * * 4 5 .1 4 9 9 0 .5 1 3 3 * * * 3 4 0 .4 2 6 8 1 .1 5 5 0 * * * 1 4 .8 0 7 1 b an k ed (r ef er en ce : n o ) 3 .9 7 2 9 * * * 6 5 .9 3 7 4 2 .0 8 2 1 * * * 1 9 .1 7 6 7 5 .1 2 1 8 * * * 7 1 1 .3 3 1 7 2 .2 8 6 7 * * * 1 4 5 .6 3 4 2 h o m eo w n er sh ip (r ef er en ce : n o ) 1 .8 7 3 5 * * * 4 8 .8 8 5 8 3 .8 8 1 8 * * * 2 1 6 .0 2 1 9 1 .5 4 3 3 * * * 1 6 9 .1 2 4 6 2 .1 3 1 2 * * * 4 8 2 .7 6 6 4 r eg io n al fi x ed ef fe ct (s ta te o f re si d en ce ) in cl u d ed in cl u d ed in cl u d ed in cl u d ed m o d el fi t m ea n co n co rd an t 7 5 .0 % 7 9 .0 % 7 4 .3 % 8 1 .0 % n o te . w ei g h te d re su lt s. * p < .0 5 , * * p < .0 1 , * * * p < .0 0 1 . m. j. wilmarth et al. / financial services review 31 (2023) 35–54 49 understand the type and content of the financial education program. we do not have data explicitly identifying what was included within financial education and also the financial arrangements within households (e.g., how are financial decisions made, who makes the decisions). additionally, the nfcs dataset does not include much information on household wealth, the only available proxy for wealth is homeownership status. wealth serves as an important determinant of financial behaviors and future work would benefit from including wealth variables. future research would benefit from use of data that specifically focuses on military households, even further those in active duty, to be able to identify specific areas of concern and possible interventions appropriate for active duty personnel. even more important would be the ability to identify the variations and implications of branch of service has in this area of research. for example, there are varying levels and implementation of financial education across the branches of services, so being able to compare between the branches would also improve this research, which is currently limited by sample size and limited information on specific financial education information. 5.2. future research and conclusion for future research, we would like to further explore the financial status of military households and contributing factors to their financial well-being. how these households are different from other occupational groups and how educators can reach out to and understand the needs of military households. further, as indicated by previous literature (bell et al., 2014; carlson et al., 2015), the connection between financial stress and financial well-being continues to grow and the unique experiences of military life (e.g., deployments, frequent moves) that link to financial outcomes should be further examined. empirical results from the 2018 nfcs provide an important insight into policymakers as well as financial practitioners and educators as they seek to gain a greater understanding of how to reach military families through education that not only increases their objective level of knowledge, but also increases their subjective financial knowledge as well. these efforts help military households and similar members in volunteer service prepare for and guard against financial challenges. practitioners need to be aware of the unique challenges experienced by military households and how they are related to finances (borden et al., 2016). however, given that military households are unique in many ways, including across branches, caution is warranted in generalizing the results to other populations. military households are many times considered to be financially vulnerable households and our results support the need to increase their financial knowledge (objective and subjective). ultimately this may serve as a way to protect them not only while they serve, but also after they leave the military. both financial education and financial knowledge increase the odds of households participating in these positive financial behaviors, so increasing the access as well as the quality of the programming may be one avenue to increasing financial knowledge and potentially financial behaviors. developing financial education programs that build systems of support and skills will help military households further develop positive financial behaviors. professionals who work with military households as clients should consider addressing not only financial knowledge with their clients, but also areas of 50 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 financial communication and financial engagement to help in increasing subjective financial knowledge, which will help serve as protective factors and coping strategies for military families. appendix description of key variables in the 2018 national financial capability study (nfcs) variable description short-term behaviors the responses were summed for index; ranging 0–4. emergency funds “have you set aside emergency or rainy day funds that would cover your expenses for 3 months, in case of sickness, job loss, economic downturn, or other emergencies?” spending less than or equal to income “over the past year, would you say your spending was less than, more than, or about equal to your income?” no overdrafts “do you overdraw your checking account occasionally?” budgeting i have money left over at the end of the month long-term behaviors the responses were summed for index; ranging 0–4. retirement planning “have you ever tried to figure out how much you need to save for retirement?” retirement account “do you have any retirement plans through a current or previous employer, like a pension plan or a 401(k)?” investments “not including retirement accounts, have any investments in stocks, bonds, mutual funds, or other securities?” having a will “do you currently have a will?” objective financial knowledge sum of correct answers to financial knowledge questions, ranging 0–6. interest “suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow?” inflation “imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account?” bond price “if interest rates rise, what will typically happen to bond prices?” mortgage “a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less.” portfolio “buying a single company’s stock usually provides a safer return than a stock mutual fund.” time value of money “suppose you owe $1,000 on a loan and the interest rate you are charged is 20% per year compounded annually. if you didn’t pay anything off, at this interest rate, how many years would it take for the amount you owe to double?” subjective financial knowledge ‘‘on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall financial knowledge?’’ financial education “was financial education offered by a school or college you attended, or a workplace where you were employed?” m. j. wilmarth et al. / financial services review 31 (2023) 35–54 51 references allgood, s., & walstad, w. b. (2016). the effects of perceived and actual financial literacy on financial behaviors. economic inquiry, 54, 675–697. https://doi.org/10.1111/ecin.12255 american psychological association. (2017). stress in america: coping with change. available at https://www. apa.org/news/press/releases/stress/2016/coping-with-change.pdf bell, c. j., gorin, d. r., & hogarth, j. m. (2009). does financial education affect soldiers’ financial behavior. networks financial institute working paper 2009-wp-08. available at http://ssrn.com/abstract= 1445635. bell, m. m., nelson, j. s., spann, s. m., molloy, c. j., britt, s. l., & goff, b. n. (2014). the impact of financial resources on soldiers’ well-being. journal of financial counseling and planning, 25, 41–52. borden, l. m., braughton, j., brown, s., davis, l., gunty, a., jaeger, e., mikal, j., otto, a., richmond, a., roddy, m., roeske, r., sherman, m. d., werner, e., williams, r., & wilcox, s. (2016). military families and financial stress. available at https://reachmilitaryfamilies.umn.edu/sites/default/files/rdoc/military_ families_and_financial_stress_dod3.pdf braunstein, s., & welch, c. (2002). financial literacy: an overview of practice, research, and policy. federal reserve bulletin, 88, november, 445–457. https://doi.org/10.17016/bulletin.2002.88-11 brunson, b. h., snow, m., & gustafson, a. w. (1998). mid-life career change: career military versus noncareer military financial well-being and financial satisfaction. journal of financial counseling and planning, 9, 41– 52. carlson, m. b., britt, s. l., & goff, b. n. (2015). factors associated with a composite measure of financial behavior among soldiers. journal of financial counseling and planning, 26, 30–42. https://doi.org/10.1891/ 1052-3073.26.1.30 carlson, m. b., nelson, j. s., & skimmyhorn, w. l. (2015). military personal finance research. in j. j. xiao (ed.), handbook of consumer finance research (pp. 251–264). northampton, ma: springer. consumer financial protection bureau. (2019). financial well-being of veterans. washington, dc. available at https://files.consumerfinance.gov/f/documents/cfpb_osa_financial-well-being-americans_report.pdf. defense manpower data center. (2016). status of forces surveys of active duty members (2013 & 2014 sofs-a): briefing on leading indicators, military onesource, financial health, family life, access to technology, impact of deployments, and permanent change of station (pcs) moves. seaside, ca. available at https://download. militaryonesource.mil/12038/mos/reports/sofs-a_briefing_20160311.pdf delavande, a., rohwedder, s., & willis, r. (2008). preparation for retirement, financial literacy and cognitive resources. working paper 2008-190. available at http://www.nber.org/papers/w15266.pdf fernandes, d., lynch, j. g., jr., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors.management science, 60, 1861–1883. https://doi.org/10.1287/mnsc.2013.1849 finra ief. (2019). the state of u.s. financial capability: the 2018 national financial capability study. washington, dc. available at https://www.usfinancialcapability.org/downloads/nfcs_2018_report_natl_ findings.pdf finra ief. (2016). financial capability in the united states 2016. washington, dc. available at http://www. usfinancialcapability.org/downloads/nfcs_ 2015_ report_natl_findings.pdf finra ief. (2013a). financial capability in the united states: 2012 report of military findings. washington, dc. available at http://www.usfinancialcapability.org/downloads/nfcs_2012_report_military_findings.pdf finra ief. (2013b). financial capability in the united states: report of findings from the 2012 national financial capability study. washington, dc. available at http://www.usfinancialcapability.org/downloads/ nfcs_2012_report_ natl_ findings.pdf gale, w. g., & levine, r. (2010). financial literacy: what works? how could it be more effective? available at ssrn: https://dx.doi.org/10.2139/ssrn.1758910 garman, e. t., & forgue, r. e. (2012). personal finance (11th ed.). mason, oh: south-western cengage learning. goetting, m. a., & martin, p. (2001). characteristics of older adults with written wills. journal of family and economic issues, 22, 243–264. https://doi.org/10.1023/a:1016699822855 52 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 griffith, j. (2015). homecoming of soldiers who are citizens: re-employment and financial status of returning army national guard soldiers from operations iraqi freedom (oif) and enduring freedom (oef). work (reading, mass.), 50, 85–96. https://doi.org/10.3233/wor-131794 hadar, l., sood, s., & fox, c. r. (2013). subjective knowledge in consumer financial decisions. journal of marketing research, 50, 303–316. https://doi.org/10.1509/jmr.10.0518 harrison, j. p., satterwhite, l. f., & ruday, w. (2010). the financial impact of post traumatic stress disorder on returning us military personnel. journal of health care finance, 36, 65–74. henager, r., & cude, b. j. (2016). financial literacy and longand short-term financial behavior in different age groups. journal of financial counseling and planning, 27, 3–19. https://doi.org/10.1891/1052-3073.27.1.3 henry, r. a., weber, j. g., & yarbarough, d. (2001). money management practices of college students. college student journal, 35, 244–249. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309–322. available at http://www. federalreserve.gov/pubs/bulletin/2003/0703lead.pdf hosek, j., & wadsworth, s. m. (2013). economic conditions of military families. the future of children, 23, 41–59. https://doi.org/10.1353/foc.2013.0009 james, j. d., boyle, p. a., bennett, j. s., & bennett, d. a. (2012). the impact of health and financial literacy on decision making in community-based older adults. gerontology, 58, 531–539. https://doi.org/10.1159/ 000339094 kaiser, t., lusardi, a., menkhoff, l., & urban, c. (2021). financial education affects financial knowledge and downstream behaviors. journal of financial economics, 145, 255–272. kim, k. t., anderson, s., & seay, m. (2019). financial knowledge and short-term and long-term financial behaviors of millennials in the united states. journal of family and economic issues, 40, 194–208. https://doi.org/ 10.1007/s10834-018-9595-2 kotlikoff, l. (1988). intergenerational transfers and savings. journal of economic perspectives, 2, 41–58. https://doi.org/10.1257/jep.2.2.41 lee, s. t., & kim, k. t. (2020). propensity to plan, financial knowledge, overconfidence, and credit card management behaviors of millennials. family and consumer sciences research journal, 49, 123–143. https:// doi.org/10.1111/fcsr.12381 london, a. s., & heflin, c. m. (2015). supplemental nutrition assistance program (snap) use among activeduty military personnel, veterans, and reservists. population research and policy review, 34, 805–826. https://doi.org/10.1007/s11113-015-9373-x lusardi, a. (2008). household saving behavior: the role of financial literacy, information, and financial education programs (working paper 14084). available at http://www.nber.org/papers/w13824 lusardi, a., & mitchell, o. s. (2007). financial literacy and retirement planning: new evidence from the rand american life panel (working paper 2007–2157). available at http://www.nber.org/papers/w17078 lusardi, a., & mitchell, o. s. (2011). financial literacy and planning: implications for retirement wellbeing (no. w17078). cambridge, ma: national bureau of economic research. luther, r. k., leech, i. e., & garman, e. t. (1998). the employer’s cost for the personal financial management difficulties of workers: evidence from the u.s. navy. personal finances and worker productivity, 2, 175– 184. mandell, l. (2008). the financial literacy of young american adults: results of the 2008 national jump$tart coalition survey of high school seniors and college students. available at http://www.jumpstart.org/assets/ files/2008surveybook.pdf military.com. (2020). free legal assistance. available at https://www.military.com/benefits/military-legalmatters/free-legal-assistance.html mottola, g. r., & kieffer, c. n. (2017). understanding and using the national financial capability study. family and consumer sciences research journal, 46, 31–39. https://doi.org/10.1111/fcsr.12227 moore, d. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences (technical report 03-39). available at https://www.sesrc.wsu.edu/sesrcsite/papers/files/dfitechreport-final2-16-04.pdf m. j. wilmarth et al. / financial services review 31 (2023) 35–54 53 olson, j. r., welsh, j. a., perkins, d. f., & ormsby, l. (2018). individual, family, and community predictors of ptsd symptoms following military deployment. family relations, 67, 615–629. https://doi.org/10.1111/ fare.12343 robb, c. a., & woodyard, a. s. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning, 22, 60–70. skimmyhorn, w. l. (2016a). assessing financial education: evidence from boot camp. american economic journal: economic policy, 8, 322–343. https://doi.org/10.1257/pol.20140283 skimmyhorn, w. l. (2016b). comparing military and civilian household finances: descriptive evidence from recent surveys. journal of consumer affairs, 50, 471–483. https://doi.org/10.1111/joca.12109 skimmyhorn, w. l. (2014). the financial welfare of military households: descriptive evidence from recent surveys. available at http://www.usma.edu/sosh/siteassets/sitepages/ faculty/publications/finra-paper2014.pdf tang, n., & baker, a. (2016). self-esteem, financial knowledge and financial behavior. journal of economic psychology, 54, 164–176. https://doi.org/10.1016/j.joep.2016.04.005 tang, n., & peter, p. c. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience. financial services review, 24, 119. varcoe, k. p., lees, n. b., wright, j., & emper, n. (2003). financial issues faced by marine corps families. journal financial counseling and planning, 14, 43–50. xiao, j. j. (2008). applying behavior theories to financial behavior. in j. j. xiao (ed.), handbook of consumer finance research (pp. 69–81). new york, ny: springer. xiao, j. j., chen, c., & sun, l. (2015). age differences in consumer financial capability. international journal of consumer studies, 39, 387–395. https://doi.org/10.1111/ijcs.12205 xiao, j. j., serido, j., & shim, s. (2011). financial education, financial knowledge, and risky credit behavior of college students. in d. lamdin (ed.), consumer knowledge and financial decisions (pp. 113–128). new york, ny: springer. https://doi.org/10.1111/ijcs.12205 wang, z., & pullman, l. e. (2019). financial stress, financial stability, and military spousal well-being. journal of military, veteran and family health, 5, 82–91. https://doi.org/10.3138/jmvfh.5.s1.2018-0025 zick, c. d., mayer, r. n., & glaubitz, k. (2012). the kids are all right: generational differences in responses to the great recession. journal of financial counseling and planning, 23, 3–16. 54 m. j. wilmarth et al. / financial services review 31 (2023) 35–54 perspectives on “sell in may and go away”: a look at recent evidence and implications anthony lovisceka,*, adam brodera adepartment of finance, seton hall university, 400 south orange avenue, south orange, nj 07079, usa abstract this study offers three perspectives on the quote “sell in may and go away.” first, it tests the annual performance of switching from four equity mutual funds to u.s. treasury bills during the historically low-return period of may 1 to october 31 against that of the buy-and-hold strategy from november 1 to october 31. second, it examines the switching strategy during the two bear and two bull markets occurring from 2000 to 2016. third, it tests the impact of taxes on the switching strategy. although there are signs of switching strategy effectiveness, especially during the bear markets, the results lack the statistical significance to conclude that it is superior to the buy-and-hold strategy. © 2018 academy of financial services. all rights reserved. jel classification: g11 keywords: switching strategy; large-cap fund; small-cap fund; mid-cap fund; balanced fund 1. introduction here are some memorable wall street quotes: “you can’t fight the tape” (hooke, 2010); “the market can stay irrational longer than you can stay solvent” (shilling, 1993, from j. m. keynes, 1920); “the trend is your friend” (taylor, 1988); and “be fearful when others are greedy and greedy when others are fearful” (buffet, 2008) * corresponding author. tel.: �1-973-761-9207; fax: �1-973-761-9217. e-mail address: anthony.loviscek@shu.edu financial services review 27 (2018) 303-322 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. another well-known and popular quote is “sell in may and go away.” its origin is traced to the british saying: “sell in may and go away and come on back on st. leger’s day,” (brecht, 2016), the day of the final race in britain’s mid-september triple crown. the event marked the return of high-class citizenry from their cooler country residences to the cities, and presumably to the stock-selection hunt. today, according to the stock trader’s almanac (hirsch, 2016), the quote refers to the observation that, since 1950, stock returns, as measured by the dow jones industrial average, have advanced by nearly 7.5% during the months from november through april but only 0.30% during the months from may through october. the may–october period, in particular, is highlighted by september’s historically negative monthly return of �0.75%, the worst performing month for equities. such patterns were first officially published in the stock trader’s almanac (hirsch, 1986). not surprisingly, given the high correlations across equity market measures, the more-diversified s&p 500 posts an average of 1.14% from november through april compared with 0.25% from may through october. this study does not challenge these patterns; rather, it points out three additional perspectives centered on the individual investor. first, unlike in some previous studies that use data before the official publication date of the pattern (e.g., bouman and jacobsen, 2002; kochman and badarinathi, 2008; kochman, badarinathi, and bray, 2014; lucey and zhao, 2008; maberly and pierce, 2004), or studies that test for the “sell in may” effect by using selected data after the official publication date (e.g., andrade, chhaochharia, and fuerst, 2013; jones and lundstrum, 2009), the results in this study begin in 1986, the year that the pattern was first published. thus, the results go a long way to avoid possible data mining and data-snooping biases that could significantly affect the findings in previous studies, as compellingly pointed out by sullivan, timmermann, and white (2001) and by irwin and park (2007) regarding calendar year effects in stock prices. second, the first two decades of the new century have witnessed, two bear markets, from 2000 to 2002 and from 2007 to 2009, and two bull markets, from 2002 to 2007 and from 2009 to deep into the next decade. how well the switching strategy has held-up during these times is an issue that is unexplored in the literature. as part of this test, there is the opportunity to assess the impact of the global financial crisis of 2008–2009 on the switching strategy. third, a recognized but underemphasized issue is one of taxation. the switching strategy is short-term, implying that the gains will be taxed at ordinary rates under the u.s. tax code unless they are sheltered. the buy-and-hold strategy is long-term and, therefore, receives favorable tax treatment under the u.s. tax code when gains are realized after one year (with unrealized gains theoretically being held “forever”). thus, the after-tax rate of return should be in focus for portfolios that are not tax-sheltered, another perspective that we incorporate. 2. a look at the literature the observation “sell in may” was first officially recognized and published in 1986 in the stock trader’s almanac. published annually since 1968, it reveals a rich array of detailed technical trading tactics and strategies, including the best and worst days to trade stocks 304 a. loviscek, a. broder / financial services review 27 (2018) 303-322 daily, weekly, and monthly, as well as contrarian and inside trading approaches. to assess the efficacy of these approaches, however, two salient points should be kept in mind. first, as the almanac’s author, hirsch may not be an unbiased reference. second, technical trading strategies do not align with the large body of evidence supporting the weak form of the efficient markets hypothesis. beginning with his seminal work on the subject, fama (1970) has long asserted that evidence of the efficacy of technical trading strategies is sparse, a view supported in other efficient markets studies and by efficient markets advocates (e.g., bessembinder and chan, 1998; chang and lewellen, 1984; malkiel, 1995; 2015; poterba and summers, 1987; sewell, 2012). although exceptions to it have been documented (e.g., anderson and smith, 2006; lo, 2017; prentis, 2011; schleifer and vishnay, 1997; smith, 2016), sullivan et al. (2001) point out the flaws in studies of seasonal stock price movements if the data are mined. similarly, irwin and park (2007) show that, while technical trading strategies can be profitable, as indicated in 56 of 95 studies they examine, most of the evidence is subject to data snooping biases, questionable ex-post trading rules, and inaccurate risk adjustments. for the evidence to be compelling, future research must adjust for these issues, as this study attempts to do. at the outset, the strategy “sell in may and go away” refers to the seasonal observation that the additional reward for bearing more risk is greater from november 1 to april 30 than it is from may 1 to october 31. as a result, a “pure” test of the strategy should be one of maintaining the same degree of risk as the buy-and-hold approach. this means that individual investors should increase their equity investments while the risk premium is higher, such as borrowing through margin accounts. for practical purposes, we take the position that individual investors are more likely to avoid leveraging as part of a plan to move to lower-volatile fixed-income securities, such as u.s. treasury bills. with respect to the literature, however, the switch is only a way to test the effectiveness of the observation, which does not explicitly call for it.1 as evidence favoring the “sell in may” strategy, bouman and jacobsen (2002) document the seasonal effect, also known as the halloween indicator, because it marks the end of a volatile and unsettling period—from may through october—for stock investors. they find that the effect significantly holds up between 1970 and 1998 in 36 of the 37 developed and emerging markets, with particularly strong effects in europe. in fact, they find british evidence of it dating back to 1694. they conclude that none of the conventional explanations can compellingly explain this puzzle. responding to the patterns published in the almanac in 1986, aronson and masters (2006) back-tested the switching strategy from november 1987 through april 2006 based on their own scientific technique. the evidence pointed to annualized returns of 16.3% on the s&p 500 during the high-performance period compared with 3.9% for the low-performing months. nonetheless, unless investors believe that “history repeats itself”—a characteristic associated with technical analysis and charting—back-testing can easily overstate future performance. that said, the performance differences lead the authors to conclude that the switching strategy offers the most effective way to earn above-average returns among the over 6,000 rules in the almanac. kochman and badarinathi (2008), using the s&p 500 between 1926 and 2004, document that the annualized returns for the periods november through april and may through october 305a. loviscek, a. broder / financial services review 27 (2018) 303-322 are 15.6% percent and 9.1% percent, respectively. they also find that november–april outperforms may–october by 13.6%, or 19.5% versus 5.9%. however, they recommend investors stay fully invested and switch from high-beta stocks to low-beta stocks as well as sell call options. in a follow-up study, based on eight years of data covering november 2004 through october 2012, kochman et al. (2014) show that per annum returns for november through april are 10.1%, which significantly offset the negative annualized loss of nearly 1.2% for may through october. they conclude that investors should adjust accordingly to the “sell in may” strategy. in a comprehensive study, andrade et al. (2013) present evidence consistent with the “sell in may” strategy. using msci data in local currencies and guided by bouman and jacobsen (2002), they show that it is pervasive in financial markets across 37 countries, registering a mean return that exceeds the underperforming months in all 37 countries. more important, the results are not only out-of-sample but also are persistent. their results support the findings of bouman and jacobsen (2002). the effect found is not only statistically significant but is also economically large, and holds across foreign exchange, carry trade, and risk premiums. in another international study, borowski (2015) examines the “sell-in-may” effect on 122 equity indices and 39 commodities and across several different periods. he finds, as does andrade et al. (2013), that the effect holds for november through april. as well, he discovers that it applies to the following periods: from october 15 through may, from november 15 through may, from october 1 through may and from november 1 through may. he also shows that it holds across various return calculations, including daily mean returns. the evidence suggests that markets are not as efficient as many researchers have concluded. on the other hand, maberly and pierce (2004) counter these results, at least for u.s. equities, by adjusting for two major outliers that they assert significantly drive the results of bouman and jacobsen: the stock market crash of 1987 and the failure of the hedge fund long-term capital management in 1998. they extend their analysis to u.s. futures markets and do not find evidence of market inefficiency from 1970 to 2003. lucey and zhao (2008), recognizing the possibility of an exploitable november–april effect, search for evidence of it from 1926 through 2002, but find nothing significant. they suggest that the positive results found in previous studies reflect, at least in part, the january effect. all things considered, the evidence is not strong enough to conclude that markets are inefficient. jones and lundstrum (2009) reach a similar conclusion. they rely on two periods, 1976 to 1998, as did bouman and jacobsen (2002), and 1991 to 2006. they use s&p 500 data and a full-year asset allocation strategy between stocks and u.s. treasury bills, and find that the cumulative return for a buy-and-hold strategy beats the november–april switching method, registering a return of 17.3% versus 15.0% for 1991 through 2006. thus, in contrast to studies uncovering profitable opportunities for the “sell in may” thesis, they find that it underperforms for this period. because this period overlaps with the years in the study by bouman and jacobsen (2002), jones and lundstrum (2009) cannot rule out that the findings of bouman and jacobsen (2002) may be more time-specific than noted in the literature. they point out that the s&p 500 return was 15.5% during the 1980s and 16.5% during the 1990s (each well above the annual average return on equities of approximately 10% from 1928 to 306 a. loviscek, a. broder / financial services review 27 (2018) 303-322 2017). therefore, it is not clear to what extent investors would have profited from the switching strategy, especially when considering the incentive to “buy and hold,” given the bull market conditions at those times. furthermore, and because the trading pattern was not revealed until 1986, u.s. investors could not have fully implemented the switching strategy from 1976 to 1998, for example, unless they uncovered it themselves. similarly, dichtl and drobetz (2015) revisit whether the strategy can still produce above-average market performance. in line with the work of sullivan et al., (2001) and lucey and zhao (2008), they consider which investment instruments investors had at their disposal, given limited data, in the interest of mitigating data mining and snooping biases. using a bootstrap approach, they find the “sell in may” strategy is becoming less effective over time, even disappearing in some instances. they conclude that the evidence is consistent with market efficiency. with the three perspectives in mind, does the switching strategy outperform the buy-andhold approach from 1986 through 2016? alternatively, which side of the literature do the results support? in addressing these questions, this study fits within the literature on the timing of security transactions, long of interest in the review. for example, early studies explore whether active stock trading outside january is worthwhile (mann and solberg, 1991), whether dollar-cost-averaging adds value (knight, 1992), and whether it is profitable to trade equity mutual funds (radcliffe, 1992). later studies explore tax harvesting (smithhwang and smith, 2008), batch trades (johnson and newman, 2012), and turn-of-the-month effects (chen, shin, and sun, 2015). 3. mutual funds and data the attention is on individual investors in u.s. markets, given the focus of the stock trader’s almanac and the well-documented “home bias” of u.s. investors (coval and moskowitz, 1999; li and zhao, 2016). the approach follows in the spirit of sullivan et al. (2001) and irwin and park (2007) by avoiding data mining and data-snooping biases as much as possible. it also follows the recommendation of jones and lundstrgum (2009) by beginning the analysis at the 1986 publication date and extending it through october 2016. in the spirit of jones and lundstrum (2009), who take a critical stand against indices that individual investors can never realistically own, such as the s&p 500 or the russell 2000, four mutual funds are in focus. each has a history dating back to at least 1986 and would be available to investors during the study period. as a move toward completeness and representation across categories, the four mutual funds encompass large-cap stocks, small-cap stocks, and mid-cap stocks. they also include a balanced fund to determine if the “sell-inmay” effect holds for investors who pursue a blended investment strategy of stocks and bonds. the four funds are as follows: vanguard 500 index fund (“large-cap fund”) clearbridge small-cap fund (“small-cap fund”) dreyfus mid-cap fund (“mid-cap fund”) vanguard wellesley income fund (“balanced fund”) 307a. loviscek, a. broder / financial services review 27 (2018) 303-322 the choice of funds, while realistic and prudent, is not random. it is influenced by the availability of data. because the small-firm effect was not widely acknowledged until the 1980s, there are few small-cap and mid-cap funds that have existed from 1986 through 2016. that said, we cannot rule out some degree of survivorship bias, especially regarding the small-cap and mid-cap funds, each of which is actively managed, as opposed to the passively-managed large-cap and balanced funds. to be as objective as possible, in the spirit of harvey, liu, and hequing (2016), with the individual investor in mind, the choice of the small-cap and mid-cap funds is random from the website of fidelity investments (2016). with the exception of the large-cap fund, the choice of the other funds largely confines generalizations to them. whether the results hold for similar-styled mutual funds that began at later dates and still exist today is an empirical issue left to future research. from the perspective of individual investors, the data are “friendly.” they come from the finance section of yahoo, where historical price data, incorporating stock splits and dividends, are readily found and are easily downloaded to compute the respective rates of return for each fund. rates of return are compounded daily for each of them. aligned with the view of jones and lundstrum (2009), the data from academic sources, such as from crsp or compustat, are not used, because the average active investor does not have access to them. in step with the literature, three-month u.s. treasury bill returns apply for the period may through october. they come from the historical series of the federal reserve bank of chicago. three points guide the three perspectives. first, while brokerage fees can vary widely across financial advisory firms, the variety of low-cost, discount brokerage fee services readily available leads us to assume that brokerage fees have a minimal impact on fund performance. second, the results from the switching strategy, both in terms of returns and risk-adjusted returns, are compared with those of the traditional buy-and-hold approach. third, as described in more detail in the next section, the wilcoxson signed-rank test, which is designed for comparative results of the kind in this study, is used to test the effectiveness of the switching strategy. these points, along with the two contrasting strands of literature, motivate the following hypothesis to be tested: the switching strategy, as drawn from the stock trader’s almanac, does not outperform the buy-and-hold strategy annually across the four mutual funds. the next section presents the results with three subsections, one for the switching strategy, one for the bear-bull analysis, and one for the tax considerations. additional subsections under the switching strategy cover annual returns and annual risk-adjusted returns based on the sharpe portfolio performance ratio. the last section concludes the study. 4. results to further motivate the analysis, the examination first centers on the performance of the switching strategy relative to that of the wilshire 5000, the broadest of all domestic market indices. the results are presented in table 1, which includes means, medians, maximums, minimums, and volatilities, as measured by annual standard deviations. from a return perspective, the switching strategy outperforms the buy-and-hold strategy on nine occasions, or in 30% of the years. the buy-and-hold approach registers a higher 308 a. loviscek, a. broder / financial services review 27 (2018) 303-322 mean return, at 12.33%, than the 10.63% for the switching strategy, a 16% difference. the volatility of the buy-and-hold approach, however, is 16.07%, or 63% higher than that of the switching strategy at 9.86%. these results indicate that the additional return earned from holding an index fund that tracks the wilshire 5000 does not fully compensate for the additional risk incurred compared with that of the switching strategy. this difference is an incentive for individual investors to search for an alternative to the buy-and-hold strategy. is the “sell in may” approach a worthy alternative? to answer this question, the analysis begins with an overview of the returns and volatilities for both strategies. 4.1. first perspective: returns and volatilities the starting date, following the publication of the pattern for the first time in the stock trader’s almanac (1986), is november 1, 1986. the ending date is october 31, 2016. the switch into three-month u.s. treasury bills occurs from may 1 through october 31 of a given year. a summary of the results is in table 2 for both the switching strategy and the buy-and-hold approach. to motivate statistical tests of the difference in the results between the two strategies, in terms of salient features, for the large-cap fund, the switching strategy outperforms the buy-and-hold approach on a return basis, which is a periodic stochastic process that relies on the arithmetic mean, in 8 of the 30 years.2 second, at 7.90%, the buy-and-hold mean at 10.66% is 276 basis points higher, or 34%, and the larger median difference is 643 basis points. third, the biggest difference in return performance by year occurs during the global financial crisis, 2,429 basis points, or �36.13 versus the switching strategy at �11.84%. fourth, and similar to the results found with the wilshire 5000, the mean and median return differences favoring the buy-and-hold approach come at the expense of significantly higher volatility, or 15.77% versus 9.08%, a 66% increase. given the difference in returns, this signals that the additional return earned from the buy-and-hold approach does not fully table 1 summary statistics of switching from holding the wilshire 5000 from november 1 of each year to april 30 of the next year to holding three-month u.s. treasury bills from may 1 to october 31 of each year wilshire 5000 no. of switching outperform years 9 years percentage of time (over 30 years) 30.00% mean return (switching) 10.63% mean return (buy-and-hold) 12.33 median return (switching) 11.94 median return (buy-and-hold) 15.26 max return (switching) 32.77 min return (switching) �10.24 max return (buy-and-hold) 39.41 min return (buy-and-hold)) �34.94 volatility (�) (switching) 9.86 volatility (�) (buy-and-hold) 16.07 included are the means, medians, maximums, minimums, and standard deviations for the switching strategy and the buy-and-hold strategy. 309a. loviscek, a. broder / financial services review 27 (2018) 303-322 compensate for the additional risk incurred. while the “what if” claim that the buy-and-hold strategy would have shown more positive results had the financial crisis of 2008–2009, a rare event that is essentially an outlier, not occurred, the position in this study is “what occurred is what occurred.” for the small-cap fund, the switching strategy outperforms the buy-and-hold approach on a return basis in 15 of the 30 years. while the buy-and-hold approach registers the highest return during the high-return period of november through april, at 57.51%, it also incurs the largest loss, as seen in the 52.90% drop during the global financial crisis. these results are not particularly surprising when viewed in the light of the volatility of small-cap stocks. historically, according to ibbotson & associates (2018), it averages 23%, and when compared with approximately 17% for large-cap stocks, the expectation is for a larger low-high range. there is also a difference of 108 basis points between the median returns, and at over 40% less volatility, additional evidence that the buy-and-hold approach is not compensating investors sufficiently for the additional risk incurred. the mid-cap fund switching strategy beats the buy-and-hold approach in 11 of the 30 years. as with the large-cap and small-cap funds, the switching strategy carries much lower volatility than the buy-and-hold approach, 10.93% versus 19.33%. this is yet more evidence of compensation falling short of that needed to cover the additional risk for the buy-and-hold strategy. the balanced fund displays a large difference in mean and median returns between the two strategies, at 4.62% versus 9.00%, a difference of 438 basis points for the means, with the spread in the medians of 449 basis points. the switching strategy outperforms the buy-and-hold method in five of the 30 years. it can be argued that this fund, while passively managed, already contains a risk-reduction element in the approximately 60% of fixedincome securities it has historically held. while the switching strategy lowers the risk of the fund from 8.19% to 3.86%, the reduction comes with a much lower return, 4.62% versus 9.00%. table 2 summary statistics of switching from holding the four mutual funds from november 1 of each year to april 30 of the next year to holding three-month u.s. treasury bills from may 1 to october 31 of each year large-cap fund small-cap fund mid-cap fund balanced fund no. of switching outperform years 8 years 15 years 11 years 5 years percentage of time (over 30 years) 26.67% 50.00% 36.67% 16.67% mean return (switching) 7.90% 11.78% 10.38% 4.62% mean return (buy-and-hold) 10.66 12.85 11.48 9.00 median return (switching) 8.04 10.50 11.31 4.29 median return (buy-and-hold) 14.47 9.50 16.00 8.78 max return (switching) 25.98 43.24 29.77 12.94 min return (switching) �11.84 �23.85 �11.93 �5.30 max return (buy-and-hold) 33.10 57.51 49.40 22.07 min return (buy-and-hold) �36.13 �52.90 �46.22 �14.25 volatility (�) (switching) 9.08 13.73 10.93 3.86 volatility (�) (buy-and-hold) 15.77 24.33 19.33 8.19 included are the means, medians, maximums, minimums, and volatilities, as measured by annual standard deviations, for the switching strategy and the buy-and-hold strategy. 310 a. loviscek, a. broder / financial services review 27 (2018) 303-322 4.1.1. yearly returns: switching strategy versus buy-and-hold approach for a detailed look at the summary annual rates of return in table 2, table 3 corresponds to the switching strategy while table 4 refers to the buy-and-hold approach. the respective returns provide a means to test statistically the effectiveness of the switching strategy as an alternative to the buy-and-hold approach on a return basis for each of the four funds. as expected from their respective compositions, the balanced fund is the steadiest performer while the small-cap fund shows the most fluctuation. to test for differences, the nonparametric wilcoxon signed-rank test comes into focus. anderson and loviscek (2005), derrick and white (2017), and higgins and peterson (1998) show its power for pairwise tests. moreover, it is more robust with respect to outliers (e.g., the large drop in stock returns during the financial crisis of 2008–2009) and heavy-tail distributions than the “t” test. under the null hypothesis that no difference exists in the average signed ranks of the returns between the switching strategy and the buy-and-hold strategy leads to the following “z” statistics and “p-values” for each of the four funds: table 3 this table provides the annual rates of return for the switching strategy for each of the four funds, november 1, 1986 to october 31, 2016 large-cap fund small-cap fund mid-cap fund balanced fund 1986 11.42% 18.93% 19.98% �1.16% 1987 6.04 20.74 22.57 6.64 1988 10.88 21.59 15.61 6.38 1989 �0.58 �5.34 0.00 �1.10 1990 25.98 41.66 29.77 12.94 1991 7.55 5.71 5.46 3.98 1992 6.82 9.09 11.58 10.20 1993 �2.00 �0.68 �1.93 �5.30 1994 10.94 2.05 4.97 9.41 1995 14.26 25.10 15.73 4.11 1996 15.20 5.60 �0.61 4.38 1997 22.93 13.67 12.66 9.65 1998 22.87 43.24 25.59 3.84 1999 7.82 6.23 28.90 0.37 2000 �11.84 �3.92 �7.73 8.83 2001 2.39 28.97 19.20 3.59 2002 4.50 11.91 1.48 5.13 2003 6.32 4.03 7.25 4.19 2004 3.54 6.63 3.84 3.85 2005 10.05 12.73 14.68 4.15 2006 8.95 17.17 11.56 6.09 2007 �9.56 �23.85 �11.93 �1.13 2008 �8.50 3.53 �3.83 0.99 2009 15.64 27.78 21.21 7.67 2010 16.28 16.87 20.81 5.98 2011 12.71 13.30 11.06 6.72 2012 14.33 18.77 21.15 6.55 2013 8.26 6.09 7.08 5.51 2014 4.35 3.16 7.55 2.96 2015 �0.65 2.62 �2.15 3.24 for example, 11.42% represents the rate of return on the large-cap fund from november 1, 1986 to april 30, 1987, followed by the switch into three-month u.s. treasury bills from may 1, 1987 to october 31, 1987. 311a. loviscek, a. broder / financial services review 27 (2018) 303-322 the results indicate that the buy-and-hold strategy for the large-cap fund, which registers higher returns than the switching strategy in 22 of the 30 years, and that for the balanced fund, which outperforms it in 25 of the 30 years, display significant results at the 5% level. no significant difference appears to exist for the small-cap and mid-cap funds. thus, at this level of analysis, the buy-and-hold approach appears preferable to the switching strategy for the large-cap and balanced funds for investors who can tolerate the higher risks of not switching to u.s. treasury bills. for investors holding the small-cap and mid-cap tests of return differences: switching strategy vs. buy-and-hold approach large-cap: z � �1.84; p � 0.03 small-cap: z � �0.31; p � 0.37 mid-cap: z � �0.84; p � 0.20 balanced: z � �3.61; p � 0.00 table 4 this table provides the annual rates of return for the buy-and-hold strategy for each of the four funds, november 1, 1986 to october 31, 2016 large-cap fund small-cap fund mid-cap fund balanced fund 1986 �2.01% �17.54% �8.13% �3.93% 1987 14.51 20.45 23.47 15.53 1988 22.96 37.60 28.07 17.05 1989 �7.62 �14.68 �14.03 0.90 1990 33.10 53.06 49.40 22.07 1991 9.78 8.05 6.22 11.27 1992 14.75 32.80 27.53 19.40 1993 3.72 �6.20 1.20 �5.36 1994 26.27 9.01 20.53 21.68 1995 23.98 26.01 17.89 12.87 1996 32.00 33.93 26.78 17.84 1997 21.92 �5.61 �15.14 13.48 1998 25.66 57.51 30.65 0.32 1999 6.21 9.06 31.07 8.71 2000 �25.00 �24.95 �21.40 12.01 2001 �15.16 9.15 �0.37 2.10 2002 20.63 52.26 25.41 8.72 2003 9.25 2.08 12.48 8.84 2004 8.60 18.59 16.58 4.80 2005 16.18 8.34 16.70 10.86 2006 14.43 10.82 8.38 8.41 2007 �36.13 �52.90 �46.22 �14.25 2008 9.86 38.93 15.41 17.53 2009 16.39 25.75 19.98 13.50 2010 7.91 �8.96 4.52 8.15 2011 15.05 9.46 10.67 11.72 2012 26.97 33.37 34.93 7.98 2013 17.07 10.97 17.12 8.55 2014 5.07 �0.26 6.79 3.26 2015 3.30 9.53 �2.00 5.89 for example, �2.01% represents the rate of return on the large-cap fund from november 1, 1986 to october 31, 1987. 312 a. loviscek, a. broder / financial services review 27 (2018) 303-322 funds, the results suggest no difference in the results from using the switching strategy or holding the funds for the entire 12 months. 4.1.2. switching and buy-and-hold strategies: risk-adjusted returns as table 2 shows, however, the higher returns come at the price of higher risk. as a result, table 5 and table 6 illustrate annual sharpe ratios for the switching strategy and the buy-and-hold approach, respectively. the smallest sharpe ratios are in the recession years of the first decade of the 21st century, as low as �1.87 for the large-cap fund, with the largest ratios in the post-financial crisis period, especially in 2012, with a high of 3.21 for the small-cap fund. between the two tables, the means are higher for the large-cap and balanced funds but smaller for the small-cap and mid-cap funds. the difference in the means of the balanced fund between the two strategies is table 5 this table illustrates annual sharpe ratios for each fund for the switching strategy, november 1, 1986 to october 31, 2016 large-cap fund small-cap fund mid-cap fund balanced fund 1986 0.29 0.96 0.45 �0.62 1987 �0.03 1.20 1.01 0.02 1988 0.23 1.29 0.45 �0.27 1989 �1.52 �1.13 �0.50 �2.03 1990 2.14 2.13 0.78 1.50 1991 0.37 0.21 0.14 0.11 1992 1.27 0.55 0.65 1.26 1993 �0.78 �0.48 �0.22 �1.00 1994 0.74 �0.52 �0.06 0.58 1995 1.73 2.79 0.30 �0.16 1996 0.77 0.05 �0.39 �0.06 1997 2.29 0.81 0.35 0.47 1998 1.93 1.99 1.23 �0.07 1999 0.17 0.03 0.73 �0.42 2000 �1.04 �0.35 �0.16 0.97 2001 0.10 1.28 1.24 0.25 2002 0.26 0.67 0.04 0.58 2003 0.88 0.32 0.99 0.89 2004 0.09 0.34 0.09 0.24 2005 1.23 0.36 1.14 �0.08 2006 0.79 0.62 1.01 0.22 2007 �1.18 �1.43 �1.24 �0.61 2008 �0.42 0.12 �0.19 0.11 2009 1.76 3.11 2.12 1.83 2010 2.22 1.94 2.97 1.39 2011 1.93 1.40 1.32 1.85 2012 2.24 2.21 3.02 1.30 2013 1.28 0.73 0.87 1.07 2014 0.56 0.13 0.99 0.59 2015 �0.14 0.14 �0.26 0.44 mean 0.67 0.72 0.63 0.35 for example, 0.29 represents the sharpe ratio for the large-cap fund from november 1, 1986 to october 31, 1987, with the switch into three-month u.s. treasury bills occurring from may 1, 1987 to october 31, 1987. all successive years follow this pattern. annual u.s. treasury bill rates of return measure the risk-free rates of return and annual standard deviations of the fund returns measure the risks. 313a. loviscek, a. broder / financial services review 27 (2018) 303-322 relatively large, 0.84 versus 0.35. to test for the differences between the two tables, the “z” and “p-value” statistics, according to the wilcoxon signed-rank test, are as follows: at the 5% level, only the balanced fund registers a statistically significant difference, and one that favors the buy-and-hold approach. for perspective with the large-cap fund, in a series of simulations with the standard deviation of the switching strategy returns held table 6 this table illustrates annual sharpe ratios for each fund for the buy-and-hold strategy, november 1, 1986 to october 31, 2016 large-cap fund small-cap fund mid-cap fund balanced fund 1986 �0.25 �0.70 �0.32 �0.80 1987 0.54 0.89 0.92 0.81 1988 0.97 1.97 0.95 1.01 1989 �0.97 �1.37 �0.98 �0.73 1990 2.17 2.81 1.37 2.15 1991 0.50 0.32 0.16 1.03 1992 1.92 2.67 1.58 2.63 1993 �0.02 �0.70 �0.10 �0.81 1994 2.52 0.26 1.05 1.92 1995 2.19 1.55 0.34 1.03 1996 1.50 1.73 1.12 0.97 1997 0.83 �0.31 �0.54 0.76 1998 1.54 1.67 1.32 �0.40 1999 0.03 0.18 0.59 0.23 2000 �1.52 �0.96 �0.33 1.24 2001 �0.80 0.24 �0.09 0.04 2002 1.32 2.45 1.64 1.08 2003 1.14 0.06 1.11 1.89 2004 0.69 1.06 1.11 0.44 2005 1.71 0.14 0.96 0.92 2006 1.21 0.27 0.43 0.51 2007 �1.87 �1.60 �1.67 �1.69 2008 0.42 1.02 0.64 1.63 2009 0.87 1.14 1.07 2.00 2010 0.45 �0.37 0.26 1.10 2011 1.46 0.69 0.77 2.48 2012 2.84 3.21 3.09 1.10 2013 2.14 0.80 1.61 1.41 2014 0.34 �0.02 0.50 0.40 2015 0.29 0.57 �0.21 0.80 mean 0.81 0.66 0.61 0.84 for example, �0.25 represents the sharpe ratio for the large-cap fund from november 1, 1986 to october 31, 1987. all successive years follow this pattern. annual three-month u.s. treasury bill rates of return measure the risk-free rates of return and annual standard deviations of the fund returns measure the risks. tests of sharpe ratio differences: switching strategy vs. buy-and-hold approach large-cap: z � �1.22; p � 0.11 small-cap: z � �0.44; p � 0.33 mid-cap: z � �0.03; p � 0.49 balanced: z � �3.47; p � 0.00 314 a. loviscek, a. broder / financial services review 27 (2018) 303-322 constant at 9.08%, as taken from table 2, the mean value of the sharpe ratios has to increase to at least 1.04 and have sharpe ratios that exceed those of the buy-and-hold approach in at least 22 of the 30 years to outperform statistically the buy-and-hold approach. this is a large increase from the mean of 0.67, indicating the difficulty of outperforming a broad market index. although the evidence indicates that the switching strategy does not outperform the buy-and-hold approach of the large-cap, small-cap, and mid-cap funds, it also indicates that the buy-and-hold approach for each of these funds does not offer superior performance. as a result, investors who have been drawn to the switching strategy have probably not realized significant risk-adjusted return underperformance with respect to these three funds. it is conceivable that other small-cap and mid-cap funds, including etfs, offer superior switching strategy effectiveness. finding these funds, however, given the body of evidence supporting market efficiency, is unlikely to be easy. the balanced fund’s results signal that investors drawn to the switching strategy would be served well to combine this fund with the buy-and-hold approach. it is the only fund that appears to compensate for the additional risk of not switching into u.s. treasury bills. so far, the results indicate that while the buy-and-hold strategy falls short of compensating investors for the additional risk incurred from may 1 to october 31, the switching strategy does not outperform it. they lead to the conclusion not to reject the hypothesis under test; namely, that the switching strategy does not outperform the buy-and-hold strategy, as drawn from the stock trader’s almanac (2016). the evidence is in contrast to the results in previous studies, such as in andrade et al., (2013), borowski (2015), and bouman and jacobsen (2002). overall, the findings here align with those of, for example, dichtl and drobetz (2015), lucey and zhao (2008), and jones and lundstrum (2009). their results lead them to question the effectiveness of following the “sell in may” observation. collectively, they conclude that the results supporting it use time-specific data that do not fully align with the publication dates of the observation, and that u.s. markets have largely adjusted to it. 4.2. second perspective: bear-bull analysis for the second perspective, there are two bear markets and two bull markets from december 1999 through october 2016. in all four instances, price changes exceed 20%. while 20% is somewhat arbitrary, it is the standard benchmark (thestreet.com, 2018). given the small number of observations in each case, only descriptive statistics are reported. moreover, given the uniqueness in bear and bull markets, caution is in order when making generalizations. the first bear market is from march 2000 through october 2002. within this span, there are three switching opportunities: november 1999 to april 2000, followed by the may–october switch; november 2000 to april 2001, followed by the may–october switch; and november 2001 to april 2002, followed by the may–october switch. the analysis for the first bear market is provided in table 7. for the large-cap fund, the switching strategy shows a return of 7.82% from november 1999 to october 2000, outperforming the buy-and-hold approach by 161 basis points. (although not shown in the table, the large-cap fund registers a decline of over 6% in the first quarter of 1999, followed by a rise of over 10% by the third quarter of 2000 before continuing its descent into the fourth quarter and into 2002.) given the returns of �11.84% 315a. loviscek, a. broder / financial services review 27 (2018) 303-322 and 2.39% in the following periods, compared with �25.00% and �15.16% for the buy-and hold approach—a mean difference of over 1000 basis points across the three switching periods—the switching strategy was an effective hedge during this bear market. a similar impact occurs with the small-cap and mid-cap funds. in fact, they display double-digit gains during the period november 2001–october 2002 with the switching strategy. the balanced fund posts gains throughout the bear market, with the buy-and-hold approach registering increases that are greater overall than those of the switching strategy. all told, the switching strategy helped to limit losses during this period, at least with respect to the four funds, compared with the buy-and-hold method, expectedly so given the “riskfree” attributes of u.s. treasury bills. table 8 represents the rates of return on the four funds in the second bear market. it records two switching strategies. each fund posts losses during the switching period, with the small-cap fund registering the largest, at �52.90%, and the balanced fund displaying the smallest, at �1.13%. the difference in losses between the switching and buy-and-hold strategies is striking at over 7,800 basis points during the first switching strategy, with the mid-cap fund registering the largest difference at over 3,400 basis points. when combined table 7 this table provides the results for each of the four funds during the first bear market of the 21st century, or from march 2000 through october 2002 large-cap fund small-cap fund mid-cap fund balanced fund nov. 1999 to apr. 2000 may 2000 to oct. 2000 total return (switching) 7.82% 6.23% 28.90% 0.37% total return (buy-and-hold) 6.21 9.06 31.07 8.71 nov. 2000 to apr. 2001 may 2001 to oct. 2001 total return (switching) �11.84 �3.92 �7.73 8.83 total return (buy-and-hold) �25.00 �24.95 �21.40 12.01 nov. 2001 to apr. 2002 may 2002 to oct. 2002 total return (switching) 2.39 28.97 19.20 2.59 total return (buy-and-hold) �15.16 9.15 �0.37 2.10 three switching periods occur during this time. table 8 this table provides the results for each of the four funds during the second bear market of the 21st century, or from october 2007 through march 2009 large-cap fund small-cap fund mid-cap fund balanced fund nov. 2007 to apr. 2008 may 2008 to oct. 2008 total return (switching) �9.56% �23.85% �11.93% �1.13% total return (buy-and-hold) �11.23 �52.90 �46.22 �14.25 nov. 2008 to apr. 2009 may 2009 to oct. 2009 total return (switching) �8.50 3.53 �3.83 0.99 total return (buy-and-hold) 9.86 38.93 15.41 17.53 two switching periods occur during this time. 316 a. loviscek, a. broder / financial services review 27 (2018) 303-322 with the results in table 7, the switching strategy helped to limit losses during this period and may serve as a guide in future bear markets. table 9 displays the results from the first bull market of the 21st century, with four switching strategies occurring. all returns are positive. the buy-and-hold strategy records higher returns in 12 of the 16 comparisons, displaying double-digit returns in eight instances compared with the five for the switching strategy. for the small-cap and mid-cap funds, the buy-and-hold approach registers impressive double-digit gains in the second switching period, or 18.59% and 16.58%, respectively, with the mid-cap fund recording a higher gain by 1,274 basis points. both the buy-and-hold large-cap and balanced funds show higher returns in all four switching periods. table 10 illustrates the returns across the four funds during the second bull market, or from march 2009 through october 2016. across all 32 returns, the buy-and-hold outperforms the switching strategy 23 times. the small-cap fund shows a jump of 38.93% in the immediate post-financial crisis period, followed by a return of 25.75% in the next period. together, they help to offset most (but not all) of the loss of 52.90% during the bear market of 2007–2009. although it does not post the highest returns, the steadiest performer during this period is the balanced fund, registering double-digit gains on three occasions, with the highest at 17.53%. however, the degree of risk tolerance of the investor comes into play, considering the higher risk of the buy-and-hold approach. these results suggest that the switching strategy can be a worthy alternative to the buy-andhold approach during risk-off periods, at least with respect to the four mutual funds, tempering somewhat the failure to reject the hypothesis under test. because the strategy calls for a move into u.s. treasury bills during the flat-to-down months of the year, it shows the increase in portfolio performance through risk reduction. as a result, individual investors may want to “buy and hold” table 9 this table provides the results of the four funds during the first bull market of the 21st century, or from november 2003 through october 2007 large-cap fund small-cap fund mid-cap fund balanced fund nov. 2003 to apr. 2004 may 2004 to oct. 2004 total return (switching) 6.32% 4.03% 7.25% 4.19% total return (buy-and-hold) 9.25 2.08 12.48 8.84 nov. 2004 to apr. 2005 may 2005 to oct. 2005 total return (switching) 3.54 6.63 3.84 3.85 total return (buy-and-hold) 8.60 18.59 16.58 4.80 nov. 2005 to apr. 2006 may 2006 to oct. 2006 total return (switching) 10.05 12.73 14.68 4.15 total return (buy-and-hold) 16.18 8.34 16.70 10.86 nov. 2006 to apr. 2007 may 2007 to oct. 2007 total return (switching) 8.95 17.17 11.56 6.09 total return (buy-and-hold) 14.43 10.82 8.38 8.41 four switching periods occur during this time. 317a. loviscek, a. broder / financial services review 27 (2018) 303-322 during bullish periods and use the switching strategy during bearish periods. however, this comes with the caveat that some market timing is in order, a difficult practice for individual investors, as barber and odean (2000) and barber et al. (2001) demonstrate. 4.3. third perspective: tax considerations for the third perspective, investors need to be circumspect of the impact of taxes on their returns. while this consideration is obvious, barber and odean (2004) find that investors trading activities undercut their after-tax returns, leading to subpar after-tax performance. by realizing gains faster than losses, it seems they do not capitalize fully on tax-avoidance strategies in their quest to earn high returns, often failing to defer the realization of capital gains to lower rates in the united states. as insight into this issue, table 3, which provides the annual returns for the switching strategy for each of the four funds, becomes the focus. table 10 the table provides the results across the four funds during the second bull market of the 21st century, or from march 2009 through october 2016 large-cap fund small-cap fund mid-cap fund balanced fund nov. 2008 to apr. 2009 may 2009 to oct. 2009 total return (switching) �8.50% 3.53% �3.83% 0.99% total return (buy-and-hold) 9.86 38.93 15.41 17.53 nov. 2009 to apr. 2010 may 2010 to oct. 2010 total return (switching) 15.64 27.78 21.21 7.67 total return (buy-and-hold) 16.39 25.75 19.98 13.50 nov. 2010 to apr. 2011 may 2011 to oct. 2011 total return (switching) 16.28 16.87 20.81 5.98 total return (buy-and-hold) 7.91 �8.96 4.52 8.15 nov. 2011 to apr. 2012 may 2012 to oct. 2012 total return (switching) 12.71 13.30 11.06 6.72 total return (buy-and-hold) 15.05 9.46 10.67 11.72 nov. 2012 to apr. 2013 may 2013 to oct. 2013 total return (switching) 14.33 18.77 21.15 6.55 total return (buy-and-hold) 26.97 33.37 34.93 7.98 nov. 2013 to apr. 2014 may 2014 to oct. 2014 total return (switching) 8.26 6.09 7.08 5.51 total return (buy-and-hold) 17.07 10.97 17.12 8.55 nov. 2014 to apr. 2015 may 2015 to oct. 2015 total return (switching) 4.35 3.16 7.55 2.96 total return (buy-and-hold) 5.07 �0.26 6.79 3.26 nov. 2015 to apr. 2016 may 2016 to oct. 2016 total return (switching) �0.65 2.62 �2.15 3.24 total return (buy-and-hold) 3.30 9.53 �2.00 5.89 eight switching periods occur during this time. 318 a. loviscek, a. broder / financial services review 27 (2018) 303-322 the maximum marginal income tax rate is applied on short-term capital gains to each of the positive returns, with changes noted in the maximum rate, beginning with 38.5% in 1987 (that is the tax rate applied to gains from november 1, 1986 through october 31, 1987). there is a drop to 31% between 1988 and 1991 before an increase to 39.6% from 1993 to 2000. a drop to 35% follows from 2003 to 2012 before the rate increases back to 39.6% through 2016. (for ease of calculation, the medicare tax surcharge, the phase-out of partial itemized deductions, carryover losses, and state and local income taxes are set aside.) because the tax rates lower the returns in table 3, and because the statistical evidence favors the buy-and-hold approach (that can apply for many years) with large-cap and balanced funds, the switching strategy should not be expected to outperform its alternative, at least on a return basis. the wilcoxon signed-rank test relative to the returns on the buy-and-hold results listed in table 4 leads to the following results: the statistics indicate outperformance by the buy-and-hold approach. the implication is that, particularly in bull markets, investors should use the switching strategy in a taxsheltered account if it is to be a satisfactory alternative to the buy-and-hold method. it aligns more closely with the preferential treatment given to long-term capital gains in the united states than a non-sheltered tax account. 5. conclusion with the focus on individual investors, this study offers three perspectives on the quote “sell in may and go away.” the data apply from november 1, 1986, the publication year of the quote, to october 31, 2016, as drawn from the stock trader’s almanac (2016), centered on four mutual funds. they represent large-cap equity (vanguard 500 index), small-cap equity (clearbridge), mid-cap equity, (dreyfus), and balanced equity-fixed income (vanguard wellesley income). first, each of the four funds is held from november 1 through april 30, followed by a switch to u.s. treasury bills for the remaining months. the performance on a return basis and on a risk-adjusted return basis is compared with the performance of the buy-and-hold approach. although the results suggest that the buy-andhold approach does not fully compensate investors for the increase in risk incurred from may 1 through october 31, statistical tests do not lead to the rejection of the hypothesis under test; namely, that the switching strategy does not outperform the buy-and-hold strategy. this finding aligns with the body of literature that questions the effectiveness of the statement “sell in may and go away” as a market-beating method. as recent research demonstrates, tests of after-tax return differences: switching strategy vs. buy-and-hold approach large-cap: z � �2.62; p � 0.00 small-cap: z � �1.63; p � 0.05 mid-cap: z � �1.94; p � 0.03 balanced: z � �3.96; p � 0.00 319a. loviscek, a. broder / financial services review 27 (2018) 303-322 evidence favoring the strategy in prior studies may not only be time-specific, but it also may no longer hold, as u.s. markets have largely adjusted to it. the findings from the balanced fund may be a draw for individual investors. the buy-and-hold approach, given the fund’s composition of equities and fixed-income securities, registers statistically superior performance, both on a return basis and on a risk-adjusted return basis, compared with that of the switching strategy. second, this study examines the performances of the two bear markets and the two bull markets that occurred from 2000 through 2016 to test the switching strategy during each period. with caution about making generalizations given the uniqueness of each bear and bull market, the switching strategy registers superior return performance over the buy-andhold approach during these bear markets, while the buy-and-hold approach offers superior return performance during bullish periods. although the results from the bear markets are not strong enough to reject the null hypothesis under test, they signal that individual investors might want to consider switching to u.s. treasury bills during flat-to-bearish times but move to the buy-and-hold approach during bullish periods. this requires, however, accurate market timing of turning points, a difficult task. third, this study provides some concise statistical evidence that the application of the switching strategy should be in a tax-sheltered account if the switching strategy is to at least keep pace with the buy-and-hold method. this is because the gains from the switching strategy are short-term, which are taxed currently at ordinary income rates in the u.s. preferential tax treatment is given to long-term gains earned from assets held for over one year, which aligns more closely with the buy-and-hold approach than with a short-term trading strategy. notes 1 we are grateful to an anonymous referee for these clarifications. 2 we are grateful to an anonymous referee for these clarifications. references anderson, r. i., & loviscek, a. l. (2005). searching for information content: portfolio performance of the 100 best stocks to own in america. financial services review, 14, 197–209. anderson, j., & smith, g. (2006). a great company can be a great investment. financial analysts journal, 62, 86–93. andrade, s., chhaochharia, v., & fuerst, m. e. (2013). “sell in may and go away” just won’t go away. financial analysts journal, 69, 94–105. aronson, d., & masters, t. (2006). evidence-based technical analysis. new york, ny: wiley. barber, b., lehavy, r., mcnichols, m., & trueman, b. (2001). can investors profit from the prophets? analyst recommendations and stock returns. journal of finance, 56, 531–563. barber, b., & odean, t. (2000). trading can be hazardous to your wealth: the common stock performance of individual investors. journal of finance, 55, 773–806. barber, b., & odean, t. (2004). are individual investors tax savvy? evidence from retail and discount brokerage accounts. journal of public economics, 88, 419–442. 320 a. loviscek, a. broder / financial services review 27 (2018) 303-322 bessembinder, h., & chan, k. (1998). market efficiency and the returns to technical analysis. financial management, 27, 5–17. borowski, k. (2015). analysis of sell-in-may-and-go-away among 122 equity indices and 39 commodities. international journal of economics and finance, 7, 119–129. bouman, s., & jacobsen, b. (2002). the halloween indicator, ‘sell in may and go away’: another puzzle. american economic review, 92, 1618–1635. brecht, k. (2016). u.s. news and world report. (available at http://money.usnews.com/investing/articles/201604-26/why-you-should-not-sell-in-may-and-go-away). buffet, w. (2008). be fearful when others are greedy and greedy when others are fearful. interview with charlie rose, pbs (october 1, 2008). chang, e. c., & lewellen, w. c. (1984). market timing and u.s. investment fund performance. journal of business, 57, 57–72. chen, h., shin, h. s., & sun, x. (2015). return-enhancing strategies with international etfs: explaining the turn-of-the-month effect. financial services review, 24, 271–288. coval, j., & moskowitz, t. (1999). home bias at home: local equity preference in domestic portfolios. journal of finance, 54, 2045–2073. derrick, b., & white, p. (2017). comparing two samples from an individual likert question. international journal of mathematics and statistics, 18, 1–13. dichtl, h., & drobetz, w. (2015). sell in may and go away: still good advice for investors? international review of financial analysis, 38, 29–43. fama, e. (1970). efficient capital markets: a review of theory and empirical work. journal of finance, 25, 383–417. fidelity investments. (2016). fidelity mutual funds. (available at https://fidelity.com/viewpoints/active-trader/ sell-in-may). harvey, c., liu, y., & heqing, z. (2016). … and the cross-section of expected returns. review of financial studies, 29, 5–68. higgins, e. j., & peterson, d. r. (1998). the power of one and two-sample t statistics given event-induced variance increases and non-normal stock returns: a comparative study. quarterly journal of business and economics, 37, 27–49. hirsch, y. (1986). stock trader’s almanac. new york, ny: reed business information. hirsh, y. (2016). stock trader’s almanac. new york, ny: reed business information. hooke, j. c. (2010). security analysis and business valuations on wall street: a comprehensive guide to today’s valuation methods (2nd. ed.). new york, ny: wiley ibbotson & associates (2018). classic yearbook 2018. chicago, il: morningstar. irwin, s., & park, c-h. (2007). what do we know about the profitability of technical analysis? journal of economic surveys, 21, 786–826. johnson, r. m., & newman, j. a. (2012). direct stock purchase plans using batch trades: do investors buy high and sell low? financial services review, 21, 227–237. jones, c. p., & lundstrum, l. l. (2009). is ‘sell in may and go away’ a valid strategy for u.s. equity allocation? the journal of wealth management, 12, 104–112. knight, j. r. (1992). nobody gains from dollar-cost-averaging: analytical, numerical, and empirical results. financial services review, 2, 51–61. kochman, l., & badarinathi, r. (2008). the halloween effect: an enduring market anomaly. finance india, 22, 443–446. kochman, l, badarinathi, r., & bray, d. (2014). sell in may and go away revisited. american economist, 59, 90–91. li, m., & zhao, x. (2016). neighborhood effect on stock price co-movement. north american journal of economics and finance, 35, 1–22. lo, a. w. (2017). adaptive markets: financial evolution at the speed of thought. princeton, nj: princeton university press. 321a. loviscek, a. broder / financial services review 27 (2018) 303-322 lucey, b. m., & zhao, s. (2008). halloween or january? yet another puzzle. international review of financial analysis, 17, 73–101. maberly, e. d., & pierce, r. m. (2004). stock market efficiency withstands another challenge: solving the “sell in may/buy after halloween” puzzle. economic journal watch, 1, 29–46. malkiel, b. j. (1995). returns from investing in equity mutual funds 1971 to 1991. journal of finance, 50, 549–572. malkiel, b. j. (2015). a random walk down wall street (11th ed.). new york, ny: w.w. norton. mann, s. v., & solberg, d. (1991). should individual investors avoid the stock market outside of january? financial services review, 1, 101–108. poterba, j. m., & summers, l. (1987). mean reversion in stock prices: evidence and implications. journal of financial economics, 22, 27–59. prentis, e. l. (2011). evidence on a new stock trading rule that produces higher returns with lower risk. international journal of economics and finance, 3, 92–104. radcliffe, r. (1992). active trading decisions of equity mutual funds. financial services review, 2, 21–39. schleifer, a., & vishny, r. (1997). the limits of arbitrage. journal of finance, 52, 35–55. sewell, m. (2012). the efficient market hypothesis: empirical evidence. international journal of statistics and probability, 1, 164–178. shilling, a. g. (1993). markets can remain irrational a lot longer than you and i can remain solvent. (quote from j.m. keynes, 1920) forbes, 151, 236. smith-hwang, m., & smith, g. (2008). harvesting capital gains and losses. financial services review, 17, 309–321. smith, g. (2016). great company, great investment returns revisited. journal of wealth management, 19, 34–39. sullivan, r., timmermann, a., & white, h. (2001). dangers of data mining: the case of calendar effects in stock returns. journal of econometrics, 105, 249–286. taylor, j. (1988). the trend is your friend. new york times magazine, 21, 26. thestreet.com (2018). what is a bear market and what does it mean in 2018? (available at https://www. thestreet.com/markets/what-is-a-bear-market-14713949). 322 a. loviscek, a. broder / financial services review 27 (2018) 303-322 academy of financial services officers president robert moreschi virginia military institute president-elect duncan williams western carolina university executive vice president-program swarn chatterjee university of georgia vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations claire matthews massey university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president william chittenden texas state university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama halil kiymaz rollins college frances lawrence louisiana state university tom potts baylor university janine scott massey university martin seay kansas state university past presidents thomas coe, 2015-2016 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $75 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. membership forms may be accessed at the journal website at http://www.academyfinancial.org. or for membership, subscription, and address change notification, please contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email: smichels@stetson.edu. editorial: authors should submit their papers electronically (word format, no pdfs please) as an e-mail attachment to the editor at smichels@stetson.edu. afs member submission fees are $50. the afs non-member submission fee is $125, which includes a one year membership to afs. concurrent with the submission, please pay online or mail a check (for us funds) payable to afs to stuart michelson at the address above. should a manuscript revision be invited, no additional fees will be required. style information for manuscripts is on the inside back cover of this journal. copyright © 2017 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 26, no. 1, 2017 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning ning tang san diego state university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy larry prather southeastern oklahoma state university insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university real estate international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price for the year 2014 is $100. personal subscription price for the year 2014 is $75 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2017 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. the impact of age differences and race on the social security early retirement decision for married couples: an extension with gender role reversals diane scott dockinga,*, rich fortinb, stuart michelsonc anorthern illinois university, department of finance, dekalb, il 60115, usa bnew mexico state university, msc 3 fin, box 30001, las cruces, nm 88003, usa cstetson university, school of business – unit 8398, 421 n. woodland boulevard, deland, fl 32723, usa abstract the purpose of this study is to examine the impact of age differences on the social security early and delayed retirement decision for married couples. this article extends the analysis of docking, fortin, and michelson’s 2015 study that assumed the working spouse (male) was older than the non-working spouse (female). in this current study we reverse the spouse and spouse employment role and ages. we now assume a working spouse (female) who is older than her non-working spouse (male). we analyze the nine married couple combinations for the following races: whites (w), hispanics (h), and blacks (b). we develop an excel model to compute the breakeven internal rate of return (be irr) for each of nine race combinations. three claiming scenarios are considered: receiving benefits early (e.g., at age 62 vs. 66), the maximum realistic delay period (e.g., at age 62 vs. 70), and delaying benefits past full retirement age (e.g., age 66 vs. 70). within these three claiming scenarios we examine couples by race combination who retire at the same age and at different ages, and with age differences of 0, 4, 7, and 10 years. we compare the results of the two studies. the primary substantive conclusions from this study depends on the age comparisons that are being made. for couples who retire at the same age or at different ages, the greater the age difference the greater the incentive to retire early as the hurdle rate is lower to overcome. this is true irrespective who is older and the breadwinner (earning spouse). women almost always have higher be irrs than men. the implication is that in marriages where the spouse is the breadwinner and the older partner, it is more difficult for the couple to retire early, as compared to marriages where the spouse is the breadwinner and the older partner. irrespective of who is the breadwinner, hispanics have higher hurdle rates, whereas whites have lower hurdle rates. for a given retirement age comparison or age difference the results can be interpreted as follows: the high (low) breakeven group would prefer to * corresponding author. tel.: �1-815-753-6396; fax: �1-815-753-0504. e-mail address: ddocking@niu.edu (d.s. docking) financial services review 26 (2017) 87–111 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. retire later (earlier) because the hurdle rate is more difficult (less difficult) to overcome. thus, hispanics have a more difficult time retiring early, whereas whites have a less difficult time retiring early. © 2017 academy of financial services. all rights reserved. keywords: social security; retirement; financial planning; retirement age 1. introduction the purpose of this study is to examine the impact of age differences on the social security early and delayed retirement decision for married couples. this article extends the analysis of docking, fortin, and michelson (2015), where we assumed the working spouse (male) is older than the non-working spouse (female). in this current study, we reverse the spouse and spouse employment role and ages. we now assume the working spouse (female) is older than the non-working spouse (male). this extended analysis is done for married couples by race. more specifically, we analyze the nine married couple combinations for the following races: whites (w), hispanics (h), and blacks (b). the nine spouse (male, m)/spouse (female, f) combinations are: wm_wf, bm_bf, hm_hf, wm_bf, bm_wf, wm_hf, hm_wf, bm_hf, and hm_bf. we develop an excel model to compute the breakeven irr for each of the nine race combinations. following blanchett (2013), three claiming scenarios are considered: receiving benefits early (e.g., at age 62 vs. 66), the maximum realistic delay period (e.g., at age 62 vs. 70), and delaying benefits past full retirement age (e.g., age 66 vs. 70). within these three claiming scenarios we examine couples by race combination who retire at the same age and with age differences of 0, 4, 7, and 10 years. we also look at a specific scenario where the spouses retire at different ages and the impact of age differences on their retirement decision. we then compare these results with the results of the docking, fortin, and michelson 2015 study. 2. literature review there has been an extensive number of studies on the early versus delayed social security retirement decision for married couples although none have explicitly addressed the age difference issue across race categories as this study does. for a review of prior literature, see docking, fortin, and michelson (2012, 2013). only a few studies have looked at the age difference between the spouses in determining the optimal retirement age. results have been mixed. coile, diamond, gruber, and jousten (2002) find that if the spouse is older than the spouse, then he should delay retirement to age 65; but if the spouse is five years older than her spouse, he should retire early at age 62. munnell and soto (2007) show that as the age difference between the spouses (spouse minus spouse) increase, the spouse should claim earlier (age 62) and the spouse should claim later (age 69). sun and webb (2009) show that if the spouse is three or more years older than her spouse, he should retire at 69 and she at 66. tucker (2009) says both should retire at age 62 no matter the age difference. mccormack and perdue (2006) assume the spouse is seven 88 d.s. docking et al. / financial services review 26 (2017) 87–111 years older than his spouse and the spouse has the higher earnings. they conclude that both should retire at age 62. blanchett (2013), examined three claiming scenarios for married couples: 62 versus 66, 62 versus 70, and 66 versus 70. they find that most retirees would be best served by delaying social security benefits (ssbs) until at least full retirement age (fra) or later. docking, fortin, and michelson (2013) look at the impact of race on the retirement decision for married couples of the same age who retire at the same time. they compute a breakeven (be) internal rate of return (irr) for each of nine race combinations from age 62 through age 70. if a couple’s opportunity cost of capital (that can be considered a hurdle rate) is greater than (less than) the computed be irr, the couple should retire at the earlier (later) age., that is, the greater (lower) the be irr, the more optimal for a couple to retire later (earlier). results are fairly uniformly consistent across the nine race combinations: be irrs for a given base age are, in general, monotonically decreasing compared with older ages. the highest be irrs are for couples with a hispanic spouse, and the lowest be irrs are for couples with a white spouse. they conclude that, from a given base age, it is generally more optimal to retire now with a longer time horizon since the hurdle rate is lower and later with a short time horizon because the hurdle rate is higher. docking, fortin, and michelson (2015) look at the impact of race and age difference on the retirement decision of married couples retiring at the same age and at different ages. in this study they assume the spouse is older than the spouse and is the only earner or breadwinner in the family. assuming couples retire at the same age, for the retirement ages 62 versus 66, the be irrs uniformly decrease as the age difference increases. this implies that the greater the age difference between couples, the earlier the couples should retire. be irrs range from a low of 5.0862% for the wm_bf combination with a 10-year age difference, to a high of 5.8151% for the hm_hf combination with a zero year age difference. for the retirement ages 62 versus 70, the be irrs uniformly increase as the age difference increases. this implies that the greater the age difference between couples, the later the couples should retire. however, the be irrs are much lower in this retirement decision, ranging from a low of 2.8949% for the wm_wf combination with a zero year age difference, to a high of 3.8750% for the hm_hf with a 10-year age difference. results for the 66 versus 70 retirement age decision are similar to the 62 versus 70 decision; however, the be irrs are even lower. be irrs range from a low of �0.123% for the wm_hf combination with a zero year age difference, to a high of 2.0017% for the hm_bf with a 10-year age difference. overall, the highest be irrs have an older hispanic male as the breadwinner; whereas the lowest be irrs have an older white male as the breadwinner. this would suggest early retirement at all age differences and all race-gender combinations, given the lower hurdle rates (be irrs) to overcome. assuming couples retire at different ages, docking, fortin, and michelson (2015) examined a specific scenario of the impact of age differences on an early male/female retirement of 66 and 62, respectively, versus a late male/female retirement of 70 and 66, respectively. in all nine race combinations the be irrs decline as the age differences increase. this suggests that the greater the age difference the greater the incentive to retire early as the hurdle rate is lower to overcome. 89d.s. docking et al. / financial services review 26 (2017) 87–111 3. how social security works a detailed description of how social security works can be found in docking, fortin, and michelson (2012, 2013). briefly, individuals aged 62 or older who had earned income that was subject to the social security payroll tax for at least 10 years (40 quarters) since 1951 are eligible for retirement benefits. individuals born between 1946 and 1954 can retire with full ssbs at their fra of 66. the fra gradually rises until it reaches 67 for people born in 1960 or later. however, individuals have the option to retire earlier or later than their fra. the earliest one can retire is age 62, and the latest is age 70. early retirement is attractive for many reasons: ssbs and rules can change, health concerns, and increased demand for leisure. however, ssbs are permanently reduced by an actuarial reduction factor (5/9 of 1% for the first 36 months and 5/12 of 1% per month thereafter for early retirement). for example, a worker with a fra of 66 who claims early at age 62 receives 75% of their fra benefit amount; a worker with a fra of 67 who claims at age 62 receives only 70% of their fra benefit amount. delayed retirement is attractive because ssbs are increased by a delayed retirement credit (drc) of 8% for each year of delay after fra up to age 70. in this case a delayed retirement credit (drc) will be added to the fra benefit. for example, a worker with a fra of 66 who delays claiming until age 70 receives 132% of their fra benefit amount; a worker with a fra of 67 who claims at age 70 receives only 124% of their fra benefit amount. workers who claim early retirement benefits, but continue to work, may have their ssbs reduced. this is referred to as the earnings test (et). however, since 2000, there has been no et above the fra.1 that is, ssbs are not reduced if the worker is of fra and continues to work. a spouse has dual entitlements to ssbs. a spouse is entitled to the larger of 100% of benefits at fra based on his or her own earnings record or up to 50% of the spouse’s benefits at fra. ssbspouse1 � max {ssbown; .5(ssbspouse2)} for example, a spouse receives one-half of her spouse’s full retirement benefit unless the spouse begins collecting benefits before her fra. if the spouse begins collecting benefits before her fra, the amount of the spouse’s benefit is reduced by a percentage base on the number of months before she reaches fra.2 for example, based on the fra of 66, if the spouse begins collecting benefits: at age 65, the benefit amount would be about 45.8% of the retired worker’s (spouse’s) full benefit; at age 64, it would be about 41.7%; at age 63, 37.5%; and at age 62, 35%. if the spouse’s fra is greater than 66, spousal benefits are further reduced for early retirement. for example, assume dean is 66 and dorothy is 62, both with a fra of 66. dean retires at 66 with ssb at fra of $2,000 per month. dorothy retires early at 62 and receives 35% of $2,000 or $700 per month in spousal ssbs. if dorothy waits and retires at her fra of 66 she receives 50% of $2,000 or $1,000 in spousal ssbs. 90 d.s. docking et al. / financial services review 26 (2017) 87–111 once one begins ssbs based on his or her own work record they cannot later switch to ssbs based on the spouse’s record. also, one cannot begin ssbs based on the spouse’s record and then later switch to ssbs based on his or her own work record. however, there is an exception: a spouse (spouse) can retire and begin collecting her (his) own ssbs while her (his) spouse (spouse) still works and delays benefits. upon her (his) spouse’s (spouse’s) retirement, she (he) can switch over to 50% of his (her) benefits, if spousal benefits are greater than her (his) own benefits. spouse’s benefits do not include any accrued delayed retirement credits. for example, assume richard and jane, are both 62 with a fra of 66. currently, richard’s ssbs at fra are $2,000 per month and jane’s ssbs at fra are $1,000. jane retires at 62 and receives 75% of 1,000 or $750 per month. richard continues to work until age 66. his ssbs at fra are still $2,000 per month and he retires at fra. assuming no cola for jane’s ssbs, she can now switch over to spousal benefits of 50% � $2,000 � $1,000 per month. 4. model similar to mccormack and perdue (2006) we avoid the problem of an uncertain discount rate (dr) by computing the internal rate of return (irr) equating two retirement options. following docking, fortin, and michelson (2015), for married couples of the same age, the irr can be solved for by using the following equation: %benefit_1 � � 1 i � 1 1� irr 12 � i � %benefit_2 � � 1 j � 1 1� irr 12 � j � %benefit_3 � � 1 m � 1 1� irr 12 � m � � 1 1� irr 12 � n3�n1 � %benefit_4 � � 1 n � 1 1� irr 12 � n � � 1 1� irr 12 � n4�n2 where: benefit_x � percent of ssb received based on retirement age i � 1 to months to life expectancy for retirement age 1 of spouse_1 (n1) j � 1 to months to life expectancy for retirement age 1 of spouse_2 (n2) m � 1 to months to life expectancy for retirement age 2 of spouse_1 (n3) n � 1 to months to life expectancy for retirement age 2 of spouse_2 (n4) 91d.s. docking et al. / financial services review 26 (2017) 87–111 n3 – n1 and n4 – n2 � difference in months between retirement age 1 and retirement age 2, where retirement age 2 is greater than retirement age 1. the two terms on the left-hand side of the equation, %benefit_1 � � 1 i � 1 1� irr 12 � i and %benefit_2 � � 1 j � 1 1� irr 12 � j , represent the present value of initiating receipt of benefits at retirement age 1. the two terms on the right-hand side of the equation, %benefit_3 � � 1 m � 1 1� irr 12 � m and %benefit_4 � � 1 n � 1 1� irr 12 � n , represent the present value of initiating receipt of benefits at retirement age 2; the two second terms on the right-hand side, � 1 1 � irr 12 � n3�n1 and � 1 1� irr 12 � n4�n2 discount the present value of benefits at retirement age 2 back to retirement age 1 so that the irr can be computed at the same point in time. for example, if the first retirement age is 62 and the second retirement age is 66, the irr computation for the age 66 term must be discounted back to the same point in time as the age 62 term. it should be noted that this model is appropriate only for same aged couples retiring at the same age. when the couples are different ages but still retire at the same age, an additional discount factor � 1 1� irr 12 � d is required to discount all expected cash flows back to the initial start of benefits. the model now becomes: %benefit_1 � � 1 i � 1 1� irr 12 � i � %benefit_2 � � 1 j � 1 1� irr 12 � j �� 1 1� irr 12 � d � %benefit_3 � � 1 m � 1 1� irr 12 � m � � 1 1� irr 12 � n3�n1 � %benefit_4 � � 1 n � 1 1� irr 12 � n � � 1 1� irr 12 � n4�n2 � � 1 1� irr 12 � d 92 d.s. docking et al. / financial services review 26 (2017) 87–111 where: d � the age difference in months between the spouses (agespouse_1 – agespouse_2) and agespouse_1 � agespouse_2. in addition, if the couples are different ages and retire at different ages, additional discounting complications are introduced. the model now becomes: %benefit_1 � � 1 i � 1 1� irr 12 � i � %benefit_2 � � 1 j � 1 1� irr 12 � j �� 1 1� irr 12 � d��n�n2� � %benefit_3 � � 1 m � 1 1� irr 12 � m � � 1 1� irr 12 � n3�n1 � %benefit_4 � � 1 n � 1 1� irr 12 � n � � 1 1� irr 12 � n4�n2 � � 1 1� irr 12 � dd � �n3 � n4� in our 2015 study, we assumed spouse_1 was the male/spouse and spouse_2 was the female/spouse and spouse_1 male was older than spouse 2_female. 4.1. assumptions in the model like our 2015 study, the following assumptions are made: 1. ssb are received monthly. 2. the retirement decision is made annually because life expectancy tables only provide annual data. 3. the 2006 united states life tables and the 2010 national center for health statistics provide life expectancies.3 life expectancy is adjusted for when a worker retires. for example, a white male who retires at age 62 is expected to live approximately 19 more years to age 81; whereas if he waits and retires at age 66 he is expected to live approximately 16 more years to age 82. we look at life expectancies based on gender and race. 4. we assume excess earnings are $0 and that early retirement ssb are not further reduced by the earnings test. 5. if a retiree has substantial income (earned and unearned) in addition to his ssb, up to 85% of his annual benefits may be subject to federal income tax. in our analysis we assume other income is below the minimum such that 0% of ssb are taxed. however, by using the irr method to find the optimal retirement age, taxation of ssb really becomes irrelevant, since (1-tax rate of ssb) shows up on both the leftand right-hand sides of our equation, effectively cancelling out one another. 93d.s. docking et al. / financial services review 26 (2017) 87–111 6. since 1983, the ssa provides for an automatic increase in ssb if there is an increase in the cpi-w from third quarter last year to third quarter of the current year. spitzer (2006) finds that only longevity and expected rates of return are determining factors as the optimal time to retire and that inflation and taxes play no significant role. as a consequence, we assume cola is zero. 7. we also assume the couple has no dependents, and that neither party receives a government pension. furthermore, the couple may be forced into a higher federal or state tax bracket because of other income; this, too, is irrelevant in our analysis and is ignored. in this study we make two different assumptions from out 2015 study: 8. we assume spouse_1 is the spouse (female) and spouse_2 is the spouse (male). the spouse (female) is older than the spouse (male). we look at age differences (agefemale – agemale) of 0, 4, 7, and 10. table 1 average life expectancy given current age age all males white males black males hispanic males avg # years remaining expected age to die avg # years remaining expected age to die avg # years remaining expected age to die avg # years remaining expected age to die 62 19.19 81.19 19.32 81.32 16.90 78.90 21.26 83.26 63 18.46 81.46 18.57 81.57 16.29 79.29 20.48 83.48 64 17.73 81.73 17.83 81.83 15.69 79.69 19.71 83.71 65 17.01 82.01 17.10 82.10 15.10 80.10 18.96 83.96 66 16.30 82.30 16.38 82.38 14.51 80.51 18.21 84.21 67 15.60 82.60 15.67 82.67 13.93 80.93 17.48 84.48 68 14.90 82.90 14.97 82.97 13.36 81.36 16.77 84.77 69 14.22 83.22 14.28 83.28 12.80 81.80 16.07 85.07 70 13.55 83.55 13.60 83.60 12.25 82.25 15.38 85.38 age all females white females black females hispanic females avg # years remaining expected age to die avg # years remaining expected age to die avg # years remaining expected age to die avg # years remaining expected age to die 62 22.11 84.11 22.18 84.18 20.72 82.72 24.24 86.24 63 21.30 84.30 21.37 84.37 19.99 82.99 23.39 86.39 64 20.50 84.50 20.56 84.56 19.27 83.27 22.55 86.55 65 19.71 84.71 19.76 84.76 18.57 83.57 21.72 86.72 66 18.93 84.93 18.97 84.97 17.87 83.87 20.90 86.90 67 18.15 85.15 18.18 85.18 17.17 84.17 20.10 87.10 68 17.38 85.38 17.41 85.41 16.48 84.48 19.30 87.30 69 16.62 85.62 16.64 85.64 15.80 84.80 18.51 87.51 70 15.87 85.87 15.89 85.89 15.14 85.14 17.74 87.74 source: national vital statistics report (2010), june 28, 2010, volume 58, number 21, united states life tables, 2006; and arias (2010), united states life tables by hispanic origin. national center for health statistics. vital health stat 2(152). 2010. 94 d.s. docking et al. / financial services review 26 (2017) 87–111 9. we assume a one-earner family. the spouse is the working spouse_1, and the spouse is the non-working spouse_2. 5. examples 5.1. spouse earner, spouse and spouse same age, spouse and spouse retire at same time michael, a black male born in 1948, is married to angela, a black female born in 1948. they are trying to decide if they should retire early at age 62 or wait until fra of 66. michael is the sole breadwinner of the family. angela has no ssbs of her own. according to table 1, michael’s life expectancy at age 62 is an additional 16.90 years (202.8 months) to age 78.9; while his life expectancy at age 66 is an additional 14.51 years (174.12 months) to age 80.51. angela’s life expectancy at age 62 is an additional 20.72 years (248.64 months) to age 82.72; whereas her life expectancy at age 66 is an additional 17.87 years (214.44 months) to age 83.87. based on current social security requirements, michael will receive 100% of his ssb at age 66, but only 75% of his fra benefits at age 62. angela is able to claim up to 50% of michael’s ssb if she is at fra, but only 35% at age 62. using excel we can find the irr that will equate both sides of the following equation: 75% � � 1 202.8 � 1 1� irr 12 � i � 35% � � 1 248.64 � 1 1� irr 12 � j � 100% � � 1 174.12 � 1 1� irr 12 � m � � 1 1� irr 12 � �66�62��12 � 50% � � 1 214.44 � 1 1� irr 12 � n � � 1 1� irr 12 � �66�62� � 12 the irr that equates both sides is equal to 5.5291% (see table 2a) or approximately 5.53%. if the couple’s opportunity costs are less (greater) than 5.53%, then they should retire at the later (earlier) age. assume michael’s ssb at fra of 66 is $1,600 per month and his early retirement benefit is 75% or $1,200 per month at age 62. based on michael’s fra benefit of $1,600 per month, angela’s ssb will be 35% of $1,600 or $560 per month at age 62. at age 66 michael will receive $1,600 per month and angela will receive 50% of $1,600 or $800 per month. if the current market interest rate is 5%, then the present value (pv) of the left-hand side of the equation (retire early at age 62) is $164,070 (michael) plus $86,603 (angela) for a total of $250,673. the pv of the right-hand side of the equation (delay retirement to age 66) is $162,038 (michael) and $92,787 (angela) for a total of $254,825. this results in a difference 95d.s. docking et al. / financial services review 26 (2017) 87–111 of $4,152, implying that michael and angela should wait until age 66 to retire. if michael and angela believe they could invest their monthly ssb at 5.53% or greater over the next four years, then they should retire early, at age 62; if not, they should delay retirement until age 66. of course, this assumes they do not need any of their ssb on which to live; a highly unlikely assumption. 5.2. spouse earner, spouse age � spouse age, spouse and spouse retire at same time now, assume angela was born in 1952 and is four years younger than michael. there is an additional four years of discounting required (48 months) for the spouse’s spousal benefits at both age 62 and 66. this is reflected in the following formula: 75% � � 1 202.8 � 1 1� irr 12 � i � 35% � � 1 248.64 � 1 1� irr 12 � j � � 1 1� irr 12 � �4�12� � 100% � � 1 174.12 � 1 1� irr 12 � m � � 1 1� irr 12 � �66�62��12 � 50% � � 1 214.44 � 1 1� irr 12 � n � � 1 1� irr 12 � �66�62��12 � � 1 1� irr 12 � �4�12� note that the age 62 spousal benefits are now discounted 48 months (instead of none previously) and the age 66 spousal benefits are now discounted 96 months instead of 48 months. using excel we find the be irr is 5.4042% (approximately 5.40%) that is reflected in table 2a with a four year age difference. if the couple’s opportunity costs are less (greater) than 5.40%, then they should retire at the later (earlier) age. again, if we assume michael’s ssb at fra of 66 is $1,600 per month and if the current market interest rate is 5%, then the pv of the left-hand side of the equation (retire early at age 62) is $164,070 (michael) plus $70,933 (angela) for a total of $235,003. the pv of the right-hand side of the equation (delay retirement to age 66) is $162,038 (michael) and $75,999 (angela) for a total of $238,037. this results in a difference of $3,034, implying that michael and angela should wait until age 66 to retire. if michael and angela believe they could invest their monthly ssb at 5.40% or greater over the next four years, then they should retire early, at age 62; if not, they should delay retirement until age 66. again, this assumes they do not need any of their ssb on which to live. now, assume angela was born in 1958 and is 10 years younger than michael. there is an additional 10 years of discounting required (120 months) for the spouse spousal benefits at 96 d.s. docking et al. / financial services review 26 (2017) 87–111 both age 62 and 66. based on current social security requirements, michael will receive 100% of his ssb at age 66, but only 75% of his fra benefits at age 62. angela is able to claim up to 47.2% of michael’s ssb at age 66, but only 33.3% at age 62.4 this is reflected in the following formula: 75% � � 1 202.8 � 1 1� irr 12 � i � 33.3% � � 1 248.64 � 1 1� irr 12 � j � � 1 1� irr 12 � �10�12� � 100% � � 1 174.12 � 1 1� irr 12 � m � � 1 1� irr 12 � �66�62��12 � 47.2% � � 1 214.44 � 1 1� irr 12 � n � � 1 1� irr 12 � �66�62��12 � � 1 1� irr 12 � �10 � 12� note that the age 62 spousal benefits are now discounted 120 months and the age 66 spousal benefits are now discounted 168 months. using excel we find the be irr is 5.1618% that is reflected in table 2a with a 10 year age difference. 5.3. spouse earner, spouse age � spouse age, spouse and spouse retire at different times to illustrate an example from table 3a again consider the same couple above with a four year age difference but with the h/w early retirement ages of 66/62 and delayed retirement ages of 70/66. the formula to solve this example would be: 100% � � 1 174.12 � 1 1� irr 12 � i � 35% � � 1 248.64 � 1 1� irr 12 � j � �� 1 1� irr 12 � 4��66�62� �12 � � 132% � � 1 147 � 1 1� irr 12 � m � � 1 1� irr 12 � �70�66��12 � 50% � � 1 214.44 � 1 1� irr 12 � n � � 1 1� irr 12 � �66�62��12 � �� 1 1� irr 12 � 4��70�66� �12 � using excel to solve for the be irr yields 4.7984% or approximately 4.80% (see table 3a). 97d.s. docking et al. / financial services review 26 (2017) 87–111 6. results 6.1. retiring at same age the results presented in tables 2a are from docking, fortin, and michelson (2015) and are based on applying the previously described excel models for a representative baby boom birth year of 1948 for both the spouse and spouse initially and progressively later years for the non-working female spouse. the tables provide the be irrs for the nine race/gender combinations where w � white, b � black, h � hispanic, m � male, and f � female. the nine race/gender combinations are: wm_wf, bm_bf, hm_hf, wm_bf, bm_wf, wm_hf, hm_wf, bm_hf, and hm_bf. three claiming scenarios are considered: receiving benefits early (e.g., at age 62 vs. 66); the maximum realistic delay period (e.g., at age 62 vs. 70) and delaying benefits past fra (e.g., age 66 vs. 70). within these three claiming scenarios we examine couples by race combination who retire at the same age with age differences of 0, 4, 7, and 10 years with the non-working spouse younger than the working spouse. 6.1.1. results of role switching table 2b shows be irrs when the spouse is the earner and is older than the spouse. the results presented in tables 2b are based on applying the previously described excel models for a representative baby boom birth year of 1948 for both the spouse and spouse initially and progressively later years for the non-working male spouse. keep in mind that the be irrs can be viewed as “hurdle rates” where if a couple’s expected return or opportunity cost of capital is greater than (less than) the computed be irr over the given time horizon, the couple should retire at the earlier (later) age. this analysis also assumes that the couple does not need the ssbs to live on and can invest the benefits in the capital markets if the decision is made to retire early. results are similar to when we assume the spouse was the earner and was older than the spouse. assuming couples retire at the same age, for the retirement ages 62 versus 66, the be irrs uniformly decrease as the age difference increases. this implies that the greater the age difference between couples, the earlier the couples should retire. be irrs range from a low of 5.26% for the bm_wf combination with a 10-year age difference, to a high of 5.87% for the hm_hf combination with a zero year age difference. for the retirement ages 62 versus 70, the be irrs uniformly increase as the age difference increases. this implies that the greater the age difference between couples, the later the couples should retire. however, the be irrs are much lower in this retirement decision, ranging from a low of 3.25% for the wm_wf combination with a zero year age difference, to a high of 4.20% for the bm_hf with a 10-year age difference. results for the 66 versus 70 retirement age decision are similar to the 62 versus 70 decision, however the be irrs are even lower. be irrs range from a low of 0.82% for the hm_wf combination with a zero year age difference, to a high of 2.66% for the bm_hf with a 10-year age difference. overall the highest be irrs have an older hispanic female as the breadwinner; whereas the lowest be irrs have an older white female as the breadwinner. this would suggest early retirement 98 d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 2a b re ak ev en ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith in cr ea si ng ag e di ff er en ce s w ith m al e as br ea dw in ne r an d m al e ol de r (m al e bo rn 19 48 ; fe m al e bo rn 19 48 , 19 52 , 19 55 , 19 58 ) a ge di ff er en ce m al e re tir em en t ag e 1 fe m al e re tir em en t ag e 1 m al e re tir em en t ag e 2 fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 62 62 66 66 5. 46 48 % 5. 52 91 % 5. 81 51 % 5. 45 66 % 5. 53 71 % 5. 60 61 % 5. 68 42 % 5. 67 86 % 5. 67 73 % 4 62 62 66 66 5. 33 77 % 5. 40 42 % 5. 68 54 % 5. 33 01 % 5. 41 15 % 5. 46 01 % 5. 57 30 % 5. 53 42 % 5. 56 67 % 7 62 62 66 66 5. 22 30 % 5. 29 16 % 5. 57 06 % 5. 21 68 % 5. 29 77 % 5. 33 17 % 5. 47 17 % 5. 40 65 % 5. 46 66 % 10 62 62 66 66 5. 09 07 % 5. 16 18 % 5. 44 03 % 5. 08 62 % 5. 16 61 % 5. 18 51 % 5. 35 54 % 5. 26 06 % 5. 35 20 % 0 62 62 70 70 2. 89 49 % 3. 10 06 % 3. 39 19 % 2. 98 63 % 3. 00 80 % 2. 98 30 % 3. 31 66 % 3. 09 57 % 3. 39 98 % 4 62 62 70 70 2. 97 99 % 3. 18 92 % 3. 49 17 % 3. 06 56 % 3. 10 25 % 3. 06 15 % 3. 42 30 % 3. 18 36 % 3. 50 00 % 7 62 62 70 70 3. 10 74 % 3. 31 66 % 3. 62 63 % 3. 18 48 % 3. 23 85 % 3. 18 64 % 3. 56 07 % 3. 31 66 % 3. 62 94 % 10 62 62 70 70 3. 35 49 % 3. 56 21 % 3. 87 50 % 3. 42 07 % 3. 50 05 % 3. 43 91 % 3. 80 89 % 3. 57 89 % 3. 86 30 % 0 66 66 70 70 0. 04 97 % 0. 46 23 % 0. 71 19 % 0. 31 48 % 0. 19 22 % � 0. 01 23 % 0. 77 10 % 0. 12 75 % 0. 99 83 % 4 66 66 70 70 0. 05 50 % 0. 50 83 % 0. 78 65 % 0. 34 46 % 0. 21 37 % � 0. 01 37 % 0. 85 04 % 0. 14 20 % 1. 09 12 % 7 66 66 70 70 0. 30 27 % 0. 77 09 % 1. 07 86 % 0. 58 73 % 0. 48 28 % 0. 25 25 % 1. 12 70 % 0. 42 97 % 1. 35 86 % 10 66 66 70 70 1. 05 09 % 1. 48 49 % 1. 82 12 % 1. 27 27 % 1. 26 29 % 1. 05 92 % 1. 82 20 % 1. 26 86 % 2. 00 17 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. 99d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 2b b re ak ev en ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith in cr ea si ng ag e di ff er en ce s w ith fe m al e as br ea dw in ne r an d fe m al e ol de r (f em al e bo rn 19 48 ; m al e bo rn 19 48 , 19 52 , 19 55 , 19 58 ) a ge di ff er en ce m al e re tir em en t ag e 1 fe m al e re tir em en t ag e 1 m al e re tir em en t ag e 2 fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 62 62 66 66 5. 53 % 5. 55 % 5. 87 % 5. 57 % 5. 51 % 5. 74 % 5. 68 % 5. 72 % 5. 72 % 4 62 62 66 66 5. 45 % 5. 47 % 5. 78 % 5. 49 % 5. 43 % 5. 67 % 5. 57 % 5. 65 % 5. 62 % 7 62 62 66 66 5. 37 % 5. 40 % 5. 70 % 5. 41 % 5. 35 % 5. 60 % 5. 48 % 5. 58 % 5. 52 % 10 62 62 66 66 5. 27 % 5. 31 % 5. 60 % 5. 32 % 5. 26 % 5. 51 % 5. 37 % 5. 51 % 5. 41 % 0 62 62 70 70 3. 25 % 3. 47 % 3. 68 % 3. 34 % 3. 38 % 3. 61 % 3. 33 % 3. 74 % 3. 42 % 4 62 62 70 70 3. 36 % 3. 58 % 3. 80 % 3. 45 % 3. 48 % 3. 74 % 3. 43 % 3. 86 % 3. 52 % 7 62 62 70 70 3. 49 % 3. 70 % 3. 94 % 3. 59 % 3. 60 % 3. 88 % 3. 56 % 3. 99 % 3. 66 % 10 62 62 70 70 3. 73 % 3. 92 % 4. 18 % 3. 83 % 3. 82 % 4. 12 % 3. 80 % 4. 20 % 3. 90 % 0 66 66 70 70 0. 86 % 1. 36 % 1. 39 % 1. 01 % 1. 21 % 1. 43 % 0. 82 % 1. 75 % 0. 97 % 4 66 66 70 70 0. 94 % 1. 47 % 1. 52 % 1. 11 % 1. 31 % 1. 56 % 0. 89 % 1. 89 % 1. 06 % 7 66 66 70 70 1. 19 % 1. 72 % 1. 80 % 1. 37 % 1. 54 % 1. 83 % 1. 15 % 2. 14 % 1. 33 % 10 66 66 70 70 1. 80 % 2. 26 % 2. 42 % 1. 99 % 2. 08 % 2. 42 % 1. 81 % 2. 66 % 2. 00 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. 100 d.s. docking et al. / financial services review 26 (2017) 87–111 at all age differences and all race-gender combinations, given the lower hurdle rates (be irrs) to overcome. from table 2c and 2d we can see that in all but seven scenarios, the be irrs are greater when the working spouse is female. women almost always have higher be irrs. when deciding to retire early or later, the higher the hurdle rate, the more difficult it is to retire early. the implication is that in marriages where the spouse is the breadwinner and is the older partner, it is more difficult for the couple to retire early, as compared to marriages where the spouse is the breadwinner and is the older partner. it is also interesting and useful to compare the results across race categories at key comparison ages. from tables 2a and 2b the high and low be irrs for the following retirement age comparisons are evident. retirement age comparison male as breadwinner and older female as breadwinner and older high breakeven irr low breakeven irr high breakeven irr low breakeven irr 62/62 vs. 66/66 age difference 0 hm_hf wm_bf hm_hf bm_wf 4 hm_hf wm_bf hm_hf bm_wf 7 hm_hf wm_bf hm_hf bm_wf 10 hm_hf wm_bf hm_hf bm_wf 62/62 vs. 70/70 age difference 0 hm_bf wm_wf bm_hf wm_wf 4 hm_bf wm_wf bm_hf wm_wf 7 hm_bf wm_wf bm_hf wm_wf 10 hm_hf wm_wf bm_hf wm_wf 66/66 vs. 70/70 age difference 0 hm_bf wm_hf bm_hf hm_wf 4 hm_bf wm_hf bm_hf hm_wf 7 hm_bf wm_hf bm_hf hm_wf 10 hm_bf wm_wf bm_hf wm_wf remember that a higher (lower) be irr would imply retiring later (earlier) because the hurdle rate opportunity cost is more difficult (less difficult) to overcome. when the male is the breadwinner and older, the high be irr column is dominated by hm_bf (seven occurrences) and hm_hf (five occurrences). the low be irr column has 5 wm_wf lows, 4 wm_bf lows and 3 wm_hf lows. the most obvious patterns here are that the high be irr group consistently has a hispanic spouse and low be irr group consistently has a white spouse. when the female is the breadwinner and older, the high be irr column is dominated by bm_hf (eight occurrences) and hm_hf (four occurrences). the low be irr column has 5 wm_wf lows, 4 bm_wf lows and 3 hm_wf lows. the most obvious patterns here are that the high be irr group consistently has a hispanic spouse and low be irr group consistently has a white spouse. thus, irrespective of who is the breadwinner, hispanics have higher hurdle rates; while whites have lower hurdle rates. for a given retirement age comparison or age difference the 101d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 2c pe rc en ta ge di ff er en ce in br ea ke ve n ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith in cr ea si ng ag e di ff er en ce s a ge di ff er en ce m al e re tir em en t ag e 1 fe m al e re tir em en t ag e 1 m al e re tir em en t ag e 2 fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r a b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 62 62 66 66 1. 26 % 0. 45 % 1. 00 % 2. 13 % � 0. 42 % 2. 33 % � 0. 08 % 0. 73 % 0. 73 % 4 62 62 66 66 2. 09 % 1. 28 % 1. 70 % 3. 01 % 0. 37 % 3. 77 % 0. 03 % 2. 12 % 0. 87 % 7 62 62 66 66 2. 79 % 1. 99 % 2. 27 % 3. 72 % 1. 06 % 4. 96 % 0. 12 % 3. 29 % 0. 98 % 10 62 62 66 66 3. 60 % 2. 83 % 2. 93 % 4. 56 % 1. 89 % 6. 36 % 0. 23 % 4. 67 % 1. 10 % 0 62 62 70 70 12 .2 0% 12 .0 6% 8. 56 % 11 .7 7% 12 .5 2% 21 .0 8% 0. 38 % 20 .8 2% 0. 52 % 4 62 62 70 70 12 .6 3% 12 .2 3% 8. 86 % 12 .6 1% 12 .2 8% 22 .0 9% 0. 21 % 21 .1 2% 0. 70 % 7 62 62 70 70 12 .3 3% 11 .7 0% 8. 67 % 12 .7 6% 11 .3 1% 21 .7 9% 0. 02 % 20 .1 8% 0. 85 % 10 62 62 70 70 11 .1 2% 10 .0 8% 7. 79 % 12 .0 3% 9. 06 % 19 .6 9% � 0. 25 % 17 .3 1% 1. 00 % 0 66 66 70 70 16 29 .4 7% 19 4. 45 % 94 .9 2% 22 1. 74 % 52 9. 66 % 11 74 8. 89 % 5. 80 % 12 71 .8 8% � 3. 08 % 4 66 66 70 70 16 09 .7 8% 18 9. 82 % 92 .7 1% 22 2. 03 % 51 1. 98 % 11 51 4. 12 % 4. 99 % 12 29 .7 2% � 2. 79 % 7 66 66 70 70 29 2. 53 % 12 2. 85 % 66 .8 8% 13 3. 24 % 21 9. 35 % 62 5. 71 % 2. 43 % 39 7. 94 % � 1. 78 % 10 66 66 70 70 71 .3 0% 52 .4 6% 32 .7 1% 56 .2 6% 64 .7 8% 12 8. 09 % � 0. 71 % 10 9. 54 % � 0. 33 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. a pe rc en ta ge di ff er en ce ca lc ul at ed as ab so lu te va lu e of (f em al e br ea dw in ne r ir r m al e br ea dw in ne r ir r ) di vi de d by m al e br ea dw in ne r ir r or (t ab le 2b ir r t ab le 2a ir r )/ a b s( t ab le 2a ir r ). 102 d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 2d pe rc en t di ff er en ce in br ea ke ve n ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith in cr ea si ng ag e di ff er en ce s a ge di ff er en ce m al e re tir em en t ag e 1 fe m al e re tir em en t ag e 1 m al e re tir em en t ag e 2 fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r a b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 62 62 66 66 0. 07 % 0. 03 % 0. 06 % 0. 12 % � 0. 02 % 0. 13 % 0. 00 % 0. 04 % 0. 04 % 4 62 62 66 66 0. 11 % 0. 07 % 0. 10 % 0. 16 % 0. 02 % 0. 21 % 0. 00 % 0. 12 % 0. 05 % 7 62 62 66 66 0. 15 % 0. 11 % 0. 13 % 0. 19 % 0. 06 % 0. 26 % 0. 01 % 0. 18 % 0. 05 % 10 62 62 66 66 0. 18 % 0. 15 % 0. 16 % 0. 23 % 0. 10 % 0. 33 % 0. 01 % 0. 25 % 0. 06 % 0 62 62 70 70 0. 35 % 0. 37 % 0. 29 % 0. 35 % 0. 38 % 0. 63 % 0. 01 % 0. 64 % 0. 02 % 4 62 62 70 70 0. 38 % 0. 39 % 0. 31 % 0. 39 % 0. 38 % 0. 68 % 0. 01 % 0. 67 % 0. 02 % 7 62 62 70 70 0. 38 % 0. 39 % 0. 31 % 0. 41 % 0. 37 % 0. 69 % 0. 00 % 0. 67 % 0. 03 % 10 62 62 70 70 0. 37 % 0. 36 % 0. 30 % 0. 41 % 0. 32 % 0. 68 % � 0. 01 % 0. 62 % 0. 04 % 0 66 66 70 70 0. 81 % 0. 90 % 0. 68 % 0. 70 % 1. 02 % 1. 45 % 0. 04 % 1. 62 % � 0. 03 % 4 66 66 70 70 0. 89 % 0. 96 % 0. 73 % 0. 77 % 1. 09 % 1. 58 % 0. 04 % 1. 75 % � 0. 03 % 7 66 66 70 70 0. 89 % 0. 95 % 0. 72 % 0. 78 % 1. 06 % 1. 58 % 0. 03 % 1. 71 % � 0. 02 % 10 66 66 70 70 0. 75 % 0. 78 % 0. 60 % 0. 72 % 0. 82 % 1. 36 % � 0. 01 % 1. 39 % � 0. 01 % w � w hi te , b � b la ck , h � h is pa ni c, m � m al e, f � fe m al e. a pe rc en t d if fe re nc e ca lc ul at ed as (f em al e br ea dw in ne r ir r m al e br ea dw in ne r ir r ) or (t ab le 2b ir r t ab le 2a ir r ). 103d.s. docking et al. / financial services review 26 (2017) 87–111 results can be interpreted as follows: the high (low) breakeven group would prefer to retire later (earlier) because the hurdle rate is more difficult (less difficult) to overcome. thus, hispanics have a more difficult time retiring early and whites have a less difficult time retiring early. 6.2. retiring at different ages the results presented in table 3a are from docking, fortin, and michelson (2015) and are based on applying the previously described excel models for a representative baby boom birth year of 1948 for both the spouse and spouse initially and progressively later years for the non-working female spouse. the spouses are assumed to retire at different ages. a specific scenario of the impact of age differences on an early male/female retirement of 66 and 62, respectively, versus a late male/female retirement of 70 and 66, respectively, is examined. there is a not applicable (na) in the table for an age difference of 0 because spousal benefits cannot be claimed by the female until the male retires. in all nine race combinations the be irrs decline as the age differences increase. this suggests that the greater the age difference the greater the incentive to retire early as the hurdle rate is lower to overcome. 6.2.1. results of role switching table 3b shows be irrs when the spouse is the earner and is older than the spouse. results are similar to when we assume the spouse was the earner and was older than the spouse. again, the be irrs uniformly decrease as the age difference increases. this implies that the greater the age difference between couples, the earlier the couples should retire. from table 3c and 3d we can see that in all but seven scenarios, the be irrs are greater when the working spouse is female. again, women almost always have higher be irrs. the implication is that in marriages where the spouse is the breadwinner and is the older partner, it is more difficult for the couple to retire early, as compared with marriages where the spouse is the breadwinner and is older partner. it is also interesting to examine the high and low be irrs for this comparison for each age difference by race category. from tables 3a and 3b the high and low be irrs for the following retirement age comparisons are evident. retirement age comparison male as breadwinner and older female as breadwinner and older high breakeven irr low breakeven irr high breakeven irr low breakeven irr 66/62 vs. 70/66 age difference 0 na na na na 4 hm_hf wm_bf hm_hf bm_wf 7 hm_hf wm_bf hm_hf bm_wf 10 hm_hf wm_bf hm_hf bm_wf remember that a higher (lower) be irr would imply retiring later (earlier) because the hurdle rate opportunity cost is more difficult (less difficult) to overcome. the high be irr 104 d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 3a b re ak ev en ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith di ff er en t re tir em en t ag es an d in cr ea si ng ag e di ff er en ce s. m al e as br ea dw in ne r an d m al e ol de r (m al e bo rn 19 48 ; fe m al e bo rn 19 48 , 19 52 , 19 55 , 19 58 ) a ge di ff er en ce m al e re tir em en t ag e 1 fe m al e re tir em en t ag e 1 m al e re tir em en t ag e 2 fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 66 62 70 66 n a n a n a n a n a n a n a n a n a 4 66 62 70 66 4. 60 % 4. 80 % 5. 08 % 4. 59 % 4. 81 % 4. 75 % 4. 95 % 4. 95 % 4. 94 % 7 66 62 70 66 4. 46 % 4. 66 % 4. 94 % 4. 45 % 4. 67 % 4. 59 % 4. 82 % 4. 80 % 4. 81 % 10 66 62 70 66 4. 29 % 4. 50 % 4. 78 % 4. 28 % 4. 51 % 4. 41 % 4. 68 % 4. 62 % 4. 67 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. n a � no t ap pl ic ab le be ca us e if hu sb an d( m al e) an d w if e( fe m al e) ar e sa m e ag e, w if e ca nn ot re tir e an d dr aw sp ou sa l be ne fit s be fo re he r hu sb an d re tir es . 105d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 3b b re ak ev en ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith di ff er en t re tir em en t ag es an d in cr ea si ng ag e di ff er en ce s. fe m al e as br ea dw in ne r an d fe m al e ol de r (f em al e bo rn 19 48 ; m al e bo rn 19 48 , 19 52 , 19 55 , 19 58 ) a ge di ff er en ce fe m al e re tir em en t ag e 1 m al e re tir em en t ag e 1 fe m al e re tir em en t ag e 2 m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 66 62 70 66 n a n a n a n a n a n a n a n a n a 4 66 62 70 66 4. 67 % 4. 78 % 5. 12 % 4. 81 % 4. 64 % 4. 99 % 4. 82 % 4. 96 % 4. 95 % 7 66 62 70 66 4. 56 % 4. 68 % 5. 01 % 4. 70 % 4. 54 % 4. 89 % 4. 69 % 4. 87 % 4. 83 % 10 66 62 70 66 4. 43 % 4. 56 % 4. 88 % 4. 58 % 4. 41 % 4. 78 % 4. 55 % 4. 76 % 4. 69 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. n a � n ot ap pl ic ab le be ca us e if hu sb an d( m al e) an d w if e( fe m al e) ar e sa m e ag e, hu sb an d ca nn ot re tir e an d dr aw sp ou sa l be ne fit s be fo re hi s w if e re tir es . 106 d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 3c pe rc en ta ge di ff er en ce in br ea ke ve n ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith di ff er en t re tir em en t ag es an d in cr ea si ng ag e di ff er en ce s a ge di ff er en ce fe m al e/ m al e re tir em en t ag e 1 m al e/ fe m al e re tir em en t ag e 1 fe m al e/ m al e re tir em en t ag e 2 m al e/ fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r a b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 66 62 70 66 n a n a n a n a n a n a n a n a n a 4 66 62 70 66 1. 55 % � 0. 38 % 0. 85 % 4. 80 % � 3. 48 % 5. 06 % � 2. 57 % 0. 21 % 0. 30 % 7 66 62 70 66 2. 35 % 0. 38 % 1. 46 % 5. 76 % � 2. 88 % 6. 59 % � 2. 64 % 1. 51 % 0. 37 % 10 66 62 70 66 3. 33 % 1. 30 % 2. 19 % 6. 93 % � 2. 14 % 8. 47 % � 2. 72 % 3. 08 % 0. 44 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. a pe rc en ta ge di ff er en ce ca lc ul at ed as (f em al e br ea dw in ne r ir r m al e br ea dw in ne r ir r ) di vi de d by m al e br ea dw in ne r ir r or (t ab le 3b ir r t ab le 3a ir r )/ t ab le 3a ir r ). r ac e an d ge nd er he ld co ns ta nt . 107d.s. docking et al. / financial services review 26 (2017) 87–111 t ab le 3d pe rc en t di ff er en ce in br ea ke ve n ir r s fo r a sa m pl e of m ar ri ed re tir em en t ag es w ith di ff er en t re tir em en t ag es an d in cr ea si ng ag e di ff er en ce s a ge di ff er en ce fe m al e/ m al e re tir em en t ag e 1 m al e/ fe m al e re tir em en t ag e 1 fe m al e/ m al e re tir em en t ag e 2 m al e/ fe m al e re tir em en t ag e 2 w m _w f br ea ke ve n ir r a b m _b f br ea ke ve n ir r h m _h f br ea ke ve n ir r w m _b f br ea ke ve n ir r b m _w f br ea ke ve n ir r w m _h f br ea ke ve n ir r h m _w f br ea ke ve n ir r b m _h f br ea ke ve n ir r h m _b f br ea ke ve n ir r 0 66 62 70 66 n a n a n a n a n a n a n a n a n a 4 66 62 70 66 0. 07 % � 0. 02 % 0. 04 % 0. 22 % � 0. 17 % 0. 24 % � 0. 13 % 0. 01 % 0. 01 % 7 66 62 70 66 0. 10 % 0. 02 % 0. 07 % 0. 26 % � 0. 13 % 0. 30 % � 0. 13 % 0. 07 % 0. 02 % 10 66 62 70 66 0. 14 % 0. 06 % 0. 10 % 0. 30 % � 0. 10 % 0. 37 % � 0. 13 % 0. 14 % 0. 02 % n ot es : ir r � in te rn al ra te of re tu rn ; w � w hi te ; b � b la ck ; h � h is pa ni c; m � m al e; f � fe m al e. a pe rc en t di ff er en ce ca lc ul at ed as (f em al e br ea dw in ne r ir r m al e br ea dw in ne r ir r ) or (t ab le 3b ir r t ab le 3a ir r ). r ac e an d ge nd er he ld co ns ta nt . 108 d.s. docking et al. / financial services review 26 (2017) 87–111 column is dominated by hm_hf, no matter which spouse is older or which spouse is the breadwinner. when the male is older and the breadwinner, the low be irr is dominated by wm_bf. when the female is older and the breadwinner, the low be irr is dominated by bm_wf. the most obvious pattern here that the low be irr group has an older white breadwinner and a younger black non-earning spouse. irrespective of who is the breadwinner, hispanics have higher hurdle rates; whereas whites have lower hurdle rates. for a given retirement age comparison or age difference the results can be interpreted as follows: the high (low) breakeven group would prefer to retire later (earlier) because the hurdle rate is more difficult (less difficult) to overcome. thus, hispanics have a more difficult time retiring early and whites have a less difficult time retiring early. 7. applications and implications the practical applications/implications of our results primarily depend on the couple’s opportunity cost of capital and available other resources. if the couple’s portfolio expected return or opportunity cost of capital is greater than (less than) the computed be irr, this would suggest that this couple retire at the earlier (later) date in the comparative analysis. these results should be useful for couples of different ages facing the social security early versus delayed retirement decision and financial planners. using the analytics described in this paper, couples and/or their financial planners could first compute their breakeven internal rates of return at various comparison ages and then compare this be irr to their expected portfolio return over the comparison period. if their expected portfolio return was greater than (less than) their be irr then they should consider retiring at the earlier (later) age. 8. conclusions the primary substantive conclusions from this study depends on the age comparisons that are being made. for different aged couples, irrespective who is older and the breadwinner, couples who retire at the same chronological age, the age 62 versus 66 comparisons show be irrs uniformly decrease as the age difference increases. because these be irrs are hurdle rates, this implies that greater age difference couples should retire earlier because the hurdle rate is less to overcome than at a smaller age difference. these results reverse for the age 62 versus 70 comparison and age 66 versus 70 comparisons where the irrs uniformly increase with age differences across all race combinations. this implies that greater age differences involve a greater hurdle and the smaller the age difference the greater the incentive to retire earlier because the hurdle rate is lower. for couples who have an early male/female retirement of 66 and 62, respectively, versus a late male/female retirement of 70 and 66, respectively, the be irrs consistently decline as the age differences increase across all race 109d.s. docking et al. / financial services review 26 (2017) 87–111 combinations. this suggests that the greater the age difference the greater the incentive to retire early as the hurdle rate is lower to overcome. for couples who retire at different ages, the greater the age difference the greater the incentive to retire early as the hurdle rate is lower to overcome. this is true irrespective who is older and the breadwinner. women almost always have higher be irrs. the implication is that in marriages where the spouse is the breadwinner and is the older partner, it is more difficult for the couple to retire early, as compared with marriages where the spouse is the breadwinner and is older partner. irrespective of who is the breadwinner, hispanics have higher hurdle rates; whereas whites have lower hurdle rates. for a given retirement age comparison or age difference the results can be interpreted as follows: the high (low) breakeven group would prefer to retire later (earlier) because the hurdle rate is more difficult (less difficult) to overcome. thus, hispanics have a more difficult time retiring early and whites have a less difficult time retiring early. notes 1 http://www.socialsecurity.gov/pubs/10003.html. 2 the percentage reduction for spousal benefits is 25/36 of 1% per month for the first 36 months and 5/12 of 1% for each additional month. 3 national vital statistics report, june 28, 2010, volume 58, number 21; united states life tables, 2006 provides life expectancies for black and white males and females. e. arias, united states life tables by hispanic origin. national center for health statistics. vital health stat 2(152). 2010 provides life expectancies for hispanic males and females (arias, 2010). 4 because angela is born in 1958, her fra is 66 and 8 months. thus her spousal benefits are reduced by an extra 8 months. references arias, e. (2010). united states life tables by hispanic origin. national center for health statistics. vital health stat 2(152). blanchett, d. (2012). when to claim social security benefits. journal of personal finance, 11, 36–87. coile, c., diamond, p., gruber, j., & jousten, a. (2002). delays in claiming social security benefits. journal of public economics, 84, 357–385. docking, d. s., fortin, r., & michelson, s. (2012). the influence of gender and race on the social security early retirement decision for single individuals. journal of economics and economic education research, 13, 87–104. docking, d. s., fortin, r., & michelson, s. (2013). the influence of race on the social security early retirement decision for married couples. financial services review, 22, 133–150. docking, d. s., fortin, r., & michelson, s. (2015). the impact of different ages and race on the social security early retirement decision for married couples. journal of economics and economic education research, 16, 72–88. mccormack, j. p., & perdue, g. (2006). optimizing the initiation of social security benefits. financial services review, 15, 335–348. 110 d.s. docking et al. / financial services review 26 (2017) 87–111 munnell, a. h., & soto, m. (2007). when should women claim social security benefits? journal of financial planning, 20, 58–65. national vital statistics report. (2010). united states life tables, 2006; volume 58, number 21. june 28, 2010. social security administration. (2017). national vital statistics report (available at http://www.socialsecurity.gov/pubs/10003.html). spitzer, j. j. (2006). delaying social security payments: a bootstrap. financial services review, 15, 233–245. sun, w., & webb, a. (2009). how much do households really lose by claiming social security at age 62? center for retirement research at boston college, crr wp 2009–11, march, 1–28. tucker, m. (2009). optimal retirement age under normal and negative market conditions considering social security and private savings. journal of financial planning, july, 42–49. 111d.s. docking et al. / financial services review 26 (2017) 87–111 financial adviser background checks bhanu balasubramniana, eric r. briskera, suzanne gradishera,* adepartment of finance, the university of akron, college of business administration, akron, oh 44325, usa abstract using the 2009 national financial capability survey, we identify demographic characteristics associated with financial adviser users who conduct adviser background checks and/or consider more than one adviser before making a choice, and if these activities improve their trust in financial advisers. we find that very few financial adviser users check backgrounds, but there is a positive relationship between adviser background checks and trust levels. overall, these findings indicate that having a reliable background check system in place, allowing financial consumers to conduct adviser background checks in an easy and efficient manner, will help improve trust in financial advisers. © 2014 academy of financial services. all rights reserved. jel classification: d18; d14; g18 keywords: consumer protection; personal finance; government policy and regulation; financial advisers 1. introduction we investigate characteristics associated with financial consumers who conduct financial adviser background checks and/or consider more than one financial adviser before making a choice, and if these activities improve the level of trust financial consumers have in their financial advisers. consumers typically search for information on products and services before they buy or sell them. when a product is more expensive, the time and cost associated with the information search usually increases. financial decision-making is complex because consumers must understand the risks associated with their financial products and have the ability to project future economic scenarios and possible outcomes (lin and lee, 2004). as * corresponding author. tel.: �1-330-972-6330; fax: �1-330-972-5970. e-mail address: smg16@uakron.edu (s. gradisher) financial services review 23 (2014) 305–324 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. many retirement plans continue to transition from employer-managed defined benefit plans to employee-managed defined contribution plans, consumers are forced to make their own financial decisions. consumers can make these financial decisions themselves, or they can rely on financial advisers to assist them. as financial products and decision-making become more complex, an increasing proportion of financial consumers depend on the advice of financial professionals or advisers. previous literature from the late 1990s and early 2000s reports that between 21% and 25% of households use financial advisers (elmerick, montalto, and fox, 2002; lin and lee, 2004). however, the 2009 financial industry regulatory authority (finra) investor education foundation’s national financial capability survey (nfcs) shows that 56.7% of households responding to the survey used a financial adviser between 2004 and 2009. as an increasing number of consumers rely on financial advisers in their financial decision-making, there has been an increase in the number of people offering financial adviser services, making it imperative that financial consumers spend time in selecting a financial adviser who can best serve their needs. consumers have the ability to verify that financial advisers are licensed or registered, determine if they have been involved in professional misconduct, and ensure that they have adequate education and professional experience to give a reasonable assurance about the adviser’s competence, conduct, and reliability.1 some of this information can be obtained from the securities and exchange commission (sec). organizations such as aarp,2 individual state agencies,3 and the certified financial planner (cfp) board4 provide information regarding how to conduct background checks and what should be considered. despite the availability of this information to facilitate financial adviser background checks, we find that only 14.2% of financial consumers responding to the nfcs survey who have used financial advisers have checked the background of a financial adviser in the last five years. however, we find that nearly 46% responded that they considered more than one adviser before making a choice, indicating they see the importance associated with choosing a competent financial adviser. to the best of our knowledge, there is no academic study regarding financial adviser background checks. our study fills this gap while also adding to the growing body of literature pertaining to financial advisers and financial consumer decision-making (elmerick et al., 2002; finke, huston, and waller, 2009; lachance and tang, 2012; ligon, 2003; lin et al., 2004; jones, lesseig, and smythe, 2005). using the nfcs survey, we identify the characteristics of consumers who have conducted a background check of a financial adviser in the last five years and/or considered more than one financial adviser before making a choice. we also consider if these activities improve the level of trust financial consumers have in their financial advisers.5 throughout our analysis we consider five different types of financial advisers to include debt counselors, savings and investments, mortgage and loan, insurance, and tax planning advisers. our article is structured as follows. we provide an overview of the background check systems available for financial consumers in section 2. we explain the nfcs survey and methodology used in our analysis in section 3, and present empirical results in section 4. we summarize and conclude in section 5. 306 b. balasubramnian et al. / financial services review 23 (2014) 305–324 2. background information based on the investment advisers act, 1940, sec. 211 (g) (1), investment advisers are required to register with the sec, or with state agencies, based on the level of assets they manage. these registered investment advisers have fiduciary responsibilities meaning they must act in the best interest of their clients, and disclose any conflict of interest they may have to their clients. what is troubling is that not all types of financial professionals performing financial services have fiduciary responsibilities. for example, registered investment advisers have fiduciary responsibilities, but brokers do not have this higher fiduciary standard, only a relatively lower “suitability” standard.6 some financial planners, such as a certified financial planner (cfp) have fiduciary responsibilities, but several other similar designations do not. the dodd-frank act, 2010, section 913 requires the sec to consider changes in how different financial professional designations are required to have fiduciary responsibilities with their clients. the sec is currently in the process of making new rules for retail investors.7 many financial consumers are not aware of these subtle legal differences. for example, bernard madoff was a registered broker-dealer having a significant amount of assets under management, but he did not register as an investment adviser until the sec conducted an investigation in 2004. the madoff fraud case highlights the fact that many investors blindly trusted him, not realizing he had no fiduciary responsibility towards them. this highlights how background verification of financial advisers is an important activity for financial consumers. the costs associated with using financial advisers are usually transaction fees or fees related to total assets under management. because the financial adviser serves as an agent for the consumer (principal) there is also an agency cost associated with using financial advisers. finke et al. (2009) categorize transaction costs and asset management fees as direct cost and the additional agency monitoring costs as indirect costs. jensen and meckling (1976) subdivide the agency costs into monitoring costs, bonding costs, and residual losses.8 monitoring costs are incurred by the financial consumer (principal) when they go through the process of checking the background of financial advisers (agents) they are considering, and/or consider more than one financial adviser, before establishing a contractual relationship with a single adviser. we will refer to these activities as pre-selection monitoring to indicate they occur before choosing a financial adviser. theory predicts that the financial consumer (principal) will stop monitoring their financial adviser (agent) when the marginal costs of monitoring equal the marginal benefits. if the cost of searching for a financial adviser is likely to be too high in terms of money, time, and effort it is very likely that pre-selection monitoring will not be completed. prior literature considers who actually uses financial advisers and what types of information is used by financial consumers when making financial decisions. elmerick et al. (2002), using the 1998 survey of consumer finances, find people with higher education, income, and personal net worth are more likely to use a financial adviser to assist them in making financial decisions. lin and lee (2004), using the 2000–2001 macromonitor dataset, find that education, income, risk tolerance, and the dollar amount of investment positively influence financial consumer’s information search behavior to include what specific information sources (i.e., internet, professional advisers, etc.) are used in helping make financial 307b. balasubramnian et al. / financial services review 23 (2014) 305–324 decisions. lachance and tang (2012), using the 2009 nfcs survey, examine determinants of trust in financial professionals and the impact that trust has on the use of financial advisers. they find that trust declines with age and increases with willingness to take investment risk; having some financial literacy increases trust but having too much decreases trust; and that trust and cost are the two most important determinants of financial advice-seeking behavior. in this article we extend this prior literature by examining the characteristics associated with financial consumers who are willing to incur pre-selection monitoring costs before establishing a contractual relationship with a financial adviser, and if pre-selection monitoring improves the level of trust financial consumers have in their financial advisers. we consider two pre-selection monitoring activities associated with choosing a financial adviser: (1) conducting financial adviser background checks, and (2) considering more than one financial adviser before making a choice. 2.1. current system of background check elmerick et al. (2002) show that 50% of households use stock brokers, 25% use financial planners, 6% use accountants, 4% use bankers, and 2% use attorneys for financial advice. standards of professional conduct exist for cpas, attorneys, insurance agents, cfps with fiduciary responsibilities, and financial brokers without fiduciary responsibilities.9 zweig and pilon (2010) point out that according to finra, there are more than 95 professional designations for financial advisors, and 115 others not tracked by finra. ligon (2003) predicts that in the long run these various qualitative credentials will distinguish themselves through performance. brokers must be registered with the security and exchange commission (sec) and maintain membership with the finra. finra is committed to investor protection and market integrity through effective and efficient self-regulation of the securities industry. they accomplish this through enforcing rules governing the activities of security firms and brokers, enforcing educational standards such as the series 6 licensing examinations, promoting market transparency, and educating investors. finra also maintains the brokercheck10 system that can be used to check background information on brokers. brokercheck has information on 1.3 million current and former finra-registered brokers, 17,400 current and former finra-registered brokerage firms, �441,000 current and former investment adviser representatives, and 45,700 current and former investment adviser firms. for financial advisers, the sec adopted rule 204a-1 requiring sec-registered investment advisers to adopt and enforce codes of ethics11 that establish standards of conduct expected of supervised persons and reflect the adviser’s fiduciary duties. the sec offers the investment adviser public disclosure (iapd)12 website that can be used by consumers to check the backgrounds of registered investment advisers and the firms with which they are associated. the sec also maintains the investment advisers registration depository (iard) jointly with the north american securities administrators association (nasaa). finra is also responsible for maintaining the iard website, and distributes information on financial advisers from the iard database through their brokercheck system. out of 50 states, only 21 states have enacted regulations, or issued special notices, regarding the use of professional designations by registered investment advisers.13 the sec 308 b. balasubramnian et al. / financial services review 23 (2014) 305–324 permits financial advisers to satisfy their filing and registration obligations under state and federal law using a single electronic filing available at their website.14 the nasaa also advises that consumers should check with state regulators15 when conducting background checks of financial advisers. this would require the consumer to check 50 different regulators to verify whether the financial adviser has any complaints filed anywhere in the united states. considering the numerous and varied databases of financial professionals and firms that are maintained, it becomes apparent how time consuming it would be to complete a thorough background check of financial advisers. in a september 19, 2013, wall street journal article, daisy maxey (maxey, 2013) reports that there are several online directories that aggregate basic financial adviser background information from finra, iapd, and various state databases, but she warns about the “impartiality, conflict of interest, and possibility of abuse” that may exist in these online directories because of financial advisers having the ability to pay money to have their names included. in addition, it is possible for financial advisers to influence the priority of search results on these websites. financial consumers must pay a fee to access these aggregated background check lists, and it is possible they are getting inaccurate or incomplete information from these unofficial lists. besides the fragmentation of data for background verification among the sec, finra, and the various states, eaglesham and barry (2014a), in their wall street journal article, point out that securities brokers and investment advisors fail to disclose their personal bankruptcies and other criminal charges. such critical information is not recorded in brokercheck. numerous educational and job designations mislead financial consumers to trust financial advisors, which is exacerbated particularly among senior citizens. improving the financial literacy of consumers is also emphasized in the dodd-frank act, 2010 along with establishment of institutions for consumer protection such as the consumer financial protection bureau (cfpb). we suggest including financial adviser background checks as part of the financial literacy campaign. 3. finra survey and methodology 3.1. finra 2009 financial capability survey we use the 2009 finra investor education foundation national financial capability survey (nfcs)16 that was developed in consultation with the u.s. department of treasury and the president’s advisory council on financial literacy (finra, 2009). the finra investor education foundation (finra foundation) conducted the online survey of 28,146 respondents (�500 respondents per state and the district of columbia) over a five-month period between june and october of 2009. the survey provides an unprecedented level of data pertaining to financial behaviors across all 50 states and the district of columbia. the first survey question we consider is if the respondent had sought any advice from a financial adviser in one of five specific areas within the past five years. the five specific adviser areas included are debt counseling, savings and investments, mortgage or loan, insurance, and tax planning. possible responses included: yes, no, do not know, and 309b. balasubramnian et al. / financial services review 23 (2014) 305–324 prefer not to say. after limiting our sample to 27,273 respondents (from a total of 28,146 surveyed) who answered either yes or no, we find that 15,466 of the respondents (56.7%) used at least one of the five types of financial advisers in the past five years. focusing on these respondents who have used a financial adviser in the past five years, we consider the survey questions that ask if they have ever checked with a state or federal regulator regarding the background, registration, or license of a financial professional, and if they typically consider more than one financial adviser before making a choice. possible responses to these questions include yes and no only. we then consider how these respondents answer the question of if they trust financial professionals and accept what they recommend with possible responses being on a 1 (strongly disagree) to 7 (strongly agree) scale with 4 (neither) being the middle option. from the 15,466 respondents who have used a financial adviser in the past five year, we have a final sample of 15,188 respondents who also answered these additional questions of interest that we will use throughout our analysis. 3.2. methodology using a univariate analysis, we first identify the percentage of respondents in several demographic characteristic categories who have used a financial adviser within the past five years and have checked the background of a financial adviser, considered more than one adviser before making a choice, and have done both. we report these results for respondents who have used at least one type of financial adviser (fa user) and for each specific type of financial adviser (debt counselor, savings or investments, mortgage of loan, insurance, or tax planning) separately. the demographic characteristics considered include gender, age, ethnicity, education, marital status, income, employment, and region. these demographic characteristics have been shown to impact personal financial behavior and the probability of using a financial adviser when making financial decisions (barber and odean, 2001; elmerick et al., 2002). we then estimate a multinomial logit regression model using the same sample of respondents who used at least one type of financial adviser to determine what demographic characteristics impact the probability of a financial adviser user checking the background of financial advisers. the dependent variable, financial adviser background checks, is set equal to 1 if the financial adviser user checked the background of a financial adviser in the past five years, and 0 otherwise. we also estimate a multinomial logit model to determine what demographic characteristics impact the probability of a financial adviser user considering more than one adviser before making a choice. in this model, the dependent variable, considered �1 financial adviser, is set equal to 1 if the financial adviser user considered more than one financial adviser before making a choice, and 0 otherwise. the demographic characteristics included are female, age, black (non-hispanic), education, marital status, income, employment, and region. the categories within each demographic characteristic (i.e., high school within the education demographic characteristic) are set equal to 1 if the respondent identifies themselves to be in that categorical group, and 0 otherwise. positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to check the background of their adviser, or consider more than one financial adviser before making a choice, compared to the reference category. 310 b. balasubramnian et al. / financial services review 23 (2014) 305–324 we also consider the likelihood of checking the background of a financial adviser, or considering more than one financial adviser before making a choice, based on the type of financial adviser the survey respondent has used in the past five years while controlling for demographic characteristics. this is accomplished by estimating a similar multinomial logit model for both financial adviser background checks and considered �1 financial adviser where we include independent dummy variables for debt counseling, savings or investments, mortgage or loan, insurance, and tax planning that are set equal to 1 if the respondent used that type of financial adviser in the past five years, and 0 otherwise. we include the demographic characteristic variables as control variables in this model. finally, we consider the impact that conducting a background check on financial advisers, or considering more than one financial adviser before making a choice, has on the level of trust survey respondents have in their financial advisers. we use a tobit regression model where the dependent variable, trust in financial adviser, is a categorical variable ranging from 1 (strongly disagree) to 7 (strongly agree) based on the survey respondent’s answer to the question “i would trust financial professionals and accept what they recommend.” the independent variable of interest in the first tobit model is background check that is set equal to 1 if the respondent checked the background a financial adviser in the past five years, 0 otherwise. we also estimate a second tobit model where the independent variable of interest is considered �1 financial adviser that is set equal to 1 if the respondent considered more than one financial adviser before making a choice, 0 otherwise. we also include the demographic characteristic variables as control variables in both models as it would be expected that these characteristics would impact trust levels in financial advisers. 4. empirical results 4.1. summary statistics we present the summary statistics in table 1. of the 15,188 surveyed who responded that they used the services of a financial adviser in the last five years (fa users), only 14.2% claimed to have checked the background of a financial adviser (% checked background) during that same time period. however, 45.9% of financial adviser users considered more than one adviser before making a choice of who they used (% considered �1 adviser). there were 10.5% of financial adviser users who did both of these pre-selection monitoring activities (do both). we also provide these details for each category of financial adviser that was used such as debt counseling, savings and investments, mortgage or loan, insurance, and tax planning. higher percentages of financial consumers using debt counselors and tax planning services engage in pre-selection monitoring than those using other financial adviser types. in table 2 we present the number of financial adviser users (#) and percentages that checked financial adviser backgrounds (% bg check), considered more than one adviser (% �1 adviser), and that did both (do both) in each demographic characteristic. the demographic characteristics considered include gender, age, ethnicity, education, marital status, income, employment, and region. out of 8,115 (7,073) females (males) who have 311b. balasubramnian et al. / financial services review 23 (2014) 305–324 used a financial adviser, 12.7% (15.9%) conducted background checks, 42.9% (49.5%) considered more than one adviser, and 9% (12.3%) did both, indicating that males engage in more pre-selection monitoring when choosing a financial adviser compared with females. fewer financial adviser users in both the youngest (18–24) and oldest (65�) age groups engage in the pre-selection monitoring, while financial consumers between 35 and 54 engage in the most pre-selection monitoring when choosing a financial adviser. even though there are fewer blacks (3,477) than whites (11,711) using financial advisers, a much higher percentages of blacks (18.2%) conduct pre-selection monitoring compared with whites (13%). there is a monotonically increasing percentage of financial consumers engaging in pre-selection monitoring across higher levels of both education and income. there is a larger number of married financial adviser users (9,542) compared with single users (3,092), but there is a larger percentage of single users (15.2%) conducting pre-selection monitoring. there are a larger number of financial adviser users who have full-time employment, but a much higher percentage of users who are self-employed engage in pre-selection monitoring (19.9%). results are mixed across the geographical regions with a higher percentage of financial adviser users in the northeast who check backgrounds (16.7%), a higher percentage in the south (47.6%) and west (48.1%) who consider more than one adviser, and the highest percentage in the northeast who do both (12.7%). we present the same information for each demographic characteristic for each financial adviser type in table 3. the results are generally consistent across the different financial adviser types when considering gender, ethnicity, education, marital status, income, table 1 panel a: presents the analysis of the 15,188 survey respondents who have used a financial adviser in the last five years panel a yes no have you ever checked with state or federal regulators regarding the background, registration, or license of a financial professional? 14.2% 84.3% did you meet with or talk to more than one adviser before making a choice? 45.9% 45.7% panel b: presents the results for each of these adviser types separately and whether the respondents verified the background and considered more than one adviser (do both) panel b variable # % checked background % considered �1 adviser do both fa user 15,188 14.2% 45.9% 10.5% debt counselor 2,720 20.4% 54.0% 16.3% savings or investments 8,647 17.5% 49.8% 13.1% mortgage or loan 7,294 15.2% 49.0% 11.4% insurance 9,291 16.6% 50.3% 12.6% tax planning 5,060 21.0% 52.7% 16.0% we report the percentage of survey respondents that answered yes or no to the specified question. fa user indicates the survey respondent used at least one or more of the five different financial adviser types to include debt counselor, savings or investments, mortgage or loan, insurance, and tax planning. in panel a, we present the percentage of respondents who used financial adviser and checked the background of their adviser (% checked background) or considered more than one adviser (% considered �1 adviser). 312 b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 2 percent of the 15,188 survey respondents that used at least one of the five types of financial advisers (fa user) in the last five years and claimed to have checked the background of their adviser (% bg check), considered more than one adviser (% �1 adviser), and did both (do both) in each of the demographic characteristic categories to include gender, age, ethnicity, education, marital status, income, employment, and region fa user # % bg check % �1 adviser do both gender male 7,073 15.9% 49.5% 12.3% female 8,115 12.7% 42.9% 9.0% age 18–24 1,277 13.7% 46.0% 10.2% 25–34 2,688 14.4% 49.0% 10.8% 35–44 2,998 15.0% 49.6% 12.1% 45–54 3,283 14.7% 46.9% 10.9% 55–64 2,601 14.6% 44.1% 10.5% 65� 2,341 12.1% 38.7% 7.8% ethnicity white 11,711 13.0% 43.5% 9.2% black 3,477 18.2% 54.4% 14.9% education � high school 250 10.4% 37.2% 7.2% high school 2,876 10.8% 43.0% 8.0% some college 5,163 13.4% 47.0% 10.0% college graduate 4,135 15.4% 46.3% 11.5% post graduate 2,764 17.7% 47.4% 12.9% marital status married 9,542 14.2% 45.6% 10.4% single 3,092 15.2% 48.5% 11.5% separated 214 14.5% 50.0% 10.7% divorced 1,682 12.4% 44.9% 9.3% widowed 658 14.9% 40.4% 10.8% income �$15,000 1,071 11.3% 43.1% 8.7% $15k–$24,999 1,502 11.6% 43.1% 8.6% $25k–$34,999 1,697 11.8% 44.1% 8.8% $35k–$49,999 2,426 13.0% 45.3% 9.7% $50k–$74,999 3,252 13.7% 45.8% 9.6% $75k–$99,999 2,081 16.4% 47.4% 12.3% $100k–$149,000 1,930 16.8% 47.7% 12.6% $150,000 1,229 19.3% 51.3% 14.5% employment self employed 1,536 19.9% 53.1% 14.8% full-time 6,422 14.2% 46.4% 11.0% part-time 1,376 11.4% 44.1% 8.6% homemaker 1,205 12.3% 45.4% 9.0% student 536 14.9% 49.3% 11.4% disabled 462 14.3% 43.9% 8.9% unemployed 1,070 14.5% 48.7% 10.6% retired 2,581 13.0% 40.4% 8.6% region northeast 2,637 16.7% 45.3% 12.7% midwest 3,655 11.6% 42.0% 8.1% south 4,927 14.0% 47.6% 10.6% west 3,969 15.2% 48.1% 11.1% the five adviser types include debt counselor, savings or investments, mortgage or loan, insurance, and tax planning. 313b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 3 percent of the survey respondents that used a specific type of financial adviser in the past five years and claimed to have checked the background of their adviser (% bg check), considered more than one adviser (% �1 adviser), and did both (do both) in each of the demographic characteristic categories to include gender, age, ethnicity, education, marital status, income, employment, and region panel a: debt counselors and savings or investments debt counselor savings or investments # % bg check % �1 adviser do both # % bg check % �1 adviser do both gender male 1,238 23.8% 58.3% 19.6% 4,199 19.4% 53.9% 15.3% female 1,482 17.6% 50.4% 13.4% 4,448 15.7% 46.0% 11.1% age 18–24 236 26.3% 60.2% 20.3% 694 17.7% 50.4% 13.7% 25–34 660 21.7% 55.2% 18.5% 1,415 20.1% 55.9% 15.4% 35–44 659 20.2% 55.7% 15.8% 1,541 19.8% 54.6% 16.2% 45–54 626 19.0% 51.6% 15.0% 1,770 17.3% 51.4% 13.2% 55–64 350 19.1% 51.7% 14.6% 1,635 17.3% 48.0% 12.4% 65� 189 16.9% 48.7% 12.2% 1,592 13.3% 39.6% 8.7% ethnicity white 1,815 18.0% 49.9% 13.4% 6,779 15.8% 47.1% 11.3% black 905 25.4% 62.2% 21.9% 1,868 23.8% 59.5% 20.0% education �high school 55 12.7% 47.3% 5.5% 94 19.1% 43.6% 11.7% high school 584 16.6% 55.5% 13.4% 1,301 13.8% 46.3% 10.5% some college 1,045 17.7% 53.0% 13.9% 2,778 17.0% 51.7% 12.6% college graduate 692 24.3% 53.9% 19.8% 2,534 18.6% 49.7% 14.1% post graduate 344 28.8% 55.8% 23.0% 1,940 19.1% 50.0% 14.5% marital status married 1,553 20.5% 53.6% 15.9% 5,526 17.2% 49.2% 12.7% single 669 22.4% 58.4% 18.8% 1,748 19.2% 53.8% 15.2% separated 63 20.6% 60.3% 17.5% 93 21.5% 60.2% 15.1% divorced 352 15.3% 48.3% 11.9% 878 15.7% 48.4% 12.1% widowed 83 24.1% 44.6% 19.3% 402 16.4% 41.0% 12.2% income �$15,000 255 22.0% 51.4% 16.9% 488 15.6% 48.8% 11.9% $15k–$24,999 371 15.9% 51.8% 12.7% 663 14.8% 48.7% 11.2% $25k–$34,999 401 15.5% 55.9% 12.2% 836 14.6% 46.8% 11.4% $35k–$49,999 497 18.9% 53.3% 15.5% 1,250 17.0% 49.9% 13.0% $50k–$74,999 600 18.3% 52.5% 14.0% 1,872 17.2% 48.8% 12.3% $75k–$99,999 311 24.4% 57.6% 19.6% 1,304 19.2% 50.7% 14.5% $100k–$149,000 201 33.3% 55.2% 26.9% 1,335 18.5% 50.9% 14.2% �$150,000 84 38.1% 61.9% 32.1% 899 20.5% 53.2% 15.2% employment self employed 284 26.1% 64.8% 21.1% 929 22.9% 58.2% 17.3% full-time 1,257 20.3% 52.4% 16.6% 3,689 17.9% 51.2% 14.2% part-time 236 15.7% 52.5% 13.1% 779 14.8% 47.0% 11.4% homemaker 205 17.1% 47.8% 11.2% 586 15.0% 49.3% 10.8% student 107 28.0% 58.9% 20.6% 320 18.1% 54.1% 13.8% disabled 121 17.4% 38.0% 10.7% 170 20.0% 44.7% 10.6% unemployed 259 23.2% 62.9% 18.1% 506 19.6% 52.6% 14.0% retired 251 17.5% 52.6% 14.7% 1,668 14.7% 42.4% 10.0% region northeast 436 26.8% 53.4% 21.6% 1,613 21.1% 49.3% 16.2% midwest 669 14.5% 51.4% 11.7% 2,116 14.0% 45.4% 9.6% south 934 21.7% 55.6% 17.2% 2,712 17.3% 52.1% 13.5% west 681 20.4% 54.8% 16.0% 2,206 18.4% 51.6% 13.9% 314 b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 3 continued panel b: mortgage or loan and insurance mortgage or loan insurance # % bg check % �1 adviser do both # % bg check % �1 adviser do both gender male 3,374 17.4% 52.0% 13.6% 4,347 18.8% 53.6% 14.9% female 3,920 13.2% 46.5% 9.4% 4,944 14.7% 47.3% 10.6% age 18–24 563 16.0% 47.4% 12.4% 694 16.4% 50.6% 12.5% 25–34 1,622 15.0% 50.6% 11.0% 1,756 16.5% 53.4% 12.8% 35–44 1,718 15.2% 51.3% 12.1% 1,896 18.2% 53.3% 14.5% 45–54 1,574 16.3% 49.9% 12.5% 2,125 16.6% 50.6% 12.6% 55–64 1,058 15.9% 47.3% 12.0% 1,563 17.7% 49.1% 13.1% 65� 759 11.5% 42.6% 6.5% 1,257 13.2% 42.2% 9.0% ethnicity white 5,662 13.4% 46.7% 9.6% 7,116 15.1% 47.8% 11.0% black 1,632 21.2% 57.2% 17.5% 2,175 21.6% 58.3% 18.0% education �high school 91 8.8% 35.2% 4.4% 153 11.1% 34.6% 7.2% high school 1,213 12.4% 45.6% 9.3% 1,701 12.6% 47.5% 9.6% some college 2,430 14.4% 50.0% 10.9% 3,168 15.0% 51.0% 11.4% college graduate 2,160 16.0% 49.9% 12.0% 2,564 18.4% 51.1% 14.3% post graduate 1,400 17.9% 49.9% 13.3% 1,705 21.5% 51.8% 15.9% marital status married 5,014 14.3% 48.4% 10.5% 5,988 16.4% 49.7% 12.4% single 1,264 18.2% 52.6% 14.4% 1,789 17.7% 53.7% 13.9% separated 103 18.4% 49.5% 14.6% 131 19.1% 50.4% 13.0% divorced 684 14.9% 49.1% 11.4% 1,016 15.3% 49.3% 11.5% widowed 229 16.2% 43.2% 11.4% 367 17.2% 45.8% 12.5% income �$15,000 331 15.7% 47.7% 11.5% 625 12.8% 47.0% 9.8% $15k–$24,999 552 13.4% 47.6% 10.3% 897 12.9% 46.9% 9.8% $25k–$34,999 676 12.9% 46.6% 9.5% 1,019 12.9% 50.2% 10.1% $35k–$49,999 1,165 13.9% 48.1% 10.6% 1,494 15.7% 49.6% 11.8% $50k–$74,999 1,684 13.5% 49.3% 9.8% 1,954 16.6% 51.1% 11.9% $75k–$99,999 1,163 17.3% 48.9% 13.2% 1,305 18.8% 51.5% 14.6% $100k–$149,000 1,081 16.4% 50.7% 12.3% 1,203 19.5% 49.7% 15.2% �$150,000 642 19.5% 52.0% 14.6% 794 22.8% 54.7% 17.1% employment self employed 793 21.2% 54.9% 15.9% 1,037 22.1% 55.5% 16.3% full-time 3,552 14.6% 48.5% 11.5% 3,967 17.2% 51.1% 13.7% part-time 559 13.1% 48.3% 10.6% 839 13.6% 47.3% 10.6% homemaker 647 12.7% 48.5% 9.4% 758 13.5% 49.9% 10.0% student 219 18.3% 53.4% 14.6% 293 17.1% 58.4% 13.7% disabled 195 18.5% 48.2% 10.8% 310 15.8% 45.5% 9.4% unemployed 464 16.4% 51.5% 11.0% 646 16.9% 53.7% 12.5% retired 865 13.2% 44.7% 8.3% 1,441 14.6% 44.0% 10.1% region northeast 1,195 18.6% 48.2% 14.5% 1,535 19.9% 49.1% 15.8% midwest 1,689 12.1% 44.6% 8.5% 2,262 13.3% 45.9% 9.3% south 2,361 15.0% 50.7% 11.6% 3,025 16.5% 52.5% 13.0% west 2,049 15.9% 51.3% 11.7% 2,469 17.8% 52.3% 13.2% 315b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 3 continued panel c: tax planning tax planning # % bg check % �1 adviser do both gender male 2,457 23.8% 56.0% 18.8% female 2,603 18.3% 49.6% 13.3% age 18–24 383 23.0% 53.0% 17.5% 25–34 905 23.5% 56.5% 18.0% 35–44 959 22.7% 55.5% 18.1% 45–54 1,034 21.3% 55.1% 16.6% 55–64 903 19.6% 51.6% 15.0% 65� 876 16.7% 44.1% 11.1% ethnicity white 3,937 18.5% 49.8% 13.4% black 1,123 29.6% 62.9% 25.0% education � high school 56 19.6% 33.9% 10.7% high school 713 19.1% 48.8% 15.1% some college 1,522 19.6% 55.0% 14.5% college graduate 1,521 21.4% 52.5% 16.2% post graduate 1,248 23.2% 53.4% 18.2% marital status married 3,496 19.9% 52.4% 15.0% single 867 24.8% 56.2% 19.4% separated 58 34.5% 55.2% 25.9% divorced 425 20.5% 52.9% 16.2% widowed 214 21.5% 43.5% 15.4% income �$15,000 217 23.5% 52.5% 16.6% $15k–$24,999 331 18.1% 49.2% 13.0% $25k–$34,999 392 19.9% 51.8% 16.1% $35k–$49,999 691 20.8% 53.1% 15.5% $50k–$74,999 1,099 19.4% 52.0% 13.7% $75k–$99,999 794 21.9% 52.0% 18.0% $100k–$149,000 830 21.7% 53.9% 17.3% �$150,000 706 22.9% 55.2% 17.1% employment self employed 729 22.2% 58.4% 17.7% full-time 2,055 22.4% 54.1% 17.7% part-time 464 18.8% 49.1% 14.2% homemaker 377 15.9% 52.3% 10.9% student 151 25.8% 58.3% 18.5% disabled 74 28.4% 47.3% 16.2% unemployed 286 25.2% 52.8% 19.9% retired 924 17.4% 46.6% 12.1% region northeast 964 24.6% 52.8% 19.2% midwest 1,212 16.1% 47.9% 11.3% south 1,547 21.3% 54.3% 16.7% west 1,337 22.5% 55.3% 17.1% panel a reports results for debt counselor and savings or investments advisers, panel b reports results for mortgage or loan and insurance advisers, and panel c reports results for tax planning advisers. 316 b. balasubramnian et al. / financial services review 23 (2014) 305–324 employment, and region. however, the results are more mixed when considering age. there is a higher percentage of the 18–24 age group conducting pre-selection monitoring when considering debt counselors (panel a). all other results regarding age are mixed. 4.2. regression results in table 4 we present multinomial logit regression results that indicate how the different demographic characteristics impact the probability of conducting a background check on a financial adviser in the past five years (left side of table 4) and the probability of considering more than one financial adviser before making a choice (right side of table 4). we find that females are 20% less likely to have checked the background of financial advisers compared with males. we do not observe any significant differences among age groups except in the 65� age group that is, on average, 21.7% less likely to check financial adviser backgrounds compared with the 18–24 age group. we also find there are no significant differences among the different education groups except that financial adviser users with post-graduate education are 50% more likely to check financial adviser backgrounds compared with those who did not complete high school. compared with married financial adviser users, living with partner (single) users are nearly 20% (14%) more likely to check financial adviser backgrounds. financial adviser users with more than $35,000 are more likely to check financial adviser backgrounds compared with users earning less than $35,000, with this probability increasing monotonically across higher income levels. all the employment categories are less likely to check financial adviser backgrounds compared with self-employed users with full-time users being 37.4% less likely to check financial adviser backgrounds compared with self-employed users. finally, financial adviser users living in the midwest, south, or west are all less likely to check financial adviser backgrounds compared with those living in northeast, with those in midwest region being 30% less likely compared with those in the northeast region. the results for the multinomial logit regression that considers the probability of considering more than one financial adviser are reported on the right side of table 4. consistent with the background check results, females are 20% less likely than males, and blacks are over 40% more likely than whites, to consider more than one financial adviser before making a choice. there also remains a monotonic increase in the probability of considering more than one adviser across higher income levels, no statistically significant difference across education levels, and married couples remain less likely to consider more than one adviser compared with those that are living with a partner or single. self-employed financial adviser users remain more likely to consider more than one adviser which is also consistent with the background check results. however, we find financial consumers between the ages of 25 and 44 are more likely to consider more than one financial adviser compared with the 18–24 age group whereas those 65� are almost 22% less likely to consider more than one financial adviser. this is somewhat different for the 25 to 44 age group, but also points out that the 65� age group is the least likely to conduct any pre-selection monitoring before choosing a financial adviser. we also find that financial adviser users from the midwest are less likely to consider more than one adviser compared with those from the northeast that is consistent with our background check tests, however, there is no significant difference between the 317b. balasubramnian et al. / financial services review 23 (2014) 305–324 remaining regions indicating that users in the south and west are just as likely to consider more than one financial adviser as those in the northeast. table 4 multinomial logit regression using 15,188 survey respondents who have used a financial adviser in the past five years where the dependent variable, financial adviser background checks, is set equal to 1 if the respondent checked the background of a financial adviser in the past five years, 0 otherwise financial adviser background checks considered �1 financial advisers coefficient z-stat standard error odds ratio coefficient z-stat standard error odds ratio constant �1.849*** �7.21 0.256 0.157 �0.152 �0.85 0.179 0.859 female �0.219*** �4.45 0.049 0.803 �0.227*** �6.31 0.036 0.797 age group (reference category: 18–24) 25–34 �0.045 �0.42 0.107 0.956 0.175** 2.26 0.078 1.192 35–44 �0.043 �0.40 0.108 0.958 0.163** 2.09 0.078 1.177 45–54 �0.042 �0.39 0.108 0.959 0.057 0.74 0.078 1.059 55–64 �0.070 �0.61 0.115 0.932 �0.056 �0.67 0.083 0.946 65� �0.244* �1.83 0.133 0.783 �0.242** �2.57 0.094 0.785 black 0.421*** 7.63 0.055 1.523 0.345*** 8.11 0.043 1.413 education (reference category: did not complete high school) high school 0.013 0.06 0.218 1.013 0.137 0.93 0.147 1.147 some college 0.197 0.91 0.215 1.217 0.261* 1.79 0.146 1.298 college graduate 0.291 1.34 0.217 1.338 0.171 1.16 0.148 1.186 post graduate 0.405* 1.84 0.220 1.499 0.191 1.27 0.151 1.211 marital status (reference category: married) living with partner 0.181** 1.98 0.091 1.198 0.181*** 2.60 0.070 1.198 single 0.128** 2.10 0.061 1.136 0.073* 1.65 0.044 1.076 income (reference category: less than $15,000) $15,000–$24,999 0.119 0.92 0.130 1.126 0.058 0.66 0.088 1.060 $25,000–$34,999 0.186 1.44 0.129 1.205 0.134 1.51 0.089 1.144 $35,000–$49,999 0.296** 2.38 0.124 1.345 0.159* 1.85 0.086 1.172 $50,000–$74,999 0.355*** 2.86 0.124 1.426 0.162* 1.88 0.086 1.176 $75,000–$99,999 0.529*** 4.04 0.131 1.697 0.216** 2.32 0.093 1.241 $100,000–$149,999 0.546*** 4.04 0.135 1.726 0.197** 2.04 0.096 1.218 $150,000 or more 0.685*** 4.80 0.143 1.985 0.360*** 3.45 0.104 1.433 employment (reference category: self employed) employed full-time �0.468*** �6.21 0.075 0.626 �0.345*** �5.64 0.061 0.709 employed part-time �0.509*** �4.67 0.109 0.601 �0.275*** �3.45 0.080 0.760 homemaker �0.309*** �2.68 0.115 0.734 �0.129 �1.50 0.086 0.879 full-time student �0.287* �1.88 0.152 0.751 �0.167 �1.44 0.116 0.846 disabled �0.133 �0.87 0.153 0.876 �0.205* �1.76 0.116 0.815 unemployed �0.219** �1.96 0.112 0.803 �0.122 �1.41 0.086 0.885 retired �0.244** �2.37 0.103 0.783 �0.277*** �3.54 0.078 0.758 region (reference category: northeast) midwest �0.357*** �4.79 0.074 0.700 �0.127** �2.35 0.054 0.881 south �0.213*** �3.15 0.068 0.808 0.072 1.41 0.051 1.075 west �0.139** �1.99 0.070 0.870 0.080 1.49 0.053 1.083 the dependent variable, considered �1 financial adviser, is set equal to 1 if the respondent considered more than one financial adviser before making a choice, 0 otherwise. the independent variables include female, age, black (non-hispanic), education, marital status, income, employment, and region. positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to check the background of their adviser, or consider more than one financial adviser before making a choice, compared with the indicated reference category. ***,**,* denote significance at the 1%, 5%, and 10% levels, respectively. 318 b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 5 reports results where we estimate the same multinomial logit model with additional dummy variables included that indicate if the financial adviser user used that specific type of financial adviser. this allows us to estimate the probably of checking the background of a financial adviser (left side of table), or considering more than one financial adviser (right side of table), based on the type of financial adviser that was used while controlling for the demographic characteristics included earlier that have already been shown to impact these same probabilities. there are five different financial adviser types included: debt counseling, savings and investments, mortgage or loan, insurance, and tax planning. consistent with the univariate results reported in table 1, we find, after controlling for all demographic characteristics, those using debt counseling or tax planning services are more likely to check the backgrounds of their financial advisers. those using insurance advisers are the least likely to check backgrounds, whereas there is no statistically significant relationship between checking backgrounds and using mortgage or loan services. however, we find that those using debt counselors and insurance advisers as most likely to consider more than one financial adviser before making a choice. those using savings or investments and mortgage or loan advisers are least likely to consider more than one adviser before choosing. all of these results are statistically significant at the 1% level. 4.3. do financial consumers who conduct pre-selection monitoring have higher levels of trust in their financial advisers? next, we consider if those financial adviser users who conduct some form of pre-selection monitoring before choosing a financial adviser have a higher level of trust in the financial advisers they use. we use the responses to the survey question: “i would trust financial professionals and accept what they recommend.” we report the percentage of financial adviser users who selected each of the available responses to include strongly disagree, disagree, neither agree nor disagree, agree, and strongly agree in table 6. overall, out of the 15,188 financial adviser users, there are nearly 40% who agree, 34% who neither agree nor disagree, and over 23% who disagree with this statement (some users did not answer this question). in table 7, we estimate a tobit regression model using the full sample of 15,188 financial adviser users to determine the impact that pre-selection monitoring has on the answer to this question. we control for the all the demographic characteristics used earlier and include either background check (left side of table 7) as an additional dummy variable indicating if the user conducted a financial adviser background check, or considered �1 financial adviser (right side of table 7) as an additional dummy variable indicating if the user considered more than one financial adviser before making a choice. overall, we find many demographic characteristics are not statistically related to financial adviser trust levels such as gender, education, and income. consistent with lachance and tang (2012), we find that trust levels generally decline with age. we also find that living with partner couples have lower trust levels, and people in the midwest have higher trust levels than those from other regions. most importantly, we find a positive and highly significant (significant at the 1% level) relationship between checking the background of financial advisers and financial 319b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 5 multinomial logit regression using 15,188 survey respondents who have used a financial adviser in the past five years where the dependent variable, financial adviser background checks, is set equal to 1 if the respondent checked the background of a financial adviser in the last five years, 0 otherwise financial adviser background checks considered �1 financial adviser coefficient z-stat standard error odds ratio coefficient z-stat standard error odds ratio constant �2.760*** �10.43 0.265 0.0633 �0.746*** �4.06 0.184 0.474 debt counseling 0.602*** 10.21 0.059 1.825 0.364*** 7.75 0.047 1.439 savings or investments 0.485*** 9.02 0.054 1.624 0.221*** 5.96 0.037 1.248 mortgage or loan 0.077 1.55 0.050 1.080 0.214*** 5.94 0.036 1.239 insurance 0.416*** 7.85 0.053 1.516 0.388*** 10.78 0.036 1.474 tax planning 0.561*** 11.01 0.051 1.752 0.247*** 6.44 0.038 1.281 female �0.209*** �4.16 0.050 0.812 �0.234*** �6.44 0.036 0.791 age group (reference category: 18–24) 25–34 �0.095 �0.87 0.110 0.909 0.119 1.51 0.079 1.127 35–44 �0.020 �0.18 0.110 0.980 0.153* 1.93 0.079 1.165 45–54 �0.009 �0.08 0.110 0.991 0.063 0.79 0.079 1.065 55–64 �0.020 �0.17 0.117 0.980 �0.024 �0.29 0.084 0.976 65� �0.180 �1.32 0.136 0.835 �0.185* �1.93 0.096 0.831 black (non-hispanic) 0.368*** 6.49 0.057 1.445 0.329*** 7.60 0.043 1.390 education (reference category: did not complete high school) high school �0.012 �0.05 0.221 0.988 0.139 0.93 0.149 1.149 some college 0.110 0.50 0.218 1.116 0.222 1.51 0.147 1.248 college graduate 0.179 0.81 0.220 1.196 0.119 0.80 0.149 1.126 post graduate 0.254 1.14 0.224 1.290 0.122 0.80 0.153 1.130 marital status (reference category: married) living with partner 0.212** 2.28 0.093 1.236 0.200*** 2.84 0.070 1.221 single 0.156** 2.50 0.062 1.168 0.109** 2.41 0.045 1.115 income (reference category: less than $15,000) $15,000–$24,999 0.107 0.81 0.132 1.112 0.042 0.47 0.089 1.043 $25,000–$34,999 0.149 1.13 0.132 1.160 0.109 1.21 0.090 1.115 $35,000–$49,999 0.234* 1.84 0.127 1.263 0.115 1.32 0.087 1.122 $50,000–$74,999 0.261** 2.06 0.127 1.298 0.106 1.20 0.088 1.112 $75,000–$99,999 0.415*** 3.09 0.134 1.514 0.146 1.53 0.095 1.157 $100,000–$149,999 0.411*** 2.96 0.139 1.508 0.123 1.25 0.099 1.131 $150,000 or more 0.497*** 3.38 0.147 1.643 0.263** 2.46 0.107 1.301 employment (reference category: self employed) employed full-time �0.344*** �4.44 0.077 0.709 �0.280*** �4.51 0.062 0.756 employed part-time �0.415*** �3.73 0.111 0.661 �0.206** �2.55 0.081 0.814 homemaker �0.165 �1.40 0.118 0.848 �0.046 �0.53 0.087 0.955 full-time student �0.162 �1.05 0.155 0.850 �0.079 �0.67 0.118 0.924 disabled 0.038 0.25 0.156 1.039 �0.141 �1.19 0.118 0.869 unemployed �0.108 �0.94 0.114 0.898 �0.052 �0.60 0.088 0.949 retired �0.145 �1.38 0.105 0.865 �0.209*** �2.63 0.079 0.812 region (reference category: northeast) midwest �0.372*** �4.91 0.076 0.689 �0.143*** �2.61 0.055 0.867 south �0.203*** �2.93 0.069 0.816 0.070 1.35 0.052 1.072 west �0.120* �1.68 0.071 0.887 0.071 1.32 0.054 1.074 the dependent variable, considered �1 financial adviser, is set equal to 1 if the respondent considered more than one financial adviser before making a choice, 0 otherwise. the independent variables debt counseling, savings or investments, mortgage or loan, insurance, tax planning are variables set equal to 1 if the respondents used that type of financial adviser in the past five years, 0 otherwise. the other independent variables include female, age, black (non-hispanic), education, marital status, income, employment, and region. positive (negative) coefficient estimates indicate that demographic characteristic is more (less) likely to check the background of their adviser, or consider more than one financial adviser before making a choice, compared with the indicated reference category. ***,**,* denote significance at the 1%, 5%, and 10% levels, respectively. 320 b. balasubramnian et al. / financial services review 23 (2014) 305–324 adviser trust levels. however, we find no significant relationship between considering more than one financial adviser and trust levels in financial advisers. in summary, these results may indicate that checking a financial adviser’s background helps develop trust in that financial adviser. however, simply considering more than one adviser does not necessarily develop higher trust levels in the financial adviser that is eventually chosen. perhaps post-selection monitoring is more important for financial adviser users who only consider more than one adviser as their only pre-selection monitoring activity. overall, our results indicate that having a reliable system in place that allows financial consumers to conduct background checks of financial advisers in an easy and efficient manner may help develop trust in financial advisers. 5. summary and conclusions using the 2009 nfcs conducted by finra we identify demographic characteristics associated with financial consumers who conduct prescreening monitoring when choosing financial advisers. the two prescreening monitoring considered include checking the background of financial advisers and considering more than one financial adviser before making a choice. we also test if these prescreening monitoring improve the level of trust financial consumers have in their financial advisers. we consider five different financial adviser types throughout our study to include debt counseling, savings and investments, mortgage and loan, insurance, and tax planning. we find that only 14.2% of financial adviser users responding to the nfcs survey have checked the background of a financial adviser in the past five years, but nearly 46% of these financial adviser users considered more than one adviser before making a choice. higher percentages of financial consumers using debt counselors and tax planning services engage in pre-selection monitoring than those using other financial adviser types. based on multinomial logit regression estimates, we find that females are less likely, and blacks (compared with whites) more likely, to conduct pre-selection monitoring. we also find the probability of conducting pre-selection monitoring is monotonically increasing at higher income levels above $35,000/year, but decreases significantly for consumers over the age of 65. finally, financial consumers who are self-employed, and those living in the northeast (compared with the midwest, west, and south) are more likely to conduct pre-selection monitoring. table 6 responses from 15,188 survey respondents who have used a financial adviser in the past five years to the question regarding their level of trust in financial professionals and if they would accept what they recommend i would trust financial professionals and accept what they recommend. strongly disagree 5.4% disagree 18.2% neither agree nor disagree 34.0% agree 32.7% strongly agree 6.6% total percentages to answer in each category ranging from 1 (strongly disagree) to 7 (strongly agree), with the middle category 4 indicating that they neither agree nor disagree. 321b. balasubramnian et al. / financial services review 23 (2014) 305–324 table 7 tobit regression using 15,188 survey respondents who have used a financial adviser in the past five years where the dependent variable, trust financial adviser, is a categorical variable ranging from 1 (strongly disagree) to 7 (strongly agree) trust in financial adviser trust in financial adviser coefficient t-stat standard error coefficient t-stat standard error constant 4.642*** 31.97 0.145 4.671*** 32.12 0.145 background check 0.108*** 2.60 0.041 considered �1 financial adviser �0.001 �1.30 0.001 female 0.034 1.14 0.030 0.033 1.11 0.030 age group (reference category: 18–24) 0 0 0 25–34 �0.194*** �3.01 0.064 �0.194*** �3.00 0.064 35–44 �0.506*** �7.78 0.065 �0.505*** �7.78 0.065 45–54 �0.680*** �10.53 0.065 �0.680*** �10.53 0.065 55–64 �0.776*** �11.28 0.069 �0.776*** �11.28 0.069 65� �0.698*** �8.86 0.079 �0.701*** �8.90 0.079 black (non-hispanic) �0.007 �0.19 0.036 �0.003 �0.09 0.036 education (reference category: did not complete high school) 0 0 0 high school �0.051 �0.43 0.118 �0.055 �0.47 0.118 some college �0.024 �0.21 0.117 �0.027 �0.23 0.117 college graduate �0.008 �0.06 0.118 �0.009 �0.07 0.118 post graduate �0.021 �0.17 0.121 �0.022 �0.18 0.121 marital status (reference category: married) 0 0 0 living with partner �0.170*** �2.95 0.057 �0.167*** �2.91 0.057 single 0.022 0.58 0.037 0.022 0.60 0.037 income (reference category: less than $15,000) 0 0 0 $15,000–$24,999 0.011 0.15 0.073 0.011 0.15 0.073 $25,000–$34,999 0.085 1.16 0.073 0.085 1.17 0.073 $35,000–$49,999 0.090 1.27 0.071 0.091 1.28 0.071 $50,000–$74,999 0.114 1.61 0.071 0.115 1.62 0.071 $75,000–$99,999 0.131* 1.70 0.077 0.134* 1.74 0.077 $100,000–$149,999 0.095 1.18 0.080 0.096 1.20 0.080 $150,000 or more 0.124 1.43 0.087 0.128 1.47 0.087 employment (reference category: self employed) 0 0 0 employed full-time 0.190*** 3.74 0.051 0.184*** 3.62 0.051 employed part-time 0.077 1.16 0.066 0.069 1.04 0.066 homemaker �0.009 �0.13 0.071 �0.013 �0.19 0.071 full-time student 0.233** 2.42 0.096 0.229** 2.38 0.096 disabled 0.009 0.09 0.096 0.008 0.08 0.096 unemployed 0.112 1.56 0.072 0.108 1.51 0.072 retired 0.117* 1.78 0.066 0.111* 1.70 0.066 region (reference category: northeast) 0 0 0 midwest 0.118*** 2.62 0.045 0.114** 2.52 0.045 south 0.026 0.60 0.043 0.023 0.53 0.043 west �0.036 �0.80 0.045 �0.037 �0.83 0.045 the independent variable of interest are background check that is set equal to 1 if the respondent checked the background of a financial adviser in the last five years, 0 otherwise, and considered �1 financial adviser, which is set equal to 1 if the respondent considered more than one financial adviser before making a choice, 0 otherwise. the other independent variables include female, age, black (non-hispanic), education, marital status, income, employment, and region. positive (negative) coefficient estimates indicate that variable is more (less) likely to trust their financial adviser compared with the indicated reference category. ***,**,* denote significance at the 1%, 5%, and 10% levels, respectively. 322 b. balasubramnian et al. / financial services review 23 (2014) 305–324 when considering the specific financial adviser types, we find that those using debt counseling or tax planning services are more likely to check the backgrounds of their financial advisers, and those using insurance advisers are the least likely to check backgrounds. those using debt counselors and insurance advisers are most likely, and those using savings or investments and mortgage or loan advisers are least likely, to consider more than one financial adviser before making a choice. we also consider how conducting pre-selection monitoring, controlling for demographic characteristics, impact the level of trust financial adviser users have in their financial advisers. we find that there is a positive relationship between financial adviser background checks and financial adviser trust levels. however, we find no significant relationship between considering more than one financial adviser and trust levels in financial advisers. we also find that trust levels generally decline with age, which is consistent with lachance and tang (2012) findings. we also find that living with partner couples have lower trust levels (compared with married couples and those who are single), and people in the midwest have higher trust levels than those from other regions. in summary, these results show that checking a financial adviser’s background helps develop trust in that financial adviser. however, simply considering more than one adviser does not necessarily develop higher trust levels in the financial adviser that is eventually chosen. overall, our results indicate that having a reliable and well-known background check system in place that allows financial consumers to conduct background checks of financial advisers in an easy and efficient manner may help improve trust in financial advisers. notes 1 http://www.sec.gov/investor/brokers.htm. 2 http://assets.aarp.org/www.aarp.org_/articles/bulletin/money/financialquestionnaire. pdf. 3 http://www.oag.state.md.us/forms/checklist.pdf. 4 http://www.letsmakeaplan.org/cfp-pros-their-expertise/cfp-experts-corner/article/letsmake-a-plan-blogs/things-to-think-about-when-choosing-a-financial-adviser; http:// www.letsmakeaplan.org/working-with-a-financial-planner/what-to-ask. 5 we do not use the 2012 financial capability survey data because this survey does not contain any questions on whether the respondents verified the credentials or how much they trust their financial advisers or their advice. 6 http://online.wsj.com/news/articles/sb10001424052702304679404579459831342132 534. 7 http://www.sec.gov/spotlight/investor-advisory-committee-2012/fiduciary-dutyrecommendation.pdf. 8 bonding costs are incurred by the agent (financial adviser) to signal to the principal of his fiduciary responsibilities through certifications and other binding constraints. residual loss will include losses suffered by the principal (consumer) because of agent’s (financial adviser’s) action that may not maximize the consumer’s wealth. 323b. balasubramnian et al. / financial services review 23 (2014) 305–324 9 appendix a provides details about various certification and maintenance requirements for various professional designations. 10 http://www.finra.org/investors/toolscalculators/brokercheck/. 11 http://www.sec.gov/rules/final/ia-2256.htm. 12 http://www.adviserinfo.sec.gov/iapd/content/search/iapd_search.aspx. 13 http://www.finra.org/investors/protectyourself/beforeyouinvest/p120759. 14 http://www.sec.gov/foia/docs/invafoia.htm. 15 http://www.nasaa.org/2709/how-to-check-out-your-broker-or-investment-adviser/. 16 please see the following link for the complete questionnaire: http://www.usfinancial capability.org/downloads/nfcs_2009_natl_qre_eng.pdf. references barber, b. m., & odean, t. (2001). boys will be boys: gender, overconfidence, and common stock investment. the quarterly journal of economics, 116, 261–292. bundrick, h. m. (2013). hiring a financial adviser: making sense of certifications. (available at http://www. thestreet.mobi/story/12036906/1/hiring-a-financial-adviser-making-sense-of-certifications.html?puc�yahoo &cm_ven�yahoo). consumer financial protection board. (2013). senior designations for financial advisers: reducing consumer confusion and risks. report to the congress by the consumer financial protection bureau (available at http://files.consumerfinance.gov/f/201304_cfpb_olderamericans_report.pdf). eaglesham, j., & barry, r. (2014). stockbrokers fail to discluse red flags. the wall street journal, march 5. elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217–231. finke, m., huston, s., & waller, w. (2009). do contracts impact comprehensive financial advice? financial services review, 18, 177–194. finra. (2009). financial capability in the united states national survey. http://usfinancialcapability.org/ downloads/nfcs_2009_state_by_state_qre.pdf jensen, m., & meckling, w. h. (1976). theory of the firm: managerial behavior, agency costs, and ownership structure. journal of financial economics, 3, 305–360. jones, m. a., lesseig, v. p., & smythe, t. i. (2005). financial advisers and multiple share class mutual fund. financial services review, 14, 1–20. lachance, m., & tang, n. (2012). financial advice and trust. financial services review, 21, 209–226. ligon, j. a. (2003). search, adverse selection, and the services of financial experts. financial services review, 12, 275–291. lin, q., & lee, j. (2004). consumer information search when making investment decisions. financial services review, 13, 319–332. maxey, d. (2013). adviser registries draw questions about impartiality. the wall street journal, september 19. zweig, j., & pilon, m. (2010). is your advisor pumping up his credentials? those fancy initials after your financial advisor’s name might not be as impressive as they seem. the wall street journal, october 16. 324 b. balasubramnian et al. / financial services review 23 (2014) 305–324 investment strategies when selecting sustainable firms todd m. shank, ph.d.a,*, benjamin shockey, mbab adepartment of finance, kate tiedemann college of business, university of south florida st. petersburg, 140 7th avenue, south, st. petersburg, fl 33701, usa braymond james financial, inc., 880 carillion parkway, st. petersburg, fl 33716, usa abstract “sustainability” is the most recent construct to describe efforts by modern corporations to include environmental, social, and economic governance considerations in business operations. such nonfinancial areas of firm performance have received increased focus, especially in recent years as firms seek to differentiate themselves from competitors. of central concern to analysts and investors is whether the emerging emphasis on sustainability is financially rewarded by market participants. research on this question has generally supported the idea that firms with greater attention to sustainable business practices outperform their peers financially; presumably because of the perception of lesser future risk. this study examines the efficacy of passive versus active investment strategies when selecting sustainable firms for inclusion within an equity portfolio. utilizing two groups of “sustainability-focused” firms of varying degrees, this study finds financial support for an active selection of sustainable firms on a risk-adjusted basis. specifically, the financial performance of the dow jones sustainability index (djsi) and a group of sustainability leaders are compared with the broader market over a 10-year period. thus, it can be understood whether an investor would have received a greater risk-adjusted financial return through either (1) an active “sustainability-focused” approach, (2) a passive sustainability-focused approach, or (3) a passive broad-market approach. the evidence presented here supports efforts to identify global sustainability leaders by industry, as they collectively showed greater financial performance over the past decade than both the djsi and the broader market. © 2016 academy of financial services. all rights reserved. jel classification: g11 (portfolio choice: investment decisions) keywords: sustainability; dow jones sustainability index (djsi); investment selection; financial performance; active portfolio management * corresponding author. tel.: �1-727-873-4894; fax: �1-727-873-4571. e-mail address: tshank@mail.usf.edu (t.m. shank) financial services review 25 (2016) 199–214 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. 1. introduction over the past two decades, numerous researchers within the financial industry have attempted to establish a link between a firm’s financial performance and their proclivity to embrace sustainable business practices. in doing so, these researchers have attempted to find significant variance between the financial performance of companies that embrace sustainable business practices and those whose concern for environmental, social, and governance competencies (esg) is less evident. the results of these studies have relevance not only for managers, analysts, and investors, but also policymakers and regulators who are increasingly turning their attention to non-financial outcomes of corporate performance. given that non-financial aspects of corporate performance are less quantifiable, and thus less readily identified by investors (compared with the traditional financial performance outcomes), informational inefficiencies can be expected in stock prices of firms who assign greater emphasis to esg factors. presumably, finance theory supports the argument that a company’s overall value increases as it better incorporates sustainable business strategies into its operations. this is based on the logical connection between the implementation of such strategies and the stabilization of future cash flows and reduction of corporate risk exposure. to date, there has been a disparity in the results of existing corporate sustainability performance (csp)– corporate financial performance (cfp) studies, much of which can be attributed to the ways in which corporate environmental, social, and governance performance are determined and how financial performance is measured. while there are many established metrics for financial performance, including the market price of a firm’s stock or financial ratio analyses of historical performance, the metrics for non-financial corporate financial performance are still in the early stages of development. for those investors that believe csp is relevant to the firm selection aspect of their investment strategy, and that markets for sustainable firm equities is semi-strong–form efficient, they have the option of passively investing in one of the hundreds of “socially responsible” mutual funds. such mutual funds invest in a basket of sustainable companies, selected based on one of the myriad of established rubrics assessing and identifying these companies. alternatively, an investor can adopt a more active investing approach, based on their belief that there are semi-strong–form inefficiencies for csp-focused firms. investors using this approach attempt to exploit informational inefficiencies in the market for sustainability-focused firm equities by selecting “sustainable” stocks, based on a more in-depth assessment of a firm’s csp. this study will (1) review existing literature for the measurement of csp and cfp, (2) examine how csp data are assimilated into the marketplace, and (3) attempt to determine if one of these two approaches to sustainable investing (active vs. passive; not taking into account transaction costs) would have yielded better results on a risk-adjusted basis over the past decade in global equity markets. our approach utilizes a comparison of market returns to those of two sustainable firm portfolios using four, common risk-adjusted portfolio performance statistics. here, sustainable firms are separated into two groups: (1) a portfolio of all firms selected for inclusion in the dow jones sustainability index (djsi), and (2) a separate portfolio of only those djsi firms that were selected as “supersector leaders.” 200 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 the potentially unique role of firms that the djsi identifies as “leaders” among the group of sustainable firms has scarcely to be studied. while we do not find evidence of inefficiencies in the market for all djsi firms, we do find that there are likely semi-strong inefficiencies in the market for “leading” csp firms, as evidenced by their ability to earn abnormal returns over the period studied. these results imply that an active investing approach may be useful in this segment of the sustainable-firm equity market. 2. previous research current theoretical discussions of the relationship between a company’s financial performance and broader-based assessments of firm performance attempt to yield insights into whether non-financial performance measures add any new information for corporate stakeholders, especially investors. existing csp–cfp studies normally seek to determine if a firm’s choice of sustainable business practices enables it to achieve better financial results. because researchers have used a variety of traditional metrics for financial, and non-financial performance, results have been mixed (see orlitzky et al., 2003, for meta-analysis of 52 studies). some studies have found that companies that demonstrate stronger performance in non-financial areas, such as environmental, social, and governance (esg) also have comparatively more financial accomplishments, whereas others have failed to establish that link. where the data have supported this link, the financial valuation premise follows that firm value is increased because sustainability practices stabilize future cash flows and lower forward risk. the question that investors must consider is whether such information is already reflected in market prices. if not, investors should endeavor to identify those firms whose equity values do not yet fully reflect their non-financial efforts. in investment analysis, such firms would be considered “undervalued” and good candidates for addition to a value-focused portfolio of corporate stocks. 2.1. measuring corporate sustainability performance: existing studies the growing market for ratings of non-financial performance is partially driven by the investor’s need for more (and better) information about public companies’ environmental, social, and governance activities. as analysts seek insight into future financial performance, they are increasingly examining other aspects of firm output that are less-obvious, and perhaps, not yet priced by financial market participants. such non-financial measures face greater scrutiny as investors attempt to (1) identify which aspects of sustainability most highly correlate with inclusion in “sustainable funds” and (2) discover which of these aspects most frequently contribute to future financial success. for example, berry and junkus (2013) found that (1) investors generally consider environmental practices to be the most important aspect of a firm’s sustainability performance, and (2) that investors prefer to reward firms that display positive social behavior (i.e., “positive screening”) rather than excluding firms for particular products or practices (“negative screening”). flammer (2013) finds “that companies reported to have behaved responsibly toward the environment experience a significant stock increase.” 201t.m. shank, b. shockey / financial services review 25 (2016) 199–214 one of the shortcomings of such existing studies is that social and environmental performances are distinct, non-financial outputs. moreover, non-financial performance has not, historically, been measured and reported on a large scale to the investing public. therefore, comprehensive assessments of csp have been difficult. there have been a number of recent attempts, however, to develop broad-based measures of sustainability to better inform investors about a firm’s non-financial performance. these measures represent an improvement on existing measures that are based on a single sustainability component— social, environmental, or governance. the inclusion of all three focus areas seems to be, in part, motivated by the hierarchical structure of the 1997 global reporting initiative (see labuschagne et al., 2004) as many of the most popular comprehensive rating systems were developed in the early 2000s. while previous studies that focus on a single facet of non-financial performance are important to the evolution of thought on the topic, more recent studies of the csp–cfp relationship utilize a more inclusive and multifaceted approach to rating a firm’s social/ environmental/governance performance. some of the more notable rating systems which attempt to measure sustainability performance by considering all three of these corporate outputs include the domini social 400 stock index (see baird et al., 2012), the djsi, and the financial times sustainability index (ftse4good), as well as a large number of other comprehensive benchmarks dedicated to corporate stocks of a particular country or geographic region (www.world-exchanges.org/sustainability/m-7–0.php). as indicated by the types of firms that have developed these comprehensive assessments, the primary driving force behind these new measures was investor’s unmet need for better information upon which to select sustainable companies for inclusion in their portfolios. more recent studies of these sustainability indices attempt to assess the financial impact of a firm’s addition to or removal from a particular index. from an investor’s point of view, we assume the addition of a company to a list would result in a financial market benefit stemming from the recognition of the firm’s sustainability efforts, and enhanced future value. robinson et al. (2011) find that although there is a sustained increase in the financial value of a firm that is added to the djsi, there is an insignificant effect when a firm is removed. conversely, doh et al. (2010) assert that there is more meaning in being removed from a comprehensive sustainability index (i.e., the domini social 400), than being added; which results in no significant impact. with such mixed results, it seems that little informational value is gleaned by simply examining the financial performance firms that are being added to or removed from these sustainable indices. 2.2. measuring corporate financial performance the measurement of firm financial performance is well-established, although there are two distinct approaches to financial metrics used in prior csp–cfp studies. existing studies focus on accountingor market-based measures. the findings of such studies do often show that sustainability practices are more apparent in one financial metric versus the other type. for example, orlititzky’s (2003) meta-analysis that examined 52 prior studies dating back to 1975 found that accounting based measures of financial performance better reflected csr efforts compared to market-based measures. many studies, such as rodgers et al. (2013) find 202 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 that a firm’s csr commitment leads to better financial performance on both accountingbased and market-based financial metrics. overall, most recent studies, whether using accounting-based measures (see ameer et al., 2012; lopez et al., 2007) or market-based measures (see hill et al., 2007; shank et al., 2005) have found that financial value is created when firms are recognized as being relatively more sustainability-focused than their peers. the question that arises for sustainability-focused investors is how to best identify firms that are making a long-term strategic commitment to sustainability. investors that want to reward such commitment by a firm, and seek to achieve above-market performance by purchasing these potentially undervalued stocks would benefit significantly from greater insight into the risk-return tradeoffs of such securities. commitment to sustainability would seem a continuum, with some firms embracing sustainable business practices to a greater extent than others. several recent csp–cfp studies have found that a firm’s level of commitment does matter in whether the firm enjoys the financial rewards of a sustainable strategy. barnett and salomon (2012) find, after categorizing companies as “low,” “moderate,” or “high” csp firms, that those with high csp have the highest cfp, arguing that whether it pays to be good depends on how well a company is able to capitalize on its sustainability efforts. similarly, lourenco et al. (2012, 2014) find that financial success is greater for firms with a reputation for sustainability leadership when compared with firms not considered leaders in non-financial performance. 2.3. the valuation issue and market efficiency as previously noted, many existing studies have found evidence for the financial market relevance of non-financial performance information. this information is utilized in the formation of many of the existing stock indices that positively screen companies with dedicated sustainability efforts (i.e., the djsi, domini social 400, and ftse4good). the purpose of the development of these indices was to assist investors in better identifying firms with exceptional sustainability records and a clear commitment to the continuation of such efforts. finance theory, which posits that the goal of a manager is to maximize shareholder wealth, would support activities by the firm to increase future cash flows or to reduce forward risk. support of the link between more sustainable firms and lower corporate risk can be found in lee (2009) and ghoul et al. (2011). moreover, investors who believe that a firm’s long-term commitment to sustainability impacts the value of a stock, would value new sources of non-financial information for assimilation into their stock screening methodologies. currently, the lack of consensus on the impact of csp and limited availability of non-financial market factors represents a market inefficiency juxtaposed to the established and highly efficient nature of existing financial market factors. this information asymmetry is narrowed as measures for non-financial factors become more comprehensive and more widely used in indices such as the djsi and djgi. therefore, the question increasingly becomes whether sustainability-focused investors can depend on these indices to identify and fully differentiate the most sustainable firms from less sustainable firms and the broader equity market. 203t.m. shank, b. shockey / financial services review 25 (2016) 199–214 2.4. passive versus active investment strategies according to north and stevens (2015), a passive investment approach is supported by efficient market arguments that active trading of stocks will not consistently beat the market on a risk-adjusted basis. our study looks for evidence that there may be semi-strong form inefficiencies in the market for sustainable firm equities. for individual investors, especially those that value firms adopting sustainable business practices, the opportunity to outperform passive selections of sustainable firm portfolios may exist. in our study, we found that actively selecting a portfolio of only firms determined to be industry leaders in sustainable business practices over the past 11 years earned positive alphas compared to both the sustainability index fund, and the general market. according to the efficient markets hypothesis (emh), those firms identified as “supersector leaders” by the djsi should not outperform a portfolio of sustainable firms as a whole (all firms comprising the djsi) if this information is quickly priced in the market for sustainable firm stocks. like north and stevens, we believe our approach offers a rigorous test of semi-strong market efficiency for a particular subsector of equities, and the performance of two, specially selected portfolios. passive investment strategies are more appropriate for markets where there is high informational efficiency; meaning that sustainability-focused investors may be best served by choosing to merely buy shares in a diversified portfolio of stocks of sustainable firms. such firms are normally included in broad-based portfolios that have been constructed by selecting companies implicitly labeled as sustainability-focused by being included in a sustainability index. theoretically, the newer measures, such as the djsi, which utilizes a number of non-financial performance data simultaneously, would increase the informational efficiency in the market for sustainable firm equities. several recent studies have compared the financial performance of firms selected for inclusion in one of the major sustainability indices to benchmarks of the broader market. of those, consolandi et al. (2009) found that the djsi (focused on sustainable european firms) slightly underperformed market benchmarks from 2001 to 2006. xiao et al. (2013) investigated the role of corporate sustainability (using the djsi) investment in asset pricing from 2001 to 2007 and found no significant impact on equity returns. such results would suggest there is no additional informational value to investors of a firm being included in the comprehensive benchmarks of dow jones, inc. that identify sustainable company stocks. however, the potentially unique role of firms that the djsi identifies as the leaders among the group of sustainable firms has scarcely to be studied. conventional investing wisdom suggests stock investors adopt an active portfolio management strategy in markets that are not relatively efficient, whereby market participants do not possess all relevant information when selecting securities for inclusion in their portfolios. as investors seek to identify sustainable companies, the question is whether the information currently used by such investors is adequate. although studies have suggested that little relevant information is provided when a firm is selected as a member of the djsi, it is possible that a firm’s identification as a top-performer in the non-financial arena does have investment value. similar to the way in which lin (2014) sought to determine if an individual investor might benefit from active management in equity sector allocations (compared to passively managed 204 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 sector index funds), the goal of our study is to determine if (and how) individual investors might benefit from active management in the market for sustainable firm equities. like lin, we also find support for active management based on superior risk adjusted returns for one of our two portfolios (as measured by the sharpe ratio) compared to the passive portfolios tested. in summary, our study seeks to examine the csp–cfp relationship by comparing the risk adjusted stock market returns over the past decade of two distinct, sustainability-focused portfolios constructed using djsi selected companies. the first portfolio is comprised of all firms that have been added to the index over the past 10 years, and the second is comprised of those firms identified as supersector leaders of the djsi. in this way, we will determine whether an investor would have performed better than the overall equity market by passively “buying the index” versus actively buying stocks of only those firms identified by djsi as sustainability leaders in their respective industries. since prior studies have found that the level of a firm’s sustainability commitment is relevant, our study will determine whether those companies specifically identified as sustainability leaders do perform better, using a market-based, risk-adjusted approach. in addition, since barnett (2007) found that financial impacts of csr vary over time, we will divide the 2002–2012 period into three subperiods to discover any temporal differences. 3. data and methodology two methodologies were adopted and utilized for assessing returns for supersector leaders against the broader market (for this study, the s&p 500 is used as a proxy for the performance of the broader market). the first methodology (“annual supersector leader portfolio”) consisted of creating a balanced (equally weighted) portfolio out of the 18 to 19 companies named in each “class” from 2002 through 2012 and regressing their mean (arithmetic) monthly returns from the time of each classes’ inception through march 2013 against the monthly returns of two indexes, the s&p 500 and the dow jones sustainability index world diversified (w1sgi). the second methodology (“cumulative portfolio”) consisted of creating a balanced cumulative portfolio of supersector leaders. beginning in 2002, a portfolio was created out of the 18 original supersector leaders. each time a new company was named as a supersector leader in a subsequent annual robecosam corporate sustainability assessment, it would be added to the portfolio from that date. meanwhile, any company previously named a supersector leader would remain within the portfolio regardless of whether or not it was again named as a supersector leader in any subsequent year. hence, the cumulative portfolio annually increased in size as demonstrated in figure 1. 3.1. annual supersector leader portfolio the first step in analyzing the returns of the supersector leaders, as ranked by the djsi (based on the robecosam corporate sustainability index), was to obtain the monthly returns of each individual company named as a supersector leader, by year (i.e., the monthly 205t.m. shank, b. shockey / financial services review 25 (2016) 199–214 returns of the 2002 supersector leader class were gathered from september 4, 2002 through march 1, 2013 and averaged). second, the mean (arithmetic) monthly returns of each supersector leader class, by year, were compared to the monthly returns of the s&p 500 in terms of jensen’s alpha, m-square, the treynor measure, and the sharpe ratio. in essence, a portfolio was created from each annual class of supersector leaders from 2002 through 2012 and compared with the monthly returns of the broad s&p 500. the results of this comparison were then regressed against the returns of the s&p 500 to test the significance of the results at a 95% confidence level. third, as a means of comparison, the monthly returns of the djsi world were compared to the s&p 500 by the same measures and time periods and regressed to test significance at a 95% confidence level. finally, the jensen’s alpha, m-square, the treynor measure, and the sharpe ratio of the djsi world (w1sgi) and each annual supersector leader class were compared relative to their respective returns against the s&p 500, thereby creating a means to determine whether an individual investor would have achieved greater returns from the djsi world, the s&p 500, or each annual supersector leader class, all else equal, from the inception date of each annual class through march 1, 2013. 3.2. measures of performance1 jensen’s alpha: �p � rp � �rf � �p�rm � rf�� (1) m-square: m2 � (sp � sm) � �m (2) treynor’s measure: tp � rp � rf �p (3) sharpe’s measure: sp � rp � rf �p (4) fig. 1. graphic demonstration of cumulative supersector leader portfolio size (2002 to 2012). 206 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 3.3. cumulative supersector leader portfolio analysis the second means by which the returns of the djsi supersector leaders were analyzed was by creating a cumulative list of supersector leaders and obtaining the monthly returns of each company named since 2002 from the date in which the company was first identified as a supersector leader until march 1, 2013. in this way, a cumulative portfolio of companies was created in which all supersector leaders are held in a portfolio beginning on each company’s respective date of initial pronouncement. the average monthly returns of all companies held in this cumulative portfolio during each given month were then compared with the monthly returns of the s&p 500 (broad market proxy) in terms of jensen’s alpha, m-square, the treynor measure, and the sharpe ratio for the four time periods (1-year, 3-year, 5-year, and the total time period from september 4, 2002 through march 1, 2013). the results of this comparison were then regressed against the returns of the s&p 500 to test the significance of the results at a 95% confidence level. the cumulative holdings of supersector leaders are demonstrated in table 1. second, as a means of comparison, the monthly returns of the djsi world were compared to the s&p 500 by the same measures and time periods and regressed to test significance at a 95% confidence level. finally, the jensen’s alpha, m-square, the treynor measure, and the sharpe ratio of the djsi world (w1sgi) and the cumulative supersector leader portfolio were compared relative to their respective returns against the s&p 500, thereby creating a means to determine whether an individual investor would have achieved greater returns from the djsi world, the s&p 500, or the cumulative supersector leader portfolio from september 1, 2002 through march 1, 2013, all else equal. for investors, this methodology would reflect a belief that once a company is named as a supersector leader, it remains a leader of corporate sustainability, which is supported by the fact that 44 companies were named as a supersector leader more than once from 2002 through 2012 and that 76 companies account for the 203 supersector leaders named from 2002 through 2012. additionally, this methodology allows for reduced transaction costs table 1 tabular demonstration of cumulative supersector leader portfolio size (2002 to 2012) year number of companies added total number of companies in portfolio 2002 18 18 2003 5 23 2004 5 28 2005 6 34 2006 5 39 2007 5 44 2008 8 52 2009 7 59 2010 4 63 2011 8 71 2012 5 76 table 1 demonstrates the annual growth of the cumulative supersector leader portfolio from 2002 to 2012 and specifically shows the number of companies added each year. 207t.m. shank, b. shockey / financial services review 25 (2016) 199–214 compared with the annual supersector leader portfolio methodology as the only transactions occurring would be the addition of the approximately 4 to 8 companies named as supersector leaders that are not already included in the cumulative portfolio (this does not account for portfolio rebalancing). 4. results the results of this study were statistically significant for the annual supersector classes from 2002 through 2011. during these years, average annualized monthly returns ranged from 1.73% (2007 class) to 13.75% (2011 class) and outperformed the s&p 500 in gross return from 2002 through 2009 and in 2011, with the s&p 500 outperforming the supersector class in 2010. during this time span, the djsi world index was consistently more highly correlated to the returns of the s&p 500, while the supersector classes had correlation (r2) ranging from 0.22 for the 2002 class to a high of 0.86 for the 2006 class, with a � for each class ranging from 0.87 for the 2004 class to 1.16 for the 2006 class. for the cumulative portfolio, from september 2002 through march 2013, the average annual return was 3.68%, comparing favorably to the average annual return of the s&p 500 of 2.11% during this timeframe. this also compares favorably to the return of the djsi world index of 2.07%. once again, the djsi world index was more highly correlated with an r2 of 0.89 against the s&p 500, compared with an r2 of 0.82 for the cumulative supersector leader portfolio. the � measure for the cumulative portfolio during this timeframe was 1.05 against the s&p 500, while the djsi world index produced a � measure of 1.12. as discussed in the methodology, the findings below will be presented using four risk-adjusted measures of return. 4.1. annual supersector leader portfolio the results of the annual supersector leader portfolio analysis are presented in tables 2, 3, 4, and 5. the jensen’s alpha and m-square performance measures (represented in table 2) provide risk-adjusted comparative statistics whereby the performance each annual supersector class is compared with the baseline market proxy (the s&p 500). it then follows that a positive jensen’s alpha or m-square indicates that the supersector portfolio outperformed the s&p500 and vice versa. therefore, it can be interpreted that with 95% significance that the annual supersector leader portfolio outperformed the s&p500 each year from 2002 through 2008 and underperformed from 2009 to 2011. the 2012 results did not provide a statistically significant indicator of performance. second, this portfolio was tested against the s&p500 using treynor’s measure and sharpe’s measure, the results of which are shown in table 3. these risk-adjusted measures for the supersector portfolio can be directly compared to the s&p500 baseline, as shown in table 3, with the higher return representing the better statistical performance. therefore, it can again be interpreted that with 95% significance that the annual supersector leader 208 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 portfolio outperformed the s&p500 each year from 2002 through 2008 and underperformed from 2009 to 2011. the 2012 results once again did not provide a statistically significant indicator of performance. third, tables 4 and 5 also compared the performance of the annual supersector leader portfolio to the performance of the w1sgi, again using the s&p500 as the baseline market proxy. as demonstrated in tables 4 and 5, it can be concluded that the annual supersector portfolio outperformed the djsi in all years on a risk-adjusted basis with 95% statistical significance (with the exception of the 2012 results, which were not statistically significant because of sample size). table 2 annual supersector leader portfolio performance vs. market proxy (2002 to 2012) year jensen’s m-square significant at 95%? p value 2012 0.1476 0.0559 no 0.87 2011 �0.0010 �0.0275 yes 2.3 � 10�4 2010 �0.0219 �0.0194 yes 5.5 � 10�9 2009 �0.0077 �0.0088 yes 1.3 �10�14 2008 0.0141 0.0111 yes 4.4 �10�22 2007 0.0130 0.0118 yes 5.1 �10�27 2006 0.0167 0.0140 yes 1.1 �10�34 2005 0.0185 0.0184 yes 5.1 �10�37 2004 0.0260 0.0245 yes 2.7 �10�22 2003 0.0201 0.0174 yes 6.8 �10�27 2002 0.0258 0.0210 yes 2.7 � 10�8 table 2 demonstrates the annual performance of each supersector leader class portfolio against the baseline market proxy (s&p 500) using jensen’s alpha and m-square as performance measures. statistical significance of the results is shown in the columns to the far right column, including p values, which apply to jensen’s alpha, m-square, treynor’s measure (table 3), and sharpe’s measure (table 3). table 3 annual supersector leader portfolio performance vs. market proxy (2002 to 2012) year treynor (leaders) treynor (s&p 500) sharpe (leaders) sharpe (s&p 500) 2012 1.4763 0.0755 0.7737 0.4444 2011 0.1121 0.1131 0.3410 0.4503 2010 �0.0012 0.0184 �0.0077 0.1421 2009 0.0072 0.0151 0.0430 0.1028 2008 0.0163 0.0037 0.0790 0.0198 2007 �0.0024 �0.0140 �0.0118 �0.0761 2006 0.0055 �0.0088 0.0298 �0.0512 2005 0.0124 �0.0070 0.0701 �0.0433 2004 0.0248 �0.0051 0.1253 �0.0331 2003 0.0177 �0.0031 0.0955 �0.0211 2002 0.0288 0.0011 0.1478 0.0073 table 3 demonstrates the annual performance of each supersector leader class portfolio against the baseline market proxy (s&p 500) using treynor’s measure and sharpe’s measure to gauge performance. for the benefit of the reader, the greater return between the supersector portfolio and market proxy has been bolded. 209t.m. shank, b. shockey / financial services review 25 (2016) 199–214 4.2. cumulative supersector leader portfolio analysis the results of the cumulative supersector leader portfolio analysis are presented in tables 6, 7, 8, and 9. for the cumulative supersector portfolio analysis, the jensen’s alpha and m-square performance measures were once again utilized to provide risk-adjusted comparative statistics whereby the performance of each annual supersector class is compared with the baseline market proxy (the s&p 500). it then follows that a positive jensen’s alpha or m-square indicates that the supersector portfolio outperformed the s&p500 and vice versa. therefore, it can be intertable 4 annual supersector leader portfolio performance vs. w1sgi (2002 to 2012) year jensen’s (leaders) jensen’s (djsi) m-square (leaders) m-square (djsi) leaders p value djsi p value 2012 0.1476 0.0456 0.0559 0.0199 0.87 0.21 2011 �0.0010 �0.0637 �0.0275 �0.0567 2.3 � 10�4 1.1 � 10�7 2010 �0.0219 �0.0260 �0.0194 �0.0206 5.5 � 10�9 2.8 �10�14 2009 �0.0077 �0.0222 �0.0088 �0.0190 1.3 �10�14 3.4 �10�19 2008 0.0141 �0.0106 0.0111 �0.0090 4.4 �10�22 4.9 �10�29 2007 0.0130 �0.0101 0.0118 �0.0078 5.1 �10�27 7.7 �10�34 2006 0.0167 �0.0063 0.0140 �0.0048 1.1 �10�34 6.2 �10�39 2005 0.0185 �0.0035 0.0184 �0.0025 5.1 �10�37 6.8 �10�44 2004 0.0260 �0.0013 0.0245 �0.0008 2.7 �10�22 4.1 �10�49 2003 0.0201 �0.0005 0.0174 �0.0002 6.8 �10�27 2.5 �10�55 2002 0.0258 �0.0005 0.0210 �0.0005 2.7 � 10�8 3.8 �10�62 table 4 demonstrates the annual performance of each supersector leader class portfolio against the dow jones sustainability index (w1sgi), while utilizing the s&p 500 as the baseline market proxy. performance measures in table 5 are jensen’s alpha and m-square. statistical significance of the results is shown in the columns to the far right column, including p values that apply to jensen’s alpha, m-square, treynor’s measure (table 5), and sharpe’s measure (table 5). for the benefit of the reader, the greater return between the supersector portfolio and w1sgi has been bolded. table 5 annual supersector leader portfolio performance vs. w1sgi (2002 to 2012) year treynor (leaders) treynor (djsi) sharpe (leaders) sharpe (djsi) 2012 1.4763 0.1590 0.7737 0.5613 2011 0.1121 0.0616 0.3410 0.2245 2010 �0.0012 �0.0024 �0.0077 �0.0176 2009 0.0072 �0.0042 0.0430 �0.0265 2008 0.0163 �0.0055 0.0790 �0.0281 2007 �0.0024 �0.0230 �0.0118 �0.1183 2006 0.0055 �0.0144 0.0298 �0.0789 2005 0.0124 �0.0101 0.0701 �0.0587 2004 0.0248 �0.0063 0.1253 �0.0383 2003 0.0177 �0.0036 0.0955 �0.0228 2002 0.0288 0.0006 0.1478 0.0041 table 5 demonstrates the annual performance of each supersector leader class portfolio against the dow jones sustainability index (w1sgi), while utilizing the s&p 500 as the baseline market proxy. performance measures in table 5 are treynor’s measure and sharpe’s measure. for the benefit of the reader, the greater return between the supersector portfolio and w1sgi has been bolded. 210 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 preted that with 95% significance the cumulative supersector leader portfolio outperformed the s&p500 in the 3-year, 5-year, and 11-year maximum investment periods, and underperformed the s&p-500 in the 1-year investment period from 2002 to 2003. second, this portfolio was also tested against the s&p500 using treynor’s measure and sharpe’s measure, the results of which are shown in table 7. these risk-adjusted measures for the supersector portfolio can be directly compared with the s&p500 baseline, as shown table 6 cumulative supersector leader portfolio performance vs. market proxy (2002 to 2012) investment period m-square (leaders) jensens (leaders) significant at 95%? p value september 2002 to march 2013 (total) 0.0135 0.0156 yes 5.1�10�48 september 2002 to september 2003 (1-year) �0.0086 �0.0078 yes 8.2� 10�8 september 2002 to september 2005 (3-year) 0.0125 0.0149 yes 2.0�10�14 september 2002 to september 2007 (5-year) 0.0184 0.0253 yes 1.2�10�14 table 6 demonstrates the performance of the cumulative supersector leader class portfolio against the baseline market proxy (s&p 500) using jensen’s alpha and m-square as performance measures. the investment periods for this analysis are detailed in the far left column. statistical significance of the results is shown in the columns to the far right column, including p values which apply to jensen’s alpha, m-square, treynor’s measure (table 7), and sharpe’s measure (table 7). table 7 cumulative supersector leader portfolio performance vs. market proxy (2002 to 2012) investment period treynor (leaders) treynor (s&p 500) sharpe (leaders) sharpe (s&p 500) september 2002 to march 2013 (total) 0.0160 0.0011 0.1243 0.0073 september 2002 to september 2003 (1-year) 0.0336 0.0414 0.2128 0.2686 september 2002 to september 2005 (3-year) 0.0368 0.0214 0.3064 0.1936 september 2002 to september 2007 (5-year) 0.0445 0.0177 0.3730 0.1831 table 7 demonstrates the performance of the cumulative supersector leader class portfolio against the baseline market proxy (s&p 500) using treynor’s measure and sharpe’s measure the gauge performance. the investment periods for this analysis are detailed in the far left column. statistical significance of the results is shown in the far right column. for the benefit of the reader, the greater return between the supersector portfolio and market proxy has been bolded. table 8 cumulative supersector leader portfolio performance vs. w1sgi (2002 to 2012) investment period m-square (leaders) m-square (djsi) jensens (leaders) jensens (djsi) leaders p value djsi p value september 2002 to march 2013 (total) 0.0135 �0.0005 0.0156 �0.0005 5.1�10�48 3.8�10�62 september 2002 to september 2003 (1-year) �0.0086 0.0018 �0.0078 0.0033 8.2� 10�8 1.6� 10�7 september 2002 to september 2005 (3-year) 0.0125 0.0064 0.0149 0.0081 2.0�10�14 7.1�10�20 september 2002 to september 2007 (5-year) 0.0184 0.0094 0.0253 0.0117 1.2�10�14 2.7�10�27 table 8 demonstrates the performance of the cumulative supersector leader class portfolio against the dow jones sustainability index (w1sgi), while utilizing the s&p 500 as the baseline market proxy. performance measures in table 8 are jensen’s alpha and m-square. statistical significance of the results is shown in the columns to the far right column, including p values that apply to jensen’s alpha, m-square, treynor’s measure (table 9), and sharpe’s measure (table 9). for the benefit of the reader, the greater return between the supersector portfolio and w1sgi has been bolded. 211t.m. shank, b. shockey / financial services review 25 (2016) 199–214 in table 7, with the higher return representing the better statistical performance. therefore, these measures confirm the results of the jensen’s alpha and m-square measures in finding that with 95% significance the cumulative supersector leader portfolio outperformed the s&p500 in the 3-year, 5-year, and 11-year maximum investment periods, and underperformed the s&p-500 in the 1-year investment period from 2002 to 2003. finally, tables 8 and 9 also compared the performance of the cumulative supersector leader portfolio to the performance of the w1sgi, again using the s&p500 as the baseline market proxy. as is demonstrated in tables 8 and 9, it can be concluded that the cumulative supersector portfolio outperformed the dow jones sustainability index in the 3-year, 5-year, and 11-year maximum investment periods, and underperformed the djsi in the 1-year investment period from 2002 to 2003. 5. interpretation of results these results, which show statistically significant superior returns to the market (on a risk adjusted basis) over longer term periods, are consistent with the supposition that stocks of sustainable firms would show most benefits over a long-term time horizon. the results of this study point to a potential inefficiency as the market may not fully understand how to incorporate the effects of sustainability practices into securities pricing. these results also indicate that an investor could have achieved above-average returns by investing in sustainability leaders over the last 10 years based on this inefficiency. such information could provide an incentive and positive market pressure to be a leader in sustainability, which would provide newfound energy and headwinds in the sustainability movement, which has previously been driven through disincentives and regulation. such positive market pressure is also a more economically efficient means of effecting change in the market and in the allocation of supply and demand. for investors with an inclination to invest in or reward sustainable businesses, this study provides a comparison of active versus passive strategies as well as a potential blueprint for achieving competitive or even above-average returns. the data in this study demonstrate that investments in sustainability leaders over the last 10 years has been profitable compared with the broad market, but has vastly outperformed sustainability indexes over the same time period, which utilize a lesser degree of selectivity than the sample portfolios of supersector leaders. table 9 cumulative supersector leader portfolio performance vs. w1sgi (2002 to 2012) investment period treynor (leaders) treynor (djsi) sharpe (leaders) sharpe (djsi) september 2002 to march 2013 (total) 0.0160 0.0006 0.1243 0.0073 september 2002 to september 2003 (1-year) 0.0336 0.0445 0.2128 0.2806 september 2002 to september 2005 (3-year) 0.0368 0.0290 0.3064 0.2513 september 2002 to september 2007 (5-year) 0.0445 0.0177 0.3730 0.2798 table 9 demonstrates the performance of the cumulative supersector leader class portfolio against the dow jones sustainability index (w1sgi), while utilizing the s&p 500 as the baseline market proxy. performance measures in table 9 are treynor’s measure and sharpe’s measure. for the benefit of the reader, the greater return between the supersector portfolio and w1sgi has been bolded. 212 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 6. summary and conclusions our results shed additional light on the ongoing debate over whether an investor can “do well by doing good.” moreover, the results provide information regarding not only “if” sustainable investing is rewarded, but also “how” this financial payoff would have been earned in the past decade when selecting sustainable firms stocks. we find that an investor can achieve superior financial returns by actively selecting sustainable firm stocks, and that the method of firm selection does have impact, judging by the performance of sustainable firm equity investments over the 2002–2012 period. by measuring financial performance based on the risk-adjusted stock returns of two distinct groups of sustainable firms, our analysis reveals that djsis supersector leaders outperformed the overall market on this basis. alternatively, the w1sgi underperformed the broader market over the period studied. the nuances of the approach taken may explain why these results yield new insights into how investors might increase the likelihood that a sustainable investing strategy would reap financial rewards. first, we measure financial payoff using four, comprehensive return metrics that all measure a portfolio’s success by simultaneously considering stock returns and the level of risk taken to earn those returns. prior work has hypothesized that firms embracing sustainable business practices are comparatively more successful financially, either because they earned higher returns, or were seen by markets as having lower risk. therefore, by including all of these measures, we were able to check the results for consistency since each metric considers a different aspect of overall corporate risk. all four yielded consistent results: a portfolio comprised of companies identified as sustainability leaders would have earned an investor superior returns during the period studied. a second nuance of our approach was the separation of the 10-year performance of the sustainability portfolios into subperiods/time “buckets” to help determine any relevance of shorterversus longer-term investment horizons for sustainable investors. these results, which show statistically significant superior returns to the market (on a risk adjusted basis) over longer term periods, are consistent with the supposition that stocks of sustainable firms would show most benefits over a long-term time horizon, and may not earn abnormally high returns for short-term focused investors. in other words, embracing sustainable business practices is a long-term, value-maximizing strategy. notes 1 see bodie, kane, and marcus (2014), pp. 837–847. references ameer, r., & othman, r. (2012). sustainability practices and corporate financial performance: a study based on the top global corporations. journal of business ethics, 108, 61–79. baird, p., geylani, p., & roberts, j. (2012). corporate social and financial performance re-examined: industry effects in a linear mixed model analysis. journal of business ethics, 109, 367–388. 213t.m. shank, b. shockey / financial services review 25 (2016) 199–214 barnett, m. (2007). stakeholder influence capacity and the variability of financial returns to corporate social responsibility. academy of management review, 32, 794–816. barnett, m., & salomon, r. (2013). does it pay to be really good? addressing the shape of the relationship between social and environmental performance. strategic management journal, 33, 1304–1320. berry, t., & junkus, j. (2013). socially responsible investing: an investor perspective. journal of business ethics, 112, 707–720. bodie, z., kane, a., & marcus, a. (2014). investments (10th ed.) new york, ny: mcgraw-hill education. cloninger, d., aggarwal, r., karpoff, j., lacey, n., & smith, c. (2000). doing well by doing good: do financial markets ‘price’ ethical behavior? financial practice and education, fall/winter, 24–33. consolandi, c., jaiswal-dale, a., poggiani, e., & vercelli, a. (2009). global standards and ethical stock indexes: the case of the dow jones sustainability index. journal of business ethics, 87, 185–197. doh, j. (2010). does the market respond to an endorsement of social responsibility? the role of institution, information, and legitimacy? journal of management, 36, 1461–1485. flammer, c. (2013). corporate social responsibility and shareholder reaction: the environmental awareness of investors. academy of management journal, 56, 758–781. ghoula, s. (2011). does corporate social responsibility affect the cost of capital? journal of banking and finance, 35, 2388–2408. griffin, j., & mahon, j. (1997). the corporate social performance and corporate financial performance debate: twenty-five years of incomparable research. business and society, 36, 5–31. hamilton, s., jo, h., & statman, m. (1993). doing well while doing good? the investment performance of socially responsible mutual funds. financial analysts journal, 49, 62–66. hill, r., ainscough, t., shank, t., & manullang, d. (2007). corporate social responsibility and socially responsible investing: a global perspective. journal of business ethics, 70, 165–174. labuschagne, c., brent, a., & van erck, r. (2004). assessing the sustainability performance of industries. journal of cleaner production, 13, 1–13. lee, d. (2009). corporate sustainability performance and idiosyncratic risk: a global perspective. financial review, 44, 213–237. lin, c. (2014). does active management work? evidence from equity sector funds. financial services review, 23, 249–271. lopez, m., garcia, a., & rodriquez, l. (2007). sustainable development and corporate performance: a study based on the dow jones sustainability index. journal of business ethics, 75, 285–300. lourenco, i., branco, m., curto, j., & eugenio, t. (2012). how does the market value corporate sustainability performance? journal of business ethics, 108, 417–428. lourenco, i., callen, j., branco, m., & curto, j. (2014). the value relevance of reputation for sustainability leadership. journal of business ethics, 119, 17–28. north, d., & stevens, j. (2015). investment performance of aaii stock screens over diverse markets. financial services review, 24, 157–176. orlitzky, m. (2003). corporate social and financial performance: a meta-analysis. organization studies, 24, 403–441. robinson, m., kleffner, a., & bertels, s. (2011). signaling sustainability leadership: empirical evidence of the value of djsi membership. journal of business ethics, 101, 493–505. rodgers, w., choy, h., & guiral, a. (2013). do investors value a firm’s commitment to social activities? journal of business ethics, 114, 607–623. shank, t., manullang, d., & hill, r. (2005). “doing well while doing good” revisited: a study of socially responsible firms’ short-term versus long-term performance. managerial finance, 30, 33–45. world federation of exchanges: exchanges and sustainable investment. (available at www.world-exchanges.org/ sustainability/m-7-0.php). xiao, y., faff, r., gharghori, p., & lee, d. (2013). an empirical study of the world price of sustainability. journal of business ethics, 114, 297–310. 214 t.m. shank, b. shockey / financial services review 25 (2016) 199–214 untitled ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires december 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. a. b. c. d. 2. a. policyholder b. beneficiaries c. insurance company d. all of the above 3. a. unexpected needs of dependents b. medical expense shocks c. death of the insured d. unemployment 5. do large 401(k) plans generally have lower fees than small 401(k) plans? a. they generally have the same level of fees. b. small 401(k) plans generally have lower fees than large 401(k) plans. c. large 401(k) plans generally have lower fees than small 401(k) plans. d. the level of fees varies for both sizes of plans so that you cannot make a generalization. 4. are workers with low financial literacy generally better off in iras or 401(k) plans? a. b. d. persons with low financial literacy are generally equally well off in either an ira or a 401(k) plan. in west, de zwaan, and johnson, the authors examine the financial literacy performance of adults at an australian university and find______. there are no gender differences in financial literacy performance. women perform better in tests of financial literacy performance than men. more women select the non-response option than men. financial literacy performance is related to confidence with money. in mulholland and finke, which stakeholder is economically affected by life insurance policy lapsation ? in mulholland and finke, all of the following are non-mortality background risks that lead to a need for liquidity, except: it depends on the plan, so one cannot generalize that either is better than the other. persons with low financial literacy are generally better off in iras. c. persons with low financial literacy are generally better off in 401(k) plans. finser_23_3 using the new portability election of deceased spouses: a pedagogical example de’arno de’armond, ph.d.a,*, darlene pulliam, ph.d.b, robin patterson, j.d.c aedwards professor of financial planning, west texas a&m university, box 60809, canyon, tx 79016-001, usa bregents professor of accounting, mccray professor of business, west texas a&m university, box 60809, canyon, tx 79016-001, usa cpatterson professor of business law, west texas a&m university, box 60809, canyon, tx 79106-001, usa abstract the united states has a unified system that taxes transfers of property during an individual’s lifetime (gifts) and property transferred as a result of the individual’s death. the tax relief, unemployment insurance reauthorization, and job creation act of 2010 (the act) contains a provision that will allow the unused portion of a decedent’s exclusion (taxable estate protected by the unified credit) to be used upon the subsequent death of the surviving spouse. the portability election is simple for situations where it appears the surviving spouse will not remarry, however, becomes much more complicated if the surviving spouse should remarry. © 2014 academy of financial services. all rights reserved. jel classification: m410, accounting; k11, estate planning keywords: tax accounting; estate planning; financial planning; lifetime gifts; portability election 1. introduction on the minds of many financial planners, advisors, practitioners, estate planners, tax accountants, academics, and general individuals as well, resides pending implications and effects of the “fiscal cliff” facing americans (dudley, 2013). finding fiscal balance between political parties may be challenging for the united states at best (grunwald, 2012). whereas * corresponding author. tel.: !1-806-651-2490; fax: !1-806-651-2488. e-mail address: ddearmond@wtamu.edu (d. de’armond) financial services review 23 (2014) 239–248 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. many facets of the “cliff” are known and commonly explored, there exist many smaller facets such as a “slew of temporary tax policies” that have, for the most part, been underexposed with exception to discussions often happening within the confines of the academic classroom and or certain individual prevailing circumstances (suderman, 2013). as political agreements have surfaced during this time, estate planning for married couples has been impacted, as evidenced with the passing of the tax relief act of 2010 (deener, 2012). the passing of the “the act” brought with it one such facet, dubbed dsuea, the “deceased spousal unused exclusion amount.” dsuea provides a surviving spouse an option to utilize the unused applicable exclusion of his or her deceased spouse for decedents dying after 2010 (dunn, 2011). this transfer or “portability” election requires special care and consideration in its application and should not be seen as a “cure-all” as many potential complex issues may arise (gallo, 2011). careful attention should be given the caveats of the portability election in an effort to utilize portability in the most efficient and effective manner (katz, 2011). as one might expect, the internal revenue service issued bulletins appropriate to the filings of dsuea via form 706 as a vehicle (internal revenue bulletin, 2011). although information has surfaced cautioning the use of the portability election, it remains a viable option given certain estate conditions (joseph, 2012). most recently, with the passing of the american taxpayer relief act of 2012 (atra), congress made permanent the two significant provisions discussed within this article, the $5,000,000 indexed basic exclusion amount and dsuea, as well as introduced a maximum transfer tax rate of 40% (miller, 2013). as other authors and academics alike have indicated in their writings, the public tends to not possess the financial knowledge to make good financial decisions all of the time (beierlein and neverett, 2013). it is in the spirit of encouraging basic and applied research in the area of personal financial planning that this work is presented. this article serves to enlighten and inform readers on one such “portability of deceased spouse unused exclusion tax provision” providing a pedagogical example utilized within the university setting in an effort to disseminate information and assist the development of curricula in financial services. in much the same “learning by doing” spirit that thomas eyssell wrote about the financial planning process, this particular work contributes to the literature (eyssell, 1999). this particular work will serve as a quick review of the u.s. estate and gift law, with brief discussions of the gift tax, annual exclusion, marital deduction, generation skipping tax, tax planning credits, and the new portability elections. the work concludes with examples and an excel application model developed for use in the academic and/or learning environment, and summary of the possible implications. 2. a quick review of u.s. estate and gift law the united states has a unified system that taxes transfers of property during an individual’s lifetime (gifts) and property transferred as a result of the individual’s death. the system is structured so that the sum of an individual’s taxable gifts and net assets transferred at death must exceed an applicable exemption amount before any gift or estate taxes are imposed. the 240 d. de’armond et al. / financial services review 23 (2014) 239–248 applicable exemption amount is $5,000,000 for 2011, $5,120,000 in 2012, $5,250,000 in 2013, and $5,340,000 in 2014. table 1, unified transfer tax rate schedule, after 2012 provides information for the sum of taxable estate and adjusted taxable gifts and their applicable tentative tax. as an example, utilizing table 1 below, the tentative tax on $5,340,000 using unified transfer tax rate schedule is found as $2,081,800 less applicable unified credit for 2014 of $2,081,800, yielding a net transfer tax payable of $0. note that in 2014 $5,340,000 of net estate is exempted from tax, but the marginal tax rate on the first dollar over $5,340,000 is 40%. 2.1. the gift tax the gift tax is imposed on a calendar year basis on transfers of property by gift by any individual. internal revenue code section (sec.) 2501. it is cumulative during an individual’s life. the gift tax is the excess of a tentative tax, computed on the aggregate sum of the taxable gifts for the calendar year and the preceding calendar years, over a tentative tax computed on the aggregate sum of the taxable gifts for the preceding calendar years. sec. 2502(a). the tax is computed with the same tax rates used for estate purposes in sec. 2001(c). 2.2. the annual exclusion in defining the term “taxable gifts,” sec. 2503 allows for an annual exclusion of $14,000 (in 2014) for gifts to any number of persons each year. since 1998, the amount of the annual exclusion has been indexed annually for inflation in increments of $1,000. gifts of a future interest do not generally qualify for the exclusion. however, transfers for the benefit of individuals under the age of 21 will not be considered transfers of a future interest if the property and the income therefrom may be expended by, or for the benefit of the donee before attainment of the age of 21 and any amounts not so expended will pass to the donee upon that donee’s 21st birthday. sec. 2503(c). table 1 unified transfer tax rate schedule after 2012 sum of taxable estate and adjusted taxable gifts tentative tax over but not under $0 $ 10,000 18% of such amount $10,000 $ 20,000 $1,800 plus 20% of excess over $10,000 $20,000 $ 40,000 $3,800 plus 22% of excess over $20,000 $40,000 $ 60,000 $8,200 plus 24% of excess over $40,000 $60,000 $ 80,000 $13,000 plus 26% of excess over $60,000 $80,000 $ 100,000 $18,200 plus 28% of excess over $80,000 $100,000 $ 150,000 $23,800 plus 30% of excess over $100,000 $150,000 $ 250,000 $38,800 plus 32% of excess over $150,000 $250,000 $ 500,000 $70,800 plus 34% of excess over $250,000 $500,000 $ 750,000 $155,800, plus 37% of the excess over $500,000 $750,000 $1 million $248,300, plus 39% of the excess over $750,000 $1 million $345,800, plus 40% of the excess over $1 million 241d. de’armond et al. / financial services review 23 (2014) 239–248 this exclusion allows for the transfer of substantial property over a period of years without the imposition of either the gift or estate tax. for example, an individual with five children can transfer $70,000 (5 " $14,000) per year to his or her children. a couple electing gift-splitting under sec. 2513 can transfer $140,000 per year to five children with no gift or estate tax liability. these amounts should increase because of the annual exclusion inflation adjustments for gifts made after 1998. a gift-giving program can result in the transfer of large amounts of property over several years with no gift or estate tax implications. 2.3. the marital deduction a donor spouse is allowed an unlimited deduction for lifetime gifts made to his or her spouse (sec. 2523). this allows for gifts between spouses with no gift or estate tax implications. certain terminable interests, interests that will terminate after a certain period or on the occurrence of some event, do not qualify for the marital deduction (sec. 2523). however, if it meets the requirements of sec. 2523(f), a life estate may qualify for the marital deduction if it is “qualified terminable interest property.” property is qualified terminable interest property (qtip) if: y the donee spouse is entitled to the income from the property for life; y income from the property is payable to the donee spouse at least annually; y no person has the power to appoint any part of the property to any person other than the donee spouse; and y an irrevocable election to have all or part of the trust qualify for the marital deduction is made by the donor and attached to the gift tax return (sec. 2523). 2.4. the generation skipping tax to remove the benefit of transferring property to grandchildren or otherwise skipping the estate or gift tax that will be paid by the immediately after generation, a generation-skipping transfer tax is imposed at a flat 40% rate on generation-skipping transfers. the most common generation-skipping transfers are direct transfers from a grandparent to a grandchild and life estates from the grandparent to the child with the remainder interest to the grandchild. generation-skipping transfers are defined in sec. 2603. these transfers include taxable distributions, taxable terminations, and direct skips. a taxable distribution is a distribution to a transferee who is a member of a generation at least two generations younger than the transferor. the amount received is subject to the tax (sec. 2611). the tax is paid by the transferee (sec. 2603). a taxable termination is the termination by death, lapse of time, release of power, or similar event of an interest held in trust that passes to a transferee who is a member of a generation at least two generations younger than the transferor. the value of the property in which the interest terminates is subject to the tax (sec. 2611). the tax is paid by the trustee (sec. 2603). a direct skip is a transfer of property to, or for the benefit of persons two or more generations below the transferor. the value of the property transferred is subject to the tax (sec. 2611). the tax is paid by the transferor (sec. 2603). there is a lifetime generation-skipping tax exemption of 242 d. de’armond et al. / financial services review 23 (2014) 239–248 $5,340,000 (in 2014) per grantor (sec. 2631). this amount will be indexed for inflation for decedents dying and gifts made after 2011 (code sec. 2631(c)). 2.5. couples or individuals with $5,340,000 or less. because of the structure of the unified transfer system, couples with net assets of $5,340,000 or less (smaller applicable unified credit for years earlier than 2014) will pay no federal estate tax. the only transfer tax planning that might need to be done would relate to state estate or inheritance taxes. the primary assistance that the tax advisor can provide is a strong recommendation that the individual execute a will, a durable power of attorney and an advance directive to physician. a revocable trust might also be considered. even with small estates, nontax considerations might warrant estate planning to protect the surviving spouse and other family members. there are some situations where one or both of the spouses might not want his or her assets to pass directly to the surviving spouse. one of the spouses might not be capable of managing the couple’s assets. there could be concern that the surviving spouse’s remarriage might result in the original couple’s assets coming under the control of the new spouse, and consequently, not being available for the offspring of the first marriage. in estate planning for a second marriage, one or both of the parties might want to provide for the surviving spouse during his or her lifetime, but want the assets to be available for the offspring of an earlier marriage upon the death of the surviving spouse. in these cases, creation of a trust might be in order. one option would be the creation of a revocable trust with a provision that all or part of the trust becomes irrevocable upon the death of one of the spouses. another option would be a testamentary trust created under the will of one or both of the spouses. in either case, the trust could provide for the income of the assets and some limited portion of the assets to be available to the surviving spouse during his or her lifetime, with the remaining assets to be distributed to their beneficiaries upon that spouse’s death. 3. former tax planning to use the unified credit of both spouses before the act discussed below, the most basic estate planning technique was to use both spouses’ unified credit by using the marital deduction effectively. this was not accomplished if all of the assets of the first spouse to die pass freely to the surviving spouse. in this case, the first to die has no net estate because of the marital deduction and no need for the unified credit. however, unless the surviving spouse manages to spend enough of the combined assets to reduce his or her estate to less than $5,340,000, this second spouse’s unified credit will not be enough to shield that estate from estate taxes. one approach was to leave all of the assets except $5,340,000 (or, if less, enough assets to reduce the second estate to $5,340,000) to the surviving spouse. the marital deduction would reduce the first estate to $5,340,000 and the first unified credit would offset all of the taxes. the second unified credit could then offset up to $5,340,000 of the second estate, with maximum use being made of both unified credits. except for very large estates, a direct transfer of assets to someone other than the surviving 243d. de’armond et al. / financial services review 23 (2014) 239–248 spouse might not be acceptable because the surviving spouse needs the assets, or at least the income from the assets, for living expenses. however, the use of a trust with a terminable interest that does not qualify as qualified terminable interest property could yield the same results. an effective, flexible tool is a qualified terminable interest property trust giving the executor the right to make the election as to how much of the life estate will be treated as qualified terminable interest property qualifying for the marital deduction. the executor will choose to not make the election for the amount of the property equal to the applicable exemption amount. a proper election will result in just enough property being left in the taxable estate to use up the decedent’s applicable unified credit. if a couple owned all of its property jointly, this technique could not be used successfully. jointly owned property passes automatically to the surviving owner. this property would qualify for the marital deduction, leaving nothing in the estate of the first spouse to die and everything in the estate of the surviving spouse. marital deduction planning could not be achieved if one of the spouses owned none or very little of the assets of the couple and that spouse dies first. there was no way for that spouse to retain assets to use up his or her unified credit. to facilitate this planning technique, the spouse owning the assets could make appropriate gifts to the other spouse. the unlimited marital deduction would prevent any taxable gifts. trusts serve estate planning purposes other than marital deduction planning. a trust with the income to be paid to the surviving spouse and the remaining assets to be paid to the couple’s children or other named beneficiaries upon the second death can be used to protect the assets for the benefit of both the surviving spouse and the beneficiaries. putting the assets in trust keeps the assets out of the control and out of the estate of a subsequent spouse of the decedent’s surviving spouse. if the trusts are set up so that they would qualify as qualified terminable interest property if the executor so elects, the executor of the estate can make the proper election to result in a marital deduction that maximizes the use of the couple’s applicable unified credits. 4. availability of new portability elections the tax relief, unemployment insurance reauthorization, and job creation act of 2010 (the act) contains a provision that will allow the unused portion of a decedent’s exclusion (taxable estate protected by the unified credit) to be used upon the subsequent death of the surviving spouse. the american taxpayer relief act of 2012 extended the portability election indefinitely. the act fixed the amount of taxable estate protected by the exclusion—the basic exclusion amount—at $5,000,000 through december 31, 2011. this amount has been indexed annually for decedents dying after 2011—resulting in $5,120,000 for 2012, $5,250,000 in 2013 and $5,340,000 in 2014. under irc sec. 2010(c)(2) the applicable exclusion amount is the sum of y the basic exclusion amount and, 244 d. de’armond et al. / financial services review 23 (2014) 239–248 y if a proper election has been made for a surviving spouse, the deceased spouse’s unused exclusion amount. under irc sec. 2010(c)(4) the deceased spouse’s unused exclusion amount is the lesser of: y the basic exclusion amount or y the excess of ! the basic exclusions amount of the last deceased spouse of the surviving spouse over, ! the tentative tax computed on the estate of that deceased spouse. the act goes on to require that the executor of the estate of the deceased must file an estate tax return computing the tax and making an election to make the unused exclusion available to the surviving spouse’s estate or the deceased spouse unused exclusion cannot be taken into account. the act also allows that the returns for which a portability election has been made can be examined after the statute of limitations has expired. the deceased spousal unused exclusion can be reduced or eliminated with this examination. however, no additional taxes can be assessed on the original return after the statute has expired. 4.1. if surviving spouse is not expected to remarry the portability election is fairly simple for situations where it appears the surviving spouse will not remarry. if all of the decedent’s assets pass to the surviving spouse, the marital deduction will reduce the taxable estate to zero and no exclusion will be needed to reduce the taxes due. bequests to children or other parties will reduce the marital deduction and increase the taxable estate—using some or all of the exclusion. if any of the exclusion is not used, the executor should make the election on the 706 when it is filed. 4.2. example mr. and mrs. smith have net assets of $7,500,000, owned equally by each spouse. mr. smith dies in 2014, leaving all of his assets directly to mrs. smith. mr. smith’s estate will have no net estate because of the marital deduction. the executor of mr. smith’s estate should make the election to make the unused exclusion available to mrs. smith’s estate. if mrs. smith still has $7,500,000 at death, her estate will use the combined exclusions from both estates to offset of the tentative tax on her estate. 4.3. surviving spouse is expected to remarry in the past, in marriages where the spouse died first, employment history of the couple, property laws and estate planning often resulted in the spouse owning more assets than the spouse. consequently, if the spouse died first owning assets equaling less than the exclusion amount less than the total exclusion was used. although times have changed and it is possible that more women have accumulated more assets individually, this situation may still be true for many couples. 245d. de’armond et al. / financial services review 23 (2014) 239–248 there are unique problems if the surviving spouse remarries. if the executor of the decedent’s estate makes the election to make the unused exclusion available to the surviving spouse’s estate and (s)he remarries, his or her estate will only be able to use the first spouse’s additional exclusion if (s)he dies before the death of the new spouse. there may also be family members concerned that the new spouse may acquire family assets upon the surviving spouse’s death. the solution is for the surviving spouse to make gifts to the children of the first marriage sufficient to use up the unused exclusion of the first decedent’s estate. the gifts must be completed before the death of the new spouse. 4.4. example mr. and mrs. smith have net assets of $7,500,000, owned equally by each spouse. mr. smith dies in 2014, leaving all of his assets directly to mrs. smith. mr. smith’s estate will have no net estate because of the marital deduction. the executor of mr. smith’s estate should make the election to make the unused exclusion available to mrs. smith’s estate. if mrs. smith remarries, she should make consider making gifts equal to mr. smith’s unused exemption. if mr. smith died in 2014, the unused exemption would be $5,340,000 and gifts of that amount would leave her with assets of only $2,160,000. the desire to use tax benefits should not overshadow good financial planning. she should not make gifts to the point that that she excessively depletes her assets. 4.5. should a protective election be made? in a situation where the combined net assets of a couple are less than $5,340,000, the question arises as to the advisability of making the election upon the death of one of the individuals. it is expensive to file a 706–estate tax return. however, it will be expensive if the surviving spouse’s assets increase (somebody wins the lottery every time) and the second spouse ends up with a net estate greater than $5,340,000. a protective election should be carefully considered. a situation where no election is filed and the second-to-die spouse’s estate ends up paying taxes could result in a lawsuit concerning the executor’s fiduciary responsibility to consider the election. 4.6. example mr. and mrs. smith have net assets of $7,500,000, owned equally by each spouse. mr. smith dies in 2014, leaving all of his assets to the couple’s children. because mrs. smith retains only $3,750,000 it appears that her exemption will cover her estate. the executor does not make the election. mrs. smith lives on her social security and is able to invest her assets well. she dies with an estate of $6,000,000—more that the exemption amount at that time. the executor should have made the protective election. reg. sec. 20.2010–2t(a)(7)(ii) provides some relief for executors in this situation. for items reported on schedules a – real estate, b – stocks and bonds, c – mortgages, notes and cash, d – insurance on the decedent’s life, e – jointly owned property, f – other miscellaneous property, g – transfers during decedent’s life, h – powers of appointment and i – annuities the 246 d. de’armond et al. / financial services review 23 (2014) 239–248 executor is allowed to estimate the value of the assets rather than acquire an acceptable appraisal. all other amounts on the return are subject to the normal rules of filing the 706. this provision significantly reduces the time and effort required to file a normal 706. 4.7. microsoft excel model to further explain and provide examples, we have created a microsoft excel model, see the appendix, portability of deceased spouse unused exclusion model. the model serves the purpose of allowing the reader to understand the numbers more specifically by entering information and numbers to exhibit how the basic premise of portability functions relative to other pertinent variables interact and function. the appendix includes additional notes pertaining to the model as well as a publicly available (live) download link. 5. conclusion time is of the essence when detangling tax issues and implications. it is with no doubt that the act has had lasting effects on how individuals learn, teach, and consider tax policies and provisions as they arise from tax code revision. the underexposed tax policies tend to provide excellent opportunity for academic exploration in and out of the classroom. as eyssell wrote in 1999, there is credibility and viability in “learning by doing,” such that pedagogy should use the creativity of teaching to further explore these avenues as they occur (eyssell, 1999). in the case of using the portability election of deceased spouses, relatively little has been written to this point, and even less attention has been turned to classroom curricula encompassing such opportunities, as for the most part, this is new and unchartered territory. although we are in the first stages of the post-cliff era, many future decisions made at governmental and regulating body levels remain to be made, and, each decision carries the potential to create changes along the way. the portability provision was extended and should be included in estate planning. accounting, law, and financial planning educators, especially within the tax courses, have a great opportunity to explore the exciting and changing realm of tax code with students given the post-cliff decisions, scenarios, implications, and conclusions. appendix: portability of deceased spouse unused exclusion model notes about excel model: available for public download at http://bit.ly/1tfm97h. the excel model is designed to illustrate the operation of the portability provisions of the act in the most basic terms. readers can input values in the highlighted fields, and the model will make calculations that demonstrate how portability operates on a couple’s estate. the model is designed for illustrative and educational purposes only and does not represent an exhaustive calculation of the estate tax liability for an estate. of particular note, the model requires certain assumptions for the purpose of making 247d. de’armond et al. / financial services review 23 (2014) 239–248 calculations. specifically, the model does not take into account lifetime gifts and their impact on the estate tax, and thus, assumes that no lifetime gifts have been made. additionally, the model assumes that property is owned by spouses equally and does not give any consideration to the impact of state property law regimes like community property. finally, in the interest in comparing “apples to apples”—more specifically 2024 dollars to 2024 dollars— the model makes an inflationary calculation to the exemption amount available in 2014. the model provides a growth calculation for the surviving spouse’s estate and the surviving spouse’s life expectancy. as a result, the model must inflate the exemption amount to the year of surviving spouse’s predicted death. this inflated amount provides a “best guess” of what the exemption may look like in a future year. in actuality, the irs acts each year to set the exemption amount based on an inflationary adjustment then rounds to the nearest $1,000. even with the above assumptions, the model provides an effective tool for illustrating portability and its effect on a couple’s estate. references beierlein, j. j., & neverett, m. (2013). who takes personal finance? financial services review, 22, 151–171. deener, j. (2012). new portability rules: a cure for incomplete estate planning. journal of accountancy, 214, 48–52. dudley, s. (2013). fear the regulatory cliff, too. reason, 44, 31–32. dunn, d. v. (2011). bypass the bypass trust? trusts & estates, 150, 24. eyssell, e. (1999) learning by doing: offering a university practicum in personal financial planning. financial services review, 8, 293–303. gallo, j. j. (2011). why portability isn’t a cure-all. journal of financial planning, 24, 32–33. grunwald, m. (2012). how to avoid the fiscal cliff. time, 180, 40. internal revenue service. (2011). guidance on electing portability of deceased spousal unused exclusion amount. internal revenue bulletin, 2011, 516–518. internal revenue code sec. 2001(c), as amended by 2012 taxpayer relief act §101(c)(1) code sec. 2010(c)(4)(b)(i), as amended by 2012 taxpayer relief act §101(c)(2). joseph, b. m. (2012). beware the portability election. california cpa, 80, 13–16. katz, r. d. (2011). what planners need to know about portability. cpa journal, 81, 50. miller, j. a. (2013). wealth transfer tax planning for 2013 and beyond. brigham young university law review, 2013, 878. suderman, p. (2013). the fiscal cliff. reason, 44, 26–28. 248 d. de’armond et al. / financial services review 23 (2014) 239–248 does active management work? evidence from equity sector funds crystal y. lina,* aschool of business, eastern illinois university, charleston, il 61920, usa abstract this study presents considerable evidence that equity sector mutual funds, the nine fidelity select portfolios here, have provided better after-expense returns against broader market etf, spy, and their peer sector etfs, the nine select sector spdr funds, over the sample period 1999–2010. not only do they achieve higher nominal returns over the 12 years, except for few sector mutual funds, some of the funds also generate higher risk-adjusted returns measured by sharpe ratio and ! from various asset pricing models. more important, none of the sector mutual funds generates a significant negative ! for the sample period no matter that asset pricing model is used. the results suggest that actively managed sector funds be considered by individual investors and/or their financial planners for mutual fund selection. © 2014 academy of financial services. all rights reserved. jel classification: g11; g12 keywords: mutual fund performance; active mutual fund management; sector investing 1. introduction the mutual fund industry has seen tremendous growth in the past few decades. the 2013 investment company institute fact book shows that 44.4% of u.s. households owned mutual funds in 2012, up significantly from 4.6% in 1980 and 23.4% in 1990, modestly down from 48.6% in 2000. bulk of individual’s mutual fund assets, $5.4 trillion out of $11.1 trillion, are invested in equities. individual accounts hold 90.3% of total equity mutual fund assets ($5.9 trillion). with the expansion of defined contribution plan assets, almost threefold from $1.7 * corresponding author. tel.: !1-217-581-2227; fax: !1-217-581-6642. e-mail address: cylin@eiu.edu (c.y. lin) financial services review 23 (2014) 249–271 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. trillion in 1995 to $5.1 trillion in 2012, knowledge in equity mutual funds becomes more and more important for individual investors. today many retirement savings plans offer a broader range of investment vehicles for individual investors. sector funds appear on the investment menu for many plan participants. do they deserve individual investors’ attention? how is their historical performance against their benchmarks? should financial planners recommend such mutual funds to their clients? using sector funds with the longest history, the fidelity select portfolios and the select sector spdr funds, this article tries to answer these questions for individual investors and/or their financial planners. once investors decide to allocate their assets to the u.s. public equity market, they must evaluate possible ways to implement the allocation in their subportfolios. there are two dimensions of this implementation. the first one is indexing or not: do investors want to passively invest so that their returns closely track selected indexes and at the same time investors pay less fees and experience less turnover? or do investors want to actively manage their subportfolios either in-house or through external fund managers? sullivan and xiong (2012) estimate that $1.2 trillion out of $3.5 trillion assets in the u.s. equity mutual funds and etfs was passively managed as of september 2010; equity index mutual funds and equity etfs split the share of passively managed equity index funds.1 the second dimension of this implementation is whether to make allocation decisions at the sector level: do investors want to further divide stocks by sector/industry and set up weight limits to these groups? or do investors not care about sector issues at all. porter (1985) and mcgahan and porter (1997) make a strong case for sector investing: they demonstrate that a company’s performance is influenced by the growth and structure of its industry. in addition, groysberg et al., (2011) show that forecasted industry growth is the most important explanatory variable when analysts construct their forecasts on companies. in this article, i examine the performance of actively managed equity sector funds and their passively managed counterparties to assist investors in their decisions in implementing equity sector asset allocations. the research question i try to answer is: when an investor wants to use external fund managers to allocate assets among u.s. equity sectors, are sector index funds a better choice than actively managed funds or vice versa? two parallel analyses regarding equity sector fund performance are provided in this article. one is their performance against a broad u.s. equity market index fund, which tests the efficient market hypothesis (emh). this analysis serves as a general empirical study on u.s. equity market efficiency. the other is sector mutual funds’ performance against sector index funds, which tests emh in a smaller territory: equity sector. it is arguable that using a broad equity market index is not appropriate when evaluating a sector fund because it has a much smaller universe of securities from which to choose. an index for the same sector would be more appropriate when used as a benchmark. in reality, this is what many actively managed funds do: they select an index that best matches their investment universe as their benchmark. 2. sector mutual funds and etfs a variety of equity sector funds have been introduced to the market place during the past several decades. these sector funds allow investors to custom tailor asset allocations to fit 250 c.y. lin / financial services review 23 (2014) 249–271 their particular investment needs or goals. like their broad equity market counterparties, index sector funds emerged later than actively managed sector funds. two groups of sector funds with the longest history in each category are used in this study: the fidelity select portfolios and the select sector spdr funds. in addition, spdr s&p 500 etf is used as an investable benchmark for the study. 2.1. fidelity select portfolios the fidelity’s web site2 listed 38 mutual funds under its stock funds/sector funds category at the time of writing this article. the inception date of the earliest three sector funds (energy, healthcare, and technology) is july 14, 1981. i identified nine broader sector funds based on fund prospectus: select consumer discretionary portfolio (fscpx), select consumer staples portfolio (fdfax), select energy portfolio (fsenx), select financial services portfolio (fidsx), select health care portfolio (fsphx), select industrials portfolio (fcyix), select materials portfolio (fsdpx), select technology portfolio (fsptx), and select utilities portfolio (fsutx).3 the above funds match nine sector index funds that are described later. these sector mutual funds were started between july 1981 and march 1997. most of the other funds listed are narrower focused industry funds. according to fidelity’s fund prospectuses, these funds seek capital appreciation, invest in domestic and foreign issuers, normally invest primarily in common stocks, and invest at least 80% of assets in securities of companies principally engaged in the selected sector. in other words, these funds are actively managed. fidelity management & research company is the fund’s manager. the fidelity select portfolios have an expense ratio between 0.80% (healthcare) and 1.41% (materials) as of february 29, 2012. the portfolio turnover rate is between 35% (consumer staples) and 384% (financial services). the funds have net assets between $0.28 billion (consumer discretionary) and $2.50 billion (energy).4 fidelity charges a short-term redemption fee, 0.75%, when money is withdrawn from a sector fund within 30 days of purchase to reduce short-term mutual fund trading.5 all these sector funds are open to new investors. 2.2. select sector spdr funds the select sector spdr trust was organized as a massachusetts business trust on june 10, 1998. state street global advisors serves as the fund manager. the trust consists of nine separate investment portfolios (each a “select sector spdr fund”) incepted in december 1998: the consumer discretionary select sector spdr fund (xly), the consumer staples select sector spdr fund (xlp), the energy select sector spdr fund (xle), the financial select sector spdr fund (xlf), the health care select sector spdr fund (xlv), the industrial select sector spdr fund (xli), the materials select sector spdr fund (xlb), the technology select sector spdr fund (xlk), and the utilities select sector spdr fund (xlu). these sector funds seek to provide investment results that, before expenses, correspond generally to the price and yield performance of publicly traded equity securities of companies in certain “select sector indexes”: the consumer discretionary 251c.y. lin / financial services review 23 (2014) 249–271 select sector index, the consumer staples select sector index, the energy select sector index, the financial select sector index, the health care select sector index, the industrial select sector index, the materials select sector index, the technology select sector index, and the utilities select sector index. each stock in the s&p 500 is allocated to one and only one select sector index. the combined companies of the nine select sector indexes represent all of the companies in the s&p 500. that is, the select sector spdr funds unbundle the s&p 500.6 these passively managed sector funds use a replication strategy, attempting to track the performance of an unmanaged index of securities. according to select sector spdr fund annual report, the ratio of expenses to average net assets is 0.19% for each individual fund as of september 30, 2011. the turnover rate is between 3.20% (utilities) and 13.86% (materials). the funds have net assets between $1.64 billion (materials) and $6.64 billion (utilities).7 2.3. spdr s&p 500 etf the spdr s&p 500 etf (spy) is an exchange traded fund designed to generally correspond to the price and yield performance of the s&p 500 index. utilizing a full replication approach, the trust owns all 500 securities of the s&p 500 index in their approximate market capitalization weight. the fund was incepted on january 22, 1993. the portfolio has an expense ratio of 0.09%, a turnover rate of 3.72%, and $80.87 billion net assets as of september 30, 2011.8 both equity mutual funds and etfs are pooled investments that represent ownership in a basket of stocks. however, etfs can be traded like individual stocks. they are also shortable, marginable, and optionable. index etfs normally have lower fees by eliminating many of the operating, research, and transaction expenses incurred by active money managers. they also provide greater transparency: one can get a holding list more frequently than with mutual funds. for example, fidelity select portfolios publish monthly holdings whereas the select sector spdrs update their online information daily. 3. a first look at equity sector funds: raw returns my analysis starts in january 1999 and ends in december 2010,9 since the earliest price data available for select sector spdr funds is mid-december 1998. i downloaded price and dividend data from yahoo!finance web site.10 monthly, annual, and 12-year holding period return is calculated for each fund as: ri,t " pi,t # di,t pi,t"1 $ 1, (1) where ri,t is the return for fund i during period t, pi,t is the price for fund i at the end of period t, pi,t"1 is the price for fund i at the end of period t " 1, and di,t is the total dividend/cash distribution of fund i during period t. 252 c.y. lin / financial services review 23 (2014) 249–271 the purpose of this study is to assist investors implement asset allocations at the sector level, therefore, i use fund price instead of fund net asset value to calculate fund return. this return is net of expenses and is attainable. as argued by jones and wermers (2011), i compare actively managed sector mutual fund performance to their passive alternative and not to the index itself. for the same reason, i use spy as an investable broad u.s. equity market benchmark. 3.1. twelve-year return which group of funds generates higher returns during the sample period? the results are summarized in table 1. in the rest of the article, i use mf to represent fidelity select portfolios, etf to represent select sector spdr funds, and spy to represent the spdr s&p 500 etf for easier reference. panel a of table 1 lists the 12-year (1999–2010) holding period return for the 19 funds. the highest return is 295.2% for the energy mf and the lowest return is "17.4% for the technology etf, and the return for spy is 20.8%. seven of nine mfs, except for the utilities and consumer discretionary mf, have a higher return than that of their etf counterparties; the average outperformance is 51.2%. when spy is used as the benchmark, eight out of nine mfs (except utilities) and seven out of nine etfs (except financial and technology) outperform. the average mf holding period return across all sectors is 109.2% versus 58.0% for etfs, and the difference is significant at the 5% level using one-tailed t test. results are slightly different when i compound annual holding period returns through the 12 years. panel b of table 1 shows that the only underperforming mf against spy is financial, not utilities. the same seven mfs beat peer etfs with an average outperformance of 63.3%. the average mf 12-year return with annual compounding across all sectors is 127.6% versus 64.3% for etfs, and the difference is also significant at the 5% level using one-tailed t test. these results show that on average sector mfs outperform their etf counterparties for the whole sample period. none of the nine mfs suffers a loss during the sample period; however, the financial and technology etfs generate a negative return under both calculation methods. the tech bubble in early 2000s and the financial crisis in 2008 contribute to the negative 12-year return for these two sectors. although sector etfs mimicked their benchmark indexes during these bear markets, it seems sector mf managers made the right decisions against trend changes. fig. 1 shows that the annual return for the technology mf is 119.07%, "28.18%, and "31.70% for the year 1999, 2000, and 2001, whereas the annual return for its peer etf is 65.13%, "41.89%, and "23.34%, respectively. the annual return for the financial mf is "13.42%, "49.83%, and 23.71% for the year 2007, 2008, and 2009, whereas the annual return for its peer etf is "18.89%, "54.06%, and 16.98%, respectively. 3.2. average annual return seven out of nine mfs (except for utilities and consumer discretionary) outperform their peer etfs when measured with average annual holding period return depicted in fig. 2. an interesting finding in fig. 2 is that the ranking of performance is different for mfs and etfs. energy mf has the highest average annual return of 18.3%, followed by materials (17.0%), 253c.y. lin / financial services review 23 (2014) 249–271 technology (14.9%), industrials (11.5%), consumer staples (7.7%), healthcare (6.3%), utilities (4.7%), consumer discretionary (4.6%), and financial (3.4%). in the etf group, energy (13.8%) and materials (9.7%) are still ranked first and second, whereas financial (1.9%) is again at the bottom. however, the other six sectors are ranked differently. fig. 2 also shows that for the materials, industrials, technology, and consumer staples sector, mf average annual holding period returns are higher than that of corresponding etfs at the 1% or 5% significance level using one-tailed t test. table 1 twelve-year return, 1999–2010 sector mf etf mf beats etf difference mf beats spy etf beats spy a: holding period return b 287.6% 105.1% yes 182.5% yes yes e 295.2% 221.3% yes 73.9% yes yes f 23.4% "8.0% yes 31.4% yes no i 125.9% 62.1% yes 63.8% yes yes k 73.5% "17.4% yes 90.9% yes no p 82.0% 26.4% yes 55.7% yes yes u 20.6% 41.3% no "20.7% no yes v 50.4% 36.2% yes 14.2% yes yes y 24.0% 55.0% no "30.9% yes yes average fund return 109.2% 58.0% p-value average mf return higher than average etf return 0.022 spy 20.8% b: return with annual compounding b 325.9% 118.7% yes 207.2% yes yes e 354.3% 243.2% yes 111.0% yes yes f 8.4% "13.7% yes 22.0% no no i 148.8% 67.8% yes 81.0% yes yes k 70.5% "16.2% yes 86.7% yes no p 111.2% 31.6% yes 79.6% yes yes u 25.7% 51.8% no "26.1% yes yes v 67.6% 37.4% yes 30.1% yes yes y 36.4% 58.4% no "21.9% yes yes average fund return 127.6% 64.3% p-value average mf return higher than average etf return 0.015 spy 23.9% note: b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. holding period return in panel a is calculated by adding ending price and all dividends paid in 12 years then divided by beginning price. return with annual compounding in panel b is calculated by compounding annual holding period returns for each individual fund. 254 c.y. lin / financial services review 23 (2014) 249–271 -100% -50% 0% 50% 100% materials -60% -40% -20% 0% 20% 40% 60% energy -60% -40% -20% 0% 20% 40% 60% industrials -40% -20% 0% 20% 40% consumer staples -60% -40% -20% 0% 20% 40% healthcare -60% -40% -20% 0% 20% 40% financial -60% -40% -20% 0% 20% 40% 60% 80% 100% 120% 140% technology -40% -20% 0% 20% 40% utilities -60% -40% -20% 0% 20% 40% 60% consumer discretionary mf etf spy fig. 1. annual holding period return. 255c.y. lin / financial services review 23 (2014) 249–271 when annual holding period return is compared between mfs and etfs, for the same sector, table 2 shows that seven out of nine mfs (except utilities and consumer discretionary) generate a higher return in seven or more years in the 12 year sample period. the average number of years of outperforming is 7.9 and the percentage of years of outperforming is 66%. using spy as the benchmark, on average in 7.4 out of 12 years mfs beat spy (62% of years); only in 6.4 years does the etf group beat spy (54% of years). table 2 annual return comparison, 1999–2010 sector mf beats etf mf beats spy etf beats spy number of years % of years number of years % of years number of years % of years b 9 75% 10 83% 9 75% e 7 58% 9 75% 7 58% f 9 75% 6 50% 7 58% i 10 83% 9 75% 7 58% k 8 67% 6 50% 4 33% p 10 83% 9 75% 5 42% u 6 50% 6 50% 6 50% v 7 58% 7 58% 5 42% y 5 42% 5 42% 8 67% average 7.9 66% 7.4 62% 6.4 54% note: b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. total number of years: 12. p-value average mf annual holding period return higher than average etf annual holding period return b e f i k p u v y 0.012 0.082 0.092 0.002 0.046 0.013 0.430 0.300 0.222 0% 5% 10% 15% 20% b e f i k p u v y mf etf spy fig. 2. arithmetic mean of annual holding period return. b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. 256 c.y. lin / financial services review 23 (2014) 249–271 3.3. decomposition of 12-year holding period return what portion of that 12-year holding period return is contributed by capital gains? what portion is contributed by dividend yield (regular dividend and special cash distribution)? there are four mfs (financial, utilities, healthcare, and consumer discretionary) and two etfs (financial and technology) have negative capital gains during the period. the spy has a capital gains yield of 2.0%. actually, only three mfs (materials, energy, and industrials) have a higher capital gains yield than dividend yield. that number is five for etfs (materials, energy, industrials, healthcare, and consumer discretionary). the spy has a dividend yield of 18.8%. with 2.0% capital gains yield from spy, the 12-year sample period is pretty flat. the s&p 500 price index has two peaks, 1552.87 on march 24, 2000 and 1576.09 on october 17, 2007, and two troughs, 768.63 on october 10, 2002 and 666.79 on march 6, 2009. this range provides a good testing field for performance analysis. after examining 12-year returns and average annual returns, i find that most sector mfs outperform both their etf peers and spy for most years during the 1999–2010 sample period. the exceptions are the utilities and consumer discretionary mf. 4. a closer look at equity sector funds: risk adjusted returns before drawing a conclusion on sector mfs’ performance, one must investigate the risk dimension of the returns. here, i look at risk adjusted returns that incorporate both total risk and systematic risk based on monthly holding period returns. table 3 summarizes statistics of both monthly holding period returns and excess returns, which are calculated as monthly holding period return minus monthly 1-month t-bill rate. panel a shows the maximum monthly holding period return is 31.446% (technology mf, february 2000) and the minimum monthly return is "28.379% (technology mf, february 2001). because sector funds focus on specific investment areas, it is expected that both sector mfs and etfs have a higher volatility when compared to a broader market benchmark. that table 3 summary statistics of monthly return (%) maximum minimum mean median sd skewness kurtosis n sharpe ratio a: raw return mf 31.446 "28.379 0.695 1.014 6.316 "0.238 2.738 1289 etf 24.768 "26.198 0.468 0.796 5.982 "0.255 1.762 1296 spy 9.935 "16.519 0.260 0.741 4.652 "0.485 0.594 144 overall 31.446 "28.379 0.564 0.897 6.081 "0.246 2.326 2729 b: excess return mf 31.016 "28.769 0.474 0.759 6.321 "0.229 2.696 1289 0.075 etf 24.628 "26.198 0.245 0.533 5.985 "0.246 1.730 1296 0.041 spy 9.925 "16.599 0.038 0.526 4.660 "0.454 0.535 144 0.008 overall 31.016 "28.769 0.342 0.652 6.085 "0.236 2.288 2729 0.056 257c.y. lin / financial services review 23 (2014) 249–271 is true as shown in table 3. the sector mfs have the widest span of monthly returns (59.825%), followed by sector etfs (50.966%); both are much higher than that of spy (26.454%). the standard deviation for sector mf, etf, and spy is 6.316%, 5.982%, and 4.652%, respectively. all fund returns are negatively skewed. these characteristics are similar in excess returns as presented in panel b. the group sharpe ratio is 0.075, 0.041, 0.008, and 0.056 for the mfs, etfs, spy, and overall funds, respectively. the sector mf group has the highest sharpe ratio. these small but positive sharpe ratios reflect the flat u.s. equity market during the sample period. 4.1. sharpe ratio does the sharpe ratio comparison between each pair of sector funds echo the group sharpe ratio results in table 3? table 4 shows that for the full sample period, january 1999 through december 2010, seven out of nine sector mfs have a higher sharpe ratio than that of their peer etfs, except for the utilities and consumer discretionary mf. the materials table 4 individual fund sharpe ratio, january 1999 through december 2010 fund mean excess return sd sharpe ratio mf sr higher than etf mf sr higher than spy etf sr higher than spy bbb 1.030 6.786 0.152 yes yes xlb 0.555 6.760 0.082 yes eee 1.100 7.390 0.149 yes yes xle 0.853 6.504 0.131 yes fff 0.035 6.224 0.006 yes no xlf "0.089 6.886 "0.013 no iii 0.669 6.157 0.109 yes yes xli 0.314 5.868 0.054 yes kkk 0.671 10.126 0.066 yes yes xlk "0.015 8.069 "0.002 no ppp 0.367 3.728 0.098 yes yes xlp 0.042 3.709 0.011 yes uuu 0.066 5.021 0.013 no yes xlu 0.184 4.671 0.039 yes vvv 0.224 4.175 0.054 yes yes xlv 0.093 4.280 0.022 yes yyy 0.112 4.896 0.023 no yes xly 0.272 5.899 0.046 yes spy 0.038 4.660 0.008 average mf sr 0.074 average etf sr 0.041 p-value average mf sr higher than average etf sr 0.020 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. xlb, xle, xlf, xli, xlk, xlp, xlu, xlv, and xly represent the select sector spdr etfs for these nine sectors, respectively. spy represents the spdr s&p 500 trust. 258 c.y. lin / financial services review 23 (2014) 249–271 and energy mf have the highest sharpe ratios, 0.152 and 0.149, respectively; whereas the financial and technology etf have a negative sharpe ratio for the same period. the average mf sharpe ratio across all sectors is 0.074 versus 0.041 for sector etfs, and the difference is significant at the 5% level using one-tailed t test. when the sample period is divided into two equal-length subperiods, the results are slightly different.11 for the first half sample period, from january1999 through december 2004, the healthcare mf does not have a higher sharpe ratio than its peer etf. for the second half sample period, from january 2005 through december 2010, the energy mf does not have a higher sharpe ratio, but the consumer discretionary mf does outperform its peer etf. subsample analysis also shows that the significant higher mf average sharpe ratio for the whole sample period is mainly because of higher mf average sharpe ratio in the second half sample period, which is significant at the 1% level, whereas the p-value is greater than 5% for the first half sample period. table 4 also compares sharpe ratio between individual sector funds and spy. only the financial mf underperforms spy for the full sample period, whereas both the financial and technology etf underperform spy for the same time period. using total risk as the measurement, i find that most sector mfs outperform their peer etfs. sector mfs also have a higher number of funds outperform spy. 4.2. performance against s&p 500 etf systematic risk is always the part of risk that gets more attention because many argue that unsystematic risk can be diversified away at a relatively low cost.12 capm has been the standard model to test fund performance. i modify the model by replacing excess market return with excess spy return because an investable benchmark makes more sense for comparing attainable returns: ri,t $ rf,t " !i # %i#rspy,t $ rf,t$ # &i,t (2) where ri,t is the return of fund i in month t, rf,t is the return of one-month t-bill in month t, rspy,t is the spy return in month t, and &i,t is an error term. table 5 shows that the materials mf generates a 0.988% monthly abnormal return (11.856% annually) and the energy mf generates a 1.064% monthly abnormal return (12.768% annually) during the sample period, significant at the 1% and 5% level, respectively. the industrial etf has a significant ! of 0.672% at the 1% level. all %s are significant at the 1% level. both mfs and etfs for the materials sector, the financial sector, the industrials sector, and the technology sector have a % greater than 1. all other funds have a % less than 1 except for the consumer discretionary etf. the average adjusted r2 is 0.563 for the mfs and 0.555 for the etfs. for the first half sample period, none of the !s is significant at the 5% level. the average adjusted r2 declines to 0.438 for the mfs and 0.461 for the etfs. the average adjusted r2 are higher for the second half sample period: 0.719 for the mfs and 0.692 for the etfs. the materials mf, the consumer staples mf, and the industrials etf generate 0.940%, 0.495%, and 0.543% monthly abnormal returns, respectively, from january 2005 through december 2010, at the 5% level. 259c.y. lin / financial services review 23 (2014) 249–271 fama-french three-factor model is also modified by replacing excess market return with excess spy return: ri,t $ rf,t " !i # %i#rspy,t $ rf,t$ # hihmlt # sismbt # &i,t (3) where hml (high minus low) is the average return on two value portfolios minus the average return on two growth portfolios and smb (small minus big) is the average return on three small portfolios minus the average return on three big portfolios (fama and french 1993). the data are downloaded from kenneth r. french data library.13 results in table 6 show that over the full sample period two sector mfs (materials and technology) have a positive ! at the 5% level. none of the etfs has a significant ! at the 5% level. all %s are significant at the 1% level with the same above/below 1 % distribution as in table 5. hml is not significant for the healthcare mf and etf, and it is significantly negative for three funds (technology mf and etf, and utilities mf). the smb, size factor, is the least significant factor of the three: only seven out of 18 funds have a significant coefficient. during the first half sample period, none of the mfs or etfs generates a positive ! at the 5% level. fewer hml and smb coefficients are significant. during the second half sample table 5 one factor model results using spy excess return, january 1999 through december 2010 fund ! % adjusted r2 estimate se t-value estimate se t-value bbb 0.988 0.362 2.727*** 1.121 0.078 14.379*** 0.590 xlb 0.512 0.347 1.475 1.145 0.075 15.321*** 0.620 eee 1.064 0.491 2.164** 0.962 0.106 9.086*** 0.363 xle 0.821 0.436 1.882 0.834 0.094 8.880*** 0.352 fff "0.005 0.309 "0.018 1.074 0.067 16.126*** 0.644 xlf "0.134 0.345 "0.387 1.183 0.074 15.921*** 0.638 iii 0.272 0.230 1.183 1.164 0.052 22.589*** 0.789 xli 0.672 0.242 2.782*** 1.112 0.050 22.444*** 0.778 kkk 0.606 0.522 1.162 1.711 0.112 15.228*** 0.617 xlk "0.071 0.359 "0.197 1.466 0.077 18.977*** 0.715 ppp 0.349 0.248 1.407 0.485 0.053 9.092*** 0.363 xlp 0.025 0.256 0.097 0.451 0.055 8.186*** 0.315 uuu 0.035 0.278 0.127 0.807 0.060 13.463*** 0.557 xlu 0.165 0.339 0.486 0.497 0.073 6.807*** 0.240 vvv 0.202 0.268 0.755 0.574 0.058 9.950*** 0.406 xlv 0.066 0.223 0.296 0.718 0.048 14.963*** 0.609 yyy 0.078 0.209 0.374 0.903 0.045 20.058*** 0.737 xly 0.231 0.256 0.902 1.082 0.055 19.611*** 0.728 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. xlb, xle, xlf, xli, xlk, xlp, xlu, xlv, and xly represent the select sector spdr etfs for these nine sectors, respectively. spy represents the spdr s&p 500 trust. ***significant at the 1% level. **significant at the 5% level. 260 c.y. lin / financial services review 23 (2014) 249–271 t ab le 6 t hr ee fa ct or re gr es si on re su lts us in g sp y ex ce ss re tu rn ,j an ua ry 19 99 th ro ug h d ec em be r 20 10 fu nd ! % h m l sm b a dj us te d r 2 e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e b b b 0. 71 9 0. 34 0 2. 11 7* * 1. 14 5 0. 07 2 15 .9 01 ** * 0. 50 4 0. 09 3 5. 39 6* ** 0. 13 7 0. 09 1 1. 50 0 0. 65 6 x l b 0. 32 7 0. 33 0 0. 99 0 1. 17 4 0. 07 0 16 .7 50 ** * 0. 43 0 0. 09 1 4. 73 6* ** 0. 04 0 0. 08 9 0. 44 8 0. 67 2 e e e 0. 92 5 0. 49 7 1. 86 1 0. 98 2 0. 10 5 9. 31 4* ** 0. 31 4 0. 13 7 2. 29 9* * 0. 03 6 0. 13 4 0. 26 6 0. 37 9 x l e 0. 73 3 0. 43 8 1. 67 5 0. 86 1 0. 09 3 9. 27 3* ** 0. 29 6 0. 12 0 2. 45 7* * " 0. 04 1 0. 11 8 " 0. 35 1 0. 37 8 ff f " 0. 20 9 0. 24 7 " 0. 84 5 1. 12 4 0. 05 2 21 .4 22 ** * 0. 59 5 0. 06 8 8. 74 7* ** " 0. 03 6 0. 06 6 " 0. 54 0 0. 78 3 x l f " 0. 35 2 0. 26 5 " 1. 32 6 1. 24 3 0. 05 6 22 .0 85 ** * 0. 68 7 0. 07 3 9. 41 2* ** " 0. 07 1 0. 07 1 " 0. 99 0 0. 79 6 ii i 0. 37 3 0. 20 5 1. 81 8 1. 17 2 0. 04 3 27 .3 37 ** * 0. 45 4 0. 05 6 8. 05 7* ** 0. 17 7 0. 05 5 3. 20 4* ** 0. 85 6 x l i 0. 11 8 0. 21 2 0. 55 5 1. 13 3 0. 04 5 25 .2 00 ** * 0. 33 8 0. 05 8 5. 79 3* ** 0. 04 7 0. 05 7 0. 83 0 0. 82 0 k k k 0. 61 4 0. 27 1 2. 26 2* * 1. 56 5 0. 05 8 27 .1 83 ** * " 1. 01 4 0. 07 5 " 13 .5 80 ** * 0. 65 0 0. 07 3 8. 90 3* ** 0. 90 1 x l k 0. 12 6 0. 22 4 0. 56 4 1. 38 6 0. 04 7 29 .1 97 ** * " 0. 80 4 0. 06 2 " 13 .0 57 ** * 0. 18 4 0. 06 0 3. 06 1* ** 0. 89 4 pp p 0. 23 9 0. 21 5 1. 11 3 0. 52 2 0. 04 6 11 .4 61 ** * 0. 39 0 0. 05 9 6. 60 9* ** " 0. 06 5 0. 05 8 " 1. 12 4 0. 54 4 x l p 0. 05 1 0. 23 1 0. 21 9 0. 49 2 0. 04 9 10 .0 34 ** * 0. 25 0 0. 06 4 3. 93 6* ** " 0. 20 8 0. 06 2 " 3. 35 5* ** 0. 46 7 u u u 0. 17 5 0. 27 9 0. 62 7 0. 80 0 0. 05 9 13 .5 28 ** * " 0. 22 4 0. 07 7 " 2. 92 4* ** " 0. 09 5 0. 07 5 " 1. 27 0 0. 57 7 x l u 0. 13 9 0. 31 4 0. 44 2 0. 54 5 0. 06 7 8. 18 8* ** 0. 36 0 0. 08 6 4. 17 3* ** " 0. 19 0 0. 08 4 " 2. 25 7* * 0. 38 1 v v v 0. 17 4 0. 27 6 0. 63 0 0. 57 9 0. 05 8 9. 90 5* ** 0. 06 9 0. 07 6 0. 90 9 0. 00 5 0. 07 4 0. 06 5 0. 40 2 x l v 0. 15 6 0. 22 7 0. 68 8 0. 72 6 0. 04 8 15 .0 86 ** * " 0. 05 9 0. 06 2 " 0. 94 7 " 0. 11 8 0. 06 1 " 1. 93 1 0. 61 4 y y y " 0. 18 3 0. 18 7 " 0. 98 0 0. 90 2 0. 04 0 22 .7 62 ** * 0. 32 2 0. 05 1 6. 26 8* ** 0. 24 3 0. 05 0 4. 83 2* ** 0. 80 0 x l y 0. 03 1 0. 24 6 0. 12 5 1. 09 0 0. 05 2 20 .8 55 ** * 0. 30 9 0. 06 8 4. 55 3* ** 0. 14 6 0. 06 6 2. 19 9* * 0. 76 0 n ot e: b b b ,e e e ,f ff ,i ii ,k k k ,p pp ,u u u ,v v v ,a nd y y y re pr es en t th e fi de lit y se le ct po rt fo lio m ut ua l fu nd s fo r th e m at er ia ls se ct or ,t he e ne rg y se ct or , th e fi na nc ia l se ct or , th e in du st ri al s se ct or , th e t ec hn ol og y se ct or , th e c on su m er st ap le s se ct or , th e u til iti es se ct or , th e h ea lth ca re se ct or , an d th e c on su m er d is cr et io na ry se ct or ,r es pe ct iv el y. x l b ,x l e ,x l f, x l i, x l k ,x l p, x l u ,x l v ,a nd x l y re pr es en t th e se le ct se ct or sp d r e t fs fo r th es e ni ne se ct or s, re sp ec tiv el y. sp y re pr es en ts th e sp d r s& p 50 0 t ru st . ** *s ig ni fic an t at th e 1% le ve l. ** si gn ifi ca nt at th e 5% le ve l. 261c.y. lin / financial services review 23 (2014) 249–271 period, three sector mfs (materials, industrials, and consumer staples) have a significant positive !. the financial etf generates a significant negative ! at the 5% significance level. the four-factor model is modified by replacing excess market return with excess spy return: ri,t $ rf,t " !i # %i#rspy,t $ rf,t$ # hihmlt # sismbt # mimomt # &i,t (4) where mom (momentum) is the average return on the two high prior return portfolios minus the average return on the two low prior return portfolios (see carhart 1997). the data are downloaded from kenneth r. french data library. the results presented in table 7 resemble those from the three-factor model. table 7 shows that over the full sample period the same two sector mfs (materials and technology) have a positive ! at the 5% level. for the momentum factor, mom, only five out of 18 funds have a significant coefficient. generally speaking, more often, sector mfs generate significant higher !s than peer etfs no matter whether a one-factor, three-factor, or four-factor model is adopted when spy is used as the proxy for market portfolio. all fund returns are sensitive to the overall equity market movements as the % coefficients for excess spy return are all significant at the 1% level. fund returns are less sensitive to the value/growth factor, the size factor, and the momentum factor in the order of listing. the findings on performance against spy are not consistent with emh. they support the argument by kacperczyk, sialm, and zheng (2005) that more sector/industry concentrated funds perform better. 4.3. sector mutual fund performance against sector index funds let us examine mf performance against peer etf within each sector. excess return of peer etf is used as the independent variable for the one-factor model: rmfj,t $ rf,t " !j # %j#retfj,t $ rf,t$ # &j,t (5) where rmfj,t is the return of mf of sector j in month t, and retfj,t is the return of etf in the same sector j in month t. table 8 reports that three sector mfs (materials, industrials, and technology) generate significant positive !s, 0.506%, 0.446%, and 0.689%, respectively, during the full sample period. that is equivalent to annualized outperformance of 6.072%, 5.352%, and 8.268%, respectively. subsample analysis shows that none of the sector mfs outperforms in the first half sample period, whereas the materials and the industrials mf outperform in the second half sample period. all %s are positive and significant at the 1% level. which model can explain most of the return variances of sector funds? i summarize the adjusted r2 for the four models in table 9. most of the time, adding hml and smb does increase the adjusted r2 when using spy as the benchmark. only one out of 18 regressions for the full sample period suffers a slight decrease of explaining power, 0.004. adding mom, however, does not increase adjusted r2 across the board. over the full sample period, the average adjusted r2 across all funds is 0.559, 0.649, 0.653, and 0.764 for the one/three/four-factor model using spy and one-factor 262 c.y. lin / financial services review 23 (2014) 249–271 t ab le 7 fo ur fa ct or re gr es si on re su lts us in g ex ce ss sp y re tu rn ,j an ua ry 19 99 th ro ug h d ec em be r 20 10 fu nd ! % h m l sm b m o m a dj us te d r 2 e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e b b b 0. 70 6 0. 34 1 2. 07 2* * 1. 17 1 0. 08 1 14 .5 04 ** * 0. 51 5 0. 09 5 5. 43 1* ** 0. 12 8 0. 09 2 1. 38 7 0. 04 2 0. 05 8 0. 71 7 0. 65 5 x l b 0. 31 6 0. 33 2 0. 95 3 1. 19 6 0. 07 9 15 .2 14 ** * 0. 44 0 0. 09 2 4. 76 4* ** 0. 03 2 0. 09 0 0. 35 9 0. 03 5 0. 05 7 0. 62 7 0. 67 e e e 0. 87 8 0. 49 4 1. 77 8 1. 07 7 0. 11 7 9. 19 8* ** 0. 35 6 0. 13 8 2. 58 6* * 0. 00 3 0. 13 4 0. 02 4 0. 15 2 0. 08 4 1. 80 4 0. 38 9 x l e 0. 68 4 0. 43 3 1. 58 2 0. 96 1 0. 10 3 9. 36 2* ** 0. 33 9 0. 12 1 2. 81 4* ** " 0. 07 5 0. 11 7 " 0. 64 3 0. 15 9 0. 07 4 2. 15 9* * 0. 39 4 ff f " 0. 18 8 0. 24 6 " 0. 76 3 1. 08 0 0. 05 8 18 .5 13 ** * 0. 57 6 0. 06 9 8. 40 2* ** " 0. 02 1 0. 06 7 " 0. 31 5 " 0. 07 0 0. 04 2 " 1. 65 8 0. 78 6 x l f " 0. 31 9 0. 26 1 " 1. 22 1 1. 17 7 0. 06 2 18 .9 88 ** * 0. 65 8 0. 07 3 9. 03 7* ** " 0. 04 8 0. 07 1 " 0. 67 7 " 0. 10 6 0. 04 5 " 2. 38 1* * 0. 80 2 ii i 0. 37 0 0. 20 6 1. 79 6 1. 17 8 0. 04 8 24 .3 18 ** * 0. 45 6 0. 05 7 7. 99 0* ** 0. 17 4 0. 05 6 3. 11 6* ** 0. 00 9 0. 03 5 0. 25 3 0. 85 5 x l i 0. 13 2 0. 21 2 0. 62 3 1. 10 4 0. 05 0 21 .9 92 ** * 0. 32 5 0. 05 9 5. 51 3* ** 0. 05 7 0. 05 7 0. 99 6 " 0. 04 6 0. 03 6 " 1. 27 3 0. 82 1 k k k 0. 61 0 0. 27 3 2. 23 5* * 1. 57 5 0. 06 5 24 .3 50 ** * " 1. 01 0 0. 07 6 " 13 .2 95 ** * 0. 64 7 0. 07 4 8. 75 1* ** 0. 01 5 0. 04 7 0. 31 6 0. 90 0 x l k 0. 15 9 0. 21 8 0. 72 8 1. 31 9 0. 05 2 25 .4 62 ** * " 0. 83 3 0. 06 1 " 13 .6 80 ** * 0. 20 7 0. 05 9 3. 49 6* ** " 0. 10 7 0. 03 7 " 2. 86 8* ** 0. 89 9 pp p 0. 22 2 0. 21 4 1. 03 9 0. 55 5 0. 05 1 10 .9 42 ** * 0. 40 5 0. 06 0 6. 78 8* ** " 0. 07 6 0. 05 8 " 1. 31 6 0. 05 4 0. 03 7 1. 46 9 0. 54 8 x l p 0. 03 5 0. 23 1 0. 15 3 0. 52 3 0. 05 5 9. 55 8* ** 0. 26 4 0. 06 4 4. 10 2* ** " 0. 21 9 0. 06 3 " 3. 50 4* ** 0. 05 0 0. 03 9 1. 27 7 0. 46 9 u u u 0. 15 8 0. 27 9 0. 56 5 0. 83 4 0. 06 6 12 .6 28 ** * " 0. 20 9 0. 07 8 " 2. 69 2* ** " 0. 10 7 0. 07 6 " 1. 41 7 0. 05 6 0. 04 8 1. 17 0 0. 57 8 x l u 0. 11 4 0. 31 3 0. 36 5 0. 59 5 0. 07 4 8. 02 9* ** 0. 38 2 0. 08 7 4. 38 6* ** " 0. 20 8 0. 08 5 " 2. 45 0* * 0. 08 1 0. 05 3 1. 51 4 0. 38 7 v v v 0. 14 5 0. 27 3 0. 53 0 0. 63 8 0. 06 5 9. 85 5* ** 0. 09 5 0. 07 6 1. 24 3 " 0. 01 5 0. 07 4 " 0. 20 7 0. 09 4 0. 04 7 2. 02 3* * 0. 41 5 x l v 0. 16 3 0. 22 8 0. 71 6 0. 71 2 0. 05 4 13 .1 81 ** * " 0. 06 5 0. 06 3 " 1. 03 0 " 0. 11 3 0. 06 2 " 1. 82 9 " 0. 02 3 0. 03 9 " 0. 59 2 0. 61 2 y y y " 0. 17 6 0. 18 7 " 0. 94 1 0. 88 9 0. 04 4 19 .9 86 ** * 0. 31 6 0. 05 2 6. 05 3* ** 0. 24 8 0. 05 1 4. 87 3* ** " 0. 02 2 0. 03 2 " 0. 69 5 0. 79 9 x l y 0. 06 3 0. 24 2 0. 26 2 1. 02 4 0. 05 7 17 .8 48 ** * 0. 28 0 0. 06 7 4. 15 0* ** 0. 16 8 0. 06 6 2. 56 8* * " 0. 10 6 0. 04 1 " 2. 56 9* * 0. 76 9 n ot e: b b b ,e e e ,f ff ,i ii ,k k k ,p pp ,u u u ,v v v ,a nd y y y re pr es en t th e fi de lit y se le ct po rt fo lio m ut ua l fu nd s fo r th e m at er ia ls se ct or ,t he e ne rg y se ct or , th e fi na nc ia l se ct or , th e in du st ri al s se ct or , th e t ec hn ol og y se ct or , th e c on su m er st ap le s se ct or , th e u til iti es se ct or , th e h ea lth ca re se ct or , an d th e c on su m er d is cr et io na ry se ct or ,r es pe ct iv el y. x l b ,x l e ,x l f, x l i, x l k ,x l p, x l u ,x l v ,a nd x l y re pr es en t th e se le ct se ct or sp d r e t fs fo r th es e ni ne se ct or s, re sp ec tiv el y. sp y re pr es en ts th e sp d r s& p 50 0 t ru st . ** *s ig ni fic an t at th e 1% le ve l. ** si gn ifi ca nt at th e 5% le ve l. 263c.y. lin / financial services review 23 (2014) 249–271 model using peer etf, respectively. on average, peer etf benchmarking provides the best model fit for fund returns. this result is similar with dellva, demaskey, and smith (2001) and kaushik, pennathur, and barnhart (2010). table 8 one factor model results using peer etf excess return, january 1999 through december 2010 fund ! % adjusted r2 estimate se t-value estimate se t-value bbb 0.506 0.196 2.588** 0.943 0.029 32.566*** 0.881 eee 0.167 0.167 0.998 1.095 0.026 42.875*** 0.927 fff 0.113 0.137 0.820 0.872 0.020 43.574*** 0.929 iii 0.446 0.158 2.815*** 1.004 0.027 36.969*** 0.909 kkk 0.689 0.306 2.249** 1.170 0.038 30.712*** 0.868 ppp 0.332 0.175 1.895 0.831 0.047 17.529*** 0.681 uuu "0.064 0.316 "0.204 0.710 0.068 10.480*** 0.432 vvv 0.166 0.269 0.618 0.623 0.063 9.901*** 0.404 yyy "0.095 0.160 "0.595 0.764 0.027 28.011*** 0.845 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. ***significant at the 1% level. **significant at the 5% level. table 9 adjusted r2 comparison for different models, january 1999 through december 2010 fund 1-factor model (spy) 3-factor model (spy) 4-factor model (spy) 1-factor model (peer etf) bbb 0.590 0.656 0.655 0.881 xlb 0.620 0.672 0.670 eee 0.363 0.379 0.389 0.927 xle 0.352 0.378 0.394 fff 0.644 0.783 0.786 0.929 xlf 0.638 0.796 0.802 iii 0.789 0.856 0.855 0.909 xli 0.778 0.820 0.821 kkk 0.617 0.901 0.900 0.868 xlk 0.715 0.894 0.899 ppp 0.363 0.544 0.548 0.681 xlp 0.315 0.467 0.469 uuu 0.557 0.577 0.578 0.432 xlu 0.240 0.381 0.387 vvv 0.406 0.402 0.415 0.404 xlv 0.609 0.614 0.612 yyy 0.737 0.800 0.799 0.845 xly 0.728 0.760 0.769 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. xlb, xle, xlf, xli, xlk, xlp, xlu, xlv, and xly represent the select sector spdr etfs for these nine sectors, respectively. spy represents the spdr s&p 500 trust. 264 c.y. lin / financial services review 23 (2014) 249–271 comparing the adjusted r2 of one-factor model with peer etf to that of one-factor model with spy, i find that seven out of nine mf regressions using peer etf have a higher explanation power for the full and first half sample periods with the exceptions of the utilities sector and the healthcare sector. all nine mf regressions improve their adjusted r2 for the second half sample period using peer etf benchmark. unlike previous research on sector mutual fund performance, this article is one of the first that provides detailed analysis on individual sector funds. zheng and tower (2005), for example, examine performance of asset-weighted and equal-weighted fidelity sector fund portfolios and find that they performed less well than corresponding indexes. dellva, demaskey, and smith (2001) only list number of sector funds in their analysis. kaushik, pennathur, and barnhart (2010) report sector aggregate performance results. for individual investors and/or their financial planners, however, it is important to examine individual sector fund performance and allocate their assets accordingly. 4.4. changing dynamic do mf and etf in the same sector move together? are they highly correlated with the general market index? pairwise correlation is charted in fig. 3. in all but two cases (utilities and healthcare) the correlation between mf and etf in the same sector, the black bar, is the highest compared with the correlation between mf/etf and spy. six sectors have a correlation between mf and etf higher than 0.9 (materials, energy, financial, industrials, technology, and consumer discretionary), whereas the other three sectors have a correlation of 0.827 (consumer staples), 0.661 (utilities), and 0.638 (healthcare). an interesting find was that these three sectors are often considered defensive sectors. common factors in one sector appear to affect the returns of the mf and etf in this sector more than that of general factors affecting the overall equity market. the returns of mfs and etfs in the same sector tend to go hand-in-hand. the regression results during different sample periods discussed earlier, however, hint that 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 b e f i k p u v y mf vs. etf mf vs. spy etf vs. spy fig. 3. monthly return correlation between funds, 1999–2010. b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. 265c.y. lin / financial services review 23 (2014) 249–271 the correlation might not be stable throughout the whole sample period. fig. 4 plots 36-month rolling correlation for the nine sectors. consistent with the results in fig. 3, for most of the sample period, in all but two sectors (utilities and healthcare), the correlation between mf and etf in the same sector, the black line, is the highest compared with the correlation between mf/etf and spy. the industrial sector, the technology sector, and the consumer discretionary sector have the most consistent correlation among three pairs of correlation plotted, mf versus etf, mf versus spy, and etf versus spy. the three lines are close to each other with the black one on the top. however, the utilities sector and the healthcare sector show great time varying correlation. the lines cross each other and the spread is huge. for example, for the healthcare sector, the correlation between the etf and spy was the highest (0.864 vs. 0.127 and 0.189) for the first 36 months, january 1999 through december 2001, but it turns out to be the lowest (0.817 vs. 0.931 and 0.869) for the last 36 months, january 2008 through december 2010. a noticeable phenomenon is that the correlations tend to converge overtime. almost all sectors have tighter correlation spreads moving into the end of the sample period. the correlation between the mf/etf and spy, the gray dashed line and the black dotted line, almost overlap for most of the sectors during the last quarter of the chart period. this changing dynamic explains why regression based results are sensitive to sample period selection. the open question is whether the correlation convergence will continue into the future. if yes, the diversification benefit one can enjoy through sector investing may diminish overtime. 5. discussion 5.1. benchmarking the fidelity select portfolios cited both the s&p 500 and a msci u.s. im sector 25/50 index in the “management’s discussion of fund performance” section of the annual reports. however, the earliest etfs based on the msci u.s. im sector 25/50 indexes were launched in january 2004 by vanguard, which is five years later than the select sector spdr funds. the price correlation between these two etfs in the same sector ranges from 0.975 to 0.999, with seven out of nine above 0.99, for the period from 2004 through 2010. this high correlation justifies the use of the select sector spdr funds as benchmarks for performance evaluation. one obvious explanation of the results that sector mfs outperform their peer etfs somehow is that these actively managed mutual funds can invest outside of the s&p 500 basket. they can invest in foreign issuers as the prospectuses of the fidelity select portfolios state. domestically they can invest in any stock that is not in the s&p 500 index, mainly smaller capitalization stocks. they also only “normally … invest at least 80% of assets in securities of companies principally engaged in the selected sector,” which gives them some wiggle room across sector borders. however, it is hard to imagine they deviate very far from their benchmark. only eight out 266 c.y. lin / financial services review 23 (2014) 249–271 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m a y -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 materials 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 energy 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m a y -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 financial 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m a y -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 industrials 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 technology 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 consumer staples 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 utilities 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 healthcare 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m a y -1 0 o ct -1 0 consumer discretionary mf vs. etf mf vs. spy etf vs. spy fig. 4. thirty-six months rolling correlation between funds. 267c.y. lin / financial services review 23 (2014) 249–271 of 90 stocks of the top 10 holdings of the fidelity select portfolios at the end of september 2012 are not in the s&p 500 index.14 half of that eight are foreign stocks. seven out of the eight has a weight between 1.59% and 4.05% of corresponding sector mfs; the other one weights 13.66%.15 the top weighted index stocks also make the backbone of the sector mutual funds. the number of stocks in the top 10 holdings of both sector mfs and etfs varies from three to seven. seven out of top 10 holdings of the materials, industrials, and healthcare mf are also in the etf top 10 list, and the top 10 holdings make up between 46% and 64% of these three mutual funds. that number is six for the consumer staples and consumer discretionary mf, five for the energy and technology mf, four for the utilities mf, and three for the financial mf. the top 10 holdings make up between 29% and 68% of the funds in these six sectors. 5.2. fidelity fidelity had been the largest mutual fund family that mainly provides actively managed mutual funds for several decades. pozen and hamacher (2011) argue that fidelity, among several other fund families, has two characteristics that contribute to its success in the u.s. mutual fund business: dedication primary to asset management and control by investment professionals. the fidelity fund family has maintained top market shares in the past decade: 10.2%, 11.8%, and 11.3% in year 1990, 1995, and 2010, respectively. it was no. 1 in 1990 and 1995, but passed by vanguard (12.1%) in 2010. fidelity’s stock is effectively controlled by members of its funding family, which relieves the short-term performance pressure from public shareholders to increase quarterly earnings. they also stated that fidelity can develop compensation programs that promote top performance. fidelity’s megellan fund has long been used as an evidence to defy emh (see kochman and badarinathi 1993 and marcus 1990). fidelity’s large size can also potentially gain an insider edge as golec (2007) shows some evidence on informed trades made by fidelity funds. the large size of fidelity’s assets under management can provide economy of scale for securities research, which aids the key element of active management: security selecting. unlike passive managers, who often do little on stock selection, active managers can beat their benchmark by overweighting future winners, underweighting future losers, or some combination of both. they also have the freedom of holding cash. focusing on only one sector, sector fund managers can potentially gain growing knowledge and experience dealing with stocks in that sector, which could help their performance. elton, gruber, and green (2007) explain why funds may be more similar inside than outside fund families: portfolio managers within families are likely to have access to the same research analysis produced either by internal analysts or by a particular set of external research firms; portfolio managers may begin the security selection process with an economic forecast that is shared by other fund managers within the firm. it is not a surprise that fidelity sector funds as a group can outperform when the shared research and macro view work well. 268 c.y. lin / financial services review 23 (2014) 249–271 6. conclusion this article is one of the first that reports detailed analysis on individual equity sector fund performance. it also contributes to the literature by adding evidence of active equity management outperformance. i find considerable evidence that sector mutual funds, the nine fidelity select portfolios here, have provided better after-expense returns against broader market etf, spy, and their peer sector etfs, the nine select sector spdr funds, during the sample period 1999–2010. not only do they achieve higher nominal returns over the 12-year period except for few sector mfs, some of the funds also generate higher riskadjusted returns measured by sharpe ratio16 and ! from various asset pricing models. none of the sector mfs generates a significant negative ! for the full sample period no matter which asset pricing model is used. that is, the sector mfs do not underperform spy or peer etfs measured by !. the materials and the technology sector mf stand out in the analysis. the materials mf beats its peer etf and spy across the board for the full sample period: second highest 12-year return, highest sharpe ratio, and significant positive !s in all four regression models. the annualized abnormal return of the materials sector mf is 11.856%, 8.628%, and 8.472% against spy and 6.072% against xlb using factor models. the technology mf also beats both spy and peer etfs on 12-year return, sharpe ratio, and most regression-based measures. it generates annualized abnormal return of 7.368% and 7.320% against spy using the three-factor and four-factor models, and 8.268% against xlk. the energy mf outperforms on 12-year return, sharpe ratio, and one-factor model against spy with an annualized abnormal return of 12.768%. the industrials mf outperforms on 12-year return, sharpe ratio, and one-factor model against xli with an annualized abnormal return of 5.352%. the utilities and consumer discretionary mf have the weakest results. they do not beat their peer etfs measured by 12-year return and sharpe ratio, however, they do not underperform when asset pricing models are adopted. many researches have showed that on average active managers do not add value after fees and expenses (e.g., barras, scaillet, and wermers 2010). some argue that the value of active management lies in making market more efficient by improving asset allocation (jones and wermers 2011). xiong et al. (2010) document that both asset allocation and active management are critical to performance. the need for active managers to set the price was emphasized by a large index fund manager: “passive management is a free-ride strategy; it piggybacks on active management. you need to have active managers out there, and they need to be paid.”17 this article, however, presents evidence on outperformance of equity sector funds. if one considers active managers together are playing a zero-sum game, this study finds some of the winners. for individual investors who are interested in sector investing, this study shows that actively managed mutual funds can be a good candidate for sector allocation. most of these mutual funds do a better or equivalent job measured by both nominal return and total/ systematic risk adjusted return, after higher expenses and fees. the bottom line is that they do not underperform when measured with ! from asset pricing models, at least during the sample period. this study also provides evidence for financial planners when they help their 269c.y. lin / financial services review 23 (2014) 249–271 clients select mutual funds. it may be appropriate for financial planners to recommend sector mutual funds to their clients who have risk appetite for sector investing. notes 1 most etfs, although not all, are passively managed to track a specific index, such as the s&p 500. 2 https://www.fidelity.com/. 3 because telecommunication stocks are covered in the technology select sector spdr fund, i do not consider telecommunication a separate sector. 4 fidelity select portfolio annual report, february 29, 2012. 5 fidelity select portfolio prospectuses. 6 however, all nine select sector spdrs are diversified funds with respect to the internal revenue code. as a result, each sector index is modified so that an individual security does not comprise more than 25% of the index. source: http:// www.sectorspdr.com/. 7 select sector spdrs annual report, september 30, 2011. 8 spdr s&p 500 etf trust annual report, september 30, 2011. 9 the available price data for fidelity select industrials portfolio starts on july 8, 1999. 10 http://finance.yahoo.com/. 11 subsample period results are discussed but not reported throughout the article. results are available upon request. 12 for example, ross, westerfield, and jordan, corporate finance, 9th ed., 2010, mcgraw-hill. 13 http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html. 14 from fidelity and sector spider web sites. 15 british american tobacco plc adr represents 13.66% of the fidelity select consumer staples portfolio as of september 28, 2012. 16 eling (2008) shows that choosing a performance measure is not critical to fund evaluation and the sharpe ratio is generally adequate. 17 frank j. fabozzi, sergio m. focardi, and caroline jonas, 2010, investment management after global financial crisis, research foundation publications, cfa institute, page 26. references barras, l., scaillet, o., & wermers, r. (2010). false discoveries in mutual fund performance: measuring luck in estimated alphas. journal of finance, 65, 179–216. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. dellva, w. l., demaskey, a. l. & smith, c. a. (2001). selectivity and market timing performance of fidelity sector mutual funds. financial review, 36, 39–54. eling, m. (2008). does the measure matter in the mutual fund industry? financial analysts journal, 64, 54–66. elton, e. j., gruber, m. j. green, c. (2007). the impact of mutual fund family membership on investor risk. journal of financial and quantitative analysis, 42, 257–277. 270 c.y. lin / financial services review 23 (2014) 249–271 fabozzi, f. j., focardi, s. m. & jonas, c. (2010). investment management after the global financial crisis. charlottesville, va: research foundation publications, cfa institute. fama, e. f., & french, k. r. (1993). common risk factors in the returns on stocks and bonds. journal of financial economics, 33, 3–53. golec, j. (2007). are the insider trades of a large institutional investor informed? financial review, 42, 161–190. groysberg, b., healy, p., nohria, n., & serafeim, g. (2011). what factors drive analyst forecasts? financial analysts journal, 67, 18–29. jones, r. c., & wermers, r. (2011). active management in mostly efficient markets. financial analysts journal, 67, 29–45. kacperczyk, m., sialm, c. & zheng, l. (2005). on the industry concentration of actively managed equity mutual funds. journal of finance, 60, 1983–2011. kaushik, a., pennathur, a. & barnhart, s. (2010). market timing and the determinants of performance of sector funds over the business cycle. managerial finance, 36, 583–602. kochman, l. m., & badarinathi, r. (1993). net selectivity revisited. journal of economics and finance, 17, 73–79. marcus, a. j. (1990). the magellan fund and market efficiency. journal of portfolio management, 17, 85–88. mcgahan, a. m. & porter, m. e. (1997). how much does industry matter, really? strategic management journal, 18, 15–30. porter, m. e. (1985). competitive advantage: creating and sustaining superior performance. new york, ny: free press. pozen, r., & hamacher, t. (2011). most likely to succeed: leadership in the fund industry. financial analysts journal, 67, 21–28. sullivan, r. n., & xiong, j. x. (2012). how index trading increases market vulnerability. financial analysts journal, 68, 70–84. xiong, j. x., ibbotson, r. g., idzorek, t. m., & chen, p. (2010). the equal importance of asset allocation and active management. financial analysts journal, 66, 22–30. zheng, w., & tower, e. (2005). fidelity versus vanguard: comparing the performance of the two largest mutual fund families. international review of economics and business, 52, 433–465. 271c.y. lin / financial services review 23 (2014) 249–271 finser_31-2_complete_issue afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires december 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in davis et al., which of the following statements is true based on the article’s findings: (select one best option) a. students with high measured financial literacy and low self-assessed financial literacy are the most interested in financial education. b. students with high measured financial literacy and high self-assessed financial literacy are the most interested in financial education. c. students with low measured financial literacy and low self-assessed financial literacy are the most interested in financial education. d. students with low measured financial literacy and high self-assessed financial literacy are the most interested in financial education. 2. authors augustin and martin found this type of exposure to financial literacy has the greatest impact on positive financial behaviors? a. high school exposure b. college exposure c. seminar exposure d. work exposure 3. according to the findings in financial planning time horizon and end-of-life mortality expectations, everything else equal, people who have more wealth are more likely to a. plan their finances relatively longer into the future. b. not engage in any financial planning. c. plan with a focus limited to the next six months. d. rent rather than own their homes. 4. according to the article “the changing assessment of risk for young investors”, which of the following are drivers of risk tolerance for gen z? a. willingness to take risk, self-reported investment risk tolerance, and ownership of investment accounts. b. willingness to take risk and selfreported investment risk tolerance. c. ownership of investment accounts. d. none of the above. 5. loan modification options depend on current interest rates and changes in home values. as discussed by mcclatchey, a refinance is most valuable when interest rates have ________ and home values have ________. a. decreased; increased b. decreased; decreased c. increased; increased d. increased; decreased ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: terrance k. martin, e-mail: martintk@wssu.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services terrance k. martin college of arts, sciences, business, and education winston-salem state university reynolds center, rm 111 601 s. martin luther king jr. drive winston-salem, nc 27110 (address service requested) finser_23_1 downside risk: what the consumer sentiment index reveals mark a. johnsona,*, atsuyuki nakab adepartment of finance, the sellinger school of business and management, loyola university maryland, 4501 north charles street, baltimore, md 21210, usa bdepartment of economics and finance, college of business administration, university of new orleans, 2000 lakeshore drive, new orleans, la 70148, usa abstract this article examines the ability of consumer sentiment for different age groups to forecast short-term as well as long-term equity returns. using a long-horizon asymmetric response regression format, we show that negative changes in sentiment have a greater influence on stock returns than positive changes in sentiment. our findings are supportive of the prospect theory. however, we observe that younger individuals appear to be less risk-averse than older individuals. we provide evidence that reminds individual investors and financial planners that risk is an important consideration when investing, and that demographic characteristics matter when determining appropriate investing approaches and risk tolerance. © 2014 academy of financial services. all rights reserved. jel classifications: g10; d03; d12 keywords: consumer sentiment; downside risk; asymmetric response; long-horizon regression 1. introduction consumer sentiment is based on factors such as unemployment rates, wages, expected inflation, interest rates, and other prevailing market circumstances. research indicates that consumer sentiment concerning current and future economic conditions can affect the outcomes of financial assets.1 further clarification of the dynamics of this relationship will improve the understanding of how the sentiment affects equity returns. the importance of * corresponding author. tel.: !1-410-617-2473 fax: !1-410-617-5035 e-mail address: majohnson@loyola.edu (m.a. johnson) financial services review 23 (2014) 45–61 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. this relationship has the potential to aid investors, especially individual investors, and financial planners as they seek to understand and identify useful economic indicators that can potentially help forecast stock returns. consumer sentiment surveys generally ask individuals how they feel about their current economic situation and how they perceive their future economic situation. by incorporating such data into economic models, we ask the following questions: assuming consumer sentiment affects equity returns, how long of a holding period can be predicted? does negative sentiment have a larger effect on equity returns than positive sentiment? do the sentiments of different age groups have the same or similar effects on equity returns? do different sizes of firms react to the sentiment equally? by answering these questions, we can deepen our understanding of the relationships between consumer behavior and stock returns in the united states. furthermore, because household consumption is such a significant portion of gross domestic product (gdp) in the united states, understanding the outlook of consumers can impact economic growth and corporate profitability. to address these questions, we utilize the asymmetric response model with long-horizon regressions. to test asymmetric responses with respect to consumer sentiment, the consumer sentiment index compiled by the university of michigan is divided into positive changes and negative changes, where negative consumer sentiment indicates pessimistic feelings about future economic prospects, or aggravation of losing wealth and positive sentiment indicates the opposite. according to the prospect theory developed by kahneman and tversky (1979), losses matter more to individuals than gains because of the risk averse nature of investors, and is closely related to the concept of downside risk. this specific type of risk is critical for individual investors and financial planners to monitor, given the findings of the prospect theory. furthermore, downside risk has been studied in other scenarios (e.g., when is the ideal time for an individual to receive social security benefits (friedman and phillips, 2010). the current article seeks to add to the downside risk literature by looking at individuals through the lenses of consumer sentiment and how changes in sentiment may foretell forthcoming equity returns. we look at 1, 3, 6, 12, and 24 month holding period horizons to investigate whether changes in consumer sentiment have the ability to forecast equity returns. if changes in consumer sentiment have this predictive ability, this can assist individual investors and financial planners, regardless of whether their investment horizon is short-term or long-term. the sentiments of three different age groups are investigated to address the idea of life cycle investment hypothesis.2 existing research has shown that an aging population results in higher average risk aversion and subsequently, higher risk premiums. for example, an individuals’ age and their appetite (or tolerance) for risk has been studied by hariharan et al. (2000), gibson et al. (2013), larkin et al. (2013), and schooley and worden (1996). the current article contributes to the literature by presenting comprehensive analysis of the effects of consumer sentiment on the u.s. stock returns that includes data from the recent financial crisis. by including the recent financial crisis in our sample period, we are able to provide results that include a time period that has been referred to as an once-in-a-lifetime event. fisher and statman (2003) successfully show that there is a relationship between consumer sentiment and stock returns. specifically, they find evidence that changes in consumer sentiment and contemporaneous stock returns have a positive, statistically signif46 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 icant relationship. additionally, they present that the relationship between changes in consumer sentiment and the returns of small-cap stocks is stronger than the relationship between sentiment and a broad market index such as the s&p 500.3 baker and wurgler (2006) study the relationship between cross-sectional differences of stock returns and investor sentiment by constructing the investor sentiment index based on six market variables, and categorizing sentiment as either optimism or pessimism about stocks markets.4 they show that investor sentiment has a larger effect on hard-to-price securities such as small stocks, young stocks, high volatility stocks, unprofitable stocks, non-dividend paying stocks, growth stocks, and distressed stocks. akhtar et al. (2011) and qui and welch (2006) and show that consumer sentiment is a good proxy for investor sentiment. lemmon and portniaguina (2006) present a way of capturing pessimism and optimism by using a consumer sentiment index (csi) as a function of a large number of macroeconomic variables such as inflation, the default spread, changes in personal consumption expenditures, gross domestic product, and the unemployment rate. they argue that any consumer sentiment based on fundamental economic conditions is reasonable, justifiable, and rational. on the other hand, consumer sentiment based on factors other than economic conditions is interpreted as unjustifiable and irrational, where they define the residual from the regression as a proxy for unjustifiable sentiment. their results support the noise trader hypothesis that sentiment will have a greater effect on asset returns held by individuals. ho and hung (2009) examine the performance of asset pricing models by using sentiment as conditioning information as it reflects investors’ expectations about future forecasts of financial markets. by including various sentiment measures, they show that the capital asset pricing models perform better when sentiment is included, and conclude that sentiment helps explain market anomalies such as momentum, liquidity, size, and book-to-market effects. schmeling (2009) utilizes consumer confidence as a proxy for individual investor sentiment to investigate whether lagged sentiment explains stock returns for eighteen industrialized countries as examined through a long-run horizon regression model. the author shows that international investor sentiment predicts future aggregate market returns, and the impact of sentiment on returns is stronger for countries that have less developed markets since they are more prone to investor overreaction. the results indicate a statistically significant coefficient for the sentiment variable, and this significance holds for various forecast horizons. schmeling finds that lagged sentiment has a stronger effect on stock returns in countries such as germany, japan, italy, and the united states, but little or no evidence of such a relationship in countries such as the united kingdom, australia, and new zealand. further, akhtar et al. (2011) study the reaction of the australian stock market by using the asymmetric response model, and show that negative changes in the australian csi have statistically significant effects on market returns, whereas positive changes do not. other recent studies such as chung et al. (2012), stambaugh et al. (2012), and yu and yuan (2011) look at sentiment’s impact on stock returns as well. yu and yuan (2011) show the influence of sentiment on the market’s mean–variance tradeoff using the baker and wurgler (2006) sentiment index. stambaugh et al. (2012) also use the baker and wurgler (2006) sentiment index and test sentiment’s role when accounting for documented anomalies in investing strategies such as a long-short strategy. both of these studies show evidence related to the stock return predictive ability of sentiment. chung et al. (2012) incorporates 47m.a. johnson, a. naka / financial services review 23 (2014) 45–61 business cycles (i.e., economic expansions and recessions) into their modeling to show that during recessions, the predictive power of sentiment is not nearly as strong and can possibly be insignificant. our benchmark model provides evidence that sentiment matters more in forecasting near-term market risk premiums versus farther out time horizons. by including various time horizons of equity holding-period returns into consideration, we find that sentiment has a greater ability to explain market risk premiums and forecast equity returns in the short-run than in the long-run. as a result, individual investors and financial planners that track consumer sentiment as an economic indicator used in investing should exercise caution when using sentiment as an indicator to forecast equity returns too far into the future. we show that negative changes in sentiment matters more in explaining market risk premiums than positive changes in sentiment based on asymmetric response models that account for unequal responses to negative events and positive events. this finding supports the prospect theory developed by kahneman and tversky (1979) as to possible losses mattering more to individuals than possible gains. this can benefit individual investors and financial planners because if consumer sentiment begins to decrease (increase) because of macroeconomic conditions, more (less) credence can be placed on consumer sentiment as a potential forecaster of future equity returns. however, we observe that younger individuals appear to be less risk-averse than older individuals. finally, the current article examines the noise trader hypothesis by using 10 size-sorted portfolio returns, and finds that smaller market cap stocks are more impacted by changes in consumer sentiment, a result consistent with the previous literature. our final finding further reminds individual investors and financial planners that small-cap stocks have the potential to be riskier than large-cap stocks. 2. methodology and data 2.1. empirical models to test if consumer sentiment has a near-term or long-term impact on the stock market, we utilize long-horizon regressions. by testing different future time periods (horizons), we investigate whether or not changes in csi have such future explanatory power. long-horizon financial models are popular and have been used in prior studies. for example, rich and reichenstein (1993) use a type of long-horizon model to see if individual investors can time the stock market. our benchmark long-horizon regression model is given as:5 ri,t!1 ! . . . ! ri,t!k " #"k# ! $"k#$csit ! %t!k, k " 1, 3, 6, 12, 24, (1) where ri,t!k is the excess return of equity index i over k month periods, and excess returns are defined as the equity returns minus the risk free rate. the risk free rate is the monthly yield on a 30 day treasury bill. $csit is the monthly percentage change in the csi, that is, $csit % (csit!1 & csit)/csit and %t!k is the residual term for horizon k. a direct test of downside risk within the context of the prospect theory would be to 48 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 observe how negative or positive changes in consumer sentiment alter the stock returns. harlow and rao (1989) provide a simple way of measuring asymmetric responses to account for individuals exhibiting downside risk. in the context of this article, the asymmetric response model is one way of blending the prospect theory and downside risk based on the documented fact that responses to losses and gains are not the same. by using an asymmetric response model, we isolate improvements and deteriorations in sentiment and allow for interpretations of how changes in csi identify downside risk for different holding period returns. the following long-horizon asymmetric response regression model is estimated: ri,t!1 ! . . . ! ri,t!k " #"k# ! $&"k#$csit & ! $!"k#$csit ! ! %t!k, k " 1, 3, 6, 12, 24, "2# where $csit & is the negative change in csi ($csit ' 0) or zero otherwise, and $csit ! is the positive change in csi ($csit ( 0) or zero otherwise.6 the coefficients $& and $! indicate the downsize $ and upside $, respectively. if stock returns respond similarly to positive changes in consumer sentiment as they do to negative changes in consumer sentiment, $& would be equal to $!. however, if investors are averse to downside risk, $& will be positive, representing a positive risk premium (e.g., the higher the downside risk, the higher the stock returns). this effect is a type of risk that investors tend to be more sensitive towards when discussing the potential for loss in the value of an asset even though many investors may anticipate a particular asset’s value to increase over a long time. the estimates of $! will also be smaller in magnitude and positive if individuals prefer upside potential (e.g., the higher the upside potential, the lower the stock returns). 2.2. data description the primary sources of consumer sentiment data in the united states are the university of michigan’s surveys of consumers (csi) and the conference board’s consumer confidence index. fisher and statman (2003) and lemmon and portniaguina (2006) confirm that the indices are highly correlated and provide similar empirical results, despite survey design differences. ludvigson (2004) recognizes that many studies use csi. studies that utilize csi in relation to stock returns include fisher and statman (2003), ho and hung (2009), lemmon and portniaguina (2006), and schmeling (2009). in this study, we use monthly csi data from january 1978 to december 2010, encompassing more than 30 years of sentiment observations. we segment the data into the widely cited composite index as well as indices for three age groups, respectively: persons 18 to 34 years old; persons 35 to 54 years old; and persons 55 years old and older. this partition of age groups within the csi allows for a demographic investigation that has previously been unexplored regarding how stock markets can be understood in relation to the sentiment of persons of different ages. the center for research in security prices (crsp) equally weighted returns (crsp ew) and crsp value-weighted returns (crsp vw) are used to capture the overall stock market performance. we use 30 day u.s treasury bill yields to 49m.a. johnson, a. naka / financial services review 23 (2014) 45–61 represent risk-free rates to estimate the risk free returns. these data sets are obtained from wharton research data services (wrds). table 1 presents the summary statistics of the variables used in this study. panel a shows all of the csi data for the composite index as well as for the three age groups. the mean sentiment value for the youngest age group, 18 to 34 years old, is the highest among all age groups (95.169) and the mean value for the oldest age group, persons 55 years old and older, is the lowest among all age groups (78.297). to test whether the mean sentiment values across age groups are the same, we test the null hypothesis that csiage group 18–34 % csiage group 35–54 % csiage group 55 and older. on the other hand, the alternative hypothesis is that at least one of the mean age group sentiment values is different from the others. when tested, we find an f-statistic of 160.257 (p % 0.000) and reject the null hypothesis. this finding implies that differences exist in the various age groups’ mean sentiment values. the results suggest that over the sample period, younger consumers tend to be more optimistic than older consumers and this is consistent with the life cycle investment hypothesis. one behavioral explanation for this would be that younger individuals have more years of their life to participate in the labor force and earn money, resulting in hopeful current or future saving and consumption whereas older individuals have fewer years of their life to participate in the labor force. at the same time, families with children see their children enter adulthood and as a result, they may not foresee increased consumption. with respect to volatility of the csi index for the age groups, the 35 to 54 years old age group has the largest standard deviation (14.413) and the oldest age group has the lowest volatility (11.791). and although the percentage changes of csi are similar across different groups and close to zero on average, there is a large range between the minimum and the maximum of percentage changes of csi.7 the corresponding standard deviations are relatively large and are somewhat similar to those for the stock returns. according to ando and modigliani (1963), the desire or propensity to consume and invest is higher in the lives of younger people, whereas middle to older-aged individuals tends to have higher incomes with lower propensities to consume. those in the younger age group are more likely to be in the early professional years of their careers, attempting to acquire more permanent assets, possibly preparing to start a family. panel b presents the excess returns of the crsp ew and crsp vw indices over five horizons. for both indices, as the number of horizons increases, the mean returns increase as well as the sds. panel c presents the returns data for the crsp market capitalization of 10 size-sorted portfolios. a brief look at panel c provides data on the well-documented fact that small cap stocks have higher returns than large cap stocks but are also riskier, as evident by the standard deviations of the two portfolios. the mean returns for size-sorted portfolios deciles 3 through decile 8 appear the same because they represent the mid cap stock segment. as a result, these six portfolios have similar risk-return characteristics. we use all 10 size-sorted portfolios to see if the size of the firm matters in terms of the short-run or long-run effects of sentiment. akhtar et al. (2011), baker and wurgler (2006), lemmon and portniaguina (2006), schmeling (2009), and others find that sentiment does impact firms of various sizes differently. fig. 1 displays the time series movements of csi for different groups over the sample period. the major declines in csi over the sample period occur in the late 1970s, early 50 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 1990s, early 2000s, and in 2007 and 2008 during the recent financial crisis. of interestingly to the authors, these periods correspond closely with the dates of recessions in the united states as defined by the national bureau of economic research (nber). for example, table 1 summary statistics mean sd minimum maximum panel a: csi and percent changes csi composite 86.120 13.096 51.700 112.000 csi age group 18–34 95.169 13.456 60.400 120.000 csi age group 35–54 86.409 14.413 43.600 113.300 csi age group 55 and older 78.297 11.791 43.400 105.300 $csicomposite 0.001 0.050 &0.181 0.246 $csi18–34 age group 0.002 0.061 &0.174 0.287 $csi35–54 age group 0.002 0.066 &0.231 0.241 $csi55 and older age group 0.002 0.069 &0.196 0.339 $csicomposite ' 0 &0.018 0.029 &0.181 0.000 $csicomposite ( 0 0.019 0.031 0.000 0.246 $csi18–34 age group ' 0 &0.022 0.033 &0.174 0.000 $csi18–34 age group ( 0 0.023 0.040 0.000 0.287 $csi35–54 age group ' 0 &0.024 0.037 &0.231 0.000 $csi35–54 age group ( 0 0.026 0.042 0.000 0.294 $csi55 and older age group ' 0 &0.025 0.039 &0.196 0.000 $csi55 and older age group ( 0 0.027 0.043 0.000 0.339 panel b: stock index returns for different horizons crsp ew index returns k % 1 0.011 0.056 &0.273 0.224 k % 3 0.033 0.111 &0.470 0.417 k % 6 0.065 0.158 &0.565 0.588 k % 12 0.128 0.212 &0.613 0.736 k % 24 0.245 0.242 &0.751 0.796 crsp vw index returns k % 1 0.008 0.046 &0.227 0.127 k % 3 0.024 0.083 &0.374 0.258 k % 6 0.046 0.118 &0.539 0.364 k % 12 0.092 0.168 &0.565 0.485 k % 24 0.179 0.222 &0.599 0.542 risk-free rate (30 day t-bill) 0.004 0.003 0.000 0.014 panel c: returns of equity size portfolios capitalization decile 1 0.014 0.064 &0.277 0.329 capitalization decile 2 0.011 0.065 &0.303 0.285 capitalization decile 3 0.012 0.063 &0.289 0.261 capitalization decile 4 0.012 0.061 &0.294 0.226 capitalization decile 5 0.012 0.061 &0.281 0.255 capitalization decile 6 0.012 0.056 &0.259 0.222 capitalization decile 7 0.012 0.055 &0.259 0.224 capitalization decile 8 0.012 0.054 &0.241 0.188 capitalization decile 9 0.011 0.051 &0.225 0.225 capitalization decile 10 0.010 0.048 &0.204 0.136 note. all data is monthly and is from january 1978 until december 2010. panel a consumer sentiment data was obtained from the university of michigan surveys of consumers (http://www.sca.isr.umich.edu). panel b and panel c stock index returns and equity size-sorted portfolio returns data were obtained from center for research in security prices (crsp). k stands for monthly cumulated k horizon returns. 51m.a. johnson, a. naka / financial services review 23 (2014) 45–61 recessionary periods in the united states occurred from july 1990 until march 1991, march 2001 until november 2001, and december 2007 until june 2009.8 the most recent decline in csi corresponds with the widely regarded “great recession,” which was sparked by the abrupt contraction in home prices and spike in home foreclosures because of the subprime crisis and other banking problems. the financial crisis ultimately resulted in the sharp increase in the unemployment rate and a rise in the number of nonperforming loans on the balance sheets of financial institutions among other sudden negative economic conditions. consistent with panel a of table 1, fig. 1 shows that the younger consumers (18 to 34 years old) tend to be more optimistic than the older consumers, and this gap is persistent for much of the sample period. the 55 years old and older age group tends to record the lowest sentiment reading throughout the sample period. further, we observe substantial movements of these indices over time, resulting in large sds. this figure also shows how the sentiment of individuals, regardless of age, tends to move in tandem. 3. empirical results 3.1. benchmark case table 2 reports the results of the benchmark regression model from eq. (1). the coefficients for $csit are positive as well as statistically significant at the 1% level for both stock fig. 1. time-series plot of consumer sentiment based on age groups. this figure plots the consumer sentiment indexes of three age groups: between 18 to 34 years old; between 35 and 54 years old; and 55 years old and older. 52 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 market indices used for six month horizons or less. for the 12 month horizon, the coefficients for $csit are significant at the 5% level for the crsp ew index.9 the results are also economically significant. for example, for the three month horizon for the crsp ew (vw) index, a 10% change in overall consumer sentiment (consumers of all ages) results in a market risk premium change of 7.36% (5.34%). it can be interpreted from these findings that a positive change in csi results in positive future excess stock returns for subsequent holding period returns. however, for the 24 month horizon, changes in consumer sentiment are no longer statistically significant in forecasting stock returns. the results indicate that sentiment is more important in the nearer term instead of the distant future that can help individual investors and financial planners as they consider using this sentiment indicator to make equity investment decisions. the long-horizon regressions that schmeling (2009) uses also show that the impact of sentiment on average future returns declines as the forecast horizon increases. furthermore, given that csi data are released monthly, recent data points provide better insights pertaining to consumers and their outlooks regarding the stock returns. another immediate observation present in table 2 is that the values of the slope coefficients are much larger for the crsp ew excess returns compared to the crsp vw excess returns. we attribute the lower excess returns of the vw index $ coefficients to the documented findings of lemmon and portniaguina (2006) and schmeling (2009). they show that sentiment affects the stocks of firms of various sizes differently. in the vw index, large firm stock returns are inherently weighted more heavily than those returns of smaller firms. thus, the effect of changes in sentiment becomes harder to disentangle because of this size effect. the ew index $ coefficients shows that holding firm size constant, changes in csi are still important. we also observe that the values of r2 are the largest for the one month and three month horizons, and they gradually decrease as the horizon increases. as we found in table 2, table 2 benchmark model dependent variable forecast horizon ( k) 1 3 6 12 24 ew returns vw returns ew returns vw returns ew returns vw returns ew returns vw returns ew returns vw returns intercept # 0.006 0.003 0.019 0.010 0.038 0.019 0.073 0.037 0.135 0.069 se (0.003) (0.002) (0.009) (0.006) (0.016) (0.012) (0.024) (0.019) (0.029) (0.026) p value [0.032] [0.151] [0.025] [0.131] [0.017] [0.113] [0.003] [0.053] [0.000] [0.008] slope $ 0.362 0.204 0.736 0.534 0.635 0.425 0.487 0.287 0.338 0.092 se (0.064) (0.054) (0.107) (0.081) (0.142) (0.098) (0.214) (0.161) (0.256) (0.188) p value [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.023] [0.076] [0.187] [0.626] r2 0.104 0.049 0.109 0.103 0.040 0.033 0.013 0.008 0.005 0.000 note. this table represents long-horizon ordinary least square regressions of one month excess returns (ri,t) of various indices on changes in the consumer sentiment index (csi). the dependent variable is the excess return of either the crsp equally weighted portfolio or the crsp value weighted portfolio. k stands for monthly cumulated k horizon returns and when k is greater than one, the regressions use overlapping monthly data. all data is monthly and is from january 1978 until december 2010. statistical significance is determined by the newey-west p values. 53m.a. johnson, a. naka / financial services review 23 (2014) 45–61 nearer term economic data impacts sentiment more, this in turn, is then priced into the stock market shortly thereafter. campbell et al. (1997) presents long-horizon regression results with increasing r2’s for longer horizons. however, valkanov (2003) shows that the interpretation of the r2’s for long-horizon regressions is not straightforward and can be misleading for comparisons of regression results. 3.2. asymmetric response model table 3 presents the findings of the asymmetric response models in eq. (2) for the overall csi composite. we observe that all of these coefficients of the downside $s ($&) are positive and statistically significant, with most at the 1% level of significance, and two coefficients at the 5% level of significance (ew index, 12 and 24 month horizons). the results support that the ideas of a positive risk premium since investors are averse to downside risk and the higher the downside risk, the higher the stock returns. these results are consistent with studies such as ang, chen, and xing (2006) and akhtar et al. (2011), and capture the essence of the prospect theory of kahneman and tversky (1979). referring to the estimates of downside $s, as the forecast horizon increases, so does the magnitude of $&. for example, the values are 0.428 (ew returns) and 0.338 (vw returns) for the one month horizon but 1.319 (ew returns) and 1.463 (vw returns) for the 24 month horizon. this implies that risk premiums that can compensate for downside risk become increasingly important as the time horizon increases. ang et al. (2006) show that cross-sectional stock returns indicate evidence of reflecting a premium for downside risk as measured by negative market returns because individuals place greater emphasis on downside risk and less emphasis on potential gains. further, they find that past downside $ has forecasting ability for future stock returns for most of their cross-sectional sample. turning our attention to upside $s ($!), the coefficients are not equal in magnitude and are consistently smaller compared to the downside $s for all horizons studied. for the five different horizons estimated, many of the $! coefficients are statistically insignificant, and in fact, only three are significant at the 5% level. akhtar et al. (2011) also report the statistical insignificance of the upside $. an interesting observation regarding the $! coefficients is their declining nature over the forecasting horizons. the one month forecast horizon magnitudes are 0.303 (ew returns) and 0.084 (vw returns) respectively, and when compared to the 24 month forecast horizon, the upside $ coefficients are &0.493 (ew returns) and &1.070 (vw returns). thus, the upside $ magnitudes decrease and even turn negative as the time horizons increase. the results imply that as forecast horizons increase, upside potential becomes of less importance, and even more so when compared to downside risk. individuals start to exhibit behavior that is consistent with focusing less on increasing wealth and more on not losing wealth. when we test the null hypothesis of $& % $! for all horizons, we do not reject the null hypothesis for the majority of these tests. the results are because of the statistical insignificance of the estimates of $! and the fact that the denominators of the t-statistics tend to be large relative to the numerator. however, this should not undermine the results of our long-horizon asymmetric regressions that show statistically significant downside $s for all cases, and provides support to the concepts of the prospect theory. akhtar et al. (2011) state 54 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 that negative (positive) consumer sentiment news will induce a negative (zero) stock market reaction. as a result, they hypothesize that the coefficient for the negative change in sentiment would be negative, and their baseline regression results support this hypothesis. because their estimated coefficient on the positive change in sentiment variable is not significant, they argue that the result provides more support for the negativity effect hypothesis. however, if investors are averse to downside risk, the coefficient for downside $ would be positive because of the idea that a higher downside risk would correspond to higher returns. furthermore, we find that as the forecast horizon increases, the magnitudes of the downside $s increase while the magnitudes of the upside $s decrease. this divergence between the general trends of these coefficients shows an increasing importance of downside risk to investors, individual and institutional, over time. we now turn our attention to shedding light on the life cycle investment hypothesis by incorporating the changes in sentiment of different age groups. panel a of table 4 presents the empirical results of eq. (2) for the csi index of individuals 18 to 34 years old. focusing on the downside $ for this youngest age group that we study, we observe that $& has the expected positive sign and is statistically significant at the 1% or 5% level of significance in the majority of the one, three, and six month forecast horizons. however, we do not observe in table 4 an increasing tendency of the downside $ coefficients as we presented in table 3 when the forecast horizon increases, implying that aversion to losses among younger individuals does not necessarily increase over longer periods with respect to excess market table 3 asymmetric response model: csi composite dependent variable forecast horizon (k) 1 3 6 12 24 ew returns vw returns ew returns vw returns ew returns vw returns ew returns vw returns ew returns vw returns intercept #i 0.009 0.008 0.025 0.021 0.048 0.035 0.099 0.073 0.168 0.115 se (0.004) (0.003) (0.009) (0.007) (0.015) (0.012) (0.024) (0.019) (0.027) (0.028) p value [0.041] [0.018] [0.006] [0.001] [0.002] [0.004] [0.000] [0.000] [0.000] [0.000] slope $&1 0.428 0.338 0.907 0.862 0.937 0.892 1.216 1.324 1.319 1.463 se (0.150) (0.129) (0.239) (0.190) (0.357) (0.286) (0.552) (0.401) (0.658) (0.461) p value [0.005] [0.009] [0.000] [0.000] [0.009] [0.002] [0.028] [0.001] [0.046] [0.002] slope $! 0.303 0.084 0.584 0.243 0.370 0.014 &0.152 &0.621 &0.493 &1.070 se (0.106) (0.082) (0.212) (0.151) (0.295) (0.229) (0.445) (0.355) (0.560) (0.442) p value [0.004] [0.303] [0.006] [0.109] [0.210] [0.950] [0.733] [0.081] [0.380] [0.016] r2 0.105 0.058 0.111 0.118 0.043 0.048 0.025 0.046 0.021 0.040 note. this table represents long-horizon ordinary least square regressions of asymmetric response model incorporating changes in consumer sentiment. k stands for monthly cumulated k horizon returns and when k is greater than one, the regressions use overlapping monthly data. the dependent variable is the excess return of either the crsp equally weighted portfolio or the crsp value weighted portfolio. the independent variables are defined in the following manner: $& is the change in csi if sentiment decreases (i.e., $csi ' 0) and zero otherwise, and $! is the change in csi if sentiment increases (i.e., $csi ( 0) and zero otherwise. all data is monthly and is from january 1978 until december 2010. statistical significance is determined by the newey-west p values. 55m.a. johnson, a. naka / financial services review 23 (2014) 45–61 returns. the magnitudes of the estimated coefficients $& are in general greater than those of $! for long-horizon models estimated except for two cases (both estimates are not statistically significant). the upside $ for this age group is statistically significant at the 5% level of significance in only 2 out of the 10 long-horizon models estimated, whereas the downside $ is statistically significant in 5 out of the 10 long-horizon models estimated. we conjecture that the focus on avoiding losses is present among younger individuals, especially given that they are in the accumulation phase that reilly and brown (2008) describe and possibly cannot afford to lose what assets they do have at this particular point in their lives. table 4 asymmetric response model: csi for individuals 18–34 years old, 35–54 years old, and 55 years old and older dependent variable forecast horizon (k) 1 3 6 12 24 ew returns vw returns ew returns vw returns ew returns vw returns ew returns vw returns ew returns vw returns panel a: csi for individuals 18–34 years old slope $& 0.336 0.283 0.441 0.459 0.394 0.492 0.217 0.525 0.137 0.633 se (0.128) (0.105) (0.184) (0.142) (0.266) (0.190) (0.456) (0.301) (0.480) (0.326) p value [0.009] [0.007] [0.017] [0.001] [0.139] [0.010] [0.635] [0.082] [0.775] [0.053] slope $! 0.120 0.007 0.389 0.147 0.303 0.025 0.412 &0.047 0.187 &0.452 se (0.058) (0.047) (0.159) (0.096) (0.186) (0.135) (0.332) (0.190) (0.317) (0.245) p value [0.041] [0.886] [0.015] [0.126] [0.104] [0.852] [0.215] [0.805] [0.555] [0.066] r2 0.060 0.043 0.050 0.048 0.017 0.020 0.009 0.010 0.002 0.011 panel b: csi for individuals 35–54 years old slope $& 0.273 0.195 0.740 0.667 0.738 0.695 0.982 0.980 1.151 1.104 se (0.110) (0.093) (0.220) (0.177) (0.341) (0.272) (0.481) (0.369) (0.614) (0.424) p value [0.014] [0.038] [0.001] [0.000] [0.031] [0.011] [0.042] [0.008] [0.061] [0.010] slope $! 0.191 0.066 0.248 0.070 0.206 &0.020 &0.248 &0.494 &0.557 &0.807 se (0.083) (0.063) (0.175) (0.128) (0.244) (0.186) (0.362) (0.293) (0.472) (0.330) p value [0.023] [0.302] [0.156] [0.586] [0.398] [0.916] [0.493] [0.093] [0.239] [0.015] r2 0.072 0.035 0.086 0.097 0.040 0.046 0.025 0.042 0.026 0.038 panel c: csi for individuals 55 years old and older slope $& 0.238 0.178 0.547 0.495 0.443 0.370 0.560 0.602 0.733 0.806 se (0.102) (0.083) (0.164) (0.134) (0.236) (0.188) (0.348) (0.255) (0.396) (0.317) p value [0.020] [0.033] [0.001] [0.000] [0.061] [0.050] [0.109] [0.019] [0.065] [0.011] slope $! 0.169 0.049 0.264 0.119 0.195 0.047 &0.074 &0.295 &0.307 &0.611 se (0.075) (0.061) (0.137) (0.101) (0.182) (0.135) (0.285) (0.210) (0.338) (0.258) p value [0.025] [0.426] [0.055] [0.239] [0.284] [0.728] [0.795] [0.160] [0.364] [0.018] r2 0.061 0.030 0.062 0.069 0.019 0.017 0.009 0.017 0.011 0.022 note. this table represents long-horizon ordinary least square regressions of asymmetric response model incorporating changes in consumer sentiment. k stands for monthly cumulated k horizon returns and when k is greater than one, the regressions use overlapping monthly data. the dependent variable is the excess return of either the crsp equally weighted portfolio or the crsp value weighted portfolio. the independent variables are defined in the following manner: $&1 is the change in csi if sentiment decreases (i.e., $csi ' 0) and zero otherwise, and $! is the change in csi if sentiment increases (i.e., $csi ( 0) and zero otherwise. all data is monthly and is from january 1978 until december 2010. statistical significance is determined by the newey-west p values. 56 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 panel b of table 4 presents results that pertain to changes in consumer sentiment among individuals 35 to 54 years old and their relationship to excess market returns. we observe that all downside $s are greater than the upside $s in all long-horizon models estimated for this age group. additionally, 9 out of the 10 $& coefficients are positive and statistically significant at the 1% or 5% level of significance, and mostly increase in magnitude as the forecast horizons increase. for the forecast horizons of 3, 6, 12, and 24 months, the downside $s for individuals 35 to 54 years old are greater than the corresponding downside $s for individuals 18 to 34 years old. the results support bakshi and chen (1994) who show that risk aversion increases with age and conclude that demographic changes will cause price fluctuations in the capital markets as they affect macroeconomic variables. furthermore, the consolidation phase described by reilly and brown (2008) involves the time when it is most likely that an individual’s earnings exceed their expenses and as a result, they can invest this difference in additional assets such as stocks. reilly and brown state that “because individuals in this phase are concerned about capital preservation, they do not want to take very large risks that may put their current nest egg in jeopardy.” as their careers and age advance, individuals typically begin investing their wealth more conservatively as they near retirement. our asymmetric response model results for individuals 55 years of age and older are presented in panel c of table 4. they indicate that 6 of the 10 downside $s estimated are statistically significant. additionally, $& for the oldest age group exhibits a mostly increasing trend that is consistent with the other asymmetric response models, as well as all downside $s being greater than the upside $s in all long-run horizons. the upside $s exhibit a mostly decreasing trend as the forecast horizon increases, but only 2 out of the 10 $! coefficients are statistically significant at a 5% level of significance. it should be noted that the $& coefficients for this age group are all smaller than the corresponding horizon $& coefficients for the 35 to 54 age group. this implies that individuals 35 to 54 years old exhibit more aversion towards downside risks than individuals 55 and older. given that stock market excess returns are used as the dependent variable in our asymmetric models, we conjecture that older individuals are less likely to have significant stock market exposure because of their nearer-term focus upon retirement and inclination to possibly hold a larger portion of fixed income securities. in summary, we observe that the magnitudes of the downside $ coefficients are greater than those of the upside $s because downside risk matters more. this finding holds true in the majority of our results for the composite csi, as well as the csi for each respective age group. further, our results show that the sizes of the downside $s are generally greater for consumers 35 to 54 years old and 55 years and older, than those of the downside $ coefficients for the youngest age group, consumers 18 to 34 years old. the results confirm that the older we get, the more risk averse we become. our findings are also consistent with riley and chow (1992) and halek and eisenhauer (2001) who show how risk aversion can be explained by demographic attributes. thus, financial planners should continue to thoroughly assess the risk tolerance levels of their clients as they recommend risky assets for their portfolios. individual investors should also recognize this risk-return relationship and take the necessary steps to diversify their holdings, consider risk-reduction strategies as they 57m.a. johnson, a. naka / financial services review 23 (2014) 45–61 approach retirement, and consider seeking professional investment assistance to ensure appropriate investing approaches. 3.3. size effects baker and wurgler (2006) and lemmon and portniaguina (2006) find that sentiment has more of an impact on small-firm stocks versus large-firm stocks. however, does this noise trader argument hold among changes in consumer sentiment for all ages? to empirically test this question, we estimate eq. (1) but use a one month forecast horizon (k % 1) and use crsp market capitalization portfolios. the dependent variable in the modified eq. (1) is the stock return of the decile i portfolio minus the risk-free rate, and the independent variable is used and defined in the same manner as in eq. (1). crsp segments these sized-sorted decile portfolios into deciles based on a firm’s market capitalization, whereby small firms will appear in lower decile portfolios and larger firms appear in higher decile portfolios. table 5 presents the results of the size effects. our findings are consistent with lemmon and portniaguina (2006) and baker and wurgler (2006) for all age groups; smaller firms’ market risk premiums are affected more by changes in sentiment than those of larger firms. for example, the coefficient for $csit for the csi composite (panel a) is 0.430 for the decile 1 portfolio and 0.185 for the decile 10 portfolio. the implication of this is important, both statistically and economically; a 10% change in overall consumer sentiment (consumers of all ages) results in a market risk premium change of 4.30% in the following month for the smallest firms versus a market risk premium change of 1.85% for the largest firms. lee, shleifer, and thaler (1991) suggest that small-cap stocks are more closely followed by individual investors. this size effect is, for the most part, linear in that the smallest firms are affected the most and this effect gradually decreases as firm size increases. this pattern is true for changes in consumer sentiment for all age groups. our finding thus confirms to individual investors and financial planners that investing in small-cap stocks can be risky and proper risk-return tradeoff analyses must be performed to confirm suitability for one’s portfolio and risk tolerance. 4. concluding remarks in this article, we examine the ability of consumer sentiment to forecast short-term as well as long-term equity returns and whether or not negative sentiment has a larger effect on the returns than positive sentiment for three different age groups. we observe that negative changes in sentiment have a greater influence on stock returns than positive changes in sentiment. further, sentiment is more effective forecasting short-term holding period returns than long-term holding period returns, and these results are consistent across all ages. these empirical findings are consistent with the documented behavior that agents place greater weight on downside risk than they place on upside gains, and are agreeable with the notion of the prospect theory proposed by kahneman and tversky (1979). our results also document a difference in age groups’ sentiment, and that younger individuals appear to be less risk-averse than older individuals. the finding supports the 58 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 existing literature and the concept of the life cycle investment hypothesis. finally, we present additional evidence for the notion that the risk premiums of smaller firms are more affected by sentiment than larger firms; a result that is in line with the noise trader hypothesis. we provide evidence that reminds individual investors and financial planners that risk is an important consideration when investing. moreover, demographic characteristics such as age matter when determining appropriate investing approaches and risk tolerance. this article applies well-known concepts in behavioral economics and finance, and deepens our understanding of the relationship and importance of consumer sentiment in evaluating the equity returns. table 5 size effects approach decile 1 decile 2 decile 3 decile 4 decile 5 decile 6 decile 7 decile 8 decile 9 decile 10 panel a: all age groups: csi composite intercept # 0.009 0.007 0.008 0.007 0.008 0.008 0.008 0.007 0.007 0.006 se (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.002) p value [0.008] [0.032] [0.009] [0.011] [0.007] [0.004] [0.004] [0.008] [0.007] [0.015] slope $ 0.430 0.417 0.383 0.346 0.323 0.283 0.270 0.245 0.225 0.185 se (0.064) (0.073) (0.077) (0.075) (0.076) (0.068) (0.070) (0.066) (0.066) (0.055) p value [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.001] [0.001] panel b: age group 18–34 years old intercept # 0.009 0.007 0.008 0.007 0.008 0.008 0.008 0.007 0.007 0.006 se (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.002) p value [0.012] [0.041] [0.013] [0.014] [0.010] [0.006] [0.005] [0.011] [0.009] [0.017] slope $ 0.252 0.247 0.219 0.202 0.186 0.165 0.153 0.145 0.139 0.121 se (0.060) (0.061) (0.060) (0.058) (0.057) (0.053) (0.053) (0.051) (0.047) (0.042) p value [0.000] [0.000] [0.000] [0.001] [0.001] [0.002] [0.004] [0.004] [0.004] [0.004] panel c: age group 35–54 years old intercept # 0.009 0.007 0.007 0.007 0.008 0.008 0.008 0.007 0.007 0.006 se (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.002) p value [0.011] [0.042] [0.013] [0.014] [0.010] [0.006] [0.005] [0.010] [0.009] [0.017] slope $ 0.266 0.264 0.253 0.226 0.208 0.186 0.183 0.155 0.147 0.109 se (0.045) (0.057) (0.061) (0.056) (0.059) (0.052) (0.053) (0.050) (0.051) (0.040) p value [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.001] [0.002] [0.004] [0.006] panel d: age group 55 years old and older intercept # 0.009 0.007 0.007 0.007 0.008 0.008 0.008 0.007 0.007 0.006 se (0.004) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) (0.002) p value [0.012] [0.043] [0.013] [0.015] [0.010] [0.006] [0.006] [0.011] [0.009] [0.018] slope $ 0.245 0.231 0.207 0.190 0.177 0.148 0.138 0.129 0.112 0.096 se (0.043) (0.044) (0.046) (0.046) (0.044) (0.041) (0.041) (0.040) (0.037) (0.034) p value [0.000] [0.000] [0.000] [0.000] [0.000] [0.000] [0.001] [0.002] [0.003] [0.005] note. this table represents ordinary least square regressions of one month excess returns of various size-sorted decile portfolios on lagged one month changes in the consumer sentiment index (csi). the indices used are the crsp annual rebalanced indices based on individual stock market capitalization values. the market capitalization portfolios are formed from stocks listed on the new york stock exchange (nyse), american stock exchange (amex), and the national association of securities dealers automated quotations (nasdaq) and are rebalanced each year and separated based on deciles. decile 1 represents firms that have smaller market capitalizations versus the firms with higher market capitalizations (higher decile indices). decile 10 represents firms that have some of the largest market capitalizations of listed firms. all data is monthly and is from january 1978 until december 2010. statistical significance is determined by the newey-west p values. 59m.a. johnson, a. naka / financial services review 23 (2014) 45–61 notes 1 for example, these studies include akhtar et al. (2011), baker and wurgler (2006), baker and wurgler (2007), chung et al. (2012), fisher and statman (2003), lemmon and portniaguina (2006), schmeling (2009), stambaugh et al. (2012), and yu and yuan (2011). 2 reilly and brown (2008) identify four life cycle phases: the accumulation phase, consolidation phase, spending phase, and gifting phase. 3 fisher and statman (2003) define small-cap stocks as the average of the returns on the bottom three deciles of crsp decile 1 to decile 10 portfolios formed based on market capitalization. 4 the market variables used are the closed-end fund discount, nyse share turnover, the number of initial public offerings (ipos), the average first day return of ipos, the dividend premium, and the equity share in new issues. 5 we follow the description of the long-horizon regression model presented by campbell et al. (1997). 6 these are not dummy variables and provide the magnitude of sentiments’ negative and positive effects. 7 we note that the csi composite is not the average of three different age groups because of slightly different survey methods. 8 u.s. business cycle expansions and contractions according to the national bureau of economic research (http://www.nber.org/cycles.html). 9 newey-west methods are used to obtain robust parameter estimates by correcting the possible autocorrelation and heterocedasticity. acknowledgments we are grateful for helpful comments from tarun mukherjee, arja turunen-red, wei wang, seminar participants at the 2012 midwest finance association annual meeting, and the federal reserve bank of chicago. we also thank the editor, stuart michelson, and the anonymous reviewers for their comments and suggestions that improved this study. this research was supported by a summer research grant from the sellinger school of business and management at loyola university maryland. references akhtar, s., faff, r., oliver, b., & subrahmanyam, a. (2011). the power of bad: the negativity bias in australian consumer sentiment announcements on stock returns. journal of banking and finance, 35, 1239–1249. ando, a. & modigliani, f. (1963). the “life cycle” hypothesis of saving: aggregate implications and tests. american economic review, 53, 55–84. ang, a., chen, j., & xing, y. (2006). downside risk. review of financial studies, 19, 1191–1239. baker, m., & wurgler, j. (2006). investor sentiment and the cross-section of stock returns. journal of finance, 61, 1645–1680. 60 m.a. johnson, a. naka / financial services review 23 (2014) 45–61 baker, m., & wurgler, j. (2007). investor sentiment in the stock market. journal of economic perspectives, 21, 129–152. bakshi, g., & chen, z. (1994). baby boom, population aging, and capital markets. journal of business, 67, 165–202. campbell, j. y., lo, a. w., & mackinlay, a. c. (1997). the econometrics of financial markets. princeton, nj: princeton university press. chung, s., hung, c., & yeh, c. (2012). when does investor sentiment predict stock returns? journal of empirical finance, 19, 217–240. fisher, k. l., & statman, m. (2003). consumer confidence and stock returns. journal of portfolio management, 30, 115–128. friedman, j., & phillips, h. e. (2010). the downside risk of postponing social security benefits. financial services review, 19, 285–294. gibson, r., michayluk, d., & van de venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review, 22, 23–50. halek, m., & eisenhauer, j. g. (2001). demography of risk aversion. journal of risk and insurance, 68, 1–24. hariharan, g., chapman, k. s., & domian, d. l. (2000). risk tolerance and asset allocation for investors nearing retirement. financial services review, 9, 159. harlow, w. v., & rao, r. k. s. (1989). asset pricing in a generalized mean-lower partial moment framework: theory and evidence. journal of financial and quantitative analysis, 24, 285–311. ho, c., & hung, c. (2009). investor sentiment as conditioning information in asset pricing. journal of banking and finance, 33, 892–903. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–291. larkin, c., lucey, b. m., & mulholland, m. (2013). risk tolerance and demographic characteristics: preliminary irish evidence. financial services review, 22, 77–91. lee, c. m. c., shleifer, a., & thaler, r. h. (1991). investor sentiment and the closed-end fund puzzle. journal of finance, 46, 75–109. lemmon, m., & portniaguina, e. (2006). consumer confidence and asset prices: some empirical evidence. review of financial studies, 19, 1499–1529. ludvigson, s. c. (2004). consumer confidence and consumer spending. journal of economic perspectives, 18, 29–50. qiu, l. x., & welch, i. (2006). investor sentiment measures. working paper, brown university. reilly, f. k., & brown, k. c. (2008). investment analysis and portfolio management (9th ed.). boston, ma: south-western publishing. rich, s. p., & reichenstein, w. (1993). market timing for the individual investor: using the predictability of long-horizon stock returns to enhance portfolio performance. financial services review, 3, 28. riley, w., & chow, k. v. (1992). asset allocation and individual risk aversion. financial analysts journal, 48, 32–37. schmeling, m. (2009). investor sentiment and stock returns: some international evidence. journal of empirical finance, 16, 394–408. schooley, d. k., & worden, d. (1996). risk aversion measures: comparing attitudes and asset allocation. financial services review, 5, 87. stambaugh, r., yu, j., & yuan, y. (2012). the short of it: investor sentiment and anomalies. journal of financial economics, 104, 288–302. valkanov, r. (2003). long-horizon regressions: theoretical results and application. journal of financial economics, 68, 201–232. yu, j., & yuan, y. (2011). investor sentiment and the mean–variance relation. journal of financial economics, 100, 367–381. 61m.a. johnson, a. naka / financial services review 23 (2014) 45–61 do women have lower levels of financial literacy, or are they opting out? a look at the non-response gender bias in financial literacy measurement tracey westa,*, laura de zwaana, di johnsonb adepartment of accounting, finance, and economics, griffith university, gold coast campus qld 4222 bdepartment of accounting, finance, and economics, griffith university, nathan campus qld 4111 abstract men consistently appear to outperform women on standard financial literacy tests. however, could the results be because of inherent gender bias in measurement tools? this study investigates the reasons for women selecting the non-response option in financial literacy questions, including numerical self-efficacy, risk aversion, and confidence. our analysis finds evidence that women answer more questions than men utilizing the non-response option. a sustained lack of confidence with financial information is the primary reason. these results are important for shaping policy and providing resources that close the gap in measurement and ability. © 2023 academy of financial services. all rights reserved. jel classifications: g53; g40 keywords: financial literacy; women; gender studies; measurement bias; financial literacy education 1. introduction financial education is recognized as a core component of the financial empowerment of individuals and the overall stability of the financial system (oecd, 2017). the organization for economic co-operation and development (oecd) defines financial literacy as “a combination of awareness, knowledge, skill, attitude and behavior necessary to make sound financial decisions and ultimately achieve individual financial well-being” (2017, p. 13). *corresponding author: tel.: +61-7-555-29769. e-mail address: t.west@griffith.edu.au 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 55–71 measuring financial literacy aptitude is important for benchmarking progress and to inform strategies for financial education. a large survey of adult financial literacy in oecd countries shows that, on average, fewer than half of adults (48%) could answer 70% of the financial knowledge questions correctly, or in other words, meet the minimum target score (oecd, 2017). respondents need help understanding diversification and compounding interest, with only 60% able to understand diversification and 27% able to calculate both simple interest and recognize the added benefit of compounding over five years. these two concepts are essential for people to recognize the consequences of financial decisions, such as paying only the minimum repayment on credit cards and saving for financial security in retirement. the difference between the percentage of men and women achieving the minimum target score for financial knowledge in g20 countries stands at 11 percentage points, with men significantly more likely to achieve this score than women in all but three of the countries with comparable data (china, indonesia, and the russian federation; oecd, 2020a). this finding is consistent across many studies, with reasons for the disparity pointing largely to men and women having different levels of interest in money (chen & volpe, 2002; lusardi et al., 2010). this study recognizes widespread gender differences in financial literacy but takes a closer look at gender differences in the measurement tools. accordingly, we offer insight as to why women may score lower on financial literacy question sets, particularly if the question set is multiple-choice and provides an “unsure” option. there is a growing body of literature that identifies different factors leading to a gender difference in the decision to skip a question (riener & wagner, 2017). the mathematics education literature is of particular interest, given the mathematical skills required in financial decisions. gender differences in mathematics learning in favor of males have persisted since the 1970s (see seminal work by fennema & sherman [1977] and a comprehensive summary of early research by leder, 1992). multiple explanations for the difference, such as early experience, biological constraints, educational policy, and cultural context contribute to performance gaps, the ability to solve challenging mathematical problems, and interest and participation in mathematics learning (halpern et al., 2007). as assessment of mathematical performance is usually via an examination under time constraints, some research has posited that it is the performance under the conditions, rather than knowledge, that explains the difference. research consistently evidences reduced performance for females under time constraints, with recent studies positing increased risk aversion as a potential cause (dilmaghani, 2020). therefore, selecting an answer when the respondent is not sure if it correct, is risky behavior more likely to be demonstrated by men, than avoiding the risky behavior and selecting a “do not know” response, more often demonstrated by women. accordingly, we draw on research from studies conducted on gender differences in mathematical testing, such as that by niederle and vesterlund (2010) who find that women tend to have a higher number of skipped questions in a multiple-choice test, especially when the task is competitive. they posit that there is a negligible gender difference in math ability, but men embrace a competitive environment and have higher self-perceived skill levels. therefore, when completing a test under a competitive setting, men pick an answer and 56 t. west et al. / financial services review 31 (2023) 55–71 tackle difficult questions, and women tend to choose the less challenging option more often, that is, skipping questions or selecting the ‘i don’t know’ answer. niederle and vesterlund (2010) note that further research is needed to strengthen conclusions that can be drawn from women’s responses to the math tests, including as an example in the area of single-sex schooling. the novel contribution of this study is to investigate gender bias in one set of questions that measures financial literacy performance. we study the non-response option as the decision science and education literature identifies women tending to pick this option, while men do not. we regress purported determinants of non-response, such as gender, numerical ability, risk aversion, and confidence on the binary non-response for each of the three financial literacy questions and a sum of non-responses. we find that there is a large gender effect, that is, women choose the non-response more often than men, and urge researchers to consider more deeply the formulation of questions. this paper is structured as follows. section 2 briefly reviews the literature. section 3 explains the empirical methodology and the data employed in the analysis. section 4 presents the results and a discussion of the findings. 2. literature review research in the area of gender differences in financial literacy is growing, with many studies reporting sizeable gender differences in adult financial literacy scores (hung et al., 2009; lusardi & mitchell, 2009; lusardi & tufano, 2009a, 2009b; lusardi et al., 2010). for example, zissimopoulos, karney, and rauer (2008) find that less than 20% of middleaged college-educated women can answer a basic compound interest question compared with about 35% of college-educated men of the same age. gender gaps in financial literacy exist for a variety of demographic and socioeconomic groups, including teenagers (bottazzi & lusardi, 2016; driva et al., 2016), university students (gerrans & heaney, 2019), and migrants (karunarathne & gibson, 2014). gender gaps may also be related to low levels of education, income, wealth, and health, as these are other factors associated with poor financial literacy (west & worthington, 2018); however, rothwell and wu (2019) find that gender differences are persistent even when accounting for exposure to financial education. many researchers suggest that women score lower because they either do not know basic facts, terminology or personal finance concepts, or they do not perform well in mathematicsrelated questions (agnew & cameron-agnew, 2015; chen & volpe, 1998, 2002; goldsmith & goldsmith, 1997; volpe et al., 1996). goldsmith and goldsmith (1997) suggest that women have lower scores than men because women, in general, are less interested in the topics of “investments” and “personal finances.” they find that people’s financial literacy is related to their self-perception of their knowledge in personal finance. men have higher self-perceived education of investments than women, and men are found to be more knowledgeable than women. gerrans et al. (2014) investigate the link between financial literacy and participants’ satisfaction with their own financial situation and find financial knowledge drives good t. west et al. / financial services review 31 (2023) 55–71 57 financial behavior in men and financial satisfaction, but for women, this link is not evident; instead, financial status has a strong impact on financial satisfaction for women. interestingly, there is no evidence of a gender gap in financial literacy in children, despite the prevalence for adults. for example, agnew and cameron-agnew (2015) did not find a significant difference by gender in financial literacy quiz scores for 14-year-olds, but do observe a gender difference in first-year tertiary students. the recent program for international students assessment (pisa) report on financial literacy performance in 15year-old high school students in oecd countries finds that boys scored two points higher than girls, although this difference is not regarded as statistically significant (oecd, 2020b). emerging research, however, finds that boys and girls are treated differently by their parents regarding money conversations which have a long-lasting effect (agnew & cameronagnew, 2015). for example, a new zealand study found boys receive more pocket money than girls even though both spend 2.4 hours a week doing chores (wade, 2013). west and de zwaan (2020) identify differing financial socialization pathways to building financial literacy in men and women, with higher levels of financial literacy in women linked to more frequent money conversations in the home when they were children. this suggests that while performance metrics indicate equality on average in boys and girls, their experiences with money in these formative years are different and may have lasting impacts. the widening financial literacy deficit we observe in women as they age may be due to contextual factors. several structural features in our society affect women’s relationship to money, including the gender pay gap and women’s contribution to unpaid and caring roles. it is well-documented that women, on average, earn less than men for the same work (cassells et al., 2009). other factors, including, gender stereotypes, lower wages for female-dominated industries, inflexible working conditions, time out of the workforce due to caring roles, and gender discrimination are found to contribute to this difference (alcon, 1999; anthes & most, 2000; timmermann, 2000). consequently, comparatively low incomes of women mean fewer resources to support the growth of confidence in future opportunities, and more attention to everyday financial issues, thwarting retirement planning activity (larisa et al., 2021). labor market status and occupation have been found to explain 16% of the financial literacy gender gap in a study of australian adults (preston & wright, 2019). family economists contend that gender-based labor division within the household means that men traditionally and frequently make the majority of financial decisions for families, and it is through these activities that they are more likely than women to be exposed to financial information (bucher-koenen et al., 2017; fonseca et al., 2012; hsu, 2011). therefore, the gender gap is magnified as those making financial decisions further enhance their financial knowledge through feedback effects (fonseca et al., 2012; lusardi & mitchell, 2014; lusardi et al., 2017; ward & lynch, 2019). this brief review provides insight into the challenges faced by policymakers and educators seeking to improve overall financial literacy, when barriers exist for half the population. this study does not downplay the importance of addressing broader gender biases by accusing measurement tools of being inaccurate but rather aims to focus attention on the unique role gender plays in the field of financial knowledge. by so doing, we hope to position gender as a key consideration in the measurement and analysis of financial literacy. 58 t. west et al. / financial services review 31 (2023) 55–71 3. conceptual framework while a variety of financial literacy measures have emerged in recent years (such as that by knoll & houts, 2012), the “big three” questions initially designed by lusardi and mitchell persist. the three questions were originally designed for the 2004 us health and retirement study and incorporate principles of simplicity, relevance, brevity and capacity to differentiate. they were consequently added to the 2007–08 national longitudinal survey of youth (nlsy), the 2008 american life panel, and the 2009 financial capability study (lusardi & mitchell, 2011). in australia, they were added to the household, income and labor dynamics in australia (hilda) panel data in 2016 (wilkins & lass, 2018). many studies of smaller cohorts have adopted these questions in a wide variety of contexts. since lusardi and mitchell published an overview of financial literacy studies in 2011, there have been another half a million journal articles published on the broad topic of financial literacy. the first question is designed to measure the capacity to do a simple calculation; the second question measures an understanding of inflation; and the third question gauges knowledge of risk diversification: 1. suppose you had $100 in asavings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow? more than $102 exactly $102 less than $102 do not know refuse to answer 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account? more than today exactly the same less than today do not know refuse to answer 3. please tell me whether this statement is true or false. “buying a single company’s stock usually provides a safer return than a stock mutual fund.” true false do not know refuse to answer criticisms of the questions include the reliance on self-assessment, misunderstanding the question (especially if data are collected via phone interview), and sensitivity to question t. west et al. / financial services review 31 (2023) 55–71 59 framing (lusardi & mitchell, 2011; worthington, 2013). further, there are criticisms over whether these three questions assess domain knowledge or general mathematical skills, particularly the first question (schuhen & schurkmann, 2014). with the second question, respondents need to understand inflation and interest concepts, and apply mathematical skills. it is not clear whether the respondents apply their knowledge or take a random guess, given the multiplechoice format of the questions (schuhen & schurkmann, 2014). as with all data collection, these criticisms need to be considered when interpreting results. there is also very little agreement on how to measure financial knowledge. we propose that there is a further important factor to consider and draw from the decision science literature. this literature is building evidence that women and men are distinctly different (in aggregate) in their approach to decision-making under uncertainty, such as presented in multiple-choice questions (baldiga, 2014; espinosa & gardeazabal, 2010; stumpf & stanley, 1996; tannenbaum, 2012). decision science research uses experiments or competitions to assess gender differences in question responses. research finds the importance of the result affects performance significantly for women, such as high stakes university entrance exams. this may be because women dislike high-pressure situations or competitive settings, especially in mathematical tasks (niederle & vesterlund, 2007, 2010). for example, pekkarinen (2015) finds that women are less likely than men in finland to gain entry to university because in the entrance exam, women skipped more questions than is optimal to maximize the probability of acceptance. baldiga (2014) also finds that when a penalty is imposed for wrong answers, women answer fewer questions than men, even after controlling for knowledge of the material, levels of confidence, and risk preferences. espinosa and gardeazabal (2010) conducted a field experiment by switching the incentive to provide a reward for skipping questions. this was found to increase the number of omissions, but mostly by women. even with the incentive to provide a non-response, men tend to provide an answer. further evidence of women behaving differently in high-stakes tasks is provided by the risk aversion literature, where women are consistently found to be more risk-averse than men (west & worthington, 2013, 2014). if answering questions at random risks losing an incentive or points, individuals can avoid this risk by not answering (charness & gneezy, 2012; croson & gneezy, 2009). many researchers find that women are consistently risk averse in many domains, and in the domain of test-taking women display risk aversion when avoiding answering a question if they are not sure and will lose points or an incentive (charness & gneezy, 2012; west & worthington, 2014). confidence is another factor being considered. men have been found to be more overconfident than women in male-specific tasks, like numeracy (barber & odean, 2001). this relates to fulfilling stereotype expectations, as individuals (conditional on ability) are less willing to contribute ideas in areas that are stereotypically outside of their gender’s domain (coffman, 2014). this behavior is found in some studies of school-age children in which boys are shown to outperform girls when facing novel problems presented in standardized tests and girls are more confident answering questions about familiar material (kimball, 1989; loewen et al., 1988). riener and wagner (2017) find that girls are underconfident and underestimate their ability to provide the correct answer, and this is consistent with the stereotype threat (riener & wagner, 2017). 60 t. west et al. / financial services review 31 (2023) 55–71 finally, there is limited evidence that socioeconomic status is significantly linked to decision-making behavior, but riener and wagner (2017) find a small indication that it is. in their experiment involving a standardized test in schools and the kahneman (2003) dual model of thinking, they find pupils in vocational schools answer questions according to the automatic system 1 while pupils in high schools seem to answer according to the effortful system 2. synthesis of the decision science evidence directs our investigative attention to including factors relating to self-perceived numeracy ability (niederle & vesterlund, 2007), risk aversion (charness & gneezy, 2012), overconfidence (barber & odean, 2001; coffman, 2014), and socioeconomic background (riener & wagner, 2017). this results in the following model of financial literacy non-response: 4. methodology this study uses data from a survey of 420 students from an australian university in 2019. students were invited to participate via a broadcast email, and we use a monetary prize draw to incentivize and increase participation (yu et al., 2017). the survey consisted of 64 questions on various topics involving money, including financial stress, financial advice, student loans, and financial knowledge. the final sample consists of 266 respondents after screening for full responses to the variables of interest for this study. the descriptive statistics of the respondents shown in table 1 are overweight to women (74%), being 23 years old or younger and earning less than $19,999 per annum. participants who identified gender as non-binary were excluded due to the small number of respondents. we also note the difficulty in controlling for bias due to students being confident in money matters because they have chosen to study related topics like a business. an inspection of the fields of the study shows a statistical difference in participation in “business and management” by men t. west et al. / financial services review 31 (2023) 55–71 61 (i.e., more men than women as a proportion of the gender group). conversely, participation in “law and paralegal studies” and “social work” is statistically significant in favor of women. accordingly, the interpretation of results needs to consider the limitation that the men in our sample are mainly in the business and management field of study, which may table 1 descriptive statistics of sample parameter women (n = 196) men (n = 70) proportion of total sample proportion of total sample age what age category are you in? 1–23 or younger 0.500 0.443 2–24 to 29 0.250 0.314 3–30 to 39 0.158 0.171 4–40 to 49 0.056 0.043 5–50 to 59 0.031 0.014 6–60 or over 0.005 0.014 inc what is your current annual income, including paid work, government benefits, and other financial support? 0–don’t know/prefer not to answer 0.097 0.157 1–$0 0.056 0.086 2–$1–$19,999 0.357 0.400 3–$20,000–$39,999 0.255 0.143 4–$40,000–$59,999 0.117 0.086 5–$60,000–$79,999 0.061 0.029 6–$80,000–$99,999 0.041 0.043 7–above $100,000 0.015 0.057 field of study agriculture and environment 0.015 0.014 architecture and built environment 0.005 0.014 business and managementa 0.209 0.500 communications 0.015 0.000 computing and information systems 0.010 0.029 creative arts 0.056 0.043 dentistry 0.015 0.014 engineering 0.010 0.029 health services and support 0.102 0.043 humanities, culture, and social sciences 0.097 0.143 law and paralegal studiesa 0.122 0.014 medicine 0.056 0.029 nursing 0.031 0.000 psychology 0.077 0.029 rehabilitation 0.020 0.014 science and mathematics 0.077 0.043 social worka 0.056 0.000 teacher education 0.026 0.043 tourism, hospitality, personal services, sport, and recreation 0.026 0.000 veterinary science 0.005 0.000 note. at tests show that the participation in this field of study is significantly different between women and men in this sample. 62 t. west et al. / financial services review 31 (2023) 55–71 have influenced their financial knowledge and field of study bias is proposed as an area of further research. we are also not able to infer from the data the reasons why participants select the answer they do. investigation using larger data sets and other research methods—such as interviews and focus groups—would contribute to the body of knowledge. the financial literacy questions are presented in this survey so that answers collected are “yes,” “no,” “unsure,” and “prefer not to answer,” a slight variation to the lusardi and mitchell (2011) specification. “no” is the correct answer for all three questions. we code the responses to “unsure” as a binary variable and include this binary for answers to each financial literacy question (i.e., interest “unsure” is intuns, inflation “unsure” is infuns and diversification “unsure” is divuns), and a total out of the three (uns), which can take the form of 0 (no non-responses) to 3 (3 non-responses). we use proxies from the available survey questions to quantify the direction and magnitude of the effect of the hypothesized determinants of uns. the decision-making literature highlights the importance of self-perceived ability in numerical tasks as a determining factor for non-response. we identify a proxy for this variable from the survey questions and take the opportunity to broaden the variable set, as the question asks for self-perceived ability in math, reading, writing, and financial knowledge at school age. the oecd pisa study of 15-year-olds finds a close link between financial literacy, numeracy, and literacy. on average, girls outperform boys in reading, and boys outperform girls in mathematics and the averaging out explained the small variation between genders (oecd, 2020a). we include all three proxies, as women may have marginally different pathways to being financially literate than men (reading vs. math). self-perceived ability (num) is proxied by the question below regarding math (num), reading (read), writing (writ), and financial knowledge (fin): “thinking about when you were in year 7 at school, how did you perform in the topics below compared with other children in your class? much better better about the same worse much worse.” we recode a score of 1 through 5, with 5 indicating a high level of self-perceived ability for the responses. a risk aversion metric (risk) is taken from a well-known question that appears in the hilda survey: “which of the following statements comes closest to describing the amount of financial risk that you are willing to take with your spare cash? that is, cash used for savings or investments. i take substantial financial risks expecting to earn substantial returns i take above-average financial risks expecting to earn above-average returns i take average financial risks expecting average returns i am not willing to take any financial risks i never have any spare cash.” we recode a score of 1 through 5, where 5 indicates a high willingness to take risks. t. west et al. / financial services review 31 (2023) 55–71 63 finally, we use the following question to proxy for confidence (conf), and appears in the ohio national student financial wellness study: “i am confident i can manage my finances. describes me completely describes me very well somewhat describes me describes me very little does not describe me at all not applicable” we recode from 1 through 6, with 6 indicating a high level of confidence in managing finances. unfortunately, the data does not contain a reasonable proxy for socioeconomic status (ses). measures considered included using the postcode of current accommodation that could be linked to the socio-economic indexes for areas (seifa) index; however, this indicator would likely not be representative of family ses as students often move to city areas to study. we also tested this question: “again thinking about when you were in year 7 at school, on a scale of 1 to 10, with 10 being a high level of financial satisfaction, rate the level of household financial satisfaction as best you can at this time.” however, regression results were not significant and due to the limitation of the question it was dropped from further analysis. the probit model fits a maximum likelihood model with a dichotomous dependent variable coded as 0/1, and is given as pr yj 6¼ 0jxj � � ¼ w xjbð þ where yj is uns, xj is the set of predictor variables (num, risk, conf) and w is the standard cumulative normal. we report coefficients, robust standard errors, and marginal effects. 5. analysis the financial literacy question set, coding of variables and descriptive statistics are provided in table 2. the responses to the three financial literacy questions show that more respondents choose the correct answer for compound interest (71.9% of women and 80% of men), followed by inflation (42.3% women and 65.7% men), and diversification (29.1% women and 58.6% men). the percentage of wrong answers is marginally higher for men for all three questions, with the highest percentage of wrong answers given for inflation (15.7% of men and 14.8% of women). overall, we observe knowledge of the concept of diversification is particularly lacking, followed closely by knowledge of inflation. there is a marked gender difference in the responses to “unsure.” women chose the “unsure” option more frequently than men. for compound interest, the difference is over 10 percentage points (16.8% vs. 5.7%), for inflation the difference is over 15 percentage points (37.2% vs. 15.7%), and for diversification, it is nearly 30 percentage points (60.2% vs. 31.4%). what is interesting is that the ratio of women choosing “unsure” relative to men is falling as the level of difficulty increases. for compound interest, the ratio is 2.9 (0.168/ 0.057), for inflation it is 2.3 (0.372/0.157), and for diversification it is 1.9 (0.602/0.314). 64 t. west et al. / financial services review 31 (2023) 55–71 the descriptive statistics for self-perceived numerical ability, risk aversion, and confidence are provided in table 3. men report higher levels of ability in num, and fin, while women report higher levels in read and writ. men are much more willing to take financial risks (24.3% of men report taking above-average risk compared with 6.1% of women) and have more confidence than women with managing finances (31.4% of men compared with 21.9% of women). the probit model estimates, including coefficients, robust standard errors, and levels of significance are reported in table 4. the first model reported (1) is on the total number of uns responses. this is followed by the “unsure” responses to each of the three financial literacy questions such as with regard to interest (intuns) (2), inflation (infuns) (3), and diversification (divuns) (4). the pseudo r2 shows the best fit is for model (4), followed by (1), (3), and (2). the regressions show that being wom is the overwhelmingly dominant characteristic for choosing “unsure.” the coefficients for wom are significant at the 0.05 level or higher, and coefficients range from 0.510 to 0.706, indicating that if a non-response is recorded there is more than a 50% chance the respondent is a woman. self-perceived ability with financial knowledge, fin, is significant and negative in models (1), (3), and (4), increasing in magnitude in the model (4). thus, respondents who recall having better financial knowledge than their peers at school are less likely to select a nonresponse answer to a financial knowledge question. hence, we infer, as other studies do, that table 2 descriptive statistics of financial literacy questions financial literacy questions women men compound interest if you invested $100 today and the interest rate was 2% per year your bank account balance after five years would be exactly $102. yes 0.066 0.114 no 0.719 0.800 unsure (intuns) 0.168 0.057 prefer not to answer 0.046 0.029 after 1 year you would be able to buy more than today if you invested $100 in your bank account today at an interest rate of 1% per year when inflation is 2% per year. yes 0.148 0.157 no 0.423 0.657 unsure (infuns) 0.372 0.157 prefer not to answer 0.056 0.029 diversification buying shares in a single company usually provides a safer return than buying units in a managed share fund. yes 0.056 0.100 no 0.291 0.586 unsure (divuns) 0.602 0.314 prefer not to answer 0.051 0.000 total unsure responses (uns) 0 0.321 0.571 1 0.327 0.343 2 0.240 0.071 3 0.112 0.014 t. west et al. / financial services review 31 (2023) 55–71 65 having an early interest in money enables people to acquire and build knowledge, especially regarding key financial concepts. further to this point, risk is significant only in the model (4). remembering that diversification is the most difficult question and requires an understanding of managing risk through diversification, it makes sense that those who indicate a willingness to take risks are willing to select an answer response rather than a non-response. in this sample, the risk parameter may proxy for both risk-taking knowledge and risk-taking behavior, or it is hard to distinguish between the two. numerical ability is significant in model (2) with a negative coefficient. respondents who have higher levels of self-perceived ability in mathematics are less likely to record an “unsure” response to the question with the least difficulty—interest. in this model for the “unsure” response to the interest question, only wom and num are significant. finally, table 3 descriptive statistics of explanatory parameters parameters women men self-perceived ability maths (num) 5 much better 0.250 0.386 4 better 0.321 0.243 3 about the same 0.281 0.214 2 worse 0.133 0.129 1 much worse 0.015 0.029 reading (read) 5 much better 0.485 0.286 4 better 0.270 0.429 3 about the same 0.199 0.271 2 worse 0.046 0.014 1 much worse 0.000 0.000 writing (writ) 5 much better 0.403 0.300 4 better 0.270 0.257 3 about the same 0.250 0.357 2 worse 0.077 0.071 1 much worse 0.000 0.000 financial knowledge (fin) 5 much better 0.082 0.071 4 better 0.194 0.271 3 about the same 0.571 0.543 2 worse 0.138 0.100 1 much worse 0.015 0.014 risk aversion (risk) 5 i take substantial financial risks expecting to earn substantial returns 0.005 0.100 4 i take above-average financial risks expecting to earn above-average returns 0.061 0.243 3 i take average financial risks expecting average returns 0.306 0.357 2 i a.m. not willing to take any financial risks 0.367 0.214 1 i never have any spare cash 0.260 0.086 overconfidence (conf) 6 describes me completely 0.219 0.314 5 describes me very well 0.321 0.386 4 somewhat describes me 0.332 0.229 3 describes me very little 0.082 0.057 2 does not describe me at all 0.041 0.014 1 not applicable 0.005 0.000 66 t. west et al. / financial services review 31 (2023) 55–71 age records a significant and negative relationship with uns in the model (1). therefore, older people are less likely to record an “unsure” response. 6. discussion investigation into the disparity between women and men in measuring financial literacy performance is important. women tend to live longer than men, have lower earnings, and have fewer retirement savings and are thus a vulnerable cohort (cassells et al., 2009). the research documenting low levels of financial literacy provides evidence that investing in financial literacy is one pathway to improving saving and spending behaviors and overall financial stability (oecd, 2020b). in this study, we take a closer look at how financial literacy is measured. we draw our hypothesis that women make more non-responses than men from the decision science literature, and find that more women select the non-response option than men in financial literacy questions. most men select an option other than “unsure” (57%); whereas, only 32% of women select an option other than “unsure.” the proportion of women who select unsure responses is about double that of men for all three questions. we infer from the regression results that having an interest in money matters at school age is important. an understanding of risk concepts or risk-taking behavior is also important for selecting an answer as opposed to a non-response option. this confirms prior work by charness and gneezy (2012). we were surprised that early confidence does not translate table 4 probit model parameter estimates of non-responses parameter (1) (2) (3) (4) uns intuns infuns divuns fem 0.510** 0.695** 0.686*** 0.613 0.198 0.302 0.224 0.200 num �0.064 �0.205** �0.064 0.007 0.084 0.106 0.087 0.085 read 0.061 �0.038 0.078 0.062 0.156 0.183 0.157 0.152 writ 0.035 �0.070 �0.111 �0.002 0.139 0.171 0.144 0.138 fin �0.263** �0.018 �0.272** �0.305 0.115 0.148 0.123 0.115 risk �0.129 0.057 �0.048 �0.153 0.088 0.107 0.093 0.087 conf �0.051 �0.078 �0.033 �0.049 0.080 0.094 0.081 0.080 age �0.130* �0.071 �0.102 �0.105 0.079 0.096 0.083 0.078 inc �0.065 �0.025 �0.081 �0.048 0.056 0.070 0.060 0.056 cons 1.592* �0.013 0.851 1.237 0.639 0.772 0.650 0.635 pseudo r2 0.095 0.071 0.087 0.100 lr x2 33.74*** 15.16*** 28.97*** 36.89 notes. *p< .10, **p< .05, ***p< .01. t. west et al. / financial services review 31 (2023) 55–71 67 into current money confidence as the conf parameter is not significant in any models. however, this may be because it is an accurate self-assessment of the respondent’s ability. 7. implications for researchers and educators, more work is to be done on understanding gender bias in financial literacy metrics. is it the style of questions that is problematic for women? is it the language used? perhaps open-ended questions would elicit a more contextual understanding of knowledge about money matters by allowing qualitative aspects of respondents’ social experience, values, and cultural diversity to be included (hunter & sawatzki, 2019). or, perhaps in adulthood, women are focused on financial management tasks different from investing, but nonetheless important. thus, a sophistication may exist that is not captured by financial literacy metrics such as balancing budgets and making financial decisions for the immediate and extended family (blue & o’faircheallaigh, 2018). in addition, work continues in schooling on how to make effective links between knowledge and financial decisions in practice. some researchers have raised concerns on the efficacy of financial education in schools and there is still work to be done on the integration of money matters across a range of topics (not just mathematics) and incorporation of social justice and cultural issues (blue et al., 2018; sawatzki, 2014). gender-inclusive measurement of knowledge, literacy, and capability in financial matters is one step in liberating women from predesigned social structures and norms that may be enforcing myths about women’s capability in the realm of finance. this study has found that a shift in mindset may be needed when developing methodologies for measuring financial literacy to account for gender differences, and when interpreting the results of biased measurement instruments. these results contribute to this urgent area of research in framing a more inclusive financial research agenda and finance system for current and future generations. acknowledgments griffith university new researcher grant supported the data collection for this study. references agnew, s., & cameron-agnew, t. (2015). the influence of consumer socialisation in the home on gender differences in financial literacy: gender impact of financial socialisation in home. international journal of consumer studies, 39, 630–638. https://doi.org/10.1111/ijcs.12179 alcon, a. (1999). financial planning and the mature woman. journal of financial planning, 12, 82–88. anthes, w. l., & most, b. w. (2000). frozen in the headlights: the dynamics of women and money. journal of financial planning, 13, 130–142. baldiga, k. (2014). gender differences in willingness to guess. management science, 60, 434–448. https://doi. org/10.1287/mnsc.2013.1776 barber, b., & odean, t. (2001). boys will be boys: gender, overconfidence and common stock investment. the quarterly journal of economics, 116, 261–292. https://doi.org/10.1162/003355301556400 68 t. west et al. / financial services review 31 (2023) 55–71 blue, l., & o’faircheallaigh, c. (2018). indigenous autonomy and financial decision-making in communities. financial planning research journal, 4, 39–50. blue, l., o’brien, m., & makar, k. (2018). exploring the classroom practices that may enable a compassionate approach to financial literacy education. mathematics education research journal, 30, 143–164. https://doi. org/10.1007/s13394-017-0223-5 bottazzi, l., & lusardi, a. (2016).gender differences in financial literacy: evidence from pisa data in italy. available at https://institute.eib.org/wp-content/uploads/2016/10/gender-diff.pdf bucher-koenen, t., lusardi, a., alessie, r., & van rooij, m. (2017). how financially literate are women? an overview and new insights. journal of consumer affairs, 51, 255–283. https://doi.org/10.1111/joca.12121 cassells, r., miranti, r., nepal, b., & tanton, r. (2009). she works hard for the money: australian women and the gender divide. amp natsem income and wealth report, 22. available at http://hdl.voced.edu.au/ 10707/153977 charness, g., & gneezy, u. (2012). strong evidence for gender differences in risk taking. journal of economic behavior & organization, 83, 50–58. https://doi.org/10.1016/j.jebo.2011.06.007 chen, h., & volpe, r. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. https://doi.org/10.1016/s1057-0810(99)80006-7 chen, h., & volpe, r. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. coffman, k. b. (2014). evidence on self-stereotyping and the contribution of ideas. the quarterly journal of economics, 129, 1625–1660. https://doi.org/10.1093/qje/qju023 croson, r., & gneezy, u. (2009). gender differences in preferences. journal of economic literature, 47, 448– 474. https://doi.org/10.1257/jel.47.2.448 dilmaghani, m. (2020). gender differences in performance under time constraint: evidence from chess tournaments. journal of behavioral and experimental economics, 89, 101505. https://doi.org/10.1016/j. socec.2019.101505 driva, a., lührmann, m., & winter, j. (2016). gender differences and stereotypes in financial literacy: off to an early start. economics letters, 146, 143–146. https://doi.org/10.1016/j.econlet.2016.07.029 espinosa, m. p., & gardeazabal, j. (2010). do students behave rationally in multiple choice tests? evidence from a field experiment. journal of mathematical psychology, 54, 415–425. https://doi.org/10.1016/j.jmp.2010. 06.001 fennema, e., & sherman, j. (1977). sex-related differences in mathematics achievement, spatial visualization and affective factors. american educational research journal, 14, 51–71. https://doi.org/10.3102/00028312014001051 fonseca, r., mullen, k. j., zamarro, g., & zissimopoulos, j. (2012). what explains the gender gap in financial literacy? the role of household decision-making. the journal of consumer affairs, 46, 90–106. https://doi. org/10.1111/j.1745-6606.2011.01221.x gerrans, p., & heaney, r. (2019). the impact of undergraduate personal finance education on individual financial literacy, attitudes and intentions. accounting & finance, 59, 177–217. https://doi.org/10.1111/acfi.12247 gerrans, p., speelman, c., & campitelli, g. (2014). the relationship between personal financial wellness and financial wellbeing: a structural equation modelling approach. journal of family and economic issues, 35, 145–160. https://doi.org/10.1007/s10834-013-9358-z goldsmith, r. e., & goldsmith, e. b. (1997). gender differences in perceived and real knowledge of financial investments. psychological reports, 80, 236–238. https://doi.org/10.2466/pr0.1997.80.1.236 halpern, d. f., benbow, c. p., geary, d. c., gur, r. c., hyde, s. h., & gernsbacher, m. a. (2007). the science of sex differences in science and mathematics. psychological science in the public interest: a journal of the american psychological society, 8, 1–51. https://doi.org/10.1111/j.1529-1006.2007.00032.x hsu, j. w. (2011). aging and strategic learning: the impact of spousal incentives on financial literacy. finance and economics discussion series 2011–53.washington, dc: board of governors of the federal reserve system. hung, a. a., parker, a. m., & yoong, j. k. (2009). defining and measuring financial literacy. rand labor and population working paper series wr-708. available at https://www.rand.org/content/dam/rand/pubs/ working_papers/2009/rand_wr708.pdf t. west et al. / financial services review 31 (2023) 55–71 69 hunter, j., & sawatzki, c. (2019). discovering diverse students’ funds of knowledge related to finance: pasifika students in new zealand. mathematics education research journal, 31, 419–439. https://doi.org/10.1007/ s13394-019-00259-0 kahneman, d. (2003). a perspective on judgement and choice: mapping bounded rationality. the american psychologist, 58, 697–720. karunarathne, w., & gibson, j. (2014). financial literacy and remittance behaviour of skilled and unskilled immigrant groups in australia. journal of asian economics, 30, 54–62. https://doi.org/10.1016/j.asieco.2013.12.004 kimball, m. (1989). a new perspective on women’s math achievement. psychological bulletin, 105, 198–214. https://doi.org/10.1037/0033-2909.105.2.198 knoll, m. a., & houts, c. r. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. journal of consumer affairs, 46, 381–410. https://doi.org/10.1111/ j.1745-6606.2012.01241.x larisa, l. e., njo, a., & wijaya, s. (2021). female workers’ readiness for retirement planning: an evidence from indonesia. review of behavioral finance, 13, 566–583. https://doi.org/10.1108/rbf-04-2020-0079 leder, g. c. (1992). mathematics and gender: changing perspectives. in d. a. grouws (ed.), handbook of research on mathematics teaching and learning (pp. 597–622). new york, ny: macmillan. loewen, j., rosser, p., & katzman, j. (1988). gender bias in sat items. paper presented at the annual meeting of the american educational research association. new orleans, la. lusardi, a., michaud, p., & mitchell, o. s. (2017). optimal financial knowledge and wealth inequality. the journal of political economy, 125, 431–477. lusardi, a., & mitchell, o. s. (2009). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness.working paper 15350. national bureau of economic research. lusardi, a., & mitchell, o. s. (2011). financial literacy around the world: an overview. journal of pension economics & finance, 10, 497–508. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. https://doi.org/10.1111/j.1745-6606.2010.01173.x lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lusardi, a., & tufano, p. (2009a). debt literacy, financial experiences, and overindebtedness. working paper 14808. national bureau of economic research. lusardi, a., & tufano, p. (2009b). teach workers about the perils of debt. harvard business review, 22–24. november, niederle, m., & vesterlund, l. (2007). do women shy away from competition? do men compete too much? the quarterly journal of economics, 122, 1067–1101. niederle, m., & vesterlund, l. (2010). explaining the gender gap in math test scores: the role of competition. journal of economic perspectives, 24, 129–144. https://doi.org/10.1257/jep.24.2.129 oecd. (2017). g20/oecd infe report on adult financial literacy in g20 countries. report presented at the 2017 g20 leader’s summit meeting in hamburg on july 8, 2017. available at http://www.oecd.org/daf/fin/ financial-education/g20-oecd-infe-report-adult-financial-literacy-in-g20-countries.pdf oecd. (2020a). oecd/infe 2020 international survey of adult financial literacy. available at https://www. oecd.org/financial/education/oecd-infe-2020-international-survey-of-adult-financial-literacy.pdf oecd. (2020b). pisa 2018 results (volume iv): are students smart about money? paris: oecd publishing. pekkarinen, t. (2015). gender differences in behaviour under competitive pressure: evidence on omission patterns in university entrance examinations. journal of economic behavior & organization, 115, 94–110. https://doi.org/10.1016/j.jebo.2014.08.007 preston, a. c., & wright, r. e. (2019). understanding the gender gap in financial literacy: evidence from australia. economic record, 95, 1–29. https://doi.org/10.1111/1475-4932.12472 riener, g., & wagner, v. (2017). shying away from demanding tasks? experimental evidence on gender differences in answering multiple-choice questions. economics of education review, 59, 43–62. https://doi.org/ 10.1016/j.econedurev.2017.06.005 70 t. west et al. / financial services review 31 (2023) 55–71 rothwell, d., & wu, s. (2019). exploring the relationship between financial education and financial knowledge and efficacy: evidence from the canadian financial capability survey. journal of consumer affairs, 53, 1725– 1747. https://doi.org/10.1111/joca.12259 sawatzki, c. (2014). connecting social and mathematical thinking: using financial dilemmas to explore children’s financial problem-solving and decision-making (doctoral dissertation). monash university. schuhen, m., & schurkmann, s. (2014). construct validity of financial literacy. international review of economics education, 16, 1–11. https://doi.org/10.1016/j.iree.2014.07.004 stumpf, h., & stanley, j. (1996). gender-related differences on the college board’s advanced placement and achievement tests, 1982–1992. journal of educational psychology, 88, 353–364. https://doi.org/10.1037/ 0022-0663.88.2.353 tannenbaum, d. (2012). do gender differences in risk aversion explain the gender gap in sat scores? uncovering risk attitudes and the test score gap. (unpublished paper) chicago: university of chicago. timmermann, s. (2000). planning for the future: a focus on mature women. journal of financial service professionals, 54, 22–24. volpe, r., chen, h., & pavlicko, j. (1996). investment literacy among college students: a survey. financial practice and education, 6, 86–94. wade, a. (2013). boys get paid more pocket money. new zealand herald, october 15. available at http:// www.nzherald.co.nz/lifestyle/news/article.cfm?c_id=6&objectid=11140064 ward, a. f., & lynch, j. g. (2019). on a need-to-know basis: how the distribution of responsibility between couples shapes financial literacy and financial outcomes. journal of consumer research, 45, 1013–1036. https://doi.org/10.1093/jcr/ucy037 west, t., & de zwaan, l. (2020). financial socialisation of australian university students: differences in gender. consumer interests annual, 66, 1–14. west, t., & worthington, a. c. (2013). a canonical correlation analysis of financial risk-taking by australian households. consumer interests annual, 59, 1–4. west, t., & worthington, a. c. (2014). personal attributes and financial risk-taking in australia. jassa, 25, 25–31. west, t., & worthington, a. c. (2018). synthesizing an index of financial literacy using not-for-purpose survey data. economic notes, 47, 331–352. https://doi.org/10.1111/ecno.12107 wilkins, r., & lass, i. (2018). the household, income and labour dynamics in australia survey: selected findings from waves 1 to 16. melbourne institute: applied economic & social research, university of melbourne. available at https://melbourneinstitute.unimelb.edu.au/__data/assets/pdf_file/0009/2874177/ hilda-report_low-res_10.10.18.pdf worthington, a. c. (2013). financial literacy and financial literacy programmes in australia. journal of financial services marketing, 18, 227–240. https://doi.org/10.1057/fsm.2013.18 yu, s., alper, h. e., nguyen, a., brackbill, r. m., turner, l., walker, d. j., maslow, c. b., & zweig, k. c. (2017). the effectiveness of a monetary incentive on survey offer response rates and response completeness in a longitudinal study. bmc medical research methodology, 17, 77–91. zissimopoulos, j.m., karney, b., & rauer, a. (2008). marital histories and economic well-being. michigan retirement research center research paper no. wp, 180. ann arbor, mi: university of michigan, michigan retirement research center. t. west et al. / financial services review 31 (2023) 55–71 71 financial services review, 32(1) 1 a dynamic analysis of the impact of household portfolio allocation decisions on the demand for life insurance ning wang1, yiling deng2, and ruohan wu3 abstract prior research on the demand for life insurance in household portfolio holdings has not made a clear distinction between portfolio shifts resulting from active allocation decisions and those resulting from passive acceptance. our study examines the relationship between household portfolio allocation decisions and the demand for life insurance in a dynamic setting, using panel data before and after the 2008 financial crisis. the study provides the first evidence that household decisions to invest in cash and cash equivalents, bonds, retirement assets, and pay off debts significantly affect life insurance ownership. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation wang, n., deng, y., & wu, r. (2024). a dynamic analysis of the impact of household portfolio allocation decisions on the demand for life insurance. financial services review, 32(1), 1-28. introduction household financial decisions are complex and interdependent, and central to the functioning of the financial system; however, an important asset class for households that has received comparatively less attention is life insurance (gomes et al., 2021). life insurance can provide households with financial protection and help them pay off debts in the event of the premature death of a wage earner. whole life insurance, in particular, allows households to borrow against its cash value through a loan option. if households reduce or drop life insurance coverage during an economic downturn, they may experience financial hardships that can have significant economic consequences (scott & gilliam, 2014). 1 corresponding author (nwang@ung.edu). mike cottrell college of business, university of north georgia, dahlonega, usa. 2 college of business, university of central arkansas, conway, usa. 3 mike cottrell college of business, university of north georgia, dahlonega, usa. household financial decisions should be assessed and analyzed within the broader context of the entire portfolio, rather than focusing on individual assets, securities, or accounts (rabbani et al., 2017; gomes et al., 2021). during a recession, because of rising unemployment and declining income, households have a tight budget for various financial assets including stocks, bonds, and life insurance, and are less likely to pay off debts. the holdings of life insurance can be associated with the holdings of other financial assets and debts as households rebalance and adjust their financial portfolios. the allocation of household financial portfolios is essential in explaining the purchase of life insurance (lin & grace, 2007; shi et al., 2015). accordingly, it is https://creativecommons.org/licenses/by-nc/4.0/ mailto:nwang@ung.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(1) 2 crucial to explore household choices of life insurance within the framework of household portfolios. previous studies have addressed the importance of considering household portfolio holdings when analyzing the determinants of the demand for life insurance. fortune (1973) indicates that life insurance is a substitute for primary financial assets. in contrast, outreville (2013) suggests a positive relationship between the demand for life insurance and household holdings of primary financial assets. luciano et al. (2016) explain this relationship by the increased participation in both the stock market and life insurance market due to financial market proximity and familiarity. lin and grace (2007) find that individual retirement accounts are complements to total life insurance for youngand middle-aged households. in addition to the holdings of financial assets, prior research has suggested that debt holdings can impact the demand for life insurance (ferber & lee, 1980). while lin and grace (2007) note that the effect of debts on life insurance could be ambiguous, hau (2000) documents a positive effect, implying that life insurance benefits can serve as immediate cash to pay debts and costs of death. more recently, wang (2023) finds that changes in household portfolio holdings were more significant than life events in determining life insurance ownership. changes in household portfolio holdings may result from the passive acceptance of portfolio shifts or from active decision-making for portfolio allocation and rebalancing (bricker et al, 2011). passive acceptance refers to situations where households simply accept portfolio shifts driven by external factors, such as changes in asset prices. this could result in household portfolios that may not necessarily align with their risk and return preferences. on the other hand, portfolio allocation involves households actively making decisions to rebalance the distribution of diverse financial assets within their portfolios to fulfill their financial needs. this process involves the consideration of factors such as financial objectives, risk tolerance, and financial literacy or guidance. under standard models of portfolio choices, households in efficient markets are expected to invest passively with a low frequency of portfolio rebalancing (gomes et al., 2021). nevertheless, campbell et al. (2009) demonstrate that households trade more frequently and excessively than is good for their financial wellbeing. during the 2008 financial crisis, bricker et al. (2011) indicate that the majority of families passively accepted changes in portfolio shares. hoopes et al. (2016) demonstrate that households in the very highest income groups rebalanced financial portfolios more actively than others throughout the 2008 financial crisis. in response to business cycles, active portfolio allocation enables households to adjust their financial strategies as circumstances evolve. this guarantees that household portfolios stay aligned with their evolving needs. consequently, it becomes important to differentiate between choices of active portfolio allocation and passive acceptance. however, this distinction has not been addressed within the existing body of insurance literature. we aim to address this significant gap in the literature by investigating the relationship between active decision-making for household portfolio allocation and the demand for life insurance. to the best of our knowledge, our study is the first to explore this correlation. using data from the survey of consumer finances (scf), we examine different portfolio allocation decisions that can have an impact on life insurance ownership and coverage. this unique focus provides valuable implications for financial advisors, policymakers, and households to enhance their overall financial planning strategies. active engagement in portfolio allocation is expected to empower households with betterinformed financial decisions, insights into diverse investment opportunities, and a deeper comprehension of overall household financial planning. additionally, our study contributes to the existing literature by highlighting a distinction of life insurance demand based on type. we distinguish between term life insurance, primarily designed as a pure safeguard against potential future income loss, and whole life insurance, which provides an additional vehicle for investment and wang et al. 3 tax-efficient asset accumulation. notably, prior research has emphasized the importance of treating the purpose and features of term life insurance differently from those of whole life insurance (lin & grace, 2007; grable, 2016; heo et al., 2021). moreover, our study makes a valuable contribution to the existing literature by utilizing panel data to explore the dynamics of life insurance demand. liebenberg et al. (2012) suggest that a dynamic analysis based on panel data may provide the basis for a better understanding of the determinants for life insurance demand. using the 2007-2009 scf panel data, our paper investigates the impact of changes in household portfolio allocation decisions on life insurance ownership and coverage decisions in a dynamic setting. specifically, life insurance ownership decisions examined in this paper include initiating, increasing, decreasing, and dropping life insurance holdings, and the dollar amount of these changes are included in life insurance coverage decisions (also see liebenberg et al., 2012; wang, 2023). household portfolio allocation decisions include the decisions of debt repayment and investment in cash and cash equivalents, bonds, stock, and retirement assets. we apply a two-part cragg model with probit and truncated regressions to identify the variables that significantly impact life insurance ownership and coverage decisions. lastly, our analysis is based on the context of the 2008 financial crisis period, which provides valuable insights into household portfolio decision-making in response to business cycles. bricker et al. (2011) report that during the financial crisis, the ownership and median value of bonds and life insurance increased, while stocks experienced sharp declines in median value. the life insurance markets were more severely impacted by the financial crisis as they had relatively larger asset portfolios than other insurance markets (baranoff & sager, 2011). our results indicate that, during the financial crisis, the ownership of term life insurance serves as a replacement for investments in cash and cash equivalents yet complements investments in bonds. our results also indicate that term (whole) life insurance ownership is likely to be a complement (substitute) rather than a substitute (complement) for investment in retirement assets. we also find that households that pay off debts are more likely to initiate or increase whole life insurance ownership. these findings can provide valuable insights for financial planning practitioners, life insurance agents, and social policymakers, by guiding them to deliver adaptive and comprehensive financial services to households. particularly relevant during economics downtowns, these insights can assist households that are adjusting their decisions of portfolio allocation to align with changing situations. the remainder of the article is organized as follows. section 2 provides a further literature review of the demand for life insurance. section 3 outlines the methodology used to examine the impact of household portfolio allocation decisions on life insurance demand in the 2008 financial crisis, using panel data. section 4 presents summary statistics of the data and empirical results. finally, section 5 concludes the study and discusses further implications for financial practitioners and agents. literature review prior research has examined how household demographic characteristics and economic status influence the demand for life insurance (ferber & lee, 1980; miller, 1985; bernheim, 1991; lin & grace, 2007; inkmann & michaelides, 2012; luciano et al., 2016; wang, 2019). the results of these studies have been mixed in terms of financial assets, wealth, and retirement, as noted by hau (2000), zietz (2003), and outreville (2014). grable (2016) and heo and grable (2017) suggest that if a financial planner leans only on the functional characteristics of life insurance purchases, the consumer’s needs may ultimately not be fulfilled. meanwhile, heo (2020) and heo et al. (2021) examine the factors related to life insurance ownership by type, focusing on the role of financial status characteristics, psychological characteristics, and demographic characteristics. hau (2000) particularly highlights that financial portfolios may have a greater influence than household demographic characteristics in shaping the demand for life insurance. it is evident in literature that the allocation of financial services review, 32(1) 4 financial assets can have an impact on the demand for life insurance. for instance, fortune (1973) indicates that life insurance is a substitute for primary financial assets. households may invest more in primary financial assets during bullish financial markets, depressing the flow into life insurance. headen and lee (1974) propose a household portfolio model with four interrelated components, including primary securities, cash and cash equivalents, savings, and life insurance sales. their findings provide limited evidence that life insurance increases with savings. ferber and lee (1980) argue that the purchase of life insurance is determined by household spending and saving habits, financial assets, and debts. brown (1999) presents that households tend to hold term life insurance and private annuities simultaneously. hau (2000) investigates the impact of life insurance as a liquid financial asset on an insured's death, and finds that saving accounts, bonds, and stocks negatively impact life insurance holdings. the study also reveals evidence that debts have a positive effect on life insurance. lin and grace (2007) identify the impact of various types of financial assets on both term and whole life insurance. more recently, outreville (2013) finds a positive relationship between primary financial assets and life insurance, which is attributed to the good performance of financial markets during a booming economy with high household savings. shi et al. (2015) suggest that life insurance complements, rather than substitutes, other financial assets in a household's asset allocation decisions. in addition to financial portfolios, previous research has found that household wealth, as measured by net worth, is an important determinant of the demand for life insurance. lewis (1989) suggests that life insurance is negatively associated with net worth. nevertheless, bernheim (1991), hau (2000), and eisenhauer and halek (1999) have reported a positive correlation between household wealth and life insurance purchases, attributing this to bequest motives and the increasing absolute risk aversion hypothesis. maremont and scism (2010) report that the ownership of cash-value life insurance dropped dramatically over the last decade, but the total face value fell at a slower rate. heo et al. (2013) indicate that cash-value life insurance acts as a complement to, rather than substitute for, wealth. mulholland et al. (2016) predict that cash-value life insurance ownership will continue to decline among younger households, while the wealthiest households have increased it as an estate planning instrument. luciano et al. (2016) find that the ratio of income to net worth negatively affects the demand for term life insurance. they suggest that focusing solely on income variables can be misleading, as wealth is playing an increasing role in determining the demand for life insurance. moreover, previous research has examined the impact of retirement assets, including social security, on the demand for life insurance. pissarides (1980) indicates that retirement savings and bequests are dependent on the ability to purchase life insurance. bernheim (1991) finds that social security benefits lead to a drop in the acquisition of annuities but an increase in life insurance purchases. hubener et al. (2016) demonstrate that social security rules and family risk have important effects on the optimal life cycle household saving and asset allocation patterns, retirement decisions, and life insurance purchases. limited literature has explored the demand for life insurance during economic recessions. the 2007-2009 financial crisis severely impacted global insurance markets, with the segments of annuities and life insurance hit harder than health lines, as they had relatively larger asset portfolios (baranoff & sager, 2011). during the crisis, low long-term interest rates, which serve as the valuation basis to determine premiums, guaranteed rates of return, profit-sharing, and policy reserves, created significant financial problems and pressure on profits for life insurers nationwide (holsboer, 2000). this resulted in an increased use of paid-up options or high lapses on life insurance policies and a large drop in the demand for new policies. for households, the financial crisis led to rising unemployment and declining value of financial assets and household wealth. swiss re (2009) reports that sales of equity-linked products declined tremendously during the financial crisis, while non-equitylinked savings products, including fixed annuities and traditional life savings, continued to increase. wang et al. 5 after the 2008 financial crisis, households demonstrated an increased risk aversion attitude, with higher precautionary savings (bricker et al., 2011). households also tended to have better performance in their behaviors of budgeting, saving, and spending, and have more frequent performance of commonly recommended financial practices, including holding adequate insurance (o'neill & xiao, 2012). additionally, there are insufficient research attempts that employ panel data to investigate the determinants of the demand for life insurance in a dynamic context. liebenberg et al. (2012) suggest that a dynamic analysis based on panel data may be better suited to explain the determinants of the demand for life insurance than a static analysis based on cross-sectional data. they analyze the 1983-1989 scf panel data and find that life events, such as finding a new job and becoming unemployed, have a significant and dynamic impact on changes in life insurance holdings. the same dataset is used by bertaut (1998) to examine the determinants of stock ownership decisions in a dynamic setting at the household level, and also used by liebenberg et al. (2010) to explore policy loans of whole life insurance. more recently, wang (2023) uses the 2007-2009 scf panel data and reveals a positive relationship between changes in the holdings of primary assets and the ownership choices of life insurance during the financial crisis. changes in household portfolio holdings may result from the passive acceptance of portfolio shifts driven by changes in asset prices, or from the active decision-making to rebalance various financial assets in household portfolios (bricker et al, 2011). bricker et al. (2011) indicate that the majority of families passively accepted changes in portfolio shares driven by changes in asset 4 see liebenberg et al. (2012) for the advantages of studying changes in financial portfolios using the scf panel data, rather than using other household panel data sources including the panel study of income dynamics, consumer expenditure surveys, survey of income and program participation, and health and retirement survey. 5 see the official scf website, https://www.federalreserve.gov/econres/scf-previoussurveys.htm, for more information. prices during the financial crisis with a low frequency of portfolio rebalancing. hoopes et al. (2016) find that households in the very highest income groups rebalanced their financial portfolios more actively than others throughout the financial crisis, implying that portfolio rebalancing decisions vary across households during recessions. however, prior insurance literature has not distinguished between household portfolio shifts resulting from passive acceptance and those resulting from active decisions of portfolio restructuring. our study aims to bridge this gap by focusing on examining the impact of households' active portfolio allocation decisions on the demand for life insurance. data and methodology data this paper examines household portfolio allocation decisions that are hypothesized to impact the demand for life insurance in a dynamic setting during the 2008 financial crisis. we use the scf data, a nationwide household survey that has been used extensively in prior literature to explore the demand for life insurance (hau, 2000; lin & grace, 2007; liebenberg et al., 2012; glumov, 2013; scott & gilliam, 2014; wang, 2023)4. the scf data includes information about household characteristics and economic status, as well as the allocation of household portfolios. the data oversamples wealthier households which are expected to hold a variety of financial assets and rebalance their financial portfolios as situations evolve (see glumov, 2013; hoopes et al., 2016). in the history of the scf, only two panel datasets are available: one spanning from 1983 to 1989, and the other from 2007 to 20095. our study utilizes the more recent dataset, as it holds greater relevance and value6. in the 2007 6 we have omitted the years spanning from 1983 to 1989 from our paper due to the lack of relevant inquiries regarding active decision-making for financial portfolios within the 1983-1989 scf panel survey. the 1983-1989 scf codebook can be accessed through this website link, https://www.federalreserve.gov/econres/files/1989p_ codebk89p.txt. for the specific inquiries regarding household portfolio allocation decisions discussed in this paper, we have sourced data from the 2007-2009 https://www.federalreserve.gov/econres/scf-previous-surveys.htm https://www.federalreserve.gov/econres/scf-previous-surveys.htm https://www.federalreserve.gov/econres/files/1989p_codebk89p.txt https://www.federalreserve.gov/econres/files/1989p_codebk89p.txt financial services review, 32(1) 6 scf, 89% (3,857) of eligible households agreed to complete a panel interview in 2009, which was included in the 2007-2009 scf panel data set. this panel data enables us to identify the decision-making for household portfolio allocation as well as the life insurance ownership and coverage for the same household since 2007. therefore, this data set is well-suited to exploring the dynamic effects of household portfolio allocation decisions on life insurance ownership and coverage decisions. methodology to study life insurance ownership: the probit models we apply the two-part cragg model to estimate the determinants of the demand for life insurance (cragg, 1971; liebenberg et al., 2012; wang, 2023), considering that many households in the survey sample do not hold life insurance. this modeling approach allows for a separate analysis of household decisions on life insurance ownership status and coverage amount. specifically, in the first part, we use probit regression models to identify the determinants of life insurance ownership decisions for the full sample of households. in the second part, we develop truncated regression models to explore the determinants of life insurance coverage decisions for the subsamples of households that experienced changes in life insurance ownership. the first part of the cragg model examines the likelihood of life insurance ownership changes, in relation to those that did not change their life insurance ownership decisions. the changes in life insurance ownership include initiating, increasing, decreasing, and dropping term or whole life insurance. the analysis focuses on four types of households that experienced changes in life insurance ownership from 2007 to 2009: 1) those that initiated or increased term life insurance, 2) those that initiated or increased whole life insurance, 3) those that decreased or dropped term life insurance, and 4) those that decreased or dropped whole life insurance. equations (1) to (4) represent the probit regression models that are developed to identify scf panel survey. see appendix a for these inquiries. while the data are not as current as desired, they are the only source of the necessary information for a the factors that influence these four life insurance ownership decisions. during the financial crisis, it is hypothesized that households that have made decisions to invest more in cash and cash equivalents, bonds, and retirement assets but less in stocks, pay off debts, spend less, become less aggressive, and invest less for the long term were more likely than other households to change their life insurance ownership decisions. in these regression models, newincrterm (newincrwhole) is an indicator variable that equals 1 for households that initiated a term (whole) life insurance policy or increased term (whole) life insurance coverage since 2007, and 0 otherwise. dropdecrterm (dropdecrwhole) is an indicator variable that equals 1 for households that dropped a term (whole) life insurance policy or decreased term (whole) life insurance coverage since 2007, and 0 otherwise. the variables of household portfolio allocation decisions include six indicator variables of whether households made decisions to invest more in cash and cash equivalents (invmorecash), invest more in bonds (invmorebonds), invest less in stock (invlessstock), invest more in retirement assets (invmorera), invest less in retirement assets (invlessra), and pay off debts (paydebt). three additional variables of financials decisions that have changed the ways households arrange their money or investments are also hypothesized to affect the demand for life insurance. moreriskaver is an indicator variable equal to 1 if households chose to have more conservative or disciplined investments, and 0 otherwise. spendless is an indicator variable equal to 1 if households chose to spend less, and 0 otherwise. invlesslong is an indicator variable equal to 1 if households chose to invest less for the long term, and 0 otherwise. control variables are the same in each equation and represent household characteristics and economic status examined in prior research (see liebenberg et al., 2012; heo & grable, 2017; heo, 2020; wang, 2023). specifically, we control for employment (work), household income (lnincome), household net worth (lnnetworth), broad sample of households that is crucial for our study. wang et al. 7 marital status (married), risk attitude (risk), stock holdings (stockshare), age (age3549, age5064, and age65_), the number of kids (kid), race (white), education (college), and homeownership (homeowner), with their definitions addressed in table 1. newincrtermi =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invmorerai +  5 moreriskaveri +  6 invlesslongi +  7 dropdecrwholei + control variables + 𝑖, (1) newincrwholei =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invmorerai +  5 paydebti +  6 spendlessi +  7 moreriskaveri +  8 invlesslongi +  9 dropdecrtermi + control variables + 𝑖, (2) dropdecrtermi =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invlessrai +  5 invlesslongi +  6 newincrwholei + control variables + 𝑖, (3) dropdecrwholei =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invlessrai +  5 paydebti +  6 spendlessi +  7 invlesslongi +  8 newincrtermi + control variables + 𝑖, (4) methodology to study life insurance coverage: the truncated models in the second part of the cragg model, we develop truncated regression models to investigate the subsamples of households that have experienced changes in life insurance ownership. as long as the life insurance ownership decisions have been made, households may adjust the amount of their life insurance coverage, resulting in four categories of life insurance coverage decisions: 1) the amount of newly purchased or increased term life insurance coverage, 2) the amount of newly purchased or increased whole life insurance coverage, 3) the amount of decreased or dropped term life insurance coverage, and 4) the amount of decreased or dropped whole life insurance coverage (liebenberg et al., 2012; wang, 2023). equations (5) to (8) describe the four truncated regression models, developed as the second part of the cragg model, to explore the determinants of the four categories of life insurance coverage decisions conditional on their ownership decisions. it is hypothesized that, among the subsamples of households that have made decisions to change their life insurance ownership, those that chose to invest more in cash and cash equivalents, bonds, and retirement assets but less in stocks, pay off debts, spend less, become less aggressive, and invest less for the long term made significantly more adjustments to their life insurance coverage, compared to other households during the financial crisis. here, lnnewincrterm (lnnewincrwhole) is the natural log of the amount of newly purchased or increased term (whole) life insurance coverage, specifically applicable to the households that have chosen to initiate or increase term (whole) life insurance since 2007, where newincrtermi = 1 (newincrwholei = 1). lndropdecrterm (lndropdecrwhole) is the natural log of the amount of decreased or dropped term (whole) life insurance coverage, specifically applicable to the households that have chosen to drop or decrease term (whole) life insurance since 2007, where dropdecrtermi = 1 (dropdecrwholei = 1). the variables of household portfolio allocation decisions and control variables are the same as the previous regressions described by equations (1) to (4). lnnewincrtermi =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invmorerai +  5 moreriskaveri +  6 invlesslongi +  7 lndropdecrwholei + control variables + 𝑖, (5) lnnewincrwholei =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invmorerai +  5 paydebti +  6 spendlessi +  7 moreriskaveri +  8 invlesslongi +  9 lndropdecrtermi + control variables + 𝑖, (6) financial services review, 32(1) 8 lndropdecrtermi =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invlessrai +  5 invlesslongi +  6 lnnewincrwholei + control variables + 𝑖, (7) lndropdecrwholei =  +  1 invmorecashi +  2 invmorebondsi +  3 invlessstocki +  4 invlessrai +  5 paydebti +  6 spendlessi +  7 invlesslongi +  8 lnnewincrtermi + control variables + 𝑖, (8) to address potential endogeneity issues arising from substitution effects, we conducted wald exogeneity tests for all four probit models and four truncated models 7 . these tests failed to reject exogeneity for any of the equations. hence, we were able to directly apply probit regressions and truncated regressions to identify the variables related to household portfolio allocation decisions that have a significant and dynamic impact on life insurance ownership and coverage decisions. in addition, we checked for robustness by incorporating instrument variables (iv) for dropdecrwhole, dropdecrterm, newincrwhole, and newincrterm into the probit regressions described by equations (1), (2), (3), and (4) respectively, through a two-stage process. in the first stage, for instance, we estimated the predicted values of dropdecrwhole using all the explanatory variables except newincrterm and the control variables specified in equation (4). in the second stage, we employed the predicted values derived from the first stage as the iv for dropdecrwhole to conduct the probit regression described by equation (1). this approach effectively addresses the endogeneity of the independent variable dropdecrwhole, while ensuring that the iv is not correlated with the dependent variable newincrterm in equation (1). it is noted that our robustness analyses demonstrated no significant deviations from the results we initially obtained regarding the determinants of the demand for life insurance8. results summary statistics summary statistics and variable definitions of the full sample are reported in table 1. it shows that around 28% of the households that were reinterviewed in 2009 had initiated or increased their holdings of term life insurance, and 18% had initiated or increased their holdings of whole life insurance since 2007. additionally, about 29% of the households that owned term life insurance in 2007 had dropped or decreased their coverage in 2009, while 19% of the households that owned whole life insurance in 2007 had dropped or decreased their coverage in 2009 (also see wang, 2023). table 2 presents summary statistics for the four subsamples of households that experienced changes in life insurance ownership since 2007. 7 for example, in equation (1), dropdecrwhole is included as an explanatory variable that could impact household decisions to purchase term life insurance (newincrterm) in equation (1). however, newincrterm is included as an explanatory variable that could impact household decisions to drop whole life insurance in equation (4). this inclusion of substitution effects in both equations may give rise to potential endogeneity concerns. the same as liebenberg et al. (2012), wald exogeneity test was applied to check for endogeneity in each equation. if the test indicates that those variables are exogenous, we can estimate the equation using the original probit or truncated model as they can provide efficient and consistent results. if the regressors are endogenous, the approach of two-stage least squares should be used for the equation. we have reported the results of the wald exogeneity tests in tables 3, 4, 5, and 6. 8 we conducted additional analysis to check for robustness. we ran the probit regressions described by equations (1) to (4) by adjusting the values of the control variables from 2007 to 2009, given that these four regressions provide the primary source of our key findings. it is noted that the results of this analysis also present no significant deviations from the results outlined in this paper, indicating that the analysis and the results of this paper are robust. the results of these two robustness analyses are presented in appendix b. wang et al. 9 table 1. variable definitions and summary statistics for life insurance ownership decisions (full sample) variables definitions mean std. dev. life insurance ownership decisions newincrterm equal to 1 for households that initiated a term life insurance policy or increased term life insurance coverage since 2007, and 0 otherwise 0.28 0.45 newincrwhole equal to 1 for households that initiated a whole life insurance policy or increased whole life insurance coverage since 2007, and 0 otherwise 0.18 0.39 dropdecrterm equal to 1 for households that dropped a term life insurance policy or decreased term life insurance coverage since 2007, and 0 otherwise 0.29 0.45 dropdecrwhole equal to 1 for households that dropped a whole life insurance policy or decreased whole life insurance coverage since 2007, and 0 otherwise 0.19 0.39 portfolio allocation decisions invmorecash equal to 1 for households that have made decisions to invest more in cds, other deposits, "cash" since 2007, and 0 otherwise 0.02 0.13 invmorebonds equal to 1 for households that have made decisions to invest more in tax-exempt bonds, treasury bills/bonds, other bonds since 2007, and 0 otherwise 0.01 0.11 invlessstock equal to 1 for households that have made decisions to invest less in stocks since 2007, and 0 otherwise 0.04 0.19 invmorera equal to 1 for households that have made decisions to invest more in retirement assets (ira, keogh, 401(k), roth, etc.) since 2007, and 0 otherwise 0.01 0.11 invlessra equal to 1 for households that have made decisions to invest less in retirement assets (ira, keogh, 401(k), roth, etc.) since 2007, and 0 otherwise 0.01 0.11 paydebt equal to 1 for households that have made decisions to pay off debt since 2007, and 0 otherwise 0.02 0.14 other financial decisions variables spendless equal to 1 if households that have made decisions to spend less, cut back since 2007, and 0 otherwise 0.16 0.37 moreriskaver equal to 1 if households that have made decisions to have more conservative or disciplined investments, be less risk/aggressive since 2007, and 0 otherwise 0.09 0.29 invlesslong equal to 1 if households that have made decisions to invest less for the long term since 2007, and 0 otherwise 0.01 0.12 financial services review, 32(1) 10 table 1 (continued). variable definitions and summary statistics for life insurance ownership decisions (full sample) control variables work equal to 1 if either spouse was employed in 2007, and 0 otherwise 0.76 0.43 lnincome the natural log of household income in 2007 11.57 2.01 lnnetworth the natural log of household net worth in 2007 12.12 4.36 married equal to 1 for married households in 2007, and 0 otherwise 0.69 0.46 risk equal to 1 if households preferred no financial risk in 2007, and 0 otherwise 0.30 0.46 stockshare the ratio of stock value to household wealth in 2007 5.06 39.02 age3549 equal to 1 if the age of household respondent was between 35 and 49, and 0 otherwise 0.30 0.46 age5064 equal to 1 if the age of household respondent was between 50 and 64, and 0 otherwise 0.34 0.47 age65_ equal to 1 if the age of household respondent was 65 or older, and 0 otherwise 0.21 0.41 kid the number of children in the household in 2007 0.88 1.20 white equal to 1 for white households, and 0 otherwise 0.77 0.42 college equal to 1 if either spouse had college education in 2007, and 0 otherwise 0.57 0.50 homeowner equal to 1 if households owned their primary residence in 2007, and 0 otherwise 0.76 0.43 observations 3857 regression models (1) (4) note: all data variables are taken or calculated from the 2007-2009 scf panel survey. see appendix a for more details. wang et al. 11 table 2. summary statistics for the life insurance coverage decisions (subsamples) variables m sd m sd m sd m sd lnnewincrterm 11.54 1.80 4.72 5.99 lndropdecrwhol e 3.02 5.17 11.15 2.36 lnnewincrwhole 11.07 2.37 2.83 4.97 lndropdecrterm 4.75 5.92 11.55 1.92 invmorecash 0.01 0.10 0.02 0.14 0.02 0.13 0.02 0.16 invmorebonds 0.01 0.12 0.01 0.11 0.01 0.09 0.02 0.14 invlessstock 0.03 0.18 0.05 0.21 0.03 0.18 0.05 0.21 invmorera 0.02 0.14 0.01 0.08 invlessra 0.02 0.14 0.01 0.08 paydebt 0.03 0.16 0.02 0.13 spendless 0.17 0.37 0.16 0.37 moreriskaver 0.08 0.28 0.10 0.31 invlesslong 0.02 0.14 0.01 0.10 0.01 0.09 0.02 0.13 work 0.84 0.36 0.77 0.42 0.83 0.37 0.76 0.43 lnincome 11.67 1.72 12.00 2.14 11.76 1.93 12.18 2.13 lnnetworth 12.26 3.84 13.44 3.56 12.55 3.92 13.76 3.25 married 0.75 0.43 0.75 0.44 0.76 0.43 0.80 0.40 risk 0.24 0.43 0.25 0.43 0.27 0.44 0.18 0.39 stockshare 3.99 14.09 4.64 11.80 5.94 69.43 5.79 13.72 age3549 0.38 0.48 0.27 0.45 0.30 0.46 0.24 0.43 age5064 0.33 0.47 0.38 0.48 0.41 0.49 0.42 0.49 age65_ 0.12 0.33 0.26 0.44 0.16 0.37 0.26 0.44 kid 1.01 1.21 0.82 1.15 0.94 1.19 0.88 1.20 white 0.79 0.41 0.80 0.40 0.78 0.42 0.84 0.37 college 0.62 0.49 0.64 0.48 0.63 0.48 0.68 0.47 homeowner 0.79 0.40 0.86 0.35 0.81 0.39 0.89 0.32 observations 1078 700 1103 724 subsamples newincrterm =1 newincrwhole =1 dropdecrterm =1 dropdecrwhole =1 regression models (5) (6) (7) (8) note 1: as described in section 3, lnnewincrterm (lnnewincrwhole) is defined as the natural log of the amount of newly purchased or increased term (whole) life insurance coverage, specifically applicable to the households that have chosen to initiate or increase term (whole) life insurance since 2007, where newincrtermi = 1 (newincrwholei = 1). lndropdecrterm (lndropdecrwhole) is defined as the natural log of the amount of newly purchased or increased whole life insurance coverage, specifically applicable to the households that have chosen to drop or decrease term (whole) life insurance since 2007, where dropdecrtermi = 1 (newincrwholei = 1). note 2: all data variables are taken or calculated from the 2007-2009 scf panel survey. see appendix a for more details. financial services review, 32(1) 12 results for new life insurance ownership and coverage tables 3, 4, 5, and 6 report the variables of household portfolio allocation decisions that have a significant and dynamic impact on life insurance ownership and coverage decisions, while holding the other determinants of the demand for life insurance constant. table 3 shows the results for new life insurance ownership. it indicates that households that have made decisions to invest more in cash and cash equivalents are less likely to initiate or increase term life insurance. by holding more cash and cash equivalents, households could gain financial security and confidence in relying on the accumulated cash savings should unforeseen circumstances arise. consequently, they may exhibit reduced risk aversion and choose to selfinsure instead of holding term life insurance to offset potential financial losses resulting from premature death. furthermore, from the perspective of the total portfolio picture, households might perceive their most liquid financial assets as an avenue to explore investments in alternative assets that possess the potential for higher returns compared to life insurance. so they could be reluctant in directing their surplus cash towards life insurance, particularly when confronted with other financial priorities during recessions. in table 3, the results also indicate that households that have determined to increase their investment in retirement assets are more likely to initiate or increase term life insurance. given that retirement assets represent a substantial portion of household wealth, this finding suggests that household wealth may play a positive role in driving the demand for term life insurance (headen & lee, 1974; bernheim, 1991; hau 2000; lin & grace 2007; shi et al., 2015; mulholland et al., 2016). this insight implies that improved financial wellbeing may lead to an increase in the demand for term life insurance, particularly during periods of economic downturns. the proximity to and familiarity with financial markets could potentially contribute to this trend, as these factors tend to bolster both engagement in financial markets and the insurance market (luciano et al., 2016). additionally, diversification could emerge as an additional factor. term life insurance can add another layer of protection beyond the relatively less liquid retirement investments, thereby mitigating overall financial risk for households. furthermore, the results for new life insurance coverage shown in table 4 demonstrate that, among the households that chose to initiate or increase term life insurance, those that invested more in retirement assets tend to increase significantly more coverage than others. it is implied that the flow of household funds into risky assets can positively impact their holding of new or more term life insurance during recessions. increasing investments for retirement in conjunction with the acquisition or expansion of term life insurance coverage for risk management can constitute fundamental elements of a comprehensive and prudent financial planning strategy, particularly in the context of a financial crisis. table 3 also indicates that households that have determined to increase their investment in retirement assets are less likely to initiate or increase whole life insurance. it suggests that household wealth tends to diminish the demand for whole life insurance that has a saving function. households focusing on bolstering their retirement investments could find that the cost of whole life insurance can limit the amount of funds they can allocate to retirement assets during recessions. these households might have a higher risk tolerance and a greater propensity to invest in assets offering potentially higher returns compared to whole life insurance, which emphasizes stability and guaranteed or moderate returns. in addition to the active allocation of financial assets in household portfolios, the results of table 3 also demonstrate that household decisions to pay off debts can impact the ownership of whole life insurance. it shows that households that have decided to pay off debts are more likely to initiate or increase whole life insurance. the cash value of whole life insurance holds the potential to serve as a resource for debt or loan repayment (ferber & lee, 1980; hau, 2000; lin & grace, 2007; wang, 2023). incorporating whole life insurance into a debt repayment strategy can wang et al. 13 contribute to enhancing household financial stability during economic recessions. we also explore the impact of other financial decisions made by households on their new life insurance ownership. according to table 3, households that have reduced their spending during recessionary periods are more likely to initiate or increase whole life insurance. this suggests that as households become more risk averse due to reduced spending during a financial crisis, they tend to purchase whole life insurance and uphold sufficient life insurance coverage (bricker et al., 2011; scott & gilliam, 2014). additionally, table 3 reveals that households demonstrating more conservative or disciplined financial behaviors are less likely to initiate or increase term life insurance during recessions. tables 3 also shows that households that have opted to invest less for the long term are more likely to initiate or increase term life insurance. this suggests that term life insurance may not be considered an essential element of long-term household financial planning. results for dropped life insurance ownership and coverage table 5 shows the results for dropped life insurance ownership. it indicates that households that have made decisions to increase their investment in bonds are less likely to decrease or drop term life insurance during the financial crisis. this suggests a positive relationship between allocating more funds to bonds and the ownership of term life insurance. bonds represent a less risky investment choice compared to stocks and other higher-risk assets. households concentrating on risk mitigation or financial stability might recognize the value of maintaining term life insurance coverage in conjunction with their bond investments during periods of recession. while prior studies find a substitution effect between term life insurance and lower-risk assets like bonds (fortune, 1973; hau, 2000; lin & grace, 2007), our paper implies that this effect may result from passive acceptance of household portfolio shifts rather than active decisionmaking for portfolio allocation in bonds. table 5 also indicates that households might perceive a reduced need for term life insurance if they are investing less in retirement assets. a reduction of retirement investments might arise due to budget constraints, leading to a decline in term life insurance. table 5 also shows the impact of other financial decisions made by households on the dropped life insurance ownership. it reveals that households spending less are more likely to drop or reduce their existing whole life insurance. this relationship can be attributed to liquidity constraints faced by households (bernheim et al., 1999). as households have tightened their budgets and prioritized essential expenses during economic recessions, they may consider their current whole life insurance as an option for cost reduction. the cash value component of whole life insurance can also serve as an asset that can be accessed during recessions. consequently, they may choose to reduce or terminate whole life insurance, or borrow the cash value (cole & fire, 2021). additionally, table 5 shows that households that have made decisions to invest less for the long term are less likely to drop or reduce term life insurance. furthermore, these households tend to drop or reduce significantly more term life insurance coverage than other households once they have made decisions to drop, as indicated in table 6. this finding can be attributed to a lower level of familiarity and engagement with financial markets among households that lack long-term financial planning (luciano et al., 2016). financial services review, 32(1) 14 table 3. results for new life insurance ownership dependent variable newincrterm newincrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.4109 0.1878 ** -0.0569 0.1807 invmorebonds 0.2112 0.2064 -0.2030 0.2323 invlessstock -0.0853 0.1271 0.1238 0.1294 invmorera 0.3857 0.1932 ** -0.4559 0.2747 * paydebt 0.3035 0.1642 * spendless 0.1529 0.0666 ** moreriskaver -0.1422 0.0796 * -0.0932 0.0835 invlesslong 0.3415 0.1809 * -0.3390 0.2268 dropdecrwhole 0.4188 0.0556 *** dropdecrterm 0.3553 0.0517 *** work 0.1409 0.0645 ** -0.0135 0.0694 lnincome -0.0095 0.0147 0.0175 0.0165 lnnetworth -0.0041 0.0082 0.0411 0.0107 *** married 0.1455 0.0551 *** 0.0113 0.0608 risk -0.1051 0.0569 * -0.0210 0.0633 stockshare -0.0007 0.0011 -0.0041 0.0021 * age3549 0.0216 0.0707 0.0816 0.0883 age5064 -0.2609 0.0765 *** 0.1391 0.0919 age65_ -0.5752 0.0961 *** 0.2936 0.1080 *** kid -0.0149 0.0207 -0.0076 0.0237 white 0.0856 0.0571 -0.0710 0.0643 college 0.0471 0.0526 0.0585 0.0587 homeowner 0.1681 0.0695 ** 0.0909 0.0799 intercept -0.7018 0.1543 *** -1.9193 0.1743 *** observations 3857 3857 pseudo r-square 0.057 0.043 wald exogeneity test 0.541 0.555 regression models (1) (2) note 1: the model for each type of life insurance is a probit based on the full sample. we reported the probit results here, as the wald test for exogeneity is not rejected. note 2: variables definitions and summary statistics in these two regressions were listed in table 1. note 3: statistical significance at the 1, 5, and 10 percent levels is denoted by ***, ** and *, respectively. wang et al. 15 table 4. results for new life insurance coverage dependent variable lnnewincrterm lnnewincrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.2925 0.4126 -0.0189 0.5600 invmorebonds 0.4328 0.3656 1.0510 0.7432 invlessstock 0.0919 0.2502 -0.3983 0.3908 invmorera 0.5878 0.3175 * 0.3033 0.9882 paydebt -0.4682 0.4601 spendless -0.2011 0.2088 moreriskaver 0.0213 0.1562 0.0094 0.2550 invlesslong -0.1420 0.3072 -0.8170 0.7599 lndropdecrwhole 0.0550 0.0084 *** lndropdecrterm 0.0137 0.0127 work 0.5749 0.1329 *** 0.2486 0.2131 lnincome 0.3164 0.0331 *** 0.3464 0.0468 *** lnnetworth 0.0663 0.0163 *** 0.1678 0.0332 *** married 0.7232 0.1082 *** 0.5705 0.1927 *** risk -0.2905 0.1154 ** 0.0113 0.2017 stockshare 0.0038 0.0031 -0.0090 0.0069 age3549 -0.0292 0.1314 -0.9426 0.3117 *** age5064 -0.5671 0.1416 *** -1.4542 0.3255 *** age65_ -1.0423 0.1895 *** -1.9376 0.3672 *** kid 0.0722 0.0403 * 0.1033 0.0750 white -0.0369 0.1107 -0.2393 0.2045 college 0.1792 0.0985 ** 0.2785 0.1886 homeowner -0.2024 0.1320 -0.4381 0.2580 * intercept 6.2106 0.3415 *** 5.6857 0.5477 *** observations 1078 700 subsamples newincrterm = 1 newincrwhole = 1 wald exogeneity test 0.861 0.219 regression models (5) (6) note 1: the model for each type of life insurance is a truncated regression, based on the subsample of households with new policies. we reported the truncated results here, as the wald test for exogeneity is not rejected. note 2: these variables were defined in tables 1 and 2. summary statistics was reported in table 2. note 3: statistical significance at the 1, 5, and 10 percent levels is denoted by ***, ** and *, respectively. financial services review, 32(1) 16 table 5. results for dropped life insurance ownership dependent variable dropdecrterm dropdecrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.0600 0.1688 0.1154 0.1714 invmorebonds -0.3818 0.2244 * 0.1549 0.2080 invlessstock -0.0467 0.1245 -0.0124 0.1303 invlessra 0.3744 0.1886 ** -0.3394 0.2527 paydebt -0.0707 0.1832 spendless 0.1514 0.0676 ** invlesslong -0.3908 0.2046 * -0.0986 0.1988 newincrwhole 0.3724 0.0549 *** newincrterm 0.4085 0.0529 *** work 0.2817 0.0637 *** -0.0808 0.0693 lnincome -0.0031 0.0145 0.0116 0.0163 lnnetworth -0.0104 0.0083 0.0333 0.0109 *** married 0.1878 0.0546 *** 0.1645 0.0628 *** risk 0.0212 0.0561 -0.2369 0.0662 *** stockshare 0.0004 0.0006 -0.0011 0.0019 age3549 0.1153 0.0736 0.0069 0.0914 age5064 0.3058 0.0775 *** 0.3206 0.0934 *** age65_ 0.0555 0.0954 0.4174 0.1105 *** kid 0.0047 0.0208 0.0546 0.0235 ** white -0.0403 0.0565 0.0018 0.0669 college 0.0935 0.0523 * 0.0386 0.0592 homeowner 0.1488 0.0693 ** 0.1504 0.0828 * intercept -1.1260 0.1530 *** -1.9970 0.1742 *** observations 3857 3857 pseudo r-square 0.044 0.069 wald exogeneity test 0.520 0.864 regression models (3) (4) note 1: the model for each type of life insurance is a probit regression to examine the determinants of dropped policy ownership status based on the full sample. we reported the probit results here, as the wald test for exogeneity is not rejected. note 2: variables definitions and summary statistics in these two regressions were listed in table 1. note 3: statistical significance at the 1, 5, and 10 percent levels is denoted by ***, ** and *, respectively. wang et al. 17 table 6. results for dropped life insurance coverage dependent variable lndropdecrterm lndropdecrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.3747 0.3349 -0.1317 0.4902 invmorebonds -0.1010 0.4960 -0.5793 0.5468 invlessstock -0.1350 0.2465 0.4601 0.3697 invlessra 0.0263 0.3165 -0.3573 0.8963 paydebt 0.3427 0.5738 spendless -0.1017 0.2027 invlesslong 1.4451 0.4570 *** 0.3396 0.5768 lnnewincrwhole 0.0298 0.0087 *** lnnewincrterm 0.0567 0.0125 *** work 0.2246 0.1361 * 0.4849 0.2052 ** lnincome 0.3002 0.0295 0.2311 0.0493 *** lnnetworth 0.1222 0.0172 *** 0.2518 0.0364 *** married 0.4033 0.1134 *** 0.4452 0.2052 ** risk -0.2903 0.1111 *** 0.1514 0.2114 stockshare -0.0001 0.0006 -0.0002 0.0059 age3549 -0.5806 0.1579 *** -0.5270 0.3059 * age5064 -0.5850 0.1624 *** -0.9374 0.3079 *** age65_ -1.3084 0.2002 *** -1.1187 0.3514 *** kid 0.1496 0.0416 *** 0.0820 0.0743 white -0.0638 0.1097 -0.4781 0.2161 ** college 0.4972 0.1043 *** 0.1054 0.1774 homeowner 0.0331 0.1434 -0.2345 0.2819 intercept 6.1824 0.3201 *** 5.1278 0.5338 *** observations 1103 724 subsamples dropdecrterm = 1 dropdecrwhole = 1 wald exogeneity test 0.841 0.59 regression models (7) (8) note 1: the model for each type of life insurance is a truncated regression to examine the determinants of dropped policy coverage, based on the subsample of households with dropped policies. we reported the truncated results here, as the wald test for exogeneity is not rejected. note 2: these variables were defined in tables 1 and 2. summary statistics was reported in table 2. note 3: statistical significance at the 1, 5, and 10 percent levels is denoted by ***, ** and *, respectively. financial services review, 32(1) 18 conclusions and implications household financial decisions are interdependent and essential to both household financial wellbeing and social welfare (gomes et al., 2021). bhamra and uppal (2019) suggest a multiplier effect that small biases in household financial decisions can lead to large economic losses, not just for individual households, but also for society. previous research has examined the demand for life insurance as a function of household demographic characteristics and economic status, including household portfolio holdings. household portfolio shifts can arise from active decision-making for portfolio allocation, or from passive acceptance of asset price changes (bricker et al., 2011). the differentiation between these two aspects has not been explored in the insurance literature. our paper aims to fill the gap by exploring the relationship between the active decision-making for portfolio allocation and the demand for life insurance. in addition, the disparities observed in demand determinants between term and whole life insurance in our study indicate the importance of classifying life insurance by type, in line with previous literature (lin & grace, 2007; grable, 2016; heo et al., 2021). moreover, the scarcity of research employing household panel data and the limited attention given to recessionary periods in the literature further demonstrate the significance of our study (liebenberg et al., 2012). our results indicate that household portfolio allocation decisions have a significant and dynamic impact on life insurance ownership, while having a limited impact on life insurance coverage during recessions. the significant and influential factors of household portfolio allocation decisions found in this paper can be useful predictors of changes in the demand for life insurance at the household level during recessionary times. specifically, the results indicate that households deciding to invest more in cash and cash equivalents are less likely to initiate or increase term life insurance during the financial crisis. the results also suggest that during recessions, the ownership of term life insurance is likely to be a complement for the allocation of household portfolios in retirement assets, while the ownership of whole life insurance is likely to be a substitute. additionally, the results demonstrate that households deciding to invest more in bonds are less likely to drop or reduce term life insurance. the results also demonstrate that households that have decided to pay off debts are more likely to initiate or increase their whole life insurance. the insights from this study can be beneficial for life insurers, financial planning practitioners, and social policymakers in estimating the demand for life insurance and developing production and marketing strategies for different economic conditions. the findings highlight the importance of considering households’ active decisions of portfolio allocation in creating these strategies, which can help life insurance agents and financial planning practitioners offer adaptive and comprehensive financial services to households whose financial decisions may change during recessions. the findings are especially valuable for financial practitioners who serve wealthier households with a diverse range of financial assets and a propensity to adjust their portfolio compositions in response to evolving economic conditions. first, our results indicate that promoting the allocation of household portfolios in retirement assets and fostering the participation in financial markets can potentially boost the demand for term life insurance. from the perspective of the overall household portfolios, concurrently managing retirement planning through investments in risky assets and risk mitigation through term life insurance could represent a more comprehensive and prudent strategy for household financial planning. this holds particular importance for wealthier households in need of tailored advice to optimize their investments and risk management strategies. second, our findings suggest that households can be better served when financial researchers and practitioners more carefully identify and accommodate the demand for different types of life insurance. it is recommended that term life insurance can be promoted to households that have decided to increase their investments in retirement assets and bonds, while whole life insurance can be targeted towards households aiming to pay off debts. these recommendations wang et al. 19 can help financial service companies and practitioners tailor their strategies to align with the needs of households during economic downturns, ensuring that households have adequate life insurance coverage to protect against financial hardships. third, our findings suggest that the growth of household wealth may stimulate the demand for term life insurance as a safeguard against financial losses during periods of economic downturns. nevertheless, it is suggested that the growth of household wealth may crowd out the new demand for whole life insurance which has a function of savings. fourth, our results imply that the efforts to enhance household liquidity status, promote conservative or disciplined financial behavior, and encourage long-term investments may not necessarily have the intended positive impact on boosting the demand for term life insurance. instead, these strategies might have an adverse effect or fail to produce the anticipated results during economic downturns. finally, the analysis of determinants influencing household decisions to drop or reduce life insurance can assist in identifying households that are more prone to experiencing financial hardships following a financial crisis. our results indicate that households deciding to invest less in retirement assets are more likely to drop or reduce term life insurance, and households deciding to spend less are more likely to drop or reduce whole life insurance compared to other households. prioritizing these households becomes crucial for social policymakers seeking to mitigate household financial hardships due to inadequate life insurance during recessions. one limitation of this study is that the scope of active decision-making for portfolio allocation could include the scenario where households actively choose to retain their current allocation, an aspect not explicitly addressed in our research. a future direction of our study involves establishing connections between household financial decisions, including choices in insurance markets, and the potential impacts of macroeconomic policies. this endeavor aims to enhance household financial well-being and overall social welfare. financial services review, 32(1) 20 appendix a survey questions about the demand for life insurance and the household portfolio allocation decisions in the codebook of the 2007-2009 survey of consumer finances panel data set are listed as follows. for more information, see https://www.federalreserve.gov/econres/files/codebk2009p.txt. 1. the ownership of term life insurance x4002: question text same as 2009 version p4002: the two major types of life insurance are term and cash-value policies. term policies pay a benefit if the insured person dies, but otherwise have no value. they are often provided through an employer or union, but may also be bought by individuals. cash-value policies also pay a death benefit, but differ in that they build up a value as premiums are paid. are any of your (family's) policies term insurance? 2. the coverage of term life insurance x4003: question text same as 2009 version p4003: what is the current face value of all the term life policies that you (and your family living here) have? (the face value of a policy is what the policy would pay in the event of death) 3. the ownership of whole life insurance x4004: question text same as 2009 version p4004: do you have any policies that build up a cash value or that you can borrow on? these are sometimes called "whole life", "straight life", or "universal life" policies. 4. the coverage of whole life insurance x4005: question text same as 2009 version p4005: what is the current face value of all of the policies that build up a cash value? (the face value of a policy is what the policy would pay in the event of death.) x4006: question text same as 2009 version p4006: if you cancelled these policies now, how much would you receive from the insurance company for the payments you have made up to now? that is, what is the current "cash value" of the policies? https://www.federalreserve.gov/econres/files/codebk2009p.txt wang et al. 21 5. household portfolio allocation decisions p091460: over this time, have you (and your family) made decisions to change the ways you arrange your money or investments? p091421: generally, what were those decisions? code all that apply 2) invest more in cds, other deposits, "cash" 3) invest less in stocks 12) invest more in tax-exempt bonds 14) invest more in treasury bills/bonds 16) invest more in other bonds 17) invest less in retirement assets (ira, keogh, 401(k), roth, etc.) 18) invest more in retirement assets (ira, keogh, 401(k), roth, etc.) 27) invest less for the long term 46) more conservative/disciplined investments; less risk/aggressive 65) spend less, cut back 88) pay off debt financial services review, 32(1) 22 appendix b table b1. robustness test results for new life insurance ownership using the iv approach dependent variable newincrterm newincrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.3532 0.1783 ** -0.0457 0.1799 invmorebonds 0.4087 0.2058 ** -0.0956 0.2911 invlessstock -0.1164 0.1247 0.1323 0.1281 invmorera 0.3705 0.1947 * -0.4277 0.2662 paydebt 0.3190 0.1685 * spendless 0.1465 0.0668 ** moreriskaver -0.1436 0.0793 * -0.0884 0.0824 invlesslong 0.2979 0.1764 * -0.2173 0.2760 dropdecrwhole (iv) -2.8799 1.0204 *** dropdecrterm (iv) 1.2009 1.3608 work 0.0852 0.0668 -0.0885 0.1392 lnincome 0.0107 0.0153 0.0177 0.0177 lnnetworth 0.0143 0.0100 0.0427 0.0125 *** married 0.2826 0.0701 *** -0.0427 0.1034 risk -0.2871 0.0797 *** -0.0227 0.0639 stockshare -0.0013 0.0013 -0.0044 0.0021 * age3549 0.0176 0.0706 0.0490 0.1018 age5064 -0.0226 0.1048 0.0462 0.1704 age65_ -0.2841 0.1315 * 0.2676 0.1104 * kid 0.0316 0.0246 -0.0077 0.0235 white 0.0945 0.0578 -0.0607 0.0672 college 0.0860 0.0550 0.0304 0.0746 homeowner 0.2677 0.0760 *** 0.0475 0.1059 intercept -0.8122 0.1503 *** -1.9914 0.2378 *** observations 3857 3857 pseudo r-square 0.046 0.032 regression models (1) (2) note: the results of this robustness check by using the instrumental variable (iv) approach presented in table b1 show that new term life insurance ownership is negatively associated with household decisions to invest more in cash and cash equivalents and to become more risk averse, and positively associated with household decisions to invest more in retirement assets and less for the long term; and new whole life insurance ownership is positively associated with household decisions to pay off debts and to spend less during recessions. all these results largely confirm our findings in this paper. wang et al. 23 table b2. robustness test results for dropped life insurance ownership using the iv approach dependent variable dropdecrterm dropdecrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.0810 0.1702 0.1050 0.2051 invmorebonds -0.4368 0.2241 * 0.1551 0.2275 invlessstock -0.0267 0.1272 -0.0094 0.1327 invlessra 0.3908 0.1829 ** -0.3009 0.2641 paydebt -0.0587 0.1873 spendless 0.1501 0.0669 ** invlesslong -0.4431 0.2146 ** -0.0959 0.2277 newincrwhole (iv) -0.3176 0.8378 newincrterm (iv) 0.4317 1.0116 work 0.2845 0.0629 *** -0.0822 0.0799 lnincome 0.0001 0.0148 0.0117 0.0176 lnnetworth -0.0050 0.0100 0.0313 0.0114 ** married 0.1911 0.0550 *** 0.1618 0.0803 * risk 0.0199 0.0569 -0.2324 0.0775 ** stockshare 0.0003 0.0004 -0.0012 0.0017 age3549 0.1272 0.0742 . 0.0060 0.0927 age5064 0.3311 0.0836 *** 0.3185 0.1236 ** age65_ 0.1078 0.1134 0.4120 0.2008 * kid 0.0046 0.0202 0.0554 0.0231 * white -0.0556 0.0596 0.0059 0.0732 college 0.1033 0.0539 0.0350 0.0638 homeowner 0.1679 0.0735 * 0.1454 0.1025 intercept -1.1304 0.1498 *** -1.9594 0.3301 *** observations 3857 3857 pseudo r-square 0.033 0.055 regression models (3) (4) note: the results of this robustness check by using the iv approach presented in table b2 show that dropped term life insurance ownership is negatively associated with household decisions to invest more in bonds and less for the long term, and positively associated with household decisions to invest less in retirement assets; and dropped whole life insurance ownership is positive associated with household decisions to spend less. all these results largely confirm our findings in this paper. financial services review, 32(1) 24 table b3. robustness test results for new life insurance ownership using the 2009 control variables dependent variable newincrterm newincrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.4342 0.1883 ** -0.0672 0.1804 invmorebonds 0.2006 0.2065 -0.2031 0.2315 invlessstock -0.0774 0.1272 0.1012 0.1295 invmorera 0.3804 0.1936 ** -0.4773 0.2762 * paydebt 0.2801 0.1652 * spendless 0.1677 0.0674 ** moreriskaver -0.1729 0.0798 ** -0.1041 0.0835 invlesslong 0.3453 0.1815 * -0.3413 0.2278 dropdecrwhole 0.4119 0.0555 *** dropdecrterm 0.3620 0.0519 *** work09 0.1653 0.0592 *** -0.0546 0.0633 lnincome09 0.0349 0.0121 *** 0.0017 0.0115 lnnetworth09 -0.0035 0.0067 0.0421 0.0085 *** married09 0.2059 0.0528 *** 0.1001 0.0581 * risk09 -0.0837 0.0532 -0.0722 0.0595 stockshare09 -0.0028 0.0023 -0.0025 0.0024 age3549_09 0.1097 0.0771 -0.1405 0.0961 age5064_09 -0.1313 0.0811 0.0137 0.0971 age65_09 -0.4638 0.0976 *** 0.1618 0.1103 kid09 0.0001 0.0207 0.0149 0.0242 white09 0.0769 0.0576 -0.1137 0.0651 * college09 -0.0099 0.0502 0.0081 0.0557 homeowner09 0.1195 0.0683 * 0.2456 0.0815 *** intercept -1.2581 0.1478 *** -1.7069 0.1607 *** observations 3857 3857 pseudo r-square 0.066 0.052 regression models (1) (2) note: the results of the robustness check by using the 2009 control variables presented in table b3 show that new term life insurance ownership is negatively associated with household decisions to invest more in cash and cash equivalents and to become more risk averse, and positively associated with household decisions to invest more in retirement assets and less for the long term; and new whole life insurance ownership is negatively associated with household decisions to invest more in retirement assets, and positively associated with household decisions to pay off debts and to spend less during recessions. all these results largely confirm our findings in this paper. wang et al. 25 table b4. robustness test results for dropped life insurance ownership using the 2009 control variables dependent variable dropdecrterm dropdecrwhole independent variables coefficient estimate standard error coefficient estimate standard error invmorecash -0.0459 0.1685 0.1449 0.1705 invmorebonds -0.3525 0.2255 0.1711 0.2077 invlessstock -0.0600 0.1246 -0.0073 0.1294 invlessra 0.4102 0.1882 ** -0.3486 0.2517 paydebt -0.0973 0.1825 spendless 0.1631 0.0673 ** invlesslong -0.3342 0.2037 * -0.0617 0.1984 newincrwhole 0.3773 0.0550 *** newincrterm 0.3991 0.0527 *** work09 0.1501 0.0575 *** -0.0240 0.0635 lnincome09 -0.0139 0.0107 0.0171 0.0121 lnnetworth09 -0.0109 0.0065 * 0.0189 0.0080 ** married09 0.1592 0.0516 *** 0.1166 0.0581 ** risk09 0.1334 0.0519 ** -0.1452 0.0597 ** stockshare09 0.0009 0.0007 0.0008 0.0007 age3549_09 0.0483 0.0796 0.0723 0.1000 age5064_09 0.2570 0.0816 *** 0.3277 0.1007 *** age65_09 -0.0300 0.0965 0.4873 0.1143 *** kid09 -0.0162 0.0211 0.0200 0.0241 white09 -0.0183 0.0566 0.0198 0.0665 college09 0.1318 0.0493 *** 0.0184 0.0549 homeowner09 0.1855 0.0675 *** 0.2135 0.0807 *** intercept -0.9006 0.1396 *** -1.9695 0.1653 *** observations 3857 3857 pseudo r-square 0.040 0.059 regression models (3) (4) note: the results of the robustness check by using the 2009 control variables presented in table b4 show that dropped term life insurance ownership is negatively associated with household decisions to invest less for the long term, and positively associated with household decisions to invest less in retirement assets; and dropped whole life insurance ownership is positive associated with household decisions to spend less. all these results largely confirm our findings in this paper. financial services review, 32(1) 26 references baranoff, e. g., & sager, t.w. (2011). the interplay between insurers’ financial and asset risks during the crisis of 2007-2009. the geneva papers on risk and insurance issues and practice, 36(3), 348-379. bernheim, b. d. (1991). how strong are bequest motives? evidence based on estimates of the demand for life insurance and annuities. journal of political economy, 99(5), 899-927. bernheim, b. d., forni, l., gokhale, j., & kotlikoff, l. j. (1999). the adequacy of life insurance: evidence from the health and retirement survey (no. w7372). national bureau of economic research. bertaut, c. c. (1998). stockholding behavior of us households: evidence from the 1983–1989 survey of consumer finances. review of economics and statistics, 80(2), 263-275. bhamra, h. s., & uppal, r. (2019). does household finance matter? small financial errors with large social costs. american economic review, 109(3), 1116-1154. bricker, j., bucks, b., kennickell, a. b., mach, t., & moore, k. b. (2011). surveying the aftermath of the storm: changes in family finances from 2007 to 2009. federal reserve board working paper. brown, j. r. (1999). are the elderly really overannuitized? new evidence on life insurance and bequests (no. w7193). national bureau of economic research. campbell, j. y., ramadorai, t., & schwartz, a. (2009). caught on tape: institutional trading, stock returns, and earnings announcements. journal of financial economics, 92(1), 66-91. cole, c. r., & fier, s. g. (2021). an examination of life insurance policy surrender and loan activity. journal of risk and insurance, 88(2), 483-516. cragg, j. g. (1971). some statistical models for limited dependent variables with application to the demand for durable goods. econometrica, 39(5), 829-844. eisenhauer, j. g., & halek, m. (1999). prudence, risk aversion, and the demand for life insurance. applied economics letters, 6(4), 239-242. ferber, r., & lee, l. c. (1980). acquisition and accumulation of life insurance in early married life. journal of risk and insurance, 47(4), 713-734. fortune, p. (1973). a theory of optimal life insurance: development and test. the journal of finance, 28(3), 587-600. gomes, f., haliassos, m., & ramadorai, t. (2021). household finance. journal of economic literature, 59(3), 919-1000. glumov, d. (2013). how did the great financial crisis affect portfolio allocations and attitudes towards risk? evidence from the survey of consumer finances. retrieved from http://papers.ssrn.com/sol3/papers. cfm?abstract_id=2293189. grable, j. e. (2016). the case approach to financial planning bridging the gap between theory and practice (3rd ed.). the national underwriter company. hau, a. (2000). liquidity, estate liquidation, charitable motives, and life insurance demand by retired singles. journal of risk and insurance, 67(1), 123-141. headen, r. s., & lee, j. f. (1974). life insurance demand and household portfolio behavior. journal of risk and insurance, 41(4), 685-698. heo, w. (2020). the demand for life insurance: dynamic ecological systemic theory using machine learning techniques. springer nature. heo, w., & grable, j. e. (2017). demand for life insurance: a consumer perspective. journal of financial service professionals, 71(3), 19-23. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2293189 http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2293189 wang et al. 27 heo, w., grable, j. e., & chatterjee, s. (2013). life insurance consumption as a function of wealth change. financial services review, 22(4), 389-404. heo, w., lee, j. m., & park, n. (2021). who demands which type of life insurance? various factors in life insurance ownership. financial services review, 29(2), 101-119. holsboer, j. h. (2000). the impact of low interest rates on insurers. the geneva papers on risk and insurance issues and practice, 25(1), 38-58. hoopes, j., langetieg, p., nagel, s., reck, d., slemrod, j., & stuart, b. (2016). who sold during the crash of 2008-9? evidence from tax-return data on daily sales of stock. nber working paper no. w22209. national bureau of economic research. hubener, a., maurer, r., & mitchell, o. s. (2016). how family status and social security claiming options shape optimal life cycle portfolios. the review of financial studies, 29(4), 937-978. inkmann, j., & michaelides, a. (2012). can the life insurance market provide evidence for a bequest motive? journal of risk and insurance, 79(3), 671-695. lewis, f. (1989). dependents and the demand for life insurance. the american economic review, 79(3), 452-467. liebenberg, a. p., carson, j. m., & hoyt, r. e. (2010). the demand for life insurance policy loans. journal of risk and insurance, 77(3), 651-666. liebenberg, a. p., carson, j. m., & dumm, r. e. (2012). a dynamic analysis of the demand for life insurance. journal of risk and insurance, 79(3), 619-644. lin, y., & grace, m. f. (2007). household life cycle protection: life insurance holdings, financial vulnerability, and portfolio implications. journal of risk and insurance, 74(1), 141-173. luciano, e., outreville, j. f., & rossi, m. (2016). life insurance ownership by italian households: a gender-based differences analysis. the geneva papers on risk and insurance issues and practice, 41(3), 468-490. maremont, m., & scism, l. (2010). shift to wealthier clientele puts life insurers in a bind. the wall street journal. http://online.wsj.com/article/sb1000142 40527487034351045754214114495552 40. miller, m. a. (1985). age-related reductions in workers' life insurance. monthly labor review, 108, 29-34. mulholland, b., finke, m., & huston, s. (2016). understanding the shift in demand for cash value life insurance. risk management and insurance review, 19(1), 7-36. o'neill, b., & xiao, j. j. (2012). financial behaviors before and after the financial crisis: evidence from an online survey. journal of financial counseling and planning, 23(1), 33-46. outreville, j. f. (2013). the relationship between insurance and economic development: 85 empirical papers for a review of the literature. risk management and insurance review, 16(1), 71-122. outreville, j. f. (2014). risk aversion, risk behavior, and demand for insurance: a survey. journal of insurance issues, 37(2), 158-186. pissarides, c. a. (1980). the wealth-age relation with life insurance. economica, 47(188), 451-457. rabbani, a. g., grable, j. e., heo, w., nobre, l., & kuzniak, s. (2017). stock market volatility and changes in financial risk tolerance during the great recession. journal of financial counseling and planning, 28(1): 140154. scott, j. k., & gilliam, j. (2014). boomers' life insurance adequacy pre-& post the 2008 financial services review, 32(1) 28 financial crisis. financial services review, 23(4), 287-304. shi, x., wang, h. j., & xing, c. (2015). the role of life insurance in an emerging economy: human capital protection, assets allocation and social interaction. journal of banking & finance, 50(c), 19-33. swiss re. (2009). world insurance in 2008. sigma no. 3. wang, n. (2019). the demand for life insurance in a heterogeneous-agent life cycle economy with joint decisions. the geneva risk and insurance review, 44(2): 176-206. wang, n. (2023). a dynamic analysis of the demand for life insurance during the 2008 financial crisis: evidence from the panel survey of consumer finances. the geneva papers on risk and insurance issues and practice, 48(4), 733-759. https://doi.org/10.1057/s41288-02100262-2. zietz, e. n. (2003). an examination of the demand for life insurance. risk management and insurance review, 6(2): 159-191. zhu, y. (2007). one-period model of individual consumption, life insurance, and investment decisions. journal of risk and insurance, 74(3): 613-636. pii: 1057-0810(92)90011-z from the editor investors today are faced with an abundance of investment vehicles and a bewildering number of claims made by the vendors of these vehicles. individuals are urged to purchase mutual funds or variable annuities based upon a variety of performance measures including alphas, betas, and five and ten year compounded annual returns. some vendors tout investment “systems”, such as dollar cost averaging, as best meeting the investor’s needs. the four articles in this issue call into question many of these claims. the first article, “a multicriteria approach to mutual fund selection,” by kevin hebner and wade cook looks beyond a fund’s risk-adjusted rate of return. they suggest a multicriteria methodology matches a set of fund attributes to the preferences of the individual. these attributes include such things as front and back-end load fees, the level of diversification, the quality of service and the standard deviation of the fund’s alpha. the second article, “active asset allocation decision of professional equity managers,” by robert brooks, haim levy and robert radcliffe, is related in that it tells us not to trust a single measure of a managed portfolio’s risk, notably its beta. the widely reported time series beta is found to be a poor predictor of a portfolio’s current equity risk exposure since active and passive asset allocation will change that beta from one quarter to another. they recommend that investors be given the fund’s current cross-sectional beta as well since this best describes the market risk of the current portfolio. incidentally, this information is already supplied to the sophisticated consultants of large investors such as pension funds. the third article is “long-run returns on stock and bond portfolios: impli cations for retirement planning” by kirt butler and dale domian. since retirement funds are accumulated and disaccumulated over long periods of time, the authors look at the effect of time diversification on the choice of asset classes. by resampling ibbotson data they find, for example, that over a 30 year period of accumulation, the probability that corporate bonds will outperform common stocks is less than 4 percent. this type of information is critically important to individuals who are forced to make asset allocation decisions about their pensions with little substantive guidance. v vi financial services review, 2(l) 1992/1993 the fourth article “nobody gains from dollar cost averaging,” takes on one of the most cherished and widely advertised investment techniques recommended by retail brokerage firms. written by john knight and myself, the article contrasts dollar cost averaging with optimal rebalancing and a buy and hold strategy through graphical and empirical analysis and numerical simulation. dollar cost averaging, even without including its high transactions costs, turns out worst for investors, but perhaps not for its promoters. while all of the articles are academically rigorous, they also have immediately usable implications. hopefully they will be used by the better educated financial planners as well as by financial product vendors seeking to enlarge their share of an increasingly sophisticated market. a great deal of additional research needs to be done in the area of individual investment. we do not know, for example, the effect of mean reversion on optimal portfolio construction for the individual. nor do we know why so many consumers purchase mutual funds and variable annuities with timing services that almost consistently underperform the indices. little has been written about hedging pur chasing power risk in retirement portfolios, particularly for those living on non adjustable defined benefit pension plans. and what about the tradeoffs of adding an employer’s stock to your portfolio in a subsidized purchase plan when you are already dependent upon that employer for your salary and pension? this column will be used to suggest research topics in individual financial management. the editors of financial services review are accessible and welcome calls to discuss research ideas. -lewis mandell finser_31-2_complete_issue refinance, pay additional principal, or recast? a home borrower’s dilemma christine mcclatcheya,* amonfort college of business, university of northern colorado, campus box 128, greeley, co 80639, usa abstract home borrowers encounter important financial decisions well beyond their mortgage’s origination date. two common examples are refinancing or paying additional principal. a third option, the decision to recast, has received far less attention. we review the mechanics and availability of recasting to discover why this omission exists. the combination of (a) recent job losses from covid, (b) expectations of higher interest rates, and (c) rising home prices suggest recasting may be beneficial for many homeowners going forward. our discussion has implications for finance professionals, faculty, and anyone with a mortgage as they should be aware of the benefits of this timely option. © 2023 academy of financial services. all rights reserved. jel classifications: g21; d83 keywords: recast; refinance; mortgage; loan amortization; principal reduction 1. introduction forty-four percent of u.s. consumers had a mortgage in 2020. while this population did not change significantly from the prior year, the average loan balance increased after 2019 which indicates individuals were borrowing more than usual, likely due to rising home prices (stolba, 2021). decisions regarding a mortgage in no way end after the complicated and time-intensive process that takes place before closing. in contrast, astute borrowers continuously reevaluate their circumstances, alongside market conditions, to identify opportunities that may improve their financial well-being. *corresponding author. tel.: +1-970-302-9058. e-mail address: christine.mcclatchey@unco.edu 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 197–209 a refinance (alt-1) involves taking out a new mortgage and using the proceeds to repay the existing one. the new mortgage will (a) have a new maturity, interest rate, and loan conditions; (b) be subject to underwriting approval; and (c) trigger significant upfront closing costs. refinancing is usually exercised in a falling interest rate environment. with a lower rate, the borrower’s monthly payments decline and if the monthly savings exceed the new mortgage’s up-front closing costs in a timely manner, the borrower ultimately benefits. refinancing may also be used by borrowers wanting to extract home equity that exists because they have paid down the original loan’s principal and/or because the home price has significantly increased from the time of purchase. refinancing might also be used to reduce or extend the loan’s term (brady, canner, & maki, 2000). others may refinance to replace an adjustable-rate loan (arm) with a fixed-rate loan to eliminate the risk of rising interest rates. one downside of refinancing is that the borrower’s current credit history, income, debt, and assets will be reviewed as a part of the transaction. furthermore, refinancing may not be an option if rates have increased or, in the case of flat or even slightly declining rates, the new mortgage’s up-front closing costs cannot be recouped in a reasonable time frame. most mortgages offer borrowers (without penalty) the opportunity to pay extra principal (alt-2). this option is typically used by a borrower who has extra monthly income or has received a one-time cash inflow (e.g., inheritance, bonus, or tax refund). the benefit of paying extra principal is that the mortgage will be repaid before its original maturity (e.g., creating a new, shorter “effective maturity”). compared to a refinance, the advantage of paying extra principal is that there is no new credit check, income/asset/debt verification, or additional fees. a potential drawback is that the mortgage’s interest rate and minimum payment remain unchanged. one can imagine a situation whereby a one-time windfall is used to pay additional principle, only to be followed by a change in circumstances that halts or reduces future monthly income. in this situation, no relief is in sight as the original higher monthly payment is still due. a recast (alt-3) is viable only for borrowers who have already paid additional principal or currently have a lump sum available to apply to the original loan’s principal. alt-3 calculates a new monthly payment based on the reduced principal amount using the original interest rate and remaining loan term. the mortgage is essentially reamortized. the immediate benefit of a recast is that the borrower’s monthly payment declines, although the mortgage will not be paid off before its original maturity. importantly, the fee to recast is significantly lower (almost negligible) compared to the closing costs triggered when originating a new mortgage. table 1 summarizes the important differences between these three options. 2. literature review the decision to refinance is the only option that does not require a prior or current principal reduction. when interest rates fall refinancing reduces a borrower’s required monthly payment, assuming the borrower does not extract home equity or roll closing costs into the new loan that increase the loan balance. figure 1 depicts 30-year fixed mortgage rates dating back to 1971. since their peak in 1982, mortgage rates have followed a downward trend which explains why alt-1 received so much attention over the past three decades. 198 c. mcclatchey / financial services review 31 (2023) 197–209 early discussions on refinancing focused on “back-of-the-envelope” payback calculations that simply divide the up-front costs of the new mortgage by the reduction in monthly payment. if the homeowner expects to live in the home longer than the calculated breakeven period, refinancing is optimal. various heuristics followed. noyan and eugene (1993) and bennett, keane, and mosser (1999) concluded that if the new mortgage had an interest rate 1-2% less than the original interest rate, and the borrower planned to stay in the home for a specified minimum number of years, refinancing was optimal. subsequent research recognized the importance of the time value of money and taxes on the breakeven calculation. chen (1997) adds time value of money considerations to the simplified payback calculation but does not include the tax implications of a change in the monthly interest expense paid and a change in the amortization of certain closing costs. rose (1992) incorporates the time value of money using the new mortgage’s interest rate but importantly adds the tax implications of a change in monthly interest payments and closing costs, which at that time would be amortized for tax purposes over the new loan’s life. johnson and randle (1996, 2003) created an excel model that iteratively solved for the breakeven period, although they did not account for the ability to deduct certain closing costs for tax purposes. hoover (2003) provides a closed-form model that assumes the mortgage is an interest-only loan and considers amortized discount points, a change to interest tax shields, and a new loan balance. fortin, michelson, smith, and weaver (2007) extend table 1 refinancing, paying additional principal, and recasting features/considerations refinance pay additional principal recast new mortgage required (additional closing costs, credit check, and income/asset/debt verification) yes no no requires prior or current principal reduction no yes yes allows borrower to “cash out” increased home equity yes no no change to the original mortgage’s maturity yesa yesb no change to the original mortgage’s interest rate yes no no decrease in future minimum monthly payment maybec no yes athe new mortgage maturity could be shorter or longer than the original mortgage’s maturity. bthe “effective maturity” of the new mortgage will be shorter than the original mortgage’s maturity. cthe future payment depends on the new mortgage’s rate, principal, and terms. monthly payments could increase or decrease. fig. 1. u.s. 30-year fixed mortgage loan rate. april 1971 – december 2021. source: freddie mac, primary mortgage market survey, 30-year fixed rate mortgages. c. mcclatchey / financial services review 31 (2023) 197–209 199 hoover’s (2003) closed-form model by adding an iterative approach that includes lost tax deductions for interest. fortin et al. (2007) find the breakeven period in months is about 35– 40% longer than a calculation ignoring this important variable. virmani and murphy (2010) took a different approach to traditional breakeven analysis and use an option pricing model. they find the performance of an option pricing model is not significantly different from that of a simple 1% heuristic model when using data from 1980 to 2007, although their model performed better than the 2% heuristic. while an options-based approach may appeal to an academic audience, the traditional breakeven method is likely preferred by financial planners and homeowners, as it is easy to explain and understand. the literature remained relatively quiet until the financial crisis of 2008. as the topic resurfaced, attention shifted from evaluating the merits of refinancing viewed from the lens of a mortgage-level transaction to various market frictions in the refinancing process that impede monetary policy attempts to stimulate the economy. these market frictions are more severe during an economic recession. defusco and mondragon (2020) examine loan-level fha mortgage data that span march 2009 to july 2010 (1.3 million fha loans or 15.6 million loan-months). their study comprises six months before and six months after the implementation of the fha’s streamline refinance (slr) program. the late-2009 program implemented two key changes in response to the general deterioration in the mortgage market that imposed restrictions on many borrowers’ eligibility to refinance. first, slr reduced the maximum loan amount for streamlines without an appraisal. borrowers with negative or little equity were now required to pay additional upfront closing costs out-of-pocket. second, slr implemented new income documentation that prohibited unemployed borrowers from refinancing altogether. defusco and mondragon’s (2020) event study finds these frictions resulted in a large decline in refinancing activity that if continued, would have prevented many foreclosures that clearly would have benefited homeowners and the overall u.s. economy. they commented: [these] frictions are likely to bind most for precisely the households whose expenditures may be most sensitive to reduced rates – those with little cash on hand or who recently experienced a negative income shock. this fact may exacerbate the already unequal impacts of recessions by limiting the extent to which reductions in interest rates or other policies that operate through mortgage refinancing benefit lower-income households directly. our results suggest there are a significant number of borrowers that would refinance their mortgages when lower rates are offered, but who cannot do so because of these large frictions in the mortgage market (defusco & mondragon, 2020, p. 2372). following the u.s. economic recovery from the 2008 financial crisis, discussion on market frictions and their effect on the pass-through of monetary policy reemerged as covid-19 triggered a global pandemic that disrupted household employment, income, and the ability of many homeowners to refinance. the circumstances of homeowners facing unemployment due to covid-19 are like those in the 2008 financial crisis in that both experienced an exogenous shock that placed them in a distressed financial situation. two important differences between the 2008 financial crisis and the current environment make the recent landscape unique. 200 c. mcclatchey / financial services review 31 (2023) 197–209 1. the u.s. unemployment rate peaked around 10% in october 2009 but took more than five years to reverse to prior levels (fig. 2). during the covid-19 crisis, unemployment levels peaked much higher in april 2020 at 14.7%, but recovered within one year; though more homeowners experienced an employment or income disruption during covid-19, the effects were short-lived. 2. during covid-19, far fewer borrowers found themselves in homes worth less than their mortgage. homeowners paid down their mortgage from 10 years ago and increases in house prices were robust (golding, goodman, green, & wachter, 2021). this home equity could be used to pay closing costs incurred when refinancing, without pushing a borrower’s loan-to-value ratio (ltv) to an unacceptable level or allow for a “cash out” refinance to cover short-term expenses. following the lead of defusco and mondragon (2020), two studies sought to examine how relaxing employment and income restrictions during the covid-19 crisis might have unleashed restricted refinancing activity. both aimed to hypothesize and to quantify how reducing market frictions could have benefited borrowers and the u.s. economy without an outright bailout to taxpayers. golding et al. (2021) propose a streamlined refinance for federal mortgages that would allow borrowers to refinance with no employment and income tests for no-cash-out refinances: a program they termed as “harp 3.0”. they estimate three million borrowers would have been eligible to refinance under their program which would have contributed $53 billion in additional stimulus per year. gerardi, loewenstein, and willen (2021) examine a no employment and income test for refinancing but add a cash-out option that would allow “in-the-money” borrowers to extract some of their increased housing equity. their program could have saved as much as $280 a month for fannie and freddie loans and $200 a month for ginnie borrowers. 3. recasting (alt-3) today’s homeowner faces a very different landscape. the federal reserve began increasing interest rates in early 2022 making refinancing considerably less attractive for many homeowners. in contrast, a recast does not require a new mortgage at current market interest fig. 2. u.s. unemployment rate. january 2000 – february 2022. source: economic research division, federal reserve bank of st. louis. c. mcclatchey / financial services review 31 (2023) 197–209 201 rates. to this end, we explore the characteristics of recasting to identify borrower-specific instances where it may be a homeowner’s optimal decision. to focus attention on this goal, we ignore tax implications, conversations about the efficient use of free cash flow (e.g., paying down credit card debt, creating an emergency fund, or investing in bonds/stocks), and time value of money concepts, although we note these important extensions would be ripe for follow-up research and discussion. 3.1. the mechanics of a recast a prerequisite to a recast is that the borrower has already paid additional principal or currently has a lump sum available to do so. that is, the loan balance must be lower than the balance calculated by the original amortization schedule. recasting is like making additional principal payments in that neither option requires a new mortgage brings a host of expensive closing costs and lender scrutiny of the borrower’s current financial condition (mclaughlin, 2019). paying additional principal incurs no cost (for most lenders), while a recast does trigger a nominal lender fee (e.g., $150 to $400). because a recast fee is so small, breakeven analysis is moot. paying additional principal shortens the borrower’s effective maturity but has no effect on the mortgage’s monthly minimum payment. under a recast, the lender computes a new payment using the original mortgage’s interest rate and remaining term, calculated on the current principal amount due (hence the requirement of a past or current principal reduction). a recast has no effect on the mortgage’s effective maturity, rather gains come immediately in the form of a lower minimum monthly payment. to illustrate, consider a 30-year (360-month), $200,000 mortgage that carries an apr of 4.99%. the monthly principal and interest payment is $1072.42. one year after closing, the homeowner receives a $40,000 windfall. table 2 provides selected entries of the mortgage’s amortization schedule for alt-2 and alt-3. in the first panel, the borrower pays the $40,000 as an additional principal. there is no change in subsequent monthly payments and the mortgage will be repaid early, between months 238-239. in the right panel, the borrower executes a recast. we see an identical drop in the loan’s beginning balance at month 13, but the required monthly payment falls from $1072 to $855 (recomputed over the mortgage’s remaining 348month life at the original apr of 4.99%). under the recast, the mortgage will remain outstanding for the full 360-month original maturity. 3.2. eligible mortgage types mortgages can broadly be classified as either conventional or government-insured. conventional mortgages generally have a higher minimum down payment, or lower loan-tovalue (ltv) ratio, and higher credit requirements compared to a government-insured loan. conventional mortgages can be conforming or nonconforming. conforming mortgages meet fannie and freddie purchase guidelines regarding the amount, credit and income requirements, down payment, credit score, and suitable property guidelines set by these governmentsponsored enterprises (gses) for the purchase and subsequent repackaging into a mortgage202 c. mcclatchey / financial services review 31 (2023) 197–209 t ab le 2 r ec as t v er su s p ay in g ad d it io n al p ri n ci p al p ay ad d it io n al p ri n ci p al r ec as t m o n th b eg in n in g b al an ce p ay m en t in te re st p ri n ci p al e n d in g b al an ce m o n th b eg in n in g b al an ce p ay m en t in te re st p ri n ci p al e n d in g b al an ce 1 $ 2 0 0 ,0 0 0 $ 1 ,0 7 2 .4 2 $ 8 3 1 .6 7 $ 2 4 0 .7 5 $ 1 9 9 ,7 5 9 1 $ 2 0 0 ,0 0 0 $ 1 ,0 7 2 .4 2 $ 8 31 .6 7 $ 2 4 0 .7 5 $ 1 9 9 ,7 5 9 2 $ 1 9 9 ,7 5 9 $ 1 ,0 7 2 .4 2 $ 8 3 0 .6 7 $ 2 4 1 .7 6 $ 1 9 9 ,5 1 7 2 $ 1 9 9 ,7 5 9 $ 1 ,0 7 2 .4 2 $ 8 30 .6 7 $ 2 4 1 .7 6 $ 1 9 9 ,5 1 7 3 $ 1 9 9 ,5 1 7 $ 1 ,0 7 2 .4 2 $ 8 2 9 .6 6 $ 2 4 2 .7 6 $ 1 9 9 ,2 7 5 3 $ 1 9 9 ,5 1 7 $ 1 ,0 7 2 .4 2 $ 8 29 .6 6 $ 2 4 2 .7 6 $ 1 9 9 ,2 7 5 4 $ 1 9 9 ,2 7 5 $ 1 ,0 7 2 .4 2 $ 8 2 8 .6 5 $ 2 4 3 .7 7 $ 1 9 9 ,0 3 1 4 $ 1 9 9 ,2 7 5 $ 1 ,0 7 2 .4 2 $ 8 28 .6 5 $ 2 4 3 .7 7 $ 1 9 9 ,0 3 1 5 $ 1 9 9 ,0 3 1 $ 1 ,0 7 2 .4 2 $ 8 2 7 .6 4 $ 2 4 4 .7 8 $ 1 9 8 ,7 8 6 5 $ 1 9 9 ,0 3 1 $ 1 ,0 7 2 .4 2 $ 8 27 .6 4 $ 2 4 4 .7 8 $ 1 9 8 ,7 8 6 6 $ 1 9 8 ,7 8 6 $ 1 ,0 7 2 .4 2 $ 8 2 6 .6 2 $ 2 4 5 .8 0 $ 1 9 8 ,5 4 0 6 $ 1 9 8 ,7 8 6 $ 1 ,0 7 2 .4 2 $ 8 26 .6 2 $ 2 4 5 .8 0 $ 1 9 8 ,5 4 0 7 $ 1 9 8 ,5 4 0 $ 1 ,0 7 2 .4 2 $ 8 2 5 .6 0 $ 2 4 6 .8 2 $ 1 9 8 ,2 9 4 7 $ 1 9 8 ,5 4 0 $ 1 ,0 7 2 .4 2 $ 8 25 .6 0 $ 2 4 6 .8 2 $ 1 9 8 ,2 9 4 8 $ 1 9 8 ,2 9 4 $ 1 ,0 7 2 .4 2 $ 8 2 4 .5 7 $ 2 4 7 .8 5 $ 1 9 8 ,0 4 6 8 $ 1 9 8 ,2 9 4 $ 1 ,0 7 2 .4 2 $ 8 24 .5 7 $ 2 4 7 .8 5 $ 1 9 8 ,0 4 6 9 $ 1 9 8 ,0 4 6 $ 1 ,0 7 2 .4 2 $ 8 2 3 .5 4 $ 2 4 8 .8 8 $ 1 9 7 ,7 9 7 9 $ 1 9 8 ,0 4 6 $ 1 ,0 7 2 .4 2 $ 8 23 .5 4 $ 2 4 8 .8 8 $ 1 9 7 ,7 9 7 1 0 $ 1 9 7 ,7 9 7 $ 1 ,0 7 2 .4 2 $ 8 2 2 .5 1 $ 2 4 9 .9 2 $ 1 9 7 ,5 4 7 1 0 $ 1 9 7 ,7 9 7 $ 1 ,0 7 2 .4 2 $ 8 22 .5 1 $ 2 4 9 .9 2 $ 1 9 7 ,5 4 7 1 1 $ 1 9 7 ,5 4 7 $ 1 ,0 7 2 .4 2 $ 8 2 1 .4 7 $ 2 5 0 .9 6 $ 1 9 7 ,2 9 6 1 1 $ 1 9 7 ,5 4 7 $ 1 ,0 7 2 .4 2 $ 8 21 .4 7 $ 2 5 0 .9 6 $ 1 9 7 ,2 9 6 1 2 $ 1 9 7 ,2 9 6 $ 1 ,0 7 2 .4 2 $ 8 2 0 .4 2 $ 2 5 2 .0 0 $ 1 9 7 ,0 4 4 1 2 $ 1 9 7 ,2 9 6 $ 1 ,0 7 2 .4 2 $ 8 20 .4 2 $ 2 5 2 .0 0 $ 1 9 7 ,0 4 4 $ 4 0 ,0 0 0 .0 0 $ 4 0 ,0 0 0 .0 0 1 3 $ 1 5 7 ,0 4 4 $ 1 ,0 7 2 .4 2 $ 6 5 3 .0 4 $ 4 1 9 .3 8 $ 1 5 6 ,6 2 5 1 3 $ 1 5 7 ,0 4 4 $ 8 5 4 .7 2 $ 6 53 .0 4 $ 2 0 1 .6 8 $ 1 5 6 ,8 4 2 1 4 $ 1 5 6 ,6 2 5 $ 1 ,0 7 2 .4 2 $ 6 5 1 .3 0 $ 4 2 1 .1 2 $ 1 5 6 ,2 0 3 1 4 $ 1 5 6 ,8 4 2 $ 8 5 4 .7 2 $ 6 52 .2 0 $ 2 0 2 .5 2 $ 1 5 6 ,6 4 0 1 5 $ 1 5 6 ,2 0 3 $ 1 ,0 7 2 .4 2 $ 6 4 9 .5 5 $ 4 2 2 .8 8 $ 1 5 5 ,7 8 1 1 5 $ 1 5 6 ,6 4 0 $ 8 5 4 .7 2 $ 6 51 .3 6 $ 2 0 3 .3 6 $ 1 5 6 ,4 3 6 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 2 3 7 $ 2 ,4 0 3 $ 1 ,0 7 2 .4 2 $ 9 .9 9 $ 1 ,0 6 2 .4 3 $ 1 ,3 4 0 2 3 7 $ 8 2 ,6 7 8 $ 8 5 4 .7 2 $ 3 43 .8 0 $ 5 1 0 .9 2 $ 8 2, 1 6 7 2 3 8 $ 1 ,3 4 0 $ 1 ,0 7 2 .4 2 $ 5 .5 7 $ 1 ,0 6 6 .8 5 $ 2 7 4 2 3 8 $ 8 2 ,1 6 7 $ 8 5 4 .7 2 $ 3 41 .6 8 $ 5 1 3 .0 4 $ 8 1, 6 5 4 2 3 9 $ 2 7 4 $ 1 ,0 7 2 .4 2 $ 1 .1 4 $ 1 ,0 7 1 .2 8 ! $ 7 9 8 2 3 9 $ 8 1 ,6 5 4 $ 8 5 4 .7 2 $ 3 39 .5 4 $ 5 1 5 .1 8 $ 8 1, 1 3 9 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * 3 5 7 $ 3 ,3 8 4 $ 8 5 4 .7 2 $ 1 4 .0 7 $ 8 4 0 .6 5 $ 2 ,5 4 3 3 5 8 $ 2 ,5 4 3 $ 8 5 4 .7 2 $ 1 0 .5 7 $ 8 4 4 .1 4 $ 1 ,6 9 9 3 5 9 $ 1 ,6 9 9 $ 8 5 4 .7 2 $ 7 .0 6 $ 8 4 7 .6 5 $ 8 5 1 3 6 0 $ 8 5 1 $ 8 5 4 .7 2 $ 3 .5 4 $ 8 5 1 .1 8 $ 0 n o te . s ha d ed ro w s h ig h li g h t im p o rt an t ch an g es to th e am o rt iz at io n sc h ed u le w h en a o n eti m e $ 4 0 ,0 0 0 w in d fa ll is ap p li ed as ei th er ad d it io n al p ri n ci p al o r u se d to re ca st a 3 0 -y ea r m o rt g ag e, o n e y ea r af te r th e lo an ’s o ri g in at io n . c. mcclatchey / financial services review 31 (2023) 197–209 203 backed security (mbs).1 recasting is allowed for conventional, conforming mortgages. eligibility for a nonconforming conventional mortgage is at the lender’s discretion. government-insured mortgage programs include fha, va, pih, and usda. while fannie and freddie actually purchase mortgages and either hold them in portfolio or subsequently securitize them and issue a mbs, ginnie mae does not purchase and securitize mortgages. lenders originate these mortgages and either hold them in portfolio or privately securitize them. ginnie’s blessing assures the investors in the mbs that they will receive payments without disruption. mortgages backed by ginnie mae are not eligible for recasting. after purchasing a home, a homeowner may borrow against the home’s equity. a home equity loan provides a lump-sum amount, often used for a major home remodel or large immediate one-time expense. it generally carries a fixed interest rate. under a home equity line of credit (heloc), the borrower qualifies for an amount that they can borrow and pay back as many times as needed until the mortgage’s draw period comes to an end. during the draw period, the lender requires only monthly interest payments (perhaps subject to a minimum amount), and the mortgage carries a variable interest rate. the ability to recast a home equity loan or a heloc is lender-specific. eligibility for all mortgage programs is detailed in table 3. 3.3. lender-specific requirements common lender-specific eligibility requirements are that (a) the mortgage is current with no outstanding amount due, (b) the borrower has had no past due payments within the last 12months, and (c) the recast application is submitted more than 90 days after the mortgage’s closing date. for an adjustable rate mortgage, the recast application must generally be submitted more than 90 days before any scheduled rate change. also common is a reduction in mortgage principal (termed principal curtailment), although rules vary from lender to lender. some lenders have no minimum, while others require curtailment of at least $5,000 to $20,000 before a mortgage can be recast. for mortgages that do qualify for recasting, the process is far simpler than refinancing. the required documentation is minimal. recast applications we generally see are 1-2 pages long, in which most of the content is generic personal and mortgage information. additionally, bank fees (in the range of $150 to $500) are nominal. a bank can limit the number of times a mortgage can be recast, so borrowers should ask if today’s decision will constrain future opportunities. table 3 recasting and allowable mortgage types mortgage type recast allowed? conventional and conforming fannie and freddie yes government-backed (ginnie) fha, va, pih, usda no home equity mortgage/heloc maybe lender discretion source: treece and cetera (2020) and motley fool (2020). 204 c. mcclatchey / financial services review 31 (2023) 197–209 4. when might a recast be optimal? a recast is a useful tool for borrowers that previously paid additional principal or currently have funds available to make an additional lump sum payment. for this subset, recasting may be beneficial for borrowers that (a) want to reduce their monthly payment, rather than pay off their mortgage faster; (b) have suffered a decline in their credit score, income, or assets and cannot qualify for a refinance; (c) want to simultaneously sell and purchase a new home, particularly in a competitive market; (d) have an original mortgage rate more competitive than current rates; or (e) want to eliminate private mortgage insurance (pmi) early. to illustrate, we consider five fictitious clients. 4.1. client 1 – i recently received a one-time “windfall” from time-to-time, individuals or families may receive an unusual one-time cash inflow. some may be anticipated, such as an annual tax refund, while others are unexpected like a work bonus, inheritance, or even gambling winnings. refinancing (alt-1) does not require additional money be paid on the original mortgage (so a one-time windfall by itself would not trigger a refinance), but if current interest rates are lower than the original mortgage rate, one could capitalize on the opportunity to both lower the principal amount and lock in a lower rate. this option may not be optimal for borrowers for which the lower rate is insufficient to satisfy their breakeven calculation (fortin et al., 2007) or for those who are unable to qualify for a new mortgage due to a change in circumstances (income, credit score, and debt/income ratio). paying additional principal (alt-2) will leave monthly payments unchanged but shorten the mortgage’s effective maturity. recasting (alt-3) lowers the monthly payment but keeps the original maturity date. at first glance, the decision appears to be borrower-specific; that is, each borrower would have to choose whether receiving the benefit later (alt-2) or now (alt-3) is best for them. however, a recast carries a valuable option that paying additional principal does not. if a borrower recasts, they can enjoy a lower required minimum monthly payment, and still retain the option to pay a higher monthly amount and shorten the mortgage’s effective maturity. the cost of recasting (again, relatively small) can be likened to the purchase price of a call option that provides the homeowner additional flexibility in the future, should it become valuable. 4.2. client 2 – i am closing on my new home, but before selling my existing home purchasing a new home and moving entails many stressful logistics. for renters, the difficulty of the transaction is certainly reduced, but for those trying to sell an existing home at the same time they are trying to purchase a new one, the complexity of the transaction quickly escalates, especially in a competitive real estate market (lerner, 2021). buyers can submit an offer contingent on the sale of their existing home, although most sellers would prioritize offers without a contingency to ensure the contract does not fall through. strong buyers that can qualify to carry two mortgage payments for a short period of time could c. mcclatchey / financial services review 31 (2023) 197–209 205 close on the new home before selling their existing home. once the original home sells, the proceeds can be used to initiate a refinance (alt-1) or a recast (alt-3) to reduce the new mortgage’s monthly payment during its remaining term. alt-1 would be very expensive as it would trigger a new set of closing costs and subject the borrower to additional underwriting scrutiny. alt-3 appears superior – just pay the nominal recast fee. 4.3. client 3 – can i afford to retire or even retire early? for many households, monthly income declines at retirement. social security benefits are payable as early as age 62 but are 29% lower than full benefits at age 67 (social security administration, 2022). while some may have supplemental sources of income, like withdrawals from a 401k or income from rental properties, reducing monthly expenses may be helpful to maintain one’s standard of living, or even present the option to retire early. refinancing (alt-1) is only advantageous if rates have declined or at least remained flat, the breakeven period can be met, and the household can qualify for a new mortgage. further, alt-1 triggers significant closing costs. while refinancing in this situation could reduce the monthly payment and allow for on-time or early retirement, it is not a viable option outside these limited circumstances. paying additional principal (alt-2) could help the household pay off their mortgage by their desired retirement date, although the monthly minimum payment will not be affected. recasting (alt-3) lowers the minimum payment, perhaps making retirement possible and/or more comfortable on a month-to-month basis for those willing to have a mortgage payment during retirement years. 4.4. client 4 – my family is facing an unexpected financial hardship alt-1 will likely not help households facing financial hardship because it requires qualifying for a new mortgage. if the borrower’s debt to income ratio increased, current income fallen, or credit score declined, they may not be able to obtain a new mortgage. if the borrower can qualify, alt-1 may be viable if rates have fallen significantly since the original mortgage’s origination date, or if the borrower can extend the mortgage’s maturity so the minimum monthly payment falls. alt-2 in this situation is likely undesirable; paying additional principal in a time of hardship is generally not feasible as the focus is likely on helping the family’s current situation. alt-3 lowers the mortgage’s minimum monthly payment. if prior principal reductions were significant, the reamortized payment may be low enough to help a family stay in their home and endure an unexpected, but short-term, hardship. 4.5. client 5 – can i terminate pmi on a conventional mortgage? a conventional mortgage carrying a ltv ratio greater than 80% requires the borrower to purchase private mortgage insurance (pmi).2 pmi helps limit the lender’s losses (not the borrower’s) if the lender must foreclose on the property. borrowers can request pmi be dropped when their ltv reaches 80% of the home’s original value. the homeowners protection act of 1998 (hoepa) requires a lender terminate pmi when (a) the mortgage 206 c. mcclatchey / financial services review 31 (2023) 197–209 balance reaches 78% of the original purchase price if the borrower is in good standing and has not missed any scheduled payments, or (b) the mortgage reaches the half-way mark of its amortization schedule. pmi may also be terminated if, two years after purchase, the home price has appreciated, and a new appraisal proves the current ltv is 75% or less of the new appraised value.3 for conventional mortgages, a refinance (alt-1) makes sense if interest rates have declined, the borrower’s credit has not significantly changed, and the breakeven calculation has been met. similar to previous situations, alt-1 reduces the monthly payment and triggers significant up-front closing costs but now carries the added benefit of eliminating the monthly pmi. if a borrower has additional funds to reach the 80%, 78%, or 75% ltv thresholds either alt-2 or alt-3 might be optimal. like client 1, recasting may be superior as it reduces the minimum payment but retains the option (for a nominal recast fee) to still pay a higher monthly amount and repay the mortgage faster should the borrower elect to do so. 5. conclusion anecdotal evidence suggests the option to recast is not well-known compared to refinancing or paying additional principal. refinancing has been the most visible and researched option, as evidenced in our literature review. this option’s popularity among homeowners and academics is likely due to the general decline in interest rates since the 1980s (fig. 1). covid-19 altered many homeowners’ financial options due to temporary employment disruptions, a decline in household credit scores, or an increase in debt-to-income ratios. though rising home values may leave refinancing a viable alternative for some, recent interest rate increases (qtrs. 1-2, 2022) indicate this avenue may dissipate for many in the near term. paying additional principal means the mortgage will be repaid before its initial maturity, but leaves required monthly payments unchanged. recasting reduces the minimum monthly payment but leaves the original mortgage maturity intact. this complex decision depends on a borrower’s individual circumstances, the forecast for future mortgage rates, lender-specific rules and requirements, and consideration of future changes to the u.s. tax code. in today’s environment, we find recasting may be a tool for households that (a) want to retire early, (b) have received a one-time financial windfall, (c) want to eliminate private mortgage insurance, (d) have experienced an unexpected hardship, or (e) have bought a new home before the sale of their existing home. overall, the option to recast may become a more commonplace term in the vocabulary of homeowners, faculty, and finance professionals. notes 1 fannie mae and freddie mac are government-created enterprises (gses) that buy mortgages from lenders and hold them in portfolios or turn them into mortgagebacked securities. conforming loans meet requirements regarding: (a) maximum loan amount (different for one-, two-, three-, and four-family dwellings, (b) credit c. mcclatchey / financial services review 31 (2023) 197–209 207 and income requirements, (c) down payment, and (d) suitable property type. loans that do not meet fannie and freddie credit requirements are called b, c, and d paper loans (versus a paper if they do meet credit requirements). loans above fannie and freddie guidelines are called jumbo loans. 2 pmi ranges from 0.25% to 2% of the loan balance per year, depending on the size of the down payment and mortgage, the loan term, and the borrower’s credit score. 3 fha loans carry similar insurance termed mortgage insurance premium (mip). if the original down payment is less than 10%, mip cannot be canceled; the only way to terminate mip is to refinance. if the original down payment is 10% or more, the borrower can cancel monthly mips after 11 years. va loans require an upfront “funding fee” and no monthly insurance premium. references bennett, p., keane, f., & mosser, p. (1999). mortgage refinancing and the concentration of mortgage coupons. current issues in economics and finance, federal reserve bank of new york, 5(4). brady, p. j., canner, g. b., & maki, d. m. (2000). the effects of recent mortgage refinancing. federal reserve bulletin, 86, 441–450. chen, r. (1997). a refinancing decision model using spreadsheet software. national public accountant, 42, 44–48. defusco, a. a., & mondragon, j. (2020). no job, no money, no refi: frictions to refinancing in a recession. the journal of finance, 75, 2327–2376. https://doi.org/10.1111/jofi.12952 fortin, r., michelson, s., smith, s. d., & weaver, w. (2007). mortgage refinancing: the interaction of breakeven period, taxes, npv, and irr. financial services review, 16, 197–209. gerardi, k., loewenstein, l., & willen, p. s. (2021). evaluating the benefits of a streamlined refinance program. housing policy debate, 31, 51–65. https://doi.org/10.1080/10511482.2020.1850014 golding, e., goodman, l. s., green, r., & wachter, s. (2021). the mortgage market as a stimulus channel in the covid-19 crisis. housing policy debate, 31, 66–80. https://doi.org/10.1080/10511482.2020.1850015 hoover, g. (2003). the mortgage refinance decision: an equation-based model. financial services review, 12, 319–337. johnson, r., & randle, p. (1996). the mortgage refinancing decision: a breakeven approach. the cpa journal, 66, 69–71. johnson, r., & randle, p. (2003). the mortgage refinancing decision: updated spreadsheet. the cpa journal, 73, 60–61. lerner, m. (2021). recasting is another way, besides refinancing, to save money on your mortgage. the washington post. available at https://www.washingtonpost.com/business/2021/03/24/recasting-is-anotherway-besides-refinancing-save-money-your-mortgage/ mclaughlin, k. (2019). to reduce payments, recast your mortgage: with an infusion of cash, some banks will re-amortize a mortgage to a lower monthly bill. the wall street journal. available at https://www.wsj.com/ articles/to-reduce-payments-recast-your-mortgage-11549466415?adobe_mc¼mcmid%3d17035339205025 685463475251611284040076%7cmcorgid%3dcb68e4ba55144caa0a4c98a5%2540adobeorg% 7cts%3d1696883385 motley fool. (2020). could mortgage loan recasting work for you? here’s how it works. available at https:// www.usatoday.com/story/money/2020/07/30/what-is-mortgage-loan-recasting/112446088/ noyan, a., & eugene, p. (1993). revisiting the economics of mortgage refinancing. journal of retail banking, 15, 45–48. rose, c. c. (1992). real estate investment refinancing. real estate finance, 8, 57–62. social security administration. (2022). retirement benefits. publication no. 05-10035. available at https:// www.ssa.gov/pubs/en-05-10035.pdf 208 c. mcclatchey / financial services review 31 (2023) 197–209 stolba, s. l. (2021). mortgage debt sees record growth despite the pandemic. available at https://www.experian. com/blogs/ask-experian/how-much-americans-owe-on-their-mortgages-in-every-state/ treece, k., & cetera, m. (2020). mortgage recasting: can it save your money? forbes advisor. available at https://www.forbes.com/advisor/mortgages/mortgage-recasting/ virmani, s., & murphy, a. (2010). an empirical analysis of residential mortgage refinancing decision-making. journal of housing research, 19, 129–138. https://doi.org/10.1080/10835547.2010.12092021 c. mcclatchey / financial services review 31 (2023) 197–209 209 the financial literacy of generation y and the influence that personality traits have on financial knowledge: evidence from canada robert n. killinsa,* aschool of accounting and financial services, seneca college, 1750 finch avenue e., toronto, ontario m2j 2x5 canada abstract this article examines the financial literacy of the generation y age cohort and explores how personality traits influence individual’s financial knowledge. using a detailed financial literacy survey, multiple areas of financial literacy are measured (investments, budgeting, economics, risk management, and retirement planning) along with the well-known big five personality traits. the findings of this article suggest that the generation y cohort is more knowledgeable in budgeting and risk management segments of financial literacy but lack knowledge in retirement planning. secondly, extraversion and conscientiousness are both important personality traits when regressed on individuals overall financial literacy levels. these finding help develop the insights into how behavioral and personality traits influence the cognitive and financial decision-making ability of individuals. © 2017 academy of financial services. all rights reserved. jel classification: a20; i20; d04 keywords: financial literacy; behavioral finance; personality; generation y 1. introduction individuals around the globe are confronted with financial decisions on a daily basis. with the evolving information technology landscape, financial information is more abundant than ever before. following conventional economic theory, one would conclude that with all this * corresponding author. tel.: �1-416-491-5050 ext. 22671; fax: �1-416-493-0210. e-mail address: robert.killins@senecacollege.ca financial services review 26 (2017) 143–165 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. readily available information, individuals should be able to make better financial decisions. however, to make better decisions, individuals need to know how to interpret and use this information effectively. further, even with the ability to understand and effectively apply the abundant amount of financial information, society is faced with several more choices of financial products and services. these changes in the financial services industry may lead to what is known as the “paradox of choice.” as schwartz (2004) suggests, the fact that some choice is good does not necessarily mean that more choice is better. the financial literacy of individuals has been assessed in many countries, and the consensus is that people lack the financial knowledge to make use of the new information that is provided to them in the new age of information sharing. the recent financial product developments now make it even more important for individuals to understand the financial decisions they are making as the “human touch” of traditional banking services are now being transitioned to internet-based platforms and algorithms. many of these new “fintech” products are adopted by the generation y age cohort, which in previous financial literacy research has been shown to exhibit low levels of financial literacy. this generation has to make critical financial decisions early in life, such as going to postsecondary education, buying a vehicle, housing decisions, among others that are vital to their future financial positions. thus, it is important for practitioners, academics, and regulators to better understand the financial literacy levels of the generation y demographic and this study aims add to the growing amount of information on this group. behavioral finance has also emerged as an alternate view on decision-making from the traditional finance theory which depended on rational decision-making models. from the behavioral finance stream, previous literature has investigated how personality traits impact investment choices and other financially related decisions, but little research has focused on how personality traits affect one’s financial literacy. this study intends to fill such gap in the literature and provides academics and practitioners with a new understanding of how personality traits influence one’s ability to digest and implement the financial information provided to them. additional stakeholders can benefit from understanding and increasing the financial literacy of canadians. financial institutions can benefit from understanding the financial literacy of the individuals whom they serve to better cater products and services to them; thus, improving customer retention and reducing the potential risks financial service firms face via litigation or debt defaults. additionally, policymakers and regulators will benefit from insights into the financial literacy of the generation y cohort in canada. beginning in 2009 the canadian federal government initiated a financial literacy task force that has further evolved into what is now the national strategy for financial literacy–count me in, canada. the national strategy focuses on bringing together various stakeholders in the financial services landscape to further the development and understanding of financial literacy in canada. this research complements the canadian government initiative by providing a detailed view of the financial literacy levels of young canadians and also testing how behavioral (personality) traits can influence their financial understanding. the remainder of this article is as follows: section 2 will provide an overview of the relevant literature on financial literacy and behavioral finance. section 3 will outline the data sources. section 4 describes the methodology. section 5 will consist of the results. section 144 r.n. killins / financial services review 26 (2017) 143–165 6 will provide a discussion of the results including the implications. section 7 will provide the conclusion and recommendations for future research. 2. background and literature review 2.1. financial literacy financial literacy is defined as the ability to process economic information and make an informed decision about financial planning, wealth accumulation, debt, and pensions (lusardi, 2015). over the past decade, financial literacy has been pushed into the spotlight by different government organizations around the globe. the organization for economic co-operation and development (oced) has initiated some financial literacy programs and studies via the program for international student assessment to assess and develop the financial literacy of youth across the globe. on a national scale, many developed countries now see the importance of financial literacy, which can bring increased financial stability to financial markets and have launched campaigns and surveys to promote and collect data on financial literacy. in the united states, beginning in 2004 financial literacy questions were added to the health and retirement study (hrs) and later added to several other surveys including the national financial capability survey (nfcs). data from these surveys indicate low levels of financial literacy by individuals in the united states (lusardi & mitchell, 2011a; lusardi & mitchell, 2011b lusardi, mitchell, & curto, 2010). other studies such as bernheim (1995, 1998), lusardi and mitchell (2014), lusardi and tufano (2009), and smith and stewart (2009) all draw similar conclusions suggesting the lack of knowledge of basic financial concepts for the general u.s. population. additional data from the jump$tart coalition for personal financial literacy and national council on economic education focuses on a younger sample of the u.s. population. using data from this survey, mandell (2008) along with mandell and klien (2007) find high school students in the u.s. also receive a poor grade for their financial literacy. outside of the united states, several studies in europe and asia has yielded similar findings.1 in canada, the quantitative survey data has started to develop but still lags behind the united states and some european countries. the primary sources of financial literacy data are the survey’s sponsored by the financial agency of canada (fcac) and statistics canada financial capability survey (fcs). the most recent fcs in 2014 indicated that 60% of adults rate their financial knowledge as “fair” or “poor” and 80% of young canadians are not confident in their financial knowledge. boisclair, lusardi, and michaud (2014) find that 42% of canadians were able to correctly answer three questions centered on interest compounding, inflation, and risk diversification. these findings put canada ahead of the united states but lagging behind some european countries. also, they also conclude that the young and the old, women, and minorities score lower in their financial literacy level. buckland (2010) draws qualitative data from a subsection of low-income canadians and finds that the “low income” population of canada is perhaps more financially literate than the quantitative data suggests. he suggests that the low-income sample he drew from cope well 145r.n. killins / financial services review 26 (2017) 143–165 with budgeting, credit, and knowledge of government programs but may lack knowledge in institutional policies and deeper financial life goals. 2.2. generation y following brosdahl and carpenter (2011) individuals born after 1981 are classified as the generation y age cohort. a key formative characteristic for generation y is early and frequent exposure to technology, which has advantages and disadvantages regarding cognitive, emotional, and social outcomes (immordino-yang, chrisodoulou, & singh, 2012). generation y consumers have also benefited from the increased availability of customized products and personalized services. they “want it all” and “want it now,” particularly in relation to work/pay and benefits, career advancement, work or life balance, interesting work and being able to make a contribution to society via their work (howe & strauss, 2009; ng, schweitzer, & lyons, 2010). within the financial landscape, youth have been targets of financial institutions growth plans for decades. advertisements from the financial services industry can be seen on display across college campuses across canada. students can sign up for credit cards on campus with little information provided and little income to pay for the credit extended to them. cudmore, patton, ng, and mcclure (2010) documents the importance of financial literacy to “protect” the millennial generation from aggressive and sometimes deceptive tactics by financial institutions. innovations by financial institutions via the technology sector has created a new avenue for financial institutions to strengthen their positions with this up and coming generation. prior studies on the financial literacy of the college-aged demographic show that they are not very knowledgeable about personal finances pillars (chen & volpe, 1998; harrison & chudry, 2011; lusardi, 2015). thus, there is a potential for severe risks to the millennial demographic but also to the financial markets as a whole if young adults turn to “fintech” resources without the necessary knowledge of the economic landscape. further, as the baby boomer transition into retirement, an estimated $750 billion dollars in wealth will be transferred to a younger generation in canada. this leads to significant implications for the entire financial system, and increased emphasis on the importance of financial literacy for the young generation is warranted. 2.3. psychological elements and financial decision making a fundamental assumption in the traditional literature on financial markets centers on rational expectations theory, which states that individuals use all the relevant information available when forming their expectations about economic decisions. in this traditional theory, to make good financial decisions, an individual requires proper information as well as the ability to process this information. in the recent past, the traditional view on financial markets has been challenged, and the behavioral characteristics of individuals have been used to better understand the choices people make in the financial markets (shiller, 2003; shleifer, 2000; thaler & de bondt, 1993). personality traits are important because they influence the way individuals interact within a particular environment. the big five taxonomy has been used to provide evidence linking 146 r.n. killins / financial services review 26 (2017) 143–165 personality traits to various elements of a person’s life including health and longevity, social, work, and academic outcomes.2 when focusing on how personality traits may influence one’s financial environment, the behavioral finance literature has provided some possible connections. mayfield, perdue, and wooten (2008) research indicate that individuals who are more extroverted intend to engage in short-term investing, while those who are higher in neuroticism and risk aversion avoid this activity. risk adverse individuals also do not engage in long-term investing. further, individuals who are more open to experience are inclined to engage in long-term investing; however, openness did not predict short-term investing. further, wilhelm, varcoe, and fridrich (1993) found that conscientious people are more likely to be savers. brandstatter (1996) found emotional instability and introversion are linked to the increased likelihood of one saving over spending. more recently, davis and runyan (2016) explore individual’s personality characteristics and financial satisfaction. their findings suggest that trait characteristics such as the need for material resources and emotional instability affect one’s overall financial satisfaction. although research linking personality traits and financial literacy is limited, previous literature does provide us with potential linkages. garcia (2011) discusses both behavioral finance and financial literacy. one discussion in her article centers on the topic of “bounded rationality” of individuals, which suggests that it is almost impossible for individuals to process the sizable amount of financial information in the current informational environment and thus individuals take “shortcuts” when making financial decisions. these shortcuts are what kahneman and tversky (2000) referred to as biases and heuristics. lusardi (2008) concluded that to improve individual’s financial decision-making ability, the process must be simplified, and barriers for processing information must be reduced. one shortcoming of the lusardi (2008) conclusion is that not all individuals are alike in their abilities to filter, understand, and interpret financial information and thus furthering the understanding of what impacts the financial literacy levels of individuals may aid financial planning practitioners in determining whether biases or shortcuts need to be monitored. previous research has identified several important control variables to include when examining the level of one’s financial literacy and behaviors. these variables include gender, age, race, marital status, presence of children, employment status, education, and income (fernandes, lynch, & netemeyer, 2014; robb & woodyard, 2011; xiao, chen, & chen, 2014; xiao, chen, & sun, 2015; zick, mayer, & glaubitz, 2012). for example, minorities and those with less education and income tend to score lower on measures of financial knowledge and women tend to score lower than men. the studies mentioned above provide evidence that financial literacy is sensitive to a variety of elements unique to each individual but does not explore the impact personality, or behavioral traits have on one’s financial literacy. previous studies have shown potential pathways between personality traits and cognitive performance. specific personality traits such as conscientiousness as well as openness are positively correlated with academic performance (goldberg, sweeney, merenda, & hughes, 1998; noftle & robins, 2007). further, davis and runyan (2016) suggest conscientious individuals might be more effective problem solvers lead to better understating of the complexities of the personal financial landscape. this leads this research to suggest that conscientiousness will be the strongest 147r.n. killins / financial services review 26 (2017) 143–165 personality trait influencing one’s financial literacy and suggests it will have a positive impact. hypothesis 1: individuals scoring high in conscientiousness will score higher in their overall financial literacy. secondly, saucier and goldberg (1996) drawing from their research on the big five trait domains, conceptualized the openness domain as “intellect,” emphasizing its connection to creativity, abstract thinking, depth of thought, and other intellective qualities. thus, this article will test the impact of openness on one’s financial literacy and suggests that openness will have a positive impact. hypothesis 2: individuals scoring high in openness will score higher in financial literacy. finally, connor-smith and flachsbart (2007) found extraversion was predictive of concrete problem-solving skills as well as coping strategies. further, extroverted individuals, who are optimistic and outgoing, can be seen as more likely to consult someone for financial advice and take the initiative to gather additional resources to become better informed. thus, the third hypothesis suggests extroverted individuals will have higher levels of financial literacy. hypothesis 3: individuals scoring high in extroversion will score higher in financial literacy. 3. data this study uses the “big five” personality framework to capture the personality characteristics of each individual. this research uses five-factor model (ffm), or the “big five,” because it is, unquestionably, the most universal and widely accepted trait framework in the history of personality psychology (john & srivastava, 1999).3 specifically, the article uses the 44-item inventory (version 4a and 54) that measures an individual on the big five factors (dimensions) of personality.4 the first dimension refers extraversion or introversion to the preference of interacting with others or being alone. the second agreeableness includes trust, moderation, altruism, cooperativeness, modesty and kindness or compassion (mccrae & costa, 2010). the third conscientiousness involves competence, order, sense of duty, a tendency toward achievement, self-discipline, and cautiousness. this dimension has been associated with academic success (mccrae & costa, 2010). the fourth neuroticism is defined as a personality dimension characterized by a tendency to experience negative emotions and is associated with emotional distress or negative effect. people scoring high in neuroticism report greater anxiety, depression, hostile anger, impulsiveness, self-consciousness, and emotional vulnerability. the fifth openness includes esthetic openness and openness to feelings, activities, ideas, and values (mccrae & costa, 2010). following the standard approach in the literature this research creates the standardized cronbach alpha reliability index to assess the internal consistency of the five items, leading to the following reliability measures for the sample: conscientiousness (0.77), extraversion (0.86), agreeableness (0.75), neuroticism (0.81), and openness (0.73). these diagnostic measures of 148 r.n. killins / financial services review 26 (2017) 143–165 reliability fall within the previous reliability findings of the various big five personality traits surveys and provide confidence (above 0.70) of the reliability of the assessment in this study. the financial literacy questions were developed by the author by incorporating previous financial literacy surveys such canadian financial capabilities survey (cfsc) conducted by statistics canada5 along with academic studies by chen and volpe (1998) and lusardi (2015). this new survey measures not only the overall financial literacy of an individual but also how knowledgeable a participant is in the different subcategories of financial literacy that include; investments, budgeting, economics, risk management, and retirement planning. thus, this survey provides further details into the strengths and weaknesses that individuals may have in regards to their financial knowledge and allows policymakers, practitioners, and educators to potentially focus more resources on specific areas of the financial services environment (see appendix for financial literacy survey). the data from the survey was collected via two different college campuses in canada during october 2016 and april 2017. both schools were located in large metropolitan centers and benefited from a broad range of ethnic diversity, educational backgrounds, and prior financial experiences. in total, 157 individuals participated in the study in which 149 of these were able to be used (eight incomplete surveys). while the sample size is somewhat smaller than other studies using national datasets, the power and generalizability of this study remain robust with the inclusion of four (base model) and nine (full model) independent variables.6 table 1 shows the descriptive statistics of the variables used in the study. panel a of table 1 illustrates the distribution of individuals based on gender, race, and employment status. panel b of table 1 outlines the financial literacy and big five personality variables. the personality traits are ranked on a five-point scale, and the data suggests we have a wide variety of different personality traits across individuals in the study. the first financial literacy variable (flo) indicates that the highest score recorded on the financial literacy survey was 97% and the lowest only 21% with a mean score of 67%. out of the five different subcategories of financial literacy, participates scored well in budgeting, economics, and risk management but are weaker in categories of investments and retirement planning. table 2 documents the correlation matrix for the financial literacy and personality measures. high degrees of correlation exists between the financial literacy variables, but little correlation is shown between the big five personality traits that will be included as independent variables in the regression which should lead to robust results. 4. empirical methodology this study uses a multiple regression approach to investigate the impact personality traits have on financial literacy. the initial model development was based on previous literature that has suggested that gender, race, and employments status tend to impact one’s level of financial literacy (fernandes et al., 2014; xiao et al., 2014; xiao et al., 2015). further, because of canada immigration policies and the high proportion of international students in postsecondary education, this study controls and tests whether 149r.n. killins / financial services review 26 (2017) 143–165 new immigrant status impacts the level of financial literacy. thus, the base model takes on the following form; fli � �0 � �1geni � �2racei � �3empi � �4resi � � (1) where fli is the financial literacy score of the participant i. this study uses an overall financial literacy score (flo) based on 28 questions survey, along with financial literacy scores for subsections of financial literacy which includes: investments, budgeting, economics, risk management, and retirement planning. geni is a dummy variable equal to 1 for male participates and 0 for female. race is a dummy variable equal to 1 if the participant is white and 0 if identified as a minority. emp is a dummy variable equal to 1 if the participant has been employed for over two years and 0 if not. finally, res is a dummy variable taking a value of 1 if the participant has been a resident of canada for over five years and zero if not. consistent with the previous literature, we would expect that �1 � 0, �2 � 0, �3 � 0 and �4 � 0. table 1 descriptive statistics panel a: categorical and dummy variables variable n % gender male 86 57.7 female 63 42.3 education high school 117 78.5 college diploma 17 11.4 bachelor degree 11 7.4 master’s degree 4 2.7 marital status single 124 83.2 married 22 14.8 divorced 3 2.0 years in canada 0–2 19 12.8 3–5 56 37.6 6–8 12 8.1 8 or more 62 41.6 employment status employed 96 35.6 unemployed 53 64.4 gross income per year �$15,000 69 46.3 $15,001 to $30,000 34 22.8 $30,001 to $65,000 22 14.8 �65,001 24 16.1 ethnicity white 49 32.9 african american/black 16 10.7 hispanic/latino 21 14.1 asian 36 24.2 other 27 18.1 150 r.n. killins / financial services review 26 (2017) 143–165 building off the base model, this research will incorporate the personality measures to explore how they may influence individual’s financial literacy. the second model takes on the following form; fli � �0 � �1geni � �2 racei � �3empi � �4resi � �5bfiei � �6bfiai � �7bfici � �8bfini � �9bfioi � � (2) where the measurement of fl, gen, race, emp, and res are as above in the base model. bfie is the measurement of extraversion/introversion, bfiai is the measurement of agreeableness, bfici is the measurement of conscientiousness, bfin is the measurement of neuroticism and bfioi is the measurement of openness. we would expect that �5 � 0, �6 � 0, �7 � 0, �8 � 0, and �9 � 0. 5. results the results from the base regression model (1) can be seen in table 3. the findings suggest that only race has a significant impact on financial literacy levels of the generation y demographic in canada. the coefficient of 0.129 indicates that minorities exhibit approxpanel b: descriptive statistics (financial literacy and personality traits) variable n minimum maximum mean standard deviation skewness kurtosis flo 149 0.19 0.97 0.71 0.13 �0.43 �0.59 fli 149 0.17 1.00 0.67 0.19 �0.31 �1.11 flb 149 0.00 1.00 0.74 0.16 �0.40 1.62 fle 149 0.11 1.00 0.73 0.24 �0.38 �0.16 flrm 149 0.00 1.00 0.72 0.26 �0.94 0.48 flrp 149 0.00 1.00 0.58 0.28 0.09 �0.17 bfie 149 2.01 4.75 3.58 0.55 �0.24 �0.24 bfia 149 1.75 5.00 3.98 0.60 �0.51 0.69 bfic 149 1.95 5.00 3.43 0.58 �0.27 0.17 bfin 149 1.00 4.45 2.89 0.61 �0.56 0.26 bfio 149 2.35 4.90 3.61 0.51 0.17 �0.39 overall financial literacy (flo) is the percentage of correct scores from 28 questions on financial literacy. fli represents the financial literacy metric focused on investments. flb represents the financial literacy metric focused on budgeting. fle represents the financial literacy metric focused on economics. frrm represents the financial literacy metric focused on risk management. flrp represents the financial literacy metric focused on retirement planning. bfie represents the level of extroversion an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (introvert) to 5.0 (extrovert). bfia represents the level of agreeableness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (not agreeable) to 5.0 (very agreeable). bfic represents the level of conscientiousness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (not conscientious) to 5.0 (very conscientious). bfin represents the level of neuroticism an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (low neuroticism) to 5.0 (high neuroticism). bfio represents the level of openness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (closed) to 5.0 (very open). 151r.n. killins / financial services review 26 (2017) 143–165 table 2 correlation matrix flo fli flb fle flrm flrp bfie bfia bfic bfin bfio flo 1.00 fli 0.73 1.00 flb 0.59 0.42 1.00 fle 0.77 0.38 0.36 1.00 flrm 0.73 0.35 0.38 0.53 1.00 flrp 0.79 0.56 0.27 0.59 0.44 1.00 bfie �0.06 �0.09 �0.19 0.02 0.14 �0.13 1.00 bfia 0.09 0.14 0.15 �0.03 0.18 �0.10 0.15 1.00 bfic 0.28 0.22 0.32 0.25 0.40 0.17 0.43 0.37 1.00 bfin �0.09 �0.14 0.08 �0.04 �0.09 �0.09 �0.32 �0.16 �0.35 1.00 bfio 0.14 0.13 0.21 0.03 �0.04 �0.04 0.35 0.18 0.33 �0.25 1.00 overall financial literacy (flo) is the percentage of correct scores from 28 questions on financial literacy. fli represents the financial literacy metric focused on investments. flb represents the financial literacy metric focused on budgeting. fle represents the financial literacy metric focused on economics. frrm represents the financial literacy metric focused on risk management. flrp represents the financial literacy metric focused on retirement planning. bfie represents the level of extroversion an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (introvert) to 5.0 (extrovert). bfia represents the level of agreeableness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (not agreeable) to 5.0 (very agreeable). bfic represents the level of conscientiousness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (not conscientious) to 5.0 (very conscientious). bfin represents the level of neuroticism an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (low neuroticism) to 5.0 (high neuroticism). bfio represents the level of openness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (closed) to 5.0 (very open). table 3 regression results (base model) variable flo fli flb fle flrm flrp constant 0.627c (0.051) 0.519c (0.061) 0.682c (0.054) 0.666c (0.074) 0.802c (0.082) 0.522c (0.081) gen �0.019 (0.043) 0.019 (0.051) 0.020 (0.047) �0.026 (0.064) �0.121a (0.071) �0.054 (0.071) race 0.129c (0.044) 0.167c (0.053) 0.085a (0.048) 0.144b (0.66) 0.149b (0.073) 0.143b (0.072) emp 0.029 (0.042) 0.059 (0.051) 0.002 (0.046) 0.065 (0.063) 0.040 (0.071) 0.041 (0.069) res �0.008 (0.044) 0.089 (0.053) �0.015 (0.048) �0.034 (0.066) �0.087 (0.073) �0.054 (0.072) r2 0.148 0.152 0.091 0.091 0.131 0.086 f-statistic 3.19c 3.24c 2.721b 2.654b 2.983b 2.457a the dependent variables include: overall financial literacy (flo) is the percentage of correct scores from 28 questions on financial literacy. fli represents the financial literacy metric focused on investments. flb represents the financial literacy metric focused on budgeting. fle represents the financial literacy metric focused on economics. frrm represents the financial literacy metric focused on risk management. flrp represents the financial literacy metric focused on retirement planning. the independent variables include: gender (gen) equal to 1 if the participant is male and zero if female. race (race) equal to 1 if the participant identifies as caucasian and zero if otherwise. employment status (emp) equal to one if the participant has been in the workforce force for more than 2 years and zero if not. residence in canada (res) equal to one if the participant has held residence in canada for more 5 years and 0 if otherwise. aindicates significance at the 10% level, bat the 5% level, cat the 1% level. 152 r.n. killins / financial services review 26 (2017) 143–165 imately 13% lower scores in the overall financial literacy survey. this result holds across all five subsections (investments, budgeting, economics, risk management, and retirement planning) of the financial literacy survey and is significant at all conventional statistical significance levels. unlike previous studies by lusardi and mitchell (2011a) and van rooij, lusardi, and alessie (2011) who find that women exhibit lower scores in financial literacy than men, this study does not find the gender variable significant in determining financial literacy levels. in fact, when measuring financial literacy surrounding risk management, males tend to exhibit lower financial literacy than females. finally, employment status or experience and residency length do not seem to influence the levels of the participants in this survey. table 4 provides the results from regression (2) which incorporates the control variables above along with the big five personality traits (extraversion, agreeableness, conscientiousness, neuroticism, and openness). the extraversion and conscientiousness variables show to be statistically significant on the various measures of financial literacy but have differing effects. the extraversion coefficient ranges from �0.120 to �0.023 across the various measures of financial literacy suggesting that individuals that exhibit more extroversive characteristics have lower levels financial literacy than those who may be more of an introvert. the conscientiousness variable ranges from 0.164 to 0.021, indicating that individuals who score higher on conscientiousness tend to have a greater degree of financial literacy. conscientiousness tends to impact risk management and economics literacy the most but has little impact on budgeting. of the other personality traits table 4 regression results (full model) variable flo fli flb fle flrm flrp constant 0.386 (0.252) 0.337 (0.311) 0.188 (0.277) 0.363 (0.398) �0.068 (0.408) 0.992b (0.434) gen �0.007 (0.042) 0.023 (0.052) 0.029 (0.046) 0.005 (0.067) �0.082 (0.068) �0.059 (0.073) race 0.134c (0.041) 0.166c (0.051) 0.079a (0.045) 0.164b (0.065) 0.162b (0.067) 0.153b (0.071) emp 0.010 (0.040) 0.043 (0.050) 0.015 (0.044) 0.031 (0.064) 0.018 (0.065) �0.006 (0.069) res 0.001 (0.042) 0.096a (0.051) �0.030 (0.045) �0.001 (0.065) �0.069 (0.066) �0.028 (0.071) bfie �0.071b (0.034) �0.089b (0.042) �0.090b (0.037) �0.034 (0.054) �0.023 (0.055) �0.120b (0.059) bfia �0.005 (0.034) 0.018 (0.042) 0.055 (0.037) �0.048 (0.054) 0.026 (0.056) �0.092 (0.059) bfic 0.103c (0.035) 0.095b (0.042) 0.021 (0.038) 0.149c (0.055) 0.164c (0.056) 0.147b (0.061) bfin 0.007 (0.032) 0.040 (0.039) 0.047 (0.035) 0.037 (0.051) 0.029 (0.052) �0.038 (0.056) bfio 0.033 (0.037) 0.036 (0.046) 0.106b (0.041) �0.008 (0.059) 0.046 (0.060) �0.032 (0.064) r2 0.276 0.268 0.252 0.197 0.271 0.196 f-statistic 4.156c 3.456c 3.032c 2.602b 3.756c 2.568b the dependent variable(s) remain as in table 3. the control variables gen, race, emp, and res remain as described in table 3. bfie represents the level of extroversion an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (introvert) to 5.0 (extrovert). bfia represents the level of agreeableness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (not agreeable) to 5.0 (very agreeable). bfic represents the level of conscientiousness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (not conscientious) to 5.0 (very conscientious). bfin represents the level of neuroticism an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (low neuroticism) to 5.0 (high neuroticism). bfio represents the level of openness an individual exhibits as measured by the big five personality survey with scores ranging from 0.1 (closed) to 5.0 (very open). aindicates significance at the 10% level, bat the 5% level, cat the 1% level. 153r.n. killins / financial services review 26 (2017) 143–165 measured by the big five, agreeableness does not seem to affect financial literacy in any definitive manner. neuroticism tends to have a positive impact (although not significant) across most financial literacy measures expect retirement planning. finally, openness tends to have a mixed impact on financial literacy levels but does have a positive and significant impact budgeting literacy. before moving on to the discussion and implications, the assumptions of the regression analysis seem to be met. the first assumption, linearity, was assessed through an analysis of residuals and these results do not exhibit any nonlinear patterns. tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (vif scores ranging from 1.10 to 1.62. finally, the data suggests that no consistent pattern was found of the residuals and normality of the error terms, thus fulfilling the independence and normality assumptions. 6. discussion and implications the findings of these study lead to several implications for the consumer, financial institutions, and regulators of the financial sector. the primary results that focus on behavioral traits of individuals add new evidence to the financial literacy literature that outlines what groups of people may be most at risk when it comes to financial decision making. the results indicate that extroverts exhibit less financial literacy than introverts, which should lead financial services practitioners to be more cautious in their dealing with these types of individuals and ensure that additional resources are provided, so they are aware of the financial products and services they are receiving. second, for firms that are seeking to employ individuals that provide financial products and services to consumers, extroverts are sometimes preferred over introverts, as interaction with people is a necessary step in providing financial services. popular human resource questionnaires that screen for behavioral traits that translate into sales roles may conflict with characteristics that ensure cognitive abilities in financial markets. thus, regulators and firms should seek a proper balance when determining which traits to value during employment assessments. as the industry moves towards technology-based platforms where traditional human interaction is becoming less important, the financial sector may benefit from introverts being more employable and strengthening the financial sector’s human capital by providing training models via alternative mediums such as social media platforms. further, conscientiousness is an important trait for financial service employees to have and this adds further evidence that employers should value this quality when making hiring decisions in the financial sector as these individuals exhibit a better understanding of the overall financial landscape and again can add more stability to the overall financial system. the lack of significance of the gender dummy provides differing results from previous studies such as, chen and volpe (2002); lusardi and mitchell (2011a); van rooij et al. (2011), among others who find that women seem to exhibit lower scores in financial literacy than men. this sample of individuals drawn from the canadian population suggests that canadian women have similar financial knowledge when compared with their male counterparts. this may be because of the increased gender neutrality that canada has been 154 r.n. killins / financial services review 26 (2017) 143–165 pushing for in the past decade in the overall workforce. as drolet (2016) outlines, women now play a greater role in the purchase of items such as houses, automobiles, insurance, and financial services. women also face different financial challenges than men. canadian women can expect to live about 4.5 years longer than men and, therefore, must finance a longer period of retirement. canadian women have higher disability rates than men and may incur costly long-term care needs as they age. these findings showcase why financial literacy is of great importance to females and that issues such as insurance and estate planning should be emphasized in future research. the result that indicates that non-whites experience lower financial literacy rates confirms previous findings such as lusardi et al. (2010) and lusardi and mitchell (2014) who find minorities have significantly less financial literacy when compared to whites. the findings from the previous research along with this article should provide financial literacy groups in canada such as the financial consumer agency of canada (fcac) to fund projects that are focused on developing the financial literacy of minority groups in canada. after controlling for length of residency in canada, it seems that new immigrants have transitioned well into the canadian financial system. this becomes increasingly important as new immigrants to canada have reached all-time highs in the past decade and the transition into a new banking system and economy can be difficult for them and could provide risks to the overall financial sector. further research focused on new immigrants (potentially on various age demographics) is warranted to ensure new financial literacy programs are directed at the individuals most at need. 7. conclusion it is important to push forward the financial literacy of canadians in many regards. lusardi and mitchell (2007, 2011a, 2011b) provided multiperiod life cycle model and evidence that shows that it is socially optimal to raise financial knowledge of everyone early in life. former federal reserve board chair, ben bernanke stated in a speech in 2013 that, “among the lessons of the financial crisis is the need for virtually everyone both young and oldto acquire a basic knowledge of finance and economics. such knowledge is necessary for anyone who will be faced with managing a household budget, making financial investments, finding reliable information about buying a car or a house and preparing financially for retirement and other life goals” (bernanke, 2013). as we push forward with financial education programs to foster the financial literacy of households and individuals it is important to understand where educators and regulators focus their attention. this research concentrates on the generation y age demographic, which are some of the early adopters of the financial technology (fintech) products and services. while focusing on this demographic, this research measures the personality traits of individuals and shines new light on how behavioral traits may influence one’s ability to understand the complexities of the financial decisions people face. the results of this research indicate that some personality traits do in fact play a role in financial literacy of an individual. the most significant personality trait shown to impact financial literacy is conscientiousness. individuals that have a higher degree of conscientiousness tend to exhibit higher levels of overall financial literacy. these results hold across 155r.n. killins / financial services review 26 (2017) 143–165 the various measures of financial literacy which include: investments, budgeting, economics, risk management, and retirement planning. another personality trait that has an influence on one’s overall financial literacy is extraversion. those who tend to be more extroverted tend to score lower in their overall financial literacy. although the primary research goal of this article was to uncover the role of personality traits on financial literacy this article also confirmed that minorities tend to exhibit a lower level of financial literacy, which is supported by previous work by lusardi and mitchell (2014), among others. in contrast to the previous finding by chen and volpe (2002) and lusardi and mitchell (2011a), the gender dummy variable did not show any statistical significance suggesting that males and females in the generation y age cohort share similar levels of financial literacy. the results of this research add to the growing literature in both the financial literacy and behavioral finance fields. policymakers and regulators can use this research to understand how to develop their financial literacy initiatives. financial institutions in canada who want to strengthen their customer base to provide additional products and services can also benefit from understanding the gaps that exist in financial knowledge of the generation y age cohort. this age group may need further assistance in developing their financial literacy levels, and financial institutions and regulators can use social media outlets to engage the generation y cohort and build a more financially savvy base of young clients. finally, financial institutions should be aware of the “paradox of choice.” as schwartz (2004) suggests, the fact that some choice is good doesn’t necessarily mean that more choice is better. understanding the limits of their customer base should lead banks and financial services providers to restrict the number of choices consumers need to make in complex financial planning areas. further research is imperative in canada and around the globe, to further the understanding of how financial literacy affects the various stakeholders in the financial markets. as the baby boomer transition into retirement, an estimated $750 billion dollars in wealth will be transferred to a younger generation in canada. this leads to significant implications for both young and old. the older generation will need to structure the transfer of this wealth properly, and the younger generation will have abundant amounts of capital that will need to be managed and preserved for future generations. this research acknowledges that estate planning literacy (outside risk management) was not addressed and future research is needed to better understand how wealth transfers in the coming decades can be managed effectively. collins (2012) has shown that financial literacy impacts one’s motivation to seek professional financial advice which should provide motivation for industry professionals to promote financial literacy as a complement rather than a substitute for their services. further, industry professionals along with regulators need to modernize their communication mediums to include the various social media platforms which will allow them to capture the various personality types of the generation y age cohort. notes 1 see lusardi and mitchell (2011b) for a detailed list of literature outside of the united states. 156 r.n. killins / financial services review 26 (2017) 143–165 2 for a detailed survey of the literature on these see john, naumann, and soto (2008). 3 although widely accepted in the psychology literature, alternative approaches to the big five approach have been put forward including approaches with fewer than five factors and approaches with more than five factors. see almlund, duckworth, heckman, and kautz (2011) for a discussion of the alternatives to and criticisms of the big five approach that have been put forward in the psychology literature. 4 see john, donahue, and kentle (1991) and john et al. (2008) for detailed review of the survey used and analysis of personality factors. 5 cfsc survey see www.23.statcan.gc.ca/imdb/p2sv.pl?function�getsurvey&sdds� 5159 6 the desired ratio of observations to independent variables is 15:1. when this level is reached, the results should be generalizable if the is representative. this research acknowledges that the sample is representative of the generation y demographic and the results should not be applied to other age groups. appendix financial literacy survey thank you for participating in our survey. this survey is intended to measure college students’ knowledge of personal finance. the results will be used to help students improve their knowledge and colleges improve curriculums. please answer all the questions the best of your ability. again, thank you for participating. investments 1. which of the following investment types carries the most risk? a. equities (stocks) b. short-term government bonds c. grade aa corporate bonds d. government insured certificates (gics) e. balanced canadian mutual fund 2. which of the following investment types would offer the highest expected return? a. grade aa corporate bonds b. short-term government bonds c. equities (stocks) d. balanced canadian mutual fund e. government insured certificates (gics) 157r.n. killins / financial services review 26 (2017) 143–165 3. what do you think the average yearly return is on the overall stock market (toronto stock exchange)? a. 30% b. 20% c. 2% d. 7% e. -5% 4. which of the following is false? a. as shareholders of a mutual fund, you have a right to tell fund managers what securities to buy. b. a mutual fund is a diversified collection of securities used as an investment vehicle. c. a mutual fund is an investment corporation that raises funds from investors and purchases securities. d. your ownership in a mutual fund is proportional to the number of shares you own in the fund. 5. a dividend is, a. distributions paid to bond-holders b. distributions paid to executives c. distributions paid to equity(stock) holders d. distributions paid from derivatives budgeting (savings and borrowing) 6. which of the following can hurt your credit rating? a. making late payments on loans and debts b. staying in one job too long c. living in the same location too long d. using your credit card frequently for purchases 7. what can affect the amount of interest that you would pay on a loan? a. your credit rating b. how much you borrow c. how long you take to repay the loan d. all of the above 8. which of the following will help lower the cost of a house? a. paying off the mortgage over a long period of time b. agreeing to pay the current rate of interest on the mortgage for as many years as possible c. making a larger down payment at the time of purchase d. making a smaller down payment at the time of purchase 9. your savings accounts in a federally insured commercial bank are insured by a. boc to the maximum amount of $10,000 per account. 158 r.n. killins / financial services review 26 (2017) 143–165 b. cdic to the maximum amount of $100,000. c. cdic to the maximum amount of $50,000 per account. d. boc to the maximum amount of $100,000. 10. if you invest $1,000 today at 4% for a year, your balance in a year will be a. higher if the interest is compounded daily rather than monthly. b. higher if the interest is compounded quarterly rather than weekly. c. higher if the interest is compounded yearly rather than quarterly. d. $1,040 no matter how the interest is computed. e. $1,000 no matter how the interest is computed. 11. which is false concerning credit cards? a. you can use your credit card to receive a cash advance. b. if your credit card balance is $1,000 and you pay $300, interest is charged on the unpaid balance of $700. c. the rate of interest on your credit card is normally higher than you can earn on a savings account. d. a credit card company will not charge you interest if you pay off the entire balance by the due date. economics 12. what is inflation? a. the rate at which the country’s exports grow b. the rate at of unemployment in the country c. the rate at which the average price level of goods and services changes d. the rate at which the stock market grows 13. how do economists classify a recession? a. two negative quarters of gdp growth b. unemployment reaches 10% c. unemployment and inflation both rise d. one (1) month of negative gdp growth 14. when the central bank of canada (the bank of canada) raises its interest rate how does this affect you? a. makes my investments more profitable b. makes my loans or mortgage more affordable c. makes my loans or mortgages more expensive d. increases inflation and make products and services more expensive 15. as the canadian dollar falls in value compared to u.s. dollar what impact does that have on domestic (canadian) prices? a. makes goods and services more expensive b. makes goods and services less costly c. has no direct effect 159r.n. killins / financial services review 26 (2017) 143–165 d. prices will go up and then go down 16. if the inflation rate was 6% and you made a 5% return on your investment in the past year you would have more purchasing power now than at the beginning of the year? a. true b. false 17. if interest rates rise, the price of a 10-year government bond will a. increase. b. decrease. c. remains the same. d. trade at a premium. e. be impossible to predict. risk management 18. if each of the following persons had the same amount of take home pay, who would need the greatest amount of life insurance? a. a young single woman with two young children b. a young single woman without children c. an elderly retired man, with a spouse who is also retired d. a young married man without children 19. the main reason to purchase insurance is to a. protect you from a loss recently incurred. b. provide you with excellent investment returns. c. protect you from sustaining a catastrophic loss. d. protect you from small incidental losses. 20. what type of insurance product is limited in length of coverage? a. whole-life insurance coverage b. term-insurance coverage c. duration-insurance coverage d. all insurance products are limited in length 21. auto insurance companies determine your premium based on a. age of insured. b. record of accidents. c. type and age of vehicle. d. all of the above. 22. life insurance products remain in force even though premiums have not been paid? a. true b. false 160 r.n. killins / financial services review 26 (2017) 143–165 retirement planning 23. high-risk and high-return investment strategy would be most suitable for a. an elderly retired couple living on a fixed income. b. a middle-aged couple needing funds for their children’s education in two years. c. a young married couple without children. d. all of the above because they all need high return. 24. upon retirement today what is the approximate maximum value that a person could receive from the canadian pension plan? (monthly payment) a. 505.00 b. 1065.00 c. 1920.00 d. 2510.00 25. your employer is obligated to provide you with retirement income if you worked for them for more than 25 years? a. true b. false 26. consider the following scenario: jack and jill are twins. at the age of 20, jack started contributing $20 a month to a savings account. after 20 years, at the age of 40, he stopped adding to his savings, but he left the money in the account. jill didn’t start to save until she was 40. then, she saved $20 a month until she retired 20 years later at age 60. suppose both jack and jill earned 6% interest per year on their savings. when they both retired at age 60, who had more money? a. jack b. jill c. they had the same amount d. don’t know 27. alice wants to invest $1,000 for retirement this year. her new employer will fully match her company pension contributions, up to $10,000 per year. all else being equal, which of the following options will give alice the highest total amount at the end of the year? a. alice contributes $1,000 to her company pension plan and invests that money in mutual fund at the end of the year with mutual fund a and has earned a 5% return. b. alice does not contribute to her company pension plan but she invests $1,000 in mutual fund x outside of her company pension plan. at the end of the year, mutual fund x has earned a 20% return. c. alice does not contribute to her company pension plan, but she invests $1,000 in mutual fund y outside of her company pension plan. at the end of the year, mutual fund y has earned a 5% return. d. don’t know. 161r.n. killins / financial services review 26 (2017) 143–165 general 28. what age category do you fall into? a. 15–19 b. 20–24 c. 25–29 d. 30� 29. indicate the highest level of formal education completed. a. high school b. some college c. 4-year bachelor degree d. master’s degree or better 30. in the above education were you enrolled in business-specific program? a. yes b. no 31. indicate which ethnic category you define yourself as, a. white b. african american c. hispanic d. asian e. middle eastern f. other 32. what is your gender/sex? a. male b. female 33. what is you marital status? a. single b. married c. separated/divorced 34. how long have you lived in canada? a. 0–2 years b. 3–5 years c. 6–8 years d. 8� years 35. indicate how many years of experience you have in the work-force. a. 0–2 b. 3–5 c. 6–8 d. 8� 36. what is your approximate gross income per year? 162 r.n. killins / financial services review 26 (2017) 143–165 a. less than $15,000 b. $15,001 to $30,000 c. $30,001 to $65,000 d. $65,000� 37. how many times in a month do you psychically go to you banking institution? a. 0–1 b. 1–3 c. 3–5 d. 5 or more 38. how many times in a month do you check your bank account online or via your banking app? a. 0–3 b. 4–7 c. 8–11 d. 11 or more 39. do you feel more comfortable interacting with someone about your finance in person or via a technology based platform? a. in person b. technology-based 40. how many credit cards do you own? a. 0 b. 1 c. 2 d. 3 e. 4 or more thank you!!! references almlund, m., duckworth, a. l., heckman, j. j., & kautz, t. d. (2011). personality psychology and economics (no. w16822). national bureau of economic research. bernheim, d. (1995). do households appreciate their financial vulnerabilities? an analysis of actions, perceptions, and public policy. tax policy and economic growth, 3, 11–13. bernheim, d. (1998). financial illiteracy, education, and retirement saving (no. 96-7). wharton school pension research council, university of pennsylvania. bernanke, b. s. (2013). financial and economic education: a speech at the 13th annual redefining investment strategy education (rise) forum, dayton, ohio (via prerecorded video), april 4, 2013 (no. 626). board of governors of the federal reserve system (us). boisclair, d., lusardi, a., & michaud, p. c. (2014). financial literacy and retirement planning in canada. journal of pension economics and finance, 16, 1–20. brandstatter, h. (1996). saving, income, and emotional climate of households related to personality structure. center for economic research, tilburg university, progress report no. 38. 163r.n. killins / financial services review 26 (2017) 143–165 brosdahl, d. j., & carpenter, j. m. (2011). shopping orientations of us males: a generational cohort comparison. journal of retailing and consumer services, 18, 548–554. buckland, j. (2010). are low-income canadians financially literate? placing financial literacy in the context of personal and structural constraints. adult education quarterly, 60, 357–376. chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. connor-smith, j. k., & flachsbart, c. (2007). relations between personality and coping: a meta-analysis. journal of personality and social psychology, 93, 1080–1107. cudmore, b. a., patton, j., ng, k., & mcclure, c. (2010). the millennials and money management. journal of management and marketing research, 4, 1–28. davis, k., & runyan, r. c. (2016). personality traits and financial satisfaction: investigation of a hierarchical approach. journal of financial counseling and planning, 27, 47–60. drolet, m. (2016). gender differences in the financial knowledge of canadians. statistics canada. available at http://publications.gc.ca/collections/collection_2016/statcan/75-006-x/75-006-20 fernandes, d., lynch jr., j. g., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors. management science, 60, 1861–1883. garcia, m. j. r. (2011). financial education and behavioral finance: new insights into the role of information in financial decisions. journal of economic surveys, 27, 297–315. goldberg, l. r., sweeney, d., merenda, p. f., & hughes, j. e. (1998). demographic variables and personality: the effects of gender, age, education, and ethnic/racial status on self-descriptions of personality attributes. personality and individual differences, 24, 393–403. harrison, n., & chudry, f. (2011). overactive, overwrought or overdrawn? the role of personality in undergraduate financial knowledge, decision-making and debt. journal of further and higher education, 35, 149–182. howe, n., & strauss, w. (2009). millennials rising: the next great generation. new york, ny: vintage. immordino-yang, m. h., christodoulou, j. a., & singh, v.(2012). rest is not idleness: implications of the brain’s default mode for human development and education, perspectives on psychological science, 7, 352–364. john, o. p., donahue, e. m., & kentle, r. l. (1991). the big five inventory–versions 4a and 54. berkeley, ca: university of california, berkeley, institute of personality and social research. john, o. p., naumann, l. p., & soto, c. j. (2008). paradigm shift to the integrative big five trait taxonomy: history, measurement, and conceptual issues. in o. p. john, r. w. robins, & l. a. pervin (eds.), handbook of personality: theory and research (pp. 114–158). new york, ny: guilford press. john, o. p., & srivastava, s. (1999). the big five trait taxonomy: history, measurement, and theoretical perspectives. handbook of personality: theory and research, 2, 102–138. kahneman, d., & tversky, a. (2000). choices, values, and frames. cambridge, uk: cambridge university press. lusardi, a. (2008). household saving behavior: the role of financial literacy, information, and financial education programs (no. w13824). national bureau of economic research. lusardi, a. (2015). financial literacy: do people know the abcs of finance? public understanding of science, 24, 260–271. lusardi, a., & mitchell, o. (2007). financial literacy and retirement preparedness: evidence and implications for financial education. business economics, 42, 35–44. lusardi, a., & mitchell, o. s. (2011a) financial literacy and retirement planning in the united states. journal of pension economics and finance, 10, 509–525. lusardi, a., & mitchell, o. s. (2011b). financial literacy around the world: an overview. journal of pension economics and finance, 10, 497–508. 164 r.n. killins / financial services review 26 (2017) 143–165 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and overindebtedness (no. w14808). national bureau of economic research. mandell, l. (2008). the financial literacy of young american adults: results of the 2008 national jump $ tart coalition survey of high school seniors and college students. washington, dc: the jump $ tart coalition for personal financial literacy. mandell, l., & klein, l. s. (2007). motivation and financial literacy. financial services review, 16, 105–116. mayfield, c., perdue, g., & wooten, k. (2008). investment management and personality type. financial services review, 17, 219–236. mccrae, r. r., & costa, p. t. (2010). neo inventories professional manual. odessa, fl: psychological assessment resources. ng, e. s., schweitzer, l., & lyons, s. t. (2010). new generation, great expectations: a field study of the millennial generation. journal of business and psychology, 25, 281–292. noftle, e. e., & robins, r. w. (2007). personality predictors of academic outcomes: big five correlates of gpa and sat scores. journal of personality and social psychology, 93, 116–130. robb, c. a., & woodyard, a. s. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning volume, 22, 60–70. saucier, g., & goldberg, l. r. (1996). the language of personality: lexical perspectives. the five-factor model. in j.s wiggins (ed.), the five-factor model of personality: theoretical perspectives (pp.21–50). new york, ny: guilford press. schwartz, b. (2004). the paradox of choice: why more is less. new york, ny: ecco. shiller, r. j.(2003). from efficient markets theory to behavioral finance. the journal of economic perspectives, 17, 83–104. shleifer, a. (2000). inefficient markets: an introduction to behavioral finance. new york, ny: oxford university press. smith, b. a., & stewart, f. (2009). learning from the experience of organization for economic co-operation and development countries: lessons for policy, programs, and evaluations. in overcoming the saving slump: how to increase the effectiveness of financial education and saving programs (pp. 345–367). chicago, il: university of chicago press. thaler, r. h., & de bondt, w. (1993). advances in behavioral finance. new york, ny: russell sage foundation. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. wilhelm, m. s., varcoe, k., & fridrich, a. h. (1993). financial satisfaction and assessment of financial progress: importance of money attitudes. financial counseling and planning, 4, 181–198. xiao, j. j., chen, c., & chen, f. (2014). consumer financial capability and financial satisfaction. social indicators research, 118, 415–432. xiao, j. j., chen, c., & sun, l. (2015). age differences in consumer financial capability. international journal of consumer studies, 39, 387–395. zick, c. d., mayer, r. n., & glaubitz, k. (2012). the kids are all right: generational differences in responses to the great recession. journal of financial counseling and planning, 23, 3–16. 165r.n. killins / financial services review 26 (2017) 143–165 anchoring, affect, and efficiency of sports gaming markets around playoff positioning kevin kriegera, r. daniel pacea,*, nicholas clarkeb, clay girdnerc adepartment of accounting and finance, university of west florida, college of business, 11000 university pkwy. pensacola, fl 32514, usa bdepartment of finance, college of business, florida state university, tallahassee, fl 32306, usa ckeybank, credit portfolio management, 127 public square, cleveland, oh 44114, usa abstract we consider the wagering market of national football league (nfl) and national basketball association (nba) games when participating teams have secured playoff positions. we use both the opening and closing lines (analogous to asset prices) of spread bets to examine if potential “letdown” effects, either psychologically or strategically, are priced. results demonstrate that the initial opening line consistently provides a profitable strategy for those betting against teams that have clinched positions in the post-season. by the close of the betting cycle, closing lines move in the expected direction as the market partially prices the letdown. many closing lines tighten to the extent that, after paying commissions, the naïve strategy of betting against clinched teams is less profitable. however, certain wagers, for example betting against nfl teams that have clinched top seeds, are statistically significantly economically profitable after paying commissions. these results support the behavioral finance concepts of anchoring, affect, in addition to lines moving towards efficiency. © 2015 academy of financial services. all rights reserved. keywords: behavioral finance; anchoring; affect; market efficiency; sports wagering 1. introduction certainly one of the most tested theories in finance is market efficiency. indeed the subject has such depth that major journals have published significant literature reviews by its first proponent (fama, 1970, 1991). proponents of inefficiency are also chronicled (hirshliefer, * corresponding author. tel.: �1-850-474-3201; fax: �1-850-473-7060. e-mail address: dpace@uwf.edu financial services review 24 (2015) 313–329 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. 2001), and an entire subfield, behavioral finance, has developed from the works of kahneman (kahneman and tversky, 1979), largely as a challenge to investor rationality. in relatively short order, researchers recognized that sports gaming markets were potentially ripe for considering direct market efficiency tests. in parallel to the traditional financial markets, the gaming-based research has streamed from identifying a potential statistical inefficiency, to expanding the research to including trading costs, and finally to providing possible explanations of the persistence of cataloged economical inefficiencies (and typically ignoring the joint hypotheses problem of the testability of market efficiency [fama, 1991]). the behavioral finance field also suggests the concept of “affect,” which is the desire to associate with perceived “good” firms and projects that may thereby influence financial judgment (macgregor, slovic, dreman, and berry, 2000). a popular anecdotal belief surrounding professional sports franchises is that once teams have secured a post-season position, they are apt to relax and lose focus, either unintentionally or as a strategic maneuver to refresh themselves for post-season tournament play.1 this “letdown effect” can be seen as a lack of motivation on the part of players who know there is nothing significant left to play for, a strategic maneuver on the part of coaches to give their players maximum rest before the start of post-season play, or both. in this article we consider whether the market makers of american professional sports betting markets, “bookmakers,” incorporate this letdown effect into the wagering lines of regular season contests, near the conclusion of the season, when exactly one team in the game has limited motivation. more important, is this limited motivation and strategic resting of players priced? early works in more traditional financial settings indicate that after reaching a goal, some degree of letdown, for either strategic or emotional reasons, may be commonplace (chevalier and ellison, 1997). we examine the most common wager: a spread bet.2 our questions are does the spread price this letdown as an efficient market suggests? is this a tradable position? what are possible explanations of the existence of these tradable positions? there is some controversy about the spread and what the purpose of the spread is to the bookmaker. is the spread similar to a local floor trader who wants, at the end of the day, a matched book (equal long and short positions)? or is the bookmaker taking an active position? the research is mixed with proponents of both the matched book (avery and chevalier, 1999) and positioning hypotheses (humphreys et al., 2013). fortunately, the structure of our study does not need to resolve this question as we examine both the opening and closing spreads of games to seek evidence of efficiency. prior research in gaming markets also gives some indication that irrational bettors may be willing to pay too high of prices (by accepting inflated spreads for wagers) to support “favorites” (see, e.g., vergin and scriabin, 1978, and paul and weinbach, 2005a).3 as most teams that have clinched playoff positions late in seasons would be favored against opponents who have not, they might draw a particularly disproportionate amount of wagers from naïve bettors, unaware of inflated spreads. there is evidence that as point spreads increase in the national football league (nfl), the number of wagers placed on the favored teams increases (paul and weinbach, 2011). some bookmakers may even intentionally set spreads off from the value they believe would result in a 50/50 split of wagered funds to increase profitability (levitt, 2004). 314 k. krieger / financial services review 24 (2015) 313–329 to help consider these possibilities further, we look separately at opening and closing lines of games with exactly one team clinched into playoff position. if an opening line creates an opportunity for systematic profitability, it should be corrected by the closing line, and the final, market clearing price (spread) should not afford systematic profitability, if the market is efficient. this may result because of the unveiling of information between the release of the opening line and the posting of a closing line, or because of order flow from bettors convincing bookmakers to update their prices (krieger and fodor, 2013). we examine the opening and closing lines for both statistical and economic efficiency of wagers when the positioning of future playoffs is fixed (in the parlance of the sports market a “clinched” position). gray and gray (1997) note that nfl wagering inefficiencies may dissipate over time, though in their results, the suggested dissipation is in terms of years rather than the opening to closing betting cycle. in fact, should profitable results be detected, the most interesting question to many observers might be whether such effects might persist into the future, especially with the popular publication of such information. in our study, we find that betting lines are set systematically too aggressively, in favor of teams that have clinched post-season berths, for the remainder of the regular season. we detect this using both a traditional statistical test evaluating the prices of point spreads in games of professional football and basketball, as well as by the profit levels recognized by wagering on such clinched teams. we also detect differences in wagering effectiveness between the opening line and the closing line. opening lines are consistently biased towards teams that have clinched playoff positions, often providing a potentially profitable bet, even after considering commissions. in nearly every case, however, the markets at least partially correct, providing support that gaming markets correct over time (see gray and gray, 1997 and krieger and fodor, 2013). in some cases, closing lines correct to the point that betting against clinched playoff teams no longer produces a viable betting position. this is evidence of the market becoming more efficient during the betting cycle. we also find some subsets (e.g., nfl teams locking top seeds or national basketball association [nba] teams locking specific seeds) where evidence of inefficiency of the betting market, because of letdown bias, persists at the closing line. these results further support “anchoring” theory, which suggests that markets are slow to update beliefs tied to inefficient prices (kahneman and tversky, 1979; kahneman, 1992; beggs and graddy 2009). 2. behavioral finance implications attempting to better understand why such opportunities persist motivates a turn to behavioral finance. anchoring (see kahneman and tversky, 1979) is the theory that the first price offered, in our case the opening spread, unduly influences the market clearing price. in another unorthodox market, beggs and graddy (2009) find support for anchoring in the case of art auctions. anchoring effects continue to develop as a source of current and relevant research. in a trio of recent studies, the 52-week highs for stocks and indices are anchors (there are many confirming empirical studies of the 52-week high referencing). george and hwang (2004) find that the current stock price coupled with the 52-week high price creates an anchor as information incorporates into the stock price, but traders are reluctant to push 315k. krieger / financial services review 24 (2015) 313–329 prices far away for the 52-week high. peng and xiong (2006) provide a theoretical structure and show that limited investor attention leads to investors processing more market and sector information than firm-specific information. li and yu (2012) find two anchors are important in explaining the price movements, both the 52-week high and the historical high. anchoring also occurs within industry earnings. cen, hillary, and wei (2013) find the industry median acts as an anchor in forecast earnings, as well as the difference between the forecast earnings and industry earnings. in the gaming literature, mcalvanah and moul (2013) examine the case of australian bookies and horseracing finding that when a horse “scratches” (i.e., is abruptly withdrawn after betting has started), the odds are not fully adjusted on the remaining horses. the adjustments recover only about 80% of the lost profit margin. we propose an anchoring-based theory as potentially explanatory of a new gaming-market result. “affect” is the theory that an individual’s motivations to associate with desirable capital projects, firms, and teams influence decisions that deviate from a pure risk-return rationale. macgregor, slovic, dreman, and berry (2000) demonstrate via experiment that the affective reaction, that is, the desire to associate with positive firms, influences financial judgment, a willingness to overvalue an asset. moreno, kida, and smith (2002) find that, in capital budgeting, affective reactions influence risk taking. the desire to associate with positive affectations can result in accepting projects that would be otherwise rejected. more recently, aspara and tikkanen (2011) find that the more positive an individual’s attitude towards a firm—affective self-affinity—can further the extra investment motivation. bernile and lyandres (2011) find, in the case of publicly traded european soccer clubs, that investors are overly optimistic about “their” team’s success. in sports gaming, early evidence suggests that bettor irrationality might generate from the desire to side with the favored team (vergin and scriabin, 1978; paul and weinbach, 2005a), which would be the team that has clinched a post-season milestone in the context of our study. the overall impression from our nfl analysis is that betting against nfl teams after clinching playoff berths or locking playoff seeds appears to be a lucrative strategy relative to the opening line. the opening point spreads set for nfl games systematically appear to underappreciate the letdown tendencies of teams that have clinched playoff positions, and thus, savvy bettors may profit. bookmakers themselves may have various perspectives on any inefficient lines they are setting. an aggressive bookmaker is aware of this inefficiency and allows it to exist (at least for now) because he or she is taking a calculated risk (in which case the informed, contrarian bettor is benefitting because of naïve public bettors) by, for example, making the line on a contest “new england-12” when the real handicapped line the bookmaker secretly believes to be fairest is “new england-11,” but naïve bettors (say, e.g., 70% of dollars wagered) will still bet new england even while laying 12 points (1 point “too many”). alternatively, the bookmaker could act more conservatively to achieve a balanced book. for example, the bookmaker may make the line to the above hypothetical game “new england-13” to achieve a more balanced volume of dollars wagered on the contest (there will be more dollars bet on new england’s opponent if it is getting 13 points [2 points “too many”] rather than 11 [the “fair” line in the bookmaker’s eyes] or 12 [the aggressive price designed to lure in additional wagers at poor odds]). in such a case, the opportunity for the contrarian bettor is even greater than in the aggressive bookmaker case. 316 k. krieger / financial services review 24 (2015) 313–329 we do not know the exact mix of aggressiveness or conservatism held by the sports books providing the lines used in our sample; furthermore, unlike a game of dice or playing cards, whose exact odds can be determined with simple probability calculations, there is no guarantee (especially for one specific game) that the bookmaker “correctly” estimates the “fair” line. regardless of bookmaker intention/philosophy the opportunity for bettors to profit from strategies based on a lack of awareness of a potential letdown effect would be present. we suspect that any notable results we discover will be based on a mix of bookmaker aggressiveness and conservatism. our results demonstrate that the sports gaming markets may initially be inefficient as the letdown effect is not priced correctly. the market does partly correct, and by the closing line, most historically available betting strategies are no longer as statistically powerful as at opening; however, some cases of statistical significance remain, even at closing lines. this may be particularly indicative of anchoring (kahneman and tversky, 1979; kahneman, 1992; beggs and graddy, 2009). contributing to this inefficiency may be the over optimism or desire to bet with favorites (see bernile and lyandres, 2011; vergin and scriabin, 1978; paul and weinbach, 2005a). even in the many cases lacking statistical significance, historical results are generally very supportive of the opportunity to successfully wager against clinched playoff teams for profit. it is possible that, in the future, such opportunities for profit may continue because of affect and anchoring effects. it is also possible that, in the future, greater attention drawn to such a successful strategy may cause more future bettors to wager against clinching teams; thus, shifting point spreads to more “correct” levels. the remainder of the article proceeds as follows: section 3 describes the data. section 4 describes our method and results. section 5 briefly concludes. 3. data our primary source of historical point spread data (opening and closing lines) for nfl and nba analysis and game score data are sportsinsights.com, which began collecting data in the middle of 2003. sportsinsights.com provided wagering data through the middle of 2012. thus, our initial focuses herein are the 2003–2012 nfl and nba seasons. nfl historical point spread and game score data for the 2012 regular season is taken from sunshine forecasts4 to complete the initial data set. opening lines are those set initially when bookmakers offer a “price” on the game to the betting public.5 betting action may move these prices up until the start of contests. this is particularly true if books are attempting to balance wagering dollars on each side of a contest. evidence of this desire is mixed (e.g., see levitt (2004) and paul and weinbach (2011)). additionally, information may develop regarding teams’ strategic and/or management intentions for upcoming games after opening lines are issued. for example, teams may not make the decision to rest important players until after the opening line of a game has been made. in such a case, a market might be perfectly efficient to introduce one opening line and change this line as information develops; thus, we also track the performance of wagering relative to the closing lines of contests, which are those in place immediately preceding the start of a game. evidence of market inefficiency 317k. krieger / financial services review 24 (2015) 313–329 relative to closing lines of contests is particularly important in demonstrating evidence of our hypotheses. we consider whether teams underperform gambling markets’ expectations after reaching benchmarks determining post-season positioning. we begin by considering whether teams that have clinched any position in the upcoming post-season tournament underperform. however, many of these teams still have considerable incentive to compete vigorously in remaining regular season games because playoff seeding is at stake.6 thus, we further narrow our sample to consider only those teams that are “locked” into specific playoff positions (i.e., regardless of all future regular season results, a team’s playoff seed can neither improve nor decline). to determine the historical dates when professional sports teams clinched playoff berths, locked in playoff positions, or more specifically clinched top seeds in their respective playoff tournaments, we conduct web searches for news stories regarding the teams that participated in each season’s playoffs. we hand collect the dates when playoff berths were secured or seeds were locked, taking care to distinguish between the actual date of the contest and the publishing date of news stories. the dates of some teams’ clinching of playoff berths or locking of positions could not be readily obtained from web searches. archived web news stories regarding clinching dates are relatively accessible in recent years but are more difficult to acquire for earlier seasons. to avoid using only a handful of contests in seasons from further back in our data set, we instead elect to utilize some date cutoffs for different portions of our analysis.7 for our purposes, clinching dates could be determined for all nfl playoff berths from 2003 to 2012 and all locked playoff seeds could be determined from 2004 to 2012; however, nba playoff berth clinching dates could only be regularly ascertained for the 2007–2012 period, though locked dates of playoff seeds for nba playoff tournaments could be determined for the 2004–2012 period. 4. method and results we analyze clinching teams’ spreads in the regular season contests after their clinching date(s) of accomplishments. we analyze the spreads of the next game, the next three games, and, if possible, all remaining games after a clinching performance.8 we question, specifically, if the letdown effect is priced into spreads, and if not, if those spreads afford a profitable wagering opportunity.9 to avoid confounding concerns, we include only games when the clinching team competes against an opponent that has not clinched its own playoff berth.10 we begin by considering the efficiency of point spreads in nfl and nba contests, after a team has clinched a playoff berth, locked in a particular playoff seed, or more particularly a top playoff seed, using the approach of zuber et al. (1985). this test requires the estimation of the simple regression model: actualpointdiffi � a � b�openinglinei� � ei (1) for the consideration of performance relative to the opening line of a contest, and 318 k. krieger / financial services review 24 (2015) 313–329 actualpointdiffi � a � b�closinglinei� � ei (2) for the consideration of performance relative to the closing line of a contest. actualpointdiffi, the dependent variable, is the opponent’s final score in game i, minus the clinched team’s final score in the game. the independent variables of the regressions, openinglinei and closinglinei, are the opening and closing bookmaker lines of game i as reported by sportsinsights.com or sunshine forecasts. to conduct the test, we estimate the unrestricted version of eqs. (1) and (2) and note the residual sum of squares of the regression. we use this information, along with the residual sum of squares when the restrictions a � 0 and b � 1 are imposed, to calculate f test statistics for each specification. a significant f-statistic denotes evidence against the efficiency of the opening or closing line of games as accurate predictors of game results. our results for the nfl and nba are shown in table 1. we first note that, in panel a, which addresses line efficiency for nba and nfl teams after clinching a playoff berth, no evidence of point spread statistical inefficiency is detected via the zuber et al. (1985) tests. we do, however, find evidence of line inefficiency in panels b and c, which provide the line efficiencies for nba and nfl teams after locking a specific playoff seed and clinching the top overall playoff seed, respectively. in panel b, in the opening line case, we find evidence of inefficiency of lines set for nba teams that have locked playoff positions. nba teams that are set in specific playoff positions have inefficient lines, as detected from f-tests, for the next game (next three games, all remaining games) at the 5% (1%, 5%) level. when we consider the closing lines, the significance of these nba results declines (also in panel b) to the 10% (5%, 10%) level for the next game (next three games, all remaining games) sample. this supports the conjecture that the letdown effect is not priced initially in the opening line, but by the time closing lines are established, results are not as strong, indicating that the letdown effect is priced somewhat more fully by the closing line than the opening line. in panel c, we note that when concentrating our focus on the subset of teams that have clinched the top seed in their upcoming playoff tournament, some significant results emerge for the nfl, as well as the nba. for opening lines, the small sample of 19 nfl games with locked top seeds shows statistical evidence of inefficient spreads at the 10% level, and nba lines are inefficient at the 5% level for the three-game sample after the locking of top playoff seeds for the opening line. when we expand to look at all nba contests after the locking of top seeds, opening and closing lines are inefficient at the 10% level. it is not surprising, given our hypothesis that the letdown effect is underappreciated, that results are stronger for the subsample of games where teams are locked into playoff position and, therefore, have nothing tangible to play for (panels b and c). conversely, many teams that have simply clinched playoff berths (panel a) are still looking to improve playoff seeding in remaining games. however, we also note that the smaller samples of panels b and c, providing lower statistical power, require a higher hurdle in the available samples to achieve significant results than would be necessary in panel a (where no evidence is seen).11 our initial indication, therefore, is that nba and nfl spreads do not fully appreciate and price the letdown effect for games after the locking in of playoff seeds. put more plainly, our conjecture is that with playoff seeds locked, betting markets do not fully appreciate the relative lack of effort that teams (or their managers) will put forth in subsequent “meaning319k. krieger / financial services review 24 (2015) 313–329 table 1 zuber et al. (1985) tests of line efficiency of nba and nfl games for teams that have clinched post-season benchmarks n a b f-stat panel a: line efficiency for teams after clinching playoff berth opening lines nfl teams, next game after clinching 70 0.32 0.98 0.77 nfl teams, all games after clinching 160 0.55 1.01 0.91 nba teams, next game after clinching 77 �0.68 0.89 0.59 nba teams, next three games after clinching 190 0.13 0.95 0.37 nba teams, all games after clinching 536 0.67 0.99 1.44 closing lines nfl teams, next game after clinching 70 0.06 0.77 0.56 nfl teams, all games after clinching 160 0.08 0.96 0.07 nba teams, next game after clinching 77 �1.09 0.73 0.88 nba teams, next three games after clinching 190 0.08 0.94 0.16 nba teams, all games after clinching 536 0.55 1.03 0.57 panel b: line efficiency for teams after locking specific playoff seed opening lines nfl teams, next game after locking seed 33 �0.76 0.88 0.06 nfl teams, all games after locking seed 41 1.75 0.86 0.32 nba teams, next game after locking seed 58 4.10*** 1.01 3.50** nba teams, next three games after locking seed 119 3.63*** 0.99 5.00*** nba teams, all games after locking seed 167 3.18*** 1.18 4.90** closing lines nfl teams, next game after locking seed 33 �1.93 0.92 0.33 nfl teams, all games after locking seed 41 0.54 0.92 0.09 nba teams, next game after locking seed 58 3.39** 1.07 2.84* nba teams, next three games after locking seed 119 2.83*** 0.95 3.57** nba teams, all games after locking seed 167 2.08** 1.12 2.53* panel c: line efficiency for teams after clinching top playoff seed opening lines nfl teams, next game after clinching conference 13 2.32* 0.93 2.49 nfl teams, all games after clinching conference 19 5.77*** 0.88 3.29* nba teams, next game after clinching conference 12 1.86 1.24 0.31 nba teams, next three games after clinching conference 35 3.31*** 1.05 3.99** nba teams, all games after clinching conference 63 2.99* 0.95 3.14* closing lines nfl teams, next game after clinching conference 13 1.71 0.91 1.55 nfl teams, all games after clinching conference 19 5.50*** 0.85 2.72* nba teams, next game after clinching conference 12 2.04 1.19 0.18 nba teams, next three games after clinching conference 35 3.01** 1.03 2.91* nba teams, all games after clinching conference 63 2.75* 0.77 2.47* table 1 presents estimates of intercepts and slopes for the regression model of actual point differential realized in nfl and nba games on the opening and closing spreads of these games (eqs. 1 and 2). the dependent variable of the simple ols regression, actualpointdiff, is the opponent team’s final score in the game, minus the clinched team’s final score in the game. the independent variables of the regression are closingline(openinglinei), which is the closing (opening) bookmaker line of game i as reported by sportsinsights or sunshine forecasts (http://www.repole.com/sun4cast/data.html). the closing (opening) line is reported relative to the team which has clinched a playoff berth, locked in a specific seed in the upcoming post-season, or more specifically the top seed. the sample is from the 2003-2012 nfl regular seasons for playoff berth clinching and 2004-2012 for locking of nfl playoff positioning (including top seeds). the sample is from the 2007-2012 nba regular seasons for playoff berth clinching and 2004-2012 for locking nba playoff positioning (including top seeds). panel a presents results for the performance of teams after clinching playoff berths. panel b presents results for the performance of teams after being “locked” into playoff position. panel c presents results for the performance of teams after clinching top seeds in playoff competition. games in which a team’s opponent has already clinched a playoff berth are omitted from the sample as our research question involves a potential relative lack of motivation for teams which have already clinched playoff-related goals. dates at which teams clinched playoff berths or top seeds or were locked into playoff positions are taken from internet searches of historical sports news stories. f-statistics are presented which test the joint hypothesis that a � 0 and b � 1, a test for efficiency of the betting line noted by zuber et al. (1985) and other authors. *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. 320 k. krieger / financial services review 24 (2015) 313–329 less” games. this lack of effort may manifest itself through the management’s resting of star players or the lack of advanced strategy or effort in execution put forth by the remaining players. it is perhaps not surprising that this result particularly appears in games after top seeds are achieved, for these are the cleanest examples of opportunities for teams to rest. by contrast, after the clinching of general playoff berths, many teams may still be motivated to put forth maximum effort to gain a higher playoff seed, which may explain the lack of statistical significance in the results of panel a.12 evaluating the efficiency of betting lines based on the zuber et al. (1985) approach is an interesting first step, but this ignores the most relevant and interesting questions for betting markets. namely, do betting lines create the anticipated dynamic in which bettors cannot overcome the commission charged on wagering by using a strategy? do sports gambling markets set efficient prices for wagers, or is it possible that the marketplace neglects the impact of the letdown effect of teams that have clinched playoff positions? these questions cannot be answered by simply considering the zuber et al. (1985) test. if, for example, teams that have just clinched playoff berths or locked seeds or top seeds are as likely to cover spreads as their opponents, but were disproportionately outscored, relative to spreads, in those games in which they failed to cover, then the rationality/ efficiency test of zuber et al. (1985) may detect a systematic error in the line. however, that line might actually be completely efficient for purposes of wagering. the reverse is also possible. we might find insignificant test statistics for line efficiency of clinching teams given the approach of zuber et al. (1985), but if this result is driven by covering relatively few spreads by large amounts and failing to cover relatively more spreads by small amounts, then a profitable opportunity for wagering, by simply betting against teams that have recently clinched playoff positions, may exist. thus, we analyze whether systematically betting against recently clinched teams, because of an under-appreciation of the letdown effect, may be a profitable endeavor. because of the typical 10% vigorish charge (the commission charged by the bookmakers on winning wagers), a bettor must actually win over 52.38% of equal wagers to make a profit;13 hence, the best way to consider the performance of a proposed betting strategy is relative to this rate. we omit contests that result in tied (or “pushed”) wagers from this portion of the analysis. tables 2 (nfl) and 3 (nba) provide results for the regular season performances of teams that have clinched a post-season benchmark. subsequently, for a strategy of betting against these teams to prove successful, the clinching team must fail to cover in a minimum of 52.38% of games (if a clinched team fails to cover, a bet against this team would result in a “win”). we present one-sided p-values for the results, relative to our hypothesis of underperformance by clinched teams.14 panel a provides results for teams that have clinched a playoff berth; panel b provides results for teams that are locked into any specific conference seed; panel c provides results for teams that are locked into the overall top conference seed. as in table 1, we consider results relative to the opening and closing spreads separately. table 2 presents our analysis of the retrospective performance of wagering against nfl teams that have clinched playoff positions. we find in panel a that wagering against nfl teams that clinched a playoff berth, at the opening line offered by bookmakers, has been a profitable strategy over the 2003–2012 seasons. when analyzing the opening line spreads, nfl teams fail to cover the spread at a 321k. krieger / financial services review 24 (2015) 313–329 table 2 betting on nfl teams that have clinched post-season benchmarks next game all remaining games panel a: teams that clinched a playoff berth opening line results n 70 160 covered 28 68 failed to cover 42 92 percentage filed to cover 60.00 57.50 one-sided p-value vs. 52.38% 0.101 0.097* closing line results n 70 160 covered 30 71 failed to cover 40 89 percentage failed to cover 57.14 55.63 one-sided p-value vs. 52.38% 0.213 0.205 panel b: teams that locked any seed in conference opening line results n 30 38 covered 13 14 failed to cover 17 24 percentage failed to cover 56.67 63.16 one-sided p-value vs. 52.38% 0.319 0.092* closing line results n 30 38 covered 15 16 failed to cover 15 22 percent failed to cover 50.00 57.89 one-sided p-value vs. 52.38% — 0.248 panel c: teams that locked top seed in conference opening line results n 13 19 covered 4 4 failed to cover 9 15 percent failed to cover 69.23 78.95 one-sided p-value vs. 52.38% 0.112 0.010** closing line results n 13 19 covered 5 5 failed to cover 8 14 percent failed to cover 61.54 73.68 one-sided p-value vs. 52.38% 0.254 0.032** table 2 presents the results for betting on nfl regular season games from 2003 through 2012 in which only one team had clinched a post-season benchmark. the opening and closing line spreads and the final scores for the games were gathered from sportsinsights or sunshine forecasts (http://www.repole.com/sun4cast/data.html). to determine whether a team with a clinched benchmark “covered” the bet, the spread was added to the locked teams score, and the opponent’s score was then subtracted. if the outcome was positive, the locked team covered the spread; if the outcome was negative, the locked team failed to cover the spread. games in which both teams had locked their respective playoff seeds were ignored, as were games that resulted in a tie (or “push”). all results are provided relative to both opening spreads and closing spreads. panel a provides the results for teams that clinched a playoff berth in their conference. results are subcategorized into next game and all remaining regular season games, which refer to the original clinching date. panel b provides the results for teams that locked the overall top seed in their conference, and panel c provides results for teams that locked any seed in their conference. the sample is from the 2003-2012 nfl regular seasons for playoff berth clinching (panel a) and 2004-2012 for locking of nfl playoff positioning (panels b and c). web searches were used to hand collect when teams clinched playoff berths and subsequently locked specific playoff seeds. a strategy must win in over 52.38% of occurrences in order to be profitable, given the “11-for-10” commission charged to sports bettors. p-values relative to this benchmark are reported. * and ** denote statistical significance at the 10% and 5% levels, respectively. 322 k. krieger / financial services review 24 (2015) 313–329 table 3 betting on nba teams that have clinched post-season benchmarks next game next 3 games all remaining games panel a: teams that clinched a playoff berth opening line results n 77 190 536 covered 36 92 258 failed to cover 41 98 278 percent failed to cover 53.25 51.58 51.87 one-sided p-value vs. 52.38% 0.457 — — closing line results n 77 190 536 covered 38 94 262 failed to cover 39 96 274 percent failed to cover 50.65 50.53 51.12 one-sided p-value vs. 52.38% — — — panel b: teams that locked any seed in conference opening line results n 57 117 165 covered 21 44 69 failed to cover 36 73 96 percent filed to cover 63.16 62.39 58.18 one-sided p-value vs. 52.38% 0.052* 0.015** 0.068* closing line results n 57 117 165 covered 21 47 73 failed to cover 36 70 92 percent failed to cover 63.16 59.83 55.76 one-sided p-value vs. 52.38% 0.052* 0.053* 0.192 panel c: teams that locked top seed in conference opening line results n 12 34 60 covered 5 13 26 failed to cover 7 21 34 percent failed to cover 58.33 61.76 56.67 one-sided p-value vs. 52.38% 0.340 0.137 0.253 closing line results n 12 34 60 covered 4 14 28 failed to fover 8 20 32 percent failed to cover 66.67 58.82 53.33 one-sided p-value vs. 52.38% 0.161 0.226 0.441 table 3 presents the results for betting on nba regular season games from 2004 through 2012 in which only one team had clinched a post-season benchmark. the opening and closing line spreads and the final scores for the games were gathered from sportsinsights or sunshine forecasts (http://www.repole.com/sun4cast/data.html). to determine whether a team with a clinched benchmark “covered” the bet, the spread was added to the locked teams score, and the opponent’s score was then subtracted. if the outcome was positive, the locked team covered the spread; if the outcome was negative, the locked team failed to cover the spread. games in which both teams had locked their respective playoff seeds were ignored, as were games that resulted in a tie (or “push”). all results are provided relative to both opening spreads and closing spreads. panel a provides the results for teams that clinched a playoff berth in their conference in the 2007-2012 seasons. results are subcategorized into next game, next 3 games, and all remaining regular season games, all of which refer to the original clinching date. panel b provides the results for teams that locked the overall top seed in their conference, and panel c provides results for teams that locked any seed in their conference. both panel b and c are from the 2004-2012 seasons. web searches were used to hand collect when teams clinched playoff berths and subsequently locked specific playoff seeds. a strategy must win in over 52.38% of occurrences in order to be profitable, given the “11-for-10” commission charged to sports bettors. one-sided p-values relative to this benchmark are reported. * and ** denote statistical significance at the 10% and 5% levels, respectively. 323k. krieger / financial services review 24 (2015) 313–329 60.00% rate (p-value � 0.101). in all remaining regular season games, teams with a clinched playoff berth fail to cover the opening line spread in 92 out of 160 games, a 57.50% rate (p-value � 0.097). these figures indicate a viable strategy is present in betting against nfl teams that have clinched a playoff berth, eclipsing the 52.38% required rate for success. we additionally consider results relative to closing lines in panel a. in the game after the clinching of a playoff berth, 40 of 70 nfl teams have failed to cover the closing line spread (57.14%). this outperforms the 52.38% mark necessary to offset the commission, but not at a significant rate, statistically (p-value � 0.213). betting against locked-seed nfl teams in all remaining regular season games would win 55.63% of wagers in our sample period, again above the profitability mark, but not statistically significantly better than the 52.38% mark (p-value � 0.205). a potential cause of the inferior results for the closing line sample is that some knowledgeable bettors might seize upon the potential letdown effect and thus push spreads away from clinched teams. if the letdown effect is not correctly priced in the opening line the spread, by the time of the closing line, may no longer afford a statistically significant profitable wager. we consider, in panel b of table 2, betting against teams that are locked into any playoff seed (thus, having no strategic motivation in the game for upcoming playoff positioning, unlike many teams in panel a). for the opening line case, results eclipse the 52.38% benchmark for the next game (17 of 30 locked teams, 56.67%, fail to cover) and for all remaining regular season games (24 of 38 locked teams, 63.16%, fail to cover). statistical evidence is weak for the next-game results (p-value � 0.319) but are marginally significant when considering all remaining games (p-value � 0.092). again, the performance of these teams is worse against opening line spreads than closing line spreads. out of the 30 games immediately after the game in which a seed was locked, teams fail to cover only half of the time. out of 38 total post-lock regular season games, the locked team fails to cover 22 times against the closing line, resulting in a p-value of only 0.248. in panel c of table 2, we see that wagering against nfl teams that have clinched the top seed in their conference has been a very profitable venture over the nine seasons from 2004 to 2012, especially for bets placed relative to the opening line spreads. of 13 games immediately after the clinching of a top seed, the locked team has failed to cover the opening spread nine times, a 69.23% rate (p-value � 0.112). of the 19 total regular season games played by nfl teams after clinching top conference seeds, the team has failed to cover 15 times relative to the opening line, a 78.95% rate (p-value � 0.010). these significance levels are impressive given the very limited sample sizes afforded by the nfl. performances of strategies betting against top seed locks, relative to closing line spreads, are similar, though significance declines given the small sample sizes at work. in the first game after locking the top seed, teams fail to cover 8 out of the 13 games, a 61.54% rate (p-value � 0.254). for all remaining regular season games, top seed locks fail to cover 14 out of 19 contests, a 73.68% rate (p-value 0.032). while all of these sample sizes are very small, the near-75% success rate of the strategy of betting against teams locked into top nfl seeds in all remaining games results in statistically significant winning rates above 52.38%. after considering the costs of placing wagers, we find this lone example of statistically significant inefficiency in nfl gaming markets, but we also note the many cases of historical 324 k. krieger / financial services review 24 (2015) 313–329 success of such wagering strategies that do not reach the high threshold of statistical significance. we present our results in table 3 using similar betting strategies, historically, for the nba. in panel a, we note that betting against nba teams after they clinch playoff berths has not proven profitable over the 2007–2012 seasons. relative to the opening line spreads, the success rate of such wagering in the next game is only 53.25%, barely eclipsing the 52.38% mark needed to demonstrate positive returns. further, the success rates of betting against teams that have clinched playoff berths in the next three games or all games after clinching exceed 50%, but they do not exceed the positive return benchmark of 52.38%. like in the nfl results of table 2, using closing line spreads results in slightly worse results. panel b, which provides the results of nba teams that have clinched any playoff seed, are strongly indicative of a lack of full pricing of the letdown effect. again, wagering against playoff teams at opening lines proves more favorable, with success rates of wagering against locked position teams in the next game 63.16% (significant at 10%) , next three games 62.39% (significant at 5%), and all remaining games 58.18% (significant at 10%). for closing spreads, these rates are 63.16% for the next game (significant at 10%), 59.83% for the next three games (significant at 10%), and 55.76% for all remaining games after clinching, indicating historical profitability. a bettor using a strategy of wagering against nba teams locked into a specific playoff seed would eclipse the 52.38% mark needed for success in all of the situations documented in panel b. in panel c, we see that wagering against nba teams after they have locked conference top seeds has indeed proven quite historically profitable (this over the 2004–2012 seasons). with small sample sizes of 12, 34, and 60 games, respectively, for the next game, next three games, and all games after the locking of top seeds, the results are not statistically significantly greater than 52.38%, for either opening or closing lines, but the success rates are still encouraging (like in panel b). as in the nfl wagering results of table 2, it appears that wagering against teams that have clinched playoff benchmarks, while they are concluding their regular seasons, is a profitable strategy in the nba, especially based on opening lines. small differences in the specific findings are, however, present. wagering against nfl teams after clinching any playoff berth (without necessarily being locked into a particular seed position) has been historically successful, while this has not been the case in the nba (panel a of tables 2 and 3). results for the nfl show the greatest statistical strength for the subsample of teams locked into the top seeds in remaining regular season games (table 2, panel c), while results for the nba show the greatest statistical strength for the sample of all teams locked into playoff position (table 3, panel b).15 5. conclusion we examine the efficiency of the nfl and nba sports gaming markets after teams secure post-season positions. teams clinching post-season berths are commonly overvalued initially in the opening lines offered by bookmakers, perhaps because of the lack of information regarding the status of certain players on these teams or uncertainty regarding such teams’ 325k. krieger / financial services review 24 (2015) 313–329 intentions for competing with maximum effort to win the remaining contests of the regular season. this effect is generally stronger for the subsample of professional teams that are locked into specific playoff positions (particularly in the nba), including the top seed in their upcoming post-season playoff tournament (particularly in the nfl). as these teams, by definition, have little true incentive to play at optimal levels (or may choose to strategically rest players), the betting public may be particularly slow to recognize this situation or unable to overcome their bias towards perceived superior teams, even when these teams do not have the same incentives they did in establishing their superior records. closing lines, analogous to the last market clearing prices, generally demonstrate some correction in the appropriate direction. in some cases, closing lines correct to the extent that a naïve betting rule does not overcome betting commissions. however, many cases persist in which closing lines persist that are statistically and economically profitable for bettors willing to wager against those teams that may experience a letdown effect. our results are consistent with markets correcting towards efficiency while also providing evidence of affect (the desire to associate with perceived winners) and anchoring theories. football and basketball are followed closely and even casual fans or gamblers can readily understand the emotional and strategic reasons for a letdown after achieving specific goals. despite this, even after trading costs, we find statistical evidence supporting behavior biases of affect and anchoring. notes 1 for some discussion from another market, consider: http://www.telegraph.co.uk/sport/ football/competitions/premier-league/10810560/do-premier-league-teams-ease-upwhen-they-have-nothing-to-play-for.html. 2 bookmakers set point spreads, or “lines,” in the most common form of handicapping games. the point spread issued by the bookmaker (often a casino or internet company) establishes the “favorite” and the “underdog” of a game. this point spread serves as a correction based on the perceived likelihood of each team winning a game. the favorite is considered more likely to win a game, and thus, the spread is instituted to place the two sides of a wager on more equivalent footing. a wager is graded based on subtracting the spread from the favorite’s final score and comparing this adjusted figure to the score of the underdog. whichever side then has the higher score is the winning team of the “against the spread” wager. the team that wins an against-thespread wager is said to have “covered” the game or the spread. 3 this is not to be confused with the “longshot bias” found in parimutuel betting (thaler and ziemba, 1988). woodland and woodland (1994) also note that, in the case of baseball, this bias is reversed but not to the point where this bias is a tradable strategy. 4 http://www.repole.com/sun4cast/data.html. 5 sportsinsights.com “opening” lines are based on overnight lines issued for nba games played the following afternoon/evening. nfl opening lines from sportsinsights. com are based on lines issued on sunday evenings for games to be played the following week (thursday through the following sunday) with the exception of teams about to 326 k. krieger / financial services review 24 (2015) 313–329 be involved in monday games the following day. for monday game participants, opening lines for thursday–sunday games are usually issued late monday night or tuesday morning. 6 typically, teams with higher playoff seeds draw supposedly weaker opposition in the opening round of post-season competition, and earn the advantage of playing the majority of an odd-numbered series of post-season games at their home venue. 7 the substance of the results shown herein remain nearly unchanged when the few contests from preceding years are included (e.g., a few nba playoff berth clinching dates are determined and these observations used as well). we merely cutoff the sample dates to limit the randomness of only including some observations from earlier years in certain categories of analysis. 8 we track performance in the next game, and all remaining regular season games for nfl contests with one clinched playoff participant (very rarely are positions determined more than three games before the 16-game nfl regular season concludes). we track performance in the next game, next three games, and all remaining regular season games for nba contests with one clinched playoff participant. 9 as mentioned previously, the letdown effect can include strategic maneuvering on the part of coaching staffs to rest players before the post-season. because these moves are sometimes delayed until after the initial spreads are set, some of the adjustments of opening lines towards closing lines is related to information realization of the betting markets. 10 thus, as noted before, the majority of these games feature “good” teams (those qualifying for post-season play) against those of perceived lower quality. hence, most games have the clinched teams favored. however, this is not always the case, particularly when news develops before an upcoming game that the clinched team might not use its best players in the contest. 11 such a framework persists in tables 2 and 3 as well. 12 in unpublished results, f-statistics are also insignificant for the full sample of nba and nfl teams after the clinching of playoff berths (beyond the next three games) and for the full sample of nba teams after clinching top seeds (beyond the next three games). 13 details of this calculation may be found in levitt (2004). 14 winning wager rates greater than 52.38% are indicative of historical profitability. statistical significance of rates being greater than 52.38% is a high threshold given the limited power available and the small sample. examples of significance above the 52.38% benchmark are quite limited in the literature. (fodor et al. 2013). 15 we note that in the nba, home court advantage in the nba finals, pitting winning teams from each conference’s post-season tournament, is granted to the team with the superior regular season record. thus, it may be possible that locked top conference seeds in the nba may still have tangible reason to compete in remaining regular season games. however, we investigate this hypothesis and find results no more significant than those of table 3, panel c. 327k. krieger / financial services review 24 (2015) 313–329 references aspara, j., & tikkanen, h. (2011). individuals’ affect-based motivations to invest in stocks: beyond expected financial returns and risks. journal of behavioral finance, 12, 78–89. avery, c., & chevalier, j. (1999). identifying investor sentiment from price paths: the case of football betting. journal of business, 72, 493–521. beggs, a., & graddy, k. (2009). anchoring effects: evidence from art auctions. american economic review, 99, 1027–1039. bernile, g., & lyandres, e. (2011). understanding investor sentiment: the case of soccer. financial management, 40, 357–380. cen, l., hilary, g., & wei, k. c. j. (2013). the role of anchoring bias in the equity market: evidence from analysts’ earnings forecasts and stock returns. journal of financial and quantitative analysis, 48, 47–76. chevalier, j., & ellison, g. (1997). risk taking by mutual funds as a response to incentives. journal of political economy, 105, 1167–1200. fama, e. (1970). efficient capital markets: a review of theory and empirical work. journal of finance, 25, 383–417. fama, e. (1991). efficient capital markets: ii. journal of finance, 46, 1575–1617. fodor, a., difilippo, m., krieger, k., & davis, j. (2013). inefficient pricing from holdover bias in the nfl point spread markets. applied financial economics, 27, 1407–1418. george, t., & hwang, c. y. (2004). 52-week high and momentum investing. journal of finance, 59, 2145–2176. gray, p., & gray, s. (1997). testing market efficiency: evidence from the nfl sports betting market. journal of finance, 52, 1725–1737. hirshleifer, d. (2001). investor psychology and asset pricing. journal of finance, 56, 1533–1597. humphreys, b., paul, r., & weinbach, a. (2013). bettor biases and the home-underdog bias in the nfl. international journal of sport finance, 8, 294–311. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–292. kahneman, d. (1992). reference points, anchors, norms and mixed feelings. organizational behavior and human decision processes, 51, 296–312. krieger, k., & fodor, a. (2013). price movements and the prevalence of informed traders: the case of line movements in college basketball. journal of economics and business, 68, 70–82. levitt, s. (2004). why are gambling markets organized so differently from financial markets? the economic journal, 114, 223–246. li, j., & yu, j. (2012). investor attention, psychological anchors, and stock return predictability. journal of financial economics, 104, 401–419. macgregor, d., slovic, p., dreman, d., & berry, m. (2000). imagery, affect, and financial judgment. journal of psychology and financial markets, 1, 104–111. mcalvanah, p., & moul, c. c. (2013). the house doesn’t always win: evidence of anchoring among australian bookies. journal of economic behavior & organization, 90, 87–99. moreno, k., kida, t., & smith, j. f. (2002). the impact of affective reactions on risky decision making in accounting contexts. journal of accounting research, 40, 1331–1349. paul, r., & weinbach, a. (2005a). bettor misperceptions in the nba: the overbetting of large favorites and the “hot hand.” journal of sports economics, 6, 390–400. paul, r., & weinbach, a. (2011). nfl bettor bias and price setting: further tests of the levitt hypothesis of sportsbook behavior. applied economics letters, 18, 193–197. peng, l., & xiong, w. (2006). investor attention, overconfidence, and category learning. journal of financial economics, 80, 563–602. thaler, r., & ziemba, w. (1988). anomalies: parimutuel betting markets: racetracks and lotteries. journal of economic perspectives, 2, 161–174. 328 k. krieger / financial services review 24 (2015) 313–329 vergin, r., & scriabin, m. (1978). winning strategies for wagering on national football league games. management science, 24, 809–818. woodland, l., & woodland, b. (1994). market efficiency and the favorite-longshot bias: the baseball betting market. journal of finance, 49, 269–279. zuber, r., gandar, j., & bowers, b. (1985). beating the spread: testing the efficiency of the gambling market of national football league games. journal of political economy, 93, 800–806. 329k. krieger / financial services review 24 (2015) 313–329 from the editor this issue contains volume 29 issue 4 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “financial (il)literacy versus individual’s behavior; evidence on credit card repayment patterns” is coauthored by gustavo barboza at loyola university of new orleans; paola bongini at university of milano-bicocca and monica rossolini at university of milanobicocca. the authors investigate the role that financial (il)literacy and personal traits on financial behavior. they assess the implications of revealed lack of financial knowledge on financial behavior regarding credit card use in comparison with two other cohorts; cohort one answering correctly, and cohort two failing to answer correctly. their empirical findings indicate that among personal-traits overspending results in lack of payment in full in credit card deb, and these effects dominate any gains derived from financial literacy. additionally financial literacy appears to only play a marginal role avoiding month-to-month credit card debt. financial knowledge derived from parents has a strong positive effect on individuals’ financial behavior especially for students characterized by a relevant financial illiteracy. the implications support that early exposure to financial education is strictly preferred and should be promoted at early stages of the educational system. the second article “framing the annuity as bequest protection: an experimental test” is coauthored by ying yan at eastern new mexico state university and russell n. james iii at texas tech university. in this article, the authors investigate how framing partial annuitization as a protection for an intended bequest against the risk of asset exhaustion due to unexpected longevity influences the desire to purchase an annuity. their results indicate that this framing argument does increase interest in purchasing an annuity. the regression results demonstrate that this framing has a larger positive effect for individuals with a greater bequest motive. the third article, “financial literacy to prevent poor borrowing choices” is coauthored by terrance martin at utah valley university, janine k. sam at shepherd university, *, and philip gibson at winthrop university. in this article the authors investigate the impact of financial literacy on the decision to access retirement plan loans before retirement or use one or more high-cost lenders. their results show that being financially literate reduces the likelihood of using high-cost lenders and using retirement-plan loans. additionally, they find evidence of a negative relation between financial literacy and myopic spending. 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) v–vii the final article, “improving collegiate financial literacy via financial education seminars” is coauthored by jonathan handy at western kentucky university, beth pontari at furman university, thomas smythe florida at gulf coast university, and suzy summers at furman university.the authors review a personal finance program developed at a private liberal arts university aimed to improve financial literacy. the authors examine program effectiveness at improving students’ financial knowledge and confidence in their financial future and find that financial knowledge and confidence improve. additionally, women (minorities) narrow their financial knowledge and confidence gaps when compared to men (caucasians). the follow-up analyses show that increases in confidence appear justified in that they are calibrated to increases in knowledge. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review s. michelson / financial services review 29 (2021) v–vii vi thank you to all the fsr reviewers over the past year! abdullah al-bahrani northern kentucky university maher alyousif texas tech sonya britt-lutter bond university gjergji cici university of kansas andrew clare university of london jack dejong nova southeastern catherine d’hondt louvain james dilellio pepperdine yaman erzurumlu bahcesehir university javier estrada iese business school daniel fernandes catholic university of portugal fred fernatt jim gilkeson u of central florida andreas hackethal goethe university frankfurt daniel hoechle u of applied sciences & arts patti j. fisher virginia tech russell james texas tech university vladimir kotomin ilinois state university kyre lahtinen wright state university camilla mazzoli politecnica delle marche steffen meyer darshak patel university of kentucky cliff robb wisconsin madison nic schaub university of st. gallen nic schaub whu martin seay kansas state university jinfei sheng uc irvine peter smith university of york carol springer sargent middle georgia state university ruilin tian north dakota state university bruce vanstone politecnica delle marche william walstad university of nebraska tom warschauer san diego state jamie weathers western michigan university terry zhang australian national university vii s. michelson / financial services review 29 (2021) v–vii from the editor this issue contains volume 28 issue 2 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “constructing tax efficient withdrawal strategies for retirees with traditional 401(k)/iras, roth 401(k)/iras, and taxable accounts” is coauthored by james dilellio at pepperdine and daniel ostrov at santa clara university. in this paper, the authors construct an algorithm for u.s. retirees that computes individualized tax efficient annual withdrawals from tax-deferred, tax-exempt, and taxable accounts. their approach determines the optimal switching times between tax-exempt and taxable account consumption, as well as between tax-deferred and taxable account consumption. their model, accommodates salient tax code features, including dividends, different taxable lots, and required minimum distributions. the second article “the perfect withdrawal amount over the historical record” is authored by e. dante suarez at trinity university. in this research, the author determines the perfect withdrawal amount (pwa) from retirement savings accounts. the pwa is that which, if taken out in the first year of retirement and used again every year adjusted by inflation, leaves exactly the desired final balance on the account. he presents a formula for obtaining the pwa and finds that safety-minded investors should enter retirement with a higher stock allocation than what is currently used in most investment funds designed to provide income during retirement. the third article, “conceptualizing financial advice in australia: the impact of business models and external stakeholders on client’s best interest practice” is coauthored by d.w. richards and e.f. morton at rmit university. in this paper, the authors examine the australian financial advice sector, determine when financial advice can be provided in a client’s best interest, and formulate a model differentiating types of financial advice. they find that some business models prioritize financial institution interests while thwarting external stakeholders encouraging best interest practice. the final article, “active vs. passive, the case of sector equity funds” coauthored by yuhong fan weber state university and crystal yan lin, eastern illinois university. the authors examine performance of 95 actively managed u.s. sector equity mutual funds from 29 fund families relative to their peer exchange-traded funds, spdr sector etfs, in the 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) v–vi period of 2008 to 2017. their results show that passive funds outperform actively managed sector mutual funds. when focusing on the 9 oldest actively managed fidelity sector mutual funds, outperformance in the period of 1999-2010, appears to fade away during the period of 2011-2017. their results indicate that u.s. sector equity market has become more efficient in the past decade. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 28 (2020) v–vi minority household size and the life insurance purchase decision michael a. guillemette, ph.d., cfp�a,*, m. monica hussein, ph.d.,b g. michael phillips, ph.d.b, terrance k. martin, jr., ph.d.c adepartment of personal financial planning, stanley hall 239, university of missouri, columbia, mo 65211, usa bcenter for financial planning and investment and department of finance, financial planning, and insurance at csu northridge, 18111 nordhoff street, northridge, ca 91330-8245, usa cdepartment of economics and finance at the university of texas-pan american, 1201 w. university drive, edinburg, tx 78539, usa abstract this study uses the 1992–2010 survey of consumer finances to analyze whether the likelihood of life insurance ownership and the face value amount of life insurance changes for minorities as household size changes. we find that the likelihood of life insurance ownership declines for black and larger hispanic families as household size increases when controlling for a variety of socioeconomic and demographic variables. there is also a significant decline in the face value amount of term life insurance purchased by black families as household size rises. we provide possible explanations for these effects and also discuss implications for financial planners. © 2015 academy of financial services. all rights reserved. keywords: life insurance; household size; term life insurance; cash value life insurance; race, minority 1. introduction a properly structured life insurance plan can be a powerful part of a family’s financial plan. it is especially useful in managing tax liabilities, unexpected expenses, lost income, and household services after the death of a family member. however, this basic observation is not a truism for all parts of american society. in particular, the existing literature suggests a very * corresponding author. tel.: �1-573-884-9188; fax: �1-573-884-8389. e-mail address: guillemettem@missouri.edu (m.a. guillemette) financial services review 24 (2015) 37–50 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. diverse usage of life insurance ownership among households in different ethnic and racial groups. if these suggestions are correct, financial planners may face special challenges when constructing plans for clients who are in a group that tends to underutilize life insurance products. alternatively, it may be that the decision to use life insurance may be a function of family size, rather than ethnicity. studies on family support networks suggest that the presence of strong family support networks are important in determining how families manage household resources. could family support networks be perceived by clients as a substitute for a properly designed life insurance program? this is an important distinction for financial planners because it could suggest a focus on family size rather than cultural background when deciding how to educate clients about insurance products. this study investigates how family support systems affect life insurance purchasing decisions across various ethnic groups. by performing this analysis, we are able to help distinguish whether possible underinsurance tends to result from strong family support networks accompanying larger family size or whether it tends to be a function of ethnic background, or whether it is an interaction between the two. specifically, we focus on the effects of household size and ethnicity on the face value of insurance policies purchased according to data from the 1992–2010 survey of consumer finances. the results of this study have practical value for financial planners by providing insight into minority households’ predilections regarding life insurance products and the perceived need for such tools in their financial plans. by recognizing minority households’ inclinations towards life insurance when crafting a financial plan, advisors can reinforce their relationships and long-term impact on their clients’ financial wellbeing. 2. literature review insurance provides an effective way for individuals to protect against a severe downturn in consumption. by pooling and sharing risks, beneficiaries will receive their insurance payouts when a covered death occurs (bajtelsmit, 2005). this is an example of smoothing consumption over time, a primary goal of financial planning (ando & modigliani, 1963). on the other hand, households with extended familial ties may plan to rely on the safety net provided by family members to offset the financial loss after the death of a household earner, illustrated in part by the increase since 1990 in the number of older individuals who live in multigenerational households (fry & passel, 2014). there is evidence of interdependence and strong family support networks among minority households (harrison, wilson, pine, chan, & buriel, 1990). minority households also receive greater financial assistance from family members compared with white households (mutran, 1985). white households have relatively higher rates of insurance coverage compared to black households (gutter & hatcher, 2008). large family support systems could be used as a substitute for life insurance ownership as earnings produced by the lost family member may be replaced by other household members. findings indicate that black, hispanic, and asian households utilize adaptive strategies of strong extended family networks and collectivism, and also exhibit group loyalty (harrison et al., 1990). these adaptive strategies foster child-rearing goals of socialization for interdependence. when comparing black and his38 m.a. guillemette et al. / financial services review 24 (2015) 37–50 panic households, black families have larger support systems (mui, 1993; wasserman, brunelli, rauh, & garcia-castro, 1990). studies on family support networks suggest they are important in determining how families manage household resources. tienda and angel (1982) analyze hispanic, black and white families to determine whether extended household structure moderates the impact of labor market disadvantages. they report the following: (1) hispanic and black households are similar in their dependence on extended household support whereas white households are less likely to rely on extended family support; (2) non-immediate members in black and hispanic households contribute significantly to total household income; (3) non-immediate members in white households do not appear to participate contribute significantly to total household income. mutran (1985) finds that when controlling for socioeconomic status, elderly black parents are more likely than elderly white parents to provide financial assistance to adult children. padgett (1997) uses the 1988 national survey on families and households to assess the extent of network involvement and its relationship with household labor. when just examining married couples, approximately half of black couples receive tangible assistance related to household production (padgett, 1997). most commercial transactions include some element of trust (arrow, 1972). if we define trust as the likelihood an individual attributes to the possibility of being cheated (as defined in guiso, sapienza, & zingales, 2008), we may better understand the greater dependence of minorities on their families in contrast to a product sold by an insurance salesman. within a principal-agent theoretical framework, an individual’s level of trust may affect financial decisions (akerlof, 1970), including the decision to purchase life insurance. exchange systems involve a series of individual actions based on assymetric information. trust helps to manage responses to the innate uncertainty of exchange relationships (tyler & stanley, 2007). brehm and rahn (1997) argue that experiences with discrimination may explain the pervasiveness of low trust among black households. discriminatory practices from agents within the financial and capital markets, combined with a history of restricted access to these markets, may lead to a negative perception by minority families of all agents and the products they sell. for example, in the early part of the 20th century major life insurance companies excluded black customers or set discriminatory rates (weems, 1996). as a result, blackowned insurance companies began selling cash value burial policies to provide affordable life insurance coverage for working class black families (weems, 1996). 3. data we use the 1992, 1995, 1998, 2001, 2004, 2007, and 2010 scf for our analysis. the scf is a triennial cross-sectional survey that provides detailed financial information on u.s. households. it contains the most detailed balance sheet information of any publicly available nationally representative dataset (campbell, 2006). it is sponsored by the united states federal reserve board in conjunction with the department of the treasury and other governmental agencies. since the scf oversamples wealthy households, the descriptive statistics in our analysis are weighted to generalize to a nationally representative population (kennickell & woodburn, 1997). the total sample size over the time period was 32,371. 39m.a. guillemette et al. / financial services review 24 (2015) 37–50 there are several factors that have been found to influence the decision to purchase life insurance. the presence of a spouse may increase the demand for life insurance as the purpose of life insurance is human capital replacement for an individual with insurable interest. truett and truett (1990) find age, income, and education level affect the demand for life insurance. because human capital declines with age, age should be negatively associated with the likelihood of owning life insurance. a bachelor’s degree is a human capital signal that proxies for a steeper earnings path that should increase the likelihood of owning life insurance. there is evidence that the demand for life insurance is positively related to the number of dependents in a household (burnett & palmer, 1984; hammond, houston, & melander, 1967). a large amount of liquid assets may decrease demand because of the ability to self-insure. campbell (1980) finds that accumulated household wealth acts as a substitute for life insurance. self-employed individuals do not have access to employer-provided group life insurance which may reduce the likelihood of ownership. households that are currently unemployed may lose access to employer-provided life insurance if the coverage is not portable, which may reduce the likelihood of ownership. campbell (1980) finds bequests to be positively associated with the demand for life insurance. planning to leave a sizable estate may also increase the demand for life insurance, particularly for cash value polices. in terms of premium pricing, a respondent who indicates that they are in fair or poor health should have a relatively higher premium payment than someone who indicates good or excellent health. females have a longer life expectancy compared to males, decreasing the cost of a policy covering females. finally, individual risk preferences should influence the demand for life insurance ownership. 4. descriptive statistics table 1 displays the ownership percentages of life insurance based on race and household size. life insurance ownership declines for black households as they move from a household size of one, three, five, or more members. that trend is not evident for any other racial groups. table 2 shows the median face value amounts of term life insurance held by those who own term life products, sorted by race and household size. table 3 displays the median face value amounts of cash value life insurance held by those who own cash value life products, also sorted by race and household size. in the larger household size categories, only table 1 ownership of life insurance by race and household size (1992–2010 scf) household size black hispanic other white 1 11.73%** 2.75%** 27.55%** 57.97%** 2 6.70%** 2.65%** 28.52%** 62.13%** 3 8.17%** 5.17%** 28.80%** 57.86%** 4 5.11%** 5.96%** 28.40%** 60.53%** 5� 5.84%** 7.34%** 27.13%** 59.69%** *p � 0.05; **p � 0.01. 40 m.a. guillemette et al. / financial services review 24 (2015) 37–50 black and “other” races show a decrease in the face value amount of term life insurance when comparing a household size of four to a household size of five or more. for the “other” race category the same trend holds for cash value life insurance. however, for black households, the face value amount of cash value life insurance increases when moving from a household size of four to a household with five or more members. for hispanic households, when comparing a household of four to a household of five, the face value amount of cash value life insurance actually declines. 5. method a logistic regression model (a) is constructed to better understand how interactions between different races and different levels of household size affect the likelihood of owning life insurance. if the respondent owns a life insurance policy (including individual and group policies, but not accidental life insurance), the variable is coded as one, with no ownership of life insurance as the reference group and coded as zero. race is broken into categories that include white, black, hispanic, and other races. respondents who identified themselves as white are used as the reference group in the multivariate analysis. household size does not include people who do not usually live in the household or who are financially independent. a household size of five includes households with five or more members. our control variables include inflation-adjusted income, inflation-adjusted net worth, and inflation-adjusted liquid assets. income, net worth, and liquid assets were indexed to 2010 dollars and sorted from lowest (q1) to highest (q4) quartile. other control variables include the age of the respondent and whether the respondent has a bachelor’s degree, owns a home, is married, is a male, is self-employed, is employed, plans to leave a sizeable estate, has a child present, self-identifies as healthy,1 and is willing to take substantial financial risk with table 2 median face value of term life insurance by race and household size (1992–2010 scf) household size black hispanic other white 1 $ 30,000 $ 66,000 $ 42,000 $ 27,300 2 $ 40,200 $ 86,250 $100,000 $ 71,550 3 $ 61,500 $ 92,250 $172,500 $135,300 4 $132,090 $115,000 $246,000 $230,000 5� $115,000 $128,790 $136,752 $230,000 table 3 median face value of cash value life insurance by race and household size (1992–2010 scf) household size black hispanic other white 1 $20,000 $ 79,950 $ 82,362 $ 15,750 2 $35,775 $ 75,040 $ 59,052 $ 43,050 3 $40,404 $ 69,930 $105,000 $ 77,700 4 $67,000 $105,000 $200,000 $115,000 5� $73,800 $ 93,800 $155,400 $141,669 41m.a. guillemette et al. / financial services review 24 (2015) 37–50 personal investments. in addition, a dummy variable was included to control for the year in which the survey was conducted. the year 1992 was used as the reference group. (a) own any life insurance � b0 � bj race dummy � bk household size dummy � � bi control variables � � where, i � control variables: income, net worth, liquid assets, bachelor’s degree, ownership of a home, age, married, male, healthy, self-employed, unemployed, plan to leave a sizeable estate, the presence of a child, willingness to take substantial financial risk and the year in which the survey was conducted. two separate tobit models were constructed to understand how interactions between different races and different levels of household size affect the face value of term life insurance (b) and cash value life insurance (c). common examples of cash value policies include whole, straight, or universal life insurance. the dependent variables contain a large percentage of zero values because of non-ownership. when a dependent variable contains a large number of zero values, the use of an ordinary least squares model may result in biased coefficient estimates (madalla, 1987). a tobit model is not subject to this same bias. the insurance face value amounts were inflation adjusted to 2010 dollars and squarerooted to reduce skewness. the face amount of cash value life insurance was broken into quartiles and included as a control variable when the dependent variable was the squarerooted face amount of term life insurance. the face amount of term life insurance was broken into quartiles and included as a control variable when the dependent variable was the square-rooted face amount of cash value life insurance. (b) face amount of term life insurance � b0 � bj race dummy � bk household size dummy � � bi control variables � � where, i � control variables: income, net worth, liquid assets, bachelor’s degree, ownership of a home, age, married, male, healthy, self-employed, unemployed, plan to leave a sizeable estate, the presence of a child, willingness to take substantial financial risk, the year in which the survey was conducted, and the face amount of cash value life insurance (c) face amount of cash value life insurance � b0 � bj race dummy � bk household size dummy � � bi control variables � � where, i � our control variables: income, net worth, liquid assets, bachelor’s degree, ownership of a home, age, married, male, healthy, self-employed, unemployed, plan to leave a sizeable estate, the presence of a child, willingness to take substantial 42 m.a. guillemette et al. / financial services review 24 (2015) 37–50 financial risk, the year in which the survey was conducted, and the face amount of term life insurance 6. results the results for model (a) are displayed in table 4. as household size rises for black and hispanic families, the likelihood of life insurance ownership declines. a black family with a household size of three is 28.85%2 less likely to own life insurance compared with a non-black family of the same household size. a black household with five or more members is 53.55% less likely to own life insurance compared with a non-black family of a comparable household size. a hispanic family with a household size of two is 24.82% less likely to own life insurance compared with a non-hispanic family of two. a hispanic family with five or more members is 59.61% less likely to own life insurance compared with a non-hispanic family of a comparable household size. fig. 1 graphically displays the likelihood of life insurance ownership for black and hispanic families as household size rises. the odds ratios listed in figure 1 are statistically significant. tables 5 and 7 display the results for model (b), or the relation between the face value amount of term life insurance and household size. for black families, as household size rises the conditional mean face amount of term life insurance declines. a black household with two members has $20,4753 less term life insurance compared with a non-black family of the same household size. a black family of three has $74,038 less term life insurance compared with a non-black family of three. the face value of term life insurance is $98,910 less for a black household with five or more members compared with a non-black family of a comparable household size. the average face amount of term life insurance is $28,560 less for a hispanic family of four compared with a non-hispanic family of the same household size. a hispanic family with five or more members has $68,583 less term life insurance compared with a non-hispanic family that is the same size. fig. 2 graphically represents the results in table 7. the odds ratios that are listed are statistically significant. the results for model (c), or the interaction between the face amount of cash value life insurance and household size, are shown in tables 6 and 8. as household size rises the conditional mean face amount of cash value life insurance has a statistically significant decline only for hispanic families with a household size of four or more members. a hispanic family with a household size of four has a face amount of cash value life insurance that is $24,092 less than a non-hispanic family of the same household size. a hispanic family with more than four members has a face amount of cash value life insurance that is $27,680 less than a non-hispanic family of a comparable household size. 7. conclusions the multivariate results indicate a negative relation between household size and the ownership of life insurance for black and hispanic families when controlling for a variety 43m.a. guillemette et al. / financial services review 24 (2015) 37–50 of demographic and socioeconomic variables. the conditional mean face value amount of term life insurance declines for black families as household size rises compared with non-black families. this relation is statistically significant for all black household sizes. the table 4 ownership of life insurance parameter estimate se odds ratio intercept �1.092** 0.0889 not employed �0.833** 0.0406 0.435 male �0.0021 0.0446 0.998 age 0.0124** 0.0012 1.012 income (q2) 0.5637** 0.0388 1.757 income (q3) 0.8704** 0.0496 2.388 income (q4) 1.0568** 0.0687 2.877 net worth (q2) 0.5274** 0.0457 1.695 net worth (q3) 0.4417** 0.0596 1.555 net worth (q4) 0.0985 0.0824 1.103 degree 0.1306** 0.0332 1.140 married 0.4226** 0.0525 1.526 child 0.2816** 0.0542 1.325 homeowner 0.4364** 0.0402 1.547 self-employed �0.4688** 0.0418 0.626 healthy 0.00267 0.0359 1.003 sizable estate 0.0261 0.0314 1.026 liquid assets (q2) 0.5815** 0.0404 1.789 liquid assets (q3) 0.5562** 0.0492 1.744 liquid assets (q4) 0.404** 0.0617 1.498 substantial risk �0.126* 0.0619 0.882 1995 �0.0583 0.0575 0.943 1998 �0.2641** 0.0566 0.768 2001 �0.3721** 0.0561 0.689 2004 �0.5121** 0.0554 0.599 2007 �0.5783** 0.0556 0.561 2010 �0.5856** 0.0513 0.557 black (b) 1.0399** 0.0838 2.829 hispanic (h) �0.3993** 0.1400 0.671 other (o) �0.3267* 0.1542 0.721 household size 2 (hhs2) 0.1528** 0.0536 1.165 household size 3 (hhs3) 0.1066 0.0825 1.112 household size 4 (hhs4) 0.251** 0.0919 1.285 household size 5� (hhs5p) 0.1794 0.0975 1.197 bhhs2 �0.2786* 0.1174 0.8818 bhhs3 �0.447** 0.1352 0.7115 bhhs4 �0.7646** 0.1535 0.5983 bhhs5p �0.9463** 0.1569 0.4645 hhhs2 �0.4381* 0.1751 0.7518 hhhs3 �0.5052** 0.183 0.6713 hhhs4 �0.6937** 0.1804 0.6423 hhhs5p �1.0859** 0.1791 0.4039 ohhs2 0.1891 0.1971 1.4076 ohhs3 0.0341 0.2177 1.1511 ohhs4 �0.1363 0.2356 1.1215 ohhs5p �0.2618 0.2383 0.9209 *p � 0.05; **p � 0.01. 44 m.a. guillemette et al. / financial services review 24 (2015) 37–50 conditional mean face value amount of term and cash value life insurance declines for larger hispanic families as household size rises from four to five or more members. 8. implications and future research our findings suggest that black and hispanic families used household members as a substitute for life insurance between 1992 and 2010. the use of household members as a life insurance substitute is not utility maximizing and a contingent claim should be purchased to replace the lost human capital in the event of a household member’s death. financial planners or life insurance agents can use these findings to help ensure that black households and larger hispanic households are adequately insured. one possible explanation for why black families act as though household members are a substitute for the face value amount of term life insurance, but not cash value life insurance, is that black families purchase cash value policies specifically for burial purposes. the low median face value amounts of cash value policies among black households, as compared with all other racial groups, provides some additional indication that these cash value policies may be used for that purpose.4 however, black families have also been found to have lower risky asset ownership compared with white families as household size rises (gutter, fox, & montalto, 1999) and so it is possible that black families are using cash value life insurance as an alternative investment. financial planners have many tools available to help their clients allocate resources over time. life insurance is an important part of a financial plan as it protects families against a sharp decline in consumption. however, when families have alternative arrangements to meet the needs of the household, such as through a large family network, purchasing an adequate amount of life insurance might appear to be excessive. clients who believe they have a strong family network may be underinsured. heo, grable, and chatterjee (2013) find fig. 1. likelihood of life insurance ownership by household size. 45m.a. guillemette et al. / financial services review 24 (2015) 37–50 table 5 face value of term life insurance parameter estimate se t value intercept �564.9753** 36.214530 �15.60 cash value policy face amount (q2) �271.9539** 19.338508 �14.06 cash value policy face amount (q3) �372.1336** 19.815740 �18.78 cash value policy face amount (q4) �279.7793** 20.650291 �13.55 not employed �253.0329** 16.341234 �15.48 male �11.0729 19.311746 �0.57 age �3.8124** 0.477248 �7.99 income (q2) 165.9838** 17.076209 9.72 income (q3) 281.7889** 19.989716 14.10 income (q4) 605.0263** 25.720547 23.52 net worth (q2) 132.6310** 18.898002 7.02 net worth (q3) 88.1648** 23.424179 3.76 net worth (q4) 188.2409** 30.886998 6.09 degree 98.3133** 12.157400 8.09 married 182.5603** 22.248095 8.21 child 155.3139** 22.170767 7.01 homeowner 123.0864** 16.529780 7.45 self-employed �18.7360 14.569702 �1.29 healthy 32.8656* 14.642588 2.24 sizable estate �8.311813 11.949690 �0.70 liquid assets (q2) 174.3139** 16.859052 10.34 liquid assets (q3) 166.7142** 19.465660 8.56 liquid assets (q4) 223.5183** 23.296634 9.59 substantial risk 73.9107** 22.788373 3.24 1995 �4.9478 20.317286 �0.24 1998 �37.6212 20.403211 �1.84 2001 �37.0724 20.249540 �1.83 2004 �6.9363 20.193088 �0.34 2007 4.7159 20.323864 0.23 2010 14.2123 18.903550 0.75 black (b) 302.9927** 35.091726 8.63 hispanic (h) �94.7716 64.669447 �1.47 other (o) �106.9755 70.962724 �1.51 household size 2 �15.9366 22.818493 �0.70 (hhs2) household size 3 8.7066 34.430109 0.25 (hhs3) household size 4 72.5314* 36.898153 1.97 (hhs4) household size 5� 116.6348** 38.641196 3.02 (hhs5p) bhhs2 �127.1559** 47.217569 �2.69 bhhs3 �280.8052** 53.911229 �5.21 bhhs4 �357.7793** 58.620302 �6.10 bhhs5p �431.1337** 62.465342 �6.90 hhhs2 �66.4825 80.235433 �0.83 hhhs3 �145.0277 82.318090 �1.76 hhhs4 �241.5283** 80.492362 �3.00 hhhs5p �378.5186** 79.949937 �4.73 ohhs2 135.1139 85.355958 1.58 ohhs3 �22.0967 92.065171 �0.24 ohhs4 78.0881 93.930658 0.83 ohhs5p �132.6312 99.445551 �1.33 *p � 0.05; **p � 0.01. 46 m.a. guillemette et al. / financial services review 24 (2015) 37–50 table 6 face amount of cash value life insurance parameter estimate se t value intercept �1746.4416** 56.326721 �31.01 cash value policy face amount (q2) �388.6134** 25.007965 �15.54 cash value policy face amount (q3) �482.7647** 24.978475 �19.33 cash value policy face amount (q4) �454.0463** 24.623215 �18.44 not employed �196.3087** 23.796685 �8.25 male 10.8115 29.600057 0.37 age 4.9954** 0.706360 7.07 income (q2) 79.3575** 26.340385 3.01 income (q3) 123.2469** 30.008617 4.11 income (q4) 379.9408** 36.794123 10.33 net worth (q2) 372.4267** 31.083055 11.98 net worth (q3) 507.1737** 36.780105 13.79 net worth (q4) 816.6910** 45.801190 17.83 degree 62.4141** 17.690651 3.53 married 164.8562** 34.034548 4.84 child 67.8820* 33.701665 2.01 homeowner 108.7534** 25.947366 4.19 self-employed 110.9823** 20.409938 5.44 healthy 11.5404 21.558259 0.54 sizable estate 120.6049** 18.027918 6.69 liquid assets (q2) 222.8966** 27.076805 8.23 liquid assets (q3) 225.5325** 30.046516 7.51 liquid assets (q4) 284.6579** 34.369930 8.28 substantial risk �36.0096 33.015220 �1.09 1995 �47.9328 28.214129 �1.70 1998 �123.0544** 28.516715 �4.32 2001 �207.8471** 28.581703 �7.27 2004 �229.4625** 28.698742 �8.00 2007 �309.7293** 29.104553 �10.64 2010 �375.2035** 27.518708 �13.63 black (b) 336.7514** 56.580048 5.95 hispanic (h) �145.7467 114.830500 �1.27 other (o) �58.1589 109.976406 �0.53 household size 2 152.4559** 34.650184 4.40 (hhs2) household size 3 213.6898** 52.449452 4.07 (hhs3) household size 4 237.3292** 56.109391 4.23 (hhs4) household size 5� 233.6837** 58.688419 3.98 (hhs5p) bhhs2 �110.5899 74.170272 �1.49 bhhs3 �124.4946 83.166274 �1.50 bhhs4 �157.8035 89.189548 �1.77 bhhs5p �124.0041 93.680331 �1.32 hhhs2 �189.7301 140.064691 �1.35 hhhs3 �232.3589 145.041940 �1.60 hhhs4 �392.5461** 143.669057 �2.73 hhhs5p �400.0568** 141.389434 �2.83 ohhs2 �60.1170 130.539406 �0.46 ohhs3 �215.8021 143.603131 �1.50 ohhs4 �182.7971 144.790031 �1.26 ohhs5p �230.8507 153.042356 �1.51 *p � 0.05; **p � 0.01. 47m.a. guillemette et al. / financial services review 24 (2015) 37–50 evidence to suggest that life insurance acts as a compliment, rather than a substitute, for wealth. therefore, minority households may not be thinking about life insurance as a wealth replacement option upon the death of a household member. it can be detrimental to use rules of thumb (such as household size) when assessing the need for life insurance (collins & ham, 2011). recognizing potential cultural biases and the possible role of household members are topics that financial planners should explore when meeting with their clients to discuss what type of life insurance, and how much coverage, is needed to meet the goals outlined within a financial plan. there are multiple areas for future researchers to explore as it pertains to minority household size and the decision to purchase life insurance. future research should focus on why there is a divergence in the decline of the face amount of term and cash value life insurance as household size rises for black families. the question of how much human table 7 decline in term face amount as household size rises parameter estimate bhhs2 �$20,475** bhhs3 �$74,038** bhhs4 �$81,366** bhhs5p �$98,910** hhhs2 �$ 6,793 hhhs3 �$18,583 hhhs4 �$28,560** hhhs5p �$68,583** ohhs2 �$14,203 ohhs3 �$ 179 ohhs4 �$22,686 ohhs5p �$ 256 *p � 0.05; **p � 0.01. fig. 2. face value amount of term life insurance by household size. 48 m.a. guillemette et al. / financial services review 24 (2015) 37–50 capital is left uninsured by minority households as family size rises also remains unanswered. these are questions we hope will be explored in future studies. notes 1 as defined as “excellent” or “good” health. 2 the odds ratio was derived by taking the coefficient for a household size of three (0.1066) and adding it to the interaction variable coefficient (�0.4470). the interaction variable coefficient is multiplied by 1 to designate a black household (0.1066) � (�0.4470)*(1) � �0.3404. the exponential is then taken to get the odds ratio of 0.7115. 3 this value is derived by taking the coefficient for a household size of two (�15.9366) and summing it with the coefficient for a black household size of two (�127.156). the interaction variable coefficient is multiplied by 1 to indicate a black household (�15.9366) � (�127.156)*(1) � �143.092. the solution is then squared since the square root of the dependent variable was originally taken to reduce skewness (�143.092)2̂ � �20,475. the negative sign is retained after the solution is squared. 4 refer to table 3. references akerlof, g. a. (1970). the market for “lemons”: quality uncertainty and the market mechanism. quarterly journal of economics, 84, 488–500. ando, a., & modigliani, f. (1963). the life cycle hypothesis of saving; aggregate implications and tests. american economic review, 53, 55–84. arrow, k. j. (1972). gifts and exchanges. philosophy & public affairs, 1, 343–362. bajtelsmit, v. l. (2005). personal finance: skills for life. hoboken, nj: wiley. braumoeller, b. (2004). hypothesis testing and multiplicative interaction terms. international organization, 58, 807–820. table 8 decline in cash value face amount as household size rises parameter estimate bhhs2 �$ 1,753 bhhs3 �$ 7,956 bhhs4 �$ 6,324 bhhs5p �$12,030 hhhs2 �$ 1,389 hhhs3 �$ 349 hhhs4 �$24,092** hhhs5p �$27,680** ohhs2 �$ 8,526 ohhs3 �$ 4 ohhs4 �$ 2,974 ohhs5p �$ 8 *p � 0.05; **p � 0.01. 49m.a. guillemette et al. / financial services review 24 (2015) 37–50 brehm, j., & rahn, w. (1997). individual-level evidence for the causes and consequences of social capital. american journal of political science, 41, 999–1023. burnett, j., & palmer, b. (1984). examining life insurance ownership through demographic and psychographic characteristics. journal of risk and insurance, 51, 453–467. campbell, j. y. (2006). household finance. journal of finance, 61, 1553–1604. campbell, r. (1980). the demand for life insurance: an application of the economics of uncertainty. journal of finance, 35, 1155–1172. collins, p., & lam, h. d. (2011). asset allocation, human capital, and the demand to hold life insurance in retirement. financial services review, 20, 303–325. fry, r., & passel, j. s. (2014). in post-recession era, young adults drive continuing rise in multi-generational living. pew research center’s social and demographic trends project. washington, dc. http://www. pewsocialtrends.org/files/2014/07/st-2014-07-17-multigen-households-report.pdf guiso, l., sapienza, p., & zingales, l. (2008). trusting the stock market. journal of finance, 63, 2557–2600. gutter, m. s., fox, j. j., & montalto, c. p. (1999). racial differences in investor decision making. financial services review, 8, 149–162. gutter, m. s., & hatcher, c. b. (2008). racial differences in the demand for life insurance. journal of risk and insurance, 75, 677–689. hammond, j. d., houston, d. b., & melander, e. r. (1967). determinants of household life insurance premium expenditure: an empirical investigation. journal of risk and insurance, 34, 397–408. harrison, a. o., wilson, m. n., pine, c. j., chan, s. q., & buriel, r. (1990). family ecologies of ethnic minority children. child development, 61, 347–362. heo, w., grable, j. e., & chatterjee, s. (2013). life insurance consumption as a function of wealth change. financial services review, 22, 389–404. kennickell, a. b., & woodburn, r. l. (1997). consistent weight design for the 1989, 1992 and 1995 scfs and the distribution of wealth. working paper. board of governors of the federal reserve system, washington, dc. maddala, g. s. (1987). limited dependent variable models using panel data. journal of human resources, 22, 307–338. mui, a. c. (1993). self-reported depressive symptoms among black and hispanic frail elders: a sociocultural perspective. journal of applied gerontology, 12, 170–187. mutran, e. (1985). intergenerational family support among blacks and whites: response to culture or to socioeconomic differences. journal of gerontology, may, 382–389. padgett, d. l. (1997). the contribution of support networks to household labor in african american families. journal of family issues, 18, 227–250. tienda, m., & angel, r. (1982). determinants of extended household structure: cultural pattern or economic need? american journal of sociology, 87, 1360–83. truett, d. b., & truett, l. j. (1990). the demand for life insurance in mexico and the united states: a comparative study. journal of risk and insurance, 57, 321–328. tyler, k., & stanley, e. (2007). the role of trust in financial services business relationships. journal of services marketing, 21, 334–344. wasserman, g. a., rauh, v. a., brunelli, s. a., & garcia-castro, m. (1990). psychosocial attributes and life experiences of disadvantaged minority mothers: age and ethnic variations. child development, 61, 566–580. weems, r. e. (1996). black business in a black metropolis: the chicago metropolitan assurance company, 1925–1985. bloomington, in: indiana university press. zietz, e. n. (2003). an examination of the demand for life insurance. risk management and insurance review, 6, 159–191. 50 m.a. guillemette et al. / financial services review 24 (2015) 37–50 finser_31-2_complete_issue academy of financial services officers president tom potts baylor university president-elect executive vice president-program terrance k. martin winston-salem state university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations michelle cull western sydney university vice president-mktg & public relations shawn brayman financial planning research consultant immediate past president inga timmerman university of north florida editor, financial services review terrance k. martin winston-salem state university directors jason andreson university of kansas jasmine fang massey university norah feng massey university john garble university of georgia wookjae heo purdue university phillip gibson winthrop university barry mulholland university of akron utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. institutional membership to academy of financial services is $250 and includes the four annual issues of fsr. how to submit: there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academyfinancial.org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor terrance k. martin at martintk@wssu.edu. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2023 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 31, nos. 2 & 3, 2023 editor terrance k. martin, winston-salem state university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $250. academic subscription price is $125 and is available by joining the academy of financial services. academic subscription includes four emailed online issues of financial services review each year. hardcopies of the journal are available at additional cost. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. editorial office: terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. advertising information. those interested in advertising in the journal should contact terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu printed in the usa © 2023 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president janine sam shepherd university president-elect inga timmerman california state university, northridge executive vice president-program terrance k. martin utah valley university vice president-communications david nanigian california state university, fullerton vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations chris browning texas tech university immediate past president swarn chatterjee university of georgia editor, financial services review stuart michelson stetson university directors shawn brayman planplus global charles chaffin cfp board of standards victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university robert moreschi virginia military institute past presidents swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2020 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 28, no. 2, 2020 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2020 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. a theoretical examination of cash-back credit cards and their effect on consumer spending noah macdonalda,*, brent evansb adepartment of economics, georgia college & state university, 415 atkinson hall, cbx 014, milledgeville, ga 31061, usa bdepartment of economics, georgia college & state university, 415 atkinson hall, cbx 014, milledgeville, ga 31061, usa abstract the role of cash-back credit cards in personal financial strategies is highly debated. for example, dave ramsey (ramsey, 2019) urges consumers to avoid even the most lucrative cash-back cards, while others argue that these cards offer significant savings. herein, we construct models to analyze the use of cash-back cards by rational consumers, demonstrating that cash-back cards increase spending (and, thus, reduce savings) for some consumers. while prior research focuses on behavioral issues related to credit cards, our research is the first to show that some consumers will rationally increase spending when using a cash-back credit card in lieu of cash. © 2020 academy of financial services. all rights reserved. keywords: credit card; personal finance; cash back; retirement planning; churning 1. introduction credit card usage in the united states is at an all-time high. in 2018, u.s. consumers amassed $3.67 trillion in credit card spending, a 9.7% annual increase (nilson report, 2019). not surprisingly, credit card debt is also rising; the average u.s. household carries an average revolving credit card debt of $3,453.1 clearly, credit cards play a significant role in the economy. within this market, rewards-based credit cards are becoming increasingly popular. rewards credit cards offer special incentives to credit card users for initial and continued card use. often, these cards provide “cash back” or frequent flier miles to a credit card *corresponding author. tel.: +1-678-502-0677; fax: +1-478-445-5249. e-mail: noah.matthew.macdonald@gmail.com (noah macdonald) 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 223–242 user. today, about 57% of u.s. citizens hold at least one “rewards” credit card (creditcards. com, 2018) with cash-back cards cited as the most popular and most heavily utilized form of rewards cards (tsys, 2018, p. 23). despite the prominence and relevance of cash-back credit cards, the subject has drawn surprisingly little attention from academic researchers. in this article, we construct and analyze simple (but novel) graphical models to consider how rational consumers respond to the incentives offered by cash-back credit cards. our analyses consider the two styles of rewards that are typical to cash-back credit cards: minimum spend bonuses and per-dollar rewards. minimum spend bonuses provide consumers with a large, onetime bonus for reaching a “minimum spend” level within a given period of time. for example, the popular chase sapphire preferred card currently offers users 60,000 points (worth $600 or more) for spending $4,000 with the card in the first three months of card ownership. while not all rewards cards offer a minimum spend bonus, they almost always provide a per-dollar reward incentive. the aforementioned chase sapphire preferred card, for example, gives consumers a minimum of one percent cash-back on every dollar spent when using the card. by considering typical demand curves of hypothetical, rational consumers, we show that the use of cash-back credit cards will not affect all consumers equally. specifically, consumers with very inelastic demand curves will generally reduce overall expenditures (once rewards are incorporated) with a rewards credit card, relative to cash or checks. in contrast, someone with a very elastic demand curve would greatly enhance spending when using a cash-back card in lieu of cash. beyond these general findings, we find that minimum spend limits may lead to interesting consumption choices for consumers whose spending levels are typically less than the minimum spend threshold offered by the card. in aggregate, our theoretical findings suggest that the connection between the use of cash-back credit cards and one’s savings rate varies greatly across consumers. the implications from the theoretical models included herein are not trivial. if cash-back credit cards increase spending levels for some consumers, the growing popularity of such cards could lead to an overall reduction in savings rates in the united states, spurring shortterm economic growth, but delaying retirement for many. if this is indeed occurring, the burden of savings may fall more heavily on government programs such as social security. while we do not claim that our theoretical model could possibly prove this is occurring, our research will hopefully encourage others to consider the role of credit cards in determining consumption and savings decisions. while enhanced spending from credit card usage has been linked to increased consumer borrowing and money mismanagement, our simple analyses show that increased spending could merely be the result of consumers rationally responding to the incentives provided by rewards cards. 2. the logistics of rewards credit cards the credit cards rewards sphere is surprisingly complicated. virtually all credit cards seem to have their own associated rules regarding spending and redemptions. first, let us consider a relatively simple credit card, the citi double cash card. the citi double cash card effectively offers a two percent cash-back benefit on all purchases (citi, 2019). for example, if you make a $100 purchase with this credit card, you would effectively only pay 224 n. macdonald, b. evans / financial services review 28 (2020) 223–242 $98 after the cash-back rewards is attributed to your account. while there are many simple cards like the citi double cash card, the most highly-coveted rewards credit cards offer a minimum spend bonus (msb). an msb allows consumers to receive a significant reward if they reach a given spending threshold within a given period of time. for example, the wells fargo cash wise card offers a $200 cash bonus if the user makes at least $1,000 in purchases within the first three months in addition to a standard 1.5% cash-back benefit on all purchases. this is a rather lucrative offer. a user that spends $2,000 on the card earns a standard 1.5% return on all purchases ($2,000*1.5% = $30) in addition to a $200 cash return. thus, this consumer is able to spend $1,770 to buy $2,000 worth of goods—an 11.5% cost reduction. the msb is what leads the most savvy credit card users to apply for several credit cards in a given year, reaching the msb on multiple cards to maximize benefits.2 this practice of “credit card churning” has become so popular that many credit card issuers have implemented rules to slow such activities; for example, chase often denies applications for new credit cards for individuals that have applied for five or more credit cards in the last 24months (kerr, 2019). nonetheless, credit card churning is a popular hobby. for example, the message board website, reddit.com, has a forum dedicated solely to the practice of churning (reddit.com/r/churning), which currently has 210,000 members. the aforementioned credit cards are both of the “cash-back” variant. while such cards are certainly popular, many credit card enthusiasts are primarily interested in travel-based rewards cards. for example, consider the aadvantage aviator red world elite mastercard, which (despite its lengthy title) offers a straightforward method for attaining free flights. the card provides 50,000 in aadvantage (american airlines) miles after the cardholder pays the $99 annual fee and makes at least one purchase, which will likely cover two or more round-trip flights anywhere in the united states (barclays, 2019). in addition to this msb, the card also offers two aadvantage miles per dollar spent on american airline purchases and one aadvantage mile per dollar spent on any other purchases. other cards, such as the chase sapphire reserve (csr) offer ancillary benefits like auto-rental insurance, trip cancelation insurance for travel, and zero fees when used abroad, in addition to a lucrative msb and “points-back” on all purchases (chase, 2020).3 these points can be converted to cash, but typically have more value when redeemed for travel. while travel cards are an important fixture of the credit card industry, the value of the points and perks offered by these travel cards is difficult to quantify in dollar terms.4 travel rewards cards are certainly popular, but cash-back credit cards still hold the greater share of the rewards card market. about 80% of consumers primarily use a credit card that provides cash-back redemptions, while only 33% primarily use a card that allows for travel rewards (tsys, 2018, p. 24).5 given the prominence and comparatively straightforward nature of cashback credit cards, our forthcoming analysis focuses solely on cash-back rewards cards. 3. credit cards and personal financial planning it would be unfair to discuss the potential value to be gained from cash-back credit cards without first discussing a significant challenge that millions of credit card users face—debt. the connection between credit card use and debt may seem obvious; credit cards give n. macdonald, b. evans / financial services review 28 (2020) 223–242 225 consumers, who may desire to spend more than they earn, an easy and practical way to borrow money. however, it appears that this connection is actually quite complex. research suggests that credit card usage tends to increase consumer spending, even for those that do not use credit cards as an avenue for borrowing. in other words, credit cards are not borrowing vehicles for many consumers, but still lead consumers to increase their household consumption. there are many potential reasons for this effect. for example, soman (2003) finds that consumers using credit cards are more likely to purchase frivolous or “unnecessary” items than consumers using cash. soll, keeney, and larrick (2013) suggest that credit card transactions tend to be less memorable to consumers, leading consumers to increase spending. no one can be certain why credit card usage tends to spur consumer spending, but it does appear that there is a strong behavioral component. for example, chatterjee and rose (2012) find that consumers using credit cards seem to focus more on the benefit (and less on the cost) of a purchased item, relative to consumers using cash. thus, it is not a surprise that consumers will spend far more when paying with a credit card. perhaps the most notorious example of this spending bump is reported in prelec and simester (2001), who show that consumers roughly double their willingness to pay for highly sought-after sporting event tickets when credit cards are the only accepted payment option. given that credit cards tend to increase consumer spending, it is not surprising that some consumers find themselves, unexpectedly, in substantial debt. wilcox, block, and eisenstein (2011) find that consumers with typically high self-control actually increase spending if they “carry a balance” on their credit card. the authors liken this effect to a smoker that is trying to quit; once a smoker falls off the wagon and has one cigarette, the smoker will likely feel that she has failed to maintain control and will likely smoke many cigarettes after breaking through the self-control threshold. such an effect—individuals abandoning a goal after failure—has been well-documented in psychology research (e.g., cochran and tesser, 1996). research shows that the connection between credit card usage and debt appears to be a function of age and education ability (lopes, 2008), which further illustrates the complexity of credit cards as a viable option for consumers seeking to plan for retirement. for the united states as a whole, revolving credit card debt reached $444 billion in 2019, indicating that an average u.s. household carries $3,453 in such debt (isa, 2019; u.s. census bureau, total households, 2020). that being said, it is important to emphasize that credit card indebtedness is a condition that affects the minority of credit card users. surveys conducted by the federal reserve in five periods between 2004 and 2016 show that 56.2% to 64.0% of credit card users do not “carry a balance” (bricker et al., 2017). thus, for most users, credit cards are used for the convenience they provide or perhaps for some other reason, such as to seek credit card rewards. nonetheless, given that consumers label rewards as the most important feature when choosing a credit card (tsys, 2018, p. 23), there is an inherent indirect connection between credit card debt and rewards that has yet to be formally studied, theoretically or empirically. in the personal finance space, credit card usage is a highly debated topic. one of the leading personal finance gurus, dave ramsey, urges consumers to completely avoid credit cards. ramsey cites aforementioned research that suggests that credit card users spend more money than those utilizing cash. he argues that credit card use will cause consumers to unintentionally go into debt. perhaps, this fear of debt explains why many consumers choose to use 226 n. macdonald, b. evans / financial services review 28 (2020) 223–242 debit cards as their primary payment method (king & king, 2005, 2011). not surprisingly, ramsey does not believe that cash-back credit cards are as lucrative as they initially appear. he writes: the concept is two things. one is—and i’ve got a friend who makes a lot of money, and he uses his amex for everything and gets the cash back and gets the travel points and all that stuff—but i even watch this guy who’s very educated and pretty sophisticated in the handling of his money, and i watch him purchase things for $100 to get the $3 kickback. i always kind of cock my head sideways and go, “you just did that.” the motivation on the $100 spend was the $3 kick. that’s just wrong, you know? what that tells us is it’s motivating people to do things that they wouldn’t normally do because somewhere in the back of their head, they feel like they’re getting rich off this 3% kickback. (ramsey, 2019) in ramsey’s view, the benefits of a cash-back card are miniscule compared with the potential costs of credit card debt. given the tremendous size of dave ramsey’s audience—he has sold over 11 million book copies and hosts a nationally syndicated radio show—it is not surprising that his position on credit cards is the source of substantial scrutiny. responses from the credit card rewards community range from bemusement to complete outrage. jt genter, a writer for the thepointsguy.com offers a more nuanced response: dave ramsey’s argument not to use credit cards is both right and wrong. he’s right that those that can’t control their spending absolutely shouldn’t get or use credit cards. however, his argument is wrong for those who pay their balance in full and utilize the rewards. (genter, 2018) ultimately, there is no academic that could possibly settle this debate. researchers (e.g., soman, 2003 and soll et al., 2013) have found clear evidence that credit card usage is linked with higher spending. however, it is undeniably true that cash back credit cards could serve to reduce the effective cost of transactions for a responsible user. 4. prior research related to credit card rewards research regarding credit card rewards and consumer spending choices is surprisingly sparse; indeed, the lack of research in this arena was a primary motivator for the current research. that being said, there are some tangentially related manuscripts that consider credit card rewards and their impacts. perhaps the most substantial work regarding credit card rewards is a 2009 paper by fumiko hayashi. hayashi considers macroeconomic and long-run effects of the continuing expansion of credit card rewards. while it is clear that consumers can potentially benefit from seeking credit card rewards, hayashi projects that the benefits to a consumer will dissipate over time. her position is based on the various fees that are imposed on merchants that accept credit card payments.6 while all credit and debit card transactions reduce the effective price received by a merchant (after paying fees), rewards credit cards typically impose the highest fees. for example, interchange fees for a visa signature preferred rewards credit card are approximately 0.3 percentage points higher than the fees for traditional visa credit cards (visa, 2019). thus, businesses earn more revenue, net of credit card fees, on a n. macdonald, b. evans / financial services review 28 (2020) 223–242 227 traditional credit card than they do when processing a rewards card. hayashi suggests that firms will be forced to raise prices to account for expanding merchant fees related to rewards cards, eventually eroding the effective savings that consumers enjoy from using said cards. while we do not claim to refute hayashi’s findings, we show that consumers would still receive benefits from using rewards cards, relative to cash or traditional credit cards, even if her projection comes to fruition. a 2010 article from jalbert, stewart, and martin is also of significant interest to the current research. the authors consider the benefits of using credit cards relative to cash, and attempt to reconcile the choice that many consumers make to ignore the value from rewards cards and retain their use of traditional payment methods. their analysis focuses on two specific benefits that credit cards provide. first, they discuss the specific benefits afforded by rewards credit cards; in particular, they emphasize the rewards cards that provide frequent flyer miles. second, they consider a characteristic that all credit cards provide; when using a credit card, consumers can essentially borrow money to purchase an item today and not actually pay for the item for several weeks. this so called “payment float” allows consumers to borrow funds at zero percentage interest for a short period of time. in effect, a consumer could invest the money that is newly available because of payment float to take advantage of the zero percentage interest offered by a card. while the duration of payment float and the speed of rewards-earning differs among cards, jalbert et al. (2010) attempt to quantify the benefits from using a rewards card relative to cash or checks. they report that consumers can generate substantial monetary benefits. for example, an individual with $2,000 in monthly expenses and a discount rate of 10% could expect to reduce the net present value of five years of spending by $7,119 when using a rewards card in lieu of cash.7 the authors conclude the paper by pondering why so many consumers eschew rewards credit cards. while credit card use has been linked to increased spending (e.g., soman, 2003), the authors wonder if the substantial benefits from using rewards cards may dwarf the potential costs of frivolous spending. ricaldi et al. (2013) provide a potential response to jalbert et al. (2010). they suggest that many consumers lack the financial literacy to understand or properly evaluate the merits of rewards credit cards. the authors argue that informed consumers should always prefer rewards credit cards, relative to other credit cards and suggest that credit card firms provide both offerings as a way to segment the market of consumers between the price-sensitive, sophisticated users that will always opt for rewards credit cards and the, perhaps, “naı̈ve” consumers who select cards for nonfinancial or superficial reasons (p. 13). this conclusion is based on their finding that those with low financial literacy were far less likely to use rewards credit cards, even after controlling for other important characteristics such as credit score. the findings from their research suggests that the prevalence and quality of financial education will play a role in shaping the credit card market in future generations. while aforementioned research indicates that credit card rewards are lucrative and, yet, underutilized, rewards are nonetheless an important aspect of payment choices made by consumers. for consumers who hold both debit and credit cards, removing rewards increases the frequency of noncard payments by almost four percent (ching & hayashi, 2010). furthermore, it appears that consumers are somewhat sensitive to the magnitude of rewards; as rewards become more lucrative, consumers are more likely to use rewards credit cards 228 n. macdonald, b. evans / financial services review 28 (2020) 223–242 instead of other payment methods (carten et al., 2007). given that rewards cards are already heavily utilized and younger consumers are more apt to use credit cards relative to prior generations, it appears that rewards cards will continue to gain relevance in the coming years. we hope that the theoretical models constructed herein will further the literature on this meaningful subject area. 5. theory and models using cost and demand curves, we aim to construct models that clearly illustrate the consumption choices made by consumers implementing a cash-back rewards card. to do so, we use actual credit card reward systems that are currently available in the united states. these models allow us to consider how a rational consumer’s spending and savings would be affected (if at all) when the consumer switches from cash to a cash-back rewards card. to simplify our analyses, we must first construct assumptions. then, we consider the consumption choices made by several representative consumers, each with their own unique demand curves. 5.1. assumptions in the following section, we consider how an individual that traditionally uses cash as a payment method would alter consumption habits (if at all) when a rewards credit card suddenly becomes available. both the plausibility of using a rewards credit card (after all, many do not qualify for credit cards) and the spending choices that an individual would make are clearly unique to a given individual. to simplify, we consider representative consumers that are bounded by the following assumptions. 1. consumers may use cash or a cash-back credit card for all purchases. 2. consumers always pay their credit card balance in full; thus, the interest rate that a card offers is inconsequential. 3. consumers are indifferent among payment methods and expenditures will not vary across payment methods, ceteris paribus. 4. attaining a credit card is costless and credit cards charge no annual fee. 5. credit cards only offer cash-back rewards and provide no other perks. 6. consumers maximize utility, which is measured by consumer surplus. 7. consumers have a discount rate of zero. 8. consumers have a price-elasticity of demand that is neither perfectly elastic nor perfectly inelastic. each of these assumptions is crucial to developing a feasible model that is generalizable to all consumers. assumption 1 allows consumers to freely choose whether to use cash or a cash-back credit card for all purchases; thus, there is never a situation where a consumer is forced to use a given payment method.8 assumptions 2, 3, and 4 simplify the consumption choice made by consumers. if consumers do not use credit cards as a source of borrowing, are indifferent among payment options, and can freely choose to use any credit card, their payment choice will be based on the incentives (or lack-thereof) specific to a given credit n. macdonald, b. evans / financial services review 28 (2020) 223–242 229 card. we assume that rewards cards only offer cash-back rewards (assumption 5), ignoring the complex world of travel rewards. because cash-back is by far the most common rewards redemption method (tsys, 2018, p. 24), this assumption effectively matches reality for many consumers. assumption 6 provides the numerical framework for our analyses. consumer surplus, essentially, measures the value of a transaction to a consumer. for example, if a consumer is willing to spend $100 for an item, but said item only costs $70, we can infer that the consumer earned $30 worth of “consumer surplus” from the transaction. we assume that consumer utility is perfectly represented by consumer surplus and that consumers all wish to maximize consumer surplus. credit cards offer consumers the benefit of providing a “payment float.” as fully described earlier in the article, one can defer the cash payment of an item into the future by using a credit card. considering the time value of a money, a consumer is likely to gain positive utility from taking advantage of payment float. this is not the only time-value consideration for credit cards. credit card rewards are often paid in the future; that is, if an individual earns cash-back from a rewards card, it may be weeks before these funds become available to the consumer. to simplify our analysis, we include assumption 7, which states that consumers have no discount rate; thus, they value money equally across all time periods. finally, assumption 8 assumes that consumers possess “normal” downward-sloping demand curves. this assumption allows us to apply the “law of demand” in our analyses— when prices drop, consumption rises. collectively, these assumptions simplify our analysis to allow for enhanced clarity, but it is important to note that our assumptions do not necessarily reflect reality. for example, many of the most lucrative credit cards enforce an annual fee and offer significant ancillary benefits, such as access to airport lounges. while such complexities stray beyond the scope of our current research, we hope to incorporate these concepts in future endeavors. 5.2. consumption with a simple cash-back card given the assumptions described, we attempt to model the consumption choices for several consumers that are choosing between cash and a cash-back rewards credit card. we consider two different credit cards. the first credit card has a simple cash-back system, where the consumer earns a flat two percent cash-back on all purchases. the second card includes both an msb and a cash-back component. given that rational consumers consider relevant marginal benefits and costs when determining their consumption levels, we attempt to determine the efficient consumption choice for representative consumers. fig. 1 models a consumer choosing between consumption with cash or a simple cashback rewards credit card. for our purposes, we use the citi double cash card, which offers consumers a flat two percent cash-back on all purchases. this card is ideal for our analyses because of the simplicity of its rewards program and its lack of an annual fee, consistent with assumption 4. in the absence of credit card rewards, the marginal cost (mc) of $1 of goods purchased remains constant at $1, as shown by the “marginal cost w/o cc” curve. for this consumer, buying items that amount to $1,000 in “gross expenditure” will cost 230 n. macdonald, b. evans / financial services review 28 (2020) 223–242 exactly $1,000, ignoring sales tax.9 in our model, the marginal benefit (mb) gained from each additional unit of goods as measured by gross expenditure diminishes as consumption increases. this is consistent with the law of diminishing marginal utility. marginal benefit is represented by the demand curve. given a consumer’s mb and mc curves, there is some optimal consumption quantity, c0, such that mb = mc. at this level of consumption, the consumer has consumed all goods for which the mb ≥ mc and is not consuming any goods in which the mb < mc; the consumer has maximized utility. area “a” represents the consumer surplus earned by this consumer. the cost structure described above changes with the advent of cash-back credit cards. as a result of the two percent cash-back reward, the marginal cost of $1 of goods reduces to an “effective cost” of $0.98. because the marginal cost of $1 of goods purchased decreases, the optimal quantity consumed also increases from c0 to c1. this results in changes to consumer surplus as well. consumer surplus increases to include areas “b” and “c”; each of these areas has its own unique interpretation. area b results from the items that the consumer would purchase with cash or with the two percent cash-back credit card. however, with the card, she effectively reduces the marginal cost from $1.00 to $0.98 for each dollar of “gross expenditure.” while area b indicates a utility-improvement for the consumer, it is not a result of the consumer changing behavior; she is merely enjoying a cost reduction. on the contrary, area c is a result of the new consumption that a consumer chooses to make since all items are now, effectively, two percent cheaper. the benefit received (area c) is determined by consumers’ price elasticity of demand. consumers with more elastic demand curves are more responsive to price changes, and are also very responsive to credit card rewards, which effectively lowers the price of consumption. fig. 1. consumption using a simple cash-back credit card. n. macdonald, b. evans / financial services review 28 (2020) 223–242 231 to further explore the changes in consumption made by a consumer, let’s consider three representative cases. each of these cases considers how consumer spending would change as a result of the two percent cash-back card with varying levels of demand elasticity. fig. 2 shows how a consumer with a unit elastic demand curve would react to a two percent unlimited cash-back rewards card. if monthly consumption using cash, c0, is $10,000, then their post-reward consumption, c1, would increase to $10,204.08.10 after the two percent cashback redemption, their effective spending remains constant at $10,000. thus, consumers with unit elastic demand curves effectively spend the same amount of money with cash and with a credit card, but they are able to consume more goods and services with the cash-back credit card. for these consumers, utilization of the citi double cash card would have a net neutral effect on their savings rate, while enabling increased consumption. now, let us consider the case of a consumer with an elastic demand curve. fig. 3 shows a consumer with an initial consumption level of $10,000/month, but a price elasticity of demand of two. given this relatively flat demand curve, this consumer would actually spend more money (even after cash-back redemptions) with the introduction of rewards opportunities. specifically, with a price elasticity of two, this consumer would spend $10,412.37 buying goods and services, effectively spending $10,204.12 after receiving the two percent cash-back. because this consumer increases their spending by a greater percentage than the effective price decreases, they are effectively spending more money after rewards implementation. thus, the use of this two percent cash-back card will reduce consumer saving for an individual with elastic demand. that being said, it is important to recognize that the additional consumption still raises consumer surplus and utility. fig. 2. consumption using a simple cash-back credit card (unit elastic demand). 232 n. macdonald, b. evans / financial services review 28 (2020) 223–242 finally, let us consider a consumer with an inelastic demand curve. fig. 4 shows a consumer with a price elasticity of demand of 0.5. if this consumer spends $10,000/month with cash, her consumption rises to $10,101.52. after the two percent cash-back reward, she now effectively spends $9.899.49. this consumer increases spending by a smaller percentage fig. 4. consumption using a simple cash-back credit card (inelastic demand). fig. 3. consumption using a simple cash-back credit card (elastic demand). n. macdonald, b. evans / financial services review 28 (2020) 223–242 233 than the effective price decreases, which means she spends less money as a result of credit card rewards. thus, her savings rate increases. considering the results from figs. 1, 2, 3, and 4, in unison, it is clear that cash-back credit cards can only have a positive effect on consumer utility, if our (strenuous) assumptions hold. for a rational consumer, consumer surplus, and thus, utility, increases once a consumer takes advantage of the two percent cash-back offer provided by the citi double cash card. however, it also plainly evident that this gain in utility could serve to hamper retirement planning for consumers with relatively elastic demand. 5.3. consumption with a cash-back card that offers a minimum spend bonus not all cash-back credit cards are as straightforward as the citi double cash card; rewards become much more complicated with the introduction of minimum spend bonuses (msb). for example, let’s consider the capitol one savor rewards card, once again chosen for its relative simplicity. the capital one savor rewards card offers a $300 cash bonus if the consumer spends $3,000 over the first three months of card ownership (capital one, 2019). in addition, the card also offers one percent cash-back on all purchases.11 because of the design of minimum spend bonuses, the graphical analyses employed must be reconsidered. consider the marginal cost of a dollar spent while using this card. technically, the 3,000th dollar spent would possess a marginal cost of -$299. in other words, by spending the 3,000th dollar, the consumer is able to reduce their overall cost by $299 once they receive the $300 cash bonus. as a result, traditional marginal analysis is no longer appropriate. to more accurately model how consumers would approach consumption decisions with an msb card, we instead consider average cost per dollar spent through the msb spending range. the need for this for this adjustment should reveal itself as we traverse through the following examples. fig. 5 illustrates a consumer’s consumption choice given three options. while demand (marginal benefit) shows willingness to pay in a typical fashion, the consumer now faces three payment choice options. first, this consumer could use cash (marginal cost w/o cc), which would lead to no rebate on consumption—every $1 spent on items would have an effective cost of $1. the consumer could also choose to use the credit card but fail to meet the msb. in this case, the consumer receives a flat one percent cash-back reward on every dollar spent leading to an effective cost of $0.99 per $1 of spending. lastly, the consumer could choose to spend enough money to reach the minimum spend (ms) threshold. if the consumer surpasses the minimum spend threshold, the consumer’s average cost would equal $0.89 per dollar spent on the first $3,000 of consumption. how did we arrive at this value? if the consumer spends exactly $3,000, he receives one percent cash-back on all purchases ($30) in addition to $300 for reaching the msb. because he effectively pays only $2,670 to purchase $3,000 worth of goods, he is effectively paying $0.89 in average cost per $1 of normal spending. since the bonus is only applied if he reaches $3,000 in consumption, he will consider the average costs and benefits of spending choices to determine if it is worthwhile to reach the msb. after he reaches the msb, however, his choice is less complicated as credit card will now function as a simple cash-back card with a $0.99 effective marginal cost. 234 n. macdonald, b. evans / financial services review 28 (2020) 223–242 in fig. 5, the consumer has adequately high demand (ms < c0) leading to a straightforward consumption choice. if the consumer uses cash, he will consume at level c0, leading to a consumer surplus of area a. however, if he instead uses the card, he will effortlessly reach the msb because his current consumption already surpasses the $3,000 level. however, as with figs. 2-5, his consumption will slightly increase from c0 to c1 as a result of the flat one percent cash-back reward. in total, the consumer surplus increases by areas b, c, d, and e. areas b and c result from the consumer earning one percent cash-back on the items he would buy with cash or card. area d results from new consumption that was incentivized by the one percent cash-back reward. finally, area e is the $300 minimum spend bonus earned by the consumer. given that this consumer is earning a $300 cash reward, it is likely that this credit card would have a positive effect on savings. although the consumer is making more purchases (as evidenced by the increase from c0 to c1), his demand would need to be absurdly elastic for this consumption increase to outstrip the $300 in new income that was afforded by the msb. thus, for consumers with high demand, it appears that the use of cash-back cards with msb will have a positive effect on both spending and saving. consumption choices become more complex when a consumer’s typical spending level is below the msb threshold. fig. 6 shows a consumer that will not meet the msb if they use cash. however, this consumer would indeed choose to increase spending to reach the msb. using cash, this consumer would not choose to purchase items that would extend their consumption from c0 to c1. however, the consumer will recognize that the average cost per dollar spent would drop tremendously if they increase consumption to c1 to capture the msb. for example, if this consumer typically spends $2,800 every three months, it is abundantly clear that they would benefit from spending an extra $200 on their card in order to receive a $300 cash bonus. once the consumer reaches the msb, however, they would then choose to cease all consumption since the marginal cost ($0.99 per $1 of goods and services) surpasses fig. 5. consumption with a minimum spend bonus (high demand). n. macdonald, b. evans / financial services review 28 (2020) 223–242 235 the marginal benefit. using the specific values prescribed above, this consumer would reduce effective spending by using the card, leading to increased savings. however, this is not necessarily true for all consumers. if, instead, a similar consumer increased consumption from $2,000 (c0) to $3,000 (c1) to reach the msb, this consumer would increase effective spending, reducing savings. thus, the effect on savings would be a function of the magnitude and elasticity of the demand curve. for the specific consumer displayed in fig. 6, their spending would equal c1 with a flat one percent cash-back credit card or with the msb card with one percent cash-back. thus, in this specific case, this consumer would save an extra $300 (the full value of the msb) if they use this credit card instead of a one percent cashback card with no msb. clearly, this would boost savings. low demand consumers, like those shown in fig. 7, are the most difficult to analyze when using an msb credit card. such consumers do not come close to reaching the minimum spend threshold when using cash. to analyze consumption for this individual, we must introduce a novel concept—negative consumer surplus. typically, consumer surplus is defined by consumption for which the marginal benefit exceeds the marginal cost (price). however, in the forthcoming scenario, the consumer willingly consumes units where the marginal cost exceeds marginal benefit leading to consumption that yields negative consumer surplus. on its own, such consumption reduces utility; however, through this increased spending the consumer is able to reduce the effective cost of consumption for all purchases, potentially leading to an overall increase in utility. the following paragraphs discuss this unique scenario. without credit cards, this consumer would choose to buy a small amount of goods and services, reaching a consumption level of c0. in a typical benefit-cost model, this consumer would not want to spend more than c1, given that the marginal cost exceeds marginal benefit fig. 6. consumption with a minimum spend bonus (medium demand). 236 n. macdonald, b. evans / financial services review 28 (2020) 223–242 for these goods. however, this consumer would, yet again, consider the average cost of spending if they spent exactly $3,000 to reach the msb. certainly, this consumer (and all consumers bounded by the assumptions) would prefer to use the credit card in lieu of cash, regardless of spending level. thus, if this consumer fails to reach the msb, he will face an effective average cost of $0.99 per dollar of spending, leading to a consumption level of c1 and a consumer surplus of areas a and b. however, the consumer could also choose to consume at level c2 to reach the msb. in this case, the average cost is $0.89 per $1 spent. the consumer would gain new consumer surplus equal to area c. however, they would also need to consume units for which the marginal costs exceed the marginal benefits, leading to a negative consumer surplus of area d. while this seems strange, the logic of such a choice is feasible. perhaps this consumer only spends $1,000 when using a simple one percent cash-back card, but would choose to spend $2,200 in total if the effective cost per dollar spent were reduced to $0.89. this consumer may be willing to spend an additional $800 to buy goods and services that they only value at $600 to reduce their average cost. graphically, the consumer is left with an interesting choice. if area c is greater than area d, the consumer should choose to spend enough money to reach the msb (and then cease consuming). if area d, however, is greater than c, this consumer will merely consume at level c1, earning the flat cash-back rate of one percent. in the former case, it is possible that consumers could increase spending to such a degree that this could greatly reduce their savings rate. however, as in prior cases, this reduction in savings would be a conscious choice made by the consumer that would increase consumer surplus, and in turn, yield higher utility levels. fig. 7. consumption with a minimum spend bonus (low demand). n. macdonald, b. evans / financial services review 28 (2020) 223–242 237 6. discussion in light of the novel, and at times, complicated, models described in section 4, we hope that the overall findings will leave the reader with an enhanced understanding of how credit card use effects consumer spending. using simple demand curves for all products, we clearly show that even for rational consumers, cash-back credit cards, will, in some scenarios, incentivize increased consumer spending (even after the consumer receives the cashback award). thus, it appears that cash-back credit cards could lead to reduced savings for some consumers, potentially delaying retirement. prior research has explored the influence of behavioral phenomena on consumer spending. for instance, chatterjee and rose (2012) show that credit card users tend to overvalue the benefits of goods and services when they use a credit card to make a purchase. such findings rely on distorted behavior that aligns with irrationality on the part of consumers. however, our findings show that even fully rational agents may consume more when utilizing rewards credit cards. for example, consider the effects of a simple one percent cash-back credit card. given that this credit card reduces the cost of spending, we posit that any consumer with a nonperfectly inelastic demand curve would spend more in the face of this new incentive to spend. however, the overall effects on spending, once cash-back rewards are considered, depend on how much extra spending the consumer chose to make. if the one percent cashback rewards card caused a consumer to increase consumption by 0.5%, this consumer will actually spend less money once the cash-back reward is attributed to her account. this consumer has an inelastic demand curve; thus, her effective spending drops and her savings rate increases. in contrast, if the one percent cash-back card causes the consumer to spend three percent more on goods and services, this consumer will, effectively, increase spending by two percent and reduce savings. thus, for the consumer using a simple cash-back credit card, the more elastic their demand for all goods and services, the more spending will increase as a result of cash-back credit card utilization. elastic demand curves are not the only potential culprit of increased spending. minimum spend bonuses offered by many rewards credit cards may also lead to increased spending. for example, the capital one cash rewards savor card offers $300 cash-back reward for a consumer that manages to spend $3,000 over their first three months of card ownership. for a rational consumer that already spends this much money quarterly, this card will likely lead to reduced effective spending and an increased savings rate. however, some consumers that typically spend much less may rationally choose to spend just enough to reach the threshold. for example, a consumer that spends $2,000 every three months may find it beneficial to spend $1,000 extra to earn the $300 reward, leading to a $700 increase in effective spending. in both cases described above, it is important to understand that these consumption increases made by the consumers are still rational—these choices make the consumer better off. we show in our graphical analyses that consumers will increase consumer surplus by making these savings-reducing consumption choices. thus, it seems that credit cards may put many consumers in an awkward position—the rational consumer may choose to spend more, increasing utility, but reducing savings. 238 n. macdonald, b. evans / financial services review 28 (2020) 223–242 the findings from this theoretical study may do little to temper any arguments regarding the efficacy of using rewards credit cards as a personal finance tool. advocates of rewards cards can use our findings to demonstrate how some consumers can use credit cards to increase spending and consumption simultaneously; this will be true for consumers with an inelastic demand curve using a simple cash-back card and for consumers that already spend an adequate amount to reach minimum spend bonuses. furthermore, our study shows that all rational consumers are made better off by cash-back cards, even if the cards incentivized increased consumption. conversely, for dave ramsey and other credit card skeptics, the findings herein only serve to show another way that credit cards can reduce savings. if rational credit card use leads to more spending for some, just imagine how irrational and/or irresponsible consumers may react when they get their paws on a cash-back card. ultimately, digesting the findings included in our article requires nuance. for the individual user, cash-back credit cards can clearly enhance utility, when used responsibly. however, the increased spending that could result from using these cards would certainly exacerbate any irresponsible usage. 7. conclusions in an effort to gain a better understanding of the role of cash-back credit cards in a personal financial plan, we construct graphical models of consumer demand. our novel analysis shows that cash-back credit cards have varying effects on consumer spending choices. focusing only on rational consumers that always pay off their balance in full, we show that cash-back cards will cause some consumers to increase spending and reduce savings. in particular, consumers with elastic demand are likely to increase spending when they earn a set cash-back rate on their credit cards. relative to cash, a consumer with a demand elasticity of four would spend four percent more on goods services when using a one percent cash-back card; effectively, this consumer spends about three percent more, net of cash-back rewards. additionally, some consumers, especially those with low expenses, may be incentivized to increase spending to reach minimum spend bonuses often offered by credit card firms. prior research has focused on the behavioral effects of credit cards, suggesting that consumers may irrationally increase spending when using a credit card, perhaps because credit card purchases are less memorable (soll et al., 2013), which potentially causes consumers to underestimate the true cost of a good or service (chatterjee and rose, 2012). our findings are novel since they show that consumer spending may increase, rationally, as a response to the rewards offered by credit cards. as a result, cash-back credit cards, while utility-improving, may reduce savings rates and impede one’s progress towards retirement—this is true even for the most responsible and rational of consumers. given the frequent debates over the viability of credit cards in a personal finance plan, our research offers new insights that will hopefully spark more empirical scrutiny. this is the primary contribution of our research. in addition to providing important insight into the consumption choices of rewards cardholders, the models above constitute a novel approach to quantifying rewards benefits and a unique application of marginal cost and benefit curves for modeling consumption behavior. n. macdonald, b. evans / financial services review 28 (2020) 223–242 239 these unique models could easily be adapted for use in undergraduate economics lessons surrounding marginal benefits and costs, consumer surplus, and price elasticity of demand. given that rewards credit cards are now a primary method of payment in the united states and many other developed nations, it is imperative that more research be conducted in the field. we hope that our theoretical methods encourage more research examining the wide-reaching effects of expanding rewards credit card usage. notes 1 this figured was found by dividing $444 billion in total credit card debt (issa, 2019) by the total number of us households (census.gov, 2019). 2 for a colorful example of credit card churning, see lozano (2019). 3 cards with significant ancillary benefits typically carry high annual fees. for example, the chase sapphire reserve enforces a $550 annual fee, which is among the highest in the industry. given the complexity of such cards, we choose to focus our forthcoming analysis on cards with no annual fees and minimal ancillary benefits. 4 for an academic discussion of the value of points and miles in credit card rewards programs, see jalbert et al. (2010). 5 note that many cards, such as the aforementioned chase sapphire preferred, provide cash and travel rewards redemptions. 6 her analysis considers technical matters related to credit cards that aren’t relevant to our current research. however, we encourage interested parties to read her paper which fully explains the credit card payment process and potential ramifications. 7 however, it’s important to remember that credit card use has been linked to increased spending (e.g. soman 2003). 8 generally, consumers are unable to use credit cards for some spending categories. for example, mortgage payments often require cash, check, or a bank transfer. our assumption simplified a consumer’s choices, focusing only on the scenarios in which a consumer can freely choose his or her payment method. 9 credit card points are indeed earned on sales tax. thus, the sales tax rate levied by a locality has no effect on the relative benefit of using a rewards credit card. 10 calculations are made using the mid-point formula. 11 technically, the card offers “4% cash back on dining and entertainment, 2% at grocery stores, and 1% on all other purchases” (capital one, 2019). for simplicity, we only consider the 1% cash back return that would apply to the majority of purchases for a typical consumer. references barclays. (2019). aadvantage aviator red world elite mastercard. (available at https://cards.barclaycardus. com/banking/cards/aadvantage-aviator-red-world-elite-mastercard/). bricker, j., dettling, l. j., henriques, a., hsu, j. w., jacobs, l., moore, k. b., pack, s., sabelhaus, j., thompson, j., &., windle, r. a. of the board’s division of research and statistics prepared this article with 240 n. macdonald, b. evans / financial services review 28 (2020) 223–242 assistance from p. hansen and e. llanes. (2017). changes in u.s. family finances from 2013 to 2016: evidence from the survey of consumer finance. (available at https://www.federalreserve.gov/publications/ files/scf17.pdf). capital one. (2019). savor rewards from capital one. capitalone.com. (available at https://www.capitalone. com/credit-cards/savor-dining-rewards/). carten, m., littman, d., schuh, s., & stavins, j. (2007). consumer behavior and payment choice: 2006 conference summary (public policy discussion paper no. 07-4). boston, ma: federal reserve bank of boston. census.gov. (2019). 2019 u.s. population estimates continue to show the nation’s growth is slowing. (available at https://www.census.gov/newsroom/press-releases/2019/popest-nation.html). chase. (2020). chase sapphire reserve. chase.com. (available at https://creditcards.chase.com/rewards-creditcards/sapphire/reserve). chatterjee, p., & rose, r. l. (2012). do payment mechanisms change the way consumers perceive products? journal of consumer research, 38, 1129-1139. ching, a. t., & hayashi, f. (2010). payment card rewards programs and consumer payment choice. journal of banking & finance, 34, 1773-1787. citi. (2019). citi credit cards. citi.com. (available at https://www.citi.com/credit-cards/credit-card-details/citi. action?id=citi-double-cash-credit-card). cochran, w., & tesser, a. (1996). the ‘what the hell effect’: some effects of goal proximity and goal framing on performance. in l. l. martin & a. tesser (eds.), striving and feeling: interactions among goals, affect, and self-regulation (pp. 99-120). mahwah, nj: lawrence erlbaum & associates, creditcards.com. (2018). americans love their cash back credit cards. article by ted rossman. (available at https://www.creditcards.com/credit-card-news/survey-americans-prefer-cash-back-credit-cards.php). genter, j. t. (2018). why dave ramsey is both right and wrong on credit cards. (available at https:// thepointsguy.com/2018/03/dave-ramsey-right-wrong-credit-cards/). hayashi, f. (2009). do u.s. consumers really benefit from payment card rewards?. economic review, 94, 37-63. isa, e. e. (2019). 2019 american household credit card debt study. (available at https://www.nerdwallet.com/ blog/average-credit-card-debt-household/#footnote-two). jalbert, t., stewart, j. d., & martin, d. (2010). the value of credit card benefits. financial services review, 19, 227-244. kerr, r. (2019). chase’s 5/24 rule: everything you need to know. (available at https://thepointsguy.com/guide/ ultimate-guide-chase-5-24-rule/). king, a. s., & king, j. t. (2011). golden eggs versus plastic eggs: hyperbolic preferences and the persistence of debit. journal of economics and finance, 35, 93-103. king, a. s., & king, j. t. (2005). the decision between debit and credit: finance charges, float and fear. financial services review, 14, 21-36. lopes, p. (2008). credit card debt and default over the life cycle. journal of money, credit and banking, 40, 769-790. report, n. (2019). u.s. card spending at merchants. the nilson report. (available at https://nilsonreport.com/ publication_the_current_issue.php). o’shea, b. (2019). what is a fico score? (available at https://www.nerdwallet.com/blog/finance/comparingcredit-scores/). prelec, d., & simester, d. (2001). always leave home without it: a further investigation of the credit-card effect on willingness to pay.marketing letters, 12, 5-12. ramsey, d. (2019). the philosophy of credit card use. daveramsey.com. (available at https://www. daveramsey.com/askdave/debt/the-philosophy-of-credit-card-use). ricaldi, l., finke, m. s., & huston, s. j. (2013). financial literacy and shrouded credit card rewards. journal of financial services marketing, 18, 177-187. soll, j. b., keeney, r. l., & larrick, r. p. (2013). consumer misunderstanding of credit card use, payments, and debt: causes and solutions. journal of public policy & marketing, 32, 66-81. soman, d. (2003). the effect of payment transparency on consumption: quasi-experiments from the field. marketing letters, 14, 173-183. n. macdonald, b. evans / financial services review 28 (2020) 223–242 241 tsys. (2018). consumer payment study. 2018 tsys u.s. (available at https://www.tsys.com/assets/tsys/ downloads/rs_2018-us-consumer-payment-study.pdf). u.s. census bureau, total households. (2020). [ttlhh], retrieved from fred, federal reserve bank of st. louis. (available at https://fred.stlouisfed.org/series/ttlhh). visa. (2019). visa usa interchange reimbursement fees. (available at https://usa.visa.com/dam/vcom/ download/merchants/visa-usa-interchange-reimbursement-fees.pdf). wilcox, k., block, l. g., & eisenstein, e. m. (2011). leave home without it? the effects of credit card debt and available credit on spending. journal of marketing research, 48, s78-s90. 242 n. macdonald, b. evans / financial services review 28 (2020) 223–242 college student interest in personal finance education christine harrington, ph.d.a,*, walter smith, ph.d.b aassistant professor of finance, department of business administration, college of business, auburn university at montgomery, p.o. box 244023, montgomery, al 36124-4023, usa bassociate professor of accounting, school of accountancy, college of business, auburn university at montgomery, p.o. box 244023, montgomery, al 36124-4023, usa abstract this study investigates demand for investing in financial literacy while in college using survey responses from a cross-section of students at a medium-size, private university. results indicate that student interest in personal finance education is largely a function of perceived return, time cost, financial independence, and gender where female students have relatively more interest. income, patience in consumption, credit experience, numerical ability, and other factors are not consistently significant to demand. the results support offering learning opportunities for individual personal finance topics in addition to a personal finance course. © 2016 academy of financial services. all rights reserved. jel classification: d1; i2 keywords: college student; demand; personal finance education; financial literacy 1. introduction financial education has long been thought to improve financial outcomes. efforts to educate u.s. youth in primary and secondary schools resurged in the 1970s after falling out of favor (e.g., bernheim, garrett, and maki, 2001). however, u.s. state-mandated personal finance education in grades k-12 is unevenly applied (council for economic education, 2014), and its efficacy on financial outcomes is mixed (e.g., hastings, madrian, and skimmyhorn, 2013). * corresponding author. tel.: �1-334-244-3513; fax: �1-334-244-3792. e-mail address: charrin1@aum.edu (c. harrington) financial services review 25 (2016) 351–372 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. college may be one of the last opportunities to help the young become more financially knowledgeable and confident. however, personal finance education is voluntary for many college students. the challenge is to increase student participation, but little is known about the drivers of student interest in personal finance education. for example, lyons (2004) finds that financially at-risk students are relatively more interested in campus-provided money management information. more recently, beierlein and neverett (2013) examine the characteristics of students who take a personal finance course and find that gender, major, verbal sat score, and high school gpa influence enrollments. this study examines undergraduate college student demand for personal finance education using classic determinants of demand and gender to gain insights into increasing student participation in financial education. the sample is from a medium-size, private university that does not require personal finance as part of the curriculum. the campus has a small but emerging financial literacy program limited to salt, a one-week speaker series each semester, and individual counseling from the financial aid office and the wellness center. because of these limited opportunities for students, we characterize demand as stated interest in personal finance education. based on the theory of investment in financial literacy in jappelli and padula (2013), we model college student demand for personal finance education (the general body of knowledge) as a function of income, patience in consumption, time or effort cost as a price variable, initial stock of financial literacy, perceived return on financial literacy, and financial independence. gender is added as an experimental variable. we also examine the demand for learning about specific topics within personal finance that are likely to benefit college students such as personal budgeting, credit cards, student loans, and an “other topics” category to capture interest in learning about retirement planning, leasing versus buying, and so forth.1 prior studies indicate that numerical ability is associated with patience in consumption, the cost of acquiring financial literacy, and the initial stock of financial literacy (e.g., agarwal and mazumder, 2013; bansak and starr, 2010; cole, paulson, and shastry, 2016; jappelli and padula, 2013). we explore the influence of math ability on demand for personal finance education by using a student’s self-reported math ability, highest level of math course completed, and score on a five-question numeracy quiz based on banks and oldfield (2007). we survey students and find that perceived return and time cost explain most of the variation in the demand for personal finance education while in college. these results are robust to alternative models and empirically support the theory in jappelli and padula (2013). however, student characteristics have varying influences on demand for specific topics within personal finance. for example, financial independence is positively and significantly related to all topics except credit cards. students without student loans are relatively uninterested in learning about student loans. self-reported math ability positively influences the demand for learning about personal budgeting and topics within the “other” category. gender has independent and robust explanatory power for interest in all personal finance topics except the “other” category. however, the significance of gender is reduced when math ability is included in regressions. on average, sample females self-report lower math ability and achieve lower scores on the numeracy quiz compared to males. sample females are more likely to work part time, have student loans, and are significantly more interested 352 c. harrington, w. smith / financial services review 25 (2016) 351–372 in personal finance education. our gender results conflict with the findings in chen and volpe (2002) and beierlein and neverett (2013). this study contributes to understanding the demand for personal finance education and is conducted in the spirit of increasing college student participation. we use a theory-based approach to investigate student interest in personal finance education and provide an empirical test of the investment function in jappelli and padula (2013). we also add to the literature on the influence of gender and numerical ability on interest in financial education. our study is similar to chen and volpe (2002) and beierlein and neverett (2013). chen and volpe collect gender, student income, and other attributes to examine college student personal financial knowledge and perceptions of the importance of financial knowledge. beierlein and neverett examine characteristics of students who enroll in a personal finance course including gender, race, age, standardized test scores, major, and other attributes. our study attempts to discover factors that drive interest in specific personal finance topics that are likely to benefit college students. many campuses are implementing financial literacy programs that include workshops and peer mentoring. our study supports the idea of offering opportunities for students to learn about individual personal finance topics through various delivery methods in addition to a more comprehensive personal finance course. the findings in our study may help to target specific types of students and increase participation in financial literacy programs and services. we acknowledge the limitations of this study. we sample from a single, private university largely attended by 18to 22-year-old students, and our results may not apply to a more diverse student population. we also characterize demand as stated interest in personal finance education. however, utilization of personal finance education resources may be a more precise measure of demand. some measures of variables used to test demand are compromises rather than ideals. for example, we include a single measure of patience in consumption. this attribute may be better measured by creating an index from responses related to patience in consumption. however, we opt for a single question about patience in consumption for survey brevity. we also do not include parental income as an explanatory variable because of concern about reporting accuracy. finally, the inclusion of math ability as an explanatory variable for interest in personal finance education ignores the possibility that the subject may be presented in a way that minimizes math. the article continues as follows: section 2 discusses the literature related to college student demand for personal finance education and states hypotheses. section 3 describes the survey used to collect data for hypotheses tests. the sample is described in section 4. in section 5, we present the empirical methods used to test the hypotheses. empirical results are discussed in section 6, and section 7 concludes. 2. literature review and hypotheses jappelli and padula (2013) present an intertemporal consumption model in which an individual chooses retirement savings and financial literacy investment given an initial stock of financial literacy and other parameters. initial financial literacy is characterized as financial knowledge before entering the labor market that jappelli and padula suggest is 353c. harrington, w. smith / financial services review 25 (2016) 351–372 related to schooling and parental background. financial literacy is treated as human capital and is assumed to increase asset returns at a decreasing rate. consumers may choose to purchase additional financial literacy at a cost of money, time, or effort. financial literacy and savings are jointly determined and positively correlated over the life cycle. optimal investment in financial literacy is a function of income, patience in consumption, monetary and time/effort cost, the initial stock of financial literacy, and the return on financial literacy. investment in financial literacy increases in income, patience in consumption, and the return on financial literacy and decreases in the initial stock of financial knowledge and cost. the generosity of a financial safety net (such as social security) lowers the need to accumulate wealth and invest in financial literacy. the predictions in jappelli and padula (2013) form the basis of our hypotheses related to student demand for financial education while in college. below, we relate each variable in the model of financial literacy investment—income, patience in consumption, cost, the initial stock of financial literacy, return on financial literacy, and a safety net—to proposed measures that may explain college student demand for personal finance education and state empirical hypotheses. 2.1. income income is a conventional determinant of demand and is predicted to positively influence investment in financial literacy in jappelli and padula (2013). sources of income for college students may include labor income, gifts, savings, or scholarships. however, some students may trade off employment for studying and work part-time (if at all). an alternative view of income is to classify students as working or not. the timing of full-time labor force entry may influence the choice to invest in financial education while in school. for example, most college seniors expect to earn full-time income within one year, whereas full-time labor force entry is not imminent for most freshmen. other things constant, year in school may approximate the role of income in the decision to invest in financial literacy. we hypothesize that income (alternatively measured as current income, employment status, and year in school) is positively and independently associated with interest in personal finance education while in college. 2.2. patience in consumption jappelli and padula (2013) predict that patience in consumption drives investment in financial literacy. their empirical results support this prediction where more patient consumers choose to invest in financial literacy and accumulate relatively more wealth. meier and sprenger (2013) also support this notion in a study of low-income individuals who are offered a free credit counseling session as a complementary service to the volunteer income tax assistance (vita) program. empirical results suggest that vita customers who opt for credit counseling are more patient in consumption. based on the theory in jappelli and padula and prior empirical results, we predict that patience in consumption will positively influence student interest in personal finance education while in college. 354 c. harrington, w. smith / financial services review 25 (2016) 351–372 2.3. cost of financial education jappelli and padula (2013) suggest that the cost of investing in financial literacy is related to money, time, or effort. for college students, we assume that the marginal monetary cost is relatively low if a personal finance course is taken as a nonoverload elective or the student uses a campus-provided personal finance resource (counseling, workshop, web site, etc.). however, acquiring financial education by taking a personal finance course, attending a free campus counseling session, workshop, or using online resources requires time and effort. because the monetary cost is likely low for college students, we focus on time and effort costs of acquiring financial literacy while in college. we predict that higher perceived time and effort costs will reduce the demand for personal finance education. 2.4. initial stock of financial literacy the model in jappelli and padula (2013) predicts that a higher initial stock of financial literacy reduces investment in financial literacy. the initial stock of financial literacy may result from prior personal finance education while in high school, from parents, or other sources. while many u.s. college students are exposed to formal training in personal finance in grades k–12, the state-mandated depth and quantity of the exposure may vary greatly. according to the 2014 survey of the states conducted by the council for economic education, 43 states include personal finance in the k–12 education standards. of these 43 states, only 19 require a stand-alone personal finance course to be offered in high school, and 17 of the 19 states require the course as part of the curriculum. states that require personal finance in high school change from year to year. for example, new york drops and florida adds the requirement between 2011 and 2014 (council for economic education, 2014). the uneven state mandates for personal finance education in grades k–12 may endow the typical high school graduate with a small stock of financial literacy (e.g., chen and volpe, 1998; cummins, haskell, and jenkins, 2009). mandell and klein (2009) find no significant difference in the spending and saving habits of individuals who receive personal finance education while in high school compared with a control sample. in chen and volpe (2002), most students do not recognize personal finance education in high school as a source of financial education and instead indicate parents or prior financial mistakes as the leading sources of financial education. however, prior financial education may increase interest in acquiring additional financial education (goetz, cude, nielsen, chatterjee, and mimura, 2011). experience with debt may contribute to the initial stock of financial literacy. students who are responsible for paying credit card bills or have other personal loans may have relatively more financial knowledge gained through experience with debt products. a student with accumulated debt may be interested in financial education as a way to improve debt management. lyons (2004) finds that students who are delinquent on credit card payments have the highest demand for financial education. however, meier and sprenger (2013) do not find that prior financial experience and knowledge influence demand for additional financial education in the form of credit counseling. 355c. harrington, w. smith / financial services review 25 (2016) 351–372 in summary, prior studies suggest two measures of the initial stock of financial literacy: prior financial education and debt experience. however, empirical results for these measures do not clearly support the jappelli-padula model prediction that the initial stock of financial literacy reduces the incentive to invest in financial literacy. therefore, the influence of prior financial education and debt experience on interest in investing in personal finance education while in college is ambiguous. 2.5. numerical ability prior studies suggest that numerical ability may summarize patience in consumption and time or effort cost. for example, agarwal and mazumder (2013) find that patience in consumption is strongly correlated with math ability. students with low actual or perceived numerical ability are generally turned off by subjects thought to involve math (e.g., bansak and starr, 2010). however, beierlein and neverett (2013) find that math sat scores do not influence college student enrollment in an elective personal finance course. numerical ability is also associated with the initial stock of financial literacy in prior studies. jappelli and padula (2013) find that math and reading skills at the age of 10 are strong predictors of current financial literacy. results from the 2012 oecd programme for international student assessment (pisa) show a strong positive correlation between financial literacy and math and reading skills for 15-year-old students (oecd, 2014). cole et al. (2016) present large-sample evidence that math ability and not exposure to personal finance education in high school matters to financial outcomes. the authors compare financial outcomes for preand post-mandated changes in financial education and math education programs in high schools. state mandated personal finance courses do not influence financial outcomes measured as the likelihood of having investments, level of investment income, credit score, credit card delinquency, and probability of bankruptcy or foreclosure. state requirements that high school students take additional math courses are associated with increased asset accumulation, reduced credit card delinquency, and reduced probability of foreclosure. gerardi, goette, and meier (2013) find a strong inverse relationship between numerical ability and mortgage default, independent of cognitive ability and financial literacy. based on the theory in jappelli and padula (2013), demand for financial education is expected to increase in patience in consumption and decrease in the cost of acquiring the education and the initial stock of financial literacy. if numerical ability suffices for these three determinants of demand, then its influence is a priori ambiguous. further, the mixed empirical results do not suggest a distinct influence of numerical ability on the demand for financial education. therefore, we do not make a prediction about the influence of numerical ability on student interest in personal finance education. 2.6. return on financial literacy the return on financial literacy positively influences investment in financial literacy in jappelli and padula (2013). we approximate the return on financial literacy by measuring college students’ perceptions of the return on financial education. chen and volpe (2002) 356 c. harrington, w. smith / financial services review 25 (2016) 351–372 support the idea of a perceived return by asking students to rate the importance of a menu of college courses. the majority of students view a course in personal finance as either very or somewhat important to improving the quality of their lives. we predict that a student’s perception of the benefits of financial literacy is positively associated with the demand for personal finance education. 2.7. financial safety net jappelli and padula (2013) predict and find that the generosity of a financial safety net (e.g., social security) reduces both investment in financial literacy and wealth accumulation. a parent or other source of nonlabor income may provide a financial cushion, causing students to effectively defer financial responsibility. therefore, a student who is relatively financially dependent may not view financial education as relevant. prior studies do not suggest a consistent relationship between financial independence and interest in personal finance education (goetz et al., 2011; lyons, 2004). because the empirical evidence is mixed, we defer to the jappelli-padula model and predict that the incentive to invest in financial literacy while in school is positively associated with the student’s degree of financial independence. in summary, we predict that college student willingness to invest in financial education while in school is positively related to income, patience in consumption, perceived return on financial education, financial independence, and inversely related to time cost. the influence of the initial stock of financial literacy and numerical ability on interest in acquiring financial education is ambiguous. these variables are constructed from survey results, described next. 3. survey 3.1. survey questions we survey college students at a private, midsize university to measure demand for personal finance education as a function of the variables associated with optimal investment in financial literacy from jappelli and padula (2013) and related studies. demand is measured by asking students to use a 7-point likert scale (1 � strongly disagree, 7 � strongly agree) to rate interest in learning about all personal finance topics (collectively) while in college. we also measure interest in learning about specific topics within personal finance that may appeal to college students including budgeting, credit cards, student loans, and a group of “other” topics (leasing vs. buying, retirement planning, etc.). the individual topics are chosen from the introductory personal finance textbook by kapoor, dlabay, and hughes (2013). alternative measures of income include annual income, employment status, and year in school. the survey asks students to indicate annual income with a 5-category range from $10,000 or less to $25,000 and over. students are asked to identify employment status (full-time, part-time, or not working) and year in school (freshman, sophomore, etc.). 357c. harrington, w. smith / financial services review 25 (2016) 351–372 patience in consumption is measured by an implied discount rate following agarwal and mazumder (2013) and meier and sprenger (2013). our survey includes the open-ended question, “you win a prize. you have the choice to either receive $100 today or to wait 1 year and receive a larger prize. what is the smallest prize amount that would convince you to wait 1 year to collect your winnings?” this question is a simplification of the 2006 national longitudinal survey of youth (nlsy) survey question (bureau of labor statistics, 2006). the nlsy question wording is complex and uses a larger dollar amount ($1,000). agarwal and mazumder (2013, online appendix, p. 1) find that “many respondents provided answers that were clearly unreasonable, with absurdly high implied discount rates.” the discount rates may reflect the large dollar amount used in the question, math skills, or the ability to correctly read and understand the question. time and effort cost is measured by asking students to identify the number of hours per week they are willing to spend learning about personal finance with a scale of 0, 0.5, 1, 2, on up to 5-plus hours. lower response values indicate a higher time or effort cost of investing in personal finance education. the initial stock of financial literacy has two alternative measures: prior financial education and having other debt excluding student loans (e.g., credit card or car loan). prior financial education is measured by the yes/no response to a question about personal finance education in high school, the military, or the workplace. we focus on nonstudent loan debt as an indicator of the initial stock of financial literacy because student loan payments are deferred. having nonstudent loan debt is the yes/no response to the question, “i have other, nonstudent loan debt (e.g., credit card, car loan, etc.).” we recognize that many students may not have a credit card because of current credit card laws that may prevent students under the age of 21 from obtaining a credit card without a cosigner. perceived return on financial literacy is measured by the response to the question, “learning about personal finance topics while in college will improve my ability to make good financial decisions in the future.” the responses follow a 7-point likert scale (1 � strongly disagree, 7 � strongly agree). the degree of financial independence is measured by asking students to identify the percentage of self-paid living expenses using the examples of food, housing, and so forth. the survey question has five response categories (0%, �50%, 50%, �50%, and 100%) ranked as 0% � 1 to 100% � 5. the survey includes three measures of numerical ability for robustness: self-perception of math ability, self-reported highest level of completed math course, and a rank based on the number of correct responses to a 5-question quiz modeled after banks and oldfield (2007). for the first measure, each student rates her/his mathematical ability compared with the average peer following the 5-point rating scheme in bansak and starr (2010) (bottom 10%, below average, average, above average, and top 10%). bansak and starr find that student perception of math ability closely corresponds to self-reported college entrance exam scores. the second measure is the self-reported, highest-level math course completed. cole et al., (2016) find that more math education predicts better financial outcomes. we provide a checklist of typical college math courses and a write-in response for other math courses completed including advanced placement courses. 358 c. harrington, w. smith / financial services review 25 (2016) 351–372 the five questions from banks and oldfield (2007) are open-ended and students are not permitted to use a calculator to answer the questions. because we use paper surveys and cannot control the order in which the questions are answered, we omit the first question in banks and oldfield that is offered only in the event that a survey participant incorrectly answers the subsequent three questions. survey questions are ordered by difficulty and labeled q1-q5. individual numerical ability is classified into four groups based on banks and oldfield. group 1 displays the lowest numerical ability and either incorrectly answers q1-q3, or correctly answers q1 with incorrect answers for q2-q4. group 2 has at least one correct answer to q2-q4. group 3 correctly answers q1-q4 with q5 incorrect. group 4 correctly answers q1-q5. to analyze interest in student loans and credit cards, we ask if students have a student loan (yes/no) and a credit card (yes/no). having a student loan may drive interest in learning about student loans. having a credit card in combination with financial responsibility for paying the credit card bill may be helpful to understand student interest in learning about credit cards. financial responsibility for a credit card is measured by asking students to identify the percentage of the credit card bill paid by themselves (0%, �50%, 50%, �50%, and 100%). we also collect gender. gender may capture inherent interest in personal finance. for example, beierlein and neverett (2013) find that males are more likely to enroll in a personal finance course. chen and volpe (2002) find gender differences among college students related to the perception of personal finance, where male students rank personal finance courses as relatively more important. 3.2. survey protocol this research involving human subjects is approved by the university’s institutional review board. the survey is offered to undergraduate students across the university during spring 2015 and is administered in paper form by the authors or other faculty members on their behalf to achieve a fair representation across majors, year in school, and gender. the survey does not ask for personally identifying information such as name, address, or student identification number. participants are asked to sign and date an informed consent form that describes the purpose of the research, its benefits and risks, and that participation is voluntary. each completed survey is assigned a number that is recorded on both the informed consent form and the survey form. the anonymity of the respondents is maintained by separating the informed consent form from the completed survey. survey responses for each student are recorded using only the assigned number as the observation identifier. 4. sample description in total, 546 students have completed the survey. three participants are dropped from the sample either because the participant is under 18 years old or a graduate student. another 47 participants do not fully complete the survey and are dropped from the sample. the final sample contains 496 observations and represents around 7% of undergraduate students across 359c. harrington, w. smith / financial services review 25 (2016) 351–372 the four colleges of the university. although not tabulated, the sample proportions from each college are similar to the percentage of undergraduates enrolled in each college. table 1 displays the mean interest in personal finance education measured on five dimensions (learning about budgeting, credit cards, student loans, other topics, and all topics) for each sample attribute (gender, income, etc.). survey participants are 56% female and 44% male. on average, female students have relatively more interest in learning about student loans while in college (t statistic is 2.72), but average demand for learning about budgeting, credit cards, other, and all topics is not statistically different between female and male students. mean demand by income range is reported in panel a of table 1. because most students (80.6%) report income of $10,000 or less, we redefine income categories as income �$10,000 and income �$10,000. the mean interest in personal finance topics is not statistically different between the two income groups. regarding employment status, 47.2% of students report working part time, 2.8% work full time, and 50% are not working. students who work part-time are significantly more interested in learning about student loans compared to students who are not working (t statistic is 5.46). all other differences in means are not significant. the final income measure is level in college. the sample consists of 25.4% freshmen, 24% sophomores, 30% juniors, and 20.6% seniors. no college level is statistically more interested in learning about personal finance. the survey response for patience in consumption is open ended and therefore continuous. responses for the least amount of money students would accept to wait one year to receive $100 range from $0 to $1,000,000. implied discount rates range from �100% to 999,900% with a mean of 13,891% and median of 400%. these results are similar to agarwal and mazumder (2013, online appendix) who also find high discount rates. we dichotomize the results as 0–1 by classifying responses as “more patience” (�1) if the discount rate is below the median and “less patience” otherwise. mean interest in personal finance education by patience in consumption is displayed in panel a of table 1. patience in consumption does not matter to interest in personal finance education with the exception of learning about topics in the other category. students classified as more patient are significantly more interested in learning about leasing v. buying, and so forth (t statistic is 2.90). the time and effort cost survey question contains seven possible responses for the hours per week that students are willing spend learning about personal finance topics. in panel a of table 1, responses are grouped into 0–1 hour, 2–3 hour, and 4–5� hours. students willing to spend the most time have the highest interest in learning about all personal finance topics except for student loans. these results are supported by t tests between the 0–1 and 4–5� hours groups. the initial stock of financial literacy is approximated by the student having nonstudent loan debt and alternatively as having prior financial education. in panel a of table 1, 24% of students have nonstudent loan debt. on average, these students are significantly more interested in learning about all personal finance topics but with no statistical difference for learning about specific topics. only 27.8% of sampled students report having prior financial education, but this exposure does not result in a statistically different interest in learning about personal finance. 360 c. harrington, w. smith / financial services review 25 (2016) 351–372 table 1 mean demand for personal finance education while in college by student attribute panel a n � 496 n/n% dbudget dcard dsloan dother dall female 56.0% 5.69 5.25 4.97 5.67 5.72 male 44.0% 5.61 5.00 4.52 5.83 5.68 t test (female-male) 0.68 1.95 2.72 �1.32 0.28 income � $10,000 19.4% 5.73 5.21 4.67 5.78 5.82 income � $10,000 80.6% 5.64 5.13 4.80 5.73 5.67 t test (higher-lower) 0.59 0.48 �0.64 0.33 0.95 work part-time 47.2% 5.71 5.23 5.24 5.85 5.79 work full time 2.8% 5.71 5.57 4.71 5.86 6.07 not working 50.0% 5.61 5.04 4.35 5.63 5.60 t test (part time-not) 0.86 1.53 5.46 1.79 1.67 freshman 25.4% 5.65 5.02 4.63 5.63 5.61 sophomore 24.0% 5.55 4.95 4.92 5.66 5.64 junior 30.0% 5.70 5.26 4.73 5.85 5.70 senior 20.6% 5.74 5.34 4.85 5.80 5.89 t test (freshman-senior) �0.53 �1.77 �0.94 �1.03 �1.72 more patience 49.8% 5.70 5.19 4.81 5.91 5.78 less patience 50.2% 5.61 5.10 4.75 5.57 5.62 t test (more-less) 0.85 0.75 0.36 2.90 1.39 time spent 0–1 hour 47.8% 5.33 4.81 4.55 5.38 5.32 time spent 2–3 hour 40.5% 5.85 5.40 5.01 5.95 5.93 time spent 4–5� hours 11.7% 6.33 5.64 4.88 6.45 6.48 t test (most-least) 5.33 3.90 1.18 5.32 6.23 has nonstudent loan debt 24.0% 5.62 5.31 5.03 5.90 5.94 no nonstudent loan debt 76.0% 5.67 5.09 4.70 5.69 5.63 t test (has-no) �0.34 1.46 1.70 1.52 2.37 has prior financial education 27.8% 5.64 5.02 4.70 5.64 5.70 no prior financial education 72.2% 5.66 5.19 4.80 5.77 5.70 t test (has-has not) �0.22 �1.21 �0.55 �0.97 �0.06 low return (1–5) 14.5% 4.39 4.26 3.88 4.32 4.32 high return (6–7) 85.5% 5.87 5.29 4.93 5.98 5.94 t test (high-low) 7.95 5.35 4.66 8.26 8.55 return � 6 36.7% 5.47 4.86 4.58 5.45 5.45 return � 7 48.8% 6.17 5.62 5.19 6.38 6.30 t test (7–6) 7.34 6.18 3.54 8.91 8.83 living expenses � 0% 28.8% 5.53 5.15 4.36 5.68 5.57 living expenses � 50% 35.1% 5.53 5.14 4.90 5.64 5.70 living expenses � 50% 12.7% 5.49 4.92 4.60 5.51 5.44 living expenses � 50% 12.3% 6.20 5.28 5.28 6.08 6.05 living expenses � 100% 11.1% 5.98 5.24 5.13 6.09 5.96 t test (100% to 0%) 2.03 0.38 2.51 1.99 1.82 panel b n � 496 n/n% dbudget dcard dsloan dother dall has credit card 59.3% 5.67 5.15 4.73 5.79 5.75 no credit card 40.7% 5.64 5.14 4.84 5.66 5.63 t test (has-has not) 0.28 0.06 �0.61 1.03 0.97 pays credit card � 0% 29.2% 5.52 4.97 4.40 5.64 5.57 pays credit card � 50% 9.7% 5.58 5.38 4.81 5.90 5.94 pays credit card � 50% 4.2% 5.29 4.48 3.90 5.62 5.33 pays credit card � 50% 3.6% 5.56 5.17 4.56 5.72 5.50 pays credit card � 100% 19.8% 5.82 5.40 5.22 5.93 5.93 361c. harrington, w. smith / financial services review 25 (2016) 351–372 in panel a of table 1, 85.5% of students sampled either agree or strongly agree that learning about personal finance topics while in college will improve their ability to make good financial decisions in the future. therefore, we display descriptive statistics for this variable as high perceived return on financial literacy corresponding to a response of agree or strongly agree (likert scale response of 6 or 7), and low perceived return for all other responses. students with a high perceived return have higher interest in every personal finance topic. we test for differences in interest among the agree and strongly agree responses. students who strongly agree have a statistically significant higher interest compared with students who agree. panel a of table 1 shows mean responses by percentage of living expenses paid by the student as an approximation of a financial safety net. around 29% of students pay 0% of living expenses and 11% of surveyed students fully support themselves. students who fully pay their living expenses are significantly more interested in learning about budgeting, student loans, and topics in the other category. statistics on having credit cards, paying credit card bills, and having student loans are in panel b of table 1. around 59% of sampled students report having a credit card, but only 19.8% report paying 100% of their credit card bills. having a credit card does not matter to interest in personal finance education. students responsible for paying 100% of their credit card bills are statistically more interested in learning about credit cards, student loans, and all personal finance topics. of the students surveyed, 56% have student loans. these students are significantly more interested in learning about student loans while in college. table 1 (continued) n � 496 n/n% dbudget dcard dsloan dother dall pays credit card n/a 33.5% 5.76 5.16 4.96 5.68 5.68 t test (100% to 0%) 1.66 2.29 3.41 1.60 2.06 has student loan 56.0% 5.70 5.19 5.62 5.70 5.78 no student loan 44.0% 5.60 5.09 3.71 5.79 5.60 t test (has-has not) 0.89 0.79 13.04 �0.76 1.57 ability � average 55.6% 5.54 5.08 4.86 5.57 5.65 ability � average 44.4% 5.81 5.23 4.67 5.95 5.77 t test (lower-higher) �2.46 �1.20 1.12 �3.27 �1.04 highest math is algebra 65.9% 5.61 5.07 4.77 5.61 5.64 highest math � calculus 34.1% 5.74 5.29 4.79 5.98 5.82 t test (algebra-calculus) �1.07 �1.69 �0.15 �2.96 �1.45 math quiz group 1 2.6% 5.38 4.38 4.92 5.69 5.31 math quiz group 2 40.7% 5.70 5.25 4.65 5.60 5.67 math quiz group 3 33.9% 5.60 5.20 5.10 5.79 5.79 math quiz group 4 22.8% 5.71 4.96 4.50 5.92 5.67 t test (group 2-group 4) �0.07 1.84 0.65 �2.03 0.00 notes: dbudget, dcard, dsloan, dother, and dall represent demand for learning about personal budgeting, credit cards, student loans, other personal finance topics, and all personal finance topics, respectively. time spent is the number of hours per week the student is willing commit to personal finance education. return is the perceived return from personal finance education. living expenses is the percentage of living expenses paid by the student. ability is self-reported math ability. highest math is the highest level of math course completed. group is the rank based on a 5-question numeracy quiz score. 362 c. harrington, w. smith / financial services review 25 (2016) 351–372 panel b of table 1 shows statistics for the measures of numerical ability. most students report their ability as average (48%) or above average (38.5%) with very few students reporting ability as below average (6.7%), bottom 10% (1%), or top 10% (5.8%). therefore, we display self-reported mathematical ability as average or below and above average in panel b of table 1. self-reported highest level of completed math course has a similar distribution. thirty students (6.1%) have yet to complete a college level math course. the majority of these students are freshmen (not tabulated). for students who have completed college math courses, the highest level is lower-level algebra for 12.5% and higher-level algebra for 53.4% of the sample. the first calculus course is the highest level for 29.2%, and 4.8% of students have completed the second calculus course or higher. given the smaller proportions of the sample at the extremes of this distribution, math level is dichotomized as the highest level is any algebra course (65.9%) and highest level is the first calculus course and above (34.1%). based on the results from the 5-question math quiz, 2.6% of students are ranked as group 1, 40.7% as group 2, 33.9% as group 3, and 22.8% as group 4. regardless of measurement, students with higher numerical ability are statistically more interested in other personal finance topics. only students who self-report high math ability are relatively more interested in learning about personal budgeting. table 2 presents correlations between the measures of numerical ability, patience in consumption, time cost, and prior financial education. the correlation coefficients for self-reported ability and highest math course completed, self-reported ability and group, and highest math course completed and group are 0.43, 0.33, and 0.19, respectively (p values � 0.0001). the three measures of math ability share weaker correlations with the dichotomized expression of patience in consumption (1 � more patient). the self-reported highest math course completed is most strongly correlated with patience in consumption (p value � 0.0004) and the group ranking based on the math quiz score is least correlated with patience table 2 correlations between numeracy measures ability math level group patience cost fined ability 1.00 0.43 0.33 0.13 0.04 �0.06 (�.0001) (�.0001) (0.0046) (0.3716) (0.1928) math level 1.00 0.19 0.16 0.08 �0.01 (�.0001) (0.0004) (0.0814) (0.8056) group 1.00 0.10 0.01 �0.02 (0.0211) (0.8318) (0.7177) patience 1.00 0.11 �0.01 (0.0178) (0.7983) cost 1.00 �0.02 (0.6480) fined 1.00 note: pearson correlation coefficients (p values) are presented in the table. ability is self-reported math ability measured on a 5-point scale (1 � bottom 10%, 5 � top 10%). math level is the highest level of math course completed. group is the group based on the score on a numeracy quiz (1 � lowest, 4 � highest). patience is a 0–1 indicator of the rate of time preference above (�0) or below the median (�1) based on responses to the corresponding survey question. cost is an indicator of the hours the student is willing to commit to personal finance education while in college based on a 7-point scale (1 � 0, 7 � 5�). fined is a 0–1 indicator of prior financial education (1 � no). 363c. harrington, w. smith / financial services review 25 (2016) 351–372 (p value � 0.0211). the numerical ability measures, time cost, and prior financial education are not correlated with each other. in untabulated results, the alternate measures of income are significantly correlated. income is positively associated with working and college level. working is also significantly positively correlated with having a credit card and nonstudent loan debt as well as paying a higher percentage of living expenses. the correlations also indicate gender differences with respect to income, time cost, prior financial education, and math ability. females are more likely to work part time, have lower income, and have student loans. males are willing to spend more time investing in personal finance education and have more prior financial education. males also have higher math ability as measured by self-reported ability and the score on the math quiz. 5. empirical method following jappelli and padula (2013), we model college student demand (d) for personal finance education as a function of income, patience in consumption (patience), cost of acquiring financial education (cost), the initial stock of financial literacy (stock), perceived return on financial education (return), and the existence of a financial safety net (net). gender (gen) is included as an experimental variable. the regression function is: dj � b0 � b1 incomej � b2 patiencej � b3 costj � b4 stockj � b5 returnj � b6 netj � b7 genj � ej (1) we first examine student demand for learning about all personal finance topics (collectively) while in college as the response on a 7-point scale (1 � strongly disagree, 7 � strongly agree). income is measured as the student’s stated income range because this is the most direct measure of income from the survey. employment status (part-time, full-time, and not working) and college level are used for robustness. patience in consumption is the 0–1 variable created to represent discount rates above (�0) and below the median (�1), where 1 represents relatively more patience in consumption. cost is measured as the time the student is willing to spend to learn about personal finance topics, where a low number represents a high time cost. the initial stock of financial literacy is an indicator of having nonstudent loan debt (�1, 0 otherwise). for robustness, prior financial education is the alternative measure of the initial stock of financial literacy. because the majority of students in the sample do not have prior financial education, this variable equals one if the student does not have prior financial education. return is the response to the perceived return to studying personal finance while in college, where higher responses indicate greater perceived returns. the financial safety net is measured as a categorical variable corresponding to the percentage of living expenses paid by the student and represents the degree of financial responsibility. gender is equal to 1 if the student is female. as defined in eq. (1), the demand for learning about personal finance topics is expected to increase in income, patience in consumption, time the student is willing to spend, perceived return, and the percentage of living expenses paid by the student. the relationship 364 c. harrington, w. smith / financial services review 25 (2016) 351–372 between demand and the initial stock of financial literacy is a priori ambiguous. prior studies indicate that gender is correlated with interest in financial education. however, we do not have a theoretical basis for expecting gender to influence demand in any way and therefore do not make a prediction for this variable. numerical ability is hypothesized to be correlated with patience in consumption, cost of acquiring financial literacy, and the initial stock of financial literacy. we test this hypothesis by substituting a measure of math ability for these variables. however, the expected influence of math ability is a priori ambiguous given the predicted signs for patience, cost, and stock. the regression function is: dj � b0 � b1 incomej � b2 returnj � b3 netj � b4 genj � b5 mathj � ej (2) because the dependent variables are categorical in eqs. (1) and (2), coefficient estimates are calculated via logistic regressions of demand on the hypothesized explanatory variables. estimates are with respect to the highest response measure (strongly agree), capturing the influence of the explanatory variables on being highly interested in personal finance education while in college. 6. results 6.1. demand for all personal finance topics regression results for interest in learning about all personal finance topics are in table 3. the results from estimating eq. (1) are in model 1 with a pseudo r2 of 39%. income and patience in consumption are statistically insignificant. time cost is highly significant with the predicted sign (p value � 0.0001), supporting the hypothesis that students with a higher perceived time cost will have lower demand for personal finance education. the initial stock of financial literacy (measured by having nonstudent loan debt) is positively related but marginally significant to demand (p value � 0.0231). the coefficient estimate for return is positive and highly significant (p value � 0.0001), supporting the idea that students with a higher perceived return from financial education demand more financial education. financial independence is positive and significant as predicted (p value � 0.0048). finally, gender contributes a small amount of explanatory power to demand (p value � 0.0129), where female students are marginally more interested in all personal finance topics while in college. eq. (1) is estimated using alternative explanatory variable definitions where available (not tabulated). income is alternatively defined as employment status and level in college. neither alternative income measure is significant to the demand for all personal finance topics while in college. the income measures are correlated with the percentage of living expenses paid by the student. dropping an income measure (however defined) from the regression does not change the statistical significance of the financial independence variable. prior financial education is an alternate measure of the initial stock of financial literacy but is not significant to the demand for all personal finance topics. we also drop the dichotomous definition of patience in consumption and use its continuous version but this variable remains insignificant. 365c. harrington, w. smith / financial services review 25 (2016) 351–372 sample results for the dependent variable in eq. (1) have relatively few responses for the strongly disagree and disagree categories. these responses are collapsed into a single category and eq. (1) is re-estimated as a robustness check. the results, represented as model 2 in table 3, suggest that the estimates are not sensitive to the inclusion of relatively few responses for the two categories. model 3 in table 3 gives the regression results for eq. (2). demand for all personal finance topics is regressed on each alternative measure of math ability (self-reported ability, highest level of math course completed, and the score from the numeracy quiz). the results using self-report math ability are shown in the table. none of the math ability measures is significant to demand for all personal finance topics. further, the results from eq. (1) are not sensitive to the substitution of math ability for patience in consumption, cost, and the initial stock of financial literacy except for gender. gender in eq. (2) is insignificant to the demand for all personal finance topics. 6.2. demand for topics within personal finance education results from investigating demand for specific topics within personal finance are in tables 4 through 7. demand for learning about personal budgeting is regressed on the variables as defined in eq. (1) and the results are shown in model 1 in table 4. the pseudo r2 is 35%. cost, perceived return, and financial independence are all significant with the expected signs (p values range from �0.0001 to 0.0002). gender is also significant to interest in learning table 3 demand for learning about all personal finance topics while in college (1) (2) (3) estimate p value estimate p value estimate p value intercept �10.47 �0.0001 �10.63 �0.0001 �9.83 �0.0001 intercept �8.45 �0.0001 �8.60 �0.0001 �7.97 �0.0001 intercept �6.95 �0.0001 �7.10 �0.0001 �6.57 �0.0001 intercept �5.22 �0.0001 �5.39 �0.0001 �4.90 �0.0001 intercept �4.66 �0.0001 �4.34 �0.0001 intercept �4.05 �0.0001 �3.70 �0.0001 income �0.06 0.6032 �0.05 0.6540 0.05 0.6518 patience 0.08 0.6408 0.09 0.6182 cost 0.59 �0.0001 0.60 �0.0001 stock 0.47 0.0231 0.48 0.0208 return 1.18 �0.0001 1.20 �0.0001 1.31 �0.0001 net 0.19 0.0048 0.19 0.0055 0.17 0.0093 gender 0.44 0.0129 0.43 0.0155 0.16 0.3600 math �0.02 0.8420 psuedo r2 0.39 0.40 0.31 note: the dependent variable is interest in learning about all personal finance topics while in college measured on a 7-point scale (1 � strongly disagree, 7 � strongly agree). model (1) is the estimation of eq. (1). model (2) is the estimation of eq. (1) combining the lowest two response categories for the dependent variable. model (3) contains estimates from eq. (2). the results are maximum likelihood estimates with the highest value of the dependent variable [�7 in (1) and (3), �5 in (2)] as the reference. the intercepts reference the lower values for the dependent variable. the p values are for wald �2 statistics. 366 c. harrington, w. smith / financial services review 25 (2016) 351–372 about personal budgeting (p value � 0.0093), where females are relatively more interested than males. income, patience in consumption, and the initial stock of financial literacy are not significant in model 1. the alternate measures of income and the stock of financial literacy are all insignificant when included in model 1 (untabulated). model 2 displays the results from estimating eq. (2). self-reported math ability is positive and marginally significant as a summary measure for patience in consumption, cost, and the initial stock of financial literacy (p value � 0.0510). perceived return remains highly significant but the statistical significance of gender is sensitive to the inclusion of self-reported math ability. the alternate measures of math ability are insignificant in eq. (2). model 3 adds self-reported math ability to eq. (1). cost, perceived return, financial independence, and gender remain highly significant as in model 1, yet math ability has independent explanatory power. table 5 displays estimates for interest in learning about credit cards while in college. the results in model 1 are from estimating eq. (1). the pseudo r2 is 21%. cost, return, and gender are significant with females having relatively more interest in this topic (p values range from �0.0001 to 0.0012). having nonstudent loan debt is positive and marginally significant to interest in learning about credit cards (p value � 0.0614). income range, patience in consumption, and financial independence are insignificant. income range is alternatively replaced with employment status and college level in eq. (1) (not tabulated). college level is marginally significant to interest in learning about credit cards and employment status is not significant. we also replace nonstudent loan debt with prior financial education but this variable remains insignificant. in model 2 college level approximates expected income for estimating the coefficients in eq. (2). college level is significant as are table 4 demand for learning about personal budgeting while in college (1) (2) (3) estimate p value estimate p value estimate p value intercept �10.08 �0.0001 �10.29 �0.0001 �10.97 �0.0001 intercept �8.01 �0.0001 �8.33 �0.0001 �8.89 �0.0001 intercept �6.46 �0.0001 �6.84 �0.0001 �7.33 �0.0001 intercept �4.91 �0.0001 �5.32 �0.0001 �5.77 �0.0001 intercept �4.08 �0.0001 �4.49 �0.0001 �4.93 �0.0001 intercept �3.34 �0.0001 �3.71 �0.0001 �4.18 �0.0001 income 0.04 0.7052 0.07 0.5171 0.04 0.6935 patience 0.08 0.6250 0.05 0.7724 cost 0.46 �0.0001 0.47 �0.0001 stock �0.22 0.2739 �0.26 0.2005 return 1.11 �0.0001 1.19 �0.0001 1.10 �0.0001 net 0.25 0.0002 0.20 0.0018 0.24 0.0003 gender 0.46 0.0093 0.27 0.1141 0.51 0.0039 math 0.22 0.0510 0.27 0.0201 psuedo r2 0.35 0.30 0.36 note: the dependent variable is interest in learning about personal budgeting while in college measured on a 7-point scale (1 � strongly disagree, 7 � strongly agree). model (1) is the estimation of eq. (1). model (2) is the estimation of eq. (2). model (3) contains estimates from eq. (1) also including numerical ability. the results are maximum likelihood estimates with the highest value of the dependent variable (�7) as the reference. the intercepts reference the lower values for the dependent variable. the p values are for wald �2 statistics. 367c. harrington, w. smith / financial services review 25 (2016) 351–372 perceived return and gender (p values range from �0.0001 to 0.0167), but any measure of math ability is insignificant. in model 3 we augment eq. (1) with college level as the proxy for income, include indicators of having a credit card (�1), responsibility for paying a credit card bill (�1), and an interaction term that equals one if the student has and pays a credit card. none of these indicators is significant to interest in learning about credit cards while in college. table 6 shows results for estimating the demand for learning about student loans while in college. the pseudo r2 from estimating eq. (1) in model 1 is 15%. both income range and employment status are positive and significant (p values are 0.0001 and 0.0025, respectively), but college level is not significant when used as an alternate measure of income. time cost is marginally significant (p value � 0.0678) and perceived return, financial independence and gender are highly significant (p values range from �0.0001 to 0.0006) with females relatively more interested in this topic. having nonstudent loan debt and the alternate measure of prior financial education are insignificant to demand for learning about student loans. because few students are employed full time, we collapse employment status into an indicator of working (�1) or not as a measure of income and estimate eq. (2). the results are displayed in model 2. the working indicator is positive and highly significant (p value � 0.0001) but reduces the significance of financial independence and gender. the alternate table 5 demand for learning about credit cards while in college (1) (2) (3) estimate p value estimate p value estimate p value intercept �7.53 �0.0001 �8.13 �0.0001 �7.93 �0.0001 intercept �5.76 �0.0001 �6.41 �0.0001 �6.14 �0.0001 intercept �4.57 �0.0001 �5.27 �0.0001 �4.95 �0.0001 intercept �2.98 �0.0001 �3.71 �0.0001 �3.34 �0.0001 intercept �2.49 �0.0001 �3.22 �0.0001 �2.85 �0.0001 intercept �1.25 0.0471 �1.94 0.0069 �1.60 0.0218 income �0.02 0.8778 0.20 0.0084 0.15 0.0605 patience 0.12 0.4603 0.12 0.4789 cost 0.38 �0.0001 0.37 �0.0001 stock 0.37 0.0614 0.25 0.2127 return 0.72 �0.0001 0.84 �0.0001 0.74 �0.0001 net �0.03 0.6666 �0.05 0.3924 gender 0.55 0.0012 0.40 0.0167 0.56 0.0011 math 0.11 0.3034 has card �0.30 0.4150 pays card �0.01 0.7914 has x pays 0.07 0.2950 psuedo r2 0.21 0.17 0.22 note: the dependent variable is interest in learning about credit cards while in college measured on a 7-point scale (1 � strongly disagree, 7 � strongly agree). model (1) is the estimation of eq. (1). model (2) is the estimation of eq. (2) with college level approximating expected income. model (3) contains estimates from eq. (1) with college level for income, has card as an indicator that the student has a credit card (�1), pays card as the percentage of the credit card bill paid by the student measured on a 5-point scale (1 � 0%, 5 � 100%), and has x pays as an interaction term for has credit card and pays credit card. the results are maximum likelihood estimates with the highest value of the dependent variable (�7) as the reference. the intercepts reference the lower values for the dependent variable. the p values are for wald �2 statistics. 368 c. harrington, w. smith / financial services review 25 (2016) 351–372 math ability measures are not significant. we retain the working indicator for income in model 3 and add an indicator for having a student loan (�1). the student loan indicator is highly significant (p value � 0.0001) and dominates the significance of working and financial independence. results from estimating the demand for learning about other personal finance topics (leasing, retirement planning, etc.) are displayed in table 7. the results from estimating eq. (1) are in model 1 with a pseudo r2 of 36%. cost, perceived return, and financial independence are positive and significant (p values range from �0.0001 to 0.0033). patience in consumption is marginally significant (p value � 0.0275) with the predicted sign. income, nonstudent loan debt as the stock variable, and gender are insignificant. in untabulated results, eq. (1) is estimated with the alternative measures of income and prior financial education as a proxy for the initial stock of financial literacy. prior financial education is positive and marginally significant indicating that students without prior financial education are marginally more interested in learning about other personal finance topics. in model 2, self-reported math ability is positive and marginally significant when this variable replaces cost, patience in consumption, and the stock of financial literacy (p value � 0.0846). the other numeracy variables are insignificant. model 3 is the reconfiguration of eq. (1) with prior financial education approximating the stock variable and self-reported math ability as an additional variable. self-reported math ability is marginally significant (p value � 0.0572) table 6 demand for learning about student loans while in college (1) (2) (3) estimate p value estimate p value estimate p value intercept �5.68 �0.0001 �5.47 �0.0001 �7.86 �0.0001 intercept �4.50 �0.0001 �4.27 �0.0001 �6.45 �0.0001 intercept �3.69 �0.0001 �3.46 �0.0001 �5.43 �0.0001 intercept �2.69 �0.0001 �2.45 0.0002 �4.21 �0.0001 intercept �2.46 �0.0001 �2.23 0.0008 �3.95 �0.0001 intercept �1.27 0.0324 �1.04 0.1164 �2.64 �0.0001 income �0.30 0.0025 0.65 �0.0001 0.23 0.1928 patience 0.06 0.6901 0.04 0.7905 cost 0.13 0.0678 0.11 0.1126 stock 0.31 0.1070 �0.48 0.0161 return 0.54 �0.0001 0.58 �0.0001 0.67 �0.0001 net 0.23 0.0002 0.15 0.0174 0.02 0.7268 gender 0.57 0.0006 0.42 0.0112 0.41 0.0173 math �0.17 0.1244 loan 2.25 �0.0001 psuedo r2 0.15 0.16 0.37 note: the dependent variable is interest in learning about student loans while in college measured on a 7-point scale (1 � strongly disagree, 7 � strongly agree). model (1) is the estimation of eq. (1). model (2) is the estimation of eq. (2) with working as the income variable, an indicator equal to 1 if the student is employed, 0 otherwise. model (3) contains estimates from eq. (1) with working as the income variable and loan as an indicator equal to 1 if the student has a student loan, 0 otherwise. the results are maximum likelihood estimates with the highest value of the dependent variable (�7) as the reference. the intercepts reference the lower values for the dependent variable. the p values are for wald �2 statistics. 369c. harrington, w. smith / financial services review 25 (2016) 351–372 as is initial stock (p value � 0.0262), where the significance of patience in consumption, cost, return, financial independence, and gender is relatively unaffected. prior research indicates that business majors are relatively more knowledgeable about personal finance (e.g., chen and volpe, 1998). regarding interest in personal finance education, beierlein and neverett (2013) find that business majors (as well as human ecology and social science majors) are relatively more likely to complete a personal finance course. to test if being a business student influences demand for personal finance education, we include an indicator variable for business majors in all of the regressions. however, this indicator is only marginally significant to interest in learning about student loans and is insignificant to all other dependent variables. we find that business students are relatively less interested in learning about student loans compared to nonbusiness majors. 7. conclusion this study employs the jappelli and padula (2013) model of investment in financial literacy to examine student interest in personal finance education while in college. although we find some support for this model, our results suggest that students are not uniformly interested in every personal finance topic. in our study, perceived return and time cost are key drivers of interest in personal finance education. student interest in specific personal finance topics is associated with circumtable 7 demand for learning about other personal finance topics while in college (1) (2) (3) estimate p value estimate p value estimate p value intercept �9.86 �0.0001 �10.03 �0.0001 �10.43 �0.0001 intercept �7.98 �0.0001 �8.24 �0.0001 �8.52 �0.0001 intercept �6.86 �0.0001 �7.17 �0.0001 �7.40 �0.0001 intercept �5.26 �0.0001 �5.59 �0.0001 �5.79 �0.0001 intercept �4.53 �0.0001 �4.85 �0.0001 �5.06 �0.0001 intercept �3.53 �0.0001 �3.80 �0.0001 �4.07 �0.0001 income �0.10 0.3684 �0.04 0.6943 �0.06 0.5522 patience 0.38 0.0275 0.34 0.0462 cost 0.41 �0.0001 0.41 �0.0001 stock 0.26 0.2058 0.42 0.0262 return 1.20 �0.0001 1.30 �0.0001 1.19 �0.0001 net 0.20 0.0033 0.17 0.0093 0.20 0.0031 gender 0.04 0.8276 �0.10 0.5676 0.03 0.8451 math 0.20 0.0846 0.22 0.0572 psuedo r2 0.36 0.31 0.37 note: the dependent variable is interest in learning about other personal finance topics while in college measured on a 7-point scale (1 � strongly disagree, 7 � strongly agree). model (1) is the estimation of eq. (1). model (2) is the estimation of eq. (2). model (3) contains estimates from eq. (1) also including numerical ability and the initial stock of financial literacy (stock) as an indicator of prior financial education (0 � yes, 1 � no). the results are maximum likelihood estimates with the highest value of the dependent variable (�7) as the reference. the intercepts reference the lower values for the dependent variable. the p values are for wald �2 statistics. 370 c. harrington, w. smith / financial services review 25 (2016) 351–372 stances or characteristics. for example, female students who pay their living expenses are relatively more interested in learning about budgeting. upper-level students with nonstudent loan debt are relatively more interested in learning about credit cards. students with existing student loans are highly interested in learning more about these loans. while a traditional personal finance course offers a broad base of knowledge, some students may not view all topics as relevant and other students may not have time for a 3-credit-hour course. designing a financial education program as a menu of personal finance learning opportunities allows students to self-select interests and time commitments. to provide students with control over their time commitment, the menu might include multiple education delivery options for each topic such as individual counseling, workshops, web resources and videos, reduced credit hour courses, and a traditional course. individual counseling may appeal to students who prefer private instruction. periodic workshops on specific topics may attract students who prefer brief, face-to-face, hands-on instruction. workshops are also an opportunity to partner with the local business community. for students who prefer internet delivery, the institution could offer a web page containing links to topic-focused articles, videos, and tools such as student loan calculators. articles, videos, and tools provide instruction and application that may better accommodate student schedules. reduced credit-hour courses that cover a related set of topics may appeal to specific students and may not impact tuition. finally, a traditional course may appeal to students who are willing to commit larger amounts of time and wish to learn about a broader range of topics. increasing participation in personal finance education may also involve outreach to specific groups of students. as an example, we find that financial independence influences interest in learning about budgeting. an attention-grabbing message such as, “supporting yourself while in school? budgeting makes it easier to balance school and work!” may attract students with these characteristics. including a link to a short video on budgeting in the message may stimulate interest in attending a workshop or visiting a web site containing budgeting resources (e.g., apps, articles, etc.). because personal finance education is voluntary for most college students, designing campus offerings around student characteristics and preferences may increase participation while in college. future research may examine the effectiveness of this approach in increasing college student financial self-efficacy and literacy. note 1 we intentionally omit investing as a candidate topic. in the authors’ prior experience with surveying students about interest in personal finance education, investing is the most popular choice. our intention with the current survey is to understand interest in subjects that may benefit students during the college experience. acknowledgment the authors acknowledge the university of tampa dana foundation grant and karla borja for helpful comments. 371c. harrington, w. smith / financial services review 25 (2016) 351–372 references agarwal, s., & mazumder, b. (2013). cognitive abilities and household financial decision making. american economic journal: applied economics, 5, 193–207. banks, j., & oldfield, z. (2007). understanding pensions: cognitive function, numerical ability and retirement saving. fiscal studies, 28, 143–170. bansak, c., & starr, m. (2010). gender differences in predispositions towards economics. eastern economic journal, 36, 33–57. beierlein, j. j., & neverett, m. (2013). who takes personal finance? financial services review, 22, 151–171. bernheim, b. d., garrett, d. m., & maki, d. m. (2001). education and saving: the long-term effects of high school financial curriculum mandates. journal of public economics, 80, 435–465. bureau of labor statistics, u.s. department of labor. (2006). national longitudinal survey of youth 1979 cohort, (round 22) (available at: https://www.nlsinfo.org/sites/nlsinfo.org/files/attachments/121210/19.consumption %20impatience%20risk.html). chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. cole, s., paulson, a., & shastry, g. k. (2016). high school curriculum and financial outcomes. journal of human resources, 51, 656–698. council for economic education. (2014). survey of the states: economic and personal finance education in our nation’s schools 2014 (available at: http://www.councilforeconed.org/wp/wp-content/uploads/2014/02/2014survey-of-the-states.pdf). cummins, m., haskell, j. h., & jenkins, s. j. (2009). financial attitudes and spending habits of university freshmen. journal of economics & economic education research, 10, 3–20. gerardi, k., goette, l., & meier, s. (2013). numerical ability predicts mortgage default. proceedings of the national academy of sciences, 110, 11267–11271. doi:10.1073/pnas.1220568110. goetz, j., cude, b. j., nielsen, r. b., chatterjee, s., & mimura, y. (2011). college-based personal finance education: student interest in three delivery methods. journal of financial counseling & planning, 22, 27–42. hastings, j. s., madrian, b. c., & skimmyhorn, w. l. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5, 347–373. jappelli, t., & padula, m. (2013). investment in financial literacy and saving decisions. journal of banking & finance, 37, 2779–2792. kapoor, j. r., dlabay, l. r., & hughes, r. j. (2013). focus on personal finance: an active approach to help you develop successful financial skills (4th ed.). new york, ny: mcgraw-hill. lyons, a. c. (2004). a profile of financially at-risk college students. journal of consumer affairs, 38, 56–80. mandell, l., & klein, l. (2009). the impact of financial literacy education on subsequent financial behavior. journal of financial counseling & planning, 20, 15–24. meier, s., & sprenger, c. d. (2013). discounting financial literacy: time preferences and participation in financial education programs. journal of economic behavior & organization, 95, 159–174. oecd. (2014). pisa 2012 results: students and money: financial literacy skills for the 21st century (vol. vi). paris: pisa, oecd publishing. 372 c. harrington, w. smith / financial services review 25 (2016) 351–372 conflicted advice about portfolio diversification sally shena, john a. turnerb,* aglobal risk institute in financial services, 55 university avenue, suite 1801, toronto, on m5j 2h7, canada bpension policy center, 3713 chesapeake street nw, washington, dc 20016, usa abstract we investigate the validity of the argument by the financial services industry to “roll over your ‘old’ 401(k) plan” because 401(k)-type plans have a limited number of investment options, while individual retirement accounts (iras) have a virtually unlimited number of options. we empirically analyze the diversification of a large 401(k)-type plan with only five basic investment options. financial advisers with a conflict of interest may use strategic complexity to encourage rollovers, recommending complex portfolios to impress naïve clients who have a weak understanding of the concept of diversification, while not weighing the cost of the complex portfolios against any added benefits of diversification. © 2018 academy of financial services. all rights reserved. jel classification: g2; g4 keywords: portfolio diversification; financial advice; conflicts of interest 1. introduction because of pension rollovers, individual retirement accounts (iras) have become the most important source of pension income in the united states. thus, pension rollovers play a key role in the u.s. retirement income system. the argument generally made to support the campaign by the financial services industry to “roll over your ‘old’ 401(k)” is that 401(k)-type plans have a limited number of investment options, while iras have a virtually unlimited number of options. this paper investigates the validity of the advice that better diversification is a reason for rolling over to an ira by empirically analyzing * corresponding author. tel.: �1-202-686-6775. e-mail address: jaturner49@aol.com (j.a. turner) financial services review 27 (2018) 47-81 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. the diversification a 401(k)-type plan with only five basic investment options (and five target date funds composed of those options)—the thrift savings plan for federal government workers. to study the issue of complexity in pension portfolios, we have purposively chosen a pension plan with a small number of investment options. our argument is not that a plan with a relatively small number of investment options is superior to a plan with say 30 options. rather, we address the question of whether a plan with only five basic investment options can be adequately diversified. if that is the case, the much larger number of investment options generally found in 401(k) plans indicates that the argument favoring the diversification benefits of the substantially larger number of investment options available in iras is generally not valid. we argue in this paper that conflicted advisers focus on the aspects that are favorable to the case they are making (e.g., “only five funds”), but do not consider whether pension participants need more choice in funds to improve diversification. in addition, they do not weigh the costs associated with their advice to leave a low-cost plan to obtain more investment options. we argue that it is psychologically less costly for advisers to make a true statement that is incomplete than to make a false statement. in addition, some advisers may simply follow the industry standard argument, without considering its merit. we argue that many participants are susceptible to this argument relating to portfolio complexity because of their naïve understanding of diversification. they think that more investment options are always better, not understanding the characteristics required of new options to improve diversification, and not considering the costs in higher fees of the added options. this paper thus relates to the more general issue of how many diversified mutual funds are needed to form a diversified investment portfolio. for example, do target date funds need more than a dozen different investments in their portfolios, or would a smaller number be better in part because it would involve less costly funds. it can be argued that an investment that is a small share of the portfolio does not materially affect the risk-return characteristics of the portfolio and should not be included if it is a relatively expensive investment in terms of fees. some financial advisers and financial products companies may engage in strategic complexity in their portfolios, providing complex portfolios to impress naïve clients. the remainder of the paper is structured as follows. we first provide background information about roll overs to iras and why we focus on the thrift savings plan with its five basic options. focusing on the tsp provides a test of the hypothesis that pension participants should roll over from their 401(k) plans to obtain greater portfolio diversification. we then review the relevant literature concerning portfolio diversification and pension rollovers. following that, we discuss the investment options available in the thrift savings plan. the main section of the paper follows in which we analyze the effect of adding more investment options and investigate the validity of the advice to roll over from the tsp. last, we offer our conclusions relating to the quality of the advice from some conflicted financial advisers, the nature of the arguments some conflicted financial advisers make, and the susceptibility of some pension participants with low financial literacy to making decisions based on these arguments. 48 s. shen, j.a. turner / financial services review 27 (2018) 47-81 2. rollovers to iras iras are the largest type of pension plan in terms of assets in the united states, having overtaken 401(k) plans and defined benefit plans. rollovers from 401(k)-type plans are the primary source of funding for iras, with relatively few people contributing to iras. iras had an estimated $7.9 trillion in assets at the end of 2016 and represented 31% of total u.s. retirement market assets, compared with 18% two decades earlier. in 2012, $335 billion was rolled over from employer-provided plans to iras (investment company institute, 2016b, 2017). the council of economic advisers (cea 2015), which advises the president on economic policy, surveyed the literature on the quality of financial advice provided in the united states. the cea finds that advice concerning pension rollovers from employer-provided plans to iras, and stemming from conflicts of interest, costs u.s. pension participants $17 billion a year in higher fees and lower rates of return. supporting these conclusions, a study by munnell, aubrey and crawford (2015) finds that iras on average receive net rates of return that are about 1 percentage point less than do employer-provided defined contribution plans, such as 401(k) plans and 403(b) plans, in part because of higher fees. advisers giving bad advice presumably make an argument to their clients as to why their advice is good advice. typically, that argument in this context is that pension participants will have more investment options in iras than in 401(k)-type plans. that argument has become the industry standard for advice. for example, tiaa (2016) indicates an advantage of rolling over to an ira is that a pension participant has “a virtually unlimited array of investments.” similarly, fidelity (2016) indicates an advantage to rolling over is that you have “a wide range of investment options.” according to a survey of persons making pension rollovers, while improved diversification is not the only reason workers give for why they rolled over to an ira, it is the primary reason for 21% and one of the reasons for 61% of those making rollovers (investment company institute, 2016b). this argument supporting roll overs seemingly runs counter to the requirements of u.s. pension law. u.s. pension law (erisa section 404(c)) requires that 401(k) plans that allow participants the opportunity to make their own investment choices must provide investment options that permit adequate diversification. brightscope and investment company institute (2014) find that in 2012, 401(k) plans on average provided participants 25 investment options. 3. thrift savings plan the thrift savings plan (tsp) is the 401(k)-type plan for u.s. federal government employees, members of congress and the military. in terms of assets, it is the largest pension plan in the united states (towers watson, 2014) and the largest defined contribution plan in the world (white, 2011). it has more participants than the social security systems of more than 90 countries (world bank, 2014). we focus on the tsp because it only offers five basic investment options, plus target date funds. it also charges extremely low fees—three basis 49s. shen, j.a. turner / financial services review 27 (2018) 47-81 points for all its funds, including its international equity fund and its target date funds, which tend to be higher fee funds. the average fee for target date funds outside the tsp is roughly 30 times higher than for the target date funds the tsp provides (vanguard, 2014). a survey of tsp participants who made a withdrawal in 2013 finds that an estimated 16,400 participants (about one-third of those making withdrawals) made a withdrawal of all or part of their tsp account because they were advised by their financial adviser to do so (aonhewitt, 2014). advisers frequently advise tsp participants to roll over from their low-fee account to an ira that the adviser would manage. a survey of financial advisers finds that advisers who advise their clients to roll over their tsp accounts commonly use the argument that because the tsp only offers five funds (plus lifecycle or target date funds based on those five funds), the client can obtain greater diversification outside of the tsp (turner, klein, and stein 2016). for example, ric edelman, who was three times named the top independent financial adviser in the united states by barron’s, states, “the downside to the thrift plan is the fact that you have only five investment choices. none of them are particularly exciting in terms of their performance relative to what’s available elsewhere, so we are not terribly thrilled with the choices in the thrift plan although we do acknowledge it’s really cheap” (tergesen, 2014). 4. literature review 4.1. quality of advice because of the importance of the 401(k) rollover decision in retirement planning, many people seek financial advice. one survey finds that 61% of the people with rollover iras received advice to roll over from a financial adviser (investment company institute, 2016b). a small but growing literature focuses on the quality of investment advice that some financial advisers provide their clients as being a factor leading to poor investment outcomes. mullainathan, noeth, and schoar (2012) find that people who initially are invested in low-fee, diversified portfolios are frequently advised to invest in higher-fee, less-welldiversified portfolios. dvorak (2015) compares the 401(k) plan investment choices in the plans of financial advisory firms with the plans of the companies they advise. he finds that the investment options in the advisee firms’ plans but not in the adviser firms’ plans tend to be high-fee options. christoffersen, evans, and musto (2013) find that brokers tend to sell higher-cost funds that give them higher compensation. an argument counter to these findings is that individual investors tend to underperform the stock market because of bad financial decisions, and that they could do better in avoiding their mistakes if they had the assistance of a financial adviser (anspach, 2016). the fundamental explanation for bad advice is the conflict of interest that many advisers have. however, several theories go further to investigate why advisers act on that conflict of interest. akerlof and shiller (2015), two nobel prize laureates, in their book phishing for phools: the economics of manipulation & deception, argue that many financial advisers take advantage of the behavioral biases of their clients that lead to poor decision making. 50 s. shen, j.a. turner / financial services review 27 (2018) 47-81 a related strand of literature relates to psychological underpinnings of bad advice. di tella et al., (2015) analyze instances of self-serving biases, which occur when people take actions that benefit themselves but that harm other people. in such instances, the people taking the action negatively distort their views of the other person (think badly of the other person) to make it psychologically less costly to treat them poorly. di tella et al.’s main hypothesis is that “people manage their self image while trying to earn money.” this purposeful bias in one’s views of another person reduces the psychological cost of taking an action that is favorable to oneself but harmful to the other person. in our paper, we make a slightly different argument. we argue that some financial advisers exhibit self-serving biases in that they make true but incomplete statements to their clients because it is psychologically less costly to make those statements than it is to make false statements. because of self-serving biases, the advisers may believe that their advice concerning improved diversification is good advice. as di tella et al., (2015) note, “the possibility that beliefs exhibit a self-serving bias has been studied since the development of the theory of cognitive dissonance (e.g., festinger, 1957; hastorf and cantril, 1954).” chen and gesche (2016) in an experiment find that some people induced to provide bad advice through use of a cash incentive are likely to continue to provide that advice when the incentive is removed. they argue that the subjects continue to provide bad advice because the subjects have adopted a self-serving bias that justifies the objectively bad advice as actually being good advice. 4.2. diversification relating to the issue of portfolio diversification, fama (1976) analyzes the effect on the standard deviation of a portfolio of adding an additional stock. he finds a large decline in standard deviation up to 20 stocks, but relatively little further reduction when adding further stocks. specifically, he finds that about 95% of the reduction in standard deviation in going from a portfolio of one stock to a portfolio with more stocks is achieved with a portfolio of 20 stocks. according to betterment (2016), an investment adviser, “many investors know that they should be diversified, but don’t understand what that really means.” lusardi and mitchell (2011), in a survey of older americans, find that only half of respondents know that holding stock in a single company is riskier than investing in a mutual fund. benartzi and thaler (2001) present experimental evidence suggesting the tendency of investors to engage in naïve diversification, splitting their investments evenly among the available options when a small number of options are provided. this approach is called the 1/n approach. fisch and wilkinsonryan (2014) present further experimental evidence that unsophisticated investors may be attracted to naïve diversification strategies, which may explain the appeal of the advice that they can have more investment options if they roll over their employer-sponsored defined contribution plan to an ira. for example, in their experiment, 75% of those participants who invested in a low-fee equity index fund also invested in an identical high-fee fund. money magazine (2015) identifies as a myth some investors believe that investing in a large number of different mutual funds guarantees diversification, writing “breadth of holdings alone does not guarantee diversification.” that myth directly relates to the success of the advice to roll over to have access to a larger number of investment options. 51s. shen, j.a. turner / financial services review 27 (2018) 47-81 for some pension participants, however, having many options may make investment decisions more difficult. behavioral economics does not support the idea that having unlimited choice by rolling over to an ira is a good feature. “the paradox of choice” refers to the negative effects of having too many choices. several studies document the problems that people have in making decisions when facing a large number of options (carosa, 2014; iyengar and lepper, 2000). despite the concept from traditional economics that more options are always better, research has documented that for psychological reasons of mental overload, above a certain level, fewer choices are better for many people when the available options allow for a sufficient range of choice. relating specifically to pension investment options, a study finds that having many investment options in 401(k) plans lowers participation rates (iyengar, huberman, and jiang, 2004). in this paper, we do not explore the issue of what the optimal number of investment options is from the perspective of the participants’ ability to make investment decisions, but rather what is the minimal number of investment options needed to provide adequate diversification. another aspect of too much choice, in the context of iras, is the tradeoff between quantity and quality of choice. a large number of choices that are not preselected by a financial expert with a fiduciary obligation, as is the case with iras, will include more options that are of poor quality, are poorly diversified, have high fees and poor rates of return (goldreich and halaburda, 2011). a substantial literature demonstrates that the cognitive costs of greater choice, beyond a certain number of choices, can lead to worse savings and retirement investment choices (ashraf, karlan, and yin, 2006; choi et al., 2007, 2010; duarte and hastings, 2009; hastings and tajeda-ashton, 2008; madrian and shea, 2001). 5. investment options in the tsp the thrift savings plan for federal government employees, members of the military, and members of congress uses passively managed index funds. the tsp offers a choice of 10 funds, five of which are lifecycle or target date funds, based on the participant’s expected date of retirement. in the empirical analysis, we focus on individual portfolios constructed from the five basic funds. the five basic funds are: (1) the government securities investment fund (g fund, which is based on medium-term and long-term government bond rates); (2) the fixed income index investment fund (f fund, which tracks the barclays capital u.s. aggregate bond index), which includes treasury securities, government-agency bonds, mortgage-backed bonds, corporate bonds and a small amount of foreign bonds traded in the united states; (3) the common stock index investment fund (c fund, which tracks the standard & poor’s 500 index); (4) the small capitalization index fund (s fund, which tracks the dow jones u.s. completion total stock market index, it represents all u.s. equities other than those in the standard & poor’s 500 index); and (5) the international stock index investment fund (i fund, which tracks the morgan stanley capital international eafe, which is the europe, australasia, and far east index; thrift savings plan, 2015), which includes securities from more than 20 developed countries. 52 s. shen, j.a. turner / financial services review 27 (2018) 47-81 6. empirical analysis the tsp stock funds do not cover emerging markets, canada, and international small capitalization stocks. they also do not include real estate, commodities, and international bond funds. copeland (2013) finds that, in aggregate, ira participants invest 13.8% of their assets in the category “other,” which refers to investments not in stocks, bonds, or target date funds. this finding suggests that ira participants do hold a wider range of investments, because the tsp does not have any investment that would be in that category. the tsp also does not offer actively managed funds. the spanning test introduced by huberman and kandel (1987) is a well-known method for investigating whether a set of funds is adequately diversified. it uses a likelihood ratio test to examine whether additional risky assets can span the minimum–variance frontier. seminal work by tang et al. (2010), and elton et al. (2006) also adopt this method for examining the adequacy and efficiency of 401(k) investment options. kan and zhou (2012) impose a comprehensive test for a mean-variance spanning method. however, bessler and wolff (2015) argue that the spanning method is limited to in-sample tests which may exaggerate the benefits of additional assets. we adopt the bessler and wolff (2015) method to test whether the optimal portfolio constructed by the basic five tsp funds is fully diversified. first, we investigate whether the tsp participants can benefit from greater diversification when extra investment options are available in addition to the five existing options. second, we investigate the closely related question of whether rolling over from the tsp with its five basic investment options results in better diversification. 6.1. asset allocation strategies suppose an investor can allocate her wealth among n risky funds. we use four different asset allocation strategies–the mean-variance approach, the minimum-variance portfolio; the 1/n naïve diversification rule, and the risk parity asset allocation strategy. we impose borrowing constraints on each of the asset allocation strategies; hence, a short position is prohibited. in the mean-variance approach (markowitz, 1952), the investor faces a trade-off between risk and return. the investor maximizes the mean-variance utility function max� u � ��� � � 2 ���� (1) where � measures risk aversion. portfolio weights on risky assets are indicated by the n – vector �, with � � (�1 . . . �n)�. the expected excess return (relative to the risk-free rate) is denoted by the n – vector � with � � (�1 . . . �n)�, and the n � n covariance matrix � is given by � � � �1 2 · · · �1n··· · · · ··· �n1 · · · �n 2 �. the fraction of wealth not invested in risky assets, 1 � i��, is invested in the risk-free asset. if borrowing is allowed, the optimal investment in risky assets is 53s. shen, j.a. turner / financial services review 27 (2018) 47-81 �mv � 1 � ��1 � (2) if borrowing is not allowed (�i � 0), we need to solve the quadratic optimization problem numerically. for the purpose of in-sample analysis, there is no need to analyze different asset allocation strategies because the markowitz mean-variance strategy dominates all other strategies under the assumption that the in-sample analysis includes a perfect forecast of all asset returns. however, this assumption does not reflect reality, as this approach is limited to the condition that future performance of the return series is known in advance and input parameters are estimated without error. therefore, we also introduce alternative asset allocation strategies that rely less on the estimation performance of input variables and analyze the out-of-sample benefit of having additional funds. the expected asset returns are notoriously difficult to estimate from historical data (merton, 1980). hence, it can be beneficial to exclude return estimates and to focus solely on risk estimates. therefore, we also introduce alternative strategies that require only risk estimates and a naïve strategy that requires no estimates. the minimum-variance portfolio is obtained by solving the following optimization problem min� u � ���� (3) subject to ��l � 1. without the borrowing constraint, the global minimum variance portfolio is given by �gmv � ��1l l���1l the 1/n naïve portfolio denoting �na � 1/n can be attractive to some private investors (benartzi and thaler, 2001)). demiguel, garlappi, and uppal (2009) show that the 1/n naive strategy might outperform the mean-variance portfolios when the input mean and covariance parameters are estimated with errors because the 1/n naïve portfolio is independent of the input data, hence is estimation error neutral. the risk parity strategy is commonly implemented by index providers, defined benefit pension funds and long-term investors (see anderson, bianchi, and goldberg 2012). the idea of the risk parity strategy is to adjust the risk allocation of each fund so that each fund has the same risk level. the risk parity portfolio, denoted by �rp, is given by �rp(i) � 1 �i �i�1 n 1 �i where i � 1, . . ., n and �i is the volatility of asset i. 54 s. shen, j.a. turner / financial services review 27 (2018) 47-81 6.2. estimation methods mean-variance portfolios are highly sensitive to estimates of means, volatilities, and correlations. mis-specified inputs make mean-variance portfolios problematic. in this paper, we introduce three robust estimators, including factor models and the bayesian shrinkage method, aiming at enhancing the quality of input estimators. define an n-variate random vector for each time period rt � (r1,t . . . rn,t)�. we first introduce the least robust method to estimate � and �, namely using their historical sample analogues: �sm � 1 t �t�1 t rt �sm � 1 t � n � 2 �t�1 t (rt � �sm) (rt � �sm)�. where t refers to the sample observations. historical averages are very noisy estimates of the mean return. hence, we also adopt two indexed-model based estimation approaches and a bayesian shrinkage method recommended by ledoit and wolf (2003) to estimate the input parameters. the capital asset pricing model (capm), a well-known linear one-factor model that analyzes equilibrium expected returns of risky assets, is used for our analysis. the model assumes the rate of excess return of asset i is given by ri,t � �i � �irm,t � ei,t (4) where rm,t is the excess market return at time t. the difference between the fair return and the actually expected rate of excess return on asset i is captured by the constant term, �i. the constant coefficient term �i measures the contribution of asset i to the variance of the market portfolio. the error term of the regression is denoted by ei,t. therefore, the capm estimates for � and � are given by �capm � �capm � �capm �m (5) �capm � �m 2 �capm�capm� � capm (6) where �capm � (�1 · · · �n)� and �capm � (�1 · · · �n)� are coefficient vectors. capm is the n � n diagonal matrix with entries (�2(e1) . . . �2(en)). �m and �m 2 are the mean and variance of the market excess return series. we also use the fama-french three-factor model to estimate expected asset returns: rit � �i � �imrmt � �ismbsmbt � �ihmlhmlt � eit (7) where smb � small minus big is the return of a portfolio of stocks with a high book-to-market ratio in excess of the return on a portfolio of large stocks. hml � high minus low is the return of a portfolio of stocks with a high book-to-market ratio in excess of the return of portfolio of stocks with a low book-to-market ratio. 55s. shen, j.a. turner / financial services review 27 (2018) 47-81 under the fama-french multifactor model, the expected excess return and covariance matrix of the assets are �ff3 � �ff3 � �ff3 �f �ff3 � �ff3�f�ff3� � �ff3 where �ff3 � ��1 · · · �n�� and �ff3 � ��m, �smb, �hml� is a n � 3 matrix with �f � ��1f · · · �nf�� for f � m, smb and hml, respectively. �f � ��m, �smb, �hml�� and �f � e��ft � �f��ft � �f�� are the mean and covariance matrix of the three factors, with ft � �rmt, smbt, hmlt��. �ff3 is the n � n diagonal matrix with entities equal to the variance of the regression residual for each asset i. for both the capm and fama-french models, we use least-squares estimates �̂i, �̂i for the time-series regression for each asset i. we use moment methods to estimate the remaining parameters. investors can also improve the estimates of inputs by using robust statistical estimators. the bayesian shrinkage approach by stein (1956) and james and stein (1961) is one of the most well received robust estimators. the shrinking estimators take care of outliers and extreme values that may jeopardize estimation performance. we follow bessler, opfer, and wolff (2017) to formulate the bayesian shrinkage methods. the idea behind the bayesian shrinkage method is to shrink the sample mean �sm towards the expected return of the minimum variance portfolio �min with: �min � �gmv�sm � � 1l l�� 1l �sm the return estimates based on the bayesian shrinkage approach take the form �bs � (1 v)�sm � v�minl with v � n � 2 �n � 2� � t��sm � �minl ��� 1��sm � �minl � , where n is the number of assets and t is the sample size. we set bayesian shrinkage estimates of the covariance matrix to be the same as sample moment methods. with �bs � �sm. 6.3. performance measures to measure the out-of-sample performance of optimal portfolios under various estimation methods, we compute the portfolio’s net return ��, and volatility � �� as well as the net sharpe ratio �� � �� in excess of transaction cost for each asset allocation strategy. 56 s. shen, j.a. turner / financial services review 27 (2018) 47-81 6.4. data tsp funds track the performance of various stock and bond indices, so we use those indices to run the diversification analysis. we collect monthly returns from january 1993 to april 2015. we use the barclays capital u.s. aggregate bond index for the f fund; standard and poor’s (s&p) 500 stock index for the c fund; the dow jones u.s. completion total stock market index for the s fund; and the msci eafe stock index for the i fund. because the g fund has different risk and return characteristics from publicly available u.s. government bonds, we use monthly-rate-of-return data provided to us by the thrift savings board. we use three-month u.s treasury discount bond yields as a proxy for the risk-free rate. we consider four additional investment options. first, we include a real estate fund. we use the ftse nareit u.s. real estate index series (reit). previous studies, such as burns and epley (1982), ennis and burik (1991), and giliberto (1993), use reit data to show that investing in real estate funds improves diversification for u.s investors. second, we add an emerging market fund. li, sarkar, and wang (2003) find substantial international diversification benefit for u.s equity investors. the data we use are from the msci emerging markets index. third, we consider the commodity market. daskalaki and skiadopoulos (2011) show that only nonmean-variance investors can benefit from commodity investment, and this result only holds in sample. we use the s&p goldman sachs commodity index to calculate the return from investing in the commodity market. fourth, we add an international bonds fund. we use the citi non-usd non-gbp world government bond index as a proxy. table 1 shows summary statistics for the data. the upper panel presents the sample moments of the five tsp funds. the g fund is almost risk free while providing an average annual return of 4.525%. the 10-year government bonds are roughly comparable with the g fund in terms of average return and volatility. other tsp funds and additional funds all have higher annualized returns compared with the g fund and also higher levels of risk. the average return of reits during the sample period is 9.926%, which is comparable with the expected stock index return (c fund). the sharpe ratios of all the rest of the additional funds are lower than the sharpe ratios of most of the tsp funds, suggesting that the additional funds are not attractive as a stand-alone investment. the jarque-bera statistic of most funds is significant at the 5%-level besides the fixed income indices, rejecting the assumption of normal distribution of returns for all funds except bonds. even if the additional funds do not appear to be attractive in terms of stand-alone investments, they may still improve the risk-return profile if the correlations with the tsp funds are low or negative. to gain insights in terms of potential diversification benefit, we present the pair-wise correlation matrix in table 2. we find low but significantly positive correlation between the international bond index and most of the tsp funds. there is also a low but significant correlation between the real estate index and the f fund. the emerging market index is highly correlated with most of the tsp funds. based on our correlation analysis, an international bond index fund might be able to bring additional diversification benefit to the tsp portfolio. 57s. shen, j.a. turner / financial services review 27 (2018) 47-81 6.5. analysis we start the empirical analysis by examining the in-sample benefit of adding extra funds to the tsp portfolio. table 3a and 3b reports the optimal weight on each risky asset class for the mean-variance portfolio. as a benchmark, we set the risk aversion level at � � 5. both table 1 descriptive statistics of asset returns (january 1993 to april 2015) mean (%) standard deviation (%) skewness kurtosis sharpe var (99%) jb (p-value %) observations tsp g fund 4.525 0.504 �0.190 1.940 0.000 0.643 0.084 268 tsp f fund 5.686 3.595 �0.238 3.966 0.821 2.881 0.154 268 tsp c fund 8.160 14.579 �0.711 4.300 0.372 8.966 0.000 268 tsp s fund 10.090 18.567 �0.634 4.634 0.396 13.752 0.000 268 tsp i fund 5.684 16.462 �0.659 4.351 0.179 9.816 0.000 268 10-year bond 4.508 0.449 �0.057 2.167 0.000 0.640 1.925 268 real estate 9.926 19.174 �1.649 13.110 0.375 11.851 0.000 268 emerging 8.116 23.118 �0.710 5.071 0.233 14.854 0.000 268 commodity 6.201 21.224 �0.269 4.603 0.163 14.943 0.000 268 intern. bond 5.203 8.300 0.223 3.897 0.297 6.510 0.368 268 3-month t-bill 2.734 0.633 0.025 1.387 0.000 0.520 0.000 268 this table provides sample moments, sharpe ratios, value-at-risk, and jarque-bera statistics of the five tsp-fund indices, the 10-year government bond, the four additional fund indices and the risk-free rate used in the empirical analysis. the evaluation period covers 268 months from january 1993 to april 2005. mean and standard deviation represent annualized time-series mean and annualized standard deviation of monthly returns. skewness and kurtosis denote the third and the fourth moment of the return distribution. sharpe represents the annualized sharpe ratios of the respective asset classes. we treat g fund index and 10-year government bond as riskless assets; therefore, their sharpe ratios are zeros. var (99%) shows the nonparametric 99% value-at-risk of the monthly returns during the sample period. jb (p-value) is the p-value of the jarque-bera statistics for testing normality of sample returns. var � value at risk; jb � jarque-bera; tsp � thrift savings plan; intern. � international. table 2 correlation matrix of asset returns (january 1993 to april 2015) correlation tsp funds additional funds g fund f fund c fund s fund i fund 10 year bond real estate emerging commodity intern. bond g fund 1.000 0.197** �0.017 �0.048 �0.031 0.969** �0.027 �0.066 �0.009 0.108* f fund 0.197** 1.000 0.039 �0.017 0.033 0.163** 0.176** 0.002 0.005 0.210** c fund �0.017 0.039 1.000 0.851** 0.801** �0.012 0.556** 0.718** 0.255** 0.002 s fund �0.048 �0.017 0.851** 1.000 0.753** �0.045 0.581** 0.737** 0.308** 0.010 i fund �0.031 0.033 0.801** 0.753** 1.000 �0.019 0.527** 0.784** 0.380** 0.154* 10 year bond 0.969** 0.163** �0.012 �0.045 �0.019 1.000 �0.034 �0.061 0.021 0.064 real estate 0.001 0.182** 0.563** 0.585** 0.532** �0.005 1.000 0.476** 0.178** 0.078 emerging �0.066 0.002 0.718** 0.737** 0.783** �0.061 0.472** 1.000 0.361** 0.028 commodity �0.009 0.005 0.255** 0.308** 0.380** 0.021 0.163** 0.361** 1.000 0.074 intern. bond 0.108* 0.210** 0.002 0.010 0.154* 0.064 0.072 0.028 0.074 1.000 the table displays the correlation matrix for the five tsp-fund indices, the 10-year government bond and the four additional fund indices used in the empirical analysis over the period from january 1993 to april 2015. * and ** represent the correlation values significantly different from zero at the 5% and 1% level, respectively. intern. � international. 58 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 3a in -s am pl e op tim al po rt fo lio w ith ou t co ns tr ai nt (i n pe rc en ta ge ) m et ho d sa m pl e m om en ts c a pm t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp f fu nd 92 .7 5 92 .6 4 92 .8 4 91 .5 1 83 .8 1 82 .6 7 82 .6 7 79 .9 9 75 .4 7 77 .4 9 77 .9 7 71 .3 6 64 .9 2 64 .9 2 c fu nd 6. 14 6. 11 6. 13 7. 01 7. 57 8. 31 8. 31 8. 81 8. 09 8. 46 8. 57 7. 83 6. 87 6. 87 s fu nd 11 .9 2 11 .8 2 11 .7 1 11 .2 2 11 .5 3 10 .5 0 10 .5 0 7. 43 6. 77 7. 12 7. 23 6. 60 5. 73 5. 73 i fu nd � 10 .8 2 � 10 .9 0 � 11 .2 1 � 11 .9 8 � 12 .3 6 � 14 .0 5 � 14 .0 5 3. 77 3. 39 3. 60 3. 66 3. 34 2. 85 2. 85 r ea l 0. 32 0. 32 0. 32 6. 28 5. 28 5. 28 e m er . 0. 53 0. 71 0. 71 3. 33 2. 66 2. 66 c om . 2. 24 2. 08 2. 08 2. 57 2. 06 2. 06 in t. b 9. 46 9. 46 9. 46 10 .8 7 9. 63 9. 63 m et ho d fa m afr en ch 3 fa ct or s b ay es ia n sh ri nk ag e t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp f fu nd 80 .0 5 75 .4 3 77 .6 2 77 .9 9 71 .3 1 64 .9 4 64 .9 4 92 .6 3 92 .5 2 92 .8 0 91 .4 7 83 .7 6 82 .8 5 82 .8 5 c fu nd 8. 78 8. 11 8. 42 8. 55 7. 82 6. 91 6. 91 6. 35 6. 23 6. 18 7. 07 7. 63 8. 12 8. 12 s fu nd 7. 41 6. 76 7. 09 7. 21 6. 58 5. 72 5. 72 12 .3 8 12 .0 7 11 .8 3 11 .3 4 11 .6 7 10 .0 8 10 .0 8 i fu nd 3. 76 3. 38 3. 58 3. 65 3. 33 2. 84 2. 84 � 11 .3 5 � 11 .2 5 � 11 .3 7 � 12 .1 2 � 12 .5 3 � 13 .3 2 � 13 .3 2 r ea l 6. 33 5. 28 5. 28 0. 44 0. 10 0. 10 e m er . 3. 28 2. 69 2. 69 0. 55 0. 64 0. 64 c om . 2. 60 2. 03 2. 03 2. 24 2. 06 2. 06 in t. b 10 .9 6 9. 60 9. 60 9. 47 9. 47 9. 47 t hi s ta bl e re po rt s th e po rt fo li o w ei gh ts (% ) of ri sk y as se ts fo r th e in -s am pl e op ti m iz ed m ea nva ri an ce po rt fo li o. w it ho ut bo rr ow in g or sh or t co ns tr ai nt , po rt fo li o w ei gh ts ca n be be yo nd 1 or be ne ga ti ve . in pu t pa ra m et er s, m ea n an d co va ri an ce m at ri x, ar e es ti m at ed us in g fo ur di ff er en t m et ho ds . t he op ti m al po rt fo li o is ne t of ad m in is tr at io n co st th at is se t at 3 bp fo r th e t s p pa rt ic ip an ts an d 60 bp fo r th e no nt s p pa rt ic ip an ts . r is k av er si on le ve l is se t at � � 5. t sp � th ri ft sa vi ng s pl an ; r ea l � re al es ta te ; e m er � em er gi ng m ar ke t; c om � co m m od ity ; in t.b . � in te rn at io na l bo nd . 59s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 3b in -s am pl e op tim al po rt fo lio w ith co ns tr ai nt (i n pe rc en ta ge ) m et ho d sa m pl e m om en ts c a pm t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp f fu nd 71 .6 2 67 .0 3 71 .6 2 71 .6 2 71 .6 2 67 .0 2 67 .0 2 45 .9 1 27 .6 5 38 .0 3 43 .0 9 45 .9 1 18 .5 3 18 .5 3 c fu nd 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 25 .5 7 23 .8 0 24 .9 3 25 .3 8 25 .5 7 23 .0 9 23 .0 9 s fu nd 28 .3 7 21 .9 7 28 .3 8 28 .3 8 28 .3 8 21 .9 7 21 .9 7 26 .9 7 25 .5 4 26 .4 8 26 .8 4 26 .9 7 25 .0 2 25 .0 2 i fu nd 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 1. 56 0. 05 1. 02 1. 40 1. 56 0. 00 0. 00 r ea l 11 .0 1 11 .0 1 11 .0 1 22 .9 7 22 .3 9 22 .3 9 e m er . 0. 00 0. 00 0. 00 9. 54 8. 64 8. 64 c om . 0. 00 0. 00 0. 00 3. 29 2. 33 2. 33 in t. b 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 m et ho d fa m afr en ch 3 fa ct or s b ay es ia n sh ri nk ag e t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp t sp � r ea l � e m er � c om � in t. b � a ll n on -t sp f fu nd 46 .1 3 27 .4 9 38 .5 0 43 .1 9 46 .1 3 18 .4 3 18 .4 3 73 .7 1 70 .1 4 74 .3 8 74 .4 0 74 .4 1 72 .3 4 72 .3 4 c fu nd 25 .4 6 23 .8 6 24 .7 6 25 .2 9 25 .4 6 23 .1 0 23 .1 0 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 s fu nd 26 .9 1 25 .5 2 26 .3 8 26 .7 8 26 .9 1 24 .9 5 24 .9 5 26 .2 9 20 .3 2 25 .6 2 25 .6 0 25 .6 0 19 .1 5 19 .1 5 i fu nd 1. 51 0. 00 0. 97 1. 35 1. 51 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 r ea l 23 .1 3 22 .6 2 22 .6 2 9. 54 8. 50 8. 50 e m er . 9. 39 8. 65 8. 65 0. 00 0. 00 0. 00 c om . 3. 40 2. 25 2. 25 0. 00 0. 00 0. 00 in t. b 0. 00 0. 00 0. 00 0. 00 0. 00 0. 00 t hi s ta bl e re po rt s th e po rt fo lio w ei gh ts (% ) of ri sk y as se ts fo r th e in -s am pl e op tim iz ed m ea nva ri an ce po rt fo lio w ith co ns tr ai nt on bo rr ow in g or sh or t po si tio ns . in pu t pa ra m et er s ar e es tim at ed us in g fo ur di ff er en t m et ho ds . t he op tim al po rt fo lio is ne t of ad m in is tr at io n co st th at is se t at 3 bp fo r th e t sp pa rt ic ip an ts an d 60 bp fo r th e no nt sp pa rt ic ip an ts . a s a be nc hm ar k, w e se t ri sk av er si on le ve l at � � 5. t sp � th ri ft sa vi ng s pl an ; r ea l � re al es ta te ; e m er � em er gi ng m ar ke t; c om � co m m od ity ; in t.b . � in te rn at io na l bo nd . 60 s. shen, j.a. turner / financial services review 27 (2018) 47-81 tables contain four panels representing four different estimates of parameter inputs. the first column of each panel reports the portfolio weight of the four tsp risky funds. table 3a does not impose a borrowing constraint, while table 3b does. for both cases, the f-fund dominates the optimal portfolio because of its high sharpe ratio (see table 1) compared with the other funds. we do not include the g-fund in our diversification analysis since it is risk free. the next four columns report the optimal weight after introducing one more instrument to the tsp funds, while keeping the low administration fee of 3 bps. the sixth column of each panel gives the optimal portfolio assuming all additional funds are included the tsp plan. the same market instruments are included in the last column while based on the non-tsp members cost which is 20 times more expensive. the cost of investments does not influence the mean-variance optimal portfolio; hence the last two columns are always the same. the optimal result demonstrates the attractiveness of the f-fund under various estimation methods. the capm and fama-french models make the real estate fund a valuable additional instrument with more than 20% portfolio weight for the constraint case. the international bond is also an attractive investment vehicle with around a 10% portfolio weight if borrowing is allowed. the results also demonstrate how sensitive the portfolio weights are to the input estimates and the consideration of the borrowing constraint. the in-sample performance is presented in table 4. in general, tsp risky funds are sufficiently diversified, with limited room for further diversification. table 4 shows that adding more investment options to the tsp benchmark portfolio may only bring a small while significant improvement to the optimal mean-variance portfolio under capm or the fama-french approach when borrowing is permitted. however, for non-tsp investors, the effect of adding more funds is not appealing because of the high administration cost. in other words, investors can only benefit from rolling over from the tsp while accessing alternative instruments if a substantial result can be achieved to cover the high administration cost. however, with the borrowing constraint, the possibility of having substantially high returns is fairly low. the in-sample test assumes a perfect forecast of all asset returns. this assumption does not reflect reality as it is limited to the condition that future performance of the return series is known in advance. therefore, we also analyze the out-of-sample benefit of having additional funds. we use the rolling estimation window method used by demiguel, garlappi, and uppal (2009) to compare the performance of various asset allocation strategies. it is not necessary to consider other asset allocation strategies in the in-sample analysis since the markowitz (1952) mean-variance strategy dominates any alternative strategies if investors only care about portfolio risk and return. the alternative asset allocation strategies we consider include the global minimum-variance strategy, risk parity strategy, and 1/n naïve strategy. table 5a through 5d provide the average portfolio weight of each risky asset as well as the respective standard deviation for the various optimization strategies and for the different estimation methods. the optimal portfolio for non-tsp participants is displayed in the last column of each panel in table 5, in which all the eight risky assets are included. the standard deviation of the optimal mean-variance portfolio (table 5a and 5b) is much larger than other asset allocation strategies, indicating a strong fluctuation of portfolio shares over time, and 61s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 4 in sa m pl e pe rf or m an ce of op tim al m ea nva ri an ce po rt fo lio e st im at io n m et ho d sa m pl e m om en ts c a pm fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith ou t co ns tr ai nt t sp 20 .9 1 19 .0 8 0. 95 23 .1 8 20 .2 4 1. 01 23 .1 6 20 .2 3 1. 01 17 .2 2 17 .0 4 0. 85 � r ea l es ta te 20 .9 2 19 .0 9 0. 95 26 .0 3 21 .6 0 1. 08 * 26 .0 6 21 .6 1 1. 08 * 17 .1 0 16 .9 7 0. 85 � e m er gi ng 20 .9 2 19 .0 9 0. 95 24 .2 7 20 .7 7 1. 04 24 .2 2 20 .7 4 1. 04 17 .1 0 16 .9 7 0. 85 � c om m od ity 21 .1 6 19 .2 1 0. 96 23 .7 2 20 .5 0 1. 02 23 .7 1 20 .5 0 1. 02 17 .2 6 17 .0 6 0. 85 � in te rn . bo nd 21 .6 8 19 .4 8 0. 97 24 .9 5 21 .0 9 1. 05 24 .9 6 21 .1 0 1. 05 17 .5 8 17 .2 5 0. 86 � a ll 21 .9 3 19 .6 1 0. 98 * 29 .3 1 23 .0 7 1. 15 ** 29 .3 2 23 .0 7 1. 15 ** 17 .3 9 17 .1 4 0. 86 n on -t sp 21 .3 6 19 .6 1 0. 95 28 .7 4 23 .0 7 1. 13 ** 28 .7 5 23 .0 7 1. 13 ** 16 .8 2 17 .1 4 0. 82 m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith co ns tr ai nt t sp 6. 91 5. 83 0. 72 7. 48 6. 47 0. 73 7. 47 6. 46 0. 73 6. 76 5. 86 0. 69 � r ea l es ta te 7. 09 6. 20 0. 70 8. 34 7. 51 0. 75 8. 35 7. 52 0. 75 6. 86 6. 12 0. 68 � e m er gi ng 6. 91 5. 83 0. 72 7. 67 6. 68 0. 74 7. 66 6. 66 0. 74 6. 70 5. 76 0. 69 � c om m od ity 6. 91 5. 83 0. 72 7. 48 6. 45 0. 74 7. 48 6. 44 0. 74 6. 70 5. 76 0. 69 � in te rn . bo nd 6. 91 5. 83 0. 72 7. 48 6. 47 0. 73 7. 47 6. 46 0. 73 6. 70 5. 76 0. 69 � a ll 7. 09 6. 20 0. 70 8. 50 7. 64 0. 76 8. 51 7. 64 0. 76 * 6. 69 5. 83 0. 68 n on -t sp 6. 52 6. 20 0. 61 7. 93 7. 64 0. 68 7. 94 7. 64 0. 68 6. 76 5. 86 0. 69 t hi s ta bl e di sp la ys th e in -s am pl e op tim al m ea nva ri an ce po rt fo lio pe rf or m an ce ne t of ad m in is te ri ng co st fo r th e fu ll sa m pl e fr om ja nu ar y 19 93 to a pr il 20 15 . t he up pe r pa ne l pr es en ts th e re su lts w ith ou t bo rr ow in g co ns tr ai nt w hi le th e lo w er pa ne l re st ri ct s sh or t po si tio ns . t he fir st ro w of ea ch pa ne l is th e be nc hm ar k t sp 5fu nd po rt fo lio an d th e ne xt fiv e ro w s sh ow th e t sp po rt fo lio s co m pl em en te d w ith ad di tio na l fu nd s. t he ad m in is tr at iv e ex pe ns e fo r th e t sp pl an pa rt ic ip an ts is 0. 03 % pe r ye ar , w hi ch is ap pl ie d to th e fir st si x ro w s. t he la st ro w sh ow s th e pe rf or m an ce fo r th e no nt sp po rt fo lio w ith an nu al co st 20 tim es m or e ex pe ns iv e th an th e t sp pa rt ic ip an ts . m ea n de no te s th e an nu al iz ed m on th ly (i n pe rc en ta ge ) re tu rn s. st an da rd de vi at io n re pr es en ts th e as so ci at ed an nu al iz ed st an da rd de vi at io n of po rt fo lio re tu rn s. sh ar pe is th e an nu al iz ed sh ar pe ra tio .t he m ea n an d th e co va ri an ce -m at ri x ar e es tim at ed us in g fo ur di ff er en t m et ho ds . * an d ** in di ca te th e si gn ifi ca nt hi gh er va lu es of sh ar pe ra tio in co m pa ri so n to th e va lu e of th e be nc hm ar k t sp po rt fo lio at th e 10 % an d 5% le ve l, re sp ec tiv el y. t sp � th ri ft sa vi ng s pl an ; in te rn . � in te rn at io na l. 62 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5a o ut -o fsa m pl e op tim al po rt fo lio w ei gh ts (i n pe rc en ta ge ) of m ea nva ri an ce po rt fo lio w ith ou t co ns tr ai nt m et ho d sa m pl e m om en ts c a pm t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith ou t bo rr ow in g co ns tr ai nt f fu nd 99 .3 6 93 .5 9 10 2. 34 94 .7 7 91 .0 4 82 .6 8 89 .9 8 82 .5 3 86 .4 1 85 .0 1 77 .8 1 67 .3 6 (7 .5 9) (1 2. 14 ) (1 1. 71 ) (7 .0 7) (8 .0 4) (9 .7 4) (1 2. 30 ) (1 4. 34 ) (1 0. 78 ) (1 0. 29 ) (7 .7 2) (7 .9 2) c fu nd � 13 .0 6 � 11 .9 1 � 10 .3 6 � 6. 64 � 10 .0 8 � 2. 38 3. 01 2. 31 2. 88 2. 83 2. 83 2. 01 (2 4. 23 ) (2 2. 90 ) (2 3. 45 ) (2 0. 35 ) (2 2. 48 ) (1 7. 31 ) (7 .3 2) (6 .7 5) (7 .0 5) (6 .8 5) (6 .2 0) (5 .3 6) s fu nd 23 .5 0 22 .5 5 18 .8 1 20 .0 0 22 .0 0 14 .9 1 5. 60 4. 80 5. 18 5. 12 4. 84 3. 67 (1 5. 58 ) (1 8. 09 ) (1 3. 28 ) (1 3. 72 ) (1 3. 80 ) (1 3. 44 ) (2 .4 1) (2 .2 8) (2 .4 0) (2 .3 7) (2 .1 5) (2 .0 1) i fu nd � 9. 80 � 9. 55 � 25 .0 8 � 14 .4 0 � 11 .4 0 � 24 .7 2 1. 41 1. 03 1. 15 1. 20 1. 24 0. 65 (1 1. 50 ) (1 0. 72 ) (1 6. 48 ) (1 0. 19 ) (1 3. 08 ) (2 1. 24 ) (4 .0 5) (3 .7 3) (3 .8 3) (3 .8 1) (3 .4 6) (2 .9 5) r ea l 5. 32 6. 82 9. 33 7. 66 (1 1. 63 ) (1 2. 07 ) (6 .0 8) (5 .9 9) e m er . 14 .3 0 9. 11 4. 38 3. 16 (1 9. 32 ) (1 8. 36 ) (3 .4 2) (2 .4 5) c om . 6. 27 5. 28 5. 84 4. 31 (3 .2 1) (3 .0 2) (3 .2 3) (2 .2 7) in t. b 8. 44 8. 29 13 .2 8 11 .1 9 (9 .8 7) (1 3. 13 ) (5 .9 7) (5 .0 8) (c on ti nu ed on ne xt pa ge ) t hi s ta bl e re po rt s th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e op tim iz ed m ea nva ri an ce po rt fo lio w ith ou t bo rr ow in g co ns tr ai nt . t he in pu t va ri ab le s ar e es tim at ed us in g sa m pl e m om en t m et ho d an d c a pm m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he si s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od .r em ar k: th e op tim al po rt fo lio w ei gh t is no t af fe ct ed by th e ad m in is tr at io n co st . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 63s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5a (c on tin ue d) m et ho d fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith ou t bo rr ow in g co ns tr ai nt f fu nd 90 .5 0 83 .0 1 87 .2 5 85 .5 0 78 .5 1 68 .0 9 10 0. 00 93 .5 5 10 4. 26 95 .1 5 90 .7 1 81 .9 9 (1 2. 06 ) (1 4. 52 ) (1 0. 55 ) (1 0. 22 ) (7 .3 5) (8 .3 3) (8 .4 9) (1 3. 47 ) (1 5. 34 ) (7 .6 5) (8 .7 8) (1 2. 18 ) c fu nd 2. 75 2. 15 2. 52 2. 62 2. 54 1. 77 � 14 .0 5 � 12 .9 2 � 12 .9 2 � 7. 40 � 10 .3 8 � 4. 02 (7 .2 9) (6 .7 6) (7 .1 3) (6 .8 3) (6 .1 8) (5 .4 8) (2 7. 83 ) (2 6. 58 ) (2 9. 44 ) (2 3. 19 ) (2 5. 69 ) (2 2. 35 ) s fu nd 5. 50 4. 72 5. 00 4. 99 4. 67 3. 47 25 .1 1 24 .0 1 20 .6 1 21 .3 0 23 .0 8 16 .0 9 (2 .3 2) (2 .1 9) (2 .3 7) (2 .3 3) (2 .1 0) (2 .0 6) (1 6. 72 ) (1 8. 81 ) (1 4. 35 ) (1 4. 59 ) (1 4. 81 ) (1 3. 67 ) i fu nd 1. 25 0. 91 0. 93 1. 04 1. 06 0. 48 � 11 .0 7 � 11 .0 7 � 29 .0 7 � 15 .6 4 � 12 .7 6 � 29 .3 5 (3 .9 7) (3 .6 5) (3 .7 8) (3 .7 6) (3 .3 9) (2 .9 4) (1 3. 92 ) (1 2. 97 ) (2 1. 18 ) (1 2. 16 ) (1 5. 58 ) (2 6. 75 ) r ea l 9. 22 7. 47 6. 43 8. 37 (6 .1 3) (6 .0 8) (1 2. 63 ) (1 3. 17 ) e m er . 4. 30 3. 14 17 .1 3 11 .5 3 (3 .4 4) (2 .4 2) (2 5. 04 ) (2 3. 73 ) c om . 5. 86 4. 32 6. 59 5. 84 (3 .1 9) (2 .2 4) (3 .6 8) (3 .7 2) in t. b 13 .2 1 11 .2 6 9. 36 9. 55 (6 .1 5) (5 .2 8) (1 0. 21 ) (1 4. 69 ) t hi s ta bl e co nt in ue s t ab le 5a re po rt in g th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e op tim iz ed m ea nva ri an ce po rt fo lio w ith ou t bo rr ow in g co ns tr ai nt .t he in pu t va ri ab le s ar e es tim at ed us in g th e fa m afr en ch 3fa ct or m od el an d th e b ay es ia n sh ri nk ag e m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he se s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 64 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5b o ut -o fsa m pl e op tim al po rt fo lio w ei gh ts of m ea nva ri an ce po rt fo lio w ith co ns tr ai nt m et ho d sa m pl e m om en ts c a pm t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith bo rr ow in g an d no -s ho rt co ns tr ai nt f fu nd 79 .2 6 58 .3 1 72 .6 4 65 .3 8 60 .2 9 27 .5 4 67 .9 3 41 .1 2 52 .9 6 49 .3 7 49 .3 5 11 .4 7 (1 2. 23 ) (2 6. 08 ) (8 .2 7) (1 0. 31 ) (2 4. 63 ) (2 4. 95 ) (2 1. 13 ) (3 3. 21 ) (2 1. 05 ) (1 8. 73 ) (2 0. 45 ) (1 8. 08 ) c fu nd 3. 44 2. 11 3. 44 3. 89 3. 44 2. 13 11 .1 2 8. 84 10 .5 2 10 .4 1 10 .9 1 6. 62 (7 .1 3) (4 .7 5) (7 .1 3) (7 .3 7) (7 .1 4) (4 .7 5) (1 0. 41 ) (8 .5 8) (1 0. 30 ) (1 0. 22 ) (1 0. 42 ) (8 .2 2) s fu nd 17 .0 1 10 .3 7 10 .0 0 11 .3 8 16 .3 0 4. 22 17 .9 8 15 .9 4 16 .8 2 16 .7 3 17 .1 4 12 .1 3 (1 2. 91 ) (1 4. 60 ) (1 1. 89 ) (1 0. 99 ) (1 2. 76 ) (9 .7 1) (1 1. 49 ) (1 1. 07 ) (1 1. 12 ) (1 1. 05 ) (1 1. 48 ) (1 0. 56 ) i fu nd 0. 29 0. 26 0. 00 0. 17 0. 28 0. 00 2. 98 2. 25 2. 44 2. 50 2. 77 0. 88 (2 .0 5) (1 .8 0) (0 .0 0) (1 .2 0) (1 .9 6) (0 .0 0) (5 .2 4) (4 .3 8) (4 .5 1) (4 .7 0) (5 .0 5) (2 .0 2) r ea l 28 .9 5 27 .0 3 31 .8 5 27 .0 8 (2 9. 89 ) (3 1. 32 ) (2 4. 37 ) (2 4. 18 ) e m er . 13 .9 2 7. 18 17 .2 6 14 .3 6 (1 4. 55 ) (9 .4 6) (1 2. 88 ) (1 1. 42 ) c om . 19 .1 8 17 .0 9 20 .9 9 17 .6 4 (1 3. 35 ) (1 1. 88 ) (1 3. 02 ) (1 1. 62 ) in t. b 19 .6 9 14 .8 2 19 .8 3 9. 82 (2 5. 89 ) (2 0. 96 ) (2 4. 40 ) (1 4. 89 ) (c on ti nu ed on ne xt pa ge ) t hi s ta bl e sh ow s th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e op tim iz ed m ea nva ri an ce po rt fo lio w ith bo rr ow in g an d sh or tpo si tio n co ns tr ai nt .t he in pu ts ar e ba se d on th e sa m pl e m om en ts an d th e c a pm ap pr oa ch .t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he se s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 65s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5b (c on tin ue d) m et ho d fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith bo rr ow in g an d no -s ho rt co ns tr ai nt f fu nd 68 .9 4 42 .2 6 54 .7 7 50 .3 1 50 .8 5 12 .5 3 80 .8 9 62 .1 6 74 .9 5 68 .7 8 63 .5 7 33 .7 5 (2 0. 17 ) (3 3. 34 ) (1 9. 51 ) (1 7. 95 ) (1 9. 35 ) (1 8. 40 ) (1 1. 47 ) (2 3. 37 ) (8 .1 8) (9 .2 0) (2 3. 74 ) (2 3. 39 ) c fu nd 10 .5 4 8. 41 9. 84 9. 95 10 .3 6 6. 38 3. 25 1. 99 3. 23 3. 64 3. 21 2. 02 (1 0. 17 ) (8 .4 2) (1 0. 09 ) (1 0. 09 ) (1 0. 19 ) (8 .1 3) (6 .7 0) (4 .4 7) (6 .6 5) (6 .8 0) (6 .6 2) (4 .3 8) s fu nd 17 .7 7 15 .6 9 16 .2 2 16 .3 9 16 .8 3 11 .8 3 15 .5 8 9. 76 8. 80 10 .1 9 14 .5 3 3. 85 (1 1. 33 ) (1 0. 78 ) (1 0. 94 ) (1 0. 92 ) (1 1. 35 ) (1 0. 35 ) (1 2. 03 ) (1 3. 48 ) (1 0. 61 ) (9 .8 8) (1 1. 60 ) (8 .7 3) i fu nd 2. 75 2. 06 2. 18 2. 32 2. 56 0. 78 0. 28 0. 24 0. 00 0. 17 0. 28 0. 00 (5 .0 1) (4 .1 6) (4 .2 4) (4 .5 3) (4 .8 3) (1 .9 7) (1 .8 8) (1 .6 9) (0 .0 0) (1 .2 0) (1 .7 7) (0 .0 0) r ea l 31 .5 8 26 .5 9 25 .8 5 23 .6 7 (2 4. 67 ) (2 4. 52 ) (2 6. 70 ) (2 7. 37 ) e m er . 16 .9 9 14 .2 3 13 .0 2 6. 58 (1 2. 80 ) (1 1. 34 ) (1 3. 63 ) (8 .6 6) c om . 21 .0 3 17 .5 2 17 .2 1 15 .4 6 (1 2. 95 ) (1 1. 64 ) (1 2. 12 ) (1 0. 66 ) in t. b 19 .4 1 10 .1 4 18 .4 0 14 .6 8 (2 4. 44 ) (1 5. 41 ) (2 4. 31 ) (2 0. 69 ) t hi s ta bl e sh ow s th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e op tim iz ed m ea nva ri an ce po rt fo lio w ith bo rr ow in g an d sh or tpo si tio n co ns tr ai nt .t he in pu ts ar e ba se d on th e fa m afr en ch 3fa ct or m od el an d th e b ay es ia n sh ri nk ag e m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he se s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 66 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5c o ut -o fsa m pl e op tim al po rt fo lio w ei gh ts of gl ob al m in im um va ri an ce po rt fo lio m et ho d sa m pl e m om en t c a pm t sp � r ea l � e m er � c om � in t.b � a ll t sp � r ea l � e m er � c om � in t. b � a ll m in im um -v ar ia nc e as se t al lo ca tio n st ra te gy f fu nd 93 .2 3 92 .4 3 93 .2 7 91 .9 0 89 .0 5 86 .4 6 87 .5 3 84 .4 4 85 .9 6 85 .7 8 76 .1 8 71 .8 3 (1 .8 2) (2 .4 4) (1 .6 2) (1 .9 3) (3 .6 4) (3 .7 6) (0 .9 5) (3 .0 3) (1 .1 3) (1 .4 4) (1 .4 4) (3 .6 8) c fu nd 7. 09 7. 31 6. 20 8. 60 7. 52 7. 86 4. 94 4. 66 4. 79 4. 78 4. 23 3. 82 (6 .2 7) (6 .3 1) (6 .1 2) (6 .9 6) (6 .8 8) (7 .0 1) (0 .2 2) (0 .1 9) (0 .2 0) (0 .2 1) (0 .2 3) (0 .1 8) s fu nd 1. 05 3. 89 1. 40 0. 30 0. 89 4. 06 3. 08 2. 86 2. 96 2. 97 2. 62 2. 29 (2 .0 1) (5 .6 1) (3 .0 5) (1 .6 9) (1 .9 0) (6 .7 4) (0 .2 7) (0 .2 5) (0 .2 5) (0 .2 2) (0 .1 5) (0 .1 2) i fu nd � 1. 37 � 1. 16 � 0. 24 � 2. 99 � 1. 82 � 0. 67 4. 45 4. 14 4. 29 4. 32 3. 85 3. 43 (8 .3 3) (7 .9 9) (6 .2 7) (9 .3 1) (9 .0 9) (7 .0 1) (0 .9 8) (0 .8 5) (0 .9 6) (1 .0 1) (0 .9 7) (0 .8 6) r ea l � 2. 47 � 2. 51 3. 90 3. 20 (5 .0 6) (5 .1 1) (2 .4 2) (2 .2 2) e m er . � 0. 63 � 1. 99 2. 00 1. 58 (2 .6 3) (3 .4 9) (0 .1 8) (0 .1 8) c om . 2. 19 2. 09 2. 15 1. 66 (0 .7 6) (0 .6 8) (0 .4 2) (0 .4 0) in t. b 4. 35 4. 70 13 .1 3 12 .1 9 (5 .5 6) (5 .9 9) (0 .5 9) (0 .5 2) (c on ti nu ed on ne xt pa ge ) t hi s ta bl e re po rt s th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e gl ob al m in im um va ri an ce po rt fo lio .t he in pu tv ar ia bl es ar e ba se d on th e sa m pl e m om en t m et ho d an d th e c a pm m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he si s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 67s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5c (c on tin ue d) m et ho d fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll m in im um -v ar ia nc e as se t al lo ca tio n st ra te gy f fu nd 88 .0 4 85 .0 1 86 .7 0 86 .3 0 76 .8 3 72 .6 3 92 .6 3 92 .3 6 92 .4 9 91 .5 8 88 .7 9 86 .7 8 (1 .4 7) (3 .6 9) (1 .7 5) (2 .1 1) (2 .0 0) (4 .6 2) (1 .6 9) (1 .7 2) (1 .6 8) (1 .6 8) (3 .9 6) (4 .5 9) c fu nd 4. 77 4. 52 4. 59 4. 64 3. 99 3. 63 9. 32 9. 53 7. 77 10 .5 6 9. 70 9. 15 (0 .3 3) (0 .3 2) (0 .3 2) (0 .3 7) (0 .4 3) (0 .4 2) (9 .3 4) (9 .3 2) (8 .6 0) (9 .9 4) (1 0. 03 ) (9 .3 9) s fu nd 2. 88 2. 66 2. 74 2. 77 2. 36 2. 04 � 1. 52 1. 49 � 0. 45 � 2. 17 � 1. 70 2. 53 (0 .1 7) (0 .1 5) (0 .1 8) (0 .2 0) (0 .2 9) (0 .3 2) (1 .2 7) (3 .7 3) (2 .0 0) (1 .3 2) (1 .4 3) (5 .6 0) i fu nd 4. 31 4. 03 4. 15 4. 19 3. 69 3. 30 � 0. 42 � 0. 10 2. 16 � 1. 71 � 0. 77 1. 97 (1 .1 0) (1 .0 0) (1 .0 7) (1 .1 3) (1 .1 0) (0 .9 9) (9 .4 7) (9 .1 2) (5 .9 9) (1 0. 29 ) (1 0. 27 ) (6 .4 4) r ea l 3. 78 3. 07 � 3. 28 � 3. 53 (2 .4 2) (2 .3 1) (4 .3 7) (4 .4 2) e m er . 1. 82 1. 44 � 1. 96 � 3. 14 (0 .3 1) (0 .3 3) (3 .9 6) (4 .9 1) c om . 2. 11 1. 63 1. 74 1. 74 (0 .5 2) (0 .5 0) (0 .7 6) (0 .7 5) in t. b 13 .1 3 12 .2 7 3. 99 4. 51 (0 .6 2) (0 .5 6) (5 .5 8) (5 .6 8) t hi s ta bl e co nt in ue s t ab le 5c re po rt in g th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e gl ob al m in im um -v ar ia nc e po rt fo lio .t he in pu t va ri ab le s ar e es tim at ed us in g th e fa m afr en ch 3fa ct or m od el an d th e b ay es ia n sh ri nk ag e m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he si s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 68 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5d o ut -o fsa m pl e op tim al po rt fo lio w ei gh ts of ri sk pa ri ty po rt fo lio m et ho d sa m pl e m om en t c a pm t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll r is k pa ri ty as se t al lo ca tio n st ra te gy f fu nd 60 .9 6 54 .2 9 55 .8 1 55 .4 9 48 .4 6 38 .8 1 60 .9 3 54 .2 7 55 .7 8 55 .4 6 48 .5 0 38 .8 2 (1 .1 8) (2 .7 9) (1 .2 1) (1 .5 4) (1 .0 9) (1 .9 9) (1 .0 1) (2 .5 7) (1 .0 3) (1 .3 5) (0 .8 9) (1 .7 6) c fu nd 14 .4 0 12 .8 2 13 .1 8 13 .1 1 11 .4 5 9. 16 14 .4 2 12 .8 4 13 .2 1 13 .1 3 11 .4 8 9. 19 (0 .2 5) (0 .4 9) (0 .2 1) (0 .2 4) (0 .1 9) (0 .3 3) (0 .2 5) (0 .5 2) (0 .2 1) (0 .2 6) (0 .2 0) (0 .3 4) s fu nd 11 .2 0 9. 98 10 .2 5 10 .2 0 8. 90 7. 13 11 .2 1 9. 99 10 .2 6 10 .2 0 8. 92 7. 14 (0 .5 4) (0 .7 4) (0 .4 9) (0 .5 3) (0 .4 5) (0 .5 1) (0 .5 9) (0 .7 8) (0 .5 3) (0 .5 7) (0 .4 7) (0 .5 3) i fu nd 13 .4 5 11 .9 3 12 .3 1 12 .2 3 10 .6 9 8. 53 13 .4 4 11 .9 3 12 .3 1 12 .2 3 10 .7 0 8. 54 (1 .3 6) (0 .8 5) (1 .2 3) (1 .1 5) (1 .0 6) (0 .6 1) (1 .2 7) (0 .7 7) (1 .1 5) (1 .0 7) (1 .0 0) (0 .5 8) r ea l 10 .9 8 7. 85 10 .9 7 7. 85 (3 .0 6) (2 .1 9) (2 .9 9) (2 .1 6) e m er . 8. 44 5. 87 8. 45 5. 87 (0 .2 6) (0 .2 0) (0 .2 4) (0 .2 2) c om . 8. 98 6. 26 8. 99 6. 28 (0 .8 6) (0 .4 5) (0 .8 1) (0 .4 3) in t. b 20 .5 0 16 .4 0 20 .4 0 16 .3 2 (0 .4 1) (0 .4 5) (0 .3 4) (0 .5 0) (c on ti nu ed on ne xt pa ge ) t hi s ta bl e sh ow s th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e ri sk pa ri ty po rt fo lio . t he in pu t va ri ab le s ar e es tim at ed us in g th e sa m pl e m om en t an d th e c a pm m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he se s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 69s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 5d (c on tin ue d) m et ho d fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e t sp � r ea l � e m er � c om � in t. b � a ll t sp � r ea l � e m er � c om � in t. b � a ll r is k pa ri ty as se t al lo ca tio n st ra te gy f fu nd 60 .9 3 54 .2 7 55 .7 8 55 .4 6 48 .5 0 38 .8 2 58 .6 8 52 .0 0 53 .4 6 53 .1 5 46 .2 6 36 .7 3 (1 .0 1) (2 .5 7) (1 .0 3) (1 .3 5) (0 .8 9) (1 .7 6) (1 .5 2) (3 .0 5) (1 .5 6) (1 .9 0) (1 .4 2) (2 .2 5) c fu nd 14 .4 2 12 .8 4 13 .2 1 13 .1 3 11 .4 8 9. 19 15 .2 2 13 .4 9 13 .8 7 13 .7 9 12 .0 0 9. 52 (0 .2 5) (0 .5 2) (0 .2 1) (0 .2 6) (0 .2 0) (0 .3 4) (0 .2 7) (0 .4 7) (0 .2 2) (0 .2 3) (0 .2 0) (0 .3 4) s fu nd 11 .2 1 9. 99 10 .2 6 10 .2 0 8. 92 7. 14 11 .8 2 10 .4 8 10 .7 7 10 .7 1 9. 32 7. 40 (0 .5 9) (0 .7 9) (0 .5 3) (0 .5 7) (0 .4 7) (0 .5 3) (0 .5 0) (0 .7 0) (0 .4 5) (0 .4 9) (0 .4 2) (0 .5 0) i fu nd 13 .4 4 11 .9 4 12 .3 1 12 .2 3 10 .7 0 8. 54 14 .2 8 12 .6 1 13 .0 1 12 .9 2 11 .2 5 8. 90 (1 .2 7) (0 .7 7) (1 .1 5) (1 .0 7) (1 .0 0) (0 .5 8) (1 .5 6) (0 .9 7) (1 .3 9) (1 .2 9) (1 .1 8) (0 .6 7) r ea l 10 .9 7 7. 85 11 .4 2 8. 06 (2 .9 9) (2 .1 6) (3 .1 2) (2 .1 9) e m er . 8. 45 5. 87 8. 88 6. 09 (0 .2 4) (0 .2 2) (0 .3 3) (0 .1 8) c om . 8. 99 6. 28 9. 43 6. 49 (0 .8 1) (0 .4 3) (1 .0 0) (0 .5 2) in t. b 20 .4 0 16 .3 2 21 .1 9 16 .8 1 (0 .3 4) (0 .5 0) (0 .5 4) (0 .3 9) t hi s ta bl e co nt in ue s t ab le 5d re po rt in g th e av er ag e po rt fo lio w ei gh ts (% ) of ri sk y as se ts in th e ou tof -s am pl e ri sk pa ri ty po rt fo lio .t he in pu tv ar ia bl es ar e es tim at ed us in g th e fa m afr en ch 3fa ct or m od el an d th e b ay es ia n sh ri nk ag e m et ho d. t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. n um be rs in pa re nt he se s de no te th e as so ci at ed st an da rd de vi at io ns of th e op tim al w ei gh ts ov er th e te st in g pe ri od . c a pm � ca pi ta la ss et pr ic in g m od el ;t sp � th ri ft sa vi ng s pl an ;r ea l� re al es ta te ;e m er � em er gi ng m ar ke t; c om � co m m od ity ;i nt .b � in te rn at io na l bo nd . 70 s. shen, j.a. turner / financial services review 27 (2018) 47-81 also implicitly showing the noise of estimating expected returns. the results demonstrate the attractiveness of the f-fund because it dominates all asset allocation strategies for all estimation methods. real estate fund is also attractive under the mean-variance portfolio, but is less attractive for the minimum-variance strategy, indicating the high riskiness of the reits. the international bond index fund is also attractive, with almost 20% of average portfolio weight, especially under the minimum variance and risk parity strategies. both the minimum variance strategy and risk parity strategy are stable over time with much lower standard deviations. this result is because for both strategies, asset volatility is the only input parameter, and the volatility term is less volatile compared with the first moment. in table 6a, we present the out-of-sample performance of the optimal mean-variance portfolio under various estimation methods and for different portfolio combinations. without the borrowing constraint, the optimal mean-variance portfolio can generate substantially high returns but also substantially high volatility, which leads to a lower sharpe ratio compared with the constraint case. under the mean-variance strategy, adding extra funds in most of the cases can enhance the portfolio returns by approximately 2.5%, while the risk of the portfolio increases by more than about 5%. if an investor only cares about portfolio return regardless of the underlying risk, having additional investment vehicles can indeed bring much higher expected returns even with high investment fees. however, the marginal gain from the risk-return trade-off is negligible. therefore, unless there is a dramatic increase in the portfolio returns, non-tsp investors can benefit from investing in a larger class of assets even with much higher investment fees. however, with the borrowing constraint, a dramatic increase in portfolio returns is less likely to happen. hence, the chance to benefit from rolling-over from tsp plan is also small. table 6b reports the out-of-sample performance under the minimum variance strategy, risk parity strategy and 1/n naïve strategy. in many cases, additional investment options increase the portfolio returns by less than 0.5% and can also reduce the portfolio volatility by 2–3%. the result demonstrates that there is still limited but significant room for further diversification. however, the marginal increase of portfolio returns is too small to cover the high investment expense for the non-tsp investors. as a result, almost none of the sharpe ratios of the non-tsp scenarios beat the tsp benchmark scenario. as a robustness check, in table 7 we present the out-of-sample analysis in three subperiods. only during the subperiod 2001–2008 (table 7b), we observe significant improvement by adding extra funds under various asset allocation strategies. the out of sample performance using capm and fama-french estimation approaches is more stable than other estimation methods. the mean-variance strategy (with constraint) persistently presents a significant improvement of the sharpe ratio for tsp portfolios complemented with additional funds, especially using factor-model based estimation approaches across subsamples. therefore, table 7 shows that rolling over from the tsp plan may give investors an opportunity to enhance their portfolio performance in terms of diversification benefit even with much higher administration fees. however, such benefit highly relies on the asset allocation strategy, market timing, estimation approach of input variables as well as investors’ risk preference. in other words, this result indicates that a larger number of investment options 71s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 6a o ut -o fsa m pl e po rt fo lio be ne fit s of ad di tio na l fu nd s us in g m ea nva ri an ce as se t al lo ca tio n st ra te gi es m et ho d sa m pl e m om en ts c a pm fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith ou t co ns tr ai nt t sp 15 .8 0 21 .0 8 0. 62 18 .8 1 19 .0 2 0. 89 18 .2 3 18 .9 0 0. 87 12 .0 5 29 .9 5 0. 32 � r ea l es ta te 18 .7 6� 22 .8 0 0. 70 � 22 .1 3� 21 .1 9 0. 97 � 21 .3 9� 21 .0 2 0. 94 � 14 .7 1� 30 .9 9 0. 40 � � e m er gi ng 16 .4 0� 23 .4 6 0. 58 17 .7 2 20 .3 1 0. 79 16 .9 9 20 .1 1 0. 77 13 .0 8� 30 .9 5 0. 34 � � c om m od ity 16 .0 6� 22 .3 3 0. 60 18 .8 6� 20 .6 0 0. 85 18 .3 6� 20 .4 6 0. 83 11 .9 0 30 .7 4 0. 30 � in te rn . bo nd 15 .7 0 21 .7 2 0. 60 20 .4 5� 20 .5 5 0. 92 � 19 .7 8� 20 .4 0 0. 90 � 11 .0 6 30 .4 2 0. 27 � a ll 20 .5 4� 26 .6 8 0. 71 � 22 .3 5� 24 .7 9 0. 92 � 21 .5 5� 24 .4 7 0. 90 � 15 .3 2� 33 .0 4 0. 39 � n on -t sp 19 .9 7� 26 .6 8 0. 69 � 21 .7 8� 24 .7 9 0. 89 20 .9 8� 24 .4 7 0. 88 � 13 .8 3� 33 .0 6 0. 34 � m ea nva ri an ce as se t al lo ca tio n st ra te gy w ith co ns tr ai nt t sp 6. 73 5. 03 0. 77 6. 94 5. 31 0. 77 6. 92 5. 29 0. 76 6. 61 5. 19 1. 23 � r ea l es ta te 8. 11 � 6. 92 0. 76 8. 53 � 7. 14 0. 79 � 8. 50 � 7. 10 0. 79 � 7. 75 � 6. 82 1. 10 � e m er gi ng 7. 59 � 6. 37 0. 73 8. 19 � 7. 00 0. 74 8. 14 � 6. 96 0. 74 7. 37 � 6. 38 1. 13 � c om m od ity 8. 06 � 6. 98 0. 73 8. 51 � 7. 45 0. 75 � 8. 48 � 7. 44 0. 75 7. 71 � 6. 84 1. 10 � in te rn . bo nd 7. 09 � 5. 58 0. 74 7. 27 � 5. 67 0. 77 7. 24 � 5. 65 0. 77 � 6. 89 � 5. 62 1. 19 � a ll 9. 98 � 8. 99 0. 78 � 10 .7 9� 9. 25 0. 84 � 10 .7 4� 9. 25 0. 84 � 9. 30 � 8. 66 1. 04 n on -t sp 9. 41 � 8. 99 0. 71 10 .2 2� 9. 25 0. 78 � 10 .1 7� 9. 25 0. 77 � 8. 73 � 8. 66 0. 66 t hi s ta bl e di sp la ys th e ou tof -s am pl e po rt fo lio pe rf or m an ce fo llo w in g m ea nva ri an ce as se ta llo ca tio n st ra te gi es w ith an d w ith ou tb or ro w in g co ns tr ai nt fo r th e t sp 5fu nd po rt fo lio an d po rt fo lio s co m pl em en te d w ith ad di tio na l fu nd s du ri ng th e pe ri od fr om ja nu ar y 19 93 to a pr il 20 15 . t he re su lts ar e ne t of ad m in is tr at io n co st w hi ch is 0. 03 % fo r th e t sp pa rt ic ip an ts an d 0. 6% fo r th e no nt sp in ve st or s. pa ra m et er s ar e es tim at ed us in g fo ur di ff er en tm et ho ds .t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. im pr ov em en ts in co m pa ri so n to th e t sp be nc hm ar k po rt fo lio ar e hi gh lig ht ed w ith � . c a pm � ca pi ta l as se t pr ic in g m od el ; t sp � th ri ft sa vi ng s pl an ; in te rn . � in te rn at io na l. 72 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 6b o ut -o fsa m pl e po rt fo lio be ne fit s of ad di tio na l fu nd s us in g m in im um -v ar ia nc e, ri sk pa ri ty , an d 1/ n na ïv e as se t al lo ca tio n st ra te gi es m et ho d sa m pl e m om en ts c a pm fa m afr en ch 3 fa ct or b ay es ia n sh ri nk ag e m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m ea n st an da rd de vi at io n sh ar pe m in im um -v ar ia nc e as se t al lo ca tio n st ra te gy t sp 5. 03 3. 41 1. 04 5. 06 3. 38 1. 06 5. 02 3. 40 1. 05 5. 11 3. 78 0. 97 � r ea l es ta te 5. 24 � 3. 29 � 1. 16 � 5. 30 � 3. 31 � 1. 15 � 5. 26 � 3. 33 � 1. 13 � 5. 30 � 3. 69 � 1. 06 � � e m er gi ng 5. 19 � 3. 39 � 1. 10 � 5. 22 � 3. 35 � 1. 12 � 5. 18 � 3. 37 � 1. 10 � 5. 26 � 3. 75 � 1. 02 � � c om m od ity 5. 05 � 3. 37 � 1. 06 � 5. 13 � 3. 35 � 1. 09 � 5. 12 � 3. 36 � 1. 08 � 5. 14 � 3. 76 � 0. 98 � � in te rn . bo nd 4. 82 3. 35 � 1. 00 4. 93 3. 16 � 1. 10 � 4. 89 3. 17 � 1. 08 � 4. 91 3. 74 � 0. 93 � a ll 5. 22 � 3. 17 � 1. 20 � 5. 32 � 3. 05 � 1. 25 � 5. 29 � 3. 08 � 1. 23 � 5. 26 � 3. 56 � 1. 10 � n on -t sp 4. 65 3. 17 � 1. 02 4. 75 3. 05 � 1. 07 � 4. 72 3. 08 � 1. 05 4. 69 3. 56 � 0. 93 r is k pa ri ty as se t al lo ca tio n st ra te gy t sp 6. 36 6. 52 0. 71 6. 37 4. 46 1. 07 6. 37 4. 54 1. 05 6. 49 7. 06 0. 68 � r ea l es ta te 6. 84 � 7. 12 0. 70 6. 85 � 4. 51 1. 15 � 6. 85 � 4. 60 1. 13 6. 99 � 7. 67 0. 67 � e m er gi ng 6. 92 � 7. 66 0. 68 6. 92 � 4. 60 1. 15 � 6. 92 � 4. 72 1. 13 7. 09 � 8. 22 0. 66 � c om m od ity 6. 52 � 6. 73 0. 71 6. 53 � 4. 58 1. 07 6. 53 � 4. 66 1. 06 6. 66 � 7. 28 0. 68 � in te rn . bo nd 5. 94 5. 75 � 0. 76 � 5. 94 3. 99 � 1. 10 � 5. 94 4. 07 � 1. 08 6. 03 6. 20 � 0. 72 � � a ll 6. 83 � 7. 11 0. 74 � 6. 83 � 4. 17 � 1. 25 � 6. 83 � 4. 29 � 1. 23 6. 95 � 7. 61 0. 71 � n on -t sp 6. 26 7. 11 0. 65 6. 26 � 4. 17 � 1. 11 � 6. 26 4. 29 � 1. 10 6. 38 7. 61 0. 63 1/ n na iv e as se t al lo ca tio n st ra te gy t sp 8. 19 12 .1 6 0. 53 8. 19 7. 75 0. 84 8. 19 7. 91 0. 82 8. 19 12 .3 0 0. 52 � r ea l es ta te 8. 71 � 12 .4 8 0. 53 8. 71 � 7. 59 � 0. 89 � 8. 71 � 7. 76 � 0. 87 8. 71 � 12 .6 5 0. 53 � � e m er gi ng 9. 14 � 13 .9 8 0. 53 9. 14 � 7. 99 0. 93 � 9. 14 � 8. 20 0. 91 9. 14 � 14 .1 2 0. 53 � � c om m od ity 8. 02 11 .8 0� 0. 54 � 8. 02 7. 81 0. 83 8. 02 7. 93 0. 82 8. 02 11 .9 8� 0. 53 � � in te rn . bo nd 7. 41 10 .0 4� 0. 57 � 7. 41 6. 47 � 0. 89 � 7. 41 6. 62 � 0. 87 � 7. 41 10 .2 0� 0. 56 � � a ll 8. 51 � 11 .6 6� 0. 60 � 8. 51 � 6. 64 � 1. 04 � 8. 51 � 6. 82 � 1. 03 � 8. 51 � 11 .8 7� 0. 59 � n on -t sp 7. 94 11 .6 6� 0. 55 � 7. 94 6. 64 � 0. 95 � 7. 94 6. 82 � 0. 95 � 7. 94 11 .8 7� 0. 54 � t hi s ta bl e di sp la ys th e ou t of sa m pl e po rt fo lio pe rf or m an ce fo llo w in g va ri ou s al te rn at iv e as se t al lo ca tio n st ra te gi es in cl ud in g th e m in im um va ri an ce ,t he ri sk pa ri ty an d th e na ïv e st ra te gi es du ri ng th e pe ri od fr om ja nu ar y 19 93 to a pr il 20 15 . t he re su lts ar e ne t of ad m in is tr at io n co st w hi ch is fo r th e t sp pa rt ic ip an ts an d fo r th e no nt sp in ve st or s. pa ra m et er s ar e es tim at ed us in g fo ur di ff er en t m et ho ds . t he es tim at io n w in do w le ng th is 12 0 m on th s an d th e te st in g w in do w ha s th e le ng th of 16 8 m on th s. im pr ov em en t in co m pa ri so n to th e t sp be nc hm ar k po rt fo lio ar e hi gh lig ht ed w ith � . c a pm � ca pi ta l as se t pr ic in g m od el ; t sp � th ri ft sa vi ng s pl an ; in te rn . � in te rn at io na l. 73s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 7a o ut -o fsa m pl e an al ys is fo r su bpe ri od fr om ja nu ar y 19 93 to ja nu ar y 20 01 m ea nva ri an ce as se t al lo ca tio n st ra te gy m in im um -v ar ia nc e as se t al lo ca tio n st ra te gy sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e t sp 1. 19 5 1. 36 0 1. 32 9 1. 01 0 � 0. 16 4 1. 03 9 0. 94 5 � 0. 30 5 � r ea l es ta te 1. 26 2� 1. 44 7� 1. 41 5� 1. 05 3� � 0. 11 8� 1. 19 7� 1. 03 8� � 0. 29 2� � e m er gi ng 1. 19 6� 1. 36 6� 1. 33 4� 1. 00 8 � 0. 49 9 0. 89 0 0. 82 7 � 0. 70 9 � c om m od ity 1. 22 4� 1. 38 8� 1. 35 9� 1. 02 1� 0. 00 5� 0. 95 9 0. 86 5 � 0. 10 3� � in te rn . bo nd 1. 22 6� 1. 39 0� 1. 35 5� 1. 02 4� � 0. 13 5 0. 84 7 0. 80 8 � 0. 25 1� � a ll 1. 32 3� 1. 50 6� 1. 47 1� 1. 09 0� � 0. 36 4 0. 81 6 0. 71 8 � 0. 68 4 n on -t sp 1. 27 0� 1. 44 5� 1. 41 1� 1. 04 1� � 0. 56 8 0. 61 3 0. 52 3 � 0. 87 0 r is k pa ri ty as se t al lo ca tio n st ra te gy 1/ n na iv e as se t al lo ca tio n st ra te gy sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e t sp 0. 70 4 1. 07 1 1. 00 1 0. 61 7 0. 59 8 0. 92 5 0. 86 8 0. 54 9 � r ea l es ta te 0. 74 8� 1. 19 0� 1. 07 5� 0. 63 8� 0. 64 1� 1. 04 9� 0. 94 5� 0. 56 5� � e m er gi ng 0. 42 2 0. 69 9 0. 63 0 0. 37 0 0. 21 7 0. 35 3 0. 31 9 0. 22 1 � c om m od ity 0. 60 1 0. 93 3 0. 87 4 0. 53 5 0. 47 5 0. 75 8 0. 69 4 0. 45 7 � in te rn . bo nd 0. 48 9 0. 73 3 0. 70 9 0. 42 6 0. 42 7 0. 66 8 0. 64 6 0. 39 5 � a ll 0. 30 0 0. 53 8 0. 44 3 0. 24 9 0. 14 3 0. 28 5 0. 20 1 0. 14 3 n on -t sp 0. 20 0 0. 37 2 0. 28 9 0. 15 9 0. 07 2 0. 16 7 0. 09 0 0. 07 6 t hi s ta bl e di sp la ys th e ou tof -s am pl e po rt fo lio sh ar pe ra tio s fo r di ff er en ta ss et al lo ca tio n st ra te gi es an d di ff er en tp ar am et er es tim at io n m et ho ds fo r th e t sp 4ri sk yfu nd po rt fo lio an d po rt fo lio s co m pl em en te d w ith ad di tio na l fu nd s du ri ng th e su b pe ri od fr om ja nu ar y 19 93 to ja nu ar y 20 01 . t he re su lts ar e ne t of ad m in is tr at io n co st .t he le ng th of th e es tim at io n w in do w is 30 m on th s. im pr ov em en ts in co m pa ri so n to th e t sp be nc hm ar k po rt fo lio ar e hi gh lig ht ed w ith � . c a pm � ca pi ta l as se t pr ic in g m od el ; t sp � th ri ft sa vi ng s pl an ; in te rn . � in te rn at io na l. 74 s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 7b o ut -o fsa m pl e an al ys is fo r su bpe ri od fr om fe br ua ry 20 01 to fe br ua ry 20 08 m ea nva ri an ce as se t al lo ca tio n st ra te gy m in im um -v ar ia nc e as se t al lo ca tio n st ra te gy sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e t sp 1. 42 3 1. 53 9 1. 53 7 1. 23 9 0. 59 2 0. 83 7 0. 86 6 0. 49 6 � r ea l es ta te 1. 49 7� 1. 71 2� 1. 71 8� 1. 28 3� 0. 37 5 0. 89 8� 0. 91 3� 0. 33 1 � e m er gi ng 1. 43 4� 1. 57 0� 1. 55 8� 1. 25 3� 0. 34 3 1. 02 6� 1. 01 9� 0. 19 1 � c om m od ity 1. 53 4� 1. 60 5� 1. 60 0� 1. 31 2� 0. 71 1� 0. 96 0� 0. 97 3� 0. 59 3� � in te rn . bo nd 1. 55 9� 1. 67 8� 1. 67 6� 1. 33 0� 0. 82 5� 1. 03 2� 1. 06 2� 0. 67 9 � a ll 1. 76 0� 1. 90 5� 1. 89 7� 1. 47 3� 0. 40 5 1. 35 0� 1. 30 4� 0. 25 4 n on -t sp 1. 49 4� 1. 85 6� 1. 84 8� 1. 15 5 0. 18 2 1. 14 8� 1. 10 1� 0. 05 0 r is k pa ri ty as se t al lo ca tio n st ra te gy 1/ n na iv e as se t al lo ca tio n st ra te gy sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e t sp 0. 71 0 0. 95 9 0. 94 8 0. 67 2 0. 57 8 0. 87 4 0. 84 1 0. 56 3 � r ea l es ta te 0. 72 8� 1. 03 2� 1. 02 8� 0. 68 9� 0. 61 2� 0. 93 2� 0. 91 4� 0. 58 8� � e m er gi ng 0. 90 1� 1. 41 1� 1. 37 1� 0. 83 9� 0. 89 6� 1. 52 0� 1. 44 6� 0. 85 0� � c om m od ity 1. 03 2� 1. 28 4� 1. 27 2� 0. 95 5� 1. 07 0� 1. 32 9� 1. 31 1� 1. 00 4� � in te rn . bo nd 0. 85 6� 1. 15 2� 1. 15 3� 0. 80 5� 0. 66 0� 1. 00 1� 0. 98 5� 0. 64 8� � a ll 1. 17 7� 1. 78 6� 1. 76 6� 1. 07 9� 1. 21 9� 1. 83 1� 1. 79 7� 1. 12 2� n on -t sp 1. 07 5� 1. 62 2� 1. 60 5� 0. 99 0� 1. 14 7� 1. 72 1� 1. 68 9� 1. 06 0� t hi s ta bl e co nt in ue s t ab le 5a ,s ho w in g th e ou tof -s am pl e po rt fo lio sh ar pe ra tio s fo r di ff er en ta ss et al lo ca tio n st ra te gi es an d di ff er en tp ar am et er es tim at io n m et ho ds du ri ng th e su b pe ri od fr om fe br ua ry 20 01 to fe br ua ry 20 08 .i m pr ov em en ts in co m pa ri so n to th e t sp be nc hm ar k po rt fo lio ar e hi gh lig ht ed w ith � . c a pm � ca pi ta l as se t pr ic in g m od el ; t sp � th ri ft sa vi ng s pl an ; in te rn . � in te rn at io na l. 75s. shen, j.a. turner / financial services review 27 (2018) 47-81 t ab le 7c o ut -o fsa m pl e an al ys is fo r su bpe ri od fr om m ar ch 20 08 to a pr il 20 15 m ea nva ri an ce as se t al lo ca tio n st ra te gy m in im um -v ar ia nc e as se t al lo ca tio n st ra te gy sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e t sp 1. 35 6 1. 54 4 1. 52 0 1. 25 3 1. 75 8 1. 39 2 1. 37 9 1. 33 0 � r ea l es ta te 1. 34 4 1. 65 4� 1. 61 2� 1. 22 8 1. 89 2� 1. 44 0� 1. 42 2� 1. 45 1� � e m er gi ng 1. 36 2� 1. 57 2� 1. 54 8� 1. 25 8� 2. 11 9� 1. 37 4 1. 35 3 1. 54 5� � c om m od ity 1. 37 1� 1. 60 0� 1. 57 5� 1. 26 4� 1. 57 6 1. 28 5 1. 20 9 1. 10 6 � in te rn . bo nd 1. 37 1� 1. 55 3 1. 52 7� 1. 26 5� 1. 56 3 1. 18 6 1. 18 7 1. 15 3 � a ll 1. 37 8� 1. 73 8� 1. 69 4� 1. 24 9 2. 50 2� 1. 11 2 1. 03 4 1. 72 6� n on -t sp 1. 70 2� 1. 67 2� 1. 62 9� 1. 01 7 2. 19 7� 0. 88 1 0. 80 2 1. 49 4� r is k pa ri ty as se t al lo ca tio n st ra te gy 1/ n na iv e as se t al lo ca tio n st ra te gy sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e sa m pl e m om en ts c a pm fa m afr en ch b ay es ia n sh ri nk ag e t sp 1. 16 6 1. 62 1 1. 58 4 1. 07 4 0. 85 3 1. 30 6 1. 25 5 0. 84 0 � r ea l es ta te 1. 08 1 1. 68 2� 1. 62 6� 1. 00 6 0. 83 9 1. 37 6� 1. 32 0� 0. 82 1 � e m er gi ng 0. 91 1 1. 44 3 1. 38 1 0. 85 9 0. 63 6 1. 05 9 0. 99 3 0. 64 1 � c om m od ity 0. 81 6 1. 28 1 1. 22 4 0. 76 3 0. 49 9 0. 85 2 0. 78 6 0. 51 8 � in te rn . bo nd 1. 06 3 1. 40 3 1. 35 0 0. 97 4 0. 81 9 1. 22 8 1. 18 2 0. 81 0 � a ll 0. 64 8 1. 14 8 1. 07 3 0. 61 7 0. 44 3 0. 83 4 0. 77 3 0. 45 9 n on -t sp 0. 55 6 0. 99 0 0. 92 4 0. 53 2 0. 39 1 0. 73 9 0. 68 5 0. 40 8 t hi s ta bl e co nt in ue s t ab le 5a an d 5b ,r ep or tin g th e ou tof -s am pl e po rt fo lio pe rf or m an ce ne to f ad m in is tr at io n co st du ri ng th e su b pe ri od fr om m ar ch 20 08 to a pr il 20 15 . fo r di ff er en t as se t al lo ca tio n st ra te gi es an d di ff er en t pa ra m et er es tim at io n m et ho ds . im pr ov em en ts in co m pa ri so n to th e t sp be nc hm ar k po rt fo lio ar e hi gh lig ht ed w ith � . c a pm � ca pi ta l as se t pr ic in g m od el ; t sp � th ri ft sa vi ng s pl an ; in te rn . � in te rn at io na l. 76 s. shen, j.a. turner / financial services review 27 (2018) 47-81 does not necessarily result in a better performance if the investor fails to obtain correct input parameters or does not follow the optimal asset allocation strategy. in summary, three factors affect our conclusions—portfolio volatility, gross rates of the portfolio return, and fees. the empirical analysis shows the tsp 5-fund plan is fully diversified in term of reducing volatility. however, one can still make a small improvement in net investment returns by including additional instruments if the newly added funds charge low investment fees. this additional benefit fades away when the investment fee is roughly 60 basis points. thus, although rolling over from the tsp plan allows an investor to have more investment opportunities, the marginal cost of investing cancels out the marginal benefit. if the advice to roll over to an ira also involves an ongoing investment management fee of 60 basis points or more, or if the investment fees in total are roughly 65 basis points or higher, the advice will lead to reduced net returns. 7. policy results the main result concerns the quality of advice that pension participants are receiving. the advice to roll over from the tsp, or from 401(k) or 403(b) plans with more investment options, for better diversification is generally not valid. thus, we document that many participants in the tsp, and in defined contribution plans with more investment options, are receiving bad advice that is costly. the advice to roll over can result in present-value losses of thousands of dollars (turner, klein, and stein 2016). it should be noted that some pension participants, particularly those in defined contribution plans provided by small employers, are in plans with relatively high fees, and they can reduce their fees by rolling over to a low-fee ira provider. in addition, for some participants, financial advisers may add value by keeping them from engaging in panic selling when there is a market downturn. in addition, in some circumstances, participants may benefit from a partial roll over, in particular when the disbursement options are limited within the pension plan, which has been the case for the tsp. more generally, our results indicate that for participants in large 401(k) plans, which typically have lower fees than small plans or iras, the advice to roll over for better diversification is based on a true statement that it may be possible to obtain better diversification, but ignores the costs. because the improvement in diversification is generally relatively small, the increase in costs from rolling over to an ira outweighs the improvement in diversification. thus, bad advice is supported by bad analysis. in particular, the analysis focuses on only one aspect of the situation, in this case portfolio diversification, without adequately considering costs. the second main result is that pension plans can be well diversified with a relatively small number of funds. for example, with its five basic investment options, the tsp is well diversified. adding an additional four investment options results in slightly better sharpe ratios using some investment strategies. thus, this result suggests that defined contribution plans and other funds of funds, such as target date plans, can provide their participants the opportunity to have well diversified funds while still retaining the simple choice menu of a small number of funds that are themselves well diversified and that are selected to work well together in a portfolio. this result has relevance for litigation as to the adequacy of the investment options offered by pension plans, 77s. shen, j.a. turner / financial services review 27 (2018) 47-81 where some plaintiffs have charged imprudent management based on a small number of investment options being offered (fisch and wilkinson-ryan, 2016). in addition, the results suggest that some financial advisers and some providers of financial products may use strategic complexity to impress naïve investors, recommending or providing complex investment portfolios when simpler portfolios may be superior, once fees are taken into account. 8. conclusions mounting evidence documents that some financial advisers with conflicts of interest provide advice that is costly to their clients. nevertheless, these advisers presumably have arguments that they use to persuade their clients to follow their advice. we conclude that bad advice is sometimes supported by bad analysis. this paper analyzes one such argument. in doing so, it investigates the hypothesis that some advisers with a conflict of interest, in communicating with their clients, focus on the benefits of their advice without weighing the marginal benefits against the marginal costs. we argue that it is psychologically less costly to make a true statement that is incomplete than to make a false statement. many advisers, however, may simply be following the industry standard advice. we present evidence against the industry standard advice concerning 401(k) plan rollovers to have more investment options. a plan with as few as five well diversified investment options can provide adequate diversification, while leaving an employer-provided plan and rolling over to an ira for more options will generally result in higher fees. that is not always the case because small plans tend to have relatively high fees, and sophisticated investors can find low-fee investment options outside of employer-provided pension plans. however, the typical 401(k) plan has far more than five investment options, so our results suggest that the industry standard advice to roll over for greater diversification is generally not valid. our results suggest that some advisers and financial service providers may engage in strategic complexity in the portfolios they provide to impress naïve clients as to their complex diversification. acknowledgments we thank two anonymous referees and the editor for valuable suggestions and comments. we express our appreciation to kim weaver of the tsp for providing us with rate of return data for the g fund. we have received helpful comments from louis raes at the 2016 canadian economic association meetings, from participants at the 2016 behavioral finance working group conference at queen mary university in london, and from jill fisch. 78 s. shen, j.a. turner / financial services review 27 (2018) 47-81 references akerlof, g. a., & shiller, r. j. (2015). phishing for phools: the economics of manipulation & deception. princeton, nj: princeton university press. anderson, r. m., bianchi, s. w., & goldberg, l. r. (2012). will my risk parity strategy outperform? financial analysts journal, 68, 75–93. anspach, d. (2016). why average investors earn below average market returns: investor overreactions cause poor historical returns. the balance. (available at https://www.thebalance.com/why-average-investors-earn-belowaverage-market-returns-2388519). aon hewitt. (2014). 2014 tsp withdrawal survey results. lincolnshire, il: aon hewitt. ashraf, n., karlan, d., & yin, w. (2006). tying odysseus to the mast; evidence from a commitment savings product in the philippines. quarterly journal of economics, 121, 635–667. benartzi, s., & thaller, r. h. (2001). naive diversification strategies in defined contribution saving plans. american economic review, 91, 79–98. bessler, w., & wolff, d. (2015). do commodities add value in multi-asset portfolios? an out-of-sample analysis for different investment strategies. journal of banking & finance, 60, 1–20. bessler, w., opfer, h., & wolff, d. (2017). multi-asset portfolio optimization and out-of-sample performance: an evaluation of black–litterman, mean-variance, and naïve diversification approaches. the european journal of finance, 23, 1–30. betterment. (2016). four myths about diversification. (available at https://www.betterment.com/resources/4myths-about-diversification/). brightscope & investment company institute. (2014). the brightscope/ici defined contribution plan profile: a close look at 401(k) plans. san diego, ca: brightscope and washington, dc: investment company institute. burns, w. l., & epley, d. r. (1982). the performance of portfolios of reits � stocks. journal of portfolio management, 8, 37–42. carosa, c. (2014). exclusive interview: quinn curtis reveals true intent of the curtis/ayres 401k fee paper. fiduciary news, march 18. (available at http://fiduciarynews.com/2014/03/exclusive-interview-quinn-curtisreveals-true-intent-of-the-curtisayres-401k-fee-paper/). chen, z., & tobias, g. (2016). persistent bias in advice-giving. paper presented at the 2016 behavioural finance workshop, queen mary university, london. choi, j. j., laibson, d., & madrian, b. c. (2007). $100 bills on the sidewalk: suboptimal investment in 401(k) plans. nber wp 11554. choi, j. j., laibson, d., & madrian, b. c. (2010). why does the law of one price fail? an experiment on index mutual funds. the review of financial studies, 23, 1405–1432. christoffersen, s. e. k., evans, r., & musto, d. k. (2013). what do consumers’ fund flows maximize? evidence from their brokers’ incentives. journal of finance, 68, 201–235. doi: 10.2139/ssrn.1393289. (available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id�1393289). copeland, c. (2013). ira asset allocation, 2011. ebri notes, 34, 8–22. (available at http://www.ebri.org/pdf/ notespdf/ebri_notes_10_oct-13_retsvgs-iras.pdf). council of economic advisers. (2015). the effects of conflicted investment advice on retirement savings. washington, dc: executive office of the president of the united states. (available at http://www.whitehouse. gov/sites/default/files/docs/cea_coi_report_final.pdf). daskalaki, c., & skiadopoulos, g. (2011). should investors include commodities in their portfolios after all? new evidence. journal of banking & finance, 35, 2606–2626. demiguel, v., garlappi, l., & uppal, r. (2009). optimal versus naive diversification: how inefficient is the 1/n portfolio strategy? review of financial studies, 22, 1915–1953. di tella, r., perez-truglia, r., babino, a., & sigman, m. (2015). conveniently upset: avoiding altruism by distorting beliefs about others’ altruism. american economic review, 11, 3416–3442. duarte, f., & hastings, j. (2009). fettered consumers and sophisticated firms: evidence from mexico’s privatized social security system. manuscript, yale university department of economics. 79s. shen, j.a. turner / financial services review 27 (2018) 47-81 dvorak, t. (2015). do 401(k) plan advisors take their own advice? journal of pension economics and finance, 14, 55–75. elton, e. j., gruber, m. j., & blake, c. r. (2006). the adequacy of investment choices offered by 401 (k) plans. journal of public economics 90.67, 1299–1314. ennis, r. m., & burik, p. (1991). pension fund real estate investment under a simple equilibrium pricing model. financial analysts journal, 47, 20–30. fama, e. f. (1976). foundations of finance: portfolio decisions and securities prices. new york, ny: basic books. festinger, l. (1957). a theory of cognitive dissonance. stanford, ca: stanford university press. fidelity. (2016). roll over your old 401(k). (available at https://www.fidelity.com/go/401k-rollover-hub/ ?imm_pid�700000001008509&immid�100063&imm_eid�e5548494616&gclid�cng1jpj57s0cfuro6 qodb1akjq&gclsrc�ds). fisch, j. e., & wilkinson-ryan, t. (2014). why do retail investors make costly mistakes? an experiment on mutual fund choice. university of pennsylvania law review, 162, 605–647. (available at http://papers. ssrn.com/sol3/papers.cfm?abstract_id�2086766). giliberto, s. m. (1993). measuring real estate returns: the hedged reit index. journal of portfolio management, 19, 94–99 . goldreich, d., & halaburda, h. (2011). when smaller menus are better: variability in menu-setting ability. harvard business school working paper no. 11-086. cambridge, ma: harvard university. (available at http://www.hbs.edu/research/pdf/11-086.pdf). hastings, j., & tejeda-ashton, l. (2008). financial literacy, information and demand elasticity: survey and experimental evidence from mexico. nber wp 14538. hastorf, a. h., & cantril, h. (1954). they saw a game: a case study. journal of abnormal and social psychology, 49, 129–134. huberman, g., & kandel, s. (1987). mean-variance spanning. the journal of finance, 42, 873–888. investment company institute. (2016a). retirement assets total $24.5 trillion in second quarter 2016. (available at https://ici.org/research/stats/retirement/ret_16_q2). investment company institute. (2016b). the role of iras in u.s. households’ saving for retirement, 2015. ici research perspective, 22, february. (available at https://www.ici.org/pdf/pe_r22-01.pdf). investment company institute. (2017). investment company 2017 factbook. (available at http://www. icifactbook.org/). iyengar, s. s., huberman, g., & jiang, w. (2004). how much choice is too much? contributions to 401(k) retirement plans. in o. s. mitchell & s. p. utkus (eds.), pension design and structure: new lessons from behavioral finance (pp. 83–95). oxford: oxford university press. iyengar, s. s., & lepper, m. r. (2000). when choice is demotivating: can one desire too much of a good thing? journal of personality and social psychology, 79, 995–1006. (available http://psycnet.apa.org/index. cfm?fa�buy.optiontobuy&id�2000-16701-012). james, w., & stein, c. (1961). estimation with quadratic loss. in proceedings of the fourth berkeley symposium on mathematical statistics and probability (vol. 1, no. 1961, pp. 361–379). kan, r., & zhou. g. (2012). tests of mean-variance spanning. annals of economics & finance 13.1. (available at http://aeconf.com/articles/may2012/aef130105.pdf). ledoit, o., & wolf, m. (2003). improved estimation of the covariance matrix of stock returns with an application to portfolio selection. journal of empirical finance, 10, 603–621. li, k., sarkar, a., & wang, z. (2003). diversification benefits of emerging markets subject to portfolio constraints. journal of empirical finance, 10, 57–80. lusardi, a., & mitchell, o. s. (2011). financial literacy and planning: implications for financial wellbeing. nber wp 17078. (available at file:///c:/users/owner/downloads/financial�literacy�and� planning-��implications�for�retirement�wellbeing.pdf). madrian, b. c., & shea, d. (2001). the power of suggestion: inertia in 401(k) participation and savings behavior. quarterly journal of economics, 116, 1149–1187. markowitz, h. (1952). portfolio selection. journal of finance, 7, 77–91. 80 s. shen, j.a. turner / financial services review 27 (2018) 47-81 merton, r. c. (1980). on estimating the expected return on the market: an exploratory investigation. journal of financial economics, 8, 323–361. money magazine. (2015). three myths about diversification. august 24. (available at http://moneymag.com.au/ three-myths-diversification/). mullainathan, s., noeth, m., & schoar, a. (2012). the market for financial advice: an audit study. nber wp 17929. cambridge, ma: national bureau of economic research. (available at http://www.nber.org/papers/ w17929.pdf). munnell, a. h., aubrey, j-p., & crawford, c. v. (2015). investment returns: defined benefit vs. defined contribution plans. issue in brief no. 15–21, december. (available at http://crr.bc.edu/wp-content/uploads/ 2015/12/ib_15-21.pdf). stein, c. (1956). inadmissibility of the usual estimator for the mean of a multivariate normal distribution. stanford, ct: stanford university. tang, n., mitchell, o. s., mottola, g. r., & utkus, s. p. (2010). the efficiency of sponsor and participant portfolio choices in 401 (k) plans. journal of public economics, 94, 1073–1085. tergesen, a. (2014). ditch a super-cheap 401(k) for an ira? sure, ira sellers say. the wall street journal, december 4. (available at http://blogs.wsj.com/totalreturn/2014/12/04/ditch-a-super-cheap-401k-for-an-irasure-ira-sellers-say/). thrift savings plan. (2015). summary of the thrift savings plan: your plan, your future. (available at http://www.cafc.uscourts.gov/sites/default/files/employment/benefits/tsp-summary.pdf). tiaa. (2016). ira: rollovers. (available at https://www.tiaa.org/public/offer/products/iras/rollovers). towers watson. (2014). pensions & investments/towers watson 300 analysis. september. (available at http:// www.towerswatson.com/en/insights/ic-types/survey-research-results/2014/09/the-worlds-300-largestpension-funds-year-end-2013). turner, j. a., klein, b. w., & stein, n. p. (2016). financial illiteracy meets conflicted advice: the case of thrift savings plan rollovers. journal of retirement, 3, 47–66. white, a. (2011). us federal employees’ plan embarks on giant investment tender. top 1000 funds, august 31. (available at http://www.top1000funds.com/profile/2011/08/31/us-federal-employees-plan-embarks-on-giantinvestment-tender/). world bank. (2014). pensions: data. (available at http://web.worldbank.org/wbsite/external/topics/ extsocialprotection/extpensions/0,contentmdk:23231994�menupk:8874064�pagepk:148956� pipk:216618�thesitepk:396253,00.html). vanguard. (2014). vanguard target retirement funds. malvern, pa: vanguard group. (available at https:// investor.vanguard.com/mutual-funds/target-retirement/#/). 81s. shen, j.a. turner / financial services review 27 (2018) 47-81 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 financial services review, 33(1) 67 resilient personality or financial resilience framework for coping with physical and mental health during the covid-19 pandemic megan mccoy,1 ives machiz,2 portia johnson,3 kenneth white,4 kimberly watkins,5 and chet bennetts6 abstract crises events such as the covid-19 pandemic can have a profound impact on consumers’ financial, physical, and mental health. this study explores the role of two resilience frameworks, namely the financial resilience framework and the resilient personality, in coping with physical and mental health challenges during the pandemic. the financial resilience framework encompasses economic resources, access to financial resources, financial knowledge and behavior, and social capital, while the resilient personality focuses on cognitive flexibility and the ability to tolerate ambiguity. the study aims to investigate whether these frameworks act as complements or substitutes in promoting resilience. glm anova is employed in this research to examine the effects of financial resilience and a resilient personality on physical and mental health outcomes. findings from this study indicate that both the financial resilience framework and resilient personality may contribute to one’s mental and physical health. however, the financial resilience framework is a stronger predictor of a positive self-assessment for both health factors than a resilient personality. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation mccoy, m., machiz, i., johnson, p., white, k., watkins, k., & bennetts, c. (2025). resilient personality or financial resilience framework for coping with physical and mental health during the covid-19 pandemic. financial services review, 33(1), 67-84. introduction the past two decades have witnessed a confluence of unprecedented global crises impacting not only financial well-being but also physical and mental health. since the 2008 1 corresponding author (meganmccoy@k-state.edu). kansas state university, manhattan, ks, usa 2 arizona state university, phoenix, az, usa. 3 auburn university, auburn, al, usa. 4 university of arizona, tucson, az, usa. 5 university of georgia, athens, ga, usa. 6 kansas state university, manhattan, ks, usa. housing bubble financial crisis, coupled with the emergence of cyber threats and geopolitical instability, individuals have faced heightened levels of uncertainty. these factors have had ripple effects on economic activity and public https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(1) 68 health (burgard & kalousova, 2015; shandler et al., 2023). most notably, the covid-19 pandemic strongly illustrates how a global health crisis can have cascading economic and psychological consequences. although the pandemic has ended, the ripple effects continue to be significant. in terms of physical health, there was not only increased morbidity and mortality stemming from covid-19 infections but also reduced access to healthcare services for other conditions (shadmi et al., 2020). this resulted in the subsequent years marked by an increase in chronic health conditions resulting from delayed or canceled medical appointments and decreased use of preventive services (bambra et al., 2021; patel et al., 2021). moreover, physical distancing measures and social isolation spawned reductions in physical activity, increased sedentary behaviors, and changes in dietary patterns, all of which can negatively impact physical health outcomes today (ahmed et al., 2021; meyer et al., 2020). additionally, the pandemic has also resulted in significant mental health challenges that have not subsided as quickly as covid-19 infection rates (gruber et al., 2023). research shows that the pandemic had a negative impact on mental health outcomes, with increased rates of anxiety, depression, and substance abuse symptomatology and diagnosis (gao et al., 2020; pfefferbaum & north, 2020). the emotional stress caused by the covid-19 pandemic (salari et al., 2020; tsamakis et al., 2020) was compounded by the financial stress that lockdown policies, unemployment, layoffs, and furloughs facilitated (coibion et al., 2020; crayne, 2020; faria-e-castro, 2021; kochhar, 2020; pappas, 2020; tran et al., 2020). the adverse outcomes of the pandemic did not have a homogenous effect on all. although many people struggled with the effects of the pandemic, others demonstrated resilience in light of stressors and have experienced an increase in financial, physical, and mental health during the endemic (prati & mancini, 2021). the purpose of this study is to identify protective factors that allowed some to be resilient in light of the covid-19 crisis. two key theories of resilience propose the protective factors necessary to be resilient: (a) the financial resilience framework (morrow, 2008; salignac et al., 2019) and (b) the resilient personality (asendorpfet al., 2001). the financial resilience framework is composed of four multidimensional components: economic resources, access to financial resources, financial knowledge and behavior, and social capital (morrow, 2008; salignac et al., 2019). it posits that individuals are best equipped to cope with adversity when they have knowledge of an adverse event and the resources to adapt successfully (morrow, 2008; salignac et al., 2019). in comparison, a resilient personality is characterized by cognitive flexibility and an ability to tolerate ambiguity well (asendorpf et al., 2001). an individual with this personality type has the inner vision, calmness, intelligence, maturity, and self-esteem needed to see a challenge not as a threat but as a time to gather internal resources to enact positive and effective resistance. individuals with a resilient personality are often seen as assertive, verbally expressive, energetic, personable, open-minded, smart, and self-confident (asendorpfet al., 2001). the overarching research question guiding this study is whether the financial resilience framework and the resilient personality are complements (e.g., have an additive effect) or substitutes (have the same impact and it is not cumulative) as they relate to resilience in one’s physical and mental health in light of covid-19. the examination used glm anova to analyze the effects of the financial resilience framework and a resilient personality on physical and mental health. through insights into resiliency, the hope is the findings will be generalizable to other stressors and crises. theoretically informed literature review biopsychosocial model the biopsychosocial model (engel, 1977) proposes that health and disease are determined by the interaction of biological, psychological, and social factors. the state of a person’s biological condition (e.g., organs, tissue, cells) is strongly influenced by psychological factors (e.g., cognition, emotions, motivation) and social interactions (e.g., society, community, family; serafino, 2011). the biopsychosocial model has mccoy et al. 69 been extensively adopted in medical research. findings from these studies have led to the development of a more comprehensive approach to healthcare to include mental health professionals, patients, families, and support systems. the theory was foundational to the study of covid-19’s short-term and long-term impacts across many domains. for example, kop (2021) found that attention to psychological and social factors is 74% higher in covid-19-related articles compared to all other physical healthrelated scientific articles published during the same period. the biopsychosocial model can be applied to the study of personal finance, as financial resilience and financial well-being are also complex constructs influenced by biological, psychological, and social factors (hughes, 2021). however, the relationship between biopsychosocial factors and financial well-being is largely unexplored (kannadhasan et al., 2016). the model has been applied to some aspects of personal finance, namely financial risk-taking and tolerance (fong, 2005; grable & joo, 2004; grable & webb, 2008; kannadhasan et al., 2016), and oniomania or compulsive overbuying (faber, 1992). additional evidence for the complex relationship of the elements of the biopsychosocial model was found in a longitudinal study of married couples (lee et al., 2021). using structural equation modeling (sem), the authors discovered that during the middle years of adulthood, the existence of family financial hardships was associated with reduced marital stability, which was linked to heightened mental health difficulties. moreover, the results reaffirmed the influential role of psychological distress in shaping subsequent physical health outcomes. specifically, anxiety symptoms reported by both husbands and wives during their early middle years contributed to the decline of their physical health in later adulthood (lee et al., 2021). while some results seem intuitive, further exploration of each nuanced component of the biopsychosocial model is in order. biological chou et al. (2016) established a connection between financial well-being, economic hardships, and adverse health outcomes including heightened physical pain, reduced pain tolerance, and an elevated risk of coronary heart disease. in another study, individuals who reported substantial financial debt experienced poorer self-reported general health and higher diastolic blood pressure (sweet et al., 2013). these associations persisted even after controlling for previous socioeconomic status, psychological and physical health, and various demographic factors (sweet et al., 2013). a meta-analysis found that being in debt was related to poor health behaviors including increases in smoking, problem drinking, and drug dependence (richardson et al., 2013). these findings underscore the interplay between financial wellbeing and physical health, highlighting the importance of considering the biological implications of economic hardships and financial stress. psychological ryu and fan (2023) find that financial stress greatly contributes to one’s psychological distress, and the relationship between financial stress and distress is moderated by socioeconomic factors such as gender, marital status, employment, income, and homeownership. the association between financial stress and psychological distress was significantly stronger among women, people who were separated, divorced, widowed, or never married, unemployed persons, those with a household income under $75,000, and those who rent versus own a home. a meta-analysis found statistically significant associations between debt and the presence of various mental health conditions, including but not limited to mental health disorders, suicide completion or attempt, as well as psychotic disorders. (richardson et al., 2013). depression and anxiety are widely studied psychological conditions associated with financial stress. a longitudinal study among cancer patients found that financial burden significantly predicted depressive symptoms and general anxiety. however, depressive symptoms and general anxiety during the initial survey did not predict subsequent financial burden, suggesting that financial difficulties are indicative of future distress (jones et al., 2020). at the extreme end of psychological issues, after financial services review, 33(1) 70 controlling for demographic and clinical covariates specific to the population of the study, results suggested that for each progressive increment in financial strain, the predicted probabilities of suicide attempts and suicidal ideation experienced significant escalation. respondents who acknowledged all four financial strain variables measured exhibited a predicted probability of future suicide attempts that were 20 times higher in comparison to respondents who did not endorse any of the financial strain variables (elbogen et al., 2020). in summary, there is a critical need to address financial well-being as a key factor in promoting and maintaining mental health. sociological the sociological implications of financial resilience and financial well-being have been relatively well examined from the lens of several disciplines. looking at many sociological elements, one study explored the relationship between financial wellness, personal well-being, and gender, finding that men scored higher in financial satisfaction and knowledge than women. however, women demonstrated higher levels of personal well-being, affirming the multi-dimensional aspect of financial wellness proposed by joo (2008). this also underscored the mediating role of financial satisfaction in the relationship between financial satisfaction and knowledge (gerrans et al., 2014). in another study, kim et al. (2003) found that after accounting for the initial financial stressor score, age, and household income, credit counseling had a positive impact on reducing financial stressors for clients who remained in the program for 18 months. although a person’s general degree of optimism can affect their resilience (muir & strnadova, 2014), the relationship between resilience and optimism is ambiguous, as the impact of one on the other remains unclear. while individuals' optimism levels may influence their ability to recover from adverse events, it is also plausible that their confidence in coping abilities influences their level of optimism (salignac et al., 2019). individuals classified as optimistic tend to have greater social capital (friends and family on whom they can rely for financial knowledge and assistance) and greater access to financial products and services (such as bank accounts, affordable credit and insurance products) than individuals classified as neutral or pessimistic (salignac et al., 2019). they note that since those experiencing sociological limitations such as mental illness may have higher barriers to amass formal and informal social supports and community resources, they are in turn also likely to express lower financial resilience. the literature reviewed here highlights the sociological factors that can influence financial well-being, such as gender, social support, and community resources. therefore, an approach based on the biopsychosocial model (engel, 1977) is useful for understanding financial stress and financial resilience’s role in promoting financial well-being. financial resilience framework the financial resilience framework (morrow, 2008; salignac et al., 2019) is intended to aid individuals and families in building financial resilience in the face of economic uncertainty and financial shocks, such as the period marked by covid-19 (norris, 2010). morrow (2008) conceptualized financial resilience in the context of measuring financial inclusion/exclusion as dependent on one’s knowledge of events, ability to predict risks, access to and knowledge of available alternatives, and resources to adapt. norris (2010) offered five essential financial elements of financial resilience: saving, budgeting, debt management, insurance, and investment. sherraden (1991) added that acquiring assets (savings, investments, and property) is a precursor to household and individual financial resilience. finally, salignac et al. (2019) expanded on the financial resilience framework by outlining four essential components: (a) economic resources, (b) financial resources and products, (c) financial knowledge and behaviors, and (d) social capital which serve as the operationalization of the financial resilience framework. the first component, economic resources, includes savings, income, and the ability to meet cost-of-living expenses. resilience is influenced by one’s ability to meet their cost of living (jacobs et al., 2014), and an inability to do so mccoy et al. 71 contributes to financial stress (orthner et al., 2004). the economic resources component captures an individual’s ability to cope with adversity and deal with unexpected expenses given their monetary inputs (demirgüç-kunt et al., 2015). the second component, financial resources and products, measures access to financial products and services. individuals may experience several different types of exclusion from financial products and services (cnaan et al., 2012; gomez-barroso & marban-flores, 2013; marron, 2013; salignac et al., 2016). financial exclusion arises from a confluence of factors. one key factor is condition mismatch, where the products and services offered simply don't align with the target population’s needs or interests. another factor is access limitations, where individuals do not meet minimum requirements to qualify for desired products or services (salignac et al., 2019; kempson & poppe, 2018). physical and geographic barriers also play a role, as the absence of local branches or service availability can significantly hinder access. price can be a significant hurdle as well, with costs exceeding the budgets of potential users. self-selected exclusion can occur when individuals voluntarily choose not to participate due to cultural reasons, psychological factors, or lack of financial literacy. finally, marketing gaps contribute to the problem when marketing strategies fail to effectively reach the target group, leading to a lack of awareness about available financial products and services. the third component, financial knowledge, builds on literature from financial literacy (lusardi & mitchell, 2014) and financial capability research (kempson & poppe, 2018; serido et al., 2013). given increasingly complex financial systems, an individual’s financial security is based on an adequate understanding of the system, along with positive financial skills and behaviors (lusardi & mitchell, 2014). this component of the financial resilience framework fills a necessary gap in the literature as neither financial literacy nor financial capability alone can well-explain one’s ability to cope with financial stressors or economic shocks (salignac et al., 2022). the fourth component is social capital. social capital is the connections among individuals— social networks and the norms of reciprocity and trustworthiness that arise from them (putnam, 2000). social capital depends on networks— specifically on the payoff from network membership—in terms of access to resources and opportunities that would be otherwise unavailable (scrivens & smith, 2013). individuals draw on family, friends, and community as sources of financial support and information in times of emergency (demirgüç kunt et al., 2015; seccombe, 2002). the component of social capital theory integrates this framework into the biopsychosocial model (engel, 1977) as it falls in the social domain of this model. resilient personality personality is made up of behavioral predispositions (i.e., temperament), cognitive attributes, and emotional qualities (skodol, 2010). research on the resilient personality type, a concept developed by werner & smith (1982), involves the traits and qualities that help people deal well with problems, stress, and other life difficulties. derived from the big five personality dimensions (agreeableness, conscientiousness, extraversion, neuroticism, and openness to experience), this simplified approach to personality typology characterizes individuals as resilient, undercontrolled, or overcontrolled (werner & smith, 1982). unlike under and overcontrolled personality types, resilient personality types can recover from difficult events or situations, adapt to changes, and maintain a positive attitude despite obstacles (morrow, 2008). resilient personality type ranges from having the skills to cope with adversity to being able to thrive in the face of adversity (bonanno, 2004). although one’s resilience is a dynamic internal process that changes and is affected by personal, familial, community, and cultural factors (masten, 2014), the resilient personality type has internal developmental assets that promote stronger resilience despite external factors (benson et al., 1999; masten, 2001). a relationship may exist between the financial resilience framework factors and resilient financial services review, 33(1) 72 personality; however, this examination is limited in the literature. zahedi et al. (2022) propose that future studies explore the ways that financial resilience might be moderated by personality traits. this study aims to expand on these findings by better understanding how having a resilient personality is associated with coping with financial adversity. methods data the data were collected by the qualtrics partner network of panel providers from november 17, 2021 to december 15, 2021 using several different avenues of recruitment (e.g., email, social media platforms). this data were part of a larger study that aimed to collect data on resilience in the aftermath of the covid-19 pandemic. the targeted population included adults living in the united states (n = 3,598) with a particular emphasis on low and moderate income households and people of color, as this population was most negatively impacted by the pandemic (kantamneni, 2020). the participants obtained through this data collection process were 51% white, 22% black/african american, 10% asian american, and 17% other. regarding ethnicity, 20% of the sample identifies as hispanic. also, among the respondents, approximately 30% have a high school diploma (or equivalent), 28% were some college, 31% were college degree holders (associate’s or bachelor's), 8% have earned graduate degrees, and 3% had less than high school educations. the majority (52%) were employed at least part-time, 24% were unemployed, and the remainder were self-employed, students, or other. the married or cohabiting respondents account for 58% of the sample. demographic data is summarized in table 1. table 1. client demographic data total sample analyzed cases n % n % age (n = 3,597) 18 – 24 years 654 18.18 529 17.12 25 – 34 years 1,107 30.78 994 32.17 35 – 44 years 869 24.16 761 24.63 45 – 54 years 441 12.26 381 12.33 55 – 64 years 256 7.12 209 6.76 65 years or older 270 7.51 216 6.99 gender female 1,767 49.11 1,498 48.48 male 1,753 48.72 1,527 49.42 other 78 2.17 65 2.10 race white 1,820 50.58 1,595 51.62 black 799 22.21 697 22.56 asian 362 10.06 290 9.39 other 617 17.15 508 16.44 hispanic no 2,879 80.02 2,464 79.74 yes 719 19.98 626 20.26 mccoy et al. 73 education (n = 3,596) less than high school 124 3.45 90 2.91 high school graduate or equivalent 1,095 30.45 934 30.23 some college, or degree or in progress 976 27.14 834 26.99 associate degree 442 12.29 385 12.46 bachelor's degree 682 28.97 603 19.51 graduate degree (master's, professional, doctorate) 277 7.70 244 7.90 marital status married or cohabiting 2,085 57.95 1,741 56.34 not married nor cohabiting 1,513 42.05 1,349 43.66 employment status (n = 3,597) employed full-time (40 hours per week) 1,457 40.51 1,349 43.66 employed part-time (less than 40 hours per week) 397 11.04 326 10.55 self-employed 301 8.37 250 8.09 full-time student 201 5.59 168 5.44 part-time student 67 1.86 56 1.81 unemployed 878 24.41 730 23.62 other 296 8.23 211 6.83 resilient personality no 2,037 56.61 1,748 56.57 yes 1,561 43.39 1,342 43.43 n = 3,598 unless otherwise noted for the total sample; n = 3,090 for the analyzed cases measures the two dependent variables were self-reported measures of physical and mental health. the physical health measure was a single question that asked respondents to rate their overall health. respondents were then asked in a separate single question to rate their overall mental health. responses for both questions range from 1 to 5, with 1 being “excellent” and 5 being “poor”. the independent variables of interest were the resilient personality typology and the financial resilience framework. to assess a resilient personality, respondents self-reported whether overcontrolled, undercontrolled, or resilient comes closest to their personality. respondents were allowed to select only one typology. the financial resilience framework variable was a scale composed of 10 questions representing the four components of the framework. seven of the ten resilience questions were given scores ranging from 0 to 1, with 0 representing the absence of a resilient feature and 1 indicating the highest level of resilience. financial knowledge was coded on a scale ranging from 0 to 1. the survey had three financial knowledge questions, and the total number of correct answers for each respondent was calculated. the score was then recoded as follows: 0 correct = 0.00; 1 correct = 0.33; 2 correct = 0.67; and 3 correct = 1.00. by recoding the objective financial knowledge variable [0, 1] interval, the construct of objective financial knowledge was weighted the same as the other constructs in the financial resilience index. social capital was measured by asking respondents to identify sources of financial financial services review, 33(1) 74 support, as was done in similar articles (e.g., scrivens & smith, 2013). respondents who identified a family member, friend, neighbor, faith-based community, service provider, institution, or organization as a source for urgently needed monetary support to face an emergency, were coded 1 indicating presence of social capital. if respondents said there was nobody they could ask if they urgently needed $1,000 for an emergency, then they were coded 0 to indicate no social capital. the total scores for the scale range from 0 to 8. responses to the questions for the financial resiliency framework are summarized in table 2. several covariates (i.e., age, gender, race, ethnicity, education, marital status, employment status, and employment change during covid19) were included in the analyses. table 2. financial resilience framework—survey results n % annual household income less than $15,000 643 18.37 $15,000 – $24,999 540 15.43 $25,000 – $34,999 539 15.40 $35,000 – $49,999 698 19.94 $50,000 – $74,999 360 10.29 $75,000 – $99,999 360 10.29 $100,000 – $149,999 360 10.29 greater than $150,000 0 0.00 savings before covid-19 pandemic i had no savings 984 28.20 i had very little savings (1 month of income or less) 886 25.39 i had limited savings (1 to 2 months of income) 763 21.87 i had moderate savings (3 or more months of income) 856 24.53 access to any form of credit before covid-19 pandemic no access to any form of credit 808 22.46 access to any form of credit 2,790 77.54 access to financial accounts no access to financial accounts 82 2.28 access to financial accounts 3,516 97.72 before making major financial decisions… almost no research 380 10.58 a little bit of research 996 27.72 moderate amount of research 1,168 32.51 a great deal of research 1,049 29.20 total financial literacy questions answered correctly 0 895 24.87 1 1,294 35.96 mccoy et al. 75 2 854 23.74 3 555 15.43 knowledge of financial products and services no knowledge of financial products and services 422 11.75 basic knowledge of financial products and services 1,749 48.68 good knowledge of financial products and services 1,069 29.75 very good knowledge of financial products and services 353 9.82 could you ask someone if you urgently needed $1,000 for an emergency? no 981 27.27 yes 2,617 72.73 missing data handled by listwise deletion for each question analysis glm anova was used to analyze the effects of the financial resilience framework and a resilient personality on physical and mental health. glm anova was chosen because it can account for continuous covariates and yet allow for greater interaction analysis than ols. this is important for substitute/complement analysis. for analysis, responses to the financial resilience variables were summed, and the scores were grouped into three categories with one category comprising the lowest quartile, one representing the middle two quartiles (interquartile range), and one representing the highest quartile. the quartile cutoff scores were less than 4.00 for the lower quartile (n = 810), 4.00 to 5.76 for the middle two quartiles (iqr) (n = 1,770), and greater than 5.76 for the upper quartile (n = 824). missing data on key variables reduced the number of complete cases to n = 3,090. results for mental health, glm anova analysis indicated significant results for the main effects of both the financial resilience framework score (χ2[2] = 34.99, p < .001) and the resilient personality indicator (χ2[1] = 51.60, p < .001). the glm anova results for mental health are presented in table 3. the interaction between resilient personality and the financial resilience framework score was also significant (χ2[2] = 11.24, p < .01). table 4 summarizes the main and interaction effects for the model. all interactions with a resilient personality and a high quartile indicator of the financial resilience framework were significantly stronger than any other combination at p < .001. a resilient personality combined with the financial resilience framework in the iqr was also significant at p < .001 compared to the same financial resilience framework level without a resilient personality. given a significant interaction, the analysis supports a resilient personality and the financial resilience framework as complements concerning their effect on mental health. both are associated with a significant difference in mental health, and the two factors interact to associate with even greater change. financial services review, 33(1) 2 table 3. mental health—glm anova results factors df χ2 significance indicator of resilience 2 34.99 *** resilient personality 1 51.60 *** interaction 2 11.24 ** n = 3,090 * p < .05; ** p < .01; *** p < .001 for physical health, glm anova analysis showed significant main effects for both the financial resilience framework score (χ2[2] = 77.50, p < .001) and a resilient personality (χ2[1] = 15.76, p < .01). the glm anova results for physical health are presented in table 5. however, the interaction term was not significant (χ2[2] = 4.21, p = 0.12). the main effects difference for a resilient personality was significant at p < .001, and all three possible main effects comparisons for the financial resilience framework score were also significant at p < .001. the main effects contrast is presented in table 6. glm anova supports the financial resilience framework and a resilient personality as substitutes concerning physical health. each factor is associated with a significant difference in physical health, with changes in the financial resilience framework associated with the greatest changes in physical health. the interaction between the two factors was not significant, indicating a substitution effect, where changing one factor alone would not be expected to affect the other factor’s impact on physical health. discussion this study examined whether the financial resilience framework and the resilient personality work as complements or substitutes to impact one’s mental and physical health in light of the health and financial crisis created by the covid19 pandemic. findings from this study indicate that both the financial resilience framework and resilient personality may contribute to one’s mental and physical health. however, the financial resilience framework is a stronger predictor of both mental and physical health outcomes than a resilient personality. findings from our study provide several essential contributions to the literature. first, the current study provides an important finding regarding mental health and resilience. the components that make up the financial resilience framework (salignac et al., 2019) have been linked to mental health outcomes in prior research. for example, financial knowledge and behavior have been associated with lower levels of psychological distress and depression (lim et al., 2019; seay et al., 2019). similarly, social capital has been linked to improved mental health outcomes, such as greater social support, reduced stress, and better overall well-being (kim & garman, 2019; moksnes et al., 2018). taken together, this study’s finding that the financial resilience framework is essential for mental health is important in designing personal finance interventions and policies that target financial health that can also aid a client’s mental health. mccoy et al. 1 table 4. mental health contrasts (main and interactions) (indicator group, personality group) contrast se z significance main: indicator 1 vs 0 0.1711 0.05 3.25 ** main: indicator 2 vs 0 0.3790 0.06 5.89 *** main: indicator 2 vs 1 0.2079 0.05 4.01 *** main: personality 1 vs 0 0.3174 0.04 7.18 *** (0, 1) vs (0, 0) 0.1341 0.09 1.56 (1, 0) vs (0, 0) 0.0917 0.06 1.43 (1, 1) vs (0, 0) 0.3845 0.07 5.56 *** (2, 0) vs (0, 0) 0.1834 0.08 2.17 * (2, 1) vs (0, 0) 0.7087 0.08 8.96 *** (1, 0) vs (0, 1) -0.0424 0.08 -0.54 (1, 1) vs (0, 1) 0.2541 0.08 3.05 ** (2, 0) vs (0, 1) 0.0493 0.10 0.52 (2, 1) vs (0, 1) 0.5746 0.09 6.35 *** (1, 1) vs (1, 0) 0.2929 0.05 5.13 *** (2, 0) vs (1, 0) 0.0917 0.07 0.04 (2, 1) vs (1, 0) 0.6170 0.07 9.25 *** (2, 0) vs (1, 1) -0.2012 0.08 -2.61 ** (2, 1) vs (1, 1) 0.3242 0.07 4.61 *** (2, 1) vs (2, 0) 0.5254 0.08 6.42 *** n = 3,090 * p < .05; ** p < .01; *** p < .001 indicator group: 0 = low quartile financial resilience, 1 = iqr financial resilience, 2 = high quartile financial resilience personality group: 0 = not resilient, 1 = resilient table 5. physical health—glm anova results factors df χ2 significance indicator of resilience 2 77.50 *** resilient personality 1 15.76 ** interaction 2 4.21 n = 3,090 * p < .05; ** p < .01; *** p < .001 financial services review, 33(1) 78 table 6. physical health—contrasts (main only) main group contrast se z significance indicator 1 vs 0 0.1621 0.05 3.56 *** indicator 2 vs 0 0.4735 0.06 8.51 *** indicator 2 vs 1 0.3114 0.04 6.96 *** personality 1 vs 0 0.1517 0.04 3.97 *** n = 3,090 * p < .05; ** p < .01; *** p < .001 indicator group: 0 = low quartile financial resilience, 1 = iqr financial resilience, 2 = high quartile financial resilience personality group: 0 = not resilient, 1 = resilient secondly, our findings support the link between factors related to the financial resilience framework (salignac et al., 2019) and physical health. for example, economic resources have been associated with improved access to healthcare, better nutrition, and better overall physical health (morrow, 2011; zhu et al., 2019). access to financial resources has also been linked to improved physical health outcomes, including lower levels of chronic disease and better overall health status (kobayashi et al., 2015; zhu et al., 2019). social capital through informal networks can provide individuals with information, support, and motivation to engage in healthpromoting behaviors such as exercise, healthy eating, and smoking cessation (berkman et al., 2000). similarly, our findings show the added benefit of increased physical health and higher levels of financial resilience. furthermore, social networks can facilitate access to healthcare services and encourage compliance with medical treatments (kim et al., 2017). yet, to the authors’ knowledge, no study has directly examined how the entirety of the financial resilience framework impacts physical health domains, and the link is important to explore further in future studies. finally, our study found that the financial resilience framework and resilient personality were complementary in explaining mental health but not the physical health results. the complementary nature of the resiliency factors and mental health is consistent with previous research that suggests that personality traits can moderate the relationship between financial stressors and mental health outcomes (rothmann & coetzer, 2003). for instance, individuals with a resilient personality may be better able to cope with financial stressors and maintain positive mental health outcomes (windle, 2011). interestingly, this study did not find support for a complementary relationship between the financial resilience framework and resilient personality traits when it came to physical health. this may be because financial factors are more directly related to physical health outcomes, particularly in the united states where medical care costs are high, financial resources become even more essential to access healthcare, purchase prescriptions, and buy healthy food options (todorova et al., 2016). the impact of inner vision, calmness, intelligence, maturity, and self-esteem on financial setbacks will not overcome the need for financial resources to be physically healthy. however, more research is needed to gather deeper insights into why resilient personality did not have a complementary role when it came to physical health. these findings suggest that financial professionals should continue to highlight the importance of financial security and capability and the role of social support networks in promoting both mental and physical health to their clients. moreover, the findings underscore the importance for financial professionals to advocate for increased access to mental and physical health resources so individuals can maintain a healthy lifestyle. financial services mccoy et al. 79 providers should be trained on how to identify financial difficulties that may be linked to psychological and physical stressors to help these clients receive the proper care, thus increasing their ability to cope with financial shocks. limitations there are several limitations to note. this study did not have a random selection of respondents, which can impact generalizability to the larger population. next, none of the households sampled had an income greater than $150,000. this was an intentional decision as the study was designed to study at-risk households and oversample households based on race/ethnicity and income. results from our study cannot be applied directly to households that were excluded based on the research design. future research should focus on high-income groups to see if these findings will hold. our study used a singleitem question to measure a resilient personality. future studies should consider using multi-item measure to improve robustness and possibly provide results for different levels of the resilient personality trait. similarly, where data allowed, we included the individual’s position before the stressor event (savings before covid and access to credit before covid), but longitudinal data would improve the reliability and validity of this study. lastly, nearly one-fourth of respondents were unemployed. although this employment status is overrepresented, it aligns with the situation at hand during the height of the pandemic. the unemployment rate for black and hispanic/latino communities was significantly higher than the rate for white communities during the covid-19 pandemic. according to data from the bureau of labor statistics (bls), the unemployment rate for black americans peaked at 16.8% in april 2020. by march 2021, this rate gradually declined to 7.8%. similarly, the unemployment rate for hispanic/latino americans peaked at 18.9% in april 2020 and declined to 7.9% by march 2021. in comparison, the unemployment rate for white americans peaked at 14.2% in april 2020 and declined to 5.4% by march 2021. it is worth noting that these unemployment rates do not account for individuals who dropped out of the labor force due to pandemic-related factors, such as caregiving responsibilities or health concerns. thus, the true impact of the pandemic on employment may be even greater than these statistics suggest. implications results from this study provide an opportunity to re-examine the biopsychosocial model (engel, 1977) as the model originally incorporated basic financial aspects (e.g., socioeconomic status and household income) within the social component construct. however, the intersection between more nuanced financial health factors, as described in the financial resilience framework (i.e., economic resources, access to financial resources, financial knowledge and behavior, and social capital; salignac et al., 2019) suggests there is potential for the biopsychosocial model to include a separate financial component. adding a separate financial component will allow research to examine the unique contributions of social health (e.g., friends and community) and one’s overall well-being, including financial health. this is consistent with previous research that has emphasized the need for a more comprehensive understanding of the factors that contribute to overall well-being, including financial factors (moffitt et al., 2018). as stated by kannadhasan et al. (2016), “there is no specific theory on the role of biopsychosocial factors in the financial services domain” (p. 118). our findings suggest that this is an oversight in the theorizing of the connection between pillars of well-being. this may be especially true in the united states, where the costs of medical care are staggering relative to other countries around the world (papanicolas et al., 2018), and one’s physical and mental health may be even more directly dictated by one’s financial health (todorova et al., 2016). this study has several practical implications for various personal finance and mental health stakeholders across education, advising, coaching, planning, and counseling domains. for mental health professionals, interventions that increase resilience could help clients better cope with mental health issues such as trauma, depression, and anxiety. for financial professionals, understanding the traits that make financial services review, 33(1) 80 people resilient can help identify individuals at risk for mental health problems. the findings of this study suggest that financial professionals should persist in highlighting to their clients the critical role of financial stability, financial literacy, and social support networks in promoting both mental and physical health. moreover, the findings underscore the importance for financial professionals to advocate for increased access to services and resources that influence mental and physical well-being, thereby supporting individuals in maintaining a healthy lifestyle. it is crucial for financial professionals to acknowledge and address financial difficulties that may be linked to psychological and physical stressors among their clients. finally, financial professionals can utilize the insights from this study to enrich discussions with clients about how personality factors contribute to their overall well-being. our study provides several reasons for the inclusion of resilience methods in financial and mental health interventions. first, findings from this study support the idea that interventions should include complementary mechanisms that focus on helping clients enhance their financial resiliency. the need for holistic and multifaceted interventions is evident given that those in this study with a resilient personality and the multiple components of the financial resilience framework may have better mental and physical health. by creating interventions that address resilience across the five elements (saving, budgeting, debt management, insurance, and investment; norris, 2010) and incorporating strategies that support the four financial resilience framework components (economic resources, financial resources, and products, financial knowledge and behaviors, and social capital; salignac et al., 2019), practitioners and researchers will strengthen their effectiveness in helping individuals cope and have better overall wellbeing. second, lingering chronic stress, secondary to adverse covid-19 pandemic related outcomes, is an issue that many mental health and personal finance practitioners are still helping their clients manage. resources have been made available from federal and state agencies, professional organizations (e.g., afcpe, fpa), financial institutions, and nonprofits. for example, the consumer financial protection bureau (cfpb, n.d.) developed web-based and printed materials to inform consumers on how they can protect and manage their finances during covid-19. these resources focus on financial management, mortgage and housing assistance, avoiding fraud, and student loan relief. practitioners can familiarize themselves with such resources and incorporate them as tools to help their clients build a mindset to better endure future economic downturns and protect their mental and financial health. for personal finance practitioners, these results also hint at the importance of financial planners incorporating the biopsychosocial model when working with clients. this inclusion may provide a more thorough assessment of a client’s physical and mental health, particularly for financial planning components such as cash flow, estate planning, and insurance planning. given the difficulty many practitioners experience getting their clients to implement recommendations, the biopsychosocial model coupled with assessing financial resilience could help planners and counselors better assist their clients with meeting their goals. more research should be conducted to test these interventions in hopes of informing professional practice. finally, the findings support an intersectional approach to research that could potentially expand our understanding of how people cope during economic uncertainty in times of crisis and the tools they need to cope and recover. future studies can focus on examining how having financial resilience is associated with other outcomes of well-being, such as parenting, environmental, or relationship health. to do this, measures that capture resilient personalities and the financial resilience framework could be included in data collection. additionally, the findings suggest a need for empirically based interventions that are inclusive of targeted audiences (e.g., gender, race, culture) and lead to valid and reliable outcomes. resilience may also help expand insights regarding factors associated with other research areas such as consumer decision-making and behavioral economics. mccoy et al. 81 references ahmed, o., hossain, k. n., siddique, r. f., & jobe, m. c. (2021). covid-19 fear, stress, sleep quality and coping activities during the lockdown, and personality traits: a person-centered approach analysis. personality and individual differences, 178, 110873. asendorpf, j. b., borkenau, p., ostendorf, f., & van aken, m. a. g. (2001). carving personality description at its joints: confirmation of three replicable personality prototypes for both children and adults. european journal of personality, 15, 169–198. https://doi.org/10.1002/per.408. bambra, c., riordan, r., ford, j., & matthews, f. (2021). the covid-19 pandemic and health inequalities. journal of epidemiology and community health, 75(6), 461-463. benson, p. l., scales, p. c., leffert, n., & roehlkepartain, e. c. (1999). a fragile foundation: the state of developmental assets among american youth. minneapolis, mn: search institute. berkman, l. f., glass, t., brissette, i., & seeman, t. e. (2000). from social integration to health: durkheim in the new millennium. social science & medicine, 51(6), 843-857. bonanno, g. a. (2004). loss, trauma, and human resilience: have we underestimated the human capacity to thrive after extremely aversive events? american psychologist, 59(1), 20–28. https://doi.org/10.1037/0003066x.59.1.20 burgard, s. a., & kalousova, l. (2015). effects of the great recession: health and wellbeing. annual review of sociology, 41, 181-201. chou, e. y., parmar, b. l., & galinsky, a. d. (2016). economic insecurity increases physical pain. psychological science, 27(4), 443–454. https://doi.org/10.1177/0956797615625 640 cnaan, r. a., moodithaya, m. s., & handy, f. (2012). financial inclusion: lessons from rural south india. journal of social policy, 41, 183–205. coibion, o., gorodnichenko, y., & weber, m. (2020). labor markets during the covid-19 crisis: a preliminary view (nber working paper 27017). national bureau of economic research. https://10.3386/w27017 consumer financial protection bureau. (n.d.). protecting your finances during the coronavirus pandemic. https://www.consumerfinance.gov/coron avirus/ crayne, m. p. (2020). the traumatic impact of job loss and job search in the aftermath of covid-19. psychological trauma: theory, research, practice, and policy, 12(s1), s180–s182. https://doi.org/10.1037/tra0000852 elbogen, e. b., lanier, m., montgomery, a. e., strickland, s., wagner, h. r., & tsai, j. (2020). financial strain and suicide attempts in a nationally representative sample of us adults. american journal of epidemiology, 189(11), 1266–1274. https://doi.org/10.1093/aje/kwaa146 engel, g. l. (1977). the need for a new medical model: a challenge for biomedicine. science, 196(4286), 129-136. faber, r. j. (1992). money changes everything: compulsive buying from a biopsychosocial perspective. american behavioral scientist, 35(6), 809-819. faria-e-castro, m. (2021). fiscal policy during a pandemic. journal of economic dynamics and control, 125, article 104088. https://doi.org/10.1016/j.jedc.2021.1040 88 fong, t. w. (2005). the biopsychosocial consequences of pathological gambling. psychiatry, 2(3), 22–30. https://doi.org/10.1037/0003-066x.59.1.20 https://doi.org/10.1037/0003-066x.59.1.20 https://doi.org/10.1037/tra0000852 https://doi.org/10.1037/tra0000852 https://doi.org/10.1037/tra0000852 https://doi.org/10.1016/j.jedc.2021.104088 https://doi.org/10.1016/j.jedc.2021.104088 https://doi.org/10.1016/j.jedc.2021.104088 https://doi.org/10.1016/j.jedc.2021.104088 financial services review, 33(1) 82 gao, j., zheng, p., jia, y., chen, h., mao, y., chen, s., wang, w., fu., h. & dai, j. (2020). mental health problems and social media exposure during the covid-19 outbreak. plos one, 15(4), e0231924. gerrans, p., speelman, c., & campitelli, g. (2014). the relationship between personal financial wellness and financial well-being: a structural equation modeling approach. journal of family and economic issues, 35(2), 145–160. https://doi.org/10.1007/s10834-0139358-z gomez-barroso, j. l., & marban-flores, r. (2013). basic financial services: a new service of general economic interest? journal of european social policy, 23, 332–339. grable, j. e., & joo, s.h. (2004). environmental and biopsychosocial factors associated with financial risk tolerance. journal of financial counseling and planning, 15(1), 73–82. grable, j., britt, s., & webb, f. (2008). environmental and biopsychosocial profiling as a means for describing financial risk-taking behavior. journal of financial counseling and planning, 19(2). gruber, j., hinshaw, s. p., clark, l. a., rottenberg, j., & prinstein, m. j. (2023). young adult mental health beyond the covid-19 era: can enlightened policy promote long-term change? policy insights from the behavioral and brain sciences, 10(1), 75-82. hughes, c. (2021). the impact of creditworthiness on financial well-being, anxiety, depression, hopelessness, and suicide. journal of accounting & finance (2158-3625), 21(3), 143–160. https://doi.org/10.33423/jaf.v21i3.4400 jacobs, d., perera, d., & williams, t. (2014). inflation and the cost of living. bulletin. reserve bank of australia. joo, s. (2008). personal financial wellness. in j. j. xiao (ed.), handbook of consumer finance research (pp. 21–33). springer. kannadhasan, m., aramvalarthan, s., mitra, s. k., & goyal, v. (2016). relationship between biopsychosocial factors and financial risk tolerance: an empirical study. vikalpa, 41(2), 117-131. kantamneni, n. (2020). the impact of the covid-19 pandemic on marginalized populations in the united states: a research agenda. journal of vocational behavior, 119, 103439. kempson, e., & poppe, c. (2018). understanding financial well-being and capability a revised model and comprehensive analysis. oslo metropolitan university. kim, h., & garman, e. t. (2019). financial socialization and the role of social capital in reducing financial stress. journal of financial counseling and planning, 30(2), 239-253. kim, j., han, y., & lee, h. (2017). social network types and health status among elderly koreans. journal of community health nursing, 34(4), 173-182. kobayashi, l. c., wardle, j., wolf, m. s., & von wagner, c. (2015). aging and functional health literacy: a systematic review and meta-analysis. journals of gerontology series b: psychological sciences and social sciences, 71(3), 445-457. kochhar, r. (2020, june 11). unemployment rose higher in three months of covid-19 than it did in two years of the great recession. pew research center. https://www.pewresearch.org/facttank/2020/06/11/unemploymentrosehigher-in-three-months-of-covid-19than-it-did-intwo-years-of-the-greatrecession/ kop, w. j. (2021). biopsychosocial processes of health and disease during the covid-19 pandemic. psychosomatic medicine, 83(4), 304-308. mccoy et al. 83 lee, s., wickrama, k. k. a. s., lee, t. k., & o’neal, c. w. (2021). long-term physical health consequences of financial and marital stress in middle-aged couples. journal of marriage and family, 83(4), 1212–1226. https://doi.org/10.1111/jomf.12736 lim, v. k., teo, t. s., & loo, g. l. (2019). effects of financial knowledge, behavior, and well-being: an exploratory study of individuals in singapore. journal of financial counseling and planning, 30(1), 110-124. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. marron, d. (2013). governing poverty in a neoliberal age: new labour and the case of financial exclusion. new political economy, 18, 785. masten, a. s. (2001). ordinary magic: resilience in development. american psychologist, 56(3), 227-238. doi: 10.1037/0003066x.56.3.227 masten, a. s. (2014). global perspectives on resilience in children and youth. child development, 85(1), 6-20. doi: 10.1111/cdev.12205 meyer, j., mcdowell, c., lansing, j., brower, c., smith, l., tully, m., & herring, m. (2020). changes in physical activity and sedentary behavior in response to covid-19 and their associations with mental health in 3,052 us adults. international journal of environmental research and public health, 17(18), 6469. moksnes, u. k., lazarewicz, m., & løvgren, m. (2018). social support and mental health among young norwegian adults. scandinavian journal of psychology, 59(6), 623-629. morrow, s. l. (2008). conceptualizing financial resilience. journal of family and economic issues, 29(2), 612-623. morrow, s. l. (2011). financial capability and asset holding in later life: a life course perspective. journal of gerontological social work, 54(8), 815-829. muir, k., & strnadova, i. (2014). whose responsibility? resilience in families of children with developmental disabilities. disability and society, 29(6), 922–937. norris, f., (2010). behavioural science perspectives on resilience. carri research paper, 11, community and regional resilience institute oak ridge: tennessee, usa, 50 pp. orthner, d. k., jones-sanpei, h., & williamson, s. (2004). the resilience and strengths of low-income families. family relations, 53, 129–167. papanicolas, i., woskie, l. r., & jha, a. k. (2018). health care spending in the united states and other high-income countries. journal of the american medical association, 319(10), 10241039. pappas, s. (2020). how will people react to the new financial crisis? american psychological association. patel, j. a., nielsen, f. b. h., badiani, a. a., assi, s., unadkat, v. a., patel, b., ravindrane, r. & wardle, h. (2021). poverty, inequality and covid-19: the forgotten vulnerable. public health, 183, 110-111. pfefferbaum, b., & north, c. s. (2020). mental health and the covid-19 pandemic. new england journal of medicine, 383(6), 510-512. prati, g., & mancini, a. (2021). the psychological impact of covid-19 pandemic lockdowns: a review and meta-analysis of longitudinal studies and natural experiments. psychological medicine, 51(2), 201–211. https://doi .org/10.1017/s0033291721000015 putnam, r. (2000). bowling alone: the collapse and revival of american community. simon and schuster. financial services review, 33(1) 84 richardson, t., elliott, p., & roberts, r. (2013). the relationship between personal unsecured debt and mental and physical health: a systematic review and metaanalysis. clinical psychology review, 33(8), 1148–1162. https://doi.org/10.1016/j.cpr.2013.08.00 9 rosenström, t., & jokela, m. (2017). a parsimonious explanation of the resilient, undercontrolled, and overcontrolled personality types. european journal of personality, 31(6), 658-668. rothmann, s., & coetzer, e. p. (2003). the big five personality dimensions and job performance. sa journal of industrial psychology, 29(1), 68-74. ryu, s., & fan, l. (2023). the relationship between financial worries and psychological distress among us adults. journal of family and economic issues, 44(1), 16-33. salari, n., hosseinian-far, a., jalali, r., vaisiraygani, a., rasoulpoor, s., mohammadi, m., rasoulpoor, s., & khaledi paveh, b. (2020). prevalence of stress, anxiety, depression among the general population during the covid-19 pandemic: a systematic review and meta-analysis. globalization and health, 16(1), article 57. https://doi.org/10.1186/ s12992-02000589-w salignac, f., hanoteau, j., & ramia, i. (2022). financial resilience: a way forward towards economic development in developing countries. social indicators research, 160(1), 1-33. salignac, f., marjolin, a., reeve, r., & muir, k. (2019). conceptualizing and measuring financial resilience: a multidimensional framework. social indicators research, 145(1), 17–38. https://doi.org/10.1007/s11205-01902100-4 salignac, f., muir, k., & wong, j. (2016). are you really financially excluded if you choose not to be included? insights from social exclusion resilience and ecological systems. journal of social policy, 45(2), 269–286. https://doi.org/10.1017/s004727941500 0677 scrivens, & smith, c. (2013). four interpretations of social capital. (oecd statistics working papers, no. 2013/6). seay, m. c., kubik, m. y., & grymes, m. r. (2019). financial education, financial behavior, and psychological distress: an examination of the mediating influence of financial behaviors. journal of financial counseling and planning, 30(1), 86-99. serido, j., shim, s., & tang, c. (2013). a developmental model of financial capability: a framework for promoting a successful transition to adulthood. international journal of behavioural development, 37, 287–297. shadmi, e., chen, y., dourado, i., faran-perach, i., furler, j., hangoma, p., piya hanvoravongchai, obando, c., petrosyan, v., rao, k.d., ruano, a.l., shi, l. de souza, l.e., spitzer-shohat, s., sturgiss, e., suphanchaimat, r., uribe, m.v., willems, s. & shi, l. (2020). health equity and covid-19: global perspectives. international journal for equity in health, 19(1), 1-16. shandler, r., gross, m. l., & canetti, d. (2023). cyberattacks, psychological distress, and military escalation: an internal metaanalysis. journal of global security studies, 8(1), ogac042. sherraden, m. (1991). assets and the poor: a new american welfare policy. armonk, ny: m.e. sharpe. skodol, a.e. the resilient personality. handbook of adult resilience; guilford press: new york, ny, usa, 2010; volume 112 sweet, e., nandi, a., adam, e. k., & mcdade, t. w. (2013). the high price of debt: household financial debt and its impact on mental and physical health. social mccoy et al. 85 science & medicine, 91, 94–100. https://doi.org/10.1016/j.socscimed.201 3.05.009 todorova, i. l., tucker, k. l., jimenez, m. p., lincoln, a. k., arevalo, s., & falcón, l. m. (2013). determinants of self-rated health and the role of acculturation: implications for health inequalities. ethnicity & health, 18(6), 563-585. tran, b. x., nguyen, h. t., le, h. t., latkin, c. a., pham, h. q., vu, l. g., le, x., nguyen, t. t., pham, q. t., ta, n., nguyen, q. t., ho, c., & ho, r. (2020). impact of covid-19 on economic wellbeing and quality of life of the vietnamese during the national social distancing. frontiers in psychology, 11, article 565153. https://doi .org/10.3389/fpsyg.2020.565153 tsamakis, k., triantafyllis, a. s., tsiptsios, d., spartalis, e., mueller, c., tsamakis, c., chaidou, s., spandidos, d.a., fotis, l., economou, m., & rizos, e. (2020). covid‑19 related stress exacerbates common physical and mental pathologies and affects treatment. experimental and therapeutic medicine, 20(1), 159-162. werner, e. e., & smith, r. s. (1982). vulnerability and resiliency in children: a longitudinal study. american journal of orthopsychiatry, 52(4), 559-567. doi: 10.1111/j.1939-0025.1982.tb01456.x windle, g. (2011). what is resilience? a review and concept analysis. reviews in clinical gerontology, 21(2), 152-169. https://doi.org/10.1017/s095925981000 0420. zahedi, j., salehi, m., & moradi, m. (2022). identifying and classifying the contributing factors to financial resilience. foresight, 24(2), 177-194. zhu, h., luo, x., & wang, d. (2019). how does financial access affect health outcomes? evidence from china. applied economics letters, 26(10), 825-831. https://doi.org/10.1017/s0959259810000420 https://doi.org/10.1017/s0959259810000420 individual estimates of life expectancy and consumption patterns derek r. lawson, cfpa*, stuart j. heckman, ph.d., cfpa adepartment of personal financial planning, kansas state university, 1324 lovers lane, 303 justin hall, manhattan, ks 66506, usa abstract previous research reveals a paucity of retirement preparedness in the united states. with the shift to defined contribution plans, retirement planning now falls on the shoulders of consumers. using the 2013 survey of consumer finances (scf), the relationship between subjective life expectancy and consumption is investigated. specifically, regression analyses examine the relationship between subjective life expectancy and three indicators of consumption: financial assets, credit card debt, and saving behavior. a secondary analysis separated respondents near retirement (i.e., within 10 years) and far from retirement (i.e., more than 10 years) to determine if retirement saliency affects the relationship between life expectancy and consumption. the influence of life expectancy on consumption is analyzed by separating life expectancy into two periods: remaining work life and retirement life. results indicate that remaining work life expectancy and retirement life expectancy are associated with financial assets and that retirement life expectancy is associated with savings behavior. © 2017 academy of financial services. all rights reserved. jel classification: d14; e21 keywords: consumption; life expectancy; retirement; saving 1. introduction one result of the shift away from defined benefit plans to defined contribution plans is that household consumption decisions have a substantial and lasting impact on a household’s ability to prepare for retirement. retirement planning is a complex task for the average * corresponding author. tel.: �1-515-707-3729; fax: �1-785-532-5505. e-mail address: drlawson@ksu.edu (d.r. lawson) financial services review 26 (2017) 1–18 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. consumer given all the uncertainty involved, for example, length of time in retirement, inflation, rates of return, and so forth. additionally, consumers have difficulty in making decisions under uncertainty and are prone to many biases and mistakes (tversky & kahneman, 1974). now that most consumers bear the investment and longevity risk associated with retirement funding, consumption decisions throughout the life cycle have an important role to play in shaping retirement adequacy. the current study focuses on one particular source of uncertainty, an individual’s life expectancy, and examines how subjective estimates of life expectancy influence household consumption. because of the dynamic relationship between current age, retirement age, and life expectancy and consumption choice, the current study analyzes the influence of life expectancy on consumption by dividing the respondents’ remaining life expectancy into two periods—remaining work life expectancy and retirement life expectancy. from a life cycle perspective, examining these two periods of time (i.e., remaining working period and retirement period) highlights the tradeoff between saving and dissaving. 1.1. literature review some studies find subjective life expectancy to be a valuable predictor of actual mortality (hamermesh, 1985; hurd & mcgarry, 2002). elder’s (2007) findings confirm individual estimates of life expectancy to be accurate but noted an interesting pattern in that individuals nearing retirement had higher discount rates for their future. his results show that individuals nearing or just starting retirement were pessimistic about living to age 75, but became optimistic about living to age 85 once they survived to age 75. this creates a misalignment in the savings patterns needed to smooth consumption because the initial retirement period planned for is too short and needs to be corrected for a longer horizon. there is some evidence to support the notion that subjective life expectancy influences saving and consumption patterns. for example, bloom, canning, moore, and song’s (2006) findings show that increased subjective life expectancy yields increased savings at any age. they also look at differences between a husband’s and wife’s subjective life expectancy estimates on household wealth and find that a 10% increase in a husband’s life expectancy results in an increase of household wealth by $27,600, and a 10% increase in the wife’s life expectancy results in an increase in household wealth by $32,700. other findings show that an increase in life expectancy is associated with an increase in savings, but most of these findings are based on data from east asian countries because of their increases in life expectancies (bloom, canning, & graham, 2003; tsai, chu, & chung, 2000). yet, other research finds contradictory results indicating that rises in life expectancy mostly increase expected retirement age when individuals are not subject to mandatory retirement ages (deaton & paxson, 2000). bloom et al. (2003) find improved health has an ambiguous effect on savings rates because it may increase remaining work life expectancy. because of this, savings rates do not necessarily need to increase because compounded interest on current savings can help cover expenditures later on (bloom, canning, & moore, 2004). spaenjers and spira (2015) analyze the empirical relationship between subjective life expectancy and portfolio choice. their findings show that life expectancy positively influenced the amount allocated towards equities within portfolios. that is, after controlling for 2 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 effects related to gender-specific age, health, and other socioeconomic characteristics, people with longer life expectancies have greater amounts of equities than their counterparts. they also compare people with the same subjective life expectancies to see if bequest motives make a difference on portfolio allocations. they find that those without bequest motives had a decreased equity allocation as compared to those with bequest motives, indicating that people with bequest motives may view their life horizons as “indefinite” (spaenjers & spira, 2015, p. 104). an important consideration for retirement planning and savings decisions is an individual’s planned retirement age. montalto, yuh, and hanna (2000) examine the determinants of planned retirement age before retirement. they find that as people get older their planned retirement age also increases. specifically, they find that as full-time workers age from 35 to 45, their expected retirement age increases by 1.08 years, while workers aging from 45 to 55, increase their expected retirement age by 3.41 years. finally, those workers aging from 55 to 65 increase their expected retirement date by 5.72 years. they also find that life expectancy, non-investment income, and educational attainment are all positively associated with planned retirement age. financial assets (excluding ira/keogh and defined contribution plans), having a defined benefit plan, and low-skilled occupations are negatively associated with expected retirement age (montalto et al., 2000). 1.2. summary and gap while subjective life expectancy has been shown to influence consumer savings (bloom et al., 2003), there is little empirical research examining the influence on household consumption behavior, particularly when using the survey of consumer finances (scf). other research utilizing the scf finds subjective life expectancy influences annuity purchase decisions (gupta & li, 2012) and construction of investment portfolios (spaenjers & spira, 2015). therefore, the current study explores the following research question: rq: what effect do subjective estimates of life expectancy have on consumption decisions? 1.3. theoretical framework 1.3.1. life cycle hypothesis first posited by modigliani and brumberg (1954), the life cycle hypothesis is driven by individual utility—a function of an individual’s aggregate consumption. in this model, it is assumed that individuals aim to maximize their lifetime utility given their total lifetime wealth, w, which is the summation of one’s current level of resources (a0) and their future earnings (h) (yuh & hanna, 2010). w � a0 � h if there is no uncertainty about remaining work life expectancy (rwle) and retirement life expectancy (rle), and a zero real rate of interest, as initially assumed in modigliani and brumberg’s model, then optimal consumption, c, is equal to total wealth divided by total life expectancy (where le � rwle � rle): 3d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 c � w le to smooth consumption, individuals may borrow initially until income is high enough to support expenses. they then save when income is above expenses and dissave during retirement. ando and modigliani (1963) present the following equation for total consumption, c, in year t, for a person that is age t: ct t � �t t�t t where �t t represents what ando and modigliani (1963) call a proportionality factor, which depends on the rate of return an individual can earn on assets, the individual’s current age, and the form of the utility function to be used. the second part, �t t, represents total resources available to that individual, which can be expressed as net worth. based on this framework, holding all else (especially retirement age) constant, it is expected that an increase in life expectancy would decrease the amount of consumption in each year, t. in other words, if total resources remain constant and an individual lives longer, consumption would need to decrease, assuming that the compounded interested earned on assets from the previous periods are insufficient to fund extra years of income needs.1 1.3.2. behavioral life cycle hypothesis to consider some of the behavioral tendencies inherent in humans, shefrin and thaler (1988) developed a derivative of the life cycle model known as the behavioral life cycle hypothesis. while they state that the life cycle hypothesis was useful in its own right, particularly as a prescriptive model, they acknowledge that individuals do not always act in an optimal manner. therefore, modifications to the model were made to make the “theory more behaviorally realistic” (shefrin & thaler, 1988, p. 1). they modify the life cycle model to account for three human behavior factors often missing from economic theory: (a) self-control, (b) mental accounting, and (c) framing. according to shefrin and thaler (1988), self-control is made up of three components: (a) internal conflict, (b) temptation, and (c) willpower. internal conflict is described as an inherent battle between the planner and the doer within each individual. the planner is concerned with the long term while the doer aims to maximize utility in the present. temptation deals with the choices that are presented to individuals, recognizing that for each choice there are differing degrees of enticement. willpower is the mental cost associated with resisting temptation and the degree of temptation influences how much willpower the individual must expend to make a choice that aligns with their utility-maximizing preference for both the longand short-term. the more an individual is tempted to indulge in something, the more willpower they must expend to resist the temptation. although willpower can be taxing, it is necessary to reduce consumption (shefrin & thaler, 1988). individuals will engage in willpower if the utility lost from less consumption (i.e., not succumbing to temptation) is less than the utility cost of engaging in willpower. hence, willpower is exerted when the net marginal cost of using it is greater than zero (shefrin & thaler, 1988). self-control, then, is the relationship among internal conflict, temptation, and willpower. 4 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 consumption in today’s economy is heavily influenced by advertising, which is designed to convince consumers that they need products they previously had not considered. approximately $180 billion is spent annually on advertising in the united states (and projected to grow to nearly $220 billion by 2018) (statista, 2016), which results in nearly constant exposure to advertisements. consumers are constantly at conflict with themselves because they are tempted to buy things that they may not actually need or want. shefrin and thaler (1988) argue that an economic savings model that does not factor in self-control is a mis-specified model. thus, willpower—the mental costs of resisting temptation—must be considered. a lack of self-control can lead to financial problems down the road (e.g., inadequate retirement preparation), whereas too much self-control can deprive someone of their wants and needs. although individuals may expect to live a long time in retirement, they may lack the ability to exert self-control and defer consumption from today to future periods by saving. from this perspective, it is predicted that an increase in life expectancy would not result in a change in consumption because of willpower and self-control issues among working households for which retirement is not salient. that is to say, retirement may be too far in the future to exert self-control and reduce consumption for young households. however, as retirement becomes more salient (i.e., as households near retirement) it is expected that longer life expectancies would have a negative effect on consumption. 1.3.3. summary subjective life expectancy should have a negative effect on consumption according to the life cycle hypothesis. in general, those who expect to live longer should save more and spend less. however, consumers may struggle with self-control according to the behavioral life cycle hypothesis. in the case of self-control problems, life expectancy is expected to have a negligible or inconsistent effect on consumption for individuals far from retirement but may have a negative effect on consumption for individuals near retirement. that is to say, for individuals near retirement, those who expect to live longer should save more and spend less than individuals far from retirement. 2. method 2.1. data the 2013 scf, a triennial survey sponsored by the federal reserve board (fed), is used and our sample is limited to respondents who work full-time, expect to retire, and have a retirement life expectancy period greater than zero (n � 2,632). the reason for this restriction is that full-time workers have the greatest flexibility in adjusting saving and consumption decisions based on their life expectancy. retired households or those who work less than full-time may systematically differ from full-time workers. additionally, the sample is limited to those with a rle period greater than zero because a few respondents had a negative rle (i.e., the respondent expected to retire at a date later than his or her life expectancy), possibly indicating that they did not understand the questions. the switch 5d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 variable in the scf is used to obtain respondent characteristics (lindamood, hanna, & bi, 2007). for consistency, weighted statistics and regression models are presented, although multivariate results are robust to weighting. the scf relies on a complex sampling design (nielsen & seay, 2014; shin & hanna, 2016) and uses multiple imputation to handle missing data and to help protect participants’ privacy (montalto & sung, 1996). therefore, an available stata program (see nielsen, 2015) is used to implement the repeated imputation inference (rii) method (montalto & sung, 1996) while also using the replicate weight file provided by the fed to bootstrap the standard errors (nielsen & seay, 2014). this procedure provides estimates of variance that are closer to the true variance. 2.2. variables three dependent variables were chosen to represent consumption decisions in the scf: (a) financial assets, (b) credit card balance, and (c) saving behavior. both financial assets and credit card balances are log-transformed to correct for the highly skewed distribution. our third dependent variable is saving behavior (whether or not the respondent saved or not).2 this is measured by whether or not the household reported spending less than income, after considering durable goods purchases (in particular, costs associated with purchasing a car, home, or investments are considered savings in the scf). households who spent all or more of their income are described as non-savers. the key independent variables are remaining work life expectancy (rwle) and retirement life expectancy (rle). rwle is computed by subtracting respondent age from respondent’s estimated retirement age. rle is computed by subtracting respondent’s retirement age from respondent’s estimated life expectancy. from a life cycle perspective, examining rwle and rle is reasonable as it highlights the tradeoff between saving (during rwle) and dissaving (during rle). delaying retirement while holding life expectancy constant has the beneficial effect (assuming a life cycle hypothesis perspective) of increasing wealth while reducing the number of years without earned income, which has an overall positive effect on consumption during the remainder of the life. the opposite also holds— retiring early reduces wealth while increasing the number of years without earned income. therefore, coding each period should yield clearer results. to identify the ceteris paribus influence of life expectancy on consumption decisions, a number of variables were included as controls, including relationship status and gender (married, single male, single female, or partner), racial or ethnic status (white, black, hispanic, or other), education (less than high school, high school, some college, bachelor’s degree, or graduate degree), whether or not the respondent has children in the household that are under the age of 18, and health status (poor health vs. good health). other controls are if the respondent contributes to a retirement plan at work, if they have access to $3,000 from a family member or friend in the event of an emergency, their planning horizon (short, intermediate, or long), income (log), net worth (log; only for the saved model), and the degree to which the survey administrator believed that the respondent understood the scf questions (poor/fair, good, or excellent). to account for negative values of net worth, log (0.01) is used. additional controls are whether or not the respondent owns a home, plans to 6 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 give a bequest (spaenjers & spira, 2015), and if they expect to receive an inheritance in the future. 2.3. models ordinary least squares is used to model log credit card balance and log financial assets and binary logistic regression is used to model whether or not the household saved. the coefficients on rle and rwle are the key parameters of interest. after running each of the models on the full sample of full-time workers who expect to retire and have a rle greater than zero (n � 2,632), each model is also run to compare those with a rwle of less than or equal to 10 years (n � 617) with those who have a rwle greater than 10 years (n � 1,212). the secondary analysis helps examine whether there were differences between the two groups based on the saliency of retirement, as suggested by the behavioral life cycle model, and is limited to those between the ages of 35 and 62. other recent studies have used 35 to 60 (kim & hanna, 2015a) and 35 to 70 (kim, hanna, & chen, 2014), stating that those under the age of 35 are likely to have significant changes in jobs and/or marital status, which reduces the quality of the expected retirement measure. given that many respondents expect to retire in their mid-60s (about 30% expect to retire at age 65 and about 56% expect to retire between the ages of 63 and 70), age 62 is used as the cutoff age for the secondary analysis. 3. results 3.1. descriptive results most respondents are in a relationship—53% are married and 11% have a partner. most of the respondents are white (70%) and 48% have completed a college degree. fifty-nine percent of the respondents have a retirement plan at their workplace. most respondents (72%) have access to at least $3,000 in case of an emergency through a family member or friend. half of the respondents (50%) plan for the intermediate-term (2–9 years) while many (35%) plan for the short-term (within a year). approximately half of the respondents have children under the age of 18 (51%). almost all of the respondents are in good health (85%). nearly 48% of the respondents have an excellent understanding of the scf questions while 45% have a good understanding and approximately 6% have a poor or fair understanding of the scf questions. a majority of respondents spend less than they make (63%), while nearly 11% say that they spend more than they earned over the last year. respondents report a median income of $68,988 with a median net worth of $103,980, median financial assets totaling approximately $225,000, and a majority (69%) owned a home. approximately 36% plan to give a bequest while 17% are planning on receiving an inheritance. most continue to carry some level of debt (86%) with a median debt level of $65,600. nearly 47% hold credit card debt with an average balance of $2,792. the average age of the respondent is in the mid-40s (m � 44.45, sd � 5.52) with an average expected retirement age of 63.56 (sd � 2.92), and an average subjective life expectancy of 84.60 (sd � 4.36). 7d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 fig. 1 shows the cumulative distribution of life expectancy for the full sample, and for those near and far from retirement. the range on life expectancy is 39 years to 150 years with approximately 95% of respondents expecting to live somewhere between 70 and 100 years old. given the purpose of this study, even if people have unreasonable expectations about their life expectancy, outliers should not affect our results because we are interested in how their life expectancy perceptions influence consumption decisions. fig. 2 presents the cumulative distribution of expected retirement age for the full sample, and for those near and far from retirement. the range for expected retirement age is 24 years to 102 years of age with nearly 60% expecting to retire between age 63 and 75, and 95% of respondents expect to retire between ages 49 and 79. again, it was considered important to keep the outliers on retirement age because the purpose to see how estimates of life expectancy and retirement age influence consumption decisions during each period. finally, mean remaining work life expectancy (rwle) is 19.11 (sd � 5.06) and the average retirement life expectancy (rle) is 21.05 (sd � 4.96). in the secondary analysis, respondents who are within 10 years of retirement (i.e., near retirement) have an average age of 55.44 (sd � 2.20) with an expected retirement age of 61.91 years (sd � 2.17), whereas those who are further from retirement (11 or more years away) have an average age of 45 years (sd � 2.65) and expect to retire just before age 65 (m � 64.88, sd � 2.16). those who are closer to retirement expect to live a little longer (m � 84.10, sd � 4.17) than those who are further from retirement (m � 83.72, sd � 4.08). fig. 1. cumulative distribution of subjective life expectancy for the full sample (n � 2,632), for those near retirement (n � 617), and for those further from retirement (n � 1,212). fig. 2. cumulative distribution of expected retirement age for the full sample (n � 2,632), for those near retirement (n � 617), and for those further from retirement (n � 1,212). 8 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 respondents expecting to retire in the next 10 years expect to have a longer retirement period (m � 22.19, sd � 4.70) than those who are further from retirement (m � 18.85, sd � 4.21). respondents who expect to retire within the next 10 years expect to work an average of 6.47 years more (sd � 1.12) while those expecting to retire later expect to work an average of 19.85 more years (sd � 2.70). respondents near retirement have slightly more than twice the median net worth ($260,000) than those who are further from retirement ($128,770), and have approximately 42% more financial assets (median � $367,400) than those further from retirement (median � $257,300). about 84% of those near retirement own their home whereas nearly 72% of those 11 or more years away from retirement own their home. of those near retirement, approximately 66% report spending less than they made in the prior year and 11% report spending more than they make, whereas only 60% of those further from retirement report saving and nearly 13.5% spend more than they make. an interesting find was that both groups have a higher prevalence of having debt (87% for those near retirement, 88% those further from retirement) than the full sample (86%) and holding credit card debt (49% for those near retirement, 48% for those further from retirement) than the full sample (47%). absolute dollar amounts of total debt ($72,000 for those near retirement, $94,000 for those further from retirement) and credit card debt ($3,150 for those close to retirement, $3,203 for those further from retirement) are also higher than the full sample ($65,600, $2,792, respectively). approximately 35% of those within 10 years of retirement have minor children in the home whereas 64% of those 11 or more years away from retirement have minor children at home. almost 70% of those nearing retirement have a retirement plan at their place of employment, while only 60% of those not expecting to retire within the decade have such a plan. finally, proportionately more of those who are nearing retirement (56%) plan for the intermediate-term than did those who were further from retirement (49%), while those further from retirement are planning for the short-term proportionately more (35%) than those near retirement (32%). full descriptive results can be seen in table 1. 3.2. multivariate results the regression models are presented in tables 2 through 4 with each table containing the overall model first, followed by the model representing full-time workers ages 35 to 62 expecting to retire in the next 10 years, and ending with the model for full-time workers ages 35 to 62 expecting to retire beyond 10 years. 3.2.1. financial assets considering first the factors that influence the log of financial assets, rle and rwle are significant. contrary to expectations, rle is negatively associated, indicating that a one-year increase in rle is associated with a 1% decrease in financial assets (b � �0.01, se � 0.00, p � 0.01). rwle is also negatively associated with log financial assets—a one-year increase in a respondent’s working career is associated with a 4% reduction in financial assets (b � �0.04, se � 0.00, p � 0.01). however, this makes conceptual sense given that those with a longer time until retirement are likely to have less financial assets. owning a home and indicating an intent to leave a bequest are both positively associated with financial asset 9d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 table 1 sample descriptives of full-time workers: full sample, sample of those near retirement, and sample of those more than 10 years from retirement full model 0–10 years 11� years (n � 2,632) (n � 617) (n � 1,212) variable (reference group) m/% sd m/% sd m/% sd age 44.45 5.42 55.44 2.20 45.03 2.65 retirement age 63.56 2.92 61.91 2.17 64.88 2.16 life expectancy measures subjective life expectancy 84.60 4.36 84.10 4.17 83.72 4.08 retirement life expectancy 21.05 4.96 22.19 4.70 18.85 4.21 remaining work life expectancy 19.11 5.06 6.47 1.12 19.85 2.70 income and asset info incomea $68,988 $83,192 $78,119 net wortha $103,980 $260,000 $128,770 financial assetsa $224,950 $367,400 $257,300 owns home 69.18% 84.29% 71.96% bequest motive 36.02% 35.48% 32.25% receive inheritance 17.41% 15.44% 18.18% debt burden have debt 86.06% 87.05% 88.28% have credit card 76.55% 83.35% 76.81% have credit card balance 46.83% 49.35% 48.41% total debta $65,600 $72,000 $94,000 total credit card balanceb $2,792 $2,932 $3,150 $3,194 $3,203 $3,049 spending behavior spending � income 10.98% 11.02% 13.45% spending � income 25.56% 23.03% 26.28% spending � income 63.45% 65.95% 60.27% relationship status married 53.36% 59.00% 58.26% single male 13.42% 10.51% 12.62% single female 22.78% 25.13% 19.44% partner 10.44% 5.36% 9.69% racial/ethnic status white 69.96% 73.21% 70.07% black 13.96% 13.74% 13.60% hispanic 11.07% 9.60% 10.68% other 5.01% 3.46% 5.65% education no high school diploma 5.96% 8.08% 5.51% high school 20.05% 20.74% 19.20% some college 25.78% 26.04% 25.76% bachelor’s 28.58% 24.79% 29.97% graduate 19.62% 20.36% 19.57% have children � 18 years 50.58% 35.65% 64.27% perceived poor health 14.74% 16.69% 14.22% 10 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 ownership. unmarried individuals have less financial assets than their married counterparts. when compared to white workers, non-white workers also have less financial assets. education is positively associated with financial assets, as is income. those with minor children and/or in poor health are found to have less financial assets than those without children and in good health. full-time workers that are perceived to have an excellent understanding of the scf questions have greater financial assets than those with a good understanding of the questions, while those with a poor-to-fair understanding of the scf questions have less financial assets than do respondents who have a good understanding of the questions. workers with a retirement plan and access to $3,000 are found to have a positive association with financial assets, too. finally, those that have planning horizons of less than 10 years are less likely to accumulate financial assets compared with those who plan for the long term. there are some slight differences in the secondary analysis where the sample is split into those who are near retirement (within 10 years) and those who are far from retirement (more than 10 years). rle and rwle are not significantly associated with financial assets for those who expect to retire within 10 years. however, rle and rwle are significant in determining financial assets for respondents that expect to retire beyond 10 years. again, results seem odd for rle, as an increase of rle by one year is associated with a decrease in financial assets by 2% (b � �0.02, se � 0.01, p � 0.01). an increase in rwle by one more working year results in a 3% decrease in financial assets (b � �0.03, se � 0.01, p � 0.01). 3.2.2. credit card debt among credit card holders, rle and rwle have no significant effect on outstanding credit card balance. home ownership is positively associated with the amount of credit card debt held. single males hold less amounts of credit card debt when compared to their married counterparts. compared to white households, households of other races and ethnicities table 1 (continued) full model 0–10 years 11� years (n � 2,632) (n � 617) (n � 1,212) variable (reference group) m/% sd m/% sd m/% sd understanding of scf’s poor/fair 6.36% 6.45% 5.02% good 45.37% 45.61% 44.32% excellent 48.27% 47.94% 50.66% has retirement plan 58.98% 69.21% 60.54% has access to $3,000 72.71% 74.80% 71.31% planning horizon short 35.49% 31.54% 34.59% intermediate 49.89% 55.50% 49.25% long 14.63% 12.95% 16.17% source: weighted analysis of the 2013 survey of consumer finances. note: analysis includes only those that are full-time, expect to retire, and have a retirement life expectancy greater than zero. a results are reported as the median. b reported as mean because of the median equaling zero. 11d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 (e.g., asian, native american, etc.) have lower balances and hispanic households have higher balances. more educated workers hold more credit card debt than did their peers (with the exception of those having a graduate degree). those who are in worse health have less credit card debt, and those who have a retirement plan have more credit card debt than those who did not have a retirement plan at work. lastly, workers with shorter planning horizons hold more credit card debt than do their counterparts with longer planning horizons. similar results are found in the secondary analysis—rle and rwle are not significantly associated with credit card balances for respondents near or far from retirement. table 2 ols regressions on the log of financial assets, full sample of full-time workers, sample of those near retirement, and sample of those more than 10 years from retirement full model 0–10 years 11� years (n � 2,632) (n � 617) (n � 1,212) variables (reference group) b se. b p b se. b p b se. b p retirement life expectancy �0.01 (0.00) �.001 0.01 (0.01) 0.408 �0.02 (0.01) 0.007 remaining work life expectancy �0.04 (0.00) �.001 �0.02 (0.03) 0.499 �0.03 (0.01) 0.004 own’s home 0.79 (0.10) �.001 0.53 (0.23) 0.020 0.95 (0.14) �.001 bequest motive 0.52 (0.07) �.001 0.32 (0.15) 0.040 0.72 (0.09) �.001 receive inheritance 0.15 (0.09) 0.074 �0.01 (0.15) 0.962 0.09 (0.12) 0.449 relationship status (married) single male �0.31 (0.11) 0.004 �0.58 (0.29) 0.048 �0.46 (0.16) 0.005 single female �0.50 (0.13) �.001 �0.81 (0.31) 0.009 �0.60 (0.19) 0.002 partner �0.26 (0.12) 0.031 �0.27 (0.34) 0.425 �0.01 (0.18) 0.955 racial/ethnic status (white) black �0.64 (0.13) �.001 �0.94 (0.30) 0.002 �0.37 (0.18) 0.042 hispanic �0.48 (0.13) �.001 �0.95 (0.36) 0.009 �0.50 (0.23) 0.029 other 0.04 (0.11) 0.717 0.16 (0.39) 0.673 0.08 (0.14) 0.546 education (no high school diploma) high school 0.87 (0.19) �.001 0.95 (0.39) 0.016 0.56 (0.30) 0.063 some college 1.55 (0.21) �.001 1.52 (0.37) �.001 1.34 (0.31) �.001 bachelor’s 1.97 (0.22) �.001 1.62 (0.42) �.001 1.95 (0.31) �.001 graduate 2.29 (0.22) �.001 1.81 (0.44) �.001 2.28 (0.34) �.001 log of income 0.62 (0.13) �.001 0.73 (0.27) 0.007 0.61 (0.16) �.001 have kids �0.16 (0.04) �.001 �0.26 (0.09) 0.004 �0.17 (0.05) 0.001 perceived poor health �0.21 (0.10) 0.040 0.13 (0.21) 0.540 �0.26 (0.18) 0.148 understand scf questions? (good) poor/fair �0.45 (0.18) 0.015 �0.49 (0.38) 0.202 �0.87 (0.40) 0.029 excellent 0.15 (0.07) 0.024 �0.11 (0.16) 0.495 0.18 (0.10) 0.088 having retirement plan 1.26 (0.07) �.001 1.27 (0.15) �.001 1.38 (0.10) �.001 access to $3,000 0.91 (0.10) �.001 0.75 (0.22) 0.001 0.87 (0.15) �.001 planning horizon (long) short �0.64 (0.11) �.001 �0.84 (0.25) 0.001 �0.48 (0.16) 0.003 intermediate �0.29 (0.09) 0.001 �0.31 (0.21) 0.132 �0.24 (0.11) 0.031 constant 1.28 (1.43) 0.369 0.56 (3.09) 0.857 1.03 (1.81) 0.569 r2 0.57 0.58 0.56 f 142.68 34.58 63.86 source: weighted analysis of the 2013 survey of consumer finances; rii technique and bootstrapping used. note: analysis includes only those that are full-time, expect to retire, and have a retirement life expectancy greater than zero. 12 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 3.2.3. saving behavior looking at factors that influence whether or not a worker saved, rle is again significant and positively associated with the likelihood of saving (b � 0.01, se � 0.00, p � 0.05, odds ratio [or] � 1.01). other positive factors are having a bequest motive, being a single male or in a partnership, income, and net worth. hispanics, having kids, and being in poor health are negatively associated with saving. having access to $3,000 is positively associated with saving, while those with short-term planning horizons (i.e., less than one year), intermediateterm planning horizons (i.e., two to nine years), and credit card debt are less likely to save. a few differences exist in the secondary analysis. for those near retirement, rle is not significant but rwle is significantly and negatively associated with saving. as rwle increases by one year for individuals near retirement, the odds of saving are about 12% lower table 3 ols regressions on the log of credit card balance, full sample of full-time workers, sample of those near retirement, and sample of those more than 10 years from retirement full model 0–10 years 11� years (n � 2,632) (n � 617) (n � 1,212) variables (reference group) b se. b p b se. b p b se. b p retirement life expectancy �0.01 (0.01) 0.338 �0.01 (0.01) 0.399 �0.01 (0.01) 0.469 remaining work life expectancy 0.00 (0.01) 0.627 �0.04 (0.06) 0.529 0.01 (0.02) 0.410 own’s home 0.73 (0.14) �.001 0.42 (0.43) 0.329 0.52 (0.23) 0.024 bequest motive �0.11 (0.14) 0.455 �0.48 (0.29) 0.103 0.10 (0.22) 0.653 receive inheritance �0.24 (0.19) 0.196 0.56 (0.43) 0.197 �0.30 (0.25) 0.228 relationship status (married) single male �0.93 (0.19) �.001 �0.22 (0.41) 0.596 �0.99 (0.39) 0.010 single female �0.41 (0.22) 0.056 �0.94 (0.40) 0.021 �0.01 (0.28) 0.969 partner 0.34 (0.21) 0.116 0.93 (0.72) 0.193 0.50 (0.33) 0.134 racial/ethnic status (white) black 0.22 (0.24) 0.370 1.56 (0.47) 0.001 �0.46 (0.35) 0.199 hispanic 0.73 (0.24) 0.003 0.32 (0.83) 0.699 0.25 (0.37) 0.506 other �0.92 (0.33) 0.005 0.87 (0.73) 0.232 �1.24 (0.45) 0.006 education (no high school diploma) high school 0.80 (0.29) 0.006 �0.16 (0.66) 0.806 1.02 (0.40) 0.010 some college 1.44 (0.31) �.001 0.83 (0.59) 0.158 1.46 (0.41) �.001 bachelor’s 1.24 (0.34) �.001 �0.14 (0.66) 0.833 1.20 (0.45) 0.007 graduate 0.70 (0.36) 0.056 �0.30 (0.72) 0.680 0.99 (0.50) 0.048 log of income 0.07 (0.07) 0.316 �0.53 (0.18) 0.004 0.03 (0.09) 0.783 have kids �0.06 (0.06) 0.324 �0.04 (0.19) 0.840 �0.06 (0.09) 0.483 perceived poor health �0.58 (0.17) 0.001 �1.02 (0.46) 0.027 �0.33 (0.28) 0.234 understand scf questions? (good) poor/fair �0.36 (0.25) 0.157 �1.41 (0.79) 0.075 0.13 (0.44) 0.761 excellent �0.06 (0.14) 0.691 0.07 (0.33) 0.839 �0.12 (0.24) 0.605 having retirement plan 0.46 (0.14) 0.002 0.92 (0.34) 0.006 0.57 (0.21) 0.007 access to $3,000 0.27 (0.15) 0.074 0.82 (0.33) 0.013 0.51 (0.25) 0.039 planning horizon (long) short 0.99 (0.21) �.001 1.25 (0.51) 0.015 0.65 (0.38) 0.087 intermediate 0.59 (0.18) 0.001 0.52 (0.45) 0.253 0.69 (0.36) 0.054 constant 0.78 (0.86) 0.364 8.24 (2.11) �.001 0.95 (1.26) 0.447 r2 0.05 0.09 0.04 f 5.39 2.30 2.06 source: weighted analysis of the 2013 survey of consumer finances; rii technique and bootstrapping used. note: analysis includes only those that are full-time, expect to retire, and have a retirement life expectancy greater than zero. 13d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 (b � �0.13, se � 0.03, p � 0.001, or � 0.88). for individuals further from retirement, rle is significant and positively associated with saving while rwle did not have a significant effect. 4. discussion when breaking down subjective estimates of life expectancy into a retirement life expectancy period and a remaining work life expectancy period, there is confusing, contradictory, and ultimately very little influence on household consumption decisions. retirement table 4 logistic regressions on the likelihood of saving, full sample of full-time workers, sample of those near retirement, and sample of those more than 10 years from retirement full model 0–10 years 11� years (n � 2,632) (n � 617) (n � 1,212) variables (reference group) b se. b p or b se. b p or b se. b p or retirement life expectancy 0.01 (0.00) 0.014 1.01 �0.02 (0.01) 0.054 0.98 0.01 (0.01) 0.045 1.01 remaining work life expectancy 0.01 (0.00) 0.159 1.01 �0.13 (0.03) �.001 0.88 �0.02 (0.01) 0.093 0.98 own’s home 0.10 (0.10) 0.344 1.10 0.46 (0.29) 0.120 1.58 �0.11 (0.16) 0.480 0.89 bequest motive 0.19 (0.08) 0.015 1.21 �0.01 (0.24) 0.959 0.99 0.02 (0.13) 0.894 1.02 receive inheritance �0.14 (0.12) 0.255 0.87 0.29 (0.33) 0.392 1.33 �0.03 (0.16) 0.840 0.97 relationship status (married) single male 0.37 (0.16) 0.017 1.45 1.33 (0.48) 0.006 3.77 0.10 (0.23) 0.676 1.10 single female �0.13 (0.13) 0.328 0.88 0.07 (0.35) 0.848 1.07 �0.34 (0.17) 0.043 0.71 partner 0.28 (0.13) 0.030 1.32 0.92 (0.66) 0.161 2.50 0.16 (0.18) 0.379 1.17 racial/ethnic status (white) black �0.13 (0.12) 0.302 0.88 �0.27 (0.34) 0.430 0.77 0.01 (0.18) 0.947 1.01 hispanic �0.34 (0.11) 0.002 0.71 �0.26 (0.45) 0.558 0.77 �0.45 (0.19) 0.018 0.64 other 0.04 (0.18) 0.818 1.04 �0.89 (0.40) 0.027 0.41 0.35 (0.26) 0.177 1.42 education (no high school diploma) high school �0.09 (0.18) 0.622 0.92 0.47 (0.47) 0.318 1.59 �0.28 (0.33) 0.397 0.76 some college �0.11 (0.21) 0.607 0.90 0.50 (0.47) 0.287 1.64 �0.41 (0.37) 0.272 0.67 bachelor’s 0.34 (0.21) 0.110 1.40 1.34 (0.52) 0.010 3.84 0.08 (0.38) 0.827 1.09 graduate 0.27 (0.22) 0.220 1.31 1.00 (0.55) 0.067 2.72 0.37 (0.40) 0.360 1.45 log of income 0.48 (0.09) �.001 1.61 0.65 (0.19) 0.001 1.91 0.55 (0.13) �.001 1.73 log of net worth 0.04 (0.01) �.001 1.04 0.02 (0.03) 0.513 1.02 0.05 (0.02) 0.004 1.05 have kids �0.20 (0.03) �.001 0.82 �0.27 (0.12) 0.026 0.77 �0.21 (0.05) �.001 0.81 perceived poor health �0.30 (0.12) 0.010 0.74 �0.56 (0.31) 0.067 0.57 �0.56 (0.20) 0.005 0.57 understand scf questions? (good) poor/fair �0.22 (0.16) 0.171 0.81 �0.02 (0.52) 0.962 0.98 �0.02 (0.29) 0.952 0.98 excellent 0.07 (0.08) 0.421 1.07 �0.22 (0.22) 0.319 0.80 �0.03 (0.13) 0.794 0.97 having retirement plan 0.14 (0.09) 0.111 1.15 �0.23 (0.21) 0.268 0.79 0.37 (0.12) 0.002 1.45 access to $3,000 0.27 (0.10) 0.008 1.31 0.13 (0.26) 0.621 1.14 0.16 (0.17) 0.351 1.17 planning horizon (long) short �0.67 (0.15) �.001 0.51 �1.17 (0.45) 0.010 0.31 �0.53 (0.22) 0.016 0.59 intermediate �0.46 (0.14) 0.001 0.63 �0.99 (0.37) 0.008 0.37 �0.30 (0.21) 0.152 0.74 have credit card balance �0.62 (0.07) �.001 0.54 �1.13 (0.18) �.001 0.32 �0.64 (0.11) �.001 0.53 constant �4.80 (1.01) �.001 0.01 �4.77 (2.25) 0.034 0.01 �5.06 (1.33) �.001 0.01 pseudo r2 0.14 0.21 0.17 log likelihood 473.51 166.81 277.85 source: weighted analysis of the 2013 survey of consumer finances; rii technique and bootstrapping used. note: or � odds ratio; analysis includes only those that are full-time, expect to retire, and have a retirement life expectancy greater than zero. 14 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 life expectancy is not a significant predictor in our primary analysis of credit card debt. additionally, rwle is not a significant predictor in our primary analysis of credit card debt and whether or not a respondent saved. even when separating individuals into those who are near and far from retirement, rle is not significantly associated with financial assets, credit card debt, and savings behavior for those near retirement, and it is not significantly associated with credit card debt for those further from retirement. rwle, on the other hand, is not significantly associated with financial assets and credit card debt for those near retirement, and is not associated with credit card debt and savings behavior for those further from retirement. rle exhibits effect directions contrary to expectations in financial assets for both the primary and secondary analysis. subjective life expectancy was only positively associated with saving behavior for those far from retirement. although the life cycle hypothesis suggests that life expectancy should influence household consumption choices, the behavioral life cycle model suggests that there may be other factors that prevent households from smoothing consumption according to the life cycle hypothesis. for example, even if a household realizes they may live longer, holding retirement age constant, they may lack the self-control or willpower to change consumption today to smooth consumption over the remainder of their lifetime. the negative direction of the rle factor on financial assets for the overall model may indicate what previous literature has found—u.s. consumers are not doing a good job at preparing for retirement (aegon, 2016; kim & hanna, 2015b; yuh, montalto, & hanna, 1998). additionally, the fact that rle and rwle are not significant for those near retirement may show that people are not able to account for their life expectancy when making consumption choices. alternatively, it could be that these households believe that, although retirement is salient, there is nothing they can do, so they do not factor in their rle and rwle when making consumption decisions related to their financial assets. finally, because rle is significant, but negatively associated with financial assets for those who plan to retire more than a decade from now, these individuals may be making consumption choices that are not in-line with the life cycle hypothesis. that is, there may be a self-control or willpower problem associated with their ability to properly plan for retirement. 4.1. limitations the most serious limitations center on measurement issues. although the authors feel that the best available consumption measures in the scf are chosen, the analysis would be improved with better measures of consumption. the saving measure is fairly limited as it is simply whether or not the respondent spent less than they made last year, and it assumes that the respondent actually knows whether or not they really did spend less than they made. another limitation is that the secondary analysis age group restriction may be limited because of the fact that those in their mid-to-upper 30s, and even those in their mid-to-low 40s, are likely considerably different than those in their 50s and early 60s, regardless of rwle. for example, someone that is 40 years of age with an rwle of 10 may be quite different than a 55-year old with an rwle of 10. 15d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 4.2. implications the results of this study have implications for professionals who help individuals and families plan for retirement, employers who offer retirement plans, and policymakers implementing legislation on retirement plans. financial planners, financial counselors, and financial therapists have an important role to play in helping clients plan for retirement and/or make decisions around consumption choices that influence retirement preparedness. given that it is likely that clients are not taking into account their life expectancy, particularly the amount of time they expect to spend in retirement, these professionals need to carefully help clients consider how their consumption choices impact their future financial resources available to them and how these choices may negatively influence their retirement lifestyle. having conversations early and often, and holding clients responsible for their actions is of utmost importance. professionals can help clients stick to the plan and illustrate the longterm impact of short-term consumption. this may help the client to think twice before making a choice that is detrimental to their financial assets available for retirement. second, these results may have meaningful impact for human resources directors and those who manage retirement plans. given that we have shifted from a defined benefit (i.e., pensions) system to a defined contribution (i.e., 401(k)s) system, and that willpower and self-control may be an issue, incorporating programs that automatically opt-in employees to retirement plans, as well as automatically increase the employees’ initial level of savings to some maximum reasonable percentage, may be a welcome adjustment. previous research has found that this is a significant way to help increase employees’ savings (beshears, choi, laibson, & madrian, 2009; thaler & benartzi, 2004). although our results are not conclusive, the general lack of impact of life expectancy on consumption choices suggest there is a need for policies aimed at improving retirement preparation in the united states. if our result is explained by willpower and self-control problems, legislation requiring employers to, at minimum, automatically enroll their employees into defined contribution plans seem beneficial. according to aegon (2016), three-quarters (76%) of respondents agreed on some level that the government should mandate employers to auto-enroll employees into the defined contribution plans. this article contributes to the literature by showing that even if consumers are considering subjective life expectancy and their work life and retirement periods, they are doing a poor job. this may be due, in part, to self-control and willpower issues in delaying current consumption to provide consumption during retirement. the most likely explanation is that consumers do not consider their life expectancy and associated work and retirement periods when making consumption choices, resulting in a failure to adjust their consumption levels over the course of their life. subsequently, this results in a lack of retirement preparedness for these individuals. notes 1 for example, using an 80% retirement income replacement rate for our sample’s median income ($68,988) results in $55,190 in annual income needs during retirement. 16 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 based on the median financial assets of our sample ($224,950), a 24.5% annual return would be needed to fund this retirement lifestyle in perpetuity ($55,190/$224,950). alternatively, a portfolio of approximately $1 million would be needed to earn the amount necessary for retirement spending, given the assumptions above (based on a more realistic 5.5% annual return; $1m*.055 � $55,000). hence, it is plausible that many households do not have the resources available to fund such a retirement income requirement, and that current financial asset income would not be sufficient to fund retirement income needs. 2 spending relative to income (less, equal to, or more) in a multinomial logit was modeled. results did not differ so the simpler (binary) model is presented. references aegon. (2016). a retirement wake-up call: the aegon retirement readiness survey 2016: united states country report. aegon: center for longevity and retirement (available at http://www.aegon.com/ documents/aegon-com/research/2016-retirement-survey/2016-aegon-retirement-readiness-report-usa.pdf). ando, a., & modigliani, f. (1963). the “life cycle” hypothesis of saving: aggregate implications and tests. the american economic review, 53, 55–84. beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2009). the importance of default options for retirement savings outcomes: evidence from the united states. in j. r. brown, j. b. liebman, & d. a. wise (eds.), social security policy in a changing environment (pp. 167–195). chicago, il: university of chicago press. bloom, d. e., canning, d., & graham, b. (2003). longevity and life-cycle savings. scandinavian journal of economics, 105, 319–338. bloom, d. e., canning, d., & moore, m. (2004). health, longevity, and optimal retirement (available at ssrn: http://dx.doi.org/10.2139/ssrn.594801). bloom, d. e., canning d., moore, m., & song, y. (2006). the effect of subjective survival probabilities on retirement and wealth in the united states (no. 13688). cambridge, ma: national bureau of economic research. deaton, a., & paxson, c. (2000). growth, demographic structure, and national savings in taiwan. population and development review, 26, 141–173. elder, t. e. (2007). subjective survival probabilities in the health and retirement study: systematic biases and predictive validity. working paper 2007–159, university of michigan, michigan retirement research center. gupta, a., & li, z. (2013). optimal annuity purchase decisions under uncertain lifetime. journal of family economic issues, 34, 447–459. doi:10.1007/s10834–012-9340–1 hamermesh, d. (1985). expectations, life expectancy, and economic behavior. quarterly journal of economics, 100, 389–408. hurd, m. d., & mcgarry, k. (2002). the predictive validity of subjective probabilities of survival. economic journal, 112, 966–985. kim, k. t., & hanna, s. d. (2015). does financial sophistication matter in retirement preparedness? journal of personal finance, 14, 9–20. kim, k. t., & hanna, s. d. (2015). do u.s. households perceive their retirement preparedness realistically? financial services review, 24, 139–155. kim, k. t., hanna, s. d., & chen, s, c-c. (2014). consideration of retirement income stages in planning for retirement. journal of personal finance, 13, 52–64. lindamood, s., hanna, s. d., & bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41, 195–222. 17d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. the collected papers of franco modigliani (vol. 6, pp. 3–45). cambridge, ma: the mit press. montalto, c. p., yuh, y., & hanna, s. (2000). determinants of planned retirement age. financial services review, 9, 1–15. montalto, c. p., & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances. financial counseling and planning, 7, 133–141. nielsen, r. b. (2015). scf complex sample specification for stata. technical note. department of financial planning housing and consumer economics, university of georgia, athens, ga. nielsen, r. b., & seay, m. c. (2014). complex samples and regression-based inference: considerations for consumer researchers. the journal of consumer affairs, 48, 603–619. doi:10.1111/joca.12038. shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609–643. shin, s. h., & hanna, s. d. (2016). accounting for complex sample designs in analysis of the survey of consumer finances. the journal of consumer affairs, 1–15. doi:10.1111/joca.12106. spaenjers, c., & spira, s. m. (2015). subjective life horizon and portfolio choice. journal of economic behavior & organization, 116, 94–106. statista (2016). media advertising spending in the united states from 2011 to 2018 (in billion u.s. dollars; available at https://www.statista.com/statistics/272314/advertising-spending-in-the-us/). thaler, r. h. & benartzi, s. (2004). save more tomorrow: using behavioral economics to increase employee saving. journal of political economy, 112, s164–s187. tsai, i-j., chu, c. y. c., & chung, c-f. (2000). demographic transition and household saving in taiwan. population and development review, 26, 174–193. tversky, a., & kahneman, d. (1974). judgement under uncertainty: heuristics and biases. science, 185, 1124–1131. yuh, y., & hanna, s. d. (2010). which households think they save? the journal of consumer affairs, 44, 70–97. yuh, y., montalto, c. p., & hanna, s. d. (1998). are americans prepared for retirement? financial counseling and planning, 9, 1–13. 18 d.r. lawson, s.j. heckman / financial services review 26 (2017) 1–18 ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. cfp board recently adopted revisions to several provisions of its ce policies, including changing the minimum number of questions for self-study assessments from 10 to 5 per full ce credit hour. therefore, financial services review ce exams will have 5 questions. ce credit for this issue expires september 30, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in "the relationship between objective financial knowledge, financial management, and financial self-efficacy among african american students” by white, park, watkins, mccoy, and morris," the authors suggest ______ may be most useful to help african american students' increase their financial literacy levels. a. a traditional introductory course on personal finance b. a series of university sponsored personal finance workshops c. better crafted student loan seminars d. experiential learning opportunities 2. in white, park, watkins, mccoy, and morris, which of the following is/are associated with african american college students’ financial self-efficacy? a. objective financial knowledge only b. financial management only c. both objective financial knowledge and financial management d. none of the above 3. in white, park, watkins, mccoy, and morris note that higher self-efficacy affects the following: a. saving behavior b. increases in net worth c. healthy financial management d. all of the above 4. in “enumerating the value of financial advice in a competitive market – a dual structure approach and analysis” by fraser, payne & schatzle.all but which of the following do the authors suggest is a potential advantage of a consolidated dual-fee structure (cdfs)? a. a cdfs can improve the transparency in pricing for separate services b. a cdfs can quickly and completely solve the fiduciary debate c. a cdfs recognizes the scope of financial planning complexity d. a cdfs can reduce the incentive for planners to increase aum 5. in fraser, payne & schatzle. the consolidated dualfee structure (cdfs) proposed and illustrated in the study identifies various “levers” a planner or planning firm might consider and adjust to suit their specific practice. which of the following do the authors suggest should be considered when operationalizing a cdfs? a. the break points and respective fees for each of the parallel, regressive fee structures b. the percentage reduction in the planning fee after the initial year c. the size and scope (ratio) of net worth (nw) to investable assets (aum) a. all of the above are important considerations when implementing a cdf manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) the impact of using financial technology on positive financial behaviors qianwen bia, lukas r. dean*,b, tao guoc, xu sund adepartment of finance and economics, woodbury school of business, utah valley university, 800 w. university parkway, orem, ut 84058, usa bdepartment of finance and economics, woodbury school of business, utah valley university, 800 w. university parkway, orem, ut 84058, usa cdepartment of economics, finance and global business, cotsakos college of business, william paterson university, 1600 valley road, wayne, nj 07470, usa ddepartment of finance and economics, woodbury school of business, utah valley university, 800 w. university parkway, orem, ut 84058, usa abstract this study uses 2013 survey of consumer finances data to explore the impact of financial technologies on households’ positive financial behaviors. after controlling for variables on general capitals, financial literacy capitals, and financial resources, we find that only planning technologies (e.g., direct deposit and computer software) are positively related to households’ engagement in positive financial behaviors. in contrast, the impact of transaction technologies (e.g., using atm card, credit card, phone banking, and computer banking) is negative. policymakers and financial service providers should assist consumers with better financial tools and help them manage financial resources and behaviors. © 2021 academy of financial services. all rights reserved. keywords: financial technology; financial software; positive financial behaviors 1. introduction technological innovation plays an important role in the development of the financial services industry. the way households manage finances now has changed rapidly over the past decade as a variety of technologies, such as electronic banking and automated advisers, have been designed to help households achieve better financial well-being. however, today’s *corresponding author. tel.: +1-801-863-8236; fax: +1-801-863-7218. e-mail: luke.dean@uvu.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 29–54 financial world is much more complicated because of the wide range of financial products and services (parrish & servon, 2006). most of these products and services are associated with e-banking products and services provided by banks and financial institutions, such as automated teller machine (atm), credit cards, direct deposit, preauthorized debit, phone banking, online banking, and so forth. a recent report on consumer mobile banking finds that over three-fourths of the u.s. population now has a smartphone (pew research center, 2018), and about 50% of users have used mobile banking in the past 12months (merry, 2018). it is said that electronic banking technologies have improved the effectiveness of distribution channels by reducing the transaction costs and service time (lee & lee, 2001) and expanding credit access in consumer lending (jagtiani & lemieux, 2018). however, there is mixed evidence regarding whether e-banking technologies are helpful in managing household finance. some research finds that electronic banking technologies help consumers save time in managing their finances through easier access to financial services (anguelov, hilgert, & hogarth, 2004). other studies find that consumers are concerned about security issues associated with online banking (hamlet & strube, 2000). although many researchers have investigated the determinants for consumers to adopt internet banking (kim, widdows, & yilmazer, 2005; jun & cai, 2001; lee, lee, & eastwood, 2003; lee & lee, 2000), the literature on the impact of electronic technology is very limited. son and hanna (2011) find that the internet affects how consumers use financial services and how consumers search and evaluate financial information before making financial decisions. evidence shows that technological training and e-banking support financial literacy (servon & kaestner, 2008). while the question is: are households who adopt these transaction-based technologies improving their financial well-being status? in this article, we use the term “financial technology” to include both electronic technology and computer software technology, and examine the impact of using financial technology on households’ positive financial behaviors. we hypothesize that different types of financial technologies affect household financial behaviors differently. specifically, we use a life cycle and human capital theoretical framework and differentiate two types of financial technologies: transaction-based financial technologies (i.e., atm card, credit card, phone banking, and computer banking), and planningbased financial technologies (i.e., preauthorized debit, direct deposit, and computer software use). we hypothesize that transaction-based financial technologies should be negatively related to positive financial behaviors as they greatly increase the ease of accessing financial capital and the risks of overspending as well, especially for customers with self-control issues. in contrast, planning-based financial technologies should be positively associated with financial behaviors as households with clear planning goals are more likely to save for the future efficiently. to test the hypotheses, we construct our sample using the data from the 2013 survey of consumer finances (scf). following previous studies in dew and xiao (2011), xiao et al. (2007), worthy et al. (2010), and hayhoe et al. (2000), we first measure positive financial behaviors based on 13 financial behaviors. we then create three proxies for positive financial behaviors using different methodologies including principal component analysis based on these thirteen financial behaviors. to test how different types of financial technologies affect household financial behaviors, we also use principal component analysis to construct a 30 q. bi et al. / financial services review 29 (2021) 29–54 measure of transaction-based technology usage and a measure of planning-based technology usage, respectively. last, we include a vector of other variables to control for the impact of other factors that are likely to affect household financial behaviors. our univariate analysis suggests that 81.9% of respondents use an atm card, 72.4% of respondents use a credit card, 21.2% of respondents use phone banking, 67.5% of respondents use computer banking, 57% of respondents used preauthorized debit, 86.4% of respondents use direct deposit, and 21% of respondents use computer software to manage their household finances. these findings are consistent with previous studies demonstrating that household adoption of electronic technologies has expanded substantially (anguelov, hilgert, & hogarth, 2004; servon & kaestner, 2008). for the multivariate analysis, the results are consistent with our hypotheses that not all currently used financial technologies contribute to positive financial behaviors and personal financial wellbeing. households that use planning purposed financial technologies are significantly related to higher positive financial behaviors while households using transactional purposed financial technologies are negatively related to positive financial behaviors. these results are robust after controlling for various household-level and economic factors, as well as to different measures of financial behaviors. our findings suggest that planningbased financial technologies appear to create a more positive environment and enhance household financial well-being. previous research has shown that the very act of monitoring progress promotes better goal attainment and behavioral moderation (harkin et al., 2016). the present study demonstrates that technological innovations like computer banking that assimilate planningbased financial technology have the potential to improve households’ financial wellbeing and promote savings behavior and progress towards long-term goals (e.g., saving for retirement). u.s. households are facing an increasingly complex world that places more responsibility on their shoulders. our findings in this study suggest that policymakers, financial planning professionals, employers, and financial service providers should find means and methods to assist consumers with better financial tools and help households manage their financial resources and behaviors so that households can enjoy the peace of mind and security from financial well-being. the rest of the article is organized as follows. section 2 reviews related literature and develops our hypotheses. we show data and variable constructions in section 3. section 4 presents the results and discusses the findings. finally, section 5 provides the conclusions. 2. literature review and hypothesis development 2.1. positive financial behavior garman, leech, and grable (1996) define poor financial behaviors as personal and family money management practices that have consequential, detrimental, and negative impacts on one’s life at home and/or work. prior research finds that positive financial behaviors are associated with positive life outcomes (shim et al., 2009) and negative financial behaviors tend to cede to more negative financial behaviors (dean et al., 2013). given the common q. bi et al. / financial services review 29 (2021) 29–54 31 financial activities that households need to deal with, we consider positive financial behaviors in three categories: debt management, planning activities, and risk management. 2.1.1. debt management consumer credit use plays an important role in how modern households handle their debt. among all household debt management activities, the most obvious positive financial behavior is to pay bills on time. however, almost seven percentage of u.s. households reported having at least one payment in the past year that was at least 60 days late (hogarth & anguelov, 2004). other than loans and mortgages, credit card use is another financial activity that will cause consumers into debt. in 2008, the total outstanding credit debt carried by americans is about $976 billion (federal reserve, 2009). according to the federal reserve (2013), credit card transactions increased at a 7.6% annual rate, rising from $21 billion in 2009 to $26.2 billion in 2012. the average number of credit cards that u.s. credit users hold is more than five and the average balance for each card is at least $1,000 (experian, 2009). recently, a report by transunion (2019) shows that bank-issued (private) credit card balances increased to $5,668 ($2,022) on a personal level as of q3 2019.1 rutherford and devaney (2009) define two types of credit card users: convenience users and revolvers. convenience users are those who pay the balance in full on a regular basis while revolvers are those who pay only a portion of the balance and let the remaining balance accrue interest. other research suggests that high credit card balances are a result of behavior problems instead of liquidity problems (gross & souleles, 2001). thus, carrying a balance in credit cards while having money in the checking account is not considered a positive financial behavior because one has to pay high interest on the credit card balance and the checking account provides no or little interest. moreover, making late payments on credit cards will lead to late fees and a negative remark on the credit report. it is worth mentioning that consumers with a history of late payments are less likely to be convenience users (rutherford & devaney, 2009), as they tend to pay off small debts first even when the larger debt have higher interest rates (amar et al., 2011). the high debt-payment-to-income ratio is another detriment to households’ lives. according to the scf, 11% of all families in the united states had debt-payment-toincome ratios greater than 40% in 2001. this number increased to 15% in 2007 and 18.5% in 2010. bricker et al. (2017) examined changes in financial management of u.s. families and reported that over 20% of families felt constrained in credit though their access to consumer credit has been increased. aizcorbe et al. (2003) point out that most of the debtors who had greater than 40% debt-payment-to-income ratios were from lower-income families. severe consequences such as bankruptcy might happen to these families in the long-run if they are unable to make adjustments on their debt management. late payments along with high debt-payment-to-income ratios will negatively affect credit scores and limit future possibilities (i.e., access to credit, housing, or employment) of exhibiting positive debt management behaviors. 32 q. bi et al. / financial services review 29 (2021) 29–54 2.1.2. planning activities according to life cycle theory, individuals are saving and borrowing to smooth out consumption over lifetime (modigliani & brumberg, 1954). the assumption of life cycle theory is that people are forward-looking and making plans for the future. planning for the future is a positive financial behavior as it allows households to smooth out consumption to maximize lifetime utility. to effectively transfer financial resources from one period to another, individuals need to have an extensive planning process. rutherford and devaney (2009) suggest that financial advisors and educators must encourage and assist households in preparing financial plans that extend beyond five years. they find that households who have financial planning horizons of at least five years are more likely to be convenience users of credit. stawski et al. (2007) find that goal clarity serves as an important psychological mechanism, which motivates individuals to plan for the future. for example, neukam and hershey (2003) demonstrate that financial goals have a significant impact on retirement savings contributions. specifically, goals help individuals structure perceptions and form expectations about future resource needs so they help increase both actual savings levels and the intention to save (stawski et al., 2007). moreover, households who had not engaged in planning activities are significantly less likely to accumulate wealth (ameriks et al., 2002; lusardi, 2010). another class of financial planning activities involves information-seeking, especially when individuals are shopping for credit, savings, and investment products. lee and hogarth (1999) find that households who extensively search when shopping for credit are more likely to have lower aprs and more likely to solve their credit card debts. consumers who take more effort to shop for credit are likely to find loans with good terms and conditions as well as being more likely to be convenience credit users (rutherford & devaney, 2009). the more exploration a household does when making financial decisions on credit, savings, and investments, the more likely it is going to get a better deal. 2.1.3. risk management households are vulnerable when facing a variety of unexpected events that could lead to serious financial difficulties. insurance is a tool to protect against substantial financial losses when unplanned perils or health circumstances occur. thus, insurance is an important aspect of personal financial management. lin and grace (2007) find that financial vulnerability has a significant impact on the amount of term life or total life insurance purchased. they argue that the key determinant of the demand for life insurance is the impact of the insured’s death on the future consumption of other household members. households with dependent children are more financially vulnerable because children consume most resources but generally contribute little to the household income. preparing for the potential loss of the breadwinner of the household is very beneficial especially for households with children under 18 present (lewis, 1989). evidence shows that around two-thirds of poverty among surviving women and more than one-third of poverty among surviving men result from failures to insure survivors against sudden loss of household head (bernheim et al., 2001). although other research finds that life insurance q. bi et al. / financial services review 29 (2021) 29–54 33 is essentially uncorrelated with financial vulnerability at every stage of the life cycle (bernheim et al., 2003), we consider protection for your dependents as a positive financial behavior because it protects the family from the financial shock of losing a breadwinner. households need to insure against the loss of health-related human capital of its earners to ensure viability. however, health insurance products are too complicated for most households. according to data from the 1977 national medical expenditure survey, 4.3% of nonelderly families spent more than 20% of their income on health care (feenber & skinner, 1994). health expenditure shocks can lead to households’ bankruptcy (livshits, tertilt, & macgee, 2007) yet only high-income households accumulate precautionary savings to shield themselves from catastrophic health expenditures (jeske & kitao, 2009). other than health-related human capital loss, households are facing temporary or permanent disability risks as well. in theory, disability insurance provides benefits to workers who are physically unable to find suitable work. although programs like social security disability insurance (ssdi) and supplemental security income (ssi) are designed to help workers with disabilities, these benefits are not enough to lift incomes above the poverty line (stapleton et al., 2006). between 1985 and 2004, the number of disabled individuals receiving disability insurance increased by over 100% (chen & van der klaauw, 2008). however, ssdi only pays benefits to “insured” workers and certain members of their family. in other words, workers with a disability need to have worked and paid social security taxes for a long enough period to receive disability benefits from ssdi. a recent fact sheet from the social security website shows that about over one in four of today’s 20 year-olds will become disabled before reaching age 67 (ssa, 2019) (https://www.ssa.gov/pubs/en-0510029.pdf). in addition, 68% of the private sector workforce has no long-term disability insurance (ssa, 2019). because the purpose of disability insurance is to provide substitute income to workers with disabilities, individuals need to be covered with disability insurance if they are currently working and should not need coverage if they have retired. in addition to protecting against health issues, disability, and potential loss of life—unexpected events could also place households into financial difficulties. setting aside a bucket of money to prepare for rainy days is imperative so households do not have to sell off their cars, appliances, and other household durables (huston & chang, 1997). a three-month income reserve is used as an adequate holding of an emergency fund in household emergency fund research (chang & huston, 1995; huston & chang, 1997). 2.2. financial technology electronic banking technologies include atm, online banking, debit (or check) card, direct deposit, direct payment (also electronic bill payment), electronic bill presentment and payment (ebpp), electronic check conversion, electronic fund transfer (eft), payroll card, preauthorized debit (or automatic bill payment), prepaid card, smart card, and stored-value card (anguelov, hilgert, & hogarth, 2004). because the diffusion of innovation has not been applied to financial innovations, the current understanding of electronic banking technology, such as atm card, debit card, direct deposit, and direct payment is very limited (lee & lee, 2000). 34 q. bi et al. / financial services review 29 (2021) 29–54 anguelov, hilgert, and hogarth (2004) use three specific technologies to represent different types of e-banking technologies at different stages in their development: debit cards, preauthorized debits, and electronic banking. computer ownership and internet access are related to the adoption of electronic banking but many studies have been unable to control for those variables. consumers’ acceptance of technological innovations is influenced by socioeconomic characteristics, demographic characteristics, perceptions of specific technologies, and the characteristics of different products and services. electronic banking technologies can be classified as either “passive” or “active” (kolodinsky, hogarth, & hilgert, 2004). passive technologies (i.e., direct deposit and preauthorized debit) do not require any behavioral changes or continuous effort by the consumer so it is easier to spread. in contrast, active technologies (i.e., electronic banking) require new behaviors or repeated effort so they are hard to spread (kolodinsky et al., 2004; servon & kaestner, 2008). davis (1989) created the technology acceptance model (tam) that shows that perceived usefulness and ease of use are factors associated with the adoption of a system. interconnections between technologies exist because the diffusion of any technology is not independent of the diffusion of another technology (stoneman & kwon, 1994). moreover, a consumer’s prior pattern of adopting related technology will affect his or her willingness to adopt new technology (bayus, 1987). consumers with good knowledge of computers are generally more likely to engage in electronic banking usage. demographic factors such as age, income, education, occupation are significant factors for internet banking adoption as well (kim et al., 2005). when relating financial technology to financial behaviors, we can categorize financial technologies into two main functions—technologies that are fundamentally transactional and technologies that aid in planning. transaction-based technologies include atm cards, credit cards, phone, and electronic banking. households use an atm card to access their bank account at an electronic terminal without the limitations of finding the nearest branch of their local bank. this is especially useful for transactions when traveling. it is more common to see consumers use credit cards to complete transactions for online purchases now than 50 years ago. credit cards make online transactions convenient and safe by allowing households to set up transaction alerts on their mobile device and monitor spending instantaneously. households also use preauthorized debit to set up electronic auto-payments on loans. phone banking and electronic banking provides 24/7 financial service with almost no cost. using these e-banking technologies, households can access their account information with little or no cost and conduct financial transactions conveniently (lee & lee, 2001). another important reason for using financial technology is to plan for the future. research finds that computer-based mediated interventions contribute to a variety of behavior changes. behavioral modification strategies include self-monitoring, goal-setting, shaping, reinforcement, and stimulus control (butryn, webb, & wadden, 2011). the very act of monitoring progress towards goals has demonstrated evidence of significant improvement towards behavioral changes and the actual outputs desired (harkin et al., 2016). household finance-related computer software can help households make better financial decisions by providing financial knowledge and information, enhancing numerical ability on calculations, and monitoring finance on a regular basis. governments and organizations use direct deposit as the preferred way to make reimbursements and distributions on salaries. households who q. bi et al. / financial services review 29 (2021) 29–54 35 use preauthorized debit and direct deposit are trying to simplify their financial management, and automate their savings for the future. the conceptual framework used for this study builds upon life cycle theory (modigliani & brumberg, 1954) and human capital theory (becker & tomes, 1994). according to the life cycle theory, an individuals’ objective is to maximize lifetime utility. individuals try to achieve higher lifetime utility through improved financial well-being. in our article, we measure household financial well-being through positive financial behaviors regarding debt management, planning activities, and risk management. the positive financial behavior (pfb) index in fig. 1 is a set of household financial behaviors that will lead to positive financial outcomes. household financial behaviors are affected by household human capital, the endowed and acquired knowledge and skills a household has (huston, 2010). following previous studies, we consider two types of household human capital: specific human capital and general human capital (becker & tomes, 1994), as illustrated in fig. 1 specific human capital represents knowledge and skills that a household has towards specific areas such as financial technology and financial management. specifically, financial technology indicates how well a household can understand and potentially use technology-related products and services to increase the probability of increased positive financial behaviors. financial literacy indicates how well an individual can understand and potentially use personal finance-related information to increase expected lifetime utility from consumption (huston, 2010). in contrast, general household capital represents the knowledge and skills that a household has and fig. 1. household financial behavior conceptual framework. 36 q. bi et al. / financial services review 29 (2021) 29–54 could be used in many areas. for example, households who have more education are more likely to perform better in a wide variety of tasks. last, household financial behaviors are also influenced by other factors such as behavioral biases, self-control, family, peer, cultural, environmental, and economic conditions (huston, 2010). financial resources also impact the household’s financial well-being. households with more positive financial behaviors are more likely to have a higher level of financial well-being, after controlling for financial resources and other influences. based on our theoretical framework in fig. 1, positive financial behavior is a function of household capital, cultural/environmental influences, economic status: positive financial behaviour pfbð þ ¼ ffspecifichouseholdcapital; generalhousehold capital; cultural= environmental influences; economic stat using this conceptual framework, we test the impact of specific household capital (financial technology) on positive financial behaviors, controlling for all other factors, using the following empirical regression model. pfb = b 0 + b 1(specifichouseholdcapital ) + b 2(generalhousehold capital) + b 3-6(financial sophistication level) + b 7(homeownership) + b 8-10(age groups) + b 10-12(education level) + b 13(married) + b 14-17(income quintiles) + b 17-20 (net worth quintiles) + b 21-22(household size) + b 23(presence of children under 18) + b 24(female) + b 25-26(race groups) + b 27(economic expectation) + b 28(interest rate expectation) + b 29-30(risk tolerance) + e (2). 3. data and variables descriptions this study uses data from the 2013 survey of consumer finances (scf).2 the scf is a triennial survey of u.s. households sponsored by the federal reserve, in cooperation with the internal revenue service, statistics of income division, and collected by norc at the university of chicago. the survey data includes information on families’ balance sheets, pensions, income, and demographic characteristics. information is also included from related surveys of pension providers and the earlier such surveys conducted by the federal reserve board.3 in the 2013 scf survey, 6,015 households were available in the public dataset. while our study focuses on the use of financial technology and requires households to have at least a checking or a savings account. as a result, we censored the data to include banked households only for the purpose of this study. after applying this filter, our final sample includes a total of 5,447 households. 3.1. positive financial behaviors proxies: the dependent variable based on the theoretical framework, we identify thirteen financial behaviors from the scf to construct our positive financial behavior proxies. these behaviors cover three q. bi et al. / financial services review 29 (2021) 29–54 37 t ab le 1 p er ce n ta g e o f b an k ed h o u se h o ld s en g ag in g in p o si ti v e fi n an ci al b eh av io r p o si ti v e fi n an ci al b eh av io r m ea su re m en t b an k ed h o u se h o ld s (% ) s ta n d ar d iz ed sc o ri n g co ef fi ci en ts d eb t m an ag em en t n o la te p ay m en ts 1 if al l lo an an d m o rt g ag e p ay m en ts m ad e o n ti m e o r ah ea d o f ti m e, 0 o th er w is e. 8 5 .9 0 .2 1 9 g o o d cr ed it re p o rt 1 if n o t b ee n tu rn ed d o w n fo r cr ed it o r if tu rn ed d o w n b u t re ce iv ed fu ll am o u n t w h en th ey re ap p li ed o r n ev er ap p ly w it h in th e p as t fi v e y ea rs b ec au se o f af ra id o f b ei n g tu rn ed d o w n , 0 o th er w is e. 7 6 .4 0 .2 2 7 c re d it ca rd b al an ce 1 if n o t ca rr y in g cr ed it ca rd b al an ce w h en h av in g m o n ey in b an k ac co u n ts , 0 o th er w is e. 6 1 .5 0 .1 0 1 n o b an k ru p tc y 1 if n ev er fi le d fo r b an k ru p tc y , 0 o th er w is e. 8 6 .5 0 .1 4 4 t o ta l d eb t p ay m en t ra ti o 1 if to ta l d eb t p ay m en t ra ti o sm al le r th an 3 6 % , 0 o th er w is e. 8 9 .6 0 .1 1 1 p la n n in g ac ti v it ie s p la n n in g h o ri zo n 1 if p la n n in g h o ri zo n is a fe w y ea rs o r m o re , 0 o th er w is e. 5 8 .5 0 .2 2 7 c u rr en tl y sa v in g 1 if h av e at le as t o n e re as o n to sa v e, sp en d in g is le ss th an in co m e, an d ac tu al ly h av e sa v ed fo r th at re as o n , 0 o th er w is e. 5 8 .6 0 .2 7 6 l ev el o f sh o p p in g fo r cr ed it 1 if w h en m ak in g m aj o r d ec is io n s ab o u t cr ed it o r b o rr o w in g , d o a m o d er at e to a g re at d ea l o f sh o p p in g , 0 o th er w is e. 7 5 .1 0 .0 9 9 l ev el o f sh o p p in g fo r sa v in g s an d in v es tm en ts 1 if w h en m ak in g m aj o r d ec is io n s ab o u t sa v in g o r in v es ti n g , d o a m o d er at e to a g re at d ea l o f sh o p p in g , 0 o th er w is e. 6 9 .8 0 .1 4 5 r is k m an ag em en t c h il d re n u n d er 1 8 co v er ed b y li fe in su ra n ce 1 if ch il d re n u n d er 1 8 co v er ed b y li fe in su ra n ce o r n o ch il d re n u n d er 1 8 p re se n t in th e h o u se h o ld , 0 o th er w is e. 8 9 .0 0 .1 8 4 h ea lt h in su ra n ce 1 if ev er y o n e in th e h o u se h o ld is co v er ed u n d er h ea lt h in su ra n ce , 0 o th er w is e. 8 1 .1 0 .2 1 9 d is ab il it y in co m e in su ra n ce 1 if h ea d o f h o u se h o ld is w o rk in g an d co v er ed b y d is ab il it y in su ra n ce o r is re ti re d an d n o t co v er ed b y d is ab il it y in su ra n ce , 0 o th er w is e. 2 5 .4 0 .1 2 6 e m er g en cy fu n d 1 if h av in g $ 3 ,0 0 0 em er g en cy fu n d p re p ar ed o r b ei n g co n fi d en t ab o u t b o rr o w $ 3 ,0 0 0 fr o m fr ie n d s o r re la ti v es , 0 o th er w is e. 8 6 .6 0 .2 0 7 38 q. bi et al. / financial services review 29 (2021) 29–54 decision-making domains in debt management, planning activities, and risk management, respectively. table 1 shows the descriptions and measures of these financial behaviors and the summary statistics. for debt management activities, our results show that more than 85% of respondents have no late payments, no bankruptcy, and a total debt payment ratio that is smaller than 36%. further, 76.4% of respondents indicate having a good credit report; 61.5% of respondents do not carry credit card balances when they have money in their bank accounts. for financial planning activities, 58.5% of respondents indicate that they are planning a few years ahead and have saved for specific goals. 75.1% of respondents report that they do a lot of shopping when making decisions on credit, savings, and investments. additionally, households get involved in risk management; 89% of respondents who have children under 18 are covered by some sort of life insurance, as compared with 81.1% of respondents reporting that everyone in the household is covered by health insurance. in addition, 86.6% of respondents have an emergency fund (or are confident that they can borrow in an urgent situation), but only 25.4% of respondents have appropriate disability insurance coverage. based on these variables, we construct three positive behavior proxies. the first proxy is the number of positive financial behaviors adopted, that is, a simple summation of these 13 behavior dummies. the second proxy is a positive behavior index created through principal component analysis. finke and huston (2013) use the principal component analysis and construct a measure of time preference through eight financial decisions that indicate individual time preference. letkiewicz, robinson, and domian (2016) apply the same approach and create a measure of financial self-efficacy and a measure of financial stress based on individuals’ financial circumstances. following their studies, we construct component factors based on the thirteen positive financial behaviors and select the first factor as our proxy for the positive financial behavior. this factor has an eigenvalue of 2.27 and is the only factor with an eigenvalue larger than 1.50. the standardized scoring coefficient to each behavior dummy variable is provided in table 1. the majority of the behavior dummy variables carry a scoring coefficient in a narrow range of 0.10 and 0.25, suggesting that most of the positive financial behaviors included in this study contribute to the measure of a single positive behavior proxy. last, we also create a dummy variable to indicate above-average positive behavior if a respondent’s positive behavior index score is greater than the median index score of the full weighted sample. to ease our interpretation, for all proxies, the higher the score, the more the adoption of financial behaviors, and the higher lifetime utility. table 2 presents summary statistics of our three positive financial behavior proxies for the full weighted sample. the range of the number of positive financial behaviors is from zero to 13. all households in the sample report engaging in at least two positive financial behaviors. the mean and median number of positive financial behaviors is nine and 10, respectively. the majority of households engage in eight to 12 positive financial behaviors but only 4.28% of them have engaged in all of them. the mean and median of the index are provided in table 2. q. bi et al. / financial services review 29 (2021) 29–54 39 3.2. independent variables on personal finance-specific human capital personal finance-related human capital illustrates how well an individual understands and potentially uses personal finance-related information to increase expected lifetime utility from consumption (huston, 2010). from this perspective, financial technology is part of the personal finance-related human capital, which indicates how well an individual can use technology to help manage household resources or finances. in this study, we use the following financial behaviors available from scf and create seven dummy variables to measure personal finance-specific human capital. 3.2.1. transaction-based financial technologies atm card: “an electronic terminal provided by financial institutions and other firms that permits consumers to withdraw cash from their bank accounts, make deposits, check balances, and transfer funds” (anguelov, hilgert, & hogarth, 2004). credit card: using credit cards allows households to borrow up to the credit limit, build good credit, reap rewards, and make payments for online merchandise easier. however, we consider borrowing too much without an appropriate repayment schedule as bad financial behavior because of the high interest and fees on the unpaid amount and the negative impact on their credit record. table 2 descriptive statistics of dependent variables: positive financial behavior proxies number of positive financial behaviors banked households % panel a: number of positive financial behaviors zero 0 one 0 two 0.03 three 0.33 four 0.91 five 2.05 six 5.62 seven 9.4 eight 13.33 nine 16.48 ten 17.35 eleven 17.36 twelve 12.88 thirteen 4.28 mean 9.44 median 10 panel b: principle component index mean �0.153 median �0.011 panel c: above-average positive behavior mean 0.500 median 0.000 40 q. bi et al. / financial services review 29 (2021) 29–54 phone banking: phone banking provides households an immediate solution to emergency issues such as reporting a stolen or lost card, applying for new credit cards, checking account balances, etc. mobile banking has also been developed very fast recently. phone banking makes it very convenient for households to solve banking related financial problems so they can better manage their money. computer banking: “banking services that consumers can access, by using an internet connection to a bank’s computer center, to perform banking tasks, receive and pay bills, and so forth (anguelov, hilgert, & hogarth, 2004). computer banking allows the households to manage their accounts wherever they are as long as they have computers and internet access. computer banking allows households to check account balances, make electronic transfers, transfer money into designated saving accounts, keep themselves updated on new banking services or receive important warnings from financial institutions, and so on. it provides a convenient channel for households to manage their finance with almost no costs. 3.2.2. planning-based financial technologies preauthorized debit: “a form of payment that allows a consumer to authorize automatic payment of regular, recurring bills from his or her account on a specific date, and usually for a specific amount” (anguelov, hilgert, & hogarth, 2004). for example, households use pre-authorized debit to set up automatic car payments, housing payments, utility bills, and so on. this method makes it easier for the households to make their payments on time. direct deposit: “a form of payment by which an organization pays funds via an electronic transfer” (anguelov, hilgert, & hogarth, 2004). direct deposit makes it easier for households to save because it transfers directly into designated accounts before the consumer has a chance to spend the money elsewhere. computer software: households who use computer software to manage their finance are more likely to make the right decisions for savings and investments. computer software can help households calculate how much they need to save for each goal, such as retirement savings, college fund savings, and so on. it provides more accurate spending and saving information, which can assist households with better data to make proper financial decisions. using computer software can increase the probability of households reaching financial goals by facilitating consumers to save amounts that are more appropriate and aiding consumers in considering multiple goals simultaneously. the descriptive statistics of financial technology variables are presented in panel a of table 3. direct deposit is the most commonly used technology, as 81.9% of banked households report that they use it. with the new technologies coming out, phone banking is not as popular as before. our results suggest only 21.2% of banked households report using phone banking while 67.5% report using computer banking. moreover, 21.2% of banked households report using computer software to help managing their finance. similar to how we construct the positive financial behavior index, we also adopt principal component analysis to form two composite technology usage measures: transactional tech usage index and planning tech usage index. the transactional tech usage index is a measure of transaction-based technology usage based on atm card use, credit card use, phone banking use, and computer banking use. the planning tech usage index measures planning-based technology usage based on the last three technology-use behaviors. both composite measures use the first component factor from principal component analysis as the proxy. the standardized scoring coefficients are provided in table 3. q. bi et al. / financial services review 29 (2021) 29–54 41 t ab le 3 d es cr ip ti v e st at is ti cs o f fi n an ci al te ch n o lo g y v ar ia b le s an d o th er co n tr o l v ar ia b le s p an el a : k ey in d ep en d en t v ar ia b le s o n sp ec ifi c h o u se h o ld ca p it al v ar ia b le s m ea su re m en t b an k ed h o u se h o ld s (% ) s ta n d ar d iz ed sc o ri n g co ef fi ci en ts t ra n sa ct io n te ch n o lo g ie s a t m ca rd 1 if u se a t m ca rd as o n e o f th e m ai n w ay s y o u d o b u si n es s w it h b an k o r if y o u h av e a ca rd th at al lo w s y o u to d ep o si t o r w it h d ra w m o n ey fr o m y o u r b an k u si n g an a t m , 0 o th er w is e 8 1 .9 0 .4 3 c re d it ca rd 1 if h av e an y cr ed it ca rd o r ch ar g e ca rd , 0 o th er w is e 7 2 .4 0 .3 8 9 p h o n e b an k in g 1 if u se au to m at ed p h o n e sy st em as o n e o f th e m ai n w ay s to d o b u si n es s w it h b an k , 0 o th er w is e 2 1 .2 0 .2 2 2 c o m p u te r b an k in g 1 if u se co m p u te r as o n e o f th e m ai n w ay s to d o b u si n es s w it h b an k , 0 o th er w is e 6 7 .5 0 .5 7 1 p la n n in g te ch n o lo g ie s p re au th o ri ze d d eb it 1 if h av e u ti li ty b il ls , m o rt g ag e o r re n t p ay m en ts , o r o th er p ay m en ts au to m at ic al ly p ai d d ir ec tl y fr o m b an k ac co u n ts w it h o u t h av in g to w ri te a ch ec k , 0 o th er w is e 5 7 0 .6 0 8 d ir ec t d ep o si t 1 if h av e p ay ch ec k s o r s o ci al s ec u ri ty b en efi ts o r o th er m o n ey au to m at ic al ly p ai d d ir ec tl y in to ac co u n ts , 0 o th er w is e 8 6 .4 0 .4 2 6 c o m p u te r so ft w ar e 1 if u se co m p u te r so ft w ar e to m an ag e m o n ey , 0 o th er w is e 2 1 0 .5 2 p an el b : c o m p o si te te ch n o lo g y u sa g e m ea su re s fr o m p ri n ci p al co m p o n en t an al y si s m ea n m ed ia n t ra n sa ct io n al te ch u sa g e in d ex �0 .0 9 5 0 .2 8 6 p la n n in g te ch u sa g e in d ex �0 .0 5 8 0 .3 8 5 p an el c : o th er co n tr o l v ar ia b le s o n g en er al h o u se h o ld ca p it al v ar ia b le s m ea su re m en t b an k ed h o u se h o ld s (% ) f in an ci al so p h is ti ca ti o n 0 – 2 0 p er ce n ti le 1 if h o u se h o ld fi n an ci al so p h is ti ca ti o n le v el is in th e 0 % to 2 0 % ra n g e, 0 o th er w is e 2 0 2 1 – 4 0 p er ce n ti le 1 if h o u se h o ld fi n an ci al so p h is ti ca ti o n le v el is in th e 2 0 % to 4 0 % ra n g e, 0 o th er w is e 2 0 4 1 – 6 0 p er ce n ti le 1 if h o u se h o ld fi n an ci al so p h is ti ca ti o n le v el is in th e 4 0 % to 6 0 % ra n g e, 0 o th er w is e 2 0 6 1 – 8 0 p er ce n ti le 1 if h o u se h o ld fi n an ci al so p h is ti ca ti o n le v el is in th e 6 0 % to 8 0 % ra n g e, 0 o th er w is e 2 0 (c o n ti n u ed o n n ex t p a g e) 42 q. bi et al. / financial services review 29 (2021) 29–54 t ab le 3 (c o n ti n u ed ) p an el c : o th er co n tr o l v ar ia b le s o n g en er al h o u se h o ld ca p it al v ar ia b le s m ea su re m en t b an k ed h o u se h o ld s (% ) 8 1 – 1 0 0 p er ce n ti le 1 if h o u se h o ld fi n an ci al so p h is ti ca ti o n le v el is in th e 8 0 % to 1 0 0 % ra n g e, 0 o th er w is e 2 0 h o m eo w n er sh ip 1 if h o u se h o ld o w n s o r p ar ti al ly o w n s a h o m e, 0 o th er w is e 6 8 .1 a g e 1 8 – 3 4 1 if th e h ea d o f h o u se h o ld is ag e 1 8 -3 4 , 0 o th er w is e 2 1 .3 3 5 – 4 9 1 if th e h ea d o f h o u se h o ld is ag e 3 5 -4 9 , 0 o th er w is e 2 6 .4 5 0 – 6 4 1 if th e h ea d o f h o u se h o ld is ag e 5 0 -6 4 , 0 o th er w is e 2 9 6 5 an d o v er 1 if th e h ea d o f h o u se h o ld is ag e 6 5 an d o v er , 0 o th er w is e 2 3 .3 e d u ca ti o n le v el l es s th an h ig h sc h o o l 1 if h o u se h o ld h ig h es t le v el o f sc h o o l co m p le te d is < 1 2 , 0 o th er w is e 6 h ig h sc h o o l/ g e d 1 if h o u se h o ld h ig h es t le v el o f sc h o o l co m p le te d is = 1 2 o r h ig h sc h o o l d ip lo m a/ g e d , 0 o th er w is e 2 4 .8 s o m e co ll eg e 1 if h o u se h o ld h ig h es t le v el o f sc h o o l co m p le te d is > 1 2 b u t n ev er g et a co ll eg e d eg re e, 0 o th er w is e 2 6 .8 b ac h el o rs o r h ig h er 1 if h o u se h o ld h ig h es t le v el o f sc h o o l co m p le te d is b ac h el o rs o r h ig h er , 0 o th er w is e 4 2 .4 m ar ri ed 1 if h ea d o f h o u se h o ld is cu rr en tl y m ar ri ed o r li v in g w it h a p ar tn er , 0 o th er w is e 5 9 f in an ci al re so u rc es in co m e 0 – 2 0 p er ce n ti le 1 if h o u se h o ld in co m e is b et w ee n $ 1 an d $ 2 5 ,0 0 0 , 0 o th er w is e (r ef er en ce ) 2 3 .9 2 1 – 4 0 p er ce n ti le 1 if h o u se h o ld in co m e is b et w ee n $ 2 5 ,0 0 1 an d $ 4 8 ,0 0 0 , 0 o th er w is e 2 5 .5 4 1 – 6 0 p er ce n ti le 1 if h o u se h o ld in co m e is b et w ee n $ 4 8 ,0 0 1 an d $ 8 5 ,0 0 0 , 0 o th er w is e 2 3 .2 6 1 – 8 0 p er ce n ti le 1 if h o u se h o ld in co m e is b et w ee n $ 8 5 ,0 0 1 an d $ 1 9 2 ,0 0 0 , 0 o th er w is e 2 0 .8 8 1 – 1 0 0 p er ce n ti le 1 if h o u se h o ld in co m e is b et w ee n $ 1 9 2 ,0 0 1 o r m o re , 0 o th er w is e 6 .7 n et w o rt h 0 – 2 0 p er ce n ti le 1 if h o u se h o ld n et w o rt h is le ss th an $ 0 to $ 1 0 ,6 8 2 , 0 o th er w is e (r ef er en ce ) 2 3 2 1 – 4 0 p er ce n ti le 1 if h o u se h o ld n et w o rt h is b et w ee n $ 1 0 ,6 8 3 to $ 9 0 ,1 1 9 , 0 o th er w is e 2 5 4 1 – 6 0 p er ce n ti le 1 if h o u se h o ld n et w o rt h is b et w ee n $ 9 0 ,1 2 0 to $ 3 5 2 ,7 4 1 , 0 o th er w is e 2 6 .8 6 1 – 8 0 p er ce n ti le 1 if h o u se h o ld n et w o rt h is b et w ee n $ 3 5 2 ,7 4 2 to $ 1 ,8 2 5 ,4 6 8 , 0 o th er w is e 1 9 .8 8 1 – 1 0 0 p er ce n ti le 1 if h o u se h o ld n et w o rt h is b et w ee n $ 1 ,8 2 5 ,4 6 9 o r m o re , 0 o th er w is e 5 .5 h o u se h o ld si ze o n e p er so n 1 if n u m b er o f p eo p le in h o u se h o ld = 1 , 0 o th er w is e 2 5 .3 t w o p er so n 1 if n u m b er o f p eo p le in h o u se h o ld = 2 , 0 o th er w is e 3 4 .2 t h re e o r m o re 1 if n u m b er o f p eo p le in h o u se h o ld > = 3 , 0 o th er w is e 4 0 .5 p re se n ce o f ch il d re n u n d er ag e 1 8 1 if ch il d re n u n d er ag e 1 8 ar e p re se n t in th e h o u se h o ld , 0 o th er w is e 3 4 .4 o th er in fl u en ce s f em al e 1 if h ea d o f h o u se h o ld is fe m al e, 0 is m al e 5 1 .9 (c o n ti n u ed o n n ex t p a g e) q. bi et al. / financial services review 29 (2021) 29–54 43 t ab le 3 (c o n ti n u ed ) p an el c : o th er co n tr o l v ar ia b le s o n g en er al h o u se h o ld ca p it al v ar ia b le s m ea su re m en t b an k ed h o u se h o ld s (% ) r ac e an d et h n ic it y w h it e an d “o th er ” 1 if h o u se h o ld d es cr ib es it se lf as w h it e, a si an , p ac ifi c is la n d er , o r n at iv e a m er ic an , 0 o th er w is e 7 7 .7 b la ck 1 if h o u se h o ld d es cr ib es it se lf as b la ck , 0 o th er w is e 1 2 .7 h is p an ic 1 if h o u se h o ld d es cr ib es it se lf as h is p an ic , 0 o th er w is e 9 .6 e co n o m ic ex p ec ta ti o n s 1 if ex p ec t th e u .s . ec o n o m y to p er fo rm b et te r o v er th e n ex t 5 y ea rs , 0 o th er w is e 4 5 .4 in te re st ra te ex p ec ta ti o n s 1 if ex p ec t in te re st ra te s w il l b e h ig h er 5 y ea rs fr o m n o w , 0 o th er w is e 7 7 r is k to le ra n ce n o ri sk 1 if n o t w il li n g to ta k e an y fi n an ci al ri sk s, 0 o th er w is e 4 4 .4 m o d er at e ri sk 1 if w il li n g to ta k e av er ag e o r ab o v e av er ag e fi n an ci al ri sk s ex p ec ti n g to ea rn av er ag e o r ab o v e av er ag e re tu rn , 0 o th er w is e 5 2 .7 s u b st an ti al ri sk 1 if w il li n g to ta k e su b st an ti al fi n an ci al ri sk s ex p ec ti n g to ea rn av er ag e o r ab o v e av er ag e re tu rn , 0 o th er w is e 2 .9 44 q. bi et al. / financial services review 29 (2021) 29–54 based on shefrin and thaler’s (1981) “doer versus planner” model, consumers experience a costly intrapersonal conflict between a “planner” and a “doer” (gul & pesendorfer, 2001) where the planner is concerned with lifetime utility but the doer exists for only one time period and would consume most of their resources today. the authors suggest that to shift intertemporal choice (benabou & pycia, 2002) and prevent the doer from consuming total lifetime income in the first period, some psychic technology capable of affecting the doer’s behavior is required.4 based on the theoretical model of shefrin and thaler (1981), planning purposed financial technologies have the potential to fulfill all three ways the authors identify for shifting the myopic doer to more of a “planner” mindset: (1) modify the doers’ preferences, (2) force doer to input to a savings program or budget, “simply keeping track seems to act as a tax on any behavior the planner views as deviant,” and (3) alter incentives. we expect that using transaction-based financial technologies may not have a positive impact on households’ engagement in positive financial behaviors because transaction-based technologies basically enhance consumer discretion and ability to myopically overextend themselves by consuming too much today. in contrast, we expect that the use of planning purposed financial technologies will have a positive impact on positive financial behaviors. in panel b of table 3, it shows that the mean (median) of the transitional technology usage index and planning technology usage index is�0.095 (0.286) and�0.058 (0.385), respectively. 3.3. other control variables to control for other factors that are likely to be related to households’ positive financial behaviors, we include a range of other measures as shown in panel b of table 3. financial sophistication: a score from four questions in scf that represents the financial literacy of a household (huston et al., 2012). we expect that households with higher financial sophistication levels are more likely to engage in positive financial behaviors. homeownership: households with homeownership have more personal finance experiences, such as payments of mortgage loans, and refinance options. we expect that such households are more likely to engage in positive financial behaviors. 3.3.1. general human capital age: as age goes up, years of experience become valuable human capital. education level: generally speaking, more years of schooling will increase households’ general productivities. we expect that older households and those with more education are more likely to engage in positive financial behaviors. marital status: from the whole household level, married households have more general human capital because one partner can access the other’s resources. we expect married households are more likely to engage in positive financial behaviors. 3.3.2. financial resources income: the more income a household has, the more resources that could be managed to achieve higher lifetime utility. q. bi et al. / financial services review 29 (2021) 29–54 45 net worth: if the households have higher net worth, they are more likely to be in good shape with their finances but it does not necessarily mean that they are better financial managers compared with the others. household composition: having children under 18 influences household expenditure. we expect households without children under 18 are more likely to engage in positive financial behaviors. 3.3.3. cultural/environmental influences gender: gender differences have been explored, identified, and established for different financial behaviors from savings (strömbäck et al., 2017; fisher, 2010) to willingness to take risks (fisher & yao, 2017). for example, hayhoe et al. (2000) found that gender was more influential in predicting financial management practices than was affective credit attitude, with female students employing a greater number of financial practices. gender differences have also been demonstrated for objective financial knowledge and numeric ability, where males typically perform better than females (chen & volpe, 2002; fonseca et al., 2012; lusardi & mitchell, 2008; powell & ansic, 1997). lind et al. (2020) attempted to explore how gender impacted broader measures of financial behavior while controlling for differences in relevant cognitive abilities and demographic statistics—their research discovered that women reported a lower level of subjective financial wellbeing even though they reported a more prudent financial behavior than men when controlling for socio-demographics and cognitive abilities. race and ethnicity: cultural biases and behaviors affect households’ financial behaviors. for example, asian households are more likely to save and more encouraged to attain higher education. economic expectations: different economic expectations affect households’ financial decisions such as saving and consumption. interest rate expectations: different interest rate expectations affect households’ financial decisions such as saving and consumption, and chosen loan products. risk tolerance: willingness to take risk affects households’ investment related financial decisions. we expect households who are willing to take on some risk are more likely to engage in positive financial behaviors. 4. multivariate analysis, results, and discussions to determine the impact of using financial technology on positive financial behavior, we use ordinary least square (ols) by regressing the independent variables on the positive financial behavior indexes as specified in equation (2). results of the regression analysis are shown in table 4. models (1), (2), and (3) show the regression results using three different measures of positive financial behaviors as mentioned previously. our results suggest that transactional technology usage is significantly and negatively related to positive financial behaviors. for example, in model (2), it indicates that a one unit increase in the transactional technology index will cause the composite positive financial behavior index to decrease significantly by �0.051 units. this result is consistent across all three models regardless of how we measure positive financial behaviors. the results indicate that transaction-based financial technologies like atm card, credit card, phone banking, and computer banking negatively affect households’ engagement in positive financial behaviors. for example, the convenience of using atm cards to withdraw money at any location might increase the probability of households consuming now instead of saving for the future. overuse of credit cards could lead to high-interest expenses or over-purchasing behaviors 46 q. bi et al. / financial services review 29 (2021) 29–54 t ab le 4 t h e im p ac t o f te ch n o lo g y ad o p ti o n o n p o si ti v e fi n an ci al b eh av io rs , es ti m at ed th ro u g h m u lt ip le re g re ss io n sp ec ifi ca ti o n s v ar ia b le s n u m b er o f p o si ti v e fi n an ci al b eh av io rs (1 ) p ri n ci p le c o m p o n en t in d ex (2 ) a b o v eav er ag e p o si ti v e b eh av io r (3 ) c o n st an t 8 .1 3 9 * * * �0 .8 0 4 * * * �1 .7 1 2 * * * s p ec ifi c h o u se h o ld ca p it al f in an ci al te ch n o lo g ie s t ra n sa ct io n al te ch u sa g e in d ex �0 .1 6 9 * * * �0 .0 5 1 * * * �0 .1 0 2 * * * p la n n in g te ch u sa g e in d ex 0 .0 7 9 * * * 0 .0 3 5 * * * 0 .0 8 2 * * * f in an ci al so p h is ti ca ti o n (r el at iv e to 4 1 – 6 0 % ) 1 – 2 0 p er ce n ti le �0 .6 5 * * * �0 .3 2 3 * * * �0 .7 0 4 * * * 2 1 – 4 0 p er ce n ti le �0 .2 3 4 * * * �0 .1 2 2 * * * �0 .2 2 7 * * * 6 1 – 8 0 p er ce n ti le 0 .4 8 * * * 0 .2 4 3 * * * 0 .7 7 6 * * * 8 1 – 1 0 0 p er ce n ti le 0 .7 8 * * * 0 .3 6 9 * * * 1 .0 8 1 * * * h o m eo w n er sh ip �0 .3 2 * * * �0 .1 0 5 * * * �0 .1 2 7 * * * g en er al h o u se h o ld ca p it al a g e (r el at iv e to 3 5 – 4 9 y ea rs o ld ) 1 8 – 3 4 0 .5 1 * * * 0 .2 3 2 * * * 0 .6 0 8 * * * 5 0 – 6 4 �0 .1 8 6 * * * �0 .0 6 8 * * * �0 .0 4 4 6 5 an d o v er 0 .0 1 7 0 .1 0 7 * * * 0 .2 6 9 * * * e d u ca ti o n le v el (r el at iv e to h ig h sc h o o l/ g e d ) l es s th an h ig h sc h o o l �0 .0 3 4 �0 .0 2 2 �0 .3 0 7 * * * s o m e co ll eg e 0 .0 4 8 * 0 .0 3 1 * * 0 .0 2 4 b ac h el o rs o r h ig h er 0 .2 8 6 * * * 0 .1 5 6 * * * 0 .3 2 2 * * * m ar ri ed 0 .0 5 7 * 0 .0 5 2 * * * 0 .1 5 6 * * * f in an ci al re so u rc es in co m e (r el at iv e to 0 to 2 0 p er ce n ti le ) 2 1 – 4 0 p er ce n ti le 0 .1 0 4 * * * 0 .0 5 5 * * * 0 .1 4 1 * * * 4 1 – 6 0 p er ce n ti le 0 .3 6 8 * * * 0 .1 9 2 * * * 0 .5 2 3 * * * 6 1 – 8 0 p er ce n ti le 0 .8 3 8 * * * 0 .4 1 8 * * * 1 .0 1 5 * * * 8 1 – 1 0 0 p er ce n ti le 1 .2 8 8 * * * 0 .6 0 5 * * * 1 .7 9 7 * * * n et w o rt h (r el at iv e to 0 to 2 0 p er ce n ti le ) 2 1 – 4 0 p er ce n ti le 0 .7 5 1 * * * 0 .3 8 4 * * * 0 .8 1 7 * * * 4 1 – 6 0 p er ce n ti le 1 .4 2 7 * * * 0 .7 0 6 * * * 1 .4 3 3 * * * 6 1 – 8 0 p er ce n ti le 1 .8 4 3 * * * 0 .8 8 9 * * * 1 .9 2 9 * * * 8 1 – 1 0 0 p er ce n ti le 1 .7 5 7 * * * 0 .8 0 2 * * * 2 .0 2 3 * * * (c o n ti n u ed o n n ex t p a g e) q. bi et al. / financial services review 29 (2021) 29–54 47 t ab le 4 (c o n ti n u ed ) v ar ia b le s n u m b er o f p o si ti v e fi n an ci al b eh av io rs (1 ) p ri n ci p le c o m p o n en t in d ex (2 ) a b o v eav er ag e p o si ti v e b eh av io r (3 ) h o u se h o ld si ze (r el at iv e to tw o p er so n ) o n e p er so n 0 .4 8 4 * * * 0 .2 8 4 * * * 0 .6 1 6 * * * t h re e o r m o re �0 .1 3 5 * * * �0 .0 7 3 * * * �0 .1 9 4 * * * p re se n ce o f ch il d re n u n d er ag e 1 8 �0 .3 7 5 * * * �0 .2 3 4 * * * �0 .4 2 8 * * * o th er in fl u en ce s f em al e �0 .1 2 3 * * * �0 .0 6 3 * * * 0 .0 0 8 r ac e an d et h n ic it y (r el at iv e to w h it e an d “o th er ”) b la ck 0 .0 3 6 �0 .0 0 6 �0 .1 2 1 * * h is p an ic �0 .2 0 3 * * * �0 .1 0 5 * * * �0 .6 3 4 * * * e co n o m ic ex p ec ta ti o n s 0 .0 5 7 * * * 0 .0 1 3 �0 .0 0 7 in te re st ra te ex p ec ta ti o n s 0 .0 6 2 * * * 0 .0 0 6 0 .0 7 8 * * r is k to le ra n ce (r el at iv e to n o ri sk ) m o d er at e ri sk �0 .0 1 7 �0 .0 5 8 * * * �0 .2 4 6 * * * s u b st an ti al ri sk �0 .3 2 4 * * * �0 .2 1 3 * * * �0 .6 0 1 * * * c o ef fi ci en ts fr o m o rd in ar y l ea st s q u ar e re g re ss io n s ar e re p o rt ed in th e fi rs t tw o co lu m n s, co ef fi ci en ts fr o m lo g is ti c re g re ss io n ar e re p o rt ed in th e la st co lu m n . * * * ,* * ,* in d ic at e si g n ifi ca n ce at th e 0 .0 1 , 0 .0 5 , an d 0 .1 0 le v el s, re sp ec ti v el y . 48 q. bi et al. / financial services review 29 (2021) 29–54 that households cannot afford to repay. using phone banking and computer banking seems to have negative effects on positive financial behaviors. a recent study by transunion showed that both the volume and balance of personal unsecured loans have been increasing significantly in the past few years as consumers become more likely to choose fintech than traditional lenders for borrowing.5 overall, our results are consistent with these findings by showing that, though the convenience provided by transaction-based e-banking technology saves time and costs for households to complete financial transactions, it also increases the ease of accessing funds and can be detrimental to household financial well-being in the long run. in contrast, the regression results in models (1), (2), and (3) also indicate that planningbased financial technologies, including preauthorized debit, direct deposit, and computer software use, positively affect households’ engagement in financial behaviors. our results show that the coefficients of the planning technology usage index are significantly positive across all models. for example, the positive coefficient of 0.082 in model (3) suggests that planning technology usage is more likely to be positively related to households’ positive financial behaviors. the results hence support our hypothesis that planning-based financial technologies will have a positive impact on households’ financial behaviors.6 planningbased financial technologies appear to create a more positive environment for enhancing household financial well-being. for example, preauthorized debit is used to set up future automatic payments on loans and bills while direct deposit makes it easier to save and budget. computer software helps households plan for the future by providing financial knowledge, calculation help, and action plans. our results hence are consistent with previous studies showing that planning behaviors have a significant impact on personal savings practices (lusardi, 2010). also, our findings are consistent with shefrin and thaler’s (1981) theory that individuals will be more likely to save or become a “planner” when preferences or incentives are altered, behaviors are tracked, or the doer’s set of choices is limited with constraints. households with clear goals are more likely to save for the future, which will lead to higher levels of financial well-being and life satisfaction. consistent with other research findings (smith, finke, & huston, 2011, 2012a, 2012b), we also find that financial sophistication significantly affects households’ financial behaviors. specifically, low financial sophistication in households is negatively related to positive financial behaviors, while high financial sophistication is significantly and positively associated with positive financial behaviors. households with high education, high income, and high net worth are also related to high positive financial behaviors. households with homeownership or less risk-averse (i.e., who are willing to take a substantial risk) are engaging in low positive financial behaviors. consistent with our expectation, households with children under 18 are associated with low positive financial behaviors. for the impact of age, households with young individuals between ages 18 and 34 show higher positive financial behaviors than those with older individuals with ages between 50 and 64. 5. conclusions and implications given the rapid growth of technology innovation in the finance sector, it is natural to ask whether these technologies help households engage in positive financial behaviors. using q. bi et al. / financial services review 29 (2021) 29–54 49 the 2013 survey of consumer finances commissioned by the federal reserve board, this study found that not all kinds of financial technologies are helpful for households’ engagement in positive financial behaviors. in this study, we use a life cycle and human capital theoretical framework to illustrate the impact of financial technology on household financial behavior. consistent with the theoretical framework, we find that financial technology-specific household capital has a significant impact on positive financial behaviors but not all types of financial technology will enhance positive financial behaviors. specifically, we found transaction-based financial technologies like atm card use, credit card use, phone banking, and computer banking have a negative impact on the number of positive financial behaviors reported. providing easy access to bank accounts may encourage people to overspend in current periods, especially for those with self-control issues. in contrast, planning-based financial technologies like direct deposit and computer software use have positive impact on the number of positive financial behaviors. financial sophistication, general household capital such as age and education, financial resources, and other resources such as expectation on the economy and risk tolerance are also found to have a significant impact on positive financial behavior. given the findings of this study, we suggest financial planners and financial educators encourage clients and individuals to use planning-based financial technologies such as computer software. financial planners must focus on improving clients’ personal finance management skills by emphasizing the importance to think from a long-term perspective when making financial decisions. although the effectiveness of financial education programs is mixed (willis, 2008), it is important to educate households on the effective use of tools to change their financial behaviors, rather than simply delivering financial education. to take full advantage of financial technology, consumers and professionals that assist consumers need to be aware of what types of technology will help in achieving higher financial satisfaction over the long run. our findings in this study suggest that simply providing financial technology to complete transactions does not appear to improve household financial well-being. only planning-based financial technologies have a positive impact on household financial well-being and hence should be given more attention in terms of technology development and marketing perspectives. for example, hyperbolic consumers are those who know that they should save for the future but it is hard for them to give up current consumption (angeletos et al., 2001). guiding these myopic consumers through planning-based financial technology may be an effective way to enhance their financial behaviors because it helps them create commitment devices to realize the benefit from engaging long-term financial practices that are consistent with maximizing lifetime utility. we realize that there are limitations in our study. because of data constraints, the positive financial behavior indexes used in this study cover only a few, not all, positive financial behaviors. also, we do not look at the various types of computer software that are used by households to help manage their finance use in this study. future research could explore households that switch financial technology and examine the impact of this change on their positive financial behaviors engagement. 50 q. bi et al. / financial services review 29 (2021) 29–54 notes 1 see https://newsroom.transunion.com/consumers-poised-to-continue-strong-creditactivity-this-holiday-season 2 the data is available on federal reserve at https://www.federalreserve.gov/econres/ scfindex.htm 3 the 2013 scf collect data using computer-assisted personal interviewing (capi). thus, there is no questionnaire in the usual sense. scf uses a dual-frame sample design consisting of a standard, geographically based random sample and an oversample of affluent households. missing values are imputed by making multiple estimates of the missing data and creating five implicate data sets. we use all five implicates to avoid inaccurate results on the significant test (rubin, 1987). 4 two main techniques are available for this: (1) the doer can be given discretion in which case either his preferences must be modified or his incentives must be altered, or (2) the doer’s set of choices may instead be limited by imposing rules that change the constraints the doer faces” (shefrin & thaler, 1981). 5 available at https://www.transunion.com/blog/consumer-credit-origination-balanceand-deliquency-trends 6 in untabulated results, we also regress three dependent variables on individual financial technology variables instead of the composite indexes. the results are consistent with these reported in the main texts. references aizcorbe, a. m., kennickell, a. b., & moore, k. b. (2003). recent changes in u.s. family finances: evidence from the 1998 and 2001 survey of consumer finances. federal reserve bulletin, 89, 1-32. amar, m., ariely, d., ayal, s., cryder, c. e., & rick, s. i. (2011). winning the battle but losing the war: the psychology of debt management. journal of marketing research, 48, s38-s50. ameriks, j., caplin, a., & leahy, j. (2002). wealth accumulation and the propensity to plan. national bureau of economic research, no. w8920, angeletos, g. m., laibson, d., repetto, a., tobacman, j., & weinberg, s. (2001). the hyperbolic consumption model: calibration, simulation, and empirical evaluation. journal of economic perspectives, 15, 47-68. anguelov, c. e., hilgert, m. a., & hogarth, j. m. (2004). us consumers and electronic banking, 1995-2003. federal reserve bulletin, 90, 1–18. bayus, b. (1987). forecasting sales of new contingent products: an application to the compact disc market. journal of product innovation management, 4, 243-255. becker, g. s., & tomes, n. (1994). human capital and the rise and fall of families. in human capital: a theoretical and empirical analysis with special reference to education (3rd ed., pp. 257–298). chicago, il: the university of chicago press. bénabou, r., & pycia, m. (2002). dynamic inconsistency and self-control: a planner–doer interpretation. economics letters, 77, 419-424. bernheim, b. d., carman, k. g., gokhale, j., & kotlikoff, l. j. (2001). the mismatch between life insurance holdings and financial vulnerabilites: evidence from the survey of consumer finances. nber working paper, no. 8544, october. bernheim, b. d., carman, k. g., gokhale, j., & kotlikoff, l. j. (2003). are life insurance holdings related to financial vulnerabilities? economic inquiry, 41, 531-554. q. bi et al. / financial services review 29 (2021) 29–54 51 bricker, j., dettling, l. j., henriques, a., hsu, j. w., jacobs, l., moore, k. b., & windle, r. a. (2017). changes in us family finances from 2013 to 2016: evidence from the survey of consumer finances. federal reserve bulletin, 103, 1-42. butryn, m. l., webb, v., & wadden, t. a. (2011). behavioral treatment of obesity. the psychiatric clinics of north america, 34, 841-859. chang, y. r., & huston, s. j. (1995). patterns of adequate household emergency fund holdings: a comparison of households in 1983 and 1986. financial counseling and planning, 6, 119-128. chen, s., & van der klaauw, w. (2008). the work disincentive effects of the disability insurance program in the 1990s. journal of econometrics, 142, 757-784. chen, h., & volpe, r. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289-307. davis, f. (1989). perceived usefulness, perceived ease of use and user acceptance of information technology. mis quarterly, 13, 319-340. dean, l. r., joo, s. h., gudmunson, c. g., fischer, j. l., & lambert, n. (2013). debt begets debt: examining negative credit card behaviors and other forms of consumer debt. journal of financial service professionals, 67, 72-84. dew, j., & xiao, j. j. (2011). the financial management behavior scale: development and validation. journal of financial counseling and planning, 22, 43-59. experian. (2009). with age comes wisdom . . . and a higher credit score. experian. (available at http://www. experian.com/assets/business-resources/studies/snapshot-seniors-debt.pdf) federal reserve. (2009). consumer credit november 2008. federal reserve statistical release. washington, dc: board of governors of the federal reserve system. federal reserve. (2013). the 2013 federal reserve payments study: recent and long-term payment trends in the united states: 2003–2012. washington, dc: board of governors of the federal reserve system. feenber, d., & skinner, j. (1994). the risk and duration of catastrophic health care expenditures. the review of economics and statistics, 76, 633-647. finke, m., & huston, s. (2013). time preference and the important of saving for retirement. journal of economic behavior & organization, 89, 23-34. fisher, p. j. (2010). gender differences in personal saving behaviors. journal of financial counseling and planning, 21, 14-24. fisher, p. j., & yao, r. (2017). gender differences in financial risk tolerance. journal of economic psychology, 61, 191-202. fonseca, r., mullen, k. j., zamarro, g., & zissimopoulos, j. (2012). what explains the gender gap in financial literacy? the role of household decision making. the journal of consumer affairs, 46, 90-106. garman, e. t., leech, i. e., & grable, j. e. (1996). the negative impact of employee poor personal financial behaviors on employers. financial counseling and planning, 7, 157-168. gross, d. b., & souleles, n. s. (2001). do liquidity constraints and interest rates matter for consumer behavior? evidence from credit card data (no. w8314). cambridge, ma: national bureau of economic research. gul, f., & pesendorfer, w. (2001). dynamic inconsistency and self-control. econometrica, 69, 1403-1436. hamlet, c., & strube, m. (2000). community banks go online. aba banking journal, 92, 61-65. harkin, b., webb, t. l., chang, b. p., prestwich, a., conner, m., kellar, i., benn, y., & sheeran, p. (2016). does monitoring goal progress promote goal attainment? a meta-analysis of the experimental evidence. psychological bulletin, 142, 198-229. hayhoe, c. r., leach, l. j., turner, p. r., bruin, m. j., & lawrence, f. c. (2000). differences in spending habits and credit use of college students. journal of consumer affairs, 34, 113-133. hogarth, j. m., & anguelov, c. e. (2004). are families who use e-banking better financial managers? financial counseling and planning, 15, 61-78. huston, s. j., & chang, y. r. (1997). adequate emergency fund holdings and household type. financial counseling and planning, 8, 37-46. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296-316. 52 q. bi et al. / financial services review 29 (2021) 29–54 huston, s. j., finke, m. s., & smith, h. (2012). a financial sophistication proxy for the survey of consumer finances. applied economics letters, 19, 1275-1278. jagtiani, j., & lemieux, c. (2018). do fintech lenders penetrate areas that are underserved by traditional banks? journal of economics and business, 100, 43-54. jeske, k., & kitao, s. (2009). us tax policy and health insurance demand: can a regressive policy improve welfare? journal of monetary economics, 56, 210-221. jun, m., & cai, s. (2001). the key determinants of internet banking service quality: a content analysis. international journal of bank marketing, 19, 276-291. kim, b. m., widdows, r., & yilmazer, t. (2005). the determinants of consumers’ adoption of internet banking. in proceedings of the consumer behavior and payment choice 2005 conference. (pp. 1-34). boston, ma. kolodinsky, j. m., hogarth, j. m., & hilgert, m. a. (2004). the adoption of electronic banking technologies by u.s. consumers. international journal of bank marketing, 22, 238-259. lee, j., & hogarth, j. m. (1999). returns to information search: consumer credit card shopping decisions. journal of financial counseling and planning, 10, 23-35. lee, e., & lee, j. (2000). haven’t adopted electronic financial services yet? the acceptance and diffusion of electronic banking technologies. financial counseling and planning, 11, 49-60. lee, e., & lee, j. (2001). consumer adoption of internet banking: need-based and/or skill based? marketing management journal, 11, 101-113. lee, e., lee, j., & eastwood, d. (2003). a two-step estimation of consumer adoption of technology-based service innovations. journal of consumer affairs, 37, 256-282. letkiewicz, j., robinson, c., & domian, d. (2016). behavioral and wealth considerations for seeking professional financial planning help. financial services review, 25, 105-126. lewis, f. (1989). dependents and the demand for life insurance. the american economic review, 79, 452-467. lin, y., & grace, m. f. (2007). household life cycle protection: life insurance holdings, financial vulnerability, and portfolio implications. journal of risk & insurance, 74, 141-173. livshits, i., tertilt, m., & macgee, j. (2007). consumer bankruptcy: a fresh start. american economic review, 97, 402-418. lind, t., ahmed, a., skagerlund, k., strömbäck, c., västfjäll, d., & tinghög, g. (2020). competence, confidence, and gender: the role of objective and subjective financial knowledge in household finance. journal of family and economic issues. lusardi, a. (2010). information, expectations, and savings for retirement. in h. j. aaron (ed.), behavioral dimensions of retirement economics (pp. 81-115). washington, dc: brookings institution press and russell sage foundation. lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review, 98, 413-417. merry, e. a. (2018). mobile banking: a closer look at survey measures. feds notes. washington: board of governors of the federal reserve system, march 27, 2018. modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. franco modigliani, 1, 388-436. neukam, k. a., & hershey, d. a. (2003). financial inhibition, financial activation, and saving for retirement. financial services review, 12, 19-37. parrish, l., & servon, l. (2006). policy options to improve financial education: equipping families for their financial futures. washington, dc: asset building program, new america foundation. pew research center. (2018). mobile fact sheet. (available at http://www.pewinternet.org/fact-sheet) powell, m., & ansic, d. (1997). gender differences in risk behavior in financial decision making: an experimental analysis. journal of economic psychology, 18, 605-628. rubin, d. (1987).multiple imputation for nonresponse in surveys. new york, ny: john wiley & sons. rutherford, l., & devaney, s. a. (2009). utilizing the theory of planned behavior to understand convenience use of credit cards. journal of financial counseling and planning, 20, 48-63. servon, l. j., & kaestner, r. (2008). consumer financial literacy and the impact of online banking on the financial behavior of lower-income bank customers. journal of consumer affairs, 42, 271-305. q. bi et al. / financial services review 29 (2021) 29–54 53 shefrin, h. m., & thaler, r. h. (1981). an economic theory of self-control. journal of political economy, 89, 392-406. shim, s., xiao, j. j., barber, b., & lyons, a. (2009). pathways to success: a model of financial well-being of young adults. journal of applied developmental psychology, 30, 708-723. smith, h. l., finke, m. s., & huston, s. j. (2011). the impact of financial sophistication on adjustable rate mortgage ownership. journal of financial counseling and planning, 22, 3-15. smith, h. l., finke, m. s., & huston, s. j. (2012a). financial sophistication and housing leverage among older households. journal of family and economic issues, 33, 315-327. smith, h. l., finke, m. s., & huston, s. j. (2012b). the influence of financial sophistication and financial planners on roth ira ownership. journal of financial service professionals, 66, 69-81. social security administration (2019). disability benefits. ssa publication no. 05-10029. icn 456000. baltimore, md (july). son, j., & hanna, s. (2011). the relation between internet use and financial planner use for savings and investment decisions. proceedings of the academy of financial services. stapleton, d. c., o’day, b. l., livermore, g. a., & imparato, a. j. (2006). dismantling the poverty trap: disability policy for the twenty-first century. the milbank quarterly, 84, 701-732. stawski, r. s., hershey, d. a., & jacobs-lawson, j. m. (2007). goal clarity and financial planning activities as determinants of retirement savings contributions. international journal of aging & human development, 64, 13-32. stoneman, p., & kwon, m. j. (1994). the diffusion of multiple process technologies. the economic journal, 104, 420-431. strömbäck, c., lind, t., skagerlund, k., västfjäll, d., & tinghög, g. (2017). does self-control predict financial behavior and financial well-being? journal of behavioral and experimental finance, 14, 30-38. willis, l. e. (2008). against financial literacy education. iowa law review, 94, 08-10. worthy, s. l., jonkman, j., & pike, l. b. (2010). sensation-seeking, risk-taking, and problematic financial behaviors of college students. journal of family and economic issues, 31, 161-170. xiao, j. j., shim, s., barber, b., & lyons, a. (2007). academic success and well-being of college students: financial behaviors matter (tcai report). tucson, az: university of arizona. 54 q. bi et al. / financial services review 29 (2021) 29–54 who uses robo-advisory services, and who does not? martha fulka,*, john e. grable, ph.d., cfpa, kimberly watkinsa, michelle krugera auniversity of georgia, 300 dawson hall, athens, ga 30602, usa abstract the purpose of this study was to compare the demographic, attitudinal, and behavioral characteristics of u.s. consumers in their current and expected use of robo-advisory services, traditional financial planning services, or a combination of the two services. findings showed a difference between those who used robo-advisory services and those who used traditional financial planning services. overall, those who used a traditional financial planner were older and reported higher levels of net worth, while users of robo-advisors, on average, reported lower levels of net worth. in addition, those who used traditional financial planning services reported a larger percentage of their total net worth from an inheritance, whereas a lower percentage of net worth from an inheritance was reported by robo-advisor users. results showed that users of robo-advisory services generally (1) had lower income, (2) had lower net worth, (3) had received no or less inheritance, and (4) were less impulsive financially. © 2018 academy of financial services. all rights reserved. keywords: financial planning; financial planner; robo-advisory services; robo-advisor 1. introduction the term “robo-advisor” encompasses any automated investment or financial planning service that is designed to appeal to the mass market by providing easy and inexpensive access to financial information, advice, products, and services (berger, 2015). those who promote the value and use of robo-advisors point to the following features: * corresponding author. tel.: �1-706-542-4758; fax: �1-706-542-4397. e-mail address: mfulk@uga.edu (m. fulk) financial services review 27 (2018) 173-188 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. 1. costs tend to be lower than what a consumer would pay for services provided by a traditional financial planner (investment management services generally top out at 50 basis points for robo-advisors); 2. investment recommendations are generally implemented using low-cost exchange traded funds (etfs), which helps keep annual management expenses low; 3. all aspects of the investing process, from risk assessment to ongoing rebalancing, are completed quickly; 4. nearly all robo-advisors provide tax tracking features; and 5. access to financial and other information is easily and quickly accessible. some of the primary arguments against the use of robo-advisors include: 1. lack of flexibility; 2. lack of personalization to a client’s unique needs or desires; and 3. lack of personal interaction and relationship with a financial planner. the value provided by robo-advisors can be gauged, in part, by the growth in assets under management among robo-advisory firms. according to research disseminated by a. t. kearney (2015), assets under management held at robo-advisory firms in 2016 exceeded $300 billion. this figure is expected to increase to more than $2 trillion by 2020. if the growth projection holds true, this will signal a significant shift from traditional human interfacing services to one that includes an increasingly significant role for automated advice. although little is known about the types of people drawn to robo-advisory services, nearly all indicators suggest that adopters of fully automated systems tend to be younger (under age 35) and relatively knowledgeable about personal and household financial issues, whereas non-users tend to be older consumers who are nearing or in retirement. those who avoid robo-advisory services are thought to be more cautious (a. t. kearney, 2015). if this general assessment is accurate, the future growth of robo-advisory services will depend on less experienced and cautious consumers’ willingness to transfer assets and information to robo-advisors. if this occurs, there could be profound disruptions to the revenue of firms operating as traditional asset managers and firms providing direct person-to-person financial planning advice (regan, 2015). the purpose of this study was to compare the demographic and attitudinal characteristics of affluent u.s. consumers in their current and expected use of robo-advisory services, traditional financial planning services, or a combination of services. the results from this study add to the literature by providing an empirically robust description of current and potential users of robo-advisory services. the remainder of this paper is focused on providing a background review of robo-advisory services, a description of the research methodology, a report of the findings, and an applied discussion of results. 2. literature review when first conceptualized, robo-advisors were brought to market as an alternative for small investors to gain access to financial advice and portfolio management services in a cost-effective manner (fein, 2015). a key feature associated with robo-advisors is the 174 m. fulk et al. / financial services review 27 (2018) 173-188 delivery of services at low costs. cost savings are generated by automating the process of collecting and analyzing client data, creating and implementing recommendations, and monitoring outcomes. today, robo-advisors have expanded to include noninvestment money management services, such as budgeting, planned saving, expense tracking, mortgage financing services, legal advice, tax services, and general financial planning advice. although federal regulators have yet to determine whether robo-advisory services meet either, both, or neither the suitability or fiduciary standard, there is a general recognition among policymakers that automated services may be a mechanism to expand investment advice and services to more individuals and families (idzelis, 2016). it is important to note, however, that some have expressed concerns about the role of robo-advisors in the financial marketplace. for example, fein (2015) published a critique of robo-advisors and concluded that robo-advisors do not always provide advice that is in the best interest of consumers. fein argued that there is a general lack of transparency in the way consumer information is used. this lack of transparency can lead to undisclosed conflicts of interests that can increase apparent and hidden costs paid by consumers. regardless of the debate that continues over the fiduciary status of robo-advisory firms, it is clear that an increasing number of consumers have embraced the notion of allowing an automated system to manage some or most of their finances. the limited literature on the use of robo-advisors suggests that primary adopters of this technologically driven money management approach tend to be younger consumers who have a high level of trust in online platforms (cutler, 2015; pisani, 2016). what is less well known is how well the adoption of robo-advisory services matches with the help-seeking behavior of consumers for traditional financial planning services. 2.1. help-seeking behavior the financial help-seeking literature is relatively robust, meaning that researchers have been able to describe the general characteristics of those who seek out financial advice from professionals. a recent review by gentile, linciano, and soccorso (2016) summarized much of this literature. gentile et al. noted that basic demographic factors often help differentiate those who seek professional help from those who do not. for example, given the pricing models associated with financial planning and investment management services, household income and net worth tend to be important factors in describing help-seeking behaviors. specifically, those with greater income and wealth tend to be more likely to pay for financial advice and, thus, seek out such advice (finke, huston, and winchester, 2011; miller and montalto, 2001). whether these factors are also important in shaping who may prefer to work with a robo-advisor is less well known; however, it is reasonable to expect that those with less income and lower levels of wealth may be the primary users of robo-advisory services. age, gender, education, and marital status are also known to be associated with helpseeking behavior (auslander and litwin, 1990; elmerick, montalto, and fox, 2002; fischer and farina, 1995; hackethal, haliassos, and jappelli, 2012; gall, kratzer, jones, and decooke, 1990; kaskutas, weisner, and caetano, 1997; phillips and murrell, 1994; robb, babiarz, and woodyard, 2012; salter, harness, and chatterjee, 2010). when asked where 175m. fulk et al. / financial services review 27 (2018) 173-188 financial services are obtained, older individuals are more likely, compared with younger individuals, to report seeking help from traditional financial planners. this is logical given that households managed by older individuals often have more complex financial questions and concerns, as well as the resources needed to pay for services. while there is little consensus on the gender effect, gentile et al. (2016) did note that women seek the help of professional service providers more so than men. generally speaking, those with high levels of attained education also seek out and use professional financial advice more so than those with less attained education (elmerick et al., 2002). similarly, help seekers tend to be married. this may be (1) because of increased complexities faced by married couples, (2) because of higher levels of net worth with combined income, or (3) the need to optimize the use of services to increase time together as a couple. previous research has shown that behavioral and attitudinal characteristics also help describe help-seeking behavior. in the gentile et al. (2016) study, the researchers noted that factors such as saving behavior, impulsivity, financial satisfaction, self-assessed financial knowledge, positive perceptions of one’s financial situation, level of engagement in a household’s financial affairs, and feelings of regret can be used to describe who is likely to seek financial advice from a professional. in general, the profile of help-seekers includes those who (1) take proactive steps to plan and save for the future, (2) use and stick to a budget, (3) are more knowledgeable, (4) are confident, and (5) are less regretful or disappointed over past mistakes (grable and joo, 2001). this does not mean that these factors are causal characteristics. it is possible that the use of professional services enhances knowledge attainment, confidence, and other positive personal characteristics. instead, what the previous literature does show is that those who seek and use help from professional service providers exhibit different characteristics from those who obtain help in other forms. when viewed holistically, the profile of someone who is most likely to work with a traditional financial planner is that of an affluent, well educated, knowledgeable person who has a longer planning horizon (chatterjee and zahirovic-herbert, 2010; salter et al., 2010). of interest, beyond acknowledging that younger, technologically savvy individuals are more likely to use robo-advisory services, little is known about the shared demographic, attitudinal, and behavioral characteristics of those who work with financial planners and/or roboadvisors. a primary outcome associated with this study was to address this gap in the literature. 2.2. methodology data for this study were obtained from the online mechanical turk (mturk) survey system. survey participants were recruited during fall 2015. criteria for inclusion in the study included a requirement that participants be primarily or jointly responsible for the management of household financial decisions. additionally, participants needed to have at least $25,000 in household income and be able to identify the basic definition of net worth. specifically, potential participants needed to answer the following question correctly: “how is net worth calculated?” in total, 608 individuals met each requirement and responded to the survey questions. it is important to note that the sample was not designed to be nationally representative, but rather, the sample was designed to capture the current and future 176 m. fulk et al. / financial services review 27 (2018) 173-188 help-seeking characteristics of relatively affluent consumers (i.e., those who have the financial capacity to pay for financial planning services). the outcome variable of interest in this study was developed based on participants’ answers to two questions. the first asked participants about their use of robo-advisory services (automated investment technology). specifically, each participant was asked to indicate whether she or he “invests with a robo-advisor (automated investment technology).” the second question asked about each participant’s experience working with a traditional financial planner. participants were asked to indicate if they “currently work directly with a financial advisor, wealth manager, or other professional advisor to help you manage your finances and/or manage your investments.” three response categories were provided for each question: (1) i currently use this service; (2) i do not use this service but plan to; and (3) i do not use this service, and i do not plan to. answers were recoded dichotomously so that 1 � currently use or plan to use the service and 0 � do not use and do not plan to use the service. the final outcome variable was then coded as: 1 � only use a robo-advisor, 2 � only use a financial planner, 3 � use both a robo-advisor and financial planner, and 4 � use neither a robo-advisor nor financial planner. approximately 9%, 29%, 11%, and 51% of survey participants fell into each category, respectively. fifteen independent variables were used to identify the help-seeking characteristics of participants. the choice of the variables was based on a review of the financial help-seeking literature. age was measured as a continuous variable. the mean, median, and standard deviation of the age variable was 35.85, 33.00, and 10.76 years, respectively. gender was coded 1 � female and 0 � male. the sample was split almost evenly between women and men. education was measured on an ordinal scale with eight categories: (1) high school graduate, (2) some college, no degree, (3) associate’s degree, occupational, (4) associate’s degree, academic, (5) bachelor’s degree, (6) master’s degree, (7) doctoral degree, and (8) professional degree. over 60% of study participants held a bachelor’s degree or higher level of education. over 95% of the respondents stated they were the primary financial decision maker or that they jointly participated in making household financial decisions. household income and net worth were measured in u.s. dollars. given skewness in the data, income and net worth were normalized using templeton’s (2011) two-step approach for transforming continuous variables. the mean and standard deviation for household income was $79,017.74 and $63,406.31, respectively. the mean and standard deviation for household net worth was $154,650.11 and $388,594.68, respectively. an inheritance variable was included to account for the possibility that a large asset windfall in the past might have prompted someone to seek professional financial advice. each participant was asked to indicate the percentage of her or his overall net worth that was a result of an inheritance. responses ranged from 2% to 58%, with a mean of 8.64% and a standard deviation of 12.20%. marital status was coded into three mutually exclusive variables: (1) single, never married (39%); (2) married, never divorced (47%); and (3) other, which included those who were remarried, widowed, and divorced (14%). saving behavior was measured by asking participants to indicate what percentage of their income was being saved each year. responses ranged from none to 77%, with the mean response being 12.80% (sd � 12.87%). impulsivity was measured by asking: “how often have you ignored budgets or plans when making large-scale purchases?” a five-point 177m. fulk et al. / financial services review 27 (2018) 173-188 likert-type scale was used to record responses, which ranged from 1 � never to 5 � very often/always. the mean and median score was 2.22 and 2.00, respectively. financial satisfaction was measured with the following item: “in general, how satisfied are you with your current financial situation?” a five-point likert-type scale using 1 � very unsatisfied and 5 � very satisfied was used to code answers. the mean and median for the question was 2.95 and 3.00, respectively. a participant’s perceived ability as an investor was assessed by asking: “in the past, how often have you thought of yourself as a smart/savvy investor?” a five-point likert-type scale was used to code responses, which ranged from 1 � never to 5 � very often/always. the mean and median for the item was 2.84 and 3.00, respectively. whether a participant felt she or he currently earned enough money was measured by asking: “my family’s current income is sufficient for most needs and wants.” a five-point likert-type agreement scale was used to code responses, with 1 � strongly disagree to 5 � strongly agree. the mean and median for the question was 3.62 and 4.00, respectively. financial and investment acumen was measured by asking the following question: “i often do not know how my investments are performing.” a five-point likert-type agreement scale was used to code responses, with 1 � strongly disagree to 5 � strongly agree. the mean and median for the question was 2.33 and 2.00, respectively. regret and disappointment was assessed with the following item: “how often have you been disappointed by the financial decisions you have made?” a five-point likert-type scale was used to record responses, which ranged from 1 � never to 5 � very often/always. the mean and median response was 2.75 and 3.00, respectively. three tests were used to identify the demographic, attitudinal, and behavioral characteristics of those who, at the time of the survey, worked with a traditional financial planner, a robo-advisor, both, or neither. two tests were used to screen the 15 independent variables to identify which of the variables were statistically significantly associated with help-seeking behavior. these tests were based on bivariate relationships. first, given the categorical nature of gender and marital status, a series of �2 analyses were conducted. second, analysis of variance (anova) tests were run with the variables coded at the continuous or ordinal level. as the first step in the analytical procedure, results from the �2 and anova analyses were used to identify participant characteristics that were highly associated with help-seeking behavior. these variables were then used in the third test, a multinomial regression with type of financial service used as the dependent variable in the model. this two-step procedure revealed the most important individual and household level characteristics associated with the use of financial planning services. 3. results as shown in table 1, and based on the �2 results, the association between gender and help-seeking was not significant, �2 (3) � 0.904, p � 0.824. also shown in table 1 are the �2 results for the association between marital status categories and help-seeking. no significant association was noted for those who were married (�2 (3) � 2.598, p � 0.458), single never married (�2 (3) � 3.306, p � 0.347), or remarried, widowed, or 178 m. fulk et al. / financial services review 27 (2018) 173-188 divorced (�2 (3) � 2.890, p � 0.409). as a result, these variables were not included in the final multinomial regression analysis. table 2 shows the results from the anova tests. the table shows the variation in means across the four groups. it is important to note that the differences in means should not be interpreted as predictors of help-seeking behavior. instead, the results simply show what characteristics were associated with help-seeking behavior. consider, for example, the age result. data in table 2 provide a more nuanced insight into who was using or considering using a financial planning service. those who used or planned to use a financial planner were found to be older than those who used or planned to use the table 1 cross tabulations showing differences in help-seeking by gender and marital status robo planner both neither female count 27 82 33 152 expected count 27.8 86.8 31.7 147.7 % 9.2% 27.9% 11.2% 51.7% standardized residual �.1 �.5 .2 .4 male count 29 93 31 146 expected count 28.2 88.2 32.3 150.3 % 9.7% 31.1% 10.4% 48.8% standardized residual .1 .5 �.2 �.3 married count 21 87 29 139 expected count 25.9 80.9 29.6 139.6 percentage within married � 1 7.6% 31.5% 10.5% 50.4% standardized residual �1.0 .7 �.1 �.1 not married count 35 88 35 163 expected count 30.1 94.1 34.4 162.4 percentage within married � 1 10.9% 27.4% 10.9% 50.8% standardized residual .9 �.6 .1 .0 single never married count 25 62 30 122 expected count 22.4 70.1 25.6 120.9 percentage within single never married � 1 10.5% 25.9% 12.6% 51.0% standardized residual .5 �1.0 .9 .1 other than single marital status count 31 113 34 180 expected count 33.6 104.9 38.4 181.1 percentage within single never married � 1 8.7% 31.6% 9.5% 50.3% standardized residual �.4 .8 �.7 �.1 remarried, widowed, divorced count 10 26 5 41 expected count 7.7 24.0 8.8 41.5 percentage within remarried, widowed, divorced � 1 12.2% 31.7% 6.1% 50.0% standardized residual .8 .4 �1.3 �.1 other than remarried, widowed, divorced count 46 149 59 261 expected count 48.3 151.0 55.2 260.5 percentage within remarried, widowed, divorced � 1 8.9% 28.9% 11.5% 50.7% standardized residual �.3 �.2 .5 .0 179m. fulk et al. / financial services review 27 (2018) 173-188 table 2 anova results related to help-seeking behavior n mean sd f post hoc test age robo 56 34.3750 10.36351 3.328* planner � both planner 174 37.0230 11.05109 both 64 32.4531 8.74131 neither 300 36.2267 10.91447 education robo 56 5.4643 1.61768 1.113 no differences planner 175 5.4743 1.69454 both 64 5.5156 1.67135 neither 301 5.2292 1.70995 normalized household income robo 55 73915.7866 64190.15510 3.313* planner � neither planner 157 91060.7398 60051.27519 both 55 84177.0123 61229.55156 neither 258 71676.4239 65032.61515 normalized household net worth robo 54 63105.0342 405920.06781 5.911*** planner � robo planner � neitherplanner 160 258430.8394 400112.05336 both 56 150437.9558 350868.55523 neither 265 113753.6146 376795.38976 normalized percent of net worth because of inheritance robo 55 3.4282 6.94182 7.876*** planner � robo planner � neitherplanner 175 12.0661 14.64660 both 64 8.0936 12.10211 neither 300 7.8226 12.25119 percent of income saved robo 56 14.4643 13.98826 2.460 no differences planner 175 13.5829 12.17604 both 64 15.4531 13.89979 neither 301 11.5017 12.79782 ignore budget when making large purchases robo 56 1.9464 .74881 3.657* both � robo planner 175 2.2857 .80841 both 63 2.4286 .92831 neither 300 2.1933 .87091 financial satisfaction robo 56 2.6786 1.06356 1.762 no differences planner 175 3.0286 1.08505 both 63 3.0635 1.04531 neither 302 2.9238 1.09552 consider self to be savvy investor robo 56 2.9464 .94233 1.510 no differences planner 174 2.8563 .89122 both 64 3.0000 .79682 neither 302 2.7616 1.02272 (continued on next page) 180 m. fulk et al. / financial services review 27 (2018) 173-188 services of both a robo-advisor and a financial planner. the age of those using a robo-advisor was less than the age of those who used a financial planner, but the difference was not significant. based on what has been reported in the literature, it was not surprising that household income was highest for those working with a traditional financial planner. a statistical difference in the test was noted between those who used a financial planner and those who had no intention of working with either a robo-advisor or financial planner. the net worth finding provides more insight into the help-seeking issue. those who used a traditional financial planner had the highest net worth, while users of robo-advisors had the lowest net worth (although not the lowest income). also, the net worth of those who worked with a financial planner was higher than the net worth of those who had no plans to work with either a robo-advisor or traditional financial planner. an almost identical finding was noted for the inheritance variable. overall, those who reported working with a traditional financial planner reported a larger percentage of their total net worth from an inheritance. the lowest percentage of net worth from an inheritance was reported by robo-advisor users. this may be related to the tendency among robo-advisor users to be younger, which means they may not have yet received an inheritance. according to the anova results, impulsivity was found to be associated with helpseeking behavior. those who more frequently ignored their budget when making a large purchase reported using the services of both a robo-advisor and a financial planner. the least impulsive were users of robo-advisors. one attitudinal variable was found to be related to help-seeking behavior. those least disappointed by their financial decisions tended to use the table 2 (continued) n mean sd f post hoc test household income sufficient to meet needs robo 56 3.5357 1.11133 1.337 no differences planner 174 3.7471 1.07755 both 64 3.6563 1.02692 neither 302 3.5464 1.13952 do not know how own investments are performing robo 56 2.1607 .86921 0.783 no differences planner 174 2.3851 1.01776 both 64 2.2500 1.16837 neither 301 2.3289 1.03350 disappointed by financial decisions you have made robo 55 2.8909 .71162 2.617* both � planner planner 175 2.6743 .79667 both 64 2.9531 .86244 neither 301 2.7143 .81533 anova � analysis of variance. *p � .05, **p � .01, ***p � .001. 181m. fulk et al. / financial services review 27 (2018) 173-188 services of a traditional financial planner. those who reported the highest levels of disappointment were those who used the services of both robo-advisors and financial planners. several variables were not significant in the anova tests. specifically, percent of income saved, financial satisfaction, being a savvy investor, having enough income to meet needs, and knowing about investment performance were not associated with help-seeking behavior. the results from the �2 and anova tests provided insights into the differences among those who used only a robo-advisor, only a financial planner, both a robo-advisor and a financial planner, and those who used neither. based only on bivariate relationships, the services of traditional financial planners appear to appeal to those with greater income and net worth. there was a statistically significant difference in the average income and net worth of those who used a financial planner compared with those who did not use either service. additionally, the use of a traditional financial planner was related to receiving an inheritance. those who had received an inheritance reported working with a traditional financial planner rather than a robo-advisor. it may be that a substantial increase in net worth because of an external event prompts an individual or family to seek the help of a traditional financial planner. attitudinally, those who worked with a traditional financial planner reported being less disappointed with past financial decisions. it is worth noting, however, that when tested together in one model, some of these relationships described above may change. those who used robo-advisors tended to be younger, but the difference in mean age compared with those who used traditional financial planning services was not significant. users of robo-advisory services also reported having lower income and lower net worth. additionally, robo-advisor users reported less wealth from inheritances, but they also reported being less impulsive financially. participants who reported the combined use of robo-advisors and financial planners were, on average, the youngest and most impulsive. those fitting this profile may be attempting to explore the most appropriate type of advice and guidance given their situation, or impulsive clients might be signing up for both services because they lack the patience to research alternatives. on the other hand, the propensity for younger clients to use both services might be best explained by a possible quest to determine which of the two services provides the best benefit—based on each participant’s personality, lifestyle choices, and long-term goals—at the lowest cost. those currently in this dual use category may eventually gravitate toward one service provider and drop the other. household income was found to be a driving factor in describing help-seeking behavior for those who were not currently using the services of a robo-advisor, a traditional financial planner, or did not plan to use services of either type of provider in the future. among the four groups of help seekers, household income was the lowest for those in the “neither” category. similarly, those in this group reported a lower net worth than those who reported working with traditional financial planners. these respondents also reported holding a smaller proportion of their net worth from an inheritance. however, it is interesting to note that in comparison to those who reported working with a robo-advisor, those in the neither category held more wealth. each of the statistically significant variables from the first two tests were combined into a multivariate model as a way to identify the most important help-seeking characteristics. table 3 shows the results from the multinomial regression. the regression used 182 m. fulk et al. / financial services review 27 (2018) 173-188 t ab le 3 m ul tin om ia l re gr es si on sh ow in g fa ct or s as so ci at ed w ith he lp -s ee ki ng be ha vi or (n � 50 3) r ef er en ce ca te go ry � ne ith er r ob o pl an ne r b ot h b (s e ) o dd s ra tio b (s e ) o dd s ra tio b (s e ) o dd s ra tio in te rc ep t � 1. 33 5 (. 84 4) � 1. 56 2 (. 54 0) ** � 1. 30 6 (. 85 7) a ge � 0. 02 4 (. 01 7) 0. 97 6 0. 00 2 (. 01 1) 1. 00 2 � 0. 05 5 (. 01 9) ** * 0. 94 7 n or m al iz ed ho us eh ol d in co m e 0. 00 0 (. 00 0) 1. 00 0 0. 00 0 (. 00 0) 1. 00 0 0. 00 0 (. 00 0) 1. 00 0 n or m al iz ed ne t w or th 0. 00 0 (. 00 0) 1. 00 0 0. 00 0 (. 00 0) * 1. 00 0 0. 00 0 (. 00 0) 1. 00 0 n or m al iz ed pe rc en t of ne t w or th be ca us e of in he ri ta nc e 0. 00 0 (. 00 0) * 1. 00 0 0. 00 0 (. 00 0) * 1. 00 0 0. 00 0 (. 00 0) 1. 00 0 ig no re bu dg et w he n m ak in g la rg e pu rc ha se s � 0. 67 3 (. 22 4) ** 0. 51 0 0. 24 2 (. 14 4) 1. 27 4 0. 09 6 (. 21 0) 1. 10 1 d is ap po in te d by fin an ci al de ci si on s yo u ha ve m ad e 0. 75 3 (. 23 3) ** * 2. 12 3 � 0. 06 0 (. 16 0) 0. 94 1 0. 41 5 (. 23 3) 1. 51 4 r 2 � 0. 14 (c ox an d sn el l) , r 2 � 0. 16 (n ag el ke rk e) . m od el � 2 (1 8) � 78 .0 34 , p � .0 01 . *p � .0 5, ** p � .0 1, ** *p � .0 01 . 183m. fulk et al. / financial services review 27 (2018) 173-188 the help-seeking categories as the dependent variable, with the neither category as the reference group. the advantage associated with using a multinomial regression, compared with a �2 or anova test, in this type of study is that the model can be designed to more precisely identify the variables that are most highly associated with help-seeking behavior, controlling for the other variables in the model. the following summarizes the results of the multinomial logit. compared with those who use neither a robo-advisor nor a traditional financial planner: y younger participants were more likely to report using services from both robo-advisors and traditional financial planners. y participants with a higher net worth were more likely to use the services of a traditional financial planner. y participants who reported a higher proportion of their net worth was from an inheritance were more likely to use the services of a robo-advisor. y participants who reported a higher proportion of their net worth was from an inheritance were more likely to use the services of a traditional financial planner. y participants who reported being impulsive with large purchase decisions or who tended to ignore their budget were less likely to use the services of a robo-advisor. y participants who reported feeling disappointed by past financial decisions were more likely to use the services of a robo-advisor. 4. discussion findings from the logit analysis generally supported the results from the �2 and anova tests. on the whole, users of robo-advisory and financial planning services were similar in reporting more of their net worth coming from an inheritance when compared with those who had no interest in seeking help. this suggests that the receipt of an inheritance or sudden windfall is a trigger that prompts help seeking. the finding showing that younger participants were more likely to use services from both robo-advisors and traditional financial planners, compared with those who reported working with neither a robo-advisor or financial planner, suggests that these “do-it-yourself” investors may be comprised of older individuals. those who used (or planned to use) a robo-advisor were less likely to exhibit impulsivity. whether this was a result of following robo-advisory recommendations or a tendency for nonimpulsive types to seek out automated systems is a question that deserves additional study. for example, what may be occurring is a change in behavior prompted or mediated by robo-advisory services. although a reasonable hypothesis, all that can be said from this study is that current users of robo-advisory services do exhibit less impulsivity. robo-advisor users were also found to be more regretful (disappointed) in their past financial behavior compared with those who did not use a financial planning service. this finding hints that the hypothesized effect of robo-advisory services on impulsivity, described above, may not hold when evaluated because one would expect to see a similar result in relation to disappointment. as a robustness check, a correlation analysis was conducted to 184 m. fulk et al. / financial services review 27 (2018) 173-188 ensure that multicollinearity was not present in the logit model. the only relationship among the independent variables that was of medium or larger effect was the association between impulsivity and disappointment (r � 0.45). no other issues of high correlation or multicollinearity were observed. findings from this study partially validate what has been previously reported in the literature. those who used or planned to use a robo-advisor exhibited certain characteristics; however, it is important to note that the list of characteristics is not fully inclusive. in many respects, the type of person who was at the time of the survey working with a robo-advisor was similar to the profile of those who were working with a traditional financial planner. there were, however, four distinguishing help-seeking characteristics that differentiated the two groups: (1) older clients generally reported working with traditional financial planners (a human interaction); (2) those with high incomes reported working with traditional financial planners; (3) those with more wealth reported working with traditional financial planners; and (4) those with more wealth attributable to an inheritance reported working with a traditional financial planner. it appears, at least among those who participated in this study, that robo-advisors fill a unique niche for consumers who (1) do not meet asset under management (aum) minimums for traditional financial planning firms or (2) desire a financial planning service that provides fast and low-cost service. the projected growth of robo-advisory services, and the firms operating in this segment of the market, is expected to develop at a rapid pace. this indicates that robo-advisory services are not transient. it is worth considering, however, that those who currently use robo-advisory services may find services to be inadequate or inappropriate in the future, particularly as the complexity of a help-seeker’s financial situation changes over time. the lack of flexibility, personalization, and human interaction will continue to be a challenge for robo-advisors as the financial needs and net worth of existing clients increase. if the demographic and attitudinal profile of help-seekers remains consistent across time, opportunities for traditional financial planners may emerge. consumers may abandon roboadvisors for more traditional firms as the real or perceived need for a more personal service grows in tandem with household income, wealth, and financial planning complexity. it would be a mistake to assume consumers will always opt in to the least expensive and quickest product or service on the market. clients generally consider many factors when pursuing financial help, often weighing the costs and benefits (both financial and emotional) associated with maintaining a financial planner and/or robo-advisor relationship. thus, traditional financial planners can add value by developing meaningful client-financial planner relationships and creating personalized solutions based on each client’s financial situation. keep in mind, however, that this is only speculation at this point. more time and research will be needed to better understand consumers’ preferences for robo-advisory services over traditional financial planning firm offerings. the findings from this study call out for additional research efforts to better understand the similarities and differences between those who use (or plan to use) robo-advisory and/or traditional financial planning services. an important question that remains unanswered is: “how fluid is the use of robo-advisory services?” data from this study confirm previous findings in the literature; namely, those who use robo-advisors tend to be younger, have less household income, and lower household net worth. hence, it is reasonable to ask if these 185m. fulk et al. / financial services review 27 (2018) 173-188 current users will migrate to traditional financial planners as they age and experience an increase in income and net worth. it is also reasonable to ask if those who currently use robo-advisors switch to traditional financial planning services upon the receipt of an inheritance or other financial windfalls that significantly increases household net worth. one answer to this question is that the financial planning marketplace is undergoing a transformational change, and there is no reason to expect clients will shift services over time. the robo-advisory business may be counting on the inelasticity of consumer preferences to dominate future help-seeking decisions. rather than shifting to a traditional financial planner as financial circumstances improve, robo-advisory firms may be counting on either offering new services in the future or hoping that the satisfaction level of current clients outweighs the advantages of migrating to a financial planner. essentially, the argument is that millennial consumers are so technologically savvy that they will continue to prefer automated advice regardless of their household financial circumstances. the counterargument is that some shift in help-seeking behavior is likely to occur over time. when individuals are starting out on their lifetime financial journey, they may feel that a fully automated system is the best option in terms of obtaining appropriate financial advice at a very low-cost entry point. the current marketplace for financial planning services appears to reinforce this perception. while robo-advisory minimums for assets under management range from a few thousand dollars to approximately $50,000, nearly all asset management platforms offered by traditional financial planners, and nearly all advisorybased financial service firms, have minimums that start at six figures. this requirement effectively shuts out most americans and nearly all young professionals. financial planners who are interested in preempting younger prospective clients from beginning the financial process via robo-automation might consider using the findings from this study to create access to services for younger and less wealthy prospective clients. this might require a financial planner to provide services at a discount or even a potential loss (e.g., hiring younger financial planners as a loss-lead expense), with the hope that younger clients will maintain the relationship as their income and wealth grows over time. 5. limitations and future directions limitations associated with the data and research methodology do exist in this study. as noted earlier, the sample was not designed to be nationally representative, hence findings from the study should not be applied nationally across the population. in many respects, the use of mturk samples, even if precisely developed, does introduce a degree of sample selection bias. while steps were taken to limit participation to those who had investable assets, basic financial knowledge, and the financial capacity to pay for financial advice, it is possible that the final sample was biased in unknown ways. as such, it is important to note that the sample and results describe the help-seeking behaviors of relatively affluent consumers. the issue of endogeneity and dual causality is something new help-seeking studies should explore. it is possible, although not likely, that the income and net worth results found in this study may have been influenced through the use of financial planning services. specifically, 186 m. fulk et al. / financial services review 27 (2018) 173-188 some may argue that income and net worth were higher among participants who used a traditional financial planner because the advice received from a financial planner lead to income and wealth gains. this is certainly a possibility but one that may not hold true when tested. for example, financial planning, as a profession, is plagued by a high asset minimums and relatively high planning costs, even among those who provide services by the hour. sadly, it takes existing financial capacity to pay for financial planning services, which makes it likely that the profile of those who use traditional financial planning services, as described in this paper, must exist before a financial planning engagement. the issue of dual causality may be more likely in relation to attitudinal and behavioral characteristics. knowledge, disappointment, regret, impulsivity, and other characteristics may be influenced by the type of service a help-seeker obtains. results from the current study can only be used to say that those who currently use or plan to use a service share a particular profile. it is also worth noting that the growth in robo-advisory services may just be the result of access issues. given the demand for financial planning services and the limited supply of traditional financial planners, consumers with limited resources may, by default, be pushed toward the use of robo-advisory services, even if a personal relationship is desired. the tracking of consumers over time, either through longitudinal or panel study, would be very helpful in identifying help-seeking trends. it is possible that as a consumer’s financial situation improves, access to person-to-person financial planning will become more feasible, primarily because the consumer will be in a position to pay for services. a personal relationship may also become a more valued attribute associated with good service for millennials and others in the future. of course, both arguments are conjecture at this point. what is needed is further research on the dynamics of help-seeking in the financial planning marketplace. one way to do this is through a longitudinal study of consumer preferences, attitudes, and behaviors in relation to help-seeking behavior. it behooves financial planning practitioners, policymakers, and regulators to better understand the attitudes and preferences of consumers in light of technological changes. without additional data and empirical work, the financial planning profession may find the professional services (and potential consumers of such services) migrate to other platforms as new viable practice models begin to emerge in the future. acknowledgments the authors wish to thank dr. sarah fallaw for her assistance in collecting data and providing valuable feedback regarding this manuscript. references auslander, g. k., & litwin, h. (1990). social support networks and formal help-seeking: differences between applicants to social services and a non-applicant sample. journal of gerontology, 45, 112–119. berger, r. (2015, february 5). 7 robo-advisors that make investing effortless. forbes. (available at https://www. forbes.com/sites/robertberger/2015/02/05/7-robo-advisors-that-make-investing-effortless/#1c81cf9c4ae4) 187m. fulk et al. / financial services review 27 (2018) 173-188 chatterjee, s., & zahirovic-herbert, v. (2010). retirement planning of younger baby-boomers: who wants financial advice? financial decisions, 22, 1–12. cutler, n. e. (2015). millennials and finance: the ‘amazon generation’. journal of financial services professionals, 69, 33–39. elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217–231. fein, m. l. (2015). robo-advisors: a closer look. ssrn working paper 2658701. (available at from: http://ssrn.com/abstract�2658701) finke, m. s., huston, s. j., & winchester, d. d. (2011). financial advice: who pays? journal of financial counseling and planning, 22, 18–26. fischer, e. h., & farina, a. (1995). attitudes toward seeking professional psychological help: a shortened form and considerations for research. journal of college student development, 36, 368–373. gall, n.-l., kratzer, l., jones, e., & decooke, p. (1990). children’s self-assessment of performance and task related help-seeking. journal of experimental child psychology, 49, 245–263. gentile, m., linciano, n., & soccorso, p. (2016). financial advice seeking, financial knowledge and overconfidence: evidence from the italian market. consob, commissione nazionale per le societa e la borsa, working paper 83. rome: consob. grable, j. e., & joo, s.-h. (2001). a further examination of financial help-seeking behavior. journal of financial counseling and planning, 12, 55–74. hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking and finance, 36, 509–524. idzelis, c. (2016, april 7). new dol fiduciary rule, robo-advice dominate adviser concerns at sifma conference. investment news. (available at http://www.investmentnews.com/article/20160407/free/160409937/new-dolfiduciary-rule-robo-advice-dominate-adviser-concerns-at) kaskutas, l. a., weisner, c., & caetano, r. (1997). predictors of help-seeking among a longitudinal sample of the general population, 1984–1992. journal of studies on alcohol, 58, 155–161. kearney, a. t. (2015, june). hype vs. reality: the coming waves of “robo” adoption. (available at https:// www.atkearney.com/documents/10192/7132014/hype�vs.�reality_the�coming�waves�of�robo� adoption.pdf/9667a470-7ce9-4659-a104-375e4144421d) miller, s. a., & montalto, c. p. (2001). who uses financial planners? evidence from the 1998 survey of consumer finances. consumer interests annual, 47, 1–9. phillips, m. a., & murrell, s. a. (1994). impact of psychological and physical health, stressful events, and social support on subsequent mental health help-seeking among older adults. journal of consulting and clinical psychology, 62, 270–275. pisani, j. (2016). robo-advisors: a disruptive force in the wealth management industry (doctoral dissertation). university of malta, malta. (available at https://www.um.edu.mt/library/oar/handle/123456789/13703) regan, m. p. (2015, june 18). robo advisers to run $2 trillion by 2020 if this model is right. bloomberg. (available at https://www.bloomberg.com/news/articles/2015-06-18/robo-advisers-to-run-2-trillion-by-2020if-this-model-is-right) robb, c., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. salter, j. r., harness, n., & chatterjee, s. (2010). utilization of financial advisors by affluent retirees. financial services review, 19, 245–263. templeton, g. f. (2011). a two-step approach for transforming continuous variables to normal: implications and recommendations for is research. communications of the association for information systems, 28, 41–58. 188 m. fulk et al. / financial services review 27 (2018) 173-188 academy of financial services officers president robert moreschi virginia military institute president-elect swarn chatterjee university of georgia executive vice president-program janine scott university shepherd vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations philip gibson winthrop university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president thomas coe quinnipiac university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards inga chira california state university, northridge victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university terrance martin university of texas–rio grande valley martin seay kansas state university past presidents thomas coe, 2015-2016 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial.org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2018 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 27, no. 2, 2018 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2018 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. pii: 1057-0810(91)90054-3 187 index volume 1,1991/1992 authors chinloy, peter, “real estate income and relocation,” l(1): 45-59 cox, larry, “review: disability income insurance and the individual,” l(1): 61-78 crabb, ronald r., “probabilistic estate planning,” l(2): 143-157 kleiman, robert t., and anandi p. sahu, “life insurance companies as invest ment managers: new implications for consumers,” l(1): 23-34 madura, jeff, and thomas j. o’brien, “international diversification for the individual: a review,” l(2): 159 175 mann, steven v., and donald p. solberg, “should individual investors avoid the stock market outside of janu ary?” l(2): 101-107 markowitz, harry m., “individual versus institutional investing,” l(1): l-8 murphy, neil b., “determinants of household check writing: the impacts of the use of electronic banking services and alternative pricing of services,” l(i): 35-44 nunnally, bennie h., jr., see plath, d. anthony o’brien, thomas j., see madura, jeff plath, d. anthony, and nunnally, bennie h., jr., “effective credit costs in retail financial markets: leasing versus borrowing,” l(2): 109-129 potts, tom l., and william reichenstein, “the optimal allocation of pension fund assets: an individual’s perspective,” l(1): 9-22. puelz, amy v., and robert puelz, “personal financial planning and the allocation of disposable wealth,” l(2): 87-99 puelz, robert, see puelz, amy v. reichenstein, william, see potts, tom l. sahu, anandi p , see kleiman, robert t. solberg, donald p, see mann, steven v. yohannes, arefaine g., “comparing mortgages with different payment frequencies,” l(2): 131-142 articletitles ‘comparing mortgages with different payment frequencies,” arefaine g. yohannes, l(2): 131-142 “determinants of household check writ ing: the impacts of the use of elec tronic banking services and alternative pricing of services,” neil b. murphy, l(1): 35-44 financial services review, l(2) 1991 “effective credit costs in retail financial markets: leasing versus borrow ing,” d. anthony plath and bennie h. nunnally, jr., l(2): 109-129 “individual versus institutional invest ing,” harry m. markowitz, l(1): 1-8 “international diversification for the individual: a review,” jeff madura and tom o’brien, l(2): 159-175 “life insurance companies as investment managers: new implications for consumers,” robert t. kleinman and anadi p sahu, l(1): 23-34 “optimal allocation of pension fund assets: an individual’s perspective, the,” tom l. potts and william reichenstein, i( 1): 9-22. “personal financial planning and the allocation of disposable wealth,” amy v. puelz and robert puelz, l(2): 87-99 “probabilistic estate planning,” ronald r. crabb, l(2): 143-157 “real estate income and relocation,” peter chinloy, i( 1): 45-59 “review: disability income insurance and the individual,” larry cox, i( i): 61-78 “should individual investors avoid the stock market outside of january?” steven v. mann and donald p. solberg, l(2): 101-107 portfolio insurance using leveraged etfs jeffrey georgea, william j. trainor, jr.a,* adepartment of economics and finance, east tennessee state university, box 70686, johnson city, tn 37614, usa abstract this study examines the use of leveraged exchange traded funds (letfs) within a constant proportional portfolio insurance (cppi) strategy. the advantage of using letfs in such a strategy is that it allows a greater percentage of the portfolio to be invested in the risk-free rate relative to a traditional cppi. where a standard cppi strategy may require 50% of the portfolio to be invested in equities, using a 2x letf only requires 25%, and a 3x letf only requires 16.7% to attain the same effective exposure to equities. results show when the risk-free asset is yielding at least 3% or the 1 year minus 90-day treasury exceeds 1%, the use of letfs within a cppi framework results in annual returns approximately 1–2% higher with better sharpe, sortino, omega, and cumulative prospect values while reducing value at risk (var) and excess shortfall (es) below var. © 2017 academy of financial services. all rights reserved. jel classification: g11; g17 keywords: portfolio insurance; leveraged exchange traded funds 1. introduction with two major market crashes in the last 17 years, portfolio insurance, or the protection of downside risk, has become increasingly important as “once in a century” events are occurring multiple times instead. the two main types of portfolio insurance are option based (leland and rubinstein, 1976) and constant proportionate portfolio insurance (cppi) strategies set forth by black and jones (1987). most option based ideas are premised on purchasing or creating synthetic puts on an index, effectively a protective put. the cost of * corresponding author. tel.: �1-423-439-5668; fax: �1-423-439-8583. e-mail address: trainor@etsu.edu (w.j. trainor) financial services review 26 (2017) 387–403 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. this protection is usually quite high and although it reduces downside exposure, gains tend to be moderated significantly. cppi strategies are based on a portfolio floor value where a percentage of the portfolio is invested in the risky asset and the remainder in a risk-free asset. if the risky asset value declines, the percentage in the risky asset is reduced. this exposure can decrease to zero if a decline in the risky asset causes the portfolio value to reach the floor, where the floor is defined as the minimum value that the portfolio can fall to over an investment period. research on portfolio insurance strategies is extensive with most research coming down on the side of cppi. cesari and cremonini (2003) test different dynamic strategies including cppi and option based portfolio insurance among buy-and-hold and constant mix strategies. they find cppi strategies are dominant in bear and no-trend markets and considered more beneficial overall. zieling, mahayni, and balder (2014) review an extensive research list and show that using a time varying multiple to dictate the amount of market exposure improves cppi results. pezier and scheller (2014) show cppi strategies are superior to option based strategies when implemented in discrete time. annaert, osselaer, and verstraete (2009) compare strategies under a stochastic dominance approach and generally find one strategy does not outperform another, including buy-and-hold when considering first, second, and third order stochastic dominance. however, maalej and prigent (2016) finds cppi outperforms option based portfolio insurance based on third order stochastic dominance. bertrand and prigent (2011) analyze option based portfolio insurance and cppi strategies using downside risk measures and performance measures that consider the nonnormality of returns, otherwise known as kappa performance measures. they find the cppi method outperforms option based portfolio insurance using the omega measure. kahneman and tversky’s (1979) prospect theory, which assumes investors weigh losses more than gains, is particularly relevant in terms of portfolio insurance for mitigating downside risk. tversky and kahneman (1992) expand this idea with the introduction of cumulative prospect theory. dichtl and drobetz (2011) use this idea to show portfolio insurance strategies are superior to buy-and-hold based on higher cumulative prospect values (cpvs). this study attempts to combine aspects of both cppi and option based portfolio insurance. specifically, by incorporating leveraged exchange traded funds (letfs) within a cppi strategy, this study shows cppi results can be improved. in a typical cppi strategy where 50% of the portfolio is invested in equities, the use of a 2x letf only requires 25% to be invested in equities while attaining the same effective equity exposure. thus, rather than having only 50% earning the risk-free rate, 75% of the portfolio is earning the risk-free rate. using a 3x letf only requires 16.67% in the risky asset to attain the same 50% exposure. the drawbacks of using letfs are their higher expense ratios, inherent financing costs through their use of derivatives, and return decay over time relative to what the daily leverage ratio might imply. for instance, a 2x letf usually falls short of multiplying the index return by two over longer holding periods. however, with more of the portfolio earning the risk-free rate relative to a standard cppi using the underlying index, the gains on average exceed the costs. this study finds that if the risk-free rate is at least 3% or the one year minus the 90-day treasury exceeds 1%, the simple substitution of letfs for the underlying index within a cppi strategy results in higher returns with better sharpe, sortino, omega, and cpv values 388 j. george, w.j. trainor / financial services review 26 (2017) 387–403 while reducing value at risk (var) and excess shortfall (es) below var. treasury rates less than 3% or flat yield curves generally eliminate the value of using letfs in a cppi framework. however, even under low interest rate conditions, using letfs within a cppi framework does not result in extreme underperformance (less than 0.4% annually) relative to a standard cppi. the drawback is with low t-bill rates, the advantage of having a letf effectively “borrow” short so the investor can earn a higher return with excess funds over the year is eliminated. this drawback can be overcome to the extent an investor is willing to take on more risk by investing excess funds in higher performing but riskier assets such as a diversified bond fund. 2. review of letfs although leveraged mutual funds have been around since at least 1993, they did not gain traction until proshares introduced the first 2x letf in 2006. since that time, they have expanded dramatically and as of 2017, there are more than 170 letfs with $40� billion in assets on a variety of assets and indexes including, gold, oil, foreign currencies, treasurybond futures, and a myriad of equity indexes. in general, most letfs magnify the daily return of an underlying index up to �3.0x, although there are few funds that magnify the monthly return. recently, several funds have been proposed to deliver �4.0x, (hunnicutt and mccrank, 2017). strictly speaking, the proposed 4x funds will magnify index futures, but the effect will generally be the same because the cost of carry is mitigated because of the underlying dividends paid. letfs attain their leverage by using derivative assets such as futures and swaps. it should be noted that although the swaps are based on the underlying daily index returns and libor, there is counterparty risk. thus, a large gain to the letf could theoretically not be paid by the counterparty. although an unlikely scenario, not an impossible one. the primary drawback to letfs is realized leverage over time is usually less than what the daily multiple might imply (trainor and baryla, 2008). thus, while the realized return over time often falls short of the daily leverage ratio, the risk does not. to enumerate, assume an underlying index falls 5% on day 1, and increases 10% on day 2 for a 2-day return of 4.5%. a 3.0x letf would lose 15%, then gain 30% for a 2-day return of 10.5% resulting in an effective leverage ratio of 2.3 (10.5%/4.5%) instead of 3.0. this is often referred to as leverage decay and is a function of time, leverage, return trend, and volatility, with volatility usually being the significant determinant. realized leverage can mathematically be expressed by eq. (1): lrt xrt � � � t ��2 � �� 2 � ��r 2�t � 1� � �r 2� xrt (1) where lrt is the return to the leveraged fund, xrt is the underlying index return, � is the daily leverage ratio, t is time in days, �r is the mean daily return, and �r 2 is the standard daily 389j. george, w.j. trainor / financial services review 26 (2017) 387–403 population variance (avellaneda and zhang, 2010; cheng and madhaven, 2009; trainor and caroll, 2013). when ��r 2�t � 1� � �r 2� is negative, the realized leverage over time will be less than the daily leverage ratio �. this effect is greater with higher leverage since a daily leverage ratio of 2x multiplies this term by one, [(22 – 2)/2], but a daily leverage ratio of 3x multiplies this term by three. lu, wang, and zhang (2012) generalized this leverage decay and suggest over holding periods no greater than one month, an investor can assume that a 2x/-2x letf will maintain its leverage ratio and provide the expected return applied to the underlying index. most research generally concludes letfs should be used for short-term trading strategies only, and the providers market them as such. however, if the trend � is high enough, trainor (2011) shows an investor can end up with a great deal more than the daily leverage ratio might indicate. a perfect example of this is proshare’s 3.0x ultrapro (upro) fund that magnifies the daily return of the s&p 500. since the fund was introduced in june of 2009, the s&p 500 increased 183% through december 2016, while the 3.0x fund increased 1,045% for an effective ratio of 5.7x. this occurred over a period with high return trend and lower than average volatility. within a portfolio setting, dilellio, hesse, and stanley (2014) suggest there may be a place for long-term holdings of letfs as their results show letfs may reduce a portfolio’s standard deviation. from an option based portfolio insurance strategy, trainor and gregory (2016) show the results of using covered calls and protective puts with letfs. both strategies reduce risk, but significant returns are often sacrificed. scott and watsun (2013) suggest a floor-leverage rule where an investor places 85% of wealth in a risk-free asset and 15% in a 3x letf. with annual rebalancing, they find this strategy can be used to manage risk and appears to be optimal for sustainable investment in retirement. this study considers the use of letfs in a cppi context. the logic for using letfs within a cppi setting is straight forward. in a standard cppi portfolio, the investor may start out with 50% in the risk-free rate and 50% in the risky asset. with a 2x letf, the investor only needs to invest 25% in the letf leaving 75% to earn the risk-free rate. if a 4.0x letf becomes available, only 12.5% is needed. if the benefits of the additional return from having a greater percentage of the portfolio in the risk-free asset exceeds letf’s decay and higher expenses, then a cppi using letfs will outperform a standard cppi strategy using the underlying index. this study determines this is indeed the case when the risk-free asset yields at least 3% or the one year minus 90-day treasury exceeds 1%. 3. methodology this study explores the benefits of letfs within a cppi format. the s&p 500 is used as the underlying index and the 1-year treasury is set as the risk-free asset. four different cppis labeled cppi s&p, cppi 2x, cppi 3x, and cppi 4x are compared using a 90% floor. although there currently are no 4x funds, both forceshares and proshares have proposed them. the 90% floor is used to account for a typical 50/50 stock/bond portfolio where an investor wants to limit a stock loss to approximately 20% over any given year. assuming no change in the value of a bond fund, a 20% loss in equities would hit a 90% floor value. 390 j. george, w.j. trainor / financial services review 26 (2017) 387–403 the proportion in the risky asset for the cppi s&p at time t is calculated as the max{min[(m(vp – f),vp],0}/vp where m is the multiplier, vp is the value of the portfolio, and f is the floor. the index multiplier is set at 5 which implies an initial 50% investment in the risky asset. following cesari and cremonini (2003), the portfolio is rebalanced only when the underlying risky-asset (the s&p 500 in this study) increases or decreases by 2.5% since the last rebalance. rebalancing once a week or once a month is also tested. for the latter, even if the market sheds 20% over the month, the floor would not be breached. this seems reasonable even for very risk averse individuals since a 20% loss in a single month has only occurred once post-wwii in october 1987 (�21%). the drawback of monthly rebalancing for letfs is the leverage decay that can be experienced over a month. with weekly or 2.5% price limits, this is less likely to be an issue. the benefit of using letfs is 50% exposure can be attained with a smaller equity position. in the case of a 2x, the multiplier only needs to be 2.5 to attain the same 50% exposure. for a 3x, the multiplier only needs to be 1.67, and for the 4x, 1.25. because it is assumed the investment in the risky asset is capped at 100%, the letfs maximum exposure must be additionally constrained. the proportion in the risky asset for cppi 2x at time t is calculated as the min{max(min[(m(vp – f),vp],0)/vp,0.5} while the proportion in the risky asset for cppi 3x at time t is min{max(min[(m(vp – f),vp],0)/vp,0.33}. for a 4x, the maximum exposure is 25%. the four initial positions using a portfolio value (vp) of $100,000 for exposition are shown in table 1 below. while table 1 shows the initial positions, table 2 demonstrates how the cppis are adjusted after each rebalance using the s&p 500 spy etf and proshare’s 3x s&p 500 table 1 initial positions for cppi strategies cppi floor (f) cushion multiplier (m) % in risky asset % in risk-free cppi s&p $90,000 $10,000 5 $50,000 $50,000 cppi 2x $90,000 $10,000 2.5 $25,000 $75,000 cppi 3x $90,000 $10,000 1.67 $16,667 $83,333 cppi 4x $90,000 $10,000 1.25 $12,500 $87,500 constant proportional portfolio insurance (cppi) positions are initially set based on a 90% floor and are rebalanced with every 2.5% move in the s&p 500. floor values are reset each year based on the portfolio value at the end of the previous year. table 2 percentage changes in the cppi strategies 2016 spy ret % in s&p cppi s&p vp 3x upro ret % in 3x cppi 3x vp 1/4-1/7 �4.82% 50.00% $0.98 �14.12% 16.67% $0.98 1/7-1/13 �2.69% 39.00% $0.97 �8.16% 13.12% $0.97 1/13-1/29 2.59% 34.26% $0.98 7.29% 11.52% $0.98 1/29-2/5 �2.98% 38.75% $0.96 �9.15% 12.99% $0.96 2/5-2/11 �2.71% 33.91% $0.96 �7.71% 11.16% $0.96 results show the returns of the underlying risky asset designated as the spy etf, the allocation to the risky asset (% in s&p, % in 3x), and the portfolio value with a start value of $1 for two of the constant proportional portfolio insurance cppi strategies (cppi s&p vp, cppi 3x vp) from 1/4/16 to 2/11/16. upro is proshare’s 3x s&p letf. 391j. george, w.j. trainor / financial services review 26 (2017) 387–403 (ticker upro) as an example. with each 2.5% change in the underlying index, the exposure is adjusted. the floor is rebalanced annually to 90% of the value of the portfolio at the end of the year. this implies an investor could be 100% in equities if the returns are positive enough. because the floor is only reset annually, more than a 10% loss could occur within the year, but not for the entire year assuming there are no historic losses in any given day. the floor could be breached if equities declined by at least 20% in any given day before the portfolio could be rebalanced. even this loss is likely not enough to breach the floor as the value of the bond portion of the portfolio would increase dramatically in such a scenario. in addition, even if a 100% equity position was held because of the increase in the portfolio value over the year, a 20% daily decline in equities would still be required to breach the 90% floor. as an example of how the rebalancing is implemented, table 2 shows the spy falls �4.8% in the first 3 days of 2016. this breaches the 2.5% limit and the portfolios are rebalanced. this results in a reduction in the risky asset from 50% of the portfolio to 39% for the cppi s&p. a similar type of reduction is made in the cppi 3x portfolio as upro’s 3x return is �14.12%. alternatively, from 1/13/16 to 1/29/16, the market increases 2.6% leading to a percentage increase in the risky asset. rebalancing occurs every time the index changes by 2.5% or more since the previous rebalancing. on average, 23 trades a year are required using a 2.5% barrier. in january of the following year, the percentages are reset to their original values with a 90% floor based on the value of the portfolio at the end of december. it should be noted the results in this study do not explicitly account for brokerage costs or for bid-ask spreads, although the latter are usually a few cents a share at most. because all the cppi strategies require equal number of transactions, the relative results between the cppi strategies are not biased, but depending on the size of the portfolio, may overstate results relative to buy-andhold. for a $100,000 portfolio, brokerage costs would be approximately 0.2% annually. because the two major letfs on the s&p 500 were not introduced until 2006 and 2009, respectively, theoretical letf returns are calculated to attain a clearer picture of the risk/return characteristics from the cppi strategies. the center for research in security price’s (crsp) s&p 500 value weighted portfolio is used as the risky asset and 2x, 3x, and 4x returns are calculated assuming a 1.2% annual expense ratio which is approximately 0.2 percentage points higher than the expense ratio for s&p 500 letfs. the reason for the higher expense ratio is explained below. following scott and watsun (2013), the daily returns for the 2x, 3x, and 4x letfs are calculated as: rl � l � rs&p � rexp � (l-1) � rf (2) where rl is the daily return to the letf with a daily leverage ratio of l, rs&p is the daily return of the s&p, rexp is the daily expense ratio, and rf is the borrowing rate using the 90-day t-bill rate as a proxy. strictly speaking, the one-week/month libor rate should be used, but libor data begins in 1986 and to remain consistent with sampled returns before this date, the 90-day t-bill rate is used. the 90-day t-bill has a 98% correlation with libor and averages 0.2% less than libor. 392 j. george, w.j. trainor / financial services review 26 (2017) 387–403 the logic behind eq. (2) is a 2x has increased exposure by borrowing $1 for every $1 invested. a 3x borrows $2 for every $1 invested. since letfs primarily attain their exposure using swaps, there are imbedded financing costs increasing with leverage (charupat and miu, 2014). in addition, there are additional transactions costs not reflected in letfs expense ratio that are generally higher for these funds because of the use of derivative contracts. to test this pricing equation, theoretical daily, monthly, and annual returns for a 2x and 3x are compared with the actual daily, monthly, and annual returns of proshare’s 2x sso and 3x upro on the s&p 500, and their 2x sqd and 3x tqqq on the nasdaq 100. to create near equivalence for daily, monthly, and annual returns between the letfs and the simulated returns, the annual expense ratio is increased from 1.0 to 1.2% which coincides with the exact amount of the average difference between the 90-day t-bill and libor. this results in average daily and monthly differences at or close to zero and average annual differences less than 0.1%. as an example of the differences, for 2016, the 2x sso return is 21.5% while the theoretical return was 21.6%. for the nasdaq 2x, both the 2x sqd and simulated 2x had returns of exactly 10.2%. based on the reliability of the results above, letf returns are calculated from jan. 1947 to dec. 2016 using daily data based on eq. (2). in addition to these results, empirical data using proshare’s 2x sso is presented for 2007–2016. finally, to determine the robustness of the results, block bootstrapping is used to resample 252 daily return windows to create 10,000 unique annual returns. the 252-day blocks are used to keep the continuity of the interest rate environments associated with the stock returns during that 252-day period, although 5, 22, and 63-day blocks are also examined with no substantial change in the average results. because the floor is reset each year, this covers thousands of historically relevant 252-day periods. the drawback of using shorter blocks is that the interest rates experienced both in terms of financing and the risk-free asset used for investing excess funds tends to average out, and does not show what may occur in sustained extreme interest rate environments. thus, sampling shorter blocks would bias the results. results are analyzed using a variety of measures. these include the average return, minimum return, maximum return, standard deviation, sharpe ratio, sortino ratio, var, es, omega ratio, and cumulative prospect values. the risk metrics are explained in appendix 1. 4. results 4.1. historical results table 3 shows the annual performance and risk measures from 1947 to 2016 for the cppi s&p, cppi 2x, cppi 3x and cppi 4x. the multipliers of 5, 2.5, 1.67, and 1.25, respectively, determine the exposure to the risky asset that is rebalanced with a 2.5% or greater move in the s&p 500 relative to the last rebalance. the remaining allocation is invested in one-year t-bills. the floor is set to 90% of the portfolio value and is reset annually. return data are provided for the respective risky assets with a 1.2% annual expense ratio and daily financing costs assumed for the letfs based on eq. (2). 393j. george, w.j. trainor / financial services review 26 (2017) 387–403 the first item to note is that all the letfs suffer decay and their returns are positively skewed. for example, the s&p 3x average annual return of 33.97% is only 2.71 times the average annual return of the s&p. the riskiness of the letfs is also apparent with extreme minimums, vars, and es values. based on sharpe ratios, there is not a huge difference between them although letf’s minimums are daunting ranging from �67% to �95%. however, there does appear to be benefits for letfs within a cppi strategy as all the letf cppis display better average annual returns over the standard s&p cppi ranging from 0.47% for the cppi 2x to 1.35% for the cppi 4x. there is some increase in the standard deviation for these higher returns, but the minimums, vars, and es are lower as reflected in their higher sortino, omega, and cpv values. from a risk-return perspective, all four cppi strategies dominate the s&p 500 based on risk metrics, but they do give up 2.46% to 1.37% annually moving from the cppi s&p to the cppi 4x. because the letfs are rebalanced after every 2.5% move, the �95% possibility of buying and holding a 4x is eliminated along with the fact the 4x can only be a maximum of 25% in the portfolio. downside protection for the leveraged cppis is confirmed as the 90% floor is never breached for any of the cppi strategies. for prospect theory type investors, the cpvs of the cppis do exceed the s&p 500 with increasing levels of cpv from cppi s&p to the cppi 4x. graphically, fig. 1 demonstrates the value of cppi strategies relative to a 100% investment in the s&p, and the value of using letfs in a cppi strategy as opposed to a standard cppi portfolio. all of the cppi strategies avoided the major drawdowns of wealth in the early 1970s, as well as in 2001 and 2008. there is a cost for cppi strategies as their cumulative values all fall short of the s&p 500. it is also the case all the letf cppis outperform the standard s&p cppi. results thus far suggest letf cppi strategies appear to outperform a standard s&p cppi strategy. however, part of the outperformance of letf cppi strategies has been the greater percentage of wealth that earns the risk-free rate. in effect, the letfs are borrowing to attain table 3 annual returns from 1947 to 2016 for the underlying indexes and cppi portfolios s&p s&p 2x s&p 3x s&p 4x cppi s&p cppi 2x cppi 3x cppi 4x average 12.53% 21.91% 33.97%b 47.55%b 10.13% 10.60% 11.15% 11.48% standard deviation 17.02% 36.11% 58.89% 87.00% 11.79% 12.23% 12.58% 12.78% median 13.81% 24.51% 33.07% 42.43% 8.18% 8.38% 8.87% 9.24% minimum �36.65% �66.82% �85.24% �94.55% �8.65% �8.64% �8.63% �8.64% maximum 52.85% 127.45% 239.66% 402.99% 46.32% 46.65% 47.54% 48.15% sharpe ratio 0.49 0.49 0.51 0.50 0.51 0.53 0.56 0.57 sortino 0.80 0.78 0.87 1.02 2.21 2.39 2.62 2.78 var 5% �15.47% �34.35% �49.77% �62.56% �6.58%b �6.64%b �6.34%b �6.31%b es �5.62% �12.87% �19.01% �24.55% �0.23%b �0.40%b 0.33%b 0.38%b omega 5.96 4.67 4.66 4.79 13.52 14.56 16.06 17.11 cpv 7.47 5.76 5.34 5.33 14.20 14.56 15.01 15.24 results show average annual returns from january 1947 to december 2016 for the s&p 500 along with three leveraged exchange traded funds (letfs) and four constant proportional portfolio insurance (cppi) strategies. 2x, 3x, and 4x cppi strategies replace the standard s&p 500 with 2x, 3x, and 4x letfs. asignificantly better than the cppi s&p at the 5% level. bsignificantly better than the s&p 500. 394 j. george, w.j. trainor / financial services review 26 (2017) 387–403 increased exposure to the index. an investor using a letf cppi strategy needs to attain additional return with the excess funds to overcome the financing costs from the leverage and the associated higher expense ratio. to ascertain how well this works in different interest rate environments, table 4 shows historical subperiods from 1947 to 1959, 1960–1978, 1979– 1991, 1992–2008, and 2009–2016. these periods correspond to average rates of 2.0%, 5.59%, 10.87%, 4.71%, and 0.39% respectively. as might be expected, when rates are low with a relatively flat yield curve as seen in 1947–1959 and especially during 2009–2016, there is little advantage to using a letf cppi with average returns less or not much greater than a traditional cppi. during more “normal” interest rate periods such as 1960–1978 and 1992–2008, letf cppis outperform a traditional cppi by 0.3% to as much as 1.5%. an interesting find was that the cppi portfolios have higher average returns than the s&p 500 during the 1960–1978 period as they avoided 0 1 2 3 4 5 6 7 8 19 46 19 49 19 52 19 55 19 58 19 61 19 64 19 67 19 70 19 73 19 76 19 79 19 82 19 85 19 88 19 91 19 94 19 97 20 00 20 03 20 06 20 09 20 12 20 15 cumula�ve ln growth of $1 from 1947 to 2016 s&p s&p cppi cppi 2x cppi 3x cppi 4x fig. 1. cumulative ln growth of $1 for the four cppi strategies and the s&p 500 from 1947 to 2016. table 4 average annual returns for different interest rate environments during 1947–2016 time period 1-year t-bill s&p cppi s&p cppi 2x cppi 3x cppi 4x 1947–2016 5.10% 12.53% 10.13% 10.60% 11.15% 11.48% 1947–1959 2.00% 18.30% 13.57% 13.38% 13.78% 14.06% 1960–1978 5.59% 7.51% 7.94% 8.50% 9.11% 9.51% 1979–1991 10.87% 17.56% 14.96% 16.72% 17.80% 18.40% 1992–2008 4.71% 8.81% 7.93% 8.20% 8.60% 8.82% 2009–2016 0.39% 14.81% 6.59% 6.20% 6.36% 6.41% results show average annual returns from january 1947 to december 2016 for the s&p 500 along with four constant proportional portfolio insurance (cppi) strategies for various sub-periods corresponding to changing interest rate environments. 395j. george, w.j. trainor / financial services review 26 (2017) 387–403 large losses and gave up little relative performance by investing in one-year t-bills. letf cppi’s performed best during the 1979–91 period when the one-year t-bill return averaged 10.87%. although this was much less than the s&p’s 17.56% average, the 3x and 4x cppi still outperformed the s&p itself. relative to the standard s&p cppi, all the letf cppis outperformed by 1.76% to 3.44%. thus, during low interest rate or flat yield curve environments, there appears to be little advantage of creating letf cppi portfolios. in average or high interest rate environments, they outperform a standard cppi. with letf etfs not being introduced until mid-2006, the results using actual letfs are unlikely to be favorable relative to a traditional cppi since the federal reserve has followed a near zero interest rate policy since the 2008 financial crisis. however, it is informative to examine how a letf cppi performs in practice. table 5 shows the returns from 2007 through 2016 using the spy for the s&p 500 and proshares 2x sso. although not shown, proshares 3x upro has similar results. both cppi strategies perform as advertised as downside risk is mitigated and the �10% floor is not breached even in 2008 when the s&p fell 37%. both cppi strategies track each other, but the s&p cppi strategy outperforms the cppi 2x every year with the exception of near equal results in 2008. these results are mainly because of the fact the 2x still must deal with decay along with higher expenses and cannot make up the difference with additional earnings on the risk-free rate. these results reinforce the conclusions from table 4 when rates are low. in addition, when the risk-free rate was still relatively high in 2007 and 2008, the sso 2x still underperforms because of the high volatility during that period causing serious return decay. in fact, for 2007, the 2x sso only returns 1.04% compared with spy’s 5.14%, although half of the 4% underperformance is because of the large tracking error this fund had when first introduced. this tracking error has not been observed since. furthermore, 2008 had the next largest tracking error when sso returned �67.94% relative to the predicted �67.06% from eq. (2). this tracking error has continually fallen since then and over the last 3 years has been less than 0.1% on an annual basis. table 5 cppi annual returns based on spy and proshares 2x sso letf from 2007–2016 year 1-year t-bill spy sso 2x cppi s&p cppi 2x cppi s&p w/agg cppi 2x w/agg 2007 5.94% 5.14% 1.04% 2.45% 2.26% 2.92% 3.23% 2008 4.60% �36.81% �67.94% �8.62% �8.60% �2.94% �2.51% 2009 0.56% 26.37% 47.26% 8.25% 8.13% 9.70% 10.58% 2010 0.55% 15.06% 26.84% 5.81% 5.66% 11.37% 13.70% 2011 0.37% 1.89% �2.92% �3.40% �3.58% �1.06% 0.40% 2012 0.16% 15.99% 31.04% 8.04% 7.64% 9.33% 10.68% 2013 0.34% 32.31% 70.47% 23.66% 22.92% 23.10% 20.34% 2014 0.24% 13.46% 25.53% 6.25% 5.81% 9.77% 12.08% 2015 0.16% 1.25% �1.19% �1.49% �1.79% �1.16% �1.36% 2016 0.76% 12.00% 21.55% 6.22% 6.03% 9.41% 9.89% geo. ann. ret. 1.35% 6.87% 7.06% 4.40% 4.15% 6.79% 7.47% results show annual returns from 2007 to 2016 for the s&p 500 spy, the s&p 500 2x (sso), along with constant proportional portfolio insurance (cppi) strategies using the s&p and s&p 2x. the last two columns replace the one-year t-bill with ishares agg aggregate bond portfolio within the cppi portfolios. 396 j. george, w.j. trainor / financial services review 26 (2017) 387–403 it is also interesting to note that both cppi strategies lost money in 2011 and 2015 even though the spy was slightly positive, 1.89% and 1.15%, respectively. this occurs because cppi strategies are reactionary. when the market falls, the percentage of the portfolio in equities is reduced. when it bounces back, there is less wealth in the portfolio to regain the losses. similarly, if the market increases (more is allocated to equities) then decreases. thus, cppi strategies tend to do poorly in volatile flat markets. the same holds true for letfs because of their daily rebalancing. in fact, the sso 2x lost money both in 2011 and 2015 despite the market increasing. this is a perfect demonstration of letf’s return decay. for the less risk averse, table 5 shows one alternative to overcoming extremely low interest rates. the last two columns show cppi returns by combining a composite bond etf within a cppi strategy. instead of using the one-year treasury from 2007 to 2016, ishares agg bond etf is used as a proxy for a relatively safe asset with higher expected returns. both cppi strategies show improvement and the letf cppi outperforms the standard cppi except for 2013 and is only slightly worse in 2015. in 2013, both cppi strategies using the ishares bond etf underperform cppi strategies using the one-year treasury. this is because of the �1.98% return the ishares bond etf experienced that year. it was the only negative year for the aggregate bond fund, but highlights the fact that increasing expected return, even with a relatively safe bond fund, does have risks. 4.2. bootstrapped results to check on the robustness of the data, the historical data are blocked bootstrapped by sampling 252-day trading periods to create 10,000 annual returns to remove any bias from the january to january annual returns that may have favored one method or another. in addition, this will hopefully encompass most future possibilities while maintaining the relationship between interest rates and returns along with systemic low or high interest rate environments that may be experienced. table 6 shows the results. table 6 bootstrapped annual returns for the underlying indexes and cppi portfolios s&p s&p 2x s&p 3x s&p 4x cppi s&p cppi 2x cppi 3x cppi 4x average 12.31% 21.60%b 33.30%b 46.41%b 10.07% 10.59% 11.09%a 11.39%a standard deviation 16.72% 35.68% 57.99% 84.45% 11.88% 12.40% 12.74% 12.93% median 12.97% 21.22% 29.35% 35.74% 8.11% 8.36% 8.81% 9.06% minimum �46.99% �77.09% �91.77% �97.60% �9.79% �9.79% �9.79% �9.79% maximum 71.24% 175.88% 327.36% 530.57% 60.62% 66.33% 68.88% 70.19% sharpe ratio 0.49 0.49 0.50 0.50 0.50 0.52ab 0.54ab 0.56ab sortino 0.95 1.08 1.29 1.50 1.37 1.51 1.66 1.75 var 5% �15.63% �34.94% �53.34% �70.40% �5.51%b �5.38%b �5.26%ab �5.17%ab es �25.53% �50.09% �68.56% �82.70% �7.40%b �7.32%b �7.25%b �7.20%b omega 6.18 4.75 4.73 4.86 14.10 15.18 16.40 17.17 cpv 8.70 9.43 8.71 7.17 14.83 15.48 15.98 16.28 results show annual statistics based on 10,000 bootstrapped simulations for the s&p 500 along with three leveraged exchange traded funds (letfs) and four constant proportional portfolio insurance (cppi) strategies. asignificantly better than the cppi s&p at the 5% level. bsignificantly better than the s&p 500. 397j. george, w.j. trainor / financial services review 26 (2017) 387–403 the returns and risk statistics confirm the earlier results with the letf cppi portfolios showing better returns and risk metrics relative to the standard cppi based on sortino, value at risk (var), expected shortfall (es), omega, and cpv. although there are no statistical tests for significance for the sortino and omega values, the differences are monotonically increasing. the sortino measure also suggests the 4x by itself has better return relative to downside risk even over the cppi s&p, but this is mainly because of the very positively skewed returns for this asset resulting in a high average return. the median return shows the more likely result and is significantly less than the average for the 4x fund. reconfirming the historical results, all four cppi portfolios have better cpv values relative to the s&p 500, with the letf cpvs greater than the cppi s&p as well. both the cppi 3x and cppi 4x have significantly greater returns than the cppi s&p. the cppi 4x gives the best results despite using the riskiest asset. the caveat to the cppi 4x results is there are no 4x letfs currently in existence, and when and if they do make it to market, the expense ratio and cost of running these funds may be more expensive than what is assumed in this study. with no way to test the pricing model on an actual 4x fund, the 4x results should be interpreted with caution. from an absolute return standpoint, cppi portfolios have annual average returns 1% to 2% less and median returns 4% to 5% less than the s&p 500. thus, cppi portfolios do not provide “free” downside protection. however, the amount of average return sacrificed to avoid large losses would seemingly be appealing to risk-averse investors. the reduction in the minimums, var, and es values bear this out along with much higher cpv values that measure the value to prospect theory type investors. for the risk-speculators, the average returns to the letfs by themselves are enticing, up to an average 46% return with the 4x. however, these returns are coupled with up to �98% losses in any given year. the question going forward is which cppi strategy is likely to do best? noting what has happened since 2007 when the t-bill rate is close to zero, a letf cppi loses its advantage if the additional funds are simply invested in the risk-free rate if rates are very low. to ascertain what risk-free rate of return one would need to overcome decay and the higher expenses from using a letf, 40,000 bootstrapped annual returns are sorted based on the average t-bill return attained each year. the average t-bill return was 5.13% with a standard deviation of 3.84%. the top section of table 7 shows the return data for when the t-bill rate is below 1% to greater than 7%. two conclusions are apparent from examining the top section of table 7. for a letf cppi to outperform, a rate of 2.5% or more is required to make up for the additional costs from using letfs. two, when the rate is less than 2%, for those constructing a cppi, a standard cppi will be slightly better. alternatively, an investor could redefine the “risk-free” asset to accommodate a relatively “risk-free” bond fund providing higher yield. interest rates above 6% are very favorable to letf cppis with returns up to 3% greater over the standard cppi. in addition, rates over 7% seem to favor any cppi, even over the s&p itself. they are particularly favorable to letf cppis with average returns 1.5% to 3% greater than both the s&p and the cppi s&p. by using letfs in a cppi, the investor is in effect borrowing short via the letf and investing in one-year treasuries. thus, the slope of the yield curve is a critical issue. generally, lower rates are associated with a flatter yield curve and vice versa. thus, returns 398 j. george, w.j. trainor / financial services review 26 (2017) 387–403 are also sorted by the yield curve slope as measured by the one year minus 90-day treasury. the bottom of table 7 shows the results. when the 90-day treasury exceeds the return from the one-year treasury, the standard cppi outperforms a cppi 2x by 0.7% to 0.4% moving from a cppi 2x to a cppi 4x. as soon as the slope exceeds 1%, all the letf cppis outperform a tradition cppi and this outperformance increases the greater the difference in treasury rates. these results reaffirm what may be obvious; the advantage of using a letf cppi strategy is what return an investor can attain with the excess funds relative to the intrinsic financing costs of the letfs leverage. 4.3. rebalancing rules there is nothing magic about using a 2.5% market move to rebalance. in fact, using a 2% or 3% market move gives virtually the same results. taken to the extreme, daily rebalancing could be implemented, but trading costs would increase. although this study did not account for brokerage costs explicitly in the return calculations, with two trades a day at $5 a trade, even a $1,000,000 account faces an additional 0.25% in trading cost if using daily rebalancing. table 8 shows the average returns from 10,000 bootstrapped simulations for the cppi strategies based on the 2%, 2.5%, and 3.0% rule along with daily, weekly, and monthly rebalancing. weekly rebalancing is not significantly different from the 2.5% rule, nor is monthly worse suggesting the decay drag, even over a month, is not significant confirming lu, wang, and zhang (2012) findings that investors can generally expect to earn the leverage ratio for up to a month. daily rebalancing does improve results, but not by enough to overcome transaction costs. with the exception of daily rebalancing, the average returns for the cppi s&p varied from 10.01% to 10.05%. the same type of range held for the letf cppis. table 7 bootstrapped annual returns sorted by 1-year t-bill return and slope of the yield curve s&p s&p 2x s&p 3x s&p 4x cppi s&p cppi 2x cppi 3x cppi 4x t-bill return 0–1% 17.41% 34.46%b 53.96%b 74.70%b 10.21% 9.81% 9.99% 10.09% 1–2% 17.92% 35.14%b 55.24%b 77.22%b 11.31% 11.05% 11.27% 11.39% 2–3% 6.62% 11.54%b 19.84%b 31.19%b 6.06% 6.03% 6.28% 6.44%a 3–4% 6.87% 11.64%b 19.02%b 28.32%b 6.36% 6.60% 6.97%a 7.22%a 4–5% 8.39% 13.59%b 20.55%b 28.08%b 7.05% 7.38% 7.80%a 8.06%a 5–6% 8.62% 12.45%b 16.97%b 20.99%b 6.92% 7.37% 7.84%a 8.15%a 6–7% 13.49% 23.08%b 34.69%b 47.07%b 10.90% 11.65% 12.26%a 12.62%a �7% 13.44% 21.32%b 31.42%b 42.69%b 13.79% 15.43%ab 16.35%ab 16.87%ab 1 year to 90-day �0% 23.47% 47.97%b 79.43%b 117.76%b 17.03% 16.32% 16.51% 16.65% 0–1% 12.55% 23.08%b 35.23%b 47.77%b 7.61% 7.42% 7.64% 7.76% 1–2% 10.39% 17.51%b 26.64%b 36.83%b 8.18% 8.43% 8.83%a 9.08%a 2–3% 10.48% 17.26%b 26.23%b 36.41%b 9.41% 10.01% 10.54%a 10.86%a �3% 12.02% 19.40%b 29.07%b 40.00%b 12.51% 14.12%ab 15.02%ab 15.54%ab results show average annual returns based on 40,000 bootstrapped simulations sorted by 1-year t-bill returns and the slope of the yield curve measured by the 1-year minus 90-day t-bill return. asignificantly different from constant proportional portfolio insurance cppi s&p at the 5% level. bsignificantly better than the s&p 500. 399j. george, w.j. trainor / financial services review 26 (2017) 387–403 in summary, the additional return from using a letf cppi strategy relative to just using the index appears robust to the rebalancing method chosen with a cppi 2x, cppi 3x, and a cppi 4x earning approximately 0.5%, 1.0% and 1.3% more, respectively. 5. conclusion letfs are proclaimed to be risky-short term trading vehicles with plenty of warnings, (carver, 2009; justice, 2009; zweig, 2009, 2017). as an individual asset, there is no denying the extremes that can be experienced by buy-and-hold investors. however, more active traders can moderate this risk by periodic rebalancing which fits in perfectly with a cppi strategy. by managing the exposure as letfs change in value, downside losses can be mitigated. one of the disadvantages of a cppi is the need for constant rebalancing. however, with only periodic rebalancing based on market movements, this study shows using letfs instead of the underlying risky asset in a cppi portfolio leads to greater returns with less risk. this outcome is possible because the same amount of exposure to the risky asset can be attained with a smaller percentage of the portfolio, leaving a larger amount available to earn the risk-free rate. if the return from the additional amount in the risk-free asset exceeds the letf decay and higher expenses, the letf cppi will outperform. results suggest risk-free yields of 3% or if the one-year exceeds the 90-day treasury by more than 1% appear to be sufficient for letf cppis to outperform a standard cppi using the index itself. both simulated results from 1947 to 2016 and bootstrapped data show cppi strategies created with letfs outperform a cppi strategy using the underlying index. average annual returns over all interest rate environments are 0.5% to 1.3% higher with better minimums, sharpe ratios, sortino ratios, omegas, vars, es, and cpvs. using letfs in a cppi strategy will underperform slightly when the risk-free rate is extremely low as it has been for the last seven years with the yield below 1%. there is simply no additional return from having a greater percentage of wealth in the risk-free rate to compensate for letfs higher expenses and leverage decay. table 8 average annual returns for different rebalancing rules rebalance avg. no. of trades s&p cppi s&p cppi 2x cppi 3x cppi 4x 2% rule 31 12.33% 10.04% 10.53% 11.01%a 11.28%a 2.5% rule 23 12.31% 10.01% 10.53% 11.03%a 11.32%a 3.0% rule 17 12.35% 10.05% 10.52% 10.98%a 11.24%a daily 252 12.30% 10.27% 10.74% 11.17%a 11.39%a weekly 50 12.28% 10.03% 10.53% 11.00%a 11.28%a monthly 12 12.34% 10.01% 10.50% 10.95%a 11.22%a results show average annual returns based on 10,000 bootstrapped simulations for the s&p 500 along with four constant proportional portfolio insurance (cppi) strategies based on various rebalancing rules. percentage rules based on absolute return of the s&p 500 relative to previous rebalance. asignificantly better than the cppi s&p at the 5% level. 400 j. george, w.j. trainor / financial services review 26 (2017) 387–403 for the less risk averse, more aggressive cppi portfolios could be created by using a smaller multiplier, lower floors, letting the amount in the 2x, 3x, or 4x letfs exceed 50%, 33%, or 25%, respectively, and/or using a more risky “risk-free” asset such as longer-term treasuries or some type of composite bond etf as demonstrated in the last two columns of table 5 where a bond etf was substituted for the one-year treasury. the latter adjustment is likely the safest when yields are extremely low, while allowing the amount in the letf to increase up to 100% is certainly the riskiest. two days like october 16 and october 19, 1987 when the market fell 5.1% then 19.5% would see a cppi 4x lose 84% of its value if exposure to a 4x letf is allowed to increase to 100%. this would seem to defeat the purpose of a cppi strategy in the first place. letfs are relatively new instruments. the reception from investors has been mostly positive despite their risk as seen by the phenomenal growth both in number and in asset growth. like options and futures, letfs can be used for highly speculative gambles, hedging, or risk management. this study demonstrates that letfs, even a 4x letf if it becomes available, can be used to enhance return, and reduce risk within the right context. appendix 1 the risk metrics used to evaluate the results are described below: 1. the sharpe ratio is the excess return divided by the standard deviation and is reported for completeness, even though it is not the best measure for assessing the downside risk portfolio insurance portfolios are attempting to mitigate, (sharpe, 1964). opdyke’s (2007) testing procedure is used to determine whether the letfs cppi sharpe ratios are better than the s&p 500 and the cppi s&p. 2. the sortino ratio is a modification of the sharpe ratio and only considers the downside deviation removing the aspect of “good volatility” (sortino and price, 1994). it is more appropriate for analyzing portfolio insurance strategies that are designed to mitigate large losses and whose returns may not be normally distributed. the sortino ratio is written as: s � r � t tdd where tdd � �1 n � i�1 n �min(0,xi � t))2 (3) where r is the return, t is the target return, n is the total number of returns and xi is the ith return. the higher the ratio, the greater the return per unit of downside risk. t is set at the one-year treasury return for this study. 3. the omega ratio, reported by keating and shadwick (2002) also measures downside risk. omega is the sum of the returns above a certain threshold divided by the sum of the returns below that threshold, which is also set at zero in this study. the omega ratio is written as, 401j. george, w.j. trainor / financial services review 26 (2017) 387–403 �x��� � � � � �1 � fx� x�dx � � � fx� x�dx (4) where f is the cumulative distribution function and � is the threshold return. in this study, it is simply the sum of the returns above zero divided by the absolute value of the sum of the returns below zero. 4. value at risk measures the expected maximum loss with a given confidence level over a specific time; 5% of the observations are less than the var. to test for differences in var, an unconditional coverage test is applied as put forth in annert, osselaer, and verstraete (2009) and is written as: � 1 n � n 1 n hitn � �� ���1 � �� n (5) where hitn equals one if the letf cppi return is lower than the s&p cppi var, zero otherwise, n is the number of returns, and � is the 5% var. all statistical tests for significant differences are set at the 95% confidence level. 5. as pointed out in acerbi, nordio, and sirtori (2001), excess shortfall (es) shows the average loss beyond the var threshold and represents the severity of a dramatic loss. it addresses the “what if” factor and makes up for the discrepancies with the var calculation. acerbi and tasche (2002) confirm the appropriateness of this definition compared with other shortfall calculations. annaert et al. (2009) is followed to test for differences in es values. 6. because there is an expectation portfolio insurance appeals more to prospect theory investors, cumulative prospect values (cpv) are calculated using the function and parameters set forth in tversky and kahnemen (1992). the probability weighting parameter for gains is 0.61 and 0.69 for losses. dichtl and drobetz (2011b) use a similar methodology in comparing dollar cost averaging to lump sum investing. references acerbi, c., nordio, c., & sirtori, c. (2001). expected shortfall as a tool for financial management. (available at https://www.researchgate.net/publication/1825314_expected_shortfall_as_a_tool_for_financial_risk_ management) acerbi, c., & tasche, d. (2002). on the coherence of expected shortfall. journal of banking and finance, 26, 1487–1503. 402 j. george, w.j. trainor / financial services review 26 (2017) 387–403 annaert, j., osselaer, s., & verstraete, b. (2009). performance evaluation of portfolio insurance strategies using stochastic dominance criteria. journal of banking and finance, 33, 272–280. avenllaneda, m., & zhang, s. (2010). path-dependence of leverage etf returns. society for industrial and applied mathematics, 1, 586–603. bertrand, p., & prigent, j. (2011). omega performance measure and portfolio insurance. journal of banking and finance, 35, 1811–1823. black, f., & jones, r. (1987). simplifying portfolio insurance. journal of portfolio management, 14, 48–51. carver, a. (2009). do leveraged and inverse etfs converge to zero? the journal of derivatives, 2009, 144–149. cesari, r., & cremonini, d. (2003). benchmarking, portfolio insurance and technical analysis: a monte carlo comparison of dynamic strategies of asset allocations. journal of economic dynamics and control, 27, 987–1011. charupat, n., & miu, p. (2014). a new method to measure the performance of leveraged exchange-traded funds. financial review, 49, 735–763. cheng, m., & madhavan, a. (2009). the dynamics of leveraged and inverse exchange-traded funds. journal of investment management, 7, 43–62. dichtl, h., & drobetz, w. (2011). portfolio insurance and prospect theory investors: popularity and optimal design of capital protected products. journal of banking and finance, 35, 1683–1697. dichtl, h., & drobetz, w. (2011b). dollar-cost averaging and prospect theory investors: an explanation for a popular investment strategy. journal of behavioral finance, 12, 41–52. dilellio, j., hesse, r., & stanley, d. (2014). portfolio performance with inverse and leveraged etfs. financial services review, 23, 123–149. hunnicutt, t., & mccrank, j. (2017). more quadruple-leveraged etfs proposed despite sec review. reuters. (available at http://www.reuters.com/article/us-sec-etfs-iduskbn19b2dp) justice, p. (2009). warning: leveraged and inverse etfs kill portfolios. morningstar investing specialists. (available at http://news.morningstar.com/articlenet/article.aspx?id�271892) kahneman, d., & tversky, a. (1979). prospect theory of decisions under risk. econometrica, 47, 263–291. keating, c., & shadwick, w. (2002). a universal performance measure. london: the finance development centre. leland, h. e., & rubinstein, m. (1976). the evolution of portfolio insurance. in d. l. luskin (ed.), portfolio insurance: a guide to dynamic hedging. new york, ny: wiley. lu, l., wang, j., & zhang, g. (2012). long term performance of leveraged etfs. financial services review, 21, 63–80. maalej, h., & prigent, j. (2016). on the stochastic dominance of portfolio insurance strategies. journal of mathematical finance, 6, 14–27. opdyke, j. (2007). comparing sharpe ratios: so where are the p-values? journal of asset management, 8, 308–336. pezier, j., & scheller, j. (2013). best portfolio insurance for long-term investment strategies in realistic conditions. insurance: mathematics and economics, 52, 263–274. scott, j., & watsun, j. g. (2013). the floor-leverage rule for retirement. financial analysts journal, 69, 45–60. sharpe, w. (1964). capital asset prices: a theory of market equilibrium. journal of finance, 19, 425–442. sortino, f. a., & price, l. n. (1994). performance measurement in a downside risk framework. the journal of investing, 3, 59–64. trainor, w., & baryla, e. (2008). leveraged etfs: a risky double that doesn’t multiply by two. journal of financial planning, 21, 48–55. trainor, w. (2011). solving the leveraged etf compounding problem. journal of index investing, 1, 66–74. trainor, w., & carroll, m. (2013). forecasting holding periods for leveraged etfs using decay thresholds. journal of financial studies and research, 2013, 179–190. trainor, w., & gregory, r. (2016). leveraged etf option strategies. managerial finance, 42, 438–448. tversky, a., & kahneman, d. (1992). advances in prospect theory: cumulative representation of uncertainty. journal of risk and uncertainty, 5, 297–323. zieling, d., mahayni, a., & balder, s. (2014). performance evaluation of optimized portfolio insurance strategies. journal of banking & finance, 43, 212–225. zweig, j. (2009). how managing risk with etfs can backfire. wall street journal, feb. 27. zweig, j. (2017). are you really crazy enough to buy a quadruple-leveraged etf? wall street journal, may 19. 403j. george, w.j. trainor / financial services review 26 (2017) 387–403 are “fun” sources of windfalls destined to be spent hedonistically? eugene blanda*, valrie chambersb atexas a&m university – corpus christi, 6300 ocean drive, corpus christi, tx 78412, usa bstetson university, 421 n woodland boulevard unit 8398, deland, fl 32723, usa abstract recently, richard thaler was awarded a nobel prize for his work in developing behavioral economics. while much of economics assumes that people act rationally, areily (2008), expanding on thaler’s body of work, proves that we are not only often irrational, but we are predictably irrational. when an interviewer asked thaler how he would spend the roughly $1.1 million in prize money, he responded, “this is quite a funny question.” thaler added, “i will try to spend it as irrationally as possible.” we know that affective tags for money exist but what specifically are those affective tags? more specifically still, is one of those tags for sources of income “fun,” and if so, does that affect whether the money will be spent on fun? classical economics would assume that satisfaction comes from the consumption of goods and services, that money is a medium of exchange, and that the source of that medium of exchange does not enter into the choice of the goods or services consumed. thaler’s (1999) works show that people create mental accounts, indicating that the source of the money may not be as completely irrelevant as classical economics predicted. this is important because where irrational behavior is suboptimal behavior, if we can anticipate it, we can construct environments to support better choices. we find that fun sources of income are significantly more likely to be spent on fun expenditures. however, as the amount of the windfall increases, the amount spent on fun levels off, indicating that this affect may be bounded. we were unable to find statistically significant support that more “adult” sources of income are more likely to be spent on more adult uses, but money from adult sources was significantly more likely to be invested. this is important because understanding more about affective tags and how they affect decisions to use money, we become better predictors of irrational behavior. © 2020 academy of financial services. all rights reserved. jel classification: g40; d140 keywords: income; income source; behavioral finance; consumer behavior; mental accounting * corresponding author. tel.: �1-361-825-2829; fax: �1-361-825-5609. e-mail address: eugene.bland@tamucc.edu (e. bland) financial services review 28 (2020) 17-34 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. 1. introduction economic researchers have traditionally assumed that people’s behavior is rational. classical economics would assume that satisfaction comes from the consumption of goods and services, that money is a medium of exchange, and that the source of that medium of exchange does not enter into the choice of the goods or services consumed. dan ariely’s (2008) work asserts however that we are predictably irrational. however, while deviations from rational behavior abound, behavioral economics is relatively new and there is much to learn about how rationality is bounded and how people make financial decisions. frederick (2005) found that the more cognitive reflection that occurred, the less nonrational behavior occurred. thaler’s (1999) work shows that people create mental accounts, indicating that the source of the money may not be as completely irrelevant as classical economics predicted. that is, there is much to be learned about how rationality and irrationality interact. specifically, we know that affective tags for money exist (bradford, 2008; henderson and peterson, 1992; levav and mcgraw, 2009; winkelmann et al., 2011,) but we do not know what specifically those affective tags are, nor do we know the strength of those tags. more specifically still, one of those affective tags for sources of income may be “fun,” and if so, that may affect how the money will be spent. there may be a direct connection between the affective tag on income and the affective tag on the disposition of that income. this connection has yet to be directly studied. by understanding more about what affective tags are and how affective tags on income affect decisions to use money, we become better predictors of people’s economic behavior. such understandings in aggregate can help lead people to create systems whereby they make better financial choices, which in turn reduces their financial stress and increases their wealth and quality of life. this research is important as part of a broader field of research because where irrational behavior is suboptimal behavior, if we can anticipate it, we can construct environments to support better choices. alternatively, by understanding where irrational behavior may occur, policymakers may be able to provide alternative choices that may result in better outcomes. this research is important individually because as individuals better understand their own behavior, they can reflect upon it and adjust their behavior to better achieve their goals. for example, if a person knows that she generally receive a tax refund and she is prone to use such a windfall for adult purposes, she can incorporate a savings plan directly attached to that refund that will build her wealth more quickly than she currently does. 2. literature review according to thaler’s (1999) mental accounting theory, people create different mental accounts like long-term savings and have different marginal propensities to consume from each account. academic literature supports mental accounting theory from a regular income flow or from an irregular, lump-sum, windfall (adamopoulou and zizza, 2017; johnson et al., 2006; o’curry, 1999; souleles, 2002), and supports the periodic reconciliation of 18 e. bland, v. chambers / financial services review 28 (2020) 17-34 people’s mental accounts for income and expense (camerer et al., 1997; heath and soll, 1996; read et al., 1999; rizzo and zeckhauser, 2003). characteristics of the use of mental accounts have been studied by karlsson et al. (1999), who reported that cash spending on a durable good depended on compatible reasons for saving. abeler and marklein (2016) and benjamin (2006) found that math aptitude affects mental budgeting, and cheema and soman (2006) and wertenbroch (2001), concluded that mental budgeting was a matter of self-control. arkes et al. (1994) found that a greater percentage of a small windfall was spent than that from the same amount of anticipated ordinary income, indicating that foreknowledge of income is a factor in saving, supporting the findings of rucker (1984) and consistent with the findings of karlsson et al. (1999). trump et al. (2015) found that individuals made riskier choices with a stranger’s money than with a friend’s money. whether income was earned affected how compliant taxpayers were after a tax audit (boylan, 2010), and whether earned income was a windfall or restores a status quo was found to be significant (agarwal and qian, 2014; epley and gneezy, 2007). the size of the income can also be significant. rucker (1984) studied the retroactive payment of a raise approved by a university, reversed by the federal pay board but then reinstated by the u.s. supreme court and found that the size of the windfall was the most important factor for deciding how the funds were used, although the length of time that the recipient had to anticipate the income was also significant. chambers et al. (2009) studied responses to small hypothetical tax rebates and found that at some amount over $600, materiality was significant in how the money would be used. karlsson et al. (1999) noted that individuals considered the future consequences of spending in their mental budgeting, indicating that the permanence of the income might be significant. friedman’s (1957) permanent income hypothesis says that people will spend money consistent with their perceived permanent income level. similarly, blinder (1981) posited that a permanent tax decrease would elicit more spending than a temporary tax cut. parker (1999) found that a temporary, end-of-year reduction in the social security tax for high-income wage earners was spent when received, not averaged evenly over the fiscal year. hsieh (2003) studied large, regular bonuses associated with the annual alaska permanent fund payment, which was fully anticipated and found no spike in consumption. however, consumption by the same households was very responsive to income tax refunds, suggesting that predictable and regular payments are built into consumption decisions (hsieh, 2003). browning and collado (2001) studied spanish panel data to measure the effect of customary, predictable bonus payments and like hsieh (2003), did not find changes in consumption. in contrast, studies of the spending from nonrecurring, nonpermanent sources of income are fairly rare. bodkin (1959) estimated the marginal propensity to consume from a one-time dividend paid in 1950 to world war ii veterans by the national service life insurance to be between 0.72 and 0.97. kreinin (1961) analyzed the consumption of a sample of israeli citizens receiving restitution payments from germany in 1957 and 1958 and estimated that 35% was spent. shapiro and slemrod (1995) found that almost half of the respondents receiving decreased periodic tax withholding refunds in 1992 would spend them, even though the total yearly tax liability remained unchanged, resulting in a lower end-of year tax refund. however, in 2001, when a tax cut took the form of a lump-sum rebate, only about one-fourth of the respondents surveyed expected to spend the payment (shapiro and 19e. bland, v. chambers / financial services review 28 (2020) 17-34 slemrod, 2003). chambers and spencer (2008) found that the timing of payments (whether paid as a lump-sum or spread out in equal monthly installments for a year) was significant, and sahm et al. (2012), confirmed that finding. however, whether people were in the habit of saving versus spending also mattered (spencer and chambers, 2012). the framing of payments seems to matter: baker et al. (2007) found that more money was spent from likely recurring income (dividends) than from less recurring capital gain income. hershfield et al. (2015) found that consumers placed savings and debt into different mental accounts, making them insensitive to the significant differences between the interest rates on these accounts. shefrin and thaler (1988) found that more of a lump sum bonus was saved than if the same amount increased regular income, even when the bonus is fully anticipated. (1) is the source of the income important in mental accounting? windfall sources in prior literature include: inheritance (baker and nofsinger, 2002; zagorsky, 2013), bonus (henderson and peterson, 1992; hsieh, 2003), tax rebate (chambers and spencer, 2008; meekin et al., 2015), and lottery (winkelmann et al., 2011). some evidence suggests that the source of one’s income is important in mental accounting. winkelmann et al. (2011) found that spending from different sources of income conferred different marginal utilities. sources of income may be tied to uses of income. for example, henderson and peterson (1992) reported that individuals were more likely to spend $2,000 on a vacation if the source of the funds was a gift rather than a work bonus. bradford (2008) found that individuals allocate gifted and inherited assets consistent with their goals in the relationship. epley et al. (2006) found that people spent more from an income source of the same amount and timing labeled “bonus” than they did one labeled “rebate.” milkman and beshears (2009) found that consumers who received $10 windfalls in the form of grocery coupons spent an additional $1.59 on groceries that the consumer did not typically buy. chambers et al. (2017) found that people given a hypothetical payment from one of five different sources would spend the funds differently, depending on the source of the money, and that less of the windfall would be saved overall from a game show payment than from a tax rebate. similarly, chambers et al. (2017) found that people tended not to shift away from their spending habits. the goal of this article, given that money is fungible, is to test to see whether the affective tag of the spending significantly mirrors the affective tag of the windfall source. (2) affective tags and mental accounting levav and mcgraw (2009) proposed that windfalls in mental accounting may have a feeling attached to that sum of money, or “affective tag.” they found that when a windfall that is negatively tagged is received, the associated negative feelings influenced respondents to consume the windfall either reluctantly or virtuously to cope with those negative feelings. o’curry and strahilevitz (2001) found that those receiving lotto payments spent it hedonistically. this study focuses on one of those questions: does an income source affectively tagged as “fun” result in significantly more spending on fun? will less fun sources be used 20 e. bland, v. chambers / financial services review 28 (2020) 17-34 more for less fun uses? additionally, as a research question, how is the spending on fun bounded, if at all? (3) demographic factors several demographic factors might be significant. chen and volpe (2002) found that gender was significant to personal finance, but education and experience could have a significant impact on the financial literacy of both genders. fisher et al. (2015) found that income, income uncertainty, wealth, high-risk tolerance, and savings also differed significantly by gender, as did being nonwhite and having other household members. fisher (2010) found that certain race differences in savings were explained by the individual determinants of saving, including: receiving government assistance, feeling that credit use is bad, being turned down for credit in the past five years, or having a shorter saving horizon; race also significantly affected risk tolerance. 3. hypothesis and research questions this study examines whether people spend the same proportion of a distribution on fun categories when the windfall source is a fun source, like from a gameshow, as they do when the source is less fun, like from a tax rebate. this study examines the spending from tax rebates, game show winnings and bonuses, which might be a more neutral benchmark. only windfall earnings will be explored, as literature indicates that amounts spent from windfall income is spent differently from one’s regular income (arkes et al., 1994; karlsson et al., 1999). how might the recipient consider these sources as similar or different? tax systems are run by a government or its appointed agency and are largely outside the respondent’s control, whereas bonuses and game show winnings are generally run by private enterprises and may have more elements of respondent’s control. to what extent the money is “earned” is debatable, but bonuses and game show winnings require some personal skill, knowledge, and effort. tax rebates sometimes differ from the other sources of payment because the tax rebate is a return of the taxpayer’s withholdings. that is, outside of refundable credits tied to specific performance, respondents generally cannot materially profit from a tax rebate because it is a return of the taxpayer’s own money already paid in but can profit from a game show winning or bonus. some political rhetoric frames taxes as money belonging fundamentally to taxpayers, not the government, whereas bonuses and game show winnings come with less of an entitlement. bonuses are likely to be closely tied to an individual’s performance, however. game show winnings might be as well, if the winner attributes success to having a higher skill level than fellow contestants. in addition, collecting a bonus or a tax rebate may be repeatable. one could not count on or control repeating a game show winning. additional differences in affective “euphoria” surround these payments. it is unlikely that there will be a tv commercial asking, “you just got a tax rebate, what are you going to do?” “i’m going to disney world!” however, winning a game show, or perhaps even earning a bonus may be cause for celebration. if the mental frame of the windfall is celebratory, then 21e. bland, v. chambers / financial services review 28 (2020) 17-34 perhaps the spending will be directed more toward celebrations and fun than if the windfall was from a tax rebate. alternatively, if the recipient were looking to brag about or show off their good fortune, they may be more likely to spend it conspicuously on fun than on regular household expenses. they may allocate more toward an infrequent expense such as a vacation, bigger holiday gifts, or something they have always wanted. differences in the amount spent by classification and by source are to be expected, but no source is absolute and completely separate in characteristics from the other sources, biasing against finding any differences. basically, our hypothesis is that the more euphoric and hedonistic the source, like game show winnings, the more one would spend on fun. alternatively, the more adult the source of the windfall, such as a tax rebate, the more one would allocate to more “responsible” uses, like investing in stocks and bonds, or household expenses and durable goods, such as a car or washing machine. in testing these hypotheses, the amount of the income in dollars and relative to household income, the amount of the payment, the respondent’s habit of spending or saving, the order of presentation, the frequency of payments and the demographic characteristics of the respondents will be controlled for. with that in mind, the null hypotheses are: hypothesis 1: there will be no difference in spending on “fun” by source of windfall. hypothesis 2: there will be no difference in allocations for regular expenses, credit card payments, durable assets or investing in stocks, bonds and savings account (adult uses of funds) by source of windfall. additionally, if either of these hypotheses produce significant findings, the sensitivity of the spending pattern will be analyzed to answer the research question: rq1: is the amount spent on fun or adult uses bounded at a fixed dollar amount or a relative percentage of the amount received, or is it relatively elastic? this research question, we believe, has been previously unexplored in research literature and represents a contribution to the knowledge of the field. 4. method this study examined respondents’ intended uses of hypothetical windfalls. sheppard et al.’s (1988) meta-analysis of 86 theory-of-reasoned-action studies found a 0.53 correlation between intention and behavior, indicating that intent is a strong predictor of action. in this study, the intended spending/saving patterns of respondents were gathered through 80 different paper-and-pencil instruments through students’ professors at seven universities. professors familiar with this type of research gathered the responses, sometimes providing a negligible amount of extra credit, and returned the responses to the authors. each participant was given one of these 80 instruments at random, which contained identical questions except for the source of the income and the amount of the hypothetical cash transferred, and asked how she would use the funds, both if it were received as a lump-sum and if the same amount were received spread out over 12 equal monthly payments (within-subject design) from two 22 e. bland, v. chambers / financial services review 28 (2020) 17-34 of these five sources: bonus, game show winnings, inheritances, lottery winnings, and tax rebates (between-subjects design). the amount of payments on the instrument was one of these four different amounts: $300, $600, $1,500, and $3,000. some instruments presented the lump-sum amounts first and some presented the periodic amounts first to control for the order effect. consistent with chambers and spencer (2008), the instruments asked how much of a lump sum rebate would be used for: (1) investing, (2) paying off credit card debt, (3) paying off notes, (4) regular monthly expenses, (5) buying a durable asset, (6) saving for an infrequent expense, and/or (7) used for fun. the instrument also asked how much of a monthly payment, equal to one-twelth of the lump sum amount, would be used for each of these seven purposes. similarly, the flip side of each instrument asked these same questions, changing only the source of the payment from one source to another—such as from a tax rebate to a game show payment. students were considered provisionally acceptable respondents per walters-york and curatola (1998) and ashton and kramer (1980). as such, instruments were distributed to university students at these institutions: coastal carolina university, francis marion university, longwood university, metropolitan state university of denver, texas a&m university – corpus christi, university of alabama – birmingham, and university of houston – clear lake. notably, at least four of these universities have a large nontraditional student population which adds external validity to this study beyond that expected from a traditional student population examined in the academic studies just listed. all research questions were analyzed with descriptive statistics, and then were converted to a percentage of the total payment received for each of the seven categories: (1) investing, (2) paying off credit card debt, (3) paying off notes, (4) regular monthly expenses, (5) buying a durable asset, (6) saving for an infrequent expense, and (7) used for fun were coded as spending. because the examples listed in category (6) were “a vacation, bigger holiday gifts, or something you’ve been wanting,” the percentages for items (6) and (7) were added together as were the dependent variable for “fun spending.” the independent variables were log of income, materiality of payment, spending default (that is whether the respondent habitually saves or spends unexpected money received), dummy variables for the total amount of payment, and dummy variables for the source of the windfall: game show winnings, bonus, or tax rebate. demographic and other control variables were included to control for order effect, risk-taking variables gender, age, business experience level, and education level. the complete regression model was of the form: percent spent on fun � f(income, zero income, amount, education, gender, age, importance, seatbelt use, smoker, spend1, experience level, dummy variables for the source of the payment (tax rebate, bonus, or game show), and a dummy for the order of presentation (monthly payment first, or lump sum payment first)). “income” is the log of the respondent’s income plus one. “amount” is the hypothetical amount of the distribution in dollars. rather than use a continuous variable for the total payments, dummy variables were created for the four discrete payment amounts. education was divided into four categories: high school, associate degree, undergraduate degree, and 23e. bland, v. chambers / financial services review 28 (2020) 17-34 graduate degree. “gender” was a categorical male/female variable, where female was coded as “1.” “age” was the participant’s age in years. “materiality” was defined as the total payment divided by the income of the respondent. the “seatbelt” and “smoker” dummy variables were included as proxies for respondents’ risk preference; seatbelt wearers and smokers were coded as “1.” the “spend1” variable is a measure of respondents’ habit of using extra money; the respondents were asked “when you get ‘extra money,’ do you spend it or save it?” for those that answered “spend,” the dummy was set to 1. in testing these hypotheses, the order of presentation and the frequency of payments were also controlled for. interaction effects were also run as a control measure. ultimately, the monthly payments were considered immaterial and dropped from the model. basic regressions were run matching the two extremes of (un)fun: tax rebate and game show winnings, but eliminating nonsignificant control variables except for income, materiality, spend 1 and level of payment. that model is: hypothesis 1: percent spent on fun � f(income, materiality, spend 1, dummy variables for amount, and a dummy variable for game show). hypothesis 2: percent spent on adult sources � f(income, materiality, spend 1, dummy variables for amount, and a dummy variable for game show). 5. results there were approximately 1,800 returned instruments in total, of which 601 were usable and pertained to the tax rebate, bonus and game show sources of income. most of the remaining instruments measured responses for inheritances and lottery winnings, which were not used in this analysis. some of the instruments were not sufficiently completed, perhaps because some students were trying to get extra credit without doing the work, and because responses were anonymous, turning in a partially completed instrument would produce that effect. if we are correct in reading this situation, however, that would bias this study against findings because of the noise introduced in hastily completed instruments. twenty-two of these observations had at least one missing value. the results of the regression are shown in table 1. as shown in table 1, the results of this regression were highly significant at p � 0.0005, although the r2 is 0.0652 and the adjusted r2 is 0.0481. likewise, the source of the payment was highly significant at p � 0.0151 and the coefficient was a positive 0.06933, indicating that respondents spent more on fun when they received the same amount of payment from a fun source (game show) than from a less fun source (tax rebate), and rejecting the null hypothesis 1. materiality, which is the relative size of the total payment, was also statistically significant at p � 0.0106, however, the coefficient of 0.00009549 is economically quite low. the total amount of the payment for each dummy variable was significant at p � 0.05, with all amount coefficients being negative, indicating that the higher the payment, the less was spent on fun. spend1, which was the dummy variable equal to one for those that indicated that they would normally spend extra funds, was significant at p � 0.05 as this would be a 24 e. bland, v. chambers / financial services review 28 (2020) 17-34 one-tailed test for this variable. the results of this regression indicate that the first null hypothesis was rejected. in the combined, three-source regression shown in table 2, the model continues to be highly significant at p � 0.01. the amounts of the payment continue to be highly significant with negative coefficients, and the game show source dummy continues to be highly significant and results in higher spending on fun. when comparing payments from bonus and tax rebates (the omitted variable), the source of the payment was not marginally significant at p � 0.10, indicating that a bonus was more neutral than either tax rebate or game show sources. table 1 regression of hedonistic “fun” spending between game show and tax rebate analysis of variance source df sum of squares mean square f value pr � f r2 model 7 2.03636 0.29091 3.82 0.0005 0.0652 error 383 29.18922 0.07621 corrected total 390 31.22558 parameter estimates variable df parameter estimate standard error t-value pr � t intercept 1 0.22191 0.06037 3.68 0.0003 lnincome 1 0.00778 0.00583 1.33 0.1829 materiality 1 0.00009549 0.00003720 2.57 0.0106 spend1 1 0.05509 0.02934 1.88 0.0611 level600 1 �0.10282 0.03625 �2.84 0.0048 level1500 1 �0.09549 0.04242 �2.25 0.0249 level3000 1 �0.13641 0.04201 �3.25 0.0013 gameshow dummy 1 0.06933 0.02839 2.44 0.0151 table 2 regression of hedonistic “fun” spending among bonus, game show, and tax rebate analysis of variance source df sum of squares mean square f value pr � f r2 model 8 1.97917 0.24740 3.21 0.0014 0.0431 error 570 43.97676 0.07715 corrected total 578 45.95592 parameter estimates variable df parameter estimate standard error t-value pr � t intercept 1 0.27961 0.05228 5.35 �.0001 lnincome 1 0.00184 0.00486 0.38 0.7046 materiality 1 0.00004837 0.00003032 1.60 0.1111 spend1 1 0.03874 0.02435 1.59 0.1121 level600 1 �0.08007 0.03120 �2.57 0.0105 level1500 1 �0.11248 0.03409 �3.30 0.0010 level3000 1 �0.11368 0.03441 �3.30 0.0010 bonus dummy 1 0.03351 0.02911 1.15 0.2502 gameshow dummy 1 0.06745 0.02840 2.38 0.0179 25e. bland, v. chambers / financial services review 28 (2020) 17-34 regressions were also run directly comparing spending on fun from game show winnings with a bonus. no statistically significant differences were found. similarly, regressions were run directly comparing spending from a tax rebate and a bonus. no statistically significant differences were found and tables for these results are omitted. this may mean that a bonus has characteristics of both a tax rebate and a game show winning. like a tax rebate, it is derived from work, but like a game show winning, a bonus may have euphoric qualities one would celebrate. though not statistically significant from zero, the coefficient for bonus is positive and about half the size of the coefficient for the game show dummy. in the end, while some of a tax rebate would be spent on fun, the regression results indicate that the amount spent on fun from a bonus is not different from the amount spent on fun when the source is either a tax rebate or a game show winning. regressions were also run to see if spending on adult uses would differ by the source of the income. various definitions of “adult uses” to mean (1) spending on regular monthly expenses, or (2) the sum of regular monthly expenses and paying off credit cards, or (3) the sum of regular monthly expenses, paying off credit cards, and to buy a durable asset (such as a car, boat, washing machine, or furniture) were used. regardless of the form of the measure used for “adult spending,” none of these regression models produced results significant enough to reject the second null hypothesis and are not presented as a table. however, when regressions were run to see if saving (investing in stocks, bonds, savings accounts, etc.) increased significantly when the source was a tax rebate instead of a game show, the results were significant (see table 3). recipients of tax rebates would allocate more to this type of savings than they would if they received the same windfall amount from either bonuses or game show winnings, confirming that at least to some extent, people tend to use fun sources of income for hedonistic uses and adult sources of income for adult, utilitarian uses, consistent with o’curry and strahilevitz (2001). like o’curry and strahitable 3 regression of percent invested in stocks, bonds, and savings comparing bonus, game show and tax rebate analysis of variance source df sum of squares mean square f value pr � f r2 model 8 4.23598 0.52950 7.51 �.0001 0.0953 error 570 40.19711 0.07052 corrected total 578 44.43309 parameter estimates variable df parameter estimate standard error t-value pr � t intercept 1 0.26953 0.04999 5.39 � .0001 lnincome 1 �0.00043260 0.00465 �0.09 0.9259 materiality 1 �0.00002566 0.00002898 �0.89 0.3764 spend1 1 �0.15297 0.02328 �6.57 � .0001 level600 1 0.04130 0.02983 1.38 0.1667 level1500 1 0.08352 0.03259 2.56 0.0106 level3000 1 0.08799 0.03290 2.67 0.0077 bonus dummy 1 �0.06468 0.02783 �2.32 0.0205 gameshow dummy 1 �0.05112 0.02715 �1.88 0.0602 26 e. bland, v. chambers / financial services review 28 (2020) 17-34 levitz (2001), our amounts varied in value; however, unlike o’curry and strahilevitz (2001), we did not ask respondents to assume complete financial independence and allowed respondents to allocate money rather than choose among a selection of prizes that we believe results in a more comprehensive and divisible allocation. income in both absolute terms and relatively (as measured by materiality) were extremely small and insignificant, indicating that the tendency to save and invest transcends income levels, but is strongly dependent on the respondent’s savings habits, as indicated by the spend1 variable. we then examined the pattern of responses further: were the coefficients for various levels of hedonistic spending linear by source, or did they display a different pattern? when looking at the coefficients for each level in the game show/tax rebate model and the significance of the materiality variable in model 1, the incremental amount spent appeared to be both significant and nonlinear. to confirm, the average percentage spent on fun was calculated for the monthly amounts of $25, $50, $125, and $250 and the yearly amounts of spending on fun. next, the average percentage spent on fun for game show, bonus and tax rebate windfalls was graphed. the results, shown in fig. 1, indicate that for small rebates, the percentage spent on rebates varied, and varied by source. for larger rebates of $1,500 and $3,000, spending on fun leveled out and began to converge at around 30%, regardless of source and then began to slowly fall. 6. discussion overall, this model lends significant support to o’curry and strahilevitz (2001) findings of people placing affective tags on money and expands the body of knowledge that one affective tag is fun. generally, these findings also support thaler’s (1999) mental accounting theory. however, the size of the effects also supports the neo-classical fig. 1. percent of windfall spent on fun. 27e. bland, v. chambers / financial services review 28 (2020) 17-34 economic notion that money is more fungible than not, and/or people are more rational than not with their money when it comes to fun sources and fun uses. the effect of affective tags may be bounded. regression results presented in table 2 suggest that survey respondents did not spend the windfall differently if the source was a tax rebate or a work bonus. additionally, the regressions comparing only windfalls from a bonus and a game show did not show a significant difference between the source. these results seem to indicate that there may be a hierarchy of fun sources. game shows winnings are likely more fun than work, and work is not much different from taxes, but game shows are clearly more fun than taxes. regression results presented in table 3 suggest that more of tax rebates, which are likely more predictable than bonuses and especially game show winnings, are invested, indicating that clients might be open to making investments during a predictable tax season, providing a greater demand for astute financial advisers like certified financial planners and certified public accountants. combined with other academic literature on the anticipation of a receipt discussed in the literature review section of this article, early, increased communication, especially in the february to april “tax season” might be beneficial to addressing and servicing clients’ financial needs. advisors that are already tax professionals may be at an advantage in serving clients because they know the timing of the receipt, and the amount as well. they also would have the means, with client permission, to split a direct deposit of a tax refund among up to three different accounts with up to three different u.s. financial institutions. splitting the refund can be accomplished electronically or through the irs’ form 888, allocation of refund (including savings bond purchases). we found the results of the research question enlightening. we know that respondents have separate mental accounts, or “buckets” (thaler, 1999). we know those accounts can get full (chambers, spencer, and mollick, 2009). this appears to be what is happening through roughly the $600 payment level. as income rises, so does lifestyle, ceteris paribus. however, not all uses of income necessarily rise proportionately. for example, if one’s income doubled, that person would not necessarily incur twice as much in medical expenses. a similar increase in income might result in more than doubling a household’s federal income tax bill because federal income tax rates are progressive. therefore, how do the allocations for fun change with an increase in income? apparently, at small amounts of affectively tagged windfalls, enough money is spent to fill the current bucket for fun, and then the size of the bucket increases proportionately. the first part of this graph, then, suggests that people can have “enough fun” for their standard of living, confirming chambers et al. (2009) that buckets get full. the second part of this graph describes the elasticity of fun as windfalls increase, which is an important contribution to literature, which we believe has not yet tested how the components of allocating income, and in particular fun, shift, if at all, with respondents’ increase in income. this leads to several questions for further study. how does the allocation of income, and in particular income from fun sources, shift, if at all, with respondents’ increase in income? it appears that while affective tags can produce significant results, hedonistic spending from an affectively tagged source may be, if not 28 e. bland, v. chambers / financial services review 28 (2020) 17-34 absolutely bounded, relatively bounded. people’s rationality, more than not, seems to keep exuberance in check. additionally, if people use adult sources like tax rebates for adult uses, then financial professionals can incorporate these findings into their financial advising practices. because tax filing is an at least annual event, and because most taxpayers receive refunds, tax season may be a robust time for financial professionals to encourage saving from a windfall. to enhance this practice, financial professionals might consider encouraging savings from the current refund and also the coming year’s tax refund. from what we know about the power of commitment, those who commit to saving in the future save less than what they commit to saving, but more than those who did not make a commitment at all (thaler and sunstein, 2009). then remind clients of their future commitment throughout the year. mullainathan and shafir (2013) showed that by sending a monthly reminder to save by either text or letter, for example, savings increased 6%. this method of future client commitment in the tax setting combined with reminders has not been tested though, so questions still remain, which leads to several questions for further study. 7. limitations and opportunities for future research this research showed that there might be a limit to how much of a windfall people are willing to spend on fun, even if the windfall is from a fun source. future research could explore the elasticity or shape of the spending on fun. other questions also lend themselves to further research: had windfalls increased further, would the percentage of income allocated to fun stay relatively flat? to what extent is hedonistic spending bounded when the source of the money is affectively tagged as either fun or adult? had windfalls increased further, would the percentage of income allocated to fun stay relatively constant? do other allocations of income to, for example, monthly expenses and investments also grow proportionately, or do some level off or even reverse? what are the other affective tags? how does tagging affect income allocations currently, and as the amount of windfall income rises? one limitation of this article is that it focuses on the changes in behavioral intent when presented with modest windfalls from different sources, and does not examine the latent mental processes (or lack thereof) that are used to reach that intent. we do not disentangle the stimulus, or priming, from the mental accounting that produces the behavioral intent. priming can be exhibited through what thaler and sunstein (2009) would call a “nudge,” for example when setting up certain financial defaults to encourage individuals to save for their retirement. mental accounting on the other hand is an internal construct, but it is connected to nudges that others may use in the environment to improve the choices of people who process information through their mental accounting systems. that is, priming is a cause that when processed with another’s mental accounting system may yield a different behavior than that displayed by those who were not primed. in this particular study, we are less concerned with the nuance of disentangling the prime from the latent mental accounting, and more concerned with the type of stimulus and differences in intent. 29e. bland, v. chambers / financial services review 28 (2020) 17-34 in this instrument, we asked if respondents smoked and if they wear seatbelts as a proxy for risk aversion. we also asked them for the extent of their business experience. we had seen these questions in previous studies, sometimes in multiple studies. in hindsight, these questions were too general to yield meaningful results. for example, an item on “personal financial expertise” would likely have yielded more information than the more general “business experience.” similarly, we analyzed the differences in uses among hypothetical receipts from the following sources: game show earnings, bonuses, and tax rebates. underlying this analysis is that winning money on a game show is more fun than doing one’s taxes. that might not be universally true, and that assumption biases against results in this this study. for future research, it might be useful to test the extent that respondents find game show winnings to be more fun than receiving a tax rebate. for example, some people may be so thrilled that they are (hypothetically) getting any money that the source is irrelevant, and a separate analysis of those respondents may result in further interesting, significant findings. another way to examine the effects of fun earnings relative to adult earnings would be to ask respondents what categories of activities they saw as fun rather than adult. these responses could be incorporated into fun or adult scenarios and, among subjects design, could test for different uses of those funds. such a design may be a better test of these concepts and produce stronger results. 8. conclusion our findings are consistent with the idea that fun sources of income are more likely to be spent on a fun expenditure. money won on a game show would be spent more on fun than money received from a tax rebate. this provides support for rejecting the first null hypothesis that there will be no difference in spending on fun by source of windfall. however, there may be a hierarchy of fun sources: game shows winnings are likely more fun than work, and work is not much different from taxes, but results show that game shows winnings are clearly more fun than tax rebates. while some of the regression results were unable to reject null hypothesis 2 as it pertains to adult spending, regression results for additional investing provide support for rejecting null hypothesis 2, that there is no difference in allocations for regular expenses, credit card payments, durable assets, or investing in stocks, bonds and savings account (adult uses of funds) by source of windfall. though various combinations of “adulting” (spending on adult causes) were used, we were unable to show that more adult sources of income are spent on adult spending like paying down a credit card or paying regular household expenses. however, there is significant evidence that there is a difference in investing based on the source of the windfall, validating a specific kind of adulting. finally, the percentage of the windfall spent on fun levels out. people will apparently spend significantly more on fun when a fun windfall is received, but that spending on fun is not limitless. additionally, as the amount of the total payment increased, the percentage spent on fun appears to level, indicating that at least within this range of payments, there may be such a thing as “enough (spending on) fun.” 30 e. bland, v. chambers / financial services review 28 (2020) 17-34 acknowledgment the authors wish to acknowledge and thank stetson university for supporting eugene bland’s development leave. appendix sample survey instrument “what would you do if . . .?” (fill in the amounts): by participating in a game show, you won a prize that would result in you receiving $600.00 for 2012. if received, how much of these winnings would you plan to: 1. invest (in stocks, bonds, savings account, and so forth)? $ 2. use to pay off credit card debt? $ 3. use to pay off notes (such as mortgage, car note, and so forth)? $ 4. use up about evenly every month for expenses? ______/month. � 12 months. � $ 5. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 6. use to save for an infrequent expense (such as a vacation, bigger holiday gifts, or something you’ve been wanting)? $ 7. spend right away on something fun? $ amount must total $600.00————-3 if instead, by participating in a game show, you won a prize that would result in you receiving $50.00/month for the next 12 months. if received, how much of this monthly increase would you plan to: 8. invest (in stocks, bonds, savings account, and so forth)? $ 9. use to pay off credit card debt? $ 10. use to pay off notes (such as mortgage, car note, and so forth)? $ 11. use up for regular monthly expenses? $ 12. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 13. use to save for an infrequent yearly expense (such as a vacation, bigger holiday gifts, and/or something you’ve been wanting)? $ 14. spend right away on something fun? $ amount must total $50.00————-3 please list your: zip code________ years of work experience ____ highest education level: high school ___ associate degree ___ undergraduate ___ graduate or above ___ occupation: __________________ gender: female ___ male___ age ____ race/ethnicity ___________________ # of college-level accounting classes completed college major (if applicable) __________________ industry where you work ______________________________ approx. yearly household income (from all wage and salary earners and other sources of income) $__________________ credit card debt: $_______________ other debt: $_______________ do you smoke? yes ___ no ___ do you normally wear your seatbelt? yes —___ no ___ when you normally get “extra money,” do you spend it or save it? spend ___ save ___ i rate my level of business experience as: high ___ fairly high ___ moderate ___ fairly low ___ low___ none ___ complete other side, please. thank you for your participation!!! 31e. bland, v. chambers / financial services review 28 (2020) 17-34 references abeler, j., & marklein, f. (2016). fungibility, labels, and consumption. journal of the european economic association, 15, 99–127. adamopoulou, e., & zizza, r. (2017). regular versus lump-sum payments in union contracts and household consumption. working paper series 2013, european central bank. agarwal, s., & qian, w. (2014). consumption and debt response to unanticipated income shocks: evidence from a natural experiment in singapore. american economic review, 104, 4205–4230. ariely, d. (2008). predictably irrational. new york, ny: harpercollins. arkes, h. r., joyner, c. a., pezzo, m. v., nash, j. g., siegel-jacobs, k., & stone, e. (1994). the psychology of windfall gains. organizational behavior and human decision processes, 59, 331–347. ashton, r. h., & kramer, s. s. (1980). students as surrogates in behavior accounting research. journal of accounting research, 18, 1–15. baker, m., nagel, s., & wurgler, j. (2007). the effect of dividends on consumption. brookings papers on economic activity, 1, 231–276. baker, h. k., & nofsinger, j. r. (2002). psychological biases of investors. financial services review, 11, 97–116. benjamin, d. j. (2006). psychology, preferences, and markets. proquest dissertations and theses, pp. 231–231. (available at http://search.proquest.com/docview/305344915?accountid�7084) blinder, a. s. (1981). temporary income taxes and consumer spending. journal of political economy, 89, 26–53. bodkin, r. (1959). windfall income and consumption. the american economic review, 49, 602–614. “what would you do if . . .?” (fill in the amounts): you got a bonus at work that would result in you receiving $600.00 which for 2012 will automatically be mailed to you as a check from your employer. if enacted, how much of this monthly increase would you plan to: 15. invest (in stocks, bonds, savings account, and so forth)? $ 16. use to pay off credit card debt? $ 17. use to pay off notes (such as mortgage, car note, and so forth)? $ 18. use up about evenly every month for expenses? ______/month. � 12 months. � $ 19. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 20. use to save for an infrequent expense (such as a vacation, bigger holiday gifts, or something you’ve been wanting)? $ 21. spend right away on something fun? $ amount must total $600.00————-3 another work bonus would result in you receiving $50.00/month after taxes; that is, your paychecks would go up $50.00/month. if received, how much of this monthly increase would you plan to: 22. invest (in stocks, bonds, savings account, and so forth)? $ 23. use to pay off credit card debt? $ 24. use to pay off notes (such as mortgage, car note, and so forth)? $ 25. use up for regular monthly expenses? $ 26. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 27. use to save for an infrequent yearly expense (such as a vacation, bigger holiday gifts, and/or something you’ve been wanting)? $ 28. spend right away on something fun? $ amount must total $50.00————-3 thank you for your participation!!! 32 e. bland, v. chambers / financial services review 28 (2020) 17-34 boylan, s. (2010). prior audits and taxpayer compliance: experimental evidence on the effect of earned versus endowed income. journal of the american taxation association, 32, 73–88. bradford, t. w. (2008). intergenerationally gifted asset dispositions. journal of consumer research, 36, 93–111. browning, m., & collado, m. d. (2001). the response of expenditures to anticipated income changes: panel data estimates. the american economic review, 91, 681–692. camerer, c., babcock, l., loewenstein, g., & thaler, r. (1997). labor supply of new york city cabdrivers: one day at a time. quarterly journal of economics, cxiii (may), 407–442. chambers, v., & spencer, m. (2008). does changing the timing of a yearly individual tax refund change the amount spent vs. saved? journal of economic psychology, 29, 856–862. chambers, v., bland, e., & spencer, m. (2017). does the source of a cash flow affect spending versus saving? financial services review, 26, 291–313. chambers, v., spencer, m., & mollick, j. (2009). goldilocks rebates: complying with government wishes only when the rebate amount is “just right.” journal of economics and economic research, 10, 101–119. cheema, a., & soman, d. (2006). malleable mental accounting: the effect of flexibility on the justification of attractive spending and consumption decisions. journal of consumer psychology, 16, 33–44. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. (available at https://manowar.tamucc.edu/login?url�http://search. proquest.com/docview/212012102?accountid�7084) epley, n., mak, d., & idson, l. (2006). bonus or rebate?: the impact of income framing on spending and saving. journal of behavioral decision making, 19, 213–227. epley, n., & gneezy, a. (2007). the framing of financial windfalls and implications for public policy. the journal of socio-economics, 36, 36–47. fisher, p. j. (2010). black-white differences in saving behaviors. financial services review, 19, 1–16. fisher, p. j., hayhoe, c. r., & lown, j. m. (2015). gender differences in saving behaviors among lowto moderate income households. financial services review, 24, 1–13. frederick, s. (2005). cognitive reflection and decision making. the journal of economic perspectives, 19, 25–42. friedman, m. (1957). a theory of the consumption function. princeton, nj: national bureau of economic research. heath, c., & soll, j. b. (1996). mental accounting and consumer decisions. journal of consumer research, 23, 40–52. henderson, p. w., & peterson, r. a. (1992). mental accounting and categorization. organizational behavior and human decision processes, 51, 92–117. hershfield, h. e., sussman, a. b., rourke, l. o., & bryan, c. j. (2015). leveraging psychological insights to encourage the responsible use of consumer debt. perspectives on psychological science, 10, 749–752. hsieh, c. (2003). do consumers react to anticipated income changes? evidence from the alaska permanent fund. the american economic review, 93, 397–405. johnson, d. s., parker, j. a., & souleles, n. s. (2006). household expenditure and the income tax rebates of 2001. american economic review, 96, 1589–1610. karlsson, n., gärling, t., & selart, m. (1999). explanations of effects of prior income changes on buying decisions. journal of economic psychology, 20, 449–463. kreinin, m. e. (1961). windfall income and consumption: additional evidences. the american economic review, 51, 388–390. levav, j., & mcgraw, p. (2009). emotional accounting: how feelings about money influence consumer choice. journal of marketing research, 46, 66–80. meekin, k. e., tach, l., & sykes, j. (2015). it’s not like i’m poor: how working families make ends meet in a post-welfare world. berkeley, ca: university of california press. milkman, k. l., & beshears, j. (2009). mental accounting and small windfalls: evidence from an online grocer. journal of economic behavior & organization, 71, 384–394. (available at https://manowar.tamucc.edu/ login?url�http://search.proquest.com/docview/230848483?accountid�7084) mullainathan, s., & shafir, e. (2013). scarcity: why having too little means so much. london: macmillan. o’curry, s. (1999). income source effects. unpublished working paper, depaul university. 33e. bland, v. chambers / financial services review 28 (2020) 17-34 o’curry, s., & strahilevitz, m. (2001). probability and mode of acquisition effects on choices between hedonic and utilitarian options. marketing letters, 12, 37–49. parker, j. (1999). the reaction of household consumption to predictable changes in social security taxes. american economic review, 89, 959–973. read, d., loewenstein, g., & rabin, m. (1999). choice bracketing. journal of risk and uncertainty, 19, 1–3, 171–197. rizzo, j. a., & zeckhauser, r. j. (2003). reference incomes, loss aversion, and physician behavior. the review of economics and statistics, 85, 909–922. rucker, m. h. (1984). allocation of windfall income: a case study of a retroactive pay increase to university employees. the journal of consumer affairs, 18, 101–118. sahm, c. r., shapiro, m. d., & slemrod, j. (2012). check in the mail or more in the paycheck: does the effectiveness of fiscal stimulus depend on how it is delivered? american economic journal-economic policy, 4, 216–250. shapiro, m. d., & slemrod, j. (1995). consumer response to the timing of income: evidence from a change in tax withholding. american economic review, 85, 274–283. shaprio, m. d., & slemrod, j. (2003). did the 2001 tax rebate stimulate spending? evidence from taxpayer surveys. in j. m. poterba (ed.), tax policy and the economy (83–109). cambridge, ma: national bureau of economic research and mit press. shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609–643. sheppard, b., hartwick, j., & warshaw, p. (1988). the theory of reasoned action: a meta-analysis of past research with recommendations for modifications and future research. journal of consumer research, 15, 325–343. souleles, n. s. (2002). consumer response to the reagan tax cuts. journal of public economics, 85, 99–120. spencer, m. k., & chambers, v. (2012). national heuristic shift toward saving any form of tax rebate. accounting and the public interest, 12, 106–136. thaler, r. h. (1999). mental accounting matters. journal of behavioral decision making, 12, 183–206. thaler, r. h., & sunstein, c. r. (2009). nudge: improving decisions about health, wealth, and happiness. london: penguin. trump, r. k., finkelstein, s. r., & connell, p. m. (2015). i will risk a stranger’s money, but not my own or my friend’s money: effect of proximity of the money source to the self on financial risk-taking. marketing letters, 26, 501–512. walters-york, l. m., & curatola, a. p. (1998). recent evidence on the use of students as surrogates subjects. advances in accounting behavioral research, 1, 123–143. wertenbroch, k. (2001). self-rationing: self-control in consumer choice. insead working paper no. 2001/63/mkt. winkelmann, r., oswald, a. j., & powdthavee, n. (2011, august). what happens to people after winning the lottery? in european economic association & econometric society parallel meetings (august 25–29). oslo, norway. zagorsky, j. l. (2013). do people save or spend their inheritances? understanding what happens to inherited wealth. journal of family economic issues, 34, 64–76. 34 e. bland, v. chambers / financial services review 28 (2020) 17-34 wealth and credit compliance: does economic literacy matter? celeste varuma,*, alla kolybana adepartment of economics, management and industrial engineering at the university of aveiro, campus universitário de santiago, 3810-193, aveiro, portugal abstract the purpose of this work is to empirically examine the influence of economic literacy upon individuals’ over-indebtedness and households’ wealth. it may be argued that the lack of economicfinancial knowledge may have detrimental consequences, in particular reflected in higher exposure to credit and financial risk. there is scarce literature testing the importance of economic-financial literacy for individuals’ over-indebtedness and household wealth. this article provides empirical evidence on the importance of financial literacy, for both individuals’ over-indebtedness and household wealth, exploring the case of portugal, a country about which there is scant empirical evidence on these matters. © 2014 academy of financial services. all rights reserved. jel classification: b26; d8; d14; d3; h63 keywords: economic literacy; financial literacy; over-indebtedness; wealth 1. introduction a recent report on financial stability from the bank of portugal shows that the degree of failure in the portuguese economy continues to increase to record levels, with bad loans reaching the highest levels in 15 years. in total lending to households, the amount classified as doubtful debts amounted to €5,031 million eurosthe highest on records. the latest figures show that out of €134,15 million of loans to households, 3.73% are of bad loans, meaning deemed uncollectible by financial institutions. in loans to households, out of the * corresponding author. tel.: 00351 234 370 200; fax: 00351 234 370 215. e-mail address: camorim@ua.pt (c. varum). financial services review 23 (2014) 325–339 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. €5,031 million classified doubtful, €2,229 million are in home loans and €1,539 million in consumer credit (e.g., to buy cars or appliances). moreover, owing to the large debt ratio and job instability, personal bankruptcies jumped to 2,400, a 79% increase in 2010. two in five bankruptcies in portugal are no longer in business, but individuals. these facts are replicated in many organization for economic co-operation and development countries. devaney and lytton (1995) explored the problem with household bankruptcy. however, more detailed analyses are necessary to reduce this trend. the rapid growth in household debt and its link to the current financial crisis has highlighted consumer weaknesses, and raises the question of whether individuals’ lack of economic and financial knowledge led them to take out mortgages and revolving credit they could not afford. indeed, the acknowledged widespread lack of financial and economic literacy casts serious doubts on the ability of individuals to make financial decisions. in this article we seek to understand the relationship between both economic and/or financial literacy and financial decision-making, as well as how they may combine to cause over-indebtedness. the study is based upon data collected through a new survey focused specifically on economic literacy, including a component of financial literacy. this article contributes to the existing literature in three ways. first, our research allows us to measure economic literacy, including financial matters, as well as individuals’ perception of overindebtedness. rather than relying on existing debt indicators, we ask individuals to judge their own debt levels (evaluate their capacity to pay their debts). finally, we assess how economic and financial literacy is linked to over-indebtedness and monthly household income. our sample comprises portuguese citizens with respect to their economic and financial literacy, and their judgments about the extent of their indebtedness and levels of income. in the next section we explore the nexus between economic literacy and over-indebtedness. section 3 describes the methodology, while the results are reported in section 4. section 5 concludes. 2. economic literacy and over-indebtedness there is wide recognition that people in general hold low levels of economic and financial literacy. as the organization for economic co-operation and development (2005) report on financial literacy documents, the low levels of financial literacy are not exclusive to portugal, it extend to several countries. similarly, the survey of health, aging and retirement in europe (share) shows that respondents score poorly on financial numeracy and literacy scales (christelis, jappelli, & padula, 2010). rapid growth in household debt raises the question of whether individuals’ lack of economic and financial knowledge led them to take out mortgages and revolving credit they could not afford. in recent years, a growing body of literature has shown that financial knowledge affects a wide range of financial behaviors (e.g., stock market participation, portfolio diversification, participation and asset allocation, indebtedness, etc.) that in turn can affect wealth accumu326 c. varum, a. kolyban / financial services review 23 (2014) 325–339 lation. indeed, research in the area has typically focused on individuals’ knowledge of economics and finance and its effects on financial decisions, usually related to savings, retirement planning, or portfolio choice (lusardi & tufano, 2009). there is some indication that economic and financial literacy may affect debt as well. we expect people with higher levels of economic literacy lato sensu, to have better capacity to avoid excess indebtedness. this knowledge influences ones’ ability to make simple decisions (e.g., regarding debt contracts and other decisions in the context of everyday financial choices). however, little research has been done on the relationship between economic and financial literacy and indebtedness. economic literacy comprises a set of knowledge and competences allowing agents to improve their personnel decisions in daily activities, this is a wide concept comprising economic and financial concepts. in addition, economic literacy is about making trade-offs between deficient resources (dahl, 1998), getting a job and combatting inflation (buchholz, 1998), understanding the forces that influence the quality of households lives (farrell, 1999), and increasing competence skills in complex global markets (gupta, 2006). in summary, economic knowledge will apply to daily decisions about financial to money management; it comprises the ability to understand, communicate, manage, decide, and forecast the financial issues (remund, 2010). the foundation of financial decisions is general economic knowledge. according to lusardi and mitchell (2007) and pang (2010) financial literacy will give the agent knowledge about economic concepts that will be used to plan, evaluate, and accurately decide about financial aspects. in contrast, insufficient economic or financial knowledge leads to low savings, mortgage defaults and financial mistakes, extra fees, or excessive interest rates on credit card debt (agarwal et al., 2009; agarwal & mazumder, 2011; banks & oldfield, 2007; gerardi, goette, & meier, 2010). moreover, economically illiterate individuals may be intimidated by politicians (gupta, 2006). moore (2003) reports that respondents with lower levels of financial literacy were more likely to have costly mortgages. the results from lusardi and tufano (2009) reinforce this concept as they have found a strong negative relation between literacy about debt-related issues and debt loads. generally, individuals with lower levels of debt literacy tended to conduct high-cost transactions (incurring fees and high borrowing rates); the less financially literate were either unable to judge their debt position or reported excessive debt loads. along this line, perry and morris (2005) found individuals who were more knowledgeable about financial matters were generally more likely to engage in financially responsible behavior, such as controlling their spending, budgeting, and planning for the future. other works explore the relationship between financial knowledge and money management (hilgert, hogarth, & beverly, 2003), retirement savings (bucher-koenen, 2009; lusardi, 2004; lusardi & mitchell, 2009), stock market participation (christelis, jappelli, & padula, 2010; van rooij, lusardi, & alessie, 2011; yoong, 2011), financial results (banks, 2010; smith, mcardle, & willis, 2010), and household wealth management (bateman et al., 2011; lusardi, & mitchell, 2007; van rooij, lusardi, & alessie, 2012; and others). hence, the effect of economic and financial literacy may reflect upon wealth. bateman et al. (2011) found that financial literacy score increases with household income. van rooij, lusardi, and alessie (2012) show that financial literacy is positively correlated with household wealth accumulation because economic and financial comprehension enhances the 327c. varum, a. kolyban / financial services review 23 (2014) 325–339 ability to form plans and make decisions and reduces the cost of collecting and treating information. in this way, analysis of relationship between financial literacy and its effects on household wealth has become important and relevant in the current policy environment. based upon the above we expect economic literacy to contribute to explanation regarding over-indebtedness and wealth. these expectations are tested in the following sections. 3. methodology 3.1. survey design we implement an extensive survey, fielded in april 2012, collecting information on individuals’ economic and financial knowledge and demographic characteristics (e.g., age, gender, nationality, education, and employment). we also assess attitudes towards economics as well as data on individuals’ judgments about their indebtedness. in addition, the survey collects self-reported information on household income. according to the literature, financial literacy is a branch of economic literacy; the first is essentially money management, and the second is the ability to accurately understand and decide about daily aspects. our survey encompasses questions regarding economic and financial knowledge. there are national surveys that measure financial knowledge; nevertheless, few specifically focus on economics. the assessment of economic and financial knowledge comprises 29 multiple-choice questions, with 22 that measure economic knowledge and seven focuses upon financial aspects specifically. the economic questions comprise the following economic topics: consumer economy, economy and production, financial economy, and the economic role of government and international economy. the financial questions cover the concepts of: euribor and spread, the degree of risk of a term deposit, relationship between inflation and interest rate, ability to pay a loan, identification of the balance of demand deposit, and identification of changes in the balance of the demand deposit. we follow an adapted version of the economic literacy test developed by the national council on economic education (ncee) to evaluate economic literacy. for the question on financial literacy section we use the questions applied in a survey conducted by the bank of portugal. 3.2. the sample the study was applied to a sample of portuguese adults with children at their charge, regarding their economic knowledge and their judgments about the extent of their indebtedness. the questionnaires were collected among the parents of children attending five different primary schools in portugal. in total, 618 questionnaires, out of the 1,061 questionnaires distributed, were returned, which means a response rate of 58%. table 1 reports frequencies (percent) of demographic and socioeconomic variables. most respondents in our sample are women (70.9%) and have portuguese nationality (93.2%). the majority of participants (68%) have between 36 and 45 years. about 26% of respondents reported at least 328 c. varum, a. kolyban / financial services review 23 (2014) 325–339 some kind of economic or financial education and 58.6% of them have higher education. about 66% of survey respondents replied they were able to meet their credit commitments 3.3. model to explore the influence of economic literacy upon the ability to pay individual’s debts we use a probit model. the dependent variable is binary, equaling 1 if the individual is able to fulfill all his debts, and 0 if he reports difficulties in meeting his debts. table 1 sample: demographic and socioeconomic characteristics demographic and socioeconomic variables % gender female 70.9 male 28.4 no answer 0.7 nationality portuguese 93.2 other 6.6 no answer 0.2 age 26–35 14.9 36–45 67.7 46–55 15.5 56–67 1.5 no answer 0.4 education level from 0 to 9th grade 15.8 12th grade 24.4 higher education 58.6 no answer 1.2 attended some kind of training in economics or finance yes 26.3 no 62.4 no answer 11.3 national qualification levels 1st group: legislators, senior officials, and managers 4.4 2nd group: professionals, including teaching professionals 37.8 3rd group: technicians and associate professionals 11.8 4th group: administrative staff 12.1 5th group: service workers and shop and market sales workers 10.2 6th group: skilled agricultural workers and fishery workers 0.5 7th group: craft and related trade workers 3.2 8th group: plant and machinery operators and assemblers 0.2 9th group: elementary occupations 4.5 no answer 15.3 income �1,000€ 19.4 1,001€–2,000€ 30.5 2,001€–6,000€ 37.3 6,001€–10,000€ 0.8 �10,001€ 0.3 no answer 11.6 can you assess your capacity to pay the debt? i can meet my commitments 66.1 i cannot meet my commitments 20.2 i don’t have credits 12.0 no answer 1.8 329c. varum, a. kolyban / financial services review 23 (2014) 325–339 the main explanatory variables in the model are the level of economic-financial literacy, economic literacy, and financial literacy measured by the score in the questionnaire (percentage of correct answers). additionally, we have considered variables likely to affect the probability of over-indebtedness. concerning demographic variables we have: gender (dummy: 1 for male and 0 for female), age and nationality (1 for portuguese, and 0 otherwise). income was measured regarding a likert scale from 1 to 5. education was proxied by edc1, edc2, and edc3 and previous training in economics or finance (dummy: 1 for yes, and 0 for no). the ranking of qualifications were defined in accordance to the national classification of professional activities, being that 1 corresponds to the highest level of qualification and 9 the lowest. a vector of variables was included to grasp individuals’ attitudes, interest towards economic matters, and motivations. follow is a dummy with value 1 if the respondent table 2 description of variables variables description dependent variable able to meet credit 1 � meet credits; 0 � over-indebted explanatory variables economic and/or financial literacy economic or financial literacy percentage of correct answers in economic or financial questions (x*100/29) economic literacy percentage of correct answers in economics (x_eco*100/22) financial literacy percentage of correct answers in finance (x_fin*100/7) demographic characteristics age discrete variable for individual’s age age2 age squared variable edc1 education level from 0 to 9 years schooling: 1 yes; 0 no edc2 12 years schooling: 1 yes; 0 no edc3 higher education: 1 yes; 0 no gender gender 1 � male; 0 � female nac nationality 1 � portuguese; 0 � otherwise socioeconomic characteristics income 1 to 5, from lower levels of income to higher levels of income econ attended some kind of training in economics or finance: 1 yes; 0 no cnp 1 to 9: 1 corresponds to the activities corresponding to the highest level of qualifications and 9 to the lowest interest and motivation towards economic matters follow frequently follows economic matters in the media: 1 yes; 0 no import degree of importance of economic knowledge to several situations: 0 to 29 attitudes regarding saving save saves: 1 yes; 0 no. aqgoods saves with medium objectives saves to acquire durable goods: 1 yes; 0 no retirement saves with long term objectives saves for retirement: 1 yes; 0 no 330 c. varum, a. kolyban / financial services review 23 (2014) 325–339 follows various media regarding economic affairs and 0 otherwise. aqgoods and retirement are two variables that indicate the motivation to save. aqgoods is a dummy variable that takes value 1 if the respondent saves essentially for purchase of durable goods and 0 otherwise, and retirement takes value 1 when the respondent saves for retirement and 0 otherwise. save is and a dummy variable that takes value 1 if the respondent is able to save a percentage of his disposable income and 0 otherwise. we measure the importance of the respondents knowledge of economics for various situations. for each situation we considered a likert scale of 1 (not important) to 4 (very important). the importance attributed to economics is measured by a weight factor (from 1 to 29, in an increasing scale). table 2 presents a summary of the variables. 4. results 4.1. descriptive findings table 3 reports percentage of correct answers on economic and financial literacy of respondents according all explanatory variables. based upon statistics presented in table 3 youngest respondents and minorities exhibit poorer performance in economic and financial literacy, which is particularly troubling. the average percentage of correct responses increases with post-school education and monthly household income. the average percentage of correct answers for economic questions is significantly higher than the average for financial questions, showing that portuguese citizen s have higher literacy level in economics than in finance. table 4 presents the descriptive statistics on the capability to pay credit payments and economic and/or financial literacy. there is a strong correlation between ability to meet commitments and economic or financial knowledge. furthermore, a significant differences between those that meet commitments and those that are over-indebted. descriptive statistics on capability to pay debts and demographic characteristics are presented in table 5. we find a significant relationship between demographic characteristics and over-indebtedness. descriptive findings show that men, portuguese respondents, and those with higher education are more likely to meet their credit obligations. table 6 corroborates the descriptive findings on capacity to pay their credit obligations and socioeconomic characteristics. it is observed that fewer difficulties in paying debts can be found among individuals who are able to save some money and among those who frequently follow economic issues in media. individuals with lower income and qualification levels are more likely to have difficulties paying credits. 4.2. econometric results in our study we seek to understand whether people face difficulties paying their debts and if this situation is related to their knowledge about economic matters. to proxy debt levels, we ask individuals about their capacity to meet their credit obligations. at first, we try to understand which factors influence the ability to meet credit. table 7 331c. varum, a. kolyban / financial services review 23 (2014) 325–339 table 3 percentage of correct answers to economic and financial questions characteristics % correct answers on economic literacy % correct answers on financial literacy gender female* 72.4 60.7 male* 83.3 71.3 nationality portuguese* 76.9 65.7 other* 56.9 36.6 age 26–35* 61.7 52.0 36–45* 78.7 67.0 46–55* 75.8 61.8 56–67** 73.9 59.4 education level from 0 to 9th grade** 46.8 44.5 12th grade* 72.8 60.5 higher education* 84.0 70.5 attended some kind of training in economics or finance yes* 85.6 73.6 no* 73.6 62.2 national qualification levels 1st group: legislators, senior officials, and managers* 84.7 69.2 2nd group: professionals, including teaching professionals* 84.2 71.3 3rd group: technicians and associate professionals* 78.8 66.0 4th group: administrative staff* 78.8 70.2 5th group: service workers and shop and market sales workers* 65.7 53.3 6th group: skilled agricultural workers and fishery workers 71.0 21.8 7th group: craft and related trade workers 64.5 55.3 8th group: plant and machinery operators and assemblers 27.0 0.00 9th group: elementary occupations** 47.9 42.5 income �1,000€* 61.5 50.2 1,001€–2,000€* 77.4 65.7 2,001€–6,000€* 86.3 74.7 6,001€–10,000€ 94.0 84.5 �10,001€ 95.0 74.5 frequently follows economic issues in the media yes* 80.8 70.3 no* 65.7 49.3 to understand political decisions very important* 83.0 71.1 important* 77.7 66.6 not very important* 70.2 58.7 not important* 68.4 52.8 (continued) 332 c. varum, a. kolyban / financial services review 23 (2014) 325–339 presents the econometric results. our dependent variable is binary; therefore, we will consider binary choice models; assuming a homogeneous sample, probit estimation tends to be the most accurate. this model shows no evidence of lack of fit based on the hosmerlemeshow statistic. in column 1 (first-stage estimation) we observe that higher schooling years and monthly income also positively influence the probability to meet credits; furthermore, assigning a proportion of the disposable income to saving also increases the probability of meeting credit obligations. in column 2 we infer that higher schooling levels, higher monthly income, and saving increase the probability of meeting credit commitments. table 3 continued characteristics % correct answers on economic literacy % correct answers on financial literacy to get a better paid job very important* 74.8 63.6 important* 76.6 66.2 not very important* 79.1 64.9 not important* 68.0 49.9 to become an active citizen and play a fuller part in society very important* 77.6 67.7 important* 75.9 64.3 not very important* 79.0 62.7 not important* 66.0 50.0 to make optimal choices about investment and saving very important* 80.9 69.0 important* 69.5 59.2 not very important* 58.3 40.8 not important* 41.2 16.5 to make better decisions on consumption now and in the future very important* 78.8 67.1 important* 76.0 64.5 not very important* 72.8 58.2 not important* 42.8 21.9 to make optimal choices about loans or credits very important* 80.6 68.5 important* 72.7 62.3 not very important** 61.4 49.9 not important* 45.4 15.5 to improve my wealth and well-being very important* 77.7 65.8 important* 77.5 66.2 not very important* 74.9 63.0 not important* 55.4 34.7 are you able to save some money? yes* 81.8 69.2 no* 69.2 58.8 save to acquire durable goods save* 78.6 65.9 do not* 74.7 63.8 retirement savings save* 78.7 67.5 do not* 75.5 63.3 note. we calculate an average percentage of economic and financial literacies for each category of explanatory variables. maximum score is 100%. moreover, we use paired t-test to compare financial and economic literacy scores. *1% significance level; **5% significance level. 333c. varum, a. kolyban / financial services review 23 (2014) 325–339 column 3 shows that financial knowledge, higher schooling levels, higher saving increase the probability of meeting contracts. we find that financial literacy shows a very strong inverse relationship with overindebtedness. those who report higher levels of financial literacy are more likely to belong to the group who report having no difficulties paying off debt. the effect is not only sizable, but it also increases with higher scores for self-assessed literacy. conversely, those who are less literate are much more likely to report having difficulties with debt and there is an inverse relationship between financial literacy and too much debt. although the estimates are less sizable than for those who may have difficulties with debt, the unsure also are much less likely to display high levels of literacy. these results are consistent with results of gerardi, goette, and meier (2010), which show significant correlation between mortgage delinquency and numerical ability. mortgage delinquency table 4 capacity to pay credits and economic and/or financial literacy characteristics % correct answers on economic or financial literacy % correct answers on economic literacy % correct answers on financial literacy meet credits 77.99 80.82 69.07 over-indebted 65.30 68.47 55.31 f-test 50.54 44.70 39.25 p-value 0.00 0.00 0.00 note. we use one-way analysis of variance to test the hypothesis that the means of economic and/or financial literacy score among two groups are equal. the first group represents respondents, who can meet credits and second who are over-indebted. table 5 capacity to pay credits across demographic characteristics characteristics % of respondents who are able to pay credits % of over indebted respondents gender male 81.88 18.13 female 74.12 25.88 pearson’s �2(1) 3.73 p-value 0.05 nationality portuguese 77.47 22.53 other 60.71 39.29 pearson’s �2(1) 4.15 p-value 0.04 education from 0 to 9th grade 46.05 53.95 12th grade 74.05 25.95 higher education 84.83 15.17 pearson’s �2(2) 52.25 p-value 0.000 note. we use �2 test of association to establish relationship between capacity to meet credits and demographic factors. 334 c. varum, a. kolyban / financial services review 23 (2014) 325–339 rates are greater among borrowers with lower financial literacy. demographic variables are correlated to debt loads as well. those who are employed and have higher income and higher wealth are much more likely to report they have the appropriate amount of debt. finally, women, and those with low income and wealth are more likely to be unable to judge their debt load. monthly income is likely to influence the ability to meet debt. we explore what determines asymmetries of income among families, and investigates the effect of economic knowledge. the relationship between monthly household income and economic and/or financial literacy is confirmed in ols regression analysis including the same explanatory variables as used previously (see table 8). table 6 capacity to pay credits across socioeconomic characteristics characteristics % of respondents who are able to pay credits % of over indebted respondents attended some kind of training in economics yes 82.43 17.57 no 75.99 24.01 pearson’s �2(1) 2.47 p-value 0.12 national qualification levels 1st group 64.00 36.00 2nd group 85.12 14.88 3rd group 76.56 23.44 4th group 84.06 15.94 5th group 63.27 36.73 6th group 33.33 66.67 7th group 66.67 33.33 8th group – – 9th group 45.00 55.00 pearson’s �2(7) 33.92 p-value 0.00 income level �€1,000 41.41 58.59 €1,001–€2,000 79.17 20.83 €2,001–€6,000 91.67 8.33 €6,001–€10,000 100.00 0.00 � €10,000 100.00 0.00 pearson’s �2(4) 99.72 p-value 0.00 frequently follows economic issues in the media yes 81.16 18.84 no 68.25 31.75 pearson’s �2(1) 11.34 p-value 0.00 are you able to save some money? yes 95.17 4.83 no 55.79 44.21 pearson’s �2(1) 108.94 p-value 0.00 note. we use �2 test of association to establish relationship between capacity to meet credits and socioeconomic factors. 335c. varum, a. kolyban / financial services review 23 (2014) 325–339 the ols regression shows that both economic and financial knowledge is positively correlated with monthly income. age, nationality, schooling, and saving are shown to be statistically significant. more concretely, they are positively related to the achievement of higher income levels. 5. conclusion the results of this study enable us to conclude that many portuguese households are not able to meet their credit commitments. moreover, our findings show that portuguese individuals have limited financial knowledge. in particular, the youngest and those respondents with lower level of education displayed the lowest levels of economic and/or financial knowledge. these findings are consistent with results of chen and volpe (1998). most important, those who have the lower levels of financial literacy are more likely to report problems with debt. moreover, respondents with lower levels of economic and financial literacy reported lower levels of monthly household income. all together, these findings point to the fact that widespread lack of economic and financial knowledge is a reasonable cause for concern. table 7 capacity to meet credits and economic and/or financial literacy: probit estimation explanatory variables model 1 (z-statistic) model 2 (z-statistic) model 3 (z-statistic) c �1.652 (�0.476) �1.832 (�0.532) �1.463 (�0.415) economic or financial literacy 0.012 (1.698) economic literacy 0.009 (1.252) financial literacy 0.011** (2.096) age �0.031 (�0.203) �0.023 (�0.151) �0.033 (�0.210) age2 �8.38e-05 (�0.046) �0.0002 (�0.094) �7.04e-05 (�0.038) gen 0.310 (1.412) 0.334 (1.527) 0.315 (1.438) nac �0.372 (�0.657) �0.292 (�0.525) �0.433 (�0.767) edc2 0.934** (2.346) 0.941** (2.362) 1.002** (2.517) edc3 0.871** (1.979) 0.891** (2.025) 0.922** (2.100) econ �0.266 (�1.229) �0.253 (�1.166) �0.237 (-1.106) cnp �0.047 (0.709) 0.045 (0.682) 0.040 (0.597) income 0.536* (3.193) 0.549* (3.285) 0.535* (3.191) follow 0.154 (0.735) 0.157 (0.752) 0.140 (0.667) aqgoods 0.336 (1.396) 0.340 (1.412) 0.375 (1.562) retirement 0.020 (0.083) 0.018 (0.074) 0.051 (0.209) save 1.161* (5.570) 1.151* (5.540) 1.188* (5.681) import 0.019 (0.702) 0.023 (0.859) 0.021 (0.768) total obs 357 357 357 mcfadden r2 0.34090 0.33721 0.34526 lr statistic 121.4201 120.1074 122.9730 prob(lr statistic) 0.00000 0.00000 0.00000 hosmer-lemeshow �2 (8) 11.72 8.15 12.90 prob � �2 0.1641 0.4188 0.1153 note. dependent variable is “able to meet credits.” all explanatory variables described in table 2. *1% significance level; **5% significance level. 336 c. varum, a. kolyban / financial services review 23 (2014) 325–339 education concerning economic principles and credit issues are very important in helping people avoid excess indebtedness, mortgage delinquencies and foreclosures, and bankruptcies and borrowing that are costly. the proliferation of credit and debt instruments, often with extensive information from written provisions and salespeople, can overwhelm borrowers. first, we focus attention on the important effect of economic and financial knowledge upon individuals’ debt. secondly, we consider the rich set of variables, attitude, and experiences. thirdly, we listen to individuals about their own debt levels. finally, we design a collaborative research project that blends scholarly research with timely market research. our conclusions suggest a complex set of interactions among literacy, experience, demographics, and debt loads. our work suggests that financial literacy is related to the choices that people make, and this in turn affects their income and capacity to meet debt. we interpret this to mean that additional research on economic and financial literacy—and education to enhance financial literacy—can complement, and not substitute for, auto-default and other comparable approaches. table 8 household income and economic and/or financial literacy: ols estimation explanatory variables model 1 ( t-statistic) model 2 ( t-statistic) model 3 ( t-statistic) c �3.114* (�2.750) �3.183* (�2.804) �3.178* (�2.800) economic or financial literacy 0.008* (3.269) economic literacy 0.006* (2.858) financial literacy 0.005* (2.875) age 0.129** (2.512) 0.132** (2.561) 0.138* (2.708) age2 �0.001** (�2.115) �0.001** (�2.159) �0.001** (�2.298) gen 0.004 (0.062) 0.016 (0.246) 0.011 (0.176) nac 0.757* (4.725) 0.792* (4.975) 0.763* (4.736) edc2 0.281** (2.172) 0.286** (2.199) 0.310** (2.403) edc3 0.822* (5.977) 0.836* (6.066) 0.856* (6.270) econ �0.066 (�1.014) �0.065 (�0.989) �0.043 (�0.672) cnp 0.005 (0.222) 0.005 (0.218) �0.003 (�0.121) follow 0.026 (0.387) 0.033 (0.498) 0.021 (0.317) aqgoods 0.053 (0.748) 0.050 (0.694) 0.082 (1.145) retirement �0.009 (�0.126) �0.012 (�0.156) 0.006 (0.082) save 0.399* (6.608) 0.397* (6.530) 0.421* (6.984) import 0.007 (0.816) 0.009 (0.953) 0.010 (1.151) total obs 394 394 394 r2 0.48551 0.48216 0.48229 adjusted r2 0.46650 0.46303 0.46316 f-statistic 25.5461 25.2060 25.2191 prob(f-statistic) 0.00000 0.00000 0.00000 breusch-pagan �2(1) 0.26 0.32 0.10 prob � �2 0.6068 0.5700 0.7542 note. dependent variable is “income.” all explanatory variables described in table 2. *1% significance; **5% significance. 337c. varum, a. kolyban / financial services review 23 (2014) 325–339 acknowledgment the study has been conducted under research project “economicando” (ptdc/egeeco/100923/2008), financed by feder funds through the programa operacional fatores de competitividade compete and by national funds through the fct fundação para a ciência e tecnologia. references agarwal, s., driscol, j., gabaix, x., & liabson, d. (2009). the age of reason: financial decisions over the lifecycle with implications for regulation. brookings papers on economic activity, autumn, 51–101. agarwal, s., & mazumber, k. (2011). cognitive abilities and household financial decisions making. paper presented at the 2nd boulder conference on consumer financial decision making, (june). boulder, colorado. banks, j., & oldfield, z. (2007). understanding pensions: cognitive function, numerical ability and retirement saving. fiscal studies, 28, 143–170. banks, j. (2010). cognitive function, financial literacy and financial outcomes at older ages: introduction. economic journal, 120, 357–362. bateman, h., eckert, c., geweke, j., louviere, j., thorp, s., & satchell, s. (2011). financial competence and expectations formation: evidence from australia. the economic records, 88, 39–63. bucher-koenen, t. (2009). financial literacy and private old-age provision in germany: evidence from save 2008. mannheim research institute for the economics of aging discussion, paper 192-2009. university of mannheim. buchholz, t. (1998). hope and danger for economic literacy. minneapolis, mn: the region, federal reserve bank of minneapolis. chen, h., & volp, r. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54, 18–39. dahl, d. (1998). why johnny can’t choose. minneapolis, mn: the region, federal reserve bank of minneapolis. devaney, s. a., & lytton, r. t. (1995). household insolvency: a review of household debt repayment, delinquency and bankruptcy. financial services review, 4, 137–156. farrell, c. (1999). we’re all economists. presented at the economic literacy symposium, 13 may. federal reserve bank of minneapolis. gerardi, k., goette, l., & meier, s. (2010). financial literacy and subprime mortgage delinquency: evidence from a survey matched to administrative data. federal reserve bank of atlanta, working paper 2010-10. federal reserve bank of atlanta. gupta, a. (2006). why johnny can’t choose: economic illiteracy in america. mid-american journal of business, 21, 3–5. hilgert, m., hogarth, j., & beverly, s. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309–322. lusardi, a. (2004). saving and the effectiveness of financial education. pension design and structure: new lessons from behavioral finance (pp. 157–84). oxford: oxford university press. lusardi, a., & mitchell, o. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. (2009). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness. national bureau of economic research, working paper 15350. national bureau of economic research, cambridge. 338 c. varum, a. kolyban / financial services review 23 (2014) 325–339 lusardi, a., & tufano, p. (2009). debt literacy, financial experience and over-indebtedness. national bureau of economic research working paper no. w14808. cambridge, ma: national bureau of economic research. moore, d. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences. technical report n. 03-39, social and economic sciences research center, washington state university, washington. pang, m. f. (2010). boosting financial literacy: benefits from learning study. instructional science, 38, 659–677. perry, v. g., & morris. m. d. (2005). who is in control? the role of self-perception, knowledge and income in explaining consumer financial behavior. journal of consumer affairs, 39, 299–313. organization for economic co-operation and development. (2005). improving financial literacy: analysis of issues and policies, paris, france. oecd, paris, france. remund, d. l. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. smith, j., mcardle, j., & willis, r. (2010). financial decision making and cognition in a family context. economic journal, 120, 363–380. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–72. van rooij, m., lusardi, a., & alessie, r. (2012). financial literacy, retirement planning and household wealth. the economic journal, 122, 449–478. yoong, j. (2011). financial illiteracy and stock market participation: evidence from the rand american life panel. in a. lusardi & o. mitchell (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 76–100). oxford: oxford university press. 339c. varum, a. kolyban / financial services review 23 (2014) 325–339 financial services review, 31(4) 283 households’ decision on capital market participation— what are the drivers? a multi-factor contribution to the participation puzzle andreas oehler1 and matthias horn2 abstract stock market investments are in the spotlight of the household finance literature, although realworld households make other financial decisions of higher relevance. we widen the scope and include decisions related to voluntary pension plans, whole life insurance contracts, housing, and investments in risky assets other than stocks (e.g., bonds or mutual funds). further, we provide a methodology that goes beyond regression analysis by employing a structural equation analysis (sea) and apply it on data from a broad and representative survey of the german central bank. our sea allows us to investigate and quantitatively estimate complex relationships and to test several hypotheses simultaneously. our structural equation model captures about 60% of the variation in the capital market participation decision. the results show that although households' financial literacy and risk aversion are most strongly related to investments in risky assets, further factors such as wealth, voluntary pension plans and whole life insurance contracts, financial advice, and investment experience should also be considered. financial literacy is negatively related to risk aversion (i.e., the higher the financial literacy, the lower is the risk aversion). age and gender are directly related to capital market participation and indirectly via financial literacy and risk attitude. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation oehler, a., & horn, m. (2023). households’ decision on capital market participation—what are the drivers? a multi-factor contribution to the participation puzzle. financial services review, 31(4), 283-304. 1 corresponding author (andreas.oehler@uni-bamberg.de). bamberg university, bamberg, germany. we thank john grable (the editor), daniel oehler, julian schneider, stefan wendt, and anonymous referees as well as participants at the afs academy of financial services 2023 conference, psychonomic society 2023 annual meeting, and 2023 society for judgment and decision making conference for valuable comments and suggestions. all remaining errors are our own. this paper uses data of the panel on household finances (phf) that is compiled by the german central bank (deutsche bundesbank). the results published and the related observations and analyses may not correspond to results or analyses of the data producers. the authors would like to thank german central bank for providing the dataset. 2 bamberg university, bamberg, germany https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 31(4) 284 introduction the participation of households in stock markets is one of the key issues in the literature on empirical financial markets in general and in the emerging field of household finance in particular (campbell, 2006; cocco et al., 2005; guiso & sodini, 2013; halko et al., 2012; kaustia et al., 2023; oehler & horn, 2021; oehler et al., 2022a). households without stock investments do not receive the equity premium and fail to invest efficiently (see mehra & prescott, 1985). studies have postulated theoretically and empirically that financial literacy, or the lack thereof, is a key driver of whether and to what extent people participate in stock markets (see chatterjee et al., 2017; lusardi & mitchell, 2008; lusardi & mitchell, 2014; oehler et al., 2018b; oehler et al., 2022a; van rooij et al., 2011; von gaudecker, 2015;). although stock market investments are in the spotlight of the household finance literature, realworld households make other financial decisions of higher relevance (kaustia et al., 2023). decisions related to housing or human capital investments in earlier stages in the life-cycle are also important as are investments in risky assets other than stocks (e.g., bonds or mutual funds). kaustia et al. (2019) note that data availability has likely been a factor in the number of studies conducted on stock market participation versus other aspects of household finance. we contribute to addressing this gap in the literature. more specifically, we widen the scope of previous studies from stock market to capital market investments and provide a multi-factor structural analysis with data from a broad and representative survey of the german central bank. our data allow both a differentiated analysis of capital market participation (i.e., not only equities but also other risky assets such as mutual funds or bonds, and a consideration of financial and non-financial factors of households' capital market participation). our structural equation analysis (sea) allows us to investigate and quantitatively estimate complex relationship structures between manifest and/or latent variables. in contrast to a regression analysis, sea tests complex variable relationships that reflect causal conjectures about the relationship structures among the variables under consideration. “complex” in this context means that several causal hypotheses are considered simultaneously. our structural equation model explains about 60% of the variation in the capital market participation decision. our results show that although households' financial literacy and risk aversion are the dominant drivers of investments in risky assets, wealth, voluntary pension plans and whole life insurance contracts, financial advice, and investment experience play a remarkable role. financial literacy has a negative influence on risk aversion (i.e., the higher the financial literacy, the lower is the risk aversion). we do not find significant results of housing on capital market participation, but age and gender play a role, directly, and indirectly via financial literacy and risk attitude. our study is organized as follows: in the next section, we review the literature. in the third section, we introduce the dataset and the methodology of our structural equation analysis. we then present the variables of the structural equation analysis, provide descriptive statistics, and outline our hypotheses. the paper concludes with the presentation and discussion of findings. literature review according to neoclassical models, capital market participation is only determined by risk attitude and each household should participate to get a share of the equity premium (guiso & sodini, 2013, pp. 1424 et sqq.; merton, 1969). hence, households’ risk attitude is considered as the most important determinant in both theoretical asset pricing models and studies that aim to empirically explain households’ investment decisions (cohn et al., 1975; dorn & huberman, 2005). studies often try to find explanations for solving the equity premium puzzle (i.e., the phenomenon that many households actually do not invest in stocks at all) (mehra & prescott, 1985). it has been well established that further factors in addition to risk attitude are relevant for stock market participation. an explanation for oehler & horn 285 households’ non-participation in the stock market is that they do not know the benefits of an investment. this is backed up by studies that find a significant influence of financial literacy on stock market participation (beshears et al., 2018; kaustia et al., 2023; laurinaityte, 2018; van rooij et al., 2011; von gaudecker, 2015). in addition, households with a higher total wealth face lower relative fixed participation costs and are, therefore, more likely to own stocks (campbell, 2006; bilias et al., 2010; calvet et al., 2007; haliassos & bertaut, 1995; kaustia et al., 2023; vissing-jorgensen, 2004). households’ monthly income determines the amount of money a households is able to save. hence, it is not surprising that households with higher monthly income are more likely to participate in the stock market (haliassos & bertaut, 1995; kaustia et al., 2023; laurinaityte, 2018; mankiw & zeldes, 1991). however, it is important to notice that these factors are not independent from each other (e.g., households’ willingness to take risk increases with wealth) (calvet & sodini, 2014; oehler & horn, 2021). moreover, further factors have an influence on households’ risk attitude and financial literacy (e.g., higher educated individuals are more likely to be financially literate) (bucher-koenen et al., 2021; bucherkoenen & knebel, 2021; hammer et al., 2022; kaustia et al., 2023). regression analyses can hardly cover these complex relations. hence, we employ sea. this allows us to include a battery of further factors that have an influence on cmp. it is evident that the age and gender of a decision maker is related to her risk attitude and financial literacy. according to calvet et al. (2009) age is negatively related to the sophistication of a household’s financial decisions. korniotis and kumar (2011) explain this effect with adverse effects of cognitive aging. hence, we predict that financial literacy will be higher at a lower age. moreover, older decision makers show a higher degree of risk aversion (dohmen et al., 2011, 2017; oehler et al., 2018a, 2022a). nevertheless, stock market participation increases with age (athreya et al., 2023; oehler et al., 2018a). in addition, previous studies find significant gender differences. the financial literacy of men is higher than that of women (bannier & schwarz, 2018; fey et al., 2020; guiso & zaccaria, 2023; hanna et al., 2021). men show a lower degree of risk aversion and invest more in risky assets (croson & gneezy, 2009; eckel & grossman, 2008; halko et al., 2012). the awareness and use of financial advice and financial planning tools enhance the probability to invest in capital markets (chien & morris, 2017; fey et al., 2020). von gaudecker (2015) reports that nearly all households that rely on professional contacts for advice achieve reasonable investment outcomes, particularly because financial advice leads to better diversified portfolios. financial advice, however, has no influence on the relation between financial literacy and stock market participation (hermansson et al., 2022). hence, financial advice is not a substitute for financial literacy. of course, the individual situation of a household, which is usually linked to the status in the lifecycle, has a major impact on investment decisions. oehler and horn (2021) build on the behavioral portfolio theory of shefrin and statman (2000) and show that households assign their assets into different mental accounts. the mental accounts build up on each other in a hierarchical structure (i.e., as layers of a pyramid). direct investments in financial markets such as stocks or bonds are in the highest layer, whereas residential property, pension plans, and whole life insurance contracts are in the layer below. this means that most households will only invest in stocks when they have financed their residential property and/or their pension plan. studies with a focus on housing decisions support this pyramid structure and real estate is by far the most popular investment vehicle for households in europe (efama, 2020, p. 29; kaustia et al., 2023). cocco et al. (2005) and gomes et al. (2021) argue that house ownership discourages saving in financial assets as households usually want to first repay their mortgage loan (see guiso & zaccaria, 2023; calvet & sodini, 2014). however, the influence of voluntary pension plans and whole life insurance contracts is understudied. our study caters to this gap in the literature. we assess investments in voluntary pension plans and whole life insurance contracts as a quasi-safe addition to a household portfolio. hence, households could spend further free financial services review, 31(4) 286 budget for riskier investments in the next higher layer instead of investing in additional risk-free assets. consequently we expect higher capital market participation when households already have a voluntary pension plans and/or a whole life insurance contract. investment decisions are always linked to a planning horizon, an assessment of the current situation, and expectations for the future. if applicable, past experiences additionally have an influence as households learn from their last decisions. households with a longer planning horizon and more positive expectations for the future are more likely to invest in stocks as they can better bear the short-term crash risk (ameriks & zeldes, 2004; barberis, 2000; calvet et al., 2007). households that are more satisfied with their current lifestyle should have more financial resources available for investments in risky assets and be less risk averse (xiao, 2016). investors with positive past investment outcomes usually get more confident and subsequently invest higher amounts and trade at higher frequency (choi et al., 2009; de et al., 2010). further, malmendier and nagel (2011) show that investors who have experienced higher stock market returns throughout their lives are less risk averse and more likely to invest in stocks. data and methodology phf survey data the panel on household finances (phf) by the german central bank (deutsche bundesbank) covers data for all the influential factors mentioned in the previous literature review. we use the dataset of the third wave. the dataset covers a variety of financial and behavioral variables at the household level and personal data on all household members. each household is represented by a financially knowledgeable person (fkp) who can provide the necessary information about the household and is assumed to be mainly responsible for the household’s financial decisions (see altmann et al., 2020; phf survey team, 2019a, 2019b; von kalckreuth et al., 2012). information about the fkp comprises age, gender, graduation, professional qualification, and financial literacy. the third wave of the phf started in march 2017, and the collection process ended in november 2017. the total number of households that participated was 4,962. following von gaudecker (2015), we exclude households with less than 1,000 euros in financial assets, which yields an initial sample of 4,538 households. methodology we use structural equation analysis (sea), which allows us to investigate and quantitatively estimate complex relationship structures between manifest and/or latent variables (byrne, 2016; hair et al., 2010). the aim of sea is to represent the a priori formulated relationships in a system of equations and to estimate the model parameters in such a way that the initial data collected on the variables are reproduced as well as possible. structural equation modeling has been used in many disciplines and has become an important method of analysis in academic research (e.g., byrne, 2001; hair et al., 2010; kline, 2005; savalei & bentler, 2006). in contrast to regression analysis (ra, ols), sea tests complex variable relationships that reflect causal conjectures about the relationship structures among the variables under consideration. “complex” in this context means that several causal hypotheses are considered simultaneously. in this context, individual variables in the different hypotheses may represent both independent and dependent variables. furthermore, bilateral relationships (interrelationships) between variables are also possible. thus, multi-equation systems are used, which represent the presumed effect relationships in several regression equations, which are estimated simultaneously (i.e., a non-recursive model) (weiber & mühlhaus, 2014). while a ra makes a clear distinction between a dependent and one or more independent variables, sea does not require such a clear distinction. another key difference from ra is that a ra considers only empirically directly measurable variables (manifest variables), whereas sea can analyze relationships between manifest variables as well as between latent variables (i.e., variables that are not directly observable). latent variables are also referred to as hypothetical constructs (e.g., risk attitude, oehler & horn 287 competence, financial literacy, trust, reputation). we use amos 29 and thus a covariance structure analysis based on confirmatory factor analysis. the latent variables are interpreted as factors that are “behind” the measurement variables and are assigned to the different measurement variables according to the formulated hypothesis system. factor analysis is then used to estimate the factor loadings (i.e., correlations between measured variables and factors) in such a way that the empirical variance-covariance matrix or correlation matrix can be reproduced as accurately as possible (weiber & mühlhaus, 2014). accordingly, manifest and latent variables are to be distinguished in sea. manifest variables are directly observable, and their manifestations can be recorded directly with the help of suitable measurement instruments. latent variables (i.e., hypothetical constructs) are characterized by the fact that they elude direct observability. therefore, suitable measurement models are needed to capture the manifestations of a latent variable in reality. if a structural model consists only of manifest variables and if there were no interactions between the variables, ra would be the classical method of analysis. if, on the other hand, there are interactions between the manifest variables, then path analysis is used. it allows complex structural models to be tested using multiple ra. for structural models that formulate relationships between latent variables, suitable measurement models are first required, which can be used to obtain empirical observed values for the latent variables. using these measurement values, the presumed structure between the latent variables can then be empirically tested, analogous to the case of manifest variables. for structural models with latent variables, the term causal analysis is also common in the literature (weiber & mühlhaus, 2014). the sea with latent variables thus consists of three sub-models. 3 a second possibility is to fix the variance of a latent variable to 1. in our analysis both types of metric determination lead to at least similar parameter estimates. hence, it can be assumed that the (1) the core is the structural model, which represents the theoretically assumed relationships between the latent variables. here, the endogenous variables are explained by the causal relationships assumed in the model, with the exogenous variables serving as explanatory variables, but not themselves explained by the causal model. (2) the measurement model of the latent exogenous variables contains the empirical measurements from the operationalization of the exogenous variables and reflects the assumed relationships between the measurements and the exogenous variables. (3) the measurement model of the latent endogenous variables contains the empirical measurements from the operationalization of the endogenous variables and reflects the presumed relationships between these measurements and the endogenous variables. accordingly, the relationships discussed in the literature review are the basis for our structural model. the two associated measurement models and the variables within them are discussed in the following section. we use amos 29 (arbuckle, 2019; byrne, 2016) to apply the structural equation model. as widely recommended in the literature, we apply the maximum-likelihood (ml) method (e.g., backhaus et al., 2015; weiber & mühlhaus, 2014; weston & gore, 2006). byrne (2001) notes that amos automatically imposes the value of one to the first of each set of factor loadings and to the regression coefficients associated with each error term. accordingly, they do not estimate these values. byrne (2001) explains that the factor loadings set to a value of one address the issues of model identification and the scaling of the unobserved factors, while those associated with the error terms represent values that are considered to be known (see backhaus et al., 2015; weston & gore, 2006).3 parameter estimates also provide reliable measurements of the unobservable variables (byrne, 2001; weiber and mühlhaus, 2014). for the single-item constructs it is assumed that the financial services review, 31(4) 288 as recommended in the literature, we examine the standardized estimates as they are considered most informative. because different variables may have different scales, determining which variable has the greatest effect can only be done by comparing the standardized parameter estimates (backhaus et al., 2015; weston & gore, 2006). we use the standardized total effects in general and differentiate between the direct and the indirect effects for some variables. to evaluate the goodness-of-fit between the hypothesized model and the sample data, we calculate several fit indexes. as recommended in the literature, we use the root mean square error of approximation (rmsea); the standardized root mean square residual (srmr); the adjusted goodness-of-fit index (agfi), the incremental-fit index (ifi), and the comparative-fit index (cfi) for baseline comparisons between the default model and independence model (see backhaus et al., 2015; browne & cudeck, 1993; byrne, 1989; byrne, 2016; hair et al., 2010; haughton et al., 1997; hu & bentler, 1999; maccallum et al., 1996; schermelleh-engel et al., 2003; weston & gore, 2006; weiber & mühlhaus, 2014). according to the literature, the two main indexes are the rmsea and the srmr. as an index of fit, rmsea corrects for a model’s complexity. as a result, when two models explain the observed data equally well, the simpler model will have the more favorable rmsea value. a rmsea value of zero indicates that the model fits the data exactly. weston and gore (2006) suggest using the 90% ci (confidence interval) for the rmsea that incorporates the sampling error associated with the estimated rmsea. the srmr index is based on covariance residuals in which smaller values indicate a better fit. the srmr is a summary of how much difference exists between the observed data and the model. indicator can measure the construct without error, which is why the variance of the associated error variable is fixed at a value of 0 (byrne, 2001; weiber & mühlhaus, 2014). single-item constructs: variables and hypotheses for the structural equation analysis dependent variables in the structural model and measurement concepts a sea, unlike a ra, allows individual variables in the different hypotheses to be both independent and dependent variables. these variables are also referred to as intervening variables. in our analysis, the variables risk attitude and financial literacy act as intervening variables between the independent and exogenous variables and the main dependent variable cmp. capital market participation the main dependent variable, capital market participation (cmp), comprises the following wealth positions of a household (phf survey team, 2019b, pp. 4-5): mutual funds, bonds, publicly traded shares (bucciol et al., 2019; calvet & sodini, 2014; calvet et al., 2007; halko et al., 2012). hence, we widen the scope of previous studies and focus not only on stock market participation. in order to cover the participation in that three main categories of capital market assets, we define dummy variables for the holding of publicly traded shares, bonds, and mutual funds (fey et al., 2020). as a hypothetical construct, the latent endogenous variable cmp influences these three measurement variables. the three dummy variables contain the empirical measurement from the operationalization of the endogenous variable cmp. financial literacy for the definition of financial literacy, we follow the growing strand of literature that uses the concept of financial capability with the key element of practical skills (see aubram et al., 2016; bernheim et al., 2001; deepak et al., 2015; dixon, 2006; oehler & werner, 2008; oehler et al., 2018b; xiao & o’neill, 2016) and the related concept of financial competencies by the oecd (oecd/infe, 2016). to measure financial literacy empirically, lusardi and mitchell age, gender, net wealth, net income, satisfaction with life (present), planning horizon (future), financial advice, and investment experience). oehler & horn 289 develop three questions that are suitable for surveys, although the questions reflect a rather narrow definition of financial literacy (see bucher-koenen & knebel, 2021; bucher-koenen et al., 2017; bucher-koenen et al., 2021; lusardi & mitchell, 2011; lusardi & mitchell, 2014). the three questions are on compound interest, inflation, and risk diversification. according to rieger (2020), the cronbach’s alpha of this scale is .43, which is “acceptable, given that it consists of only three items” (p. 4). the phf survey also follows this measurement of financial literacy. however, in its third wave, a fourth question on compound interest and debt is added (phf survey team, 2019a, pp. 164-165). the phf study allows us yet another perspective on financial literacy, namely the possibility of taking economic literacy courses while in school. this variable is determined by the answer to the question whether the respondent participated in courses or training sessions on household finances or asset management (phf survey team, 2019a, p. 32). in order to cover both categories of financial literacy, we define two variables for operationalizing the extent of financial literacy. as a hypothetical construct, the latent endogenous variable financial literacy influences that two measurement variables. the both variables contain the empirical measurement from the operationalization of the endogenous variable financial literacy. we determine our first measurement variable on financial literacy, fin_lit_score, with the answers to the four questions mentioned above and code the answers as indicator variables (van rooij et al., 2011). fin_lit_score equals four if all answers are correct, three if three out of four answers are correct, two if two out of four questions are correct, one if only one answer is correct, and zero otherwise (oehler et al., 2022a). the second measurement variable, fin_lit_training, equals one if a member of the respective household participated in courses or training sessions on household finances or asset management and zero otherwise. risk attitude in addition to financial literacy, the other dependent variable, risk attitude, acts as a main influencing variable in the analysis of capital market participation. the risk appetite, usually measured as degree of risk aversion, is a crucial determinant, and acts as an intervening variable between the independent variables and capital market participation. risk aversion is covered by two different concepts in the financial domain (schoemaker, 1993). one strand of literature relies on the neoclassical assumption that the financial risk taken by an individual mirrors exactly her risk aversion (e.g., arrow, 1965; pratt, 1964). hence, risk attitude can be measured by the self-selected level of financial risk. this concept is considered as objective risk aversion (see nosic & weber, 2010). other studies use the terms risk-taking (schooley & worden, 1996), observed risktaking (schoemaker, 1993), risk tolerance (wang & hanna, 1997), or relative risk aversion (riley & chow, 1992). we do not employ this concept due to possible endogeneity issues. the second strand of literature assumes that investment decisions are the result of a process that is additionally influenced by individuals’ subjective perception, heuristics, and bounded rationality (hirshleifer, 2015). therefore, the investment decisions, and likewise the measured objective risk aversion, are most likely driven by partially unobservable factors (schoemaker, 1993). in this framework, researchers consequently can only measure an individual’s risk aversion by directly asking them to selfassess their willingness to take financial risk (chaulk et al., 2003; dohmen et al., 2011; nosic & weber, 2010; oehler & horn, 2019). since individuals’ self-assessment always includes subjective components, it is a subjective risk aversion. other studies use the terms such as financial risk aversion (kaustia et al., 2023) or intrinsic attitude toward risk (schoemaker, 1993). since both concepts are not mutually exclusive, some studies combine both in one framework. for example, nosic and weber (2010) differentiate between subjective and objective risk aversion and find that the subjective risk aversion is a significant determinant of the objective risk aversion (see also schooley & financial services review, 31(4) 290 worden, 1996; chaulk et al., 2003; halko et al., 2012; kaustia et al., 2023). oehler et al. (2018a) conclude from a simultaneous analysis of both measures of risk aversion in an experimental setting that the subjective risk aversion is a better predictor for the objective risk aversion than a set of commonly used socio-demographic and economic factors such as age or income. hence, we assume that a measure of subjective risk aversion shall be a good predictor for cmp. dohmen et al. (2011) add to this discussion and use a question asking people about their willingness to take risks “in general”. they confirm the behavioral validity of this measure in an experiment that uses paid lottery choices and conclude that this question is the best all-round predictor of risky behavior. following the main findings of the literature, we use the measure of dohmen et al. (2011) and a measure of subjective risk aversion within the measurement model for the latent variable risk attitude. in order to cover both categories of risk attitude, we define two variables for operationalizing the extent of risk aversion. as a hypothetical construct, the latent endogenous variable risk attitude influences that two measurement variables. the both variables contain the empirical measurement from the operationalization of the endogenous variable risk attitude, the self-assessment of risk aversion in the financial domain, riskfin; and the selfassessment of general risk-taking, riskgen (oehler et al., 2022a). riskfin is determined by the answer to the question, “if savings or investment decisions are made in your household, which of the statements best describes the attitude toward risk?” (phf survey team, 2019a, p. 153), on a scale from one to four. one means that “we take significant risks and want to generate high returns”; two means that “we take above-average risks and want to generate above-average returns”; three means that “we take average risks and want to generate average returns”; and four means that “we are not ready to take any financial risks”. riskgen is determined by the answer to the question, “how do you view yourself? are you in general a risk-taking person or do you try to avoid risks?” on a scale from 0 to 10. zero means that you are “very willing to take risks”; 10 means that you are “not at all ready to take risks” (the original scale is recoded to align in the same direction as in the question on risk aversion in the financial domain) (oehler et al., 2022a). independent variables in the structural model and measurement concepts the measurement model of the latent exogenous variables contains the empirical measurements from the operationalization of the exogenous variables and reflects the assumed relationships between the measurements and the exogenous variables. net wealth we use a household’s net wealth as proxy for its wealth position (total household assets minus total outstanding liabilities, measured in euros) (phf survey team, 2019b, p. 8). among others, households’ financial assets include the total value of deposits, mutual funds, bonds, non selfemployment private businesses, publicly traded shares, managed accounts, money owed to the household, ‘other’ financial assets, voluntary pension plans and whole life insurance contracts. the main share lies in deposits and “households need to keep enough of their wealth in deposits to manage their everyday spending and meet any unforeseen needs; the lower their overall wealth, the more they will need to rely on easily accessible cash” (efama, 2020, p. 28). the other main wealth component is the real asset position including the household’s main residence. home loan saving and hmr mortgage outstanding for a clearly differentiated analysis, we additionally use the amount saved in euros for a household main residence (hmr) via home loan saving contracts (phf survey team, 2019a, p. 139) as an alternative investment in the household’s portfolio as well as the current level of outstanding debt in euros for existing hmr as the major share of household’s debt (phf survey team, 2019b, p. 5). voluntary pension plan and whole life insurance oehler & horn 291 investments as precautions, in particular for oldage provision, are a diversifying addition to the portfolio of financial assets. these investment alternatives through financial intermediaries such as insurance companies typically represent no direct investments in capital markets. it could be argued that such products of financial intermediaries are also related to the capital market, because a part of the clients’ insurance premiums are likely to be invested in bonds. for this analysis, however, the perception of households is crucial. most households probably understand such investments as so-called safe investments, similar to deposits. according to efama (2020, p. 5 & p. 24), the “strong market position of insurance-based products can be explained by the preference of many citizens for products with a nominal capital guarantee and a strong preference by households for saving in bank deposits and insurance products that offer some form of guarantee.” we employ the total amount in euro invested by a household in voluntary pension plans and whole life insurance contracts (phf survey team, 2019b, p. 3) as independent variables. net income household income originates from different sources, in particular from employment, selfemployment, and pensions (phf survey team, 2019b, pp. 4-5). for a more realistic analysis, we calculate a household’s net income position in euro to approximate a possible volume for the cmp. the net income takes the estimate of monthly net disposable income into account (after the deduction of taxes and social security contributions; phf survey team, 2019a, p. 38), minus total expenditures of the household typically spend per month on consumer goods and services (without financial payments) (e.g.. loan repayments) (phf survey team, 2019a, p. 35), and minus payments for household’s total debt (phf survey team, 2019b, p. 6). satisfaction with life (present) the phf survey provides a subjective measure of satisfaction that captures the self-perceived overall satisfaction status. to map the current life situation in the life cycle, we use the question about the current satisfaction with life as a proxy (satisfied overall with life at present, 0 = totally dissatisfied, 10 = entirely satisfied; phf survey team, 2019a, p. 119). planning horizon (future) to map the future life situation in the life cycle, we use the question about the planning horizon as a proxy. we code the answers as follows: 0 = “we do not make plans in advance”; 1 = “a few months”; 2 = “one year”; 3 = “a few years”; 4 = “5 to 10 years”; 5 = “more than 10 years”. age we use the variable age as an additional proxy for the life-cycle status (e.g. oehler et al., 2022a). age is calculated as the difference between 2017 (the year when the third wave of the survey was conducted) and the year of birth of the fkp. some empirical studies have also used the squared age and higher moments of age (ameriks & zeldes, 2004; cocco et al., 2005; guiso & sodini, 2013; fagereng et al., 2017; fey et al., 2020; kaustia et al., 2023; poterba & samwick, 2001). when we use the age squared, the results of our model differ only marginally. hence, we only use age for a more intuitive interpretation. gender following the literature we control for gender effects by including the gender of the financially knowledgeable person (fkp) as dummy variable (1 = male, 2 = female; phf survey team, 2019a, p. 168). the analysis of hanna et al. (2021) on whether the husband or wife was the financially knowledgeable person (fkp) showed a strong effect of the spouse with more education being the respondent. education according to cole et al. (2012), guiso and sodini (2013), calvet and sodini (2014), bannier and schwarz (2018), laurinaityte (2018), kaustia et al. (2019), bucher-koenen et al. (2021), bucherkoenen and knebel (2021), and hammer et al. (2022) the basic and main drivers of financial literacy are the formal level of education in school and the formal level of professional education. we combine the highest level of school education completed (scale from 6 = ”general or specific upper level secondary school permitting admission to university” to 1 = ”currently still a pupil” with 0 = no answer/no school degree) and financial services review, 31(4) 292 the highest level of professional education completed (scale from 7 = ”phd” to 1 = ”currently in vocational training or degree program” with 0 = no answer/no higher education degree) in our variable education (phf survey team, 2019b, pp. 30-31; oehler et al., 2022a). financial advice within the phf survey, households are asked about the financial advice obtained from the household’s main bank in the three years prior to the interview. we code their answers as a dummy variable (1 = advice, 2 = no advice; phf survey team, 2019a, p. 157). while we have no information on the frequency at which households consulted their banks, the content of these meetings or if the household ever acted upon the advice that it receives, this variable gives a good proxy about the household’s general willingness to seek professional advice (fey et al., 2020). in addition, it can be argued that the possible use of the bank’s consulting service also implies that a direct approach is made by the main bank concerned to its customers. investment experience another predictor variable for the cmp is the household’s own experience with investments in risky assets. the phf survey provides us with answers to the question on significant gains or losses from trading with financial assets in the three years prior to the interview (1 = gains, 2 = neither, 3 = losses; phf survey team, 2019a, p. 156). descriptive statistics and hypotheses in the structural model table 1 displays the descriptive statistics for the variables of the measurement concepts of the surveyed households. oehler & horn 293 table 1. descriptive statistics for the variables of the measurement concepts of the surveyed households (n = 4,538) mean median sd min max net wealth (in euros) 518,552 260,081 1671,257 -1271,354 92691,570 home loan saving and hmr mortgage outstanding (in euros) home loan saving 584 0 1,848 0 60,000 hmr mortgage outstanding 27,223 0 71,701 0 980,00 voluntary pension plan and whole life insurance (in euros) voluntary pension plan 22,804 1,400 48,246 0 662,600 whole life insurance 15,828 0 43,241 0 800,000 net income (in euros) 2,842 1,900 9,039 0* 149,600 satisfaction with life (present) 7.57 8 1.80 0 10 planning horizon (future) 2.20 2 1.53 0 5 age 58 59 16.08 19 90 gender 1.42 1 .49 1 2 education level of education in school 4.21 4 1.67 0 6 level of professional education 3.66 3 2.02 0 7 financial advice 1.71 2 .45 1 2 investment experience 1.94 2 .012 1 3 capital market participation (cmp) mutual funds .27 0 .44 0 1 bonds .07 0 .26 0 1 stocks .23 0 .42 0 1 financial literacy fin_lit_score 3.22 3 .95 1 4 fin_lit_training .27 0 .45 0 1 risk attitude riskfin 3.63 4 .53 1 4 riskgen 5.86 6 2.16 0 10 notes: table 1 displays the descriptive statistics for the variables of the measurement concepts of the surveyed households. for each variable we provide mean value (mean), median value (median), standard deviation (sd), minimum value (min), and maximum value (max). example: the mean value of fin_lit_score is 3.22 with a standard deviation of .95, the median is 3 with a range from 1 to 4. *variable net income: negative values were replaced by the value 0 because these were likely caused by inputations (n = 120). figure 1 illustrates an overview of the hypothesized relationships and the expected effects between the dependent variables (capital market participation (cmp), risk attitude, financial literacy) and the independent variables. financial services review, 31(4) 294 figure 1. hypotheses in the structural model consistent with the literature, we expect higher cmp with higher financial literacy and a lower degree of risk aversion. financial literacy should be higher among better educated, younger male fkps. risk aversion should decrease with higher financial literacy, net income and wealth, satisfaction in life, positive investment experiences, and financial advice. older and female fkps should show a higher degree of risk aversion. cmp should be higher among older, male fkps that received financial advice, have a higher degree of satisfaction in life, longer planning horizon, better investment experience, investments in voluntary pension plans / whole life insurances, and higher net income and wealth. home loan savings or an outstanding mortgage should have a negative influence on cmp. results structural model the results support the hypothesized relationships in our structural equation model presented in figure 1. we provide the results of the structural equation model in table 2. oehler & horn 295 table 2. results of the structural equation model (standardized estimates) panel a model fit indices value rmsea .041 panel b squared multiple correlations (smc) of the endogenous (dependent) variables value capital market participation (cmp) .59 financial literacy .64 risk attitude .39 panel c standardized total effects dependent variables independent variables value capital market participation financial literacy .43*** capital market participation risk attitude -.53*** risk attitude financial literacy -.47*** capital market participation net wealth .11*** capital market participation home loan saving and hmr mortgage outstanding -.06* capital market participation voluntary pension plan and whole life insurance .15*** capital market participation net income .00 capital market participation satisfaction with life (present) .08 capital market participation planning horizon (future) .09*** capital market participation age .17*** capital market participation gender -.09** capital market participation financial advice -.21*** capital market participation investment experience -.20*** financial services review, 31(4) 296 table 2 (continued). results of the structural equation model (standardized estimates) panel c (continued) standardized total effects dependent variables independent variables value financial literacy education .69*** financial literacy gender -.15*** financial literacy age -.25*** risk attitude net wealth -.09*** risk attitude net income -.04 risk attitude investment experience .19*** risk attitude financial advice .08*** risk attitude satisfaction with life (present) -.09*** risk attitude gender .21*** risk attitude age .14** notes: we provide the fit indices for the full model in panel a. given the benchmark values from the literature (see section 4.1) our model has a very good fit. panel b shows the squared multiple correlations (smc) of the latent endogenous variables (proportion of explained variance) in which 59% of the variance is in capital market participation, 39% is in risk attitude, and 64% in financial literacy are explained by the latent variables. according to the reference in the literature (see section 4.2), this is a substantial value for capital market participation and for financial literacy, and a moderate value for risk attitude. panel c displays the standardized total effects within the structural model. most of the coefficients are in the proposed direction and significant. for example, financial literacy has a great impact on capital market participation (.43) and risk attitude has a great impact on capital market participation (-.53), too. both are significant at the one per mill level. the symbols ***, **, and * denote significance at the one per mill, 1%, and 5% levels, respectively. the benchmarks for a good model fit that are recommended in the literature are below or equal to .06 for the rmsea index; below or equal to .08 for the srmr index; and above or equal to .9 for the agfi, ifi, and cfi. given that our results for the rmsea equal .041 (above the lower 90% confidence estimate: .039; below the upper 90% confidence estimate: .043) and the srmr equals .028, our model has a very good fit. capital market participation: financial literacy, risk attitude, and other predictors the squared multiple correlation (smc) of cmp is calculated as one minus the value of the respective residual term and amounts to 0.59. this is the proportion of the variance in participation that the latent variables can explain. according to the example in chin (1998), these results show a substantial smc. within this part of the model, financial literacy (.43) and risk attitude (-.53) are the dominant drivers of investments in risky assets (funds, bonds, publicly traded shares). the influence of both variables is strong and significant with p < .001. as hypothesized and in accordance with the literature (beshears et al., 2018; fey et al., 2020; kaustia et al., 2023; laurinaityte, 2018; thomas & spataro, 2015; von gaudecker, 2015), higher financial literacy leads to higher cmp. in addition to financial literacy, as proposed, the second major variable, risk attitude, has a strong oehler & horn 297 impact on capital market participation. the degree of risk aversion is a crucial determinant and acts as an intervening variable between the independent variables and cmp. as expected, lower risk aversion leads to higher cmp. households’ net wealth has a positive impact on cmp with p < .001. this is in line with the literature (calvet and sodini, 2014; fey at al. 2020; laurinaityte, 2018), but the magnitude of the influence is not strong (.11). in addition, our analysis shows that the impact of the two correcting variables in the context of wealth, the home loan saving and hmr mortgage outstanding, and the voluntary pension plan and whole life insurance, act in the expected direction, however, only with moderate impact (coefficients of -.06 and .15., respectively). the positive impact of the investment in voluntary pension plans and whole life insurances is statistically significant with p < .001. overall, housing or hmr influence cmp (cocco et al., 2005; efama, 2020; gomes et al., 2021; kaustia et al., 2023). however, housing discourages saving in financial assets not in a crucial manner. households’ perception that the investment in voluntary pension plans and whole life insurances act as so-called safe investments (efama, 2020), similar to deposits, may lead to the moderate influence on capital market participation. financial advice has a positive impact on cmp. households who use the consulting service of their main bank invest more in capital markets with p < .001. hence, our findings provide further support for those of von gaudecker (2015), chien and morris (2017), and fey et al. (2020). additionally, investment experience has the hypothesized impact. significant previous gains result in higher cmp, while significant losses have the opposite effect (.20). the influence is statistically significant with p < .001. age has a positive influence on cmp (.17, p < .001). it is plausible that investments in human capital and cmp compete for the limited resources of younger people (athreya et al., 2023; poterba & samwick, 2001). consistent with the literature, our results provide evidence that men invest more in capital markets than women (i.e., the gender gap) (fey et al., 2020; guiso & zaccaria, 2023; hanna et al., 2021). we will discuss this result more deeply in the light of literacy and education below. while satisfaction with life (present) hardly has an impact on capital market participation (.08, not significant), planning horizon positively influences the participation decision, however, with only small magnitude (.09, p < .001). the longer the planning horizon is aligned, the larger the cmp (ameriks & zeldes, 2004; barberis, 2000; calvet et al., 2007). contrary to our expectation, higher net income results not in a larger cmp. this could be due to the fact that households’ income stems from different sources, in particular from employment, self-employment, and pensions. financial literacy in order to cover both categories of financial literacy, we defined two measurement variables for operationalizing the extent of financial literacy: fin_lit_score, and fin_lit_training. their standardized total effects (i.e. their influences) are quite similar: .42 for fin_lit_score and .35 for fin_lit_training. overall, the squared multiple correlation (smc) of financial literacy amounts to .64, which means that more than 60% of the variance is explained by the assumed predictors education, gender, and age. according to chin (1998), these results show a substantial smc. our results indicate that higher schooling education and higher professional qualification will contribute to higher financial literacy. hence, education has a positive impact of more than 69% on financial literacy (standardized total effect: .69, p < .001), and increases with the level of school education completed and the level of professional qualification completed. age (.25) and gender (.15) also play a highly significant role, but the effects are weaker than the influence of education. as expected, financial literacy will be higher at a lower age (guiso & sodini, 2013; calvet et al., 2009; korniotis & kumar, 2011), and men seem to be more financial literate than women (bannier & schwarz, 2018; fey et al., 2020; guiso & zaccaria, 2023; hanna et al., 2021). financial services review, 31(4) 298 risk attitude in order to cover both categories of risk attitude, we define two measurement variables: riskfin, the self-assessment of risk aversion in the financial domain, and riskgen, the selfassessment regarding general risk-taking. their standardized total effects are substantially different: .67 on riskfin and .37 on riskgen. overall, the squared multiple correlation (smc) of risk attitude amounts to .39, which means that about 40% of the variance is explained by the assumed predictors financial literacy, investment experience, financial advice, gender, age, net wealth, net income, and satisfaction with life. according to chin (1998), these results show a moderate smc. our results indicate that higher financial literacy will contribute to lower risk attitude (i.e. lower degree of risk aversion) (standardized total effect: .47, p < .001). moreover, risk aversion decreases with the level of school education and the level of professional qualification. as expected, positive investment experience (i.e. significant gains) result in a lower degree of risk aversion, while significant losses have the opposite effect (.19, p < .001). with respect to the usefulness of professional advice, we find that financial advice leads to lower risk aversion, but the effect is rather weak (.08, p < .001). although age and gender play a significant role, their influences are weaker than the influence of financial literacy. as expected, risk aversion is lower for younger fkps (.14, p < .01) (calvet et al., 2009; guiso & sodini, 2013; korniotis & kumar, 2011), and women are more risk averse than men (.21, p < .001) (fey et al., 2020; guiso & zaccaria, 2023; halko et al., 2012; hanna et al., 2021). our results indicate that higher net wealth will contribute to lower risk attitude (i.e., lower risk aversion) (standardized total effect: .09, p < .001) (calvet & sodini, 2014; fey at al., 2020; laurinaityte, 2018). however, the latter effect is not strong in magnitude. contrary to our expectation, higher net income does not result in a lower degree of risk aversion. as mentioned above this could be due to the fact that household’s income stems from different sources, in particular from employment, self-employment, and pensions. calvet et al. (2021) estimate a lower degree of risk aversion for households with riskier labor income. further, reduced income in old age from low pensions may led to higher risk aversion. both effects could explain our findings. satisfaction with life (present) has only a small impact on risk aversion (.09, p < .001). discussion stock market investments are in the spotlight of the household finance literature, although realworld households make other financial decisions of higher relevance (kaustia et al., 2023). we widen the scope and include decisions related to voluntary pension plans, whole life insurance contracts, housing, and investments in risky assets other than stocks, (e.g., bonds and mutual funds). further, we provide a methodology that goes beyond regression analysis by employing a multi-factor structural analysis and apply it on data from a broad and representative survey of the german central bank. our sea allows us to investigate and quantitatively estimate complex relationship structures between manifest and/or latent variables. in contrast to regression analysis, sea tests several causal hypotheses simultaneously. our structural equation model explains about 60% of the variation in the capital market participation decision. yet, our study uses a cross-sectional dataset. hence, we discuss our findings with caution in terms of causality. the results show that although households’ financial literacy and risk aversion are the dominant drivers of investments in risky assets, further factors such as wealth, voluntary pension plans and whole life insurance contracts, financial advice, and investment experience play a remarkable role. financial literacy reduces risk aversion (i.e., the higher the financial literacy, the lower is the risk aversion). since cmp is related to investment experiences, financial advisors should take care of clients that suffered from losses, explain that temporary losses are part of risky investments, and that a complete divestment from capital markets would harm the clients’ future wealth accumulation severely. we do not find significant oehler & horn 299 results of housing on capital market participation, but age and gender play a role, directly, and indirectly via financial literacy and risk attitude. as proposed in our structural model, financial literacy has a strong impact on capital market participation (.43). our analysis additionally allows us to attribute this influence on cmp to a direct component and to an indirect effect via risk attitude. the direct effect amounts to the smaller part (.18 or 42% of the effect), while the indirect part via risk attitude is higher (.25 or 58%). the higher the financial literacy, the lower the risk aversion is; and lower risk aversion is associated with higher investments in risky assets. contrary to our expectation, higher net income results not in a larger cmp. therefore, we dig deeper and differentiate different income types instead of only looking at total net income. if we examine the entire sample of 4,538 households, only 1,202 households report investing in mutual funds (26.5%), only 7.4% use bonds (n = 334), and 22.6% (n = 1,026) invest in stocks. if we now differentiate the cmp according to the three types of income mentioned, we find that the cmp for households with employee income (n = 2,916) is rather lower than in the entire sample (funds: 26.5, bonds: 6.5, stocks: 21.7%). in contrast, households with pension income (n = 2,018) show a higher participation rate in stocks and bonds (funds: 26.3, bonds: 9.1, stocks: 24.3%; intervening effect from higher wealth with higher age), which explains the relative neutrality of the variable net income. households with income from self-employment (n = 913) show an even stronger cmp (funds: 30.7, bonds: 9.9, stocks: 27.3%). these results confirm the assessment of georgarakos and inderst (2011) which state that self-employed people are more likely to invest in stocks. moreover, the income variable is likely to be biased in the case of positive or negative wealth shocks (oehler et al., 2022a). for example, households that have inherited a substantial amount of money or assets but tend to have lower incomes are more likely to behave like high-asset households than low-income households. on the other hand, households with high income and low wealth (e.g. shortly after starting a job or getting divorced) are more likely to behave like households with low wealth by first building up precautionary liquidity as insurance against income shocks (job loss or similar) and to be able to cover unexpected expenses (efama 2020); additionally, measurement errors are claimed (calvet & sodini, 2014; guiso & sodini, 2012; fagereng et al., 2017). nevertheless, financial planners, financial counselors, and policy makers should try to convince more households with employee income to participate in capital markets, maybe with opt-out programs for capital market linked pension plans. another strand in the literature on household finance analyzes the question whether there is a gender effect on participation in the capital market (bannier & schwarz, 2018; fey et al., 2020; guiso & zaccaria, 2023; hanna et al., 2021). most of the studies conclude that the socalled gender gap disappears once risk aversion is considered (halko et al., 2012). consistent with the literature, our results provide evidence that men invest more in capital markets than women, but the total effect is not strong (.09). in the context of halko et al.’s (2012) findings, we take a deeper look at the phf data and our results show that women have a higher risk aversion than men (.21, significant at the one per mill level). the variable for risk in the financial domain, riskfin (median: 4, scale from 1 to 4), is equally distributed for men (n = 2,650) and women (n = 1,888). however, regarding the variable covering risk in a general context, riskgen (median: 6, scale from 0 to 10) we find a difference. while the sample of men shows a median of 5, women show a median value of 6. at the same time, men show a higher financial literacy (fin_lit_score: 4 vs. 3 correct answers; fin_lit_training: 31 vs. 23% passed a course). further differences concern the level of education in school (median 5 vs. 4) and the professional education completed (median 4 vs. 2). these differences add up to higher risk aversion among women as risk aversion decreases with the level of school education and the level of professional education. referring to the results in fey et al. (2020) and hanna et al. (2021), we additionally consider the marital status when analyzing a possible gender gap. we find a difference in the cmp depending on whether the fkp as respondent of the household is married, divorced, or widowed. financial services review, 31(4) 300 divorced and, to a smaller extant, widowed persons invest below average in funds, bonds, and stocks. in contrast, households with married fkp have above-average cmp. a more detailed analysis also shows that widowed and divorced fkps are mostly women (fey et al., 2020; georgarakos & inderst, 2011; hanna et al., 2021). policy makers should elaborate on measures to enable these women to participate in capital markets. many of these households are in challenging economic situations. the situation only gets worse when they do not receive a share of the equity premium. conclusion our multi-factor structural model explains about 60% of the variation of households’ capital market participation and, hence, solves major aspects of the so-called participation puzzle. although households’ financial literacy and risk aversion are the dominant drivers of investments in stocks, bonds, or mutual funds, further factors such as net wealth, voluntary pension plans and whole life insurance contracts, financial advice, and investment experience play a remarkable role. financial literacy reduces risk aversion (i.e., the higher the financial literacy, the lower is the risk aversion). we do not find significant results of housing on capital market participation, but age and gender play a role, directly, and indirectly via financial literacy and risk attitude. the socalled gender gap can be mainly explained by more risk averse women and their role as financially knowledgeable person (fkp), if at the same time it is taken into account that it is above average women who were interviewed as widowed or divorced fkp. in addition, any effort to promote the capital market participation, and the financial literacy to that end, should keep in mind that many households, but in particular younger ones, do not seem to be in a position to have financial funds available for capital market participation at all due to their tight budget (campbell, 2006; vissing-jorgensen, 2002, 2004). given the economic consequences of progressively higher inflation, but also given the nexus of physical health aspects and financial health, further analysis should also clarify the extent to which capital market participation may be permanently impaired. when policymakers and academics elaborate on concepts to increase the engagement of households in capital markets, they should be aware of households’ challenging economic situations as a determining factor. if policymakers and academics only focus on enhancing financial literacy without considering the households’ financial restrictions, the interventions would most probably fail. practitioners such as financial advisors should better point out to low net wealth households that participation in the capital market is already possible and feasible with diversified investments as low as five dollars/euros per month, for example, via exchange traded funds (d’acunto & rossi, 2020; horn & oehler, 2020; oehler et al., 2022a, 2022b; rossi & utkus, 2020). references altmann, k., bernard, r., le blanc, j., gabortoth, e., hebbat, m., kothmayr, l., schmidt, t., tzamourani, p., werner, d., & zhu, j. (2020). the panel on household finances (phf) microdata on household wealth in germany. german economic review, 21(3), 373–400. ameriks, j., & zeldes s. p. (2004). how do household portfolio shares vary with age?. working paper. https://business.columbia.edu/sites/default/f ilesefs/pubfiles/16/ameriks_zeldes_age_sept_ 2004d.pdf arbuckle, j. l. (2019). ibm® spss® amos™ 26 user’s guide. ibm. arrow, k. j. (1965). aspects of the theory of riskbearing. yrjö jahnssonin säätiö. athreya, k., ionescu, f., & neelakantan, u. (2023). stock market participation: the role of human capital. review of economic dynamics, 47(january), 1-18. aubram, t., kovarova-simecek, m., & wanzenried, g. (2016). financial literacy, pension planning, and investment behavior. working paper, uas st. pölten. oehler & horn 301 https://papers.ssrn.com/sol3/papers.cfm?abs tract_id=2834869 backhaus, k., erichson, b., & weiber, r. (2015). fortgeschrittene multivariate analysemethoden, eine anwendungsorientierte einführung (3rd ed.). springer. bannier, c. e., & schwarz, m. (2018). gender and education-related effects of financial literacy and confidence on financial wealth. journal of economic psychology, 67(c), 6686. barberis, n. (2000). investing for the long run when returns are predictable. journal of finance, 55(1), 225–264. bernheim, b. d., garrett, d. m., & maki, d. m. (2001). education and saving: the longterm effect of high school financial curriculum mandates. journal of public economics, 80(3), 435–465. beshears, j., choi, j. j., laibson, d., madrian, b. c. (2018, behavioral household finance. nber. working paper 24854. bilias, y., georgarakos, d., & haliassos, m. (2010). portfolio inertia and stock market fluctuations. journal of money, credit and banking, 42(4), 715-742 browne, m. w., & cudeck, r. (1993). alternative ways of assessing equation model fit. in k. a. bollen, k.a. and j. s. long (eds.), testing structural equation models. sage. bucciol, a., cavasso, b., & zarri, l. (2019). can risk-averse households make risky investments? the role of trust in others. scandinavian journal of economics, 121(1), 326–352. bucher-koenen, t., & knebel, c. (2021). finanzwissen und finanzbildung in deutschland—was wissen wir eigentlich? zew discussion paper, no. 21-016. https://ftp.zew.de/pub/zewdocs/dp/dp21016.pdf bucher-koenen, t., alessie, r., lusardi, a., & van rooij, m. (2021). fearless woman: financial literacy and stock market participation. zew discussion paper, no. 21-015. https://ftp.zew.de/pub/zewdocs/dp/dp21015.pdf bucher-koenen, t., lusardi, a., alessie, r., & van rooij, m. (2017). how financially literate are women? an overview and new insights. journal of consumer affairs, 51(2), 255–283. byrne, b. m. (1989). a primer of lisrel: basic applications and programming for confirmatory factor analytic model. springer. byrne, b. m. (2001). structural equation modeling with amos, eqs, and lisrel: comparative approaches to testing for the factorial validity of a measuring instrument. international journal of testing, 1(1), 55–86. byrne, b. m. (2016). structural equation modeling with amos: basic concepts, applications, and programming (3rd ed). taylor and francis. calvet, l., & sodini, p. (2014). twin picks: disentangling the determinants of risktaking in household portfolios. journal of finance, 69(2), 867–906. calvet, l., campbell, j., gomes, f., & sodini, p. (2021). the cross-section of household preferences. cepr discussion paper dp16105. https://cepr.org/publications/dp16105 calvet, l., campbell, j., & sodini, p. (2007). down or out: assessing the welfare costs of household investment mistakes. journal of political economy, 115(5), 707–747. calvet, l., campbell, j., & sodini, p. (2009). measuring the financial sophistication of households. american economic review, 99(2), 393–398. campbell, j. (2006). household finance. journal of finance, 61(4), 1553–1604. chatterjee, s., fan, l., jacobs, b., & haas, r. (2017). risk tolerance and goals-based savings behavior of households: the role of financial literacy. working financial services review, 31(4) 302 paper. https://papers.ssrn.com/sol3/papers.cfm? abstract_id=2918014 chaulk, b., johnson, p. j., & bulcroft, r. (2003). effects of marriage and children on financial risk tolerance: a synthesis of family development and prospect theory. journal of family and economic issues, 24(3), 257–279. chien, y., & morris, p. (2017), household participation in stock market varies widely by state. federal reserve bank of st. louis report. https://www.stlouisfed.org/publications/ regional-economist/third-quarter2017/household-participation-in-stockmarket-varies-widely-by-state chin, w. w. (1998). the partial least squares approach for structural equation modeling in g. a. marcoulides (ed.), modern methods for business research. lawrence erlbaum associates. choi, j. j., laibson, d., madrian, b., & metrick, a. (2009). reinforcement learning and savings behavior. journal of finance, 64(6), 2515–2534. cocco, j. f., gomes, f. j., & maenhout, p. j. (2005). consumption and portfolio choice over the life cycle. review of financial studies, 18(2), 491–533. cohn, r. a., lewellen, w. g., lease, r. c., & schlarbaum, g. g. (1975). individual investor risk aversion and investment portfolio composition. journal of finance, 30(2), 605-620 croson, r., & gneezy, u. (2009). gender differences in preferences. journal of economic literature, 47(2), 448-474 d’acunto, f., & rossi, a. (2020). roboadvising. cesifo working paper no. 8225. https://www.cesifo.org/en/publications/2 020/working-paper/robo-advising de, s., gondhi, n. r., & pochiraju, b. (2010). does sign matter more than size? an investigation into the source of investor overconfidence. indian school of business working paper. https://papers.ssrn.com/sol3/papers.cfm? abstract_id=1657926 deepak, c. a., singh, p., & kumar, a. (2015). financial literacy among investors: theory and critical review of literature. international journal of research in commerce, economics, and management, 5(4), 99–103. dixon, m. (2006). rethinking financial capability: lessons from economic psychology and behavioural finance. institute for public pension policy research. york. https://www.ippr.org/publications/rethin king-financial-capabilitylessons-fromeconomic-psychology-and-behaviouralfinance dohmen, t., falk, a., golsteyn, b., huffman, d., & sunde, u. (2017). risk attitudes across the life course. economic journal, 127(605), 95-116 dohmen, t., falk, a., huffman, d., schupp, j., sunde, u., & wagner, g. (2011). individual risk attitudes: measurement, determinants, and behavioral consequences. journal of the european economic association, 9(3), 522–550. dorn, d., & huberman, g. (2005). talk and action: what individual investors say and what they do. review of finance, 9(4), 437-481 eckel, c. c., & grossman, p. j. (2008). men, women and risk aversion: experimental evidence in c. r. plott and v. l. smith, v. l. (eds.), handbook of results in experimental economics. north holland/elsevier press efama. (2020). household participation in capital markets—assessing the current state and measuring future progress. european fund and asset management association. brussels. https://www.efama.org/sites/default/files /files/kpi%20report_final%20versio n%20%281%29_1.pdf oehler & horn 303 fagereng, a., gottlieb, c., & guiso, l. (2017). asset market participation and portfolio choice over the life-cycle. journal of finance, 72(2), 705–750. fey, j.-c., lerbs, o., schmidt, c., & weber, m. (2020). risk attitude and capital market participation: is there a gender investment gap in germany? discussion paper 20-080, zew. https://www.zew.de/publikationen/riskattitude-and-capital-marketparticipation-is-there-a-genderinvestment-gap-in-germany georgarakos, d., & inderst, r. (2011). financial advice and stock market participation. ecb working paper. https://www.ecb.europa.eu/pub/pdf/scp wps/ecbwp1296.pdf gomes, f., haliassos, m., & ramadorai, t. (2020). household finance. journal of economic literature, 59(3), 919-1000. guiso, l., & sodini, p. (2013). household finance: an emerging field in constantinidis, g., harris, m., stulz, r. (ed.) handbook of the economics of finance volume 2b. elsevier. guiso, l., & zaccaria, l. (2023). from patriarchy to partnership: gender equality and household finance. journal of financial economics, 147(3), 573-595.. hair, j. f., black, w., babin, b. j., & anderson, r. e. (2010). multivariate data analysis (7th ed.). prentice hall. haliassos, m., & bertaut, c. (1995). why do so few hold stocks? the economic journal, 105(432), 1110-1129 halko, m.-l., kaustia, m., & alanko, e. (2012). the gender effect in risky asset holdings. journal of economic behavior and organization, 83(1), 66–81. hammer, t., krahnhof, p., & zureck, a. (2022). correlation of financial literacy and demographics on home ownership in germany. working paper. https://papers.ssrn.com/sol3/papers.cfm? abstract_id=4211298 hanna, s. d., kim, k. t., lindamood, s., & lee, s. t. (2021). husbands, wives, and perception of relative knowledge about household finances. financial planning review, 4(1), e1110. haughton, d. m. a., oud, j. h. l., & jansen, r. a. r. g. (1997). information and other criteria in structural equation model selection. communications in statistics— simulation and computation, 26(4), 1477–1516. hermansson, c., jonsson, s., & liuc, l. (2022, the medium is the message: learning channels, financial literacy, and stock market participation. international review of financial analysis, 79(january), 101996. horn, m., & oehler, a. (2020). automated portfolio rebalancing: automatic erosion of investment performance? journal of asset management, 21(6), 489–505. hu, l., & bentler, p. m. (1999). cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. structural equation modeling, 6(1), 1–55. kaustia, m., conlin, a., & luotonen, n. (2023). what drives the heterogeneity in portfolio choice? the role of institutional, traditional, and behavioral factors. journal of banking & finance, 148(march), 106743. kline, r. b. (2005). principles and practice of structural equation modelling. the guilford press, korniotis g., & kumar, a. (2011). do older investors make better investment decisions? review of economics and statistics, 93(1), 244–265. laurinaityte, n. (2018). household financial risk tolerance in europe. zbw working paper. https://www.zbw.eu/econisarchiv/bitstream/11159/368371/1/ebp0 75847094_0.pdf lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women financial services review, 31(4) 304 fare? american economic review: papers & proceedings, 98(2), 413–417. lusardi, a., & mitchell, o. s. (2011). financial literacy around the world: an overview. journal of pension economics & finance, 10(4), 497–508. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52(1), 5–44. maccallum, r. c., browne, m. w., & sugawara, h. m. (1996). power analysis and determination of sample size for covariance structure modeling. psychological methods, 1(2), 130–149. malmendier, u., & nagel, s. (2011). depression babies: do macroeconomic experiences affect risk taking? quarterly journal of economics, 126(1), 373–416. mankiw, n., & zeldes, s. (1991). the consumption of stockholders and nonstockholders. journal of financial economics, 29(1), 97-112. mehra, r., & prescott, e. (1985). the equity premium—a puzzle. journal of monetary economics, 15(2), 145-161. merton, r. c. (1969). lifetime portfolio selection under uncertainty: the continuous-time case. review of economics and statistics, 51(3), 247-257 nosic, a., & weber, m. (2010). how risky do i invest: the role of risk attitudes, risk perceptions and overconfidence. decision analysis, 7(3), 282–301. oecd/infe. (2016). survey of adult financial literacy competencies. paris. https://www.oecd.org/daf/fin/financialeducation/oecd-infe-internationalsurvey-of-adult-financial-literacycompetencies.pdf oehler, a., & horn, m. (2019). does households' wealth predict the efficiency of their asset mix? empirical evidence. review of behavioral economics, 6(3), 249–282. oehler, a., & horn, m. (2021). behavioral portfolio theory revisited: lessons learned from the field. accounting and finance, 61(1), 1743–1771. oehler, a., horn, m., & wedlich, f. (2018a). young adults’ subjective and objective risk attitude in financial decision making: evidence from the lab and the field. review of behavioral finance, 10(3), 274–294. oehler, a., horn, m., & wendt, s. (2022a). investment in risky assets and financial market participation: does financial literacy matter? international review of economics (forthcoming) oehler, a., horn, m., & wendt, s. (2022b). investor characteristics and their impact on the decision to use a robo-advisor. journal of financial services research, 62(1), 91-125. oehler, a., horn, m., wendt, s., reisch, l. a., & walker, t. j. (2018b). young adults and their finances: an international comparative study on applied financial literacy. economic notes, 47(2-3), 305– 330. oehler, a., & werner, c. (2008). saving for retirement—a case for financial education in germany and uk? an economic perspective. journal of consumer policy 31, 253–283. phf survey team. (2019a). phf-codebookwave3-en-data. frankfurt. https://www.bundesbank.de/resource/blo b/798120/2b459c836ccd6eb62e0a13879 bad399e/ml/phf-codebook-wave3-endata.pdf phf survey team. (2019b). list of derived variables. frankfurt. https://www.bundesbank.de/resource/blo b/617436/7c64aa272805cf3c73d96cb71 261730e/ml/phf-derived-variablesdata.pdf poterba, j. m., & samwick, a. a. (2001). household portfolio allocation over the life cycle. in s. ogura et al. (eds.), aging issues in the united states and japan. nber. oehler & horn 305 rieger, m. o. (2020). how to measure financial literacy? journal of risk financial management, 13(12), 324. riley, w. b., & chow, k. v. (1992). asset allocation and individual risk aversion, financial analyst journal, 48(6), 32–37. rossi, a. g., & utkus, s. (2020). who benefits from robo-advising? evidence from machine learning. working paper. https://papers.ssrn.com/sol3/papers.cfm? abstract_id=3552671 savalei, v., & bentler, p. m. (2006). structural equation modeling, in r. grover and m vriens )eds.), the handbook of marketing research: uses, misuses, and future advances. sage. schermelleh-engel, k., moosbrugger, h., & müller, h. (2003). evaluating the fit of structural equation models: tests of significance and descriptive goodnessof-fit measures. methods of psychological research online, 8(2), 23–74. schoemaker, p. (1993). determinants of risktaking: behavioral and economic views. journal of risk and uncertainty, 6(1), 49–73. schooley, d. k., & worden, d. d. (1996). risk aversion measures: comparing attitudes and asset allocation. financial services review, 5(2), 87–99. shefrin, h. m., & statman, m. (2000). behavioral portfolio theory. journal of financial and quantitative analysis, 35(2), 127-151 thomas, a., & spataro, l. (2015). financial literacy, human capital and stock market participation in europe: an empirical exercise under endogenous framework. working paper. http://internationalpension-workshop.com/paperspdf/thomas.pdf van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101(2), 449–472. vissing-jørgensen, a. (2002). limited asset market participation and the elasticity of intertemporal substitution. journal of political economy 110(4), 825–853. vissing-jorgensen, a. (2004). perspectives on behavioral finance: does "irrationality" disappear with wealth? evidence from expectations and actions. nber macroeconomics annual, 18, 139–194. von gaudecker, h.-m. (2015). how does household portfolio diversification vary with financial literacy and financial advice? journal of finance, 70(2), 489– 507. von kalckreuth, u., eisele, m., le blanc, j., schmidt, t., & zhu, j. (2012). the phf: a comprehensive panel survey on household finances and wealth in germany. discussion paper. deutsche bundesbank, frankfurt am main. https://papers.ssrn.com/sol3/papers.cfm? abstract_id=2796868 wang, h., & hanna, s. (1997). does risk tolerance decrease with age? journal of financial counseling and planning, 8(2), 27–31. weiber, r., & mühlhaus, d. (2014). strukturgleichungsmodellierung. eine anwendungsorientierte einführung in die kausalanalyse mit hilfe von amos (2nd ed.). smartpls und spss. springer. weston, r., & gore, p. a. (2006). a brief guide to structural equation modeling. the counseling psychologist, 34(5), 719– 751. xiao, j. j. (2016). consumer financial capability and wellbeing in j. j. xiao (ed.), handbook of consumer finance research (2nd ed.). springer. xiao, j. j., & o'neill, b. (2016). consumer financial education and financial capability. international journal of consumer studies, 40(6), 712–721. do as i tell you, not as i do: financial advisors and personal financial decision-making negin azamian1, kristine beck1, hsin-hui chiu1, inga timmerman1,* 1department of finance, financial planning and insurance, california state university, northridge, california abstract we describe the financial behavior of financial advisors and whether they follow the advice they give clients. we focus on the following areas of comprehensive financial planning as they relate to advisor behavior: (1) cash flow, (2) debt, (3) retirement planning, (4) investments, and (5) estate planning. the primary goal is to investigate whether financial planners practice what they preach. a secondary goal is to identify the characteristics associated with the advisors that best plan their own financial lives. we find that financial advisors generally follow their own advice; as a group they are more likely to be prepared for retirement, have less debt, higher liquidity, covered insurance needs, and have an estate plan in place. © 2022 academy of financial services. all rights reserved. keywords: financial advisors; retirement planning; investments; debt; estate planning 1. introduction and motivation the use of financial advisors has been associated with better preparedness for retirement, higher financial confidence, and an increased sense of financial well-being. for example, a 2014 survey conducted by the insured retirement institute claims that baby boomers who use financial advisors are twice as likely to feel confident about their retirement savings as those who do not use an advisor.1 a strand of literature suggests that financial advisors provide value to their clients with regard to behavioral biases and investments. shapira and venezia (2001) analyze investment patterns of clients of a major israeli brokerage house and compare investment decisions of those making independent decisions to those managed by professionals. they conclude corresponding author: tel.: 818-677-4615; fax: 818-677-6079. e-mail address: inga.timmerman@csun.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 57–68 professional training and experience may reduce judgmental biases, and that professionally managed accounts are more diversified and less correlated with the market/more profitable than those of independent accounts. cici, kempf, and sorhage (2017) compare the taxavoidance behavior of investors who operate under the guidance of financial advisors with investors who do not have financial advisors. they document tangible benefits in the form of useful tax-management advisory services to mutual fund investors, helping those investors engage in tax-avoidance strategies such as tax-loss selling. financial planners also help clients match self-indicated risk tolerance with ownership of investment assets. park and yao (2016) suggest that financial planners provide significant value to households on the consistency of their financial risk attitude and behavior. park and yao (2016) find that financial planners provide significant value to households by matching financial risk attitude to actual savings and investments behavior. in addition, professional financial planners can still benefit from a third-party assessment to provide more objective recommendations on personal financial planning or to de-bias behavioral issues. common behavioral biases include under-diversification, local and home bias, and the disposition effect.2 for instance, seasholes and zhu (2010) point out that investors who demonstrate local stock bias do not earn superior returns, and. hoechle et al (2017) find that financial planners help better diversify portfolios and reduce local bias. at the same time, another strand of literature questions the effectiveness of financial advisors to bring value. this could be explained by the agency problem; the inherent conflict of interest in the advisor-client relationship that makes it hard for the advisor to align her interests with the client. although advisors should act in the client’s best interest, some research questions advisor motivations. for example, hackethal, haliassos, and jappelli (2012) point out that advisors with commission-based incomes prefer to devote time to customers likely to trade on a bigger scale. mullainathan, noeth, and schoar (2012) show that advisors fail to de-bias their clients and often reinforce biases that are in their interests. they also find that advisors encourage return-chasing behavior and push for actively managed funds that have higher fees, even if the client starts with a well-diversified, low-fee portfolio. hoechle et al. (2017) examine the performance of advised and independent trades by comparing them trade-by-trade with in-person analysis. they conclude financial planners help reduce the behavioral biases to which retail investors are subject, but that advised trades still perform worse than independent trades. similarly, chalmers and reuter (2020) use changes in retirement plans to examine the choice between broker advice and target date funds. they find that brokers recommend higher-commission options and that investors most worried about bear market risk will invest in target date funds when available, with better outcomes than the broker-advised portfolios. a 2013 survey from the society of actuaries shows that 52% of pre-retirees and 44% of retirees consult with a financial advisor.3 alyousif and kalenkoski (2017) examine five types of financial advice sought by the general population: debt counseling, saving/investment, mortgage/loans, insurance, and tax planning. they find no significant differences across subsamples defined by gender, age, and financial literacy and that income and risk tolerance are positively related to demand for financial advice. they also find that low awareness of financial knowledge, perhaps a proxy for self-confidence, and financial fragility decrease the probability of seeking financial advice. 58 n. azamian et al. / financial services review 30 (2022) 57–68 overall, financial literacy is found to have a significant impact on portfolio diversification and investment outcomes. a high degree of financial literacy increases the usage of financial planning services, although investors who seek advice may not strictly follow guidance and therefore do not improve their portfolio efficiency. von gaudecker (2015) examines portfolio diversification and finds a significant relationship between good investment outcomes and financial literacy and/or reliance on professional financial advice. compared with those groups, households with below-median financial literacy that trust their own decision-making capabilities underperform those who do not. calcagno and monticone (2015) analyze the effect of investors’ financial literacy on their decision to seek financial advice. they conclude a high degree of financial literacy increases the likelihood that investors consult with financial advisors. battacharya et al. (2012) use german brokerage data to examine the efficacy of unbiased investment advice. they find investors who most need financial advice are the least likely to obtain it. in addition, their research suggests that the 5% of investors who do seek advice barely follow the advice and do not significantly improve portfolio efficiency. financial planners are a group of individuals with a high degree of financial literacy. according to nofsinger and varma (2007), who survey over 100 financial planners to assess their reasoning mode, financial planners are more analytical than the general population with regard to intertemporal choices, risk aversion and preferences, and framing focus. however, many financial planning professionals do not have business plans, retirement plans, or successions plans in place, and doviak (2016) discusses how advisors struggle to cope with emotional stress using behavioral finance.4 it is useful to examine whether financial planners handle their personal finances as well as they advise their clients and whether they follow through with execution plans. this article investigates how advisors make their own financial decisions and whether the advice they give is consistent with their own behavior. we focus on the following areas of comprehensive financial planning as they relate to advisor behavior: (1) cash flow, (2) debt, (3) retirement planning, (4) investments, and (5) estate planning. a secondary goal is to identify characteristics associated with the advisors that plan their own financial lives according to best financial practices. we examine financial planners’ financial decisions with respect to debt and savings, and whether they handle their own personal finances efficiently. in addition, we investigate whether financial planners rely on professional services, such as hiring tax professionals or financial planners. we find that financial planners mostly preach what they practice. as a group, planners are more likely to be prepared for retirement, have less debt, higher liquidity, covered insurance needs, and an estate plan in place. as a result, the general population could benefit financially by hiring planners. these results are consistent with those found by linnainmaa, melzer, and previtero (2021), who examine a sample of financial planners in canada. they conclude that the personal investments of advisors are similar to client advice, even when the advice may be expensive and inefficient. dvorak (2015) also finds that advisors’ plans are comparable to their clients’ plans; they tend to hold identical funds and use the same fund families and fund categories. outlaw and outlaw (2017) focus on the investment aspect and compare the advisors’ own trading activity with that of their clients. they n. azamian et al. / financial services review 30 (2022) 57–68 59 find that advisors do their best for the clients, as they do for themselves, but sales incentives may influence the quality of advice. 2. survey design we collect information from financial advisors via survey in summer 2018.5 advisors are recruited through targeted facebook pages (such as xypn), napfa, and fpa. the responses are anonymous and voluntary, with no renumeration. the survey consists of 33 questions, categorized as follows: demographic information, cash flow questions (budget existence and use, emergency savings and consumer debt), insurance questions (need assessment and implementation), estate planning (existence and household preparedness for emergencies), investments (existence and decision-making in terms of time and investment style), taxes (knowledge and preparation), and an overall assessment of the satisfaction with past financial choices. we obtained 124 complete responses during the summer of 2018, a response rate of 82%. by design, the sample is biased towards planners who do not exclusively charge commissions. our sample mirrors the overall gender distribution of financial advisors well. of the respondents, 68% are male and 83% are married, 34% of the respondents are under 34, and 5% are over 65. 88% consider themselves a comprehensive financial advisor and 63% have earned the certified financial planner (cfp) designation. 3. results 3.1. descriptive statistics table 1 presents descriptive statistics by category. for cash flow, our expectation is that everyone will have a budget, given how consistently this topic is enforced in financial planning. we find that 67% of advisors do have a personal budget, but out of those, despite having a budget in name, 20% do not track or enforce it consistently. similarly, 67% of u.s. households prepare a monthly budget.6 overall, only 47% of planners elevate their own budget to the same level of responsibility they ask their clients to follow.7 in terms of liquidity, 9% of the respondents have less than $3,000 in liquid assets saved for emergencies, 23% have somewhere between $3,000-$10,000, and 23% have more than $50,000. although liquidity is an important component of financial planning, we do not have an expectation for an optimal level. still, the 9% that have less than $3,000 accessible for emergencies is much less than the typical advice of three to six months of liquid assets. by comparison, 45% of u.s. adults have no savings, and 70% have less than $1,000 in savings.8 as a side note, we ask questions about the ability of partners to find the financial records of their spouse. for example, if a spouse were to die, would the second partner be able to access all the accounts, and know who to call for pension plans and insurance, and so forth? 60 n. azamian et al. / financial services review 30 (2022) 57–68 given that one person in the relationship is a financial advisor, we expect the respondents to be able to easily access the information for their spouse. we find that 67% of advisors have a document in place for their spouse’s accounts but 10% do not know what is available or where to access files. on the flip side, 54% of respondents have a document in place for their non-advisor spouse, 21% have no formal document but have shared the accounts and accessibility, and 25% have not prepared the information for their spouse. across older americans, 32% have not informed their family where to find legal, medical, and financial documents.9 we ask advisors to compare the amount of time and effort they spend on their client portfolios compared with time spent on their own portfolios: 56% of advisors spend the same amount of time on their client portfolios as they do on their own, and only 8% spend significantly more time. interestingly, 35% stated that they spend significantly less time on their own portfolio. because financial advisors sell professional services and experience, we also expect advisors to use such services. even though advisors have the expertise to manage their own finances, adding a neutral, unbiased third party would be very beneficial for the behavioral aspect of money management. to assess this topic, we ask if they (1) prepare their own taxes table 1 descriptive statistics variables mean standard deviation min max n gender 0.6820 0.4672 0 1 123 budget 0.8644 0.7151 0 2 118 age 2.3220 1.2529 1 5 124 marriagestatus 0.8307 0.37658 0 1 124 advisortype 0.8065 0.60723 0 1 124 cfpcode 0.6363 0.4830 0 1 121 liquidassets 3.3559 1.3173 0 5 118 income 3.5213 1.3808 1 5 117 note. independent variables are as follows: male is equals 1 and otherwise, 0. budget is represented by a code where 1 identifies advisors who have a personal budget and review is regularly, 2 represents advisors who have a budget but do not review is regularly and 0, advisors who do not have a formal budget for themselves. age is represented by a code from 1 to 5 where one is less than 35, two is 36–45, three is 46-64, four is 55-64, and five is over 65. marriagestatus is equal to 1 if married and 0 otherwise. advisortype is equal to 1 if the person is a comprehensive financial planner and 0 otherwise. cfpcode is equal to one if the person is a cfp and 0 otherwise. liquid assets range between 1 to 5, depending on the amount of available assets. one is less than $3,000, two is between $3,000-$10,000, three is $10,000-$20,000, four is $20,000 to $50,000, and five is more than $50,000. income is a range between one and five where one represents less than $50,000 per year, two represents $50,000-$100,000, three represents $100,000-$150,000, four represents $150,000-$200,000, and five represents more than $200,000. n. azamian et al. / financial services review 30 (2022) 57–68 61 and (2) use a financial advisor themselves; 45% of respondents have someone else to do their taxes while 55% prepare their own. of the advisors who prepare their own taxes, only 33% have a tax qualification like an ea or cpa. by comparison, only 10% of advisors have their own financial advisor. for the general population, 33% file their own taxes and 75% manage their own finances.10,11 we also ask advisors if they had made a financial decision in the past that was different from the advice they disperse to their clients; 50% of the respondents answered yes. the most common mistakes are (1) not avoiding debt, and particularly accumulating credit card debt, (2) having investments they would not include in their clients’ portfolios, (3) buying a house with a very low-down payment while on a strict budget (house poor), and (4) cashing out a roth ira. exhibit 1 summarizes the differences in financial planning behavior between our sample of financial advisors and the general population 3.2. multivariate analysis table 2 presents the regression results. the dependent variables are as follows: the existence and enforcement of a budget in model 1, the amount of liquid assets in model 2, the assessment of insurance needs in model 3, the amount of credit card debt in model 4, and the existence of an estate plan in model 5. models 6 and 7 break the existence of an estate plan into questions regarding whether the spouse knows about how to access financial sources: cnbc, debt.com, gobankingrates, people press, wells fargo 62 n. azamian et al. / financial services review 30 (2022) 57–68 t ab le 2 f in an ci al p la n n in g v ar ia b le s m o d el 1 co ef fi ci en t p -v al u e m o d el 2 co ef fi ci en t p -v al u e m o d el 3 co ef fi ci en t p -v al u e m o d el 4 co ef fi ci en t p -v al u e m o d el 5 co ef fi ci en t p -v al u e m o d el 6 co ef fi ci en t p -v al u e m o d el 7 co ef fi ci en t p -v al u e g en d er �0 .0 0 2 �0 .2 9 4 �0 .3 1 5 �0 .0 3 8 0 .1 9 2 �0 .3 2 2 �0 .0 6 2 (0 .9 9 1 ) (0 .3 3 2 ) (0 .1 0 3 ) (0 .8 4 2 ) (0 .4 0 4 ) (0 .1 1 2 ) (0 .7 2 0 ) a g e �0 .1 2 7 0 .1 8 4 0 .0 3 4 0 .0 6 3 0 .3 1 8 0 .0 0 1 0 .0 2 6 (0 .0 5 2 )* (0 .0 6 4 )* (0 .6 8 7 ) (0 .3 7 4 ) (0 .0 0 0 )* * * (0 .9 9 0 ) (0 .7 5 8 ) m ar ri ag es ta tu s �0 .2 5 6 0 .5 1 5 �0 .3 4 4 �0 .3 8 4 0 .0 9 5 (0 .2 9 2 ) (0 .4 1 5 ) (0 .4 0 1 ) (0 .2 4 7 ) (0 .8 4 0 ) a d v is o rt y p e �0 .3 4 2 �0 .0 1 1 �0 .3 0 5 �0 .1 9 2 �0 .5 8 3 �0 .8 6 8 �0 .5 8 2 (0 .2 6 0 ) (0 .9 7 8 ) (0 .3 1 6 ) (0 .3 7 5 ) (0 .1 1 5 ) (0 .0 0 1 )* * * (0 .0 2 9 )* * c f p c o d e 0 .0 2 0 0 .4 3 2 �0 .6 6 1 0 .4 5 9 0 .0 8 8 0 .2 0 3 �0 .2 6 5 (0 .9 1 8 ) (0 .2 3 0 ) (0 .0 0 5 )* * * (0 .0 4 7 )* * (0 .7 1 1 ) (0 .0 1 4 )* * (0 .1 6 9 ) l iq u id a ss et s �0 .1 0 5 0 .0 9 2 �0 .4 7 3 0 .1 4 7 0 .2 0 3 0 .1 1 9 (0 .1 6 9 ) (0 .1 6 3 ) (0 .0 0 0 )* * * (0 .1 1 6 ) (0 .3 0 3 ) (0 .1 6 9 ) in co m e 0 .0 5 7 0 .3 0 4 �0 .0 2 9 �0 .0 0 6 �0 .0 1 7 0 .2 3 1 0 .0 0 4 (0 .3 8 2 ) (0 .0 2 9 )* * (0 .6 6 5 ) (0 .9 2 9 ) (0 .8 4 8 ) (0 .7 8 5 ) (0 .9 5 2 ) c o n st an t 1 .9 0 4 1 .4 3 1 2 .3 7 6 2 .3 2 8 1 .3 3 0 1 .6 7 3 1 .8 5 3 (0 .0 0 0 )* * * (0 .1 3 1 ) (0 .0 0 0 )* * * (0 .0 0 0 )* * * (0 .0 7 9 )* (0 .0 1 0 )* * * (0 .0 0 3 )* * * o b se rv at io n s 7 9 7 9 7 9 7 8 7 9 7 2 7 2 p ro b > f 0 .0 4 8 9 * * 0 .0 0 1 0 * * * 0 .0 0 6 6 * * * 0 .0 0 0 0 * * * 0 .0 0 0 2 * * * 0 .0 0 2 3 * * * 0 .0 3 6 3 * * r 2 0 .1 2 7 8 0 .2 3 4 5 0 .2 2 6 8 0 .4 1 2 7 0 .2 6 9 8 0 .2 0 7 1 0 .1 1 9 4 n o te . t h e d ep en d en t v ar ia b le s ar e as fo ll o w s: t h e ex is te n ce an d en fo rc em en t o f a b u d g et in m o d el 1 , th e am o u n t o f li q u id as se ts in m o d el 2 , th e as se ss m en t o f in su ra n ce n ee d s in m o d el 3 , th e am o u n t o f cr ed it ca rd d eb t in m o d el 4 , an d th e ex is te n ce o f an es ta te p la n in m o d el 5 . m o d el s 6 an d 7 b re ak th e ex is te n ce o f an es ta te p la n in to th e q u es ti o n s w h et h er th e sp o u se k n o w s ab o u t h o w to ac ce ss fi n an ci al in fo rm at io n in ca se o f th e ad v is o r’ s d ea th , an d w h et h er th e ad v is o r k n o w s h o w to ac ce ss th e sp o u se ’s fi n an ci al in fo rm at io n in ca se o f h is o r h er d ea th . in d ep en d en t v ar ia b le s ar e as fo ll o w s: m al e is eq u al s 1 an d o th er w is e, 0 . a g e is re p re se n te d b y a co d e fr o m 1 to 5 w h er e o n e is le ss th an 3 5 , tw o is 3 6 -4 5 , th re e is 4 6 -6 4 , fo u r is 5 5 -6 4 , an d fi v e is o v er 6 5 . m ar ri ag es ta tu s is eq u al to 1 if m ar ri ed an d 0 o th er w is e. a d v is o rt y p e is eq u al to 1 if th e p er so n is a co m p re h en si v e fi n an ci al p la n n er an d 0 o th er w is e. c f p c o d e is eq u al to o n e if th e p er so n is a c f p an d 0 o th er w is e. l iq u id as se ts ra n g e b et w ee n 1 to 5 , d ep en d in g o n th e am o u n t o f av ai la b le as se ts . o n e is le ss th an $ 3 ,0 0 0 , tw o is b et w ee n $ 3 ,0 0 0 -$ 1 0 ,0 0 0 , th re e is $ 1 0 ,0 0 0 -$ 2 0 ,0 0 0 , fo u r is $ 2 0 ,0 0 0 to $ 5 0 ,0 0 0 , an d fi v e is m o re th an $ 5 0 ,0 0 0 . in co m e is a ra n g e b et w ee n o n e an d fi v e w h er e o n e re p re se n ts le ss th an $ 5 0 ,0 0 0 p er y ea r, tw o re p re se n ts $ 5 0 ,0 0 0 -$ 1 0 0 ,0 0 0 , th re e re p re se n ts $ 1 0 0 ,0 0 0 -$ 1 5 0 ,0 0 0 , fo u r re p re se n ts $ 1 5 0 ,0 0 0 -$ 2 0 0 ,0 0 0 , an d fi v e re p re se n ts m o re th an $ 2 0 0 ,0 0 0 . n. azamian et al. / financial services review 30 (2022) 57–68 63 ce 1-hour general principles of financial planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. cfp board recently adopted revisions to several provisions of its ce policies, including changing the minimum number of questions for self-study assessments from 10 to 5 per full ce credit hour. therefore, financial services review ce exams will have 5 questions. ce credit for this issue expires january 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in “using investor utility to determine portfolio choice with reits” by feng, jones, & allen, investors with lower risk aversion ____ . a. obtain higher risk-return benefits from adding reits b. obtain lower risk-return benefits from adding reits c. obtain no risk-return benefits from adding reits d. obtain equal risk-return benefits from adding reits, compared to fixed income 2. in feng, jones, & allen, what benefit do reits provide individual investors? a. ability to purchase mortgage insurance in the public market b. ease of investing in real estate using publicly traded shares c. ease of locking in real estate prices using publicly traded shares d. ability to competitively purchase specific real estate in the public market 3. feng, jones, & allen use ____ to examine portfolio decisions regarding reits. a. risk measures, such as vix b. risk of the s&p 500 c. risk-neutral probabilities d. risk preferences 4. in “financial, demographic and psychological differences between chapter 13 bankruptcy fil and non-filers,” by kehiaian, williams, and bir the following psychological variables are found be significantly different for filers and non-filers except: a. self-efficacy b. self-control c. motivation d. locus of control 5. kehiaian, williams, and bird, found that for chapter 13 bankruptcy filers, a. males and females file equally. b. blacks are less likely to file than white c. homeowners are less likely to file than renters. d. single people are less likely to file than married people manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) call for papers: virtual conference deadline extended due july 1, 2020 the academy of financial services 34th annual meeting september 29-30, 2020 virtual conference the academy of financial services will hold its annual conference in conjunction with the fpa be annual conference. in light of recent events, we have restructured pricing for this year’s virtual annual meeting: $199 for academics and practitioners and $99 for students. the afs conference will feature speakers, symposia, several special sessions, posters, and a reception. among them, we will introduce a new panel session for phd students, highlighting how to best navigate the job market and a panel for women in financial services. with the generous support of our sponsors, the academy has awarded several best paper awards during past meetings and we anticipate continuing best paper awards in 2020. we will continue with our emerging scholar award to a current graduate student for promising research work on a paper or poster presented at the conference. in addi�on, in 2020, we will ini�ate a new, program directors track. the goal is to allow program directors to present and discuss program issues and best practices in a panel environment, such as “working with your university’s foundation”, “capstone course cases: what’s the right content?”, “understanding career paths and student fit” and “developing a passionate program in a box: scholarships, competitions, student organizations”. we welcome other panel topics deemed beneficial to program directors. submission information: research papers and abstracts covering all aspects of individual financial management and education are sought for inclusion in the program. papers in the areas of estate planning, insurance, tax accounting aspects of financial planning, investments, and retirement planning are encouraged. proposals for panel discussions and tutorials devoted to current issues in individual financial management or the practice of financial planning will also be considered for inclusion in the program. several sessions will be registered for continuing education (ce) credit with the cfp ® board. submit your paper or abstract here: https://proposalspace.com/calls/d/1181 submissions are due july 1st the review period ends on july 31st, 2020 with the selection period and formulation of the agenda estimated to be completed by august 31st, 2020. notice of acceptance as an oral session or a poster is targeted for september 5 th , 2020. note that the terms and conditions of this call-for are outlined in the online submission form. only accepted presentations are included in the subsequent proceedings, which are posted on the afs website. thus, the proceedings publication is refereed in order to accommodate the rules of the american association of intercollegiate schools of business-international (aacsb) on table 2-1 (intellectual contributions). for further information: visit the afs website at academyfinancial.org that will be frequently updated. for content ques�ons contact program chair, dr. terrance k. mar�n jr. at terrance.martin@uvu.edu manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) how risky is your retirement income risk model? patrick j. collins, ph.d., clu, cfaa, huy lam, cfab,*, josh stampfli, ma, eesorc aschool of management, university of san francisco, 2130 fulton street, san francisco, ca 94117, usa bschultz collins, inc., 455 market street, #1250, san francisco, ca 94105, usa c151 rock creek lane, scarsdale, ny 10583, usa abstract adequately sustaining lifetime income is a critical portfolio objective for retired investors. this article provides a brief review of various retirement income modeling approaches including historical back testing, monte carlo simulations, and other more advanced risk modeling techniques. implausible assumptions underlying common risk models may mislead investors concerning the risk and return expectations of their retirement investment strategies. we compare risk models, evaluate their credibility, and demonstrate how an oversimplified model may distort the risks retired investors face. differences in sustainability rates are stark: 4% failure at the low end versus 49% failure at the high end. the article ends with general comments regarding model risk and practitioner investment advice. © 2015 academy of financial services. all rights reserved. keywords: the 4% rule; monte carlo simulation; portfolio sustainability; retirement income; risk modeling 1. introduction sustainability of adequate lifetime income is a critical portfolio objective for retired investors. commentators often define sustainability in terms of (1) a portfolio’s ability to continue to make withdrawals throughout the applicable planning horizon, or (2) a portfolio’s ability to fund a minimum level of target income at every interval during the planning horizon. the first approach focuses on the likelihood of ending with positive wealth, or, if wealth is depleted before the end of the planning horizon, on the magnitude and duration of * corresponding author. tel.: �1-415-354-3704; fax: �1-415-291-3015. e-mail address: huy@schultzcollins.com (h. lam) financial services review 24 (2015) 193–216 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. the shortfall; the second focuses on the likelihood of consistently meeting all period-byperiod minimum cash flow requirements. risk models help advisors assess a portfolio’s ability to provide adequate cash flow throughout retirement. conclusions about cash flow sustainability are usually reached by determining the likelihood that withdrawals (fixed amounts, percentage of corpus, or “dynamic”) can be maintained for either deterministic or stochastic time periods under various asset allocations and longevity assumptions. expressed in terms of a venn diagram, portfolio success lies at the intersection of the three elements in fig. 1: “sustainability” differs from the concept of “feasibility.” feasibility depends on an actuarial calculation to determine if a retirement income portfolio is technically solvent—current market value of assets equals or exceeds the stochastic present value of the cash-flow liabilities. if the current market value of assets is less than the cost of a lifetime annuity, the targeted periodic withdrawals exceed the resources available to fund them. in short, the portfolio violates the feasibility condition. determination of the feasibility of retirement income objectives is not subject to model risk because the determination rests on current observables—annuity cost versus asset value—rather than on projections of financial asset evolutions and the distribution of longevity. although it is important to track both risk metrics—sustainability and feasibility—as part of prudent portfolio surveillance and monitoring, the remainder of this article focuses on the sustainability or shortfall probability risk metric. fig. 1. 194 p.j. collins et al. / financial services review 24 (2015) 193–216 2. sources of retirement income model risk probability assessments are only as good as the models upon which they are based—that is to say, assessments are prone to “model risk.” in general, model risk arises from several sources: 1. variables of interest: projected model outcomes may differ considerably depending on the range of input variables. health shocks, inheritance expectations, life insurance availability, and other variables may or may not improve the calculated probability of a successful retirement investment and consumption strategy. 2. model sensitivity to changes in the value of input variables: output can be notoriously sensitive to small changes in input values; likewise, compounding over long planning horizons can produce large differences in outcome values and likelihoods given small changes in input values. 3. model structure: deterministic inputs will likely project outcomes different than those generated by a model that treats investment, inflation, and longevity variables stochastically. the nature of the model’s covariance matrix may be an additional source of estimation error. 4. model assumptions: the choice of utility function can influence model output. for example, assumption of constant relative risk aversion may rank outcomes differently from those flowing from models assuming a hyperbolic risk aversion function.1 likewise, the functional form chosen to generate inflation or investment returns will often influence investment recommendations. econometricians often discuss model risk in terms of specification error. errors may arise as a result of including irrelevant variables in the model, failure to incorporate relevant variables, and inaccurate estimation of input variable values. specification errors may account for different models producing dissimilar outputs when considering the same problem. this is an underlying reason why any single retirement income risk model may be unable to provide a good assessment of retirement risk.2 this article focuses on postretirement shortfall risk from assessments derived from modeling investment, longevity, and inflation related risks. model-based probability assessments rely, in part, on outputs generated by computer algorithms that approximate, with varying degrees of accuracy, the processes that drive financial asset price changes. thus, any assessment of the sustainability of a retirement income investment program should not over rely on outputs produced by a single risk model; and, when using model outputs to monitor the portfolio, practitioners should take care to ascertain that the model is academically defensible. beyond the above-listed sources of model risk are two other considerations: 1. bonini’s paradox: models that explain the workings of complex systems are seemingly impossible to construct. as a model of a complex system becomes more complete, it becomes less understandable; for it to be more understandable it must be less complete and, therefore, less accurate. when a model becomes accurate, it is just as difficult to understand as the real-world processes it represents. 195p.j. collins et al. / financial services review 24 (2015) 193–216 2. ambiguity in how the investor should preference rank heterogeneous outcomes: outcomes from equally credible models may differ significantly even when each model uses identical inputs and input values. interpretation of calculated results becomes difficult and there is no clear “winning strategy” or preferred solution path. 3. modeling approaches there are a number of modeling approaches to ascertain the likelihood that a portfolio’s investment strategy is suitable to its cash flow requirements—that is, estimating the probability that a jointly determined asset allocation/retirement spending strategy is sustainable throughout the planning horizon absent significant, and possibly difficult to implement, midcourse corrections: y analytic formulae (closed form solutions usually within a life-cycle model context) y historical back testing of empirical returns y bootstrapping (reshuffled historical returns) y monte carlo simulation (assuming a two-parameter normal or lognormal distribution) y simulating non-normal distributions (student’s t, pareto, truncated levy flight, gamma, logistic, exponential, etc.) y vector auto-regression y regime-switching simulation models. a brief discussion of each method follows. 3.1. analytic formula the analytic formulae approach attempts to solve the sustainability question by (1) describing retirement planning as complex systems of equations, (2) transforming descriptive formulae with algebraic manipulation to achieve closed formed solutions, and (3) plugging in assumed values for the independent variables to arrive at a final conclusion. most formulaic solutions pile assumption upon assumption regarding the functional forms and parameterized values of the numerous variables included in the model. to make the mathematics tractable, many models assume that stock returns are independent, identically distributed, and, as a consequence, that the underlying distribution of stock returns is stable. input variables may include rates of returns and volatilities for financial assets, inflation rate behavior, interest rate term structure, and the form of an investor’s utility function. specific input variables are often estimated using econometric techniques that, although critically important to an assessment of a model’s credibility, are, nevertheless, tangential to the focus of this article.3 generally speaking, most studies of stock price change reject the hypothesis that the return series is normally distributed, with the most often cited deficiency as a failure to capture the “volatility of volatility.”4 finding the closed formed solutions to analytic formula models is a daunting task that often requires applying highly sophisticated integral calculus or solving intricate partial differential equations.5 huang, milevsky, and wang (2004), for example, use a formulaic 196 p.j. collins et al. / financial services review 24 (2015) 193–216 approach to conclude that an inflation adjusted withdrawal rate equal to 3.33% of a 65-year-old investor’s starting portfolio value exhibits a 95% sustainability rate. despite its mathematical elegance, application of the analytic formula approach to sustainability analysis has been quite limited. solutions often seem enigmatic, and generally require a complex array of equations. another weakness of using analytic formulas to arrive at retirement success rates lies in the fact that analytic models often ignore randomness in the independent variables. asset returns are the prime example. for instance, to assert that an individual willing to run out of money after 10 years can withdrawal nearly four times as much as one determined to preserve capital forever, orszag (2002) assumes a constant dollar return of 3% in his equations. returns, however, are not constant; and the sequence of returns can play a major role in portfolio depletion rates. 3.2. historical back testing perhaps the simplest approach for determining the sustainability rate of a retirement income plan is historical back testing (also known as rolling period analysis or overlapping period analysis). as the name implies, this approach relies on a sufficiently long set of historical returns data. the historical returns used are the actual returns an investor’s portfolio would have experienced, given its asset allocation. many retirement income risk models specify a withdrawal strategy throughout the planning horizon– often a fixed 20, 25 or 30 years. a commonly evaluated strategy is the 4% rule that withdraws an annual inflation-adjusted amount equal to 4% of the portfolio’s initial dollar value. the historical back testing method tests the success or failure of the retirement plan for each unique planning horizon in the data set of historical returns. the number of unique periods is determined by rolling up the start date of each planning horizon by a single increment of time. for example, bierwirth (1994) begins his analysis in 1926 and uses a one year rolling window to calculate 42 unique 27-year rolling periods, ending in 1992. each sample period is “unique” by virtue of the fact that its start year drops out of the data set as a new ending year enters the data set. intervening years, however, continue to appear in multiple samples. assuming that the past is indicative of the future, the historical model calculates the likelihood of retirement income sustainability by dividing the number of successful planning periods by the total number of rolling periods for any given asset allocation/retirement spending strategy combination. the combination with the highest success rate is considered optimal when optimality is measured by the likelihood that the unmodified or “autopilot” withdrawal strategy is sustainable over the applicable horizon. the acceptable retirement income sustainability rate is highly subjective, and depends on investor circumstances and risk tolerance. however, for the purposes of this discussion, we will benchmark model outputs relative to the 75% guideline of cooley, hubbard, and walz (2011). that is to say, a retirement risk model incorporating an asset management strategy exhibiting a 75% or greater likelihood of success is acceptable to a retired investor.6 the key question is: how credible is the success probability derived from a particular risk model; or, when is a model’s 75% or greater success rate not really indicative of a 75% or greater likelihood for success? 197p.j. collins et al. / financial services review 24 (2015) 193–216 historical back testing is easy to understand, and is a simple way to calculate relatively accurate assessments of what would have happened.7 however, an investor relying on such an approach should proceed with caution. decisions based solely on historical data force an investor to have faith in the highly dubious assumption that future returns will mimic realized past returns.8 furthermore, as stated, the rolling period method overweighs observations in the middle of the time period relative to observations occurring at the beginning or end. such over weighting creates statistical bias. extreme observations found in overweighted middle time horizons can cause clusters of failed sustainability. o’flinn and schirripa (2010) attribute one such cluster of failures in their study of withdrawal plans to significant inflation during the 1970s and 1980s. although modeling the serial and cross correlation of asset returns is desirable, the rolling period approach fails to provide a sufficient number of independent samples for credible portfolio sustainability testing. a looping method is one way to deemphasize the importance of the middle observations in the data set (davis, wicas, and kinniry, 2004). instead of stopping at the last observation date, the looping method carries calculations back to the beginning of the historical return sequence. however, the looping method distorts the value of the autocorrelation statistic in the historical dataset and presents its own statistical difficulties. although interesting, the pure history model fails to provide assurance that past conditions are sufficiently similar to current conditions so that they act as conditions precedent. 3.3. bootstrapping retirement income risk models sometimes use a bootstrap approach to develop a broader set of financial asset returns. bootstrapping is a process that develops return sequences by randomly drawing, typically with replacement, from the historically realized set of returns. randomly drawn sequences serve as possible future economic paths for testing the sustainability of retirement spending. a large number of economic paths can be bootstrapped, thus providing a larger set of scenarios for testing sustainability than is possible with the historical back testing process. bootstrapping, depending on the structure of the risk model, can either preserve correlations across asset classes—by making random draws which take a period’s realized returns across two or more asset classes (“cross-sectional” random draws)—or eliminate correlations across asset classes by taking random draws of asset returns from differing periods. it is beneficial to preserve the covariance of asset class returns; and so studies based on bootstrapping often preserve cross-sectional correlations. spitzer (2008), for example, uses the bootstrapping method to discern the best withdrawal rate and asset allocation over a pre-specified time horizon given an acceptable portfolio sustainability rate. however, unlike historical back testing, the bootstrapped scenarios do not preserve serial correlations evident in empirical returns, because the bootstrapped returns for each time period are independent draws from the set of historical outcomes. much like the rolling period technique, bootstrapping requires a long history of asset returns. without a long history, the sequences created from a small set of possible outcomes will be too similar, with the result that the retirement income risk model performs sustainability tests on scenarios that do not credibly reflect potential future economies. even with 198 p.j. collins et al. / financial services review 24 (2015) 193–216 a large history, sampled outcomes cannot differ from the pre-specified set of observables; therefore, propagated series over rely on the past to predict the future. 3.4. monte carlo simulation: normal distribution the monte carlo method further expands the set of possible outcomes for sustainability testing. it overcomes the limitation of relying on realized past returns as the basis of potential outcomes inherent in both the historical back testing and bootstrapping routines. monte carlo simulators generate sequences of potential economic paths by drawing random samples from probability density functions meant to represent the true underlying distribution of financial asset returns. this allows for a much greater range of potential outcomes. most commonly used monte carlo simulation engines assume asset class returns adhere to the well-known, bell-shaped normal, or log-normal, probability distribution that is often parameterized by the historical mean and variance. furthermore, a well behaved correlation matrix is used to preserve cross correlation in simulated outcomes. using a simple two asset class monte carlo simulation model with log-normally distributed returns, klinger (2011) reports sustainability rates above 85% for all retirement strategies he analyzed. klinger’s results, however, are based on the assumptions that stock returns average 6.9% per annum with a standard deviation of 15.7%, and that bonds return 6.6% on average with a standard deviation of 2.4%. the values of distribution parameters in simple monte carlo simulation models have been a point of contention.9 blanchett and blanchett (2008) point out that sustainability rates derived from simple monte carlo simulations are very sensitive to changes in the assumed expected rates of returns and standard deviations. however, the debate over what values constitute the most reliable parameters for the mean and standard deviation may be a secondary concern if asset price returns are not normally distributed. moreover, much like the bootstrapping method, simple monte carlo simulations destroy serial correlations evident in historical asset price returns. as we shall see, more complex simulation engines can imitate the auto-correlated nature of returns and the time-varying behavior of risk; but for now, we turn our attention to the topic of non-normally distributed returns. 3.5. monte carlo simulation: non-normal distributions the normality assumption implies that returns are stationary, symmetric and, at reasonable values for the standard deviation statistic, have low probabilities of realizing extreme deviations from the mean. simulated returns based on the gaussian distribution exhibit, on average, neither skewness nor excess kurtosis. statistical analysis of historical returns by lee (2009), however, indicate that realized returns are slightly skewed (asymmetric) and have higher likelihood of extreme events than predicted by a normal distribution (leptokurtic, fat-tailed, or heavy-tailed). however, if the bell curve does not accurately represent the true underlying distribution of returns, what other stable distributions can monte carlo engines use? levy and duchin (2004) fit monthly historical asset returns to eleven different probability distributions and infer that the logistic distribution is the best fit for many asset classes. athavale and goebel (2011) simulate asset returns based on 10 different distributions (beta, 199p.j. collins et al. / financial services review 24 (2015) 193–216 extreme, gamma, laplace, logistic, lognormal, pert, rayleigh, wakeby, and weibull) to test the 4% withdrawal rule. they conclude that a 4% withdrawal rate tested in non-normal distributions generally results in a lower sustainability rate when compared to test results using a simple monte carlo method that assumes distributional normality. the major econometric weakness of simulating one or more stable distributions, however, is that each distribution assumes that periodic returns are independent and stationary with the result that the model fails to capture autocorrelation. 3.6. vector auto-regression simulations using non-normal distributions address some problematic assumptions; but such simulations fail to account for autocorrelation in asset price returns. the time-dependent nature of asset prices manifests itself through momentum and mean reversion in the return series. more complex simulations based on vector autoregressive processes, however, can better reflect serial correlations. pang and warshawsky (2009), for example, use a vector autoregressive simulation model to compare six retirement plans using combinations of mutual funds and annuities. in addition to addressing the serial correlation issue, simulations based on vector autoregressive models can also incorporate what campbell and viceira (2005) call “state variables,” which are variables useful for forecasting asset returns. however, the vector autoregressive approach is often complex and, for investment advisors, difficult to implement. furthermore, the coefficients of the vector autoregressive equation must be estimated, and adding numbers of economic variables significantly reduces the precision of estimated parameters (campbell and viceira, 2005). with additional suitable but complex extensions, a vector autoregressive model can model heteroskedasticity and, therefore, capture a portfolio’s time-varying risk.10 3.7. regime switching an alternative to complex vector autoregressive conditional heteroskedastic models is a regime switching model.11 regime switching models assume returns come from two, or more, sets of probability distributions—one representing asset price behavior during states of normalcy, and the other(s) during financial crises. when financial crises occur, markets are afflicted with a flight to liquidity and with the contagion of fear. expected returns fall, volatility increases, and correlations converge towards one.12 the model we present in this article generates asset returns from two separate market regimes [bull and bear], with inflation modeled as an autoregressive process. assuming a two-state model, for each simulated return path, the regime switching algorithm calculates the probability either that the underlying economy will transition to the other regime in the next period, or a corresponding probability (1 – probability of switching) that the economy will remain in the same regime. both ang and bekaert (2004), and kritzman, page, and turkington (2012), for example, utilize a variable markov process with a constant transition probability. as simulated economies evolve through bull and bear markets of various lengths and magnitudes, the ang and bekaert model generates the 200 p.j. collins et al. / financial services review 24 (2015) 193–216 volatility cluttering and correlation breakdowns that characterize asset price behaviors in turbulent market conditions. as stated, an advantage of a regime switching approach is its ability to capture dynamic correlation and time-varying risk premia. thus, instead of using average unconditional correlation values determined by the historical data, the approach applies correlation values conditioned on the economic state of nature. for example, over the entire sample period, an asset class may exhibit a mean of 10% and a standard deviation of 20%. however, during bull markets, the parameter values may be �18% mean and 15% standard deviation; while, during bear markets, the parameter values may be �23% and 25% respectively. thus, simply using the unconditional mean, standard deviation and correlation values for the aggregate historical period cannot capture realistic asset price behavior. 4. inflation and model risk given the number of modeling approaches, it should not be surprising to find that there is a correspondingly wide range of model outputs. however, the investor, or advisor, may base decision making on the output from only one type of modeling approach; and, furthermore, may not realize that even this single-perspective view of retirement risk may flow from a model that incorporates oversimplified assumptions regarding critical factors such as inflation, investment costs, and rebalance frequency. here is the critical point: variations in a model’s mathematical structure and input assumptions can lead to outputs suggesting drastically different conclusions regarding the suitability of current asset management policies to a client’s financial objectives. for example, even within the simplistic rolling period analysis approach, using historical returns spanning different time periods or varying the size of the rolling window can generate substantially different success or failure probabilities. understanding such sensitivities is essential to discerning the trustworthiness of outputs from a retirement income risk model. to illustrate model risk under a sustainability risk metric, we present outputs from our proprietary risk modeling system which has the capacity to illustrate simulated outputs under varying asset management and modeling approaches. the initial model incorporates only a few basic variables and reflects a simple bell-curve structure for the distribution of future investment returns. we then utilize risk models that incorporate more variables and that allow for greater modeling flexibility. we demonstrate how an oversimplified model—many of which form the basis for normative articles in the financial press—may seriously distort the risks faced by retired investors. consider a simple, annually rebalanced, two asset class portfolio, allocated 70% to u.s. equities and 30% to u.s. bonds. initially, the model ignores fees, taxes, and transaction costs. the model assumes normally distributed asset returns parameterized by historical averages and standard deviations. portfolio price evolutions are multivariate normal, where the process derives from a single variance/covariance matrix assuming static (average historical) correlation values for all future economies. the initial portfolio value is $1,250,000 with an annual inflation-adjusted withdrawal of $60,000 for exactly 30 years. the portfolio consists of only two asset classes and, therefore, is not well diversified. 201p.j. collins et al. / financial services review 24 (2015) 193–216 to best illustrate the risk of relying on over-simplistic retirement income risk models, we focus on the inflation variable. a common way to incorporate inflation into a risk model is to assume a constant rate. however, retirement income risk models are hypersensitive to the level of assumed inflation with the result that success or failure probability assessments may differ widely. table 1 exhibits the inflation adjusted results of our two-asset class model under three different levels of fixed inflation: 3%, 4%, and 5%. a one percentage point difference in the assumed rate generates a marked divergence in ending portfolio values. at the median,13 the model assuming 3% inflation yields an ending value worth roughly a million dollars more in goods and services than the model assuming 4% inflation. in terms of the 75% probability of sustainability benchmark, the outcomes of both the 3% and 4% inflation models are acceptable (sustainability rate � [1 – bankruptcy rate]). the 3% inflation input yields a 91% sustainability rate, which is fully 9% higher than the 4% inflation input’s 82% sustainability rate. inflation, however, is not constant; and, dealing with inflation in such an oversimplified manner produces implausible outputs. more credible risk models treat inflation as a random variable. alongside the three constant inflation outputs in table 1 are two additional outputs, one based on inputting inflation starting at its long term average (4.32% during the period from 1973 to 2012), and the other based on inputting inflation starting at its previous 12 month average (1.74% in 2012). furthermore, the enhanced risk model generates paths of future inflation by treating it as a stochastic variable exhibiting a mean reversion factor. the inflation process has an expected value (drift) factor, and an innovation (diffusion) factor. the result is an output that accounts for serial correlation in a random but “sticky” time series of inflation rates. because the 2012 rate is so much lower than its historical average, the stochastic model of inflation using 2012 inflation as its starting point leads to a more plausible estimation of a portfolio’s sustainability rate. this result, in turn, contrasts with the results generated by a stochastic process starting off in the historically average inflationary environment.14 we direct the interested reader to the 2006 study by paul kaplan (2006) for additional discussion of how success probabilities differ depending on the retirement income risk model’s inflation assumptions. kaplan models inflation in three ways: constant, single-period lagged auto-correlated process, and two-period lagged auto-correlated process. table 1 various models of inflation risk model with inflation constant 3% inflation constant 4% inflation constant 5% inflation stochastic long term average inflation stochastic previous 12 months inflation ending wealth at the 50th percentile $2,417,712 $1,347,812 $665,967 $1,050,470 $1,423,391 ending wealth at the 30th percentile $1,232,951 $499,252 $33,002 $129,020 $397,651 ending wealth at the 10th percentile $71,029 $0 $0 $0 $0 bankruptcy 9% 18% 29% 26% 21% assets ever � $750k inflation adjusted 27% 43% 57% 51% 44% 202 p.j. collins et al. / financial services review 24 (2015) 193–216 5. illustrating risk 5.1. a simple two-asset class model having illustrated how five distinct inflation behaviors yield significantly different assessments of portfolio sustainability, we now turn our attention to the assumptions underlying the modeling of asset price evolutions. in our earlier discussion, we introduced the monte carlo method for generating the distribution of future asset prices. risk models using simple monte carlo simulations assume normally distributed asset returns parameterized with historic means and standard deviations. the simple monte carlo method also utilizes a historical correlation matrix to account for the co-movement of asset prices. we designate this model the “nh” model, for “normal historical.” it is a model commonly used by financial advisors.15 next, we modify the historical parameters of the basic monte carlo model to conform to the single-index, capital asset pricing model’s assumption that all investments have the same long-term expected real sharpe ratio in efficient markets. this modification assumes that expected returns plot on the capital market line. however, the model maintains the assumption of distributional normality. we designate this model ne, for “normal efficient.” although the nh and ne models differ in the values they use to parameterize the normal distribution assumption, both assume time-invariant parameters. finally, we consider modeling variations that mitigate many statistical difficulties arising when assuming distributional normality. we previously demonstrated that it is possible to originate economic evolutions from non-normal distributions using a number of techniques: bootstrapping, monte carlo simulations sampling from non-normal distributions, vector autoregressive models, and regime switching simulations. despite the fact that each approach has its pros and cons, we prefer the regime switching approach when cumulating dollar values over long planning horizons. mary hardy (2003), a prominent canadian actuary, stresses the importance of using credible risk models when cumulating dollar values over lengthy planning horizons. she provides strong support for using a regime-switching model. after an in-depth survey of various modeling tools and techniques, she concludes that the best way to approximate the range of future portfolio dollar value within an asset/liability matching context is through a two state regime-switching lognormal model. unlike simple monte carlo simulations, return series produced in a regime switching engine exhibit all of the following empirical asset price behaviors: skewness, fat-tails, autocorrelation, volatility clustering, and dynamic correlations. the first of our regime switching models assumes an investor with an agnostic view of the capital markets. this means that there is no attempt to predict whether the immediately forthcoming returns will start in either a bull market or a bear market. outputs, therefore, do not depend on the accuracy of the investor’s forecasting ability. this variation of the risk model randomly selects the underlying initial state of the economy where the selection of an initial bull or bear market state reflects historical relative frequency. we identify this randomly selected initialstate model “bb” for bull/bear. however, if the investor wishes to impose a market viewpoint, there is an opportunity to specify the forthcoming initial underlying economic state from which financial asset returns 203p.j. collins et al. / financial services review 24 (2015) 193–216 are generated. for example, if the investor strongly believes that asset prices will rise in the next period, we start our regime switching procedure in a bull market, and we identify the nonrandomly selected initial-state model as the “bull” model. on the other hand, if the investor wishes to reflect a pessimistic outlook for forthcoming returns; or, if the investor simply wishes to see how starting retirement in adverse economic conditions impacts portfolio values, then we start the return-generating process in a bear market. we identify this non-randomly selected initial-state model as the “bear” model. putting these models into the context of retirement planning, we generate return evolutions for an annually rebalanced, two-asset class portfolio, without accounting for taxes, fees, and other investment costs. inflation follows a stochastic process using a trailing 12 month average starting value. we present the inflation-adjusted outputs generated for this “simple” portfolio under the five distinct asset price models introduced above: 1. multivariate normal distribution historical model (nh) 2. multivariate normal efficient returns model (ne) 3. regime switching model market agnostic (bb) 4. regime switching model bull market prediction (bull) 5. regime switching model bear market prediction (bear) the risk-metric of interest is the likelihood that a $1.25 million portfolio distributing an inflation-adjusted $60,000 annually (4.8% of its initial value) will become fully depleted— that is, bankrupt—before the end of the planning horizon.16 we designate the series of model outputs by the symbol “s,” for their streamlined inputs, in fig. 2. fig. 2 also presents outputs from models we will discuss later. projected portfolio sustainability rates for the five models of the s series clearly depict a wide range of possible outcomes. although the commonly used nh model’s bankruptcy rates are less than 25%, the investor is left to ponder the extent to which this fig. 2. 204 p.j. collins et al. / financial services review 24 (2015) 193–216 favorable picture is merely an artifact of a risk model that fails to incorporate critical elements of asset pricing behaviors. the ne model modifies parameters fit directly from historical data; it produces a 6% higher portfolio risk profile. the market agnostic retirement income portfolio risk model (bb) indicates that the risk of portfolio depletion by myopically following a fixed asset allocation/4.8% withdrawal strategy is significantly greater than suggested by the simple monte carlo model. not surprisingly, by starting the simulations off in a predicted bull market environment, the bankruptcy rate decreases— but only by 3% over the agnostic view of future markets. it is still 6% higher than projected by the simple monte carlo model. on the other hand, having a pessimistic view of near term market returns drastically increases portfolio bankruptcy rates to approximately 49%. this outcome is illustrative of return sequence risk faced by retires.17 if a retiree wishes to impose a market viewpoint on his or her investment strategy—always a dangerous and uncertain proposition; or, if a retiree wishes insight into financial asset performance in “worst-case” economies, the bear model option illustrates a distribution of future results reflecting implied pessimism. 5.2. a diversified portfolio luckily, investors do not live in a two asset class world. the portfolio returns generated in our simplified set of models suffer from unsystematic market risk. a savvy investor would diversify away the unsystematic risk inherent in their portfolio, and move it to, say, a 14 asset class allocation.18 doing so allows us to illustrate the effect of diversification within each of the five return-generating models. in fig. 2, we label the new series “d,” for diversified. by taking advantage of broad-scope diversification, three of the five models in series d exhibit acceptable sustainability rates, compared to just one in series s. furthermore, each diversified portfolio in model series d has a lower bankruptcy rate compared with their two asset class counterparts in series s. however, it is interesting to see the relative benefits of diversification abate in models incorporating regime-shifting bb methodologies.19 when measured by the improvement in failure rates, there is only a diversification advantage of just 4% in the bb approach versus a 9% improvement in the nh model. the relative diversification benefits diminish in a regime shifting model. turbulent market conditions tend to raise the correlation between asset classes above their long term historic averages, thus reducing the benefit of diversification. neither the nh or ne models can replicate such conditions since their returns generating mechanism uses only a single correlation matrix. 5.3. longevity thus far, the model series assumes a fixed 30-year planning horizon. this assumption is not realistic for individual investors with uncertain life spans. a 30-year planning horizon may overstate shortfall risk for many post age 65 retired individuals. to get a more realistic view of risk, the uncertain nature of longevity is integrated into the next series of simulations. we term the series “dl” for a diversified portfolio reflecting an uncertain life span. the lengths of simulated trials are no longer preset at 30 years. rather, the distribution of life 205p.j. collins et al. / financial services review 24 (2015) 193–216 span reflects the society of actuaries’ mortality table for the subpopulation of high-income, white collar retirees—the group most likely to use the services of a financial advisor. longevity risk (the likelihood of outliving resources) is a stochastic variable, not simply an average. the distribution of life span for this population group differs significantly from the distribution of life span for the general population. inputting social security general population mortality data, for example, decreases the failure rate probabilities significantly. although beyond the scope of this discussion, the issue of expected versus actual life span is important when modeling retirement income portfolios.20 one result is that the uncertainty of an investor’s remaining lifetime increases with age despite a reduction in expected longevity. in fact, actuarial life expectancy is conditional on attained age—the longer you live, the longer you are expected to live. the dl model assumes a 68 year old female investor in excellent health. average life expectancy for such an investor is roughly 19 years (mean � 18.7; median � 19.1); and, therefore, we expect bankruptcy rates to fall. trial length is the lesser of 360 months (30 years), the date of portfolio depletion (bankruptcy), or the month of death, whichever event comes first. as fig. 2 shows, the mortality-adjusted time horizon lowers bankruptcy risk dramatically. all five market models incorporating longevity exhibit portfolio sustainability rates in excess of the 75% acceptability benchmark. an investor using a simple monte carlo simulation program anticipates that there is an approximately 4% chance that the investment strategy, operating across time without modifications, will be unable to meet critical needs. an investor with a pessimistic near-term market outlook, anticipates an approximately 25% chance that the retirement portfolio will be unable to provide the target lifetime income. the empirical data underlying each model’s output is exactly the same. this means that the differences in risk measurement is due solely to the structure of the retirement income risk models—that is, model risk. this does not imply that the model generating worst-case results is the most credible. however, it does suggest that inappropriate investment advice may be offered to investors based on outputs from overly simplistic models. 5.4. portfolio frictions (fees and taxes) the simulations presented to this point have ignored investment costs such as trading commissions, custodial/trustee fees, mutual fund/etf expenses, investment advisor fees, and so forth. the next series of simulations reflect the costs of investing for a diversified portfolio paying management fees21 and transaction costs. fig. 3 labels the series “dfl”—diversified portfolio paying fees and expenses and reflecting uncertain life span. as seen in fig. 3, incorporating fees and transaction costs increases bankruptcy rates. just as death is certain, so are taxes. taxes, however, are particularly difficult to model because of (1) myriad nuances within the tax code, and (2) variations in investor tax circumstances. the type of investment account often determines the nature and extent of taxation. additionally, the assets themselves may be tax exempt; interest and dividend income are often taxed differently than capital gains income; turnover rates may determine whether long or short term capital gains rates apply, and so forth. nonetheless, taxes are a cost factor either for setting the threshold target budget or for determining the drag on 206 p.j. collins et al. / financial services review 24 (2015) 193–216 portfolio growth. usually, investors calculate their income needs by including an estimate of the taxes that they must pay. however, advisors may, from time-to-time, need to incorporate taxes directly into the retirement income model. rather than ignore taxes, we have made some simplistic assumptions that allow us to evaluate, to a reasonable extent, the impact of taxes. our model maintains a constant tax regime in that it does not forecast changes in tax law over the investor’s lifetime. we do not include an allocation to municipal bonds, and we assume that no assets are held in tax sheltered accounts such as 401(k)s, 403(b)s, iras, and so forth.22 interest income is taxed at the ordinary income rate—assumed to be 20%, and all dividends are tax qualified and are taxed at a long-term rate of 15%. long-term rates are also used for the taxation of investment gains when assets are sold for both withdrawals and rebalancing. assets are taxed at a blend of short and long-term rates, depending on the annual turnover rates. we assume that an asset’s initial cost basis is half that of value in the start year, and that portfolio investment positions are low cost, low turnover, passive indexed funds. the model further assumes that tax loses and or tax payments are accounted for monthly. both operations are reasonably consistent with the notion of prepaying estimated taxes quarterly. fig. 3 also compares previous pretax results to results from a model encompassing taxation. in this case, the investor specifies that she requires a $5,000 minimum monthly income net of income tax liabilities. we label the new outputs “dftl” for a diversified portfolio paying investment fees, paying taxes, and incorporating an uncertain planning horizon by virtue of investor mortality. because taxes are an additional cost, the dftl bankruptcy rate probabilities all plot above the previous outputs that ignored taxes. the effect of our tax model is quite significant: more than doubling the failure rates in the normal distribution models, and raising bankruptcy rates from 4 to 7 percentage points in the regime switching models. fig. 3. 207p.j. collins et al. / financial services review 24 (2015) 193–216 5.5. summary of results fig. 4 puts all five previously examined series together on a single graph. our example started with the investor, holding a streamline two-asset class portfolio (the s series), demanding a constant retirement income stream for 30 years. bankruptcy rates from those simulations were somewhat alarming. the investor then sought help from an investment advisor who recommended diversifying the portfolio. the benefits of diversification are evident in the lower failure rates in the d series. however, diversification is less beneficial in the regime switching models. this result is not surprising given the tendency for correlation values to move towards one in times of market turbulence. however, it is unlikely that a retired investor will need income for exactly 30 years. incorporating the variability of life span into the model yields results presented in the dl series. the drastic fall in failure rates across all risk models indicates the importance of accounting for uncertain life span. when unknown life span is considered, bankruptcy rates fall to a third of their level in the nh and ne risk models: 12% to 4% and 16% to 5%, respectively. failure rates are only cut in half in the bb and bull models: 26% to 14% and 22% to 11%, respectively. the bear model exhibits the least benefit from the force of mortality: a 40% reduction in shortfall rates from 43% to 25%. the dl series depicts the lowest level of failure rates among model outputs. mortality affects sustainability rates in the normal-distribution models much more so than it does in the regime switching models. investment advice is not free and asset management generates trading cost; therefore, we further expand the models to include investment costs and advisory fees. the dfl series incorporates the force of mortality into a diversified portfolio model decremented for investment costs. the dfl model is a credible retirement income risk model when the investor’s budget is defined in pretax dollars. however, retirement income may come primarily from a trust established for the benefit of the investor. this may result in a tax fig. 4. 208 p.j. collins et al. / financial services review 24 (2015) 193–216 liability that must be paid from the trust portfolio rather than by the individual investor. therefore, we extend our models to cover taxation in the dftl series. taxation has a lesser effect on sustainability rates under regime switching models than under models assuming normality in the return distribution. fig. 4 clearly illustrates that the most simple and commonly used risk model, the nh model, understates risk given any set of underlying inputs or assumptions. on the other hand, assuming markets start out in a bear regime may overstate the risk of retirement failure. nevertheless, it is always interesting to examine the “worst case” environment. the differences in sustainability rates are stark: a 4% failure rate at the low end versus a 49% failure rate at the high end. decisions based on implausible risk models may not be appropriate and, thus, may mislead investors in assessing the risks and return expectations of their retirement investment strategies. 6. utility discussions concerning strategies to enhance portfolio sustainability differ from discussions concerning how to optimize aggregate utility of consumption for a retired investor with finite resources. portfolio sustainability, when defined as the ability to fund a minimum periodic target income, implies a state preference utility function. that is to say, a retired investor may have a strong preference for avoiding periods during which consumption falls below a minimum acceptable threshold. such an investor is willing to sacrifice greater utility in higher-portfolio-value states in favor of assuring a minimum standard of living in lower-portfolio-value states. aggregate utility—summed over all consumption/investment states—takes a back seat to assuring, at least, a minimum living standard in each state. optimization of expected utility, in most retirement income risk models, is a probability-weighted value taken over the entire distribution of outcomes— that is, over all possible economic states from depression to prosperity. summing utility values over all states assumes a separable utility function. furthermore, conclusions derived from optimization procedures may differ drastically from those drawn from sustainability analysis. for example, an optimal withdrawal rate in a study by finke, pfau, and williams (2012) requires that the utility-maximizing investor accept only a 43% sustainability rate. tomlinson (2012) notices similar observations. typically, commentators tracking shortfall risk metrics would consider such an optimal withdrawal strategy to be unacceptable. fortunately, whenever a threshold level of consumption must be maintained, the two approaches sometimes coincide. the investor may apply a utility penalty—negative utility— for failing to meet threshold income requirements. some retirement income risk models impose an additional penalty for exceeding a periodic income or ending wealth target.23 the risk model imposes a penalty for overachieving under the supposition that target financial objectives could have been met at a lower level of risk. in terms of consumption objectives, some commentators view surplus ending wealth merely as a missed opportunity to enjoy a higher standard of living throughout retirement. 209p.j. collins et al. / financial services review 24 (2015) 193–216 7. conclusion the normal (bell curve) distribution is not a good fit for most financial asset return series. quantifying investment risk by the first two moments of multivariate symmetric distributions (gaussian, student’s t, etc.) is often misleading. furthermore, monte carlo simulations based on a normal distribution cannot realistically capture asymmetry in the distribution (skew) or the frequency and magnitude of tail-risk events (leptokurtosis). risk models inputting return distributions with differing assumptions, operating under different stochastic processes, often produce significantly different results. to mitigate deficiencies, we use a hybrid autoregressive, regime-switching risk model using two state-dependent normal distributions with separately calculated means and variances. the distributions, according to the markov transition probabilities, capture the frequency and magnitudes of outlier results better than distributions produced by single-parameter input variables. given a large number of simulation paths, our retirement income risk model provides a rich set of future asset returns. although no investment risk model can predict the future, one hallmark of a credible model is that it enables investors to make good decisions within a wide range of possible futures.24 success or failure should never be evaluated in terms of just a single model—nor in terms of just a single metric. a model, in many respects, is just one person’s—that is, the model builder’s—opinion about how the future may unfold. given the computer power currently available to investment advisors, and given the approximately 30 years of research into econometric topics such as time series analysis, what accounts for the propensity to use oversimplified risk models? perhaps the financial advice profession suffers from what paul kleindorfer (2010) terms “legitimation.” he defines the term as follows: “. . . a credible anticipation of being held accountable not just for outcomes but for the logic that led to them will have predictable effects on the nature of the choice process itself.” a disgruntled investor’s demands for explanations regarding why and how an advisor’s recommendation went sour may lead advisors to “play it safe.” in the context of this article, advisors may tend to use only models commonly used throughout the profession—for example, historical back testing or monte carlo simulation models—so that if the investment strategy fails to produce its intended result, the advisor can take comfort in the fact that most other advisors also got it wrong: “. . . the mere thought of making choices of consequence under conditions of ambiguity and ignorance calls out for company.” undoubtedly, there are other explanations for the financial industry’s slow adoption of academic advances in risk modeling. however, if the central focus of each investor remains the intelligent consideration of risk/return tradeoffs, then the tools of the advisory profession should be evaluated in terms of their ability to indicate the consequences of portfolio management elections. a model is an imperfect representation of a more complex reality. in this case, there are (at least) two embedded “model risks” to consider: 1. investors are interested in forecasts of a price change process. however, the time series of asset prices is not statistically stationary (i.e., exhibits the potential for infinite variance). it is only by differentiating the logarithm of prices on a period-by-period basis that the creation of a stationary series of returns is possible. that is to say, it is 210 p.j. collins et al. / financial services review 24 (2015) 193–216 only possible to model asset returns; but an investor measures wealth based on asset prices. this is a subtle but important distinction. returns are based on the single historical path of price changes, the realization of which is merely a manifestation of an unknown price generating process. simulation analysis greatly broadens our perspective about possible future outcomes; but any model of such a process must remain only a crude approximation of reality. indeed, calculation of investment return is a function of the measurement interval (yearly, monthly, daily, intraday, continuous time) and, at the limit, may be meaningless in a statistical context. 2. the single historically realized return path for each asset class may be “representative” of the unknown price generating process; or, may merely be an outlier result unlikely to persist. for example, an asset allocation tilt towards small and value stocks is justified based on historical return data. if the premium for investing in small and value stocks reflects a reward for systematic risks, then investors have some confidence that they will continue to be rewarded for making these investments. if, however, the premium for such investments is merely an artifact of a chance historical price change process, then investors may be increasing risk as they move their asset allocation deeper into the small/value gradient. furthermore, investment volatility is measured by the variance statistic—the squared difference between actual returns and average return. however, if the historical return path is not representative, then the concept of average becomes meaningless and statistical measures are not illuminating. beyond this, a savvy investor should be aware of the limitations of basing decisions on the outcome of a risk model. if optimal decisions are model-dependent, how does an investor make the best decision when the outputs of various models differ significantly? this means that the investor must consider both the credibility of each risk model as well as the economic consequences of the various choices that the risk models present. investors are rewarded for taking prudent and calculated risks. investors may use historical data to make inferences concerning the interrelationships between asset allocation, risk, and reward. however, in designing and implementing a portfolio, it is always wise to remain aware of uncertainties in both data and the risk models that incorporate it. this is why it is important to track both sustainability and feasibility as part of a prudent assessment of retirement income strategies. past performance is not a guarantee of future results. we are grateful to an anonymous referee for a series of helpful comments. notes 1 technically, crra is nested in hyperbolic risk aversion functions. the critical distinction is between relative and absolute risk aversion. hyperbolic risk aversion functions [hara] encompass decreasing absolute risk aversion [dara]—as wealth increases the investor is more comfortable committing dollars to the risky asset; increasing absolute risk aversion [iara]—as wealth increases the investor pulls back on the number of dollars put at risk; and constant absolute risk aversion cara]—as wealth increases the investor keeps the same amount of dollars at risk. whenever the applicable risk metric defines the percentage of wealth put at risk, the risk model 211p.j. collins et al. / financial services review 24 (2015) 193–216 incorporates a relative risk aversion measure; whenever the applicable risk metric defines the level of dollar wealth put at risk, the risk model incorporates an absolute risk aversion measure. it is, however, the rare investor who remains indifferent to changes in the level of wealth when evaluating investment and distribution strategies. this is a central criticism of incorporating relative risk aversion into a retirement income model. 2 the society of actuaries and the actuarial foundation’s review of a cross-section of financial planning software determined that “. . . programs vary considerably regarding when the user runs out of assets, if at all. because of this finding, the study recommends that people run multiple programs, use multiple scenarios within programs, and rerun the programs every few years to reassess their financial position” (turner and witte, 2009). 3 an excellent review of econometric issues in the modeling of asset price returns is carol alexander’s (2008) four volume series market risk analysis. 4 see, for example, a. g. karolyi (2001) “why stock return volatility really matters,” strategic investor relations. 5 analytic models often assume lognormality when returns are measured in discrete time; when returns are measured in continuous time, the models often assume that returns follow a geometric brownian motion process. many analytical models must incorporate geometric brownian motion as a pre-condition to using the techniques of integral calculus. over the small intervals serving as units of time for continuous finance models, return differences between normal and non-normal distributions are minimal and seemingly inconsequential. aggregation of results over time often relies on the central limit theorem’s tenet that the mean of a sequence of normal random variables is, itself, a normal random variable that, at the limit, exhibits the mathematical property of convergence. however, for longer term planning horizons, assuming normality in the distribution of investment returns may have severe economic consequences if returns are, in fact, not normally distributed. additionally, the central limit theorem characterizes the distribution of sample means and provides only limited insight into the value of the variance statistic. 6 see dicarlo jr. and fast (2008) for a survey of opinions found in financial advice literature regarding the acceptability of various levels of portfolio shortfall risk. the article focuses primarily on standards of prudence for management of trust-owned investment portfolios. 7 we note additional complications surrounding continuous index availability throughout the planning period as well as portfolio drift in the absence of constant rebalancing to the designated asset allocation target. 8 mcgoun (1995) argues that the empirical distribution of financial asset price returns is not a measure of risk. it is merely a measure of historical realizations that may or may not be applicable to the current economic situation. mcgoun’s article presents a history of risk measurement by economists. it provides a good theoretical basis for a monitoring and surveillance system using current observables to supplement shortfall risk measures. 9 milevsky and abaimova (2006) observe that different commercially-available monte carol simulation programs produce different solutions even when given the same inputs. 212 p.j. collins et al. / financial services review 24 (2015) 193–216 10 research by kopcke, webb, hurwitz, and li (2013) compares outputs generated by three vector autoregressive models. 11 for an introduction to markov chain transition probability matrices, see laverty, miket, and kelly (2002). hybrid models utilize a regime switching mechanism in conjunction with a vector autoregressive conditional heteroskedastic process. for example, a variety of autoregressive processes are incorporated into regime switching models by hamilton and susmel (1994). litzenberger and modest (2010) develop a model with eight different states in which the financial asset behavior within each state is modeled by a normal distribution with state-dependent means and standard deviations. ameriks, caplin, and van nieuwerburgh (2008) utilize a four-state health model to examine issues in retirement income planning that seeks to preserve a threshold income level. the matrix is a markov chain transition matrix with an age-varying, one-period” transition probability. the evolution of health status is also an important variable in the model described by gupta and li (2013). 12 smith and gould (2006) discuss the differences between unlucky draws from a stable probability distribution versus a substantial change (for the worse) in the probability distribution itself. 13 the 50th percentile value of the distribution when ranking outputs (trials) according to terminal value from low (1st percentile) to high (100th percentile). 14 our risk model also incorporates cpi into the variance/co-variance matrix of asset classes used to generate asset price evolutions. this means that there is a complex interaction between increases in cpi—that are more probable under a mean-reversionary process when the inflation rate is below its historical average—and asset price evolutions which may be negatively correlated to inflation rate increases. 15 since the realized path of history is a single vector of results—a sample of one—the nh model produces a credible distribution of investment outcomes only under the assumption that future investment conditions will mirror previous economic environments. 16 in this example, we arbitrarily fix the planning horizon at 30 years. we later expand the example to incorporate the force of mortality. 17 the risk that the combination of (1) unfavorable portfolio returns realized early in retirement and, (2) portfolio withdrawals will deplete the portfolio’s dollar value to the point where future favorable returns operating on smaller dollar values are insufficient to offset the early losses. 18 the diversified portfolio invests in broad range of domestic and developed foreign equities with a tilt towards value oriented and small capitalization stocks. for fixed income, the portfolio invests in diversified pools of high quality domestic and foreign bonds with short and intermediate maturities. other holdings in the portfolio include positions in diversified baskets of us reits (real estate investment trusts) and emerging market stocks. 19 an early study of the effect of diversification on portfolio sustainability is collins, savage, and stampfli (2000). this article presents results from simulating multiasset class portfolios consisting of globally diversified stock, bond and real estate investments. 213p.j. collins et al. / financial services review 24 (2015) 193–216 20 one half of the population will live longer than the average life span; and, in some cases, the individual’s life span may be many years above the average. distributions of life spans approach exponential distributions with long tails. the research of brown & mcdaid (2003), sponsored by the soa, reviews 45 research articles that examine factors, including preand post-retirement income, that affect mortality. 21 in place of a simple flat percentage fee, we elect to use the following progressive fee schedule: 1% on the first million, 65 basis points (bps) on the second million, 35 bps on each dollar between two and five-million, and finally 25 bps on any amount above five-million. if the portfolio grows, this laddered fee structure is less detrimental to the portfolio than a flat 1% fee. a minimum fee of $5,000 is charged to the portfolio if it ever falls below half a million dollars. the portfolio is passively managed with index funds and rebalanced annually. 22 these assumptions allow us to bypass the issue of asset location. gordon pye (2001) discusses at length how a portfolio’s asset location affects the withdrawal rate adjustment amount necessary to maintain portfolio sustainability in after tax dollars. incorporating asset location into a risk model adds another dimension of analysis, and model risk, to retirement planning. although there are well known rules of “conventional wisdom” that address withdrawal strategies from various accounts, such as “draw down taxable accounts first, then turn to taxed-deferred accounts,” coopersmith and sumutka (2011) demonstrate the benefits of a tax-efficient optimization approach over conventional wisdom. 23 hughen, laatsch, and klein (2002) view positive terminal wealth as a potential “opportunity cost.” the authors develop a concept called “the equivalent payment value.” this is a way to convert terminal wealth into extra monthly payments throughout the planning horizon. blanchett, kowara, and chen (2012) optimize based on a utility function that penalizes portfolio depletion more heavily than it penalizes excess bequests. see also, scott, sharpe, and watson (2009), which advances the proposition that investors should not try to generate “unneeded”/low-utility surplus. 24 thomas j. sargent’s noble prize winning research deals with how investors make decisions when they doubt the accuracy of their model. when confronted with ambiguity, they tend to use a family of models and to over-weight bad outcomes as a mechanism for exercising caution: see hansen and sargent (2008) for example. for a discussion on the wisdom of using multiple models for security valuation, see collins (2007). references alexander, c. (2008). market risk analysis. chichester: john wiley & sons. ameriks, j., caplin, a., & van nieuwerburgh, s. (2008). annuity valuation, long-term care, and bequest motives. in j. ameriks & o. mitchell (eds.), recalibrating retirement spending and saving (pp. 251–275). oxford: oxford university press. ang, a., & bekaert, g. (2004). how regimes affect asset allocation. financial analyst journal, 60, 86–99. athavale, m., & goebel, j. m. (2011). a safe withdrawal rate using various return distributions. journal of financial planning, 24, 36–43. 214 p.j. collins et al. / financial services review 24 (2015) 193–216 bierwirth, l. (1994). investing for retirement: using the past to model the future. journal of financial planning, 7, 14–24. blanchett, d. m., & blanchett, b. c. (2008). data dependence and sustainable real withdrawal rates. journal of financial planning, 21, 70–84. blanchett, d., kowara, m., & chen, p. (2012). optimal withdrawal strategy for retirement income portfolios. chicago, il: morningstar. brown, r. l., & mcdaid, j. (2003). factors affecting retirement mortality. north american actuarial journal, 7, 24–43. campbell, j. y., & viceira, l. m. (2005). long-horizon mean-variance analysis: a user guide, appendix for “the term structure of the risk-return tradeoff.” financial analysis journal. (available at http://scholar. harvard.edu/files/campbell/files/long-horizon_appendix_2005.pdf). campbell, j. y., & viceira, l. m. (2005). the term structure of the risk-return trade-off. financial analysts journal, 61, 34–44. collins, p. j. (2007). prudence. the banking law journal, 124, 29–96. collins, p. j., savage, s. l., & stampfli, j. (2000). financial consequences of distribution elections from total return trusts. real property, probate, and trust journal, 35, 243–304. cooley, p. l., hubbard, c. m., & walz, d. t. (2011). portfolio success rates: where to draw the line. journal of financial planning, 24, 48–60. coopersmith, l. w., & sumutka, a. r. (2011). tax efficient retirement withdrawal planning using a linear programming model. journal of financial planning, 24, 50–59. davis, j. h., wicas, n. w., & kinniry, f. m. (2004). the strengths and weaknesses of various financial simulation methods. journal of wealth management, 6, 33–42. dicarlo jr., d. p., & fast, s. m. (2008). prudence: what are the odds? ali-aba course of study: representing estate and trust beneficiaries and fiduciaries, 477–479. finke, m., pfau, w. d., & williams, d. (2012). spending flexibility and safe withdrawal rates. journal of financial planning, 25, 44–51. gupta, a., & li, z. (2013). optimal annuity purchase decisions under uncertain lifetime. journal of family economics, 34, 447–459. hamilton, j., & susmel, r. (1994). autoregressive conditional heteroskedasticity and changes in regime. journal of econometrics, 64, 307–333. hansen, l. p., & sargent, t. j. (2008). robustness. princeton, nj: princeton university press. hardy, m. (2003). investment guarantees: modeling and risk management for equity-linked life insurance. new york, ny: john wiley & sons. huang, h., milevsky, m. a., & wang, j. (2004). ruined moments in your life: how good are the approximations? insurance: mathematics and economics, 34, 421–447. hughen, c. j., laatsch, f. e., & klein, d. p. (2002). withdrawal patterns and rebalancing costs for taxable portfolios. financial services review, 11, 341–366. kaplan, p. d. (2006). asset allocation with annuities for retirement income management. the journal of wealth management, 8, 27–40. karolyi, a. g. (2001). why stock return volatility really matters. strategic investor relations, 1, 1–20. kleindorfer, p. r. (2010). reflections on decision-making under uncertainty. in f. x. diebold, n. a. doherty, & r. j. herring (eds.), the known, the unknown, and the unknowable in financial risk management. princeton, nj: princeton university press. klinger, w. (2011). in search of the “best” retirement strategy. journal of financial service professionals, 65, 62–73. kopcke, r. w., webb, a., hurwitz, j., & li, z. (2013). rethinking optimal wealth accumulation and decumulation strategies in the wake of the financial crisis. center for retirement research at boston college crr wp 2013-1. kritzman, m., page, s., & turkington, d. (2012). regime shifts: implications for dynamic strategies. financial analysts journal, 68, 22–39. 215p.j. collins et al. / financial services review 24 (2015) 193–216 laverty, w. h., miket, m. j., & kelly, i. w. (2002). simulation of hidden markov models with excel. journal of the royal statistical society: series d (the statisticican). 51, 31–40. lee, m. i. (2009). model risk in retirement simulations (available at http://us.dimensional.com/pdf/model_risk_ in_retirement_simulations.pdf). levy, h., & duchin, r. (2004). asset return distributions and the investment horizon. journal of portfolio management, 30, 47–62. litzenberger, r. h., & modest, d. m. (2010). crisis and noncrisis risk in financial markets. in f. x. diebold, n. a. doherty, & r. j. herring (eds.), the known, the unknown and the unknowable in financial risk management (pp. 74–102). princeton, nj: princeton university press. mcgoun, e. g. (1995). the history of risk “measurement.” critical perspectives on accounting, 6, 511–532. milevsky, m. a., & abaimova, a. (2006). will the true monte carlo number please stand up? journal of financial planning, online edition. o’flinn, c., & schirripa, f. (2010). social science research network (available at http://papers.ssrn.com/sol3/ papers.cfm?abstract_id�1641382). orszag, m. (2002). ruin in retirement: running out of money in drawdown programs. watson wyatt worldwide technical report ru06. pang, g., & warshawsky, m. j. (2009). comparing strategies for retirement wealth management: mutual funds and annuities. journal of financial planning, 22, 36–47. pye, g. b. (2001). adjusting withdrawal rates for taxes and expenses. journal of financial planning, 14, 126–136. scott, j., sharpe, w., & watson, j. (2009). the 4% rule: at what price? journal of investment management, 7, 31–48. smith, g., & gould, d. (2006). measuring and controlling shortfall risk in retirement. the journal of investing, 16, 82–95. spitzer, j. j. (2008). retirement withdrawals: an analysis of the benefits of periodic “midcourse” adjustments. financial services review, 17, 17–29. tomlinson, j. a. (2012). a utility-based approach to evaluating investment strategies. journal of financial planning, 25, 53–60. turner, j. a., & witte, h. a. (2009). retirement planning software and post-retirement risks. society of actuaries, p. 20. 216 p.j. collins et al. / financial services review 24 (2015) 193–216 develop a retirement plan and stick to it: it will improve both your attitude and behavior with money gizelle d. willowsa,* acollege of accounting, university of cape town, rondebosch, 7001, south africa abstract this study explores the relationship between financial literacy, behavior and attitude, and retirement savings decisions. members of a south african tertiary institution’s retirement fund were surveyed and multistage multivariate regression and mediation analyses showed that developing and conforming to a retirement plan positively influenced financial attitude and behavior. this indicates that interventions should focus on the specific behaviors which drive retirement planning, rather than financial literacy in isolation. use of formal tools such as consulting with a financial planner also increases the relative risk of successful retirement planning. this presents a tangible and linear approach towards improved financial behavior. © 2020 academy of financial services. all rights reserved. jel classifications: c4; d9; e2; g4 keywords: financial literacy; financial behavior; financial attitude; planner; retirement savings 1. introduction many investors lack a basic understanding of financial concepts and a sizeable proportion of the population in a number of countries lack financial literacy (atkinson & messy, 2012). an awareness of this creates the need to understand the potentially deleterious effects of this lack of financial literacy on an individual’s financial wealth. this study focuses on retirement savings as one aspect of financial wealth, with specific focus on the attitude, behavior, and retirement savings decisions that are made by individuals. thus, this study aims to * corresponding author. tel.: +2721 650 2292 e-mail address: gizelle.willows@uct.ac.za 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 243–271 assess more than just financial literacy, but also the impact of several variables on financial attitude and financial behavior. the social and broader macroeconomic context within south africa is such that consumers are unable to only rely on the state and/or their employer for financial security in retirement. the national treasury (2014) has stated that only an estimated six percentage of south africans will be able to replace their full income at retirement and maintain their lifestyle. the situation south africa finds itself in is not too dissimilar to other countries. in the united states, a shift from a few decades ago shows americans needing to rely on 401(k)type accounts to supplement social security in retirement (morrissey, 2016). additionally, this retirement wealth has been unable to grow at the same pace that the aging population has. in canada, the threat of widening intergenerational inequality exists (mckenna, 2015). almost half of canadian families nearing retirement have no accrued employer pension benefits (shillington, 2016). the cost of health care and old age security, which is paid out of current revenues, is rising at a faster rate than the tax base. this study is important because it goes beyond the measurement of financial literacy and assesses the connections with retirement savings decisions and behaviors. it does this by assessing not just the impact of financial literacy on thinking and planning for retirement, but also the interconnectedness between financial literacy, financial attitude, planning for retirement and financial behavior. focusing on south africa, which is rich in diversity with large inequality and unemployment rates (statistics south africa, 2018; taylor, 2002), also provides a unique perspective on a community not too dissimilar to comparable developing economies across the world. the research approach includes multivariate regression and mediation analysis on data obtained from surveying members of a south african tertiary institution’s retirement fund. socioeconomic information for each respondent was obtained directly from the retirement fund, allowing for the necessary objective controls when assessing the proposed relationships. the results show that it is not financial literacy that is positively related to financial behavior. rather, developing and conforming to a retirement plan, is the predictive variable. furthermore, a positive financial attitude is shown to be positively associated with financial behavior. however, further mediation analysis showed that being a more successful planner was associated with a better financial attitude and in turn better financial behavior. furthermore, a formal approach to retirement planning, such as consulting with a financial planner, can assist in being a more successful planner. these results are helpful as they provide the basis for a tangible plan to improve retirement saving for individuals. furthermore, educational interventions can be targeted to specific behaviors, rather than financial literacy in isolation. this study continues by reviewing literature on financial literacy and financial behavior. thereafter, the research method is presented, which includes descriptive statistics to understand the socioeconomic breakdown of the respondents. the measurement of variables is also discussed. the results continue to present the output from multivariate regression and mediation analyses that critically assess the relationship among the measured variables. discussion on the outcomes are presented and conclusions and recommendations made which are broadly applicable across the globe. 244 g.d. willows / financial services review 28 (2020) 243–271 2. literature review analyses of the level of financial literacy of individuals will be reviewed and potential sources from which retirement savings information is received will be summarized. after that, connections with a consumer’s associated financial behavior (i.e., their propensity to save for retirement) will be examined. studies that have looked at education interventions will also be considered as a possible means to improve financial literacy. 2.1. financial literacy financial literacy includes financial knowledge, awareness, and skills and capability, with the last of those being inclusive of financial planning (xu & zia, 2012). lusardi and mitchell (2011) measured how workers made their savings decisions, how they collected information to make such decisions and whether these workers possessed the financial literacy needed to make such decisions. lusardi and mitchell (2011) found financial illiteracy to be most common among older americans and that minorities, women, and consumers without a college or high school degree were most susceptible to having low financial knowledge. this was supported by xu and zia (2012) who found women to have lower levels of financial literacy in almost all the countries they investigated. using a ten-question financial literacy experiment, agnew and szykman (2005) found that individuals tended to gain financial experience as they age, which increased their financial literacy. the researchers found that married consumers performed better than those who were single. however, individuals with children had lower levels of financial literacy than those without children. the variable with the most significant effect on the test scores was salary levels, which were found to decrease test scores as salary levels decreased. the worst performers were those without a college degree. furthermore, investigating financial literacy across a number of countries, xu and zia (2012) found higher-income countries to perform better on financial literacy tests than lower-income countries, and that levels of financial literacy followed an inverted-u shape when plotted against age. agnew and szykman (2005) noted that consumers with the lowest results on a financial exam were the same types of consumers who were not saving enough for retirement. this literature presents a potential correlation between financial literacy and retirement savings decisions. furthermore, certain socioeconomic characteristics such as age, gender, marital status, and salary appear to influence financial literacy. however, further insight from meta-analyses and studies using methods to arrive at causal estimates is required. showing that planning and financial literacy are interconnected, lusardi and mitchell (2011) found that those individuals who displayed financial literacy were more inclined to plan and, furthermore, to succeed in that plan. this demonstrates a possible association between financial literacy and planning for retirement, which might result in successful retirement savings outcomes. furthermore, olsen and whitman (2007) found that poor savings can be ascribed to a lack of suitable retirement goals. g.d. willows / financial services review 28 (2020) 243–271 245 2.2. source of information the source from which a consumer receives financial information could play an integral part in their overall financial welfare (olsen & whitman, 2007). in the united states, men, older and better-educated consumers, non-hispanic whites, and those earning $70 000 or more a year were found to be significantly more likely to use formal advisors. those earning less than $20 000 a year were more likely to rely on an informal advisor (olsen &whitman, 2007). research has shown that people with low basic literacy capabilities rely on informal sources of information, such as friends and family. conversely, the proportion of households that rely on inter alia newspapers, books, financial magazines and financial information on the internet increases considerably as they move from a low level to a higher level of basic literacy. professional financial advisors are also more likely to be trusted by those households displaying higher financial literacy (van rooij et al., 2011). moving from basic to advanced financial literacy, the result is similar but more pronounced. those individuals who display high levels of advanced financial literacy are less likely to rely on informal sources of information and are more likely to consult financial advisors and seek information in newspapers and on the internet (van rooij et al., 2011). these findings propose a possible link between the tools that individuals use as a source of retirement saving information and their level of financial literacy. next, studies investigating the effect of financial education interventions will be reviewed, in so much as its potential impact on financial literacy and financial behavior. 2.3. influencing financial behavior miller et al. (2015) conducted a meta-analysis on 188 papers and articles that presented impact results of interventions designed to improve financial literacy and/or financial behavior. key findings were that financial literacy and capability interventions can have a positive impact in certain areas such as increasing savings and promoting certain financial skills, such as record keeping (miller et al., 2015). in south africa specifically, financial messages delivered through a popular soap opera improved certain financial behaviors such as borrowing from a formal financial institution rather than retailers which come with higher costs (berg & zia, 2013). a study that focused on the effect of financial education on debt outcomes in early adulthood was performed by brown et al (2016). the interventions themselves were related to various changes to state-level high-school curricula which varied between financial literacy, economics, and mathematics course offerings. the results showed significant positive effects of financial literacy exposure on increased debt knowledge of youth—in that it increases the occurrence of credit reports thus inferring an understanding of the importance of building a credit record (brown et al, 2016). furthermore, for those who had a credit report, math and financial literacy reduced average debt balances and the likelihood of carrying debt. the net effect of both math and financial literacy education was an increase in average creditworthiness. this positive impact of mathematics education is confirmed by goodman (2009) and cole et al. (2013) who show that students exposed to more math training have higher average incomes and savings. 246 g.d. willows / financial services review 28 (2020) 243–271 less promising findings by brown et al. (2016) showed that economic education resulted in increased debt balances, specifically debt for the purpose of supporting consumption. this suggests that while economics training might clarify borrowing and credit markets, it does not improve the ability to make appropriate financial decisions. these results suggest that financial education programs might have a significant impact on the financial decision-making of youth. however, the content of these programs should be carefully constructed. there appears to be different roles for different types of quantitative education in influencing young adults’ debt experiences. brown et al. (2016) do acknowledge a shortcoming in the research in that they were unable to disassociate effects by demographics. this provides impetus for the research that this paper is attempting which controls for the socioeconomic breakdown of respondents. a further natural experiment looked at the effect of a mandatory eight-hour personal financial management course (pfmc) on newly enlisted soldiers in the u.s. army (skimmyhorn, 2016). the pfmc covered principles (e.g., the time value of money), rules of thumb (e.g., how to obtain a copy of your credit report), and the financial decisions young workers are most likely to face (e.g., buying a car). attendance at the course succeeded in increasing average monthly contributions and retirement savings rates, the effect of which lasted at least two years. furthermore, the probability of having adverse legal actions or delinquencies in the first year after the course was reduced. however, this effect did not hold in the second year. these results are somewhat contrary to brown et al. (2016). potential explanations could be that brown et al. (2016) focused on the decision’s youth make shortly after leaving high school. skimmyhorn (2016) present several reasons explaining the success of the pfmc. for one, the course is well-timed for individuals who are becoming increasingly responsible for their own financial wealth. second, the course has a focused curriculum which covers only the most relevant topics for its cohort of students. finally, the course offers tailored advice (e.g., use credit when purchasing assets not consumables) rather than broad principles (e.g., how to prepare a net-present-value calculation), which helps maintain the students’ attention and interest. another notable financial education intervention is the supplemented high-school curriculum in brazil. using a randomized control trial, bruhn et al. (2016) studied the impact of this comprehensive financial education program in six states, 868 schools, and approximately 20,000 high school students in brazil. the program used new textbooks with interactive classroom exercises, practical homework (e.g., creating a household budget with parents), and relevant role-playing exercises. teachers were supported with training, instructor handbooks, and internet-based learning tools. the program increased the financial knowledge of the students that led to an increase in saving for purchases, a greater likelihood of financial planning, and greater participation of students in household financial decisions. positive effects were also noted on students’ intertemporal preferences and attitudes. furthermore, there was evidence of “trickle-up” whereby the parents of these students also showed improved financial knowledge, saving, and spending behavior (bruhn et al, 2016). these results suggest that financial education might be a valuable complement to the typical high-school curriculum. however, such education requires greater intensity than that which is seen in characteristic once-off financial education training. g.d. willows / financial services review 28 (2020) 243–271 247 a final meta-analysis of the relationship of financial literacy and education on financial behavior was performed by fernandes et al. (2014) across 201 prior studies. the results showed that interventions to improve financial literacy explained only 0.1% of the variance in the financial behaviors studies. additionally, financial education deteriorated over time. also, the effects of financial literacy reduce when controlling for certain psychological traits that were omitted in the prior studies investigated. this emphasizes the need for further research in the field of financial literacy to control for these psychological traits. this will be considered in this article. building the case on the importance of psychological traits, miller et al. (2015) propose that omitted variables related to such psychological traits (such as impulse control, self-efficacy and delayed gratification) contribute to the variance in the results for financial education interventions. miller et al. (2015) posit that the variance is because people with certain psychometric profiles are more likely to engage in activities that might improve their financial literacy and, in turn, their financial outcomes. however, these behaviors (specifically self-control) are not the typical focus of financial education interventions. 3. method the findings from the literature review inform the following three research questions for this study: 1. does an individual’s level of financial literacy influence him/her thinking about retirement? 2. does an individual’s level of financial literacy influence him/her planning for retirement? 3. does an individual’s level of financial literacy and/or financial attitude and/or the type of planner he/she is influence his/her financial behavior? thus, the results will present findings on the connection between financial literacy, financial behavior, financial attitude and retirement savings decisions. of which the latter encapsulates thinking and planning for retirement. staff at the university of cape town (uct) who were members of the uct retirement fund (uctrf) were targeted for a survey. the uctrf was established on january 1, 1995 and is a defined contribution provident fund. all permanent and fixed-term contract employees of uct, inclusive of academic and support or administrative staff, who have not yet reached the normal retirement age, automatically become members of the uctrf upon their appointment. the uctrf is relevant to the south african landscape in that it is a large fund (both in terms of number of members and accumulated funds) and that it combines academics and administrative staff that are from very different walks of life. thus, while the results may not be generalizable to the rest of south africa in a statistical sense, as a single case study, it provides relevant information that is theoretically or logically generalizable. 248 g.d. willows / financial services review 28 (2020) 243–271 participants were asked questions to determine their level of financial literacy, financial behavior, financial attitude, and whether they thought about and planned for their retirement. most of the questions used were drawn from previous studies (notably atkinson & messy, 2012; lusardi & mitchell, 2011, 2017; and van rooij, lusardi, & alessie, 2011). these questions have been used in a number of studies worldwide, including: australia (agnew, bateman, & thorp, 2013), switzerland (brown & graf, 2013), france (arrondel, debbich, & savignac, 2013), russia (klapper & panos, 2011), romania (beckmann, 2013), and japan (sekita, 2011). certain socioeconomic characteristics were also garnered from the survey questions. appendixes a– c show the various questions. furthermore, the uctrf agreed to provide certain sociodemographic information for each participant such as age, gender, race, cost of employment (coe) and highest qualification (all as at july 31, 2014). the inclusion of these characteristics addresses some of the variances noted by brown et al. (2016) and miller et al. (2015) in the literature reviewed. 3.1. research strategy the survey was cognitively tested by 11 people (five males and six females of differing ages) to evaluate the wording and design of the survey (willimack, lyberg, martin, japec, & whitridge, 2004). the suggestions from this testing were analyzed and, where appropriate, changes were made to update and improve the original survey questions. the cognitive testing was done in stages; not all 11 testers tested the survey at the same time. this was done to enable each subsequent tester to test any changes suggested by the previous testers. the changes made and the reasoning for each amendment were as follows: • changing certain americanisms to south african terminology (e.g., replacing “firm” with “company” and “stock” with “share”). • some of the ordered response options were re-ordered from “disagree to agree” to “agree to disagree,” as most testers seemed to anticipate that construction. • references in financial knowledge questions to “savings accounts” and “shares” were changed to include “savings accounts/cash” and “shares/equity” to avoid incorrect responses owing to differing terminology. • references to “moderate” risk companies were changed to “medium” risk companies for more universal/easier language use. the uctrf had 3 602 members as at may 31, 2014 to whom surveys were sent. of these 3,602 members, 3,333 had access to email. the survey was emailed to these members on tuesday, august 26, 2014 and remained open until the end of that week. the remaining 269 members had hard copies of the survey (with self-addressed return envelopes) posted to them. a total of 764 responses were received (of which 23 were in hard-copy format). this equated to a total response rate of just over 21%. the responses were sent to the uctrf by the respondents. the uctrf used the staff number provided by each respondent to identify and provide relevant information such as but not limited to: age (as of july 31, 2014), race, gender, cost of employment (coe), and g.d. willows / financial services review 28 (2020) 243–271 249 highest qualification level. however, the uctrf was unable to give information for 11 of the members owing to incorrect staff numbers as provided by these respondents. consequently, these 11 members were excluded from any further analysis, resulting in a final sample of 753 members. nevertheless, the sample remained large enough to apply the central limit theorem and to assume a normal distribution for statistical testing to be performed. also, 43 respondents did not provide data as to their highest qualification. in the testing, controls were implemented for these nonresponses, and as such, the affected respondents remained in the sample. to avoid an overstatement (or understatement) of results arising from a nonrepresentative sample (as was found by lusardi & mitchell, 2017), the sample was weighted by the gender groups and racial groups, to ensure greater representativeness of the south african population. ideally, the sample should also be weighted by age, income, and education, because of the older, high earning, and highly educated sample. however, this was not possible due to the lack of accurate statistics on our variables of interest from the south african population. instead, gender and race (two important factors in the context of south africa) have been considered to address this issue. a t test was conducted with each of age, income, and education by gender, as well as a correlation analysis for race across age, income and education. these significant results indicate that weighting the data by gender and race subsequently addressed the weighting issues among the other variables for which there is no available data. table 1 provides further detail regarding the final weighted sample of respondents. 3.2. research process to address the research question, the data were analyzed in four themes, namely (1) financial literacy, (2) thinking about retirement, (3) planning for retirement, and (4) financial behavior and attitude. descriptive statistics and preliminary analyses for each of these themes are presented first. 3.2.1. financial literacy the basic financial literacy questions tested simple concepts that form the basis of basic financial decision-making and transactions. table 2 shows the percentage of correct answers for the total sample of respondents in respect of each of the four questions. table 2 shows that while most respondents can do simple calculations (calculating interest) financial literacy is not comprehensive. the percentage of correct answers is not nearly as high as the 93% and 91% found by lusardi and mitchell (2017) for the numeracy and inflation questions, respectively. the results for the advanced financial literacy questions are shown in table 3. these questions were more sophisticated than the basic questions and tested concepts regarding the share market, collective investment schemes, bonds, and risk. more than half of the respondents have some knowledge of shares, long-period returns, variability of returns, and how risk diversification works. however, an analysis of the responses in respect of knowledge of and principles relating to bonds shows a lack of competence. nearly half of the respondents (43%) responded that they did not know the answer to the bond principles question (that tested the link between bond prices and interest rates). 250 g.d. willows / financial services review 28 (2020) 243–271 a financial literacy score was then calculated for each respondent by deriving person scores from a three-parameter logistic item response model with a pseudo-guessing parameter common to all items. the financial literacy items had kuder-richardson coefficient of reliability of 0.81. two items separated from the rest in terms of both difficulty and discrimination with both having markedly greater values on both parameters (q21 and q24 in appendix b). the test information function showed that the instrument consisting of all financial literacy items provided the most information about participants with ability levels in the midrange of the presumed latent financial literacy. persons scores were transformed to a standard t-score with a mean of 50 and standard deviation of 10, for example, a score of 60 is one standard deviation above the mean, while a score of 30 is two standard deviations below the mean. t-scores are positive and fall within the interval [0, 100]. the results are shown in table 4 while classifying the sample of respondents into different subsets by socioeconomic characteristics, namely: (1) age, (2) gender, (3) race, (4) education level, (5) marital status, and (6) cost of employment. within the race subset, respondents are classified as being either white, colored (of mixed-race descent), african, indian, or other. the mean financial literacy score for the total sample is 45 (sd¼ 9.8; median 44). the distribution is somewhat negatively skewed with a minimum value of 28 and maximum of 69. when analyzed for different socioeconomic characteristics, it is noticeable that male respondents, those respondents with postgraduate qualification levels, those respondents with higher earnings, and all racial groupings other than african, have mean scores higher than the average. african respondents have a mean score of 44 (sd¼ 9.6; median 41). the noticeable racial divide between the financial literacy scores might be caused by a variety of issues. some possible explanations include: table 1 survey sample (n = 753) minimum mean median maximum age 24 43.16 44 65 coe (in zar) 80 778 338 870 273 260 1 916 158 male female gender 415 (55%) 338 (45%) african colored white indian other race 600.42 (79.74%) 59.84 (7.95%) 58.11 (7.72%) 16.93 (2.25%) 17.70 (2.35%) less than high school high school higher certificate/diploma college postgraduate education 0.33 (0.05%) 120.87 (18.31%) 66.78 (10.12%) 133.96 (20.30%) 338.05 (51.22%) married single marital status 477.40 (63.40%) 275.60 (36.60%) source: author’s calculations. g.d. willows / financial services review 28 (2020) 243–271 251 1. the survey was in the english language. for black respondents in particular, this might have been their second, or perhaps third language (gough, 1996). the wording of the financial literacy questions was important in their construction and this could have been a disadvantage to those who are less fluent in english. 2. as a result of the racially exclusionist education policies implemented during the apartheid years in south africa, many non-white south africans might have experienced substandard schooling, or have grown up in households with parents who might never have had a formal qualification or exposure to finance and financial instruments (draper and spaull, 2015). as financial literacy is often acquired over time, this might have been another disadvantage to non-white respondents. furthermore, financial literacy appears to be at its highest for those over 65 years of age. however, there were only two respondents over the age of 65 years. respondents between the ages of 35 years to 65 years had a mean financial literacy score lower than the average. this suggests a decrease in financial literacy with age, which is dissimilar to the finding of xu and zia (2012). this will be further analyzed in the regression analysis in table 7. xu and zia (2012) found higher-income countries to perform better on financial literacy tests than lower-income countries, and that levels of financial literacy followed an inverted-u shape when plotted against age. 3.2.2. thinking about retirement to determine what influences retirement planning, survey respondents were asked a question assessing how much they thought about retirement. the results are shown in table 5. the largest proportion of the full sample of respondents have thought about retirement “a lot.” this result is consistent within most of the socioeconomic cohorts into which the full sample is divided. the older the respondent, the greater the proportion of respondents who thought about retirement a lot. it might be expected that a consumer would be more inclined to think about retirement as they get older and near retirement. table 2 percentage of correct answers by basic financial literacy question (n= 753) numeracy inflation time value of money money illusion correct 72% 55% 59% 45% incorrect 7% 12% 28% 51% do not know 21% 33% 13% 4% source: author’s calculations. sample weighted by gender and race. table 3 percentage of correct answers by advanced financial literacy question (n= 753) function of share market knowledge of shares knowledge of cis’s knowledge of bonds longperiod returns highest variability risk diversification bond principles correct 46% 75% 44% 46% 50% 58% 66% 25% incorrect 19% 12% 14% 28% 38% 18% 20% 32% do not know 35% 13% 42% 26% 12% 24% 14% 43% source: author’s calculations. sample weighted by gender and race. 252 g.d. willows / financial services review 28 (2020) 243–271 3.2.3. planning for retirement to determine what “type of planner” each respondent is, a series of questions were asked assessing whether each respondent had tried to determine how much to save for retirement, whether he or she had developed a plan to do so, and if so, whether he or she had conformed to that plan consistently. an analysis of responses has enabled the grouping of respondents in a manner similar to that used by lusardi and mitchell (2011), as shown in table 6. approximately 27% of respondents had tried to determine how much to save for retirement. these results are somewhat less promising than the 31.3% noted by lusardi and mitchell (2011) and reflect a disappointing result in the sense that most respondents have not done this exercise. however, several studies show a similar trend with few consumers undertaking or understanding retirement planning. further questioning in the survey of these respondents (i.e., those who had indeed tried to determine out how much to save for retirement) revealed that almost 15% of such respondents had never developed a plan to undertake this saving. these respondents are termed “simple planners.” one-third had developed a plan, while the majority (53.8%) had “more or less” developed a plan. these two groups were then further questioned to determine table 4 summary statistics of financial literacy score mean median standard deviation total sample (n = 753) 45 44 9.8 age <35 (n= 175) 47 46 9.3 35 to 49 (n= 397) 44 42 9.5 50 to 64 (n= 179) 44 44 10.6 >=65 (n= 2) 55 57 9.9 gender male (n= 415) 47 46 10.4 female (n= 338) 43 42 8.7 race african (n= 600) 44 41 9.6 colored (n= 60) 46 46 7.6 white (n= 58) 55 56 8.1 indian (n= 17) 51 48 9.9 other (n= 18) 52 54 5.8 education less than high school (n= 1) 34 34 n/a high school (n= 112) 42 40 7.7 higher certificate or diploma (n= 62) 41 38 7.8 college (n= 125) 40 40 7.0 postgraduate (n= 315) 51 52 9.3 unknown (n= 138) 40 40 8.2 marital status married (n= 477) 45 43 10.4 single (n= 276) 45 44 8.8 cost of employment <r400 000 p.a. (n= 520) 42 40 7.8 >r400 000 p.a. (n= 233) 53 54 9.5 source: author’s calculations. sample weighted by gender and race. g.d. willows / financial services review 28 (2020) 243–271 253 whether they were able to conform to the plan that they had (or more or less had) developed. a cumulative 22.5% of these respondents had rarely or never conformed to the plan. these respondents are termed “serious planners,” that is, they developed a plan, but just could not conform to it. the remaining respondents, who “always” or “mostly” conformed to their plans are termed “successful planners.” these “successful planners” represent 17% of the total sample. most of the total sample (73%) have never tried to determine how much they will need for their retirement. this grouping is termed “not a planner.” 3.2.4. financial behavior and attitude a score representing each respondent’s financial behavior and attitude was calculated to assess whether the positive outcomes of being financially knowledgeable are driving behavior and whether a respondent’s financial attitude drives his/her behavior. these scores were derived using a graded response model (see appendix c for the questions). persons scores were again transformed to t-scores. most of the items had great overlap among many of the response categories, resulting in those categories offering little in terms of placing people on the scale. the test information function showed that for both the behavior and attitude instrument, all seven behavior items table 5 responses to “thinking about retirement” by socioeconomic characteristics how much have you thought about retirement? a lot some only a little not at all full sample (n = 753) 35.4% 27.0% 29.4% 8.2% age <35 (n= 175) 24.2% 22.7% 35.5% 17.6% 35 to 49 (n= 397) 26.8% 33.3% 34.0% 5.8% 50 to 64 (n= 179) 64.8% 17.0% 13.7% 4.5% >=65 (n= 2) 74.1% 19.7% 6.2% 0.0% gender male (n= 415) 29.3% 30.0% 32.7% 8.0% female (n= 338) 42.8% 23.3% 25.4% 8.5% race african (n= 600) 34.2% 23.3% 32.9% 9.6% colored (n= 60) 40.7% 33.4% 20.0% 6.0% white (n= 58) 41.5% 45.9% 11.1% 1.5% indian (n= 17) 41.9% 41.4% 16.7% 0.0% other (n= 18) 30.6% 53.0% 16.3% 0.0% education less than high school (n= 1) 0.0% 100% 0.0% 0.0% high school (n= 112) 40.8% 14.8% 44.3% 0.2% higher certificate or diploma (n= 62) 29.2% 23.4% 30.2% 17.1% college (n= 125) 37.9% 35.6% 17.3% 9.3% postgraduate (n= 315) 33.9% 35.3% 28.3% 2.4% unknown (n= 138) 34.9% 11.5% 30.6% 23.1% marital status married (n= 477) 35.0% 31.3% 23.8% 10.0% single (n= 276) 36.0% 19.5% 39.2% 5.3% cost of employment <r400 000 p.a. (n= 520) 33.9% 21.7% 32.8% 11.7% >r400 000 p.a. (n= 233) 38.8% 38.8% 22.0% 0.5% source: author’s calculations. sample weighted by gender and race. 254 g.d. willows / financial services review 28 (2020) 243–271 and all three attitude items provided the most information about participants. trait levels were in the lower to midrange of the presumed latent financial behavior continuum and in the mid to upper range of the presumed latent financial attitude continuum. the cronbach’s a on the financial behavior items and financial attitude items was 0.76 and 0.70, respectively. the financial behavior and attitude scores had a mean of 48 and 50, and median of 49 and 50, respectively (with a maximum of 66 on the behavior scale and a maximum of 62 on the attitude scale). a higher score indicates better behavior and attitude towards financial decisions. both scores showed negatively skewed distributions. 4. results the results are presented in the same four themes, but with specific emphasis on the factors influencing those variables. multivariate regression analyses are performed at each stage to assess the relationship among variables. following that, a mediation analysis is performed. 4.1. factors influencing financial literacy after determining each respondent’s financial literacy score, an ordinary least squares (ols) regression was performed to determine whether any factors have a relationship with financial literacy. the regression analyses were performed using robust standard errors to account for heteroskedasticity. the assumption of homoskedasticity was evaluated using the breusch-pagan/cook-weisberg test (x2(1)¼ 43.62, p¼<0.001). the independent variables listed are standard variables which the literature has found to be explanatory (atkinson & messy, 2012; lusardi & mitchell, 2011, 2017; van rooij et al., 2011). the reference groups for the racial categories grouping and education level variables will be the white racial group and high school education, respectively. the results are shown in table 7. the model is as follows: table 6 proportion of planners in respective sub-groups successful planner serious planner simple planner not a planner did you ever try to figure out how much to save for retirement? yes no 26.9% (n= 203) 73.1% (n= 550) did you develop a plan? yes more or less no 31.3% (n= 64) 53.8% (n= 109) 14.9% (n= 30) were you able to stick to the plan? always mostly rarely never 23.1% (n= 40) 51.4% (n= 89) 19.1% (n= 33) 6.4% (n= 11) successful planner serious planner simple planner not a planner source: author’s calculations. sample weighted by gender and race. g.d. willows / financial services review 28 (2020) 243–271 255 financial literacy score ¼ b 0 þ b 1 ageþ b 2 age-squared þ b 3marital status þ b 4 financial dependentsþ b 5gender þ b 6-10 racial grouping dummies þ b 11 log of cost of employment þ b 12-17 level of qualification dummiesþ e (1) the statistically significant factors affecting financial literacy are: being male, race, cost of employment, and qualification levels. having a higher cost of employment or being male, as opposed to female, is associated with a higher financial literacy score on average, holding all else constant. being of a different racial grouping (to white) is expected to have a lower financial literacy score on average. the african racial grouping has the largest negative coefficient on financial literacy, in relation to the white racial grouping, which is significant at the one percentage level. the potential reasons for this are discussed earlier. lastly, a qualification less than high school, when compared with a high school qualification, is negatively associated with an increased financial literacy score. however, given there is only one respondent in this category, limited interpretation can be made of this result. 4.2. factors influencing thinking about retirement to determine the factors that influence thinking about retirement, a multinomial logistic regression that controls for a range of socioeconomic factors is shown in table 8. this table 7 ols regression of financial literacy score coefficient (std error) age �0.575 (0.781) age-squared 0.004 (0.008) marital status �1.538 (1.309) dependents 2.614 (2.782) gender: male 2.803* (1.352) race: colored �3.473* (1.437) race: african �7.166*** (2.011) race: indian �3.424* (1.540) race: other �3.970 (2.331) log coe 8.744*** (2.071) qualification: less than high school �6.301* (2.553) qualification: higher certificate/diploma 0.500 (2.878) qualification: college �2.743 (2.144) qualification: post graduate 2.359 (2.616) qualification: unknown 1.483 (2.442) constant �46.38 (27.91) observations 753 r2 0.484 source: author’s calculations. sample weighted by gender and race. standard errors in parentheses. ***p< .01, ** p< .05, * p< .1. 256 g.d. willows / financial services review 28 (2020) 243–271 model was applied in favor of an ordered logistic regression as the proportional odds assumption was not met. because of the very small numbers in the less than high school education level group and both the indian and other race groups a subsample excluding these cases were considered for the model to avoid issues of multicollinearity and or complete or quasi-complete separation that may result in large standard errors and consequently misinterpretation of the results. the reference groups for the racial categories grouping and education level variables will be the white racial group and higher certificate or diploma education, respectively. the model is as follows: thinking about retirement ¼ b 0 þ b 1financial literacyþ b 2 ageþ b 3 age-squared þ b 4marital status þ b 5 financial dependents þ b 6gender þ b 7-9 racial grouping dummies þ b 10 log of cost of employment þ b 11-15 level of qualification dummies þ e (2) table 8 shows that an increase in financial literacy, age, cost of employment, or education level, increases the relative risk of thinking more about retirement. for age, it is understandable that as a consumer is closer to retirement (in most cases equating to being older) the more tangible retirement becomes necessitating more thoughts about it. an increased cost of employment suggests that having more disposable income enables the means to save for retirement and thus allows thoughts about it. for education level, the relative risk ratio of a respondent with a college or postgraduate degree thinking about retirement a lot, compared with not at all, is 4.7 and 3.8 times higher than the relative risk ratio for someone with only a higher certificate or diploma. this might indicate that higher education creates a greater awareness of retirement. for those respondents who are married, have dependents or are male, the relative risk of thinking about retirement is lower. the gender variable is only statistically significant (at the five percentage level) when comparing thinking about retirement not at all, to a lot. a possible explanation for married respondents thinking less about retirement than single respondents might be explained by financial dependency on their partner. the presence of dependents in turn might necessitate thinking more about daily living expenses and providing for those dependents, rather than retirement. returning to financial literacy, the statistically significant positive association (mostly at the 1% level) with thinking about retirement suggests possible endogeneity. this might result in “thinking about retirement” influencing “financial literacy” rather than the inverse. a question was asked to assess each respondent’s level of accounting and/or economics knowledge (q4 in appendix a). based on the responses, each respondent was placed into one of three ordinal groups. the purpose of this was to use this ordinal variable as an instrumental variable associated with the financial literacy score. an instrumental multinomial probit model (first stage uncensored, second stage probit) was performed (wald x 2(57)¼ 1014.74, pvalue¼<0.001) (roodman, 2011). the output is given in appendix d. the results show that financial literacy remains statistically significant in increasing the relative risk of a respondent thinking more about retirement. this emphasizes the importance of financial literacy. however, although thinking about retirement is a step in the right direction, it is the g.d. willows / financial services review 28 (2020) 243–271 257 transcendence of this thinking into actively planning for retirement that will assist in improving retirement savings. an analysis of this progression is examined next. 4.3. factors influencing planning for retirement a multivariate ordered logistic regression analysis showing the relationship between planning and financial literacy and planning tools, after controlling for several factors, is shown in table 9. two tests are performed. test 1 only assesses the 351 respondents who were table 8 multinomial logistic regression of thinking about retirement only a little some a lot coefficient (std error) risk ratio coefficient (std error) risk ratio coefficient (std error) risk ratio financial literacy 0.093** (0.033) 1.097** (0.036) 0.156*** (0.033) 1.169*** (0.039) 0.111*** (0.033) 1.117*** (0.037) age 0.173*** (0.330) 1.188*** (0.039) 0.141*** (0.034) 1.151*** (0.039) 0.224*** (0.033) 1.251*** (0.042) marital status �1.242** (0.424) 0.289** (0.122) 0.0897 (0.452) 1.094 (0.494) �0.973** (0.427) 0.378** (0.161) dependents �1.738 (1.144) 0.176 (0.201) �2.979** (1.154) 0.051** (0.059) �2.333** (1.155) 0.097** (0.112) gender: male �0.320 (0.400) 0.726 (0.290) �0.519 (0.410) 0.595 (0.244) �1.041** (0.401) 0.353** (0.142) race: colored 1.370 (1.537) 3.936 (6.051) 1.144 (1.497) 3.140 (4.700) 1.316 (1.503) 3.728 (5.604) race: african 1.885 (1.397) 6.588 (9.204) 1.308 (1.359) 3.699 (5.027) 1.691 (1.370) 5.425 (7.431) log coe 1.009 (0.628) 2.742 (1.724) 1.813** (0.641) 6.126** (3.926) 1.408** (0.637) 4.087** (2.605) qualification: high school 5.949** (2.415) 383.55** (926.35) 5.055** (2.432) 156.79** (381.31) 6.033** (2.422) 416.82** (1009.69) qualification: college 0.649 (0.633) 1.914 (1.211) 1.666** (0.631) 5.294** (3.340) 1.551** (0.628) 4.715** (2.960) qualification: postgraduate 1.458** (0.743) 4.297** (3.191) 0.613 (0.760) 1.845 (1.402) 1.338* (0.757) 3.811* (2.887) qualification: unknown 0.143 (0.520) 1.154 (0.600) �0.388 (0.574) 0.679 (0.390) 0.379 (0.540) 1.461 (0.788) constant �21.67** (7.334) <0.001** (<0.001) �32.22*** (7.488) <0.001*** (<0.001) �28.72*** (7.470) <0.001*** (<0.001) observations 708 pseudo r2 0.2414 source: author’s calculations. sample weighted by gender and race. standard errors in parentheses. ***p< .01, ** p< .05, * p< .1. 258 g.d. willows / financial services review 28 (2020) 243–271 classified as a type of planner (i.e., “simple,” “serious,” or “successful planner”). this is done to assess the influence of different planning tools, as only those respondents who indicated that they did try to determine how much to save for retirement were asked which planning tools they used. test 2 assesses all 753 respondents, that is, simple, serious, and successful planners as well as those classified as not a planner. an approximate likelihoodratio test of proportionality of odds across response categories was performed and the results showed no serious violation of the assumption (test 1: x2(13)¼ 15.92, p¼ .2536 and test 2: x 2(16)¼ 22.48, p¼ .1283). the models are as follows: type of planner ¼ b 0 þ b 1 talk to family=relatives þ b 2 talk to co-workers=friends þ b 3 attemded retirement seminars þ b 4used calculators or worksheets þ b 5 consulted a financial planner þ b 6 financial literacy þ b 7 ageþ b 8marital status þ b 9 financial dependents þ b 10gender þ b 11-15 racial grouping dummies þ b 16 log of cost of employment þ b 17-22 level of qualification dummies þ e (3) type of planner ¼ b 0 þ b 1 financial literacyþ b 2 ageþ b 3marital status þ b 4 financial dependentsþ b 5 gender þ b 6-10 racial grouping dummies þ b 11 log of cost of employment þ b 12-17 level of qualification dummies þ e (4) consulting a financial advisor or talking to family or relatives statistically significantly increases the odds of being a successful planner versus the combined simple and serious planner. the former shows that a respondent is 40 times more likely to be a more successful planner if he or she consults a financial planner. however, talking to co-workers or friends as a tool for retirement planning decreases the odds of being a more successful planner. this suggests that a more formal approach is a more successful approach. this is supported by bashall, willows, and west (2018) who found that investors acting with the assistance of professional advisors showed negative behavioral biases to a lesser extent. however, this does not extend as far as to improve investment returns (allie, west, & willows, 2016). the benefit of involving a professional might be limited to the behavioral element. however, the difference between financial planners and financial advisors are also pervasive. g.d. willows / financial services review 28 (2020) 243–271 259 t ab le 9 t w o -s ta g e o rd er ed lo g is ti c re g re ss io n o f ty p e o f p la n n er t es t 1 : p la n n er s o n ly t es t 2 : a ll re sp o n d en ts c o ef fi ci en t (s td er ro r) o d d s ra ti o c o ef fi ci en t (s td er ro r) o d d s ra ti o f in an ci al li te ra cy 0 .0 3 3 (0 .0 4 0 ) 1 .0 3 4 (0 .0 4 1 ) 0 .0 1 8 (0 .0 1 3 ) 1 .0 1 8 (0 .0 1 3 ) t al k to fa m il y /r el at iv es 1 .6 3 4 * * (0 .5 5 2 ) 5 .1 2 4 * * * (2 .8 2 7 ) t al k to co -w o rk er s/ fr ie n d s �1 .1 8 9 * (0 .4 9 6 ) 0 .3 0 5 * * (0 .1 5 1 ) a tt en d ed re ti re m en t se m in ar s �0 .2 1 9 (0 .4 7 0 ) 0 .8 0 4 (0 .3 7 7 ) u se d ca lc u la to rs o r w o rk sh ee ts 0 .1 1 0 (0 .5 4 1 ) 1 .1 1 6 (0 .6 0 4 ) c o n su lt ed fi n an ci al p la n n er 3 .7 0 2 * * * (0 .6 3 2 ) 4 0 .5 3 1 * * * (2 5 .5 9 9 ) a g e 0 .1 0 2 * * * (0 .0 2 8 ) 1 .1 0 7 * * * (0 .0 3 1 ) 0 .0 4 5 * * * (0 .0 1 1 ) 1 .0 4 6 * * * (0 .0 1 1 ) m ar it al st at u s 0 .9 6 9 * (0 .4 1 5 ) 2 .6 3 5 * * (1 .0 9 4 ) �0 .3 7 1 (0 .2 1 5 ) 0 .6 9 0 * (0 .1 4 8 ) d ep en d en ts �1 .1 8 5 (0 .7 9 0 ) 0 .3 0 6 (0 .2 4 2 ) 0 .8 9 8 * * (0 .3 2 6 ) 2 .4 5 4 * * * (0 .7 9 9 ) g en d er : m al e �1 .0 8 3 * (0 .4 9 3 ) 0 .3 3 9 * * (0 .1 6 7 ) 0 .1 6 4 (0 .1 9 1 ) 1 .1 7 8 (0 .2 2 5 ) r ac e: c o lo re d �0 .7 9 8 (0 .7 3 3 ) 0 .4 5 0 (0 .3 3 0 ) �0 .1 3 9 (0 .4 2 7 ) 0 .8 7 1 (0 .3 7 2 ) r ac e: a fr ic an 0 .8 5 2 (0 .6 9 7 ) 2 .3 4 5 (1 .6 3 3 ) �0 .6 9 5 (0 .3 5 7 ) 0 .4 9 9 * (0 .1 7 8 ) r ac e: in d ia n 3 .5 3 8 * (1 .6 1 3 ) 3 4 .4 0 9 * * (5 5 .5 0 6 ) �0 .3 4 9 (0 .6 2 3 ) 0 .7 0 5 (0 .4 3 9 ) r ac e: o th er �1 .7 7 2 (0 .9 4 8 ) 0 .1 7 0 * (0 .1 6 1 ) �0 .9 3 0 (0 .5 9 5 ) 0 .3 9 5 (0 .2 3 5 ) lo g c o e 0 .5 8 3 (0 .7 0 3 ) 1 .7 9 2 (1 .2 5 9 ) 1 .6 2 8 * * * (0 .3 2 3 ) 5 .0 9 3 * * * (1 .6 4 3 ) q u al ifi ca ti o n : l es s th an h ig h sc h o o l �1 8 .2 2 (2 9 6 5 5 .4 6 ) < 0 .0 0 1 (< 0 .0 0 1 ) q u al ifi ca ti o n : h ig h er ce rt ifi ca te /d ip lo m a 0 .2 8 6 (1 .1 0 4 ) 1 .3 3 1 (1 .4 6 9 ) �1 .0 0 7 (0 .5 2 4 ) 0 .3 6 5 * (0 .1 9 1 ) q u al ifi ca ti o n : c o ll eg e 3 .3 4 3 * * * (0 .8 5 0 ) 2 8 .3 0 0 * * * (2 4 .0 6 0 ) 0 .2 3 7 (0 .3 2 3 ) 1 .2 6 7 (0 .4 0 9 ) q u al ifi ca ti o n : p o st -g ra d u at e 1 .4 4 3 (0 .7 9 2 ) 4 .2 3 1 * (3 .3 5 0 ) �0 .4 0 9 (0 .3 3 7 ) 0 .6 6 4 (0 .2 2 4 ) q u al ifi ca ti o n : u n k n o w n 0 .6 5 0 (0 .9 1 8 ) 1 .9 1 6 (1 .7 5 8 ) �0 .3 7 6 (0 .3 7 3 ) 0 .6 8 7 (0 .2 5 7 ) c o n st an t cu t1 1 2 .6 9 8 (8 .6 4 6 ) 2 4 .2 3 * * * (3 .8 7 1 ) c o n st an t cu t2 1 5 .2 7 5 (8 .6 6 1 ) 2 4 .5 0 * * * (3 .8 7 4 ) c o n st an t cu t3 2 4 .9 7 * * * (3 .8 7 9 ) o b se rv at io n s 3 5 1 7 5 3 p se u d o r 2 0 .1 4 4 9 0 .1 9 5 6 s o u rc e: a u th o r’ s ca lc u la ti o n s. s am p le w ei g h te d b y g en d er an d ra ce . s ta n d ar d er ro rs in p ar en th es es . * * * p < .0 1 , * * p < .0 5 , * p < .1 . 260 g.d. willows / financial services review 28 (2020) 243–271 emphasis is placed on test 2 for the discussion of the remaining variables. this includes those respondents who have never tried to determine how much they need to save for their retirement (i.e., those designated as not a planner). the results show that having dependents statistically significantly increases the odds of being a planner. this might be explained by the need to manage financial affairs to avoid problems for dependents (dave, 2017). either in the form of providing for their day-to-day need or upon the respondent’s death. the odds of being a planner increases as the respondent ages. this positive relationship was also seen in table 8, showing the heightened awareness of both thinking and planning for retirement the closer it becomes a reality. another noteworthy finding which mirrors the results from table 8 is the positive relationship with cost of employment. 4.4. factors influencing financial behavior and attitude an ols regression analysis was performed to determine whether a consumer’s financial attitude and/or financial literacy and/or the type of planner he or she is influences his or her financial behavior. the regression analyses were performed using robust standard errors to account for heteroskedasticity. the assumption of homoskedasticity was evaluated using the breusch-pagan/ cook-weisberg test (x2 (1)¼ 6.12, p¼<0.013). not being a planner is the reference group for the type of planner variable. these results are shown in table 10. the model is as follows: financial behavior ¼ b 0 þ b 1 financial attitude þ b 2 financial literacy þ b 3-6 type of planner dummies þ b 7 ageþ b 8 age-squared þ b 9marital status þ b 10 financial dependents þ b 11gender þ b 12-16 racial grouping dummies þ b 17 log of cost of employment þ b 18-23 level of qualification dummies þ e (5) the effect of financial attitude is statistically significant (at the 1% significance level) in influencing financial behavior after controlling for a range of factors. this result supports the finding of atkinson and messy (2012) that a person with a negative attitude towards saving for his or her future will be less inclined to actually save. being a serious planner (i.e., developing a plan) or successful planner (i.e., sticking to that plan), compared with those that are not a planner, results in a more positive financial behavior. this suggests that considered decisions in terms of figuring out how much to save for retirement and developing a plan to do so will influence an individual’s financial behavior. this might be because careful consideration on spending behavior is required to stick to the plan. an increase in age positively influences financial behavior, but only after the age of 44 years. a negative coefficient is noted before the age of 43 years. however, both these coefficients are small. lastly, indian respondents (when compared with white respondents), on average, had a higher financial behavior score, holding all else constant. g.d. willows / financial services review 28 (2020) 243–271 261 4.5. financial attitude as a mediator the results of table’s 9 and 10 both show that financial literacy is not associated with the type of planner a respondent is, nor whether he or she will have a more positive financial behavior. however, financial attitude and being a serious or successful planner does show a positive relationship with financial behavior. further consideration is given as to whether there is connection among these three variables. this is diagrammatically explained as follows: a mediation model, as proposed by baron and kenny (1986) suggests that rather than hypothesizing a direct causal relationship between the iv and dv (that is represented by path c in diagram 1), that the iv affects the dv through its association with m (represented by paths a and table 10 ols regression of financial behavior coefficient (std error) financial attitude 0.501*** (0.090) financial literacy 0.004 (0.012) simple planner 0.258 (0.182) serious planner 0.610** (0.305) successful planner 0.563** (0.219) age �0.122* (0.066) age-squared 0.001** (<0.001) marital status 0.0208 (0.213) dependents 0.185 (0.173) gender: male 0.160 (0.168) race: colored 0.228 (0.162) race: african 0.163 (0.210) race: indian 0.425** (0.167) race: other 0.233 (0.255) log coe 0.170 (0.184) qualification: less than high school �0.205 (0.293) qualification: higher certificate/diploma 0.154 (0.431) qualification: college 0.103 (0.305) qualification: post-graduate �0.0887 (0.236) qualification: unknown �0.151 (0.276) constant �0.531 (2.979) observations 753 r2 0.390 source: author’s calculations. sample weighted by gender and race. standard errors in parentheses ***p< .01, ** p< .05, * p< .1. m = financial attitude; iv = type of planner; dv = financial behavior. 262 g.d. willows / financial services review 28 (2020) 243–271 b in diagram 1). to test for such mediation, the direct and indirect effects were derived after fitting a multiple linear regression. the direct and indirect effects of being a more successful planner on financial behavior are statistically significant. however, the coefficient for the direct effect (0.173) is greater than the indirect effect (0.042). the results indicate that part of the reason why a positive financial attitude positively influences financial behavior is because being a more successful planner positively influences a respondent’s financial attitude. the root mean squared error of approximation is very small that suggests that this model was able to reproduce the covariance matrix accurately. furthermore, the standardized root mean squared residual was also very small. the model fit was satisfactory with a relatively small root mean square error of approximation (rmsea; <0.001) and standardized root mean square residual (srmr; <0.001). the r2 and adjusted r2 are 0.2485 and 0.2465, respectively. 5. conclusion the positive effects of planning for retirement are apparent. it positively influences an individual’s financial attitude and, in turn, his or her financial behavior. increased focus on assisting individuals to try figure out how much to save for retirement is necessitated. further emphasis on then developing a plan and sticking to that plan is even more beneficial. the results showed that consulting with a financial planner statistically significantly increases the odds of being a more successful planner. this is noteworthy as it indicates an active step that can be taken to positively influence an individual’s financial behavior. additionally, individuals should be discouraged from talking to co-workers or friends to get advice on how to save for retirement. the key finding should be that a formal approach is the preferred approach. while the level of financial literacy is seen to show significant differences among respondents of different gender, race, cost of employment, and qualification levels, the relevance of financial literacy in thinking about retirement, planning for retirement or financial behavior is not as apparent. this is a somewhat hopeful outcome. there is little an individual can do to change those factors that determine his or her level of financial literacy. but, the steps to becoming a more successful planner are linear and achievable. proposed educational interventions should be focused on the decisions around determining and conforming to retirement plans. the improvement in such understanding can in turn, assist consumers to behave better financially and make better retirement savings decisions. fernandes et al. (2014) propose an immediate approach to financial education by targeting specific behaviors at the point in time when the need arises. however, consideration should also be given to other interventions, such as auto-enrolment plans and default options, to “nudge” individuals towards improved savings (thaler & sunstein, 2008, p. 67). note 1 the user written command for a multinomial model with instrumental variable by roodman (2011) is unable to cater for weighted data. thus, this output is seen as exploratory and preliminary. g.d. willows / financial services review 28 (2020) 243–271 263 appendix a socioeconomic questions: select the answer that most correctly reflects your situation question source q1 please indicate your marital status: h single h married h separated h divorced h widowed h living with a partner original to this study q2 do you have any dependents that is, someone who relies on you to financially support them? h yes h no original to this study answer if do you have any dependents that is, someone who relies on you to financially support them? yes is selected q3 what is your relationship with this dependent? (you may choose more than one option.) h my child/children h my spouse/partner h my brother/sister h my mother/father h other ____________________ original to this study q4 have you ever taken economics or accounting as a course or subject? (you may choose more than one option.) h yes, in high school. h yes, for a short period of time at university/college/technicon. h yes, as one of my majors at university/college/technicon. h no lusardi and mitchell (2017), amended. q5 how much have you thought about retirement? h a lot h some h only a little h not at all lusardi and mitchell (2017), response options tailored slightly (continued on next page) 264 g.d. willows / financial services review 28 (2020) 243–271 appendix a (continued) question source q6 have you ever tried to figure out how much your household would need to save for retirement? h yes h no lusardi and mitchell (2011) answer if have you ever tried to figure out how much your household would need to save for retirement? yes is selected q7 did you develop a plan for retirement saving? h yes h more or less h no lusardi and mitchell (2011) answer if did you develop a plan for retirement saving? yes is selected or did you develop a plan for retirement saving? more or less is selected q8 how often were you able to stick to this plan? h always h mostly h rarely h never lusardi and mitchell (2011) answer if have you ever tried to figure out how much your household would need to save for retirement? yes is selected q9 in what ways have you tried to figure out how much your household would need? you may select more than one option (if applicable) h i talked to family and relatives h i talked to co-workers or friends h i attended retirement seminars h i used calculators or worksheets that are computer or internet-based h i consulted a financial planner or advisor or an accountant h other ____________________ lusardi and mitchell (2011), “attended retirement seminars:” included by author as it is an option with the uctrf. g.d. willows / financial services review 28 (2020) 243–271 265 appendix b financial literacy questions: for each question, select the answer that you think is most correct question source numeracy q10 suppose you had r100 in a savings account and the interest rate was two percentage per year. after five years, how much do you think you would have in the account if you left the money to grow? h more than r102 h exactly r102 h less than r102 h do not know lusardi and mitchell (2011) inflation q11 imagine that the interest rate on your savings account was one percentage per year and inflation was two percentage per year. after one year, how much would you be able to buy with the money in this account? h more than today h exactly the same h less than today h do not know lusardi and mitchell (2011) time value of money q12 assume a friend inherits r10,000 today and his brother inherits r10,000 three years from now. who is richer because of the inheritance? h my friend h his brother h they are equally rich h do not know lusardi and mitchell (2017), “sibling” changed to “brother” for clarification. money illusion q13 suppose that in the year 2020, your income has doubled and prices of all goods have doubled too. in 2020, how much will you be able to buy with your income? h more than today h the same as today h less than today h do not know lusardi and mitchell (2017) function of the share market q14 which of the following statements describes the main function of the share market (also referred to as the “stock market” or “equity market”)? h the share market helps to predict share earnings h the share market results in an increase in the price of shares h the share market brings people who want to buy shares together with those who want to sell shares h none of the above h do not know lusardi and mitchell (2017), ‘stock’ (american terminology) replaced with “share” (south african terminology), and further example’s given to ensure clarity. (continued on next page) 266 g.d. willows / financial services review 28 (2020) 243–271 appendix b (continued) question source knowledge of shares q15 which of the following statements is correct? if somebody buys a share of company b in the share market: h he owns a part of company b h he has loaned money to company b h he is liable for company b’s debts h none of the above h do not know van rooij et al. (2011), “stock” (american terminology) replaced with “share” (south african terminology). knowledge of cis’s q16 which of the following statements is most correct? h once you invest in a collective investment scheme that is, “unit trust,” you cannot withdraw the money in the first year h unit trusts can invest in several asset classes, for example; shares/equity, bonds, property, and cash. h unit trusts pay a guaranteed rate of return that depends on their past performance h none of the above h do not know lusardi and mitchell (2017), “mutual fund” (american terminology) replaced with “collective investment scheme” (south african terminology), and lay terminology of “unit trust” also given to ensure clarity. knowledge of bonds q17 which of the following statements is correct? if somebody buys a bond issued by company b: h he owns a part of company b h he has loaned money to company b h he is liable for company b’s debts h none of the above h do not know van rooij et al. (2011) long-period returns q18 considering a long time period (e.g., 10 or 20 years), which asset normally gives the highest return? h savings accounts/cash h bonds h shares/equity h do not know lusardi and mitchell (2017), multiple terms for cash and equity given to ensure clarity. highest variability q19 normally, which asset displays the highest variability of return over time? h savings accounts/cash h bonds h shares/equity h do not know lusardi and mitchell (2017), multiple terms for cash and equity given to ensure clarity. risk diversification q20 complete the sentence. when an investor spreads his or her money among different assets, the risk of losing money should: h increase h decrease h stay the same as if the investor hadn’t spread his or her money h do not know lusardi and mitchell (2017), response for “stay the same” extended to make complete sentence. (continued on next page) g.d. willows / financial services review 28 (2020) 243–271 267 appendix b (continued) question source bond principles q21 true or false? if you buy a 10-year bond, it means you cannot sell it after five years without incurring a major penalty. h true h false h do not know van rooij et al. (2011) q22 true or false? equity/shares are normally riskier than bonds. h true h false h do not know lusardi and mitchell (2017), multiple terms for equity given to ensure clarity. q23 true or false? buying a share of a consumer company usually provides a safer return than a general equity unit trust. h true h false h do not know lusardi and mitchell (2017), “mutual fund” (american terminology) replaced with “unit trust” (south african terminology). q24 if the interest rate rises, what should happen to bond prices? h rise h fall h stay the same h none of the above h do not know lusardi and mitchell (2017) 268 g.d. willows / financial services review 28 (2020) 243–271 appendix c financial behavior and attitude questions: for each statement, circle your answer (or use the slider) on a scale from 1 to 5, where 1 means “never” and 5 means “always”. (source: atkinson & messy, 2012) statement never always before i buy something i carefully consider whether i can afford it 1 2 3 4 5 i pay my bills on time 1 2 3 4 5 i keep a close personal watch on my financial affairs 1 2 3 4 5 i set long term financial goals and strive to achieve them 1 2 3 4 5 i’m personally (or jointly) responsible for day to day money management decisions in my household 1 2 3 4 5 i live in a household with a budget 1 2 3 4 5 i borrow money to make ends meet 1 2 3 4 5 i find it more satisfying to spend money than to save it for the long term 1 2 3 4 5 i tend to live for today and let tomorrow take care of itself 1 2 3 4 5 money is there to be spent 1 2 3 4 5 appendix d instrumental variable: instrumental variable probit of thinking about retirement1 only a little some a lot financial literacy financial literacy 0.099** (0.044) 0.142*** (0.041) 0.098** (0.041) age 0.006 (0.014) 0.0320** (0.013) 0.071*** (0.013) 0.025 (0.027) marital status �0.188 (0.264) �0.022 (0.252) 0.015 (0.256) �0.062 (0.559) dependents 0.546* (0.305) 0.198 (0.284) 0.341 (0.289) �0.001 (0.582) gender: male �0.514* (0.284) �0.609** (0.267) �0.487* (0.270) 2.980*** (0.535) race: colored 0.369 (0.436) 0.081 (0.418) 0.195 (0.414) �4.589*** (0.670) race: african 0.873 (0.592) 0.167 (0.585) 0.535 (0.569) �6.687*** (0.985) race: indian 4.900 (245.6) 4.800 (245.6) 4.913 (245.6) �3.406** (1.191) race: other 3.860 (736.8) 3.939 (736.8) 3.654 (736.8) �1.657 (2.414) log coe 0.143 (0.447) �0.064 (0.427) 0.149 (0.426) 5.015*** (0.644) qualification: less than high school 1.072* (0.562) 1.006* (0.556) 1.131** (0.545) 8.243 (6.791) qualification: higher certificate/diploma 0.766 (0.516) 0.605 (0.492) 0.548 (0.491) 7.789 (6.804) qualification: college 0.445 (0.409) 0.397 (0.390) 0.340 (0.384) 7.604 (6.790) qualification: post-graduate 0.213 (0.471) 0.0238 (0.455) �0.025 (0.450) 9.892 (6.813) qualification: unknown 6.446 (6.780) accounting and economics score 3.506*** (0.404) constant �6.847 (4.359) �6.363 (4.195) �9.004** (4.194) �25.52** (10.09) lnsig_5 constant 1.904*** (0.0258) atanhrho_25 constant �0.496* (0.289) atanhrho_35 constant �0.687** (0.327) atanhrho_45 constant �0.317 (0.235) observations 753 wald x2(57) 1014.74 source: author’s calculations. standard errors in parentheses. 1the user written command for a multinomial model with instrumental variable by roodman (2011) is unable to cater for weighted data. thus, this output is seen as exploratory and preliminary. ***p< .01, ** p< .05, * p< .1. g.d. willows / financial services review 28 (2020) 243–271 269 acknowledgments the work is based on the research supported in part by the national research foundation of south africa for the grant, unique grant no. 94145. references agnew, j., bateman, h., & thorp, s. (2013). financial literacy and retirement planning in australia. numeracy, 6, article 7. agnew, j., & szykman, l. r. (2005). asset allocation and information overload: the influence of information display, asset choice, and investor experience. the journal of behavioral finance, 6, 57–70. allie, j., west, d., & willows, g. (2016). the value of financial advice: an analysis of the investment performance of advised and non-advised individual investors. investment analysts journal, 45, s63–s64. arrondel, l., debbich, m., & savignac, f. (2013). financial literacy and financial planning in france. numeracy, 6, article 8. atkinson, a., & messy, f. a. (2012). measuring financial literacy: results of the oecd/international network on financial education (infe) pilot study. oecd working papers on finance, insurance and private pensions, 15, 1–73. baron, r. m., & kenny, d. a. (1986). the moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. journal of personality and social psychology, 51, 1173–1182. bashall, j., willows, g. d., & west, d. (2018). the extent to which professional advice can reduce the disposition effect: an emerging market study. journal of emerging market finance, 17, 229–249. beckmann, e. (2013). financial literacy and household savings in romania. numeracy, 6, article 9. berg, g., & zia, b. (2013). financial literacy through mainstream media: evaluating the impact of financial messages in a south african soap opera. world bank working paper, washington, dc. brown, m., & graf, r. (2013). financial literacy and retirement planning in switzerland. numeracy, 6, article 6. brown, m., grigsby, j., van der klaauw, w., wen, j., & zafar, b. (2016). financial education and the debt behavior of the young. review of financial studies, 29, 2490–2522. bruhn, m., de souza leáo, l., legovini, a., marchetti, r., & zia, b. (2016). the impact of high school financial education: experimental evidence from brazil. american economic journal: applied economics, 8, 256–295. cole, s., paulson, a., & shastry, g. (2013). high school and financial outcomes: the impact of mandated personal finance and mathematics courses. boston, ma: manuscript, harvard business school. (available at http://www.hbs.edu/faculty/publication%20files/13-064_c7b52fa0-1242-4420-b9b6-73d32c639826. pdf) dave, r. (2017). here’s how to set your financial life in order to avoid problems for dependents. the economic times. (available at https://economictimes.indiatimes.com/wealth/plan/heres-how-to-set-yourfinancial-life-in-order/articleshow/58546285.cms) draper, k., & spaull, n. (2015). schooling in south africa: how low quality education becomes a poverty trap. south african journal of childhood education, 5, 34–41. fernandes, d., lynch, j. g., & netemeyer, r. g. (2014). financial literacy, financial education and downstream financial behaviors.management science, 60, 1861–1883. gough, d. (1996). english in south africa. in c. muller & m. wright (eds.), dictionary of south african english on historical principles. cape town: oup. goodman, j. (2009). the labor of division: returns to compulsory mathematics coursework. working paper, harvard kennedy school, cambridge, ma. klapper, l., & panos, g. a. (2011). financial literacy and retirement planning: the russian case. journal of pension economics and finance, 10, 599–618. lusardi, a., & mitchell, o. s. (2011). financial literacy and planning: implications for retirement wellbeing. in o. s. mitchell & a. lusardi (eds.), financial literacy. implications for retirement security and the financial marketplace (pp. 17–39). oxford: oxford university press. 270 g.d. willows / financial services review 28 (2020) 243–271 lusardi, a., & mitchell, o. s. (2017). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness. quarterly journal of finance, 07, 1750008. mckenna, b. (2015). as the baby boomers retire, the threat of intergenerational inequality looms. the globe and mail, 06 november. (available at https://www.theglobeandmail.com/globe-investor/retirement/as-thebaby-boomers-retire-the-threat-of-intergenerational-inequality-looms/article27163694/) miller, m., reichelstein, j., salas, c., & zia, b. (2015). can you help someone become financially capable? a meta-analysis of the literature. the world bank research observer, 30, 220–246. morrissey, m. (2016). the state of american retirement. retirement inequality chartbook, 03 march. (available at http://www.epi.org/publication/retirement-in-america/) national treasury. (2014).statement on the impact of the proposed retirement reforms. (available at http:// www.treasury.gov.za/comm_media/press/2014/2014070901-statement on the impact of the proposed retirement reforms.pdf) olsen, a., & whitman, k. (2007). effective retirement savings programs: design features and financial education. social security bulletin, 67, 53–72. roodman, d. (2011). fitting fully observed recursive mixed-process models with cmpo. the stata journal: promoting communications on statistics and stata, 11, 159–206. sekita, s. (2011). financial literacy and retirement planning in japan. journal of pension economics and finance, 10, 637–656. shillington, r. (2016).we’re facing a wave of seniors living in poverty and we’re not ready. ipolitics, 17 february. (available at https://ipolitics.ca/2016/02/17/were-facing-a-wave-of-seniors-living-in-poverty-andwere-not-ready/) skimmyhorn, w. (2016). assessing financial education: evidence from boot camp. american economic journal: economic policy, 8, 322–343. statistics south africa. (2018). quarterly labour force survey. quarter 2. (available at http://www.statssa.gov. za/publications/p0211/p02112ndquarter2018.pdf) taylor, v. (2002). transforming the present protecting the future. report of the committee of inquiry into a comprehensive system of social security for south africa. (available at http://www.ci.org.za/depts/ci/pubs/ pdf/poverty/news/report_release_of_coi.pdf) thaler, r., & sunstein, c. r. (2008). nudge: improving decisions about health, wealth and happiness (p. 6). (available at https://assets1c.milkeninstitute.org/assets/publication/mireview/pdf/67-85mr40.pdf) van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. willimack, d. k., lyberg, l., martin, j., japec, l., & whitridge, p. (2004).evolution and adaptation of questionnaire development, evaluation, and testing methods for establishment surveys. in: s. presser, j. m. rothgeb, m. p. cooper, j. t. lessler, e. martin, j. martin & e. singer (eds.),methods for testing and evaluating survey questionnaires. new york: wiley xu, l., & zia, b. (2012). financial literacy around the world: an overview of the evidence with practical suggestions for the way forward. policy research working paper, 6107, 1–56. g.d. willows / financial services review 28 (2020) 243–271 271 value line quarterly eps forecast error: analyst credibility or management appeasement? philip bairda,* aduquesne university, palumbo/donahue school of business, 600 forbes avenue, pittsburgh, pa 15282, usa abstract a study of value line quarterly earnings forecast errors from 1999 through q3 2016 shows that the direction of forecast bias and forecast efficiency with respect to earnings news depend on investment rating. patterns of bias and inefficiency indicate that value line analysts are primarily motivated to maintain credibility with investors than to appease company managers. for buy-rated stocks, forecast bias is pessimistic, and forecasts are inefficient with respect to good earnings news. when news is bad for buy-rated stocks, forecasts are unbiased and efficient. for sell-rated stocks, forecast bias is optimistic, and forecasts are inefficient with respect to bad earnings news. when news is good for sell-rated stocks, forecasts are unbiased and efficient. © 2017 academy of financial services. all rights reserved. jel classification: g11; g14 keywords: value line; earnings forecasts; earnings management; earnings forecast bias; earnings forecast inefficiency 1. introduction with roots dating to 1931, value line, inc. (symbol: valu) provides independent investment research that for many years has had substantial influence with individual investors. value line’s core business is producing investment periodicals and underlying research, and its target audiences are individual investors, colleges, libraries, and investment * corresponding author. tel.: �1-412-396-6246; fax: �1-412-396-4764. e-mail address: bairdp@duq.edu financial services review 26 (2017) 37–54 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. management professionals. in its annual report for 2016, value line, inc. reported total revenue from investment periodicals and related publications of $34.5 million. its flagship publication, the investment survey, has been published weekly since 1965. it has been recognized by hulbert financial digest in its newsletter honor roll, most recently in 2016. the investment survey delivers the research and ratings of value line’s analyst team for 1,700 public companies, which comprise roughly 90% of the market capitalization of u.s. common stocks. investment ratings are conveyed by value line’s timeliness rating system, which ranks stocks for relative 6-to-12-month-ahead stock price performance. the performance of value line’s timeliness ratings has been referred to in prior research as the “value line enigma.” in addition, the investment survey delivers both shortand long-term forecasts of 23 financial variables for the companies in its coverage universe. although investors spend tens of millions of dollars annually subscribing to value line publications, relatively little has been reported about the quality of the financial forecasts produced by the company’s analysts. the aim of this article is to help close this gap in the literature by reporting an analysis of the bias and inefficiency in value line’s quarterly earnings per share forecasts. there is a large literature on analysts’ earnings forecasts generally. perhaps the salient findings are that they are biased and inefficient—short-range forecasts are characterized by pessimism, long-range forecasts by optimism, and forecasts generally do not fully reflect available information. observed distributions of eps forecast error typically display a preponderance of large negative forecast errors vis-à-vis large positive errors (a “tail asymmetry”) and a preponderance of small positive forecast errors vis-à-vis small negative errors (a “middle asymmetry”). these asymmetries reflect that companies more often than not report earnings that meet or beat analysts’ forecasts and that when they miss they tend to miss by wide margins. they have been attributed to (a) incentives for reporting firms to take ‘earnings baths’ (i.e., large losses) when they cannot meet current earnings benchmarks, (b) incentives for reporting firms to meet or beat earnings forecasts by managing reported earnings and/or by guiding analysts to beatable forecasts, and (c) incentives for analysts to play along with the firms they cover in the earnings guidance game. among analysts employed by sell-side firms, incentives to produce optimistic forecasts arise from pressures to generate investment banking and trading revenue. among analysts generally, the pressure for optimistic forecasts also has been attributed to analysts’ need to appease the firms they cover for fear of losing access to management as the result of too-critical coverage. tempering these incentives to optimism are management guidance to analysts to issue beatable forecasts and analysts’ desire to maintain credibility with investor clients. in sum, the observed distribution of earnings forecast error is the outcome of a complex interplay of incentives faced by managers of public companies and the analysts who cover them. value line earnings forecasts are uniquely useful for unraveling, at least in part, this confounded knot of incentives. because value line is an independent research firm, its analysts do not support either investment banking or brokerage operations. hence, only two of the theories of analyst motivation cited in the literature can potentially explain the bias and inefficiency in value line forecasts, namely, management appeasement and analyst credibility. in addition, unlike most analysts who produce simultaneously both investment ratings 38 p. baird / financial services review 26 (2017) 37–54 and earnings forecasts, value line analysts produce forecasts only, and they do so with prior knowledge of stocks’ investment ratings. in this setting, patterns of forecast bias and inefficiency arguably differ between management appeasement and analyst credibility incentives. the literature indicates that one reason analysts issue biased forecasts is to appease or flatter management to obtain underwriting business and for fear of losing access to management information. if managers prefer that analysts issue beatable earnings forecasts, then analysts will generally oblige them by issuing forecasts that most firms will meet or beat. moreover, when revising their forecasts, analysts will generally respond more fully to bad earnings news than to good news. further, if beatable forecasts serve to counteract ill will stemming from undesirable investment ratings, analysts will respond more fully to bad news for low-rated stocks than they will for high-rated stocks, and when news is good they will respond more fully for high-rated stocks than for low-rated stocks. conversely, if, instead of appeasing company management, analysts seek to maintain or build credibility with investor clients, they will issue forecasts that favorably rated companies will most often meet or exceed, and they will issue forecasts that unfavorably rated companies will most often fail to meet. hence, forecast pessimism will increase with investment rating. additionally, analysts will respond more fully to bad news than to good news for favorably rated stocks, and they will respond more fully to good news than to bad news for unfavorably rated stocks. empirical results reported herein shed new light directly on the motivations of value line analysts in particular and indirectly on the motivations of analysts generally. the distribution of value line quarterly eps forecast error from 1999 through the third quarter of 2016 is consistent with prior studies of analysts’ earnings forecast error. the overall mean (median) forecast error is �1.4 (0.0) cents per share, and the frequency of positive errors exceeds that of negative errors by 49 to 43%. these results and others reported herein indicate the presence of both middle and tail asymmetries in the forecast error distribution, similar to that documented in prior research. what has not been reported previously, however, is that, although these asymmetries appear in the entire sample, segmenting by investment rating shows that the middle asymmetry is confined primarily to stocks rated buy and the tail asymmetry is confined to stocks rated sell. as well, value line earnings forecast bias and inefficiency display distinctly different patterns across buy, hold, and sell rating categories. forecast pessimism is increasing in investment rating; the mean (median) forecast errors for stocks rated buy, hold and sell are 2.3 (2.0), �1.0 (0.0), and �6.5 (�2.0) cents per share, respectively. moreover, among stocks rated buy, value line forecasts are unbiased and efficient with respect to prior bad news, while at the same time they are biased and inefficient with respect to prior good news. the opposite pattern emerges among stocks rated sell, for which forecasts are unbiased and efficient with respect to prior good news and biased and inefficient with respect to prior bad news. these patterns of bias and inefficiency point to analysts’ desire to maintain credibility with investor clients, as opposed to appeasing company management, as the dominant force motivating analyst behavior. additionally, the time pattern of forecast error over the one-year forecast horizon before quarter end is better understood in terms of analyst credibility than in terms of management appeasement. 39p. baird / financial services review 26 (2017) 37–54 2. prior research 2.1. earnings forecast bias the body of published research on analysts’ earnings forecasts is extensive. a dominant theme emerging from this literature is that because reported earnings, on average, fail to meet analysts’ forecasts, researchers typically conclude that analysts issue optimistic earnings forecasts (e.g., abarbanell & lehavy, 2003b; agrawal & chen, 2006; bradshaw et al., 2006; brous, 1992; brous & kini, 1993; butler & lange, 1991; dreman & berry, 1995a; easterwood & nutt, 1999; francis & philbrick, 1993; fried & givoly, 1982; kang et al., 1994; and o’brien, 1988). this conclusion, however, is confounded by the fact that the frequency with which reported earnings meet or beat forecasts typically exceeds the frequency with which they fail to meet forecasts. analysts’ incentives to issue optimistic forecasts have been attributed to their employers’ investment banking and brokerage operations and to their fear of losing access to management as the result of negative research (e.g., cowen et al., 2006; dugar & nathan, 1995; francis & philbrick, 1993; ljungqvist et al., 2007; michaely & womack, 1999; richardson et al., 2004). mitigating these incentives for optimism are analysts’ desire to establish and maintain credibility with investor clients (cowen et al., 2006; lin & mcnichols, 1998; raedy et al., 2006) and management guidance to lower forecasts as quarter end approaches (richardson et al., 2004). richardson et al. (2004) report that in the late 1990s institutional and regulatory changes increased managers’ incentives to guide analysts to beatable forecasts, so that optimism at longer forecast horizon becomes pessimism as quarter end approaches. managers’ incentives to meet or beat analysts’ forecasts derive from the differential stock price impact of a quarterly earnings miss versus that of a meet or beat (skinner & sloan, 2002) and from valuation premia for firms that consistently meet or beat estimates (bartov, givoly, & hayn, 2002; kasznick & mcnichols, 2002). firms seek to meet or beat forecasts by managing reported earnings (abarbanell & lehavy, 2003a, 2003b; bartov et al., 2002; burgstahler & dichev, 1997; degeorge et al., 1999; matsumoto, 2002) and by guiding analysts to beatable forecasts (cotter et al., 2006; matsumoto, 2002; richardson et al., 2004). analysts’ incentives to participate in the earnings guidance game stem from their dependence on management for future information and from their employers’ underwriting activities (cotter et al., 2006; dugar & nathan, 1995; lin & mcnichols, 1998; michaely & womack, 1999; richardson et al., 2004). some research shows that earnings forecast bias depends on investment rating. francis and philbrick (1993) test whether analysts’ earnings forecasts are more optimistic for stocks rated sell than for those rated hold and similarly for those rated hold versus buy. they assume that sell ratings undermine relations with management and that analysts issue relatively optimistic earnings forecasts to compensate. if, however, managers prefer that analysts issue beatable forecasts, it seems likely the impact of rating on forecast bias would run in the direction opposite to that hypothesized by francis and philbrick (1993). that is, earnings forecast optimism (pessimism) would be lower (higher) for sells than for holds and similarly for holds versus buys, because analysts would issue earnings forecasts that lower-rated stocks are more likely to meet or beat. francis and philbrick (1993) find that 40 p. baird / financial services review 26 (2017) 37–54 mean unscaled forecast error for buys is not significantly different from zero and that mean forecast errors for holds and sells are not significantly different from buys or from each other.1 additional evidence that investment rating influences earnings forecast error is provided by abarbanell and lehavy (2003a) who argue that stock rating affects firms’ incentive to manage reported earnings. their results, along with abarbanell and lehavy (2003b) and cohen and lys (2003), can explain the preponderance of small positive versus small negative forecast errors (the “middle” asymmetry) and the preponderance of large negative versus large positive forecast errors (the “tail” asymmetry) that are typically observed in earnings forecast error distributions.2 abarbanell and lehavy (2003a) find that firms rated buy (sell) are more (less) likely to meet or beat analysts’ forecasts. they explain this in terms of firms’ incentives to manage reported earnings, but they do not consider that this pattern of forecast error across buy/hold/sell ratings might be attributable to analyst forecast bias. prior research indicates forecast bias depends on the analyst’s employer type. cowen et al. (2006) find that analysts employed by firms with significant underwriting and trading operations make less optimistic forecasts than those at brokerage firms that do no underwriting and that forecast optimism is especially low among bulge underwriter firm analysts. their results imply that the importance of analyst (firm) reputation reduces forecast optimism. ljunqvist et al. (2007) find that the presence of institutional investors in stocks is associated with more accurate earnings forecasts. lin and mcnichols (1998) find that earnings forecasts issued by affiliated analysts are generally not more optimistic around seasoned equity offerings than those issued by unaffiliated analysts.3 these results suggest that analysts’ desire to establish and maintain credibility among investor clients can impact their forecasts. as explained below, the present research tests the hypothesis that analysts’ desire to maintain credibility with investors leads them to issue forecasts that firms rated buy will generally meet or beat and that firms rated sell will generally miss. 2.2. earnings forecast efficiency if analysts’ forecasts efficiently incorporate relevant information about future earnings, then their forecast errors are uncorrelated with available information. forecast errors have been found to be correlated with past stock returns, past earnings changes, and prior forecast errors (ali et al., 1992; abarbanell & bernard, 1992; shane & brous, 2001). cohen and lys (2003) state “analysts underreact to both prior good news and prior bad news and are, thus, inefficient” (p.155). raedy et al. (2006) provide a rational economic explanation for analyst underreaction to forecast error; namely, for an error of given magnitude, analyst credibility is damaged when later information causes a forecast revision of the opposite sign than the analyst’s prior revision. hence, analysts’ loss functions are asymmetric with respect to the sign of earnings forecast error. extending this thought, it seems reasonable that analyst credibility is weakened when stocks on their buy lists fail to meet earnings forecasts and, conversely, when stocks on sell lists meet or exceed forecasts. this suggests a tendency to maintain forecasts that buy-rated firms are likely to meet or beat and that sell-rated firms are not likely to meet. as a result, analysts will efficiently incorporate bad earnings news into their forecasts for stocks on their buy lists, and they will respond more slowly to good news. 41p. baird / financial services review 26 (2017) 37–54 for sell-rated stocks, analysts will efficiently incorporate good news into their forecasts and respond more slowly to bad news. the present research explores the possibility that the nature of the asymmetry in analysts’ loss function, and hence of the inefficiency in their forecasts, depends on investment rating. basu and markov (2004) show that if analysts’ loss function is linear, as opposed to quadratic, then ordinary least squares (ols) regression tests of rationality are misspecified. if analysts seek to minimize absolute forecast error, then regression tests of rationality should be estimated by the method of least absolute deviation (lad). ols and lad estimators are analogues of the sample mean and median, and in regression they are consistent and asymptotically normal estimators of the population mean and median of the dependent variable conditional on the explanatory variables. however, because the sample median is relatively insensitive to extreme observations, the lad estimator is a robust estimator for skewed, fat-tailed distributions. in the present research, the unbiasedness and efficiency of analysts’ forecasts are examined in lad regression framework in which model parameters are allowed to vary by investment rating. this econometric setting enables tests of competing theories of analyst behavior, management appeasement and analyst credibility, which possess clearly distinguishable implications for forecast bias and inefficiency. 2.3. value line earnings forecasts published research on value line’s quarterly eps forecasts is scant. philbrick and ricks (1991) find that from 1984 to 1986 mean and median quarterly earnings forecast errors are a statistically significant �0.37% and �0.04% of stock price, respectively. francis and philbrick (1993) find that from 1987 to 1989 mean unscaled quarterly eps forecast errors for stocks rated buy, hold and sell are not significantly different from zero.4 ramnath et al. (2005) find that from 1993 to 1997 mean and median quarterly earnings forecast errors are a statistically significant �0.054% and 0.011% of stock price, respectively. das et al. (1998) investigate value line’s annual earnings forecasts from 1989 to 1993 and report that the mean forecast error is a statistically significant �1.5% of stock price. szakmary et al. (2008) find that 3–5 year ahead earnings forecasts display large optimistic bias. 3. methodology and data 3.1. quarterly eps forecast error and investment rating value line analysts produce multiple forecasts for a given quarter with initial forecasts typically issued more than a year before quarter end. quarterly eps forecast error is calculated as actual eps minus the eps forecast issued at horizon h before quarter end: feit h � eit � fit h (1) eit denotes actual eps for company i in quarter t. fit h denotes the value line forecast issued at horizon h. feit �1 denotes the error of the forecast issued approximately one year before 42 p. baird / financial services review 26 (2017) 37–54 quarter end. feit 0 denotes the error of the latest forecast before quarter end. the mean (median) number of days from the date of the latest forecast to quarter end is 16 (17). the horizon profile of forecast error is characterized below in section 4 in intervals from h � �1 to 0. forecast errors are often scaled (i.e., deflated, normalized, divided) by stock price even though the practice can confound interpretation. for example, in abarbanell and lehavy (2003a) the apparent relation between unexpected accruals and earnings forecast error could be an artifact of scaling by stock price.5 the distribution of the p/e ratio reflects investor expectations for future performance. consequently, earnings forecast error of a given dollar amount likely has a larger impact on the price of a high-p/e stock than of a low-p/e stock. yet, scaling the error by stock price reverses its measured impact. scaling by stock price also can introduce time dependence in measured forecast error to the extent that common stock earnings multiples vary over the sample period. cohen and lys (2003) show that scaling by stock price magnifies the left tail of the distribution of earnings forecast error. results reported below in section 4 show the same effect of price scaling, which also has the effect of increasing the clustering of observations near the mean of the scaled forecast error distribution. value line’s timeliness rank conveys the predicted relative 6-to-12-month price performance of the approximately 1,700 stocks in its coverage universe. each stock is ranked from 1 (highest) to 5 (lowest) by a quantitative model of ex post earnings and stock price performance. as such, the timeliness rank is independent of value line analysts’ input. because timeliness is a relative rank, the number of stocks in each rank at each point in time remains constant as shown in table 1. the total number of sample firm-quarter observations in each rank is shown in the rightmost column. in the present study, it is assumed that rank 1 and 2 stocks are recommended buys, rank 3 stocks are holds, and rank 4 and 5 stocks are sells. 3.2. rationality of analysts’ forecasts bias and inefficiency in analysts’ forecasts are investigated in the regression framework proposed by basu and markov (2004): et � �0 � �1ft 0 � �2fet�1 0 � �t (2) if analysts’ forecasts are unbiased, then �0 � 0 and �1 � 1. �0 � 0 captures forecast bias that is uncorrelated with the forecast, and �1 � 1 captures forecast bias that is correlated with the forecast. if forecasts are efficient with respect to prior earnings news then �2 � 0. prior table 1 timeliness rank number of stocks predicted 6–12 month stock performance firm-quarter observations 1 100 highest 3,843 2 300 above average 11,105 3 �900 average 27,923 4 300 below average 9,969 5 100 lowest 3,487 �1,700 56,327 43p. baird / financial services review 26 (2017) 37–54 earnings news is assumed to be captured by prior-quarter earnings forecast error such that fet�1 0 � 0 (fet�1 0 � 0) indicates good (bad) news. the possibility exists that forecasts are unbiased (�0 � 0; �1 � 1) yet inefficient (�2 � 0). subtracting the latest current-quarter forecast from both sides of (2) results in: fet 0 � et � ft 0 � �0 � �1�ft 0 � �2fet�1 0 � �t (3) where �1� � �1 � 1. for �1� � 0, forecast optimism (pessimism) is decreasing (increasing) in the forecast; that is, the signed forecast error is increasing in the forecast. for �1� � 0, forecast optimism (pessimism) is increasing (decreasing) in the forecast. for �2 � 0 (�2 � 0), analysts underreact (overreact) to prior earnings news. coefficients in eq. (3) are estimated in lad regression with dummy variables to capture quarterly calendar effects.6 for a discussion of lad estimation, see portnoy and koenker (1997). hypothesis tests on lad-estimated coefficients are conducted based on the resampling method described in chen et al. (2008).7 in interpreting regression results, it is assumed that managers prefer that analysts issue beatable forecasts (richardson et al., 2004). if analysts seek to appease managers, they will bias their forecasts so as to mitigate ill will created by unfavorable investment ratings. analysts’ forecasts will display a pessimistic bias for lower-rated stocks such that stocks rated sell will more often meet or beat forecasts than will stocks rated hold and similarly for stocks rated hold versus those rated buy. hence, by the management appeasement hypothesis (mah) it will be the case that in eq. (3) �0 sell � �0 hold � �0 buy � 0. that is, the estimated constant term for sells will exceed its value for holds, which will exceed its value for buys, which will be non-negative if managers prefer that analysts issue beatable forecasts and analysts seek to appease managers. moreover, if analysts generally strive to maintain beatable forecasts, they will respond more efficiently to bad news than to good news, and if they are more strongly incented to maintain beatable forecasts for sells than for buys, they will respond more efficiently to bad news for sells than to bad news for buys and more efficiently to good news for buys than to good news for sells. if the mah describes analyst behavior, then the pattern of inefficiency in analysts’ forecasts will be reflected in the coefficient relations shown in table 2. if, instead of seeking to appease managers, analysts seek to maintain credibility among investors, they will bias earnings forecasts such that firms rated buy will more often meet or beat forecasts than will firms rated hold, and those rated hold will more frequently meet or beat forecasts than will those rated sell. that is, by the analyst credibility hypothesis (ach) an overall optimistic (pessimistic) bias will be decreasing (increasing) in investtable 2 management appeasement hypothesis (mah) implications for forecast efficiency prior bad news (fet�1 0 � 0) prior good news (fet�1 0 � 0) �2 sell � �2 hold � �2 buy � �2 buy � �2 hold � �2 sell 44 p. baird / financial services review 26 (2017) 37–54 ment rating, and the constant term in eq. (3) will be such that �0 sell � 0 � �0 buy and �0 sell � �0 hold � �0 buy. further, if analysts seek to maintain forecasts that firms rated buy are likely to meet or beat, they will respond more efficiently to bad news than to good news for buys, and if they seek to maintain forecasts that firms rated sell are likely to miss, they will respond more efficiently to good news than to bad for sells. hence, for stocks rated buy (sell), analysts’ forecasts will be relatively inefficient with respect to good (bad) prior earnings news. if the ach describes analyst behavior, then the patterns of inefficiency in their forecasts will be reflected in the coefficient relations shown in table 3. 3.3. data and sample each weekly issue of the investment survey contains reports on a set of companies organized by industry so that updated reports are published quarterly for all companies in value line’s universe. each company report contains historical financial data, target stock price, and forecasts of quarterly and annual earnings, sales, cash flow, dividends, and more. these data are obtained from value line’s earnings and projections file beginning 1988 through the third quarter of 2016. the sample excludes firms with non-december fiscal years and firms in agriculture, financial services, fishing, forestry, and public administration industries. from 1988 through 1998 value line forecasts display significant optimistic bias that largely disappears beginning in 1999 and thereafter. value line, inc. annual reports from 1999 and 2000 mention technological initiatives underway at the time to upgrade information systems and revisions to the salary structure in the firm’s research department. because these initiatives and revisions presumably led to substantial reduction in optimistic forecast bias beginning in 1999, the sample period of analysis begins with that year. 4. results and discussion 4.1. the forecast error distribution restricting the sample to short-horizon (h � 0) forecasts, there are 56,898 firm-quarter observations for 2,170 unique firms. panel a of table 4 reports selected percentiles of the distribution of unscaled forecast error both with (row 1) and without (row 2) extreme 0.5% tails. it is not possible to know if extreme observations in the untrimmed data (row 1), particularly those in the left tail, are the result of data errors, so ensuing analyses utilize trimmed data. however, it should be noted that lad regression results reported below hold table 3 analyst credibility hypothesis (ach) implications for forecast efficiency prior good news (fet�1 0 � 0) �2 sell � �2 hold � �2 buy prior bad news (fet�1 0 � 0) �2 sell � �2 hold � �2 buy 45p. baird / financial services review 26 (2017) 37–54 up in the untrimmed sample. row 2 shows that negative skewness remaining in the trimmed sample is greatly reduced vis-à-vis the untrimmed sample. the median forecast error is $0.00, and 50% of sample errors are of magnitude less than or equal to 5 cents per share. panels b–d of table 4 report percentiles of standardized distributions of trimmed forecast error and its components, earnings and forecasts, both unscaled and scaled by stock price. each distribution is standardized by subtracting its mean and dividing by its standard deviation so that table entries show percentiles in terms of units of standard deviation from the mean. comparable statistics are shown for the standard normal distribution in panel e. panel b shows that scaling forecast error by stock price drastically increases negative skewness and the length of both tails while simultaneously exacerbating the central clustering of observations. the minimum unscaled error lies 8.4 standard deviations below its mean, whereas the minimum scaled error lies more than 47 standard deviations below its mean. scaling has a similar effect for the components of forecast error, except that in the case of earnings and forecasts scaling by stock price changes skewness from positive to negative. comparison with the standard normal distribution shows that scaling by price also increases clustering of observations near the mean. for example, panel b shows that 50% of unscaled errors lie within �/�0.8 standard deviations of the mean, while 50% of scaled errors are within �/�0.3 standard deviations of its mean. similarly, inspection of panels c and d indicates that scaling by stock price increases the central clustering of the distributions of both actual earnings and forecasts. the minimum stock price in the sample is $1.60; hence, these effects of scaling cannot be attributed to very low stock prices. remaining analyses focus on trimmed, unscaled forecast errors. table 4 the distribution of value line quarterly eps forecast error and components percentiles n min 0.5 5 25 50 75 95 99.5 max skew a: forecast error ($) 1-untrimmed 56,898 �231.0 �1.64 �0.33 �0.05 0.00 0.05 0.24 0.87 9.38 �156.1 2-trimmed 56,327 �1.6 �1.03 �0.31 �0.05 0.00 0.05 0.22 0.63 0.87 �2.0 b: forecast error, standardized unscaled �8.4 �5.3 �1.5 �0.8 0.1 0.8 1.2 3.3 4.6 �2.0 scaled �47.5 �4.6 �0.7 �0.3 0.1 0.3 0.5 1.8 20.2 �15.4 c: earnings, standardized unscaled �13.5 �2.4 �1.0 �0.7 �0.2 0.9 1.5 4.3 33.3 5.1 scaled �90.5 �4.5 �0.7 �0.3 0.1 0.4 0.6 1.4 9.2 �25.9 d: forecasts, standardized unscaled �15.2 �2.0 �0.9 �0.7 �0.2 0.9 1.4 4.4 34.4 5.9 scaled �95.7 �4.2 �0.8 �0.4 0.1 0.5 0.7 1.8 15.2 �25.4 e: normal (0,1) �2.6 �1.6 �1.3 0.0 1.3 1.6 2.6 0 panel a shows percentiles of the distribution of unscaled quarterly earnings forecast error from 1999 to 2016 q3 for the entire sample (row 1) and after trimming extreme 0.5% tails (row 2). panels b, c, and d report percentiles of standardized distributions of trimmed forecast error and its components, earnings and forecasts, both unscaled and scaled by stock price and multiplied by 100. table entries for panels b–d show percentiles of each standardized distribution in terms of units of standard deviation from the mean. comparable statistics are shown for the standard normal distribution in panel e. 46 p. baird / financial services review 26 (2017) 37–54 table 5 illustrates the presence of middle and tail asymmetries in the forecast error distribution. row a shows the proportional distribution of forecast errors in intervals of absolute forecast error centered on zero. for example, 8% of forecast errors equal zero, 5% of errors are less than or equal to 1 cent in absolute value, and so forth, 51% of the sample is comprised of forecast errors of magnitude less than or equal to 5 cents. row b shows that in the complete sample the ratio of positive errors to negative errors p/n equals 1.15 and that only among errors larger than 10 cents magnitude does the number of negative errors exceed the number of positive errors. the relative frequency of small positive errors versus small negative errors evidences the middle asymmetry, and the relative frequency of large negative errors versus large positive errors evidences the tail asymmetry. 4.2. the horizon profile of forecast bias richardson et al. (2004) demonstrate that earnings forecast bias switches from optimism to pessimism as the forecast horizon diminishes. this pattern of bias over the forecast horizon is attributed to analysts’ desire to support their employers’ underwriting and trading operations at longer horizons and to play along with managers in the earnings guidance game in which managers guide analysts to beatable forecasts just before quarter end. in the case of value line, which has no underwriting or trading operations, the horizon profile of forecast bias reflects analysts’ desire to either appease managers or to maintain credibility with investors. the mah implies that the horizon profile will resemble that in richardson et al. (2004), where long-horizon optimism becomes short-horizon pessimism. moreover, the profile will be more pronounced for sells than for buys if analysts bias their earnings forecasts so as to mitigate the ill will that is presumably attributable to poor investment ratings. that is, long-horizon optimism and short-horizon pessimism will both be exaggerated for sells vis-à-vis buys. conversely, the ach implies that forecast bias will be such that stocks rated buy will most often meet or beat forecasts and those rated sell will most often fail to meet forecasts. ach implications for the shape of the horizon profile are not clear. fig. 1 displays the horizon profile of mean forecast error by investment rating from one year prior (h � �1) to within 30 days (h � 0) of quarter end. consistent with mah, mean table 5 value line quarterly eps forecast error by magnitude of error intervals of absolute forecast error ($) �fet 0� 0 [0–.01) [.01–.02) [.02–.03) [.03–.04) [.04–.05) [.05–.10) �0.10 total a 8 5 14 11 7 7 20 29 100% b 1.26 1.33 1.54 1.41 1.38 1.23 0.84 1.15 row a shows the proportional distribution of forecast errors in intervals of absolute forecast error around zero. for example, 8% of forecast errors equal zero, 5% of errors are less than or equal to $0.01 in absolute value, and so forth. row b shows the ratio of positive to negative forecast errors p/n in each interval of absolute error. for example, among errors larger than $0.10 in magnitude p/n � 0.84, and in the entire sample p/n � 1.15. the sample is trimmed at the extreme 0.5% tails of the error distribution. 47p. baird / financial services review 26 (2017) 37–54 forecast error at long-horizon indicates optimism that is decreasing in investment rating; that is, analysts are optimistic for all stocks, but they are more optimistic for sells versus holds and for holds versus buys. if, however, analysts seek to appease managers, and if managers prefer that analysts issue beatable forecasts, then forecast optimism will become pessimism before quarter end. clearly, except for buys, this does not happen. hence, it does not appear from the horizon profile of forecast bias that analysts seek to appease managers. rather, the horizon profile suggests that, consistent with analyst credibility, forecast bias is such that, on average, stocks rated buy meet or beat forecasts, stocks rated sell fail to meet, and stocks rated hold lie between these cases. 4.3. earnings forecast bias table 6 characterizes the forecast error distribution in the complete sample and segmented by investment rating. panel a shows that in the entire sample the mean (median) forecast error equals �1.4 (0.0) cents, while positive forecast errors exceed negative errors by a ratio fig. 1. the horizon profile of mean eps forecast error by investment rating: 1999–2016 q3. notes: h � �1 denotes 1-year before quarter end; h � 0 denotes �30 days before quarter end. table 6 quarterly eps forecast error distribution in the complete sample and by investment rating forecast error (fet 0) ratio of positive to negative forecast errors (p/n) mean median [0–.01) [.01–.02) [.02–.03) [.03–.04) [.04–.05) [.05–.10) �0.10 total a: complete sample �0.014 0.000 1.26 1.33 1.54 1.41 1.38 1.23 0.84 1.15 b: investment rating buy 0.023 0.020 1.76 1.72 2.14 2.21 2.13 2.20 1.88 1.98 hold �0.010 0.000 1.21 1.29 1.51 1.32 1.37 1.22 0.89 1.16 sell �0.065 �0.020 0.84 1.01 1.05 0.95 0.87 0.70 0.39 0.65 table entries show mean and median quarterly eps forecast error and ratios of positive to negative forecast error (p/n) by magnitude of absolute forecast error. panel a pertains to the entire sample. panel b segments results by investment rating. 48 p. baird / financial services review 26 (2017) 37–54 of 1.15 to 1 (i.e., by 49 to 43%). as can be seen in ratios of positive-to-negative errors, positive errors are most frequently of small magnitude (�10 cents), and negative errors are most frequently of large magnitude (�10 cents). this is consistent with the presence of both middle and tail asymmetries. panel b shows that by a ratio of nearly 2-to-1 positive forecast errors for stocks rated buy exceed negative errors. among stocks rated sell, the ratio of positive-to-negative errors equals 0.65, which indicates more than 1.5 negative errors for each positive error. among stocks rated buy, the prevalence of positive versus negative errors is apparent across all magnitudes of error, and there is no evidence of a tail asymmetry. among stocks rated sell, there is virtually no evidence of a middle asymmetry; only a tail asymmetry is apparent. these results are consistent with the ach, which implies that analysts bias their forecasts such that buy-rated stocks will likely meet or beat their forecasts and that sell-rated stocks will likely miss forecasts. 4.4. earnings forecast efficiency the mah implies that analysts bias their forecasts so that firms generally meet or beat them and that the pessimistic bias is more pronounced among stocks rated sell than hold and buy. further, the mah implies that analysts generally respond more efficiently to bad earnings news than to good news. moreover, when news is bad analysts respond more efficiently for sells than for buys and when news is good they respond more efficiently for buys than sells. the ach implies that analysts bias their forecasts so that stocks rated buy (sell) most often meet or beat (fail to meet) forecasts. the ach implies further that when news is bad analysts respond more efficiently for buys than sells, and when news is good they respond more efficiently for sells than buys. these implications are examined empirically in results for eq. (3). table 7 presents regression results for eq. (3), for which parameters are estimated by the method of lad. hypothesis tests on the estimated coefficients are based on methods table 7 earnings forecast bias and inefficiency n �̂0 �̂1 �̂2 a: complete sample 51,956 0.002 0.004a 0.233a b: buys 13,647 0.005a 0.011a 0.259a c: holds 25,929 0.003a 0.003 0.187a d: sells 12,380 �0.004a �0.014a 0.235a e: prior bad news (fet�1 0 � 0) 22,143 0.002 �0.007a 0.205a buy 3,032 0.002 �0.002 0.056 hold 10,977 0.002 �0.007 0.154a sell 8,134 �0.003 �0.021a 0.246a f: prior good news (fet�1 0 � 0) 29,813 0.002a 0.009a 0.231a buy 10,633 0.004a 0.008a 0.320a hold 14,943 0.004a 0.007a 0.175a sell 4,237 �0.001 �0.002 0.089 regression results for eq. (3): fet 0 � �0 � �1�ft 0 � �2fet�1 0 � �t. parameters are estimated by least absolute deviation regression—see portnoy and koenker (1997). aindicates statistical significance at 1% based on the resampling method described in chen et al. (2008). 49p. baird / financial services review 26 (2017) 37–54 described in chen et al. (2008). if analysts’ forecasts are unbiased and efficient with respect to prior earnings news, then all coefficients in eq. (3) equal zero. panel a presents results for the entire sample of trimmed observations.8 the constant �̂0 � .002 is not significantly different from zero. the forecast coefficient �̂1� � .004 is significantly positive, which indicates the presence of forecast bias that is correlated with the forecast; that is, forecast pessimism (optimism) is increasing (decreasing) with the forecast. the prior news coefficient �̂2 � .233 is significantly positive, which indicates that analysts underreact to (and, hence, their forecasts are inefficient with respect to) prior earnings news. panels b–d of table 7 present results for stocks rated buy, hold and sell, respectively, and they indicate the presence of forecast bias and inefficiency in each rating category. the relation among the estimated constant terms is consistent with the ach: �̂0 sell � 0 � �̂0 buy and �̂0 sell � �̂0 hold � �̂0 buy. also, the prior news coefficients suggest that forecast inefficiency is more pronounced for stocks rated buy or sell than for stocks rated hold. hence, investment rating seems to impact the manner in which analysts respond to earnings news. panels e and f of table 7 present results for observations segmented by prior earnings news, where non-negative (negative) prior-quarter forecast error is taken to be good (bad) news. within each prior news subset, observations are further segmented by investment rating. panel e of table 7 indicates there are 22,143 observations of prior bad news. for these observations, the estimated constant �̂0 � 0.002 is not significantly different from zero. the slope �̂1� � �0.007 is significantly negative, which indicates forecast optimism (pessimism) that is increasing (decreasing) in the forecast. the slope �̂2 � 0.205 is significantly positive, which indicates that analysts’ forecasts are inefficient with respect to prior bad news. segmenting prior bad news observations by investment rating shows that the bias and inefficiency in analysts’ forecasts can be traced primarily to stocks rated sell. for these stocks, results indicate both bias and inefficiency in analysts’ forecasts. for stocks rated hold, analysts’ forecasts appear unbiased (�̂1 � �̂1� � 0) but inefficient with respect to prior bad news (�̂2 � 0). for stocks rated buy, none of the coefficients is significantly different from zero, which indicates that for these stocks analysts’ forecasts are unbiased and efficient with respect to bad news. panel f of table 7 shows there are 29,813 observations of prior good news. for these observations, the estimated constant �̂0 � 0.002 is significantly positive, which indicates a pessimistic bias that is uncorrelated with the earnings forecast and with prior news. the slope �̂1� � 0.009 is significantly positive, which indicates forecast pessimism (optimism) that is increasing (decreasing) in the forecast. the slope �̂2 � 0.231 is significantly positive, which indicates that analysts’ forecasts are inefficient with respect to good news. segmenting prior good news observations by investment rating shows that the bias and inefficiency can be traced to stocks rated buy and hold. for these stocks, all of the estimated coefficients are significantly different from zero. for stocks rated sell, however, none of the coefficients is significantly different from zero, which indicates that for these stocks analysts’ forecasts are unbiased and efficient. the patterns of bias and inefficiency indicated in table 7 are consistent with analysts seeking to maintain credibility with investors rather than to appease managers of the companies they follow. 50 p. baird / financial services review 26 (2017) 37–54 5. summary and conclusion equity analysts occupy a precarious position balancing the often conflicting expectations of investor clients and managers of public companies. investors expect analysts will produce unbiased, objective research on the companies they cover. doing so, however, puts analysts at risk of losing access to company managers who provide information analysts need in their work but who may be sensitive to critical coverage. analysts can be further conflicted in striving to support their employers’ investment banking and brokerage activities. in the end, their research reports and recommendations embody the subjective tradeoffs they make in balancing these conflicts. bias and inefficiency in earnings forecasts have been studied extensively by researchers seeking to understand analysts’ motives. the present research continues in this vein. its primary contribution is that it sheds new light on analysts’ motives in a setting in which they are free from the pressures of supporting underwriting and trading activities. because value line is not involved in these activities, and because stocks’ investment ratings are assigned by value line with no analyst input, value line earnings forecast errors are particularly useful for assessing analysts’ incentives to maintain credibility with investors and to appease company management. empirical results reported in the present research support several inferences. first, analysts appear to bias their forecasts primarily to maintain credibility with investors. if it were the case that analysts strive to appease company managers, and assuming that managers prefer that analysts issue beatable forecasts, then most firms will report earnings that meet or beat forecasts. prior research establishes that this is indeed the case generally, and the present research shows that this finding holds among value line earnings forecasts as well. however, the present research also shows that the propensity for reported earnings to meet or beat forecasts is confined primarily to stocks rated buy. more generally, the direction of earnings forecast bias depends on investment rating. the bias is pessimistic for buy-rated stocks, which by a ratio of almost 2 to 1 report earnings that meet or beat forecasts. the bias is optimistic for sell-rated stocks, for which earnings misses prevail by a ratio of more than 1.5 to 1. hence, it does not appear that value line analysts strive to appease the managers of companies to which sell ratings are attached. rather, it appears that analysts strive to maintain credibility with investors by issuing earnings forecasts that buy-rated stocks are likely to meet or beat and that sell-rated stocks are likely to miss. the present research documents a pattern of forecast inefficiency with respect to earnings news that further supports the view that analysts strive to maintain credibility. although prior research establishes that analysts underreact to earnings news, the present research shows that they do so under certain conditions. value line analysts underreact to good news for buy-rated stocks and to bad news for sell-rated stocks. however, they respond efficiently to bad news for buy-rated stocks and to good news for sell-rated stocks. these results are consistent with analysts striving to maintain pessimistic forecasts for buys and optimistic forecasts for sells, which again suggests that analysts seek primarily to maintain credibility. this interplay among investment rating and earnings forecast bias and inefficiency illuminate the incentives motivating analysts to an extent not elaborated heretofore. 51p. baird / financial services review 26 (2017) 37–54 notes 1 see their table 3, panel b. 2 evidence of these asymmetries is also apparent in bartov et. al. (2002), burgstahler and dichev (1997), brown (2001), cotter, et. al. (2006), degeorge et. al. (1999), kasznick and mcnichols (2002), and matsumoto (2002). 3 in a similar vein, walker and claasen (2006) find that analyst affiliation has no effect on stock price response to ratings changes. 4 their investment ratings are derived from value line timeliness ratings: 1, 2 � buy, 3 � hold, 4, 5 � sell. from their table 3 panel b, the mean unscaled forecast error for buys -$0.003 is not significantly different from zero, and the mean unscaled errors for holds and sells are not significantly different from buys or from each other. 5 if stock price and p/e are increasing in investment rating, then results in their table 4 panel a can be attributed to a stock price effect, not to earnings management. 6 eq. (3) is also estimated via quarterly lad regressions, and the significance of the coefficients is assessed from the time series of the quarterly estimates. results reported herein are robust to this alternative estimation procedure. 7 see their eq. (8), p. 111. 8 results reported in table 7 are robust to inclusion of extreme observations. references abarbanell, j., & bernard, v. (1992). tests of analysts’ overreaction/underreaction to earnings information as an explanation for anomalous stock price behavior. the journal of finance, 47, 1181–1207. abarbanell, j., & lehavy, r. (2003a). can stock recommendations predict earnings management and analysts’ earnings forecast errors? journal of accounting research, 41, 1–31. abarbanell, j., & lehavy, r. (2003b). biased forecasts or biased earnings? the role of reported earnings in explaining apparent bias and over/underreaction in analysts’ earnings forecasts. journal of accounting and economics, 36, 105–146. agrawal, a., & chen, m. (2006). analysts conflicts and research quality. american law and economics association annual meetings. ali, a., klein, a., & rosenfeld, j. (1992). analysts’ use of information about permanent and transitory earnings components in forecasting annual eps. the accounting review, 67, 183–198. bartov, e., givoly, d., & hayn, c. (2002). the rewards to meeting or beating earnings expectations. journal of accounting and economics, 33, 173. basu, s. & markov, s. (2004). loss function assumptions in rational expectations tests on financial analysts’ earnings forecasts. journal of accounting & economics, 38, 171. bradshaw, m. t., richardson, s. a., & sloan, r. g. (2006). the relation between corporate financing activities, analysts’ forecasts and stock returns. journal of accounting and economics, 42, 53–85. brous, p. a. (1992). common stock offerings and earnings expectations: a test of the release of unfavorable information. the journal of finance, 47, 1517–1536. brous, p. a., & kini, o. (1993). a reexamination of analysts’ earnings forecasts for takeover targets. journal of financial economics, 33, 201–225. brown, l. (2001). a temporal analysis of earnings surprises: profits versus losses. journal of accounting research, 39, 221–241. burgstahler, d., & dichev, i. (1997). earnings management to avoid earnings decreases and losses. journal of accounting and economics, 24, 99. 52 p. baird / financial services review 26 (2017) 37–54 butler, k. c., & lang, l. h. p. (1991). the forecast accuracy of individual analysts: evidence of systematic optimism and pessimism. journal of accounting research, 29, 150. chen, k., ying, z., zhang, h., & zhao, l. (2008). analysis of least absolute deviation. biometrika, 95, 107. cohen, d., & lys, t. (2003). a note on analysts’ earnings forecast errors distribution. journal of accounting and economics, 36, 147. cotter, j., tuna, i., & wysocki, p. (2006). expectations management and beatable targets: how do analysts react to explicit earnings guidance? contemporary accounting research, 23, 593. cowen, a., groysberg, b., & healy, p. (2006). which types of analyst firms are more optimistic? journal of accounting and economics, 41, 119. das, s., levine, c. b., & sivaramakrishnan, k. (1998). earnings predictability and bias in analysts’ earnings forecasts. the accounting review, 73, 277. degeorge, f., jayendu, p., & zeckhauser, r. (1999). earnings management to exceed thresholds. the journal of business, 72, 1. dreman, d. n., & berry, m. a. (1995). analyst forecasting errors and their implications for security analysis. financial analysts journal, 51, 30–41. dreman, d. n., & berry, m. a. (1995). overreaction, underreaction, and the low-p/e effect. financial analysts journal, 51, 21. dugar, a., & nathan, s. (1995). the effect of investment banking relationships on financial analysts’ earnings forecasts and investment recommendations. contemporary accounting research, 12, 131. easterwood, j. c., & nutt, s. r. (1999). inefficiency in analysts’ earnings forecasts: systematic misreaction or systematic optimism? the journal of finance, 54, 1777–1797. francis, j., & philbrick, d. (1993). analysts’ decisions as products of a multi-task environment. journal of accounting research, 31, 216. fried, d., & givoly, d. (1982). financial analysts’ forecasts of earnings: a better surrogate for market expectations. journal of accounting & economics, 4, 85. kang, s., o’brien, j., & sivaramakrishnan, k. (1994). analysts’ interim earnings forecasts: evidence on the forecasting process. journal of accounting research, 32, 103–112. kasznik, r., & mcnichols, m. (2002). does meeting earnings expectations matter? evidence from analyst forecast revisions and share prices. journal of accounting research, 40, 727. lin, h., & mcnichols, m. (1998). underwriting relationships, analysts’ earnings forecasts and investment recommendations. journal of accounting and economics, 25, 101. ljungqvist, a., marston, f., starks, l., wei, k., & hong, y. (2007). conflicts of interest in sell-side research and the moderating role of institutional investors. journal of financial economics, 85, 420. matsumoto, d. (2002). management’s incentives to avoid negative earnings surprises. the accounting review, 77, 483. michaely, r., & womack, k. (1999). conflict of interest and the credibility of underwriter analyst recommendations. the review of financial studies, 12, 653. o’brien, p. (1988). analysts’ forecasts as earnings expectations. journal of accounting & economics, 10, 1988. philbrick, d. r., & ricks, w. e. (1991). using value line and ibes analyst forecasts in accounting research. journal of accounting research, 29, 397. portnoy, s., & koenker, r. (1997). the gaussian hare and the laplacian tortoise: computability of squared-error versus absolute-error estimators, with discussion. statistical science, 12, 279. raedy, j., shane, p., & yang, y. (2006). horizon-dependent underreaction in financial analysts’ earnings forecasts. contemporary accounting research, 23, 291. ramnath, s., rock, s., & shane, p. (2005). value line and i/b/e/s earnings forecasts. international journal of forecasting, 21, 185. richardson, s., teoh, s., & wysocki, p. (2004). the walk-down to beatable analyst forecasts: the role of equity issuance and insider trading incentives. contemporary accounting research, 21, 885. shane, p., & brous, p. (2001). investor and (value line) analyst underreaction to information about future earnings: the corrective role of non-earnings-surprise information. journal of accounting research, 39, 387. 53p. baird / financial services review 26 (2017) 37–54 skinner, d., & sloan, r. (2002). earnings surprises, growth expectations, and stock returns or don’t let an earnings torpedo sink your portfolio. review of accounting studies, 7, 289–312. szakmary, a., conover, c., & lancaster, c. (2008). an examination of value line’s long-term projections. journal of banking & finance, 32, 820. walker, m., & claasen, b. (2006). what drives sell-side recommendation announcement returns? financial services review, 15, 315. 54 p. baird / financial services review 26 (2017) 37–54 pii: 1057-0810(92)90010-a volume 2 number 1 199u1993 financial services review the journal of individual financial management editor lewis mandell university of connecticut managing editor barbara poole university of connecticut associate editors neil g. cohen george washington university mona j. gardner illinois wesleyan university benton e. gup university of alabama jean louis heck villanova university david s. kidwell university of minnesota neil b. murphy virginia commonwealth university phyllis s. myers virginia commonwealth university george c. philippatos university of tennessee chris j. prestopino california state university chico s. travis pritchett university of south carolina william reichenstein baylor university frank k. reilly university of notre dame arthur l. schwartz university of south florida peter l. struck washington mutual savings bank g. c. uselton texas a & m university thomas waschauer san diego state university walt woerheide rochester institute of technology greenwich, connecticut jai press inc. london, england pii: 1057-0810(91)90028-w financial services review, l(2): 109-129 copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. effective credit costs in retail financial markets: leasing versus borrowing d. anthony plath bennie h. nunnally, jr. this study examines reported credit cost information in the automobile sales market to deter mine if vehicle leasing really is cheaper than installment borrowing. in addition, the study evaluates the accuracy of credit cost data furnished to consumers by commercial banks, vehicle leasingjrms, and automobile dealers to gauge whether any systematic differences exist in the accuracy of reported credit cost information. results of the study suggest that the cost of leasing is significantly different from borrowing, yet neither financing alternative is unilaterally cheaper than the other. in addition, suppliers of credit in consumerjnance markets routinely and signif?cantly understate effective credit costs reported to consumers. introduction vehicle leasing has become an increasingly popular financing alternative for new car buyers, reaching almost $42 billion in mid1991. according to recent federal reserve (1991) data, retail automobile lease contracts accelerated at an average annual rate equal to 19.2 percent between 1985 and 1990, while traditional automobile financing contracts advanced just 13 percent each year during the same period. vehicle lease contracts, which represented 44 percent of total automobile sales supported by retail credit arrangements in 1985, captured nearly 5 1 percent of the new automobile credit market in 1990. in addition, industry analysts estimate that 1.3 million new vehicles will be leased in 1991, representing a 20 percent growth in leased vehicles from 1990 and a 100 percent growth in leased vehicles since 1985 (koretz, 1990). in years preceding the tax reform act of 1986, the tax deductibility of lease payments was often used to explain the growth in vehicle leasing. under current tax regulations, however, most consumers classified by the internal revenue service as employees are not permitted a tax deduction for personally leased vehicles, and the d. anthony plath and bennie h. nunnally, jr. l college of business administration, university of north carolina at charlotte, charlotte, nc 28223. 110 financial services review, l(2) 1991 continued growth in leasing must be explained in other ways. for example, lessors frequently promote the benefits of leasing by arguing that (1) lease financing often requires lower monthly payments than an installment purchase, (2) leasing permits consumers to acquire more costly vehicles for a given monthly payment, and (3) leasing simplifies the disposal of used vehicles at the maturity of the lease. while these benefits might provide the impression that leasing represents the lower cost financing alternative, this conclusion is not necessarily supported by principles of financial management. unfortunately, consumers who remain unac customed to the application of these principles are not able to make accurate lease versus borrow comparisons. federal legislation designed to inform consumers about the effective costs of different financing options actually impedes the lease borrow comparison. the truth-in-lending act (1969) requires installment lenders to express the effective cost of borrowing in a standardized way (the annual percentage rate of interest), so that consumers can accurately compare the cost of alternative borrow ing contracts. in contrast, the consumer leasing act (1976) does not require lessors to report effective leasing costs in a manner comparable to the apr. therefore, a simple and direct comparison of leasing versus borrowing costs becomes impos sible. the purpose of this study is to examine reported credit cost information in the automobile market and determine the accuracy of pricing information furnished to consumers by a varied array of credit suppliers. the study examines the behavior of three credit suppliers-commercial banks, vehicle leasing firms, and automobile dealers-and surveys credit costs in different urban markets to gauge whether any systematic differences in the accuracy of reported credit cost information exist. second, the study investigates credit cost differences between leasing and borrow ing, and draws from the extant finance literature to explain observed credit cost differences. finally, the study provides a simple analytical framework for evaluating the effective cost of leasing that is directly comparable to the effective cost of borrowing. 1. credit costs and consumer leasing the typical consumer vehicle lease contract represents an operating lease, or a financing arrangement which is cancelable at the option of the lessee and is not fully amortized over the term of the lease. in addition, most consumer contracts are closed-end, characterized by lease agreements specifying a fixed number of payments for a finite period of time. when the leased asset is returned undamaged and vehicle mileage is within the limits established in the lease contract, the lessee has no further financial obligation to the lessor. as such, leasing contracts transfer some of the ownership risks associated with debt financing to the creditor. effective credit costs in retail financial markets: leasing versus borrowing 111 i, 1. the cost of leasing the extant finance literature offers a variety of different frameworks to evalu ate the lease versus borrow question. following the work of myers, dill, and bautista (mdb hereafter, 1976) and weingartner (1987)) these valuation models begin by comparing the value of two different financing transactions: (1) leasing, which involves purchasing the necessary cash to acquire an asset by giving up the asset’s depreciation tax shields, salvage value, and investment tax credit, and agreeing to make a fixed series of cash payments to the lessor; and (2) purchasing, which involves the acquisition of cash by selling an optimal package of financing securities exclusive of the lease contract. mathematically, the valuation of financial leases from the lessee’s perspective takes the general form v, = itc + c l,(l t)[l + k(1 t)]-j + j=o ” c td,[l + k(l t)]-j + f[l + k(1 t)]p /=i where itc = the investment tax credit available to the lessee at time zero; lj = the lease payment in periodj; t = the lessee’s marginal tax rate; k = the before-tax cost of debt to the lessee; dj = the depreciation expense displaced by the lessee in periodj; and f = the after-tax residual value of the leased asset, which occurs in period n. the general lease valuation model is remarkably versatile. brick, fung, and subrahmanyam (1987) use it to explain pricing differences between manufacturer lessors and third-party lessors, copeland and weston (1982) use it to value cancel able operating leases, and franks and hodges (1978) use it to value corporate lease contracts. unfortunately, the general model shown in equation (1) is less helpful in evaluating consumer financing transactions, because installment borrowing costs are frequently expressed in percentage terms. following the work of beechy (1969 and 1970), roenfeldt and osteryoung (1973), and doenges (1974), however, the mdb lease valuation framework can be easily modified to provide the annual percentage cost of lease financing. in this case, the quantity v, in equation (1) is set equal to the value of the leased asset acquired by the lessee, reduced by the after-tax 112 financial services review, l(2) 1991 cash costs required at the inception of the lease, and the equation is solved in terms of the unit-period percentage cost of leasing (k). given monthly lease payments, k can be transformed exponentially to yield the annual percentage cost of leasing (apl): apl = (1 + k)‘* 1 (2) under current tax regulations, which prohibit the deduction of vehicle lease payments from taxable income for most consumers, t is set equal to zero in equation (1). the absence of an investment tax credit in the current tax code also requires that itc equal zero in equation (1). sorenson and johnson (1977) and crawford, harper, and mcconnell (198 1) offer similar percentage-cost lease valuation frameworks to evaluate the cost of corporate lease agreements. in general, these studies find that the cost of leasing corporate assets far exceeds concomitant borrowing costs. crawford, harper, and mcconnell establish the before-tax yield on leases as 22.72 percent in 1975, while the average yield on bbb-rated corporate debt in the same period was 10.61 percent. they offer three possible explanations for this difference: (1) there is a greater probability of default among firms which use lease financing, (2) the market for leased assets is relatively inefficient; and (3) lease contracts are functionally differ ent from debt contracts in ways that are not well understood by financial managers. this empirical research provides only indirect evidence regarding the cost of consumer leasing, because it deals with financial lease contracts negotiated between corporate borrowers and lenders. while such transactions might be considered arms-length agreements between informed buyers and informed sellers of credit claims, anderson and martin (1977) show that many large corporate lessees inaccu rately value lease contracts. while the degree of informational asymmetry between buyers and sellers of consumer credit may be even more extreme, the finance literature offers little evidence concerning the valuation of retail lease contracts. given the recent growth in consumer leasing, it is important to consider an accessible, straightforward methodology allowing consumers to compare directly the costs of borrowing versus leasing durable goods. at present, consumer decisions are frequently governed by differences in out-of-pocket costs. some consumers believe, and some creditors promote, that leasing is the optimal means of asset acquisition simply because it results in lower monthly payments. this logic runs counter to the principles of financial theory, and more important, it encourages inefficiencies in the market for consumer credit. i. 2. the cost of borrowing according to regulation z of the federal reserve (truth-in-lending regula tions, 1984), annual percentage rate computations for closed-end borrowing trans effective credit costs in retail financial markets: leasing versus borrowing 113 actions take the form of the familiar internal rate of return calculation. in particular, the unit-period cost of credit, k, is determined by solving v, = 5 cf,/(l + k)’ t-l iteratively for k. v, represents the principal balance advanced to the borrower, and cf, represents the contractual loan payment required in period t. according to federal regulations, this unit-period cost is transformed to the annual percentage rate of interest by multiplying the unit-period rate by the number of periods in one year. while the federal reserve provides annual percentage rate tables to facili tate compliance with regulation z, lenders are free to use any computational tool in determining aprs, provided these tools conform to the mathematical framework outlined by the fed. as a general rule, lenders must report percentage financing costs within f 0.125 percent of the actual apr to comply with regulation z. 1.3. i%e lease versus borrow comparison simple comparison of the apr and apl data identifies the less costly financ ing alternative. advocates of the installment purchase option might argue, however, that credit costs alone do not adequately capture differences between leasing and borrowing alternatives. lessees must surrender leased assets at the expiration of the financing contract, while borrowers gain title to assets purchased over time. for this reason, nunnally and plath (1989) provide a second index to compare the wealth position of lessees and installment borrowers at the maturity of these respective financing contracts. this lease hurdle rate (lhr) provides the unit period rate of return necessary to transform the net cash savings from leasing into the residual value of the leased asset at the maturity of the lease. in the absence of tax-deductible lease payments, financing costs, and the depreciation and investment tax credit tax shields on consumer purchases, the unit-period lhr may be defined by solving r,, = c( 1 + k)’ + t+, m, (1 + k)’ for k. (4) in this expression, r, represents the net residual value of the leased asset at the expiration of the lease, c represents the difference between initial cash outlays under the lease and installment purchase options, and it4t represents the difference between monthly cash costs associated with leasing and debt financing, respec tively. the unit-period lhr is transformed into an effective annual rate through the use of equation (2). this annualized hurdle rate represents the after-tax return which 114 financial services review, l(2) 1991 lessees must earn to realize cash wealth equal to the leased asset’s terminal residual value. lessees unable to obtain this rate of return will possess smaller total wealth than installment purchasers at the maturity of the respective financing contracts. accordingly, the annual lhr defines the minimum rate of return that lessees must earn on the investment of residual cash flows from leasing to justify the selection of this financing alternative. the study used a random sampling procedure to gather financing cost data from providers of new vehicle credit within major metropolitan markets. credit suppliers included financial institutions, specialized leasing firms, and franchised automobile dealerships. restricting the research population to large urban markets insured a high degree of competition in the credit markets examined, and provided a straightforward method to determine the size and composition of the population. urban population centers included the 25 most populous metropolitan statisti cal areas within the u.s., and consumer credit vendors included firms that main tained 1989 telephone directory listings under the “banks”, “automobile leasing”, or “automobile dealers-new cars” categories in the yellow pages. lessors specializing in short-term vehicle rentals were excluded from the research population. the study used a disguised written questionnaire to obtain specific credit cost information. shown in the appendix, this questionnaire appeared as a request for lease and installment purchase credit information initiated by a typical consumer in the market for personal transportation. the professional identity of the researchers and the purpose of the inquiry were not revealed in the questionnaire, because these disclosures might severely bias the information reported by respondents. while each questionnaire requested information concerning 48-month installment borrowing and leasing contracts to standardize the data set, the vehicles described in individual questionnaires varied according to the manufacturer affiliation of differ ent auto dealerships. 2.1. the sampling plan the study used a two-stage, probability-proportionate-to-size sampling plan in order to generalize the credit cost data reported by respondents to the larger research population. in the first stage of sampling, five urban markets were selected at random from the overall research population. table 1 shows population data and market composition for each of these five markets. in the second stage of sampling, 300 individual firms were randomly chosen for inclusion in the sample. in keeping with the sampling methodology, the number of firms selected from each geographic market reflected the population density of that market, and the distribution of banks, lessors, and auto dealers within each effective credit costs in retail financial markets: leasing versus borrowing 115 table 1. description of sampling plan i. population characteristics market area 1987 population percent of (thousands) total sample size philadelphia 5,891 37% 110 miami 2,954 18 54 cleveland 2,767 17 52 atlanta 2,657 16 48 denver 1,861 12 36 ‘iwtal 16,130 100% 00 market area ii. sampling plan banks lessors dealers total no. pet. no. pet. no. pet. no. pet. a. market composition philadelphia 68 miami 229 cleveland 51 atlanta 130 denver 160 total 638 b. respondent composition philadelphia 3 miami 3 cleveland 2 atlanta 7 denver 4 jwal 19 response rate 19 10% 35 7 17 22 18% 7% 27 11 37 44 19% 37% 130 20% 35 5 228 33 187 25 158 22 738 21% 8 0 7 4 1 20 20 20% 0 37 21 11 20% 32% 459 70% 399 60 415 60 431 58 403 56 2,107 60% 657 19% 663 19 694 20 748 21 721 21 3,483 100% 30 73% 41 41% 8 73 11 11 10 53 19 19 8 42 19 19 4 44 9 9 - 60 61% 99 100% 60 32% 99 33% market paralleled the composition of that market in the research population. table 1 also reports useable responses across the five markets and three classes of credit suppliers. in most cases, response rates reflect the composition of the original research population, providing geographic and creditor diversity within the data set. in order to qualify as a useable response, the study required that respon dents report the annual percentage cost of both leasing and borrowing, as well as all dollar-cost items shown in equations (1) and (3). while the consumer leasing act does not require lessors to report the annual percentage cost of leasing to consumers, 88 percent of the respondents (99 of 113) provided this information in response to the original survey. 116 financial services review, l(2) 1991 2.2. statistical tests the study examined the accuracy of reported credit cost information in a variety of ways. first, the apr on installment borrowing transactions and the apl associated with vehicle leases were calculated from the reported credit cost data using the methodology described above. in calculating lease cost data, the values of ztc, dju = 1 to n), and t in equation (1) were set equal to zero in conformity with current tax regulations for individual taxpayers. values for f, representing the expected residual value of leased assets in period iz, were approximated by the purchase option prices offered to lessees at the conclusion of the lease. in calculating the annual percentage cost of leasing, equation (1) contains an implicit assumption that the riskiness of the expected residual value term is equal to the risk of contrac tual lease payments. while this clearly oversimplifies the lease valuation problem, the random selection of different vehicle manufacturers and various leased assets in the sample suggests that residual value estimation errors will contain no systematic bias. effective annual credit costs for both borrowing and leasing transactions were determined using the exponential transformation shown in equation (2). next, calculated annual credit costs were compared with reported credit costs to identify the direction and magnitude of individual credit cost reporting errors. finally, these reporting errors were evaluated using various statistical significance tests, including one-way analysis of variance (anova) procedures. given the unbalanced anova designs present in the study, heterogeneity of variance can severely bias the reported statistical results. bartlett’s test (neter, wasserman, and kutner, 1985) for the equality of group variances provides a means of testing this condition. test results from this procedure are shown in table 2. in table 2. homogeneity of variance hypothesis testing. bartlett’s test results dependent variable calc. loan cost calc. lease cost calc. loan cost calc. lease cost loan error lease error loan error lease error lease risk premia lease risk premia treatment market area market area creditor type creditor type market area market area creditor type creditor type market area creditor type anoh table 4-i-a 4-i-b 4-11-a 4-11-b s-ii-a 5-11-a 5-11-b 5-11-b 6-11-a 6-11-b chi-square values df calculated critical 4 16.7** 13.3 4 18.9** 13.3 2 10.4** 9.2 1 0.6 6.6 4 49.4** 13.3 3 4.0 11.3 2 36.2** 9.2 1 10.5** 6.6 4 2.3 13.3 1 0.9 6.6 noret ** indicates significance at least at the 1% level. effective credit costs in retail financial markets: leasing versus borrowing 117 cases where the homogeneity of variance condition is violated, welch’s (195 1) f” statistic replaces the conventional f statistic in the reported statistical results. keselman, games, and rogan (1979) demonstrate that welch’s procedure provides adequate control of type i errors in cases where unbalanced designs exhibit unequal variance structures. 3. results table 3 provides a summary of the reported credit costs furnished by respon dents, and the calculated credit costs and lease hurdle rates derived from these reported data. in general, reported loan costs exceed reported lease costs, and the variance within reported loan costs exceeds the variance in reported lease costs. in contrast, calculated lease costs frequently exceed calculated loan costs, while the variance within calculated credit costs is greater for lease transactions. these data suggest that larger credit cost reporting errors characterize vehicle lease agree ments . variance in the derived lease hurdle rates exceeds variance in both reported and calculated credit costs, indicating that in the specific lease and borrow alterna tives provided by respondents, it is impossible to declare either financing option as unilaterally optimal. in some cases, the lease hurdle rate is actually negative, which table 3. profiling consumer credit costs mean standard deviation minimum value maximum value i. composite sample reported loan cost calculated loan cost reported lease cost calculated lease cost lease hurdle rate 11.02% 1.76% 6.90% 12.95% 12.47 2.26 7.18 18.35 10.56 0.95 8.75 12.00 12.65 2.97 0.31 18.29 9.02 7.90 -23.45 105.86 ii. financing costs by market area philadelphia reported loan cost 10.99% calculated loan cost 12.70 reported lease cost 10.39 calculated lease cost 12.35 lease hurdle rate 7.52 miami reported loan cost 11.40% calculated loan cost 13.46 reported lease cost nc calculated lease cost nc lease hurdle rate nc 1.66% 6.90% 12.95% 2.57 7.18 17.58 0.76 9.00 11.50 3.19 0.31 16.58 8.18 -23.45 20.62 1.81 7.90% 12.50% 0.71 12.40 14.49 nc nc nc nc nc nc nc nc nc 118 financial services review, l(2) table 3. (continued) profiling consumer credit costs 1991 mean standard minimum maximum deviation value value cleveland reported loan cost calculated loan cost reported lease cost calculated lease cost lease hurdle rate atlanta reported loan cost calculated loan cost reported lease cost calculated lease cost lease hurdle rate denver reported loan cost calculated loan cost reported lease cost calculated lease cost lease hurdle rate 10.66% 11.99 10.47 12.00 8.14 10.84% 11.99 10.20 13.73 11.73 11.45% 12.38 11.33 13.80 16.23 iii. financing costs by credit source banks reported loan cost 11.96% calculated loan cost 12.70 reported lease cost nc calculated lease cost nc lease hurdle rate nc lessors reported loan cost 10.68% calculated loan cost 12.59 reported lease cost 10.15 calculated lease cost 12.23 lease hurdle rate 9.96 dealers reported loan cost 10.86% calculated loan cost 12.38 reported lease cost 10.68 calculated lease cost 12.77 lease hurdle rate 8.78 2.12% 6.90% 12.95% 2.22 8.11 15.28 0.98 9.00 12.00 3.21 5.30 18.29 73.20 8.59 105.86 2.09% 6.90% 12.90% 2.75 8.07 18.35 1.45 8.75 12.00 1.77 10.60 15.90 4.72 6.38 20.14 1.49% 6.90% 12.50% 1.22 8.79 13.71 0.52 11.00 12.00 0.76 12.90 14.57 3.10 10.99 18.84 0.52% 0.63 nc nc nc 11.00% 11.47 nc nc nc 12.50% 13.44 nc nc nc 2.17% 6.90% 12.95% 2.72 8.20 18.35 1.00 8.75 11.25 2.63 6.20 15.90 5.97 1.17 20.62 1.81% 6.90% 12.90% 2.42 7.18 17.58 0.94 9.00 12.00 3.08 0.31 18.29 8.35 -23.45 21.95 note: nc-insufficient responses for calculation. effective cm&t costs in retain finances markets: leasing verssus borrowing 119 table 4. calculated loan and lease credit costs, analysis of variance i. calculated credit costs by market area a. loan data: omnibus test (welch’s f”) sv 4 ss treatment 4 15.2 error 79 409.4 total s3 424.6 b. lease data: omnibus test (welch’s f”) sv 4 ss treatment 4 32.0 error 69 614.1 total 13 646.1 ms 3.8 5.2 ms 8.0 8.9 f” p>f 3.0 > 0.05 f” p>f 3.8 > 0.10 ii. calculated credit costs by creditor type a. loan data: omnibus test (welch’s f”) sv df ss treatment 2 1.4 error 81 423.2 total ‘is? 424.6 b. lease data: omnibus test (traditions f) sv df ss treatment 1 3.8 ms 0.7 5.2 ms 3.8 f” p=-f’ 0.1 > 0.25 f” p>f 0.4 > 0.51 72 642.3 8.9 total 73 646.1 implies that the undiscounted periodic cash savings provided by leasing generates cumulative cash flows in excess of the vehicle’s residual value at the maturity of the ~nancing contracts. in other cases, the lhr exceeds 100 percent, indicating that the small difference between periodic lease and installment purchase costs makes it virtually impossible for consumers to invest these cash flows to provide terminal wealth equal to the vehicle’s residual value. while these data may be dismissed as erroneously reported information, they are not treated as statistical outliers and removed from the data set. rather, reported credit costs are shown in unadulterated form to illustrate the nature of credit information routinely provided to consumers. the results contain substantial vari ance, and they may provide grossly inaccurate and misleading info~ation. table 4 provides summary anova results examining whether calculated credit costs differ according to market area or creditor type. these categorical variables are not significant in explaining variance in calculated credit costs. given the depth, breadth, and communications linkages within the market for consumer credit, this market can be described as national in scope. regional credit cost differences explain very little of the variance in either calculated lease or calculated borrowing costs, 120 financial services review, l(2) 1991 similar results emerge from anova testing of the creditor type variable. competition within the metropolitan credit markets examined would suggest that credit costs remain stable across different suppliers of credit. the data support this hypothesis, as calculated credit costs do not vary significantly across banks, lessors, or auto dealers. 3.1. credit cost reporting errors credit cost reporting errors, representing the difference between calculated and reported credit costs, are significantly different from zero for both lease and table 5. analysis of credit cost reporting errors (reporting error = calculated credit cost reported credit cost) i. tests of significance (unequal population variance) absolute value of proportion of sample where reporting error reported cost > reported cost < mean std. error calculated cost calculated cost a. aggregate sample 1. loan data 1.39%** 2. lease data 3.22 ** b. errors by market area 1. loan data phila. 1.71%** miami 2.16 * cleveland 1.43 atlanta 0.74 ** denver 0.71 2. lease data phila. 3.54%qq miami nc cleveland 2.78 * atlanta 3.56 ** denver 2.09 ** c. errors by creditor type i. loan data banks 0.83%** lessors 1.35 dealers 1.55 2. lease data banks nc lessors 1.80%** dealers 3.62 ** 0.19% 0.33 0.31% 0% 0.75 0 0.71 9 0.08 0 0.12 0 0.59% 25% nc 0.76 0 0.52 0 0.38 0 0.07% 0% 0.71 9 0.23 0 nc 0.17% 0.39 1% 9 14% 86% 8 92 99% 91 100% 100 91 100 100 75% 100 100 100 100% 91 100 notes: nc insufficient responses for calculation. ** indicates significance at least at the i % level. * indicates significance at least at the 5 % level. effective credit costs in retail financial markets: leasing versus borrowing 121 installment borrowing results. table 5 reports the significance test statistics for these data. the absolute value of reporting errors are used in statistical tests to evaluate the magnitude and dispersion of these errors independent of their direction. table 5 also reports directional characteristics of reporting errors. in virtually all cases, calculated credit costs exceed reported costs. in other words, creditors systematically understate the effective cost of consumer credit. on average, loan reporting errors are smaller and more narrowly concentrated about their mean than lease errors. this may occur because the truth-in-lending act establishes explicit guidelines for creditors reporting borrowing cost data to consumers, while similar standards do not apply to lease transactions. the report ing errors shown for loan transactions occur because the truth-in-lending act permits creditors to use an arithmetic transformation of unit-period credit costs provided in equation (3) to obtain annualized credit costs. in keeping with financial theory, this study generates calculated credit costs using the exponential transforma tion of unit-period returns shown in equation (2). commercial banks understate effective borrowing costs by an average 0.83 percent. the significance of this error represents a statistical artifact, which is ii. analysis of variance a. reporting errors by market area 1. loan data: omnibus test (welch’s f”) sv df ss treatment 4 17.8 error 71 194.8 total 75 212.7 2. lease data: omnibus test (conventional f) sv df ss ms 4.5 2.7 ms f” 4.8 f p>f > 0.05 p>f treatment 4 41.6 10.4 1.4 > 0.27 error 27 203.2 7.5 total 31 244.8 b. reporting errors by creditor type (omnibus test (welch s f”) 1. loan data: omnibus test (welch’s f”) sv df ss ms treatment 2 5.8 2.9 error 73 206.8 2.8 total 75 212.7 2. lease data: omnibus test (welch’s f”) sv df ss ms treatment 1 12.3 12.3 error 30 232.4 7.5 total 31 244.8 f” 1.3 f” 2.6 p>f > 0.25 p>f > 0.10 122 financial services review, l(2) 1991 explained by the different compounding methods used to produce calculated and reported borrowing costs in the study. the modest variance associated with report ing errors for banks suggests that these financial institutions use a consistent methodology to calculate effective borrowing costs. while this methodology conforms to federal regulations, it modestly understates effective borrowing costs. in addition, reporting errors provided by lessors and auto dealers are not statistically significant, due to the relatively large variance in these data. while the truth-in-lending act applies to all lenders who regularly extend consumer credit, lessors and auto dealers do not appear to follow a common methodology in deter mining and reporting effective credit costs. the statistical significance associated with credit cost reporting errors is not attributable to market area or creditor type. table 5 shows that anova tests performed on the data do not yield significant results. credit cost reporting errors occur systematically throughout the consumer credit industry; they are not isolated within a particular geographic area or unique to a particular type of credit supplier. 3.2. explaining credit cost differences: leasing i/et-sus borrowing the finance literature demonstrates that long-term lease contracts are similar to secured debt in many respects, but as smith and wakeman (1985) and ang and peterson (1984) illustrate, these financing alternatives are clearly different from one another. in spite of their differences, mdb explain that both leasing and borrowing represent a means for lessees/borrowers to acquire the necessary cash to support asset acquisition plans. as such, miller and upton (1976) point out that in efficient and competitive capital markets, the financial costs of leasing and borrowing must be equal. informed lessees/borrowers seek out the lowest-cost source of credit, while competition among different credit providers drives economic rents toward zero. table 6 reveals, however, that under the current tax treatment of personal borrowing and leasing transactions, the effective annual percentage costs of these alternative financing arrangements are not the same. in addition, the effective cost of leasing is neither unilaterally higher, nor lower, than the corresponding borrow ing cost. in roughly one-half of the sample cases, borrowing costs exceeded leasing costs, while the remaining cases exhibit the opposite cost preference. this result counters many previous empirical studies in the finance literature examining leasing costs. for example, sorensen and johnson (1977), mcgugan and caves (1974), gudikunst and roberts (1978), and crawford, harper and mcconnell (198 1) estimate the internal rates of return associated with a variety of different corporate financial lease contracts. in general, these studies report that lease yields unilaterally exceed equivalent debt financing costs. in contrast to this evidence, the finance literature offers a number of reasons to explain why in some cases, the cost of leasing should exceed that of borrowing, while in other cases, the reverse is true. this literature is useful for interpreting the results shown in table 6. effective credit costs in retail financial markets: leasing versus borrowing 123 table 6. credit cost differences: leasing versus borrowing (difference = calculated lease cost calculated loan cost) i. tests of significance (unequal population variances) absolute value of credit cost difference mean std. error proportion of sample where difference > 0 difference < 0 a. aggregate sample 2.92%** 0.27% 52% 48% b. credit cost difference by market area philadelphia 2.72%** 0.41% 45% 55% miami 3.63 * 0.95 33 67 cleveland 3.58 ** 0.65 38 62 atlanta 3.25 ** 0.89 87 13 denver 2.08 ** 0.50 78 22 c. credit cost difference by creditor rvpe banks nc nc lessors 2.53%** 0.51% 36% 64% dealers 3.01 ** 0.32 56 44 ii. analysis of variance a. absolute value of credit cost difference by market area: omnibus test (conventional f) sv df ss ms f” p>f treatment 4 15.3 3.8 0.7 0.58 error 65 344.1 5.3 total 70 359.4 b. absolute value of credit cost difference by creditor type: omnibus test (conventional f) sv df ss ms f r>f treatment i 2.6 2.6 0.5 0.49 error 68 356.8 5.2 total 69 359.4 notes: nc insufficient responses for calculation. ** indicates significance at least at the 1% level. * indicates significance at least at the 5 % level. first, the credit cost estimation framework provided by equations (1) and (3) may not capture all of the relevant financial costs associated with leasing. the lease borrow comparisons in table 6 assume that equivalent risk premia apply to lessees and installment borrowers. this assumption may be invalid. if lessees, as a group, exhibit higher rates of expected delinquency or nonperformance, then the financial cost of leasing should exceed borrowing costs to compensate lessors for the addi tional default risk of leasing. this argument seems unlikely, however, because suppliers of credit generally offer the lease-borrow choice to consumers on the basis of similar credit standards. individuals classified as creditworthy for borrowing are also offered the lease 124 financial services review, l(2) 1991 alternative, and consumers who fail to qualify for installment credit are also rejected as lessees. while credit quality undoubtedly varies across different debtors, it seems unlikely that it differs across the entire class of lessees versus borrowers. a second financial cost affecting lease yields, the non-diversifiable residual value risk premium, is not considered in equation (1). miller and upton (1976), mcconnell and schallheim (1983), and schallheim, johnson, lease, and mccon nell (1987) demonstrate that financial lease yields are negatively related to the non diversifiable residual value risk of leased assets. assets exhibiting higher rates of systematic depreciation, measured by the time-series covariance between market returns and asset depreciation rates, are priced by lessors to provide higher returns. if this residual value risk differs substantially across the leased vehicles represented in table 6, then these results may represent equilibrium financing costs in a competi tive market. lease yields for vehicles containing relatively high residual value risk will exceed equivalent borrowing costs, while leases supporting vehicles that exhibit lower residual value risk will be priced below the corresponding borrowing cost. another explanation for differences between observed leasing and borrowing costs concerns the presence of imperfections in the market for consumer credit. these imperfections include differential tax rates between lessors and lessees, transactions costs, information costs, and contract monitoring costs. in some cases, particular imperfections imply that leasing costs will exceed borrowing costs, while in other cases a different set of imperfections offers reasons why leasing is less costly than installment borrowing. several imperfections suggest that lease yields will exceed borrowing costs. in cases where information is not costlessly available to all participants in consumer finance markets, lessors may be able to exploit lessees’ inability to determine the true cost of leasing. anderson and martin (1977) report that even among fortune 200 firms, many analytical and methodological errors occur in the valuation of financial leases. this problem is perhaps even more acute in consumer finance transactions, where many individuals unfamiliar with the principles of financial analysis make financing choices. residual value uncertainty provides another imperfection explaining why lease yields exceed borrowing costs. in cases where lessors and lessees do not share a similar view of a given asset’s residual value distribution, lessees may be willing to pay lessors a premium to avoid residual value uncertainty. given that lessors are more familiar with secondary markets for leased assets, paying this premium allows lessees to avoid the search, information, and transactions costs associated with the disposal of owned assets. monitoring costs can also explain the higher cost of leasing. smith and wakeman (1985) point out that an asset’s value is affected by its history of use and maintenance. since lessees do not acquire disposal rights in connection with lease transactions, they have less incentive to maintain the asset to maximize its resale value. if the lease contract cannot effectively bind the lessee to provide mainte effective credit costs in retail financial markets: leasing versus borrowing 125 nance, or if it is relatively expensive for lessors to detect asset abuse caused by undermaintenance, then lease yields will reflect a premium for the added monitor ing costs and adverse selection problems created by leasing. finally, differences in marginal tax rates between lessors and lessees can explain differences between leasing and borrowing yields. in cases where lessees are unable to use the tax shields provided by lease payments, but lessors must treat lease payments as taxable income, then the cost of leasing may exceed borrowing costs to compensate lessors for the disproportionate tax burden they bear. current tax regulations attenuate this cost, because most consumers are unable to use lease payments to shield income from taxes. on the other hand, differences in lessor and lessee tax rates provide one explanation why the cost of leasing can fall below borrowing costs. franks and hodges (1978) and miller and upton (1976) show that when lessees maintain lower marginal tax rates than lessors, depreciation tax shields are more valuable to lessors. in this circumstance, firms seeking to acquire depreciable assets can exchange depreciation tax shields for reduced financing costs through the use of leasing. in consumer markets, a similar exchange of tax shields for reduced financ ing costs might be expected, because most personal assets cannot be depreciated against taxable income. through leasing, consumers can pass these tax shields to lessors who can realize their value, and in return, lessors may offer reduced financ ing costs of consumers. finally, transaction cost advantages can explain why borrowing costs exceed leasing costs. in cases where lessors have a comparative advantage in the disposal of used assets, lewellen, long, and mcconnell (1976) note that lessors can promote the lease alternative in competitive markets by offering reduced financing costs of lessees. in consumer markets, this reduction in lease yields is made possible by lessors’ comparatively lower search, information, and transaction costs in provid ing centralized locations for second-hand asset sales. given the offsetting nature of various imperfections, and the absence of a single market imperfection which dominates the determination of lease yields, different lease-borrow comparisons produce significantly different results. this leads lewellen, long, and mcconnell to note that corporate lease-borrow compari sons must be evaluated on a case-by-case basis to determine the lower cost financing method. the same general conclusion applies to consumer credit transactions. identifi cation of the less costly retail financing method is difficult, however, because reported lease cost data tends to confuse, rather than clarify, true financing costs, and federal credit cost disclosure requirements are not structured to permit the direct comparison of leasing and borrowing costs. moreover, the need for case-by case credit cost comparison is not limited to particular geographic areas or specific credit suppliers. as table 6 indicates, the market area and creditor type variables are not significant in explaining the variance in calculated credit cost differences. while leasing and borrowing costs differ across a variety of geographic markets and credit 126 financial services review, l(2) 1991 suppliers, the cost of leasing is neither generally higher, nor generally lower, than associated borrowing costs in specific metropolitan areas. similarly, vehicle leasing firms do not generally provide superior leasing terms, and commercial banks do not necessarily provide borrowing rates that favor installment purchase. 4. conclusions this article surveys reported consumer credit costs associated with new vehi cle lease and installment borrowing transactions to measure the credit cost advan tages associated with lease financing, and assess the accuracy of credit cost data provided to consumers. in general, the study finds that major suppliers of credit in consumer markets-including banks, leasing firms, and auto dealers-consistently and significantly understate effective annual credit costs. the magnitude of these reporting errors is larger for leasing transactions, which may occur because federal consumer credit regulations do not require lessors to report the effective cost of leasing to consumers in a uniform manner. these credit cost reporting errors are not unique to specific geographic areas, and they are not confined to particular types of credit suppliers. the study also finds that neither leasing nor installment borrowing provides a unilaterally cheaper consumer financing alternative. while leasing and borrowing costs are significantly different from one another in every credit market examined, in approximately one-half of the sample cases, borrowing costs exceeded leasing costs, and the remaining cases exhibited the opposite cost preference. this result counters previous empirical research concerning the effective cost of leasing, which concludes that leasing is generally more expensive than borrowing. the results of the present study, however, are compatible with a wide body of basic research in the finance literature that explains why the cost of different lease agreements may fall above and below the associated cost of borrowing. finally, the study concludes that comparative analysis of consumer leasing and borrowing costs must be handled on a case-by-case basis to make appropriate financial choices. unfortunately, a variety of impediments in the market for consumer credit-including participants who are unskilled in the application of modern financial theory, inaccurate credit cost information reported by credit suppliers, and federal reporting standards which make direct comparisons between leasing and borrowing costs difficult-can lead to inappropriate financing deci sions. in addition, these impediments encourage allocational inefficiency in the market for consumer credit. acknowledgments: the authors gratefully acknowledge funding from the foundation of the university of north carolina at charlotte and from the state of north carolina in support of this research. effective credit costs in retail financial markets: leasing versus borrowing 127 appendix research questionnaire residential street address city, state zip date creditor name creditor address city, state zip sales manager: i will be relocating to the area in september, and i plan to acquire a new car upon my arrival there. in order to research local financing costs before my arrival, i would appreciate some information from your firm regarding: (1) the cost of a 4%month, closed-end lease with zero capital cost reduction provided at the beginning of the lease; and (2) the cost of a 48-month installment loan contract with a 20 percent down payment. i would like to acquire for personal use a new, 1989 sedan containing air condition ing, am/fm stereo cassette, and the custom appearance package. please provide a summary of the initial costs, lease payments, and final costs your firm would charge to lease this car. in addition, i would like to know the tot1 price of the car used to determine the lease payments, the estimated residual value of the vehicle at the termination of the lease, and the effective annual percentage cost of the lease contract. in order to compare leasing with installment purchase, i would also like to know the total cost necessary to purchase the vehicle, the monthly payments necessary to amortize an installment loan over a 4%month period, and the annual percentage rate of interest for this transaction. i am planning to make a 20 percent down payment on this purchse. i have enclosed a postage-paid envelope for your convenience. thank you for your time, and i look forward to discussing the details of this transaction with you in september. sincerely, d. anthony plath 128 financial services review, l(2) 1991 references anderson, p and j. martin. 1977. “lease vs. purchase decisions: a survey of current practice,” financial management, 6: 4 l-47. ang, j. and i? peterson. 1984. “the leasing puzzle,” journal offinance, 39: 10551065. beechy, t. 1970. “the cost of leasing: comment and correction,” the accounting review, 45: 769 773. beechy, t. 1969. “quasi-debt analysis of financial leases,” the accounting review, 44: 375-381. brick, i., w. fung, and m. subrahmanyam. 1987. “leasing and financial intermediation: comparative tax advantages,” financial management, 16: 55-59. consumer leasing act of 1976. 1984. 15 u.s.c.a. 1601. copeland, t. and j. weston. 1982. “a note on the evaluation of cancelable operating leases,” financial management, 11: 60-67. crawford, i?, c. harper, and j. mcconnell. 1981. “further evidence on the terms of financial leases, ” financial management, 10: 7-15. doenges, r. 1974. “the cost of leasing,” 7&e engineering economist, 17: 31-44. federal reserve system board of governors. 1991. “domestic finance companies: business credit outstanding,” federal reserve bulletin, 77: 10, a35. franks, j. and s. hodges. 1978. “valuation of financial lease contracts: a note,” journal of finance, 33: 657-669. gudikunst, a. and g. roberts. 1978. “equipment financial leasing practices and costs: comment,” financial management, 7: 79-81. keselman, h., p. games, and j. rogan. 1979. “protecting the overall rate of type i errors for pairwise comparisons with an omnibus test statistic,” psychological bulletin, 86: 884-888. koretz, g. 1991. “how the car-leasing explosion burns uncle sam,” business week, september 2, 16. lewellen, w., m. long, and j. mcconnell. 1976. “asset leasing in competitive capital markets,” journal of finance, 3 1: 787-798. mcconnell, j. and j. schallheim. 1983. “valuationofasset leasing contracts,” journaloffinancial economics, 12: 237-261. mcgugan, v. and r. caves. 1974. “integration and competition in the equipment leasing industry,” journal of business, 47: 382-396. miller, m. and c. upton. 1976. “leasing, buying, and the cost of capital services,” journal of finance, 31: 761-786. myers, s., d. dill, and a. bautista. 1976. “valuation of financial lease contracts,” journal of finance, 3 1: 799-820. neter, j., w. wasserman, and m. kutner. 1985. app(ied linear statistical models, second edition. homewood, il: irwin, 618-622. nunnally, b. and d. plath. 1989. “leasing versus borrowing: evaluating alternative forms of consumer credit,” journal of consumer affairs, 23: 383-392. roenfeldt, r. and j. osteryoung. 1973. “analysis of financial leases,” financial management, 2: 74-87. schallheim, j., r. johnson, r. lease, and j. mcconnell. 1987. “the determinants of yields on financial lease contracts,” journal of financial economics, 19: 45-68. smith, c., and l. wakeman. 1985. “determinants of corporate leasing policy,” journal offinance, 40: 895908. sorenson, i. and r. johnson. 1977. “equipment financial leasing practices and costs: an empirical study,” financial management, 6: 33-40. truth-in-lending act of 1969. 1987. 15 u.s.c.a. 1667a. truth-in-lending regulations, regulation z, app. j. 1984. 15 u.s.c.a. foll. 1700. effective credit costs in retail financial markets: leasing versus borrowing 129 weingartner, h. 1987. “leasing, asset lives, and uncertainty,” financial management, 16: 5-12. welch, r. 1951. “on the comparison of several mean values: an alternative approach,” biometrika, 38: 330-336. pii: 1057-0810(91)90025-t from the editor financial planning for the family or individual is far more difficult than financial planning for the corporation, in large part because the former lacks anything resembling a viable optimization criterion. those of us interested in individual financial management cannot seek simply to maximize shareholder wealth. this is partly because the “shareholders” of a family are simultaneously its time-constrained consumers and producers, partly because there is no legal or viable market for a family’s wealth, and partly because inter-family agency problems make it difficult to identify the ultimate shareholder. economists suggest that we broaden the optimization criterion to that of utility maximization. while theoretically elegant, this solution lacks practicality. ignoring the agency problem of how family members trade off their individual utilities in favor of maximizing the family’s utility, we are still faced with complex problems including the allocation of wealth to present consumption or future consumption through compatible investment vehicles. in our opening article, “personal financial planning and the allocation of disposable wealth,” by amy and robert puelz, the authors state that “in the personal financial planning literature, however, there is no decision model that fully integrates an individual’s subjective valuation of all objectives and constraints relevant to this time dependent wealth allocation problem. ” since conflicting goals cannot be simultaneously achieved, the authors propose the “satisficing” of multiple goals, a solution first proposed by herbert simon in 1955. they achieve this through two decision-making techniques, goal programming and the analytical hierarchy process. the second article addresses a more practical and immediate problem. small investors are susceptible to “get rich quick” investment schemes based upon alleged anomalies in otherwise reasonably efficient investment markets. one such anomaly suggests that it only pays to be in the market in january. in their article entitled “should the small investor be out of the market outside of january?,” steven mann and donald solberg conclude that evidence supporting the january anomaly is not strong and that the investor should be wary of the claim. in recent years, the proportion of consumers leasing cars has increased dramatically, extending the lease versus buy decision from the corporation to the v vi financial services review, l(2) 1991 household. some have attributed this new development to tra86 which phased out the tax deductibility of consumer automobile loans not secured by home equity. in their article entitled “effective credit costs in retail financial markets: leasing vs. borrowing,” d. anthony plath and bennie h. nunnally use empirical analysis to come up with two very significant results. first, neither form of financing is superior. in about half the cases, leasing is less expensive and in the other half, borrowing proves cheaper. second, the cost of financing is substantially understated by both lenders and leasers, particularly the latter who are not bound by the uniform reporting standards of the truth-in-lending act. the fourth article, “comparing mortgages with different payment frequencies,” by arefaine yohannes is a behavioral study of consumers designed to find out who uses the biweekly payment mortgage. the use of this option tends to reduce the term of a 30 year mortgage by about a third. not surprisingly, he finds through logistic regression that frequency of payday is a very important determinant with those who are paid weekly or biweekly the more likely to utilize this innovation. he also finds that utilization varies as well with level of education, the better educated more willing to avail themselves of savings offered by the biweekly plan. an interesting new approach to estate planning is taken by ronald crabb in his article “probabilistic estate planning. ” this article challenges the traditional approach to estate planning which uses life expectancy to calculate likely estate value and insurance needs. since few people die at the time of their actuarially-predicted demise, there is a range of estate values and strategies implied. using markowitz’s mean variance portfolio approach, estate planning comes down to a risk-return tradeoff which is determined by the risk preferences of the decision maker. our review article, “international investing for the individual,” by jeff madura and tom o’brien, examines the rapidly expanding world of inter national investing and asks how individual investors may best participate. three key issues are examined: the benefits of international diversification, the key problem of currency exposure, and the effective means of achieving international diversification. this issue of financial services review concludes with abstracts of articles on individual financial management, edited by phyllis myers. i am pleased to report that the flow of high quality submissions keeps increasing. in order to attract the best possible articles relating to individual financial management, we have established a sixty day review turnaround as our goal for articles received between september and the beginning of april. turnaround slows during the summer months as reviewers become somewhat more scarce. we are seeking review articles for our forthcoming issues. readers who are interested in reading or writing about a particular topic should contact me. -lewis mandell does the source of a cash flow affect spending versus saving? valrie chambersa,*, eugene blandb, marilyn spencerc astetson university, 421 n. woodland blvd., deland, fl 32723, usa btexas a&m university–corpus christi, 6300 ocean drive, corpus christi, tx 78412-5808, usa ctexas a&m university–corpus christi, 6300 ocean drive, corpus christi, tx 78412-5808, usa abstract this study examines whether people use different mental accounts for different types of hypothetical revenue windfalls rather than viewing them as fungible in their use consistent with neoclassical economics. this study finds that the income source sometimes influenced the amount spent/saved and a respondent’s general default as a spender or saver was highly significant in all regressions. this article adds to the literature by responding to epley and gneezy’s (2007) call for “a broader sample of participants, varying amounts of payment, and alternative frames” to identify moderators of windfall framing effects with implications for behavioral economic theory and financial planning. © 2017 academy of financial services. all rights reserved. jel classification: d140 keywords: income; income source; decision making; consumer behavior; mental accounting 1. introduction a significant volume of research has established that people use mental accounts for budgeting their savings and expenditures. there is ample literature that suggests that income may be segregated into different mental accounts according to its source (e.g., thaler, 1999). there is very little literature on whether income from a specific mental account flows directly to savings, or alternatively whether income from a specific source is comingled with income * corresponding author. tel.: �1-386-822-7999; fax: �1-386-822-7426. e-mail address: valrie.chambers@stetson.edu financial services review 26 (2017) 291–313 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. from other sources and then the combined amount is allocated into mental accounts for savings and expenditures. this study is important because if there is a significant difference in how much windfall income people save based on the source of that income alone, then people may be making savings decisions earlier in the process than many of us assume. that is, if people save more of an inheritance than of a bonus, they are arguably not just saving a fixed percentage of income or saving from excess funds, they may be attaching a kind of purpose to the source itself, in which case the mental accounts used for budgeting expenditures does not originate with the collection of fungible income, but rather sometimes earlier, at the creation of the distinct sources of the income themselves. by understanding the process of allocating money to savings better, and specifically, the point in the process at which they make that allocation, individuals can better understand why they are (not) meeting their savings goals. as we better understand how people collectively allocate money to savings, societies can better predict the effects of different types of income structures on societal goals, like stimulating an economy during a recession. if income source causes significant differences, we can further study questions like these: from a macroeconomic standpoint, what effect on amount saved does an ordinary income pay structure have versus a lower ordinary income plus a bonus structure that keeps total compensation equal with the alternative? as the baby boomer generation dies off, what effect will additional inheritances have on spending versus saving? what role might “fun” play in people’s spending patterns? is greater personal control in earnings a factor in savings choices, such as when one receives an earned tax refund from the withholding that she established versus a bonus, which is also earned but more dependent on employer discretion? however, first, we need to know whether this path is worth exploring further, and to do that, we must explore whether any differences in saving because of the source of the income appear to exist. this is the purpose and importance of the study. 2. literature review 2.1. overview according to mental accounting theory, people create different mental expenditure accounts (e.g., long-term savings), and have different marginal propensities to consume from each account. numerous studies support mental accounting from a regular income flow or from an irregular, lump-sum windfall (johnson et al., 2006; o’curry, 1999; souleles, 2002). informally, people periodically reconcile their mental accounts for income and expense (camerer et al., 1997; heath & soll, 1996; read et al., 1999; rizzo & zeckhauser, 2003). karlsson et al. (1999) reported that cash spending on a durable good depended on compatible reasons for saving. generally, math aptitude affects mental budgeting (abeler & marklein, 2008; benjamin, 2006). cheema and soman (2006) and wertenbroch (2003) concluded that mental budgeting is a matter of self-control. sprenger and stavins (2010) found that one way of exercising control over spending on revolving credit cards is to substitute debit card use. 292 v. chambers et al. / financial services review 26 (2017) 291–313 frederick (2005) reported a negative relationship between non-rational behavior and cognitive reflection. baker and nofsinger (2002) found that, when inheriting a hypothetical portfolio, people tend to hold the investments that they inherited, rather than spend them or change the risk tolerance of these investments to match their own risk tolerance. milkman and beshears (2009) found that consumers who receive $10 windfalls in the form of grocery coupons spend an additional $1.59 on groceries that the consumer does not typically buy, perhaps buying something she thought that she otherwise could not afford. 2.2. source literature some evidence suggests that the source of one’s income does affect the use of those funds. henderson and peterson (1992) reported that individuals were more likely to spend $2,000 on a vacation if the source of the funds was a gift rather than a work bonus. arkes et al. (1994) found that a greater percentage of a small windfalls was spent than that from the same amount of anticipated income, indicating that foreknowledge of income is a factor in saving, consistent with rucker (1984) and karlsson et al. (1999). dobbelsteen and kooreman (1997) found that individuals were more sensitive to changes in a child’s allowance than to other income sources for the child when deciding how much to spend on their child’s clothing. winkelmann et al. (2011) used evidence from german lottery winners to show that spending from different sources of income conferred different marginal utilities. trump et al. (2015) found that individuals would make riskier choices with a stranger’s money than with a friend’s money. similarly, bradford (2008) found that individuals allocate gifted and inherited assets in support of relational goals. levav and mcgraw (2009) proposed that tracking expenditures from windfalls in mental accounts can be complemented by examining one’s “affective tag,” which is how one feels about a sum of money. for example, the affective tag one places on the money associated with negative feelings may influence them to consume the windfall either reluctantly or virtuously to cope with those negative feelings. levav and mcgraw further postulated that income is spent in a way that matches that source; money won on a bet may be spent frivolously, but money back from the irs would be used to pay for something of greater importance or lasting value. still, the framing of payments seems to matter: baker et al. (2007) found that more money was spent from likely recurring income (dividends) than less regular capital gain income. epley et al. (2006) found that people spent more from an income source labeled “bonus” than they did of a “rebate” of the same amount and timing. similarly, shefrin and thaler (1988) found that more of a lump sum bonus is saved than if the same amount increases regular income, even when the bonus is fully anticipated. meekin et al. (2015) found that nearly four in 10 recipients of a tax refund initially save it, with about one in five taxpayers receiving the earned income tax credit for working families. this refund is large for low-income families, and sometimes exceeds the amount of federal income taxes withheld, but families view it as “earned,” just as the name of the credit states. those families anticipate it and intend to save about 17% of it “to get ahead.” where the refund is spent on everyday expenses, it tends to be spent on bills that are past due. tax refunds saved are frequently used to pay down debt that cannot be paid down through ordinary income. meekin et al. (2015) further estimated that 21% of the refund is devoted to 293v. chambers et al. / financial services review 26 (2017) 291–313 expenditures like education, home repairs, purchasing or repairing a car, or buying a durable good like a freezer that will produce savings on expenses over future periods. 2.3. earning whether income is perceived as earned may affect responses. boylan (2010) found that compliance with the tax system is influenced by whether taxable income is earned or endowed. epley and gneezy (2007) reported that a windfall that positively deviates from the status quo, like a bonus, is more likely spent than a windfall that restores the status quo. zagorsky (2013), studied consumption of inherited money and found that over 40% of those who inherited less than $1,000 spent their bequest. only 18.7% of those receiving $100,000 or more spent it all. in all, this research indicates that only about one half of inherited money was retained, the remainder was reduced by capital losses or is spent. agarwal and qian (2013) studied how consumers responded to an exogenous income shock, and found that consumption rose significantly at the rate of $0.80 per $1 received. spending began with the announcement of the income shock. low-liquidity consumers and low-credit consumers consumed more. 2.4. frequency of distribution neoclassical economics assumes that the decision to spend, and how to spend one’s income would not depend on the way in which it is received. yet, the difference in spending patterns from a limited number of monthly payments and a lump-sum tax rebate of the same amount is well documented. rucker (1984) studied the retroactive payment of a raise approved by a university, reversed by the federal pay board but reinstated by the u.s. supreme court. the size of the windfall was found to be the most important factor for deciding how the funds were spent, with smaller checks more likely to be consumed. in addition, the length of time that the recipient had to anticipate the receipt of the funds also influenced the use of the money. the shorter the time before the receipt of the money was anticipated, the more likely that the money was consumed. karlsson et al. (1999) noted that individuals considered the future consequences of spending in their mental budgeting, which may indicate a contemplation of permanent income. shapiro and slemrod (1995) found that almost half of the respondents surveyed would spend the 1992 decreased tax withholding refunded to them, even though the total yearly tax liability remained unchanged, resulting in a lower end-of year tax refund. however, in 2001, when a tax cut took the form of either a $300 or $600 lump-sum rebate, only about one-fourth of those surveyed expected to spend the payment (shapiro & slemrod, 2003). slemrod and bakija (2004) attributed the change in behavior of taxpayers between the differently distributed rebates to changes in economic conditions. however, applying thaler’s (1999) mental accounting theory, chambers and spencer (2008) found that the timing of payments (whether paid as a lump-sum, or spread out in equal monthly installments for a year) mattered. sahm et al. (2012), confirmed this finding. 294 v. chambers et al. / financial services review 26 (2017) 291–313 2.5. permanence of distribution neoclassical economics tells us that neither the marginal cost nor the marginal benefit of a purchase is dependent on the source of the income from which it is spent. however, the permanence of payments may be a factor in how much people choose to save. blinder (1981) posited that a permanent tax decrease would elicit more spending than a temporary tax rebate, which he surmised would be treated as one half from a normal income tax change and the other half from a windfall. parker (1999) studied tax cuts, finding that a temporary, end-of-year reduction in social security tax for high-income wage earners was spent when received, not averaged evenly over the fiscal year. friedman’s (1957) permanent income hypothesis says that people will spend money consistent with what they believe to be their permanent income level, but stopped short of examining the source of the income or testing the spending on amounts of limited duration. studies of unique, one-time payments are rare. however, bodkin (1959) estimated the marginal propensity to consume to be between 0.72 and 0.97 of a one-time dividend paid in 1950 to world war ii veterans by the national service life insurance. the payments averaged $175, roughly $1,723.39 in 2015 dollars (bureau of labor statistics, 2016). similarly, kreinin (1961) analyzed the spending of a sample of israeli citizens receiving restitution payments from germany in 1957 and 1958 and estimated that 35% was spent while 65% of the restitution payment was saved, with 45% saved in liquid assets and 20% in real estate (kreinen, 1961). 2.6. materiality of amount chambers et al. (2009) studied responses to small hypothetical tax rebates of the size distributed in 2008, $300 and $600, as well as larger amounts, $1,500 and $3,000. they found that at some amount over $600, materiality mattered greatly in how the money would be used. under the $600 amount, individuals were likely to spend a rebate if that was the government’s intent for distributing it, but at or above $600, the government’s wishes were ignored (chambers et al., 2009). research on large, regular bonuses includes hsieh (2003) who studied household consumption associated with receipt of the annual alaska permanent fund payment, which was fully anticipated; and no spike in consumption was found. however, consumption by the same households was very responsive to income tax refunds, suggesting that sizable, predictable, and regular payments are built into consumption decisions (hsieh, 2003). browning and collado (2001) studied spanish panel data to measure the effect of customary bonus payments. usually workers are paid 1/14th of their annual wage per month for the 10 months other than december and june or july, when they received bonuses of 2/14th of their salary. similar to hsieh, they did not find changes in consumption patterns (browning & collado, 2001). 2.7. demographic factors several demographic factors might affect the savings intent as well. chen and volpe (2002) sampled multiple colleges and universities and found that women generally knew less 295v. chambers et al. / financial services review 26 (2017) 291–313 about personal finance topics but that education and experience can have a significant impact on the financial literacy of both men and women. spencer and chambers (2012) studied the lump sum tax rebates of 2008, relative to the 2009 tax rebates distributed in small amounts in take-home pay. their findings indicated that the small periodic distributions accomplished the stimulus significantly better. many taxpayers were not aware of receiving the 2009 rebate, and those who did realize it, saved a lower average percentage than with the 2008 rebate. income, wealth, high-risk tolerance, and savings differed significantly by gender, as did being non-white and having other household members (fisher et al., 2015). specifically, income uncertainty was associated with a significantly lower likelihood of saving for men. fisher’s (2010) findings indicated that black–white differences in savings were explained by the individual determinants of saving—like receiving government assistance, feeling that credit use is bad, being turned down for credit in the past 5 years, or having a shorter saving horizon. black households also had a lower risk tolerance and were more likely to save for a bequest. 3. hypothesis this study tests whether people spend a distribution from a hypothetical tax rebate as they would if the distribution came from other windfall sources, such as a bonus from work, a game show winning, an inheritance or a lottery winning. specifically, the goal was to test whether there is any difference in the savings rate among different sources of income. regular income was omitted, relying on literature that amounts saved from regular income differed from that saved from windfalls (arkes et al., 1994; karlsson et al., 1999; rucker, 1984). some sources of windfall were included, as in earlier literature, among them: inheritance (e.g., baker & nofsinger, 2002), bonus (e.g., henderson & peterson, 1992; hsieh, 2003), tax rebate (e.g., chambers & spencer, 2008; hsieh, 2003) and lottery (winkelmann et al., 2011). these factors are expected to vary in their affective tags (levav & mcgraw, 2009): perceived deservedness (earnings, not endowment), and outcomes that are susceptible to changes in self-control (cheema & soman, 2006; sprenger & stavins, 2010; wertenbroch, 2003). for example, an inheritance is likely to have a strong affective tag, is less earned and less susceptible to one’s own self-control. a bonus has a more moderate affective tag, is presumably earned, but is only moderately susceptible to one’s own self-control versus the employer’s control. a tax rebate should have a lesser affective tag, is presumably earned, but less susceptible to one’s own self-control than that of the irs. lottery winnings have a low affective tag, are unearned and very unsusceptible to one’s own self-control. this study introduces a new, exploratory variable, game show winnings, which has a moderate affective tag, is moderately earned and moderately in one’s self-control. based on earlier literature, a cash inheritance would be expected to stay primarily in the form of cash savings (baker & nofsinger, 2002). a bonus would be expected to be saved (henderson & peterson, 1992; hsieh, 2003). a tax rebate was assumed to be split between savings and spending (chambers & spencer, 2008; meekin et al., 2015). lottery winnings were anticipated to be spent (winkelmann et al., 2011). without relying on earlier literature 296 v. chambers et al. / financial services review 26 (2017) 291–313 for game show winnings, those earnings were expected to be spent because, whereas such winnings arguably result from some earnings effort and more self-control than a lottery, those winnings are closer in source to a lottery than to the other sources tested here. although a lottery winner is in control of the decision to buy another ticket, the decision to be a gameshow contestant does not solely rest with the player. it is unlikely a gameshow winning could be repeated, and that makes it different from a lottery. how might the recipient consider some of these sources as similar and others as different? lottery winnings are similar to tax rebates in the united states, in that both lottery systems and tax systems are run by a government or its appointed agency. both types of payment amounts are largely outside the respondent’s control. to what extent the money is “earned” is debatable in both cases, but bonuses and game show winnings—and sometimes inheritances—require some personal skill, knowledge, and effort. tax rebates sometimes differ from the other four sources of payment because the tax rebate is a refund or return of withholdings the taxpayer has previously paid in. that is, outside of refundable credits tied to specific performance, respondents generally cannot materially profit from a tax rebate because it is a refund of money already paid in, but can profit from a lottery, game show, or bonus. an inheritance is not a profit, per se, but is generally not a return of one’s own capital. inheritances might be property or money that carries with it memories of the decedent, and those emotions (affective tags; levav & mcgraw, 2009) might carry over to how the respondent intends to use the inheritance. further, some political rhetoric frames taxes as money belonging fundamentally to taxpayers, not the government, whereas lottery winnings come with no similar sense of entitlement. bonuses are likely to be closely tied to an individual’s performance, however. game show winnings might be as well, if the winner attributes success to having a higher skill level than fellow contestants. that is, difference in amount saved by source is to be expected, but no source is absolute and completely separate in characteristics from the other sources, biasing against finding any differences. with that in mind, the null hypothesis is: h1: there will be no difference in savings rates by source of windfall. in testing this hypothesis, the amount of the income was controlled for, as were the order of presentation, the frequency of payments and the demographic characteristics of the respondents. 4. methodology sheppard et al.’s (1988) meta-analysis of 86 theory-of-reasoned-action studies found a 0.53 correlation between intention and behavior, indicating that intent is a good predictor of action. for this study, on each survey instrument, two of the five different sources of windfall were presented, resulting in 10 unique pairings. these 10 pairings were tested at four different gain amounts found in chambers and spencer (2008). these 40 instruments were then tested with the total amount paid in a lump sum, as well as paid out in 12 monthly installments; but to control for order effect, half showed the monthly installments first and the other showed the lump sum first, resulting in 80 different instruments. each participant 297v. chambers et al. / financial services review 26 (2017) 291–313 was given one of these 80 instruments at random and asked how the funds would be used, both if received as a lump-sum and if the same amount were received over 12 equal monthly payments (within-subject design), from two of these five sources: bonus, game show winnings, inheritances, lottery winnings, and tax rebates (between-subjects design). each instrument hypothesized one of these four different amounts: $300, $600, $1,500, and $3,000. see appendix for a sample survey instrument. the instruments asked how much of a lump sum refund would be used for: (1) investing, (2) paying off credit card debt, (3) paying off notes, (4) regular monthly expenses, (5) buying a durable asset, (6) saving for an infrequent expense, and/or (7) used for fun. hershfield et al. (2015) found that consumers’ tendency to place savings and debt into separate mental accounts makes them insensitive to the significant differences between the interest rates on these accounts. the instrument also asked how much of a monthly payment (equal to 1/12 of the lump sum amount) would be used for each of these seven purposes, consistent with chambers and spencer (2008). similarly, the opposite side of each instrument asked these same questions, changing only the source of the payment from one source to another—such as from a tax rebate to a lottery, work bonus, inheritance, or game show payment. experimental questionnaires were distributed to university students at these universities: coastal carolina university, francis marion university, longwood university, metropolitan state university of denver, texas a & m university-corpus christi, university of alabamabirmingham, and university of houston-clear lake. students were considered provisionally acceptable respondents per walters-york and curatola (1998) and ashton and kramer (1980). all research questions were analyzed with descriptive statistics, converted to percentages, and then then analyzed using four sets of ols regressions. the choices were (1) investing, (2) paying off credit card debt, (3) paying off notes, and (6) saving for an infrequent expense were coded as savings, and choices (4) regular monthly expenses, (5) buying a durable asset, and (7) used for fun were coded as spending. two of the sets of regressions, one where the monthly distribution was shown first and the other where the lump sum was shown first, used “longer-term savings” as its dependent variable, which excluded savings from item (6), saving for an infrequent expense. the other two sets of regressions (one where the monthly distribution was shown first) used dependent variables dubbed “total savings” included in item (6). the four dependent variables were each regressed against the source of the windfall income, and demographic variables were included to control for income, gender, age, importance to the budget, business experience level, and education level. the regression models were of the form: percent saved � f(income, zero income, amount, education, gender, age, importance, seatbelt use, smoker, spend1 (default for spender), experience level, dummy variables for the source of the payment (lottery, tax rebate, inheritance, game show, or bonus), and a dummy for the order of presentation (monthly payment first, or lump sum payment first)). “income” is the log of the respondent’s income plus one. “amount” is the hypothetical amount of the distribution, in dollars. as four discrete values were possible for the amount, dummy variables were created for each amount rather than treat this variable as continuous. the education variable is divided into four categories: high school, associate degree, undergraduate degree, and graduate degree. “gender” is a categorical male/female variable, 298 v. chambers et al. / financial services review 26 (2017) 291–313 where female was coded as “1.” “age” is the participant’s age in years. as there may be some non-linearity in the age variable, the square of age, “agesq” was added to the model to measure the non-linear contribution to the dependent variable that occurs as the reported age increases. “importance” was defined to be the payment divided by the income of the survey participant. the “seatbelt” and “smoker” dummy variables were included as proxies for respondents’ risk preference; seatbelt wearers and smokers were coded as “1.” for the variable “spend1” the participants were asked “when you get ‘extra money,’ do you spend it or save it?” the dummy was set to 1 for those that answered “spend.” business experience was a categorical, self-reported measure coded as “0” for “none,” “1” for responses of “low,” and progressing upward to “5” for “high.” this categorical variable was then transformed to dummy variables for use in the regression as described below. various formulations of the credit card debt variable were also introduced to observe whether debt in dollars or as some proportion would affect the results; the results were not affected, however. 5. results the data were gathered in 2013. of the 1,844 responses, 984 had complete data for regression analysis. a separate analysis using 1,719 of the responses, which were missing some of the control variables but not data for the variables of interest, was also run. these results were substantially the same as the analysis from the more pristine, complete data presented below. table 1 presents the descriptive statistics for the variables collected. the average income was $47,628, which compares with an average $57,706 for 2010 from the irs statistics of income (internal revenue service, 2012). respondents averaged 5.15 years of work experience and had some college education (that is to be expected as the sample was collected primarily from college students); 54% of the respondents were women. these respondents perceived themselves to have moderate business experience, as indicated by a 2.80 average score out of a possible 5.0. further discussion of the following non-significant variable (at p � 0.10) is omitted for parsimony, except where noted: monthly pmt (order effect), gender, smoke, seatbelt, and all expl levels. to test the null hypothesis, does the source of the payment matter, four sets of regressions were run. in the first set, the dependent variable longer term savings (ltsavyr) for the lump sum payment was regressed. the results are presented in table 2. the percentage saved, long term, when given a single lump sum was found to be significant and positively related to the log of income (p � 0.018) and zero income (p � 0.0203). respondents who reported a zero income saved a greater percentage of the windfall than those that reported earning an income. the parameter estimate indicates that those reporting zero income saved roughly 23% more of the payment than those reporting an income. though the variable importance is significant only at the 10% level, the sign of the coefficient was negative, indicating that as the distribution as a percentage of respondent’s income decreased, more was saved as a percentage (p � 0.096). in other words, if the payment was significantly larger than the income, savings decreased. 299v. chambers et al. / financial services review 26 (2017) 291–313 where: income log of the respondent’s reported income (plus 1). zeroincome dummy variable � 1 for those reporting zero income. level(x) dummy variables representing the lump sum for the different amounts used in the survey. monthly pmt dummy variable equal to one if the survey began with the monthly payment or zero if the lump sum payment was presented first. hsed1 dummy variable representing the respondent indicating high school as their highest education level. ased2 dummy variable representing the respondent indicating an associate degree as their highest education level. baed3 dummy variable representing the respondent indicating an ba/bs degree as their highest education level. graded4 dummy variable representing the respondent indicating the graduate level as their highest education level. gender equal to one for female respondents. age the participant’s age in years. agesq the square of the age reported. importance the monthly payment given in the survey divided by the participant’s income. smoke dummy variable equal to one if the response was “yes” to the questions about smoking. seatbelt dummy variable equal to one if the response was “yes” to the questions about seatbelt use. spend1 dummy set to one for those that answered “spend” to the question, “when you get ‘extra money,’ do you spend it or save it?” table 1 descriptive statistics for sample participants variable n mean 25th percentile median 75th percentile maximum standard deviation income 1,350 47627.84 6000.00 20000.00 55000.00 300000.00 122264.64 zeroincome 1,844 0.1035792 0 0 0 1.0000000 0.3047965 level300 1,844 0.2559653 0 0 1.0000000 1.0000000 0.4365208 level600 1,844 0.2478308 0 0 0 1.0000000 0.4318702 level1500 1,844 0.2180043 0 0 0 1.0000000 0.4130024 level3000 1,844 0.2781996 0 0 1.0000000 1.0000000 0.4482338 monthlypmtmt 1,844 0.4826464 0 0 1.0000000 1.0000000 0.4998343 hsed1 1,791 0.3417085 0 0 1.0000000 1.0000000 0.4744149 ased2 1,791 0.1072027 0 0 0 1.0000000 0.3094572 baed3 1,791 0.4684534 0 0 1.0000000 1.0000000 0.4991432 graded4 1,791 0.0826354 0 0 0 1.0000000 0.2754072 gender 1,812 0.5413907 0 1.0000000 1.0000000 2.0000000 0.4995280 age 1,447 23.0352453 20.0000000 21.0000000 24.0000000 80.0000000 6.6002851 agesq 1,447 574.1561852 400.0000000 441.0000000 576.0000000 6400.00 468.1131076 importance 1,350 200.1580834 0.0149999 0.0499983 0.2998501 3000.00 633.8898301 smoke 1,798 0.1173526 0 0 0 1.0000000 0.3219295 seatbelt 1,782 0.9595960 1.0000000 1.0000000 1.0000000 1.0000000 0.1969602 spend1 1,713 0.3607706 0 0 1.0000000 1.0000000 0.4803643 expl1 1,794 0.0562988 0 0 0 1.0000000 0.2305620 expl2 1,794 0.1984392 0 0 0 1.0000000 0.3989359 expl3 1,794 0.4665552 0 0 1.0000000 1.0000000 0.4990193 expl4 1,794 0.1755853 0 0 0 1.0000000 0.3805730 expl5 1,794 0.0490524 0 0 0 1.0000000 0.2160377 footerbonus 1,844 0.1827549 0 0 0 1.0000000 0.3865703 footerlotto 1,844 0.2152928 0 0 0 1.0000000 0.4111368 footertax 1,844 0.1881779 0 0 0 1.0000000 0.3909602 footerinherit 1,844 0.1979393 0 0 0 1.0000000 0.3985542 footergamg 1,844 0.2158351 0 0 0 1.0000000 0.4115120 ltsavmo 1,774 0.4127889 0 0.4000000 0.6600000 1.0000000 0.3670234 ltsavyr 1,766 0.4701184 0.2000000 0.5000000 0.6666667 1.0000000 0.3286521 totasavmoo 1,774 0.5575938 0.2800000 0.6000000 1.0000000 1.0000000 0.3616980 totalsavyr 1,766 0.6414166 0.5000000 0.6666667 0.9100000 1.0000000 0.2943912 300 v. chambers et al. / financial services review 26 (2017) 291–313 the level of the payment was also positively related to savings. those receiving $600 saved 7.5% more of the payment than those receiving $300 (p � 0.012), whereas those receiving $3,000 (10 times more), saved about 7.4% more of that payment than those who hypothetically received the $300 payment (p � 0.018). those receiving the $1,500 payment saved about 5.7% more of the payment than those receiving the $300 payment, but that result is only significant at the 10% level (p � 0.077). respondents were asked to report their “highest education level: high school ___ associate degree ___ undergraduate ___ graduate or above ___.” dummy variables were expl variable for the answer to the respondent’s evaluation of her business experience. the values range from 0 (none) to 5 (high). these responses were transformed to these dummy variables: expl1 group respondents reporting “low” and “none.” expl2 report “fairly low,” expl3 report “moderate,” expl4 report “fairly high” and expl5 report “high.” footer a dummy variable of interest, representing the source of the payment, where bonus � bonus, inherit � inheritance, lotto � lottery, game � game show, tax � tax rebate. ltsavmo the percentage of the sum that is saved in when there is a monthly payment for one year. ltsavyr the percentage of the sum that is saved in when the payment is a single lump sum. totalsavmo the percentage of the sum that is saved when there are monthly payments for one year. totalsavyr the percentage of the sum that is saved when the payment is a single lump sum. table 2 longer term savings is the dependent variable, for lump sum payment analysis of variance source df sum of squares mean square f value pr � f model 24 9.12929 0.38039 3.52 �.0001 error 959 103.75996 0.10820 corrected total 983 112.88925 root mse 0.32893 r2 0.0809 dependent mean 0.46931 adj. r2 0.0579 coefficient variable 70.08904 parameter estimates variable df parameter estimate standard error t value pr � �t� intercept 1 �0.09641 0.15288 �0.63 0.5284 lnincome 1 0.02008 0.00846 2.37 0.0177 zeroincome 1 0.22799 0.09807 2.32 0.0203 level600 1 0.07490 0.02971 2.52 0.0119 level1500 1 0.05671 0.03206 1.77 0.0773 level3000 1 0.07354 0.03114 2.36 0.0184 ased2 1 0.07236 0.03711 1.95 0.0515 baed3 1 0.06408 0.02481 2.58 0.0100 graded4 1 0.07688 0.04608 1.67 0.0955 age 1 0.01532 0.00754 2.03 0.0425 agesq 1 �0.00018961 0.00010129 �1.87 0.0615 importance 1 �0.00004912 0.00002956 �1.66 0.0969 spend1 1 �0.09609 0.02232 �4.30 �.0001 full model was run, but only (marginally) significant variables are shown for parsimony. 301v. chambers et al. / financial services review 26 (2017) 291–313 created to determine if education level influenced the level of savings. those that answered “undergraduate” saved 6.4% more than those that reported “high school” (p � 0.01). those that reported “associate degree” or “graduate or above” also saved more, 7.2% and 7.7%, respectively, but those amounts are significant only at the 10% level (p � 0.051 and p � 0.096, respectively). that is, higher levels of education were associated with higher levels of savings. older respondents did save more of the payment. the coefficient for the variable age was 0.015, indicating that for each year older the respondent was, the savings was higher by 1.5% (p � 0.043). the variable spend1, indicating whether a respondent’s default behavior pattern was to spend any extra money received, was economically and statistically significant. those that answered “spend” saved almost 10% less than those that answered “save” (p � 0.0001). table 2a are lottery, tax, inheritance, or game show sources saved as much as bonus? (omitted variable bonus) df parameter estimate standard error t value pr � �t� footerlotto 1 0.02365 0.03455 0.68 0.4939 footertax 1 0.06604 0.03450 1.91 0.0559 footerinherit 1 0.02145 0.03341 0.64 0.5211 footergame 1 �0.04372 0.03343 �1.31 0.1912 table 2b are game show, tax, inheritance, or bonus sources saved as much as lottery? (omitted variable lottery) df parameter estimate standard error t value pr � �t� footergame 1 �0.06737 0.03380 �1.99 0.0466 footertax 1 0.04240 0.03487 1.22 0.2244 footerinherit 1 �0.00220 0.03367 �0.07 0.9479 footerbonus 1 �0.02365 0.03455 �0.68 0.4939 table 2c are lottery, game show, inheritance, or bonus sources saved as much as tax? (omitted variable tax) df parameter estimate standard error t value pr � �t� footerlotto 1 �0.04240 0.03487 �1.22 0.2244 footergame 1 �0.10977 0.03360 �3.27 0.0011 footerinherit 1 �0.04460 0.03374 �1.32 0.1866 footerbonus 1 �0.06604 0.03450 �1.91 0.0559 table 2d are lottery, tax, game show, or bonus sources saved as much as inheritance? (omitted variable inheritance) df parameter estimate standard error t value pr � �t� footerlotto 1 0.00220 0.03367 0.07 0.9479 footertax 1 0.04460 0.03374 1.32 0.1866 footergame 1 �0.06517 0.03253 �2.00 0.0454 footerbonus 1 �0.02145 0.03341 �0.64 0.5211 302 v. chambers et al. / financial services review 26 (2017) 291–313 the results in table 2a indicated marginally significant, lower savings from bonus payments than from tax rebates. the results in table 2b indicated that savings from lottery payments were higher than savings from game show payments at the 5% significance level. the results in table 2c indicated that savings from tax rebate payments were higher than savings from game show payments at the 1% significance level, and from bonus payments at the 10% significance level. the results in table 2d indicated that savings from inheritance payments were higher than savings from game show payments at the 5% significance level. table 3 provides the results for the regression when the (totalsavyr) total savings from a lump sum payment is used as the dependent variable. this variable is composed of the longer term savings variable plus the amount saved “for infrequent expenses such as vacations, bigger holiday gifts, or something you’ve been wanting;” thus, representing the total amount devoted to savings. similar to the results in table 2, the variables income, zero income, education at the associate’s degree level, importance to the budget, and spend1 were all significant at the 5% level. age was marginally significant, at the 10% level (p � 0.070). the dummy variables from the different amounts of the payments were no longer significant. unlike for table 2, the percentage saved in table 3 is not statistically different for those receiving $300 than those getting the higher amounts. those reporting an undergraduate or graduate or higher level of education did not save a statistically significant amount more than those reporting a high school level of education when presented a single lump sum. respondents’ default behavior, whether they saw themselves as spenders or savers, again was highly significant. table 3 total savings is the dependent variable—payment is made as a lump sum analysis of variance source df sum of squares mean square f value pr � f model 24 7.28237 0.30343 3.43 �.0001 error 959 84.78328 0.08841 corrected total 983 92.06565 root mse 0.29733 r2 0.0791 dependent mean 0.63737 adj. r2 0.0561 coefficient variable 46.65046 parameter estimates variable df parameter estimate standard error t value pr � �t� intercept 1 0.21583 0.13820 1.56 0.1187 lnincome 1 0.02259 0.00764 2.96 0.0032 zeroincome 1 0.28155 0.08865 3.18 0.0015 ased2 1 0.07724 0.03355 2.30 0.0215 age 1 0.01235 0.00682 1.81 0.0704 importance 1 �0.00005784 0.00002672 �2.16 0.0306 spend1 1 �0.11056 0.02018 �5.48 �.0001 full model was run, but only (marginally) significant variables are shown for parsimony. 303v. chambers et al. / financial services review 26 (2017) 291–313 when comparing whether source of income was significant in windfalls distributed monthly for a year, there were no significant differences at the p � 0.05 level. for total savings with a lump sum payment, game show winnings were saved less than a tax rebate, but only at a marginally significant level of p � 0.0632. the analysis for tables 4 and 5 are similar to tables 2 and 3; however, the hypothetical total amount received was distributed over 12 equal monthly payments instead of as a lump sum. the total of the monthly payments equaled the lump sum payment. for instance, if the lump sum amount was a one-time payment of $600, the monthly payment was $50 for one year. in table 4, longer term savings with a monthly payment, the variable for the log of income was still positive and significant at the 5% level. as more income was earned, more of the amount received was saved. however, the dummy variable for those reporting a zero income was no longer significant, and neither was the variable for importance. as the monthly amount was only 1/12 of the amount received in the lump sum question (table 2), though, this was not surprising because the monthly amount was likely immaterial per chambers et al. (2009). consistent with the findings in table 2 (longer-term savings from a lump sum payment), more was saved as the payment becomes larger. those who received a total amount of $600 ($50 a month) saved 9.7% more of the payment than those who received $300 ($25 a month; p � 0.004). those who received $3,000 per year ($250 per month) saved 8.2% more of the payment than those receiving $300 ($25 a month; p � 0.017). similar to the results in table 2, those that received $1,500 ($125 a month) saved 6.1% than those receiving $300 ($25 a month), but this was significant only at the 10% significance level (p � 0.086). contrary to the results in table 2, none of the education variables were significant at even the 10% level. the age variables were also insignificant in table 4, though they were table 4 longer term savings is the dependent variable–payment made monthly analysis of variance source df sum of squares mean square f value pr � f model 24 9.88387 0.41183 3.06 �.0001 error 964 129.57707 0.13442 corrected total 988 139.46094 root mse 0.36663 r2 0.0709 dependent mean 0.40857 adj. r2 0.0477 coefficient variable 89.73351 parameter estimates variable df parameter estimate standard error t value pr � �t� intercept 1 0.00882 0.16955 0.05 0.9585 lnincome 1 0.01995 0.00932 2.14 0.0325 level600 1 0.09672 0.03315 2.92 0.0036 level1500 1 0.06139 0.03567 1.72 0.0856 level3000 1 0.08242 0.03433 2.40 0.0165 spend1 1 �0.08324 0.02490 �3.34 0.0009 monthlypmt 1 0.06055 0.02377 2.55 0.0110 full model was run, but only (marginally) significant variables are shown for parsimony. 304 v. chambers et al. / financial services review 26 (2017) 291–313 significant in table 2. again, the difference between the independent variables in tables 2 and 4 were that the payment amounts in table 4 were 1/12th the size of the payment amounts in table 2; the total payment amount was the same, but spread over the 12 months. the variable representing the respondents’ default behavior for spending or saving additional money continues to be highly significant. the results indicate that if a respondent said that she generally spends additional money, she did. when the respondent indicated that she was a spender, on average she saved 8.3% less (p � 0.001) than one who indicated that she was a saver. a new finding is exposed in tables 4 and 5. the order of presentation now matters. on the survey, half of the forms had the question for lump sum payment first, and the other half had the monthly payment first. when the lump sum values were used (tables 2 and 3), this variable was insignificant. when the smaller, but recurring, values are used in tables 4 and 5, those receiving the monthly payment saved (both economically and statistically) more. for longer term savings with monthly payments (table 4), when the monthly question was provided first, respondents saved 6.1% (p � 0.01) more than when the annual question was presented first. in table 5, representing total savings, 13.4% more was saved (p � 0.001) when the monthly amount was presented first than when the lump sum amount was presented first. for the dependent variable longer term savings with monthly payments, the percentage saved from game show winnings were significantly lower than those from a lottery at the p � 0.0355 significance level. the percentage saved from game show winnings were significantly lower than from a tax rebate at the p � 0.0046 significance level, and inheritance savings were significantly lower than a tax rebate at the p � 0.0664 significance level. the percentage saved from game show winnings were lower than that of a bonus, but table 5 total savings is the dependent variable–payment made monthly analysis of variance source df sum of squares mean square f value pr � f model 24 15.56019 0.64834 5.24 �.0001 error 964 119.30184 0.12376 corrected total 988 134.86203 root mse 0.35179 r2 0.1154 dependent mean 0.55978 adj. r2 0.0934 coefficient variance 62.84473 parameter estimates variable df parameter estimate standard error t value pr � �t� intercept 1 0.00213 0.16269 0.01 0.9895 lnincome 1 0.02589 0.00894 2.90 0.0039 zeroincome 1 0.17478 0.10470 1.67 0.0954 level600 1 0.10244 0.03181 3.22 0.0013 level1500 1 0.07867 0.03423 2.30 0.0218 level3000 1 0.08060 0.03294 2.45 0.0146 spend1 1 �0.09852 0.02389 �4.12 �.0001 monthlypmt 1 0.13449 0.02281 5.90 �.0001 full model was run, but only (marginally) significant variables are shown for parsimony. 305v. chambers et al. / financial services review 26 (2017) 291–313 only at the p � 0.0646 significance level. the results displayed in table 5 were similar to those above in many respects. income was still significant (p � 0.004) and positive. the coefficient for those reporting a zero income was positive as in all of the other regressions, but significant only at the 10% level (p � 0.095). (it was insignificant in table 4, but significant at the 5% level in tables 2 and 3.) as in table 4 but contrary to tables 2 and 3, the variable for importance (monthly payment/income) was insignificant at any conventional level. compared with the $300 payment ($25 per month), as a greater monthly payment was received, a greater percentage of the payment was saved. those receiving $600 ($50 per month) saved 10.2% more of the payments than those who received $300 ($25 per month; p � 0.001). those receiving $1,500 ($125 a month) saved 7.9% more of the payment than those receiving $25 a month, while those receiving $3,000 ($250 per month) saved 8.1% more of the payment (p � 0.015). the self-reported “spenders” saved 9.9% less than the “savers” (p � 0.001). as with table 4, the order variable was statistically and economically significant. when the questionnaire ordered the monthly payment first and the lump sum second, participants’ total savings increased 13.4% more than those where the lump sum payment questions were asked first (p � 0.001). the significant results from regression output similar to tables 2a through 2d follow and summarized again in table 6. for the dependent variable total savings with monthly table 6 summary of significance of the sources of payment a. longer-term savings, lump sum � pr � �t� at the 1% significance level game show vs. tax rebate �0.110 0.001 at the 5% significance level game show vs. lottery �0.067 0.047 game show vs. inheritance �0.065 0.045 at the 10% significance level bonus vs. tax rebate �0.066 0.056 b. total savings, lump sum payment � pr � �t� at the 10 % significance level game show vs. tax rebate �0.056 0.063 c. longer-term savings, monthly payment � pr � �t� at the 1% significance level game show vs. tax rebate �.1058 .005 at the 5% significance level game show vs. lottery �.079 .036 at the 10% significance level game show vs. bonus �.068 .065 inheritance vs. tax rebate �.069 .066 d. total savings, monthly payment � pr � �t� at the 1% significance level game show vs. lottery �.096 .008 game show vs. bonus �.108 .002 at the 5% significance level game show vs. tax rebates �.076 .033 inheritance vs. bonus �.077 .033 at the 10% significance level inheritance vs. lottery �.065 .072 306 v. chambers et al. / financial services review 26 (2017) 291–313 payments for one year, game show winnings were saved at a lower rate than bonus payments at the p � 0.0024 significance level, and an inheritance was saved at a lower level than a bonus at the 0.00325 significance level. the percentage saved from game show winnings were lower than savings from lottery payments at the p � 0.0077 significance level, and inheritance saving was lower than that from lottery payments, but only at the p � 0.0721 significance level. savings from game show winnings were less than savings from a tax rebate at the p � 0.0329 significance level. the percentage saved from a bonus was higher than inheritance savings at the p � 0.0325 significance level, and savings from a lottery winning was higher than that of from an inheritance, but only at the p � 0.0721 significance level. the results summarized in part a of table 6 shown an economically and statistically significant lower amount was saved from game show winnings than were saved from lottery winnings, inheritance and tax rebates. less of a bonus was saved than a tax rebate, but this was significant at slightly more than the 5% level. part b of table 6 provides weak support (p � 0.063) indicating that 5.6% less was saved from game show winnings than from tax rebates. part c of table 6 shows strong economic and statistically significant results showed that savings from game show winnings were less than those from lottery and tax rebates. weak statistical support indicated that savings from game show winnings were lower than that from bonus payments; and that savings from inheritances were lower than that from tax rebates. finally, part d of table 6 summarizes the results providing strong economic and statistically significant results showed that savings from game show winnings were less than those from lottery, bonus or tax rebates. in addition, less of inheritance payments were saved than from bonus payments. weak evidence suggested that less of inheritance payments were saved than from monthly lottery payments. the results summarized in table 6 negate the null hypothesis. for example, in at least one set of regressions, the results indicated that savings from game show winnings were significantly lower than from bonuses, inheritances, lottery winnings, or tax rebates. 6. discussion the results show multiple instances where windfalls, whether distributed as a lump sum or spread out in the form of monthly payments for a year affect longer-term savings and total savings, negating the null hypothesis. this is a clear exception to neoclassical economic theory but consistent with mental accounting theory and behavioral economics. for game show winnings and inheritance, a greater percentage was spent than was money from a bonus, tax rebate or lottery winning. as the amounts used in this study ranged from $300 to $3000, these results are consistent with the findings of zagorsky (2013), who found that over 40% of those who inherited less than $1,000 spent their entire bequests. because there are significant differences in how much windfall income people save based on the source of that income alone, the source of a windfall may be a factor in saving, not the amount of the windfall itself. in particular, game show winnings, which are arguably associated with “fun” 307v. chambers et al. / financial services review 26 (2017) 291–313 are much more likely to be spent—a hedonistic pursuit. further, there were significant findings that inheritances were treated differently than a bonus, tax rebate, or lottery winning. this finding seems to support and extend the finding that there is some non-rational, perhaps affective tag associated with some sources of income, as presented by levav and mcgraw (2009). with a regular income stream, people generally treat money as fungible when it comes in, but then use mental accounting “buckets” to determine where it will go out, combining elements of both neoclassical economic theory and behavioral economics. neoclassical economic theory and behavioral economics are seen as somewhat competing theories when in fact there may be a place for both. people could decide that money from one source will go to one set of bills, and money from another will go to savings or another set of bills. it appears that they do not. it appears that often, but not always, they mix their revenues together, and then decide how to allocate the mixed pool of money. that is, people tend to be neoclassical, but not rigidly neoclassical when making revenue decisions, and follow behavioral economics when making expenditure decisions. finding a balance between competing theories materially adds to current literature. that the answer to whether the source matters seems to be “sometimes,” and is itself important and an opportunity for more nuanced study. 7. limitations and further research more nuanced studies could include: which people are more likely to affectively tag income, what kind of affective tags are assigned, at what point (notice, receipt, sometime in between) are affective tags assigned, are affective tags limited to certain kinds of income, or also expenses, and what factors influence whether a tag is assigned at all? these studies are warranted, but they are extensive and left for further research. the effect of these tags are also left for further study. for example, from a macroeconomic standpoint, what effect on amount saved does an ordinary income pay structure have versus a lower ordinary income plus a bonus structure that keeps total compensation equal with the alternative? as the baby boomer generation dies off, what effect will additional inheritances have on spending versus savings? what role might “fun” play in people’s spending patterns? is greater personal control in earnings a factor in savings choices, such as when one receives an earned tax rebate from the withholding that they established versus a bonus that is also earned but more dependent on employer discretion? this study was conducted during a time when the economy was recovering from a shock that was severe enough to disrupt people’s normal spending or savings habits. in more stable economic times, results may differ. the recession of 2007–2008 was the worst since the great depression, and researchers found that spending and saving behavior changed dramatically, perhaps permanently (spencer & chambers, 2012). further study over many years is needed to measure the long-term effects. the different sources of the funds yielded some significant differences, raising more questions for further study. the order difference should also be investigated. when the annual payment was viewed first, spending was higher for the monthly sum. this finding may have the greatest practical implication for those in the financial planning arena. presenting clients with annual values 308 v. chambers et al. / financial services review 26 (2017) 291–313 (e.g., for retirement income) may entice the client to increase current spending compared with providing estimates for monthly income. 8. conclusion it is well-documented that people use mental accounts for budgeting their savings and expenditures; and ample literature that suggests that income may be segregated into different mental accounts according to its source (e.g., thaler, 1999). however, very little has been published on whether income from a specific mental account flows directly to saving, or instead whether income from a specific source is comingled with income from other sources and then the combined amount is allocated in to mental accounts for savings and expenditures. the results of this study indicate that sometimes income appears to be affectively tagged and flows directly to savings; other times it appears to be comingled with income from other sources. while this is left for further study, that income that has a highly emotional or affective tag, like inheritances that can be associated with love and grief and game shows that can be associated with fun tend to be tagged nearly instantly, whereas other sources of windfalls associated with less affect tend to be comingled and spent or saved more rationally. with a better understanding the process of allocating money to savings, and specifically, the point in the process at which individuals make that allocation, they can better understand why they are (not) meeting their savings goals. agarwal and qian’s (2013) study found that spending began as soon as the individual learned of the windfall, but they did not study all of these types of unexpected income. as we better understand how people collectively allocate money to savings, societies can better predict the effects of different types of income structures on societal goals, like how inheritances are likely to be spent or saved with the passing of the baby boom generation. that is the purpose and importance of this study. appendix: sample survey instrument “what would you do if . . .?” (fill in the amounts): by participating in a game show, you won a prize that would result in you receiving $600.00 for 2012. if received, how much of these winnings would you plan to: 1. invest (in stocks, bonds, savings account, and so forth)? $ 2. use to pay off credit card debt? $ 3. use to pay off notes (such as mortgage, car note, and so forth)? $ 4. use up about evenly every month for expenses? ______/month. � 12 months. � $ 5. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 6. use to save for an infrequent expense (such as a vacation, bigger holiday gifts, or something you’ve been wanting)? $ 7. spend right away on something fun? $ amount must total $600.003 if instead, by participating in a game show, you won a prize that would result in you receiving $50.00/month for the next 12 months. 309v. chambers et al. / financial services review 26 (2017) 291–313 if received, how much of this monthly increase would you plan to: 8. invest (in stocks, bonds, savings account, and so forth)? $ 9. use to pay off credit card debt? $ 10. use to pay off notes (such as mortgage, car note, and so forth)? $ 11. use up for regular monthly expenses? $ 12. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 13. use to save for an infrequent yearly expense (such as a vacation, bigger holiday gifts, and/or something you’ve been wanting)? $ 14. spend right away on something fun? $ amount must total $50.003 please list your: zip code_______________ years of work experience _______ highest education level: high school ___ associate degree ___ undergraduate ___ graduate or above ___ occupation: __________________ gender: female ___ male___ age ____ race/ethnicity ___________________ number of college-level accounting classes completed ___ college major (if applicable) __________________ industry where you work ______________________________ approx. yearly household income (from all wage and salary earners and other sources of income) $__________________ credit card debt: $_______________ other debt: $_______________ do you smoke? do you normally wear your seatbelt? yes ___ no ___ when you normally get “extra money,” do you spend it or save it? spend ___ save ___ i rate my level of business experience as: high ___ fairly high ___ moderate ___ fairly low ___ low___ none ___ complete other side, please. thank you for your participation!!! “what would you do if . . .?” (fill in the amounts): you got a bonus at work that would result in you receiving $600.00 which for 2012 will automatically be mailed to you as a check from your employer. if enacted, how much of this monthly increase would you plan to: 15. invest (in stocks, bonds, savings account, and so forth)? $ 16. use to pay off credit card debt? $ 17. use to pay off notes (such as mortgage, car note, and so forth)? $ 18. use up about evenly every month for expenses? ______/month. � 12 months. � $ 19. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 20. use to save for an infrequent expense (such as a vacation, bigger holiday gifts, or something you’ve been wanting)? $ 21. spend right away on something fun? $ amount must total $600.003 another work bonus would result in you receiving $50.00/month after taxes; that is, your paychecks would go up $50.00/month. if received, how much of this monthly increase would you plan to: 22. invest (in stocks, bonds, savings account, and so forth)? $ 23. use to pay off credit card debt? $ 24. use to pay off notes (such as mortgage, car note, and so forth)? $ 310 v. chambers et al. / financial services review 26 (2017) 291–313 references abeler, j., & marklein, f. (2008). fungibility, labels, and consumption. (available at http://papers.ssrn.com/ sol3/papers.cfm?abstract_id�1139870) agarwal, s., & qian, w. (2013). consumption and debt response to unanticipated income shocks: evidence from a natural experiment in singapore. unpublished working paper. national university of singapore. arkes, h. r., joyner, c. a., pezzo, m. v., nash, j. g., siegel-jacobs, k., & stone, e. (1994). the psychology of windfall gains. organizational behavior and human decision processes, 59, 331–347. ashton, r. h., & kramer, s. s. (1980). students as surrogates in behavior accounting research. journal of accounting research, 18, 1–15. baker, m., nagel, s., & wurgler, j. (2007). the effect of dividends on consumption. brookings papers on economic activity, 1, 231–276. baker, h. k., & nofsinger, j. r. (2002). psychological biases of investors. financial services review, 11, 97–116. (available at https://manowar.tamucc.edu/login?url�http://search.proquest.com/docview/212053247?accountid� 7084) benjamin, d. j. (2006). psychology, preferences, and markets. (doctoral dissertation, harvard university.) blinder, a. s. (1981). temporary income taxes and consumer spending. journal of political economy, 89, 26–53. browning, m., & collado, m. d. (2001). the response of expenditures to anticipated income changes: panel data estimates. the american economic review, 91, 681–692. bodkin, r. (1959). windfall income and consumption. the american economic review, 49, 602–614. boylan, s. (2010). prior audits and taxpayer compliance: experimental evidence on the effect of earned versus endowed income. journal of the american taxation association, 32, 73–88. bradford, t. w. (2008). intergenerationally gifted asset dispositions. journal of consumer research, 36, 93–111. bureau of labor statistics. (2016). databases, tables & calculators by subject – cpi inflation calculator. (available at http://bls.gov/data/inflation_calculator.htm) camerer, c., babcock, l., loewenstein, g., & thaler, r. (1997). labor supply of new york city cabdrivers: one day at a time. quarterly journal of economics, cxiii, 407–442. chambers, v., & spencer, m. (2008). does changing the timing of a yearly individual tax refund change the amount spent vs. saved? journal of economic psychology, 29, 856–862. chambers, v., spencer, m., & mollick, j. (2009). goldilocks rebates: complying with government wishes only when the rebate amount is “just right.” journal of economics and economic research, 10, 101–119. cheema, a., & soman, d. (2006). malleable mental accounting: the effect of flexibility on the justification of attractive spending and consumption decisions. journal of consumer psychology, 16, 33–44. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. (available at https://manowar.tamucc.edu/login?url�http://search. proquest.com/docview/212012102?accountid�7084) dobbelsteen, s., & kooreman, p. (1997). financial management, bargaining and efficiency within the household; an empirical analysis. de economist, 145, 345–366. epley, n., mak, d., & idson, l. (2006). bonus or rebate?: the impact of income framing on spending and saving. journal of behavioral decision making, 19, 213–227. 25. use up for regular monthly expenses? $ 26. use to buy a durable asset (such as car, boat, washing machine, furniture)? $ 27. use to save for an infrequent yearly expense (such as a vacation, bigger holiday gifts, and/or something you’ve been wanting)? $ 28. spend right away on something fun? $ amount must total $50.003 complete other side, please. thank you for your participation!!! 311v. chambers et al. / financial services review 26 (2017) 291–313 epley, n., & gneezy, a. (2007). the framing of financial windfalls and implications for public policy. the journal of socio-economics, 36, 36–47. fisher, p. j. (2010). black-white differences in saving behaviors. financial services review, 19, 1–16. fisher, p. j., hayhoe, c. r., & lown, j. m. (2015). gender differences in saving behaviors among lowto moderate income households. financial services review, 24, 1–13. frederick, s. (2005). cognitive reflection and decision making. the journal of economic perspectives, 19, 25–42. friedman, m. (1957). a theory of the consumption function. princeton, nj: national bureau of economic research. heath, c., & soll, j. b. (1996). mental accounting and consumer decisions. journal of consumer research, 23, 40–52. henderson, p. w., & peterson, r. a. (1992). mental accounting and categorization. organizational behavior and human decision processes, 51, 92–117. hershfield, h. e., sussman, a. b., rourke, l. o., & bryan, c. j. (2015). leveraging psychological insights to encourage the responsible use of consumer debt. perspectives on psychological science, 10, 749–752. hsieh, c. (2003). do consumers react to anticipated income changes? evidence from the alaska permanent fund. the american economic review, 93, 397–405. internal revenue service. (2012). (available at http://www.irs.gov/uac/soi-tax-stats-individual-income-taxreturns) johnson, d. s., parker, j. a., & souleles, n. s. (2006). household expenditure and the income tax rebates of 2001. american economic review, 96, 1589–1610. karlsson, n., gärling, t., & selart, m. (1999). explanations of effects of prior income changes on buying decisions. journal of economic psychology, 20, 449–463. kreinin, m. e. (1961). windfall income and consumption: additional evidences. the american economic review, 51, 388–390. levav, j., & mcgraw, p. (2009). emotional accounting: how feelings about money influence consumer choice. journal of marketing research, 46, 66–80. (available at http://search.proquest.com/abicomplete/docview/ 235223437/64e7aed8026742e2pq/1?accountid�7084) meekin, k. e., tach, l., & sykes, j. (2015). it’s not like i’m poor: how working families make ends meet in a post-welfare world. berkley, ca: university of california press. milkman, k. l., & beshears, j. (2009). mental accounting and small windfalls: evidence from an online grocer. journal of economic behavior & organization, 71, 384–394. (available at https://manowar.tamucc.edu/ login?url�http://search.proquest.com/docview/230848483?accountid�7084) o’curry, s. (1999). income source effects. unpublished working paper, depaul university. parker, j. (1999). the reaction of household consumption to predictable changes in social security taxes. american economic review, 89, 959–973. read, d., loewenstein, g., & rabin, m. (1999). choice bracketing. journal of risk and uncertainty, 19, 1–3, 171–197. rizzo, j. a., & zeckhauser, r. j. (2003). reference incomes, loss aversion, and physician behavior. the review of economics and statistics, 85, 909–922. rucker, m. h. (1984). allocation of windfall income: a case study of a retroactive pay increase to university employees. the journal of consumer affairs, 18, 101–118. sahm, c. r., shapiro, m. d., & slemrod, j. (2012). check in the mail or more in the paycheck: does the effectiveness of fiscal stimulus depend on how it is delivered? american economic journal-economic policy, 4, 216–250. shapiro, m. d., & slemrod, j. (1995). consumer response to the timing of income: evidence from a change in tax withholding. american economic review, 85, 274–283. shaprio, m. d., & slemrod, j. (2003). did the 2001 tax rebate stimulate spending? evidence from taxpayer surveys. in j. m. poterba (ed.), tax policy and the economy (pp. 83–109). cambridge, ma: national bureau of economic research and mit press. shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609–643. 312 v. chambers et al. / financial services review 26 (2017) 291–313 sheppard, b., hartwick, j., & warshaw, p. (1988). the theory of reasoned action: a meta-analysis of past research with recommendations for modifications and future research. journal of consumer research, 15, 325–343. slemrod, j., & bakija, j. (2004). taxing ourselves: a citizen’s guide to the debate over taxes (3rd ed.). cambridge, ma: the mit press. souleles, n. s. (2002). consumer response to the reagan tax cuts. journal of public economics, 85, 99–120. spencer, m., & chambers, v. (2012). national heuristic shift toward saving any form of tax rebate. accounting and the public interest, 12, 1–31. sprenger, c., & stavins, j. (2010). credit card debt and payment use. financial services review, 19, 17–35. thaler, r. h. (1999). mental accounting matters. journal of behavioral decision making, 12, 183–206. trump, r. k., finkelstein, s. r., & connell, p. m. (2015). i will risk a stranger’s money, but not my own or my friend’s money: effect of proximity of the money source to the self on financial risk-taking. marketing letters, 26, 501–512. walters-york, l. m., & curatola, a. p. (1998). recent evidence on the use of students as surrogates subjects. advances in accounting behavioral research, 1, 123–143. wertenbroch, k. (2003). “self-rationing: self-control in consumer choice,” in g. loewenstein, d. read, r. baumeister (eds), time and decision: economic and psychological perspectives on intertemporal choice, russell sage foundation, new york (2003), pp 491–516. winkelmann, r., oswald, a. j., & powdthavee, n. (2011, august). what happens to people after winning the lottery? european economic association & econometric society parallel meetings (pp. 25–29). zagorsky, j. l. (2013). do people save or spend their inheritances? understanding what happens to inherited wealth. journal of family economic issues, 34, 64–76. 313v. chambers et al. / financial services review 26 (2017) 291–313 expense ratios and net alphas of large cap funds: do expenses add value? abhay kaushika,*, raymond boisvertb adepartment of accounting, finance and business law, college of business and economics, radford university, 801 east main street, radford, virginia, 24142, usa bdepartment of accounting, finance and business law, college of business and economics, radford university, 801 east main street, radford, virginia, 24142, usa abstract global equity markets witnessed a tumultuous time-period of decline followed by growth, starting from the early part of 2000 to the beginning of the current period. investment in actively managed mutual funds also experienced a decent growth over the same period though there were times when investors especially retail investors sat on the sidelines and seemed reluctant to invest in equity funds. this study analyzes the performance of large cap equity funds between january 2000 and december 2013. the main objective is to assess the performance as reflected in the alpha of funds conditioned on expenses. we apply both ols regression and ranked portfolio approaches to estimate the abnormal performance. results of this study show that large cap funds underperform against benchmarks after incorporating expenses. results also suggest that expenses are not the only reason behind their underperformance. © 2018 academy of financial services. all rights reserved. keywords: mutual funds; performance; expense ratios 1. introduction mutual funds are important vehicles used by retail investors not just to diversify their portfolios, but also to generate higher returns. within the space of mutual funds, large cap funds tend to be more stable. the question that many participants ask is whether these fees can be justified by the returns, or more importantly, by the net alphas generated by these funds. there is overwhelming support in the existing literature that actively managed funds * corresponding author. tel.: �1-540-831-6426; fax: �1-540-831-6209. e-mail address: akaushik@radford.edu (a. kaushik) financial services review 27 (2018) 99-113 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. underperform passive indices especially after incorporating expenses. the rise of low cost exchange traded funds (etfs) also echoes the same sentiment. however, despite all that noise actively managed funds have been growing at a remarkable rate for the past several decades, which raises the question of how the investment community especially retail investors view actively managed mutual funds. do they consider them expenses or strategic investments? according to the investment company institute’s 2013 fact book mutual funds are a $13.5 trillion industry and are currently held by 47.1 percent of american households. the growth in the number of mutual funds over the last 40 years has been impressive—from 361 in 1970 to 7,596 in 2012. mutual funds are attractive investment vehicles to novice and sophisticated investors alike because they offer a highly diversified investment and access to professional level management without having to commit to an extensive amount of capital. research indicates that the rise in popularity of mutual funds is related to the manner in which they are advertised. the marketing emphasis is placed on past performance versus other funds and the market as a whole, with little discussion of the associated fees. mutual fund fee structure usually includes some sort of sales charge, or load, administered at the purchase, sale, or during the life of an investment in a mutual fund. subsequently, the expense ratio of the fund is calculated by the sum of operational and administrative fees (also 12-b1 fees) divided by the average asset base of the fund. over the last 20 years, the trend in expense ratios has been in a steady decline. expense ratios for equity funds decreased more than 20% during the period 1990–2012, from 99 basis points in 1990 to 77 basis points by year-end 2102. many fund costs are fixed; therefore, as the net asset value of a fund rises the expense ratio tends to fall. according to the investment company institute, at the end of 2012 equity funds with expense ratios in the lowest quartile managed 72% of all equity funds’ total net assets. equity index funds had 80% and target date funds held 79% of total net assets in the lowest quartile. the investment company institute contends that the growth in popularity of index funds has driven down the expense ratios of actively managed funds so that they can remain competitive. there are quite a number of studies in the existing literature that document the relationship between performance of funds and expenses. this article extends the scope of earlier research. the objective of this research is to determine whether the difference in management fees associated with large cap equity mutual funds should be viewed as an investment or an expense from the perspective of investors. do funds that charge higher expense ratios offer commensurately higher risk-adjusted returns? are higher expense ratios “dead costs” that are purely detrimental to portfolio performance, or do they signify that the funds hire and retain better managers who can accurately time the market or otherwise provide greater riskadjusted returns, therefore allowing them to outperform the market in adverse conditions? this study contributes to the existing literature in a variety of ways. first, we analyze the impact of expenses on funds’ performance at a period when expense ratios charged by active funds are far less than they used to be. the major benefit of using the current period is to see whether the impact is still the same or if it does not really matter anymore. secondly, we are not using the expense ratio as another variable in cross-sectional analysis to see the impact of expenses on a fund’s alpha. we are using the rank-portfolio approach to construct portfolios purely based on the expense ratio charged by different funds, and then comparing the alpha of extreme portfolios to analyze whether expenses are really good or bad for fund 100 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 investors. we are using a dynamic strategy to capture the impact of expenses on alpha. our study spans 14 years and funds might change their expense ratio over time; therefore, we rebalanced portfolios every year in order to capture the real-time impact of expenses on the funds’ alpha. since expenses are vastly blamed for the underperformance of actively managed funds and for the growth of low-cost exchange traded funds (etfs), we believe that the rank-portfolio methodology and more recent time period provide interesting implications and expand the existing research on this topic. the answers to these questions could make a huge difference in an investor’s choice in fund and bottom-line. what if higher expenses also generate better returns and not just returns but better alpha? if that is the case, then higher expenses should not be viewed as expenses but as strategic investments, especially in the case of portfolio management, where alpha is more relevant than raw return. existing literature is divided on the implications of expenses on the net alpha. for example, a recent research study conducted by barber, odean, and zheng (2005) indicates a negative correlation between expense ratios and fund performance; enhancing the notion that lower expense ratio funds are the best choice for investors. gruber (1996) and carhart (1997) also show that a fund’s net return is negatively affected by the expense ratio. sirri and tufano (1998) further suggest that expenses are not only bad for existing investors, but they also act as deterrents for new investors. many other studies also concluded that higher expenses are not justified by fund managers and that the net performance depreciates significantly for funds with higher expenses (e.g., bogle, 1998; bollen and busse, 2006; chalmers, edelen, and kadlec 1999; edelen, evans, and kadlec, 2013; hooks, 1996; ippolitio, 1989; wermers, 2000; among others). however, not all studies view expenses negatively. funds that charge higher expenses may generate higher net of cost returns, help existing shareholders by deterring those investors who are not in for the long haul, or attracting and retaining managers that have the potential to outperform regardless of economic conditions. for example, droms and walker (1995) analyzed equity mutual funds by using a pooled cross-section/time series regression methodology to evaluate the impact of fund characteristics on fund performance. nanigian (2012) suggests that the adverse relationship between expenses and performance tends to dissipate, specifically for funds that serve sophisticated investors. these findings support the notion that higher expenses might be able to attract managers who can outperform passive indices based on risk-adjusted performance. higher expenses offer incentives to portfolio managers to use that extra cash to invest for superior research, retain quality managers, generate better investment ideas, and screen out short-term investors; thus, giving higher net returns to shareholders (e.g., golec, 1996; grinblatt and titman, 1994; among others). chordia (1996) and nanda, narayanan, and warther (2000) find support for the theory that higher loads should accompany better performances because this helps managers screen out investors, especially those who are motivated by short-term gains, and limits the fund managers’ unnecessary trading activities; thus, reducing the trading and transaction costs significantly. lin (2014) advocates that sector specific actively managed funds not only achieve higher returns, but also higher risk adjusted returns; thus, supporting the value added component of actively managed funds. tufano and sevick (1997) suggest that expense ratio declines with fund size. their findings are interesting for our research because we are specifically analyzing the interplay between expenses 101a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 and performance of large cap funds. moreover, given the excessive volatility experienced in the last 15 years, more current research is needed to determine whether fund managers that charge higher expense ratios have “earned their keep” in the face of large market fluctuations. we select january 2000 through december 2013 for this research because this period witnessed some of the most volatile and important events in the u.s. economy. global markets suffered several cataclysmic events. in particular, the u.s. economy suffered two recessions during this time. on october 31, 2002, the s&p 500 bottomed at 768.63 and then on march 31, 2009 it bottomed even lower at an ominous 666.79. this time period also felt the lingering effects of the .com bubble, the september 11 attack on the world trade center and pentagon, hurricane katrina in 2005, the collapse of the mortgage industry in 2007 and the subsequent shift in the fed’s monetary policy, the highest unemployment since 1983, and the may 6, 2010 flash crash, among many other market moving occurrences. specifically, in this study, we analyze large cap equity funds because we believe that large cap equity fund managers of high expense funds would have used their superior market timing and stock picking abilities to navigate their investors through adverse market conditions compared with their low-fee counterparts thus justifying the premium they charge investors. this paper analyzes u.s. domiciled large cap actively managed equity funds across all three morningstar classifications: value, growth, and blend. the funds are sorted based on their expense ratios and then compared on the basis of their risk-adjusted return during the period january 2000 to december 2013 in an effort to determine if there is any value to investing in funds with higher expense ratios or if they are just a means to take advantage of unsophisticated investors. this paper is organized as follows: section 2 reviews the existing literature on equity mutual funds; section 3 describes the data; section 4 describes the methodology; section 5 summarizes the empirical results; and section 6 concludes the paper. 2. literature review prior research shows mixed results on the performance of actively managed mutual funds. in a seminal paper, jensen (1968) shows that fund managers do not possess any superior forecasting abilities to outperform the market and any abnormal performance by an individual fund is merely by chance. jensen’s (1968) findings are further strengthened by grossman and stiglitz (1980), who examine the mutual funds’ performance and propose that the funds’ excess gross return is zero because the costs associated with obtaining that superior information cancel out any benefits earned from getting that information. malkiel (1995) echoes the similar sentiments that markets are efficient. however, grinblatt and titman (1989 and 1993) show that not all but some managers do possess superior skills and ability to beat the passive benchmark on a gross return basis. gruber (1996) and wermers (2000) show that funds underperform corresponding benchmarks after considering expenses; wermers (2000) findings show that funds outperform corresponding benchmarks by 1.3%; however, the same funds on a net return basis underperform corresponding benchmarks by 1%. however, ciccotello and grant (1996) analyze performance based on size and show that large funds, especially those that increase in assets because of their past performance, might not be the right choice for individual investors. 102 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 mutual fund literature on actively managed funds has extensively examined the performance against passive benchmarks either inclusive of expenses or net of expenses, and, as previously noted, generally finds that actively managed funds underperform the benchmarks on a net basis. wermers (2000) advocates that even though funds with higher turnover ratios experience higher transaction costs, managers tend to rebalance their portfolios with stocks with substantially higher returns compared to those with fewer turnover funds. the higher return is more than enough to offset any additional expenses these higher turnover funds might experience because of trading costs. on the other hand, a sizeable number of studies such as those conducted by carhart (1997), chalmers, edelen, and kadlec (2000), among others, find the opposite effect: turnover ratios have a significantly negative impact on the funds’ performance. carhart (1997) shows a negative relationship between a fund’s expenses and performance. similar to carhart (1997), gruber (1996) and wermers (2000) find that funds underperform passive indices after considering expenses. droms and walker (1995) analyzed 150 equity funds. they constructed portfolios based on riskiness of funds’ returns. their research shows that funds with higher expenses tend to have higher risk, but the findings also suggest that funds with higher expense ratios also tend to generate higher risk-adjusted returns. their results find support in the notion that funds that charge higher expense ratios are able to allocate more money into quality research and, thus, are able to generate superior returns. in short, their research suggests that higher expenses are not really bad for investors. detzel and weigand (1998) analyze equity funds and conclude that persistence of performance is very much dependent on size. chrodia (1996) advocates that funds that are less susceptible to redemptions tend to be more profitable. his research findings indicate that higher expenses can be used as tools by funds to dissuade short-term investors who redeem their investments quickly, thus adding costs to other shareholders. 3. data and descriptive statistics we use the morningstar direct database to collect data points related to our sample funds during the period january 2000–december 2013. any fund that is classified as u.s. large cap value/growth/blend by morningstar’s global category is selected as a sample fund. since we are interested in analyzing domestic retail funds, any fund that is classified as an international or institutional fund is removed from the selection process. we also ensured that all funds are u.s. domiciled and have at least 90 percent of money invested in stocks. consistent with the existing literature, our final sample consists of only those funds that have at least 36 monthly observations. many times funds are sold with multiple share classes. however, all different classes have the same claims and holdings, and they only differ in terms of their fee structure. therefore, if a fund has multiple share classes, we select the oldest share class for empirical purposes. our final sample consists of 431 unique funds. out of 431 funds, 230 are large cap growth funds, while 149 are large cap value funds and 146 are large cap blend funds. descriptive statistics of sample funds are reported in table 1 panel a. on average, sample funds invest nearly 96 percent in stocks and 3 percent in cash holdings. average fund size is around $1.2 billion over the 14-year period. the highest average net assets are observed in 2000, whereas the lowest are observed in 2009. the u.s. equity market saw a steep decline 103a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 t ab le 1 pa ne l a : d es cr ip tiv e st at is tic s fo r sa m pl e fu nd s t hi s ta bl e sh ow s th e fu nd sp ec ifi c va ri ab le s fo r sa m pl e fu nd s in ex is te nc e be tw ee n th e ye ar s 20 00 an d 20 13 . r et is th e av er ag e m on th ly re tu rn fo r th e gi ve n ye ar . t ur n is th e av er ag e tu rn ov er ra tio , ca lc ul at ed as pu rc ha se s or sa le s (w hi ch ev er is le ss ) di vi de d by av er ag e m on th ly ne t as se ts . c as h is th e av er ag e pe rc en ta ge of po rt fo lio s th at ar e al lo ca te d to ca sh . e qu ity re pr es en ts th e av er ag e pe rc en ta ge of po rt fo lio s th at ar e al lo ca te d to eq ui ty . t t o p is av er ag e po rt io n of po rt fo lio s al lo ca te d to th e to p 10 ho ld in gs . e x p is th e av er ag e ex pe ns e ra tio ch ar ge d by th e fu nd . t n a is th e av er ag e as se ts un de r m an ag em en t. h ol di ng s sh ow th e av er ag e nu m be r of ho ld in gs th e fu nd s co nt ai n. r an ki ng is th e av er ag e sc or e (o ut of 5) as si gn ed to th e fu nd s by m or ni ng st ar . t en ur e is th e av er ag e tim e in ye ar s th at m an ag er s ha ve be en w ith th ei r re sp ec tiv e fu nd s. r a ss is th e ri sk ad ju st ed su cc es s ra tio . n is th e nu m be r of fu nd s in ex is te nc e pe r ye ar . y ea r r et t ur n c as h e qu ity t t o p e x p t n a (i n m ill io n) h ol di ng s r an ki ng t en ur e r a ss n 20 00 � 1. 44 % 83 .7 6% 4. 01 % 94 .5 4% 35 .7 6% 1. 20 % $ 2, 26 7. 35 98 .3 7 3. 50 8. 40 43 .0 1 37 9 20 01 � 10 .9 2% 87 .0 0% 3. 82 % 95 .2 4% 34 .8 9% 1. 20 % $ 1, 61 4. 20 10 0. 15 3. 34 8. 28 43 .1 5 30 6 20 02 � 22 .8 0% 86 .7 6% 3. 41 % 94 .8 1% 33 .2 7% 1. 32 % $ 1, 21 0. 47 10 7. 43 3. 14 8. 28 43 .3 4 31 8 20 03 26 .2 8% 75 .8 4% 3. 43 % 95 .4 3% 33 .3 4% 1. 20 % $ 1, 07 2. 64 10 5. 01 3. 17 8. 23 43 .2 0 33 0 20 04 11 .2 8% 12 7. 56 % 3. 02 % 96 .1 1% 31 .8 4% 1. 20 % $ 1, 21 5. 63 10 9. 41 3. 17 8. 23 43 .3 0 33 9 20 05 7. 20 % 71 .1 6% 2. 92 % 96 .3 2% 31 .9 6% 1. 20 % $ 1, 22 3. 27 10 4. 97 3. 08 8. 19 43 .0 8 36 0 20 06 11 .8 8% 75 .3 6% 2. 92 % 96 .3 2% 31 .4 2% 1. 20 % $ 1, 21 0. 53 98 .9 0 3. 04 8. 10 43 .1 0 37 6 20 07 8. 88 % 73 .0 8% 2. 61 % 96 .6 8% 31 .3 8% 1. 20 % $ 1, 29 9. 31 97 .0 4 3. 01 8. 07 42 .9 5 39 1 20 08 � 44 .1 6% 82 .6 8% 3. 03 % 96 .2 2% 31 .7 7% 1. 20 % $ 1, 07 1. 73 10 2. 46 3. 07 8. 00 42 .8 7 40 2 20 09 28 .2 0% 85 .8 0% 2. 64 % 96 .6 2% 30 .9 0% 1. 20 % $ 72 1. 51 10 3. 94 3. 05 7. 94 42 .7 6 41 3 20 10 15 .7 2% 73 .5 6% 2. 21 % 97 .2 1% 30 .1 9% 1. 20 % $ 82 8. 10 10 4. 61 2. 96 7. 83 42 .9 4 43 0 20 11 0. 60 % 68 .6 4% 2. 25 % 97 .2 3% 30 .5 8% 1. 20 % $ 91 0. 47 10 2. 26 2. 95 7. 80 42 .7 7 43 1 20 12 14 .5 2% 65 .4 0% 2. 07 % 97 .5 7% 32 .1 6% 1. 20 % $ 89 3. 59 10 0. 44 2. 94 7. 80 42 .7 7 43 1 20 13 29 .4 0% 63 .3 6% 1. 95 % 97 .8 0% 30 .8 1% 1. 08 % $ 98 0. 42 99 .7 3 2. 88 7. 80 42 .7 7 43 1 a ve ra ge 5. 33 % 80 .0 0% 2. 88 % 96 .2 9% 32 .1 6% 1. 20 % $ 1, 17 9. 94 10 2. 48 3. 09 8. 07 43 .0 0 38 1 104 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 in the mid-2000s, and we believe that is reflected in the lowest tna under management in 2009. most of these funds seem to be well-diversified as reflected in the average number of holdings during the 14-year long period. on average, sample funds manage 102 holdings per year and this number is nearly consistent every year from 2000 to 2014. large cap funds are supposed to be stable funds and it is further strengthened from the fact that, on average, managers stayed with the fund for eight years when we look at the average managerial tenure of all 431 funds over the 14 years of this study. the data further suggest that large cap funds charge less fees from their investors. on average, the monthly expense ratio is 0.10 percent table 1 panel b: descriptive statistics of individual categories of large cap funds the table shows the mean values of large cap value/growth/blend funds over the 14-year period (01/2000 to 12/2013). ret is annual return generated by sample funds. exp is the annual expense ratio. turn is the turnover ratio. ttop is the funds’ investment in their top 10 assets. cash is the annual cash holdings. equity is the funds’ investments in stocks. holdings is the number of holdings in an average fund. tenure is the average managerial tenure with a funs. large cap value large cap growth large cap blend ret 7.37% 5.50% 6.20% exp 1.11% 1.24% 1.15% turn 62.55% 79.76% 73.63% ttop 31.50% 33.10% 28.73% cash 2.49% 2.48% 2.30% equity 96.57% 97.19% 96.41% holdings 91.13 87.49 136.22 tenure 7.96 8.24 7.69 table 1 panel c: descriptive statistics for market factors this table shows the return on the market, risk free rate, small minus big, high minus low, and momentum factors. rm is the average monthly return of the crsp value weighted index. rf is the average monthly yield for a one-month treasury bill. smb is the difference in returns between small and large cap stocks. hml is the difference in returns between high and low book-to-market stocks. mom is the difference in returns between stocks with high and low past returns. n is the number of monthly observations. year rm rf smb hml mom n 2000 �0.93% 0.48% �0.02% 3.06% 1.64% 168 2001 �0.82% 0.31% 1.64% 1.29% �0.27% 168 2002 �1.79% 0.13% 0.37% 1.03% 2.33% 168 2003 2.38% 0.08% 1.69% 0.30% �1.51% 168 2004 0.97% 0.10% 0.44% 0.71% 0.03% 168 2005 0.53% 0.25% �0.12% 0.69% 1.16% 168 2006 1.22% 0.39% 0.06% 1.02% �0.53% 168 2007 0.49% 0.38% �0.68% �1.02% 1.78% 168 2008 �3.56% 0.13% 0.60% 0.18% 1.59% 168 2009 2.30% 0.01% 0.68% �0.02% �5.33% 168 2010 1.52% 0.01% 1.05% �0.15% 0.50% 168 2011 0.14% 0.00% �0.42% �0.54% 0.69% 168 2012 1.31% 0.01% 0.05% 0.56% �0.05% 168 2013 2.57% 0.00% 0.50% 0.03% 0.53% 168 average 0.45% 0.16% 0.42% 0.51% 0.18% 168 105a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 (1.2 percent annually) and this average is almost constant per year from 2000 to year-end 2013. although most of these funds manage a good number of holdings, it appears that fund managers tend to invest a large portion in their top picks. the data show that, on average, almost one-third (32 percent) of investment is made in the top 10 percent holdings. turnover ratio is less than 100 percent over the 14-year period. highest turnover ratio of 127 percent is seen in 2004, whereas the lowest of 64 percent is observed in year-end 2013. the highest number of funds (431) exist in each year from 2010 to 2013 whereas the lowest number of 306 funds is observed in 2001. on average, 381 funds are observed per year over the 14-year period. table 1 panel b shows a few key descriptive statistics for individual categories. on average, large cap growth funds earn 5.50 percent return per year over the 14-year period. on average, during the 14-year period, large cap growth funds combined charge an expense ratio of approximately 1.23 percent per year, manage 88 holdings worth $1.445 billion in assets, and use roughly 2.5 percent of assets as cash holdings. similar statistics are given in table 1 panel b for large cap blend and value funds. for example, on average per year, a large cap blend fund earns 6.20 percent return, managing 136 holdings with over $833 million in assets and charges 1.15 percent expense ratio during the 14-year period, whereas the annual statistics for value funds are 7.37 percent return, 63 percent turnover, 91 holdings, and $680 million in assets per fund and charge an expense ratio of 1.11 percent over the same period. these statistics offer good insight into each category. growth funds manage higher dollar amount in assets and lowest return, whereas value funds manage lower amount with highest return. also, the value category charge lowest expenses and the growth charge the highest. these descriptive statistics show sharp differences among all these different categories and support the idea of this study to evaluate all of these categories as one group and also separately. table 1 panel c shows a few key statistics of the market and market related portfolios. the average return of market portfolio (the crsp value weighted index) is 0.45 percent per month (5.40 percent annual equivalent) whereas the average annual return on the u.s 1-month t-bills rate is 1.92 percent. proxy portfolios returns that mimic risk premium generated from small stocks, value stocks, and momentum are 5.04 percent, 6.12 percent, and 2.16 percent, respectively. returns on the crsp value weighted index, the one month u.s. treasury bills, and market mimicking portfolios smb, hml, and mom are taken from kenneth french’s website. the highest monthly market return is 2.57 percent observed in year-end 2013. the lowest of �3.56 percent is seen in the 2008. interestingly, during 2008, risk premium generated by small and value stocks are positive and so is the risk premium generated by the momentum effect. a simple comparison of returns between sample funds and the market index shows that in 2013, sample funds generate 2.45 percent return and �3.68 percent in 2008. this comparative analysis clearly shows a direct relationship between sample funds and the market. 4. methodology the sharpe (1964) – lintner (1965) capital asset pricing model (capm) is the primary and most often used tool to price assets. the capm states that in equilibrium, expected returns are linearly related to their level of risk, more specifically, their beta or systematic 106 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 risk. many tests and models have been developed over the years to measure performance of the mutual funds’ managers. jensen’s (1968) alpha is perhaps the best known primary model. rit � rft � �i � �i * rmrft (1) where: rit � rft is the excess return on fund i over the 1-month t-bill rate, �i is the measure of the portfolio’s performance (jensen’s alpha), rmrft � rmt � rft is the excess return on the market, and �i � is the unconditional measure of risk. we use monthly observations over the period january 2000 to december 2013. it is argued that the market model alone does not reflect true risk inherent in the funds’ performance and therefore, in order to control the funds’ selectivity performance for investment style, we also use the four-factor model of carhart (1997), which adjusts fund excess return for the fama-french (ff) factors smb, hml, and carhart’s momentum in addition to difference in returns between small and large capitalization stocks and the difference in returns between high and low book-to-market stocks. rit � rft � �i � �1i * rmrft � �2i * smbt � �3i * hmlt � �4i * momt � �i, t (2) where: rmrft is the excess monthly return (market return net of monthly t-bill return) on the crsp value weighted index, smbt is the difference in returns between small and large capitalization stocks and hmlt is the difference in returns between high and low book-to-market stocks. momt is the difference in returns between stocks with high and low past returns. monthly smbt, hmlt, and momt factors are taken from the kenneth french web site. a positive alpha in equation (2) shows the positive stock picking abilities or selectivity of the fund managers, whereas a negative alpha corresponds to negative selectivity of fund managers. a zero alpha means that fund managers have no selectivity ability and they are not able to either underperform or overperform the comparative benchmark. since funds may differ from each other, first we estimate alpha for each fund separately using equation (2) and then estimate asset-weighted average alpha of the entire portfolio. further, we use newy-west regression methodology to control for any heteroscedasticity and auto-correlation in error terms. we estimate alphas overall and separately across each large cap category. 4.2. abnormal performance based on expense ratio the main objective of this study is to evaluate abnormal performance conditioned on expenses; therefore, we use portfolio approach to estimate expense ratio conditioned alpha. 107a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 we form quintiles each year based on the expense ratio in the previous year. in other words, we rank funds based on expense ratio1; for example, year 2000 is the first formation year. at year-end 2000, we sort funds based on expense ratios and form quintiles. q1 portfolio consists of funds that belong to the lowest 25 percent of funds based on expense ratio whereas q5 contains funds with highest expense ratio. since expense ratio may change overtime, we follow dynamic strategy and repeat the step at the end of each year. doing so gives us a time series of funds based on their expense ratios. we estimate alpha of q1 and q5 using four-factor carhart model (equation 2) mentioned above. the main idea is to estimate different alphas for funds with the highest and lowest expense ratios. in order to have an even better understating of those alphas, we also estimate statistically the difference between those extreme alphas.2 finally, to get an even clearer understanding of large cap funds and expenses charged, we re-estimate expense ratio conditioned alphas for value, growth, and blend categories separately. 5. results results are reported in tables 2, 3, and 4. interesting results are obtained when we run the four-factor model across all 431 large cap funds. funds in our sample differ in terms of assets under management; therefore, we estimate asset-weighted alpha for each fund. out of 431 funds, 244 funds exhibit negative abnormal performance, whereas 187 funds have positive alpha estimates. out of these funds, 71 funds (16.47 percent of sample funds) have alphas that are statistically significant. out of 71 statistically significant funds, 19 funds generate positive abnormal performance and 52 funds (73 percent) are poorly performing funds. we run the same model across each individual category. table 2 reports these statistics overall and across individual categories. maximum number of funds, 179, is available in growth category; 13 percent of funds (17 funds) show significant alphas. out of 17 funds, 10 funds outperform, and 7 funds underperform the passive index statistically. these numbers vary across categories. in the large cap growth category, 23 out of 179 funds exhibit statistically significant alphas whereas a majority, 156 funds, fail to earn alpha statistically different from table 2 alphas overall and across categories the following table shows number of funds that participated in the four-factor model and the number of funds that exhibited positive, negative, significant, and insignificant alphas. overall value growth blend n 431 134 179 118 positive 187 75 67 45 negative 244 59 112 73 significant 71 17 23 24 insignificant 360 117 156 94 positive and significant 19 10 3 6 negative and significant 52 7 20 18 significant @ 1% 15 1 3 7 significant @ 5% 25 10 7 10 significant @ 10% 31 6 13 7 108 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 zero. three funds outperform whereas 20 funds show negative performance. in the large cap blend category, all together 118 funds participated in regressions and only 6 outperform the market whereas 18 underperform against the passive index. next, we estimate the average alpha and beta coefficients for the four-factor model. on average, large cap funds generate �0.007 percent alpha per month (�0.084 percent annually), however, the alpha coefficient is statistically insignificant. regression analysis shows positive and significant association between the excess market returns and sample funds’ returns. looking at the premium attached to various groups, the sample funds’ returns are negatively related to small stocks and positively related to excess market returns, value stocks, and momentum effects. all these coefficients are statistically significant. we are not surprised by these results because our sample funds are large cap stocks so their inverse relationship to the premium attached to small stocks is quite obvious and the positive relationship to market and value stocks is also very apparent. we run the same regressions for individual categories. startling differences are observed when we run the same regressions across individual categories. alpha of large cap value funds is positive (0.0503 percent per month or 0.603 percent annually) and it is statistically significant. sample funds’ returns are positively and statistically significant related to market returns and value stocks and negatively (and statistically significant) related to small stocks premium and momentum effects. these results are also very obvious given that we are analyzing large cap value stocks. the coefficients of rmrf and hml are not really surprising because existing literature also suggests a positive and significant association between funds and market returns and since we are analyzing the value funds; therefore, it is obvious to have a positive relationship between value premium and returns. in the growth category, alpha estimate is negative. rmrf and mom effects are positively related to funds’ returns and hml coefficient is negatively attached to funds’ returns. a negative relationship between value premium and growth stocks is quite expected. table 3 panel d reports the same statistics for large cap blend funds. the alpha estimate is insignificant whereas rmrf, smb, hml, and mom parameter estimates are statistically highly significant. results are reported in table 3 panel a to d. finally, we estimate the performance of funds based on the expense ratios they charge their customers. we use ranked portfolio approach to estimate the performance conditioned on expenses. at the beginning of each year, we rank the entire portfolio based on the annual expense ratio and construct quintiles. the sorting is done based on prospectus based expense ratio. q1 is the quintile (group of funds) that charges the lowest expense ratio and q5 is the group of funds that charges the highest expenses. since funds’ expense ratio may change over time, we use dynamic rebalancing as opposed to static approach and repeat the same step at the beginning of each year. year-end 2000 is our first formation period whereas year-end 2012 is the last formation period. results are tabulated in table 4. findings show a good difference between the extreme portfolios. on average, funds that charge lower expenses deliver higher alphas overall and across individual categories except value category. however, results also show that, on average, funds in these two extreme groups underperform against the passive index. another interesting observation is the difference in alphas between two extreme groups. results are interesting because they seem to suggest that the difference in alpha is not just because of expenses. for example, the average monthly 109a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 alpha difference between funds that charge higher expenses and lower expenses is �0.067 percent (�0.804 percent annually) statistically significant, whereas the average annual difference in their expense ratio is just 30 basis points. in other words, the majority of the difference (50 basis points difference) is coming from factors other than expense ratio. results further strengthen the notion that many times performance deteriorates not because of expenses charged by funds, but by over-diversification or managers’ behavior of taking more risk in anticipation of better alpha. 6. conclusion mutual funds are supposed to offer low-cost diversification and consistent returns to retail investors and it is perhaps one of the main reasons of steep growth of mutual funds year by year for the past several decades. the question of expenses and alphas has taken center stage in academic and practitioner research. one of the main questions investors want to know is table 3 panel a: abnormal performance of funds the following table shows the abnormal performance of the sample funds over the period 1/2000 to 12/2013. the abnormal performance (�) is based on the four-factor model. rit is the excess monthly return of fund i over one month u.s. t-bill return. rmrf is the excess monthly return of the crsp value weighted index over the one month u.s. t-bill return. smb, hml, and mom are monthly returns of size (the difference in returns between small and large cap stocks), book to market (the difference in returns between high and low book-to-market stocks), and momentum (the difference in returns between stocks with high and low past returns) portfolios, respectively. the dependent variable is the individual fund’s monthly excess return over the corresponding one month t-bill rate. alpha is expressed in percentage per month. regression estimates are based on newey-west adjusted standard errors. n is the number of fund month observations. model: rit � rft � �i � �1i * rmrft � �2i* smbt � �3i * hmlt � �4i * momt � �i, t variable name mean standard error t-value p-value alpha �.007 0.000082 �0.89 0.3732 rmrf 0.983155*** 0.00319 308 0.0000 smb �0.04745*** 0.00775 �6.12 0.0000 hml 0.011741* 0.00667 1.76 0.0782 mom 0.014201*** 0.00311 4.57 0.0000 n 62,497 ***, **, and * show the significance at 1%, 5%, and 10% level, respectively. table 3 panel b: abnormal performance of value funds variable name mean standard error t-value p-value alpha 0.0503*** 0.000147 3.41 0.0006 rmrf 0.919338*** 0.00526 174.69 0.000 smb �0.13051*** 0.00889 �14.68 0.0000 hml 0.285684*** 0.00995 28.71 0.0000 mom �0.01315** 0.0053 �2.48 0.0131 n 15,465 ***, **, and * show the significance at 1%, 5%, and 10% level, respectively. 110 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 whether expenses charged by these funds are justified by their performance. while there is overwhelming support in the existing literature that actively managed funds underperform passive indices and the underperformance is more visible after incorporating expenses, still, there are studies that documented positivity of expenses. their arguments include from dissuading short-term investors to using extra cash for better research to attracting and retaining quality portfolio managers. equity markets have seen a roller coaster ride since the beginning of 2000 and the global markets have tested a tumultuous time-period in the past decade. based on all of these circumstances, we realize that a thorough analysis of expensebased performance is warranted. we use only large cap funds in our research because they tend to be more stable and are supposed to withstand the market volatility. in this study, we analyze large cap funds from january 2000–december 2013. we analyze large cap funds as one group and reexamine the performance across individual categories, namely: value, growth, and blend categories. we document that not all types of large cap funds underperform against the passive index. our results indicate superior performance by value funds. in our study, they earn on average, 600 basis points alpha over the 14-year period whereas, on average, large cap growth funds underperform by nearly 240 basis points during the same period. we divide the entire portfolio into quintiles based on the expense ratio they charge from investors and our results show amazing differences in their performance across those divisions. generally, expenses are blamed for lower net returns to investors, but our results suggest that expenses are not the only reasons behind the underperformance of funds. expense ratio does decrease the net alpha earned by a fund’s investor, however, a lot of other fund-specific factors contribute to the majority of negative alpha. table 3 panel c: abnormal performance of growth funds variable name mean standard error t-value p-value alpha �0.024* 0.000129 �1.84 0.0661 rmrf 1.029688*** 0.00495 208.07 0.000 smb 0.003744 0.0131 0.29 0.7755 hml �0.16509*** 0.01 �16.44 0.000 mom 0.024633*** 0.00485 5.08 0.000 n 24,678 ***, **, and * show the significance at 1%, 5%, and 10% level, respectively. table 3 panel d: abnormal performance of blend funds variable name mean standard error t-value p-value alpha �0.018 0.000127 �1.44 0.1486 rmrf 0.953652*** 0.00469 203.21 0.0000 smb �0.09858*** 0.00832 �11.84 0.0000 hml 0.082804*** 0.00856 9.67 0.0000 mom 0.012906*** 0.0044 2.94 0.0033 n 14,463 ***, **, and * show the significance at 1%, 5%, and 10% level, respectively. 111a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 notes 1 we use raw expense ratio of each fund in the sorting process. 2 we estimate the difference in returns of the two series and use it as the dependent variable to run the four-factor regression. references barber, b., odean, t., & zheng, l. (2005). out of sight, out of mind: the effects of expenses on mutual fund flows. journal of business, 78, 2095–2119. bogle, j. (1998). the implications of style analysis for mutual fund performance evaluation. the journal of portfolio management, 24, 34–42. bollen, n., & busse, j. (2006). tick size and institutional trading costs: evidence from mutual funds. journal of financial and quantitative analysis, 41, 915–937. carhart, m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. chalmers, j., edelen, j., & kadlec, g. (1999). mutual fund returns and trading costs: evidence on the value of active fund management. working paper. chalmers, j., edelen, j., & kadlec, g. (2000). an analysis of mutual fund trading costs. working paper. chordia, t. (1996). the structure of mutual fund charges. journal of financial economics, 41, 3–39. ciccotello, c., & grant, c. (1996). equity fund size and growth: implications for performance and selection. financial services review, 5, 1–12. detzel, l., & weigand, r. (1998). explaining persistence in mutual fund performance. financial services review, 7, 45–55. droms, w., & walker, d. (1995). determinants of variation in mutual fund returns. applied financial economics, 5, 383–389. edelen, r., evans, r., & kadlec, g. (2013). shedding light on “invisible” costs: trading costs and mutual fund performance. financial analysts journal, 69, 33–44. fama, e., & french, k. (1993). common risk factors in the return on bonds and stocks. journal of financial economics, 33, 3–53. table 4 performance of funds based on expense based quintiles the table reports the fund’s abnormal performance conditioned on expense ratio over the period 01/2000 to 12/2013 with year-end 2000 as the first formation period. quintiles are formed at the beginning of each year based on previous year’s annual expense ratio. quintile 5 consists of funds that belong to top 20% expense ratio level in the formation year and quintile 1 consists of funds that are at the bottom of the 20% of the group based on annual expense ratio in the formation year. alpha is the average alpha (in percentage per month) of each series. regression estimates are based on newey-west adjusted standard errors. rmrf is the excess monthly return of the crsp value weighted index over the corresponding one month risk free (t-bill rate). the table provides the mean alphas for the quintiles and the difference in the mean alpha in percentage per month between the extreme portfolios q1 and q5. model: rit � rft � �i � �1i * rmrft � �2i * smbt � �3i * hmlt � �4i * momt � �i, t quintile 1 quintile 5 quintile 1 to quintile 5 overall �0.036** �0.087*** 0.051 value �0.049* �0.020 �0.029 growth �0.056** �0.123*** 0.067 blend �0.035* �0.074** 0.039 ***, **, and * show the significance at 1%, 5%, and 10% level, respectively. 112 a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 golec, j. (1996). the effects of mutual fund managers’ characteristics on their portfolio performance, risk, and fees. financial services review, 5, 133–148. grinblatt, m., & titman, s. (1989). mutual fund performance: an analysis of quarterly portfolio holdings. journal of business, 62, 393–416. grinblatt, m., & titman, s. (1993). performance measurement without benchmarks: an examination of mutual fund returns. journal of business, 66, 47–68. grinblatt, m., & titman, s. (1994). a study of monthly mutual fund returns and performance evaluation techniques. journal of financial and quantitative analysis, 29, 419–444. grossman, s., & stiglitz, j. (1980). on the impossibility of informationally efficient markets. american economic review, 70, 393–408. gruber, m. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51, 783–810. hooks, j. (1996). the effect of loads and expenses on open-end mutual fund returns. journal of business research, 36, 199–202. investment company institute. (available at https://www.ici.org). ippolito, r. (1989). efficiency with costly information: a study of mutual fund performance, 1965–1984. the quarterly journal of economics, 104, 1–23. jensen, m. (1968). the performance of mutual funds in the period 1945–1964. journal of finance, 23, 389–416. lin, c. (2014). does active management work? evidence from equity sector funds. financial services review, 23, 249–271. lintner, j. (1965). the valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. the review of economics and statistic, 47, 13–37. malkiel, b. (1995). returns from investing in equity mutual funds 1971 to 1991. journal of finance, 50, 549–72. nanda, v., narayanan, m., & warther, v. (2000). liquidity, investment ability, and mutual fund structure. journal of financial economics, 57, 417–443. nanigian, d. (2012). why do mutual fund expenses matter? financial services review, 21, 239–257. sharpe, w. (1964). capital asset prices: a theory of market equilibrium under conditions of risk. journal of finance, 19, 425–442. sirri, e., & tufano, p. (1998). costly search and mutual fund flows. journal of finance, 53, 1589–1622. tufano, p., & sevick, m. (1997). board structure and fee-setting in the mutual fund industry. journal of financial economics, 46, 321–355. wermers, r. (2000). mutual fund performance: an empirical decomposition into stock-picking talent, style, transactions costs, and expenses. journal of finance, 55, 1655–1695. 113a. kaushik, r. boisvert / financial services review 27 (2018) 99-113 financial services review, 33(1) 50 protecting well-being through financial shocks emily koochel,1 megan mccoy,2 and sonya lutter3 abstract the ability to provide more than problem-solving interventions is useful in reducing client stress. protective features can be built and amplified during financial uncertainty that may increase individuals’ resilience against factors that have the potential to cause damage to their financial well-being. to understand the predictive relationship between financial stressors, prior financial experience and exposure, and resources impact on financial well-being, a three-model hierarchical multiple regression was conducted with financial well-being as the dependent variable. greater availability of resources increased financial well-being above and beyond the effects of stressors and prior exposure and experience. however, it is important to note that greater availability of resources was not measured just by income, rather it was variables that assessed other forms of capital. specifically, individual qualities such as self-control and perceived health positively contributed to financial well-being. results indicate that maintaining financial well-being is about more than knowledge and skill. increasing opportunities for financial socialization and building clients’ sense of control may serve as a key buffer during times of financial stress. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation koochel, e., mccoy, m., & lutter, s. (2025). protecting well-being through financial shocks. financial services review, 33(1), 50-66. introduction in recent times, the phrase "unprecedented times'' has become an almost ubiquitous part of our daily vocabulary due to its frequent use. headlines concerning political upheaval (apostolakis et al., 2021), allegations of russian war crimes (oxford analytica, 2022), disruptions in supply chains (ramelli & wagner, 2020), and surging inflation rates (dunsmir, 2022) have collectively contributed to a sense of impending financial crises. financial counselors and planners have always been dedicated to helping clients deal with both large-scale financial shocks like these 1 corresponding author (ekoochel@emoneyadvisor.com). emoney, new york, ny, usa. 2 kansas state university, manhattan, ks, usa. 3 texas tech university, lubbock, tx, usa. and deeply personal household shocks like divorce and health crises. the responses of individual clients to financial shocks are diverse from resilience and nonadaptation to destabilizing shocks that erode their financial well-being (fox & bartholomae, 2020). which begs the question, what are the best ways of supporting clients to be able to buffer financial shocks? the purpose of this paper was to address strategies for maintaining and improving financial well-being among financial counseling and planning clients. https://creativecommons.org/licenses/by-nc/4.0/ mailto:ekoochel@emoneyadvisor.com https://creativecommons.org/licenses/by-nc/4.0/ koochel et al. 51 while financial professionals can attempt to prepare households for potential financial shocks through the elements and strategies of a financial plan, they may also need to complement the provision of technical financial advice with emotional and social support when these shocks occur through counseling-like skills (byram et al., 2023; dubofsky & sussman, 2011; fox & bartholomae, 2020). the capacity of financial professionals to address the psychological aspects of their clients, including financial stress, alongside relationship and behavioral issues, has become an integral facet associated with constructing and implementing a financial planning framework that leads to a trusting planner-client relationship (byram et al., 2023; mccoy et al., 2022). the primary objective of this study is to highlight the protective factors that financial counselors and planners can evoke in their clients, factors that could enhance clients' resilience when faced with financial shocks. framed within the stress and coping theory (lazarus & folkman, 1986), this paper aims to explore how prior financial experiences and exposures might influence an individual's assessment and interpretation of financial shocks. additionally, we seek to understand how an individual's past may affect their capacity to identify and utilize the financial, personal, and relational resources at their disposal, resources that foster financial wellbeing in the wake of financial shocks. framework and related literature lazarus and folkman’s (1986) stress and coping theory serves as a theoretical lens for understanding an individual’s reactions during a time of financial stress. the theory posits that our ability to cope with a stressor event is dependent on our unconscious appraisal of the event that gauges (a) how large of a threat the stressor is to them (primary appraisal) and (b) their available resources (secondary appraisal) that could be utilized to overcome the stress (coping) (lazarus & folkman, 1986). the theory views stress and coping from a transactional perspective as stress is seen as a result of our unique psychological, social, and cultural make up that come into play in determining our stress experience. stress is not seen as a universal experience but rather a byproduct of the interactions between our unique complex systems. this theory was developed to explain why individuals may have dissimilar reactions to the same stressor event. but, more importantly, this theory enables researchers to explore potential protective factors that increase resilience in light of stress. financial stressors and financial shocks financial shocks and financial stressors are related concepts, but they are not identical. “household financial shocks can result from decreases in income, such as job loss or reduced hours, or from increases in expenses due to emergencies, such as illness, injury, or damage to household possessions in natural disasters” (sun et al., 2022, p. 1). these shocks often occur unexpectedly. in contrast, financial stress refers to ongoing hardships, such as a household facing challenges in meeting basic needs due to a shortage of money. material financial hardships typically have an immediate impact on households' consumption and their ability to cover essential expenses such as food, clothing, utilities, and transportation. households experiencing high levels of financial stress may be more vulnerable and particularly exposed to adverse conditions. therefore, financial shocks can be characterized by their sudden and acute nature, necessitating immediate response, while financial stressors are generally ongoing issues that require long-term management and planning. more specifically for the project at hand, utilizing this theory will allow us to understand why some financial planning clients can persevere and be resilient in light of financial shocks (e.g., sudden job loss or unemployment, major medical expenses due to accident or illness, natural disasters) or stressor events (e.g., income volatility, major debt, major car/home repair, ongoing medical bills), while others experience higher rates of emotional distress or crisis. primary appraisal primary appraisal is the interpretation of the stressor event. it is the process of determining whether the stressor event is an irrelevant event, a positive event, or a dangerous event. a stressor financial services review, 33(1) 52 event can be defined as any event (real or perceived) that ignites the human stress response process. it is not the stressor event that is stressful but the bidirectional relationship between the stressor event and the primary appraisal process that determines if the stressor event will adversely affect one’s well-being (folkman, 1984). primary appraisal is shaped by personal and situational factors. personal factors impacting appraisal of a stressor include values, ideals, and goals. this is how a person assesses the “stakes” that are involved with the stressor event (folkman, 1984). for instance, compare a single individual experiencing financial stress with someone who sees themselves as a breadwinner for their family. a single individual experiencing financial stress might see it as a challenge primarily affecting their lifestyle or future plans. in contrast, someone who identifies strongly with the role of a breadwinner views their ability to provide for their family as central to their identity. therefore, financial stress for a breadwinner may feel much more consequential, impacting their sense of selfworth and responsibility. additionally, societal expectations can exacerbate the stress experienced by individuals. failing to meet financial obligations might be perceived as a failure to fulfill societal and familial expectations, further increasing their stress. situational factors that may impact the primary appraisal processes are thoughts about how likely it is that the threat will actually happen, how much damage could result from the threat, and potentially most importantly, have they ever faced a similar threat before and overcome it (e.g., personal exposure or experience with the threat) (folkman, 1984). increased personal exposure and experience is an asset that protects against financial stressors ultimately creating an increased sense of well-being (lazarus & folkman, 1986). gudmunson and danes (2011) theorized about personal exposure and experiences as part of their model of financial socialization. their model posits that our prior experiences and exposures around money (e.g., financial socialization processes) are key contributing factors to our financial attitudes, knowledge, and capabilities (gudmunson and danes, 2011). an individual may rely on their experience (socialization) in a time of financial stress. financial stress can lead families feeling vulnerable and uneasy amid job loss and insecurities with basic needs. many would agree that economic crises of this scale do not discriminate, as most families experience some form of economic stress or change in financial status because of the altering economic environment (dew et al., 2012). however, falconier (2015) suggested experiencing concerns about one’s finances is not only limited to those who are facing objective economic hardship, but it extends to any individual that perceives that his or her resources are insufficient or inadequate to meet his or her financial needs. lazarus and folkman (1986) described how the primary appraisal processes can end with one of three outcomes: threat, harm, and challenge. a threat response is the anticipation that this stressor event will harm them while the harm response is the belief that the stressor event has already damaged them. on the other hand, the challenge response is responding to the stressor event as something that will be faced and conquered. this is the resilience stress response. in our study, we incorporated a variable within the primary appraisal stage that looked at the participant’s belief that they had a fair shot at economic mobility in the hopes this would be an indicator of their interpretation of financial shocks as a challenge rather than a threat or harm. secondary appraisal the secondary appraisal process is the analysis of the available resources. individuals who have access to greater resources will feel less stress after assessing a situation compared to a person with fewer resources. the secondary appraisal is a crucial supplement to the initial assessment/appraisal, as it is the cognitive process when an individual evaluates his or her ability to take action to improve the stressful event (lazarus & folkman, 1986). for example, a widow who has been involved in the financial decision-making process and has a good relationship with a financial professional and/or lawyer, will have greater confidence in her financial security. as a result, the individual will koochel et al. 53 be more prepared to adapt during a time of financial stress, ultimately lessening the effect on their overall financial well-being. it may be obvious to state, but financial resources are going to be key in increasing clients’ ability to be resilient in light of financial shocks. of course, high incomes and the ability to raise funds quickly will make financial shocks less detrimental, but resources may go beyond just income statements. social capital theory argues that our relationships are a form of resources that can lead to economic benefits (bourdieu, 1986). in a sense, social capital theory contends that relationships and networks can serve as invaluable resources, even yielding economic benefits (hellerstein & neumark, 2020). this theory encompasses various components, including marital status and access to community services, which collectively contribute to an individual's resilience in the face of financial shocks. more specifically, this theory showed how having strong relationships in your community with your informal networks (e.g., being married and/or friends to support you) and formal networks (e.g., knowledge of and access to organizations that can help you in your community) are powerful protective and resilient factors for individuals (mancini et al., 2018). social capital theory also argues that health is a key resource within individuals. good health can enable greater social interaction and participation in economic activities, indirectly contributing to an individual's social capital. additionally, access to healthcare resources and improved well-being associated with good health can further enhance an individual's capacity to build and leverage their social networks (nieminen et al., 2013). one additional personal resource that was included in our study was self-control. selfcontrol is often a measure included in studies examining financial behaviors. early theorists in behavioral economics recognized self-control as essential in understanding why people do not act rationally when it comes to money (shefrin & thaler, 1988). self-control has been linked to retirement planning (kim et al., 2016), compulsive shopping (horváth et al., 2015), and debt (pelier et al., 2016). in a comprehensive study of self-control and financial well-being, strömbäck et al. (2017) found that self-control impacted both their respondents’ financial health (e.g., savings and behaviors), and also their emotional well-being related to money (e.g., financial anxiety and confidence). impact on financial well-being in the end, the theory of stress and coping (lazarus & folkman, 1986) attempts to understand how individuals perceive and respond to stressor events and how these processes influence well-being, in this case financial wellbeing. financial well-being can be defined as the state wherein an individual has a sense of (a) control over day-to-day and month-to-month finances; (b) the capacity to absorb a financial shock; (c) being on track to meet financial goals; and (d) ability to make financial choices to enjoy life (cfpb, 2017). this definition was based on in-depth interviews with a diverse group of consumers, contains a subjective element that reflects people’s expectations, preferences, and satisfaction with their financial situation. the stress and coping theory posits that coping is a subjective experience based on appraisals of oneself and one’s resources. numerous studies have utilized this scale to explore a wide range of topics, including consumer financial literacy, financial inclusion, materialism, personality traits, self-control, spending decisions, and financial shocks (nanda & banerjee, 2021). findings from these studies highlight the need for financial practitioners to support client’s in recognizing their self-appraisal and the resources within themselves and their networks. conceptual model and hypotheses the conceptual model for this study is based on lazarus and folkman’s (1986) stress and coping theoretical framework. evidence has shown that financial stress has been found to be directly and adversely related to financial well-being, and experiencing stressful events such as economic shocks, job security, and monetary loss have been associated with increased financial stress among individuals (choi, et al., 2020; kelley, et al., 2023). therefore, it is hypothesized that stressor events will have a negative effect on financial well-being. h1: stressor events (i.e., financial shocks) are negatively related to financial well-being. financial services review, 33(1) 54 during a stressor event, an individual may rely on their prior experiences and exposures, which are the financial socialization instances in their past that provided them with the knowledge, capabilities, and personal factors (e.g., values) to assess the “stakes” of the stressor event. early financial socialization experience has been found to be directly and positively associated with financial knowledge, and indirectly and positively associated with financial skill and subjective well-being (fan & park, 2021). therefore, it is hypothesized that increased personal exposure and experience will help protect against financial stressors, creating an increased sense of well-being. h2: increased prior experience and exposure are positively related to financial well-being. financial resources objectively increase an individual’s ability to be resilient in light of financial shocks, as high incomes and the ability to raise funds quickly may make financial shocks less detrimental; however, resources go beyond financial resources and extend to the ability to draw upon social capital. social capital, our relationships and networks, serve as invaluable resources (hellerstein & neumark, 2020). therefore, it is hypothesized that increased resources will help protect against and reduce the impact of stressors on financial well-being. h3: increased resources will significantly be positively related to financial well-being. in summary, to address the impact of financial stressor events on well-being, it is important to examine past experiences and exposures and how that influences an individual’s ability to call upon available resources to cope with the shock. to see a visual of this conceptual model please see figure 1. figure 1. lazarus and folkman (1986) stress and coping theoretical framework methods data this study used data from the cfpb’s national financial well-being survey public use file (puf), which includes the financial well-being scale. the cfpb designed the financial wellbeing scale based on extensive interviews, which emerged with four main themes including, having “control of day-to-day and month-to-month finances, having capacity to absorb a financial shock, being on track to meet financial goals, and having the freedom to make choices that allow enjoyment of life” (cfpb, 2017, p. 6) all of the measures and variables used in this study were gathered from the cfpb public use file (puf) published in september 2017. the sampling strategy for this study was designed to ensure adequate representation across populations, as well as an oversample of adults over the age of 62. the general sample (5,000 surveys) was used for this study. sixteen surveys in the general sample were 60% incomplete and therefore removed from the sample, for a total sample size of 4,984 (table 2.1). according to the cfpb (2017), the data were weighted for age, race/ethnicity, sex, education, household income, census region, home ownership status, and metropolitan area. the general sample generally stressor event appraisal of situation based on prior experience and exposure appraisal of resources available to address stressors impact of stressors on well-being coping (building experience & resources) to aid with future stressors koochel et al. 55 consisted of 2,261 males (53.4%) and 2,323 females (46.6%). the most represented groups were the 25 – 34-year-old (20.9%), white, nonhispanics (73.4%), full-time employees (45.1%), and married respondents (58.9%). variable measurement well-being. cfpb’s financial well-being scale served as the outcome variable. the measure consists of 10 items assessing an individual’s current financial situation, financial obligations, and how securely respondents feel about their financial future. respondents were asked to indicate how statements such as ability to handle a major unexpected expense, never having the things one wants in life, concern about money lasting, feeling that finances control one’s life, etc. applied to them where 1 = not at all/never and 5 = completely/always. the internal reliability is good (α = .80). stressors. we incorporated three variables to encompass stressors: household income volatility, financial shocks, and stress levels. household income volatility was coded as 1 for respondents who stated their household income varies quite a bit from one month to the next and 0 for those who responded that income was roughly the same each month with some variability throughout the year. financial shocks captured within the data included losing a job, reduced work hours, foreclosure, major car/home repair, health emergency, divorce/separation, added child to household, death of primary breadwinner, receipt of large sum beyond normal income, child started daycare/college, and providing unexpected financial support to others. respondents who did not experience any financial shock were coded as 0 versus 1 for respondents who experienced any shock. respondents were asked the degree to which they agree they had a lot of stress in their life where 1 = strongly disagree and 5 = strongly agree. prior experience/exposure. for simplicity of model interpretation, age and education were treated as continuous variables in the regression analyses since the intent was simply to capture increased experience/exposure. age was measured in the survey as 1 = 18-24 2 = 25-34, 3 = 35-44, 4 = 45-54, 5 = 55-61, 6 = 62-69, 7 = 70-74, 8 = 75 or older. education was measured in the survey as 1 = less than high school, 2 = high school degree/ged, 3 = some college/associates, 4 = bachelor’s degree, 5 = graduate/professional degree. as indicators of socialization, parent/guardian education was measured in the same five educational categories. financial socialization was measured with seven items of a dichotomous response of yes or no in regard to the house in which one was raised: discussed family financial matters, spoke about importance of savings, discussed credit, taught how to be a smart shopper, taught that actions determine success, provided allowance, and provided savings account. the seven items were summed and treated as a continuous variable in the regression analyses. belief in economic mobility captured respondents’ degree to which they agreed that everyone has a fair chance of moving up the economic ladder where 1 = strongly disagree and 7 = strongly agree. resources. once again for simplicity of model interpretation, household income was treated as continuous variables in the regression analyses since the intent was simply to capture increased resources. the ability to absorb a financial shock (i.e., one’s confidence in ability to raise $2,000 in 30 days) was captured in the survey as 1 = i am certain i could not come up with $2,000 2 = i could probably not come up with $2,000, 3 = i could probably come up with $2,000, and 4 = i am certain i could come up with the full $2,000. certainty in one’s ability to raise $2,000 was the reference category in the regression analyses. respondents who were married or living with a partner were coded 1, otherwise 0. self-reported health was coded as 1 = poor 2 = fair, 3 = good, 4 = very good, and 5 = excellent. the coding was left as continuous for purposes of this study to indicate increasing health. self-control was measured by summing three items: i often act without thinking through all the alternatives, i am good at resisting temptation, and i am able to work diligently toward long-term financial services review, 33(1) 56 goals where 1 = not at all 2 = not very well, 3 = very well, and 4 = completely well. access to resources was measured by summing two items: there are services in this area to help me and there are good work opportunities for me, if i choose to take them where 1 = strongly disagree and 5 = strongly agree. control variables as described in the literature review, gender and presence of financially dependent children was expected to impact perception of magnitude of financial shock as it would be related to the breadwinner identity. therefore, these were included as control variables. gender was coded 1 = male, 0 = female. presence of financially dependent children was coded 0, otherwise 1 if no financially-dependent children. in addition, race and ethnicity was captured within four categories of white, non-hispanic; black, non-hispanic; other, non-hispanic; hispanic due to the welldocumented racial wealth gap found in the u.s. (oliver & shapiro, 2019). white, non-hispanic was used as the reference group in the regression analyses. sample the final sample size was 4,451. approximately half of the sample was male (53%), majority reported being white and non-hispanic, and in their late 30s to early 50’s (age is categorical and the mean was 3.90 meaning in between category 3 (35-44 years old) and category 4 (45-54 years old). interestingly, the majority were married/living with their partner (66%) but many did not have financially dependent children at the time of the survey (60%). for full descriptive information see the descriptive statistics shown in table 1. see appendix 1 for a correlation matrix (only correlations of p < .05 are shown in the matrix to aid in readability). analyses the general purpose of a multiple regression is to model the relationship between two or more explanatory variables and outcome variables by fitting the linear equation to the observed data. the outcome of the multi linear regression then represents the best prediction of the dependent variable (jeger et al., 2014). therefore, to better understand the predictive relationship between financial stressors, prior financial experience and exposure, and resources and their effects on financial well-being, hierarchical regression models were computed. the first model controlled for participants gender, ethnicity, and if they had financially dependent children. the second model included prior financial experience and exposure, measured by age, education, parent’s education, financial socialization, and beliefs in economic mobility. lastly, the third model included resources, assessed by household income, the ability to absorb shocks, marriage / partner status, self-assessed health, self-control, and access to resources. financial services review, 33(1) 57 table 1. descriptive statistics (n = 4,451) variables mean sd range outcome financial well-being score 55.05 13.81 14-95 stressors volatile income 0.07 0.25 0-1 experienced financial shock(s) 0.52 0.50 0-1 stress 3.25 1.07 1-5 prior experience / exposure age category 3.90 1.95 1-8 education category 3.19 1.18 1-5 parent education category 3.04 1.23 1-5 received financial socialization 3.65 2.20 0-7 believe in economic mobility 4.68 1.67 1-7 resources household income category 5.70 2.65 1-9 ability to absorb shock i am certain i could not come up with $2k 0.13 0.34 0-1 i could probably not come up with $2k 0.08 0.26 0-1 i could probably come up with $2k 0.15 0.36 0-1 i am certain i could come up with full $2k 0.66 0.47 0-1 married / living with partner 0.66 0.47 0-1 self-assessed health 3.48 0.92 1-5 self-control 7.87 1.24 3-12 access to resources 7.08 1.878 2-10 controls male 0.53 0.50 0-1 race/ethnicity white, non-hispanic 0.74 0.44 0-1 black, non-hispanic 0.08 0.28 0-1 other, non-hispanic 0.06 0.23 0-1 hispanic 0.12 0.33 0-1 no financially dependent children 0.60 0.49 0-1 notes: we used the computed cfpb’s financial well-being scale score. please see text for categorical variables. all variables were normally distributed except for income volatility, which has a higher peak and right skewness. before dichotomizing income volatility, skewness was 1.52 and kurtosis was 1.18. financial services review, 33(1) 58 table 2. summary of hierarchical regression analysis for variables predicting financial wellbeing (n = 4,451) model 1 model 2 model 3 variable b b b controls male 0.00 -0.80* -1.30*** race/ethnicity (ref = white, non-hispanic) black, non-hispanic -3.89*** -2.109*** -0.63 other, non-hispanic -2.10** -1.39 -0.40 hispanic -3.18*** -0.60 1.01* no fin. dep. child 0.81* 0.76* 1.62*** stressors volatile income -4.30*** -2.70*** -1.66** exp. fin.shock(s) -3.06*** -2.60*** -1.31*** stress -5.48*** -4.48*** -3.19*** prior exp./exposure age 1.24*** 0.90*** education 2.49*** 0.47** parent education 0.44* 0.12 received fin. socialization 0.64*** 0.17* belief in economic mobility 1.51*** 0.74*** resources household income 0.71*** absorb shock (ref = could raise $2k) could not raise $2k -12.22*** probably could not raise $2k -9.02*** probably could raise $2k -7.21*** married/ partner 0.99** self-assessed health 1.00*** self-control 0.56*** access to resources 0.86*** r2 .24*** .36*** .53*** note: *p < .05. **p < .01. ***p < .001 results to understand the predictive relationship between financial stressors, prior financial experience and exposure, and resources impact on financial well-being, a three-model hierarchical multiple regression was conducted with financial well-being as the dependent variable. model 1, testing the control variables and financial stressors, was significant (r2 = .24, p <.001). across each model black, non-hispanic, other, non-hispanic, and hispanic maintained negative relationships with financial well-being, with the exception of hispanic later in model 3. gender was not a significant predictor of wellbeing when controlling for race/ethnicity, dependent children, and financial stressors. having no financially dependent children in the home was statistically associated with higher financial well-being (b = 0.81, p < .05). koochel et al. 59 as expected, the presence of financial stressors were among highest contributing factors to financial well-being in model 1 and throughout model 3. specifically, volatile income (b = -4.30, p < .001), experiencing a financial shock (b = 3.06, p < .001), and general stress (b = -3.18, p < .001) were all negatively associated with wellbeing. the addition of prior exposure and experience (represented by age, education, parent’s education, financial socialization, and beliefs in economic mobility) resulted in a statistically significant model 2 explaining 36% of the variance (r2 = .36, p <.001). older age (b = 1.24, p < .001), more education (b = 2.49, p < .001), greater parental educational attainment (b = 0.44, p < .001), greater financial socialization (b = 0.64, p < .001), and belief in economic mobility (b = 1.51, p < .001) were all statistically positively associated with financial well-being. gender became statistically significant in model 2 with males being correlated with negative financial well-being as compared to females (b = -0.08, p < .05). lastly, model 3 was significant with 17% higher explained variance (r2 = .53, p <.001) with the inclusion of resources in explaining financial well-being. naturally, greater availability of resources increased financial well-being above and beyond the effects of stressors and prior exposure and experience. it is important to note that greater availability of resources was not measured just by income, rather it was variables that assessed other forms of capital. income from households was positively associated with higher financial well-being (b = 0.71, p < .001) and the lack of ability to absorb a financial stock of $2,000 was the largest overall contributor to financial well-being with betas ranging from 7.21 to -12.22 (p < .001). further, individual qualities such as having a partner within the home (b = 0.99, p < .001), being in good health (b = 1.00, p < .001), selfcontrol (b = 0.56, p < .001), and having people to turn to for resources, if needed (b = 0.86, p < .001), were all statistically significant in predicting higher financial well-being. limitations the results should be interpreted with caution, as populations different from the sample used in this study may have different experiences with financial shocks and well-being. additionally, while the cfpb (2017) financial well-being scale operationalized financial well-being as a multidimensional construct, measures and definitions of financial well-being remain consistent across the literature. this study aimed to broaden our understanding of the multidimensional nature of financial well-being. despite these efforts, we must acknowledge the limitations of our study, particularly regarding the diversity of experiences across different populations. these variations highlight the need for further research to ensure that our understanding and measurement of financial well-being are inclusive and accurately reflect the experiences of diverse groups that differ from the sample used in this study may have different experiences to financial shocks and well-being. finally, longitudinal data would offer insights into the current study’s limitation of crosssectional data that cannot truly assess if appraisals preceded behavior. discussion in this study, lazarus and folkman's (1986) stress and coping theory offered a comprehensive framework for understanding individual reactions to financial stress. the theory's core premise is that the way individuals appraise stressors and the resources they have at their disposal significantly influences their ability to cope with financial shocks. this appraisal is influenced by personal factors, such as values and goals, and situational factors, such as past experiences with financial difficulties, as well as financial assets, social capital, and personal attributes like self-control and self-efficacy. to understand the predictive relationship between financial stressors, prior financial experience and exposure, and resources’ impact on financial well-being, a three-model hierarchical multiple regression was conducted with financial well-being as the dependent variable. our hypotheses that individuals experiencing a stressor event or financial shock would subconsciously engage in a primary appraisal of their past experiences and exposures to finances to determine if the financial shock was truly a threat, and that additional resources would financial services review, 33(1) 60 be considered during a secondary appraisal process if the event was identified as a threat was supported by the results. empirical results from the hierarchical multiple regression analysis underscore the importance of these appraisal processes in shaping financial well-being. model 1 included control variables (race/ethnicity, gender, and dependent children) and financial stressors, and was significant (r2 = .24, p <.001). the analysis revealed significant disparities in financial well-being among different demographic groups, with black, nonhispanic, other, non-hispanic, and hispanic individuals showing negative relationships with financial well-being, except for hispanic individuals later in model 3. unfortunately, the financial literacy racial/ethnic gap among americans is well documented (e.g., al-bahrani et al., 2019; lee & kim, 2022; kim et al., 2011). previous research reveals that black and hispanic individuals are more likely to fall within the lower half of the income distribution, report diminished financial health, and exhibit lower financial literacy scores (al-bahrani et al., 2019). these disparities are associated with reduced financial well-being (kim et al., 2011). this disparity is also observed in differences by socioeconomic status, children from higher socioeconomic backgrounds are exposed to better experiential learning of finances (e.g., financial socialization). these inequalities are often observed by differences in access to opportunities (e.g., social mobility) and available financial resources (explored in model 3). the presence of dependent children was included as a control due to the increased financial responsibility associated with having children (sun et al., 2022). race and gender were included as controls as research has found differences in financial socialization due to these factors (e.g., cameron-agnew, 2015; kim et al., 2011) research exploring financial socialization and race found evidence that there are systematic factors of financial socialization that disproportionately disadvantaged people who are non-white (gutter et al., 2010). notably, white children and emerging adults have been found to receive more direct parent-child financial discussion (e.g., explicit financial socialization) and are more likely to have savings accounts than their non-white peers (kim et al., 2011). regarding gender, there have been mixed results for differences between genders and financial socialization. interestingly in our study males had an increasingly significantly negative relationship with financial well-being when controlling prior experience and exposure and resources. cameron-agnew (2015) suggested that parents seem to engage in more direct financial conversations (i.e., explicit socialization) with males younger than females. while tang et al. (2015) found parental influence improves females’ financial behavior more than men. yet, more recently agnew et al. (2019) did not find any gender differences in rates of financial socialization. falahati and colleagues (2015) also looked at determinants of financial well-being, examining gender, and found that among asian college students males and females perceive different levels of financial strain, and that financial management is the strongest predictor of well-being for males, while among females it is financial knowledge and literacy. furthermore, their findings revealed that peers, media, and other socializing agents had a positive effect on financial strain among men while females were more affected by the negative effects of financial attitudes in managing their finances. these equivocal secondary effects may further reveal the influence of additional socialization agents impacting financial wellbeing and financial strain between genders. the results of this study are interpreted to mean that gender differences in well-being are better explained, in part, by increased experience and access to resources. these findings are consistent with lazarus and folkman’s (1986) stress and coping theory, which highlights how both personal and situational factors shape the primary appraisal of financial stressors. disparities in financial wellbeing among different racial and ethnic groups, as well as those with different socioeconomic statuses, highlight the role of personal and situational factors in shaping how financial stressors are appraised and managed. children from higher socioeconomic backgrounds, for example, may be exposed to better experiential learning of finances (e.g., financial socialization), koochel et al. 61 which underscores the theory’s assertion that past experiences and exposures play a crucial role in the primary appraisal process. these inequalities are often observed by differences in access to opportunities (e.g., social mobility) and available financial resources, further explored in subsequent models. these findings lead to our next model, which incorporates prior exposure and experience regarding financial socialization. model 2, which added variables such as age, education, parent’s education, financial socialization, and beliefs in economic mobility, was statistically significant, explaining 36% of the variance in financial wellbeing (r² = .36, p <.001). this suggests that financial socialization is a critical factor in building resilience and reducing overall stress related to financial well-being. much of the learning attributed to socialization occurs through behavior modeling and the implicit transfer of information observed in one’s environment. parents who explicitly and positively influence their children’s financial knowledge and skills have been found to enhance their children’s future financial competencies, such as effective money management (van campenhout, 2015). additionally, while parental influence is crucial during childhood, financial socialization continues throughout the lifespan, with age and education also showing statistically significant positive effects on well-being. model 2’s findings support the theory's notion that past experiences shape how financial stressors are appraised, reinforcing the role of financial socialization. individuals with positive financial experiences and effective financial socialization during their upbringing demonstrated greater resilience in the face of financial stress. the inclusion of resources in model 3 further underscored the critical role of secondary appraisal. access to financial resources, strong social networks, and personal traits like self-control significantly enhanced financial well-being, highlighting the importance of evaluating and mobilizing resources to cope with financial stress. this comprehensive approach, integrating primary and secondary appraisal processes, underscores the multifaceted nature of financial resilience and well-being. lastly, model 3 was also significant (r2 = .53, p <.001) based on resources. with the addition of resources in model 3, models 1 and 2 remained significant. naturally, a greater availability of resources increased financial well-being beyond the effects of stressors and prior exposure and experience. however, it is important to note that the greater availability of resources was not measured solely by income but also by other forms of capital. research has shown that having strong relationships within your community, including informal networks (e.g., a partner or friends to support you) and formal networks (e.g., knowledge of and access to helpful organizations in your community), are powerful protective and resilient factors for individuals (mancini et al., 2018). this web of support appears to be a strong indicator of resilience in light of financial stressors. furthermore, individual qualities such as selfcontrol (e.g., the ability to resist urges, selfregulation) and self-efficacy have been noted as key attributes to social mobility and are further associated with better financial behavior and financial well-being (lind et al., 2020; stromback et al., 2017). dare et al. (2022) found that financial self-efficacy was positively related to positive financial behaviors and financial wellbeing; relatedly our findings confirmed a statistically significant relationship between selfcontrol and financial well-being. interestingly, there was not a significant relationship between parents’ education and financial well-being, but perhaps individuals who are married are more likely to turn to their partner as a resource and rely more on their own financial household income and resources. further, having no financially dependent children was found to remain statistically significant across all three models. the findings indicate that financial well-being is a multifaceted construct influenced by a combination of demographic factors, financial socialization, and resource availability. persistent disparities observed across race and gender suggest that systemic issues, such as unequal access to financial education and resources, continue to impact financial well-being. moreover, the significant role of financial financial services review, 33(1) 62 socialization highlights the importance of early financial education and positive parental influence in building financial resilience. access to resources, both financial and social, emerged as a crucial factor in mitigating the impact of financial stress, underscoring the need for policies and programs that enhance resource availability and support networks. combining the insights from the stress and coping theory with these empirical findings provides a deeper understanding of financial resilience. it underscores the importance of addressing systemic disparities in financial education and resource access, promoting early financial socialization, and fostering strong social support networks. by doing so, we can enhance individuals' ability to cope with financial stress, ultimately improving their overall financial wellbeing. this integrated approach not only aids researchers in identifying protective factors that increase resilience but also informs policymakers and practitioners in developing targeted interventions to support vulnerable populations in navigating financial challenges. given the known associations between financial skills and financial well-being (haynes-bordas et al., 2008), the findings in this study are consistent with previous research: maintaining financial well-being is about more than knowledge and skill. the individual’s beliefs and behaviors about money also play a crucial role in sustaining financial well-being during times of financial stress. as the current findings suggest, financial socialization matters. the key drivers of individuals' financial beliefs significantly impact their financial well-being and may serve as critical buffers during times of financial stress. with explicit talk of socialization, individuals are better prepared to adapt to financial shocks, ultimately lessening the effect of the shock on their overall financial well-being. practical implications findings of this study provide support for financial counselors and planners to engage in more holistic and personalized financial conversations. as financial professionals engage with clients about their financial past, the client may discover patterns and behaviors that were previously unknown to them (positive or negative). such discoveries may serve as motivators for the client, thus encouraging the client to further invest in their skills and selfefficacy. in addition, these conversations regarding clients’ money stories may enrich the planner-client relationship and improve trust, commitment, and retention (e.g., britt, 2016; kahler, 2012; sharpe et al., 2007). for example, financial professionals can ask questions similar to ones developed by mumford and weeks (2003) such as: ● what is your earliest money memory? ● what was your parents’ role around money in your household? is your role in your relationship similar or different? ● as a kid, did you see yourself as rich or poor and how accurate were those perceptions? ● in what ways did your parents align with their financial beliefs, and where were some areas of disagreement? there is an important caveat to asking questions that aid clients in exploring their financial socialization. lurtz (2022) described how sensitive and emotionally laden these conversations can be for some clients. lurtz described how important it is to have a strong and trusting relationship with your client and foreshadow these conversations. the goal is to facilitate their own self-exploration and experiences to discover their internal and external resources that can aid in future stressful situations, not to cause more stress for the client. so, although a financial professional does not need to become a mental health professional to ask these types of questions, it can also be beneficial to seek out training in client psychology to ensure you have the skills to navigate these conversations and the potential emotions that may arise as a result. grubman et al. (2023) described wealth 3.0, in part, as the next evolution of financial planning focused on more collaborative interaction among professions and cross-discipline training. with enhanced awareness and comfort in talking across professions, financial planners can make mental health referrals as seamlessly as they make tax or estate referrals. koochel et al. 63 when stress arises, we tend to focus on problem solving. we want to face the stressor head on and create a game plan to resolve financial problems. this is an important coping strategy and will benefit many of our clients (e.g., zhang et al., 2019). however, lazarus and folkman (1986) theory of stress and coping and the results of our study suggest that sometimes just focusing on the numbers will not be enough. that is why financial counselors and planners need to be aware of a second type of coping, emotion-focused coping. emotion focused coping focuses on regulating your feelings and emotions to the problem. emotion-focused coping is advantageous when you can’t solve the problem right away or it is outside your control to change. emotion-focused coping strategies are stress management techniques such as talking about your stressors with someone you trust, journaling, talk therapy, and mindfulness. these can be a powerful tool to supplement problem solving. in essence, the financial professional’s job is to help clients use problem solving strategies to address problems they can control. however, it is also important for professionals to help clients address underlying thoughts and emotions that are arising for them during a crisis situation. the cfp board (2022) recently added the psychology of financial planning to their educational and exam requirements and released a book on how financial planners can improve their understanding of psychological tools and techniques to address underlying thoughts, beliefs, and biases especially during crises. that said, financial planners are trained to be financial professionals rather than mental health professionals. in addition to focusing on acquiring skills to aid in emotional coping (in general and considering a financial shock), it is also important to create a referral network to mental health professionals and financial therapists that can provide additional support to your clients. individuals with more positive financial behaviors are associated with more stable financial situations, which is consequently associated with an improved sense of well-being. improving clients’ skills and sense of confidence is a pathway toward happier clients. references agnew, s., maras, p., & moon, a. (2018). gender differences in financial socialization in the home—an exploratory study. international journal of consumer studies, 42(3), 275-282. apostolakis, g. n., floros, c., gkillas, k., & wohar, m. (2021). political uncertainty, covid-19 pandemic and stock market volatility transmission. journal of international financial markets, institutions and money, 74, 101383. bourdieu, p. (1986). the forms of capital. in j.g. richardson (ed.), handbook of theory and research for the sociology of education (pp. 241-258). greenwood. byram, j. l., mccoy, m., kruger, m., & grable, j. (2023). financial planning counseling skills. national underwriter company. choi, s. l., heo, w., cho, s. h., & lee, p. (2020). the links between job insecurity, financial well‐being and financial stress: a moderated mediation model. international journal of consumer studies, 44(4), 353–360. cfp board. (2022). cfp board adds 'psychology of financial planning' to exam requirements. https://www.cfp.net/news/2021/03/cfpboard-adds-psychology-of-financialplanning-to-exam-requirements. consumer financial protection bureau (cfpb). (2017). financial well-being in america. https://www.consumerfinance.gov/dataresearch/research-reports/financial-wellbeing-america/ dunsmir, l. (2022). no inflation relief in sight for u.s. as impact of ukraine war intensifies. reuters. https://www.reuters.com/markets/europe /us-inflation-set-heat-up-further-impactukraine-war-intensifies-2022-03-07/. falahati, l., & sabri, m. f. (2015). an exploratory study of personal financial wellbeing determinants: examining the financial services review, 33(1) 64 moderating effect of gender. asian social science, 11(4), 33. fan, l., & park, n. (2021). factors mediating the association between financial socialization and well-being of young adults: testing a conceptual framework. journal of financial counseling and planning, 32(2), 202– 216. folkman, s. (1984). personal control and stress and coping processes: a theoretical analysis. journal of personality and social psychology, 46(4), 839-852. fox, j., & bartholomae, s. (2020). household finances, financial planning, and covid‐19. financial planning review, 3(4), e1103. grubman, j., jaffe, d. t., & keffeler, k. (2023). wealth 3.0: the future of family wealth advising. family wealth consulting. hellerstein, j. k., & neumark, d. (2020). social capital, networks, and economic wellbeing. the future of children, 30(1), 127-152. horváth, c., büttner, o. b., belei, o. b., & adıgüzel, f. (2015). balancing the balance: self-control mechanisms and compulsive buying. journal of economic psychology, 49, 120-132. kelley, h. h., lee, y., lebaron-black, a., dollahite, d. c., james, s., marks, l. d., & hall, t. (2023). change in financial stress and relational wellbeing during covid-19: exacerbating and alleviating influences. journal of family & economic issues, 44(1), 34–52. kim, k. t., lee, j. m., & hong, e. (2016). the role of self-control on retirement preparedness of us households. international journal of human ecology, 17(2), 31-42. lazarus, r. s. & folkman, s. (1986). cognitive theories of stress and the issue of circularity. in dynamics of stress. springer. lee, s. t., & kim, k. t. (2022). a decomposition analysis of racial/ethnic differences in financial knowledge and overconfidence. journal of family and economic issues, 43(4), 815-831. lurtz, m. (2019). finding your own money story to better communicate with clients. https://www.kitces.com/blog/margaretatwood-payback-debt-book-moneyscript-egg-narrative/ mancini, j. a., o'neal, c. w., martin, j. a., & bowen, g. l. (2018). community social organization and military families: theoretical perspectives on transitions, contexts, and resilience. journal of family theory & review, 10(3), 550565. mccoy, m., machiz, i., harris, j., lynn, c., lawson, d., & rollins-koons, a. (2022). the science of building trust and commitment in financial planning: using structural equation modeling to examine antecedents to trust and commitment. journal of financial planning, 35(12), 68-89. mumford, d. j. & weeks, g. r. (2003). the money genogram. journal of family psychotherapy 14(3), 33-44. nanda, a. p., & banerjee, r. (2021). consumer’s subjective financial well‐being: a systematic review and research agenda. international journal of consumer studies, 45(4), 750-776. nieminen, t., prättälä, r., martelin, t., härkänen, t., hyyppä, m. t., alanen, e., & koskinen, s. (2013). social capital, health behaviours and health: a population-based associational study. bmc public health, 13(1), 1-11. oliver, m. l., & shapiro, t. m. (2019). disrupting the racial wealth gap. contexts, 18(1), 16-21. oxford analytica. (2022). financial market volatility will rise as risks pile up. emerald expert briefings. https://doi.org/10.1108/oxandb267655 https://www.kitces.com/blog/margaret-atwood-payback-debt-book-money-script-egg-narrative/ https://www.kitces.com/blog/margaret-atwood-payback-debt-book-money-script-egg-narrative/ https://www.kitces.com/blog/margaret-atwood-payback-debt-book-money-script-egg-narrative/ koochel et al. 65 ramelli, s. & wagner, a. (2020). what the stock market tells us about the consequences of covid-19. mitigating the covid economic crisis: act fast and do whatever. working paper. shefrin, h. m., & thaler, r. h. (1988). the behavioral life‐cycle hypothesis. economic inquiry 26(4), 609-643. strömbäck, c., lind, t., skagerlund, k., västfjäll, d., & tinghög, g. (2017). does self-control predict financial behavior and financial well-being? journal of behavioral and experimental finance, 14, 30-38. sun, l., small, g., huang, y. h., & ger, t. b. (2022). financial shocks, financial stress and financial resilience of australian households during covid-19. sustainability, 14(7), 3736. tang, n., baker, a., & peter, p. c. (2015). investigating the disconnect between financial knowledge and behavior: the role of parental influence and psychological characteristics in responsible financial behaviors among young adults. journal of consumer affairs, 49(2), 376-406. van campenhout, g. (2015). revaluing the role of parents as financial socialization agents in youth financial literacy programs. journal of consumer affairs, 49(1), 186-222. zhang, y., zhang, y., ng, t. w., & lam, s. s. (2019). promotion-and preventionfocused coping: a meta-analytic examination of regulatory strategies in the work stress process. journal of applied psychology, 104(10), 1296. financial services review, 33(1) 66 appendix a table a1. correlations for study variables, income volatility, shocks, stress, age, education, parent education, financial socialization, mobility, household income, absorb shock, married, health, self-control, access to resources, and financial wellbeing (n = 4,451) volatile inc. shocks stress age educ. parent edu. fin. soc. mobil ity hh inc. absorb shock married healt h contro l access vol. inc. shocks -0.09 stress 0.08 -0.17 age -0.06 0.09 -0.27 education -0.10 parent ed. -0.07 -0.15 0.67 fin. soc. -0.07 0.04 -0.05 -0.14 0.27 0.29 mobility 0.07 -0.16 0.05 -0.07 -0.06 0.11 hh inc. -0.12 0.08 -0.08 0.04 0.51 0.39 0.22 0.08 ab. shock -0.05 0.13 -0.18 0.13 0.13 0.10 0.10 0.11 0.22 married -0.04 -0.08 0.13 0.13 0.05 0.07 0.29 0.07 health -0.05 0.11 -0.22 0.13 0.26 0.24 0.23 0.14 0.29 0.14 0.08 control -0.03 0.04 -0.09 -0.04 0.04 0.05 0.13 0.15 0.06 0.09 0.05 0.22 access -0.05 0.04 -0.12 -0.13 0.20 0.17 0.24 0.25 0.20 0.12 0.04 0.26 0.20 fin. well -0.13 0.20 -0.45 0.27 0.26 0.17 0.19 0.25 0.39 0.36 0.17 0.30 0.18 0.30 only correlations of p <.05 are shown. pii: 1057-0810(91)90030-3 financial services review, i(2): 143-157 copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserwl. probabilistic estate planning ronald r. crabb probabilistic estate planning is based on the principle of maximizing expected net present value commensurate with the riskassumed. rather than assuming that death occurs at life expectancy, probabilistic estate planning treats death as a random variable. compounded to randomly chosen ages of death, estate assets are taxed and distributed to heirs. the purpose ofprobabilis tic estate planning is to find the estate plan and asset/liability combination that maximizes the expected net present value of assets passing to heirs and to convey some idea of the risk associated with that estate plan. “in america, it was the tax statutes of the 19th and 20th centuries which acted as the catalysts for the tremendous interest in estate planning” (ackerman, 1973). prior to the existence of those taxes, estate planning, while not unnecessary, was uncomplicated. assets were simply passed by will or by gift from one generation to another. incorporated farms, complex trusts, and other legal fictions were unneces sary. as the burden of estate, gift, and inheritance taxes and probate fees increased, estate planning and estate planners became more sophisticated. “for over 25 years estate planning concepts have been ‘sold’ to clients. typically, the clients want to concentrate on one numerical result-the amount of tax dollars saved” (miller, 1987). in his article on “selling estate planning with a computer screen and without reams of paper,” ralph gano miller (1987) describes a computer screen approach to explaining estate planning to his clients. assumptions are made regarding the date of death of the first decedent and the number of years until the remaining spouse dies. “based on statistics it is generally concluded that the husband will die first. ” a growth rate for the estate is assumed, and the computer does the tax calculations to determine how much the heir(s) will receive after the first decedent and the remaining spouse have died. various estate planning scenarios are depicted (all to spouse, use of different trusts, gifting of ronald r. crabb l assistant professor of finance, university of wisconsin-whitewater. whitewater, wi 53 190. 144 financial services review, i(2) 1991 assets), and the computer screens “show” the clients which type of estate plan, given the assumptions made, is “best” for their situation. whether done on a computer screen or handed to the client in a twenty pound bound paper volume, the principles underlying current estate planning are similar: choose an age of death for the husband (wife), choose a survival period for the wife (husband), choose a growth rate for the estate, analyze alternate estate plans, and determine, given the assumptions made, which estate plan minimizes the tax liabil ity for the heir(s). probabilistic estate planning introduces risk into this process. rather than assuming a specific date of death for the husband (wife) and then a survivorship period for the widow (widower), ages of death for both spouses are treated as random variables. the process involves the use of a mortality table and a random number generator. the methodology follows. methodologyunderlying probabilistic estate planning the 1984 u.s. life table states that of an original population of 100,000 females age 0, at age 50 those still alive number 95,26 1. similarly, for 100,000 age 0 males, 9 1,207 remain alive at age 50 (national center for health statistics, 1987). a portion of the mortality table is shown in table 1 (see columns 1, 2, 3, 6, 7, 8). by dividing the number of people who survive to age 5 1. . .62. . .73. . .84. . . by the number alive at age 50, the resulting quotients compute the probability of survival to age 5 1 . . .62.. .73.. .84. those quotients, one for a male age 50 and one for a female age 50, underlie the random age death selection process (see columns 5 and 10 in table 1). a computer generates two random numbers (with a uniform distribution on the unit interval). assume that the computer generates 0.676 for the male and 0.834 for the female. in table 1, male data, column 5, the number 0.676 lies between 0.70040 and 0.67325, indicating that the male has attained age 70 when he dies; the number 0.834 indicates that the female has attained age 69 when she dies. the computer also notes that the female, in this trial, dies first. a graph of the data (see figure 1) in columns 5 and 10 versus age shows that, initially, the survival curves are decreasing at an increasing rate, and, as the popula tion of persons who had attained age 50 shrinks, decreasing at a decreasing rate (note: the graphs are extended beyond the data shown in columns 5 and 10 through the end of the 1984 u.s. life mortality table). since the random numbers are generated on the unit interval, they must fall on the vertical axes of the graphs in figure 1. starting from 0.676 (on the male curve) and 0.834 (on the female curve), go across each graph horizontally to the point where the random number generated intersects the survival curve; drop down probabilistic estate planning 145 table 1. 1984 u.s. life mortality data male data female data probability probability probability probability number number of of survival number number of of survival age alive dying dying* to next age age alive dying dying * to next age t11 121 131 [41 151 [61 171 [81 191 [loi 50 91207 570 0.00625 0.99375 50 95261 330 0.00346 0.99654 51 90637 628 0.00689 0.98687 51 94931 364 0.00382 0.99271 52 90009 692 0.00759 0.97928 52 94567 400 0.00420 0.98852 53 89317 762 0.00835 0.97092 53 94167 437 0.00459 0.98393 54 88555 837 0.00918 0.96175 54 93730 478 0.00502 0.97891 55 87718 916 0.01004 0.95170 55 93252 521 0.00547 0.97344 56 86802 999 0.01095 0.94075 56 92731 567 0.00595 0.96749 57 85803 1085 0.01190 0.92885 57 92164 617 0.00648 0.96101 58 84718 1176 0.01289 0.91596 58 91547 672 0.00705 0.95396 59 83542 1270 0.01392 0.90204 59 90875 731 0.00767 0.94628 60 82272 1367 0.01499 0.88705 60 90144 795 0.00835 0.93794 61 80905 1467 0.01608 0.87096 61 89349 861 0.00904 0.92890 62 79438 1569 0.01720 0.85376 62 88488 931 0.00977 0.91913 63 77869 1672 0.01833 0.83543 63 87557 1002 0.01052 0.90861 64 76197 1777 0.01948 0.81595 64 86555 1076 0.01130 0.8973 1 65 74420 1881 0.02062 0.79532 65 85479 1154 0.01211 0.88520 66 72539 1986 0.02177 0.77355 66 84325 1237 0.01299 0.87221 67 70553 2099 0.02301 0.75053 67 83088 1327 0.01393 0.85828 68 68454 2222 0.02436 0.72617 68 81761 1425 0.01496 0.84333 69 66232 2351 0.02578 0.70040 69 80336 1531 0.01607 0.82725 70 63881 2476 0.02715 0.67325 70 78805 1641 0.01723 0.81003 71 61405 2595 0.02845 0.64480 71 77164 1756 0.01843 0.79159 72 58810 2706 0.02967 0.61513 72 75408 1878 0.01971 0.77188 73 56104 2808 0.03079 0.58434 73 73530 2005 0.02 105 0.75083 74 53296 2898 0.03 177 0.55257 74 71525 2138 0.02244 0.72839 75 50398 2977 0.03264 0.51993 75 69387 2275 0.02388 0.70451 76 47421 3041 0.03334 0.48659 76 67112 2417 0.02537 0.67913 77 44380 3089 0.03387 0.45272 77 64695 2562 0.02689 0.65224 78 41291 3119 0.03420 0.41852 78 62133 2709 0.02844 0.62380 79 38172 3128 0.03430 0.38422 79 594242860 0.03002 0.59378 80 35044 3116 0.03416 0.35006 80 56564 3014 0.03164 0.56214 81 31928 3079 0.03376 0.31630 81 53550 3171 0.03329 0.52885 82 28849 3015 0.03306 0.28325 82 50379 3330 0.03496 0.49390 83 25834 2921 0.03203 0.25122 83 47049 3493 0.03667 0.45723 84 22913 2797 0.03067 0.22055 84 43556 3658 0.03840 0.41883 *given survival until age 50. vertically to the male age axis or female age axis to read the attained age of the male and the female at death. for all ages there exist curves similar to those shown in figure 1. the vertical axis, computed by dividing the number of survivors to some future age by the 146 financial services review, l(2) 1991 1.00 0.95 0.90 0.85 0.80 0.75 probability 0.70 of survival 0.65 tosxne 0.60 future age, 0.55 given survival 0.50 toacje5o 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0. io 0.05 0.00 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 50 55 60 65 70 75 80 85 90 95 100 105 110 mple age 50 55 60 65 70 75 80 85 90 95 100 105 i io femne age figure 1. probability of survival curves. probabilistic estate planning 147 attained age of the person who is interested in probabilistic estate planning, is graphed versus attained age. for a person age 65, the series of points computed by dividing the number of survivors to age 66 . . .75. . .84. . . to the end of the mortality table would be graphed versus age 66 . . .75. . .84. . . to the end of the mortality table. the graph would be similar in shape to that shown in figure 1, but the slope of the survival curve would be steeper. the older the person, the steeper the slope of the survival curve; the steeper the slope of the curve, the shorter the horizontal axis; in the limiting case (the last year in the mortality table) the survival curve would be vertical, and no matter what random number was generated, death would be predicted to occur in the upcoming year. (the only distribution assumption required for probabilistic estate planning is that time until death has a distribution defined by the 1984 u.s. life table. it should be noted that this distribution is not normal.) for any pair of random numbers, three possible outcomes exist. the male dies first, the female dies first, or both the male and female die in the same future year. for example, if the computer generated 0.825 as the male random number and 0.901 as the female random number, the data in table 1 (or the graph in figure 1) would indicate that both the 50 year old male and the 50 year old female would die in their 63rd year of life. if the computer generated 0.585 for the male and 0.590 for the female, the male would be predicted to die in his 73rd year of life, the female in her 80th year of life. in an earlier example the female died first, the male second. assuming a 100 trial simulation, the computer generates two hundred random numbers, one hundred to predict the age of the male at death and one hundred to predict the age of the female at death. sometimes the male will die first, sometimes the female will die first, and sometimes death will occur at the same age. the 200 random numbers result in 100 combinations of death for which estate taxation and distribution can be computed. estate growth and taxation are straightforward processes. assume some growth rate for the assets comprising the estate. allow those assets to grow until the death of the first person. tax the estate. allow the remaining estate assets to grow until the death of the second person. tax the estate. finally, compute the present value of the net estate passing to the heir(s). this process will yield 100 net-to-the heir(s) numbers, and a variance associated with those numbers. different estate planning techniques and/or asset/liability mixtures may produce different expected net present values and variances. theadvantagesofaprobabilisticapproach perhaps the most common assumptions in family estate planning are that the husband dies first, that the husband dies at his life expectancy, and that the surviving spouse dies at her life expectancy. according to the 1984 u.s. life table, at age 60 life expectancy for a female is 23 years, 18 years for a male (rounded to the nearest 148 financial services review, l(2) 1991 integer). however, the probability that a female age 60 dies in her 83rd year of life is 0.0387 and the probability that the male age 60 dies in his 78th year of life is 0.0379 [female probability = (3493/90144) and male probability = (3119/82272)]. few people die in the year of their life expectancy. table 2, based on the 1984 u.s. life table, shows the probability of dying in a seven-year period centered on life expectancy. for example, a male age 50 has a life expectancy of 26 years. the probability that the male dies in the seven-year period between 73 and 79 is only 23%. [note: sum the numbers 0.03079 + 0.03177 + 0.03264 + 0.03334 + 0.03387 + 0.03420 + 0.03430 in column (4) in table 1 to compute the seven-year death probability; round the result of 0.23091 to 23 % .] similarly, the age 50 female expectation of death in a seven-year period centered on life expectancy is also 23 % . data for other ages for males and females show that most people do not die anywhere close to the age computed by adding their life expectancy to their attained age. note that the data in table 1 can not be used to compute seven year death probabilities for any age except age 50. to compute the other probabilities shown in table 2, interested readers can create six tables similar to table 1, with columns 4 and 9 modified by dividing the number of persons dying in a given year by the number of persons who survived to ages 53, 56, 60, 63, 66, and 70. since the number surviving decreases as age goes from 53 to 70, but the number dying at a given age (columns 3 and 8) stays constant, the probability of dying (columns 4 and 9) in a given year increases as age goes from age 53 to 70, and the sum of a seven year series (of the one year death probabilities) increases. yet even at age 70, only one of every three decedents will die within a seven year period centered on life expect ancy. computations based on death at life expectancy are clearly inconsistent with table 2. death in a seven year period centered on life expectancy age male expectation of life * probability of death in a seven year period** female expectation of life* probability of death in a seven year period** 50 26 23% 31 23% 53 23 24% 28 24% 56 21 25% 26 26% 60 18 26% 23 29% 63 16 28% 20 30% 66 14 30% 18 33% 70 12 33% 15 36% *rounded to nearest integer; **rounded to nearest percent. probabilistic estate planning 149 reality and computations based on death near life expectancy are little better; and, contrary to popular belief, actuaries do not base their calculations on life expectan cies. “it is popularly believed that the expectation of life is widely used in actuarial calculations. in reality, it is of interest to actuaries only because it affords an index for comparing different mortality tables” (jordan, 1952). in the case of life annuities, “one of the persistent misconceptions is that the present value of a life annuity at agex is equal to the value of an annuity-certain for a term equal to the life expectancy at age x. . . . the annuity-certain for the term of the life expectancy always exceeds the life annuity value” (jordan, 1952). the previous statement is based on jensen’s inequality. when jensen’s inequality is applied to estate planning, three possible outcomes are possible. if the growth rates of the estate assets exceed the discount rate, then the mean of the expected value(s) will exceed the expected value of the mean(s). that is, the probabilistic mean will be larger than the traditional life expectancy point estimate. if the growth rates of the estate assets equal the discount rate, then the mean of the expected value(s) will be equal to the expected value of the mean(s), and the probabilistic mean will equal the traditional point estimate. if the discount rate exceeds the growth rates of the estate assets, then the mean of the expected value(s) will be less than the expected value of the mean(s), and the traditional point estimate will be larger than the probabilistic mean. in only one case (a highly unlikely case) is the traditional point estimate equal to the probabilistic mean. the magnitude of the error associated with the traditional estimate is directly proportional to the magnitude of the difference between the growth rates of the estate assets and the discount rate and directly proportional to the remaining lifetime(s) of the estate owner(s). the probabilistic estate planning mean is actuarially sound, and it presents a clearer picture of reality than the traditional life expectancy point estimate. even when spouses are the same age, and, as everyone knows, females have a larger life expectancy than males of the same age, there is a significant chance that a wife will predecease her husband. table 3 is based on a simulation, where each cell table 3. probability that wife predeceases husband age of husband 50 53 56 age of wife 60 63 66 70 50 39% 43% 53% 61% 66% 73% 79% 53 30% 33% 46% 57% 61% 66% 75% 56 23% 30% 37% 49% 53% 62% 70% 60 18% 21% 27% 37% 44% 52% 60% 63 14% 20% 23% 28% 36% 44% 51% 66 11% 15% 13% 23% 30% 38% 47% 70 8% 9% 12% 17% 22% 29% 38% 1.50 financial services review, l(2) 1991 is the result of choosing 1000 random ages of death for men agex and women age y, and x and y vary from 50 to 70. holding the age of men and women equal (see the equal age diagonal of table 3), the chance that a woman predeceases her husband varies from 39 % to 33 % , with a mean of 37 % . for a woman six years older than her husband that probability increases to a mean of 53%. for a woman six years younger, it decreases to mean of 23 % . estate planning based on death at life expectancy (or even near life expectancy) is not realistic. estate planning based on the assumption that the male dies first is often incorrect. probabilistic estate planning, using realistic mortality rates, allows an estate planner to choose a mathematically optimal plan. since the computations are being made net of taxes (or net of costs), and since the computations are being made on a present value basis, the probabilistic estate planning methodology is, in part, essentially an application of the basic financial principle “maximize expected net present value. ” simplicity is the primary value of life expectancy estate planning. usually expressed on a future value basis, life expectancy estate planning simultaneously overstates the value of the tax savings, overstates the size of the inheritance of the heir(s), and ignores the riskiness associated with the estate plan. while a present value analysis (of the future taxes saved or the size of the inheritance) could be easily incorporated into traditional estate planning, variance based risk analysis cannot, and to ignore variance based riskiness is to implicitly ignore markowitz. in his classic article on “portfolio selection,” markowitz (1952) stated “we saw that the expected returns or anticipated returns rule is inadequate. let us now consider the expected returns-variance of returns (e-v) rule.” choosing an estate plan simply because it has the largest tax savings (like choosing a security with the highest expected return) without investigating the variance associated with that estate plan (security) is inconsistent with modern portfolio theory. as markowitz (1952) closed his classic article, he stated “i believe that what is needed is essentially a ‘probabilistic’ reformulation of security analysis. i will not pursue this subject here, for this is ‘another story. ’ it is a story of which i have read only the first page of the first chapter.” as defined in this paper, probabilistic estate planning permits modern portfolio theory (mean, variance tradeoffs) to be used in selecting an optimal estate plan, and presents, in this author’s opinion, the first page in the first chapter of the reformulation of traditional estate planning. thedisadvantagesofprobabilisticestateplanning the primary disadvantages of this approach are two: computational complex ity and the choice of the appropriate discount rate for computing present values. while the logic underlying the computations demanded by probabilistic estate planning is not particularly complex, the computer program required to execute that logic is both complex and lengthy. for the simple two estate plan analysis (status quo and a single change from status quo) that was presented at the second annual probabilistic estate planning 151 meeting of the academy of financial services in new orleans in the fall of 1988, 1000 lines of basic code were required. for the analysis which follows this section, 500 more lines of code were added. the problem of computational complexity pales when compared to the choos ing of the appropriate discount rate. an expenditure of hard work can solve the programming problem; the conceptual problems underlying the choice of the appropriate discount rate could keep a financial philosopher in deep thought for the remainder of his/her lifetime. is the proper rate used in discounting the expected inflation rate, the time preference rate of the beneficiary, or the long term u.s. treasury bond rate? what adjustment should be made when different asset/liability combinations or estate plans have different levels of risk? the inability of this author to unequivocally answer these questions does not invalidate the probabilistic estate planning model, but in practical applications these questions must be addressed and answered. since the goal of probabilistic estate planning is to maximize the net present value of the after-tax estate passing to the heir(s), perhaps the most logical choice of discount rates is the time preference rate of the heir(s) of the estate. acomparisonoftraditionalandprobabilistic estate planning assume that a couple, both age 60, have a $2,000,000 estate. assume that $1 $00,000 is invested in husband owned real estate which is appreciating at 7 % per year and that $1 ,ooo,ooo is invested in wife owned bonds paying 9% (taxable at a constant 28 %) per year. assume that the estate has no debts. assume their wills are spouse to spouse, remainder to child. assume that the couple has a single child. assume that the couple lives in wisconsin. assume that the child has a time preference for money of 5 % . figure 2 depicts the traditional estate planning analysis for the above couple and their child. the husband is assumed to die at age 78, his wife at age 83 (their 1984 u.s. life table life expectancies, rounded to the nearest integer). wisconsin, like many states, provides for an estate tax equal in amount to the allowable federal estate state death tax credit. since that tax has no real cost to the heir (what the federal government loses in tax revenue is equal to that paid to the state of wiscon sin), only the total estate tax is shown in the analysis below and in all subsequent analyses. assume that administration expenses are equal to 5% of the gross estate. the traditional analysis demonstrates that the total estate tax payable is $3,735,233, that the net to the child is $4,569,233. since that tax is not payable for 23 years and the net to the child is not receivable for 23 years, those numbers exaggerate both the inheritance of the child and the total estate tax on the estate of the mother. in traditional analysis, the next step is usually to show how to lower the total estate tax. by changing the wills and using a trust (which passes legal title) or 152 financial services review, l(2) 1991 spouse to spouse, remainder to child gross estate of father $3,379,932 gross estate of mother administration expenses $168,996 administration expenses adjusted gross estate $3,210,936 taxable estate marital deducation $3,210,936 $0 total estate tax payable taxable estate total estate tax payable $0 net to child figure 2. traditional estate planning analysis-status quo. $8,741,543 $437,077 $8,304,466 $3,735,233 $4,569,233 an outright gift (which passes both legal and equitable title) of $600,000 to the child on the death of the father, significant future tax savings (in an amount of $399,727) are possible. see figure 3. as an alternative to changing the wills, an estate planner might suggest the purchase of insurance and the placing of that insurance in a trust. hence, assume that the wife takes $500,000 of her assets and purchases a single premium life insurance policy with a guaranteed face amount of $1,567,570.1 although tax reform has effected the investment aspects (for the policy owner who takes posses sion of the cash value through loans, withdrawals, or terminations) of owning single premium whole life, “the new tax treatment will in no way decrease the return to policy owners who leave the funds with the insurer and look to the death benefit as the primary benefit of the contract” (leimberg et al., 1989). single premium whole life insurance continues to be a viable product ifused to provide death benefits. assume that this policy is on the life of the husband. assume that the policy is held in trust for the benefit of the child, and held in such a manner (e.g., no retained interest) such that the policy proceeds are not included in the mother’s estate. (note that the gift itself will need to be included in her estate for federal estate tax purposes.) assume the trustee charges l/2 of 1% annually to administer the trust $600,000 to child on first death, remainder to child on second death gross estate of father $3,379,932 gross estate of mother $7,900,012 administration expenses $168,996 administration expenses $395,001 adjusted gross estate $3,210,936 taxable estate $7,505,011 marital deducation $2,610,936 total estate tax payable $3,335,505 taxable estate $600,000 net to child, mother’s total estate tax payable $0 death !$4,169,506 net to child, father’s death* $841,531 total future value to child $5,011,037 figure 3. traditional estate planning analysis-revised will. (*) compounded to mother’s death. probabilistic estate planning 153 after the policy benefit is paid into the trust. assume that the trustee projects a 7 % rate of return on trust corpus. see figure 4 for a visual picture of the restructured estate. the total estate tax payable using the life insurance alternative is $2,931,392, significantly less than under a revised will. the net to the child is $5,235,751, significantly more than under a revised will. on a traditional estate planning life expectancy point estimate basis, the life insurance option is clearly optimal. initial estate plan husband real estate $1 ,ooo,@oo wife bond portfolio $1,000,000 total estate $2,000,000 will: wife’s assets to husband, husband’s assets to wife, survivor’s assets to child revised estate plan parent’s estate child’s trust estate husband real estate $1,000,000 life insurance trust* $500,000 wife bond portfolio $481,800 proceeds at death of total estate $1,481,800 father $1567,570 life insurance trust $500,000 *this is the initial cash value/premium state gift taxes $17,200 for the policy. the policy proceeds will legal and trust compound until the mother’s death expenses $1,000 since the father is assumed to die first. total estate + trust + expenses $2,ooo,ooo will: wife’s assets to husband, husband’s assets to wife, survivor’s assets to child using a life insurance trust funded with the wife’s assets gross estate of father $3,379,932 gross estate of mother administration expenses $168,996 administration expenses adjusted gross estate $3,210,936 taxable estate marital deducation $3,210,936 total estate tax payable taxable estate $0 net to child, mother’s total estate tax payable $0 death insurance trust $1,567,570 -+ net to child, father’s death* total future value to child $6522,929 $326,146 $6,196,783 $2,93 1,392 $3,265,391 $1,970,360 $5,235,751 figure 4. traditional estate planning analysis-life insurance. (*) compounded to mother’s death. 154 financial services review, l(2) 1991 the value (and purpose) of probabilistic estate planning becomes evident when the same estate scenarios are replayed on a probabilistic basis. a flow chart for the probabilistic estate planning process is shown in figure 5. a 100 trial simulation of the status quo estate plan yielded a net present value to the child of $1,558,154 i start 1 read the mortality table into memory 1 input other variables: husband’s age/wife’s age number of trials joint and separate assets rates of return on those assets income taxability of those assets 1 determine random death age combinations 1 tax and distribute estate to heirs for each combination of death ages 1 input the discount rate 1 compute net present value of assets passing to heirs 1 output results to screen 1 change estate plan or change asset mixture 1 recompute taxes and distribute estate to heirs for each combination of death ages chosen above 1 i i compute net present value of assets passing to heirs 1 l compare results under initial plan to that of revised plan 1 1 output results to screen 1 got01 or stop figure 5. computer program flow chart. flow chart for initial estate plan flow chart for revised estate plans probabilistic estate planning 155 and a standard deviation of $108,438. (for comparative purposes, the present value of the traditional point estimate was $1,487,611.) the life insurance policy option, clearly optimal on a point estimate’basis, is no longer clearly optimal on a probabilistic basis. using the same death age combinations as in the first simulation yielded a net to the child of $1,825,164, with a standard deviation of $267,547. on a probabilistic basis there is a large increase in thesizeoftheexpectedestate (fromameanof$1,558,154 toameanof$l,825,164), but accompanying that expected increase is a large rise in the level of risk (from a standard deviation of $108,438 to a standard deviation of $267,547). the decision to purchase the life insurance now depends on the risk preferences of the child. the risk to the child in the probabilistic analysis arises from the fact that if the husband dies shortly after the policy is issued, then the net present value of the future estate of the child (after his mother has died) will be significantly larger than if no insurance were purchased; conversely, if the father lives for a long time (into his late 80s or beyond), much of the interest income earned by the life insurance company will be consumed by death benefits paid to other beneficiaries, and the net present value of the estate of the child will be smaller than if life insurance had not been purchased. since the time at which this occurrence happens is well beyond the 23 year life expectancy of the mother, this risk is not communicated to the estate owners under the traditional estate planning methodology. to further complicate matters, when the probabilistic process (using the same death age combinations as in the first simulation) is applied to the estate plan using the revised will (to take advantage of passing $600,000 total estate tax free to the child on the death of the first parent), the net present value of the estate of the child after both parents are deceased was $1,792,267, the standard deviation $140,29 1. while the expected value of this option is less than the expected value of the life insurance option, there is also significantly less risk. the trade-off between the expected return and the variance associated with that expected return (plotted versus the square root of the variance) for the three estate plans just discussed is depicted in figure 6. 300 250 jv 200 (,ooo) 150 a newwlll 100 a statusaa 0 0 1000 1200 1400 1600 1800 expected value (,ooo) figure 6. three alternate estate plan e-v combinations. 156 financial services review, l(2) 1991 figure 7. attainable estate plan e-v combinations. now the child must decide what risk level s/he is comfortable with, and that decision will allow the child, from amongst those three estate plans, to chose the optimal one. the life insurance purchase is no longer a clearly optimal plan. whether or not any or all of the three estate plans just examined fall on the optimal estate plan frontier or lie above and to the left of that frontier is unimportant; the object of probabilistic estate planning is now clear. on a probabilistic basis, the net present value of assets passed to the child can be graphed versus the variance (or square root thereof) associated with that estate plan. by investigating other asset reallocation possibilities and alternative estate plans (gifting of assets without purchasing insurance, combinations of will changes, trusts, use of insurance, et al.) and plotting means and variances for different estate plans and asset/liability combinations, the estate planner can identify the optimal estate plan frontier. (see figure 7.) attainable estate plans which are inferior to those on the frontier can be avoided. the plan which maximizes the expected net present value of assets passed to the child consistent with the risk level with which the child is comfortable can be implemented. conclusions probabilistic estate planning allows those decision makers who are interested in risk to make optimal estate planning decisions. if the plan which maximizes the net present value of assets passed to the heir(s) simultaneously has the smallest variance, then clearly that plan is optimal. if, however, the plan which maximizes the net present value of assets passed to the heir(s) simultaneously has the largest variance, and other plans having lower expected values have lower variances, then the e-v rule applies, and the heir(s) must choose a plan consistent with his/her/their risk level(s) from among those plans falling on the optimal estate plan frontier. probabilistic estate planning 157 combining the basic principle in finance of maximizing expected net present value with jensen’s inequality and markowitz’s e-v rule results in a process referred to by this author as probabilistic estate planning. the primary advantages of this methodology over current methodology are that it incorporates the element of risk, that the probabilistic mean is unbiased with respect to growth rates, discount rates and age(s) of the estate owner(s), and that results expressed on a present value basis (rather than on a future value basis) are more meaningful to estate owners. in summary, probabilistic estate planning permits modern portfolio theory (mean, variances tradeoffs) to be used to select an optimal estate plan. there are other variables influencing an estate plan that could be considered as random variables. considering time until death is a first step. acknowledgments: thanks are extended to the two anonymous referees and james c. hickman for their numerous helpful comments and suggestions. notes 1. policy issued by a major life insurance company on a no-load basis; valedictorian i-single premium life; dec. 1987, 6% guaranteed, 9% projected. references ackerman, laurence j. 1973. “estate planning principles,” life and health insurance handbook, third edition. richard d. irwin, p. 845. jordan, chester wallace jr. 1952. life contingencies. the society of actuaries, pp. 246-247. leimberg, stephan r., et al. 1989. the financial services professional’s guide to the state of the art/ 1989. bryn mawr, pa: american college, page 6.9. markowitz, harry m. 1952. “portfolio selection,” journal of finance, march: 312, 324. miller, ralph gano. 1987. “selling estate planning with a computer screen and without reams of paper,” journal of american society of clu & chfc, january: 88-89. national center for health statistics. 1987. vital statistics of the united states, 1984, vol. ii, mortality part a. dhhs pub. no (phs) 87-l 122. public health service. washington: u.s. government printing office, section 6, p. 11. finser_31-2_complete_issue the changing assessment of risk for young investors kristine l. becka, hsin-hui chiub, inga timmermanc,* adepartment of finance, financial planning and insurance, california state university, northridge, 18111 nordoff street, northridge, ca 91330-8379, usa bdepartment of finance, financial planning and insurance, california state university, northridge, 18111 nordoff street, northridge, ca 91330-8379, usa cdepartment of accounting and finance, university of north florida, 1 unf drive, jacksonville, fl 32224, usa abstract investment advice is changing to incorporate new products and platforms, and the rate of change is likely to accelerate as millennials and gen z increase their involvement in investment markets. using survey methodology, we examine the changing landscape of risk tolerance for young people, concluding that the typical risk assessment tools advisors use may not be as applicable to the next generation of investors. we find that the components that drive willingness to take risk are interest in investments, self-reported investment risk tolerance, and ownership of investment accounts. our findings indicate that it is time to start assessing risk differently.1,2,3 © 2023 academy of financial services. all rights reserved. jel classifications: g11; g59; d81 keywords: investment risk; risk assessment; risk tolerance 1. introduction and literature review according to mckinsey & company research, the wealth management industry in north america has been evolving tremendously over the last 20 years. between 2000 and 2010, total assets grew by approximately 45%, from $13 trillion to $19 trillion, and by 2018 client assets were up to $30.5 trillion.4 more importantly, mckinsey determined that both the demographic composition of those who invest and their investment choices are very different today than they were 20 years ago. millennials and gen z segments of the population control more assets than ever, and their ease of use with digital tools is extremely high. for example, while 31% of affluent gen z use a robo-advisor, only 13% of gen x do so.5 but *corresponding author. tel.: (904) 620-5354. e-mail address: inga.timmerman@unf.edu (i. timmerman) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 97–106 how do gen z investors decide on investments and how do they measure their risk tolerance and capacity? many instruments are available to assess risk tolerance for portfolio construction, some linking risk tolerance to financial knowledge. it is unclear, however, how well these tools capture the changing world of investment risk. when a 20-year-old has a cryptocurrency wallet, access to a brokerage account in the form of a robinhood app, and is having conversations about gamestop, one wonders how well the traditional risk profiling tools capture both financial knowledge and risk tolerance.6 what is becoming clear is that the large segment of the population that is moving into the asset accumulation phase has a very different perception of risk than the prior generation of advisory clients. as a result, it is important to modify the tools used to assess those risks to better capture the shifts in financial risk tolerance. understanding a client’s risk tolerance is the most pressing issue in a successful financial planning process (moreschi, 2005). we argue that traditional risk-based portfolio construction methods may not be the most suitable way to build portfolios for divergent generations. several research papers have already questioned whether traditional questionnaires could truly assess a client’s risk tolerance (yook & everett, 2003; bouchey, 2004; roszkowski, davey, & grable 2005). risk assessment tools developed by academics and used by advisors for decades may not accurately assess the risk tolerance and risk inclination of incoming investors. to partially solve this problem, roszkowski and grable (2005) propose to introduce psychology and psychometrics into designing a more appropriate questionnaire to assess true risk tolerance. more recent literature proposes to design an instrument to assess various aspects of risk tolerance (risk knowledge, risk capacity, risk attitude, and risk propensity) from psychological perspectives (wahl & kirchler 2020). traditional factors such as gender, age, and household income are determinants of investor risk tolerance (sung & hanna 1996; yao & hanna 2005). education level is also found to be positively related with risk tolerance in these studies.7 in particular, research shows the financial knowledge of college students is related to greater risk tolerance (park et al., 2020; rabbani et al., 2022; sjöberg & engelberg, 2009). given changing investment demographics and the way investment decisions are now made, we seek to understand risk tolerance firsthand by surveying college students enrolled in courses that cover investment fundamentals. we measure changing perceptions young investors have toward investment risk to understand how young adults who have been exposed to modern investment tools see risk and the components that make up their risk profile. we analyze the impact of the individual willingness of each survey participant to take risks. specifically, we explore the following questions: (1) does enrollment in additional finance classes change the perception of risk tolerance? in other words, do advanced finance classes matter to the individual in the analysis of risk? (2) does the knowledge component (both objective and subjective) still have an impact on risk tolerance given the changing dynamics of the investment landscape? (3) does interest and experience with investments affect risk tolerance for current college students? this study is a first step aiming at informing academics of the need to develop new and adaptive tools or questionnaires to accurately 98 k. l. beck et al. / financial services review 31 (2023) 97–106 assess the way young adults make investment decisions. the results will also help advisors serve their future gen z clients. 2. survey design and results during the fall 2021 semester, we conducted a survey of students who have taken at least one finance class in the college of business at a large public university. we limit the sample to business majors/minors for two reasons: first, we are targeting individuals who have taken at least one introductory finance class and have a basic understanding of financial markets and investments. second, we want to capture shifting demographics and believe the university environment provides the best access to individuals who participate in the modern investment landscape. we survey juniors and seniors who by the end of 2021 have taken at least one finance course. grable, heo, and kruger (2016) identify factors that are associated with financial risk tolerance and recommend financial planners gather information such as gender, age, education, marital status, household income, household size, and net worth during the planning process. using the same variables, we conduct a survey consisting of questions designed to assess the following independent variable categories associated with risk: (1) demographic data, (2) investment knowledge/acumen, both objective and subjective, (3) interest and experience with investments, and (4) control variables that have been shown to be associated with risk tolerance in prior research. we measure each respondent’s risk tolerance by asking both hypothetical risk assessment questions and current situational questions about the ability to undertake risk.8 we collected results during the first two weeks of december 2021 for 262 respondents who were finishing the fall semester. we measure risk tolerance across four categories/ dimensions and present the results in the first four rows of table 1. all four dependent variables are measures of risk derived from answers to a specific survey question. for example, the first question is “protecting my money is more important than high returns.” the answers range from 1 to 5, depending on the degree of agreement with the statement. the second assessment asks the respondents to invest a hypothetical $1,000 from a relatively conservative target date fund to a portfolio of cryptocurrency. the third question assesses what they would do in case of a sharp stock market decline, and finally, the last measure is an aggregate score (aggriskscore) of the first three questions. aggriskscore is our main measure of risk assessment. the results in table 1 show that aggregate risk ranges from 3 to 13, with a mean of 7.9 and a standard deviation of 1.97. the next two panels of table 1 summarize the independent variables used. it is noteworthy that about 40 of the students took another finance course beyond the basic finance class. respondents are neither overly optimistic nor pessimistic in their assessment of investment knowledge, with the mean at 4.97 out of 10. the objective knowledge measures at 3.04 out of 5. interest in investments averages 2.58 out of 3, which shows that this group of students has considerable interest in investments. about 47% of respondents own stock investments k. l. beck et al. / financial services review 31 (2023) 97–106 99 like mutual funds, etfs, or individual stocks. we also measured the self-reported risk attitude from 1 (very conservative/do not know) to 5 (aggressive); the average is 2.7. along with demographic variables, the respondents have a higher gpa than the average student, are roughly even gender-wise, have a range of investment accounts already, and about 75% have some savings and some debt.9 table 1 descriptive statistics for the sample variable mean sd min max n protectmoneycat 2.4847 0.9697 1 5 262 1kinvestcat 2.3587 1.0654 1 4 262 stockmarketdowncat 3.0801 0.8516 1 4 262 aggriskscore 7.9236 1.9695 3 13 262 advancedclasscat 0.4083 0.49247 0 1 262 self-knowldcati 2.8549 0.8034 1 5 262 selfratedknowledgeinvestments2 4.973 2.0799 0 10 261 knowledgescore 3.0419 1.4836 0 5 262 interestedinvcat 2.5839 0.5593 1 3 262 ownetfsmutualcat 0.4656 0.4997 0 1 262 invattcat 2.7022 1.0735 1 5 262 gpacategory 3.5725 0.9099 1 5 262 gendercat 0.5190 0.5005 0 1 262 agecat 1.4541 0.70289 1 4 262 hhincomecat 2.6436 1.5760 1 6 261 debtcat 2.3282 1.4030 1 6 262 savingscat 0.7786 0.4159 0 1 262 accountsheldcat 1.8320 1.6077 0 4 262 note. the first four rows represent dependent variables, and the rest indicate independent variables. protectmoneycat ranges from 1 to 5, where 1 is strongly agree and 5 is strongly disagree with the statement “protecting my money is more important than high returns.” 1kinvestcat ranges from 1 to 4, where a respondent chooses how to invest $1,000 from 1, most conservative in a target date fund, to 4, most aggressive in cryptocurrency. stockmarketdowncat ranges from 1 to 4, where 1 is equal to selling all investments immediately in response to a sharp market decline, and 4 to immediately buying more. aggriskscore is the sum of the previous three variables, and ranges from 3 to 13. advancedclasscat is equal to 1 if the respondent completed an advanced finance class and 0, otherwise. self-knowldcati represents self-reported perception of investment knowledge, ranging from 1 to 5, where 1 is poor and 5 is expert. selfratedknowledgeinvestments2 is a wider measurement of self-reported investment knowledge, ranging from 0 to 10 where 0 is not knowledgeable at all and 10 is an expert. knowledgescore is an aggregate objective investment score ranging from 0 to 5 on five investment-related questions. interestedinvcat represents self-reported interest in investing, ranging from 1 to 3, where 1 is not interested and 3 is very interested. ownetfsmutualcat is equal to 1 if the respondent owns any stock, etfs or mutual funds and 0, otherwise. invattcat is a self-reported measure of attitude in investments, ranging from 1 to 5, where 1 is very conservative/do not know to 5, very aggressive. gpacategory ranges from 1 to 5, where 1 is a gpa of 2.0 and 5 represents a gpa of 4.5. gendercat equals 1 if the respondent identifies as a female and 0, otherwise. agecat ranges from 1 to 4, with 1 representing younger respondents. hhincomecat ranges from 1 to 6, based on the combined income of the family, from lowest to highest. debtcat represents debt held, from 0 for category 1 to more than $50,000 for category 6. savingscat represents individual savings, where 1 represents existing savings designated for emergencies and 0, otherwise. accountsheldcat represents the complexity and range of investment accounts held, where 0 means no accounts, 1 represents work-sponsored accounts and ira types only, and 4 includes brokerage and cryptocurrency accounts. 100 k. l. beck et al. / financial services review 31 (2023) 97–106 table 2 presents the multivariate analysis for the three risk questions, exploring how risk is influenced by category. we further differentiate aggregate risk in table 3. models 1-3 in table 2 present the results for each of the three individual risk questions. as each of the individual questions measures a slightly different risk dimension, it is important to look at the results by individual question. in model 1, we look at the agreement with the statement that “protecting money is more important than high returns.” we find that the only measure with a significant impact on the result is the self-reported assessment of one’s risk tolerance, invattcat. the higher the self-reported risk tolerance, the more risk the person is willing to take, as evidenced by the strongest disagreement with this statement. there appears to be a positive relationship between willingness to take investment risk (look for higher return vs. conservative money protection) and the way an individual perceives themself in relation to risk. when asked to rate a different category of risk, hypothetically investing $1,000 in a range of options, the answers ranged from a relatively conservative target date fund to a risky allocation of 100% cryptocurrency. we found that the main variable influencing investment is prior ownership of financial assets. the more account diversity the investor has, the more likely the person is to make a risky investment. this is a key finding that has implications for how we teach finance and the type of practical knowledge students should be exposed to in a finance class. the final individual risk question presents another hypothetical scenario in which the stock market declines sharply. the measure of risk ranges from conservative, where a respondent decides to sell the rest of the position, to aggressive, where a respondent decides to buy more of the same, now low, investment. we find several variables that influence this risk tolerance dimension. the same two variables mentioned in the previous questions are table 2 regression analysis. influence of risk by category variable model 1 model 2 model 3 advancedclasscat 0.05988 (0.659) 0.08012 (0.565) !0.0229 (0.828) selfratedknowledgeinvestments2 0.06633 (0.128) 0.0044 (0.914) 0.02821 (0.357) knowledgescore !0.04883 (0.298) 0.0638 (0.176) 0.03902 (0.302) interestedinvcat 0.17453 (0.142) 0.1071 (0.375) 0.2307 (0.022)** ownetfsmutualcat 0.08036 (0.585) 0.2789 (0.112) 0.1144 (0.387) invattcat 0.17935 (0.012)** 0.0994 (0.169) 0.03731 (0.471) gpacategory 0.02548 (0.735) !0.0206 (0.799) 0.1548 (0.007)*** gendercat !0.1336 (0.350) !0.0586 (0.692) !0.1923 (0.121) agecat !0.07843 (0.426) !0.0278 (0.757) 0.12962 (0.090)* hhincomecat !0.06061 (0.123) !0.0305 (0.462) 0.01238 (0.676) debtcat 0.04026 (0.439) 0.0181 (0.724) 0.0497 (0.121) savingscat !0.1764 (0.249) !0.1072 (0.511) 0.0559 (0.633) accountsheldcat !0.06035 (0.218) 0.12049 (0.044)** 0.0784 (0.072)* constant 1.7172 (0.000)*** 1.4883 (0.001)*** 1.1004 (0.001)*** model p-value .0031*** .0000*** .000*** adj r2 0.1222 0.1834 0.2530 n 260 260 260 note. models 1–3 have the following dependent variables: protectmoneycat, 1kinvestcat, and stockmarketdowncat. see table 1 for variable definitions. ***significant at 1%, **significant at 5%, *significant at 10%. k. l. beck et al. / financial services review 31 (2023) 97–106 101 also of interest here; the more interest one has in investments and the more accounts one already has, the more likely the person is to hold the existing positions and to see the decline as an opportunity. this is a key finding that can also be addressed when teaching finance. given the challenge of timing the market and the fact that a typical investor on average realizes only about 38% of the broad market return, being comfortable with taking the risk and holding on to a position when the market declines would result in significantly higher wealth over the lifetime of the individual.10 additionally, we find some demographic variables have an association with willingness to stay invested. specifically, an older survey respondent who has a higher gpa is more likely to see this as an opportunity to take risks than an opportunity to cut losses. this can be explained by older individuals having more experience with investments, and the fact that students with a higher gpa may have more knowledge regarding financial markets and, as a result, are more willing to take risks. to further examine the relationship between aggregate risk and impacting factors, in table 3 we assess individual components that may influence aggregate risk. the dependent variable is an aggregate risk measure constructed from the three individual risk questions. in model 1, we test only demographic variables; in model 2 we add existing savings and debt, variables that show the ability to undertake risk; in model 3 we take a separate look at the most significant variables in the previous analyses, such as interest in investments and investment knowledge. the last model combines all variables. we find that the demographic and knowledge variables have almost no relationship to the risk tolerance of a college student with some knowledge of finance; the only demographic characteristic with explanatory power is gender. as previously documented in literature, females are more prone to risk aversion (see, e.g., olsen & cox, 2001; lascu, babb, & phillips, 1997). unlike prior literature, we do not find that financial knowledge is linked to risk. the difference can be attributed to the sample used. it is possible that once someone has a basic understanding of finance and investments through exposure to introductory finance, the significance of the knowledge differential disappears. our sample consists of students who already have a basic understanding of finance and investments. this aspect needs to be explored further as it affects the way financial risk should be taught in college and discussed with clients. research is mixed on whether risk aversion increased during the pandemic so we cannot assume a higher or lower level of risk aversion for this sample.11 of interest in our results is the focus on the variables that are not usually studied in the context of risk tolerance. the components that drive risk-taking willingness are the interest in investments, the self-reported risk tolerance attitude, and existing investment accounts. the more interest one has in investments, the more willing one is to take on investment risks, at least in hypothetical scenarios. the same holds true for how one perceives their own risk attitude. the more risk-tolerant one self-identifies, the more likely they are to choose riskier investments. finally, someone who has several accounts, including riskier cryptocurrency and individual stock investments, and someone who already has investments is more likely to be comfortable with risk. these few factors explain about 31% of the variability in aggregate risk. 102 k. l. beck et al. / financial services review 31 (2023) 97–106 3. conclusions and implications the requirements of investment advice are changing due to new products and trading platforms, and the expectation is that they will change even faster as millennials and gen z access investment markets. in this paper, we look at the changing landscape of risk tolerance and perceptions for current college students, concluding that the typical risk assessment tools advisors use may not be as applicable to the investors who will be coming through their doors in the next decade. although new risk assessment tools are beyond the scope of this paper, our study indicates that it is important to change the way we perceive and measure investment risk. future research could also survey intergenerational changes in risk tolerance as well as include non-college gen z investors to address possible sample selection bias. another implication of our study is the importance of updating how finance courses are currently taught in most universities. for example, we find that the type of investment accounts already owned is the major driver of how a young investor will choose to invest additional money. it is important to point out that many finance courses while succeeding in the delivery of theoretical knowledge, do not go into the practical aspect of opening investment accounts. familiarizing students with investments, even if hypothetical through an investment simulation game, is a great way to create exposure and the willingness to increase risk. advisors should also devote time to educating clients on risk tolerance. overall, we find that the components that drive willingness to take risks are interest in investments, self-reported investment risk tolerance, and ownership of investment accounts. our findings point out the need to assess risk in finance, investment, and personal finance table 3 aggregate risk results variable model 1 model 2 model 3 model 4 advancedclasscat 0.36027 (0.108) 0.3191 (0.169) self-knowldcati !0.2309 (0.173) !0.1691 (0.319) knowledgescore 0.1373 (0.085)* !0.1691 (0.319) interestedinvcat 0.5905 (0.005)*** 0.6089 (0.004)*** ownetfsmutualcat 0.5655 (0.046)** 0.5280 (0.065)* invattcat 0.38895 (0.001)*** 0.4035 (0.001)*** accountsheldcat 0.5908 (0.018)** 0.1663 (0.066)* gpacategory 0.1034 (0.445) 0.2002 (0.177) 0.1584 (0.245) gendercat !1.3034 (0.000)*** !1.2695 (0.000)*** !0.4546 (0.066)* agecat 0.27856 (0.122) 0.1634 (0.386) 0.0665 (0.708) hhincomecat 0.03264 (0.669) 0.01323 (0.861) !0.0518 (0.473) debtcat 0.2304 (0.014)** 0.1032 (0.236) savingscat 0.1617 (0.580) !0.1728 (0.535) constant 7.7455 (0.000)*** 6.9372 (0.000)*** 4.7929 (0.000)*** 4.6703 (0.000)*** model p-value .0000*** .000*** .0000*** .000*** adj r2 0.1115 0.1345 0.2940 0.3047 n 261 261 262 261 note. models 1–4 have the following dependent variables: protectmoneycat, 1kinvestcat, stockmarketdowncat, and aggriskscore. see table 1 for variable definitions. ***significant at 1%, **significant at 5%, *significant at 10%. k. l. beck et al. / financial services review 31 (2023) 97–106 103 courses to match the reality of the modern investment landscape. many students are already holding assets that are traditionally perceived as risky. creating a better understating of risk tolerance, how modern investments relate to risk, and the potential downfalls would be valuable to college finance students as well as young investment clients. in addition to new generations of investors, trading platforms or channels of financial advisement have changed and raised high attention from media and clients. large financial institutions such as fidelity and vanguard initiated roboadvisor to provide automated and algorithm-driven investment services and advise. with minimum human supervision, roboadvisor helps to save cost and keep account minimum low. with the rise of roboadvising services, it is especially crucial to create a survey that could accurately capture true investor risk preference. based on this study, financial institutions could consider practical variables such as ownership of investment accounts instead of testing clients finance course knowledge when assessing risk. in addition, providing clients with finance simulation courses or games to create exposure to investments will help educate concepts in risk assessment. even in an environment with minimum human involvement of roboadvising, clients will be able to increase their financial literacy through practical courses and have a better understanding in their own risk tolerance. notes 1 traditional risk tolerance questionnaires do not assess risk well for younger individuals. 2 the drivers of risk tolerance for gen z are willingness to take risks, self-reported investment risk tolerance, and ownership of investment accounts. 3 academia needs to update how risk is assessed in finance, investment, and personal finance courses to match the reality of the modern investment landscape. many students are already holding assets that are traditionally perceived as risky. 4 trillion: https://www.mckinsey.com/industries/financial-services/our-insights/ on-the-cusp-of-change-north-american-wealth-management-in-2030 5 https://www.investopedia.com/study-affluent-millennials-are-warming-up-to-robo-advisors4770577. 6 according to pew research center, 41% of men aged 18-29 have invested in cryptocurrency as of 2022: https://www.pewresearch.org/short-reads/2023/04/10/majority-ofamericans-arent-confident-in-the-safety-and-reliability-of-cryptocurrency/#:;:text¼overall %2c%2017%25%20of%20u.s.%20adults,and%20women%20of%20any%20age. 7 see sung and hanna (1996), grable (2000), ardehali et al. (2005) and halek and eisenhauer (2001). finke and guillemette (2020) provide a though review on theories and factors related to measuring risk tolerance. 8 a full survey is available upon request; please email the authors for a copy. 9 the correlation matrix and vif measures show no significant correlation between any of the variables. 104 k. l. beck et al. / financial services review 31 (2023) 97–106 10 nick murray of behavioral investment counseling estimates that for the period ending in 2007, the average equity fund returned 10.81%, while the average investor in the same fund realized 4.18% due to poor market timing. 11 see, for example, yue et al. (2020), shachat et al. (2020), angrisani et al. (2020), heo et al. (2020), and iqbal and li (2022). references angrisani, m., cipriani, m., guarino, a., kendall, r., & ortiz de zarate, j. (2020). risk preferences at the time of covid-19: an experiment with professional traders and students. frb of new york staff report, 927. ardehali, p. h., paradi, j. c., & asmild, m. (2005). assessing financial risk tolerance of portfolio investors using data envelopment analysis. international journal of information technology & decision making, 04, 491– 519. https://doi.org/10.1142/s0219622005001660 bouchey, p. (2004). questionnaire quest: new research shows that standard questionnaires designed to reveal investors’ risk tolerance levels are often flawed or misleading. financial planning, 1, 97–99. finke, m., & guillemette, m. (2020). measuring risk tolerance: a review of literature. journal of personal finance, 15, 63–76. grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625–630. https://doi.org/10.1023/a:1022994314982 grable, j. e., heo, w., & kruger, m. (2016). the intertemporal persistence of risk tolerance scores. journal of financial planning, 29, 42–51. halek, m., & eisenhauer, j. (2001). demography of risk aversion. the journal of risk and insurance, 68, 1–24. https://doi.org/10.2307/2678130 heo, w., grable, j., & rabbani, a. (2020). a test of the association between the initial surge in covid-19 cases and subsequent changes in financial risk tolerance. review of behavioral finance, 13, 3–19. https://doi.org/ 10.1108/rbf-06-2020-0121 iqbal, m., & li, l. (2022). does covid-19 really make people risk aversion in investment decision-making? shs web of conferences, 132, 01021. https://doi.org/10.1051/shsconf/202213201021 lascu, d.-n., babb, h., & phillips, r. (1997). gender and investment: the influence of gender on investment preferences and practices.managerial finance, 23, 69–83. https://doi.org/10.1108/eb018652 moreschi, r. w. (2005). a pressing issue for financial planning. journal of personal finance, 4, 43–47. olsen, r., & cox, c. (2001). the influence of gender on the perception and response to investment risk: the case of professional investors. journal of psychology and financial markets, 2, 29–36. https://doi.org/10.1207/ s15327760jpfm0201_3 park, j., kim, d., & oh, s. (2020). explicit and implicit stock investment: differences in psychological characteristics and risk behavior between college students majoring in financial engineering or business. current psychology, 39, 1954–1969. https://doi.org/10.1007/s12144-019-00555-9 rabbani, a. g., heo, w., & lee, j. m. (2022). a latent profile analysis of college students’ financial knowledge: the role of financial education, financial well-being, and financial risk tolerance. journal of education for business, 97, 112–118. https://doi.org/10.1080/08832323.2021.1895046 roszkowski, m. j., davey, g., & grable, j. (2005). insights from psychology and psychometrics on measuring risk tolerance. journal of financial planning, 18, 66–77. roszkowski, m. j., & grable, j. e. (2005). estimating risk tolerance: the degree of accuracy and the paramorphic representations of the estimate. journal of financial counseling and planning, 16, 29–47. shachat, j., walker, m. j., & wei, l. (2020). the impact of the covid-19 pandemic on economic behaviours and preferences: experimental evidence from wuhan. esi working paper, 20. sjöberg, l., & engelberg, e. (2009). attitudes to economic risk taking, sensation seeking and values of business students specializing in finance. journal of behavioral finance, 10, 33–43. https://doi.org/10.1080/15427560902728712 k. l. beck et al. / financial services review 31 (2023) 97–106 105 sung, j., & hanna, s. (1996). factors related to risk tolerance. journal of financial counseling and planning, 7, 11–20. wahl, i., & kirchler, e. (2020). risk screening on the financial market (risc-fm): a tool to assess investors’ financial risk tolerance. cogent psychology, 7, 1–33. https://doi.org/10.1080/23311908.2020.1714108 yao, r., & hanna, s. d. (2005). the effect of gender and marital status on financial risk tolerance. journal of personal finance, 4, 66–85. yook, k., & everett, r. (2003). assessing risk tolerance: questioning the questionnaire method. journal of financial planning, 16, 48–55. yue, p., gizem korkmaz, a., & zhou, h. (2020). household financial decision making amidst the covid-19 pandemic. emerging markets finance and trade, 56, 2363–2377. https://doi.org/10.1080/1540496x.2020.1784717 106 k. l. beck et al. / financial services review 31 (2023) 97–106 pii: 1057-0810(91)90027-v financial services review, l(2): 101-108 copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. should individual investors avoid the stock market outside of january? steven v. mann donald p. solberg recent studies suggest that there is no rewardfor bearing risk outside of january, implying that individuals should invest in common stocks only in january. the purpose of this study is to demonstrate that this conclusion is far too strong given existing empirical evidence. our results suggest that inferences drawnffom the evidence can be altered greatly through small changes in the way the empirical question is addressed. there is suficient evidence to doubt the conclusion that individuals are not compensatedfor the risk ofparticipating in the stock market outside of january. recent studies present a disturbing picture of the relationship between risk and return. i simply put, researchers suggest that there is no reward for bearing risk outside of january. the return foregone by holding treasury bills rather than risky stocks in months other than january is virtually zero. if true, the message to individual investors is clear, albeit surprising. based on these studies, the rational investor would invest in common stocks during the month of january, then shift funds into riskless assets for the remainder of the year. the validity of this strategy is especially important to the individual due to transaction costs incurred during portfolio rebalancing. clearly, if there is no reward to risk outside of january, it is puzzling that rational investors would choose to hold stocks in the remaining eleven months of the year. in this paper, we demonstrate that this conclusion is far too strong given existing empirical evidence. the most troubling studies are by chang and pinegar (1988a and 1988b), which examine excess monthly returns of long-term corporate over government bonds, stocks over government bonds, and individual stocks over bonds of the same firm. they conclude that investors receive no risk premia february through decem ber. chang and pinegar’s results appear especially powerful because their tests rely steven v. mann l college of business administration, university of south carolina, columbia, sc 29208. donald f’. solberg l federal home loan mortgage corporation, mclean, va 22102. 102 financialservicesreview,1(2) 1991 on the least restrictive model imaginable-that investors demand incremental returns for taking greater risk. our analysis suggests that this approach is deceptively simple. the test of whether a risk premium differs significantly from zero is really a test of whether the average risk premium is “large” relative to its standard error. this test is sensitive to both the market price of risk (which varies over time) and the size of the return sample. furthermore, relatively minor changes in the research question can produce the opposite conclusion. given these conflicting results, one must attempt to make a sensible interpre tation of the available evidence. one could conclude that investors are irrational and require compensation for risk bearing only in january, even if it runs counter to our intuition. alternatively, the empirical tests and interpretation of these tests may be at fault. our own results place in doubt the conclusion that equity risk premia are zero in february through december. we leave it to the reader to decide if investors are irrational for holding equity in non-january months. data we examine excess returns for bearing risk with two equity portfolios. the first, common, is the standard and poor’s composite index. the risk premium is the difference between monthly returns on this index and returns on one-month treasury bills. the second portfolio, small, is a value-weighted portfolio of the smallest quintile stocks in terms of market value on the new york stock exchange.* similarly, the risk premium is monthly returns on this portfolio less returns on one month treasury bills. return series for these portfolios are taken from the stocks, bonds, bills and inflation 1989 yearbook.3 empiricalresults table 1 presents average risk premia and their standard deviations for each month over the period 1926-1988. table 1 also presents t-statistics for the null hypothesis that mean monthly risk premia are zero using a one-tail test. statistical significance at the five percent level requires t-statistics of at least 1.645. while january returns safely meet this benchmark, the other months meet the standard infrequently. generally, the results of table 1 are similar to those reported in other studies. although the methodology in these tests appears straightforward, similar statistical tests can produce widely differing conclusions. we use two different approaches. the first examines the question: are monthly risk premia positive if we pool the eleven monthly risk premia for all non-january months into a single sample? this method increases the number of observations in the test and increases the precision of the sample estimates accordingly. furthermore, this test answers should individual investors avoid the stock market outside of january? 103 table 1. summary measures for the individual monthly risk premia (sample period: 1926-1988) jan feb mar apr may jun jul aw scp ott nov dee arithmetic mean 0.014 0.002 0.001 0.011 -0.00 1 0.012 0.011 0.018 -0.012 -0.002 0.010 0.013 cowllllotl standard deviution 0.049 0.043 0.054 0.015 0.062 0.059 0.068 0.066 0.063 0.065 0.054 0.038 tarithmetic statistic mean 2.19 0.068 0.43 0.014 0.09 -0.001 1.12 0.010 -0.14 -0.001 1.68 0.009 2.00 0.022 2.15 0.011 -1.55 -0.010 -0.23 -0.013 1.47 0.009 2.64 0.006 small standard deviation 0.091 0.064 0.083 0.103 0.112 0.085 0.086 0.108 0.102 0.088 0.014 0.064 t statistic 5.93 1.14 -0.14 0.11 -0.04 0.82 2.01 1.28 -0.80 -1.11 0.95 0.11 table 2. summary measures for the monthly non-january risk premia (sample period: 1926-1988) portfolio arithmetic standard mean deviation t-statistic common-bills 0.0062 0.0599 2.72 small-bills 0.0056 0.0895 1.64 table 3. summary measures for the eleven month february-december risk premia (sample period: 1926-1988) portfolio arithmetic standard mean deviation t-statistic common-bills 0.070 0.203 2.14 small-bills 0.068 0.327 1.66 104 financial services review, l(2) 1991 more directly the question of whether risk premia are positive on average over all non-january months. table 2 presents means and standard deviations for monthly non-january risk premia of our two equity portfolios. utilizing a one-tail test, table 2 also presents t statistics for the null hypothesis that mean risk premia are zero. the mean risk premia for the common and small equity portfolios are .62 % and .56 % , respec tively; they are positive and statistically significant at conventional levels. contrary to the findings of chang and pinegar and others, these results suggest that equity investors are rewarded on average for bearing risk in non-january months. a second set of tests asks if investors are rewarded for risk bearing over eleven month holding periods, february through december. a series of eleven-month returns are computed for the equity portfolios for each year 1926-1988.4 returns for investing in the corresponding series of eleven one-month treasury bills are also computed. finally, we compute mean differences between equity returns and trea sury bill returns over these eleven-month periods. table 3 presents means and standard deviations for the eleven-month return spreads for the two equity portfolios. the t-statistics, under the null hypothesis that the mean differences are zero using a one-tail test, indicate that mean differences are positive and statistically significant. on average, equity investors earn an excess return over treasury bills of approximately seven percent.5 like the pooled one month returns, these eleven-month returns also suggest that equity investors are compensated for bearing risk outside the month of january. previous researchers’ conclusions that there is no reward to bearing risk in non-january months are sensitive to the specific hypothesis being tested. however, our results do not settle the issue about whether investors earn returns appropriate for the level of risk they bear, period. in the next section, we argue that researchers are asking for more than empirical tests (such as those used here and in previous studies) can deliver. insearchof therisk-returnrelationship results in table 1 lead to opposite conclusions from results in tables 2 and 3 for equity risk premia, even though the samples are drawn from the same raw return series. surely, the tests used to detect whether investors are compensated for bearing risk deserve more careful scrutiny. the question motivating these tests is simple enough. is the difference between the return on risky assets and the return on a “riskless” asset positive on average in each calendar month of the year? unfortunately, there are some pitfalls in this seemingly direct test. the t-statistic in our hypothesis tests is given by where e(z?p) is the sample mean excess return or risk premium of a particular should individual investors avoid the stock market outside of january? 105 portfolio, a(rp) is the estimated standard deviation of the risk premium, and n is the number of sample observations. the denominator in this expression is the standard error, which is simply the standard deviation divided by &. rewrite the expression above as (2) the ratio e(rp)/a(rp) can be viewed as the reward to risk demanded by investors. the other component of the t-statistic, 4, is important from a purely statistical perspective. as sample size increases, one can obtain more precise estimates of the sample mean risk premium. therefore, any ratio of e(w) to a(w) is more likely to be statistically different from zero if the two parameters are estimated with a large number of observations. in principle, even a relatively small reward to risk would be judged significant if it persisted over a long period of time. in contrast, even a “substantial” reward to risk might not be significantly different from zero if there were very few observations. the influence of sample size explains why the inferences drawn from examin ing the pooled set of february through december risk premia in table 2 would differ from examining each of the eleven calendar months individually, as in table 1. there are only sixty-three observations for each calendar month reported in table 1, but 693 observations for each risk premium series reported in table 2. in contrast, chang and pinegar’s (1988b) sample has twenty observations per return series; their study would require a very high risk premium to standard deviation ratio to produce statistical significance. sample size, however, does not explain the contradictory results in tables 1 and 3. the return series in table 1 represent eleven series of sixty-three one-month risk premia while the return series in table 3 represents sixty-three eleven-month risk premia. this apparent contradiction is resolved if we consider the relationship between return interval and the standard deviation of returns. for return series that are well-behaved statistically (zero serial correlation and constant variance), the standard deviation of the eleven-month returns will be approximately m or 3.16 times as large as the standard deviation of monthly returns.6 the standard deviations increase at the rate of &as the interval over which returns are computed increases, while returns increase at the rate oft. as a result, average eleven-month returns will be approximately eleven times as large as average one-month returns. since the numerator grows at a rate oft, while the denominator grows at ji, the ratio e(z?p)/a(z?p) should increase at the rate of aas the return interval increases and t-statistics should reflect this behavior. for example, the t statistic for eleven-month returns would be fi times the t-statistic for one-month returns if the number of observations is held constant. this is precisely what we see in table 3. both the one-month average risk premia in table 1 and the eleven-month average risk premia in table 3 have sixty-three observations. however, the eleven month average risk premia in table 3 are positive and statistically significant while most of the one-month average risk premia are not. 106 financial services review, l(2) 1991 standard deviation has a dual role in these tests. first, standard deviation is simply a measure of dispersion. obviously, the lower the standard deviation, the more likely a researcher will reject the hypothesis that a given average risk premium is zero. this raises the question, “how high should we expect the average risk premium to be relative to its standard deviation?” there is no simple answer. the ratio [e(rp) / a(w)] measures the excess return per unit of standard deviation. from portfolio theory, we know that efficient portfolios would maximize this ratio for each standard deviation. for an efficient portfolio, this reward to total risk measure is the slope of the capital market line.’ the slope of the capital market line depends on investors’ aggregate risk aversion, the reward demanded per unit of risk. theory provides no direct guidance on the size of this slope-only that investors demand a positive risk premium. let us suppose we have identified a mean-variance efficient portfolio and wish to test whether the average risk premium is positive in a particular month, say august. in this context, we can state a corollary research question, “is the slope of the capital market line (the market price of risk) sufficient to yield a f-statistic of 1.645 in each calendar month?“8 the expectation that the slope will be sufficient to yield a t-statistic of 1.645 over all arbitrary calendar months and all arbitrary time periods seems to be an unreasonable restriction on the data. moreover, since the portfolios in our study and others are not necessarily efficient, the risk premium to standard deviation ratios will almost certainly be smaller than those of efficient portfolios. in well functioning capital markets, investors will price such portfolios to yield a risk premium sufficient to compensate for risk. the failure to reject the hypothesis that average risk premia equal zero says more about the power of the statistical tests to distinguish between zero and some positive number in small samples than it does about the true risk premium. another important component of empirical tests is the level of statistical significance demanded by researchers. a five or ten percent level of significance represents an aversion on the part of researchers to making type i errors (rejecting the null hypothesis when it is true). apparently, researchers would prefer to avoid inferring that there is a positive risk premium when, in fact, there is none. however, investors may be more concerned about making type ii errors (accepting the null hypothesis of a zero average risk premia when, in fact, there is a positive average risk premium). one could plausibly conclude that investors might take a chance of falsely believing there is a common stock risk premium if they can expect to earn an additional return of seven percent in the february through december period. this is certainly more plausible than investors choosing to forego additional return because it is not statistically different from zero at the five percent level.” conclusion our research indicates that reports of the death of the risk/return relationship are premature. inferences drawn from the evidence can be altered greatly through should individual investors avoid the stock market outside of january? 107 small changes in the way the empirical question is addressed. the precision of the mean risk premia can be increased by expanding the number of observations. since the test statistic is sensitive to the number of observations, a high excess return to risk ratio is required to find statistical significance when there are few observations. portfolio theory places no constraints on the size of the risk premium to standard deviation measure. one reasonable constraint is that it is positive, which is true even outside of january for the two equity portfolios examined in this study. investors should also note that the evidence suggests that the reward for bearing risk does vary over the calendar year. however, for equity investors that have longer holding periods, this variability in risk premia is of little importance and a rational investor would participate in the market throughout the year, not only during the month of january. in summary, we believe there is sufficient evidence to be skeptical about the conclusion that the risk/return relationship is severed outside of january. acknowledgments: the helpful comments of robin grieves, scott harrington and ted moore are gratefully acknowledged. the usual disclaimer applies. 1. 2. 3. 4. 5. 6. 7. notes see, for example, tinic and west (1984), gultekin and gultekin (1987), and chang and pinegar (1988a, 1988b). tinic and west find that january is the only month in which there is a reliable relationship between beta and expected returns in the context of the capital asset pricing model. gultekin and gultekin find an analogous result for risk measures associated with the arbitrage pricing theory. of course, the empirical difficulties associated with testing these models makes these conclusions more tenuous. see roll (1977) and shanken (1982) for discussions of these empirical difficulties. starting in 1982, this portfolio of small capitalization stocks contains some stocks that are listed on the american stock exchange and on nasdaq. stocks, bonds, bills and injlation is published annually by ibbotson and associates. for further details on the construction of these portfolios, the interested reader is directed to the 1989 yearbook. each monthly return (in decimal form) is augmented by one. the eleven-month return is then simply the product of these eleven individual monthly returns. since the term structure of interest rates is generally upward sloping, the choice of one-month bills may bias our results in favor of finding a positive and statistically significant equity risk premium. an investor would be able to capture an additional term premium in a typical year by investing in longer term government securities. the choice of one-month bills is made purely as a matter of convenience due to data availability. the properties of the relationship between standard deviation of returns and time are discussed by young (1971) and mcenally (1985). first order serial correlations for the common and small equity portfolios are .i0 and .16, respectively. positive serial correlation in returns will lead r-period standard deviations to be slightly more than jt times as large as single period standard deviations. of course, this is not literally true since the capital market line is developed in the context of a single period model. the slope of the theoretical capital market line is given by (e(r,) rf) / urm. it measures the ex ante reward to risk of the market portfolio. our measure is an ex 108 financial services review, l(2) 1991 8. 9. post measurement approximating this slope. if the risk-free rate were non-stochastic over all return intervals and r, represented the true market portfolio, our measure would provide a sample estimate ofthe slope of the capital market line. for our purposes, these differences are relatively minor. recall that 1.645 is the critical value for a one-tail test at the five percent level of significance. thus, the r-statistic must be at least 1.645 or we would fail to reject the null hypothesis that the risk premium is zero. for a general discussion of how researchers misuse tests of significance, see mccloskey (1985). references chang, eric, and j. michael pinegar. 1988. “a fundamental study of the seasonal risk-return relationship: a note,” journal offinance, 43: 1035-1039. chang, eric, and j. michael pinegar. 1988. “does the market reward risk in non-january months?” journal of portfolio management, 15: 55-57. gultekin, bulent, and mustafa n. gultekin. 1987. “stock return anomalies and tests of the apt,” journal offinance, 42: 1213-1224. mccloskey, donald. 1985. “the loss function has been mislaid: the rhetoric of significance tests,” american economic review, 75: 201-205. mcenally, richard w. 1985. “time diversification: the surest route to lower risk?” journal of portfolio management, 11: 24-26. roll, richard w. 1977. “a critique of the asset pricing theory’s tests: part i: on the past and potential testability of the theory,” journal of financial economics, 4: 129-176. shanken, jay. 1982. “the arbitrage pricing theory: is it testable?” journal offinance, 37: i129 1140. tinic, seha m., and richard west. 1984. “risk and return: january vs. the rest ofthe year,” journal of financial economics, 13: 561-574. young, william. 1971. “random walk of stock prices: a test of the variance-time function,” econometrica, 39: 797-812. academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2021 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 29, no. 4, 2021 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2021 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. the effects of fund commonality in mutual fund families on fund operating expenses and return correlations: evidence from u.s. equity mutual funds youngkyun parka,* acollege of business and economics, university of idaho, 875 perimeter drive ms 3161, moscow, id 838443161, usa abstract this article examines the effects of the commonality of mutual funds within a fund family, which is measured by common stock holdings and multi-fund management, on fund operating expenses and return correlations. for u.s. equity funds during the period of 2001–2006, we find that common stock holdings and multi-fund management are negatively related to fund operating expenses but positively related to the correlation of fund return residuals, which increases the correlation of fund returns. additionally, we find that the fund commonalities can have negative net effects on risk-adjusted returns of a portfolio with equity funds that have different investment objectives. © 2016 academy of financial services. all rights reserved. jel classification: g23; g11 keywords: mutual fund family; fund commonality; fund operating expense; fund return correlation 1. introduction many people in the united states use mutual funds as investment and retirement saving vehicles. over 90 million individual investors owned mutual funds and held about 87% of the total mutual fund assets in 2014 (investment company institute [ici], 2015). mutual fund investors often limit their transactions to one fund family to save search costs (ciccotello, miles, and walsh, 2007; sirri and tufano, 1998), to simplify fund management (elton, * corresponding author. tel.: �1-208-885-7154; fax: �1-208-885-5347. e-mail address: youngpark@uidaho.edu financial services review 25 (2016) 29–49 1057-0810/16/$ – see front matter © 2016 academy of financial services. all rights reserved. gruber, and green, 2007), or as a result of following a fund family’s reputation (ciccotello, greene, and walsh, 2007; gerken, starks, and yates, 2014). in addition, mutual fund investors have become sensitive to a fund’s expense ratio.1 fig. 1 shows the simple average expense ratios, the asset-weighted average expense ratios, and the difference between the two expense ratios of equity funds for the period of 1996 to 2014 (ici, 2015). the figure indicates that even if the mutual fund industry has recently provided more low-cost equity funds, equity fund investors have kept their sensitivity to fund expense ratios. because long-term returns are significantly affected by fund expenses (e.g., carhart, 1997; gil-bazo and ruiz-verdu, 2009; haslem, baker, and smith, 2008), it is natural that fund investors focus on a fund’s expense ratio. however, if a lower expense ratio results from increased fund commonality within a family, seeking only low-cost funds from one family may result in increasing an investor’s portfolio risk. in this study, we measure the commonality of funds by how much a fund holds the same stocks for other funds in the same family and whether a fund is managed by multi-fund managers who simultaneously manage multiple funds in the family. common stock holdings and multi-fund management have been recognized in the recent literature that emphasizes mutual funds as members of a fund family rather than as stand-alone entities (e.g., agarwal, ma, and mullally, 2015; choi, kahraman, and mukherjee, 2013; elton et al., 2007; yadav, 2010). common stock holdings and multi-fund management may reduce fund operating expenses (excluding 12b-1 fee), which are associated with portfolio management and administrative services. the two fund commonalities, however, may also enhance return correlations between funds and thereby increase an investor’s portfolio risk. this study examines the effects of common stock holdings and multi-fund management of funds on fund operating expenses and return correlations. common stock holdings and multi-fund management are assessed only for pairs of funds that have different investment objectives, by excluding pairs of funds that have the same investment objective. because of fig. 1. expense ratios in equity mutual funds. source: investment company institute fact book. 30 y. park / financial services review 25 (2016) 29–49 this exclusion, the fund commonalities are conservatively evaluated. we analyze 154 actively managed u.s. equity funds from 46 fund families for the period of 2001 to 2006, reconciling key data items (such as a fund’s stock holdings and managers) in the center for research in security prices (crsp), the thomson financial mutual fund holdings, and the morningstar databases. we find that common stock holdings and multi-fund management are negatively related to fund operating expenses but positively related to the correlation of fund return residuals, which increases the correlation of fund returns. for a portfolio that is constructed with funds in the same family, the findings indicate that an increase in a portfolio’s risk-adjusted return net of expenses (such as a sharpe ratio using expenseadjusted returns) because of decreased fund operating expenses can be wiped out by an increase in the portfolio risk because of increased return correlations. because of the opposite effects on a portfolio’s risk-adjusted return, we additionally investigate the effects of the fund commonalities on risk-adjusted returns of a portfolio that consists of equity funds with different investment objectives and find that the fund commonalities can have negative net effects on risk-adjusted returns of the portfolio. our findings contribute to the financial services literature in three ways. first, we expand previous research on mutual fund expenses by showing that a fund’s operating expense can be reduced by its common stock holdings and managers who simultaneously manage other funds in the family. second, we extend extant research on interfamily dynamics by showing that the fund commonalities within a family can deteriorate an investor’s portfolio diversification by increasing the correlation of fund returns. third, our findings provide practical implications to individual mutual fund investors. when individual investors construct a portfolio with low-cost equity funds within a family, they should be aware of an investment risk that fund commonalities that lower fund operating expenses can increase fund return correlations and thereby can reduce the portfolio’s risk-adjusted return. the remainder of the article is organized as follows. section 2 addresses hypotheses on relations between fund commonalities and operating expense and between fund commonalities and return correlations. section 3 describes the data and the sample. section 4 presents the methodology. section 5 presents empirical findings. section 6 concludes with a summary of findings and implications for individual mutual fund investors. 2. fund commonalities, operating expense, and return correlation funds in a family may have common features even if they have different investment objectives (elton et al., 2007). for example, different funds may hold the same stocks as a result of fund managers’ common view on individual companies from a common security selection process at a family level (elton et al., 2007) or of fund managers’ efforts to reduce the costs of monitoring the performance of the stocks held in their portfolios (shawky and smith, 2005). in addition, different funds may be managed simultaneously by a manager(s) in particular when the manager(s) exhibits superior past performance (agarwal, ma, and mullally, 2015). these common features would increase the commonality of funds in the family. the increased fund commonality, in turn, may influence the characteristics of funds, such as fund operating expenses and the correlation of fund returns. 31y. park / financial services review 25 (2016) 29–49 2.1. fund commonalities and operating expense mutual funds incur ongoing charges for managing fund assets. fund expenses paid out of fund assets consist of three broad components: management fees, other expenses including administrative fees, and 12b-1 fees (collins, 2003; khorana & servaes, 2007; latzko, 1999). the first and typically largest is the management fee paid to a fund manager, which compensates a fund manager for expenses incurred in providing services, including asset allocation and security selection. the second is other expenses including administrative fees, which mainly relate to recordkeeping and transactions services to shareholders.2 the third is the 12b-1 fees spent on advertising, marketing, and distribution services or commissions to brokers. a fund’s expense ratio may matter to fund families that want to increase assets under management because mutual fund investors increasingly purchase funds with low expenses (khorana & servaes, 2007). for example, collins (2007) documents that 90% of net new cash flow to stock funds went to those funds whose expense ratios were below the average expense ratio of stock funds offered in the marketplace from 1997 to 2006. thus, fund families or managers may increase their efforts to reduce a fund’s expense ratio. we focus on fund operating expenses, defined by fund expenses net of 12b-1 fees, because fund operating expenses account for the largest share of a fund’s expense ratio (collins, 2003) and because 12b-1 fees are mainly related to distribution fees for the services provided by brokers rather than the services by fund families (ici, 2015). fund operating expenses, consisting of management fees and other expenses including administrative fees (collins, 2003), may be reduced by an increase in common stock holdings or multi-fund management. first, common stock holdings may be able to lower a fund’s management fee. fund managers in the same family are likely to use the same research analysis produced by either internal analysts or external research firms (elton et al., 2007). the sharing of the research analysis within a family may increase common stock holdings across different funds, but it may also reduce the costs incurred in the security selection process, which are related to a fund’s management fees. thus, we expect that a fund’s common stock holdings are negatively related to its management fee. second, the management of multiple funds (i.e., the multifund management) may be able to lower a fund’s other expenses including administrative fees. fund families often assign multiple funds to the same portfolio manager. agarwal et al. (2015) find that, for the period of 1980 to 2012, 48% of mutual fund managers managed multiple funds simultaneously. this multi-fund management may enable a fund manager to reduce administrative fees by efficiently allocating her resources among the funds she manages. in this respect, we expect that multi-fund management is negatively related to a fund’s other expenses. thus, for a fund’s management fee and other expenses, we hypothesize the following: hypothesis 1-a: common stock holdings are negatively related to a fund’s management fee. hypothesis 1-b: multi-fund management is negatively related to a fund’s other expenses. 32 y. park / financial services review 25 (2016) 29–49 2.2. fund commonalities and return correlation except for common exposure to market and systematic factors, return correlations of funds in a family may be increased by nonmarket or nonsystematic factors that are shared between funds, such as common stock holdings and the management of multiple funds. to measure the correlation of fund returns caused by nonmarket or nonsystematic factors, we use the correlation of return residuals. following elton et al. (2007), we decompose the correlation of fund returns into two parts: (1) the return correlation caused by market and systematic factors, such as size, book-to-market, and momentum factors, and (2) the correlation of return residuals, which are not captured by market and systematic factors. then we focus on the correlation of return residuals and examine whether return residual correlations are related to common stock holdings and multi-fund management. first, common stock holdings of funds may enhance the commonality of funds in a family. elton et al. (2007) find that return correlations of funds are higher within families than between families and that about 60% of the increased correlation is because of common stock holdings. thus, we expect that common stock holdings are positively related to the correlation of return residuals, which increases the correlation of fund returns. second, multi-fund management may also increase the commonality of funds. funds within a family tend to pursue a similar investment strategy (e.g., lowor high-risk strategies) across different investment objectives (elton et al., 2007). similarly, funds managed by a multi-fund manager may adopt a similar investment strategy even if they have different investment objectives. a multi-fund manager may pursue her investment strategy for her different funds by selecting stocks with similar risk characteristics (but not necessarily the same stocks). thus, funds managed by a multifund manager would be more likely to have similar risk characteristics and, as a result, to have a higher correlation of return residuals than those managed by different fund managers. in this respect, for the correlation of return residuals, we hypothesize the following: hypothesis 2-a: common stock holdings are positively related to the correlation of return residuals. hypothesis 2-b: multi-fund management is positively related to the correlation of return residuals. 3. data and sample the data come from the crsp, the thomson financial mutual fund holdings, and the morningstar databases for the period of 2001 and 2006. first, we merge the crsp and the thomson financial mutual fund holdings and collect the information on funds’ stock holdings.3 to identify fund investment objectives, we use the investment company data, inc.’s (icdi) classifications that are matched with the strategic insight (si) classifications because the broad icdi categories enable us to develop a parsimonious model (table 1). our sample includes only actively managed u.s. equity funds that are classified as aggressive growth (ag), growth and income (gi), and long-term growth (lg). second, we use the morningstar to collect the information on fund managers and management fees, instead of 33y. park / financial services review 25 (2016) 29–49 the crsp. the crsp provides less reliable information on fund managers (massa, reuter, and zitzewitz, 2010) and reports management fees that include waivers and reimbursements, which can lead to negative management fees. we merge the morningstar and the merged dataset above, using funds’ tickers, cusips, net asset values, and returns that prior studies use to merge the morningstar and the crsp databases (e.g., berk and van binsbergen, 2015; pástor, stambaugh, and taylor, 2015).4 third, we include only funds that have distinct portfolios. mutual funds offer multiple shares, but they usually have the same portfolio managers, the same pool of securities, and the same returns before expenses and loads (zhao, 2004). the morningstar designates the oldest share class as a fund that has the distinct portfolio.5 relying on the information provided by the morningstar, we include only funds that are identified as ones that have distinct portfolios. fourth, we include only funds that existed for the period of 2001 to 2006. last, we exclude 14 observations where a fund’s expense ratio is less than its management fee.6 after all of these data-cleaning steps, we have a sample of 154 u.s. equity funds from 46 fund families. 4. methodology 4.1. fund commonalities: common stock holding and multi-fund management as proxies to capture the commonality of funds within a family, we use common stock holdings and multi-fund management (or the management of multiple funds). first, we define common stock holdings of funds within a family as follows: matchhijt � � k min�wkit, wkjt� (1) where wkit and wkjt are the weights of stock k held by funds i and j at year-end t, respectively, and the sum is taken over all the stocks in both funds. funds i and j have different investment objectives. a fund’s stock holding is determined by using the information reported on the date closest to the end of the year t. matchhijt ranges from 0 to 1. it is equal to 0 when the two funds do not have any common stock holdings in their portfolios at year-end t; it is equal to 1 when they hold exactly the same portfolios at year-end t. in addition, to indicate fund table 1 investment objective classifications for equity mutual funds icdi classification si classification aggressive growth (ag) aggressive growth (agg) and small company growth (scg) growth and income (gi) growth and income (gri) and income and growth (ing) long-term growth (lg) growth (gro) and growth mid-cap (gmc) notes: this article uses the investment company data, inc.’s (icdi) classifications that are matched with the strategic insight (si) classifications. the matching is conducted on the basis of a two-way table of icdi and si classifications for the period of january 2001 to june 2003. the icdi classifications are not available from the crsp database after july 2003. 34 y. park / financial services review 25 (2016) 29–49 i’s common stock holdings for all other investment objective funds in the same family, we modify eq. (1), which is defined for a pair of funds, as follows: comhit � �j�1 l matchhijt l (2) where matchhijt is the common stock holdings of funds i and j at year-end t and the sum is taken over l, which indicates the total number of funds in all other investment objectives that are different from fund i’s investment objective. thus, comhit indicates the average weight of common stock holdings in a portfolio of fund i at year-end t. second, we define multi-fund management as follows: matchmijt � �1, if there is at least one multi-fund manager between funds i and j 0, otherwise (3) where funds i and j have different investment objectives. to determine a fund’s managers, we use a measure of whether or not they manage the fund at the end of the year. matchmijt takes a value of 1 if there is at least one multi-fund manager between funds i and j at year-end t and 0 if there is no multi-fund manager between the funds at year-end t. in addition, to indicate whether or not fund i is managed by fund managers who simultaneously manage funds in other investment objectives, we modify matchmijt, which is defined for a pair of funds, as follows: commit � �1, if � j�1 l matchmijt � 0 0, if � j�1 l matchmijt � 0 (4) where matchmijt indicates the multi-fund management that is defined for a pair of funds i and j at year-end t and the sum is taken over l, which indicates the total number of funds in all other investment objectives that are different from fund i’s investment objective. commit takes a value of 1 if fund i is managed by at least one multi-fund manager at year-end t and 0 if fund i is managed by fund managers who do not manage other investment objective funds at year-end t. 4.2. model specification 4.2.1. fund commonalities and operating expenses to test the hypotheses on relationships between fund commonalities and fund operating expenses (h1-a and h1-b), we estimate the following model: expenseit � �0 � �1comhit � �2commit � �3zt � vit (5) where the dependent variable expenseit indicates fund i’s operating expense, management fee, or other expenses in year t. first, we use as the dependent variable a fund’s operating expense ratio that is computed by subtracting a 12b-1 fee from a fund’s expense ratio 35y. park / financial services review 25 (2016) 29–49 (collins, 2003). to calculate a fund’s operating expense ratio, we use fund expense ratios from the crsp database and 12b-1 fees from the morningstar database. this enables us to include more valid observations in the sample. for our sample, the crsp provides fewer missing values for a fund’s expense ratio than the morningstar, while the morningstar provides fewer missing values for a fund’s 12b-1 fee than the crsp. for example, if we use the data on 12b-1 fees from the crsp, we would lose about 47% of the sample. second, we use a fund’s management fee as the dependent variable. for management fees, we use the morningstar database because the management fee provided by the crsp can be offset by fee waivers and reimbursements, which can lead to negative management fees. these negative management fees might distort a relationship between fund commonalties and management fees. third, we use a fund’s other expense ratio as the dependent variable. a fund’s other expenses include fees for administrative and business services, such as legal, accounting, and administrative services (collins, 2003; mahoney, 2004). a fund’s other expense ratio is computed by subtracting a management fee from a fund’s operating expense ratio. eq. (5) also includes a set of control variables (z) at fund and family levels. following prior studies, we first include a fund’s annual turnover ratio that represents trading activity. a fund with a higher turnover ratio is likely to have a higher expense ratio because of active trading (malhotra and mcleod, 1997). the higher expense ratio caused by the active trading, however, may result from an increase in a fund’s other expenses because they include registration fees on the fund shares sold each year and custodial fees that cover the costs of settling trades (latzko, 2003). second, we control for a fund’s load, using a dummy variable, which indicates whether or not a fund has a sales load (either a front-end or back-end load). the variable takes a value of 1 if a fund has any sales loads and 0 otherwise. third, we control for a fund’s investment objective, using two dummy variables with a base category of ag. fourth, we control for fund size and family size, using the natural logarithm of one plus a fund’s year-end total net assets (in billions) and the natural logarithm of one plus the year-end total net assets (in billions) managed by a family, respectively. last, we include family fixed effects, which control for unobserved time-invariant family characteristics such as different strategic behaviors at a family level (e.g., clare, o’sullivan, and sherman, 2014; gaspar, massa, and matos, 2006), and year fixed effects. 4.2.2. fund commonalities and return residual correlation to test the hypotheses on relationships between fund commonalities and the correlation of return residuals (h2-a and h2-b), we first decompose the correlation of returns into two parts, following elton et al., (2007): corr�ri, rj� � corr�fi, fj� � corr�ei, ej� (6) where corr(ri, rj) is the correlation of returns of funds i and j; corr(fi, fj) is the correlation of fund returns caused by systematic movements; and corr(ei, ej) is the correlation of return residuals. we calculate corr(ei, ej) with residuals that are obtained from estimating the carhart four-factor model (carhart, 1997): 36 y. park / financial services review 25 (2016) 29–49 rit � rft � �i � �iem�emt� � �ismb�smbt� � �ihml�hmlt� � �imom�momt� � �it (7) where rit is the return of fund i in month t; rft is the one-month treasury-bill rate in month t; emt is the monthly excess return on the market, which is calculated by the monthly return of the crsp value-weighted index less the one-month treasury-bill rate in month t; smbt, hmlt, and momt denote the size-factor, the book-to-market factor, and the momentum factor in month t, respectively; and �it is a random error. second, we regress return residual correlations (corr(ei, ej)) on the average common stock holding and the multi-fund management between funds as follows: corr�ei, ej� � �0 � �1amatchhij � �2amatchmij � �3w � uij (8) where amatchhij is the average common stock holding between funds i and j for the period of 2001 to 2006; amatchmij is a dummy variable to indicate whether there exists any multi-fund management between funds i and j for the period; and w indicates a set of control variables at fund and family levels.7 we control for funds’ investment objectives, using two dummy variables with a base category of a combination of gi and lg. in addition, at a family level, we include (1) the natural logarithm of the total number of funds having distinct portfolios that are offered in ag, gi, and lg investment objectives by a family over the period and (2) family size defined by the natural logarithm of one plus the average total net assets (in billions) managed by a family over the period. last, we include family fixed effects, which control for unobserved time-invariant family characteristics. 5. results 5.1. summary statistics table 2 presents summary statistics of the sample for the period of 2001 to 2006. the average expense ratio is 1.17%; the average 12b-1 fee is 0.17%; and the average operating expense ratio is 1.00%. a fund’s management fee and other expenses, on average, account for 71% and 29%, respectively, of a fund’s operating expense ratio. funds hold, on average, 17% of their portfolios with the same stocks across other funds in different investment objectives. about 39% of the funds in the sample are managed by multi-fund managers over the period. most funds in the sample have sales loads. this is because we use funds that have distinct portfolios defined by the morningstar and they are likely to be class a shares that generally have a front-end load. table 2 also provides summary statistics by fund investment objectives: ag, gi, and lg. funds seeking ag, on average, have a higher expense ratio, a higher operating expense ratio, and a higher management fee than those seeking gi or lg. however, 12b-1 fees and other expenses are similar across the fund investment objectives. in addition, funds seeking ag have significantly less common stock holding than those seeking gi or lg, but ag funds are more likely to be managed by multi-fund managers than are lg funds.8 last, the average total net asset of ag funds is less than that of gi or lg funds. 37y. park / financial services review 25 (2016) 29–49 table 2 sample summary statistics observation mean standard deviation min max full sample (154 funds) expense ratio (%) 903 1.17 0.34 0.40 2.31 12b-1 fee (%) 913 0.17 0.22 0.00 1.00 operating expense (%) 901 1.00 0.29 0.15 1.90 management fee (%) 913 0.71 0.22 0.24 1.50 other expenses (%) 901 0.29 0.21 0.00 1.13 fraction of common stock holding in a fund portfolio 915 0.17 0.16 0.00 0.96 multi-fund management (yes � 1) 915 0.39 0.49 0.00 1.00 turnover ratio 897 0.77 0.73 0.03 7.76 load fund (yes � 1) 915 0.98 0.13 0.00 1.00 fund size ($ billion) 915 3.88 10.61 0.001 120.84 family size ($ billion) 276 52.11 158.43 0.17 1,149.30 aggressive growth (43 funds) monthly gross return (%) 3,096 0.88 5.50 �34.28 25.95 expense ratio (%) 254 1.33 0.31 0.79 2.31 12b-1 fee (%) 255 0.19 0.27 0.00 1.00 operating expense (%) 253 1.13 0.22 0.36 1.80 management fee (%) 255 0.85 0.18 0.46 1.50 other expenses (%) 253 0.29 0.22 0.00 1.03 fraction of common stock holding in a fund portfolio 256 0.13 0.16 0.00 0.96 multi-fund management (yes � 1) 256 0.45 0.50 0.00 1.00 turnover ratio 253 0.76 0.63 0.04 3.58 load fund (yes � 1) 256 0.98 0.15 0.00 1.00 fund size ($ billion) 256 0.89 1.25 0.004 9.51 growth and income (43 funds) monthly gross return (%) 3,079 0.61 3.68 �14.39 18.46 expense ratio (%) 246 1.04 0.33 0.55 2.27 12b-1 fee (%) 252 0.16 0.19 0.00 1.00 operating expense (%) 245 0.88 0.30 0.32 1.90 management fee (%) 252 0.61 0.21 0.24 1.00 other expenses (%) 245 0.28 0.19 0.00 1.13 fraction of common stock holding in a fund portfolio 253 0.17 0.15 0.00 0.81 multi-fund management (yes � 1) 253 0.38 0.49 0.00 1.00 turnover ratio 243 0.54 0.39 0.03 2.17 load fund (yes � 1) 253 1.00 0.06 0.00 1.00 fund size ($ billion) 253 5.84 12.71 0.014 79.33 long-term growth (68 funds) monthly gross return (%) 4,877 0.46 4.99 �24.85 29.00 expense ratio (%) 403 1.15 0.33 0.40 2.09 12b-1 fee (%) 406 0.16 0.20 0.00 1.00 operating expense (%) 403 0.99 0.28 0.15 1.79 management fee (%) 406 0.69 0.20 0.31 1.00 other expenses (%) 403 0.30 0.21 0.00 1.13 fraction of common stock holding in a fund portfolio 406 0.20 0.15 0.00 0.81 multi-fund management (yes � 1) 406 0.35 0.48 0.00 1.00 turnover ratio 401 0.92 0.89 0.03 7.76 load fund (yes � 1) 406 0.98 0.15 0.00 1.00 fund size ($ billion) 406 4.53 12.00 0.001 120.84 notes: the sample consists of 154 actively managed u.s. equity funds from 46 fund families for the period of 2001 and 2006. summary statistics are obtained from fund-year observations except for family size and monthly gross return. summary statistics of family size are obtained from family-year observations and summary statistics of monthly gross return from fund-month observations. 38 y. park / financial services review 25 (2016) 29–49 5.2. regression results 5.2.1. effects of fund commonalities on fund operating expenses regression results in table 3 present that fund operating expenses are significantly negatively related to two fund commonalities: common stock holdings and multi-fund management.9 first, common stock holdings are significantly negatively related to a fund’s operating expense ratio. a 10% increase in common stock holding, for example, reduces a fund’s operating expense ratio by 1.3–1.4%. to further examine how common stock holdings are related to the components of fund operating expenses, we estimate eq. (5) for funds’ management fees (columns 3 and 4) and other expenses (columns 5 and 6). regression results show that common stock holdings are significantly negatively related to a fund’s management fee, supporting hypothesis h1-a. however, common stock holdings are not significantly related to a fund’s other expenses. thus, the results indicate that a decrease in a fund’s operating expense ratio caused by an increase in common stock holdings occurs through a decrease in its management fee, not through its other expenses. second, multi-fund management is significantly negatively related to a fund’s operating expense ratio (columns 1 and 2). for example, a fund’s operating expense ratio is reduced by 4.2–5.1% when a fund table 3 effects of fund commonalities on fund expenses operating expense management fee other expenses (1) (2) (3) (4) (5) (6) common stock holding (comh) �0.138** �0.134* �0.158** �0.076* 0.020 �0.047 (0.047) (0.063) (0.032) (0.035) (0.042) (0.060) multi-fund management (comm, yes � 1) �0.042** �0.051* 0.014 0.018 �0.056** �0.069** (0.014) (0.025) (0.012) (0.016) (0.013) (0.024) turnover ratio 0.022 0.019 �0.018** �0.012 0.041** 0.033* (0.014) (0.015) (0.007) (0.007) (0.012) (0.015) load fund (yes � 1) 0.035 �0.070 0.032 �0.014 0.001 �0.058 (0.087) (0.049) (0.033) (0.014) (0.065) (0.042) growth and income (gi) �0.157** �0.204** �0.180** �0.165** 0.023 �0.042* (0.019) (0.018) (0.015) (0.010) (0.018) (0.016) long-term growth (lg) �0.060** �0.068** �0.087** �0.079** 0.025 0.007 (0.018) (0.014) (0.013) (0.008) (0.016) (0.014) fund size �0.119** �0.105** �0.043** �0.028** �0.076** �0.076** (0.010) (0.012) (0.006) (0.006) (0.009) (0.011) family size �0.026** �0.083** �0.038** 0.011 0.012* �0.091** (0.005) (0.017) (0.004) (0.016) (0.005) (0.021) year fixed effects yes yes yes yes yes yes family fixed effects no yes no yes no yes observations 895 895 895 895 895 895 adjusted r2 0.426 0.641 0.468 0.807 0.148 0.392 notes: the table reports the results of the regressions of fund expenses on fund commonalities. the dependent variables are operating expense (columns 1 and 2), management fee (columns 3 and 4), and other expense (columns 5 and 6) of mutual funds. a fund’s operating expense is divided into a fund’s management fee and other expenses. fund size and family size indicate the natural logarithm of one plus a fund’s year-end total net assets (in billions) and the natural logarithm of one plus the year-end total net assets (in billions) managed by a family, respectively. robust standard errors are in parentheses. *p � 0.05, **p � 0.01. 39y. park / financial services review 25 (2016) 29–49 is managed by multi-fund managers who simultaneously manage other funds that have different investment objectives. in addition, regression results on management fees (columns 3 and 4) and other expenses (columns 5 and 6) show that multi-fund management is not significantly related to a fund’s management fee but significantly negatively related to a fund’s other expenses, supporting hypothesis h1-b. thus, the results indicate that a decrease in a fund’s operating expense ratio caused by multi-fund management occurs through a decrease in its other expenses, not through its management fee. for the effects of the control variables on fund operating expenses, table 3 shows similar results reported in prior studies. first, a fund’s turnover ratio, indicating trading activity, is positively related to a fund’s other expenses (columns 5 and 6), which include administrative fees such as registration fees on the fund shares sold and the costs of settling trades (latzko, 2003). second, a fund’s load is not significantly related to a fund’s operating expenses. this result is consistent with khorana and servaes (2007) claiming that a positive relationship between loads and fund expenses is driven by 12b-1 fees rather than by fund operating expenses. third, funds seeking gi or lg are likely to have a lower operating expense ratio than those seeking ag. the lower operating expense ratios of gi and lg funds result mainly from their lower management fees. last, fund size and family size are significantly negatively related to fund operating expenses, the negative relation that indicate economies of scale at fund and family levels, respectively (e.g., dowen and mann, 2004; latzko, 1999). in particular, consistent with elton, gruber, and blake (2012), we find that a negative relationship of other expenses and fund size is stronger than a negative relationship of management fee and fund size. in summary, the regression results in table 3 indicate that a fund’s operating expense ratio can be reduced by an increase in common stock holdings or by multi-fund management.10 the influences of the two commonalities on a fund’s operating expense ratio, however, occur through different channels: common stock holdings reduce a fund’s management fee while multi-fund management lowers a fund’s other expenses. 5.2.2. effects of fund commonalities on return residual correlations next, we examine whether common stock holdings and multi-fund management affect the correlation of fund returns within families. for doing this, we focus on the correlation of return residuals, which is a component of the correlation of fund returns (eq. 6). panel a of table 4 presents summary statistics for pairs of funds that have different investment objectives, such as ag-gi, ag-lg, and gi-lg. in particular, gi-lg funds have, on average, a higher return correlation but a lower return residual correlation than the other pairs of funds. the higher return correlations between gi and lg funds may result from their similar investment objectives (gi vs. lg), while the lower return residual correlations may be caused by a lower level of nonsystematic commonality between the funds. panel b of table 4 presents regression results of return residual correlations, which are obtained after estimating the carhart four-factor model (eq. 7). first, results in columns (1) and (2) show that average common stock holding (amatchh) and multi-fund management (amatchm) are significantly positively related to the correlation of return residuals, supporting hypotheses h2-a and h2-b, respectively. for example, a 10% increase in common stock 40 y. park / financial services review 25 (2016) 29–49 table 4 effects of fund commonalities on return residual correlation panel a: summary statistics of pairs of funds within families observation mean standard deviation min max return correlation: 256 0.82 0.18 �0.15 0.98 between ag and gi funds (ag-gi) 38 0.76 0.23 �0.15 0.91 between ag and lg funds (ag-lg) 100 0.80 0.16 �0.03 0.98 between gi and lg funds (gi-lg) 118 0.84 0.17 �0.07 0.98 return residual correlation: 256 0.16 0.25 �0.66 0.87 between ag and gi funds (ag-gi) 38 0.17 0.25 �0.31 0.69 between ag and lg funds (ag-lg) 100 0.21 0.24 �0.39 0.87 between gi and lg funds (gi-lg) 118 0.11 0.24 �0.66 0.58 average common stock holding 256 0.18 0.16 0.00 0.79 multi-fund management for the period (yes � 1) 256 0.17 0.37 0.00 1.00 family size ($ billion) 46 52.11 153.54 0.33 872.43 total number of funds offered by a family (tnfo) 46 3.83 3.95 2.00 27.00 notes: ag, gi, and lg indicate fund investment objectives of aggressive growth, growth and income, and long-term growth, respectively. average common stock holding indicates the average of the common stock holdings between funds for the period of 2001 to 2006. tnfo indicates the total number of funds having distinct portfolios that are offered in ag, gi, and lg investment objectives by a family over the period. panel b: regression results of return residual correlation (1) (2) (3) (4) (5) (6) average common stock holding (amatchh) 0.695** 0.767** 0.778** 0.789** 0.706** 0.775** (0.090) (0.118) (0.114) (0.141) (0.091) (0.117) multi-fund management for the period (amatchm, yes � 1) 0.191** 0.195** 0.249** 0.223** (0.038) (0.057) (0.054) (0.086) amatchh � amatchm �0.278 �0.100 (0.163) (0.222) amatchm1 [a] 0.164** 0.133* (0.044) (0.060) amatchm2 [b] 0.281** 0.395** (0.042) (0.089) fund objectives: ag-gi 0.153** 0.136** 0.159** 0.138** 0.154** 0.130* (0.044) (0.052) (0.045) (0.053) (0.044) (0.051) fund objectives: ag-lg 0.145** 0.111** 0.157** 0.114** 0.149** 0.121** (0.028) (0.034) (0.030) (0.036) (0.028) (0.035) log(tnfo) 0.078* 0.077* 0.065* (0.032) (0.033) (0.033) family size �0.050** �0.050** �0.044** (0.014) (0.014) (0.014) family fixed effects no yes no yes no yes observations 256 256 256 256 256 256 adjusted r2 0.387 0.451 0.390 0.449 0.391 0.466 p-value for f-test: [a] � [b] 0.025* 0.007** notes: ag, gi, and lg indicate fund investment objectives of aggressive growth, growth and income, and long-term growth, respectively. tnfo indicates the total number of funds having distinct portfolios that are offered in ag, gi, and lg investment objectives by a family over the period of 2001 to 2006. family size indicates the natural logarithm of one plus the average total net assets (in billions) managed by a family for the period. observations indicate the number of different pairs of funds. robust standard errors are in parentheses. *p � 0.05, **p � 0.01. 41y. park / financial services review 25 (2016) 29–49 holdings increases the correlation of return residuals by 7%, while the multi-fund management between funds increases the correlation of return residuals by 19%. second, we include an interaction term of average common stock holdings (amatchh) and multi-fund management (amatchm) in a model to see if they interact with each other on return residual correlations (columns 3 and 4). the coefficient of the interaction term (matchh � matchm), however, is not statistically significant while the main effects of the two commonalities remain significant. the results, thus, indicate that common stock holdings and multi-fund management independently influence the correlation of return residuals. third, we further examine whether or not different multi-fund management structures are related to return residual correlations. we divide multi-fund management into two categories (columns 5 and 6). one is that a pair of funds is managed by at least one multi-fund manager, but not managed by exactly the same managers (matchm1). the other is that a pair of funds is managed by exactly the same fund manager(s) (matchm2). because the management by exactly the same managers is more likely to increase similarity between funds, we expect that the coefficient of matchm2 is higher than that of matchm1. matchm1 and matchm2 are dummy variables that have the same base category, which is that a pair of funds is managed by two completely different groups of fund managers. results in columns (5) and (6) show that both variables, matchm1 and matchm2, are significantly positively related to return residual correlations and that the coefficient of matchm2 is significantly greater than that of matchm1 (p-values for f-tests � 0.05), as expected. in summary, the regression results in table 4 show that common stock holdings and multi-fund management are positively related to the correlation of return residuals. the results indicate that, when an investor constructs a portfolio with equity funds that have different investment objectives within a family, the fund commonalities can deteriorate the investor’s portfolio diversification by increasing the correlation of return residuals, which is a component of the correlation of fund returns. 5.3. fund commonalities and risk-adjusted returns regression results presented in tables 3 and 4, taken together, indicate that common stock holdings and multi-fund management decrease funds’ operating expense ratios, but increase their return correlations. these results imply that an increase in a portfolio’s risk-adjusted return net of expenses caused by a decrease in fund operating expenses can be wiped out by an increase in the portfolio risk resulting from increased return correlations. in this section, we examine net effects of the fund commonalities on a portfolio’s sharpe ratio net of expenses. for the examination, we employ a hypothetical investor who constructs a diversified portfolio with equity funds that have ag, gi, and lg investment objectives, as suggested by moreno and rodriguez (2013). suppose that a portfolio consists of two different investment objective funds with equal weights and that the two funds are not managed by multi-fund managers. for each portfolio, we use a set of assumptions presented in panel a of table 5. first, we compute a portfolio’s annual gross return (before expenses) and expense ratio, using the average monthly gross returns and the average expense ratios presented by fund investment objectives in table 2. second, we calculate a portfolio’s risk, using annualized 42 y. park / financial services review 25 (2016) 29–49 standard deviations of monthly gross returns in table 2 and the average return correlations in panel a of table 4. third, we use a risk-free rate. last, we calculate a portfolio’s sharpe ratio net of expenses, which is defined by sp � �rp � erp � rf�/�p where rp is a portfolio’s annual gross return; erp is a portfolio’s weighted expense ratio with equal weights; rf is a risk-free rate of 2.57%, which is the average rate of one month treasury bills for 2001–2006; and �p is a portfolio’s risk. panels b and c present changes in a portfolio’s expense ratio, risk, and sharpe ratio because of an increase in fund commonality such as a 10% increase in common stock table 5 increased fund commonalities and risk-adjusted returns panel a: annual gross returns, expense ratios, risks, and sharpe ratios of portfolios constructed with different investment objective funds portfolio annual gross return (rp) portfolio expense ratio (erp) portfolio risk (�p) portfolio sharpe ratio (sp) ag-gi 9.33% 1.19% 14.96% 0.372 ag-lg 8.38% 1.24% 17.25% 0.265 gi-lg 6.65% 1.10% 14.44% 0.207 notes: the panel presents a set of assumptions on three different portfolios of ag-gi, ag-lg, and gi-lg funds. ag, gi, and lg indicate fund investment objectives of aggressive growth, growth and income, and long-term growth, respectively. each portfolio consists of two different investment objective funds with equal weights. panel b: common stock holding increased by 10% portfolio: ag-gi ag-lg gi-lg (1) (2) (1) (2) (1) (2) �portfolio expense ratio �1.14% �1.14% �1.15% �1.15% �1.14% �1.14% �portfolio risk 0.36% 0.45% 0.22% �sharpe ratio 0.24% �0.12% 0.31% �0.14% 0.42% 0.20% additional gross return (basis points) 2.0 2.1 0.7 notes: the panel reports changes in a portfolio’s expense ratio, risk, and sharpe ratio due to a 10% increase in common stock holding between funds in a portfolio. additional gross return indicates how much additional gross return is needed to have the same increase in a sharpe ratio (reported in column 1 for each portfolio) caused by a decrease in a portfolio’s expense ratio. panel c: multi-fund management by exactly the same fund managers portfolio: ag-gi ag-lg gi-lg (1) (2) (1) (2) (1) (2) �portfolio expense ratio �4.33% �4.33% �4.36% �4.36% �4.35% �4.35% �portfolio risk 1.73% 2.30% 1.27% �sharpe ratio 0.92% �0.80% 1.18% �1.09% 1.60% 0.32% additional gross return (basis points) 9.7 10.6 3.9 notes: the panel reports changes in a portfolio’s expense ratio, risk, and sharpe ratio when funds in a portfolio are managed by exactly the same fund managers. additional gross return indicates how much additional gross return is needed to have the same increase in a sharpe ratio (reported in column 1 for each portfolio) caused by a decrease in a portfolio’s expense ratio. 43y. park / financial services review 25 (2016) 29–49 holdings (panel b) or multi-fund management by exactly the same fund managers (panel c). the results of portfolios of ag-gi, ag-lg, and gi-lg funds are obtained by combining the assumptions in panel a and the regression results in tables 3 and 4. we compute a change in a portfolio’s expense ratio, using the results on fund operating expenses (reported with year and family fixed effects) in column (2) of table 3 and assuming that a 12b-1 fee is not affected by a change in fund commonality.11 next, we calculate a change in a portfolio’s risk, using the results on return residual correlations (reported with family fixed effects) in column (6) of panel b, table 4. to investigate a change in a sharpe ratio net of expenses (panels b and c, table 5), we examine two cases: one considers only a decrease in a portfolio’s expense ratio (column 1), and the other considers both a decrease in a portfolio’s expense ratio and an increase in a portfolio’s risk (column 2). the first case is to reflect that an investor disregards an enhanced portfolio risk resulting from an increase in the fund commonalities, while the second case is to reflect that an investor incorporates the enhanced risk into her portfolio return. panel b of table 5 presents changes in a portfolio’s expense ratio, risk, and sharpe ratio when common stock holdings between two funds in a portfolio are increased by 10%. first, for a portfolio of ag-gi funds, a 10% increase in common stock holdings has a negative net effect on the portfolio’s risk-adjusted return. the increased fund commonality decreases the portfolio’s expense ratio by 1.14%, a decrease which raises the portfolio’s sharpe ratio by 0.24% (column 1). however, the increased common stock holding also increases the portfolio risk by 0.36%; as a result, the portfolio’s sharpe ratio is eventually decreased by 0.12% (column 2). thus, the increased common stock holding results in a negative net effect on the portfolio’s risk-adjusted return. for an increase in the sharpe ratio caused by the decreased portfolio expense ratio, we further investigate how much additional gross return is needed to have the same increase in the risk-adjusted return. we find that the portfolio should additionally earn a 2.0 basis point return to have the same increase of 0.24% in the sharpe ratio (reported in column 1). second, for a portfolio of ag-lg funds, a 10% increase in common stock holdings results in a negative net effect on the portfolio’s risk-adjusted return, a result which is similar to that of ag-gi funds. the increased fund commonality reduces the portfolio’s expense ratio by 1.15%. the lower portfolio expense ratio increases the portfolio’s sharpe ratio by 0.31% (column 1). however, the increased fund commonality increases the portfolio risk by 0.45% (column 2). as a result, the portfolio’s sharpe ratio is decreased by 0.14% (column 2), indicating a negative net effect on the portfolio’s riskadjusted return. for an increase of 0.31% in the sharpe ratio because of the decreased portfolio expense ratio, the portfolio needs an additional gross return of 2.1 basis points. third, in contrast to portfolios of ag-gi and ag-lg funds, a 10% increase in common stock holding results in a positive net effect on the risk-adjusted return of a portfolio of gi-lg funds. the increased fund commonality reduces the portfolio’s expense ratio by 1.14%, which increases the sharpe ratio by 0.42% (column 1). however, the increased fund commonality raises the portfolio risk by 0.22% (column 2). as a result, the sharpe ratio is increased only by 0.20% (column 2). for an increase of 0.42% in the sharpe ratio because of the decreased portfolio expense ratio, the portfolio needs an additional gross return of 0.7 basis points. in summary, an increase in common stock holdings of a portfolio decreases the portfolio’s expense ratio, but increases its risk. the increased portfolio risk can cause a 44 y. park / financial services review 25 (2016) 29–49 negative net effect on the portfolio’s risk-adjusted return, as shown in the portfolios of ag-gi and ag-lg. in addition, because of the increased portfolio risk, the portfolio needs an additional return to have the same sharpe ratio that an investor would have from the decreased portfolio expense ratio. panel c of table 5 presents changes in a portfolio’s expense ratio, risk, and sharpe ratio when funds in a portfolio are managed by exactly the same fund managers. results in panel c are similar to those reported in panel b, though the magnitudes of the changes are greater in panel c than in panel b. the multi-fund management reduces a portfolio’s expense ratio by 4.43% for ag-gi, 4.36% for ag-lg, and 4.35% for gi-lg funds, and the decreased expense ratios raise the portfolios’ sharpe ratios by 0.92%, 1.18%, and 1.60%, respectively (column 1). however, the multi-fund management also increases the portfolio risk by 1.73% for ag-gi, 2.30% for ag-lg funds, and 1.27% for gi-lg funds. as a result, the sharpe ratios are decreased by 0.80% for ag-gi funds and 1.09% for ag-lg funds, while the sharpe ratio of gi-lg funds is increased by 0.32%. the results indicate that the multi-fund management has negative net effects on risk-adjusted returns for portfolios of ag-gi and ag-lg funds, while it has a positive net effect on risk-adjusted returns for a portfolio of gi-lg funds. finally, each portfolio needs an additional return to have the same sharpe ratio that the portfolio would have from the decreased portfolio expense ratio. the ag-gi portfolio needs an additional gross return of 9.7 basis points; the ag-lg portfolio needs an additional gross return of 10.6 basis points; and the gi-lg portfolio needs an additional gross return of 3.9 basis points. in summary, similar to an increase in common stock holdings, the multi-fund management by exactly the same managers decreases a portfolio’s expense ratio, but increases its risk. the increased risk can bring a negative net effect on the portfolio’s risk-adjusted return, as shown in the portfolios of ag-gi and ag-lg. 6. conclusion in this study, we provide extensive evidence that the commonality of funds within a family can influence funds’ operating expenses and return correlations. using common stock holdings and multi-fund management as proxies for the fund commonality in a family, we have analyzed a sample of 154 actively managed u.s. equity funds from 46 fund families for the period of 2001 to 2006. we first find that common stock holdings and multi-fund management can decrease funds’ operating expenses and that the influences on fund operating expenses occur through different channels: common stock holdings reduce a fund’s management fee while multi-fund management lowers a fund’s other expenses that include administrative fees. second, we find that common stock holdings and multi-fund management can increase the correlation of return residuals, which raises the correlation of fund returns. third, because the results indicate opposite effects of the fund commonalities on risk-adjusted returns, we have investigated net effects on risk-adjusted returns of portfolios with equity funds that have different investment objectives and found that the fund commonalities can have negative net effects on the portfolios’ risk-adjusted returns. this study enhances an understanding of the commonality of funds within a family. we expand previous research by showing that fund commonalities within a family can influence 45y. park / financial services review 25 (2016) 29–49 fund operating expenses and return correlations. this study also provides practical implications to individual mutual fund investors. when investors construct a portfolio with different low-cost equity funds from a single family, they should be aware of an investment risk that fund commonalities that lower fund operating expenses can increase the correlation of fund returns and, as a result, can reduce the portfolio’s risk-adjusted return. in other words, when expense-conscious investors invest in equity funds in the same family, they need to pay additional attention to fund commonalities such as funds’ common stock holdings and management structure. notes 1 mutual fund investors have also invested largely in low-cost funds. for example, at year-end 2014, equity funds with expense ratios in the lowest quartile held 74% of the funds’ total net assets, whereas those with expense ratios in the upper three quartiles held only 26% (ici, 2015). 2 the transaction services include providing statements and reports, disbursing dividends, and paying state and local taxes and custodial, legal, audit, registration, and directors’ fees (collins, 2003; latzko, 1999). 3 to merge the crsp and the thomson financial mutual fund holdings, we use the mflinks files from the wharton research data services (wrds). 4 when using fund returns to merge the two databases, we focus on monthly returns in december and compare fund returns reported from the two databases. when the difference in returns is less than or equal to 10 basis points in any year, we include the funds in the sample. however, when the return difference is greater than 10 basis points, we drop the funds from the merging process. 5 the morningstar adds a note that funds having distinct portfolios are likely to have the a share class. 6 an inclusion of the 14 dropped observations does not make regression results significantly different from the results reported in this article. 7 year-end stock holdings that are used to calculate the average common stock holding between funds are not matched with monthly fund returns that are used to compute the correlations of return residuals. however, using the annual holdings data would not significantly affect results reported in this study because it focuses on common stock holdings of funds in the same family rather than funds’ trading behavior in which a more precise measurement of the timing of trades, such as monthly holdings data, is needed (elton et al., 2010). 8 for common stock holdings by fund investment objectives, pairwise t-test results show that common stock holdings are significantly different among the three investment objectives: the difference between ag and lg funds is significant at 1% level (t-statistics�5.03) and the differences between ag and gi and between gi and lg funds are significant at 5% level (t-statistics � 2.41 for ag-gi and t-statistics � 2.45 for gi-lg). for the multi-fund management, pairwise t-test results show that the difference between ag and lg funds are significant at 5% level (t-statistics � 2.31) 46 y. park / financial services review 25 (2016) 29–49 but the differences between ag and gi funds and between gi and lg funds are not statistically significant. 9 common stock holding and multi-fund management are weakly positively correlated with a correlation coefficient of 0.137 (p � 0.001). the positive correlation, however, does not cause a serious multi-collinearity in estimating the regression models of eqs. (5) and (8). for example, the values of the variance inflation factor (vif) for both variables are lower than 10, which is usually used to indicate whether a serious multi-collinearity problem exists or not (e.g., hair et al., 1995; kennedy, 1992; neter, wasserman, and kutner, 1989). 10 we check the robustness of the results presented in table 3, using the data on fund expense ratios from the morningstar, instead of the crsp. appendix a reports regression results. the results are not significantly different from those reported in table 3. thus, our results are robust to the use of different data sources for fund expense ratios. 11 the assumption of no change in 12b-1 fees would be valid because a 12b-1 fee that relates to distribution services is less likely to be affected by a change in fund holdings or fund management structure. appendix a robustness check for the effects of fund commonalities on fund expenses operating expense management fee other expenses (1) (2) (3) (4) (5) (6) common stock holding �0.248** �0.275** �0.158** �0.076* �0.037 �0.128** (0.047) (0.064) (0.032) (0.035) (0.036) (0.043) multi-fund management (yes � 1) �0.067** �0.054* 0.014 0.018 �0.074** �0.054** (0.016) (0.025) (0.012) (0.016) (0.012) (0.020) turnover ratio �0.018 �0.001 �0.018** �0.012 �0.002 0.016* (0.012) (0.010) (0.007) (0.007) (0.009) (0.008) load fund (yes � 1) 0.044 �0.093 0.032 �0.014 0.012 �0.065 (0.103) (0.066) (0.033) (0.014) (0.081) (0.059) growth and income (gi) �0.200** �0.186** �0.180** �0.165** �0.020 �0.038** (0.019) (0.017) (0.015) (0.010) (0.016) (0.013) long-term growth (lg) �0.087** �0.092** �0.087** �0.079** �0.001 �0.030** (0.019) (0.013) (0.013) (0.008) (0.016) (0.011) fund size �0.088** �0.068** �0.043** �0.028** �0.044** �0.042** (0.010) (0.012) (0.006) (0.006) (0.007) (0.008) family size �0.033** 0.004 �0.038** 0.011 0.004 0.008 (0.005) (0.015) (0.004) (0.016) (0.004) (0.012) year fixed effects yes yes yes yes yes yes family fixed effects no yes no yes no yes observations 817 817 895 895 817 817 adjusted r2 0.432 0.701 0.468 0.807 0.103 0.546 notes: this table is to check the robustness of the results presented in table 3, using the data on fund expense ratios from the morningstar, instead of the crsp. the dependent variables are operating expense (columns 1 and 2), management fee (columns 3 and 4), and other expense (columns 5 and 6) of mutual funds. fund size and family size indicate the natural logarithm of one plus a fund’s year-end total net assets (in billions) and the natural logarithm of one plus the year-end total net assets (in billions) managed by a family, respectively. robust standard errors are in parentheses. *p � 0.05, **p � 0.01. 47y. park / financial services review 25 (2016) 29–49 acknowledgments the author is grateful to sojung park and jesse thomas and would like to thank the editor and anonymous referees for constructive comments. references agarwal, v., ma, l., & mullally, k. (2015). managerial multitasking in the mutual fund industry. working paper (available at http://ssrn.com/abstract�1741282). berk, j. b., & van binsbergen, j. h. (2015). measuring skill in the mutual fund industry. journal of financial economics, 118, 1–20. http://dx.doi.org/10.1016/j.jfineco.2015.05.002. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. choi, d., kahraman, c. b., & mukherjee, a. (2013). performance-chasing behavior and mutual funds: new evidence from multi-fund managers. working paper (available at http://ssrn.com/abstract�2219846). ciccotello, c. s., miles, j. a., & walsh, l. s. (2006). should investors choose funds from focused families? financial services review, 15, 247–264. ciccotello, c. s., greene, j. t., & walsh, l. s. (2007). supermarket distribution and brand recognition of open-end mutual funds. financial services review, 16, 309–326. clare, a., o’sullivan, n., & sherman, m. (2014). family status and mutual fund performance. j of asset management, 15, 163–175. collins, s. (2003). the expenses of defined benefit pension plans and mutual funds fees. investment company institute perspective, 9. collins, s. (2007). fees and expenses of mutual funds 2006, investment company institute research fundamentals, 16. dowen, r. j., & mann, t. (2004). mutual fund performance, management behavior, and investor costs. financial services review, 13, 79–91. elton, e., gruber, m. j., & green, t. c. (2007). the impact of mutual fund family membership on investor risk. journal of financial and quantitative analysis, 42, 257–278. elton, e. j., gruber, m. j., blake, c. r., krasny, y., & ozelge, s. o. (2010). the effect of holdings data frequency on conclusions about mutual fund behavior. journal of banking & finance, 34, 912–922. elton, e. j., gruber, m. j., & blake, c. r. (2012). does mutual fund size matter? the relationship between size and performance. review of asset pricing studies, 2, 31–55. gaspar, j.-m., massa, m., & matos, p. (2006). favoritism in mutual fund families? evidence on strategic cross-fund subsidization. journal of finance, 61, 73–104. gil-bazo, j., & ruiz-verdu, p. (2009). the relation between price and performance in the mutual fund industry. journal of finance, 64, 2153–2183. gerken, w. c., starks, l. t., & yates, m. c. (2014). the importance of family: the role of mutual fund family reputation in investment decisions. working paper (available at ssrn: http://ssrn.com/abstract�2022503). hair, j. f., jr., anderson, r. e., tatham, r. l., & black, w. c. (1995). multivariate data analysis (3rd ed.). new york, ny: macmillan. haslem, j. a., baker, h. k., & smith, d. m. (2008). performance and characteristics of actively managed retail equity mutual funds with diverse expense ratios. financial services review, 17, 49–68. investment company institute. (2015). 2015 investment company fact book (55th edition). washington, dc. kennedy, p. (1992). a guide to econometrics. oxford: blackwell. khorana, a., & servaes, h. (2007). competition and conflicts of interest in the u.s. mutual fund industry. working paper (available at http://faculty.london.edu/hservaes/competition.pdf). latzko, d. a. (1999). economies of scale in mutual fund administration. journal of financial research, 22, 331–339. latzko, d. a. (2003). mutual fund expenses: an econometric investigation. working paper, pennsylvania state university, york campus. 48 y. park / financial services review 25 (2016) 29–49 mahoney, p. g. (2004). manager-investor conflicts in mutual funds. journal of economic perspective, 18, 161–182. malhotra, d. k., & mcleod, r. w. (1997). an empirical analysis of mutual fund expenses. journal of financial research, 20, 175–190. massa, m., reuter, j., & zitzewitz, e. (2010). when should firms share credit with employees? evidence from anonymously managed mutual funds. journal of financial economics, 95, 400–424. moreno, d., & rodríguez, r. (2013). optimal diversification across mutual funds. applied financial economics, 23, 119–122. neter, j., wasserman, w., & kutner, m. h. (1989). applied linear regression models. homewood, il: irwin. pástor, l’uboš, stambaugh, r. f. & taylor, l. a. (2015). scale and skill in active management. journal of financial economics, 116, 23–45. shawky, h. a., & smith, d. m. (2005). optimal number of stock holdings in mutual fund portfolios based on market performance. financial review, 40, 481–495. sirri, e. r., & tufano, p. (1998). costly search and mutual fund flows. journal of finance, 53, 1589–1622. yadav, v. (2010). portfolio matching by multi-fund managers: effects of fund performance and flows. insead university working paper. zhao, x. (2004). why are some mutual funds closed to new investors? journal of banking and finance, 28, 1867–1887. 49y. park / financial services review 25 (2016) 29–49 local creative culture and dividend policy erdem ucara, arsenio staera,* acalifornia state university fullerton, mihaylo college of business and economics, department of finance, 800 n. state college boulevard, fullerton, ca 92831, usa abstract this paper examines the role of local risk-taking propensity on dividend demand by using local creative culture as a measure of local risk-taking. we find that firms located in areas with a strong creative culture are less likely to pay and initiate dividends and exhibit lower levels of dividend yield. the empirical findings also remain robust after addressing endogeneity and a series of robustness checks. furthermore, our paper highlights the local component of corporate dividend policies and offers additional evidence supporting dividend catering theory. our results underscore the importance of cultural determinants of investors’ risk-taking for the financial industry participants. © 2018 academy of financial services. all rights reserved. jel classification: g35; g40 keywords: dividend policy; creative culture; risk-aversion; dividend catering 1. introduction this study investigates the role of risk-taking for local dividend demand and corporate dividend policies. prior literature has investigated the determinants of dividend demand and investors’ payout preferences since miller and modigliani (1961). previous studies suggest that risk aversion plays an important role for investors’ choice between dividends and capital gains (i.e., gordon, 1963; lintner, 1963) and the financial planning process (guillemette and nanigian, 2014). moreover, recent studies show the impact of local factors on dividend demand and different corporate payout policies that cater to investor demand (i.e., becker, ivković, and weisbenner, 2011; ucar, 2016). we introduce a new measure of risk-taking to * corresponding author. tel.: �1-657-278-3957; fax: �1-657-278-216. e-mail address: astaer@fullerton.edu (a. staer) financial services review 27 (2018) 367-389 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. dividend literature and examine the impact of risk-taking behavior associated with local culture on dividend demand and corporate dividend policies. specifically, we use the fraction of local creative class that includes people employed in occupations that require creative thinking as a proxy for creative culture and investigate the effect of local risk-taking propensity induced by local creative culture on geographically varying dividend demand and corporate dividend policies. the empirical findings show that firms located in areas with a stronger creative culture are less likely to be dividend payers and to initiate dividends. in addition, these firms have lower dividend yields. our findings are consistent with prior literature that highlights the notion that creative culture is associated with higher degrees of risk-taking behavior and creative people are risk-takers (e.g., amabile, 1983; dewett, 2004, 2006; gardner, 1993; heilman, 2016; tesluk, farr, and klein, 1997). furthermore, our results are consistent with prior literature that underscores the link between risk-aversion and dividend demand (e.g., gordon, 1963; lintner, 1963; ucar, 2016). in addition, our results are consistent with the studies that examine the importance of the determinants of spatial and temporal variation in investors’ risk aversion (guillemette and nanigian, 2014; kuzniak and grable, 2017.1 creativity and creative thinking require higher degrees of risk-taking tendency and previous studies highlight the link between perceived risk and creativity in organizations (e.g., fidler and johnson, 1984; jalan and kleiner, 1995; shalley, gilson, and blum, 2000; zhou and george, 2001). quoting adams (1986): “creativity involves risk because it involves embracing the unknown and deviating from norms.” from marade, gibbons, and brinthaupt (2007) we have: “taking risks and encountering failure in expressing their novel ideas freely comes with the territory for truly creative individuals.” risk-taking can actually be considered as a part of creativity as stated by gardner (1993) and amabile (1983). in professional settings, a willingness to take risks is a major predictor of employee’s creative behavior (dewett, 2004, 2006), and when shown en masse by the employees it is also a major contributor to creative environments (tesluk, farr, and klein, 1997). from the neuropsychological perspective, creative people are often risk-takers as it is one of the behaviors that activate their ventral striatal reward system (heilman, 2016). therefore, one expects that creative culture encourages the risk-taking behavior. this paper uses local creative share, which is the fraction of creative class—people employed in occupations that require creative thinking—as a measure of local creative culture and investigates the role of creative culture on dividend demand and policies. the creative class measure is also consistent with the creative class theory introduced by richard florida (e.g., florida, 2002, 2003, 2005). this theory examines people who work in knowledge-intensive industries and similar occupations, such as intellectuals and artists, with the focus on their innovative culture and their contribution to economic growth. in addition, recent studies show that local investors’ risk-taking induced by creative culture has an impact on corporate innovation (ucar, 2018a) and corporate risk-taking behavior and other policies (ucar, 2018b). our paper is consistent with these studies and demonstrates the impact of local creative culture on dividends through its impact on local investors’ risk-taking characteristics. psychological biases that affect investors’ risk aversion have been extensively documented in the literature (baker and nofsinger, 2002, hirshleifer, 2015). however, not only 368 e. ucar, a. staer / financial services review 27 (2018) 367-389 retail investors but the financial industry professionals with local clienteles may be subject to behavioral and psychological biases that influence risk aversion as well (baker, filbeck, and ricciardi, 2017; baker and ricciardi, 2015; nofsinger and varma, 2007). furthermore, hirshleifer, jian, and zhang (2018) and statman (2018) suggest that cultural factors too can affect financial decision making and in particular investors’ risk aversion. our paper sheds light on the cultural factors that may influence risk tolerance of financial planners and their clienteles yielding a deeper understanding of the financial planning process biases that have been documented (nofsinger and varma, 2007). risk-taking is one of the important factors in shaping dividend demand, and it has attracted attention in the literature. previous studies suggest that investors prefer dividends over capital gains because dividends are perceived as safe current income compared with future risky capital gains (gordon, 1963; lintner, 1963). guillemette and nanigian (2014) examine components of investors’ risk aversion and kuzniak and grable (2017) find geographical and temporal variation in risk aversion of investors. furthermore, ucar (2016) demonstrates a dividend demand effect based on local religion consistent with differences in risk aversion among different religious groups. consistent with this literature, we investigate whether local risk-taking behavior induced by creative culture affects geographically varying dividend demand and corporate dividend policies. another strand of literature investigates the notion that corporations provide payout policies in line with investors’ dividend demand and dividend clienteles. in particular, baker and wurgler (2004a) and baker and wurgler (2004b) suggest that investors consider dividends as more valuable compared with capital gains and firms cater to investors’ dividend preferences through their corporate dividend policies. becker, ivković, and weisbenner (2011) and ucar (2016) demonstrate that firms cater to dividend demand by providing dividend payouts consistent with local dividend clienteles. our paper contributes to this literature by showing that firms provide dividend payouts and policies in line with local dividend demand shaped by the effect of the local creative culture on the local investors’ risk-taking. this paper is also related to the strand of literature that investigates the impact of local factors on financial outcomes and the role of local bias. ivković and weisbenner (2005) demonstrate that individual investors have a local bias and a higher likelihood of investing in local firms. massa and simonov (2006) and grinblatt and keloharju (2001) document local bias in other countries as well. pirinsky and wang (2006) show a higher degree of co-movement of stock returns for firms headquartered in the same location. hilary and hui (2009) find the impact of local religion on corporate risk-taking and corporate policies. we demonstrate the role of local creative culture for corporate payout policies. our paper is closely related to recent studies that examine local dividend clientele effect. becker, ivković, and weisbenner (2011) show how corporations determine their dividend policies based on the age of the local dividend clienteles. ucar (2016) finds that geographical variation in local religions leads to a dividend clientele effect and firms shape their dividend payouts consistent with this clientele effect. our paper contributes to this literature by introducing the role of local risk-taking characteristics induced by a new local factor—creative culture—in the determination of the dividend demand and showing a geographically varying dividend clientele effect consistent with local creative risk-taking. 369e. ucar, a. staer / financial services review 27 (2018) 367-389 using previously mentioned venues of research, we build our intuition in the following way. we start with the literature on dividend demand and dividend clientele. the origins of this literature go back to miller and modigliani (1961) who suggest that transaction costs, taxes, or other market imperfections create differences in investor preferences and dividend demand, and help to form dividend clienteles. this view was developed further by baker and wurgler (2004a) who formalize the catering theory of dividends in three basic components. first, some investors exhibit an uninformed and time-varying demand for dividend-paying stocks perhaps from sentiment, mental accounting (shefrin and statman, 1984), or investment constraints, resulting in a dividend premium. second, arbitrage is limited and the uninformed dividend premium is allowed to persist. third, managers rationally cater to investor demands by paying higher dividends when the premium on the dividend payers is high and vice versa. baker and wurgler (2004a) and baker and wurgler (2004b) are generally agnostic as to the causality flow2 in their theory, although they highlight the investor sentiment as the most probable explanation for the time-varying dividend premium to which firms respond by catering and adjusting their dividend payout policies. next, we introduce the literature on the local bias pervasive among investors (coval and moskowitz, 1999; ivković and weisbenner, 2005) who document that local investors overweight local companies in their portfolios. subsequently, we apply local bias to the dividend catering theory to isolate the magnitude of the dividend demand using the prevalence of local investors among firm shareholders. finally, we investigate the creative culture as a determinant of the local investors’ dividend demand that shapes a firm’s dividend policy through rational catering to local investors. becker, ivković, and weisbenner (2011) propose a very similar approach to test whether it is actually shareholder demand for dividends that influences firm payout policy and not the other way around by using the evidence on the heterogeneity of the dividend clienteles from graham and kumar (2006) and the existence of local retail bias from ivković and weisbenner (2005). the authors use the geographical variation in the proportion of senior citizens in the areas close to the firms’ headquarters as an instrument in their identification strategy and find that the firms do respond to the changes in the dividend demand of the local retail senior investors by adjusting their dividend payouts. the increase in dividend demand is exemplified for instance by the higher jump in the price at the initiation of the dividend (baker and wurgler, 2004a) and the lower drop in the stock price at the announcement of a decrease in the dividend. firms rationally observe that time-varying dividend premium and react accordingly to capture the investor demand for dividends which is further corroborated by the direct tests of dividend catering by kumar, lei, and zhang (2016). on the other hand, anon-trivial part of firm shares are held by local investors (about 4% of total stock ownership was held by local senior investors during 1991 to 1996 period (becker, ivković, and weisbenner, 2011a) who in turn are subject to time-varying risk tolerance determined by a multitude of factors. the importance of the retail investors in the determination of the dividend policy is further corroborated by kumar, lei, and zhang (2016) who perform a direct test of dividend catering using historical google searches.3 in other words, we do not need the interaction of the board of directors of the firm with the local creative class directly,4 we can observe the effect of the creative culture on the corporate 370 e. ucar, a. staer / financial services review 27 (2018) 367-389 decisions via catering to the changes in the dividend premium driven in part by the local investors who are subject to the effect of the creative culture. we believe it is important to distinguish between retail and institutional dividend demands and highlight the fact that the paper focuses on the retail part of the dividend demand. furthermore, we do not assume that the firms disregard institutional investors when managing their dividend payout policy. institutional investors actually do form dividend clienteles with concomitant dividend catering of their own as shown by hotchkiss and lawrence (2007) and they are subject to local bias as well (coval and moskowitz, 1999). however, consistent with the mission of fsr, we focus on the role of retail investors. the literature (for instance, baker and wurgler, 2004a, 2004b; brown, stice, and white, 2015; graham and kumar, 2006; loughran and schultz, 2004; peress, 2014; shive, 2012) finds that retail investors, and in particular, local retail investors, in general exert substantial influence on asset prices and firm policies while controlling for the influence of the institutional investors.5 furthermore, ucar (2016) shows that investor characteristics such as religion can influence dividend clienteles while pantzalis and ucar (2018) examine the incidence of allergy bouts affecting local investors and link it with lower trading volume and stock returns of the firms headquartered in the area. additionally chi and shanthikumar (2016) discover that an increase in local google searches, most commonly associated with retail investors, before earnings announcements is linked to higher bid-ask spreads, trading volume and stronger post-earnings announcement drift. these studies show that even though, retail investors hold a smaller share of stocks than the institutional investors their effect on the equity prices and corporate policies is still considerable and merits scientific scrutiny. the remainder of the paper is organized as follows. the next section presents a summary of the data and the sample selection method along with the summary statistics. section 3 provides the main empirical tests along with a set of additional detailed tests and robustness checks. section 4 presents a conclusion. 2. data, sample selection, and summary statistics we follow a sample selection method similar to the one used in recent studies (i.e., grullon et al., 2011; ucar, 2016). we exclude the firms in the utilities and financial industries (sic codes 4900–4999 and 6000–6999) and the firms with issue codes other than 10 or 11. we exclude utilities because they tend to be regulated with little discretionary control over the amount and frequency of dividend payouts.6 furthermore, regulation may influence firm characteristics like volatility or debt that show up as control variables on the right-hand side of our regressions potentially biasing our inferences. because of these reasons, financial literature tends to avoid including financial and utilities firms as mentioned in baker and wurgler (2004) and fama and french (2001).7 our sample obtains accounting and firm information from compustat and stock price information from crsp databases. we use the firm address information from compustat in the main tests. our sample requires the sample firms to have one year of lagged information because we use one year lagged firm information in constructing some variables. to measure local creative culture and risk-taking propensity associated with creative culture, 371e. ucar, a. staer / financial services review 27 (2018) 367-389 we use a variable called creativeshare, which measures the fraction of the creative class in a given firm-county similar to recent studies (ucar, 2018a, 2018b). the creative share data are from the u.s. department of agriculture economic research service (usda ers) website.8 the usda ers presents detailed information on county-level creative share information and the creative class occupations that are used in the dataset. the ers website reports that the occupations in constructing the creative share are the occupations “that involve a high level of creative thinking” such as architecture, engineering, arts, design, entertainment, media, computer, and mathematics.9 this website provides 1990, 2000, and 2007 county-level creativeshare information that measures the fraction of the local creative class. we use the interpolations of this dataset for the sample years without available data. our sample includes creativeshare variable for the years between 1990 and 2007. we use creativeshare in our empirical tests. therefore, our final sample spans the years between 1990 and 2007. we use an empirical model similar to the one used in the related literature (e.g., becker, ivković, and weisbenner (2011) and ucar (2016)). this model entails a set of three ols and logit regressions with the three dependent variables describing dividend policy: dividend payer, dividend yield and dividend initiation, and the independent variables comprising the variable of interest and a host of the firm, time and locale control variables. logit regressions are used when the dependent variable is binary, that is, takes 1 or 0 as values, which in the case of this model are dividend initiation and dividend payer. the set of three regressions is then considered as the baseline model and additional robustness and subsample tests then are performed.10 we construct dividend payout and firm characteristics variables used in our empirical tests by following previous studies (becker, ivković, and weisbenner, 2011a; grullon et al., 2011; ucar, 2016). the dependent variables used in our empirical tests are dividend payer, dividend yield, and dividend initiation. dividend payer is a dummy variable that takes a value of one if the total amount of dividends is greater than zero for a given year, and zero otherwise. dividend yield is the ratio of total dividends to lagged market value. dividend initiation is a dummy variable that takes a value of one if a non-dividend payer firm in the previous year becomes a dividend payer in the current year, and zero if a non-dividend payer firm in the previous year stays as non-dividend payer firm in the current year. we use the following set of main control variables and define them by following ucar (2016) and becker, ivković, and weisbenner (2011a). we define net income as the net income divided by total assets for a given year. cash is the cash divided by total assets for a given year. we define q as the sum of the market value of equity and the book value of liabilities divided by total assets for a given year. debt is the long-term debt divided by total assets for a given year. we define log of mv as the logarithm of a firm’s market value for a given year and log of assets as the logarithm of total assets. we define volatility as the standard deviation of monthly stock returns for the previous two-year period and lagged return as the monthly stock returns for the previous two-year period.11 asset growth is the logarithm of the total assets growth rate calculated using both the current and previous year’s figures. all accounting and firm variables are winsorized at the 1% and 99% levels. firm age is the time between the date that a firm is listed on the crsp and the current year. we use the following firm age-group indicator variables in our empirical tests: age 1–5, age 6–10, 372 e. ucar, a. staer / financial services review 27 (2018) 367-389 age 11–15, and age 16–20 with age 21 being the omitted category. the main empirical tests also control for state, industry,12 and year indicator variables. our regression model is similar to the general form of the model in becker, ivković, and weisbenner (2011a) and ucar (2016) in the example of the dividend payer test and this model can be represented as the model (1) below: divpayeri,t � � � �cscreativesharei,t � �ninii,t � �cashcashi,t � �qqi,t � �debtdebti,t � �volvoli,t � �lagretlagreti,t � �logmvlogmvi,t � �assetgrassetgrowthi,t � agegroupi,t � loccontri,t � indfei,t � yearfet � � i,t (1) moreover, we use the following alternative control variables used in the literature in some robustness tests and define these variables by following grullon et al. (2011) and ucar (2016). we define nye as the measure of firm size based on the nyse equity (market capitalization) percentiles for the corresponding period. it is important to note that although nyse equity percentiles are calculated via sorting the nyse stock universe into market capitalization percentiles, they are used to create an alternative firm size (market capitalization) variable for all firms in the sample.13 nyse equity percentiles are widely used in the financial literature, for instance, by baker and wurgler (2004) and kumar, lei, and zhang (2016) who adopt this approach from fama and french (2001) with all three papers analyzing dividend policy. we do not limit our sample to any particular exchange as neither did the aforementioned studies. furthermore, the intuition behind using nyse instead of saying nasdaq percentiles is described in fama and french (2001) on page 76: “… instead of forming equal groups by size, however, we use the 20th and 50th percentiles of market capitalization for nyse firms to assign nyse, amex, and nasdaq firms to portfolios. this prevents the growing population of small nasdaq firms from changing the meaning of small, medium, and large over the sample period. (the 20th and 50th nyse percentiles lead to similar average numbers of firms in the medium and large groups, and many more in the small group.)” it could also be argued that there might be a potential industry selection bias when using nyse size deciles to assign a nasdaq firm to a percentile. however, we believe that a possible bias is mitigated for the following reasons. first, this is set of control variables alternative to our main specification which includes the log of mv defined as the logarithm of a firm’s market value for a given year and does not employ sorting the stock universe into deciles and hence avoids any potential bias from using any ranking procedure. we only use nyse equity percentiles in the alternative set of controls in one test as a robustness check and we use our main controls variables in our other tests with the findings in both tests being broadly similar. second, we include industry fixed effects based on fama-french 48 industry classification as controls in all of the main regressions controlling for the heterogeneity between industries in our samples. while there may benon-linear patterns in firm industry distribution over time, existing corporate financial literature usually considers the imple373e. ucar, a. staer / financial services review 27 (2018) 367-389 mentation of industry fixed effects using fama-french 48 industry classification to be a reasonable approach for the majority of cases. we define m/b as the ratio of the market to book value of assets in which the market value of assets is calculated as the market value of equity plus total assets minus total equity. roa is the return on assets as calculated by income before depreciation divided by the total assets for a given year. we define sales growth as the sales growth rate calculated as the percentage change from the previous to the current year’s sales. we also use local control variables in some empirical tests consistent with the related literature. becker, ivković, and weisbenner (2011a) use the fraction of local seniors and find an age-based local dividend clientele effect. we include local seniors variable which is the proportion of individuals who are 65 years old or older within a county where a firm is headquartered by following becker, ivković, and weisbenner (2011a). ucar (2016) uses local religion and shows a local dividend clientele effect induced by local religion. therefore, we also include cpratio which is the ratio of catholics to protestants in the county where a firm is located by following ucar (2016). we also use local income, which is the median household income in a given firm county. local controls also include log of population, which is the logarithm of the population for a given county, and local education, which is the fraction of people 25 years and over having a bachelor’s, graduate, or professional or some college degree. table 1 reports summary statistics of some important firm characteristics and local creative culture as measured by the fraction of local creative class. on average, 27% of the sample firms are dividend payer firms with about 1% dividend yield. on average, 2% of the table 1 summary statistics mean 25th percentile median 75th percentile standard deviation creativeshare 0.292 0.248 0.279 0.335 0.070 dividend payer 0.275 0.000 0.000 1.000 0.446 dividend yield 0.006 0.000 0.000 0.005 0.012 dividend initiation 0.023 0.000 0.000 0.000 0.149 nye 24.547 3.000 12.000 39.000 27.563 m/b 2.029 1.085 1.465 2.235 1.745 roa 0.054 0.028 0.109 0.169 0.238 sales growth 0.208 �0.024 0.089 0.250 0.647 total assets ($mil) 1,222.124 32.146 129.586 588.833 3,799.233 age 14.332 4.441 9.422 19.641 14.475 note: this table provides summary statistics of the following variables. creativeshare measures the fraction of the creative class in a given firm county. dividend payer is a dummy variable that takes a value of one if the total amount of dividends is greater than zero for a given year, and zero otherwise. dividend yield is the ratio of total dividends to lagged market value. dividend initiation is a dummy variable that takes a value of one if a non-dividend payer firm in the previous year becomes a dividend payer in the current year, and zero if a non-dividend payer firm in the previous year stays as non-dividend payer firm in the current year. nye is defined as the measure of firm size based on the nyse equity percentiles for the corresponding period. m/b is defined as the ratio of the market to book value of assets in which market value of assets is calculated as the market value of equity plus total assets minus total equity. roa is defined as the return on assets as calculated by income before depreciation divided by the total assets for a given year. sales growth is defined as the sales growth rate calculated as the change in the previous and current year’s figures. firm age is the time between the date that a firm is listed on the crsp and the current year. 374 e. ucar, a. staer / financial services review 27 (2018) 367-389 sample firms initiate dividends during the sample years. the fraction of the creative class, as measured by creativeshare, in an average firm location is about 29%. on average, sample firms have an equity value that is equal to about the 25th percentile of the nyse equity size distribution in a given year. the average sample firm’s market-to-book ratio is about two with 5.4% roa and 21% sales growth. on average, the sample has about $1.2 billion in total assets. the average firm age is about 14.3 years. overall, table 1 presents summary statistics consistent with prior literature. 3. empirical results 3.1. main tests first, in this section, we present the main tests of the impact of local risk-taking induced by creative culture on dividend payout policy variables in table 2. we employ an empirical model similar to the one used in the related literature (e.g., becker, ivković, and weisbenner, 2011a; ucar, 2016). the main control variables include net income, cash, q, debt, volatility, lagged return, log of mv, log of assets, asset growth, and also firm age indicator variables.14 the main tests also include state, industry, and year fixed effects. standard errors are adjusted for heteroskedasticity and clustered at the firm level. the dependent variables are dividend payer, dividend yield, and dividend initiation for columns 1, 2, and 3, respectively. the main variable of interest is creativeshare that is a measure of local creative culture in a given year and defined as the fraction of the creative class in a given firm’s county. we use a logit regression model for dividend payer and dividend initiation tests and an ols model for dividend yield test in this table as well as the following tables. creativeshare is negative and statistically significant in all three columns. this result demonstrates a negative relationship between dividend payout variables and local risktaking. firms located in areas with a more prominent creative culture are less likely to be dividend payers and to initiate dividends, and they have lower levels of dividend yields. the economic significance of coefficients cannot be directly interpreted by looking at coefficient magnitudes in logit regressions. to understand the economic importance of variables, it is easier and better to focus on the change in odds for the dependent variable by using a one standard deviation change in a given independent variable. we use this approach in interpreting economic values of coefficients in this table and also the other tables of this paper. column 1 of table 2 suggests that a one standard deviation increase in creative share in a firm’s county is associated with a 17.4% less likelihood in the odds that a firm pays dividends compared with another firm located in a county with lower creative share. similarly, column 3 indicates that a one standard deviation increase in creative share in a firm’s county is associated with 10.9% less likelihood in the odds that a firm initiates dividends. these findings demonstrate the economic significance of the impact of the local creative culture on dividend demand and corporate dividend payout. column 2 also presents a similar result. column 2 suggests that a one standard deviation increase in local risk-taking behavior as measured by local creative culture leads to an almost 0.07 standard deviation decrease in dividend yield. 375e. ucar, a. staer / financial services review 27 (2018) 367-389 table 2 main tests (1) (2) (3) dependent variable dividend payer dividend yield dividend initiation creativeshare �2.746*** �0.013*** �1.692** (�3.66) (�4.75) (�2.13) net income 3.824*** �0.001*** 4.017*** (14.24) (�4.21) (6.35) cash �0.792*** 0.001 0.343* (�3.66) (1.14) (1.69) q �0.163*** �0.000*** �0.153*** (�4.11) (�8.12) (�3.62) debt �1.035*** �0.004*** �0.372* (�5.72) (�8.69) (�1.71) volatility �16.240*** �0.019*** �4.504*** (�24.55) (�18.75) (�5.67) lagged return �0.007 0.000 0.185*** (�0.31) (0.65) (6.22) log of mv 0.398*** 0.001*** 0.272*** (7.12) (8.01) (3.69) log of assets 0.050 0.000** �0.079 (0.85) (2.05) (�1.05) asset growth �0.594*** �0.001*** �0.277** (�10.74) (�16.69) (�2.39) local controls yes yes yes industry and year fe yes yes yes observations 65,239 65,239 47,014 r2 0.438 0.280 0.117 note: *significant at 10%; **significant at 5%; ***significant at 1 %. this table reports the main tests for the years between 1990 and 2007. the dependent variables are dividend payer, dividend yield, and dividend initiation for columns 1, 2, and 3, respectively. dividend payer is a dummy variable that takes a value of one if the total amount of dividends is greater than zero for a given year, and zero otherwise. dividend yield is the ratio of total dividends to lagged market value. dividend initiation is a dummy variable that takes a value of one if a non-dividend payer firm in the previous year becomes a dividend payer in the current year, and zero if a non-dividend payer firm in the previous year stays as non-dividend payer firm in the current year. creativeshare measures the fraction of the creative class in a given firm county. this table uses an empirical setting, as well dependent variables and main control variables similar to the ones used in the related literature (i.e., becker, ivkovi&cacute;, and weisbenner, 2011). this table has the following main controls: net income is defined as the net income divided by total assets for a given year. cash is the cash divided by total assets for a given year. q is defined as the sum of the market value of equity and the book value of liabilities divided by total assets for a given year. debt is defined as the long-term debt divided by total assets for a given year. log of mv is defined as the logarithm of a firm’s market value for a given year. log of assets is defined as the logarithm of total assets. volatility is defined as the standard deviation of monthly stock returns for the previous two-year lagged return is defined as the monthly stock returns for the previous two-year period. asset growth is the logarithm of the total assets growth rate calculated using both the current and previous year’s figures. these are the variables reported in the table. this table also controls for local control variables for religion (cp ratio), seniors (local seniors), population (log of population), education (local education), and income (local income); however, the coefficient estimates are not reported for brevity. the main tests also include the following age-group indicator variables: age 1–5, age 6–10, age 11–15, and age 16–20. age 21 and over is the dropped category in the tests. the main empirical tests also control for industry and year indicator variables. intercept, firm age indicators, industry, and year dummy variables are not displayed for brevity. standard errors are adjusted for heteroskedasticity and clustered at the firm level. robust t and z stats are in parentheses. 376 e. ucar, a. staer / financial services review 27 (2018) 367-389 table 2 presents empirical findings consistent with risk-taking effect associated with creativity and creative culture highlighted in previous social science studies. table 2 also shows evidence in line with the relationship between risk aversion and investors’ dividend preferences suggested in the related dividend literature. in addition, this table’s results are consistent with dividend clientele argument which suggests a variation in dividend demand associated with differences in investor characteristics. overall, this table indicates that risk-taking characteristics induced by local creative culture and environment play an important role in investors’ demand for dividends and payout policies of local firms that cater to this demand. 3.2. additional tests and robustness checks to shed more light on the main results reported in the previous section, in this section we perform additional tests and robustness checks and report the results in table 3. panel a, b, and c in table 3 display results for the robustness tests for the dependent variables: dividend payer, dividend yield, and dividend initiation, respectively. furthermore, we include the main and local control variables, and year and industry dummies in all regressions performed in this section. tests in column 1 investigate whether local factors or state-related variables drive the results reported in the main dividend policy tests in table 2. across all panels a through c, creativeshare in column 1 has a negative sign consistent with the earlier main dividend test results. therefore, the negative sign and magnitude for creativeshare in column 1 provide additional support to the findings in table 2 and demonstrate that the effect of local risk-taking, as measured by local creative culture, is robust to local factors and state effects and is the main driver of the results shown in the earlier findings. to shed more light on our previous findings and demonstrate that local risk-taking induced by creative culture is effective on not only some areas with a prominent creative culture but also on all the other areas, we exclude firms located in areas with a highly prominent creative culture and repeat the main regressions in column 2. specifically, we exclude firms that are located in the silicon valley area and re-examine the empirical findings in column 2 with the underlying goal to investigate the extent of the local risk-taking effect. creativeshare is negative and statistically significant for all the three dividend payout tests in panels a through c in column 2 demonstrating that local risk-taking effect on dividend payout holds not only for the areas with a well-known and strong creative culture but also for the other areas. the results in column 2 also indicate economically important effects. a one standard deviation increase in local risk-taking where a firm is located, as measured by creativeshare, leads to a 16.6% decrease in the odds that a firm pays dividends as presented in panel a. a one standard deviation increase in local creative share is associated with almost 0.065 standard deviation decrease in dividend yield in panel b. in addition, a one standard deviation increase in local creative share leads to a 9.7% decrease in the likelihood of a firm initiating dividends as indicated in panel c. these findings provides additional supporting evidence and highlight the influence of local risk-taking propensity induced by creative culture on dividend demand and corporate payout policies of local firms. 377e. ucar, a. staer / financial services review 27 (2018) 367-389 table 3 additional tests and robustness checks (1) (2) (3) panel a: dividend payer tests dependent variable: dividend payer creativeshare �2.807 (�3.05)*** �2.674 (�3.55)*** �2.085 (�2.70)*** main controls y y y year variables y y y industry variables y y y local controls y y y state fixed effects y n n excluding areas with a famous creative culture n y n alternative location data n n y observations 65,239 62,335 49,245 r2 0.447 0.436 0.432 panel b: dividend yield tests dependent variables: dividend yield creativeshare �0.010 (�3.12)*** �0.012 (�4.46)*** �0.012 (�4.08)*** main controls y y y year variables y y y industry variables y y y local controls y y y state fixed effects y n n excluding areas with a famous creative culture n y n alternative location data n n y observations 65,239 62,335 49,245 r2 0.289 0.280 0.280 panel c: dividend initiation tests dependent variable creativeshare �2.032 (�2.09)** �1.532 (�1.93)* �0.952 (�1.04) main controls y y y year variables y y y industry variables y y y local controls y y y state fixed effects y n n excluding areas with a famous creative culture n y n alternative location data n n y observations 46,831 44,381 34,458 r2 0.124 0.115 0.113 note: *significant at 10%; **significant at 5%; ***significant at 1 %. this table reports the additional tests and robustness checks for the main tests with controls for state effects, strong creative culture areas and an alternative location dataset. the dependent variables are dividend payer, dividend yield, and dividend initiation in panels a, b, and c, respectively. the main variable of interest is creativeshare. column 1 controls for state dummies. column 2 re-examines the tests after excluding firms located in areas with a strong creative culture, and more details are provided in the text. column 3 re-examines the tests by using an alternative firm location information provided by the compact disclosure data as well as the firm location information from bill mcdonald’s website. this table has the following main controls: net income, cash, q, debt, log of mv, log of assets, volatility, lagged return, asset growth. the main controls also include the following age-group indicator variables: age 1–5, age 6–10, age 11–15, and age 16–20. age 21 and over is the dropped category in the tests. these variables are defined in table 1 and table 2. the empirical tests also control for the state, industry, and year indicator variables. only creativeshare is displayed for brevity. local controls comprise the following variables: local seniors, cpration, income, log of population, education. local seniors is defined as the proportion of individuals who are 65 years old or older within a county where a firm is headquartered. cpratio is defined as the ratio of catholics to protestants in the county where a firm is located. income is the median household income in a given firm county. log of population is the logarithm of population for a given county. education is the fraction of people 25 years and over having a bachelor’s, graduate, or professional or some college degree. standard errors are adjusted for heteroskedasticity and clustered at the firm level. robust t and z stats are in parentheses. 378 e. ucar, a. staer / financial services review 27 (2018) 367-389 in column 3, we rerun the main regressions for the dividend policy variables using an alternative firm location dataset and analyze the coefficient estimates. the previous tests use firm location information provided by compustat. prior literature shows that there is a small number of headquarter moves (e.g., pirinsky and wang, 2006). one might suggest that the fact that compustat does not consider any corporate relocation may bias our findings. to take into account corporate relocations and to ascertain whether our findings are driven by compustat firm location information or not, we use an alternative firm location dataset. in particular, we use firm location information from the compact disclosure database as well as bill mcdonald’s website15 and repeat our main tests in column 3. the results presented in column 3 exhibit a negative and statistically significant coefficient for creativeshare for all the three dividend payout variables in line with the previous tests after using an alternative firm location information dataset which considers firm relocations. our empirical findings remain economically robust too. column 3 estimates demonstrate that a one standard deviation increase in creative share leads to an almost 14% decrease in the odds that a firm becomes a dividend payer. in addition, a one standard deviation increase in creative share is associated with almost 0.068 standard deviation decrease in dividend yield. panel c suggests a negative relationship between creative share and dividend initiation analysis consistent with the previous results. however, the coefficient is statistically insignificant. the reason for this can be the smaller sample size in the alternative location data sample for the dividend analysis. overall, table 3 provides additional supporting evidence for the role of local risk-taking induced by local creative culture in the determination of the investors’ dividend demands and, hence, of the corporate dividend policies of local firms that cater to these demands. 3.3. tests using an alternative set of controls in this section, we provide a series of robustness tests, as well as some additional tests, to shed more light on local risk-taking effect induced by creative culture on geographically varying dividend demand and corporate dividend policies. previous studies employ a slightly different set of variables in examining certain dividend payout variables. now, first, we repeat our main tests for dividend payer, dividend yield, and dividend initiation after controlling an alternative set of control variables used in prior literature (e.g., fama and french, 2001; grullon et al., 2011) by following ucar (2016). specifically, we control for market-to-book ratio, roa, sales growth, and firm size, by following definitions used by fama and french (2001) and grullon et al. (2011), as well as local controls, and year, industry, and state dummies in table 4. table 4 reports a negative and statistically significant creative share coefficient in all columns as consistent with our earlier results. the empirical findings are also economically important. a one standard deviation increase in creative share is associated with a 20.9% decrease in the likelihood of becoming a dividend payer for a firm. similarly, a one standard deviation increase in creative share leads to a 0.086 standard deviation decrease in dividend yield and an 11.5% decrease in the likelihood of initiating dividends. this table shows that our empirical results are robust to alternative control variables and supports the main findings. this table suggests that local risk-taking behavior encouraged by local creative 379e. ucar, a. staer / financial services review 27 (2018) 367-389 culture is the main driver of the negative impact on dividend payout as indicated by the previous sections. 3.4. identification tests one might suggest that a firm’s location choice is endogenous or there might be an omitted variable that affects the results presented in the earlier sections. therefore, we use a series of tests controlling for endogeneity. first, we use a matched sample analysis similar to the one used in ucar (2016) and re-examine dividend payout variables. matching sample analysis allows us to control for the firm characteristics while observing the exogenous variation in creativeshare.16 we take a closer look at pairwise comparisons between firms located in areas with a high creative share and a matched sample of firms located in areas with a low creative share. first, we divide the sample into five based on creativeshare, and examine firms in the highest quintile of creativeshare as the firms located in areas with a high creativeshare and the lowest quintile of creativeshare as the firms located in areas with a low creativeshare. we determine a firm-year observation with the same year, industry, and age group from low creativeshare area firms for each firm-year observation from high creativeshare area firms. next, we use a matching process based on the firm characteristics including total assets, market value, net income, cash, q, debt, volatility, table 4 tests with an alternative set of controls (1) (2) (3) dependent variable dividend payer dividend yield dividend initiation creativeshare �3.360*** �0.015*** �1.779** (�4.59) (�5.33) (�2.21) nye 0.040*** 0.000*** 0.017*** (30.40) (23.14) (11.98) m/b �0.414*** �0.001*** �0.109*** (�11.93) (�12.99) (�3.94) roa 5.880*** 0.002*** 4.370*** (21.90) (4.20) (11.53) sales growth �1.028*** �0.001*** �0.005 (�15.28) (�18.42) (�0.06) local controls yes yes yes year, industry, state dummies yes yes yes observations 65,118 65,118 46,767 r2 0.331 0.215 0.099 note: ***, **, and *indicate significance at the 1%, 5%, and 10% levels, respectively. this table reports the main tests with an alternative set of controls. the dependent variables are dividend payer, dividend yield, and dividend initiation for columns 1, 2, and 3, respectively. dividend payer and dividend initiation tests have logit regressions whereas dividend yield has ols regression. the main variable of interest is creativeshare. this table has the following set of controls: nye, m/b, roa, and sales growth. only creativeshare and the control variables are displayed for brevity. the empirical tests also control for the state, industry, and year indicator variables as well as local controls (local seniors, cp ratio, income, education, and log of population). intercept, year and industry dummies are not reported for brevity. robust t statistics and z statistics are reported in parentheses. pseudo r2 values are reported for dividend payer and dividend initiation tests. 380 e. ucar, a. staer / financial services review 27 (2018) 367-389 and lagged return. in particular, we match every firm-year observation of high creativeshare area firms with a firm-year observation from a low creativeshare area firm from the same year, industry, and age group, with the closest matched values for total assets, market value, net income, cash, q, debt, volatility, and lagged return. the matched sample analysis includes the firms from high creativeshare areas with a match from low creativeshare areas in table 5 and examines differences in dividend payout variables between the high creativeshare area firms and their matches from the low creativeshare areas. table 5 presents the mean values for the matched sample analysis. the findings are consistent with the earlier findings. the difference in dividend payer between high creativeshare area firms and low creativeshare area firms is negative and statistically significant. similarly, the differences in dividend yield and dividend initiation are negative and statistically significant. this table shows that firms located in areas with a greater degree of local risk-taking induced by creative culture are less likely to pay and initiate dividends and have lower dividend yields compared with firms located in areas with a lower degree of local risk-taking as measured by creative culture after using matched sample tests. in summary, table 5 is consistent with the earlier findings and provides supporting evidence by using a matched sample analysis. to shed more light on the local risk-taking effect induced by creative culture and to take a further step in addressing endogeneity, now we use an instrumental variable (iv) approach and re-examine dividend payout variables. in particular, we repeat the earlier logit regression analyses of dividend payer and dividend initiation by using a probit regression with an iv approach and the earlier ols regression analysis of dividend yield by using a two-stage least squares (2sls) analysis with an iv approach. we use the following variables as an iv for creativeshare to address possible endogeneity: creativesharet-10, the creative share lagged by 10 years, in table 6 panel a and artshare, the fraction of people used in the arts for a given county, in table 6 panel b. in table 6 panel a, we use creativesharet-10, the creative share lagged by 10 years, as the first iv for the creativeshare. the creative share lagged by 10 years can be considered correlated with the current creative share. on the other hand, one expects that the creative table 5 matched sample analysis variable n high creativeshare low creativeshare difference p-value dividend payer 11,777 0.203 0.297 �0.093 (0.000)*** dividend yield 11,777 0.004 0.006 �0.002 (0.000)*** dividend initiation 7,317 0.014 0.021 �0.007 (0.001)*** note: *significant at 10%; **significant at 5%; ***significant at 1%. this table presents the mean values for dividend payout variables for firms that are located in high creativeshare and a matched sample of firms that are located in low creativeshare areas. dividend payout variables—dividend payer, dividend yield, and dividend initiation—defined in tables 1 and 2. the sample is divided into five based on creativeshare, and the firms in the highest quintile of creativeshare are assigned to the high creativeshare area and the firms in the lowest quintile of creativeshare are assigned to the low creativeshare area. matched low creativeshare area firms are identified after matching each firm-year observation of a high creativeshare area firm with a firm-year observation of a low creativeshare area firm that is from the same year, industry, and age group with the closest asset size, market value, net income, cash, q values, debt, volatility, and lagged return. all the variables that are used in matching are defined in tables 1 and 2. robust p-values are in parentheses. 381e. ucar, a. staer / financial services review 27 (2018) 367-389 share lagged by 10 years is not correlated with any omitted variables in the current year. furthermore, using a local variable lagged by 10 years can be regarded as a good iv candidate considering that hilary and hui (2009) use local religion lagged by three years as an iv for current local religion in their setting. therefore, we implement creativesharet-10 as an iv in the first stage of 2sls in table 6 panel a to predict creativeshare before running the main tests for corporate decision and risk-taking variables during the second stage. table 6 instrumental variable approach (1) (2) (3) dependent variable dividend payer dividend yield dividend initiation panel a: iv approach (iv: creativesharet-10) creativeshare �1.240*** �0.014*** �0.589 (�4.40) (�3.75) (�1.03) main controls y y y local controls y y y year fixed effects y y y industry fixed effects y y y observations 25,691 25,691 19,049 r2 0.400 0.229 0.137 panel b: iv approach (iv: artshare) creativeshare �1.688*** �0.011*** �0.412 (�7.59) (�3.14) (�0.95) main controls y y y local controls y y y year fixed effects y y y industry fixed effects y y y observations 65,239 65,239 47,014 r2 0.429 0.280 0.110 note: *significant at 10%; **significant at 5%; ***significant at 1%. this table reports instrumental variables (iv) tests for the dependent variables dividend payer, dividend yield, and dividend initiation defined earlier. panels a and b use an iv probit analysis with an iv approach for logit regression analyses of dividend payer and dividend initiation used in the main tests and a two-stage least squares (2sls) analysis with iv approach for the ols regression analysis of dividend yield used in the main tests. panels a and b present coefficients of the instrumented creative share variable from second stages of these iv analyses. panels a and b use all the main control variables used the earlier dividend payout analyses along with the year and industry dummies and local controls. panel b uses creativesharet-10, creative share lagged by five years, as iv whereas panel c uses artshare, the fraction of people employed in the arts in a county in a year, as iv. more details on iv approach are provided in the text. this table has the following main controls: net income, cash, q, debt, log of mv, log of assets, volatility, lagged return, asset growth. local controls comprise the following variables: local seniors, cpration, income, log of population, education. local seniors is defined as the proportion of individuals who are 65 years old or older within a county where a firm is headquartered. cpratio is defined as the ratio of catholics to protestants in the county where a firm is located. income is the median household income in a given firm county. log of population is the logarithm of the population for a given county. education is the fraction of people 25 years and over having a bachelor’s, graduate, or professional or some college degree. the main controls also include the following age-group indicator variables: age 1–5, age 6–10, age 11–15, and age 16–20. age 21 and over is the dropped category in the tests. these variables are defined in table 1 and table 2. only creativeshare is reported for brevity. robust t and z values are in parentheses. pseudo r2 values are reported for dividend payer and dividend initiation tests. 382 e. ucar, a. staer / financial services review 27 (2018) 367-389 we report the coefficients of the instrumented creativeshare variable from the second stages of these iv analyses in table 6. first, in panel a, we use an iv approach that uses the ten years lagged creative share variable, creativesharet-10. panel a reports creative share coefficients with a negative sign consistent with the earlier findings. creative share is statistically significant for dividend payer and dividend yield tests that provides additional support to the earlier findings and highlights the role of local risk-taking induced by creative culture for dividend demand and dividend policy. creative share is not statistically significant for dividend initiation although it has a negative sign as expected. lower statistical significance may be because of a smaller sample of observations used in the dividend initiation tests. overall, panel a provides some additional supporting evidence for the earlier empirical results. next, in panel b, we use artshare, the fraction of people used in the arts for a given county, as a second iv variable for creativeshare to further address endogeneity concerns and shed more light on the impact of creative culture on corporate decisions. people who are employed in the arts include “art and design workers, painters, musicians, composers, sculptors, photographers, and so forth”17 artshare is a subset of creativeshare—creative class—that includes the people who work in the arts. usda ers reports that creative class dataset identifies occupations that involve a high level of “thinking creatively” and this skill element is defined as “developing, designing, or creating new applications, ideas, relationships, systems, or products, including artistic contributions.” the usda ers creativeshare—creative class—definition includes occupations such as architecture, engineering, arts, design, entertainment, sports, media, computer and mathematical science, advertising, top executives, physical scientists, and social scientists. artists or people from art occupations are considered as creative people and risk-takers18 as the other occupations in the creative class—creativeshare. therefore, artsshare and creativeshare are correlated in terms of creativity and risk-taking. although one might suggest that other occupations that constitute creative culture—creativeshare—such as architecture, engineering, media, computer and mathematical science, and advertising, might be considered as related to corporate decisions or factors affecting corporate decisions, this reasoning does not hold for people employed in the arts. artsshare, which represents the local fraction of artists or people from art occupations, is directly related to local creative culture while artsshare is not considered as correlated with the factors related to local corporate decisions or it is not expected to influence corporate policies. therefore, artshare can be considered a good iv because it is correlated with creativity and creative-risk taking but not correlated with any potential omitted variables related to corporate decisions. the evidence in panel b demonstrates that creativeshare instrumented by artshare exhibits negative coefficients in each one of the dividend policy tests consistent with the earlier results. except for the dividend initiation test, all the other tests have a statistically significant creative culture effect. lower statistical significance for creativeshare in the dividend initiation test is most likely because of a much smaller sample used in that particular test. in summary, after using ivs, creativeshare has signs and coefficient magnitudes consistent with the earlier results. creativeshare is statistically significant for dividend payer and dividend yield tests. this result provides additional support to the earlier findings and 383e. ucar, a. staer / financial services review 27 (2018) 367-389 highlights the role of local risk-taking induced by creative culture for dividend demand and dividend policy. creativeshare is not statistically significant for dividend initiation although it has a negative sign as expected. this result might come from a smaller sample of observations used in the dividend initiation tests. overall, the evidence in panels a and b lends support to the earlier findings and shows that local creative culture and creative risk-taking have a negative effect on dividend demand and corporate dividend payout after addressing endogeneity concerns. the endogeneity, for instance, can present itself via geographical clustering.19 one example could be the stock options explanation where the highly educated local investors are comprised in large part by the long-term employees of once-young startups that usually use stock options-based compensation in the early stage of their lifecycle. in this case, more young technology-oriented firms would locate close to the education clusters and use riskier stock options-based compensation to attract employees with low risk aversion. we recognize that causality is difficult to ascertain in corporate finance research and that is why we implement several robustness tests and corrections for endogeneity. first, we use several control variables that should capture the geographical clustering effect: education, to control for local high-education high-tech clusters, firmage, to control for the heterogeneity in dividend policy across firm’s lifecycle, and industry fixed effects, to control for the variation in dividend payout policies across industries. we believe that younger firms in tech industries in our sample would prefer stock options-based compensation to the dividend one and that our control variables should account for that pattern in the data. second, we address possible endogeneity concerns by using two iv: creativesharet-10, the creative share lagged by 10 years, and artshare, the fraction of people employed in the arts for a given county. iv regressions like 2sls help to elucidate the causality in regressions with possible endogeneity. one of the requirements for a valid iv is that the iv is correlated with the potentially endogenous independent variable of interest but not correlated with the possible omitted variables as stated on pages 89–90 in wooldridge (2010). hilary and hui (2009) use local religion lagged by three years as an iv for current local religion and we employ a similar approach with creativesharet-10, that is, creative share lagged by 10 years, which is correlated with the current creative share but is unlikely to correlate with the current year omitted variables. artshare, the fraction of people employed in the arts for a given county, is also correlated with creative share and is unlikely to be correlated with the omitted variables, say, for instance, the frequency of using stock options compensation plans by the local companies. the results in 2sls tests show that after accounting for possible omitted variables, the relationship between creative share and dividend policy remains significant and in line with our prior hypotheses. third, we also perform a matching sample analysis where we match firms from low and high creative share areas on their characteristics and compare the sample means for the three dividend policy variables. this approach allows us to control for potentially omitted patterns in the variation of firm characteristics and to isolate the relationship between creative share and dividend policy from confounding factors. in general, we believe that the aforementioned tests offer sufficient evidence in favor of our hypothesis even in the presence of possible endogeneity, for instance, because of geographical clustering. 384 e. ucar, a. staer / financial services review 27 (2018) 367-389 4. conclusion studies in social literature argue that creativity is associated with risk-taking behavior and creative people are risk-takers. we use a novel measure of risk-taking, local creative culture, and examine the relationship between investors’ willingness to take risks and investor’s payout preferences and, through the dividend catering channel, corporate payout policies. in particular, we empirically investigate the effect of the local creative share, which is a proxy for local creative culture, on geographically varying dividend demand and corporate dividend policies. we show that firms located in areas with a greater creative share are less likely to become dividend payers and to initiate dividends. in addition, firms located in areas with a more pronounced creative culture have lower dividend yields. these results are consistent with higher risk-taking behavior associated with creativity and creative culture. these results are also consistent with previous studies in the dividend and financial planning (guillemette and nanigian, 2014; kuzniak and grable, 2017) literature that highlight the role of risk aversion for dividend demand and investor preferences for risky assets. furthermore, our findings offer underscore the importance of the dividend catering theory and argue that firms shape their payout policies consistent with variation in dividend demand and cater to dividend demand of dividend clienteles. we contribute to this literature by showing the impact of risk-taking behavior induced by creative culture on corporate dividend policies consistent with dividend demand shaped by this risk-taking behavior. our paper shows the role of local risk-taking behavior associated with creative culture for corporate dividend policies of local firms that cater to investors’ dividend demand. moreover, recent studies underline the importance of local factors on dividend policies and show local dividend clienteles based on local factors such as local senior effect (becker, ivković, and weisbenner, 2011a) or local religion effect (ucar, 2016). this study contributes to this literature by introducing a new local factor associated with local risk-taking20 on dividend demand. furthermore, the tests in this paper yield additional evidence on the importance of the cultural factors (hirshleifer, jian, and zhang, 2018) that influence risk tolerance of financial industry professionals and their clienteles (baker and ricciardi, 2015; nofsinger and varma, 2007), which is a key factor in the financial planning process. we demonstrate that the empirical results remain robust after addressing potential endogeneity concerns about local creative culture and firm location by employing a matched sample analysis and an instrumental variable approach. the empirical findings hold after a series of robustness tests. the local risk-taking effect is robust to local controls, county fixed effects, and the use of alternative firm location dataset. the results are also robust to use of an alternative set of control variables. the empirical results remain robust after excluding firms located in areas with a well-known strong creative culture. this point demonstrates that the impact of local-risk-taking on dividends affects not only the firms located in areas with a highly prominent creative environment like silicon valley but also the firms located in other areas. overall, this paper introduces a new local factor—creative culture and risktaking associated with creative culture—to dividend literature and shows that investors’ local risk-taking characteristics affect corporate dividend policies. this finding highlights the 385e. ucar, a. staer / financial services review 27 (2018) 367-389 notion that firms cater to investors’ dividend preferences determined by local risk-taking characteristics. notes 1 the empirical results hold after addressing endogeneity concern by using a matched sample analysis and an iv approach. our findings also remain robust after a series of robustness tests. the local risk-taking effect holds after controlling for local factors, county effects, and it also remains robust after using an alternative firm location dataset. in addition, our findings hold after using an alternative set of control variables including return on assets and sales growth. return on assets and sales growth are used to control for profitability and investment opportunities that determine dividend supply so that we can focus on the dividend demand effect on firm dividend policy. we thank anonymous referee for pointing this distinction out. these results support the notion that the effect shown in this paper comes through the local culture channel. our findings remain robust after excluding areas with a highly prominent local creative culture. the association between dividend policies and the local risk-taking propensity proxied by the local creative culture is significantly positive in not only the areas with a highly prominent creative culture but also all the other areas. this result highlights the extent of the association between the dividend policy and local risktaking. we thank the anonymous referee for the suggestion to move the discussion about robustness tests to a footnote. 2 we thank anonymous referee for bringing the role of causality in dividend catering to our attention. 3 we thank the anonymous referee for highlighting the importance of additional tests of dividend catering theory. 4 we would like to thank anonymous referee for the recommendation to clarify the importance the board of directors in setting firm’s dividend policy. 5 we thank anonymous referee for raising the issue of differences in shareholdings of retail and institutional investors. 6 we thank the anonymous referee for bringing this point to our attention. 7 quoting from fama and french (2001) on page 68: “we begin by examining the incidence of dividend payers among nyse, amex, and nasdaq firms. we exclude utilities from the sample to avoid the criticism that their dividend decisions are a byproduct of regulation. we exclude financial firms because our data on the characteristics of dividend payers are from compustat and compustat’s historical coverage of financial firms is spotty.” 8 please see the dataset specifics at http://www.ers.usda.gov/data-products/creativeclass-county-codes/. 9 please see the data documentation at http://www.ers.usda.gov/data-products/creativeclass-county-codes/documentation/. 10 we thank the anonymous referee for the suggestion to clarify the models used in this paper. 386 e. ucar, a. staer / financial services review 27 (2018) 367-389 11 following ucar (2016), we require volatility and lagged return to have stock return information to be non-missing for at least the previous 12 months for firms with stock returns available for less than 24 months. 12 we use the fama and french (1997) 48 industry classifications. 13 we thank the anonymous referee for pointing out the potential bias when using nyse percentiles. 14 the addition of firm controls related to profitability and investment opportunities of the firm allows us to isolate dividend supply determined for instance by factors like lagged return, q, and asset growth as in becker, ivković, and weisbenner (2011) and focus on the dividend demand by the local investors. we thank anonymous referee for bringing this issue to our attention. 15 the 10k header dataset is provided by bill mcdonald’s website at http://www3.nd. edu/�mcdonald/10-k_headers/10-k_headers.html. 16 we thank the anonymous referee for bringing this interpretation of the matching sample tests to our attention. 17 https://www.ers.usda.gov/data-products/creative-class-county-codes/ and https:// www.ers.usda.gov/data-products/creative-class-county-codes/documentation.aspx. 18 for example, marade, gibbons, and brinthaupt (2007), poorsoltan (2012), tyagi et al. (2017), and fillis (2000). 19 we thank the anonymous referee for bringing this potential source of endogeneity to our attention. 20 we should emphasize that creativeshare is not a complete proxy for local investor risk taking. there is, however, a demonstrably strong association between creative culture and risk-taking (see section 1: introduction), even though the former is not a complete proxy for the latter. we thank anonymous referee for making this distinction. references adams, j. l. (1986). the care and feeding of ideas: a guide to encouraging creativity. first paperback printing edition (da capo lifelong books, reading, mass). amabile, t. m. (1983). the social psychology of creativity: a componential conceptualization. journal of personality and social psychology, 45, 357–376. baker, h. k., filbeck, g., & ricciardi, v. (2017). financial behavior: players, services, products, and markets. new york, ny: oxford university press. baker, h. k., & nofsinger, j. r. (2002). psychological biases of investors. financial services review, 11, 97. baker, h. k., & ricciardi, v. (2015). understanding behavioral aspects of financial planning and investing. journal of financial planning, 28, 22–26. baker, m., & wurgler, j. (2004a). a catering theory of dividends. the journal of finance, 59, 1125–1165. baker, m., & wurgler, j. (2004b). appearing and disappearing dividends: the link to catering incentives. journal of financial economics, 73, 271–288. becker, b., ivković, z., & weisbenner, s. (2011). local dividend clienteles. the journal of finance, 66, 655–683. brown, n. c., stice, h., & white, r. m. (2015). mobile communication and local information flow: evidence from distracted driving laws. journal of accounting research, 53, 275–329. 387e. ucar, a. staer / financial services review 27 (2018) 367-389 chi, s. s., & shanthikumar, d. m. (2016). local bias in google search and the market response around earnings announcements. the accounting review, 92, 115–143. coval, j. d., & moskowitz, t. j. (1999). home bias at home: local equity preference in domestic portfolios. the journal of finance, 54, 2045–2073. dewett, t. (2004). employee creativity and the role of risk. european journal of innovation management, 7, 257–266. dewett, t. (2006). exploring the role of risk in employee creativity. the journal of creative behavior, 40, 27–45. fama, e. f., & french, k. r. (1997). industry costs of equity. journal of financial economics, 43, 153–193. fama, e. f., & french, k. r. (2001). disappearing dividends: changing firm characteristics or lower propensity to pay? journal of applied corporate finance, 14, 67–79. fidler, l. a., & johnson, j. d. (1984). communication and innovation implementation. the academy of management review, 9, 704–711. fillis, i. (2000). being creative at the marketing entrepreneurship interface: lessons from the art industry. journal of research in marketing and entrepreneurship, 2, 125–137. florida, r. (2002). the rise of the creative class: and how it’s transforming work, leisure, community and everyday life. new york, ny: basic books. florida, r. (2003). cities and the creative class. city & community, 2, 3–19. florida, r. (2005). cities and the creative class. london: taylor and francis. gardner, h. (1993). creating minds: an anatomy of creativity seen through the lives of freud, einstein, picasso, stravinsky, eliot, graham, and gandhi. new york, ny: basic books. gordon, m. j. (1963). optimal investment and financing policy. the journal of finance, 18, 264–272. graham, j. r., & kumar, a. (2006). do dividend clienteles exist? evidence on dividend preferences of retail investors. the journal of finance, 61, 1305–1336. grinblatt, m., & keloharju, m. (2001). how distance, language, and culture influence stockholdings and trades. the journal of finance, 56, 1053–1073. grullon, g., paye, b., underwood, s., & weston, j. p. (2011). has the propensity to pay out declined? the journal of financial and quantitative analysis, 46, 1–24. guillemette, m., & nanigian, d. (2014) what determines risk tolerance? financial services review, 23, 207–218. heilman, k. m. (2016). possible brain mechanisms of creativity. archives of clinical neuropsychology, 31, 285–296. hilary, g., & hui, k. w. (2009). does religion matter in corporate decision making in america? journal of financial economics, 93, 455–473. hirshleifer, d. (2015). behavioral finance. annual review of financial economics, 7, 133–159. hirshleifer, d., jian, m., & zhang, h. (2018). superstition and financial decision making. management science, 64, 235–252. hotchkiss, e. s., & lawrence, s. (2007). empirical evidence on the existence of dividend clienteles. ssrn scholarly paper, social science research network, rochester, ny. ivković, z., & weisbenner, s. (2005). local does as local is: information content of the geography of individual investors’ common stock investments. the journal of finance, 60, 267–306. jalan, a., & kleiner, b. h. (1995). new developments in developing creativity, journal of managerial psychology, 10, 20–23. kumar, a., lei, z., & zhang, c. (2016). a direct test of the dividend catering hypothesis. ssrn scholarly paper, social science research network, rochester, ny. kuzniak, s., & grable, j. e. (2017). does financial risk tolerance change over time? a test of the role macroeconomic, biopsychosocial and environmental, and social support factors play in shaping changes in risk attitudes. financial services review, 26, 315–338. lintner, j. (1963). the cost of capital and optimal financing of corporate growth. the journal of finance, 18, 292–310. loughran, t., & schultz, p. (2004). weather, stock returns, and the impact of localized trading behavior. the journal of financial and quantitative analysis, 39, 343–364. 388 e. ucar, a. staer / financial services review 27 (2018) 367-389 marade, a. a., gibbons, j. a., & brinthaupt, t. m. (2007). the role of risk-taking in songwriting success. the journal of creative behavior, 41, 125–149. massa, m., & simonov, a. (2006). hedging, familiarity and portfolio choice. the review of financial studies, 19, 633–685. miller, m. h., & modigliani, f. (1961). dividend policy, growth, and the valuation of shares. the journal of business, 34, 411–433. nofsinger, j. r., & varma, a. (2007). how analytical is your financial advisor? financial services review, 16, 245–260. pantzalis, c., & ucar, e. (2018). allergy onset and local investor distraction. journal of banking & finance, 92, 115–129. peress, j. (2014). the media and the diffusion of information in financial markets: evidence from newspaper strikes. the journal of finance, 69, 2007–2043. pirinsky, c., & wang, q. (2006). does corporate headquarters location matter for stock returns? the journal of finance, 61, 1991–2015. poorsoltan, k. (2012). artists as entrepreneurs. international journal of entrepreneurship, 16, 77–94. shalley, c. e., gilson, l. l., & blum, t. c. (2000). matching creativity requirements and the work environment: effects on satisfaction and intentions to leave. the academy of management journal, 43, 215–223. shefrin, h. m., & statman, m. (1984). explaining investor preference for cash dividends. journal of financial economics, 13, 253–282. shive, s. (2012). local investors, price discovery, and market efficiency. journal of financial economics, 104, 145–161. statman, m. (2018). culture matters to clients, and it should matter to planners. journal of financial planning, 31, 32–34. tesluk, p. e., farr, j. l., & klein, s. r. (1997). influences of organizational culture and climate on individual creativity. the journal of creative behavior, 31, 27–41. tyagi, v., hanoch, y., hall, s. d., runco, m., & denham, s. l. (2017). the risky side of creativity: domain specific risk taking in creative individuals. frontiers in psychology, 8, 145. ucar, e. (2016). local culture and dividends. financial management, 45, 105–140. ucar, e. (2018a). local creative culture and corporate innovation. journal of business research, 91, 60–70. ucar, e. (2018b). what creates corporate risk-taking? local creative culture and corporate decisions. european financial management, forthcoming. wooldridge, j. m. (2010). econometric analysis of cross section and panel data (2nd ed.). cambridge, ma: mit press. zhou, j., & george, j. m. (2001). when job dissatisfaction leads to creativity: encouraging the expression of voice. the academy of management journal, 44, 682–696. 389e. ucar, a. staer / financial services review 27 (2018) 367-389 financial services review, 32(2) 53 factors mediating the association between financial socialization and well-being: an african american perspective john h. young,1 crystal r. hudson,2 and c. w. copeland3 abstract this study examines the relationship between financial socialization and well-being (financial and subjective) mediated by three motivations: financial knowledge, goal setting, and self-control, either directly on financial behaviors or indirectly through financial skills on financial behaviors of african americans compared with european americans. we used an integrated concept derived from gudmunson and dane’s (2011) financial socialization framework and fisher and fisher’s (1992) information motivation behavior model to examine national financial well-being survey data. we found a significant difference between african americans and european americans where “self-control” mediates directly through financial behaviors and indirectly through financial skills. there was also a significant difference in the relationship where “goal setting” mediated the relationship indirectly through skills in financial behaviors. finally, there was a significant difference when “financial knowledge” directly mediated financial behaviors. the findings from this study imply that self-control is one of the most impactful factors on african americans’ wellbeing. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation young, j. h., copeland, c. w., hudson, c. r. (2024). factors mediating the association between financial socialization and well-being: an african american perspective. financial services review, 32(2), 53-76. introduction financial well-being is a holistic view of one’s financial life and how one feels and thinks about money (yakoboski et al., 2019). specifically, financial well-being is one’s ability to meet current and ongoing financial obligations to maintain anticipated and desired living standards, 1 clark atlanta university, atlanta, ga, usa 2 clark atlanta university, atlanta, ga, usa 3 corresponding author (cwcopeland2@cau.edu). clark atlanta university, atlanta, ga, usa as well as feeling secure in one’s financial future and having the ability to make choices that allow one to enjoy life (brüggen et al., 2017). according to kim and chatterjee (2013), family financial socialization is related to how individuals develop the attitudes, beliefs, knowledge, and skills necessary to manage their https://creativecommons.org/licenses/by-nc/4.0/ mailto:cwcopeland2@cau.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(2) 54 finances. financial socialization is associated with financial well-being (drever et al., 2015), but are there factors within financial socialization that are related to financial well-being more so than others? every community within the united states should experience financial well-being at a similar level (brown & robinson, 2016). however, some segments of the population lag behind others in terms of financial well-being. therefore, it is essential to understand what impacts or what limits financial well-being. wealth has been shown to impact financial well-being (atalay & edwards, 2022). thus, those with enough wealth to cover their day-to-day expenses, as well as enough wealth to accomplish their future goals, are more likely to experience financial freedom than those who do not have an adequate level of wealth (atalay & edwards, 2022). other factors, such as employment status and education, are directly linked to financial well-being (federal reserve board of governors, 2022). thus, those who are financially literate most often better manage their money and financial affairs and are more likely to have a higher level of financial well-being than those who do not (brugger et al., 2017). also, subjective constructs, such as selfcontrol and confidence, can impact financial well-being (porto & xiao, 2022). those with more confidence in their financial affairs and selfcontrol most likely experience higher levels of financial well-being than those who exhibit under-confidence and less self-control (porto & xiao, 2022). african americans have experienced financial well-being at a different rate than other racial groups (cobb, 2022). numerous studies have examined the black–white wealth gap, which highlights the reality that african americans have not generated as much wealth as european americans4 (mcintosh et al., 2020). in the context of this observation, it is worth remembering that wealth is associated with financial well-being (atalay & edwards, 2022). discrimination has certainly impacted african americans’ ability to secure and maintain 4 the use of this term aligns with recommendations made by the american psychological association (2015). employment, establish excellent credit histories, and accumulate wealth (atalay & edwards, 2022). financial knowledge and literacy likewise impact african americans’ financial well-being more than that of european americans (yakoboski et al., 2019). it is vital for african americans to experience financial well-being at the same rate as other segments for the united states to be vibrant (brown & robinson, 2016). fan and park (2021) combined the theoretical model of financial socialization by gudmunson and dane (2011) with fisher and fisher’s (1992) information motivation behavior model to develop a new model that examined the impact of financial socialization mediated by financial knowledge, goal setting, and self-control on an individual’s well-being (financial and subjective). fan and park (2021) used this model on young adults. in contrast, in this study, we used these models to assess the impact of financial socialization on well-being (financial and subjective) to determine if there was a difference in the relationships between african americans versus european americans. this study advances the literature related to the blackwhite wealth gap by concentrating explicitly on african americans’ financial well-being. this study provides insight into understanding why african americans have yet to experience financial well-being as much as others. literature review financial socialization at what point does a person choose to be better off than they are now? is it after they fail to pay a bill? is it when they cannot feed their children? is it when they are a child observing the family’s financial dynamics? jorgensen and savla (2010) surveyed college students about the perceived influence of parents on their financial attitudes and behaviors. they found the level of influence was a significant external factor since the degree of impact had magnitude and direction. according to shim and serido (2011), parental influence is 1.5 times greater than that of financial education and more than twice that of friends. young et al. 55 they showed that students who reported discussing financial matters with their parents and learning about managing money from them also reported healthier financial attitudes. financial socialization refers to “acquiring and developing values, attitudes, standards, norms, knowledge, and behaviors” (danes, 1994, p. 128) that provide the context for one’s financial practices. according to kim and chatterjee (2013), financial socialization is related to how individuals develop the attitudes, beliefs, knowledge, and skills necessary to manage finances. among these financial socialization agents, family influence, especially the influence of parents, is predominant (danes & haberman, 2007; fan & chatterjee, 2019). in this literature review, we will identify research on the various socialization agents (i.e., family, peers, and financial literacy programs) and their impact on financial behavior ultimately leading to a specific financial well-being status. it only makes sense that family influences are predominantly based on the typical length of time (birth to age 18) family members have to plant, nourish, and grow the ideas/ideals expected. most financial literacy materials aimed at children support parental involvement; however, there are not enough such programs, so instances where parents initiate socialization gives young people a tremendous head start. for example, pliner et al. (1996) found that children whose mothers gave them financial guidance and warmly communicated their economic expectations exhibited positive financial behaviors. similarly, children whose parents oversaw their spending were more likely to have confidence in their abilities as money managers (kim & chatterjee, 2013). while role modeling is the most prevalent method of parental financial socialization via day-to-day interaction, parental teaching is also a meaningful way to explicitly transfer financial knowledge and skills (serido & deenanath, 2016). in the poem, live your creed by langston hughes (hughes, n.d.), the following passage, “for i may misunderstand you and the fine advice you give, but there is no misunderstanding how you act and how you live,” emphasizes the effect of explicit education. not all research is pointed in the direction of the passage, as mentioned above. according to van campenhout (2015), the importance of family financial socialization has been further confirmed, as implicit learning is more prevalent in impacting financial behavior than explicit learning. the role of peers is also important. peers often influence college students; however, the influences received during their formative years may generally dominate other sources. further evidence of a parent’s role in explicit and implicit financial socialization comes from hibbert et al. (2004) who asked college and graduate students questions about what kind of financial behavior was modeled in their homes while they were growing up. they found that being raised in a financially prudent household, where parents saved and paid their bills on time, resulted in less self-reported engagement in negative financial behaviors, such as misusing credit cards and making unaffordable purchases, even after controlling for socioeconomic background. whether in line with explicit or implicit learning, most research indicates that parents play a critical role in enhancing knowledge of economic matters, which can positively affect financial well-being (agnew et al., 2018). in explicit learning, it is more critical; for example, sansone et al. (2018) evaluated the relationship between receiving an allowance (pocket money) in childhood and financial confidence in adulthood. they concluded that pocket money given by parents to their children became a crucial informal vehicle for developing a young person’s financial habits in later life. according to utkarsh et al. (2020), students who discussed their parents’ spending behavior, financial investments, and the importance of savings when growing up were more likely to display a positive attitude toward saving and tracking expenses, resulting in expected improved financial wellbeing. the work of utkarsh et al. supports the family financial socialization theory (ffst) by showing that family financial socialization indirectly affects financial behavior and financial well-being through financial literacy (zhao & zhang, 2020). utkarsh et al. (2020) further emphasized that parents must engage their children in financial discussions, and they added that this socialization at a young age will improve financial services review, 32(2) 56 a child’s positive outlook on future financial situations. financial literacy financial literacy5 refers to understanding finance and the capability to use knowledge to make sound financial decisions (hogarth & hilgert, 2002). financial education facilitates literacy; that is, mastery of finance-related knowledge and expertise is essential in undertaking daily transactions and wealth accumulation investments. it empowers people to manage their finances and provides long-lasting financial security for themselves and their families (sundarasen et al., 2016). as sundarasen et al. (2016) highlighted, financial literacy has been linked to savings and portfolio decisions. for example, people with a low level monetary education are more inclined to face issues with financial obligations (lusardi & tufano, 2009), less likely to take an interest in value ventures (christelis et al., 2010; van rooij et al., 2007), less apt to pick high performing mutual funds with lower expenses (hastings & tejeda-ashton, 2008), less inclined to aggregate and oversee wealth successfully (hilgert et al., 2003; stango & zinman, 2007), and less likely to antedate retirement (lusardi & mitchell, 2007). as these studies document, it cannot be denied that financial literacy is a crucial part of sound financial decision-making. further, and many young people wish they had more money-related information (lusardi et al., 2009). sundarasen et al. (2016) found that financial literacy contributed positively to financial satisfaction. low financial literacy is associated with poor financial decisions in equity investments, debt financing, and long-term retirement planning (lusardi & tufano, 2009). these financial decisions can decrease welfare (chu et al., 2017). households with a lower level of financial literacy can also make suboptimal decisions when choosing loans or mortgages (lusardi & tufano, 2009; moore, 2003; utkus & young, 2011) and suffer from problems such as debt accumulation (lusardi &tufano, 2009), 5 conceptually, financial literacy includes elements of skill. however, financial literacy programs and measures of financial literacy, grounded in the cognitive paradigm, tend to focus on knowledge bankruptcy, and foreclosure (gerardi et al., 2010). according to ammerman and stueve (2019), characterizing learning as implicit or explicit refers to the degree to which the agent’s intentions are made clear. implicit or vicarious learning occurs when attitudes, beliefs, and behaviors develop through the observation of a social agent rather than from direct experience or instruction. explicit learning, by contrast, refers to developing attitudes and beliefs through direct experience or intentional instruction (rettig & mortenson, 1986). an example of implicit learning would be when a child observes the financial behavior of their parents and compares it against the financial outcomes and well-being of the household. an example of explicit learning would be when a child manages their own money (in the case of an allowance). ammerman and stueve (2019) stated that since financial counseling and literacy programs aim to help clients enhance their financial well-being by adopting healthy financial behaviors, early intervention and socialization would be critical components of these programs. however, many financial literacy programs are grounded in cognitive learning approaches, emphasizing the role of information transmission and knowledge attainment in changing behavior (lusardi et al., 2015). this approach is not to dismiss financial literacy programs because some financial knowledge is better than none, but it is a necessary critique of the shortcomings of such programs. the cognitive learning paradigm assumes that individuals with more financial knowledge will better manage their financial resources, resulting in enhanced financial well-being (efland, 1995; greenwald, 1968; huston, 2010). this concept is easily applied to other favorable skill development, which would apply to developing financial behaviors. as discussed by tatom (2010), two of the most critical and early works on the effectiveness of financial education are attainment and ignore experiential learning, such as skill attainment through socialization (ammerman & stueve, 2019). thus, academic research has yet to resolve the dichotomy of this philosophy. young et al. 57 bernheim et al. (1997) and bernheim and garrett (2003). self-control zakaria et al. (2012) showed that savings, an indicator of responsible behavior, is also determined by financial literacy. they argued that the finding regarding the relationship between these two variables is conclusive, with all studies finding that having financial knowledge influences individuals to behave more responsibly. another important variable is selfcontrol. although it seems intuitive, only some realize the relationship between financial wellbeing and self-control. one way to define selfcontrol is that it constitutes the ability of one’s future self to control their current self. when selfcontrol failure occurs, people act non-optimally. they might, for example, procrastinate in their work, even though they know they would be better off spreading the workload over time (ariely & wertenbroch, 2002; fudenberg & levine, 2006). according to stromback et al. (2017), the behavioral life-cycle hypothesis further states that people’s financial behavior throughout life is determined by their ability to control impulses and the costs of exercising such self-control. the ability to control impulses is undoubtedly a key factor for long-term success in many areas of life. stromback et al. (2017) confirmed that studies that have explored the link between self-control and financial behavior have focused on specific financial decisions, such as retirement planning or credit use. achtziger et al. (2015) found that people with low self-reported self-control are more likely to engage in compulsive shopping, while gathergood (2012) found that people with self-control problems in the financial domain are likelier to suffer from credit withdrawals and unforeseen expenses on durables, leading to overindebtedness. respondents with good selfcontrol are more likely to save money regularly from their paychecks. this finding means they are better prepared to manage unforeseen expenses and more likely to have enough money for retirement (stromback et al., 2017). summary as highlighted in this review, the issue of financial well-being is a critical policymaking issue. increasing financial well-being can result in reduced poverty (iramani & lutfi, 2021). according to the consumer financial protection bureau (2015b), financial well-being entails having control over one’s day-to-day and monthto-month finances, having the capacity to absorb financial shocks, being on track to meet financial goals, and having the financial freedom to make choices that allow one to enjoy life (consumer financial protection bureau 2015b). if financial well-being is the ultimate goal, developing a positive attitude toward money (which includes inculcating the habit of savings, tracking expenses, and being prudent with money) is a strong predictor of financial well-being (utkarsh et al., 2020). it is reasonable to ask why would financial wellbeing not be a policy goal? given the choice between being better off and not, a prudent person should choose to be better off; consequently, they should take the relevant actions to achieve that goal. according to drever et al. (2015), healthy attitudes about saving and some frugality are necessary for skillful money management— positive views on budgeting support financial goal setting and planning. thus, a lack of materialism likely leads to an ability to live within one’s means. a conceptual model of family financial socialization the conceptual framework illustrated in figure 1 proposes interrelationships and relationships that influence parental financial socialization on financial and subjective well-being through three motivators: financial knowledge, goal setting, and self-control, either directly on financial behaviors or indirectly through financial skills on financial behaviors. with this theoretical foundation, we introduce a conceptual model that embeds family financial socialization through gudmunson and danes’ (2011) processes within personal finance. the model offers a unique perspective on how financial socialization can influence financial and subjective well-being (fan & park, 2021; fisher et al., 2006; fung, 2017; gudmunson & danes, 2011; limbu, 2017). additionally, the information motivation behavioral skills model, originally developed to financial services review, 32(2) 58 predict health behavior by fisher and fisher (1992), which has been shown to predict and change several behaviors, has been included. despite being initially developed in different disciplines, these two models provide theoretical support for constructing our conceptual framework. the literature has documented evidence that parental financial socialization, either explicitly or implicitly, can have long-term influences on individual financial behavior (fan & park, 2021; kim & chatterjee, 2013; norvilitis & maclean, 2010; tang et al., 2015). furthermore, financial well-being is achieved by promoting financial knowledge and perceived behavioral control (shim et al., 2009). the information motivation behavior skills model, originally developed to predict health behavior by fisher and fisher (1992), has been shown to describe several behaviors, including financial behaviors (fan & park, 2021; limbu, 2017). however, fan and park (2021) evaluated the influences of motivations to learn and perform that were significant predictors of financial behavior. based on their work, the conceptual framework aims to incorporate the behavioral motivation element into examining financial socialization and extends the current understanding of family socialization into subjective financial well-being as an outcome of family financial socialization. the theoretical framework (figure 1) proposes that financial socialization is significantly and positively associated with financial and subjective well-being, mediated by financial knowledge, goal setting, and self-control. these three motivators mediate financial and subjective well-being either directly through financial behaviors or indirectly through financial skills and then through financial behaviors. this study uses this framework to posit that there is a significant difference in this relationship between african americans and european americans. thus, we propose the following hypotheses: hypothesis 1. there is a significant difference in the relationship between financial socialization and well-being (financial and subjective), mediated by financial knowledge directly on financial management behaviors for african americans compared with european americans. hypothesis 2. there is a significant difference in the relationship between financial socialization and well-being (financial and subjective), mediated by financial knowledge indirectly through financial skills on financial behaviors for african americans compared with european americans. hypothesis 3. there is a significant difference in the relationship between financial socialization and well-being (financial and subjective), mediated by goal setting directly on financial management behaviors for african americans compared with european americans. hypothesis 4. there is a significant difference in the relationship between financial socialization and well-being (financial and subjective), mediated by goal setting indirectly through financial skills on financial behaviors for african americans compared with european americans. hypothesis 5. there is a significant difference in the relationship between financial socialization and well-being (financial and subjective), mediated by self-control directly on financial management behaviors for african americans compared with european americans. hypothesis 6. there is a significant difference in the relationship between financial socialization and well-being (financial and subjective), mediated by self-control indirectly through financial skills on financial behaviors for african americans compared with european americans. young et al. 59 figure 1. conceptual framework and hypotheses note: finsoc = financial socialization, knowl = financial knowledge, fingoal = financial goal-setting, selfcont = financial self-control, skill = financial skills, mgmt = financial management behavior, fwb = financial well-being, swb = subjective well-being. methodology data analysis this study used partial least squares (pls) path modeling, a variance-based structural equation modeling (sem) method, to examine the research model empirically (albort-morant et al., 2018). this methodology tests and estimates causal relations between latent or unobserved variables using a combination of data measured through observed items or variables. pls–sem includes the following multistage analyses. first, measurement model evaluation and second, structural model evaluation (ringle et al., 2018; wong, 2013). the measurement model requirement guarantees that the structural model will use only the constructs with reasonable indicator loadings, convergent validity, composite reliability (cr), and discriminant validity. structural model knowledge evaluation is meant to evaluate path coefficients and test their magnitude through management in the bootstrapping method. additionally, pls can readily operate with formative variables (chin et al., 2003). this study used the smart pls 4.0.9.5 software suite (ringle et al., 2015) to estimate the pls-sem model. data this study used data from the 2016 national financial well-being survey (nfwbs), which was designed and administered by the consumer financial protection bureau (cfpb). the data comprised respondents’ financial attitudes, skills, knowledge, experiences, behaviors, individual characteristics, and household and family financial status hypothesized to affect financial well-being. the nfwbs sample represents the noninstitutionalized adult population in the united states through multiple steps of the design process. initially, to represent the u.s. population proportionally, the nfwbs survey targeted 5,000 adults. the 2016 annual socioeconomic supplement of the current population survey (cps) was used to compare the proportions. finally, to offset the lower response rates for crucial population segments, the survey oversampled those below 200% of the federal financial services review, 32(2) 60 poverty level, those aged 62 and older, and african americans, non-hispanics, or hispanic americans (cfpb, 2015a). consequently, the total number of respondents who completed the survey was 6,394. the current study follows the cfpb recommendation of using weights to adjust for oversampling issues so that the results represent the total population. in the sample, we used self-identified african american and european american respondents. a final dataset of 4,872 participants was used in this study. sample weights were applied in the analysis. variables financial and subjective well-being variables. the primary outcome variable in this study was subjective well-being which was created as a latent construct comprising three observed variables that measured respondents’ levels of subjective well-being (cfpb, 2015a). the items included were, “i am satisfied with my life,” “i am optimistic about my future,” and “if i work hard today, i will be more successful in the future.” the answers for each item were coded using a 7-point likert scale ranging from 1 = strongly disagree to 7 = strongly agree. financial well-being was constructed as a latent variable composed of 10 indicators. each of the 10 items was on a likert agreement scale of 1–5, with some items reverse-coded. the items were (a) “because of my money situation, i feel like i will never have the things i want in life;” (b) “i am just getting by financially;” (c) “i am concerned that the money i have or will save won’t last;” (d) “giving a gift for a wedding, birthday, or other occasion would put a strain on my finances for the month;” (e) “i could handle a major unexpected expense;” (f) “i am securing my financial future;” (g) “i can enjoy life because of the way i’m managing my money;” (h) “i have money left over at the end of the month;” (i) “i am behind with my finances;” and (j) “my finances control my life.” financial socialization, financial knowledge, financial goal-setting, financial self-control. financial socialization was a latent construct comprising five observed variables regarding how respondents were financially socialized while growing up (cfpb, 2015a). the items were, (a) “while growing up at home, did your family do any of the following?” (b) “discussed family financial matters with me,” (c) “spoke to me about the importance of saving,” (d) “discussed how to establish a good credit rating,” (e) “taught me how to be a smart shopper,” and (f) “taught me that my actions determine my success in life.” the answers to each item were coded as 1 if respondents had such an experience and 0 otherwise. financial knowledge was an observed variable measured by responses to four items using knoll and houts’ (2012) financial knowledge questions (cfpb, 2015a). the items were (a) “understanding of long-term returns on investment,” (b) “understanding of stocks vs. bond vs. savings volatility,” (c) “understanding of the possibility of housing market losses,” and (d) “understanding of credit card minimum payments.” financial goal setting was an observed variable measured by a latent construct comprising four observed variables (cfpb, 2015a) (answers were given dichotomously). the items were (a) “do you have a current or recent financial goal?”, (b) “i set financial goals for what i want to achieve with my money,” (c) “i prepare a clear plan of action with detailed steps to achieve my financial goals” (each was on a 5-point likert scale), and (d) “confidence in own ability to achieve financial goals” (each was on a 4-point likert) (cfpb, 2015a). self-control was an observed variable measured by two items. the respondents were asked, “i am good at resisting temptation,” and “i am able to work diligently toward long-term goals.” the answers were coded with a 4-point scale ranging from 1 = not at all to 4 = completely well. financial skills and financial management behavior. the nfwbs survey provides modules of questions with predefined themes for both financial skills and financial management behavior variables. a latent construct was created for financial skills, comprising nine indicators reflecting respondents’ subjective evaluations of their financial skills and capabilities. the items in the survey were self-assessed financial skills, including: (a) “i know how to get myself to young et al. 61 follow through on my financial intentions,” (b) “i know where to find the advice i need to make decisions involving money,” (c) “i know how to make complex financial decisions,” (d) “i am able to make good financial decisions that are new to me,” (e) “i am able to recognize a good financial investment,” (f) “i know how to keep myself from spending too much,” and (g) “i know how to make myself save,” (h) “i know when i do not have enough information to make a good decision involving my money,” and (i) “i know when i need advice about my money” (cfpb, 2015a). each item was coded on an agreement scale of 1–5, with some items reverse-coded. financial management behavior was a latent construct, with four observed variables reflecting positive and desirable financial management behaviors. respondents were asked to indicate how often they had engaged in the following activities in the past six months: (a) “paid all the bills on time,” (b) “stayed within the budget or spending plan,” (c) “paid off credit card balance in full each month,” and (d) “checked the statements, bills, and receipts to make sure there were no errors” (cfpb, 2015a). the answers were coded with a 5-point scale ranging from 1 = never to 5 = always. results respondent profile table 1 provides descriptive statistics for the survey sample. the weighted percentages reflect the u.s. population profile. the sample was comprised of 84.4% european americans and 15.6% african americans. (the race of participants was categorized as european american, african american, asian american, or latin american in accordance with the standards established by the american psychological association (2015).) gender identification indicated that the dataset was close to half, with female respondents having a slight advantage (52.4%) with male respondents at 47.6%. overall, the 35 to 54 age group (32.2%) represented a more significant percentage of the dataset, with the 70 or older age group having the lowest representation at 15.1%. a slightly higher percentage of respondents were married (56.6%). approximately 67% of participants had some college education or less. forty-two percent of the sample was employed, 23% was retired, and 35% was unemployed for numerous reasons. table 1. respondent profile variable frequency percent gender male 2,317 47.6 female 2,555 52.4 racial group african american 758 15.6 european american 4,114 84.4 age group 34 and younger 1,383 28.4 35–54 1,571 32.2 55–69 1,182 24.3 70 or older 737 15.1 marital status married 2,755 56.6 not married 2,268 43.4 education level some college or less 3,277 67.3 bachelor’s degree 995 20.4 graduate degree 600 12.3 employment status full-time 2,057 42.2 not full-time 1,693 34.8 retired 1,121 23.0 financial services review, 32(2) 62 measurement model all constructs in the overall model satisfied the requirements for composite reliability (cr), and cronbach’s alpha6 was greater than 0.60 (gefen et al., 2000; nunnally & bernstein, 2007). acceptable convergent validity and discriminant validity was noted with each loading being greater than 0.50, average variance extracted (ave) was greater than 0.50, and the square root of ave being greater than each correlation coefficient (see tables 2 and 3 and bagozzi & yi, 1988; chin, 1998; hair et al., 2011). in order to perform a multigroup analysis according to racial identity in a later stage, convergent validity and discriminant validity for each racial group were also tested to insure the consistency and rigor of the measuring instrument. table 2 shows the adequate first-order constructs’ reliabilities and convergent validities. values were above the suggested thresholds suggested by hair and hult (2017). although the ave for financial knowledge of the african american group was below 0.50, the construct was retained due to the acceptable cr and cronbach’s alpha values. the heterotrait-monotrait (htmt) ratio of correlations was used to assess discriminant validity. henseler et al. (2015) argued a strong case for using this approach. in order to distinguish between the two factors, the htmt should be smaller than 0.90 (henseler et al., 2016). as table 3 shows, all correlations complied with this criterion. discriminant validity was established across all latent variables using the htmt criterion, as shown in table 3. the values of htmt’s confidence intervals of the correlations between constructs were less than 0.80 and did not include the value of 1.0 (hair et al., 2017), which supports the adequacy of discriminant validity. 6 cronbach’s alpha reliability coefficient typically ranges from 0 to 1.0. however, there is no lower limit to the coefficient. the closer cronbach’s alpha coefficient is to 1.0, the greater the internal consistency of the items in the scale. young et al. 63 table 2. measures of convergent validity overall african european construct items loadings alpha cr (ave) loadings alpha cr (ave) loadings alpha cr (ave) fgoals fgoal1 0.523 0.715 0.822 0.542 0.481 0.706 0.817 0.538 0.520 0.718 0.822 0.542 fgoal2 0.736 0.703 0.752 fgoal3 0.855 0.852 0.851 fgoal4 0.790 0.837 0.780 fmbeh fmbeh1 0.788 0.760 0.847 0.581 0.799 0.779 0.858 0.603 0.775 0.747 0.839 0.567 fmbeh2 0.804 0.855 0.785 fmbeh3 0.750 0.722 0.761 fmbeh4 0.701 0.722 0.687 fskills fskill1 0.847 0.896 0.916 0.551 0.841 0.902 0.921 0.571 0.853 0.894 0.915 0.550 fskill2 0.738 0.812 0.739 fskill3 0.780 0.806 0.769 fskill4 0.817 0.867 0.821 fskill5 0.739 0.742 0.746 fskill6 0.777 0.766 0.780 fskill7 0.803 0.835 0.808 fskill8 0.569 0.507 0.549 fskill9 0.552 0.532 0.536 fsoc fsoc1 0.668 0.806 0.865 0.563 0.717 0.826 0.877 0.588 0.665 0.796 0.860 0.551 fsoc2 0.795 0.777 0.789 fsoc3 0.718 0.760 0.717 fsoc4 0.787 0.816 0.769 fsoc5 0.776 0.759 0.767 fwb fwb1 0.820 0.870 0.902 0.606 0.772 0.851 0.887 0.569 0.835 0.875 0.905 0.616 fwb2 0.695 0.577 0.702 fwb3 0.733 0.734 0.745 fwb4 0.844 0.814 0.847 fwb5 0.789 0.823 0.778 fwb6 0.781 0.778 0.791 khk khk1 0.782 0.678 0.802 0.505 0.733 0.622 0.677 0.394 0.781 0.655 0.791 0.489 khk2 0.737 0.695 0.711 khk3 0.675 0.742 0.684 khk4 0.641 0.081 0.610 scon scon1 0.856 0.709 0.872 0.773 0.901 0.814 0.914 0.842 0.839 0.670 0.857 0.750 scon2 0.902 0.934 0.893 swb swb1 0.901 0.792 0.869 0.691 0.943 0.740 0.815 0.602 0.899 0.805 0.878 0.707 swb2 0.866 0.690 0.890 swb3 0.715 0.663 0.722 financial services review, 32(2) 64 table 3. discriminant validity coefficients race fgoals fmbeh fskills fsoc fwb khk scon swb overall fgoals fmbeh 0.714 fskills 0.761 0.650 fsoc 0.330 0.231 0.305 fwb 0.416 0.547 0.407 0.157 khk 0.220 0.315 0.209 0.188 0.191 scon 0.603 0.497 0.624 0.326 0.270 0.315 swb 0.454 0.350 0.434 0.241 0.405 0.098 0.377 african fgoals fmbeh 0.790 fskills 0.739 0.625 fsoc 0.337 0.224 0.316 fwb 0.427 0.452 0.292 0.103 khk 0.199 0.177 0.082 0.078 0.276 scon 0.573 0.444 0.485 0.237 0.165 0.280 swb 0.316 0.240 0.438 0.182 0.236 0.075 0.312 european fgoals fmbeh 0.717 fskills 0.778 0.681 fsoc 0.341 0.237 0.292 fwb 0.445 0.621 0.502 0.204 khk 0.296 0.368 0.271 0.205 0.179 scon 0.685 0.593 0.679 0.329 0.379 0.340 swb 0.466 0.379 0.453 0.276 0.480 0.188 0.408 structural model bootstrapping was used to determine whether the path connections for overall and racial-based models were significant. to estimate the model for each subsample, a set of 7,000 cases of bootstrapped sub-samples were created for the procedure (hair et al., 2011). the hypothetical testing and t values for each path relationship are shown in table 4. young et al. 65 table 4. path coefficient racial group path relationship beta value std. error t-value decision overall fgoals → fmbeh 0.327 0.023 13.973 *** supported fgoals → fskills 0.519 0.016 32.860 *** supported fmbeh → fwb –0.470 0.019 25.258 *** supported fskills → fmbeh 0.297 0.020 14.903 *** supported fsoc → fgoals 0.250 0.018 13.560 *** supported fsoc → khk 0.146 0.028 5.142 *** supported fsoc → scon 0.253 0.026 9.570 *** supported fwb → swb –0.377 0.020 18.933 *** supported khk → fmbeh 0.132 0.017 7.962 *** supported khk → fskills 0.038 0.017 2.243 ** supported scon → fmbeh 0.045 0.021 2.196 ** supported scon → fskills 0.262 0.021 12.486 *** supported african fgoals → fmbeh 0.424 0.048 8.827 *** supported fgoals → fskills 0.541 0.038 14.197 *** supported fmbeh → fwb –0.405 0.050 8.015 *** supported fskills → fmbeh 0.260 0.060 4.356 *** supported fsoc → fgoals 0.261 0.043 6.064 *** supported fsoc → khk 0.000 0.089 0.002 not supported fsoc → scon 0.203 0.055 3.666 *** supported fwb → swb –0.241 0.037 6.468 *** supported khk → fmbeh 0.096 0.088 1.088 not supported khk → fskills –0.112 0.076 1.475 not supported scon → fmbeh 0.043 0.051 0.839 not supported scon → fskills 0.207 0.043 4.762 *** supported european fgoals → fmbeh 0.283 0.019 14.663 *** supported fgoals → fskills 0.521 0.016 32.644 *** supported fmbeh → fwb –0.526 0.024 21.738 *** supported fskills → fmbeh 0.316 0.021 15.011 *** supported fsoc → fgoals 0.263 0.018 14.537 *** supported fsoc → khk 0.153 0.026 5.940 *** supported fsoc → scon 0.247 0.024 10.272 *** supported fwb → swb –0.443 0.029 15.215 *** supported khk → fmbeh 0.122 0.019 6.438 *** supported khk → fskills 0.042 0.018 2.408 ** supported scon → fmbeh 0.086 0.019 4.505 *** supported scon → fskills 0.263 0.018 14.291 *** supported notes. *p < .10, **p < 0.05, ***p < 0.01. the most important factors in this study were the cross-validated redundancy (q2) estimates of the latent construct. a blindfolding procedure was performed to assess the model’s predictive capability, given the parameters of pls-sem (chin, 1998). a q2 result above 0 indicates predictive relevance in overall and race-based models (fornell & cha, 1994). r-squared (r2) values for financial socialization, financial goals, financial skills, financial knowledge, self-control, financial services review, 32(2) 66 and financial management behavior ranged from substantial to moderate, respectively (cohen, 1988). the findings for the models are shown in table 5. table 5. r2 and cross-validated redundancy racial group constructs r2 q² (comm) q² (red) overall fgoals 0.062 0.266 0.033 fmbeh 0.393 0.306 0.224 fskills 0.471 0.443 0.256 fwb 0.221 0.450 0.128 khk 0.021 0.190 0.011 scon 0.064 0.300 0.048 swb 0.142 0.393 0.087 african fgoals 0.068 0.270 0.036 fmbeh 0.430 0.344 0.250 fskills 0.427 0.473 0.240 fwb 0.164 0.405 0.082 khk 0.000 0.010 –0.003 scon 0.041 0.446 0.033 swb 0.058 0.258 0.022 european fgoals 0.069 0.266 0.036 fmbeh 0.409 0.284 0.228 fskills 0.495 0.443 0.268 fwb 0.277 0.463 0.165 khk 0.023 0.165 0.011 scon 0.061 0.252 0.044 swb 0.196 0.422 0.125 to stringently compare the results across two racial groups, t statistics were calculated to evaluate the differences in path coefficients across models. as shown below in figure 2, a procedure described by chin et al. (2003) was used to perform a multigroup analysis. as there were two ethnic groups, two separate comparisons were tested in the analysis. figure 2. formula and multigroup analysis 𝑡 = 𝑃𝑎𝑡ℎ𝑔𝑟𝑜𝑢𝑝1 − 𝑃𝑎𝑡ℎ𝑔𝑟𝑜𝑢𝑝2 [√ (𝑚 − 1)2 (𝑚 + 𝑛 − 2) ∗ 𝑆. 𝐸.2𝑔𝑟𝑜𝑢𝑝1+ (𝑛 − 1)2 (𝑚 + 𝑛 − 2) ∗ 𝑆. 𝐸.2𝑔𝑟𝑜𝑢𝑝2 ] ∗ [√ 1 𝑚 + 1 𝑛] table 6 compares african and european americans’ financial socialization values based on their t value results. the effects of financial socialization, mediated by financial knowledge, self-control, financial goals, financial management behavior, financial skills, financial well-being, and subjective well-being were significantly different. young et al. 67 table 6. multigroup comparison between african americans and european americans relationship african americans european americans t value beta std. error beta std. error fsoc → fgoals → fmbeh 0.111 0.024 0.074 0.007 1.462 fsoc → fgoals → fmbeh → fwb -0.045 0.011 -0.039 0.004 0.485 fsoc → fgoals → fmbeh → fwb → swb 0.011 0.003 0.017 0.002 1.562 fsoc → fgoals → fskills 0.141 0.028 0.137 0.010 0.141 fsoc → fgoals → fskills → fmbeh 0.037 0.012 0.043 0.004 0.520 fsoc → fgoals → fskills → fmbeh → fwb -0.015 0.005 -0.023 0.002 1.332 fsoc → fgoals → fskills → fmbeh → fwb → swb 0.004 0.002 0.010 0.001 3.155 *** fsoc → khk → fmbeh 0.000 0.008 0.019 0.004 2.031 ** fsoc → khk → fmbeh → fwb 0.000 0.003 -0.010 0.002 2.497 ** fsoc → khk → fmbeh → fwb → swb 0.000 0.001 0.004 0.001 3.287 *** fsoc → khk → fskills 0.000 0.009 0.006 0.003 0.658 fsoc → khk → fskills → fmbeh 0.000 0.003 0.002 0.001 0.770 fsoc → khk → fskills → fmbeh → fwb 0.000 0.001 -0.001 0.000 0.952 fsoc → khk → fskills → fmbeh → fwb → swb 0.000 0.000 0.000 0.000 1.387 fsoc → scon → fmbeh 0.009 0.010 0.021 0.005 1.098 fsoc → scon → fmbeh → fwb -0.004 0.004 -0.011 0.003 1.542 fsoc → scon → fmbeh → fwb → swb 0.001 0.001 0.005 0.001 2.468 ** fsoc → scon → fskills 0.042 0.015 0.065 0.007 1.392 fsoc → scon → fskills → fmbeh 0.011 0.005 0.021 0.003 1.610 fsoc → scon → fskills → fmbeh → fwb -0.004 0.002 -0.011 0.001 2.245 ** fsoc → scon → fskills → fmbeh → fwb → swb 0.001 0.001 0.005 0.001 3.704 *** notes. *p < .10, **p < 0.05, ***p < 0.01. discussion relations between financial socialization, financial controls, and financial management the findings from this study confirm the link between financial socialization effects, financial controls, and financial management abilities in universally determining financial and subjective well-being. all the mediating factors toward financial well-being and subjective well-being were found to be significantly related to financial socialization toward financial well-being and subjective well-being. as such, subjective financial well-being was also significantly associated with financial well-being. from the five mediating factors, the overall model indicates that financial knowledge, financial goals, financial self-control, financial skills, and financial management behaviors are significant financial services review, 32(2) 68 predecessors to financial well-being. the results correspond to earlier studies conducted in an american context (e.g., danes, 1994; danes & haberman, 2007; fan & chatterjee, 2019; kim & chatterjee, 2013; van campenhout, 2015). it is therefore supposed that african americans, like the general community, maintain that their financial socialization mainly develops their financial and subjective well-being. the findings relate well to african americans’ concern about the financial socialization process. moreover, being african american, they generally follow the same path to financial well-being and subjective well-being. furthermore, they also tended to relate financial well-being to subjective well-being. the findings generally have a few exceptions with additional details when the overall model is divided into two based on ethnicity. except for financial knowledge and self-control, the predictors of financial socialization toward financial well-being across ethnic groups were mixed. however, for european americans, the findings match the overall model supporting financial well-being. it is evident that despite being broadly similar in their financial socialization process, european americans’ financial well-being can be predicted by additional financial control factors. hence, it is essential to look at the findings of the latter stage. multi-group comparison according to racial identity the findings of the multigroup analysis indicate significant differences between african and european americans’ financial socialization when mediated by financial skill and financial management behavior factors. when financial goals mediate financial socialization, only financial skills show significant differences between groups. financial skills indicate no significant differences when socialization is mediated by financial knowledge. however, financial management behaviors demonstrate significant differences. more surprisingly, financial self-control was a significant factor for financial skills and financial management behaviors in mediating the differences between african and european americans gaining financial well-being, leading to subjective wellbeing. considering the findings, hypotheses two, three, five, and six were supported and cannot be rejected in this study. the first and fourth hypotheses were not supported and were rejected in the study. irrespective of the insightful influence of ethnicity and culture on americans’ financial socialization process, the findings imply that behavior self-control is one of the most impactful ways for african americans to close the wellbeing (financial and subjective) gap. in essence, african americans’ financial self-control as a factor affecting well-being differs from that of european americans. the findings support the theories of earlier studies, which state that selfcontrol influences financial behavior and subjectively perceived financial well-being (strömbäck et al., 2017; vuković & pivac, 2021). over the years, a drastic increase in financial literacy focus and easier access to financial services have increased the likelihood of developing financial goals (hudson et al., 2017). however, setting financial goals when considering financial skills remains a significant gap between african and european americans. moreover, the findings suggest that financial knowledge alone cannot overcome financial management behavior changes. access to financial services remains an issue for african americans in narrowing the gap with european americans (sun et al., 2022). setting financial goals and following through with the necessary financial skill set continue to separate african and european americans in terms of their wellbeing (financial and subjective). as a result, african americans have continued to be socialized in their inherited cultures and those of european americans, which sometimes conflicts. this study clarifies some differences between the impact of financial socialization on african american and european americans’ well-being (financial and subjective). limitations this study has the following limitations that must be considered. first, the results do not imply any causal relationship when studying the crosssectional dataset. second, the study focused on financial socialization, which is only one financial and subjective well-being sub-domain. young et al. 69 other major domains, such as past financial exclusion and health access-related well-being, were not included in this study, but can be examined when applicable datasets are available. future studies are needed to investigate financial socialization’s effect and long-term influences on african americans’ financial behavior and wellbeing through a lens of historical discrimination. finally, the findings of this analysis can only be applied to financially excluded americans, mainly african americans. relationships among the variables, including the significant direct influences of generational financial socialization on financial knowledge, goal setting, and selfcontrol and its indirect impact on financial skill and behavior and financial and subjective wellbeing, may not be generalized to younger generations of african americans. the influences of financial self-control and behavior undoubtedly require further investigation by other groups of americans. however, the relationships in this study applied to african americans can shed light on and provide a starting point for future studies on younger generations of african americans. thus, this conceptual framework could be a starting point for future research to understand other groups’ financial socialization and financial and subjective well-being. implications and conclusion financial socialization is influential in developing and bringing about change in wellbeing (financial and subjective). at the same time, it provides insights into the reasoning as to what drives well-being (financial and subjective) and why. consequently, understanding the wellbeing (financial and subjective) gap between 7 although the effects of racism on well-being have been document in the literature, the models and methodological approach utilized in this study were designed to hold racism constant. the methodological approach was made strategically to enable the exploration of specific impacts of other variables within environments where the level of racial bias or discrimination is assumed to be uniform. this methodology allows for a clearer understanding of how various factors—such as socio-economic status, education, or geographic location—operate independently of the direct effects of racism. by examining outcomes under this controlled condition, americans and their respective financial socialization experiences remains essential to bridging this gap. however, the complexity of the financial services industry and society has directed americans toward solidifying the wellbeing gap. increased focus on the well-being gap may expedite the process of increasing the level of financial literacy across all age groups, enhancing americans’ financial socialization. as such, it is imperative for educators, employers, and policymakers to strategically address what sets ethnic groups apart from and what integrates them into their well-being (financial and subjective).7 this study highlights the importance of understanding the financial socialization of americans’ pathways to well-being (financial and subjective). results also provide insight into ways to improve financial skills development and financial management behaviors for all americans, despite ethnic diversity. since african americans comprise the principal share of low-income americans (creamer, 2021), knowing more about their financial inclusion will push financial literacy to the next level to embrace all americans. such knowledge will also help to introduce financial literacy at an earlier stage in life, thus reinforcing financial self-control and impacting financial management behaviors later in life. thus, meticulously implementing financial literary at different stages of development and having a clear understanding of what makes ethnic groups different will prove pivotal to addressing the well-being (financial and subjective) gap between african and european americans. it is possible to identify patterns, disparities, and potential causal relationships that the pervasive influence of racial discrimination might otherwise obscure. this approach is crucial in social science, sociology, psychology, and financial well-being, where isolating the distinct effects of systemic and interpersonal racism from other variables is essential for devising effective interventions and policies. it provides a nuanced understanding of inequality and injustice, facilitating targeted strategies that address specific barriers to equity and inclusion while acknowledging the broader context of racism. financial services review, 32(2) 70 references achtziger, a., hubert, m., kenning, p., raab, g., & reisch, l. (2015). debt out of control: the links between selfcontrol, compulsive buying, and real debts. journal of economic psychology, 49, 141–149. https://doi.org/10.1016/j.joep.2015.0 4.003 agnew, s., maras, p., & moon, a. (2018). gender differences in financial socialization in the home—an exploratory study. international journal of consumer studies, 42(3), 275–282. https://doi.org/10.1111/ijcs.12415 albort-morant, g., leal-rodríguez, a. l., & de marchi, v. (2018). absorptive capacity and relationship learning mechanisms as complementary drivers of green innovation performance. journal of knowledge management, 22(2), 432–452. https://doi.org/10.1108/jkm-07-20170310 american psychological association. (2015). guidelines for psychological practice with transgender and gender nonconforming people. american psychologist, 70(9), 832–864. https://doi.org/10.1037/a0039906 ammerman, d. a., & stueve, c. (2019). childhood financial socialization and debt-related financial well-being indicators in adulthood. journal of financial counseling and planning, 30(2), 213–230. https://doi.org/10.1891/10523073.30.2.213 ariely, d., & wertenbroch, k. (2002). procrastination, deadlines, and performance: self-control by precommitment. psychological science, 13(3), 219–224. https://doi.org/10.1111/14679280.00441 atalay, k., & edwards, r. (2022). house prices, housing wealth, and financial well-being. journal of urban economics. 129(c). https://doi.org/10.1016/j.jue.2022.10 3438 bagozzi, r., & yi, y. (1988). on the evaluation of structural equation models. journal of the academy of marketing science, 16(1), 74–94. https://doi.org/10.1007/bf02723327 bernheim, b. d., & garrett, d. m. (2003). the effects of financial education in the workplace: evidence from a survey of households. journal of public economics, 87(7–8), 1487– 1519. bernheim, b. d., garrett, d. m., & maki, d. m. (1997). education and savings: the long-term effects of high school curriculum mandates. nber working paper, (6085). brown, c., & robinson, l. (2016). breaking the cycle: from poverty to financial security for all. policylink. https://www.policylink.org/sites/defa ult/files/breakingthecycle_0.pdf brüggen, e. c., hogreve, j., holmlund, m., kabadayi, s., & löfgren, m. (2017). financial well-being: a conceptualization and research agenda. journal of business research, 79, 228–237. https://doi.org/10.1016/j.jbusres.201 7.03.013 chin, w. (1998). the partial least squares approach for structural equation modeling. in g. a. marcoulides (ed.), modern methods for business https://doi.org/10.1016/j.joep.2015.04.003 https://doi.org/10.1016/j.joep.2015.04.003 https://doi.org/10.1037/a0039906 https://doi.org/10.1111/1467-9280.00441 https://doi.org/10.1111/1467-9280.00441 young et al. 71 research (pp. 295–336). lawrence erlbaum associates publishers. chin, w., marcolin, b., & newsted, p. (2003). a partial least squares latent variable modelling approach for measuring interaction effects: results from a monte carlo simulation study and an electronic-mail emotion/adoption study. information systems research, 14(2), 189–217. https://doi.org/10.1287/isre.14.2.189. 16018 christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. white economic review, 54(1), 18–38. https://doi.org/10.1016/j.euroecorev. 2009.04.001 chu, z., wang, z., xiao, j. j., & zhang, w. (2017). financial literacy, portfolio choice and financial well-being. social indicators research, 132, 799–820. cobb, a. (2022). financial literacy education for and its effect of financial wellbeing. university of south california proquest dissertation publishing 2022, 29325135. cohen, j. (1988). statistical power analysis for the behavioral sciences (2nd ed.). routledge. https://doi.org/10.4324/97802037715 87 consumer financial protection bureau. (2015a). measuring financial wellbeing: a guide to understanding the cfpb financial well-being scale. retrieved from https://files.consumerfinance.gov/f/2 01512_cfpb_financial-well-beinguser-guide-scale.pdf. consumer financial protection bureau. (2015b). financial well-being: the goal of financial education. report, iowa city, ia: consumer financial protection bureau. creamer, j. (2021, december 9). inequalities persist despite decline in poverty for all major race and hispanic origin groups. census.gov. https://www.census.gov/library/stori es/2020/09/poverty-rates-for-blacksand-hispanics-reached-historic-lowsin-2019.html danes, s. m. (1994). parental perceptions of children’s financial socialization. journal of financial counseling and planning, 5, 127– 149. danes, s. m., & haberman, h. (2007). teen financial knowledge, self-efficacy, and behavior: a gendered view. journal of financial counseling and planning, 18(2), 48–60. drever, a., white, e., kalish, c., quest, n., hoaglan, e., & nelms, e. (2015). foundations of financial well-being: insights into the role of executive function, financial socialization, and experience-based learning in childhood and youth. the journal of consumer affairs, 49(1), 13–38. https://doi.org/10.1111/joca.12068 efland, a. d. (1995). the spiral and the lattice: changes in cognitive learning theory with implications for art education. studies in art education, 36(3), 134–153. https://doi.org/10.2307/1320905 fan, l., & park, n. (2021). factors mediating the associations between financial socialization and well-being of young adults: testing a conceptual framework. journal of financial planning and counseling, 2(2), 202– https://doi.org/10.1287/isre.14.2.189.16018 https://doi.org/10.1287/isre.14.2.189.16018 https://doi.org/10.1111/joca.12068 https://doi.org/10.2307/1320905 financial services review, 32(2) 72 216. https://doi.org/10.1891/jfcp20-00056 fan, l., & chatterjee, s. (2019). financial socialization, financial education, and student loan debt. journal of family and economic issues, 40(1), 74–85. https://doi.org/10.1007/s10834-0189589-0 federal reserve board of governors. (2022). economic well-being of us households in 2021. federal reserve board publications. board of governors of the federal reserve system. https://www.federalreserve.gov/publi cations/files/2021-report-economicwell-being-us-households202205.pdf fisher, j. d., & fisher, w. a. (1992). changing aids-risk behavior. psychological bulletin, 111(3), 455– 474. https://doi.org/10.1037/00332909.111.3.455 fisher, j. d., fisher, w. a., amico, k. r., & harman, j. j., (2006). an information-motivation-behavioral skills model of adherence to antiretroviral therapy. health psychology, 25(4), 462–73. https://doi.org/10.1037/02786133.25.4.462 fornell, c., & cha, j. (1994). partial least squares. advanced methods of marketing research, 407, 52–78. fudenberg, d., & levine, d. k. (2006). a dual-self model of impulse control. american economic review, 96(5), 1449–1476. fung, m. s. k. (2017). an imb model testing via endorser types and advertising appeals on young people’s attitude towards cervical cancer prevention advertisement in hong kong. young consumers, 18(1), 1–18. https://doi.org/10.1108/yc-07-201600620 gathergood, j. (2012). self-control, financial literacy and consumer overindebtedness. journal of economic psychology, 33, 590–602. https://doi.org/10.1016/j.joep.2011.1 1.006 gefen, d., straub, d., & boudreau, m. (2000). structural equation modeling and regression: guidelines for research practice. communications of the association of information systems, 4(7). https://doi.org/10.17705/1cais.0040 7 gerardi, k. s., rosen, h. s., & willen, p. s. (2010). the impact of deregulation and financial innovation on consumers: the case of the mortgage market. the journal of finance, 65(1), 333–360. https://doi.org/10.1111/j.15406261.2009.01531.x greenwald, a. g. (1968). cognitive learning, cognitive response to persuasion, and attitude change. in a. g. greenwald, t. c. brock, & t. m. ostrom (eds.), psychological foundations of attitudes (pp. 147-170). academic press. https://doi.org/10.1016/b978-1-48323071-9.50012-x gudmunson, c. g., & danes, s. m. (2011). family financial socialization: theory and critical review. journal of family and economic issues, 32(4), 644–667. https://doi.org/10.1007/s10834-0119275-y hair, j., & hult, g. (2017). a primer on partial least squares structural https://doi.org/10.1891/jfcp-20-00056 https://doi.org/10.1891/jfcp-20-00056 https://doi.org/10.17705/1cais.00407 https://doi.org/10.17705/1cais.00407 young et al. 73 equation modeling (pls-sem). sage publications. hair, j., ringle, c., & sarstedt, m. (2011). pls-sem: indeed a silver bullet. journal of marketing theory and practice, 19, 139–151. https://doi.org/10.2753/mtp10696679190202 hastings, j. s., & tejeda-ashton, l. (2008). financial literacy, information, and demand elasticity: survey and experimental evidence. nber working paper no. 14538. henseler, j., hubona, g., & ray, p. (2016). using pls path modeling in new technology research: updated guidelines. industrial management & data systems, 116(1), 2–20. https://doi.org/10.1108/imds-092015-0382 henseler, j., ringle, c. m. & sarstedt, m. (2015). a new criterion for assessing discriminant validity in variancebased structural equation modeling. journal of the academy of marketing science, 43(1), 115–135. https://doi.org/10.1007/s11747-0140403-8 hibbert, j. r., beutler, i. f., & martin, t. m. (2004). financial prudence and next generation financial strain. journal of financial counseling and planning, 15(2), 51–59. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309. hogarth, j. m., & hilgert, m. a. (2002). financial knowledge, experience and learning preferences: preliminary results from a new survey on financial literacy. consumer interest annual, 48(1), 1–7. hudson, c. r., young, j., anong, s., hudson, e., & davis, e. (2017). african american financial socialization. the review of black politically economy, 44, (3–4). https://doi.org/10.1007/s12114-0179258-9 hughes, l. (n.d.). live your creed poem. success minded. https://successminded.co/liv e-your-creed-poem/ huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44(2), 296–316. https://doi.org/10.1111/j.1745 6606.2010.01170.x iramani, r., & lutfi, l. (2021). an integrated model of financial well-being: the role of financial behavior. accounting, 7(3), 691–700. https://doi.org/10.5267/j.ac.2020.12. 007 jorgensen, b. l., & savla, j. (2010). financial literacy of young adults: the importance of parental socialization. family relations, 59(4), 465–478. https://doi.org/10.1111/j.1741-3729. 2010.00616.x. kim, j., & chatterjee, s. (2013). childhood financial socialization and young adults’ financial management. journal of financial counseling and planning, 24(1), 61–79, 91–92. knoll, m. a. z., & houts, c. r. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. the journal of consumer affairs, 46(3), 381–410. https://doi.org/10.1007/s12114-017-9258-9 https://doi.org/10.1007/s12114-017-9258-9 https://successminded.co/live-your-creed-poem/ https://successminded.co/live-your-creed-poem/ financial services review, 32(2) 74 https://doi.org/10.1111/j.17456606.2012.01241.x limbu, y. b. (2017). credit card knowledge, social motivation, and credit card misuse among college students: examining the informationmotivation-behavioral skills model. the international journal of bank marketing, 35(5), 842–856. https://doi.org/10.1108/ijbm-042016-0045 lusardi, a., michaud, p. c., & mitchell, o. s. (2015). using a life cycle model to evaluate financial literacy program effectiveness (working paper no. 2015-5). global financial literacy excellence center, george washington university. retrieved from: https://gflec.org/wpcontent/uploads/2016/08/wp-20155-using-a-life-cycle-model-toevaluate-financial-literacyprogram-effectiveness.pdf lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54(1), 205–224. https://doi.org/10.1016/j.jmoneco.20 06.12.001 lusardi, a., mitchell, o. s., & curto, v. (2009). financial literacy and financial sophistication among older americans (no. w15469). national bureau of economic research. retrieved from: https://repository.upenn.edu/prc_pap ers/346/ lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and over indebtedness. nber working paper 14808, national bureau of economic research, cambridge, ma. retrieved from: https://www.nber.org/papers/w14808 mcintosh, k., moss, e., nunn, r., & shambough j. (2020). examining the black-white wealth gap. upfront. brookings, washington, d.c. retrieved from: https://www.brookings.edu/blog/upfront/2020/02/27/examining-theblack-white-wealth-gap/ moore, d. l. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences. washington state department of financial institutions. norvilitis, j. m., & maclean, m. g. (2010). the role of parents in college students’ financial behaviors and attitudes. journal of economic psychology, 31, 55–63. https://doi.org/10.1016/j.joep.2009.1 0.003 nunnally, j. c., & bernstein, i. h. (2007). psychometric theory (3rd ed.). nunnally and bernstein. pliner, p., freedman, j., abramovitch, r., & darke, p. (1996). children as consumers: in th e laboratory and beyond. in p. lunt & a. furnham (eds.), economic socialization: the economic beliefs and behaviors of young people (pp. 35–46). edward elgar. porto, n., & xiao, j. j. (2022). the role of consumer financial confidence on financial well-being. consumer interest annual. american council of consumer interest, 68(1), 1–5. https://www.consumerinterests.org/a ssets/docs/cia/cia2022/portonilto ncia2022.pdf rettig, k. d., & mortenson, m. (1986). household production of financial https://doi.org/10.1016/j.jmoneco.2006.12.001 https://doi.org/10.1016/j.jmoneco.2006.12.001 https://www.brookings.edu/blog/up-front/2020/02/27/examining-the-black-white-wealth-gap/ https://www.brookings.edu/blog/up-front/2020/02/27/examining-the-black-white-wealth-gap/ https://www.brookings.edu/blog/up-front/2020/02/27/examining-the-black-white-wealth-gap/ https://www.consumerinterests.org/assets/docs/cia/cia2022/portoniltoncia2022.pdf https://www.consumerinterests.org/assets/docs/cia/cia2022/portoniltoncia2022.pdf https://www.consumerinterests.org/assets/docs/cia/cia2022/portoniltoncia2022.pdf young et al. 75 management competence. human resources research, 1887-1987: proceedings/edited by ruth e. deacon and wallace e. huffman. ringle, c. m. (2015). partial least squares structural equation modelling (plssem) using smartpls 3. computational data analysis and numerical methods vii wcdanm, portugal. ringle, c. m., sarstedt, m., mitchell, r., & gudergan, s. p. (2018). partial least squares structural equation modeling in hrm research. the international journal of human resource management, 31(12), 1617–1643. https://doi.org/10.1080/09585192.20 17.1416655 sansone, d., rossi, m., & fornero, e. (2018). four bright coins shining at me: financial education in childhood, financial confidence in adulthood. journal of consumer affairs, 53(2), 630–651. https://doi.org/10.1111/joca.12207 serido, j., & deenanath, v. (2016). financial parenting: promoting financial selfreliance of young consumers. in j. j. xiao (ed.), handbook of consumer finance research (2nd ed.) (pp. 291– 300). springer. shim, s., & serido, j. (2011). young adults’ financial capability https://www.nefe.org/_images/resear ch/aplus-wave-2/aplus-wave2-final-report.pdf (accessed 9 march 2020). shim, s., xiao, j. j., barber, b. l., & lyons, a. c. (2009). pathways to life success: a conceptual model of financial well-being for young adults. journal of applied developmental psychology, 30(6), 708–723. https://doi.org/10.1016/j.appdev.200 9.02.003 stango, v., & zinman, j. (2007). fuzzy math and red ink: when the opportunity cost of consumption is not what it seems. working paper, dartmouth college. strömbäck, c., lind, t., skagerlund, k., vastfjail, d., & tanghog, g. (2017). does self-control predict financial behavior and financial well-being? journal of behavioral and experimental finance., 14(1), 30–38. https://doi.org/10.1016/j.jbef.2017.0 4.002 sun, s., chen, y., ansong, d., huang, j., & sherraden, m. s. (2022). household financial capability and economic hardship: an empirical examination of the financial capability framework. journal of family and economic issues, 43(4), 716–729. https://doi.org/10.1007/s10834-02209816-5 sundarasen, s. d. d., rahman, m. s., othman, n. s., & danaraj, j. (2016). impact of financial literacy, financial socialization agents, and parental norms on money management. journal of business studies quarterly, 8(1), 137. https://doi.org/10.35484/pssr.2020(4i)56 tang, n., baker, a., & peter, p. c. (2015). investigating the disconnect between financial knowledge and behavior: the role of parental influence and psychological characteristics in responsible financial behaviors among young adults. the journal of consumer affairs, 49(2), 376. https://doi.org/10.1111/joca.12069 tatom, j. (2010). financial wellbeing and some problems in assessing its link to https://doi.org/10.1080/09585192.2017.1416655 https://doi.org/10.1080/09585192.2017.1416655 https://doi.org/10.1016/j.appdev.2009.02.003 https://doi.org/10.1016/j.appdev.2009.02.003 https://doi.org/10.1111/joca.12069 financial services review, 32(2) 76 financial education. john tatom networks financial institute at indiana state university 1 october 2010. retrieved from: https://mpra.ub.unimuenchen.de/26411/1/mpra_paper _26411.pdf utkarsh, pandey, a., ashta, a., spiegelman, e., & sutan, a. (2020). catch them young: impact of financial socialization, financial literacy and attitude towards money on financial well‐being of young adults. international journal of consumer studies, 44(6), 531–541. https://doi.org/10.1111/ijcs.12583 utkus, s. p., & young, j. (2011). financial literacy and 401 (k) loans. financial literacy: implications for retirement security and the financial marketplace, 59. https://repository.upenn.edu/cgi/view content.cgi?article=1214&context=p rc_papers van campenhout, g. (2015), revaluing the role of parents as financial socialization agents in youth financial literacy programs. journal of consumer affairs, 49(1), 186–222. https://doi.org/10.1111/joca.12064 van rooij, m. c., kool, c. j., & prast, h. m. (2007). risk-return preferences in the pension domain: are people able to choose? journal of public economics, 91(3–4), 701–722. https://doi.org/10.1016/j.jpubeco.200 6.08.003 vuković, m., & pivac, s. (2021). does financial behavior mediate the relationship between self-control and financial security? croatian operational research review, 12(1), 27–36. https://doi.org/10.17535/crorr.2021.0 003 wong, k. k. (2013). partial least squares structural equation modeling (plssem) techniques using smartpls. marketing bulletin, 24, 1–32. yakoboski, p. j., lusardi, a., & hesler, a. (2019). financial literacy and wellbeing among african–americans: new insights from the personal finance (p-fin) index. optimizing financial education utilization. financial education. accesslex resource collections. https://gflec.org/wpcontent/uploads/2019/11/tiaainstitute-gflec_african-americanp-fin-index_november-2019.pdf zakaria, r. h., jaafar, n. i. m., & marican, s. (2012). financial behavior and financial position: a structural equation modelling approach. middle-east journal of scientific research, 12(10), 1396–1402. https://doi.org/10.5829/idosi.mejsr.2 012.12.10.79 zhao, h., & zhang, l. (2020). talking money at home: the value of family financial socialization. international journal of bank marketing, 38(7), 1617–1634. https://doi.org/10.1108/ijbm-042020-0174 https://gflec.org/wp-content/uploads/2019/11/tiaa-institute-gflec_african-american-p-fin-index_november-2019.pdf https://gflec.org/wp-content/uploads/2019/11/tiaa-institute-gflec_african-american-p-fin-index_november-2019.pdf https://gflec.org/wp-content/uploads/2019/11/tiaa-institute-gflec_african-american-p-fin-index_november-2019.pdf https://gflec.org/wp-content/uploads/2019/11/tiaa-institute-gflec_african-american-p-fin-index_november-2019.pdf financial services review, 32(1) 29 altruistic bequests: giving motive and receipt expectation using the 2019 survey of consumer finances jason anderson1, jeffrey furlong2, and stuart heckman3 abstract this study uses altruistic bequest theory to update research by kao et al. (1997) exploring the factors driving the giving and receiving of bequests using data collected from the 2019 survey of consumer finances (scf). the bequest gift motive multinomial logit model uncovered evidence in support of altruistic bequest theory, specifically regarding family savings priorities and charitable giving. on the other hand, volunteerism was not associated with the bequest gift motive. respective to the base categories, receiving an inheritance, self-employment, marital status, race, attitude toward leaving a bequest, charitable giving, risk tolerance, family savings priorities, poor health status, age, income, financial assets, and nonfinancial assets were more likely to predict giving a bequest versus no bequest. for bequest receipt expectations, a binomial logit model showed receiving an inheritance, education, marital status, race, presence of living parents, age, income, and financial assets were the most important predictors of receiving a bequest. for both models, results show economic and attitudinal variables are important drivers for the giving and receiving of bequests. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation anderson, j., furlong, j., & heckman, s. (2024). altruistic bequests: giving motive and receipt expectation using the 2019 survey of consumer sciences. financial services review, 32(1), 2946. introduction as the baby boomer generation (those born between 1946 and 1964) continues its journey into retirement, the transfer of wealth is a noteworthy topic of study in the field of personal financial planning due to, at least in part, the estimated value of the transfer. a 2010 study from metlife (2010) using the survey of 1 corresponding author (jasonanderson@ksu.edu). kansas state university, manhattan, usa. 2kansas state university, manhattan, usa. 3kansas state university, manhattan, usa. consumer finances (scf) estimated the baby boomer generation inherited $11.6 trillion from the former generation, money that will transfer to the next if not spent. this vast sum has undoubtedly garnered interest from researchers and practitioners alike. indeed, grable (2013) noted that this generation is “one of the most discussed, studied, and evaluated groups of https://creativecommons.org/licenses/by-nc/4.0/ mailto:jasonanderson@ksu.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(1) 30 people the world has ever known” (p.7). this current study continues this discussion on a critical aspect of wealth transfer: the giving and receiving of bequests. the purpose of this research is threefold. the first is to update the bequest research conducted by kao et al. (1997). kao et al.’s research used altruistic bequest theory to explore the relationship between respondents’ sociodemographic characteristics on their expectations of receiving and leaving bequests. the authors found significant relationships between certain respondent sociodemographic traits (education, marital status, race, surviving parents, and the number of siblings) and expecting to receive a bequest. an exploration of the expectation to leave a bequest uncovered both positive (self-employment, middle age, nonliquid asset holdings, education, marital status, and a favorable attitude toward bequests) and negative relationships (total number of children and disability). many of the original variables found in the kao et al. (1997) study are further investigated here using the newer 2019 scf dataset. the second purpose of this study is to improve the model by adding variables within the scf that better align with altruistic bequest theory. finally, this study explores the relationship between longevity expectation and bequest motive. literature review generational transfers rossi and rossi (1990) examined the various beneficiary types stated in respondents' wills using information from a sectional threegeneration survey conducted in the years 1984– 1985. both the likely (hypothetical) beneficiaries and the beneficiaries of the respondents' actual wills, if any, were inquired about in the poll. the most frequent beneficiaries of wills were wives and children, according to descriptive statistics from the study (rossi & rossi, 1990, p. 475). when the actual beneficiaries of wills were examined, it was found that women were more likely than males to leave provisions for close relatives (such as a child, parent, sibling, or niece). it was suggested that for those who are childless, nieces and nephews might stand in for one's children, and for those who are not married, siblings could stand in for a spouse. coleman and ganong (1998) explored the effect of divorce and remarriage on leaving bequests. the authors found that genetic relationships, patriarchal lineage, and family ties influenced perception on regarding which family members should be included in the will. reciprocity, a concept closely aligned with altruism, was found to highly influence the individual’s definition of what constitutes a family. in contrast to the need to leave a bequest based on proximity and reciprocity, the obligation to include next of kin was stronger. even when reciprocity was a crucial contextual factor in establishing the limit of family, relatives who voluntarily chose to be included in a parent's will were less likely to do so (e.g., daughters-in-law and stepgrandchildren). grandchildren were more likely to be listed as beneficiaries following a parent's divorce or remarriage. engler-bowles and kart (1983) surveyed 60 wills from rural northwest ohio (wood county) from probate records from 1820-1967, focusing on how changes in intergenerational relationships affected patterns of inheritance. based on the degree of familial duty, inheritance patterns were categorized into three groups: familistic inheritance, articulated inheritance, and disinheritance. familistic inheritance patterns comprised the bulk of inheritance patterns, meaning only family members received the estate. an articulated inheritance that emphasized non-family members was relatively uncommon (only 3 out of 60 wills). it was referred to as disinheritance when a family member was not listed in a will. only one case demonstrated a trend of disinheritance (today, most state laws have protections for spouses in such cases, see levmore, 2020). the researchers concluded that the relationships between parents and children were based on mutual affection, understanding, and respect, even if testators favored spousal ties above lineal ones in all periods surveyed. bequest gift motive the studies investigating the bequest gift motive are numerous. key study themes include the relationship between giving a bequest to annuities (particularly the "annuity puzzle") and savings. anderson et al. 31 friedman and warshawsky (1990) concluded that bequest motives and yield differentials influence annuity purchase behaviors. hansen and i̇mrohoroğlu (2008) used the bequest motive to study the annuity puzzle and hump-shaped consumption patterns within the united states. finally, personal financial planning researchers williams and james (2019) demonstrated that bequest provisions, in addition to mortality salience, can drive annuity type selection. the relationship between savings and bequest motives has been explored by several researchers as well, especially within the context of the theoretical framework of the life-cycle hypothesis. hurd (1987) is an older example of a study that fails to find broad-reaching evidence for a bequest motive. hurd (2002) later observed that the marginal utility from bequests was much lower than that from consumption. davies (1981) argued that the unknown lifespan is a main driver of post-retirement expenditure rates while de nardi et al. (2009a; 2009b) concluded that uncertain longevity and high medical costs are more explanatory in elderly saving than bequests. in contrast, dynan et al. (2002) looked at saving in terms of contingencies and the bequest motive. additional perspectives can be found in browning and lusardi (1996) and spencer and fan (2002), who explored the motivations for bequests through the lenses of savings and debt. philanthropic bequests a charitable bequest occurs when someone leaves assets to charities or nonprofit organizations rather than to family, close friends, or other relatives. according to a review of 319 wills by schwartz (1993), most who make philanthropic bequests leave around five percent of their assets to charities and unrelated persons. rossi and rossi (1990), a study mentioned previously, also explored when charities and institutions were named as beneficiaries in respondents' wills. it was discovered that women were more likely than men to include non-kin beneficiaries—friends, charities, and institutions—in their wills (rossi & rossi, 1990, p. 475). more women than men named a friend as a beneficiary. key indicators for providing for an institution in a will included age, marital status, previous experience receiving a bequest, education, and religiosity with an unmarried status having the largest effect (rossi & rossi, 1990, p. 479). to explore charitable bequests, boskin (1976) created two separate models: the economic estate (i.e., the gross estate less debts and expenses) and the adjusted disposable estate (i.e., economic estate less the taxes paid where there is no charitable bequest). boskin showed that bequests to charities were significantly lower for decedents who passed away before the age of 65, had spouses or children, had small estates, were single, and lived in states with community property laws. the main finding of boskin's study was that the deductibility of inheritance taxes significantly impacted the number of charitable bequests. barthold and plotnick (1984) explored connecticut estates to look at the demographics of additional beneficiaries as well as the inheritance tax. decedents without surviving spouses or children were more likely to leave sizeable charitable bequests. the volume of philanthropic bequests was significantly and favorably influenced by religious choice. in contrast to previous studies, the magnitude of charitable bequests was not significantly correlated with inheritance tax rates. auten and joulifan (1996) developed a model of philanthropic bequests and gifts to consider intergenerational wealth transfers from parents to their offspring. the aim of the study was to determine the effects of children's income and bequest taxes on parents' charitable contributions. the main finding of this study was that the size of a charitable bequest was influenced by the “tax price” of the bequest due to estate tax rates. those who were older, married, and wealthy left more money to charity. further evidence that parents of financially better-off children made more lifetime charitable contributions than did parents of children with lower incomes can be seen in the insight that the children's wages greatly enhanced the amounts of charitable contributions. receiving bequests fewer studies have investigated receiving bequests. grawe (2010) used the panel study of income dynamics (psid) to study bequest financial services review, 32(1) 32 receipt, family size, and earnings. zagheni and wagner (2015) also used this dataset to explore the interplay between age and bequest receipt and found that the timing of a bequest’s receipt impacts wealth inequality. stark and nicinska (2015) used the survey on health, ageing and retirement in europe (share) to show that those who receive a bequest are more likely to expect to give one. conceptual model bequests have been studied using several theoretical frameworks and models including economic theory, the overlapping-generations model, and the economic theory of family size effect (brown et al., 2010; grawe, 2010; weil, 1996). this paper uses the same theory found in the study by kao et al. (1997): altruistic bequest theory. altruistic bequest theory holds that, in addition to their own consumption, parents' utility is influenced by the resources and wealth of their offspring. parents maximize their lifetime usefulness and feel contentment by enhancing their children's financial security through bequests (becker, 1974; becker and tomes, 1979; menchik and david, 1983; tomes, 1981). altruistic bequest theory was created chiefly by becker (1974), who used economic theory to analyze interactions between members of the same family. becker’s model had two basic concepts: "social environment" (i.e., the monetary value of other people's traits) and "social income" (i.e., the sum of an individual's income). due to the interconnectivity of the individual and family within this framework, family members behave altruistically and are motivated to maximize the wealth of the entire family rather than their own individual incomes. the family unit has one utility function. becker goes on to state that “the major, and somewhat unexpected, conclusion is that if a head exists, other members also are motivated to maximize family income and consumption, even if their welfare depends on their own consumption alone” (becker, 1974, p. 19). he concluded that parents commonly give their children proportionally larger bequests of wealth than the amount of their own income increase since transfers are thought to be responsive to changes in parental income (i.e., high-income elasticity). becker expanded the altruistic model of wealth transfer to include charitable transfers, which are motivated by the desire to improve the well-being of unrelated people. becker and tomes (1979) developed a cogent theory of intergenerational inequality and the intergenerational mobility of wealth on the premise that each family's goal is to maximize its utility across several generations. tomes (1981) empirically investigated the altruistic bequest model of intergenerational transmission of inequality proposed by becker (1974), blinder (1973), and ishikawa (1975). according to this study, parental investments in human capital and transfers of monetary wealth were undertaken as a form of altruism. because parents leave different amounts as compensation for economic inequality among the children, a larger share of wealth is passed down to low-income children than to children in higher income categories, according to tomes (1981), who saw this finding as supporting altruistic bequests. an empirical study on the quantity of bequests and the desire to leave bequests was conducted by menchik and david (1983). they concluded from tax return data in wisconsin between 1946-1964 that the quantity of the bequest and the parents' propensity to do so were closely connected to the parents' ages at death, likely due to a desire to bequest and/or risk aversion. those with higher incomes left greater bequests than by lower discounted lifetime earnings. the study also found that self-employed decedents left larger bequests than non-self-employed. in summary, altruistic bequest theory posits that parents gain utility from giving a bequest to their children, similar to the utility begotten from consumption. furthermore, parents will bequeath larger amounts to children of low earnings, in other words, bequests are compensatory (wilhelm, 1996). this paper argues that further insight can be gained into the dynamics of altruistic bequest theory through proxies or indicators of altruism, namely, charitable giving, the prioritization of saving for the benefit of children, and volunteerism. anderson et al. 33 data and sample as mentioned previously, this study updates kao et al.'s (1997) initial analysis using the 1989 scf to the 2019 wave. the scf is a cross-sectional survey sponsored by the federal reserve system, which employs a complex sampling technique (board of governors of the federal reserve system, 2022). the scf employs multiple imputations that include five times the actual observations (board of governors of the federal reserve system, 2022; hanna et al., 2018; montalto and sung, 1996). therefore, the repeated imputation inference (rii) method is used to estimate the correct standard errors (hanna et al., 2018; montalto and sung, 1996). the survey is conducted every three years and is meant to be representative of the united states population after applying the appropriate statistical weighting (board of governors of the federal reserve system, 2022). dependent variables the first dependent variable for this study captures those who expect to receive an inheritance, which is a dichotomous variable in the scf (“do you (or your husband/wife) expect to receive a substantial inheritance or transfer of assets in the future?”). label 1 signifies that the respondent expects an inheritance (n = 881) and label 5 if not (n = 4,895). table 1. cross tabulation of bequest gift motive and receipt expectation bequest receipt expectation bequest gift motive no yes total yes 78% 22% 100% no 86% 14% 100% maybe 93% 7% 100% total 85% 15% 100% the second dependent variable is for those expecting to leave a bequest coming from variable x5825 (“do you (and your {husband/wife/partner/spouse}) expect to leave a sizable estate to others?”). this is a three-level variable with yes (1), possibly (3), and no (5) as answers. there were approximately n = 2,522 households who expected to give a bequest. a cross-tabulation of these two dependent variables is shown in table 1. bequest gift motive independent variables economic measures. bequest gift motive economic measures include the log of household income (a continuous variable), the log of liquid and non-liquid asset holdings (continuous variables), if the respondent had received an inheritance (binary where 1 indicates having received an inheritance and 0 otherwise), and self-employment status (binary where 1 selfemployment and 0 otherwise). sociodemographic measures. bequest gift motive respondent sociodemographic measures include age (continuous), education (recoded as a categorical variable with levels of high school or less, some college, and bachelor’s degree or above), marital status (recoded as a categorical variable with married, partner relationship, single female, and single male), race (recoded as a categorical variable with white, black, hispanic, and other), number of children within the household (categorical with a range of 0 to 7 which includes kids of respondent and reference person), and if the respondent had a living parent (binary where 1 represents the presence of a living mother or father and 0 otherwise). attitudinal measures. bequest gift motive sociodemographic measures include attitude toward leaving a bequest, having made a charitable contribution, and risk aversion. attitude toward a bequest is a categorical variable with the answer to the following question: “some people think it is important to leave an estate or inheritance to their surviving heirs, while others don't. which is closer to your (and your husband/wife/partner/spouse's) feelings? would you say it is very important, important, somewhat important, or not important?” the answers represent the levels for this variable. the charitable contribution variable is dichotomous with 1 representing ever making a charitable contribution and 0 otherwise. finally, risk aversion is a categorical variable representing the answer to the following question: “some people are fully prepared to take financial risks when they save or make investments, while others try to avoid taking financial risks. on a scale from financial services review, 32(1) 34 zero to ten, where zero is not at all willing to take risks and ten is very willing to take risks, what number would you (and your husband/wife/partner) be on the scale?" to simplify the reporting of the results, this variable was recoded into low (up to 4), medium (5 to 7), and high (8 to 10) risk categories. health measures. bequest gift motive health measures include self-reported health and disability status and longevity expectations. the self-reported health status categorical variable stores the three categories from the following question: “would you say your (husband/wife/partner/spouse)'s health in general is excellent, good, fair, or poor?” the answers represent the levels in the variable. disability status is dichotomous with 1 representing a disabled job status and 0 otherwise. longevity expectation is a continuous variable representing how long the respondent expects to live, ranging from 40 to 150 years. altruistic measures. altruistic measures were added to the current study to test altruistic bequest theory. bequest gift motive altruistic measures include household volunteerism, saving for college as a savings priority, and helping kids as a savings priority. volunteerism is dichotomous with 1 representing someone in the household volunteering at least one hour or more a week and 0 otherwise. the second two variables measure savings attitudes. the scf asks the following question: “people have different reasons for saving, even though they may not be saving all the time. what are your most important reasons for saving?” the saving for college variable is dichotomous, with 1 representing the answer of “children's education; education of grandchildren” to this question and 0 otherwise. the helping kids variable is dichotomous, with 1 representing the desire to save "for the children/family," "to help the kids out," or "estate” and 0 otherwise. bequest receipt expectation independent variables economic measures. bequest receipt expectation economic measures include the log of household income, the log of liquid and nonliquid asset holdings, and if the respondent had received an inheritance. these variables were coded the same as previously mentioned for the bequest gift motive. sociodemographic measures. bequest receipt sociodemographic measures include age, education, marital status, race, number of siblings, and if the respondent was living with parents. age, education, marital status, race, and if the respondent had living parents were coded the same as previously mentioned for the bequest gift motive. the number of siblings variable was recoded to capture a range from 0 to 6. health measures. bequest receipt health measures include self-reported health and disability status. these variables were coded the same as previously mentioned for the bequest gift motive. table 2 shows a summary of the various independent variables for each part of the study. anderson et al. 35 table 2. independent variables for bequest gift motive and bequest receipt expectation bequest gift motive bequest receipt expectation group 1: economic measures group 1: economic measures a.1 household income b.1 household income a.2 liquid and non-liquid asset holdings b.2 liquid and non-liquid assets a.3 amount of inheritance ever received – changed to binary b.3 ever received inheritance a.4 self-employment status group 2: sociodemographic measures group 2: sociodemographic measures a.5 age b.4 age a.6 education b.5 education a.7 marital status b.6 marital status a.8 race b.7 race a.9 number of children b.8 number of children under 18 – dropped from the study a.10 respondent has living parents b.9 number of siblings b.10 respondent has living parents group 3: attitudinal measures group 3: health-related measures a.11 attitude toward leaving a bequest b.11 self-reported health a.12 ever having made a charitable contribution b.12 disability status a.13 extent of risk aversion group 4: health-related measures a.14 self-reported health a.15 disability status a.16 longevity expectation group 5: altruistic measures a.17 household volunteerism a.18 saving for college is a savings priority a.19 helping kids is a savings priority note. bolded variables are new or modified variables compared to kao et al. (1997). empirical model there is much debate regarding the use of weighted vs. unweighted regression modeling. according to shin and hannah (2017), using unweighted models can produce more conservative significance test results when the focus is on structural relationships. we followed this guidance and chose to study the bequest gift motive dependent-independent variable relationships using an unweighted multinomial logit model. three categories were compared: (a) bequest versus maybe bequest, (b) bequest versus no bequest, and (c) no bequest versus maybe bequest. the result of this analysis is found in financial services review, 32(1) 36 table 4. the receipt expectation dependentindependent variable relationships were studied using an unweighted binary logit model. the result of this analysis is found in table 5. results descriptive statistics the descriptive statistics for the variables in the current study are presented in table 3. unweighted results show 27% of the sample received an inheritance and 22% are selfemployed. the sample was highly educated with 73% having at least some college education and 48% having a bachelor’s degree or higher. married persons represented 54% of the sample. sample races included 72% white, 13% black, 10% hispanic, and 6% other. nearly 55% of the sample thought giving a bequest was “important” or “very important.” risk tolerance categories included 40% for low, 44% for medium, and 16% for high. this sample’s self-reported health was high with 28% reporting “excellent” and 50% reporting “good.” the mean for longevity expectation was 86 years old. regarding altruistic and additional variables for this study, 47% of the sample have given to charity at some point in the past, while 30% were volunteers. the college savings and child savings priorities represented a relatively minor portion of the sample at 4% and 5%. the mean age was 53 years old. the median income for the sample was $79,4131, with a median of $64,500 for financial assets and $252,600 for nonfinancial assets. (these figures are unweighted results across all five implicates; weighted results show the mean income for the sample was $106,251, with a mean of $358,116 for financial assets and $496,302 for nonfinancial assets). anderson et al. 37 table 3. descriptive statistics unweighted weighted mean/proportion mean/proportion received an inheritance 0.2707 0.2361 self-employed 0.2176 0.1111 education high school or less 0.2667 0.3156 some college 0.2557 0.2991 bachelor's and above 0.4777 0.3853 marital status married 0.5380 0.4607 partner relationship 0.0860 0.0997 single female 0.2164 0.2614 single male 0.1596 0.1782 race white 0.7167 0.6800 black 0.1301 0.1565 hispanic 0.0966 0.1093 other 0.0566 0.0542 number of children 0.7483 0.7224 number of siblings 2.4395 2.4714 living parents 0.3578 0.3802 attitude toward bequest very important 0.2811 0.2520 important 0.2679 0.2771 differ 0.0071 0.0064 somewhat important 0.2778 0.2882 not important 0.1661 0.1764 charitable giving 0.4730 0.3592 risk tolerance low 0.3945 0.4700 medium 0.4424 0.4179 high 0.1631 0.1121 health excellent 0.2787 0.2349 good 0.4913 0.5000 fair 0.1906 0.2174 poor 0.0394 0.0477 disabled 0.0590 0.0713 financial services review, 32(1) 38 longevity expectation 85.5674 85.0849 volunteerism 0.2941 0.2324 saving for college priority 0.0417 0.0424 saving for children priority 0.0542 0.0527 age 52.8039 51.2769 income 79413* 106251 financial assets 64500* 358116 nonfinancial assets 252600* 496302 n = 5,777 (28,885 across five implicates) *median reported due to skewedness bequest gift motive multinomial logit model for the first column in table 4, expecting to give a bequest versus maybe give a bequest, the multinomial logit model showed that those who received an inheritance had 1.30 times the odds of expecting to give a bequest versus maybe give a bequest compared to those who had not received an inheritance. self-employed individuals had 1.36 times the odds of expecting to give a bequest versus maybe give a bequest compared to those who were not self-employed. single females and single males had 1.27 and 1.31 times the odds, respectively, of expecting to give a bequest versus maybe give bequest compared to those who were married. blacks had 1.27 times the odds of expecting to give a bequest versus maybe give bequest compared to whites. attitudes toward giving a bequest had a strong directional relationship with expecting to leave a bequest (across all three categories studied). those with lower attitudes toward giving a bequest compared to those who thought giving a bequest was “very important” had lower odds of expecting to give a bequest versus maybe give a bequest. respondents who reported “disabled” as their employment status had 1.65 times the odds of expecting to give a bequest versus maybe give a bequest compared to those who did not list this as their work status. charitable givers had 1.39 times the odds of expecting to give a bequest versus maybe give a bequest compared to those who had not given to charity. log of income, log of financial assets, and log of nonfinancial assets were positively associated with expecting to give a bequest versus maybe give a bequest. longevity expectation, volunteerism, and the savings priority variables were not significant in the first column of results. the focus of this paper rests on the second column of the multinomial logit model, namely, giving a bequest versus not giving a bequest. this column is likely the most important for interpreting the results given the study’s research question. the multinomial logit model showed that those who received an inheritance had 1.84 times the odds of expecting to give a bequest versus no bequest compared to those who had not received an inheritance. this is in line with a recent study by deboer and hoang (2017). the self-employed had 1.47 times the odds of expecting to give a bequest versus no bequest compared to those who were not self-employed. single males had 1.32 times the odds of expecting to give a bequest versus no bequest compared to those who were married. blacks had 1.80 times the odds and hispanics 1.81 times the odds of expecting to give a bequest versus no bequest compared to whites. those with lower attitudes toward giving a bequest compared to those who thought giving a bequest was “very important” had lower odds of expecting to give a bequest versus no bequest. charitable givers had 1.62 times the odds of expecting to give a bequest versus no bequest compared to those who had not given to charity. risk tolerance also played a role in bequest expectations; those ranked at medium risk tolerance and high risk tolerance had 1.31 and 1.92 times the odds of giving a bequest versus no bequest compared to those with low risk tolerance. those reporting poor health had 0.47 times the odds of giving a bequest versus no bequest compared to those with excellent health. anderson et al. 39 log of income, log of financial assets, and log of nonfinancial assets were positively associated with expecting to give a bequest versus no bequest. longevity expectation and volunteerism were not significant in the second column of results, but the two savings priority variables were. respondents with the saving for college priority had 1.93 times the odds of expecting to give a bequest versus no bequest compared to those who did not have this saving priority. respondents with the saving for children priority had 1.84 times the odds of expecting to give a bequest versus no bequest compared to those who did not have this saving priority. for the final column in table 4, no bequest versus maybe bequest, the model showed that those who received an inheritance had 0.71 times the odds of expecting to not give a bequest versus maybe give a bequest compared to those who had not received an inheritance. those who held the two savings priorities exhibited lower odds of expecting to not give a bequest versus maybe give a bequest compared to those who did not have these savings priorities. unlike the previous two categories, charitable giving was not statistically significant for the no bequest versus maybe bequest category. longevity expectation had a significant relationship with an odds ratio close to 1 (of 0.99) for the no bequest versus maybe bequest category. finally, the log of financial assets and the log of nonfinancial assets had a negative relationship with the odds of being in the no bequest category rather than the maybe bequest category. financial services review, 32(1) 40 table 4. multinomial logit for bequest gift motive modeled response v. base bequest v. maybe bequest bequest v. no bequest no bequest v. maybe bequest est. coef se odds ratio est. coef se odds ratio est. coef se odds ratio received an inheritance 0.2608 ** 0.0884 1.2980 0.6100 *** 0.0965 1.8404 -0.3492 *** 0.1001 0.7053 self-employed 0.3055 ** 0.0979 1.3573 0.3852 *** 0.1093 1.4699 -0.0797 0.1153 0.9234 education (ref= hs or less) some college -0.2471 * 0.1056 0.7811 -0.0701 0.1083 0.9323 -0.1769 0.1041 0.8378 bachelor's and above -0.1069 0.1078 0.8987 -0.1077 0.1110 0.8979 0.0008 0.1106 1.0008 marital status (ref=married) 1.0000 partner relationship 0.0939 0.1372 1.0985 0.1087 0.1468 1.1148 -0.0147 0.1417 0.9854 single female 0.2389 * 0.1097 1.2698 0.1210 0.1108 1.1286 0.1178 0.1089 1.1251 single male 0.2699 * 0.1173 1.3098 0.2769 * 0.1215 1.3191 -0.0070 0.1215 0.9930 race (ref=white) 1.0000 black 0.2420 * 0.1223 1.2738 0.5892 *** 0.1287 1.8026 -0.3472 ** 0.1255 0.7067 hispanic 0.1131 0.1301 1.1197 0.5948 *** 0.1401 1.8127 -0.4817 *** 0.1337 0.6177 other -0.0803 0.1539 0.9228 0.2628 0.1763 1.3005 -0.3431 * 0.1722 0.7096 number of children -0.0121 0.0368 0.9879 -0.0711 0.0398 0.9314 0.0590 0.0396 1.0608 living parents 0.0551 0.0978 1.0567 0.0771 0.1054 1.0801 -0.0219 0.1058 0.9783 attitude toward bequest (ref= very important) important -0.7003 *** 0.0952 0.4964 -1.0548 *** 0.1141 0.3483 0.3545 ** 0.1222 1.4254 differ -1.0165 ** 0.3799 0.3618 -1.2927 0.4649 0.2745 0.2762 0.4747 1.3181 somewhat important -1.3801 *** 0.0987 0.2516 -2.2118 *** 0.1147 0.1095 0.8317 *** 0.1177 2.2973 not important -1.4065 *** 0.1377 0.2450 -3.3186 *** 0.1376 0.0362 1.9122 *** 0.1369 6.7676 charitable giving 0.3285 *** 0.0917 1.3888 0.4815 *** 0.0974 1.6185 -0.1531 0.0987 0.8581 anderson et al. 41 risk tolerance (ref=low) medium 0.1618 0.0836 1.1756 0.2720 ** 0.0867 1.3126 -0.1102 0.0852 0.8957 high 0.4074 *** 0.1155 1.5029 0.6516 *** 0.1274 1.9187 -0.2442 0.1327 0.7833 health (ref=excellent) good -0.0637 0.0872 0.9383 -0.1428 0.0968 0.8669 0.0791 0.0992 1.0823 fair -0.0992 0.1213 0.9056 -0.1735 0.1275 0.8407 0.0743 0.1264 1.0771 poor -0.1393 0.2679 0.8699 -0.7575 ** 0.2398 0.4688 0.6181 ** 0.2373 1.8555 disabled 0.5021 * 0.2042 1.6522 0.1121 0.1797 1.1186 0.3900 * 0.1801 1.4770 longevity expectation -0.0024 0.0039 0.9976 0.0061 0.0039 1.0061 -0.0085 * 0.0039 0.9916 volunteerism 0.1134 0.0859 1.1201 0.0507 0.0927 1.0520 0.0627 0.0956 1.0648 saving for college priority 0.1406 0.1708 1.1510 0.6590 ** 0.2064 1.9328 -0.5184 * 0.2047 0.5955 saving for children priority 0.2081 0.1635 1.2313 0.6084 ** 0.1866 1.8375 -0.4003 * 0.1896 0.6701 age -0.0029 0.0035 0.9971 -0.0280 *** 0.0037 0.9724 0.0250 *** 0.0036 1.0254 log of income 0.1190 *** 0.0276 1.1263 0.0919 ** 0.0325 1.0962 0.0271 0.0299 1.0275 log of financial assets 0.1095 *** 0.0195 1.1158 0.2521 *** 0.0197 1.2867 -0.1426 *** 0.0182 0.8671 log of nonfinancial assets 0.0291 * 0.0146 1.0295 0.0753 *** 0.0144 1.0782 -0.0463 *** 0.0130 0.9548 constant -1.8025 *** 0.4710 0.1649 -2.5855 *** 0.5060 0.0754 0.7830 0.4974 2.1880 model fit statistics log-likelihood 4823.2696 mcfadden pseudo r2 0.2182 *p < .05 ** p < .01 ***p<0.001 financial services review, 32(1) 42 bequest receipt expectation binary logit model for the binary logit model (table 5), those who received an inheritance had 1.85 times the odds of expecting to receive a bequest compared to those who had not received an inheritance. those with more education had higher odds of expecting a bequest compared to those with less education. black, hispanic, and other races had lower odds compared to whites to expect a bequest. those with living parents had 1.54 times the odds of expecting a bequest compared to those who did not have living parents. although the scf does not explicitly link the parent-tochild bequest type, this finding makes sense as this bequest would require living parents to execute. age had a negative relationship with bequest receipt expectation. in other words, older individuals were less likely to expect a bequest from their benefactors. the log of income had a negative relationship with bequest receipt expectation, whereas the log of financial assets had a positive relationship. discussion this paper’s discussion begins with the new variables added to the bequest gift motive multinomial logit model. first, household volunteerism was not associated with the bequest gift motive across any of the categories compared to the base category of those who do not volunteer. it is possible that this can be explained because volunteers consider their time and efforts as equivalent to a monetary bequest. second, although longevity expectation was associated with the bequest gift motive for the no bequest versus maybe bequest category, the relationship was quite weak. this result suggests that longevity expectation is not a strong predictor of bequest intention. third, the two savings priorities were associated with bequest intention in two categories (i.e., bequest vs. no bequest and no bequest vs. maybe bequest), which supports the idea that parents who prioritize the well-being of their children are more likely to leave bequests. this finding reinforces the underlying concepts of altruistic bequest theory. respective to the base categories of receiving an inheritance, selfemployed, single (male) marital status, race (black and hispanic), positive attitude toward leaving a bequest, charitable giving, higher risk tolerance, age, income, financial assets, and nonfinancial assets were more likely to predict giving a bequest versus no bequest. outside of the poor health status, health was not a key predictor in the model. this finding deserves further study to understand the dynamics underlying this relationship. regarding the bequest receipt expectation binomial logit model, receiving an inheritance, level of education, marital status, race, presence of living parents, age, income, and financial assets were the most important predictors of receiving a bequest. surprisingly, the log of income showed a negative relationship with the expectation of receiving a bequest. perhaps this means children or relatives with lower incomes are more likely to expect a bequest from their benefactors. limitations and future research the current study curated variables from the scf to test altruistic bequest theory. although large secondary datasets have great value in answering research questions, a researcher cannot go back to ask more precise questions of the respondents. as such, a few of the variables used to answer the study’s research questions have inherent limitations. first, the volunteerism variable extracted from the scf measures if someone in the household volunteers at least one hour or more a week. however, a direct link between the volunteering individual (or individuals) and the bequest receipt or bequest gift expectation is unknown. also, the charitable giving dichotomous variable gives equal weight to someone who, hypothetically, gave one small gift ten years ago versus someone who regularly gives to multiple charities. furthermore, the scf often does not go beyond an expectation to act to measure the action itself. the savings priority variables used in this study, for example, simply measure a desire to prioritize saving for children, grandchildren, or college but do not measure if this occurred. researchers in future studies could attempt to gather primary data that have less ambiguity in measuring altruistic bequest theory, ones that better define relationships and test behavior. anderson et al. 43 this investigation encourages further study in two areas. first, it points to the need for continued study of the relationship between the bequest gift motive and health, as noted previously. second, a similar study could be enhanced by studying variables across waves of the scf to show if the relationships uncovered in this study change over time. table 5. binary logit model for bequest receipt expectation coeff se odds ratio received an inheritance 0.6152 *** 0.0895 1.8519 self employed 0.1535 0.0972 1.1688 education some college 0.2689 * 0.1282 1.3112 bachelor's and above 0.5110 *** 0.1252 1.6683 marital status partner relationship -0.0175 0.1450 0.9780 single female -0.5659 *** 0.1330 0.5670 single male -0.1067 0.1213 0.8957 race black -0.8287 *** 0.1661 0.4387 hispanic -1.3017 *** 0.2104 0.2721 other -0.4919 ** 0.1754 0.6090 number of siblings -0.0271 0.0254 0.9729 living parents 0.4291 *** 0.1035 1.5386 health good 0.0364 0.0882 1.0373 fair -0.1865 0.1350 0.8276 poor -0.1953 0.2922 0.8227 disabled 0.3551 0.2188 1.4309 age -0.0369 *** 0.0038 0.9637 log of income -0.0636 * 0.0254 0.9382 log of financial assets 0.0960 *** 0.0213 1.1003 log of nonfinancial assets -0.0052 0.0156 0.9946 constant -0.5388 0.3431 0.5871 model fit statistics log-likelihood -2183.6854 mcfadden pseudo r2 0.1149 * significant at p < 0.05; ** significant at p < 0.01; *** significant at p < 0.001 conclusion this paper updated bequest research conducted by kao et al. (1997). this original study used the 1989 scf to study the two sides of bequests: the bequest gift motive and the receipt expectation. many of the original economic, sociodemographic, attitudinal, and health variables found in that study are further investigated here using the 2019 scf. further, this paper added additional variables found within the scf that align with altruistic bequest financial services review, 32(1) 44 theory. results from the addition of these altruistic variables offer support for the continued use of altruistic bequest theory in the study of bequest intention. finally, this paper explored the relationship between longevity expectation and bequest motive, showing that longevity expectation is not a strong predictor of bequest intention. references auten, g., & joulifan, d. (1996). charitable contributions and intergenerational transfers. journal of public economics, 59(1), 55-68. https://doi.org/10.1016/00472727(94)01475-2 barthold, t., & plotnick, r. (1984). estate taxation and other determinants of charitable bequests. national tax journal, 37(2), 225-237. https://doi.org/10.1086/ntj41791948 becker, g. (1974). a theory of social interactions. journal of political economy, 82(6), 1063-1093. https://doi.org/10.1086/260265 becker, g. s., & tomes, n. (1979). an equilibrium theory of the distribution of income and intergenerational mobility. journal of public economy, 87(6), 11531189. https://doi.org/10.1086/260831 blinder, a. s. (1973). a model of inherited wealth. the quarterly journal of economics, 87(4), 608-626. https://doi.org/10.2307/1882027 board of governors of the federal reserve system. (2022). survey of consumer finances (scf). retrieved february 16, 2023, from https://www.federalreserve.gov/econres/ scfindex.htm. boskin, m. j. (1976). estate taxation and charitable bequests. journal of public economics, 5(1-2), 27-56. https://doi.org/10.1016/00472727(76)90059-1 brown, j. r., coile, c., & weisbenner, s. j. (2010). the effect of inheritance receipt on retirement. the review of economics and statistics, 92(2), 425–434. doi: https://doi.org/10.1162/rest.2010.11182 browning, m., & lusardi, a. (1996). household saving: micro theories and micro facts. journal of economic literature, 34(4), 1797–1855. http://www.jstor.org/stable/2729595 coleman, m., & ganong, l. h. (1998). attitude toward inheritance following divorce and remarriage. journal of family and economic issues, 19, 289-314. https://doi.org/10.1023/a:10229737317 65 davies, j. b. (1981). uncertain lifetime, consumption, and dissaving in retirement. the journal of political economy, 89(3), 561-577. https://doi.org/10.1086/260986 deboer, d. r., & hoang, e. c. (2017). inheritances and bequest planning: evidence from the survey of consumer finances. journal of family economic issues. 38(1), 45-56. https://doi.org/10.1007/s10834-0169509-0 de nardi, m., french, e., & & jones, j. b. (2009a). why do the elderly save? the role of medical expenses. the journal of political economy, 118(1), 39-75. https://doi.org/10.1086/651674 de nardi, m., french, e., & jones, j. b. (2009b). life expectancy and old age savings. washington, d.c.: national bureau of economic research. https://www.aeaweb.org/articles?id=10. 1257/aer.99.2.110 dynan, k. e., skinner, j., & zeldes, s. p. (2002). the importance of bequests and lifecycle saving in capital accumulation: a new answer. the american economic review, 92(2), 274-278. doi: 10.1257/000282802320189393 engler-bowles, c. a., & kart, c. s. (1983). intergenerational relations and testamentary patterns: an exploration. anderson et al. 45 the gerontologist, 23(2), 167-173. https://doi.org/10.1093/geront/23.2.167 friedman, b.m., & warshawsky, m.j. (1990). the cost of annuities: implications for saving behavior and bequests. the quarterly journal of economics, 105(1), 135-154. https://doi.org/10.2307/2937822 grable, j. e. (2013). gender, wealth, and risk: why are baby boomer women less risk tolerant than baby boomer men? journal of financial services professionals, 67(3), 7-9. grawe, n. d. (2010). bequest receipt and family size effects. economic inquiry, 48(1), 156-162. https://doi.org/10.1111/j.14657295.2008.00208.x hanna, s. d., kim, k. t., & lindamood, s. (2018). behind the numbers: understanding the survey of consumer finances. journal of financial counseling and planning, 29(2), 410– 418. https://doi.org/10.1891/10523073.29.2.410 hansen, g.d., & i̇mrohoroğlu, s. (2008). consumption over the life cycle: the role of annuities. review of economic dynamics, 11(3), 566-583. https://doi.org/10.1016/j.red.2007.12.00 4 hurd, m. d. (1987). savings of the elderly and desired bequests. the american economic review, 77(3), 298-312. https://www.jstor.org/stable/1804096 hurd, m. (2002). are bequests accidental or desired? (no. dru-3010). rand corporation. https://ideas.repec.org/p/ran/wpaper/dru3010.html ishikawa, t. (1975). family structures and family values in the theory of income distribution. journal of political economy, 83(5), 987-1008. https://doi.org/10.1086/260373 kao, y. e., hong, g. s., widdows, r. (1997). bequest expectations: evidence from the 1989 survey of consumer finances. journal of family and economic issues, 18(4), 357-377. https://doi.org/10.1023/a:10249434210 55 levmore, s. (2020). death as divorce for the abandoned spouse: davis v combes and the cautious and gender-sensitive judiciary. the university of chicago law review, suppl. special issue, 87, 24212435. menchik, p. l., & david, m. (1983). income distribution, lifetime saving, and bequests. the american economic review, 73(4), 672690.https://www.jstor.org/stable/181656 6 metlife (2010). metlife mature market institutes study estimates boomer’s inheritance at $8.4 trillion. https://www.metlife.com/aboutus/newsroom/2010/december/metlifemature-market-institute-study-estimatesboomers-inherita/ montalto, c. p. & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances. journal of financial counseling and planning, 7, 143–146. rossi, a.s. & rossi, p. (1990). help beyond death: legators and beneficiaries of insurance policies and written wills. in of human bonding: parent-child relations across the life course (pp. 461– 482). aline de gruyter schwartz, t. p. (1993). testamentary behavior: issues and evidence about individuality, altruism, and social influences. the sociological quarterly, 34(2), 337-355. https://doi.org/10.1111/j.15338525.1993.tb00394.x spencer, h. l., & fan, j. x. (2002). savers, debtors, and simultaneous debtors and savers. journal of financial counseling and planning, 13(2), 25–39. https://www.proquest.com/openview/41 financial services review, 32(1) 46 ef59ff667d82fc58495165f9552ca8/1?pq -origsite=gscholar&cbl=38873 shin, s. h., & hanna, s. d. (2017). accounting for complex sample designs in analyses of the survey of consumer finances. journal of consumer affairs, 51(2), 433–447. https://doi.org/10.1111/joca.12106 stark, o., & nicinska, a. (2015). how inheriting affects bequest plans. economica, 82(1), 1126-1152. https://doi.org/10.1111/ecca.12164 tomes, n. (1981). the family, inheritance, and the intergenerational transmission of inequality. journal of political economy, 89(5), 928-958. https://doi.org/10.1086/261014 wilhelm, m. o. (1996). bequest behavior and the effect of heir’s earnings: testing the altruistic model of bequests. the american economic review, 86(4), 874892. https://www.jstor.org/stable/2118309 weil, d.n. (1996). intergenerational transfers, aging, and uncertainty. in wise, d.a. (ed.), advances in the economics of aging (pp. 321-342). university of chicago press. williams, j. a., & james, r. n. (2019). bequest provision preferences in commercial annuities: an experimental test of the role of mortality salience. journal of financial counseling and planning, 30(1), 121–131. http://dx.doi.org/10.1891/10523073.30.1.121 zagheni, e., & wagner, b. (2015). the impact of demographic change on intergenerational transfers via bequests. demographic research, 33(1), 525–534. http://www.jstor.org/stable/26331995 a portfolio of leveraged exchange traded funds william j. trainor, jr.a,*, indudeep chhachhib, christopher l. brownb aeast tennessee state university, department of economics and finance, box 70686, johnson city, tn 37614, usa bwestern kentucky university, department of finance, 1906 college heights boulevard, bowling green, ky 42101-1061, usa abstract this study demonstrates how a portfolio of leveraged exchange traded funds (letfs) targeting a unit exposure to their underlying indexes outperforms a portfolio using traditional etfs while simultaneously reducing downside risk. by extension, a 3x letf portfolio designed to mimic 2x letfs outperforms the underlying 2x letf portfolio. the results are primarily a function of letfs borrowing short while the investor lends the additional wealth generated from this leverage in oneto seven-year treasury bonds or similar type of assets. for every one percent earned above the implied borrowing rate, a portfolio of 2x and 3x letfs outperforms a traditional portfolio by 0.41% and 0.63%, respectively, corresponding roughly to the additional return on the 50% and 67% of the wealth invested in bonds. more than 90% of letfs outperformance is explained by the borrowing lending differential. © 2020 academy of financial services. all rights reserved. jel classification: g11; g17 keywords: diversified portfolios; leveraged exchange traded funds 1. introduction leveraged exchange traded funds (letfs) were first listed in 2006 by proshares, although leveraged mutual funds have been around since 1993. while proshares introduced �2x products, direxion upped the leverage ante with �3x funds in late 2008. because letfs are designed to return a daily multiple, the constant daily leverage results in uncertain realized * corresponding author. tel.: �1-423-439-5668; fax: �1-423-439-8381. e-mail address: trainor@etsu.edu (w.j. trainor) financial services review 28 (2020) 35-48 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. leverage over longer periods of time. in general, realized leverage tends to fall over time because of the volatility of returns (avenllaneda and zhang, 2010; carver, 2009; cheng and madhavan, 2009; trainor and baryla, 2008). historically, investors have used margin to create leverage in their investments. however, with letfs expansion into everything from equity indexes to oil, gold, currencies, and treasuries, it is now possible to create a diversified portfolio of letfs that mimic a typical investor’s portfolio. this can be done by creating unit exposure to the underlying asset classes freeing up wealth to enhance returns. specifically, 2x and 3x letfs only require 50% and 33% of investor’s wealth to create the same exposure as investing 100% of an investor’s wealth in the underlying etfs or mutual funds. the advantages of using letfs, instead of margin, are leverage can be theoretically increased to 3x; no explicit interest costs, no possibility of margin calls, and unlike margin, letfs can be used in most retirement accounts. however, letf’s leverage is not free as there are implicit lending costs. in addition, letfs have higher expense ratios and general leverage decay. to the extent an investor seeks to maintain a specific asset allocation, frequent rebalancing is required along with the associated trading costs and taxes on realized gains. a letf portfolio structured for unit exposure will outperform its traditional counterpart if and only if the return to the invested excess wealth exceeds the implicit financing costs and higher costs of using letfs. in contrast to previous research, this study examines how a portfolio composed almost entirely of letfs can be created to maintain unit exposure to the underlying indexes without increasing the overall risk. george and trainor (2017) use similar methodology, but where they use a single letf to demonstrate how it can be used within a portfolio insurance setting, this study uses a portfolio of letfs to maintain an asset allocation comparable with traditional buy-and-hold portfolios created from mutual funds or etfs. in addition, for risk-seeking investors looking to invest in a diversified portfolio with up to 2x leverage, this study examines if 3x letfs may be used to the same effect. investing 67% in a 3x letf is equivalent to 100% in a 2x letf. if the return to the remaining 33% of the investor’s wealth exceeds the additional borrowing cost, the 3x letf will outperform. a critical question for the 2x or 3x strategy suggested above is where to invest the freed-up wealth. this study investigates leveraging bonds instead of equity as this will maintain the risk and structural characteristics of the underlying unleveraged portfolio. because letfs typically borrow short to attain their exposure, the portfolios investigated in this study are in effect borrowing short to lend long. this strategy should moderately increase expected return and reduce downside risk as most of the portfolio is invested in less risky bonds. results suggest a portfolio composed of 2x or 3x letfs outperforms a portfolio using standard etfs based on the same underlying indexes. even in the low interest environment from 2010 to 17 and using an aggregate bond index for the remaining wealth, a portfolio of 2x or 3x letfs outperforms a portfolio of standard etfs by 0.9% and 1.8%, respectively, on an annual basis. for risk seekers, using 3x letfs to mimic the exposure of 2x letfs outperforms by 1.5% annually during this period. the critical input is the borrowing-lending differential that explains 90% or more of the letf portfolio’s greater return. the results are essentially confirmed with simulated letf returns since 1946. 36 w.j. trainor et al. / financial services review 28 (2020) 35-48 the caveat to this type of strategy is letf portfolios must be rebalanced more often as their initial positions deviate from “optimal” asset allocations even faster than standard portfolios. a 10% barrier threshold defined as a relative deviation of 10% from the initial allocation is used to determine when portfolios are rebalanced. this active management keeps the risk exposure within reasonable bounds while keeping 2x and 3x rebalancing requirements to approximately quarterly and monthly, respectively. because of tax effects, the strategies outlined above are more suited to qualified nontaxable accounts. 2. mathematics of letfs letfs magnify the daily return of an underlying index. because of this constant daily leverage, the realized leverage over any multiday period can be virtually anything and is a function of the daily leverage ratio, time, return, and variance with the latter generally having the largest effect. realized leverage can mathematically be expressed by: lrt xrt � (1 � xrt)l exp� �l � l2� �2t 2 � � 1 xrt (1) where lrt is the return to the leveraged fund, xrt is the underlying index return, l is the daily leverage ratio, t is time in days, and �2 is the standard daily population variance, (avellaneda and zhang, 2010; cheng and madhaven, 2009). on average, the variance dominates and realized leverage over time tends to decline. this effect is greater with higher leverage since a daily leverage ratio of 2x multiplies the term (1 � xrt)l by exp(-�2t) but a daily leverage ratio of 3x multiplies this term by exp(-3�2t).1 equation (1) is related to volatility drag, which is the difference between geometric and arithmetic average returns and is a major hindrance to letf returns over extended periods of time. the relationship between the geometric and arithmetic return of any asset is written as: xpt � xrt � 0.5 �t 2 (2) where xpt is the geometric return, xrt is the arithmetic return, and �t 2 is the variance of returns. as an example, assume a daily return and standard deviation of 0.045% and 1.0%, respectively, which roughly corresponds to the s&p 500 market averages from 1946 to 2017. a 3x letf multiplies these numbers by 3. thus, the geometric return over a year for the underlying index assuming 252 trading days is 252*0.045% � 0.5*2.52% � 10.08%, and for a 3x this return is 252*0.135% � 0.5*22.68% � 22.68%.2 for long-term holdings of letfs, this volatility drag is a major drawback and clearly shows why leveraged funds usually do not return the daily multiple of the underlying index over time. expanding on the work of scott and watsun (2013), ott and zimmer (2016) show the return of an investment with leverage l is: 37w.j. trainor et al. / financial services review 28 (2020) 35-48 lrt � xrt � (xrt � rb) (l � 1) � 1⁄2 (�2 l2) (3) where lrt is the leverage return, xrt is the underlying index return, rb is the borrowing rate, l is the leverage ratio, and �2 is the variance of the underlying index. to reduce volatility drag relative to the underlying index, this study suggests using only a portion of the investor’s portfolio to invest in the letf. this percentage is set so the effective exposure to the index is the same as if the investor invested only in the underlying index. in effect, the volatility of the position in a letf is no greater than investing directly in the underlying index. this results in additional wealth available to offset the implied borrowing costs and higher expense ratio of the letfs. in this way, equation (3) can be modified to account for the return of a portfolio of etfs or a portfolio of letfs. to simplify, assume a portfolio is composed of just one underlying index. accounting for the letf’s higher expense ratio along with the additional wealth created from using letfs, equation (3) becomes the following: lpt � 1 l [xrt � (xrt � rb)(l � 1) � rexp] � 1 2 �2 � �1 � 1 l� rf (4) where lpt is the return to the portfolio using the letf, rexp is the letf expense ratio, and rf is the return to the risk-free asset.3 if this asset is not risk-free, there is volatility drag to this asset’s return which could easily be adjusted for in equation (4). the returns are multiplied by 1/l because only a portion of the portfolio relative to the underlying index is invested in the letf. when l is 1, signifying no leverage, equation (4) basically reduces to equation (2). with leverage, the letf return is multiplied by the inverse of the leverage. for example, with l equal to 3, the investor attains exactly xrt assuming no borrowing costs or expenses. this result attains since one third of an investor’s wealth in a 3x is basically the same as if they had invested 100% in the underlying index. with borrowing costs and higher expenses, they are not equivalent. however, the difference is offset by the return in a risk-free or some other alternative asset represented by rf. volatility drag for wealth in the letf or in the underlying index is the same as both have equal variance. the volatility drag will only become an issue if the percentage in the underlying letf varies to the point where the effective exposure of the letf portfolio diverges from a portfolio using the underlying index. with constant rebalancing, this discrepancy can be eliminated or at least effectively managed. constant or daily rebalancing is infeasible for most investors because of transaction costs but can be mitigated by periodic rebalancing to keep the 1/l ratio relatively constant over time. lu, wang, and zhang (2012) find an investor can assume a 2x/-2x letf will maintain its leverage ratio for holding periods up to one month. thus, daily rebalancing is likely not needed. to calculate the return from using a letf relative to investing in the underlying index, equation (2) is subtracted from equation (4) to attain the following: lpt � xpt � (1 � 1 l )rf � 1/l[(rb)(l � 1) � rexp] (5) 38 w.j. trainor et al. / financial services review 28 (2020) 35-48 equation (5) can be extended to account for multiple underlying asset classes that may be used by an investor. similarly, one can compare a 3x letf with a 2x letf. however, the basic implication of equation (5) remains the same. if the excess wealth from using letfs to create a portfolio (the first term in equation [5]) is greater than the borrowing costs and higher expense ratio of the letfs, then the return from using letfs will be greater than the return from using the underlying indexes. for the comparison of two letfs, the expense ratios will be approximately the same which reduces equation (5) to whether the additional wealth gained from the higher leveraged letf earns a return that exceeds its additional financing costs. this study determines whether this is the case. 3. data and methodology because most letfs were only recently introduced to the market, empirical research is limited. however, there are now 2x and 3x letfs that cover the main asset categories found in a typical diversified portfolio including small, mid, and large cap equity funds, international funds, short and long-term bond funds, reits, and a variety of commodity funds including gold, oil, and currencies for those using less traditional portfolios. to compare portfolio results using tradeable etfs, table 1 shows a portfolio of etfs and their 2x and 3x counterparts along with the effective percentages in each asset. the percentages are based on an investor who has 50% in domestic equity, 20% in international equites, five percent in a reit, and 25% in bonds. the domestic equity is split between 30% in the s&p 500 with 10% each in mid and small caps. the international equity is split with 10% in developed and 10% in emerging markets. the 25% in bonds is further delineated by five percent in 20� year t-bonds, five percent in seven to 10-year t-bonds, and the remaining 15% in an aggregate bond portfolio. a similar mix is used by considine (2006) comparing portfolios of etfs to mutual funds. the exact percentages are not critical to the results but are created to represent what a typical investor might have. the portfolio described above can be also be created using letfs. for an investor using 2x letfs, only half as much wealth is needed in each letf to obtain the same amount of table 1 portfolio composition using standard etfs and letfs asset class etf 2x letf 3x letf s&p 500 spy, 30% sso, 15% upro, 10% mid cap ijh, 10% mvv, 5% midu, 3.33% small cap iwm, 10% uwm, 5% tna, 3.33% international developed efa, 10% efo, 5% eurl, 3.33% emerging markets eem, 10% eet, 5% edc, 3.33% real estate iyr, 5% ure, 2.5% drn, 1.67% 20-year t-bonds tlt, 5% ubt, 2.5% tmf, 1.67% 7 to 10-year t-bonds ief, 5% ust, 2.5% tyd, 1.67% aggregate bonds bnd, 15% bnd, 57.5% bnd, 71.67% note: the base percentage in each asset class is set by the initial percentages in the etfs. ticker symbols used to calculate results are shown with each percentage. 39w.j. trainor et al. / financial services review 28 (2020) 35-48 exposure using the underlying etfs. for 3x letfs, only a third of the wealth is needed. thus, to attain 30% exposure to the s&p 500, an investor needs 15% in a 2x s&p 500 letf or 10% in a 3x s&p 500 letf. all the asset classes, except for aggregate bonds, have a corresponding letf. the aggregate bond asset class is used as the alternative for the excess wealth available when using letfs to build a portfolio. this results in the 2x and 3x letf portfolios investing 57.5% and 71.67% of the portfolio, respectively, in a relatively safe bond portfolio. in absolute terms, 80% of the etf portfolio is at moderate to high risk based on volatility of the underlying assets whereas only 40% and 27% of the 2x and 3x portfolios, respectively, are exposed to equity markets. for comparison of the 3x letf portfolio to the 2x letf portfolio, it is assumed the percentage in the underlying 2x letfs are the same as the underlying index portfolio, that is, 30% is invested in the 2x s&p, and so forth. for the 3x to attain the same exposure as the 2x, 20% is invested in the 3x s&p and so forth. with limited historical letf data, additional theoretical letf returns are calculated for the period before their inception. because letfs attain their exposure using a variety of derivative assets including swaps, there are embedded financing costs increasing with leverage (charupat and miu, 2014). based on the methodology of scott and watsun (2013), letf returns are calculated based on data going back to 1946. this shows how letf portfolios are likely to perform in a variety of market environments including the very high interest rate period during the early 1980s. the equation to calculate daily returns using the s&p 500 letf as an example is: rl � l*rs&p � rexp � (l � 1)*rb (6) where rl is the daily return to the letf with a daily leverage ratio of l, rs&p is the daily return of the s&p, rexp is the daily expense ratio, and rb is the borrowing rate using the 90-day t-bill rate as a proxy. strictly speaking, the one-week/month libor rate should be used, but libor data begins in 1986 and to remain consistent with sampled returns before this date, the 90-day t-bill rate is used. the 90-day t-bill has a 98% correlation with libor and averages 0.2% less than libor. thus, the borrowing rate is set at the 90-day t-bill yield �0.2%. the logic behind equation (6) is a 2x letf increases exposure by borrowing $1 for every $1 invested. a 3x letf borrows $2 for every $1 invested. to determine the validity of equation (6), theoretical daily, monthly, and annual returns are compared with the actual daily, monthly, and annual returns for the letfs listed in table 1. all return data are from the center for research in security prices (crsp). monthly return differences as measured by simulated returns minus the letf returns average 0.01% assuming an additional 1.2% annual expense ratio. although letf’s average expense ratio is approximately one percent, there are embedded costs associated with derivatives not accounted for. using a 1.2% expense ratio reduces the average differences for daily, monthly, and annual returns to near zero for the eight asset classes. monthly differences ranged from �0.07% for the 2x ubt (20� year treasury) to 0.17% for 3x emerging markets eet. thus, equation (6) appears to approximate returns accurately enough to simulate letf returns from index data predating letf’s introduction. 40 w.j. trainor et al. / financial services review 28 (2020) 35-48 for historical index data, asset classes are defined as crsp’s s&p 500 index, the 2–4 value weighted deciles proxy for a small-cap fund, and 5–7 value weighted deciles proxy for a midcap fund. the 20-year t-bond and an average of seven to 10-year t-bonds proxy for two additional bond funds. the remainder of a letf’s portfolio is invested equally in one, two, five, and seven-year treasury bonds. no reliable daily international data are available before 1991 so the portfolio comparisons are limited to domestic data. table 2 shows the portfolio weights for the theoretical historical portfolios. as above, when comparing the 3x to the 2x letf, it is assumed 100% of the portfolio is in 2x letfs. this requires two-thirds of the 3x letf portfolio to be invested in the 3x letfs. the final question is rebalancing. if it is assumed the weights in tables 1 and 2 are optimal, the investor must determine to what extent they can deviate from those percentages. a variety of equity variance thresholds are tested to determine when the portfolios need to be rebalanced. for example, a 10% threshold implies an initial 70% equity exposure is rebalanced when the combined equity position breaches 63% or 77%. the effective exposure for the 2x and 3x portfolios is under the same constraint. for the 2x, this means the amount of wealth in equities can only deviate from 35% by �3.5% without initiating a rebalance. daily, monthly, and quarterly rebalancing is also investigated. 4. results 4.1. rebalancing to determine how often rebalancing is required to maintain a consistent risk-profile, monthly, quarterly, and allocation thresholds are tested on simulated historical data from 1946 to 2017. table 3 gives summary statistics using monthly or a 10% variance threshold for rebalancing relative to initial asset allocations described in table 2. results for the 2x and 3x are based on their effective exposure. with monthly rebalancing, a standard portfolio with 70% exposure to equities deviates from 62.65% to 73.44%. however, the 2x letf portfolio deviates from 43% to 84%, while a 3x letf portfolio deviates from 27% to 97%.4 quarterly rebalancing saw greater extremes as would be expected suggesting a variance threshold must be set for letf portfolios to keep the risk profile comparable with using standard etfs. using a 10% variance threshold demonstrates improved results in terms of absolute deviations in exposure relative to being invested in the underlying index. using a 10% table 2 portfolio composition using 1946–2017 historical data asset class portfolio 2x letf 3x letf s&p 500 50% 25% 16.67% mid cap 10% 5% 3.33% small cap 10% 5% 3.33% 20-year t-bonds 15% 7.5% 5.00% 7 to 10-year t-bonds 15% 7.5% 5.00% bond ladder, 1 to 7 years 0% 50% 66.67% 41w.j. trainor et al. / financial services review 28 (2020) 35-48 variance threshold gives results approximately equal to daily rebalancing. for portfolios using either 2x or 3x letfs, maximum exposure to equities is reduced. a 3x still reaches a maximum equity exposure of 85%, but this is the same exposure reached using daily rebalancing. thus, a tighter variance threshold is not warranted. in addition, the standard deviation of the exposure to equities is significantly reduced, especially for 3x letf portfolios. over 72 years, using a 10% threshold results in 24 rebalances for the standard portfolio, 243 for the 2x, and 651 for the 3x. this results in rebalancing for a standard portfolio, 2x, and 3x on average every three years, quarterly, and 45 days, respectively. in terms of transaction costs, assuming six trades at $5 a trade is needed at each rebalance, a standard portfolio valued at $100,000 would have 0.01% additional annual expenses, a 2x would have 0.1%, and a 3x would have 0.27%. for taxable accounts, the additional trading required using letfs will have a greater tax burden that could nullify any excess returns. these tax effect differentials are discussed in the empirical results. the reported results in this research are based on using a 10% equity variance threshold for rebalancing. it should be noted 24 rebalances over 72 years using standard etfs might seem small but consider a 20% increase in equities. with an initial $100 portfolio, the value of equities would increase from $70 to $84, but the percentage in equities only increases to $84/$114 or 73.7%. this still does not breach the 10% barrier even with no increase in the bond position. thus, a relatively large move is required to breach the barrier. a variety of variance thresholds are tested but return differences are not significantly different across portfolios based on rebalancing rules. a 10% threshold effectively controls asset exposure without requiring excessive trading. thus, only the 10% variance threshold is reported. 4.2. 2010 to 2017 portfolio returns table 4 shows the portfolio results for using the etfs and letfs shown in table 1 from 2010 to 2017. the average annual portfolio return using standard etfs is 10.26%. using 2x letfs with unit exposure to the underlying indexes increases this average return to 11.19%, while using 3x letfs increases this return to 12.06%. in any individual year, the portfolio returns are relatively similar demonstrating the use of letfs does not result in any discernable increases in risk. table 3 equity portfolio exposure for etf and letfs standard etfs 2x letfs 3x letfs monthly 10% rebal monthly 10% rebal monthly 10% rebal average 70.08% 71.59% 70.44% 71.05% 70.87% 70.94% standard deviation 0.98% 2.80% 3.91% 3.10% 6.86% 3.41% min 62.65% 62.59% 43.28% 51.22% 26.72% 36.91% max 73.44% 77.28% 83.98% 81.03% 96.92% 85.34% note: summary of portfolio exposure for etf and letfs from 1946 to 2017 based on monthly and 10% threshold rebalancing. initial and rebalanced weights are 70% equities, 30% bonds. weights for the letfs are effective exposure. 42 w.j. trainor et al. / financial services review 28 (2020) 35-48 as expected, the letf portfolio returns relative to using standard etfs rely heavily on the return to the aggregate bond portfolio (bnd). in 2013, the bnd etf has a return of �2.10% which leads to letf portfolios underperforming. in 2015, the bnd return is 0.56% also resulting in slight underperformance by letf portfolios. however, in 2010, 2011, 2012, 2014, 2016, and 2017 when the aggregate bond fund did relatively well, the letf portfolios outperform by up to five percent over a traditional portfolio (see the 3x letf portfolio in 2011). on average, even under a near zero interest rate environment over the last eight years, a portfolio of letfs outperforms a traditional portfolio. the 100% in 2x and 67% in 3x columns show the returns for doubling the exposure to the underlying indexes. returns for most years are approximately doubled, but so are the standard deviations. the compounding issue is easily seen in 2011 as the etf portfolio is slightly positive, but the fully leveraged 2x portfolio is negative. finally, the portfolio using 3x letfs to create exposure equivalent to the 2x letfs outperforms the 2x letf portfolio by approximately 1.5% annually. to further break down the performance of letfs relative to comparable etfs, table 5 shows the returns as if 100% is invested in each asset class relative to 50% in the 2x letf table 4 portfolio returns from 2010 to 2017 etfs 50% in 2x 33% in 3x 100% in 2x 67% in 3x bnd return 2010 17.11% 19.98% 20.17% 30.57% 34.52% 4.94% 2011 0.42% 2.78% 5.34% �3.77% �1.82% 7.92% 2012 13.62% 15.11% 16.62% 25.29% 28.44% 3.89% 2013 17.19% 14.85% 14.03% 34.88% 33.28% �2.10% 2014 8.43% 10.13% 12.55% 14.29% 16.46% 5.82% 2015 �1.77% �1.99% �2.06% �6.17% �6.88% 0.56% 2016 9.83% 10.25% 11.54% 16.21% 18.02% 2.53% 2017 17.21% 18.44% 18.31% 34.90% 36.62% 3.57% average 10.26% 11.19% 12.06% 18.27% 19.83% 3.39% standard deviation 7.56% 7.61% 7.31% 16.28% 16.69% 3.12% note: annual returns for a portfolio of etfs, 2x letfs, and 3x letfs along with the bnd etf returns to show what return the excess wealth from letf portfolios attained. table 5 100% effective exposure for etf and corresponding letfs etf asset etf etf 2x relative 3x relative ticker class average return standard deviation performance performance spy large cap 14.50% 10.15% �0.99% �0.67% ijh mid cap 15.11% 12.88% �0.68% �1.08% iwm small cap 14.51% 15.31% �0.88% �1.23% efa developed 7.28% 13.78% �1.49% �0.84% eem emerging 5.84% 19.76% �1.14% �1.58% iyr real estate 12.78% 11.81% �0.70% �0.18% tlt 20-year 8.20% 15.48% �0.47% �0.95% ief 7–10 year 4.29% 6.42% �0.41% �0.44% average 10.31% 13.20% �0.85% �0.87% note: comparison of annual results for 100% in the underlying etf, 50% in a 2x, and 33.3% in a 3x for each etf or letf from 2010 to 2017. remaining wealth in letfs assumed to earn zero. no statistical return differences between the etfs or letfs. 43w.j. trainor et al. / financial services review 28 (2020) 35-48 and 33.3% in the corresponding 3x letf. the remaining wealth for the letf portfolios is assumed to earn zero. the 10% variance rebalancing rule is applied to each asset class to keep the exposure close to 100% while still being able to estimate beta decay and the higher costs of letfs. on average, 50% of a 2x or 33% of a 3x underperforms its 100% etf counterpart by 0.85% for 2x letfs and 0.87% for 3x letfs. for 2x letfs, the underperformance ranges from �0.41% for 7 to 10-year treasuries to �1.49% for developed equity. the 3x letfs underperformance ranges from �0.018% for real estate to �1.58% for emerging markets. the discrepancy in ranges is due in part to differences in the volatility/return differences across assets and tracking error. it is not surprising the worst underperformance for the letfs occurs in the volatile international markets that have substandard returns given their volatility. the standard deviations for the etfs and letfs are virtually identical and thus, only the etf standard deviations are shown. in summary, these results show the inherent costs of the higher expense ratio and financing costs of letfs. from a portfolio standpoint, the remaining funds available from using letfs need to overcome this performance lag. to determine how well each letf tracks their etf counterpart, letf returns are also regressed on their corresponding etf returns from 2010 to 2017. beta coefficients range from 0.97 to 1.04 with 12 of 16 at 0.99 or 1.0. the 2x emerging market letf is the only exception with a beta of 0.90 and an r2 of 90%; r2 for all other letf regressions are 98% or better. thus, of the 16 letfs, only one did not attain almost exact unit exposure. these results contrast to tang and xu (2013) who show even realized daily leverage was less than advertised from 2006 to 2010. however, that period was more volatile and the libor rate was higher. in addition, small tracking errors are not as magnified in this study as only the inverse of the leverage is held in each fund. the market was also generally upward trending during this study’s time period resulting in a positive compounding effect, and finally, the funds themselves are possibly doing a better job of maintaining their daily leverage ratios. thus, the results suggest using letfs along with the 10% variance threshold achieves the goal of 100% exposure to the underlying indexes, albeit with an annual financing and expense ratio drag of approximately �0.85%. 4.3. 1946 to 2017 simulated historical returns to get a better idea of portfolio returns going forward, simulated returns from 1946 to 2017 are created using equation (6). table 6 shows the portfolio results using the weighting shown in table 2. subperiod returns are also shown roughly corresponding to different interest rate environments. for the entire 1946–2017 period, letf portfolios on average outperform a traditional portfolio from 0.63% to 1.41% a year. however, if this strategy is implemented in nonqualified accounts, the differences do not overcome trading costs and taxes. assuming a long-term capital gains tax of 15% and a 24% marginal income tax bracket, the approximate returns after taxes and trading costs for the 1946 to 2017 time period for the etf, 2x, and 3x portfolios are 10.54%, 9.75%, and 10.45%, respectively. this assumes all gains to the etf are taxed at 15% every three years and all 44 w.j. trainor et al. / financial services review 28 (2020) 35-48 gains to the letfs are taxed at 24% each year. thus, after taxes and trading costs, using letfs do not outperform a standard etf portfolio. for qualified nontaxable accounts, letf’s outperform, especially when the interest rate environment is relatively high such as 1979 to 1991. alternatively, during periods of extremely low rates such as the 1946–1959, a portfolio of letfs tends to underperform. however, low rates by themselves do not relegate letf portfolios to underperformance. both the empirical data in table 4 and simulated data in table 6 show letf portfolios outperforming during 2010–2017. the critical factor is the return to the freed-up wealth from using letfs relative to their implicit financing costs and higher expense ratios. the bottom of table 6 shows the standard deviation, minimum, maximum, 10% value at risk (var), along with the sharpe and sortino ratio, the latter of which measures downside risk (sortino and price, 1994). even though there is a marginal increase in the standard deviation when using a portfolio of letfs, the minimum and 10% var for the letf portfolios is better than the traditional portfolio while at the same time having higher maximums. thus, the standard deviation is misleading when measuring risk for the letf portfolios as downside risk for letfs is mitigated because of the higher percentage in a relatively safe bond ladder. the large percentage in the bond ladder is also valuable during market panics as investors flee to safer assets. results from the simulated returns reinforce the empirical data from 2010 to 2017.5 for investors seeking 2x leverage, the 100% in 2x and 67% in 3x columns show the results. average returns are less than double while the standard deviations and minimums are doubled with a tripling of the var. using the 3x letf to attain 2x exposure shows improved results. overall and in every subperiod, 67% in a 3x outperforms 100% in a 2x with a 1.6% annual average increase. the minimum and var using 3x letfs are also better than 2x letfs along with a higher sharpe and sortino ratio. table 6 portfolio average annual returns using 1946 to 2017 historical data time period portfolio 50% in 2x 33% in 3x 100% in 2x 67% in 3x treasuries 1946–2017 11.12% 11.75% 12.53% 18.96% 20.58% 5.46% 1946–1959 11.45% 10.78% 11.22% 21.91% 22.23% 1.63% 1960–1978 8.03% 8.45% 9.35% 13.26% 14.82% 5.39% 1979–1991 16.56% 18.81% 20.05% 25.40% 28.75% 11.42% 1992–2009 10.13% 11.00% 11.65% 15.40% 17.22% 5.78% 2010–2017 11.29% 11.47% 12.16% 23.80% 24.72% 1.96% 1946–2017 risk statistics standard deviation 12.67% 13.23% 13.51% 26.52% 26.93% 5.09% min �23.38% �19.45% �18.64% �46.33% �44.81% �1.66% max 38.84% 39.53% 44.18% 89.92% 90.62% 25.77% var �5.31% �4.24% �3.19% �14.57% �12.52% 0.51% sharpe 0.55 0.57 0.62 0.56 0.61 n/a sortino 2.56 2.93 3.34 1.97 2.26 n/a note: table shows average annual returns from 1946 to 2017 along with several sub-periods corresponding to different interest rate environments. because of limited data and large standard deviations, there are no statistical differences between the mean returns between the unit exposure portfolios or between the 2x exposure. 45w.j. trainor et al. / financial services review 28 (2020) 35-48 the disadvantage to this strategy is the rebalancing frequency required for a 3x. using a 10% variance threshold to maintain the initial 2x exposure to the underlying indexes requires 79 and 381 rebalances for the 2x and 3x, respectively, over the 72 years. thus, although using the 3x to mimic the exposure of a 2x increases the returns, trading costs and the more onerous tax effects in nonqualified accounts eliminate the after-tax return differential. for a letf portfolio to outperform, the returns to the t-bond ladder need to exceed the financing and expense ratio costs of letfs. however, this only occurs 45% of the time from 1946 to 2017. despite this, 2x and 3x letf portfolios outperform 54% and 61% of the time, respectively, during this time period. this is possible because a highly positive trending market allows letfs to return more over time than their daily leverage multiple implies. this can make up for small interest rate spreads. to more accurately estimate the effect of the lending minus borrowing differential on returns, the 50% in 2x and 33% in 3x letf returns minus the standard portfolio returns are regressed on the bond minus borrowing rate differential for both the 2010 to 2017 empirical data and the 1946 to 2017 simulated data. the regression results are shown in table 7. not surprisingly, the bond return minus the borrowing cost explains more than 90% of the return differences of the 2x and 3x letf portfolios relative to a standard portfolio for both the empirical and simulated historical data. based on the empirical data, a one percent return over the borrowing rate leads to a 0.41% increase for a 2x and 0.63% increase for a 3x letf over a standard portfolio. these regression results held when comparing the letfs to their corresponding etfs at an individual level and roughly correspond to the 50% and 67% freed up wealth from using 2x or 3x letf portfolios, respectively. results are similar for the 3x letf relative to the 2x letf as a one percent return over the borrowing rate corresponds to a 0.34 percentage point increase using the empirical data and 0.54 percentage point gain for the simulated data. both results are associated with moderately lower r2 as there is much more volatility with returns leveraged to 2x. thus, both the empirical and simulated data confirms the importance of the lending-borrowing spread. 5. conclusion despite the early negative press about the dangers of investing in letfs, they have become popular investment vehicles with 265� funds growing to more than $68 billion in table 7 return differences regressed on t-bond ladder or agg etf minus borrowing rate r2 intercept (t-stat) coefficient (t-stat) 2010 to 2017 2x letf–etf portfolio 92.53% �0.001 (�1.87) 0.41 (13.7)a 3x letf–etf portfolio 95.36% 0.000 (�1.17) 0.63 (17.36)a 3x letf–2x letf 64.70% 0.000 (0.35) 0.34 (4.65)a 1946 to 2017 2x letf–etf portfolio 93.13% �0.0071 (�6.14)a 0.56 (21.39)a 3x letf–etf portfolio 95.15% �0.005 (�3.54)a 0.76 (25.88)a 3x letf–2x letf 79.75% 0.003 (1.15) 0.54 (11.06)a note: astatistical difference at the one percent level. 46 w.j. trainor et al. / financial services review 28 (2020) 35-48 assets over the last 12 years (www.etf.com). while letfs have higher expense ratios and suffer from leverage decay, letfs can be used effectively to improve average returns and reduce downside risk if properly managed. one key is to periodically rebalance letfs to minimize both the impact of decay and maintain a set risk exposure. this study compares a diversified portfolio of typical etfs to portfolios comprised of 50% in 2x letfs or 33% in 3x letfs with the remainder invested in a relatively safe t-bond ladder or a bond fund. in addition, a 3x letf portfolio is compared with a 2x letf portfolio where the underlying exposure of the 3x is set equal to the 2x. using a 10% variance threshold for rebalancing, this study finds a portfolio using letfs since 2010 or one based on simulated data from 1946 to 2017 outperforms a standard portfolio. by extension, the 3x portfolio created to mimic the 2x portfolio outperforms the 2x portfolio. specifically, combining tradeable letfs since 2010 with an aggregate bond etf, a 2x letf portfolio averages 0.9% more per year while a 3x averages 1.8% more relative to a standard etf portfolio. the 3x portfolio created to mimic the 2x portfolio outperforms by 1.5% annually. using simulated data from 1946 to 2017, a 2x letf or 3x letf portfolio combined with a one, three, five, and seven-year treasury ladder outperforms a standard etf portfolio by 0.6% and 1.4%, respectively, on an annual basis with no increase in risk and better downside risk metrics as measured by minimums, var, and sortino ratios. the results extend for the 3x over the 2x with an average annual outperformance of 1.6%. the critical component for letf portfolio outperformance is how the return to the remainder of the portfolio not invested in letfs compares to the implicit borrowing costs and higher expense ratios of letfs. regression results show the difference in the bond return minus borrowing rate explains more than 90% of the difference in returns between a portfolio created with letfs versus one created with standard mutual funds. for every one percent return earned over the borrowing rate, the 2x letf portfolio outperforms by 0.41% and a 3x letf portfolio outperforms by 0.63%. this corresponds closely to the freed-up wealth invested in bonds. from a practitioner’s point of view, letfs require more active management and in any given day, these instruments are more volatile. if the rebalance rule is breached on the downside, an investor will need the “stomach” to buy more of a fund that theoretically could lose 60% or more of its value in a day (recall october 19, 1987 when the s&p 500 fell 20%). thus, a certain degree of behavioral fortitude may be needed before creating a portfolio of 2x or 3x letfs. in addition, the tax liability from the greater rebalancing neutralizes the excess returns from using letfs found in this study. thus, the implementation of the strategy should probably be limited to qualified accounts. although tax issues are detrimental, the proposed letf strategy does have the ability to be used in ira type accounts— accounts that typically do not allow margin trading. in summary, this study shows letfs can successfully be held long-term as major components of a portfolio while improving returns and reducing downside risk. whether piecemeal such as only using an s&p letf to partially or fully take the place of an s&p etf, or to completely replicate an investor’s entire portfolio of etfs, letfs, when properly managed, can enhance returns and reduce risk. 47w.j. trainor et al. / financial services review 28 (2020) 35-48 notes 1 as an example, a 10% return to the index over a 252-day trading period with a one percent daily standard deviation will result in a theoretical realized leverage ratio for a 2x and 3x letf of 1.80 and 2.34, respectively. 2 the variance over time is �t 2 � t�2 so for the index, �t 2 � 252*0.012 � 2.52%. for the 3x, the daily returns and standard deviations are 3*0.045 (.135%) and 3*1 (3%), respectively. thus, for the 3x, �t 2 � 252*0.032 � 22.68%. 3 note �2 l2 after dividing � by l becomes (�/l)2 l2 that simplifies to �2. 4 the max and min exposure levels are for the entire period and do not necessarily sum up to 100%. in addition, for the letf portfolios, the exposure does not include the allocation to the treasury ladder meaning the combined equity and bond allocation will deviate from 100% between rebalances. 5 block bootstrapping of the 1946–2017 data was performed to create 20,000 unique one-year returns. the results are qualitatively like the original historical results with the only differences being the mean returns and vars across the portfolios are statistically significantly different. references avenllaneda, m., & zhang, s. (2010). path-dependence of leverage etf returns. society for industrial and applied mathematics, 1, 586–603. carver, a. (2009). do leveraged and inverse etfs converge to zero? the journal of derivatives, 1, 144–149. charupat, n., & miu, p. (2014). a new method to measure the performance of leveraged exchange-traded funds. financial review, 49, 735–763. cheng, m., & madhavan, a. (2009). the dynamics of leveraged and inverse exchange traded funds. journal of investment management, 7, 43–62. considine, g. (2006). comparing portfolios of etfs and mutual funds, quantext. (available at https:// docplayer.net/27758586-geoff-considine-ph-d.html) george, j., & trainor, w. (2018). portfolio insurance using leveraged etfs. financial services review, 26, 1–17. lu, l., wang, j., & zhang, g. (2012). long term performance of leveraged etfs. financial services review, 21, 63–80. ott, r., & zimmer, t. (2016). determining the return-maximizing portfolio leverage and its limitations. financial services review, 25, 415–425. scott, j., & watsun, j. (2013). the floor-leverage rule for retirement. financial analysts journal, 69, 45–60. sortino, f., & price, l. (1994). performance measurement in a downside risk framework. the journal of investing, 3, 59–54. tang, h., & xu, x. (2013). solving the return deviation conundrum of leveraged exchange-traded funds. journal of financial and quantitative analysis, 8, 309–342. trainor, w., & baryla, e. (2008). leveraged etfs: a risky double that doesn’t multiply by two. journal of financial planning, 21, 48–55. 48 w.j. trainor et al. / financial services review 28 (2020) 35-48 pii: 1057-0810(91)90005-j financial services review, 1(1):23-34 copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. life insurance companies as investment managers: new implications for consumers robert t. kleiman anandi p. sahu this paper examines the attractiveness of the equity portfolios of life insurance companies as an alternative investment to mutual funds. in particular, this study analyzes the risk adjusted investment performance of the stock portfolios of ltfe insurance companies, attributable to their stock selection and market timing abilities. using conventional measures of risk-adjusted portfolio performance, we find that ltfe insurance companies exhibit performance similar to mutualfunds. the evidence suggests that the ltfe insurance companies, like their mutual fund counterparts, fail to exhibit differential stock selection or market timing abilities that are statistically significant. while the risk-adjusted investment performance of the two investment vehicles is similar, the variable annuity contracts of life insurance companies may offer an edge over mutualfunds due to their ability to defer taxes. i. introduction the investment performance of mutual fund portfolios has been the subject of extensive empirical investigation in the finance literature. according to several different studies, the average risk-adjusted performance of mutual funds rarely outperforms the market [see, for example, sharpe (1966), jensen (1968), carlson (1970), and shawky (1982)l. in addition to the analysis of the risk adjusted performance, a number of studies have examined the ability of mutual fund managers to time bull and bear market cycles and react accordingly. ideally, a portfolio manager should increase the systematic risk of the portfolio in anticipation of a market upturn and decrease the beta prior to a market downturn. however, the results of studies undertaken by treynor and mazuy (1966), fabozzi and francis (1979), veit and cheney (1982), chang and lewellen (1984), henriksson (1984) and feri, oberlhelman, and roenfeldt (1984) find no evidence that mutual fund managers are able to successfully time robert t. kleiman l assistant professor of finance, oakland university, rochester, mi 48309. anandi p. sabu l assistant professor of economics, oakland university, rochester, mi 48309. 24 financial services review, l(1) 1991 market changes and to alter their betas in anticipation of differential market conditions. consequently, the results of these studies indicate that collectively mutual fund managers are unable to outperform a passive ‘buy and hold’ investment strategy. studies which evaluate the investment performance of institutional investors other than mutual funds are more limited in number. the available evidence from these studies suggests that the performance of non-mutual fund institutional investors is no better than that of mutual funds. for example, schlarbaum (1974) found that the risk-adjusted performance of 20 property casualty insurance companies was significantly below the market averages for the 1958-1967 time period. in another study, bogle and twardowski (1980) compared the investment performance of four categories of institutional investors-banks, investment advisors, insurance companies, and mutual funds for a variety of time periods ending in 1977. their results indicated that the mutual funds achieved the best performance, followed by investment advisors, and then insurance companies, with banks achieving the poorest relative results. however, in their study, bogle and twardowski only compared the frequency distributions of returns for each category of institutional investors and did not adjust for the level of systematic risk. their conclusion is thus flawed since financial theory suggests that the evaluation of investment managers should encompass measures of both risk and return. the purpose of this paper is to provide evidence regarding the risk-adjusted investment performance of life insurance companies and to examine whether their investment performance is materially different from that of mutual funds. historically, common stocks have been a small percentage of the total assets held by this group of investors. because life insurance policies contain contractual guarantees for specified dollar amounts, bonds, rather than stocks, have been a major investment medium for these firms. moreover, in the past, there existed legal provisions which limited investments in common stock on the part of life insurance companies. however, legislation in most states now permits life insurance companies to maintain separate investment accounts for a given pension plan or group of plans. these plans maintain their assets in an account separate from the company’s other assets. separate accounts are allowed a greater latitude in making equity investments than insurance company investments in general. in addition to examining the risk-adjusted performance of this category of institutional investor in the context of jensen’s abnormal performance index, this study will also consider whether life insurance companies exhibit differential stock selection or market timing abilities in bull and bear markets. accordingly, this study will make a contribution to the finance literature in two primary ways. first, in contrast to schlarbaum’s study which analyzed property casualty insurance companies, this analysis will examine the investment performance of life insurance companies.’ second, this will be the first study life irwurance compmies as investment managers 25 to provide evidence regarding the macro-market timing abilities of a category of institutional investors other than mutual funds by comparing the systematic risk coefficient in bull and bear markets. these results, considered in conjunction with the tax deferral aspects of the life insurance company investments, will determine the competitiveness of these investment vehicles vis a-vis mutual funds. the results of this study suggest that the investment performance of life insurance companies is similar to mutual funds on a risk-adjusted basis, assuming comparable tax treatment. on average, the equity funds of life insurance companies do not appear to display significantly positive stock selection abilities. moreover, we do not find statistically significant differentials in market timing abilities in bull and bear markets. however, the evidence also suggests that the life insurance companies do not significantly underperform the market averages either, which is in contrast to previous findings for property casualty companies. thus, the findings in this paper suggest that life insurance company products should become more competitive with mutual funds in the future given the superiority of separate accounts from a tax perspective. the remainder of this paper is organized as follows. section ii provides a discussion of tax advantages and institutional characteristics of life insurance companies’ variable annuity contracts. section iii discusses the statistical techniques and empirical models used to test for stock selection and market timing abilities, and describes the data used in this study. section iv presents the empirical results of these models. finally, section v provides a summary and major conclusions. ii. tax advantagesand institutional characteristicsof lifeinsurancecompanyfunds as noted above, separate accounts are the funding vehicles for life insurance companies’ equity oriented variable annuity contracts. the separate account for a variable annuity is a unit investment trust that invests at asset values in the shares of a particular equity portfolio. both mutual funds and variable annuities provide professional management of a securities portfolio. both charge the investor for the costs of investment management and administration. like the majority of mutual funds, most variable annuities also levy a sales charge. the tax consequences of the two investment vehicles are quite different, however. unlike mutual funds, variable annuities do not act as conduits. congress enables investors to invest in the annuity contracts of life insurance companies without having to pay taxes on the dividends and capital gains until the money is withdrawn. earnings thus accumulate during the life of the annuity on a taxed deferred basis. this arrangement differs from investment in non 26 financial services review, l(1) 1991 tax qualified mutual funds, where any gain is taxed in the year it is earned, even if the gain is reinvested.’ on the other hand, the gain in the value of the annuitant’s account is not taxed as ordinary income until the payout period. other characteristics of the variable annuity contracts may, however, partially offset the tax advantages. the liquidity of the contracts is poor since irs penalties and insurance company surrender charges are imposed for early withdrawals. the irs imposes a 10% penalty charge on any withdrawals prior to age 59%. in addition, surrender charges initially can total as much as 10% of the investment, declining to zero usually through the fifth or sixth year of the contract. while variable annuities have high expenses-typically 2% of assets versus 1% for mutual funds-the difference in expenses relates to “mortality risk,” the possibility that the annuitant lives beyond what the actuarial charts anticipated. overall, the aforementioned tax and institutional characteristics appear to favor life insurance annuity contracts vis--vis mutual funds. as a result, the growth in variable annuity contracts offered by life insurance companies has been substantial. at year end 1988, 71 insurance companies were offering 391 different investment portfolios. the assets invested in variable annuity separate accounts totaled nearly $26 billion at year end 1988. this compares with 38 insurance companies offering 66 different investment accounts having total assets of $2.4 billion at year end 1978. iii. marketperformance:methodologyanddata in a seminal paper, jensen (1968) used the framework of the capital asset pricing model to investigate the investment performance of mutual funds over the 1945-1964 period. in this work, jensen developed the abnormal performance index (a!), which represents the additional return earned on a portfolio after adjusting for systematic risk. the abnormal performance index is estimated by regressing the excess returns of the portfolio on the excess returns of the market3: where &,,t the return on a given portfolio at time t; & is the risk free rate of return at time t; &,t is the average return on the market portfolio at time t; &, is the beta coefficient measuring the covariance of portfolio returns with market returns; and et is the random error term (with usual properties). a statistically significant positive value for ap can be viewed as evidence of a superior risk-adjusted performance, whereas a significant negative value is indicative of inferior risk-adjusted performance. furthermore, the coefficient of determination (r’) from the regression equation provides a measure of the diversification of the portfolio. life insurance companies as investment managers 21 fama (1972) has noted that the performance of mutual funds depends upon the ability of fund managers in two areas: (1) selectivity, i.e., selecting individual securities and (2) market timing. jensen’s measure as outlined in equation 1, however, ignores the possibility of market timing activity since &, is assumed to be stable over time.4 therefore, this study also employs a model developed by fabozzi and francis (1979) which incorporates both micro forecasting (i.e., selectivity) and macro-market timing abilities. to test whether an insurance company portfolio’s alpha intercept and/ or beta terms differ significantly during bull and bear market periods, we employ the following regression equation5: where dt is a dummy variable which is unity if the period t is a bull market and zero otherwise. the coefficient cu,’ is a measure of the differential abnormal return on the portfolio due to the manager’s security selection ability, whereas flp’ provides a measure of the differential level of systematic risk in bull versus bear markets.6 to determine if the alpha and beta parameters are equal in bull and bear markets, we examine whether the corresponding differential coefficients (op’ and pp’) are significantly different from zero. although the coefficients from the least squares (ols) estimation of equation 2 provide consistent parameter estimates, there may be a problem with heteroscedasticity in the error term (et) which causes the parameter estimates to be inefficient. therefore, we use the weighted least squares (wls) regression analysis, as suggested by henriksson (1984), to correct for heteroscedasticity.7 since the 40 insurance company funds used in this study are certainly not independent of one another, we also employ a seemingly unrelated regression (sur) model to test the market timing and stock selection ability for the sample as a whole. however, since the explanatory variables in all the 40regressions, under any one specification, are the same, the wls results are identical with sur estimates.’ thus, we refer only to wls when reporting our results. two different measures of bull and bear markets are employed in the empirical tests. the first is forbes magazine’s definition of bull and bear markets which is based on general market trends. the second measure simply categorizes any month in which the market return is positive as a bull market and any month in which the market return is less than or equal to zero as a bear market. the sample used in this study consists of 40 equity-oriented investment funds managed by life insurance companies with complete monthly return data for the 1 l-year period from october 1974 through september 1984. the specific funds included in the sample are displayed in appendix a. the returns for each 28 financial services review, l(1) 1991 insurance company portfolio are obtained from computer directions advisors (cda), and include both dividend income and capital appreciation. we confine ourselves to the 1974-1984 period as the data for the more recent period are not readily available from cda for research purposes. the value-weighted standard jz poor’s 500 stock index is employed as the proxy for the market portfolio. the yields on treasury bills having approximately one month to maturity are obtained from ibbotson associates (1986) and these serve as the measure of the risk-free rate of return. iv. ma~rketperformancezempiricalresults initially, we estimate the regression specification used by jensen (1968): which ignores the possibility of market timing activity and does not distinguish between bull and bear markets. the results using equation 1 are summarized in table 1. although cr, is, on average, positive, it is statistically insignificant for 30 out of the 40 insurance company funds at the 5% level of significance. moreover, only six of the portfolios have a statistically significant positive cu, at the 95% level of confidence-four funds are significantly negative at this level of confidence. at the 99% level of confidence, the number of funds with a statistically significant positive i+ drops to three, and one fund remains significantly negative. therefore, the results provide little evidence of a superior risk-adjusted performance by the insurance companies’ portfolio managers when market timing is ignored. however, the results suggest that life insurance companies offer superior investment performance as compared to property casualty insurance companies since schlarbaum (1974) found that the latter group significantly underperformed the market averages over the 1958-1967 period. the average &, for the life insurance companies’ portfolios is .9138, which indicates that the insurance companies undertake less systematic risk than the market as a whole. however, this figure is higher than the beta coefficient of .8013 found by schlarbaum for property-casualty companies over the 1958-1967 time period. this suggests that life insurance companies are less risk-averse than property-casualty companies. the r’ of .7735 suggests that the life insurance companies’ portfolios are well diversified. as indicated earlier, the above analysis ignores the market timing activity undertaken by investment managers and fails to distinguish between bull and bear markets. to examine the separate contributions from micro stock selection and macro market timing, the regression equation specified in equation 2 is estimated. lije insurance comjmnies as investment managers 29 table 1. regression results without using bull-bear dummies (1974:10-1984:9) &v rr.r = crp + pp(i?w rr,r) + et parameter estimates with heteroscedasticity correction ( wls) mean (sdd) range (min, max) up .0381 (.1478) (-.4960, .4196) bp .9138 (.0448) (.6225, 1.242) adj r’ .7735 (s135, .9179) test criterion reject ap = 0 at 5% reject ap = 0 at 1% a, >0 bp> 1 number of finds: 6(ap > 0) 4(ap < 0) 3cffp > 0) l(ap<o) 23 11 using forbes’ definition of bull and bear markets, the results for equation 2 based on wls are displayed in table 2. twenty-six of the insurance companies had a positive alpha differential. however, the alpha differential was significantly positive for only four of the insurance companies at the 95% level of confidence. moreover, none of the insurance company portfolios exhibited a statistically significant alpha differential at the 99% confidence level. thus, the stock selection abilities of the portfolio managers do not appear to have been materially different in bull versus bear markets. using forbes’ definition, twenty of the portfolios had a positive beta differential. only one insurance company had a statistically significant increase and one had statistically significant decrease in beta at the 95% confidence level. thus, the results based on forbes’ definition do not provide evidence for the hypothesis that insurance company portfolio managers are able to successfully alter their levels of systematic risk in anticipation of differential market conditions. the results using the second definition, in which a positive return on the market is classified as a bull market and a negative return a bear market, are displayed in table 3. the alpha differential is significantly positive for nine portfolios at the 95% level of confidence. however, only three of the portfolios display a statistically significant alpha differential at the 99% confidence level. thus, the results provide only a weak support for differential stock selection abilities in bull versus bear markets based upon the above definition of market conditions. using the second definition, a positive beta differential was found for twenty-eight of the portfolios. however, only one of the insurance companies 30 financial services review, l(1) 1991 table 2. regression results using bull-bear dummies based on forbes’ definition of bull market (general market trends determining bull/bear market) (1974:10-1984:9) &,i rj, = cu, + ap’dt + j?p(rm,r &r) + &‘[di * (rm,r &)i estimates with heteroscedasticity correction ( wls) test characteristics mean (sd) range (&fin, max) -.0423 (.2465) (-.7677, .3954) .i421 (.3212) (-.3992, x836) 9024 (.osss) f.6179, 1.353) .0070 (.1025) (-.2343, .2x01) .7757 (so99, .9208) reject a,’ = 0 at 10% reject a,’ = at 5% reject a,,’ = at 1% reject &’ = at 10% reject &’ = at 5% reject &,’ = at 1% ap’ > 0 pi>0 number of funds: 7(a/ > 0) o((ypl <o) 4(ap’ > 0) o(a/ < 0) o(a/ > 0) o(cyi < 0) 3@,’ > 0) 403; < 0) upp’ > 0) 10% < 0) 0%’ > 0) o(p/ < 0) 26 20 exhibited a significantly positive increase in beta during bull market periods at both the 95% and 99% confidence levels. therefore, the results do not suggest that insurance company managers are able to successfully time the market. while the results are sensitive to the definition of bull and bear market conditions, we find some support for the hypothesis that life insurance companies exhibit differential stock selection abilities, but not market timing abilities in bull versus bear markets.’ thus, the results reported in this paper provide further support for the efficient market hypothesis and confirm the efficacy of a passive ‘buy and hold’ investment strategy for another category of institutional investors. taken as a whole, the evidence presented in this paper is consistent with the findings of fabozzi and francis (1979) for mutual fund portfolios. this suggests that life insurance company portfolios are competitive with mutual funds on the basis of risk-adjusted measures of portfolio performance, assuming comparable tax treatment for the two investment vehicles. however, as noted in section ii, the variable annuity contracts of life insurance companies permit individual investors to defer the taxation of dividends and capital gains until the payout period. therefore, on an after-tax basis, the returns on the lie table 3. regression results using bull-bear dummies based on market movements (bull: il,* > 0, bear: rm,$ 2 0) (1974:10-1984:9) &j rf,l = cw, + c+ dt + &(rm,r rf,t) + ep'[dt * (rnt,, rdl + et estimates with heteroscedasticity correction ( wls) test churucteristics mearz (sd) range (min, max) ffp -. 1925 (.3550) (-1.986, .8853) ct,’ .2770 (.4358) (-.7960,2.399) fb .8511 (.0995) (.2709, 1.271) so654 (.1312) c-.3785, .2882) adj i?” .7722 (s372, .9180) reject cy,’ = 0 at 10% reject cl,’ = at 5% reject cup’ = at 1% reject &’ = at 10% reject &’ = at 5% reject &’ = at 1% ckipu,’ > 0 fli,‘>o number of funds: lo((w,‘> 0) o( ff,’ <o) 9(ffpl> 0) o(cx/ < 0) 3(01pl> 0) o(cu,’ < 0) 3@pr > 0) i@,‘,< 0) wpt>o) wp’< 0) w$ > 0) 25 o(pp’< o) 28 insurance companies’ equity portfolios appear to be favorable in comparison to those achieved on mutual funds if the contract holders do not withdraw funds prior to age 59%. iv. summaryandconclusions this paper provides evidence regarding the risk-adjusted investment performance of a previously unexamined category of institutional investors life insurance companies. using jensen’s measure, the results provide little evidence of superior risk-adjusted performance by the life insurance companies over the 1974-1984 period. however, in contrast to property-casualty companies, life insurance companies do not significantly underperform the market averages either. using a dummy variable regression model developed by fabozzi and francis, this paper also conducts a joint test for the presence of differential stock selection and market timing abilities on the part of the life insurance funds during bull and bear market periods. in general, the findings indicate that the stock selection and market timing abilities of life insurance companies do not differ during bull and bear markets. thus, the findings indicate that life 32 insurance companies investment strategy. these results are financial services review, l(1) 1991 are unable to outperform a passive ‘buy and hold’ similar to those obtained in previous studies of mutual fund portfolios. thus, on the basis of risk-adjusted investment performance alone, both life insurance and mutual fund portfolios yield similar returns. however, as life insurance contracts offer the opportunity for individual investors to defer taxes, these investment vehicles may offer an edge over mutual funds when performance is considered on an after-tax basis. appendix a sample of life insurance company investment funds company name description aetna life insurance separate act. 1 aetna life insurance separate act. 2 bankers life co. separate act. a fund confederation life american common stock connecticut general life separate act. 3 connecticut general life separate act. a connecticut mutual life cm equity (sa-c) first variable life fund a first variable life fund b general american life verco fund guardian insur. & annuity variable act. 1 home life equity fund jefferson std. life sep. act. a-growth div. life insur. co. virginia separate act. a maccabees mutual life separate a fund massachusetts mutual separate inv. act. a metropolitan life separate act. 1 metropolitan life separate act. 5 minnesota mutual life separate act. a minnesota mutual life separate act. b minnesota mutual life separate act. c mutual benefit life variable act. 1 mutual of new york pooled act. 2 fund new england mutual life sep. equity securities new england mutual life sep. capital growth north american life nalco inv. funds u.s. pacific mutual life equity sep. act. 1 phoenix mutual life comb. sep. act. a pilot life sep. act. a-growth div. life insurance companies as investment managers company name description 33 provident mutual life separate act. 1 fund prudential insur. vca-5 prudential insur. vca-9 security benefit life series e-l security benefit life series i-l state mutual amer. sep. inv. act. a travelers insur. sep. act. a travelers insur. sep. act. b travelers insur. sep. act. c union central life pooled equity sep. act. united of omaha var. fund-a acknowledgments: we wish to thank an anonymous referee and the editor (lewis mandell) for helpful comments an earlier draft of this paper. 1. 2. 3. 4. 5. 6. 7. 8. notes note that because the risks insured by life insurance companies are more predictable than those insured by property and casualty insurance companies, the investment strategies undertaken by the two groups of insurance companies may differ. in particular, life insurance companies have lower liquidity requirements and a greater tolerance for loss of principal than property-casualty companies. this is true for individual investors. however, institutional investors, such as pension plans, are not subject to taxation on the returns on their invested assets. it may be noted that composite measures of portfolio performance based on the capital asset pricing model are not without problems. roll (1978) has argued that the theoretical market portfolio should include other risky assets in addition to common stocks. thus, the choice of benchmark indices, such as the s&p 500, may lead to ambiguous results regarding performance measurement. stambaugh (1982), however, has indicated that the empirical findings of tests of the capm do not appear to be very sensitive to the composition of the market portfolio. the portfolio beta may change even if the manager does not change the risk of the portfolio. first, the betas of individual securities may themselves not be intertemporally stable. second, changes in the relative market value weights of the individual securities will in turn lead to a change in the portfolio beta. we use evidence from bull and bear markets to assess market timing abilities on the part of investment managers. an alternative approach to examining market timing abilities is discussed in henriksson (1984). in contrast to fabozzi and francis (1979), we estimate equation 2 in risk-premium form. the results are, however, unlikely to be materially different from those based on the specification employed by fabozzi and francis. see henriksson (1984) for details of the methodology. see johnston’s econometrics methods for a discussion of conditions under which wls and sur results become identical. 34 financial services review,l(l) 1991 9. both kon (1983) and henriksson (1984) find that there is negative correlation between measures of security selection and market timing (of mutual funds). jagannathan and korjaczyk (1986) proposed that such a negative correlation could spuriously arise as a result of investing in options or levered securities. however, our sample of life insurance companies did not exhibit this negative correlation. references bogle, j., and j. twardowski. 1980. “institutional investment performance compared,” financial analysts journal, 36( 1): 33-41. carlson, r. 1977. “aggregate performance of mutual funds 1948-1967,” journnl of financial and quantitative analysis, 5( 1): l-32. chang, e., and w. lewellen. 1984. “market timing and mutual fund performance, journal of business, 57( 1): 57-72. fabozzi, f., and j.c. francis. 1979. “mutual fund systematic risk for bull and bear markets: an empirical investigation, journal of finance, 34(5): 1243-1250. fama, e. 1972. “components of investment performance,” journal of finance, 27(3): 551-567. ferri, m., h.d. oberhelman, and r. roenfeldt. 1984. “market timing and mutual fund portfolio composition,” journal of financial research, 7(2): 143-150. francis, j.c. 1988. management of investments, 2nd edition. new york: mcgraw hill. henriksson, r. 1984 “market timing and mutual fund performance: an empirical investigation,” journal of business, 57(l) part 1: 73-96. ibbotson associates. 1986. stocks, notes, bonds and bills. chicago. jagannathan, r., and r.a. korajczyk. 1986. “assessing the market timing performance of managed portfolios,” journal of business, 59(2) part i: 217-235. jensen, m. 1968. “the performance of mutual funds in the period 1945-1964,” journal of finance, 23(2): 389-416. johnston, j. econometrics methods. new york: mcgraw-hill. kon, s.j. 1983. “the market-timing performance of mutual fund managers,” journal of business, 56(3): 323-347. life insurance fact book. 1984. washington, dc: american council of life insurance. roll, r. 1978. “ambiguity when performance is measured by the securities market line,” journal of finance, 33(4): 1051-1069. schlarbaum, g. 1974. “the investment performance of the common stock portfolios of property-liability insurance companies,” journal of financial and quantitative analysis, 9( 1): 89-106. sharpe, w. 1966. “mutual fund performance,” journal of business, 39(l) part 2: 119-138. shawky, h. 1982. “an update on mutual funds: better grades,” journal of portfolio management, 8(2): 29-34. stambaugh, r. 1982. “on the exclusion of assets from tests of the two parameter model: a sensitivity analysis,” journal of finunciuz economics, lo(3): 237-268. treynor, j., and k. mazuy. 1966. “can mutual funds outguess the market?” harvard business review,“43(1): 131-136. veit, e.t., and j. cheney. 1982. “are mutual funds market timers?” journal of portfolio management, 8(2): 35-42. ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires december 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in “mutual fund knowledge assessment for policy and decision problems” by scholl and fontes, which of the following groups had the highest levels of mutual fund knowledge? a. investors b. non-investors c. those with high levels of general financial knowledge d. those with low levels of general financial knowledge 2. based on the findings of scholl and fontes, respondents answered questions related to ________ incorrectly most frequently. a. fees b. performance history c. risk d. disclosure 3. according to scholl and fontes, respondents with higher levels of financial well-being had ______________ levels of mutual fund knowledge. a. lower b. about the same c. higher d. none of the above 4. in “financial advisor use, life events, and the relationship with beneficial intentions” by sommer, lim, and macdonald, the study conceptualized life events as a motivation for individuals to move from the precontemplation to contemplation stage of change. according to the study, life events result in: a. social liberation b. dramatic relief c. self-reevaluation d. consciousness-raising 5. according to sommer, lim, and macdonald, the most frequently cited beneficial financial planning intention planned for the upcoming 12 months is: a. establish an emergency fund b. reduce debt c. save for retirement d. reevaluate insurance manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) financial literacy and the level of financial planning individuals use lua a. v. augustina,*, terrance k. martinb aeisenberg college of business, slippery rock university, rm 308a, 1 morrow way, slippery rock, pa 16057, usa bcollege of arts, sciences, business, and education, winston-salem state university, reynolds center, rm 111, 601 s. martin luther king jr. drive, winston-salem, nc 27110, usa abstract there is little evidence in the existing literature on the relationship between financial literacy and the level of financial planning that individuals use. different financial decisions require varying levels of financial planning, ranging from simple planning to complex planning. these decisions involve tradeoffs between taxes, retirement planning, and consumption to name a few. making these decisions can be difficult without the right knowledge, and financial literacy can ease the decision-making process. this article evaluates the relationship between financial literacy and the level of financial planning behavior that individuals display. we use the 2012 and 2014 waves of the national longitudinal survey of youth (nlsy79) and found that higher levels of financial literacy are positively associated with more “comprehensive” levels of financial planning behaviors. the results of this article indicate that individuals at the highest level of financial literacy are more likely to choose “comprehensive” planning over doing nothing at all. in addition, those with higher levels of financial literacy are also more likely to display some level of financial planning behavior compared with doing nothing at all. the implications of this may support efforts related to providing broad based financial education, with a specific focus on the value of planning, in the hopes increasing the financial literacy and level of planning of the overall population. © 2022 academy of financial services. all rights reserved. jel classifications: g40; g41; g50; g53 keywords: financial literacy; financial planner; financial advisor 1. introduction although there is a lot of research on financial literacy and financial planning, there is not much in the way of how financial literacy is associated with the level of financial planning *corresponding author. tel.: +1-214-444-9631; fax: 724-738-2959. e-mail address: lua.augustin@sru.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 205–222 consumers use. we examine the role financial literacy plays in the levels of financial planning chosen by individuals. financial decisions are complex, and some level of financial planning can help individuals navigate these complex options. as people become financially literate, they should become more aware of the complexity of their financial decisions, and as a result, consume more “comprehensive” levels of financial planning. comprehensive financial planning covers a wider range of financial situations than simpler forms of planning. financial planning can range from simple to complex, depending on the knowledge, cost, and time required. creating a household budget may be classified as simple, while preparing intricate tax returns may be classified as complex. an individual with very little knowledge can set up cash inflows and outflows and mistakes would be minimal. it takes more knowledge and expertise to prepare items such as taxes, especially where there are multiple income sources or tax brackets. in this case, even what seems to be a minor error can result in large financial costs. the more complex the financial task and the higher the potential cost of mistakes, the more prudent it would be to engage in acquiring financial education. the financial decisions made by individuals will affect their quality of life. because many of these decisions are complex, making the right decisions can be a challenge for those with limited financial knowledge. individuals selecting products and strategies must understand performance and pricing characteristics, ever-changing tax laws, the value of money over long periods of time, and the basics of financial theory. all of these factors can be very confusing to the average consumer, which would suggest that individuals would benefit from increased financial knowledge and planning (bae & sandager, 1997; calcagno & monticone, 2015). a question that remains is, are people with higher levels of financial literacy more aware of the varying complexity of financial decisions and the value of financial planning? financial literacy leads to awareness of the intricacies of financial assets and this awareness should encourage people to make better plans. individuals with high financial literacy tend to make better decisions than those with low literacy. they also tend to engage in more complex forms of financial planning behavior, and better planning is associated with better outcomes (hilgert, hogarth, & beverly, 2003). in many cases, financial advice serves as a complement to financial literacy (calcagno & monticone, 2015; collins, 2012), as the more literate tend to seek advice. those with low levels of literacy, financial or otherwise, tend to be less inclined to engage in financial planning activities (cole, sampson, & zia, 2010). they are more likely to do nothing at all when it comes to managing their finances. there is evidence that those with higher levels of financial literacy are more likely to seek professional advice (calcagno & monticone, 2015; kramer, 2012), indicating financial literacy may be a complement to advice (calcagno & monticone, 2015). if financial literacy is a complement to financial planning, then financial literacy efforts may increase demand for professional financial planning services. we analyze responses from the 2012 and 2014 waves of the national longitudinal survey of youth (nlsy79) and evaluate the relationship between financial literacy and the use of financial planning at various levels and found support for our hypotheses. we found that as financial literacy levels increase, they are positively associated with demand for more comprehensive levels of financial planning. we also found that decreases in financial literacy are negatively associated with demand for more comprehensive levels of financial planning. 206 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 2. literature review human capital theory states that individuals will invest in their human capital if the benefits outweigh the costs of doing so (becker, 1994). when topics are complex, the acquisition costs of human capital are high and require large investments of time, energy, and effort. many in the financial planning profession have invested years into the education and certifications requirements that qualify them as experts in the field. for the average consumer, investments of this size may not practical or efficient. however, individuals acquire some level of financial knowledge and literacy through education and experience. for those that acquire higher levels of literacy, evidence from the literature has shown improved decisions and outcomes related to a number of financial decisions, one of which, is a higher likelihood to plan for retirement (j. r. agnew, bateman, & thorp, 2013; vinet & zhedanov, 2010). while this is positive, the average consumer likely lacks the financial sophistication to adequately address all the decisions that are required plan for this goal. to bridge the gap individuals would have to engage in higher levels of financial planning, which would require the acquisition of additional human capital or the renting the human capital from an expert. in either case, higher levels of literacy would be associate with a better understanding of the personal deficiencies related to the decisions at hand. 3. financial literacy financial literacy can be defined as the measurement of how well individuals use information related to personal finance (huston, 2010). increased financial literacy improves financial decision making (huston, 2010; meier & sprenger, 2013; sekita et al., 2011) and financial outcomes including reduced debt and increased participation in retirement saving plans (clark, ambrosio, & mcdermed, 2003; delavande, rohwedder, & willis, 2008; howlett, kees, & kemp, 2008; vinet & zhedanov, 2010; willis, 2008). human capital theory predicts that individuals will invest time in acquiring knowledge when the expected return exceeds the expected time and transaction costs (blundell, dearden, meghir, & sianesi, 1999). individuals will invest in their own financial literacy if they believe the benefits from improved financial decisions exceed the costs of increased knowledge. this knowledge can be used to manage their own finances, but increased knowledge may also allow an individual to better recognize which decisions are too complex and are best delegated to a professional (collins, 2012; meier & sprenger, 2013). however, not everyone has an adequate level of financial literacy or is willing to seek a professional who has that financial literacy. individuals with lower financial knowledge may not recognize the value received from increased planning. those with lower levels of financial literacy are less likely to adequately plan for retirement and make poor decisions during their earning years (van rooij, lusardi, & alessie, 2012). poor and less educated households are more likely to be financially illiterate (sekita et al., 2011) and tend to make investment mistakes more frequently than their counterparts (campbell, 2006). most individuals are still unable to understand simple l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 207 financial concepts like compound interest (collins, 2012; lusardi, 2008) and this affects tasks such as portfolio choice and retirement planning. low levels of financial literacy may even lead to nonparticipation in the stock market (van rooij, lusardi, & alessie, 2011). in contrast, those with higher levels of financial literacy make better financial decisions. they are more likely to be diversified in their portfolios and have retirement savings (de bassa scheresberg, 2013; vinet & zhedanov, 2010). financially literate individuals display better reasoning and numeracy skills and are also able to make more complex decisions. this may translate to demanding more complex levels of financial planning. 3.1. what drives demand for professional advice services? uncertainty about the results of financial outcomes may drive individuals to seek professional financial advice. because people are generally risk averse, a recognition of the uncertainty related to financial outcomes would likely lead individuals to seek financial advice to hedge against variations in financial outcomes. in fact individuals who admit they lack knowledge about their finances were more likely to consult a professional (bae & sandager, 1997). in the financial world, demand for professional services is driven by a variety of factors including but not limited to the complexity of financial decisions and the level of financial resources available (hanna, 2011). people also may rely on advice from friends and relatives (hanna, 2011; kramer, 2012), especially spouses (chen & volpe, 1998; hackethal, haliassos, & jappelli, 2012), and may use them as substitutes for professional advice. in this case, they may not see the need for professional financial advice due to their reliance on other resources. it is also possible that the direction of effect may flow from increased literacy to increased demand, or may arise if financial planners target higher wealth clients who happen to be more financially literate (collins, 2012; hackethal et al., 2012; vinet & zhedanov, 2010). 4. varying levels of financial planning in this study, we consider four different levels of financial planning. “no plan” refers to respondents who have not calculated their retirement needs, nor consulted a financial planner. these are individuals who are not actively involved in managing their financial situation. “retirement only” refers to respondents who have only calculated their retirement needs but have not consulted a financial planner. these are individuals who decide to manage their finances without the use of outside help. “plan only” refers to respondents who have gotten advice from a financial planner, instead of relying solely on personal human capital. these are individuals who have taken the extra step of consulting a professional. in this case they may have realized that their financial situation is too complex to manage on their own. “comprehensive” refers to respondents who have both calculated their retirement needs and have consulted a financial planner. these are the individuals 208 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 who take an active role in managing their finances as well as consulting a professional for advice (fig. 1). using almost any level of financial planning is associated with increased saving for retirement (van rooij et al., 2012), greater savings in general (jappelli & padula, 2013), increased stock market participation (van rooij et al., 2011), greater portfolio diversification, and improved investment performance (hackethal et al., 2012). individuals may choose to not plan for their financial futures because they are overwhelmed with information (j. r. agnew & szykman, 2005; schwab et al., 2008). this principle can be applied to individuals with formal, detailed plans, versus those who did nothing and or indulged in noncomprehensive planning. doing nothing and having no plan are not the same. in situations with automatic enrollment into retirement plans, individuals “do nothing” but still have some level of financial planning (beshears, choi, laibson, & madrian, 2009). in this case it is done for them instead of them being an active participant. individuals who do nothing would not choose to enroll in retirement plans where it is not automatic. they would fare worse than those who auto enrolled as they would not even have retirement savings put in place. individuals may choose to have no plan, either as a result of ignorance or from being overwhelmed by options (sethi-iyengar, huberman, & jiang, 2004). there is a lot of information widely available to the public when it comes to financial planning. in addition to the lengthy prospectuses from investments, there is a barrage of advertising. individuals are also bombarded with mobile apps that state individuals can do all their own planning. this information overload can lead to paralysis and inaction when it comes to financial planning (j. r. agnew & szykman, 2005; sethi-iyengar, huberman, & jiang, 2004). individuals who become overwhelmed may choose to do nothing because they cannot sift through the relevant information. individuals may choose to calculate their retirement needs on their own and forgo seeking professional advice. individuals who either calculate their own retirement needs or seek professional advice tend to have higher net worth (hanna, 2011), compared with those who do no planning. individuals who calculate their retirement needs can more effectively manage their 401(k) contributions than those who do no planning. those who contribute regularly are more likely to be better prepared for retirement than those who do not. contributions can affect future taxation (gokhale & kotlikoff, 2003), so in this case it may be wise to seek additional levels of financial planning, like from a professional. individuals who calculate their retirement needs only may shortchange themselves as they are only focused on one area of their financial lives. they may be overlooking items such as wills and healthcare if they only focus on retirement. healthcare costs can skyrocket as people age and this can be fig. 1. levels of financial planning. l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 209 burdensome with limited income during retirement (butrica, goldwyn, & johnson, 2005). working with a professional may have the same pitfalls as focusing solely on retirement, if the professional does not offer comprehensive advice. due to the complex nature of the financial arena, individuals may choose to rent the human capital of a professional rather than invest in increasing their own financial literacy. individuals who seek a financial planner may not wish to do their own financial planning. there are multiple definitions of financial planners and each one offers different services (elmerick, montalto, & fox, 2002). while there is much to be said about variation in the types of advice and advice quality, we focus on the evidence related to engaging a financial professional and how that decision is related to the outcomes that individuals experience. financial planners are specialists in their field. they have invested considerable amounts of time studying the complexities of the financial area so they can give advice. use of financial planners is associated with positive financial outcomes. financial planners can help individuals understand the risk associated with holding on to losing stocks. they can also help set up investment plans that prevent individuals from constantly buying and selling; thereby, losing money (barber & odean, 2000). finally, individuals may choose a comprehensive level of financial planning. at this point, individuals would have looked at their own financial situation as well as recognizing the need for outside help. they may have decided that they should take an active part in planning their financial future but also recognize the need for professional services. even when seeking the help of a professional it would be wise for individuals to maintain some level of financial literacy so they can be aware of the level of assistance they are getting. individuals would still need to understand the workings of their financial situation instead of simply leaving it all in the hands of the professional. individuals who are aware of their financial standing tend to make better informed decisions than those who are not. individuals who choose the comprehensive level of financial planning seek help from some financial planning professional. financial planners can guide individuals to the right investments for their given level of risk, avoid frequent trades, and avoid behavioral mistakes such as the disposition effect (collins, 2012). financial planners can provide expertise to help individuals make better choices (hackethal et al., 2012), but few individuals actually use a planner (hanna, 2011) when they need one. although individuals can make these decisions through great mental effort, it would be better left to a professional. we analyze data from the 2012 and 2014 waves of the nlsy79 to examine the relationship between financial literacy and levels of financial planning. we found support for the hypothesis that increased financial literacy is positively associated with higher levels of financial planning. 5. data, hypotheses, and methods 5.1. data this article uses the 2012 and 2014 waves of the nlsy79. this is a nationally representative longitudinal study of americans ranging from age 14 to age 22 at date of first interview in 1979. the united states bureau of labor statistics sponsors the survey and the interviews are collected by the ohio state university. the full survey contains 12,868 respondents and our final sample contains 9,432 observations. the study contains questions related to the level of financial planning individuals use, making it an excellent choice for analysis. it explicitly asks if 210 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 individuals have calculated their retirement needs, along with asking if they have sought any type of financial planning advice. the study also contains questions about the relevant independent variables such as income and age, which we include in our analysis. 5.2. sample the sample used in this article contains 9,432 respondents after we censor the data. we initialize the sample by removing any observations who did not provide a response to the following two questions: 1. have you [or] [spouse/partner’s name]: “. . . ever calculated how much retirement income you would need at retirement?” 2. have you [or] [spouse/partner’s name]: “. . . consulted a financial planner about how to plan your finances after retirement?” we further censor the sample by removing nonresponses to the independent variable of interest, financial literacy. the three questions used are as follows: 1. suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow: more than $102, exactly $102, less than $102? 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, would you be able to buy more than, exactly the same as, or less than today with the money in this account? 3. do you think that the following statement is true or false? “buying a single company stock usually provides a safer return than a stock mutual fund.” we finalize the sample by removing nonresponses to other independent variables, arriving at the final sample size of 9,432 observations. 6. hypothesis hypothesis 1: financial literacy is not related to the level of financial planning individuals choose. hypothesis 1a: financial literacy is related to the level of financial planning individuals choose. we expect to find a positive relationship between financial literacy and comprehensive financial planning. theory indicates that individuals would invest in financial literacy if the benefits outweigh the costs. this could mean increasing their own financial literacy or using the financial literacy of a professional as a proxy. prior literature indicates the more financially literate tend to be more educated, wealthier, and use financial planning at more comprehensive levels. 7. method to test the question, we use the following conceptual model: l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 211 level of financial planning ¼ f financial� literacy, demographics � � we use the multinomial probit model as follows: y�ij ¼ b jx 0 i þ « ij � � here y�ij represents the probability of choosing one of the four levels of financial planning (lfp), x0i represents a vector of variables including financial literacy, race/ethnicity, gender, income, net worth, education level, age, and marital status, and « ij represents the error term. 7.1. dependent variable the dependent variable is lfp and there are four categories, namely comprehensive, plan only, retirement only, and no plan. we create the lfp variables by combining two questions from the survey (see appendix b). one question asks if the respondent consulted a financial planner and they could answer yes or no, as well as refuse to answer the question. the second question asks if the respondent has calculated the amount needed for retirement, and the available choices are the same as for the question on whether they consult a planner. we combine these two questions to form the four lfp options. respondents are given a value of 1 if they answered yes to the questions and 0 if they did not. respondents who consult a planner and calculate retirement income needs are coded as “comprehensive plan.” respondents who only consult a planner but do not calculate retirement needs are coded as “plan only” (non-comprehensive). respondents who do not see a planner but have calculated retirement income are coded as retirement only (non-comprehensive). finally, respondents who have neither consulted a planner nor calculated retirement income needs are classified as no plan. 7.2. independent variables the independent variables comprise financial literacy (generated using three questions), race/ethnicity, sex, income, net-worth, education, age, and marital-status. the level of risk aversion is an important factor in the demand for financial advice (hanna, 2011); however, only 247 responded to the question and after censoring it dropped the responses to 214. we decided to omit the risk variable due to the small number of viable responses. 7.3. financial literacy variable we create the financial literacy variable using a prior model (lusardi & mitchell, 2005) by combining three questions from the nlsy79. the three questions are a measure of inflation, numeracy, and stock fund knowledge. the financial literacy measure was developed in 2005 and has been used in various studies (lusardi & mitchell, 2006). this measure of financial literacy is divided into three groups as follows: low literacy, moderate literacy, and 212 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 high literacy. low literacy is defined as getting zero questions correct, moderate literacy is defined as getting one to two questions correct, and high literacy is defined as getting all three questions correct. the income variable is a self-reported, continuous variable and will be transformed into four quartiles for the regression. net worth is also self-reported and is grouped into four quartiles for the analysis. the age variable is a self-reported linear variable and will be transformed into the following age ranges: 47–50, 51–55, and 56 and above. education is an ordinal variable ranging from 0 to 20. a value of 0 represents no school completed, 1 means first grade is the highest grade completed and so on with 20 representing the eighth year of college. the varying levels of education are broken down into categories based on college degree versus no degree for the purpose of this article. the gender variable is a binomial value with 1 being male and 0 being female. race/ethnicity is a discrete variable ranging from 1 through 3 with 1 being hispanic, 2 black, and 3 being non-hispanic non-black. race/ethnicity will be used as a dummy variable for the regression. 8. model we use the multinomial probit regression method to measure the relationships between the independent variables and the level of financial planning used by respondents. this regression method allows for choice when there is no specified order to the options. while one could argue that having some level of financial planning is better than doing nothing at all we cannot make a case for a ranking between retirement only and plan only. in fact lusardi and mitchell (lusardi & mitchell, 2005) refer to those who calculate their retirement needs as using simple planning. the simple level of financial planning would fall in between no planning and comprehensive planning. it would be difficult to make a hierarchical case for retirement only versus plan only at this point. 9. results 9.1. univariate analysis the sample for analysis contains 9,432 observations. table 1 shows the descriptive statistics for the sample. about 18% of the sample use a comprehensive level of financial planning, about 7% use the plan only level of financial planning. about 17% use the retire only level, and about 58% use no plan. about 50% of the sample have high financial literacy, meaning they got all three questions correct. about 35% of the sample have moderate literacy, with the remaining 14% of the sample scoring in the low literacy range. about 19% of the final sample is hispanic, 31% is black, and the remaining 50% is nonblack non-hispanic. although we recognize that hispanic is not a race classification, this is l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 213 the title that is used in the data collection method of the survey (see appendix 3: demographics questions). about 51% of the sample is male, while the remaining 49% is female. the sample has a mean age of 52 years and ranged from 47 years to 58 years. about 71% of the sample have less than a college degree, with the remaining 28% having a college degree or higher. the portion of those reporting no college degree includes respondents currently in college who have not completed their degree at the time of the survey. this sample has a mean income of about $44,000, with a maximum income of about $370,000. the mean net worth for the sample is about $270,000, with a maximum net worth of slightly above $5,000,000. the income and net worth of the respondents are divided into quartiles for the purpose of analysis. finally, about 62% of the sample is married, including those who are currently separated. the remaining 38% are unmarried and this includes divorced, widowed, and never married individuals. table 1 descriptive statistics for level of financial planning sample (n = 9,432) variable frequency level of financial planning comprehensive 17.90% plan only 6.81% retire only 17.26% no plan 58.04% financial literacy low 14.40% moderate 35.41% high 50.19% race/ethnicity hispanic 17.37% black 31.09% non-black non-hispanic 51.54% sex male 51.48% female 48.52% income quartiles $0 to $10,000 25.59% $10,001 to $35,000 24.87% $35,001 to $64,000 24.71% $64,001 and above 24.83% net worth quartiles $0 to $2,000 25.48% $2,001 to $67,000 24.54% $67,001 to $270,750 24.98% $290,751 and above 25.00% education college degree 28.11% less than college 71.89% marital status married 62.12% not married 37.88% 214 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 table 2 shows the cross tabulations of the sample by level of financial planning. it is important to note that across all categories most individuals choose the no plan level of financial planning. we see that among those with low levels of financial literacy, about 70% choose the no plan level of financial planning. this drops to about 50% for those with high levels of financial literacy. as financial literacy levels increase, there is a drop in the chance of respondents using the no plan level financial planning. about 24% of those with high financial literacy use the comprehensive level of financial planning, about 8% use plan only, about 19% use retire only, and about 50% use no plan. about 14% of hispanics use the comprehensive level of financial planning, while about 63% use no plan. about 12% of blacks use the comprehensive level of financial planning, while about 66% use no plan. about 23% of non-black non-hispanics use the comprehensive level of financial planning, compared with about 52% who use no plan. about 17% of males and about 19% of females use the comprehensive level of financial planning. about 58% of both males and females use the no plan level of financial planning. table 2 cross tabulations of sample by level of financial planning (n = 9,432) variables level of financial planning comprehensive plan only retire only no plan financial literacy low 8.98% 6.41% 14.43% 70.18% moderate 13.35% 5.63% 16.62% 64.40% high 23.66% 7.75% 18.53% 50.06% race/ethnicity hispanic 13.61% 5.92% 17.58% 62.88% black 12.04% 5.87% 15.89% 66.20% non-black non-hispanic 22.87% 7.67% 17.98% 51.48% sex male 16.95% 6.55% 18.22% 58.28% female 18.90% 7.08% 16.24% 57.78% income quartiles $0 to $10,000 13.30% 4.72% 12.60% 69.37% $10,001 to $35,000 9.93% 5.67% 14.49% 69.91% $35,001 to $64,000 17.55% 8.54% 20.29% 53.63% $64,001 and above 30.96% 8.37% 21.82% 38.86% net worth quartiles $0 to $2,000 8.11% 4.54% 13.20% 73.82% $2,001 to $67,000 9.29% 5.01% 14.95% 70.76% $67,001 to $270,750 17.36% 7.72% 19.10% 55.81% $290,751 and above 36.85% 9.97% 21.50% 31.68% education college degree 33.42% 10.00% 19.16% 37.42% less than college 11.83% 5.56% 16.52% 66.10% age 47–50 15.05% 5.58% 15.46% 63.91% 51–55 18.21% 7.30% 17.59% 56.90% 56+ 22.54% 6.86% 19.48% 51.11% marital status married 20.74% 7.15% 18.36% 53.75% not married 13.24% 6.24% 15.45% 65.07% l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 215 as income increases, we see an increase in the proportions of individuals who use the comprehensive level of financial planning except for the lowest quartile of income. as income increases, we see a reduction in the proportion of individuals who use the no plan level of financial planning. about 69% of those in the lowest income bracket use the no plan level of financial planning, compared with about 39% in the highest income bracket. we see a similar trend with net worth with those in the lowest bracket choose the no plan level of financial planning. as net worth increases, we see a significant drop in the proportion of the sample that use the no plan level of financial planning. about 74% of those in the lowest wealth bracket use the no plan level of financial planning, while about 32% of those in the highest wealth bracket do so. of individuals with a college degree, about 33% use the comprehensive level of financial planning, while about 37% use no plan. as age increased the proportion of individuals who use the comprehensive level of financial planning increases. about 64% of individuals aged 47-50 use no plan and this falls to about 51% around age 65+. about 21% of individuals use the comprehensive level of financial planning compared with about 13% of individuals who are not married. about 54% of those who are married use no plan, compared with about 65% who are not married. table 3 shows the summary statistics of the three financial literacy questions. although about 75% of the full sample got the interest question correct, only about 33% of individuals with low financial literacy and about 57% with moderate literacy got it correct. about 82% of the full sample got the inflation question correct. about 36% of those with low literacy and about 76% of those with moderate literacy were able to answer correctly. the risk question seemed to be the most difficult of the three to get correct. only about 23% of the full sample got this question correct, compared with about 82% who got the interest question correct and about 75% who got the interest question correct. 9.2. multivariate analysis table 4 shows the marginal effects of the multinomial probit regression of the independent variables on the level of financial planning. each marginal effect estimates the table 3 summary statistics for three financial literacy questions (n = 9,432) full sample low literacy moderate literacy high literacy interest more than $102 (correct response) 75.08% 33.28% 56.77% 100.00% exactly $102 14.40% 39.62% 24.55% 0.00% less than $102 10.52% 27.10% 18.68% 0.00% inflation more than today 8.02% 28.13% 11.20% 0.00% exactly the same 9.78% 35.94% 12.99% 0.00% less than today (correct response) 82.21% 35.94% 75.81% 100.00% risk true 77.18% 21.65% 67.43% 0.00% false (correct response) 22.82% 78.35% 32.57% 100.00% 216 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 table 4 marginal effects of independent variables on the level of financial planning (n = 9,432) group marginal effect standard error significance comprehensive literacy (none) moderate 0.0253 (0.0344) high 0.0697 (0.0346) ** race/ethnicity (hispanic) black 0.0156 (0.0116) non-black non-hispanic 0.0204 (0.0104) ** gender (male) female 0.0405 (0.0078) *** income quartile (1st) 2 �0.0329 (0.0107) *** 3 0.0108 (0.0111) 4 0.0448 (0.0122) *** net worth quartile (1st) 2 0.0130 (0.0095) 3 0.0709 (0.0106) *** 4 0.1925 (0.0133) *** education (no degree) college degree 0.1038 (0.0098) *** age (47–50) 51–55 0.0288 (0.0083) *** 56+ 0.0710 (0.0136) *** status (not married) married 0.0000 (0.0083) plan only literacy (none) moderate 0.0084 (0.0238) high 0.0136 (0.0239) race/ethnicity (hispanic) black 0.0058 (0.0081) non-black non-hispanic 0.0024 (0.0073) gender (male) female 0.0069 (0.0054) income quartile (1st) 2 0.0104 (0.0074) 3 0.0280 (0.0078) *** 4 0.0107 (0.0079) net worth quartile (1st) 2 0.0028 (0.0068) 3 0.0241 (0.0076) *** 4 0.0451 (0.0092) *** education (no degree) college degree 0.0293 (0.0069) *** age (47–50) 51–55 0.0163 (0.0057) *** 56+ 0.0125 (0.0090) status (not married) married �0.0076 (0.0058) (continued on next page) l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 217 probability of using a level of financial planning given a change in one independent variable, holding all others constant. the base category is no plan and all categories will be compared with it. 9.3. comprehensive level of financial planning we found that individuals with high financial literacy were more likely than those with low literacy to use comprehensive financial planning over no plan. this is consistent with our hypothesis that increased financial literacy is associated with using the comprehensive level of financial planning. non-black non-hispanic individuals were more likely than hispanics to use comprehensive financial planning over no plan. our results indicate that females were more likely than males to use comprehensive financial planning over no plan. individuals in the second quartile of income were less likely than those in the first quartile to use comprehensive over no plan. as individuals move up to the top quartile of income they are more likely than those in the first quartile to use comprehensive over no plan. individuals in the third and fourth quartiles of wealth are more likely than those in the first quartile to use comprehensive over no plan. we also found that individuals with a college degree were more likely than those with no degree to use comprehensive over no plan. this result is consistent with the hypothesis as table 4 (continued) group marginal effect standard error significance retire only literacy (none) moderate 0.0419 (0.0327) high 0.0466 (0.0329) race/ethnicity (hispanic) black �0.0015 (0.0122) non-black non-hispanic �0.0155 (0.0111) gender (male) female �0.0055 (0.0081) income quartile (1st) 2 0.0187 (0.0105) * 3 0.0638 (0.0113) *** 4 0.0686 (0.0126) *** net worth quartile (1st) 2 0.0049 (0.0109) 3 0.0323 (0.0117) *** 4 0.0599 (0.0136) *** education (no degree) college degree �0.0036 (0.0096) age (47–50) 51–55 0.0209 (0.0088) ** 56+ 0.0409 (0.0141) *** status (not married) married 0.0061 (0.0085) – no plan (base outcome) note. denotes significance at the following levels: *p < .1, **p < .05, ***p < .01. 218 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 financial literacy and education tend to be correlated. finally, we found that older individuals were more likely than younger ones to use comprehensive over no plan. 9.4. plan only level of financial planning we found that individuals in the third income quartile were more likely than those in the first quartile to use plan only over no plan. we also found that individuals in the third and fourth wealth quartiles were more lkeily than those in the first quartile to use plan only over no plan. individuals with a college degree were more likely than those with no degree to use plan only over no plan. finally, we found that individuals in the accumulation stage (age 51–55) were more likely than those at the acquisition stage (age 47–50) to use plan only over no plan. 9.5. retire only level of financial planning we found that individuals in the second, third, and fourth income quartiles were more likely than those in the first quartile to use retire only over no plan. these results are consistent with the life cycle theory which indicates these individuals are aware of the need for consumption smoothing. we also found that individuals in the third and fourth wealth quartiles were more likely than those in the first quartile to use retire only over no plan. individuals at the accumulation and decumlation stages were more likely than those at the acquisition stage to use retire only over no plan. 10. discussion the current study examines the association between financial literacy on the level of financial planning individuals use. we found that individuals with high levels of financial literacy were more likely to engage in the comprehensive level of financial planning than those with no literacy. we also found that as financial literacy increases individuals were less likely to use no plan than those with no financial literacy. these results are consistent with our hypothesis that financial literacy is related to the level of financial planning individuals use. human capital theory would indicate that more educated individuals may also invest in either their own financial literacy or renting someone else’s human capital. by choosing the the comprehensive level of financial planning they are combining their own human capital with that of a professional to obtain the best decision-making set. our results indicate that there may be some value in providing broad-based financial education that is tilted towards the value of planning and how to select a quality advisor where the costs of human capital acquisition are high.this may be especially valuable for those that do not want to make heavy investments in the human capital required to make quality decision in complex domains. this type of education could make the delegation decision more efficient, while also increasing the demand for quality professional advice. while we note that increased financial literacy is associated with comprehensive financial planning, we also note that respondents have seen the questions in prior surveys. we may be observing a bit of the learning effect here if respondents have simply remembered the l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 219 questions. this is a limitation of the current study. future research may want to assess the level of financial planning with different literacy questions or with a different sample. the respondents in this survey have seen the questions in prior years and the responses may not be capturing true financial literacy but rather memorization of the answers. appendix a: financial literacy questions 1. suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow: more than $102, exactly $102, less than $102? 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, would you be able to buy more than, exactly the same as, or less than today with the money in this account? 3. do you think that the following statement is true or false? “buying a single company stock usually provides a safer return than a stock mutual fund.” appendix b: levels of financial planning questions 1. have you [or] [spouse/partner’s name]: “. . . ever calculated how much retirement income you would need at retirement?” 2. have you [or] [spouse/partner’s name]: “. . . consulted a financial planner about how to plan your finances after retirement?” appendix c: correlation matrix of level of financial planning data group literacy12 race sex incqt netqt edlevel ages1 marstatus group 1 literacy12 �0.1787 1 race �0.1306 0.1563 1 sex �0.0213 �0.0987 �0.0191 1 incqt �0.2341 0.2142 0.1504 �0.2254 1 netqt �0.3372 0.2265 0.2732 �0.0443 0.4175 1 edlevel �0.2962 0.242 0.1792 0.0353 0.3256 0.3177 1 ages1 �0.075 0.0151 �0.0087 0.0145 �0.0075 �0.0082 0.0135 1 marstatus �0.1141 0.1045 0.1402 �0.0444 0.1784 0.3187 0.1295 0.0148 1 note. literacy12 = literacy questions; incqt = income quartiles; netqt = net worth quartiles; edlevel = education level; ages1 = age groups; marstatus = marital status. 220 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 references agnew, j. r., & szykman, l. r. (2005). asset allocation and information overload: the influence of information display, asset choice, and investor experience. journal of behavioral finance, 6, 57–70. https://doi.org/ 10.1207/s15427579jpfm0602_2 agnew, j. r., bateman, h., & thorp, s. (2013). financial literacy and retirement planning in australia. numeracy, 6(2), article 7. https://doi.org/10.5038/1936-4660.6.2.7 agnew, j., & szykman, l. (2011). annuities, financial literacy, and information overload. in o. s. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 158–178). oxford university press. bae, s. c., & sandager, j. p. (1997). what consumers look for in financial planners. association for financial counseling and planning education, 8, 9–16. barber, b. m., & odean, t. (2000). trading is hazardous to your wealth: the common stock investment performance of individual investors. the journal of finance, 55, 773–806. https://doi.org/10.1111/0022-1082.00226 becker, g. (1994). human capital revisited. journal of chemical information and modeling, 53, 1689–1699. beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2009). the importance of default options for retirement saving outcomes: evidence from the united states. in social security policy in a changing environment (pp. 167–195). national bureau of economic research, inc. blundell, r., dearden, l., meghir, c., & sianesi, b. (1999). human capital investment: the returns from education and the economy. fiscal studies, 20, 1–23. https://doi.org/10.1111/j.1475-5890.1999.tb00001.x butrica, b. a., goldwyn, j. h., & johnson, r. w. (2005). understanding expenditure patterns in retirement. working paper. center for retirement research at boston college. calcagno, r., & monticone, c. (2015). financial literacy and the demand for financial advice. journal of banking & finance, 50, 363–380. https://doi.org/10.1016/j.jbankfin.2014.03.013 campbell, j. y. (2006). household finance. the journal of finance, 61, 1553–1604. https://doi.org/10.1111/ j.1540-6261.2006.00883.x chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. https://doi.org/10.1016/s1057-0810(99)80006-7 clark, r. l., & d’ambrosio, m. (2003). financial education and retirement savings. social science research network. available at ssrn 390642. http://dx.doi.org/10.2139/ssrn.390642 cole, s., sampson, t., & zia, b. (2010). prices or knowledge? what drives demand for financial services in emerging markets? the journal of finance, 66, 193. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. de bassa scheresberg, c. (2013). financial literacy and financial behavior among young adults: evidence and implications financial literacy and financial behavior among young adults: evidence. numeracy, 6, article 5. https://doi.org/10.5038/1936-4660.6.2.5 delavande, a., rohwedder, s., & willis, r. (2008). preparation for retirement, financial literacy and cognitive resources. michigan retirement research center research paper no. 2008-190. available at https://ssrn.com/abstract=1337655 elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217–231. gokhale, j., & kotlikoff, l. j. (2003). who gets paid to save? tax policy and the economy, 17, 111–139. https://doi.org/10.1086/tpe.17.20140506 hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10(1), 36–62. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 106, 309–322. howlett, e., kees, j., & kemp, e. (2008). the role of self regulation, future orientation, and financial knowledge in long term financial decisions. journal of consumer affairs, 42, 223–242. https://doi.org/10.1111/j.17456606.2008.00106.x huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. https://doi.org/ 10.1111/j.1745-6606.2010.01170.x l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 221 jappelli, t., & padula, m. (2013). investment in financial literacy and saving decisions. journal of banking & finance, 37, 2779–2792. https://doi.org/10.1016/j.jbankfin.2013.03.019 kramer, m. m. (2012). financial literacy, cognitive ability, and financial advice-seeking. working paper, june 11. lusardi, a. (2008). financial literacy: an essential tool for informed consumer choice? national bureau of economic research (no. w1408). lusardi, a., & mitchell, o. s. (2005). financial literacy and planning: implications for retirement wellbeing. michigan retirement research center research paper, no. wp, 10. lusardi, a., & mitchell, o. s. (2006). financial literacy and planning: implications for wellbeing. pension research council. working paper, wp2006-01. meier, s., & sprenger, c. d. (2013). discounting financial literacy: time preferences and participation in financial education program. journal of economic behavior & organization, 95, 159–174. https://doi.org/10.1016/j. jebo.2012.02.024 schwab, c., iannicola, d., jr., beck, t., hira, t., byrant, j. h., parker, j., . . . dawson, c. (2008). president’s advisory council on financial literacy (pacfl). washington, dc: the department of the treasury. sekita, s., horioka, c. y., kajitani, s., kan, m., kureishi, w., mck, c. r., . . . yoshida, k. (2011). financial literacy and retirement planning in japan. journal of pension economics and finance, 10, 637–656. https:// doi.org/10.1017/s1474747211000527 sethi-iyengar, s., huberman, g., & jiang, g. (2004). how much choice is too much? contributions to 401(k) retirement plans. pension design and structure. new lessons from behavioral finance, 83, 84–87. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. https://doi.org/10.1016/j.jfineco.2011.03.006 van rooij, m. c. j., lusardi, a., & alessie, r. j. m. (2012). financial literacy, retirement planning, and household wealth. the economic journal, 122, 449–478. https://doi.org/10.1111/j.1468-0297.2012.02501.x vinet, l., & zhedanov, a. (2010). a “missing” family of classical orthogonal polynomials. quarterly journal of finance, 25(1), 6–22. https://doi.org/10.1088/1751-8113/44/8/085201 willis, l. e. (2008). against financial literacy education. iowa law review, 94, 8–10. 222 l. a. v. augustin and t. k. martin / financial services review 30 (2022) 205–222 academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar-asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2020 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 28, no. 4, 2020 editor stuart michelson, stetson university special issue co-editor terrance martin, utah valley university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2020 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. pii: 1057-0810(91)90002-g from the editor many people date the origins of modern finance theory to the publication of harry markowitz’s 1952 article, “portfolio selection,” in the journal of finance. it is therefore appropriate to begin our new journal with an article by that same author who shared the 1990 nobel prize in economics for his work. professor markowitz seeks to apply portfolio selection theory to the individual investor. he reveals that his original work was focused on an investment company, a much simpler entity than an individual or family. in dealing with individual financial management, one must recognize the inherent illiquidity of important assets, such as the home and its furnishings. furthermore, given the random nature of events that might affect a family, a single plan must give way to adaptive decision rules. professor markowitz concludes by suggesting ways in which the “game of life” may be simulated, and encourages the financial services review to “publish research with various approaches to various aspects of its topic area.” he concludes by saying that “such pluralism is desirable in research, as it is in politics and the marketplace.” in starting this, the first scholarly journal to focus on individual financial management, the academy of financial services formalizes the creation of a new subdiscipline of finance. however, unlike other subdisciplines-such as investments, banking, real estate, and insurance-the scope of individual financial management focuses not on instruments and institutions but on the application and effect of all aspects of finance on the individual. the “pluralism” described by professor markowitz encompasses both theoretical and empirical studies which seek to explain individual financial behavior and present ways in which the welfare of families and individuals may be improved. this inaugural issue reflects the pluralistic approach of both the journal and the new field, and serves as a model for subsequent issues. the second article, by professors tom potts and william reichenstein, asks the rather practical question whether it is better to tax shelter assets designated for retirement in a pension account or maintain them outside of such a shelter. surprisingly, the answer depends on whether the owner of those assets intends to manage the nonpension portion actively or passively. y vi financial services review, l(1) 1991 a somewhat related issue is considered empirically in the next article, by professors robert kleiman and anandi sahu, in which they ask whether individual investors are better off putting their funds in equity portfolios of life insurance companies or in mutual funds, and conclude that there is very little difference in risk-adjusted performance. this implies that variable annuity contracts offered by life insurance companies may offer an advantage over mutual funds to some investors by deferring taxes. the fourth article, by professor neil murphy, examines household financial behavior in the payments area. his article uses the survey of currency and transaction account usage conducted by the federal reserve system to find out why increased use of automated teller machines has not reduced the number of checks written by consumers. he concludes that checks are written for necessary third party transactions, while atms are merely a convenient way to obtain cash from one’s bank. therefore, the only effective method of reducing the number of checks written by consumers is the imposition of a marginal cost for each check written. next, professor peter chinloy examines the extent to which real estate income influences the location decisions of individuals. since real estate constitutes more than 50% of the personal (non-human) wealth of americans and since, as professor markowitz points out (above), it is rather illiquid, expected returns to the real estate component of the portfolio should be an important component of job relocation decisions. his empirical results support this assertion. the last article is a review piece by professor larry cox that raises some major questions relating to disability insurance and assesses the adequacy with which each is addressed. as we develop the field of individual financial management, it is important to review what has already been done and what still needs to be done. for that reason, we plan to run a detailed review article as a regular feature, and encourage scholars to contact us with subjects. given the diverse nature of the fields which constitute individual financial management, it is difficult to stay abreast of developments which might affect it. to help readers with this problem, our review editor, professor phyllis myers, presents abstracts of articles relating to individual financial management that have appeared in other journals. this, also, will be a regular, ongoing feature of the financial services review. -lewis mandell exploring differences in african-americans’ financial well-being based on financial security factors c. w. copelanda,*, john h. younga, crystal r. hudsona adepartment of finance, clark-atlanta university, 223 james p. brawley drive, atlanta, ga 30314, usa abstract what impacts the financial well-being of african americans, compared with other ethnic groups, has been a mystery beyond basic socio-economic factors. however, when explored through the lens of homeownership and employment, two variables that have been latent due to historical racism, african americans fare far worse than other ethnic groups. this study utilized data from the 2016 national financial well-being survey (nfwbs) including the cfp financial well-being scale, and specifically targeted middle-income african americans. researchers found that when efforts are made to pull themselves up by their bootstraps through long-term savings, investing, and education, african americans only show statistical significance if they are middle-income because student loans tend to create a drag on financial well-being levels. © 2022 academy of financial services. all rights reserved. jel classifications: i310 (general welfare; well-being) keywords: financial well-being; general welfare; living standards; middle class; middle income; quality of life; standard of living; stress; homeownership; employment; savings 1. introduction many african americans enjoy a middle-class lifestyle, typically a result of a college education. a college education is a means to improve people’s professional status and economic stability (blalock, 2017). nevertheless, is this middle-class lifestyle built on wealth or simply income? in other words, if these african americans lost their income, would they be able to maintain this lifestyle and live on the wealth that they have saved? otherwise, what *corresponding author: tel.: +1-404-913-1413, fax: +1-404-880-8458. e-mail address: cwcopeland2@cau.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 321–336 factors prevent them from experiencing this financial well-being? according to mcintosh et al. (2020), the benchmark of economic security includes employment, homeownership, savings, retirement security, and financial literacy. this present study explores whether middleclass african americans experience financial security or financial well-being compared with middle-class white americans. african americans do not generate as much wealth as the general u.s. population, and the black-white wealth gap is widening (kochhar, 2014). in 2013, white households had 13 times the wealth of african american households. this wealth gap had increased since 2010 when white households’ wealth was only 10 times that of african american households (kochhar, 2014). it is more alarming that african americans have not generated enough wealth and have been unable to pass on that wealth to their children and grandchildren, which leaves them at an economic disadvantage (pfeffer & killewald, 2018). this transfer of wealth could be in the form of paying for their children’s college tuition or an outright lump sum of cash (pfeffer & killewald, 2018). 2. financial well-being and wealth financial behaviors, financial stressor events, and individual characteristics are functions of financial well-being. financial well-being is also an outcome of financial behaviors (kim et al., 2003). whether we as researchers address financial well-being at the household or the individual level, it is apparent that the factors mentioned above affect the financial wellbeing of african americans more than other racial or ethnic groups (irrespective of income). the distinction may not be as prevalent if individual characteristics across groups are accounted for; the issue is well recognized when we reduce the number of financial behaviors and financial stressor events being addressed (as associated issues). one factor identified as an essential determinant of individual success in saving money is the ability to delay gratification and exercise self-control. although economists like to assume that homo economicus can postpone short-term gratification for the sake of long-term need, prior research has proven the opposite (brounen et al., 2016). brounen et al. (2016) further support the premise with the empirical proof that these time preferences (short-term gratification/long-term need) can be partly transferred from one generation to the next. financial discipline is—at least partly—the result of how parents raise their children. in other words, financial education and financial upbringing may well be two routes to the same destination—taking individual financial responsibility later in life. the premise of wahla and colleagues’ (wahla et al., 2019) study is that human beings do not act rationally. they tend to be involved in decisions based on heuristics and mental shortcuts. they could be frame-dependent due to their restricted ability to absorb excessive information in complex learning environments. the irrational approach of individuals could lead them to exercise negative investment behaviors. negative investment behaviors by individuals could also adversely influence their financial well-being. the obvious conclusion is that poor financial behaviors can lead to vulnerability due to a lack of resources in an emergency (also classified as a financial stressor event). according to mello (2018), first, 322 c. w. copeland et al. / financial services review 30 (2022) 321–336 consistent with previous research regarding the literature demonstrating widespread financial fragility among u.s. households, the results suggest that many households are underinsured against even small financial shocks. when faced with a traffic fine of less than $200, individuals accrue collections and delinquencies on their credit reports, suggesting their inability to cover an unexpected expense. second, individuals exhibiting minimal distress at the baseline are mainly unaffected by nuisance fines, while those already facing several unpaid bills experience the most significant declines in financial well-being (mello, 2018). 3. literature review 3.1. discrimination in employment and homeownership although it is easy to point the finger at the individual, numerous external forces play a pivotal role in african americans’ low level of financial well-being. although the order hierarchy is random, the severity is felt across these external forces, for example, the lack of housing equity and low employment prospects resulting from discrimination against african americans. some of the best evidence regarding the persistence of employment discrimination comes from audit studies conducted by the urban institute (fix & struyk, 1993). in these studies, white applicants were favored over black applicants with identical qualifications 20% of the time. thus, negative racial stereotypes of african americans appear to play a role, both when individual employers evaluate potential applicants and when corporate decision-makers deliberate about the possible locations of employment facilities (fix & struyk, 1993). there is a long-documented history of employment discrimination in the united states; however, consumer-level discrimination can be equally damaging. if consumer discrimination exists, growth in consumer contact may help explain recent declines in blacks’ relative earnings and employment [to whites] (holzer & ihlanfeldt, 1998). federal laws have been implemented over the years to combat racial discrimination by employers; however, how is the battle won when the discriminating group (i.e., customers) cannot be subjected to legislation? the discrimination experienced by black workers due to the preferences of white customers is likely to have more negative effects on their wages and employment than any discrimination experienced by white workers due to black customers, as whites are more able to find employment in sectors without customer discrimination against them (holzer & ihlanfeldt, 1998). stable employment is at the foundation of positioning oneself for wealth accumulation. without it, the premise of homeownership and investment does not exist. nevertheless, even when stable employment is legislated, barriers to homeownership may still exist. because much of the wealth of most american families take the form of home equity, a substantial part of this race-based inequity in homeownership is linked to housing policies and institutional discrimination experienced in the past (oliver & shapiro, 2006). in an optimally functioning housing market, it could be assumed that each household selects the type of tenure that maximizes its utility. people who move frequently are highly risk-averse or dislike the responsibilities of homeownership would-be renters. people who desire to use small housing services rent because small housing units are not generally available for purchase. as c. w. copeland et al. / financial services review 30 (2022) 321–336 323 suggested by mcdonald (1974), if it is presumed that black and white americans do not differ in their risk aversion and taste for the responsibilities of homeownership, some urban black households rent but would own if they faced the same housing market whites face. the white housing submarket is characterized by a higher relative number of single-family houses and less difficulty obtaining mortgage credit (mcdonald, 1974). despite u.s. policies to increase minority homeownership (or because of them), the housing market and foreclosure debacle, fueled by lending discrimination, has further exacerbated disparities in homeownership between caucasians and minorities. in some cases, homeownership rates are worse than those that existed nearly 25 years ago (williams, 2015). for these groups, the american dream has become too elusive. furthermore, lending discrimination has trajectorial effects on entire communities (williams, 2015). because of the operation of these large-scale societal processes, indicators of socioeconomic status are not equivalent across racial groups; this is true at the community, the household, and the individual levels. because of residential segregation, black and white neighborhoods dramatically differ in the availability of jobs, family structure, opportunities for marriage, educational quality, and exposure to conventional role models. they also differ in their quality of life and access to resources and amenities that sustain health (williams, 1999). these two external forces alone have had a tremendous impact on african americans’ financial wellbeing. unfortunately, they have compounded the issue, leaving room for many external forces to thrive. 3.2. savings and investing in the stock market african americans do not invest in the stock market as much as white americans. furthermore, this could contribute to african americans’ lack of accumulated wealth and the fact that they do not experience as much financial well-being as white americans (herring & henderson, 2016; kochhar, 2014). moreover, african americans do not save or replenish their savings as much as white americans, which could again be a factor contributing to african americans’ limited wealth and financial well-being (kochhar, 2014). herring and henderson (2016) explored possible reasons for the black-white wealth gap and conducted ordinary least squares analysis and quantile regression analysis of the survey of consumer finances data. they found that african americans had significant income, stock ownership, and business ownership disadvantages. the researchers also found that african americans received lower returns on education, stock ownership, and business ownership than white americans (herring & henderson, 2016). similarly, gutter and copur (2011) examined the relationship between financial behaviors and financial wellbeing. the researchers used data obtained from 15,797 college students for this study and found that budgeting, saving, risky credit card behavior, and compulsive buying behavior were significantly related to financial well-being (gutter & copur, 2011). stromback et al. (2017) investigated what psychological characteristics influenced individuals’ positive financial behavior and financial well-being. the authors surveyed 2,063 participants from the swedish population. they found that individuals with good self-control were more likely to save money from every paycheck, had better general financial behaviors, 324 c. w. copeland et al. / financial services review 30 (2022) 321–336 felt less anxious about financial matters, and felt more secure in their current and future financial situations (stromback et al., 2017). sivaramakrishnan and srivastava (2019) wanted to understand the influence of risk avoidance and financial well-being on investing in equity products. sivaramakrishnan and srivastava’s (2019) research team chose urban, retail, and middle-class investors from four cities in india. they found that financial well-being or the feeling of financial security did not embolden individuals to invest in the stock market. instead, it proved to deter individuals from participating in the stock market (sivaramakrishnan & srivastava, 2019). moreover, sabri et al. (2020) investigated the relations among financial management, savings, investment behavior, and financial well-being. they surveyed 722 working women in the malaysian public sector and used a multistage random sampling method. according to their study, 39.2% of the women indicated that their assets were more than their debts, and 44.3% said their salary was sufficient to meet their basic requirements (sabri et al., 2020). moreover, more than 80% practiced good financial management behaviors. the researchers found that the malaysian working women had good financial management practices, which were indicated by their ability to engage in savings and investment behavior to manage their surplus money wisely to achieve a higher level of financial wellbeing (sabri et al., 2020). 3.3. financial literacy and education evidence also shows a strong relationship between financial knowledge and the likelihood of engaging in desirable financial practices: paying bills on time, tracking expenses, budgeting, paying credit card bills in full each month, saving out of each paycheck, maintaining an emergency fund, diversifying investments, and setting financial goals (hilgert et al., 2003). low financial literacy (an outcome of inadequate financial knowledge) is associated with poor financial decisions in equity investment, debt financing, as well as long-term retirement planning, and these decisions can lead to decrease in welfare (chu et al., 2017). according to chu et al. (2017), households with lower levels of financial literacy can also make suboptimal decisions when choosing loans or mortgages, as well as suffer from problems, such as debt accumulation, bankruptcy, and foreclosure. suboptimal decisions can clearly lead to not only lower amounts of wealth but perhaps even its absence. people’s attitudes toward money rely on different variables, such as individuals’ adolescence experiences, education, and economic and societal status. depending on these variables, the attitudes toward money differ from person to person (qamar et al., 2016). studies on financial issues reveal that an individual’s attitudes toward money play a significant role in deciding one’s financial management and level of financial well-being (shim et al., 2010). based on the depth and consistency of research in this area, this leads to the conclusion that responsible financial behavior is strongly related to strong financial knowledge (zakaria et al., 2012). qamar et al. (2016) found that individual financial efficacy had a strong positive association with financial well-being. they tested this hypothesis: “there is a relationship between c. w. copeland et al. / financial services review 30 (2022) 321–336 325 financial efficacy and financial well-being” (1458). the conclusions from this research were in favor of the hypothesis. shim et al. (2010) found that financial literacy alone was insufficient to guarantee control over individual finances; financial self-efficacy was similarly important. what is the difference between financial literacy and financial efficacy? while researchers have a good grasp of financial literacy, financial efficacy is a much less publicized concept. the most common definition of financial efficacy is a person’s perceived capability to control one’s personal finances (lapp, 2010; postmus, 2011). with this understanding, it makes sense to move forward with the attitude that financial literacy and financial efficacy must work in tandem to create a positive outcome; however, before closing the door on this discussion, a few other terms used in industry discussions can be explored. as noted by brounen et al. (2016), financial knowledge involves understanding key financial terms and ideas needed to function day by day in society. they also state that the terms financial literacy, financial knowledge, and financial education have regularly been used interchangeably in both academic literature and the mainstream media. financial education facilitates literacy, that is, mastery of finance-related knowledge and expertise, which are essential in undertaking daily transactions and wealth accumulation investments. it empowers people to manage their own finances and provide long-lasting financial security for themselves and their families (sundarasen et al., 2016). based on the latter two cited studies’ findings, the respondents who are financially interested, keep a tight household administration, have a strong locus of control, and have a positive economic outlook are all more prone to postpone immediate consumption for the sake of future needs. households that save money share a certain set of personality variables (brounen et al., 2016). with theoretical and empirical support, the results of sundarasen and colleagues’ (sundarasen et al., 2016) study indicate that parental norms, socialization proxies, and financial literacy play a significant role in money management. 3.4. conceptual framework and theory from the start, the life-cycle theory (see figure 1) reports that individuals consume resources in various amounts throughout their lives based on income and family dynamics. in light of this theory, we can assume that individuals emphasize consumption to provide utility, leading to financial well-being. variables such as financial knowledge and employment seem to have an equal influence on financial behaviors and financial well-being. additionally, financial behaviors such as home acquisition and savings are an expected part of the life-cycle and should positively influence financial well-being. based on the life cycle theory, the researchers have put forward the following hypothesis: 4. hypotheses h1: african americans have lower levels of financial well-being compared with white americans. h2: african americans who are employed full time are more likely to have higher levels of financial well-being than african americans who are not employed full time. 326 c. w. copeland et al. / financial services review 30 (2022) 321–336 h3: african americans who own their homes are more likely to have higher levels of financial well-being than african americans who do not own their homes. h4: african americans who are financially knowledgeable are more likely to have higher levels of financial well-being than african americans who are not financially knowledgeable. h5: african americans with savings greater than $5,000 and with investments are more likely to have higher levels of financial well-being than african americans who have savings of less than $5,000 and have no investments. 5. methods 5.1. data and sample selection this study explores the financial security of african american households with an income of $40,000 or more, based on the u.s. census’ median african american income of $45,438 in 2019 (semega et al., 2020). the study’s target population consists of households belonging to the income baseline of $40,000 or greater, representing 68% of the sample set. the study sample comprised 4,339 respondents from the income range of $40,000 to $150,000 plus, divided into three ethnic groups. of the 4,339 households, 3,070 (70.7%) identified themselves as white americans, 407 (9.4%) as african americans, and 862 (19.9%) as other americans. this study used the 2016 national financial well-being survey (nfwbs) that the consumer financial protection bureau (cfpb) administers and releases. the nfwbs was designed to measure the levels and distribution of financial well-being among the u.s. adult population and to provide household characteristics, income levels, and employment characteristics, financial experiences, and financial behaviors, skills, and fig. 1. this model was produced by pettinger (2019), the life-cycle hypothesis. from economicshelp.org (https://www.economicshelp.org/blog/27080/concepts/life-cycle-hypothesis/). c. w. copeland et al. / financial services review 30 (2022) 321–336 327 attitudes. both descriptive and multivariate results were weighted according to the nfwbs study weights to represent the u.s. population. table 1 exhibits the summary statistics of the dataset. 6. measurement of variables [database] 6.1. dependent variables a cfpb financial well-being scale score is a standardized number between 0 and 100 that represents the respondent’s underlying level of financial well-being. a higher score indicates a higher level of measured financial well-being, but there is no specific cutoff score for table 1 descriptive statistics of the weighted sample overall sample (n = 4,339) variables white americans (n = 3,070) african americans (n = 407) other americans (n = 852) observations % observations % observations % financial well-being (fwb) low fwb 753 24.5 123 30.2 277 32.2 average fwb 1,030 33.6 142 34.9 336 39.0 high fwb 1,287 41.9 142 34.9 249 28.8 household income $40,000–49,999 268 8.7 55 13.5 107 12.4 $50,000–59,999 318 10.4 67 16.5 96 11.1 $60,000–74,999 425 13.8 51 12.6 118 13.7 $75,000–99,999 591 19.2 80 19.7 179 20.8 $100,000–149,999 755 24.6 79 19.5 183 21.2 $150,000 or more 713 23.3 75 18.2 179 20.8 gender male 1,505 49.0 164 40.3 448 52.0 female 1,565 51.0 243 59.7 414 48.0 age group 34 and under 862 28.1 133 35.5 382 44.3 35–54 1,071 34.8 161 39.7 310 36.0 55–69 767 25.0 83 20.4 134 15.5 70 and over 370 12.1 30 7.4 36 4.2 marital status unmarried 838 27.3 172 42.3 323 37.5 married 2,232 72.7 235 57.7 539 62.5 education level some college or less 1,772 57.7 230 56.5 541 62.8 college degree 799 26.0 107 26.3 215 24.9 post-graduate degree 499 16.3 70 17.2 106 12.3 employment status self-employed 193 6.3 21 5.2 74 8.6 full-time 1,531 49.9 242 59.4 432 50.1 retired 617 20.1 59 14.5 77 8.9 other 729 23.7 85 20.9 279 32.4 328 c. w. copeland et al. / financial services review 30 (2022) 321–336 a “good” or a “poor” score. the respondents were classified into three categories, namely low financial well-being (0–49.9), average financial well-being (50–60.9), and high financial well-being (61–100); (consumer financial protection bureau, 2015, 2017). in this study, financial well-being is the dependent variable. it was converted into an ordinal qualitative variable using the visual binning technique to identify appropriate cutoff points to break the variables into three approximately equal groups. equal percentiles were used based on the scanned cases (equal intervals with two cutoff points). thus, this study considered the dependent variable a multinomial ordinal-dependent variable (1 = low, 2 = average, and 3 = high; lobos et al., 2016; nielsen, 2015). 6.2. independent variables the primary focus of this research was homeownership. the dataset comprised four housing groups: owners, renters, neither, and refused to answer. the respondents were coded as 1 if they were homeowners and 0 otherwise. four demographic characteristics were included as control variables in the analysis. age group, education level, gender, and employment status were coded as several binary variables: 34 and under, 35–54, 55–69, and 70 and over; some college or less, college degree, and post-graduate degree; self-employed, full-time, retired, and not employed, respectively. the 70 and over variable was the reference group for the age group; the post-graduate degree variable was the reference group for education level, and the not employed variable was the reference group for employment status. gender was dichotomized and defined as male = 1 and female = 0. the financial security variables were defined to reflect the conceptual framework. household savings was based on a single question, with the responses measured on a scale from 1 ($0 saved) to 7 ($75,000 or more). student loan (the subjectively assessed probably have variable) was based on the responses yes = 1 and no = 0. non-retirement investments (the subjectively assessed probably have variable) were based on a question with the responses yes = 1 and no = 0. retirement investments (the subjectively assessed probably have variable) were based on a question with the responses yes = 1 and no = 0. the financial knowledge scale (the objectively assessed financial knowledge) was based on the number of correct answers to the nine questions formulated by houts and knoll (2020). 6.3. model specifications a multinomial logistic regression model was used in this study. the financial well-being level was assumed to be a function of financial security factors, including financial knowledge, household savings, homeownership, and investments, as well as socioeconomic factors, such as age group, marital status, education level, and employment status. financial wellbeing ¼ f financial security factors and socioeconomic factorsð þ (1) c. w. copeland et al. / financial services review 30 (2022) 321–336 329 logit rð þ ¼ log r high well beingð þ r middle well beingð þ � � ¼ b 0 þ b 1x 1 þ b 2x 2 þ � � � þ b k xk (2) logit rð þ ¼ log r low well beingð þ r middle well beingð þ � � ¼ b 0 þ b 1x 1 þ b 2x 2 þ � � � þ b k xk , (3) where b denotes a vector of coefficients to be determined, and x represents a vector of african americans’ financial security factors and socioeconomic characteristics. 7. results 7.1. descriptive statistics table 1 provides descriptive statistics of the survey sample set. the sample comprised 70.8% white americans, 9.6% african americans, and 19.6% other americans. surprisingly, at the middle (average) financial well-being level, white americans had the lowest (33.6%) percentage compared with african americans (34.9%) and other americans (39%). income levels between $100,000 and $149,999 were the highest represented group, with about 25% white americans, 20% african americans, and 21% other americans. females dominated the sample set, with 51% white americans and 59.7% african americans, except for (48%) other american households. the same pattern emerged in the 35–54 age group, while the other americans (44.3%) had a higher share of the 34 and younger group. most of the respondents were married, comprising 72.7% white americans, 57.7% african americans, and 62.5% other americans. regarding education levels across all groups, having some college education was the most influential statistic. full-time employment was evenly spread across the groups: approximately 50% white americans, 60% african americans, and 50% other americans. table 2 presents the multinominal logistical estimates of the likelihood of obtaining a low score in financial well-being by ethnicity group. when comparing the average-level respondents with the low-level respondents, no significant differences were discovered, with white americans as the reference group. more interestingly, the african americans’ financial well-being level was significantly related to the high level among white americans compared with the low level among the reference families. however, when comparing the likelihood of obtaining a high score in financial well-being compared with a low score, african americans comprised the most unlikely group (47.6%, p < .001) at the high level compared with white americans. other americans comprised the next unlikely group (32.2%, p < .01) compared with white americans. therefore, ethnicity seems relevant in households classified as having a low level of financial well-being— african americans to a large extent compared with white americans. based on the results, hypothesis h1 is not rejected (table 2). 330 c. w. copeland et al. / financial services review 30 (2022) 321–336 7.2. multinomial logistic results the purpose of the second stage of the analysis was to isolate better the link between the financial well-being levels and financial security factors (e.g., full-time employment, homeownership, financial knowledge levels, savings rate, and investor status). the multinomial logistic regression analysis results are presented in table 3. 7.3. average-level versus low-level financial well-being the first model (table 3, low [column heading]) specifically investigated the differences between american households with low scores in financial well-being and those with average scores across ethnic groups (white americans, african americans, and other americans). an analysis of the education level showed that white americans with a low level of financial well-being and some college education or less were 42.7% (p < .05) more likely than post-graduate respondents to have low scores in financial literacy. surprisingly, the education level was not significant for african americans and other americans compared with post-graduate respondents. retired respondents were less likely to be classified as having a low level of financial well-being compared with respondents who had averagelevel scores. more specifically, white american retirees (52.5%, p < .01) and african americans (80.2%, p < .05) were less likely to be included in the group with a low level of financial well-being. other americans’ scores were not significant when compared with the average level scores. other americans (22.4%, p < .05) comprised the only group affected by the financial knowledge scale, indicating an unlikelihood of being classified in the lowlevel financial well-being group compared with the average-level group. financial knowledge levels were not significant for white americans and african americans. household savings of less than $5,000 appeared to be the most impactful factor for americans’ likelihood of being classified as belonging to the group with a low level of financial well-being. white americans (261.6%, p < .001), african americans (467.1%, p < .001), and other americans (122.2%, p < .001) were more likely to be included in the group with a low level of financial well-being. only african american respondents (91.3%, p < .05) table 2 multinominal logistical analysis of the likelihood of financial well-being levels among ethnic groups financial well-being levels (n = 4,339) average high b exp(b ) b exp(b ) intercept 0.314 0.536 ethnicity group: reference category = white americans african americans �0.121 0.886 �0.647*** 0.524 other americans �0.167 0.846 �0.388** 0.678 notes. the exponentiated coefficient minus one and times 100 gives the percentage increase or decrease due to a one-unit change in the independent variable. the reference category for the model is the low financial wellbeing group. **p < .01, ***p < .001. c. w. copeland et al. / financial services review 30 (2022) 321–336 331 t ab le 3 m u lt in o m in al lo g is ti ca l an al y si s o f th e li k el ih o o d o f fi n an ci al w el lb ei n g le v el s (fi n an ci al se cu ri ty v ar ia b le s) f in an ci al w el lb ei n g le v el s (n = 4 ,3 3 9 ) p ar am et er es ti m at es w h it e a m er ic an s (n = 3 ,0 7 0 ) a fr ic an a m er ic an s (n = 4 0 7 ) o th er a m er ic an s (n = 8 6 2 ) l o w h ig h l o w h ig h l o w h ig h b e x p (b ) b e x p (b ) b e x p (b ) b e x p (b ) b e x p (b ) b e x p (b ) in te rc ep t �1 .9 1 1 0 .6 9 7 �1 .1 2 4 2 .2 5 5 �2 .7 9 9 0 .7 5 7 a g e g ro u p : r ef er en ce ca te g o ry = 7 0 an d o v er 3 4 an d u n d er 0 .3 5 4 1 .4 2 5 �0 .0 9 0 0 .9 1 4 �0 .2 9 3 0 .7 4 6 �0 .4 0 5 0 .6 6 7 0 .8 9 8 2 .4 5 4 �1 .0 7 3 0 .3 4 2 3 5 – 5 4 0 .3 2 7 1 .3 8 7 �0 .5 5 9 * * 0 .5 7 2 �0 .5 8 6 0 .5 5 7 �0 .8 5 8 0 .4 2 4 0 .8 8 8 2 .4 3 0 �1 .0 0 1 0 .3 6 7 5 5 – 6 9 0 .2 5 8 1 .2 9 4 �0 .3 3 5 0 .7 1 5 �1 .0 3 0 0 .3 5 7 �0 .8 4 5 0 .4 3 0 1 .0 8 8 2 .9 7 0 �0 .6 5 6 0 .5 1 9 m ar it al st at u s: r ef er en ce ca te g o ry = u n m ar ri ed m ar ri ed �0 .0 3 7 0 .9 6 4 0 .0 5 4 1 .0 5 6 0 .3 9 2 1 .4 8 0 0 .3 1 3 1 .3 6 7 0 .2 0 1 1 .2 2 3 0 .0 0 7 1 .0 0 7 e d u ca ti o n le v el : r ef er en ce ca te g o ry = p o st -g ra d u at e d eg re e s o m e co ll eg e o r le ss 0 .3 5 6 * 1 .4 2 7 �0 .1 3 0 0 .8 7 8 �0 .2 8 1 0 .7 5 5 �0 .6 7 2 0 .5 1 1 0 .0 3 5 1 .0 3 6 0 .3 0 1 1 .3 5 2 c o ll eg e d eg re e 0 .2 0 0 1 .2 2 2 0 .0 0 0 1 .0 0 0 0 .1 3 9 1 .1 4 9 �0 .8 5 4 * 0 .4 2 6 0 .1 3 2 1 .1 4 1 �0 .3 2 4 0 .7 2 3 e m p lo y m en t ty p e: r ef er en ce ca te g o ry = n o t em p lo y ed s el fem p lo y ed 0 .1 9 6 1 .2 1 7 0 .3 3 5 1 .3 9 8 �1 .1 2 2 0 .3 2 6 �1 .3 9 3 * 0 .2 4 8 0 .1 2 9 1 .1 3 8 0 .1 2 6 1 .1 3 4 f u ll -t im e 0 .0 9 8 1 .1 0 3 0 .0 4 7 1 .0 4 8 �0 .5 7 9 0 .5 6 1 �0 .6 1 7 0 .5 3 9 0 .3 9 3 1 .4 8 1 0 .3 6 7 1 .4 4 4 r et ir ed �0 .7 4 4 * * 0 .4 7 5 0 .6 4 5 * * * 1 .9 0 7 �1 .6 2 1 * 0 .1 9 8 �0 .7 5 7 0 .4 6 9 �0 .0 0 1 0 .9 9 9 0 .4 4 4 1 .5 6 0 f in an ci al k n o w le d g e sc al e: r ef er en ce ca te g o ry = s ca le : �2 .0 to 1 .5 k n o w le d g e sc al e 0 .0 0 0 1 .0 0 0 0 .4 2 3 * * * 1 .5 2 7 �0 .3 0 2 0 .7 3 9 0 .1 0 2 1 .1 0 8 �0 .2 5 3 * 0 .7 7 6 0 .2 8 5 * 1 .3 2 9 h o u se h o ld sa v in g s: r ef er en ce ca te g o ry = $ 5 ,0 0 0 o r m o re $ 0 – 4 ,9 9 9 1 .2 8 5 * * * 3 .6 1 6 �0 .7 9 3 * * * 0 .4 5 3 1 .7 3 5 * * * 5 .6 7 1 �0 .5 2 6 0 .5 9 1 0 .7 9 8 * * * 2 .2 2 2 �0 .9 3 0 * * * 0 .3 9 4 h o m eo w n er sh ip st at u s: r ef er en ce ca te g o ry = h o m eo w n er n o n -h o m eo w n er 0 .0 8 5 1 .0 8 9 �0 .4 1 8 * * 0 .6 5 8 0 .6 4 9 * 1 .9 1 3 �0 .5 8 3 0 .5 5 8 0 .4 1 9 1 .5 2 1 �0 .0 5 4 0 .9 4 8 in v es tm en ts (r et ir em en t an d n o n -r et ir em en t) : r ef er en ce ca te g o ry = y es n o 0 .4 7 8 * * * 1 .6 1 3 �0 .2 7 3 * 0 .7 6 1 0 .1 6 8 1 .1 8 3 �0 .1 5 0 0 .8 6 0 0 .8 1 4 * * * 2 .2 5 8 0 .1 2 2 1 .1 2 9 n o te s. t h e ex p o n en ti at ed co ef fi ci en t m in u s o n e an d ti m es 1 0 0 g iv es th e p er ce n ta g e in cr ea se o r d ec re as e d u e to a o n eu n it ch an g e in th e in d ep en d en t v ar iab le . t h e re fe re n ce ca te g o ry fo r th e m o d el is th e av er ag e fi n an ci al w el lb ei n g g ro u p . * p < .0 5 , * * p < .0 1 , * * * p < .0 0 1 . 332 c. w. copeland et al. / financial services review 30 (2022) 321–336 who were not homeowners had a score significant enough to be included in the group with a low level of financial well-being compared with the average level. homeownership was not significant for white americans and other americans. not having an investment account (retirement and non-retirement) was significant for white americans (61.3%, p < .001) and other americans (125.8%, p < .001), indicating their membership in the group with a low level of financial well-being. surprisingly, not having an investment account was not significant for african americans. unexpectedly, the age group and marital status had no significant impact on the respondents’ financial well-being levels compared with the average-level and the low-level financial well-being. 7.4. average-level versus high-level financial well-being the second model (table 3, high [column heading]) in the multinomial logistic analysis estimated the likelihood that a person would belong to the average-level financial well-being group compared with the high-level financial well-being group. an analysis of americans’ age groups and financial well-being revealed that the relation between these variables was only significant for white americans with high-level financial well-being compared with the average group. white respondents in the 35–54 age group were 42.8% (p < .01) less likely to be classified in the high-level financial well-being category than the respondents aged 70 and older. the scores of african americans and other americans were not significant enough to be in the range of high-level financial well-being. conversely, when reviewing high-level financial well-being, education level played a significant role solely for african americans; college-graduate respondents were 57.4% (p < .05) less likely to be classified as having a high level of financial well-being than post-graduate respondents. as expected, white retired respondents were 90.7% (p < .001) more likely to be included in the high-level financial well-being group than their unemployed counterparts. however, selfemployed african american respondents were 75.2% (p < .05) less likely to be members of the said group than their unemployed counterparts. employment levels were not significant for the other americans. interestingly, the respondents with savings of less than $5,000 were less likely to belong to the high-level financial well-being group than the respondents with more than $5,000 in savings (white americans at 54.7%, other americans at 60.6%, and not significant for african americans). surprisingly, financial knowledge level (52.7% more likely, p < .001), homeownership (34.2% less likely, p < .01), and investment account status (23.9% less likely, p < .05) were significant solely for white americans in the high-level financial wellbeing group when compared with the reference groups. furthermore, marital status did not appear to significantly affect the respondents’ financial well-being at a high level. 8. discussion consistent with past studies (herring & henderson, 2016; kochhar, 2014), this study found that high-level financial well-being was unlikely associated with african americans. c. w. copeland et al. / financial services review 30 (2022) 321–336 333 regarding savings of $5,000 or greater, the findings suggest that compared with savers with less than $5,000 in savings, african americans had a stronger association with financial security and financial freedom of choice. this saving gap may be because african americans lack trust in the financial markets or the banking system, which delivers limited financial knowledge, such as interest on savings, inflation, bond prices, and risk diversification. limited financial knowledge coupled with non-homeownership occurs at the stage of life when african americans are working and are less likely to adhere to financial planning for long-term goals. compared with demographic factors (formal education level and employment status), each financial well-being level examined in this study (low, middle, or average, high) showed a relatively small association with objective african american financial wellbeing. this finding implies that understanding one’s financial situation and financial capability can address financial wellness deficiencies for all households, regardless of their levels of financial well-being. this study’s findings that financial well-being had a positive association with the selfreported savings level and the self-reported homeownership status underscore the importance of shaping trust in the financial sector’s impact on african americans’ upward financial mobility. interestingly, savings of $5,000 or greater had a tremendous positive impact on self-reported savings, suggesting that ethnicity affects whether an individual is exposed to stocks versus savings at a young age. such lack of exposure plays a fundamental role in african americans’ financial well-being gap. this finding corroborates past studies’ results showing that savings at the expense of investing have a critical influence on children’s lives as they grow up and continue to influence the wealth-building gap of their own children when they become adults (asli & elif, 2019; jorgensen & savla, 2010). 9. conclusion and implications this research found that african americans who had a post-graduate degree, owned a home, and had savings of over $5,000 were more likely to have a higher level of financial well-being than african americans who lacked these characteristics. therefore, most of the financial security variables in this study proved to have impacts on african americans’ financial well-being. the only financial security variables that had no effect were financial knowledge and investing in the stock market. the implications for financial institutions are enormous. african americans are willing to save because of the sense of financial security associated with that behavior. the more educated cohort of this ethnic group tends to save at higher rates and has more satisfaction from the activity. financial institutions have a chance to leverage multiple products to satisfy this cohort because there is an underlying desire to transition from “saving” to “investing.” the implications for policy-makers are also looming; as african americans become more financially literate, their capacity to become homeowners increases, which positively impacts the tax base of the municipalities where they reside. the implications for researchers intensify as we understand the differences in financial behavior of the various financial cohorts of african americans and their contrasts with other ethnic groups. 334 c. w. copeland et al. / financial services review 30 (2022) 321–336 references asli, e. a., & elif, a. s. (2019). an investigation of financial literacy, money ethics and time preferences among college students: a structural equation model. the international journal of bank marketing, 37, 880–900. https://www.emerald.com/insight/content/doi/10.1108/ijbm-05-2018-0120/full/html blalock, c. (2017). african americans experiences of managing college debt. minneapolis, mn: walden university proquest dissertation publishing. brounen, d., koedijk, k., & pownall, r. a. j. (2016). household financial planning and savings behavior. discussion paper no. 11033. london: center for economic policy research. chu, z., wang, z., xiao, j. j., & zhang, w. (2017). financial literacy, portfolio choice and financial well-being. social indicators research, 132, 799–820. https://doi.org/10.1007/s11205-016-1309-2 consumer financial protection bureau. (2015). measuring financial well-being. available at https://files. consumerfinance.gov/f/201512_cfpb_financial-well-being-user-guide-scale.pdf consumer financial protection bureau. (2017). cfpb financial well-being scale. available at https://files. consumerfinance.gov/f/documents/201705_cfpb_financial-well-being-scale-technical-report.pdf fix, m. e., & struyk, r. j. (1993). clear and convincing evidence: measurement of discrimination in america. washington, dc: urban institute press. https://webarchive.urban.org/publications/105136.html gutter, m., & copur, z. (2011). financial behaviors and financial wellbeing of college students: evidence from a national survey. journal of family and economic issues, 32, 699–714. https://doi.org/10.1007/s10834-011-9255-2 herring, c., & henderson, l. (2016). wealth inequality in black and white: cultural and structural sources of the racial wealth gap. race and social problems, 8, 4–17. https://doi.org/10.1007/s12552-016-9159-8 hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309–322. holzer, h., & ihlanfeldt, k. r. (1998). customer discrimination and employment outcomes for minority workers. the quarterly journal of economics, 113, 835–867. https://doi.org/10.1162/003355398555766 houts, c. r., & knoll, m. a. z. (2020). the financial knowledge scale: new analyses, findings, and development of a short form. journal of consumer affairs, 54, 775–800. https://doi.org/10.1111/joca.12288 jorgensen, b. l., & savla, j. (2010). financial literacy of young adults: the importance of parental socialization. family relations, 59, 465–478. https://doi.org/10.1111/j.1741-3729.2010.00616.x kim, j., garman, e. t., & sorhaindo, b. (2003). relationships among credit counseling clients’ financial wellbeing, financial behaviors, financial stressor events, and health. journal of financial counseling and planning, 14, 75–87. kochhar, r. (2014). 10 projections for the global population in 2050. washington, dc: pew research center. available at https://www.pewresearch.org/fact-tank/2014/02/03/10-projections-for-the-global-population-in2050/ lapp, w. m. (2010). behavioral models for driving prosperity for low-income people: earn’’s model of financial self-efficacy. san francisco, ca: earned assets resource network. available at https:// d1wqtxts1xzle7.cloudfront.net/43046252/5_-_behavioral_models_for_prosperity-_a_statistical_assessment_ of_savings_and_behavioral_change-1-libre.pdf?1456400832=&response-content-disposition=inline%3b +filename%3dmonth_year_title_of_paper_behavioral_mod.pdf&expires=1686743562&signature= gu0exrue1ef~h8zznjsfbexc-yeowcnkixznndvxnnj1wzo2grlejb7l0ykm5miwa-0yu8u-q0u9gueu qpjjt2qe7rj55nzuihgemcbpywi0zmcihgnxwsraxzcu1fgc-ucra5wicc4wuuqqcgwwk-bkv6-cj6wu58su42j65do2zahhdmrfruihuqlrlis~kgevro6wt~q9d8tbtrh6sgpboze4c0-ytzuwg7xbcy ifwcnqylgejyeqlkh7tjkuqyh~i0qrpi2six8ng25hgpt7lje4jb5omjfv3n7sn5nfjmr2j4~qlwxmshzu uggeessjqd3ghrs35eeqzfx6g__&key-pair-id=apkajlohf5ggslrbv4za lobos, g., grunert, k. g., bustamante, m., & schnettler, b. (2016). with health and good food, great life! gender differences and happiness in chilean rural older adults. social indicators research, 127, 865–885. https://doi.org/10.1007/s11205-015-0971-0 mcintosh, k., moss, e., nunn, r., & shambaugh, j. (2020). examining the black and white wealth gap. washington, dc: brookings. available at https://memphis.uli.org/wp-content/uploads/sites/49/2020/07/examiningthe-black-white-wealth-gap.pdf c. w. copeland et al. / financial services review 30 (2022) 321–336 335 mcdonald, j. f. (1974). housing market discrimination, homeownership, and savings behavior. the american economic review, 64, 225–229. mello, s. (2018). speed trap or poverty trap? fines, fees, and financial wellbeing. new york, ny: fines and fees justice center. available at https://finesandfeesjusticecenter.org/articles/speed-trap-or-poverty-trapfines-fees-and-financial-wellbeing/ nielsen, r. (2015). scf complex sample specification for stata. athens, ga: department of financial planning housing and consumer economics, university of georgia. oliver, m. l., & shapiro, t. m. (2006). black wealth/white wealth: a new perspective on racial inequality (2nd ed.). new york, ny: routledge. pettinger, t. (2019). life-cycle hypothesis. economics help. available at https://www.economicshelp.org/blog/ 27080/concepts/life-cycle-hypothesis/ pfeffer, f. t., & killewald, a. (2018). generations of advantage. multigenerational correlations in family wealth. social forces; a scientific medium of social study and interpretation, 96, 1411–1442. https://doi.org/ 10.1093/sf/sox086 postmus, j. l. (2011). understanding financial literacy with survivors of intimate partner violence. madison, wi: center for financial security at the university of wisconsin. available at http://www.cfs.wisc.edu/ briefs/postmus2011_understandingbrief.pdf qamar, m., nadeem khemta, m., & hassan, j. (2016). how knowledge and financial self-efficacy moderate the relationship between money attitudes and personal financial management behavior. european online journal of natural and social sciences, 5, 296–308. sabri, m. f., syahrul, r., & wijekoon, r. (2020). financial management, savings and investment behavior and financial well-being of working women in the public sector. majalah ilmiah bijak, 17, 135–153. https://doi. org/10.31334/bijak.v17i2.1008 semega, j., kollar, m., shrider, e. a., & creamer, j. (2020). household income: 2019 census.gov. income and poverty in the united states: 2019. available at https://www.census.gov/content/dam/census/library/ publications/2020/acs/acsbr20-03.pdf shim, s., barber, b. l., card, n. a., xiao, j. j., & serido, j. (2010). financial socialization of first-year college students: the roles of parents, work, and education. journal of youth and adolescence, 39, 1457–1470. https://doi.org/10.1007/s10964-009-9432-x sivaramakrishnan, s., & srivastava, m. (2019). financial well-being, risk avoidance and stock market participation. international journal of financial services management, 9, 326–344. https://doi.org/10.1504/ijfsm.2019.102456 stromback, c., lind, t., skagerlund, k., vastfjall, d., & tinghog, g. (2017). does self-control predict financial behavior and financial wellbeing? journal of behavioral and experimental finance, 14, 30–38. https://doi. org/10.1016/j.jbef.2017.04.002 sundarasen, s. d. d., rahman, m. s., othman, n. s., & danaraj, j. (2016). impact of financial literacy, financial socialization agents, and parental norms on money management. journal of business studies quarterly, 8, 140–156. https://www.researchgate.net/publication/315816479_impact_of_financial_literacy_financial_socialization_ agents_and_parental_norms_on_money_management wahla, k. r., aktar, h., & ali shah, s. z. (2019). framing effects and financial well-being: role of investment behaviors as mediator. review of economic and development studies, 5, 343–354. williams, a. p. (2015). lending discrimination, the foreclosure crisis and the perpetuation of racial and ethnic disparities in homeownership in the u.s.william & mary business literature review, 6, 601–661. williams, d. r. (1999). race, socioeconomic status, and health. the added effects of racism and discrimination. annals of the new york academy of sciences, 896, 173–188. https://doi.org/10.1111/j.1749-6632.1999. tb08114.x zakaria, r. h., jaafar, n. i. m., & marican, s. (2012). financial behavior and financial position: a structural equation modeling approach.middle east. journal of scientific research, 12, 1396–1402. 336 c. w. copeland et al. / financial services review 30 (2022) 321–336 west, t., & worthington, a. c. (2018). life events and portfolio rebalancing of the family home. journal of financial counseling and planning, 29, 103–113. xiao, j. j., o’neill, b., prochaska, j. m., kerbel, c. m., brennan, p., & bristow, b. j. (2004). a consumer education programme based on the transtheoretical model of change. international journal of consumer studies, 28, 55–65. xiao, j. j., tang, c., serido, j., & shim, s. (2011). antecedents and consequences of risky credit behavior among college students: application and extension of the theory of planned behavior. journal of public policy & marketing, 30, 239–245. 88 m. sommer et al. / financial services review 30 (2022) 69–88 financial professionals and financial well-being: evidence from the national financial well-being survey richard stebbinsa,*, kyoung tae kima, martin seayb adepartment of consumer sciences, university of alabama, tuscaloosa, al 35487, usa bdepartment of personal financial planning, kansas state university, manhattan, ks 66506, usa abstract this study examined the association between financial professional use and financial well-being using the 2016 national financial well-being survey. we tested financial well-being across various sources of financial advice such as financial professionals, family, employer, community, financial institution, and government. results from the logistic regression showed that those who received advice from financial professionals had higher levels of financial well-being than those who did not receive advice from a financial professional. additional analyses with those who received financial advice from any source also showed that use of a financial professional had a stronger positive association with financial well-being. this study provides important insights to help consumers and educators better understand the value of financial professionals. © 2022 academy of financial services. all rights reserved. jel classifications: d12; d14 keywords: source of financial advice; financial professionals use; financial well-being; national financial well-being survey 1. introduction seeking out and using financial advice is a form of help-seeking behavior driven by the need to “solve problems, meet needs, and reach goals related to financial subjects” (fan, 2021). prior research created a framework for financial help-seeking behavior using help-seeking processes in health care decisions (e.g., grable & joo, 1999). this framework provides five *corresponding author: tel.: 205-348-6057; fax: 205-348-8721. e-mail address: rstebbins@ua.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 191–204 stages for consideration: (1) financial behaviors, (2) self-evaluation of financial behaviors, (3) the identification of the causes of financial behaviors, (4) decision to seek help, and (5) choosing an assistance option (grable & joo, 1999). age, education, income, and self-esteem tend to be the factors that most directly influence choice of help-provider (grable & joo, 2003). there are also certain characteristics that those likely to help-seek seem to share across studies. older adults with higher levels of financial literacy and cognitive ability are more likely to seek advice from a financial professional (kim, maurer, & mitchell, 2019). among students, older students with less net worth and financial knowledge were more likely to seek advice (britt, grable, cumbie, cupples, henegar, schindler, & archuleta, 2011). finding a source of financial advice is relatively easy. a public awareness campaign launched by the certified financial planner board of standards, inc. (cfp board) in 2010 increased awareness of the cfp certification in the united st from 17% in 2011 to 34% in 2015 (cfp board, 2015). the cfp board has spent over $90 million on marketing since 2011, with another campaign, with a cfp professional, launched in 2018 (cfp board, 2021). the percentage of households using financial advice has been reported at 27% (elmerick, montalto, & fox, 2002; hanna, 2011), similar to the 26.55% found in the current study. however, not all sources of advice are equal. financial professionals may provide a wide range of services for clients, from helping plan for a child’s education to creating a comprehensive financial plan (elmerick, montalto, & fox, 2002). the relationship may be limited in scope and time or extend for a lifetime. research has also connected the use of a financial planner with increased financial knowledge (robb, babiarz, & woodyard, 2012), which influences the ability to meet financial obligations and make investment decisions (hilgert, hogarth, & beverly, 2003). planning for the future, specifically holding a retirement saving goal, has also been linked with the use of a financial planner (kim, pak, shin, & hanna, 2018). combined, these individual elements suggest that the use of a financial professional might be connected with financial well-being. family and friends might not have the same level of expertise, leading to different outcomes. previous studies have recognized the importance of financial well-being on a variety of outcomes, with a special emphasis on the young (gutter & copur, 2011; shim, xiao, barber, & lyons, 2009). despite the importance placed on financial well-being, u.s. adults average a score of 54 out of 100 on the financial well-being scale devised by the consumer financial protection bureau (cfpb, 2017). according to the cfpb, approximately 33% of adults score below 51 on the scale that means that these adults have a high probability of worrying about food, running out of food, having utilities shut off, being unable to afford medical treatment, or even becoming homeless (cfpb, 2017). governmental action during the coronavirus disease 2019 (covid-19) pandemic, including stimulus checks, student loan interest and payment freeze, student loan forgiveness, and eviction moratoriums, highlight the importance policymakers place upon financial well-being. research examining financial well-being has used varying terminology to describe the concept including financial wellness (joo & garman, 1998), financial satisfaction (joo & grable, 2004), and financial stress (kim & garman, 2003). in particular, previous studies have used financial satisfaction as a mediator between wages and happiness (diener & biswas-diener, 2002), a dimension of life satisfaction or well-being (vera-toscano, atecaamestoy, & serrano-del-rosal, 2006), and as subjective well-being and life satisfaction overall (archuleta, dale, & spann, 2013; hsieh, 2001; plagnol, 2011). while there is 192 r. stebbins et al. / financial services review 30 (2022) 191–204 general agreement on the broad definition of financial well-being, there has not been a consensus on how it should be specifically measured or defined. recent research has viewed financial well-being as a construct dealing with current money management and expected future financial security (netemeyer et al., 2018). financial well-being can also be broadly defined as an individual’s satisfaction with his or her personal financial situation (cfpb, 2015a). financial well-being is considered the goal of financial education and the “ultimate measure of success for financial literacy efforts” (cfpb, 2015a). the incharge financial distress/financial well-being scale (ifdfw scale) was developed in 2004 (prawitz, garman, sorhaindo, o’neill, kim, & drentea, 2006). this was the first scale created to specifically measure financial well-being and it was created based on the literature and suggestions from financial educators. this scale has been improved with the inclusion of input from financial practitioners (cfpb, 2015a). the main object of this study is to examine the association between the source of financial advice and financial well-being. in this study, we used the cfpb’s measure of financial wellbeing defined as “a state of being wherein a person can fully meet current and ongoing financial obligations, can feel secure in their financial future, and is able to make choices that allow enjoyment of life” (cfpb, 2015b). this was measured using four elements: (1) control over your day-to-day, month-to-month finances; (2) financial freedom to make choices to enjoy life; (3) capacity to absorb a financial shock; and (4) on track to meet your financial goals. the exact questions used to operationalize these elements can be found in the appendix table a1. for empirical analyses, we used the 2016 national financial well-being survey, which is the first national survey released by the cfpb. this study contributes to the literature by using an explicit and comprehensive definition of financial well-being that incorporates feedback from financial practitioners in addition to financial education experts. this study also makes an important contribution but assessing the role of any financial advice and then breaking out those who received financial advice to examine the impact of different sources of financial advice. this research will provide important insights into financial advice and financial well-being. this contribution to the literature will also aid consumers in the valuation of financial advice and financial literacy. 2. method 2.1. dataset and sample selection the 2016 national financial well-being survey (nfwbs) is a nationally representative dataset that was collected by the consumer financial protection bureau (cfpb) in 2016. the main goal of this dataset was to understand financial well-being in u.s. adults, how financial knowledge and consumer behavior contribute to financial well-being, and support research. respondents were emailed and completed the survey online. the survey asked questions related to financial well-being, skill, knowledge, and demographic information using existing scales when possible. after we dropped cases where respondents chose “response not written to database,” or “refused to answer,” the final analytic sample includes 6,248 respondents. for robustness, we analyzed a subsample of 5,097 respondents who used any source financial advice. r. stebbins et al. / financial services review 30 (2022) 191–204 193 2.1. measurement of variables 2.1.1. dependent variable. in this study, a dependent variable is a financial well-being scale score developed by nfwbs. financial well-being scale is constructed based on four elements: (1) control over daily and monthly finances, (2) capacity to absorb a financial shock, (3) on track to meet financial goals, and (4) the financial freedom to make choices that allow enjoyment of life. an individual with a high level of financial well-being feels that he can meet current and future financial obligations, is secure in his financial future, and the ability to make choices that allow enjoyment of life. financial well-being is operationalized by asking 10 questions on a likert-type scale that are combined to create a single score. this single score ranges from 0 to 100 and the mean is 52.30. fig. 1 shows a distribution of financial well-being scale. 2.1.2. independent variables. 2.1.2.1. source of financial advice. respondents were questioned “do you seek advice on matters involving money from any of the following types of people or organizations?” with several options to choose. respondents could choose as many or as few of the options listed. the options included family, employers, friends/co-workers, community, financial institution, financial professionals and government. family included parents, spouses (or partners), and even extended family members such as cousins. the friends/co-workers option included co-workers and those friends outside of the workplace and different types of employers combined into a single employer option. the community option included community or faith-based organizations. the financial professional option included financial advisors, planners, counselors, or coaches. fig. 1. distribution of financial well-being scale, 2016 nfwbs. note: weighted results. nfwbs = national financial well-being survey. 194 r. stebbins et al. / financial services review 30 (2022) 191–204 2.1.2.2. control variables. according to the cfpb, financial well-being may be influenced by: (1) income and employment; (2) savings and safety nets; (3) past financial experience; and (4) financial behaviors, skills, and attitudes. in addition to the source of financial advice, financial skill and financial knowledge are included in our model. the financial knowledge is measured based on three financial knowledge questions related to personal finance topic of compound interest, inflation and stock (lusardi & mitchell, 2008).the number of correct answers was summed, ranging from 0 to 3. the financial skill scale is a 10-item likert type scale that asks respondents to answer questions about their perceived skill in learning about finances, making financial decisions, and recognizing when they need more information or help to make a financial decision. the questions are combined to create a single score that ranges from 0 to 100 and the mean is 49.89. this study also included the following set of control variables such as age (18-24; 25–34; 35–44; 45–54; 55–64; 65–74; 75, or older), gender (male, female), marital status (married, partner, single, or separated/divorced/widowed), race/ethnicity (white, black, hispanic, or others), employment status (self-employed, employee, homemaker, student, disabled, or retired), education (less than high school, high school diploma, some college, bachelor’s degree, or postbachelor’s degree), household income (0–$20,000; $20,000–$29,9000; $30,000–$49,900; $50,000–$74,900; $75,000–$99,900; or $100,000 or more) and census division (new england, mid-atlantic, east-north central, west-north central, south atlantic, east-south central, west-south central, mountain, or pacific). 2.1.2.3. statistical analysis. for descriptive results, we conducted several t-tests to compare the value of financial well-being across different sources of financial advice. further, this study used an ordinary least squares (ols) regression model to analyze the association between financial professional use and financial well-being, controlling for various household characteristics. for robustness check, we conducted regression analyses on a subsample that used any source of financial advice. all of our results were weighted using the survey weight provided by the 2016 nfwbs. 3. results 3.1. descriptive results fig. 2 shows descriptive results of different sources of financial advice. the results showed that most of the respondents, 63%, received financial advice from family members. less than 4% of the respondents relied on the government or the community for financial advice. financial professionals were the third-largest source of financial advice at 21.8%. lastly, only 18.9% of the respondents did not consult with any of the given options for financial advice. note that because respondents could select multiple sources, the total is well over 100%. as shown in table 1, respondents who used a financial professional for advice reported the highest levels of financial well-being at with a mean score of 60.7. the mean financial wellbeing score ranged from 60.7 down to 50.3, which was associated with taking advice from the government. those respondents who did not receive any financial advice fared slightly better than those taking advice from the government, with a mean financial well-being score of 52.2. r. stebbins et al. / financial services review 30 (2022) 191–204 195 mean financial well-being scores went up to 52.7 for community advice, 53.2 for employer advice, 54.7 for advice from family, and 58 for advices from financial institutions. the overall mean score on the financial well-being scale was 52.3. table 1 also presents t-test results, including six group comparisons of financial well-being scale with a reference group of financial professional use. all six pairwise comparisons are statistically significant.1 with a mean financial well-being score of 52.3, the final analytic sample was 6,248 respondents. the mean financial knowledge score was 2.44 out of 3, and the mean financial skills score was 49.9 out of 100. this is high relative to other studies (lusardi & mitchell, 2014); however, given our sample consisting of financially literate and older white males, it is expected. all respondents were 18 or older with fewer younger participants (9.7% under the age of 24) and a larger group of older participants (24.3% were over age 61). the sample was 48.4% male and almost 55% were married. education levels were spread fairly fig. 2. descriptive statistics, source of financial advice, 2016 nfwbs. note: weighted results. nfwbs = national financial well-being survey. table 1 mean financial well-being scale by different sources of financial advice, 2016 nfwbs source of financial advice distribution mean financial well-being scale p-value financial professional (reference) 21.8% 60.7 n/a family 62.6% 54.7 <0.0001 employer 26.8% 53.2 <0.0001 community 3.6% 52.7 0.0003 financial institution 18.2% 58.0 0.0341 government 3.3% 50.3 0.0035 no advice 18.9% 52.2 <0.0001 note. weighted results. t-tests were conducted for six group comparisons (reference: financial planner). nfwbs = national financial well-being survey. 196 r. stebbins et al. / financial services review 30 (2022) 191–204 evenly between those holding a graduate degree, bachelor’s degree, some college, or having no college. the majority of the sample, 64.9% was white and 44.6% earned over $75,000 per year. almost half of the respondents were salaried workers and 20.9% of were retired. more detailed information is available in appendix table a2. 3.2. multivariate results table 2 presents baseline results from the ols regression. results from both the reduced and full model showed that use of a financial professional had a positive effect on financial well-being (reduced model) even after controlling for various types of financial advice (full model). in particular, the use of a financial professional was positively associated with financial well-being and specifically increased the level of financial well-being by 1.85–1.95. advice from family was positively related, while advice from employer and government were both negatively related to the level of financial well-being. financial knowledge and skills increased the financial well-being score. as the age of the respondent increased, the level of financial well-being increased. married respondents had higher financial well-being scores than partner and separated, divorced, or widowed. compared with salaried workers, disabled respondents had lower levels while homemakers and retired had higher levels of financial well-being. the stronger negative associations for financial well-being were found in respondents with a disability and those that were separated, widowed, or divorced. this is similar to previous research which found financial difficulties for divorced individuals (west & mitchell, 2022). the educational attainment of the respondent was positively associated with financial well-being. not surprisingly, as income level increased, the level of financial well-being increased. in fact, the largest coefficient was associated with incomes higher than $150,000 as compared with respondents with less than $20,000 in annual income. it is also important to note that both age and income both had stronger positive associations than other variables with financial well-being. this is consistent with previous research finding age and income were the most significant contributors to financial well-being (west et al., 2021). as shown in table 3, we conducted similar analyses with respondents who used any type of financial advice as a robustness check. results from additional analyses were consistent with our main results. use of a financial professional and use of family advice were positively while advice from employer and government were negatively associated with financial well-being. full results are available from authors upon request. 4. discussion and implications this study analyzed the association between financial professional use and financial wellbeing. among various sources of financial advice, respondents who used a financial professional for advice reported the highest levels of financial well-being while lowest for those who took advice from the government. further, results from the 2016 nfwbs showed a positive association between the use of a financial professional and financial well-being even after controlling for various types of financial advice. overall, our empirical results r. stebbins et al. / financial services review 30 (2022) 191–204 197 table 2 ols regression of financial well-being, all respondents (n = 6,248), 2016 nfwbs variables coefficient standard error p-value coefficient standard error p-value source of financial advice financial professional 1.8545 0.3479 <0.0001 1.9453 0.3679 <0.0001 family — — — 0.8701 0.3876 0.0248 employer — — — �0.9531 0.3290 0.0038 community — — — �0.1789 0.7340 0.8075 financial institution — — — 0.2831 0.3717 0.4463 government — — — �2.9460 0.7732 0.0001 no advice — — — 0.0102 0.5015 0.9837 financial knowledge score 0.8225 0.1906 <0.0001 0.7891 0.1912 <0.0001 financial skills score 0.4111 0.0110 <0.0001 0.4114 0.0110 <0.0001 age of respondent (ref: age 45–54) age 18–24 �1.1863 0.6657 0.0748 �1.4060 0.6718 0.0364 age 25–34 �1.4710 0.4492 0.0011 �1.4418 0.4510 0.0014 age 35–44 �0.8361 0.4780 0.0803 �0.7889 0.4776 0.0986 age 55–61 1.0994 0.5022 0.0286 1.1602 0.5023 0.0209 age 62–69 4.7919 0.6035 <0.0001 4.7489 0.6037 <0.0001 age 70–74 6.5155 0.8019 <0.0001 6.4250 0.8019 <0.0001 age 75 or older 7.0357 0.7622 <0.0001 6.9418 0.7636 <0.0001 male (ref: female) 0.3734 0.2823 0.186 0.4461 0.2824 0.1143 marital status (ref: married) partner �1.6284 0.5873 0.0056 �1.5036 0.5883 0.0106 single (never married) 0.5362 0.4275 0.2098 0.7868 0.4330 0.0692 separated/divorced/widowed �1.9218 0.4163 <0.0001 �1.6136 0.4250 0.0001 race/ethnicity (ref: white) black �0.0049 0.4488 0.9913 0.1117 0.4505 0.8042 hispanic 0.6181 0.4131 0.1346 0.6171 0.4139 0.1360 others �1.3211 0.5148 0.0103 �1.2575 0.5144 0.0145 employment status (ref: salaried workers) self-employed 0.0141 0.5519 0.9796 �0.1143 0.5518 0.8360 homemaker 1.6479 0.5816 0.0046 1.4503 0.5828 0.0129 student �1.4246 0.7291 0.0507 �1.5847 0.7297 0.0299 disabled �3.6130 0.5020 <0.0001 �3.7266 0.5042 <0.0001 retired 2.2690 0.5684 <0.0001 2.2063 0.5679 0.0001 education (ref: less than high school diploma) high school 2.3484 0.6994 0.0008 2.3881 0.6983 0.0006 some college 0.8066 0.7053 0.2528 0.8591 0.7039 0.2223 bachelor degree 1.4935 0.7420 0.0442 1.5274 0.7407 0.0392 post-bachelor degree 2.2965 0.7667 0.0028 2.3735 0.7654 0.0019 household income (ref: less than $20,000) $20,000–$29,999 1.1599 0.5949 0.0512 1.0676 0.5942 0.0724 $30,000–$39,999 2.4896 0.5860 <0.0001 2.3136 0.5859 <0.0001 $40,000–$49,999 4.1452 0.6690 <0.0001 3.9420 0.6686 <0.0001 $50,000–$59,999 5.7668 0.6501 <0.0001 5.5721 0.6499 <0.0001 $60,000–$74,999 6.5514 0.6274 <0.0001 6.3844 0.6270 <0.0001 $75,000–$99,999 7.2807 0.5903 <0.0001 7.1220 0.5900 <0.0001 $100,000–$149,999 8.9721 0.5925 <0.0001 8.7319 0.5936 <0.0001 $150,000 or more 11.4253 0.6267 <0.0001 11.2671 0.6274 <0.0001 constant 22.1398 1.0630 <0.0001 21.9743 1.1222 <0.0001 regional fixed effect (census division) included included adjusted r2 0.4047 0.4073 note. weighted results. nfwbs = national financial well-being survey; ols = ordinary least squares. 198 r. stebbins et al. / financial services review 30 (2022) 191–204 support the positive association between the use of a financial professional and financial well-being. interestingly, advice from employer and government were negatively related to the level of financial well-being, which implies an effect is offset by other types of financial advice. this may also be due to the likely types of advice offered by the government and the specificity of advice employers offer. government advice tends to cover government entitlements such as unemployment benefits, welfare, food stamps, or children’s health insurance program. employer financial advice might only cover retirement plan options for employees. if financial well-being is viewed as a domain of overall well-being or life satisfaction, policymakers should support financial planning education. this has been done at the k–12 level in several states with a positive impact of improving the credit scores and lowering the probability of delinquency in young adults (urban, schmeiser, collins, & brown, 2015). three years after the programs began, ‘credit scores increased by 10.89 points in georgia, 16.19 points in idaho and 31.71 points in texas’ (urban, schmeiser, collins, & brown, 2015). instructors teach these k–12 programs with various backgrounds with minimal training requirements in georgia and no formal training requirements in texas (urban, schmeiser, collins, & brown, 2015). a college degree in financial planning could help improve the curriculum. however, in 2018, only 4.5% of four-year universities offer degrees in financial planning (iacurci, 2018). funding for more programs can be encouraged through policy and consumer demand. the advice of financial planners might help vulnerable consumers better understand the market and the value of planning for retirement (hilgert et al., 2003; robb et al., 2012). an alarming portion of millennials are saving for retirement in conservative investments such as bonds or money market funds (tepper, 2018). on the other end of the pendulum, predatory retail investor trading of gamestop highlighted the volatility investors, primarily in their mid-30s, were willing to take on (hasso et al., 2021). in addition, there are fewer financial planners under the age of 30, who might better understand millennials, than there are over the age of 70 (iacurci, 2018). table 3 ols regression of financial well-being, respondents who used financial advice (n = 5,097), robustness check, 2016 nfwbs variables coefficient standard error p-value coefficient standard error p-value source of financial advice financial professional 1.8504 0.3593 <0.0001 2.0015 0.3668 <0.0001 family — — — 0.7598 0.3868 0.0495 employer — — — �0.9785 0.3259 0.0027 community — — — �0.0734 0.7231 0.9191 financial institution — — — 0.2684 0.3673 0.4649 government — — — �2.9334 0.7621 0.0001 constant 21.9445 1.2037 <0.0001 21.8242 1.2513 <0.0001 control variables included included regional fixed effect (census division) included included adjusted r2 0.4008 0.4040 note. weighted results. control variables are the same as table 2. nfwbs = national financial well-being survey; ols = ordinary least squares. r. stebbins et al. / financial services review 30 (2022) 191–204 199 in this study, there are some important limitations to note. first, this study used a crosssectional dataset, which makes it difficult to make a causal inference on the association between financial professional use and financial well-being. at present, no other national survey dataset is available that contains the full range of information needed for ideal analyses of the research questions, especially for the solid measurement of financial well-being. however, the use of a longitudinal dataset allows researchers to account for some methodological concerns. in addition, heckman, seay, kim, and letkiewicz (2016) discussed a significant concern about the content validity of financial planner measurement among publicly available u.s. household datasets. given the nfwbs dataset’s limitation, financial professional use is defined broadly incorporating financial advisors, planners, counselors, or coaches. there may be different levels of interaction within each category with clients, ranging from biannual meetings over a lifetime for a comprehensive financial planner to a few meetings in a single month to cover an emergency situation with a financial counselor. self-selection bias may also exist as clients seeking professional financial advice might be more likely to be in good financial situations and potentially possess greater knowledge than those that do not seek professional financial advice. the positive association between financial professionals and financial well-being supports the cfp board’s public awareness campaign and financial institutions’ increased offerings of robo advisors (fisch, laboure, & turner, 2017). robo advisors are more popular with millennials than the boomers and more people now have access to some sort of financial planning (cutler, 2015). however, a weakness of the robo advisor is that it does not educate the users on financial planning topics and that goes hand-in-hand with the fact that the robo advice is only as good as the information the user supplies (wharton, 2018). future research could examine the outcomes of clients that use a robo-advisor compared with clients that use a financial professional. future studies should use a well-developed scale of financial professional use that may capture the various aspects of financial advisory services to meet the validity requirement as well as capture specific aspects of comprehensive financial planning services. in addition, the potential differences in financial well-being of different cohorts using different types of financial professionals for advice could yield interesting results. future research should also examine clients’ financial well-being before and after engagement of a financial professional to potentially speak to causation and impact. note 1 we conducted similar t-tests as a reference group of no advice. five pair-wise comparisons are found to be significant except for the pair of community—no advice. 200 r. stebbins et al. / financial services review 30 (2022) 191–204 appendix table a1 descriptive statistics of sample characteristics, 2016 nfwbs variables percentage mean (median) financial well-being scale 52.30 (54.00) source of advice financial professional 21.8% family 62.6% employer 26.8% community 3.6% financial institution 18.2% government 3.3% no advice 18.9% mean (median) financial knowledge score 2.44 (3.00) mean (median) financial skills score 49.89 (49.00) age of respondent age 18–24 9.67 age 25–34 21.13 age 35–44 14.07 age 45–54 18.93 age 55–61 11.92 age 62–69 10.83 age 70–74 5.35 age 75 or older 8.10 gender male 48.43 female 51.57 marital status married 55.39 partner 6.71 single (never married) 22.24 separated/divorced/widowed 15.66 race/ethnicity white 64.89 black 11.65 hispanic 15.45 others 8.02 employment status salaried workers 50.41 self-employed 6.94 homemaker 6.78 student 5.05 disabled 9.91 retired 20.90 education less than high school 4.77 high school 20.72 some college 28.45 bachelor’s degree 24.26 post-bachelor’s degree 21.81 (continued on next page) r. stebbins et al. / financial services review 30 (2022) 191–204 201 references archuleta, k. l., dale, a., & spann, s. m. (2013). college students and financial distress: exploring debt, financial satisfaction, and financial anxiety. journal of financial counseling and planning, 24, 50–62. britt, s. l., grable, j. e., cumbie, j., cupples, s., henegar, j., schindler, k., & archuleta, k. (2011). student financial counseling: an analysis of clinical and non-clinical sample. journal of personal finance, 10(2), 172–186. cfp board. (2015). 2015 annual brand tracking report: executive summary. available at https://www.cfp.net/ docs/default-source/news-events–-research-facts-figures/2015-brand-tracking-report-executive-summary.pdf cfp board. (2021). increasing awareness. available at https://www.cfp.net/initiatives/increasing-awareness table a1 (continued) variables percentage household income less than $20,000 13.08 $20,000–$29,999 8.86 $30,000–$39,999 9.95 $40,000–$49,999 6.63 $50,000–$59,999 7.55 $60,000–$74,999 9.37 $75,000–$99,999 13.41 $100,000–$149,999 15.97 $150,000 or more 15.18 census division new england 5.06 mid-atlantic 12.87 east-north central 14.42 west-north central 6.59 south atlantic 21.16 east-south central 5.15 west-south central 11.27 mountain 6.91 pacific 16.57 note. weighted results. nfwbs = national financial well-being survey. table a2 financial well-being survey questions, 2016 nfwbs no. questions how well does this statement describe you or your situation? 1 i could handle a major unexpected expense 2 i am securing my financial future 3 because of my money situation, i feel like i will never have the things i want in life 4 i can enjoy life because of the way i’m managing my money 5 i am just getting by financially 6 i am concerned that the money i have or will save won’t last how often does this statement apply to you? 7 giving a gift for a wedding, birthday, or other occasion would put a strain on my finances for the month 8 i have money left over at the end of the month 9 i am behind with my finances 10 my finances control my life note. nfwbs = national financial well-being survey. 202 r. stebbins et al. / financial services review 30 (2022) 191–204 consumer financial protection bureau. (cfpb). (2015a). financial well-being: the goal of financial education. iowa city, ia: consumer financial protection bureau. available at https://www.consumerfinance. gov/data-research/research-reports/financial-well-being/ consumer financial protection bureau (cfpb). (2015b). measuring financial well-being: a guide to using the cfpb financial well-being scale. available at https://files.consumerfinance.gov/f/201512_cfpb_financialwell-being-user-guide-scale.pdf consumer financial protection bureau (cfpb). (2017). financial well-being in america. available at https:// files.consumerfinance.gov/f/documents/201709_cfpb_financial-well-being-in-america.pdf cutler, n. e. (2015). millennials and finance: the “amazon generation.” journal of financial service professionals, 69(6), 33–39. diener, e., & biswas-diener, r. (2002). will money increase subjective well-being? social indicators research, 57, 119–169. https://doi.org/10.1023/a:1014411319119 elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by us households. financial services review, 11, 217. fan, l. (2021). a conceptual framework of financial advice-seeking and short-and long-term financial behaviors: an age comparison. journal of family and economic issues, 42, 90–112. https://doi.org/10.1007/s10834020-09727-3 fisch, j. e., laboure, m., & turner, j. a. (2017). the economics of complex decision making: the emergence of the robo adviser. available at https://www.geog.ox.ac.uk/events/170911/robo-vs-human-advisers-aug-28.pdf grable, j. e., & joo, s. h. (1999). financial help-seeking behavior: theory and implications. journal of financial counseling and planning, 10, 14. grable, j. e., & joo, s. h. (2003). a snapshot view of the help-seeking market. journal of financial planning, 16, 88. gutter, m., & copur, z. (2011). financial behaviors and financial well-being of college students: evidence from a national survey. journal of family and economic issues, 32, 699–714. https://doi.org/10.1007/s10834-0119255-2 hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. hasso, t., müller, d., pelster, m., & warkulat, s. (2021). who participated in the gamestop frenzy? evidence from brokerage accounts. finance research letters, 45(2), 102140. heckman, s., seay, m., kim, k. t., & letkiewicz, j. (2016). household use of financial planners: measurement considerations for researchers. financial services review, 25, 427–446. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309. hsieh, c.-m. (2001). correlates of financial satisfaction. international journal of aging & human development, 52, 135–153. https://doi.org/10.2190/9yde-46pa-mv9c-2jrb iacurci, g. (2018). few colleges offer a financial planning program. investmentnews. available at https:// www.investmentnews.com/few-colleges-offer-a-financial-planning-program-75860 joo, s.-h., & garman, e. t. (1998). the potential effects of workplace financial education based on the relationship between personal financial wellness and worker job productivity. personal finances and worker productivity, 2, 163–173. joo, s.-h., & grable, j. e. (2004). an exploratory framework of the determinants of financial satisfaction. journal of family and economic issues, 25, 25–50. https://doi.org/10.1023/b:jeei.0000016722.37994.9f kim, j., & garman, e. t. (2003). financial stress and absenteeism: an empirically derived model. journal of financial counseling and planning, 14, 31. kim, h. h., maurer, r., & mitchell, o. s. (2019). how cognitive ability and financial literacy shape the demand for financial advice at older ages (no. w25750). national bureau of economic research. available at https://www.nber.org/papers/w25750 kim, k., pak, t., shin, s., & hanna, s. (2018). the relationship between financial planner use and holding a retirement saving goal: a propensity score matching analysis. financial planning review, 1, e1008. https:// doi.org/10.1002/cfp2.1008 r. stebbins et al. / financial services review 30 (2022) 191–204 203 lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review, 98, 413–417. https://doi.org/10.1257/aer.98.2.413 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. https://doi.org/10.1257/jel.52.1.5 netemeyer, r. g., warmath, d., fernandes, d., & lynch, j. g., jr. (2018). how am i doing? perceived financial well-being, its potential antecedents, and its relation to overall well-being. journal of consumer research, 45, 68– 89. https://doi.org/10.1093/jcr/ucx109 plagnol, a. c. (2011). financial satisfaction over the life course: the influence of assets and liabilities. journal of economic psychology, 32, 45–64. https://doi.org/10.1016/j.joep.2010.10.006 prawitz, a. d., garman, e. t., sorhaindo, b., o’neill, b., kim, j., & drentea, p. (2006). incharge financial distress/financial well-being scale: development, administration, and score interpretation. financial counseling and planning, 17, 34–50. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21(4), 291–305. shim, s., xiao, j. j., barber, b. l., & lyons, a. c. (2009). pathways to life success: a conceptual model of financial well-being for young adults. journal of applied developmental psychology, 30, 708–723. https://doi.org/ 10.1016/j.appdev.2009.02.003 tepper, t. (2018). millennials prefer cash over stocks–and it could cost them millions. bankrate. available at https://www.bankrate.com/investing/financial-security-july-2018/ urban, c., schmeiser, m., collins, j. m., & brown, a. (2015). state financial education mandates: it’s all in the implementation. washington, dc: finra investor education foundation. available at https://www. finrafoundation.org/files/state-financial-education-mandates-its-all-implementation vera-toscano, e., ateca-amestoy, v., & serrano-del-rosal, r. (2006). building financial satisfaction. social indicators research, 77, 211–243. https://doi.org/10.1007/s11205-005-2614-3 west, t., cull, m., & johnson, d. (2021). income more important than financial literacy for improving wellbeing. financial services review, 29(3), 187–207. west, t., & mitchell, e. (2022). australian women with good financial knowledge fare better in divorce. australian journal of management, 47, 203–224. https://doi.org/10.1177/03128962211022041 wharton. (2018). the rise of the robo-advisor: how fintech is disrupting retirement. knowledge@wharton. available at https://knowledge.wharton.upenn.edu/article/rise-robo-advisor-fintech-disrupting-retirement/ 204 r. stebbins et al. / financial services review 30 (2022) 191–204 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) a framework for analyzing defined benefit pension insurance: the survivor benefit plan for veterans william w. jenningsa, jeffrey c. merrellb, thomas c. o’malleyc, brian c. payned,* adepartment of management, united states air force academy, co 80840, usa bleeds school of business, university of colorado, 995 regent drive, boulder, co 80309, usa cdepartment of management, united states air force academy, co 80840, usa dcollege of business, university of colorado-colorado springs, 1420 austin bluffs parkway, colorado springs, co 80918, usa abstract millions of defined benefit pensioners must select a pension insurance method. we present a framework for making this decision within the context of u.s. military veterans’ survivor benefit plan (sbp). federal government subsidies generate a positive expected net payout for sbp. while insurance outcomes are typically skewed, the asymmetry of sbp outcomes is stark. in a common scenario, five percent of participants receive 60% of benefits. an alternative financial planning approach incorporates private insurance and investments and often bests the sbp. actuarially correct life expectancy, moral hazard, taxes, and individual financial needs all play important roles in selecting a pension insurance program. © 2018 academy of financial services. all rights reserved. jel classification: g23; h55; j38 keywords: survivor benefit plan; insurance; valuation; pension; retirement; simulation 1. introduction tens of millions of people in the united states with defined benefit (db) pensions will face a key financial decision when they retire: whether and how to insure their db pension for their spouse. these insured plans are common among state employees, certain federal * corresponding author. tel.: �1-719-255-3186; fax: �1-719-255-3494. e-mail address: bpayne3@uccs.edu (b.c. payne) financial services review 27 (2018) 147-172 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. employees, and, central to this analysis, military veterans. richwine (2013) asserts that about 86% of local and state government employees have db pensions. there are tens of millions of first responders and state employees as well as over 1.5 million military members who at retirement can opt for a life insurance product that pays beneficiaries a life annuity upon the insured’s death. additionally, although it covers private pensions, the employee retirement income security act of 1974 (erisa) requires db plans to provide a survivor’s benefit that gives pension payments over both the retiree’s and spouse’s lifetimes. these programs generally have “opt-out” features, meaning couples are defaulted into selecting joint benefits but can elect not to pay the costs for the prospective survivor benefits. consequently, these couples and their advisors need a framework to analyze this key decision.1 this article provides such a framework using the u.s. military’s pension insurance program as a specific case. one can tailor the approach presented here to any of the tens of millions facing this decision. for military service members the military pension insurance program is called the survivor benefit plan (sbp). for years the conventional wisdom has been that military retirees should embrace the sbp. among the reasons: y it requires no medical exam, y it is government-subsidized insurance (that implies it is more than actuarially fair), y it takes care of a surviving spouse and/or children for their lifetime in real terms, and y as davis and fraser (2012) summarize, it can provide substantial risk-free real returns. while these points have merit, the analyses to date are admittedly incomplete. we expand earlier findings with this study and provide a more granular investigation of the sbp decision, particularly with respect to military retiree and spouse life expectancies, the skewness embedded in average returns, and market-based insurance and investment alternatives. we also approach this narrow decision from a holistic financial planning perspective and believe it is imperative that retirees deliberately and thoughtfully consider an sbp-like decision in light of their complete financial planning and life circumstances. based on our analysis, for a broad swath of the retiring military population, we reach the opposite conclusion of prior studies: that is, we contend sbp is not an ideal approach to pension protection or life insurance. more strongly, our default position would be to recommend against sbp unless an individual has a compelling reason or need to select it.2 because the sbp program is subsidized by the federal government, we posit that other db pension recipients could reach the same conclusion when considering their pension insurance options. finally, because this conclusion is not the default option during the busy time when one is transitioning to retirement, we advocate analyzing one’s alternatives in advance of retirement by two to five years. before reaching these overarching conclusions, we map the distribution of potential outcomes sbp participants can expect to experience and summarize this comprehensive picture with three major findings: 1. the expected returns are more skewed than previously documented, with more than 60% of retirees in a common scenario expected to pay more into sbp than they receive, thus, losing wealth; 148 w.w. jennings et al. / financial services review 27 (2018) 147-172 2. from a holistic financial planning perspective, sbp is largely irrelevant for the most-likely mortality scenarios; and 3. in circumstances where sbp is most valuable, there exist other life insurance mechanisms to provide similar, or even better, financial security. our article augments current tools available to federal employees making this critical financial decision. the department of defense office of the actuary maintains a sbp calculator on its website that quantifies the expected benefit based on user-customized inputs (http://actuary.defense.gov/, “survivor benefit plan” tab). the general takeaway from this calculator is that sbp provides a financially valuable payout to those who subscribe. in contrast, this study provides more context for these calculations and for this important decision. we identify viable alternatives to the sbp and provide a novel calculator that allows individuals with unique circumstances to compare our proposed alternative and the sbp (this calculator is located at www.financialcheckpoints.com). 2. background and literature review as poterba et al. (2007) show using a simulation methodology, public defined benefit pension plans are very generous. according to the analysis presented in jennings and reichenstein (2001), the present value equivalent of a retired military officer pension can readily exceed $1 million. currently, military members receive a monthly pension amounting to 2.5% of their final 36 months’ average pay (nominal) for every year of service.3 for example, after 20 years of service at 2.5% per year, a retiree receives 50% times their high-36 month average pay, which is usually calculated from the retiree’s last 36 months of pay. for the typical 20-year enlisted career this amounts to $2,100 per month in 2015; it is approximately $4,000 per month for a typical officer. given the longevity of the typical military retirement, these monthly cash flows are well worth protecting. the life cycle hypothesis holds that households seek stability in consumption and, thus, tend to save during their labor income years and dissave in retirement (ando and modigliani, 1963). with the diminishing marginal utility of wealth, losses are disproportionately harmful. in the event the military retiree dies without sbp protections, the spouse is not entitled to the retiree’s retirement income. annuitization stops as does the associated welfare gains found with annuities in standard life cycle models. this loss potential is problematic; people who are forward-looking, utility-maximizing, and understand longevity risk should, therefore, choose to protect this income stream to smooth consumption through time. to this end, the military’s sbp helps; it protects a portion of the db pension cash flows that military members receive after 20� years of service. to protect this pension for the military member’s family, the sbp permits military retirees to pay up to 6.5% of their monthly pension pretax to guarantee their beneficiary up to 55% of the monthly retirement benefit should the military retiree die.4 while the historical opt-in rates for sbp are not available to the authors, the department of defense (dod) actuary plans for a majority of retirees to enroll going forward, using a 50–60% election rate. of these individuals, most 149w.w. jennings et al. / financial services review 27 (2018) 147-172 retirees (80%) who elect coverage choose the maximum amount (“report of the military compensation and retirement modernization commission,” final report, january 2015; military compensation and retirement modernization commission 2015). under the typical enlisted retiree scenario, the retiree pays $137 pretax monthly for their surviving spouse to receive a taxable benefit of $1,155 per month. the typical officer’s figures are $260 and $2,100, respectively. unlike the scenario in lachance et al. (2003), which evaluates pensioners’ option to buy back into a db plan from a defined contribution plan, sbp provides very limited optionality. similar to many pension insurance programs, the sbp is an opt-out program where the member and spouse must affirmatively elect to not pay for these benefits. new retirees and their spouses can opt out immediately or between the second and third year of retirement. opt-out decisions are irrevocable, unless the retiree subsequently remarries or adopts a child. sbp spousal beneficiary payments continue for the remainder of the surviving spouse’s life, or until the spouse remarries if he or she remarries before age 55. premiums and payouts are adjusted with inflation annually. premium payments cease after 30 years; however, the coverage continues for the duration of the member’s life. if the spouse predeceases the member, both premiums and spousal coverage cease immediately. there are no refunds of premiums paid. for more extensive details on the sbp program, see higdon (2009). the sbp pension protection program is essentially an inflation-adjusted annuitized life insurance product. as such, the analysis here is relevant to many annuity options presented to public employees. to be clear, the sbp is not an investment, which makes comparing its return to investments inappropriate. unlike comprix and muller (2011), for example, who find evidence private firms adjust their db pension discount rate to suit their purposes, we are not concerned with a discount rate (i.e., expected investment returns beyond a risk-free rate) for u.s. government pension assets. instead, insurance such as sbp operates under the principle of indemnity. its intent is to make the insured whole after experiencing a loss. as with most insurance products, consumers should expect—and in most cases hope—that they will not have to realize any benefit from paying insurance premiums. insurance products exist to prevent harmful large losses. as one author states, “it makes sense to buy insurance where the premiums will result in a known small loss” (kitces, 2014). the question this study examines is whether sbp provides the best “known small loss” option for military retirees. while recent work on pension insurance programs is limited (e.g., bell and graham, 1984, analyze 900 private pension plans’ characteristics more than three decades ago), some analysis of the sbp program has been conducted. jennings and reichenstein (2001) demonstrate sbp is actuarially fair under the then-current rules. using cost-benefit analysis, beatty and kang (2007) question the sbps value in limited illustrative scenarios. sbp has been enhanced since these analyses, making it likely that sbp is advantageous for the typical participant. davis and fraser (2012) investigate two changes to sbp that took effect in 2008. first, a 30-year maximum payment timeframe was then implemented. second, an offset feature that reduced sbp benefits for social security income was eliminated. the authors find that with these changes the average implied rate of return to sbp is a riskless 6.8% for a 45-year-old couple, suggesting the program returns make it quite appealing. implied returns are especially generous for male retirees who have younger female spouses. 150 w.w. jennings et al. / financial services review 27 (2018) 147-172 this study extends the sbp analysis of davis and fraser (2012) in ways they recommend and reaches more muted conclusions. first, we utilize life expectancy tables for military members rather than using those for the general population. the difference is nontrivial. service members are subject to extensive initial health screenings and then mandatory annual physical fitness tests required to remain in the service. military members also undergo routine preventive health exams that sometimes lead to preretirement medically-based diagnoses that can lead to separation from service. accounting for these factors leads to a life expectancy for military retirees (i.e., those who serve 20-plus years) that is significantly higher than for the general population. the sbp experiences a smaller adverse selection problem than other insurance programs. further, given that most military members are not married to other military members, the typical retiree spouse has a life expectancy more closely approximated by the general population. as a result, the probability of a spouse collecting sbp diminishes substantially. in contrast to prior analyses, using military-relevant life expectancy tables reduces the financial appeal of sbp. other contributions of this study address recommendations proposed in beatty and kang (2007). that is, we provide a rich set of life expectancy distribution data, absolute payoff data, and expected payoff distribution data for a variety of member-spouse mortality combinations, to include child-only sbp coverage (the online calculator allows myriad age and status combinations; however, this article depicts the most statistically common scenario for illustration purposes). we provide this analysis for both the sbp option as well as a viable cost-neutral alternative, which includes purchasing privately-issued level-term life insurance coupled with investing the difference between these private premiums and the sbp premiums. instead of simulating possible member-spouse mortality scenarios, we provide the joint life expectancy distribution for individual couples. this allows military members to identify the tradeoffs associated with sbp decision by forecasting a very wide spectrum of possible outcomes based on their unique circumstances. furthermore, we present the entire outcome space, which accurately depicts the proportion of probable sbp “winners” and “losers” for the first time. this approach enables one to better view the mortality landscape and the associated skewness in the payoff distribution. 3. data and methodology it is important to recognize that on average sbp-insured military officers live longer lives. just as the social security administration (ssa) calculates life expectancy tables for the general u.s. population, the dod actuary generates tables for military retirees.5 fig. 1 shows that military officer retirees have markedly higher survival rates at every age relative to general population men and women. while data for first responders are not available to us, we posit that their life expectancies approximate those of military members more closely than those of the general population given similarities in health screening, physical training, job-related physical requirements, and job-related risk. on the beneficiary side of the equation, it is important to recognize that according to the dod actuary, military spouses unfortunately do not experience the same expected longevity as the retired military members. currently, most military retirees are males (90%�),6 and 151w.w. jennings et al. / financial services review 27 (2018) 147-172 married military members tend to wed non-military spouses (90%�).7 these unbalanced figures aid our analysis, because the dod actuary does not report retiree and spouse tables by gender. because the vast majority of spouses are female, we can compare ssa female probabilities to military retiree spouses. fig. 1 shows the probability of survival for military retiree spouses relative to the general population females. retiree spouses have slightly higher survival probabilities than ssa females before age 70, but after age 70 they are comparable. because military retirees and their spouses differ from the general population, this warrants an analysis that focuses on their unique life expectancies. because prior studies use ssa life expectancy for both military retirees and spouses, they potentially overstate the sbp value. to illustrate our method and showcase our results we use a currently common officer retirement scenario. our baseline scenario involves a military male retiree age 44 married to a non-military female who is also 44 years old. at this point a typical officer would have a 20–22 year career, while the typical enlisted member would have a 22–26 year career. the officer retiree’s insured amount is $55,000. we further assume the spouse does not remarry once the service member dies. in fact, if a beneficiary spouse remarries before age 55, then sbp payouts cease unless the remarriage ends in divorce or the new spouse dies. we do not incorporate this remarriage probability. the value of sbp would decrease relative to the baseline scenario shown here if remarriage were considered. as for our dollar value calculations, because future sbp premiums and payouts adjust with inflation, we do not use a discount factor to adjust them.8 finally, given our personal biases, we demonstrate the typical officer retiree calculations, summarize some others, and allow the reader to explore fig. 1. survival probability of u.s. citizens, u.s. military retirees, and retirees’ spouses. this plot shows the probability of survival by age for an average u.s. male and female according to the social security administration. it also shows the probability of survival for u.s. military officer servicemembers who retire and their spouses, according to the department of defense actuary office. 152 w.w. jennings et al. / financial services review 27 (2018) 147-172 alternative scenarios using an online calculator. obviously, there are many permutations to this scenario; we summarize a few at the end of this study. coupling these assumptions with the ssa and dod life expectancy tables allows us to generate a joint probability density function for expected mortality at any age combination the military retiree and spouse might experience. fig. 2 depicts this joint distribution for our age 44 officer retiree scenario, with the highest (or most likely) mortality combination occurring when the retiree is 88 and the surviving spouse is 89. as expected, the joint distribution is left-skewed toward lower ages given the rather abrupt upper-bound on mortality (i.e., the virtual impossibility of living much beyond 110). for ease of interpretation, most of the subsequent figures incorporate the same basic format and perspective as fig. 2. 4. results while the joint life expectancy information is necessary, the payouts associated with these life expectancy combinations is equally important. we calculate these net sbp payouts in fig. 3. while we calculate payouts based on $55,000 of retired officer pay, payout calculations are the same on a percentage basis for any retiree because of the percentage-based calculations of sbp premiums and payouts. fig. 3 shows absolute amounts for the typical officer retiree. it implements the same xand y-axes as fig. 2; however, it quantifies spouse’s fig. 2. joint probability of mortality for u.s. military retirees and spouses. this plot shows the joint probability density function for mortality age of u.s. military retirees and their spouses when both are currently age 44, according to the according to the department of defense actuary office. the military member is male; the spouse is female in this illustration. the highest expected joint probability occurs when the member is age 88 and the spouse is age 89. 153w.w. jennings et al. / financial services review 27 (2018) 147-172 payout minus the member’s premiums for each retiree-spouse mortality combination. for instance, if our 44-year-old military retiree lives to age 50, then he has paid six years’ worth of sbp premiums, or $21,450. if the officer’s equivalent-aged spouse then lives to age 70, she collects 20 years’ worth of sbp payouts, or a total of $605,000.9 taking the difference between these payouts and premiums yields a net real payout of approximately $583,550 for the officer who opts for sbp pension protection in this scenario. fig. 3 shows who can expect to win and to lose—in strictly financial terms—from the sbp program. the extreme sbp “winners” are, sadly, the spouses of sbp-covered military members who die immediately after retiring (as we discuss later, this analysis does not account for the fact that retirees carry their active duty sgli [life insurance] coverage for 120 days into retirement at no cost). we hope that the reader understands we put the word winners in quotations for obvious reasons, as we would rather nobody experiences this situation! clearly, the untimely death of a spouse is a personal and financial catastrophe we wish everyone could avoid. the extreme winning combinations are those in the circled area of fig. 3. the net payout value for officer retirees’ spouses in this plot approaches $2 million in the extreme case where a retiree dies immediately, the surviving widow stays single or remarries after 55, and then she lives to well over 100. the net payoff function for the sbp “losers” has a lower bound at 30 years’ worth of premiums, or $107,250 in our baseline scenario. the maximum negative net payouts—that represent the worst “losing” situations—occur when the military member pays 30 years’ worth of sbp premiums and the military retiree’s spouse does not survive the retiree. using dod mortality assumptions, this occurs more often than not. in this scenario where the retiree and spouse are the same age, the probability that the spouse outlives the military fig. 3. survivor benefit plan (sbp) net payout (benefits – costs) for u.s. military retirees and spouses. this plot shows the sbp net payouts, equal to the insurance payout minus premium costs, for various mortality combinations of a military member and nonmilitary spouse who elect to purchase sbp pension protection. 154 w.w. jennings et al. / financial services review 27 (2018) 147-172 retiree is around 42%, markedly lower than the almost 60% calculated using ssa mortality tables (davis and fraser, 2012). yet, this result is not surprising given the survival probability differences shown in fig. 1. the triangle over the trough in fig. 3 highlights the negative payout scenarios. the cost-benefit breakeven point for a spouse to recoup all 30 years’ of premiums is 3.55 years ($107,250 worth of 30-year premiums divided by $30,250 annual payout for officers), meaning a spouse must outlive the retiree by approximately three and one-half years to recoup all sbp premiums. more generally, for sbp payouts to offset premiums paid (i.e., breakeven) at any point in time, the spouse beneficiary must outlive the insured member by 12% of the total duration that the couple has paid sbp premiums.10 the upward sloping payout on the right side of the triangle represents the mortality combinations where the spouse predeceases the retiree. in these situations, the retiree has paid some premiums—but not the full 30 years’ worth—and ceases paying upon the death of his spouse. having laid the groundwork for understanding joint life expectancy and absolute payoffs, we can now quantify the expected costs versus benefits of the sbp program for an officer retiree. this is done by simply multiplying the values in figs. 2 and 3. doing so yields fig. 4, which plots the expected net payout contour for the couple in our scenario. visually, the joint life expectancy probabilities dominate the surface relative to the payouts. because there is a high probability of both a military retiree and his spouse living around 80 years, these probabilities tend to dominate the net gains and losses shown in fig. 3. the likelihood of “winning” in the sbp—again, the member dying fig. 4. expected survivor benefit plan (sbp) net payout for u.s. military retirees and spouses. this plot shows the probability-weighted sbp net payouts for various mortality combinations of a military member and nonmilitary spouse who elect to purchase sbp pension protection. 155w.w. jennings et al. / financial services review 27 (2018) 147-172 tragically young and spouse living for a long time—is so low that the multimillion-dollar payout scenario scarcely registers. recognizing that the volume under the curve in fig. 4 represents the aggregate expected value of sbp (i.e., probability � net payout, where sum of probabilities equals one), some noteworthy observations result. for instance, as the sbp program and prior authors have noted, the program is not actuarially-neutral. in fact, its expected net payout (i.e., benefits minus premiums) for this scenario is over $61,000. thus, as many claim, sbp is a subsidized program for the benefit of military retirees. also, the knife-edge between winning and losing in the center of the plot stands out. probabilistically, it is almost a toss-up between whether (1) the retiree dies early enough for his spouse to collect enough sbp for a positive net payout (represented by the peak) and (2) the retiree either lives too long for his spouse to collect enough sbp or that she predeceases him, represented by the deep depression. we now ask where the sbp value comes from and whether that value most effectively addresses the retiree’s needs. it turns out that almost 60% of the sbp value comes in an expanded definition of who are the sbp winners. specifically, let us expand the definition of winners to include those scenarios where the military retiree dies before age 66, and his spouse lives at least 10 years more. in other words, the retiree dies relatively young, and the spouse then survives for at least another decade. while this particular scenario occurs less than five percent of the time, it represents almost 60% of the sbp total expected value. stated differently, for a common retiree-spouse scenario, non-winners represent over 95% of the scenarios but only collect 40% of the expected sbp benefit. the returns to sbp are exceedingly skewed toward those who die relatively young and have long-lived surviving spouses who do not remarry. a minority of people get more back than they spend, and a very select few get exceedingly high returns. there is over a 60% probability that sbp participants will pay more into the sbp program than they ever receive, despite its federally-subsidized status. this finding surprises us, given that when asked, “is sbp a good buy?”, the military pay and benefits website sponsored by the office of the under secretary of defense for personnel and readiness responds, “. . . the answer for most retirees is yes!”11 having defined the winner as a low-odds 1-in-20 scenario, we now expand the definition and determine who are the “usual winners” of sbp. these are couples in which the military retiree dies after age 66 (i.e., dies later) and the spouse then lives at least 12% longer than the retiree has been retired (i.e., to get past the breakeven point discussed earlier). because these are more likely outcomes, it is critical to consider these cases. by age 67, a beneficiary spouse will have social security options and often other savings (retirement and other), and perhaps a pension from the spouse’s own job. as a result, the sbp decision becomes one of marginal wealth. that is, given the couple’s other endowments (savings) and income (social security and pension), will the marginal sbp protection be necessary in the most probable situations? we contend the answer to this question is: probably not. thus, our results directly counter the argument that one should buy sbp because the expected return to sbp is positive. note that we do not contend pension insurance is universally a bad idea. rather, in contrast to prior findings, we argue that the default position should be against sbp based strictly on the financial calculations so long as other alternatives exist. we detail some of these alternatives next. 156 w.w. jennings et al. / financial services review 27 (2018) 147-172 5. alternatives to the sbp if not sbp or any analogous pension insurance, then what? given the statistical unlikelihood of winning the sbp or even experiencing the previously-advertised positive expected return, what is a (military) retiree to do to protect his or her family? after all, it would be imprudent to forego pension protection in the early retirement years, which programs like sbp provide. and the typical military member remains poised and inevitably wants to act on the information provided here. fortunately, there are some alternatives to consider. 5.1. invest for the long term the first—and most basic—alternative is simply to invest the sbp premiums for longterm growth rather than using it for this insurance that will most likely never pay out. according to davis and fraser’s (2012) analysis, if a retiree were to earn a real investment return of 6.8%, he or she would be equally well-off relative to participating in sbp. however, considering the level of risk required to earn 6.8% real returns on investments compared with earning these returns in an essentially riskless government benefit (i.e., sbp), a fiduciary would be hard-pressed to recommend a pure investment strategy over a virtually riskless sbp strategy. relying on uncertain investment returns to pay off sufficiently when needed presents a risky strategy. additionally, it simply ignores the nonfinancial utility enhancement life insurance can provide. we do not recommend this approach because it does not protect the surviving spouse in the rare but catastrophic situation where the retiree pensioner dies prematurely. even if one assumes investment returns that are exceedingly high by historical standards, portfolio values would likely be insufficient given the short investment horizon. indeed, a pure investment strategy would only be able to provide sbp equivalent protection for a retiree who dies much later in life. truthfully, this option represents a so-called straw man. from a holistic financial planning perspective, it would be imprudent to not protect one’s db pension in the event of an untimely death, especially if one’s dependents rely on the pension for nondiscretionary living expenses. a risky investments strategy is exactly that—risky. 5.2. buy term and invest the difference the more responsible option, therefore, becomes the old financial planning adage: “buy term and invest the difference,” which we call the term-and-invest strategy. one typically sees this phrase when comparing the purchase of whole life, or cash value, insurance to the option of buying level-term life insurance. notably, we do not model a whole life policy option in this analysis. the reasons for doing so are two-fold. first, in our military retiree example, these individuals are already conditioned to paying for a term life insurance policy called servicemembers’ group life insurance (sgli) during their time in service. transitioning to another term policy is a familiar construct. secondly, and more importantly, we are considering an illustrative option that is cost-neutral for the military retiree couple when compared with sbp. whereas cash value insurance guarantees coverage for life, doing so 157w.w. jennings et al. / financial services review 27 (2018) 147-172 involves much higher premiums—on the order of 10 times the cost of term life insurance for the same face value (we depict these premiums in fig. 6). the typical level-term insurance we model has a fixed premium for a fixed amount of insurance benefit for a finite period of time. term premiums and the insurance coverage amounts are nominally fixed but decrease over time in real terms. once the level-term contract expires, if the individual is insurable, the insured can typically renew a policy for higher premiums, lower death benefit, or a combination of the two.12 because both sbp and a term-and-invest strategy provide the utility of having life insurance, it is not necessary to quantify this value explicitly. instead, for this analysis, the relevant considerations are (1) the cost of the sbp coverage, (2) the insurance needs of the retiring military member, and (3) the cost of these needs from a private insurance provider. we once again consider the case of our typical officer, a lieutenant colonel (o-5) retiring after 20 years of service at age 44, who has a 44-year-old spouse. making some reasonable assumptions for these considerations allows us to quantify this termand-invest alternative. using an income replacement strategy, foregoing sbp and using private life insurance means our male officer retiree would need to replace $30,250 in annual sbp income for his nonmilitary spouse. if we use the well-documented bengen (1994) four percent payout rule, the amount of insurance coverage required to provide a real lifetime annuity of $30,250 is approximately $756,000. given recent skepticism of the four percent rule as too aggressive, one might plausibly decrease it to 3.5% or lower. if we use 3.5%, the insurance value increases to $864,000. for this analysis we use $800,000 in coverage as a happy medium, which is conveniently the maximum insurance policy our insurance quote source will write for many active duty officers. we recognize this payout rule is a hotly debated heuristic and do not intend to provide a recommendation here. instead, we leave it to an advisor or planner to conduct sensitivity analyses based on their own philosophies and/or their clients’ situations. note that the four percent payout rule and its kin assume inflation adjustments, just as the sbp does.13 well-known private companies servicing military members provide level-term life insurance quotes via their websites.14 a 44-year-old male nontobacco user in good health can purchase an $800,000, 30-year level-term policy for approximately $125 per month. assuming the insurer has sufficient financial strength, this insurance policy can serve the same need as the revenue stream provided by sbp at less than half the cost of sbp (i.e., $125/month vs. $297/month). it is important to acknowledge that insurance payouts are generally reliable but have more risk than sbp payouts that the federal government backs. therefore, we consider two companies’ products. one company has either the highest or second-highest of 16 or 21 possible ratings from a.m. best, moody’s, and s&p; the other has been in business for over 100 years but does not purchase ratings. in the long-term, a key part of this buy term and invest the difference option is to invest the savings. rather than consuming the $172 monthly savings if purchasing life insurance instead of sbp, one can save and invest these funds to offset the loss of insurance in 30 years when the term life contract expires. we cannot overemphasize the fact that maintaining budget and investment discipline is required for this option to succeed. while it would be appealing to use an expected market return on an investment portfolio, doing so is not 158 w.w. jennings et al. / financial services review 27 (2018) 147-172 consistent with finance theory. although they analyze the pension problem from the payer side (vs. our focus on the recipient), brown and wilcox (2009) cogently argue for using returns that are risk-equivalent when analyzing investment performance. to ensure we are comparing risk-equivalent alternatives, we compound the savings component at three percent annually and use a 2.5% inflation rate to obtain real dollar values from the buy-term-andinvest strategy.15 fig. 5, panel a, shows the present value to the surviving spouse of this term-and-invest strategy under the aforementioned assumptions for all previously considered mortality scenarios. in the early years the value of this strategy comes from the level-term policy. however, because of its nominal $800,000 value, it decreases in real terms until it expires at 30 years. correspondingly, the nominal premium also decreases in real terms over time. the sharp drop off in benefits occurs at the 30-year point when the policy expires. at this point we see the value of the investments increasing in an exponential manner. to be clear, we assume all investment contributions cease upon the retiree’s death. however, in the case of a spouse’s death we assume the retiree continues saving and paying the life insurance premium since he still receives a pension and might have other desired beneficiaries, such as children or charities. also, we see no difference between the winner scenario (i.e., retiree dies young; spouse lives for decades) and the scenario where both partners die young. this present value equivalent payout is independent of the surviving spouse’s life span; it hinges only on when the retiree dies. as we see, dying immediately after the term policy expires (age 74�) presents the worst situation. at this point the investment amount is modest, and the level-term coverage is gone. to the extent one can extend the level-term policy at a reasonable rate or buy a level-term policy that extends beyond 30 years, these might represent even more appealing options. the overall expected net payoff for the term-and-invest strategy is $162,214 in our baseline scenario, $100,946 greater than the expected net payoff ($61,268) of the sbp option. a subsequent figure summarizes these expected net payoffs for many retiree-spouse age combinations. the next figure depicts the horse race, directly comparing payouts to the sbp and the term-and-invest strategy. fig. 5, panel b, expands the earlier figures by differencing fig. 5, panel a, and fig. 3. that is, it shows the net payout of term-and-invest minus sbp for various mortality combinations. anywhere the plot is positive, the term-and-invest strategy bests the sbp strategy, and vice versa. fig. 5, panel c shows the mortality probability weighted version of the panel b. some valuable observations emerge from these figures. first, when sbp really matters— that is, if a retiree dies early and the beneficiaries need immediate funds to pay off a mortgage, and so forth, then the term-and-invest option has a higher present value than sbp (see circle in panel b). second, there is once again a large payout reduction in the term-and-invest option when the level-term policy expires. this result induces us to recommend that a prospective retiree consider stretching the term policy as far as possible within reason. next, there clearly remain mortality combinations where the individual is “lesscovered” by insurance than by sbp, shown in panel c by the negative expected payout areas of the contour. however, these points occur later in life, at around age 75, when one would expect other income streams—like retirement accounts, pensions, or social security—to 159w.w. jennings et al. / financial services review 27 (2018) 147-172 fig. 5. panel a: payout for the buy-term-and-invest strategy. this plot shows, for a variety of mortality combinations, the present value of payouts for a strategy of buying a term insurance policy and investing the difference between the cost of the term insurance policy and sbp. the term and invest strategy assumes an $800,000 30-year policy costing $125/month, 3% return on investments, and 2.5% inflation. panel b: payout difference between the term-and-invest strategy and survivor benefit plan (sbp). this plot shows, for a variety of mortality combinations, the difference in payouts between sbp and a strategy of buying a term insurance policy and investing the difference between the cost of the term insurance policy and sbp. if positive, the term/invest policy has a higher payout than sbp. the term and invest strategy assumes an $800,000 30-year policy costing $125/month, 3% return on investments, and 2.5% inflation. panel c: expected payout difference between the term-and-invest strategy and sbp. this plot shows, for a variety of probabilistically determined mortality combinations, the difference in payouts between sbp and a strategy of buying a term insurance policy and investing the difference between the cost of the term insurance policy and sbp. if positive, the term/invest policy has a higher expected payout than sbp. the term and invest strategy assumes an $800,000 30-year policy costing $125/month, 3% return on investments, and 2.5% inflation. 160 w.w. jennings et al. / financial services review 27 (2018) 147-172 provide some financial support. additionally, by investing, a retiree becomes self-insured much later in life because of the compounded investment returns (see the ramp as the member’s age increases). overall, during the highest-probability mortality combinations, the term-and-invest strategy is preferred over the sbp db insurance program. critically, unlike the sbp, the value of the term insurance and the investments remains intact even if the spouse dies. that is, the term policy and investments are available to children, charities, or any future beneficiaries. besides covering the retiree’s family during the vital early years, the term-and-invest option provides almost certain non-negative future wealth, regardless of the mortality order of the retiree and spouse. finally, and very importantly, these values represent the marginal benefits (and costs) of the term-and-invest strategy relative to sbp. even in those mortality combinations where the term-and-invest strategy underperforms relative to the sbp strategy, it does not mean the military retiree’s spouse is without financial resources. on the contrary, fig. 5, panel a shows that in these very scenarios the widow does indeed receive positive payouts in many of the scenarios. the reason these payouts are negative in panel b is because the payouts in the sbp winner scenario are exceedingly high. in summary, relative to sbp, a term-and-invest strategy has many advantages. to recap the two options succinctly, for the couple identified here, there is a 60% probability they will pay into sbp more than they ever receive, despite the program’s government subsidy and despite the program’s expected positive real returns. conditional upon falling into this 60% net loss scenario, the couple can expect a net “loss” of $55,232. on the other hand, if they happen to see any sbp payout, which has a 40% probability of occurring, their conditional expected payout is $116,500. while this proposition might sound reasonable in isolation for an insurance product, it loses appeal relative to the term-and-invest alternative. in the same 60% of the time that the couple would lose with sbp, their conditional expected payout under term-and-invest is $85,421; for the 40% of the time that fig. 5. (continued). 161w.w. jennings et al. / financial services review 27 (2018) 147-172 sbp pays off positively, term-and-invest pays off as well, with $76,792. all of these figures involve comparable risk, which is the appropriate way to analyze financial decisions. predictably, the differences become even more biased toward the term-and-invest strategy’s benefit if one posits real investment returns higher than 0.49%.16 having quantified the most common retiree and spouse scenario and presented the case for a term-and-invest strategy, it is important to understand there are cases where sbp clearly makes financial sense. using the same methodology presented in the previous scenario, fig. 6 summarizes other selected officer retiree-spouse age combinations at retirement. with estimates of term insurance premiums and insured sbp coverage amounts, it provides the expected values of sbp and the term-and-invest strategy, along with the probability of receiving more from sbp than paid in premiums. given the relative life expectancies, it is evident that military retirees with younger spouses should strongly consider purchasing the sbp program. the shaded combinations show the retiree-spouse age combinations where the expected payout to sbp exceeds that of a possible term-and-invest strategy. if a retiree and spouse differ in age by approximately six years, they will want to look closely at sbp, as eight (four) years difference makes sbp (term-and-invest) the dominant strategy. fig. 6. expected value of survivor benefit plan (sbp) for various retiree-spouse age combinations at retirement. this table shows the expected value based on life expectancies for the sbp program available to military retirees and their spouses, e(sbp). it also shows the expected value for a term insurance plus investment strategy, e(t&i), where commercial level term life insurance is purchased and the difference between these premiums and sbp premiums is invested in risk-free treasuries. it also shows the probability that the couple will receive sbp payouts that surpass premiums, pr(sbp � 0). the age combinations represent the retiree and spouse ages at the time of retirement. the final row shows the amount of annual retirement pay the retiree insures. shaded cells represent retiree-spouse combinations where the expected sbp payout is greater than the expected term plus investment payout, or e(sbp) � e(t&i). 162 w.w. jennings et al. / financial services review 27 (2018) 147-172 5.3. mentor those junior to you about the sbp decision for those with clients who have already chosen sbp, it is important to stress that it is not an ineffective or malicious program. it does provide income protection for the spouse should the retiree predecease the spouse. sbp is an annuitization strategy that goes a long way in addressing the longevity risk of a surviving spouse. for many, sbp is a viable solution to the annuity puzzle, where too few retirees annuitize retirement wealth as life cycle models would suggest they should (yaari, 1965). sbp largely addresses an inadequacy in life insurance that leaves many widows experiencing a dramatic reduction in standard of living (auerbach and kotlikoff, 1991). with potentially better options available to military retirees, we want to urge mentorship of future generations about their choices. the finra (finra investor education foundation, 2012) study entitled, “financial capability in the united states, 2012 report of military findings,” finds that many active military members fail to plan for retirement. sharing this sbp analysis with individuals earlier in the preretirement phase of their careers may be beneficial. in some cases, doing so will allow them to buy level-term insurance at lower rates and while they are insurable. as good as they are at strategic, operational, and tactical military planning, many military members allow retirement planning—including the sbp decision—to flank them. we posit that young workers with private db plans could benefit from similar mentoring. finally, conflicting incentives converge at retirement, making an early insurance decision even more important. typically, military members are incentivized to report every ailment at retirement for disability qualification purposes under veterans’ administration (va) benefits programs. some of these ailments could clearly affect one’s insurability for any program besides sbp or the military’s term insurance counterpart, veterans’ group life insurance (vgli), neither of which requires a medical exam. therefore, making the “to sbp or not sbp” decision earlier could prove tremendously beneficial beyond the figures shown here. we recommend that service members specifically—and db recipients more generally—consider the pension insurance decision two to five years before retirement. 6. caveats and additional considerations we have presented a scenario-based analysis using the most common officer retirement scenario. this means unambiguously that our recommendations are not one-size fits all. instead, they come with the following caveats and additional considerations. the online calculator provides individual-specific insights that this paper does not depict. 6.1. competent financial planning first and foremost, this sbp decision is a tactical one that is part of an individual’s and family’s broader financial strategy. having a comprehensive strategy is absolutely critical and the sbp decision should complement this bigger picture. it would be tragic to see 163w.w. jennings et al. / financial services review 27 (2018) 147-172 destitute war widows and widowers win the tactical financial skirmish but lose the broader campaign. 6.2. taxes as mentioned earlier, our analysis ignores taxes. the basis for this decision is that sbp premiums are paid pretax, yet benefits are taxed. conversely, term life insurance payments are taxed; the proceeds are not. the tax status of investments varies based on whether they are in tax-advantaged retirement accounts or not, and if so, which flavor (e.g., traditional or roth). for these reasons, we urge readers to perform a situation-specific tax analysis before making any final decisions.17 see the prior recommendation. 6.3. personal differences clearly our scenario-based analysis analyzes a particular couple—again, we understand it to be the most common officer retiree scenario inasmuch as one exists. obviously, individual retiree and spouse health at retirement varies and the couple’s mortality expectations can deviate significantly—both positively and negatively—from the actuarial table estimates. for instance, because of a medical condition, it is possible the retiree cannot get private insurance. one needs to make this determination before declining sbp. 6.4. age differences marked age differences can certainly swing this analysis in favor of the sbp decision, particularly if the retiree is a male married to a much younger female spouse. in that case it is more likely that the male retiree will predecease his spouse, generating much higher expected payoffs to the spousal beneficiary. fig. 6, as well as our online calculator can help here. 6.5. children along similar lines as the previous caveat, retirees with young children can consider opting for the child-only benefit portion of the sbp. this a la carte option permits the retiree to buy sbp at a very low rate (e.g., on the order of $10–$20 per month) with children as the beneficiaries. eligible children share the full 55% benefit equally until they turn 18, or 22 if a student, at which time they become ineligible. if both spouse and child benefits are elected, then children are not eligible for these benefits until the spouse is ineligible through remarriage or death. in short, when a retiree’s youngest child is very young, the sbp child option can serve as an inexpensive partial substitute for the private insurance coverage in the term-and-invest option we present. notably, untabulated analysis shows the child-only coverage is not actuarially favorable to participants; however, the premiums can be a small price to pay for the peace of mind this coverage provides. 164 w.w. jennings et al. / financial services review 27 (2018) 147-172 6.6. the “talk” as discussed early on, the sbp is an opt-out program. a married member who wishes to decline coverage must get their spouse’s approval, whereby the spouse recognizes this decision to forego the annuity upon the member’s death. anecdotally, we cannot overstate the difficulties associated with this decision if a couple has not had prior conversations about their plans. as part of this talk, a spouse must recognize the potential implications of foregoing sbp. for instance, if a couple declines sbp and subsequently divorces, the spouse could lose control over the insurance and investments unless they protect themselves carefully in the divorce proceedings. it is important the spouse recognize this potential liability. narrowly, we recommend conversing with one’s spouse early about the sbp decision. more broadly, we encourage couples to discuss general financial planning matters early and often! 6.7. investment and budgetary discipline discussing this article with others, including insurance sales representatives, made us consider putting this item first on the list. as one astute reviewer highlighted, the term-andinvest alternative depicted here relies on the invest component of the strategy, particularly later on when returns are compounded over decades. behaviorally, it takes discipline to keep investing the difference versus enhancing one’s quality of life. unfortunately, for some individuals it is very easy to stop paying a term policy premium or stop investing the difference if a new car or house upgrade beckons. if a couple is not committed to investing the difference, then they might consider sbp more carefully. despite the low return (i.e., risk-free rate) assumptions in our scenario, the investments represent fully 75% of the term-and-invest strategy’s value. thus, a term insurance only strategy underperforms sbp ($40,700 expected payout vs. $61,300 for sbp), which is not surprising given the sbp subsidy. the real surprise is that a competitively priced life insurance policy from a reputable insurer competes quite well with sbp at this time. furthermore, if allowed to deviate from risk-equivalent returns, a broad market portfolio of investments will perform much better over decades than shown in our analysis. additionally, a beneficiary of a lump-sum insurance payout requires budgetary discipline to ensure the funds last as if they were an annuity. financial advisors can assist with this discipline aspect . . . again, see the first recommendation. 6.8. assumptions we make known our assumptions about our scenario, interest rates, and inflation. however, we must add this item for completeness. of course, the outcomes are sensitive to these assumptions. caveated sufficiently, we do stress that the essence of our recommendations are not highly sensitive to changes in these assumptions. 165w.w. jennings et al. / financial services review 27 (2018) 147-172 6.9. marital status the younger a widow(er), the more likely he or she is to remarry. fig. 7 shows these probabilities. if an sbp recipient widow(er) remarries before age 55, then the sbp benefit disappears. consequently, if we added these probabilities of remarriage into this analysis, the expected payouts to the winners (i.e., the long-lived spouse whose sbp-insured spouse died young) would decrease. adding this caveat would on-balance, enhance the appeal of the term-and-invest strategy relative to sbp. for divorced retirees, higdon (2009) outlines multiple family law cases involving sbp. another application of this analysis results from these legal situations: divorced service members might find a level-term life insurance policy a more financially appealing way to provide for their ex-spouse’s insurable interest in the service member’s retirement pay. 6.10. credit quality (both ways) directly comparing our term-and-invest strategy with sbp assumes the private insurer has credit quality equal to the u.s. government. we could avoid this assumption by adjusting the insurance payouts with a discount rate reflecting this risk. we have chosen not to do so at this time because (1) the private insurers we quote have top ratings and (2) we do not think using such a discount rate would necessarily obviate any related criticism. 6.11. female military member for brevity, this article focused on the most common member-spouse gender scenario; clearly others exist. obtaining by-gender mortality data from the dod actuary would facilitate this relevant analysis. given the relative longevity of military members versus their spouses along with the typical male-female age differences in marriage we suspect that our term-and-invest strategy dominates sbp in many female military member cases. fig. 7. probability of a widowed spouse remarriage. this plot shows the probability of a widowed military spouse remarrying at a given age conditional upon the age at which they were widowed (i.e., 35-, 38-, or 41-years-old). if remarrying before age 55, a widowed spouse foregoes future survivor benefit plan (sbp) benefits. 166 w.w. jennings et al. / financial services review 27 (2018) 147-172 6.12. death indemnity compensation (dic) offset and servicemembers group life insurance (sgli) currently, the government’s death indemnity compensation program compensates beneficiaries of military members who die while on active duty or from a disability compensable under veterans administration laws. this dic currently offsets spousal sbp payments dollar-for-dollar such that in the aggregate beneficiaries receive up to the maximum of either dic or sbp benefits. we have excluded this dic offset impact in our analysis. however, if we included it, sbp becomes even less appealing than we portray here, as it is theoretically possible to pay sbp premiums but never receive an sbp benefit if the dic maximum payment subsumes the sbp benefits. recent military compensation reform proposals recommend allowing retirees the option of removing this offset; however, in doing so retirees would pay higher sbp premiums. the net effect can be addressed when it materializes. relatedly, servicemembers should know that their sgli coverage that they carry on active duty continues for 120 days after retirement at no cost. currently, the maximum coverage is $400,000. we do not model this coverage, but if we did, it would temper the relative benefit of sbp (or strengthen the term-and-invest option) in those scenarios where the servicemember dies within four months of retirement. 7. conclusion tens of millions of retirees—state employees, federal employees, and even private db pensioners—face decisions about whether and how to insure their defined benefit pension. the u.s. military has a transparent and well-defined pension protection program called the sbp. the sbp presents a unique opportunity to present a framework for analyzing this key financial decision. prior analysis has simulated the payoffs for various scenarios, ultimately concluding that the sbps average real return is quite remarkable, between six and seven percentage for a common scenario. while we concur that this average return makes the sbp program very attractive, we update the results using more relevant and appropriate life expectancy data. we show that a 1-in-20 winner scenario is embedded in any rate of return calculation, masking the highly skewed distribution of outcomes. we ultimately recommend a default position against the federally-insured pension protection program for the typical retiree given currently available market alternatives. one alternative that is currently more appropriate for the typical military retiree considering sbp-like programs is to buy privately-available level-term insurance and invest the difference between the term cost and the sbp premiums. based on an illustration of a common retireespouse combination and risk-equivalent return assumptions, a strategy of buying term insurance and investing the savings generally provides more preferred outcomes than sbp. our analysis couples the department of defense actuary life expectancy tables with realistic demographic and economic assumptions to first determine the expected payoffs to the sbp program. it turns out that over 60% of sbp clients should expect to pay more in premiums than they ever receive. while we might anticipate this breadth of expected negative returns with private insurers because of profit and overhead requirements, it is somewhat surprising 167w.w. jennings et al. / financial services review 27 (2018) 147-172 given the sbp is federally subsidized and endorsed. further, the sbp insures relatively healthy individuals—u.s. military members who have served at least 20 years—as evidenced by their required preventive health exams and routine physical fitness tests. this study further characterizes the expected sbp winners versus typical recipients, finding that while the winners (i.e., the insured dies young and the beneficiary lives long) occur only a small fraction of the time (5%), almost 60% of the sbps expected value goes to the surviving spouse in these couples (as noted in the paper, we do not mean literal fig. 8. results from survivor benefit plan (sbp) calculator for one common scenario. this figure shows the inputs available and output summary from a calculator that permits individuals to assess the sbp versus term-and-invest pension protection strategies given their unique circumstances. results are based on militaryretiree specific mortality tables. calculator is available at www.financialcheckpoints.com. 168 w.w. jennings et al. / financial services review 27 (2018) 147-172 winners. we genuinely hope nobody “wins” early sbp payments). in contrast, the typical sbp recipients receive their payouts relatively late in the game for relatively few years, at a time when they likely have other sources of income besides the sbp annuity. for these reasons, we explore an alternative to sbp, which is buying term insurance and investing the difference between the cost of the term policy and the sbp premium. this term-and-invest strategy is superior to the sbp in many situations. it protects the surviving spouse from an unexpected early death of the insured retiree. unlike sbp, the term insurance policy benefits are not conditional upon the beneficiary spouse’s mortality. in other words, children, charities, or the retiree’s estate will still benefit if the retiree’s spouse passes away. the investments also provide a certain benefit to the spouse, children, or couple whether the retiree predeceases them early or not. in fact, in a common scenario of a 44-year-old male military officer retiring with great health and under plausible nominal investment return (7%) and inflation (2.5%) assumptions, the expected payout to a term-and-invest strategy is $330,000, which is over five times the expected payout from sbp. a summary of various scenarios is found in fig. 6, which comes from our calculator that we provide online. fig. 8 shows a snapshot of this calculator. in closing, it is universally important to recognize certain things about sbp: y the expected returns are more skewed than previously documented, y well over one-half of retirees in common situations are expected to pay more in premiums than their spouse collects; y from a holistic financial planning perspective, the most likely mortality scenarios make sbp largely irrelevant; and y in circumstances where sbp is most valuable, there exist other mechanisms to provide similar, or even enhanced, financial security. yet, because of the caveats we list, which include taxes, individual-unique circumstances, behavioral considerations, our assumptions, and so forth, the sbp decision is truly situationspecific. therefore, we recommend readers use this article’s framework as a baseline for discussing this important topic with their own clients or planners before making a potentially life-altering decision. we recommend having this conversation two to five years before retirement. for our part, we are eating our own cooking: those of us who are eligible have purchased term life policies and declined sbp.18 notes 1 the analysis in this article clearly assumes that the couple needs to protect the db recipient’s pension out of a need for the income replacement. this would not be the case if, for instance, the couple had substantial wealth (e.g., recent inheritance or longstanding trust) or income (e.g., spouse earns high income independently) that would provide for the family’s needs upon the death of the db recipient. 2 although the outcome is still an actuarial loss to the military retiree, the one exception we would recommend is strong consideration of the a la carte child coverage when military retirees have young children. 169w.w. jennings et al. / financial services review 27 (2018) 147-172 3 this option is now changing under the new blended retirement system. for analysis of this new option, see payne et al, 2018. 4 while lesser amounts of coverage for the beneficiary are possible, it is worth comparing the benefits of comparable private insurance that we describe later. details on this lesser coverage and benefit appear here: http://militarypay.defense.gov/benefits/ survivor-benefit-program/costs-and-benefits/spouse-coverage/. 5 source: “valuation of the military retirement system,” september 30, 2013 and revised january 2015, produced by the department of defense (dod) actuary (department of defense office of the actuary, 2013). note that we had to extrapolate mortality probabilities for retirees beyond age 109, as that is the age the dod actuary ended its tables. we assumed a maximum lifespan of 120 years, with exceedingly low probabilities of living beyond 109. our extension was informed by soa rp-2014. we obtained social security data from the social sec. office of the chief actuary 2014. 6 http://actuary.defense.gov/, “statistical report on the military retirement systemseptember 30, 2013,” table: military retirees by gender and branch of service as of september 30, 2013. 7 http://www.militaryonesource.mil/12038/mos/reports/2012_demographics_report. pdf, “2012 demographics: profile of the military community,” table 2.57: number and percentage of active duty officers and enlisted members in dual-military marriages by service branch. 8 some (e.g., jennings and reichenstein, 2003) might argue we should use the tips rate for the real discount rate; however, staying on the other side of that argument makes our analysis more conservative (i.e., more favorable to sbp than it would be if we used a tips rate). the online calculator permits the user to make this adjustment. 9 for the officer, monthly premiums equal 6.5% � $55,000/12 � $297.92 � 72 months � $21,450. annual benefits are 55% � $55,000 � $30,250 � 20 years � $605,000. the enlisted calculations are identical, except they begin with a $2,100 per month pension benefit. 10 for benefits to outweigh costs, the ratio of spouse lifespan after retiree death/insured retiree lifespan after retirement must be greater than the ratio of total premiums � 6.5%/total benefits � 55%, which equals 0.12. thus, the spouse lifespan after retiree death must be greater than 0.12 � insured retiree’s lifespan as a military pension recipient and sbp premium payer. 11 http://militarypay.defense.gov/benefits/survivor-benefit-program/overview/ 12 because pfau (2015) makes a compelling case for considering whole life insurance as a part of holistic retirement planning, we emphasize our recommendation that couples seek competent financial advice when considering this option. 13 as an astute reviewer pointed out, if one were not comfortable assuming a 3.5– 4% payout assumption since bengen (1994) evaluates a risky investment portfolio, then doing so would require upward adjustments for insurance premiums and downward adjustments for the invested difference relative to what we present here. we do not explore these alternatives in this analysis, but they are easy to compute using the accompanying online calculator. the authors could also provide them upon request. 170 w.w. jennings et al. / financial services review 27 (2018) 147-172 14 we also calibrated these costs with a broad group of general life insurance providers to ensure their reasonableness. the military-centric coverage is more expensive, presumably because of the consideration of such factors as high-risk work and the absence of war exclusion clauses. evaluating military-specific policies is, therefore, conservative. 15 we assume 2.5% inflation and 3% return on investments for a 0.49% real return on investments. doing so is consistent with the jennings and reichenstein (2003) approach to valuing defined benefit pensions. note the results are relatively less sensitive to these return assumptions than they are to the insurance coverage amount, term, and cost. 16 for instance, assuming a reasonable investment portfolio return of 7% (4.39% real), the conditional expected payouts for the term-and-invest scenario become $125,665 and $205,335 in the 40% and 60% spaces, respectively, for a total expected payout to term-and-invest of $331,000. 17 using the illustrative example in the article, we assume a 25% pre-death income tax rate (federal and state), 17% post-death income tax rate (federal and state), and 15% capital gains rate and that any investments are in non-tax-advantaged accounts. under these conditions, the expected value of sbp decreases by $2,620 or 4.3% while the expected value of term-and-invest decreases by $3,465 or 2.1%. the magnitude of the tax impacts is similar under both options. obviously, there are countless permutations based on individual circumstances, leading us back to the first consideration in this list. 18 full disclosure: one of us with young children opted for the child-only option. acknowledgment the views in this article are solely the authors’ and do not reflect those of the u.s. government, department of defense, u.s. air force, or u.s. air force academy. it was previously circulated as “alternate options to the survivor benefit plan: important considerations for us veterans.” we gratefully acknowledge comments and assistance with this project from participants at the 2014 and 2017 academy of financial services annual meetings, seminar participants from the department of management at the u.s. air force academy, lieutenant colonel (retired) stephanie bruce, and lieutenant colonel joseph suhajda. references ando, a., & modigliani, f. (1963). the “life cycle” hypothesis of saving: aggregate implications and tests. the american economic review, 53, 55–84. auerbach, a. j., & kotlikoff, l. (1991). life insurance inadequacy-evidence from a sample of older widows (working paper no. 3765). retrieved from national bureau of economic research website: https://www. nber.org/papers/w3765.pdf. 171w.w. jennings et al. / financial services review 27 (2018) 147-172 beatty, s. e., & kang, t. (2007). the survivor benefit plan: a cost-benefit analysis mba professional report, thesis and dissertation collection, monterey, california. (available at https://calhoun.nps.edu/handle/10945/ 10185) bell, d., & graham, a. (1984). surviving spouse’s benefits in private pension plans. monthly labor review, april, 23–31. bengen, w. p. (1994). determining withdrawal rates using historical data, journal of financial planning, october, 171–180. brown, j. r., & wilcox, d. w. (2009). discounting state and local pension liabilities. american economic review, 99, 538–542. comprix, j., & muller, k. (2011). pension plan accounting estimates and the freezing of defined benefit pension plans. journal of accounting and economics, 51, 115–133. davis, k., & fraser, s. p. (2012). insuring defined-benefit plan value: an examination of the survivor benefit plan (sbp) decision. financial services review, 21, 275–290. department of defense office of the actuary. (2013). valuation of the military retirement system. september 30, 2013, revised january 2015. (available at https://actuary.defense.gov/portals/15/documents/mrf% 20valrpt%202013.pdf?ver�2015-02-24-102857-617) finra investor education foundation. (2013). financial capability in the united states: 2012 report of military findings. (available at http://www.usfinancialcapability.org/downloads/nfcs_2012_report_ military_findings) higdon, j. n. (2009). the survivor benefit plan: its history, idiosyncracies, coverages, cost, and applications. family law quarterly, 43, 439–488. jennings, w. w., & reichenstein, w. r. (2001). the value of retirement income streams: the value of military retirement. financial services review, 10, 19–35. jennings, w. w., & reichenstein, w. r. (2003). valuing defined-benefit plans. financial services review, 12, 179–199. kitces, m. (2014). insurance rule: buy the losers. financial planning, december, 73–76. lachance, m. e., mitchell, o., & smetters, k. (2003). guaranteeing defined contribution pensions: the option to buy back a defined benefit promise. journal of risk and insurance, 70, 1–16. military compensation and retirement modernization commission. (2015). report of the military compensation and retirement modernization commission. final report. (available at https://www.ngaus.org/sites/default/ files/mcrmc%202015_0.pdf) payne, b. c., jennings, w. w., & o’malley, t. c. (2018). implied discount rates for the u. s. military blended retirement system. journal of retirement, 6, 22–32. pfau, w. d. (2015). optimizing retirement income by combining actuarial science and investments. owrs. (available at http://www.owrsfirm.com/wp-content/uploads/oa_wp_opt-ret_inc_05–15_web.pdf). poterba, j. m., rauh, j., venti, s., & wise, d. (2007). defined contribution plans, defined benefit plans, and the accumulation of retirement wealth. journal of public economics, 91, 2062–2086. richwine, j. (2013). nine fallacies used to defend public-sector pensions. backgrounder, the heritage foundation, no. 2765, washington, dc. (available at https://www.heritage.org/social-security/report/ninefallacies-used-defend-public-sector-pensions) social security office of the chief actuary. (2014). statistical tables, period life table. (available at https:// www.ssa.gov/oact/stats/table4c6.html) yaari, m. e. (1965). uncertain lifetime, life insurance, and the theory of the consumer. the review of economic studies, 32, 137–150. 172 w.w. jennings et al. / financial services review 27 (2018) 147-172 financial literacy, attitudes, and financial satisfaction: an assessment of credit card debt-taking behavior of australians muhammad s. tahira,*, daniel w. richardsb, abdullahi d. ahmedc acollege of business, rmit university, melbourne, victoria, australia 3000 bcollege of business, rmit university, melbourne, victoria, australia 3000 ccollege of business, rmit university, melbourne, victoria, australia 3000 abstract unpaid credit card debt can be problematic; people should avoid it where possible. unlike prior studies, this article examines the relative strength of the association of financial literacy, attitude toward balancing spending and savings, and financial satisfaction with credit card debttaking behavior by analyzing the 2016 wave of the household, income and labor dynamics in australia (hilda) survey. we find that higher financial literacy is associated with less credit card debt. however, incorporating the other factors reduces this relationship. our results advise policy-makers to include components in the financial literacy curricula that encourage savings attitude to reduce problematic debt-taking. © 2020 academy of financial services. all rights reserved. jel classification: g51; g53; g40; d14; d10 keywords: hilda; financial literacy; credit card debt; financial satisfaction; savings attitude 1. introduction credit cards provide convenient shopping facility enabling consumers to utilize their future income. however, in case of future income volatility, the repayment might be delayed allowing financial institutions to charge high interest on the due amount. recent research *corresponding author. +61 3 9925 1694. e-mail address: muhammad.tahir@rmit.edu.au 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 273–301 undertakings highlight factors associated with credit card debt. gorbachev and luengoprado (2019) show that those with savings in low-interest liquid assets and no credit card debt are more financially literate than those who simultaneously hold low-interest liquid assets and high-interest credit card debt. lin, revindo, gan, and cohen (2019) find that those who prefer to make payments using a card are more likely to have more credit card debt as compared with those who use other modes of payments. they find that a person’s attitude towards money is more related to credit card debt than the person’s demographic factors, including gender, education level, and employment type. in this article, we extend research on credit card debt. we explore the relative strength of the association of financial literacy, attitude towards balancing spending and savings, and financial satisfaction with credit card debt-taking behavior in australia. credit card use is topical in australia because the australian securities and investments commission (asic, 2018) explicates that credit card debt is continuously rising. according to the australian bureau of statistics (abs), 74% of australian households held debt in 2015–2016 (abs, 2017b) and credit card debt was the most common type of household debt with 55% of households holding it. according to the asic (2018), more than 14 million credit card accounts (with 21.4 million cards) exist in australia at june 2017 with a total outstanding balance of almost $45 billion. out of the total outstanding balance of $45 billion, the overdue amount is $31.7 billion being eligible for interest charges. accordingly, financial institutions charged almost $1.5 billion in annual fees and late payment surcharges to the credit card users in 2016–2017 (asic, 2018). the asic (2018) shows that 18.5% of credit card holders are in a problematic condition due to either severe/serious delinquency or having persistent debt or low repayments. from 2003–2004 to 2015–2016, the abs (2017a) found an increase of 41% in the credit card debt of middle and high wealth households and a 25% increase in the credit card debt of low wealth households. households, like businesses, can access many types of debt including mortgages, personal loans, credit card debt, student loans, vehicle loans, and so forth (cassells, duncan, kelly, & ong, 2015). understandably, not all types of debt have a negative impact on households. in the academic literature, a method of distinguishing whether a debt is problematic is whether it has been collateralized (secured by an asset) or not (berger, collins, & cuesta, 2016; harari, 2018; tippett, 2010). dunn and mirzaie (2016) contend that non-collateralized debt is more stressful for households than collateralized debt. when households do not have any collateral to surrender, lenders or collection agencies can adopt aggressive behavior in collecting debt, and this increases stress felt by households (dunn & mirzaie, 2016). among non-collateralized debt types (that include credit card debt, student loans, pay day loans etc.), dunn and mirzaie (2012, 2016) argue that credit card debt is relatively more stressful and problematic due to the added penalties (high interest rates) in addition to the aggressive behavior of the collection agencies. furthermore, debt taken beyond the means to repay it also becomes stressful and problematic (cecchetti, mohanty, & zampolli, 2011). this implies that non-collateralized debt (especially credit card debt) and over indebtedness should be avoided as they can have negative consequences such as exorbitant repayments and interest costs that become burdensome for households to meet. other 274 m. s. tahir et al. / financial services review 28 (2020) 273–301 negative consequences of problematic debt and over indebtedness include health problems (jacoby, 2002), mental-illness leading to the attempt of suicide (turunen & hiilamo, 2014), depression and psychological disorder (richardson, 2013), heart malfunction, ulcers, and migraine headaches (jarl, cantor-graae, chak, sunbaunat, & larsson, 2015). therefore, this article specifically researches credit card debt-taking because it is a problematic type of debt-taking. credit cards facilitate purchasing of household goods with the option of “buy now, pay later.” however, when income is volatile, households often utilize savings to meet their consumption needs (kaplan & violante, 2014). a risky income stream and fewer savings may result in an increased credit card repayment liability. an overdue credit card amount often leads to a high associated interest charge. research shows that those who have fewer savings and are less financially satisfied, are more vulnerable to economic shocks and tend to access more credit (reyers, 2019). further, those who are less financially literate and are not aware of the credit market terms, are most likely to take on credit card debt. research associates low financial literacy with a high probability of taking on debt (brown, grigsby, van der klaauw, wen, & zafar, 2016; disney & gathergood, 2013; norvilitis et al., 2006; ottaviani & vandone, 2018). however, to our knowledge, little research is published that empirically explores the relative weight of the association of financial literacy, attitude towards balancing spending and savings, and financial satisfaction with credit card debt-taking behavior. unlike prior studies, this study researches the relative weight of the association and analyzes the 2016 wave of household, income, and labor dynamics in australia (hilda) survey, which reflects australian household financial behavior and many other household characteristics.1 in our empirical analysis, we find a negative association between financial literacy and credit card debt-taking behavior. in the sensitivity analysis, where we include respondents’ attitude towards balancing spending and savings and respondents’ financial satisfaction, the association between financial literacy and credit card debt-taking behavior remains negative. a comparison of the various factors using the average marginal effects (ames) reveals that attitude toward spending and savings has the strongest relationship with credit card debt-taking behavior, followed by financial satisfaction and financial literacy, respectively. this implies that financial literacy has a relatively weaker magnitude of relationship with credit card debt-taking behavior, yet remains a significant factor to have an association with credit card debt-taking behavior. furthermore, the analysis shows that the concept of simple interest (one of the five financial literacy concepts) has a highly negative association with credit card debt-taking behavior. the research has policy implications suggesting that financial education has some, but limited association. financial literacy curricula that include financial behaviors, such as attitude towards balancing spending and savings will be more influential at reducing credit card debt-taking behavior. the rest of the article reviews relevant prior studies, explains the data and method of data analysis, presents empirical results, discusses these results, and concludes with a direction for future research. m. s. tahir et al. / financial services review 28 (2020) 273–301 275 2. literature review 2.1. financial literacy and financial decisions as debt-taking is a part of household financial decisions, our review includes studies that are related to explaining the relationship between financial literacy and financial decisions. literature specifically related to credit card debt-taking behavior is considered, but few studies focus on this. remund (2010, p. 284) provides a synthesized definition of financial literacy in the following words: “financial literacy is a measure of the degree to which one understands key financial concepts and possesses the ability and confidence to manage personal finances through appropriate, short-term decision-making and sound, long-range financial planning, while mindful of life events and changing economic conditions.” this definition addresses that financial literacy helps to improve household financial decisions (alhenawi & elkhal, 2013). one research avenue has been to investigate if financial literacy is related to improved financial planning for retirement. lusardi and mitchell (2007) survey those who are above 50 years of age to know about their financial literacy skills and retirement preparedness. they find that most of the respondents who are unable to answer the basic financial literacy questions are unprepared for their retirement. as a result, they conclude that being financially illiterate may be a reason behind retirement unpreparedness. martin and finke (2014) find that the use of a financial planner has a large impact on retirement preparedness. moreover, van rooij, lusardi, and alessie (2011) analyze if financial knowledge and retirement planning are associated with each other. they utilize data from the netherlands and find a strong positive association between financial knowledge and retirement planning. financial literacy is related to other apt financial behaviors. worthington (2006) uses australian data to explore the association of demographic, socioeconomic, and financial factors with the financial literacy of australian households. among other results, he finds a positive association between higher levels of mortgage debt and financial literacy. in addition, he also finds a positive association between higher levels of household savings and financial literacy. similarly, davutyan and öztürkkal (2016) study turkish households and find that an increase in literacy level increases the probability of saving more for unseen future needs and, therefore, reducing the use of debt in emergency situations. moreover, grinstein-weiss, spader, yeo, key, and freeze (2012) contend that financial literacy plays a pivotal role in shaping the future financial behavior of individuals. they find that those who are taught financial literacy and financial skills because their childhood are more likely to avoid loan delinquency in the future. alhenawi and elkhal (2013) analyze data from the united states and find that those who acquire financial knowledge through formal academic experience secure developed financial planning skills. however, their results also suggest that those who accumulate financial knowledge over time do not have good financial planning skills. lusardi and tufano (2015) develop a survey to test debt literacy skills. they survey individuals from the united states and report a low debt literacy level of the respondents. they 276 m. s. tahir et al. / financial services review 28 (2020) 273–301 further contend that those who do not know the concept of compound interest are more likely to be overindebted. the results are consistent after controlling for demographic factors as well. furthermore, huston (2012) uses data of american consumers to analyze the relationship between financial literacy and the cost of borrowing via mortgage loans and credit cards. their results suggest that financial literacy skills equip a consumer with the ability to minimize the cost of borrowing for both mortgage loans and credit cards. some studies specifically investigate an association between financial literacy and credit card debt. norvilitis et al. (2006) study college students from the united states and find that lack of financial knowledge is related to the credit card debt. brown et al. (2016) use american data of young consumers and find that financial education improves repayment behavior and decreases debt dependence. robb (2011) studies credit card usage behavior of american students and concludes that those with high financial knowledge use their credit cards more responsibly. in summary, extant of the literature suggests that individuals are more likely to engage in high-cost credit and problematic debt (especially credit card debt) when they are less exposed to the credit market terms (e.g., simple interest, compound interest etc.). worthington (2013) states that the concept of financial literacy has been unknown to australians until the end of the 20th century. however, with the start of the 21st century, efforts were started to make australians financially literate. state-level interventions were also introduced, and the australian government initiated the national financial literacy strategy (nfls) in 2011. the motive of the government initiatives was to equip australians with financial literacy skills enabling them to efficiently manage their finances and make informed financial decisions. however, the use of credit cards and credit card debt are still rising in australia (asic, 2018). while research also reports a low level of financial literacy among australians (ali, anderson, mcrae, & ramsay, 2014), we explore the association between financial literacy and credit card debt-taking behavior in australia. we expect that high financial literacy will be related to less credit card debt-taking behavior in australia. we test the following hypothesis: hypothesis 1: financial literacy negatively relates to credit card debt-taking behavior. 2.2. financial literacy, attitude towards balancing spending and savings, financial satisfaction, and financial decisions prior research describes some factors other than financial literacy that could be more relevant to financial decisions. garcı́a (2013) critically reviews prior studies and concludes that financial literacy is an important phenomenon to learn, but the prior beliefs, mental abilities, and cognitive factors dominate financial decision-making. fernandes, lynch, and netemeyer (2014) conduct a meta-analysis of 168 papers covering 201 prior studies. they find that financial literacy and financial behavior are strongly positively associated with each other. however, when the behavioral factors are incorporated in an empirical analysis, the effect size of financial literacy on financial behavior diminishes dramatically. they include five types of financial behaviors in their study (1) saving for an emergency fund, (2) understanding how much is needed for retirement, (3) having a good credit score, (4) m. s. tahir et al. / financial services review 28 (2020) 273–301 277 assessing credit and checking fees, and (5) adopting a positive savings or investment behaviors. fünfgeld and wang (2009) survey 1,282 individuals from switzerland. the first part of their analysis identifies savings attitude and spending attitude as financial behavioral factors. the second part of their analysis finds that the savers make financial decisions analytically by comparing and calculating the risk before making any financial decision. they define savings attitude as an attitude towards savings for future emergency needs. shih and ke (2014) use data from taiwan and explore the relationship between attitude towards money and financial decision-making. their empirical findings show that those who have an attitude of savings and financial planning make risky financial decisions. like other studies, they also focus on one’s attitude towards savings and planning to meet future financial needs. soman and cheema (2002) state that people either utilize their current income in the future in the form of savings or utilize future income in the present. one way to utilize future income in the present is the use of credit cards as credit cards allow people to “buy now, pay later.” if a person uses a credit card and faces income shock, she or he will be unable to make timely repayment. households may utilize their savings to smooth their current consumption pattern when experiencing income volatility (reyers, 2019). however, soman and cheema (2002) argue that easy access to credit facilities allows people to think that they would have comparable earnings in the future and would make timely repayment. here, a need to adjust current spending and saving patterns arise to avoid being in future debt. those who spend more and do not keep a balance between spending and saving end up, usually, struggling with timely repayments. hence, this argument suggests an association between spending and savings attitude and credit card debt-taking behavior. most studies in the literature measure savings attitude as an attitude towards savings for unseen future circumstances. unlike prior research, our focus is on one’s attitude towards balancing spending and savings. essentially, we test the following hypothesis: hypothesis 2: a person’s attitude towards balancing their spending and savings negatively relates to their credit card debt-taking behavior. another important factor with a relationship to financial decisions is financial satisfaction. xiao, sorhaindo, and garman (2006) use data from the united states to research consumer financial behavior. their findings suggest that a lower credit card debt is associated with increased financial satisfaction. contrary to xiao et al. (2006), zhang and kemp (2009) use data from the university of canterbury, new zealand. their empirical analysis suggests that there is no association between student debt and student life satisfaction. students seem satisfied irrespective of whether they have debt or not. however, solis and ferguson (2017) analyze student data from the united states and conclude that students with loans and credit card debt are more likely to be financially dissatisfied. brown and gray (2016) conclude similar results for consumers by analyzing the hilda survey. their empirical analysis finds a negative association between all types of debt and financial satisfaction. similarly, we expect in this article a negative association between financial satisfaction and credit card debt-taking behavior. hypothesis 3: higher financial satisfaction negatively relates to credit card debt-taking behavior. 278 m. s. tahir et al. / financial services review 28 (2020) 273–301 in the context of our study, we expect that both financial literacy and the other factors (attitude towards balancing spending and savings and financial satisfaction) are related to reduced credit card debt-taking behavior. however, we also expect that the relative weight of the association of the other factors with credit card debt-taking behavior is greater than financial literacy. 3. data we use a nationally representative dataset of australia namely hilda. hilda survey is a household panel survey that observes a change in the characteristics and behaviors of the same sample over time (watson & wooden, 2010). commenced in 2001, 18 waves of hilda have been released to the date. initial details of hilda are documented by wooden, freidin, and watson (2002). we use wave 16 for this article. the data for wave 16 were collected in 2016 and access was granted in 2018. wave 16 contains a special module of financial literacy measures along with other regular modules, making it apt for this study. the financial literacy skills were not tested in hilda before wave 16 or after. hence, it makes this study cross-sectional. table 1 lists five items used to measure the financial literacy skills of the respondents. these are objective questions with one correct answer. similar worldwide research studies have also used these measures (cude, chatterjee, & tavosi, 2019; kadoya, khan, hamada, & dominguez, 2018; lusardi & mitchell, 2007, 2011, 2013; van rooij et al., 2011). table 2 shows the percentage of valid correct responses. most of the respondents (85.5%) correctly answer the question related to simple interest, while the concept of inflation got the lowest percentage of correct responses (70.4%). furthermore, it is important to note that those who got assistance to answer these questions are excluded. additionally, we treat “don’t know” as an incorrect answer because these questions imply a correct response. next, we create the “financial literacy” variable by adding the correct responses of these five questions for each participant as applied in prior studies (ali, rahman, & bakar, 2015; xiao & o’neill, 2016). this transformed variable depicts the levels of financial literacy, where “0” represents that the respondent has answered incorrectly to all the five questions, while “5” depicts that respondent has answered all the questions correctly. table 3 below shows the distribution of financial literacy skills among the respondents. this table excludes the missing values, and it shows that around 44% of respondents have answered all the financial literacy questions correctly. this is low in comparison to the average percentage of high financial literacy for oecd countries that was 62% (oecd, 2016). to measure the attitudes of respondents towards balancing their spending and savings, wave 16 of hilda contains the item with the description “i do a good job of balancing my spending and savings.” this item is measured on a 7-point likert scale with option 1 = strongly disagree while 7 = strongly agree. this item has 1,725 missing respondents (refused/not stated/multiple responses/not asked) out of 17,694 total respondents, which is only about 9.75% of the total. m. s. tahir et al. / financial services review 28 (2020) 273–301 279 t ab le 1 it em s o f fi n an ci al li te ra cy in w av e 1 6 o f h il d a f in an ci al li te ra cy co n ce p t it em s p o ss ib le re sp o n se s (c o rr ec t an sw er in b o ld ) s im p le in te re st “s u p p o se y o u p u t $ 1 0 0 in to a n o -f ee sa v in g s ac co u n t w it h a g u ar an te ed in te re st ra te o f 2 % p er y ea r. h o w m u ch w o u ld b e in th e ac co u n t at th e en d o f th e fi rs t y ea r? ” d o n ’t k n o w /r ef u se d /$ 10 2/ o th er v al u e in fl at io n “i f th e in te re st ra te o n y o u r sa v in g s ac co u n t w as 1 % p er y ea r an d in fl at io n w as 2 % p er y ea r. a ft er o n e y ea r, w o u ld y o u b e ab le to b u y m o re /t h e sa m e/ le ss th an to d ay ?” d o n ’t k n o w /r ef u se d /m o re /s am e/ le ss th an to da y r is k an d re tu rn “a n in v es tm en t w it h a h ig h re tu rn is li k el y to b e h ig h ri sk .” d o n ’t k n o w /r ef u se d /t ru e/ fa ls e p o rt fo li o ch o ic e “b u y in g sh ar es in a si n g le co m p an y u su al ly p ro v id es a sa fe r re tu rn th an b u y in g sh ar es in a n u m b er o f d if fe re n t co m p an ie s. ” d o n ’t k n o w /r ef u se d /t ru e/ fa ls e t im e v al u e o f m o n ey “i f b y th e y ea r 2 0 2 0 y o u r in co m e h as d o u b le d , b u t th e p ri ce s o f al l o f p u rc h as es h av e al so d o u b le d . in 2 0 2 0 , w il l y o u b e ab le to b u y m o re /t h e sa m e/ le ss th an to d ay ?” d o n ’t k n o w /r ef u se d /m o re /s am e/ le ss th an to d ay n o te : s o m e o f th es e it em s ar e b as ed o n l u sa rd i an d m it ch el l (2 0 0 7 ). e ac h fi n an ci al li te ra cy q u es ti o n h as a co rr ec t an d in co rr ec t an sw er (s ). a fi n an ci al li ter ac y m ea su re is d ev el o p ed b y ta k in g th e su m o f co rr ec t an sw er s p ro v id ed b y ea ch in d iv id u al . 280 m. s. tahir et al. / financial services review 28 (2020) 273–301 moreover, respondents are asked to depict their satisfaction level with their financial situation through the item “i am now going to ask you some questions about how satisfied or dissatisfied you are with some of the things happening in your life. i am going to read out a list of different aspects of life and, using the scale on showcard k13, i want you to pick a number between 0 and 10 that indicates your level of satisfaction with each. the more satisfied you are, the higher the number you should pick. the less satisfied you are, the lower the number. . . c) your financial situation?”. this item has 31 missing respondents (refused/ not stated/don’t know) out of 17,694 total respondents. this item is measured on an 11points scale with 0 = totally dissatisfied while 10 = totally satisfied. to measure credit card debt-taking behavior, two relevant items in wave 16 of hilda are; “do you have any credit cards, charge cards or store accounts? do not include debit cards.” and “how often is the entire balance on all your credit cards paid off each month?” the first item only asks whether the respondents have a credit card or not. this does not seem relevant since having a credit card does not mean having credit card debt. instead, the second item explicitly taps the debt-taking behavior of the respondents. because the focus of this article is on credit card debt-taking behavior, we opt for the second item. the options available in response to the second item are “(1) pays off entire balance hardly ever/never (2) pays off entire balance not very often (3) pays off entire balance about half the time (4) pays off entire balance most months (5) pays off entire balance always/ almost always.” we combine the first three options of this item into one category (as “1”) and the last two options into the other category (as “0”); thus, making it a binary response variable. the reason for doing this is the variable now tracks the behavior of credit card debt-taking that we specifically research. if a respondent falls into any of the first three options, it implies his or her behavior as incurring credit card debt. whereas the last two table 2 respondents of financial literacy items in wave 16 of hilda measuring concept correct responses wrong responses got assistance missing values correct percentagea simple interestb 14,773 2,506 198 217 85.5 inflationc 12,228 5,152 76 238 70.4 risk and returnd 14,569 2,817 137 171 83.8 portfolio choicee 13,325 4,102 91 176 76.5 time value of moneyf 13,727 3,650 90 227 79.0 a correct percentage is calculated after omitting all those who got someone’s assistance regardless of a correct or wrong answer. moreover, those who replied as “don’t know” are treated as wrong answers due to being numerical and general conceptual questions. b those who correctly answered after getting assistance are 176, while those who wrongly answered after getting assistance are 22. c those who correctly answered after getting assistance are 51, while those who wrongly answered after getting assistance are 25. d those who correctly answered after getting assistance are 115, while those who wrongly answered after getting assistance are 22. e those who correctly answered after getting assistance are 46, while those who wrongly answered after getting assistance are 45. f those who correctly answered after getting assistance are 66, while those who wrongly answered after getting assistance are 24. m. s. tahir et al. / financial services review 28 (2020) 273–301 281 options depict that the respondent avoids having credit card debt. as stated earlier in the introduction section of this article, the researchers report credit card debt as a most stressful and problematic type of household debt (dunn & mirzaie, 2012, 2016; richards, ahmed, & tahir, 2019). the debt becomes more problematic and burdensome when monthly balances linger on and are not paid off. the initial valid percentage of respondents in the raw dataset (before cleaning and filtering of the whole dataset) falling into the “0” category (not taking on credit card debt) is 74.4%, while those who fall into the other category named as “1” (taking on credit card debt) is 25.6%. it shows that one out of every four individuals is entering into a credit card debt arrangement in australia. in our analysis, we control for demographic factors to analyze if the association between the main variables of interest remains the same. specifically, we include age, gender, marital status, employment status, and educational status. gender, marital status, employment status, and educational status are included as dummy variables having two categories each as males and females, those who are currently in a registered marriage and those who are not in a registered marriage, those who are full time or part time employed and those who are not working, and those who have earned bachelor/graduate diploma/postgraduate degree and those who have earned year 11 or below/year 12/certificate iii or iv/advanced diploma degree, respectively. in our analyses, we include age as a continuous variable. in addition, we include the age-squared variable to examine the quadratic relationship of age to credit card debt-taking behavior. finally, as each variable consists of some missing respondents, we removed all the missing respondents from the data and created a cleaned and filtered dataset where we included only those respondents who responded to each of our interested variables. the variable with the most missing responses is credit card debt-taking behavior. in wave 16, this variable had 8,136 missing respondents (refused/not stated/don’t know/not asked) out of 17,694 total respondents. after omitting missing responses from other variables, 6,661 respondents remained out of a total of 17,694. the sample of 6,661 individuals consists of 51.6% females, 60.3% married, and 89.9% are employed full time or part time. moreover, 40% are identified as having earned bachelor/graduate diploma/postgraduate degree, while 60% are identified as having earned year 11 or below/year 12/certificate iii or iv/advanced diploma degree. furthermore, the minimum age of respondents is 15 and the maximum age is 92 with a mean value of 45. the average income of sample individuals is $66,092. table 3 percentage of respondents in each level of financial literacy (wave 16 of hilda) financial literacy levels frequencya percent no correct answer 324 1.91 one correct answer 577 3.40 two correct answers 1,283 7.55 three correct answers 2,631 15.48 four correct answers 4,692 27.61 five correct answers (high financial literacy) 7,487 44.06 total 16,994 100.0 a missing values are excluded. 282 m. s. tahir et al. / financial services review 28 (2020) 273–301 4. methodology the binary nature of the dependent variable allows using a binary logit model to test the relationship between the variables (kennedy, 2003; long & freese, 2006). appendix 1 shows the relevant econometrics equations of a binary logit model. the results of the binary logit model can only indicate a positive or negative relationship between the variables as we cannot interpret the estimated coefficients. to interpret the magnitude of the relationship, we often calculate the odds ratio. an odds ratio can be interpreted as an expected change in the odds of the dependent variable due to a 1-unit change in the independent variable. however, we can also calculate the standardized odds ratio of logistic regression that is relatively easier to interpret as compared with the odds ratio. a standardized odds ratio states the odds ratio from a one standard deviation change in the independent variable. in addition to calculating the odds ratio and standardized odds ratio, we also calculate ames. an advantage of calculating the ames is that the values of the ames can be interpreted to compare the relative strength of each variable (mussida & sciulli, 2019; stewart, 2007; west & worthington, 2014). therefore, the computation of the ames aligns with the motive of this article, that is, to test the relationship of financial literacy, attitude towards balancing spending and savings, and financial satisfaction with credit card debt-taking behavior, and to explain the relative strength of the association of each variable. we use the statistical analysis program “stata” to use a binary logit model. we ran the analyses in three phases. first, we only include financial literacy in the model. second, we add attitude towards balancing spending and savings in the model. finally, we include financial satisfaction in the model. we also control for demographic factors in all the models. in each of the phases, we run two models. the first model of each phase includes financial literacy accumulated score variable, whereas the second model of each phase includes five financial literacy concepts as separate dummy variables. this activity would help us to identify the relative association of five financial literacy concepts with credit card debt-taking behavior. the five financial literacy concepts are simple interest, inflation, risk and return, portfolio choice, and the time value of money. there is a twofold purpose of running analyses in three phases. first, we analyze if the association between financial literacy and credit card debt-taking behavior changes over the phases. second, we analyze the relative strength of each variable in each phase. 5. results 5.1. descriptive statistics the following figures show different types of descriptive comparisons between variables for the filtered sample of 6,661 individuals. fig. 1 below shows the distribution of credit card debt-taking among males and females. our data contain 28% of males and around 33% of females with credit card debt-taking. next, fig. 2 shows the percentage of males and females who correctly answered each of the five financial literacy questions. fig. 2 identifies m. s. tahir et al. / financial services review 28 (2020) 273–301 283 that most of the males (97.89%) and females (90.21%) could correctly answer the concept of simple interest, while the least percentage of the correct answer of males relates to the concept of the time value of money (85.52), whereas the least percentage of the correct answer of females relates to the concept of inflation (73.77). fig. 3 shows the percentage of respondents who take on debt and could correctly answer each of the five financial literacy questions. according to fig. 3, most percentage of those who take on debt (91.58) could correctly answer the concept of simple interest, while the lowest percentage (76.32) could correctly answer the concept of inflation. next, fig. 4 shows that most of the respondents could correctly answer the concept of simple interest regardless of the degree they have earned. fig. 5 and fig. 6 take into account the correct response rate of all (five) financial literacy questions. fig. 5 below shows that 44% of respondents with zero financial literacy score have credit card debt. in contrast, 26% of respondents who correctly answered all five questions are identified as those who take on debt. these descriptive statistics indicate that financially illiterate people are relatively more inclined to credit card debt as compared with financially literate people. next, for a comparison purpose, we make seven blocks of age as below 25 years, 25–34, 35–44, 45–54, 55–64, 65–74, and above 74. fig. 6 shows that those who are below 25 years of age contain least percentage of respondents who could correctly answer all the five financial literacy questions, whereas the age (in years) block 55–64 fig. 1. percent comparison of credit card debt-taking behavior by males and females. fig. 2. percent comparison of correct answers to the financial literacy concepts by males and females. 284 m. s. tahir et al. / financial services review 28 (2020) 273–301 contains most of the respondents who are highly financially literate. the young generation in australia is not highly financially literate and may require financial literacy education. finally, fig. 7 shows the age comparison of those who have credit card debt. our data indicate that those who are between 25 and 34 years of age are relatively more inclined to credit card debt as compared with other age groups. the overall trend in fig. 7 explicates that debt increases up to a certain age, then it decreases. 5.2. correlation analysis table 4 below shows the pairwise correlation analysis. credit card debt-taking behavior is negatively correlated with financial literacy, attitude towards balancing spending and fig. 3. percent of those take on debt and could correctly answer the questions relating to the financial literacy concepts. fig. 4. percent comparison of correct answers to the financial literacy concepts by education status. m. s. tahir et al. / financial services review 28 (2020) 273–301 285 savings, and financial satisfaction as expected from the literature review. among demographic factors, employment status is uncorrelated with attitude towards balancing spending and savings, problematic debt-taking, and age. furthermore, gender is uncorrelated with attitude towards balancing spending and savings and age, while less correlated with credit card debt-taking behavior. all other variables are statistically significantly correlated with each other. 5.3. empirical association of financial literacy, attitude towards balancing spending and savings, and financial satisfaction with credit card debt-taking behavior table 5, table 6, and table 7 show the results of empirical analyses after controlling for demographic factors. table 5 below contains the analysis of the financial literacy variable, table 6 adds attitude towards balancing spending and savings variable, and table 7 adds the financial satisfaction variable. the first model of each table contains financial literacy fig. 5. levels of financial literacy and percentage comparison of those who take on debt. fig. 6. percent comparison of correct answers to the financial literacy concepts by age (in years). 286 m. s. tahir et al. / financial services review 28 (2020) 273–301 accumulated score variable, whereas the second model of each table contains five financial literacy concepts as separate dummy variables. table 5 below shows the results of the two models. the first model contains only financial literacy variable, and it negatively relates to credit card debt-taking behavior at the statistical significance level. this analysis supports hypothesis 1. it implies that those who are financially literate are less likely to have credit card debt. the odds ratio of financial literacy indicates that for a 1-unit change in financial literacy, the log odds of credit card debt-taking behavior are expected to change by a factor of 0.914, holding all other variables unchanged. the standardized odds ratio indicates that for a one standard deviation increase in financial literacy score, one could expect 0.920 factors change in the log odds of credit card debt-taking. unlike the odds ratio and standardized odds ratio, the ames provide ease in the interpretation as the values of the ames of each variable are comparable to know the relative strength of the association. the ame of financial literacy in the first model of table 5 can be interpreted as a higher financial literacy score decreases the probability of having credit card debt by 1.8%. the second model adds dummy variables of each financial literacy concept. the analysis shows that two of the five dummy variables (simple interest and risk and return) negatively relate to credit card debt. the other three financial literacy dummy variables are statistically insignificant. the ames imply that those who have knowledge about simple interest and risk and return concepts are less likely to have credit card debt by 6.4% and 3.8%, respectively. table 6 below adds the attitude towards balancing spending and savings variable in the analysis. the results are comparable to table 5 above. in the first model of table 6, we note the same statistical significance level of financial literacy variable as noted in the first model of table 5 above. the ame of financial literacy in the first model of table 6 can be interpreted as a higher financial literacy score decreases the probability of having credit card debt by 1.5%, which is 0.3 percentage points lower than the one in table 5. further, the second model of table 6 shows that the dummy variables of simple interest and risk and return concepts are statistically significant. the ames imply that those who have knowledge about fig. 7. percent comparison of those who take on debt by age (in years). m. s. tahir et al. / financial services review 28 (2020) 273–301 287 t ab le 4 p ai rw is e co rr el at io n an al y si s o f v ar ia b le s 1 2 3 4 5 6 7 8 9 1 . f in an ci al li te ra cy 2 . a tt it u d e to w ar d s b al an ci n g sp en d in g an d sa v in g s 0 .0 4 5 * * * 3 . f in an ci al sa ti sf ac ti o n 0 .0 6 0 * * * 0 .3 5 7 * * * 4 . c re d it ca rd d eb tta k in g b eh av io r �0 .0 7 8 * * * �0 .3 1 6 * * * �0 .3 5 2 * * * 5 . g en d er �0 .1 5 5 * * * �0 .0 1 8 �0 .0 3 3 * * * 0 .0 2 5 * 6 . a g e 0 .1 4 7 * * * 0 .1 3 9 * * * 0 .0 5 9 * * * �0 .0 7 6 * * * �0 .0 1 8 7 . in co m e 0 .1 4 8 * * * 0 .0 6 0 * * * 0 .1 6 2 * * * �0 .0 9 7 * * * �0 .1 8 1 * * * 0 .0 9 3 * * * 8 . e m p lo y m en t st at u s 0 .8 2 8 * * * 0 .0 1 6 0 .1 1 5 * * * �0 .0 0 6 �0 .1 8 1 * * * �0 .0 1 1 0 .1 4 6 * * * 9 . m ar it al st at u s 0 .0 8 4 * * * 0 .0 6 4 * * * 0 .1 5 9 * * * �0 .0 9 6 * * * �0 .0 2 8 * 0 .1 9 9 * * * 0 .0 4 9 * * * �0 .0 2 8 * 1 0 . e d u ca ti o n st at u s 0 .1 8 1 * * * 0 .0 4 0 * * 0 .1 2 3 * * * �0 .1 5 0 * * * 0 .0 8 5 * * * �0 .0 6 3 * * * 0 .1 8 1 * * * 0 .0 5 4 * * * 0 .0 8 7 * * * * * * p < .0 0 1 , * * p < .0 1 , * p < .0 5 . n = 6 ,6 6 1 . 288 m. s. tahir et al. / financial services review 28 (2020) 273–301 t ab le 5 b in ar y lo g is ti c re g re ss io n an al y si s o f cr ed it ca rd d eb tta k in g b eh av io r: p h as e 1 m o d el 1 m o d el 2 a m e o r s o r a m e o r s o r f in an ci al li te ra cy �0 .0 1 8 * * 0 .9 1 4 * * (0 .0 2 8 ) 0 .9 2 0 * * in fl at io n �0 .0 0 4 0 .9 8 1 (0 .0 7 2 ) 0 .9 9 2 p o rt fo li o ch o ic e �0 .0 2 9 0 .8 6 0 (0 .0 6 8 ) 0 .9 4 9 s im p le in te re st �0 .0 6 4 * * 0 .7 2 0 * * (0 .0 8 3 ) 0 .9 2 5 * * t im e v al u e o f m o n ey 0 .0 2 6 1 .1 4 1 (0 .0 9 3 ) 1 .0 4 7 r is k an d re tu rn �0 .0 3 8 * 0 .8 2 5 * (0 .0 7 6 ) 0 .9 4 5 * f em al e 0 .0 0 5 1 .0 2 8 (0 .0 6 1 ) 1 .0 1 4 0 .0 0 2 1 .0 0 9 (0 .0 6 1 ) 1 .0 0 5 a g e 0 .0 2 7 * * * 1 .1 4 7 * * * (0 .0 1 9 ) 5 .7 5 7 * * * 0 .0 2 8 * * * 1 .1 4 9 * * * (0 .0 1 9 ) 5 .9 1 0 * * * a g esq u ar ed �0 .0 0 0 3 * * * 0 .9 9 8 * * * (0 .0 0 0 ) 0 .1 4 0 * * * �0 .0 0 0 3 * * * 0 .9 9 8 * * * (0 .0 0 0 ) 0 .1 3 8 * * * in co m e �0 .0 0 0 0 0 1 * 1 .0 0 0 * (0 .0 0 0 ) 0 .6 0 6 * �0 .0 0 0 0 0 1 * 1 .0 0 0 * (0 .0 0 0 ) 0 .6 2 0 * a g e * in co m e 0 .0 0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 1 .2 2 3 0 .0 0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 1 .1 9 2 e m p lo y ed 0 .2 2 3 1 .1 2 1 (0 .1 0 7 ) 1 .0 3 5 0 .0 2 2 1 .1 1 7 (0 .1 0 7 ) 1 .0 3 4 m ar ri ed �0 .0 8 1 * * * 0 .6 6 3 * * * (0 .0 3 9 ) 0 .8 1 8 * * * �0 .0 8 0 * * * 0 .6 6 5 * * * (0 .0 3 9 ) 0 .8 1 9 * * * m in im u m b ac h el o r d eg re e �0 .1 2 4 * * * 0 .5 3 0 * * * (0 .0 3 3 ) 0 .7 3 3 * * * �0 .1 2 4 * * * 0 .5 3 0 * * * (0 .0 3 3 ) 0 .7 3 2 * * * c o n st an t 0 .0 9 6 * * * (0 .0 3 4 ) 0 .1 0 2 * * * (0 .0 3 7 ) n 6 ,6 6 1 6 ,6 6 1 p se u d o r 2 0 .0 4 7 0 .0 4 8 x 2 v al u e 3 7 5 .0 4 * * * 3 8 8 .2 6 * * * p ea rs o n x 2 6 9 5 6 .2 6 * * 6 9 6 6 .9 9 * * * * * p < .0 0 1 , * * p < .0 1 , * p < .0 5 . a m e = av er ag e m ar g in al ef fe ct s; o r = o d d s ra ti o ; s o r = fu ll y st an d ar d iz ed o d d s ra ti o . s ta n d ar d er ro rs in p ar en th es es . m. s. tahir et al. / financial services review 28 (2020) 273–301 289 t ab le 6 b in ar y lo g is ti c re g re ss io n an al y si s o f cr ed it ca rd d eb tta k in g b eh av io r: p h as e 2 m o d el 1 m o d el 2 a m e o r s o r a m e o r s o r f in an ci al li te ra cy �0 .0 1 5 * * 0 .9 1 7 * * (0 .0 3 0 ) 0 .9 2 4 * * in fl at io n �0 .0 0 5 0 .9 7 1 (0 .0 7 5 ) 0 .9 8 8 p o rt fo li o ch o ic e �0 .0 2 5 0 .8 6 9 (0 .0 7 2 ) 0 .9 5 2 s im p le in te re st �0 .0 5 5 * 0 .7 3 2 * (0 .0 8 9 ) 0 .9 2 8 * t im e v al u e o f m o n ey 0 .0 2 9 1 .1 8 1 (0 .1 0 1 ) 1 .0 6 0 r is k an d re tu rn �0 .0 4 0 * 0 .8 0 0 * (0 .0 7 8 ) 0 .9 3 6 * a tt it u d e to w ar d s b al an ci n g sp en d in g an d sa v in g s �0 .0 8 0 * * * 0 .6 3 7 * * * (0 .0 1 3 ) 0 .5 0 5 * * * �0 .0 8 0 * * * 0 .6 3 7 * * * (0 .0 1 3 ) 0 .5 0 4 * * * f em al e 0 .0 0 6 1 .0 3 5 (0 .0 6 5 ) 1 .0 1 7 0 .0 0 2 1 .0 1 2 (0 .0 6 4 ) 1 .0 0 6 a g e 0 .0 2 3 * * * 1 .1 4 1 * * * (0 .0 1 9 ) 5 .3 7 1 * * * 0 .0 2 4 * * * 1 .1 4 4 * * * (0 .0 2 0 ) 5 .5 8 4 * * * a g esq u ar ed �0 .0 0 0 3 * * * 0 .9 9 8 * * * (0 .0 0 0 ) 0 .1 6 8 * * * �0 .0 0 0 3 * * * 0 .9 9 8 * * * (0 .0 0 0 ) 0 .1 6 4 * * * in co m e �0 .0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 0 .7 1 4 �0 .0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 0 .7 3 3 a g e * in co m e 0 .0 0 0 0 0 0 0 0 3 1 .0 0 0 (0 .0 0 0 ) 1 .0 5 6 0 .0 0 0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 1 .0 2 6 e m p lo y ed 0 .0 2 6 1 .1 5 4 (0 .1 1 6 ) 1 .0 4 4 0 .0 2 4 1 .1 4 7 (0 .1 1 5 ) 1 .0 4 2 m ar ri ed �0 .0 6 7 * * * 0 .0 6 8 6 * * * (0 .0 4 2 ) 0 .8 3 2 * * * �0 .0 6 6 * * * 0 .6 8 9 * * * (0 .0 4 2 ) 0 .8 3 3 * * * m in im u m b ac h el o r d eg re e �0 .1 1 2 * * * 0 .5 3 3 * * * (0 .0 3 5 ) 0 .7 3 5 * * * �0 .1 1 2 * * * 0 .5 3 2 * * * (0 .0 3 5 ) 0 .7 3 4 * * * c o n st an t 0 .6 6 3 (0 .2 5 5 ) 0 .6 9 2 (0 .2 7 1 ) n 6 ,6 6 1 6 ,6 6 1 p se u d o r 2 0 .1 1 8 0 .1 2 0 x 2 v al u e 9 4 7 .9 0 * * * 9 6 1 .8 7 * * * p ea rs o n x 2 6 6 8 4 .9 7 6 7 1 2 .7 4 * * * p < .0 0 1 , * * p < .0 1 , * p < .0 5 . a m e = av er ag e m ar g in al ef fe ct s; o r = o d d s ra ti o ; s o r = fu ll y st an d ar d iz ed o d d s ra ti o . s ta n d ar d er ro rs in p ar en th es es . 290 m. s. tahir et al. / financial services review 28 (2020) 273–301 t ab le 7 b in ar y lo g is ti c re g re ss io n an al y si s o f cr ed it ca rd d eb tta k in g b eh av io r: p h as e 3 m o d el 1 m o d el 2 a m e o r s o r a m e o r s o r f in an ci al li te ra cy �0 .0 1 5 * * 0 .9 1 5 * * (0 .0 3 0 ) 0 .9 2 2 * * in fl at io n �0 .0 1 1 0 .9 3 7 (0 .0 7 4 ) 0 .9 7 4 p o rt fo li o ch o ic e �0 .0 3 2 * 0 .8 2 4 * (0 .0 7 0 ) 0 .9 3 5 * s im p le in te re st �0 .0 4 4 * 0 .7 6 9 * (0 .0 9 7 ) 0 .9 3 9 * t im e v al u e o f m o n ey 0 .0 3 7 * 1 .2 4 9 * (0 .1 1 1 ) 1 .0 8 1 * r is k an d re tu rn �0 .0 3 7 * 0 .8 0 3 * (0 .0 8 1 ) 0 .9 3 7 * a tt it u d e to w ar d s b al an ci n g sp en d in g an d sa v in g s �0 .0 5 8 * * * 0 .7 0 7 * * * (0 .0 1 5 ) 0 .5 9 1 * * * �0 .0 5 8 * * * 0 .7 0 6 * * * (0 .0 1 5 ) 0 .5 9 0 * * * f in an ci al sa ti sf ac ti o n �0 .0 5 4 * * * 0 .7 2 2 * * * (0 .0 1 3 ) 0 .5 4 4 * * * �0 .0 5 4 * * * 0 .7 2 1 * * * (0 .0 1 3 ) 0 .5 4 2 * * * f em al e 0 .0 1 1 1 .0 6 7 (0 .0 6 8 ) 1 .0 3 3 0 .0 0 7 1 .0 4 3 (0 .0 6 8 ) 1 .0 2 1 a g e 0 .0 1 8 * * * 1 .1 1 6 * * * (0 .0 2 0 ) 4 .0 7 7 * * * 0 .0 1 9 * * * 1 .1 2 1 * * * (0 .0 2 0 ) 4 .3 0 1 * * * a g esq u ar ed �0 .0 0 0 2 * * * 0 .9 9 9 * * * (0 .0 0 0 ) 0 .2 1 9 * * * �0 .0 0 0 2 * * * 0 .9 9 9 * * * (0 .0 0 0 ) 0 .2 1 2 * * * in co m e �0 .0 0 0 0 0 0 4 1 .0 0 0 (0 .0 0 0 ) 0 .8 5 6 �0 .0 0 0 0 0 0 3 1 .0 0 0 (0 .0 0 0 ) 0 .8 7 8 a g e * in co m e �0 .0 0 0 0 0 0 0 0 0 3 1 .0 0 0 (0 .0 0 0 ) 0 .9 9 4 �0 .0 0 0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 0 .9 6 8 e m p lo y ed 0 .0 6 0 * * * 1 .4 3 0 * * * (0 .1 5 1 ) 1 .1 1 4 * * * 0 .0 5 8 * * * 1 .4 2 1 * * * (0 .1 5 0 ) 1 .1 1 2 * * * m ar ri ed �0 .0 3 3 * * 0 .8 1 9 * * (0 .0 5 2 ) 0 .9 0 7 * * �0 .0 3 2 * * 0 .8 2 5 * * (0 .0 5 3 ) 0 .9 1 0 * * m in im u m b ac h el o r d eg re e �0 .0 9 8 * * * 0 .5 5 7 * * * (0 .0 3 7 ) 0 .7 5 0 * * * �0 .0 9 7 * * * 0 .5 5 8 * * * (0 .0 3 7 ) 0 .7 5 1 * * * c o n st an t 3 .6 3 2 * * (1 .4 7 9 ) 3 .5 8 0 * * (1 .4 8 6 ) n 6 ,6 6 1 6 ,6 6 1 p se u d o r 2 0 .1 6 3 0 .1 6 5 x 2 v al u e 1 3 1 2 .2 9 * * * 1 3 2 8 .8 7 * * * p ea rs o n x 2 6 6 4 4 .3 4 6 6 7 7 .4 7 * * * p < .0 0 1 , * * p < .0 1 , * p < .0 5 . a m e = av er ag e m ar g in al ef fe ct s; o r = o d d s ra ti o ; s o r = fu ll y st an d ar d iz ed o d d s ra ti o . s ta n d ar d er ro rs in p ar en th es es . m. s. tahir et al. / financial services review 28 (2020) 273–301 291 simple interest and risk and return concepts are more likely to avoid credit card debt by 5.5% and 4.0%, respectively, when attitude towards balancing spending and savings is controlled in the model. moreover, the attitude towards balancing spending and savings variable is highly statistically significant in both the models of table 6 supporting hypothesis 2. the ame shows that if a person shows a one-point positive attitude towards balancing their spending and savings, it would decrease their likelihood of having credit card debt by 8%. moreover, the comparison of the ames reveals that the relative strength of the attitude towards balancing spending and savings variable (ame = �0.080) is higher than that of the financial literacy score variable (ame = �0.015). table 7 below adds the financial satisfaction variable in the analysis. the results are comparable to table 5 and table 6 above. the financial literacy variable in the first model of table 7 shows no change in the statistical significance level, nor in the ame value. it implies that the financial satisfaction variable does not have any effect on the relative strength of the financial literacy variable towards credit card debt-taking. however, the second model of table 7 shows that, unlike the results of table 5 and table 6, the dummy variables of portfolio choice and time value of money are statistically significant. the results further show a positive association of the time value of money dummy variable. it implies that those who know the concept of time value of money, are more likely to have credit card debt when the financial satisfaction variable is incorporated in the model. moreover, in both the models of table 7, the financial satisfaction variable is highly statistically significant. this analysis supporting hypothesis 3. the ame shows that if a person indicates a one-point increase in their financial satisfaction level, it would decrease their likelihood of having debt by 5.4%. furthermore, the attitude towards balancing spending and savings variable is still highly statistically significant in both the models of table 7 as similar to the models of table 6. the ame shows that if a person shows a one-point positive attitude towards balancing their spending and savings, it would decrease their likelihood of having credit card debt by 5.8%, which is 1.2 percentage points lower than the one in table 6. however, the relative strength of the attitude towards balancing spending and savings variable is still higher than the other main variables of interest in both the models of table 7. among the demographic factors, those who are married and earned minimum bachelor degree are less likely to have debt, while the results for gender, employment status, and income variables are mostly insignificant across the models. the results for the interaction term of age and income are not statistically significant across all models. furthermore, age is positively associated with problematic debt-taking. however, the quadratic relationship for age in relation to credit card debt-taking behavior is negative. according to grable, lyons, and heo (2019), this statistical position of age and age-squared variables suggests an inverted u-shaped downward relationship with the credit card debt-taking behavior supporting the descriptive statistics shown in fig. 7 of this article. in the case of binary logistic regression, adjusted r2 is replaced by pseudo r2, which is not as same as the former (long & freese, 2006). nonetheless, a comparison of pseudo r2 across the models can indicate the explanatory value of each model relative to another. the second model of table 7, where all the variables are included in the model, has the highest pseudo r2 value of 0.165 among the six models presented in table 5, 292 m. s. tahir et al. / financial services review 28 (2020) 273–301 table 6, and table 7. the least pseudo r2 value of 0.047 is noted in the first model of table 5 where only financial literacy is included in the model. furthermore, the x2 value is highly significant across the models of all the tables implying that the regression models are significant as a whole. in addition, we test for the issue of multicollinearity using the variance inflation factor (vif) and find that all vifs are below the threshold of 10 that guarantees no multicollinearity issue (kennedy, 2003). we append the vif results in appendix 2 below. across the analyses presented above, we had a sample size of 6,661 individuals that is about one-third of the original sample of 17,694 individuals surveyed in the 2016 wave of the hilda survey. this reduction in the sample size was the result of filtering the dataset and omitting the retired respondents from the analyses. however, we run additional two models (as similar to those presented in table 7) without filtering the dataset for a purpose to check the robustness of the existing results. this analysis incorporates 8,542 individuals. the sample consists of 52% females, 62% married, and 70% are employed full time or part time. moreover, 36% are identified as having earned bachelor/ graduate diploma/postgraduate degree, while 64% as having earned year 11 or below/ year 12/certificate iii or iv/advanced diploma degree. furthermore, the minimum age of respondents is 15 and the maximum age is 98. the average income is $60,150. overall, the age and income average of this sample is different from the sample included in the main results because we include retired respondents in this robustness check. we append the results in appendix 3 below. the statistical significance of the variables in appendix 3 is similar to those in table 7. it implies that the main results of this article are robust implying that the filtering of the dataset and omitting the retired respondents in the main results do not influence our conclusion. in addition to these analyses, we analyze if financial literacy is associated with attitude towards balancing spending and savings and financial satisfaction. though the correlation analysis (see table 4 above) shows that there is a positive association between financial literacy, attitude towards balancing spending and savings, and financial satisfaction. however, we empirically analyze this relationship and control for demographic factors. our analysis reveals that financial literacy is neither associated with attitude towards balancing spending and savings (p > .05) nor with financial satisfaction (p > .05) when we control for demographic factors. these results indicate that the concept of financial literacy is independent of the other two concepts. the results are omitted for brevity. 6. conclusion existing research reports a low financial literacy level in australia (ali et al., 2014). in response to the financial literacy survey of bourova, anderson, ramsay, and ali (2018), only 30.9% of respondents could answer all the financial literacy questions correctly. the australian government has been making efforts to equip australians with financial literacy skills to give households superior financial knowledge to make better judgments and enable them to make informed financial decisions. however, the financial services regulator in m. s. tahir et al. / financial services review 28 (2020) 273–301 293 australia, the asic (2018), states that the credit card debt is still rising in australia and it is also the most common type of household debt in australia (abs, 2017b). the relevant existing literature revealed that credit card debt is relatively, as compared with other non-collateralized types of debt, a more stressful and problematic type of household debt. this issue motivated this research, and we expected that an increase in financial literacy level may be relevant to a reduced credit card debt-taking behavior in australia. in addition, in contrast to the literature that includes savings for emergency needs as a measure of savings attitude, we analyzed the attitude of individuals towards balancing spending and savings. we hypothesized that those who keep a balance between their spending and savings are more likely to avoid credit card debt. we also expected a relationship between the financial satisfaction level of a person and their credit card debt-taking behavior. our contributions to the literature are many fold. unlike prior studies, we analyzed the relative weight of the association of financial literacy, attitude towards balancing spending and savings, and financial satisfaction with credit card debt-taking behavior. we used the 2016 wave of a nationally representative dataset namely the hilda survey. in our analysis, we, first, included a variable of accumulated financial literacy scores and then five dummy variables for each financial literacy concept to separately analyze their relationship with the credit card debt-taking behavior. we undertook analyses in different phases. we, first, included only financial literacy variable, then we added attitude towards balancing spending and savings and financial satisfaction, respectively. the purpose was to analyze whether the significance level of financial literacy and the magnitude of its relationship with the credit card debt-taking behavior remains the same or does change. this also helped us to examine the consistency and robustness of the relationship among the key variables. our analyses support all hypotheses of this article. in all the phases of analyses, financial literacy remained negatively significant to show a relationship with credit card debttaking behavior. however, the magnitude of the relationship decreased when other factors were included in the model. moreover, the comparison of the ames revealed that attitude towards balancing spending and savings has more relevance to the reduced credit card debt-taking behavior, followed by the financial satisfaction variable and then financial literacy. this implies that financial literacy has a relatively weaker magnitude of relationship with credit card debt-taking behavior, yet a significant factor to have an association with credit card debt-taking behavior. among the five financial literacy concepts, the concept of simple interest showed a highly negative association with credit card debttaking behavior. our results suggest that prior financial knowledge has a relatively weaker magnitude of association with financial decisions. instead, attitude towards money matters more than the prior financial knowledge. the results are comparable with the prior studies. fernandes et al. (2014) find that the incorporation of factors other than financial literacy in analyzing the relationship between financial literacy and financial behavior diminishes the effect size of financial literacy. moreover, garcı́a (2013) shows that prior beliefs, mental abilities, and cognitive factors dominate the decision-making process. our results partially support this conclusion because the magnitude of the association of financial literary diminishes when other factors are included. however, the relevance of financial literacy does not disappear 294 m. s. tahir et al. / financial services review 28 (2020) 273–301 suggesting some importance for the role of financial literacy in credit card debt-taking behavior. in our results, the relevance of attitude towards balancing spending and savings and credit card debt-taking behavior is twofold. first, we show that those who keep a balance between spending and savings have relatively less probability of having credit card debt. second, we argue that people with a higher credit limit tend to spend more because they think that their future income will cover their credit card debt (soman & cheema, 2002). our results have an important implication for policies on improving the financial literacy of households in australia. the research suggests that education of this nature will have some, but limited relevance. curricula that focus on financial behaviors and attitude towards personal finance will be more relevant to reduced problematic debt-taking. in this regard, although the australian government has already designed a new website of financial capability that includes contents related to improving attitude towards personal finance, but there is still room for improvement as there is a global emergence of the concept of financial wellbeing.2 kempson, finney, and poppe (2017) explain the shift of focus from financial literacy to financial capability, and from financial capability to financial wellbeing. further, in light of the importance of attitude towards balancing spending and savings, people need to be educated through focused media campaigns that promote an environment of not to overspend and keep a balance between spending and savings. moreover, although the asic is continuously reviewing and regulating the credit card market, but there is a need to regulate the media advertisements of financial institutions where they promote their financial products and attract consumers to open a credit card account. the financial services regulating institutions should continuously monitor the media advertising strategies of financial institutions and make people aware of any possible “debt-trap.” finally, as credit cards are a source of utilizing future income, financial institutions should set a credit limit after keeping in view the future income of consumers. some interventions must be introduced in case of any possible economic shock that could disturb the spending and savings attitude of consumers. our research has some limitations that future researchers may wish to address. we could not conclude a causal relationship between the variables because we used a crosssectional dataset. future researchers may conduct a longitudinal analysis to see if the suggested relevance turns into a causal relationship. further, future analyses may include more relevant behavioral and personality factors such as self-control, future-oriented behavior, and non-impulsive behavior to conclude a relationship with credit card debttaking behavior. as we used a secondary dataset for this article, we were unable to capture the influence of external factors that remained unobserved and might have a larger influence on the attitude and behavior of individuals/households. for instance, education from parents or society norms might have more relevance with the behavior of households that may affect their financial decisions. moreover, although we used popular measures of financial literacy, but different measures could produce different results. a guide for future researchers may be to subjectively measure financial literacy and analyze if the same results pop-up. finally, financial capability and financial wellbeing are emerging concepts in the field of personal finance that may have relevance with reduced credit card debt-taking behavior. m. s. tahir et al. / financial services review 28 (2020) 273–301 295 notes 1 “this article uses unit record data from the household, income and labour dynamics in australia (hilda) survey. the hilda project was initiated and is funded by the australian government department of families, housing, community services and indigenous affairs (fahcsia) and is managed by the melbourne institute of applied economic and social research (melbourne institute). the findings and views reported in this article; however, are those of the author and should not be attributed to either fahcsia or the melbourne institute.” (summerfield et al., 2017). 2 https://financialcapability.gov.au/ acknowledgment this research article is a part of the project funded by the think forward initiative (tfi). the technical report submitted to the tfi can be viewed here: https://www.thinkforwardinitiative. com/research/addressing-the-challenge-of-problematic-debt-australia-and-eurozone. the authors would like to thank the editor(s) and three anonymous reviewers for their comments on this article. the authors believe that the comments received have improved the readability and quality of the article. in addition, the corresponding author would like to thank his fellow phd candidates at rmit university (melbourne) for reviewing the article and providing constructive feedback. appendix 1: equations of binary logit regression the main dependent variable of this study—problematic debt-taking—is a binary variable. this variable was coded to “0” for those who pay off their monthly credit card balance each month (that means not take debt) and coded to “1” for those who are unable to pay off their monthly credit card balance each month (demonstrating debt-taking behavior). the binary nature of the dependent variable allows using a binary logit model to test the relationship between the variables (kennedy, 2003; long & freese, 2006). the results of the binary logit model can only indicate a positive or negative relationship between the variables as we cannot interpret the estimated coefficients. to interpret the magnitude of the relationship, we often calculate the odds ratio, which can be defined as: s tð þ ¼ et et þ 1 ¼ 1 1 þ e�t (1) where, t is the linear combination of all the explanatory variables of the study and e represents their exponential value. in this article, we have financial literacy, attitudes towards spending and savings, financial satisfaction, and demographic factors as explanatory variables. the inclusion of explanatory variables is: g p xð þð þ= ln p xð þ 1 � p xð þ � � = b 0 þ b 1x (2) this equation interprets the probability of the dependent variable equaling or approaching to a specific “case,” where x represents the list of explanatory variables as described above. 296 m. s. tahir et al. / financial services review 28 (2020) 273–301 moreover, the above equation states that the logit (log odds – natural log of the odds) is as same as the linear regression equation. this model can also be transformed into the following equation after taking the exponential values of both sides. p xð þ 1 � p xð þ � � ¼ eb 0 þ b 1x (3) in the case of the logistic function, the odds of the dependent variable being equal to a case are interpreted and the exponential value truly represents the odds as shown by the following equation: odds ¼ eb 0 þ b 1x (4) appendix 2 variance inflation factor (vif) of the variables of interest variable vif credit card debt-taking behavior 1.22 financial literacy 1.10 attitude towards balancing spending and savings 1.22 financial satisfaction 1.30 female 1.10 age 1.10 income 1.12 employment status 1.07 marital status 1.08 education achieved 1.13 m. s. tahir et al. / financial services review 28 (2020) 273–301 297 a p p en d ix 3 r o b u st n es s ch ec k m o d el 1 m o d el 2 a m e o r s o r a m e o r s o r f in an ci al li te ra cy �0 .0 1 4 * * 0 .9 0 9 * * (0 .0 2 8 ) 0 .9 1 4 * * in fl at io n �0 .0 0 8 0 .9 4 4 (0 .0 7 1 ) 0 .9 7 8 p o rt fo li o ch o ic e �0 .0 3 0 * 0 .8 1 7 * (0 .0 6 5 ) 0 .9 3 1 * s im p le in te re st �0 .0 3 8 * 0 .7 7 0 * (0 .0 8 8 ) 0 .9 3 5 * t im e v al u e o f m o n ey 0 .0 2 5 * 1 .1 8 7 * (0 .0 9 6 ) 1 .0 6 7 * r is k an d re tu rn �0 .0 3 0 * 0 .8 1 3 * (0 .0 7 8 ) 0 .9 4 3 * a tt it u d e to w ar d s b al an ci n g sp en d in g an d sa v in g s �0 .0 4 9 * * * 0 .7 1 8 * * * (0 .0 1 4 ) 0 .6 0 5 * * * �0 .0 4 9 * * * 0 .7 1 8 * * * (0 .0 1 4 ) 0 .6 0 5 * * * f in an ci al sa ti sf ac ti o n �0 .0 5 0 * * * 0 .7 1 2 * * * (0 .0 1 2 ) 0 .5 2 7 * * * �0 .0 5 0 * * * 0 .7 1 1 * * * (0 .0 1 2 ) 0 .5 2 6 * * * f em al e 0 .0 1 4 1 .0 9 7 (0 .0 6 6 ) 1 .0 4 7 0 .0 1 1 1 .0 7 8 (0 .0 6 5 ) 1 .0 3 8 a g e 0 .0 1 3 * * * 1 .0 8 9 * * * (0 .0 1 5 ) 3 .9 5 7 * * * 0 .0 1 3 * * * 1 .0 9 2 * * * (0 .0 1 5 ) 4 .1 7 7 * * * a g esq u ar ed �0 .0 0 0 1 * * * 0 .9 9 9 * * * (0 .0 0 0 ) 0 .1 9 0 * * * �0 .0 0 0 1 * * * 0 .9 9 9 * * * (0 .0 0 0 ) 0 .1 8 3 * * * in co m e 0 .0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 1 .0 6 8 0 .0 0 0 0 0 0 2 1 .0 0 0 (0 .0 0 0 ) 1 .0 8 5 a g e * in co m e �0 .0 0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 0 .7 6 5 �0 .0 0 0 0 0 0 0 1 1 .0 0 0 (0 .0 0 0 ) 0 .7 5 3 e m p lo y ed 0 .0 7 9 * * * 1 .7 0 9 * * * (0 .1 4 8 ) 1 .2 7 8 * * * 0 .0 7 8 * * * 1 .7 0 1 * * * (0 .1 4 8 ) 1 .2 7 5 * * * m ar ri ed �0 .0 3 3 * * * 0 .8 0 0 * * (0 .0 4 8 ) 0 .8 9 7 * * * �0 .0 3 2 * * * 0 .8 0 5 * * * (0 .0 4 8 ) 0 .9 0 0 * * * m in im u m b ac h el o r d eg re e �0 .0 8 4 * * * 0 .5 6 5 * * * (0 .0 3 6 ) 0 .7 6 0 * * * �0 .0 8 4 * * * 0 .5 6 5 * * * (0 .0 3 6 ) 0 .7 6 0 * * * c o n st an t 4 .8 6 9 * * * (1 .6 8 5 ) 4 .8 8 3 * * * (1 .7 2 5 ) n 8 ,5 4 2 8 ,5 4 2 p se u d o r 2 0 .1 9 1 0 .1 9 3 x 2 v al u e 1 8 3 6 .3 5 * * * 1 8 5 2 .0 7 * * * p ea rs o n x 2 8 3 7 7 .9 2 8 4 1 6 .2 9 * * * p < .0 0 1 , * * p < .0 1 , * p < .0 5 . a m e = av er ag e m ar g in al ef fe ct s; o r = o d d s ra ti o ; s o r = fu ll y st an d ar d iz ed o d d s ra ti o . s ta n d ar d er ro rs in p ar en th es es . 298 m. s. tahir et al. / financial services review 28 (2020) 273–301 references abs. (2017a). household income and wealth, australia, 2015-16: changes in wealth over time. available at https://www.abs.gov.au/ausstats/abs@.nsf/lookup/by%20subject/6523.0;2015-16;main%20features;changes %20in%20wealth%20over%20time;11 abs. (2017b). household income and wealth, australia, 2015-16: household debt and over-indebtedness in australia. available at http://www.abs.gov.au/ausstats/abs@.nsf/lookup/by%20subject/6523.0;201516;feature%20article;household%20debt%20and%20over-indebtedness%20(feature%20article);101 alhenawi, y., & elkhal, k. (2013). financial literacy of u.s. households: knowledge vs. long-term financial planning. financial services review, 22, 211–244. ali, a., rahman, m. s. a., & bakar, a. (2015). financial satisfaction and the influence of financial literacy in malaysia. social indicators research, 120, 137–156. ali, p., anderson, m., mcrae, c., & ramsay, i. (2014). the financial literacy of young australians: an empirical study and implications for consumer protection and asic’s national financial literacy strategy. company and securities law journal, 32, 334–352. asic. (2018). report 580: credit card lending in australia. available at https://asic.gov.au/regulatoryresources/find-a-document/reports/rep-580-credit-card-lending-in-australia/ berger, l. m., collins, j. m., & cuesta, l. (2016). household debt and adult depressive symptoms in the united states. journal of family and economic issues, 37, 42–57. bourova, e., anderson, m., ramsay, i., & ali, p. (2018). impacts of financial literacy and confidence on the severity of financial hardship in australia. australasian accounting, business and finance journal, 12, 4–23. brown, m., grigsby, j., van der klaauw, w., wen, j., & zafar, b. (2016). financial education and the debt behavior of the young. the review of financial studies, 29, 2490–2522. brown, s., & gray, d. (2016). household finances and well-being in australia: an empirical analysis of comparison effects. journal of economic psychology, 53, 17–36. cassells, r., duncan, a., kelly, s., & ong, r. (2015). beyond our means? household savings and debt in australia. bankwest curtin economics centre. focus on the states report series, (no. 2). curtin university, perth, western australia. cecchetti, s., mohanty, m., & zampolli, f. (2011).the real effects of debt. bis working papers, (no. 352). bank for international settlements, basel, switzerland. cude, b. j., chatterjee, s., & tavosi, j. (2019). financial knowledge among iranian investors. international journal of consumer studies, 43, 503–513. davutyan, n., & öztürkkal, b. (2016). determinants of saving-borrowing decisions and financial inclusion in a high middle income country: the turkish case. emerging markets finance and trade, 52, 2512–2529. disney, r., & gathergood, j. (2013). financial literacy and consumer credit portfolios. journal of banking & finance, 37, 2246–2254. dunn, l. f., & mirzaie, i. a. (2012). determinants of consumer debt stress: differences by debt type and gender. department of economics, ohio state university, columbus, oh. dunn, l. f., & mirzaie, i. a. (2016). consumer debt stress, changes in household debt, and the great recession. economic inquiry, 54, 201–214. fernandes, d., lynch, j. g., & netemeyer, r. g. (2014). financial literacy, financial education and downstream financial behaviors.management science, 60, 1861–1883. fünfgeld, b., & wang, m. (2009). attitudes and behaviour in everyday finance: evidence from switzerland. international journal of bank marketing, 27, 108–128. garcı́a, m. j. r. (2013). financial education and behavioral finance: new insights into the role of information in financial decisions. journal of economic surveys, 27, 297–315. gorbachev, o., & luengo-prado, m. j. (2019). the credit card debt puzzle: the role of preferences, credit access risk, and financial literacy. the review of economics and statistics, 101, 294–309. grable, j. e., lyons, a. c., & heo, w. (2019). a test of traditional and psychometric relative risk tolerance measures on household financial risk taking. finance research letters, 30, 8–13. grinstein-weiss, m., spader, j. s., yeo, y. h., key, c. c., & freeze, e. b. (2012). loan performance among low-income households: does prior parental teaching of money management matter? social work research, 36, 257–270. m. s. tahir et al. / financial services review 28 (2020) 273–301 299 harari, d. (2018). household debt: statistics and impact on economy. house of commons library briefing paper, (no. 7584). london, united kingdom. huston, s. j. (2012). financial literacy and the cost of borrowing. international journal of consumer studies, 36, 566–572. jacoby, m. b. (2002). does indebtedness influence health? a preliminary inquiry. the journal of law, medicine & ethics : a journal of the american society of law, medicine & ethics, 30, 560–571. jarl, j., cantor-graae, e., chak, t., sunbaunat, k., & larsson, c. a. (2015). trauma and poor mental health in relation to economic status: the case of cambodia 35 years later. plos one, 10, e0136410. kadoya, y., khan, m. s. r., hamada, t., & dominguez, a. (2018). financial literacy and anxiety about life in old age: evidence from the usa. review of economics of the household, 16, 859–878. kaplan, g., & violante, g. l. (2014). a model of the consumption response to fiscal stimulus payments. econometrica, 82, 1199–1239. kempson, e., finney, a., & poppe, c. (2017). financial well-being: a conceptual model and preliminary analysis. oslo and akershus university college of applied sciences, norway. kennedy, p. (2003). a guide to econometrics (5th ed.). cambridge, ma: the mit press. lin, l., revindo, m. d., gan, c., & cohen, d. a. (2019). determinants of credit card spending and debt of chinese consumers. international journal of bank marketing, 37, 545–564. long, j. s., & freese, j. (2006). regression models for categorical dependent variables using stata (2nd ed.). college station, tx: stata press. lusardi, a., & mitchell, o. s. (2007). financial literacy and retirement preparedness: evidence and implications for financial education. business economics, 42, 35–44. lusardi, a., & mitchell, o. s. (2011). financial literacy and retirement planning in the united states. journal of pension economics and finance, 10, 509–525. lusardi, a., & mitchell, o. s. (2013). the economic importance of financial literacy: theory and evidence. working paper, (no. 18952). cambridge, ma. lusardi, a., & tufano, p. (2015). debt literacy, financial experiences, and overindebtedness. journal of pension economics and finance, 14, 332–368. martin, t. k., & finke, m. (2014). a comparison of retirement strategies and financial planner value. journal of financial planning, 27, 46–53. mussida, c., & sciulli, d. (2019). does the presence of a disabled person in the household affect the employment probabilities of cohabiting women? evidence from italy, france and the uk. journal of family and economic issues, 40, 338–351. norvilitis, j. m., merwin, m. m., osberg, t. m., roehling, p. v., young, p., & kamas, m. m. (2006). personality factors, money attitudes, financial knowledge, and credit-card debt in college students. journal of applied social psychology, 36, 1395–1413. oecd. (2016). oecd/infe international survey of adult financial literacy competencies. available at https:// www.oecd.org/finance/oecd-infe-international-survey-of-adult-financial-literacy-competencies.pdf ottaviani, c., & vandone, d. (2018). financial literacy, debt burden and impulsivity: a mediation analysis. economic notes, 47, 439–454. remund, d. l. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. reyers, m. (2019). financial capability and emergency savings among south africans living above and below the poverty line. international journal of consumer studies, 43, 335–347. richards, d. w., ahmed, a. d., & tahir, m. s. (2019). addressing the challenge of problematic debt: australia and eurozone. available at https://www.thinkforwardinitiative.com/research/addressing-the-challenge-ofproblematic-debt-australia-and-eurozone richardson, s. (2013). a reflection on the household, income and labour dynamics in australia survey. the australian economic review, 46, 216–222. robb, c. a. (2011). financial knowledge and credit card behavior of college students. journal of family and economic issues, 32, 690–698. 300 m. s. tahir et al. / financial services review 28 (2020) 273–301 shih, t., & ke, s. (2014). determinates of financial behavior: insights into consumer money attitudes and financial literacy. service business, 8, 217–238. solis, o., & ferguson, r. (2017). the relationship of student loan and credit card debt on financial satisfaction of college students. college student journal, 51, 329–336. soman, d., & cheema, a. (2002). the effect of credit on spending decisions: the role of the credit limit and credibility.marketing science, 21, 32–53. stewart, m. b. (2007). the interrelated dynamics of unemployment and low-wage employment. journal of applied econometrics, 22, 511–531. summerfield, m., bevitt, a., freidin, s., hahn, m., la, n., macalalad, n., . . . wooden, m. (2017). hilda user manual release 16. melbourne institute of applied economic and social research, university of melbourne, melbourne. tippett, r. m. (2010). household debt across the life course: an analysis of the late baby boomers (doctor of philosophy). duke university, durham, nc. turunen, e., & hiilamo, h. (2014). health effects of indebtedness: a systematic review. bmc public health, 14, 489. van rooij, m. c. j., lusardi, a., & alessie, r. j. m. (2011). financial literacy and retirement planning in the netherlands. journal of economic psychology, 32, 593–608. watson, n., & wooden, m. (2010). the hilda survey: progress and future developments. the. australian economic review, 43, 326–336. west, t., & worthington, a. c. (2014). macroeconomic conditions and australian financial risk attitudes, 2001– 2010. journal of family and economic issues, 35, 263–277. wooden, m., freidin, s., & watson, n. (2002). the household, income and labour dynamics in australia (hilda) survey: wave 1. the australian economic review, 35, 339–348. worthington, a. c. (2006). predicting financial literacy in australia. financial services review, 15, 59–79. worthington, a. c. (2013). financial literacy and financial literacy programmes in australia. journal of financial services marketing, 18, 227–240. xiao, j. j., & o’neill, b. (2016). consumer financial education and financial capability. international journal of consumer studies, 40, 712–721. xiao, j. j., sorhaindo, b., & garman, e. t. (2006). financial behaviours of consumers in credit counselling. international journal of consumer studies, 30, 108–121. zhang, j., & kemp, s. (2009). the relationships between student debt and motivation, happiness, and academic achievement. new zealand journal of psychology, 38, 24–29. m. s. tahir et al. / financial services review 28 (2020) 273–301 301 age when first employed and retirement wealth of baby boomers hyungsoo kim, ph.d.a, serah shin, ph.d.a,*, qun zhanga, martie gillen, ph.d., mbab adepartment of family sciences, university of kentucky, 315 funkhouser building, lexington, ky 40506-0054, usa bdepartment of family, youth and community sciences, university of florida, po box 110310, gainesville, fl 32611-0310, usa abstract this study examines how age when first employed is related to retirement savings in later years. using data from the health and retirement study, we investigate two specific questions: has age when first employed affected the retirement wealth of baby boomers? if so, to what extent? the results show that age when first employed is negatively associated with accumulated retirement wealth in later years. for college graduates (high school graduates), delaying the start of employment cost $35,103 ($7,534) per year in retirement savings after controlling for demographic characteristics, number of working years, and occupation types. © 2018 academy of financial services. all rights reserved. jel classification: d14; d3; j21 keywords: retirement savings; timing of employment; baby boomers; young adults 1. introduction the importance of starting early to save for retirement has been emphasized in introductory personal finance textbooks and in everyday life (garman and forgue, 2015; munnell, webb, and hou, 2014). the three major determinants of savings outcome at retirement are * corresponding author. tel.: �1-859-539-7136; fax: �1-859-257-3212. e-mail address: serah.shin3@gmail.com (s. shin) financial services review 27 (2018) 29-45 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. (1) amount of retirement savings contributions every month or year (how much), (2) rate of return from the contributions, and (3) number of years to contribute (how long). individuals can increase contribution rates to their retirement savings. however, this may not be easy because of competing needs under limited income. currently, american workers’ average contribution rate including matching contribution from employers for retirement savings is 10.9% of their income (utkus and young, 2017). this rate is below the recommended 15% to maintain their living standard before retirement (garman and forgue, 2015). average and median retirement account balances are low. according to a recent survey, in 2016, the mean and median retirement savings are $96,495 and $24,713, respectively. even near retirees ages 55–64 have only $178,963 (mean) and $66,643 (median) in retirement savings (utkus and young, 2017). individuals may be able to increase the rate of returns by accepting more risk through investing in stocks and bonds. however, accepting more risk can result in losing money for retirement savings (markowitz, 1952). empirical studies report the cost of losing investments from bad outcomes (bridges, gesumaria, and leonesio, 2010) despite the possibility of high returns from a riskier asset allocation (poterba, rauh, venti, and wise, 2003). in fact, american workers already have substantially higher portions of risky assets (stocks) in their investment portfolio from 87% (younger than age 30) to 56% (near retirees ages 60–64). however, despite this asset allocation with risky assets, their outstanding balances especially in near retirees are considerably low as aforementioned. having more years of contribution unambiguously can bring about better retirement savings outcomes compared with the limitations of contribution rate and rate of return. having more years of saving or investing can further be benefited from amount of time for a compound return that produces additional returns from the returns earned during previous periods in addition to the investment principal. thus, starting to save early for retirement, as a young adult, is of great importance. if people start saving at younger ages (e.g., early 20s) for retirement, a small amount of money can make a big difference in the accumulation of retirement savings because of the effects of compounding. for example, a worker who starts saving $50 per week at age 22 will have a million dollars given an 8% rate of return by age 66 whereas someone who starts saving at age 30 will have about $540,000 or about half of the age 22 amount (garman and forgue, 2015). starting early to save for retirement may require individuals to be employed as soon as possible. americans’ age when first employed, however, has increased over time. according to the current population survey, the average age of american men (women) entering the job market has increased by about two full years during the recent 14 year span from age 19.1 (21.9) in 2000, 19.5 (22.5) in 2007 and 21 (22.8) in 2014 (kamenov, 2016). temporary economic slowdowns could contribute to this trend. many people also decide to further their education in hopes of being more competitive in the job market and obtaining higher pay. in fact, college (graduate school) enrollment rates have increased from 45% (18%) in 1990 to 57% (26%) in 2015 (davis, kimball, and gould, 2015). despite the importance, empirical research to investigate the actual effect of age when first employed on retirement wealth is limited. this study examines determinants of retirement savings focusing on how age when entering the labor market is related to retirement savings in later years. two specific questions are investigated: has age when first employed affected 30 h. kim et al. / financial services review 27 (2018) 29-45 retirement wealth of baby boomers? if so, to what extent? answering these questions provides significant implications for retirement savings not only for near retirees but also for young americans with regards to labor force entry and whether to start saving for retirement at younger ages. the findings from this study also contribute to understanding a new and important determinant of retirement savings. the remainder of this article is organized as follows. section 2 describes the retirement savings of baby boomers and timing of first employment. section 3 introduces the theoretical overview. section 4 provides an overview of the methods. section 5 provides details of the results. finally, section 6 includes a discussion. 2. background 2.1. baby boomers and retirement savings baby boomers continue to receive considerable attention from researchers and policy makers regarding retirement savings. their current status of low retirement savings amounts could contribute to this attention. in fact, 40% of baby boomers have no retirement savings (insured retirement institute, 2015). about 62% of working households ages 55–64 (a core group of baby boomers) have retirement savings amounts less than one times their annual income (rhee and boivie, 2015). this is far below the recommendations for retirement savings from experts (11 times annual income), causing individual financial insecurity among this group (kadlec, 2012). in addition, baby boomers are also on the verge of experiencing a critical change in their financial life stage according to life cycle theory of savings: a shift from the saving domain to the dissaving domain regarding retirement savings. the majority of baby boomers are also the first generation to experience defined contribution plans such as a 401(k). before 1978, a defined benefit plan was the dominant employee pension plan provided by employers (employee benefit research institute, 2007). under defined benefit plans, retirees receive a fixed portion (e.g., 50%) of their pre-retired salary until their death. from 1978 when baby boomers were 19 to 30 years old, however, employers started to provide defined contribution plans instead of defined benefit plans. in fact, 69% of baby boomers do not have a defined benefit plan (insured retirement institute, 2015). under defined contribution plans, employees make their own decisions and take sole responsibility for their retirement savings outcomes. the baby boomer experience provides an opportunity to examine the relationship between when to start saving and retirement savings outcomes, but also can provide more meaningful implications for younger generations who share the same types of retirement plans. 2.2. issues concerned with timing of labor force entry age when first employed can also be important in projecting the growth of retirement savings contributions specifically related to the benefits of compound return. for example, let’s assume two individuals (person a and person b) work for the same number of years, 31h. kim et al. / financial services review 27 (2018) 29-45 (i.e., 30 years). if person a saves $500 each year starting at age 22 for 30 years (age 22 to 52) at a 7% annual rate of return and then allows the balance to grow without further contribution for an additional 10 years, she would have $99,400 in savings at age 62. however, if person b starts saving at age 32 (10 year later) and continues saving $500 each year for 30 years (age 32 to 62), she would have $50,500 at age 62; only half of the amount saved compared with person a who started saving at a younger age. many americans tend to change employers more frequently than prior generations, particularly male workers (copeland, 2010; u.s. bureau of labor statistics, 2016c). in addition, they may experience underemployment or have unpaid leave for child or elder care (economic policy institute, 2017). during these unemployed periods, savings contributed for retirement in earlier years will keep growing because of compound return, boosting retirement savings in later years. in recent years the timing of first entry in the job market shows a trend of individuals entering the job market at older ages. late entry into the job market may have pros and cons for retirement savings through a person’s life cycle. for example, obtaining a higher level of education can increase an individual’s earning capacity. a worker with a high school diploma can expect to earn $1.3 million over their lifetime, whereas a worker with a bachelor’s or a master’s degree will earn $2.3 million or $2.7 million, respectively (carnevale, rose, and cheah, 2011). at the same time, the pursuit of higher education may decrease the number of working years and benefits of compound return for retirement savings. for example, life transition events may be delayed such as marriage or having children along with economic activities such as first time home or car buying. these sequential delays may make retirement savings more difficult in middle working years. however, the research on age when first employed is limited when examining retirement savings outcomes despite the importance of starting to save at a young age and the benefits of compound return. 2.3. other issues on accumulated wealth many previous studies show that demographic characteristics such as gender, race, and marital status are associated with wealth. women tend to have lower wealth than men mostly because of lower life time earnings (ruel and hauser, 2013). racial minorities tend to have lower wealth than white adults because they have lower income amounts, have less access to financial services, and lower levels of financial literacy (employee benefit research institute, 2003; oliver and shaprio, 2006). married couples tend to have higher net worth and larger wealth accumulation than single households because of the well-known marriage benefits including economics of scale (di, belsky, and liu, 2007; schmidt and sevak, 2006). 3. theoretical overview life-cycle and permanent income theory, which is a dominant economic theory used to explain savings and consumption, suggests that age (life-cycle) and income are main predictors of accumulated wealth (friedman, 1957; modigliani and brumberg, 1954). the 32 h. kim et al. / financial services review 27 (2018) 29-45 findings from extensive empirical studies overall have supported the two important determinants of accumulated wealth as this theory predicts (attanasio and weber, 2010; browning and crossley, 2001; xiao, ford, and kim, 2011). this theory indicates that individuals maximize their lifetime utility under resource constraint by allocating a certain proportion of their lifetime resources to their consumption and savings at each period of time. this allocation may require individuals to borrow in earlier years of their life, save and pay off debt in middle years, and dis-save for living expenses in later years over their life-cycle (age) (modigliani and brumberg, 1954). the resource constraint implies that individuals should consume within the extent of the sum of net worth inherited or carried over from previous years if any and earned income over their working years (lifetime/permanent income) (friedman 1957; modigliani and brumberg, 1954). this is called the standard intertemporal budget constraint from which asset evolution over time is derived (deaton and paxson, 2000). the asset evolution equation is used to estimate retirement savings at the current or retirement age after considering the age when first employed. an individual’s retirement wealth (rw) can be expressed with a common asset evolution equation (azar, 2012; deaton and paxson, 2000): rwt�1 � �1 � r� � rwt � � yt � ct� (1) where rwt�1 is retirement wealth at one year later than year t, r is the real interest rate, yt is income, ct is consumption, and �yt � ct� indicates savings at year t. cumulating this equation to current or retirement age (a) from a starting age (a0) when first employed, we can express rwa � �1 � r�a�a0 � rwa0 � �k�0 a�a0�1�1 � r�k� ya�k�1 � ca�k�1� (2) where rwa is retirement wealth at age a, rwa0 is retirement wealth at age a0. let’s illustrate the asset evolution equation above. if a person starts working at age 25, retirement wealth at the beginning of age 60 from the asset evolution equation above is the sum of the initial wealth at age 25 and its growth for 35 years ��1 � r�60�25 � rw25�, and savings each year from age 25 to 59 and its growth during associated numbers of years. for example, the money saved at age 25 is evolved to �1 � r�34�y25 � c25�, and the money saved at age 26 is evolved to �1 � r�33�y26 � c26�. likewise, the money saved at age 59 is �1 � r�0�y59 � c59� at the beginning of age 60. all other things being constant (e.g., current or retirement age (a) and (y � c)), their starting age (a0) at which first employed is a major determinant of retirement wealth, and the levels of retirement savings at current ages are lower as age when first employed is higher. 4. method 4.1. empirical model based on the eq. (2), retirement savings are theoretically calculated from the initial wealth (rwa0 ), yearly contributions to savings (y � c), current or retirement age (a), age when first 33h. kim et al. / financial services review 27 (2018) 29-45 employed (a0), and investment returns (r). empirically, however, the initial wealth (rwa0 ), yearly contributions to savings (y � c) are barely available from general survey data sets including hrs data. thus, we substitute those theoretical concepts with more tangible measurements using related variables and proxies. rwa � � � �1 a0 � �2 l � x� � � (3) where rwa is retirement wealth at current age a, a0 is age when first employed, which is the main interest of this study. l is the number of working years at age a, which also represents the number of years to save for retirement from eq. (2). for the initial wealth (rwa0 ), we may assume most people do not have any retirement savings when they have a first job after high school and college. x is a set of other control variables including demographic characteristics and occupation information, which are assumed to be major influential factors on individual’s contributions to savings (y � c). in this empirical model, �1 � 0 indicates that an older age when first employed is related to lower retirement wealth. 4.2. data in this study, the 2014 health and retirement study (hrs) is used. the hrs is a representative sample of middle-aged and older americans with age 50 or older sponsored by the national institute on aging (nia). the hrs provides comprehensive information about the work history and financial status such as levels of income or current retirement wealth. the 2014 hrs interviewed 18,748 respondents who were born before 1961. for the study purpose described above, the sample was restricted to respondents who belong to the baby boomer generation (born in 1948 to 1959) based on the hrs classification. among the 18,748 respondents, identified baby boomers were 8,002 of which 6,772 provided the information about their age when first employed. we further consider two criteria in constructing the final analysis samples: educational attainment and associated age when first employed. lifetime earnings have significantly differed by high school graduates and college graduates, which are not homogenous. potential confounding effects also arise from respondents with unconventional life transitions (e.g., first employed at age 16 and graduated college at age 30). to mitigate effects from educational differences and unconventional life transitions, we focus on the respondents who followed relatively normative education to employment transitions. these respondents are characterized into two subgroups: (1) high school graduate and age 18 or older when first employed, (2) college graduate� and age 22 or older when first employed. accordingly, we construct two subsample sets for our analysis. of the 6,772 respondents who were baby boomers and provided age when first employed, 3,923 had educational attainment with high school or more and were first employed at age 18 or older up to age 35. from the 3,923 respondents, our two final subsamples include: (1) high school graduates and age 18 or older when first employed (n � 2,211) and (2) college or more and age 22 or older when first employed (n � 883). for the college or more graduates, age 22 was based on the average or median age of current and past college graduation according to the u.s. census bureau: currently about 60% of college graduates were ages 22 to 23 when graduating 34 h. kim et al. / financial services review 27 (2018) 29-45 (spreen, 2013), and the median age when graduating college in 1960 was 22.9 years (u.s. census bureau, 1963). in an era when 70% of high school graduates in the u.s. are enrolled in colleges or universities (u.s. department of labor, 2016b), focusing on this population who were first employed at age 22 or older can provide more meaningful implications for young college graduates about retirement savings. in our sample, the mean age when first employed was 20.8 years for the high school graduate group and 23.9 years for the college graduate group. 4.3. measures 4.3.1. age when first employed in the hrs, a respondent is asked about past jobs retrospectively at their first interview, including the information about the earliest year worked. the specific question to measure age when first employed is “in what year did you first work for six months or more?” this question is included in each wave of the hrs. we calculated age when first employed by subtracting the birth year from the year at first work. 4.3.2. dependent variables: retirement wealth household net worth (including secondary residence) is used as a main dependent variable. net worth is calculated as the sum of all assets less all debt. 4.3.3. other control variables the following demographic characteristics are included as control variables; age in 2014, gender, race, marital status in 2014, and the number of people in the household. the number of people in the household includes the respondent, spouse, residents and nonresident children. we also include the respondent’s and spouse’s total number of years worked. total number of years worked is the total number of years that the respondents actually worked. this variable is derived from the respondent’s retrospective job history and all jobs reported since the first interview and considers all related factors affecting individual’s labor force status such as a layoff or voluntary unemployment. for those who are single or have a stay at home spouse, the value of the spouse’s total number of years worked is coded as 0. occupation type or annual salary of the longest tenured job is also included as a variable to control the individual’s lifetime income, which is classified into 23 categories based on the 2010 u.s. census occupational classification system. 4.4. analysis descriptive analyses and ols regressions are used to estimate the impact of age when first employed on retirement wealth. descriptive analyses explore the overall trend of the relationship between age when first employed and retirement wealth. ols regressions are used to specify the effect of age when first employed on accumulated retirement wealth in later years. two separate regressions are conducted with the two different groups of high school graduates and college graduates. weighted data are used to represent the u.s. middle aged and older population for our analysis.1 for sensitivity checks, the main model is 35h. kim et al. / financial services review 27 (2018) 29-45 estimated with several alternative specifications (unweighted regression, natural logarithm transformation of net worth2 and financial wealth3 as a dependent variable, or using occupation salary level4 instead of occupation type as a control variable). 5. results relevant characteristics of the sample based on levels of education are presented in table 1. the mean age in 2014 is 59.9 years for the high school group and 60.2 years for the college group, respectively. more females are in the high school group (56.7%) than the college group (50.0%), whereas less white and married individuals are in the high school group than the college group. the working years (about 30 years) are similar. not surprisingly, there are huge gaps of net worth and financial wealth between the two groups. the mean net worth of the college group is $798,700 which is 3.3 times larger than that ($241,100) of the high school group. the gap in financial wealth is 4.9 times ($214,500 college group vs. $44,100 high school group). fig. 1 illustrates current accumulated retirement wealth levels in 2014 by age when first employed for each group. panel a shows accumulated wealth of the high school group. overall, the level of wealth demonstrates a declining pattern by age when first employed: the older the entry age is, the less the wealth tends to be. in particular, when respondents were first employed at age 18, wealth accumulated by 2014 is more than $200,000. however, if age when first employed was 26 or older wealth on average is about $100,000. panel b presents accumulated wealth of the college group, showing a similar pattern to the high school group where wealth declines with respect to age when first employed. when first table 1 characteristics of the sample high school (age when first employed 18 to 35) college� (age when first employed 22 to 35) statistics (p)* n 2,211 883 age (mean, sd) 59.9 (3.5) 60.2 (3.4) �1.941 (.052) female (%) 56.7 50.0 12.877 (.000) white (%) 79.3 84.9 14.483 (.000) married/partnered (%) 63.0 73.5 35.459 (.000) n of people in the household† 2.3 (1.2) 2.3 (1.1) �.250 (.802) years worked (mean, sd) 30.4 (11.7) 30.5 (10.0) �1.745 (.081) spouse’s years worked (mean, sd) 19.9 (18.7) 22.3 (17.7) �3.239 (.001) net worth ($1,000) mean (sd) 241.1 (354.8) 798.7 (907.0) �15.960 (.000) median 116.0 460.0 financial wealth ($1,000) mean (sd) 44.1 (135.2) 214.5 (430.5) �9.817 (.000) median 3.0 40.7 age when first employed (mean, sd) 20.8 (3.3) 23.9 (2.2) * �2 for age, female, white, and married, t for the other variables. † number of people in the household including the respondent, spouse, and residents and non-resident children. 36 h. kim et al. / financial services review 27 (2018) 29-45 employed at age 22, right after college graduation, current accumulated wealth is about $700,000. when age increased to 29 or older, wealth levels are about $300,000. the results of the regression analyses by each group are presented in table 2 to estimate the relationship between age when first employed and retirement wealth. panel a shows the results of the high school group. model 1 includes a relevant variable of age when first employed only. model 2 adds the number of working years to model 1, and model 3 includes demographic characteristics as well as occupation type to model 2. the coefficients of age when first employed are statistically significant across all models and qualitatively consistent. thus, we present the results of model 3. age when first employed is negatively related fig. 1. accumulated retirement wealth (net worth) distribution by age when first employed. (a) high school (age when first employed 18 to 35). (b) college� (age when first employed 22 to 35). note: mean and 95% confidence interval. 37h. kim et al. / financial services review 27 (2018) 29-45 t ab le 2 r eg re ss io n re su lts (a ) h ig h sc ho ol (a ge w he n fir st em pl oy ed 18 to 35 ) m od el 1 m od el 2 m od el 3 b (s e ) � p b (s e ) � p b (s e ) � p a ge w he n fir st em pl oy ed � 14 ,0 97 (2 ,3 93 ) � .1 26 .0 00 � 8, 97 7 (2 ,4 14 ) � .0 81 .0 00 � 7, 53 4 (2 ,5 17 ) � .0 68 .0 03 y ea rs w or ke d 5, 99 9 (6 56 ) .1 98 .0 00 3, 60 0 (7 62 ) .1 19 .0 00 a ge 2, 19 2 (2 ,4 66 ) .0 22 .3 74 g en de r fe m al e � 18 ,8 16 (1 9, 25 1) � .0 26 .3 28 r ac e n on -w hi te � 89 ,6 38 (1 9, 77 2) � .1 03 .0 00 m ar ita l st at us si ng le /u nm ar ri ed � 48 ,7 84 (2 4, 75 8) � .0 67 .0 49 n of pe op le in th e ho us eh ol d � 10 ,1 96 (7 ,1 47 ) � .0 34 .1 54 sp ou se ’s ye ar s w or ke d 3, 95 3 (6 46 ) .2 09 .0 00 o cc up at io n ty pe m an ag em en t 63 ,9 28 (3 3, 96 9) .0 46 .0 60 b us in es s/ fin an ci al op er at io n 3, 54 1 (4 5, 59 2) .0 02 .9 38 c om pu te r/ m at he m at ic al � 80 ,9 46 (7 4, 79 2) � .0 24 .2 79 a rc hi te ct ur e/ en gi ne er in g 39 ,5 49 (6 7, 91 7) .0 13 .5 60 l if e/ ph ys ic al /s oc ia l sc ie nc e � 76 5 (1 76 ,5 34 ) .0 00 .9 97 c om m un ity /s oc ia l se rv ic e � 64 ,7 43 (9 5, 47 4) � .0 15 .4 98 l eg al 57 ,2 59 (2 18 ,8 39 ) .0 06 .7 94 e du ca tio n/ tr ai ni ng /li br ar y � 3, 48 6 (5 2, 60 9) � .0 01 .9 47 a rt s/ de si gn /s po rt s/ m ed ia � 30 ,1 27 (8 1, 15 2) � .0 08 .7 11 h ea lth ca re pr ac tit io ne r 46 ,6 13 (5 1, 24 5) .0 21 .3 63 h ea lth ca re su pp or t � 86 ,8 45 (4 6, 79 3) � .0 42 .0 64 pr ot ec tiv e se rv ic e � 77 ,0 09 (5 3, 98 9) � .0 33 .1 54 fo od pr ep ar at io n/ se rv in g 91 2 (4 8, 80 0) .0 00 .9 85 b ui ld in g/ gr ou nd cl ea ni ng � 10 2, 59 9 (3 6, 14 8) � .0 68 .0 05 pe rs on al ca re /s er vi ce 53 ,2 99 (4 5, 04 6) .0 27 .2 37 sa le s 48 ,0 56 (2 9, 57 0) .0 41 .1 04 fa rm in g/ fis hi ng /f or es tr y � 37 ,7 02 (1 41 ,4 01 ) � .0 06 .7 90 c on st ru ct io n/ ex tr ac tio n � 16 ,0 87 (3 5, 06 9) � .0 12 .6 46 in st al la tio n/ m ai nt en an ce /r ep ai r 41 ,1 34 (4 3, 86 0) .0 23 .3 48 pr od uc tio n � 98 ,0 76 (2 9, 28 9) � .0 89 .0 01 t ra ns po rt at io n/ m at er ia l m ov in g � 31 ,5 76 (3 5, 87 8) � .0 23 .3 79 m ili ta ry � 13 6, 18 3 (1 18 ,9 95 ) � .0 25 .2 53 in te rc ep t 52 6, 59 2 (4 9, 06 5) .0 00 24 ,6 41 (5 7, 40 0) .0 00 16 0, 50 5 (1 47 ,9 89 ) .2 78 a dj us te d r 2 .0 16 .0 52 .1 52 (c on ti nu ed on ne xt pa ge ) m al e, w hi te , m ar ri ed , an d oc cu pa tio n ty pe of of fic e/ ad m in is tr at io n su pp or t ar e re fe re nc e gr ou ps . d ep en de nt va ri ab le is th e do lla r am ou nt s of ho us eh ol d ne t w or th . 38 h. kim et al. / financial services review 27 (2018) 29-45 t ab le 2 (c on tin ue d) (b ) c ol le ge � (a ge w he n fir st em pl oy ed 22 to 35 ) m od el 1 m od el 2 m od el 3 b (s e ) � p b (s e ) � p b (s e ) � p a ge w he n fir st em pl oy ed � 64 ,4 68 (1 5, 71 1) � .1 45 .0 00 � 46 ,1 30 (1 5, 96 8) � .1 04 .0 01 � 35 ,1 03 (1 6, 89 2) � .0 80 .0 38 y ea rs w or ke d 15 ,2 94 (3 ,2 06 ) .1 71 .0 00 8, 15 8 (3 ,8 93 ) .0 91 .0 36 a ge � 6, 86 2 (1 1, 35 9) � .0 26 .5 46 g en de r fe m al e 63 ,2 72 (7 4, 61 8) .0 35 .3 97 r ac e n on -w hi te � 22 3, 29 3 (9 1, 83 4) � .0 91 .0 15 m ar ita l st at us si ng le /u nm ar ri ed � 36 8, 42 1 (1 08 ,1 70 ) � .1 81 .0 01 n of pe op le in th e ho us eh ol d � 88 ,4 63 (3 3, 60 9) � .1 07 .0 09 sp ou se ’s ye ar s w or ke d 8, 81 7 (2 ,6 45 ) .1 73 .0 01 o cc up at io n ty pe m an ag em en t 27 6, 11 6 (1 43 ,0 02 ) .1 22 .0 54 b us in es s/ fin an ci al op er at io n 51 8, 51 4 (1 67 ,1 29 ) .1 58 .0 02 c om pu te r/ m at he m at ic al 44 9, 06 2 (2 12 ,3 60 ) .0 90 .0 35 a rc hi te ct ur e/ en gi ne er in g 34 8, 68 0 (2 37 ,9 73 ) .0 59 .1 43 l if e/ ph ys ic al /s oc ia l sc ie nc e 29 8, 95 2 (2 47 ,1 48 ) .0 48 .2 27 c om m un ity /s oc ia l se rv ic e 23 6, 65 2 (2 05 ,1 11 ) .0 49 .2 49 l eg al 47 1, 80 5 (2 00 ,7 86 ) .1 04 .0 19 e du ca tio n/ tr ai ni ng /li br ar y 79 ,0 24 (1 42 ,6 10 ) .0 33 .5 80 a rt s/ de si gn /s po rt s/ m ed ia � 6, 93 5 (2 38 ,3 89 ) � .0 01 .9 77 h ea lth ca re pr ac tit io ne r 12 6, 62 7 (1 59 ,9 26 ) .0 40 .4 29 h ea lth ca re su pp or t � 15 3, 40 5 (5 59 ,2 65 ) � .0 10 .7 84 pr ot ec tiv e se rv ic e 26 4, 66 3 (4 01 ,2 28 ) .0 24 .5 10 fo od pr ep ar at io n/ se rv in g � 27 0, 17 2 (4 08 ,5 22 ) � .0 24 .5 09 b ui ld in g/ gr ou nd cl ea ni ng � 11 8, 75 9 (3 75 ,2 28 ) � .0 12 .7 52 pe rs on al ca re /s er vi ce � 13 4, 13 5 (2 31 ,2 84 ) � .0 24 .5 62 sa le s 28 1, 46 2 (1 64 ,3 55 ) .0 87 .0 87 c on st ru ct io n/ ex tr ac tio n 18 3, 31 2 (3 26 ,3 61 ) .0 21 .5 75 in st al la tio n/ m ai nt en an ce /r ep ai r � 50 9, 46 3 (4 36 ,5 72 ) � .0 43 .2 44 pr od uc tio n � 46 ,2 50 (2 37 ,9 10 ) � .0 08 .8 46 t ra ns po rt at io n/ m at er ia l m ov in g � 11 ,5 41 (3 02 ,2 59 ) � .0 01 .9 70 m ili ta ry � 54 9, 54 7 (6 33 ,1 95 ) � .0 31 .3 86 in te rc ep t 2, 31 3, 57 7 (3 70 ,5 63 ) .0 00 1, 41 7, 08 5 (4 11 ,0 10 ) .0 04 1, 70 8, 77 3 (7 54 ,4 63 ) .0 24 a dj us te d r 2 .0 20 .0 46 .1 57 m al e, w hi te , m ar ri ed , an d oc cu pa tio n ty pe of of fic e/ ad m in is tr at io n su pp or t ar e re fe re nc e gr ou ps . d ep en de nt va ri ab le is th e do lla r am ou nt s of ho us eh ol d ne t w or th . 39h. kim et al. / financial services review 27 (2018) 29-45 to retirement wealth. this result indicates that as age when first employed increases, accumulated retirement wealth declines after controlling for the number of working years, occupation types, and relevant demographic characteristics. the coefficient of age when first employed is -7,534, indicating that starting employment one year later in their 20s or 30s leads to $7,534 less retirement wealth in later years after controlling for the number of years employed. panel b presents the results of the college group. consistent with the high school group in models 1, 2 and 3, the results show that accumulated wealth declines with an increase in age when first employed. the coefficient of age when first employed in model 3 is -35,103, indicating that starting employment one year later in their 20s or 30s leads to $35,103 less retirement wealth in later years after controlling for the number of years employed. not surprisingly, non-white and unmarried individuals have lower retirement wealth for both groups. current age and the number of years employed have a positive relationship with accumulated wealth for the high school group but not for the college group after controlling for other variables. among occupations, management is related to higher accumulated retirement wealth (reference group: office and administration support), whereas service and production occupations lead to lower accumulated wealth in the high school group. for the college group, occupations in management, business/finance, and sales show significantly higher accumulated retirement wealth. model 3 is also used to conduct several sensitivity analyses with alternative specifications such as unweighted regression, natural logarithm transformation of net worth, financial wealth as a dependent variable, and average salary level of occupation rather than type of occupation. table 3 shows the regression results of these alternative specifications for high school (panel a) and college (panel b) groups, highlighting that age when first employed has a consistent negative impact on accumulated retirement wealth. sensitivity of the results is further investigated by including lump sum money such as inheritance or life insurance benefits, and replacing the occupation types representing lifetime earnings with current income. the results of these additional analyses are consistent with our main findings (the results are available upon request). 6. discussion the extent to which age when first employed is related to accumulated retirement wealth in later years is examined. using data from the health and retirement study, this study focuses on baby boomers who were first employed at or after the age of 18 (high school graduates) or 22 (college or more graduates). overall findings reflect that age when first employed has a negative relationship with accumulated retirement wealth in later years. for college or more graduates (high school graduates), delaying employment cost $35,103 ($7,534) per year in retirement savings after controlling for demographic characteristics, the number of working years, and occupation types. to the extent of our knowledge, this finding is new and a result of the first empirical study to examine the impact of age when first employed on retirement savings in later years. previous studies primarily considered the number of years worked but ignored age when first 40 h. kim et al. / financial services review 27 (2018) 29-45 t ab le 3 r eg re ss io n re su lts w ith al te rn at iv e sp ec ifi ca tio ns (a ) h ig h sc ho ol (a ge w he n fir st em pl oy ed 18 to 35 ) u nw ei gh te d l n (n et w or th ) o cc up at io n sa la ry * fi na nc ia l w ea lth b (s e ) p b (s e ) p b (s e ) p b (s e ) p a ge w he n fir st em pl oy ed � 4, 59 4 (1 ,8 81 ) .0 15 � .0 06 (. 00 3) .0 20 � 7, 08 8 (2 ,5 19 ) .0 05 � 2, 35 8 (7 14 ) .0 01 (d em og ra ph ic s in cl ud ed ) y es y es y es y es (o cc up at io n ty pe s in cl ud ed ) y es y es n o y es in te rc ep t 11 9, 82 0 (1 29 ,3 40 ) .3 54 13 .2 07 (. 16 7) .0 00 93 ,7 39 (1 47 ,6 45 ) .5 26 12 ,5 04 (4 1, 97 4) .7 66 a dj us te d r 2 .1 74 .2 08 .1 44 .0 81 (b ) c ol le ge � (a ge w he n fir st em pl oy ed 22 to 35 ) u nw ei gh te d l n (n et w or th ) o cc up at io n sa la ry * fi na nc ia l w ea lth b (s e ) p b (s e ) p b (s e ) p b (s e ) p a ge w he n fir st em pl oy ed � 23 ,9 32 (1 2, 13 7) .0 49 � .0 29 (. 01 1) .0 12 � 37 ,7 22 (1 6, 33 6) .0 21 � 16 ,2 36 (7 ,9 27 ) .0 41 (d em og ra ph ic s in cl ud ed ) y es y es y es y es (o cc up at io n ty pe s in cl ud ed ) y es y es n o y es in te rc ep t 1, 19 8, 42 9 (6 12 ,2 45 ) .0 51 14 .4 30 (. 51 1) .0 00 1, 86 6, 60 6 (7 29 ,1 44 ) .0 11 63 4, 29 0 (3 54 ,0 80 ) .0 74 a dj us te d r 2 .1 63 .2 44 .1 56 .0 65 * r eg re ss io n re su lt w he n oc cu pa tio n w as co nt ro lle d by av er ag e sa la ry le ve l. o cc up at io n sa la ry le ve l w as cl as si fie d th e 23 ca te go ri es in to fo ur gr ou ps (q ua rt er ly ) ba se d on th ei r m ed ia n an nu al sa la ry at 20 14 (h ttp :// w w w .b ls .g ov /o es /2 01 4/ m ay /o es _n at .h tm ). 41h. kim et al. / financial services review 27 (2018) 29-45 employed in calculating lifetime earnings and estimating related retirement savings (dornbusch and fischer, 1994; modigliani and brumberg, 1954; munnell et al., 2014). this study takes into account age when first employed as well as the number of years worked to estimate retirement savings in later years. starting employment as a young adult is important for retirement savings because it provides resources that can be saved for retirement thereby providing the benefits of compound return from savings in early years. in particular, the benefit of compound return is more important. in fact, our results imply the importance of compound return for starting early. one year delay of starting work in their younger years cost college graduates $35,103 per year ($7,534 for high school graduates) in retirement savings while one additional year of work only adds $8,158 ($3,600 for high school graduates) to retirement savings. the findings provide several implications for policy, education, and research about retirement savings. starting employment as a young adult can be channeled into starting to save for retirement at an early age, leading to better retirement savings outcomes in later years. the results show that the financial losses of entering the job market one year later equates to about $35,103 ($7,534). this finding suggests that a more institutional and systematic effort is needed in the job search process to support students in securing employment as soon as possible after graduation. several colleges have recently implemented three-year acceleration graduation plans for students. this option would decrease the cost of college by eliminating one year and would have graduates into the labor market a year earlier. however, according to a survey of college students and recent graduates before the financial crisis in 2008, more than half of the respondents reported that they did not expect a job offer by graduation, and 16% went to graduate schools to postpone entry into a tight job market (kennedy, 2004). such delays on starting work as a young adult may cause insufficient retirement savings in later years. furthermore, college education may be one of the important opportunities to help young adults become more financially knowledgeable and confident. however, personal finance education is voluntary for many college students (harrington and smith, 2016). intensive supports to increase student participation may be needed. especially, for those students who lack financial awareness or literacy, more advertising targeted at these groups and offering more sections should be considered (beierlein and neverett, 2013). study findings also suggest that a change in the voluntary enrollment option often given to young employees in participating in employer sponsored retirement savings plans might be needed. currently among young employees who are at or under age 29, only 34% participate in retirement savings plans despite their employers providing retirement savings plans (the pew charitable trusts, 2016). employers who provide retirement savings plans such as a 401(k) have commonly mandated their employees’ contributions for their retirement savings. however, many employers do not require their employees including young employees, aged 30 or less, to contribute savings for retirement. even some employer sponsored retirement savings plans do not allow employees who are under age 21 or work less than a year to participate in their plan (u.s. department of labor, 2016a). this policy discourages young employees to start saving early for retirement even though they start work at a young age, and prevents them from taking advantage of the benefits of compound return. study findings suggest that this policy may need to be reconsidered or adjusted by employers 42 h. kim et al. / financial services review 27 (2018) 29-45 or policy makers for young americans to better prepare for retirement financially at younger ages. despite meaningful implications of study findings for individual retirement savings and policy makers, this study has several limitations resulting from the hrs data. a direct measure of savings could not be included in our estimation: whether to save and how much to save for retirement savings each year. the saving amount each period is a relevant factor of retirement savings outcomes in the conceptual model along with how long to work and save. the variable of periodic saving amount is not available in the data set. instead, education, career job type, or the associated average salary was controlled. although these proxy variables may not eradicate the impacts, the variables can help to mitigate the bias from missing periodic savings amount. empirical evidence suggests that higher education and associated higher salary increase the propensity to save for retirement (lusardi and mitchell, 2007; warne, 2013). as mentioned in the empirical model, we do not estimate a direct effect of compound return on retirement savings. instead, we approximate the effect with age when first employed. in other words, we assume that an individual will start to save for retirement when they have a first job. in addition, we cannot include existent savings at age when first employed if any because of data unavailability. although, most american young adults do not have any savings for retirement. further studies are necessary with related variables to capture compound return and savings for retirement at the age of first employment. additionally, further clarifications regarding first employment are needed. the measure of age when first employed is based on the following question: in what year did you first work for six months or more? this question can capture the age when first employed, but does not articulate whether it was full-time or part-time employment and career or temporary job. to mitigate potential bias from this limitation, respondents’ occupation type for the longest tenured job is included. further studies are needed to take into account these issues. notes 1 in the hrs sample, racial minorities are oversampled compared with the actual racial composition of all u.s. households. in addition, the sample contains an oversample of florida residents. the sample weights are constructed in a way to make the hrs weighted sample representative of all u.s. households or of all individuals in the u.s. population in the age eligible range (hrs, 2011). 2 net worth data tend to have a right-skewed distribution which can lead to biased estimates. thus, we conduct sensitivity tests using natural logarithm to mitigate bias from the right-skewed distribution. 3 financial wealth means the net value of non-housing financial wealth, which is calculated as the sum of stocks, bonds, funds, cds, checking and saving accounts, and other savings less debt. 4 occupation salary level was classified the 23 categories into four groups (quarterly) based on their median annual salary at 2014 (http://www.bls.gov/oes/2014/may/oes_nat.htm). 43h. kim et al. / financial services review 27 (2018) 29-45 occupation salary is an alternative way to control individuals’ lifetime income with less dummy variables. references attanasio, o. p., & weber, g. (2010). consumption and saving: models of intertemporal allocation and their implications for public policy. journal of economic literature, 48, 693–751. http://dx.doi.org/10.1257/ jel.48.3.693 azar, s. a. (2012). on the specification of the asset evolution equation in consumption models. applied economics letters, 19, 113–116. beierlein, j. j., & neverett, m. (2013). who takes personal finance? financial services review, 22, 151–171. bridges, b., gesumaria, r., & leonesio, m. v. (2010). assessing the performance of life-cycle portfolio allocation strategies for retirement saving: a simulation study. social security bulletin, 70, 23–43. (available at http://www.ssa.gov/policy/docs/ssb/) browning, m., & crossley, t. f. (2001). the life-cycle model of consumption and saving. journal of economic perspectives, 15, 3–22. http://dx.doi.org/10.1257/jep.15.3.3 carnevale, a. p., rose, s. j., & cheah, b. (2011). the college payoff: education, occupations, lifetime earnings. technical report. (available at https://repository.library.georgetown.edu/bitstream/handle/10822/559300/collegepayoff-complete.pdf?sequence�1) copeland, c. (2010). employee tenure trend lines, 1983–2010. employee benefit research institute. (available at https://www.ebri.org/pdf/notespdf/ebri_notes_12-dec10.tenure-cehcs.pdf) davis, a., kimball, w., & gould, e. (2015). the class of 2015: despite an improving economy, young grads still face an uphill climb. economic policy institute. (available at http://www.epi.org/publication/the-class-of2015/) deaton, a., & paxson, c. (2000). growth and saving among individuals and households. the review of economics and statistics, 82, 212–225. di, z. x., belsky, e., & liu, x. (2007). do homeowners achieve more household wealth in the long run? journal of housing economics, 16, 274–290. dornbusch, r., fischer, s., & startz, r. (1994). macroeconomics. new york, ny: mcgraw hill. economic policy institute. (2017). the state of working america: both genders suffer from sustained high underemployment. (available at http://www.stateofworkingamerica.org/charts/underemployment-gender/) employee benefit research institute. (2003). the 2003 minority retirement confidence survey: summary of findings. retrieved from https://www.ebri.org/surveys/rcs/2003/ employee benefit research institute. (2007). history of 401(k) plans: an update. (available at https://www. ebri.org/pdf/publications/facts/0907facts.pdf) friedman, m. (1957). a theory of the consumption function. princeton, nj: princeton university press. garman, t., & forgue, r. (2015). personal finance. boston, ma: cengage learning. harrington, c., & smith, w. (2016). college student interest in personal finance education. financial services review, 25, 351–372. hrs. (2011). updates to hrs sample weights. (available at https://hrs.isr.umich.edu/sites/default/files/biblio/ dr-013.pdf) insured retirement institute. (2015). baby boomers and retirement planning strategy. (available at https:// www.myirionline.org/docs/default-source/research/boomerstrategyoct2015.pdf?sfvrsn�2) kadlec, d. (2012, july 17). retirement savings: 11 times final pay is the new target. time. (available at http://business.time.com/2012/07/17/retirement-savings-11-times-final-pay-is-the-new-target/) kamenov, a. (2016, june 29). how hard is it to enter the job market? city-data. (available at http://www. city-data.com/blog/2456-hard-enter-job-market/) kennedy, d. g. (2004, june 1). eureka! american demographics. (available at http://adage.com/article/americandemographics/eureka/44994/) 44 h. kim et al. / financial services review 27 (2018) 29-45 lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. markowitz, h. (1952). portfolio selection. the journal of finance, 7, 77–91. http://dx.doi.org/10.2307/2975974 modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. in kurihara, k. k. post-keynesian economics (pp. 388–436). new brunswick, nj: rutgers university press. munnell, a. h., webb, a., & hou, w. (2014). how much should people save? center for retirement research at boston college. (available at http://crr.bc.edu/wp-content/uploads/2014/07/ib_14-11.pdf) oliver, m. l., & shapiro, t. m. (2006). black wealth/white wealth: a new perspective on racial inequality. new york, ny: routledge. poterba, j., rauh, j., venti, s., & wise, d. (2003). utility evaluation of risk in retirement saving accounts. national bureau of economic research. (available at http://www.nber.org/papers/w9892.pdf) rhee, n., & boivie, i. (2015). the continuing retirement savings crisis. national institute on retirement security. (available at http://www.nirsonline.org/storage/nirs/documents/rsc%202015/final_rsc_2015.pdf) ruel, e., & hauser, r. m. (2013). explaining the gender wealth gap. demography, 50, 1155–1176. http:// dx.doi.org/10.1007/s13524-012-0182-0 schmidt, l., & sevak, p. (2006). gender, marriage, and asset accumulation in the united states. feminist economics, 12, 139–166. spreen, t. l. (2013). recent college graduates in the u.s. labor force: data from the current population survey. monthly labor review, 136, 1–11. the pew charitable trusts. (2016). who’s in, who’s out: a look at access to employer-based retirement plans and participation in the states. (available at http://www.pewtrusts.org/�/media/assets/2016/01/retirement_ savings_report_jan16.pdf) u.s. bureau of labor statistics. (2016c). employee tenure in 2016. (available at https://www.bls.gov/news. release/tenure.nr0.htm) u.s. census bureau. (1963). current population report. (available at https://www.census.gov/hhes/socdemo/ education/data/cps/1960/p23-09.pdf) u.s. department of labor. (2016a). automatic enrollment 401(k) plans for small businesses. (available at https://www.irs.gov/pub/irs-pdf/p4674.pdf) u.s. department of labor. (2016b). college enrollment and work activity of 2015 high school graduates. (available at https://www.bls.gov/news.release/hsgec.nr0.htm) utkus, s. p., & young, j. a. (2017). how america saves 2017. (available at https://institutional.vanguard.com/ iam/pdf/has17.pdf) warne, l. (2013). more education, better retirement. national center for policy analysis. (available at http:// www.ncpa.org/pdfs/ib125.pdf) xiao, j. j., ford, m. w., & kim, j. (2011). consumer financial behavior: an interdisciplinary review of selected theories and research. family and consumer sciences research journal, 39, 399–414. http://dx.doi. org10.1111/j.1552-3934.2011.02078.x 45h. kim et al. / financial services review 27 (2018) 29-45 the effect of risk literacy and visual aids on portfolio choices among professional financial planners meghaan r. lurtz1, michael g. kothakota1,*, stuart j. heckman1, kristy archuleta2 1department of personal financial planning, kansas state university, 312 justin hall, 1324 lovers lane, manhattan, ks 66505, usa 2department of financial planning, housing & consumer economics, university of georgia, 205 dawson hall, athens, ga 30605, usa abstract financial planners and their clients come together regularly to discuss financial decisions, which are inherently risky. yet, financial planning research has not explored the impact of risk literacy (i.e., objective numeracy)—the ability to understand and interpret probabilistic trade-offs—and graph literacy on client-planner decision-making quality. this study uses an experimental design to test financial planners’ risk literacy and their ability to select the most resilient portfolio based on whether they were given probabilistic information and a visual representation or only probabilistic information. results indicate that visual representation do help financial planners determine the appropriate choice, but risk literacy does not. implications for financial planners and future research in this area are discussed. © 2021 academy of financial services. all rights reserved. keywords: financial planners; risk literacy; decision-making; client communication; experimental design 1. risk literacy among professional financial planners making risky choices is central to professional financial planning. the majority of financial decisions involve some risk. professional financial planners create financial plans and present financial information to help clients make financial decisions in the presence of risk. research from the field of judgment and decision-making highlight the importance of the concept known as risk literacy, which is defined as “the ability to accurately interpret and act on information *corresponding author. tel.: 919 247 9854; fax: 919 267 6732. e-mail address: michael.kothakota@wolfbridgefinancial.com 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 209–225 about risk” and is a synonym for statistical numeracy (cokely, galesic, schulz, ghazal, & garcia-retamero, 2012, p. 26). in the same way previous studies point out the need for risk literacy in the medical profession so doctors can better assist patients in making medical decisions involving risk (e.g., surgeries, new treatments, or new drugs), is the need for risk literacy is also in financial planning. compared with the average consumer, professional financial planners are educated in risk and deal with probabilistic outcomes on a regular basis (e.g., chance of disability, monte carlo simulation results of adequacy of retirement funds, etc.). this suggests professional financial planners may possess a higher level of risk literacy. understanding the risk literacy of financial planners is a critical step in the development of the financial planning profession because financial planners help clients make better financial decisions. financial planners that go through the certified financial planner (cfp) certification process are required to learn about risk. for example financial knowledge topics, including d22, d23, and e34, specifically address risk in the areas of insurance and investments. in addition, cfp professionals learn the seven-step financial planning process, which inherently includes the discussion of risk with clients, reflecting the steps found with the skilled decision theory framework. another aspect of risk literacy and enhancing understanding of risks stems from visual representations of risk. previous research in financial literacy indicates visual representation of financial concepts allows for better processing of financial information (kothakota & kiss, 2020). specifically, groups that have historically underperformed on financial literacy showed a much larger increase in financial literacy when the literacy concept is explained visually. properly formatted visualizations may assist financial planners and their clients in understanding portfolio risk. most visualization practice in the financial services industry have been applied to more complex concepts, such as bond duration and convexity or accounting analytics (rodriguez & kaczmarek, 2016). moreover, a growing body of literature suggests that while there is an expectation professionals who deal with risk on a regular basis have a high degree of risk literacy, this is not always the case. other professions, such as surgeons and senior-level police officers (garcia-retamero, cokely, wicki, & joeris, 2016; garcia-retamero & dhami, 2013), do not always exhibit high risk literacy. subsequently, an objective of this study is to extend the risk literacy research to a new profession by examining the risk literacy of professional financial planners (hereafter referred to as financial planners for convenience). the research question is: does risk literacy and visual representation of a risk-related scenario help financial planners to select an appropriate portfolio? 2. literature review risk literacy can be defined as “the ability to accurately interpret and act on information about risk” (cokely et al., 2012, p. 26) and has a well-researched history in judgment and decision-making (lurtz & heckman, 2018). risk literacy is separate and distinct from similar constructs such as: subjective numeracy, risk perception, and financial literacy. subjective numeracy is how a person feels about numerical information and their perception of their ability to use or understand numerical information (gamliel, kreiner, & garcia-retamero, 210 m. r. lurtz et al. / financial services review 29 (2021) 209–225 2016). risk perception is also subjective and relates to one’s ability to accurately understand the risks associated with a behavior or an event (roszkowski & davey, 2010). financial literacy has both an objective and subjective component as it is a combination of knowledge (objective) and perceived ability to apply one’s knowledge (subjective; huston, 2010). risk literacy is a solely objective measure of statistical numeracy. objective numeracy can be related to numeric skills, including basic arithmetic and statistics (garcia-retamero & galesic, 2010; gamliel et al., 2016). risk literacy in this investigation is the statistical, objective numeracy and will be referred to as risk literacy throughout the remainder of the study. as an example, testing individual risk literacy may involve calculating the probability of an event occurring, such as how often a die loaded to land on “6” should happen on a given number of rolls. ongoing work by researchers involved in the development of risk literacy measures have proposed a framework known as skilled decision theory (see cokely, feltz, ghazal, allan, petrova, & garcia-retamero, in press). the theory highlights visual aids and risk literacy as the two constructs that, “support skilled decision making both directly and indirectly through metacognitive effects” (p. 34). this theory details what individuals need (e.g., visual aids and risk literacy) to make skilled decisions and how those constructs impact the way in which individuals deliberate, build confidence, comprehend, and feel (affect) when making decisions (cokely et al., in press). the literature review focuses on risk literacy, visual aids, and the studies that have applied these important constructs to financial planning. 2.1. risk literacy risk literacy has been used to investigate the ways professionals understand probabilities and how they then, in turn, help others to understand or work with that information (e.g., surgeons and how they engage with clients when making surgical decisions; garcia-retamero, cokely, wicki, & hanson, 2014). a few studies have been published looking at risk literacy and financial planning topics. these studies have found that higher risk literacy is linked to better insurance decision-making, higher net worth, a desire for shared financial decisionmaking, and lower risk tolerance (campara, paraboni, da costa, saurin, & lopes, 2017; garcia-retamero & galesic, 2013; petrova, van der pligt, & garcia-retamero, 2014; smith, mcardle, & willis, 2010). 2.2. visualization and visual aids visualization is a wide and growing field encompassing studies that investigate the impact of data visualization and constructs such as spatial ability and graph literacy. spatial ability refers to one’s capability to form mental representations and/or manipulate these representations of objects in one’s mind (hegarty & kozhevnikov, 1999). studies of visual-spatial skill and numerical skill find a positive correlation between visual-spatial ability and numerical ability (hegarty & kozhevnikov, 1999; tosto et al., 2014). graph literacy is the ability to understand information that has been presented graphically and make decisions or draw conclusions based on that information (okan, galesic, garcia-retamero, 2016; shah & m. r. lurtz et al. / financial services review 29 (2021) 209–225 211 freedman, 2011) and has been linked to higher levels of education (galesic & garciaretamero, 2011). data visualization is the field of study provides insight into just how information can be presented to enhance understanding (knaflic, 2015). research on the usefulness of visual aids in financial planning, such as mind-mapping and the happiness risk/reward pyramid, have assisted financial planners and their clients to better communicate about decisions across all seven steps of the personal financial planning process (rouillier, 2011; van zutphen, 2010). a powerful visual aid used by financial advisors, coaches and therapists to help clients connect with their future and plan over a life span takes a tape-measure that the client cuts and manipulates to represent the life that they have yet to live (klontz, kahler, & klontz, 2016). narrowing the focus to just portfolio risk, a study using finvis, built to help financial planning clients visualize portfolio decisions— found that individuals using the software (1) improved decision-making, (2) increased learning and reduced ambiguity, and (3) increased confidence in understating of the financial decisions they were making (rudolph, savikhin, & ebert, 2009). these results are similar to the results from garcia-retamero, cokely, wicki, & joeris, (2016) who found that lownumerate surgeons when provided with an icon describing the risks associated with a surgery were not only more accurate choosing the correct assessment of risk, but also spent more time making decisions. this literature points to how visual aids can increase confidence, understanding, and impact resulting decision quality. 3. theoretical framework skilled decision theory details the decision-making process through which “novices” or non-experts and experts travel through to arrive at a well-informed or skilled decision (cokely et al., in press). the theory was developed based on numerous previous studies of how average individuals as well as experts process and arrive at a decision, and what can be done to influence arrival at a “skilled decision,” across a wide range of contexts (e.g., surgery, insurance, or precautionary health) (cokely et al., in press). as such, the theory organizes the decision-making process linearly. the decision-maker begins the decision process with a certain amount of risk literacy and/or the use of visual aids (fig. 1). visual aids may range in type or style but are tools that help individuals understand probabilities, percentages, and proportions. other constructs include deliberation, confidence, comprehension, and affect, each having a relationship with the use of visual aids and risk literacy. the deliberation construct is thinking about the problem at hand. both indiosyncratic risk literacy and visual aids/tools may help or hinder individual understanding of the problem. confidence follows deliberation. the confidence construct is related to one’s confidence in one’s knowledge and one’s confidence in their ability to carry out any subsequent behavior related to the decision. confidence is influenced by visual aids and risk literacy. comprehension is the next construct and it is also influenced by visual aids and risk literacy. visual aids and one’s level of risk literacy impact comprehension; high-risk literacy and use of a visual aid would make comprehending a risky decision easier as opposed to low risk literacy and no visual aid. affect, which pertains to how 212 m. r. lurtz et al. / financial services review 29 (2021) 209–225 “good” or “not good” a person feels about their decision-making ability is only influenced by risk literacy. the final construct, decision quality, is related to comprehension and affect. essentially, the best decisions are the ones we understand and that we feel good about. moreover, skilled decision theory is an appropriate theory for investigating risk literacy and decision-making in financial planners. the theory organizes how both expert and nonexperts make decisions. financial planners can be considered “expert” decision-makers. it is their job to help “non-expert” decision-makers (i.e., clients) to arrive at quality financial decisions. expert decision-makers are assumed to use this decision-making process even if they do not work with another individual. the study posits that within the financial planning process, financial planners and clients are regularly going through the decision-making process outlined by skilled decision theory. the financial planning process (appendix 1) promotes ongoing dialogue—deliberation between the client and the practitioner. scenario planning, like going over a market crash or likelihood of taking an early retirement, includes steps two through four of the financial planning process, which takes clients to a point where they are willing to implement (step 5). furthermore, this could be interpreted as evidence that the client and the planner have, at the same time, inadvertently moved through skilled decision theory, where they become confident, they understand (i.e., comprehension), and they feel good (i.e., affect) about moving forward to implementation. the current study did not test for the constructs that are ultimately related to decision quality (i.e., comprehension and affect), rather it focused solely on visual aids and risk literacy. previous investigative work on risk literacy in other professions also focused solely on visual aids and risk literacy, and used skilled decision theory as a theoretical framework (garcia-retamero et al., 2016; garcia-retamero & dhami, 2013). moreover, even without a direct theoretical connection from visual aids and risk literacy to decision quality, it is assumed that investigating the impact these two constructs alone still provide insight into decision-making ability and needs of financial planners as a first step. furthermore, cfp board registered financial planning education programs place an emphasis on risk literacy and often use visual aids in teaching materials. this additional, formalized education may assist cfp professionals to deliberate, build confidence, gain comprehension and handle affect by way of the seven-step financial planning process. fig. 1. skilled decision-making theory (cokely, feltz, ghazal, allan, petrova, & garcia-retamero, in press). m. r. lurtz et al. / financial services review 29 (2021) 209–225 213 utilizing skilled decision theory as a framework and previous literature as motivation for this study, five hypotheses were developed: hypothesis 1: financial planners who are certified financial planner (cfp) professionals will have higher risk literacy than financial planners lacking the cfp designation. hypothesis 2: use of a visual aid will be positively associated with having selected the correct risk portfolio. hypothesis 3: risk literacy will be positively associated with having selected the correct risk portfolio without use of a visual aid. 4. method 4.1. sample to gather data for this study, a 79-item survey was emailed to 106 u.s.-based financial planners via three sources: (1) an advisor-only forum (advisorheads.com), (2) a list-serve created and maintained by a popular financial planning practitioner-blogger, and (3) personal emails to financial planners. personal emails were limited in number, six emails in total. advisors from all three sources received the same email explaining the project and inviting them to complete the questionnaire. participants were not incentivized to participate, but participation was made simpler by only requiring them to click on the link provided to them in the invitation email. the computer-based questionnaire was administered in english and included basic demographic characteristics, firm characteristics, professional qualifications, financial literacy, and risk literacy. response rate from the forum, listserv, and personal emails combined totaled 65%. of those completed, less than 5% had missing items and those that had missing items were deleted (fowler, 1995). all told, 69 completed surveys were part of the research sample. 4.2. experimental design and dependent variable the experimental design of the current study is based on the work by garcia-retamero et al. (2016) who studied surgeons. the current design is similar in the following ways: (1) both tested risk literacy scores; (2) an almost identical icon array was used as a visual representation of the risk; (3) both tested the accuracy in answering a probability question; and (4) both were asked a question related to their domain of expertise. however, the studies were dissimilar in the type of risk presented. negative outcomes in surgery are death or other complications, whereas portfolio risk is not directly related to death. also, the question posed to the surgeons had known probabilities, whereas the type of market events posed to the financial planners in the current study are less precise (taleb, 2004). financial planners were randomly assigned to one of two survey instruments. one group received only written probabilistic information and the other received written probabilistic 214 m. r. lurtz et al. / financial services review 29 (2021) 209–225 information plus a visual aid. using the given information, participants were asked a specific question about the likelihood of failure for a portfolio given conditions similar to the great recession, a time period resulting in prolonged capital market decline. the text-only format asked the participant to calculate which of the two portfolios would be more resilient given the proposed market conditions: you have a client who is fearful of another great recession affecting their portfolio. based upon the fact-finding you have done you have narrowed the possible portfolio strategies to two. the first portfolio strategy is an asset allocation that is based upon an investment management strategy you have been using for years, while the second is based upon a newer investment strategy. you stress test the portfolios using 100 simulations. portfolios using the first strategy failed the client’s goals 27 times. compared with the first portfolio, the new strategy resulted in seven fewer failures. which portfolio strategy do you use? in the second condition, the respondent was given the same question, but the success and failure of the portfolios was also represented by a visual aid comprised of an icon array (fig. 2). as such, the condition being applied in this study was the presence, or lack thereof, of the icon array. the outcome of interest or dependent variable was the correct choice of portfolio given the situation. this was identified as the portfolio failing the fewest number of times. for the purposes of this study, if the respondent were in the non-visual aid condition, they were coded as “0.” if they were in the visual aid condition, they were coded as 1. if the respondent chose the correct portfolio, they were coded as 1 for correct answer, and 0 if they chose incorrectly. 4.3. independent variable of interests risk literacy was measured using the berlin numeracy test (bnt; cokely et al., 2012). the bnt is a psychometrically valid survey, which measures risk literacy and has been used fig. 2. portfolio selection icon. m. r. lurtz et al. / financial services review 29 (2021) 209–225 215 on various populations (www.riskliteracy.org). this test has been used in over 15 countries and has been shown to be both valid and reliable (cokely et al., in press; schwartz, woloshin, black, & welch, 1997). previous work differentiated risk literacy as a unique predictor of ability to understand and work with probabilities even after controlling for intellectual ability and numerical literacy (låg, bauger, lindberg, & friborg, 2014). the bnt has since become the strongest predictor of an individual’s ability to assess and understand everyday risk (cokely et al., 2012). all seven questions were asked in this survey, as used in the more comprehensive risk literacy tests. the analysis only used the four asked on the pen and paper bnt. this more closely aligns with other studies of surgeons (garcia-retamero, 2016) and the general population (cokely et al., 2012). this measure is scored as a 0–3 variable based upon the number of correct responses. categories were collapsed with scores of 0 and 1 = low numeracy, 2 and 3 = moderate numeracy, and 4 = high numeracy. the cfp certification was a self-reported measure. participants indicated whether or not they held the certification. thus, a binary variable of cfp certification was used. if an individual held the cfp certification, they were coded as a 1, and if not, they were coded as 0. no other demographic variables were used in the final regression. 4.4. demographic variables table 1 outlines demographic variable descriptive statistics according to treatment. demographic characteristics included gender, cfp certification, title, compensation method, firm type, education, personal income, and specialty. compensation structures included: (1) assets under management (aum), (2) aum fees and flat fees, (3) combination of salary, profit share, and commission, (4) commission and aum fees, (5) hourly and flat fees, (6) hourly, (7) flat, and (8) aum fees. firm types included: (1) commercial bank advisor, (2) independent b/d affiliations, (3) independent registered investment advisor (ria) of varying sizes, (4) brokerage firms, and (5) wire-houses. professional qualifications included education and professional specialties, such as financial planning and investment management, financial planning only, or life planning. 5. analyses descriptive analysis was conducted using r, in conjunction with the ide exploratory.io. univariate and bivariate tests were conducted using r in conjunction with rstudio. regression was conducted using r and rstudio, including the “tidy” packages (wickham, 2018). first, to determine if the group presented with the visual aid was similar to the group presented without a visual, a t-test across risk literacy levels by condition was conducted. to investigate whether or not a visual representation of portfolio risk increased accuracy in assessing and selecting the most resilient portfolio, parametric bootstrap logistic regression was conducted. both parametric and non-parametric bootstrap analyses were run and produced similar results. 216 m. r. lurtz et al. / financial services review 29 (2021) 209–225 table 1 descriptive statistics of sample descriptives visual (n = 69) variable visual percent no visual percent gender male 97.50% 91.30% female 2.50% 4.35% rather not say 0.00% 4.35% certified financial planner (cfp)? yes 50.00% 53.62% no 50.00% 46.38% title junior financial advisor 3.50% 13.04% assistant financial advisor 3.50% 4.35% broker/financial advisor 32.14% 26.09% senior advisor/firm owner 46.43% 47.83% senior financial advisor 14.29% 8.70% compensation method aum fees and flat fees 10.71% 17.39% aum fees only 10.71% 4.35% combination of salary, profit share, and commission 10.71% 17.39% commission and aum fees 46.43% 30.43% hourly and flat fees 7.14% 0.00% hourly, flat fees, and aum fees 14.29% 26.09% advisor channel commercial bank advisor 3.57% 5.88% independent b/d affiliation, large (>15) 3.57% 3.92% independent b/d affiliation, small (<10) 17.86% 21.57% independent ria, large 10.71% 7.84% independent ria, medium 10.71% 7.84% independent ria, small 28.57% 33.33% large regional brokerage firm 21.43% 11.76% large wire-house 3.57% 7.84% higher education masters 28.57% 30.43% phd, masters 7.14% 4.35% phd 3.57% 4.35% no higher education/chose not to respond 60.71% 60.87% income $0–$20,000 7.14% 4.35% $20,001–$50,000 3.57% 13.04% $50,001–$100,000 17.86% 17.39% $100,001–$200,000 21.43% 21.74% $200,001–$300,000 17.86% 30.43% $300,001–$400,000 7.14% 4.35% $400,001–$500,000 7.14% 0.00% $500,001–$600,000 3.57% 0.00% $600,001–$700,000 3.57% 3.92% $900,001–$1,000,000 0.00% 4.35% $1,000,001–$1,500,000 3.57% 0.00% (continued on next page) m. r. lurtz et al. / financial services review 29 (2021) 209–225 217 therefore, parametric bootstrap was used in this study. parametric bootstrap provides narrower confidence intervals and more power than non-parametric bootstrap (adjei & karim, 2016). bootstrap logistic regressions have been used in social science and the medical field to estimate a population by resampling the observations (fitrianto & cing, 2014).the independent variables included whether or not the participant was a cfp certificant and the financial planners’ bnt score. the dependent variable was whether or not the financial planner selected the most resilient portfolio. given the small sample size, a bootstrap logistic regression was also conducted. to obtain robustness with respect to logistic regression, the observations were resampled at random specific intervals (fitrianto & cing, 2014). in this case, at each iteration 10% of observations were resampled, for a total of four iterations. the more iterations, the higher the standard error and model specification is more difficult as the number of iterations increases. the suggested number of iterations is calculated by taking the number of observations in the sample and dividing by the minimum variable requirement for the type of regression used, which in this case is 18. five iterations are the maximum number of resamples recommended (fitrianto & cing, 2014). 6. results in terms of risk literacy, most participants had moderately high risk literacy. the mean risk literacy score for financial planners was 2.20 (sd = 0.99). the cronbach’s a for the bnt was .79. the following groups were compared with see if they contained similar profiles: (1) cfp status and (2) risk literacy. crosstab information (table 2) indicated 44.44% of non-cfp holders had “low” risk literacy compared with 21.43% of cfp holders. onethird (33%) of non-cfp holders had “medium” risk literacy, compared with 50% of cfp certificants. those respondents in the “high” group were 22.23% for non-cfp holders and 28.57% for cfp holders. a x2 test of cfp certificant status and choosing the correct portfolio was conducted. results indicated that choosing the correct portfolio was not significantly different and independent of whether the participant was a cfp certificant or not. results in table 3 indicate a p-value of 0.16. of participants who received the portfolio information with no visual, 64% of participants chose the correct portfolio (table 4). for participants who received the portfolio information and a visual representation, 87% chose the correct portfolio. a x2 test was conducted table 1 (continued) descriptives visual (n = 69) variable visual percent no visual percent specialty financial planning and investment management 96.43% 95.65% financial planning only 3.57% 0.00% life planning 0.00% 4.35% note: aum ¼ assets under management; b/d ¼ broker-dealer; ria ¼ registered investment advisor. 218 m. r. lurtz et al. / financial services review 29 (2021) 209–225 indicating a significant difference (p= .016) and having the visual aid increased accuracy. a robustness check using a t test for the percentages was also conducted, confirming the results of using the count data from the x 2-test. 6.1. bootstrap logistic regression parametric bootstrap logistic regression results are presented in table 5. results indicated that participants in the group that saw the visual representation had 2.45 times (p = .03) greater odds of choosing the correct portfolio. holding a cfp certificate (p = .25) nor possessing numeracy (p = .29) were significant in selecting the correct portfolio. moreover, these results do not support hypothesis 1 and 3, but does provide support for hypothesis 2. visually presented information had a significant and positive impact on correct portfolio selection. univariate models were run with each independent variable in the full model. results are similar to the multivariate model and are contained within tables 6, 7, and 8. whether or not a participant was a cfp certificant was not significant as it relates to selecting the correct portfolio (p= .26). numeracy was also not significant in the univariate model (p= .35) as it relates to selecting the correct portfolio. whether the participant received a visual aid was significant (p= .02) and had 2.65 greater odds of choosing the correct portfolio. 7. discussion the research question was: does risk literacy and visual representation of a risk-related scenario help financial planners to select an appropriate portfolio strategy? in short, risk literacy did not impact appropriate portfolio strategy, but visual representation did. to table 2 cross-tab of risk literacy by certified financial planner (cfp) status risk literacy level non-cfp cfp low 44.44% 21.43% moderate 33.32% 50.00% high 22.23% 28.57% total 100.00% 100.00% table 3 v2 test of participants’ risk literacy on whether they were a certified financial planner (cpf) certificant v2 degree of freedom p-value 6.65 4 0.16 significance levels *p < .10, **p < .05, ***p < .01. m. r. lurtz et al. / financial services review 29 (2021) 209–225 219 investigate this question, the study measured risk literacy using the bnt and an experimental design, which consisted of a randomly assigned visual aid component. high risk literacy may be related to the unique nature of financial planners’ work, as it is inherently involved with discussing, understanding, and measuring risk. the study suggests those planners who further their education and obtain the cfp certification have higher risk literacy scores. the type of work, education, and use of the seven-step process may be related to higher levels of risk literacy among those who hold a cfp certification. on the other hand, it is also important to recognize that this finding may also be a selection effect, and those with higher risk literacy scores opt-in to obtaining cfp certification. selecting the correct portfolio was not linked to risk literacy, education, or professional certification as demonstrated by the logistic regression. this may be due to a small sample size, the convenience nature of the sample, or a commonality of industry training (e.g., series 7, 63 and/or 65 exam). unlike garcia-retamero et al. (2016), this study found that risk literacy was not linked to the likelihood of choosing the appropriate portfolio strategy, which could be an artifact of testing risk literacy in financial planners. said another way, a reason for investigating risk literacy in financial planners was to examine how risk literacy in financial planners may be different from that of other professionals (e.g., surgeons or high-level police officers). financial planners, by way of their training on portfolio selection, may still be able to select the right portfolio regardless of risk literacy. another possible reason for this finding is that risk literacy and the knowledge needed to select the correct portfolio are not one in the same as originally thought by the researchers. lastly, it is important to remember that this group, as a whole, was very risk literate. therefore, the sample may not have enough variation to detect the importance of risk literacy. although these findings differ from previous work in this area, it can be argued that the findings in this study still support the new theoretical framework of skilled decision theory (cokely et al., in press). financial planners may ultimately choose to become financial planners not only due to their natural ability to understand risk, but also due to their education and training. either way, financial planners’ higher levels of risk literacy cannot be ignored table 4 v2 test of participants’ risk portfolio on condition v2 degree of freedom p-value effect size 5.84 1 0.016** 0.19 significance levels *p < .10, **p < .05, ***p < .01. table 5 results for bootstrap logistic regression on portfolio selection (n = 69) variable coefficient se p or lower ci upper ci intercept 1.01 0.54 0.62 — — — certified financial planner (cfp) �0.38 0.42 0.25 0.61 0.26 1.42 risk literacy 0.02 0.20 0.29 1.17 0.87 1.58 visual aid 0.94 0.39 0.029** 2.45 1.28 5.04 source: four resampling intervals. cfp = certified financial planner; or = odds ratio; ci = confidence interval. significance levels *p < .10, **p < .05, ***p < .01. 220 m. r. lurtz et al. / financial services review 29 (2021) 209–225 and the use of visual aids, which was the only significant predictor of the correct portfolio selection, does demonstrate the importance of visual aids for risk decisions—perhaps even in high risk literacy populations. 7.1. implications this study demonstrates visual icons help individuals at all levels of risk literacy to improve their decision-making. financial planners, even cfp certificants, may want to test themselves and then take pro-active steps to become better at interpreting and explaining risk information. financial planning programs registered with the cfp board may wish to start adding a component of visual aid literacy to their curriculums. larger regulatory financial institutions, like the security and exchange commission (sec) as well as financial industry regulatory authority (finra), may want to request that financial planners, in addition to measuring client’s risk tolerance, also display risk-reward tradeoffs in a visual format. financial planners, especially those acting as fiduciaries, can consider using visual techniques in their workflow process. this information will not only help the financial planner to assess what they should be discussing when they discuss risks with clients, but also how they explain recommendations and actions taken as it relates to risk. investing in financial software that utilizes visual best practices may also be advantageous. displaying information graphically during reviews and illustrating portfolio stress tests via charts may be useful in helping clients comprehend what advisors are attempting to communicate. 7.2. limitations there are limitations inherent in this study. first, the sample is small and has less power than it would have if the sample were larger. this has implications for the results and a table 6 results for bootstrap logistic regression of certified financial planner (cfp) status on portfolio selection (n = 69) variable coefficient se p or lower ci upper ci intercept 1.09 0.25 <0.01*** — — — cfp 0.46 0.41 0.26 1.58 �0.34 1.25 source: four resampling intervals. cfp = certified financial planner; or = odds ratio; ci = confidence interval. significance levels *p < .10, **p < .05, ***p < .01. table 7 results for bootstrap logistic regression of risk literacy on portfolio selection (n = 69) variable coefficient se p or lower ci upper ci intercept 0.65 0.7 0.36 — — — risk literacy 0.13 0.14 0.35 1.14 �0.15 0.41 source: four resampling intervals. cfp = certified financial planner; or = odds ratio; ci = confidence interval. significance levels *p < .10, **p < .05, ***p < .01. m. r. lurtz et al. / financial services review 29 (2021) 209–225 221 more robust sample may have different conclusions. bootstrap logistic regression can increase standard errors, but this was limited by limiting the number of iterations. however the sample mirrors the population of financial planning advisors in two key ways. participants in the study were mostly male (94.12%) with high incomes, which are consistent traits of financial planners based on recent industry research (tharp, lurtz, melitz, ammerman, & kitces, in press). this could have had an impact on the insignificance of risk literacy, resulting in different findings than past research. however, this may also point to the fact that financial planners possibly have received more education geared toward understanding and interacting with probabilities (i.e., risk literacy) than the general population. overall, 30.43% of the sample scored low on the risk literacy assessment. financial planners’ style of work, deliberating, and working with the client to understand risk, may also drive up risk literacy scores. financial planners may, in general, have higher risk literacy than some other professional groups due to the way their education and the practice of financial planning, is conducted. however, as this was administered online, with no time limit, the participants could have answered using online calculators or internet searches. in addition, this was a convenience sample drawn from willing participants through listservs and email lists known by the researchers. while most financial planners will choose portfolios based upon a variety of factors, the portfolio selection task in this study was simple, constrained, and narrowly defined. 7.3. future research directions to our knowledge, this study is the first of its kind in financial planning research. findings suggest that further exploration with a larger, more diverse sample may result in different findings that could reflect previous work conducted with other populations. testing other types of visual aids utilizing the same probabilistic information could lead to a better understanding of best practices for visual aid use. it would also be helpful to uncover how, if at all, risk literacy does increase or change as financial planners grow in their careers and position. the same could be said for clients of financial planners: does working with a financial planner increase risk literacy skill? finally, future research would also be enhanced by measuring the other constructs of skilled decision theory. 7.4. conclusion literature from judgment and decision-making highlights the importance of risk literacy and visual aids in making decisions. this study (1) compares risk literacy levels of financial table 8 results for bootstrap logistic regression of visual aid on portfolio selection (n = 69) variable coefficient se p or lower ci upper ci intercept 0.83 0.26 <0.01*** — — — visual aid 0.97 0.41 0.02** 2.65 1.18 4.21 source: four resampling intervals. significance levels *p < .10, **p < .05, ***p < .01. 222 m. r. lurtz et al. / financial services review 29 (2021) 209–225 planners to previously published studies of risk literacy levels and (2) replicates the study by garcia-retamero et al. (2016) that investigated risk literacy and the effect of visual aids among surgeons. this study contributes to the literature by (1) investigating risk literacy levels among financial planners, and (2) examining the effect of visual aids on portfolio choices among financial planners. the findings provide support for skilled decision-making theory (cokely et al., in press) as well as previous work conducted by garcia-retamero et al. (2016) and kothakota & kiss (2020), which suggest the use of visual aids in financial planning is useful and potentially necessary. not all visual aids are created equally, however, and a poorly constructed visual, like a simple verbatim task description, may not help in explaining risks to clients. while this study only investigated the use of an icon array, the direct application to understanding portfolio failure risk is something financial planners address every day. in summary, financial planners help clients make appropriate financial decisions that involve risk and the need to interpret probability appropriately. while most financial planners in this study scored moderate or high in risk literacy, the combination of a visual aid and written probabilistic information versus only written probabilistic information significantly helped planners choose the appropriate portfolio strategy. appendix 1 the seven-step financial planning process 1. understanding the client’s personal and financial circumstances 2. identifying and selecting goals 3. analyzing the client’s current course of action and potential alternative courses of action 4. developing the financial planning recommendation(s) 5. presenting the financial planning recommendation(s) 6. implementing the financial planning recommendation(s) 7. monitoring progress and updating adapted from the cfp board of standards (www.cfp.net) berlin numeracy test traditional paper and pencil format instructions: please answer the questions below. do not use a calculator but feel free use the space available for notes (i.e., scratch paper). 1. imagine we are throwing a five-sided die 50 times. on average, out of these 50 throws how many times would this five-sided die show an odd number (1, 3, or 5)? 2. out of 1,000 people in a small town 500 are members of a choir. out of these 500 members in the choir 100 are men. out of the 500 inhabitants that are not in the choir 300 are men. what is the probability that a randomly drawn man is a member of the choir? _____ % (please indicate the probability in percentage) 3. imagine we are throwing a loaded die (6 sides). the probability that the die shows a 6 m. r. lurtz et al. / financial services review 29 (2021) 209–225 223 is twice as high as the probability of each of the other numbers. on average, out of these 70 throws, how many times would the die show the number 6? _____ 4. in a forest 20% of mushrooms are red, 50% brown, and 30% white. a red mushroom is poisonous with a probability of 20%. a mushroom that is not red is poisonous with probability of 5%. what is the probability that a poisonous mushroom in the forest is red? _____% scoring = count total number of correct answers. correct answers: 1 = 30; 2 = 25; 3 = 20; 4 = 50. references adjei, i. a., & karim, r. (2016). an application of bootstrapping in logistic regression model. open access library journal, 3(9), 1–9. campara, j. p., paraboni, a. l., da costa, n., saurin, v., & lopes, a. (2017). subjective risk tolerance and numeracy skills: a study in brazil. journal of behavioral and experimental finance, 14, 39–46. cokely, e. t., feltz, a., ghazal, s., allan, j. n., petrova, d., & garcia-retamero, r. (in press). decision making skill: from intelligence to numeracy and expertise. in k. a. ericsson, r. r. hoffman, a. kozbelt, & a. m. williams (eds.), cambridge handbook of expertise and expert performance (2nd ed., pp. 476–505). new york, ny: cambridge university press. available at http://www.riskliteracy.org/files/1414/8312/0711/2016_ cokely_et_al_decision_making_skill_camhandbook_final_pre-print.pdf cokely, e. t., galesic, m., schulz, e., ghazal, s., & garcia-retamero, r. (2012). measuring risk literacy: the berlin numeracy test. judgment and decision making, 7, 25. fitrianto, a., & cing, n. m. (2014). empirical distributions of parameter estimates in binary logistic regression using bootstrap. international journal of mathematical analysis, 8, 721–726. fowler jr, f. j., & fowler, f. j. (1995). improving survey questions: design and evaluation. sage publishing. thousand oaks, ca. galesic, m., & garcia-retamero, r. (2011). graph literacy: a cross-cultural comparison. medical decision making, 31(3), 444–457. gamliel, e., kreiner, h., & garcia-retamero, r. (2016). the moderating role of objective and subjective numeracy in attribute framing. international journal of psychology : journal international de psychologie, 51, 109–116. garcia-retamero, r., cokely, e. t., wicki, b., & joeris, a. (2016). improving risk literacy in surgeons. patient education and counseling, 99, 1156–1161. garcia-retamero, r., & dhami, m. k. (2013). on avoiding framing effects in experienced decision makers. quarterly journal of experimental psychology, 66, 829–842. garcia-retamero, r., & galesic, m. (2013). does young adults’ preferred role in decision making about health, money, and career depend on their advisors’ leadership skills? international journal of psychology, 48, 492– 501. garcia-retamero, r., & galesic, m. (2010). how to reduce the effect of framing on messages about health. journal of general internal medicine, 25, 1323–1329. garcia-retamero, r., wicki, b., cokely, e. t., & hanson, b. (2014). factors predicting surgeons’ preferred and actual roles in interactions with their patients. health psychology, 33(8), 920. hegarty, m., & kozhevnikov, m. (1999). types of visual-spatial representations and mathematical problem solving. journal of educational psychology, 91, 684–689. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. klontz, b., kahler, r., & klontz, t. (2016). facilitating financial health: tools for financial planners (2nd edition), coaches, and therapists. national underwriter company. erlanger, ky. 224 m. r. lurtz et al. / financial services review 29 (2021) 209–225 knaflic, c. (2015). storytelling with data: a data visualization guide for business professionals.boken: john wiley & sons. hoboken, nj. kothakota, m., & kiss, d. e. (2020). use of vizualization tools to improve financial knowledge: an experimental approach. journal of financial planning and counseling, 31(2), 193–208. låg, t., bauger, l., lindberg, m., & friborg, o. (2014). the role of numeracy and intelligence in health-risk estimation and medical data interpretation. journal of behavioral decision making, 27, 95–108. lurtz, m., & heckman, s. (2018). risk literacy. in c. r. chaffin (ed.). financial psychology (pp. 155–165). hoboken, nj: john wiley & sons, inc. okan, y., galesic, m., & garcia-retamero, r. (2016). how people with low and high graph literacy process health graphs: evidence from eye-tracking. journal of behavioral decision making, 29, 271–294. petrova, d. g., van der pligt, j., & garcia-retamero, r. (2014). feeling the numbers: on the interplay between risk, affect, and numeracy. journal of behavioral decision making, 27, 191–199. rodriguez, j., & kaczmarek, p. (2016). visualizing financial data. john wiley & sons. hoboken, nj. roszkowski, m. j., & davey, g. (2010). risk perception and risk tolerance changes attributable to the 2008 economic crisis: a subtle but critical difference. journal of financial service professionals, 64(4), 42–53. rouillier, p. (2011). mind mapping: visualize money in motion. journal of financial planning,may/june, 20–21. rudolph, s., savikhin, a., & ebert, d. s. (2009). finvis: applied visual analytics for personal financial planning. 2009 ieee symposium on visual analytics science and technology, atlantic city, nj. shah, p., & freedman, e. g. (2011). bar and line graph comprehension: an interaction of top-down and bottomup processes. topics in cognitive science, 3, 560–578. schwartz, l. m., woloshin, s., black, w. c., & welch, h. g. (1997). the role of numeracy in understanding the benefit of screening mammography. annals of internal medicine, 127(11), 966–972. taleb, n. n. (2004). bleed or blowup? why do we prefer asymmetric payoffs?. the journal of behavioral finance, 5(1), 2–7. tharp, d. t., lurtz, m., mielitz, k., kitces, m., & ammerman, d. a. (in press). technology use and time spent in the financial planning process. journal of financial counseling and planning. tosto, m. g., hanscombe, k. b., haworth, c. m., davis, o., petrill, s., dale, p., malykh, s., plomin, r., & kovas, y. (2014). why do spatial abilities predict mathematical performance? developmental science, 17, 462–470. van zutphen, n. (2010). a visual aid for successful financial planning: the happiness risk/reward pyramid. journal of financial planning, 23, 54–57. wickham, h. (2018). the tidyverse. r package ver, 1(1), 836. m. r. lurtz et al. / financial services review 29 (2021) 209–225 225 academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2020 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 29, no. 2, 2021 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2021 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. academy of financial services officers president janine sam shepherd university president-elect inga timmerman california state university, northridge executive vice president-program terrance k. martin utah valley university vice president-communications david nanigian california state university, fullerton vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations chris browning texas tech university immediate past president swarn chatterjee university of georgia editor, financial services review stuart michelson stetson university directors shawn brayman planplus global charles chaffin cfp board of standards victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university robert moreschi virginia military institute past presidents swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2020 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 28, no. 3, 2020 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2020 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. the association between financial risk and retirement satisfaction blain pearsona,*, michael guillemetteb adepartment of personal financial planning, kansas state university, 343 justin hall, manhattan, ks 66506-1403, usa bdepartment of personal financial planning, texas tech university, box 41210, lubbock, tx 79409-1210, usa abstract a higher level of risky financial assets that a retiree holds may produce higher returns, resulting in utility gains. to test this hypothesis, a variable is constructed measuring retirees’ ratio of risky assets to total assets (risk ratio). next, the association between the risk ratio and retiree utility is examined using a retirement satisfaction variable from the 1992-2014 waves of the health and retirement study. the findings suggest that increases in retirees’ risk ratio is associated positively with increases in their retirement satisfaction. the results and ensuing discussion offer a new perspective for retiree asset management. © 2020 academy of financial services. all rights reserved. jel classification: d8 keywords: financial planning; retirement satisfaction; risk aversion introduction financing consumption in advanced age is paramount when planning for the transition into retirement. in the absence of non-labor income sources, such as social security income, pension income, and annuity income, retirees utilize their saved assets to finance consumption. saved assets can take varying forms of financial risk, and the traditional “time-horizon” approach to asset management suggests shifting from risky assets to less risky assets when transitioning into retirement. however, risky assets, such as equities, have historically *corresponding author: tel.: +1-828-455-2617; fax: 1-785-532-5505. e-mail address: bmpearson@k-state.edu. 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 341–350 provided higher returns when compared with their less-risky alternatives, such as bonds, money market accounts, and cds. risky assets have the potential for higher returns and may provide retirees with more income to finance their consumption in retirement, and more income to finance consumption in retirement may lead to a more satisfactory retirement experience. retirement satisfaction is affected by many different factors, including a retiree’s financial situation (bonin et al., 2007; diener et al., 2010; seccombe & lee, 1986), marital status (easterlin, 2003; van solinge, 2008), health status (barfield & morgan, 1978; price & balaswamy, 2009), and pre-retirement feelings about retirement (elder, 1999; kimmel, 1978). planning for retirement, reading about retirement, and exposure to radio or television programs about retirement also are significant correlates of retirement satisfaction (dorfman, 1989; taylor-carter, 1997). many studies have analyzed how risk preference affects utility (bachmann et al., 2017; hanna & chen, 1997; pratt, 1964; pålsson, 1996). however, most of these studies assume homogeneity in their sample and do not consider the structural-grouping differences among the population, such as how their results would apply to a retired sample. analyzing risk preferences and retiree utility minimizes human capital and employment considerations that researchers argue should be considered when measuring risk preferences. for example, individuals with higher levels of human capital are more likely to have higher risk tolerance (shaw, 1996), suggesting that human capital investment is an inverse function of risk aversion. thus, the potential for human capital development may affect risk preference, as fully retired individuals will not invest in their human capital for purposes of future labor income. there are a variety of ways to measure risk preference. one approach is questionderived assessments. however, risk-preference questionnaires may not be reflective of actual investment behavior (bouchey, 2004; corter & chen, 2006; yook & everett, 2003). objective risk tolerance, or an individual’s ratio of risky financial assets relative to either their net worth or assets, is another measure used in a variety of studies (cordell, 2002; hanna & chen, 1997; sung & hanna, 1996). relative risk is another measure of risk preference. relative risk provides a coefficient of individuals’ level of risk relative to their total wealth and may be a better measure when conducting a comparative analysis (dyer & sarin, 1982). the goal of this study is to measure the effect of the risk ratio on retirement satisfaction levels. this research question will help shed light on the association between risky asset holdings in retirement, relative to total wealth, and retirement satisfaction. we posit that assets with higher risk also come with the potential for higher returns, and higher returns provide retirees with more income to finance their consumption in retirement. therefore, we hypothesize that a higher level of risky assets, relative to total wealth, may lead to higher retirement satisfaction. data longitudinal data that are collected from the health and retirement study (hrs) are used for hypothesis testing.1 the hrs is a household survey conducted by the institute for 342 b. pearson, m. guillemette / financial services review 28 (2020) 341–350 social research at the university of michigan. the rand hrs 2014 fat file (v2a) is used, which includes the 1992-2014 waves.2 the sample only includes individuals who are fully retired. to focus solely on retirees, the subset of hrs respondents who answer “retired” when asked, “are you working now, temporarily laid off, unemployed and looking for work, disabled and unable to work, retired, a homemaker, or what?” respondents who state anything other than “retired,” as well as incomplete responses, are dropped from the analysis. respondents who state that they are retired, yet still reported earned income, are also dropped from the sample. this is done to ensure that the sample is fully retired and that the only income that the sample receives is non-labor income. the presence of labor income would add complexities to the analysis that are difficult to control for given the data limitations within the hrs. for example, the riskiness of labor income varies based on factors such as occupation and tenure, and there are either limited or no data available to control for these differences among respondents within the hrs. utility is measured as a retiree’s level of retirement satisfaction. retirement satisfaction is measured using the following question: “all in all, would you say that your retirement has turned out to be very satisfying, moderately satisfying, or not at all satisfying?” using a likert method, the observations are coded as 1 (not at all satisfied 3,520), 2 (moderately satisfied 20,234), and 3 (very satisfied 35,640). the average satisfaction score is 2.54. this implies that “very satisfied” individuals were more likely to be found in the data and therefore selection bias might influence the results. the sample size is 17,672 and there are 59,404 observations. the risk ratio variable the risk ratio (rr) variable is constructed by dividing retirees’ stock assets by their total wealth: rrit ¼ tseai on i¼0 tseaið þ þ tbeaið þ þ tcaið þ þ e heið þ� � where tseai ¼ total stock equity assets. tbeai ¼ total bond assets. tcai ¼ total market and non-market cash assets. e heið þ ¼ summation of all residences – all mortgage liabilities. total stock equity assets (tsea) are considered risky assets, which include the net value of stocks, mutual funds, and investment trusts. total bond assets (tbea) include the net value of bonds and bond funds. the market and non-market cash assets (tca) include the net value of cds, government savings bonds, t-bills, checking accounts, savings accounts and money market accounts. b. pearson, m. guillemette / financial services review 28 (2020) 341–350 343 home equity is included as a riskless asset in the denominator. the results may be sensitive to the treatment of whether or not home equity is included as a riskless or risky asset (hanna et al., 2001; pålsson, 1996). home equity is treated as a riskless asset for the reasons noted by bellante and green (2004). they suggest that home equity is riskless because the older segments of the u.s. population are more likely to own their homes and carry very little debt. therefore, the low leverage levels provide a barrier to the effects of home-equity value fluctuations. it should be noted that housing debt has been increasing slightly among older americans (lusardi et al., 2018) since bellante and green’s (2004) study. additionally, home equity provides a hedge against systemic inflationary risks. human capital is a substantial part of an individual’s wealth (schultz, 1961). as noted in hanna and chen’s (1997) study, human capital should be analyzed as a part of the total wealth portfolio when developing objective risk measures, such as the rr variable. however, human capital estimates are difficult to assess, with many of the methods under heavy academic scrutiny (chenet al., 2004; fitz-enz, 2000; mulligan & sala-i-martin, 2000). because retirees are analyzed, it is assumed that they will not use their human capital for income nor invest in their human capital. human capital, therefore, is assumed to have a value of zero for the retirees in the sample. retirees may consider their non-labor income sources when deciding on the level of risk of their saved assets. thus, arguments could be made for the inclusion of the net present value (npv) of pension, annuity, social security, and other non-labor incomes in the rr total-wealth denominator. the rr variable does not include the npv of non-labor incomes because non-labor income is controlled for in the quantitative analysis. other variables dummy variables are created and coded with a value of “1” if the respondent is white, married, or male. a “0” is coded otherwise. continuous variables are created to measure the retirees’ age, non-labor income (income), non-housing wealth (wealth), and years of education. nominal values are used for income and wealth. a categorical variable measuring health status also is created. the health status variable can take the following values: 1 (poor), 2 (fair), 3 (good), 4 (very good), and 5 (excellent). table 1 provides the descriptive statistics of the sample. the average rr is 0.09. the average age of the retirees in the sample is 73. the average income and wealth of the retirees are $49,720 and $187,068, respectively. respondents in our sample are wealthier than average and, therefore, have more resources for consumption, which may be a possible explanation for why the mean retirement satisfaction score is 2.54 out of 3. therefore, selection bias may be present. the sample includes 87.81% white retirees, 45.74% male retirees, and 61.11% of the retirees are married. table 2 provides a further breakdown of the descriptive statistics by retirement satisfaction levels. there is a positive relation between the rr variable and retirement satisfaction. retirees being “not at all” satisfied, “moderately” satisfied, and “very” satisfied with their retirements have average risk ratios of 0.05, 0.07, and 0.11, respectively. additionally, there 344 b. pearson, m. guillemette / financial services review 28 (2020) 341–350 is a positive relation between being “very satisfied” and higher levels of health, income, and wealth. method to test the hypothesis, a random-effects ordered probit model is estimated on the unbalanced panel: sat� it ¼ b 0i þ b 1rrit þ b jdvit þ ai þ eit satit ¼ 1 if sat� it < m1 not at all satisfiedð þ satit ¼ 2 if m1 ≤ sat� it < m2 moderately satisfiedð þ satit ¼ 3 if sat� it ≥ m2 very satisfiedð þ where sat� it is a latent measure of retiree i’s satisfaction in wave t. the unknown thresholds, m1 and m2, are to be estimated. table 2 summary of data by satisfaction measures “not at all satisfied” “moderately satisfied” “very satisfied” respondents 3,530 20,234 35,640 risk ratio 0.05 0.07 0.11 health 2.10 2.78 3.34 years of education 11.70 12.27 13.03 age 69.90 73.61 73.70 white 0.81 0.86 0.90 male 0.43 0.44 0.47 married 0.51 0.55 0.66 income $32,926 $42,133 $55,689 wealth $69,911 $132,131 $229,867 note: n = 59,404 observations from 17,672 retirees. table 1 descriptive statistics of sample mean standard error min max risk ratio 0.0905 0.1801 0 0.9999 satisfaction 2.5405 0.6060 1 3 age 73.4445 9.2105 50 106.6667 education 12.6939 2.8713 8 17 white 0.8781 0.3271 0 1 male 0.4574 0.4982 0 1 health 3.0764 1.0948 1 5 married 0.6111 0.4875 0 1 income $49,720 $85,822 $0 $7,307,860 wealth $187,068 $643,298 $0 $42,300,000 note: n = 59,404 observations from 17,672 retirees. b. pearson, m. guillemette / financial services review 28 (2020) 341–350 345 b 0 represents the y-intercept of the model. the intercept value for individual i is expressed as b 0i ¼ b 0 þ ei, where i ¼ 1,. . .,n and e(eiþ ¼ 0 and varð= s 2 eþ. below is the assumption concerning the composite error component: ei�n 0,s 2 e � � e eietð þ ¼ 0 for i 6¼ t an ordered probit model is used due to the non-linear nature of the dependent variable. to the authors’ knowledge, a reliable computation for a fixed-effects ordered probit model does not currently exist. thus, a random effects model is utilized. if a fixed-effects estimator becomes available, future research should reconsider this study’s findings to better understand how the with-in subject variation of the rr variable is associated with retirement satisfaction. rrit represents the risk ratio variable, rr, and b 1 is the coefficient associated with rrit. it is expected that rrit will result in a positive coefficient for the “very satisfied” category and a negative coefficient for the “moderately satisfied” and “not at all satisfied” categories. a positive coefficient for the very satisfied category would suggest that retirees are more likely to respond very satisfied if the risk of their saved assets is increased. a granger causality test is utilized to test for reverse causality between sat�it and rrit (granger, 1969). the results of the test suggest that there are no statistically significant reverse causality issues. the matrix dvit contains all of the demographic variables used as control variables in the model, including white, married, male, health, age, income, wealth, and years of education. b j is the vector of coefficients related to the matrix dvit. it is expected that higher levels of income, wealth, education, health, and being married will result in positive coefficients for the very satisfied category. ai is the unknown intercept for each retiree i. average marginal effects provide the magnitudes for each of the effects on observed retirement satisfaction. the error term is assumed to follow a standard normal distribution. results the average marginal effects from the random-effects ordered probit regression are reported in table 3. as the rrit variable increases from 0 to 1, the results suggest that the probability of retirees being very satisfied with their retirement increases by 0.0762. as the rr variable increases from 0 to 1, the results suggest that the probability of retirees being moderately satisfied and not at all satisfied with their retirement decreases by 0.0523 and 0.0239, respectively. one consideration to note is that a higher rr means that a retiree has a higher percentage of assets held in stocks. thus, stock returns may be endogenous in the estimated model. to adjust for the potential impact that stock assets may have on the results, annualized nominal 346 b. pearson, m. guillemette / financial services review 28 (2020) 341–350 s&p 500 returns are included in a new model. the hrs data are biannual and therefore the s&p 500 data are biannual for each wave of the survey (refer to table 4). the returns analyzed include both the returns generated from asset price changes and dividends. a sensitivity analysis is conducted by including the series of s&p 500 returns as a variable in the random-effects regression. the average marginal effects are reported in table 5. the results indicate that as the rrit variable increases from 0 to 1, the probability of retirees being very satisfied with their retirement increases by 0.0223. as the rrit variable increases from 0 to 1, the results suggest that the probability of retirees being moderately satisfied and not at all satisfied with their retirement decreases by 0.0135 and 0.0093, respectively. although these results are still statistically significant, the new model has reduced the magnitude of the rrit variable. it is important to note that satisfied individuals were more likely to be in the sample, and this possible selection bias might influence the results. table 3 average marginal effects of the risk ratio on retirement satisfaction “not at all” “moderately satisfied” “very satisfied” risk ratio �0.0239* (0.0038) �0.0523* (0.0084) 0.0762* (0.0122) health (poor as base outcome) fair �0.0633* (0.0036) �0.0624* (0.0030) 0.1257* (0.0063) good �0.0959* (0.0036) �0.1263* (0.0036) 0.2222* (0.0065) very good �0.1181* (0.0037) �0.1963* (0.0045) 0.3145* (0.0071) excellent �0.1275* (0.0038) �0.2385* (0.0065) 0.3661* (0.0089) male (female as base outcome) 0.0002 (0.0018) 0.0005 (0.0038) �0.0007 (0.0056) white (non-white as base outcome) �0.0121* (0.0024) �0.0529* (0.0052) 0.0387* (0.0076) married (non-married as base outcome) �0.0242* (0.0016) �0.0529* (0.0033) 0.0772* (0.0048) education �0.0042* (0.0003) �0.0092* (0.0006) 0.0133* (0.0009) age �0.0012* (0.0000) �0.0026* (0.0001) 0.0039* (0.0002) income (10k) �0.0003* (0.0000) �0.0006* (0.0001) 0.0009* (0.0002) wealth (10k) �0.0001* (0.0000) �0.0002* (0.0001) 0.0003* (0.0002) note: significance is defined as follows: *significant at the one-percent level. wealth and income means and standard errors reported in $10,000s. n ¼ 59,404 observations from 17,672 retirees. table 4 s&p 500 returns wave (year) s&p 500 return wave 1 (1992) 7.62 wave 2 (1994) 1.32 wave 3 (1996) 22.96 wave 4 (1998) 28.58 wave 5 (2000) �9.10 wave 6 (2002) �22.10 wave 7 (2004) 10.88 wave 8 (2006) 15.79 wave 9 (2008) �37.00 wave 10 (2010) 15.06 wave 11 (2012) 16.00 wave 12 (2014) 13.69 note: returns include both price changes and dividends generated the s&p 500 index for the year analyzed. b. pearson, m. guillemette / financial services review 28 (2020) 341–350 347 conclusions the only financial resources available to fully-retired individuals are from saved assets and income from sources such as annuities, pensions, and government transfers. aside from reentering the labor force, retirees have few options to increase their income in retirement. one-way retirees can increase income is from the management of their saved assets. for example, one option available to retirees is to convert their saved assets into non-labor income. retirees often facilitate this by annuitizing their saved assets. however, this option may not be an optimal solution for retirees with bequest motives. in addition, illiquidity brought about by annuitization may decrease a retiree’s ability to afford larger unexpected expenses, such as the occurrence of a health shock. increasing financial risk is potentially another option to increase retiree income. however, traditional asset management approaches advocate for the opposite, suggesting that retirees decrease financial risk in retirement. most of these arguments are rooted in the traditional time-horizon asset management approach. this approach suggests that when a goal, such as retirement, approaches, individuals should transition their financially risky assets into lessfinancially risky assets. during the pre-retirement stage of the lifecycle, individuals should reduce risk in response to a decline in human capital. because human capital resembles a “bond-like” asset, increasing bond exposure as an investor approaches retirement is a theoretically sound decision. the rationale behind glide paths within target date or lifecycle funds fits within this theoretical framework, as these funds reduce equity exposure and increase bond exposure as an individual approaches a retirement date. once a retiree exits the labor market, and the present value of future earnings from labor is zero, human capital is no longer a significant asset within a holistic portfolio, and a future reduction in portfolio risk may not be warranted from a theoretical standpoint. absent human capital, there is not a strong theoretical rationale for the reduction of risk during the table 5 average marginal effects of the risk ratio on retirement satisfaction including s&p 500 returns “not at all” “moderately satisfied” “very satisfied” risk ratio �0.0092* (0.0021) �0.0135* (0.0032) 0.0223* (0.0053) health (poor as base outcome) fair �0.0694* (0.0034) �0.0461* (0.0020) 0.1116* (0.0051) good �0.1100* (0.0034) �0.1037* (0.0027) 0.2137* (0.0054) very good �0.1372* (0.0035) �0.1676* (0.0037) 0.3048* (0.0061) excellent �0.1492* (0.0037) �0.2077* (0.0059) 0.3569* (0.0083) male (female as base outcome) 0.0008 (0.0021) 0.0013 (0.0029) �0.0022 (0.0052) white (non-white as base outcome) �0.0277* (0.0026) �0.0410* (0.0038) 0.0687* (0.0064) married (non-married as base outcome) �0.0307* (0.0018) �0.0455* (0.0026) 0.0762* (0.0044) education �0.0050* (0.0003) �0.0075* (0.0005) 0.0125* (0.0008) age �0.0022* (0.0001) �0.0032* (0.0001) 0.0054* (0.0002) income (10k) �0.0006* (0.0001) �0.0008* (0.0002) 0.0014* (0.0003) wealth (10k) �0.0001* (0.0000) �0.0002* (0.0000) 0.0003* (0.0001) s&p 500 returns 0.0002* (0.0000) �0.0003* (0.0000) �0.0004* (0.0000) note: probit model with random effects. significance is defined as follows: *significant at the one-percent level. wealth and income means and standard errors reported in $10,000s. n = 59,404 observations from 17,672 retirees. 348 b. pearson, m. guillemette / financial services review 28 (2020) 341–350 retirement stage of the lifecycle. instead, retirees should decide how much variation in asset returns and consumption they are willing to accept for a given level of risk. if retirees reduce the ratio of stocks held relative to total financial assets, it may decrease retirement satisfaction. on the other hand, assuming a retiree is willing to accept the risk of higher return variation, increasing the ratio of stocks in a portfolio may increase income for consumption and enhance satisfaction in retirement. it is important to note that there is a potential downside to increasing portfolio risk for a retiree to achieve higher future returns. higher portfolio risk should theoretically increase asset return variation, which may lead to greater uncertainty about consumption outcomes in the future. enhancing portfolio risk may be a satisfactory investment solution if a retiree is willing to accept the possibility of higher consumption uncertainty in exchange for the potential for higher returns. however, if a retiree is not willing to bear additional portfolio risk for greater return potential in the future, then an increase in portfolio risk may not be appropriate. it is important for retirees, and their financial planners, to consider risk tolerance when making investment decisions. in this study, we only consider objective risk preferences in our analyses. future studies should explore whether controlling for subjective risk preferences alters the association between stockholding and retirement satisfaction. notes 1 health and retirement study, (rand hrs 2014 fat file (v2a)) public use dataset. produced and distributed by the university of michigan with funding from the national institute on aging (grant number nia u01ag009740). ann arbor, mi, (2018). 2 rand hrs 2014 fat file (v2a). produced by the rand center for the study of aging, with funding from the national institute on aging and the social security administration. santa monica, ca (2018). references bachmann, k., hens, t., & stössel, r. (2017). which measures predict risk taking in a multi-stage controlled investment decision process. financial services review, 26, 339–365. barfield, r. e., & morgan, j. n. (1978). trends in satisfaction with retirement. the gerontologist, 18, 19–23. bellante, d., & green, c. a. (2004). relative risk aversion among the elderly. review of financial economics, 13, 269–281. bonin, h., dohmen, t., falk, a., huffman, d., & sunde, u. (2007). cross-sectional earnings risk and occupational sorting: the role of risk attitudes. labour economics, 14, 926–937. bouchey, p. (2004). questionnaire quest. financial planning, 34, 97–99. chen, j., zhu, z., & yuan xie, h. (2004). measuring intellectual capital: a new model and empirical study. journal of intellectual capital, 5, 195–212. cordell, d. m. (2002). risk tolerance in two dimensions. journal of financial planning, 15, 30. corter, j. e., & chen, y. j. (2006). do investment risk tolerance attitudes predict portfolio risk? journal of business and psychology, 20, 369. diener, e., ng, w., harter, j., & arora, r. (2010). wealth and happiness across the world: material prosperity predicts life evaluation, whereas psychosocial prosperity predicts positive feeling. journal of personality and social psychology, 99, 52–61. b. pearson, m. guillemette / financial services review 28 (2020) 341–350 349 dorfman, l. t. (1989). retirement preparation and retirement satisfaction in the rural elderly. journal of applied gerontology, 8, 432–450. dyer, j. s., & sarin, r. k. (1982). relative risk aversion.management science, 28, 875–886. easterlin, r. a. (2003). explaining happiness. proceedings of the national academy of sciences, 100, 11176– 11183. elder, h. w., & rudolph, p. m. (1999). does retirement planning affect the level of retirement satisfaction? financial services review, 8, 117–127. fitz-enz, j. (2000). the roi of human capital: measuring the economic value of employee performance. new york, ny: amacom: a division of american management association. granger, c. w. (1969). investigating causal relations by econometric models and cross-spectral methods. econometrica, 37, 424–438. hanna, s. d., & chen, p. (1997). subjective and objective risk tolerance: implications for optimal portfolios. financial counseling and planning, 8, 17–26. hanna, s., gutter, m., & fan, j. (2001). a measure of risk tolerance based on economic theory. journal of financial counseling and planning, 12, 53. kimmel, d. c., price, k. f., & walker, j. w. (1978). retirement choice and retirement satisfaction. journal of gerontology, 33, 575–585. lusardi, a., mitchell, o. s., & oggero, n. (2018). the changing face of debt and financial fragility at older ages. aea papers and proceedings, 108, 407–411. mulligan, c. b., & sala-i-martin, x. (2000). measuring aggregate human capital. journal of economic growth, 5, 215–252. pålsson, a. m. (1996). does the degree of relative risk aversion vary with household characteristics? journal of economic psychology, 17, 771–787. pratt, j. w. (1964). risk aversion in the small and in the large. econometrica, 32, 122–136. price, c. a., & balaswamy, s. (2009). beyond health and wealth: predictors of women’s retirement satisfaction. the international journal of aging and human development, 68, 195–214. schultz, t. w. (1961). investment in human capital. the american economic review, 51, 1–17. seccombe, k., & lee, g. r. (1986). gender differences in retirement satisfaction and its antecedents. research on aging, 8, 426–440. shaw, k. l. (1996). an empirical analysis of risk aversion and income growth. journal of labor economics, 14, 626–653. sung, j., & hanna, s. d. (1996). factors related to risk tolerance. financial counseling and planning, 7, 11–20. taylor-carter, m. a. ., cook, k., & weinberg, c. (1997). planning and expectations of the retirement experience. educational gerontology: an international quarterly, 23, 273–288. van solinge, h., & henkens, k. (2008). adjustment to and satisfaction with retirement: two of a kind? psychology and aging, 23, 422–434. yook, k. c., & everett, r. (2003). assessing risk tolerance: questioning the questionnaire method. journal of financial planning, 16, 48. 350 b. pearson, m. guillemette / financial services review 28 (2020) 341–350 financial (il)literacy vs. individual’s behavior: evidence on credit card repayment patterns gustavo barbozaa,b, paola bonginic, monica rossolinic,* ajack and vada reynolds endowed chair professor of international business, loyola university of new orleans, 6363 st charles avenue, new orleans, la 70118, usa bdepartment of business and law, university of milano-bicocca, via bicocca degli arcimboldi, 8 20126, milan, italy cdepartment of business and law, university of milano-bicocca, via bicocca degli arcimboldi, 8 20126, milan, italy abstract we explore the role that financial (il)literacy and personal traits have on financial behavior. using a sample of 156 college students from the united states, we provide unique empirical evidence by specifically differentiating between individuals with higher levels of financial literacy versus individuals declaring not knowing the answers to financial literacy questions and those answering incorrectly. thus, we assess the implications of revealed lack of financial knowledge on financial behavior regarding credit card use in comparison with two other cohorts; cohort one answering correctly, and cohort two failing to answer correctly. a novelty of our study is that we contrast these results to the behavioral factors of over spending and surprised levels of spending—proxies for personality traits—when using credit card. our exploratory empirical findings indicate that among personal-traits considered in this study overspending results in lack of payment in full in credit card debt, and more importantly these effects dominate any gains derived from financial literacy. to this extent financial literacy appears to only play a marginal role avoiding month-to-month credit card debt. furthermore, financial knowledge derived from parents has a strong positive effect on individuals’ financial behavior especially for students characterized by a relevant financial illiteracy. the implications of this research support the argumentation that early financial literacy may have the strongest effect in shaping individuals inherent behavior patterns; that is, early exposure to financial education is strictly preferred and should be promoted at early stages of the educational system. © 2021 academy of financial services. all rights reserved. jel classification: i22; g41 keywords: financial literacy; credit card debt; college students; financial behavior *corresponding author. tel.: +39 02 6448 3014; fax: +39 02 6448 3072 e-mail address: monica.rossolini@unimib.it 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 247–276 1. introduction issues relating to financial knowledge and individual behavioral patterns are at the core of every individual’s economic decision (thaler & shefrin, 1981). it is well documented that financially illiterate individuals incur in costly or even improper financial decisions with significant negative spillover effects on other aspects of their life (see lusardi & mitchell, 2007; and van rooij, lusardi, & alessie, 2011, 2012, among others). for instance, previous studies demonstrate (see lusardi & mitchell 2014, for an excellent review of the literature) that when young people possessing low levels of financial literacy and combined with a widespread use and easy access to credit cards (especially among college students), results in making the wrong financial decisions; most notably carrying over large balances in credit card debt. as a matter of fact, credit card debt coupled with financial illiteracy is associated with unhealthy behaviors (adams & moore 2007; berg et al., 2010; lyons & rogers, 2004), lower academic performance (pinto et al., 2001), and lower financial well-being (grable & joo 2006; norvilitis et al., 2006; nelson et al., 2008). the negative effects of poor financial decisions are well stated; however, what causes and fuels this behavior is a much lesser study field. in this context, the role of personality traits, individual preferences with hyperbolic discounting, and attitudes in influencing financial decision-making have been less investigated to this date in the literature. this is more interesting as these issues are increasingly attracting the attention of psychologists and, and more recently yet to a lesser extent economists (see brown & taylor, 2014 and barboza, 2018, for instance). similarly, issues relating to one self’s perception (i.e., confidence about one’s knowledge or ability to conduct financial decision in particular) has been explored in the realm of behavioral economics and finance (thaler, 1980; thaler & shefrin, 1981). however, despite these advances, it is only until recently that the combined elements have become a subject of analysis in conjunction. more specifically the actual financial knowledge showing that overconfident individuals, or those with high self-assessed knowledge but low actual knowledge, have a higher propensity to engage in risky (costly) financial behaviors (tokar asaad, 2015; brown & taylor, 2014; chu et al., 2017; kramer, 2016; porto & xiao, 2016; xia et al., 2014). on the basis of these considerations and given the importance that financial literacy plays in optimal decision-making processes, our paper aims at studying the relationship between individual behavior and financial (il)literacy. more specifically, we aim at providing robust empirical evidence useful to fill the gap where little attention has been given to study the impact of both financial literacy and of personality traits, preferences and attitudes on overall financial behavior measured as repayment patterns in credit card debt. specifically, we hypothesize that regardless to one’s financial knowledge level, specific personality traits, such as present-bias preferences or impulse behavior, may negatively impact the financial decision-making processes and overpower the (potential) positive effects from higher level of financial literacy. secondly, we hypothesize that self-awareness of not possessing a strong financial knowledge (admitted by choosing the do not know option in a financial literacy questionnaire) may lead to more conservative financial 248 g. barboza et al. / financial services review 29 (2021) 247–276 decisions. in turn these decisions could prevent individuals from assuming too much financial risk, resulting in a “correct” or less costly financial decision. this paper studies credit cards use among college students as we investigate the financial behavior and decisions that college students make in terms of credit card repayment patterns. that is, whether students select to pay the credit card balance in full every billing cycle, or decide to carry a month-to-month balance, either by paying less than the full balance, by only the minimum payment (anchoring) or by falling behind on their payments. we use a sample of 156 college students from the united states and consequently estimate a series of ordered probit models. our results provide unique empirical evidence by creating a distinctive differentiation between individuals with higher levels of financial literacy versus individuals declaring not knowing the answers to financial literacy questions and those answering incorrectly. due to the relevance that perception about oneself has on actual behavior, we bring forth the hypothesis that incorrectly answering financial questions versus answering “i do not know” may have potentially large implications relating to the way individuals react in the face of making—important—financial decisions. besides, we account for a series of personality traits (namely overspending, anxiety, and mental accounting issues) to test for their effects on financial behavior (credit card repayment). an important finding of our study is that these personality traits may have stronger (negative) effects not easily overcome by higher levels of financial literacy. as a spoiler alert, our exploratory results find robust evidence indicating that individuals’ personal traits are the main driver for individuals to accumulate and carry over a month-tomonth balance in their credit card(s). specifically, attitude towards overspending is the main factor negatively affecting the capability of individuals to repay in full credit card debt every billing cycle. this result is in line and provides statistical support to the importance of mental accounting issues faced by individuals with present bias preferences. we also find, contrary to previous literature (norvilitis et al., 2006; robb, 2011; shim et al., 2009, 2010), that a higher level of financial literacy is not a fundamental and it only controls on the margin for personal traits and attitudes (overspending, lack of self-control, and issues related to poor mental accounting). furthermore, financial knowledge derived from parents has a strong positive effect on individuals’ financial behavior especially for people characterized by high levels of financial illiteracy. our findings reinforce the evidence of those studies that specifically investigated the relationship between the role of parents and the (mis)use of credit cards by college students (hancock et al., 2013; xiao et al., 2011). our findings may serve policy makers when designing specific policies aimed at avoiding, or reducing, debt traps and socio-economic vulnerability of the borrower. our paper contributes to the extant literature in at least three strands. at this point we do not make any claim to make a theoretical contribution to the field. our contributions are empirical and they are as follows. first, we provide a unique and innovative empirical break down between financial literacy and financial illiteracy. in particular we separate wrong answering to a standard financial literacy questionnaire (five questions in total; see section 2 for more details on the questionnaire) between incorrect answering and do not know responses. the former is classified as an attempt to respond and measures an overconfidence about one’s actual knowledge which at the end is revealed to be poor (overconfidence). whereas the latter represents a direct answer to not knowing and therefore to not possess overconfidence on his or her g. barboza et al. / financial services review 29 (2021) 247–276 249 knowledge level. it is relevant to point out that while at present previous research has studied the role of financial literacy (correct/incorrect) in decision making, little has been dedicated to study the relation to how much people actually do not know or believe they do not know. therefore, our central contributing focuses on accounting the magnitude and explicit separation between an inaccurate perception of knowledge level leading to incorrectly answering, versus individuals actually not knowing and clearly revealing this lack of knowledge.1 to the best of our knowledge, we are the first to explore this issue.2 secondly, we study the effect of consumption and spending patterns, as they relate to overspending (present bias) and surprised factors (inadequate mental accounting) as they lead to credit card debt accumulation. the issues at stake here are similar to those first introduced by thaler (1980) and thaler and shefrin (1981) on the different effects that individual mental accounting patterns may have between a planner (sophisticated self) and doer (naı̈ve self). more specifically, the decisions individuals have to make include but are not limited to borrowing money, using credit cards, and adjusting consumption to income flows and avoiding falling prey of present-bias behavior and consequently accumulate a month-to-month balance on their credit card. in this context we then hypothesize that those with higher levels of self-awareness have the knowledge to answer correctly (basic financial literacy questions) or recognize their limitations and face the reality of not knowing the answer, and openly recognizing that by answer do not know. consequently, here our interest is to uncover the differences between those that are aware of their limitations versus those that are or possess excessive self-confidence, while being wrong. it is in this difference that we expect to reach relevant findings as to why individuals with certain personality traits are more likely to make incorrect and costly financial decisions. we believe this is the first paper that directly addresses this very important and timely issue. finally, our third contribution is to study the role of financial literacy in ameliorating negative effects of costly personality traits on credit card repayment patterns. the extant literature assumes that higher levels of financial literacy are a predominantly determinant of superior financial performance. however, while significant contributions have been made in advancing the role of financial literacy, previous works in this field have paid little attention to the role that personality traits (aptitude and attitude) play in financial decision and secondly what the role of financial (il)literacy is in shaping these inherent behavioral patterns.3 we organize the rest of the paper as follows. the next section reviews the most relevant literature on financial literacy and personal traits and sets forth our testing hypotheses. the third section describes the research design and the estimation model. the fourth section analyzes estimation results, which are then discussed in the fifth section, while the last section presents some general conclusions and policy recommendations. 2. literature review recently the topic of financial literacy has received increased and extensive attention in the literature. the literature argues in favor of the relevance of adequate and timely financial literacy as paramount on individual decision-making process and their outcomes. in fact, the evidence points out in support of the argument that low levels of financial literacy are not only linked with high levels of personal and household debt (lusardi & tufano, 2009; 250 g. barboza et al. / financial services review 29 (2021) 247–276 moore, 2003; stango & zinman, 2009), inadequate retirement planning (hilgert, hogarth, & beverly 2003; lusardi & mitchell, 2007), or inadequate stock market participation (van rooij, lusardi, & alessie 2011), but also to poor health (joo & garman, 1998) or adverse health choices (peters et al., 2007) and in general poorer overall life outcomes. in this respect, widespread financial illiteracy among young people is of particular concern for two main reasons. first, as they enter adulthood, a number of important financial decisions are to be undertaken (such as financing college studies; moving away from home; purchasing their first car; using credit cards; saving for retirement; etc.), for which they might not be adequately prepared. misguided financial decisions in the early stage of their lives could have potentially disastrous consequences (huge debt, a poor credit rating, and inadequate retirement plans) for the remaining of their whole life (schagen & lines, 1996; lusardi, mitchell, & curto 2010). second, a lack of financial literacy seems to impact students’ university performance as noted by kezar and yang (2010) whom suggest that a student’s academic achievement is negatively affected by financial distress, which, in turn, is a more likely outcome in presence of low levels of financial literacy. in particular the literature documents that inadequate financial skills (especially in the area of cash management) result in higher level of stress and even anxiety with significant (negative) impact on academic performance (kapoor et al., 2006; razafimahasolo et al., 2016; xiao et al., 2011). low financial capabilities generate financial stress and anxiety, which in turn create negative spillover effects into other life dimensions. in more recent research, particularly in the last decade, the majority of theoretical and empirical literature on financial literacy has addressed and investigated many different topics. these topics cover from the influencing factors that drive financial literacy to the methodological approaches to best treat survey questions that try to effectively measure a latent variable such as financial literacy. a general agreement has been reached such as on the necessary prerequisites to gauge financial literacy (including the types of knowledge that best motivate and facilitate financial action). it has now become common knowledge that financial literacy among both adults and the young is low; and influenced by socio-demographic factors, such as gender, education, income, employment status, and age. it is also recognized that informal sources of education such as family background and interaction with peers are of particular importance in this regard (see lusardi & mitchell, 2014, for a review of theory and empirical evidence). more specifically, individuals’ financial literacy seems to be significantly and positively associated with parental educational attainment and with the presence of forms of financial socialization within the family and the group of peers (for instance children observing their parents’ saving behavior or receiving a more formal financial education from them). in general, the literature also agrees that young adults receive financial literacy through two main sources, parents and the educational system (lusardi et al., 2009). in particular, lusardi et al (2009) note that young adults with college-educated parents tend to have a better understanding of financial concepts. mandell (1997) provides evidence on the role that proper types of financial education play as a significant factor in achieving financial literacy, while others indicate that much of the financial education is being conducted through business and community organizations, and not through educational institutions (fox et al., 2005). g. barboza et al. / financial services review 29 (2021) 247–276 251 however, despite that the role of diversity (e.g., gender, ethnicity, age, and education) has been well documented, including considering different settings and time spans, our understanding of the mechanisms and channels of how these forms of diversity continue to result in significant gaps in financial literacy and consequently results in pervasive effects, remains a conundrum. the quest to find solutions to these differences is far from complete. this challenge is more complicated given the plethora of methodologies used to assess financial literacy. to date, the issue of assessing financial literacy has concentrated on the conceptual definition of this latent variable. houston (2010) and remund (2010) provide a thorough literature review that helps frame the issue of the conceptual definition of what financial literacy is or should be. indeed, financial literacy has been variably defined as specifically referring to a form of knowledge (e.g., hilgert, hogarth, & beverly, 2003), the ability to apply that knowledge (e.g., mandell, 2008), or good financial behavior (e.g., moore, 2003). the constructs used to measure financial literacy vary quite substantially according to the different conceptual definitions adopted. in fact, the construct either covers a wide variety of financial topics, including debt, insurance, spending, investments and retirement savings, budgeting, and inflation, or focuses on a few financial issues. accordingly, the number of questions used to assess financial knowledge levels varies widely, ranging from three to 45 total items. across studies, both multiple-choice questionnaires and self-report questions have been employed to measure financial literacy; where the former are knowledge based and the latter assess perceived knowledge. more recently, surveys have been designed to gauge both objective knowledge and perceived knowledge. in general, considerable progress has been achieved in the design of surveys aimed at identifying individual levels of financial literacy through the effort made by the oecd and its international network on financial education (infe). jointly they develop and promote a common questionnaire based on the experience of a large number of previous rigorous national and international surveys. the oecd/infe (2012) report describes the questionnaire’s underlying methodology and kempson (2009) provides further details on this subject.4 due to its importance, research on financial literacy has in fact inspired numerous public initiatives at both the national and international level. several countries now have financial literacy initiatives and strategies in place to increase the levels of financial understanding and knowledge among all citizens. in contrast, it is not until very recently, that the process of data analysis (i.e., of analyzing the information obtained through questionnaires) has taken a central role in exploring the difference between financial literacy and illiteracy. that is to say, the emphasis has been on what is known but not in what it is not known. in other words, the research focus has been placed on people thinking that they know enough to be correct, but not realizing that they do not know enough to be correct. due to the large arrays of data sources, it has become necessary to use both bivariate (ordered data from less to more) and multivariate techniques to quantify financial (il)literacy. in general, responses to the stated questions are simply summed to generate an index (score) of financial literacy, which typically ranges between zero and the maximum number of correct answers.5 a common practice to most studies is to cluster the “do not know” responding with “incorrect” answers, in opposition to “correct” answers, notwithstanding the fact that it is 252 g. barboza et al. / financial services review 29 (2021) 247–276 also widely recognized that these two types of responses (dnk and incorrect) might refer to two distinct kind of respondents, with diverse (financial) educational needs. manton et al. (2006) were the first to signal that college women tend to select “don’t know” response more frequently than men, especially on more numerically oriented subjects. in the same vein, lusardi and mitchell (2014) pointed out that “one twist on the differences by sex, (. . .) is that while women are less likely to answer financial literacy questions correctly than men, they are also far more likely to say they ‘do not know’ an answer to a question, a result that is strikingly consistent across countries.” however, so far, the common choice throughout the empirical literature is to include them in the same cluster of wrong answers for methodological issues. exceptions to this practice are two recent papers by chen and garand (2018) and kim and mountain (2019). kim and mountain specifically address the econometric issue of misleading results obtained from ignoring dnk responses and suggest the use of binomial-latent regression models to prevents distortions from dk/rf responses. chen and garand (2018) deepen the well-known issue of gender gap in financial literacy, by giving specific attention to dnk answers. in particular, after having ascertained that women may exhibit lower levels of financial knowledge because they lose the opportunity to hazard a guess and arrive at a correct answer based either on partial knowledge or on random chance, they consider the possibility that there are psychological processes at work involving risk acceptance and confidence in financial knowledge that prompt women to give dk responses at a rate higher than men. as a result, they suggest that future research should consider the relative roles of dk and incorrect responses in measuring financial knowledge. in this study, we follow such a suggestion and aim to contribute to the extant literature by treating respondents, who admit not knowing, differently from those respondents who implicitly consider themselves as knowledgeable but in fact possess a “wrong” knowledge. as a matter of fact, when self-assessed questionnaires are included in multiple choice questionnaire (testing objective knowledge), the evidence indicates that most people are unaware of their own shortcomings, as there is often a substantial mismatch between people’s selfassessed knowledge versus their actual knowledge (lusardi & mitchell, 2014). this incorrect assessment is also directly related to the research on behavior as presented by thaler and shefrin (1981). in particular, we take particular interest in separately scoring “correct,” “incorrect,” and “do not know” answers and conjecture that admitting of not possessing a proper financial knowledge may lead to more conservative financial decisions, that could prevent taking too much financial risk. in addition, we approximate the measures for present bias and mental accounting biases as noted next. we also take a cue from a growing body of literature acknowledging that the drivers of financial choices are not constraint just by knowledge and the acquisition of basic information. in this new stream of research cognitive biases, individual psychological traits and aptitudes, motivations, and timing of the choice to be undertaken are all examples of the behavioral and psychological constraints that interact with economic decisions in general and financial choices in particular. one specific cognitive bias that has attracted much attention in empirical studies connecting financial decisions, behavioral biases, and financial knowledge is overconfidence on one’s actual ability, performance, level of control, or chance of success (moore & healy, 2008). financial confidence, in particular, reflects a self-assessed level of financial knowledge, which may or may not coincide with measured financial g. barboza et al. / financial services review 29 (2021) 247–276 253 knowledge. financial literacy overconfidence has been linked to various risky behaviors such as higher stock market participation (xia et al., 2014), less use of financial advice (kramer, 2016; porto & xiao, 2016) and a preference for direct stock investment rather than in less risky/more diversified mutual funds (chu et al., 2017) or greater likelihood of engaging in risky (costly) financial behaviors, such as taking out a title-loan, or a short-term payday loan (tokar asaad, 2015). from this perspective, we take particular interest in the literature on behavioral economics which assumes, and dictates consequently, that individuals may acquire debt, adjust repayment capabilities, and adhere to differentiated patterns of repayment away from the neoclassical rational expectations teaching. the general consensus is that rationally behaving individuals have perfect foresight, are rational and apply consistent discounting rules on consumption. on the other hand, the literature on behavioral biases proposes the existence of several factors governing individuals’ behavior. particularly, issues relating to hyperbolic discount functions in consumption – present bias –, naı̈ve behavior, lack of self-control and impatience (akerlof, 1991; kuchler 2013; laibson, 1997; o’donoghue & rabin, 1999; thaler & shefrin 1981; thaler 2018, among others) may lead to patterns of consumption, which fueled by easy access to credit cards may result into too much debt accumulation and patterns of procrastination on repayment (barboza, 2018). although aware that the psychological literature on personality traits is not confined to and is richer than that just mentioned on behavioral biases, in this study we are interested in highlighting and measuring those individual features that characterize one’s pattern of consumption and savings and manifest: (1) in a tendency to overvalue immediate rewards (i buy what i like now, disregarding the issue of affordability of the purchase), while putting less worth in long-term consequences (i will think tomorrow how to find the money to afford the purchase); (2) in a process, known as mental accounting, whereby individual expenses will not be considered in conjunction with the present value of one’s total wealth.6 instead it is considered in the context of the current budgetary period and the category of expenses, leading to constraints/relaxations of purchases irrespective to the whole—same—fungible resource that is income plus (eventually) personal net worth ( cheema & soman, 2006; zhang & sussman, 2018) and a greater willingness to pay for goods when using credit cards than cash (prelec & simester, 2001). at the same time the procrastination of debt payment can cause pain and anxiety concerns which feedback into the repayment behavior individuals display next period. we follow a recent strand of literature (andrews & wilding, 2004; fiksenbaum et al, 2017; marjanovic et al., 2013) who define an emotional state, such as financial threat, referring to self-reported fearful-anxious uncertainty regarding one’s current and future financial situation. in this study we take particular interest in the effect that higher financial anxiety may have on credit card repayment behavior. thus, we propose to study the combined and interaction effects of financial (il)literacy and personality traits, as they relate to credit card repayment patterns. with these considerations in mind, then the backbone of our analysis is driven by the following set of hypotheses. hypothesis 1: higher levels of financial literacy—measured as higher number of correct answers— lead to better repayment rates in credit card debt. 254 g. barboza et al. / financial services review 29 (2021) 247–276 hypothesis 2: the higher the transgenerational transmission of financial education from parents to children results in improved repayment in credit cards. hypothesis 3: individuals with self-assessed lower level of financial literacy (dk answers) display diverse behaviors (credit card repayment patterns) as opposed to individuals who do not possess basic financial literacy (wrong answers). hypothesis 4: individuals displaying present bias, improper mental accounting and or impulse behavior and financially derived anxiety are less likely to display proper financial behavior (i.e., pay credit cards in full) and consequently carry month-to-month balance. hypothesis 5: positive effects from higher levels of financial literacy could be overpowered by personal traits. this is to say, that financial literacy may or may not be enough to counter inherent negative traits that individuals already possess. 3. method 3.1. data and variables data for this research comes from a survey administered to three samples of business college students in the united states, attending a midwest higher education university and a mid-atlantic university. the survey was paper based and administered in person to a total of 1,149 students. data were collected in 2015 and the total complete sample size useful for our analysis is of 156 respondents. the sample is composed as follow: 45.50% are female students and 54,50% are male students; minority represents 21.2% whereas white race students are 78.8% of the sample. in terms of academic status, freshman students are 32.05% of the sample, sophomore students 8.97%, junior students are 20.51%, senior students are 36.54%, and graduate students only 1.93%. the main interest of this study is to explore the effects that financial (il)literacy and individual personal traits related to purchasing behavior have on their financial behavior as manifested in credit card repayment patterns. thus, we measure financial behavior as an individual’s credit card repayment pattern and ask questions regarding credit card repayment behavior, such as: pay if full every month; pay some in full and then only the minimum required; pay more than minimum required but not in full; pay minimum required; or pay less than minimum required. we create the categorical variable used to estimate the ordered probit model and corresponding probabilities of occurrence. the survey also includes a combination of questions related to personal traits and financial (il)literacy elements. regarding financial literacy we post five financial literacy questions to students (see fig. 1 for details) that tests the basic financial concepts traditionally tested in the financial literacy literature since lusardi and mitchell (2007) proposed them: inflation, time value of money, diversification, and interest rate compounding. the survey also includes a set of questions designed to capture specific personality traits. these variables serve as proxies to assess individuals’ time preferences (present bias issues), g. barboza et al. / financial services review 29 (2021) 247–276 255 f ig . 1 . f in an ci al li te ra cy q u es ti o n s. 256 g. barboza et al. / financial services review 29 (2021) 247–276 issues relating to mental accounting, and potential issues relating to procrastination and lack of commitment. based on the review of the literature we argue that these topics may manifest in turn in behavioral delays on credit card repayment, as purchase may be in excess to monthly repayment capabilities. therefore, to further understand individuals’ decision-making process, subjects were asked if they were surprised at the end of the billing cycle with the balance the credit cards has reached, and secondly if they have engaged in purchases knowing that they did not have money to pay it in full when the balance was due. we define these two variables as surprised and overspending, respectively. in both cases, our tentative hypothesis is to expect that the higher the level of surprised and the higher the amount of overspending lead to a worsen in credit card repayment patterns in the next billing cycle. to identify one’s personal attitude toward debt repayment we include the variable anxiety where the variable takes the values according to the following scale: 1= not anxious at all and 5 = extremely anxious. in our survey, students self-report the level of their financial anxiety answering to a specific question of the survey. anxiety is in relation to how the person feels when he or she has to pay the credit card bill at the end of the billing cycle. finally, in addition to financial literacy questions our survey also considers a financial education question. in particular we analyze the role that parents may play in the buildup of the financial literacy levels of their children. the survey asks whether their parents were the primary source of financial education and we define a dummy variable fepar when parents are reported as the main source of financial education. in line with the transgenerational effect found elsewhere (barboza, smith, & pesek 2016), we argue that if parents are the main source of financial education, parents’ financial experiences serve as a mechanism to develop knowledge spillover effects and possibly avoiding painful self-experiences that could result in lower levels of financial-based anxiety and translate onto better credit card repayment patterns. based on the number of correct answers to the financial literacy questions, we first proceed to construct the cumulative number of total correct answers. the number of correct answers is thus the actual level of financial literacy (that we called financial literacy rate), as it is presented in most of the extant literature. we take particular interest in the objective separation between correct, incorrect, and do not know answers. therefore, we construct the variable total number of incorrect answers given by each individual, and call it financial illiteracy level a. finally, we construct the total number of answers under the do not know category. as noted earlier, we want to emphasize that answering incorrectly is different than answering do not know. correspondingly, answering dnk is labeled as financial illiteracy level b. we argue that those answering incorrectly (incorrect) have attempted to answer assuming that they have an adequate level of knowledge; however, clearly failing to achieve a correct answer. on the other hand, those answering i do not know (dnk) openly acknowledge that they are not prepared to even attempt answering recognizing a higher level of financial illiteracy with no fear to state it as such. the fundamental difference is that the dnk group may be less ready to engage in financial decision-making processes, as their self-awareness indicates a clear lack of preparation. conversely, those responding incorrectly (level a) implicitly assume knowing the answer but failing. the potential implications of the separation of answering not correctly into these two distinctive groups could be relevant in understanding the implications of different levels of financial illiteracy. we also g. barboza et al. / financial services review 29 (2021) 247–276 257 provide a decomposition of answers (under the three categories) by question as these relate to different financial concepts. 3.2. descriptive statistics table 1 presents a summary of results from the questionnaire. these results indicate that a large majority of individuals answer correctly questions on compound interest, inflation and purchasing power. a large number of students were not capable of answering correctly the question regarding inheritance. in terms of the inflation question, a relative similar proportion of individuals answer dnk or incorrectly. for the risk question a significant number of students responded dnk. in general, the data appears to indicate that dnk answering is a prevalent issue. in addition, a large number of student answer incorrectly. the combination of both levels of financial illiteracy confirms the prevalence of wide spread lack of financial knowledge. when we look at the cumulative answering (see table 2, below), we observe that 5% of the population are not capable of answering any question correctly. by the same token, only 12% of the individuals are capable of answering all questions correctly. in addition, a large proportion of individuals answer dnk to one and two questions. when combined with those also answering incorrectly one to two questions, we are able to observe a large proportion of the population struggling with about 50% of the questions asked. in general, this evidence indicates that financial illiteracy is highly present among the population under study. this statistic is in line with previous literature and with most recent valuations made across the u.s. population (see lusardi & mitchell 2014). table 3 disassembles financial literacy rate by demographics. the interpretation of the results thus indicates that females are more likely to answer correctly since people able to provide five correct answers are mainly females (62.5%); this is a result that contradicts the empirical evidence found in the literature for the united states (borden et al. 2008; chen & volpe, 2002; danes & tahira, 1987; lusardi, table 2 financial (il)literacy indicators (total answers) no. of answers correct do not know incorrect 0 4.9% 52.9% 22.1% 1 10.0% 31.4% 41.9% 2 18.0% 9.2% 25.9% 3 27.4% 3.1% 7.6% 4 27.7% 2.2% 1.5% 5 11.9% 1.0% 1.0% table 1 financial (il)literacy indicators per question correct do not know incorrect compound interest 70.3% 6.0% 23.7% inflation 66.8% 16.1% 17.1% inheritance 33.5% 3.1% 63.3% purchasing power 77.2% 8.2% 14.6% risk 51.0% 40.1% 8.9% 258 g. barboza et al. / financial services review 29 (2021) 247–276 mitchell, & curto, 2010; markovich & devaney, 1997) where it is argued that females possess lower levels of financial literacy. however, outside the united states, this is not a standard evidence (bongini et al, 2016; koshal et al., 2008; wagland & taylor, 2009) when specifically studying business students. in addition, older students are more likely to answer correctly, as well as upper classmen or classwomen. among different academic status graduate students provide at least three correct answers and the 33% of students answers correctly to all the financial literacy questions. this interpretation of the results speaks in favor of the educational process. as expected, one can argue that attending college should increase knowledge in field specific subjects, financial literacy being one of them. finally, the evidence does indicate that minorities are at a disadvantage and more likely to answer incorrectly or do not know. this last piece is in agreement with the extant literature. table 4 below presents some basic descriptive statistics on the variables used in the empirical estimation section. a full description of each variable and its corresponding coding could also be found there. with these considerations in mind, we then proceed to outline the model specification and the corresponding model estimations expectations and restrictions. 3.3. the model as we research the effects of personal traits and financial literacy on financial behavior, the basic model description has the following general specification: y�i ¼ x 0 ib þ « i (1) where « i are assumed independent and identically distributed random variables as usual, x 0 i is the matrix of explanatory variables (financial (il)literacy and personality traits), b is the vector of coefficients to be estimated, and y�i is unobserved yet described by the answers to our survey questions relating to credit card repayment habits. in fact, we code students’ responses on credit card repayment capability using a discrete categorical scale as follows: 1 = pay if full every month, 2 = pay some in full and other only table 3 financial literacy rate by demographics financial literacy rate gender mean as mean age race 0 0.333 1.939 20.152 0.697 0.2 0.313 2.119 20.373 0.731 0.4 0.430 2.421 20.554 0.777 0.6 0.492 2.415 21.196 0.837 0.8 0.478 2.595 21.978 0.892 1 0.625 2.838 23.213 0.938 source: demographic variables are defined as follows. gender g is a dummy variable taking value of 1 if female and 0 otherwise. academic status-as is the 1 = freshman, 2 = sophomore, 3 = junior, 4 = senior, and 5 = grad student. race is 1 if white and 0 if minority. age is expressed in years at the time of the survey. flitrate is the percentage of correct answers provided, based on a five question questionnaire as presented in figure 1. g. barboza et al. / financial services review 29 (2021) 247–276 259 t ab le 4 d es cr ip ti v e st at is ti cs o f v ar ia b le s b y ca te g o ry , d es cr ip ti o n , an d co d in g c at eg o ry d es cr ip ti o n c o d e m ed ia n m ea n s d m ax m in o b s d em o g ra p h ic s a ca d em ic st at u s (f = 1 , s = 2 , j = 3 , s r = 4 , g ra d = 5 ) a s 3 2 .6 7 3 1 .3 1 1 5 1 1 5 6 g en d er (f em al e = 1 ) g 0 0 .4 5 5 0 .5 1 2 2 0 1 5 6 a g e a g e 2 0 2 0 .6 0 3 2 .8 5 3 4 2 1 8 1 5 6 r ac e (m in o ri ty = 0 ) r a c e 1 0 .7 8 8 0 .4 1 0 1 0 1 5 6 f in an ci al ed u ca ti o n an d li te ra cy p ar en ts ar e m ai n so u rc e o f fi n an ci al ed u ca ti o n f e p a r 1 0 .6 2 8 0 .4 8 5 1 0 1 5 6 n u m b er s o f co rr ec t an sw er s t o tc o r 3 .0 0 2 .8 7 8 1 .2 6 2 5 0 1 5 6 n u m b er s o f in co rr ec t an sw er s t o t_ in c 1 .0 0 1 .3 9 1 1 .0 3 2 5 0 1 5 6 n u m b er o f “d o n o t k n o w ” an sw er s t o td k 1 .0 0 0 .7 3 1 0 .8 6 8 4 0 1 5 6 b eh av io ra l l ev el o f an x ie ty o n re p ay m en t ca p ac it y a n x 2 2 .0 1 9 1 .1 1 6 5 1 1 5 6 f re q u en cy o f p ay m en t o n cr ed it ca rd r fr eq 1 1 .6 2 2 1 .0 1 2 4 1 1 5 6 s u rp ri se d o n c c b al an ce le v el s u rp 1 0 .8 7 8 0 .9 1 8 3 0 1 5 6 f re q u en cy o f o v er sp en d in g o v er sp d 0 0 .4 6 2 0 .8 2 2 3 0 1 5 6 n u m b er o f cr ed it ca rd n c c 1 1 .5 6 4 1 .0 6 1 6 1 1 5 6 s o u rc e: d em o g ra p h ic v ar ia b le s ar e d efi n ed as fo ll o w s. g en d er g is a d u m m y v ar ia b le ta k in g v al u e o f 1 if fe m al e an d 0 o th er w is e. a ca d em ic s ta tu sa s is th e 1 = f re sh m an , 2 = s o p h o m o re , 3 = ju n io r, 4 = s en io r, an d 5 = g ra d s tu d en t. r ac e is 1 if w h it e an d 0 if m in o ri ty . a ge is ex p re ss ed in y ea rs at th e ti m e o f th e su rv ey . f in an ci al l it er ac y , e d u ca ti o n , an d re la te d v ar ia b le s in cl u d e th e fo ll o w in g . t ot c o r is th e n u m b er o f fi n an ci al li te ra cy q u es ti o n an sw er s co rre ct ly ; t ot d k is th e n u m b er o f an sw er s w it h a "d o n o t k n o w " an sw er ; an d t ot _i n c is th e n u m b er o f in co rr ec t an sw er s. f o r ea ch in d iv id u al th e su m o f t o t c o r + t o t d k + t o t _ in c = 5 . t h e f in an ci al l it er ac y r at ef l itr at e m ea su re s th e p er ce n ta g e o f co rr ec t an sw er s o u t o f th e fi v e li te ra cy k n o w le d g e q u es ti o n s. a n x ie ty -a nx is th e le v el o f an x ie ty re p o rt ed o n a sc al e fr o m 1 = n o an x ie ty to 5 = ex tr em el y an x io u s in re la ti o n to h is o r h er ca p ac it y to re p ay y o u r cr ed it ca rd m o n th ly b il l. f e p ar ta k es a v al u e o f 1 if p ar en ts ar e th e m ai n so u rc e o f fi n an ci al ed u ca ti o n , 0 o th er w is e. r f re q d efi n es th e re p ay m en t b eh av io r o n cr ed it ca rd s. t h e v ar ia b le ta k es th e v al u es o f 1 = p ay in f u ll ; 2 = p ay m o re th an m in im u m , b u t ca rr y m o n th -t o -m o n th b al an ce ; 3 = p ay o ff so m e cr ed it ca rd s, b u t p ay s m in im u m o n th e re st ; 4 = p ay m in im u m in al l; 5 = p ay le ss th an m in im u m in al l. 6 = 1 = p ar en ts p ay in fu ll w il l b e th e as su m p ti o n . su rp ri se d d efi n es th e le v el o f su rp ri se th e in d iv id u al re p o rt s o n h o w h ig h o f a b al an ce is o n th e m o n th ly st at em en t. t h e v ar ia b le ta k es th e v al u es o f 0 = n ev er , 1 = r ar el y , 2 = s o m et im es , 3 = f re q u en tl y . o ve rs pd d efi n es th e b eh av io r o f th e in d iv id u al re g ar d in g th e in d iv id u al k n o w in g th at sh e o r h e u se th e cr ed it ca rd k n o w in g th at sh e o r h e d id n o t h av e m o n ey to p ay w h en th e b il l ca m e d u e. t h e v ar ia b le ta k es th e v al u es o f 0 = n ev er , 1 = ra re ly , 2 = so m et im es , 3 = fr eq u en tl y . f in al ly , n c c is a co u n t v ar ia b le m ea su ri n g th e n u m b er o f cr ed it ca rd s an in d iv id u al p o ss es se s. 260 g. barboza et al. / financial services review 29 (2021) 247–276 minimum required, 3 = pay more than minimum required but not in full, 4 = pay minimum required, 5 = pay less than minimum required. therefore, our dependent variable (rfreq) takes discrete values along the scale 1 to 5. notice that rfreq of values ranging from 2 to 5 results in accumulation of month-to-month balance, with the risk of incurring high financial costs related to interest rates and possibly other assessed fees. in consideration of the ordinal nature of the dependent variable, the most appropriate model to use is an ordered probit model. the main difference among the models regards the financial literacy variable. due to our interest to test for both knowledge and lack of knowledge, we decompose the responses given by individuals between correct, incorrect and do not know classifications, per our discussion in the previous section. in the first model we consider as financial literacy variable the number of correct answers. in the second model we consider the number of questions where the students said that he does not know the answers, and in the last model we consider the number of incorrect answers. furthermore, because each of the financial literacy questions measures knowledge of different financial issues (more or less directly related to our dependent variable, i.e., credit card repayment), we also proceed to conduct estimations by question.7 thus, we propose to estimate the impact of each financial issue on the financial behavior under investigation: our intuition is that some questions may be more relevant (have a larger impact) and the separation of effects may yield relevant results, when understanding credit card repayment patterns. in all the specifications, we consider all the personality trait variables and the presence of financial education from parents; plus, as control variable we include the total number of credit cards possessed by our sample students (ncc). this latter variable proves to be very much correlated to the usual demographics variables (age, gender, race, and academic status) used as controls; we choose to use ncc as our single control variable because the number of observations suggested to contain the number of covariates.8 we then proceed to incorporate interaction effects between the overall scores of financial literacy, financial illiteracy level a (incorrect) and level b (dnk), in relation to the financial education from parents and personality trait variables. in addition, and due to the possibility of endogeneity in the data, we conduct hausman testing as a robustness indicator. the argument regarding endogeneity is justified as the possibility of a self-selection problem in terms of the characteristics of those selecting to apply and obtain a credit card. however, the counter argument indicates that in the u.s. market, young adults need to apply for credit cards as a requirement to begin building their credit history and create a credit score. due to this apparent controversy, we apply the endogeneity testing. 4. results and discussion estimations are presented in table 5 where each of the three models is further decomposed to study the impact of each financial issue included in the financial literacy index on credit card repayment behavior of individuals. when looking at the coefficients (sign and magnitude) of the total number of questions answered (correctly, incorrectly, and dnk, respectively) we observe that correct/dkn hold g. barboza et al. / financial services review 29 (2021) 247–276 261 both a negative sign, while only correct being statistically significant at the 10% level of confidence. the negative sign indicates that higher levels of literacy and/or declaring dkn result in better repayment patterns. higher financial literacy having a positive impact on repayment is an expected outcome (confirming hypothesis 1); however, the second component (dnk) is clearly a puzzling result. notice that higher levels of financial literacy are only marginally statistically significant at the 10%. regarding incorrect answering (financial illiteracy) the result is statistically significant and with the expected positive sign. this result supports the hypothesis that lack of financial literacy creates a negative effect on individuals’ financial performance, resulting in an accumulating month-to-month balance in their credit card debt (hancock et al., 2013; robb, 2011; xiao et al., 2011). furthermore, results in table 5 consumer behavior and financial decision making process with credit card repayment as dependent variable model 1 correct answers model 2 incorrect answers model 3 do not know answers 1.1 1.2 2.1 2.2 3.1 3.2 financial educ parents �0.939 �0.973 �0.930 �0.969 �0.912 �0.888 (0.001)*** (0.001)*** (0.001)*** (0.001)*** (0.001)*** (0.001)*** anxiety 0.143 0.124 0.158 0.141 0.163 0.159 (0.18) (0.25) (0.13) (0.18) (0.12) (0.11) surprised 0.143 0.130 0.141 0.133 0.182 0.175 (0.28) (0.33) (0.25) (0.27) (0.16) (0.17) overspending 0.533 0.533 0.506 0.497 0.509 0.511 (0.001)*** (0.001)*** (0.001)*** (0.001)*** (0.001)*** (0.001)*** ncc 0.161 0.156 0.164 0.169 0.135 0.144 (0.08)* (0.10)* (0.04)** (0.04)** (0.13) (0.08)* financial literacy questions total answers �0.150 0.222 �0.022 (0.10)* (0.03)** (0.88) compound interest �0.017 0.020 �0.051 (0.95) (0.94) (0.91) inflation �0.385 0.501 �0.014 (0.12) (0.05)** (0.97) inheritance �0.107 0.163 �0.184 (0.68) (0.47) (0.41) purchasing power �0.046 0.118 �0.125 (0.87) (0.72) (0.82) risk �0.139 0.297 0.022 (0.55) (0.35) (0.92) pseudo r2 0.209 0.212 0.212 0.217 0.197 0.199 obs 158 157 156 156 156 156 lr statistic 65.06 65.71 65.59 67.29 61.10 61.66 probability (lr stat) (0.001)*** (0.001)*** (0.001)*** (0.001)*** (0.001)*** (0.001)*** akaike info criterion 1.674 1.726 1.681 1.722 1.710 1.758 schwarz criterion 1.848 1.979 1.857 1.976 1.886 2.012 hannan-quinn criterion 1.744 1.829 1.753 1.825 1.782 1.861 note: ***, **, * statistically significant at the 1%, 5%, and 10% level, respectively. 262 g. barboza et al. / financial services review 29 (2021) 247–276 models 2.1 and 3.1 clearly indicate that individuals responding incorrectly are significantly more likely to incur in poor credit card repayment behavior. the estimated coefficient is larger than the correct counterpart and holds a positive sign and has a higher level of statistical significance. perhaps the most revealing result of the estimations in table 5 is the support to our third hypothesis that incorrect answering and dnk answering have different implications on credit card repayment patterns. that is, the common practice in previous research to cluster any answering that is not correct as incorrect, seems to be not an appropriate way to understand the effects of different levels (a and b) of financial illiteracy. in summary, our findings confirm hypothesis 1: higher levels of financial literacy (measured as higher number of correct answers) lead to better repayment rates in credit card debt but also hypothesis 3 individuals with self-assessed lower level of financial literacy (dkn answers) display diverse behaviors (credit card repayment patterns) as opposed to individuals who do not possess basic financial literacy (wrong answers). our findings also confirm hypothesis 2 and to a lesser extent hypothesis 4. first, we observe that the variable fepar, being the main source of financial education, is highly significant in all models and negatively associated with repayment habits. this result indicates that individuals that receive their education from parents are more likely to have better repayment behavior on her or his credit card.9 this result also confirms the importance of parental involvement in kids overall financial education aspects. our findings confirm hypothesis 2: the higher the transgenerational transmission of financial education from parents to children results in improved repayment in credit cards. with respect to personality traits also control for the presence of personality trait (surprised, anxiety, and overspending) our estimates highlight that individuals’ capability to repay their credit cards is driven by the use factor and only marginally by the knowledge component. in all the models, the role of overspending impacts negatively the capability to repay credit card debt. the other variables, anxiety and surprised, are not statistically significant. as a consequence, our hypothesis 4 is partially confirmed, at least with respect to present-bias preferences. recall that overspending is the response to individuals actually buying when they knew a priori that he or she would not have enough money to repay the credit card at the end of the billing cycle. we argue, thus, that the overspending coefficient serves as a measure of present bias behavior as well as an indicator of lack of self-control. this lack of self-control becomes a materialized purchase due to the availability of credit cards. if the individual were not to have a credit card, he or she could not complete the purchase. while this may seem obvious, it is relevant to point out that this behavior clearly results in increased financial costs of the purchases, as individuals know ahead of time that they will incur in a rolling debt. notice also that the estimated coefficient value for overspending is consistent across all alternative models, even when controlling for financial (il)literacy. finally, the ncc yields the expected positive sign indicating that individuals having more credit cards are more likely to hold a month-to-month balance: in our estimations an increase in the number of credit cards made the capacity of repayment worse. as hypothesized earlier, a higher number of credit cards may be the result of maxing out of credit on one card and consequently apply for more; or using new credit cards to transfer balances (at promotional interest rates). g. barboza et al. / financial services review 29 (2021) 247–276 263 due to the possible gains in understanding individuals’ behavior on credit card repayment patterns we proceed with a further decomposition of the financial (il)literacy by question. first, notice that under the correct model (1.2), all five questions have the expected negative sign (answering correctly results in better credit card repayment patterns) even though none individually is statistically significant. in contrast, model (2.2), yields opposite signs in all five questions in relation to the correct model version. these differences in results are consistent with expectations, and apparently strong enough to explain the overall positive sign of the total number of incorrect answers. lastly, the dnk model (3.2) provides results that are more congruent with the correct estimations than the incorrect estimations; thus providing further evidence in favor of the hypothesis that clustering incorrect with do not know is not an appropriate way to understand different levels of financial illiteracy. however, given the results from the alternative estimations in table 5, we confirm that the role of financial literacy, or lack of, appears not to dominate behavioral/personal traits (hypothesis 5). this is to say, that financial decisions seem to be made primarily along pattern of behavioral traits and financial education from parents; and in this context only marginally ameliorated by financial education. assuming that these results are consistent and robust (after accounting for possible sample issues) then, they indicate that while financial literacy is an important or fundamental element for all individuals to learn, behavioral variables drive the accumulation of credit card debt and result in less than optimal repayment. in this respect, we confirm the findings of xiao et al. (2011) who, applying the theory of planned behavior to investigate risky credit behavior among college students, found that behavioral intentions were the single most important factor in whether students make responsible credit decisions (risky borrowing behavior, risky paying behavior, and holding credit card debt). under these conditions, it then becomes relevant to explore the drivers of overspending. given the robustness of the personal trait variables, in conjunction with the lack of strong statistical significance of the financial literacy estimates, several possible scenarios come to mind. for instance, we can argue as mentioned above that behavioral traits are just too strong and clearly individuals have a hard time controlling them. particularly, issues relating to present-bias, preferences and gaps in mental accounting are strong and present. in addition, one can argue that this behavior may be prevented or ameliorated with early intervention in the form of exposure to financial literacy. that is, there is the possibility that students in our sample may be receiving financial education too late in life, and thus personality traits are harder to counterbalance. this interpretation is compliant also with the relevance of financial education transmitted by parent in ameliorating the credit card repayment. 4.1. interaction effects the results thus far have provided very useful information in the advancing the understanding of the relationship between financial (il)literacy and personality traits into credit card repayment decisions. however, the relationships may also be shaped by the presence of interaction effects deriving from financial (il)literacy and the behavioral or personality variables. as noted, we uncover that personality traits are dominant in the decision individuals 264 g. barboza et al. / financial services review 29 (2021) 247–276 make, particularly overspending (+) and financial education deriving from parents (�). in this context, and given the strong interest in the effects that financial (il)literacy may have on financial decision making, we proceed to compute and report the alternative models including interaction effects between financial (il)literacy and financial education from parent and all four behavioral variables. the results are presented in tables 6, 6a, and 6b, for correct, incorrect (level a), and dnk (level b), respectively. the first set of estimations in table 6 corresponds to the financial literacy-correct answering sample. in these estimations, we observe that all previous results hold, overspending having a negative effect and worsening repayment, increased the number of credit cards also worsening repayment. fepar has a significant effect improving credit card repayment when the level of financial literacy is low whereas higher financial literacy resulting in better repayment patterns and lower chance of accumulation month-to-month debt when students did not receive financial education from their parents. in addition, the interaction effect of financial literacy and fepar has a positive and statistically significant coefficient: when students receive education from parents the positive effect of financial literacy disappears. notice also that the akaike information criterion (aic) for the interaction model is lower than the corresponding model in table 5 without the interaction. this is to say that the interaction model specification is preferred. in addition, none of the other interaction effects prove to be statistically significant, and therefore the models are inferior. for the second set of interaction models, reported in table 6a, for the level of financial illiteracy type a, we also observe results that are in accordance to the estimations reported in table 6. however, under the interaction effect model 1, the level of fepar is now not statistically significant, despite holding the expected sign. when the financial illiteracy is low the role of financial education from parents is not significant in improving the credit card repayment. the increase in the number of total incorrect answers results in credit card repayment worsening. when we interact financial illiteracy and education from parent the coefficient is negative and statistically significant the 5% level of confidence. our intuitive interpretation of this effect is that for student characterized by high level of financial illiteracy the education received from parents is fundamental in improving the credit card repayment. notice that as it was the case for the financial literacy models, all other interaction effects are not statistically significant and the aic values are also higher. we argue that the model 1 is the preferred model. for the third estimations, where financial illiteracy is measured through dnk answers (table 6b), we observe that the interaction effects model for fepar and financial illiteracy level b is significant and holds a negative sign. our intuitive rationale for the coefficient in this model is that students that mainly answered dnk are able to improve their credit card repayment habits if they received proper financial education from parents. in addition, it is relevant to point out that the financial illiteracy coefficient is not statistically significant (as it was in model 3.1–table 5), but it now present a reversal of sign. in addition, the aic for model 1–table 6b is lower than the corresponding model in table 5, without interaction effects. thus, we preferred the interaction model specification in this regard. as it was with other models, interaction effects for all other variables are not statistically significant. our findings do not confirm completely hypothesis 5. we do not detect that positive effects from higher levels of financial literacy could be overpowered by personal traits. furthermore, g. barboza et al. / financial services review 29 (2021) 247–276 265 t ab le 6 c o n su m er b eh av io r w it h cr ed it ca rd re p ay m en t fr eq u en cy as d ep en d en t v ar ia b le (c o rr ec t an sw er s) c o rr ec t an sw er s m o d el s w it h in te ra ct io n ef fe ct s 1 2 3 4 5 f in an ci al ed u c p ar en ts �2 .1 4 5 (0 .0 0 1 )* * * �0 .9 3 4 (0 .0 0 1 )* * * �0 .9 3 6 (0 .0 0 1 )* * * �0 .9 2 7 (0 .0 0 1 )* * * �0 .9 3 9 (0 .0 0 1 )* * * a n x ie ty 0 .1 2 5 (0 .2 4 ) 0 .2 3 8 (0 .4 0 ) 0 .1 4 0 (0 .1 8 ) 0 .1 4 4 (0 .1 7 ) 0 .1 4 4 (0 .1 7 ) s u rp ri se d 0 .1 2 0 (0 .3 7 ) 0 .1 3 5 (0 .3 1 ) 0 .1 4 7 (0 .2 7 ) 0 .3 9 5 (0 .1 3 ) 0 .1 4 4 (0 .2 8 ) o v er s p en d in g 0 .5 7 7 (0 .0 0 1 )* * * 0 .5 3 3 (0 .0 0 1 )* * * 0 .4 3 9 (0 .1 0 )* 0 .5 2 8 (0 .0 0 1 )* * * 0 .5 3 0 (0 .0 0 1 )* * * n c c 0 .1 7 1 (0 .0 6 )* 0 .1 6 2 (0 .0 8 )* 0 .1 5 5 (0 .0 9 )* 0 .1 6 7 (0 .0 7 )* 0 .0 8 1 (0 .7 5 ) f in an ci al li te ra cy q u es ti o n s t o ta l an sw er s �0 .4 1 0 (0 .0 0 1 )* * * �0 .0 8 6 (0 .6 7 ) �0 .1 7 4 (0 .1 1 ) �0 .0 4 2 (0 .7 5 ) �0 .1 9 1 (0 .2 1 ) in te ra ct io n ef fe ct s o f fi n an ci al li te ra cy w it h p ar en ts fi n an ci al ed u ca ti o n 0 .4 2 8 (0 .0 2 )* * a n x ie ty �0 .0 3 2 (0 .7 2 ) o v er sp en d in g 0 .0 3 5 (0 .6 9 ) s u rp ri se d �0 .0 9 7 (0 .2 6 ) n c c 0 .0 2 6 (0 .7 3 ) o b s 1 5 8 1 5 8 1 5 8 1 5 8 1 5 8 p se u d o r 2 0 .2 2 7 0 .2 0 9 0 .2 0 9 0 .2 1 3 0 .2 0 9 l r st at is ti c 7 0 .6 9 2 6 5 .1 9 2 6 5 .2 2 3 6 6 .3 4 6 6 5 .1 7 5 p ro b ab il it y (l r st at ) (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * a k ai k e in fo cr it er io n 1 .6 5 1 1 .6 8 5 1 .6 8 5 1 .6 7 8 1 .6 8 6 s ch w ar z cr it er io n 1 .8 4 4 1 .8 7 9 1 .8 7 9 1 .8 7 2 1 .8 7 9 h an n an -q u in n cr it er io n 1 .7 2 9 1 .7 6 4 1 .7 6 4 1 .7 5 7 1 .7 6 4 n o te : * * * , * * , * st at is ti ca ll y si g n ifi ca n t at th e 1 % , 5 % an d 1 0 % le v el , re sp ec ti v el y . 266 g. barboza et al. / financial services review 29 (2021) 247–276 t ab le 6 a c o n su m er b eh av io r w it h cr ed it ca rd re p ay m en t fr eq u en cy as d ep en d en t v ar ia b le (i n co rr ec t an sw er s) f in an ci al il li te ra cy l ev el a in co rr ec t an sw er s m o d el s w it h in te ra ct io n ef fe ct s 1 2 3 4 5 f in an ci al ed u c p ar en ts �0 .2 8 3 (0 .4 5 ) �0 .9 4 0 (0 .0 0 1 )* * * �0 .9 2 2 (0 .0 0 1 )* * * �0 .9 4 2 (0 .0 0 1 )* * * �0 .9 1 9 (0 .0 0 1 )* * * a n x ie ty 0 .1 3 2 (0 .2 0 ) 0 .2 2 2 (0 .1 3 ) 0 .1 4 9 (0 .1 5 ) 0 .1 6 3 (0 .1 1 ) 0 .1 6 0 (0 .1 3 ) s u rp ri se d 0 .1 4 2 (0 .2 5 ) 0 .1 4 9 (0 .2 3 ) 0 .1 4 4 (0 .2 4 ) �0 .0 5 1 (0 .7 9 ) 0 .1 4 6 (0 .2 3 ) o v er s p en d in g 0 .5 3 7 (0 .0 0 1 )* * * 0 .5 0 6 (0 .0 0 1 )* * * 0 .6 1 2 (0 .0 0 1 )* * * 0 .4 9 8 (0 .0 0 1 )* * * 0 .4 8 8 (0 .0 0 1 )* * * n c c 0 .1 8 5 (0 .0 2 )* * 0 .1 6 4 (0 .0 4 )* * 0 .1 5 4 (0 .0 5 )* * 0 .1 6 7 (0 .0 4 )* * 0 .2 7 1 (0 .0 1 )* * * f in an ci al li te ra cy q u es ti o n s t o ta l an sw er s 0 .5 3 0 (0 .0 0 1 )* * * 0 .3 4 8 (0 .2 1 ) 0 .2 8 0 (0 .0 4 )* * 0 .0 7 6 (0 .6 3 ) 0 .3 6 8 (0 .0 3 )* * in te ra ct io n ef fe ct s o f fi n an ci al li te ra cy w it h p ar en ts fi n an ci al ed u ca ti o n �0 .4 3 4 (0 .0 5 )* * a n x ie ty �0 .0 5 4 (0 .6 0 ) o v er sp en d in g �0 .0 6 1 (0 .4 5 ) s u rp ri se d 0 .1 2 3 (0 .2 2 ) n c c �0 .0 8 0 (0 .1 6 ) o b s 1 5 6 1 5 6 1 5 6 1 5 6 1 5 6 p se u d o r 2 0 .2 2 3 0 .2 1 3 0 .2 1 3 0 .2 1 6 0 .2 1 5 l r st at is ti c 6 9 .1 7 2 6 5 .8 7 0 6 6 .0 5 0 6 6 .9 8 1 6 6 .5 9 0 p ro b ab il it y (l r st at ) (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * a k ai k e in fo cr it er io n 1 .6 7 1 1 .6 9 2 1 .6 9 1 1 .6 8 5 1 .6 8 8 s ch w ar z cr it er io n 1 .8 6 7 1 .8 8 8 1 .8 8 7 1 .8 8 1 1 .8 8 3 h an n an -q u in n cr it er io n 1 .7 5 1 1 .7 7 2 1 .7 7 1 1 .7 6 5 1 .7 6 7 n o te : * * * , * * , * st at is ti ca ll y si g n ifi ca n t at th e 1 % , 5 % , an d 1 0 % le v el , re sp ec ti v el y . g. barboza et al. / financial services review 29 (2021) 247–276 267 t ab le 6 b c o n su m er b eh av io r w it h cr ed it ca rd re p ay m en t fr eq u en cy as d ep en d en t v ar ia b le (d o n o t k n o w an sw er s) f in an ci al il li te ra cy l ev el b d o n o t k n o w an sw er s m o d el s w it h in te ra ct io n ef fe ct s 1 2 3 4 5 f in an ci al ed u c p ar en ts �0 .5 6 0 (0 .0 5 )* * �0 .9 1 4 (0 .0 0 1 )* * * �0 .9 1 6 (0 .0 0 1 )* * * �0 .9 1 3 (0 .0 0 1 )* * * �0 .9 0 8 (0 .0 0 1 )* * * a n x ie ty 0 .1 8 3 (0 .0 9 )* 0 .1 3 0 (0 .3 7 ) 0 .1 6 7 (0 .1 2 ) 0 .1 6 3 (0 .1 2 ) 0 .1 6 5 (0 .1 2 ) s u rp ri se d 0 .1 4 7 (0 .2 7 ) 0 .1 8 0 (0 .1 6 ) 0 .1 8 5 (0 .1 5 ) 0 .1 8 4 (0 .2 8 ) 0 .1 8 4 (0 .1 5 ) o v er s p en d in g 0 .5 1 9 (0 .0 0 1 )* * * 0 .5 1 0 (0 .0 0 1 )* * * 0 .5 5 1 (0 .0 0 1 )* * * 0 .5 0 9 (0 .0 0 1 )* * * 0 .5 0 4 (0 .0 0 1 )* * * n c c 0 .1 3 0 (0 .1 5 ) 0 .1 3 6 (0 .1 3 ) 0 .1 3 8 (0 .1 3 ) 0 .1 3 4 (0 .1 4 ) 0 .1 0 8 (0 .3 4 ) f in an ci al li te ra cy q u es ti o n s t o ta l an sw er s 0 .2 7 5 (0 .2 0 ) �0 .1 0 6 (0 .7 1 ) 0 .0 2 3 (0 .8 9 ) �0 .0 1 8 (0 .9 3 ) �0 .0 9 8 (0 .6 9 ) in te ra ct io n ef fe ct s o f fi n an ci al li te ra cy w it h p ar en ts fi n an ci al ed u ca ti o n �0 .5 4 8 (0 .0 7 )* a n x ie ty 0 .0 4 1 (0 .7 4 ) o v er sp en d in g �0 .0 7 3 (0 .6 1 ) s u rp ri se d �0 .0 0 3 (0 .9 8 ) n c c 0 .0 4 6 (0 .7 0 ) o b s 1 5 6 1 5 6 1 5 6 1 5 6 1 5 6 p se u d o r 2 0 .2 0 9 0 .1 9 8 0 .1 9 8 0 .1 9 7 0 .1 9 8 l r st at is ti c 6 4 .6 7 4 6 1 .2 1 4 6 1 .3 5 9 6 1 .1 0 1 1 .7 2 2 p ro b ab il it y (l r st at ) (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * (0 .0 0 1 )* * * a k ai k e in fo cr it er io n 1 .7 0 0 1 .7 2 2 1 .7 2 1 1 .7 2 3 1 .7 2 2 s ch w ar z cr it er io n 1 .8 9 6 1 .9 1 8 1 .9 1 7 1 .9 1 8 1 .9 1 7 h an n an -q u in n cr it er io n 1 .7 7 9 1 .8 0 2 1 .8 0 1 1 .8 0 2 1 .8 0 1 n o te : * * * , * * , * st at is ti ca ll y si g n ifi ca n t at th e 1 % , 5 % , an d 1 0 % le v el , re sp ec ti v el y . 268 g. barboza et al. / financial services review 29 (2021) 247–276 financial knowledge derived from parents has a strong positive effect on individuals’ financial behavior especially for people characterized by a relevant financial illiteracy. based on the findings from the interaction effects models, and the corresponding superiority of models 1, in tables 6, 6a, and 6b, we decided to use these series of models to compute the marginal effects and corresponding probabilities. we present the overall probabilities for each model in table 7, followed by the computation of the marginal changes in the overall probabilities when the independent variables change by one unit. it is also relevant to note that the estimated probabilities are computed at the mean value of the independent variables. for the especial case of the fepar variable (dummy 1,0) we compute the marginal effects by taking the difference between the overall probabilities when fepar = 1, minus fepar = 0. there are several interesting results that spring out the analysis of table 7. for instance, the first element to notice is that all probabilities are very similar across alternative models of financial (il)literacy with the probability to paying in full being in the range of 67.27 (dnk answers – level b) to 68.82% (incorrect answers – level a), with 67.55% (correct answers). and the rest of probabilities being almost identical for the other ranges across models. in other words, when evaluated at the mean values, we observe very little discrepancies across alternative models, when the only difference is based on the level of financial (il) literacy. this result on itself is surprising, as one would expect that different levels of financial (il)literacy would yield much larger differences in repayment patterns. secondly, when we look at the changes in probabilities due to variations in rhs variables, we now observe potentially large differences across models. the first effect to study here is that of fepar. notice that the effect of this dummy variable has a significant marginal effect on the probability of repaying in full and has the largest effect of all variables. in this sense, we also observe that those suffering the most from level b financial illiteracy are the group that would benefit the most with an increase learning deriving from their parents. the counter result indicates that those suffering from level a financial illiteracy would still benefit from further interaction with their parents as the main source of financial literacy but with the smallest effect among the three possible groups. in addition, the counter result of variations in the fepar variable indicate that both level b financial illiteracy and those answering correctly would be the groups that would suffer the most in their repayment capabilities should fepar were not to be the main source of financial literacy as reflected by the negative sign of the probability changes for the remaining categories. the fourth element that we pay attention in this analysis is the effect of changes in the level of overspending. in this case, we observe that changes in overspending have the second largest effect of all variables. more specifically, an increase of one unit in overspending (following the stated categories in the descriptive statistics) have a negative effect in repayment capabilities of about 20% decrease in repayment in full; and incidentally an increase in all other repayment categories with paying less than full balance every month being the most affected. in this sense, a marginal change in overspending patterns, results in a large increase in month-to-month debt accumulation and falling behind in repayment patterns. furthermore, this negative effect is larger than an increase in financial literacy as we will discuss next. financial literacy has been championed as the main way to improve financial decisionmaking process, and specifically research studying credit card debt emphasize on its g. barboza et al. / financial services review 29 (2021) 247–276 269 t ab le 7 p ro b ab il it ie s an d m ar g in al ef fe ct s, fo r m o d el w it h in te ra ct io n ef fe ct s p ro b (y = 1 jx) /∂ p 1 /∂ x p ro b (y = 2 jx) /∂ p 2 /∂ x p ro b (y = 3 jx) /∂ p 3 /∂ x p ro b (y = 4 jx) /∂ p 4 /∂ x f in an ci al li te ra cy m o d el 6 7 .5 5 % 1 9 .6 7 % 7 .6 9 % 5 .0 9 % p ar en ts fi n an ci al ed u c 0 .3 1 3 �0 .1 1 6 �0 .0 9 1 �0 .1 0 6 o v er sp en d in g �0 .2 0 7 0 .0 8 7 0 .0 6 0 0 .0 6 0 n c c �0 .0 6 1 0 .0 2 6 0 .0 1 8 0 .0 1 8 c o rr ec t an sw er s 0 .1 4 7 �0 .0 6 2 �0 .0 4 3 �0 .0 4 3 in te ra ct io n ef fe ct �0 .1 5 4 0 .0 6 4 0 .0 4 5 0 .0 4 5 l ev el a f in an ci al il li te ra cy 6 8 .8 2 % 1 9 .1 4 % 7 .3 2 % 4 .7 2 % p ar en ts fi n an ci al ed u c 0 .2 9 5 �0 .0 5 5 �0 .0 8 5 �0 .0 9 5 o v er sp en d in g �0 .1 9 0 0 .0 8 2 0 .0 5 5 0 .0 5 3 n c c �0 .0 6 5 0 .0 2 8 0 .0 1 9 0 .0 1 8 in co rr ec t an sw er s �0 .1 8 7 0 .0 8 1 0 .0 5 4 0 .0 5 2 in te ra ct io n ef fe ct 0 .1 5 4 �0 .0 6 7 �0 .0 4 4 �0 .0 4 3 l ev el b f in an ci al il li te ra cy 6 7 .2 7 % 1 9 .2 5 % 7 .8 4 % 5 .6 4 % p ar en ts fi n an ci al ed u c 0 .3 4 8 �0 .1 2 0 �0 .0 9 8 �0 .1 3 0 o v er sp en d in g �0 .1 8 7 0 .0 7 5 0 .0 5 4 0 .0 5 9 n c c �0 .0 4 7 0 .0 1 9 0 .0 1 3 0 .0 1 5 d o n o t k n o w an sw er s �0 .0 9 9 0 .0 4 0 0 .0 2 8 0 .0 3 1 in te ra ct io n ef fe ct 0 .1 9 8 �0 .0 7 9 �0 .0 5 7 �0 .0 6 2 s o u rc e: m ar g in al ef fe ct s o n ly re p o rt ed fo r m o d el 1 t ab le 6 , 6 a, an d 6 b . y = 1 re fe rs to p ay in fu ll , y = 2 re fe rs to p ay le ss th an fu ll b u t m o re th an m in im u m , y = 3 re fe rs to p ay m in im u m , an d y = 4 re fe rs to p ay le ss th an m in im u m . 270 g. barboza et al. / financial services review 29 (2021) 247–276 importance. while we do not dispute the inner importance of financial literacy in improving decision making, our results indicate that financial literacy plays a secondary role, when controlling for individuals’ personality traits and consequent behavior. in particular, marginal effects analysis from table 5, indicate that an increase in the number of correct answer in the financial literacy questionnaire result in a positive improvement (pay in full) in repayment patterns on the approximate amount of 14.7%. while a significant improvement, this amount is not enough to counterbalance the negative effects of behavior. by the same token, the negative effect of increased financial illiteracy is decomposed in a worsen repayment capability (away from full repayment every month) of 18.7% for level a and 9.9% for level b. it is relevant to point out that those overestimating their financial knowledge, yet answering incorrectly, are at the highest risk of falling behind and carry a month-to-month balance. in addition, it comes as a relative surprise that those suffering from level b illiteracy are less likely to carry an increase negative probability of repayment in full when answering more dnk to the financial literacy questions. in this context, it becomes evidence that individuals with a lack of financial literacy level a and personality traits dominated by overspending are highly more likely to fall behind, as reflected by the estimated values of the prob y = 2, 3, and 4 categories. a generalization of the results also indicate that the estimated probabilities and corresponding marginal effects are lower for the level b financial illiteracy individuals than the other two groups. for level b, it appears that they benefit the most from increased interaction with their parents (fepar), and recur less to more credit cards and consequently have a lower negative effect on repayment as the number of credit cards is increased. thus, it appears that not knowing about financial literacy (dnk) makes them act in a more cautious way in relation to credit card use. furthermore, the robustness of the personality traits manifestation in the form of overspending are statistically consistent across all estimated models. in this context, it is more relevant to note that the estimated marginal effects of changes in this behavioral variable are also consistent even after controlling for the different levels of financial literacy and financial (il)literacy levels a and b. in other words, our estimated probabilities and related marginal changes of increases in overspending seem to be independent of financial literacy. this result, as far as we are aware of, is unique and not present in the extant literature. as such, we argue that financial literacy has a limited impact on modifying personality traits and related behavior. it appears, that early intervention and an increase in financial literacy at early stages in life, primarily through parental education, may have the largest offsetting effects to personality traits leading to poor financial decision making. 5. conclusions the empirical evidence in this paper demonstrates that financial behavior, measured in terms of credit card repayment patterns, is affected more by personal traits than by financial literacy. this is to say, that financial decisions are made mainly based on personal traits or behavioral factors. behavioral variables, such overspending, drive the accumulation of credit card debt and result in less than optimal repayment. financial knowledge derived from g. barboza et al. / financial services review 29 (2021) 247–276 271 parental interaction with children seems to be the form of financial knowledge most relevant in positively influencing credit card repayment. individuals with self-assessed lower level of financial literacy (dnk answers) display diverse behaviors (credit card repayment patterns) as opposed to individuals who do not possess basic financial literacy (wrong answers) even after controlling for cognitive or personal trait factors. in our context, it appears that financial literacy has a limited positive benefit in shaping financial decisions. previous evidence is mixed on this specific issue: although there is ample evidence supporting that higher financial knowledge translate into less risky credit card use (norvilitis et al., 2006; robb, 2011; shim et al., 2009, 2010), other researches have reported greater financial knowledge was associated with lower fear about using credit cards and greater levels of debt (borden et al. 2008; lyons & rogers, 2004; robb & sharpe, 2009). one possible explanation for these mixed results could depend on how financial knowledge is operationalized and measured. for instance, our study levers up the traditional financial literacy questions developed by lusardi & mitchell (2014) while robb (2011) uses a financial knowledge score specifically designed to uncover specific knowledge about credit card use. different results could be related to imprecise measurement of a latent variable such as financial knowledge. results from the parsimonious model specification also indicate the presence of robust behavior patterns along personality traits. present-bias behavior, overconfidence and lack of control seem to be the main drivers of credit card use, and consequently of credit card repayment. now, with the existent data, we can measure the impact of behavioral differences while controlling for financial (il)literacy, but we cannot measure the possible gains in behavioral actions/responses due to increases in financial literacy. it is important to acknowledge this caveat in our study to both avoid incorrect data interpretation, and to set the future research agenda as we move forward. in essence, the long-term goal of the financial education or literacy movement is to provide mechanisms to positively affect economic behavior that currently result in costly and inefficient financial decisions. this is so, as financial behavior is highly determined by preferences, and these preferences may or may not be a function of financial literacy. if financial literacy does affect individuals’ preferences and decision-making processes then one would expect that higher levels of financial literacy result in a series of benefits such as: increased saving, wiser investment decisions, lower to no month-tomonth credit card debt, higher wealth accumulation, and higher retirement savings. our findings are useful for policy makers to implement policies able to avoid, or reduce, deb trap and socio-economic vulnerability of the borrower. for instance, it appears that early intervention in terms of financial education may provide the necessary means to control endogenous costly personality traits. once the personality trait has fully developed alternative means of positively impacting financial decision making might need to be implemented. as xiao et al. (2011) highlight, financial education programs should target the multiple psychological processes that lead to changes in attitude and the enhancement of self-confidence, that is, they should target and develop students’ positive financial intentions. finally, while being outside the scope of this study, further research on the subject of financial anxiety and its relation to credit card behavior will need to be explored, particularly the determination of the causality between credit card repayment habit and financial anxiety. 272 g. barboza et al. / financial services review 29 (2021) 247–276 the tentative hypothesis states that higher financial anxiety should result in worse credit card repayment; thus, future research could analyze in more depth the role of psychological aspects and personality traits, in particular financial anxiety, in the process of financial decision-making, that is the possibility and cause-effect that higher credit card balance may fuel anxiety and result in a freeze effect where individuals may not improve repayment. notes 1 in several studies, most individuals report having higher believed level of financial literacy than actual correct scoring on financial literacy questionnaire/surveys. see recent studies by gflec at george washington university center for some examples (almenberg et al., 2016). 2 in a recent papers kim and mountain (2019) suggest a specific statistical approach (binomial-latent regression models) to specifically tackle the issue of group differences that are hidden in dk/rf responses. 3 aptitude is different from attitude: while attitude is a way of looking at an issue or an object, a mental position or way of thinking about an issue (in our case financial matters), the concept of aptitude is akin to natural or acquired talent or ability, inclination, predisposition. in this sense, aptitude for financial matters could be learn through financial education as reflected by higher levels of financial literacy. 4 the questionnaire has been successfully used to capture the financial literacy of diverse populations since it was first piloted in 2010 as part of the first oecd international financial literacy and financial inclusion measurement exercise. in 2018, an updated version was released that takes into consideration the changes in the financial landscape and the evolving state of knowledge; therefore, including questions related to digital financial services and crypto-assets, trust, integrity and financial consumer protection. 5 more recent studies have applied factor analysis (van rooij, lusardi, & alessie, 2011). it is widely acknowledged, however, that more work, developed through rigorous psychometric analysis is needed (bongini et al., 2015, 2016; knoll & houts, 2012). 6 for a comprehensive survey on personality psychology and economics the interested reader can refer to almlund, duckworth, heckman, and kautz (2011) and to brown and taylor (2014) for a specific application that analyzes the relationship between personality traits and financial decision-making focusing on unsecured debt and financial assets. 7 while the topics of “time value of money,” “inflation,” and “interest rate compounding” do represent basics knowledge to make informed choices when deciding to pay in full or accumulating a month-to-month balance, the issue of risk diversification is less strictly correlated. 8 estimations for the ncc variables as the dependent variable, with demographic characteristics as rhs variables, are available from the authors upon request. we conducted the same estimations with demographics instead of ncc, and obtained g. barboza et al. / financial services review 29 (2021) 247–276 273 very similar results, without affecting the estimated coefficients of the other variables. 9 recall that credit card repayment behavior is measured in a reverse scale, where the lowest value implies repayment in full, and higher values otherwise. references adams, t., & moore, m. (2007). high-risk health and credit behavior among 18-to 25-year-old college students. journal of american college health, 56, 101–108. akerlof, g. (1991). procrastination and obedience. the american economic review, 81(2), 1–19. andrews, b., & wilding, j. m. (2004). the relation of depression and anxiety to life-stress and achievement in students. british journal of psychology, 95, 509–521. barboza, g. (2018). i will pay tomorrow, or maybe the day after. credit card repayment, present biased and procrastination. economic notes, 47, 455–494. barboza, g., smith, c., & pesek, j. (2016). impact of gender and academic performance differences on financial literacy. journal of financial education,winter, 205–242. berg, c., sanem, j., lust, k., ahluwalia, j., kirch, m., & an, l. (2010). health-related characteristics and incurring credit card debt as problem behaviors among college students. the internet journal of mental health, 6, 1. bongini, p., trivellato, p., & zenga, m. (2015). business students and financial literacy: when will the gender gap fade away? journal of financial management, markets and institutions, 3, 13–30. bongini, p., trivellato, p., & zenga, m. (2016). financial literacy and undergraduates. application of latent regression rasch model. journal of financial management, markets and institutions, 41, 23–42. borden, l. m., le, s., serido, j., & collins, d. (2008). changing college students’ financial knowledge, attitudes, and behavior through seminar participation. journal of family and economic issues, 29, 23–40. brown, s., & taylor, k. (2014). household finances and the ‘big five’ personality traits. journal of economic psychology, 45, 197–212. cheema, a., & soman, d. (2006). malleable mental accounting: the effect of flexibility on the justification of attractive spending and consumption decisions. journal of consumer psychology, 16, 33–44. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11(3), 289–307. chen, z., & garand, j. c. (2018). on the gender gap in financial knowledge: decomposing the effects of don’t know and incorrect responses. social science quarterly, 99(5), 1551–1571. chu, c., buchman-schmitt, j. m., stanley, i. h., hom, m. a., tucker, r. p., hagan, c. r., & joiner jr, t. e. (2017). the interpersonal theory of suicide: a systematic review and meta-analysis of a decade of crossnational research. psychological bulletin, 143(12), 1313. danes, s. m., & tahira, k. h. (1987). money management knowledge of college students. journal of student financial aid, 17, 4–16. fiksenbaum, l., marjanovic, z., & greenglass, e. (2017). financial threat and individuals’ willingness to change financial behavior. review of behavioral finance, 9, 128–147. fox, m. d., snyder, a. z., vincent, j. l., corbetta, m., van essen, d. c., & raichle, m. e. (2005). the human brain is intrinsically organized into dynamic, anticorrelated functional networks. proceedings of the national academy of sciences of the united states of america, 102, 9673–9678. grable, j., & joo, s. (2006). student racial differences in credit card debt and financial behaviors stress. college student journal, 40, 400–408. retrieved from http://www.eric.ed.gov/ericwebportal/detail?accno= ej765337 hancock, a. m., jorgensen, b. l., & swanson, m. s. (2013). college students and credit card use: the role of parents, work experience, financial knowledge, and credit card attitudes. journal of family and economic issues, 34, 369–381. 274 g. barboza et al. / financial services review 29 (2021) 247–276 hilgert, m., hogarth, j., & beverly, s. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309–322. houston, s. j. (2010). measuring financial literacy. the journal of consumer affairs, 44, 296–316. joo, s., & garman, t. e. (1998). the potential effects of workplace financial education based on the relationship between personal financial wellness and worker job productivity. personal finances and worker productivity, 2, 163–173. kapoor, a., dunn, e., kostaki, a., andrews, m. h., & matthews, s. g. (2006). fetal programming of hypothalamo-pituitary-adrenal function: prenatal stress and glucocorticoids. the journal of physiology, 572, 31–44. kempson, e. (2009). framework for the development of financial literacy baseline surveys. retrieved from oecd-library.org kezar, a., & yang, h. (2010). the importance of financial literacy. about campus: enriching the student learning experience, 14, 15–21. kim, n., & mountain, t. p. (2019). financial knowledge and “don’t know” response. journal of consumer affairs, 53(4), 1948–1969. knoll, m. a. z., & houts, c. r. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. journal of consumer affairs, 46, 381–410. koshal, r. k., gupta, a. k., goyal, a., & choudhary, v. n. (2008). assessing economic literacy of indian mba students. american journal of business, 23, 43–52. kuchler, t. (2013). sticking to your plan: hyperbolic discounting and credit card debt paydown. discussion papers 12-025. stanford institute for economic policy research. laibson, d. (1997). golden eggs and hyperbolic discounting. the quarterly journal of economics, 112, 443–477. lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the role of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic perspectives, 52, 5–44. lusardi, a., mitchell, o. s., & curto, v. (2009). financial literacy among the young: evidence and implications for consumer policy. cerp working papers 91. turin, italy: center for research on pensions and welfare policies. lusardi, a., mitchell, s., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–381. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and overindebtedness. nber working papers 14808. national bureau of economic research, inc. lyons, j. s., & rogers, l. (2004). the us child welfare system: a de facto public behavioral health care system. journal of the american academy of child & adolescent psychiatry, 43, 971–973. mandell, l. (1997). personal financial survey of high school seniors. washington, dc: jump start coalition for personal financial literacy. mandell, l. w. (2008). the financial literacy of young american adults: results of the 2008 national jump$tart coalition survey of high school seniors and college students. washington, dc: jump start coalition for personal financial literacy. manton, e. j., english, d. e., avard, s., & walker, j. (2006). what college freshmen admit to not knowing about personal finance. journal of college. teaching &. learning, 3, 43–54. marjanovic, z., greenglass, e. r., fiksenbaum, l., & bell, c. m. (2013). psychometric evaluation of the financial threat scale (fts) in the context of the great recession. journal of economic psychology, 36, 1–10. markovich, c. a., & devaney, s. (1997). college seniors’ personal finance knowledge and practices. journal of family and consumer sciences, 89, 61–65. moore, d. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences. olympia, wa: washington state department of financial institutions. moore, d. a., & healy, p. j. (2008). the trouble with overconfidence. psychological review, 115(3), 502. nelson, m., lust, k., story, mf., & ehlinger, e. (2008). credit card debts, stress and key health risk behaviors among college students. american journal of health promotion, 22, 400–412. g. barboza et al. / financial services review 29 (2021) 247–276 275 norvilitis, j. m., merwin, m. m., osberg, t. m., roehling, p. v., young, p., & kamas, m. m. (2006). personality factors, money attitudes, financial knowledge, and credit-card debt in college students 1. journal of applied social psychology, 36, 1395–1413. o’donoghue, t., & rabin, m. (1999). doing it now or later. american economic review, 89, 103–124. peters, m., seeds, k., goldstein, a., & coleman, n. (2007). parental involvement in children’s education 2007, dcsf research report rr034. london: department for children, schools and families. retrieved from www.dcsf.gov.uk/research/data/uploadfiles/dcsf-rr034.pdf pinto, m., parente, d., & palmer, t. (2001). college student performance and credit card usage. journal of college student development, 42, 49–58. retrieved from http://www.eric.ed.gov/ericwebportal/detail? accno=ej624498 porto, n., & xiao, j. j. (2016). financial literacy overconfidence and financial advice seeking. journal of financial service professionals, 70(4). prelec, d., & simester, d. (2001). always leave home without it: a further investigation of the credit card effect on willingness to pay.marketing letters, 12, 5–12. razafimahasolo, m., borromeo, r., sausa, l., carpizo, r., & sabado, j. (2016). impact of financial literacy on level of stress and academic achievement among college students. aup research journal, 19(2), 21–35. remund, d. l. (2010). financial literacy explication: the case for a clearer definition in an increasingly complex economy. the journal of consumer affairs, 44, 276–295. robb, c. a. (2011). financial knowledge and credit card behavior of college students. journal of family and economic issues, 32, 690–698. robb, c. a., & sharpe, d. l. (2009). effect of personal financial knowledge on college students’ credit card behavior. journal of financial counseling and planning, 20(1). schagen, s., & lines, a. (1996). financial literacy in adult life. london: national foundation for educational research in england and wales. natwest group charitable trust. shim, s., xiao, j. j., barber, b. l., & lyons, a. c. (2009). pathways to life success: a conceptual model of financial well-being for young adults. journal of applied developmental psychology, 30, 708–723. shim, s., barber, b. l., card, n. a., xiao, j. j., & serido, j. (2010). financial socialization of first-year college students: the roles of parents, work, and education. journal of youth and jadolescence, 39, 1457–1470. stango, v., & zinman, j. (2009). what do consumers really pay on their checking and credit card accounts? explicit, implicit, and avoidable costs. american economic review, 99, 424–429. thaler, r. (1980). toward a positive theory of consumer choice. journal of economic behavior & organization, 1, 39–60. thaler, r. (2018). from cashews to nudges: the evolution of behavioral economics. american economic review, 108, 1265–1287. thaler, r., & shefrin, h. m. (1981). an economic theory of self-control. journal of political economy, 89, 392–406. tokar asaad, c. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24(2). van rooij, m. c., lusardi, a., & alessie, r. j. (2012). financial literacy, retirement planning and household wealth. the economic journal, 122(560), 449–478. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. wagland, s. p., & taylor, s. t. (2009). when it comes to financial literacy, is gender really an issue? the australasian accounting business and finance journal, 3, 13–23. xia, t., wang, z., & li, k. (2014). financial literacy overconfidence and stock market participation. social indicators research, 119(3), 1233–1245. xiao, j. j., tang, c., serido, j., & shim, s. (2011). antecedents and consequences of risky credit behavior among college students: application and extension of the theory of planned behavior. journal of public policy & marketing, 30, 239–245. zhang, c. y., & sussman, a. b. (2018). perspectives on mental accounting: an exploration of budgeting and investing. financial planning review, 1, e1011. 276 g. barboza et al. / financial services review 29 (2021) 247–276 mutual fund knowledge assessment for policy and decision problems brian scholl1, angela fontes2,* 1office of the investor advocate, 100 f street ne, washington, dc 20549, usa 2norc at the university of chicago, 55 east monroe, chicago, il 60602, usa abstract we develop a measure of mutual fund investment knowledge that complements existing financial literacy measures. our question battery was administered to 3,444 survey respondents. we validate the index with factor analysis identifying two latent components, and descriptive regressions demonstrating the additive value of our index beyond general financial literacy in explaining variation in financial well-being, investment ownership, and fee calculation proficiency. despite mutual funds’ importance in household savings, our index suggests that the public lacks adequate understanding of them. we demonstrate the utility of our index for studying selected decision and policy problems. © 2022 academy of financial services. all rights reserved. jel classification: g53 (household finance: financial literacy); g11 (portfolio choice; investment decisions) keywords: financial literacy; mutual funds; investor decision-making; measurement; human capital; knowledge; exchange-traded funds (etfs) the securities and exchange commission disclaims responsibility for any private publication or statement of any commission employee or commissioner. the commission has expressed no view regarding the data, analyses, findings, and conclusions contained in this publication, which does not necessarily reflect the views of the commission, the commissioners, or staff of the commission. this paper is part of a series of papers that the office of the investor advocate is publishing. papers in this series investigate a broad range of issues relevant to the commission’s mission and are disseminated to stimulate discussion and critical comment. inclusion of a paper in this series does not indicate a commission determination to take any particular action. *corresponding author. tel.: +1-608-630-1373. e-mail address: angela@fontesresearch.com 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 31–56 1. introduction mutual funds are extremely important to household investment portfolios, potentially providing well-diversified investment management options for most investors, and serving as the main investment vehicle in retirement and educational savings accounts.1 with the transition to defined contribution retirement plans well underway, nearly 80% of investors hold employer-sponsored retirement plans, putting the onus of investment management largely on individuals themselves. the majority of these individuals report ownership of mutual funds and exchange traded funds (etfs)—with more investors reporting ownership of funds than stocks and bonds combined; some estimates suggest that the number of individual mutual fund shareholders was 99.5 million in 2018 (ici, 2010, 2018).2 although mutual funds are often targeted towards individual investors, their complexity may be an obstacle that inhibits investor choice. to better understand the precise knowledge gaps that may lead to potential investor decision pitfalls, we develop an easily deployable and respondent-friendly battery of 11 true-false technical mutual fund questions that allows us to measure respondent knowledge of key concepts important to investment decision making. our battery is the result of extensive input from extant literature, financial regulators, a close reading of the mandated disclosure documents intended to communicate important product features to investors, substantial cognitive interviewing, multiple rounds of testing, and considerable data analysis following standard practice in index development (in particular, see devellis, 2016). administration is comprised of a set of progressively more complicated technical questions that seek to test knowledge of basic concepts such as risk, fees, regulatory protections, marketplace characteristics and performance. we field the battery on a large, nationally representative, probability-based panel, allowing us to provide a broad perspective on u.s. households. to summarize respondent knowledge, we use these individual questions to construct a simple index that well-encapsulates the level of technical mutual fund knowledge of the investing public. we believe our index is highly useful in the context of regulatory policy and academic work on investor decision making. policymakers have an interest in making sure mutual fund products work as intended so that investors achieve their investment goals, with investor protection one of the main goals of the u.s. securities and exchange commission (sec), the body that regulates mutual funds.3 regulators operationalize policies that advance this goal with rules limiting or prohibiting certain activity and (perhaps most importantly) by mandating disclosures that provide transparency about decision-relevant features such as fees and risks, which can facilitate informed investor decision making.3 the interrelated nature of financial knowledge and disclosure is highlighted by u. s. securities & exchange commission (2012), a sec staff report on financial literacy mandated by congress as part of the dodd-frank act; the report notes troubling disparities in knowledge, particularly in selected demographic groups, but also responds at length to specific congressional mandates to probe into highly related issues in disclosure. in a disclosure-centric regulatory regime, the burden of choosing investments ultimately falls on the investor. disclosure of information about a particular product does not mean that investors can understand the concepts, understand their implications or make appropriate investment choices. such conceptual gaps may forestall potential investors from investing 32 b. scholl and a. fontes / financial services review 30 (2022) 31–56 leaving them unprepared for important goals such as college or retirement, but for those that do invest, a knowledge gap related to important technical specifics may result in choices that have real consequences for a household’s balance sheet. for example, an investor that follows conventional wisdom by investing $100,000 in one of the lowest priced s&p index funds (“fund a”) with returns averaging 7% per year would pay approximately $1,300 in fees over 20 years, and would have a final balance of approximately $385,000. but an investor that makes a single mistake in their purchase decision by selecting one of the higher priced s&p index funds (“fund b”)—a mistake that could arise by simply picking the more expensive option from a broker’s menu of funds—would pay approximately $80,000 in fees over the same period and end up with less than $240,000. the consequences of fees are well known to professionals, policymakers, and academics, but it is important to ask: how widely known is this to the public who may not be as experienced or knowledgeable?5 moreover, what other misconceptions do they hold about mutual funds that could lead to other costly mistakes? identifying these deficiencies may help in developing more targeted interventions to complement broad financial education initiatives. in the context of investor decision making, our index complements existing measures of general financial knowledge that have been associated with a broad range of economic outcomes. a large body of evidence has elevated attention to financial knowledge and its implications for various financial outcomes (lusardi, 2019, 2012, 2008; lusardi & mitchell, 2014; lusardi, mitchell, & curto, 2014; van rooij, lusardi, & alessie, 2011), such as wealth accumulation, debt management, general financial management, and uptake of financial advice (scholl & hung, 2018). typical measures of financial literacy attempt to glean knowledge of general economic principles, which may make them helpful for understanding such concerns as overall well-being and asset accumulation, but these measures may be insufficiently specialized for the policy-making context of financial market regulators. for example, the useful and widely accepted three-question financial literacy questions advanced by lusardi and mitchell (lusardi & mitchell, 2014, 2008) tests literacy in terms of questions on purchasing power and inflation, risk, and interest compounding. in extended versions of this standard battery (lusardi, 2008; lin et al., 2019; van rooij, lusardi, & alessie, 2011), further developments and refinements of these fundamental measurement concepts are made to examine the knowledge of increasingly complex principles of economics, finance, and investment. yet, in the $145,000 investment mistake illustrated above, the knowledge gap that leads to this investor decision-making error relates to technical features of funds and the market for funds, rather than general economic knowledge. leading up to this “asset selection decision” between fund a and b based on an evaluation of the characteristics of the funds, an individual presumably has previously made the decision to invest in securities (“participation decision”) and determined that a particular asset class such as mutual funds are right for them (“asset class selection decision”). of the lusardi-mitchell big 5 general financial literacy questions, the most directly relevant to the context of mutual fund investments is a truefalse question that asks: “buying a single company’s stock usually provides a safer return than a stock mutual fund?” this question seems directly related to the asset class selection decision and perhaps the participation decision, but less directly related to the asset selection decision in which the investor evaluates the merits of features of funds. to make a reasonable asset selection choice, the investor may need to be aware of facts such as: that b. scholl and a. fontes / financial services review 30 (2022) 31–56 33 differences in fees exist; that the expense ratio does not necessarily reflect all fees such as commissions or loads; that fees compound over time much like interest; and where to find information about fees. moreover, the investor is often only presented with a single fund, and may need to evaluate such a fund in reference to an unknown set of alternatives; if they are presented with fund b, they may need to engage in costly search to identify a better alternative (as in hortaçsu & syverson, 2004). a lack of knowledge about the potential to find alternatives may forestall search, possibly because they do not know that cheaper, nearly identical alternatives exist. for securities regulators, the asset selection decision is arguably the most relevant to policy levers given that disclosures tend to contain information relevant to investment selection and management decisions rather than participation decisions. these concerns motivated us to develop an index that more specifically focuses on technical knowledge geared towards asset selection decisions that are most relevant in the policy and research contexts we study. technical knowledge of investment products would seem a prerequisite to good investor decision making—after all, if an investor does not understand fees, where they may be hidden, or how to identify them in the disclosures, they ultimately may not even know the choice dimensions on which they should be optimizing. the focus on technical knowledge recognizes that disclosures are typically written by securities attorneys with specialized knowledge of the subject matter, and are often intended for multiple audiences: from mutual fund experts to ordinary investors with limited investment experience. over time, much criticism has been levied at disclosures that are not informative for retail investors.6 our broad view is that general financial literacy may be informative about respondent inclusion of mutual funds in their portfolios, but believe our measure of knowledge may provide additional insight on the respondent’s ability to distinguish between products in the marketplace. it is households’ propensity for making technical mistakes in mutual fund choice decisions that we attempt to assess in our measure of knowledge; arguably, such propensities should be of interest to policymakers and researchers interested in promoting better investor decision making. ultimately, the relative value of any measure of knowledge in a particular context is an empirical question that we study in this article; here, we provide extensive evidence that demonstrates the utility of our knowledge index. regrettably, the picture we paint about mutual fund knowledge in the population is somewhat bleak. overall knowledge scores are quite low with many respondents performing far worse than if they had guessed randomly in responding to the questions. these results are disheartening: most of the concepts that form a potential basis for informed decision making, and have been mandated for inclusion in disclosures, are not broadly understood. our findings call into question the sufficiency of existing disclosures as a vehicle for providing decision-relevant information to investors, and also raise questions about the adequacy of relying on disclosure alone as such a prominent method of promoting investor protection. the results also challenge the notion that financial education on its own has been sufficient to equip most individuals with the knowledge and skills needed to successfully manage their own investments: either that education has been insufficiently widespread, or its efficacy may be limited; whatever the reason, we do not find a sufficient level of pass-through from education to decision-relevant knowledge at the population-level. these results are particularly troubling especially in light of the importance of mutual funds to education and savings (scholl & hung, 2018). moreover, while fees are widely considered among the most 34 b. scholl and a. fontes / financial services review 30 (2022) 31–56 important and controllable feature on which to make choices, see, for example, (barber et al., 2005; choi et al., 2010; carhart, 1997; sec office of investor education and advocacy 2016), virtually every subgroup we have examined has demonstrated extremely poor performance on fee questions.6 our analysis demonstrates the potential of our knowledge index to serve as a measurement tool to explain a host of household financial concerns including investment participation (i.e., ownership of investment accounts and/or financial securities), financial well-being, and fee computation skill. other emergent work has also demonstrated the utility of the measure in other contexts (e.g., chin, scholl, & vanepps, 2021; scholl, 2020). this article proceeds as follows: section 2 provides a comprehensive review of relevant literature; section 3 describes the individual items and methodology; section 4 provides a synopsis of item and cumulative results; section 5 discusses index development and validation; section 6 concludes. 2. literature our article speaks to the literature on financial literacy, and to a lesser extent financial education, policy work on the role of financial capability and education in promoting better investment outcomes, and broader literature on financial decision making. financial literacy has been associated with a wide array of economic outcomes such as debt management, wealth accumulation, financial vulnerability, and a host of other economic outcomes in the united states and other countries (lusardi, 2008, 2012, 2019; lusardi & mitchell, 2014; lusardi & tufano, 2009, 2015; lusardi, mitchell, & curto, 2014; van rooij, lusardi, & alessie, 2011). in these studies, literacy is measured by the number of correct responses to a set of survey questions, with by far the most widely accepted being the standard set developed as the lusardi-mitchell “e” or the “big 5” (lusardi & mitchell, 2014, 2011; lusardi & mitchell, 2008). other work has extended these standard questions (e.g., lin et al., 2019; lusardi, 2008), while other approaches to measurement of financial literacy are surveyed in elan (2011). these studies overwhelmingly find that higher levels of financial literacy are associated with more favorable financial and economic outcomes (while also providing important tools for measurement of knowledge within the population). rather than using survey measures, calvet, campbell, and sodini (2007) measure the related concept of financial sophistication, backing out an index of sophistication from identifiable mistakes in observed household portfolio choices using swedish administrative data, although that approach requires data that is largely unavailable for most populations. meta-analyses that have altogether examined hundreds of studies on the topic have demonstrated mixed conclusions as to the importance of financial literacy and financial education programs (fernandes et al., 2014; kaiser & menkhoff, 2017, 2020). the topics remain of sustained interest, with a recent special issue of the economics of education review providing a wealth of articles examining related issues in financial literacy and education (including davoli & rodrı́guez-planas, 2020; kaiser & menkhoff, 2020; lusardi et al., 2020; urban et al., 2020). hastings, madrian, and skimmyhorn (2013) outline a number of issues and raise a number of outstanding questions related to financial literacy, including the b. scholl and a. fontes / financial services review 30 (2022) 31–56 35 very goals related to research and education. beshears et al. (2018) provide an extensive review of this literature in behavioral household finance. they conclude that while the literature has demonstrated the potential effectiveness of financial education and information interventions, they express some skepticism in relation to cost-effectiveness. while these are extensive contributions to the debate on topics of financial education, it is important to note that we do not take a perspective in this article on financial education per se nor its efficacy: we view our article as focusing on measurement that can help explain behavior in certain decision and policy contexts. while mutual fund products are extremely important to household investment portfolios (scholl & hung, 2018), specific knowledge of mutual fund features has apparently received far less attention than overall financial literacy. muller and turner (2021) examine knowledge in the context of the “high-fee puzzle,” or the selection of strictly dominated funds. they find that while three quarters of their sample correctly answer common financial literacy questions, only a third could correctly answer questions related to quantifying fees. kahraman (2021) examines investor mistakes in the context of purchasing inappropriate (and more expensive than necessary) share classes for mutual funds, leading to real consequences for investors. the author presents evidence to suggest that the selection of these inferior share classes is a form of exploitation of investors by professionals. in addition, the author examines fee-flow sensitivity and holding periods to test whether fund flows suggest rational or naı̈ve purchase of these funds, concluding that these are naı̈ve purchases. other recent work has examined additional barriers to investment decision making, in the context of the language used to describe funds. chin, scholl, and vanepps (2021) study linguistic barriers to understanding mutual fund fees. they conduct two studies that suggest that terminology often used is unintuitive for respondents, and that a simple set of alternative terms they test leads respondents to higher rates of identifying the true underlying fee concept. scholl, silverman, and enriquez (2020, 2021), examine prospectus readability and other textual features using natural language processing and machine learning techniques, and relate these features to ex-post fund performance. readability calculated using structural features of prospectus sentences; this concept is distinct from comprehension of the underlying concepts, which could require expert knowledge. one descriptive fact that they document is that readability of fund disclosures is extremely low. less than one percentage of fund summary prospectus documents are as readable as a u.s.a. today article, while roughly three-quarters are less readable than the u.s. tax code. the majority of these documents are at college reading levels and above. dehaan et al. (2021) provide evidence that highlights the potential for intentional obfuscation in terms of narrative complexity and the structural complexity of the securities instrument. the only study of which we are aware that specifically attempts to profile specific knowledge of mutual funds products is alexander, jones, and nigro (1998) that was conducted during a very different investing environment. that article focuses on the sources of information that investors use and the differences in knowledge based on the purchase channels investors use. the authors concluded based on survey results conducted a quarter of a century ago that there is much room for improvement in investor’s knowledge levels; they also survey an earlier literature that documents some common misconceptions of investors such as that mutual funds sold through a bank are backed by fdic insurance. our work builds off 36 b. scholl and a. fontes / financial services review 30 (2022) 31–56 some of the key features identified in alexander, jones, and nigro (1998), while delving into a broader set of mutual fund features, and formalizing an index. moreover, our knowledge index design and our survey methodology are quite distinct in several important dimensions. our sample is a general population sample, which includes both individuals residing in mutual fund owning households as well as those that reside in households that do not own mutual funds that were the focus of alexander et al. (1998).8 this sample distinction allows for us to document initial observations related to our interest in participation in the market (although a more detailed study of participation barriers in connection to knowledge deserves a separate treatment). our intention is that our work is informative in the context of literature on investor choice and decision making, and lays the groundwork for further decomposing aspects of decision making. a relatively large body of literature has emerged along these lines in recent decades relating largely to deviations from rationality or a lack of information by investors (e.g., see observational studies by barber et al. [2005], elton et al. [2011]; as well as work using behavioral experiments such as beshears et al. [2011], choi et al. [2010]; and also the extensive review in beshears et al. [2011]). in a related study, müller and weber (2010) construct a financial literacy measure, and look at participation and choice decisions in selected mutual fund markets. they note there is a positive relationship between their literacy measure and the likelihood of investing in active funds (that they argue are worse than passive management alternatives); nevertheless, despite the fact that literacy matters, it alone cannot explain the historical growth of active management. overall, they find only weak evidence that investors that performed well on their measure had superior fund selection skills. scholl (2020) examines decision making in the presence of choice set complexity in a large-scale experiment in which subjects complete an allocation problem from a menu of s&p 500 index mutual funds. that study directly uses an early version of the mutual fund knowledge index we examine here. scores on our index demonstrated strong separation and well ordering of subjects in terms of their overall performance on the investment task (where fee minimization is a strictly dominating strategy). other forthcoming behavioral research using allocation experiments by scholl and coauthors also demonstrate the utility of our mutual fund knowledge index in other contexts, such as in the classification of investor types described in chin, scholl, and vanepps (2021). the importance of the knowledge we test is highlighted in bhattacharya et al. (2017), who decomposed investor decision mistakes with respect to etf investments into “poor timing” and “poor selection”—the latter accounting for 1.69% loss per annum in investor returns; our index directly tests knowledge that could plausibly help investors avoid such selection mistakes. 3. method 3.1. data collection data from 3,444 respondents were collected over three waves of data collection (over a 10-month period) using the amerispeak panel (as), a probability-based, nationally representative u.s. panel.9 respondents were incentivized for participation in each wave of the b. scholl and a. fontes / financial services review 30 (2022) 31–56 37 t ab le 1 d em o g ra p h ic ch ar ac te ri st ic s, b y g ro u p d em o g ra p h ic ch ar ac te ri st ic s, b y g ro u p p an el a : t o ta l p an el b : g en er al fi n an ci al li te ra cy p an el c : v al id at io n v ar ia b le b re ak o u ts in d ex it em p o p to ta l l o w li te ra cy h ig h li te ra cy in v es to rs c al cu la te d fe e q u es ti o n co rr ec tl y f w b q 3 + g en d er (m al e) 5 5 .7 5 % 3 3 .7 8 % 6 4 .4 2 % 5 6 .2 5 % 6 6 .0 8 % 5 8 .5 4 % a g e 1 8 -2 9 4 .2 7 % 6 .7 6 % 4 .0 0 % 4 .2 0 % 4 .2 4 % 3 .3 3 % 3 0 -4 4 2 5 .2 6 % 3 5 .1 4 % 2 2 .2 3 % 2 5 .4 9 % 2 7 .1 8 % 1 1 .9 2 % 4 5 -5 9 2 9 .3 8 % 2 8 .3 8 % 2 8 .7 9 % 2 9 .2 9 % 2 7 .4 3 % 2 0 .6 2 % 6 0 + 4 1 .0 9 % 2 9 .7 3 % 4 4 .9 8 % 4 1 .0 2 % 4 1 .1 5 % 6 4 .1 2 % r ac e/ et h n ic it y w h it e n o n -h is p an ic 8 3 .4 2 % 6 3 .5 1 % 8 5 .4 0 % 8 3 .8 7 % 8 3 .4 2 % 8 8 .4 0 % a fr ic an a m er ic an n o n -h is p an ic 3 .7 5 % 1 2 .1 6 % 2 .7 0 % 3 .7 3 % 3 .3 7 % 2 .9 0 % h is p an ic 4 .7 6 % 1 2 .1 6 % 3 .8 6 % 4 .6 3 % 3 .2 4 % 3 .1 1 % o th er 8 .0 7 % 1 2 .1 6 % 8 .0 5 % 7 .7 6 % 9 .9 8 % 5 .5 9 % e d u ca ti o n n o h ig h sc h o o l d ip lo m a 1 .1 0 % 5 .4 1 % 0 .3 3 % 0 .7 7 % 0 .5 0 % 0 .4 3 % h ig h sc h o o l g ra d u at e o r eq u iv al en t 7 .9 3 % 1 7 .5 7 % 4 .3 3 % 6 .5 6 % 3 .9 9 % 4 .9 4 % s o m e co ll eg e 2 7 .1 8 % 4 4 .5 9 % 2 2 .4 2 % 2 5 .4 2 % 1 9 .8 3 % 2 1 .9 1 % b a o r ab o v e 6 3 .7 9 % 3 2 .4 3 % 7 2 .9 3 % 6 7 .2 4 % 7 5 .6 9 % 7 2 .7 2 % in co m e l es s th an $ 3 5 ,0 0 0 1 6 .0 6 % 3 3 .7 8 % 1 1 .1 2 % 1 1 .2 0 % 1 0 .3 5 % 5 .8 0 % $ 3 5 0 0 0 -$ 5 9 ,9 9 9 1 7 .5 7 % 2 2 .9 7 % 1 6 .1 9 % 1 6 .8 3 % 1 4 .4 6 % 1 1 .2 8 % $ 6 0 ,0 0 0 -$ 9 9 ,9 9 9 2 9 .8 8 % 2 0 .2 7 % 3 0 .5 6 % 3 1 .5 9 % 2 9 .5 5 % 3 2 .6 5 % $ 1 0 0 ,0 0 0 o r m o re 3 6 .5 0 % 2 2 .9 7 % 4 2 .1 4 % 4 0 .3 9 % 4 5 .6 4 % 5 0 .2 7 % n et w o rt h in d eb t 1 2 .2 2 % 2 2 .9 7 % 8 .8 8 % 9 .8 0 % 8 .1 0 % 2 .5 8 % z er o 9 .7 0 % 2 9 .7 3 % 6 .3 7 % 8 .2 3 % 5 .6 1 % 2 .2 6 % g re at er th an ze ro 7 8 .0 8 % 4 7 .3 0 % 8 4 .7 4 % 8 1 .9 7 % 8 6 .2 8 % 9 5 .1 7 % f ee ca lc u la ti o n (c o rr ec t) 2 3 .2 9 % 1 3 .5 1 % 2 9 .7 2 % 2 4 .8 3 % 1 0 0 .0 0 % 2 7 .3 9 % f in an ci al w el lb ei n g (q 3 + ) 2 7 .0 3 % 1 4 .8 6 % 3 1 .9 1 % 2 9 .5 2 % 3 1 .8 0 % 1 0 0 .0 0 % f in an ci al li te ra cy sc o re (m ) 2 .4 8 2 .5 3 3 .0 0 2 .5 3 2 .7 3 2 .6 6 n 3 ,4 4 4 7 4 2 ,1 5 0 3 ,0 0 1 8 0 2 9 3 1 n o te s: “l o w li te ra cy ” g ro u p an sw er ed n o l u sa rd im it ch el l b ig 3 q u es ti o n s co rr ec tl y , “h ig h li te ra cy ” g ro u p an sw er ed th re e q u es ti o n s co rr ec tl y . f w b q 3 + ar e re sp o n d en ts th at w er e in th e th ir d an d fo u rt h q u ar ti le o f f w b . f w b = c f p b ’s f in an ci al w el lb ei n g . 38 b. scholl and a. fontes / financial services review 30 (2022) 31–56 study via points redeemable for consumer goods, gift card, or cash of approximately $10. average survey length for each wave was 21-26 min, although our knowledge index only required a few minutes to complete. several aspects of the as panel contribute to data quality when investigating household finances. panelists are recruited with an enhanced addressbased sampling frame that increases coverage of u.s. households to over 97%, increasing rural enumeration. internet-phone mixed-mode data collection accommodates households without internet access.10 table 1 presents the demographics of the sample as a whole, and by subcategories used in index validation. by intention, we oversampled likely mutual fund investors in our initial recruitment; as such, the sample reported here is older and more educated than the u.s. population.11 3.2. mutual fund literacy item identification we developed a set of true/false items based on a careful review of existing work in the areas of general financial literacy (e.g., lusardi, 2008; lusardi & mitchell, 2014), investment literacy (e.g., forbes & kara, 2010), and mutual fund investment knowledge (specifically, alexander, jones, & nigro, 1998). our question areas were designed to test respondent technical knowledge on key features of mutual funds. more specifically, they were designed to elicit technical knowledge of key choice features (most notably, risks and fees) that we identified as helpful to an investor in the context of mutual fund selection problems. we believe that individuals lacking knowledge of these attributes would be impaired in investment decision-making situations. we favored this technical knowledge approach over a more generalized set of economic concepts that are typically found in financial literacy batteries. once identified, items were then refined with extensive expert input from individuals with highly specialized knowledge and experience regarding regulation, financial education, investor advocacy, in the context of both the technical features of the funds and the regulatory tools that regulators apply to mutual fund products. some of these individuals were intimately familiar with financial literacy issues and programs as they are viewed and implemented by regulators, while others had experience in disclosure review and/or the writing of disclosure rules and regulations. an investor advocate entity that promotes pro-investor policies with regulators also provided an important perspective. we refined the questions so that they would identify and elicit knowledge gaps that would inhibit investors from utilizing the information contained in disclosures for decision making. we view the link to mutual fund disclosures, the primary method of information exchange on investment options, as extremely important. our article focuses on technical knowledge that relates to feature concepts or applications that may inhibit investors from making use of disclosed information for informed decision making. for example, for an investor to pick a mutual fund that avoids a load fee: the individual may need to be aware that a sales charge exists; what services the load pays for and what is the typical range in such charges; know that it must be disclosed in disclosures; understand that the appropriate term is “load”; be able to locate it in disclosures; potentially evaluate the fee as part of a tradeoff vis-à-vis other fees; understand that no-load mutual fund investments exist; distinguish it from alternative “sales” charges (e.g., broker commissions)—that is, know that commissions are not the only b. scholl and a. fontes / financial services review 30 (2022) 31–56 39 sales charge; and potentially to understand that loads can potentially be applied at both the time of purchase and the time of sale. without these elements of knowledge—clearly linked to disclosed information and the investor’s interaction with disclosures—the investor may not be able to make an optimal choice. the overriding importance of this linkage between knowledge or literacy and disclosures has not received much attention in the academic literature, but is highlighted by a 2012 sec staff report on financial literacy (u.s. securities and exchange commission, 2012), which was undertaken pursuant to a mandate in the doddfrank act by congress. qualitative research was used extensively during the development process. an initial round of 19 interviews provided early insights that helped us to identify broad deficiencies in misconceptions. these interviews were centered around completion of a specific mutual fund choice task, and revealed large deficiencies in knowledge for certain participants. later, once themes and the initial structure of question items were identified, we tested them with extensive cognitive interviews in subsequent rounds with a total of 23 participants drawn from a nationally representative probability sample. such testing helped to refine the phrasing of questions and fielding protocol of the questions (in particular, the need for randomization of placement on a survey instrument). a larger battery of questions was originally considered, including both multiple choice questions and true-false questions on technical features of funds. additional questions increased difficulty of the overall assessment considerably, and appeared to make interviewees—particularly those with less experience in investment—more reluctant to venture answers to the questions at all. moreover, while these questions added further richness to our perspectives of respondents, in the end, we determined that a focused battery centering on the true-false questions provided sufficient comfort to the interviewees, sufficient well-ordering in terms of sophistication, and that we had a sufficient number of questions to distinguish between respondents based on sophistication. the interviews also provided insight into dealing with incorrect answers versus question skips.12 one particular debate that cognitive testing sparked among members of the research team was related to the use of terminology. our testing revealed that some individuals were so unfamiliar with mutual funds and etfs that they did not understand the terminology used in our questions such as “loads” or the “expense ratio.” while it can be argued that this unfamiliarity with the terminology creates barriers to answering the questions (as per chin, scholl, & vanepps, 2021), these barriers directly mimic barriers that individuals would face in seeking and choosing among mutual fund investments and are related to the specific knowledge each question is testing. after all, in reading investment disclosures, investors have to do their own translation of technical language. as such, we deemed that preserving such terminology was important in the context of assessing a respondent’s overall knowledge. naturally, it is possible that individuals that have engaged with the market in terms of key investment decisions in the past will have been more motivated to understand terminology before making a decision. yet, this learning process would consequentially be captured in our stock measurements of overall investor knowledge in the population—and would analytically present itself in the difference in knowledge levels between investors and non-investors. in addition to these steps in development, we have utilized the battery in experiment and testing studies with thousands of participants (e.g., scholl, 2020). the battery has overall been 40 b. scholl and a. fontes / financial services review 30 (2022) 31–56 effective at rank ordering investor sophistication, as well as the propensity of individuals to exhibit a number of decision making and comprehension mistakes in investment settings (e.g., naı̈ve diversification, susceptibility to complexity, and failures to avoid fees). our final battery identified 11 items in four key areas: market alternatives, risk, performance history, and fees. questions were true/false, with a “don’t know” response option. these questions (correct answers in parentheses) are as follows, with labels assigned to each question for ease of reference: marketplace alternatives category: 1. financial markets offer thousands of different mutual funds to investors. (true) (label: market options) risk category: 2. mutual funds pay a guaranteed rate of return. (false) (“guaranteed return”) 3. it is possible to lose money in a stock mutual fund. (true) (“risk stockfund”) 4. it is possible to lose money in a bond mutual fund. (true) (“risk bondfund”) 5. if a mutual fund is registered with the securities and exchange commission (the sec) or state securities regulators, you cannot lose money. (false) (“risk regulation”)13 performance history category: 6. a good predictor of the future performance of a mutual fund is its past performance. (false) (“performance history”)fee category: 7. a no-load mutual fund charges yearly expenses. (true) (“yearly expenses”) 8. a load fee is charged only when the fund is initially purchased. (false) (“load”) 9. fees and expenses for the mutual fund industry are capped at a maximum level by regulatory authorities. (false) (“fee cap”) 10. fund fees are required to be reported in the fund’s prospectus document. (true) (“prospectus fees”) 11. the fees or expenses charged by the mutual fund company in a given year can be approximated by multiplying the fund’s net expense ratio by the investment gains for the year. (false) (“fee basis”) in addition to these four primary categories, questions risk regulation and fee cap implicitly ask respondents for assumptions about regulatory protections. question prospectus fees connects to disclosure requirements (and the respondent’s familiarity with the prospectus document from which much of a fund’s decision-relevant information can be gleaned). question performance history also has a direct link to the standard disclosure, sometimes referred to as the mutual fund warning label that is required by the securities and exchange commission on certain performance presentations. to minimize order effects, we randomized the presentation order of questions on the survey. as much as possible, we endeavored to develop questions that had objectively correct and incorrect answers rather than ones that might be considered situationally dependent. although the questions have varying degrees of difficulty, few, if any of the questions, can be considered “trick questions.” b. scholl and a. fontes / financial services review 30 (2022) 31–56 41 t ab le 2 it em le v el re su lt s, b y g ro u p p er ce n ta g e co rr ec t, b y g ro u p p an el a : t o ta l p an el b : g en er al fi n an ci al li te ra cy p an el c : in v es to r st at u s p o p to ta l s e l o w li te ra cy s e h ig h li te ra cy s e in v es to r s e n o n -i n v es to r s e f in an ci al m ar k et s o ff er th o u sa n d s o f d if fe re n t m u tu al fu n d s to in v es to rs . 7 1 .0 5 0 .7 7 5 6 .7 6 5 .8 0 8 0 .6 0 0 .8 5 7 4 .1 1 0 .8 0 5 0 .3 4 2 .3 8 m u tu al fu n d s p ay a g u ar an te ed ra te o f re tu rn . 6 4 .2 3 0 .8 2 2 1 .6 2 4 .8 2 7 9 .1 6 0 .8 8 6 8 .3 4 0 .8 5 3 6 .3 4 2 .2 9 it is p o ss ib le to lo se m o n ey in a st o ck m u tu al fu n d . 8 1 .2 7 0 .6 6 6 0 .8 1 5 .7 1 8 9 .0 7 0 .6 7 8 3 .5 1 0 .6 8 6 6 .1 4 2 .2 5 it is p o ss ib le to lo se m o n ey in a b o n d m u tu al fu n d . 5 5 .1 1 0 .8 5 3 6 .4 9 5 .6 3 6 4 .2 8 1 .0 3 5 7 .7 1 0 .9 0 3 7 .4 7 2 .3 0 if a m u tu al fu n d is re g is te re d w it h th e s e c o r st at e se cu ri ti es re g u la to rs , y o u ca n n o t lo se m o n ey . 6 7 .7 1 0 .8 0 2 7 .0 3 5 .2 0 8 1 .0 2 0 .8 5 7 0 .8 4 0 .8 3 4 6 .5 0 2 .3 7 a g o o d p re d ic to r o f th e fu tu re p er fo rm an ce o f a m u tu al fu n d is it s p as t p er fo rm an ce . 3 2 .9 3 0 .8 0 1 7 .5 7 4 .4 5 4 0 .4 2 1 .0 6 3 4 .9 6 0 .8 7 1 9 .1 9 1 .8 7 a n o -l o ad m u tu al fu n d ch ar g es y ea rl y ex p en se s. 2 4 .2 2 0 .7 3 1 6 .2 2 4 .3 1 3 0 .4 2 0 .9 9 2 5 .7 6 0 .8 0 1 3 .7 7 1 .6 4 a lo ad fe e is ch ar g ed o n ly w h en th e fu n d is in it ia ll y p u rc h as ed . 1 6 .4 6 0 .6 3 5 .4 1 2 .6 5 2 0 .9 8 0 .8 8 1 7 .7 9 0 .7 0 7 .4 5 1 .2 5 f ee s an d ex p en se s fo r th e m u tu al fu n d in d u str y ar e ca p p ed at a m ax im u m le v el b y re g u la to ry au th o ri ti es . 2 2 .8 2 0 .7 2 9 .4 6 3 .4 3 3 0 .3 7 0 .9 9 2 4 .6 6 0 .7 9 1 0 .3 8 1 .4 5 f u n d fe es ar e re q u ir ed to b e re p o rt ed in th e fu n d ’s p ro sp ec tu s d o cu m en t. 5 9 .9 3 0 .8 4 3 7 .8 4 5 .6 8 7 2 .0 9 0 .9 7 6 3 .0 8 0 .8 8 3 8 .6 0 2 .3 2 t h e fe es o r ex p en se s ch ar g ed b y th e m u tu al fu n d co m p an y in a g iv en y ea r ca n b e ap p ro x im at ed b y m u lt ip ly in g th e fu n d ’s n et ex p en se ra ti o b y th e in v es tm en t g ai n s fo r th e y ea r. 1 4 .3 7 0 .6 0 1 .3 5 1 .3 5 1 8 .8 8 0 .8 4 1 5 .5 3 0 .6 6 6 .5 5 1 .1 8 m ea n sc o re (a v er ag e to ta l co rr ec t) 5 .1 2 .9 6 .1 5 .4 3 .3 n 3 4 4 4 7 4 2 1 5 0 3 0 0 1 4 4 3 42 b. scholl and a. fontes / financial services review 30 (2022) 31–56 4. results 4.1. individual question responses table 2 provides the proportion of the sample who responded correctly to each item in the mutual fund index. considering that most questions for this survey were developed with the intention of reflecting basic properties of mutual funds that should be considered when making investment decisions, the results overall are not encouraging in terms of respondent knowledge. of the 11 questions we developed, only six were each answered correctly by more than half of respondents. the remaining five questions each had only a third of respondents, or fewer, responding correctly. the marketplace options question aims to determine if respondents are aware that there are many different alternative investment options available to them (and by implication, if they are not happy, they can shop around). encouragingly, over 71 percent of respondents were aware that financial markets offer thousands of different mutual fund options. of course, this does not mean that the respondents believe they have the skills to successfully navigate such a diverse choice environment, and in fact, it is conceivable that investors and non-investors may be paralyzed by choice (e.g., see carvalho & silverman, 2019, and the somewhat related and agnew & szykman, 2005). concern was reflected in cognitive testing conducted during the refinement of survey questions, where several respondents indicated that choosing funds felt overwhelming (e. g., when making choices related to their employer-sponsored retirement plan). however, these responses at least suggest a realization that alternatives exist and it may be worth additional search in the marketplace if their satisfaction is low with their current investment mix. in terms of risk, 36% of respondents thought that mutual funds pay a guaranteed rate of return. almost half thought that it is not possible to lose money in a bond mutual fund (45%), although more than three-quarters (81%) did recognize that one could lose money in a stock mutual fund. in addition, one-third (32%) thought that mutual funds that are registered with the sec or a state regulator cannot lose money. overall, these results suggest very little understanding of mutual fund risk. despite the warning label offered on mutual fund product documents, nearly two-thirds (67%) indicated that past performance is a good predictor of future performance.14 the extant literature is not supportive of this view (e.g., brown & goetzmann, 1995; carhart, 1997; goetzmann & ibbotson, 1994; malkiel, 1995).15 fee questions were constructed to be slightly more technical than risk questions, with correct responses that may require knowledge that is more specialized to mutual fund products; at the same time, these questions are germane in the context of mutual fund choice problems given the importance of fees in determining net returns. only a quarter (24%) of respondents correctly stated that no-load mutual funds charge yearly expenses (as would be reflected in the fund’s expense ratio), perhaps indicating a lack of understanding of the term load, which refers to a sales charge. only 16% correctly identified that a load is not confined to the point of purchase (“front-load”), which sales loads (“back-loads”) exist. seventy-eight percent erroneously believed that fees are capped at a maximum level by regulatory authorities. three-fifths (60%) indicated that fund fees need to be reported in the fund’s prospectus. b. scholl and a. fontes / financial services review 30 (2022) 31–56 43 perhaps most disheartening is that only 14% correctly identified our conceptual fee computation question as false. the fee computation question was intended to identify if respondents understood that fees are computed based on total account balance rather than on the basis of investment returns. this observation arose in cognitive interviews that revealed that some individuals believed that the fee basis is the much lower level (investment gain) than it actually is in practice (total balance), and the possibility that some investors believe that fees are not accrued if the fund has negative performance in a given period. respondents overwhelmingly indicated that they implicitly believe that mutual fund fees are much lower than they really are, perhaps giving insight to the bhattacharya et al. (2017) results. 4.2. cumulative index scores summing correct responses provides a composite score of the extent of a respondent’s knowledge about mutual funds, as well as a general perspective on aggregate knowledge of the public. fig. 1 presents the distribution of index performance as the sum of correct responses. fig. 2 presents the cumulative distribution of correct responses. blue vertical lines denote the expected value from coin tossing true-false responses (5.5). overall performance in the literacy index is poor. the average respondent score was 5.1, with a median score of 5.0 (table 2a, panel a). in fig. 1, 52.2% of the sample answered less than six questions correctly; 13.2% answered less than two questions correctly. only 11.3% answered at least nine questions correctly. average results in particular groups roughly align with expectations, but highlight additional deficiencies. in panel b, average scores for those with the highest score on general financial literacy (all three “big 3” questions answered correct) and low general financial literacy (no “big 3” questions answered correctly), align with expectations to a certain degree. the few respondents (n = 74) that failed to correctly answer any of the generalized fig. 1. mutual fund knowledge score distribution. 44 b. scholl and a. fontes / financial services review 30 (2022) 31–56 financial literacy questions correctly, answered less than three mutual fund knowledge questions correctly, on average. high financial literacy score respondents (n = 2,150), performed significantly better, with an average of 6.1 questions correct (two-sample t test t-statistic of difference in means is 10.5, with a p-value of 0.00). yet, the average score of 6.1 questions correct is not impressive. note that the correlation coefficient between our knowledge score and general financial literacy in the sample is moderate at 0.44. fig. 2. cumulative distribution function of scores note: cumulative number of respondents that scored less than or equal to a given score value. fig. 3. factor loadings in a two-factor solution. b. scholl and a. fontes / financial services review 30 (2022) 31–56 45 our survey collected both investor and non-investor responses. on one hand, it may be argued that non-investors are less consequential for determining overall knowledge and excluded from analysis, given that they may not have experience with these products. on the other hand, we view knowledge as a potentially important barrier for participation decisions, so application of the index to this subpopulation is of interest. nevertheless, the question remains: do non-investors drive the results reported above? the answer from table 2a (panel c) is clearly no. non-investors make up only about 13% of our respondents. while their scores on average are much lower than those for investors (3.3 vs. 5.4; 2-sample t test t-value: 15.2, p < .001), this has little effect on the average score previously reported; in short, investors do poorly enough on their own. similarly, individual item responses for investors are as much as twice as accurate as those for non-investors, but performance on some questions such as fee basis and load were extremely poor even for the more experienced group. about three quarters of the more experienced group answered these questions wrong as compared with nearly ninety percentage of the non-investor group. as with high financial literacy respondents, investors did modestly well on risk questions, but overall tended to do poorly on fee questions. fig. 4. fee calculation question. 46 b. scholl and a. fontes / financial services review 30 (2022) 31–56 4.3. index development while our 11 question battery was designed to test specific technical knowledge of key features of mutual funds, it is conceivable that the questions are really capturing a smaller set of underlying latent aspects of respondent knowledge. this might make some questions redundant. after all, if a respondent does not know that mutual funds are risky financial investments, they may answer both the risk stockfund and risk bondfund questions incorrectly so that one of these questions might be eliminated. following devellis (2016), to identify latent components, we conducted factor analysis with the 11 individual items (presented in fig. 1). to determine the number of factors, or latent variables, present in the data, we investigated both the eigenvalues of factors identified. following the standard method identified in kaiser (1960, 1970), factors with eigenvalues greater than one were retained for analysis. our analysis identified two factors with eigenvalues over 1: one corresponding to the market alternatives and risk categories, and a second corresponding to the performance history and fees categories (see figure 3). performance history loaded on to the fees factor, although this was the weakest loading of any item. we calculated item-total correlations for each item and the total index score and each item with its corresponding total factor score (presented in table 3). item-total correlations suggested strong relationships between each of the individual items and the overall index score, with the correlations ranging between 0.44 and .70. the correlation between individual items and their respective factors is quite strong, with a range between 0.59 and .76. these results argue in favor of preserving all 11 items in the index.16 4.4. index validation to demonstrate construct validity, we use descriptive regressions to examine the relationship between the index and selected outcome measures to gain more perspective on the explanatory power of the index (devellis, 2016).17 results are presented in table 4. each outcome we consider has a widening sphere of influence: as a measure of direct application, table 3 item-total and item-factor total correlations item index total correlation factor 1 correlation (market alternatives) factor 2 correlation (fees) 2 0.7036 0.7553 3 0.6259 0.7182 4 0.6235 0.6898 5 0.6200 0.6442 9 0.6693 0.7023 10 0.7054 0.7686 1 0.5086 0.6378 6 0.5235 0.6377 7 0.4504 0.5978 8 0.5316 0.6663 11 0.4410 0.6020 b. scholl and a. fontes / financial services review 30 (2022) 31–56 47 we modeled response to a fee calculation skill question (column 1) (see figure 4); we examined the value of the index in explaining ownership of any financial investments (“investors”; column 2); as our broadest outcome of interest, we examine the relationship between the index and the cfpb’s financial well-being (fwb) metric (see consumer financial protection bureau, 2017; column 3). for ease of demonstrating the complementarity between overall financial literacy and mutual fund knowledge, the sum of correct questions for each were normalized. for ease of exposition, outcomes were modeled using a linear probability model (lpm) or ordinary least squares (ols), as appropriate.18 covariates included were: gender, age, education, income, race/ethnicity, net worth, and general financial literacy (see table 5). 4.5. fee calculation skill we presented survey respondents with information showing a hypothetical fund fee presentation, and response options that offered a total fee calculation with a rationale for each response. respondents were asked to identify the response option that correctly approximated the amount of fees paid to the fund’s management company in a year. we developed this question to eliminate challenges associated with numerical ability, while at the same table 4 descriptive regression estimates fee calculation (1) investor (2) fwb (3) mutual fund knowledge (sd) 0.104*** (0.008) 0.053*** (0.006) 0.753*** (0.238) financial literacy (sd) 0.021** (0.009) 0.001 (0.007) 0.497* (0.264) female �0.043*** (0.015) 0.022** (0.011) �0.554 (0.416) age 30–44 0.008 (0.036) �0.037 (0.027) �4.023*** (1.034) age 45–59 �0.046 (0.036) �0.071*** (0.027) �4.747*** (1.039) age 60+ �0.068* (0.036) �0.096*** (0.027) 3.060*** (1.026) high school graduate or equivalent �0.038 (0.070) 0.064 (0.052) �4.132** (1.995) some college �0.027 (0.067) 0.089* (0.050) �3.582* (1.921) ba or above 0.017 (0.067) 0.124** (0.050) �1.886 (1.926) income (usd 35,000–59,999) �0.004 (0.024) 0.192*** (0.018) 2.884*** (0.688) income (usd 60,000–99,999) 0.012 (0.022) 0.259*** (0.017) 6.591*** (0.632) income (usd 100,000+) 0.032 (0.022) 0.273*** (0.017) 8.910*** (0.638) black 0.021 (0.037) 0.046* (0.028) 1.372 (1.054) hispanic �0.068** (0.033) �0.014 (0.024) �0.729 (0.931) race other 0.030 (0.026) �0.053*** (0.019) �2.378*** (0.733) net worth (breakeven) �0.008 (0.030) 0.030 (0.022) 3.056*** (0.850) net worth (positive) 0.036 (0.023) 0.138*** (0.017) 9.510*** (0.647) constant 0.233*** (0.076) 0.499*** (0.056) 47.412*** (2.158) observations 3,444 3,444 3,444 r2 0.100 0.198 0.294 adjusted r2 0.096 0.194 0.291 residual se (df = 3,426) 0.402 0.301 11.490 f statistic (df = 17; 3,426) 22.438*** 49.606*** 83.951*** note: fwb = cfpb’s financial well-being. *p < 0.1, **p < 0.05, ***p < 0.01. 48 b. scholl and a. fontes / financial services review 30 (2022) 31–56 time requiring knowledge of fee calculation and approximate valuation.19 only a single option was correct. this question proximally relates to optimal decision making in mutual fund investment contexts since fees directly affect net fund returns (e.g., see elton et al., 1993; elton & gruber, 2013; elton et al., 2011; gruber, 1996). nevertheless, this question has proved challenging for most respondents; in initial trials, we found that only 20% of survey participants correctly answered this question (scholl & fontes, 2019), and in the current survey only 21.6% answered correctly. our model suggests that higher scores on our index are associated with correct calculation of the mutual fund fee. a 1 sd increase in mutual fund knowledge scores corresponds to an increase in the likelihood of correctly answering the question by 10.4 percentage points (table 4, column 1). the complementary between our knowledge index and general financial literacy is highlighted by the fact that the coefficient on our knowledge score is strong and significant when general financial literacy and other covariates are included in the regression model. in terms of relative explanatory impact, the coefficient on mutual fund knowledge was roughly five times that of general financial literacy (0.02). the adjusted r2 table 5 specification of co-variates co-variate specification gender (male, or not) age 18–29 (omitted) 30–44 45–59 60+ race/ethnicity white non-hispanic (omitted) african american non-hispanic hispanic other education no high school diploma (omitted) hs graduate or equivalent some college ba or above income less than $35,000 (omitted) $35000–$59,999 $60,000–$99,999 $100,000 or more net worth in debt (omitted) zero greater than zero fee calculation (correct, or not) financial well-being top 40% of scores financial literacy score (0–3) low literacy = 0 correct high literacy = 3 correct b. scholl and a. fontes / financial services review 30 (2022) 31–56 49 for the regression is 0.096, and analysis of variance (anova) analysis suggests that 8.9% of the total sum of squares is explained by mutual fund knowledge, relative to 0.004% explained by variation in general financial literacy (mean squared error of 49.8 vs. 2.3; fstatistic of 308.45 and 14.45, respectively). these results suggest that the mutual fund knowledge index rank orders respondents by fee calculation skill proficiency, which is highly important in investment decision making. while general financial literacy remains an important factor in predicting a correct response to the fee calculation question, the much larger relationship with mutual fund literacy supports the idea that a specific measure of knowledge related to mutual funds is more relevant to modeling decision making. furthermore, controlling for other demographic characteristics, only with a perfect score on our knowledge index is an individual more likely to correctly answer our fee calculation question than to get it wrong—only 1.3% of respondents scored this high. 4.6. investor status (investor participation) if our index is appropriately measuring mutual fund knowledge, we expect to find that higher scores are associated with a higher likelihood of owning financial securities investments. as per scholl and hung (2018), mutual funds overwhelmingly dominate the composition of investors’ investment holdings. mutual fund owners are likely to have more experience with the products, resulting in higher knowledge scores, while those with less knowledge may be disinclined to purchase mutual funds. as presented in table 4 (column 2), a 1 sd increase in our knowledge score is related to a 5.3 percentage point increase in the probability of owning financial investments (p-value < 0.001). surprisingly, unlike our descriptive model predicting for the fee calculation question, general financial literacy did not explain variation in investor participation. the r2 for the regression is strong at 0.198. anova results suggest mutual fund knowledge alone explains approximately 7.5% of the variation in participation. these findings are somewhat surprising in that general financial literacy has been used to explain broad financial outcomes in a variety of contexts and our measure of participation is not simply a measure of ownership of mutual funds or exchange traded funds, but rather of financial investments overall. results support the idea that while general financial literacy is important in measuring many financial behaviors, a more targeted measure of mutual fund literacy may be important in understanding investment behavior. 4.7. financial well-being using the cfpb’s fwb score allows us to investigate the relationship between the mutual fund knowledge index and a much broader measure of overall financial wellness. the linkage between mutual fund knowledge and financial well-being is less direct than our prior two outcomes. as in the case of a household’s overall net worth, financial investments may only represent a portion of a household’s overall financial well-being. rent and mortgages, debt, a family’s employment situation and life circumstances all arguably play a larger role 50 b. scholl and a. fontes / financial services review 30 (2022) 31–56 than financial investments for most families; consumer financial protection bureau (2017) notes, in particular, that liquid savings provided the biggest differentiation between respondents with different levels of fwb. nevertheless, our knowledge measures may be a better proxy than other measures for latent knowledge components that are important to overall financial health. we found a positive relationship between the index score and increased financial wellbeing. table 4 column 3 reports that a 1 sd increase in our mutual fund knowledge index equates to an increase in fwb of 0.753, and the coefficient is highly significant despite the presence of several other potentially important covariates such as general financial literacy, age, income, and net worth. the point estimate on mutual fund knowledge is somewhat modest given the standard deviation of fwb in our sample is 13.6, but the r2 for this model is overall 0.29. anova results indicate that the knowledge index alone explains 8.7% of overall variation in fwb, with a mean sum of squares 2.3 times higher than the next most consequential covariate (net worth). of note, the coefficient on general financial literacy was not quite significant at the 95% confidence level (p = .06) in our regression. 4.8. conclusion and discussion we developed and deployed an 11-question index of mutual fund knowledge questions that is easy to deliver and has relatively low respondent burden, and fielded the module with a large, nationally representative, address-based probability sample to obtain credible population estimates of mutual fund knowledge. we developed the questions to reflect varying degrees of difficulty in mutual fund subject matter; all questions represented important choice-relevant topics in mutual fund selection and features that regulatory bodies routinely ascribe as important features for the investing public to consider when selecting investments. we refined the index with qualitative interviewing and extensive expert input, factor analysis and descriptive regression validation. while our results indicate the index is helpful in explaining important overall financial well-being, investor participation, and the highly important fee calculation skill, respondent performance on this battery is worrisome. a substantial fraction of respondents were no more accurate in their responses than if they had guessed at random and many respondents were unable to accurately answer a single true-false question. our estimates suggest that only the top 11.4 respondents could achieve a score of nine or higher, which we broadly consider a high level of knowledge. only the top 1.3% of respondents with a perfect score were more likely to correctly answer our fee calculation question than to get it wrong. in the context of the secular shift from defined benefit to defined contribution retirement plans in the united states, mutual funds in principle offer cheap diversification opportunities for most investors. troublingly, we find that about eighty percentage of our respondents probably do not understand enough about mutual funds to make informed choices. in the context of regulatory and disclosure efficacy, our index provides context to the realities of the regulatory environment. the regulatory environment puts the onus of investment selection to investors; our battery of questions tests knowledge we identified as crucial to avoiding poor investment selection from the pool of available mutual funds and etfs. as b. scholl and a. fontes / financial services review 30 (2022) 31–56 51 discussed, poor selection within this class of investments can have severe consequences for investors. even for subgroups one would a priori expect to perform better (e.g., investors and high general financial literacy individuals), knowledge of key fund characteristics is not very robust. while these higher literacy and experience groups performed reasonably well on our market options and risk questions, they performed extremely poorly on questions about fees and the relationship of past and future performance. the limited understanding of fees in particular is perhaps the most worrisome finding in that fees and expenses are widely viewed as perhaps the single most important aspect of the investor’s investment decision. it also suggests that financial intermediaries, the educational system, and regulators are not doing enough to prepare people to make decisions crucial to their own well-being; or alternatively, it could indicate that the investment marketplace itself is simply too complex for broad segments of the population to navigate successfully. other research, including scholl (2020) and a related set of experiments, provide evidence that more sophisticated investors have higher mutual fund knowledge (see chin, scholl, & vanepps, 2021). general financial literacy correlates with higher mutual fund knowledge (r = 0.44), but nearly one-third of high financial literacy individuals still did worse on the question battery than the expected value from answering randomly. this highlights the fact that our knowledge index can serve to supplement general financial literacy measures in selected contexts researchers in understanding decision-making pathologies, and policymakers in terms of assessing population vulnerabilities and checking assumptions about baseline investor knowledge. we believe financial regulatory authorities’ disclosure objectives relate to disclosures that help investors make informed investment decisions in investment contexts. our results suggest that in the context of inhibiting informed decision making by investors, the availability of information may be less important than investors’ (in)ability to interpret it. notes 1 this article will discuss knowledge of properties largely common to mutual funds and exchange traded funds (etfs). henceforth; we will simply refer to these as “mutual funds.” while there are technical differences between the two (e.g., how they are traded), these differences are of little consequence to our context here). 2 unless otherwise noted, facts described here are documented in (scholl & hung, 2018), including facts about ownership of mutual funds and etfs, account types, and the prevalence of funds in educational and retirement accounts. 3 direct communication with investors is only one purpose for disclosures, and disclosures are vital to helping markets to operate in other ways. for example, they also communicate critical information to financial professionals who may serve as intermediaries for the investors. 4 funds fall into the purview of other regulators in certain contexts. for example, the department of labor has jurisdiction over certain types of accounts that typically are comprised of mutual funds and exchange traded funds. 52 b. scholl and a. fontes / financial services review 30 (2022) 31–56 5 gruber (1996), elton et al. (1993), elton and gruber (2013), elton et al. (2011). 6 see, for example, the recommendation of the sec’s investor advisory committee on fee disclosures: https://www.sec.gov/spotlight/investor-advisory-committee2012/recommendation-mf-fee-disclosure-041916.pdf. 7 scholl and fontes (2019, 2020) provide subgroup analysis in greater depth. 8 alexander, jones, and nigro (1998) only sample individuals in mutual fund owning households. 9 technical documentation for amerispeak is available at: https://amerispeak.norc.org/ documents/research/amerispeak%20technical%20overview%202019%2002%2018. pdf. a list of publications using amerispeak data can be found at: https://amerispeak. norc.org/research/pages/default.aspx. 10 norc’s national frame is used for the amerispeak panel, as well as other federal surveys including the survey of consumer finances and the general social survey. 11 results presented in this article do not use survey weights that would bring these demographics more in line with the u.s. population. 12 the index presented in this article treats skipped questions and incorrect answers as incorrect answers. in previous work, we also constructed a penalty-adjusted version of the score, which penalized incorrect answers. the two scores are highly correlated, with the penalty-adjusted scores providing more separation between subgroups, but are overall highly correlated. many of the additional questions that were proposed are discussed in scholl and fontes (2019). 13 as of 1996, states do not technically register mutual funds, but they do collect certain fees associated with mutual fund filings. the intent of this question was to measure whether investors believe that the regulatory environment prevents the loss of money, not to test their knowledge of the specific responsibilities of individual regulators or the division of state/federal roles. this question phrasing was adopted because it was deemed possible that individuals might not know how to distinguish between state and federal powers and roles as they have evolved over time. testing did not reveal any particular focus on the first clause in the question (example regulators), respondents’ attention appeared to have been placed on the second clause (whether or not you “cannot lose money”). 14 “past performance does not guarantee future results.” 15 this result may suggest that mutual fund warning labels are ambiguous, ineffective, ill placed or not understood, but it could also reflect the fact that investors struggle to identify specific characteristics that help determine how a mutual fund will perform and are left to contemplate past performance when no other alternative discernment features present themselves. we leave interpretation to additional research. 16 “don’t know” responses are treated as incorrect responses. we used the oblimin rotation (see clarkson & jennrich, 1988). correlation between the two factors was �0.55. 17 see also joint research centre-european commission (2008) and consumer financial protection bureau (2017). b. scholl and a. fontes / financial services review 30 (2022) 31–56 53 18 lpm estimates provided for convenience. we also estimated models using logistic regression (not reported); these were nearly identical when marginal probabilities were calculated. 19 additional specifications on the fee calculation question (not presented) controlled for numeracy and survey design effects, but were virtually identical to those reported. we preferred to keep specifications largely consistent across our three outcome variables. references agnew, j. r., & szykman, l. r. (2005). asset allocation and information overload: the influence of information display, asset choice, and investor experience. journal of behavioral finance, 6(2), 57–70. alexander, g. j., jones, j. d., & nigro, p. j. (1998). mutual fund shareholders: characteristics, investor knowledge, and sources of information. financial services review, 7, 301–316. barber, b. m., odean, t., & zheng, l. (2005). out of sight, out of mind: the effects of expenses on mutual fund flows. the journal of business, 78, 2095–2120. beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2018). behavioral household finance. in d. bernheim, d. laibson, & s. dellavigna (eds.), handbook of behavioral economics: applications and foundations 1 (vol. 1, pp. 177–276). amsterdam: elsevier. beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2011). how does simplified disclosure affect individuals’ mutual fund choices? in d. a. wise (ed.), explorations in the economics of aging (pp. 75–96). chicago, il: university of chicago. bhattacharya, u., loos, b., meyer, s., & hackethal, a. (2017). abusing etfs. review of finance, 21, 1217– 1250. brown, s. j., & goetzmann, w. n. (1995). performance persistence. the journal of finance, 50, 679–698. calvet, l. e., campbell, j. y., & sodini, p. (2007). down or out: assessing the welfare costs of household investment mistakes. journal of political economy, 115, 707–747. carhart, m. m. (1997). on persistence in mutual fund performance. the journal of finance, 52, 57–82. carvalho, l., & silverman, d. (2019). complexity and sophistication. nber working papers 26036. cambridge, ma: national bureau of economic research, inc. chin, a., scholl, b., & vanepps, e. (2021). jargon in fund fee disclosures. office of the investor advocate working paper, 2021–2001. washington, dc: office of the investor advocate. choi, j. j., laibson, d., & madrian, b. c. (2010). why does the law of one price fail? an experiment on index mutual funds. the review of financial studies, 23, 1405–1432. clarkson, d. b., & jennrich, r. i. (1988). quartic rotation criteria and algorithms. psychometrika, 53, 251–259. consumer financial protection bureau (cfpb). (2017). financial well-being scale: scale development technical report. available at https://files.consumerfinance.gov/f/documents/201705_cfpb_financial-wellbeing-scale-technical-report.pdf dehaan, e., song, y., xie, c., & zhu, c. (2021). obfuscation in mutual funds. ssrn electronic journal, 72(2/3), 2010. davoli, m., & rodrı́guez-planas, n. (2020). culture and adult financial literacy: evidence from the united states. economics of education review, 78, 102013. devellis, r. f. (2016). scale development: theory and applications (vol. 26). thousand oaks, ca: sage publications. elan, s. (2011). financial literacy among retail investors in the united states. washington, dc: library of congress federal research division. elton, e. j., & gruber, m. j. (2013). mutual funds. in g. m. constantinides, m. harris, & r. m. stulz (ed.), handbook of the economics of finance (vol. 2, pp. 1011–1061). amsterdam: elsevier. 54 b. scholl and a. fontes / financial services review 30 (2022) 31–56 elton, e. j., gruber, m. j., & busse, j. a. (2011). are investors rational? choices among index funds. the journal of finance, 59, 261–288. elton, e. j., gruber, m. j., das, s., & hlavka, m. (1993). efficiency with costly information: a reinterpretation of evidence from managed portfolios. review of financial studies, 6, 1–22. fernandes, d., lynch, j. g. , jr., , & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors.management science, 60, 1861–1883. fontes, a., jain, a., zhu, m., & dwan, k. (2019). collecting household financial data online: a comparison of the financial well-being and literacy survey with federal benchmarks. presented at the cipher conference. washington, dc. forbes, j., & kara, s. m. (2010). confidence mediates how investment knowledge influences investing self-efficacy. journal of economic psychology, 31, 435–443. goetzmann, w. n., & ibbotson, r. g. (1994). do winners repeat? the journal of portfolio management, 20, 9–18. hastings, j. s., madrian, b. c., & skimmyhorn, w. l. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5, 347–373.e. [pmc] hortaçsu, a., & syverson, c. (2004). product differentiation, search costs, and competition in the mutual fund industry: a case study of s&p 500 index funds. the quarterly journal of economics, 119, 403–456. ici. (2010). ici research report: profile of mutual fund shareholders, 2010. washington, dc: investment company institute. ici. (2018). ici research report: profile of mutual fund shareholders, 2018. washington, dc: investment company institute. joint research centre-european commission. (2008). handbook on constructing composite indicators: methodology and user guide. paris: oecd publishing. kahraman, b. (2021). fee complexity and investor mistakes in retail financial markets. quarterly journal of finance, 11(1), 2150004. kaiser, h. f. (1960). the application of electronic computers to factor analysis. educational and psychological measurement, 20, 141–151. kaiser, h. f. (1970). a second-generation little jiffy. psychometrika, 35, 401–415. kaiser, t., & menkhoff, l. (2017). does financial education impact financial literacy and financial behavior, and if so, when? the world bank economic review, 31, 611–630. kaiser, t., & menkhoff, l. (2020). financial education in schools: a meta-analysis of experimental studies. economics of education review, 78, 101930. lin, j. t., bumcrot, c., ulicny, t., mottola, g., walsh, g., ganem, r., kieffer, c., & lusardi, a. (2019). the state of u. s. financial capability: the 2018 national financial capability study.washington, dc: finra. lusardi, a. (2008). financial literacy: an essential tool for informed consumer choice? cambridge, ma: national bureau of economic research. lusardi, a. (2011). financial literacy and planning: implications for retirement well-being. in o. s. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 17–39). cambridge, ma: oxford university press. lusardi, a. (2012). numeracy, financial literacy, and financial decision making. cambridge, ma: national bureau of economic research. lusardi, a. (2019). financial literacy and the need for financial education: evidence and implications. swiss journal of economics and statistics, 155, 1.e. lusardi, a., michaud, p. c., & mitchell, o. s. (2020). assessing the impact of financial education programs: a quantitative model. economics of education review, 78, 101899. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. lusardi, a., & mitchell, o. s. (2008). planning and financial literacy: how do women fare? american economic review, 98, 413–417. lusardi, a., mitchell, o. s., & curto, v. (2014). financial literacy and financial sophistication in the older population. journal of pension economics & finance, 13, 347–366. b. scholl and a. fontes / financial services review 30 (2022) 31–56 55 lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and overindebtedness. nber working papers 14808. washington, dc: national bureau of economic research, inc. lusardi, a., & tufano, p. (2015). debt literacy, financial experiences, and overindebtedness. journal of pension economics and finance, 14, 332–368. malkiel, b. g. (1995). returns from investing in equity mutual funds 1971 to 1991. the journal of finance, 50, 549–572. muller, l. a., & turner, j. a. (2021). financial literacy, the “high-fee puzzle,” and knowledge about the importance of fees. the journal of retirement, 8, 29–38. müller, s., & weber, m. (2010). financial literacy and mutual fund investments: who buys actively managed funds? schmalenbach business review, 62, 126–153. scholl, b. (2020). choice set complexity. working paper. washington, dc: sec office of the investor advocate. scholl, b., & fontes, a. (2019). what does the public know about mutual funds? working paper. washington, dc: sec office of the investor advocate. scholl, b., & hung, a. (2018). the retail market for investment advice. washington, dc: office of the investor advocate. available at https://www.sec.gov/files/retail-market-for-investment-advice.pdf scholl, b., silverman, d., & enriquez, m. (2020). text complexity and performance. santa monica, ca: rand befi presentation. scholl, b., silverman, d., & enriquez, m. (2021). disclosure complexity and fund performance. working paper. washington, dc: sec office of the investor advocate. urban, c., schmeiser, m., collins, j. m., & brown, a. (2020). the effects of high school personal financial education policies on financial behavior. economics of education review, 78, 10178. u. s. securities & exchange commission. (2012). study regarding financial literacy among investors. washington, dc: us security and exchange commission. available at https://www.sec.gov/news/studies/ 2012/917-financial-literacy-study-part1.pdf van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. 56 b. scholl and a. fontes / financial services review 30 (2022) 31–56 from the editor this issue contains volume 27 issue 4 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “the importance of debt for household risky asset allocation and portfolio structure” is coauthored by ran tao and yuan yuan, both at university of wisconsin, whitewater. the authors examine the joint behavior of debt and financial asset portfolio decisions. they test the relationship between debt structure and asset allocation and then estimate the determinants of debt structure and asset allocation simultaneously. they find evidence that debt structure affects households’ risky asset allocation decisions and identify the demographic and financial factors that can contribute to the household overall financial portfolio structure. the second article “the impact of the tax cuts and jobs act oon ira choice for moderate income investors, is coauthored by timothy manuel, lee tangedahl, and kent swift, all at university of montana, missoula. while he traditional ira remains the more popular choice of ira among investors and the passage of the 2018 tax cut and jobs act makes the traditional ira a better choice for many moderate income individuals who will be in the 22% marginal tax bracket during their working years. the authors compute breakeven rates of return on investment where the roth is preferred to the traditional are higher with the new tax law. the utilize a spreadsheet model for investors who do not invest the tax savings and find that in this case the traditional ira can be a better choice for many moderate income individuals. the third article, “local creative culture and dividend policy” is coauthored by erdem ucar and arsenio staer, both at california state university fullerton. the authors examine the role of local risk-taking propensity on dividend demand by using local creative culture as a measure of local risk-taking. they find that firms located in areas with a strong creative culture are less likely to pay and initiate dividends and exhibit lower levels of dividend yield. the paper also highlights the local component of corporate dividend policies and offers additional evidence supporting dividend catering theory. they indicate that these results underscore the importance of cultural determinants of investors’ risk-taking for financial industry participants. financial services review 27 (2018) v–vi 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. the fourth article, “financial capability: literacy, behavior, and distress” is coauthored by janine k. scott, at shepherd university, nghia nguyen at massey university, yuanshan cheng at winthrop university, and philip gibson at winthrop university. in this paper the authors investigate the influence of individual financial knowledge and financial behavior on the probability of experiencing financial distress. they examine three measures of financial distress related to bill payment, retirement saving, and being late with a mortgage payment. they construct financial literacy and financial behavior indices using questions from the survey that pertain to financial knowledge and financial decision-making. they conclude that financial literacy and positive behavior reduces financial distress stemming from simple financial matters. although they find the opposite result for more complex financial decisions. the final article, “risk and reward of fractionally-leveraged etfs in a stock/bond portfolio” is authored by james dilellio at pepperdine university. the author investigates using 1.25x leveraged stock and bond exchange-traded funds (etfs) as an asset allocation strategy. he analyzes performance by replicating funds from 1989–2017 and conducts simulations to assess performance under a variety of market conditions, and demonstrate opportunities for excess returns over unlevered funds in a 60/40 stock/bond allocation. these results are accomplished with a small reduction in the sharpe ratio and no need to access margin. he concludes that this asset allocation strategy could be well suited for investors more interested in total returns during upward trending markets. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 27 (2018) v–vi focusing on both sides of the balance sheet: the potential benefit of liability management zhikun liua,*, david m. blanchettb aempower retirement, 8515 e. orchard road, 4t2, greenwood village, co 80111, united states bmorningstar investment management llc, 22 w. washington street, chicago, il 60602, united states abstract debt has become a significant issue among u.s. households with average household interest payments on liabilities exceeding expected returns on investment assets by more than 50%. in this study, we explore the role of u.s. household debt and analyze the impact of different economic, demographic, and behavioral factors on household borrowing decisions, with a particular focus on “good” and “bad” debts, which depend on type and interest rate. we estimate significant potential benefits with improved liability management and find that households with lower asset, income, and education levels are likely to benefit most from assistance with debt optimization. © 2021 academy of financial services. all rights reserved. jel classification: d12; g4; d15 keywords: debt management; retirement financial planning; behavioral finance; financial decision making 1. introduction debt is an increasingly significant part of the u.s. household balance sheet. after the 2007–2009 economic recession, debt levels of american households have increased significantly (bricker et al., 2017). the total u.s. household indebtedness was approximately $14.27 trillion as of june 30, 2020, according to the federal reserve bank of new york. this is higher than the previous peak of $12.68 trillion in the third quarter of 2008 (adjusted to 2019 dollars) and has increased by 27.9% since the second quarter of 2013 (federal reserve bank of new york, 2020). additional information on this effect is shown in fig. 1. *corresponding author: tel.: +1-303-737-6207; fax: +1-303-737-6544. email address: zhikun.liu@empower-retirement.com (z. liu) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 121–145 financial firms and advisors tend to spend significantly more time focusing on the assets side of the household balance sheet compared with the liability side. this focus is consistent with the traditional skill set of financial advisors—building portfolios—and reflects how they are typically compensated (e.g., as a percentage of assets under management). however, in this study, we demonstrate that this predominant attention paid to the assets does not necessarily reflect the economic importance within the context of the household’s entire balance sheet (i.e., when liabilities are taken into consideration). for example, data from the 2016 survey of consumer finances (scf) suggest that among “low-to-affluent” u.s. households, the total interest payments on debts exceed the expected gains from their financial assets.1 therefore, spending time on “debt optimization” is likely to result in better outcomes than focusing on assets alone. in this article, we explore the composition of household balance sheets in the united states to understand the potential benefits associated with making more intelligent debt decisions. consistent with past research, we find that certain types of “bad” debts, such as credit cards, are relatively common on household balance sheets today despite their high interest rates (averaging approximately 15%).2 it is not clear to what extent interest rates could be lower had the household done more due diligence on its debt decisions, or the extent to which these debts can be refinanced, but it is likely that some, and possibly many, households’ situations can be improved (i.e., the household could reduce the interest rate on outstanding debt). this analysis suggests more work should be done to understand the potential benefits of improving household credit decisions. the objective of this study is to demonstrate the urgency, importance, and potential impact of household liability management by answering the following questions: what is the current financial situation and retirement outlook of low-to-affluent u.s. households? what factors are associated with household debts and leverage ratios? what is the difference between “good” and “bad” debts?3 will the attributes related to households carrying different types of debts be similar? what kinds of families are more likely to have higher average debt interest rates and how much could they save by accessing liability optimization? fig. 1. growth trends in u.s. consumer credit owned. source: federal reserve board ny 2020 consumer credit panel/equifax. 122 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 2. literature review using data from multiple waves of the scf, emmons and noeth (2013) report that the household leverage ratio, defined by the sum of total debts divided by total assets, is higher among younger families. also using scf data, barba and pivetti (2009) demonstrate that the rising household indebtedness is associated with a decrease in the household savings rate. this phenomenon is partially explained by lagging real wage growth and the tendency for u.s. households to sustain their relative consumption level. based on data from consumer finance monthly, jiang and dunn (2013) show that younger consumers have higher levels of credit card debt and are repaying that debt at lower rates than previous generations.4 using the health and retirement study (hrs), gustman, steinmeier, and tabatabai (2011) find that relative debt levels have been increasing for households that are near retirement since the 2007–2009 recession and that much of the growth in debt appears to be related to mortgage and housing expenses. the quarterly report on household debt and credit for the second quarter of 2020 supports this finding, reporting $9.78 trillion of mortgage balances for u.s. households (federal reserve bank of new york, 2020). using the hrs, lee, lown, and sharpe (2007) study the dissaving behavior of older americans and point out that financial debt carried into later life may result in reduced access to essential health care, restrictions on activities, and delayed retirement. among the different categories of household liabilities, high-interest debts such as consumer revolving credit debts can have significant negative impacts on household balance sheets and cash flows. based on information from the federal reserve bank of new york (2019), credit card balances stood at $870 billion as of the last quarter of 2018, with a seasonally adjusted annual growth rate of 3%. auto loan originations reached the highest amount in the 19-year recorded history of the new york fed in 2018, amounting to $584 billion. unlike certain good debts, which tend to have lower relative interest rates and are typically used to purchase assets that are expected to generate long-term income or grow in value (e.g., mortgages), bad debts such as credit cards, payday loans, and some auto loans typically have higher interest rates and are generally associated with purchases (and assets) that do not generate positive long-term returns (hanson, 2006). in other words, the cost of the good debts can often be outweighed by their potential long-term benefits, while the bad debts’ high interest costs typically have little-to-zero long-term returns. bad debts are not only expensive, but they may also negatively influence the borrowers’ credit scores, hinder their financial and retirement goals, and even cause stress and health issues. davies, montgomerie, and wallin (2015) report a positive relationship between individuals who are deeply in debt and those who report mental health problems such as depression and physical illness. behavioral studies also indicate that consumers may be more likely to accumulate a larger revolving credit card balance if they frequently pay behind schedule or miss payments (kim & devaney, 2001; wärneryd, 1999). therefore, helping consumers stay away from “bad” debt and coaching them to develop good borrowing and accumulation habits are essential approaches for advisors and financial planning firms to support their clients’ liability management. z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 123 this effect, in which households spend more on their debt than they earn on their savings, is likely to continue in the future given the growth in debt among american households, as noted in fig. 2 therefore, it is essential for financial planning firms and advisors to start putting a greater emphasis on their clients’ debt structures and help them better manage their liabilities in order to help ensure that they can achieve a successful retirement. zinman (2015) notes that research on the household debt has significantly lagged its sister literature on the asset side of the balance sheet. while one may assume that households make rational decisions regarding debt, stango and zinman (2016) find that cross-consumer dispersion in credit card borrowing costs remains substantial even after controlling for debt levels, credit risk, and product characteristics. while the share of u.s. households with debt has been relatively constant, ranging from 72.3% in 1989 to 77.1% in 2016 (bricker et al., 2017), the mean value of debt for american families has increased significantly, from $66,900 in 1989 (in 2016 dollars) to $123,400 in 2016. this magnitude of debt increase has been observed across age levels. within the 2016 scf survey wave, the percentage of households carrying debt peaked around middle age (approximately 45 years old), with the most common debt categories being mortgages, credit card debts, auto loans, and student loans, as noted in fig. 3. not surprisingly, interest rates differ significantly across different types of loans. in fig. 4 we provide context regarding the distribution of interest rates for households by loan type, again using 2016 scf data. fig. 4 shows that unsecured personal loans (such as credit card loans and other consumer loans) typically have the highest interest rates. these loans are also typically categorized as bad debts because they are not used to purchase assets that improve the long-term financial condition of the household and rather are used to purchase items that are more consumptionbased in nature. fig. 5 jointly illustrates the prevalence of different loan types and the median interest rates among the households in which the head-of-household is 45 years old. while median interest rates are relatively static across ages, age 45 is selected as the representative age because it is the approximate peak age for indebtedness, as previously noted in fig. 3. our study explores the urgency and importance of liability management for american households. the article consists of the following sections: first, this study utilizes scf data to develop a general picture of u.s. households’ financial situations in terms of their balance sheet characteristics. second, we review the liability side of households’ balance sheets to investigate the prevalence of different types of consumer debts and the interest rates associated with them. third, we analyze a number of economic and demographic factors that are associated with household debts. after exploring the attributes that potentially relate to the households carrying bad debts, we then identify the characteristics of households that have higher average debt interest rates. finally, we demonstrate the impact of liability management in terms of investment alpha-equivalent (“excess investment return”-equivalent) analysis and the potential dollar amount that can be saved through interest rate reduction relative to financial asset considerations. 124 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 3. theory the household consumption decision involves a trade-off between consuming more today (borrowing) and consuming more in the future (saving). the borrowing and saving behavior of households is largely driven by their intertemporal consumption choices, affected by their time-discounting preference, investment interest rates, and other factors. fig. 2. mean value of debt for u.s. families with debt holdings. source: federal reserve board survey of consumer finances (scf) bulletin 2017. notes: all respondents in the 2016 scf data are included in this graph. the age of the household is represented by the age of the household head. fig. 3. probability of a household having debt. source: federal reserve board survey of consumer finances 2016 survey wave. notes: weights applied. other consumer loans include loans for household appliances, furniture, hobby or recreational equipment, medical bills, friends or relatives, etc. this category does not include credit cards, margin loans, or loans against life insurance or pensions. z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 125 to better analyze the liability management of u.s. households, we structure our theoretical framework according to the life-cycle hypothesis (jappelli & pagano 1989; modigliani 1986), which holds that a household chooses a consumption path to maximize its lifetime utility fig. 4. distribution of household loan interest rates. source: federal reserve board survey of consumer finances 2016 survey wave. notes: weights applied. this figure shows the percentile distribution of interest rates across different types of loans. other consumer loans include loans for household appliances, furniture, hobby or recreational equipment, medical bills, friends or relatives, etc. this category does not include credit cards, margin loans, or loans against life insurance or pensions. fig. 5. loan prevalence and median interest rates. source: federal reserve board survey of consumer finances (scf) 2016 survey wave. notes: weights applied. this graph uses a subsample of 45-year-old respondents to illustrate the prevalence and median interest rates of different types of household debts in the scf data. 126 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 subject to an intertemporal budget constraint. we start with a simple two-period life-cycle model to understand the dynamic intertemporal choice issue. then we generalize this model to multiple periods to capture the households’ liability decisions for different life stages. in the two-period model, a household maximizes its utility described as following: u c1, c2ð þ ¼ u c1ð þ þ d u c2ð þ (1) where c1 and c2 are consumptions in periods 1 and 2, respectively.5 d is the discount factor that depicts the household’s time preference. the assumption of 0 < d < 1 illustrates the tendency that present consumption is always more preferable than future consumption. d is more close to 0 when the household is more future-discounting. if d is close to 1, the household has no preference between present and future consumptions. the two-period budget constraint that the household faces can be represented by the following inequalities: c1 þ s≤ y1 (2) c2 ≤ 1 þ rð þ s þ y2 (3) where y1 and y2 are the income of the household for period 1 and period 2, respectively. the borrowing/saving factor is symbolized by s. if s > 0, then the household saves in period 1. if s < 0, then this household borrows in period 1; thereby, forfeiting investment opportunities and reducing the consumption in period 2. r represents the prevailing interest rate in the financial markets. if s > 0, then r stands for the investment return from savings. if s < 0, then r can represent the interest charged for the debt the household borrows during period 1. substituting out the borrowing/saving factor s, we obtain the “lifetime budget constraint.” this constraint represents the fact that the discounted present value of all periods’ consumption must be less than or equal to the discounted present value of lifetime income: c1 þ c2 1 þ r ≤ y1 þ y2 1 þ r (4) now the household’s intertemporal consumption choice model can be rewritten as: max c1, c2f g u c1, c2ð þ ¼ u c1ð þ þ d u c2ð þ (5) s:t: c1 þ c2 1 þ r ≤ y1 þ y2 1 þ r using the lagrangian technique, the solution to this problem is: foc c1ð þ: u0 c1ð þ=l (6) z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 127 foc c2ð þ: b u0 c2ð þ= l 1 þ r (7) foc lð þ: c1 þ c2 1 þ r ≤ y1 þ y2 1 þ r (8) putting the first order conditions together, we arrive at the euler equation: u0 c1ð þ d u0 c2ð þ ¼ 1 þ rð þ (9) this equation describes the intertemporal optimal consumption choice between the current and future period: the marginal rate of substitution (appropriately discounted by d ) is equal to the gross interest rate, which represents the relative price between consumption in period 1 and consumption in period 2. in terms of saving (s > 0), if r is high, the price of consumption in period 1 is high because the household is forgoing a high interest rate of investment return. in the case of borrowing debt (s < 0), the interpretation still applies: if r is high, the price of consumption in period 1 is high because the household is paying a high borrowing cost due to the high interest rate. the euler equation implies that the household maximizes utility by smoothing the consumption path over the life cycle, which explains the borrowing behavior of the household. the two-period intertemporal consumption model can be generalized for multiple-period analysis. assume a household’s finite lifetime can be categorized into t different periods. in each period t, the household has income yt, saves or borrows st, and consumes ct. then the household’s intertemporal consumption choice model is as follows: max c1, c2, . . . , ctf g e o t=1 t d �tþ1u ctð þ " # (10) s:t: rt t=1 1 þ rð þ�tþ1 ct ≤rt t=1 1 þ rð þ�tþ1yt (11) where d is still the discounting factor measuring the households’ preference for present versus future, and r is the rate of return on the investment (or interest rate of borrowing on the debt). similarly, one can derive the solution to this problem and arrive at the generalized euler equation: et u 0 ctð þ d u0 ctþ1ð þ " # ¼ 1 þ rð þ (12) notice that the eq. (12) can be rearranged into: 128 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 et u 0 ctð þ u 0 ctþ1ð þ " # ¼ 1 þ rð þ d (13) then we can interpret the euler eq. (13) as the marginal rate of substitution between the period (t) and the next period (t þ 1), is equal to the product of the gross interest rate and the time discounting factor. in other words, the households smooth their lifetime consumption paths based on two factors, the interest rate (borrowing or investing) and their time discounting preference. there have been some variations to the life-cycle model since its development. for example, the presence of liquidity and borrowing constraints has been brought up to modify the model for better suitability to empirical analysis. in our analysis, we assumed that u.s. households are able to leverage from various lending sources to achieve their consumption smoothing and combine the liquidity and dollar amount borrowing constraints into the interest rate constraint (the household’s tolerance of high interest rates).7 we also consider households’ liquid assets in our analysis to investigate their debt problems. to capture the discounting preference of american families, we use the household’s financial planning horizon as a proxy in the empirical analysis. the household consumption decision model provides guidance on what to expect in the regression analysis results presented in this article. for instance, we expect to observe that interest rates significantly affect household leverages across various debt types. households with relatively longer financial planning horizons are less likely to carry debt (or they have lower debt amount, debt-to-income ratio, and debt-to-asset ratio) compared with the households whose financial planning horizons are short. liquid asset holdings should significantly reduce the household debt level. detailed discussion on the regression results will be presented in the following sections. 4. data and methodology this article uses data from the scf to analyze the characteristics of u.s. household finances. the scf, conducted by the federal reserve board, is a nationally representative crosssectional survey of u.s. households. this triennial survey collects a variety of information on income, balance sheet, and demographic characteristics from a selection of more than 6,000 american families in each survey wave. using the 2016 survey wave, we study the characteristics of the balance sheets of american households, explore the factors that are associated with high debt-to-asset ratios for certain households, and investigate the benefit of liability management for these households.8 for our analysis, we focus on “low-to-affluent” american families, which we define as households with less than $1 million in financial assets. households with very high net worth often have their own unique leveraging and investment strategies, and optimizing these strategies is beyond the scope of this article. the comprehensive perspective of the average household balance sheet (see appendix a) indicates that the average return on financial (i.e., investment) assets is approximately 62% of the debt interest charges for the average u.s. household. in other words, the average u.s. z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 129 family is spending more on interest servicing household debt than they are earning from investing their financial assets. this is despite a significant focus on managing the asset side of the household balance sheet that is common within the financial advising profession. the focus of this article is to explore how low-to-affluent american families can potentially benefit from debt restructuring and liability management with assistance from their financial planners and advisors. because of the nature of the scf data, which oversamples high-income households (aizcorbe, 2003; nielsen 2015), we apply sample weights to all the empirical analyses. in addition to focusing only on households with less than $1 million in financial assets, we also restricted the opportunity set to households whose head is between 20 and 85 years old and that had an annual family income of at least $1,000. after applying these restrictions, our analysis sample is reduced to 4,481 households (see appendix b for descriptive statistics of the analysis sample). because each household in the 2016 scf data has five implicates, the total number of observations in our analysis sample is 22,415.9 to cope with the dual-frame complex sample design and the multiple-imputation process of the scf data, this study use the “scfcombo” stata macro designed by nielsen (2015) to conduct our regression analyses.10 5. results and discussion the regression analyses used in this article follow these steps: first, we use probit and ordinary least squares (ols) regressions to study what factors are associated with household debt. we look at the economic, demographic, and behavioral factors that could potentially impact the likelihood of carrying household debt, the total debt amount, the debt-to-financial-asset ratio, and the debt-to-income ratio. second, we isolate what are frequently considered bad debts (represented by credit card debts) and compare them with debts that are typically viewed as good debts (represented by mortgages) to see whether the factors associated with different debt categories are similar. then, we utilize different interest rate measures to check the attributes that relate to high interest rates. finally, we perform alphaequivalent analyses and calculated the potential savings to demonstrate the impact of liability management and interest rate reduction from a financial asset perspective. detailed descriptions and summary statistics of the variables used in the regression analyses are presented in appendix b. table 1 presents the results of the probit and ols regressions to better understand what factors are associated with household debts. the dependent variables in these regressions include “whether the household carries debt,” “total debt amount,” “debt to financial asset ratio,” and “debt to income ratio.” the marginal effect results of the probit regression in table 1 provide a general picture of what factors are associated with low-to-affluent american families’ debt holdings. the ols regression demonstrates the impact on household debt amounts from each of these factors. in some cases, relatively wealthier families that are in good financial conditions still carry larger amount of debt due to their high income or sizeable financial asset accumulations. while some financially challenged families might not be carrying a sizable sum of debt in terms of dollar amounts, these debts are typically detrimental to their financial well-being compared with their income and asset levels. to consider these cases, we analyze the debt-to-financial-asset ratio and the debt-to130 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 income ratio in comparison with the analyses on the likelihood of having debts and the total debt amount. the intertemporal model discussed in the theory section above suggests that time discounting preference should affect households’ consumption smoothing behaviors significantly. therefore, we expect to see from the results in table 1 that households with longer financial planning horizons are less likely to carry debt, have lower debt-to-asset ratio as well as lower debt-to-income ratio. in addition, we expect to see a significant negative relationship between the households’ liquid asset levels and the likelihood of carrying debts, total debt amount as well as debt-to-income ratio. based on the results in table 1, married families and households with children are more likely to carry debts. families that own houses are much more likely to borrow, and the more real assets a family owns, the more likely this family is to carry debts. liquid assets and age are negatively related to the likelihood of having debts. this is most likely because households are less likely to borrow if they have enough liquid assets to cover their needs, which supports the advocacy of emergency savings through liquid accounts for the general public. older families are less likely to have debts because they generally have had a longer time to accumulate wealth and pay off their various household debts. it appears to be counterintuitive that education and income level, as well as reporting having savings, are positively related to carrying household debts. however, if we consider the ols results together with the marginal effects of the probit regression, the impact of these factors on household debts becomes clear. for instance, although high-income families are more likely to leverage and have larger debt sizes, their debt-to-income ratios are lower and negatively related to their income level. households that have savings demonstrate much lower debt-to-financial-asset ratios, despite the higher likelihood to borrow, with other variables such as liquid asset levels controlled. the combined results could indicate that these families may be more financially literate and leverage lower-interest debts to increase their investments in financial assets and savings. when it comes to education level, more educated households are more likely to carry debts, have higher debt balances, and have a higher debt-to-income ratio, keeping all other factors, such as income and assets, the same. this is a strong indication of the impact of student loans on these families. ceteris paribus, educated families are more likely to carry student loans compared with the less educated ones, because of the prevalence of student loans used to finance education today. a family’s financial planning horizon is also a strong behavioral indicator of household debts. households with longer financial planning horizons are much less likely to have debts. total debt amount, as well as debt-to-financial-assets ratio and debt-to-income ratio, are all negatively associated with a longer financial planning horizon. this finding supports the myopic planning hypothesis, which predicts that having a myopic financial planning horizon fuels households’ borrowing and may lead families deeper into debt. it also suggests that promoting long-term financial planning horizons serves as a good approach to help families with their liability management. “not all debt is created equal,” as the saying goes. while good debts are typically defined as those with lower interest rates that help households finance activities and purchases that provide long-term benefits (e.g., mortgages), bad debts are usually associated with higher interest rates and are used to purchase depreciating assets that do not generate long-term z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 131 t ab le 1 p ro b it an d o rd in ar y le as t sq u ar es (o l s ) re g re ss io n s o n fa ct o rs as so ci at ed w it h h o u se h o ld d eb ts v ar ia b le s p ro b it (m e )a o l s o l s o l s h av e d eb t t o ta l d eb t am o u n t ($ ) d eb tto -fi n an ci al -a ss et ra ti o b d eb tto -i n co m e ra ti o c m ar ri ed 0 .1 1 5 * * (0 .0 3 9 ) �8 1 9 .6 (3 ,5 7 7 .5 0 5 ) �1 1 9 .2 (1 8 6 .3 4 4 ) �0 .2 2 3 * (0 .1 0 2 ) n u m b er o f k id s 0 .0 7 8 7 * * * (0 .0 1 9 ) 5 ,4 3 3 .1 * * * (7 6 2 .8 2 6 ) 1 5 1 .1 (1 4 2 .5 8 7 ) 0 .0 1 7 2 (0 .0 2 5 ) e d u ca ti o n le v el 0 .0 5 9 3 * * * (0 .0 0 7 ) 2 ,8 3 0 .9 * * * (6 5 8 .9 6 3 ) �6 2 .0 1 (3 3 .7 5 1 ) 0 .1 0 2 * * * (0 .0 2 0 ) r ea l as se ts (p er $ 1 0 k ) 0 .0 1 0 9 * * * (0 .0 0 2 ) 4 ,3 1 4 .9 * * * (2 0 7 .0 3 0 ) 6 .7 6 7 (8 .7 6 0 ) 0 .0 4 1 6 * * * (0 .0 0 3 ) l iq u id as se ts (p er $ 1 0 k ) �0 .0 7 9 4 * * * (0 .0 1 0 ) �4 ,0 7 6 .7 * * * (0 .0 5 1 ) 7 .4 0 7 (0 .0 0 1 ) �0 .0 4 5 9 * * * (0 .0 0 0 ) h av e h o u se s 0 .5 8 3 * * * (0 .0 5 7 ) 6 ,7 1 1 .6 * * (2 ,5 0 9 .2 6 1 ) 1 2 3 .3 (2 5 3 .9 7 5 ) 0 .7 5 0 * * * (0 .0 6 9 ) h av e sa v in g s 0 .2 9 8 * * * (0 .0 3 5 ) 3 ,1 1 4 .6 (1 ,9 6 3 .9 5 6 ) �8 5 3 .9 * * * (1 4 9 .4 5 9 ) 0 .1 0 9 (0 .0 8 9 ) r ac e b la ck 0 .1 2 6 * (0 .0 5 2 ) 8 ,2 9 1 .1 * * * (1 ,8 6 3 .7 4 6 ) 2 1 6 .5 (3 0 0 .6 1 0 ) 0 .1 1 6 (0 .0 6 7 ) r ac e h is p an ic �0 .0 3 0 0 (0 .0 5 2 ) �5 2 5 .3 (4 ,6 7 0 .2 0 3 ) �1 7 7 .6 (3 0 0 .4 3 0 ) 0 .1 1 6 (0 .0 8 8 ) r ac e o th er �0 .0 1 7 8 (0 .0 6 2 ) 4 ,9 8 9 .8 (2 ,6 8 3 .8 8 1 ) �5 1 1 .6 * * (1 8 4 .2 1 1 ) 0 .4 4 3 (0 .2 2 6 ) in co m e (p er $ 1 0 k ) 0 .0 5 7 8 * * * (0 .0 1 1 ) 3 ,4 1 0 .2 * * (1 ,1 2 8 .4 5 8 ) �3 5 .2 4 (1 8 .2 7 2 ) �0 .0 9 5 0 * * (0 .0 2 9 ) a g e �0 .0 1 6 1 * * * (0 .0 0 1 ) �1 ,2 3 7 .7 * * * (5 2 .3 0 0 ) �1 2 .9 4 * * * (3 .9 0 4 ) �0 .0 2 3 8 * * * (0 .0 0 2 ) f in an ci al p la n n in g h o ri zo n (o m it te d b as el in e ca te g o ry “n ex t fe w m o n th s” ) n ex t y ea r �0 .1 5 0 * * (0 .0 5 2 ) �3 ,8 3 9 .2 (2 ,0 4 3 .2 4 8 ) �9 3 0 .4 * * (2 8 6 .5 3 1 ) 0 .1 3 6 (0 .1 6 0 ) n ex t fe w y ea rs �0 .0 1 9 5 (0 .0 4 3 ) �5 4 3 9 .2 * (2 ,2 6 9 .3 3 3 ) �7 3 4 .6 * (2 9 3 .1 3 4 ) �0 .1 2 3 * (0 .0 5 7 ) n ex t 5 to 1 0 y ea rs �0 .2 0 2 * * * (0 .0 5 0 ) �1 1 ,8 6 1 .2 * * * (2 ,5 3 0 .7 1 3 ) �9 6 9 .5 * * * (2 6 2 .8 3 3 ) �0 .1 9 3 * * (0 .0 7 0 ) l o n g er th an 1 0 y ea rs �0 .2 4 5 * * * (0 .0 6 0 ) �6 ,5 5 3 .9 * (3 ,1 5 6 .3 2 2 ) �6 3 0 .7 * (3 0 6 .9 9 2 ) �0 .2 7 6 * * (0 .0 9 0 ) n 4 ,4 8 1 4 ,4 8 1 4 ,4 8 1 4 ,4 8 1 n o te s: a t h is co lu m n re p o rt s th e av er ag e m ar g in al ef fe ct o f th e p ro b it re g re ss io n . t h e 2 0 1 6 s u rv ey o f c o n su m er f in an ce s (s c f ) sa m p le w ei g h ts w er e ap p li ed to th e re g re ss io n s. b t h e an al y si s sa m p le in th is re g re ss io n in cl u d es th e h o u se h o ld s w h o se d eb t to fi n an ci al as se t ra ti o eq u al s to ze ro . f o r th e re d u ce d -s iz e an al y si s sa m p le w h ic h o n ly in cl u d es th e h o u se h o ld s w h o ca rr y d eb t, p le as e se e th e re d u ce d sa m p le re g re ss io n re su lt s in a p p en d ix c . c t h e an al y si s sa m p le in th is re g re ss io n in cl u d es th e h o u se h o ld s w h o se d eb t to in co m e ra ti o eq u al s to ze ro . a lt er n at iv e re g re ss io n w it h th e re d u ce d sa m p le w h er e o n ly h o u se h o ld s w it h d eb ts ar e in cl u d ed is d is cu ss ed in a p p en d ix c . s ta n d ar d er ro rs in p ar en th es es . * p < .0 5 , * * p < .0 1 , * * * p < .0 0 1 . 132 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 benefits. the costs associated with good debts are often outweighed by the benefits. bad debts, on the contrary, carry high interest rates with little or no long-term returns (hanson, 2006). these types of debts can potentially negatively impact the borrower’s credit scores, retirement goals, and financial health, as well as family relationships. in some circumstances, bad debts can create a vicious borrowing cycle for some families and cause stress and mental as well as physical health problems (davies, montgomerie, & wallin 2015). the negative health effects of debt (i.e., the “high price of debt”) is a phenomenon noted both in the united states (sweet et al., 2013) and internationally (clayton, liñares-zegarra, & wilson, 2015). this article explores the different factors that are potentially related to households carrying bad debts. we first investigate these potential factors by separating debt categories. (in table 2, we chose credit card revolving balances as a representation of bad debt and mortgages as an example of good debt.) then, we utilize different interest rate measures to check the relationships between these factors and high interest rates (table 3). although liquid assets and interest rates are both predicted to be associated with household leveraging, we expect these factors to play different roles when it comes to “good debts” versus “bad debt.” in particular, we want to test whether interest rate has more significant negative relationships with mortgages due to the large size and long durations of the debt, and whether liquid assets level is more significantly and negatively related to credit card debts due to the “liquidity needs compromise.” in addition, we expect to observe negative significant relationships between household financial planning horizons and the amount of both types of debts. the regression results in table 2 indicate that although some household attributes are related to both good and bad debts, certain factors are particularly noteworthy when it comes to explaining what kinds of households are more likely to carry bad debts. having more children is positively associated with both credit card loans and mortgages. however, other factors such as interest rate, real assets, liquid assets, and income have different relationships with credit card debt compared with mortgages. for instance, mortgages are more sensitive to interest rate changes, but credit card loans are more sensitive to liquid assets and income. the reason behind this difference could be interpreted as “liquidity needs” compromise. credit card loans are often used to cover short-term liquidity needs. their insensitivity toward interest rates could be largely caused by a lack of liquid assets to cover certain short-term needs (such as holiday shopping, etc.). therefore, credit card debts are negatively related to liquid asset levels. on the contrary, mortgages are negatively associated with interest rates because of their relatively larger debt size (hence larger interest payments) and longer investment horizon. one interpretation of the income effect on credit card loans could be that, keeping everything else (including liquid assets) equal, households with higher incomes have the ability and resources to borrow—and pay back—more credit card loans. age is another factor that is only negatively related to mortgages. this finding indicates that older households are more likely to have had a longer time to pay off their mortgages and hence reduce the size of this type of good debt. because houses are a major component of most households’ real assets, it is not surprising that the real asset level is positively related to family mortgage loans. the financial planning horizon factor is negatively associated with both credit card loans and mortgages in z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 133 table 2. this is consistent with the previous regression results, indicating that families with longer financial planning horizons are less likely to carry both kinds of debts. a major focus of household liability management is to help the targeted families to reduce the interest rates of their debts. the following analysis seeks to explore what kind of factors are associated with higher household interest rates. we expect to see negative relationships between the weighted average interest rate and certain household characteristics such as real and liquid asset levels, household head education level, homeownership, savings, as well as being married and having a longer financial planning horizon. in table 3, we use different measures to capture the households’ average interest rates as well as the percentile ranking of the average interest rates. the weighted average interest rate takes into account the dollar amount weighted average interest rates across all loan types. for example, for each household, the dollar amount of different loans is multiplied by their interest rates to calculate the overall liability cost per year. then this liability cost is divided by the total loan amount to acquire the weighted average interest rate for this household. the simple average interest rate measure takes the arithmetic average of the interest rates across all loan types. this measurement, together with the weighted average interest rate percentile and simple average interest rate percentile measures, serves as a robustness check for the weighted average interest measurement. based on the ols regression results from table 3, households with less education, lower levels of assets, fewer savings, and older age are table 2 ordinary least squares (ols) on different debt categories variables “bad” debts (credit and store cards balance) “good” debts (mortgages) interest rate �28.40 (18.556) �1,673.5* (677.040) married 265.1 (364.032) �4,993.6 (4,070.197) number of kids 275.3* (122.659) 5,148.3*** (1,183.437) education level 76.84 (54.142) 1,626.4* (771.444) real assets 0.00197 (0.001) 0.393*** (0.020) liquid assets �0.0328*** (0.003) �0.0817 (0.067) have houses 519.9 (388.969) omitted have savings �419.8 (262.935) �823.7 (3,076.358) race black �528.6 (345.616) 7695.6 (4,149.247) race hispanic �685.3* (314.295) 6,027.6 (9,231.032) race other �247.4 (311.224) 4,493.5 (5,760.525) income 0.0201** (0.007) 0.144 (0.090) age 13.30 (6.916) �1,009.6*** (106.106) financial planning horizon (omitted baseline category “next few months”) next year �685.5 (386.068) 1,138.5 (4,312.770) next few years �861.4** (322.013) �6,064.8 (4,592.322) next 5 to 10 years �1,109.0*** (329.694) �9,363.6** (3,520.808) longer than 10 years �1,368.6** (483.781) �12,236.5** (4,396.354) n 2,808 1,661 notes: not all of the respondents in our analysis sample reported the interest of different kinds of loans. therefore, the number of observations was reduced in the regressions above. the 2016 survey of consumer finances (scf) sample weights were applied to the regressions. standard errors in parentheses. *p< .05; **p< .01; ***p< .001. 134 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 t ab le 3 o rd in ar y le as t sq u ar es (o l s ) o n av er ag e in te re st ra te m ea su re s v ar ia b le s w ei g h te d a av er ag e in te re st ra te w ei g h te d av er ag e in te re st ra te p er ce n ti le s im p le b av er ag e in te re st ra te s im p le av er ag e in te re st ra te p er ce n ti le m ar ri ed �1 .1 8 7 * * (0 .3 7 8 ) 0 .0 0 5 2 4 (0 .0 0 9 ) �0 .7 8 9 * (0 .3 5 0 ) 0 .0 0 9 0 3 (0 .0 0 9 ) n u m b er o f k id s 0 .0 3 3 7 (0 .1 0 2 ) 0 .0 1 0 8 * * * (0 .0 0 3 ) 0 .1 6 4 (0 .1 0 2 ) 0 .0 1 1 0 * * * (0 .0 0 3 ) e d u ca ti o n le v el �0 .3 1 9 * * * (0 .0 5 4 ) �0 .0 0 6 1 3 * * * (0 .0 0 2 ) �0 .2 7 7 * * * (0 .0 5 1 ) �0 .0 0 6 8 0 * * * (0 .0 0 1 ) r ea l as se ts (p er $ 1 0 k ) �0 .0 3 8 1 * * * (0 .0 0 7 ) �0 .0 0 1 2 8 * * * (0 .0 0 0 ) �0 .0 4 3 1 * * * (0 .0 0 8 ) �0 .0 0 1 4 2 * * * (0 .0 0 0 ) l iq u id as se ts (p er $ 1 0 k ) �0 .1 4 1 * * * (0 .0 0 0 ) �0 .0 0 3 4 8 * * (0 .0 0 0 ) �0 .1 6 6 * * * (0 .0 0 0 ) �0 .0 0 3 8 8 * * * (0 .0 0 0 ) h av e h o u se s �0 .8 8 8 * * (0 .3 4 2 ) �0 .0 0 4 8 4 (0 .0 1 0 ) �0 .3 6 8 (0 .4 0 2 ) �0 .0 0 7 7 4 (0 .0 0 9 ) h av e sa v in g s �0 .8 2 6 * (0 .3 6 4 ) �0 .0 1 6 4 * * (0 .0 0 6 ) �0 .8 5 1 * (0 .3 5 4 ) �0 .0 1 7 4 * * (0 .0 0 5 ) r ac e b la ck �0 .5 0 6 (0 .3 8 4 ) 0 .0 3 0 7 * * (0 .0 0 9 ) 0 .0 1 2 7 (0 .3 3 6 ) 0 .0 2 8 4 * * * (0 .0 0 8 ) r ac e h is p an ic 0 .7 4 9 (0 .4 1 3 ) 0 .0 4 5 2 * * * (0 .0 1 1 ) 1 .3 3 2 * * (0 .5 0 9 ) 0 .0 3 5 8 * * * (0 .0 1 0 ) r ac e o th er 0 .0 2 6 0 (0 .4 2 0 ) 0 .0 1 2 6 (0 .0 1 1 ) 0 .6 2 7 (0 .4 7 3 ) 0 .0 1 6 6 (0 .0 1 1 ) in co m e (p er $ 1 0 k ) �0 .0 2 5 5 (0 .0 4 0 ) �0 .0 0 1 0 9 (0 .0 0 1 ) �0 .0 3 0 2 (0 .0 4 1 ) �0 .0 0 1 6 6 (0 .0 0 1 ) a g e 0 .0 4 6 2 * * * (0 .0 1 0 ) 0 .0 0 0 5 8 9 * (0 .0 0 0 ) 0 .0 2 9 7 * * * (0 .0 0 9 ) 0 .0 0 0 5 5 9 * (0 .0 0 0 ) f in an ci al p la n n in g h o ri zo n (o m it te d b as el in e ca te g o ry “n ex t fe w m o n th s” ) n ex t y ea r �1 .3 4 3 * * * (0 .4 0 6 ) �0 .0 0 4 0 7 (0 .0 1 1 ) �0 .3 9 6 (0 .4 1 6 ) �0 .0 1 0 9 (0 .0 1 0 ) n ex t fe w y ea rs �0 .5 0 5 (0 .5 0 7 ) �0 .0 1 8 8 * (0 .0 0 9 ) �0 .3 3 5 (0 .4 4 7 ) �0 .0 1 2 8 (0 .0 0 8 ) n ex t 5 to 1 0 y ea rs �1 .3 5 9 * * * (0 .3 4 6 ) �0 .0 1 8 5 (0 .0 1 1 ) �1 .1 8 8 * * * (0 .2 5 3 ) �0 .0 1 9 4 * (0 .0 0 9 ) l o n g er th an 1 0 y ea rs �0 .6 9 2 (0 .4 3 6 ) �0 .0 1 1 1 (0 .0 1 2 ) �0 .4 7 7 (0 .3 5 6 ) �0 .0 0 9 2 4 (0 .0 1 1 ) n 3 ,3 9 8 3 ,3 9 8 3 ,3 9 8 3 ,3 9 8 n o te s: in te re st in fo rm at io n fo r so m e lo an s w as n o t re p o rt ed in th e 2 0 1 6 s u rv ey o f c o n su m er f in an ce s (s c f ) d at a, th er ef o re th e to ta l n u m b er o f h o u se h o ld s w as re d u ce d fr o m 4 ,4 8 1 to 3 ,3 9 8 . t h e 2 0 1 6 s c f sa m p le w ei g h ts w er e ap p li ed . a “w ei g h te d ” m ea n s th is in te re st ra te m ea su re ta k es th e d o ll ar am o u n t w ei g h te d av er ag e o f th e in te re st ra te s ac ro ss al l lo an ty p es in to ac co u n t. t h at is , fo r ea ch h o u se h o ld , th e d o ll ar am o u n ts o f d if fe re n t lo an s ar e m u lt ip li ed b y th ei r in te re st ra te s to ca lc u la te th e o v er al l li ab il it y co st p er y ea r. t h en th is li ab il it y co st is d iv id ed b y th e to ta l lo an am o u n t to ac q u ir e th e w ei g h te d av er ag e in te re st ra te fo r ea ch h o u se h o ld . b ™ s im p le ” m ea n s th is in te re st m ea su re is b as ed o n th e si m p le ar it h m et ic av er ag e o f th e in te re st ra te s ac ro ss al l lo an ty p es . t h is m ea su re m en t se rv es as a ro b u st ch ec k fo r th e w ei g h te d av er ag e in te re st m ea su re . s ta n d ar d er ro rs in p ar en th es es . * p < .0 5 ; * * p < .0 1 ; * * * p < .0 0 1 . z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 135 subject to higher interest rates. therefore, families with these attributes are more likely to need help with liability management and could potentially benefit significantly from interest rate reductions. finally, we perform the alpha-equivalent analysis to determine the potential savings a household would experience if it were able to reduce the interest rates on their existing liabilities. for the analysis we assume the household’s interest rates are reduced based on the distribution of household loan interest rates as noted in fig. 3 we assume each liability would be reduced by some percentile amount, based on the distribution for that respective liability. for example, let us assume a household had financial assets (i.e., a portfolio) worth $100,000 and a single liability, which was $15,000 in credit card debt at an interest rate of 15%. a 15% interest rate on credit card debt would be in the 47th percentile of interest rates according to fig. 4 if the household were able to reduce the interest rate by ten percentile points, to the 37th percentile, the interest rate would decline to approximately 13%. this results in an interest savings of 2% (15% to 13% = 2%) that would translate into $300 of total savings on the $15,000 total credit card debt ($15,000*2% = $300). if we divide the estimated $300 in annual interest savings by the total financial assets, we can estimate the “alpha-equivalent” benefit associated with liability optimization, which would be 30 bps (basis points) in this case ($300/$100,000 = 30 bps). we conduct this analysis for all households, where the rate on each loan is assumed to be reduced by some percentile level, based on the distribution of loan rates in fig. 3 for the analysis the lowest possible rate is the 1st percentile. information about the distribution of potential dollar savings and alpha-equivalent benefit are included in fig. 6 in panels a and b, respectively. the potential savings associated with improving loan rates can be significant, especially for households that have higher interest percentiles. if a household’s weighted average debt interest rate is currently in the 95th percentile, a five-percentile drop could generate 113.5% equivalent alpha, or $1,641 in annual savings. if these households achieve a 10-percentile reduction in loan rates, the total savings would be $2,614, which is equivalent to 237.5% of investment alpha. even the median household stands to benefit from even modest improvements. for example, the median households would on average save $410 if they were able to reduce their weighted average loan rates by 10 percentile points, which is equivalent to a 195 bps of investment alpha. this suggests that, for many households, making efforts to reduce the interest rates on their liabilities is more likely to result in wealth gains than attempting to construct portfolios that might outperform the market. notice that when calculating the potential savings on interest rate reductions, we use the weighted average interest rate in the discussion. lowering the household average interest rate may be achieved in two different ways. first, households can make more efforts on interest rate shopping and negotiate lower interest rate on their loans, if possible. second, even if directly lowering interest rates is not feasible, the weighted average interest rate can still be reduced through debt restructuring. households can substitute a higher interest loan with lower interest borrowings to achieve the reductions of overall weighted average interest rates. (for example, consider a household with a large revolving balance on credit card loans who cannot reduce the total amount of household debt. this household could still potentially 136 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 pay off this high interest rate credit card loan with low-interest secured-personal loans or some other type of loan. by doing so, the average interest rate of this household could be reduced, potentially significantly.) the analyses above reveal the significant potential benefits of liability management and point out the characteristics that are associated with different households’ debt problems. financial planning practitioners and financial institutions can benefit from this research not only by recognizing the potential benefits of liability management for lowto-affluent american families, but also by identifying the attributes associated with those fig. 6. benefit of reducing interest rates on debt. source: federal reserve board survey of consumer finances (scf) 2016 survey wave. notes: the subsample is restricted to households that carry loans, reported complete data on all loan types, and have more than $1 in financial assets. the number of observations is 3,371. the 2016 scf sample weights were applied. z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 137 households that most need debt assistance. this study can also encourage consumers to seek for an integrated approach to making decisions about their marginal income and benefit significantly from analyzing both sides of their balance sheet extensively and regularly.11 6. conclusion and implications debt is a significant and growing component of u.s. household balance sheets. with total interest rate payments on loans exceeding the expected returns on household financial assets for the average household, the impact of liability optimization should draw more focus from financial advisors, financial firms, and consumers. in this study, we first reviewed american families’ current financial outlook by looking at their debt situations. using the scf data, we then analyzed the different economic, demographic, and behavioral factors that are associated with household borrowing and leverage ratios. next, we separated the good and bad debts and investigated whether the attributes related to different debt categories are similar. after checking the characteristics demonstrated by the households that carry high-interest debts, we performed alpha-equivalent analyses to calculate the potential benefits of liability management. our study indicates that households with lower assets, income, and education levels need assistance the most and could significantly benefit from debt management. households’ time discounting preferences also play an important role in their borrowing decisions. families with longer financial planning horizons are less likely to carry loans. among the borrowers, a shorter financial planning horizon is usually an indicator of a higher debt amount as well as higher debt-to-asset and debt-to-income ratios. families with myopic planning horizons are also more likely to carry a higher amount of bad debts, such as credit card balances. this study can also inspire advisors and financial services firms to consider alternative approaches to helping consumers improve their financial well-being. for example, advisors could help their clients design a road map for debt restructuring and interest rate reduction along with building portfolio investment strategies. by reviewing both sides of the household balance sheet extensively and periodically, advisors can integrate both investment and liability management strategies to better improve their clients’ economic outlooks. these strategies would be particularly effective for households with lower income, education, and asset levels. large retirement firms could explore the possibility of building a bridge between their retirement plan participants and lending institutions to help their participants gain access to loans with competitive rates. participants could utilize these lower “group rate” loans to restructure and reduce the interest payments on their existing debts. financial planners could also implement different behavior coaching strategies (such as behavioral nudging devices) to help their clients increase their financial planning horizons and avoid the consequences of myopic planning. 138 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 the integration of investment and liability management strategies prompts financial advisors to help their clients to answer the question, “where should my next dollar go?” by designing a universal comparison mechanism between investing and paying off debt, financial advisors can help their customers to better manage their marginal income. an integrated model or strategy can be designed to not only educate the consumers on the importance of liability management, but also guide their decision-making process after taking each consumer’s unique financial situation into account. future studies may find it favorable to build such an integrated methodology to help answer the age-old invest or pay off debt conundrum faced by many households. notes 1 defined as households with a net worth not exceeding $1 million, have more than $1,000 annual income and have at least $1 in financial assets. high net worth households, defined as those with net worth over $1 million, often have their own unique leveraging and investment strategies, and optimizing these special strategies is beyond the scope of this paper. our definition of “low-to-affluent” households includes those in the middle-to-low income range because these households are most likely to need debt management assistance. detailed descriptions of the analysis sample can be found in the data and methodology section of this article. 2 this is the lower end of average credit card and retail store installment card interest rates. source: 2016 scf data weighted average credit card interest rate for low-toaffluent households. 3 the definition of “good” and “bad” debts is discussed in both the literature review section and the results section. 4 the consumer finance monthly study is conducted by the consumer finance research group at ohio state university. 5 the utility function satisfies monotonicity (more is preferred to less) and concavity (diminishing marginal utility) properties and assumes ct’s are normal goods for every period t. the concavity property implies the preference of smoothing consumption across time because of the love of diversity. 6 the budget constraint depicted by inequality (14) is derived from the following constraints while substituting out the borrowing factor st (8t from1 to t): ct þ st ≤ yt (11.1) ctþ1 ≤ ð1þ rþ st þ ytþ1 (11.2) 7 given the prevalence of “payday lending” and other short-term loans in the united states (caskey, 2001; stegman, 2007), we assume that american households have access to sufficient amount of lending sources despite the fact that some of the loans may have unreasonably high interest rates. while we do not recommend consumers access these short-term loans, we use their potential access abilities of these loans to z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 139 simplify the model and transform the borrowing constraints to interest rate constraints. another reason why we do not restrict the borrowing/saving factor s in the intertemporal consumption model is that this factor is canceled out when combining the two-period budget constraints together using substitutional method and lagrangian technique to solve this intertemporal optimization problem. 8 the most current wave available at the time of the analysis. 9 the survey of consumer finances uses “multiple imputation technique” to account for missing data. because each missing value in the scf is imputed five times, each scf family has five separate observations (called “implicates”) in the final data. 10 the scf data are derived from a dual-frame sample design, with one frame including households chosen via an area probability sample and the second frame including households selected from a list provided by the internal revenue service. the second selection frame has introduced the problem of oversampling wealthy families (nielsen, 2015). 11 liability optimization includes debt restructuring, loan reduction, interest rate optimization, behavior coaching, etc. there are numerous complexities associated with liability optimization at the individual household level. the objective of this article is not focused on the detailed liability optimization approaches, rather to better understand which types of households have higher debts, in particular bad debts, and the potential benefits associated with reducing the interest on those debts. 140 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 appendix a sample balance sheet for the weighted mean value of the 2016 survey of consumer finances households assets category subcategory sub-category detail amount sum total percent of population interest rate (estimateda) total annual earnings financial assets $73,122.13 98.29% transaction accounts (liquid) $13,590.40 97.75% 0.20% $27.18 cds $1,485.40 4.84% 0.96% $14.26 pooled investment funds $4,569.21 5.66% 5.61% $256.52 savings bonds $351.91 7.33% 2.62% $9.22 directly held stocks $2,791.27 9.20% 7.53% $210.13 directly held bonds $273.57 0.38% 3.70% $10.12 cash value of whole life insurance $2,746.16 17.23% 2.20% $60.42 other managed assets: $3,725.21 3.35% 4.10% $152.73 annuities $2,861.44 trusts $863.77 quasi-liquid retirement accounts $42,112.74 48.04% 4.00% $1,684.51 other misc. financial assets $1,476.26 8.29% 4.00% $59.05 nonfinancial assets $159,244.50 89.76% vehicles (rvs, planes, boats, etc.) $17,982.83 84.47% primary residence $118,573.20 59.54% residential property excluding primary residence $11,071.96 9.12% net equity in nonresidential real estate $2,780.49 4.22% businesses $7,812.36 9.23% other misc. nonfinancial assets $1,023.67 4.92% total assets $232,366.60 99.32% net worth $156,181.85 total investment assets $73,122.12 $2,484.13 total financial assets less total debt $(3,062.63) mortgages (including home equity loans, helocs) debt secured by primary residence: $53,249.90 40.67% mortgages and home equity loans secured by primary residence $51,818.03 39.27% 4.51% $2,336.99 home equity lines of credit secured by primary residence $1,431.87 3.59% 5.81% $83.19 debt secured by other residential property $3,875.32 3.87% 5.45% $211.20 other lines of credit (not secured by residential real estate) $143.90 1.80% 6.00% $8.63 credit card balances after last payment $2,581.29 47.79% 15.09% $389.52 installment loans $15,794.03 53.50% 16.27% education loans $8,390.03 24.69% 5.92% $496.69 vehicle loans $5,764.13 35.36% 6.63% $382.39 other installment loans $1,639.87 12.51% 6.00% $98.39 (continued on next page) z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 141 appendix a (continued) assets category subcategory sub-category detail amount sum total percent of population interest rate (estimateda) total annual earnings other debt (e.g., loans against pensions or life insurance, margin loans) $540.31 5.35% 6.00% $32.42 total debt $76,184.75 79.24% $4,039.43 total asset return less total interest charges $(1,547.19) financial asset to debt ratio 0.960 notes: sample weights applied. number of households: 4,481; net worth < $1million; income > $1,000; age: 20–85. a interest rate estimation sources: cds: fred, federal reserve bank of st. louis. averaged since 2008. pooled investment fund: assumes 50% stocks and 50% bonds. uses the average for mutual fund return. savings bonds: us department of the treasury, 10-year high quality market (hqm) corporate bond spot rate [hqmcb10yr]. directly held stocks: s&p 500 return calculator, with dividend reinvestment. (2019). retrieved march 22, 2019. directly held bonds: us department of the treasury, 10-year high quality market (hqm) corporate bond spot rate [hqmcb10yr]. retrieved march 22, 2019, from fred, federal reserve bank of st. louis. 142 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 appendix b descriptive statistics of the analysis sample variable definition/explanations mean sd min. max. have debt (yes = 1, no = 0) 0.7924 0.4056 0 1 debt amount dollar amount of total debt $76,185 $117,866 $0 $2,630,000 married (yes = 1, no = 0) 0.5433 0.4981 0 1 number of children in household total number of children in the household 0.7989 1.1334 0 7 education level highest level of education completed according to the scf standard categoriesa 9.2774 2.7128 0 14 real assets total value of real assetsb $150,409 $185,257 $0 $2,282,900 liquid assets all types of transaction accountsc $13,590 $32,926 $0 $572,000 leverage ratio total debt/total asset 12.8096 462.9600 0 25,750 own houses (yes = 1, no = 0) 0.5954 0.4908 0 1 have savings have more than $0 in savings? (yes = 1, no = 0) 0.5043 0.5000 0 1 race black (yes = 1, no = 0) 0.1637 0.3700 0 1 race hispanic (yes = 1, no = 0) 0.1140 0.3178 0 1 race other (yes = 1, no = 0) 0.1073 0.3095 0 1 income household income in previous calendar year $62,321 $59,020 $1,013 $2,531,591 age age of the household head 49.6815 16.5849 20 85 financial planning horizon categorical variablesd next year (yes = 1, no = 0) 0.1551 0.3620 0 1 next few years (yes = 1, no = 0) 0.2817 0.4498 0 1 next 5 to 10 years (yes = 1, no = 0) 0.2186 0.4133 0 1 longer than 10 years (yes = 1, no = 0) 0.1059 0.3077 0 1 notes: sample size is 4,481 households. sample weights applied. a 2016 survey of consumer finances (scf) codebook education level standard categories: 1. 1st, 2nd, 3rd, or 4th grade. 2. 5th or 6th grade. 3. 7th and 8th grade. 4. 9th grade. 5. 10th grade. 6. 11th grade. 7. 12th grade, no diploma. 8. high school graduate high school diploma or equivalent. 9. some college but no degree. 10. associate degree in college occupation/vocation program. 11. associate degree in college academic program. 12. bachelor’s degree (e.g., ba, ab, bs). 13. master’s degree (e.g., ma, ms, meng, med, msw, mba). 14. professional school degree (e.g., md, dds, dvm, llb, jd) and doctorate degree (e.g., phd, edd). b real assets, according to the scf bulletin category definition, include: houses, vehicles, residential properties excluding primary residence (e.g., vacation homes), and net equity in non-residential real estate. c liquid assets, according to the scf bulletin category definition, include: money market accounts, checking accounts, savings accounts, call accounts, and prepaid cards. d original survey question from scf codebook: “in planning or budgeting your (family’s) saving and spending, which of the time periods listed on this page is most important to you (and your family living here)?” z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 143 references aizcorbe, a. m., kennickell, a. b., & moore, k. b. (2003). recent changes in us family finances: evidence from the 1998 and 2001 survey of consumer finances. federal reserve bulletin, 89, 1. barba, a., & pivetti, m. (2009). rising household debt: its causes and macroeconomic implications—a long-period analysis. cambridge journal of economics, 33, 113-137. bricker, j., dettling, l. j., henriques, a., hsu, j. w., jacobs, l., moore, k. b., . . . windle, r. a. (2017). changes in us family finances from 2013 to 2016: evidence from the survey of consumer finances. federal reserve bulletin, 103, 1. caskey, j. (2001). payday lending. journal of financial counseling and planning, 12. (available at https://ssrn. com/abstract=2443130) clayton, m., liñares-zegarra, j., & wilson, j. o. (2015). does debt affect health? cross country evidence on the debt-health nexus. social science & medicine, 130, 51-58. davies, w., montgomerie, j., & wallin, s. (2015). financial melancholia: mental health and indebtedness. london: political economy research centre. emmons, w. r., & noeth, b. j. (2013). economic vulnerability and financial fragility. federal reserve bank of st. louis review, 95, 361-388. appendix c ordinary least squares (ols) regressions on factors associated with debt ratios (reduced sample) variables ols ols ols debt-to-financial-asset ratioa debt-to-income ratiob debt-to-financial-asset ratioc married �215.9 (244.237) �0.357* (0.144) 21.86 (24.820) number of kids 168.7 (166.241) �0.0148 (0.029) 6.673 (6.939) education level �117.3** (44.613) 0.105*** (0.027) 1.151 (4.360) real assets (per $10k) 8.508 (9.951) 0.0435*** (0.004) 0.342 (0.526) liquid assets (per $10k) 13.88 (9.462) �0.0343** (0.013) �3.307** (1.085) have houses �109.2 (323.537) 0.770*** (0.084) 60.14*** (16.643) have savings �1142.8*** (198.107) 0.0345 (0.109) �69.84*** (19.396) race black 314.1 (373.679) 0.113 (0.077) 65.24 (54.015) race hispanic �168.1 (385.584) 0.108 (0.099) 19.05 (25.444) race other �662.5** (237.327) 0.576* (0.275) �1.861 (10.477) income (per $10k) �43.01 (22.350) �0.120** (0.040) �4.335 (2.696) age �13.59** (4.997) �0.0239*** (0.002) �0.517 (0.292) financial planning horizon (omitted baseline category “next few months”) next year �1184.4** (379.501) 0.234 (0.198) �48.46** (16.202) next few years �917.7* (371.363) �0.150* (0.071) �4.372 (29.915) next 5 to 10 years �1196.6*** (339.037) �0.150 (0.086) �52.28** (16.884) longer than 10 years �752.0 (389.874) �0.241* (0.120) �54.19** (16.834) n 3,561 3,561 3,495 notes: a the analysis sample in this regression has been reduced and does not include the households whose total debt amount equals to zero. b the analysis sample in this regression only include the households whose total debt amount is greater than zero. c the analysis sample in this regression only include the households whose total debt amount is greater than zero. the analysis sample is further reduced by eliminating the households whose total financial assets is less than $10. 144 z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 federal reserve bank of new york center for microeconomic data. (2020). quarterly report on household debt and credit, 2020:q2. (available at https://www.newyorkfed.org/microeconomics/hhdc.html). gustman, a. l., steinmeier, t. l., & tabatabai, n. (2011). how did the recession of 2007-2009 affect the wealth and retirement of the near retirement age population in the health and retirement study? (no. w17547). national bureau of economic research. hanson, j. (2006). good debt, bad debt: knowing the difference can save your financial life. new york, ny: portfolio penguin. jappelli, t., & pagano, m. (1989). consumption and capital market imperfections: an international comparison. the american economic review, 79, 1088-1105. jiang, s. s., & dunn, l. f. (2013). new evidence on credit card borrowing and repayment patterns. economic inquiry, 51, 394-407. kim, h., & devaney, s. a. (2001). the determinants of outstanding balances among credit card revolvers. financial counseling and planning, 12, 67-77. lee, y. g., lown, j. m., & sharpe, d. l. (2007). predictors of holding consumer and mortgage debt among older americans. journal of family and economic issues, 28, 305-320. modigliani, f. (1986). life cycle, individual thrift, and the wealth of nations. science, 234, 704-712. nielsen, r. b. (2015). scf complex sample specification for stata. technical note. department of financial planning housing and consumer economics. university of georgia, athens, ga. stango, v., & zinman, j. (2016). borrowing high versus borrowing higher: price dispersion and shopping behavior in the us credit card market. review of financial studies, 29, 979-1006. stegman, m. a. (2007). payday lending. journal of economic perspectives, 21, 169-190. sweet, e., nandi, a., adam, e. k., & mcdade, t. w. (2013). the high price of debt: household financial debt and its impact on mental and physical health. social science & medicine, 91, 94-100. wärneryd, k. e. (1999). the psychology of saving. a study on economic psychology. cheltenham: edward elgar publishing. zinman, j. (2015). household debt: facts, puzzles, theories, and policies. economics, 7, 251-276. z. liu and d. m. blanchett / financial services review 29 (2021) 121–145 145 the importance of debt for household risky asset allocation and portfolio structure ran taoa,*, yuan yuanb adepartment of economics, college of business and economics, university of wisconsin, whitewater, 800 w. main street, whitewater, wi 53190, usa bdepartment of finance and business law, college of business and economics, university of wisconsin, whitewater, 800 w. main street, whitewater, wi 53190, usa abstract when households decide on risky asset holdings, they do not make the decision in isolation from their debt structure and obligations, vice versa. we examine the joint behavior of debt and financial asset portfolio decisions, while existing empirical research on debt and asset portfolio choices has proceeded separately. in this paper, we first test the relationship between debt structure and asset allocation, then estimate the determinants of debt structure and asset allocation simultaneously. using the 2016 survey of consumer finances (scf) data, we find robust evidence that debt structure affects households’ risky asset allocation decisions and identify, in this simultaneous decision-making process, the demographic and financial factors that can contribute to the household overall financial portfolio structure. © 2018 academy of financial services. all rights reserved. jel classification: g11; d14 keywords: financial asset allocation; risky asset investment; secured debt; debt structure 1. introduction there has been a massive increase in consumer assets and consumer debt over the last two decades. at the end of 2017, u.s. households’ total financial assets exceeded $80 trillion and total household debts rose to an all-time high of $15.5 trillion, more than twice of what they were in 2000 (board of governors of the federal reserve system). some striking effects * corresponding author. tel.: �1-262-472-5447; fax: �1-262-472-4863. e-mail address: taor@uww.edu (r. tao) financial services review 27 (2018) 325-344 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. observed include the broadening of the stockholder base (bilias et al., 2010), the growth in mutual fund participation (bailey et al., 2011), and consumer indebtedness accompanied by the fast growth in credit card use (basnet and donou-adonsou, 2018). empirical evidence shows that households do not follow the predictions by portfolio theories (cf. campbell, 2006; guiso et al., 2002). in addition, large variations across households in their portfolio structures are observed. for example, many households do not hold risky financial assets, while those do, many hold a large proportion of risky assets (campbell, 2006). the portfolio allocation problems are gaining attention again in the academic circle. the renewed interest generally focuses on the key aspects of portfolio structure and understanding qualitatively as well as quantitatively the role of the determinants in the optimal investment decision of individuals and households (cf. cardak and wilkins, 2009). however, the approach taken by the existing literature typically focuses on specific aspects of household finance in isolation of other aspects of the balance sheet. many studies focus on either the asset side (cf. wang and hanna, 2018) or the debt side of the household portfolio (cf. yilmazer and devaney, 2005). in this paper, we examine the overall financial portfolio structure by considering both asset and debt allocation decisions. a household’s asset structure is defined as the share of risky assets (stocks, corporate/foreign bond, mutual funds, and trust funds) in total financial assets. on the liability side, we classify the total debt as secured debt (mortgage and vehicle loans) and unsecured debt (credit card debt, education loan, and other consumer loans that are not backed by any underlying assets). a household’s debt structure is represented by the ratio of outstanding secured debt balance to total debt. there are many studies of household asset portfolio. notable contributions include bergstresser and poterba (2004), cardak and wilkins (2009), rosen and wu (2004). they find that asset allocation decisions are affected by households’ demographics, educational attainment, wealth, labor income, and health risks. therefore, the likelihood of participating in risky asset investment should be related to issues affecting access and awareness of stocks if one had some level of financial assets. drawing from both the frameworks of investment decision making and previous studies, we apply a two-stage sample selection model to examine risky asset shares conditional on participation (i.e., the decision to hold risky assets). the two-stage model allows for the individual determination of differences in the participation decision and the allocation decision, because unconditional shares cannot distinguish the effects of relevant variables on the participation decision from those on the portfolio share given that the asset is held. stage one examines the likelihood of risky asset ownership and can be indicative of access barriers. stage two examines risky asset allocation and can be reflective of attitude and preferences. our methodology improves on many of the theoretical predictions of the classical portfolio theory that refer to asset shares, not to participation decisions (cf. campbell et al., 2003). our results show that the participation and allocation decisions are determined by distinct factors. we also incorporate household debt structure in the analysis. while households’ decision to invest in risky asset may well be influenced by demographic factors, income, and risk attitude, the decision is also likely to be affected by the households’ debt holding. for example, households with large portion of mortgage debt may hold safer assets to plan for expenses with fixed payment schedule (faig and shum, 2002). furthermore, optimal household portfolio may require the households to make asset allocation and debt allocation 326 r. tao, y. yuan / financial services review 27 (2018) 325-344 decisions together. policy-makers also noted the importance of analyzing household financial assets and liabilities together (brown et al., 2015). however, the interactions between the liability and asset of household finances are largely unknown and often lack theoretical modeling. given that assets and debt each display only one side of the household’s balance sheet and the decision of asset and debt allocations cannot be separated, we look at both sides of the balance sheet simultaneously via a bivariate model. our findings strongly support the hypothesis that both simultaneity and cross-causality effects affect the portfolio composition of households. this study provides several new findings. first, the ratio of secured debt does not significantly affect the household’s participation in risky asset investment, but it does have a significant impact on the portfolio share of risky assets, conditional on holding them. second, we find higher secured debt ratio is positively correlated with conditional risky asset shares. because a large proportion of household secured debt is mortgage debt, this result seems to contradict the well-known crowding-out effect of home ownership. third, we identify a set of factors (such as age and education of household head, income, and liquidity constraints) that significantly influence the debt structure and risky asset allocation simultaneously. the remainder of this paper proceeds as follows. section 2 provides theoretical background of the joint decision of asset and debt allocations and reviews the strand of literature related to both asset and liability of the household’s portfolio. section 3 describes methodologies and the econometric models. section 4 presents data selection and variable construction. in section 5, we report and analyze the empirical results. section 6 concludes the paper and provides policy recommendations. 2. joint decision of asset and debt allocations in theory, the demand for any asset or liability can be derived from a portfolio choice model in which households maximize expected utility subject to their budget constraints. consumers’ asset allocation decision should depend on the existing liability structure. for instance, holding a mortgage leaves the household with less spendable income, and the mortgage payments require the household to maintain certain liquid and less risky assets. on the liability side of the household portfolio, debt structure is also interdependent on asset allocation decisions. therefore, the allocation of debt and the allocation of assets must be considered jointly. however, empirical research on household portfolio structure often investigates a single choice at a time. studies on household risky asset allocation do not specifically test the effects of debt structure, while the research on consumer liability often focuses on analyzing the effect of credit and liquidity constraints that households face without considering the asset structure of the households (cf. brown et al., 2005; cox and jappelli, 1993). to our knowledge, yilmazer and devaney (2005) is the only study that demonstrates significant effects of financial assets on household debt. the cross-causality between debt and asset allocations can be reflected by the interactions between financial and real assets. cheung and miu (2015) demonstrate significant interaction effect between financial assets and home ownership. beaubrun-diant and maury (2016) analyze the simultaneous decisions of the households to participate in the stock market and 327r. tao, y. yuan / financial services review 27 (2018) 325-344 own their homes. they provide evidence that homeand stock-ownership decisions are taken simultaneously, therefore, rejecting the common view that these decisions are made sequentially. waggle and johnson (2003) examine the impact of home ownership on portfolio decisions relating to stocks and bonds. they find that young homeowners with high home value to net worth ratios should decrease the amount of stocks in their portfolios. with lower home to net worth ratios, investors can maximize utility by having their houses completely paid for and by holding more stocks. hu (2005) shows that homeowners hold a higher proportion of equity in liquid financial assets than renters do. existing literature only indirectly test the effect of household debt on risky asset allocation. most of these studies attempt to empirically identify factors explaining household financial asset allocation while adding mortgage debt as an explanatory variable (cf. cardak and wilkins, 2009; cocco, 2005; fratantoni, 1998). secured debt, particularly mortgage, causes liquidity constraint on the household that influences the household’s asset allocation choice. faig and shum (2002) identify real estate as both a risky investment and a personal illiquid project, which incurs penalties if discontinued. mortgage, property tax, and utility payments regularly generate liquidity needs. their results suggest that individuals who save to invest in their homes hold safer financial portfolios. recent studies on the effects of background risks on household portfolio allocation often consider mortgage payments as part of the “committed expenses.” research in this area has identified background risks as associated with a number of factors, including labor income risks, committed expenditure, proprietary business income risks, and health risks. background risks lead households to increase precautionary savings and reduce risky asset holdings. fratantoni (2001) finds that mortgage commitments and labor income risk reduce household risky assets holdings. it is worth noting that this strand of literature does not consider the household’s debt structure, which is defined as the ratio of outstanding secured debt balance to total household debt in this research. few studies incorporate both assets and debt in the analysis. brown and taylor (2008) attempt to identify the characteristics of the households that accumulate debt and/or financial assets and the determinants of net worth (i.e., the difference between household assets and debt). although this research considers both sides of the household balance sheet, it does not inform on the structure of assets and debt. cardak and wilkins (2009) acknowledge that risky asset holdings and committed expenses are jointly determined by the household, but they proceed without estimating the two quantities simultaneously. in this paper, we add to the literature by investigating the determinants of household risky asset holdings controlling for the debt structure and by examining the determinants of households’ financial portfolio incorporating financial asset and debt structures simultaneously. 3. the empirical models 3.1. a sample selection model our dependent variable is the share of risky assets, which is defined as the proportion of financial assets held in risky assets including equity and bonds. the dependent variable is not continuous and unbounded, so ols estimates would be biased. moreover, many households 328 r. tao, y. yuan / financial services review 27 (2018) 325-344 in the sample have no risky asset investment at all. the tobit model (tobin, 1958) takes into consideration the concentration of observations at zero. it also accounts for the fact that the explanatory variables may influence the probability of whether a household invest zero dollar in risky assets, and how much they actually invest, given that they invest something. tobit models are commonly used in the literature. for example, basnet and donou-adonsou (2016) apply a tobit model to analyze credit card balance which is either positive or equal to zero. brown and taylor (2008) treat household total assets and total debt as censored variables. they apply a univariate tobit specification to model total assets and total debt independently, and a bivariate tobit model to estimate the two quantities at the household level jointly. cox and jappelli (1993) take into account the selection bias caused by borrowing constraints and the household’s decision to hold positive debt, and apply a tobit model to estimate the optimal level of household debt. based on the prototypical tobit model (tobin, 1958) censoring from below at zero, where the latent variable y* is linear in regressors with additive error that is normally distributed and homoscedastic, y* can be expressed as y* � x�� � � (1) where the error term ��n [0, �2] has variance �2, which is assumed to be constant across observations. in our model, the latent variable y* is a household’s desired holding of risky assets. the observed y is the household’s actual risky assets holding. y can be expressed as y � � y* if y* � 0, 0 if y* � 0. (2) our interest is to derive the marginal effect in correspond to the effect of a change in a regressor on the desired risky asset holding (i.e., the latent variable), e� y*�x� x � �. (3) however, the latent variable is not observable. to derive the marginal effect of observed data y, first, we introduce an indicator variable, d, and d � � 1 if y* � 0, 0 otherwise. (4) following cameron and trivedi (2005), we can derive the censored mean by first conditioning the observable y on the binary indicator d and then unconditioning. the left-censored mean is e� y� � ed�ey�d� y�d�� � pr�d � 0� e� y�d � 0� � pr�d � 1� e� y�d � 1� � 0 pr� y* � 0� � pr� y* � 0� e� y*�y* � 0� � pr� y* � 0� e� y*�y* � 0� (5) 329r. tao, y. yuan / financial services review 27 (2018) 325-344 where pr[y* � 0] � 1 � pr[y* � 0] � pr[� � �x��]. the conditional means are given by e� y�x� � pr�� � � x��� x�� � e���� � � x��� � � x��/ x�� � x��/ (6) where (.) and �(.) are the standard normal distribution pdf and cdf, respectively. the marginal effect derived from observed censored data are given by �e� y�x� ��x � � � pr� y* � 0� � �����x � (7) the above tobit model restricts the censoring mechanism to be from the same model as that generating the outcome variable. in other words, the same set of variables and coefficients determine both the probability that an observation will be censored and the value of the dependent variable. this limitation can be remedied with the use of a sample selection model. we turn to a bivariate sample selection model that is defined in cameron and trivedi (2005),1 which comprises a participation equation that d � � 1 if y1* � 0, 0 otherwise, (8) and an outcome equation that y � � y2* if y1* � 0, not observed otherwise. (9) the latent variable y1 * determines whether or not the household invests in risky assets at all. y is observed if and only if y1 * � 0. the latent variable y2 * determines how much to invest, and y1 * � y2 *. the standard model specifies a linear model with additive errors for the latent variables, y1* � x1��1 � �1 (10) y2* � x2��2 � �2 (11) while in eq. (1), it is assumed that y1 * � y2 *. the allocation of risky asset is estimated using observations on only those households with positive holdings of risky asset. the sample selection bias induced by using only observations with positive values of risky asset holding can be corrected by a standard two-step procedure (heckman, 1979). we first estimate reduced-form probit equation for the participation probabilities of risky asset investment (eq. 8) and then include the estimated hazard as an additional regressor in the outcome equation (eq. 9).2 following the same derivation as eq, (6). the expected investment in risky asset, conditional on that the household invests is given by e� y�x, y1* � 0� � e� x2��2 � �2�x2��2 � �2 � 0� � x2��2 � e��2��1 � � x1��1� 330 r. tao, y. yuan / financial services review 27 (2018) 325-344 � x2��2 � �12�� x1��1�, (12) where �(z) � ��z� ��z� is the inverse mill’s ratio term (heckman, 1979). heckman’s two-step procedure is applied to estimate the positive values of y by an ols, yi � x2i��2 � �12�� x1i��̂1� � vi (13) where �̂1 is obtained by first-step probit regression of y1 on x1, and �(x�1i�̂1) is the estimated inverse mill’s ratio. if the independent variable only appears in the participation equation, we can use probit model marginal effect. define p(x) as the probability of participating in risky asset investment given x. the marginal effect of variable xj is given by �p� x� � xj � �j�� x��� (14) the marginal effect e[y�x]/ x is �2 if the independent variable only appears in the outcome equation. if the independent variable appears in both equations, the marginal effect is given by taking the partial derivatives of eq. 12 (derivation omitted). with the above sample selection model, we test how debt structure affects risky asset allocation. our key independent variable is the proportion of secured debt in total debt. as defined in the previous section, the secured debt includes the outstanding balance of the household’s mortgage and vehicle loans. we distinguish between secured and unsecured debt given the fact that unsecured debt is typically more expensive (higher interest rates) than secured debt. however, the adverse financial shock is less likely to cause immediate financial pressure on unsecured debt than secured debt (brown and taylor, 2008). in addition, the monthly payments of secured debt can be considered as part of the household’s committed expenses, which affect the household’s liquidity needs. 3.2. the bivariate model the empirical question is whether there is a relationship between risky financial asset holding and debt structure. the causal effect is not meaningful because the two quantities are clearly jointly determined by the household. households’ asset allocation choices are obviously bound by how much debt they have; when consumers borrow they need to consider how much assets they have. cardak and wilkins (2009) express concerns on the potential endogeneity problem because asset and debt allocations are often determined simultaneously, but they proceed to estimate reduced-form regressions with the risky asset ratio on the left hand side and measures of committed expenses on the right hand side. to examine the joint decision of asset and debt allocations, we apply a bivariate model, which is developed in the context of the joint distribution, assuming a bivariate normal distribution.3 the bivariate model allows for the possibility of interdependent decision making with respect to the share of risky assets (ya) and the share of secured debts (yd), both of which are 331r. tao, y. yuan / financial services review 27 (2018) 325-344 censored. each variable can be expressed by eqs. (1) and (2). the bivariate tobit model can be specified as follows, ya � � ya* if ya* � 0, 0 otherwise. (15) yd � � yd* if yd* � 0, 0 otherwise. (16) both latent variable ya* and yd* are linear model with additive errors, ya* � xh��1 � �h1 (17) yd* � xh��2 � �h2 (18) where xh is a vector of independent variables that affect household portfolio choices; �h1 and �h2 are the error terms which are jointly normally distributed with variances �h1 2 and �h2 2 , respectively, that is: �h1, �h1 � n�0, 0, �h1 2 , �h2 2 , �� (19) where the covariance is given by �h1h2 � ��h1�h2 (20) where � is the correlation coefficient of �h1 and �h2, which measures the degree of interdependence between ya* and yb*. a maximum-likelihood estimation is carried out to derive the coefficients for each equation, the cross-equation error correlations, and the variance of the error terms. if � is zero, the joint normal density function would collapse to the product of two independent normal density functions and a univariate approach of separating eqs. (17) and (18) would be appropriate. 4. data the data used in this paper comes from the 2016 survey of consumer finances (scf). the scf are sponsored by the federal reserve board in cooperation with the department of the treasury. scf has been conducted by the national opinion research center and the university of michigan since 1983. the scf is the most comprehensive data source on household financial information in the united states. the survey data in the scf include much information on households’ balance sheet, pensions, income, as well as detailed data on demographic characteristics. the scf data are not a panel data. some of the survey interviewees were selected from a standard multistage area-probability design, and the remaining were selected from a list sample derived from tax records by internal revenue service. the scf is conducted every three years. we choose the 2016 wave because it is the most recent data available and it made many changes from the previous 2013 wave. for example, there is a major update on the education loan section, the credit card section is reworked, and a new set of risk attitude 332 r. tao, y. yuan / financial services review 27 (2018) 325-344 variables is added. in addition, by the time of the 2016 survey the economy was out of the 2008 financial crisis, the impact of the subprime mortgage crisis has been fading out the economy, consumers face fewer credit constraints, and the unemployment rate is getting down from the peak. the 2016 scf data consists of the asset and debt holdings of 6,248 households. the data are imputed to account for the variability in the data because of missing information. in our empirical estimations, rubin’s combination rule (rubin, 1987) is applied to the estimated coefficients. the standard errors are also adjusted accordingly to generate the correct inference. because we are trying to analyze asset and debt choices, we exclude the households that do not have any financial assets or debt. it is also because these two variables appear in the denominator of two key variables. we screen out the observations with an extremely high value of net worth to control the impact of outliers. because of the survey design, the list of sample from irs tax records is likely to be relatively wealthy.4 therefore, we drop out the families in the top 5 percentile of assets, in the top 5 percentile of net worth, or in the top 5 percentile of annual income. we exclude households that reported labor income in the bottom 5 percentile as the low-income families may behave quite differently in investing or acquiring debt. the final sample contains 4,049 out of the original 6,248 observations. we include demographic variables that are commonly used in the literature as controls. they include the gender, age, race, and education of the household head, and number of children in the family. previous studies suggest that demographic characteristics contribute to the portfolio decision. moreover, the age pattern of risky portfolio shares is crucial to understanding portfolio behavior over the life cycle. therefore, we use dummy variables to represent each age category instead of the continuous age variable. addoum (2017) shows that couples significantly decrease their stock allocations after retirement, whereas singles’ allocations remain relatively unchanged. in addition, family size (browning, 1992), gender (bogan 2013), bequest motives (bertaut and haliassos, 1997), ethnicity (choudhury, 2001), and education (dimmock et al., 2016) are all proven to affect household financial asset allocations. we include the household’s income in the past year and net worth as independent variables. households with more income may be less vulnerable to the risk of their financial portfolios. perraudin and sørensen (2000)’s results suggest that a 10% proportional rise in wealth leads to a 24% and a 25% increase in stock and bond demand respectively. we take the natural log of total income to minimize the effect of outliers. many previous studies analyze how risk attitude affects household financial portfolio. riley and chow (1992) explore the relationships between asset allocation and individual risk aversion. they conclude that relative risk aversion decreases as one rises above the poverty level and decreases significantly for the wealthy households. hariharan et al. (2000) confirm the capm prediction that risk-tolerant investors hold a smaller fraction of their investments in the risk-free asset. we control for risk attitude by including an ordered categorical variable (“risk averse”) with values ranging from one to four. the household chooses one if it is willing to take substantial financial risks expecting to earn substantial returns. the value 2 means that the household is willing to take above average financial risks expecting to earn above average returns. value 3 means the household is willing to take average financial risks 333r. tao, y. yuan / financial services review 27 (2018) 325-344 expecting to earn average returns. value 4 means the household is not willing to take any financial risks. therefore, the higher the value for this variable, the more risk-averse the household is. we control for background risks by including measures of health risk and labor income risk.5 rosen and wu (2004) show that health is a significant predictor of both the probability of owning different types of financial assets and the share of financial wealth held in each asset category. fan and zhao (2009) provide the evidence that health shocks shift investment from risky assets toward other financial assets. therefore, we expect that poor health has a negative impact on risky asset holding. we create a binary variable for the household’s health status. the binary variable-“health risk” equals to one if either the head or his or her spouse expressed to have fair or poor health condition. to control for the labor income risk, we consider whether the head is self-employed, or own/share ownership in any privately-held businesses (“private business”). the rich who own private businesses are bound to consider business returns in selecting their financial portfolios. owning or investing in a private business may suggest that the household substitutes for investment in the stock market. moreover, private business may constantly generate liquidity needs (faig and shum, 2002) and being self-employed may expose the household to proprietary business risks (heaton and lucas, 2000). therefore, this variable may also measure the household’s liquid constraints. “committed expense” is defined as monthly mortgage payments and car loan payments divided by monthly income. it measures both background risks and liquidity constraints. additional controls for households’ liquidity constraints include the checking and saving account balances (again, we use the natural log form to dampen the effects of extreme values), a binary variable-“high expense” that takes on the value of one if the household had unusually high overall expenses in the past 12 months, the household’s total line of credit (loc), and total liquid assets that include the balance of all types of transaction accounts (liq). in the scf, families were asked “if you experienced a financial emergency, how would you deal with it?” the respondents choose from four options: 1 � borrow from others; 2 � spend from own savings; 3 � postpone payment; 4 � cut back spending. we use this variable (labeled “liq con”) as a proxy for the household’s liquidity constraint. the higher the value, the more liquidity constraints the household faces. to control for the credit constraints that the households face, we include two binary variables. the dummy variable “turned down” equals to one if the household applied for any type of credit in the past 12 months and feared denial or was turned down. variable “late payment” equals to one if the household had a late payment in the past 12 months. both of these variables represent the easiness that the household can raise money to invest. we do not control for interest rates. although a household’s debt and asset allocation decisions are likely to be affected by interest rates and expected rate of returns, the scf data does not provide this information. furthermore, we use cross-sectional data, so we assume that all of the households are subject to the same interest rate on debt (mortgage). an austrian survey finds that interest rates only have small effects on saving, portfolio and loan decisions (beer et al., 2016). moreover, the effect of loan interest rate differentials across households because of households’ credit worthiness should be captured by the credit constraint measures. personality factors such as impulsiveness, self-esteem, self-control, sensitivity, and so forth, may play an important role in consumer financial behaviors. compulsive shoppers 334 r. tao, y. yuan / financial services review 27 (2018) 325-344 often overspent when they use credit cards (basnet and donou-adonsou, 2016). norvilitis et al. (2006) and wang et al. (2011) find that impulsiveness is significantly correlated with revolving credit card debt. hira et al. (1993) report that internal locus of control is associated with optimism about one’s financial future, but norvilitis et al. (2006) find no relationship between locus of control and amount of debt in college students. because of the limited availability of data in scf and the consideration that this study targets on the household units instead of the individual consumers, such personality factors are not included. table 1 shows the summary statistics of all the variables. risky asset ratio has a mean of 28.3% and a median of 20.8%, which is consistent with the evidence provided in the literature that many households hold few or no risky assets in their portfolio (campbell, 2006). fig. 1 is a histogram showing the distribution of our dependent variable-the share of risky assets in the household’s financial assets. on average the households in our sample hold 62.5% of total debt as secured debt and the median is higher at 84.2%. the demographic statistics summary shows 78% of household head are male, 66.5% are households with a married couple, 67.8% of household head are white and non-hispanic, and, except for 5.5% of household heads who are over 75 years old, the sample data are evenly distributed among each age group. 16.4% of households are self-employed, while 21% of households own or share ownership in privately held businesses. only 3.1% of households are unemployed. one key observation in table 1 is that, although the average households’ total assets are $1.1 million, 50% of the families have less than $280,000 assets. the sample shows an average liquid assets of $47,000 and an average net worth of $0.9 million with a lower median at $161,000. the households in our sample display an overall high degree of risk aversion (3.032 out of 4) and a medium level of liquidity constraint (1.8 out of 4); 20% of households feared denial or was turned down when applied for credit in the previous year, 15% had late payments in the last year, and 26.1% of families’ head and/or spouse self-evaluated their health condition as poor or fair. 5. results 5.1. determinants of risky asset allocation we first test the effect of debt structure on risky asset holdings by using a two-step heckman estimation scheme on the tobit model. this estimation scheme separates the decision to participate in investing risky assets from the decision on how much share of risky assets to hold. the same set of independent variables is included in both the participation equation and the outcome equation.6 this allows us to identify what factors prompt the household to enter the market for risky assets and what factors influence the portfolio share of risky assets, conditional on holding them. in addition, the inverse mill’s ratio is added in the outcome equation as an independent variable to avoid selection bias (king and leape, 1998). the results show that the inverse mill’s ratio is significant, which suggests that there would be a possible selection bias in the analysis of risky asset holding if not considering the market participation decision. table 2 presents the results. 335r. tao, y. yuan / financial services review 27 (2018) 325-344 our key independent variable, the ratio of secured debt to total debt, has a positive and significant coefficient only in the outcome equation. it implies that the share of debt in secured debt does not affect the household’s decision on whether to invest in a risky asset, but households with relatively more secured debt are likely to invest relatively more in risky assets. the marginal table 1 summary statistics variable name variable description mean median standard deviation min max risky ratio risky asset holdings as a % of financial assets 28.336 20.816 29.854 0 100 secured debt ratio secured debt holdings as a % of total debt 62.451 84.239 41.215 0 100 hhsex household head gender (1 � male) 0.781 1 0.413 0 1 age � 35 age less than 35 years 0.173 0 0.378 0 1 age 35–44 age between 35 � 44 0.205 0 0.404 0 1 age 45–54 age between 45 � 54 0.227 0 0.419 0 1 age 55–64 age between 55 � 64 0.216 0 0.411 0 1 age 65–74 age between 65 � 74 0.124 0 0.330 0 1 age �75 age beyond 75 years (omitted in regressions) 0.055 0 0.228 0 1 education education: 1 � no high school diploma/ged; 2 � high school diploma or ged; 3 � some college or associative degree; 4 � bachelor’s degree or higher 3.036 3 0.986 1 4 married marital status (1 � married; not included in regressions because of multicollinearity with hhsex) 0.665 1 0.472 0 1 kids number of kids 0.920 0 1.170 0 7 race race (1 � white and non-hispanic) 0.678 1 0.467 0 1 employed employed (1 � yes) (not included in regressions because of multicollinearity) 0.784 1 0.412 0 1 self-employed self-employed (1 � yes) 0.164 0 0.370 0 1 unemployed unemployed (1 � yes) (not included in regressions because of multicollinearity) 0.031 0 0.174 0 1 retired retired (1 � yes) 0.185 0 0.388 0 1 asset total assets (’000) 1100 280 2500 0.001 23000 ln (income) annual income (log) 11.281 11.205 0.928 9.485 14.422 ln (checking) checking account balance (log) 7.244 7.784 2.804 0 14.078 ln (saving) saving account balance (log) 4.737 5.602 4.474 0 15.464 net worth net worth (’000) 914 161 2372 �2000 23000 risk aversion risk aversion (1 � take substantial financial risks expecting to earn substantial returns; 2 � take above average financial risks expecting to earn above average returns; 3 � take average financial risks expecting to earn average returns; 4 � not willing to take any financial risks) 3.032 3 0.839 1 4 turned down applied for any type of credit in past 12 months, feared denial or was turned down (1 � yes) 0.201 0 0.401 0 1 late pay have any late payment last year (1 � yes) 0.154 0 0.359 0 1 high expense have unusually high expenses last year (1 � yes) 0.260 0 0.439 0 1 committed expense ratio of committed expenses (monthly mortgage and car loan payments) to income (%) 15.700 13 14.100 0 98.800 private business own or share ownership in any privately held businesses (1 � yes) 0.210 0 0.408 0 1 liq con if experience a financial emergency, 1 � borrow from others; 2 � spend from own savings; 3 � postpone payments; 4 � cut back spending 1.800 2 1.038 0 4 loc total lines of credit (’000) 18.742 0 154.067 �0.002 7450 liq liquid assets (’000)-balance of all types of transaction account 47.069 6.592 182.790 0 5236.5 health risk health risk (1 � either head or spouse responded “poor” or “fair” health condition 0.261 0 0.439 0 1 notes: this table reports the summary statistics for the outcome and control variables. not all variables are included in the regressions because of multicollinearity problem. the number of observations � 4,049. 336 r. tao, y. yuan / financial services review 27 (2018) 325-344 effect shows an economic significance that if the secured debt share increases by 1%, the risky assets holding by an average household who already holds risky assets will increase by $13,531.7 because most of the secured debt is mortgage, this result seems to contradict the well-known crowding-out effect of home ownership. one of the explanations could be in line with the argument made by brown and tyler (2008) that households with relatively more secured debt need higher expected returns from risky asset to pay for the debt. these households may suffer more financial distress during economic down turns because financial shocks are more likely to cause immediate financial pressure on secured debt. on the other hand, these households could take advantage of the lower cost debt in exchange for relatively higher return assets. this result also partially supports beaubrun-dian and maury (2016)’s finding that previous homeowners are more likely to become stockholders. the variables that are significant in both the participation and the outcome equations include “education,” “race,” “income,” “risk averse,” and “private business.” better educated household heads, or households with higher annual income are more likely to invest in risky assets, and are more likely to hold a larger proportion of risky assets in their financial portfolios. whites are more likely to invest in risky assets and they tend to invest 5.131% more in risky assets than non-whites. as expected, households that are more risk averse or own private businesses are less likely to invest in risky assets, and if they do invest in risky assets they tend to hold relatively less (approx. 5.781%) risky assets. our result shows that a household owing a private business holds approximately 5.407% less in risky assets shares than the household without ownership in private business. this evidence supports the argument that private business is an investment substitute as private business owners are already exposed to market risks. this finding is consistent with part of the findings by faig and shum (2002) that households saving to invest in their own businesses have significantly safer financial portfolios. fig. 1. distribution of risky asset allocation. 337r. tao, y. yuan / financial services review 27 (2018) 325-344 while the above results are mostly consistent with the existing literature, we do find new and improved evidence. our results suggest that households in the early life cycle (with a younger head less than 35 years old) and the later life cycle (retired) are less likely to participate in risky asset investment with marginal effects of �10.69% and �8.71%.8 the young typically are faced with credit constraints and with limited cash, while the retired are concerned with the easiness that stocks can be liquidated. furthermore, our results show that the age profile concerns the decision to enter and exit the market for risky assets, not managing the portfolio share. similarly, “turned down,” which is used as a proxy for the household’s credit constraint, and “health risk” are also significant determinants only in the participation equation. higher health risk reduces the probability to invest in risky assets by a marginal effect of 5.38%, but it does not significantly affect the conditional risky asset shares. table 2 heckman two-stage selection model explanatory variables (1) prob[y�0] (2) y�y�0 coefficient standard error coefficient standard error secured debt ratio 0.001 0.001 0.040** 0.018 hhsex �0.099 0.061 3.984** 1.926 age � 35 �0.299** 0.132 �4.610 4.038 age 35–44 �0.071 0.132 0.826 3.590 age 45–54 0.006 0.127 3.347 3.484 age 55–64 0.078 0.120 0.602 3.405 age 65–74 0.043 0.117 �4.286 3.328 education 0.165*** 0.027 2.769*** 0.941 kids �0.023 0.023 �0.710 0.619 race 0.348*** 0.051 5.131*** 1.759 retired �0.246*** 0.081 0.482 2.260 ln(income) 0.693*** 0.048 6.613*** 1.791 ln(checking) 0.028*** 0.006 �0.609*** 0.153 ln(saving) 0.032*** 0.010 �0.447* 0.258 net worth 0.000 0.000 0.001*** 0.000 risk aversion �0.251*** 0.030 �5.781*** 1.016 turned down �0.142** 0.062 �3.023 1.974 late pay �0.022 0.066 �1.924 2.103 high expense �0.007 0.055 0.145 1.701 committed expense 0.172 0.197 1.990 5.900 private business �0.261*** 0.069 �5.407*** 1.967 liq con �0.016 0.022 �0.813 0.744 loc 0.000 0.000 0.000 0.003 liq 0.000 0.000 �0.008** 0.003 health risk �0.154*** 0.055 �0.881 1.996 constant �7.329*** 0.509 �31.277 23.287 inverse mills ratio 14.787** 6.056 notes: ***, **, and * indicate significance at 1%, 5%, and 10% level, respectively. (1) this table shows the two-stage heckman regression results. the column 1 is the first-stage participation equation results, in which the dependent variable equals one if the respondent reports ownership of any risky financial assets. the column 2 is the second-stage outcome equation results, in which the dependent variable is the percentage of financial assets invested in risky assets. the number of observations is 4,049. (2) because the data is imputed, rubin’s combination rule (rubin, 1987) is applied to the estimated coefficients. the standard errors are also adjusted accordingly to generate the correct inference. 338 r. tao, y. yuan / financial services review 27 (2018) 325-344 gender of household head and net worth are only significant in the outcome equation. male-headed households tend to invest 3.984% more in risky assets, which is consistent with the evidence presented in the existing literature. however, they are not more likely to invest in risky assets than female-headed households. households with higher net worth invest relatively more in risky asset (they can tolerate the risk better than low net worth households), but net worth is not a significant predictor of the decision to invest in any risky asset in the first place. this result contradicts the classical prediction that, after controlling for risk attitude, the portfolio share of risky assets, conditional on holding it, should be independent of the level of wealth (guiso et al., 2002). now turning to the liquidity constraint measures (“liq,” “liq con,” “committed expense,” “high expense,” “loc”), only “liq” has a significant negative coefficient in the outcome equation. contrary to the findings of previous studies on the committed expense risks, our result suggests that committed expenses neither affect the participation nor the conditional risky asset shares. another key finding is that households with higher saving and checking account balances are more likely to enter the market for risky assets, but for those already investing in risky asset, households with more cash on hand tend to invest less in risky assets. 5.2. testing the joint decision of asset and debt allocations we apply a bivariate tobit model, which allows the potential simultaneity in the decision of households to hold risky assets and to hold secured debt (see eqs. 15 and 16). both the univariate and bivariate tobit models are estimated for comparison and as robustness check. tables 3a and 3b show the results. the interdependence of the two dependent variables is tested by the likelihood ratio test on the correlation coefficient-� defined in eq. (20). � is constrained at zero in the univariate case. the likelihood ratio test statistic follows asymptotically a �2 distribution with one degree of freedom under the null hypothesis that there is no interdependence in the data, � � 0. the test statistic is large enough to reject the null hypothesis at the 1% level. therefore, the simultaneous equation bivariate tobit model used here is appropriate to analyze the two decisions. the share of risky asset and the share of secured debt are likely to be jointly determined by the households. by taking the potential simultaneity into consideration, our bivariate model provides further evidence. compared with the coefficients from the univariate tobit model, “age � 35,” the number of kids, “retired,” health risk, “turned down,” and “committed expense” become significant predictors of unconditional risky asset holding in the bivariate model, while the coefficients of gender of household head and net worth become insignificant. for example, the retired turns to invest 5.23% less and poor health people invest 4.108% less in risky assets. checking account balance has a positive and significant coefficient of 0.598 for secured debt share. it has no impact on the risky asset share as opposed to the significant negative coefficient (�0.609) in the univariate case. contrary to the evidence provided in the literature, our bivariate model results show that one percentage higher monthly committed expenses to monthly income ratio leads to a 0.129% larger share of financial asset allocated to risky assets. this can be explained by the fact that higher debt obligations push the household to seek higher return investment opportunities. 339r. tao, y. yuan / financial services review 27 (2018) 325-344 our results identify the types of households that typically hold relatively more secured debt and more risky assets. these households earn higher income, have a white head of household, have higher committed expenses ratio, or are less credit constrained (did not fear denial or was not turned down for credit). higher income households may better take advantage of lower interest rate on secured debt and exploit the wealth-generating potential of the equity premium. the types of households that hold relatively less secured debt and less risky assets are the younger households (�35 years old), or owning a private business, which is consistent with the findings from the existing literature. the younger households tend to hold 10.381% less on risky asset and 10.572% less on secured debt. the households owning a private business hold 9.034% less risky asset as well as 5.504% less secured debt. well educated or higher net worth households tend to have relatively 5.304% more risky assets but 4.915% less secured debt. “retired,” “health risk,” “risk averse,” family size, checking account balances and total liquid assets (liq) are significant predictors for the share of risky assets, but not for the debt table 3 a estimation results of univariate tobit model explanatory variables (1) risky asset ratio (2) secured debt ratio coefficient standard error coefficient standard error hhsex 2.243 2.046 7.515*** 1.935 age � 35 �10.377** 4.058 �10.537** 4.130 age 35–44 �2.232 3.918 �5.454 4.040 age 45–54 1.042 3.766 �2.711 3.884 age 55–64 0.585 3.670 �0.764 3.666 age 65–74 �3.558 3.743 2.146 3.657 education 5.354*** 0.904 �4.902*** 0.857 kids �1.444** 0.694 1.097 0.687 race 10.590*** 1.574 5.432*** 1.622 retired �5.226** 2.378 1.906 2.453 ln(income) 17.363*** 1.266 21.655*** 1.267 ln(checking) 0.021 0.162 0.596*** 0.170 ln(saving) 0.593** 0.276 �0.115 0.288 net worth 0.000 0.000 �0.002*** 0.000 risk aversion �9.008*** 0.880 0.668 0.932 turned down �6.530*** 2.061 �9.184*** 1.992 late pay �2.095 2.264 �10.741*** 2.182 high expense �0.634 1.750 �4.530*** 1.629 committed expense 12.435** 5.395 168.956*** 5.245 private business �9.023*** 2.001 �5.497*** 1.945 liq con �0.848 0.760 2.236*** 0.698 loc �0.001 0.004 �0.002 0.005 liq �0.010*** 0.004 �0.003 0.005 health risk �4.084** 1.799 �2.629 1.730 constant �172.431*** 13.915 �206.994*** 14.181 � 0 0 �1 32.40 — �2 — 43.03 notes: ***, **, and * indicate significance at 1%, 5%, and 10% level, respectively. (1) eqs. (1) and (2) are estimated separately. (2) because the data is imputed, rubin’s combination rule (rubin, 1987) is applied to the estimated coefficients. the standard errors are also adjusted accordingly to generate the correct inference. 340 r. tao, y. yuan / financial services review 27 (2018) 325-344 structure, while gender of household head, saving account balances, net worth, “late pay,” “high expense,” “liq con” only significantly affect the secured debt share. the above evidence demonstrates that when households decide on risky asset holdings, they do not make the decision in isolation from their debt structure and obligations, vice versa. the bivariate tobit model setup allows us to identify, in this simultaneous decisionmaking process, the demographic and financial factors that can contribute to the household overall financial portfolio structure. 6. conclusions analyzing the allocation of financial portfolio, including financial assets and liabilities, is of paramount importance for economic policy making. this is especially imperative at the table 3 b estimation results of bivariate tobit model explanatory variables (1) risky asset ratio (2) secured debt ratio coefficient standard error coefficient standard error hhsex 2.189 2.044 7.498*** 1.936 age � 35 �10.381** 4.060 �10.572** 4.132 age 35–44 �2.228 3.919 �5.475 4.041 age 45–54 1.007 3.766 �2.735 3.885 age 55–64 0.562 3.670 �0.793 3.667 age 65–74 �3.541 3.743 2.128 3.658 education 5.304*** 0.907 �4.915*** 0.857 kids �1.446** 0.695 1.092 0.687 race 10.595*** 1.574 5.433*** 1.622 retired �5.230** 2.378 1.893 2.454 ln(income) 17.444*** 1.270 21.688*** 1.267 ln(checking) 0.024 0.162 0.598*** 0.170 ln(saving) 0.594** 0.276 �0.113 0.288 net worth 0.000 0.000 �0.002*** 0.000 risk aversion �8.997*** 0.880 0.672 0.933 turned down �6.457*** 2.058 �9.175*** 1.991 late pay �2.111 2.265 �10.770*** 2.184 high expense �0.602 1.750 �4.542*** 1.629 committed expense 0.129** 5.425 1.691*** 5.248 private business �9.034*** 2.001 �5.504*** 1.946 liq con �0.851 0.760 2.238*** 0.699 loc �0.001 0.004 �0.002 0.005 liq �0.010*** 0.004 �0.003 0.005 health risk �4.108** 1.802 �2.639 1.731 constant �173.274*** 13.968 �207.338*** 14.187 � 0.055*** 0.018 �1 35.694 �2 42.818 notes: ***, **, and * indicate significance at 1%, 5%, and 10% level, respectively. likehood ratio test: h0: �12 � 0. �2 � 50.450, p � 0.000. (1) eqs. (1) and (2) are estimated simultaneously. (2) because the data is imputed, rubin’s combination rule (rubin, 1987) is applied to the estimated coefficients. the standard errors are also adjusted accordingly to generate the correct inference. 341r. tao, y. yuan / financial services review 27 (2018) 325-344 household level as it indicates the financial pressure and stress faced by the households in the quick-changing economic environment. in this research, we first test the effect of secured debt on risky assets by using a two-step tobit model. we separate the decision to participate in investing in risky assets from the decision on how much risky assets to hold and provide improved evidence upon the existing literature. we find robust relationships between debt structure and asset allocation. although the debt structure does not prompt households to start investing in risky assets, it does affect the share of financial assets allocated in risky assets for those that are already investing. we then allow the potential simultaneity in the decisions of household debt structure and asset allocation by applying a bivariate tobit model. this setup clarifies some contradicting conclusions from the previous studies that neglect the simultaneity of choices and allows us to identify the demographic and financial factors that contribute to the households’ overall financial portfolio structure. it also allows us to gain insights into the factors affecting the households’ vulnerability to adverse changes in the financial market. the positive relationship between the committed expense ratio and the share of risky assets investment supports the argument that households invest in risky assets to pay off the debt (brown and tyler, 2008). these are the households that are more vulnerable in the time of financial market down turns. holding more risky assets with higher secured debt obligations leaves the household with an unbalanced financial portfolio exposed to leveraged financial risks. such households need to be targeted for financial advice. better educated household heads seem to better take advantage of higher return from risky assets without being significantly constrained by the illiquidity caused by the secured debt. lower income households are less likely to participate in investing risky assets and tend to invest less, but they hold relatively more unsecured debts, which typically incur higher interest expenses. if provided with the investment opportunities and are better informed about financial matters, they may be able to exploit the wealth-generating potential of the equity premium. this further affirms the importance of financial literacy education. notes 1 cox and jappelli (1993) introduce a three-equation generalized tobit model, where the authors add one additional participation equation. 2 for probit models with standard normal density, the hazard is equal to the inverse mill’s ratio. 3 similar setup is adopted by brown and taylor (2008). 4 to obtain a clearer picture of how aggregate holdings of various asset categories are related to household-level characteristics, the scf tends to oversample wealthy households, because the wealthy segment holds most assets (guiso et. al., 2002; perraudin and sorensen, 2000). 5 cardak and wilkins (2009) point out that labor uncertainty and health risk are the two major background risk factors. 6 some variables, for example, marital status, are omitted to avoid multicollinearity problem. the remaining independent variables in the equations all have very low correlation coefficients. 342 r. tao, y. yuan / financial services review 27 (2018) 325-344 7 as discussed in section 3.1. (under eq. 13), the marginal effect �e[y�x]/�x is �2 in the outcome equation. the economic significance of secured debt is calculated using median values of households’ financial assets ($48,690), risky assets ($8,345), riskyasset ratio (0.4093), and the marginal effect (estimated coefficient) of 0.04. 8 as discussed in section 3.1. (under eq. 13), the marginal effect of variable xj in the participation equation is given by �p�x� �xi � �j��x���. the marginal effect of a dummy variable is the change in predicted probabilities when the dummy variable changes from 0 to 1 while all other variables are kept at their means. references addoum, j. m. (2017). household portfolio choice and retirement. review of economics and statistics, 99, 870–883. bailey, w., kumar, a., & ng, d. (2011). behavioral biases of mutual fund investors. journal of financial economics, 102, 1–27. basnet, h. c., & donou-adonsou, f. (2016). internet, consumer spending and credit card balance: evidence from u.s. consumers. review of financial economics, 30, 11–22. basnet, h. c., & donou-adonsou, f. (2018). marriage between credit cards and the internet: buying is just a click away! review of financial economics, 36, 252–266. beaubrun-diant, k. e., & maury, t. p. (2016). home tenure, stock market participation, and composition of the household portfolio. journal of housing economics, 32, 1–17. beer, c., gnan, e., & ritzberger-grünwald, d. (2016). saving, portfolio and loan decisions of households when interest rates are very low-survey evidence for austrian households. monetary policy & the economy, 1, 14–32. bergstresser, d., & poterba, j. (2004). asset allocation and asset location: household evidence from the survey of consumer finances. journal of public economics, 88, 1893–1915. bertaut, c. c., & haliassos, m. (1997). precautionary portfolio behavior from a life-cycle perspective. board of governors of the federal reserve system working paper, no. 542. bilias, y., georgarakos, d., & haliassos, m. (2010). portfolio inertia and stock market fluctuations. journal of money, credit and banking, 42, 715–742. bogan, v. l. (2013). household investment decisions and offspring gender: parental accounting. applied economics, 45, 4429–4442. brown, s., garino, g., taylor, k., & price, s. w. (2005). debt and financial expectations: an individual and household level analysis. economic inquiry, 43, 100–120. brown, s., ghosh, p., su, l., & taylor, k. (2015). modelling household finances: a bayesian approach to a multivariate two-part model. journal of empirical finance, 33, 190–207. brown, s., & taylor, k. (2008). household debt and financial assets: evidence from germany, great britain and the usa. journal of the royal statistical society: series a (statistics in society), 171, 615–643. browning, m. (1992). children and household economic behavior. journal of economic literature, 30, 1434– 1475. cameron, a. c., & trivedi, p. k. (2005). microeconometrics: methods and applications. cambridge: cambridge university press. campbell, j. y. (2006). household finance. journal of finance, 61, 1553–1604. campbell, j. y., chanb, y. l., & viceira, m. (2003). a multivariate model of strategic asset allocation. journal of financial economics, 67, 41–80. cardak, b. a., & wilkins, r. (2009). the determinants of household risky asset holdings: australian evidence on background risk and other factors. journal of banking & finance, 33, 850–860. 343r. tao, y. yuan / financial services review 27 (2018) 325-344 cheung, s. c., & miu, p. (2015). home ownership decision in personal finance: some empirical evidence. financial services review, 24, 51–57. choudhury, s. (2001). racial and ethnic differences in wealth and asset choices. social security bulletin, 64, 1–15. cocco, j. f. (2005). portfolio choice in the presence of housing. review of financial studies, 18, 535–567. cox, d., & jappelli, t. (1993). the effect of borrowing constraints on consumer liabilities. journal of money, credit and banking, 25, 197–213. dimmock, s. g., kouwenberg, r., mitchell, o. s., & peijnenburg, k. (2016). ambiguity aversion and household portfolio choice puzzles: empirical evidence. journal of financial economics, 119, 559–577. faig, m., & shum, p. (2002). portfolio choice in the presence of personal illiquid projects. journal of finance, 57, 303–328. fan, e., & zhao, r. (2009). health status and portfolio choice: causality or heterogeneity? journal of banking & finance, 33, 1079–1088. fratantoni, m. c. (2001). homeownership, committed expenditure risk, and the stockholding puzzle. oxford economic papers, 53, 241–259. fratantoni, m. c. (1998). homeownership and investment in risky assets. journal of urban economics, 44, 27–42. guiso, l., haliassos, m., & jappelli, t. (2002). household portfolios. cambridge, ma: the mit press. hariharan, g., chapman, k. s., & domian, d. l. (2000). risk tolerance and asset allocation for investors nearing retirement. financial services review, 9, 159–170. heaton, j., & lucas, d. (2000). portfolio choice in the presence of background risk. the economic journal, 110, 1–26. heckman, j. j. (1979). sample selection bias as a specification error. econometrica, 43, 153–161. hira, t. k., fitzsimmons, v. s., hafstrom, j. l., & bauer, j. w. (1993). factors associated with expectations of household’s future financial condition. journal of family and economic issues, 14, 237–256. hu, x. (2005). portfolio choices for homeowners. journal of urban economics, 58, 114–136. king, m. a., & leape, j. i. (1998). wealth and portfolio composition: theory and evidence. journal of public economics, 69, 155–193. norvilitis, j. m., merwin, m. m., osberg, m. t., patricia, v., roehling, v. p., young, p., & kamas, m. m. (2006). personality factors, money attitudes, financial knowledge, and credit-card debt in college students. journal of applied social psychology, 36, 1395–1413. perraudin, w. r., & sørensen, b. e. (2000). the demand for risky assets: sample selection and household portfolios. journal of econometrics, 97, 117–144. riley jr., w. b., & chow, k. v. (1992). asset allocation and individual risk aversion. financial analysts journal, 48, 32–37. rosen, h. s., & wu, s. (2004). portfolio choice and health status. journal of financial economics, 72, 457–484. rubin, d. b. (1987). multiple imputation for nonresponse in surveys. new york, ny: john wiley and sons. tobin, j. (1958). estimation of relationships for limited dependent variables. econometrica, 26, 24–36. waggle, d., & johnson, d. t. (2003). the impact of the single-family home on portfolio decisions. financial services review, 12, 201–217. wang, g. n., & hanna, s. d. (2019). racial/ethnic disparities in high return investment ownership: a heckman selection model. applied economics letters, 26, 111–115. wang, l., lu, w., & malhotra, n. k. (2011). demographics, attitude, personality and credit card features correlate with credit card debt: a view from china. journal of economic psychology, 32, 179–193. yilmazer, t., & devaney, s. a. (2005). household debt over the life cycle. financial services review, 14, 285–304. 344 r. tao, y. yuan / financial services review 27 (2018) 325-344 pii: 1057-0810(92)90008-z financial services review, 2(2): 157-168 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any fom reserved. abstracts of articles on individual financial management edited by phyllis schiller myers virginia commonwealth university consumption-savings behavior the timing of intergenerational transfers, tax policy, and aggre gate savings, by david altig (federal reserve bank of cleveland) and steven j. davis (university of chicago). we analyze an overlapping-generations framework that accommodates two observations: (i) the interest rate on consumption loans exceeds the rate of return to savings, and (ii) private intergenerational transfers primarily occur early in the life cycle. assuming altruistically motivated transfers in at least some family lines and other plausible conditions, we prove the invariance of capital’s steady-state mar ginal product to government debt, government expenditures, and the tax rates on labor and capital income. we show that the tax treatment of household interest payments has powerful effects on capital intensity and aggregate savings in life-cycle and, especially, altruistic linkage models. the american economic re view, december 1992, 82(5): 1199-1220. (reprinted with permission of the american economic review.) the impact of the demographic transition on capital formation, by a. j. auerbach and l. j. kotlikoff. the population of the united states is aging. the authors review a variety of the implications this has for u.s. national saving rates and discuss the policy issues that they raise. after reviewing what different models would predict for household saving over the next several decades, they consider how the demographic transition may also affect national saving through changes in government behavior. ways in which the composition of household saving might change as individuals age are also analyzed along with the implications of changes in government fiscal policy for asset composition. scandinavian journal of economics, 1992, 94(2): 281-195. (reprinted with permission of the journal of economic literature.) 158 financialservicesreview,2(2) 1993 time constraints in consumption and savings behavior, by vito tanzi and howell h. zee (international monetary fund, washington, dc). this paper investigates the implications for savings behavior of time con straints in consumption in a simple life-cycle model, and finds that some standard results in the literature on savings and taxation need to be modified. journal of public economics, february 1993, 50(2): 253-259. (reprinted with permission of north-holland publishing company.) earnings uncertainty and precautionary saving, by luigi guiso (bank of italy, rome, italy), tullio jappelli (istituto universitario navale, naples, italy), and daniele terlizzese (bank of italy, rome, italy). we test for the presence of precautionary saving using a self-reported measure of earnings uncertainty drawn from the 1989 italian survey of household income and wealth. the effect of uncertainty on wealth accumulation is consistent with the theory of precautionary saving and with decreasing prudence, but explains only a small fraction of saving. the results cast doubts on the empirical relevance of precautionary saving in response to earnings uncertainty, but are not in contrast with the importance of the precautionary motive per se. beside earnings uncertainty, other risks, such as health and mortality, may be important determinants of wealth accumulation. journal ofmonetary economics, november 1992, 30(2): 307-337. (reprinted with permission of north-holland publishing company.) wealth seeking reconsidered, by kai a. konrad (university of munich). if individuals are interested in their relative wealth position, they engage in contests for wealth. this leads to a steady state with overaccumulation. in a simple growth model it is shown that, if some individuals are engaged in wealth seeking activity, but others are not, the latter might benefit. however, wealth seeking is not welfare enhancing under usual assumptions concerning production technology. the steady-state equilibrium is characterized by a marginal productivity of capital that is below the steady-state rate of time preference and by emergence of a class structure. journal of economic behavior and organization, july 1992, 18(2): 215-227. (reprinted with permission of north-holland publishing company.) estateplanningandwealthtransfer saving, wealth, and the exchange-bequest motive, by w. a. lord. a model of exchange-motivated bequests based upon childrens’ “attentions” to parents, provided when the children are young adults, is joined to a standard abshzcts of articles on individual financial mangement 159 life-cycle specification to examine the consequences for wealth and savings. the results of general equilibrium simulations show contribution to be far below some recent estimates. exchange-motivated bequests produce offsets to life-cycle savings that limit their ability to increase total savings. canadian journal of economics, august 1992, 25(3): 743-753. (reprinted with permission of the journal of eco nomic literature.) patterns of intergenerational mobility in income and earnings, by h. elizabeth peters (university of colorado). this paper characterizes the patterns of intergenerational mobility in the united states using data for matched parent/child pairs from the national longitu dinal surveys. in general, what is found is far from the extremes of either perfect mobility or perfect immobility. parents’ log income explains only about 9% to 11% of the variation in children’s log incomes. earnings exhibit more mobility than does total income, and the difference is most striking for daughters. the paper also identifies the influence of family background characteristics on mobility. the addition of these background variables adds another three to five percentage points to the r* in the intergenerational earnings and income regressions. the review of economics and statistics, august 1992,74(3): 456-466. (reprinted with permission of elsevier science publishers b.v., north-holland.) individualfinancialmanagement the effect of borrowing constraints on consumer liabilities, by donald cox (boston college) and tullio jappelli (instituto di studi economici, i.u.n., naples). this paper explores the effects of liquidity constraints on consumer liabilities. while much empirical evidence attests to the importance of liquidity constraints in the u.s. economy, evidence about the effects of borrowing constraints on consumer balance sheets is scarce. using the 1983 survey of consumer finances data we estimate desired borrowing for unconstrained households. we then evaluate the gap between predicted and observed debt for the sample of liquidity constrained consumers. predicted debt is 75 percent higher than actual debt in the liquidity constrained sample. thus, the effect of removing borrowing constraints has quantitatively important implications for the allocation of debt in the household portfolio. the removal of borrowing constraints would raise aggregate house hold liabilities by 9 percent. journal of money, credit, and banking, may 1993, 25(2): 197-213. (reprinted with permission of journal of money, credit, und bunking.) 160 financial services review, 2(2) 1993 tax timing with liquidity constraints: a heterogeneous agent model, by betty c. daniel (state university of new york, albany). this paper considers the ricardian equivalence hypothesis in a model in which some families face binding liquidity constraints and others do not. the source of heterogeneity which generates binding constraints in some families, but not in others, is shown to be the rate of intergenerational discount. in an economy populated by these two types of families, a change in tax timing has non-ricardian short-run effects, but many ricardian long-run effects. this implies that heteroge neous agent models can reconcile some of the conflicting empirical evidence on ricardian equivalence. journal uf&ney, credit, and bunking, may 1993,25(2): 176-196. (reprinted with permission of journal ofmoney, credit, and bunking.) aging and the income value of housing wealth, by steven f. venti and david a. wise (national bureau of economic research, harvard university, cambridge, ma). the potential of reverse annuity mortgages to increase the current income of the elderly is analyzed. we conclude that most low-income elderly also have little housing equity. in general, a reverse annuity mortgage would substantially affect the income only of the single elderly who are very old-whose life expectancy is short. on the other hand, if the transfer were in the form of a lump-sum payment, rather than an annuity, the payment would increase the liquid wealth of most elderly families by a large fraction. thus legislation that would facilitate the market for reverse mortgages could improve substantially the financial status of a small proportion of the elderly. journul of public economics, april 199 1,44(3): 37 l-397. (reprinted with permission of north-holland publishing company.) predicting personal debt and debt repayment: psychological, so cial, and economic determinants, by sonia m. livingstone (london school of economics and political science) and peter k. lunt (univer sity college london, uk). while personal debt has grown rapidly in the united kingdom in recent years, posing problems for individuals, families and society, little empirical research has been conducted to date on everyday experiences of debt. the present paper reports on the findings of an in-depth survey of the social, economic and psychological factors related to debt. discriminant function analysis and multiple regression analysis were used to address three questions: what discriminates debtors from nondebtors; what determines how far people get into debt; and what determines how much of their debts people repay? sociodemographic factors were found to play a relatively minor role in personal debt and debt repayment. disposable income did not differ between those in debt and not in debt, although it predicted how far people abstracts of articles on individual financial mangement 161 were in debt and was most important in determining debt repayment. attitudinal factors (being pro-credit rather than anti-debt, or seeing credit as useful but prob lematic) were found to be important predictors of debt and debt repayments. further psychological factors, focusing on economic attributions, locus of control, coping strategies and consumer pleasure were found to be important, and a range of specific economic practices were also related to experiences of debt. journul ofeconomic psychology, 1992, 13: 111-134. (reprinted with permission of north-holland publishing company). international financial investment issues deviations from purchasing power parity and capital flows, by r. uppal. this article examines the implications of deviations from purchasing power parity for an investor’s portfolio decision and the consequent capital flows in a two-country, intertemporal model with complete financial markets. in the presence of deviations from purchasing power parity, investors from different nations hold disparate portfolios and the equilibrium that results is not a pooling one. the author solves explicitly for asset demands and derives the relationship between the direc tion of capital flows and deviations from purchasing power parity. in contrast to the small country assumption in the existing literature, the world interest rate in this model is determined endogenously. journal of international money and finance, april 1992,l l(2): 126-144. (reprinted with permission of the journal ofeconomic literature.) portfolio preference uncertainty and gains from policy coordination, by p. r. masson. international policy coordination is generally considered to be made less likely-and less profitable-by uncertainty about how the economy works. this paper offers a counter example in which investors’ increased uncertainty about portfolio preference makes coordination more beneficial. without such coordina tion, monetary authorities may respond to financial market uncertainty by not fully accommodating demands for increased liquidity for fear of inducing exchange rate depreciation. coordinated monetary expansion would minimize this danger. this result is formalized in a model incorporating an equity market; then, the stock market crash of october 1987 and its implications for monetary policy coordination are discussed. international monetary fund stafs papers, march 1992, 39(l): 101-120. (reprinted with permission of the journal of economic literature.) 162 ftnancialservicesreview,2(2) 1993 determinants of international financial services, by fariborz moshirian (the university of new south wales, sydney, australia). this paper analyzes the determinants of international financial services. it argues that the main difficulty in obtaining data about international financial services is the consequence of the current treatment of financial services in the system of national accounts (sna) of the united nations which treats the imputed service charge as ‘intermediate consumption of industries.’ the empirical results of import and export demands for financial services indicate that disposable income, domestic and foreign price of financial services are the important determinants of international financial services flows. journaz of bunking and finance, february 1993, 17( 1): 7-18. (reprinted with permission of north-holland publishing com pany*) optimal portfolio selection without short sales under the full information covariance structure: a pedagogic consideration, by clarence c. y. kwan and yufei yuan (mcmaster university, hamilton, canada). this study considers a portfolio optimization problem without short sales under the full-information covariance structure of security returns from a pedagogic perspective. by applying markowitz’s (1956, 1959,1987) critical line algorithm to lintner (1965) tangency portfolios, it offers a solution method which is algebraically simpler than markowitz’s original formulation. it also illustrates that the use of spreadsheets is a convenient alternative to traditional forms of computer program ming in the implementation of a small-scale analysis on microcomputers. doing so makes the markowitz analysis more accessible to finance students and practitioners in portfolio management. journul of economics and business, february 1993, 45( 1): 91-98. (reprinted with permission of the journal of economics and busi ness.) stock prices and bond yields: can their comovements be explained in terms of present value models?, by robert j. shiller (yale univer sity) and andrea e. beltratti (university of turin, italy). real stock prices do not show the relation to long-term interest rates that a simple rational expectations present value model would imply. real stock prices drop when long-term interest rates rise (and rise when they fall) more than would be implied by this vector autoregression model. in contrast, over the last century changes in real stock prices have shown little correlation with changes in inflation abstracts of articles on individual financial mangement 163 rates, and according to the present value model they should show little correlation. these conclusions were reached from an analysis of annual data in the united states, 187 1-1989, and the united kingdom, 19 18-1989. journal ofmonetary economics, october 1992,30( 1): 2546. (reprinted with permission of north-holland publish ing company.) macroeconometrics of stock price fluctuations, by dewan a. abdullah and steven c. hayworth (eastern michigan university). this paper employs granger causality tests and sims’ innovation accounting to explain fluctuations in monthly stock returns within a vector autoregressive framework. the results show that past money growth, budget deficits, inflation, and both short-and long-term interest rates are granger causal prior to stock returns. these variables also explain a substantial proportion of the forecast error variance of stock returns. it is found that stock returns are related positively to inflation and money growth and negatively to budget deficits, trade deficits, and both shortand long-term interest rates, as economic theory would predict. quarterly journd of business and economics, winter 1993, 32( 1): 50-67. (reprinted with permission of quarterly journal of business and economics.) did regulatory actions discourage consumer demand for treas ury bills?, by harold a. black (university of tennessee) and robert l. schweitzer (university of delaware). this paper studies the demand for treasury bills by consumers during each period of significant regulatory change intended to discourage such demand. the results show that there were significant changes in the structure of the public’s demand, although the changes were not always in the direction intended by the regulatory authorities. journal of banking and finance, february 1993, 17(l): 19-26. (reprinted with permission of north-holland publishing company.) initial public offerings of equity securities: anomalous evidence using reits, by ko wang and su han chan (california state univer sity, fullerton) and george w. gau (university of texas). in contrast with numerous studies that find significant underpricing for initial public offerings of industrial firms, we document a statistically significant average return of -2.82% on the first trading day for a sample of 87 initial public offerings of real estate investment trusts during the 1971-1988 period. our overpricing result is invariant to offer price, issue size, distribution method, offer period, and under writer reputation. newly issued reits, on average, substantially underperform a matching sample of seasoned reits during the first 190 trading days. interestingly, 164 financial services review, 2(2) 1993 buyers of overpriced reits are predominantly individual or non-13(f) institutional investors. journal of financial economics, june 1992,31(3): 381-410. (reprinted with permission of north-holland publishing company.) introducing risky housing and endogenous tenure choice into a portfolio-based general equilibrium model, by patric h. hender shott and yunhi won (ohio state university). portfolio-based general equilibrium models are useful for analyzing the inter action between the structure of individual tax rates and the way particular assets are taxed, for considering the role of differential tax rules and risk in determining household portfolio choices, and for addressing distributional questions. unfortu nately, current versions of these models give housing short shrift; owner housing is assumed to be riskless, rental housing is not a separately identifiable asset, and tenure choice is of necessity exogenously determined. this paper extends one of these models to incorporate a full housing subsector and uses the model to analyze the impact of the u.s. 1986 tax reform act. journal ofpublic economics, august 1992,48(3): 293-3 16. (reprinted with permission of north-holland publishing company.) portfolio rebalancing and the effective taxation of dividends and capital gains following the tax reform act of 1986, by mark fedenia (university of wisconsin-madison) and theoharry gramma tikos (european investment bank, luxembourg). the tax reform act (ira) of 1986, among other things, equalized the nominal taxation of dividend and capital gains income. the evidence is consistent with the hypothesis that ordinary investors rebalanced their portfolio holdings in response to the new taxes. such rebalancing appears to have concentrated on liquid stocks indicating that transaction costs are an important consideration for ordinary investors involved in restructuring their portfolio holdings. the tax clientele rela tionships that existed prior to the new act were distorted temporarily but have reappeared by 1988. interestingly, investors’ post-reform effective tax rates still give a preference to capital gains as they did prior to the tax reform. journal of banking and finance, june 1991, 15(3): 501-519. (reprinted with permission of north-holland publishing company.) towards an equilibrium model of the mutual funds industry, by jean dermine, damien neven (insead fontainebleau, france), and jacques f. thisse (core, louvain-la-neuve, belgium). we consider an industry in which mutual funds can form portfolios at lower cost than individual investors. investors can gather their own portfolio from primary securities and/or shares of mutual funds. in this context, we model competition between mutual funds as a non-cooperative game in which funds select their abstiacts of articles on individual financial mangement 165 portfolios. we show that a small number of funds suffices to ensure a pareto superior equilibrium. journal of bunking and finance, june 1991, 15(3): 485499. (re printed with permission of north-holland publishing company.) bayesian and capm estimators of the means: implications for portfolio selection, by philippe jot-ion (columbia university). this paper compares active investment policies under three alternative models for estimating expected stock returns: the historical sample mean, a shrinkage or bayesian estimator and a capm-based estimator. the out-of-sample performance of actively managed u.s. industry portfolios is analyzed for these three estimators over the period 1931 to 1987. it is found that the classical method, based on historical means and covari antes, leads to the worst forecasts and out-of-sample performance, and is generally outperformed by shrinkage estimators. an active portfolio with expected returns based on the capm produced the best results among all actively managed portfo lios; this strategy, as we show, closely matches a simple buy-and-hold the market rule. journal ofbanking and finance, june 1991,1.5(3): 717-727. (reprinted with permission of north-holland publishing company.) the dividend-clientele controversy and the tax reform act of 1986, by douglas hearth and james n. rimbey (university of arkan sas). several recent studies employ the tax reform act of 1986 (tra) to reexam ine the dividend tax clientele hypothesis. because tra effectively equalized the rate of taxation between dividends and capital gains, its implementation should provide a platform for resolving this controversy. instead, reported results remain sharply divergent. this study attempts to shed new light on the dividend-clientele question through use of a method that carefully matches preand post-tra dividend activity and avoids problems associated with the distribution of data. no significant or systematic changes in the ex-dividend price behavior of common stocks as a result of the tax reform act of 1986 are reported. quarterly journal of business and economics, winter 1993, 32(l): 68-81. (reprinted with permission of quarterly journul of business and economics.) why don’t individuals speculate in the forward foreign exchange market? by c. a. e. goodhart and m. p. taylor. it is shown, using institutional evidence, economic theory, and empirical evidence, that, given reasonable estimates of individuals’ coefficients of relative risk aversion, the combination of riskiness, minimal size of contract, and transac tions costs will deter all but the wealthiest individuals from seeking to speculate in the forward foreign exchange market. scottish journal of political economy, 166 financial services review, 2(2) 1993 february 1992,39( 1): 1-13. (reprinted with permission of the joumal ofeconomic literature.) mutual fund performance hot hands in mutual funds: short-run persistence of relative performance, 1974-1988, by darryl1 hendricks, jayendu patel, and richard zeckhauser (harvard university). the relative performance of no-load, growth-oriented mutual funds persists in the near term, with the strongest evidence for a one-year evaluation horizon. portfolios of recent poor performers do significantly worse than standard bench marks; those of recent top performers do better, though not significantly so. the difference in risk-adjusted performance between the top and bottom octile portfolios is six to eight percent per year. these results are not attributable to known anomalies or survivorship bias. investigations with a different (previously used) data set and with some post-1988 data confirm the find of persistence. the journal offinance, march 1993,48( 1): 93-130. (reprinted with permission of the journal offinunce.) real estate investment reducing taxes on the disposition of a personal residence with acreage, by michael m. megaard (lt.&ins and annis, spokane, wa) and susan l. megaard (eastern washington university). after the rules for deferring gain under section 1034 and excluding gain under section 121 on dispositions of a personal residence are explained, the focus is on how to use these sections in combination to minimize taxes where a home is sold with acreage in one transaction or in separate parcels. the journal of real estate taxation, spring 1993,20(3): 269-284. (reprinted with permission of the journul of real estate taxation.) planning for the “temporary rental” of a principal residence, by cynthia e. bird (california state university, san bernardino). the tax consequences of a rental of a personal residence are discussed and suggestions are made with regard to structuring the rental to maximize the tax benefits during the rental period and preserve the opportunity for deferral of gain on disposition. the journal of real estate taxation, winter, 1993,20(2): 135-148. (reprinted with permission of the journal of real estate taxation.) abstracts of articles on individual financial mangement 167 an empirical analysis of housing price appreciation in a market stratified by size and value of the housing stock, by j. allen seward, charles j. delaney (baylor university) and marc t. smith (university of florida). this paper examines the relationship between house size and appreciation, and house value and appreciation evidenced in the date collected from a single geo graphic region. the analysis of the data suggests that high price housing appreciates at a more rapid rate than low and medium price housing during expansionary periods, and there is no statistical difference in the rates of price change in contractionary periods. further, the rate of price change for larger homes exhibits no consistent difference from the rate of change for small and medium size housing for the time period studied. the journal of real estate research, spring 1992,7(2): 195-205. (reprinted with permission of the journal of real estate research.) taxation is the child care credit progressive? by a. dunbar and s. nordhauser. the child care credit has been widely perceived as regressive because critics claim that more credit is claimed by high-income taxpayers than low-income taxpayers. the credit was changed in 1981 in an effort to make it more progressive. nevertheless, the perception that the credit is regressive persists. this paper applies three measures of tax progressivity to a sample of taxpayer data to determine whether the child care credit was regressive during 1979-86 and whether changes made in 1981 made it more progressive. the conclusion is that the credit was progressive over the entire period, becoming more progressive after 198 1. national tax journal, part 2, december 1991,44(4): 519-528. (reprinted with permission of the national tax journal.) capital gains tax and equity values: empirical test of stock price reaction to the introduction and reduction of capital gains tax exemption, by ben amoako-adu (wilfrid laurier university, water loo, canada), m. rashid (university of new brunswick, canada) and m. stebbins (mount saint vincent university, halifax, canada). this paper evaluates the differential effect on stock prices of the introduction in canada of $500,000 capital gains tax exemption and the reduction of this limit to $100,000 two years later. using the seemingly unrelated regression methodology and controlling for thin-trading and heteroscedasticity, the empirical evidence indicates that the changes in capital gains tax laws has a differential effect on 168 financial services review, 2(2) 1993 low-dividend yield and high-dividend yield stocks on both occasions. while the evidence indicates that the stock market anticipated the 1985 capital gains tax law changes, the significant market reaction to the 1987 reduction of the exemption occurred a day before and after the reading of the tax reform proposals in parliament. thus, it can be inferred from the results that, despite the presence of tax-sheltering opportunities in canada, changes in capital gains tax laws affect equity values. journal of banking and finance, april 1992, 16(2): 275-287. (reprinted with permission of north-holland publishing company). academy of financial services officers president swarn chatterjee university of georgia president-elect janine scott shepherd university executive vice president-program inga chira california state university, northridge vice president-communications david nanigian california state university, fullerton vice president-finance thomas langdon roger williams university vice president-international relations philip gibson winthrop university vice president-professional organizations frank laatsch university of southern mississippi vice president-mktg & public relations chris browning texas tech university vice president-membership sherman hanna ohio state university vp local arrangements 2016 swarn chatterjee university of georgia immediate past president thomas coe quinnipiac university editor, financial services review stuart michelson stetson university directors shawn brayman planplus global charles chaffin cfp board of standards victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university terrance martin university of texas–rio grande valley past presidents robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial.org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2018 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 27, no. 3, 2018 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama dale domian york university jim gilkeson university of central florida jason greene georgia state university william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi david lange auburn university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international bill blair macquarie university s. j. chang illinois state university lawrence rose massey university sharon taylor university of western sydney education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2018 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. v31 i4 ce quiz ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires march 31, 2024, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce onlineonly—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. terry, a certified financial planner professional, has each prospective client complete a personality assessment as an element of the client data collection process. based on the assessment, one of terry’s new clients was found to be “under controlled.” given the client’s profile, terry should a. encourage the client to relax their conscientiousness when contemplating the implementation on financial planning recommendations. b. help the client calm their risk taking tendency. c. nudge the client to be a bit more extraverted when thinking about taking portfolio risk. d. consider referring the client to a psychiatrist before moving forward in the financial planning process. 2. the typical investor perceives risk as a. an opportunity. b. the opposite of return. c. the possibility of financial loss. d. a requirement to meet financial goals. 3. when researchers talk about objective financial knowledge they are referring to a test score or a. a person’s level of financial literacy. b. an individual’s perception of their financial solvency. c. a person’s financial confidence. d. someone’s degree of financial satisfaction. 4. who is more likely to report being financially satisfied? a. those who keep their finances totally separate from their partner. b. those who combine their finances with their partner. c. those who report that their household is managed using traditional gender-role arrangements. d. those who allow one partner to make all household investment decisions. 5. of the following factors, which is most important in explaining capital market participation? a. the ownership of a personal residence. b. the age of an investor. c. net household income. d. an investor’s degree of risk aversion. the impact of superannuation fund choice legislation and the global financial crisis on australian retail fund flows rakesh guptaa,*, thadavillil jithendranathanb agriffith business school, griffith university, 170 kessels road, nathan 4111, australia bfinance department, university of st. thomas, 2115 summit avenue, saint paul, mn 55105, usa abstract we examine the extent to which cash flows into the australian superannuation funds are affected by the past performance of the fund, the riskiness of the fund, the choice of superannuation fund legislation, and the global financial crisis. both retail and wholesale investors base their investment decisions on the past performance of the funds. there is little evidence that the riskiness of the fund returns has any significance effect on the flow of funds. legislation has resulted in more inflows into managed funds. there is more inflow into managed funds and equity funds since the period of the global financial crisis. © 2015 academy of financial services. all rights reserved. jel classification: g1; g2 keywords: australia; managed funds; returns chasing 1. introduction the australian managed funds industry is one of the largest in the world. a main driver of the growth of this sector of the australian economy is the compulsory superannuation retirement scheme. in the quarter ending march 2013, the total assets managed by the industry were au$2,094 billion, out of which au$1,526 billion was comprised of superannuation funds.1 the superannuation scheme was introduced in the 1980s and today it covers more than 90% of the australian workforce. currently workers invest 9.5% of their annual salary into a superannuation fund, which will progressively increase to 12% by * corresponding author. tel.: �61 7 3735 7593; fax: �61 7 3735 7760. e-mail address: r.gupta@griffith.edu.au (r. gupta) financial services review 24 (2015) 217–248 1057-0810/15/$ – see front matter © 2015 academy of financial services. all rights reserved. 2019–2020. the objective of superannuation is to provide self-funded retirement for the public, thereby to move away from reliance on publicly funded retirement. superannuation funds are broadly divided into four categories. first, corporate funds that are set up for the employees of a specific corporation, and only employees of a particular corporation, are members of this type of fund. second, industry funds cater to those employees in a specific industry, for example, the retail employees superannuation trust (rest superannuation) specifically caters to employees within the retail industry. third, there are public sector employee funds such as the public sector superannuation scheme (that existed before june 30, 2005) and the public sector superannuation accumulation plan for new members that began operation after july 1, 2005. finally, there are retail funds for all other investors that are typically developed by financial institutions and insurance companies. the retail fund sector is the largest of all australian managed funds, with assets over au$1,129 billion as of march 2013. these funds are set up by banks, insurance companies and investment firms. this sector can be further subdivided into wholesale and retail segments. wholesale funds, which include other funds, cater to large investors. retail funds cover a variety of investments, including superannuation funds, retirement investment funds, and discretionary investment funds. at the end of march, 2013,2 the total investment in the wholesale segment was au$529 billion and in the retail segment was au$574 billion. a significant change in the superannuation investment environment was brought about by the choice of superannuation fund legislation of 2005 (known as choice legislation), which gave employees more freedom to move their superannuation savings from one fund to another. the other significant economic event was the global financial crisis (gfc) of 2008–2009. in this article we look at the impact of these two events on the net fund flows into the retail funds sector. during the initial days of the superannuation scheme, the investment choices available to employees and their ability to switch from one fund to another was limited. choice of investment class and the freedom to switch between asset classes were available to members of the superannuation funds at varying levels (most funds allowed one or two switches in a year free of cost and charged fees for any additional switch). to increase flexibility for members, the government introduced choice of superannuation fund legislation which allowed members to choose their superannuation fund provider, with effect from july 1, 2005. we extend the study by gupta and jithendranathan (2012) by testing whether there was a significant change in asset allocation across asset classes after the introduction of choice legislation.3 studies into the impact of the introduction of choice legislation on australian managed funds are scant. a recent study by gupta and jithendranathan (2012) did not look at the impact of choice legislation, whereas the study by gerrans (2012) looks at the relationship between different type of members and their switching behavior in a sample of superannuation funds. our study aims to look at the changes at aggregate levels and see the impact of legislation on the asset allocation behavior of members after the introduction of the legislation. the findings of the study will be of importance for investors and fund managers who seek to maximize their risk-adjusted returns. the results indicate that investors use past returns in making investment choices and this can lead to risky investments with the potential 218 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 for significant losses. an understanding of the asset allocation behavior of fund members will be of importance for policy makers, as the primary objective of having superannuation savings is to move people away from publicly funded retirement to self-funded retirement. a shortfall in the superannuation fund accounts of individuals will mean that the government will have a responsibility to provide for the retirement of these people. before the choice legislation, members were restricted to one superannuation fund4 that they joined at the commencement of their employment, with the default option being a balanced fund. however, members could choose to make a proportional allocation into available asset classes within the fund. they were also allowed to make periodical reallocations of the asset classes. after the introduction of the choice legislation, members had the freedom to allocate their balance across funds. the government’s rhetoric explaining the introduction of the legislation was that it now provided choices to individuals in terms of their investment. there was a significant amount of media attention and because of this, individuals may have become more aware of their ability to make investment choices. we hypothesize that if this contributed to investors making more conscious decisions in their asset allocation, there will be a move away from balanced funds to other investment classes. our research found a significant shift away from balanced funds after the introduction of the legislation. another significant economic event that occurred during the study period was the gfc that originated in the united states in 2008–2009 and quickly spread to other markets around the world. throughout the world, stock markets declined and investors suffered substantial losses. it is the common perception that during financial crises investors may move away from risky assets such as stocks, into safer assets like fixed income and cash. we also study the impact of the gfc in fund members’ asset allocation preferences. we find that the superannuation fund members moved away from risky assets to less risky assets in the period immediately after the crisis. retail funds are a suitable class of managed funds to investigate to study the effect of choice legislation and the global financial crisis. we have chosen the retail managed funds for the study for two reasons. first, it is the largest managed funds category and second, the retail fund provides more freedom to its members compared with other categories. membership of the corporate funds, public sector funds and industry funds are restricted because of the way in which these funds are created. currently there are more than 14,0005 active retail individual funds offering a multitude of asset classes to superannuation, retirement income, discretionary and wholesale investors. superannuation investors are required by law to invest part of their salaries and hence their choice of funds will be based on the perceived return and risk they are taking. discretionary6 investors may have less money to invest during financial crises and may choose to move their money away from risky assets to less risky ones. decisions by wholesale investors may be partially driven by the contributions of their constituents as well as by decisions taken by fund managers. investors base their investment decisions on their perceptions of expected returns and risk of each asset category, and in turn, this will be reflected in the net funds flow into each asset class. this study will use a panel regression model to analyze the relationships between the funds flows with past return and risk of the funds, the impact of the choice legislation, and the global financial crisis. the results of our analysis show a significant difference in the allocation of assets between investors of retail and wholesale funds. retail investors have a 219r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 preference for less risky investments compared with the wholesale investors, and are biased in favor of domestic investments. a similar study by gupta and jithendranathan (2012) found a significant relationship between fund flow and past performance that could not be explained by risk. this current study overcomes the limitations of the previous study, which did not incorporate the impact of choice legislation because of limited data. the other weakness of the previous study was in ignoring the impact of the gfc in its analysis. this study overcomes both of these limitations. it is a common perception that during periods of crises investors tend to move away from more risky assets to less risky assets. this study, by incorporating the impact of the introduction of the choice legislation and the gfc, will provide a better understanding of the risk perceptions of australian investors. the findings of the study will contribute to theory by shedding light on the risk tolerance of australian superannuation fund members. because the primary objective of the superannuation fund is to provide a self-funded retirement for superannuation fund members, these findings have important implications for policy makers who seek to develop policies to achieve the objective of a self-funded retirement for most people. if fund members make poor investment choices, they may not have accumulated enough wealth to provide a sufficient retirement income. our findings also have implications for investors who seek to maximize their net wealth for retirement. the rest of the article is organized as follows. section 2 covers the literature review, section 3 describes the data, and section 4 outlines the methodology used in this article. section 5 analyzes the results and section 6 concludes the article. 2. literature review investors commonly use the past performance of the managed fund (or superannuation funds) in making their investment decisions. as such, superannuation funds with superior past performance will attract an inflow of funds while funds with poor performance may experience outflows. advertisers frequently use the past performance of the funds to attract inflows, suggesting that past performance will continue in the future. several studies examine the relationship between the fund flows and the performance of the u.s. mutual fund industry. the relationship between the performance and fund flows was studied by ippolito (1992) for the u.s. mutual fund industry for the period of 1965 to 1984. this study found a clear underlying movement of investments from recent poor performing managed funds toward recent good performers. sirri and tufano (1998) studied the inflows to the good performing funds and outflows from the poorly performing funds and found that consumers based their investment decisions on information about the prior performance of the funds. however, investors do this asymmetrically by investing significantly more in funds that performed well during the prior period(s). goetzmann and peles (1996) found a similar, significant relationship between past performance and flows. lynch and musto (2003) provided a theoretical framework to analyze this asymmetry between the flows of funds among funds. intuitively one would assume that market and management would address poor performance by either changing the 220 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 advisors and/or strategies that result in poor performance. this intuition helps to explain the asymmetry between the flow of funds among poor and superior performing funds. for investors, an understanding of the relationship between flow and performance is important from the investment decision standpoint, and this has been researched in number of empirical studies. for example, sawicki (2001) studied the flow-performance relationship of wholesale balanced pooled australian superannuation funds and found a positive, statistically significant relationship between recent performance and flow of funds, but did not find the convexity observed in the u.s. markets. drew, stanford, and veeraraghavan (2002) found that the raw risk adjusted returns showed mean reversions, and that previous period performance was not associated with future performance. bilson, frino, and heaney (2005) tested performance persistence using a sample of managed growth and managed stable australian retail funds. the performance persistence was tested using five different matrices and it was found that the inadequate adjustment of risk may cause spurious persistence in excess fund returns. frino, heaney, and service (2005), using a sample of 398 australian managed-growth and managed-stable funds, found a positive relationship between current net cash flows and past performance. the asset allocation strategies within the australian equities, fixed interest and listed property class of funds was tested by benson, gallagher, and teodorowski (2007) who found evidence of momentum investing by fund managers. gharghori, sujoto, and veeraraghavan (2008) found little evidence that australian investors are able to identify high performing superannuation funds. a similar study by del guercio and tkac (2002) compared the difference in cash flows between the u.s. mutual funds and pension funds. the results of the study indicated that there is significant difference between the behaviors of the two groups of investors. pension fund investors tend to move money away from poorly performing funds but they do not reward better performing funds by disproportionately investing into past years good performers. they also use risk measures, such as jensen’s � and tracking error to evaluate the performance of the funds. evidence from thenon-retirement mutual funds is different. these investors tend to move into newly acclaimed high performers with no consideration of risk adjusted performance measures. agnew and balduzzi (2010) use daily net aggregate fund transfers of the voluntary retirement contribution funds (401k plans) in the united states and find that, in response to market movements, these investors shift funds between equities, cash, and bonds. more recent studies find that defined contribution plans are more sensitive to past performance as compared with the rest of the mutual funds (sialm, starks, and zhang, 2012). investors (or members) of managed funds may view ratings by agencies such as morningstar as important sources of information, base their investments decisions on the ratings and move in and out of different funds based on announcements of rating changes. this effect was studied by del guercio and tkac (2008). they use a sample of morningstar’s rating changes from 1996 to 1999, and find that the changes in ratings have an impact on investment allocation decisions by retail mutual fund investors. gerrans (2004) reported that three quarter of investors base their investment decisions on rating changes. faff, parwada, and poh (2007) find similar evidence of australian retail investors moving money away from recently downgraded funds to recently upgraded funds. watson, wickramanayke, and 221r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 premachandra’s (2012) evaluation of australian equity funds using efficient portfolio models finds that efficient funds are likely to receive an upgrade in the medium to long term.7 behavioral finance literature argues that investment decisions may be influenced by behavioral biases in the decision making process. researchers have looked at different aspects of human behavior, for example, prospect theory argues that investors look at perceived gains and losses rather than relying on perceived outcomes. various studies have investigated the overconfidence hypothesis, wherein it is argued that investors who are overconfident overrate signal precision and overreact to these signals and miss-price economic factors (daniel, hirshleifer, and subrahanmanyam, 2001). bailey, kumar, and ng (2011) study investors from the u.s. discount brokerage firms to gain an understanding of the disposition and narrow framing effects. using proxies for the two behavioral biases— disposition effect and narrow framing—they find that investors may choose investments based on individual factors rather than based on the impact on the individual’s overall portfolio. gender bias is an important issue in behavioral finance literature and speelman, clark-murphy, and gerrans (2007) investigate this in the different groups of australian retirement funds. the results show that female investors are more risk averse than male investors, and young female investors exhibit the highest level of risk aversion. as investors age, there is an indication that they become return-chasers. a survey of australian investors by fry, heaney, and mckeown (2007) found that only a few investors showed interest in changing their superannuation; thus, supporting the behavioral theory of investor inertia. phillips (2011) looked at the relative risk aversion coefficient that characterizes representative self-managed superannuation fund investors and finds that these investors may be too risk averse to maximize their expected growth rate of wealth share accumulation. the effect of choice legislation on investor behavior was studied by fear and pace (2009). the results of this study indicated that superannuation investors are mostly apathetic about their pension savings, and the study did not find any strong evidence of fund switching behavior after the introduction of the choice legislation. a possible reason for lack of fund switching can be attributed to the cost of switching. a study by sy (2011) found that the choice legislation had the opposite effect on superannuation investors, as the rate of superannuation members making an active choice about their investments fell after the introduction of the legislation, and more members chose the default strategy for asset allocation. a recent australian study by gerrans (2012) looks at the changes in the investment strategies after the gfc and found that members did not make any changes to their investment strategies in the period following the crisis. the findings of these studies are contrary to common perception that investors when given greater choice ought to exercise these choices to enhance their risk adjusted returns for maximizing net wealth for retirement. a series of crises results in investors exercising more caution, moving funds into less risky investments. the results of the studies by sy (2011) and gerrans (2012) may have been limited by data.8 there are a few studies that have investigated investor behavior during the recent global financial crisis. gerrans (2012) studied the behavior of australian retirement savings investors during the crisis using a sample of 3.6 million investors from five superannuation funds, with assets totaling over au$74 billion. the results indicated that an overwhelming number of investors did not change their investment strategy in response to the crisis. gerrnas, faff, and hartnett (2012) tested the individual financial risk tolerance during the crisis using a risk 222 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 tolerance survey. the results indicated that risk-taking behavior was impacted by the crisis, but the researchers could not conclude whether or not it had any significant effect on the actual asset allocations. related studies of australian managed funds are as follows; holmes and faff (2007) who look at the style drift and fund performance over time. stout (2008) examines the relationship between withdrawals and portfolio asset allocations and brooks and porter (2012) examine the performance of mutual funds for the period of 1994 to 2005 using an attribution model. taking a global perspective, guercio and reuter (2014) examine the performance of the u.s. mutual funds for the period 1992 to 2004, and fang, kemf, and trapp (2014) examine the manner in which fund managers are allocated within the fund family, based on the market efficiency of each asset class. if investors rely purely on the past performance of funds in reallocating their investment portfolios they may be misallocating their investments. our literature survey found that the issue of impact of choice legislation and the gfc has not been sufficiently studied in relation to australian superannuation funds. these two events have been significant in their impact on the australian investment environment. gaining an understanding the impact these events have had may help investors and policy makers in more appropriately managing the future course of the superannuation environment. this research aims to fill this gap by answering three inter-related questions. the first question is, is there a significant relationship between past performance of the asset classes and the future allocation into asset classes that is not explained by risk of the individual asset class? the second question we attempt to answer is, is there a significantly increased reallocation of assets after the introduction of the choice legislation?9 finally, we test the impact of the gfc on investors’ reallocation of assets. 3. data data for this study was obtained from plan for life (qds retail and wholesale platform), plan for life is a firm of actuaries and researchers. the data covers the period from 1991 to 2013 and contains quarterly information about the funds under management, net cash flows, cash inflows, cash outflows, and investment earnings. according to the data vendor, the information is consistent with the investment and financial services definitions. because data predating september 1998 is sparse, and significant growth in the managed funds has occurred since that date, this current study uses fund data for the period from september 1998 to march 2013 only. the number of funds included in the study is 27,965, out of which 14,188 were active as of march 2013, whereas 5,873 were terminated and 7,904 were transferred to other funds. there were more than 150 fund families that offered these products. a summary of assets under management for different types of funds (superannuation and retirement), non-superannuation and wholesale funds is given in table 1. it shows a growth of 20% innon-superannuation, 23% in superannuation and retirement funds, and 24% in the wholesale category. the asset allocation trends among various types of retail funds are given in figs. 1 through 4. fig. 1 shows asset allocation trends among wholesale funds. in 1998, the largest asset category was managed balanced funds (20.03%), which reduced to a mere 3.03% by march 2013. the asset category that had the highest growth rate during this period is overseas equity 223r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 investments, which went up from 11.12% in 1998 to 20.45% in 2013. cash as percentage of total assets remained fairly stable, between 10.97% in 1998 to 9.73% in 2013. australian equities started at 13.97% in 1998, went up to 25.87% in june 2005 and came down to 18.18% in 2013. the other asset category that had significant change is mixed portfolios, which went up from less than 1% in 1998 to 24.7% in 2013. part of the reason for the increase in this category may be because of reclassification of some of the other categories to a mixed category. asset allocation trends in retail funds are given in fig. 2. in 1998, the single largest category of retail funds was capital guaranteed (18.31%), which dropped to only 4.6% in 2013. as is the case with wholesale funds, the category with the highest growth was mixed portfolios, which increased from 9.13% in 1998 to 41.22% in 2013. unlike wholesale funds, overseas equity is not a significant part of the allocation of the retail funds. in 1998, overseas equity was 4.14% of total assets in this category, and it was little changed at 3.68% in 2013. the share of australian equity also remained fairly stable; 8.87% in 1998 and 9.5% in 2013. the share of cash saw a decrease of more than 3% between 1998 and 2013. within retail funds the asset trends in allocation of superannuation funds is given in fig. 3. in 1998, the single largest class of asset was capital guaranteed (24.06%), followed by table 1 total investments in retail and wholesale products (in million australian dollars) category average funds under management funds under management on september 1998 funds under management on march 2013 annual growth rate superannuation and retirement 255,882.39 97,526.86 429,802.34 23.62% non-superannuation 140,988.45 66,707.47 144,666.00 20.65% wholesale 224,569.49 58,514.45 529,469.32 24.32% total 645,986.87 222,748.78 1,103,937.66 25.69% fig. 1. asset allocation trends in wholesale funds. 224 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 managed balanced (19.51%) and managed growth (19.19%). in 2013, the share of capital guaranteed fell to 4.87%, while the share of managed balanced (14.32%) and managed growth (15.65%) fell slightly. the category with the highest growth rate is mixed portfolios, which increased from 11.25% in 1998 to 37.95% in 2013. the share of australian equity also saw significant increase from 3.9% in 1998 to 7.92% in 2013. the overall trends in asset allocation indicate that retirement investors are fairly stable with their asset allocation choices. asset allocation trends innon-superannuation retail funds (discretionary investors) are given in fig. 4. compared with other categories of investors, discretionary investors have more investment in cash. in 1998 their allocation in cash was 21.48%, which came down to fig. 2. asset allocation trends in retail funds. fig. 3. asset allocation trends in superannuation funds. 225r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 10.2% in 2013. as in the case of superannuation and retirement investment funds, the highest growth is in mixed portfolios, which increased from 6.14% in 1998 to 50.92% in 2013. unlike superannuation fund investors, discretionary investors have less money in managed funds and more investments in australian equities. the share of asset allocation into australian equities for this category of investors remained fairly stable. one interesting aspect of asset allocation of discretionary investors is the flight to safety during the global financial crisis. the share of australian equity started at 15.85% in 1998 and fell to its lowest in december 2008 (11.11%) and then increased to 21.4% by march 2013. 4. empirical methodology superannuation fund-level rates of return (ror) are calculated as follows: rori,t � net earnings after taxi,t sizei,t�1 � 1 2 nfi,t (1) where, sizei,t�1 is the funds under management for the ith fund for the quarter t-1 and nfi,t is the net fund flows for the ith fund for the quarter t. to calculate the average returns for each category of investment, a value weighted index of individual fund returns is created. the average returns and standard deviations of these individual indices are given in table 2. among various investment categories, australian small company equity funds had the highest quarterly returns of 3.18% followed by australian equity funds (2.22%) and asia pacific equity funds (2.17%). the mixed portfolio funds had the lowest average quarterly returns at 1%, followed by cash with 1.01% average quarterly returns. managed funds also are at the lower end of returns with a range of 1.07% to 1.20%. fig. 4. assets allocation trends innon-superannuation retail funds. 226 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 risk, as measured by the standard deviation of the quarterly returns, are the highest at 9.6% for overseas property funds, followed by australian small company equity funds (8.54%) and australian property security funds (8.47%). managed funds and mixed portfolio funds in general had low return standard deviations in the range of 1.76% to 4.19%. the lowest return standard deviations were observed for cash (0.16%) and mortgage funds (0.25%). for regression analysis purposes, a panel data are created for each individual investment category. to avoid the survivorship bias, funds that were terminated or transferred are included in the dataset, provided each fund had at least 24 continuous quarterly data points, of which at least 12 quarters occurred post june 2005 to incorporate the effects of choice of superannuation fund legislation and the global financial crisis. one of the issues that arose in the analysis was the sudden increase or decrease in net flows during a quarter. to eliminate the outliers in the net flows, the data for a quarter where the net flow is greater than 0.75 of the funds under management at the beginning of the quarter, is dropped. this eliminates sudden increases in cash flows that may be because of interfund transfers. a typical mutual fund investment decision involves two steps—first the choice of an asset category and then the selection of a fund within that asset category. the choice of asset category depends on the investors’ risk preferences, and the choice of the fund may depend on its past performance. thus, individual investors are concerned about the returns and risks involved in investing. there are various agencies which report the performance of the funds, and also individual fund managers advertise their performance measures. if investors follow the performance of funds carefully, they will tend to invest more into the funds that perform well in a particular category and to withdraw funds from the poorly performing funds. the most commonly used measure of fund performance is to compare the fund return with that of a benchmark. in the case of u.s. markets, sensoy (2009) pointed out that more than one third of the benchmarks used in the performance analysis are incorrect indicators of the funds’ true style. to avoid this problem, we use a value weighted index of fund returns for table 2 summary statistics of quarterly returns (september 1998–march 2013) investment category name mean standard deviation minimum maximum alternatives 1.20% 3.04% �7.77% 7.81% cash 1.01% 0.16% 0.70% 1.44% diversified fixed interest 1.01% 2.07% �8.37% 8.27% australian equity 2.22% 6.79% �15.06% 20.73% australian equity small companies 3.18% 8.54% �22.19% 22.81% australian fixed interest 1.19% 0.91% �0.71% 2.82% managed balanced 1.17% 3.78% �10.70% 10.05% managed growth 1.17% 4.19% �11.45% 11.32% mortgage 1.20% 0.25% 0.42% 1.60% managed stable 1.07% 1.76% �5.02% 5.95% overseas asia pacific 2.17% 9.96% �18.18% 32.29% overseas fixed interest and currency 1.34% 1.39% �2.21% 4.33% overseas global 0.50% 7.02% �14.71% 20.06% overseas property 1.13% 9.60% �36.22% 25.07% australian property 1.64% 2.27% �7.59% 6.64% australian property securities 1.49% 8.47% �30.06% 27.47% mixed portfolios 1.00% 3.89% �10.62% 11.42% 227r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 each of the fund categories. the quarterly return of each fund and the index are calculated using eq. (1) and the active returns are calculated as: ri,t � rori,t � rorindex,t (2) the next important factor is to find an appropriate measure for risk. the most commonly used measures of risk in empirical analysis are jensen’s �, the sharpe ratio, and tracking error. the australian bureau of statistics and the australian prudential regulation authority report the return of assets and the standard deviation of the returns as a measure of risk. the key questions are how many investors are aware of the investment style of the fund, and how do they evaluate the risk of the fund they are investing in? a survey of u.s. mutual fund investors by capon, fitzsimons, and prince (1996) showed that only 25% of mutual fund investors knew the investment style of the fund, and that only 26.7% of those surveyed compared the fund return with the benchmark. this survey also found that 14% of the respondents used standard deviation as the measure of risk and only 4% used either the � or the sharpe measure to identify the risk. in this study we use tracking error as the measure of risk.10 an important factor that can influence the performance of a fund is its size. chen, hong, huang, and kubik (2004) showed that the performance of the funds decreases with size, and therefore, to control for the size effect, we use the lagged log size of the fund as a control variable. to adjust for the momentum effect of net flows, the lagged value of the net flows is included as an independent variable. to control for the unusual flows because of transfers, and so forth, the following control variable is used. eint � nfi,t sizet�1 � marketnetflowst marketsizet�1 (3) in time series regressions it is important to check whether the series are stationary. regressing non-stationary time series can lead to spurious regressions. on the other hand if the time series variables that are non-stationary are co-integrated, then they have a long-term, or equilibrium relationship between them. in this study we test the long-term relationship between the variables using the pedroni (1999, 2004) heterogeneous panel cointegration test. this test involves regressing the variables along with cross-section specific intercepts, and examining whether the residuals are integrated in the order of one. the pedroni test calculates two sets of statistics: (1) panel co-integration test statistics and (2) group mean panel co-integration statistics. for the first test, four test statistics are calculated: panel v-statistic, panel �-statistic, panel pp-statistic, and panel adf-statistic. for the second test three statistics are calculated: group r-statistic, group pp-statistic, and group adf-statistic. because panel and group pp-and adf-statistics have the best small sample properties, we report these in the article.11 the results of co-integration tests are given in table 3. the results strongly reject the null-hypothesis that the variables are not co-integrated. because the variables are cointegrated, the long-term relationship between the variables is tested using the dynamic ordinary least square (dols) estimator proposed by mccoskey and kao (1999) and kao and chiang (2000). the dols is an expanded ordinary least square estimation including not 228 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 only the explanatory variables, but also the leads and lags of the first difference terms to control endogeneity and to calculate the standard deviations using covariance matrix of errors that is robust to serial correlation. the regression model used is: nfi,t � bjrji,t � c�i,t�1 � dsizei,t�1 � enfi,t�1 � fnfi,t�2 � geini,t � hcht � jgfct � � j��1 1 ki, j�xi,t � �t (4) where nfi,t are the net flow in to the fund, rji, is the trailing excess returns for either 1, 2, or 3 quarters, �i,t is the tracking error, sizei,t, is the log of the fund size, cht is the choice dummy, which has a value of 0, until the third quarter of 2005 and 1 for rest of the time period, gfct is the crisis dummy, which has a value of 0, until the third quarter of 2008 and 1 for the rest of the time period, eint is a variable to capture the overall flow into the market, and fei,t is the fixed effect of ith fund. we estimated three separate regressions with the variable rji,t being 1, 2, or 3 quarter excess returns. �xi are the first difference terms of the regression variables and we use one lead and lag in the model. the choice of three quarters excess return is to see if investors look at the most recent or past returns. we do not find significance beyond three quarters.12 5. results the regression results for various categories of funds are given in tables 4 through 7. table 4 gives the results for the total retail fund net flows for various categories of assets. if investors exhibit return chasing behavior, then the past excess returns should have a table 3 pedroni panel co-integration test for the regression variables investment category name panel-pp panel-adf group-pp group-adf alternatives �10.0981*** �8.3714*** �10.4713*** �7.0531*** cash �46.2852*** �33.6419*** �51.0968*** �34.9540*** diversified fixed interest �34.2109*** �24.7624*** �41.3084*** �28.7055*** australian equity �41.1697*** �26.1176*** �46.0305*** �28.6325*** australian equity small companies �16.0264*** �11.3203*** �19.1526*** �13.0535*** australian fixed interest �32.9618*** �24.6268*** �38.0533*** �27.9730*** managed balanced �33.6623*** �19.6978*** �38.7136*** �20.9555*** managed growth �49.0109*** �30.0976*** �61.1529*** �36.1550*** mortgage �9.1042*** �3.8491*** �12.2357*** �5.9000*** managed stable �33.0024*** �19.1801*** �39.9762*** �22.2885*** overseas asia pacific �4.8391*** �3.4357*** �4.4818*** �2.8588*** overseas fixed interest and currency �17.9625*** �14.3272*** �19.4624*** �15.8561*** overseas global �37.0328*** �25.3996*** �45.6936*** �33.1615*** overseas property �6.9439*** �6.4980*** �9.0963*** �8.4064*** australian property �21.8242*** �15.6943*** �23.5065*** �16.1208*** australian property securities �28.2284*** �19.4943*** �34.9844*** �23.0703*** mixed portfolios �25.6963*** �17.2480*** �32.1449*** �21.4459*** *** significant at 1%. 229r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 4 fi xe d ef fe ct s pa ne l re gr es si on of re ta il m an ag ed fu nd ne t flo w s w ith pa st ex ce ss re tu rn s an d ri sk r eg re ss io n eq ua tio n: n f i, t � b j r j t, t � c� i, t� 1 � ds iz e i ,t � 1 � en f i, t� 1 � fn f i, t� 2 � ge in i, t � hc h t � jg f c t � � j� � 1 1 k i ,j � x i ,t � � t w he re n f i, t ar e th e ne t flo w in to th e fu nd , r j i, is th e tr ai lin g ex ce ss re tu rn s fo r ei th er 1, 2, or 3 qu ar te rs , � i, t is th e tr ac ki ng er ro r, si ze i, t, is th e lo g of th e fu nd si ze , c h t is th e ch oi ce du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 05 an d 1 fo r re st of th e tim e pe ri od , g f c t is th e cr is is du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 08 an d 1 fo r th e re st of th e tim e pe ri od , e in t is a va ri ab le to ca pt ur e th e ov er al l flo w in to th e m ar ke t, an d f e i, t is th e fix ed ef fe ct of ith fu nd . � x i ar e th e fir st di ff er en ce te rm s of th e re gr es si on va ri ab le s. w e es tim at ed th re e se pa ra te re gr es si on s w ith th e va ri ab le r j i, t be in g 1, 2, or 3 qu ar te r ex ce ss re tu rn s. fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) a lte rn at iv es o ne qu ar te r � 3. 16 77 8. 86 91 � 0. 61 56 0. 29 72 0. 14 28 21 .7 41 6 � 2. 89 20 � 0. 41 26 0. 47 08 (� 0. 70 59 ) (1 .3 54 4) (� 1. 52 25 ) (4 .7 32 9) ** * (1 .9 55 4) * (4 .5 88 6) ** * (� 3. 04 26 )* ** (� 0. 65 80 ) (1 1. 46 50 )* ** t w o qu ar te rs � 3. 23 81 9. 18 17 � 0. 61 68 0. 29 65 0. 14 31 21 .8 81 3 � 2. 89 20 � 0. 43 29 0. 47 08 (� 0. 99 52 ) (1 .3 33 0) (� 1. 52 58 ) (4 .7 11 6) ** * (1 .9 59 1) * (4 .7 69 0) ** * (� 3. 04 95 )* ** (� 0. 70 23 ) (1 1. 46 62 )* ** t hr ee qu ar te rs � 2. 65 97 8. 23 83 � 0. 65 54 0. 28 75 0. 13 98 22 .3 91 0 � 2. 88 91 � 0. 48 09 0. 46 96 (� 0. 80 00 ) (1 .2 09 8) (� 1. 62 26 ) (4 .5 52 0) ** * (1 .9 11 6) * (4 .8 56 2) ** * (� 3. 04 11 )* ** (� 0. 77 73 ) (1 1. 41 23 )* ** c as h o ne qu ar te r � 5. 56 27 1. 55 11 � 0. 72 31 0. 15 11 � 0. 12 05 49 .5 78 3 � 0. 05 13 � 1. 05 85 0. 69 24 (� 0. 15 01 ) (0 .3 11 2) (� 2. 55 38 )* * (9 .2 26 7) ** * (� 6. 50 65 )* ** (1 4. 81 95 )* ** (� 0. 12 02 ) (� 2. 82 45 )* ** (6 4. 77 03 )* ** t w o qu ar te rs � 8. 90 39 1. 55 60 � 0. 72 08 0. 15 11 � 0. 12 05 49 .5 61 7 � 0. 05 13 � 1. 05 81 0. 69 24 (� 0. 24 11 ) (0 .3 12 2) (� 2. 54 02 )* * (9 .2 26 7) ** * (� 6. 50 62 )* ** (1 4. 82 25 )* ** (� 0. 12 02 ) (� 2. 82 35 )* ** (6 4. 77 00 )* ** t hr ee qu ar te rs � 2. 17 48 1. 70 17 � 0. 72 69 0. 15 10 � 0. 12 05 49 .5 45 1 � 0. 05 28 � 1. 05 78 0. 69 24 (� 0. 07 69 ) (0 .2 53 8) (� 2. 56 93 )* * (9 .2 24 7) ** * (� 6. 50 91 )* ** (1 4. 81 94 )* ** (� 0. 12 37 ) (� 2. 82 27 )* ** (6 4. 76 80 )* ** d iv er si fie d fix ed in te re st o ne qu ar te r � 0. 91 72 � 0. 60 61 � 0. 20 42 0. 57 17 0. 15 94 6. 11 05 � 0. 12 38 0. 04 19 0. 75 97 (� 0. 67 55 ) (� 0. 34 70 ) (� 2. 77 07 )* ** (2 2. 31 98 )* ** (6 .5 77 8) ** * (7 .6 96 8) ** * (� 0. 67 57 ) (0 .4 39 4) (7 3. 89 93 ) t w o qu ar te rs 0. 24 02 � 0. 96 56 � 0. 20 06 0. 57 23 0. 16 01 6. 16 40 � 0. 12 71 0. 03 72 0. 75 96 (0 .1 88 3) (� 0. 53 45 ) (� 2. 72 56 )* ** (2 2. 34 43 )* ** (6 .6 05 7) ** * (7 .7 87 6) ** * (� 0. 69 37 ) (0 .3 89 6) (7 3. 86 39 ) t hr ee qu ar te rs 1. 07 42 � 1. 15 79 � 0. 20 20 0. 57 24 0. 16 04 6. 13 97 � 0. 12 46 0. 03 94 0. 75 97 (0 .8 14 0) (� 0. 66 48 ) (� 2. 74 55 )* ** (2 2. 35 15 )* ** (6 .6 20 8) ** * (7 .7 53 0) ** * (� 0. 68 04 ) (0 .4 13 5) (7 3. 89 95 ) a us tr al ia n eq ui ty o ne qu ar te r 3. 00 22 � 0. 34 47 � 0. 22 41 0. 54 27 0. 24 12 11 .9 91 8 � 0. 13 33 0. 15 88 0. 73 75 (3 .6 18 5) ** * (� 0. 32 36 ) (� 5. 20 68 )* ** (8 7. 53 97 )* ** (3 9. 63 34 )* ** (1 9. 88 63 )* ** (� 2. 04 75 )* * (3 .0 35 2) ** * (9 6. 01 84 )* ** t w o qu ar te rs 2. 93 21 � 0. 30 66 � 0. 22 50 0. 54 25 0. 24 13 11 .9 28 6 � 0. 13 35 0. 16 11 0. 73 75 (3 .5 65 4) ** * (� 0. 28 79 ) (� 5. 22 62 )* ** (8 7. 50 99 )* ** (3 9. 63 81 )* ** (1 9. 84 09 )* ** (� 2. 04 98 )* * (3 .0 81 0) ** * (9 6. 01 73 )* ** t hr ee qu ar te rs 2. 17 56 � 0. 41 25 � 0. 21 91 0. 54 32 0. 24 08 11 .9 87 4 � 0. 13 01 0. 16 23 0. 73 73 ** * (2 .1 56 4) ** (� 0. 38 72 ) (� 5. 08 77 )* ** (8 7. 60 77 )* ** (3 9. 55 35 )* ** (1 9. 92 39 )* ** (� 1. 99 74 )* * (3 .1 02 7) ** * (9 5. 93 77 ) a us tr al ia n sm al l co m pa ny o ne qu ar te r 1. 14 27 0. 36 33 � 0. 33 28 0. 94 28 � 0. 08 44 5. 71 41 0. 08 68 0. 03 05 0. 83 64 (0 .6 62 3) (0 .1 37 7) (� 2. 61 56 )* ** (5 0. 23 63 )* ** (� 4. 67 79 )* ** (3 .8 88 6) ** * (0 .4 69 9) (0 .2 30 3) (1 35 .1 10 )* ** t w o qu ar te rs 1. 64 18 0. 40 53 � 0. 33 39 0. 94 25 � 0. 08 43 5. 69 64 0. 08 05 0. 03 32 0. 83 64 (0 .9 97 2) (0 .1 53 9) (� 2. 62 57 )* ** (5 0. 21 33 )* ** (� 4. 67 46 )* ** (3 .8 74 4) ** * (0 .4 35 5) (0 .2 50 4) (1 35 .1 17 )* ** t hr ee qu ar te rs 1. 40 84 0. 39 86 � 0. 34 04 0. 94 28 � 0. 08 48 5. 68 35 0. 08 24 0. 03 33 0. 83 64 (0 .8 62 6) (0 .1 51 3) (� 2. 67 02 )* ** (5 0. 23 40 )* ** (� 4. 70 06 )* ** (3 .8 59 0) ** * (0 .4 45 8) (0 .2 51 5) (1 35 .0 95 )* ** 230 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 4 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f st at ) a us tr al ia n fix ed in te re st o ne qu ar te r � 5. 12 75 � 6. 44 45 � 0. 11 78 0. 36 74 0. 09 94 12 .6 23 5 � 0. 38 49 0. 49 95 0. 39 92 (� 1. 19 30 ) (� 1. 12 41 ) (� 1. 25 51 ) (2 5. 81 55 )* ** (6 .1 73 3) ** * (1 0. 88 21 )* ** (� 2. 50 06 )* * (3 .8 12 6) ** * (2 0. 89 96 )* ** t w o qu ar te rs � 9. 99 28 � 6. 54 38 � 0. 11 50 0. 36 69 0. 10 01 12 .7 19 3 � 0. 38 85 0. 50 51 0. 39 94 (� 2. 35 19 )* * (� 1. 14 35 ) (� 1. 22 55 ) (2 5. 77 20 )* ** (6 .2 11 5) ** * (1 0. 93 65 )* ** (� 2. 52 36 )* * (3 .8 52 6) ** * (2 0. 91 16 )* ** t hr ee qu ar te rs � 12 .5 71 7 � 6. 60 61 � 0. 11 71 0. 36 67 0. 10 03 12 .6 24 4 � 0. 39 17 0. 50 43 0. 39 94 (� 2. 44 74 )* * (� 1. 15 32 ) (� 1. 24 73 ) (2 5. 74 12 )* ** (6 .2 22 0) ** * (1 0. 86 31 )* ** (� 2. 54 37 )* * (3 .8 45 5) ** * (2 0. 91 53 )* ** m an ag ed ba la nc ed o ne qu ar te r � 3. 74 36 � 0. 46 96 � 0. 54 94 0. 47 93 0. 17 20 43 .0 50 3 0. 40 71 0. 29 01 0. 59 70 (� 0. 70 26 ) (� 0. 93 47 ) (� 3. 61 71 )* ** (5 4. 93 00 )* ** (1 8. 69 48 )* ** (1 7. 15 83 )* ** (1 .9 57 6) ** (1 .5 15 1) (5 7. 06 99 )* ** t w o qu ar te rs � 2. 42 37 � 0. 46 97 � 0. 54 81 0. 47 93 0. 17 20 43 .1 31 9 0. 41 12 0. 28 87 0. 59 70 (� 0. 45 25 ) (� 0. 93 49 ) (� 3. 60 89 )* ** (5 4. 92 63 )* ** (1 8. 69 48 )* ** (1 7. 18 54 )* ** (1 .9 77 4) ** (1 .5 08 1) (5 7. 06 80 )* ** t hr ee qu ar te rs � 0. 80 49 � 0. 50 06 � 0. 54 74 0. 47 93 0. 17 20 43 .0 94 9 0. 41 12 0. 28 84 0. 59 70 (� 0. 20 70 ) (� 0. 95 38 ) (� 3. 60 42 )* ** (5 4. 92 19 )* ** (1 8. 69 26 )* ** (1 7. 18 09 )* ** (1 .9 77 3) ** (1 .5 06 2) (5 7. 06 67 )* ** m an ag ed gr ow th o ne qu ar te r 1. 24 72 0. 29 70 � 0. 13 21 0. 49 38 0. 15 93 30 .4 40 0 0. 14 41 0. 27 95 0. 62 06 (1 .2 52 1) (0 .5 40 7) (� 1. 61 77 ) (7 6. 47 15 )* ** (2 3. 88 97 )* ** (2 2. 68 17 )* ** (1 .3 61 9) (2 .7 84 5) ** * (6 4. 65 34 )* ** t w o qu ar te rs 1. 38 92 0. 24 58 � 0. 13 27 0. 49 38 0. 15 93 30 .4 15 0 0. 14 41 0. 27 94 0. 62 06 (1 .3 49 6) (0 .4 28 8) (� 1. 62 61 ) (7 6. 46 98 )* ** (2 3. 89 08 )* ** (2 2. 76 10 )* ** (1 .3 62 4) (2 .7 84 1) ** * (6 4. 65 37 )* ** t hr ee qu ar te rs 0. 28 07 0. 37 80 � 0. 13 15 0. 49 39 0. 15 92 30 .4 39 1 0. 14 21 0. 28 06 0. 62 06 (0 .3 36 0) (0 .6 03 7) (� 1. 61 11 ) (7 6. 47 46 )* ** (2 3. 88 03 )* ** (2 2. 77 91 )* ** (1 .3 43 4) (2 .7 95 9) ** * (6 4. 64 24 )* ** m or tg ag e o ne qu ar te r 37 .3 47 8 � 17 .8 64 4 � 0. 89 62 0. 65 01 0. 10 23 34 .2 95 9 � 2. 53 74 � 0. 86 81 0. 64 96 (1 .1 98 1) (� 0. 31 99 ) (� 1. 83 07 )* (3 0. 36 24 )* ** (4 .8 81 3) ** * (5 .1 84 5) ** * (� 3. 37 42 )* ** (� 1. 19 64 ) (5 7. 59 91 )* ** t w o qu ar te rs 47 .7 89 8 � 21 .7 10 4 � 0. 92 80 0. 64 98 0. 10 24 34 .2 30 7 � 2. 57 15 � 0. 84 68 0. 64 97 (1 .5 43 5) (� 0. 38 72 ) (� 1. 88 85 )* (3 0. 34 74 )* ** (4 .8 86 7) ** * (5 .1 79 3) ** * (� 3. 41 68 )* ** (� 1. 16 66 ) (5 7. 62 15 )* ** t hr ee qu ar te rs 59 .5 57 4 � 17 .4 46 6 � 0. 89 80 0. 64 99 0. 10 19 34 .3 75 8 � 2. 57 82 � 0. 84 61 0. 64 98 (1 .6 68 1) (� 0. 31 06 ) (� 1. 83 23 )* (3 0. 34 44 )* ** (4 .8 62 5) ** * (5 .1 97 9) ** * (� 3. 41 63 )* ** (� 1. 16 54 ) (5 7. 63 76 )* ** m an ag ed st ab le o ne qu ar te r 13 .6 79 0 � 0. 79 42 � 0. 25 76 0. 26 23 0. 27 32 32 .6 05 0 � 0. 13 00 0. 11 57 0. 44 38 (2 .8 64 8) ** * (� 0. 29 84 ) (� 2. 73 76 )* ** (2 8. 27 92 )* ** (2 4. 06 44 )* ** (2 2. 71 13 )* ** (� 0. 92 27 ) (0 .9 46 2) (2 9. 32 19 )* ** t w o qu ar te rs 5. 00 98 � 1. 45 33 � 0. 25 75 0. 26 27 0. 27 29 32 .5 65 5 � 0. 13 07 0. 12 11 0. 44 36 (1 .2 36 5) (� 0. 51 07 ) (� 2. 73 52 )* ** (2 8. 31 87 )* ** (2 4. 03 60 )* ** (2 2. 67 62 )* ** (� 0. 92 73 ) (0 .9 89 3) (2 9. 29 14 )* ** t hr ee qu ar te rs � 1. 18 01 � 1. 35 41 � 0. 25 34 0. 26 31 0. 27 29 32 .6 00 6 � 0. 12 91 0. 12 30 0. 44 34 (� 0. 42 34 ) (� 0. 47 25 ) (� 2. 69 29 )* ** (2 8. 36 85 )* ** (2 4. 03 45 )* ** (2 2. 69 42 )* ** (� 0. 91 60 ) (1 .0 05 1) (2 9. 27 50 )* ** o ve rs ea s: a si a pa ci fic o ne qu ar te r � 3. 88 90 � 4. 22 44 � 1. 42 71 0. 65 27 � 0. 15 16 3. 32 62 2. 04 71 0. 03 31 0. 77 98 (� 0. 81 78 ) (� 0. 65 70 ) (� 3. 91 38 )* ** (1 2. 83 45 )* ** (� 3. 02 98 )* ** (1 .1 01 3) (3 .9 44 6) ** * (0 .1 04 8) (4 8. 50 03 )* ** t w o qu ar te rs � 5. 88 09 � 4. 69 05 � 1. 44 21 0. 65 16 � 0. 14 91 3. 29 33 2. 03 80 0. 00 22 0. 78 05 (� 1. 28 34 ) (� 0. 72 91 ) (� 3. 96 46 )* ** (1 2. 84 30 )* ** (� 2. 97 87 )* ** (1 .0 92 2) (3 .9 33 8) ** * (0 .0 07 1) (4 8. 69 73 )* ** t hr ee qu ar te rs � 6. 20 28 � 4. 84 90 � 1. 48 08 0. 65 00 � 0. 14 68 3. 39 01 1. 97 26 0. 05 55 0. 78 07 (� 1. 36 38 ) (� 0. 75 35 ) (� 4. 07 21 )* ** (1 2. 82 52 )* ** (� 2. 93 28 )* ** (1 .1 25 3) (3 .7 88 9) ** * (0 .1 77 9) (4 8. 74 31 )* ** 231r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 4 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f st at ) o ve rs ea s: fi xe d in co m e o ne qu ar te r 2. 04 28 � 0. 18 12 � 0. 29 97 0. 52 42 0. 23 73 5. 35 93 0. 36 82 � 0. 20 87 0. 59 45 (0 .6 98 6) (� 0. 23 62 ) (� 2. 58 66 )* ** (1 4. 59 40 )* ** (6 .6 72 1) ** * (4 .1 76 4) ** * (1 .8 20 0) (� 1. 30 46 ) (3 3. 78 13 )* ** t w o qu ar te rs � 3. 43 01 � 0. 06 39 � 0. 27 91 0. 52 28 0. 23 95 5. 01 74 0. 34 89 � 0. 18 23 0. 59 55 (� 1. 39 31 ) (� 0. 08 38 ) (� 2. 41 75 )* * (1 4. 58 51 )* ** (6 .7 35 7) ** * (3 .9 56 7) ** * (1 .7 25 2) (� 1. 15 39 ) (3 3. 91 71 )* ** t hr ee qu ar te rs � 4. 58 69 0. 48 02 � 0. 28 12 0. 52 23 0. 23 93 5. 03 09 0. 35 25 � 0. 19 97 0. 59 59 (� 1. 84 88 ) (0 .4 75 2) (� 2. 44 15 )* * (1 4. 57 52 )* ** (6 .7 35 3) ** * (3 .9 89 8) ** * (1 .7 45 2) (� 1. 25 67 ) (3 3. 96 40 )* ** o ve rs ea s: g lo ba l o ne qu ar te r 1. 46 04 1. 97 14 � 0. 18 54 0. 68 70 0. 08 86 10 .2 69 0 � 0. 02 32 � 0. 01 79 0. 63 35 (1 .2 13 3) (1 .1 27 5) (� 3. 16 07 )* ** (7 6. 84 57 )* ** (9 .8 79 1) ** * (1 1. 30 03 )* ** (� 0. 22 06 ) (� 0. 20 42 ) (5 5. 75 48 )* ** t w o qu ar te rs 0. 41 56 2. 01 20 � 0. 18 93 0. 68 71 0. 08 85 10 .1 66 9 � 0. 02 16 � 0. 01 72 0. 63 34 (0 .3 62 0) (1 .1 50 1) (� 3. 22 83 )* ** (7 6. 84 59 )* ** (9 .8 69 4) ** * (1 1. 16 70 )* ** (� 0. 20 52 ) (� 0. 19 60 ) (5 5. 73 79 )* ** t hr ee qu ar te rs � 0. 44 74 2. 03 42 � 0. 19 03 0. 68 71 0. 08 88 10 .2 32 3 � 0. 02 03 � 0. 01 78 0. 63 34 (� 0. 39 86 ) (1 .1 63 5) (� 3. 24 52 )* ** (7 6. 88 24 )* ** (9 .8 91 1) ** * (1 1. 22 76 )* ** (� 0. 19 31 ) (� 0. 20 29 ) (5 5. 74 55 )* ** o ve rs ea s: pr op er ty o ne qu ar te r � 2. 51 99 � 1. 11 38 � 0. 53 09 � 0. 01 33 0. 29 61 1. 82 00 � 0. 56 78 0. 04 17 0. 30 97 (� 0. 30 44 ) (� 0. 18 28 ) (� 0. 73 08 ) (� 0. 25 42 ) (6 .1 81 8) ** * (0 .9 39 1) (� 0. 17 55 ) (0 .0 39 9) (6 .5 25 5) ** * t w o qu ar te rs � 1. 35 22 � 1. 40 20 � 0. 54 65 � 0. 01 32 0. 29 61 1. 81 03 � 0. 48 09 0. 02 36 0. 30 96 (� 0. 18 13 ) (� 0. 22 85 ) (� 0. 75 49 ) (� 0. 25 27 ) (6 .1 81 4) ** * (0 .9 34 0) (� 0. 14 94 ) (0 .0 22 6) (6 .5 22 3) ** * t hr ee qu ar te rs � 1. 13 57 � 1. 43 29 � 0. 54 44 � 0. 01 32 0. 29 61 1. 80 78 � 0. 41 17 0. 02 14 0. 30 96 (� 0. 15 46 ) (� 0. 23 07 ) (� 0. 75 06 ) (� 0. 25 34 ) (6 .1 81 8) ** * (0 .9 32 7) (� 0. 12 90 ) (0 .0 20 5) (6 .5 21 9) ** * a us tr al ia n pr op er ty o ne qu ar te r 1. 93 91 0. 82 23 � 0. 26 23 0. 68 85 0. 12 72 12 .8 37 1 � 0. 79 76 0. 27 64 0. 74 20 (0 .8 83 6) (0 .2 98 7) (� 1. 35 19 ) (3 3. 23 26 )* ** (6 .4 67 1) ** * (5 .4 45 9) ** * (� 3. 15 89 )* ** (1 .1 76 3) (8 2. 46 68 )* ** t w o qu ar te rs 2. 48 50 0. 04 73 � 0. 27 63 0. 68 77 0. 12 73 12 .5 81 0 � 0. 81 53 0. 35 84 0. 74 23 (1 .3 56 2) (0 .0 17 0) (� 1. 42 37 ) (3 3. 24 96 )* ** (6 .4 74 6) ** * (5 .3 60 0) ** * (� 3. 23 95 )* ** (1 .5 51 1) (8 2. 60 19 )* ** t hr ee qu ar te rs 0. 97 11 � 0. 23 94 � 0. 27 16 0. 68 87 0. 12 68 12 .4 87 6 � 0. 79 27 0. 32 87 0. 74 26 (0 .5 26 9) (� 0. 08 66 ) (� 1. 40 42 ) (3 3. 29 34 )* ** (6 .4 53 2) ** * (5 .3 23 4) ** * (� 3. 15 30 )* ** (1 .4 23 3) (8 2. 70 67 )* ** m ix ed po rt fo lio s o ne qu ar te r � 0. 46 14 � 0. 62 24 � 4. 74 30 0. 21 96 0. 11 23 43 9. 48 62 15 .0 89 0 � 15 .1 51 8 0. 50 68 (0 .4 85 6) (� 0. 56 00 ) (� 2. 01 63 )* ** (1 3. 25 22 )* ** (5 .8 38 6) ** * (1 2. 93 49 )* ** (4 .4 49 1) ** * (� 5. 66 43 )* ** (3 2. 34 35 )* ** t w o qu ar te rs � 0. 46 72 � 0. 42 56 � 4. 72 13 0. 21 95 0. 11 22 44 0. 98 21 15 .2 29 2 � 15 .2 73 8 0. 50 67 (� 0. 25 90 ) (� 0. 31 61 ) (� 2. 00 75 )* ** (1 3. 24 91 )* ** (5 .8 34 1) ** * (1 3. 02 68 )* ** (4 .5 06 4) ** * (� 5. 73 33 )* ** (3 2. 34 04 )* ** t hr ee qu ar te rs � 0. 32 31 � 0. 43 70 � 4. 72 19 0. 21 95 0. 11 22 44 1. 00 13 15 .2 16 9 � 15 .2 62 7 0. 50 67 (� 0. 18 57 ) (� 0. 27 93 ) (� 2. 00 77 )* ** (1 3. 24 85 )* ** (5 .8 34 4) ** * (1 3. 02 56 )* ** (4 .5 03 3) ** * (� 5. 73 06 )* ** (3 2. 33 99 )* ** ** * si gn ifi ca nt at 1% , ** si gn ifi ca nt at 5% . 232 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 5 fi xe d ef fe ct s pa ne l re gr es si on of su pe ra nn ua tio n fu nd ne t flo w s w ith pa st ex ce ss re tu rn s an d ri sk r eg re ss io n eq ua tio n: n f i, t � b j r j t, t � c� i, t� 1 � ds iz e i ,t � 1 � en f i, t� 1 � fn f i, t� 2 � ge in i, t � hc h t � jg f c t � � j� � 1 1 k i ,j � x i ,t � � t w he re n f i, t ar e th e ne t flo w in to th e fu nd , r j i, is th e tr ai lin g ex ce ss re tu rn s fo r ei th er 1, 2, or 3 qu ar te rs , � i, t is th e tr ac ki ng er ro r, si ze i, t, is th e lo g of th e fu nd si ze , c h t is th e ch oi ce du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 05 an d 1 fo r re st of th e tim e pe ri od , g f c t is th e cr is is du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 08 an d 1 fo r th e re st of th e tim e pe ri od , e in t is a va ri ab le to ca pt ur e th e ov er al l flo w in to th e m ar ke t, an d f e i, t is th e fix ed ef fe ct of ith fu nd . � x i ar e th e fir st di ff er en ce te rm s of th e re gr es si on va ri ab le s. w e es tim at ed th re e se pa ra te re gr es si on s w ith th e va ri ab le r j i, t be in g 1, 2, or 3 qu ar te r ex ce ss re tu rn s. fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f st at ) a lte rn at iv es o ne qu ar te r � 0. 06 38 0. 12 25 0. 05 43 0. 57 11 � 0. 09 08 4. 68 21 � 0. 16 48 0. 14 79 0. 85 04 (� 0. 12 18 ) (0 .1 61 4) (0 .7 56 7) (7 .9 65 0) ** * (� 1. 83 10 )* (5 .9 34 9) ** * (� 0. 22 38 ) (1 .8 92 3) (4 6. 20 19 )* ** t w o qu ar te rs 0. 03 05 0. 33 53 0. 05 75 0. 57 65 � 0. 08 85 4. 71 92 � 0. 13 25 0. 13 49 0. 85 04 (0 .0 84 2) (0 .4 21 1) (0 .8 07 5) (8 .0 96 4) ** * (� 1. 78 46 )* (5 .9 67 0) ** * (� 0. 22 42 ) (1 .7 87 7) (4 6. 20 19 )* ** t hr ee qu ar te rs � 0. 23 49 0. 50 22 0. 05 91 0. 57 56 � 0. 08 68 4. 77 55 � 0. 13 49 0. 13 34 0. 85 04 (� 0. 66 14 ) (0 .6 45 1) (0 .8 32 7) (8 .0 77 8) ** * (� 1. 74 60 )* (5 .9 69 7) ** * (� 0. 22 14 ) (1 .7 65 7) (4 6. 20 19 )* ** c as h o ne qu ar te r 0. 30 59 � 32 .4 33 3 � 0. 79 40 0. 25 00 � 0. 04 15 45 .0 97 8 0. 41 38 � 1. 29 66 0. 41 81 (0 .0 06 9) (� 0. 86 77 ) (� 3. 12 51 )* ** (1 4. 06 47 )* ** (� 2. 20 70 )* * (1 4. 73 57 )* ** (1 .0 85 4) (� 3. 80 51 )* ** (2 1. 76 92 )* ** t w o qu ar te rs � 9. 96 31 � 30 .8 35 9 � 0. 78 81 0. 25 02 � 0. 04 16 44 .9 22 2 0. 41 16 � 1. 29 43 0. 41 80 (� 0. 22 59 ) (� 0. 82 74 ) (� 3. 09 19 )* ** (1 4. 07 11 )* ** (� 2. 21 32 )* * (1 4. 70 52 )* ** (1 .0 79 3) (� 3. 79 82 )* ** (2 1. 76 39 )* ** t hr ee qu ar te rs 5. 56 24 � 30 .8 71 0 � 0. 80 27 0. 25 00 � 0. 04 17 44 .9 14 3 0. 40 62 � 1. 29 23 0. 41 79 (0 .1 41 5) (� 0. 77 99 ) (� 3. 15 02 )* ** (1 4. 06 23 )* ** (� 2. 21 79 )* * (1 4. 70 12 )* ** (1 .0 65 4) (� 3. 79 09 )* ** (2 1. 75 78 )* ** d iv er si fie d fix ed in te re st o ne qu ar te r � 1. 02 30 � 0. 63 72 � 0. 23 32 0. 55 39 0. 17 07 6. 15 38 0. 15 11 � 0. 02 03 0. 78 22 (� 0. 66 61 ) (� 0. 33 43 ) (� 2. 67 26 )* ** (1 8. 12 82 )* ** (5 .9 62 0) ** * (6 .6 09 6) ** * (0 .7 23 9) (� 0. 17 93 ) (7 8. 48 31 )* ** t w o qu ar te rs 0. 37 44 � 1. 00 59 � 0. 22 46 0. 55 55 0. 17 06 6. 27 55 0. 14 09 � 0. 03 27 0. 78 20 (0 .2 66 6) (� 0. 50 93 ) (� 2. 58 47 )* ** (1 8. 19 77 )* ** (5 .9 52 7) ** * (6 .7 89 9) ** * (0 .6 75 6) (� 0. 28 97 ) (7 8. 40 20 )* ** t hr ee qu ar te rs 0. 65 78 � 1. 01 18 � 0. 22 41 0. 55 55 0. 17 06 6. 26 57 0. 13 99 � 0. 03 25 0. 78 21 (0 .4 53 3) (� 0. 53 26 ) (� 2. 57 92 )* ** (1 8. 19 99 )* ** (5 .9 55 0) ** * (6 .7 78 3) ** * (0 .6 70 7) (� 0. 28 78 ) (7 8. 41 25 )* ** a us tr al ia n eq ui ty o ne qu ar te r 0. 69 72 0. 68 64 � 0. 13 77 0. 55 29 0. 22 38 9. 75 99 � 0. 15 67 0. 06 96 0. 71 24 (1 .2 11 6) (0 .8 96 8) (� 5. 37 66 )* ** (7 3. 90 76 )* ** (3 0. 13 20 )* ** (2 4. 21 54 )* ** (� 3. 64 82 )* ** (2 .1 01 0) ** (8 3. 42 68 )* ** t w o qu ar te rs 0. 97 54 0. 65 87 � 0. 13 83 0. 55 28 0. 22 39 9. 74 83 � 0. 15 72 0. 07 02 0. 71 24 (1 .7 25 0) (0 .8 60 5) (� 5. 39 95 )* ** (7 3. 88 32 )* ** (3 0. 14 65 )* ** (2 4. 26 87 )* ** (� 3. 66 03 )* ** (2 .1 22 7) ** (8 3. 43 23 )* ** t hr ee qu ar te rs 0. 15 17 0. 75 84 � 0. 13 67 0. 55 31 0. 22 35 9. 76 10 � 0. 15 62 0. 07 00 0. 71 24 (0 .2 77 0) (0 .9 92 0) (� 5. 33 87 )* ** (7 3. 94 36 )* ** (3 0. 10 58 )* ** (2 4. 28 83 )* ** (� 3. 63 69 )* ** (2 .1 16 0) ** (8 3. 41 54 )* ** a us tr al ia n sm al l co m pa ny o ne qu ar te r � 0. 86 76 � 0. 19 43 � 0. 11 45 0. 49 06 0. 17 17 4. 97 87 � 0. 06 45 � 0. 02 36 0. 59 26 (� 1. 20 87 ) (� 0. 18 20 ) (� 2. 30 32 )* * (1 8. 06 49 )* ** (6 .7 99 5) ** * (7 .7 80 5) ** * (� 0. 79 59 ) (� 0. 45 28 ) (3 2. 44 86 )* ** t w o qu ar te rs 0. 28 89 � 0. 02 94 � 0. 11 52 0. 49 38 0. 17 08 4. 93 07 � 0. 07 45 � 0. 01 71 0. 59 15 (0 .4 22 4) (� 0. 02 76 ) (� 2. 30 68 )* * (1 8. 19 17 )* ** (6 .7 54 3) ** * (7 .6 93 7) ** * (� 0. 91 76 ) (� 0. 32 53 ) (3 2. 30 49 )* ** t hr ee qu ar te rs � 0. 19 51 � 0. 05 04 � 0. 11 07 0. 49 36 0. 17 05 4. 99 12 � 0. 07 21 � 0. 01 91 0. 59 17 (� 0. 29 16 ) (� 0. 04 73 ) (� 2. 20 87 )* * (1 8. 18 92 )* ** (6 .7 47 8) ** * (7 .7 74 6) ** * (� 0. 88 75 ) (� 0. 36 38 ) (3 2. 32 65 )* ** 233r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 5 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) a us tr al ia n fix ed in te re st o ne qu ar te r 0. 79 35 0. 18 10 � 0. 09 77 0. 55 90 0. 20 71 5. 35 45 � 0. 24 11 0. 23 52 0. 80 91 (0 .3 84 6) (0 .0 72 2) (� 2. 20 98 )* ** (3 0. 74 41 )* ** (1 1. 26 29 )* ** (9 .7 54 8) ** * (� 3. 20 06 )* ** (3 .7 63 7) ** * (1 12 .7 15 )* ** t w o qu ar te rs � 0. 14 43 0. 13 21 � 0. 09 86 0. 55 89 0. 20 71 5. 36 66 � 0. 24 11 0. 23 77 0. 80 20 (� 0. 07 03 ) (0 .0 52 6) (� 2. 23 94 )* ** (3 0. 73 92 )* ** (1 1. 26 27 )* ** (9 .7 53 7) ** * (� 3. 20 06 )* ** (3 .7 93 4) ** * (1 12 .7 17 )* ** t hr ee qu ar te rs � 1. 55 24 0. 37 80 � 0. 09 94 0. 55 87 0. 20 74 5. 38 05 � 0. 24 15 0. 23 86 0. 80 20 (� 0. 81 49 ) (0 .1 49 5) (� 2. 25 70 )* ** (3 0. 72 71 )* ** (1 1. 27 91 )* ** (9 .7 89 1) ** * (� 3. 20 71 )* ** (3 .8 11 6) ** * (1 12 .7 55 )* ** m an ag ed ba la nc ed o ne qu ar te r � 4. 94 83 � 0. 50 91 � 0. 45 70 0. 46 51 0. 16 79 45 .1 14 2 0. 46 48 0. 26 53 0. 58 77 (� 0. 74 85 ) (� 0. 91 44 ) (� 2. 56 97 )* * (4 6. 44 81 )* ** (1 5. 69 11 )* ** (1 4. 91 74 )* ** (1 .7 87 9) * (1 .1 30 1) (5 3. 89 38 )* ** t w o qu ar te rs � 3. 11 90 � 0. 51 00 � 0. 45 51 0. 46 50 0. 16 80 45 .2 51 8 0. 47 08 0. 26 25 0. 58 77 (� 0. 46 80 ) (� 0. 91 61 ) (� 2. 55 94 )* * (4 6. 44 09 )* ** (1 5. 69 53 )* ** (1 4. 96 40 )* ** (1 .8 09 8) * (1 .1 18 2) (5 3. 89 02 )* ** t hr ee qu ar te rs � 1. 47 46 � 0. 55 06 � 0. 45 43 0. 46 50 0. 16 80 45 .2 15 0 0. 46 97 0. 26 25 0. 58 77 (� 0. 30 78 ) (� 0. 94 49 ) (� 2. 55 46 )* * (4 6. 43 74 )* ** (1 5. 69 00 )* ** (1 4. 96 16 )* ** (1 .8 05 7) * (1 .1 18 4) (5 3. 88 94 )* ** m an ag ed gr ow th o ne qu ar te r 0. 91 26 0. 50 74 � 0. 05 46 0. 44 84 0. 17 89 31 .9 33 4 0. 16 64 0. 32 37 0. 61 21 (0 .3 10 2) (0 .7 02 2) (� 0. 57 58 ) (5 8. 49 96 )* ** (2 2. 30 74 )* ** (1 9. 20 04 )* ** (1 .2 32 1) (2 .5 89 8) ** * (6 1. 40 25 )* ** t w o qu ar te rs 0. 16 50 0. 50 35 � 0. 05 39 0. 44 84 0. 17 89 32 .0 18 2 0. 16 92 0. 32 26 0. 61 21 (0 .0 54 6) (0 .6 96 8) (� 0. 56 88 ) (5 8. 50 01 )* ** (2 2. 30 68 )* ** (1 9. 27 69 )* ** (1 .2 52 5) (2 .5 81 5) ** * (6 1. 40 53 )* ** t hr ee qu ar te rs � 1. 71 37 0. 37 67 � 0. 05 33 0. 44 84 0. 17 88 32 .0 64 4 0. 16 98 0. 32 44 0. 61 21 (� 0. 77 73 ) (0 .4 71 3) (� 0. 56 23 ) (5 8. 50 48 )* ** (2 2. 29 43 )* ** (1 9. 31 27 )* ** (1 .2 56 5) (2 .5 95 6) ** * (6 1. 40 06 )* ** m or tg ag e o ne qu ar te r 9. 02 19 25 .5 92 8 � 0. 03 81 0. 50 90 0. 11 32 11 .3 56 4 � 0. 56 91 � 0. 46 11 0. 76 49 (1 .0 36 6) (1 .4 29 9) (� 0. 58 22 ) (1 5. 31 29 )* ** (4 .2 37 0) ** * (1 0. 49 23 )* ** (� 4. 45 88 )* ** (� 4. 35 28 )* ** (7 4. 58 99 )* ** t w o qu ar te rs 6. 03 71 23 .9 35 2 � 0. 03 09 0. 51 18 0. 11 25 11 .2 99 4 � 0. 55 28 � 0. 44 78 0. 77 85 (0 .6 64 9) (1 .3 50 8) (� 0. 47 01 ) (1 5. 37 04 )* ** (4 .2 17 8) ** * (1 0. 44 13 )* ** (� 4. 31 16 )* ** (� 4. 22 91 )* ** (7 4. 71 67 )* ** t hr ee qu ar te rs � 6. 02 39 21 .5 53 7 � 0. 00 03 0. 51 74 0. 11 14 10 .9 97 0 � 0. 48 18 � 0. 43 96 0. 76 36 (� 0. 64 42 ) (1 .2 01 1) (� 0. 00 43 ) (1 5. 49 45 )* ** (4 .1 61 5) ** * (1 0. 13 40 )* ** (� 3. 72 95 )* ** (� 4. 09 97 )* ** (7 4. 05 60 )* ** m an ag ed st ab le o ne qu ar te r 10 .2 88 6 � 1. 05 00 � 0. 06 79 0. 22 65 0. 27 60 34 .6 21 9 0. 00 94 0. 07 09 0. 41 00 (1 .5 49 7) (� 0. 21 15 ) (� 0. 59 92 ) (2 0. 79 51 )* ** (1 9. 43 27 )* ** (1 8. 55 08 )* ** (0 .0 51 8) (0 .4 52 2) (2 5. 19 89 )* ** t w o qu ar te rs 0. 42 35 � 1. 27 92 � 0. 06 73 0. 22 67 0. 27 62 34 .6 18 7 0. 01 59 0. 07 26 0. 40 99 (0 .0 63 2) (� 0. 25 79 ) (� 0. 59 34 ) (2 0. 81 91 )* ** (1 9. 44 05 )* ** (1 8. 54 63 )* ** (0 .0 87 2) (0 .4 62 8) (2 5. 18 83 )* ** t hr ee qu ar te rs � 3. 68 48 � 1. 03 25 � 0. 06 51 0. 22 69 0. 27 63 34 .6 45 8 0. 01 91 0. 07 25 0. 40 98 (� 0. 63 68 ) (� 0. 18 16 ) (� 0. 57 38 ) (2 0. 83 48 )* ** (1 9. 44 26 )* ** (1 8. 56 24 )* ** (0 .1 05 2) (0 .4 61 5) (2 5. 18 33 )* ** 234 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 5 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f st at ) o ve rs ea s: fi xe d in co m e o ne qu ar te r 5. 00 56 � 0. 29 39 � 0. 46 72 0. 50 40 0. 28 17 4. 85 37 0. 75 42 � 0. 40 25 0. 65 20 (1 .0 03 9) (� 0. 34 72 ) (� 2. 89 76 )* ** (1 1. 26 39 )* ** (6 .3 12 6) ** * (2 .4 81 4) ** (2 .5 18 1) ** (� 1. 84 47 ) (3 5. 80 54 )* ** t w o qu ar te rs � 3. 47 50 � 0. 14 31 � 0. 44 60 0. 50 34 0. 28 49 4. 05 35 0. 72 46 � 0. 36 89 0. 65 22 (� 0. 82 50 ) (� 0. 17 05 ) (� 2. 77 45 )* ** (1 1. 25 69 )* ** (6 .3 82 7) ** * (2 .1 41 8) ** (2 .4 19 9) ** (� 1. 70 57 ) (3 5. 83 18 )* ** t hr ee qu ar te rs � 4. 57 91 0. 52 52 � 0. 44 79 0. 50 33 0. 28 46 3. 98 46 0. 72 98 � 0. 38 90 0. 65 27 (� 1. 45 83 ) (0 .4 61 0) (� 2. 79 36 )* ** (1 1. 27 64 )* ** (6 .3 82 7) ** * (2 .1 48 2) ** (2 .4 42 5) ** (� 1. 78 94 ) (3 5. 90 62 )* ** o ve rs ea s: g lo ba l o ne qu ar te r 1. 00 29 1. 52 52 � 0. 08 91 0. 67 20 0. 09 66 7. 65 98 � 0. 06 21 0. 00 17 0. 73 04 (1 .8 26 0) (1 .6 48 9) (� 3. 35 22 )* ** (6 4. 14 93 )* ** (9 .8 98 1) ** * (1 7. 43 63 )* ** (� 1. 26 44 ) (0 .0 43 9) (8 4. 73 90 )* ** t w o qu ar te rs 0. 24 48 1. 63 83 � 0. 09 00 0. 67 21 0. 09 65 7. 65 86 � 0. 06 07 0. 00 02 0. 73 04 (0 .4 64 3) (1 .7 70 0) (� 3. 38 82 )* ** (6 4. 14 54 )* ** (9 .8 84 1) ** * (1 7. 42 46 )* ** (� 1. 23 70 ) (0 .0 05 8) (8 4. 73 42 )* ** t hr ee qu ar te rs 0. 15 79 1. 65 97 � 0. 08 96 0. 67 27 0. 09 57 7. 64 93 � 0. 06 17 0. 00 09 0. 73 04 (0 .3 04 4) (1 .7 93 8) (� 3. 37 54 )* ** (6 4. 22 07 )* ** (9 .8 03 3) ** * (1 7. 38 91 )* ** (� 1. 25 67 ) (0 .0 22 5) (8 4. 72 62 )* ** o ve rs ea s: pr op er ty o ne qu ar te r 0. 28 92 0. 85 96 � 0. 12 89 0. 34 07 0. 20 11 3. 19 07 0. 02 31 � 0. 45 24 0. 59 53 (0 .4 47 1) (1 .7 94 1) (� 1. 36 17 ) (4 .7 97 2) ** * (3 .0 65 5) ** * (3 .5 50 5) ** * (0 .1 14 8) (� 4. 54 03 )* ** (1 9. 30 84 )* ** t w o qu ar te rs 1. 00 60 0. 97 87 � 0. 12 54 0. 33 35 0. 20 53 3. 30 43 0. 01 22 � 0. 47 26 0. 64 60 (1 .7 45 3) (2 .0 23 4) ** (� 1. 33 03 ) (4 .7 05 1) ** * (3 .1 40 2) ** * (3 .6 78 2) ** * (0 .0 17 1) (� 4. 74 96 )* ** (1 9. 56 77 )* ** t hr ee qu ar te rs 0. 66 62 0. 97 21 � 0. 12 64 0. 32 77 0. 21 03 3. 27 50 0. 02 55 � 0. 46 90 0. 64 32 (1 .1 38 9) (1 .9 59 0) (� 1. 33 15 ) (4 .5 87 1) ** * (3 .1 91 3) ** * (3 .6 32 1) ** * (0 .1 57 9) (� 4. 66 67 )* ** (1 9. 34 05 )* ** a us tr al ia n pr op er ty o ne qu ar te r 0. 77 09 1. 58 20 � 0. 08 58 0. 56 56 0. 11 39 7. 70 38 � 0. 02 81 � 0. 17 70 0. 68 95 (1 .1 91 7) (1 .7 51 1) (� 1. 46 86 ) (1 8. 61 73 )* ** (4 .2 34 6) ** * (9 .5 81 7) ** * (� 0. 37 93 ) (� 2. 55 82 )* * (5 6. 33 78 )* ** t w o qu ar te rs 0. 52 31 1. 45 53 � 0. 08 98 0. 56 45 0. 11 28 7. 69 38 � 0. 03 36 � 0. 15 06 0. 70 29 (0 .9 58 0) (1 .5 90 9) (� 1. 53 93 ) (1 8. 59 27 )* ** (4 .1 91 8) ** * (9 .5 67 6) ** * (� 0. 45 38 ) (� 2. 21 21 )* * (5 6. 30 59 )* ** t hr ee qu ar te rs � 0. 17 61 1. 32 09 � 0. 09 10 0. 56 15 0. 11 44 7. 67 52 � 0. 02 29 � 0. 16 15 0. 69 04 (� 0. 31 26 ) (1 .4 55 7) (� 1. 56 36 ) (1 8. 49 23 )* ** (4 .2 57 2) ** * (9 .5 60 0) ** * (� 0. 31 00 ) (� 2. 37 10 )* * (5 6. 55 15 )* ** a us tr al ia n pr op er ty se cu ri tie s o ne qu ar te r � 1. 75 22 � 0. 00 03 � 0. 24 80 0. 60 68 0. 02 93 11 .3 41 8 � 0. 34 41 � 0. 11 50 0. 59 45 (� 1. 03 07 ) (� 0. 01 62 ) (� 4. 40 98 )* ** (3 9. 50 03 )* ** (1 .9 71 7) ** (1 3. 58 66 )* ** (� 4. 05 57 )* ** (� 1. 47 11 ) (4 7. 58 62 )* ** t w o qu ar te rs 0. 68 24 � 0. 00 05 � 0. 25 23 0. 60 69 0. 02 98 11 .3 93 3 � 0. 34 66 � 0. 11 63 0. 59 44 (0 .4 02 0) (� 0. 02 24 ) (� 4. 48 96 )* ** (3 9. 50 33 )* ** (2 .0 03 7) ** (1 3. 66 96 )* ** (� 4. 08 65 )* ** (� 1. 48 76 ) (4 7. 56 58 )* ** t hr ee qu ar te rs 0. 79 05 0. 00 54 � 0. 25 29 0. 60 69 0. 02 98 11 .3 89 0 � 0. 34 61 � 0. 11 74 0. 59 44 (0 .7 12 1) (0 .1 80 1) (� 4. 51 31 )* ** (3 9. 51 02 )* ** (2 .0 10 6) ** (1 3. 67 72 )* ** (� 4. 08 19 )* ** (� 1. 50 53 ) (4 7. 56 73 )* ** m ix ed po rt fo lio s o ne qu ar te r � 74 .4 34 7 � 26 4. 17 74 � 2. 34 57 0. 24 21 0. 08 30 37 6. 50 76 10 .8 89 3 � 9. 78 48 0. 57 94 (� 0. 67 54 ) (� 2. 10 44 )* * (� 1. 19 41 ) (1 1. 75 24 )* ** (3 .8 19 9) ** * (1 2. 56 72 )* ** (3 .6 60 9) ** * (� 4. 21 54 )* ** (4 0. 31 96 )* ** t w o qu ar te rs � 11 7. 07 23 � 26 8. 82 99 � 2. 34 09 0. 24 20 0. 08 36 37 7. 55 46 10 .9 90 2 � 9. 68 39 0. 57 95 (� 1. 07 25 ) (� 2. 15 05 )* * (� 1. 19 21 ) (1 1. 74 90 )* ** (3 .8 47 4) ** * (1 2. 62 11 )* ** (3 .7 02 7) ** * (� 4. 20 29 )* ** (4 0. 33 81 )* ** t hr ee qu ar te rs � 52 .2 44 6 � 26 5. 25 39 � 2. 36 97 0. 24 19 0. 08 34 37 5. 31 31 10 .7 74 6 � 9. 61 44 0. 57 94 (� 0. 52 50 ) (� 2. 10 43 )* * (� 1. 20 68 ) (1 1. 74 40 )* ** (3 .8 35 2) ** * (1 2. 54 83 )* ** (3 .6 35 7) ** * (� 4. 17 59 )* ** (4 0. 32 74 )* ** ** * si gn ifi ca nt at 1% , ** si gn ifi ca nt at 5% . 235r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 6 fi xe d ef fe ct s pa ne l re gr es si on of no nsu pe ra nn ua tio n fu nd ne t flo w s w ith pa st ex ce ss re tu rn s an d ri sk r eg re ss io n eq ua tio n: n f i, t � b j r j t, t � c� i, t� 1 � ds iz e i ,t � 1 � en f i, t� 1 � fn f i, t� 2 � ge in i, t � hc h t � jg f c t � � j� � 1 1 k i ,j � x i ,t � � t w he re n f i, t ar e th e ne t flo w in to th e fu nd , r j i, is th e tr ai lin g ex ce ss re tu rn s fo r ei th er 1, 2, or 3 qu ar te rs , � i, t is th e tr ac ki ng er ro r, si ze i, t, is th e lo g of th e fu nd si ze , c h t is th e ch oi ce du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 05 an d 1 fo r re st of th e tim e pe ri od , g f c t is th e cr is is du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 08 an d 1 fo r th e re st of th e tim e pe ri od , e in t is a va ri ab le to ca pt ur e th e ov er al l flo w in to th e m ar ke t, an d f e i, t is th e fix ed ef fe ct of ith fu nd . � x i ar e th e fir st di ff er en ce te rm s of th e re gr es si on va ri ab le s. w e es tim at ed th re e se pa ra te re gr es si on s w ith th e va ri ab le r j i, t be in g 1, 2, or 3 qu ar te r ex ce ss re tu rn s. fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) a lte rn at iv es o ne qu ar te r � 7. 27 78 20 .6 85 0 � 1. 02 54 0. 31 06 0. 11 03 27 .2 16 4 � 2. 24 73 � 1. 08 71 0. 51 91 (� 0. 82 32 ) (1 .5 73 1) (� 1. 51 83 ) (3 .4 20 3) ** * (1 .0 82 3) (3 .2 12 3) ** * (� 1. 68 88 ) (� 0. 94 86 ) (1 0. 01 84 )* ** t w o qu ar te rs � 3. 76 02 19 .1 68 8 � 1. 03 53 0. 31 34 0. 10 81 27 .5 59 4 � 2. 28 90 � 1. 04 68 0. 51 98 (� 0. 57 19 ) (1 .4 29 0) (� 1. 54 22 ) (3 .4 51 2) ** * (1 .0 60 5) (3 .4 64 6) ** * (� 1. 73 21 ) (� 0. 91 97 ) (1 0. 04 48 )* ** t hr ee qu ar te rs � 1. 26 70 17 .0 46 7 � 1. 06 09 0. 30 40 0. 10 34 27 .6 89 8 � 2. 32 15 � 1. 11 94 0. 51 51 (� 0. 17 72 ) (1 .2 76 7) (� 1. 57 34 ) (3 .3 22 0) ** * (1 .0 09 3) (3 .4 63 9) ** * (� 1. 74 77 ) (� 0. 97 48 ) (9 .8 74 6) ** * c as h o ne qu ar te r � 9. 09 59 2. 04 52 � 0. 32 90 � 0. 06 67 � 0. 39 98 50 .2 83 8 � 3. 76 77 1. 29 22 0. 81 11 (� 0. 12 33 ) (0 .2 80 2) (� 0. 34 39 ) (� 1. 76 76 ) (� 8. 39 08 )* ** (4 .7 02 8) ** * (� 2. 60 51 )* ** (1 .0 95 6) (8 1. 60 39 )* ** t w o qu ar te rs � 3. 82 80 1. 99 96 � 0. 32 57 � 0. 06 66 � 0. 39 98 50 .4 05 6 � 3. 75 89 1. 28 70 0. 81 11 (� 0. 05 23 ) (0 .2 74 0) (� 0. 34 02 ) (� 1. 76 68 ) (� 8. 39 03 )* ** (4 .7 11 7) ** * (� 2. 59 94 )* ** (1 .0 90 8) (8 1. 59 63 )* ** t hr ee qu ar te rs � 8. 88 60 2. 10 10 � 0. 32 22 � 0. 06 67 � 0. 39 98 50 .4 74 2 � 3. 76 31 1. 28 80 0. 81 11 (� 0. 14 83 ) (0 .2 12 9) (� 0. 33 67 ) (� 1. 76 77 ) (� 8. 38 97 )* ** (4 .7 23 2) ** * (� 2. 60 30 )* ** (1 .0 91 6) (8 1. 59 45 )* ** d iv er si fie d fix ed in te re st o ne qu ar te r � 1. 67 85 � 0. 12 25 � 0. 13 68 0. 61 27 0. 14 55 5. 26 15 � 1. 14 14 0. 26 08 0. 63 90 (� 0. 53 96 ) (� 0. 02 78 ) (� 0. 85 30 ) (1 2. 73 81 )* ** (3 .0 82 0) ** * (3 .3 19 5) ** * (� 2. 89 31 )* ** (1 .2 84 8) (2 9. 45 49 )* ** t w o qu ar te rs 0. 06 39 � 0. 84 01 � 0. 14 22 0. 61 32 0. 14 69 5. 18 83 � 1. 13 21 0. 26 59 0. 63 91 (0 .0 20 9) (� 0. 18 54 ) (� 0. 88 46 ) (1 2. 74 54 )* ** (3 .1 22 8) ** * (3 .2 61 9) ** * (� 2. 86 79 )* ** (1 .3 06 8) (2 9. 46 85 )* ** t hr ee qu ar te rs 2. 69 73 � 1. 71 62 � 0. 16 22 0. 61 40 0. 14 83 5. 03 80 � 1. 10 29 0. 27 06 0. 63 94 (0 .8 45 4) (� 0. 38 72 ) (� 1. 00 38 ) (1 2. 76 86 )* ** (3 .1 52 4) ** * (3 .1 66 4) ** * (� 2. 78 71 )* ** (1 .3 31 7) (2 9. 51 02 )* ** a us tr al ia n eq ui ty o ne qu ar te r 5. 83 84 � 0. 51 59 � 0. 46 61 0. 54 69 0. 24 05 15 .7 01 7 � 0. 15 71 0. 23 15 0. 74 57 (2 .9 35 2) ** * (� 0. 20 69 ) (� 3. 07 10 )* ** (5 0. 78 77 )* ** (2 2. 86 64 )* ** (9 .8 15 7) ** * (� 0. 92 02 ) (1 .5 25 1) (9 8. 33 78 )* ** t w o qu ar te rs 4. 99 30 � 0. 42 40 � 0. 46 88 0. 54 67 0. 24 06 15 .5 60 8 � 0. 15 93 0. 23 60 0. 74 56 (2 .4 91 7) ** (� 0. 16 93 ) (� 3. 08 85 )* ** (5 0. 76 48 )* ** (2 2. 87 20 )* ** (9 .7 45 3) ** * (� 0. 93 28 ) (1 .5 55 1) (9 8. 32 25 )* ** t hr ee qu ar te rs � 0. 91 41 � 1. 44 05 � 0. 46 18 0. 54 76 0. 24 02 15 .6 94 0 � 0. 15 70 0. 23 96 0. 74 54 (� 0. 46 05 ) (� 0. 57 65 ) (� 3. 04 15 )* ** (5 0. 83 05 )* ** (2 2. 82 12 )* ** (9 .8 18 8) ** * (� 0. 91 92 ) (1 .5 78 4) (9 8. 19 26 )* ** a us tr al ia n sm al l co m pa ny o ne qu ar te r 2. 18 42 1. 44 55 � 0. 58 12 0. 96 55 � 0. 10 58 10 .0 73 7 � 0. 02 83 � 0. 06 64 0. 84 81 (0 .6 83 8) (0 .2 74 1) (� 2. 16 18 )* * (3 5. 96 51 )* ** (� 4. 14 05 )* ** (3 .5 48 8) ** * (� 0. 08 11 ) (� 0. 24 21 ) (1 44 .0 29 )* ** t w o qu ar te rs 2. 33 31 1. 52 74 � 0. 58 80 0. 96 53 � 0. 10 60 10 .0 04 4 � 0. 03 08 � 0. 06 23 0. 84 81 236 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 6 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) (0 .7 68 6) (0 .2 89 2) (� 2. 19 23 )* * (3 5. 95 58 )* ** (� 4. 15 15 )* ** (3 .5 23 6) ** * (� 0. 08 86 ) (� 0. 22 73 ) (1 44 .0 40 )* ** t hr ee qu ar te rs 2. 94 15 1. 78 70 � 0. 60 16 0. 96 54 � 0. 10 67 9. 92 64 � 0. 03 59 � 0. 06 10 0. 84 81 (0 .9 68 3) (0 .3 37 5) (� 2. 24 12 )* * (3 5. 96 94 )* ** (� 4. 17 75 )* ** (3 .4 94 0) ** * (� 0. 10 31 ) (� 0. 22 25 ) (1 44 .1 02 )* ** a us tr al ia n fix ed in te re st o ne qu ar te r � 11 .5 88 6 � 9. 87 98 � 0. 09 64 0. 33 81 0. 04 65 22 .2 04 8 � 0. 79 01 0. 81 75 0. 30 64 (� 1. 14 11 ) (� 0. 62 19 ) (� 0. 39 83 ) (1 4. 95 70 )* ** (1 .8 02 0) * (7 .7 63 2) ** * (� 2. 18 58 )* * (2 .5 50 3) ** (1 3. 42 47 )* ** t w o qu ar te rs � 21 .3 82 7 � 8. 13 48 � 0. 08 85 0. 33 72 0. 04 76 22 .4 00 8 � 0. 78 62 0. 79 44 0. 30 65 (� 2. 16 08 ) (� 0. 51 59 ) (� 0. 36 48 ) (1 4. 89 49 )* ** (1 .8 38 8) * (7 .8 01 2) ** * (� 2. 17 55 )* * (2 .4 86 2) ** (1 3. 42 88 )* ** t hr ee qu ar te rs � 14 .1 09 5 � 12 .8 55 2 � 0. 09 68 0. 33 61 0. 04 76 22 .0 64 6 � 0. 80 13 0. 83 23 0. 30 53 (� 1. 41 32 ) (� 0. 81 57 ) (� 0. 39 82 ) (1 4. 82 39 )* ** (1 .8 37 3) v* (7 .6 90 4) ** * (� 2. 21 46 )* * (2 .6 05 6) ** (1 3. 35 92 )* ** m an ag ed ba la nc ed o ne qu ar te r 2. 04 80 � 0. 18 66 � 1. 00 40 0. 59 57 0. 15 98 29 .2 98 6 0. 15 59 0. 10 68 0. 67 49 (0 .2 67 3) (� 0. 10 22 ) (� 3. 51 23 )* ** (3 2. 47 61 )* ** (8 .9 34 9) ** * (7 .2 91 4) ** * (0 .5 34 2) (0 .3 46 5) (7 2. 36 90 )* ** t w o qu ar te rs � 0. 98 72 � 0. 20 78 � 1. 01 04 0. 59 53 0. 16 02 29 .1 26 7 0. 14 03 0. 10 91 0. 67 48 (� 0. 12 96 ) (� 0. 11 38 ) (� 3. 53 62 )* ** (3 2. 46 65 )* ** (8 .9 60 7) ** * (7 .2 38 3) ** * (0 .4 81 8) (0 .3 54 0) (7 2. 34 90 )* ** t hr ee qu ar te rs � 0. 98 22 � 0. 20 36 � 1. 01 00 0. 59 51 0. 16 04 29 .1 70 7 0. 14 35 0. 10 73 0. 67 48 (� 0. 12 94 ) (� 0. 11 15 ) (� 3. 53 48 )* ** (3 2. 46 49 )* ** (8 .9 75 9) ** * (7 .2 36 6) ** * (0 .4 93 5) (0 .3 48 3) (7 2. 34 90 )* ** m an ag ed gr ow th o ne qu ar te r 0. 86 16 0. 09 02 � 0. 61 75 0. 67 81 0. 05 03 21 .1 80 1 0. 05 99 � 0. 10 63 0. 66 74 (0 .8 03 8) (0 .1 15 3) (� 3. 36 59 )* ** (5 4. 20 23 )* ** (4 .0 69 1) ** * (9 .4 30 6) ** * (0 .3 48 6) (� 0. 57 62 ) (7 5. 74 57 )* ** t w o qu ar te rs 1. 94 57 � 0. 31 90 � 0. 60 87 0. 67 79 0. 05 05 21 .3 45 2 0. 06 49 � 0. 10 18 0. 66 74 (1 .6 60 9) (� 0. 34 55 ) (� 3. 33 68 )* ** (5 4. 18 13 )* ** (4 .0 82 4) ** * (9 .6 61 7) ** * (0 .3 78 2) (� 0. 55 21 ) (7 5. 75 45 )* ** t hr ee qu ar te rs 0. 02 75 0. 45 57 � 0. 60 69 0. 67 84 0. 05 03 21 .3 21 1 0. 05 71 � 0. 10 39 0. 66 73 (0 .0 24 0) (0 .4 29 6) (� 3. 32 59 )* ** (5 4. 21 42 )* ** (4 .0 68 9) ** * (9 .6 48 6) ** * (0 .3 33 1) (� 0. 56 34 ) (7 5. 70 52 )* ** m or tg ag e o ne qu ar te r 14 .5 86 8 12 .7 18 9 � 2. 23 15 0. 63 78 0. 09 22 53 .6 78 4 � 4. 63 94 � 2. 05 35 0. 65 99 (0 .3 45 6) (0 .1 72 7) (� 2. 50 30 )* * (2 2. 93 58 )* ** (3 .4 81 6) ** * (4 .7 63 3) ** * (� 3. 82 66 )* ** (� 1. 66 99 )* (5 6. 51 42 )* ** t w o qu ar te rs 30 .5 21 3 10 .0 03 0 � 2. 25 57 0. 63 79 0. 09 23 53 .5 61 7 � 4. 65 85 � 2. 02 96 0. 65 99 (0 .7 33 9) (0 .1 35 2) (� 2. 52 52 )* * (2 2. 94 35 )* ** (3 .4 85 7) ** * (4 .7 50 8) ** * (� 3. 84 07 )* ** (� 1. 64 80 )* (5 6. 52 19 )* ** t hr ee qu ar te rs 26 .9 53 2 12 .9 82 6 � 2. 24 27 0. 63 77 0. 09 20 53 .8 53 7 � 4. 68 66 � 2. 02 23 0. 65 99 (0 .6 58 1) (0 .1 75 4) (� 2. 51 26 )* * (2 2. 93 53 )* ** (3 .4 75 8) ** * (4 .7 81 1) ** * (� 3. 85 92 )* ** (� 1. 64 12 ) (5 6. 51 49 )* ** m an ag ed st ab le o ne qu ar te r 16 .9 70 9 � 0. 61 71 � 0. 95 77 0. 55 85 0. 12 79 22 .0 16 9 � 0. 60 52 0. 06 82 0. 65 57 (3 .3 58 0) ** * (� 0. 29 89 ) (� 6. 13 49 )* ** (3 1. 51 69 )* ** (7 .6 63 2) ** * (1 1. 88 57 )* ** (� 3. 38 85 )* ** (0 .4 17 7) (6 2. 70 63 )* ** t w o qu ar te rs 6. 70 24 � 1. 14 74 � 0. 95 87 0. 55 96 0. 12 61 21 .5 43 5 � 0. 63 00 0. 08 79 0. 65 47 (1 .6 72 3) (� 0. 48 12 ) (� 6. 13 33 )* ** (3 1. 54 01 )* ** (7 .5 53 2) ** * (1 1. 66 23 )* ** (� 3. 52 63 )* ** (0 .5 37 4) (6 2. 45 48 )* ** t hr ee qu ar te rs � 0. 64 99 � 1. 01 55 � 0. 94 55 0. 56 17 0. 12 51 21 .4 61 9 � 0. 64 06 0. 09 97 0. 65 43 (� 0. 22 78 ) (� 0. 34 37 ) (� 6. 05 01 )* ** (3 1. 69 94 )* ** (7 .4 92 1) ** * (1 1. 61 23 )* ** (� 3. 58 44 )* ** (0 .6 08 9) (6 2. 33 32 )* ** 237r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 6 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) o ve rs ea s: fi xe d in co m e o ne qu ar te r � 1. 59 83 4. 57 86 � 0. 00 63 0. 43 33 � 0. 03 58 8. 38 65 0. 13 56 0. 03 58 0. 34 41 (� 0. 49 09 ) (1 .1 92 1) (� 0. 03 60 ) (6 .3 36 8) ** * (� 0. 56 09 ) (5 .1 66 3) ** * (0 .5 25 4) (0 .1 57 3) (8 .6 45 1) ** * t w o qu ar te rs � 2. 66 95 5. 21 43 0. 00 86 0. 44 01 � 0. 02 39 8. 45 39 0. 12 54 0. 00 93 0. 34 73 (� 1. 02 59 ) (1 .3 48 2) (0 .0 49 2) (6 .5 50 6) ** * (� 0. 37 62 ) (5 .2 38 8) ** * (0 .4 87 0) (0 .0 41 6) (8 .7 56 5) ** * t hr ee qu ar te rs � 2. 17 22 5. 09 98 0. 00 50 0. 44 13 � 0. 02 46 8. 44 57 0. 12 68 0. 00 98 0. 34 74 (� 0. 83 54 ) (1 .3 29 4) (0 .0 28 7) (6 .5 73 7) ** * (� 0. 38 76 ) (5 .2 35 8) ** * (0 .4 92 1) (0 .0 43 7) (8 .7 57 4) ** * o ve rs ea s: g lo ba l o ne qu ar te r 2. 00 94 2. 82 51 � 0. 38 70 0. 69 05 0. 08 51 13 .8 62 6 � 0. 07 75 � 0. 14 89 0. 62 48 (0 .5 59 6) (0 .6 58 6) (� 2. 05 71 )* * (4 2. 51 20 )* ** (5 .2 01 0) ** * (5 .2 17 7) ** * (� 0. 24 74 ) (� 0. 51 53 ) (5 1. 33 49 )* ** t w o qu ar te rs 0. 61 10 2. 72 78 � 0. 39 25 0. 69 07 0. 08 49 13 .7 07 2 � 0. 07 80 � 0. 14 31 0. 62 47 (0 .1 80 8) (0 .6 35 9) (� 2. 08 62 )* * (4 2. 52 18 )* ** (5 .1 83 3) ** * (5 .1 43 4) ** * (� 0. 24 91 ) (� 0. 49 52 ) (5 1. 31 97 )* ** t hr ee qu ar te rs � 2. 53 28 2. 56 14 � 0. 40 97 0. 69 06 0. 08 58 14 .0 50 5 � 0. 08 35 � 0. 14 89 0. 62 50 (� 0. 77 47 ) (0 .5 97 3) (� 2. 17 53 )* * (4 2. 55 41 )* ** (5 .2 35 7) ** * (5 .2 66 2) ** * (� 0. 26 69 ) (� 0. 51 53 ) (5 1. 37 70 )* ** o ve rs ea s: pr op er ty o ne qu ar te r � 6. 44 12 � 5. 10 92 � 0. 62 17 � 0. 01 50 0. 30 12 1. 73 60 � 0. 70 40 0. 60 47 0. 24 74 (� 0. 26 81 ) (� 0. 29 42 ) (� 0. 38 23 ) (� 0. 17 66 ) (3 .8 28 5) ** * (0 .5 48 3) (� 0. 12 84 ) (0 .2 22 6) (3 .6 14 7) ** * t w o qu ar te rs � 2. 04 78 � 5. 56 21 � 0. 68 57 � 0. 01 41 0. 30 14 1. 71 16 � 0. 40 99 0. 53 11 0. 24 71 (� 0. 12 10 ) (� 0. 31 94 ) (� 0. 42 46 ) (� 0. 16 63 ) (3 .8 29 6) ** * (0 .5 40 7) (� 0. 07 62 ) (0 .1 96 1) (3 .6 10 4) ** * t hr ee qu ar te rs � 0. 57 92 � 5. 68 66 � 0. 76 87 � 0. 01 41 0. 30 17 1. 71 28 � 0. 21 93 0. 45 68 0. 37 84 (� 0. 04 88 ) (� 0. 32 76 ) (� 0. 47 35 ) (� 0. 16 64 ) (3 .8 34 2) ** * (0 .5 41 3) (� 0. 04 08 ) (0 .1 68 9) (3 .6 15 7) ** * a us tr al ia n pr op er ty o ne qu ar te r 2. 69 63 0. 03 66 � 0. 47 23 0. 69 58 0. 12 05 20 .6 93 2 � 1. 58 78 0. 78 25 0. 75 54 (0 .5 05 2) (0 .0 06 8) (� 1. 12 62 ) (2 2. 11 95 )* ** (4 .0 93 8) ** * (4 .0 29 0) ** * (� 2. 76 08 )* ** (1 .5 10 6) (7 1. 59 21 )* ** t w o qu ar te rs 3. 59 49 � 1. 27 13 � 0. 50 22 0. 69 47 0. 12 04 19 .5 83 2 � 1. 60 31 0. 91 44 0. 75 60 (0 .9 35 8) (� 0. 23 29 ) (� 1. 19 73 ) (2 2. 15 81 )* ** (4 .1 00 5) ** * (3 .9 56 4) ** * (� 2. 80 07 )* ** (1 .8 10 4) (7 1. 81 13 )* ** t hr ee qu ar te rs 1. 37 17 � 1. 33 61 � 0. 48 26 0. 69 61 0. 11 99 19 .4 65 1 � 1. 56 42 0. 84 08 0. 75 60 (0 .2 80 6) (� 0. 24 50 ) (� 1. 14 75 ) (2 2. 18 33 )* ** (4 .0 82 2) ** * (3 .9 31 4) ** * (� 2. 73 82 )* ** (1 .6 75 8) (7 1. 83 79 )* ** a us tr al ia n pr op er ty se cu ri tie s o ne qu ar te r � 3. 84 47 � 0. 95 94 � 0. 17 24 0. 48 39 0. 03 75 20 .8 62 3 � 1. 12 17 0. 00 22 0. 39 89 (� 1. 11 46 ) ( � 0. 20 32 ) (� 0. 72 25 ) (2 1. 54 38 )* ** (1 .5 85 8) (7 .9 90 3) ** * (� 4. 50 07 )* ** (0 .0 06 4) (2 1. 60 98 )* ** t w o qu ar te rs � 0. 71 47 � 0. 94 24 � 0. 18 55 0. 48 49 0. 03 66 20 .8 95 8 � 1. 11 58 � 0. 01 70 0. 39 92 (� 0. 21 47 ) (� 0. 19 97 ) (� 0. 77 67 ) (2 1. 57 66 )* ** (1 .5 48 9) (8 .0 10 0) ** * (� 4. 47 74 )* ** (� 0. 04 84 ) (2 1. 63 64 )* ** t hr ee qu ar te rs � 2. 77 84 � 1. 33 59 � 0. 17 22 0. 48 46 0. 03 81 20 .5 73 3 � 1. 10 68 � 0. 00 89 0. 39 90 (� 0. 66 50 ) (� 0. 28 29 ) (� 0. 71 79 ) (2 1. 56 14 )* ** (1 .6 13 5) (7 .8 99 6) ** * (� 4. 44 33 )* ** (� 0. 02 54 ) (2 1. 61 66 )* ** m ix ed po rt fo lio s o ne qu ar te r 7. 60 36 � 1. 34 28 � 22 .5 61 9 0. 21 77 0. 09 49 49 5. 04 57 31 .8 19 4 � 35 .0 12 5 0. 47 76 (1 .1 16 4) (� 0. 77 52 ) (� 2. 91 30 )* ** (6 .6 79 2) ** * (2 .3 69 7) ** (4 .8 74 4) ** * (3 .0 00 3) ** * (� 4. 12 59 )* ** (2 2. 25 40 )* ** t w o qu ar te rs � 0. 67 22 � 0. 99 43 � 22 .6 67 1 0. 21 69 0. 09 28 50 5. 07 39 31 .9 75 0 � 35 .4 03 8 0. 47 69 (� 0. 24 08 ) (� 0. 47 48 ) (� 2. 92 48 )* ** (6 .6 52 8) ** * (2 .3 17 6) ** (4 .9 88 3) ** * (3 .0 13 2) ** * (� 4. 17 23 )* ** (2 2. 19 50 )* ** t hr ee qu ar te rs � 1. 04 80 � 0. 73 20 � 22 .6 76 9 0. 21 69 0. 09 28 50 5. 53 84 31 .9 47 5 � 35 .3 93 6 0. 47 69 (� 0. 23 63 ) (� 0. 24 99 ) (� 2. 92 59 )* ** (6 .6 53 2) ** * (2 .3 17 9) ** (4 .9 90 2) ** * (3 .0 12 3) ** * (� 4. 17 31 )* ** (2 2. 19 54 )* ** ** * si gn ifi ca nt at 1% , ** si gn ifi ca nt at 5% . 238 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 7 fi xe d ef fe ct s pa ne l re gr es si on of w ho le sa le m an ag ed fu nd ne t flo w s w ith pa st ex ce ss re tu rn s an d ri sk r eg re ss io n eq ua tio n: n f i, t � b j r j t, t � c� i, t� 1 � ds iz e i ,t � 1 � en f i, t� 1 � fn f i, t� 2 � ge in i, t � hc h t � jg f c t � � j� � 1 1 k i ,j � x i ,t � � t w he re n f i, t ar e th e ne t flo w in to th e fu nd , r j i, is th e tr ai lin g ex ce ss re tu rn s fo r ei th er 1, 2, or 3 qu ar te rs , � i, t is th e tr ac ki ng er ro r, si ze i, t, is th e lo g of th e fu nd si ze , c h t is th e ch oi ce du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 05 an d 1 fo r re st of th e tim e pe ri od , g f c t is th e cr is is du m m y, w hi ch ha s a va lu e of 0, un til th ir d qu ar te r of 20 08 an d 1 fo r th e re st of th e tim e pe ri od , e in t is a va ri ab le to ca pt ur e th e ov er al l flo w in to th e m ar ke t, an d f e i, t is th e fix ed ef fe ct of ith fu nd . � x i ar e th e fir st di ff er en ce te rm s of th e re gr es si on va ri ab le s. w e es tim at ed th re e se pa ra te re gr es si on s w ith th e va ri ab le r j i, t be in g 1, 2, or 3 qu ar te r ex ce ss re tu rn s. fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) a lte rn at iv es o ne qu ar te r 35 .7 29 1 23 .8 42 0 � 3. 97 46 0. 24 38 0. 23 51 15 3. 88 83 � 1. 99 64 1. 61 50 0. 52 20 (1 .1 98 2) (0 .4 29 3) (� 2. 10 16 )* * (4 .5 51 9) ** * (4 .7 25 8) ** * (5 .7 06 4) ** * (� 0. 40 93 ) (0 .4 25 0) (1 3. 89 59 )* ** t w o qu ar te rs 24 .8 60 7 25 .6 08 4 � 3. 97 04 0. 24 60 0. 23 32 14 8. 04 43 � 1. 76 89 1. 42 99 0. 52 11 (0 .9 03 6) (0 .4 59 1) (� 2. 09 00 )* * (4 .5 96 8) ** * (4 .6 89 6) ** * (5 .6 37 2) ** * (� 0. 36 32 ) (0 .3 76 8) (1 3. 85 18 )* ** t hr ee qu ar te rs 9. 68 59 29 .8 05 6 � 3. 86 77 0. 24 82 0. 23 22 14 6. 99 91 � 1. 68 30 1. 38 26 0. 52 07 (0 .2 83 2) (0 .5 35 4) (� 2. 03 24 )* * (4 .6 44 5) ** * (4 .6 67 3) ** * (5 .6 14 4) ** * (� 0. 34 53 ) (0 .3 64 3) (1 3. 83 20 )* ** c as h o ne qu ar te r � 17 .8 62 0 � 11 .5 23 0 � 3. 67 36 0. 11 71 0. 02 94 19 0. 81 62 � 3. 62 77 0. 97 82 0. 54 87 (� 0. 37 98 ) (� 0. 18 56 ) (� 3. 54 09 )* ** (3 .0 79 6) ** * (0 .8 16 5) (9 .1 46 7) ** * (� 1. 26 04 ) (0 .3 87 1) (1 6. 25 55 )* ** t w o qu ar te rs 2. 69 41 � 9. 53 99 � 3. 63 96 0. 11 83 0. 02 95 19 0. 99 48 � 3. 48 11 0. 86 92 0. 54 84 (0 .0 59 8) (� 0. 15 30 ) (� 3. 51 14 )* ** (3 .1 16 1) ** * (0 .8 19 0) (9 .1 51 0) ** * (� 1. 21 21 ) (0 .3 43 1) (1 6. 23 69 )* ** t hr ee qu ar te rs � 2. 85 12 � 9. 37 48 � 3. 64 08 0. 11 83 0. 02 96 19 1. 06 99 � 3. 48 70 0. 85 89 0. 54 84 (� 0. 05 14 ) (� 0. 15 03 ) (� 3. 51 22 )* ** (3 .1 14 3) ** * (0 .8 20 2) (9 .1 51 8) ** * (� 1. 21 48 ) (0 .3 38 6) (1 6. 23 71 )* ** d iv er si fie d fix ed in te re st o ne qu ar te r 6. 56 07 � 11 .5 43 2 0. 21 55 0. 73 21 � 0. 07 53 37 .5 58 3 � 2. 53 49 1. 52 63 0. 60 86 (0 .1 86 6) (� 0. 29 77 ) (0 .1 84 4) (1 5. 73 03 )* ** (� 1. 68 06 ) (2 .2 41 5) ** (� 0. 90 40 ) (0 .6 39 9) (2 3. 41 55 )* ** t w o qu ar te rs 10 .0 92 0 � 11 .9 78 2 0. 19 83 0. 73 22 � 0. 07 67 37 .7 85 2 � 2. 52 38 1. 46 16 0. 60 87 (0 .2 99 1) (� 0. 30 89 ) (0 .1 69 7) (1 5. 74 34 )* ** (� 1. 71 80 ) (2 .2 60 6) ** (� 0. 90 02 ) (0 .6 11 9) (2 3. 42 83 )* ** t hr ee qu ar te rs 19 .4 46 1 � 15 .3 91 0 0. 17 74 0. 73 19 � 0. 07 66 37 .5 33 5 � 2. 51 72 1. 47 67 0. 60 87 (0 .4 72 7) (� 0. 37 67 ) (0 .1 52 0) (1 5. 73 19 )* ** (� 1. 71 66 ) (2 .2 49 1) ** (� 0. 89 79 ) (0 .6 18 3) (2 3. 43 05 )* ** a us tr al ia n eq ui ty o ne qu ar te r � 0. 19 17 56 .5 79 9 � 0. 86 82 0. 32 73 0. 20 13 21 9. 03 45 � 2. 93 02 0. 68 03 0. 40 94 (� 0. 00 67 ) (1 .5 61 5) (� 1. 10 25 ) (2 0. 60 11 )* ** (1 1. 99 67 )* ** (1 5. 24 49 )* ** (� 1. 65 68 ) (0 .4 37 8) (2 0. 93 17 )* ** t w o qu ar te rs 50 .7 29 0 55 .7 85 8 � 0. 80 98 0. 32 72 0. 20 28 22 1. 15 55 � 3. 04 62 0. 82 39 ) 0. 40 92 (1 .8 41 7) * (1 .5 39 5) (� 1. 02 83 ) (2 0. 58 30 )* ** (1 2. 07 97 )* ** (1 5. 47 91 )* ** (� 1. 72 25 ) (0 .5 30 1 (2 0. 91 51 )* ** t hr ee qu ar te rs 53 .3 19 9 54 .2 18 2 � 0. 81 60 0. 32 70 0. 20 25 22 0. 88 92 � 3. 06 98 0. 83 45 ) 0. 40 91 (1 .6 45 8) * (1 .4 87 1) (� 1. 03 62 ) (2 0. 57 08 )* ** (1 2. 06 38 )* ** (1 5. 44 76 )* ** (� 1. 73 54 )* (0 .5 36 5) (2 0. 90 82 )* ** a us tr al ia n sm al l co m pa ny o ne qu ar te r � 7. 72 48 0. 44 60 � 2. 02 07 0. 31 58 0. 09 86 72 .5 26 4 � 0. 29 08 1. 87 79 0. 50 73 (� 0. 88 32 ) (0 .2 76 3) (� 3. 23 41 )* ** (9 .3 48 9) ** * (3 .0 75 0) ** * (9 .2 49 3) ** * (� 0. 25 35 ) (1 .9 71 6) ** (2 2. 57 34 )* ** t w o qu ar te rs � 0. 10 91 0. 41 87 � 1. 97 87 0. 31 76 0. 09 78 73 .1 95 0 � 0. 40 53 1. 93 40 0. 50 68 (� 0. 04 41 ) (0 .2 16 1) (� 3. 17 47 )* ** (9 .4 11 2) ** * (3 .0 48 6) ** * (9 .3 73 4) ** * (� 0. 35 53 ) (2 .0 34 1) ** (2 2. 53 39 )* ** t hr ee qu ar te rs 0. 99 19 0. 99 26 � 1. 97 59 0. 31 77 0. 09 79 73 .1 01 8 � 0. 41 42 1. 93 99 0. 50 68 (0 .2 60 2) (0 .3 78 2) (� 3. 17 32 )* ** (9 .4 15 3) ** * (3 .0 51 6) ** * (9 .3 59 0) ** * (� 0. 36 30 ) (2 .0 42 3) ** (2 2. 53 81 )* ** 239r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 7 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) a us tr al ia n fix ed in te re st o ne qu ar te r 0. 65 42 � 2. 45 26 � 0. 49 16 0. 23 26 0. 03 69 17 1. 13 05 � 3. 60 20 2. 96 01 0. 44 41 (0 .2 63 5) (� 1. 30 39 ) (� 0. 56 60 ) (8 .4 47 1) ** * (1 .2 58 8) (1 1. 86 45 )* ** (� 1. 73 11 )* (1 .6 16 0) (1 7. 85 45 )* ** t w o qu ar te rs 1. 40 48 � 1. 76 53 � 0. 48 54 0. 23 29 0. 03 68 17 0. 92 60 � 3. 60 81 2. 94 36 0. 44 42 (0 .4 69 6) (� 0. 78 04 ) (� 0. 55 91 ) (8 .4 58 2) ** * (1 .2 55 8) (1 1. 85 97 )* ** (� 1. 73 54 )* (1 .6 07 2) (1 7. 86 04 )* ** t hr ee qu ar te rs 0. 17 11 � 2. 40 77 � 0. 48 29 0. 23 30 0. 03 68 17 0. 86 35 � 3. 62 49 2. 97 39 0. 44 42 (0 .0 36 1) (� 0. 76 62 ) (� 0. 55 62 ) (8 .4 62 8) ** * (1 .2 57 2) (1 1. 85 64 )* ** (� 1. 74 30 )* (1 .6 20 8) (1 7. 86 26 )* ** m an ag ed ba la nc ed o ne qu ar te r 17 .4 95 1 � 10 2. 63 28 � 1. 81 78 0. 39 45 0. 17 85 12 1. 58 46 1. 92 40 1. 80 81 0. 52 61 (0 .4 93 5) (� 1. 99 66 )* * (� 2. 35 13 )* * (1 6. 90 50 )* ** (8 .0 77 4) ** * (9 .5 45 3) ** * (1 .4 29 1) (1 .4 58 1) (3 3. 66 33 )* ** t w o qu ar te rs 62 .8 74 8 � 10 9. 27 06 � 1. 86 72 0. 39 41 0. 17 83 12 1. 03 53 1. 84 42 1. 87 12 0. 52 73 (1 .7 63 9) (� 2. 13 00 )* * (� 2. 41 73 )* * (1 6. 92 23 )* ** (8 .0 87 4) ** * (9 .5 27 4) ** * (1 .3 71 2) (1 .5 10 7) (3 3. 82 20 )* ** t hr ee qu ar te rs 67 .9 14 2 � 10 8. 51 75 � 1. 86 55 0. 39 38 0. 17 87 12 1. 05 23 1. 85 00 1. 86 45 0. 52 72 (1 .5 58 3) (� 2. 11 57 )* * (� 2. 41 36 )* * (1 6. 89 35 )* ** (8 .1 03 3) ** * (9 .5 26 7) ** * (1 .3 74 5) (1 .5 04 9) (3 3. 80 80 )* ** m an ag ed gr ow th o ne qu ar te r 21 .5 70 2 � 9. 31 67 � 0. 16 30 0. 28 08 0. 24 21 10 9. 59 69 3. 10 36 2. 77 71 0. 57 80 (1 .0 91 2) (� 0. 42 88 ) (� 0. 31 85 ) (1 4. 04 78 )* ** (1 2. 53 32 )* ** (1 0. 86 35 )* ** (2 .8 59 3) ** * (2 .8 39 6) ** * (4 1. 66 22 )* ** t w o qu ar te rs 29 .7 46 7 � 4. 50 59 � 0. 16 48 0. 28 06 0. 24 19 10 9. 46 50 3. 09 66 2. 79 55 0. 57 82 (1 .6 70 3) (� 0. 20 16 ) (� 0. 32 23 ) (1 4. 04 10 )* ** (1 2. 53 12 )* ** (1 0. 85 12 )* ** (2 .8 54 9) ** * (2 .8 58 9) ** * (4 1. 68 76 )* ** t hr ee qu ar te rs 19 .5 85 4 � 1. 91 88 � 0. 16 82 0. 28 07 0. 24 17 10 9. 65 67 3. 09 13 2. 76 02 0. 57 81 (0 .9 09 9) (� 0. 08 47 ) (� 0. 32 88 ) (1 4. 03 99 )* ** (1 2. 52 10 )* ** (1 0. 85 98 )* ** (2 .8 49 6) ** * (2 .8 22 3) ** * (4 1. 66 31 )* ** m or tg ag e o ne qu ar te r 15 .0 03 1 � 16 5. 34 49 � 2. 78 10 0. 55 64 � 0. 00 75 11 5. 01 36 � 2. 91 12 � 12 .6 53 3 0. 63 78 (0 .0 61 0) (� 0. 56 49 ) (� 1. 78 77 ) (1 1. 33 60 )* ** (� 0. 18 86 ) (4 .9 63 6) ** * (� 1. 11 83 ) (� 4. 78 18 )* ** (3 0. 94 83 )* ** t w o qu ar te rs � 18 6. 98 21 � 15 5. 51 66 � 2. 85 65 0. 55 52 � 0. 00 76 11 4. 73 43 � 3. 05 39 � 12 .5 71 7 0. 63 90 (� 0. 76 44 ) (� 0. 53 54 ) (� 1. 84 44 ) (1 1. 33 04 )* ** (� 0. 19 37 ) (4 .9 61 3) ** * (� 1. 17 56 ) (� 4. 76 19 )* ** (3 1. 09 62 )* ** t hr ee qu ar te rs � 18 3. 97 94 � 15 5. 01 41 � 2. 82 18 0. 55 68 � 0. 00 82 11 4. 45 35 � 3. 02 24 � 12 .6 03 5 0. 63 87 (� 0. 66 08 ) (� 0. 51 62 ) (� 1. 81 65 ) (1 1. 36 01 )* ** (� 0. 20 88 ) (4 .9 48 1) ** * (� 1. 16 32 ) (� 4. 77 07 )* ** (3 1. 05 71 ) m an ag ed st ab le o ne qu ar te r 36 .1 69 5 � 9. 96 24 0. 12 00 0. 36 45 0. 20 64 44 .1 69 0 0. 09 99 0. 84 55 0. 52 28 (2 .2 71 0) ** (� 0. 42 50 ) (0 .6 04 4) (1 5. 55 71 )* ** (8 .9 00 2) ** * (1 1. 83 84 )* ** (0 .2 40 8) (2 .1 64 8) ** (3 1. 26 14 )* ** t w o qu ar te rs 51 .4 63 7 � 0. 75 34 0. 14 16 0. 36 05 0. 20 78 44 .1 57 0 0. 09 74 0. 83 12 0. 52 42 (3 .2 72 9) ** * (� 0. 03 19 ) (0 .7 13 4) (1 5. 38 09 )* ** (8 .9 75 7) ** * (1 1. 85 32 )* ** (0 .2 35 0) (2 .1 31 1) ** (3 1. 42 90 )* ** t hr ee qu ar te rs 47 .7 31 0 � 0. 58 24 0. 14 15 0. 35 99 0. 20 79 44 .2 00 1 0. 09 43 0. 83 26 0. 52 40 (2 .4 97 2) ** (� 0. 02 44 ) (0 .7 12 0) (1 5. 34 50 )* ** (8 .9 78 1) ** * (1 1. 86 35 )* ** (0 .2 27 5) (2 .1 33 9) ** (3 1. 40 54 )* ** o ve rs ea s: a si a pa ci fic o ne qu ar te r 4. 00 34 � 0. 03 16 � 3. 44 21 0. 71 54 � 0. 20 31 49 .8 57 4 3. 32 68 � 3. 21 15 0. 43 56 (0 .1 16 7) (� 0. 00 08 ) (� 1. 53 03 ) (1 3. 08 42 )* ** (� 3. 76 46 )* ** (2 .0 53 1) ** (0 .7 71 2) (� 0. 98 05 ) (1 2. 43 42 )* ** t w o qu ar te rs � 10 .2 66 8 � 10 .5 14 0 � 3. 48 06 0. 71 70 � 0. 20 40 52 .4 62 4 3. 30 14 � 3. 21 93 0. 43 68 (� 0. 37 43 ) (� 0. 25 37 ) (� 1. 55 31 ) (1 3. 13 69 )* ** (� 3. 78 39 )* ** (2 .1 99 3) ** (0 .7 67 9) (� 0. 98 73 ) (1 2. 49 34 )* ** t hr ee qu ar te rs � 15 .6 32 5 � 8. 48 92 � 3. 53 47 0. 71 78 � 0. 20 55 51 .4 15 6 3. 29 63 � 3. 24 34 0. 43 60 (� 0. 44 93 ) (� 0. 19 89 ) (� 1. 57 61 ) (1 3. 13 16 )* ** (� 3. 80 80 )* ** (2 .1 54 3) ** (0 .7 65 8) (� 0. 99 35 ) (1 2. 45 38 )* ** 240 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 t ab le 7 (c on tin ue d) fu nd ca te go ry b j (t -s ta t) c (t -s ta t) d (t -s ta t) e (t -s ta t) f (t -s ta t) g (t -s ta t) h (t -s ta t) j (t -s ta t) a dj . r 2 (f -s ta t) o ve rs ea s: fi xe d in co m e o ne qu ar te r � 17 3. 39 13 46 .4 37 4 1. 04 61 0. 48 32 � 0. 20 08 13 8. 03 40 � 6. 15 60 7. 48 97 0. 34 50 (� 2. 12 62 )* ** (0 .4 26 4) (0 .5 44 8) (1 2. 65 43 )* ** (� 5. 38 15 )* ** (5 .1 59 0) ** * (� 1. 34 58 ) (1 .9 10 0) (1 0. 52 15 )* ** t w o qu ar te rs � 11 4. 15 42 52 .3 96 5 1. 05 71 0. 48 63 � 0. 20 12 13 9. 98 70 � 6. 22 27 7. 40 56 0. 34 49 (� 1. 40 76 ) (0 .4 76 0) (0 .5 50 2) (1 2. 76 12 )* ** (� 5. 39 09 )* ** (5 .2 72 3) ** * (� 1. 36 12 ) (1 .8 84 7) (1 0. 51 63 )* ** t hr ee qu ar te rs � 51 .9 15 6 52 .2 68 9 0. 66 86 0. 48 75 � 0. 20 03 14 1. 79 57 � 6. 70 35 7. 42 78 0. 33 99 (� 0. 52 27 ) (0 .4 67 8) (0 .3 48 0) (1 2. 74 50 )* ** (� 5. 33 40 )* ** (5 .3 22 8) ** * (� 1. 46 21 ) (1 .8 78 0) (1 0. 31 02 )* ** o ve rs ea s: g lo ba l o ne qu ar te r � 0. 83 66 0. 03 60 � 0. 99 87 0. 44 71 0. 20 74 19 4. 36 97 � 2. 72 64 � 1. 37 75 0. 49 11 (� 0. 18 79 ) (0 .0 59 6) (� 0. 95 89 ) (2 5. 39 02 )* ** (1 1. 15 76 )* ** (1 1. 32 13 )* ** (� 1. 06 97 ) (� 0. 63 43 ) (2 5. 10 54 )* ** t w o qu ar te rs 0. 16 11 0. 10 95 � 0. 99 88 0. 44 71 0. 20 74 19 4. 63 64 � 2. 73 88 � 1. 37 00 0. 49 11 (0 .1 71 2) (0 .1 52 2) (� 0. 95 91 ) (2 5. 39 25 )* ** (1 1. 15 73 )* ** (1 1. 36 84 )* ** (� 1. 07 47 ) (� 0. 63 10 ) (2 5. 10 53 )* ** t hr ee qu ar te rs 0. 13 32 0. 09 92 � 0. 99 91 0. 44 71 0. 20 74 19 4. 63 73 � 2. 73 66 � 1. 37 06 0. 49 11 (0 .0 92 2) (0 .1 03 0) (� 0. 95 92 ) (2 5. 39 21 )* ** (1 1. 15 73 )* ** (1 1. 36 85 )* ** (� 1. 07 40 ) (� 0. 63 12 ) (2 5. 10 53 )* ** o ve rs ea s: pr op er ty o ne qu ar te r 23 .3 79 7 33 .1 71 8 � 5. 02 60 0. 19 15 0. 21 44 77 .4 30 3 � 11 .3 35 6 � 7. 45 89 0. 31 44 (1 .0 29 1) (0 .7 25 5) (� 1. 62 48 ) (2 .8 06 5) ** * (3 .0 97 8) ** * (3 .5 02 1) ** * (� 0. 68 10 ) (� 1. 86 18 ) (5 .0 93 1) ** * t w o qu ar te rs 15 .6 96 3 30 .0 71 0 � 5. 15 85 0. 18 83 0. 21 44 77 .3 47 7 � 10 .7 99 6 � 7. 67 21 0. 31 27 (0 .7 01 3) (0 .6 34 9) (� 1. 65 31 ) (2 .7 55 9) ** * (3 .0 92 5) ** * (3 .4 58 5) ** * (� 0. 64 85 ) (� 1. 90 53 ) (5 .0 61 5) ** * t hr ee qu ar te rs 3. 49 41 22 .6 55 8 � 4. 91 97 0. 18 87 0. 21 66 76 .5 73 9 � 10 .6 83 2 � 7. 51 77 0. 31 27 (0 .1 19 0) (0 .4 79 0) (� 1. 55 59 ) (2 .7 61 5) ** * (3 .1 17 9) ** * (3 .4 17 3) ** * (� 0. 64 16 ) (� 1. 86 04 ) (5 .0 60 1) ** * a us tr al ia n pr op er ty o ne qu ar te r � 64 .7 03 4 89 .0 70 6 � 0. 23 67 0. 17 08 0. 10 07 20 6. 07 42 1. 46 58 � 10 .1 79 6 0. 25 66 (� 1. 01 68 ) (1 .1 20 7) (� 0. 15 41 ) (3 .9 68 4) ** * (1 .9 09 9) * (4 .5 55 8) ** * (0 .3 36 8) (� 2. 26 85 )* * (7 .9 63 1) ** * t w o qu ar te rs � 51 .1 63 9 93 .4 83 2) � 0. 15 54 0. 17 16 0. 10 36 21 8. 43 37 1. 61 70 � 11 .2 83 6 0. 25 67 (� 1. 09 26 ) (1 .1 71 4 (� 0. 10 09 ) (3 .9 92 9) ** * (1 .9 71 3) ** (5 .2 29 4) ** * (0 .3 72 1) (� 2. 59 67 )* ** (7 .9 69 1) ** * t hr ee qu ar te rs � 62 .8 82 3 94 .2 78 8 � 0. 10 65 0. 17 04 0. 10 35 21 8. 89 33 1. 68 02 � 11 .4 69 1 0. 25 68 (� 1. 05 05 ) (1 .1 80 8) (� 0. 06 89 ) (3 .9 59 2) ** * (1 .9 71 2) ** (5 .2 49 8) ** * (0 .3 86 7) (� 2. 64 08 )* ** (7 .9 73 7) ** * a us tr al ia n pr op er ty se cu ri tie s o ne qu ar te r 0. 63 09 0. 91 80 � 2. 44 94 0. 12 74 0. 17 53 14 7. 55 32 � 7. 06 56 � 2. 34 95 0. 31 70 (0 .0 31 4) (0 .5 44 1) (� 2. 17 20 )* * (4 .4 43 6) ** * (5 .2 49 2) ** * (1 0. 81 12 )* ** (� 3. 84 47 )* ** (� 1. 15 54 ) (1 2. 39 89 )* ** t w o qu ar te rs � 0. 62 42 (0 .8 99 4 � 2. 45 25 0. 12 74 0. 17 53 14 7. 58 00 � 7. 06 15 � 2. 35 86 0. 31 70 (� 0. 10 88 ) 0. 51 23 ) (� 2. 17 70 )* * (4 .4 42 9) ** * (5 .2 50 5) ** * (1 0. 81 47 )* ** (� 3. 84 46 )* ** (� 1. 16 24 ) (1 2. 39 88 )* ** t hr ee qu ar te rs � 0. 16 67 1. 00 97 � 2. 44 93 0. 12 74 0. 17 54 14 7. 63 51 � 7. 06 81 � 2. 34 83 0. 31 70 (� 0. 03 64 ) (0 .4 29 7) (� 2. 17 47 )* * (4 .4 42 5) ** * (5 .2 51 0) ** * (1 0. 82 06 )* ** (� 3. 84 85 )* ** (� 1. 15 82 ) (1 2. 39 82 )* ** ** * si gn ifi ca nt at 1% , ** si gn ifi ca nt at 5% . 241r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 positive effect on net inflows into those funds, with high excess returns over the index. for retail funds, there is evidence of such return chasing behavior for australian equity funds and managed stable funds. it is also possible that investors may try to move away from funds that are successful, on the assumption that such recent high performance is an aberration, and that the law of averages will eventually catch up. under this assumption, investors may actually withdraw funds and the sign for past excess returns will be negative. there is some evidence of this in the case of australian fixed interest funds, and australian property security funds. risk is another factor in investment decisions that ordinary investors are not fully capable of estimating. if risk is a factor in investment decisions then tracking error should have a negative effect on net investment flows. there is no evidence of this in the case of retail funds in all categories of assets as this variable is not significant. the average retail investor seems to be more influenced by the past performance of the fund rather than by the variability of the returns measured by the tracking error. a possible explanation can be found in the way the investors choose their funds using the fund ratings. unlike the u.s. mutual fund rating methodology, which until recently was completely based on quantitative assessment, the two major rating agencies in australia (morningstar and standard & poor’s) use a mixture of quantitative and qualitative factors (faff, parwada, and poh, 2007). because risk is one of the quantitative variables, its relative weight within the australian mutual fund rating system may be lower than in the united states. size had a negative effect on many of the asset categories. the momentum of past net flows, as measured by the lagged net flows, shows a statistically significant positive effect. as noted by frino, heaney, and service (2005), this relationship between the past net flows and current quarter net flows is rather difficult to explain. a possible explanation is that the positive relationship between the past net flows and the current period net flows can be because of the general growth in the particular asset class. gruber (1996) explained that this relationship may be because of the fact that investors are locked into a particular fund because of restricted choices allowed by their superannuation accounts, and this variable may also capture the effect of marketing and the reputation of the fund. the effect of choice of superannuation fund legislation on net fund flows gives an interesting perspective on how investors switched their asset allocation when they were given the freedom to do so. a significant part of the retail managed fund industry is comprised of superannuation funds, and thus, choice of fund legislation can have some impact on fund flows to different asset classes. the choice variable has a negative effect on alternative investments, australian equity, australian fixed interest, mortgage, and australian property. the asset categories with positive signs for this variable are managed balanced and mixed funds. these results are an indication that given the choice to determine their asset allocation by the choice of fund legislation, investors might have chosen to reduce their exposure to more risky assets and move their funds into managed funds. the effect of crisis variable is also equally interesting for the retail funds. the conventional wisdom is that during the crisis, investors will move their investments into safe assets and reduce exposure to risky assets, such as equities. of interest to the authors, the results are the just the opposite in the case of retail managed funds. the only two categories of funds with negative signs for the crisis variable are the low risk asset class of cash and mortgages. australian equity, 242 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 managed funds, and various overseas funds all had positive signs for the crisis variable. a possible explanation is that the dramatic drop in stock prices during the crisis gave an opportunity for investors to hunt for bargains as well as to seek the shelter that managed funds provide. the regression results for superannuation funds are given in table 5. the excess return variable has a positive sign for australian equities, mortgages, managed stable, overseas fixed interest, and currencies as well as for global equity funds. these results indicate that, in general, superannuation fund investors are similarly prone to return chasing compared with retail investors. the sign for tracking error is positive for overseas property funds and negative for mixed portfolio funds. none of these results indicate any strong relationship between the risk and funds flows. the size of the fund has a negative effect on the net fund flows of cash, diversified fixed interest, australian equity, australian small companies, managed balanced, overseas fixed income, and australian property securities. this is an indication that investors in these categories prefer larger funds. choice of fund legislation has a significant effect on almost all categories of superannuation funds. managed balanced, managed growth, managed stable, overseas fixed interest and currency, and mixed portfolio funds all had a positive effect for this variable. on the other hand, this legislation had a negative effect on equity funds—both domestic and overseas. choice of fund legislation has resulted in superannuation investors moving more of their investments into managed assets. the crisis variable has a positive effect on australian equities, australian fixed interest, and managed growth funds. as discussed earlier, this may be because of investors finding bargains in equities after the crisis, as well as moving more towards managed funds to take care of the asset allocations. the categories of assets on which the crisis variable has a negative effect are cash, mortgages, overseas property funds, australian property, and mixed portfolios. the link between the housing crisis and the global financial crisis can be a possible explanation for investors reducing their investments in the property sector. regression results for non-superannuation discretionary investment funds are given in table 6. active returns have a positive effect on fund flows only for australian equity and managed stable funds. tracking error had no significant effect on any of the fund categories, whereas size has a negative effect on most of the categories. because these are discretionary investments, it can be assumed that choice of fund legislation should not have much influence on the discretionary investment fund flows. contrary to this assumption, the net fund flows of a significant number of fund categories are affected by the choice variable. mixed portfolio funds are positively affected by the choice legislation, whereas most of the other fund categories are negatively affected by the choice variable. the signs of choice variable are similar for both superannuation and discretionary investments. this may be because of the fact that many superannuation investors have discretionary investments, and their investment choices are similar for both types of investments. the variable of the global financial crisis has a negative effect only on mortgage and mixed portfolio funds. other categories are not affected by the crisis. the significant difference between the superannuation funds and discretionary funds is that both australian property and property security fund net fund flows are positively affected by the crisis, which is opposite to what is observed for superannuation funds. 243r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 the investor profiles of wholesale investors are significantly different from that of retail fund investors. many of the investors in this category are other funds including superannuation funds, high net worth individuals; do-it-yourself superannuation investors (self-managed superannuation funds). the regression results for wholesale funds are given in table 7. unlike retail funds, there is less evidence of return chasing among the wholesale fund investors. only managed balanced and managed stable fund net flows are positively affected by past returns. the net fund flows of overseas fixed income funds are negatively affected by the excess return variable. the tracking error is significant only for managed balanced funds. size has a negative effect on most of the fund categories. because some of the wholesale funds are superannuation funds, it can be assumed that the choice of fund legislation will have some effect on the net fund flows in this category. the choice variable has a positive effect on the net fund flows of managed growth funds and is negative for australian property securities. there are some previous studies that indicate that many superannuation investors choose a default strategy when it comes to choosing the asset categories for their superannuation investments. in many instances the default strategy is a managed fund. with choice of fund legislation, one might expect more movement among asset categories, but as prior research has indicated, investors did not make significant asset reallocation after the choice of superannuation fund legislation was passed. australian fixed income, mortgage, and overseas fixed income funds, overseas global equity, and property security fund net fund flows are negatively affected by choice variable. there is some similarity in the way the global financial crisis affected the fund flows into wholesale and retail funds. the crisis variable has a positive effect on the equity funds net fund flows, and a negative effect on property and mortgage funds. the fluctuation in the housing market that led the financial crisis has dissuaded investors from moving into real estate related assets, whereas bargain hunting has moved them into equity markets. 6. conclusions the australian managed fund industry has grown considerably in the past two decades. the compulsory superannuation scheme can be considered to be one of the primary drivers behind this growth. this article looks at the various segments of the managed funds market to see whether there is any significant difference in the way the assets are allocated into various asset categories and whether investors base their investment decisions only on the past performance of the funds, in light of the choice of fund legislation and the recent global financial crisis. the results show that there is significant difference in asset allocation between the retail and wholesale segments. retail investors prefer less risky investments compared with wholesale investors and have lower preference for overseas investments. there are also differences in the way the superannuation, retirement income, and discretionary funds are invested. discretionary investors preferred cash as the most preferred asset category, whereas the superannuation investors preferred the managed investment category. the results of our study are similar to the evidence which has emerged from the united states managed funds industry. choice of fund legislation has had an interesting impact on the market, and in general, investors have increased their allocation into managed funds from 244 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 other equity based categories. conversely, the impact of the gfc varies from the common perception and investors generally moved their investments into shares. this provides an interesting contradiction in the sense that, as a result of choice, they have moved their investments into managed funds, realizing their lack of understanding of the market. however, after the global financial crisis, investors moved back into equity, attempting to exploit lower prices in the market place. this is consistent with the return chasing, where the temptation of past returns results in cash inflows in better performing funds. there is clear evidence that investors base their decisions primarily on the past performance of funds, with the retail segment showing a higher level of reallocation of investments based on past performance compared with the wholesale segment. the return-chasing behavior was more pronounced in the preferred asset category of each of these groups. for example, discretionary investors preferred assets over other classes of investments and showed strong return chasing behavior in this category. there is relatively less evidence of reaction to risk among the managed fund investors. part of the problem may be the difficulty in defining the risk as it may be seen by the investors. the australian managed funds industry is dominated by superannuation investments. superannuation is the primary source of retirement savings for most australians who wish to rely on self-funded retirement. the australian federal government by its policies has been pushing to cut individuals’ reliance on public pensions and has encouraged self-funded retirement objectives. evidence of investments based on past performance only in the australian managed funds industry may be of concern to policy makers, as misallocation of investments by uniformed investors may erode their retirement savings, thereby leaving the public sector with a pension liability that was not provided for in the government’s budget. choice of fund legislation has made it easier for investors to move money across funds, and evidence suggests that investors have actually moved money into managed funds after the introduction of choice of superannuation fund legislation. this study makes an important contribution to an understanding of investor behavior in the australian managed funds industry. this may be of value to fund managers who seek to inform investors of their performance, and who use performance information for marketing purposes. investors in managed funds will also benefit from gaining an understanding of how this industry works. finally, the findings of this study have importance for policy makers in understanding investor behavior when developing policies, especially in light of recent policy changes that are aimed at providing investors with more choices and ease of movement of their investments across funds. notes 1 managed funds, australia, march 2013, australian bureau of statistics. 2 market overview by plan for life actuaries & researchers, june 19, 2013. 3 this research does not test whether investors switch across funds within the same asset class after the introduction of the choice of superannuation fund legislation. 4 in the australian superannuation industry, a fund refers to the portfolio of investment classes where a member can nominate a proportion of his or her savings to be allocated across different asset classes that are available within the fund. typically each fund 245r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 will provide options such as balanced, cash, australian shares, international shares, fixed income securities, etc. 5 these are the active funds as of march 31, 2013. 6 members can choose to contribute more than that required by superannuation guarantee contributions (currently 9.5%) and these can be contributed as pre-tax or post-tax contributions. 7 watson, wickramanayke, and premachandra (2012) find that the rating agencies are unable to distinguish between highly performing and moderately performing superannuation funds. 8 the studies by sy (2011 and gerrans (2012) are based on only four years’ data. our study aims to use a much longer data series and more sophisticated empirical models for analysis. 9 choice legislation was to allow choice of superannuation fund for investors. the choice across investment within the same fund was available to most investors before the introduction of the choice legislation in 2005. we are testing for a relationship between past performance and movement across asset classes without differentiating whether this movement is within the same fund or across funds. by using a choice dummy we hypothesize that if there is a significant difference after the introduction of choice legislation, it is likely to be either across funds or because of increased awareness after the introduction of the legislation. 10 tracking error is measured as the annualized standard deviation of the differences between the portfolio return and the benchmark return. benchmark return in this case is the mean return of the industry. we have also tried using the australian stock market index in the model, but found that this creates multi-collinearity problems. 11 we thank the anonymous referee for this suggestion. 12 for brevity we report results for only one quarter and omit for two and three quarters. references agnew, j., & balduzzi, p. (2010). the response of aggregate 401(k) trading to asset returns. unpublished manuscript, boston college. bailey, w., kumar, a., & ng, d. (2011). behavioral biases of mutual fund investors. journal of financial economics, 102, 1–27. benson, k. l., gallagher, d. r., & teodorowski, p. (2007). momentum investing and the asset allocation decision. accounting and finance, 47, 571–598. bilson, c., frino, a., & heaney, r. (2005). australian retail fund performance persistence. accounting and finance, 45, 25–42. brooks, l. d., & porter, g. r. (2012). mutual fund performance attribution: 1994–2005. financial services review, 21, 259–273. capon, n., fitzsimons, g. j., & prince, r. a. (1996). an individual level analysis of the mutual fund investment decision. journal of financial services research, 10, 59–82. chen, j., hong, h., huang, m., & kubik, j. d. (2004). does fund size erode mutual fund performance? the role of liquidity and organization. american economic review, 94, 1276–1302. daniel, k. d., hirshleifer, d., & subramanyam, a. (2001). overconfidence, arbitrage and equilibrium asset pricing. the journal of finance, lvi, 921–964. 246 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 del guercio, d., & tkac, p. a. (2002) the determinants of the flow of funds of managed portfolios: mutual funds vs. pension funds. journal of financial and quantitative analysis, 37, 523–557. del guercio, d., & tkac, p. a. (2008) star power: the effect of morningstar ratings on mutual funds flow. journal of financial and quantitative analysis, 43, 907–936. drew, m. e., stanford, j. d., & veeraraghavan, m. (2002). selecting australian superannuation funds: a retail investor’s perspective. journal of financial services marketing, 7, 115–128. faff, r. w., parwada, j. t., & poh, h. l. (2007). the information content of australian managed fund ratings. journal of business finance & accounting, 34, 1528–1547. fang, j., kemf, a., & trapp, m. (2014). fund manager allocation. journal of financial economics, 111, 661–674. fear, j., & pace, g. (2009) australia’s ‘choice of fund’ legislation: success or failure? rotman international journal of pension management, 2, 26–35. frino, a., heany, r., & service, d. (2005). do past performance and past cash flows explain current cash flows into retail superannuation funds in australia? australian journal of management, 30, 229–244. fry, t., heaney, r., & mckeown, w. (2007). will investors change their superannuation fund given the choice? accounting & finance, 47, 267–283. gharghori, p., sujoto, c., & veeraraghavan, m. (2008). are australian investors smart? australian journal of management, 32, 525–544. gerrans, p. (2004). australian fund ratings and individual investors. australian journal of management, 29, 87–107. gerrans, p. (2012) retirement savings investment choices in response to the global financial crisis: australian evidence. australian journal of management, 37, 415–439. goetzmann, w. n., & peles, n. (1996). cognitive dissonance and mutual fund investors. journal of financial research, 20, 145–158. gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds. the journal of finance, 51, 783–810. guercio, d. d., & reuter, j. (2014). mutual fund performance and incentive to generate alpha. the journal of finance, lxix, 1673–1704. gupta, r., & jithendranathan, t. (2012). fund flows and past performance in australian managed funds. accounting research journal, 25, 131–157. holmes, k. a., & faff, r. w. (2007). style drift, fund flow and fund performance: new cross-sectional evidence. financial services review, 16, 55–71. ippolito, r. a. (1992). consumer reaction to measures of poor quality: evidence from the mutual fund industry. journal of law and economics, 35, 45–70. kao, c., & chiang, m. h. (2000). on the estimation and inference of a cointegrted regression in panel data. advances in econometrics, 15, 1–44. lynch, a. w., & musto, d. k. (2003). how investors interpret past fund returns. the journal of finance, 58, 2033–2058. mccoskey, s., & kao, c. (1999). testing the stability of a production function with urbanization as a shift factor. oxford bulletin of economics and statistics, special issue, 671–690. pedroni, p. (1999). critical values for cointegration tests in heterogeneous panels with multiple regressors. oxford bulletin of economics and statistics, 61, 653–70. pedroni, p. (2004). panel cointegration: asymptotic and finite sample properties of pooled time series tests with an application to the ppp hypothesis. economic theory, 20, 597–625. phillips, p. j. (2011). will self-managed superannuation fund investors survive? australian economic review, 44, 51–63. sawicki, j. (2001). investors’ differential response to managed fund performance, the journal of financial research, 24, 367–384. sensoy, b. a. (2009). performance evaluation and self-designated benchmark indexes in the mutual fund industry. journal of financial economics, 92, 25–39. 247r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 sialm, c., starks, l., & zhang, h. (2012). defined contribution pension plans: sticky or discerning money? unpublished manuscript, university of texas, austin. sirri, e. r., & tufano, p. (1998). costly search and mutual fund flows. the journal of finance, 53, 1589–1622. speelman, c. p., clark-murphy, m., & gerrans, p. (2007). decision making clusters in retirement saving: preliminary findings. australian journal of labour economics, 10, 115–127. stout, r. g. (2008). stochastic optimization of retirement portfolio asset allocations and withdrawals. financial services review, 17, 1–15. sy, w. (2011). redesigning choice and competition in australian superannuation. rotman international journal of pension management, 4, 52–62. watson, j., wickramanayke, j., & premachandra, i. m. (2012). the value of morningstar ratings: evidence using stochastic data envelopment analysis. managerial finance, 37, 94–116. 248 r. gupta, t. jithendranathan / financial services review 24 (2015) 217–248 academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university terrance k. martin utah valley university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. institutional membership to academy of financial services is $250 and includes the four annual issues of fsr. how to submit: there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academyfinancial.org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor terrance k. martin at terrance.martin@uvu.edu. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2022 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 30, no. 2, 2022 editor stuart michelson, stetson university terrance k. martin, utah valley university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $250. academic subscription price is $125 and is available by joining the academy of financial services. academic subscription includes four emailed online issues of financial services review each year. hardcopies of the journal are available at additional cost. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to terrance k. martin, academy of financial services, utah valley university, woodbury school of business, 800 w. university pkwy., orem, ut 84058. editorial office: terrance k. martin, woodbury school of business, utah valley university, 800 w. university pkwy., orem, ut 84058. phone: (940) 225-5027. email address: terrance.martin@uvu.edu. advertising information. those interested in advertising in the journal should contact terrance k. martin, woodbury school of business, utah valley university, 800 w. university pkwy., orem, ut 84058. email address: terrance.martin@uvu.edu, (940) 225-5027. printed in the usa © 2022 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, terrance k. martin, woodbury school of business, utah valley university, 800 w. university pkwy., orem, ut 84058. phone: (940) 225-5027. email address: terrance.martin@uvu.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. who demands which type of life insurance? various factors in life insurance ownership wookjae heoa,*, jae min leeb, narang parkc adepartment of consumer sciences, south dakota state university, swg 149, box 2275a, brookings, sd 57007, united states bdepartment of family consumer science, minnesota state university, mankato, 102 wiecking center, mankato, mn 56001, united states cschool of family and consumer sciences, texas state university, 601 university drive, san marcos, tx 78666, united states abstract this study examined factors related to the ownership of life insurance by focusing on the role of the psychological characteristics of the respondents. using a recent online consumer survey, logistic regression analyses were utilized based on four groups: (a) not having any term life insurance or cash value life insurance; (b) having term life insurance policy only; (c) having cash value life insurance policy only; and (d) having both term life insurance policy and cash value life insurance policy. we found that all of the financial status and psychological characteristics were significant, while some of the demographic characteristics were not significant. the specific effects of the characteristics differ by types of ownership. the ownership of term value life insurance was better explained by financial and psychological characteristics, whereas demographic characteristics factored more in the ownership of cash value life insurance. discussion and implications are provided. © 2021 academy of financial services. all rights reserved. keywords: life insurance; term life insurance; cash value life insurance; psychological characteristics 1. introduction: the need to understand the demand for life insurance life insurance is primarily designed to protect the insured against the possibility of losing an income stream, such as the premature death of the family breadwinner, which is the purpose of all types of life insurance (thoyts, 2010). term life insurance is specifically *corresponding author. tel.: +1-605-688-5835; fax: 1-605-668-5578. e-mail address: wookjae.heo@sdstate.edu (w. heo) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 101–119 designed to provide a death benefit with a relatively cheaper insurance premium (gitman, joehnk, & billingsley, 2014). on the contrary, cash value life insurance charges a higher premium to provide a savings element in addition to a death benefit; this savings element is called the “cash value” element (rejda & mcnamara, 2016). the main differences between term life insurance and cash value life insurance, as well as suggestions for consumers, can be discussed with a financial planner. whereas term life insurance has the sole purpose of protecting lost income, cash value life insurance is more likely to be the product of financial planning. for example, cash value life insurance can have a strategic financial planning purpose, with tax-deferred savings on estates, income, and bequests (clark, 2010; cymbal, 2013; kait, 2012; whitelaw, 2014). in addition, cash value life insurance can be a financial option for retirement savings (tannahill, 2012) and investment tools (cordell & landgon, 2013). therefore, many strategies in financial planning can be related to the purchase of cash value life insurance (grable, 2016). theoretically, the choice between term life insurance and cash value life insurance depends on consumers’ specific financial needs and situations (gitman et al., 2014; grable, 2016). however, the decision regarding which life insurance policy is the best option is still in question for many individual consumers. although roughly three types of life insurance can be discussed and evaluated—term life insurance, cash value life insurance, and group life insurance—the focus of such an argument made by practitioners is often placed on the first two types of life insurance, considering the fact that group life insurance provided by employers often comes without multiple options for individual purchases (rejda & mcnamara, 2016). although this question is repeatedly asked by numerous individual consumers, it is debatable even among financial practitioners as to who actually needs term life insurance versus cash value life insurance. the question is unlikely to be answered when factors come into play beyond classical demographic and functional characteristics and expected demand, accordingly. the functional characteristics of life insurance purchases, such as maintaining or improving financial security, as well as the psychological characteristics of such purchases, including feelings of comfort and recognition, can lead to these purchases being made (grable & goetz, 2017). song, park, park, and heo (2019) also emphasized that the consumer’s personal experience (i.e., death of a family member) can spark a life insurance purchase, in addition to his or her financial circumstances. if a financial planner leans only on the functional characteristics of the purchases, there may arise communication conflicts between the financial planner and the consumer, meaning that the consumer’s needs will ultimately not be fulfilled (grable & goetz, 2017). therefore, it is important to better estimate the effect of consumers’ psychological characteristics on demand for life insurance purchases to increase the accuracy of the suggestions and relevance to consumers’ different life situations. therefore, the primary goal of this study is to analyze the role of psychographic factors in the ownership of life insurance through the use of an online consumer survey. specifically, to investigate how these various factors are related to the ownership of life insurance, this study distinguished the type of ownership into four categories: (a) none of either term life insurance or cash value life insurance; (b) term life insurance only; (c) cash value life insurance only; and (d) both types of life insurance. in this study, three aspects of explanatory factors of life insurance ownership were assessed: financial status characteristics, 102 w. heo et al. / financial services review 29 (2021) 101–119 psychological characteristics, and demographic characteristics. financial status characteristics include household net balance, the ownership of emergency funds, financial risk tolerance, and subjective financial knowledge. psychological characteristics include locus of control, financial satisfaction, financial self-efficacy, and life satisfaction. demographic characteristics include the gender of the respondent, household income level, the working status of the respondent, the number of children in the household, education level, age, race, and the marital status of the respondent, as well as the perceived health condition of the respondent. the findings of this study, based on psychographics that include a combination of demographic factors and psychological factors (heo, 2020), can provide an empirical understanding of their effects on the ownership of life insurance by type. through this study’s analysis of the psychographic factors’ associations with life insurance, it is expected that financial practitioners and researchers will be able to better accommodate and identify various consumer needs for the ownership of life insurance. the research questions in this study are as follows: in what ways are financial status characteristics, psychological characteristics, and demographic characteristics associated with life insurance ownership by type of life insurance? understanding the associations between various factors and life insurance ownership, as well as the ways in which the associations differ across the subsamples by type of life insurance, will provide financial professionals with important insights when they offer financial counseling and planning services in the future. 2. literature review 2.1. different needs of life insurance by types there should be different levels of consumer demand, pure protection, or an additional investment purpose. term life insurance is well-known for its pure protection purpose, whereas cash value life insurance can offer an additional saving purpose beyond the pure protection of the lost income of the insured (gitman et al., 2014; grable, 2016; rejda & mcnamara, 2016). thus, there can be conceptual and empirical differences in understanding influential factors in the demand for life insurance by type. however, to our knowledge, there has been little empirical and conceptual research distinguishing the different demand for life insurance by type and applying psychological characteristics to an empirical analysis for financial practitioners. based on studies from a conventional perspective with regards to life insurance, life insurance has been considered to be a substitute for future savings or a potential income source of the deceased (li, moshirian, nguyen, & wee, 2007). many studies based on this perspective in the current literature have identified factors related to life insurance purchases regardless of type. predictors of the demand for overall life insurance ownership or purchase primarily include three areas (anderson, & nevin, 1975; heo & grable, 2017; liebenberg, carson, & dumm, 2012; zietz, 2003): (a) socio-demographic characteristics; (b) financial characteristics; and (c) psycho-behavioral characteristics. other studies (e.g., heo, grable, & chatterjee, 2013) have discussed the idea that cash value life insurance should be analyzed differently from term life insurance because it can function as a complement to savings. heo et al. (2013) suggested that it is important to w. heo et al. / financial services review 29 (2021) 101–119 103 further discuss practical considerations for financial services when financial planners and advisors provide suggestions to clients. thus, placing a greater emphasis on the distinction of life insurance ownership by type can extend the current discussion of the literature and add more empirical evidence to existing body of knowledge. 2.2. determinants generally known for term life insurance term life insurance policies provide insurance coverage for limited periods of time at fixed payment rates. the policies only pay death benefits to the beneficiaries if the person insured dies within the time period. after that period, term life policies do not provide any additional benefits to the insured, and the policies must be renewed with different payments or conditions if coverage is desired for another time period (brown & goolsbee, 2002). this is why term life insurance policies are often described as pure protection and can act as car or homeowner’s insurance (garman & forgue, 2018). although extensive studies in the current literature have examined the determinants of life insurance purchasing without distinguishing between type, some studies have found that the ownership of term life insurance is related to various household characteristics. anderson and nevin (1975) found that young married couples were more likely to purchase term life insurance when they had greater net worth, the spouse already had life insurance before marriage, and there was no influence of an insurance agent on the life insurance decision. goldsmith (1983) also examined the demand for term life insurance by focusing on the human capital of spouses in married-couple households. he found that the spouse’s higher educational level and the spouse’s employment status (e.g., the spouse’s participation in the labor force) decreased the spouse’s likelihood to purchase term life insurance. the spouse’s existing insurance coverage exceeding the sample mean, a greater household asset level, and larger household size were also negatively associated with the spouse’s likelihood to purchase life insurance. however, a greater household income level increased the likelihood of the insurance purchase. liebenberg et al. (2012) used the 1983-1989 survey of consumer finances (scf) panel dataset to investigate the determinants of the demand for a new life insurance policy, as well as a change in life insurance policy. they found that households whose statuses changed over the two periods (i.e., 1983 and 1989)—including having new children, experiencing relatively high levels of income growth, and launching new jobs—were more likely to purchase new term life insurance policies in the next period if they had new children. in addition, a large increase in income and a net worth increase were also related to the larger face value of the term life insurance purchased. 2.3. determinants generally known for cash value life insurance cash value life insurance policies (also known as whole life or permanent life policies) are not term-dependent (brown & goolsbee, 2002). these policies provide insurance over the lifetime of the policyholder and pay a death benefit upon the death of the insured. cash value insurance life policies also provide a savings element that is invested separately under the policy and builds up over time, either at a fixed rate or at a variable rate (garman & 104 w. heo et al. / financial services review 29 (2021) 101–119 forgue, 2018). the cash value element can pay a living benefit to the policyholders before the death of the insured, and policyholders can cborrow against the accumulated cash value and pay policy premiums using cash value. cash value life insurance policies typically charge higher premiums and come with less homogenous options because they come with a greater variety of options and plans (e.g., premium payment, borrowing, rate of return, investment types, fees, and charges) than term life insurance does (brown & goolsbee, 2002). few studies have empirically examined the effects of household characteristics on the ownership of cash value life insurance. mulholland, finke, and huston (2016) examined the determinants of ownership of cash value life insurance using the scf dataset, but despite changes in the effects of the determinants across different survey years, they found that the following were more consistently related to the likelihood of the ownership of cash value life insurance: net worth, educational level, whether or not the insured was married, having retirement saving plans (e.g., ira/roth, dc, or db plans), having a child, and financial sophistication. meanwhile, younger age and ownership of term life insurance were negatively related to the likelihood of owning cash value life insurance. liebenberg et al. (2012) found that changes in the statuses of households over the two periods, such as newly married couples’ households experiencing relatively high income growth, were more likely to have new whole life insurance policies. meanwhile, new employment, growth in income, and amount of term life insurance dropped were each positively related to the amount of new cash value life insurance. using data from the national longitudinal survey of youth 1979 (nlsy79), song et al. (2019) examined the changes in life insurance ownership during the two waves of the survey (i.e., 2008 and 2012). the researchers found that the respondents with increases in savings, as well as those who had experienced the recent deaths of family members, were more likely to purchase cash value life insurance policies. 2.4. risk tolerance, financial knowledge, and perceived health condition risk attitude has attracted significant attention from researchers as a determinant of life insurance demands, and it is conceptualized by various terms, including risk tolerance, risk aversion, risk preference, and risk-taking, depending on the definitions presented in the studies. however, no matter which term is used, risk attitude has been found to be closely associated with life insurance ownership (outreville, 2014). however, the effect of risk attitude has shown mixed results. for example, some researchers have found that consumers with less risk tolerance were likely to buy life insurance in their asset allocation (chen, ibbotson, milevsky, & zhu, 2006; finke & huston, 2003). others found that people with a higher tendency to take risks were likely to purchase life insurance because they sought greater risk exposure (burnett & palmer, 1984; xiao, 1996). baek and devaney (2005) found that an above-average level of risk-taking was positively associated with term life insurance ownership, but it was not associated with cash value life insurance ownership. song et al. (2019) also discovered that the impact of risk-taking was not significant in purchasing cash value life insurance. financial knowledge has been considered to be an important element of financial decision-making. however, there are limited studies that have investigated the impact of w. heo et al. / financial services review 29 (2021) 101–119 105 financial knowledge on life insurance demand. using the sample of adult residents of a single state, tennyson (2011) assessed the respondents’ insurance knowledge about general insurance principles, as well as the features of specific types of insurance policies. although the level of insurance knowledge was relatively low among the respondents, it was significantly related to their confidence in insurance decision-making. in life insurance markets, health status and medical history are the basic factors for pricing on term policies. thus, some researchers believe that subjectively evaluated health status is one of the most important predictors of determining the ownership of life insurance, as well as the choice of the insurance type. for example, baek and devaney (2005) found that excellence in health status was negatively associated with cash value life insurance ownership; however, they did not find any significant influence of perceived health status on term life insurance ownership decisions. to fill in the gap in the existing literature, this study included risk tolerance, financial knowledge, and perceived health condition as independent variables of consumers’ financial status characteristics in determining the purchase of life insurance. 2.5. locus of control, financial satisfaction, financial self-efficacy, and life satisfaction locus of control has been defined as the concept of a person’s perceived controllability about a situation, which can be explained with internal or external control of reinforcement (rotter, 1966). while the external locus of control denotes that any outcomes took place because of external reasons, such as fate and luck, the internal locus of control indicates that lifetime consequences occurred because of a person’s own actions (cobb-clark et al., 2016). because the locus of control drives a person to believe that an outcome occurred based on a certain circumstance (i.e., internal or external factors), locus of control is expected to be associated with financial behavior. specifically, locus of control was found to be related to personal finance and financial decision-making (e.g., cobb-clark et al., 2016; danes & rettig, 1993; nowicki, ellis, ilescaven, gregory, & golding, 2018; perry & morris, 2005; prawitz & cohart, 2016; tokunaga, 1993). for example, cobb-clark et al. (2016) found that internal locus of control was significantly associated with the tendency toward saving, which can be explained by perry and morrison’s (2005) argument that a person with a higher level of external locus of control tended to have a lower willingness to manage their financial situation. however, the association between locus of control and the ownership of life insurance was rarely found in the existing literature. therefore, in this study, the association between external locus of control and the ownership of life insurance was explored further. financial satisfaction indicates a perceived assessment of one’s own financial situation (hira & mugenda, 1998; xiao, chen, & chen, 2014; xiao & o’neill, 2018). there is no universal consensus on the measurement of financial satisfaction, which has been previously measured using either single or multiple items. for example, some studies used a single measure, such as: “overall, thinking of your assets, debts, and savings, how satisfied are you with your current personal financial condition?” (e.g., robb & woodyard, 2011; xiao et al., 2014; xiao & o’neill, 2018). others have used multiple items (e.g., loibl & hira, 2005; montalto, phillips, 106 w. heo et al. / financial services review 29 (2021) 101–119 mcdaniel, & baker, 2019), including financial situation and the ability to understand and make sound financial decisions. as an element of general life satisfaction and well-being, a positive relationship between financial behaviors and financial satisfaction has been found (robb & woodyard, 2011; xiao et al., 2014; xiao & o’neill, 2018). in particular, robb and woodyard (2011) found that a higher financial satisfaction level was associated with more positive financial practices defined by six items including having an emergency fund, high credit report, no overdraft, credit card payoff, having a retirement account, and effective risk management. however, despite extensive studies having been done about financial satisfaction, the existing literature about the association between financial satisfaction and ownership of life insurance is limited. self-efficacy refers to an individual’s belief in their ability to perform a specific task successfully (bandura, 1977, 2006). a person with a strong sense of self-efficacy can execute the cognitive and behavioral efforts required to obtain the desired outcome (bandura, 1977). thus, the feeling of self-efficacy has been revealed to be one of the major factors influencing human behavior (ajzen, 2002; bandura, 1977). researchers in personal finance have been aware of the importance of financial self-efficacy to improve financial capability (amatucci & crawley, 2011; lown, 2011). generally defined as “one’s sense of being prepared and able to handle financial responsibility” (montalto, et al., 2019, p. 15), financial self-efficacy has been found to be significantly and positively associated with responsible financial behaviors, including help-seeking, making investments, and saving (asebedo, & seay, 2018; farrell, fry, & risse, 2016; lim, heckman, montalto, & letkiewicz, 2014). furthermore, financial self-efficacy is linked to financial well-being and subjective well-being (robb, 2017). based on previous findings, it is hypothesized that financial self-efficacy can be related to the likelihood of life insurance ownership, which is a type of financial management. to assess one’s level of financial self-efficacy, lown (2011) developed a financial self-efficacy scale (fses) that measures an individual’s self-efficacy specific to certain financial behaviors. directly modeled on the general self-efficacy scale (gses), the financial self-efficacy scale consists of six items that ask questions about respondents’ confidence in terms of managing their personal finances. higher scores on the financial self-efficacy scale indicate higher confidence in personal financial management. life satisfaction is a cognitive judgment regarding a person’s own quality of life (diener, emmons, larsen, & griffin, 1985). the assessment of quality of life is dependent upon the comparison of one’s perceived life circumstances with self-imposed standards (pavot & diener, 2009). thus, ones’ degree of satisfaction with life is highly up to the individual person (diener, emmons et al., 1985; pavot & diener, 2009). researchers have found that life satisfaction is closely related to one’s financial status proxied by income, wealth, or financial satisfaction. the major argument was whether or not money could buy happiness. although researchers have pointed out that the effects of financial factors on life satisfaction were only minimal, studies tended to confirm that there was a positive association between financial status and life satisfaction (cheung & lucas, 2015; diener & biswas-diener, 2002; diener & diener, 2009; heo, lee, & park, 2020; johnson & krueger, 2006; park, lee, & heo, 2020). in other words, individuals who earned a higher income, who had more wealth, and who reported both higher levels of financial satisfaction and lower levels of financial stress were more likely to exhibit higher levels of life satisfaction. w. heo et al. / financial services review 29 (2021) 101–119 107 however, the association between financial status and life satisfaction can also be inversive. by reviewing a variety of panel studies, diener and biswas-diener (2002) concluded that the relationship between subjective well-being and financial status is bidirectional, indicating that happy people could be more proactive in managing their finances. based on this assumption, it is reasonable to think that the subjective perception of life can influence one’s decision-making regarding life insurance ownership. if a person is highly satisfied with one’s own life, the person would be more likely to have life insurance, either as a means of transferring financial resources or as a savings vehicle. however, if a person is unsatisfied with her/his life, the person might not look beyond the present. nevertheless, there is a lack of existing literature that examines how an individual’s life satisfaction determines his/her decision to buy life insurance. 3. data and methodology 3.1. data and analyses this study conducted an online consumer survey of 1,000 respondents across the united states, collected with a random sampling method in september 2019. the questionnaire measured both individualand household-level characteristics regarding various financial decisions and circumstances, including life insurance type, net balance status, emergency fund ownership, financial risk tolerance, two types of financial knowledge (subjective and objective), perceived health condition, gender, income, working status, number of children in a household, education, age, race, and relationship status. after excluding respondents who did not complete the questions used by this study, the final sample consisted of 997 respondents. the survey was funded by the national institute of food and agriculture (nifa) as the hatch project. the details of the psychographics of the respondents are shown in the next section. in this study, a four subsample analysis was conducted based on the type of life insurance ownership. specifically, out of the total sample (n¼ 997), 521 respondents answered that they did not have any term life insurance or cash value life insurance. among the 476 respondents who had either term or cash value life insurance, 227 respondents reported that they only had term life insurance, whereas 89 respondents reported that they only had cash value life insurance. lastly, 160 respondents answered that they had both term life insurance and cash value life insurance. to answer the research question, this study utilized four logistic regression analyses to investigate the marginal effect of each influential factor on the demand for life insurance by type. the dependent variables were binary indicators of whether or not the respondent had a life insurance policy: (a) not having any term life insurance or cash value life insurance; (b) having term life insurance policy only; (c) having cash value life insurance policy only; and (d) having both term life insurance policy and cash value life insurance policy. thus, the marginal effects of the influential factors were checked against the odds ratio, which was calculated by the probability of “having life insurance” versus “not having life insurance,” so that the odds ratio of each factor in the model denotes the tendency of having life insurance influenced by the factor. 108 w. heo et al. / financial services review 29 (2021) 101–119 to ensure the robustness of the estimation, a seemingly unrelated estimation method was utilized when executing four logistic regression models. specifically, seemingly unrelated estimation executed multiple models promptly, considering the covariances and distribution simultaneously (rogers, 1993; white, 1982). by using stata 15.0 with a reliable code (i.e., suest) for seemingly unrelated estimation (weesie, 1999), four logistic models were simultaneously executed for the robustness of the model results. in addition, the significance criterion was set as alpha¼ 10% (p< .10) because the sample size of each subsample was relatively small, such as 227, 89, and 160. when the sample size is small, there is a tendency that type ii error increases (banerjee, chitnis, jadhiv, bhawalkar, & chaudhury, 2009). when alpha is strict to 5% (p< .05), a small sample size is more likely to produce insignificant results, even though they are significant. therefore, in this study, the alpha level was set at 10% (p< .10). 3.2. variables we used four dependent variables: (a) not having any term life insurance or cash value life insurance; (b) having term life insurance policy only; (c) having cash value life insurance policy only; and (d) having both term life insurance policy and cash value life insurance policy. the dependent variable was measured as a dichotomous variable (yes or no) based on the answer to the question that follows this brief explanation: “the two major types of life insurance are term and cash value policies. term policies pay a benefit if the insured person dies, but otherwise, they have no value. they are often provided through an employer or union, but they may also be bought by individuals. cash value policies also pay death benefits, but differ in that they build up value as premiums are paid. are any of your (or your spouse/partner’s) policies term insurance?” the other dependent variable was measured as a dichotomous variable (yes or no) by the answer to the following question: “do you (or your spouse/partner) have any policies that build up cash value or that you can borrow on? these are sometimes called ‘whole life,’ ‘straight life,’ or ‘universal life’ policies.” finally, the answers to all three questions were coded as binary variables: “yes” was coded as 1, and 0 otherwise. the key independent variables in this study include financial status characteristics, psychological characteristics, and demographic characteristics. as the financial status characteristics, a net balance of respondents’ emergency funds, financial risk tolerance, and subjective financial knowledge were used. the net balance was measured by a categorical variable: zero net balance, negative net balance, and positive net balance. the ownership of an emergency fund was measured as a binary question, and these two financial status factors were coded as binary variables (yes¼ 1; no¼ 0). for financial risk tolerance, grable and lytton’s 13 items were used (grable & lytton, 1999), which were considered to be reliable and valid measurements for financial risk tolerance (grable, lyons, & heo, 2019). financial risk tolerance ranged from 13 points to 40 points, where the lowest number meant the lowest level of financial risk tolerance. subjective financial knowledge was measured with a question regarding the self-assessment of one’s own level of financial knowledge, ranging from 1 point (lowest level) to 7 points (highest level). the psychological characteristics include four variables: external locus of control, financial satisfaction, financial self-efficacy, and life satisfaction. external locus of control was w. heo et al. / financial services review 29 (2021) 101–119 109 measured by using eight items (perry & morrison, 2005) that the questionnaires asked to answer with a 5-point scale (1¼ almost never, 5¼ almost always), so that the maximum number of summing the eight items (i.e., 40) denotes the highest level of external locus of control. otherwise, the minimum number of summing the eight items (i.e., 8) indicates the lowest level of external locus of control. financial satisfaction was measured with seven items from loibl & hira (2005), in which a 5-point scale was utilized (1¼ very dissatisfied, 5¼ very satisfied). the maximum number of summing these seven items (i.e., 35) denotes the highest level of financial satisfaction, while the minimum number of summing seven items (i.e., 7) indicates the lowest level of financial satisfaction. in terms of financial self-efficacy, six items were utilized from lown (2011), where a 5-point scale was utilized (1¼ strongly disagree, 5¼ strongly agree). the maximum number of summing these six items (i.e., 30) denotes the highest level of financial self-efficacy, while the minimum number of summing six items (i.e., 6) indicates the lowest level of financial self-efficacy. lastly, life satisfaction utilized five items of satisfaction with a scale derived from diener, emmons, larsen, & griffin (1985). for life satisfaction, a 7-point scale was utilized (1¼ strongly disagree, 7¼ strongly agree). therefore, the maximum number of summing these five items (i.e., 35) denotes the highest level of life satisfaction, while the minimum number of summing five items (i.e., 5) indicates the lowest level of life satisfaction. demographic characteristics consist of both respondent-level and household-level variables. for the respondent-level variables, four variables were measured as binary variables: gender (male¼ 0, female¼ 1), working status (not working¼ 0, working¼ 1), race/ethnicity (nonwhite¼ 0, white¼ 0), and relationship status (not married¼ 0, married/coupled¼ 1). the education level of the respondent was measured with a categorical variable: high school graduate or lower, associate degree, bachelor’s degree, and graduate-level or higher. the lowest education level (i.e., high school graduate or lower) is used as a reference group in an analytic procedure. the age of the respondent was measured by number of years. household-level variables include income and the number of children in the household. income was measured between eight different categories: lower than $15,000, $15,000-24,999, $25,000-34,999, $35,000-49,999, $50,000-74,999, $75,000-99,999, $100,000-149,999, and greater than $150,000. the lowest income level (i.e., lower than $15,000) was used as a reference group. the number of children in the household was measured as a continuous variable. finally, perceived health condition was the self-assessment of one’s perceived health condition, and this was measured as a binary variable (not good¼ 0, good¼ 1). 4. findings 4.1. descriptive information of samples tables 1 and 2 displayed the descriptive statistics of both the total sample and the subsamples. for example, among the total sample result (n¼ 997), approximately 40% of the respondents had zero net balances or negative net balances, and around 45% of respondents had emergency funds. the income levels of approximately half of the respondents fell 110 w. heo et al. / financial services review 29 (2021) 101–119 between $35,000 and $99,999, which means that the sample was not highly skewed to highor low-income respondents, the majority of which were working white females who had completed at least an associate degree. half of the respondents were either married or living with their partners. the average number of children and the average age of the respondents were less than one and 47 years old, respectively. for the physio-psychological characteristics, the average level of financial risk tolerance, subjective financial knowledge, and objective financial knowledge were measured as 20.67, 3.71, and 1.96, respectively. the levels of financial knowledge were both slightly higher than the median value of each, measuring for example as 3.5 and 1.5, respectively. the majority of respondents responded that they were in good health, and this is a limitation of the study caused by random sampling through an online survey. in terms of the psychological factors, the average value of external locus of control was 24.87, with a standard deviation of 4.91; the average value of financial satisfaction was 21.16, with a standard deviation of 7.34; the average value of financial self-efficacy was 15.15, with a standard deviation of 5.04; and the average value of life satisfaction was 20.92, with a standard deviation of 8.35. table 1 descriptive statistics of samples: categorical factors total sample (n= 997) no insurance (n= 521) term only (n= 227) cash-value only (n= 89) both insurance (n= 160) freq. (%) freq. (%) freq. (%) freq. (%) freq. (%) financial status characteristics negative nb 263 (26.38) 173 (33.21) 56 (24.67) 9 (10.11) 25 (15.63) emer. funds (yes) 453 (45.44) 177 (33.97) 105 (46.26) 61 (68.54) 110 (68.75) demographic characteristics gender (= female) 776 (77.83) 423 (81.19) 186 (81.94) 53 (59.55) 114 (71.25) income lower than $15k 113 (11.33) 89 (17.08) 11 (4.85) 7 (7.87) 6 (3.75) $15k $25k 126 (12.64) 97 (18.62) 8 (3.52) 10 (11.24) 11 (6.88) $25k $35k 144 (14.44) 87 (16.70) 29 (12.78) 12 (13.48) 16 (10.00) $35k $50k 157 (15.75) 87 (16.70) 36 (15.86) 16 (17.98) 18 (11.25) $50k $75k 182 (18.25) 88 (16.89) 50 (22.03) 18 (20.22) 26 (16.25) $75k $100k 128 (12.84) 39 (7.49) 43 (18.94) 11 (12.36) 35 (21.88) $100k $150k 105 (10.53) 27 (5.18) 38 (16.74) 13 (14.61) 27 (16.88) over $150k 42 (4.21) 7 (1.34) 12 (5.29) 2 (2.25) 21 (13.13) work-status (yes) 657 (65.90) 315 (60.46) 167 (7357) 44 (49.44) 131 (81.88) education high school/lower 235 (23.57) 154 (29.56) 37 (16.30) 15 (16.85) 29 (18.13) associate degree 303 (30.39) 167 (32.05) 72 (31.72) 29 (32.58) 35 (21.88) bachelor degree 321 (32.20) 148 (28.41) 90 (39.65) 26 (29.21) 57 (35.63) graduate/higher 138 (13.84) 52 (9.98) 28 (12.33) 19 (21.35) 39 (24.38) race/ethnicity white 809 (81.14) 414 (79.46) 193 (85.02) 78 (87.64) 124 (77.50) relationship status married/coupled 544 (54.46) 227 (43.57) 153 (67.40) 54 (60.67) 111 (69.38) note: nb = net balance; emer. funds = emergency funds. w. heo et al. / financial services review 29 (2021) 101–119 111 4.2. results from logistic estimation and seemingly unrelated estimation table 3 displays the results from the four logistic regression analyses. financial status characteristics were partially associated with having life insurance. the negative net balance was positively associated with having no insurance (coefficient¼ 0.31, p< .10), meaning that those with negative net balances were less likely to have any of either term life insurance and cash value life insurance. however, the ownership of emergency funds works in the opposite direction, as those with emergency funds tended to have no term life insurance or cash value life insurance (coefficient = �0.47, p< .01). they were also more likely to have both term life insurance and cash value life insurance than those without emergency funds (coefficient¼ 0.72, p< .01). subjective financial knowledge showed a similar pattern to the emergency fund variable. those who had a higher level of subjective financial knowledge were less likely to have no life insurance (coefficient = �0.13, p< .05), and were also more likely to have both types of insurance (coefficient¼ 0.26, p< .01). in terms of financial risk tolerance, those who had a higher level of financial risk tolerance tended to be less likely to have no life insurance (coefficient = �0.03, p< .01) and ownership of term life insurance (coefficient = �0.04, p< .10). the psychological characteristics were also partially associated with the ownership of life insurance. first, locus of control showed a negative association with no-ownership of any term life insurance or cash value life insurance (coefficient = �0.03, p< .10), but a positive association with the ownership of both term life insurance and cash value life insurance (coefficient¼ 0.05, p< .05). this means that a person with a higher level of external locus table 2 descriptive statistics of samples: continuous factors total sample (n5 997) no insurance (n5 521) term only (n5 227) cash-value only (n5 89) both insurance (n5 160) mean (sd) mean (sd) mean (sd) mean (sd) mean (sd) financial status characteristics frt 20.67 (4.26) 20.22 (4.09) 20.26 (3.69) 20.54 (4.20) 22.80 (4.92) sub. fk. 3.71 (1.58) 3.39 (1.50) 3.68 (1.50) 3.88 (1.42) 4.71 (1.61) demographic characteristics health status 0.72 (0.45) 0.65 (0.48) 0.74 (0.44) 0.87 (0.34) 0.86 (0.35) no. children 0.71 (1.15) 0.63 (1.17) 0.78 (1.10) 0.38 (0.89) 1.04 (1.20) age 47.02 (15.90) 45.55 (16.08) 47.56 (15.19) 56.06 (13.84) 44.36 (14.46) psychological characteristics loc 24.87 (4.91) 24.85 (4.57) 24.20 (3.81) 23.73 (4.33) 26.56 (6.89) f-satisfaction 21.16 (7.34) 19.62 (7.10) 21.04 (6.75) 24.37 (6.89) 24.59 (7.50) f-self 15.15 (5.04) 15.59 (5.00) 14.90 (4.82) 12.70 (4.55) 15.44 (5.36) l-satisfaction 20.92 (8.35) 18.90 (8.28) 22.08 (7.53) 22.46 (7.53) 25.00 (8.15) note: frt = financial risk tolerance; sub. fk. = subjective financial knowledge; char. = characteristics; loc = locus of control; f-satisfaction = financial satisfaction; f-self = financial self-efficacy; l-satisfaction = life satisfaction. 112 w. heo et al. / financial services review 29 (2021) 101–119 table 3 results from logistic regressions by types of life insurance ownership with seemingly unrelated estimation method for robustness model 1 model 2 model 3 model 4 no insurance term life only cash-value only both insurances coefficient (robust se) coefficient (robust se) coefficient (robust se) coefficient (robust se) fs characteristics negative nb 0.31† (0.18) �0.11 (0.21) �0.54 (0.41) �0.37 (0.27) emer. funds (yes¼ 1) �0.47** (0.18) �0.16 (0.20) 0.51 (0.35) 0.72** (0.23) frt 0.02 (0.02) �0.04† (0.02) �0.04 (0.03) 0.02 (0.03) sub. fk �0.13* (0.06) 0.01 (0.07) �0.05 (0.09) 0.26** (0.09) psychological characteristics loc �0.03† (0.02) �0.03 (0.02) 0.03 (0.03) 0.05* (0.02) f-satisfaction �0.00 (0.02) �0.03* (0.02) 0.01 (0.03) 0.02 (0.02) f-self �0.04† (0.02) �0.00 (0.02) �0.05† (0.03) 0.08** (0.03) l-satisfaction �0.03* (0.02) 0.03* (0.01) �0.03 (0.02) 0.02 (0.02) demographic characteristics gender (female¼ 1) 0.14 (0.19) 0.22 (0.22) �0.90** (0.28) 0.22 (0.23) income $15k $25k �0.07 (0.33) �0.58 (0.49) 0.06 (0.59) 0.91 (0.55) $25k $35k �0.74* (0.30) 0.70† (0.39) 0.14 (0.55) 0.95† (0.51) $35k $50k �0.88** (0.30) 0.88* (0.39) 0.10 (0.53) 0.99† (0.52) $50k $75k �0.97** (0.30) 1.04** (0.39) �0.03 (0.56) 1.06* (0.50) $75k $100k �1.65*** (0.33) 1.38** (0.41) �0.52 (0.62) 1.84*** (0.51) $100k $150k �1.64*** (0.35) 1.73*** (0.42) �0.27 (0.61) 1.18* (0.54) over $150k �2.05*** (0.50) 1.42** (0.52) �1.41 (0.92) 2.13*** (0.58) working (work¼ 1) �0.24 (0.18) 0.31 (0.21) 0.03 (0.30) 0.21 (0.25) health status �0.22 (0.18) �0.27 (0.21) 0.95** (0.33) 0.43 (0.27) number of children �0.11 (0.07) �0.04 (0.07) �0.10 (0.17) 0.24** (0.08) education level associate degree �0.30 (0.20) 0.27 (0.24) 0.48 (0.36) �0.09 (0.30) bachelor degree �0.29 (0.21) 0.32 (0.24) 0.18 (0.39) �0.04 (0.31) graduate/higher �0.14 (0.27) �0.41 (0.32) 0.88* (0.42) 0.19 (0.37) age �0.02** (0.01) 0.00 (0.01) 0.04*** (0.01) �0.00 (0.01) race (white¼ 1) 0.01 (0.19) 0.14 (0.23) 0.13 (0.34) �0.27 (0.26) married/coupled �0.31† (0.17) 0.26 (0.19) 0.40 (0.30) 0.09 (0.23) constant 4.57*** (0.83) �1.06 (0.86) �4.03** (1.30) �8.63*** (1.14) x2 226.80*** 95.41*** 104.30*** 185.49*** pseudo r 2 0.16 0.09 0.17 0.21 note: in model 1, the dependent variable was ownership of neither term or cash-value life insurance; in model 2, the dependent variable was ownership of only term life insurance; in model 3, the dependent variable was ownership of only cash-value life insurance; and in model 4, the dependent variable was ownership of both term life and cash-value life insurances. reference for net balance is equal to or greater than zero net balance; reference for gender is male; reference for income category is lower than $15,000; reference for working status is not working; reference for education level was lower than high school; reference for the race is if non-white; and reference for marital status is single. fs = financial status; nb = net balance; emer. funds = emergency funds; frt = financial risk tolerance; sub. fk. = subjective financial knowledge; char. = characteristics; loc = locus of control; f-satisfaction = financial satisfaction; f-self = financial self-efficacy; and l-satisfaction = life satisfaction. †p< .10, *p< .05, **p< .01, ***p < .001. w. heo et al. / financial services review 29 (2021) 101–119 113 of control (e.g., blaming external reasons) tended to have both term life insurance and cash value life insurance. similarly, financial self-efficacy showed a negative association with noownership of any term life insurance or cash value life insurance (coefficient = �0.04, p< .10), but a positive association with the ownership of both term life insurance and cash value life insurance (coefficient¼ 0.08, p< .01), which implied that a person with a higher level of financial self-efficacy was more likely to have both term life insurance and cash value life insurance. however, higher-level financial self-efficacy lowered the likelihood of ownership of only cash value life insurance (coefficient = �0.05, p< .10). third, financial satisfaction was associated with only the ownership of term life insurance negatively (coefficient = �0.03 p< .05), which means that those who had a higher level of financial satisfaction would not purchase term life insurance. finally, life satisfaction showed a negatively significant association with no-ownership of both term life insurance and cash value life insurance (coefficient = �0.03, p< .05), which implied that those who were more satisfied with their life were more likely to have either kind of life insurance. the likelihood of having ownership of only term life insurance increased by level of life satisfaction (coefficient¼ 0.03, p< .05). demographic characteristics also demonstrated different effects of factors on ownership by type of life insurance. some variables were model-specific. females showed a negatively significant association only with cash value life insurance (coefficient = �0.90, p< .01), while the highest level of education (i.e., graduate or higher; coefficient = 0.88, p< .05) and good health status (coefficient¼ 0.95, p< .01) were positively associated with ownership of cash-value life insurance. working status (i.e., currently working), lower levels of education (i.e., associate degree and bachelor’s degree), and race (i.e., whites) did not show any significant relevance to the ownership of life insurance. the number of children in a family showed a positive association with the ownership of both types of life insurance (coefficient¼ 0.24, p< .01), and the married or coupled respondents were less likely to have no life insurance at all (coefficient = �0.31, p< .10). older respondents were less likely to have no life insurance and more likely to have both term and cash value life insurance. there was an interesting finding stemming from the income level variable. the higher the income level, the greater the probability of having life insurance (less likelihood of no ownership, term life, both term life insurance, and cash value life insurance), except for the ownership of cash value life insurance. 5. discussion and implication this study examined factors related to the ownership of life insurance by type (none, term, cash value, and both term and cash value life insurance) by focusing on the role of the financial status characteristics, psychological characteristics, and demographic characteristics. this study found that some financial status and psychological characteristics show similar and opposite patterns across models. the presence of emergency funds, locus of control, subjective knowledge, and financial self-efficacy decreased the likelihood of non-ownership of life insurance and increased the likelihood of owning both types of life insurance. among those characteristics, the ownership of a specific type of life insurance was not significant. 114 w. heo et al. / financial services review 29 (2021) 101–119 those who were more sensitive to external circumstances in terms of determining life events and were prepared for financial emergency tended to have one or both forms of life insurance (=lower likelihood of non-ownership of life insurance and a higher likelihood of ownership of both forms of life insurance). these respondents were also identified as those who perceived their financial knowledge levels to be higher and who feel more confident about themselves in terms of achieving their financial goals. all of these financial and psychological characteristics show that they would value the fundamental core of life insurance purchases more than other additional characteristics of life insurance (i.e., they would like to transfer the risk of loss of income) based on their perceptions and knowledge about the need for life insurance. thus, it appears that they would fully understand the importance of life insurance purchases. in particular, however, those with a higher level of financial self-efficacy had a lower likelihood of cash value life insurance ownership, reflecting the fact that they would need financial advice about a further decision regarding life insurance (e.g., type of life insurance) beyond the basic purpose. the traits of financial risk tolerance and financial satisfaction both reduced the likelihood of ownership of term life insurance, but increased that of ownership of both insurance policies, while life satisfaction increased the ownership of term life insurance and reduced nonownership of life insurance. in other words, those who take greater financial risk and are more satisfied financially recognize the need for life insurance (i.e., they are less likely to not own life insurance), but these characteristics about financial attitude and evaluation undervalued term life insurance. this result was the opposite of that of life satisfaction. those who were satisfied with their lives but not necessarily satisfied with its financial aspects would prefer ownership of term life insurance. although financial satisfaction has been known to be an element of life satisfaction (e.g., robb & woodyard, 2011; xiao et al., 2014), its effects on financial decisions, such as term life insurance decisions, were not necessarily in the same direction. this implies the importance of taking various aspects of psychological characteristics into consideration when investigating the determinants of life insurance ownership. our findings warn about the generalization of the effect of seemingly related characteristics on financial decisions and suggest that financial practitioners consider each of them closely when working with clients. other financial status characteristics, such as negative net balances and income, were also significant in some models. negative net balances reduced the likelihood of ownership of life insurance, while income levels were generally positively related to the ownership of life insurance; more financially stable respondents tended to have either term life or both term and cash value life insurance. as the income level went up, the likelihood of having a term life insurance policy generally increased, but this pattern was not significant in the pattern of cash value life insurance ownership. the results may suggest that term life insurance is considered to be a substitute for income. while the ownership of term value life insurance was better explained by financial and psychological characteristics, more demographic characteristics factored into the ownership of cash value life insurance, which provides an additional investment vehicle and tax-wise asset accumulation. cash value life insurance was in greater demand by those in good health condition (thus, having relatively less demand for a savings chunk for sudden medical costs), or older respondents (who might have limited eligibility for a term, or might be more w. heo et al. / financial services review 29 (2021) 101–119 115 familiar with investment vehicles). however, females were less likely to have cash value life insurance, implying that the potential gender difference in financial decisions, and financial professionals can help them correct this aversion toward cash value life insurance if it is not beneficial to them. married or coupled respondents were less likely to have no life insurance at all, while those with more children tended to have both types of life insurance, all of which identifies who might possibly better perceive the significance of life insurance because of the presence of financial dependents. however, work status, lower levels of education, and race were not significant in all models. although researchers have identified the various factors that contribute to the likelihood of life insurance ownership, few studies have attempted to approach the subgroup analyses. our findings confirmed the importance of conducting separate analyses by policy type (i.e., no ownership, only term life insurance, only cash value life insurance, and both term and cash value life insurance), and deepened the discussion about the factors related to life insurance ownership. our findings are in line with the different characteristics and purposes of term and cash value life insurance. cash value life insurance was likely to be used to complement other wealth accumulation vehicles rather than as a simple substitute for lost income from premature death, whereas term life insurance was still seen as pure protection for the potential loss of future income. the results are consistent with previous work (e.g., heo et al., 2013), which suggests that the purpose and characteristics of cash value life insurance should not be treated in the same way as those of term life insurance. thus, the findings regarding the differences in consumer demand between cash value life insurance and term life insurance in terms of the influential characteristics and underlying purpose of the life insurance purchase suggest a need for a more careful approach to the study of life insurance ownership and the incorporation of psychological factors in the analysis. the findings suggest that consumers can be better served when financial practitioners and researchers more carefully identify and accommodate consumers’ needs for different types of life insurance. various idiosyncratic characteristics of consumers, including financial, psychological, and demographic factors, should be considered along with the functional purposes of the life insurance being purchased. finally, there are some study limitations of the study to mention here. first, this study used an online consumer survey based on random sampling methods, as an online survey has been shown to better reach a broader population in different geographic locations. however, our sample distribution was somewhat skewed in terms of some demographic variables, such as gender, race/ethnicity, and income. therefore, the findings of the research need to be generalized with caution, and future studies can extend the discussion by examining a more diverse population. the survey used for the study also does not include questions about the amount of the life insurance policies or how the insurance was purchased (e.g., through an employer, such as group life insurance, or individually), which can possibly be related to the ownership of life insurance and might be helpful to financial professionals when identifying the needs of their clients before advising them in their decision of what type of life insurance policy to pursue. thus, future studies are desired to examine additional characteristics of life insurance purchasing with different data. 116 w. heo et al. / financial services review 29 (2021) 101–119 acknowledgment this work was supported by the usda national institute of food and agriculture, hatch project 1017028. references ajzen, i. (2002). perceived behavioral control, self-efficacy, locus of control, and the theory of planned behavior 1. journal of applied social psychology, 32, 665-683. amatucci, f. m., & crawley, d. c. (2011). financial self-efficacy among women entrepreneurs. international journal of gender and entrepreneurship, 3, 23-37. anderson, d. r., & nevin, j. r. (1975). determinants of young marrieds’ life insurance purchasing behavior: an empirical investigation. the journal of risk and insurance, 42, 375-387. asebedo, s. d., & seay, m. c. (2018). financial self-efficacy and the saving behavior of older pre-retirees. journal of financial counseling and planning, 29, 357-368. baek, e., & devaney, s. a. (2005). human capital, bequest motives, risk, and the purchase of life insurance. journal of personal finance, 4, 62-84. bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84, 191-215. bandura, a. (2006). guide for constructing self-efficacy scales. in f. pajares & t. urdan (eds.), self-efficacy beliefs of adolescents (pp. 307-337). greenwith, ct: information age publishing banerjee, a., chitnis, u. b., jadhav, s. l., bhawalkar, j. s., & chaudhury, s. (2009). hypothesis testing, type i and type ii errors. industrial psychiatry journal, 18, 127-131. brown, j. r., & goolsbee, a. (2002). does the internet make markets more competitive? evidence from the life insurance industry. journal of political economy, 110, 481-507. burnett, j. j., & palmer, b. a. (1984). examining life insurance ownership through demographic and psychographic characteristics. the journal of risk and insurance, 51, 453-467. chen, p., ibbotson, r. g., milevsky, m. a., & zhu, k. x. (2006). human capital, asset allocation, and life insurance. financial analysts journal, 62, 97-109. cheung, f., & lucas, r. e. (2015). when does money matter most? examining the association between income and life satisfaction over the life course. psychology and aging, 30, 120-135. clark, b. z. (2010). spousal life time access trusts: planning opportunities using second�to�die life insurance policies. journal of financial services professionals, 64, 50-56. cobb-clark, d. a., kassenboehmer, s. c., & sinning, m. g. (2016). locus of control and savings. journal of banking & finance, 73, 113-130. cordell, d. m., & landgon, t. p. (2013). using life insurance to fund special needs trusts. journal of financial planning, 26, 34-35. cymbal, k. m. (2013). choosing a family member as trustee of an irrevocable life insurance trust. journal of financial services professionals, 67, 41-52. danes, s. m., & rettig, k. d. (1993). the role of perception in the intention to change the family financial situation. journal of family and economic issues, 14, 365-389. diener, e., & biswas-diener, r. (2002). will money increase subjective well-being? a literature review and guide to needed research. social indicators research, 57, 119-169. diener, e., & diener, m. (2009). cross-cultural correlates of life satisfaction and self-esteem. in e. diener (ed.), culture and well-being (pp. 71-91). dordrecht, netherland: springer. diener, e., emmons, r. a., larsen, r. j., & griffin, s. (1985). the satisfaction with life scale. journal of personality assessment, 49, 71-75. farrell, l., fry, t. r., & risse, l. (2016). the significance of financial self-efficacy in explaining women’s personal finance behaviour. journal of economic psychology, 54, 85-99. finke, m. s., & huston, s. j. (2003). the brighter side of financial risk: financial risk tolerance and wealth. journal of family and economic issues, 24, 233-256. w. heo et al. / financial services review 29 (2021) 101–119 117 garman, e., & forgue, r. (2018). personal finance (13th ed.). mason, oh: south-western cengage learning. gitman, l. j., joehnk, m. d., & billingsley, r. s. (2014). personal financial planning (13th ed.). mason, oh: cengage. goldsmith, a. (1983). household life cycle protection: human capital versus life insurance. the journal of risk and insurance, 50, 473-486. grable, j. e. (2016). the case approach to financial planning bridging the gap between theory and practice (3rd ed.). erlanger, ky: the national underwriter company. grable, j. e., & goetz, j. w. (2017). communication essential for financial planners: strategies and techniques. hoboken, nj: wiley. grable, j. e., lyons, a. c., & heo, w. (2019). a test of traditional and psychometric relative risk tolerance measures on household financial risk taking. finance research letters, 30, 8-13. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163-181. heo, w. (2020). the demand for life insurance: dynamic ecological systemic theory using machine learning techniques. switzerland: springer nature. heo, w., & grable, j. e. (2017). demand for life insurance: a consumer perspective. journal of financial service professionals, 71, 19-23. heo, w., grable, j. e., & chatterjee, s. (2013). life insurance consumption as a function of wealth change. financial services review, 22, 389-404. heo, w., lee, j. m., & park, n. (2020). financial-related psychological factors affect life satisfaction of farmers. journal of rural studies, 80, 185-194. hira, t. k., & mugenda, o. m. (1998). predictors of financial satisfaction: differences between retirees and nonretirees. journal of financial counseling and planning, 9, 75-84. johnson, w., & krueger, r. f. (2006). how money buys happiness: genetic and environmental processes linking finances and life satisfaction. journal of personality and social psychology, 90, 680-691. kait, r. e. (2012). one life insurance size doesn’t fit all estate planning situations. journal of financial planning, 26, 38-39. li, d., moshirian, f., nguyen, p., & wee, t. (2007). the demand for life insurance n oecd countries. journal of risk & insurance, 74, 637-652. liebenberg, a. p., carson, j. m., & dumm, r. e. (2012). a dynamic analysis of the demand for life insurance. journal of risk and insurance, 79, 619-644. lim, h., heckman, s., montalto, c. p., & letkiewicz, j. (2014). financial stress, self-efficacy, and financial help-seeking behavior of college students. journal of financial counseling and planning, 25, 148-160. loibl, c., & hira, t. k. (2005). self-directed financial learning and financial satisfaction. journal of financial counseling and planning, 16, 11-22. lown, j. m. (2011). development and validation of a financial self-efficacy scale. journal of financial counseling and planning, 22, 54-63. montalto, c. p., phillips, e. l., mcdaniel, a., & baker, a. r. (2019). college student financial wellness: student loans and beyond. journal of family and economic issues, 40, 3-21. mulholland, b., finke, m., & huston, s. (2016). understanding the shift in demand for cash value life insurance. risk management and insurance review, 19, 7-36. nowicki, s., ellis, g., iles-caven, y., gregory, s., & golding, j. (2018). events associated with stability and change in adult locus of control orientation over a six-year period. personality and individual differences, 126, 85-92. outreville, j. f. (2014). risk aversion, risk behavior, and demand for insurance: a survey. journal of insurance issues, 158-186. park, n., lee, j. m., & heo, w. (2020). life satisfaction in time orientation. applied research in quality of life, 1-15. pavot, w., & diener, e. (2009). review of the satisfaction with life scale. in e. diener (ed.), assessing wellbeing (pp. 101-117). dordrecht, netherland: springer. 118 w. heo et al. / financial services review 29 (2021) 101–119 perry, v. g., & morris, m. d. (2005). who is in control? the role of self-perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39, 299-313. prawitz, a. d., & cohart, j. (2016). financial management competency, financial resources, locus of control, and financial wellness. journal of financial counseling and planning, 27, 142-157. rejda, g. e., & mcnamara, m. j. (2016). principles of risk management and insurance (13th ed.). hoboken, nj: peason. robb, c. a. (2017). college student financial stress: are the kids alright? journal of family and economic issues, 38, 514-527. robb, c. a., & woodyard, a. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning, 22, 60-70. rogers, w. h. (1993). sg16.4: comparison of nbreg and glm for negative binomial. stata technical bulletin reprints (vol. 3, pp. 82-84). college station, tx: stata press. rotter, j. b. (1966). generalized expectancies for internal versus external control of reinforcement. psychological monographs: general and applied, 80, 1-28. song, i. j., park, h., park, n., & heo, w. (2019). the effect of experiencing a death on life insurance ownership. journal of behavioral and experimental finance, 22, 170-176. tannahill, b. a. (2012). life insurance’s role in retirement planning. journal of financial services professionals, 66, 33-35. tennyson, s. (2011). consumers’ insurance literacy: evidence from survey data. financial services review, 20, 165-179. thoyts, r. (2010). insurance theory and practice. new york, ny: routledge. tokunaga, h. (1993). the use and abuse of consumer credit: application of psychological theory and research. journal of economic psychology, 14, 285-316. weesie, j. (1999). sg121: seemingly unrelated estimation and the cluster-adjusted sandwich estimator. stata technical bulletin reprints (vol. 9, pp. 231-248). college station, tx: stata press white, h. l. (1982). maximum likelihood estimation of misspecified models. econometrica, 50, 1-25. whitelaw, e. r. (2014). how to relieve the plight of unskilled irrevocable life insurance trust trustees unfamiliar with their duties. journal of financial service professionals, 68, 44-49. xiao, j. j. (1996). effects of family income and life cycle stages on financial asset ownership. journal of financial counseling and planning, 7, 21-30. xiao, j. j., chen, c., & chen, f. (2014). consumer financial capability and financial satisfaction. social indicators research, 118, 415-432. xiao, j. j., & o’neill, b. (2018). propensity to plan, financial capability, and financial satisfaction. international journal of consumer studies, 42, 501-512. zietz, e. n. (2003). an examination of the demand for life insurance. risk management and insurance review, 6, 159-191. w. heo et al. / financial services review 29 (2021) 101–119 119 from the editor this issue contains volume 28 issue 3 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “career and education choice as central elements of long-term financial planning” is coauthored by inga timmerman at california state university, northridge and nikanor volkov at mercer university. in this paper, the authors demonstrate that human capital accounts for the majority of an individual’s wealth portfolio. they include human capital and show that the choice of career and education level has a significant effect on the sharpe ratio of an individual’s overall wealth portfolio. they illustrate how an individual can perform a simple npv-like analysis in the process of career planning. the second article “selecting a social security age to balance consumption and risk” is coauthored by colonel barry cobb and colonel jeffrey s. smith, both at virginia military institute. the authors use monte carlo simulation to determine the maximum consumption given retirement at age 62, initial wealth, risk tolerance, and social security decision. the authors show that, conditional on retirement at age 62, initial consumption is always maximized by taking social security no later than age 63; it also results in the highest simulated ending wealth at death, and the lowest amount of simulated time living on just social security. the third article, “a theoretical examination of cash-back credit cards and their effect on consumer spending” is coauthored by noah macdonald and brent evans, both at georgia college & state university. in this paper, the authors construct models to analyze the use of cash-back cards demonstrating that cash-back cards increase spending (and thus, reduce savings) for some consumers. while prior research focuses on behavioral issues related to credit cards, this research shows that some consumers will rationally increase spending when using a cash-back credit card in lieu of cash. the final article, “develop a retirement plan and stick to it: it will improve both your attitude and behavior with money” is authored by gizelle d. willows at the university of cape town. in this study, the author explores the relationship between financial literacy, behavior and attitude, and retirement savings decisions. members of a south african tertiary institution’s retirement fund were surveyed, and multi-stage multivariate regression and mediation 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) v–vi analyses showed that developing and conforming to a retirement plan positively influenced financial attitude and behavior. these results indicate that interventions should focus on the specific behaviors which drive retirement planning, rather than financial literacy in isolation. the author also shows that use of formal tools, such as consulting with a financial planner, increases the relative risk of successful retirement planning. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi s. michelson / financial services review 28 (2020) v–vi pii: 1057-0810(91)90026-u financial services review, i(2): 87-99 copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. personal financial planning and the allocation of disposable wealth amy v. puelz robert puelz in the process ofpersonalfinancial planning individuals are confronted with a time dependent wealth allocation problem. oftentimes the solution involves selectingjnancial products based on objective criteria, for example, product cost and expected return. while objective criteria are important to the selection process, an individual’s subjective valuation of all criteria, objective and subjective, relevant to the decision plays the crucial role. a goal programming model parameterized by the analytical hierarchy process is presented to determine the allocation of an individual 5 disposable wealth to present andfiture consumption bundles and investable assets, conditional on the preference ordering of the individual. the personal financial planning process has in recent years increased in complexity, requiring individuals to seek outside assistance in developing a plan for their financial needs (cooper and ulivi, 1983). yet, to date there has been relatively little rigorous academic research addressing the personal financial planning func tion. 1 one important dimension of personal financial planning involves the alloca tion of an individual’s disposable wealth to investable assets, and present and future consumption bundles. the process of personal financial planning, most often undertaken with the counsel of a financial planner, involves an individual making subjective assessments with regard to the risk and expected return of alternative assets, and consumption preferences. for example, a commonly used tool by financial planners is a risk profile evaluation which crudely gauges an individual’s level of risk tolerance. in the personal financial planning literature, however, there is no decision model that fully integrates an individual’s subjective valuation of all amy v. puelz and robert puelz l department of management information systems and decision sciences, and department of finance, insurance, and real estate, fogelman college of business and economics, memphis state university, memphis, tn 38 152. 88 financial services review, l(2) 1991 ob.jectives and constraints relevant to this time dependent wealth allocation problem. this paper posits such a decision paradigm. we present a multiple-objective model that generates a portfolio consistent with an individual’s preference toward current and future consumption and desired portfolio characteristics. our model is based on multi-objective goal programming (gp) with the parameters of the model derived through the analytical hierarchy process (ahp). thus, we offer a determi nate model that unifies the characteristics of the alternative portfolios with an individual’s preference set to ascertain an individual’s optimal allocation of disposa ble wealth. in section ii we provide a brief description of the two decision-making tech niques employed in our model, goal programming and the analytical hierarchy process. in section iii we outline our financial planning model and in section iv we discuss the model implementation and illustrate its use. ii. decision-makingmethodologies a. goal programming the concept of goal programming (gp) was introduced by charnes and cooper (196 1) and applied to the decision-making environment by lee (1972). gp is best viewed in the context of simon’s (1955) seminal work on the “satisficing” nature of managers. in an environment where conflicting goals cannot be achieved simultaneously the solution that achieves a set of goals to the manager’s satisfaction is implemented. in a gp model goals are formulated and the underachievement of these goals is minimized based on the relative priority or weighting of the goal. in other words, gp allows for a “satisficing” solution when an optimal solution with all goals attained is not feasible. in financial planning the decision-maker is faced with a number of conflicting goals, for example the maximization of return and liquidity while minimizing risk. therefore, the decision maker must derive a priority weighting scheme to obtain a satisfactory solution where goals are achieved in order of importance. in the context of financial planning and decision-making, gp has been applied to both personal and corporate financial planning problems. recent examples include batson (1989)) puelz and puelz (1989) and kvanli and buckley (1986). in these gp models the deviations from portfolio and/or consumption goals are mini mized at weights established by the decision-maker. these goal weights can be preemptive or relative in nature. the difficulty in utilizing gp to solve the wealth allocation problem for an individual is in the establishment of relative objective and subjective goal weights. we employ the analytical hierarchy process (ahp) to generate the weights for the portfolio and consumption goals in the asset allocation problem. these weights are used to parameterize the gp model which generates the portfolio that minimizes an individual’s goal unattainment. the actual gp model formulation is presented in detail in section iii. personal financial planning and the allocation of disposable wealth 89 b. analytical hierarchy process as suggested in the preceding section, the personal financial planning prob lem has associated with it many subjective and objective criteria important to an individual. for example, the value an individual places on the liquidity characteris tics of the portfolio is subjective while the expected return is objective. comparisons among subjective and objective criteria by an individual are, however, inherently judgmental reflecting a preference weighting after all comparisons have been held. in our model it is this subjective valuation of the relevant criteria in the personal financial planning process which is captured. multiple criteria weights are derived through the ahp by incorporating the individual’s judgment into an objective ratio scale through pair-wise comparisons of preference orderings.2 in the context of the time dependent wealth allocation problem, the objective is investor satisfaction which is dependent on short and long-term consumption goals, and portfolio goals which include the characteristics we model: risk, liquidity and asset preference. future consumption takes the form of m period short-term bundles and an addi tional bundle classified simply as “long-term consumption. ” this hierarchy of goals for the personal financial planning problem is presented in figure 1. the ahp is carried out by the individual through a pair-wise comparison of the goals at each level of the hierarchy. in other words, the individual compares the relative importance of one goal with respect to another which is quantified through a pair-wise comparison scale (see table 1). for example, a ratio of nine between any two goals means that one goal is absolutely more important than the other. all the ratios are stored in a criteria matrix which is positive reciprocal. that is, all diagonal elements equal one, elements above the diagonal range in integers from one to nine and their reciprocals, and thej, i element below the diagonal is the reciprocal of the i,j element above the diagonal. at each level of the hierarchy relative importance weights represented by the eigenvector are determined by the solution to the equation: (q x”z)w = 0 (1) where q is the n x n criteria matrix of pair-wise comparisons over iz goals, zis the iz x iz identity matrix, and x is the eigenvalue which is the solution to the characteris tic polynomial of q.3 for the personal financial planning model the first level of the hierarchy is comprised of two criteria (consumption goals and portfolio goals) which will reveal a 2 x 1 column vector of importance weights among these criteria when the individual undertakes the pair-wise comparison of relative importance. the eigen vectors for the second level of the hierarchy represent important weights for each of the m short-term consumption periods and the long-term consumption period, the n asset alternatives, and the risk and liquidity characteristics, with respect to the two characteristics of the first level. this will yield a (m + n + 3) x 2 matrix of i i i i n v e s t o r s ~ i s f a c t i o h ! 1 i i 1 i i i i i i i 1 c o i w j x p t i o b g o a l s 1 l p o r t f o l i 0 c o a l s i i i i i i i / i i 1 i i i 1 i l l p r e f e r e n c e 1 1 p r e f e r e n c e 1 1 p r e f e r e n c e 1 i i i i 1 fi gu re 1 . g o al h ie ra rc h y fo r fi n an ci al p la n n in g a llo ca ti o n . personal financial planning and the allocation of disposable wealth 91 intensity of irnportuncr table 1. importance scale dejinition i 3 5 7 9 2,4,6,8 reciprocals equal importance weak importance of one over another strong importance of one over another demonstrated importance absolute importance intermediate values between the two adjacent judgments if attribute i has one of the above non-zero numbers assigned when compared with activity j, thenj has the reciprocal value when compared to i. eigenvectors for the second level of the hierarchy. the overall ranking of each of the portfolio alternatives is obtained by pre-multiplying the 2 x 1 column vector of importance weights from the first hierarchy by the (m + n + 3) x 2 matrix of second level eigenvectors. the result is a (m + n + 3) x 1 vector which weights the (m + n + 3) bottom level goals from highest to lowest preference. finally, the reason for the joint ahp/gp process is two-fold. first, if the ahp is used to allocate directly to assets as, for example, in the asset allocation problem addressed by khaksari, kamath, and grieves (1989) or the life insurance selection problem by puelz (199 1) then the individual must directly compare categories based on objective and subjective goals. the typical individual who contracts for personal financial planning services does not have sufficient knowledge to perform such a comparison.4 second, the financial planning model proposed in this paper is based on a multiperiod horizon. in order to incorporate ahp into a multiperiod planning model a different model is required for each future period. by solving the multipe riod problem as a set of independent models the integration of a single portfolio decision for multiple periods is lost. by contrast, in our gp model the multiperiod nature of the personal financial planning decision is captured. iii. thefinancialplanningmodel we present the model in the following manner. first, we formulate feasibility constraints which impose the restrictions that (a) dollars invested each period do not exceed disposable income for that period plus liquidated investment from prior periods, and (b) that the liquidated amount of any asset does not exceed the principal amount plus earnings on that asset. second, we present the portfolio and consump tion goals and identify the deviation variables that are placed in the objective function of the gp model with a weight established from the ahp. the math program minimizes the weighted deviations from the portfolio and consumption 92 financial services review, l(2) 1991 goals subject to the feasibility constraints. the solution entails the quantity of assets which are selected for purchase and sale each period, consistent with the individ ual’s preference ordering. the notation used throughout the gp model formulation is presented in table 2. a. model constraints model constraints are those conditions that must be met in order for a feasible portfolio to be generated. the first set of constraints equates for each of the m periods in the short-term horizon the net dollar investment, c(rbhrfynm t,~,y,,), to disposable wealth for that period, a,n, less consumption for that period, z,,,5 g [r~j,rrn rsn y,,,) + z,, = 4 v m=ltom (2) n=l the second set of constraints assure that the amount divested from an asset in any period, y,,,/+ ,, is not greater than the principal investment and earnings on that asset, ,,t , [l + 4-‘“+i km y,,)l 2 y,, j+, v j = 1 tom1 andn = 1 ton. table 2. model notation (3) x ,m = y ,111, = 6, = 7&l = ‘t/h = o,, = a,, = p,, = a,,, = p,” = p,. = l = r = st, = (i, = d$+ = w, = n = m = the amount invested in asset n at the beginning of period m. the amount divested in asset n at the beginning of period m. idle funds used for short-term consumption in period rn. the dollar amount necessary to buy $1 .oo of asset n (includes transaction costs). the dollar amount received from the sale of $1 .oo of asset n (includes transaction costs). expected after tax annual return on asset n, liquidity parameter for asset n. risk parameter for asset n. dollars available to invest at the beginning of period m. desired consumption level period m. desired long-term consumption level. desired maximum portfolio average liquidity. desired maximum portfolio average risk. desired maximum (minimum) percent held of asset n. negative deviation from a particular goal level in goal constraint i. positive deviation from a particular goal level in goal constraint i. the ahp generated weight attached to the achievement of goal i. number of assets. number of periods in the short-term horizon. personal financial planning and the allocation of disposable wealth 93 b. model goals goals are classified in the personal financial planning model as consumption and portfolio goals. short-term consumption goals are anticipated cash expendi tures in the short-term horizon, and might include such things as the purchase of a car, or a planned vacation. the dollars required for each period’s short-term consumption is estimated by the individual with the aid of the financial planner and serves as desired short-term consumption goal attainment levels. the set of constraints for the short-term consumption goals are formulated as follows: z,,, + d,di+ = p,, v m=ltom (4) where p,,, is the desired short-term consumption and z,,, is cash available for consumption during period m. deviations from these goals are penalized at a level established in the ahp framework. in other words, if cash available during any period falls below the desired level, p,,,, penalties are assessed due to underachieve ment of the goal. in the gp model the negative deviation variable, dim , indicates the amount below the desired consumption level and is therefore minimized in the objective function at the ahp established weight.6 long-term consumption is simply the individual’s desired savings at the end of the planning period. goal levels are established and are discounted back to period m. pl is the discounted long-term consumption level and the right-hand side is the value of the portfolio at period m. the long-term consumption goal is formulated as (1 + pn)(m-m+‘)(xnm y,,) 1 + d,di’ = pl as in the short-term consumption goals, the negative deviation, die, represents underachievement of the goal and is minimized in the objective function at the weight established in the ahp model.’ in the development of an individual’s portfolio, the financial planner considers not only the individual’s desired short and long-term consumption goals but also the individual’s attitudes towards various portfolio characteristics such as risk and liquidity.8 in addition, the individual may have a preference for certain asset catego ries. these preferences are modeled in the portfolio goals. portfolio liquidity is measured by the liquidity of the underlying assets. the liquidity parameter, (y,,, is the percentage penalty required to immediately liquidate asset n.9 this parameter is estimated by the bid-ask spread for the asset. the goal constraints for liquidity are formulated for every period in the short-term horizon. n c %i ,n+, km ynm) 1 + d,d; = l t/ j=ltom (6) 1 94 financial services review, l(2) 1991 the right-hand side of equation (6) represents the average portfolio liquidity in period m. the left-hand side of equation (6) is the maximum desired portfolio liquidity desired, l, as measured by the percentage penalty required to immediately liquidate a dollar of the portfolio. the positive deviation variable, d:, measures the unattainment of the liquidity goal and is minimized in the objective function of the gp model at the established ahp weight. the risk of a portfolio is measured by the risk of the underlying assets. the asset’s beta, &, is a measure of the non-diversiliable risk associated with that asset. r is the portfolio risk level above which penalties are assessed. the goal constraints for risk are formulated for every period in the short-term horizon analogous to the liquidity goal, the right-hand side of equation (7) represents the average portfolio risk and the left-hand side is the maximum desired portfolio risk, r. in the objective function the positive deviation, d,:, represents unattainment and is minimized in the objective function at the appropriate ahp weight. finally, asset preference goals are in the form of the desired minimum or maximum proportion of the portfolio to be allocated to asset ~1. for example an individual may desire that at least twenty-percent of the portfolio be placed in growth stocks. f: wml yn,,) ,?i= i -e b (x,, ykj i + dipd: = s, v n=l tonandj=l toa (8) k=l m=i 1 the right-hand side of equation (8) is the proportion of the portfolio allocated to asset n and the left-hand side is the maximum or minimum proportion desired for asset n. these goals are formulated for every asset where a maximum or minimum percent is desired. if s,, is the minimum (maximum) proportion to be held of asset ii, then the negative (positive) deviation variable, d, (&), is minimized in the objec tive function at the ahp established weight. the gp objective function is of the form k mznz = c w.d+‘i i i=i (9) personal financial planning and the allocation of disposable wealth 95 where w, is the weight attached to the attainment of goal i. in words the objective function minimizes the sum of the weighted deviations from goal attainment levels. the deviational variables in equation (9) are those selected from each set of constraints that represent goal unattainment. the w, values are generated through the ahp process as discussed in section iib. in summary, the gp model generates the portfolio plan for the m-period short term horizon. the plan consists of the periodic amounts to buy and sell of each asset. the mode1 constraints assure that dollars invested each period do not exceed disposable income for that period plus liquidated investments from prior periods, and that the liquidated amount of any asset does not exceed the principal investment and earnings on that asset. the goal constraints consider short and long-term consumption and the portfolio characteristics of risk, liquidity and asset preference. in each goal constraint, the deviation variable representing underachievement of the goal is minimized in the objective function at the ahp established weights. iv. modelimplementation two types of data are required for our model: investment alternatives and the individual’s ahp established goal weights. the data on available investment vehi cles is maintained by the financial planner and contains estimates on expected returns, liquidity parameters, risk parameters, and all relevant transaction costs, for each investment option. investment options are grouped into broad homogeneous categories (i.e., growth stocks or insured municipal bonds). this is preferred over individual security investment options for two reasons. first, the amount of data to be maintained and the size of the goal programming mode1 are greatly reduced. second, the model will generate the portfolio in terms of broad investment catego ries, giving the planner and individual flexibility in selecting individual assets or mutual funds within the established broad categories. to illustrate the use of the ahp/gp framework, an example portfolio is structured. we consider ten asset categories that represent the choice set of invest ment alternatives (see table 4). in this example, these categories were determined by the authors to be important, however other categories may be important to another, and the choice set would be altered to reflect the addition or subtraction of such categories. we assume a short-term annual planning horizon of three years with projected disposable wealth (including current savings) for years one, two and three at $100,000, $16,000, and $17,000 respectively. the individual, with the aid of the financial planner, sets consumption and portfolio goal levels. the goal levels used in this example are presented in table 3. through the ahp, the individual performs a pair-wise comparison of goals at each level of the hierarchy represented in figure 1. for example, at the bottom level of the hierarchy, the individual compares the relative importance of the short-term consumption for years one and two, years one and three, and years two and three using the importance scale in table 1. the weights established in the ahp for this example are presented in table 3 . lo for example, the individual places primary 96 financial services review, l(2) 1991 table 3. example problem-goal levels and ahp established weights goal categories consumption short-term year 1 p, =$lo,ooo year 2 pz =$ 9,000 year 3 p3 = $20,000 ahp established relative weights* ,161 ,064 ,032 long-term p,=$130,000 ,201 portfolio liquidity l = .02 risk r=.65 asset preference .030 ,167 proportion in long-term growth > 5 % proportion in tax-exempt < 10% proportion in government and u.s. agencies < 7.5% proportion in low-risk corporate bonds < 15 % proportion in precious metals < 10% nore: *consumption weights are normalized by dividing p, or pc ,069 .069 .069 .069 .069 importance on long-term consumption (relative weight = .201), and relatively more importance on the risk of the portfolio than the portfolio’s liquidity. after all pair-wise comparisons are performed the ahp generated weights are incorporated into the objective function of the gp model in equation (9) and the gp model formulated in equations (2) through (9) is solved to generate the financial plan that maximizes the goal attainment level for the individual. all relevant transaction and tax cost are incorporated in the model. the optimal allocation of disposable wealth to short and long-term consumption bundles, and investable assets for this example is detailed in table 4. the dollar amounts of assets bought and sold in each year are listed in columns titled assets purchased and assets sold. initial investments in the first year are high because of the assumption that the individual has substantial savings available to invest the first year. each year, thereafter, the disposable income is assumed to come only from current income. assets are liquidated in year three to achieve the short term consumption goal during that year. it should be noted that five assets are liquidated during year three in order to achieve the liquidity and risk goals for that year. if the individual desires that transactions be in larger blocks, then the solution procedure for the gp model would be altered. i i the percentage goal attainments for this example, as measured by the actual level of attainment divided by the desired level of attainment for each goal, are presented in table 5. personal financial planning and the allocation of disposable wealth 97 table 4. example problem-model output asset category year amount purchased amount sold growth, small companies 1 $26,166 2 0 3 0 $ 0 2,257 growth, long-term 1 2 3 4,354 313 0 0 181 income 1 11,973 2 826 0 3 0 537 tax-exempt tax-exempt, high-risk 1 8,708 2 561 0 3 0 437 0 0 0 1 2 3 0 0 1 6,531 2 638 0 3 0 72 government and u.s. agencies corporate bonds, high-risk 1 2 3 0 0 0 0 0 corporate bonds, low-risk 1 20,639 2 2,508 0 3 347 0 1 8,708 2 1,001 0 3 79 0 precious metals real estate 1 2 3 0 0 0 0 0 both long-term consumption and asset preference goals have attainment levels below 100%. this is due to the fact that the gp technique maximizes the weighted attainment of all goals. in other words, the generated portfolio plan maximizes the individual’s overall satisfaction. financial services review, l(2) 1991 table 5. example problem-goal attainment prrcrntage goal attainment consumption short-term long-term 100% 91% portfolio liquidity risk asset preference* 100% 100% 89% now: *average attainment for all asset preference goals vi conclusion we have presented a multi-period model to allocate an individual’s disposable wealth to short and long-term consumption, and investable assets through the use of goal programming and the analytical hierarchy process. our model integrates an individual’s subjective valuation of all relevant goals associated with the allocation problem into a math programming model which generates an allocation solution consistent with the individual’s consumption and portfolio goals. we illustrated the model’s operation for a particular individual’s valuation of consumption goals, and characteristics of ten asset categories. the model, however, is sufficiently flexible to accompany a broader range of goals and assets. notes i. 2. 3. 4. 5. 6. 7. 8. 9. see cohen (1988) for a detailed literature review in the area. an exposition on ahp can be found in saaty (1980) or saaty and vargas (1982). the characteristic polynomial of q is the determinant of q xi. for example, an individual may be asked to assign relative weights to the liquidity characteristics of a municipal bond and a precious metal. disposable wealth is defined independent of sufficient funds to pay for all essential consumption and to abstain a financial emergency, and adequate insurance cover for common perils such as pre-mature death, disability, and property-liability losses. the ahp established weight is normalized by dividing by p,,,. the ahp established weight is normalized by dividing by p,.. the client may have other concerns such as capital appreciation, current income, inflation protection, tax reduction, etc. these preferences may be incorporated in the model in the same fashion as risk and liquidity. the liquidity parameter could be a percentage penalty required to immediately liquidate an asset or a measure of the number of years required to liquidate an asset without penalty. since short term consumption goals are already incorporated into the model the liquidation of the portfolio should only occur in the event of an emergency or unforseen event which would require personal financial planning and the allocation of disposable wealth 99 immediate liquidation. therefore, it is more appropriate to measure liquidity by the percentage penalty to immediately liquidate. io. expert choice software was used to carry out the pairwise comparison and solve for the relative weights. ii. in this case an integer math programming algorithm would be required to solve the gp model. references batson, r.g. 1989. “financial planning using goal programming,” bng range planning, 22: i l2 120. charnes, a., and w.w. cooper. 1961. management models and industrial application of linear programming. new york: john wiley and sons. cohen, n.g. 1988. “basic research in personal financial planning: needs and prospects.” working paper, george washington university. cooper, r.w., and r. ulivi. 1983. “comprehensive financial planning: a survey of consumer opinions,” journal of the american society of clu and chfc, 4: 40-46. expert choice. 1988. mclean, va: decision support software. khaksari s., r. kamath, and r. grieves. 1989. “a new approach to determining optimal portfolio mix,” journal of portfblio management, spring: 43-49. kvanli. a.h., and j.j. buckley. 1986. “on the use of u-shaped penalty functions for deriving a satisfactory financial plan utilizing goal programming, ” journal of business research, 14: i-18. lee, s.m. 1972. goal programming for decision analysis. philadelphia: auerbach publishers. puelz, a.v., and r. puelz. 1989. “personal financial planning: an interactive goal programming model using u-shaped penalty functions,” proceedings of the decision sciences institute, i : 327-329. puelz, r. 199 i. “a process for selecting life insurance contracts,” journal of risk and insurance, 53: 138-146. saaty, t.l. 1980. 7’he analytic hierarchy process. new york: mcgraw-hill. saaty, t.l. 1986. “axiomatic foundations of the analytical hierarchy process,” management science, 32: 841-855. saaty, t.l., and l.g. vargas. 1982. the lagic of priorities. boston: kluwer-nijhoff. simon, h.a. 1955. “a behavioral model of rational choice,” quarterly journal of economics, 69: 99-l 18. pii: 1057-0810(91)90006-k financial services review, 1(1):354! copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. determinants of household check writing: the impacts of the use of electronic banking services and alternative pricing of services neil b. murphy in recent years, there has been a simultaneous deregulation of interest ceilings on household deposits and a dramatic deployment of automatic teller machines (atm) by the banking industry. at one time, there was a belief that the u.s. would evolve into a “checkless society” because of the development of electronic funds transfer systems such as the atm. that has clearly not happened. the purpose of this paper is to estimate the impacts of both the use of a tm and the change in pricing due to deregulation on household check writing. the source of the data is a survey of households conducted by the board of governors of the federal reserve system. the results of the tests indicate that use of electronic banking services had no discernible impact on check writing while different methods of pricing checking account services did have a substantial impact on check writing. it has been almost 20 years since the prognostication of a “checkless society,“and there are still some 47 billion checks written each year in the united states. at the same time, the banking industry has deployed over 60,000 automatic teller machines (atms). the purpose of this paper is to estimate the impacts of the pricing of checking account services and the usage of electronic banking devices on the number of checks written by households in the united states. in a recent paper, humphrey and berger indicate that the lack of response on the part of payment system users is largely due to a divergence between the social and private costs of alternative payments media (humphrey and berger, 1988). specifically, the users of checks do not bear the full social cost of that method of making payment. humphrey and berger estimate that the largest divergence occurs for business users in which the float benefit exceeds the cost of processing the check. however, they indicate that households are not as able neil b. murphy l professor of finance, department of finance, school of business, virginia commonwealth university, 1015 university avenue, box 4000, richmond, va 23284-4000. 36 financial services review, l(1) 1991 to capture such benefits, and their usage patterns should reflect a more efficient use of resources. this is due to the nature of the fixed costs involved in a cash management system, first noted by baumol(l952). there is one problem with that notion. that is, the pricing of checking services in many cases does not reflect the marginal costs of providing the service, and, thus, there may not be any incentive for households to economize. a primary reason for this underpricing is a holdover from the era of regulated interest ceilings in which non-price competition emerged in the form of reduced or no service charges in lieu of explicit interest payments to depositors (spellman, 1982). at least part of the reason for the deployment of atms was a hope that households would economize on check writing. earlier studies did not indicate much, if any, support for the notion that atm usage affected check writing (murphy, 1979). more important, perhaps, was a desire to provide transaction services that displace lobby traffic in branch offices (murphy, 1990). the branch office delivery system was built up during the era of regulated interest ceilings and may be viewed as an expensive provision of convenience in lieu of explicit interest payments (taggart, 1978). the deployment of atms may then be viewed as an attempt to economize on the provision of convenience, substituting more cost effective atms for branch offices. with the introduction of now accounts in the 1970s in new england and in 1980 nationwide, along with the subsequent deregulation of all household deposit interest payments, the stage was set for a restructuring of pricing of a number of household financial services that should lead to incentives that result in a more efficient allocation of resources. banks pay explicit interest to attract deposits, creating incentives for them to recover directly the cost of providing payments services and reducing the incentives to proliferate branch offices. instead, less costly atms can be deployed. customers can then be confronted with a combination of interest payments, service charges, and delivery systems that allow them to select the combination of balances, checks written, and transactions that best suits their needs. however, the transition is not instantaneous (berger and humphrey, 1986). banks do not shut down all their redundant branch offices nor do they immediately impose full cost service charges. rather, they selectively close offices, reduce the rate of growth of new offices, and slowly increase service charges (canner and kurtz, 1985). this results in an environment with many different pricing arrangements (dunham, 1983). therefore it is possible to obtain a cross-section of households using different configurations of prices and electronic banking services. this permits tests to determine the impact of different pricing as well as the usage of atms on check writing. in section i, the data source and the model development are discussed while the results of the analysis are presented in section ii. section iii is the summary and conclusion. determinants of household check writing 37 i. data source and model development between may and august 1984, the survey of currency and transaction account usage was conducted by the survey research center of the university of michigan. that survey was commissioned by the board of governors of the federal reserve system. the descriptive results of the survey as well as the survey methods were discussed in some detail by avery, elliehausen, kennickell and spindt (1986). the original sample contained 1,946 interviews from a randomly selected sample of 2,500 families residing in the united states. the survey was a personal interview in which respondents reviewed the details of their currency and transactions account usage from their own records. the respondent was either the head of the family or a financially knowledgeable spouse. for purposes of this paper, the sample size was smaller because some households do not have checking accounts. the resulting sample size for the analysis is 1,596 households. the model is based upon the assumption that households will attempt to minimize the total cost of making payments. this includes not only the explicit costs charged by the bank, but also the costs of transportation and the use of time. considerations of acceptability, safety, control, and record-keeping also enter into the decision process in selecting a method of making payment. it is important to note that the benefits from cost minimizing behavior are relatively small for most households. in the corporate sector, there are sophisticated techniques and services available to permit corporations to collect payments quickly, control disbursements, and invest any surplus funds in the money markets. moreover, there is a professional organization, the national corporate cash managers association, and a professional designation, the certified cash manager (ccm) for corporate cash managers, implying substantial benefits from aggressive management of payments costs. thus, more sophisticated approaches to household payments management will develop more slowly, but the forces are clearly the same. the opportunity cost of household time, the availability of alternative methods of making payment, and the movement, albeit slow, to explicit pricing should all push in the direction of cost minimizing behavior on the part of households (murphy, 1977). the model contains three sets of independent variables: 1. demographic variables 2. payment system variables 3. pricing variables the dependent variable in the model is the number of checks written per month (or its logarithmic transform). there are a number of demographic variables that reflect income, age, marital status, education, etc. these variables reflect factors that affect 38 financial services review, l(1) 1991 household payments costs. for example, income is a measure of the opportunity cost of time in making payments. education is related to the opportunity cost of time as well as the ability to understand and choose payments methods that minimize cost. these variables also hold constant the influence of all other factors and allow the analysis to be focussed on payment system and pricing variables. those variables that were found to be statistically significant in explaining check writing are household income, marital and employment status, some educational categories, and the sex of the head of the household. it is expected that those with higher incomes will purchase more goods and services that will be reflected in more payments by check, although it is not clear what the nature of the relationship would be. while total consumption expenditure may show a strong proportional increase with changes in income, total checks written would not necessarily increase at the same rate as more transactions and higher amounts per transaction simultaneously occur as income rises. hence, a positive coefficient is expected, but the elasticity of checks written with respect to income would likely be less than the elasticity of consumption with respect to income. employment status and education reflect the need and ability to control expenditures and maintain payments records. hence, it would be expected that more checks would be written. there is no economic basis to the finding that gender affects the number of checks written, but the estimated coefficient is statistically significant. there are three payments system variables: 1. the use of atms by the household. 2. the use of direct deposit by the household. 3. the number of credit cards used by the household. approximately 25% of all households were active users of atms. whether or not check writing should be affected is not obvious apriori. most people use atms to obtain cash or to deposit funds in their account. for example, if cash were normally obtained when a paycheck is received, and the recipient splits the amount of the paycheck between a deposit and cash, there would be no necessary relationship between check writing and atm use if the household merely changed the location of the split deposit from the teller line to the atm. almost 21% of all households in the sample received either their paycheck or social security check via direct deposit. it is conceivable that such an arrangement may increase the amount of check writing as households must now access their accounts for cash rather than taking it at the time such a payment (check) is received. finally, the number of credit cards may affect check writing. since credit cards are used at the point of sale, they may be a substitute for either cash or checks. previous work has suggested that credit cards primarily substitute for cash and as such increase the number of checks written (mandell, 1971; murphy determinants of household check writing 39 table 1. variables used estimating model variable description lncks lnhhinc marr&emp hsdipl colldgr sex dirdep crcd atmuse intchk chkfee natural logarithm of the number of checks written per month per household natural logarithm of the annual household income takes a value of 1 if head of household is married and employed, 0 for all others takes a value of 1 if head of household has high school diploma, 0 for all others takes a value of 1 if head of household had a college degree, 0 for all others takes a value of 1 if payroll or social security payments are received via direct deposit takes a value of 1 if payroll or social security payments are received via direct deposit number of credit cards used by household takes a value of 1 if household uses atm, 0 for all others takes a value of 1 if interest is received on checking account, 0 for all others takes a value of 1 if service charges are based upon the number of checks written, 0 for all others 1979). that is, more checks are written to pay the credit card bills than are displaced at the point of sale. since the implementation of nationwide now accounts, households have faced an array of pricing arrangements. these involve interest payments, service charges imposed on a flat fee basis when a certain balance is not maintained, service charges tied to the number of checks written sometimes contingent upon a certain balance being maintained, and various combinations thereof. in some cases, the same bank offers as many as five or six configurations (dunham, 1983). because so many configurations exist, it is possible to observe differing costs to check writing in the resulting sample. for purposes of this study, two pricing variables were specified. 1. the payment of interest was noted. if a household receives interest on deposit balances, there is some incentive to keep funds invested for as long as possible. since each check removes interest bearing funds from the account, there may be an incentive to write fewer checks. of course, the same number of checks may be written for a different time pattern. 40 financial services review, l(1) 1991 2. the basing of service charges on the number of checks written was noted also. holding constant all other factors, it is expected that households that are charged on a per check basis would write fewer checks than households with either no service charges or flat rate service charges. on the one hand banks wish to institute service charges that recover costs. this would lead to service charges based on activity. on the other hand, banks wish to simplify account pricing so as not confuse and annoy customers, which would lead to flat service charges. the variables used in the model are described in table 1. ii. statistical results the model is estimated using ordinary least squares with both the dependent variable and household income in logarithmic form. this specification gives the best statistical results and is consistent with the results of previous studies (murphy, 1979). moreover, the dependent variable is measured in number of checks per month while the income variable is measured in dollars per year. in table 2, the means and standard deviations of the variables are shown. the logarithmic transformation changes the dimensions of the variables so that they are more comparable, and the coefficient is interpreted as an elasticity. as shown in table 3, the overall performance of the model indicates that check writing is subject to random determinants or is influenced by variables not measured here. the model is statistically significant as indicated by the f statistic of 36.5. however, the coefficient of determination (r-square) is only .1874. table 2. means and standard deviations of variables in equation variable mean standard deviation lncks 13.2701 2.3960 lnhhinc 20,075.46 2.1908 marr&emp .5334 .4990 hsdipl .5143 .4999 coldgr .2835 .4508 sex .7842 .4114 dirdep .2102 .4076 crcd 3.8042 4.2439 atmuse .2533 .4350 intchk .3196 .4664 chkfee .1671 .3731 note: geometric mean is computed for lncks and lnhhinc since variable is measured in natural logarithms. determinants of household check writing 41 table 3. regression results for determinants of checks written (dependent variable is lncks) variable intercept lnhhinc marr&emp coefficient (t-statistic) -.467 (-1.422) .264 (7.434) .268 hsdipl colldgr sex dirdep atmuse crcd intchk chkfee r’ -1874 f value 36.542 n 1,596 df 1,585 (4.831) ,249 (3.776) .303 (4.102) -.152 (-2.375) .i32 (2.762) .olo (.219) .034 (6.442) -.065 (-1.557) -.121 (-2.327) as expected, the demographic variable having the most explanatory power is the logarithm of household annual income with an elasticity of .26 and a reported t-statistic of 7.4. the other demographic variables are statistically significant at the 5% level. employment and marital status, educational levels and the sex of the head of household all have impacts on the number of checks written. for the payment system variables, there appears to be no substitution for checks written. on the contrary, both direct deposit and credit cards increased check writing. apparently the direct deposit creates a need for more check writing to obtain cash. the use of atms had no impact at all on the number of checks written. it was thought that the combination of direct deposit and atm use may have a different impact than either one separately. that is, the household would have payments directly deposited and then use the atm to 42 financial services review, l(1) 1991 obtain cash, removing the need to write checks for cash. when this was tested, the results were not affected. the question of how cash is obtained and the role of the atm in that overall process are interesting questions beyond the scope of this paper. the two pricing variables in the model have negative coefficients, and the coefficient on the check-based pricing variable is statistically significant at the 5% level. while the effect of any particular price is not known, the impact of shifting to any service charge based upon activity is substantial. furthermore, the impact of moving to pricing based upon activity cost is understated since the observed per item charges are much less than the social or private costs of processing a check estimated by humphrey and berger (1988). because the equation has the dependent variable in logarithmic form and the pricing variable in binary form, it is not possible to interpret directly the regression coefficient for that variable. therefore, the expected value of the dependent variable was calculated with representative values for the other independent variables and the binary pricing value taking a value of zero and then one. for the household with the geometric mean annual income, male head of household employed and married, with a college degree, with neither direct deposit nor atm usage, with the mean number of credit cards, and interest on his checking account, a shift to check based service charges results in an expected reduction of checks written per month of 1.6, a percentage reduction of 11.4%. that is, when all variables are held constant at the values indicated above, the expected value of the dependent variable is 13.9 checks written when service charges are not based upon activity while it is reduced to 12.3 checks when the service charge is based upon activity. iii. summaryand conclusion in their paper, humphrey and berger note that “consumer checks only have a small float benefit, so that the private costs are positive ($.72) and almost as high as the social costs ($.79), indicating only a minor market failure” (1988). however, the pricing of checking account services does not reflect the marginal (or average) cost of processing a check. in many cases, the service charge is not related at all to the number of checks, and, if it is, the charge is likely to be much less than $.79 (or $.72). in this paper, the determinants of check writing by households were estimated using a cross section of 1,596 households. the results suggest that electronic banking usage has contributed little to the reduction of check writing, but the use of a charge per check has a significant and substantial impact on household check writing. since households were estimated to have written 25.8 billion of the 47 billion checks written in 1987 (humphrey and berger, 1988), the potential for economizing is substantial. moreover, the results have determinants of household check writing 43 implications for the introduction of other payments services. the use of an atm represents households dealing with their own bank rather than interacting with third parties. since households have been shown to reduce their use of checks in response to pricing, it should follow that they will substitute other third-party payments services if the relative prices provide the proper incentive. of course, rational explicit pricing and true cost savings are required before this can happen. while the results from one cross-section should not be extrapolated without caution, the results suggest that the adoption of pricing based on the cost of producing checking services is an attractive candidate for improved allocation of resources in the use of the payments system. it certainly merits attention on a priori grounds as well as being supported by the results of this paper. acknowledgments: this paper is based upon the survey of currency and transaction account usage conducted by the board of governors of the federal reserve system. most of the work was conducted while the author was senior economist (visiting) in the monetary and financial studies section of the board of governors of the federal reserve system. i am indebted to robert b. avery and arthur b. kenickell for help in working with the survey and to allen n. berger, myron l. kwast, and lewis mandell for helpful comments. roberto sella provided able research assistance. the opinions expressed are those of the author and do not necessarily represent those of the board of governors of the federal reserve system. references avery, robert b., gregory e. elliehausen, arthur b. kennickell, and paul a. spindt. 1986. “the use of cash and transactions accounts by american families,” federal reserve bulletin, february: 87-108. baumol, w.j. 1952. “the transactions demand for cash,” quarterly journal of economics, november, 66: 545-556. berger, allen n., and david b. humphrey. 1986. “the role of interstate banking in the diffusion of electronic payments technology,” pp. 13-52 in colin lawrence and robert p. shay (eds.), technological innovation, regulation, and the monetary economy. cambridge, ma: ballinger. canner, glenn b., and robert d. kurtz. 1985. “service charges as a source of bank income and their impact on consumers.” staff study no. 145, board of governors of the federal reserve system. dunham, constance. 1983. “unravelling the complexity of now account pricing,” new england economic review, may/ june: 30-45. humphrey, david b., and allen n. berger. 1988. “market failure and resource use: economic incentives to use different payment instruments,” pp. 45-86 in david b. humphrey (ed.), i?ke u.s. payments system: efficiency, risk, and the role of the federal reserve. boston: kluwer. 44 financial services review, l(1) 1991 mandell, lewis. 1972. credit card use in the united states. ann arbor, mi: university of michigan press. murphy, neil b. 1977. “bank credit cards, now accounts, and electronic funds transfer: the role of explicit pricing,” magazine of bank administration, august: 111-l 14. murphy, neil b. 1979. “the impact of payments systems innovation on consumer check writing,” journal of retail banking, 1: 27-32. murphy, neil b. 1990. “one-stop financial service delivery opportunities in a world of automated self-service banking: some evidence from the united states,” pp. 283-292 in e.p.m. gardener, the future of financial systems and services. london: macmillan and co. spellman, lewis j. 1982. 7?ze depository firm and industry: theory, history and regulation. new york: academic press. taggart, robert a. jr. 1978. “effects of deposit rate ceilings: the evidence from massachusetts savings banks,” journal of money, credit and banking, 10: 139-l 57. toward constructing tax efficient withdrawal strategies for retirees with traditional 401(k)/iras, roth 401(k)/iras, and taxable accounts james dilellioa,*, daniel ostrovb apepperdine graziadio school of business, 24255 pacific coast highway, malibu, ca 90263, usa bsanta clara university, 500 el camino real, santa clara, ca 95053-0290, usa abstract we construct an algorithm for u.s. retirees that computes individualized tax efficient annual withdrawals from tax-deferred, tax-exempt, and taxable accounts. our algorithm applies a new approach using information from all years that generates an individualized strategy, in contrast to most previous approaches that chronologically generate a suboptimal strategy. results equal or improve the chronological “naı̈ve” withdrawal strategies advocated by many financial institutions, as well as the chronological “informed” strategies advanced by academics. our approach allows us to determine the optimal switching times between taxexempt and taxable account consumption, as well as between tax-deferred and taxable account consumption. it also allows us to understand the significant impact of an heir’s tax and withdrawal rate on a retiree’s optimal withdrawal strategy. our model, which can work to optimize either portfolio longevity or the bequest to an heir, accommodates many salient tax code features, including dividends, different taxable lots, and required minimum distributions. © 2020 academy of financial services. all rights reserved. jel classification: g11; h21 keywords: retirement income; tax efficiency; optimization 1. introduction u.s. retirees generally have their equity investments1 in three types of accounts: (1) taxdeferred accounts (tdas) like traditional iras or traditional 401(k)s, (2) tax-exempt accounts like roth iras or roth 401(k)s, and (3) taxable accounts. these three accounts are *corresponding author: tel.: +1-714-403-0085; fax: 949-223-2575. e-mail address: james.dilellio@pepperdine.edu 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 67–95 governed by significantly different tax rules. for example, tdas are taxed as income via progressive tax brackets and are subject to required minimum distributions (rmds); roth accounts are not subject to tax; and stock in taxable accounts, when sold, is subject to capital gains taxes, although all capital gains amassed by a retiree are forgiven when stock is inherited by an heir. these differences in tax structure, especially between the taxable account and the two retirement accounts, make answering the important question of how to optimally utilize these three accounts quite complicated. much attention has been given by non-academic institutions, as well as academic researchers, for how to best build up these accounts in preparation for retirement. brown et al. (2017), for example, provides a comprehensive list of academic papers in this area. intrinsic to answering this question, as well as of great interest in its own right, is understanding the somewhat less investigated question of how to best withdraw from these three accounts during retirement. the sequencing of withdrawals that a retiree makes among accounts with varying tax structures can have a significant effect on their portfolio’s longevity. even less investigated is the more general, and more complex question of sequencing withdrawals so as to optimize the usefulness of the retiree’s bequest to an heir, as opposed to the portfolio’s longevity. in this paper, we provide an algorithm that determines a withdrawal strategy that seeks to minimize the effect of taxes on the retiree and, should the retiree wish to make a bequest, the retiree’s heir. our algorithm approach is unusual in that it is an “all-time” approach, as opposed to a “forwards-time” approach or a “backwards-time” approach. forwards-time approaches, which are by far the most common, determine allocation strategies year by year working forwards chronologically in time. backwards-time approaches generally start from a projected date of death for the investor and then work backwards in chronological order. an all-time approach uses information from all years to determine the allocations in every year. in our case, this includes the annual rate of return for the stock, the stock’s dividend rate, the rate of inflation, all tax rates and tax brackets, as well as external annual fixed sources of income. in addition, if the goal is to optimize an heir’s bequest, it includes the projected time at which the retiree will die, the effective marginal tax rate of the heir, and the rate at which the heir will consume inherited tda or roth money. all-time approaches are able to tailor optimization decisions to each individual retiree’s financial situation in a way that forwardstime and backwards-time approaches, by their nature, cannot. the disadvantage of our alltime approach is that it is a complicated algorithm. the full details of this algorithm are found in dilellio and ostrov (2018) and are posted on the academy of financial services (afs) 2018 conference proceedings,2 where it was awarded the best paper sponsored by the cfp board of standards.3 because the details are posted there, we emphasize the novel features and results of the algorithm in this paper, providing only a brief overview of our algorithm’s approach, instead of details, in section 5. 1.1. literature review we provide a quick summary of some forwards-time approaches, which are the most common techniques explored by both non-academics and academics. non-academic advice for retirees’ withdrawal choices, which comes from investment firms, financial advisors, and books on retirement, all agree that retirees should conform with the law and take out 68 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 required minimum distributions (rmds) from their tdas to avoid the significant penalty for not doing so. otherwise, their advice often contradicts each other. for example, one very common category of strategies, termed “naı̈ve” strategies by horan (2006a and 2006b), recommends that retirees completely exhaust one account before moving to the next. the books by solin (2010), rodgers (2009), and lange (2009), suggest sequencing withdrawals so that retirees liquidate taxable accounts first, then tdas, and finally roth accounts. this strategy is also endorsed by large retail investment firms fidelity (see fidelity (2014) and fidelity (2015)) and vanguard (vanguard (2013)). in contrast, other financial authors, such as larimore, lindauer, ferri, and dogu (2011), recommend first liquidating taxable accounts, but then recommend liquidating roth accounts followed finally by tdas. another vanguard paper (jaconetti & bruno, 2008) recommends that the decision on whether to liquidate the tda or roth account immediately following the taxable account should be based on expected future marginal tax rates. coopersmith and sumutka (2011) estimate the suboptimality of these naı̈ve strategies to be approximately 16%. this agrees with dilellio and ostrov (2017), who provide illustrations for which the naı̈ve approaches are 10–26% suboptimal. a second set of strategies, termed “informed” strategies by horan (2006a and 2006b), use tda spending up to the top of a given tax bracket in every year. this is advocated in the book by piper (2013), who suggests filling any remaining consumption needs first with taxable stock, then with roth money, and lastly the tda, if the tda is not already exhausted. we note that there is a different informed strategy for each tax bracket, but each of these informed strategies can be run and then the best of these can then be selected. in this paper, for comparison purposes with our algorithm, we will always use the best of these informed strategies. within the context of withdrawal optimization for just tda and roth accounts, but no taxable accounts, horan’s informed strategies were a considerable step towards increasing a retiree’s portfolio longevity using forwards-time approaches. this seminal work was expanded and investigated further in reichenstein, horan, and jennings (2012). al zaman (2008) extended this approach to the case where a retiree’s goal is to optimize a bequest, in addition to the easier subcase of optimal portfolio longevity. horan’s approach also had a significant impact on forwards-time approaches for increasing the longevity of portfolios that included taxable accounts in addition to tdas and roth accounts. for example, sumutka, sumutka, and coopersmith (2012) compare the effectiveness of a wide variety of naı̈ve and informed withdrawal strategies for an array of portfolios. cook, meyer, and reichenstein (2015) cleverly expanded the field of possible strategies by considering the advantages of using conversions from the tda account in addition to withdrawals for consumption in the retiree’s early years. geisler and hulse (2018) investigate the effect of social security benefits within this approach. backwards-time algorithms include dynamic programing approaches. the paper by brown et al. (2017), for example, applies dynamic programing to tda and roth account withdrawals, but restricts itself to a two-period model due to the computational complexity of this problem, even without a taxable stock account. because of the well-known “curse of dimensionality” in dynamic programing, backwards time approaches cannot accommodate multiple tax lots. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 69 the all-time approach that we have found in the literature uses linear programing algorithms to optimize withdrawals. this line of research begins with ragsdale et al. (1994), which worked with a tda and taxable account. this was extended to include roth accounts and additional features in coopersmith and sumutka (2011) and coopersmith and sumutka (2017), who also concentrated on analyzing the effect of different rates of return in accounts, and in meyer and reichenstein (2013), welch (2016), and welch (2017), who “assign a single rate of return to all accounts to concentrate on the effects of taxes” as we do in this paper. these papers do not determine optimal longevity. instead they optimize the total amount of a bequest, with no distinction between the types of accounts, at a projected time of death. they assume that taxable stock is to be depleted before consuming from the roth account. in all cases, the model assumptions of these papers vary considerably from ours, as does their approach, because of the nature of the linear programing technique, which cannot be applied to non-linear phenomena. dilellio and ostrov (2017) contains an all-time approach that is not based on linear programing and yields an optimal tda and roth account withdrawal strategy for either the goal of optimizing bequests or the subcase of optimizing portfolio longevity. taxable accounts, however, have very different taxation rules than tdas and roth accounts, which necessitated an almost completely new, and far more complex, approach to obtain the all-time approach optimization algorithm in this paper. 1.2. new contributions our model includes working with required minimum distributions (rmds) for tdas, as is the case with most of the academic research above, but it also considers optimization while accounting for a number of features that are rare, if studied at all, in the above literature. among the new contributions of our algorithm for optimizing bequests include • the determination of when the retiree is best off switching from using taxable stock to roth account money for consumption. in the case of optimized longevity, it is best to switch to the roth only after the taxable stock account is depleted. however, when we optimize a bequest, the step up in cost basis to the heir creates an important question of when to stop using taxable stock. • the determination of when the retiree is best off switching from using taxable stock to tda money for consumption within a tax bracket. if taxable stock is going to be consumed, it is best done early to avoid losses from taxation on dividends. should a bequest be involved, however, there is the additional complication that it may be better for the retiree to hold onto the taxable stock to take advantage of the step up in cost basis to the heir. • the effect on the optimal strategy of taking into account the effective tax rate of the heir. • the effect on the optimal strategy of taking into account the rate at which the heir withdraws money, which changes the worth of inherited taxable stock vs. inherited tda and roth accounts, which yield further tax sheltering. • quantifying the effect of a retiree’s stock dividends on the bequest size. via the functions l tð þ and u tð þ, which we explain later, we are also able to accommodate additional projected fixed sources of money for the retiree that are taxed as income or that 70 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 have tax rates that are independent of the retiree’s allocation decisions among the tda, roth, and taxable accounts, which includes tax-free sources. using these we can accommodate social security in many cases, as will be discussed below. although nothing is stochastic in our approach, as is common in the above literature, we note that the annual rates of return, mt, at each time t can be chosen independently of each other, allowing results to be generated for any stochastically created set of annual rates of return, as is done in some of the papers above. these rates can be positive or negative as long as there are overall gains in the taxable account, because our approach is designed to minimize taxes in the presence of capital gains, not capital losses. one of the ramifications of our model being deterministic is that it will be optimal to have strictly stock, as opposed to bonds or cash, in the tdas, roth accounts, and taxable accounts. that is, because the stock returns are assumed known, the model cannot recommend having bonds or cash given their lower rates of return. of course, in practice, bonds and cash have an important role to play due to their lower volatility. we note that our model will still be able to accommodate known/projected payouts from bond and cash positions in taxable accounts via l tð þ and u tð þ, as will be discussed below. the all-time approach in this paper for optimizing withdrawal strategies in portfolios containing tdas, roth accounts, and taxable stock accounts makes improvements on previous methods in the following ways: • unlike forwards-time approaches, when determining the best allocation in a year, we are able to take into account all future projected annual consumption for the retiree, as well as the projected date of death, and the projected circumstances of the heir. this can have considerable advantages. for example, forwards-time approaches spend tda money up to the top of one tax bracket, unless rmds require more to be spent. this can lead to early depletion of tda money, meaning lower tax brackets in later years cannot be used by the investor to minimize overall taxes. it also prevents optimal strategies where the retiree jumps to the top of new tax bracket at different times so as to maximize the benefit to the heir. our all-time approach uses the additional information available to it to avoid these potential sub-optimal uses of tax brackets. in general, forwards-time approaches are more suited for optimizing longevity, because bequest information is not used to alter its allocation recommendations. our algorithm does use this information, so it can effectively optimize bequests. • unlike current linear programing approaches, our model can use all-time information to optimize portfolio longevity, whereas current linear programing approaches only look to optimize total bequest size. our algorithm considers the nature of capital gains and dividends in the taxable account, whereas current linear programing approaches consider taxable accounts only on an aftertax basis. our algorithm also makes distinctions among the worth of the tda, roth, and taxable stock money to an heir, which affects allocation decisions as we will explicitly show. our algorithm can account for non-linear effects in the tax code, which, by its nature, linear programing cannot. • unlike backwards-time approaches, our model is able to be extended to accommodate multiple tax lots, as explained in section 5 of dilellio and ostrov (2018). further, it can accommodate as many annual time periods as desired, as opposed to only a few time periods due to computational complexity. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 71 as with most of the recent literature above, we will compare the effects of our algorithm to the naı̈ve strategies (also known as the common rule or common strategy throughout the linear programing approach literature) and to the best of the informed strategies. our all-time approach equals or exceeds these forwards-time approaches in every case that we have examined, meaning it obtains a higher bequest or a longer portfolio longevity. this is because the minor approximations used in our approach are only needed to accommodate more complicated phenomena than these strategies consider. our approach is quite different, but compatible, with the conversion method in cook, meyer, and reichenstein (2015). there are cases where our method, by itself, will equal or exceed theirs, such as when taxable stock is not present, so it cannot be used for consumption, as is needed for the conversions described in their paper; there are others where their method, by itself, will do better. however, the strengths of both approaches can be combined in a simple way by using their method forwards in time for the first few years until conversions run out and then applying our algorithm to the remaining years. they can also be combined in a more complicated, but also more effective, way that integrates the two methods, such as the one detailed in section 7 of dilellio and ostrov (2018). while the algorithm works to optimally determine how much a retiree should spend each year from their tda, roth, and taxable stock accounts – that is, these are the three decision variables to be determined each year – it also accommodates, as indicated above, two other sources of external, fixed, annual income. the first source, which we will call l tð þ in this paper, because it will be geometrically represented in the lower part of our graphs, encompasses any projected, fixed sources of money, other than the tda account, that are subject to income tax. therefore, l tð þ is part of the allocation of withdrawals. these include earned income, some pensions, annuities bought with pretax money, and the earnings from annuities bought with post-tax money. the second source, which we will call u tð þ in this paper, because it will geometrically be represented in the upper part of our graphs, encompasses any projected, fixed sources of money that (a) have tax rates that are independent of the retiree’s allocation decisions among the tda, roth, and taxable accounts, and (b) unlike l tð þ, have no effect on the taxation rate of the tda or taxable account. examples include tax-free gifts and tax-free accounts like health savings accounts, some pensions, and the principal from annuities bought with post-tax money. therefore, for example, if we have fixed amounts of cash or bond money coming to the investor, our model can accommodate them, since bond coupons and interest are part of l tð þ, while consumption of principal and the par value of bonds at maturity are part of u tð þ. social security, unsurprisingly, is far more complicated.4 tax on social security is determined from the retiree’s “base amount,” defined as the investor’s agi plus non-taxable interest plus half of the retiree’s social security benefits. if the retiree has a small base amount, meaning, as of 2018, that their base amount is below $25,000 if single or $32,000 if married and filing jointly, then there is no tax. in this case, social security can be accommodated in our model by putting all social security income into u tð þ. our model can also accommodate investors with a large base amount, generally meaning, as of 2018, that 85% of their annual social security benefit is less than $4,500 + 85% of the excess of their base amount over $34,000 if single or $6,000 + 85% of the excess of their base amount over $44,000 if married and filing jointly. in this case, 85% of social security is subject to income tax, so 85% of the social security payment is incorporated into l tð þ, while the other 15% is put into u tð þ. 72 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 in cases where the investor is in the area between these amounts, the taxation rate on social security depends on the allocation between the tda, roth, and taxable stock accounts, which is outside our model. retirees in this area are subject to the “tax torpedo,” which is discussed in meyer and reichenstein (2013), who predict that there are likely less than 10 million out of a total of approximately 140 million tax returns that are subject to this issue. the organization of this paper is as follows: in section 2 we introduce some basic definitions for our model. section 3 contains our model’s assumptions. section 4 lists five guiding principles that will define how our algorithm prioritizes spending in order to maximize the retiree’s bequest. section 5 defines the algorithm’s objective, explains the bar graphs used to report our results, and summarizes our four-stage algorithm. section 6 shows how section 5’s algorithm for optimizing a retiree’s bequest can be used for the subcase where the retiree instead wishes to optimize portfolio longevity. section 7 demonstrates additional results obtained from our algorithm, such as a comparison of our algorithm’s strategy to the naı̈ve and the informed strategies, a comparison of portfolio longevity under the new 2018 tax legislation versus the previous 2017 tax law, and a sensitivity analysis of our results to varying a wide variety of the financial parameters used by our algorithm. in section 8, we discuss our main conclusions. 2. definitions definition of basic variables: t = time (in years) during retirement tdeath = value of t when the investor dies mt = annual rate of return, in real dollars, for stock in all accounts before dividends are distributed. the value of mt can be chosen to vary with time, t. dt = annual dividend rate, where all dividends are assumed to be qualified, distributed at the end of the year, and may be consumed or reinvested. the value of dt can be chosen to vary with time, t. capital gain distributions from mutual funds can also be included here if the tax rates t div and t gains, defined just below, are equal. definition of tax rates: tdiv = tax rate on qualified dividends tgains = tax rate on long-term capital gains tmarg = the marginal income tax rate associated to a given tax bracket. t heir = effective marginal income tax rate for the heir or heirs, which is applicable to distributions from an inherited tda definition of a, the heir’s discount factor for inherited taxable stock: should the heir immediately liquidate their inheritance, they will fail to take advantage of the additional tax advantages that occur over time for a tda or a roth account, though not for a taxable stock account. these advantages for the inherited tda or roth accounts are maximized by the heir only taking out rmds. the value of what we will call the heir’s discount factor a will measure how much additional tax efficiency the heir receives due to the speed at which they liquidate their inherited tda and roth account. comparing the worth of inherited roth money to inherited tda money is straightforward, because at t ¼ tdeath inheriting a dollar of j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 73 roth money is equivalent to inheriting 1 1-their dollars of tda money. comparing the worth of inherited roth money to inherited taxable stock money is where a comes in. we define the discount factor a≤ 1, so that inheriting a dollar of roth or tda money at t ¼ tdeath is equivalent to inheriting 1 a dollars of taxable money at t ¼ tdeath. if an heir immediately liquidates their inherited tdas and roth accounts, then a= 1. it is beneficial to the heir to obtain a lower value of a low by stretching out the liquidation of these accounts. in appendix 2 of dilellio and ostrov (2018), we compute explicit formulas for the lower bound on a, corresponding to an heir being wise and only taking rmds from inherited tdas and roth accounts. under typical circumstances, we find this lower bound to be near 0.75; that is, 0.75≤ a≤ 1. definition of the index j, “lot j of stock,” and jmax: because the cost basis of a stock’s lot depends on the stock’s purchase date, we attach a new index, j, to each successive stock purchase in the retiree’s taxable account. all stock purchased at the same time, indexed by j, will be referred to as the “lot j of stock.” for example, “lot 4 of stock” would be the fourth oldest lot of stock in the retiree’s taxable account. the index jmax corresponds to the total number of different tax lots held by the retiree. when dividends are not immediately used for consumption, they will be used to purchase more stock,5forming a new lot and increasing the value of jmax by one. if a lot is completely consumed by the retiree, then jmax is reduced by one. definition of ωt j: for the lot j of stock at time t, we define v j t = the fraction at time t of the worth of lot j of stock that is equal to its cost basis. for example, let’s say that $20 was used to purchase the original stock in the portfolio. if, 15 years into our algorithm’s projection, that lot of stock is projected to be worth $100, then v 1 15 = 0.2. note that when a new lot of stock such as reinvested dividends is created, we initially have v jmax t = 1 for this new lot j ¼ jmax of stock, because there are no capital gains yet. each year, each group of stock, in real dollars, becomes worth 1þmtð þ 1� dð þ times its previous year’s worth. given this, for each lot j in year t þ 1, we have that v j tþ1 ¼ v j t 1þmtð þ 1� dtð þ : for an investor that sells stocks with losses, we have that6 v 1 t ≤v 2 t ≤ � � � ≤v jmax t : 3. financial and model assumptions we make the following financial and model assumptions in this paper: 1. because we work in real dollars, we assume the inflation rate is known or projected. therefore, for example, if m = 5% and the rate of inflation is 3%, then the annual nominal rate of return for the stock is 8.15%, since 1.05*1.03 = 1.0815. 74 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 2. tax rates and tax brackets: a. we assume that t heir, tdiv, and tgains are known/projected constants. note that t heir may be a projected average or effective tax rate over time and/or over a number of heirs. b. we assume, as is typically the case in tax law, that the nominal tax bracket thresholds adjust with the rate of inflation. this means the projected tax brackets in our model are constant in real dollars, which is why our model uses real, instead of nominal, dollars. c. we assume that tmarg, the tax rates for each tax bracket, are known/projected constants. 3. we assume the investor’s total after-tax consumption needs, c tð þ, are known/projected in each year of retirement. we note that the model can easily be rerun with various projections/scenarios for c tð þ, as well as any of our other parameters, enabling an investor to experiment with these to better understand their financial implications. 4. we assume l tð þ and u tð þ are known/projected in each year t. the money from these funds is used strictly for consumption, not, for example, to purchase stock. we emphasize that l tð þ is subject to income tax rates and could include non-qualified dividends. the tax rate for all funds in u tð þ must be fixed and cannot depend on the manner in which spending is allocated among the tda, roth account, and taxable account, nor can the value of u tð þ affect the manner in which the tda or taxable account is taxed. 5. stock: a. we assume that mt, the annual rate of return for stock in real dollars, and dt, the annual dividend rate, are known or projected at each time t. the values of mt may be positive or negative as long as the stock, overall, has capital gains, not capital losses. b. we assume there are no transaction costs for buying or selling stock, and that stock can be sold in any quantity, including fractional shares, as is available with mutual funds. appendix 3 of dilellio and ostrov (2018) gives the subroutine by which our algorithm annually updates the tda, roth account, and each of the lots in the taxable stock account to address consumption, growth, dividends, and taxes. 6. we assume there are no additional contributions to the tda, roth, or taxable stock accounts, nor are there any conversions from one of these three accounts to another. however, in section 7 of dilellio and ostrov (2018), we discuss how to incorporate allowing roth conversions from non-rmd tda money. section 7 of dilellio and ostrov (2018) also considers years where rmds are greater than the consumption needs, c tð þ, in which case the excess rmds may be used to buy taxable stock, since the irs prohibits rmds from being converted into a roth account (see, e.g., rosato, 2015). 7. we assume that tdeath is a known/projected time. if desired, tdeath can be projected from irs life expectancy tables or the algorithm can easily be reapplied with various values of tdeath to obtain strategies for different tdeath scenarios. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 75 8. we assume the inheritance is not high enough such that the estate tax is relevant. we note that even under the previous law, only one in 500 estates were so large that any estate tax was due (see huang & debot, 2015). since then, the estate tax exemption has doubled due to the tax law passed by congress at the end of 2017, so this figure is now less than one in 1500 estates.7 9. we assume the rate at which the heir consumes their inheritance is known/projected. this is only needed to compute a, the heir’s discount factor, discussed earlier. 4. guiding principles at times, we will think about consuming from taxable stock and from dividends separately, even though both originate from the taxable stock account. in section 5, we will outline how our algorithm looks to optimize withdrawals from four types of money: tda, roth, taxable stock, and dividends generated by the taxable stock. this algorithm will be governed by the following guiding principles that stem from u.s. tax law: guiding principle 1: if a given amount of tda money that is taxed at a constant marginal rate t and a given amount of roth money are both going to be spent to address fixed consumption needs, c tð þ, the allocation/sequencing between the tda and the roth to address this consumption does not matter. similarly, if a given amount of tda money that is taxed at a marginal rate t1 and a given amount of tda money that is taxed at a marginal rate t 2 are both going to be spent, the allocation/sequencing between these two groups of tda money does not matter. this first statement is proven in appendix 1 of dilellio and ostrov (2018). the second statement can also easily be proven using the method presented in that appendix. this means that, given a specific amount of tda money and roth money to be consumed, we optimize these funds’ use by keeping the consumed tda money in the lowest tax brackets, be they for the retiree or for the heir, as possible. guiding principle 2: it is better to use taxable stock and dividends for earlier, rather than later, consumption by the retiree. since the taxable stock and the reinvested dividends have returns that are slowly eroded by the effects of dividends, if we know we are going to use part of our taxable account for consumption, it is better to use that part as early as possible. this means our prioritization of whether to use taxable stock/dividends versus tda/roth money to satisfy consumption may be time dependent, with more likelihood of using the taxable stock or dividends at earlier times, since taxation on tda/roth spending is not time dependent in the way taxable stock is. more specifically, if we know we are going to spend some taxable stock money for consumption, it should be prioritized to be consumed before spending any roth money. the bigger question between the roth account and the taxable account is whether or not it is worth prioritizing using more roth money for the retiree’s consumption needs so that less taxable stock is used for consumption, enabling more capital gains in the taxable account to be forgiven at death. the question of prioritizing the tda versus the taxable account can be even 76 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 more complex, as the desire to spend the tda in lower tax brackets may override the desire to spend taxable stock earlier. our algorithm shows how to answer both of these questions. guiding principle 3: when consuming taxable stock, we consume the lot with the highest cost basis, v t j, that is available at time t. one ramification of this principle is that we always consume dividends before liquidating other lots. by consuming stock with the highest cost basis, we minimize the amount of tax paid on stock that we need to consume. if we must consume the lower cost basis stock later, we have had the advantage of having a longer time to collect returns accrued from the larger capital gains in the lower cost basis stock. further, it is more desirable to have stock with a lower cost basis be in the retiree’s account when the retiree dies, because that means that more tax on the retiree’s capital gains will be forgiven, to the greater benefit of the heir. we note that should v 1 t ≤v 2 t ≤ � � � ≤ v jmax t ≤ 1, guiding principle 3 corresponds to lifo (last in, first out) being the optimal strategy for an investor. that is, we consume first from lot jmax and then, should this lot become exhausted and it is desirable to consume more taxable stock, we consume from lot jmax � 1, which is relabeled lot jmax, and we continue in this manner as long as it remains desirable to consume the lot of taxable stock with the highest remaining j (and v j t) value. further, since dividends correspond to a taxable lot where v jmax t ¼ 1, they are always prioritized for consumption before any other lot. it is worth noting that there may be a material risk associated with a strategy of selling stocks with the highest cost basis, as the taxable account can become very concentrated in stock with strong past performance. this can be partially mitigated by using a broad-based fund, rather than investing in individual stocks. guiding principle 4: we always prioritize using dividends to satisfy the retiree’s consumption needs before using roth money. choosing to prioritize consuming the roth money, which is not subject to any tax for the retiree or the heir, so that we can retain (after-tax) dividend money used to buy taxable stock8 is an inferior choice for three reasons: (1) the erosive effect of taxes on dividends over time with taxable stock, (2) the tax on capital gains should the taxable stock need to be sold before the retiree’s death, and (3) the heir is subject to tax on capital gains accrued after the taxable stock is inherited, even though capital gains are forgiven when the retiree dies. guiding principle 5:we always take out any rmds. the 50% fee levied on any rmds not taken by the retiree from their tda or the heir from their inherited tda or roth cannot be compensated by anything else in the current tax system. 5. algorithm in this section we outline how we use our guiding principles to determine the annual allocations from the tda, roth account, and taxable stock account that satisfy the retiree’s annual consumption needs and maximize the following objective function for wtotal tdeathð þ, the total worth of the bequest to an heir or heirs: wtotal tdeathð þ ¼ 1 a 1� t heirð þwtda tdeathð þ þ 1 a wroth tdeathð þ þwts tdeathð þ, (1) j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 77 where wtda tð þ, wroth tð þ, and wts tð þ are the pretax worths of the roth, tda, and taxable stock accounts at time t. in the context of this equation, the factor 1 a represents the additional benefit to the heir of having the tax advantages of the tda and the roth account before they are liquidated by the heir. in section 6, we will show how to extend this algorithm to optimizing portfolio longevity, instead of optimizing a bequest to an heir or heirs. before outlining our method, we explain the features of the bar graphs that we will use for visualization of both our method and many of our results. 5.1. bar graph visualization and basic set up we note the example bar graphs in fig. 1 below. there is a bar for each year t ¼ 1 through t ¼ tdeath. the height of the bar in year t is c tð þ, the known/projected real dollar after-tax consumption needs of the retiree in that year. just below the title of the bar graph, we present the four quantities in eq. (1): wtotal tdeathð þ, wtda tdeathð þ, wroth tdeathð þ, and wts tdeathð þ. fig. 1. annual after-tax consumption for cases 1 and 2. the parameter values for all of our cases can be found in table 1. for case 1 in the left panel, we attempt to address the retiree’s consumption needs, c tð þ, by choosing values of the three decision variables: tda spending (in light blue), roth spending (in green), and taxable stock and dividends spending (in magenta). in this case, even without considering rmds or taxation on dividends, there is too little money in these three sources, so the retiree has unmet consumption needs (in red). more specifically, any red section in the graph indicates that the investor’s consumption needs cannot all be fulfilled, no matter how the three decision variables are chosen. for case 2 in the right panel, we have two additional sources to address consumption, but these are fixed, not decision variables: l tð þ (in yellow), which is subject to income taxes and therefore affects the marginal tax rate of tda spending, and u tð þ (in dark blue), which has a fixed tax rate that does not affect, nor is affected by, the tax rates determined by the three decision variables. the tda is divided between rmds, which start at age 70 and a half and are represented by the parts of the light blue bars with vertical line segments within them, and voluntary tda consumption, which is represented by the parts of the light blue bars without vertical line segments. the horizontal lines on the graph represent tax bracket thresholds in real dollars. the solid horizontal line represents hheir, which corresponds to the effective marginal tax bracket for the heir. the tax brackets, starting from the bottom, are: 10%, 12%, 22%, 24%, 32%, 35%, and 37%. the higher brackets are out of the range shown in the figure. 78 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 because we have assumed the tax bracket thresholds are constant in real dollars over time, as is usually the case, the income bounds for each tax bracket correspond to horizontal lines on the graph. these are represented by dashed lines, with the exception of our using a solid line on the graph at the height, hheir, which we define as the unique height below which tmarg≤ their and above which tmarg > t heir. the case number, given in the graph’s vertical axis label, corresponds to specific values for parameters, which can be found in table 1. these parameters are: the initial balances for the tda, roth, and taxable stock accounts, the annual consumption needs of the retiree, the values of l tð þ and u tð þ, the values of tdeath, mt, dt, t div, t gains, t heir, and a. also, we must specify the age of the retiree at t ¼ 1, so we know when the retiree reaches the age of 70 and a half and tda rmds begin. we will restrict our computations, although not our algorithm, to the case of the retiree buying only a single lot of taxable stock prior to t ¼ 1, so we must specify the initial value of v for this lot. in all of our cases, we use the irs tax brackets for a single filer from 2018, which are also given in table 1. we note that because our bar graphs are in after-tax dollars, the after-tax values in the final row of our table correspond to the heights of the dashed and solid horizontal lines in the bar graphs. since the retiree’s consumption needs must be fulfilled with after-tax dollars, the consumption bars must be filled with after-tax money from the retiree’s five money sources: l tð þ (in yellow), u tð þ (in dark blue), tda money (in light blue), roth money (in green), and taxable tock money/dividends (in magenta). consumption needs that are not filled by any source are shown in red. because u tð þ involves known/projected sources of money for consumption with known fixed tax rates, we can determine the after-tax worth of these sources, which gives us the value of u tð þ. we then subtract u tð þ from c tð þ to determine the investor’s remaining consumption needs. that is, u tð þ essentially lowers the heights of the consumption bars, so we represent this by placing the consumption from u tð þ at the top of the bars. there are two sources of money subject to income tax, l tð þ and the tda, and we put them – again, in after-tax dollars – at the bottom of the bars, so that their income tax rate is clear. because l tð þ involves known/projected sources of money for consumption, we put it at the very bottom. because tda consumption is a decision variable, we ideally choose to spend it in the lower tax brackets, following guiding principle 1. geometrically, this can be accomplished by thinking of l tð þ as a fixed sandy shore at the bottom of the graph and the tda as calm water on top of it. vertical line segments placed within the tda spending indicate rmds from the tda. we refer to tda spending that is not a part of rmds and, therefore, does not have vertical line segments, as “voluntary tda spending.” consumption from the final two sources of money, roth and taxable stock/dividends, is represented in the graph above the sand/water geometry of the l tð þ/tda system and below u tð þ. we place taxable stock/dividend spending above roth spending when they occur in the same year. after our initial application of u tð þ to the top of the bars and l tð þ to the bottom of the bars, our algorithm looks to find the optimal strategy for the three time-dependent decision variables (the tda, roth, and taxable stock/dividends) using the four stages outlined in the next subsection: j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 79 t ab le 1 . p ar am et er v al u es fo r co m p u ta ti o n s th e fo ll o w in g ta b le g iv es th e p ar am et er v al u es fo r al l th e ca se s p re se n te d in th is p ap er p ar am et er s c as e 1 c as e 2 c as e 3 c as e 4 c as e 5 c as e 6 c as e 7 c as e 8 c as e 9 c as e 1 0 c as e 1 1 t d ea th (y ea rs ) 2 5 2 8 2 5 2 8 3 7 * 3 0 2 0 * 2 0 2 5 m t 5 % 6 % 5 % 6 % 5 % 5 % 5 % 5 % 7 % 5 % 5 % d t 2 % 3 % 2 % 3 % 2 % 2 % 2 % 2 % 4 % 1 % * t d iv 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % t g a in s 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % 1 5 % t h ei r 1 1 % 1 7 % 1 1 % 1 7 % 3 0 % 1 4 % 2 8 % 2 3 % 2 8 % 2 2 % 1 6 % a 0 .9 0 .9 1 0 .9 0 .9 1 0 .9 9 0 .9 0 .9 1 0 .9 0 .9 1 0 .8 5 1 v at t ¼ 1 0 .6 1 4 0 .3 1 2 0 .6 1 4 0 .3 1 2 0 .2 3 1 0 .6 1 4 0 .2 3 1 0 .6 1 4 0 .7 1 3 0 .2 3 1 0 .0 5 4 t d a m o n ey at t ¼ 1 $ 1 6 0 k $ 3 0 0 k $ 3 5 0 k $ 3 2 5 k $ 1 ,7 0 0 k $ 2 3 0 k $ 1 ,9 0 0 k $ 1 ,3 0 0 k $ 4 0 0 k $ 1 ,2 5 0 k $ 5 8 0 k r o th m o n ey at t ¼ 1 $ 6 0 k $ 1 6 0 k $ 9 0 k $ 1 6 0 k $ 2 7 5 k $ 5 0 k $ 3 0 0 k $ 4 0 0 k $ 1 0 0 k $ 1 ,6 0 0 k $ 1 ,2 0 0 k t ax ab le st o ck m o n ey at t ¼ 1 $ 1 8 0 k $ 5 5 0 k $ 1 5 0 k $ 5 5 0 k $ 1 0 0 k $ 1 5 0 k $ 1 ,5 0 0 k $ 3 0 0 k $ 7 9 k $ 2 ,2 5 0 k $ 1 ,0 0 0 k r et ir ee ag e at t ¼ 1 7 0 6 5 7 0 6 5 7 0 7 0 7 0 6 5 3 0 7 0 7 0 v al u e o f a in l tðþ ¼ a eb t� 1 ð þ $ 0 $ 2 0 k $ 0 k $ 2 0 k $ 1 0 k $ 1 0 k * * $ 3 0 k $ 0 k $ 0 k $ 0 k v al u e o f b in l tðþ ¼ a eb t� 1 ð þ 0 �0 .0 9 9 �0 .0 9 �0 .2 �0 .2 * * 0 0 0 0 v al u e o f a in u tðþ ¼ a eb t� 1 ð þ $ 0 $ 2 5 $ 0 $ 1 5 k $ 0 $ 5 k * * * $ 0 $ 0 $ 0 $ 0 v al u e o f b in u tðþ ¼ a eb t� 1 ð þ 0 �0 .3 0 �0 .2 5 0 �0 .2 5 * * * 0 0 0 0 v al u e o f a in c tðþ ¼ a 1 þ r ð þt $ 4 0 k $ 6 0 k $ 4 0 k $ 6 0 k $ 8 1 k $ 5 0 k * * * * $ 1 4 0 k $ 3 5 k $ 1 8 0 k $ 1 0 0 k v al u e o f r in c tðþ ¼ a 1 þ r ð þt �0 .0 2 0 .0 1 �0 .0 2 0 .0 1 5 0 .0 1 0 .0 0 5 * * * * 0 .0 1 0 .0 0 6 0 .0 1 0 .0 1 * v al u es v ar y an d ar e li st ed in th e p ap er . * * l tðþ ¼ $ 1 0 ,0 0 0 þ $ 8 0 t � 1 ð þ 2 9 :7 � t ð þ þ $ 9 ,0 0 0 1 :1 þ co s 0 :5 2 4 t ð þ ð þ . * * * u tðþ ¼ $ 2 5 ,0 0 0 1 :1 � co s 0 :2 3 0 t ð þ ð þ .* * * * c tðþ ¼ $ 2 0 0 ,0 0 0 þ $ 1 0 0 t � 1 ð þ 3 2 :1 � t ð þ . t h e fo ll o w in g ta b le g iv es th e in co m es at w h ic h ea ch ir s ta x b ra ck et b eg in s fo r a si n g le fi le r in 2 0 1 8 (s ee h tt p s: // ta x fo u n d at io n .o rg /2 0 1 8 -t ax -b ra ck et s/ ): b ra ck et ta x ra te 1 0 % 1 2 % 2 2 % 2 4 % 3 2 % 3 5 % 3 7 % b o tt o m o f th e b ra ck et in p re -t ax d o ll ar s $ 0 $ 9 ,5 2 5 $ 3 8 ,7 0 0 $ 8 2 ,5 0 0 $ 1 5 7 ,5 0 0 $ 2 0 0 ,0 0 0 $ 5 0 0 ,0 0 0 b o tt o m o f th e b ra ck et in af te rta x d o ll ar s $ 0 $ 8 ,5 7 2 .5 0 $ 3 4 ,2 4 6 .5 0 $ 6 8 ,4 1 0 .5 0 $ 1 2 5 ,4 1 0 .5 0 $ 1 5 4 ,3 1 0 .5 0 $ 3 4 9 ,3 1 0 .5 0 80 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 5.2. outline of our algorithm while the full details of the algorithm can be found at the academy of financial services online 20018 conference proceedings,9 including how to accommodate different taxable lots, our algorithm can be summarized in the following four stages: stage 1, optimizing using just the tda and roth account: we ignore the existence of the taxable stock account and the tda rmds in this stage. for our first step, we follow dilellio and ostrov (2017). specifically, we fill the bar graph system with tda “liquid” until it is exhausted or rises up to hheir, the level of the solid horizontal line corresponding to t heir. we then use roth money to fill the bars as a light gas would fill them, rising to the top of the bars until the roth money is exhausted or comes down to the hheir line. if either the tda or roth account is exhausted but the other fund is not, we continue to use the other fund as before until it is exhausted or all consumption needs are filled. this gives a graph like the left panel in fig. 2. for our second step, working under the restriction of maintaining the same wtotal tdeathð þ, we move the tda and roth money so that the unused consumption is moved to the left, that is, the earlier years, as much as possible, followed by the tda being moved to the left as much as possible. this gives a graph like the right panel in fig. 2, which is ready for taxable stock to be applied in stage 2. stage 2, optimally using taxable stock if there were no dividends: setting dt ¼ 0 temporarily, we first fill all previously unmet consumption with taxable stock. in the figures – see the left panel of fig. 3, for example – this means we replace the red section remaining after stage 1 with taxable stock in magenta. we then compute for each time and tax bracket “desirability factors” for the taxable stock, the tda, and the roth. by comparing these factors, we determine the most advantageous time and bracket to replace the tda or roth with fig. 2. stage 1 for case 3. left panel: step 1 produces an optimal strategy for the tda (in light blue) and the roth (in green). the unmet consumption needs are in red. right panel: in preparation for using taxable stock money in stage 2, in step 2 we move the unmet consumption to as early as possible in the transition tax bracket and the brackets above it. then, within the 12% tax bracket, we move the tda money to earlier than the roth money, because, being above hheir , it is less desirable for consumption than the roth. note that wtotal is unchanged by step 2. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 81 taxable stock, and then make this replacement, continuing in this manner until we run out of taxable stock or all remaining tda and roth consumption spending is more desirable than additional taxable stock consumption spending. we then do some cleaning up with some small steps that include freezing the taxable stock spending and rerunning step 1 of stage 1 on the tda and roth money again. the right panel in fig. 3 gives an example of the result after this stage. stage 3, incorporating dividends: we reset dt back to its real value, which reduces wtotal tdeathð þ. the generated dividends are first used to replace taxable stock spending. then they are used to replace roth spending, unless there is a better case for dividends to replace tda spending, in which case we do that instead. there are a few additional clean-up steps that may or may not need to occur. by the end of stage 3, we have changed a figure like the left panel of fig. 4 to the right panel of fig. 4 note that wtotal tdeathð þ increases because of using dividends to replace other spending during this stage. stage 4, incorporating rmds from the tda: we compute the rmds for the tda. if they are already satisfied, then we are done. if they are not, as is the case in the left panel of fig. 5 where the vertical bars extend past the blue tda spending, then we apply an iterative procedure to satisfy the rmds while maintaining tda spending in the lowest tax brackets possible. an example of the result of this procedure is given in the right panel of fig. 5. 6. optimizing portfolio longevity in this section, we show how determining a retiree’s optimal portfolio longevity is a subset of the problem of determining how a retiree can optimize their bequest to an heir. this is fig. 3. stage 2 continued for case 3. left panel: taxable stock fills previously unmet consumption, replacing the red of unmet consumption in the right panel of fig. 2 with the magenta that represents taxable stock (including dividends). right panel: taxable stock, until it is exhausted, has replaced tda and roth spending at times and tax brackets where it was most advantageous. after then freezing the taxable stock spending, we reapply step 1 of stage 1 giving the horizonal line between the tda and roth spending. note that wtotal increases because of these changes in stage 2. 82 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 because we can simply run our bequest algorithm from section 5 repeatedly with progressively larger values of tdeath ifwtotal tdeathð þ > 0 and progressively smaller values of tdeath if the retiree has consumption needs that cannot be met. a fraction, a, of the final year can be accommodated by multiplying the last year’s annual consumption needs of the retiree by a. fig. 4. step 3 for case 4. left panel: the bar graph here is the same as it was after the end of stage 2. however, the new inclusion of a 3% dividend rate decreases wtotal at tdeath from $667,908 to $649,231. this decrease, which is strictly through wts, happens because taxes must immediately be paid on the dividends that, starting in year t ¼ 19, are reinvested. right panel: following guiding principle 4, we apply the dividends to consumption needs instead of reinvesting them. all the magenta in the bars where t≥ 19 represent dividend spending, as opposed to other taxable stock spending. this superior strategy increases wtotal to $653,875. fig. 5. stage 4 for case 5. we note that all the consumption is below hheir, making spending tda money a priority. also, no taxable stock other than the dividends are used for consumption. left panel: the colors represent the results after stage 3, but the vertical segments representing tda rmds for this case are not contained in the light blue tda spending in years 20 to 30, so rmds are not yet satisfied. right panel: tda expenditures have been moved in such a way that all rmds are satisfied while continuing to maintain tda spending in as low tax brackets as possible. in this case, in both panels, the tda is exhausted and tda spending covers the entire 22% bracket. the way the remaining tda spending is spread out in the 24% tax bracket is immaterial to the end result, as stated in guiding principle 1, and confirmed by the identical wtotal values in the two panels. that is, in this case, the rmds are able to be satisfied without any negative repercussions to the retiree. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 83 once we have converged to the value of t ¼ tlongevity where wtotal tlongevityð þ ¼ 0, we have the optimal portfolio longevity for our algorithm. an example of this procedure is given in fig. 6. we note that the values of t heir and a are irrelevant to portfolio longevity since there is no bequest to the heir. therefore, any values for these parameters can be selected when running the algorithm; they will have no effect on the portfolio longevity, tlongevity, that is determined. 7. results we discuss computed results from our algorithm for a variety of cases, where, as before, each case’s parameter values can be found in table 1. in all of our cases, we have used the fig. 6. obtaining the optimal portfolio longevity for case 6. upper left panel: running our algorithm with tdeath ¼ 10 leads to wtotal > 0, a positive bequest to the heir, so we increase tdeath. upper right panel: running our algorithm with tdeath ¼ 13 leads to unmet consumption (in red), so we must decrease tdeath. lower center panel: continuing in this fashion, we converge on the optimal portfolio longevity of 11.35 years. we note that a ¼ 0:35 in this case, meaning that the consumption needs in the final year are reduced to 35% of their normal value. 84 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 irs tax brackets for a single filer from 2018, which are also given in table 1, although our program can also easily accommodate changes to the thresholds, rates, and number of tax brackets, should congress change any of these. the matlab r2017a computer program for our algorithm typically ran in under 4 seconds on an imac with a 4ghz intel core i7 processor and 16 gb of 1867mhz ddr3 memory. 7.1. examples of our algorithm results the three computed examples in fig. 7 demonstrate some of the wide range of behavior that our algorithm captures. the two upper panels of fig. 7 are the final products of the algorithm for case 3 and case 4, whose intermediate steps were presented in section 5. for case 3 in the upper left panel of fig. 7, we see that the solid horizontal line for hheir is at the top of the lowest tax bracket, which is the 10% bracket, because t heir = 11% here, which is less than 12%, the rate for the second bracket. ideally, the tda would only fill the area below hheir and nothing above, but rmds require that more be filled, following guiding principle 5. after that, it is hoped that the roth and taxable stock accounts can fill the remainder. the optimal strategy requires the taxable stock to be consumed as early as possible, following guiding principle 2. in this case, we fill to the point where the taxable stock is exhausted, causing wts tdeathð þ ¼ 0, then we fill with the roth. the roth also becomes exhausted, causing wroth tdeathð þ ¼ 0, so the remainder of the consumption needs must be filled with “voluntary” tda money. we note that the taxable stock fills all consumption needs that lie in the third tax bracket, which is the 22% bracket, which is optimal since tda spending in this bracket would be highly taxed. however, by guiding principle 1, the method in which the tda and the roth fill the 12% tax bracket does not matter, as long as the roth is exhausted and all consumption needs are satisfied, as is the case here. for case 4 in the upper right panel of fig. 7, we have known, fixed sources that create l tð þ and u tð þ at the bottom and the top, respectively, of the consumption bars. in this case, wtda tdeathð þ, wroth tdeathð þ, and wts tdeathð þ are all non-zero, so the tda stays below hheir and the roth stays above hheir. up through year 18, there is a stronger case to use taxable stock instead of the roth. that is, during this period, the forgiveness of capital gains for taxable stock when the retiree dies are a weaker effect than the erosive effects of dividend taxes and the inability to shield the heir’s subsequent gains from taxes. starting in year 19, the forgiveness of capital gains for taxable stock becomes the stronger factor, making using the roth preferable to using taxable stock, so, starting in year 19, only dividends are consumed from the taxable stock account, as required by guiding principle 4. similarly, up through year 8, there is a stronger case to use taxable stock instead of the tda in the 12% tax bracket, but, starting in year 9, this preference reverses. note that appendix 5 of dilellio and ostrov (2018) contains a table with the specific values for the annual consumption and remaining balances that correspond to the graph for case 4, as displayed in fig. 7. from an economic point of view, c tð þ, l tð þ, and u tð þ generally exhibit exponential growth or decay. for example, c tð þ may need to grow faster than inflation to accommodate increased medical needs; l tð þ may represent part-time work that decreases over time after retiring; or u tð þ may be a tax-free pension that grows with inflation and is therefore constant j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 85 in real dollars. however, this restriction to exponential models is only for economic reasons. our algorithm is capable of handling any functions for c tð þ, l tð þ, and u tð þ that we wish to model. case 7 in the lower panel of fig. 7, for example, uses non-exponential functions for all three. 7.2. comparison of our results with the naı̈ve and informed strategies in the introduction, we discussed previous common strategies non-academics and academics have used for drawing down funds in retirement, which horan called naı̈ve and informed strategies. in all of these strategies, rmds from the tda are first satisfied. in naı̈ve strategy 1, the retiree then drains the taxable stock account, followed by the tda, and finally the roth. in naı̈ve strategy 2, the retiree then drains the taxable stock account, followed by the roth, and finally the tda. in an informed strategy, the retiree drains the tda up to the top of one of the tax brackets, and then fills any excess consumption needs by first liquidating the taxable stock account, followed by the roth, and finally the tda if it has not fig. 7. algorithm results. upper left panel: case 3. upper right panel: case 4. lower panel: case 7. 86 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 already been drained. the best informed strategy selects the tax bracket with the best outcome. our algorithm generates results that are superior or equal to those generated by either of the naı̈ve strategies or the best informed strategy. this has held in every case that we have run, including not just the cases presented in this paper, but also the numerous other cases we have run but not included here for the sake of space. case 8, for example, which is shown in fig. 8, was produced by considering an investor who had a salary of $200,000 and followed two standard rules of thumb for retirement: (1) the investor saved 10 times their salary before retiring, and (2) the investor planned on initially spending 70% of their salary during retirement.10 comparing the four strategies, we find that the size of the bequest, wtotal tdeathð þ, for our algorithm is highest, followed by the best informed strategy, naı̈ve strategy 1, and finally, naı̈ve strategy 2. in table 2, we find similar results for case 6 and case 9, where we compare optimal portfolio longevity instead of bequest size. case 6 was previously presented in fig. 6 case 9 fig. 8. comparison of naı̈ve strategy 1 (upper left panel), naı̈ve strategy 2 (upper right panel), the best informed strategy (lower left panel), and the strategy generated by our algorithm (lower right panel). we note that the best informed strategy in this case fills the retiree’s consumption needs with tda money up to the top of the fourth tax bracket, which is the 24% bracket. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 87 somewhat resembles the pictures for case 8 in fig. 8, although it applies to an individual with a longer time horizon, consumption needs that only reach into the 22% tax bracket, no rmds, and no resources other that the tda, roth, and taxable stock. for case 6, the best informed strategy fills the retiree’s consumption needs with tda money up to the top of the 12% bracket. in case 9, the best informed strategy fills up to the top of the 10% bracket with tda money. as emphasized in the heading for table 2, the calculations throughout this paper use the 2018 tax brackets and tax rates passed into law by congress at the end of 2017. these brackets and rates are set to revert to the previous 2017 brackets and rates in 2025 if congress takes no further action. it is easy to modify our algorithm to use these 2017 brackets and rates, which we have done for table 3. comparing table 2 and table 3 demonstrates the effect of the 2018 versus 2017 brackets and rates on portfolio longevity. for table 3, the best informed strategy fills up to the top of the 12% bracket in case 6 and the 10% bracket in case 9. as with table 2, these correspond to the second lowest and the lowest tax brackets. 7.3. sensitivity analysis we return to case 4 to explore how sensitive the optimal decisions are to changes in the underlying parameters. recall from the withdrawals shown in the upper right-hand panel in fig. 7 for case 4, that, within the 12% tax bracket, it is preferable to use taxable stock instead of tda money in years 1–8 when possible, although tda money must still be used to satisfy tda rmds during these years. starting in year 9, the situation switches: using tda money becomes preferable to using taxable money for years 9–28, so we see that year 9 is the first year where voluntary tda money is applied. within the 22% and 24% tax brackets just above the 12% bracket, applying taxable stock money for consumption is preferable to roth money, until we switch at year 19, when it becomes preferable to apply dividends and then roth money instead of consuming non-dividend taxable stock money. we begin our sensitivity analysis with the effect of the dividend rate, dt on these two switching times. from table 1, we note that dt is 3% for case 4. if we reduce dt from 3% to table 3 portfolio longevity for various strategies using 2017 tax brackets and rates tlongevity (in years), naı̈ve strategy 1 tlongevity (in years), naı̈ve strategy 2 tlongevity (in years), best informed strategy tlongevity (in years), our algorithm’s strategy case 6 10.93 10.94 11.06 11.12 case 9 33.50 33.11 33.67 35.12 table 2 portfolio longevity for various strategies using 2018 tax brackets and rates tlongevity (in years), naı̈ve strategy 1 tlongevity (in years), naı̈ve strategy 2 tlongevity (in years), best informed strategy tlongevity (in years), our algorithm’s strategy case 6 11.17 11.18 11.26 11.35 case 9 36.18 35.75 35.99 37.23 88 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 1%, we get the strategy shown in the left panel of fig. 9 the corrosive effect of dividends, which we discuss more fully later, is reduced, so there is less push to avoid them by spending down taxable stock early. this means the switching year from taxable stock to tda in the 12% bracket is reduced from year 9 to year 4, while the switching year from taxable stock to the roth in the brackets above that is reduced from year 19 to year 17. in fact, ideally, the switching year in the 12% bracket might be earlier than year 4; however, because the tda is exhausted, there is no additional tda money to replace the taxable stock within the 12% bracket in earlier years. in the right panel of fig. 9, we see the opposite behavior: we have increased dt from 3% to 5%, so we see the switching year in the 12% bracket increase from year 9 to year 15 and then, within the higher tax brackets, the taxable stock exhausts itself at year 21 and consumption switches over to roth spending. quantifying where these withdrawal strategy shifts should optimally occur when there is a choice has not been done in other research, to our knowledge. because we can quantify these withdrawal switch times, we can measure their sensitivity to changes in other parameters. we note from table 1 that for case 4, we have that dt ¼ 3%, a ¼ 0:91, and mt ¼ 6%: we also have initial balances in the tda, roth, and taxable stock accounts equal to $325,000, $160,000, and $550,000, respectively. finally, t heir ¼ 17% and t div ¼ t gains ¼ 15%: in table 4, we adjust each of these parameters upwards and downwards from their base case values in table 1, and see the effect this has on the switching years within the 12% bracket and within the brackets above that. we next explain the qualitative reasons that correspond to the results seen in table 4. recall that the value of a is determined by how long the heir stretches out the removal of funds from their inherited tda and roth account. should they liquidate all of their accounts upon inheritance, then a= 1. in this case they are not taking advantage of the additional tax shielding available for the tda and roth, making these two accounts less valuable to the heir, so the retiree moves both switching years earlier to spend more of the tda and roth. as a gets smaller, both switching years move to later times. we see that this is a large effect for the tda switching year in the 12% bracket, changing from year 4 when a= 0.95 to year fig. 9. the effect of dividends on withdrawal choices for case 4: in the left panel, the dividend rate, dt ¼ 1%. in the right panel, dt ¼ 5%. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 89 9 in the base case where a= 0.91 to year 12 when a= 0.85. for the roth switch in the higher tax brackets, the effect is less pronounced, changing from years 17 to 19 to 22. we also see from table 4 that as mt is increased, the switching years both get a little earlier. this is because the capital gains increase with mt, giving a greater incentive not to use taxable stock for consumption and instead take advantage of the step up in basis enjoyed by the heir. there is also an increase in dividends as mt increases, which pushes the switching years to be later so as to lower the dividends. this effect, however, is smaller than the effect of the cost basis step up, which we see from the fact that the switching years get earlier, not later. in table 4 we see that the initial tda and roth balances have no effect on the switching years. this makes sense, since the only effect it should have is when the tda or roth accounts are completely drained, which does not happen in the range of values chosen here. additional initial taxable stock money, however, moves the switching years later because of the need to reduce the additional dividends through earlier expenditures from the taxable stock account. increasing t heir gives incentive to the retiree to spend more tda money, since it is of less worth to the heir. therefore, as t heir increases from 13% to 25%, we see the tda switching time in the 12% bracket decrease from year 17 to year 4 where the tda is completely drained. the effects of changing tdiv and tgains work in opposite directions: as t div increases, the switching years move to later years to decrease the dividends. however, as tgains increases, the switching years move to earlier years to push more of the taxable stock to the heir, where gains are forgiven. given that t div and tgains have been equal for decades, which effect is stronger? from the final rows of table 4, we see that the effect from increasing t gains to move the switching year earlier is stronger than the effect from increasing tdiv to move the switching year later, since increasing these rates in unison moves the switching years to earlier times. table 4 sensitivity analysis for case 4 description switch year: taxable to tda in the 12% bracket switch year: taxable to roth in higher brackets baseline (case 4) 9 19 dt ¼ 1% 4 17 dt ¼ 5% 15 21 a ¼ 0.85 12 22 a ¼ 0.95 4 17 mt ¼ 5% 10 20 mt ¼ 8% 6 17 initial tda balance = $300,000 9 19 initial tda balance = $350,000 9 19 initial roth balance = $120,000 9 19 initial roth balance = $400,000 9 19 initial stock balance = $450,000 8 17 initial stock balance = $650,000 9 19 sheir ¼ 13% 17 19 sheir ¼ 25% 4 19 sdiv ¼ sgains ¼ 10% 9 24 sdiv ¼ sgains ¼ 20% 7 17 90 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 7.4. additional implications of the heir’s tax rate changing the value for t heir can have more interesting implications on the optimal strategy than just moving the switching year for preferring tda to taxable stock spending that was seen in the previous example. consider case 2, which is shown in the right panel of fig. 1 and looks similar to case 4. from table 1, we note that case 2 has an initial tda balance of $300,000 and their ¼ 17%. coincidentally, as with case 4, within the 12% tax bracket, it becomes preferable to switch from using taxable stock to using tda money starting in year 9 and within the 22% tax bracket, it becomes preferable to switch from using taxable stock to using roth money and dividends starting in year 19. as we increase their, as we saw in the previous example, the switch in the 22% bracket stays at year 19, while the switch in the 12% bracket moves from year 9 to earlier years. this consumes more and more of the tda until it runs out. that is, all the tda money is spent either in the 10% bracket or in all but the earliest years of the 12% bracket. therefore, we boost the initial ira amount from $300,000 to $425,000 and continue to increase t heir. once t heir reaches 21.2%, the switch in the 12% bracket has moved to year 1. that is, there is no taxable stock, only tda, used in the 12% tax bracket. note that, unlike the roth which never displaces dividend spending, the tda displaces all dividend spending in the 12% bracket in this case (see fig. 10, left panel). once their increases past 22%, using the tda becomes preferable to using the roth in the 22% tax bracket, so the tda now becomes exhausted replacing the roth (see fig. 10, right panel). increasing the initial tda amount further causes it to eventually replace all the roth consumption. should there be additional tda money and then t heir is further increased, the tda will first move the switching year to progressively early times. after the switching year, the tda and dividends are used for consumption in the 22% tax bracket. when t heir ¼ 29:02%, the switching year is year 4. as soon as we cross t heir ¼ 29:02%, the tda becomes preferable to dividends in years 4 and after within the 22% bracket and, therefore, the tda immediately replaces all of this dividend spending. dividends, along with u(t), are fig. 10. the effect of the heir’s effective tax rate, t heir, on the retiree’s withdrawal choices for case 2, but with an initial tda balance of $425,000. in the left panel, t heir ¼ 21:2%. in the right panel, t heir is increased to 23%, which moves it from a value below the tax rate for the 22% bracket to a value above it. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 91 still used exclusively in the 24% bracket and, along with taxable stock, for years one through three in the 22% bracket. as t heir increases further, the switching year decreases, so the tda resumes pushing out the taxable stock and dividend spending in years one through three in the 22% bracket. if their is increased even further, the tda begins to push out later dividend spending in the 24% bracket. we note that forward time algorithms are, by their nature, incapable of determining these types of effects of t heir upon the optimal withdrawal strategy. nor are forward time algorithms capable of determining the effect of a, the heir’s discount factor, on the optimal strategy. in the left panel of fig. 11, we have case 10, where the value of a is 0.85. unlike case 4, the switching time in year 10 is the same for both the tda and the roth account. this switching time is remarkably sensitive to the value of a. if a is increased from 0.85 to 0.91, the tda and roth are now always preferred, and so they fill all the area, with the exception of the dividends, which are still consumed in all years. on the other hand, if a is decreased to 0.81, taxable stock is always preferred, so it fills all of the area above the tda rmds. that is, the decision of the heir to quickly spend down their inherited tda and roth versus taking out rmds from these accounts has a significant – and quantifiable – effect on how the retiree optimally withdraws funds. given a choice between two stocks with an expected return of 5%, where one stock returns 4 of that 5% as dividends and the other gives no dividends, the heavy majority of retirees and their advisors will pick the stock that returns dividends for reasons including that dividend paying firms are often considered to be more stable than non-dividend paying firms. from a taxation point of view; however, this is known to be a considerable mistake (e.g., see demuth 2016). there are many reasons for this. it is well known for long-term stock investing that the loss of deferring capital gains taxes because of an intermediate sale and repurchase of a stock can have a considerable negative impact on the total gains after the final sale of the stock. assuming t gains ¼ tdiv, dividends are like a forced intermediate sale, except that dividends are worse for two reasons: (1) the sale of normal stock is at the cost basis determined by v of the lot being sold, but dividends are like a sale on just the gains. that is, it is a sale on a part of the lot where v ¼ 0, the worst case. (2) the fact that fig. 11. left panel: case 10 for studying how sensitive the optimal strategy is to changes in the heir’s discount factor, a. right panel: case 11 for studying the corrosive effect of dividends, d. 92 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 dividends then lower v for the rest of the lot will partially compensate for this if the stock is later sold by the retiree. however, once the retiree stops using taxable stock for consumption, this lot will not be sold by the retiree. it will go to the heir, where all gains are forgiven and the fact that v was lowered for the retiree becomes irrelevant. our algorithm quantifies how problematic these tax issues with dividends are. consider, for example, case 11 shown in the right panel of fig. 11, where the investor starts with a million dollars in taxable stock, a= 1, and the dividend rate is dt = 4%. all the magenta in the panel represents dividends, which are optimally used for consumption by general principle 4. if we now change the dividend rate to dt = 0, the magenta part will be replaced by the roth in green, and the value of wtotal tdeathð þ will increase from $3,284,021 to $3,608,910. that is, in this case, if the retiree chooses the stock without dividends for their taxable account, their heir will receive $324,889 more dollars. in other words, we see from our results that the corrosive effect of dividends is considerable, especially for investors with large taxable stock holdings. 8. conclusions previous strategies for how a u.s. retiree should allocate their withdrawals in a tax efficient manner among tdas, roth accounts, and taxable stock accounts have not depended on the amounts in these three accounts, nor on the parameters governing the retiree, nor on the parameters governing their heirs. in this paper, we have presented an algorithm that uses these amounts and parameters to develop a strategy that adapts to the retiree’s specific circumstances. the development of our algorithm reveals insights into the complex structure governing the trade-offs in using these three accounts to satisfy retiree consumption needs. this is particularly challenging with the taxable account, because its tax structure differs significantly from the tax structure of the tda and the roth account. our algorithm starts by using current tax law to create five guiding principles that govern prioritizing consumption from the three accounts: (1) if the retiree consumes a given amount of roth money and consumes given amounts of tda money at various marginal tax rates, the order/allocation in which these are consumed is irrelevant; (2) in contrast, taxable stock and dividends are better spent earlier rather than spent later; (3) when consuming taxable stock, the lot with the highest cost basis should always be liquidated; (4) dividends should always be consumed before roth money; and (5) rmds should always be taken out of any tda. these five guiding principles lead to our algorithm for optimizing how to allocate from the three accounts, so as to meet the retiree’s projected yearly future total after-tax consumption needs. our algorithm incorporates a number of standard features studied in the research literature, such as working with rmds from the tda account and optimizing portfolio longevity, as well as a number of less standard features, such as incorporating the effect of dividends, optimizing a bequest to an heir, allowing for two types of additional fixed sources of money for the retiree, and accommodating different taxable lots in the taxable stock account. because our algorithm uses an all-time approach that takes information from every year to form its withdrawal strategy in each year, it is able to produce a number of results that were previously not possible with other approaches, including forwards-time approaches, j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 93 which generate their withdrawal strategy in chronological order. for example, our algorithm can avoid running out of the tda early, which means that low tax brackets in later years cannot be exploited, or having too much tda at the end, which means that the tda can be forced into higher tax brackets, should the other accounts be drained. we are also able to use the financial situation of the heir to guide withdrawals, and have seen that these often have a considerable effect on the optimal withdrawal strategy. our algorithm shows significant improvements over the previous naı̈ve and informed forwards-time strategies. it is also compatible with the advantages of forwards-time conversion strategies in the literature. notes 1 in this paper, the term “stock” will be shorthand for a portfolio of stocks that may include mutual funds, exchange-traded funds, as well as a variety of individual stocks. 2 https://academyfinancial.org/resources/documents/proceedings/2018/f2%20dilellio% 20and%20ostrov.pdf 3 https://academyfinancial.org/page-18138 4 https://www.fool.com/retirement/2018/09/29/is-social-security-taxable.aspx 5 if the retiree has no earned income, they cannot put dividend money into a tda or a roth account. even if they have earned income, but are older than 70 and a half, they cannot put dividend money into a tda. see irs rules: https://www.irs.gov/ retirement-plans/traditional-and-roth-iras. 6 we note that this is optimal in the following sense: for any stock at a loss, it is always optimal for the investor to sell the stock and then buy another stock with similar properties to immediately reap the tax advantage of realized capital losses. the replacement stock cannot be exactly identical because of wash sale rules. so, for example, a total stock market fund would be replaced with another total stock market fund that tracks a similar, but not identical, index. see, for example, ostrov and wong (2011). 7 please see tax policy center (2017). 8 we assume the retiree is not working, so the money remaining from dividends after taxes cannot be used to purchase stock in the tda or roth account. 9 https://academyfinancial.org/resources/documents/proceedings/2018/f2%20dilellio% 20and%20ostrov.pdf 10 see, for example, https://www.fidelity.com/viewpoints/retirement/how-much-money-doi-need-to-retire references al zaman, a. (2008). reconsidering ira and roth ira: keeping bequests and other options in mind. journal of wealth management, 11, 82–92. brown, d. c., cederburg, s., & o’doherty, m. s. (2017). tax uncertainty and retirement savings diversification. journal of financial economics, 126, 689-712. cook, k. a., meyer, w., & reichenstein, w. (2015). tax-efficient withdrawal strategies. financial analysts journal, 71, 16-29. coopersmith, l. w., & sumutka, a. r. (2011). tax-efficient retirement withdrawal planning using a linear programming model. journal of financial planning, 24, september, 50-59. 94 j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 coopersmith, l. w., & sumutka, a. r. (2017). the impact of rates of return on roth conversion decisions and retiree savings wealth. journal of personal finance, 16, 29-39. demuth, p. (2016). the overtaxed investor: slash your tax bill & be a tax alpha dog. los angeles, ca: alpha dog press. isbn 978-0-9970596-0-1 dilellio, j. a., & ostrov, d. n. (2017). optimal strategies for traditional versus roth ira/401(k) consumption during retirement. decision sciences, 48, 356-384. dilellio, j. a., & ostrov, d. n. (2018). constructing tax efficient withdrawal strategies for retirees with traditional 401(k)/iras, roth 401(k)/iras, and taxable accounts. proceedings of the academy of financial services 2018 annual conference. chicago, il. (available at https://tinyurl.com/ybvc8ekt or https://papers. ssrn.com/sol3/papers.cfm?abstract_id=3374590). fidelity. (2014). four tax-efficient strategies in retirement. (available at www.fidelity.com/viewpoints/ retirement/tax-savvy-withdrawals). fidelity. (2015). detailed methodology, retirement income planner. (available at http://personal.fidelity.com/ planning/retirement/pdf/rip_methodology.pdf). geisler, g., & hulse, david, s. (2018). the effects of social security benefits and rmds on tax-efficient withdrawal strategies. journal of financial planning, 31, 36-47. huang, c.-c., & debot, b. (2015). ten facts you should know about the federal estate tax. center on budget and policy priorities, 1-7. horan, s. m. (2006a). withdrawal location with progressive tax rates. financial analysts journal, 62, 77-87. horan, s. m. (2006b). optimal withdrawal strategies for retirees with multiple savings accounts. journal of financial planning, 19, 62-75. jaconetti, c. m., & bruno, m. a. (2008). spending from a portfolio: implications of withdrawal order for taxable investors. valley forge, pa: the vanguard group. lange, j. (2009). retire secure! new york, ny: wiley, isbn: 0470405317. larimore, taylor, lindauer, m., ferri, r. a., & dogu, l. f. (2011). the bogleheads’ guide to retirement planning. new york, ny: wiley. isbn: 0470919019. meyer, w., & reichenstein, w. (2013). the tax torpedo: coordinating social security with a withdrawal strategy to minimize taxes. retirement management journal, 3, 25-32. ostrov, d. n., & wong, t. g. (2011). optimal asset allocation for passive investing with capital loss harvesting. applied mathematical finance, 18, 291-329. piper, m. (2013). can i retire? st. louis, mo: simple subjects, llc. isbn: 0981454259. ragsdale, c. t., seila, a. f., & little, p. l. (1994). an optimization model for scheduling withdrawals from taxdeferred retirement accounts. financial services review, 3, 93-108. reichenstein, w., horan, s. m., & jennings, w. w. (2012). two key concepts for wealth management and beyond. financial analysts journal, 68, 14-22. rodgers, r. (2009). the new three-legged stool. new york, ny: wiley, isbn: 1592803636. rosato, d. (2015). can i put my required minimum distribution into a roth ira? money. (available at http:// time.com/money/3835309/retirement-required-minimum-distribution-roth/). solin daniel r. (2010). the smartest retirement book you’ll ever read. new york, ny: perigee trade. isbn: 0399536345. sumutka, a. r., sumutka, a. m., & coopersmith, l. w. (2012). tax-efficient retirement withdrawal planning using a comprehensive tax model. journal of financial planning, 25, april, 41-52. tax policy center (2017). https://www.taxpolicycenter.org/taxvox/only-1700-estates-would-owe-estate-tax2018-under-tcja vanguard. (2013). how to turn retirement savings into retirement income. (available at https://retirementplans. vanguard.com/ekit/pmed/retirement_income_center/pdf/rilongbroc.pdf). welch, j. s. (2016). measuring the financial consequences of ira to roth ira conversions. journal of personal finance, 15, 47-55. welch, j. s. (2017). a 3-step procedure for computing sustainable retirement savings withdrawals. journal of financial planning, 30, august, 45-55. j. dilellio, d. ostrov / financial services review 28 (2020) 67–95 95 pii: 1057-0810(91)90003-h financial services review, l(l):]-8 issn: 1057-0810 copyright @ 1991 by jai press inc. all rights of reproduction in any form reserved. individual versus institutional investing harry m. markowitz this paper first describes the analytic approach that markowitz used in developing his portfolio theory. developing a game-of-life simulation is a parallel approach for modelling individualfinancial management. to develop a realistic simulator will require deciding what goals are essential to the family planning process, formulating optimizable subproblems, using technology to interpret and record decisions, and developing decision rules which prove robust in the model and can be implemented in practice. professor mandell, editor of financial services review, invited me to contribute an article related to financial research for the individual for the first issue of this journal. since the subject is not my specialty, it was uncharacteristically risky of me to have accepted the invitation. but an evening of reflection convinced me that there were clear differences in the central features of investment for institutions and investment for individuals, that these differences suggest differences in desirable research methodology, and that a note on these differences may be of value. as i thought about the subject further, on subsequent days, i found myself of two minds. on the one hand, surely financial decisions for the individual should be considered as part of the “game as a whole” which the individual plays out-“game”in the sense of von neumann and morgenstern (1944). even reducing this game to its essentials, it has characteristics of situations for which simulation methods have proved to be the superior tool in practice. on the other hand, there are approximations to the individual’s financial situation which seem good enough, and would allow us to solve analytically for optimal action. neither train of thought succeeded in defeating the other. below, i present both views: the first in a section called “thesis,” and the second under “antithesis,” and attempt some reconciliation in a final section, “synthesis.” harry m. markowitz l marvin speiser distinguished professor of finance and economics, baruch college, cuny, 17 lexington avenue, new york, ny 10010. financial services review, l(1) 1991 thesis in markowitz (1952) i conjecture that “perhaps-for a great variety of investing institutions which consider yield to be a good thing; risk, a bad thing; gambling to be avoided-e, v efficiency is reasonable as a working hypothesis and a working maxim.” the “investing institution” which i had most in mind when developing portfolio theory for my dissertation was the open-end investment company or “mutual fund.“this familiarity was not from first hand experience (i was a student and son of a grocer) but from wiesenberger & company’s investment companies (1944). it was plausible to assume for the mutual fund that its objective is to obtain a “good” probability distribution of year-to-year (or quarter-to-quaker) percent increase in its net asset value. in addition, i argued for mean and variance as criteria in judging “goodness.” for the present discussion, the choice of mean-variance criteria is not the crux; rather it is the formulation of the problem as that of selecting a portfolio to achieve a good probability distribution of a single random variable: the return on the portfolio as a whole. this formulation turned out to be widely acceptable in practice as well as tractable analytically. in the 195os, i participated in attempts to develop advanced, but practical, methods for assisting manufacturing planning, particularly assisting equipment selection and production scheduling for job shops. we considered optimization techniques first, such as linear and dynamic programming, but found that too much reality had to be ignored to allow these techniques to be applied. simulation techniques seemed promising, and were developed for real decision problems with real job shops. experience confirmed the value of simulation analysis, but showed that programming the model was a bottleneck. this, and similar experiences in other application areas, stimulated development of the simulation programming languages of the early 1960s. in the meantime, attempts continued to apply analytic techniques to shop scheduling problems. a recent survey (lawler, lenstra, rinnooy kan, and shmoys, 1989) reports that some flow shop problems have been solved; others have been shown to be np-hard (i.e., as hard to solve as the traveling salesman problem); but results for job shops are meager, leading the authors to end with a quote from coffman, hofri, and weiss (1989), “there is a great need for new mathematical techniques useful for simplifying the derivation of results.” in the meantime, simulation analysis is increasingly used in practice. the difference between the investment company situation and that of a job shop is the number of state variables that need to be considered in a practical problem. for the investment company, it is plausible to assume that assets are liquid, therefore the state of the portfolio can be described by its total value. for the shop, its state description includes the contents of all its queues. to judge whether the problem of financial planning for the individual is amenable to analytic solution, let us sketch what a “game-of-life” model might ~n~ivi~uz versus ins~tu~onui investing 3 entail. we seek a model with sufficient realism as to be a guide to practice. for example, some economic theories find it convenient to assume that the individual is immortal, or that death is a poisson process independent of the age of the individu~. for actual fin~ci~ planning, however, aging and mortality are salient facts that must be included in the model. on the other hand, many details of life which are important to the individual may be ignored for financial planning. for example, the model should include the probability of an accident or disease which will keep the individual from work for an extended period, the probability distribution of time to recover or die, costs of treatment and probability of relapse, since these possibilities are major factors in financial planning; but medical details are not required. since time and uncertainty are at the heart of the problem, i will sketch the model as if it were a simulation. this is for the purpose of model description, and does not itself preclude the possibility that the model could be solved analytically. the description will use the simscript worldview (kiviat, villanueva, and markowitz, 1983). this says that, as of any instant in time, the model represents entities of various entity types; a given entity is characterized by the values of its attributes; also, it may own sets to which other entities belong, and belong to sets which other entities own. this status description changes at points in time called events.’ one event may cause one or more subsequent events to occur after fixed or random time delays. the essentials of the game-of-life is probably different for (a) the very wealthy, (b) the class of homeless that used to be called vagrants, and (c) most of my friends and relatives. i have the latter in mind as i sketch the model. among the types of entities of the model, we must distin~ish between the individual(i.e., human person) and the (nuclear)~a~i~y. often, at some stage this family will “own” a set of individuals whose roles are husband, wife, children and perhaps residing elder. frequently, in the course of events, the residing elder (if any) dies or is placed in a nursing facility; the children leave home to set up their own nuclear families; the original family (the subject of the model) then consists of husband and wife. when one dies the subject family consists of the survivor only. when the latter dies, the assets of the subject family are distributed to heirs, and the game-of-life is over for the subject family. in the simplest case, assets may be thought of as belonging to the family rather than the individual, to be used by husband and wife and (at their discretion) by the children until, at the last, it is used to support the survivor and then distributed. it may be sufficient to characterize financial assets as the total value of the family’s holdings in stocks, bonds, cash items and real estate [other than the family’s home(s)]. perhaps, upon further reflection, it may prove essential to disaggregate these items according to the maturity of the bonds, their tax exempt status, and the unrealized capital gains and losses of various assets. [problem: must we distinguish many individual stocks in order to characterize available capital gains and losses for tax calculations?] perhaps 4 financial services review, l(1) 1991 bonds and stocks may be treated as instantaneously marketable, perhaps with a small commission, but real estate requires greater (random) time and cost to sell. among other assets, the family may “own” [in the simscript sense, i.e., have associated with it] one or more residences. the residence may be owned {in the usual sense) or rented. if owned, the residence is characterized by original cost and a current market value; whether owned or rented, the residence has a value of owned furnishings. a home and its furnishings are clearly an illiquid asset, not only because of the time and cost to sell, but also that to move, and the mismatch between furniture needs of the old and new residences. attributes of individuals include those needed to characterize health, the employment or employability of husband and wife, and the educational objectives of each child. the assets of a residing elder can be characterized by associating with this entity his or her own nuclear family entity. events which change status include periodic events such as receiving a salary check, having a birthday, or the time when an income tax payment is due; and randomly occurring events such as becoming sick, becoming well, finding a job, losing a job, financing a house to buy, finding a buyer for a house to be sold. changes in price levels, interest rates, and stock and real estate values could be computed periodically; e.g., increments in price levels and interest rates could be drawn from a joint distribution, then the change in real estate values could be computed as a function of the former increments and other random variables. the simulated family must make decisions at various points in time, such as the level of (say) this week’s nondurable consumption, transfers from cash to other liquid assets, the decision to search for and then buy a new house, and the decision of one of its members to retire. the simulated family makes these decisions according to decision rules. a major purpose of the model is to evaluate alternate decision rules. the above is a partial sketch of a game-of-life model, rather than detailed specifications for one. the model should also include, as essential to evaluating family investment practice in fact, such things as iras, keoghs, social security payments (or the individu~‘s status with respect to future social security payments), status with respect to pension plans, various kinds of insurance, their costs and the kinds of events they insure against (e.g., house fire, car accident). the model sketched above is, in certain ways, akin to the worksheets published as guides to families; see, e.g., the wuil street journal (1989). the model differs from the worksheet in that the model allows for many of life’s random events-many more such events than one could take into account by filling out alternate, contingent worksheets. since future status is random, the simulated family must follow adaptive decision rules rather than a single plan as expressed on a worksheet. as noted above, a major function of the model is to evaluate these decision rules. individual versus htstftutional investing 5 this sketch of a game-of-life should suffice to convince one that the game is complex; most likely beyond analytic techniques. in contrast, using a good simulation language, it would not be difficult to program as a simulation model once the specs of the model are decided. it is unlikely that there will be general agreement as to what should be included in a game-of-life simulator, or how its output should be scored. therefore it may be expected that there will be more than one game-of-life simulator; and it may be hoped that their respective assumptions will be clearly documented. the various simulators will allow us to see whether rules of behavior which work well in one model will prove robust when tried in alternate models. if so, this will encourage us to recommend them in practice. in sum, i encourage readers with requisite skills to try building and using realistic game-of-life simulators; and editors to look kindly on the publication of their results. antithesis the problem with simulation analysis is that it is not very good at finding near optimum decision rules. it takes many runs of the model to estimate the excellence of a given set of rules. since the rules we seek may be adaptive i.e., may recommend different allocations of resources under different circumstances, and “circumstances” admit to countless variations-it will not be feasible to search for optimum decision rules. at various points in time in the game-of-life there are requirements for allocating resources among assets. if some of these can be formulated, at least approximately, as portfolio selection problems-where the problem is to get a good distribution of return on the allocation as a whole-then an optimum solution can be found for the approximate problem. if the approximation is a good one then, by definition of good approximation, the exact solution to the approximate problem will be part of a good solution for the more complex game. for example, consider a family with a house, children a few years from college age, life insurance policies in place based on a separate calculation, which faces the question of whether to shift resources among asset classes such as equities, long term tax exempt bonds, short term tax exempt bonds, etc. leaving aside, for the moment, the question of unrealized capital gains in the existing portfolio, and assuming that this family does not trade often enough to run up sizable brokerage commissions, then it is plausible to pretend that these assets are perfectly liquid, therefore the value of the portfolio at anytime is the sum of the market value of its constituents, and that the objective in choosing a portfolio of these assets is to get a good probability distribution of return (capital gain plus interest and dividends) for some period of analysis. 6 financial services review, l(1) 1991 whether the “goodness” of a probability distribution is to be measured by a utility function or by a mean-variance analysis, we must answer questions such as: (1) how do we measure return? clearly, the family wants return after taxes. first, if the family realizes capital gain by shifting out of an asset that has an unrealized gain at the beginning of the period, then the tax on this gain must be subtracted from holding period return. second, if an asset produces income during the period, we must subtract the tax on this income from its return. third, if an asset has a capital gain during the period then its value to the family is somewhere between its market value and the latter minus the tax if the gain is realized. for simplicity perhaps it is satisfactory to average these two values. finally, it seems appropriate for the family to seek a good distribution of real rather than nominal return. this raises no problem for the optimizer. (in particular, see markowitz 1987, chapter 11, concerning the treatment of real returns in a mean-variance analysis.) (2) what constraints limit portfolio choice? constraints should consist of those which are imposed by government agencies and brokerage houses on individuals, e.g., limited borrowing and short sales, plus perhaps self imposed constraints such as upper bounds on asset classes which are in fact less liquid than others. (3) what period of analysis should be used? do what we always do pick one. admittedly, approximations (and guesses) must be made, but they can be made plausibly. then the optimum solution can be found to the approximate model. if the approximation is satisfactory, this exact solution to the approximate problem should be part of a good solution for the real problem. synthesis the proposed optimization analysis is only an approximation. a realistic simulator could be used to test decision rules based on optimizing a simplified model as compared to rule-of-thumb decision rules. also, it is not always clear how the approximation is to be made; e.g., what time period to use for the analysis, how the family should pick a portfolio from the mean-variance frontier, how to treat unrealized capital gains, whether it is sufficient to consider nominal returns or essential to consider real returns, and the like. the simulator could be used to evaluate such alternate methods of form~ating the portfolio selection problem within the game-of-life model. also, a number of investigators zndividual versus znstitutionallnvesting 7 have evaluated the ability of a well chosen point from the mean-variance frontier to approximately maximize the expected value of a single period utility function.2 most have concluded that it does quite well for “reasonable” utility functions. this question could be re-examined within the framework of a game of-life simulation analysis. the exercise of building a realistic game-of-life simulator-deciding what is essential to the family planning process and incorporating it into a simulator without the severe constraint of producing an analytically tractable model should be highly educational, especially to the model builders. so should the process of formulating optimizable subproblems and evaluating these within the simulator. another challenge is to use modern computer technology to help understand and remember what has been done. i have in mind here the use of simulation/animation to display the workings of the simulated world (see caci, 1988) and the use of some kind of database to allow one to browse the inputs and outputs of prior runs. finally there is the process of deciding how the decision rules which prove robust in the simulated worlds can be explained and implemented in practice. i admit that this all seems a lot harder than formulating a highly simplified model that can be solved analytically. but i believe it has more chance of producing credible decision rules for practice-just as simulation analysis continues to produce credible policy recommendations for manufacturing, while analytic methods are not yet available for most sufficiently realistic models in the latter area. obviously, results of realistic game-of-life simulators will not be ready for the next issue of this journal. in the short and the long run, we should expect that the financial services review will publish research with various approaches to various aspects of its topic area. such pluralism is desirable in research, as it is in politics and the marketplace. notes 1. 2. in programming, it is often convenient to bundle events together into processes; but for the present discussion it is more convenient to describe events. markowitz (1959); young and trent (1969); levy and markowitz (1979); dexter, yu, and ziemba (1980); pulley (1981); pulley (1983); levy and markowitz (1984); reid and tew (1986); simaam (1987), grauer (1986); and tew and reid (1987). references caci. 1988. simgraphics: user’s guide and case book. la jolla, ca: caci products co. coffman, jr., e.g., m. hofri, and g. weiss. 1989. “scheduling stochastic jobs with a two point distribution on two parallel machines, ” in probability engineering and information science, forthcoming. 8 financial services review, l(1) 1991 dexter, a.s., j.n.w. yu, and w.t. ziemba. 1980. “portfolio selection in a lognormal market when the investor has a power utility function: computational results,” pp. 507-523 in m.a.h. dempster (ed.), stochastic programming. new york: academic press. ederington, l.h. 1986. “mean-variance as an approximation to expected utility maximization.” working paper 86-5, school of business administration, washington university, st. louis, missouri. grauer, r.r. 1986. “normality, solvency, and portfolio choice,” journal of financial and quantitative analysis, 21: 265-278. investment companies. 1944, new york: arthur wiesenberger & co. kiviat, p.j., r. villanueva, and h.m. markowitz. 1983. 77re simscript 11.5 programming language, e. russell (ed.). la jolla, ca: caci. kroll, y., h. levy, and h.m. markowitz. 1984. “mean-variance versus direct utility maximization,” journal of finance, 39: 47-61. lawler, e.l., j.k. lenstra, a.h.g. rinnooy kan, and d.b. shmoys. 1989. “sequencing and scheduling algorithms and complexity, ” in handbooks in operations research and management science. vol. 4, logistics of production and inventory, s.c. graves, a.h.g. rinnooy kan, and p. aipkin (eds.), forthcoming. new york: north holland. levy, h., and h.m. markowitz. 1979. “approximating expected utility by a function of mean and variance,” american economic review, 69: 308-3 17. markowitz, h.m. 1952. “portfolio selection,” the journal of finance, 7(l): 77-91. markowitz, h.m. 1959. portfolio selection: efficient diversification of investments. new york: wiley (yale university press, 1970, basil blackwell, 1991). markowitz, h.m. 1987. meanvariance analysis in portfolio choice and capital markets. new york: basil blackwell. pulley, l.b. 1981. “a general mean-variance approximation to expected utility for short holding periods,” journal of financial and quantitative analysis, 16: 361-373. pulley, l.b. 1983. “mean-variance approximations to expected logarithmic utility, operations research, 3 1: 685-696. reid, d.w., and b.v. tew. 1986. “mean-variance versus direct utility maximization: a comment,” journal of finance, 41: 1177-l 179. simaan, y. 1987. “portfolio selection and capital asset pricing for a class of non-spherical distributions of asset returns.” dissertation, baruch college, the city university of new york. tew, b.v., and d.w. reid. 1987. “more evidence on mean-variance versus direct utility maximization,” journal of financial research, 10: 249-257. von neumann, j., and 0. morgenstem. 1944. theory of games and economic behavior, 3rd edition, 1953. princeton university press. z7re wan street journal. 1989. “by the numbers,” in wsj reports: early retirement, december 8, r25-26. young, w.e., and r.h. trent. 1969. “geometric mean approximation of individual security and portfolio performance,” journal of financial and quantitative analysis, 4: 179-189. financial literacy to prevent poor borrowing choices terrance martina, janine k. samb,*, philip gibsonc adepartment of finance & economics, utah valley university, 800 west university parkway, orem ut 84058, usa bdepartment of business administration, shepherd university, 301 n king st, shepherdstown, wv 25443, usa ccollege of business administration, winthrop university, 202 carroll hall, rock hill, sc 29733, usa abstract working americans face the new reality of having to fund and manage their retirement while facing rising levels of indebtedness. a basic level of financial knowledge is essential to make good long-term financial decisions. using the 2015 national financial capacity study, we investigate the impact of financial literacy on the decision to access retirement plan loans before retirement or use one or more high-cost lenders. our results show that being financially literate reduces the likelihood of using high-cost lenders and using retirement-plan loans. furthermore, we find evidence of a negative relation between financial literacy and myopic spending. © 2021 academy of financial services. all rights reserved. jel classifications: g4; d14; d12 keywords: financial literacy; retirement plan loans; retirement planning; high-cost borrowing 1. introduction the use of defined contribution (dc) plans for retirement savings accumulation has increased significantly over the last 30 years. today, dc plans cover 90 million americans, with retirement assets totaling $6.7 trillion. acknowledging the long-run solvency issues facing the social security system in the united states, it is important and logical to assume that dc plans along with other tax-advantaged retirement accounts such as individual retirement accounts (iras) will be the main source of retirement wealth for americans in the future. the widespread adoption of plans such as a 401(k), leaves a growing number of *corresponding author: tel.: +1-304-876-5281; fax: +1-304-876-5193. e-mail address: jsam@shepherd.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 293–314 american workers with the responsibility of individually funding and managing this critical source of retirement income. munnell and webb (2015) found that individuals ten years or less from retirement had a combined average of only $111,000 in their dc plans. a possible explanation for such a modest average retirement savings balance, as outlined by munnell and webb (2015), is the american retirement savings system’s liquidity. in the united states—more than any other developed country—plan participants may access their retirement savings at any point during their life cycle (beshears, choi, hurwitz, laibson, & madrian, 2015). specifically, retirement plan participants may take a plan loan or hardship withdrawal against their retirement assets while working.1 total u.s. non-housing consumer debt reached $4 trillion at the end of 2016 (new york fed, 2017). the higher the level of debt on household balance sheets, the more likely households will be negatively impacted by economic shocks, such as a drop in income or change in home prices. to meet shortfalls while experiencing liquidity constraints, some american households have turned towards high-cost lenders such as payday and title loan providers. the pew charitable trust reported that 12 million americans spent $9 billion in payday loan fees and another $3 billion on auto title loans on an annual basis. the potential liquidity provided by retirement accounts and high-cost lenders is a doubleedged sword that offers short-term financial reprieve in the presence of an income shock, and the possibility of a loss of utility during present and future periods such as retirement (argento et al., 2015). using data from the national financial capacity study (nfcs) 2015 state-by-state tracking dataset, in this paper we study the use of two alternatives to conventional borrowing: retirement plan withdrawals (“leakage”) and accessing high-cost lenders. further, we investigate the impact of financial literacy on either accessing retirement savings before retirement or using high-cost lenders. we also study how financial literacy relates to optimal liquidity. while prior research associated with financial literacy have focused on retirement readiness from a wealth perspective, we offer an insight into household liquidity and debt management. thus, we make a notable contribution to the current literature on financial literacy. we created a measure of financial literacy and statistically linked the lack of financial knowledge to increased leakages via retirement plan loans, in addition to the inappropriate use of nonconventional borrowing. to our knowledge, this is the first study to explore this important topic. the remainder of the paper is organized as follows. we provide a review of the current literature in section 2. in section 3, we present the data and a description of how we derived our final sample. in section 4, we highlight the univariate analysis of our final sample. we then present our empirical results in section 5. in section 6, we discuss our main findings and discuss our conclusions and policy implications. 2. literature review the life-cycle hypothesis states that households attempt to maintain a constant present value of marginal utility of consumption over time to maximize expected lifetime utility 294 t. martin et al. / financial services review 29 (2021) 293–314 (modigliani & brumberg, 1954). this consumption smoothing can be achieved by transferring money from periods where the marginal utility of consumption is low, to periods where it is higher (i.e., borrowing when earnings are low—high marginal utility of consumption, saving when earnings are high—low marginal utility of consumption, and dissaving in retirement). households may need to withdraw funds from their retirement account(s) due to income shocks or special financial needs such as expenses associated with housing, college funding, a medical crisis, or meeting household needs after a job loss or change in occupation (argento, bryant, & sabelhaus, 2015; butrica, zedlewski, & issa, 2010; brady, 2011; copeland, 2009). an extensive review of the current literature on financial literacy and its impact on savings, investments, and debt management reveals that many individuals in the united states and worldwide are unfortunately financially illiterate (lusardi & mitchell, 2014). individuals need sufficient financial knowledge to make informed decisions in the present to maximize their chances for positive future outcomes; including the ability to recognize when they have made financial mistakes. current research provides ample evidence that financial mistakes are frequently made by individuals who exhibit low financial sophistication levels. bernheim (1998) demonstrated that financial literacy had a positive impact on retirement wealth accumulation. lusardi and mitchell (2007) indicated that individuals with low financial sophistication levels are less likely to think about retirement. according to van rooij, lusardi, and alessie (2007), individuals with low financial literacy levels are less likely to participate in the stock market. hastings and tejeda-ashton (2008) found that individuals with low financial sophistication levels are more likely to invest in mutual funds with high fees. it appears that american consumers, on average, are not using credit optimally. lusardi and tufano (2015) indicated that individuals with low financial knowledge levels are more likely to struggle with debt management, incurring higher fees, and using high-cost lenders. credit card debt revolvers often hold credit card debt while simultaneously holding low-interest liquid assets and retirement assets. this is a clear example of mental accounting and suboptimal credit use, a behavioral combination that conflicts with utility maximization (bertaut & haliassos, 2006). disney and gathergood (2012) used united kingdom household data to show that consumer credit customers underestimate borrowing costs. the authors also revealed that individuals who borrow on consumer credit tend to exhibit lower financial literacy. campbell (2006) highlighted a lower likelihood of refinancing mortgages during low-interest-rate periods among less educated and lower-income individuals. gerardi, goette, and meier (2013) provided evidence of a positive relation between low financial literacy and subprime mortgage adoption, as well as mortgage default. agarwal, skiba, and tobacman (2009) examined a group of payday loan and credit card users. the authors found that even in the presence of a more cost-effective liquidity option (a credit card), 66% of their sample still took out a payday loan. tang and lu (2014) used the nfcs as well as hypothetical debt scenarios to compare loan costs in funding consumption. they found that households were able to save up to 130% by switching from high-cost lenders such as payday loan companies to 401(k) plan loans. tang and lu (2014) concluded that consumers view 401(k) plan loans as a last resort when their liquidity is constrained. since loans from retirement plans carry lower interest rates than traditional sources of t. martin et al. / financial services review 29 (2021) 293–314 295 consumer loans, plan loans may be a more optimal choice (tang & lu, 2014; utkus & young, 2011). also, using the nfcs, lusardi and scheresberg (2013) examined high-cost borrowing methods and concluded that more financially literate individuals are less likely to have engaged in high-cost borrowing. the authors of that paper argue that (lack of) financial literacy plays an important role in explaining why individuals have used high-cost lenders such as payday loans. tang and lu (2014) showed the impact of optimal use of 401(k) plan loans on household balance sheets, but failed to consider the role of financial literacy in explaining why respondents were not utilizing 401(k) plans in periods of high marginal utility (high need) and low liquidity. lusardi and scheresberg (2013) made a powerful argument showing a negative relation between financial literacy and high-cost borrowing, but did not consider a comparison with low-cost borrowing. we contribute to the literature by investigating financial literacy and optimal borrowing choices. 3. data we use the 2015 nfcs state by state tracking dataset. the nfcs was commissioned and funded by the investor education foundation of the financial industry regulatory authority (finra, 2009). the research objectives of the nfcs were to benchmark key indicators of financial capacity and evaluate how these indicators vary with underlying demographic, behavioral, attitudinal, and financial literacy characteristics. consistent with surveys on financial capability that have been done in other countries (atkinson, mckay, kempson, & collard, 2007), the nfcs looks at multiple indicators of both financial knowledge and capacity, including how individuals manage their resources, how they make financial decisions, the skill sets they use in making decisions and the search-and-information elaboration that goes into making these decisions. the 2015 state by state tracking dataset pools 2009, 2012, and 2015 nfcs state-bystate surveys. for this paper, we only use the 2012 and 2015 pooled cross-sections. the new state-by-state tracking dataset provides some benefits not derived in a single-period crosssection. the observations are random and independent of each other at different points in time. consequently, serial correlation of residuals should not be an issue in the regression analysis. combining both waves of data results in a sample size of 53,703 respondents. to ensure a sufficient number of respondents for the analysis, african americans, hispanics, asian americans, and adults with less than a high school education are oversampled. 3.1. sample to ensure the internal validity of our results, we restricted our sample to respondents who reported having a retirement plan through their current or previous employer, and were able to choose the asset allocation of their retirement accounts. we also included respondents who reported having non-employer sponsored plans such as iras. to identify respondents who are in the accumulation stage of their life cycle, we further restricted our attention to 296 t. martin et al. / financial services review 29 (2021) 293–314 respondents who have a full-time job or are self-employed between the ages of 25-54. this resulted in a final sample of 10,560 respondents. 3.2. measuring financial literacy our primary predictor variable is financial literacy. respondents who participated in the 2012 and 2015 nfcs were asked five financial literacy questions. for the purposes of the study, we use the three most likely to be related to our research. the questions as stated in the survey include: 1. suppose you had $100 in a savings account and the interest rate was 2% per year. after five years, how much do you think you would have in the account if you left the money to grow? 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after one year, how much would you be able to buy with the money in this account? 3. buying a single company’s stock usually provides a safer return than a stock mutual fund. the first two questions were initially introduced in the 2004 health and retirement study by lusardi and mitchell (2011). subsequently, van rooij, lusardi, and alessie (2011) presented the question on bond pricing for a study carried out by the dutch central bank household survey. in prior studies, researchers have used the answers to some or all of these questions as proxies for financial sophistication (huston et al., 2012) by producing indices or other linear combinations (allgood & walstad, 2013). lusardi and scheresberg (2013) use these same questions to construct their proxy for financial literacy. 3.3. other key variables 3.3.1. retirement plan loan to capture defined-contribution leakage, this study uses two questions provided in the 2012 and 2015 nfcs. the first question used asks, “in the last 12 months, have you (or your spouse/partner) taken a loan from your retirement account(s)?” if respondents answered ‘yes’ to the question, they were assigned ‘1’ and if no, ‘0’. within the sample examined, 16% of respondents indicated taking a loan from their retirement account in the last 12 months.2 3.3.2. high-cost and inappropriate borrowing the survey includes a set of questions related to high-cost borrowing behavior. respondents were asked a series of questions about using any high-cost borrowing options in the past five years. the high-cost borrowing options included auto title loans, payday loans, advance tax refunds, rent-to-own consumer purchasing, and pawn shops. we focus on three areas known for excessive fees and high interests: payday loans, title loans, and t. martin et al. / financial services review 29 (2021) 293–314 297 pawnshops. if a respondent answered yes, they were assigned a ‘1’, if not a ‘0’. we then create a high-cost-of-borrowing variable that captures whether a respondent had used any of these types of loan options. we also create dummy variables to proxy for other forms of suboptimal borrowing, such as cash advances on credit cards, maxing out credit cards, and overdrafts on bank accounts. 3.3.3. other control variables the empirical literature shows that both household and market factors affect life-cycle behavior. if households forecast an increase in income, they may dissave to meet consumption needs; if households anticipate a drop in income, they may save more and consume below optimal consumption–more so if borrowing constraints persist. additionally, households with children may dissave to meet current consumption needs relative to households without children. homeownership is another factor that affects households’ life-cycle behavior. the nature of homeownership is to act as a forced savings mechanism, which reduces a family’s consumption over the mortgage term. health condition is another household characteristic that affects life-cycle behavior. if an individual has poor health with no health insurance, he/she may save more in anticipation of a health shock; contrastingly, households with members in good health can consume at higher levels.3 we also control for the effects of overspending on saving behavior by identifying those households that indicated spending more than their income within the previous 12 months. 4. descriptive results table 1 provides the frequency distribution of the full sample and by household groups. we can observe that most individuals with retirement plan loans incur high-cost borrowing (hcb), and only 22% are financially literate. hcb also varies with age; among those in the early stage of their careers (age 25–34), the probability of using hcb is 44%. furthermore, as age increases, the use of hcb decreases. we also see that hcb varies with income. the highest percentage of hcb users are earning between $50,000 and $74,999. table 2 shows the frequency distribution of respondents’ financial capacity and improper borrowing in the full sample and by household groups. we observe that individuals who do not have a retirement plan loan demonstrate greater financial capacity in other areas of their financial lives. for instance, the majority of them have a positive cash flow and an emergency fund. it is important to note that 68% of hcb are homeowners compared with the 45% who are not catgorized as hcb. this is a potential indication that owning property could lead to financial constraints, especially if the homeowner is not prepared for a sudden drop in income. we observe that individuals with retirement plan loans engage in other forms of borrowing, with 34% maxing out their credit cards, and 34% having a cash advance loan. in contrast, the proportion of individuals who do not have a retirement plan loan and engage in hcb, is small. results also show that 27% of hcb maxed out their credit cards, 32% have cash advances, and 48% reported having a recent bank overdraft. among those who did not engage in hcb, 4% maxed out their credit cards, 6% had cash advances, and 14% recently experienced a bank overdraft. 298 t. martin et al. / financial services review 29 (2021) 293–314 t ab le 1 f re q u en cy d is tr ib u ti o n s fo r th e w ei g h te d fu ll sa m p le an d b y h o u se h o ld g ro u p s v ar ia b le f u ll r et ir em en t p la n lo an n o re ti re m en t p la n lo an h ig h -c o st b o rr o w in g n o h ig h -c o st b o rr o w in g n = 1 0 ,5 6 0 n = 1 ,5 2 3 n = 9 ,0 3 7 n = 2 ,3 3 6 n = 8 ,2 2 4 l o an d ec is io n p la n lo an (l o w co st ) 1 6 % — — 3 5 % 9 % h ig h -c o st b o rr o w in g 2 4 % 5 5 % 1 9 % — — f in an ci al li te ra cy (a ll co rr ec t) 2 2 % 1 2 % 2 4 % 1 0 % 2 6 % (t o ta l co rr ec t) 3 .3 7 2 .7 3 3 .5 0 2 .7 2 3 .5 9 g en d er m al e 6 3 % 6 4 % 6 3 % 6 6 % 6 2 % f em al e 3 7 % 3 6 % 3 7 % 3 4 % 3 8 % r ac e w h it es 6 3 % 5 7 % 5 6 % 3 5 % 5 3 % n o n -w h it es 3 7 % 4 3 % 4 4 % 6 5 % 4 7 % a g e 2 5 – 3 4 3 0 % 4 1 % 2 8 % 4 4 % 2 5 % 3 5 – 4 4 3 3 % 3 2 % 3 3 % 3 3 % 3 3 % 4 5 – 5 4 3 7 % 2 7 % 3 9 % 2 3 % 4 2 % m ar ri ed 6 6 % 7 1 % 6 5 % 6 4 % 6 6 % in co m e l es s th an 1 5 k 1 % 2 % 1 % 2 % 1 % 1 5 k – 2 4 ,9 9 9 2 % 3 % 2 % 4 % 2 % 2 5 ,0 0 0 – 3 4 ,9 9 9 6 % 7 % 6 % 1 0 % 5 % 3 5 ,0 0 0 – 4 9 ,9 9 9 1 2 % 1 1 % 1 2 % 1 5 % 1 1 % 5 0 ,0 0 0 – 7 4 ,9 9 9 2 4 % 2 4 % 2 4 % 2 8 % 2 3 % 7 5 ,0 0 0 – 9 9 ,9 9 9 2 0 % 2 2 % 2 0 % 2 0 % 2 1 % 1 0 0 ,0 0 0 – 1 4 9 ,9 9 9 2 3 % 2 1 % 2 3 % 1 5 % 2 5 % 1 5 0 ,0 0 0 + 1 2 % 1 1 % 1 2 % 7 % 1 4 % n o . o f ch il d re n z er o 4 1 % 2 8 % 4 4 % 2 9 % 4 5 % o n e 2 3 % 2 8 % 2 2 % 2 6 % 2 2 % t w o 2 4 % 2 8 % 2 3 % 2 8 % 2 3 % t h re e 8 % 1 0 % 8 % 1 1 % 7 % f o u r+ 4 % 7 % 3 % 5 % 3 % (c o n ti n u ed o n n ex t p a g e) t. martin et al. / financial services review 29 (2021) 293–314 299 t ab le 1 (c o n ti n u ed ) v ar ia b le f u ll r et ir em en t p la n lo an n o re ti re m en t p la n lo an h ig h -c o st b o rr o w in g n o h ig h -c o st b o rr o w in g n = 1 0 ,5 6 0 n = 1 ,5 2 3 n = 9 ,0 3 7 n = 2 ,3 3 6 n = 8 ,2 2 4 e m p lo y m en t st at u s s el f 9 % 1 1 % 9 % 1 0 % 9 % f u ll 9 1 % 8 9 % 9 1 % 9 0 % 9 1 % f in an ci al ri sk to le ra n ce l o w es t ri sk 7 % 6 % 7 % 7 % 7 % r is k cl as s 2 1 4 % 1 1 % 1 5 % 1 1 % 1 5 % r is k cl as s 3 2 5 % 1 8 % 2 7 % 1 9 % 2 7 % r is k cl as s 4 3 5 % 2 8 % 3 7 % 3 1 % 3 7 % h ig h es t ri sk 1 8 % 3 8 % 1 4 % 3 2 % 1 3 % 300 t. martin et al. / financial services review 29 (2021) 293–314 table 3 shows correct answers to the financial literacy questions by the full sample and household groups. panel b highlights the distribution by loan type. we note that 84% of the total sample answered the interest question correctly, 68% answered the inflation question right, and 62% correctly answered the diversification question. most individuals characterized as hcb with a plan loan correctly responded to the interest rate question. however, only 44% of individuals with a plan loan correctly answered the inflation and the diversification question. similarly, only 45% of the hcb group correctly answered the inflation and the diversification question. we note that among individuals who do not engage in hcb nor have a plan loan, 88% answered the interest rate question correctly, 75% answered the inflation question correctly, and 67% answered the risk diversification question right. table 2 shows the weighted frequency distributions of respondents’ financial capacity and improper borrowing in the full sample and by household groups full retirement plan loan no retirement plan loan high-cost borrowing no high-cost borrowing variable n = 10,560 n= 1,523 n= 9,037 n= 2,336 n= 8,224 financial capacity cash surplus 49% 35% 51% 36% 53% emergency fund 57% 52% 58% 51% 59% drop in income 22% 50% 16% 42% 15% home owner 75% 79% 74% 68% 45% health insurance 95% 95% 95% 91% 96% improper borrowing maxed credit card 9% 34% 5% 27% 4% cash advance 12% 34% 8% 32% 6% bank overdraft 22% 55% 16% 48% 14% table 3 shows correct answers to the financial literacy questions by the full sample and household groups. panel b highlights the distribution by loan type panel a full plan loan no plan loan high-cost borrowing no high-cost borrowing interest rate question 84% 70% 87% 72% 88% inflation question 68% 44% 73% 45% 75% risk diversification question 62% 44% 65% 45% 67% n 10,560 1,523 9,037 2,336 8,224 panel b full both high-cost and plan loan high-cost only plan loan only neither interest rate question 84% 59% 78% 84% 89% inflation question 68% 25% 56% 67% 76% risk diversification question 62% 32% 53% 58% 68% n 10,560 790 9,037 733 7,491 t. martin et al. / financial services review 29 (2021) 293–314 301 t ab le 4 b in ar y lo g is ti c re g re ss io n o n th e li k el ih o o d o f ta k in g a re ti re m en t p la n lo an v ar ia b le o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p in te rc ep t 0 .2 6 * * * 0 .2 7 * * * 0 .1 9 * * * 0 .0 9 * * * 0 .0 9 * * * f in an ci al li te ra cy (a ll co rr ec t) 0 .3 7 * * * 0 .3 9 * * * 0 .4 2 * * * 0 .4 6 * * * 0 .5 6 * * * s o ci o -d em o g ra p h ic m al e 1 .2 5 * * * 1 .0 7 1 .1 5 * * 1 .1 5 * * w h it e 0 .7 8 * * * 0 .8 3 * * 0 .8 3 * * 0 .8 3 * * 3 5 – 4 4 0 .7 4 * * * 0 .8 1 * * 0 .8 0 * * 0 .8 1 * * 4 5 – 5 4 0 .6 4 * * * 0 .7 3 * * * 0 .7 1 * * * 0 .7 2 * * * m ar ri ed 1 .1 5 * 1 .2 0 * * 1 .1 0 1 .1 0 2 5 ,0 0 0 – 3 4 ,9 9 9 0 .7 6 0 .8 6 0 .9 3 0 .9 4 3 5 ,0 0 0 – 4 9 ,9 9 9 0 .7 4 * 0 .8 5 0 .9 6 0 .9 8 5 0 ,0 0 0 – 7 4 ,9 9 9 0 .7 3 * 0 .8 1 0 .9 6 0 .9 8 7 5 ,0 0 0 – 9 9 ,9 9 9 0 .8 2 0 .8 7 1 .1 8 1 .2 1 1 0 0 ,0 0 0 – 1 4 9 ,9 9 9 0 .7 5 0 .7 9 1 .2 0 1 .2 2 1 5 0 ,0 0 0 + 0 .7 2 * 0 .7 0 * 1 .2 0 1 .2 1 o n e 2 .0 0 * * * 1 .8 4 * * * 1 .6 2 * * * 1 .6 1 * * * t w o 1 .8 7 * * * 1 .7 2 * * * 1 .4 8 * * * 1 .4 7 * * * t h re e 1 .8 3 * * * 1 .6 7 * * * 1 .3 8 * * 1 .3 8 * * f o u r+ 2 .4 3 * * * 2 .1 6 * * * 1 .7 2 * * * 1 .7 2 * * * h o m e 1 .1 0 1 .0 9 f u ll 1 .0 7 1 .0 6 f in an ci al ri sk r is k cl as s 2 0 .9 9 1 .0 6 1 .0 5 r is k cl as s 3 1 .0 1 1 .1 7 1 .1 8 r is k cl as s 4 1 .1 1 1 .2 7 * 1 .2 7 * h ig h es t ri sk 3 .2 2 * * * 3 .1 5 * * * 3 .1 0 * * * f in an ci al ca p ac it y d ro p in in co m e 3 .5 8 * * * 4 .3 1 * * * e m er g en cy fu n d 0 .6 5 * * * 0 .6 3 * * * c as h su rp lu s 0 .6 6 * * * 0 .6 6 * * * h ea lt h in su ra n ce 1 .3 7 * * 1 .3 6 * * f in li t* d ro p in in co m e 0 .5 6 * * * y ea r 1 5 0 .9 4 0 .9 1 0 .8 3 * * 0 .8 9 * 0 .8 9 * r 2 0 .0 5 0 .0 9 0 .1 3 0 .2 1 0 .2 1 n o te : b in ar y lo g is ti cs re g re ss io n . r ef er en ce v ar ia b le s n o t in cl u d ed in re g re ss io n : ag e 2 5 – 3 4 , in co m e lo w er th an $ 2 5 ,0 0 0 , n o fi n an ci al ly d ep en d en t ch il d re n , se lf -e m p lo y ed , lo w es t ri sk to le ra n ce . * * * p < .0 1 , * * p < .0 5 , * p < .1 . 302 t. martin et al. / financial services review 29 (2021) 293–314 5. empirical results 5.1. financial literacy and retirement plan loans table 4 provides the logistic regression results, displaying the likelihood of respondents stating that they used a retirement-plan loan in the past 12 months. the first column of our table shows that those who are financially literate are 63% less likely to have a plan loan. in column 2, we add a set of demographic characteristics, such as a household’s estimated annual income and the number of children who financially depend on this income, gender, ethnicity, and age. we observe that financial literacy reduces the likelihood of taking out a plan loan by 61%. we also find that the use of plan loans varies strongly with gender, as males are 25% more likely to have a plan loan compared with females. moreover, being white decreases the likelihood of having a plan loan by 22% relative to nonwhites. we also observe a strong relationship between age and the likelihood of having a plan loan. as individuals get older, the odds of taking a retirement plan loan decreases. with regards to children, the presence of dependents within a household increases the odds of taking a retirement loan. in column 3, we add a set of variables that measure the level of risk tolerance. we find that respondents with the highest risk tolerance level are 222% more likely to have a plan loan than those in the lowest risk tolerance class. within the fourth column, we complete our model by taking into account financial capacity. therefore, we include variables that would provide some protection against an income or wealth shock. these variables include having an emergency fund, positive cash flow, and health insurance. our results suggest that individuals who have an emergency fund are 35% less likely to take a loan plan. likewise, individuals with a positive cash flow are 34% less likely to borrow from their retirement account. conversely, individuals who experience an income drop are 258% more likely to have a loan plan than those that have not reported one. respondents who have health insurance are 37% more likely to have a loan plan than those who do not have health insurance. this could be attributed to the presence of a high deductible health insurance plan. unfortunately, we are unable to compare different forms of health insurance due to data limitations. the negative relation between financial sophistication and the likelihood of having a plan loan is significant. we find that financial sophistication reduces the likelihood of having a plan loan by 44%. in column 5, we add interaction terms to see if financial literacy can offset the effects of a drop in income. we find that being financially literate and reporting a decline in income reduces the likelihood of having a plan loan by 44%, compared with those who are not financially literate. 5.2. financial literacy and hcb table 5 provides logistic regression results on the likelihood of respondents stating that they engaged in hcb. the first column accounts for hcb in general. we see that financial literacy reduces the likelihood of engaging in hcb by 58%. we observe that as income increase, the odds of accessing a high-cost loan decrease. similar results are found as the number of financially dependent children within a household increases. those who own their t. martin et al. / financial services review 29 (2021) 293–314 303 t ab le 5 b in ar y lo g is ti c re g re ss io n o n th e li k el ih o o d o f h ig h -c o st b o rr o w in g u si n g h ig h -c o st b o rr o w in g u si n g au to ti tl e lo an s u si n g a p aw n sh o p u si n g p ay d ay lo an v ar ia b le o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p in te rc ep t 0 .9 2 0 .1 2 * * * 0 .3 5 * * * 0 .3 7 * * * f in an ci al li te ra cy (a ll co rr ec t) 0 .4 2 * * * 0 .3 4 * * * 0 .3 7 * * * 0 .3 0 * * * s o ci o -d em o g ra p h ic m al e 1 .2 9 * * * 1 .4 0 * * * 1 .4 7 * * * 1 .3 5 * * * w h it e 0 .7 6 * * * 0 .9 1 0 .8 1 * * 0 .6 3 * * * 3 5 – 4 4 0 .7 4 * * * 0 .6 1 * * * 0 .7 1 * * * 0 .7 0 * * * 4 5 – 5 4 0 .5 8 * * * 0 .4 9 * * * 0 .5 2 * * * 0 .5 0 * * * m ar ri ed 0 .9 9 1 .1 8 * * 0 .8 8 1 .0 0 2 5 ,0 0 0 – 3 4 ,9 9 9 0 .8 6 0 .8 4 0 .7 0 * 0 .9 7 3 5 ,0 0 0 – 4 9 ,9 9 9 0 .6 6 * * 0 .7 2 * 0 .5 3 * * * 0 .8 9 5 0 ,0 0 0 – 7 4 ,9 9 9 0 .6 5 * * 0 .5 9 * * 0 .5 8 * * * 0 .7 7 7 5 ,0 0 0 – 9 9 ,9 9 9 0 .5 4 * * * 0 .6 3 * * 0 .4 3 * * * 0 .6 5 * * 1 0 0 ,0 0 0 – 1 4 9 ,9 9 9 0 .4 0 * * * 0 .5 4 * * 0 .3 0 * * * 0 .5 6 * * 1 5 0 ,0 0 0 + 0 .3 1 * * * 0 .5 0 * * 0 .2 4 * * * 0 .4 4 * * * o n e 1 .7 9 * * * 1 .6 1 * * * 2 .0 6 * * * 2 .0 7 * * * t w o 1 .9 9 * * * 1 .7 9 * * * 2 .1 5 * * * 2 .3 0 * * * t h re e 2 .4 2 * * * 1 .6 0 * * * 2 .5 9 * * * 2 .5 8 * * * f o u r+ 1 .8 1 * * * 1 .9 3 * * * 2 .0 2 * * * 2 .5 6 * * * h o m e 0 .6 8 * * * 1 .1 4 0 .8 0 * * 0 .6 2 * * * f u ll 1 .1 2 1 .2 0 0 .9 6 1 .1 9 f in an ci al ri sk r is k cl as s 2 1 .0 3 1 .3 5 1 .1 4 0 .8 1 r is k cl as s 3 1 .1 3 1 .4 7 * * 1 .4 0 * * 1 .0 6 r is k cl as s 4 1 .4 6 * * * 2 .2 6 * * * 1 .7 6 * * * 1 .5 3 * * h ig h es t ri sk 3 .0 8 * * * 5 .0 0 * * * 4 .0 6 * * * 3 .8 8 * * * f in an ci al ca p ac it y d ro p in in co m e 2 .6 0 * * * 3 .0 8 * * * 2 .8 9 * * * 3 .3 9 * * * e m er g en cy fu n d 0 .7 6 * * * 1 .1 1 1 .1 3 * 0 .6 3 * * * c as h su rp lu s 0 .6 9 * * * 0 .7 1 * * * 0 .6 9 * * * 0 .7 2 * * * h ea lt h in su ra n ce 0 .6 1 * * * 0 .4 6 * * * 0 .5 8 * * * 0 .6 6 * * * y ea r 1 5 1 .0 3 1 .3 0 * * * 1 .1 2 * 0 .9 8 r 2 0 .2 6 0 .2 7 0 .2 8 0 .3 1 n o te : b in ar y lo g is ti cs re g re ss io n . r ef er en ce v ar ia b le s n o t in cl u d ed in re g re ss io n : ag e 2 5 – 3 4 , in co m e lo w er th an $ 2 5 ,0 0 0 , n o fi n an ci al ly d ep en d en t ch il d re n , se lf -e m p lo y ed , lo w es t ri sk to le ra n ce . * * * p < .0 1 , * * p < .0 5 , * p < .1 . 304 t. martin et al. / financial services review 29 (2021) 293–314 home are 32% less likely to engage in hcb. consistent with the expectation that a drop in income might lead someone to participate in hcb, we find that a decline in income increases the likelihood of engaging in hcb by 160%, relative to respondents with stable or increasing incomes. accounting for financial capacity, we see that having an emergency fund decreases the likelihood of hcb by 24%. likewise, the presence of a positive cash flow and health insurance reduces the likelihood of hcb. column 2 provides results from a logistic regression on the likelihood of respondents stating that they have an auto title loan. we observe a strong negative relation between financial literacy and having a title loan. financial literacy reduces the likelihood of having a title loan by 66%. respondents with a recent drop in income are more than twice as likely to have an auto title loan than respondents with no such income shock. having a cash surplus reduces having a title loan by 29%, and having health insurance reduces the likelihood of having that type of loan by 54%. column 3 provides results from the logistic regression analysis, showing the likelihood of respondents using pawn shops in the past five years. the likelihood of taking a loan from a pawn shop decreases by 63% with financial literacy. individuals who experience a drop in income are 189% more likely to have a pawn shop loan. having a positive cash flow reduces the likelihood of taking a loan from a pawn shop by 31%. similarly, having health insurance reduces it by 42%. column 4 then provides results from a logistic regression on the likelihood of respondents stating that they have a payday loan. financial literacy decreases the likelihood of having a payday loan by 70%. as expected, those who experience a drop in income are 239% more likely to access a pawn shop loan than respondents whose income remained stable. 5.3. financial literacy and improper borrowing table 6 provides results from six binary logistic regression analyses displaying the likelihood of a respondent engaging in improper borrowing. specifically, we focus on the likelihood of evidence of myopic spending such as taking a cash advance on a credit card, incurring a bank overdraft, or maxing out a credit card. we then included an interaction term (financially literate*drop in income) and reran the analyses using our empirical models. results presented in column 1a shows that financial literacy reduces the likelihood of maxing out credit cards by 63%. those who have a drop in income in the last year are 189% more likely to max out their credit cards, while a positive cash flow reduces the likelihood by 31%. in column 1b, we observe the relation between our interaction term and the likelihood of maxing out credit cards. even when faced with an income shock, we find that being financially literate reduces the likelihood of maxing out credit cards by 38% compared with non-financially literate individuals. as shown in column 2a, financial literacy reduces the likelihood of having a bank overdraft by 43%. individuals who experience a drop in income are nearly three times as likely to have an overdraft. having an emergency fund decreases the likelihood of overdraft by 53%, and a positive cash flow reduces the likelihood by 52%. in column 2b, we add interaction terms to account for the effects of an income drop plus financial literacy on the likelihood of having an overdraft. we see that when faced with a t. martin et al. / financial services review 29 (2021) 293–314 305 t ab le 6 b in ar y lo g is ti c re g re ss io n o n th e li k el ih o o d o f th re e fo rm s o f in ap p ro p ri at e b o rr o w in g m ax ed c c m ax ed c c w / in te ra ct io n s o v er d ra ft o v er d ra ft w / in te ra ct io n s c as h ad v o n c c c as h ad v o n c c w / in te ra ct io n s v ar ia b le o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p in te rc ep t 0 .3 5 * * * 0 .1 3 * * * 0 .4 0 * * * 0 .3 8 * * * 0 .1 8 * * * 0 .1 7 * * * f in an ci al li te ra cy (a ll co rr ec t) 0 .3 7 * * * 0 .4 8 * * * 0 .5 7 * * * 0 .6 7 * * * 0 .4 4 * * * 0 .5 2 * * * f in an ci al li te ra cy (a ll co rr ec t) 0 .3 7 * * * 0 .4 8 * * * 0 .5 7 * * * 0 .6 7 * * * 0 .4 4 * * * 0 .5 2 * * * s o ci o -d em o g ra p h ic m al e 1 .4 7 * * * 1 .0 4 0 .9 6 0 .9 6 1 .2 4 * * 1 .2 4 * * w h it e 0 .8 1 * * 0 .8 3 * * 0 .8 0 * * * 0 .7 9 * * * 0 .8 1 * * 0 .8 1 * * 3 5 – 4 4 0 .7 1 * * * 0 .6 7 * * * 0 .9 1 0 .9 2 0 .7 1 * * * 0 .7 2 * * * 4 5 – 5 4 0 .5 2 * * * 0 .4 6 * * * 0 .6 8 * * * 0 .6 9 * * * 0 .6 8 * * * 0 .6 9 * * * m ar ri ed 0 .8 8 1 .0 5 1 .0 9 1 .0 9 0 .7 8 * * 0 .7 8 * * 2 5 ,0 0 0 – 3 4 ,9 9 9 0 .7 0 * 0 .6 7 * 0 .7 7 0 .7 8 0 .9 1 0 .9 2 3 5 ,0 0 0 – 4 9 ,9 9 9 0 .5 3 * * * 0 .7 0 * 0 .7 9 0 .8 0 0 .8 9 0 .9 0 5 0 ,0 0 0 – 7 4 ,9 9 9 0 .5 8 * * * 0 .5 5 * * 0 .7 4 * 0 .7 6 * 0 .7 5 0 .7 6 7 5 ,0 0 0 – 9 9 ,9 9 9 0 .4 3 * * * 0 .6 0 * * 0 .8 1 0 .8 3 0 .7 0 * 0 .7 1 * 1 0 0 ,0 0 0 – 1 4 9 ,9 9 9 0 .3 0 * * * 0 .6 3 * * 0 .7 7 0 .7 9 0 .6 7 * * 0 .6 7 * * 1 5 0 ,0 0 0 + 0 .2 4 * * * 0 .4 8 * * 0 .6 5 * * 0 .6 5 * * 0 .5 6 * * 0 .5 6 * * o n e 2 .0 6 * * * 1 .5 2 * * * 1 .5 5 * * * 1 .5 4 * * * 1 .4 9 * * * 1 .4 9 * * * t w o 2 .1 5 * * * 1 .7 4 * * * 1 .6 9 * * * 1 .6 8 * * * 1 .7 6 * * * 1 .7 6 * * * t h re e 2 .5 9 * * * 1 .6 0 * * 1 .7 4 * * * 1 .7 5 * * * 1 .6 1 * * * 1 .6 1 * * * f o u r+ 2 .0 2 * * * 2 .4 6 * * * 2 .2 8 * * * 2 .2 7 * * * 1 .8 2 * * * 1 .8 2 * * * h o m e 0 .8 0 * * 1 .1 8 * 0 .9 8 0 .9 7 1 .2 1 * * 1 .2 0 * * f u ll 0 .9 6 1 .0 2 1 .0 1 1 .0 1 0 .7 5 * * 0 .7 5 * * f in an ci al ri sk r is k cl as s 2 1 .1 4 1 .2 8 1 .3 7 * * 1 .3 7 * * 1 .2 1 1 .2 1 r is k cl as s 3 1 .4 0 * * 1 .5 5 * * 1 .4 1 * * 1 .4 2 * * 1 .3 6 * 1 .3 6 * r is k cl as s 4 1 .7 6 * * * 1 .7 5 * * 1 .6 2 * * * 1 .6 1 * * * 1 .9 2 * * * 1 .9 1 * * * h ig h es t ri sk 4 .0 6 * * * 5 .1 1 * * * 3 .1 6 * * * 3 .1 0 * * * 4 .5 6 * * * 4 .5 2 * * * f in an ci al ca p ac it y (c o n ti n u ed o n n ex t p a g e) 306 t. martin et al. / financial services review 29 (2021) 293–314 t ab le 6 (c o n ti n u ed ) m ax ed c c m ax ed c c w / in te ra ct io n s o v er d ra ft o v er d ra ft w / in te ra ct io n s c as h ad v o n c c c as h ad v o n c c w / in te ra ct io n s v ar ia b le o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p o d d s ra ti o p d ro p in in co m e 2 .8 9 * * * 4 .1 0 * * * 2 .8 3 * * * 3 .4 5 * * * 2 .4 3 * * * 2 .7 8 * * * e m er g en cy fu n d 1 .1 3 * 0 .6 9 * * * 0 .4 7 * * * 0 .4 7 * * * 0 .8 5 * * 0 .8 3 * * c as h su rp lu s 0 .6 9 * * * 0 .5 9 * * * 0 .4 8 * * * 0 .4 8 * * * 0 .6 3 * * * 0 .6 2 * * * h ea lt h in su ra n ce 0 .5 8 * * * 0 .6 8 * * 0 .9 7 0 .9 6 0 .7 9 * 0 .7 8 * f in li t* d ro p in co m e 0 .6 2 * * 0 .5 8 * * * 0 .6 0 * * y ea r 1 5 1 .1 2 * 0 .8 9 0 .9 2 0 .9 1 * 1 .0 1 1 .0 0 r 2 0 .2 6 0 .2 7 0 .2 2 0 .2 3 0 .1 9 0 .2 0 n o te : b in ar y lo g is ti cs re g re ss io n . r ef er en ce v ar ia b le s n o t in cl u d ed in re g re ss io n : ag e 2 5 -3 4 , in co m e lo w er th an $ 2 5 ,0 0 0 , n o fi n an ci al ly d ep en d en t ch il d re n , se lf -e m p lo y ed , lo w es t ri sk to le ra n ce . c c = cr ed it ca rd ; ad v = ad v an ce . * * * p < .0 1 , * * p < .0 5 , * p < .1 . t. martin et al. / financial services review 29 (2021) 293–314 307 decline in income, financially literate individuals are still 42% less likely to have an overdraft. in column 3a, we observe a strong negative relation between financial literacy and the likelihood of taking a cash advance; those who are financially literate are 56% less likely to do so. males compared with females are 24% more likely to get a cash advance loan, and whites are 19% less likely than non-whites. respondents indicating a drop in income in the last 12 months are 143% more likely to have to take a cash advance loan than those with stable incomes. individuals who report having an emergency fund are 15% less likely to take cash advance than those who have no such funds, and those who have a positive cash flow are 37% less likely relative to those who have a deficit in spending. in column 3b, we add interaction terms to account for the effects of having both a drop in income and financial sophistication simultaneously. we observe that individuals with financial literacy and a decrease in income are 40% less likely to take a cash advance loan. 5.4. financial literacy and loan choice table 7 provides the results of a multinomial logistic regression. the dependent variable is loan type with four levels: high-cost loan and plan loan, high-cost loan only, plan loan only, and no loan. ‘plan loan only’ is the base reference for this table and the rest of this section. in comparison to having only a retirement plan loan, financially literate respondents are 65% less likely to have taken a high-cost loan and a retirement plan loan, 20% less likely to have taken only a high-cost loan, but are 58% more likely to have neither loan type relative to non-financially literate respondents. in column b, we include an interaction variable, financial literacy*drop in income. we observe that financial literacy appears to exert a moderating effect on decision-making, even when individuals are faced with a decline in income. financially literate respondents with a recent drop in income are 52% less likely to take out a high-cost loan alternative such as a payday loan, in addition to a retirement plan loan. recall that in column a, a drop in income was strongly associated with choosing to have both types of loans. 6. conclusion in an era of higher debt levels, working americans bear greater responsibility for saving for retirement, while being at odds with traditional lenders. using data from the nfcs 2015 state-by-state tracking dataset, in this paper, we study the use of two alternatives to conventional borrowing: retirement plan loans (defined as “leakage”) and high-cost lenders. furthermore, we investigate the impact of financial literacy on either accessing retirement savings before retirement or using high-cost lenders. we also touch on how financial literacy relates to optimal liquidity in a mental accounting context. first, we show that one in four americans are likely to borrow money from high-cost lenders, while one in six turn to their retirement plans for loans. high-cost borrowers and plan loan users display lower levels of financial literacy. when we look specifically at how 308 t. martin et al. / financial services review 29 (2021) 293–314 table 7 multinomial regression on the likelihood of loan type with plan loan only being the reference variable odds ratio p odds ratio financial literacy (all correct) 0.35 *** 0.49 *** financial literacy (all correct) 0.80 ** 0.78 ** financial literacy (all correct) 1.58 *** 1.52 *** socio-demographic male 1.60 *** 1.60 *** male 1.27 ** 1.27 ** male 1.05 1.05 white 0.86 0.86 white 0.92 0.92 white 1.21 ** 1.21 ** 35–44 0.63 *** 0.64 *** 35–44 0.81 * 0.81 * 35–44 1.01 1.01 45–54 0.37 *** 0.37 *** 45–54 0.66 *** 0.66 *** 45–54 0.94 0.94 married 0.90 0.91 married 0.87 0.87 married 0.87 0.87 25,000–34,999 0.90 0.92 25,000–34,999 0.90 0.90 25,000–34,999 1.04 1.04 35,000–49,999 0.67 0.68 35,000–49,999 0.59 * 0.59 * 35,000–49,999 0.92 0.92 50,000–74,999 0.68 0.68 50,000–74,999 0.59 * 0.59 * 50,000–74,999 0.95 0.94 75,000–99,999 0.52 * 0.53 * 75,000–99,999 0.35 *** 0.35 *** 75,000–99,999 0.72 0.72 100,000–149,999 0.41 ** 0.41 ** 100,000–149,999 0.24 *** 0.24 *** 100,000–149,999 0.69 0.68 150,000+ 0.33 ** 0.33 ** 150,000+ 0.17 *** 0.17 *** 150,000+ 0.65 0.65 one 1.49 ** 1.49 ** one 1.12 1.12 one 0.65 *** 0.65 *** two 1.96 *** 1.96 *** two 1.47 ** 1.48 ** two 0.79 ** 0.79 ** three 1.90 ** 1.92 ** three 1.89 *** 1.89 *** three 0.76 * 0.76 * four+ 1.61 * 1.61 * four+ 1.09 1.09 four+ 0.65 ** 0.65 ** home 0.93 0.92 home 0.64 *** 0.64 *** home 1.07 1.07 (continued on next page) t. martin et al. / financial services review 29 (2021) 293–314 309 respondents answer each question, plan loan users were more likely to correctly respond to the interest rate question but not the questions on inflation and risk diversification. we observe similar results among high-cost borrowers. nevertheless, if respondents only used plan loans and failed to use any high-cost loans in the last five years, they were more likely to answer the questions correctly, when compared with respondents that had both types of loans or high-cost loans only. this provides evidence of higher financial knowledge among respondents who only use plan loans. table 7 (continued) variable odds ratio p odds ratio full 0.99 0.99 full 0.94 0.94 full 0.88 0.89 financial risk risk class 2 1.68 * 1.68 * risk class 2 1.05 1.05 risk class 2 1.15 1.15 risk class 3 1.87 ** 1.89 ** risk class 3 1.04 1.04 risk class 3 1.05 1.05 risk class 4 2.25 ** 2.25 ** risk class 4 1.28 1.28 risk class 4 0.99 0.99 highest risk 4.73 *** 4.75 *** highest risk 1.18 1.19 highest risk 0.56 *** 0.56 *** financial capacity drop in income 2.66 *** 3.00 *** drop in income 0.79 ** 0.76 ** drop in income 0.39 *** 0.36 *** emergency fund 1.62 *** 1.59 *** emergency fund 1.44 *** 1.44 *** emergency fund 2.21 *** 2.22 *** cash surplus 0.88 0.87 cash surplus 1.09 1.09 cash surplus 1.53 *** 1.53 *** health insurance �0.19 0.82 health insurance �0.78 0.46 *** health insurance �0.15 0.86 fin lit*drop income 0.48 ** fin lit*drop income 1.09 fin lit*drop income 1.21 year 15 1.45 ** 1.44 ** year 15 1.28 ** 1.28 ** year 15 1.37 *** 1.37 *** note: multinomial logistics regression. the dependent variable is loan type with four levels: high-cost loan and plan loan, high-cost loan only, plan loan only, neither loan. plan loan only is the based reference. reference variables not included in regression: age 25-34, income lower than $25,000, no financially dependent children, self-employed, lowest risk tolerance. ***p< .01, **p< .05, *p< .1. 310 t. martin et al. / financial services review 29 (2021) 293–314 our empirical results indicate that individuals using retirement plan loans are more likely to be younger males, to be non-white, to have children, to have seen a recent drop in income, and to have an affinity for financial risk-taking. when determining the likelihood of having a retirement plan loan, income appears to have no effect while having health insurance increases the use of plan loans; however, we provide evidence that users of high-cost lenders are less likely to have incomes at or above $35,000, and less likely to borrow to meet health shocks. moreover, the findings on the effect of financial capacity on these two loan alternatives show that respondents who report deficit spending and no precautionary savings are more likely to borrow from their retirement plan and/or high-cost lenders. notably, when looking individually at the three types of high-cost lenders, we observe differences in the effect of having an emergency fund. in comparison with respondents with no emergency funds, those with precautionary savings are less likely to have a payday loan. when we consider other forms of high-cost borrowing and conduct regression analyses, we find a higher likelihood of maxing out credit cards, having bank overdrafts, and taking cash advances on credit cards. this behavior is specifically displayed among nonwhites, individuals with kids, those with an increased tolerance for financial risk, and those who have had a recent drop in income. also, homeowners appear more likely to turn to credit cards to meet short-term liquidity needs. a striking finding across all of our empirical analyses is the strong effect of financial literacy on borrowing decisions. financial literacy plays a significant role in explaining why individuals use high-cost lenders and retirement plan loans. specifically, individuals who could answer all of the financial literacy questions correctly (deemed to be “financially literate”) were unlikely to turn to high-cost lenders or retirement plans as a loan option. these findings are consistent with lusardi and mitchell (2014) and disney and gathergood (2012), where both studies highlighted that users of high-cost debt are more likely to lack financial knowledge. we also find that financial literacy explains myopic spending patterns. the financially literate are unlikely to turn to credit cards and bank overdrafts when faced with income shortfalls. we do not ignore the effect of a recent drop in income on borrowing choices. noteworthy, we find that financial literacy is a moderating factor even when faced with a loss of income. this is an important finding because it provides evidence that individuals with high levels of financial literacy are unlikely to turn to high-cost borrowers even in the face of an income shock. the authors of this study acknowledge that retirement plan loans may allow households to increase consumption during periods when they have a liquidity constraint. the higher a respondent’s financial literacy level, the less likely they are to use plan loans to smooth consumption, even when faced with an income shock. however, when confronted with a choice of borrowing from either high-cost lenders or retirement plans, financially literate respondents are less likely to choose any loan option that includes a high-cost lender, even when faced with an income shock. tang and lu (2014) showcased that individuals view retirement plan loans as a last resort. it is apparent from the results of this study that financial literacy may explain reduced plan loan use among individuals with a high level of financial knowledge. t. martin et al. / financial services review 29 (2021) 293–314 311 7. implications undoubtedly, saving for retirement and effectively managing debt are involved processes that require financial knowledge. the u.s. retirement savings system provides plan participants with significant liquidity regardless of economic circumstance (beshears et al., 2015). as such, access to income via retirement saving schemes leads to significant preretirement leakage—even in the face of the 10% penalty and tax liability at one’s marginal tax rate (federal and/or state) of the amount withdrawn. by 2012, 25% of american households had some form of retirement plan leakage amounting to 70 billion dollars annually, while plan participants were only contributing $175 billion (fellowes & willemin, 2013). essentially, for every $1 saved, $0.40 is withdrawn dc plans (argento, bryant, & sabelhaus, 2015). plan loans account for a significant amount of the annual defined contribution plan leakage, but plan administrators do not adequately discuss the long-term impact on accumulated retirement wealth (gao, 2009). individuals may not be using plan loans to meet economic shocks, but due to time-inconsistent preferences. individuals who display low financial literacy show evidence of improper borrowing by maxing out credit cards and having bank overdrafts. if we were to add managing a retirement plan loan to this mix, along with a job, defaults are more likely. as such, there appears to be a need for greater financial education in the workplace, particularly among those who opt to use their retirement plans before retirement, to mitigate not just define contribution leakage, but myopic spending. in other words, some individuals might need to take a loan from their retirement account to meet their basic living expenses. however, they should not remove the funds from their account to take a vacation or buy a sports car. the pew charitable trust reports that annually 12 million americans spend $9 billion in payday loan fees and another $3 billion on auto title loans. on average, a borrower takes out eight loans per year and pays more in interest payments than the original principal. additionally, they find that payday loans are not used for unexpected income shocks or unforeseen expenses, but to meet daily living expenses possibly caused by myopic spending. agarwal, skiba, and tobacman (2009) find that even in the presence of more cost-effective liquidity options (a credit card) 66% of their sample still took out a payday loan. the results of this study indicate that financial literacy can help explain such gross debt mismanagement. from our analysis, it is apparent that most, if not all, of the fees paid to high-cost lenders are from individuals with low levels of financial literacy. therefore, there is a need for greater financial education and financial literacy among groups more likely to use these types of high-interest loans. payday and auto title loan companies argue that they provide a service for the underserved. although we do not argue that they provide a needed service, there is evidence that they are hurting those they claim to help. we agree with the recommendation by the consumer financial protection bureau to cap repayment levels, and clarify loan terms, but we also add that these lenders can be a catalyst for irresponsible money management. 312 t. martin et al. / financial services review 29 (2021) 293–314 notes 1 according to vanguard’s 2016 defined contribution survey, in 2015, 78% of all dc plans allowed plan loans and 84% allowed hardship withdrawals. 2 there are two other forms of defined contribution leakage, namely cashouts and inservice withdrawals. the nfcs allows us to explore in-service withdrawals further. 3 see yuh and hanna (2010) for a thorough discussion. references agarwal, s., skiba, p. m., & tobacman, j. (2009). payday loans and credit cards: new liquidity and credit scoring puzzles? (no. w14659). cambridge, ma: national bureau of economic research. allgood, s., & walstad, w. (2013). financial literacy and credit card behaviors: a cross-sectional analysis by age. numeracy, 6, 3. argento, r., bryant, v. l., & sabelhaus, j. (2015). early withdrawals from retirement accounts during the great recession. contemporary economic policy, 33, 1–16. atkinson, a., mckay, s., collard, s., & kempson, e. (2007). levels of financial capability in the uk. public money and management, 27, 29–36. beshears, j., choi, j. j., hurwitz, j., laibson, d., & madrian, b. c. (2015). liquidity in retirement savings systems: an international comparison. the american economic treview, 105, 420–425. bertaut, c., & haliassos, m. (2006). credit cards: facts and theories. in g. bertola, r. disney, & c. grant (eds.), the economics of consumer credit (pp. 181–238). cambridge, ma: mit press. bernheim, d. d. (1998). financial illiteracy, education, and retirement saving (no. 96-97). philadelphia, pa: wharton school pension research council, university of pennsylvania. brady, p. (2011). the role of iras in the retirement income process. washington, dc: investment company institute. butrica, b. a., zedlewski, s. r., & issa, p. (2010). understanding early withdrawals from retirement accounts. washington, dc: the urban institute. campbell, j. y. (2006). household finance. the journal of finance, 61, 1553–1604. copeland, c. (2009). lump-sum distributions at job change. ebri notes, 30(1), 1–12. disney, r., & gathergood, j. (2012). financial literacy and consumer credit use (no. 12/01). nottingham: university of nottingham, centre for finance, credit and macroeconomics (cfcm). fellowes, m., & willemin, k. (2013). the retirement breach in defined contribution plans: size, causes, and solutions.washington, dc: hello wallet. the federal reserve bank of new york. (2017, february 16). household debt increases substantially, approaching previous peak. available at https://www.newyorkfed.org/newsevents/news/research/2017/rp170216 finra (financial industry regulatory authority). (2009). initial report on the national survey component of the national financial capacity study. washington, dc: finra investor education foundation. available at http://www.finrafoundation.org/web/groups/foundation/@foundation/documents/foundation/p120536.pdf gerardi, k., goette, l., & meier, s. (2013). numerical ability predicts mortgage default. proceedings of the national academy of sciences of the united states of america, 110, 11267–11271. hastings, j. s., & tejeda-ashton, l. (2008). financial literacy, information, and demand elasticity: survey and experimental evidence from mexico (no. w14538). cambridge, ma: national bureau of economic research. huston, s. j., finke, m. s., & smith, h. (2012). a financial sophistication proxy for the survey of consumer finances. applied economics letters, 19, 1275–1278. lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. s. (2011). financial literacy around the world: an overview (no. w17107). cambridge, ma: national bureau of economic research. t. martin et al. / financial services review 29 (2021) 293–314 313 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. lusardi, a., & scheresberg, c. d. b. (2013). financial literacy and high-cost borrowing in the united states (no. w18969). cambridge, ma: national bureau of economic research lusardi, a., & tufano, p. (2015). debt literacy, financial experience, and overindebtedness. journal of pension economics and finance, 14, 332–368. modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. in k. k. kurihara (ed.), post-keynesian economics (pp. 388–436). new brunswick, nj: rutgers university press. munnell, a. h., & webb, a. (2015). the impact of leakages on 401 (k)/ira assets. chestnut hill, ma: center for retirement research at boston college brief. tang, n., & lu, t. (2014). are your clients making the right loan choice? journal of financial planning, 27, 39–47. utkus, s. p., & young, j. a. (2011). financial literacy and 401(k) loans. in financial literacy: implications for retirement security and the financial marketplace (pp. 59–75). new york, ny: oxford university press. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. yuh, y., & hanna, s. d. (2010). which households think they save? journal of consumer affairs, 44, 70–97. 314 t. martin et al. / financial services review 29 (2021) 293–314 impact of consumer perceptions of industry corruption on the choice to engage a financial advisor: does gender matter? danielle d. winchestera,*, roland l. leaka, nicole r. mccoya anorth carolina agricultural and technical state university, willie a. deese college of business and economics, greensboro, nc 27411, usa abstract consumer perceptions of industry corruption influences with whom and how they are willing to engage. this study explores the intersection of financial advisors’ gender and consumers’ industry corruption perceptions on the likelihood of using a female advisor as females are perceived as more trustworthy and less prone to corruption than their male counterparts. analyses reveal individuals prefer female advisors when corruption is low, but these preferences wane as corruption perceptions heighten. this suggests the interpersonal characteristics of females being more trustworthy and ethical do not carry as much weight for consumers when they perceive the industry as corrupt. © 2022 academy of financial services. all rights reserved. jel classification: d1 household behavior and family economics; d11 consumer economics: theory keywords: corruption; gender; financial advice; perceived misconduct 1. introduction there is an ongoing focus on the need to diversify the financial services industry, particularly increasing the availability of a diverse pool of financial advisors (see blaney, 2014; center for financial planning board, 2019, 2020; schmitt, 2021). recent estimates of the relative percentage of female financial advisors in the financial services industry from barron’s (20%), the cfp board (23%), and the bureau of labor statistics (31%; nair, 2021) suggest *corresponding author. tel.: +1-336-285-381; fax: 336-256-2274. e-mail address: ddwinche@ncat.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 179–190 that the percentage of female advisors in the financial services industry is disproportionate to the percentage of females in the u.s. population. at the same time, the boston consulting group reports that 37% of u.s. wealth is controlled by women, a rate that is expected to grow 40% faster than the growth of wealth held by men (nair, 2021); mckinsey and company reports that women control one-third of total u.s. household financial assets (baghai, howard, prakash, & zucker, 2020). these numbers provide evidence that women are underrepresented in the financial advisement space, and there may be a viable pool of potential clients looking for this type of representation either now or in the near future, assuming female clients prefer female advisors. in exploring gender diversity in financial planning, a recent study finds that female financial planners are generally trusted more than males (reiter, seay, & loving, 2021). while the potential exists for females to be seen as more trustworthy financial advisors, there is a need to further explore potential boundary conditions where these perceptions translate to client engagement. the current literature lacks investigation into a potential client’s willingness to engage with a female or male advisor under marketplace conditions, specifically acknowledging that advisors are recruiting clients possessing differential perceptions of the financial services industry related to corruption. this research is conducted to bridge that gap and specifically understand the willingness of clients to engage with specific advisors, while simultaneously accounting for clients’ held industry perceptions related to corruption. on a macro level, corruption has been shown to increase the likelihood of a financial crisis occurring (ali, fhima, & nouira, 2020), which can have significant negative, systemic effects on society’s economic underpinnings. empirical research shows persistent and systematic misconduct has significant negative implications for an industry’s reputation, credibility, and business (karpoff, lee, & martin, 2008). on a micro level, highly publicized examples of financial services industry misconduct and corruption have been reported in mass media and academic research outlets. for example, wells fargo has been sued for discriminating against black borrowers that were refinancing mortgages (rosenblatt, 2022), and this is in conjunction with the company settling a suit regarding using customer information to open fake bank accounts (prentice, schroeder, & moise, 2020). further, the u.s. securities and exchange commission (sec) frequently investigates and charges financial advisors with fraud for stealing money from investors, as in the case of michael barry carter that alleges unauthorized transfers of millions of dollars from client brokerage accounts to his personal accounts along with unapproved selling of clients’ securities (u.s. securities and exchange commission, 2020). these examples by no means provide an exhaustive depiction of financial services corruption and misconduct (see also dimmock & gerken, 2012; dimmock, gerken, & graham, 2018; duffie & stein, 2015; griffin & maturana, 2016; piskorski, seru, & witkin, 2015); yet, they are representative examples of where misconduct by financial service professionals jeopardizes the financial well-being of households through two domains—psychological consequences (i.e., the loss of confidence in financial matters) and economic consequences (i.e., the decrease in net worth; brenner, meyll, stolper, & walter, 2020). amid this, however, consumers continue to engage financial advisors. in seeking this advice, it is unknown whether there is an interplay between the gender of the individual providing the advice and how corrupt a potential client views the financial services industry. we provide some clarity here by asking if female and male advisors are likely to be engaged differently by 180 d. d. winchester et al. / financial services review 30 (2022) 179–190 potential clients that view the marketplace as corrupt versus not corrupt. findings from this study support notions as to why attracting females to the financial services industry—particularly financial planning and advising—is important for industry growth (blaney, 2014; center for financial planning board, 2019, 2020; schmitt, 2021). our findings suggest when the industry is viewed as corrupt, there is not difference in likelihood to use a female or male advisor, but when perceptions of corruption are absent female advisors are preferred to their male counterparts. regarding subsequent sections, for our review regarding individuals engaging a financial advisor when they either hold perceptions of industry corruption or not, we acknowledge that perceptions lie on a continuum. for parsimony, we present the differences expected by discussing the two extremes. the literature review is framed as looking at instances inclusive of “no perception of corruption” and “perception of corruption.” as noted earlier, corruption jeopardizes consumer financial well-being, so we further frame these extremes as the presence and absence of corruption threat. 2. information search the lack of financial knowledge and expertise can impair an individual’s ability to make well-informed financial decisions. as such, consumers may choose to rely on the knowledge and expertise of financial services professional (collins, 2012; macfarlan & zick, 2020). the search for a competent, professional service provider is costly and fraught with uncertainty, and the ultimate relationship between a financial advisor and consumer contains information asymmetries and agency costs. as consumers are faced with incomplete and asymmetrically distributed information regarding the quality and commitment of financial service providers, they will seek out signals, observable signs that provide information about unobservable attributes and likely outcomes (chatterjee, kang, & mishra, 2005; spence, 1974), to aid in the decision-making process. in the absence of prior interactions with another person, the service environment and atmospherics (i.e., servicescape) may affect perceptions of the service received. servicescape has been shown to be an effective signal of hidden or undiscernible qualities of the service offered, likely outcomes, and a reducer of information asymmetry (spence, 1974). elements of the servicescape have also been found to have an impact on the likeability and perceived competence of the service provider, anticipated satisfaction, and patronage intent (dean, 2014). according to baker (1986), the service provider’s appearance and behavior are components of the social servicescape and may serve as signals of quality. prior research has shown that consumers use facial appearance and smile type as signals of a service providers’ trustworthiness and an influencer of patronage intent (dean, 2017). chang and colleagues (2015) propose that the characteristic of being male or female can be an authentic signal for unobservable qualities of a service provider. for example, research concentrating on gender in a do it yourself (diy) retail service encounter finds that customers prefer to seek advice or help from male staff as they perceive male service employees to have better knowledge of diy and handling technical issues than female staff (foster, 2004). d. d. winchester et al. / financial services review 30 (2022) 179–190 181 3. search with no corruption threat the receipt of professional financial advisement increases consumer financial wellbeing, confidence, and decision-making prowess (hanna & lindamood, 2010; marsden, zick, & mayer, 2011; winchester & huston, 2014). to help navigate the financial services market consumers must rely on the interplay of multiple quality signals or cues of service and service-provider quality (chatterjee, kang, & mishra, 2005). service providers’ ethical behavior has been found to stimulate buyer satisfaction, trust, loyalty, and buyer commitment to the institution that the provider represents (carlander, gamble, gärling, johansson, hauff, & holmen, 2018; román & cuestas, 2008). this in turn has positive effects on sales and repurchase intentions (hansen & riggle, 2009; le & supphellen, 2017; schwepker, 2013). in fact, román (2003) contends that financial services companies, and more specifically their contact employees, need to be perceived as ethical by their customers so that relationships can be developed and maintained. ultimately, research indicates that the ethicality of professional service providers has a positive effect on profitability and corporate brand equity (izzo & langford, 2003; sierra, iglesias, markovic, & singh, 2017). ethics researchers have identified gender as an individual factor impacting moral judgements and ethicality. wang and calvano (2015) find that women are generally more inclined than men to act ethically. women score more highly on “integrity tests” (ones & viswesvaran, 1998), take stronger stances on ethical behavior (glover, bumpus, sharp, & munchus, 2002; reiss & mitra, 1998), and behave more generously when faced with economic decisions (eckel & grossman, 1998). these findings suggest that potential customers may be more likely to engage with female financial advisors because those women will be less likely to take advantage of the information asymmetry associated with the financial service provider-client relationship or act opportunistically. recall, reiter, seay, and loving (2021) show that women are in fact perceived as more trustworthy in the financial advisement space. further, using signaling theory, chang, travaglione, and o’neill (2015) find that an individual’s gender can be a simple, observable signal for unobservable personal qualities. gender has been identified as a factor that affects behaviors, perceptions, and judgements. marketing and services studies show gender as a significant moderator of service elements such as service quality and value. for example, sharma, chen, and luk (2012) find that the positive association of service quality with satisfaction and value is stronger for female service providers than males. prior studies also find that women have higher ethical standards (wang & calvano, 2015), make more ethical decisions (ho, li, tam, & zhang, 2015), and have a lower likelihood of committing fraud or engaging in misconduct (camarda, 2017; camarda, chira, & de jong, 2018). women serving as agents in principal-agent relationships reduce negative agency costs and are less likely to take advantage of or exploit the information asymmetries inherent to the customer-financial advisor relationship (politis & politis, 2018). therefore, gender may serve as a simple observable signal of ethicality that affects consumer engagement intent as they perceive the financial services industry as corrupt. ultimately, given the absence of corruption threat this trait should lead consumers to report a greater likelihood to engage with a female advisor versus an equally credentialed male counterpart. 182 d. d. winchester et al. / financial services review 30 (2022) 179–190 4. search with corruption threat when the perception of corruption is present, financial services consumers should migrate away from the notion of signals pertaining to the individual providing financial services and shift focus to the services that the advisor can provide. this is because consumer perceptions of corruption in the market not only diminish the reputation and value of a financial service provider relationship, but it also increases the agency costs, more specifically search and monitoring costs associated with the relationship (tran, 2020). further, the literature has shown that when fraud is present, consumers tend to disengage from the services of financial advisors, an occurrence that is only mitigated by the advisor actively (re)building trust through the types of services provided (e.g., financial planning advice; gurun, stoffman, & yonker, 2018). in essence, the veneer of interpersonal trust is removed in the presence of corruption, and advisors need to effortfully explain to consumers why their services are beneficial to potential clients. clients would be more prone to believing the claims of benefits if the messaging was perceived as relevant to what was being sold and credible (dunham, 2011). therefore, regardless of gender, when an industry is viewed with opprobrium the underlying messaging about potentially trusted services should dictate a client’s likelihood to engage. given this background, we would hypothesize the following: hypothesis 1: in predicting a potential client’s likelihood to use a financial advisor, the client’s perception the industry as corrupt interacts with a financial advisor’s gender such that as corruption perceptions are weakened clients will report a greater likelihood of using female financial advisors versus their male counterparts. 5. methodology and data the focus of this study is to analyze the role a financial advisor’s gender has on the likelihood of engagement given individual customer perceptions of industry corruption. 5.1. participants one hundred seventy-nine respondents (82 female, 97 male) were recruited to complete an online experimental survey for a nominal fee. respondents completed the survey through prolific that is an online subject pool primarily for behavioral research (palan & schitter, 2018; peer, brandimarte, samat, & acquisti, 2017), and the only filters in place were that respondents were 18 or older and in the united states. this median age of respondents was 31 with a range of ages from 19 to 74. self-reported ethnicities were as follows: white, 124; black, 15; hispanic, 14; asian, 20; and other, 6. demographic data regarding the various income tiers is reported in the appendix table 1a. we, unfortunately, do not have data on the education level of respondents. d. d. winchester et al. / financial services review 30 (2022) 179–190 183 5.2. experimental design and variables respondents participated in a between-subjects experiment designed to measure the extent to which individuals would engage with a focal financial advisor. all respondents were first provided with a definition for a financial advisor. they were told “financial advisors are professionals who provide objective guidance and assistance in financial decision making to clients based on their financial condition, needs, and goals. advisors are not merely product salespeople such as stockbrokers or insurance agents, but are partners that help individuals reach their financial goals.” after the definition presentation, respondents were told they would be presented with a financial advisor and asked specific questions related to them. respondents were then randomly assigned by the computer survey tool to an experimental scenario showing either a male or female advisor. eighty-seven respondents were shown the male condition, while 92 respondents were exposed to the female stimulus. head and shoulder photographs of the focal advisor were accompanied with questions informing the dependent variable—a three-item, seven-point, likert scale measuring their likelihood of using a financial advisor that was adapted from shao, baker, and wagner (2004), a = .95. the photos depicted images of actual, young financial advisors that were downloaded from a financial service’s corporate website. in both images, focal advisors had the same pose, and the images were framed with identical backgrounds. the responses for the dependent variable were averaged before data analysis. after responding to the dependent variable, all respondents advanced in the survey and responded to the scaled independent variable. after stimuli exposure and response to the dependent variable, respondents completed an ad hoc scale measuring their perception of financial services industry corruption, a = .95. the corruption measure was designed to tap into a holistic view of how dishonest individuals view the entire industry. the continuous measure of perceived industry corruption was averaged and mean centered before analysis to minimize potential issues with multicollinearity in interactions (aiken & west, 1991). all measures from this study are included in the appendix. 6. results and discussion hayes’s (2013) spss process macro, model 1, was used to analyze the data. summarized in table 1, the overall model was statistically significant, f(3, 175) = 14.40, p < .01, r2 = 0.44. investigating the main effects, respondents generally reported a higher likelihood of using an advisor as they held heightened perceptions of industry corruption, b = 0.83, se = 0.21, t = 4.04, p < .01. however, there was no main effect showing a general preference for a female or male advisor, b = 0.22, mfemale = 4.65, mmale = 4.42, se = 0.18, t = 1.23, p = .22. subsuming all lower-level effects, though, was a significant perceived corruption � advisor gender interaction, b = �0.30, se = 0.13, t = �2.39, p = .018. data are plotted in 184 d. d. winchester et al. / financial services review 30 (2022) 179–190 fig. 1 for visual purposes at 1 sd above and below the mean center for perceived corruption, respectively representing high and low perceived corruption. a probe of this interaction supports proffered notions. we use the johnson-neyman (jn) technique, which probes ranges of significance when using continuous variables, to understand this interaction (hayes & matthes 2009). at a mc perceived corruption value of less than or equal to �0.51, female advisors have significantly higher likelihood of use ratings than their male counterparts, d female – male = 0.38, se = 0.19, t = 1.97, p = .05. this perceived corruption value of �0.51 is significantly less than the mc value of zero, indicating that female advisors begin to be significantly preferred under conditions of low corruption, t = �4.78, p < .01. when perceived corruption values are greater than �0.51, differences in likelihood of using female and male advisors are not significant. again, for visual purposes the shaded area in fig. 1 highlights the range of significant differences between likelihood to use female and male advisors. supporting our hypothesis, as individuals’ perception of corruption in the financial services industry weakens, they are more likely to turn to a female financial advisor for services. table 1 summarized data analysis results model summary r2 f p 0.44 14.4 < 0.01 variable b coeff se t p perceived industry corruption 0.83 0.21 4.04 < 0.01 female versus male advisor 0.22 0.18 1.23 0.22 perceived corruption � advisor gender �0.30 0.13 �2.39 0.018 note. this table shows the significance of variables predicting likelihood of using a financial advisor. the b coefficient for female versus male advisor reports the differential in the mean likelihood of using those respective advisor types. fig. 1. plot of advisor gender � perceived financial industry corruption at 61 sd around perceived corruption mean. note: shaded area denotes range of significance. d. d. winchester et al. / financial services review 30 (2022) 179–190 185 as explored earlier, this sentiment is arguably derived from females being viewed as more trustworthy and ethical than their male counterparts. as a consumer’s perception of industry corruption becomes more prevalent, though, the preference for females as service providers wanes. the interpersonal characteristics of females being more trustworthy and ethical do not appear to carry as much weight for consumers in this situation. females and males using the same marketing messaging that positions their services as partnering with clients would seemly attract clients at the same rate. recall, before completing survey questions respondents were prompted with “financial advisors are professionals who provide objective guidance and assistance in financial decision making to clients based on their financial condition, needs, and goals. advisors are not merely product salespeople such as stockbrokers or insurance agents, but are partners that help individuals reach their financial goals.” this aspirational definition which can be translated into a marketing message that results in the likelihood for using a female advisor when corruption is perceived being at par with the likelihood of using a male advisor. together, these findings support the notions of aggressively pushing for more female inclusion in a financial advising capacity as there is no apparent downside to placing females in advisory roles given the variables measured and model created here. 7. limitations and future research in this research, we focus on the intersection of a customer’s perceived corruption in the financial services industry and the gender of individuals providing financial advisement services. we identify that females are as likely to be used as males when the industry is viewed as corrupt, and they have a better likelihood of use when the industry is viewed as not corrupt. we do acknowledge that this research is really the foundation of a potential literature stream that can be fully developed around the expansion of diversity, equity, and inclusion in the financial advisement space, as called upon by entities like the center for financial planning board (2019, 2020). while we highlight that there is no apparent downside to expanding female representation in this space given the variables we measured, there are a wealth of variables that can potentially moderate and mediate the parsimonious model presented here. of these variables, consumer held ideology around diversity may be critical in supporting (e.g., feminism) or diminishing (e.g., sexism) desires to diversify the industry. we also note that diversifying the financial advising industry is not simply a function of increasing female representation. the notion of trustworthiness and performance in a space that may or may not be perceived as corrupt is also affected by many other factors that can stand alone or intersect with one another. the natural extension of this experiment would be to investigate the intersectionality of the gender, age, sexual orientation, and race of financial advisors. in our design, the focal advisor appeared to be relatively young and white. our results, therefore, cannot speak to whether the reactions of consumers would remain consistent given, for instance, a black, asian, latinx, or multiethnic advisor, or a group of advisors more advanced in age. further, this study was fielded with an experimental design. it would be interesting or prescriptive to marry these findings to actual industry data, if available or the implementation of a field study where actual consumer behavior can be observed. 186 d. d. winchester et al. / financial services review 30 (2022) 179–190 acknowledgement this study was funded through nc a&t state university start-up funds provided to nicole r. mccoy. there are no conflicts of interest in this research. appendix part 2. variables used in study and associated scale reliability measures scaled dependent variable: likelihood of using a financial advisor (seven-point likert scale; 1 = extremely unlikely to 7 = extremely likely; a = 0.93): how likely is it that you would invest through this financial advisor? how likely is it that you would let this financial advisor help you with financial planning needs? how likely is it that you would take financial advice from this financial advisor? scaled independent variables: perceived financial industry corruption (seven-point semantic differential; a = 0.93): when thinking about the financial services industry, you believe the industry is: corrupt: ethical (r) dishonest: honest (r) deceitful: truthful (r) categorical variable advisor gender 1 = male 2 = female references aiken, l. s., & west, s. g. (1991). multiple regression: testing and interpreting interactions. newbury park, ca: sage. table 1a self-reported respondent income part 1. income demographic data income range n percentage cumulative percent less than $10,000 11 6.1 6.1 $10,000–$19,999 16 8.9 15.1 $20,000–$29,999 20 11.2 26.3 $30,000–$39,999 17 9.5 35.8 $40,000–$49,999 12 6.7 42.5 $50,000–$59,999 15 8.4 50.8 $60,000–$69,999 9 5.0 55.9 $70,000–$79,999 9 5.0 60.9 $80,000–$89,999 12 6.7 67.6 $90,000–$99,999 10 5.6 73.2 $100,000–$149,999 24 13.4 86.6 $150,000 or more 13 7.3 93.9 prefer not to say 11 6.1 100.0 total 179 100.0 d. d. winchester et al. / financial services review 30 (2022) 179–190 187 ali, m. s. b., fhima, f., & nouira, r. (2020). how does corruption undermine banking stability? a threshold nonlinear framework. journal of behavioral and experimental finance, 27, 100365. https://doi.org/10.1016/ j.jbef.2020.100365 baker, j. (1986). the role of the environment in marketing services: the consumer perspective. the services challenge: integrating for competitive advantage, 1, 79–84. blaney, e. (2014). making more room for women in the financial planning profession. washington, dc: center for financial planning board. brenner, l., meyll, t., stolper, o., & walter, a. (2020). consumer fraud victimization and financial well-being. journal of economic psychology, 76, 102243. https://doi.org/10.1016/j.joep.2019.102243 camarda, j. m. (2017). relation between financial advisory designations and finra misconduct fsr201714. financial services review, 26, 271–290. camarda, j., chira, i., & de jong, p. j. (2018). who is less likely to be involved in financial advisor misconduct? the journal of wealth management, 21, 85–96. https://doi.org/10.3905/jwm.2018.21.2.085 carlander, a., gamble, a., gärling, t., johansson, l. o., hauff, j. c., & holmen, m. (2018). the role of perceived quality of personal service in influencing trust and satisfaction with banks. financial services review, 27, 83–98. center for financial planning board. (2019). why diversity matters: how workforce diversity and inclusion accelerate business success in financial services. washington, dc: cfp. center for financial planning board. (2020). diversity in action: how to sustain the financial planning profession. washington, dc: cfp. chang, j., travaglione, a., & o’neill, g. (2015). how can gender signal employee qualities in retailing? journal of retailing and consumer services, 27, 24–30. https://doi.org/10.1016/j.jretconser.2015.07.004 chatterjee, s., kang, y. s., & mishra, d. p. (2005). market signals and relative preference: the moderating effects of conflicting information, decision focus, and need for cognition. journal of business research, 58, 1362–1370. https://doi.org/10.1016/j.jbusres.2004.07.001 collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307–322. dean, d. h. (2014). visual antecedents of patronage: personal and professional items in the servicescape. services marketing quarterly, 35, 68–83. https://doi.org/10.1080/15332969.2014.856744 dean, d. h. (2017). the benefit of a trustworthy face to a financial services provider. journal of services marketing, 31, 771–783. https://doi.org/10.1108/jsm-06-2016-0221 dimmock, s. g., & gerken, w. c. (2012). predicting fraud by investment managers. journal of financial economics, 105, 153–173. https://doi.org/10.1016/j.jfineco.2012.01.002 dimmock, s. g., gerken, w. c., & graham, n. p. (2018). is fraud contagious? coworker influence on misconduct by financial advisors. the journal of finance, 73, 1417–1450. https://doi.org/10.1111/jofi.12613 duffie, d., & stein, j. c. (2015). reforming libor and other financial market benchmarks. journal of economic perspectives, 29, 191–212. https://doi.org/10.1257/jep.29.2.191 dunham, b. (2011). the role for signaling theory and receiver psychology in marketing. in g. saad (eds), evolutionary psychology in business sciences (pp. 225–256). berlin: springer. eckel, c. c., & grossman, p. j. (1998). are women less selfish than men?: evidence from dictator experiments. the economic journal, 108, 726–735. https://doi.org/10.1111/1468-0297.00311 foster, c. (2004). gendered retailing: a study of customer perceptions of front-line staff in the diy sector. international journal of retail & distribution management, 32, 442–447. https://doi.org/10.1108/09590550410549329 glover, s. h., bumpus, m. a., sharp, g. f., & munchus, g. a. (2002). gender differences in ethical decision making.women in management review, 17, 217–227. https://doi.org/10.1108/09649420210433175 griffin, j., & maturana, g. (2016). who facilitated misreporting in securitized loans? review of financial studies, 29, 384–419. https://doi.org/10.1093/rfs/hhv130 gurun, u. g., stoffman, n., & yonker, s. e. (2018). trust busting: the effect of fraud on investor behavior. the review of financial studies, 31, 1341–1376. https://doi.org/10.1093/rfs/hhx058 hanna, s. d., & lindamood, s. (2010). quantifying the economic benefits of personal financial planning. financial services review, 19, 111–127. hansen, j. d., & riggle, r. j. (2009). ethical salesperson behavior in sales relationships. journal of personal selling & sales management, 29, 151–166. https://doi.org/10.2753/pss0885-3134290204 188 d. d. winchester et al. / financial services review 30 (2022) 179–190 hayes, a. f. (2013). introduction to mediation, moderation, and conditional process analysis: a regressionbased approach. new york, ny: guilford press. hayes, a. f., & matthes, j. (2009). computation procedures for probing interactions in ols and logistic regression: spss and sas implementations. behavior research methods, 41, 924–936. https://doi.org/10.3758/ brm.41.3.924 ho, s. s., li, a. y., tam, k., & zhang, f. (2015). ceo gender, ethical leadership, and accounting conservatism. journal of business ethics, 127, 351–370. https://doi.org/10.1007/s10551-013-2044-0 izzo, g., & langford, b. (2003). realtor designations as an indicant of cognitive moral development and success in real estate sales. journal of real estate practice and education, 6, 191–202. https://doi.org/10.1080/ 10835547.2003.12091591 karpoff, j. m., lee, d. s., & martin, g. s. (2008). the cost to firms of cooking the books. journal of financial and quantitative analysis, 43, 581–611. https://doi.org/10.1017/s0022109000004221 le, n. q., & supphellen, m. (2017). determinants of repurchase intentions of real estate agent services: direct and indirect effects of perceived ethicality. journal of retailing and consumer services, 35, 84–90. https:// doi.org/10.1016/j.jretconser.2016.12.006 macfarlan, a. w., & zick, c. d. (2020). financial advisor titles: informational complexity and the role of regulatory simplification. journal of consumer affairs, 54, 801–809. https://doi.org/10.1111/joca.12290 marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32, 625–643. https://doi.org/10.1007/s10834-011-9258-z nair, v. (2021). wealth management needs more female advisers. investmentnews. available at https://www. investmentnews.com/wealth-management-needs-more-female-advisers-204509 ones, d. s., & viswesvaran, c. (1998). gender, age, and race differences on overt integrity tests: results across four large-scale job applicant datasets. journal of applied psychology, 83, 35–42. https://doi.org/10.1037/ 0021-9010.83.1.35 palan, s., & schitter, c. (2018). prolific.ac – a subject pool for online experiments. journal of behavioral and experimental finance, 17, 22–27. https://doi.org/10.1016/j.jbef.2017.12.004 peer, e., brandimarte, l., samat, s., & acquisti, a. (2017). beyond the turk: alternative platforms for crowdsourcing behavioral research. journal of experimental social psychology, 70, 153–163. https://doi.org/ 10.1016/j.jesp.2017.01.006 piskorski, t., seru, a., & witkin, j. (2015). asset quality misrepresentation by financial intermediaries: evidence from the rmbs market. the journal of finance, 70, 2635–2678. https://doi.org/10.1111/jofi.12271 politis, j. d., & politis, d. j. (2018). examination of the relationship between servant leadership and agency problems: gender matters. leadership & organization development journal, 39, 170–185. https://doi.org/ 10.1108/lodj-01-2016-0020 prentice, c., schroeder, p., & moise, i. (2020).wells fargo to pay $3 billion to u.s., admits pressuring workers in fake-accounts scandal. reuters. available at https://www.reuters.com/article/us-wells-fargo-scandal-deal/u-sauthorities-to-announce-deal-over-wells-fargos-sales-practices-scandal-sources-iduskbn20f2kn reiss, m. c., & mitra, k. (1998). the effects of individual difference factors on the acceptability of ethical and unethical workplace behaviors. journal of business ethics, 17, 1581–1593. https://doi.org/10.1023/a:1005742408725 reiter, m., seay, m., & loving, a. (2021). diversity in financial planning: race, gender, and the likelihood to trust a financial planner. financial planning review, 5, e1134. román, s. (2003). the impact of ethical sales behaviour on customer satisfaction, trust and loyalty to the company: an empirical study in the financial services industry. journal of marketing management, 19, 915–939. https://doi.org/10.1080/0267257x.2003.9728245 román, s., & cuestas, p. j. (2008). the perceptions of consumers regarding online retailers’ ethics and their relationship with consumers’ general internet expertise and word of mouth: a preliminary analysis. journal of business ethics, 83, 641–656. https://doi.org/10.1007/s10551-007-9645-4 rosenblatt, j. (2022).wells fargo sued by black borrower for refinance ‘redlining’. bloomberg. available at https:// www.bloomberg.com/news/articles/2022-03-19/wells-fargo-sued-by-black-borrower-for-refinance-redlining schmitt, s. (2021). building a diverse practice: the value of cfp certification to female advisors. boston, ma: alite group, llc. d. d. winchester et al. / financial services review 30 (2022) 179–190 189 schwepker, c. h., jr. (2013). improving sales performance through commitment to superior customer value: the role of psychological ethical climate. journal of personal selling & sales management, 33, 389–402. https://doi.org/10.2753/pss0885-3134330403 shao, c. y., baker, j. a., & wagner, j. (2004). the effects of appropriateness of service contact personnel dress on customer expectations of service quality and purchase intention: the moderating influences of involvement and gender. journal of business research, 57, 1164–1176. https://doi.org/10.1016/s0148-2963(02)00326-0 sharma, p., chen, i. s., & luk, s. t. (2012). gender and age as moderators in the service evaluation process. journal of services marketing, 26, 102–114. https://doi.org/10.1108/08876041211215266 sierra, v., iglesias, o., markovic, s., & singh, j. j. (2017). does ethical image build equity in corporate services brands? the influence of customer perceived ethicality on affect, perceived quality, and equity. journal of business ethics, 144, 661–676. https://doi.org/10.1007/s10551-015-2855-2 spence, m. (1974). competitive and optimal responses to signals: an analysis of efficiency and distribution. journal of economic theory, 7, 296–332. https://doi.org/10.1016/0022-0531(74)90098-2 tran, q. t. (2020). corruption, agency costs and dividend policy: international evidence. the quarterly review of economics and finance, 76, 325–334. https://doi.org/10.1016/j.qref.2019.09.010 u.s. securities and exchange commission. (2020). sec charges financial advisor with fraud for stealing millions from investors. available at https://www.sec.gov/news/press-release/2020-158 wang, l. c., & calvano, l. (2015). is business ethics education effective? an analysis of gender, personal ethical perspectives, and moral judgment. journal of business ethics, 126, 591–602. https://doi.org/10.1007/ s10551-013-1973-y winchester, d. d., & huston, s. j. (2014). does a relationship with a financial service professional overcome a client’s sense of not being in control of achieving their goals. financial services review, 23, 1–24. 190 d. d. winchester et al. / financial services review 30 (2022) 179–190 pii: 1057-0810(91)90007-l financial services review, 1(1):45-59 copyright @ 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. real estate income and relocation peter chinloy this paper looks at the decision relationship between two major assets of the individual, residential real estate and human capital, the ability to generate income from labor. empirical results indicate that labor income is not sufficient for defining income within the utility fiction; real estate income must also be included. the derision to relocate must be made after considering both the return and risk in the area’s residential real estate as well as the potential income from salaries. 1. introduction the individual investor has a portfolio that is dominated by physical assets, notably real estate, rather than financial assets. personal wealth is concentrated in illiquid, largely undiversifiable form, notably in single family houses. in 1986, 41.3% of household net worth was in owner-occupied housing, with a median equity of $40,597. an additions 9% of household wealth was in rental housing, and 4.4% in secondary homes. median stock equity held, including pensions, was $3,892.’ households hold nondiversified human capital and real estate, and these two assets account for nearly all private wealth.2 while households are theoretically able to participate in any real estate market, practical and institutions considerations restrict diversi~~ation, with real estate holding this dominance, it is plausible that its income should influence personal choice decisions. one of these decisions is on where to live and work. an individual having a choice between locations, such as over competing job offers, evaluates the present value of income streams. since real estate returns vary by location, a total income package facing an individual, of job and real estate, has a locational component. this paper evaluates to what extent individuals include this real estate income in their decision to relocate and to accept a job offer. the total income an individual receives is divided into three components tied to location: a fourth, for income from financial assets, has no locational peter chinloy l american university, washington, dc 20006. 46 financial services review, l(1) 1991 variation. the three sources of income are from a job, from a market return to real estate, and from a premium accorded to owner-occupancy. since job income is in cash, individuals fully include it in allocation decisions. this paper tests whether the two returns to real estate are included in total income. real estate income, both market and owner-occupier forms, is the product of the rate of return and the quantity of equity held, and is divided into separate owner premium and investor components. the quantity of equity is constrained by the local real estate price level, and by institutional requirements on debt service and down payments. the market rate of return is available to any real estate landlord, regardless of place of personal residence. the premium rate of return comes from preferential treatment on capital gains and deductions accorded owner-occupancy by the tax code. the model is sufficiently flexible to test whether households include only the locationally constrained return, or include all real estate income in their decisions. also testable is whether one dollar of possibly illiquid real estate income is viewed as equivalent to one dollar of cash income. in section 2 real estate income is constructed. from full i~ormation balance sheets and income statements on households, the rate of return and quantity of real estate that can be purchased are determined. section 3 examines which of these three income components are included in decision making. the context is an empirical examination of a sample of jobseekers. an important advantage of the data set is having the actual job choices and locations faced, and knowing the final selection. the empirical results indicate that labor income is not a sufficient statistic in the income measure within the utility function: real estate income must be included in the relocation decision. there are implications for individual and corporate relocations. the return and risk in residential real estate is sufficient to dominate other income from salaries. it is not uncommon for households to have losses in real estate markets that exceed labor income. this paper attempts to integrate the behavior of households in these two markets. 2. real estate income 2.1, locational returns the focus is on the two assets that most individual investors hold most: real estate and human capital. these assets do not have the properties that make for elegant results in financial assets. diversification in real estate holdings and earning capacity is difficult, given wealth constraints and the inability to separate human capital from a person. there are no futures markets for physical real estate or human capital. short selling to hedge risk is not possible. there are transactions costs of real estate income and relocation 41 changing either jobs or houses. for houses, there are brokerage fees, moving expenses, discount points, escrow fees, and transfer taxes. job changes entail commuting costs, loss or restriction of pension, vesting, employee stock ownership and option benefits, and the cost of relocation. results on securities in financial markets depend on no arbitrage, continuous trading, and the existence of derivative markets to permit shifting of risk. these assumptions do not hold in either real estate or human capital markets. while the inclusion of real estate returns in total income is relevant for any decision, the objective is to compare locational choices for a person having more than one job offer. the jobseeker prices a standardized house h across locations. the price of the house of quality hat a given location is v. it produces a rent to value, or income capitalization rate of k. the house then rents for kvdollars. nonhousing goods and services have a vector of prices and quantities q and x. income determination in real estate and labor markets depends on locational features that are not relevant in financial markets. the offer of a job effectively comes with rights to buy into the local real estate market as an owner-occupier. without these rights, the individual wishing to buy in that location must rely on income from financial assets, or become an absentee landlord. these alternatives are limited, given the indivisibility of real estate purchases. the individual wants to compare a cost of living across locations. the relevant comparison is on k v, rather than k. the index k v can be viewed as a hedonic price index of a house of standardized quality h. the direct utility function isp(x : h), increasing in x and h and strictly quasiconcave. the individual has a choice over locations, each offering a package of a job and a real estate market. the indirect utility functionfhas level f(q,z) = maxj*(x: h) : kv+ q xc z x (1) where denotes an inner product. the indirect utility function is strictly quasiconcave, decreasing in prices and rents, and increasing in income z. a job offer pays a salary y and comes with a requirement for residence. the local real estate market has a price level that permits the individual, using y as the principal source of debt service, to purchase a dollar units of equity. in locations with more expensive real estate, fewer units of a can be purchased and serviced with a given income y. total income is z=(l t)y+g(ea) (2) 48 financial services review, l(1) 1991 where r is the tax rate in the location and e_4 is real estate income. the function g permits real estate income from being viewed as not a one-for-one dollar substitute for cash. if real estate income does not enter the locational choice of the individual, g(ea) = 0. the price of real estate services is kv, adjusting for the cap rate k and v, the hedonic price index of a house. the balance sheet condition that assets be equal to the sum of liabilities and equity is y=b+a (3) where bis mortgage debt. the balance sheet condition is converted to an income statement by multiplying the property by its total return, debt by its interest cost, and equity by its rate of return. the returns, and income statement, depend on tax characteristics of the owner. while the balance sheets of an owner-occupier and absentee investor are identical for the same property, their income statements depend on the occupancy status of the owner. if the property is owned by a landlord, the income statement is ea-l-(1 t)qb = rv+ &v = [(i r)(k h) + (1 &)p]f’+ 6bv. the return on equity plus debt service cost is equal to revenue. mortgage debt is at interest rate [y, and deductible as a business expense. the after tax cost per-dollar of debt is (1 r)o. the effective depreciation rate 6 is a transform of 6, the inverse of the permitted useful life, measured net of expected costs of recapture on disposition.3 the landlord return on equity r is r = (1 $(k h) + (1 8r)p where &[o,l] is an adjustment if capital gains are taxed at a preferred rate. if capital gains are taxed on accrual as ordinary income, 8 = i. if capital gains taxation is deferrable indefinitely, or gains can be rolled out, 6 = 0.4 the investor return to real estate is the sum of the rate of accrued capital gains p and the rental dividend or income capitalization rate. this capitalization rate is k, net of operating expenses except for property taxes at rate x. the pretax total return on the real estate asset isp -ik a. the return on assets for a landlord is e= (1 7.)(k h av) + (i &)p + tc% i--v k-u _ l-v real estate income and relocation 49 where v g b/ v is the leverage, or loan to value ratio. the rate of return is the leveraged difference between the income capitalization rate k and the user cost of real estate services. this user cost is u = (1 ~)(h + (uv) + t(k 6b) (1 f3~)p (4) the net expense per dollar of property value. the net expense is the cost of property taxes and interest (1 t)(a + cyv), plus income tax t(k 66), less net capital gains (1 f3~)p.~ this rate of return e, producing real estate income ea = (k u) v, is comparable with those on other assets and investments. the investor is not required to reside in any location to receive this return. real estate provides additional returns that are specific to a location, through the subsidies provided for owner-occupancy. if the property is owned by an owner-occupant, the balance sheet remains v = a •lb. by comparison with a landlord, the income statement differs, depending on preferential fiscal treatment. the income statement is where 0 subscripts apply to the owner-occupier. the return on equity is e, = k+g-(1-t)(h+av) = k-u,. (5) l-v l-v the rate of return is the leveraged difference between the rent and expenses. the return on assets is the sum of the income capitalization rate and capital gains, less property taxes, or r,=k+p-(1-t)h (6) since imputed rent in the income capitalization rate k is not taxed. for owner occupiers, 8 is virtually zero. property taxes are deductible against any income, and no tax depreciation is permitted.6 accrued capital gains can be rolled over and rolled out, to a limit, provided the owner is of a certain age. the tax preferences on capital gains and the interest deduction entail a designation of a principal residence. the owner must reside in this principal residence to claim preferred status for capital gains, as george bush discovered.7 the user cost of real estate services for an owner-occupier is i.+ = (1 ~)(h + av) p. (7) the difference between the user costs for a landlord and an owner-occupier is a premium financial services review, l(1) 1991 c=u-u,=r(k+8--sb). (8) this premium is the sum of the tax free treatment of imputed income plus preferential taxation of capital gains, less the inability to claim tax depreciation expense. in locations where rents are high relative to house prices, the premium earned by owner-occupiers increases, holding expected appreciationp constant. the premium is increasing in the expected appreciation rate, and in the capital gains tax on investment real estate 8. the premium is decreasing in depreciation 6, since owner-occupiers are not eligible to claim this expense. for an owner-o~upier, real estate income is ej=(k--u+c)v. for a landlord, the corresponding income is (k u) k 2.2. real estate asset purchases the above section determines the return on equity, which differs by type of owner. another feature of real estate markets is that equity is effectively constrained by institutional limits on leverage and debt service ratios, and that expected real estate income is largely excluded in income used to qualify for financing. secondary mortgage market institutions set a limit k as a fraction of cash income y that a borrower can spend on debt service and property taxes. debt service is wb, where for contract rate a and term y, o = a/[ 1 (1 4a)-‘] is the mortgage payment per dollar of loan. property taxes are hv, so the underwriting constraint is ky = wb+hv = w(v-/ij+w where 2 is the down payment equity at purchase. the solution of this constraint is the maximum house that the individual can purchase, given a job offer paying y, or v =t: icy + wa (9) oj+h this condition applies to new entrants to-the local real estate market. for existing owners in high return markets, y> v, and they could not repurchase table 1. variables and parameters parameters descridtion marginal tax rate property tax rate mortgage contract rate number of payments mortgage payment effective rate of depreciation capital gains tax for investor debt service ratio variables z y zt v b p k : r, ro e, e, u, ug description total income labor income, pretax prices, non-real estate goods and services real estate equity hedonic price standardized house mortgage debt outstanding capital gains rate income capitalization rate loan to value ratio basis ratio rate of return on assets rate of return on equity user cost their own property. in depressed markets, potentially v> l? for a marginal buyer satisfying qualifying standards at the constraint, real estate income is ej=(k-u)v+cv. (10) the open market real estate income, that would be available to any investor, is _ (k u)v= [(i t)(k h av) + 3% + (1 &)p] ky,‘,wff * (11) the premium income to owner-occupancy is ky + ojif cv=7(k+ep-n?) cc,+h . (12) the variables and parameters of the model are summarized in table 1. 52 3. financial services review, l(1) 1991 decisions on job and housing choice the individual m has offers j = l,... j(m), drawn from the underlying distribution of jobs. the choice set over jobs need not be common to all individuals. a job package j includes working conditions of salaries, benefits and hours, the quality of the employer, and a location. at the location, the rental price of the standardized house is /$i$. the price of other goods and services is q, an index of prices q other than real estate rents. direct compensation offered by the employer is 5, and the combined tax rate in the location is r+ after tax cash income is (1 75). total income in the location is zj = (1 7-j & ig(ej). if salary income alone is sufficient to determine the locational choice, then g = 0 and real estate income is excluded. if salary income is insufficient, then zj > 5. otherwise, estate income and wealth are unanticipated windfalls not affecting allocative and mobility decisions. over the j(m) job offers, the indirect utility level isakjf,e,zj). the location selected is that which maximizes utility. this is a qualitative choice, and differentiability conditions do not hold.* if the first job is numbered as that selected 44 i= 1 if g 24(k1 vi,@, yi) u,fkjq& &i 10 0 otherwise. (13) this is a conventional job decision, with real estate income excluded. the implied definition of income is &=c g(ej) = 0. a more general form has income including real estate return, though possibly with a discount. real estate income cannot easily be collateralized, and involves transactions costs of realization. institutions on mortgages, property taxes and depreciation act to discourage sale, and to lock in existing owners. capital markets for borrowing against real estate equity are incomplete, though they are also imperfect for borrowing against human capital. the multinomial decision, with location 1 being that selected, is (14) i 0 otherwise with zj = 5 + g(ea). restrictions on the structure are indicated in table 2. real estate income and relocation 53 table 2. hypothesis testing (with required restrictions) form restrictions lnz=ln y+/3kln(k-~)v+j3~lncf fik # 0, pc f 0 hz=h y+bk[h(k-u+c)f] pk = pc lnz=ln y+ln(k-u+c)v flk = flc = 1 in z = in y flk = j% = 0 the unrestricted form introduces parameters & and fik for the two types of real estate income. if & = & then both incomes are viewed as being identical. if the parameters are both equal to one, all income is measured homogeneously. if pc = @k = 0, then real estate has no role in relocation decisions. real estate returns e and e, are constructed for locations across the united states. a comparable quality house is priced in 161 metropolitan areas across the country, from data compiled by coldwell banker offices. where coldwell banker has several offices in a metropolitan area, prices in the least expensive sublocation are used. data on cap rates k, property tax rates h, and house prices v are obtained from the survey. financing is with a fixed rate mortgage for 30 years, payable monthly, as the payment w, and interest rate a’. the loan to value ratio v is 0.8 and debt service ratio k is 0.38. the marginal tax rate for an individual is of federally, and 7s at the state and city level. the tax rates are specified for a single person with no dependents, with income only from the job, and claiming no deductions other than for real estate. the tax schedules for each state and city are obtained from central clearing house (cch), state tax handbook. the combined marginal tax rate is where k = 1 if federal taxes are deductible in calculating state tax liability, and zero otherwise. the term (1 t~)t~ accounts for the deductibilty of state 54 financial services review, l(1) 1991 taxes in calculating federal liability. the depreciation rate 6 is l/27.5 to correspond to the rate on residential property under the 1986 tax reform act, and the purchase price set at two-thirds depreciable. the holding period is five years and the rate (y is used for discounting, yielding an effective depreciation rate 6. a real estate commission of 6% is payable on sale. the returns e and e, are constructed for a job seeker choosing between locations in 1989 and planning to purchase and hold a house for 1989-1994. the data on all variables except the cap rate k and expected appreciation p are for 1989. the average cap rate over 1986-1989 is used for k. two specifications on expected appreciationp are used. the first is that the sample mean rate for 1986-1989 will obtain for the holding period 1989-1994. the second is a truncation. for locations with above average rates of appreciation, the lowest observed increase during 1986-1989 is used to project the returns for 1989-1994. for locations with below average increases, the highest observed increase is used. the smaller of the two expected appreciation rates is used for p. the sufficient statistic tests are applied to a sample of new labor market entrants with professional degrees. the survey was administered for jobs commencing in 1989. respondents were asked to report details on all job offers, including compensation. fringe benefits such as pension plans, health coverage and moving allowances were surveyed. the location of each employer was asked. the respondents coded the jobs in the order in which offers were recieved, and indicated which offer was selected. the number of respondents was 189, and only those reporting at least two job offers in different locations were included in the sample. this reduced the sample size to 136, with 395 total offers. the largest number of job offers recieved by one individual was 12. the jobs represent the set of offers that an individual received, rather than comparing an observed wage with a hypothetical alternative. problems of self selection by employees in choosing employers, and of employers in making offers, are largely eliminated. by matching the job offer, salary and employer data with the 161 metropolitan area file, a tax rate is obtained, determining y. the locational file provides data on the real estate market, including cap rate k and returns to real estate. the remaining variable is the price of other goods and services q. for cities covered by a consumer price index, this level in july 1989 is used. for cities not covered, the cpi for the united states is used. the salary offer y determines the maximum level of house purchasable v, given the qualification algorithm. this determines the two types of real estate income (k u) v for the owner as investor, and the premium cf. the rate of return k u and ownership premium c are the same for all job offers within a location. the potential house size v differs, depending on the salary offer. real estate income and relocation 55 5. specification and empirical results the indirect utility function is akv,q,(l t) y,(k u) v, cv : w], in real estate rents, prices of other goods, cash income, investor and owner premium real estate income, and other characteristics w. the natural logarithms of the first five arguments are x,, n = i ,..., 5. the remaining xn, iz = 6 ,..., n are characteristics of w. if the indirect utility function has a logarithmic form in f = ? &xn + e (15) n=cl where xo denotes an intercept, p,,, ii = o,... ,n are parameters and e is an additive error. the coefficients of cash income y, investor real estate income (k u) v and premium income cv are, respectively, /33, /l, and ps. the parameters are identified by /34 = fikp3 p5 = pcp3 (16) the indicator variable 4 = 1 if location j is selected, and zero otherwise, withbj= pr[uj= l],j= i,... j(m). with a sample of mindividuals, the logarithm of the likelihood function is in 1 = 2 c imj in bmj. m=l j=l (17) a multinomial logit specification is used for estimation. the probability that an individual with set j(m) selects the first alternative is g (ln fi > lnfl 1 = pr i ,-j (ej el) < [xl xj] ’ fi 1 where xj e (xij,..., xsj), and fl is a parameter vector of dimension n. the characteristics disappear, since they are specific to the individual, and are the same across jobs. the xvariables are the changes in rents, prices of other goods, after tax salaries, and the two forms of real estate income. 56 financial services review, l(1) 1991 the sample has an advantage over other applications of quanta1 choice models, in that the data for all alternatives are known. in typical cases, data vary across individuals but not over alternatives. characteristics such as age and sex are known, but not the costs and returns for an individual in each alternative. this lack of data in other applications increases the difficulty of identifying parameters. the odds of selecting location 1 versus location 2 are b1/ bz = exp (xi x2) * b the x differentials are zero for household characteristics, so all parameters are unidentified if prices and incomes in alternative locations are unknown.g this identification problem occurs when the conditions in not accepted offers are unknown. by comparison, in this sample, all variables differ across alternatives, the jobs and locations. 6. empirical results the mean expected real estate preimium income in the largest 25 metropolitan statistical areas (msas) is $12,528, as compared with $4,617 outside this group. this is only a part of real estate income. the market component averages $8,942 in the largest msas and $4,621 otherwise. the mean total compensation is $37,612 in the largest msas, including cash benefits, but before taxes, and $32,815 outside. the differential in expected real estate income is relatively larger than in labor income. parameter estimates of the locational choice equation are reported in table 3. the first column indicates the unrestricted estimates, with all three income components. two control variables are for whether the job is in the 25 largest msas, and a financial ranking of the employer in qual. in columns l-3 are estimates with various forms of real estate income. the two real estate incomes are unrestricted in column 1. in column 2 are the estimates when the two forms of income are homogeneous, but still potentially discounted relative to cash. in column 3 are estimates when both incomes are identical with cash. in column 4 are estimates when neither form of real estate income is included in relocation decisions. the table reports the results as fik &/p3 and pf e ps/& where p4 and ps are the coefficients of investor and owner income in the logit specification, and /33 is the labor income coefficient. the estimates of the logit have maximum likelihood properties, including invariance under single valued transforms. then the estimates of pk and pc have maximum likelihood properties. in all specifications, dollar rent differentials a in k have a negative effect on job choices and relocation. prices of other goods and services a in q have real estate income and relocation 57 table 3. estimates, locational choice (asymptotic standard errors in parentheses) (4 (4 (3) pk = pc fik = ,$ = 1 pk ‘2~ = 0 rentalnkv pi -0.185 -0.199 -0.224 -0.141 (0.122) (0.108) (0.146) (0.098) other prices a in q p2 -0.035 -0.248 -0.213 -0.308 (0.029) (0.165) (0.144) (0.174) labor income a in y p3 0.070 0.075 0.061 0.090 (0.041) (0.034) (0.029) (0.044) re investor income” /% 0.485 0.529 1 aln(k-u)v (0.452) (0.239) re owner income” pc 0.726 0.529 1 alncv (0.289) (0.239) controls ranking a qual p6 0.065 0.077 0.061 0.083 (0.041) (0.041) (0.029) (0.045) largest 25 msas pl 0.009 0.027 0.029 0.028 (0.007) (0.013) (0.014) (0.011) x2/ df 8.2 12.2 14.2 df 1 2 2 note: ’ fit and pc are obtained from the conditions fi4 = fi$k and ps = fi&, where /& and ps are the parameters of real estate investor and owner income in the logit estimation. the equation is forced through the origin, so there is no intercept. first differences are denoted by a. the same effect. differentials in labor income a in y are always positive and significant. in all variants where real estate income is included ex ante, labor income is significant in affecting the locational choice. while labor income is significant, real estate income is also important in locational choice. the coefficients fik and pc are the weights assigned to the two forms of income by individuals in computing total income. if the weights are zero, all real estate income is excluded. when the weights are unity, both incomes are included, and homogeneous with labor income. in the unrestricted case, real estate investor income, while positive, is not significant, but the owner premium is significant. the hypothesis that the two types of income are identical fails. the critical value of x2 with one degree of freedom is 8.2, but the test statistic is 8.8. column 3 tests whether all three types of income are homogeneous. this test restricts the two discount factors at unity. the test fails, with a x2/2 test statistic of 12.2, against a critical value of 10.6. in column 4, all real estate income is excluded. real estate income cannot be removed from a definition of income, with x2/2 being 14.2, against a critical value of 10.6. location decisions are not made solely on the 58 financial services review, l(1) 1991 basis of labor income. individuals use a more broadly based definition of income, including the return to real estate. 7. concluding remarks the individual investor holds a restricted portfolio of assets, principally skills to earn salaries and real estate. this restricted portfolio is nondiversified and subject to risk in the area. a decline in the local economy increases the risk of unemployment, and depresses the real estate market. most large employers pay similar compensation, regardless of location. returns to human capital are standardized, while large location-specific differentials remain in real estate returns. the individual investor, dominated by indivisible investments, must make choices under capital market and other restrictions. the results indicate that a structure can be developed to accommodate these choices. acknowledgments: i am grateful to steven kapplin, mike miles, jean louis heck, lawrence gitman, john heineke, and participants at the academy of financial services meetings, orlando, florida, for comments. joe dinunno and steve lambert provided research assistance. notes the source is the united states bureau of the census, current population survey, p-70 7. on the expenditure side, housing costs have a 34% weight in the consumer price index. deaton (1989) notes that both from the survey of consumer finances and the consumer expenditure survey, the median financial wealth of a u.s. househod in 1987 was less than $1,000. this estimate excludes pension rights. for holding length t and interest rate (y, the asset value of one dollar of depreciation is d = 1 (1 + c$-= _ t a! (1 +a)= . multiplying by cy/ [ 1 (1 + cy)-r] converts the asset to an annuitized service flow, or -i cyt a = s l-(l+.*)t-1 i the effective depreciation rate. the depreciable basis is by, where b is the ratio of the purchase price of structural improvements to the market value of the property. examples of deferred capital gains on investment real estate are rollovers, exchanges within a given time limit for another designated property, under section 1031 of the tax code, and certain installment sales provisions. the term structure, as in cox, ingersoll, and ross (1985), and default risk are other modi~cations. these are for u.s. institutions. in canada, r0 = k ip a, as property taxes are not deductible, but capital gains can be rolled out with no restriction. in the u.k., japan, and real estate income and relocation 59 west germany, there is no capital gains taxation. in switzerland, imputed rent k is taxes, but assessments lag market values. on becoming vice president in 1981, he sold his houston house for $853,000 and purchased his kennebunkport, maine, property for $950,000, claiming the latter as his principal residence. the internal revenue service successfully denied his claim, arguing that his principal residence was the vice president’s house. he was assessed over $200,000 in back capital gains taxes and penalties. the continuous demand for real estate services is -(au/ak)/(du/dy). the relevant decision is the purchase of real estate properties. see mcfadden (1984) and judge, griffiths, hill, ltitkepohl, and lee (1985, pp. 770-771). references commerce clearing house (cch). 1989. state tax handbook. chicago: commerce clearing house. coldwell banker. 1986-1990. home price comparison index. chicago: coldwell banker. cox, j.c., j.e. ingersoll, and s.a. ross. 1985. “a theory of the term structure of interest rates,” econometrica, 53: 385407. deaton, a.s. 1989. “savings and liquidity constraints.“nber working paper 3196, december. green, j.r., and j.b. shoven. 1986. “the effects of interest rates on mortgage prepayments,” journal of money, credit and banking, 36: 41-58. judge, g.g., w.e. grifliths, r.c. hill, h. ltitkepohl, and t-s. lee. 1985. 77re theory and practice of econometrics, 2d edition. new york: john wiley. mcfadden, d. 1984. “econometric analysis of qualitative response models,” in handbook of econometrics, z. griliches and m. intriligator, eds. amsterdam: north-holland. scheffe, h. 1959. 711e analysis of variance. new york: wiley. academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin winston-salem state university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review terrance k. martin winston-salem state university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. institutional membership to academy of financial services is $250 and includes the four annual issues of fsr. how to submit: there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academyfinancial.org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor terrance k. martin at martintk@wssu.edu. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2022 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 30, no. 3, 2022 editor terrance k. martin, winston-salem state university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $250. academic subscription price is $125 and is available by joining the academy of financial services. academic subscription includes four emailed online issues of financial services review each year. hardcopies of the journal are available at additional cost. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. editorial office: terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. advertising information. those interested in advertising in the journal should contact terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu printed in the usa © 2022 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, terrance k. martin, college of arts, sciences, business, and education, reynolds center, rm 111, winston-salem state university, 601 s. martin luther king jr. drive, winston-salem, north carolina 27110. email address: martintk@wssu.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. finser_31-2_complete_issue fear and trust in financial institutions: a content analysis isha chawlaa, mia b. russellb, kenneth j. whitec,*, sharon a. devaneyd adepartment of family science, university of maryland college park, 4200 valley drive, college park, md 20742, usa bcenter for leadership education, johns hopkins university, 3400 north charles street, baltimore, md 21218, usa cnorton school of human ecology, university of arizona, 650 north park avenue, tucson, az 85721, usa dcollege of health and human sciences, division of consumer sciences, purdue university, 4000 s. westport avenue, # 352, sioux falls, sd 57106-2326, usa abstract financial institutions are pillars of the economy and play an important role in consumers’ daily lives. as such, trust between financial institutions and the consumers they serve is of paramount importance. using an online survey administered during the covid-19 pandemic, this paper uses a qualitative content analysis methodology to explore consumer fear and trust in financial institutions. stemming from fear of loss, three themes emerged: (1) history and experience, (2) perceived unfair practices/lack of knowledge of banking, and (3) general trust issues. implications for financial institutions are presented based on these results. © 2023 academy of financial services. all rights reserved. jel classifications: y1; y2; y8; y9 keywords: banks; fear; financial institutions; knowledge; trust 1. introduction financial institutions are pillars of the economy and play an important role in consumers’ daily lives. yet for decades, there have been concerns about consumer confidence and trust in banks (grable et al., 2023). the confidence and trust consumers place in banks are *corresponding author. tel.: +1-520-621-1075, fax: 520-621-9445. e-mail address: kennethwhite@arizona.edu (k. j. white) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 211–228 necessary for financial access and inclusion, for individuals, as well as for the pooling of savings and expansion of credit by banks (fungáčová et al., 2022). the issue of trust in the financial system has become of high importance among regulatory authorities (van der cruijsen, 2022) as well as researchers in academia (monferrer-tirado et al., 2016; nienaber et al., 2014). yet, according to the edelman trust barometer (campbell, 2019), financial services is one of the least-trusted sectors globally. this is not surprising, or new, given that a trust crisis—the loss of public confidence in financial markets, institutions, and other related economic agents—emerged after the global financial turmoil of 2007-2008 (uslaner, 2010). in fact, knell and stix (2015) found that financial crises influence trust due to perceptions of the economic environment. trust in financial institutions has been on the rise; increasing nearly 6% (22% to 28%) during the 10-year period of 2008 to 2018. however, the covid-induced crisis may differ from the global financial crisis as the pandemic affected both the physical and financial health of consumers (marcu, 2021). trust is the essence of transactions in banking and is foundational when building long-term customerbank relationships (buriak et al., 2019; lachance & tang, 2012; roberts-lombard & petzer, 2021). trust in financial institutions has been characterized as the expectation that financial institutions are generally dependable and can be relied on to deliver on their promises (fungáčová et al., 2022). in this context, trust provides a sense of comfort for consumers; allowing them to know that their money is safeguarded by the bank and creates a feeling of commitment to the bank because processes are in place that protect against opportunistic wrongdoing or misconduct (buriak et al., 2019). the theory of reasoned action (tra) provides a lens through which to highlight the importance of trust in the relationship between consumers and financial institutions (albarq & alsughayir, 2013; alqasa et al., 2014; fishbein & ajzen, 1975; shih & fang, 2004; zolait & sulaiman, 2008). fishbein and ajzen (1975) assert that an individual’s course of action is predicted by their behavioral intentions, which are determined by two components: attitude and subjective norms. in other words, an individual’s decision to transact with banks is rooted in their positive or negative evaluations, feelings, and perceptions of financial institutions, as well as the influences and information they receive from their social environments and networks (albarq & alsughayir, 2013). a lack of trust then would negatively impact a consumer’s attitude towards financial institutions and affect their willingness to bank with them. consumers who perceive financial institutions to be untrustworthy and a possible threat to their financial well-being may be motivated to protect themselves from potential loss by limiting the use of bank products and services, or by searching for alternative banking services (rogers, 1975). a consumer’s unwillingness to bank is a conceivable outcome, acting as a mechanism to cope with their fear of loss stemming from a negative attitude toward financial institutions (rogers, 1975). while we acknowledge that other forces, such as culture (albarq & alsughayir, 2013), could influence an individual’s willingness to use banks, theoretically, seeing financial institutions as untrustworthy is also a probable determinant. mayer et al. (1995) assert that trust is built from three core components: ability, integrity, and benevolence. as one of the most used practical determinants of bank trust, ability refers to the expertise or competence that the financial institution exhibits in the domain in which they are to be trusted; for example, technical and managerial abilities to provide financial 212 i. chawla et al. / financial services review 31 (2023) 211–228 services and relevant information, to assist consumers with their decisions, and to handle problems and complaints (van esterik-plasmeijer & van raaij, 2017). integrity, or perceived integrity, is the belief that the bank adheres to a set of acceptable principles, which are operationalized as honesty demonstrated by bank employees, fairness in the application of rules, procedures, and conditions, and visibly equal and fair treatment of consumers (dimitriadis, 2011; mayer et al., 1995; muller & turner, 2016). finally, benevolence is demonstrated through the bank’s genuine interest, empathy, and responsiveness to the consumer irrespective of the profit motive (mayer et al., 1995). 2. literature review trust is described as a dynamic and multifaceted concept in retail and banking literature (luo et al., 2010). trust in banks is critical, especially in turbulent times—and is vital for financial access, inclusion, and stability (bijlsma & koldijk, 2022). low trust has the propensity to limit financial access, inclusion, and stability for consumers as well as damage the financial services industry. individuals with lower levels of trust are less likely to have a savings account and have stronger liquidity preferences than people with higher levels of trust (van der cruijsen et al., 2021). buriak et al. (2019) suggest that low or limited levels of trust create conditions for less than optimal financial behaviors including engaging in financial alternatives and more risky financial arrangements such as payday lending, bitcoin, and peer-to-peer companies among others. when individuals exhibit a reluctance to use financial services, this often indicates a diminished sense of confidence in banks, stemming from a low level of trust (fungáčová et al., 2022). low trust in the financial sector may undermine financial stability for individuals, potentially damaging the financial services industry. when consumers have low levels of trust, a negative experience may be perceived as proof that the bank cannot be trusted (kidron & kreis, 2020). if the industry is not trusted, consumers will choose to engage less, which will in turn damage both the industry and the economy by reducing the availability of capital for productive purposes. another consequence may include consumers switching to non-financial suppliers of financial services such as fintech and alternative financial services (van der cruijsen et al., 2022). moreover, when consumer interactions seem improper, it is not only perceived as unsuccessful but also leads to low trust beliefs. kidron and kreis (2020) found that people do not believe that banks’ norms and safeguards lead the banks to be sufficiently trustworthy nor do they, as a rule, automatically trust that financial institutions act honestly and ethically. conversely, with a high level of trust, consumers feel confident that their interests are well served by the bank. guiso and minetti (2004) found that households with higher levels of trust are more likely to use checks for making payments and to invest a higher share of their financial wealth in stocks and less in cash. higher levels of trust also help to buffer against negative experiences that may arise (kidron & kreis, 2020). this buffering effect is particularly crucial because consumers generally do not have a clear understanding of financial products (van der cruijsen et al., 2021), which can make them vulnerable to adverse experiences in the financial service industry. i. chawla et al. / financial services review 31 (2023) 211–228 213 banking studies argue extensively that trust in financial institutions is developed when consumers are respected, have their needs fulfilled, and promises are delivered (boonlertvanich, 2019). roberts-lombard and petzer (2021) found that customer orientation, information sharing, and service fairness are critical to the trust relationship. customer orientation is a serviceoriented approach that focuses on identifying and addressing customer needs to enhance longterm customer satisfaction (mukherjee & nath, 2003). customer orientation refers to employees’ ability to be oriented toward customer engagement and support and address their needs and expectations. information sharing, within a financial services environment, is deemed an ethical and transparent business practice that employs accurate information sharing with customers (mukherjee & nath, 2003). information sharing refers to open communication channels that positively address customers’ emotional expectations to enhance the service experience. within a financial service environment, information sharing must be secured regularly to inform and educate the customer quickly, professionally, and efficiently (balaji et al., 2016). finally, service fairness encompasses all the elements of service quality (namkung & jang, 2010). this is especially important, considering that customers judge a service as fair or unfair (dwidienawati et al., 2018; roberts-lombard & petzer, 2021). 2.1. trust and financial experience customers’ trust in a bank is based on prior experience and strongly depends on the bank’s demonstrated ability to behave in a reliable way and to observe rules and regulations (järvinen, 2014). fungáčová et al. (2022) found that experiencing a banking crisis diminishes a person’s trust in banks and that the length of the banking crisis is negatively related to trust in banks. in fact, the longer banking crises last on average, the larger the impact on eroding trust in banks. however, even a mild banking crisis and the experience of loss can weaken trust and influence the behavior of individuals (mudd et al., 2010). 2.2. trust and financial knowledge it is expected that respondents with more knowledge will trust their financial institutions more than less knowledgeable consumers (hansen, 2012, 2014). knowledgeable consumers are better able to evaluate information and are more likely to make better decisions about which service provider to choose. furthermore, knowledge facilitates the learning of new information so that knowledgeable consumers may acquire and retain more information than less knowledgeable consumers. knowledge may also allow consumers to formulate more questions so that knowledgeable consumers may be more aware of what is possible for a financial service provider, and this may facilitate consumers’ understanding of the behavior of a financial service provider. focusing on young adults in the united states, shim et al. (2013) found that self-perceived financial knowledge has a significant positive effect on trust in banks and financial institutions. the relevance of the type of financial literacy measure used is also illustrated by the findings of nuñez letamendia and poher (2020) who found a positive correlation between financial literacy and trust (trust in financial institutions, trust in banks, perceived honesty of banks, and perceived solvency of banks). 214 i. chawla et al. / financial services review 31 (2023) 211–228 2.3. trust, age, and gender according to existing literature, levels of trust are likely to vary by age and gender (fungáčová et al., 2022; kidron & kreis, 2020; van der cruijsen et al., 2021). during the pandemic, fungáčová et al. (2022) found that banking crises detrimentally affect the trust of people under the age of 50; however, fungáčová et al. (2019) found that compared to people under 35, older people are less likely to trust the financial health of their banks. one common theme between these two studies is that an individual’s age at the time of the crisis is important and significant for individuals under the age of 35. in support of fungáčová et al. (2022), grable et al. (2023) found that being older and having less financial confidence increases the likelihood that consumers would have lower levels of trust. crises of any magnitude can diminish trust in banks, but banking crises with larger impacts on the real economy influence young people’s trust while less severe banking crises mainly degrade the trust of older people (fungáčová et al., 2022). pandemic-related research by van der cruijsen and colleagues (2022) suggests that trust in banks increases with age—specifically, trust among the elderly appears to be the most affected by the pandemic. from a gender perspective, males have more trust in their financial institutions than females (van der cruijsen et al., 2021). in the van der cruijsen et al. (2021) study both males and females trusted their banks, but males were four percentage points more likely to completely trust their own bank. these findings were confirmed by fungáčová et al. (2022). as a global concern for decades, predating the global financial crisis of 2008, the concept of trust in banks is still undertheorized. additionally, there is a scarcity of qualitative studies attempting to understand and explain customer trust toward banks (kidron & kreis, 2020). in fact, much of the research is concentrated on the identification of bank trust determinants like sociodemographic, economic, political, and other indicators providing cross-country analysis (buriak et al., 2019). further, prior studies on trust in banks focus on different notions of trust, such as trust in the financial health of banks, trust associated with online banking (jiang et al., 2022), general trust in banks, or trust in their personnel (van der cruijsen et al., 2021; van esterik-plasmeijer & van raaij, 2017). despite the importance of understanding the role of trust in banks during the pandemic, research on the covid-related economic crisis is still ongoing (van der cruijsen et al., 2022). considering this, understanding the dynamics of trust in the banking system is paramount (redhead, 2011). we examine the open-ended responses that focused on examining the trust of customers in the banking system. by stratifying our sample based on gender (males, females, and other groups) and age (35 or above and 35 or below) we aim to uncover any differential trust experiences among these groups. 3. methods 3.1. data a qualtrics partner network of panel providers was used to recruit participants and administer an online survey to collect data between november 17, 2021, and december 15, 2021. the comprehensive survey was created, and pilot tested by ten researchers from i. chawla et al. / financial services review 31 (2023) 211–228 215 varying disciplines and institutions. it consisted of 67 questions that included four openended questions. the survey questions pertained to sociodemographic information, housing, economic resources, social capital, financial capability, optimism, financial stress, and other financial behavior and decision-making characteristics. the purpose of the data collection was to gather detailed, financial well-being-related information before and during the covid-19 pandemic of respondents living in the united states. this study used the most suitable question to address our research question. the question used in the study garnered a sufficient number of responses to make it a standalone study (eriksson et al., 2006). the abundance of data allowed us to concentrate solely on one question, which also contained pertinent data points for stratification into subsamples. 3.2. participants the total included in the descriptive analysis is n ¼ 3,593. while the study began with 3,598 total respondents, five respondents skipped the prerequisite, 5-point likert scale question and were removed from the descriptive analyses. additional respondents were removed from the content analysis for answering “strongly agree,” “agree” or “neither” to the likert question or skipping the open-ended question. given the intent and study purpose to understand why financial institutions were not trustworthy, in the second wave of removals we omitted 3,177 respondents who either believed that financial institutions are trustworthy, expressed no opinion about whether financial institutions are trustworthy, or did not provide a reason why they believe financial institutions are not trustworthy. the remaining 416 respondents, the total included in the thematic analysis, all expressed that financial institutions are not trustworthy and provided a response to the open-ended question. of the 416 respondents included in the thematic analysis, there were 56.3% males, 40.1% females, 3.6% gender diverse, 47.6% below age 35, and 52.4% above age 35. 3.3. variables the analysis was based primarily on two survey items. the first item was a prerequisite to the second and asked respondents to indicate, using a 5-point likert scale, the extent to which they agree or disagree with the following statement: “for the most part, financial institutions are trustworthy.” possible responses ranged from 1 ¼ strongly disagree to 5 ¼ strongly agree. participants who expressed in the first item that financial institutions were not trustworthy would then respond to the second item which asked for a written response to the open-ended question, “if you do not agree that the financial institutions are trustworthy for the most part, could you please provide reasons for not trusting?” survey questions regarding age, gender, and financial account ownership were used as descriptive variables to compare the sample. 3.4. analysis this study uses descriptive and content analysis to gain an understanding of reasons why consumers do not find financial institutions trustworthy. this method has gained popularity 216 i. chawla et al. / financial services review 31 (2023) 211–228 rapidly, partly because of the increased use of open-ended survey questions. it provides an overview of the sentiments people hold, while also being more objective as it relies on categorizing concrete terms and content (neuendorf, 2017). the study first uses descriptive statistics to identify the proportion of respondents, by age and gender, who perceive financial institutions as untrustworthy. next, respondents’ reasons for not trusting financial institutions are analyzed. the content analysis follows protocols developed by braun and clarke (2006): (1) getting familiar with the data, (2) generating initial codes, (3) searching for themes, (4) reviewing themes, (5) defining and naming themes, and (6) producing the report (p. 87). first, to derive our sample and begin the familiarization process, we omit respondents who skipped the question or provided responses not related to the question. the sample is then grouped into six categories by intersections of age and gender. the six categories are: 35 and under men, 35 and under women, 35 and under gender diverse, 35 and over men, 35 and over women, and 35 and over gender diverse. to help ensure integrity, the researchers independently identify the data to create an analytic triangulation process (pieters & dornig, 2013). our team comprised four primary researchers. each of whom was involved in analyzing the data into designated codes. this process resulted in line-by-line coding twice, the initial coding and verification, to encourage saturation of potential codes and ensure triangulation. the investigator triangulation technique was utilized that required researchers to each conduct separate analyses of the data before their interpretations were compared and reconciled (denzin, 1978). the authors met on a regular, periodic basis to consider, discuss, and integrate various interpretations of the data into an emerging coding scheme (patton, 2002). the initial analysis generated 14 codes. these codes were then combined and consolidated into the three themes and 15 subthemes (e.g., misrepresentation, greed/profit-seeking). once initial codes were developed, the research team organized these codes into larger themes that could be reasonably expected to provide insight into the research question. in each meeting, the authors combined themes with related content to strengthen each theme and ensure saturation of the data. similarly, during team meetings, themes not found in multiple responses were discussed to determine if they should be excluded from the study. the final step was to explore the interconnectedness of the themes and select quotes that exemplified the respondents’ perspectives of why they do not trust financial institutions. notable quotes were extracted from the open-ended responses that demonstrated each code and used in the discussion section that follows. validity and reliability were emphasized and prioritized throughout the research process. the authors meticulously ensured that all procedures adhered to widely accepted research methods (kirk & miller, 1986). in particular, the study strictly followed the six steps of thematic analysis as laid out by braun and clarke (2006), enhancing confidence in the study’s outcomes. during the identification and discussion of themes and codes, the authors maintained transparency among themselves, ensuring the validity of the findings (tuval-mashiach, 2017). the research team prioritized maintaining a consistent methodology. similar guidelines were adhered to throughout the data analysis phase. any deviations or discrepancies were thoroughly discussed, which served to enhance the credibility of our findings. to further ascertain the validity of the outcomes, our research team, comprising four members, cross-checked the i. chawla et al. / financial services review 31 (2023) 211–228 217 data and interpretations at multiple stages throughout the study (morse et al., 2002). moreover, the authors engaged in periodic discussions about any preconceived notions about the banking system they might hold; thereby, ensuring the study’s reflexivity (haynes, 2012). 4. findings overall, findings indicate that respondents’ mistrust of banking can be attributed to a fear of loss. percentages of respondents who agree that institutions are trustworthy are presented in tables 1 and 2. of the 3,593 responses, 15.0% of females, 18.2% of males, and 29.9% of gender-diverse persons believed financial institutions are not trustworthy. larger percentages of younger respondents find financial institutions to be untrustworthy. the age category with the largest percentage (19.3%) that believes financial institutions are not trustworthy is 35-44 years old, and the smallest percentage (11.5%) is 65 and over. regarding bank accounts, participants were asked to indicate all the types of accounts they held. the findings showed that the checking account was the most prevalent financial product, with 23% of respondents holding only a checking account. when combined with a savings account, this figure increased to 35%, making it a very popular combination. approximately 5% of respondents held only savings accounts. around 22% of respondents had some form of retirement account in conjunction with other accounts. approximately 13% of respondents held brokerage accounts, either solely or in combination with other accounts. about 8% of respondents reported having certificate of deposits, either solely or along with other accounts. “other financial products,” such as annuities, constituted approximately 2% of the respondents. three main themes emerged: respondents fear loss due to (1) history and experience, (2) perceived unfair practices, and (3) general trust issues. a summary of the themes and subthemes across gender and age is presented in table 3. table 1 crosstabulation by gender (n 5 3,593) financial institutions are trustworthy female male gender diverse strongly agree 6.4% 8.2% 5.2% agree 39.5% 42.1% 24.7% neither 39.0% 31.6% 40.3% disagree 10.8% 11.7% 15.6% strongly disagree 4.2% 6.5% 14.3% table 2 crosstabulation by age (n 5 3,593) financial institutions are trustworthy 18–24 years 25–34 years 35–44 years 45–54 years 55–64 years 65+ years strongly agree 6.1% 7.3% 8.2% 6.6% 7.0% 7.8% agree 31.5% 39.1% 40.3% 45.1% 43.0% 58.1% neither 44.9% 36.5% 32.2% 32.2% 37.1% 22.6% disagree 11.2% 12.0% 13.9% 9.1% 8.6% 6.7% strongly disagree 6.3% 5.0% 5.4% 7.0% 4.3% 4.8% 218 i. chawla et al. / financial services review 31 (2023) 211–228 table 3 themes and subthemes across gender and age themes subthemes gender and age historic experience: systemic and personal historical events unfair practices specific experiences lack of knowledge mistrust in government/corporate lies/misrepresentation 35 or below males historical events bad experiences public opinion/“i heard” specific experiences 35 or below females public opinion/“i heard” specific experiences 35 or below gender diverse systemic issues bad experiences historical events lack of knowledge privacy concerns 35 or above males specific experiences unfair practices lack of trust fiduciary concerns mistrust/lack of trust 35 or above females greed/profit seeking 35 or above gender diverse perceived unfair practices greed/profit-seeking lack of knowledge fiduciary concerns mistrust in government lies/misrepresentation 35 or below males greed/profit-seeking systemic issues unfair practices lack of transparency privacy concerns 35 or below females unfair practices mistrust in government 35 or below gender diverse greed/profit-seeking lack of knowledge mistrust in government fiduciary concerns unfair practices 35 or above males unfair practices lack of knowledge mistrust in government fiduciary concerns/non-fiduciary 35 or above females greed/profit-seeking personal responsibility 35 or above gender diverse general trust general trust issues 35 or below males mistrust in government/corporate general trust issues 35 or below females 35 or above males (continued on next page) i. chawla et al. / financial services review 31 (2023) 211–228 219 with one exception, the largest percentages of all genders and ages fall into the category of fear of loss because of perceived unfair practices by financial institutions. more females above age 35 responded to the historic experience: systemic and personal theme. percentages of respondents (by age and gender) who fall into each of these three themes are presented in table 4. 4.1. perceived unfair practices that is, actions or policies that are viewed by consumers as unjust, discriminatory, or unethical reflect concerns relating to higher fees and lack of transparency existing in financial institutions. respondents expressed concerns about the fees charged to them and the lack of transparency of interest rates. to provide clarity on the perceived unfair practices theme, it is helpful to provide some subthemes in this category. among 35 or below males, the most prevalent subthemes are greed, profit-seeking, and fiduciary concerns. the “greed” subtheme reflected respondents’ concerns over banking institutions’ practices that are geared towards devising tactics that take their benefits. the “profit-seeking” subtheme represented sentiments about the banking system stating profit-making as their main or only priority. “fiduciary concerns” reflected consumer opinion on the fiduciary nature of the banking system. the subthemes reflected their concerns regarding the protection of their own interests. among the 35 or below females, greed and profit-seeking subthemes are also dominant. among the 35 or above males, greed and profit-seeking constitute the majority of the themes categorized into perceived unfair practices. overall, respondents demonstrated evidence of perceived unfair practices of the banking system and financial institutions. one 35 or younger male respondent: “most financial institutions are not made to help people. their main goal is to make money and make their partners money. if it were truly about assistance, then every customer would be treated equally and it wouldn’t cost an arm and a leg to pay these institutions for their services. and it would also be much easier for people to build credit. the credit system is just about the most oppressive system that was ever conceived. (but that’s a whole other discussion.)” a female respondent, 35 years or younger: “i think that banks are often only out for making profit for themselves which is why they are always trying to sell products to people when they’re coming in to just make a basic withdrawal.” table 3 (continued) themes subthemes gender and age trust issues in general historical events mistrust mistrust in government/corporate 35 or below gender diverse trust issues in general 35 or above males — 35 or above females general trust issues 35 or above gender diverse 220 i. chawla et al. / financial services review 31 (2023) 211–228 another female describes a common perception of unfair fees and charges: “they try to take your money with fees that aren’t reasonable.” two male respondents, under the age of 35 years, capture the ill-intent and lack of fiduciary responsibility of banks: “they are seeking to make money off me and will prioritize that rather than me.” “financial institutions encourage debt on people and play on their psychological well beings or addiction to spend money to debt trap them.” in addition to perceiving that banks lack a fiduciary responsibility to customers, others describe an intentional effort to confuse customers. one example comes from a respondent described as other gender, and under 35 years of age: “use of wordings that confuse the public, often times leading to some support of fee collection because people don’t understand how it works.” respondents aged 35 and above express cynicism and skepticism about financial institutions; however, they discuss fees more frequently: “i just worry about having my money in banks in the event or a financial crisis and the fees associated with most accounts are ridiculous.” (female) “fees on everything! i feel ‘nickel and dimed’ to death. financial institutions are only there to make money off you. they refuse to help you when you need it the most, even though your credit score is high and you haven’t missed a payment of any kind in decades.” (male) “well, alot of them have ways to get more money out of you even when you are already broke and struggling. overdraft fees, the percentages of interest on loans, i mean the reason you get a loan is because you don’t have enough money to pay for whatever the loan is for. and instead of helping . . . later on down the line they have taken way more.” (female) “they are ultimately more interested in making money above all else. they seem only work for those already doing well while burying the already struggling in high fees and high-interest rates.” (female) other respondents raise issues with transparency—with regard to fees and communication: “i do not believe financial institutions are transparent and the few i have had business dealings with are concerned more with the all might dollar and those who can provide it in the form of payments, deposits etc. . . . with their facility.” (female) “wording of certain terms are misleading.” (female) table 4 themes and distribution by age and gender (n 5 416) age 35 or below 35 or above gender males females gender diverse males females gender diverse historical experience: systemic and personal 4.6% 4.6% 0.5% 10.6% 9.4% 0.0% conceptual understanding 16.3% 14.4% 1.4% 19.0% 9.1% 0.7% general trust 2.6% 2.6% 0.5% 3.1% 0.0% 0.5% i. chawla et al. / financial services review 31 (2023) 211–228 221 4.2. history/experience many respondents describe historic experiences, both systemic and personal, that influence their trust in financial institutions. history/experience had prominent subthemes across the gender and age groups. among 35 or below males, the most prevalent subthemes were historical events, specific experiences, and lies/misrepresentations. the “historical events” subtheme reflected respondents’ experiences during historical events like the financial crisis of 2008, as well as the history of their own experiences. the “specific experience” subtheme represented sentiments related to their own specific experiences with banking. for example, an instance of having issues with online banking. “lies/misrepresentations” reflected consumer opinion regarding their integrity and responsibility towards their consumer base. these fall into the history/experience theme and are categorized as individuals possibly having a limited understanding of the business aspect of financial institutions. among the 35 or below females, historical events and bad experiences, that is, their instances of bad experiences with the banking system in general, were mostly dominant. among the 35 or above males, “systemic issues,” and “bad experiences” subthemes constituted the majority of the themes categorized into history/experiences. the systemic issues reflected inherent structural issues in the united states that are linked to the banking system. overall, respondents expressed their concerns with the banking system and financial institutions that were historical in nature. for example, male respondents—in both age categories, 35 years and below and above 35 years—explain that banks have historically taken advantage of consumers: “they have a long history of scamming people.” “every bank i’ve used minus online banking have ripped me off.” across genders there is also a reference to financial institutions engaging in unauthorized activity: “the[re are] scandals . . . where tellers were making fake accounts to meet goals.” “my financial institution opened a fake account in my name.” in addition to personal experiences, respondents also share examples of circumstances that may not have directly happened to them but are close enough that the situation influences their perception of banks. we code these data as “i heard”: “i’ve heard about lots of bad experiences people have had with various financial institutions, such as messing up paperwork, overreaching boundaries, etc.” “i remember the bailout of wall street.” “a history of financial scandals and bailouts doesn’t exactly show that most financial institutions are well managed.” “too many real-life stories of fraud and theft of/by financial institutions upper management.” respondents also describe a lack of care and empathy towards consumers: “there is evidence, and i have personal experience that they prey on the lowest income americans to make their money. they are careless with personal information and lack empathy for the needs of common people.” 222 i. chawla et al. / financial services review 31 (2023) 211–228 “[i don’t trust banks] because of their history and what they do with their consumers and their data.” 4.3. general trust issues despite financial experience or type of bank account ownership, some respondents expressed deep cynicism and skepticism. primary subthemes under general trust issues were trust in general life circumstances and mistrust in government. overall, respondents expressed emotions of trust related to their life in general as well as the system of banking and government. for example, respondents aged 35 years or below express more general mistrust sentiments: “i don’t trust any kind of institutions that are based off of money.” my grandma use to have a saying “green is the color of greed, that’s why money is men’s weakest weapon.” (male) “you never know who to trust if we’re being honest. specifically, when it comes to money . . . our money.” (female) “it’s just hard to trust anything or people with so much money.”(female) “i’m very skeptical about everything.” (other gender) males over the age of 35 also express a lack of general trust in these ways: “when money is involved, there is no one you can trust.” “i don’t trust anybody or any bank i can invest and hustly own money.” this study aimed to present results stratified by gender and age. however, the results did not reveal any discernible patterns based on gender or age. future research could explore potential differences in trust by gender and age if feasible. 5. discussion the primary purpose of this study is to explore the reasons consumers perceived financial institutions as not trustworthy during the global pandemic when many households were facing financial crises and uncertainty. specifically, we analyzed written responses to the open-ended survey question, “if you do not agree that the financial institutions are trustworthy for the most part, could you please provide reasons for not trusting?” from this analysis, we developed three themes: (1) history and experience, (2) perceived unfair practices, and (3) general trust issues. the first theme stems from historic systemic and personal experiences. many respondents also shared experiences that they heard about or that affected their communities. the second theme is related to individuals’ understanding of the financial services business model. many respondents feel vulnerable against large, profit-seeking financial institutions whose greed drives them to charge excessive fees and interest rather than providing customer service and help to consumers. the third theme is indirectly related to financial institutions. the third theme is a general trust issue held by individuals. many state they either do not trust anyone or only trust themselves, especially in matters involving their money. i. chawla et al. / financial services review 31 (2023) 211–228 223 findings from this study add to the existing literature by providing a qualitative look into consumers’ trust in financial institutions. although not generalizable, the findings from this diverse sample offer further evidence that more needs to be done to strengthen relationships between financial institutions and the consumers that use them. the implications of these results are important and should serve as notice to financial institutions that continuous attention to customer service remains necessary. left unchecked, negative information can spread and go against an organization’s interests, undermine public trust, and lead to increased public scrutiny (greve et al., 2016). the results of this study suggest that relationship management and mistrust could be a growing problem for financial institutions since higher percentages of individuals under age 35 believe these institutions are not trustworthy. larger proportions of younger respondents express concerns about losing their money to financial institutions and perceive them as a threat to their financial well-being. financial institutions may be missing an opportunity to provide additional services to this segment of the population and in fact, could be at risk of losing a portion of this group’s business altogether if viable alternatives arise that are viewed as more trustworthy. in conclusion, trust is important in the banking industry, especially during times of crisis. as banks around the globe strategize post-pandemic recovery and consumer customer engagement, exploring the dynamics of trust becomes crucial. this study is of importance to professionals and policymakers working within financial institutions, as well as financial educators and professionals who serve consumers. the findings from this study are likely to offer a framework to address concerns with respect to enhancing the quality of services and dealings with their consumer base. 6. limitations there are noteworthy limitations to this study. first, this study employs secondary analysis. as such, participants were not interviewed, and no follow-up was available. second, the analysis is based only on responses to one open-ended survey question and a one-item measure of trust. third, the researchers are unable to make any distinction between consumers who may have banking accounts but still underutilize their accounts. data limitations did not allow comparisons of those who favor banks to those who view banks negatively. subsequent analyses could prioritize understanding the underlying rationale for these contrasting perspectives. additionally, a more comprehensive analysis of trust across genders would be valuable in future studies. 7. conclusion the present study explored the reasons consumers perceived financial institutions as not trustworthy during the global pandemic when many households were facing financial crises and uncertainty. using content analysis, three themes emerged: (1) history and experience, (2) perceived unfair practices, and (3) general trust issues. 224 i. chawla et al. / financial services review 31 (2023) 211–228 there are several implications for individual consumers, financial institutions, policymakers, and researchers. specifically for consumers, a lack of trust may limit access to financial services. this lack of trust might lead consumers to avoid or underutilize financial services, which can limit financial inclusion. financial inclusion promotes consumers to have and use savings accounts, loans, and other services that are important to reaching financial goals. another consequence of financial exclusion is reduced access and availability to credit when needed. when consumers have limited exposure and experience in mainstream financial services, it may be more challenging to establish credit, ultimately making it harder to secure credit and loans at favorable interest rates. additionally, when consumers limit or avoid traditional banking and financial services, they may lose the safety and convenience of their funds as well as engage in alternative financial services, pay higher costs for transactions, and have a higher susceptibility to fraud and scams. from the financial institutions’ perspective, a lack of consumer trust can lead to lower deposits and transactions, potentially disrupting financial systems and economic activity. when consumers demonstrate a lack of trust in a specific institution, banks may experience issues of reputational risk. loss of reputation, for example, may create challenges when banks want to attract new customers or even merge with other institutions. in addition to regulatory scrutiny when institutions want to merge, there could also be increased oversight and potential penalties if claims of mistreatment, misrepresentation, and fraud are found to be true. financial institutions should consider how trust may influence a consumer’s willingness to expand current relationships. consumers entrust banks with their money, which often represents dreams, goals, and other aspirations, not to mention personal information. consumers need to know that their funds are safe and that transactions are secure. when trust is not achieved, consumers lack the safety and confidence to fully engage with the financial institution, potentially hindering their ability to experience financial inclusion. yet, to prevent these implications banks need to prioritize transparency, security, ethical behavior, and strong customer service to maintain and strengthen consumer trust. furthermore, banks should ensure that products are developed with the customer and not for the customer and that bank fees are transparent (roberts-lombard & petzer, 2021). the results of this study also point to a need for financial institutions and policymakers to work together to correct systemic issues (such as the historical events referenced in this study) that have eroded consumer confidence. acknowledging and addressing historical events and past inequitable personal experiences could help improve public opinion and increase public belief that financial institutions care about consumers’ financial well-being (lagarde, 2014). there is also an opportunity for researchers to conduct deeper qualitative research, including focus groups and interviews with consumers to better understand how financial institutions may be able to develop better relationships and earn consumers’ trust. identifying more nuanced reasons for the lack of consumer trust could also be translated into educational programs and other interventions to equip consumers with tools and resources to evaluate institutions and make informed financial decisions with regard to banking relationships. in conclusion, the present study supports the need to further explore the role of trust and fear between consumers and financial institutions. i. chawla et al. / financial services review 31 (2023) 211–228 225 acknowledgments this research was made possible from funding provided by the wells fargo foundation. the assistance offered by wells fargo was instrumental in the successful completion of this project. we sincerely appreciate their commitment to advancing research on financial well-being. references albarq, a., & alsughayir, a. (2013). examining theory of reasoned action in internet banking using sem among saudi consumers. international journal of marketing practices, 1, 16–30. alqasa, k. m., mohd isa, f., othman, s. n., & zolait, a. h. s. (2014). the impact of students’ attitude and subjective norm on the behavioural intention to use services of banking system. international journal of business information systems, 15(4), 105–122. balaji, m. s., khong, k. w., & chong, a. y. l. (2016). determinants of negative word-of-mouth communication using social networking sites. information & management, 53(4), 528–540. bijlsma, m., van der cruijsen, c., & koldijk, j. (2022). determinants of trust in banks’ payment services during covid: an exploration using daily data. de economist, 170, 231–256. boonlertvanich, k. (2019). service quality, satisfaction, trust, and loyalty: the moderating role of main-bank and wealth status. international journal of bank marketing, 37, 278–302. braun, v., & clarke, v. (2006). using thematic analysis in psychology. qualitative research in psychology, 3, 77–101. buriak, a., voznõáková, i., sułkowska, j., & kryvych, y. (2019). social trust and institutional (bank) trust: empirical evidence of interaction. economics & sociology, 12, 116–129. campbell, d. (2019). edelman trust barometer: 2019 trust in financial services. available at https://www. edelman.com/sites/g/files/aatuss191/files/2019-04/2019_edelman_trust_barometer_financial_services_ report_1.pdf. denzin, n. k. (1978). the research act: a theoretical introduction to sociological methods. new york: mcgraw-hill. dimitriadis, s. (2011). customers’ relationship expectations and costs as segmentation variables: preliminary evidence from banking. journal of services marketing, 25, 294–308. dwidienawati, d., arief, m., & abdinagoro, s. b, bina nusantara university (binus), indonesia. (2018). is service fairness influencing customers’ satisfaction and intention to pay insurance premium? a case in bpjs kesehatan indonesia. journal of business & retail management research, 13, 38–48. eriksson, c., westman, g., & hamberg, k. (2006). content of childbirth-related fear in swedish women and men —analysis of an open-ended question. journal of midwifery & women’s health, 51, 112–118. fishbein, m., & ajzen, i. (1975). belief, attitude, intention and behavior: an introduction to theory and research. reading, ma: addison-wesley. fungáčová, z., hasan, i., & weill, l. (2019). trust in banks. journal of economic behavior & organization, 157, 452–476. fungáčová, z., kerola, e., & weill, l. (2022). does experience of banking crises affect trust in banks? journal of financial services research, 62, 61–90. grable, j., kwak, e. j., & archuleta, k. (2023). distrust of banks among the unbanked and banked. international journal of bank marketing, 41, 1498–1520. greve, h. r., kim, j. y., & teh, d. (2016). ripples of fear: the diffusion of a bank panic. american sociological review, 81, 396–420. guiso, l., & minetti, r. (2004). multiple creditors and information rights: theory and evidence from us firms. available at ssrn 527222. hansen, t. (2012). understanding trust in financial services: the influence of financial healthiness, knowledge, and satisfaction. journal of service research, 15(3), 280–295. 226 i. chawla et al. / financial services review 31 (2023) 211–228 hansen, t. (2014). the role of trust in financial customer–seller relationships before and after the financial crisis. journal of consumer behaviour, 13(6), 442–452. haynes, k. (2012). reflexivity in qualitative research. qualitative organizational research: core methods and current challenges, 26, 72–89. järvinen, r. a. (2014). consumer trust in banking relationships in europe. international journal of bank marketing, 32, 551–566. jiang, m., rifon, n. j., cotten, s. r., alhabash, s., tsai, h. y. s., shillair, r., & larose, r. (2022). bringing older consumers onboard to online banking: a generational cohort comparison. educational gerontology, 48, 114–131. kidron, a., & kreis, y. (2020). listening to bank customers: the meaning of trust. international journal of quality and service sciences, 12, 355–370. kirk, j., & miller, m. l. (1986). reliability and validity in qualitative research. (vol. 1). newbury park, ca: sage publications, inc. knell, m., & stix, h. (2015). trust in banks during normal and crisis times—evidence from survey data. economica, 82, 995–1020. lachance, m. e., & tang, n. (2012). financial advice and trust. financial services review, 21, 209–226. lagarde, c. (2014). empowerment through financial inclusion, address to the international forum for financial inclusion. international monetary fund. available at: https://www.imf.org/en/news/articles/2015/09/28/04/ 53/sp062614a luo, x., li, h., zhang, j., & shim, j. p. (2010). examining multi-dimensional trust and multi-faceted risk in initial acceptance of emerging technologies: an empirical study of mobile banking services. decision support systems, 49, 222–234. marcu, m. r. (2021). the impact of the covid-19 pandemic on the banking sector. management dynamics in the knowledge economy, 9, 205–223. mayer, r. c., davis, j. h., & schoorman, f. d. (1995). an integrative model of organizational trust. academy of management review, 20, 709–734. monferrer-tirado, d., estrada-guillén, m., fandos-roig, j. c., moliner-tena, m. á., & garcı́a, j. s. (2016). service quality in bank during an economic crisis. international journal of bank marketing, 34, 235–259. morse, j. m., barrett, m., mayan, m., olson, k., & spiers, j. (2002). verification strategies for establishing reliability and validity in qualitative research. international journal of qualitative methods, 1, 13–22. mudd, s., pashev, k. & valev, n. t. (2010). the effect of loss experiences in a banking crisis on future expectations and behavior. the be journal of macroeconomics, 10(1). mukherjee, a., & nath, p. (2003). a model of trust in online relationship banking. international journal of bank marketing, 21, 5–15. muller, l. a., & turner, j. a. (2016). strategic complexity in investment management fee disclosures. financial services review, 25, 215–234. namkung, y., & jang, s. (2010). service failures in restaurants: which stage of service failure is the most critical? cornell hospitality quarterly, 51, 323–343. neuendorf, k. a. (2017). the content analysis guidebook. thousand oaks, ca: sage. nienaber, a. m., hofeditz, m., & searle, r. h. (2014). do we bank on regulation or reputation? a meta-analysis and meta-regression of organizational trust in the financial services sector. international journal of bank marketing, 32, 367–407. nuñez letamendia, l., & poher, b. (2020). the effect of financial literacy on trust: do financially literate individuals have more trust in the financial system? working paper, april. available at: https://fundacionmutualidadabogacia. org/wp-content/uploads/2020/05/wp_financial-literacy-trust-april2020.pdf patton, m. (2002). qualitative evaluation and research methods. newbury park, ca: sage. pieters, h. c., & dornig, k. (2013). collaboration in grounded theory analysis: reflections and practical suggestions. qualitative social work, 12, 200–214. redhead, k. (2011). behavioral perspectives on client mistrust of financial services. journal of financial service professionals, 65, 50–61. roberts-lombard, m., & petzer, d. j. (2021). relationship marketing: an s–o–r perspective emphasising the importance of trust in retail banking. international journal of bank marketing, 39, 725–750. i. chawla et al. / financial services review 31 (2023) 211–228 227 rogers, r. w. (1975). a protection motivation theory of fear appeals and attitude change1. the journal of psychology, 91, 93–114. shih, y. y., & fang, k. (2004). the use of a decomposed theory of planned behavior to study internet banking in taiwan. internet research, 14, 213–223. shim, s., serido, j., & tang, c. (2013). after the global financial crash: individual factors differentiating young adult consumers’ trust in banks and financial institutions. journal of retailing and consumer services, 20, 26–33. tuval-mashiach, r. (2017). raising the curtain: the importance of transparency in qualitative research. qualitative psychology, 4, 126–138. uslaner, e. m. (2010). trust and the economic crisis of 2008. corporate reputation review, 13, 110–123. van der cruijsen, c., de haan, j., & jonker, n. (2022). has the covid-19 pandemic affected public trust? evidence for the us and the netherlands. journal of economic behavior & organization, 200, 1010–1024. van der cruijsen, c., de haan, j., & roerink, r. (2021). financial knowledge and trust in financial institutions. journal of consumer affairs, 55, 680–714. van esterik-plasmeijer, p. w., & van raaij, w. f. (2017). banking system trust, bank trust, and bank loyalty. international journal of bank marketing, 35, 97–111. zolait, a. h. s., & sulaiman, a. (2008). incorporating the innovation attributes introduced by rogers’ theory into theory of reasoned action: an examination of internet banking adoption in yemen. computer and information science, 1, 36–51. 228 i. chawla et al. / financial services review 31 (2023) 211–228 financial services review, 32(3) 20 the effect of financial knowledge on workers’ expectation of never retiring zezhong e. zhang,1 sherman d. hanna,2 and lei xu3 abstract in this study, we extend previous research on factors related to workers stating that they would never retire by analyzing the impact of financial knowledge variables on the expectation. findings show that the never retire rate is related to objective financial knowledge, with a 20% rate for those who missed all questions, compared to 12% for those who answered all questions correctly. we find a similar pattern between subjective knowledge and the never retire rate. using logistic regressions, we find that survey respondents who missed questions for objective financial knowledge are more likely to choose a never retire response than those who do not miss any questions. we also find that overconfident respondents are more likely to give a never retire response than underconfident respondents. our results have implications for financial education and policies related to retirement, as well as for research on retirement adequacy. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation zhang, z., hanna, s. d., & xu, l. (2024). the effect of financial knowledge on workers’ expectation of never retiring. financial services review 32(3), 20-31. introduction a key issue for retirement planning is when to retire. this decision may be a challenge for many workers, however. hanna et al. (2017) reported that in the 2013 survey of consumer finances (scf) dataset, 18% of full-time workers aged 35 to 60 gave a “never retire” response when asked at what age they expect to stop working full-time. while a never retire answer might be a reasonable response for some workers, there is evidence that most workers who state that they will never retire failed to engage in retirement planning. it is worth noting that hanna et al. did not include financial knowledge as an independent variable. our 1 corresponding author (zhang.10558@osu.edu). the ohio state university, columbus, ohio, usa 2 the ohio state university, columbus, ohio, usa 3 the ohio state university, columbus, ohio, usa research analyzes a combination of the 2016 and 2019 scf datasets, with financial knowledge variables in addition to the independent variables included by hanna et al. the notion of reporting an intention to never retire has important educational and policy implications. workers who plan to never retire or retire very late may face lower risks from having inadequate retirement income compared to others. workers who choose to never retire do not need to worry much about saving money in their retirement accounts because their income levels will be less likely to drop at the normal retirement age. therefore, under certain circumstances, https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ zhang et al. 21 choosing never to retire might be a rational decision. however, from 1910 to 2001, the average retirement age of workers substantially decreased, from older than 70 to under 65 (burtless & quinn, 2002). based on this trend, it is unlikely that many workers really expect to never retire, even if they report an unwillingness to retire. some workers may fail to plan for retirement and then choose the never retire response when asked because they have no idea at what age they can retire. this suggests that providing a never retire response might be an indication of retirement inadequacy. in this regard, hanna et al. (2017) concluded that the expectation of never retiring is a signal that workers have failed to prepare for their retirement, rather than indicating a preference for working forever. therefore, investigating the relationship between financial knowledge and the never retire response may provide valuable insights for financial educators and policymakers. background financial knowledge has been defined in various ways, and sometimes has been used interchangeably with the term financial literacy (huston, 2010). financial knowledge is typically measured either objectively (i.e., being able to correctly answer questions related to financial decisions) or subjectively (e.g., a self-described perception of knowledge). many previous studies have found effects of financial knowledge on financial behavior, so in this study, we focus on the effect of objective financial knowledge, as measured by the proportion of financial questions answered correctly, combined with a person’s subjective perception of their financial knowledge. some researchers report a lack of financial knowledge among those living in the united states across age bands (lusardi & mitchell, 2007; mandell & klein, 2009). some studies find that the relationship between financial knowledge and individuals’ behavior to be complicated since financial knowledge does not automatically result in optimal financial decision-making or behavioral outcomes (braunstein & welch, 2002). understanding the relationship between financial knowledge and financial decision-making is increasingly recognized as an area of critical importance. objective financial knowledge, which reflects the actual understanding of financial matters, and subjective financial knowledge, which reflects perceived financial knowledge, is known, for example, to significantly impact workers’ important financial decisions, such as hardship withdrawals from retirement accounts (lee & hanna, 2020; utkus & young, 2011). financial decisions made by workers can also be impacted by financial knowledge overconfidence, which has been defined as exhibiting an aboveaverage subjective financial knowledge level and a below-average objective financial knowledge level. lee and hanna (2020) noted that workers with financial knowledge overconfidence are more likely to make financial decisions that may damage their future retirement income security, such as taking early withdrawals from retirement accounts. in some studies, taking early withdrawals from retirement accounts, also called leakage of retirement assets, leaves household retirement accounts underfunded and insufficient in terms of supporting support expenditures when the account owner retires (bovbjerg, 2010; engelhardt, 2002; munnell & webb 2015). although choosing never to retire could be a rational and reasonable decision, hanna et al. (2017) argued that most respondents giving a never retire response had not planned for retirement. analyses of never retire respondents may provide additional insights into retirement planning and retirement adequacy. this study extends the current literature by testing the effect of objective and subjective financial knowledge on expected retirement ages, which is a recommendation made by hanna et al. specifically, we extend the hanna et al. model by testing the effect of financial knowledge on the never retire response. we also test the relationship between overconfidence (i.e., high subjective and low objective knowledge) in financial knowledge and the never retire response. this study focuses on the moderating role of financial knowledge on the never retire response. financial services review, 32(3) 22 methodology data and sample for this paper, we used a combination of the 2016 and 2019 scf datasets (bhutta et al., 2020). the scf data has been collected every three years since 1983. the scf is sponsored by the u.s. federal reserve board and the u.s. department of the treasury. the scf data includes detailed information on household characteristics, family structure, and household financial decisions. the scf provides an ideal dataset for a retirementrelated study because respondents are asked many detailed questions about their retirement attitudes and behaviors, including their expected retirement age, characteristics of retirement pension plans, income, work status, insurance, and financial assets. combining the 2016 and 2019 scf datasets allows for more robust estimates of some effects from the two survey years that were somewhat similar.4 we did not include the recently released 2022 scf because of the confounding effects of the covid-19 pandemic. when considering the results from this study, it is important to understand how data are collected in the scf. there is an issue of using information from the household head versus the survey respondent (see lindamood et al., 2007). this is not an issue for non-couple households, but it can be an issue for partnered households. in the 2016 scf, for example, the head is not the respondent in 45% of households (hanna et al., 2018). traditionally, the head (defined by the scf as the male in mixed-sex couples) has been the focus of research in terms of labor force participation and earnings, with males traditionally being more consistent in labor force participation. the use of the household head does present some conceptual challenges for studies like this one. in the public dataset, for couple households, the racial/ethnic identification is only available for the respondent. the respondent answers all attitudinal and financial knowledge questions. the expected retirement age of the head is provided by the 4 our research approach followed hanna et al. (2017) who used a sample of households with heads who were working full-time and aged 35 to 60 years. we followed their reasoning in the sample selection, including the patterns of labor force participation by respondent, even if they are different persons. these limitations should be considered, but if the respondent can assess what the head of household would answer, in the context of this study, results should not be too biased for expected retirement age. the alternative of using a respondent’s expected retirement age would likely produce more distortions. for this study, we assumed that a respondent’s objective financial knowledge and subjective assessment of financial knowledge are close to those of the household head. in order to compare our results to those of hanna et al. (2017), we used information for the head of household for expected retirement age. we performed analyses using a respondent’s age, employment status, and expected retirement age, which resulted in a much smaller analytic sample (unweighted n of 2,667 compared to 4,607 using household heads, with 356 “never retire” responses for respondents, compared to 685 for heads). the descriptive patterns for relationships between never retire rates and financial knowledge variables are very similar for respondents and for household heads. the final analytic sample included 4,607 households. dependent variable the dependent variable for the analysis was created by using responses to a question that asked about the expected age to stop working full-time. if a respondents replied that the head would never retire, the variable “never retire” was coded as 1, otherwise 0. this coding matched hanna et al. (2017) and zhang and hanna (2011). independent variables we tested three models. model 1 matched hanna et al. (2017) in terms of the independent variables. model 2 included household characteristics, in addition to financial knowledge variables. one focal independent variable was objective financial knowledge. this variable, ranging from 0 to 3, was calculated using three questions in the scf. these questions, known as the “big three” (hastings et al., 2013), relate to compound the age of household heads. we obtained results similar to hanna et al. in that full-time employment decreased rapidly after the age of 60 while being very low before age 35. zhang et al. 23 interest, real rates of return, and risk diversification. the estimated value of the variable was based on the number of questions answered correctly. if a respondent answered all three questions correctly, their score was 3. if a respondent answered two questions correctly, their score was 2. if a respondent answered one question correctly, the variable was equal to 1. if a respondent answered the three questions incorrectly, they received a score of 0. another focal variable was subjective financial knowledge. this variable, ranging from 0 to 10, was based on a self-evaluation of financial knowledge by a respondent. if a respondent believed they were not at all financially knowledgeable, they received a score equal to 0. if a respondents believed they were very knowledgeable in the financial domain, their score was equal to 10. respondents were allowed to choose any integer between 0 and 10. the variable was recoded into four categories: (a) low, (b) some, (c) good, and (d) high. for the logistic regression (described below), we used the subjective financial knowledge score. in model 3, we defined financial confidence to align with lee and hanna (2020), which was based on whether a respondent had above or below median objective and subjective financial knowledge. the model used four categories of financial confidence: (a) appropriate high confidence, if subjective financial knowledge and objective financial knowledge were both high; (b) appropriate low confidence, if subjective financial knowledge and objective financial knowledge were both low; (c) overconfident, if subjective financial knowledge was high but objective financial knowledge was low; and (d) underconfident if subjective financial knowledge was low but objective financial knowledge was high. control variables the following control variables were included in the models: (a) racial/ethnic status of the respondent, (b) self-employment, (c) health status, (d) marital status, (e) life expectancy, (f) education, (g) job title of the household head, (h) whether the head of household had a defined benefit pension, (i) whether everyone was covered by health insurance, (j) household income, (k) net worth, and (l) the expectation to inherit a substantial amount of money. we also included the following control variables not used by hanna et al. (2017): (a) economic outlook, (b) willingness to take the financial risk, (c) satisfaction of expected retirement income, and (d) a dummy variable for survey year. empirical analysis we used descriptive analyses to examine the determinants of the never retire response without controlling for other variables. we then used weighted repeated-imputation inference (rii) means tests to ascertain whether differences in the never retire rates in the descriptive analyses were significant (see hanna et al., 2017; montalto & sung, 1996; montalto & yuh, 1998). the primary analyses were conducted using logistic regressions for models 1, 2, and 3. for the logistic regression analyses, we utilized multiple imputation procedures applying the rii technique to estimate the variances appropriately (see lindamood et al., 2007). the models were estimated as follows: model 1: log p(nri) 1−p(nri) = α0 + α1hi (1) where, 𝑛𝑟𝑖 represents whether workers’ retirement expectation is “never retire”; 𝐻𝑖 represents the set of control variables listed above. model 2: log p(nri) 1−p(nri) = α0 + α1obji + α2subi + α3ci (2) where, 𝑛𝑟𝑖 represent whether workers’ retirement expectation is “never retire”; 𝑂𝑏𝑗𝑖 represents household heads’ objective financial knowledge; 𝑆𝑢𝑏𝑖 represents household heads’ subjective financial knowledge; and 𝐶𝑖 represents a set of control variables. model 3: log p(nri) 1−p(nri) = α0 + α1confi + α2ci (3) where, 𝑛𝑟𝑖 represents whether workers retire expectation is “never retire”; 𝑂𝑣𝑒𝑟𝑖 represents household heads’ financial knowledge confidence; and 𝐶𝑖 represents a set of control variables. financial services review, 32(3) 24 results descriptive analysis the analytic sample included 4,607 households with full-time employed household heads aged from 35 to 60 years, with 2,395 from the 2016 scf, and 2,212 from the 2019 scf. the never retire rate for the combined analytic sample was 15.0%, with 15.2% in 2016 and 14.8% in 2019. the rates were not significantly different. table 1 shows the descriptive results for that sample by financial knowledge categories. table 1. rate of never retire by household characteristics, full-time worker households with head aged 35-60, 2016 and 2019 scf variable distribution % never retire significance level total sample 100.0% 15.03% 2016 scf 52.68% 15.22% reference 2019 scf 47.31% 14.82% 0.395 objective financial knowledge no knowledge (0 out of 3 right) 3.33% 18.88% <.001 poor knowledge (1 out of 3 right) 14.66% 17.41% <.001 fair knowledge (2 out of 3 right) 33.79% 18.32% <.001 high knowledge (3 out of 3 right) 48.23% 11.73% reference subjective financial knowledge no knowledge 1.96% 29.75% reference poor knowledge 17.17% 19.52% <.001 fair knowledge 53.62% 13.53% <.001 high knowledge 27.25% 14.09% <.001 financial knowledge confidence appropriate low confidence 29.47% 18.68% 0.080 underconfident 21.98% 11.76% <.001 overconfident 22.31% 17.33% reference appropriate high confidence 26.25% 11.71% <.001 notes. n = 4,607 households. weighted analyses with rii means tests. objective financial knowledge was estimated using the number of questions a respondent answered correctly. only 3.3% of respondents answered all three questions incorrectly, while 48.2% answered the three questions correctly. just under 14.7% answered one question correctly, and 33.8% answered two questions correctly. of the respondents who answered all three questions incorrectly, 18.9% said they would never retire, but only 11.7% of those who answered all three questions correctly said they would never retire. the never retire rate for those who missed one or more questions was higher 5 the p values shown in table 1 are rii means test results for comparisons to the reference group (answered all three questions correctly). for instance, the never retire rate for those who answered the three than the rate for those who answered all three questions correctly, and the difference was significantly higher than the rates for those who answered all three questions correctly.5 in terms of subjective financial knowledge, 29.8% of those who indicated that they are not financially knowledgeable (levels 0, 1, or 2) expected to never retire, while 14.1% of respondents who believe they are very knowledgeable (levels 9 or 10) gave the never retire response. the rates for “some,” “good,” and “high” levels of subjective knowledge were significantly lower than the rates for those with questions correctly was significantly lower than the rate for those who answered two questions correctly and also lower than the rate for those who answered one question correctly. zhang et al. 25 “low” perceived financial knowledge. the rates for “some” and “good” were significantly different, as were the rates for “some” and “high.” when respondents were categorized into financial knowledge confidence groups, those who were underconfident and those who had appropriately high confidence were significantly less likely to have chosen the never retire response than those who had appropriately low confidence. those with appropriately low confidence and the overconfident had never retire rates of 18.7% and 17.3%, respectively, compared to 11.8% for those with appropriately high financial confidence levels, and 11.7% for the financially underconfident. regression results the following discussion highlights results from the three logistic regression models using the never retire response as the outcome variable. model 1 included the same variables used by hanna et al. (2017) (table 2). model 2 utilized the same independent variables in model 1, plus dummy variables for objective and subjective financial knowledge (table 3). model 3 included the independent variables from model 1 plus dummy variables for the level of financial knowledge confidence (table 4). the results shown in table 2 are mostly consistent with the results of hanna et al. (2017), with one exception. hanna et al. found household income to be negatively related to the never retire response. we observed the effect of household income to be negative; however, the two-tail p value was not significant. hispanic respondents were less likely to give a never retire response than white respondents. having a defined benefit plan was negatively associated with the never retire response. expecting a substantial inheritance and spending less than income was also negatively associated with the never retire response. those who expected their retirement income to be satisfactory were less likely to give a never retire response than those who expected retirement income to be “enough.” single males and partnered couples were more likely to give a never retire response than otherwise similar married couples. being self-employed was also positively associated with the never retire response. years of education, expecting to live a longer time, and having health insurance also were negatively related to the never retire response. there was no significant difference between 2016 and 2019 in the never retire response. table 3 shows the results from model 2, with the financial knowledge variables added to the independent variable list. compared to heads in households with high objective financial knowledge levels, heads in households with fair financial knowledge levels were more likely to give a never retire response, with an odds ratio of a never retire response 1.24 times as high as the ratio for those who got all of the questions correct. the association between control variables and the never retire response were similar to the results in model 1, except that black respondents were less likely to give a never retire response than white respondents. in addition, those who were willing to take above-average or average risk were less likely to give a never retire response than those unwilling to take any risk. financial services review, 32(3) 26 table 2. logistic regression analysis of the likelihood of full-time worker household heads aged 35–60 years expecting never to retire, model 1 variable coefficient standard error chisquare p value odd ratio log (net worth) (ln [.01] if net worth ≤ 0) -0.0133 0.0096 0.1690 0.9870 log (income) (ln [.01] if income ≤ 0) -0.0282 0.0238 0.2362 0.9722 racial ethnic status of respondent (reference category = white) black -0.2183 0.1449 0.1319 0.8038 hispanic -0.3146 0.1448 0.0298 0.7300 asian/other -0.1733 0.1913 0.3649 0.8410 have a defined benefit pension plan -0.8625 0.2039 <.0001 0.4224 perception of the adequacy of retirement income (reference category = enough to maintain living standards) very satisfactory 0.0219 0.1331 0.8691 1.0224 satisfactory -0.6321 0.1586 0.0001 0.5318 inadequate -0.0039 0.1477 0.9787 0.9964 totally inadequate 0.6015 0.1233 <.0001 1.8250 head self-employed 0.6370 0.1041 <.0001 1.8910 years of education of the head -0.0691 0.0184 0.0002 0.9332 perceived health status of the head (reference category = good health) excellent health -0.0183 0.1073 0.8647 0.9818 fair health 0.0749 0.1217 0.5383 1.0780 poor health -0.2725 0.3646 0.4549 0.7618 all in household covered by health insurance -0.3675 0.1233 0.0029 0.6926 age of the head -0.0033 0.0062 0.5950 0.9966 expect a substantial inheritance or other transfer -0.3728 0.1282 0.0036 0.6888 expectations for the economy (reference category = better) worse 0.1830 0.1351 0.1757 1.2010 same -0.1740 0.1206 0.1493 0.8406 life expectancy for the head (reference category = younger than 71 years) live to 71-80 years -0.6152 0.1455 <.0001 0.5410 live to 81 years or older -0.4778 0.1409 0.0007 0.6208 household type (reference category = married) partnered couple 0.4551 0.1460 0.0018 1.5764 single male 0.4068 0.1253 0.0012 1.5020 single female 0.0067 0.1385 0.9615 1.0066 spending relative to income (reference category = same as income) more than income -0.0252 0.1348 0.8515 0.9752 less than income -0.1449 0.1061 0.1719 0.8652 job title of the head (reference category = transportation) executive, admin, manager, teachers -0.1021 0.1628 0.5306 0.9028 engineer, technician, office -0.1794 0.1779 0.3134 0.8358 protective and miscellaneous service -0.0338 0.1932 0.8612 0.9670 construction, production, repair -0.1129 0.1752 0.5195 0.8934 farming, fishing, forestry 0.2131 0.2984 0.4752 1.2380 year = 2019 0.0573 0.0876 0.5133 1.0590 intercept 0.6934 0.4858 0.1535 concordance (averaged for 5 implicates) 71.36% notes. n = 4,607 households. unweighted repeated-imputation inference (rii) analysis of combination of 2016 and 2019 survey of consumer finances dataset (scf). zhang et al. 27 table 3. logistic regression analysis of the likelihood of full-time worker household heads aged 35–60 years expecting never to retire, model 2, with financial knowledge variables variable coefficient standard error chisquare p value odd ratio objective financial knowledge (0 to 3) (reference category=high knowledge (3 right)) no knowledge (0 right) 0.0807 0.2506 0.7475 1.0840 poor knowledge (1 right) 0.1183 0.1450 0.4145 1.1256 fair knowledge (2 right) 0.2167 0.1067 0.0423 1.2422 subjective financial knowledge 0.0088 0.0228 0.7000 1.0088 log (net worth) (ln [.01] if net worth ≤ 0) -0.0065 0.0098 0.5054 0.9936 log (income) (ln [.01] if income ≤ 0) -0.0181 0.0241 0.4538 0.9822 racial ethnic status of respondent (reference category = white) black -0.2898 0.1468 0.0484 0.7484 hispanic -0.4346 0.1476 0.0032 0.6474 asian/other -0.2366 0.1942 0.2231 0.7892 have a defined benefit pension plan -0.8390 0.2025 <.0001 0.4326 perception of the adequacy of retirement income (reference category = enough to maintain living standards) very satisfactory -0.0205 0.1359 0.8802 0.9798 satisfactory -0.5788 0.1596 0.0003 0.5608 inadequate -0.0303 0.1492 0.8390 0.9702 totally inadequate 0.4970 0.1255 0.0001 1.6438 head self-employed 0.6438 0.1068 <.0001 1.9036 years of education of the head -0.0514 0.0190 0.0069 0.9500 perceived health status of the head (reference category = good health) excellent health -0.0157 0.1086 0.8848 0.9844 fair health 0.0489 0.1227 0.6901 1.0502 poor health -0.2129 0.3678 0.5627 0.8088 all in household covered by health insurance -0.3024 0.1248 0.0154 0.7392 age of the head 0.0004 0.0064 0.9535 1.0006 expect a substantial inheritance or other transfer -0.3221 0.1299 0.0132 0.7248 expectations for the economy (reference category = better) worse 0.1436 0.1366 0.2934 1.1546 same -0.1666 0.1216 0.1708 0.8464 life expectancy for the head (reference category = younger than 71 years) live to 71-80 years -0.6357 0.1473 <.0001 0.5300 live to 81 years or older -0.4959 0.1425 0.0005 0.6096 household type (reference category = married) partnered couple 0.4323 0.1481 0.0035 1.5408 single male 0.4334 0.1279 0.0007 1.5424 single female -0.0901 0.1416 0.5247 0.9138 spending relative to income (reference category = same as income) more than income 0.0225 0.1363 0.8690 1.0226 less than income -0.0977 0.1078 0.3647 0.9070 job title of the head (reference category = transportation) executive, admin, manager, teachers 0.0130 0.1663 0.9376 1.0132 engineer, technician, office -0.1105 0.1810 0.5414 0.8954 protective and miscellaneous service -0.0087 0.1958 0.9647 0.9916 construction, production, repair -0.0941 0.1776 0.5960 0.9100 farming, fishing, forestry 0.3370 0.3034 0.2666 1.4014 financial services review, 32(3) 28 variable coefficient standard error chisquare p value odd ratio saving reasons (reference category = other saving reasons) cannot save 0.5585 0.3598 0.1206 1.7482 retirement -0.4499 0.0990 <.0001 0.6376 investment -0.1330 0.3024 0.6600 0.8758 financial risk tolerance (reference category = take no risk) substantial risk 0.0758 0.1869 0.6852 1.0788 above average -0.5839 0.1380 <.0001 0.5574 average -0.5589 0.1151 <.0001 0.5718 year = 2019 0.0465 0.0891 0.6017 1.0476 intercept 0.3474 0.5287 0.5112 concordance (averaged for 5 implicates) 73.48% notes. n = 4,607 households. unweighted repeated-imputation inference (rii) analysis of combination of 2016 and 2019 survey of consumer finances dataset (scf). table 4 shows results from the model 3 logistic regression estimation, with financial knowledge confidence variables added to the independent variables from table 2. the likelihood of the never retire response was higher for those who exhibited overconfidence compared to those who were underconfident. the effects of the other control variables were similar to the results for models 1 and 2. there was no significant difference in the never retire rates between the 2016 and 2019 scf respondents. conclusion and implications in the descriptive analyses, we found that respondents who missed one or more objective financial knowledge questions were significantly more likely to have chosen the never retire response than those who answered all three questions correctly. we found a similar pattern for the subjective financial knowledge question with those who perceived themselves in the lowest subjective knowledge category being significantly more likely to choose the never retire response than those who perceived themselves to be in the highest subjective knowledge category. we also found that those with appropriately low confidence and overconfidence had higher never retire rates than those with appropriately high confidence and with under-confidence. the relationships between the never retire response and financial knowledge were weaker when other variables were controlled in the logistic regressions. with both objective and subjective financial knowledge and many other household characteristics controlled, respondents with a fair level of financial knowledge (i.e., missed one question) were more likely to give a never retire response than respondents who got all three questions correct. we also found that household characteristics, financial situation, and financial attitudes had significant relationships with the never retire response. for the model 3 logistic regression (i.e., the model that included financial knowledge confidence variables and the control variables), financially knowledgeable underconfident respondents were significantly less likely than similar overconfident respondents to give a never retire response, suggesting the importance of teaching workers not only financial knowledge but also the limits of their knowledge. financial planners, financial counselors, and financial educators should pay attention to the level of financial knowledge and confidence levels of their clients and also consider risk tolerance and other factors that are directly related to retirement expectations. educators and policymakers not only need to consider the impacts of financial knowledge on retirement plans and make relevant plans for education and policy but also need to consider how to help people build suitable confidence in financial knowledge. workers could benefit from these plans by making more rational retirement plans and financial plans based on their own situations. zhang et al. 29 in summary, evaluations of retirement adequacy need to include careful considerations of what never retire responses mean. some analyses of the projected retirement adequacy of u.s. workers (e.g., yuh et al., 1998) assume that those giving a never retire response will retire at age 70. as hanna et al. (2017) demonstrated, it is plausible that those households will retire at a much younger age, and therefore will have less retirement adequacy. therefore, some projections of the proportion of workers on track for an adequate retirement might be too optimistic. table 4. logistic regression analysis of the likelihood of full-time worker household heads aged 35–60 years expecting never to retire, model 3, with confidence in financial knowledge variables variable coefficient standard error chisquare p value odd ratio financial confidence categories (reference category=overconfident) appropriate low confidence -0.1472 0.1329 0.2678 0.8630 appropriate high confidence -0.0832 0.1237 0.5015 0.9202 underconfident -0.3212 0.1432 0.0249 0.7254 log (net worth) (ln [.01] if net worth ≤ 0) -0.0068 0.0098 0.4876 0.9930 log (income) (ln [.01] if income ≤ 0) -0.0187 0.0240 0.4357 0.9812 racial ethnic status of respondent (reference category = white) black -0.2969 0.1467 0.0430 0.7432 hispanic -0.4441 0.1475 0.0026 0.6414 asian/other -0.2311 0.1942 0.2341 0.7934 have a defined benefit pension plan -0.8378 0.2026 0.0000 0.4328 perception of the adequacy of retirement income (reference category = enough to maintain living standards) very satisfactory -0.0438 0.1368 0.7487 0.9574 satisfactory -0.5878 0.1597 0.0002 0.5558 inadequate -0.0278 0.1491 0.8518 0.9728 totally inadequate 0.4937 0.1255 0.0001 1.6382 head self-employed 0.6375 0.1067 <.0001 1.8918 years of education of the head -0.0516 0.0190 0.0065 0.9496 perceived health status of the head (reference category = good health) excellent health -0.0224 0.1088 0.8367 0.9778 fair health 0.0507 0.1227 0.6796 1.0520 poor health -0.2067 0.3675 0.5738 0.8136 all in household covered by health insurance -0.2956 0.1249 0.0179 0.7440 age of the head 0.0004 0.0064 0.9535 1.0006 expect a substantial inheritance or other transfer -0.3237 0.1299 0.0127 0.7236 expectations for the economy (reference category = better) worse 0.1426 0.1366 0.2966 1.1534 same -0.1640 0.1217 0.1778 0.8488 life expectancy for the head (reference category = younger than 71 years) live to 71-80 years -0.6340 0.1473 <.0001 0.5310 live to 81 years or older -0.4942 0.1428 0.0005 0.6106 household type (reference category = married) partnered couple 0.4441 0.1483 0.0028 1.5590 single male 0.4381 0.1278 0.0006 1.5496 single female -0.0920 0.1414 0.5155 0.9118 financial services review, 32(3) 30 variable coefficient standard error chisquare p value odd ratio spending relative to income (reference category = same as income) more than income 0.0193 0.1364 0.8873 1.0194 less than income -0.1022 0.1079 0.3435 0.9030 job title of the head (reference category = transportation) executive, admin, manager, teachers 0.0153 0.1663 0.9269 1.0156 engineer, technician, office -0.1116 0.1810 0.5373 0.8946 protective and miscellaneous service -0.0078 0.1957 0.9681 0.9924 construction, production, repair -0.0945 0.1776 0.5947 0.9102 farming, fishing, forestry 0.3356 0.3036 0.2689 1.3994 saving reasons (reference category = other saving reasons) cannot save 0.5640 0.3596 0.1168 1.7576 retirement -0.4483 0.0990 <.0001 0.6386 investment -0.1483 0.3030 0.6246 0.8624 financial risk tolerance (reference category = take no risk) substantial risk 0.0659 0.1868 0.7243 1.0682 above average -0.5841 0.1380 <.0001 0.5574 average -0.5562 0.1151 <.0001 0.5736 year = 2019 0.0433 0.0890 0.6267 1.0444 intercept 0.6523 0.4981 0.1904 concordance (averaged for 5 implicates) 73.48% notes. n = 4,607 households. unweighted repeated-imputation inference (rii) analysis of combination of 2016 and 2019 survey of consumer finances dataset (scf). references bhutta, n., bricker, j., chang, a. c., dettling, l. j., goodman, s., hsu, j. w., moore, k. b., reber, s., henriques volz, a., & windle, r. a. (2020). changes in u.s. family finances from 2016 to 2019: evidence from the survey of consumer finances. federal reserve bulletin, 106(5), 1-42. bovbjerg, b. d. (2010). 401(k) plans: policy changes could reduce the long-term effects of leakage on workers’ retirement savings. diane publishing. braunstein s., welch c. (2002). financial literacy: an overview of practice, research, and policy. federal reserve bulletin, 88, 445-457. burtless, g., & quinn, j. f. (2002). is working longer the answer for an aging workforce? working papers in economics, 82, 1-11. engelhardt, g. v. (2002). pre-retirement lumpsum pension distributions and retirement income security: evidence from the health and retirement study. national tax journal, 55(4), 665-685. hanna, s. d., kim, k. t., & lindamood, s. (2018). behind the numbers: understanding the survey of consumer finances. journal of financial counseling and planning, 29(2), 410-418. hanna, s. d., zhang, l., & kim, k. t. (2017). do worker expectations of never retiring indicate a preference or an inability to plan? journal of financial counseling and planning, 28(2), 268-284. hastings, j. s., madrian, b. c., & skimmyhorn, w. l. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5(1), 347-373. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44(2), 296-316. lee, s. t., & hanna, s. d. (2020). financial knowledge overconfidence and early withdrawals from retirement accounts. zhang et al. 31 financial planning review, 3(2), e1091. https://doi.org/10.1002/cfp2.1091. lindamood, s., hanna, s. d., & bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41(2), 195-222. lusardi, a., & mitchell, o. s. (2007). financial literacy and retirement planning: new evidence from the rand american life panel. michigan retirement research center research paper no. wp, 157. mandell, l., & klein, l. s. (2009). the impact of financial literacy education on subsequent financial behavior. journal of financial counseling and planning, 20(1), 10. montalto, c. p., & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances. journal of financial counseling and planning, 7, 133–146. montalto, c. p. & yuh, y. (1998). estimation of nonlinear models with multiply imputed data. journal of financial counseling and planning, 9 (1), 97-101. munnell, a., & webb, a. (2015). the impact of leakages from 401(k) s and iras. center for retirement research at boston college, (152). utkus, stephen p., and jean a. young. (2011) financial literacy and 401(k) loans. in financial literacy: implications for retirement security and the financial marketplace, edited by olivia s. mitchell and annamaria lusardi, pp. 59–75. oxford university press. yuh, y., montalto, c. p., & hanna, s. d. (1998). are americans prepared for retirement?. journal of financial counseling and planning, 9(1), 1-12. zhang, l., & hanna, s. d. (2011). the determinants of planned retirement age. academy of financial services proceedings. retrieved from http://www.academyfinancial.org/resources /documents/proceedings/2011/g3-zhanghanna.pdf. http://www.academyfinancial.org/resources/documents/proceedings/2011/g3-zhang-hanna.pdf http://www.academyfinancial.org/resources/documents/proceedings/2011/g3-zhang-hanna.pdf http://www.academyfinancial.org/resources/documents/proceedings/2011/g3-zhang-hanna.pdf call for papers: due july 1, 2020 the academy of financial services 34th annual mee�ng september 29-30, 2020 virtual conference the academy of financial services will hold its annual conference in conjunction with the fpa's annual conference. in light of recent events, we have restructured pricing for this year’s virtual annual meeting: $199 for academics and prac��oners and $99 for students. the afs conference will feature speakers, symposia, several special sessions, posters, and a reception. among them, we will introduce a new panel session for phd students, highlighting how to best navigate the job market. with the generous support of our sponsors, the academy has awarded several best paper awards during past meetings and we anticipate continuing best paper awards in 2020. we will continue with our emerging scholar award to a current graduate student for promising research work on a paper or poster presented at the conference. in addi�on, in 2020, we will ini�ate a new, program directors track. the goal is to allow program directors to present and discuss program issues and best practices in a panel environment, such as “working with your university’s foundation”, “capstone course cases: what’s the right content?”, “understanding career paths and student fit” and “developing a passionate program in a box: scholarships, competitions, student organizations”. we welcome other panel topics deemed beneficial to program directors. submission informa�on: research papers and abstracts covering all aspects of individual financial management and education are sought for inclusion in the program. papers in the areas of estate planning, insurance, tax accounting aspects of financial planning, investments, and retirement planning are encouraged. proposals for panel discussions and tutorials devoted to current issues in individual financial management or the practice of financial planning will also be considered for inclusion in the program. several sessions will be registered for continuing education (ce) credit with the cfp® board. submit your paper or abstract here: https://proposalspace.com/calls/d/1181 submissions are due july 1st the review period ends on july 31st, 2020 with the selection period and formulation of the agenda estimated to be completed by august 31st, 2020. notice of acceptance as an oral session or a poster is targeted for september 5th, 2020. note that the terms and conditions of this call-for are outlined in the online submission form. only accepted presentations are included in the subsequent proceedings, which are posted on the afs website. thus, the proceedings publication is refereed in order to accommodate the rules of the american association of intercollegiate schools of business-international (aacsb) on table 2-1 (intellectual contributions). for further informa�on: visit the afs website at academyfinancial.org that will be frequently updated. for content ques�ons contact program chair, dr. terrance k. mar�n jr. at terrance.martin@uvu.edu manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires december 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in “impact of consumer perceptions of industry corruption on the choice to engage a financial advisor: does gender matter?” by winchester, leak, and mccoy, the authors show that the likelihood to use female financial advisors is greater than that of male advisors when consumers perceive little corruption in the financial services industry. why was this finding hypothesized? a. hiring more females into advisement positions demonstrates that the industry is progressive and sensitive to societal demands. b. prior research has shown women to be viewed as more trustworthy and ethical than their male counterparts. c. female financial advisors have been shown to be objectively less corrupt than male advisors. d. there is a universal push for diversity, equity, and inclusion in the financial services industry. when external factors like corruption are not in play, women are preferred and demanded because of the relative scarcity that currently exists. 2. in winchester, leak, and mccoy, the authors note other factors that may need to be researched to further understand marketplace demand for female financial advisors. these unevaluated factors include: a. consumer ideological beliefs about diversity b. financial advisors’ age in conjunction with their gender c. advisors’ race/ethnicity in conjunction with their gender d. all of the above 3. in their article, “the effect of racial/ethnic differences on the financial obligations ratio of renters,” the authors find a. renter households with higher levels of financial risk tolerance tend to have higher financial obligations ratios than otherwise similar households with lower levels of risk tolerance. b. renter households with higher levels of financial risk tolerance tend to have lower financial obligations ratios than otherwise similar households with lower levels of risk tolerance. c. renter households with higher levels of financial risk tolerance tend to have lower financial obligations ratios than otherwise similar households with lower levels of risk tolerance. d. for renter households there was no statistically significant relationship between levels of financial risk tolerance and financial obligations ratios. 4. in their article, “the effect of racial/ethnic differences on the financial obligations ratio of renters,” what type of relationship exist between renter household and financial obligation ratios a. hispanic homeowner households tend to have higher financial obligations ratios than otherwise white renter households. b. hispanic renter households tend to have lower financial obligations ratios than otherwise similar white renter households. c. for renter households there was no statistically significant relationship between racial/ethnic groups and financial obligations ratios. d. hispanic renter households tend to have lower financial obligations ratios than otherwise similar white renter households. 5. winchester, leak, and mccoy present the following findings: a. female financial advisors are universally preferred to males regardless of how corrupt the financial services industry is perceived. b. female financial advisors provide better guidance to clients than males when the financial services industry is perceived as corrupt. c. when the financial services industry is perceived as corrupt, consumers are equally likely to engage with male and female advisors given they are similarly positioned as partners that can guide the client to meet financial goals. d. male financial advisors are less corrupt than female advisors, and therefore are universally preferred to female advisors. manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: terrance k. martin, e-mail: martintk@wssu.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services terrance k. martin college of arts, sciences, business, and education winston-salem state university reynolds center, rm 111 601 s. martin luther king jr. drive winston-salem, nc 27110 (address service requested) from the editor this issue contains volume 29 issue 1 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “optimism, overconfidence, and insurance decisions” is coauthored by jennifer coatsa and vickie bajtelsmit, both at colorado state university. the authors present experimental evidence regarding overconfidence, optimism and insurance decisions. they distinguish between an individual’s optimism bias and overconfidence bias, a contribution particularly important for understanding insurance decisions related to risks beyond the purchaser’s control. their results show that optimistic participants incur a higher total cost of risk and are more likely to underinsure than non-optimistic participants, even when purchasing insurance maximizing expected payoffs. they also find that overconfidence does not significantly affect the decision to insure, participants with higher overall overconfidence show larger differences in insurance behavior when the risk of loss arises from their own mistakes. the second article “the impact of using financial technology on positive financial behaviors” is coauthored by qianwen bi, utah valley university, lukas r. dean, utah valley university, tao guo, william paterson university, and xu sun, utah valley university, the authors use the 2013 survey of consumer finances data to explore the impact of financial technologies on households’ positive financial behaviors. the authors find that only planning technologies (e.g. direct deposit and computer software) are positively related to households’ engagement in positive financial behaviors. they also find that the impact of transaction technologies (e.g. using atm card, credit card, phone banking, and computer banking) is negative. the third article, “using investor utility to determine portfolio choice with reits” is coauthored by wei feng, lynn university, travis l. jones, florida gulf coast university, and marcus t. allen, florida gulf coast university. the authors examine the decision of individual investors to allocate a portion of their existing investment portfolios to reits. they derive the risk preferences of investors represented by their benchmark portfolios of stocks and bonds and then use the risk preferences to determine portfolio decisions regarding reits. their analysis shows that investors with lower risk aversion tend to have a more 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) v–vii substantial stock component in their benchmark porfolio and will obtain higher risk-return benefits from adding reits. the final article, “demographic and psychological differences between chapter 13 bankruptcy filers and non-filers” is coauthored by scott e. kehiaian, southern new hampshire university, albert a. williams, nova southeastern university, and carolyn l. bird, north carolina state university. in this article the authors find financial, demographic, and psychological differences between chapter 13 filers and non-filers. they also show that financial training reduces the likelihood of filing for personal bankruptcy and males are twice as likely as females to be filers. a single person is less likely to file than a married person and homeowners are more likely than renters to be filers. increases in education, religious commitment, and parents’ income reduce the likelihood of filing. increases in the psychological factors, self-efficacy, locus of control, and self-control, reduce the likelihood of filing for chapter 13 bankruptcies. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. fenaba addo lisa fiksenbaum thomas krueger jinfei sheng stephen agnew john gathergood marie-eve lachance william skimmyhorn abdullah al-bahrani john grable kyre lahtinen christina stoddard arthur allen adam greenberg andre liebenberg ning tang somer anderson john grigsby hanna lim sharon tennyson anders anderson michael guillemette ana luiza paraboni ruilin tian nikolaos artavanis andreas hackethal annamaria lusardi colleen tokar asaad kremena bachmann nathan harness john lynch jr sami vahamaa vickie bajtelsmit christine harrington zdravko marjanovic neal van zutphen bhanu balasubramnian stuart heckman terrance martin bruce vanstone mirco balatti robin henager greene camilla mazzoli christian walkshäusl michael batty robert henderson stephan meier william walstad levon blue thorsten hens steffen meyer tom warschauer paola bongini hal hershfield brianna middlewood jamie weathers peter brady russell james young park jaya wen sonya britt-lutter thomas jansson darshak patel melissa wilmarth j. michael collins jing jian xio cliff robb danielle winchester brenda cude jesse jurgenson david robinson jianzhong (andrew) zhang lucy delgadillo mary kabaci chris robinson terry zhang catherine d’hondt charlene kalenkoski john salter xin (jessica) zhao donna dudney elizabeth kiss marta serra-garcia timothy zimmer yaman erzurumlu vladimir kotomin daniel fernandes marc kramer fred fernatt vi s. michelson / financial services review 29 (2021) v–vii please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review s. michelson / financial services review 29 (2021) v–vii vii pii: 1057-0810(91)90031-s financial services review, l(2): 159-175 copyright 0 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. i i international diversification for the individual: a review jeff madura thomas j. o’brien thispaper reviews aspects ofthe literature on international investing that should be of interest to individual investors. three modern issues are covered: (i) the bene$ts of international diversij cation as the global markets continue to integrate; (2) the problem of currency exposure; and (3) effective means of achieving international diversijcation. strategies are discussed which enable the individual to apply suggestions from the research. by restricting the scope of the review to issues of most interest to the individual, we do not review research on international asset pricing theory and international market eficiency and “anomalies. ” individual investors have seen, and are continuing to see, significant growth in international investment opportunities. the globalization of investment ownership is part of the general integration of the world’s financial markets, made possible by technological advances and a worldwide liberalization of regulations regarding foreign ownership. domestic mutual funds now offer easy access to individuals interested in international funds, global funds, and country-specific funds. today, the existence of international mutual funds allows individual investors to purchase stocks in countries such as austria, indonesia, ireland, korea, malaysia, mexico, portugal, singapore, taiwan, thailand, and turkey. ilpotential internationalinvestmentstrategies the expanded opportunity set available to investors allows three distinct types of international investment strategies: jeff madura l sun bank professor of finance, finance and real estate department, college of business and public administration, florida atlantic university, po. box 3091, boca raton, fl 3343 i-099 1. thomas j. o’brien l associate professor of finance, department of finance, university of connecticut, 368 fairfield rd., u-41-f, storrs, ct 06268-2041. 160 financial services review, l(2) 1991 1. active strategies the first is an active “abnormal returns” strategy involving either specific security selection, or tactical asset allocation between “index” funds of various countries. the analysis behind these two strategies differs, as the tactical asset allocation strategy involves an analysis of countries’ economic resources and poli cies within the global economy, while security selection requires, in addition, traditional security-specific analysis. 2. a passive diversi$cation strategy the second strategy is international diversification, which is a passive, strate gic asset allocation strategy, where the asset classes are the “index” funds of various countries. the markowitz efficient frontier analysis, which identifies optimal port folios considering risk and return, is particularly well-suited for determining these strategic allocations in the global context. 3. currency betting the third strategy is “currency betting. ” exchange rate changes can add to or subtract from returns on international investments in their local markets. currency bets can be separated, to some extent, from investment strategies by currency hedging/borrowing strategies, or made on their own by simple bank deposits denominated in the currency desired. this paper reviews some research on international diversification. the emphasis is placed on diversification strategy, since it is likely to be the most relevant to individual investors. in addition, some relevant concepts in the separa tion of currency betting from internationally diversified investments are covered. the analyses behind active security selection and tactical asset allocation strategies are not covered, under the assumption that such strategies are more the province of professional managers, rather than individuals. after the review of some concepts and evidence related to international diversification and currency hedging, attention will be paid to methods available to individuals for making international investments. iii. internationaldiversification a. basic idea of international diversiycation the potential benefits from international diversification were originally demonstrated by grubel(l968) and levy and sarnat ( 1970). these studies showed that by diversifying across nations whose market cycles were not perfectly correl ated, investors could lower the volatility of portfolio returns at any level of expected return. the research methodology was to derive efficient portfolios using historical international zliversijication for the individual: a review 161 data on international stock markets. the efficient frontiers were shown to dominate those constructed with domestic securities only. a number of studies have followed the original ones over the last 25 years or so. see, for example solnik (1974), lessard (1976), ibbotson, cart-, and robertson (1982), solnik and noetzlin (1982), grauer and hakansson (1987) and others mentioned below. the entire line of research has consistently continued to advocate that international diversification is valuable. the efficient frontier methodology in the early international diversification research employed return parameters on a home-currency basis. that is, home currency returns included not only the performance of international investments in their local markets, but also the appreciation/depreciation of home currency versus foreign currency. thus while international diversification will appear beneficial if local market returns are high and less than perfectly correlated with home market returns, performance will also be enhanced if the foreign currencies appreciate relative to the home currency, during the data time period. let us use a numerical example to simplify the exposition and illustrate major points. a “two-country” example is employed, even though actual applications would include multiple countries. at first, the numerical example reviews some obvious concepts of international diversification; later, the example will be extended to less-obvious concepts. the numerical example is for a single-period investment horizon. the basic information for the example is given in table 1. table 1 shows that there are five (5) possible “states of nature” for the single period investment in a domestic (“united states”) stock index (u), for the invest ment in a foreign (“european”) stock index (e), and for the exchange rate expressed in terms of domestic currency per foreign currency ($/ecu). the $ returns on the european market ($e-the last column in table 1) are derived from the local european market returns (e) and the currency appreciation rates, via the well-known equation (1): table 1. “state ” market returns in local currency united states (u) europe (e) pet change $/ecu $ returns european ($e) 1 0.17 0.15 0.25 0.4374 2 0.12 0.33 0.25 0.6625 3 0.27 0.09 -0.2 -0.128 4 0.07 0.21 0 0.21 5 -0.03 -0.03 -0.2 -0.224 mean 0.12 0.15 0.02 0.1916 std 0.1 0.12 0.2015 0.3339 162 rd = (1 + rl) (1 + e) 1 financial services review, i(2) 1991 (1) where r, = the rate of return in domestic currency r, = the rate of return of the foreign market in local foreign currency e = the appreciation rate of the foreign currency in terms of its price per domestic currency from the basic information in table 1, it is possible to compute the correlation between the u.s. market and the $ returns on the european market for purposes of finding optimal portfolios. the reader can easily verify that the correlation is .1544. note also that the $/ecu rate is also positively correlated (.7692) with the local currency return in europe. this means that, on average, the european market rises as the ecu appreciates relative to the u.s. dollar. one reason for performing this analysis numerically, rather than analytically, is that since the analysis requires the multiplication of two random variables, the rate of return of the foreign market in local currency and the exchange rate, finding correlations between u and $e would be very complex if attempted analytically. the efficient frontier for u and $e is shown in figure 1 as the rightmost one; the other curves will be explained shortly. clearly, international diversification has benefitted the u.s. $-based investors, since the purely domestic u.s. market risk/ return (point u) plots below the u $e efficient frontier in figure 1. b. additional findings in international diversijkation while the analysis above provides the general idea behind international diver sification, the literature has documented some significant findings. some of these are reviewed below. (1) levy and sarnat (1974) emphasized how low or negative correlations between less developed country stock returns would allow for more e(r) .20 .lo .lo .20 r figure 1. international diversification for the individual: a review 163 effective international diversification. returns of less developed coun tries were commonly less correlated with other market returns, and generally assigned more weight in the efficient portfolios. the correla tions of returns among industrialized countries were relatively higher, so that diversification among markets offered less benefit. errunza (1977) and others provided further demonstration of the potential benefits of diversifying into less developed countries. (2) grubel and fadner (197 1) and others showed that diversification bene fits were achievable for various investment horizons. (3) solnik (1974) focused on foreign stocks rather than on indices to substantiate that a larger number of international stocks could reduce risk further. (4) biger (1979), mcdonald (1973), and others showed how the benefits from international diversification varied with the home country perspec tive, although some benefits were achievable for all perspectives. the degree of potential risk reduction varied among perspectives because exchange rate effects cause differences in the co-movements of stock returns among perspectives. (5) while the usual ‘default’ framework for international diversification is that for equities, international diversification results hold up for bond portfolios. see cholerton, pieraerts, and solnik (1986). in addition, levy and lerman (1988) and jorion (1989) demonstrated the benefits of international diversification across stocks and bonds of various coun tries, over the time period they studied. (6) generally, research has shown that correlation coefficients depend upon the time period chosen for historical analysis. see, for example shaked (1985) and jorion (1985). shaked’s study found that correlations among market returns are intertemporally unstable over a short-term invest ment horizon. however, the correlation structure was more stable over longer-term investment horizons. c. market integration and international diversification one of the most interesting, and heavily researched, ideas is the notion that correlation coefficients between countries’ market returns are increasing over time, as a result of the evolution toward integrated worldwide markets. the research of bertoneche (1979)) finnerty and schneeweis (1979)) hilliard (1979), maldonado and saunders (1981), shaked (1985), and others indicated that correlations are generally increasing over time. philippatos, christofi, and christofi (1983) and roll (1989) have viewed world markets as markets of single countries tied together with a single common “factor. ” in particular, roll’s study of the crash of 1987 demonstrated how all of the world’s stock markets fell significantly during the crash period. see also bennett 164 financial services review, l(2) 1991 and kelleher (1988). madura and mcdaniel (1989) found that correlations after the 1987 crash were generally even higher than before the crash. in related research in interest rates, kirchgassner and wolters (1987) analyzed the relationship among eurocurrency market rates over the 19741984 period. they found that a change in the eurodollar rate is followed somewhat by a similar change in the euro-deutschemark and euro-swiss franc rates. moreover, the relationship appears to be strengthening over time, presumably as a result of the evolution of global debt market linkages. see also kool and tatum (1988) and glick (1990). thus we have an interesting “catch-22” phenomenon. as the incentive to achieve international diversification leads to market globalization, the correlation between various countries’ equity and debt markets appears to be increasing, thus reducing the potential benefits of international diversification. nevertheless, markets may never be well-enough integrated, nor countries’ economic policies well-enough coordinated, to eliminate the case for international diversification. in addition to the imperfect correlations in debt markets reported by kirchgassner and wolters (1987), a study by cho, eun and senbet (1986), also found that equity markets were not yet well integrated despite the identification of some common world factors. moreover, studies by errunza and losq (1985) and jorion and schwartz (1986) confirm that some degree of segmentation of the markets continues to be present. thus, as markets tend to integrate, increases in correlation coefficients lessen the benefits of international diversification; however, international diversification is still regarded as a viable strategy, especially as markets behave in a volatile fashion. in light of this point, the study by eun and resnick (1984) comparing various methods of forecasting correlations, is still quite relevant. once individuals have diversified internationally, it may be in their interest to consider hedging the currency exposure of their international portfolios. iv. hedgingcurrency bets as noted by eun and resnick (1988), the original international diversification studies used data from a period of fixed or relatively stable exchange rates. due to the exchange rate volatility experienced in the 1980s researchers and practitioners have been to adopt the view that, while international diversification is still a viable strategy relative to purely domestic investing, investors may be able to improve on the performance of internationally diversified portfolios by attempting to eliminate some currency risk. currency exposure adds volatility to the portfolio without offering much in the way of compensating expected returns. perold and schulman (1989) regard currency exposure hedging to be a “free lunch, ” in that risk could be reduced, while expected rates of return are maintained. the empirical results of currency hedging in studies such as those by eun and resnick (1988), thomas (1988), and jorion (1989) lend support to the idea of currency hedging. international diversification for the individual: a review 165 a. the basic idea of portfolio currency hedging to help see the general idea behind currency hedging, let us suppose that an investor can sell forward the end-of-period foreign currency in his portfolio at the forward rate. actually, the idea of direct forward contracting might be unrealistic for an individual, since the global interbank forward market caters to corporate treasur ies and financial institutions. however, individuals can buy forward foreign currency by a simultaneous spot exchange and time-deposit. this simple transaction can be performed easily at citibank, which has run whole-page advertisements in the new york times for currency switching accounts. branches of a bank whose home country issues the currency would also routinely perform the transaction. to sell forward would involve a reverse transaction of borrowing the currency and spot exchanging it into the domestic currency. this loan may not be as easy as a deposit for an individual, but should be available as a brokerage service, as long as the rest of one’s portfolio is on account as collateral. if such an arrangement is not easily performed now, its simple concept and usefulness for internationally diversi fied investors should bring it about in the near future. an alternative transaction to a forward contract is a futures contract, which is the same concept, but involves some margin and “marking to the market. ” the international monetary market (imm) of the chicago mercantile exchange permits individuals to establish futures positions in one of several major currencies. the details of using futures to accomplish forward contracting are not within the scope of this paper: see smith, smithson and wilford (1990) for an introduction to currency futures. to follow the hedging ideas below, it will be sufficient to simply assume that the investor can either contract forward, or borrow/deposit the currency, at the “riskless interest rate” of that country, as a portfolio allocation choice. b. the “per$ect hedging ” strategy suppose the investor can contract to sell all his end-of-period currency expo sure forward, by borrowing today the present value of that exposure. this hedging strategy may not be possible in reality, since the investor does not know how many “ecus” the foreign allocation of his portfolio will be worth, since the local market return is a random variable. however, we will return to this issue shortly. temporar ily, it will be instructive to establish the performance of this strategy of “perfect hedging,” under the assumption that it is possible to do it. for simplicity, let us assume that the nominal risk free rate in europe is equal to the one in the u.s. thus the forward exchange rate can be assumed to be equal to the current spot rate. this simplified condition implies that, if the investor can “perfectly hedge” by contracting to sell all of the end-of-period foreign currency forward, then the returns in dollars are exactly the same as the local european returns (e). in addition to the means and standard deviations for u and e reported in 166 financial services review, l(2) 1991 table 1, the efficient frontier analysis needs the correlation between u and e, which the reader can verify is .25. figure 1 shows that the efficient frontier for the perfectly hedged (ph) strategy lies to the left of the one for the unhedged strategy. thus even though unhedged international diversification is preferred to a pure domestic investment strategy, the hedged international diversification strategy is better than the unhedged one. this is true even though the unhedged returns are less correlated with domestic market returns. (compare the correlation coefficient of. 1544 for the unhedged versus .25 for the hedged.) the fact that the perfectly hedged strategy reduces the volatility of the foreign investment is the reason for the gains in efficiency. it is worth noting that the composition of efficient portfolios is not the same for the unhedged and perfectly hedged strategies at various risk levels along the fron tiers. thus, it would not be possible to achieve optimal results by determining an “optimal” portfolio of international securities using unhedged ($e) return data, and then to hedge the portfolio components. instead, one must use hedged-return data (e) to find optimal portfolios. c. the “fully hedged” strategy as has already been noted, in reality, it may not be possible to hedge 100% of one’s local currency security returns. certainly, one cannot do this with standard forward contracts, since one does not know how much currency the foreign compon ent will represent at the end-of-period. the only way would be for the investor to create a forward contract that allows him to stipulate the contract size at the time of settlement. one feasible alternative strategy is to hedge 100% of the expected currency amount with standard forward contracting. for example, using the numerical returns from table 1, if you sell forward your initial investment plus the expected local currency return of 15 %, then the local currency return of 33 % in state 2 means that on an original investment of 100 ecus, 115 of end-of-period ecus would be hedged into u. s.$, but 33 15 = 18 would be subject to currency exposure. at the exchange rate change of .25 for that state, the overall dollar.return is .375. the formula for the rate of return of the “fully hedged” strategy is given in equation (2) below: rfll = (1 + w-j) (1 + f> + pi w-j) (1 + e) 1 (2) where f = the percentage difference between the forward exchange rate and the spot exchange rate, where both rates are expressed as domestic currency per foreign currency. in our example, the forward exchange rate is assumed to be equal to the spot exchange rate, and the $ returns for the other states from the “fully hedged” (fh) zntemational diversification for the zndividual: a review 167 strategy are .15 for state 1, . 102 for state 3, .2 1 for state 4, and .006 for state 5. the mean return for the fh strategy is .1686, the standard deviation is .1229, and the correlation with the u.s. market returns is .1855. the efficient frontier, considering the fh international investment strategy, is shown in figure 1 to the left of the two frontiers previously considered. note the significant finding that the fh strategy of forward hedging the expected currency value of the foreign components results in a more efficient set than one where an investor is assumed to be able to perfectly hedge 100% of all the currency exposure. d. optimal currency hedging the fact that the less-then-perfect fh hedging strategy actually outperforms the perfect hedging ph strategy implies that some amount of currency exposure may be beneficial and raises an important issue: is there some level of currency hedging that is optimal and are there situations where it would not be advisable to hedge currency risk? as eun and resnick (1988) point out, the results that show hedged portfolio results to be superior to unhedged results are due to the positive correlation between the local-currency returns and the value of the currency in dollar terms. if this correlation were not positive, would the perceived “free lunch” in currency hedging still hold up? (note also: if two stock markets are highly positively correl ated, it will not be possible for the exchange rate to be positively correlated with both!) lee (1987) has suggested a very general formulation for the international portfolio construction problem that considers separate components for securities and currencies. perhaps the best way to see lee’s idea is to regard foreign bills as a potential component of the portfolio. for simplicity, let us consider a numerical example which further assumes a zero risk free interest rate in the u.s. and europe. this simplification allows the $/ecu currency returns to be viewed as the returns from a nominally-riskless investment in european bills. a negative weight on this investment would mean a short sale. we must compute the correlation coefficient between the $/ecu and the u.s. market, which the reader can verify is .1241, and the correlation coefficient between the $/ecu and the unhedged u.s. returns on the european market ($e), which is .9698. the lefthandmost curve in figure 1 is labelled och (for optimal currency hedge) and shows the efficient frontier that is constructed by allowing the investor to invest in european stock and to hold a position in european bills in order to construct the optimal currency hedge for various levels of return and risk. it should be noted that the optimal currency hedge is not, in general, equal to hedging the initial investment plus the expected rate of return, and that the optimal currency hedge depends upon the risk level chosen and the parameter inputs. in other words, at a portfolio standard deviation of .ll, the amount of ecus the investor should sell forward, as a percent of investment into european stocks, is 168 financialservicesrevlew,1(2) 1991 different from the hedge proportion at a different standard deviation. depending upon the parameters and risk level, it could be that buying currency forward (a “texas hedge”) might be optimal. in applications with more markets and curren cies, the correlations between currencies also affect the optimal hedge, an effect that is known as “currency diversification. ” also see levy and sarnat (1978) for further discussion. one of the most provocative propositions in the area of international portfolio currency hedging is black’s (1989) “universal hedging” formula. black has shown that within the equilibrium conditions of a model of international portfolio investment with no cross-country investment barriers, the optimal currency hedge for an investor is neither perfect hedging (even if it could be accomplished) nor fully hedging. instead, in black’s model optimal currency hedging can be determined by a simple formula that depends upon (a) the expected return on the world market portfolio, (b) the volatilities of the various country markets, and (c) the volatilities of various exchange rates. the hedge proportion is “universal, ” in that it applies to all currencies, and to any investor in any country. black’s result is exciting in that it is a simple formula that by-passes the need for a lee-type efficiency analysis, as long as one buys into black’s assumptions. once again, the main assumption is equilibrium portfolio holdings by all investors in a model of international investments with no barriers. of course, the black result is new and controversial. its implications will no doubt be debated for years to come, as our understanding of optimal global invest ment and currency hedging evolves. of course, as black acknowledges, individuals who wish to make currency bets because they believe they have information that the equilibrium does not have, will be motivated to deviate from the “universal hedg ing” formula. indeed, individuals who believe they can forecast exchange rate movement better than “the market,” are likely to select option strategies as part of optimal portfolio allocations. like futures, exchange-traded options on some major curren cies are available to individuals. further elaboration is beyond our scope here; the interested reader can begin the study of currency options by referring to smith, smithson and wilford (1990), and of options in internationally diversified portfo lios by reading celebuski, hill, and kilgannon (1990). v. meansforindividualstoachieveinternational diversification technology is rapidly making it possible for an individual in one country to own specific foreign securities. for some large companies, global investment banking syndicates are helping the direct issue of securities in multiple countries, and in multiple currencies. for smaller companies and secondary market activity, an investor may already instruct his broker to buy a specific security in its home foreign market. to facilitate this process, global conventions for cross-border international diversification for the individual: a review 169 clearance of trades are becoming standardized, with much progress made by the g 30, the international private-sector “group of 30” bankers, investors, regulators and officials concerned with the basic mechanisms underlying the international financial system. for a discussion of the globalization of financial markets, see pave1 and mcelravey (1990)) and for an overview of policies recommended by the g-30, see degennaro and pike (1990). however, at present, the most efficient means for individuals to achieve international diversification are american depository receipts (adrs) and inter national mutual funds. in addition to reviewing some research results on these two media, we also review some research that suggests that investment in the stocks of multinational companies is not an effective means of achieving international diversi fication. a. american depository receipts american depository receipts (adrs) are certificates representing owner ship of foreign stocks. an adr typically represents 1 to 10 shares of the underlying stock. some adrs are traded on exchanges, while the vast majority are traded over the-counter. dividends on stocks represented by adrs are received by a depository bank and are transferred to investors holding adrs. banks charge fees for transactions involving the payment of dividends or the exchange of adrs for the underlying shares. some of the well-known firms for which adrs are available include porsche (germany), phillips lamp (netherlands), hachette (france), and jardine mathe son (hong kong). due to global arbitrage strategies in stocks and currencies, the price of an adr in dollars will approximately reflect the “unhedged” investment into the stock in its own local currency. see rosenthal (1983). companies represented by adrs are required by the securities and exchange commission (sec) to file financial statements consistent with the generally accepted accounting principles in the u.s. therefore, the financial information on such foreign companies is compatible with information on u.s. companies. however, some reporting rules are looser for the foreign firms. for example, foreign firms only need to provide financial reports to shareholders semi-annually and do not have to disclose salaries of top management. in addition, foreign firms can issue non-voting stock. officer and hoffmeister (1987) found that adr returns were more volatile than u.s. stock returns, no doubt due to the added uncertainty of exchange rates implicit in adr prices, but combined portfolios of adrs and u.s. stocks exhibited significantly lower variance than portfolios solely composed of u.s. stocks. these results are consistent with those of international diversification studies. thus adrs could effectively enable u.s. investors to reduce risk. related research by tucker (1987) found that a u.s. portfolio achieves marginal diversification benefits from adding adrs similar to those from adding 170 financial services review, l(2) 1991 foreign stocks. thus, adrs may be an adequate substitute for direct investment in foreign stocks. however, the limited number of adrs available and the costs of adr transactions may encourage some individual investors to consider alternative means for international diversification. b. international mutual funds international mutual funds may be the best way for the individual to diversify internationally. investors can purchase shares of such a fund with a small minimum investment, such as $1,000. investment management companies often offer a selec tion of open-end international mutual funds. some offer managed portfolios, while others are “index” funds. some will offer worldwide diversification (with or without u.s. stocks), while others offer regional or country-specific investment. several studies have demonstrated that u.s. investment in a foreign stock portfolio representing various countries exhibits less risk than u.s. investment in a foreign stock portfolio representing a single country. some u.s. investors have taken this to mean that an international mutual fund is less risky than a domestic mutual fund. such investors may be surprised to find that domestic funds typically exhibit less variance. foreign stock holdings will normally reduce a u.s. investor’s risk only if the investor continues to maintain some u.s. stocks within the portfolio. individual investors can more easily make investments that “mirror” the market of a single country when they purchase international mutual funds. there are also numerous funds invested in a single country and exchange-traded on a closed-end basis. by investing in several of these single-country funds, individual investors can create a well diversified portfolio, especially of stocks of countries whose markets are less well-developed. with as little as $30,000 and low transac tions costs, they may be able to invest in over one thousand stocks from more than ten foreign countries. essayyad and wu (1987) assessed the diversification attributes of i8 interna tional mutual funds over the 1977-1984 period. fifteen of the funds exhibited a higher mean return than the s&p 500 index. in addition, 16 of the funds exhibited a lower coefficient of variation than the s&p 500 index. essayyed and wu also found that the average percentage of variation in fund returns explained by s&p 500 movements was only about 24 % . in a related study, rao and aggarwal (1987) examined international mutual fund returns’ sensitivity to the s&p 500 index. they found that the funds’ estimated betas were less than 1 .oo and that on average only 30 % of the variation in each fund’s returns could be explained by market movements. this is significantly below the average explained variation for domestic mutual funds with similar regression applications. thus, since the international mutual fund returns were not driven by the u.s. market, such funds appear to be viable means of achieving international diversification. a recent study by essayad, madura, and marx (199 1) assessed the diversifica znternational diversification for the individual: a review 171 tion potential across international mutual funds. since many of the funds are concentrated in a particular region, they do not reflect fully diversified portfolios across the world. thus, there may be some additional benefits to be gained by investing in a portfolio of international funds. the researchers found that on average a portfolio of two funds contains 34 % less risk than a single international fund. they also found additional risk reduction as more funds were added to the portfolio. a portfolio of eight funds contained 59 % less risk than a single international fund, on average. after that point, adding more funds to the portfolio had a negligible effect. thus it appears that on average about eight international funds are needed to achieve complete global diversification. the precise number will vary with the type of funds considered. if funds focusing on a single country are used, more funds would no doubt be necessary to achieve complete global diversification. cumby and glen (1990) found that over the 19821988 period, international mutual funds did not provide superior performance relative to a broad international index. however, individual investors may still prefer international funds as the most efficient means of investing globally. c. investment in multinational corporations a multinational corporation (mnc) operates in more than one country and can be thought of as a portfolio of numerous smaller firms (subsidiaries) spread around the world. mncs should be somewhat insulated from their respective home markets, because a substantial portion of their operations are in other countries. while the stock of a u.s. -based mnc is not “international, ” it could possibly serve as an adequate substitute for an international stock portfolio. because they are easy for the individual to invest in, mncs may appear to be an appealing means of diversifying internationally. jacquillat and solnik (1978) tested whether mnc stocks are reasonable substitutes for foreign stocks. if mnc stocks behave like an international portfolio, then they should be sensitive to the stock markets of the various countries in which they operate. jacquillat and solnik applied a multiple regression model to 1966 1974 returns to assess the sensitivity of mnc returns to various stock markets. they used portfolio returns of mncs from the u.s. as the dependent variable, and returns of each national market as their independent variables; the regression coefficients represent the sensitivity of mnc returns to each national stock market. based on this analysis, jacquillat and solnik found that mncs based in the u.s. were typically affected only by the u.s. stock market and not by other stock market movements. they replicated the analysis for mncs based in other countries and typically found similar results. that is, an mnc portfolio’s returns are only sensitive to its respective local domestic stock market. this finding implies that mncs behave like local domestic stocks and are not good substitutes for foreign stocks. jacquillat and solnik also ran a complementary test, in which a simple regres 172 financial services review, l(2) 1991 sion model was applied to each mnc portfolio using the local stock market returns as the independent variable, and the explanatory power of this model was compared to that of the multiple regression model. the multiple regression model exhibited very little additional explanatory power (based on a comparison of the adjusted coefficients of determination.) this finding reinforces the conclusion that investing in a portfolio of mncs does not sufficiently achieve international diversification. madura (199 1) replicated the study by jacquillat and solnik with data from the 1974-1987 period. to the extent that mncs became more global in scope since the 1966-1974 time period of the jacquillat and solnik study, mncs might better have served as viable substitutes for foreign firms in the more recent period. madura found that even in this more recent period, mnc stock returns are sensitive only to the local stock market returns and not to the movements of any other markets. thus mncs continue to be poor substitutes for foreign stocks, even as the mncs evolve into the global, “stateless” corporations. some studies have assessed the performance of mncs by comparing them to purely domestic corporations (dcs). brewer (1982) compared both risk and return aspects of mnc and dc stocks by deriving separate security market lines. he found no statistically significant difference between the two security market lines, and therefore concluded that mncs do not offer any advantage over dcs. michel and shaked (1986) used the sharpe and treynor indices to compare the measures of return and risk for dcs and mncs. both indices were higher for dcs than for mncs. senchak and beedles (1980) compared risk reduction capabilities between portfolios of mnc stocks and portfolios of dc stocks. they also measured the degree of risk reduction resulting from an increased number of stocks for both types of portfolios. they found that the degree of risk reduction in portfolios of dcs exceeded that exhibited by portfolios of mncs. defusco, philippatos, and choi (1990) applied factor analysis to analyze the factor structures between mncs and dcs, and found no significant factor structure differences between the two types of firms. these results offer further evidence that mnc share prices are driven by the same process as dcs. overall, the research suggests that mncs are not a sufficient means for effective international diversification. vi. conclusion this paper has reviewed some aspects of the literature on international invest ing that should be of interest to individual investors. the main focus has been on the international diversification strategy. three modern issues were covered: (1) the benefits of international diversification as the global markets continue to integrate; (2) the problem of currency exposure; and (3) effective means of achieving interna tional diversification. by restricting the scope of the review to issues of most interest to the individual, we do not review interesting, but more general research on international diversification for the individual: a review 173 international asset pricing theory and international market efficiency and “anoma lies. ” as a result of the research, one can conclude that individual investors can (1) benefit from international diversification, in effect increasing the effi ciency of their portfolios, but as over time the world market continues to integrate, the benefits may decline, (2) reduce their currency exposure and improve on the performance of an internationally diversified portfolio by employing various hedging strat egies, and (3) achieve international diversification by purchasing adrs and interna tional mutual funds, and not so much by investing in mncs. references bennett, paul, and jeanette kelleher. 1988. “the international transmission of stock price disruption in october 1987,” frbny quarrerly review, summer: 17-33. bertoneche, marc l. 1979. “an empirical analysis of the interrelationship among equity markets under changing exchange rate systems, ” journal of banking and finance, december: 397 405. biger, nahum. 1979. “exchange risk implications of international portfolio diversification,” journal of international business studies, fall: 64-74. black, fischer. 1989. “universal hedging: optimizing currency risk and reward in international equity portfolios,” financial analysts journal, july/august: 16-22. brewer, j.l. 1982. “investor benefits from corporate international diversification,” journal of financial and quantitative analysis, march: 113-125. celebuski, matthew j., joanne m. hill, and john j. kilgannon. 1990. “managing currency exposures in international portfolios,” financial analysts journal, january/february: 16-23. cholerton, kenneth, pierre pieraerts, and bruno solnik. 1986. “why invest in foreign currency bonds?” journal of portfolio management, summer: pp. 4-8. cumby, robert e., and jack d. glen. 1990. “evaluating the performance of international mutual funds,” journal of finance, june: 497-521. defusco, richard, george c. philippatos, and dosoung choi. 1990. “differences in factor structures between u.s. multinational and domestic corporations: evidence from bilinear paradigm tests,” financial review, august: 395404. degennaro, ramon p., and christopher j. pike. 1990. “standardizing the world securities clearance systems,” federal reserve bank of cleveland, economic commentary, april: l-4. errunza, vihang r. 1977. “gains from portfolio diversification into less developed countries securities,” journal of international business studies, fall/winter: 83-99. errunza, vihang, and etienne losq. 1985. “international asset pricing under mild segmentation: theory and test,” journal of finance, march: 105-124. essayyad, musa, jeff madura, and donald m. marx. 1991. “diversifying among international mutual funds: how many are enough?” unpublished paper. essayyad, musa, and h.k. wu. 1988. “the performance of u.s. international mutual funds,” quarterly journal of business and economics, autumn: 32-46. eun, cheol s., and bruce g. resnick. 1984. “estimating the correlation structure of international share prices,” journal of finance, december: 131 i-1324. 174 financial services review, l(2) 1991 eun, cheol, and bruce resnick. 1988. “exchange rate uncertainty, forward contracts, and international portfolio selection,” journal of finance, march: 197-215. finnerty, joseph e., and thomas schneeweis. 1979. “the comovement of international asset returns,” journal of international business studies, winter: 66-78. click, reuven. 1990. “global interest rate linkages,” federal reserve bank of san francisco, weekly letter, may 25: l-3. grauer, robert r., and nils h. hakansson. 1987. “gains from international diversification: 1968 85 returns on portfolios of stocks and bonds, ” journal of finance, july: 72 l-74 1. grubel, herbert g. 1968. “internationally diversified portfolios: welfare gains and capital flows,” american economic review, december: 1299-1314. grubel, herbert g., and kenneth fadner. 1971. “the interdependence of international equity markets,” journal of finance, march: 89-94. hilliard, jimmy e. 1979. “the relationship between equity indices on world exchanges,” journal offinance, march: 103-l 14. ibbotson, roger g., richard c. carr, and anthony w. robertson. 1982. “international equity and bond returns,” financial analysts journal, july/august: 61-83. jacquillat, bertrand, and bruno solnik. 1978. “multinationals are poor tools for diversification,” journal of portfolio management, winter: 8-12. jorion, phillippe. 1989. “asset allocation with hedged and unhedged foreign stocks and bonds,” journal of portfolio management, summer: 49-54. jorion, phillippe. 1985. “international portfolio diversification with estimation risk,” journal of business, july: 259-278. jorion, phillippe, and edward0 schwartz. 1986. “integration vs. segmentation in the canadian stock market,” journal of finance, july: 602-616. kirchgassner, gebhard, and jurgen wolters. 1987. “u.s.-european interest rate linkage: a time series analysis for west germany, switzerland, and the united states,” review of economics and statistics, november: 675-684. kool, clemens, j.m. and john a. tatom. 1988. “international linkages in the term structure of interest rates,” federal reserve bank of st. louis, review, july/august: 30-43. lee, adrian. 1987. “international asset and currency allocation,” journal of portfolio management, fall: 68-73. lessard, donald r. 1976. “international diversification,” financial analysts journal, january/ february: 32-38. levy, haim, and zvi lerman. 1988. “the benefits of international diversification in bonds,” financial analysts journal, september/october: 56-64. levy, haim, and marshall sarnat. 1978. “exchange rate risk and the optimal diversification of foreign currency holdings,” journal of money, credit, and banking, november: 453-463. levy, haim, and marshall sarnat. 1970. “international diversification of investment portfolios,” american economic review, september: 668-675. madura, jeff. 1991. “influence of foreign markets on multinational stocks; implications for investors,” review of international business and economics, forthcoming. madura, jeff, and william r. mcdaniel. 1989. “impact of the 1987 crash on gains from international diversification,” journal of international finance, fall: 23-35. maldonado, rita, and anthony saunders. 1981. “international portfolio diversification and the inter-temporal stability of international stock market relationships, 1957-78,” financial management, autumn: 54-63. mcdonald, john g. 1973. “french mutual fund performance: evaluation of internationally diversified portfolios,” journal of finance, december: 1161-l 180. michel, allen, and israel shaked. 1986. “multinational corporations vs. domestic corporations: financial performance and characteristics,” journal of international business studies, fall: 89-100. international diversification for the individual: a review 175 officer, dennis t., and j. ronald hoffmeister. 1987. “adrs: a substitute for the real thing?” journal of portfolio management, winter: 61-65. pavel, christine, and john n. mcelravey. 1990. “globalization in the financial services industry,” federal reserve bank of chicago, economic perspectives, may/june: 3-18. perold, andre f., and evan c. schulman. 1988. “the free lunch in currency hedging: implications for investment policy and performance standards,” financial analysts journal, may/june: 45 50. philippatos, g.c., a. christofi, and p. christofi. 1983. “the inter-temporal stability of international stock market relationships: another view,” financial management, winter: 63-69. rao, ramesh p, and raj aggarwal. 1987. “performance of u.s. based international mutual funds,” akron business and economic review, winter: 98-106. roll, richard. 1988. “the international crash of october 1987,” financial analysts journal, october: 19-35. rosenthal, leonard. 1983. “an empirical test of the efficiency of the adr market,” journal of banking and finance, pp. 17-29. senchak, andrew j., jr., and w.l. beedles. 1980. “is indirect international diversification desirable?” journal of portfolio management, winter: 49-57. shaked, israel. 1985. “international equity markets and the investment horizon,” journal of portfolio management, winter: 80-84. smith, clifford w, jr., charles w. smithson, and d. sykes wilford. 1990. managing financial risk. new york: harper and row. solnik, bruno. 1974. “why not diversify internationally?” financial analysts journal, july/ august: 48-54. solnik, bruno, and bernard noetzlin. 1982. “optimal international asset location,” journal of portfolio management, fall: 1 l-2 1. thomas, lee r., iii. 1988. ‘currency risks in international equity portfolios,” financial analysts journal, march/april: 68-70. tucker, alan l. 1987. “international investing: are adrs an alternative?” aaii journal, november: 10-12. from the editor this issue contains volume 27 issue 3 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “describing investor profiles: a test of the associations among financial knowledge, confidence, and help and information sources” is coauthored by abed rabbani at university of missouri, john e. grable, ann woodyard, and zheying yao, at university of georgia. in this study, the authors use a large, nationally-drawn dataset of individuals who own financial assets to explore the relationships between and among types of investments owned, knowledge characteristics, investor confidence, and help and information sources. the authors found that investors who exhibited over-confidence in their financial knowledge were more likely to hold annuities, cash value life insurance, and commodities. they also found that financial planners play an important role in promoting diversification and mitigating portfolio risk. the second article “advisor compensation: which clients know and how do they pay?” is coauthored by somer g. anderson at, maryville university, martin c. seay at kansas state university, kyoung tae kim at university of alabama, and derek r. lawsond at kansas state university. the authors study agency theory using the 2015 national financial capability study investor survey to investigate characteristics associated with knowing how one’s financial advisor/broker is compensated. the study examines how individuals who do know the compensation method choose between financial advisors with different compensation models. their results indicate that clients who place importance on fees that are more knowledgeable about diversification, and perform background checks are more likely to know compensation methods. mixed results show that clients may not fully understand the compensation paid to their advisors. the third article, “improving long-term portfolio risk and return by using appreciated stocks for charitable donations” is authored by jeff whitworth at university of houstonclear lake. the author studies how stock investors who are charitable donors can minimize capital gains taxes and improve portfolio diversification by donating their most appreciated shares instead of cash, and then reinvesting the freed-up cash in the portfolio’s least-weighted stocks. in theory the charity is indifferent to the donation method, and the investor receives financial services review 27 (2018) v–vi 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. the same charitable deduction. using monte carlo simulations the authors shows that a donor-investor using this method results in higher wealth and lower portfolio risk, particularly over longer time horizons. he concludes that combining this strategy with tax loss selling and some limited recognition of capital gains harvesting further increases after-tax returns and reduces risk. the fourth article, “factors related to the risk tolerance of households in china and the united states: implications for the future of financial markets in china,” is coauthored by sherman d. hanna at ohio state university, kyoung tae kim at university of alabama, and lishu zhang at shenzhen university. the authors analyze factors related to the financial risk tolerance of chinese households, using the 2011 china household finance survey (chfs). the risk tolerance question was similar to one in the u.s. survey of consumer finances (scf), and the authors found that chfs respondents had slightly higher risk tolerance than scf respondents, but the percent of households with stock assets was 9%, compared to 49% in the u.s. their multivariate analyses found many household characteristics in the chfs had effects on risk tolerance similar to those found in the 2013 scf. the final article, “perspectives on “sell in may and go away a look at recent evidence and implications” is authored by tony loviscek at seton hall university. the author studies three perspectives on the quote “sell in may and go away.” he tests the annual performance of switching from four equity mutual funds to u.s. treasury bills against that of the buy-and-hold strategy, he examines the switching strategy during the two bear and two bull markets occurring from 2000 to 2016, and finally he tests the impact of taxes on the switching strategy. his findings show that although there are signs of switching strategy effectiveness, the results lack the statistical significance to conclude that it is superior to the buy-and-hold strategy. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review vi editorial / financial services review 27 (2018) v–vi from the editor this issue contains volume 30 issue 1 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing the implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning topics and issues. the lead article “retirement income beliefs and financial advice seeking behaviors” is coauthored by alejandro murguı́a at mclean asset management and wade d. pfau at mclean asset management and the american college of financial services. this paper investigates a series of salient behavioral finance and psychological constructs that influence retirement income planning. the authors show how these scales are related to each other as well as retirement income concerns and investment behaviors. they also describe how four investment personas can be linked with the advisor usefulness and retirement income self-efficacy scales to identify preferred financial implementation methods. this can assist individuals in recognizing their relative strengths and weaknesses while financial professionals can present advice in a manner that addresses a client’s concerns and preferred implementation. the second article “mutual fund knowledge assessment for policy and decision problems” is coauthored by brian scholl at the office of the investor advocate and angela fontes at norc at the university of chicago. in this paper the authors develop a measure of mutual fund investment knowledge that complements existing financial literacy measures. they validate the index with factor analysis identifying two latent components, and descriptive regressions demonstrating the additive value of our index beyond general financial literacy in explaining variation in financial well-being, investment ownership, and fee calculation proficiency. despite mutual funds’ importance in household savings, their index suggests that the public lacks adequate understanding of mutual funds. the third article, “do as i tell you, not as i do: financial advisors and personal financial decision-making” is coauthored by negin azamian, kristine beck, hsin-hui chiu, and inga timmerman, all at california state university northridge. the authors describe the financial behavior of financial advisors and whether they follow the advice they give clients. in doing so they focus on the following areas of comprehensive financial planning as they relate to advisor behavior: (1) cash flow, (2) debt, (3) retirement planning, (4) 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) v–vi investments, and (5) estate planning. the authors find that financial advisors generally follow their own advice; and as a group they are more likely to be prepared for retirement, have less debt, higher liquidity, covered insurance needs, and have an estate plan in place. the final article, “financial advisor use, life events, and the relationship with beneficial intentions” is coauthored by matt sommer at janus henderson investors, hanna lim at kansas state university, and maurice macdonald at kansas state university. this study investigates whether working with a financial advisor and experiencing a recent life event were associated with having beneficial financial planning intentions. the authors found no relationship between working with a financial advisor and beneficial intentions over the next 12 months. life events incurred within the prior year, however, were positively related to beneficial intentions and when interacted with working with an advisor, had a positive moderating effect. the results suggest that planning for difficult life transitions is an important benefit of working with a financial advisor. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submitting to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review s. michelson / financial services review 30 (2022) v–vi vi are multiple share class funds poorly governed? jonathan handya,*, thomas smytheb adepartment of business and accounting, furman university, 3300 poinsett highway, greenville, sc 29613, usa blutgert college of business, florida gulf coast university, 10501 fgcu boulevard south, fort myers, fl 33965-6565, usa abstract utilizing independent morningstar stewardship grades, this article finds that multiple share class mutual funds (ms funds) have lower quality governance. ordered probit regressions indicate ms funds are more likely to have lower board quality ratings and managerial incentive ratings, additional evidence the ms structure has not provided the benefits initially put forth by supporters. the results continue to demonstrate that less sophisticated investors seeking financial advice (those typically utilizing ms funds) may potentially be directed to funds that underperform and have higher costs. © 2020 academy of financial services. all rights reserved. jel classification: g180; g230; g340 keywords: mutual fund; sec; corporate governance; shareholders; investor protection 1. introduction with the 1995 establishment of the securities and exchange commission rule 18f-3 (securities and exchange commission, 1995), came the widespread use of multiple share class mutual funds (ms funds). ms funds are designed to have a single investment portfolio with a variety of combinations of commission structures and ongoing expense structures. individuals choose a particular combination from those available, the class, and invest in the fund. o’neal (1999) points out that such funds are complex and that choosing the most advantageous combination of fund and class may be difficult for investors. in most situations * corresponding author. tel.: �1-864-294-3139. e-mail address: jonathan.handy@furman.edu (j. handy) financial services review 28 (2020) 49-65 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. where ms funds are utilized, the investment is made with the aid of a financial advisor/ broker. however, research shows there is confusion among advisors as to the most appropriate ms fund class based on expected investor holding period (jones, lesseig, and smythe, 2005a). moreover, recent research finds that ms funds have higher expense ratios than single-class (non-ms) funds when taking fund governance quality into consideration (handy, nichols, and smythe, 2018). because fund expenses are directly correlated with investor returns, one might expect advisors/brokers to place great emphasis on them when recommending funds to clients, but research by jones, lesseig, and smythe (2005b) suggests that they rarely do. because previous research ties fund expenses to governance quality, and prior results indicate that advisors/brokers may not use fund expenses as a major selection criteria, this article investigates the relative governance quality of ms and non-ms funds so that investors and advisors/brokers are better informed regarding their investment selections. while motivated by o’neal (1999), which explores differences across fund classes, this article explores differences across fund structure. specifically, this article explores a more subtle issue in the mutual fund market: whether ms funds have differing levels of governance quality compared with non-ms funds. to do so, we use independent governance ratings data from morningstar. beginning in 2004, morningstar began publishing stewardship grades for mutual funds based on five criteria: board quality, managerial incentives, fees, corporate culture (of the fund sponsor), and regulatory ratings. each grade provides information on fund management in the context of governance and administrative functionality, and as stated by morningstar, “helps investors to assess funds based on the degree to which the funds’ parent—the management company offering the fund—has its interests aligned with those of fund shareholders.” over the last 10 years, a growing stream of research examines whether the stewardship grade, or its components, have any relationship to fund performance and/or fund costs. to date, this is the first article using stewardship grades to analyze whether there are differences in fund governance across ms and non-ms funds. our analysis utilizes components of the morningstar stewardship grade (msg) to gauge the degree of governance alignment between retail investors and fund management companies. more specifically, we analyze whether funds selected by investors choosing to use financial advisors have systematically different governance ratings from funds offered to retail investors not using financial advisors.1 to be clear, we are not suggesting financial advisors consciously recommend funds with differential governance; however, the primary pool of funds from which financial advisors make recommendations does consist of ms funds. in fact, for advisors who work for large securities firms, the pool of funds from which they choose is preselected at the firm level, and so advisors may be forced to suggest such funds. ultimately, it follows that if ms funds are generally more likely to have lower governance ratings, then retail investors utilizing advisors are more likely to be steered towards funds with lower governance quality and need to know this. our findings identify significant differences in morningstar governance ratings across fund structures. ms funds are more likely to have lower board quality ratings and lower managerial incentive ratings than non-ms funds (traditional no-load funds). examining predicted probabilities, the results indicate that ms funds have a significantly lower probability of having high governance ratings when compared with non-ms funds. additionally, 50 j. handy, t. smythe / financial services review 28 (2020) 49-65 a fund sponsor’s corporate culture, as rated by morningstar, has predictable independent influences on fund board quality and managerial incentives. finally, funds with better board quality ratings have higher managerial incentive ratings. our findings provide additional evidence that the ms structure is associated with largely detrimental effects on a less informed group of investors. the balance of the article proceeds as follows. in section 2, we review mutual fund literature to place this work in context. in section 3, we develop our hypotheses. in section 4, we introduce the data, build the empirical model, and discuss the variables of interest. in section 5, we present primary empirical results, and in section 6, we present our concluding remarks. 2. previous literature over the years, the use of mutual funds for investing has increased tremendously. fund assets increased from $11.1 trillion at year-end 2009 to $17.7 trillion at year-end 2018 (2010 and 2019 investment company fact books, respectively). accordingly, mutual fund research remains a growing specialty. for brevity, our literature review focuses on the three areas of fund research most related to this article: research focusing on the nature of multiple-share class funds, research focusing on the intersection of retail investors and financial advisors, and research focusing on morningstar stewardship grades (msg). 2.1. multi-share class mutual funds despite the fact that the sec has allowed mutual fund sponsors to offer ms funds since 1995, little analytical research into ms funds and the ms structure exists. this article, to the best of its authors’ knowledge, is the first to analyze whether there are differences in independently rated governance quality between ms and non-ms funds. the main theoretical multi-share class research comes from livingston and o’neal (1998), o’neal (1999), and nanda, wang, and zheng (2009). lesseig, long, and smythe (2002) and handy, nichols, and smythe (2018) provide empirical results. livingston and o’neal (1998) and o’neal (1999) focus on ms fund costs and the incentives provided to investors and brokers.2 livingston and o’neal (1998) concludes that, because little evidence supports mutual fund performance persistence, investors should select funds based on costs because higher costs uniformly lead to lower returns. they identify the most common ms fund cost distribution types, derive a series of mathematical equations expressing the costs as a present value, and provide a comprehensive set of optimal investment strategies for investors given specific investment time horizons. however, when considering the results of barber, odean, and zheng (2005), which demonstrates investor confusion with fund costs, the challenge facing ms fund investors becomes clear: investors must not only choose which fund to invest their assets but also which combination of commissions and ongoing expenses best suits their needs (i.e., they must choose the right class). 51j. handy, t. smythe / financial services review 28 (2020) 49-65 o’neal (1999) derives commission-based incentives for fund brokers/advisors and finds conflicts of interest between advisors and investors. o’neal (1999) indicates that this is particularly dangerous to ms fund investors, given that those most likely to seek out advisors are those who are relatively uninformed. nanda et al. (2009) largely focus their analysis around a fund’s decision to switch from single-class to the multiple-class structure and find that switching to the ms structure negatively impacts performance. while similar to o’neal (1999), this article analyzes the relationship between fund sponsors and investors by examining morningstar’s independent board quality and managerial incentive ratings to determine if there are differences between ms and non-ms funds. finally, lesseig, long, and smythe (2002) and handy, nichols, and smythe (2018) also focus on ms structure by examining differences in expense ratios between ms and non-ms funds. lesseig et al. (2002) analyzes the claim made by fund sponsors when the ms structure was introduced that it allows funds to decrease fund expenses. their results suggest the opposite—overall net expense ratios for ms funds are higher than for non-ms funds. handy et al. (2018) examines a longer and more recent sample and find results consistent with lesseig et al. (2002). 2.2. retail investors and the financial advisor while literature focusing on the ms structure is scarce, literature focusing on the intersection between financial advisors, mutual funds, and the retail investor is abundant. nofsinger and varma (2007) analyzes survey results and find that financial advisors are on average more analytical than the general population, that they are more financially patient, and that they perform better in intertemporal choice problems. bergstresser, chalmers, and tufano (2009) analyzes broker-sold (ms) and direct-sold (non-ms) funds from 1996 to 2004 and do not find any evidence brokers offer substantial benefits to clients.3 additionally, bergstresser et al. (2009) finds that broker-sold funds are no more skilled at aggregate-level asset allocation than funds sold through the direct channel. while early literature focuses on how individual investors make fund investment decisions (e.g., alexander, jones, and nigro, 1998; capon, fitsimmons, and prince, 1996), jones et al. (2005a) surveys over 500 financial advisors on what criteria and information sources they use in the fund recommendation process. they find that the two most important information types used are comprehensive data sources and independent rankings from firms such as morningstar and lipper (now part of thomson-reuters). however, they also find that advisors rank fund costs very low as recommendation criteria. jones et al. (2005b) also examines survey data from financial advisors regarding their compensation and investment recommendations as it relates specifically to ms funds. they find advisors are more likely to recommend a specific ms class based on the commission received rather than the appropriateness of the class for the client. additionally, when the funds are firm proprietary funds, advisors are more likely to recommend the class most profitable for the firm, usually to the detriment of investors. the results from jones et al. (2005b) are consistent with the cautions presented by o’neal (1999). 52 j. handy, t. smythe / financial services review 28 (2020) 49-65 2.3. morningstar stewardship grades msgs were first introduced in 2004 to “help investors further research, identify, and compare fund managers and fund companies that do a good job—or poor job—of aligning their interests with those of fund shareholders” (fact sheet, 2006). in short, the ratings are designed to help investors and advisors evaluate a fund’s effectiveness at mitigating the principal-agent problem between investors and fund management. this article brings attention to the grades as an empirical tool, but more importantly, it examines whether there are differences in the governance ratings across fund structure, ms versus non-ms funds. recent work analyzing msg ratings includes moore and porter (2017). they analyze a 2007 cross-section of funds and report that increased mutual fund governance quality, as measured by morningstar ratings, lead to lower fund expenses. cao, ghosh, goh, and ng (2014) establishes that msgs have granger causality on long-term risk adjusted returns and can offer an explanation for fund performance, even when morningstar star ratings are considered. chou, ng, and wang (2011) finds that firms with better governance practices, as measured by morningstar, tend to vote more responsibly on corporate governance proposals of portfolio firms and generally provide better return performance. this article, as it relates to msg ratings, is most similar to work by handy et al. (2018), which examines whether morningstar’s board quality and managerial incentive scores are correlated with fund net expense ratios. handy et al. (2018) argues that investors should seek to minimize fund expenses and analyze msg ratings and their relationship to fund expenses as a potential tool for investors to gauge a fund’s attractiveness. of particular interest to the current article is that handy et al. (2018) find that the relationships between msg ratings and fund expenses differ between ms funds and non-ms funds. this article should be considered a more general extension of their work. rather than focus on msg ratings in the context of fund expenses, this article looks at the more general question of whether governance ratings differ across distribution channels. given the empirical results cited above, the importance of this article should be clear. if lower governance ratings are associated with higher expenses and thereby lower returns, then investors investing in funds with such ratings are being negatively impacted and should be made aware. consequently, exploring whether or not ms funds have better or worse governance ratings than non-ms funds is a valid pursuit. 3. hypothesis development ms funds are targeted primarily to more vulnerable investors, suggesting the need for strong fund-level governance.4 thus, our focus is on whether ms funds and non-ms funds have differences in governance quality as reflected by msg ratings. when ms funds were introduced, the sec was concerned about inequitable treatment of shareholders across fund classes. however, handy et al. (2018) shows that ms funds have higher net expenses than non-ms funds and that governance measures have differential effects on expenses across fund structure. as such, our analysis is at the fund level and focuses on differences across fund structure. 53j. handy, t. smythe / financial services review 28 (2020) 49-65 the board qual rate and manager incent rate variables are evaluated independently. the variables are measured on a scale of 1 (lowest) to 5 (highest). as promoted by morningstar, each should provide information to investors/financial advisors about the relative quality of fund governance along these dimensions, each of which is important in mitigating the principal-agent problem between fund sponsors and investors. when estimating the empirical model for each dependent variable, we include the dummy variable, ms, equal to 1 if the observation is an ms fund and 0 otherwise. given that the primary investors in ms funds are considered less knowledgeable, the funds may take additional steps to promote good governance practices. if so, we expect ms to have a direct relationship with the governance metrics board qual rate and manager incent rate. as such, hypothesis 1(a): ms funds have higher board quality ratings (board qual rate) than non-ms funds. hypothesis 1(b): ms funds have higher managerial incentive ratings (manager incent rate) than non-ms funds. while examining differences in governance ratings across fund structure is our primary focus, we also are interested in whether the corporate culture of the fund the fund sponsor influences fund governance. as such, we include morningstar’s corp culture rate in the empirical models (a variable ranging from 1 [lowest] to 5 [highest]). corp culture rate is meant to “assess how seriously a firm takes its fiduciary duty to its fund shareholders.” it is an indirect measure of how fund sponsors may influence the governance process within funds they operate. we expect more highly rated fund sponsors to have boards that are of higher quality and stronger managerial incentives. as such, hypothesis 2(a): funds whose sponsor has a higher corporate culture rating (corp culture rate) have a higher board quality rating (board qual rate). hypothesis 2(b): funds whose sponsor has a higher corporate culture rating (corp culture rate) have a higher managerial incentive rating (manager incent rate). finally, once a board is in place, it has the authority to influence contracts between fund managers and the fund as it pertains to managerial incentives. while the board has sole authority to negotiate fund expenses, it is also likely that the board will have an influence on how much fund managers must own to align the interests between the two groups. therefore, we expect funds with more highly rated boards to have higher managerial incentive ratings. hypothesis 3: funds with higher board quality ratings (board qual rate) have higher managerial incentive ratings (manager incent rate). 4. data and empirical model 4.1. data the data for the analysis comes from morningstar and includes year-end data from 2005 to 2009.5 our sample only includes funds in the investment objectives growth and income, 54 j. handy, t. smythe / financial services review 28 (2020) 49-65 growth, aggressive growth, and small cap for two reasons. first, early mutual fund literature commonly examined these investment categories. second, and more importantly for this article, the identification of classes in the same fund portfolio had to be identified and coded by hand across all years in the sample, a time consuming process. morningstar observations are often referred to as “a fund,” but they are not. morningstar captures data at the class level, reflecting differences across share classes that a ms fund has. handy et al. (2018) describes the issues, both practical and statistical, of conducting analysis with class level data. that article analyzes class level net expense ratios, prompting them to conduct their initial analysis at the class level. however, handy et al., also introduce a robustness technique, whereby they examine data at the fund level by identifying all classes of a ms fund by hand. as they discuss, some variables are representative of the class, for example, commission structure (front-end load or redemption fee) and class level net assets, while others, such as ms, board quality rating, and turnover are unique to the fund, that is, is the same for all classes. we analyze data at the fund level, using the stata collapse command to create a single observation for each unique fund in the sample. the fund is the appropriate unit of analysis because governance ratings are for a fund, not for individual classes, regardless of whether the fund is ms or non-ms. while using the fund level data ignores subtlety that class level characteristics bring to the analysis, we are interested in fund governance. as such, the traditional sample of class level observations is reduced from over 8,000 to approximately 2,300. 4.2. empirical model while morningstar’s stewardship ratings are discrete rankings ranging from 1 to 5, the actual shift from one level to another is unobservable. as such, we use an ordered probit model to examine the relationship among fund characteristics and governance ratings. the model takes the following form: y* i � xi� � ei, where ei � n(0,1). (1) ‘y* i’ takes on the values 1 to 5, corresponding to the ordinal ranking values for board qual rate and manager incent rate separately. ‘xi’ is a vector of independent control variables. all regression models include year-fixed effects, and standard errors are robust to heteroscedasticity. the primary variable of interest is ms to test hypothesis 1 (a, b), but corp culture rate is also included as an independent variable of interest to test hypotheses 2 (a, b). when manager incent rate is the dependent variable, we include board qual rate as an additional variable of interest to test hypothesis 3. analyzing morningstar component ratings is new; therefore, so is the empirical model. the choice of independent variables reflects possibly predictable relationships between the variables and ratings. if morningstar’s evaluation process is perfectly efficient, then we would have no a priori expectation that fund characteristics are related to ratings. however, morningstar’s process is partially judgment based, likely introducing measurement error. as such, we include variables in the model that may be correlated with governance ratings. 55j. handy, t. smythe / financial services review 28 (2020) 49-65 there are 11 common control variables across the ratings’ models. agg growth, growth, and small cap identify funds that are in morningstar’s aggressive growth, growth, and small cap investment objectives, respectively. growth and income funds are the omitted category. each variable is a dummy equal to 1 if the fund is in the respective category and 0 otherwise. instl identifies funds attracting institutional investors and is a dummy variable equal to 1 if the fund, or a class in the fund, is targeted to institutional investors and 0 otherwise. load identifies funds attracting retail investors in the advisor-sold channel and is a dummy variable equal to 1 if the fund, or at least one class of the fund, has a front-end load, a contingent deferred sales charge, or a level load commission structure, and 0 otherwise. 12b-1 identifies funds charging a 12b-1 fee and is a dummy equal to 1 if the fund, or at least one class of a fund, has a 12b-1 fee, and 0 otherwise. 12b-1 fees have become a primary form of compensation in advisor-sold funds, and as such, we identify this characteristic separately from load. fund assets and family assets are included to capture size at the fund and fund family level, respectively. they are measured as assets under management and transformed as the natural logarithm. fund age is the age of the oldest class in the fund, and manager tenure (mgr tenure) is the longest recorded manager tenure of a class in the fund. each variable is log transformed. both variables, before log transformation, are measured in years. year is included to control for trends in the data and takes the values 2005–2009 for each year a fund appears in the sample. finally, when manager incent rate is the dependent variable, we control for the fund’s net expense ratio (netexpense), measured as the average expense ratio across all classes of ms funds.6 5. primary empirical results 5.1. summary statistics summary statistics for the sample are presented in table 1. there are over 2,300 observations across the years meeting the data analysis requirements. in column (1), approximately 73% of observations are ms funds. the results suggest that ratings are higher for non-ms funds, with board qual rate and corp culture rate statistically significant at the 1% level. the average ratings for ms funds versus non-ms funds, respectively, are: board qual rate 3.646 versus 3.926; manager incent rate 3.272 versus 3.322; and corp culture rate 3.520 versus 4.206. in the full sample, approximately 60% of funds have loads (column 1), dominated by the ms structure, where 80% of funds (column 2) have at least one load class. this is expected given that ms funds are targeted to the advisor-sold channel. however, in today’s market, funds using the ms structure also include institutional classes. sample wide, 51% of funds have an institutional representation; however, this is driven by the ms subsample, where 68% of funds have an institutional class, while only 2.6% of non-ms funds are for institutional investors. this is evidence the ms structure has broadened since its introduction. approximately 50% of funds have a 12b-1 fee, but again, this is driven by the ms subsample where 64% of funds have a class with a 12b-1 fee, further evidence the ms structure targets the advisor-sold channel. 56 j. handy, t. smythe / financial services review 28 (2020) 49-65 the full sample fund family average assets under management is $111.7 billion, but this differs across fund structure. the average size for non-ms funds (column 3) is $167.3 billion, while for the ms subsample it is $90.9 billion. the average fund has approximately $4.4 billion dollars under management, roughly equivalent across fund structure. the average fund age in the sample is 18.6 years, with ms funds having an age of 20 years versus 14.6 for non-ms funds. the difference across structure is not surprising. from the industry’s beginnings until approximately 1980, all funds were sold with a load. many of these funds converted to the ms structure upon its approval in the 1990s. the average manager tenure is approximately 5.9 years, similar across fund structure. finally, the average net expense ratio, defined as the fund’s gross expense ratio minus the 12b-1 fee, is 94 basis points and is similar across fund structure. 5.2. multi-variate analysis the univariate results suggest fund governance quality may be related to fund structure. we now examine the determinants of governance ratings in a multivariate framework using table 1 summary statistics variables full sample multiple share class funds non-ms funds (ms � 0) difference significance n (1) mean standard deviation n (2) mean standard deviation n (3) mean standard deviation board qual rate 2,349 3.719 0.755 1,724 3.644 0.713 625 3.926 0.825 �0.282 *** manager incent rate 2,349 3.285 1.116 1,724 3.272 1.109 625 3.322 1.135 �0.049 corp culture rate 2,349 3.698 0.999 1,724 3.513 0.965 625 4.206 0.913 �0.693 *** age in years 2,349 18.640 15.901 1,724 20.119 16.937 625 14.560 11.688 5.559 *** fund assets in mill 2,348 4.429 1.189 1,724 4.530 1.124 625 4.151 11.181 0.379 family assets in mill 2,349 111.263 169.974 1,723 90.923 148.013 625 167.337 209.564 �76.415 *** load 2,349 0.603 0.489 1,724 0.803 0.398 625 0.051 0.221 0.752 *** instl 2,349 0.130 0.165 1,724 0.168 0.150 625 0.026 0.158 0.143 *** 12b-1 2,349 0.303 0.248 1,724 0.401 0.210 625 0.033 0.100 0.368 *** ms 2,349 0.734 0.442 1,724 1.000 0.000 625 0.000 0.000 netexpense 2,349 0.940 0.369 1,724 0.945 0.346 625 0.926 0.424 0.019 mgr tenure in years 2,336 5.938 4.643 1,722 5.861 4.434 614 6.154 5.184 �0.293 note: ***p � 0.01. this table presents the summary statistics for the entire sample of fund level observations in the growth and income, growth, agg growth, and small cap investment objectives. we present the pooled sample results and the non-ms and ms sub-samples for comparison. board qual rate, manager incent rate, and corp culture rate are measures of board quality, managerial incentives, and fund sponsor corporate culture as evaluated by morningstar. each ranges from 1 (lowest) to 5 (highest). ms is equal to 1 if the fund is an ms fund and 0 otherwise. load is a dummy equal to one if the fund has a commission structure (fel, cdsc, or level load) associated with it and 0 otherwise. instl equals 1 if the fund is for institutional investors or has a class for institutional investors and 0 otherwise. fund assets and family assets are measured in millions of dollars and measure the size of the fund and fund sponsor, respectively. fund age and mgr tenure are measured in years and measure the age of the fund and the length of time the manager has been with the fund. 12b-1 equals one if the fund or a class of the fund has a 12b-1 fee and 0 otherwise. netexpense is the difference between the funds gross expense ratio and any 12b-1 fee. year is the year in which the fund is in the sample. the final column represents the results of a t-test between the non-ms and ms sub-samples for the variables. the t-value and the p-value are presented. the samples are not assumed to have equal variances. asterisks represent significance at the 10% (*), 5% (**), and 1% (***) level, respectively. 57j. handy, t. smythe / financial services review 28 (2020) 49-65 the ordered probit model described by equation (1). the results are presented for each dependent variable separately. 5.2.1. board qual rate table 2 presents the results analyzing the relationship between board qual rate and the fund characteristics discussed previously. our primary interest is on the variables ms and corp culture rate. based on hypothesis 1(a), we expect the coefficient estimate for ms to be positive and statistically significant. the results in column (1) do not support the hypothesis but do corroborate the univariate analysis above. ms is negative and statistically significant. the results indicate that ms funds are less likely (likelier) to have higher (lower) board quality scores than non-ms funds. this result is concerning given the sec’s focus on protecting individual investors, as investors in ms funds are likely to be most vulnerable to the principal-agent problem with fund management. table 3, panels a and b provide predicted probabilities.7 panel a indicates that ms funds are three times more likely to have the lowest board quality rating (a probability of 0.4% vs. 0.1%, respectively) and are consistently more likely to have below average board quality ratings. moreover, ms funds are significantly less likely to achieve the highest board quality rating (12.2% vs. 20%, respectively). in column (2) of table 2, we test hypothesis 2(a), that is, whether fund sponsors with better corporate culture, as rated by morningstar, are more likely to operate funds with better board quality. there is strong support for hypothesis 2(a). corp culture rate is positive and table 2 board qual rate dependent variable variable (1) (2) corp culture rate 0.417*** (0.0306) ms �0.333*** (0.0845) �0.352*** (0.0794) load 0.132 (0.100) 0.286*** (0.0946) instl �0.0403 (0.0555) 0.0222 (0.0540) family assets �0.0576*** (0.0147) �0.0883*** (0.0148) family age �0.149*** (0.0324) �0.0651** (0.0322) lnassets 0.125*** (0.0197) 0.0851*** (0.0187) 12b-1 �0.265* (0.149) �0.0582 (0.145) agg growth �0.133 (0.110) �0.0117 (0.114) growth 0.00986 (0.0540) 0.0148 (0.0541) small cap 0.100 (0.0749) 0.0527 (0.0759) mgr tenure 0.131*** (0.0240) 0.0807*** (0.0245) year �0.0263 (0.0194) �0.0340* (0.0193) year-fixed effects yes yes observations 2,324 2,324 note: robust standard errors in parentheses. *p � 0.1, **p � 0.05, ***p � 0.01. this table presents ordered probit results of estimating equation (1) with board qual rate as the dependent variable. all variables are defined as in the table 1 heading, except family assets, fund assets, fund age, and mgr tenure, which are the log transformed value of family assets, fund net assets, fund age, and fund manager tenure. year-fixed effects are included and standard errors are robust to heteroscedasticity. asterisks represent significance at the 10% (*), 5% (**), and 1% (***) level, respectively. 58 j. handy, t. smythe / financial services review 28 (2020) 49-65 significant at the 1% level. equally important is that our primary variable of interest, ms, continues to be negative and significant. panel b of table 3 further details the results. across all corporate culture ratings, ms funds are more likely to have lower board quality ratings compared with non-ms funds and in most cases are half as likely to achieve the highest board quality ratings (a rating of 4 or 5). interestingly, as corporate culture ratings increase, ms and non-ms funds are less likely to have poor board quality ratings and are more likely to have higher ratings, so fund sponsor culture clearly influences governance at the fund level. the general relationship continues to hold: ms funds are more likely to have lower board quality ratings compared with non-ms funds. when comparing funds with the worst corporate culture ratings (a score of 1), the model predicts ms funds will have the worst board quality rating (a score of 1) with a 3.8% probability. this is double the 1.7% probability associated with non-ms funds. for the same corporate culture rating, ms funds have only a 1% chance of receiving the highest board quality rating (a score of 5) while non-ms funds have a 2.3% chance. focusing on sponsors table 3 board qual rate predictive probabilities panel a: ms board qual rate 1 2 3 4 5 0 0.001 0.017 0.277 0.505 0.2 1 0.004 0.036 0.375 0.464 0.122 difference (basis points) 30 190 980 �410 �780 % difference 300.00% 111.76% 35.38% �8.12% �39.00% panel b: board qual rate corp culture rate ms 1 2 3 4 5 1 0 0.017 0.112 0.57 0.277 0.023 1 1 0.038 0.178 0.591 0.183 0.01 difference (basis points) 210 660 210 �940 �130 % difference 123.53% 58.93% 3.68% �33.94% �56.52% 2 0 0.006 0.057 0.485 0.396 0.056 2 1 0.015 0.102 0.561 0.295 0.027 difference (basis points) 90 450 760 �1010 �290 % difference 150.00% 78.95% 15.67% �25.51% �51.79% 3 0 0.002 0.025 0.361 0.493 0.119 3 1 0.005 0.05 0.467 0.414 0.064 difference (basis points) 30 250 1060 �790 �550 % difference 150.00% 100.00% 29.36% �16.02% �46.22% 4 0 0 0.009 0.235 0.536 0.219 4 1 0.001 0.021 0.341 0.504 0.132 difference (basis points) 10 120 1060 �320 �870 % difference n/a 133.33% 45.11% �5.97% �39.73% 5 0 0 0.003 0.134 0.507 0.356 5 1 0 0.008 0.218 0.536 0.238 difference 0 50 840 290 �1180 % difference n/a 166.67% 62.69% 5.72% �33.15% this table presents predicted probabilities resulting from ordered probit regressions. board qual rate takes on an integer value between 1 and 5 reflecting the morningstar board quality rating. ms equals zero if the fund is a non-ms fund and 1 if the fund is an ms fund. the values within the table, unless labelled otherwise, represent percentages in decimal form. panel a provides predicted probabilities based on fund type alone. panel b provides predicted probabilities for ms and non-ms funds conditional on their morningstar corporate culture rating. 59j. handy, t. smythe / financial services review 28 (2020) 49-65 with only the highest corporate culture rating (a score of 5), no firms have a board quality score of 1; however, ms funds are twice as likely to receive the next worse rating (0.8% vs. 0.3%, respectively) and are significantly less likely to achieve the highest board quality score (23.8% vs. 35.6%). a visual representation of predicted probabilities is useful. fig. 1 provides a graphical comparison of board qual rate across mutual fund types (ms vs. non-ms, fig. 1a) and across mutual fund type and corp culture rate (fig. 1b).8 both figures show that non-ms funds have lower probabilities of receiving lower board quality ratings and higher fig. 1. ms versus non-ms board qual rate fig. 1 compares board qual rate across fund type (ms vs. non-ms; fig. 1a) and across both fund type and corp culture rate (fig. 1b) series ms � 0 reflects non-ms funds in fig. 1a. series ms � 1 reflects ms funds in fig. 1a. series ms1c1 graphs predicted probabilities for ms funds with the lowest corp culture rate (corp culture rate � 1; fig. 1b). series ms1c5 graphs predicted probabilities for ms funds with the highest corp culture rate (corp culture rate � 5; fig. 1b). series ms0c1 graphs predicted probabilities for non-ms funds with the lowest corp culture rate (corp culture rate � 1; fig. 1b). series ms0c5 graphs predicted probabilities for non-ms funds with the highest corp culture rate (corp culture rate � 5; fig. 1b). 60 j. handy, t. smythe / financial services review 28 (2020) 49-65 probabilities of higher board quality ratings (ms0c1 and ms0c5). at the most extreme, fig. 1b shows that non-ms funds with the higher corporate culture ratings have approximately a 40% higher chance of receiving a top board quality rating than ms funds with the lowest corporate culture rating. we next discuss the results for the control variables in table 2, although for space considerations we do not provide or discuss predicted probabilities. the variables agg growth, growth, and small cap are not statistically significant. however, other fund characteristics are significantly correlated with board quality. funds with 12b-1 fees (12b-1) have significantly lower board quality ratings than those without a 12b-1 fee in column (1). however, when corp culture rate is included in column (2), 12b-1 is no longer statistically significant. in contrast, load is positive but not significant in column (1), but when corp culture rate is included in column (2), load is positive and statistically significant. looking at characteristics related to family/fund size, age, and manager tenure tells a mixed story. larger fund families (family assets) are associated with lower board qual rate, independent of the effect of corporate culture, as are older funds (fund age). both results are consistent with fund families (older funds) being complacent, possibly because of prior success. also, larger fund families may be more likely to use “captured boards,” that is, board members sit on multiple boards within a family, receiving significant levels of compensation from the family. in contrast to family size and fund age, larger funds (fund assets) and funds with longer tenured managers (mgr tenure) have higher measures of board qual rate. the finding for fund assets may be indicative of larger funds being in the spotlight and responding to implicit pressure to provide strong governance. the finding for longer tenured managers is consistent with these managers finding value in strong governance. the final control, year, indicates that on average board qual rate is declining over the sample period, although only significantly so when corp culture rate in included in the model. 5.2.2. manager incent rate results from estimating equation (1) with manager incent rate as the dependent variable are presented in table 4. as discussed above, we augment the control variables by including corp culture rate (hypothesis 2(b)) and board qual rate (hypothesis 3) in the model. additionally, we include netexpense as an additional control variable. when analyzing hypothesis 1(b), column (1) in table 4 shows that ms is negative and significant, indicating ms funds are less (more) likely to have higher (lower) managerial incentive ratings than non-ms funds, which does not support hypothesis 1(b). table 5, panels a and b provide predicted probabilities of the variables of interest, similar to table 3. panel a focuses strictly on the fund structure’s relationship to the managerial incentive rating. ms funds are twice as likely to have the lowest managerial incentive rating (7.3% vs. 3.5%, respectively) and are significantly less likely to achieve the highest managerial incentive rating (15.4% vs. 25.6%). in total, the findings suggest that managerial incentives are less likely aligned with retail investors in the advisor-sold channel, giving rise to a more pronounced principal-agent problem between shareholders and fund management. 61j. handy, t. smythe / financial services review 28 (2020) 49-65 as argued above, corporate culture and board quality may influence managerial incentive quality independently. the results in columns (2–4) of table 4 provide support for the hypotheses that higher corporate culture ratings lead to higher managerial incentive ratings (hypothesis 2(b)) and higher board quality ratings lead to higher managerial incentive ratings (hypothesis 3), independently, and when both are included in the model (column (4)). funds with higher corp culture rate and higher board qual rate are more likely to have higher managerial incentive ratings, a sign of strong sponsor and fund level governance. ms funds continue to have lower managerial incentive ratings. the significance of the ms variable in table 4 when compared with the univariate findings in table 1 highlights the importance of multivariate analysis. panel b of table 5 presents the analysis of how the predicted probability of achieving a certain managerial incentive rating changes across fund type and either, the corporate culture rating or the board quality rating. the results are similar across rating type and echo results in panel b of table 3. ms funds with the lowest corporate culture ratings are likelier to have the lowest managerial incentive ratings with a 12.9% probability (compared with non-ms’ 6.9% probability). additionally, ms funds with the lowest corporate culture ratings are also significantly less likely to have the highest managerial incentive ratings (8.8% vs. 16.2%, respectively). as corporate culture improves, both ms and non-ms funds are more likely to achieve higher managerial incentive ratings; however, ms funds are more likely to achieve lower scores and less likely to achieve higher scores compared with non-ms funds. for example, funds having a corporate culture rating of 5, non-ms funds have a 31% probability of achieving the highest managerial incentive rating, while ms funds only have a 19.5% table 4 manager incent rate dependent variable variable (1) (2) (3) (4) board qual rate 0.140*** (0.0314) 0.0955*** (0.0324) corporate cult rate 0.158*** (0.0285) 0.134*** (0.0296) ms �0.411*** (0.0830) �0.385*** (0.0830) �0.416*** (0.0842) �0.398*** (0.0844) load 0.149 (0.0983) 0.133 (0.0984) 0.198* (0.102) 0.179* (0.102) instl �0.144** (0.0573) �0.139** (0.0573) �0.118** (0.0575) �0.118** (0.0575) netexpense �0.0163 (0.0821) 0.0215 (0.0823) 0.0243 (0.0823) 0.0437 (0.0827) family assets �0.113*** (0.0164) �0.105*** (0.0164) �0.121*** (0.0165) �0.115*** (0.0167) fund assets 0.152*** (0.0172) 0.143*** (0.0171) 0.138*** (0.0173) 0.134*** (0.0173) 12b-1 0.450*** (0.121) 0.475*** (0.118) 0.542*** (0.122) 0.546*** (0.121) mgr tenure 0.182*** (0.0240) 0.171*** (0.0242) 0.163*** (0.0241) 0.158*** (0.0242) fund age �0.0956*** (0.0293) �0.0817*** (0.0296) �0.0607** (0.0302) �0.0565* (0.0303) agg growth 0.114 (0.112) 0.117 (0.112) 0.153 (0.112) 0.149 (0.112) growth 0.0752 (0.0545) 0.0681 (0.0550) 0.0707 (0.0553) 0.0666 (0.0555) small cap 0.110 (0.0753) 0.0895 (0.0762) 0.0796 (0.0765) 0.0701 (0.0769) year 0.133*** (0.0185) 0.136*** (0.0185) 0.132*** (0.0183) 0.134*** (0.0184) year-fixed effects yes yes yes yes observations 2,324 2,324 2,324 2,324 note: robust standard errors in parentheses. *p � 0.1, **p � 0.05, ***p � 0.01. this table presents ordered probit results of estimating equation (1) with manager incent rate as the dependent variable. all variables are defined as in the table 1 heading, except family assets, fund assets, fund age, and mgr tenure, which are the log transformed value of family assets, fund net assets, fund age, and fund manager tenure. year-fixed effects are included and standard errors are robust to heteroscedasticity. asterisks represent significance at the 10% (*), 5% (**), and 1% (***) level, respectively. 62 j. handy, t. smythe / financial services review 28 (2020) 49-65 t ab le 5 m an ag er in ce nt r at e pr ed ic tiv e pr ob ab ili tie s m an ag er in ce nt r at e ta ke s on an in te ge r va lu e be tw ee n 1 an d 5 re fle ct in g th e m or ni ng st ar pa ne l a : m an ag er in ce nt r at e m s 1 2 3 4 5 0 0. 03 5 0. 11 9 0. 34 2 0. 24 8 0. 25 6 1 0. 07 3 0. 18 2 0. 38 4 0. 20 9 0. 15 4 d if fe re nc e (b as is po in ts ) 38 0 63 0 42 0 � 39 0 � 10 20 % d if fe re nc e 10 8. 57 % 52 .9 4% 12 .2 8% � 15 .7 3% � 39 .8 4% pa ne l b : m an ag er in ce nt r at e m an ag er in ce nt r at e c or p c ul tu re r at e m s 1 2 3 4 5 b oa rd q ua l r at e m s 1 2 3 4 5 1 0 0. 06 9 0. 17 5 0. 38 1 0. 21 3 0. 16 2 1 0 0. 05 6 0. 15 9 0. 37 6 0. 22 5 0. 18 4 1 1 0. 12 9 0. 24 0. 38 3 0. 16 0. 08 8 1 1 0. 11 0. 22 5 0. 38 9 0. 17 4 0. 10 3 d if f. (b ps ) 60 0 65 0 20 � 53 0 � 74 0 d if f. (b ps ) 54 0 66 0 13 0 � 51 0 � 81 0 % d if f 86 .9 6% 37 .1 4% 0. 52 % � 24 .8 8% � 45 .6 8% % d if f 96 .4 3% 41 .5 1% 3. 46 % � 22 .6 7% � 44 .0 2% 2 0 0. 05 4 0. 15 3 0. 37 0. 22 8 0. 19 4 2 0 0. 04 7 0. 14 3 0. 36 6 0. 23 5 0. 20 9 2 1 0. 10 6 0. 21 9 0. 38 8 0. 17 9 0. 10 9 2 1 0. 09 4 0. 20 9 0. 39 0. 18 7 0. 11 9 d if f. (b ps ) 52 0 66 0 18 0 � 49 0 � 85 0 d if f. (b ps ) 47 0 66 0 24 0 � 48 0 � 90 0 % d if f 96 .3 0% 43 .1 4% 4. 86 % � 21 .4 9% � 43 .8 1% % d if f 10 0. 00 % 46 .1 5% 6. 56 % � 20 .4 3% � 43 .0 6% 3 0 0. 04 2 0. 13 3 0. 35 5 0. 24 1 0. 22 9 3 0 0. 03 9 0. 12 8 0. 35 4 0. 24 4 0. 23 5 3 1 0. 08 5 0. 19 7 0. 38 7 0. 19 7 0. 13 4 3 1 0. 08 1 0. 19 3 0. 38 8 0. 2 0. 13 8 d if f. (b ps ) 43 0 64 0 32 0 � 44 0 � 95 0 d if f. (b ps ) 42 0 65 0 34 0 � 44 0 � 97 0 % d if f 10 2. 38 % 48 .1 2% 9. 01 % � 18 .2 6% � 41 .4 8% % d if f 10 7. 69 % 50 .7 8% 9. 60 % � 18 .0 3% � 41 .2 8% 4 0 0. 03 2 0. 11 3 0. 33 6 0. 25 1 0. 26 8 4 0 0. 03 2 0. 11 4 0. 34 0. 25 1 0. 26 3 4 1 0. 06 8 0. 17 5 0. 38 1 0. 21 4 0. 16 2 4 1 0. 06 8 0. 17 7 0. 38 4 0. 21 2 0. 15 8 d if f. (b ps ) 36 0 62 0 45 0 � 37 0 � 10 60 d if f. (b ps ) 36 0 63 0 44 0 � 39 0 � 10 50 % d if f 11 2. 50 % 54 .8 7% 13 .3 9% � 14 .7 4% � 39 .5 5% % d if f n /a 55 .2 6% 12 .9 4% � 15 .5 4% � 39 .9 2% 5 0 0. 02 5 0. 09 5 0. 31 3 0. 25 7 0. 31 5 0 0. 02 7 0. 10 1 0. 32 4 0. 25 6 0. 29 2 5 1 0. 05 4 0. 15 3 0. 37 0. 22 9 0. 19 5 5 1 0. 05 8 0. 16 1 0. 37 7 0. 22 4 0. 18 d if f. (b ps ) 29 0 58 0 57 0 � 28 0 � 11 50 d if f. (b ps ) 31 0 60 0 53 0 � 32 0 � 11 20 % d if f 11 6. 00 % 61 .0 5% 18 .2 1% � 10 .8 9% � 37 .1 0% % d if f 11 4. 81 % 59 .4 1% 16 .3 6% � 12 .5 0% � 38 .3 6% m an ag er ia l in ce nt iv e ra tin g. m s eq ua ls ze ro if th e fu nd is a n on -m s fu nd an d 1 if th e fu nd is an m s fu nd . t he va lu es w ith in th e ta bl e, un le ss la be lle d ot he rw is e, re pr es en tp er ce nt ag es in de ci m al fo rm .p an el a pr ov id es pr ed ic te d pr ob ab ili tie s ba se d on fu nd ty pe al on e. pa ne lb pr ov id es pr ed ic te d pr ob ab ili tie s fo r m s an d n on -m s fu nd s co nd iti on al on th ei r m or ni ng st ar c or po ra te c ul tu re ra tin g or th e m or ni ng st ar b oa rd q ua lit y ra tin g. 63j. handy, t. smythe / financial services review 28 (2020) 49-65 probability. the board quality results echo the corporate culture findings. in short, governance matters and appears stronger in non-ms funds, as rated by morningstar. the results for the control variables capturing load status, fund objectives, family size, fund age, fund size, and manager tenure are all similar to those in table 2. in all columns of table 4, instl is negative and statistically significant, indicating funds having an institutional presence have lower managerial incentive ratings. in all columns, the presence of a 12b-1 fee leads to increased manager incent rate. finally, the inclusion of netexpense to the models adds no explanatory power. finally, unlike in table 2, year is positive and significant, indicating that on average managerial incentive ratings are increasing during the sample period. 6. conclusion while mutual funds have been widely studied, the bulk of the work has been in the area of performance and cost structure. there has been much less work, theoretical or empirical, examining the ms fund structure introduced widely in 1995. the evidence that does exist is not complimentary of the structure. the analysis in this article finds further evidence highlighting problems with the ms structure. ms funds have board quality and managerial incentive ratings that would suggest that they, and the companies sponsoring them, tend to align themselves with investor interests less. specifically, the governance quality appears lower than in non-ms funds as determined by an independent rating source, morningstar. while we do not attribute causality to these results, they do identify another unfavorable characteristic associated with the ms structure. the results remain consistent as we account for changes in other governance variables. the findings in this article suggest that in addition to the risk of being directed toward classes with suboptimal costs, as suggested by o’neal (1999), advisor-led retail investors are also, simply by investing in ms funds, investing in funds whose governance is subpar compared with non-ms funds. this is concerning as investors in the advisor-sold distribution channel have been shown to be less financially savvy. in total, evidence is mounting against the ms structure. notes 1 we take the morningstar ratings as given and assume that they are unbiased and useful as tools to evaluate board quality and managerial incentives, that is, this article does not evaluate board or managerial effectiveness directly. 2 here, the term “costs” encompasses commissions and ongoing fund expenses. often, the term cost is used interchangeably with “ongoing fund expenses” or the “expense ratio.” 3 it is important to note that the authors do not mean the fund’s investment advisor when referencing “advisor” but rather mean the investor’s financial advisor (broker, wealth manager etc.). 4 we recognize that a financial advisor could also fulfill some, or all, of this oversight role, although literature cited above suggests that they may not. 64 j. handy, t. smythe / financial services review 28 (2020) 49-65 5 in an effort to update the sample used in this article, the authors contacted morningstar sales and research departments and were advised that morningstar stopped providing stewardship ratings in approximately 2017. additionally, morningstar indicated that historical data was no longer available. 6 see handy et al. (2018) for how fund expenses and managerial incentives may be related. 7 the results remain consistent when evaluating predicted probabilities holding all other variables at the mean here and in the remainder of the article. 8 we provide the graphical representation for the results in table 2 only. to conserve space, we do not provide figures for table 4, although the conclusions mirror those in figure 1 and are available upon request. references alexander, g. j., jones, j. d., & nigro, p. j. (1998). mutual fund shareholders: characteristics, investor knowledge, and sources of information. financial services review, 7, 301–316. barber, b. m., odean, t., & zheng, l. (2005). out of sight, out of mind: the effects of expenses on mutual fund flows. journal of business, 78, 2095–2120. bergstresser, d., chalmers, j. m. r., & tufano, p. (2009). assessing the costs and benefits of brokers in the mutual fund industry. the review of financial studies, 22, 4129–4156. cao, j., ghosh, a., goh, j., & ng, s. (2014). governance matter: morningstar stewardship grades and mutual fund performance. working paper, singapore management university. capon, n., fitzsimons, g. j., & prince, r. a. (1996). an individual level analysis of the mutual fund investment decision. journal of financial services research, 10, 59–82. chou, j., ng, l., & wang, q. (2011). are better governed funds better monitors? journal of corporate finance, 17, 1254–1271. fact sheet. (2006). the morningstar stewardship grade for funds. (available at https://quicktake.morningstar.com/ datadefs/stewgrademethodology.pdf). handy, j., nichols, h., & smythe, t. (2018). should investors care about mutual fund governance quality?: evidence from morningstar stewardship ratings. the journal of wealth management, 21, 44–58. jones, m. a., lesseig, v. p., & smythe, t. i. (2005a). financial advisors and mutual fund selection. journal of financial planning, 18, 64–70. jones, m. a., lesseig, v. p., & smythe, t. i. (2005b). financial advisors and multiple share class mutual funds. financial services review, 14, 1–20. lesseig, v. p., long, d. m., & smythe, t. i. (2002). gains to mutual fund sponsors offering multiple share class funds. journal of financial research, 25, 81–98. livingston, m., & o’neal, e. s. (1998). the cost of mutual fund distribution fees. journal of financial research, 21, 205–218. moore, s., & porter, g. e. (2017). can investors benefit from using morningstar’s stewardship grades? journal of accounting and finance, 17, 130. nanda, v. k., wang, z. j., & zheng, l. (2009). the abcs of mutual funds: on the introduction of multiple share classes. journal of financial intermediation, 18, 329–361. nofsinger, j. r., & varma, a. (2007). how analytical is your financial advisor? financial services review, 16, 245–260. o’neal, e. s. (1999). mutual fund share classes and broker incentives. financial analysts journal, septemberoctober, 76–87. securities and exchange commission rule 18f-3 adoption release, release no. 33-7143, ic-20915, 1995. 65j. handy, t. smythe / financial services review 28 (2020) 49-65 encouraging living will completion using social norms and family benefit reem husseina, russell n. james iiia,* adepartment of personal financial planning, college of human sciences, texas tech university, lubbock, tx 79409-1210, united states abstract advance directives, such as a living will, can help families control their medical treatments and, in some cases, appropriately limit end-of-life medical expenses. however, usage of such documents remains relatively low. applying concepts from terror management theory, this study randomly assigned 1,771 online participants to living will descriptions referencing social norms, family benefit, both, or neither. references to family benefit alone significantly increased intentions to complete documents among men, but non-significantly decreased intentions among women. references to social norms alone modestly increased intentions for both groups. combining references to both family benefit and social norms generated the largest increase. © 2021 academy of financial services. all rights reserved. jel classifications: d1; d14; d15 keywords: estate planning; advance directives; living wills; terror management theory 1. introduction advance directives include statements of preferences regarding the use of life-sustaining technology (living will) and appointment of another to make health care decisions for them when they cannot (durable power of attorney for healthcare; king, 1996). estate planning in general is an important part of family financial resource management (delgadillo, 2014; kabaci & cude, 2015) and by allowing individuals to express their wishes regarding their end-of-life health care, a living will document can limit the financial impact of this end-oflife medical care (nicholas, 2011). *corresponding author. tel.: +1-806-787-5931; fax: +1-806-834-5130. e-mail address: russell.james@ttu.edu (r.n. james, iii) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 85–99 despite the benefits of advance directives, completion rates remain low (salmond & david, 2005). a review of 150 studies published from 2011 to 2016 found that only 37% of u.s. adults had completed any advance directives (yadav et al., 2017). it is possible that death anxiety and mortality salience play a role in the low completion rates of advance directives. discussions about death and dying tend to be a taboo topic in the united states (mclaughlin & braun, 1998; walter, 1991). terror management theory (tmt) provides a theoretical framework for people’s management of death-related thoughts. this study tests the effects of two messages consistent with a tmt approach, social norms and family benefit, both alone and together on intentions to complete a living will advance directive. 2. literature review death anxiety and the desire to avoid death-related topics may be one issue that prevents people from completing their advance directives. meeker and jezewski (2005) concluded that the primary reason why people do not complete end-of-life planning is to avoid facing their own mortality. duke et al. (2007) found that procrastination in completing advance directives related to denial and avoidance. 2.1. theory tmt, based on the body of work by cultural anthropologist ernest becker (1973), provides a theoretical framework for mortality salient decision making (greenberg et al., 1997). it suggests that death reminders generate two defenses, a proximal defense of avoidance and a distal defense of pursuit of symbolic immortality (pyszczynski et al., 1999). the pursuit of symbolic immortality is expressed by supporting one’s surviving “in-group” and their cultural worldviews (burke et al., 2010). both avoidance and support of in-group cultural worldviews aid in managing the fear of death (greenberg et al., 1997). as iverson and buttigieg (1997, p. 1487), explain, “the ‘in-group’ is defined as the clique with which the individual identifies.” we do not live forever, but our sources of identity, such as “our people” (family members, loved ones, or other in-group members) or our values (i.e., values supported by our identifying in-group) will continue in the world. they will survive us. in experiments, death reminders increase the importance of being positively remembered by this surviving in-group (greenberg et al., 2010). as such, in-group social norms (a.k.a., herd behavior) will tend to become more powerful in a death salient context (fritsche et al., 2010; gailliot et al., 2008; maheswaran & agrawal, 2004) as will leaving a positive impact on surviving loved ones such as family members (burke et al., 2010; james, 2016a). the following experiments explore the potential practical application of this general theoretical principle to the area of living wills by separately testing a message emphasizing a social norm, a message emphasizing a family benefit, and a message emphasizing both a social norm and a family benefit. james (2016a) presents an economic model predicting similar outcomes. mortality salience generates responses of avoidance and pursuit of “lasting social impact” simply as the result of utility maximization when such includes expectations 86 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 of future circumstances, as suggested by brunnermeier and parker (2005), and utility from the circumstances of others, as suggested by gary becker (1974). thus, with both models the two predicted outcomes triggered by a mortality salient condition are the same: avoidance and/or some form of social impact related to one’s surviving in-group. this is relevant given the plausibility of experiencing mortality salience when contemplating completion of a living will document. 2.2. avoidance and word choice in experiments previous experiments demonstrate the impact of descriptions using a more or less mortality-salient approach. results are consistent with the idea that mortality-salient descriptions tend to increase avoidance. salisbury and nenkov (2016) found that changing the description of annuity benefits from paying “each year you live” to paying “each year you live until you die” decreased interest in purchasing them. james (2016b) found that in a charitable bequest description, replacing “last will and testament” with “last will and testament that will take effect at your death” significantly decreased interest in making such gifts. in studying preferences for cancer treatments, o’connor (1989) found that a negative frame presenting the risk of dying reduced interest in more aggressive cancer treatments as compared with a positive frame presenting the chance of survival. 2.3. social impact descriptions in experiments previous research also supports the heightened impact of supporting one’s surviving ingroup and their cultural worldviews in a mortality salient context. a simple expression of this response is found in an increased desire to comply with social norms following mortality reminders (gailliot et al., 2008). fritsche et al. (2010) showed that in the presence of statements of pro-environmental social norms, mortality salience increased sustainable behaviors. maheswaran and agrawal (2004) studied the effects of mortality salience on consumer behavior. they found that “when mortality is salient, people are more willing to act in concert with the opinions of others” (p. 214). social norms have proven effective in descriptive word choice experiments related to other end-of-life planning contexts. james (2016b) found that adding a social norm statement (“many people like to leave a gift to charity in their wills”) to the description of a charitable bequest gift significantly increased interest in making the gift. sanders and smith (2016) conducted an experiment in which lawyers asked clients during the process of drafting a will if they wanted to leave a gift to charity in their will. they found that highlighting a social norm of charitable giving with the phrase, “many of our customers like to leave a gift to charity in their will” increased the number and the size of bequest gifts to charity. another expression of support for one’s surviving in-group is a desire to benefit one’s own family. this concern is paramount in estate planning. previous research studies in charitable bequests indicate that the desire to meet family needs and expectations is the most challenging barrier for such gifts. interviews with a sample of bequest fundraisers in australia found that attitudes towards estate planning were overwhelmingly influenced by r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 87 expectations of honoring family ties (baker, 2008). madden and scaife (2008) found that even those who included a charitable bequest in their estate plans explained their bequests in terms of family responsibilities. james (2015) found that resolving the conflict between family and charitable bequests by combining a reminder of family connections to a charitable cause with the opportunity to leave a charitable gift in honor of a family member was particularly effective in increasing charitable bequest intentions. 2.4. applications to medical conversations the low level of completed advance directives may relate to the presentation of or conversations around the documents. from a broad perspective, previous research has suggested the need to reconceptualize advance directives as part of a process to communicate and negotiate goals of medical care that satisfy the individual’s wishes and values (morrison et al., 1995; teno et al., 1997). this may be aided by simple descriptive wording changes. other experiments have found significant effects from slight wording changes for descriptions of various types of medical decisions. malloy et al. (1992) found that how life-sustaining interventions were described influenced whether individuals accepted or rejected the treatments in their advance directives. in a study of word choice in the context of choice of cancer treatments, mcneil et al. (1982) concluded that respondents were more willing to accept the riskier option if the outcomes of treatments were positively framed. previous studies test the need to improve and enhance the formulation and implantation of advance directives (schneiderman et al., 1992; teno et al., 1997). this study tests the extent to which social norms and/or a reference to family benefit impact the intention to complete the living will advance directive document. hypothesis 1: a social norms reference will increase intentions to complete a living will advance directive. hypothesis 2: a family benefit reference will increase intentions to complete a living will advance directive. 2.5. socio-demographic factors grounded in differing theoretical justifications, several socio-demographic factors, including age, income, race, and gender, have been consistently associated with differences in rates of advance directive completion. the following experimental study includes controls for these factors. in addition, given the documented relevance of these factors for advance directive completion, the analysis also explores how these factors interact with the experimental treatments. older age has been associated with higher advance directive completion rates and more openness to end-of-life discussions (moorman & inoue, 2013; pollack et al., 2010). moorman and inoue (2013) find that one year of age was associated with a 3% increase in the likelihood of having end-of-life planning documents. older adults would be more likely than younger adults to be knowledgeable of end-of-life planning as a product of their own life experiences, as well as those of their spouses, and family members (carr & khodyakov, 88 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 2007). this may be because as people age, they utilize medical services that gives them opportunities to learn about end-of-life planning documents. previous research studies have also found that an individual with a higher level of income is more likely to have advance directive documents (carr, 2012; moorman & inoue, 2013). rosnick and reynolds (2003) found that people whose incomes were less than $30,000 were 66% less likely to have a living will than those whose income were $30,000 or more. carr (2012) found that people were more likely to complete other end-of-life planning when they drafted a financial last will and testament, which is less likely among individuals with fewer assets. previous studies have also found that completion rates of advance directives were consistently higher among whites than other ethnicities (alano et al., 2010; pollack et al., 2010). hopp and duffy (2000) found that whites were significantly more likely to discuss treatment preferences and, as a results, were also more likely to complete advance directives than were african americans. others have found that obtaining estate planning documentation may be more of a barrier for african americans (lehman & james, 2018). several studies have found that being female increases the odds of having written advance care planning (alano et al. 2010; bravo et al., 2003). it is possible that gender differences reflect the fact that women are more likely to experience widowhood. women may also be more likely to talk about their end-of-life treatment preferences with others that may trigger documenting those wishes in advance directives. there are also gender differences in mortality or illness perception. fletcher and sarkar (2013) found that among terminal patients, women showed a better understanding that their illness was incurable and was at an advanced stage compared with men. previous studies have found that education was positively related to completion rates of advance directives (alano et al., 2010; carr & khodyakov, 2007). moorman and inoue (2013) found that individuals with a college degree were more likely to have advance directives than those who have only a high school education. individuals with lower education levels may not be aware of the importance and availability of end-of-life planning and, in addition, the technical language used in living will documents may be difficult to understand (hopp & duffy, 2000). 3. methodology 3.1. participants participants for the experiment were recruited using amazon’s mechanical turk (mturk; https://www.mturk.com). participants were recruited with the description “university survey of opinions on health/medical planning” and payment of 75 cents for completing the survey. if participants clicked on the description, they read, “survey of health/medical opinions. we are conducting an academic survey about opinions on medical planning options and opinions, this takes around 8-10minutes, and it is intended to advance research about people and their medical planning, so please make sure you can commit the time. at the end of the survey, you will receive a unique ‘completion code’ to receive credit for taking our survey.” r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 89 the analysis excluded answers from participants who reported already having completed living will documents. the outcome question about the likelihood of completing a living will document would measure a different behavior (i.e., changing current plans) if the participant already had a living will. further, the practical issue is understanding how to motivate those who do not yet have planning documents, rather than motivating a revision of existing documents. after excluding participants who already had living will documents, the sample size used in the analysis was 1,771. the study was approved by the human subjects institutional review board (irb2019-862) of the authors’ affiliated university. 3.2. mturk and participant attention experimental participants in social science research have traditionally been recruited from convenience samples such as nearby college students. locating experimental participants using mturk offers several advantages. participant diversity can be much greater across many measurements including geography, age, cognitive scores, income, and race. further, some experimental evidence finds that the attentiveness of participants recruited from mturk exceeds that of student samples. across three separate studies, hauser and schwarz (2016) found, “in all studies, mturkers were more attentive to the instructions than were college students” (p. 400). other studies have found responses collected online from participants recruited via mturk compare favorably with responses collection in-person (buhrmester et al., 2011; casler et al., 2013). there are other online sources for recruiting participants. however, mturk appears to perform well compared with these other online sources. for example, kees et al. (2017) found that qualtrics and lightspeed panel respondents performed worse on measure reliability tests compared with mturk respondents. they found, “in comparisons across five samples, results show that the mturk data outperformed panel data procured from two separate professional marketing research companies across various measures of data quality” (kees et al., 2017, p. 141). these advantages have led to the widespread use of mturk as a source for participant recruitment across the social sciences. (a recent google scholar search finds over 40,000 documents referencing this service.) this includes experimental research in financial planning in general (fulk et al., 2018; yazdanparas & alhenawi, 2017) and end-of-life financial planning decisions in particular (james, 2018; james & routley, 2016). participant attention is important in the experiments described below. participants are randomly assigned to read either control or experimental phrases. inattention would increase the likelihood that even a highly effective experimental phrase would generate no significant difference between the treatment and control groups (bates & lanza, 2013; mullinix et al., 2015; paolacci et al., 2010). thus, to the extent that inattention is a problem in the below experiments, the impact of the experimental phrases would tend to be understated in the results. to address this concern, participants were screened using an attention check task before beginning the study. the following block of text appeared, you are about to start the research survey, and we appreciate your time and effort. your honest efforts in this survey could benefit the accuracy of information provided in the financial services 90 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 industry. however, it is critically important that you actually take the time to read instructions closely and follow them; if not, our data based on your responses will be invalid. in order to demonstrate that you read instructions, several places in this survey will contain special instructions, such as here. in order to demonstrate that you read instructions, please select the option “no answer” for the next question that asks about how often you take surveys. then type exactly the following words in the next box, “i read the instructions” in the box labeled “any comments or questions before we start?” if you do not type the words “i read the instructions” exactly as they appear between the quotations you will not be allowed to complete the survey. please type this without any quotations or punctuation. thank you very much. this was followed by the multiple-choice question, “how often do you take surveys? __ often __ sometimes __ seldom __ never __ no answer” and an open text box after “any comments or questions before we start?” participants who did not answer these questions in the nonstandard way directed by the large block of text, that is, those who skipped the text and just answered the questions quickly, were excluded from participating in the experiment. 3.3. instrument respondents answered survey questions online using the qualtrics platform during october 14-15, 2019. participants were randomly assigned to one of four groups. each group read slightly different descriptions of a living will advance directive document and then estimated the likelihood that they would complete such documents in the next 30 days. the four groups are referred to as base, base + family benefit, base + social norm, and base + social norm + family benefit. the four corresponding statements are listed in table 1. all statements began with the identical base description of a living will document. a social norm was introduced by adding to the end of the description the sentence, “many people like to have a living will.” a family benefit was introduced by adding the sentence, “a living will can relieve family members of difficult decisions.” the combination of social norms and family benefit were introduced together by adding the sentence, “many people like to have a living will because it can relieve family members of difficult decisions.” finally, all respondents were asked, “if you were given the opportunity to complete a living will document at no cost to you in the next 30 days, what is the percentage likelihood that you would do so?” participants answered from 0 to 100 using a horizontal slider bar. table 1 living will phrases text base the living will is a legal document used to address certain future health care decisions only when individuals become incapacitated or unable to make the decisions on their own. the living will is only used at the end of life if a person cannot be cured (terminally ill) or is permanently unconscious. base + family benefit a living will can relieve family members of difficult decisions. base + social norm many people like to have a living will. base + social norm + family benefit many people like to have a living will because it can relieve family members of difficult decisions r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 91 3.4. control variables the independent variables for this study include the individual’s age, gender, income, education, and race. age, education, and income were translated into single variable formats by using reported range midpoints (or the lowest value for the open-ended top range and highest value for the bottom range) to transform grouped data into continuous variables. the age categories were 18-24, 25-34, 35-44, 45-54, 55-64, 65-74, 75-79, 80-84, and 85 or older, and were converted to 21, 30, 40, 50, 60, 67, 72, 77, 82, and 85. income categories were less than $10,000, and then intervals of $10,000-$19,999; $20,000-$29,999; $30,000-$39,999; $40,000-$49,999; $50,000-$50,999; $60,000-$69,999; $70,000-$79,999; $80,000-$89,999; $90,000-$99,999; $100,000-$149,999; and greater than or equal to $150,000. these income categories were converted to $10,000; $15,000; $25,000; $35,000; $45,000; $55,000; $65,000; $75,000; $85,000; $95,000, $125,000; and $150,000. education level was converted to the estimated number of years of education. the response to “what is the highest level of education that you have completed?” was converted to nine for “less than high school,” 12 for “high school,” 13 for “some college,” 14 for “associate degree,” 16 for “bachelors degree,” 18 for “master degree,” and 20 for “doctorate degree.” 4. results 4.1. descriptive statistics table 2 shows the characteristics of the survey participants by their assignment to each living will phrase from table 1. the average age for participants was 38 years old, 52% were female, and 77% were white. the average years of education for respondents was 14 years and the mean annual income was $49,000. the average reported probability that an individual would sign a living will document if given the opportunity to do so at no cost in the next 30 days across the entire sample was 67.8%. the lowest reported probability was for the base group, 65.3%. adding the family table 2 group means (n = 1,771) variable overall base group base + family benefit group base + social norms group base + family benefit + social norms group likelihood 67.78 65.26 67.72 68.52† 69.68* male 0.48 0.47 0.51 0.46 0.50 white 0.77 0.77 0.78 0.77 0.79 income 49,009 51,008 50,725 46,770 47,477 education 14.93 15.00 14.99 14.82 14.93 age 38.20 37.97 38.00 38.36 37.98 n 1,771 456 434 435 446 note: t test comparing each experimental group with base group, †p< .10, *p< .05. 92 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 benefit statement increased this to 67.7%. adding the social norms statement increased this to 68.5%. adding both at the same time increased this to 69.7%. a two-sample t test was conducted to measure the statistical significance of these differences in the reported likelihood of completing a living will document. thus, each group was compared against the base group, where the description included references to neither social norms nor family benefit. the increase in intentions to complete a living will document resulting from addition of the family benefit statement was not statistically significant (p¼ .222). adding the social norm statement generated a marginally significant increase (p¼ .093). adding both the social norm and family benefit statements generated a statistically significant increase in the intention to complete a living will document (p¼ .025). 4.2. regression results table 3 reports the coefficients (standard errors in parentheses) from ordinary least square regressions. the outcome variable in the regression is the stated probability of completing a living will document. column 1 of table 3 shows results without control variables using the base statement as the reference group. column 2 of table 3 shows results with the control variables included. in the controlled regression, the addition of either the social norm statement alone or the combined family benefit and social norm statement significantly increased intentions to complete a living will document. the increase resulting from adding the family benefit statement alone was not statistically significant. the significant associations with control variables matched the associations found in previous research. (however, this consistency is notable as previous research measured past document completion and this study measured future document completion intentions.) those who were older, female, or had higher incomes reported a greater likelihood of completing living will documents. table 3 reported likelihood of completing a living will document when adding references to family benefit, social norms, or both (ordinary least squares regression) variable coefficient coefficient intercept 65.262*** (1.3756) 53.945*** (5.8270) base (reference) base + family benefit 2.4615 (1.9698) 2.5215 (1.9126) base + social norms 3.2553 (1.9573) 3.7641* (1.9005) base + family benefit + social norms 4.4178* (1.9698) 4.6290** (1.9131) male �10.6302*** (1.3652) white 1.5320 (1.6421) income 0.00012*** (0.00002) education 0.1150 (0.3534) age 0.1896*** (0 0.0595) note: standard errors in parentheses; n¼ 1,771. ***, **, and * indicate statistical significance at p< . 001, p< . 01, and p< .05 levels, respectively. r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 93 although these messages had positive effects on intentions to complete living will documents, it is possible that some messages worked better for some socio-demographic groups than for others. to formally test for this, additional regressions were run including interaction variables between the intervention group and each control variable. no interactions were significant except for gender. in particular, as reported in table 4, the addition of family benefit statement alone had a significantly (p< .01) greater positive impact for men than for women. to further explore this relationship, table 5 reports the results of the controlled regression when the sample was restricted either to men only or women only. this shows that the addition of the family benefit statement alone significantly increased intentions to complete a living will document for men, but non-significantly decreased intentions to complete a living will document for women. following this same pattern, the coefficient for the combined family benefit and social norms statement was larger than for the social norms statement alone among men (6.23 vs. 3.34) but was smaller among women (3.00 vs. 4.09). 5. implications a living will advance directive can be an important part of end-of-life planning. however, usage of such documents is relatively low. this study tested the effects of different messages on intentions to complete a living will advance directive. completing such documents involves explicitly planning for one’s own end of life. past theoretical work suggested that mortality salience is likely to trigger responses of avoidance and pursuit of lasting social impact (a.k.a., symbolic immortality) through support of one’s surviving in-group. this second effect can be expressed by increased interest in complying with group norms and benefitting surviving family members. matching with successful interventions in other table 4 reported likelihood of completing living will document with interaction between gender and references to family benefit, social norms, or both (ordinary least squares regression) variable coefficient intercept 55.4383*** (5.947) base (reference) base + family benefit �1.2525 (2.6804) base + social norms 4.2404 (2.650) base + family benefit + social norms 3.1312 (2.6213) male �13.040*** (2.678) white 1.4128 (1.6463) income 0.00012*** (0.0000) education 0.1061 (0.3534) age 0.18962*** (0.0595) male � family benefit 7.60389** (3.8281) male � social norms �0.8689 (3.8132) male � family benefit + social norms 3.1551 (3.8294) note: standard errors in parentheses; n¼ 1,771. ***, **, and * indicate statistical significance at p< . 001, p< . 01, and p< .05 levels, respectively. 94 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 mortality salient contexts, the current study tested the effects of referencing social norms, family benefit, or both combined. references to social norms alone modestly increased intentions to complete living will documents. combining both social norms and family benefit references significantly increased intentions. referencing family benefit alone significantly increased intentions to complete documents among men, but non-significantly decreased intentions among women (i.e., the decreased intentions among women were not statistically significant). even though both the social norms and family benefit messages fit with the theoretical prediction of a desire to support one’s in-group and their values, the differences in results suggests that these two references may work through distinct mechanisms. this also fits with the overall result that the most effective approach was to combine both messages. the suggestion to combine both messages also matches an experimental result from charitable bequest decision-making. in that experiment, a family benefit message referenced both family connections with a charitable cause and provided an opportunity for a memorial or tribute bequest (james, 2015). both this family benefit intervention and a social norm intervention increased interest in charitable bequests, but the greatest impact came from using both messages together (james, 2015). 5.1. limitations and future research these results provide a first exploration of the use of these phrasing interventions to encourage living will document completion, but they are subject to various limitations such as an online sample and a hypothetical context. findings resulting from a non-probability crowdsourcing sample lack formal statistical generalizability and thus cannot be used to estimate national population means. participant inattention to the wording differences would lead to an understatement of the impact of the phrasing differences reported here. placing the benefit description interventions at the end of, rather than the beginning of, the lengthy table 5 reported likelihood of completing living will document for male only or female one (ordinary least squares regression) variable male respondents only female respondents only intercept 35.932*** (8.4710) 62.848*** (7.8822) base (reference) base + family benefit 6.2123* (2.8521) �1.3153 (2.5664) base + social norms 3.3366 (2.8598) 4.0912 (2.5431) base + family benefit + social norms 6.2281* (2.9171) 3.0018 (2.5118) white 0.2114 (2.4560) 2.522 (2.2061) income 0.0001*** (0.0000) 0.00011*** (0.0000) education 0.3258 (0.5352) �0.16991 (0.4687) age 0.2932*** (0.0918) 0.09493 (0.0776) n 858 913 note: standard errors in parentheses. ***, **, and * indicate statistical significance at p< . 001, p< . 01, and p< .05 levels, respectively. r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 95 living will description statement may also have led to a muted difference in responses across the groups (james, 2018). a post hoc analysis also found a significant gender interaction with the family benefit message. although not predicted a priori, this difference may warrant future exploration. the post hoc exploration of gender interactions, while potentially instructive for future research, is subject to multiple comparison limitations as it was part of an exploratory examination of five control variables (gender, race, income, age, and education). finally, future studies may consider using different sources for participant recruitment to see if these results replicate with alternative samples and the inclusion of additional demographic variables. 6. conclusion although subject to various limitations, these results are important for both theoretical and practical reasons. they provide the first experimental evidence on the effect of different messages on the estimated likelihood of completing living will advance directives. they are important practically not only by showing that the combination of social norms and family benefit messages can be, overall, beneficial, but because they show that the family benefit message was particularly powerful for men. additionally, they provide evidence that the insights gleaned from work completed with other forms of end-of-planning may also apply to advance directives. this suggest the promise of cross-disciplinary research to provide understanding across end-of-life decisions whether related to healthcare, life insurance, annuities, or estate planning (james, 2016a). advanced planning can ensure that patients’ preferences for medical treatment is followed; thus, accomplishing both personal and family financial goals. however, achieving this often requires memorialization in a written document. therefore, understanding how to encourage the documentation of such preferences can be an important step to achieving important life, health, and financial outcomes. references alano, g. j., pekmezaris, r., tai, j. y., hussain, m. j., jeune, j., louis, b., el-kass, g., ashraf, m., reddy, r., lessor, m., & wolf-klein, g. p. (2010). factors influencing older adults to complete advance directives. palliative & supportive care, 8, 267-275. baker, c. (2008). “family comes first”: fundraisers’ perspectives on charitable bequests, tasa conference 2008. melbourne, australia: tasa 2008 conference proceedings: re-imagining sociology, the australian sociological association. bates, j. a., & lanza, b. a. (2013). conducting psychology student research via the mechanical turk crowdsourcing service. north american journal of psychology, 15, 385-394. becker, e. (1973). the denial of death. new york: free press. becker, g. (1974). a theory of social interactions. journal of political economy, 82, 1063-1093. bravo, g., dubois, m. f., & paquet, m. (2003). advance directives for health care and research: prevalence and correlates. alzheimer disease and associated disorders, 17, 215-222. brunnermeier, m. k., & parker, j. a. (2005). optimal expectations. american economic review, 95, 1092-1118. 96 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 buhrmester, m., kwang, t., & gosling, s. (2011). amazon’s mechanical turk: a new source of inexpensive, yet high quality data? perspectives on psychological science, 6, 3-5. burke, b. l., martens, a., & faucher, e. h. (2010). two decades of terror management theory: a meta-analysis of mortality salience research. personality and social psychology review: an official journal of the society for personality and social psychology, inc., 14, 155-195. carr, d., & khodyakov, d. (2007). end-of-life health care planning among young-old adults: an assessment of psychosocial influences. the journals of gerontology series b: psychological sciences and social sciences, 62, s135-s141. carr, d. (2012). i don’t want to die like that. . .”: the impact of significant others’ death quality on advance care planning. the gerontologist, 52, 770-781. casler, k., bickel, l., & hackett, e. (2013). separate but equal? a comparison of participants and data gathered via amazon’s mturk, social media, and face-to-face behavioral testing. computers in human behavior, 29, 2156-2160. delgadillo, l. m. (2014). financial clarity: education, literacy, capability, counseling, planning, and coaching. family and consumer sciences research journal, 43, 18-28. duke, g., thompson, s., & hastie, m. (2007). factors influencing completion of advanced directives in hospitalized patients. international journal of palliative nursing, 13, 39-43. fletcher, d., & sarkar, m. (2013). psychological resilience: a review and critique of definitions, concepts, and theory. european psychologist, 18, 12-23. fritsche, i., jonas, e., kayser, d. n., & koranyi, n. (2010). existential threat and compliance with pro-environmental norms. journal of environmental psychology, 30, 67-79. fulk, m., grable, j. e., watkins, k., & kruger, m. (2018). who uses robo-advisory services, and who does not? financial services review, 27, 173-188. gailliot, m. t., stillman, t. f., schmeichel, b. j., maner, j. k., & plant, e. a. (2008). mortality salience increases adherence to salient norms and values. personality & social psychology bulletin, 34, 993-1003. greenberg, j., kosloff, s., solomon, s., cohen, f., & landau, m. (2010). toward understanding the fame game: the effect of mortality salience on the appeal of fame. self and identity, 9, 1-18. greenberg, j., solomon, s., & pyszczynski, t. (1997). terror management theory of self-esteem and cultural worldviews: empirical assessments and conceptual refinements. in m. p. zanna (ed.), advances in experimental social psychology (vol. 29, pp. 61-139). san diego, ca: academic press. hauser, d. j., & schwarz, n. (2016). attentive turkers: mturk participants perform better on online attention checks than do subject pool participants. behavior research methods, 48, 400-407. hopp, f. p., & duffy, s. a. (2000). racial variations in end-of-life care. journal of the american geriatrics society, 48, 658-663. iverson, r. d., & buttigieg, d. m. (1997). antecedents of union commitment: the impact of union membership differences in vertical dyads and work group relationships. human relations, 50, 1485-1510. james, r. n. iii., (2015). the family tribute in charitable bequest giving: an experimental test of the effect of reminders on giving intentions. nonprofit management and leadership, 26, 73-89. james, r. n. iii., (2016a). an economic model of mortality salience in personal financial decision making: applications to annuities, life insurance, charitable gifts, estate planning, conspicuous consumption, and healthcare. journal of financial therapy, 7, 62-82. james, r. n. iii., (2016b). phrasing the charitable bequest inquiry. voluntas: international journal of voluntary and nonprofit organizations, 27, 998-1011. james, r. n. iii., (2018). describing complex charitable giving instruments: experimental tests of technical finance terms and tax benefits. nonprofit management and leadership, 28, 437-452. james, r. n. , iii., , & routley, c. (2016). we the living: the effects of living and deceased donor stories on charitable bequest giving intentions. international journal of nonprofit and voluntary sector marketing, 21, 109-117. kabaci, m. j., & cude, b. j. (2015). a delphi study to identify personal finance core concepts and competencies of first-generation college students. family and consumer sciences research journal, 43, 244-258. r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 97 kees, j., berry, c., burton, s., & sheehan, k. (2017). an analysis of data quality: professional panels, student subject pools, and amazon’s mechanical turk. journal of advertising, 46, 141-155. king, n. m. (1996). making sense of advance directives: revised edition. washington, dc: georgetown university press lehman, j., & james, r. n. iii., (2018). the charitable bequest gap among african-americans: exploring charitable, religious, and family estate planning attitudes. journal of personal finance, 17, 43-56. madden, k., & scaife, w. (2008). keeping giving going: charitable bequests and australians. brisbane: centre for philanthropy and nonprofit studies, queensland university of technology. retrieved from http:// eprints.qut.edu.au/27259/1/keeping_giving_going.pdf maheswaran, d., & agrawal, n. (2004). motivational and cultural variations in mortality salience effects: contemplations on terror management theory and consumer behavior. journal of consumer psychology, 14, 213-218. malloy, t. r., wigton, r. s., meeske, j., & tape, t. g. (1992). the influence of treatment descriptions on advance medical directive decisions. journal of the american geriatrics society, 40, 1255-1260. mclaughlin, l. a., & braun, k. l. (1998). asian and pacific islander cultural values: considerations for health care decision making. health & social work, 23, 116-126. mcneil, b. j., pauker, s. g., sox, h. c. , jr., , & tversky, a. (1982). on the elicitation of preferences for alternative therapies. new england journal of medicine, 306, 1259-1262. meeker, m. a., & jezewski, m. a. (2005). family decision making at end of life. palliative & supportive care, 3, 131-142. moorman, s. m., & inoue, m. (2013). persistent problems in end-of-life planning among young-and middle-aged american couples. the journals of gerontology series b: psychological sciences and social sciences, 68, 97-106. morrison, r. s., olson, e., mertz, k. r., & meier, d. e. (1995). the inaccessibility of advance directives on transfer from ambulatory to acute care settings. journal of the american medical association, 274, 478-482. mullinix, k. j., leeper, t. j., druckman, j. n., & freese, j. (2015). the generalizability of survey experiments. journal of experimental political science, 2, 109-138. nicholas, l. h., langa, k. m., iwashyna, t. j., & weir, d. r. (2011). regional variation in the association between advance directives and end-of-life medicare expenditures. journal of the american medical association, 306, 1447-1453. o’connor, a. m. (1989). effects of framing and level of probability on patients’ preferences for cancer chemotherapy. journal of clinical epidemiology, 42, 119-126. paolacci, g., chandler, j., & ipeirotis, p. g. (2010). running experiments on amazon mechanical turk. judgment and decision making, 5, 411-419. pollack, k. m., morhaim, d., & williams, m. a. (2010). the public’s perspectives on advance directives: implications for state legislative and regulatory policy. health policy, 96, 57-63. pyszczynski, t., greenberg, j., & solomon, s. (1999). a dual-process model of defense against conscious and unconscious death-related thoughts: an extension of terror management theory. psychological review, 106, 835-845. rosnick, c. b., & reynolds, s. l. (2003). thinking ahead: factors associated with executing advance directives. journal of aging and health, 15, 409-429. salisbury, l. c., & nenkov, g. y. (2016). solving the annuity puzzle: the role of mortality salience in retirement savings decumulation decisions. journal of consumer psychology, 26, 417-425. salmond, s. w., & david, e. (2005). attitudes toward advance directives and advance directive completion rates. orthopedic nursing, 24, 117-127. sanders, m., & smith, s. (2016). can simple prompts increase bequest giving? field evidence from a legal call centre. journal of economic behavior & organization, 125, 179-191. schneiderman, l. j., kronick, r., kaplan, r. m., anderson, j. p., & langer, r. d. (1992). effects of offering advance directives on medical treatments and costs. annals of internal medicine, 117, 599-606. 98 r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 teno, j., lynn, j., connors, a. f. , jr., , wenger, n., phillips, r. s., alzola, c., murphy, d. p., desbiens, n., & knaus, w. a. (1997). the illusion of end-of-life resource savings with advance directives. journal of the american geriatrics society, 45, 513-518. walter, t. (1991). modern death: taboo or not taboo? sociology, 25, 293-310. yadav, k. n., gabler, n. b., cooney, e., kent, s., kim, j., herbst, n., mante, a., halpern, s. d., & courtright, k. r. (2017). approximately one in three us adults completes any type of advance directive for end-of-life care. health affairs, 36, 1244-1251. yazdanparas, a., & alhenawi, y. (2017). personality and borrowing behavior: an examination of the role of need for material resources and need for arousal traits on household’s borrowing decisions. financial services review, 26, 55-85. r. hussein, r. n. james iii / financial services review 29 (2021) 85–99 99 finser_23_3 does active management work? evidence from equity sector funds crystal y. lina,* aschool of business, eastern illinois university, charleston, il 61920, usa abstract this study presents considerable evidence that equity sector mutual funds, the nine fidelity select portfolios here, have provided better after-expense returns against broader market etf, spy, and their peer sector etfs, the nine select sector spdr funds, over the sample period 1999–2010. not only do they achieve higher nominal returns over the 12 years, except for few sector mutual funds, some of the funds also generate higher risk-adjusted returns measured by sharpe ratio and ! from various asset pricing models. more important, none of the sector mutual funds generates a significant negative ! for the sample period no matter that asset pricing model is used. the results suggest that actively managed sector funds be considered by individual investors and/or their financial planners for mutual fund selection. © 2014 academy of financial services. all rights reserved. jel classification: g11; g12 keywords: mutual fund performance; active mutual fund management; sector investing 1. introduction the mutual fund industry has seen tremendous growth in the past few decades. the 2013 investment company institute fact book shows that 44.4% of u.s. households owned mutual funds in 2012, up significantly from 4.6% in 1980 and 23.4% in 1990, modestly down from 48.6% in 2000. bulk of individual’s mutual fund assets, $5.4 trillion out of $11.1 trillion, are invested in equities. individual accounts hold 90.3% of total equity mutual fund assets ($5.9 trillion). with the expansion of defined contribution plan assets, almost threefold from $1.7 * corresponding author. tel.: !1-217-581-2227; fax: !1-217-581-6642. e-mail address: cylin@eiu.edu (c.y. lin) financial services review 23 (2014) 249–271 1057-0810/14/$ – see front matter © 2014 academy of financial services. all rights reserved. trillion in 1995 to $5.1 trillion in 2012, knowledge in equity mutual funds becomes more and more important for individual investors. today many retirement savings plans offer a broader range of investment vehicles for individual investors. sector funds appear on the investment menu for many plan participants. do they deserve individual investors’ attention? how is their historical performance against their benchmarks? should financial planners recommend such mutual funds to their clients? using sector funds with the longest history, the fidelity select portfolios and the select sector spdr funds, this article tries to answer these questions for individual investors and/or their financial planners. once investors decide to allocate their assets to the u.s. public equity market, they must evaluate possible ways to implement the allocation in their subportfolios. there are two dimensions of this implementation. the first one is indexing or not: do investors want to passively invest so that their returns closely track selected indexes and at the same time investors pay less fees and experience less turnover? or do investors want to actively manage their subportfolios either in-house or through external fund managers? sullivan and xiong (2012) estimate that $1.2 trillion out of $3.5 trillion assets in the u.s. equity mutual funds and etfs was passively managed as of september 2010; equity index mutual funds and equity etfs split the share of passively managed equity index funds.1 the second dimension of this implementation is whether to make allocation decisions at the sector level: do investors want to further divide stocks by sector/industry and set up weight limits to these groups? or do investors not care about sector issues at all. porter (1985) and mcgahan and porter (1997) make a strong case for sector investing: they demonstrate that a company’s performance is influenced by the growth and structure of its industry. in addition, groysberg et al., (2011) show that forecasted industry growth is the most important explanatory variable when analysts construct their forecasts on companies. in this article, i examine the performance of actively managed equity sector funds and their passively managed counterparties to assist investors in their decisions in implementing equity sector asset allocations. the research question i try to answer is: when an investor wants to use external fund managers to allocate assets among u.s. equity sectors, are sector index funds a better choice than actively managed funds or vice versa? two parallel analyses regarding equity sector fund performance are provided in this article. one is their performance against a broad u.s. equity market index fund, which tests the efficient market hypothesis (emh). this analysis serves as a general empirical study on u.s. equity market efficiency. the other is sector mutual funds’ performance against sector index funds, which tests emh in a smaller territory: equity sector. it is arguable that using a broad equity market index is not appropriate when evaluating a sector fund because it has a much smaller universe of securities from which to choose. an index for the same sector would be more appropriate when used as a benchmark. in reality, this is what many actively managed funds do: they select an index that best matches their investment universe as their benchmark. 2. sector mutual funds and etfs a variety of equity sector funds have been introduced to the market place during the past several decades. these sector funds allow investors to custom tailor asset allocations to fit 250 c.y. lin / financial services review 23 (2014) 249–271 their particular investment needs or goals. like their broad equity market counterparties, index sector funds emerged later than actively managed sector funds. two groups of sector funds with the longest history in each category are used in this study: the fidelity select portfolios and the select sector spdr funds. in addition, spdr s&p 500 etf is used as an investable benchmark for the study. 2.1. fidelity select portfolios the fidelity’s web site2 listed 38 mutual funds under its stock funds/sector funds category at the time of writing this article. the inception date of the earliest three sector funds (energy, healthcare, and technology) is july 14, 1981. i identified nine broader sector funds based on fund prospectus: select consumer discretionary portfolio (fscpx), select consumer staples portfolio (fdfax), select energy portfolio (fsenx), select financial services portfolio (fidsx), select health care portfolio (fsphx), select industrials portfolio (fcyix), select materials portfolio (fsdpx), select technology portfolio (fsptx), and select utilities portfolio (fsutx).3 the above funds match nine sector index funds that are described later. these sector mutual funds were started between july 1981 and march 1997. most of the other funds listed are narrower focused industry funds. according to fidelity’s fund prospectuses, these funds seek capital appreciation, invest in domestic and foreign issuers, normally invest primarily in common stocks, and invest at least 80% of assets in securities of companies principally engaged in the selected sector. in other words, these funds are actively managed. fidelity management & research company is the fund’s manager. the fidelity select portfolios have an expense ratio between 0.80% (healthcare) and 1.41% (materials) as of february 29, 2012. the portfolio turnover rate is between 35% (consumer staples) and 384% (financial services). the funds have net assets between $0.28 billion (consumer discretionary) and $2.50 billion (energy).4 fidelity charges a short-term redemption fee, 0.75%, when money is withdrawn from a sector fund within 30 days of purchase to reduce short-term mutual fund trading.5 all these sector funds are open to new investors. 2.2. select sector spdr funds the select sector spdr trust was organized as a massachusetts business trust on june 10, 1998. state street global advisors serves as the fund manager. the trust consists of nine separate investment portfolios (each a “select sector spdr fund”) incepted in december 1998: the consumer discretionary select sector spdr fund (xly), the consumer staples select sector spdr fund (xlp), the energy select sector spdr fund (xle), the financial select sector spdr fund (xlf), the health care select sector spdr fund (xlv), the industrial select sector spdr fund (xli), the materials select sector spdr fund (xlb), the technology select sector spdr fund (xlk), and the utilities select sector spdr fund (xlu). these sector funds seek to provide investment results that, before expenses, correspond generally to the price and yield performance of publicly traded equity securities of companies in certain “select sector indexes”: the consumer discretionary 251c.y. lin / financial services review 23 (2014) 249–271 select sector index, the consumer staples select sector index, the energy select sector index, the financial select sector index, the health care select sector index, the industrial select sector index, the materials select sector index, the technology select sector index, and the utilities select sector index. each stock in the s&p 500 is allocated to one and only one select sector index. the combined companies of the nine select sector indexes represent all of the companies in the s&p 500. that is, the select sector spdr funds unbundle the s&p 500.6 these passively managed sector funds use a replication strategy, attempting to track the performance of an unmanaged index of securities. according to select sector spdr fund annual report, the ratio of expenses to average net assets is 0.19% for each individual fund as of september 30, 2011. the turnover rate is between 3.20% (utilities) and 13.86% (materials). the funds have net assets between $1.64 billion (materials) and $6.64 billion (utilities).7 2.3. spdr s&p 500 etf the spdr s&p 500 etf (spy) is an exchange traded fund designed to generally correspond to the price and yield performance of the s&p 500 index. utilizing a full replication approach, the trust owns all 500 securities of the s&p 500 index in their approximate market capitalization weight. the fund was incepted on january 22, 1993. the portfolio has an expense ratio of 0.09%, a turnover rate of 3.72%, and $80.87 billion net assets as of september 30, 2011.8 both equity mutual funds and etfs are pooled investments that represent ownership in a basket of stocks. however, etfs can be traded like individual stocks. they are also shortable, marginable, and optionable. index etfs normally have lower fees by eliminating many of the operating, research, and transaction expenses incurred by active money managers. they also provide greater transparency: one can get a holding list more frequently than with mutual funds. for example, fidelity select portfolios publish monthly holdings whereas the select sector spdrs update their online information daily. 3. a first look at equity sector funds: raw returns my analysis starts in january 1999 and ends in december 2010,9 since the earliest price data available for select sector spdr funds is mid-december 1998. i downloaded price and dividend data from yahoo!finance web site.10 monthly, annual, and 12-year holding period return is calculated for each fund as: ri,t " pi,t # di,t pi,t"1 $ 1, (1) where ri,t is the return for fund i during period t, pi,t is the price for fund i at the end of period t, pi,t"1 is the price for fund i at the end of period t " 1, and di,t is the total dividend/cash distribution of fund i during period t. 252 c.y. lin / financial services review 23 (2014) 249–271 the purpose of this study is to assist investors implement asset allocations at the sector level, therefore, i use fund price instead of fund net asset value to calculate fund return. this return is net of expenses and is attainable. as argued by jones and wermers (2011), i compare actively managed sector mutual fund performance to their passive alternative and not to the index itself. for the same reason, i use spy as an investable broad u.s. equity market benchmark. 3.1. twelve-year return which group of funds generates higher returns during the sample period? the results are summarized in table 1. in the rest of the article, i use mf to represent fidelity select portfolios, etf to represent select sector spdr funds, and spy to represent the spdr s&p 500 etf for easier reference. panel a of table 1 lists the 12-year (1999–2010) holding period return for the 19 funds. the highest return is 295.2% for the energy mf and the lowest return is "17.4% for the technology etf, and the return for spy is 20.8%. seven of nine mfs, except for the utilities and consumer discretionary mf, have a higher return than that of their etf counterparties; the average outperformance is 51.2%. when spy is used as the benchmark, eight out of nine mfs (except utilities) and seven out of nine etfs (except financial and technology) outperform. the average mf holding period return across all sectors is 109.2% versus 58.0% for etfs, and the difference is significant at the 5% level using one-tailed t test. results are slightly different when i compound annual holding period returns through the 12 years. panel b of table 1 shows that the only underperforming mf against spy is financial, not utilities. the same seven mfs beat peer etfs with an average outperformance of 63.3%. the average mf 12-year return with annual compounding across all sectors is 127.6% versus 64.3% for etfs, and the difference is also significant at the 5% level using one-tailed t test. these results show that on average sector mfs outperform their etf counterparties for the whole sample period. none of the nine mfs suffers a loss during the sample period; however, the financial and technology etfs generate a negative return under both calculation methods. the tech bubble in early 2000s and the financial crisis in 2008 contribute to the negative 12-year return for these two sectors. although sector etfs mimicked their benchmark indexes during these bear markets, it seems sector mf managers made the right decisions against trend changes. fig. 1 shows that the annual return for the technology mf is 119.07%, "28.18%, and "31.70% for the year 1999, 2000, and 2001, whereas the annual return for its peer etf is 65.13%, "41.89%, and "23.34%, respectively. the annual return for the financial mf is "13.42%, "49.83%, and 23.71% for the year 2007, 2008, and 2009, whereas the annual return for its peer etf is "18.89%, "54.06%, and 16.98%, respectively. 3.2. average annual return seven out of nine mfs (except for utilities and consumer discretionary) outperform their peer etfs when measured with average annual holding period return depicted in fig. 2. an interesting finding in fig. 2 is that the ranking of performance is different for mfs and etfs. energy mf has the highest average annual return of 18.3%, followed by materials (17.0%), 253c.y. lin / financial services review 23 (2014) 249–271 technology (14.9%), industrials (11.5%), consumer staples (7.7%), healthcare (6.3%), utilities (4.7%), consumer discretionary (4.6%), and financial (3.4%). in the etf group, energy (13.8%) and materials (9.7%) are still ranked first and second, whereas financial (1.9%) is again at the bottom. however, the other six sectors are ranked differently. fig. 2 also shows that for the materials, industrials, technology, and consumer staples sector, mf average annual holding period returns are higher than that of corresponding etfs at the 1% or 5% significance level using one-tailed t test. table 1 twelve-year return, 1999–2010 sector mf etf mf beats etf difference mf beats spy etf beats spy a: holding period return b 287.6% 105.1% yes 182.5% yes yes e 295.2% 221.3% yes 73.9% yes yes f 23.4% "8.0% yes 31.4% yes no i 125.9% 62.1% yes 63.8% yes yes k 73.5% "17.4% yes 90.9% yes no p 82.0% 26.4% yes 55.7% yes yes u 20.6% 41.3% no "20.7% no yes v 50.4% 36.2% yes 14.2% yes yes y 24.0% 55.0% no "30.9% yes yes average fund return 109.2% 58.0% p-value average mf return higher than average etf return 0.022 spy 20.8% b: return with annual compounding b 325.9% 118.7% yes 207.2% yes yes e 354.3% 243.2% yes 111.0% yes yes f 8.4% "13.7% yes 22.0% no no i 148.8% 67.8% yes 81.0% yes yes k 70.5% "16.2% yes 86.7% yes no p 111.2% 31.6% yes 79.6% yes yes u 25.7% 51.8% no "26.1% yes yes v 67.6% 37.4% yes 30.1% yes yes y 36.4% 58.4% no "21.9% yes yes average fund return 127.6% 64.3% p-value average mf return higher than average etf return 0.015 spy 23.9% note: b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. holding period return in panel a is calculated by adding ending price and all dividends paid in 12 years then divided by beginning price. return with annual compounding in panel b is calculated by compounding annual holding period returns for each individual fund. 254 c.y. lin / financial services review 23 (2014) 249–271 -100% -50% 0% 50% 100% materials -60% -40% -20% 0% 20% 40% 60% energy -60% -40% -20% 0% 20% 40% 60% industrials -40% -20% 0% 20% 40% consumer staples -60% -40% -20% 0% 20% 40% healthcare -60% -40% -20% 0% 20% 40% financial -60% -40% -20% 0% 20% 40% 60% 80% 100% 120% 140% technology -40% -20% 0% 20% 40% utilities -60% -40% -20% 0% 20% 40% 60% consumer discretionary mf etf spy fig. 1. annual holding period return. 255c.y. lin / financial services review 23 (2014) 249–271 when annual holding period return is compared between mfs and etfs, for the same sector, table 2 shows that seven out of nine mfs (except utilities and consumer discretionary) generate a higher return in seven or more years in the 12 year sample period. the average number of years of outperforming is 7.9 and the percentage of years of outperforming is 66%. using spy as the benchmark, on average in 7.4 out of 12 years mfs beat spy (62% of years); only in 6.4 years does the etf group beat spy (54% of years). table 2 annual return comparison, 1999–2010 sector mf beats etf mf beats spy etf beats spy number of years % of years number of years % of years number of years % of years b 9 75% 10 83% 9 75% e 7 58% 9 75% 7 58% f 9 75% 6 50% 7 58% i 10 83% 9 75% 7 58% k 8 67% 6 50% 4 33% p 10 83% 9 75% 5 42% u 6 50% 6 50% 6 50% v 7 58% 7 58% 5 42% y 5 42% 5 42% 8 67% average 7.9 66% 7.4 62% 6.4 54% note: b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. total number of years: 12. p-value average mf annual holding period return higher than average etf annual holding period return b e f i k p u v y 0.012 0.082 0.092 0.002 0.046 0.013 0.430 0.300 0.222 0% 5% 10% 15% 20% b e f i k p u v y mf etf spy fig. 2. arithmetic mean of annual holding period return. b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. 256 c.y. lin / financial services review 23 (2014) 249–271 3.3. decomposition of 12-year holding period return what portion of that 12-year holding period return is contributed by capital gains? what portion is contributed by dividend yield (regular dividend and special cash distribution)? there are four mfs (financial, utilities, healthcare, and consumer discretionary) and two etfs (financial and technology) have negative capital gains during the period. the spy has a capital gains yield of 2.0%. actually, only three mfs (materials, energy, and industrials) have a higher capital gains yield than dividend yield. that number is five for etfs (materials, energy, industrials, healthcare, and consumer discretionary). the spy has a dividend yield of 18.8%. with 2.0% capital gains yield from spy, the 12-year sample period is pretty flat. the s&p 500 price index has two peaks, 1552.87 on march 24, 2000 and 1576.09 on october 17, 2007, and two troughs, 768.63 on october 10, 2002 and 666.79 on march 6, 2009. this range provides a good testing field for performance analysis. after examining 12-year returns and average annual returns, i find that most sector mfs outperform both their etf peers and spy for most years during the 1999–2010 sample period. the exceptions are the utilities and consumer discretionary mf. 4. a closer look at equity sector funds: risk adjusted returns before drawing a conclusion on sector mfs’ performance, one must investigate the risk dimension of the returns. here, i look at risk adjusted returns that incorporate both total risk and systematic risk based on monthly holding period returns. table 3 summarizes statistics of both monthly holding period returns and excess returns, which are calculated as monthly holding period return minus monthly 1-month t-bill rate. panel a shows the maximum monthly holding period return is 31.446% (technology mf, february 2000) and the minimum monthly return is "28.379% (technology mf, february 2001). because sector funds focus on specific investment areas, it is expected that both sector mfs and etfs have a higher volatility when compared to a broader market benchmark. that table 3 summary statistics of monthly return (%) maximum minimum mean median sd skewness kurtosis n sharpe ratio a: raw return mf 31.446 "28.379 0.695 1.014 6.316 "0.238 2.738 1289 etf 24.768 "26.198 0.468 0.796 5.982 "0.255 1.762 1296 spy 9.935 "16.519 0.260 0.741 4.652 "0.485 0.594 144 overall 31.446 "28.379 0.564 0.897 6.081 "0.246 2.326 2729 b: excess return mf 31.016 "28.769 0.474 0.759 6.321 "0.229 2.696 1289 0.075 etf 24.628 "26.198 0.245 0.533 5.985 "0.246 1.730 1296 0.041 spy 9.925 "16.599 0.038 0.526 4.660 "0.454 0.535 144 0.008 overall 31.016 "28.769 0.342 0.652 6.085 "0.236 2.288 2729 0.056 257c.y. lin / financial services review 23 (2014) 249–271 is true as shown in table 3. the sector mfs have the widest span of monthly returns (59.825%), followed by sector etfs (50.966%); both are much higher than that of spy (26.454%). the standard deviation for sector mf, etf, and spy is 6.316%, 5.982%, and 4.652%, respectively. all fund returns are negatively skewed. these characteristics are similar in excess returns as presented in panel b. the group sharpe ratio is 0.075, 0.041, 0.008, and 0.056 for the mfs, etfs, spy, and overall funds, respectively. the sector mf group has the highest sharpe ratio. these small but positive sharpe ratios reflect the flat u.s. equity market during the sample period. 4.1. sharpe ratio does the sharpe ratio comparison between each pair of sector funds echo the group sharpe ratio results in table 3? table 4 shows that for the full sample period, january 1999 through december 2010, seven out of nine sector mfs have a higher sharpe ratio than that of their peer etfs, except for the utilities and consumer discretionary mf. the materials table 4 individual fund sharpe ratio, january 1999 through december 2010 fund mean excess return sd sharpe ratio mf sr higher than etf mf sr higher than spy etf sr higher than spy bbb 1.030 6.786 0.152 yes yes xlb 0.555 6.760 0.082 yes eee 1.100 7.390 0.149 yes yes xle 0.853 6.504 0.131 yes fff 0.035 6.224 0.006 yes no xlf "0.089 6.886 "0.013 no iii 0.669 6.157 0.109 yes yes xli 0.314 5.868 0.054 yes kkk 0.671 10.126 0.066 yes yes xlk "0.015 8.069 "0.002 no ppp 0.367 3.728 0.098 yes yes xlp 0.042 3.709 0.011 yes uuu 0.066 5.021 0.013 no yes xlu 0.184 4.671 0.039 yes vvv 0.224 4.175 0.054 yes yes xlv 0.093 4.280 0.022 yes yyy 0.112 4.896 0.023 no yes xly 0.272 5.899 0.046 yes spy 0.038 4.660 0.008 average mf sr 0.074 average etf sr 0.041 p-value average mf sr higher than average etf sr 0.020 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. xlb, xle, xlf, xli, xlk, xlp, xlu, xlv, and xly represent the select sector spdr etfs for these nine sectors, respectively. spy represents the spdr s&p 500 trust. 258 c.y. lin / financial services review 23 (2014) 249–271 and energy mf have the highest sharpe ratios, 0.152 and 0.149, respectively; whereas the financial and technology etf have a negative sharpe ratio for the same period. the average mf sharpe ratio across all sectors is 0.074 versus 0.041 for sector etfs, and the difference is significant at the 5% level using one-tailed t test. when the sample period is divided into two equal-length subperiods, the results are slightly different.11 for the first half sample period, from january1999 through december 2004, the healthcare mf does not have a higher sharpe ratio than its peer etf. for the second half sample period, from january 2005 through december 2010, the energy mf does not have a higher sharpe ratio, but the consumer discretionary mf does outperform its peer etf. subsample analysis also shows that the significant higher mf average sharpe ratio for the whole sample period is mainly because of higher mf average sharpe ratio in the second half sample period, which is significant at the 1% level, whereas the p-value is greater than 5% for the first half sample period. table 4 also compares sharpe ratio between individual sector funds and spy. only the financial mf underperforms spy for the full sample period, whereas both the financial and technology etf underperform spy for the same time period. using total risk as the measurement, i find that most sector mfs outperform their peer etfs. sector mfs also have a higher number of funds outperform spy. 4.2. performance against s&p 500 etf systematic risk is always the part of risk that gets more attention because many argue that unsystematic risk can be diversified away at a relatively low cost.12 capm has been the standard model to test fund performance. i modify the model by replacing excess market return with excess spy return because an investable benchmark makes more sense for comparing attainable returns: ri,t $ rf,t " !i # %i#rspy,t $ rf,t$ # &i,t (2) where ri,t is the return of fund i in month t, rf,t is the return of one-month t-bill in month t, rspy,t is the spy return in month t, and &i,t is an error term. table 5 shows that the materials mf generates a 0.988% monthly abnormal return (11.856% annually) and the energy mf generates a 1.064% monthly abnormal return (12.768% annually) during the sample period, significant at the 1% and 5% level, respectively. the industrial etf has a significant ! of 0.672% at the 1% level. all %s are significant at the 1% level. both mfs and etfs for the materials sector, the financial sector, the industrials sector, and the technology sector have a % greater than 1. all other funds have a % less than 1 except for the consumer discretionary etf. the average adjusted r2 is 0.563 for the mfs and 0.555 for the etfs. for the first half sample period, none of the !s is significant at the 5% level. the average adjusted r2 declines to 0.438 for the mfs and 0.461 for the etfs. the average adjusted r2 are higher for the second half sample period: 0.719 for the mfs and 0.692 for the etfs. the materials mf, the consumer staples mf, and the industrials etf generate 0.940%, 0.495%, and 0.543% monthly abnormal returns, respectively, from january 2005 through december 2010, at the 5% level. 259c.y. lin / financial services review 23 (2014) 249–271 fama-french three-factor model is also modified by replacing excess market return with excess spy return: ri,t $ rf,t " !i # %i#rspy,t $ rf,t$ # hihmlt # sismbt # &i,t (3) where hml (high minus low) is the average return on two value portfolios minus the average return on two growth portfolios and smb (small minus big) is the average return on three small portfolios minus the average return on three big portfolios (fama and french 1993). the data are downloaded from kenneth r. french data library.13 results in table 6 show that over the full sample period two sector mfs (materials and technology) have a positive ! at the 5% level. none of the etfs has a significant ! at the 5% level. all %s are significant at the 1% level with the same above/below 1 % distribution as in table 5. hml is not significant for the healthcare mf and etf, and it is significantly negative for three funds (technology mf and etf, and utilities mf). the smb, size factor, is the least significant factor of the three: only seven out of 18 funds have a significant coefficient. during the first half sample period, none of the mfs or etfs generates a positive ! at the 5% level. fewer hml and smb coefficients are significant. during the second half sample table 5 one factor model results using spy excess return, january 1999 through december 2010 fund ! % adjusted r2 estimate se t-value estimate se t-value bbb 0.988 0.362 2.727*** 1.121 0.078 14.379*** 0.590 xlb 0.512 0.347 1.475 1.145 0.075 15.321*** 0.620 eee 1.064 0.491 2.164** 0.962 0.106 9.086*** 0.363 xle 0.821 0.436 1.882 0.834 0.094 8.880*** 0.352 fff "0.005 0.309 "0.018 1.074 0.067 16.126*** 0.644 xlf "0.134 0.345 "0.387 1.183 0.074 15.921*** 0.638 iii 0.272 0.230 1.183 1.164 0.052 22.589*** 0.789 xli 0.672 0.242 2.782*** 1.112 0.050 22.444*** 0.778 kkk 0.606 0.522 1.162 1.711 0.112 15.228*** 0.617 xlk "0.071 0.359 "0.197 1.466 0.077 18.977*** 0.715 ppp 0.349 0.248 1.407 0.485 0.053 9.092*** 0.363 xlp 0.025 0.256 0.097 0.451 0.055 8.186*** 0.315 uuu 0.035 0.278 0.127 0.807 0.060 13.463*** 0.557 xlu 0.165 0.339 0.486 0.497 0.073 6.807*** 0.240 vvv 0.202 0.268 0.755 0.574 0.058 9.950*** 0.406 xlv 0.066 0.223 0.296 0.718 0.048 14.963*** 0.609 yyy 0.078 0.209 0.374 0.903 0.045 20.058*** 0.737 xly 0.231 0.256 0.902 1.082 0.055 19.611*** 0.728 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. xlb, xle, xlf, xli, xlk, xlp, xlu, xlv, and xly represent the select sector spdr etfs for these nine sectors, respectively. spy represents the spdr s&p 500 trust. ***significant at the 1% level. **significant at the 5% level. 260 c.y. lin / financial services review 23 (2014) 249–271 t ab le 6 t hr ee fa ct or re gr es si on re su lts us in g sp y ex ce ss re tu rn ,j an ua ry 19 99 th ro ug h d ec em be r 20 10 fu nd ! % h m l sm b a dj us te d r 2 e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e b b b 0. 71 9 0. 34 0 2. 11 7* * 1. 14 5 0. 07 2 15 .9 01 ** * 0. 50 4 0. 09 3 5. 39 6* ** 0. 13 7 0. 09 1 1. 50 0 0. 65 6 x l b 0. 32 7 0. 33 0 0. 99 0 1. 17 4 0. 07 0 16 .7 50 ** * 0. 43 0 0. 09 1 4. 73 6* ** 0. 04 0 0. 08 9 0. 44 8 0. 67 2 e e e 0. 92 5 0. 49 7 1. 86 1 0. 98 2 0. 10 5 9. 31 4* ** 0. 31 4 0. 13 7 2. 29 9* * 0. 03 6 0. 13 4 0. 26 6 0. 37 9 x l e 0. 73 3 0. 43 8 1. 67 5 0. 86 1 0. 09 3 9. 27 3* ** 0. 29 6 0. 12 0 2. 45 7* * " 0. 04 1 0. 11 8 " 0. 35 1 0. 37 8 ff f " 0. 20 9 0. 24 7 " 0. 84 5 1. 12 4 0. 05 2 21 .4 22 ** * 0. 59 5 0. 06 8 8. 74 7* ** " 0. 03 6 0. 06 6 " 0. 54 0 0. 78 3 x l f " 0. 35 2 0. 26 5 " 1. 32 6 1. 24 3 0. 05 6 22 .0 85 ** * 0. 68 7 0. 07 3 9. 41 2* ** " 0. 07 1 0. 07 1 " 0. 99 0 0. 79 6 ii i 0. 37 3 0. 20 5 1. 81 8 1. 17 2 0. 04 3 27 .3 37 ** * 0. 45 4 0. 05 6 8. 05 7* ** 0. 17 7 0. 05 5 3. 20 4* ** 0. 85 6 x l i 0. 11 8 0. 21 2 0. 55 5 1. 13 3 0. 04 5 25 .2 00 ** * 0. 33 8 0. 05 8 5. 79 3* ** 0. 04 7 0. 05 7 0. 83 0 0. 82 0 k k k 0. 61 4 0. 27 1 2. 26 2* * 1. 56 5 0. 05 8 27 .1 83 ** * " 1. 01 4 0. 07 5 " 13 .5 80 ** * 0. 65 0 0. 07 3 8. 90 3* ** 0. 90 1 x l k 0. 12 6 0. 22 4 0. 56 4 1. 38 6 0. 04 7 29 .1 97 ** * " 0. 80 4 0. 06 2 " 13 .0 57 ** * 0. 18 4 0. 06 0 3. 06 1* ** 0. 89 4 pp p 0. 23 9 0. 21 5 1. 11 3 0. 52 2 0. 04 6 11 .4 61 ** * 0. 39 0 0. 05 9 6. 60 9* ** " 0. 06 5 0. 05 8 " 1. 12 4 0. 54 4 x l p 0. 05 1 0. 23 1 0. 21 9 0. 49 2 0. 04 9 10 .0 34 ** * 0. 25 0 0. 06 4 3. 93 6* ** " 0. 20 8 0. 06 2 " 3. 35 5* ** 0. 46 7 u u u 0. 17 5 0. 27 9 0. 62 7 0. 80 0 0. 05 9 13 .5 28 ** * " 0. 22 4 0. 07 7 " 2. 92 4* ** " 0. 09 5 0. 07 5 " 1. 27 0 0. 57 7 x l u 0. 13 9 0. 31 4 0. 44 2 0. 54 5 0. 06 7 8. 18 8* ** 0. 36 0 0. 08 6 4. 17 3* ** " 0. 19 0 0. 08 4 " 2. 25 7* * 0. 38 1 v v v 0. 17 4 0. 27 6 0. 63 0 0. 57 9 0. 05 8 9. 90 5* ** 0. 06 9 0. 07 6 0. 90 9 0. 00 5 0. 07 4 0. 06 5 0. 40 2 x l v 0. 15 6 0. 22 7 0. 68 8 0. 72 6 0. 04 8 15 .0 86 ** * " 0. 05 9 0. 06 2 " 0. 94 7 " 0. 11 8 0. 06 1 " 1. 93 1 0. 61 4 y y y " 0. 18 3 0. 18 7 " 0. 98 0 0. 90 2 0. 04 0 22 .7 62 ** * 0. 32 2 0. 05 1 6. 26 8* ** 0. 24 3 0. 05 0 4. 83 2* ** 0. 80 0 x l y 0. 03 1 0. 24 6 0. 12 5 1. 09 0 0. 05 2 20 .8 55 ** * 0. 30 9 0. 06 8 4. 55 3* ** 0. 14 6 0. 06 6 2. 19 9* * 0. 76 0 n ot e: b b b ,e e e ,f ff ,i ii ,k k k ,p pp ,u u u ,v v v ,a nd y y y re pr es en t th e fi de lit y se le ct po rt fo lio m ut ua l fu nd s fo r th e m at er ia ls se ct or ,t he e ne rg y se ct or , th e fi na nc ia l se ct or , th e in du st ri al s se ct or , th e t ec hn ol og y se ct or , th e c on su m er st ap le s se ct or , th e u til iti es se ct or , th e h ea lth ca re se ct or , an d th e c on su m er d is cr et io na ry se ct or ,r es pe ct iv el y. x l b ,x l e ,x l f, x l i, x l k ,x l p, x l u ,x l v ,a nd x l y re pr es en t th e se le ct se ct or sp d r e t fs fo r th es e ni ne se ct or s, re sp ec tiv el y. sp y re pr es en ts th e sp d r s& p 50 0 t ru st . ** *s ig ni fic an t at th e 1% le ve l. ** si gn ifi ca nt at th e 5% le ve l. 261c.y. lin / financial services review 23 (2014) 249–271 period, three sector mfs (materials, industrials, and consumer staples) have a significant positive !. the financial etf generates a significant negative ! at the 5% significance level. the four-factor model is modified by replacing excess market return with excess spy return: ri,t $ rf,t " !i # %i#rspy,t $ rf,t$ # hihmlt # sismbt # mimomt # &i,t (4) where mom (momentum) is the average return on the two high prior return portfolios minus the average return on the two low prior return portfolios (see carhart 1997). the data are downloaded from kenneth r. french data library. the results presented in table 7 resemble those from the three-factor model. table 7 shows that over the full sample period the same two sector mfs (materials and technology) have a positive ! at the 5% level. for the momentum factor, mom, only five out of 18 funds have a significant coefficient. generally speaking, more often, sector mfs generate significant higher !s than peer etfs no matter whether a one-factor, three-factor, or four-factor model is adopted when spy is used as the proxy for market portfolio. all fund returns are sensitive to the overall equity market movements as the % coefficients for excess spy return are all significant at the 1% level. fund returns are less sensitive to the value/growth factor, the size factor, and the momentum factor in the order of listing. the findings on performance against spy are not consistent with emh. they support the argument by kacperczyk, sialm, and zheng (2005) that more sector/industry concentrated funds perform better. 4.3. sector mutual fund performance against sector index funds let us examine mf performance against peer etf within each sector. excess return of peer etf is used as the independent variable for the one-factor model: rmfj,t $ rf,t " !j # %j#retfj,t $ rf,t$ # &j,t (5) where rmfj,t is the return of mf of sector j in month t, and retfj,t is the return of etf in the same sector j in month t. table 8 reports that three sector mfs (materials, industrials, and technology) generate significant positive !s, 0.506%, 0.446%, and 0.689%, respectively, during the full sample period. that is equivalent to annualized outperformance of 6.072%, 5.352%, and 8.268%, respectively. subsample analysis shows that none of the sector mfs outperforms in the first half sample period, whereas the materials and the industrials mf outperform in the second half sample period. all %s are positive and significant at the 1% level. which model can explain most of the return variances of sector funds? i summarize the adjusted r2 for the four models in table 9. most of the time, adding hml and smb does increase the adjusted r2 when using spy as the benchmark. only one out of 18 regressions for the full sample period suffers a slight decrease of explaining power, 0.004. adding mom, however, does not increase adjusted r2 across the board. over the full sample period, the average adjusted r2 across all funds is 0.559, 0.649, 0.653, and 0.764 for the one/three/four-factor model using spy and one-factor 262 c.y. lin / financial services review 23 (2014) 249–271 t ab le 7 fo ur fa ct or re gr es si on re su lts us in g ex ce ss sp y re tu rn ,j an ua ry 19 99 th ro ug h d ec em be r 20 10 fu nd ! % h m l sm b m o m a dj us te d r 2 e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e e st im at e se tva lu e b b b 0. 70 6 0. 34 1 2. 07 2* * 1. 17 1 0. 08 1 14 .5 04 ** * 0. 51 5 0. 09 5 5. 43 1* ** 0. 12 8 0. 09 2 1. 38 7 0. 04 2 0. 05 8 0. 71 7 0. 65 5 x l b 0. 31 6 0. 33 2 0. 95 3 1. 19 6 0. 07 9 15 .2 14 ** * 0. 44 0 0. 09 2 4. 76 4* ** 0. 03 2 0. 09 0 0. 35 9 0. 03 5 0. 05 7 0. 62 7 0. 67 e e e 0. 87 8 0. 49 4 1. 77 8 1. 07 7 0. 11 7 9. 19 8* ** 0. 35 6 0. 13 8 2. 58 6* * 0. 00 3 0. 13 4 0. 02 4 0. 15 2 0. 08 4 1. 80 4 0. 38 9 x l e 0. 68 4 0. 43 3 1. 58 2 0. 96 1 0. 10 3 9. 36 2* ** 0. 33 9 0. 12 1 2. 81 4* ** " 0. 07 5 0. 11 7 " 0. 64 3 0. 15 9 0. 07 4 2. 15 9* * 0. 39 4 ff f " 0. 18 8 0. 24 6 " 0. 76 3 1. 08 0 0. 05 8 18 .5 13 ** * 0. 57 6 0. 06 9 8. 40 2* ** " 0. 02 1 0. 06 7 " 0. 31 5 " 0. 07 0 0. 04 2 " 1. 65 8 0. 78 6 x l f " 0. 31 9 0. 26 1 " 1. 22 1 1. 17 7 0. 06 2 18 .9 88 ** * 0. 65 8 0. 07 3 9. 03 7* ** " 0. 04 8 0. 07 1 " 0. 67 7 " 0. 10 6 0. 04 5 " 2. 38 1* * 0. 80 2 ii i 0. 37 0 0. 20 6 1. 79 6 1. 17 8 0. 04 8 24 .3 18 ** * 0. 45 6 0. 05 7 7. 99 0* ** 0. 17 4 0. 05 6 3. 11 6* ** 0. 00 9 0. 03 5 0. 25 3 0. 85 5 x l i 0. 13 2 0. 21 2 0. 62 3 1. 10 4 0. 05 0 21 .9 92 ** * 0. 32 5 0. 05 9 5. 51 3* ** 0. 05 7 0. 05 7 0. 99 6 " 0. 04 6 0. 03 6 " 1. 27 3 0. 82 1 k k k 0. 61 0 0. 27 3 2. 23 5* * 1. 57 5 0. 06 5 24 .3 50 ** * " 1. 01 0 0. 07 6 " 13 .2 95 ** * 0. 64 7 0. 07 4 8. 75 1* ** 0. 01 5 0. 04 7 0. 31 6 0. 90 0 x l k 0. 15 9 0. 21 8 0. 72 8 1. 31 9 0. 05 2 25 .4 62 ** * " 0. 83 3 0. 06 1 " 13 .6 80 ** * 0. 20 7 0. 05 9 3. 49 6* ** " 0. 10 7 0. 03 7 " 2. 86 8* ** 0. 89 9 pp p 0. 22 2 0. 21 4 1. 03 9 0. 55 5 0. 05 1 10 .9 42 ** * 0. 40 5 0. 06 0 6. 78 8* ** " 0. 07 6 0. 05 8 " 1. 31 6 0. 05 4 0. 03 7 1. 46 9 0. 54 8 x l p 0. 03 5 0. 23 1 0. 15 3 0. 52 3 0. 05 5 9. 55 8* ** 0. 26 4 0. 06 4 4. 10 2* ** " 0. 21 9 0. 06 3 " 3. 50 4* ** 0. 05 0 0. 03 9 1. 27 7 0. 46 9 u u u 0. 15 8 0. 27 9 0. 56 5 0. 83 4 0. 06 6 12 .6 28 ** * " 0. 20 9 0. 07 8 " 2. 69 2* ** " 0. 10 7 0. 07 6 " 1. 41 7 0. 05 6 0. 04 8 1. 17 0 0. 57 8 x l u 0. 11 4 0. 31 3 0. 36 5 0. 59 5 0. 07 4 8. 02 9* ** 0. 38 2 0. 08 7 4. 38 6* ** " 0. 20 8 0. 08 5 " 2. 45 0* * 0. 08 1 0. 05 3 1. 51 4 0. 38 7 v v v 0. 14 5 0. 27 3 0. 53 0 0. 63 8 0. 06 5 9. 85 5* ** 0. 09 5 0. 07 6 1. 24 3 " 0. 01 5 0. 07 4 " 0. 20 7 0. 09 4 0. 04 7 2. 02 3* * 0. 41 5 x l v 0. 16 3 0. 22 8 0. 71 6 0. 71 2 0. 05 4 13 .1 81 ** * " 0. 06 5 0. 06 3 " 1. 03 0 " 0. 11 3 0. 06 2 " 1. 82 9 " 0. 02 3 0. 03 9 " 0. 59 2 0. 61 2 y y y " 0. 17 6 0. 18 7 " 0. 94 1 0. 88 9 0. 04 4 19 .9 86 ** * 0. 31 6 0. 05 2 6. 05 3* ** 0. 24 8 0. 05 1 4. 87 3* ** " 0. 02 2 0. 03 2 " 0. 69 5 0. 79 9 x l y 0. 06 3 0. 24 2 0. 26 2 1. 02 4 0. 05 7 17 .8 48 ** * 0. 28 0 0. 06 7 4. 15 0* ** 0. 16 8 0. 06 6 2. 56 8* * " 0. 10 6 0. 04 1 " 2. 56 9* * 0. 76 9 n ot e: b b b ,e e e ,f ff ,i ii ,k k k ,p pp ,u u u ,v v v ,a nd y y y re pr es en t th e fi de lit y se le ct po rt fo lio m ut ua l fu nd s fo r th e m at er ia ls se ct or ,t he e ne rg y se ct or , th e fi na nc ia l se ct or , th e in du st ri al s se ct or , th e t ec hn ol og y se ct or , th e c on su m er st ap le s se ct or , th e u til iti es se ct or , th e h ea lth ca re se ct or , an d th e c on su m er d is cr et io na ry se ct or ,r es pe ct iv el y. x l b ,x l e ,x l f, x l i, x l k ,x l p, x l u ,x l v ,a nd x l y re pr es en t th e se le ct se ct or sp d r e t fs fo r th es e ni ne se ct or s, re sp ec tiv el y. sp y re pr es en ts th e sp d r s& p 50 0 t ru st . ** *s ig ni fic an t at th e 1% le ve l. ** si gn ifi ca nt at th e 5% le ve l. 263c.y. lin / financial services review 23 (2014) 249–271 model using peer etf, respectively. on average, peer etf benchmarking provides the best model fit for fund returns. this result is similar with dellva, demaskey, and smith (2001) and kaushik, pennathur, and barnhart (2010). table 8 one factor model results using peer etf excess return, january 1999 through december 2010 fund ! % adjusted r2 estimate se t-value estimate se t-value bbb 0.506 0.196 2.588** 0.943 0.029 32.566*** 0.881 eee 0.167 0.167 0.998 1.095 0.026 42.875*** 0.927 fff 0.113 0.137 0.820 0.872 0.020 43.574*** 0.929 iii 0.446 0.158 2.815*** 1.004 0.027 36.969*** 0.909 kkk 0.689 0.306 2.249** 1.170 0.038 30.712*** 0.868 ppp 0.332 0.175 1.895 0.831 0.047 17.529*** 0.681 uuu "0.064 0.316 "0.204 0.710 0.068 10.480*** 0.432 vvv 0.166 0.269 0.618 0.623 0.063 9.901*** 0.404 yyy "0.095 0.160 "0.595 0.764 0.027 28.011*** 0.845 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. ***significant at the 1% level. **significant at the 5% level. table 9 adjusted r2 comparison for different models, january 1999 through december 2010 fund 1-factor model (spy) 3-factor model (spy) 4-factor model (spy) 1-factor model (peer etf) bbb 0.590 0.656 0.655 0.881 xlb 0.620 0.672 0.670 eee 0.363 0.379 0.389 0.927 xle 0.352 0.378 0.394 fff 0.644 0.783 0.786 0.929 xlf 0.638 0.796 0.802 iii 0.789 0.856 0.855 0.909 xli 0.778 0.820 0.821 kkk 0.617 0.901 0.900 0.868 xlk 0.715 0.894 0.899 ppp 0.363 0.544 0.548 0.681 xlp 0.315 0.467 0.469 uuu 0.557 0.577 0.578 0.432 xlu 0.240 0.381 0.387 vvv 0.406 0.402 0.415 0.404 xlv 0.609 0.614 0.612 yyy 0.737 0.800 0.799 0.845 xly 0.728 0.760 0.769 note: bbb, eee, fff, iii, kkk, ppp, uuu, vvv, and yyy represent the fidelity select portfolio mutual funds for the materials sector, the energy sector, the financial sector, the industrials sector, the technology sector, the consumer staples sector, the utilities sector, the healthcare sector, and the consumer discretionary sector, respectively. xlb, xle, xlf, xli, xlk, xlp, xlu, xlv, and xly represent the select sector spdr etfs for these nine sectors, respectively. spy represents the spdr s&p 500 trust. 264 c.y. lin / financial services review 23 (2014) 249–271 comparing the adjusted r2 of one-factor model with peer etf to that of one-factor model with spy, i find that seven out of nine mf regressions using peer etf have a higher explanation power for the full and first half sample periods with the exceptions of the utilities sector and the healthcare sector. all nine mf regressions improve their adjusted r2 for the second half sample period using peer etf benchmark. unlike previous research on sector mutual fund performance, this article is one of the first that provides detailed analysis on individual sector funds. zheng and tower (2005), for example, examine performance of asset-weighted and equal-weighted fidelity sector fund portfolios and find that they performed less well than corresponding indexes. dellva, demaskey, and smith (2001) only list number of sector funds in their analysis. kaushik, pennathur, and barnhart (2010) report sector aggregate performance results. for individual investors and/or their financial planners, however, it is important to examine individual sector fund performance and allocate their assets accordingly. 4.4. changing dynamic do mf and etf in the same sector move together? are they highly correlated with the general market index? pairwise correlation is charted in fig. 3. in all but two cases (utilities and healthcare) the correlation between mf and etf in the same sector, the black bar, is the highest compared with the correlation between mf/etf and spy. six sectors have a correlation between mf and etf higher than 0.9 (materials, energy, financial, industrials, technology, and consumer discretionary), whereas the other three sectors have a correlation of 0.827 (consumer staples), 0.661 (utilities), and 0.638 (healthcare). an interesting find was that these three sectors are often considered defensive sectors. common factors in one sector appear to affect the returns of the mf and etf in this sector more than that of general factors affecting the overall equity market. the returns of mfs and etfs in the same sector tend to go hand-in-hand. the regression results during different sample periods discussed earlier, however, hint that 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 b e f i k p u v y mf vs. etf mf vs. spy etf vs. spy fig. 3. monthly return correlation between funds, 1999–2010. b represents the materials sector; e represents the energy sector; f represents the financial sector; i represents the industrials sector; k represents the technology sector; p represents the consumer staples sector; u represents the utilities sector; v represents the healthcare sector; y represents the consumer discretionary sector; and spy represents the spdr s&p 500 trust. 265c.y. lin / financial services review 23 (2014) 249–271 the correlation might not be stable throughout the whole sample period. fig. 4 plots 36-month rolling correlation for the nine sectors. consistent with the results in fig. 3, for most of the sample period, in all but two sectors (utilities and healthcare), the correlation between mf and etf in the same sector, the black line, is the highest compared with the correlation between mf/etf and spy. the industrial sector, the technology sector, and the consumer discretionary sector have the most consistent correlation among three pairs of correlation plotted, mf versus etf, mf versus spy, and etf versus spy. the three lines are close to each other with the black one on the top. however, the utilities sector and the healthcare sector show great time varying correlation. the lines cross each other and the spread is huge. for example, for the healthcare sector, the correlation between the etf and spy was the highest (0.864 vs. 0.127 and 0.189) for the first 36 months, january 1999 through december 2001, but it turns out to be the lowest (0.817 vs. 0.931 and 0.869) for the last 36 months, january 2008 through december 2010. a noticeable phenomenon is that the correlations tend to converge overtime. almost all sectors have tighter correlation spreads moving into the end of the sample period. the correlation between the mf/etf and spy, the gray dashed line and the black dotted line, almost overlap for most of the sectors during the last quarter of the chart period. this changing dynamic explains why regression based results are sensitive to sample period selection. the open question is whether the correlation convergence will continue into the future. if yes, the diversification benefit one can enjoy through sector investing may diminish overtime. 5. discussion 5.1. benchmarking the fidelity select portfolios cited both the s&p 500 and a msci u.s. im sector 25/50 index in the “management’s discussion of fund performance” section of the annual reports. however, the earliest etfs based on the msci u.s. im sector 25/50 indexes were launched in january 2004 by vanguard, which is five years later than the select sector spdr funds. the price correlation between these two etfs in the same sector ranges from 0.975 to 0.999, with seven out of nine above 0.99, for the period from 2004 through 2010. this high correlation justifies the use of the select sector spdr funds as benchmarks for performance evaluation. one obvious explanation of the results that sector mfs outperform their peer etfs somehow is that these actively managed mutual funds can invest outside of the s&p 500 basket. they can invest in foreign issuers as the prospectuses of the fidelity select portfolios state. domestically they can invest in any stock that is not in the s&p 500 index, mainly smaller capitalization stocks. they also only “normally … invest at least 80% of assets in securities of companies principally engaged in the selected sector,” which gives them some wiggle room across sector borders. however, it is hard to imagine they deviate very far from their benchmark. only eight out 266 c.y. lin / financial services review 23 (2014) 249–271 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m a y -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 materials 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 energy 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m a y -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 financial 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m a y -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 industrials 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 technology 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 consumer staples 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 utilities 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a ug -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m ay -1 0 o ct -1 0 healthcare 0.0 0.2 0.4 0.6 0.8 1.0 ja n02 ju n02 n ov -0 2 a pr -0 3 se p03 fe b04 ju l04 d ec -0 4 m ay -0 5 o ct -0 5 m ar -0 6 a u g -0 6 ja n07 ju n07 n ov -0 7 a pr -0 8 se p08 fe b09 ju l09 d ec -0 9 m a y -1 0 o ct -1 0 consumer discretionary mf vs. etf mf vs. spy etf vs. spy fig. 4. thirty-six months rolling correlation between funds. 267c.y. lin / financial services review 23 (2014) 249–271 of 90 stocks of the top 10 holdings of the fidelity select portfolios at the end of september 2012 are not in the s&p 500 index.14 half of that eight are foreign stocks. seven out of the eight has a weight between 1.59% and 4.05% of corresponding sector mfs; the other one weights 13.66%.15 the top weighted index stocks also make the backbone of the sector mutual funds. the number of stocks in the top 10 holdings of both sector mfs and etfs varies from three to seven. seven out of top 10 holdings of the materials, industrials, and healthcare mf are also in the etf top 10 list, and the top 10 holdings make up between 46% and 64% of these three mutual funds. that number is six for the consumer staples and consumer discretionary mf, five for the energy and technology mf, four for the utilities mf, and three for the financial mf. the top 10 holdings make up between 29% and 68% of the funds in these six sectors. 5.2. fidelity fidelity had been the largest mutual fund family that mainly provides actively managed mutual funds for several decades. pozen and hamacher (2011) argue that fidelity, among several other fund families, has two characteristics that contribute to its success in the u.s. mutual fund business: dedication primary to asset management and control by investment professionals. the fidelity fund family has maintained top market shares in the past decade: 10.2%, 11.8%, and 11.3% in year 1990, 1995, and 2010, respectively. it was no. 1 in 1990 and 1995, but passed by vanguard (12.1%) in 2010. fidelity’s stock is effectively controlled by members of its funding family, which relieves the short-term performance pressure from public shareholders to increase quarterly earnings. they also stated that fidelity can develop compensation programs that promote top performance. fidelity’s megellan fund has long been used as an evidence to defy emh (see kochman and badarinathi 1993 and marcus 1990). fidelity’s large size can also potentially gain an insider edge as golec (2007) shows some evidence on informed trades made by fidelity funds. the large size of fidelity’s assets under management can provide economy of scale for securities research, which aids the key element of active management: security selecting. unlike passive managers, who often do little on stock selection, active managers can beat their benchmark by overweighting future winners, underweighting future losers, or some combination of both. they also have the freedom of holding cash. focusing on only one sector, sector fund managers can potentially gain growing knowledge and experience dealing with stocks in that sector, which could help their performance. elton, gruber, and green (2007) explain why funds may be more similar inside than outside fund families: portfolio managers within families are likely to have access to the same research analysis produced either by internal analysts or by a particular set of external research firms; portfolio managers may begin the security selection process with an economic forecast that is shared by other fund managers within the firm. it is not a surprise that fidelity sector funds as a group can outperform when the shared research and macro view work well. 268 c.y. lin / financial services review 23 (2014) 249–271 6. conclusion this article is one of the first that reports detailed analysis on individual equity sector fund performance. it also contributes to the literature by adding evidence of active equity management outperformance. i find considerable evidence that sector mutual funds, the nine fidelity select portfolios here, have provided better after-expense returns against broader market etf, spy, and their peer sector etfs, the nine select sector spdr funds, during the sample period 1999–2010. not only do they achieve higher nominal returns over the 12-year period except for few sector mfs, some of the funds also generate higher riskadjusted returns measured by sharpe ratio16 and ! from various asset pricing models. none of the sector mfs generates a significant negative ! for the full sample period no matter which asset pricing model is used. that is, the sector mfs do not underperform spy or peer etfs measured by !. the materials and the technology sector mf stand out in the analysis. the materials mf beats its peer etf and spy across the board for the full sample period: second highest 12-year return, highest sharpe ratio, and significant positive !s in all four regression models. the annualized abnormal return of the materials sector mf is 11.856%, 8.628%, and 8.472% against spy and 6.072% against xlb using factor models. the technology mf also beats both spy and peer etfs on 12-year return, sharpe ratio, and most regression-based measures. it generates annualized abnormal return of 7.368% and 7.320% against spy using the three-factor and four-factor models, and 8.268% against xlk. the energy mf outperforms on 12-year return, sharpe ratio, and one-factor model against spy with an annualized abnormal return of 12.768%. the industrials mf outperforms on 12-year return, sharpe ratio, and one-factor model against xli with an annualized abnormal return of 5.352%. the utilities and consumer discretionary mf have the weakest results. they do not beat their peer etfs measured by 12-year return and sharpe ratio, however, they do not underperform when asset pricing models are adopted. many researches have showed that on average active managers do not add value after fees and expenses (e.g., barras, scaillet, and wermers 2010). some argue that the value of active management lies in making market more efficient by improving asset allocation (jones and wermers 2011). xiong et al. (2010) document that both asset allocation and active management are critical to performance. the need for active managers to set the price was emphasized by a large index fund manager: “passive management is a free-ride strategy; it piggybacks on active management. you need to have active managers out there, and they need to be paid.”17 this article, however, presents evidence on outperformance of equity sector funds. if one considers active managers together are playing a zero-sum game, this study finds some of the winners. for individual investors who are interested in sector investing, this study shows that actively managed mutual funds can be a good candidate for sector allocation. most of these mutual funds do a better or equivalent job measured by both nominal return and total/ systematic risk adjusted return, after higher expenses and fees. the bottom line is that they do not underperform when measured with ! from asset pricing models, at least during the sample period. this study also provides evidence for financial planners when they help their 269c.y. lin / financial services review 23 (2014) 249–271 clients select mutual funds. it may be appropriate for financial planners to recommend sector mutual funds to their clients who have risk appetite for sector investing. notes 1 most etfs, although not all, are passively managed to track a specific index, such as the s&p 500. 2 https://www.fidelity.com/. 3 because telecommunication stocks are covered in the technology select sector spdr fund, i do not consider telecommunication a separate sector. 4 fidelity select portfolio annual report, february 29, 2012. 5 fidelity select portfolio prospectuses. 6 however, all nine select sector spdrs are diversified funds with respect to the internal revenue code. as a result, each sector index is modified so that an individual security does not comprise more than 25% of the index. source: http:// www.sectorspdr.com/. 7 select sector spdrs annual report, september 30, 2011. 8 spdr s&p 500 etf trust annual report, september 30, 2011. 9 the available price data for fidelity select industrials portfolio starts on july 8, 1999. 10 http://finance.yahoo.com/. 11 subsample period results are discussed but not reported throughout the article. results are available upon request. 12 for example, ross, westerfield, and jordan, corporate finance, 9th ed., 2010, mcgraw-hill. 13 http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html. 14 from fidelity and sector spider web sites. 15 british american tobacco plc adr represents 13.66% of the fidelity select consumer staples portfolio as of september 28, 2012. 16 eling (2008) shows that choosing a performance measure is not critical to fund evaluation and the sharpe ratio is generally adequate. 17 frank j. fabozzi, sergio m. focardi, and caroline jonas, 2010, investment management after global financial crisis, research foundation publications, cfa institute, page 26. references barras, l., scaillet, o., & wermers, r. (2010). false discoveries in mutual fund performance: measuring luck in estimated alphas. journal of finance, 65, 179–216. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57–82. dellva, w. l., demaskey, a. l. & smith, c. a. (2001). selectivity and market timing performance of fidelity sector mutual funds. financial review, 36, 39–54. eling, m. (2008). does the measure matter in the mutual fund industry? financial analysts journal, 64, 54–66. elton, e. j., gruber, m. j. green, c. (2007). the impact of mutual fund family membership on investor risk. journal of financial and quantitative analysis, 42, 257–277. 270 c.y. lin / financial services review 23 (2014) 249–271 fabozzi, f. j., focardi, s. m. & jonas, c. (2010). investment management after the global financial crisis. charlottesville, va: research foundation publications, cfa institute. fama, e. f., & french, k. r. (1993). common risk factors in the returns on stocks and bonds. journal of financial economics, 33, 3–53. golec, j. (2007). are the insider trades of a large institutional investor informed? financial review, 42, 161–190. groysberg, b., healy, p., nohria, n., & serafeim, g. (2011). what factors drive analyst forecasts? financial analysts journal, 67, 18–29. jones, r. c., & wermers, r. (2011). active management in mostly efficient markets. financial analysts journal, 67, 29–45. kacperczyk, m., sialm, c. & zheng, l. (2005). on the industry concentration of actively managed equity mutual funds. journal of finance, 60, 1983–2011. kaushik, a., pennathur, a. & barnhart, s. (2010). market timing and the determinants of performance of sector funds over the business cycle. managerial finance, 36, 583–602. kochman, l. m., & badarinathi, r. (1993). net selectivity revisited. journal of economics and finance, 17, 73–79. marcus, a. j. (1990). the magellan fund and market efficiency. journal of portfolio management, 17, 85–88. mcgahan, a. m. & porter, m. e. (1997). how much does industry matter, really? strategic management journal, 18, 15–30. porter, m. e. (1985). competitive advantage: creating and sustaining superior performance. new york, ny: free press. pozen, r., & hamacher, t. (2011). most likely to succeed: leadership in the fund industry. financial analysts journal, 67, 21–28. sullivan, r. n., & xiong, j. x. (2012). how index trading increases market vulnerability. financial analysts journal, 68, 70–84. xiong, j. x., ibbotson, r. g., idzorek, t. m., & chen, p. (2010). the equal importance of asset allocation and active management. financial analysts journal, 66, 22–30. zheng, w., & tower, e. (2005). fidelity versus vanguard: comparing the performance of the two largest mutual fund families. international review of economics and business, 52, 433–465. 271c.y. lin / financial services review 23 (2014) 249–271 from the editor this issue contains volume 30 issue 2 of the financial services review (fsr). issue 2 is integral because it is the first to be published since the unexpected passing of our longterm editor, dr. stuart michelson. for our readers, who may not be aware, dr. michelson passed away unexpectedly in march of this year. on behalf of the president of the academy of financial services and the entire executive team and board, we would like to express our sincere condolences to his family, friends, and colleagues. to honor dr. michelson, we will be naming an award after him at the annual conference of the academy of financial services. during this transition, i will be taking over the fsr editorship’s duties, having previously worked with dr. michelson as a guest editor. i would like to personally thank the board and academy of financial services members for their support. i would also like to thank you, our authors, and readers for your patience. as interim editor, i hope to build upon the great work that dr. michelson started, which includes broadening the scope of articles while still focusing on individual financial management and personal financial planning. beyond that, we are exploring other avenues to increase the efficiency and timeliness of the journal’s publication processes. the lead article “why are women less motivated to become financially literate?” is co-authored by jaclyn j. beierleina at east carolina university, kaleigh launsby at bank of american, and haley smith at duke university hospital. this article examines the motivations of women to become more financially literate. the authors find that women in their sample score lower on basic finance questions and report lower motivation to learn personal finance. they also find that women who expect to make decisions with a spouse report lower motivation than women who expect to make decisions alone. their findings show that women are less motivated to become financially literate due to confidence and shared financial responsibilities. the second article explores credit and financial sophistication. “financial literacy and its impact on the credit card debt puzzle” is authored by laura ricaldi at utah valley, terrance k. martin also at utah valley university, and sandra huston at texas tech university. using the 2016 survey of consumer finances and a series of multinomial regressions to investigate the credit card debt puzzle, the authors’ results suggest that financially literate households are less likely to display irrational behavior. the authors break their sample into three subgroups in the analysis: convenience users, solvent revolvers, and insolvent revolvers. the third article, “your mileage may vary” is co-authored by manoj athavale, stephen avila, and joseph goebel at ball state university. the paper centers around the topic of 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) v–vi retirement planning and retirement income management. the authors specifically define portfolio success to meet the parameters of their articles. their results show that the likelihood of success is inversely related to withdrawal rate, retirement horizon, and portfolio risk increase and directly related to portfolio return, allocation aggressiveness, and early experience. in addition, they find that portfolio success is highly sensitive to withdrawal rates, highlighting that aggressive allocations may provide more dependable portfolio outcomes for retirees. in the final article in this issue jason heller at coastal wealth, benjamin cummings at utah valley university, and jason martin at the american college of financial services coauthor the final article. in their paper “distribution channel effects on advisor managed investment performance,” the authors perform a series of analyses to determine whether advisors at registered investment advisory (ria) firms can produce higher net investment results compared to advisors employed at dually registered independent broker/dealer (ibd) firms. the authors use a proprietary dataset from a large united states investment advisory platform. using this dataset, they found that advisors at rias outperformed those at ibds in higher-risk portfolios through turnkey asset management programs (tamps) and unified managed accounts. i want to thank everyone who makes the fsr production possible, including our authors, referees, and readers. dr. michelson often praised the outstanding work of our reviewers, and i wish to echo his sentiment. if you have a manuscript and looking for a publication outlet, please consider submitting it to the financial services review. the journal welcomes articles on the areas of personal financial planning. like dr. michelson, i am committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. yours sincerely, terrance k. martin jr. interim editor financial services review vi t.k. martin / financial services review 30 (2022) v–vi financial literacy: profiling a successful high school outreach program greg filbecka,*, jason pettnerb, xin zhaoc ablack school of business, penn state behrend, erie, pa 16563, usa bmarket analyst, c.s. mckee, 420 fort duquesne blvd suite 800, pittsburgh, pa 15222, usa cblack school of business, penn state behrend, erie, pa 16563, usa abstract the cfa society pittsburgh launched a high school financial literacy campaign resulting in significant improvements in financial behavior, subjective and objective financial knowledge, and selfesteem. before the campaign, male students and students with higher grade point averages (gpas) show better objective knowledge. in addition, we find disconnect between actual and perceived financial knowledge. students exhibited gains in all aspects after completing the program. the subcategories with the lowest pre-survey scores or female students show the greatest improvements in the post-survey. students with lower gpas experienced greater improvement in financial behavior and objective knowledge, while higher gpa students improved more in subjective knowledge. © 2020 academy of financial services. all rights reserved. jel classifications: g53 keywords: financial literacy; survey; outreach; education 1. introduction based on the results of multiple academic studies, a significant lack of financial literacy exists across nearly all demographics. while financial literacy statistics are important, the implications of a lack of financial literacy and numeracy are far reaching due to their impact on financial decisions. as a result, the potential implications of financial literacy and *corresponding author. tel.: +1-814-898-6549. e-mail address: mgf11@psu.edu 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 315–340 numeracy will be explored in depth. in this paper, we demonstrate the effectiveness of introducing a program to improve financial literacy and numeracy for high school students. according to remund (2010, p. 278), “financial literacy is a measure of the degree to which one understands key financial concepts and possesses the ability and confidence to manage personal finances through appropriate, short-term decision-making and sound, longrange financial planning, while mindful of life events and changing economic conditions.” to gauge financial literacy, studies use many different knowledge-based questions; however, the overall concepts remain relatively consistent across financial literacy surveys. three common fundamental financial knowledge concepts include interest rates, inflation, and risk diversification. numeracy is an important component of financial literacy that literature often relates to financial behavior. estrada-mejia, de vries, and zeelenberg (2016, p. 53) define numeracy as “the ability to understand and use numerical information.” numeracy is closely related to multiple aspects of financial decision making. this study extends prior research on the effectiveness of financial literacy education by providing direct evidence from a financial literacy campaign launched by the cfa society pittsburgh. the study is based upon data collected from a financial literacy campaign of 53 high schools, across seven states, during the 2017-2018 academic year. the financial literacy education campaign materials were created using the book the missing semester (kabala & natali, 2013) as the main resource and curriculum. before starting the course, students were given a pre-survey to test their baseline in four major areas: subjective financial knowledge, financial behavior, objective financial knowledge, and financial self-esteem. following the completion of the course, students were administered a post-survey to test changes in the four major areas. to anonymously track the progress of students, they were assigned a unique student id code. the results display significant improvement in all areas of interest. each area of interest improves at a statistically significant level of at least 5% after the financial literacy campaign, indicating an effective effort at improving financial literacy. 2. literature review since the effects of financial literacy are evident and the high percentage of financial illiteracy has been proven, steps must be taken to improve financial literacy. green and riddell (2012) analyze data gathered from the canadian component of the international adult literacy and life skills survey (ialss), which sought to measure the skills of adults in canada. the survey asked skill-based questions that focused upon four cognitive skills: prose literacy, document literacy, numeracy, and problem solving. regression analysis reveals an increase of 3.4% in literacy and numeracy scores for an additional year of schooling. they also find that completing an extra 4 years of education results in a 24% increase in literacy, which increases a person’s literacy from the median to the 80th percentile. although understanding financial concepts is vital to financial behavior, possessing high financial self-esteem and confidence is an essential key to successful financial decisions. using data collected on 12,686 individuals by the u.s. bureau of labor statistics tracked over a 30-year period, tang and baker (2016) create four main variables: financial behavior, 316 g. filbeck et al. / financial services review 28 (2020) 315–340 self-esteem, objective, and subjective financial knowledge and covariates. their results indicate a direct and indirect relationship of self-esteem on multiple financial behaviors. as a result, the effect of self-esteem is statistically significant; thus, self-esteem is a factor of financial behavior. their results indicate that to improve financial behavior, subjective knowledge must be at least a portion of a financial education curriculum because objective knowledge by itself does not have a complete positive impact on financial behavior. an important facet of self-esteem is the extent to which a person’s confidence becomes greater than their actual knowledge. overconfident individuals have the tendency to overestimate their own knowledge, leading to a higher risk of engaging in costly and risky financial behaviors (asaad, 2015). asaad discovers confidence is an important fragment of financial literacy, but also finds that perceived knowledge without actual knowledge increases the risk of suboptimal financial decisions. mccannon, asaad, and wilson (2016) conduct a study based on subjects playing an experimental trust game (berg, dickhaut, & mccabe, 1995) and completing a risk assessment, background questionnaire, and financial literacy quiz. results of the risk assessment and financial literacy quiz were decomposed to generate overconfidence as a variable. they find a statistically significant relationship between overconfidence and trusting investments. although the actual financial literacy in the united states has declined over the past 5 years, lin, bumcrot, ulicny, lusardi, mottola, kieffer, and walsh (2016) find the percentage of u.s. citizens who have a high-self assessment has actually increased. thus, efforts to improve financial behavior must address the going-concern of overconfidence. recent studies have identified the ability to increase financial literacy through education at the university level. gerrans and heaney (2019) study the effects of financial literacy following an undergraduate personal finance course at an australian university. after completing the course, students showed improvements in both objective and subjective financial literacy. 2.1. gender gap in financial literacy several studies have found the existence of a financial literacy gender gap. in particular, cupák, fessler, schneebaum, and silgoner (2018) find that women score lower than men on financial literacy, with a more pronounced gap in developed countries. additionally, preston and wright (2019) examine the financial literacy gap in australia. while the “human capital variables” (age and education) were not significant in explaining the gap, “labor market variables” (including sector and occupation) were significant in explaining the gap. while these findings identify an initial gender gap in financial literacy, gerrans and heaney (2019) find female students benefit the greatest from financial literacy education. 3. need for economic education according to the 2015 national report card, financial literacy education in high school is insufficient. on a scale from “a” to “f,” twenty-six states received grades of c, d, or f (pelletier, 2015). an “a” grade (five states) requires the schools to offer a one-semester personal finance course as a graduation requirement. a “b” grade (20 states) requires the schools to include personal finance education within a required course, whether as a standg. filbeck et al. / financial services review 28 (2020) 315–340 317 alone course or part of another course. a “c” grade (11 states) requires schools to offer personal finance topics in schools, but does not mandate students take the course. a “d” grade (three states) means that the state has a “modest levels” of personal finance in its academic standards. finally, an “f” grade (12 states) goes to states with nearly no financial educational requirements, meaning that a student can graduate high school without an introduction to any financial literacy concepts. because formal education has a statistically significant relationship with literacy and numeracy (green & riddell, 2012), the lack of high school financial education is a potential factor of low financial literacy scores. although the effects of financial literacy education are highly debated, research conducted by filbeck and zhao (2018) illustrates that financial literacy initiatives have a positive impact on both financial knowledge and behavior. they find that teaching financial concepts to high school students had a profound and statistically significant effect on both subjective financial knowledge and financial behavior for the students involved in the survey. based upon previous research conducted, perceived knowledge may differ from actual knowledge (asaad, 2015; lin et al., 2016; mccannon et al., 2016). in addition, tang and baker (2016) prove that self-esteem and subjective knowledge have a statistically significant effect on financial behavior. as a result, further research could incorporate subjective education as a piece of the financial initiative, as well as financial knowledge questions that analyze both perceived and actual financial knowledge. this article extends the work of filbeck and zhao by including objective assessments about acquisition of financial knowledge based on educational outreach. our hypothesis is as follows: hypothesis 1: the financial literacy outreach program, which targets subjective financial knowledge, financial behavior, objective financial knowledge, and financial selfesteem, will result in statistically significant improvement for participating students. in our article, we first analyze baseline pre-survey responses in financial behavior, subjective and objective financial knowledge, and self-esteem of high school students participating in a financial literacy outreach program developed by the cfa society pittsburgh. we use path analysis to assess the strength of direct and indirect links between these four aspects of financial literacy and other control variables. next, by comparing matched preand post-survey results, we analyze the impact of the financial literacy outreach program in financial behavior, subjective and objective financial knowledge, and self-esteem. we first use a paired t test to compare the matched preand post-survey results. then, we use regression analysis to analyze how the improvements in these four aspects of financial literacy in the post-survey are impacted by control variables. 4. data sample the cfa society pittsburgh has been active in financial literacy outreach since 2010. over 50 individuals serve on the financial literacy committee, which directs curricular development and training efforts. each year, representatives from the society participate in act 318 g. filbeck et al. / financial services review 28 (2020) 315–340 48 training sessions across the state of pennsylvania as well as offering to provide an hourlong presentation on core financial literacy concepts on request. the number of high schools participating grew rapidly starting in 2015. in 2018, the state treasurer of pennsylvania endorsed the program and encouraged all pennsylvania high schools to participate. that same year, representatives from the cfa society pittsburgh led a session for the national association of state treasurers in providence, rhode island, in an effort to further expand outreach across more states. the cfa society pittsburgh provides participating schools with instructional materials for a semester-long equivalent course based on the missing semester (kabala & natali, 2012). each school determines the best way to deliver their programs (e.g., weekly for a semester, or daily lessons over fewer weeks). schools are supplied with powerpoint resources to accompany the book, along with a web-based portal (available through cfa society pittsburgh website) of best practices and exercises, submitted from previous participating schools. in addition, members of the financial literacy committee extend an invitation to present to the students in the classroom for a day. the financial literacy member uses a powerpoint presentation developed by the financial literacy committee, which gives a broad overview of the main topics discussed within the curriculum, while allowing students to ask questions relating to financial literacy subject material or real-world applications. committee members and other teachers act as a resource for the participating schools throughout the semester, to provide any assistance or feedback about the curriculum. for the 2018-2019 school year, a total of 78 high schools, spread across seven states (california, michigan, new jersey, pennsylvania, west virginia, wisconsin, and wyoming) with 173 classes/teachers, were invited to participate in both a preand post-survey to examine the effectiveness of financial literacy education. through our partnership with the cfa society pittsburgh, act 48 training in pennsylvania, multiple intermediate units, the pennsylvania state treasurer, and other connections with individual teachers, we established our subset of participating teachers. we offered each teacher financial literacy materials and access to the cfa society pittsburgh portal for free, in exchange for their participation in our survey. after gathering the list of enrolled teachers, we assigned each a unique class code. links for a preand post-surveys were provided to the participating teachers. within the introductory email, instructors were given directions to assign each student with a unique id number, allowing preand post-surveys to be matched for analysis. participating schools agreed to administer the presurvey before any instructional delivery. post-surveys were to be completed within a week after completion of the last instructional unit on financial literacy. the distributed survey was designed as an extension of the work conducted by filbeck and zhao (2018) with the addition of six objective financial knowledge questions. the full pre-survey and post-survey can be found in appendixes a and b, respectively. of the original population, 1,613 students from 31 participating schools and two states (pennsylvania and new jersey) completed pre-surveys. a total of 1,050 post-surveys were completed by students from 23 schools in pennsylvania (91.2%) and new jersey (8.8%). table 1 reports the descriptive statistics for the full sample and the test sample. the full sample consists of 1,613 students completing the pre-survey, while the matched test sample includes 829 students who submitted both a preand post-survey. the matched sample totals show attrition in the survey process of approximately one-half, despite attempts to minimize g. filbeck et al. / financial services review 28 (2020) 315–340 319 the loss through a series of six reminder emails to participating teachers during the period of the program. based on feedback from teachers, failure of students to complete post-surveys were primarily associated with a failure of completion of the financial literary program or a failure of teachers to oversee students in the completion of the post-survey. we also had 221 instances of students filling out post-surveys who previously did not complete a pre-survey. of the full sample, 1,425 (88%) students are in their junior or senior year; in the test sample, 718 (87%) students are in their junior or senior year. female students account for approximately 47% in both the full sample and the test sample. regarding their favorite subjects, students within the full sample favored science (28%) and math (27%), a trend which table 1 sample description grade 9th 10th 11th 12th total panel a. whole sample male english 4 4 23 64 95 math 18 13 59 171 261 science 8 20 51 147 226 social studies 24 22 59 171 276 female english 12 10 47 155 224 math 9 14 43 116 182 science 6 11 34 170 221 social studies 9 4 22 93 128 total 90 98 338 1,087 1,613 panel b. test sample male english 3 1 12 31 47 math 11 7 33 75 126 science 4 14 31 69 118 social studies 12 14 34 84 144 female english 6 3 27 81 117 math 6 8 27 61 102 science 1 11 16 82 110 social studies 6 4 15 40 65 total 49 62 195 523 829 panel c. school district characteristics mean standard deviation percentile min 25 50 75 max population 31,522.95 35,083 6,208 11,382 18,412 37,567 148,678 poverty 9.6% 5.9% 3.3% 5.9% 8.2% 11.3% 25.8% pct_college 29.5% 14.8% 11.2% 16.3% 29.8% 41.2% 57.8% note: number of students across different grade level and different favorite subjects for the whole sample (panel a) and the test sample (panel b). panel c shows the school district characteristics: population is the population in the school district. poverty is the poverty rate in the school district. pct_college is the percentage of residents who have attained bachelor degree or higher. 320 g. filbeck et al. / financial services review 28 (2020) 315–340 continued to the test sample with math and science as the favorite subjects both at 27.5%. favorite subject is included in our survey to determine whether academic interest area plays a significant role in financial literacy educational outreach. appendixes a and b show the preand post-survey questions addressed by program participants in order to assess financial knowledge (both objective and subjective questions), financial behavior, and self-esteem. lusardi and mitchell (2014) point out that a substantial mismatch exists between individual’s self-assessed (subjective) financial knowledge and their actual knowledge. the survey design is consistent with hastings, madrian, and skimmyhorn (2013), who argue that financial literacy should focus on competences that individuals need. the organization of the survey is constructed in a way that assesses the four major keys to financial success: financial self-esteem, perceived (subjective) financial knowledge, financial behavior, and objective financial knowledge with numeracy. the financial behavior and subjective financial knowledge test financial self-esteem and perceived financial knowledge by gauging the student’s self-reported understanding. questions are broadly categorized into two types: financial knowledge (subjective and objective) and financial behavior. the only difference between preand post-survey questions are in the objective financial knowledge questions—the same concepts are tested with different questions. objective financial knowledge questions are asked in a manner that contains a right or wrong answer. each objective financial knowledge question contains at least one wrong answer and a choice of “i don’t know.” the questions are based upon five major categories of financial literacy: risk diversification, compound interest, credit, numeracy (interest), and inflation. the questions are analyzed using two methods: correctness and willingness to answer. the first method of correctness assigns a 1 for each correct answer and 0 for any other answer. the second method to measure willingness assigns a 1 for an answer of i don’t know and 0 for any other answer. the second method is used to measure financial self-esteem as measured by the amount of questions answered with i don’t know. improved self-esteem occurs as students become less likely to answer i don’t know and instead select an answer that could be correct or incorrect, showing greater confidence after completing financial literacy education. the survey consists of 21 overall questions: four financial behavior, six objective financial knowledge, and 11 subjective financial knowledge. financial behavior and subjective financial knowledge questions are rated on a 5-point scale ranging from 1 = strongly disagree to 5 = strongly agree. the four financial behavior questions are “i like to save money more than i like to spend it,” “i have a checking and/or a savings account,” “i have conversations with my parents regarding personal finance,” and “i think it is important to contribute to a retirement plan (ex: roth ira, 401k, etc.)”. subjective financial knowledge questions involve perceived understanding of financial concepts, and include questions such as “i understand how to establish a financial plan,” “i think financial literacy is important for my future,” and “i understand the process by which my parents/guardians make financial decisions.” the final six objective knowledge questions test financial self-esteem and objective financial knowledge by assessing correctness of answers and willingness to select an answer other than i don’t know. g. filbeck et al. / financial services review 28 (2020) 315–340 321 5. test results first, t test and path analysis are used to analyze the pre-survey results of the full sample. following the pre-survey analysis, we run t test and regressions to compare preand postsurvey results within the test sample. 5.1. pre-survey results for the t test of the pre-survey responses, two student characteristics are present: gender and gpa. each characteristic divides the full sample into two groups. gender is broken into a subgroup for females and males. the median gpa of the whole sample divides the students into a higher gpa or lower gpa group. the response differences between groups are shown in table 2. the average responses are compared with the individual subgroups of gender and gpa to determine if the characteristics exhibit a statistically significant effect. the gender characteristic identifies the effect of a student being a female versus a male (gender), as well as the effect of a student having a high gpa versus low gpa. the data shows no statistically significant effect of gpa on subjective financial knowledge. however, consistent with cupák et al. (2018), female students tend to have lower subjective financial knowledge compared with male students (statistically significant at the 1% level). for financial behavior, students with a high gpa tend to be better financially behaved (statistically significant at the 1% level). in objective financial knowledge, females tend to score lower in correctness, while students with high gpas tend to score higher in correctness (both significant at the 1% level). similarly, for self-esteem, females tend to answer i don’t know more often, while students with higher gpas are less likely to select i don’t know (both statistically significant at the 1% level). 5.2. pearson correlation and path analysis table 3 reports pearson correlations between financial behavior and self-esteem (idk), subjective, and objective financial knowledge. financial behavior, objective and subjective knowledge are significantly correlated with each other. idk is significantly negatively correlated with the other three measures of financial literacy, which indicates a positive correlation of self-esteem level and the other three measures of financial literacy. naturally, due to the setup of the self-esteem measure, we find a �0.77 correlation between idk answers and objective knowledge. next, following tang and baker (2016), we use path models to analyze the relationship between characteristics and four financial literacy measures. we choose path analysis, as it forces us to specify relationships among all of the independent variables. this results in a model showing causal mechanisms, through which independent variables produce both direct and indirect effects on a dependent variable. all causal relationships between variables must go in one direction only, for path models. to identify the impact, we assign the dependent variable as behavior (subjective, idk), which is the total score for financial behavior questions (subjective questions, objective questions with i don’t know answers). 322 g. filbeck et al. / financial services review 28 (2020) 315–340 t ab le 2 d if fe re n ce s b as ed o n st u d en t ch ar ac te ri st ic s: p re -s u rv ey re su lt s a v er ag e re sp o n se f em al e h ig h er g p a p an el a . s u b je ct iv e fi n an ci al k n o w le d g e q u es ti o n s 2 . i u n d er st an d h o w to es ta b li sh a fi n an ci al p la n . 3 .0 3 9 �0 .2 0 4 * * * �0 .0 2 3 3 . i th in k fi n an ci al li te ra cy is im p o rt an t fo r m y fu tu re . 4 .3 6 5 0 .0 2 1 0 .1 8 6 * * * 6 . i u n d er st an d th e p ro ce ss b y w h ic h m y p ar en ts /g u ar d ia n s m ak e fi n an ci al d ec is io n s. 3 .3 5 8 �0 .1 4 3 * * * �0 .0 9 3 * 7 . i k n o w h o w to d et er m in e th e ap p ro p ri at e to ta l co st s as so ci at ed w it h th e co ll eg es /u n iv er si ti es i am in te re st ed in at te n d in g . 3 .0 8 6 �0 .0 6 6 0 .1 4 4 * * 8 . i u n d er st an d th e p ro ce ss b y w h ic h lo an re p ay m en ts ta k e p la ce in cl u d in g th e im p ac t o f in te re st , d el in q u en cy an d d ef au lt . 2 .7 4 3 �0 .2 5 9 * * * �0 .1 6 0 * * 9 . i u n d er st an d th e p ro ce ss b y w h ic h cr ed it ca rd ch ar g es an d re p ay m en t sc h ed u le s ca n im p ac t th e le v el o f fi n an ci al d eb t le v el s. 3 .4 2 8 0 .0 0 1 �0 .0 5 3 1 0 . w h en it co m es to p u rc h as in g a ca r, i k n o w h o w to d et er m in e h o w m u ch o f a ca r i ca n af fo rd . 3 .2 8 7 �0 .2 8 4 * * * �0 .2 5 9 * * * 1 1 . i u n d er st an d h o w to ev al u at e th e co st -b en efi t an al y si s o f tr ai n in g fo r th e jo b i w o u ld li k e to p er fo rm af te r co m p le ti n g sc h o o l. 3 .0 1 5 �0 .2 1 9 * * * �0 .1 2 3 * * 1 2 . i k n o w w h at a r o th ir a is an d h o w it w o rk s fr o m a ta x at io n st an d p o in t. 1 .9 5 1 �0 .1 8 1 * * * �0 .0 9 5 * 1 3 . i k n o w h o w to cr ea te a sa v in g s p la n b as ed o n th e ab il it y to es ti m at e m o n th ly li v in g ex p en se s. 3 .1 4 8 �0 .1 6 7 * * * 0 .0 0 5 1 4 . i k n o w h o w to p la n fi n an ci al ly fo r re ti re m en t. 2 .6 2 1 �0 .2 8 3 * * * �0 .0 5 0 t o ta l sc o re fo r su b je ct iv e fi n an ci al k n o w le d g e 3 3 .8 9 3 �1 .6 2 2 * * * �0 .3 8 1 p an el b . f in an ci al b eh av io r q u es ti o n s 1 . i li k e to sa v e m o n ey m o re th an i li k e to sp en d it . 3 .5 0 5 �0 .0 1 3 * * 0 .2 1 9 * * * 4 . i h av e a ch ec k in g an d /o r a sa v in g s ac co u n t. 4 .3 8 1 0 .0 3 4 0 .1 8 6 * * * 5 . i h av e co n v er sa ti o n s w it h m y p ar en ts re g ar d in g p er so n al fi n an ce . 3 .4 1 8 0 .0 1 2 0 .2 8 3 * * * 1 5 . i th in k it is im p o rt an t to co n tr ib u te to a re ti re m en t p la n (e x : r o th ir a , 4 0 1 k , et c. ). 4 .0 8 3 �0 .0 1 0 0 .3 1 8 * * * t o ta l sc o re fo r fi n an ci al b eh av io r 1 4 .7 4 6 0 .1 2 2 1 .2 3 1 * * * p an el c . o b je ct iv e q u es ti o n s (c o rr ec t a n sw er s) a 1 . is it sa fe r to p u t y o u r m o n ey in to o n e in v es tm en t o r p u t y o u r m o n ey in to m u lt ip le in v es tm en ts ? 0 .6 1 9 �0 .1 2 4 * * * 0 .0 9 4 * * * 2 . if y o u in v es t $ 1 0 0 in a r o th ir a an d ea rn 1 0 % p er y ea r fo r 3 y ea rs , h o w m u ch w o u ld it b e w o rt h at th e en d o f th re e y ea rs . 0 .2 8 9 �0 .1 0 8 * * * 0 .0 7 3 * * * 3 . if y o u u se a cr ed it ca rd in ja n u ar y fo r a to ta l o f $ 3 0 0 , w h ic h p ay m en t o p ti o n w il l re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d . 0 .4 9 3 �0 .0 7 5 * * * 0 .1 3 5 * * * 4 . s u p p o se y o u d ec id e to b u y a b m w fo r $ 5 0 ,0 0 0 . if y o u ta k e o u t an au to lo an fo r 5 y ea rs w it h 5 % in te re st , h o w m u ch to ta l w il l y o u p ay p er y ea r? 0 .4 2 9 �0 .1 4 2 * * * 0 .1 6 7 * * * 5 . in th e fu tu re , th e co st o f th in g s y o u b u y d o u b le s a n d y o u r in co m e al so d o u b le s. h o w m u ch w il l y o u b e ab le to b u y in th e fu tu re in co m p ar is o n to to d ay ? 0 .5 5 7 �0 .0 6 6 * * * 0 .1 7 3 * * * 6 . s u p p o se y o u h av e $ 3 0 ,0 0 0 in st u d en t lo an s. w h ic h p ay m en t o p ti o n w o u ld re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d ? 0 .5 3 0 �0 .0 5 4 * * 0 .2 3 5 * * * t o ta l sc o re fo r o b je ct iv e q u es ti o n s (c o rr ec t a n sw er s) 2 .9 1 6 �0 .5 7 0 * * * 0 .8 7 7 * * * (c o n ti n u ed o n n ex t p a g e) g. filbeck et al. / financial services review 28 (2020) 315–340 323 t ab le 2 (c o n ti n u ed ) a v er ag e re sp o n se f em al e h ig h er g p a p an el d . o b je ct iv e q u es ti o n s (" i d o n ’t k n o w " a n sw er s) b 1 . is it sa fe r to p u t y o u r m o n ey in to o n e in v es tm en t o r p u t y o u r m o n ey in to m u lt ip le in v es tm en ts ? 0 .2 6 3 0 .1 2 1 * * * �0 .0 1 2 2 . if y o u in v es t $ 1 0 0 in a r o th ir a an d ea rn 1 0 % p er y ea r fo r 3 y ea rs , h o w m u ch w o u ld it b e w o rt h at th e en d o f th re e y ea rs . 0 .3 4 5 0 .1 7 6 * * * �0 .0 8 6 * * * 3 . if y o u u se a cr ed it ca rd in ja n u ar y fo r a to ta l o f $ 3 0 0 , w h ic h p ay m en t o p ti o n w il l re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d . 0 .3 4 0 0 .0 4 0 * �0 .0 6 5 * * * 4 . s u p p o se y o u d ec id e to b u y a b m w fo r $ 5 0 ,0 0 0 . if y o u ta k e o u t an au to lo an fo r 5 y ea rs w it h 5 % in te re st , h o w m u ch to ta l w il l y o u p ay p er y ea r? 0 .2 9 5 0 .1 3 7 * * * �0 .0 6 4 * * * 5 . in th e fu tu re , th e co st o f th in g s y o u b u y d o u b le s a n d y o u r in co m e al so d o u b le s. h o w m u ch w il l y o u b e ab le to b u y in th e fu tu re in co m p ar is o n to to d ay ? 0 .1 8 6 0 .0 5 2 * * * �0 .0 7 1 * * * 6 . s u p p o se y o u h av e $ 3 0 ,0 0 0 in st u d en t lo an s. w h ic h p ay m en t o p ti o n w o u ld re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d ? 0 .2 8 7 0 .0 7 4 * * * �0 .1 2 3 * * * t o ta l sc o re fo r o b je ct iv e q u es ti o n s (" i d o n ’t k n o w " a n sw er s) 1 .7 1 7 0 .5 9 9 * * * �0 .4 2 1 * * * n o te : s h o w s th e d if fe re n ce s o f p re -s u rv ey st u d en t re sp o n se s o n fi n an ci al k n o w le d g e an d fi n an ci al b eh av io ra l q u es ti o n s ac ro ss d if fe re n t g en d er an d g p a fo r th e w h o le sa m p le . g en d er is b ro k en in to a s u b -g ro u p fo r fe m al es an d m al es . t h e m ed ia n g p a o f th e w h o le sa m p le d iv id es th e st u d en ts as a h ig h er g p a o r lo w er g p a g ro u p . t h e g en d er ch ar ac te ri st ic id en ti fi es th e ef fe ct o f a st u d en t b ei n g a fe m al e v er su s a m al e, as w el l as th e ef fe ct o f a st u d en t h av in g a h ig h g p a v er su s lo w g p a . * * * , * * , * in d ic at e st at is ti ca l si g n ifi ca n ce at 0 .0 1 , 0 .0 5 , an d 0 .1 0 le v el , re sp ec ti v el y . a im p ro v em en t is in d ic at ed b y a p o si ti v e d if fe re n ce an d tst at (m o re p eo p le se le ct in g th e co rr ec t an sw er ). b im p ro v em en t is in d ic at ed b y a n eg at iv e d if fe re n ce an d tst at (l es s p eo p le se le ct in g “i d o n ’t k n o w ”) . 324 g. filbeck et al. / financial services review 28 (2020) 315–340 three path models are used to test the effect of financial behavior scores (behavior), subjective financial knowledge scores (subjective), and i don’t know answers in the objective questions (idk). in all three path models, we include gender (female), grade level (sophomore, junior, and senior), favorite subject (english, math, and science), gpa and favorite learning method (learning by doing [lbd], listening, discussing, and visual) as independent variables. control variables lbd, listening, discussing, and visual allow for direct testing of instructional preferences of students which may impact success of the financial literacy program (amagir, groot, van den brink, & wilschut 2018). detailed definitions of variables are listed in appendix c. also, to explore whether the school district of participating classes has any effect on the pre-survey results, we include three school district characteristics as control variables. specifically, for each participating school, we collect its school district census data from https://censusreporter.org/. we exclude private schools, virtual, finance knowledge learning centers, and chartered schools from the whole sample as these schools are hard to determine their school districts. for each public high school, we collect its school district data on its population, poverty rate and percentage of residents who attained bachelor or higher degrees. also, because the average number of students who participated in the pre-surveys from each school is 20 with a median value of one (ranges from 1 to 225 students per school), we removed school districts with less than 20 student participants. this process reduces our whole sample from 1,613 to 1,441 pre-surveys with available school district data. summary statistics of school district data are reported in panel c of table 1. table 4 illustrates the results from these models. path (1) uses behavior as the dependent variable. the results show students of higher grade level (sophomore, junior, and senior), with high subjective financial knowledge, higher level of self-esteem, as well as a high gpa, have a higher probability to be better behaved financially. both the subjective financial knowledge coefficient and self-esteem measure (idk) are statistically significant at the 1% level, implying that a student who possesses more subjective knowledge and higher selfesteem are more likely to exhibit better financial behavior. on school district characteristics, table 3 shows the correlation coefficients between financial behavior and self-esteem (idk answers), objective and subjective financial knowledge subjective behavior objective idk answers subjective corr 1.000 p-value behavior corr 0.334*** 1.000 p-value <0.0001 objective corr 0.277*** 0.245*** 1.000 p-value <0.0001 <0.0001 idk answers corr �0.361*** �0.249*** �0.771*** 1.000 p-value <0.0001 <0.0001 <0.0001 ***, **, * indicate statistical significance at 0.01, 0.05 and 0.10 level, respectively. g. filbeck et al. / financial services review 28 (2020) 315–340 325 results show school districts with higher poverty rate are more likely to have worse financial behavior (statistically significant at the 1% level). the second model (path 2) uses subjective (representing the measure for subjective knowledge) as the dependent variable. the results show female students with low gpas are more likely to have lower subjective financial knowledge. additionally, students who prefer the learning styles of learning by doing (lbd), listening, or discussing are more likely to exhibit greater subjective financial knowledge (all statistically significant at the 5% level). the negative coefficient of idk (statistically significant at the 1% level) shows that a student who has higher level of self-esteem is also more likely to be subjectively knowledgeable in finance. the insignificant coefficient of objective (representing the measure for objective knowledge) shows the positive linkage between subjective and objective financial knowledge, which is consistent to the results of tang and baker (2016) who find disconnect between subjective and objective knowledge when comparing actual and perceived financial table 4 regressions on student characteristics: pre-survey results path (1) path (2) path (3) dep. var.: behavior dep. var.: subjective dep. var.: idk coefficient t-value coefficient t-value coefficient t-value idk �0.093 �3.53*** �0.330 �13.43*** objective 0.010 0.38 0.032 1.18 subjective 0.296 11.91*** female 0.042 1.59 �0.035 �1.28 0.172 6.38*** sophomore 0.062 1.82* �0.037 �1.04 �0.024 �0.65 junior 0.205 4.22*** �0.066 �1.31 �0.043 �0.82 senior 0.294 5.66*** �0.074 �1.36 �0.133 �2.37** gpa 0.150 5.72*** �0.046 �1.68* �0.142 �5.18*** english �0.024 �0.80 0.039 1.27 0.036 1.12 math �0.015 �0.51 0.015 0.49 �0.025 �0.78 science �0.031 �1.04 0.020 0.64 �0.027 �0.83 lbd 0.043 1.70* 0.066 2.52** �0.015 �0.55 listening �0.005 �0.20 0.066 2.49** �0.060 �2.18** discussing �0.005 �0.20 0.059 2.29** �0.008 �0.28 visual 0.047 1.86* �0.029 �1.09 0.048 1.77* log(population) 0.019 0.72 0.079 2.91*** �0.038 �1.34 poverty �0.130 �3.72*** 0.034 0.92 0.065 1.74* pct_college 0.041 1.22 0.008 0.24 �0.029 �0.80 note: table 4 shows the regression results on student characteristics of the whole sample. behavior (objective, subjective, idk) is the total score for financial behavior questions (subjective questions, objective questions with correct answers, objective questions with “i don’t know” answers). female is a dummy variable that is equal to 1 if the student is a female student and 0 otherwise. sophomore (junior, senior) is a dummy variable that is equal to 1 if the student is a sophomore (junior, senior), and 0 otherwise. english (math, science) is a dummy variable which is equal to 1 if the student’s favorite subject is english (math, science), and 0 otherwise. lbd (listening, discussing, visual) is a dummy variable that is equal to 1 if the student chooses learning by doing (listening, discussing with peers, features visual support) as favorite instruction method, and 0 otherwise. gpa is a student’s grade point average. log(population) is the log of the population in the school district. poverty is the poverty rate in the school district. pct_college is the percentage of residents who have attained bachelor degree or higher. ***, **, * indicate statistical significance at 0.01, 0.05, and 0.10 level, respectively. 326 g. filbeck et al. / financial services review 28 (2020) 315–340 knowledge. students in larger school district are more likely to be more subjectively knowledgeable in finance (statistically significant at the 1% level). one possible explanation for this finding is that students who live in larger school districts may be exposed to more financial knowledge/concepts. the third model (path 3) uses idk (representing the measure of self-esteem) as the dependent variable. female students exhibit lower self-esteem scores, while junior or senior students who prefer learning by listening are more likely to have higher self-esteem. on school district characteristics, results show school districts with high poverty rate are more likely to have lower level of self-esteem (statistically significant at the 10% level). overall, our results are consistent with tang and baker (2016) and suggest that selfesteem plays a statistically significant role in each of the remaining three variables being studied: objective knowledge, subjective knowledge, and financial behavior. the results in table 4 need to be interpreted with caution due to the possible endogeneity issues among four financial literacy measures and omitted variable problem. we cannot completely rule out the broader theoretical concern of a reverse causality among these variables although our hausman test statistics for endogeneity cannot reject the null hypothesis of no measurement error. as for omitted variable bias issue, some variables such as subjective knowledge and self-esteem measure may potentially cause omitted variable bias. future studies can test and expand upon our results by incorporating additional control variables when data are available. 5.3. post-survey results we compare the results of the preand post-survey using our test sample of 829 matched students. we define improvement in several ways. for subjective financial knowledge and financial behavior, we define gains as the post-survey scores minus the pre-survey response scores. for objective financial knowledge questions, we define gains in financial knowledge as the difference between the post-survey scores and the pre-survey response. to gauge financial self-esteem, we define confidence gains as a decrease in the responses of i don’t know in the post-survey minus the pre-survey. we run univariate tests on the gains in our test samples and subsamples. table 5 illustrates the t test results by question and overall score for each of the four characteristics measured: subjective financial knowledge, financial behavior, objective financial knowledge, and financial self-esteem. the results show a profound, statistically significant improvement across all areas measured, with 24 of 25 questions showing total score improvement as statistically significant at the 1% level. as a result, the data shows financial literacy educational efforts can lead to better student results in financial literacy understanding and behavior. within subjective financial knowledge, all questions result in statistically significant improvement at the 1% level. the biggest gains come from understanding of roth ira (gain of 1.682) and retirement (1.401). seven of the 12 questions result in a greater than 25% improvement. the biggest gain in financial behavior derives from conversations with parents on personal finance (gain of 0.033). these results closely mirror the research of filbeck and zhao (2018), who find the largest growth within the same three questions. g. filbeck et al. / financial services review 28 (2020) 315–340 327 t ab le 5 t -t es t re su lt s b et w ee n p re an d p o st -s u rv ey p re p o st d if f t -s ta t p an el a . f in an ci al su b je ct iv e k n o w le d g e q u es ti o n s 2 . i u n d er st an d h o w to es ta b li sh a fi n an ci al p la n . 3 .0 4 7 4 .0 0 0 0 .9 5 3 2 1 .6 2 * * * 3 . i th in k fi n an ci al li te ra cy is im p o rt an t fo r m y fu tu re . 4 .3 9 1 4 .6 9 2 0 .2 8 0 1 0 .3 0 * * * 6 . i u n d er st an d th e p ro ce ss b y w h ic h m y p ar en ts /g u ar d ia n s m ak e fi n an ci al d ec is io n s. 3 .3 6 9 3 .8 3 2 0 .4 6 3 1 1 .3 2 * * * 7 . i k n o w h o w to d et er m in e th e ap p ro p ri at e to ta l co st s as so ci at ed w it h th e co ll eg es /u n iv er si ti es i am in te re st ed in at te n d in g . 3 .1 1 9 3 .8 6 5 0 .7 4 5 1 6 .3 5 * * * 8 . i u n d er st an d th e p ro ce ss b y w h ic h lo an re p ay m en ts ta k e p la ce in cl u d in g th e im p ac t o f in te re st , d el in q u en cy an d d ef au lt . 2 .7 2 3 3 .8 3 2 1 .1 0 9 2 3 .7 6 * * * 9 . i u n d er st an d th e p ro ce ss b y w h ic h cr ed it ca rd ch ar g es an d re p ay m en t sc h ed u le s ca n im p ac t th e le v el o f fi n an ci al d eb t le v el s. 3 .4 1 8 4 .1 9 9 0 .7 7 7 1 7 .0 5 * * * 1 0 . w h en it co m es to p u rc h as in g a ca r, i k n o w h o w to d et er m in e h o w m u ch o f a ca r i ca n af fo rd . 3 .2 5 8 4 .1 5 0 0 .8 9 3 1 9 .8 3 * * * 1 1 . i u n d er st an d h o w to ev al u at e th e co st -b en efi t an al y si s o f tr ai n in g fo r th e jo b i w o u ld li k e to p er fo rm af te r co m p le ti n g sc h o o l. 2 .9 9 8 3 .8 7 7 0 .8 7 8 1 9 .2 5 * * * 1 2 . i k n o w w h at a r o th ir a is an d h o w it w o rk s fr o m a ta x at io n st an d p o in t. 1 .8 7 5 3 .5 5 5 1 .6 8 2 3 2 .3 5 * * * 1 3 . i k n o w h o w to cr ea te a sa v in g s p la n b as ed o n th e ab il it y to es ti m at e m o n th ly li v in g ex p en se s. 3 .1 1 0 4 .1 4 8 1 .0 3 6 2 2 .4 1 * * * 1 4 . i k n o w h o w to p la n fi n an ci al ly fo r re ti re m en t. 2 .5 5 5 3 .9 5 0 1 .4 0 1 2 8 .2 8 * * * t o ta l sc o re fo r fi n an ci al su b je ct iv e k n o w le d g e 3 3 .7 8 0 4 3 .6 7 7 9 .8 9 7 2 9 .1 7 * * * p an el b . f in an ci al b eh av io r q u es ti o n s 1 . i li k e to sa v e m o n ey m o re th an i li k e to sp en d it . 3 .5 3 1 3 .8 0 9 0 .2 7 9 6 .9 7 * * * 4 . i h av e a ch ec k in g an d /o r a sa v in g s ac co u n t. 4 .3 8 0 4 .2 6 8 0 .0 4 3 1 .0 6 5 . i h av e co n v er sa ti o n s w it h m y p ar en ts re g ar d in g p er so n al fi n an ce . 3 .4 6 8 3 .7 9 9 0 .3 3 0 7 .5 5 * * * 1 5 . i th in k it is im p o rt an t to co n tr ib u te to a re ti re m en t p la n (e x : r o th ir a , 4 0 1 k , et c. ) 4 .0 6 0 4 .4 8 1 0 .4 2 1 1 0 .7 8 * * * t o ta l sc o re fo r fi n an ci al b eh av io r 1 4 .7 9 5 1 6 .3 3 4 1 .5 3 9 1 3 .4 1 * * * p an el c . o b je ct iv e q u es ti o n s (c o rr ec t a n sw er s) a 1 . is it sa fe r to p u t y o u r m o n ey in to o n e in v es tm en t o r p u t y o u r m o n ey in to m u lt ip le in v es tm en ts ? 0 .2 5 0 0 .6 9 3 0 .0 5 3 2 .5 3 * * 2 . if y o u in v es t $ 1 0 0 in a r o th ir a an d ea rn 5 % p er y ea r fo r 3 y ea rs , h o w m u ch w o u ld it b e w o rt h at th e en d o f 3 y ea rs . 0 .3 6 7 0 .4 8 7 0 .2 0 2 1 0 .0 3 * * * 3 . if y o u u se a cr ed it ca rd in ja n u ar y fo r a to ta l o f $ 5 0 0 , w h ic h p ay m en t o p ti o n w il l re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d . 0 .3 5 1 0 .7 2 9 0 .2 4 5 1 2 .5 5 * * * 4 . 4 . s u p p o se y o u d ec id e to b u y an a u d i fo r $ 5 0 ,0 0 0 . if y o u ta k e o u t an au to lo an fo r 5 y ea rs w it h 5 % in te re st , h o w m u ch to ta l w il l y o u p ay p er y ea r? 0 .3 0 6 0 .5 5 5 0 .1 3 1 6 .1 8 * * * 5 . in th e fu tu re , th e co st o f th in g s y o u b u y d o u b le s b u t y o u r in co m e re m ai n s th e sa m e. h o w m u ch w il l y o u b e ab le to b u y in th e fu tu re in co m p ar is o n to to d ay ? 0 .1 9 0 0 .7 6 6 0 .1 9 4 9 .9 2 * * * 6 . s u p p o se y o u h av e $ 4 0 ,0 0 0 in st u d en t d eb t. w h ic h p ay m en t o p ti o n w il l re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d ? 0 .3 0 6 0 .7 1 4 0 .1 8 2 9 .0 1 * * * t o ta l sc o re fo r o b je ct iv e q u es ti o n s (c o rr ec t a n sw er s) 2 .9 3 6 3 .9 4 4 1 .0 0 7 1 5 .8 8 * * * (c o n ti n u ed o n n ex t p a g e) 328 g. filbeck et al. / financial services review 28 (2020) 315–340 t ab le 5 (c o n ti n u ed ) p re p o st d if f t -s ta t p an el d . o b je ct iv e q u es ti o n s (“ i d o n ’t k n o w ” a n sw er s) b 1 . is it sa fe r to p u t y o u r m o n ey in to o n e in v es tm en t o r p u t y o u r m o n ey in to m u lt ip le in v es tm en ts ? 0 .6 4 0 0 .0 5 2 �0 .1 9 8 �1 2 .4 2 * * * 2 . if y o u in v es t $ 1 0 0 in a r o th ir a an d ea rn 5 % p er y ea r fo r 3 y ea rs , h o w m u ch w o u ld it b e w o rt h at th e en d o f 3 y ea rs . 0 .2 8 6 0 .1 3 8 �0 .2 2 9 �1 2 .5 0 * * * 3 . if y o u u se a cr ed it ca rd in ja n u ar y fo r a to ta l o f $ 5 0 0 , w h ic h p ay m en t o p ti o n w il l re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d . 0 .4 8 4 0 .0 7 8 �0 .2 7 3 �1 5 .6 9 * * * 4 . s u p p o se y o u d ec id e to b u y an a u d i fo r $ 5 0 ,0 0 0 . if y o u ta k e o u t an au to lo an fo r 5 y ea rs w it h 5 % in te re st , h o w m u ch to ta l w il l y o u p ay p er y ea r? 0 .4 2 4 0 .1 1 0 �0 .1 9 6 �1 0 .7 6 * * * 5 . in th e fu tu re , th e co st o f th in g s y o u b u y d o u b le s b u t y o u r in co m e re m ai n s th e sa m e. h o w m u ch w il l y o u b e ab le to b u y in th e fu tu re in co m p ar is o n to to d ay ? 0 .5 7 1 0 .0 7 2 �0 .1 1 8 �7 .9 3 * * * 6 . s u p p o se y o u h av e $ 4 0 ,0 0 0 in st u d en t d eb t. w h ic h p ay m en t o p ti o n w il l re su lt in th e lo w es t am o u n t o f o v er al l in te re st p ai d ? 0 .5 3 2 0 .1 0 4 �0 .2 0 2 �1 1 .6 4 * * * t o ta l sc o re fo r o b je ct iv e q u es ti o n s (“ i d o n ’t k n o w ” an sw er s) 1 .7 7 0 0 .5 5 5 �1 .2 1 5 �1 8 .8 1 * * * n o te : t ab le 5 s h o w s th e tte st re su lt s o f st u d en t re sp o n se s to fi n an ci al b eh av io r an d k n o w le d g e q u es ti o n s b ef o re an d af te r th e fi n an ci al li te ra cy ed u ca ti o n al ef fo rt s fo r th e te st sa m p le . * * * , * * , * in d ic at e st at is ti ca l si g n ifi ca n ce at 0 .0 1 , 0 .0 5 , an d 0 .1 0 le v el , re sp ec ti v el y . a im p ro v em en t is in d ic at ed b y a p o si ti v e d if fe re n ce an d tst at (m o re p eo p le se le ct in g th e co rr ec t an sw er ). b im p ro v em en t is in d ic at ed b y a n eg at iv e d if fe re n ce an d tst at (l es s p eo p le se le ct in g “i d o n ’t k n o w ”) . g. filbeck et al. / financial services review 28 (2020) 315–340 329 additionally, both objective financial knowledge and self-esteem exhibit statistically significant improvements across the board. the biggest gains from objective financial knowledge are credit (gain of 0.245) and compounding interest (gain of 0.202). the biggest improvements in self-esteem also stem from credit (0.273 improvement) and compounding interest (0.229 improvement). the results show a link between confidence to answer a question (self-esteem) and correctness (objective financial knowledge). the t test analyzes hypotheses related to the objective financial knowledge questions. students experienced a positive gain in correct responses of 1.007 (statistically significant at the 1% level), which represents an improvement of 34%. furthermore, the mean total score for the test sample increased to almost four. at a statistically significance level of 1%, we reject null hypothesis and conclude that students are more likely to be more financially knowledgeable after completing financial literacy education. additionally, students experience an increase in financial self-esteem, as measured by the amount of questions answered with i don’t know. students experienced an improvement in the number of i don’t know responses of 1.215 (statistically significant at the 1% level), representing an improvement in financial self-esteem to answer the question. as a result, we reject null hypothesis and conclude that students are less likely to answer i don’t know and have greater confidence after completing financial literacy education. table 6 t-test results between preand post-survey for different subsamples financial behavior subjective questions objective questions pre post diff pre post diff pre post diff panel a. subsamples by gender male 14.797 16.246 1.449*** 34.492 43.535 9.043*** 3.173 4.041 0.868*** female 14.792 16.431 1.640*** 32.987 43.835 10.848*** 2.673 3.835 1.162*** panel b. subsamples by grade level freshman 14.143 15.918 1.776*** 35.694 44.082 8.388*** 2.878 3.980 1.102*** sophomore 14.565 16.387 1.823*** 34.903 43.774 8.871*** 2.887 3.952 1.065*** junior 14.340 16.294 1.954*** 33.010 44.340 11.330*** 2.782 3.970 1.188*** senior 15.053 16.381 1.328*** 33.758 43.379 9.621*** 3.006 3.930 0.924*** panel c. subsamples by gpa lower than median gpa 14.251 16.028 1.777*** 33.749 43.375 9.625*** 2.568 3.699 1.131*** higher than median gpa 15.389 16.668 1.279*** 33.814 44.008 10.193*** 3.339 4.211 0.872*** panel d. subsamples by preferred instructions methods learning by doing 14.956 16.491 1.535*** 33.738 44.002 10.263*** 2.969 4.044 1.075*** listening 14.752 16.238 1.486*** 34.227 43.467 9.240*** 2.934 3.826 0.891*** discussing with peers 14.756 16.261 1.505*** 34.185 43.519 9.333*** 2.919 3.878 0.959*** features visual support 14.968 16.440 1.472*** 33.680 43.984 10.305*** 2.975 4.049 1.074*** interactive with websites 14.667 16.262 1.596*** 34.191 44.018 9.827*** 3.218 4.040 0.822*** panel e. subsamples by favorite subjects english 14.494 16.165 1.671*** 33.890 43.616 9.726*** 2.451 3.506 1.055*** math 15.140 16.654 1.513*** 33.675 44.333 10.658*** 3.175 4.053 0.877*** science 14.596 16.136 1.539*** 33.645 43.193 9.548*** 3.061 4.013 0.952*** social studies 14.866 16.373 1.507*** 33.876 43.569 9.694*** 2.919 4.086 1.167*** note: shows the t-test results of student responses to financial behavior, subjective and objective questions before and after the financial literacy educational efforts for different subsamples. ***, **, * indicate statistical significance at 0.01, 0.05 and 0.10 level, respectively. 330 g. filbeck et al. / financial services review 28 (2020) 315–340 table 6 reports the t test results for different subgroups, which show statistically significant improvement across all subgroups at the 1% level, indicating a significant improvement after completing the financial literacy program. within the objective knowledge category, female students and students with low gpas exhibit greater improvement, a positive sign in learning potential, as these characteristics are more likely to lead to lower initial financial literacy (tables 2 and 4 findings). consistent with gerrans and heaney (2019), female students experience a greater improvement in financial behavior, financial knowledge, and objective knowledge in comparison to their male peers. next, we run regression analysis to examine how student characteristics and other control variables affect their knowledge and behavior gains. table 7 reports the regression results. panel a reports the regression results on student characteristics and school district characteristics after controlling for fixed effects of classes. specifically, the class fixed effects allow the class dummy variable to differ for each class and control for the variations across classes. because we include school district characteristics as control variables, and these data are the same in the same school district, we cluster the standard errors at the school district level. the dependent variable is diff_behav (diff_subj, diff_obj, diff_idk), which is the difference between the students’ preand post-study scores (postminus pre-) for the financial behavior (subjective, objective) questions. all the other variables are defined the same as in table 4 and listed in appendix c. we also add pre_behav, pre_subj pre_obj, and pre_idk in the regressions to test whether student gains in financial knowledge and behavior are affected by their presurvey knowledge and behavior. a negative coefficient for diff_idk indicates an improvement in self-esteem, as it means that students answer i don’t know less in the post-survey and select more answers that are correct or incorrect. the results show that students who are less knowledgeable or exhibit inferior financial behavior gain most in the study. similarly, students in school districts with higher poverty experience a statistically significant gain in financial behavior. this finding is encouraging as kaiser and menkhoff (2017) indicate that financial education is often less effective for lower (low and lower-middle) income clients (economics) due to lack of relatability to topics such as handling of debt, a fact also noted by fernandes and lynch (2014). our differing results may be attributable to the manner in which the curriculum from the cfa society pittsburgh directly addresses relatability in the context of typical, high-school appropriate, smaller cost purchases. such an approach is consistent with stolper and walter (2017) as they point out that the opportunity to relate financial literacy to various demographics in the context of their spending behavioral is key to program success. female students improve by a greater amount in financial knowledge across both subjective and objective questions. students in school districts with a higher poverty improve more in financial behavior. female students are more likely to experience an increase in self-esteem (statistically significant at the 1% level). the improvement is consistent with our findings in table 6 and the findings of gerrans and heaney (2019), which both find that female students benefit more from financial literacy education. next, we include other control variables such as favorite subjects and learning style and use fixed effect regressions controlling for classes differences. because there are no school district data included in these regressions, we cluster standard errors at the class level. panel b of table 7 reports the regression results. we use the same dependent variables as in panel g. filbeck et al. / financial services review 28 (2020) 315–340 331 t ab le 7 s h o w s th e re g re ss io n re su lt s o f th e te st sa m p le m o d el (1 ) m o d el (2 ) m o d el (3 ) m o d el (4 ) d ep . v ar .: d if f_ b e h a v d ep . v ar .: d if f_ s u b j d ep . v ar : d if f_ o b j d ep . v ar .: d if f_ id k c o ef fi ci en t z -s ta t c o ef fi ci en t z -s ta t c o ef fi ci en t z -s ta t c o ef fi ci en t z -s ta t p an el a . r eg re ss io n r es u lt s af te r co n tr o ll in g fo r fi x ed ef fe ct o f cl as se s, w it h st an d ar d er ro rs cl u st er ed at sc h o o l d is tr ic t le v el in te rc ep t 2 .7 6 2 1 .2 7 1 4 .5 6 1 .2 0 3 .8 0 7 2 .4 8 * * �2 .3 0 9 �1 .5 4 f em al e 0 .3 4 1 1 .5 7 1 .8 1 2 2 .1 0 * * 0 .2 8 7 1 .6 6 * �0 .4 3 5 �3 .0 1 * * * s o p h o m o re 0 .3 3 6 0 .7 3 1 .1 5 2 0 .4 1 0 .0 0 6 0 .0 2 �0 .2 1 6 �0 .5 4 ju n io r 0 .1 8 6 0 .5 5 3 .1 0 0 1 .4 0 0 .0 7 0 0 .2 8 �0 .1 8 3 �0 .5 5 s en io r �0 .4 3 1 �1 .5 5 1 .7 6 6 0 .8 2 �0 .0 4 7 �0 .1 5 �0 .0 2 9 �0 .0 8 g p a �0 .3 1 5 �2 .0 2 * * 0 .7 2 3 1 .3 9 �0 .0 9 6 �0 .9 3 0 .1 2 8 1 .1 9 l o g (p o p u la ti o n ) �0 .1 5 3 �0 .7 9 �1 .1 4 7 �0 .7 7 �0 .2 4 5 �1 .5 5 0 .0 9 3 0 .5 3 p o v er ty 7 .3 6 1 1 .9 3 * 6 .2 2 3 0 .2 3 �0 .7 1 4 �0 .2 4 �0 .1 1 7 �0 .0 4 p ct _ c o ll eg e 2 .4 1 9 1 .2 3 4 .4 9 0 0 .3 3 �0 .0 2 8 �0 .0 2 0 .0 6 7 0 .0 4 p an el b . r eg re ss io n re su lt s af te r co n tr o ll in g fo r fi x ed ef fe ct o f cl as se s, w it h st an d ar d er ro r cl u st er ed at cl as s le v el in te rc ep t 1 0 .3 4 9 1 0 .4 0 * * * 3 5 .3 6 8 1 2 .0 2 * * * 2 .1 3 0 5 .9 6 * * * 0 .8 9 1 4 .0 3 * * * p re _ b e h a v �0 .6 8 8 �1 2 .8 0 * * * p re _ s u b j �0 .8 3 6 �1 2 .8 3 * * * p re _ o b j �0 .6 4 7 �1 6 .2 3 * * * p re _ id k �0 .8 1 8 �2 5 .4 9 * * * f em al e 0 .1 5 1 0 .9 4 0 .6 6 0 0 .9 7 �0 .0 0 1 �0 .0 1 0 .0 6 8 1 .1 9 s o p h o m o re 0 .5 3 8 1 .1 6 0 .0 7 4 0 .0 4 �0 .0 3 1 �0 .1 0 �0 .3 0 7 �1 .3 9 ju n io r 0 .3 6 0 0 .6 6 0 .8 4 6 0 .5 9 0 .0 9 9 0 .5 1 �0 .1 1 6 �0 .6 3 s en io r 0 .2 5 4 0 .5 6 �0 .2 2 2 �0 .1 6 �0 .0 3 6 �0 .1 8 �0 .0 5 6 �0 .3 0 g p a 0 .2 5 6 1 .8 7 0 .4 4 4 1 .4 4 0 .1 9 9 2 .0 0 * * �0 .1 7 3 �3 .7 3 * * * e n g li sh �0 .0 7 1 �0 .3 6 �0 .2 6 0 �0 .3 3 �0 .3 8 7 �2 .8 8 * * * 0 .2 9 7 2 .0 8 * * m at h 0 .1 8 0 0 .5 9 0 .6 0 0 0 .6 7 �0 .1 2 1 �0 .7 2 0 .1 8 5 2 .1 3 * * s ci en ce �0 .1 5 0 �0 .6 3 �0 .3 5 6 �0 .4 4 �0 .0 9 4 �0 .5 7 0 .1 1 7 1 .0 5 l b d 0 .3 0 5 1 .3 7 1 .0 8 9 2 .1 0 * * 0 .3 0 3 3 .0 0 * * * �0 .1 8 5 �1 .7 9 * l is te n in g �0 .1 4 6 �0 .7 4 �0 .3 6 6 �0 .5 5 �0 .1 8 8 �1 .9 6 * * 0 .0 4 3 0 .6 2 d is cu ss in g �0 .1 1 1 �0 .4 5 �0 .3 5 1 �0 .5 1 �0 .1 0 5 �0 .7 8 0 .0 8 3 0 .9 9 v is u al 0 .1 0 7 0 .6 1 0 .8 2 1 1 .4 5 0 .2 5 0 2 .2 4 * * �0 .1 1 6 �1 .3 9 n o te : p an el a re p o rt s th e re g re ss io n re su lt s o n st u d en t ch ar ac te ri st ic s an d sc h o o l d is tr ic t ch ar ac te ri st ic s af te r co n tr o ll in g fo r fi x ed ef fe ct o f cl as se s, w it h st an d ar d er ro rs cl u st er ed at th e sc h o o l d is tr ic t le v el . p an el b re p o rt s th e re g re ss io n re su lt s o n st u d en t ch ar ac te ri st ic s af te r co n tr o ll in g fo r fi x ed ef fe ct o f cl as se s, w it h cl u st er ed st an d ar d er ro rs . p re _ b e h a v (p re _ s u b j, p re _ o b j, p re _ id k ) is th e st u d en t’ s to ta l sc o re fo r th e fi n an ci al b eh av io r (k n o w le d g e, o b je ct iv e) q u es ti o n s. d if f_ b e h a v (d if f_ s u b j, d if f_ o b j, d if f_ id k ) is th e d if fe re n ce b et w ee n th e st u d en t’ s p re an d p o st -s u rv ey sc o re s fo r th e fi n an ci al b eh av io r (k n o w le d g e, o b je ct iv e) q u es ti o n s. f em al e is a d u m m y v ar ia b le w h ic h is eq u al to 1 if th e st u d en t is a fe m al e st u d en t, an d 0 o th er w is e. s o p h o m o re (j u n io r, se n io r) is a d u m m y v ar ia b le w h ic h is eq u al to 1 if th e st u d en t is a so p h o m o re (j u n io r, se n io r) , an d 0 o th er w is e. e n g li sh (m at h , sc ie n ce ) is a d u m m y v ar ia b le w h ic h is eq u al to 1 if th e st u d en t’ s fa v o ri te su b je ct is e n g li sh (m at h , sc ien ce ), an d 0 o th er w is e. l b d (l is te n in g , d is cu ss in g , v is u al ) is a d u m m y v ar ia b le w h ic h is eq u al to 1 if th e st u d en t ch o o se s le ar n in g b y d o in g (l is te n in g , d is cu ss in g w it h p ee rs , fe at u re s v is u al su p p o rt ) as fa v o ri te in st ru ct io n m et h o d , an d 0 o th er w is e. g p a is a st u d en t’ s g ra d e p o in t av er ag e. l o g (p o p u la ti o n ) is th e lo g o f th e p o p u la ti o n in th e sc h o o l d is tr ic t. p o v er ty is th e p o v er ty ra te in th e sc h o o l d is tr ic t. p ct _ co ll eg e is th e p er ce n ta g e o f re si d en ts w h o h av e at ta in ed b ac h el o r d eg re e o r h ig h er . * * * , * * , * in d ic at e st at is ti ca l si g n ifi ca n ce at 0 .0 1 , 0 .0 5 an d 0 .1 0 le v el , re sp ec ti v el y . 332 g. filbeck et al. / financial services review 28 (2020) 315–340 a. we also add pre_behav, pre_subj pre_obj, and pre_idk in the regressions to test whether student gains in financial knowledge and behavior are affected by their pre-survey knowledge and behavior. the results show that students who are less knowledgeable or exhibit inferior financial behavior gain most in the study. students whose favorite instruction method is learning by doing tend to gain more in subjective financial knowledge. students with a high gpa, whose favorite subject is not english, whose favorite instruction method is learning by doing or visual tend to gain more in objective financial knowledge. students with lower self-esteem gain the largest amount of self-esteem. additionally, students with higher gpa and students that prefer learning by doing experience the largest self-esteem gain (statistically significant at the 1% and 10% level, respectively). 6. conclusions this research study investigates the effectiveness of a high school financial literacy campaign to significantly improve financial literacy in four areas: subjective financial knowledge, financial behavior, objective financial knowledge and self-esteem. the financial literacy campaigns within the study were launched by the cfa society pittsburgh based upon the book the missing semester. initially, the result of the pre-survey, taken by students before beginning the financial education program, are analyzed using a t test. the results show students with higher gpas are more likely to display better financial behavior and objective financial knowledge than students with lower gpas. students with lower gpas exhibit greater perceived knowledge (subjective financial knowledge) in the concepts of loans and cost-benefit analysis; however, the same students exhibit lower actual knowledge (objective financial knowledge) in the same categories. this finding is consistent with previous literature that shows a disconnect between actual and perceived knowledge, as well as a connection between poor financial understanding and negative debt implications (e.g., higher debt and higher borrowing costs). similarly, male students are more likely to exhibit better objective financial knowledge, while female students are more likely to exhibit lower financial self-esteem. subsequently, logistic regressions test the relationship of subjective financial knowledge, financial behavior, and objective financial knowledge. the results further display the link between gender and initial financial knowledge, as females score lower on pre-survey objective and subjective financial knowledge. in addition, higher gpa has a statistically significant effect on better financial behavior and objective financial knowledge, but worse subjective knowledge, reinforcing the findings about the disconnect between actual and perceived knowledge. furthermore, self-esteem is shown to play a statistically significant impact on both actual and perceived knowledge. students with higher self-esteem exhibit higher financial behavior, subjective financial knowledge, and objective financial knowledge (significant at the 1% level), signifying that self-esteem is an important part of the financial literacy equation. additionally, the initial positive link between objective and subjective financial knowledge is eliminated when we added the additional regression variable, selfesteem, which further emphasizing the importance of self-esteem. this finding is consistent g. filbeck et al. / financial services review 28 (2020) 315–340 333 with tang and baker (2016), which introduces the importance of self-esteem on financial behavior. students in school districts with lower income levels exhibit lower financial behavior and objective financial knowledge, while students in school districts with smaller populations score higher in subjective financial knowledge. to test the effectiveness of the financial literacy program, we then conducted a t test between results of the preand post-survey, taken after completion of the course. the t test analyzes the four major topic areas previously listed. total scores for financial behavior, subjective knowledge, objective knowledge, and self-esteem improve by 29.3%, 10.4%, 34.3%, and 68.6%, respectively. students experience a statistically significant improvement of an average 35.7% in all four topic areas at the 1% level. within subjective financial knowledge, the largest gains result from the concepts of roth ira and planning for retirement. financial behavior shows the largest improvement in having personal finance conversations with parents. objective financial knowledge and self-esteem improve the most for the concepts of compound interest and credit. overall, the characteristics with the lowest pre-survey scores show the greatest improvements in the post-survey scores. within gender, females exhibit the highest improvement in all four of the categories. also, students with lower gpas experience greater improvement in financial behavior and objective knowledge, while students with higher gpas improve more in subjective knowledge. students who prefer learning by doing and visual support experience the most improvement, while students whose favorite subject is english or social studies experience the largest improvement. based upon the analysis, statistically significant gains in subjective financial knowledge, financial behavior, objective financial knowledge, and financial self-esteem lead us to the conclusion that the cfa society pittsburgh financial literacy program is successful at increasing students’ chances of financial success. therefore, the analysis shows the program continues to be successful at attempting to confront the financial literacy crisis. as the program continues to improve and expand, we look forward to expanding the sample size and reach of the financial literacy efforts, especially to states where students may receive no mandated financial education during high school. as financial analysts, society will increasingly be looking to our profession to reduce the impact of financial illiteracy and to make a positive difference in our fiduciary duties for investor education. our study has obvious limitations. first, we do not control for the manner in which content is presented in the classroom—does it make a difference whether the program is spread out over an entire system or conducted in longer sessions over a shorter time period. we also do not control for the number of hours spent delivering the content. future research efforts will focus differences in program delivery. additionally, we did not introduce a true control group for comparison, as we did not want to jeopardize the main purpose of our program, giving students the tools for a better financial future, just to provide a control group. our primary recommendation, based on the experience of cfa society pittsburgh, is for financial professionals to consider taking a more active role in financial literacy outreach. who better to lead these efforts than those who are trained to understand its importance? 334 g. filbeck et al. / financial services review 28 (2020) 315–340 appendix a: presurvey state: school: teacher: student id: gender: gpa: grade: favorite subject in school: ____ english ____ math ____ social studies ____ science questions: 1. i like to save money more than i like to spend it. 2. i understand how to establish a financial plan. 3. i think financial literacy is important for my future. 4. i have a checking and/or a savings account. 5. i have conversations with my parents regarding personal finance. 6. i understand the process by which my parents/guardians make financial decisions. 7. i know how to determine the appropriate total costs associated with the colleges/universities i am interested in attending. 8. i understand the process by which loan repayments take place including the impact of interest, delinquency and default. 9. i understand the process by which credit card charges and repayment schedules can impact the level of financial debt levels. 10. when it comes to purchasing a car, i know how to determine how much of a car i can afford. 11. i understand how to evaluate the cost-benefit analysis of training for the job i would like to perform after completing school. 12. i know what a roth ira is and how it works from a taxation standpoint. 13. i know how to create a savings plan based on the ability to estimate monthly living expenses. 14. i know how to plan financially for retirement. 15. i think it is important to contribute to a retirement plan (ex: roth ira, 401k, etc.) learning preferences: i am able to master material when instruction includes: 1. learning by doing/manipulating objects 2. listening 3. discussing with peers 4. features visual support (e.g., powerpoint slides) 5. interactive with websites g. filbeck et al. / financial services review 28 (2020) 315–340 335 objective questions: 1. is it safer to put your money into one investment or put your money into multiple investments? a. one investment b. multiple investments c. i don’t know* 2. if you invest $100 in a roth ira and earn 10% per year for 3 years, how much would it be worth at the end of three years. a. more than $130 b. exactly $130 c. less than $130 d. i don’t know* 3. if you use a credit card in january for a total of $300, which payment option will result in the lowest amount of overall interest paid. a. the full amount ($300) b. the minimum payment required c. paying nothing ($0) d. i don’t know* 4. suppose you decide to buy a bmw for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? a. more than $10,000 b. exactly $10,000 c. less than $10,000 d. i don’t know* 5. in the future, the cost of things you buy doubles and your income also doubles. how much will you be able to buy in the future in comparison to today? a. less b. the same c. more d. i don’t know* 6. suppose you have $30,000 in student loans. which payment option would result in the lowest amount of overall interest paid? a. 10 years at $350 per month 336 g. filbeck et al. / financial services review 28 (2020) 315–340 b. 15 years at $270 per month c. 20 years at $230 per month d. i don’t know* *survey respondents were required to answer the question, so i don’t know answer allows students to select non correct/incorrect answer appendix b: post-survey state: school: teacher: student id: questions: 1. i like to save money more than i like to spend it. 2. i understand how to establish a financial plan. 3. i think financial literacy is important for my future. 4. i have a checking and/or a savings account. 5. i have conversations with my parents regarding personal finance. 6. i understand the process by which my parents/guardians make financial decisions. 7. i know how to determine the appropriate total costs associated with the colleges/universities i am interested in attending. 8. i understand the process by which loan repayments take place including the impact of interest, delinquency and default. 9. i understand the process by which credit card charges and repayment schedules can impact the level of financial debt levels. 10. when it comes to purchasing a car, i know how to determine how much of a car i can afford. 11. i understand how to evaluate the cost-benefit analysis of training for the job i would like to perform after completing school. 12. i know what a roth ira is and how it works from a taxation standpoint. 13. i know how to create a savings plan based on the ability to estimate monthly living expenses. 14. i know how to plan financially for retirement. 15. i think it is important to contribute to a retirement plan (ex: roth ira, 401k, etc.) learning preferences: i am able to master material when instruction includes: 1. learning by doing/manipulating objects 2. listening 3. discussing with peers 4. features visual support (e.g., powerpoint slides) 5. interactive with websites g. filbeck et al. / financial services review 28 (2020) 315–340 337 objective questions: 1. which is less risky: investing your money into one investment or multiple investments? a. one investment b. multiple investments c. i don’t know* 2. if you invest $100 in a roth ira and earn 5% per year for 3 years, how much would it be worth at the end of three years. a. more than $115 b. exactly $115 c. less than $115 d. i don’t know* 3. if you use a credit card in january for a total of $500, which payment option will result in the lowest amount of overall interest paid. a. the full amount ($500) b. the minimum payment required c. paying nothing ($0) d. i don’t know* 4. suppose you decide to buy a audi for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? a. more than $10,000 b. exactly $10,000 c. less than $10,000 d. i don’t know* 5. in the future, the cost of things you buy doubles but your income remains the same. how much will you be able to buy in the future in comparison to today? a. less b. the same c. more d. i don’t know* 6. suppose you have $40,000 in student debt. which payment option will result in the lowest amount of overall interest paid? a. 10 years at $450 per month b. 15 years at $365 per month 338 g. filbeck et al. / financial services review 28 (2020) 315–340 c. 20 years at $315 per month d. i don’t know* *survey respondents were required to answer the question, so i don’t know answer allows students to select non correct/incorrect answer appendix c: variable definitions dependent variables subjective total score for the financial subjective questions in the survey objective total score for the financial objective questions in the survey with correct answers behavior total score for the financial behavior questions in the survey idk total score for the financial objective questions in the survey with “i don’t know” answers diff_subj the difference between the students’ preand post-study scores (post minus pre) for the financial subjective questions. diff_obj the difference between the students’ preand post-study scores (post minus pre) for the financial objective questions with correct answers. diff_behav the difference between the students’ preand post-study scores (post minus pre) for the financial behavior questions. diff_idk the difference between the students’ preand post-study scores (post minus pre) for the financial objective questions with “i don’t know” answers. independent variables gender: female a dummy variable which is equal to 1 if the student is a female, and 0 otherwise. grade level: sophomore a dummy variable which is equal to 1 if the student is a sophomore, and 0 otherwise. junior a dummy variable which is equal to 1 if the student is a junior, and 0 otherwise. senior a dummy variable which is equal to 1 if the student is a senior, and 0 otherwise. favorite subject: english a dummy variable which is equal to 1 if the student’s favorite is english, and 0 otherwise. math a dummy variable which is equal to 1 if the student’s favorite is math, and 0 otherwise. science a dummy variable which is equal to 1 if the student’s favorite is science, and 0 otherwise. favorite learning style: lbd a dummy variable which is equal to 1 if the student’s favorite learning style is learning by doing (lbd), and 0 otherwise. listening a dummy variable which is equal to 1 if the student’s favorite learning style is listening, and 0 otherwise. discussing a dummy variable which is equal to 1 if the student’s favorite learning style is discussion, and 0 otherwise. visual a dummy variable which is equal to 1 if the student’s favorite learning style is visualization, and 0 otherwise. gpa: gpa: a student’s grade point average (gpa). school district characteristics: log(population) the log of the population in the school district. poverty the poverty rate in the school district. pct_college the percentage of residents who have attained bachelor degree or higher. g. filbeck et al. / financial services review 28 (2020) 315–340 339 references amagir, a., groot, w., van den brink, h., & wilschut, a. (2018). a review of financial-literacy education programs for children and adolescents. citizenship, social and economics education, 17, 56-80. asaad, c. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101-117. berg, j., dickhaut, j., & mccabe, k. (1995). trust, reciprocity, and social history. games and economic behavior, 10, 122-142. cupák, a., fessler, p., schneebaum, a., & silgoner, m. (2018). decomposing gender gaps in financial literacy: new international evidence. economics letters, 168, 102-106. fernandes, d., & lynch, j. (2014). financial literacy, financial education, and downstream financial behaviors. management science, 60, 8, 1861-1883. filbeck, g., & zhao, x. (2018). financial literacy outreach: the effectiveness of outreach efforts in high schools. journal of business and economic perspectives, 65, 87-117. gerrans, p., & heaney, r. (2019). the impact of undergraduate personal finance education on individual financial literacy, attitudes, and intentions. accounting & finance, 59, 1, 177-217. green, d., & riddell, w. (2012). understanding educational impacts: the role of literacy and numeracy skills. working paper: institute of labor economics. available at http://conference.iza.org/conference_files/ 2015_skill_mismatch/riddell_w5670.pdf hastings, j., madrian, b., & skimmyhorn, w. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5, 347-373. kabala, m., & natali, jr., , g. (2012). the missing semester. pittsburgh, pa: authors. kaiser, t., & menkhoff, l. (2017). does financial education impact financial literacy and finance behavior, and if so. the world bank economic review, 31, 611-630. lin, j., bumcrot, c., ulicny, t., lusardi, a., mottola, g., kieffer, c., & walsh, g. (2016). financial capability in the united states 2016. available at http://www.usfinancialcapability.org/downloads/nfcs_2015_report_ natl_findings.pdf lusardi, a., & mitchell, o. (2014). the economic importance of financial literacy: the evidence. journal of economic literature, 52, 5-44. mccannon, b., asaad, c., & wilson, m. (2016). financial competence, overconfidence, and trusting investments: results from an experiment. journal of economics and finance, 40, 590-606. pelletier, j. (2015). 2015 national report card on state efforts to improve financial literacy in high school. available at https://www.champlain.edu/centers-of-excellence/center-for-financial-literacy/report-nationalhigh-school-financial-literacy/2015-national-report-card preston, a., & wright, r. (2019). understanding the gender gap in financial literacy: evidence from australia. economic record, 95, 1-29. remund, d. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276-295. stolper, o., & walter, a. (2017). financial literacy, finance advice, and financial behavior. journal of business economics, 87, 581-641. tang, n., & baker, a. (2016). self-esteem, financial knowledge and financial behavior. journal of economic psychology, 54, 164-176. 340 g. filbeck et al. / financial services review 28 (2020) 315–340 the role of perceived quality of personal service in influencing trust and satisfaction with banks anders carlander, ph.d.a,*, amelie gamble, ph.d.a, tommy gärling, ph.d.a, jeanette carlsson hauff, ph.d.b, lars-olof johansson, ph.d.a, martin holmen, ph.d.b adepartment of psychology, university of gothenburg, p.o. box 100, s-405 30, gothenburg, sweden bschool of business, economics and law, university of gothenburg, p.o. box 600, s-405 30, gothenburg sweden abstract trust is of paramount importance to banks. previous research has shown that trust increases with repeated personal contacts. we investigate if this applies to the customer-employee relationship in banks. data from an on-line survey of 293 customers of swedish retail banks are used to construct indicator measures. by means of structural equation modeling we find that trust in the bank is influenced by perceived quality of personal service through employees’ perceived competence, perceived benevolence, and perceived transparency, and that satisfaction with the bank is influenced by perceived quality of personal service through perceived competence, perceived benevolence, perceived transparency, and trust. © 2018 academy of financial services. all rights reserved. jel classification: m keywords: retail banking; personal service; satisfaction; trust 1. introduction a trend during the last decades is a declining trust in other people as well as institutions (twenge, campbell, and carter, 2014). following this trend, trust in banks has likewise declined since the 1970s, further exaggerated by the financial crisis in 2008 (edelman, 2017; gallup, 2013). if trust is declining in banks, an important question is what role trust has for consumers’ financial decisions. the financial sector is often described as a “trust business” because of the * corresponding author. tel.: �46-31-786-30 92; fax: �46-31-786 46 28. e-mail address: anders.carlander@gu.se (a. carlander) financial services review 27 (2018) 83-98 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. credence qualities of the offered services, in conjunction with the frequently long time span until delivery of benefits. as a consequence, it has been argued that trust is of paramount importance for the consumer of financial services (sekhon, ennew, kharouf, and devlin, 2014). financial decisions have increased in importance to individuals—a trend that has augmented the need for research that addresses individuals’ ability to make such decisions (gough and niza, 2011). given the complexity of many financial decisions and the often replicated finding that individuals perceive they lack adequate knowledge (e.g., lusardi and mitchell, 2007), their coping strategies have naturally become an important focus. one strategy preferred by many customers is to personally interact with bank employees to obtain financial advice to compensate their lack of knowledge (e.g., hermansson, 2015). it has been demonstrated that trust increases with repeated personal contacts (e.g., jones and george, 1998; king-casas et al., 2005), and we assume that this also applies to the relationship between bank customers and bank employees. based on other research showing that attitudes towards companies are influenced by and influence attitudes towards employees (cf. plitt, savjani, and eagleman, 2015), a plausible further assumption is that customers’ trust in a bank changes with their trust in its employees with whom they interact. our aim in this study is to attempt to empirically demonstrate that the quality of personal service offered by bank employees influence trust in their bank through the most common determinants of trust. hence, as detailed below, we investigate whether perceived quality of personal service influences trust in banks through bank employees’ perceived competence, perceived benevolence, and perceived transparency. because another important question is whether personal service to some extent affects satisfaction with banks as previous research has shown (levy and hino, 2016), we also investigate the possibility that this relationship is mediated by trust that in previous research has been shown to have a direct effect on satisfaction (geyskens, steenkamp, and kumar, 1999). if personal service is important for bank customers’ trust, it implies access to an important tool for increasing trust or counteracting a declining trust. training of bank employees is one avenue to achieve this—preferably by explicitly targeting trust-building factors such as competence and benevolence. transparent communication with customers is yet another practical tool, proven in previous research to be effective (anderson and weitz, 1992). the role of personal service for trust may also be important to take into account in digital contexts, in particular when considering that younger people are more skilled using new technology and less likely to want a personal relationship with the bank (kpmg nunwood, 2016). the remaining of this article is organized as follows. first, we review relevant previous research on trust, personal service, and satisfaction. then we propose our hypotheses. after this we describe the method and results of an internet survey to bank customers. finally, we discuss the survey results and implications for theory and practice. 2. review of previous research 2.1. trust trust is a multifaceted construct with a multitude of different definitions (kantzberger and kuntz, 2010). a common feature is that trust is associated with a willingness to depend on 84 a. carlander et al. / financial services review 27 (2018) 83-98 another party (hong and cha, 2013; sekhon et al., 2014). one implication is that people substitute lack of knowledge for trust in experts (carlander, 2015; dia, 2011; siegrist and cvetkovich, 2000) such that trust will play the role of coping with uncertainty (colquitt et al., 2012). knowledge of determinants of trust is essential for banks to proactively address the issue of increasing trust through personal service. conceptualizations of trust determinants draw on variants of competence, benevolence, and integrity (e.g., ennew, kharouf, and sekhon 2011; in their study including also shared values and communication). in this study we similarly propose that competence, benevolence, and transparency are primary determinants of customers’ trust in banks, and that these determinants are all affected by personal service as follows. (1) competence as a determinant of trust in banks refers to trustees’ knowledge and skill needed to complete a specific task (malhotra and lumineau, 2011). financial services are by many customers perceived to be complex and difficult to understand. bank employees’ competence may, therefore, be valuable to customers and contribute to building trust in the bank. (2) benevolence is the expectation that the trustee acts in the trustor’s best interests (sirdeshmukh, singh, and sabol, 2002). similarly, integrity and shared values (e.g., ennew, kharouf, and sekhon 2011) refer to a sense of moral, ethics, and consistency beyond the sole interests of the trustee. benevolence may be important for customers’ trust in a bank since mass media coverage of high profits, high bonus payments to employees, and financial turmoil signal that banks are not acting in the customers’ best interests. (3) transparency (e.g., kanagaretnam, mestelman, naiar, and shehata, 2010) may be defined as providing sufficient information to satisfy the requirements of a reasonable person (rawlins, 2009). there has been a growing interest in the value of transparency and transparent communication for attempts to increase trust (rawlins, 2009). transparent information is also believed to increase rational decision making in financial markets (lusardi and mitchell, 2007). 2.2. personal service the importance of personal service has been claimed to increase in services with a high level of intangibility (berry, 2000) as well as a people-intensive character (sirdesmukh, singh, and sabol, 2002). being a customized type of service high in credence qualities, financial services have further been described as part of a service industry where the relation between employees and customers is particularly important (yim, chan, and lam, 2012). in line with this, dissatisfaction with in-branch banking is one factor that has been shown to push customers into alternative service distribution channels such as internet banking (devlin and yeung, 2003). it has been recognized that a high quality of personal service is instrumental for creating and maintaining satisfied customers (crosby, evans, and cowles, 1990). previous studies have also shown that the quality of personal service is a direct determinant of satisfaction (krishnan et al., 1999) as well as of perceived service quality (taylor and baker, 1994). in bloemer, de ruyter, and peeters (1998) perceived service quality was furthermore both directly and indirectly related to bank loyalty through satisfaction. we argue that an additional benefit of personal service is that it influences the determinants of customers’ trust in a bank, and through trust satisfaction with the bank. the 85a. carlander et al. / financial services review 27 (2018) 83-98 importance of face-to-face meetings when building trust has been documented (howcroft, hewer, and durkin, 2003). in personal contacts employees may act in a way such that they are perceived to be competent by giving customers good advice, perceived to be benevolent by prioritizing customers’ interests, and be transparent by disseminating relevant information to customers in an understandable way. as a consequence, customers may both perceive that the banks’ personal service is of high quality and develop trust in the bank, which is one of the channels through which their satisfaction with the bank is influenced. 2.3. satisfaction a positive relation between trust and satisfaction has been empirically verified in several meta-analyses (geyskens, steenkamp, and kumar, 1999). particularly relevant are the results from a survey focusing explicitly on banks, which found that trust is the most important factor affecting customers’ satisfaction (chakravarty, feinberg, and widdows, 1997). the difference between the two constructs has been suggested to rely on the duration of the relation, with trust being a key variable in novel settings, and satisfaction more important in maintaining an existing relationship (selnes, 1998). this is consistent with our conjectures that trust in a bank has the role of reducing uncertainty, and that trust subsequently is a determinant of satisfaction with the bank. additionally, because trust is believed to increase with time (jones and george, 1998), as well as become more relational and emotionally oriented over time (lewicki and bunker, 1995), customers� satisfaction would increase with length of their relation to the bank. customers are, therefore, likely to be loyal to a bank that they trust, a result that has been empirically documented (shainesh, 2012). 2.4. hypotheses previous research has shown that determinants of trust include competence, benevolence, and transparency (e.g., ennew, kharouf, and sekhon, 2011). our rationale for including these determinants of trust in banks is customers’ demand for competent advice by bank employees, customers’ demand for a benevolent advisor to prevent them from being cheated by the bank, and customers’ demand for transparency by the bank employees to facilitate accurate communication in contacts with the bank. therefore, we propose hypothesis 1: the quality of personal service increases trust in banks through its determinants perceived competence, perceived benevolence, and perceived transparency. a direct effect of quality of personal service on satisfaction with banks has been documented (levy and hino, 2016; moin, devlin, and mckechnie, 2015). however, we expect that this relationship will be weaker when measuring the indirect effect of trust. we thus expect that the relationship between quality of personal service and satisfaction is mediated by trust and its determinants perceived competence, perceived benevolence and perceived transparency. hypothesis 2: perceived quality of personal service has an indirect effect on satisfaction through trust and its determinants perceived competence, perceived benevolence, and perceived transparency. 86 a. carlander et al. / financial services review 27 (2018) 83-98 3. study 3.1. overview to test hypotheses 1 and 2 an on-line survey questionnaire was administered to bank customers asking them questions that provide indicators of the constructs entailed by the hypotheses. we then use covariance-based structural equation modeling (sem) to estimate the expected direct and indirect effects associated with the latent constructs perceived competence, perceived benevolence, perceived transparency, trust, and satisfaction (see fig. 1). in sem (see, e.g., bollen, 1989; kaplan, 2009) the regression coefficients of a theoretically specified system of linear regression equations are simultaneously estimated. this is made by comparing covariances between constructs derived from the regression equations to observed covariances computed from the indicators. several measures of goodness-of-fit of the model are computed. 3.2. method questionnaire data were collected in an on-line survey administered in january 2013 to participants randomly selected from a customer database maintained by the swedish bank skandiabanken. skandiabanken is an internet-only bank providing special services. only a minority of participants (25.3%) indicated in the questionnaire that it was their main bank, whereas approximately equal numbers of the majority indicated that their main bank was one of the four largest banks in sweden (handelsbanken, nordea, seb, and swedbank). participants were invited by e-mail to answer a questionnaire that in pretests took about 10 min to complete. the e-mail explained the general purpose of the study, informed about the funders of the research (a governmental agency and skandiabanken), and whom to contact to make enquiries. participants were guaranteed confidentiality and were not promised any monetary compensation. a reminder e-mail was sent to those who after three days had not replied. the invitations were sent to 4,968 potential participants of which 293 (5.9%) fig. 1. the hypothesized direct effects of perceived quality of personal service on perceived competence, perceived benevolence, and perceived transparency, the direct effects of perceived transparency, perceived competence and perceived benevolence on trust, and the indirect effect of perceived quality of personal service on satisfaction through perceived competence, perceived benevolence, perceived transparency, and trust. (solid lines represent the direct effects, broken lines the indirect effect.) 87a. carlander et al. / financial services review 27 (2018) 83-98 completed the questionnaire (28.0% women, age ranging from 21 to 79 years with a mean of 55.4 years). sample descriptives obtained from the questionnaire are given in table 1. a detailed description of the questionnaire is given in carlander (2015). briefly, participants were in the indicated order asked questions about frequency and types of bank contacts, quality of personal service, trust in and satisfaction with their main bank (defined as the bank with which they had the closest relation), trust in society and its institutions, table 1 sample descriptives (n � 293) variable value sex (% women) 28.0 mean age (years)a 55.4 education (spouse) elementary school (% degree) 6.1 (8.5) high-school (% degree) 23.9 (18.4) college (% degree) 69.6 (54.3) employment (spouse) employed (%) 60.4 (50.5) self-employed/entrepreneur (%) 17.4 (11.3) agriculturist (%) 0.3 (0.3) student (%) 1.4 (0.7) unemployed (%) 0.3 (1.4) retired/house-wife, -man (%) 28.7 (20.8) other type of occupation (%) 3.1 (1.0) married/cohabiting (%) 81.6 annual household income (sek)b 200,000 or less (%) 1.4 201,000–350,000 (%) 7.8 351,000–500,000 (%) 12.6 501,000–650,000 (%) 19.8 651,000–800,000 (%) 15.0 801,000–950,000 (%) 13.0 more than 950,000 (%) 25.6 disposal assets (cash, savings, investments) (sek) 10,000 or less (%) 1.4 11,000–50,000 (%) 5.5 51,000–150,000 (%) 10.2 151,000–300,000 (%) 12.6 301,000–600,000 (%) 16.7 601,000–900,000 (%) 10.2 more than 901,000 (%) 34.8 type of home house/garden apartment 63.1 apartment/flat (%) 35.2 other 1.4 type of agreement rented apartment (%) 14.7 owned apartment (%) 21.8 owned house (%) 60.8 other 2.3 a national average in sweden: 50.1 % women (statistics sweden, 2013), mean age 41.1 years (statistics sweden, 2012). other national averages are given within brackets. b 1sek was approximately 0.16 us$ and 0.11 eur€ at the time of the study. 88 a. carlander et al. / financial services review 27 (2018) 83-98 financial literacy assessed through both self-reported knowledge and true/false responses to knowledge questions, and questions about socio-demographic variables. the model test uses only the questions about quality of personal service, trust, and satisfaction described below. we based our trust questions on those developed in a larger set of studies (van raaij and van esterik-plasmeijer, 2017) aimed at measuring trust in the finance sector. we received from these authors (personal communication, november 2011) an earlier version of the questions that at that time had been used to measure trust in insurance companies. the questions we used asked participants to rate statements about their main bank. they had been translated, modified, and pretested in two pilot studies to obtain reliable indicators of the latent variables perceived competence, perceived benevolence, perceived transparency, and trust.1 similarly phrased statements about perceived quality of personal service and satisfaction were constructed by us. the participants rated each statement on a seven-point likert-type numeric scale ranging from 1 “totally disagree” to 7 “totally agree.” the statements (translated from swedish) are given in table 2. in the questionnaire the statements were presented in a counterbalanced order. 3.3. results firstly, we assess discriminant and convergent validity of the theoretical constructs (hair et al., 2010). secondly, we use covariance-based sem (bollen, 1989; kaplan, 2009) to test the relationships between the latent constructs. we report goodness-of-fit of the model as well as estimates of standardized regression coefficients corresponding to total, direct and indirect effects as specified in the model. because missing data are not permitted in the analyses, we conducted a preliminary test showing that the missing values were not systematic (little’s mcar test, c2 � 1376.47, df �1470, p � 0.960). expectation maximization was, therefore, used for replacing the missing values. table 2 reports means, standard deviations, skewness, kurtosis, and intercorrelations of the indicators. by means of the maximum likelihood method in ibm spss amos 21, we first tested a six-factor measurement model of the latent constructs satisfaction, trust, perceived competence, perceived benevolence, perceived transparency, and perceived quality of personal service. this resulted in a marginally acceptable model fit: c2 � 377.84, df � 120, p � 0.001; normed fit index (nfi) � 0.92; comparative fit index (cfi) � 0.95; and root mean square error of approximation (rmsea) � 0.09. the standardized indicator weights are given in table 2. with one exception these are uniformly high. in contrast, a one-factor measurement model yielded an unacceptable fit, c2 � 691.56, df � 90, p � 0.001; nfi � 0.84; cfi � 0.85; and rmsea � 0.15. table 3 reports the correlations between constructs, composite reliability (cr), cronbach’s �s, average variance extracted (ave), maximum shared squared variance (msv), and average shared squared variance (asv). as recommended by hair et al., (2010) and fornell and larcker (1981), all crs and �s are above 0.70. furthermore, the recommended criterion for convergent validity (hair et al., 2010) is fulfilled in that ave is above 0.50 and cr is larger than ave. however, the criterion for discriminant validity, that ave should be larger than both msv and asv, is only approximately fulfilled. in the main diagonal of the 89a. carlander et al. / financial services review 27 (2018) 83-98 t ab le 2 m ea ns (m ), st an da rd de vi at io ns (s d ), st an da rd iz ed sk ew ne ss , st an da rd iz ed ku rt os is , st an da rd iz ed w ei gh t co ef fic ie nt s (w ) of ra tin gs of st at em en ts (i nd ic at or va ri ab le s) on sc al es ra ng in g fr om 1 to 7, an d pr od uc t m om en t co rr el at io ns be tw ee n in di ca to rs l at en t va ri ab le in di ca to rs m sd sk ew ne ss k ur to si s w sa tis fa ct io n 1. i am sa tis fie d w ith m y ba nk 5. 28 1. 36 � 5. 25 1. 11 0. 89 2. i am pl ea se d w ith m y ba nk 5. 20 1. 41 � 5. 35 0. 76 0. 95 3. i am de lig ht ed w ith m y ba nk 5. 31 1. 36 � 6. 61 2. 61 0. 96 0. 96 t ru st 4. i tr us t m y ba nk 5. 50 1. 32 � 6. 18 1. 35 0. 81 5. i fe el gr ea t co nfi de nc e in m y ba nk 5. 05 1. 40 � 5. 46 1. 26 0. 93 6. m y ba nk is tr us tw or th y 5. 30 1. 28 � 4. 51 0. 76 0. 87 pe rc ei ve d co m pe te nc e 7. m y ba nk un de rs ta nd s co m pl ex fin an ci al pr ob le m s 4. 74 1. 26 � 2. 47 0. 13 0. 71 8. m y ba nk is m ak in g de lib er at e de ci si on s 4. 61 1. 12 � 0. 59 � 0. 79 0. 70 9. m y ba nk is fin an ci al ly co m pe te nt 5. 26 1. 12 � 4. 48 1. 52 0. 77 pe rc ei ve d be ne vo le nc e 10 . m y ba nk is be ne vo le nt to w ar ds m e 4. 26 1. 60 � 1. 53 � 2. 34 0. 84 11 . m y ba nk m ak es de ci si on s th at be ne fit m e 3. 95 1. 32 � 1. 52 � 0. 96 0. 87 12 . m y ba nk m ak es de ci si on s th at ar e be ne fic ia l fo r m e 4. 19 1. 51 � 1. 51 � 2. 38 0. 89 pe rc ei ve d tr an sp ar en cy 13 . m y ba nk di sc lo se s fe es an d co nd iti on s th or ou gh ly 5. 00 1. 54 � 3. 99 � 1. 30 0. 63 14 . m y ba nk pr ov id es tr ut hf ul in fo rm at io n w he n as ke d 5. 43 1. 24 � 4. 35 � 0. 62 0. 82 15 . m y ba nk pr ov id es al l th e in fo rm at io n i as k fo r 5. 42 1. 26 � 5. 68 2. 43 0. 86 pe rc ei ve d qu al ity of pe rs on al se rv ic es 16 . t he st af f at m y ba nk pr ov id es m e w ith th e ri gh t ty pe s of se rv ic es 4. 82 1. 49 � 2. 05 � 2. 96 0. 80 17 . i al w ay s fe el w el co m ed by th e st af f at m y ba nk 5. 54 1. 46 � 7. 09 2. 24 0. 77 18 . i re ce iv e ad eq ua te he lp fr om th e st af f at m y ba nk 5. 34 1. 35 � 5. 39 0. 77 0. 88 (c on ti nu ed on ne xt pa ge ) 90 a. carlander et al. / financial services review 27 (2018) 83-98 t ab le 2 (c on ti nu ed ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 2. 0. 86 3. 0. 85 0. 90 4. 0. 72 0. 73 0. 75 5. 0. 75 0. 83 0. 82 0. 77 6. 0. 64 0. 70 0. 76 0. 70 0. 80 7. 0. 48 0. 50 0. 49 0. 50 0. 54 0. 51 8. 0. 48 0. 56 0. 54 0. 44 0. 62 0. 58 0. 50 9. 0. 47 0. 53 0. 55 0. 49 0. 60 0. 65 0. 56 0. 52 10 . 0. 57 0. 66 0. 64 0. 58 0. 68 0. 65 0. 41 0. 46 0. 45 11 . 0. 57 0. 66 0. 66 0. 55 0. 67 0. 62 0. 42 0. 56 0. 45 0. 73 12 . 0. 56 0. 66 0. 67 0. 53 0. 69 0. 69 0. 48 0. 52 0. 49 0. 73 0. 77 13 . 0. 51 0. 54 0. 53 0. 52 0. 57 0. 57 0. 32 0. 42 0. 35 0. 52 0. 48 0. 47 14 . 0. 45 0. 54 0. 56 0. 58 0. 65 0. 72 0. 47 0. 49 0. 51 0. 54 0. 56 0. 64 0. 49 15 . 0. 50 0. 59 0. 63 0. 54 0. 67 0. 72 0. 45 0. 47 0. 52 0. 54 0. 54 0. 58 0. 51 0. 73 16 . 0. 63 0. 76 0. 72 0. 55 0. 65 0. 59 0. 46 0. 51 0. 49 0. 55 0. 65 0. 64 0. 47 0. 54 0. 58 17 . 0. 55 0. 63 0. 68 0. 50 0. 63 0. 63 0. 41 0. 40 0. 52 0. 54 0. 47 0. 56 0. 41 0. 52 0. 56 0. 60 18 . 0. 57 0. 66 0. 70 0. 57 0. 69 0. 70 0. 52 0. 45 0. 56 0. 56 0. 56 0. 63 0. 48 0. 70 0. 74 0. 70 0. 72 a ll co rr el at io ns ar e st at is tic al ly si gn ifi ca nt at p � 0. 00 1. 91a. carlander et al. / financial services review 27 (2018) 83-98 correlation matrix, the square roots of ave are given for each construct. as may be seen, these are lower than the correlations with the other constructs for perceived competence, perceived transparency, trust, and perceived quality of personal services. an overlap between indicators of these constructs is hence implied. adding a common latent factor improved the model fit marginally, c2 � 284.01, df � 102, p � 0.001; nfi � 0.94; cfi � 0.96; and rmsea � 0.08. still suggesting that a common latent factor may account for the covariance between the indicators, the standardized indicator weights were reduced for all the constructs (see mackenzie and podsakoff, 2012). the proposed structural model (fig. 1) tested by the maximum likelihood method resulted in a marginally acceptable fit, c2 � 443.05, df � 127, p � 0.001; nfi � 0.91; cfi � 0.93; and rmsea � 0.09. two alternative models were also tested. a poor fit was obtained for the model positing that satisfaction is directly determined by trust, perceived competence, perceived benevolence, perceived transparency, and perceived quality of personal service, c2 � 1489.23, df � 130, p � 0.001; nfi � 0.70; cfi � 0.72; and rmsea � 0.19, as well as for the other model, positing direct effects on satisfaction of trust and perceived quality of personal service, and indirect effects of perceived competence, perceived benevolence, and perceived transparency through trust, c2 � 1127.58, df � 130, p � 0.001; nfi � 0.77; cfi � 0.79; and rmsea � 0.16. table 4 gives standardized regression coefficient estimates of total, direct, and indirect effects. bootstrapping was used to obtain bias-corrected 95% confidence intervals (as recommended by hayes and scharkow, 2013) for a total of 2,000 replicates. as expected (hypothesis 1), a significant indirect effect is observed of perceived quality of personal service on trust through perceived competence, perceived benevolence, and perceived transparency. perceived competence, perceived benevolence, and perceived transparency also have significant direct effects on trust. furthermore, according to hypothesis 2 there is a significant indirect effect on satisfaction of perceived quality of personal service through perceived competence, perceived benevolence, perceived transparency, and trust. the direct effect of perceived personal service on satisfaction is positive although not significant. perceived competence, perceived benevolence, and perceived transparency have all significant indirect effects on satisfaction through trust. table 3 composite reliability (cr), cronbach’s �, average variance extracted (ave), maximum shared squared variance (msv), and average shared squared variance (asv), and correlations between constructs with square root of ave in the min diagonal construct cr � ave msv asv factor correlation matrix sat trust comp ben trans pers satisfaction (sat) 0.95 0.95 0.87 0.84 0.66 0.93 trust (trust) 0.90 0.90 0.76 0.84 0.77 0.92 0.87 competence (comp) 0.77 0.77 0.53 0.76 0.63 0.76 0.87 0.73 benevolence (ben) 0.90 0.89 0.75 0.71 0.63 0.79 0.84 0.75 0.86 transparency (trans) 0.82 0.79 0.60 0.80 0.67 0.73 0.89 0.79 0.79 0.78 personal service (pers) 0.87 0.86 0.68 0.79 0.70 0.85 0.86 0.80 0.80 0.89 0.82 92 a. carlander et al. / financial services review 27 (2018) 83-98 4. discussion our results obtained from the survey of customers of swedish retail banks support hypothesis 1 that that the quality of personal service increases the customers’ trust in their banks through perceived competence, perceived benevolence, and perceived transparency. we propose that the basis for customers’ perceptions of competence, benevolence, and transparency is personal contacts with the bank employees, but we are only able to infer this from the indirect effects in statistical model tests based on questionnaire data. additional research, therefore, needs to verify our results by observations of actual interactions between bank customers and employees. in support of hypothesis 2 we found that the quality of personal service influences satisfaction through trust and its determinants perceived competence, perceived benevolence, and perceived transparency. in a financial service context, previous studies of trust, including some aspects of the interaction between employee and customer (e.g., shainesh, 2012), have not attempted to draw implications for satisfaction. however, the focus on a satisfied customer is important from a practical perspective because customer satisfaction indices are measured regularly and attract much attention. assessing the determinants of satisfaction with specific banks has furthermore the virtue of identifying concrete measures aimed at increasing the level of customer satisfaction proximal to the management of the banks. previous research has reported a direct effect on satisfaction of personal service (crosby et al., 1990; krishnan et al., 1999; taylor and baker, 1994). a direct effect was, however, only weakly supported by our results. a direct effect may be explained by personal contacts that elicits affective (johnson and grayson, 2005) or relational (ponder, bugg holloway, and table 4 standardized estimates (�), 95% confidence intervals (ci), and p values for total, direct, and indirect effects personal service competence benevolence transparency trust � ci� p � ci� p � ci� p � ci� p � ci� p total effects competence 0.83 0.10 0.001 benevolence 0.83 0.12 0.001 transparency 0.91 0.14 0.001 trust 0.88 0.08 0.002 0.37 0.32 0.005 0.32 0.30 0.001 0.33 0.42 0.009 satisfaction 0.84 0.16 0.002 0.27 0.26 0.003 0.23 0.28 0.001 0.24 0.26 0.005 0.73 .50 0.002 direct effects competence 0.83 0.10 0.001 benevolence 0.83 0.12 0.001 transparency 0.91 0.14 0.001 trust 0.37 0.32 0.005 0.32 0.30 0.001 0.33 0.42 0.009 satisfaction 0.20 0.50 0.173 0.73 .50 0.002 indirect effects competence benevolence transparency trust 0.88 0.04 0.002 satisfaction 0.64 0.21 0.002 0.27 0.13 0.003 0.23 0.28 0.001 0.24 0.26 0.005 93a. carlander et al. / financial services review 27 (2018) 83-98 hansen, 2016) components, which are likely to be less salient when answering a questionnaire. this also speaks to the need for additional research using, for instance, in-depth interviews to disentangle direct affective or relational effects on satisfaction. we note that the sample we recruited represents customers of all the large swedish banks despite that is was obtained from the customer register of one of the smaller niche banks. this is probably because of the fact that individuals increasingly are customers of more than one bank and hence appear in several customer registers. for the purpose of our study, this creates an opportunity to generalize the results that would otherwise have not been possible. however, it also creates a potential bias since our questions pertained to the “main bank,” that is the bank with which the customers perceive having the closest relation. likely for this reason the ratings were generally positive, thus resulting in negatively skewed distributions that is a common finding in studies of customer satisfaction (moe, netzer, and schweidel, 2017). this poses a problem for the statistical analyses. yet, we found no differences in results in supplementary tests of our structural model using distribution-free estimation methods. another problem that the statistical analyses do not address is that the skewed (or j-shaped) distribution reflects that predominantly positive but also more negative than neutral customers were more likely to participate. although sampling biases cannot be discounted because of the low response rate, there are other possible causes of the positive ratings. a choice-supportive bias (mather and johnson, 2000) may have caused participants to be overly positive towards the bank of which they are regular customers. previous research has also found that bank customers judge their own financial advisor as more trustworthy than financial advisors in general (sunikka et al., 2010). it is, furthermore, possible that our somewhat older and more than average affluent participants may have closer relations to their banks than other customers, thus rating trust and satisfaction higher. in addition, 70% of the participants stated that they have been a customer with their main bank for more than 10 years. previous studies have reported that the longer a successful trusting relationship lasts, the stronger the trust is (gulati, 1995; lewicki and bunker, 1995). a halo effect (rosensweig, 2014) implies that satisfaction or trust with their own bank may have spilled over to participants’ evaluations of other aspects (sunikka et al., 2010). if the ratings we obtained have a common variance component because of the halo effect, it may account for the low discriminant validity. statistical support was, however, only partial for that a single construct explained some of the common variance of the indicators. another not mutually exclusive explanation includes a common method of data collection resulting in correlations between error variances. trust in and satisfaction with banks may be increased or decreased by, for example, media reporting about financial crises. this may also have potential spill-over effects that can affect similar financial products that are being discussed in the media (shin, 2009) as well as similar companies and institutions (dahlén and lange, 2006). it has been suggested that the 2008 financial crisis was sparked by a loss of trust (sapienza and zingales, 2012) that spread through the financial system like a pandemic outbreak. a serious outcome of such an outbreak is that it may create a self-fulfilling prophecy because banks are not fortified against a surge in liquidity outflow (diamond and dybvig, 1983). considering that both good and bad news may spill over in an interconnected financial system, an issue for future research 94 a. carlander et al. / financial services review 27 (2018) 83-98 is to understand how under such circumstances loss of trust would be regained to avoid the severity of financial meltdowns. additional research focusing on other customer segments is needed to generalize the results of the study. it is possible that a younger, less affluent sample would judge personal service to be less important considering the penetration of internet services, smartphone apps, and financial intermediaries, such as for example paypal, that makes financial errands convenient without a long-term bank relationship (kpmg nunwood, 2016). self-service technology is at least a complement that may serve as an important means of increasing product differentiation and thus satisfaction. whether automation increases or decreases trust in the provider is another issue for future research to address. 5. conclusions perceived quality of personal service deserves further empirical examination because is likely to be an important factor for both trust in and satisfaction with retail banks. competence, benevolence, and transparency are possible channels through which personal service influences trust and satisfaction. improving bank employees in these respects are, therefore, a potential means by which banks may increase their customers’ trust and satisfaction. note 1 we also included statements to measure stability as a determinant of trust as van raaij and van esterik-plasmeijer (2017) did. stability did however not show the expected correlation with trust. two other variables that we measured were “good citizenship” and “bank’s trust in customers.” “good citizenship” was excluded because it had a high correlation with benevolence, whereas the other variable did not correlate with trust. acknowledgment this research was supported by grant 2011-03872 from the swedish agency for innovation systems (vinnova) to the centre for finance, university of gothenburg, göteborg, sweden. we thank martin hedesström for comments on the manuscript. references anderson, e., & weitz, b. (1992). the use of pledges to build and sustain commitment in distribution channels. journal of marketing research, 29, 18–34. berry, l. l. (2000). cultivating service brand equity. journal of the academy of marketing science, 28, 128–137. bloemer, j., de ruyter, k., & peeters, p. (1998). investigating drivers of bank loyalty: the complex relationship between image, service quality and satisfaction. international journal of bank marketing, 16, 276–286. 95a. carlander et al. / financial services review 27 (2018) 83-98 bollen, k. a. (1989). structural equations with latent variables. new york, ny: wiley. carlander, a. (2015). trust as a strategy for handling uncertainty in private savings choices (doctoral dissertation). retrieved from gupea. (isbn: 978-91-628-9464-1). chakravarty, s., feinberg, r., & widdows, r. (1997). reasons of their discontent. bank marketing, 29, 49–52. colquitt, j. a., lepine, j. a., piccolo, r. f., zapata, c. p., & rich, b. l. (2012). explaining the justiceperformance relationship: trust as exchange deepener or trust as uncertainty reducer? journal of applied psychology, 97, 1–15. crosby, l. a., evans, k. r., & cowles, d. (1990). relationship quality in services selling: an interpersonal influence perspective. journal of marketing, 54, 68–81. dahlén, m., & lange, f. (2006). a disaster is contagious: how a brand in crisis affects other brands. journal of advertising research, 46, 388–397. devlin, j. f., & yeung, m. (2003). insights into customer motivations for switching to internet banking. international review of retail, distribution and consumer research, 13, 375–392. dia, e. (2011). uncertainty, trust, and the regulation of the banking industry. international review of economics, 58, 213–228. diamond, d. w., & dybvig, p. h. (1983). bank runs, deposit insurance, and liquidity. journal of political economy, 91, 401–419. edelman. (2017). edelman trust barometer (available at https://www.edelman.com/research/2017-trustbarometer-global-results). ennew, c., kharouf, h., & sekhon, h. (2011). trust in uk financial services: a longitudinal analysis. journal of financial services marketing, 16, 65–75. fornell, c., & larcker, d. f. (1981). evaluating structural equation models with unobservable variables and measurement error. journal of marketing research, 18, 39–50. gallup. (2013). confidence in institutions (available at http://www.gallup.com/poll/1597/confidince-institutions. aspx#1). gambetta, d. (1988). can we trust trust? in d. gambetta (ed.), trust: making and breaking cooperative relations (pp.725–737). oxford: university of oxford. geyskens, i., steenkamp, j. e. b. m., & kumar, n. (1999). a meta-analysis of satisfaction in marketing channel relationships. journal of marketing research, 36, 223–238. gough, o., & niza, c. (2011). retirement saving choices: review of the literature and policy implications. journal of population ageing, 4, 97–117. gulati, r. (1995). does familiarity breed trust? the implications of repeated ties for contractual choice in alliances. academy of management journal, 38, 85–112. hair, j. f. j., black, w. c., babin, b. j., anderson, r. e., & tatham, r. l. (2010). multivariate data analysis. upper saddle river, nj: prentice-hall. hayes, a. f., & scharkow, m. (2013). the relative trustworthiness of inferential tests of the indirect effect in statistical mediation analysis: does method really matter? psychological science, 24, 1918–1927. hermansson, c. (2015). understanding the relationships between bank-customer relations, financial advisory services and saving behaviour (doctoral dissertation). royal institute of technology, stockholm, sweden. isbn: 978-91-87111-02-0. hong, i. b., & cho, h. (2011). the impact of consumer trust on attitudinal loyalty and purchase intentions in b2c e-marketplaces: intermediary trust vs. seller trust. international journal of information management, 31, 469–479. howcroft, b., hewer, p., & durkin, m. (2003). banker-customer interactions in financial services. journal of marketing management, 19, 1001–1020. johnson, d., & grayson, k. (2005). cognitive and affective trust in service relationships. journal of business research, 58, 500–507. jones, g. r., & george, j. m. (1998). the experience and evolution of trust: implications for cooperation and teamwork. academy of management review, 23, 531–546. kantsperger, r., & kunz, w. (2010). consumer trust in service companies: a multiple mediating analysis. managing service quality, 20, 4–25. 96 a. carlander et al. / financial services review 27 (2018) 83-98 kaplan, d. (2009). structural equation modeling: foundations and extensions. thousand oaks, ca: sage. king-casas, b., tomlin, d., anen, c., camerer, c. f., quartz, s. r., & montague, p. r. (2005). getting to know you: reputation and trust in a two-person economic exchange. science, 308, 78–83. krishnan, m. s., ramaswamy, v., meyer, c. m., & damien, p. (1999). customer satisfaction for financial services: the role of products, services, and information technology. management science, 45, 1194–1209. kpmg nunwood. (2016). banking the customer experience dividend (available at https://www.nunwood.com/ excellence-centre/publications/white-papers/). levy, s., & hino, h. (2016). emotional brand attachment: a factor in customer-bank relationships. international journal of bank marketing, 34, 136–150. lewicki, r. j., & bunker, b. b. (1995). trust in relationships: a model of development and decline. in b. b. bunker, j. z. rubin, & associates (eds.), conflict, cooperation and justice: essays inspired by the work of morton deutsch (pp. 133–173). san francisco, ca: jossey-bass. lusardi a., & mitchell, o. s. (2007). financial literacy and retirement preparedness: evidence and implications for financial education. business economics, 42, 35–44. mackenzie, s. b., & podsakoff, p. m. (2012). common method bias in marketing: causes, mechanisms, and procedural remedies. journal of retailing, 88, 542–555. malhotra, d., & lumineau, f. (2011). trust and collaboration in the aftermath of conflict: the effects of contract structure. journal of the academy of management, 54, 981–998. mather, m., & johnson, m. k. (2000). choice-supportive source monitoring: do our decisions seem better to us as we age. psychology and aging, 15, 596–606. moe, w. w., netzer, o., & schweidel, d. a. (2017). social media analytics. in b. wierenga & r. van der lans (eds.), handbooks of marketing decision models (pp. 483–504). new york, ny: springer publishing. moin, s. m. a., devlin, j., & mckechnie, s. (2015). trust in financial services: impact of institutional trust on trusting belief. journal of financial services marketing, 20, 91–106. plitt, m., savjani, r. r., & eagleman, d. m. (2015). are corporations people too? the neural correlates of moral judgments about companies and individuals. social neuroscience, 10, 1–13. ponder, n., bugg holloway, b., & hansen, j. d. (2016). the mediating effect of customers’ intimacy perceptions on the trust-commitment relationship. journal of services marketing, 30, 75–87. rawlins, b. (2009). give the emperor a mirror: toward developing a stakeholder measurement of organizational transparency. journal of public relations research, 21, 71–99. rosenzweig, p. (2014). the halo effect: how managers let themselves be deceived. new york, ny: simon & schuster. sapienza, p., & zingales, l. (2012). a trust crisis. international review of finance, 12, 123–131. sekhon, h., ennew, c., kharouf, h., & devlin, j. (2014). trustworthiness and trust: influences and implications. journal of marketing management, 30, 409–430. selnes, f. (1998). antecedents and consequences of trust and satisfaction in buyer-seller relationships. european journal of marketing, 32, 305–322. shainesh, g. (2012). effects of trustworthiness and trust on loyalty and intentions: validating a parsimonious model in banking. international journal of bank marketing, 30, 267–279. (available at http://dx.doi.org/ 10.1108/02652321211236905). shin, h. s. (2009). reflections on northern rock: the bank run that heralded the global financial crisis. journal of economic perspectives, 23, 101–119. (available at http://dx.doi.org/10.1257/jep.23.1.101). siegrist, m., & cvetkovich, g. (2000). perceptions of hazards: the role of social trust and knowledge. risk analysis, 20, 713–720. sirdeshmukh, d., singh, j., & sabol, b. (2002). consumer trust, value, and loyalty in relational exchanges. journal of marketing, 66, 15–37. sunikka, a., peura-kapanen, l., & raijas, a. (2010). empirical investigation into the multi-faceted trust in the wealth management context. international journal of bank marketing, 28, 65–81. taylor, s. a., & baker, t. l. (1994). an assessment of the relationship between service quality and customer satisfaction in the formation of consumers’ purchase intentions. journal of retailing, 70, 163–178. 97a. carlander et al. / financial services review 27 (2018) 83-98 twenge, j. m., campbell, w. k., & carter, n. t. (2014). declines in trust in others and confidence in institutions among american adults and late adolescents, 1972–2012. psychological science. (advance online publication). van esterik-plasmeijer, p. w. j., & van raaij, f. (2017). banking system trust, bank trust, and bank loyalty. international journal of bank marketing, 35, 97–111. vargo, s. l., & lusch, r. f. (2004). evolving to a new dominant logic for marketing, journal of marketing, 68, 1–17. yim, c. k., chan, k. w., & lam, s. s. k. (2012). do customers and employees enjoy service participation? synergistic effects of selfand other-efficacy. journal of marketing, 76, 121–140. 98 a. carlander et al. / financial services review 27 (2018) 83-98 improving long-term portfolio risk and return by using appreciated stocks for charitable donations jeff whitwortha,* acollege of business, university of houston-clear lake, 2700 bay area boulevard, houston, tx 77058, usa abstract stock investors who are charitable donors can minimize capital gains taxes and improve portfolio diversification by donating their most appreciated shares instead of cash, and then investing the freed-up cash in the portfolio’s least-weighted stocks. the charity is indifferent to the donation method, and the investor receives the same charitable deduction. monte carlo simulations show that a donor-investor using this method enjoys substantially higher wealth and lower portfolio risk, particularly over longer time horizons. this strategy can be integrated with tax loss selling and some limited gain harvesting to further increase after-tax returns and reduce risk. © 2018 academy of financial services. all rights reserved. jel classifications: g11; h24 keywords: stock donation; capital gains; tax-efficient investing; diversification return; portfolio risk 1. introduction it has long been recognized that investors can minimize their tax liabilities and improve after-tax returns by strategically holding and selling different stocks in their portfolios according to the gains and losses each has accrued. constantinides (1983) shows that investors can experience substantial wealth gains by selling stocks that have declined to capture the tax benefit of the capital loss deduction, and argues that investors should avoid selling winning stocks for as long as possible to delay the payment of capital gains taxes. * corresponding author. tel.: �1-281-283-3218; fax: �1-281-283-3951. e-mail address: whitworthj@uhcl.edu financial services review 27 (2018) 257-277 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. while the tax advantage of selling investments that have dropped is obvious, there has been more debate on what to do with those that have gone up. in a follow-up to his original study, constantinides (1984) concludes that when the tax rate on long-term capital gains is lower than that on ordinary income (against which losses may be deducted), investors are better off if, in addition to realizing losses every year, they also realize gains in alternating years, thereby resetting the basis on the entire portfolio and planting the seeds for future capital loss realizations. however, dammon, dunn, and spatt (1989) argue that the value of the “tax option” this creates is highly dependent on the particular pattern of realized stock returns and other factors, and is less than constantinides suggests. stein, vadlamudi, and bouchey (2008) note that the effectiveness of gain harvesting depends on a number of variables, including the difference between long-term and short-term tax rates, the size of the accrued gain, the investor’s time horizon, and whether a step-up in basis is expected. in a study considering both the expected return and risk of several tax-efficient investing strategies, smith and smith (2008) note that deferring capital gains as long as possible has the undesirable effect of inhibiting portfolio rebalancing. because the stocks with accrued gains are not sold, they come to dominate a larger portion of the portfolio over time, and the investor’s risk increases because of the loss of diversification. an alternative strategy they propose is to realize all losses, but also harvest enough gains to offset any losses in excess of the $3,000 that may be deducted against ordinary income in a given year. the funds generated from stock sales are then used to purchase more shares of the stocks that make up a smaller part of the portfolio. in a long-term simulation, this produces higher after-tax returns and lower risk than any of the strategies studied by constantinides (1983, 1984). a more recent paper by whitworth and mccormack (2011) suggests that some investors can do better still. according to the statistics of income division of the internal revenue service, 37 million individuals reported a total of $201 billion in itemized charitable deductions for tax year 2015, and actual donations are almost certainly higher because many income tax filers claim the standard deduction instead of itemizing. for stock investors who are regular charitable donors, the tax code provides a way—in addition to the usual charitable deduction—to pay even less in taxes while rebalancing their portfolios, simultaneously increasing after-tax return and reducing risk. although no one knows beforehand whether a purchased stock will go up or down, an investor can dispose of it optimally after the fact to either minimize the associated tax liability or receive the greatest tax benefit. if the stock drops in value, tax loss selling still makes sense. however, if the stock rises, the investor can use it in lieu of cash to make the originally planned donation. the charity is indifferent to the method of donation, and the donor receives the same charitable deduction at full market value, so long as the stock has been held for more than a year. moreover, because the investor never actually sells the appreciated stock, its accrued gain is never realized for tax purposes, so no capital gains taxes are ever paid on those shares.1 the cash originally set aside for the donation (presumably equal in value to the appreciated stock) is now freed up to be invested in the portfolio, and the tax basis on any newly purchased shares will be the current market value. effectively, then, the investor receives a “step-up in basis,” permanently eliminating (rather than merely deferring) the accrued capital gains tax liability on the gifted shares. in choosing which shares to donate, it is generally best to begin with those having the largest percentage gains, thereby freeing the investor from as much capital gains tax liability 258 j. whitworth / financial services review 27 (2018) 257-277 as possible. this maximizes after-tax returns, but it also has another crucial benefit. by disposing of these shares and then investing any freed-up cash in the stocks which have fallen in value (or have not grown as much), funds are reallocated away from the most heavily weighted stocks in the portfolio and toward lesser-weighted ones. this prevents the portfolio from becoming too concentrated, thereby improving diversification and reducing overall risk. several studies (e.g., boyle et al., 2004; feld, 1999; welch, 2002) note the increased risk associated with highly concentrated stock positions, and welch (2002) specifically suggests charitable donation as an avenue for lessening one’s exposure. it should be noted that the stock donation strategy—like tax loss selling and gain harvesting—applies only to stocks in regular taxable accounts, because capital gains and losses are not recognized on transactions within tax-deferred or tax-exempt iras and 401(k)s. while individuals generally should take advantage of these accounts, many stocks are still optimally held in taxable accounts because (1) the amount that can be contributed to retirement accounts is limited, and (2) when one also owns tax-inefficient investments such as taxable bonds and/or actively managed mutual funds, these generally should be prioritized (i.e., ahead of stocks) for location in tax-advantaged accounts (horan, 2005; reichenstein, 2007; shoven and sialm, 2003). it is also important to remember that this strategy works only for those who already intend to make charitable contributions. it makes no sense to become a donor for the sole purpose of avoiding capital gains taxes. however, for those who view their gift as a sunk cost, the additional tax benefit from using appreciated shares in lieu of cash can be substantial, as previous studies have noted. in a detailed examination of this strategy, whitworth and mccormack (2011) demonstrate that in a one-period framework, the expected after-tax return can actually exceed the expected pretax return. this counterintuitive result derives from the fact that the donor-investor receives a tax deduction if the stock goes down but effectively enjoys a tax-free capital gain if it rises. other studies (aperio group, 2017; bakija and heim, 2011; reichenstein, 2007) have acknowledged the additional tax savings and increased return from donating appreciated stock, and it appears that investors are aware of the benefits as well. ackerman and auten (2011) report that in 2005, about 10% of all charitable deductions (26% for the top 1% of income earners) were for donations of stock, which typically had very low cost bases relative to their market values. welch (2002) also notes that private foundations and community-based donor-advised funds have become increasingly popular vehicles for philanthropically minded investors to dispose of low-basis stock. although most financial planners and tax professionals understand that donating appreciated stock can be advantageous, there have so far been no published studies quantifying the long-term benefits of doing so. using a monte carlo simulation approach, this paper demonstrates that better integration of portfolio management, charitable giving, and tax trading substantially increases returns and reduces risk. section 2 describes the simulation procedure. section 3 shows how terminal wealth, standard deviation (sd), diversification, and shortfall risk are affected by stock donation in conjunction with other tax-efficient strategies. section 4 discusses recent changes in the tax law that may be relevant to investors, and how the strategies in this paper could be affected by future changes. section 5 discusses the implications of our findings for financial planning. 259j. whitworth / financial services review 27 (2018) 257-277 2. simulation methodology we assume that an individual initially has annual wages of $100,000 that will grow at 3% per year. ordinary income is taxed at 24%, while long-term capital gains are taxed at 15%. the individual begins with a $250,000 portfolio invested equally across 25 non-dividendpaying stocks2 in a regular taxable account. in each year t, a simulated return for each stock i is generated as rit � rmt � �it �i � 1, 2, . . . , 25�, where the year t market return rmt is drawn from a normal distribution with mean 8% and sd 20%, and each idiosyncratic stock return component �it (i � 1, 2, … , 25) is drawn independently from a normal distribution with mean zero and sd 34.641%. the market and idiosyncratic volatilities �m � 0.2 and �� � 0.34641 are chosen so that the total sd of each stock i is �i � ��m 2 � �� 2 � �0.22 � 0.346412 � 0.4 � 40% and the contemporaneous correlation between returns of any two stocks i and j in the same year is �ij � �ij �i�j � �m 2 �i�j � 0.22 0.4 � 0.4 � 0.25. these are the same values for �i and �ij used in the simulation of smith and smith (2008). in general, investors attempting to profit from the asymmetric tax treatment of gains and losses should choose stocks with relatively high volatility and low correlation.3 consistent with the extensive literature on market efficiency (fama, 1970, 1991; malkiel, 2003), there is no serial correlation between the simulated returns across different years. therefore, the investor does not attempt to time the market based on a momentum or reversal strategy, but makes trading decisions each year based solely on tax efficiency and portfolio diversification. we consider three possible rules for deciding when to sell stocks: 1. buy and hold: no stocks are ever sold, regardless of gains or losses incurred. 2. realize losses only: each lot of stock that has declined below its original cost basis is sold so that up to $3,000 of capital losses can be deducted against ordinary income each year. any losses in excess of $3,000 are carried forward to future years. 3. realize losses and “rebalancing gains”: as explained in smith and smith (2008), all losses are realized, but the investor also realizes enough gains to offset any losses in excess of $3,000. to keep the portfolio as balanced as possible, shares of the stock comprising the largest portion of the portfolio are sold first until its weight equals that of the second largest stock, then shares of the first two stocks are sold until their respective weights equal that of the third, and so on until (a) all excess losses have been offset, or (b) there are no more gains in the portfolio to realize.4 at the end of each year, regardless of how the market has performed, the investor donates a fixed percentage (either 0%, 5%, or 10%) of his or her wage income to charity. 260 j. whitworth / financial services review 27 (2018) 257-277 to the maximum extent possible, the investor contributes appreciated shares in lieu of cash, and then replenishes the portfolio with any cash that otherwise would have been donated. this substitution does not change the portfolio’s value, but it does erase the accrued capital gains tax liability on the donated shares, effectively raising the basis on that part of the portfolio to current market value. to obtain the maximum tax benefit, the investor uses a “lowest in, first out” technique— donating shares from the lot with the lowest basis-to-value ratio (i.e., largest percentage gain), then the second lowest ratio, and so on until the entire planned donation has been made. if at any point there are no more stocks in the portfolio with accrued gains, then the remainder of the donation is made in cash as originally planned. along with the cash that had been earmarked but not used for donation, the individual immediately uses any proceeds from selling stocks (including the tax savings from harvesting capital losses) to purchase new shares. for optimal diversification, funds are invested first in the stock with the smallest portfolio weight until its weight equals that of the second smallest stock, then in the smallest two stocks until their respective weights equal that of the third smallest, and so on until all cash has been invested.5 this process is repeated for each year in the investment time horizon (that is assumed to be either 10, 20, 30, 40, or 50 years), after which the portfolio is liquidated and long-term capital gains taxes are finally paid on the difference between its total market value and cost basis. for each possible selling criterion, charitable contribution level, and time horizon, one million simulations are run as described above. the next section discusses the resulting statistics on terminal wealth, portfolio diversification, and risk for each scenario. 3. simulation results 3.1. mean and median effect on terminal wealth table 1 reports the mean and median ratios of an investor’s after-tax terminal wealth from alternative tax-efficient strategies versus a pure buy-and-hold strategy. results are shown for annual charitable contribution levels of up to 10%, and for time horizons up to 50 years. regardless of whether the investor realizes capital gains and/or losses, donating appreciated stock instead of cash significantly increases terminal wealth, and the more one regularly donates, the greater the benefit of doing so with stock. for example, as shown in panel a, an investor who does not harvest capital losses but does donate 10% of his annual wage income can realize a mean (median) wealth increase of 33% (26%) over a 30-year time frame compared with a pure buy-and-hold strategy. the benefits are even greater when using this strategy for longer. over 40 years, the same investor ends up on average with 62% more wealth versus buy-and-hold, and over 50 years, the mean increase is 107%. although larger charitable donors certainly have more of an opportunity to reap the benefits of this strategy, significant gains are still available to smaller donors. in this simulation, a 5% donor realizes about two-thirds of the wealth gain enjoyed by a 10% donor. for example, a 5% donor who contributes stock in lieu of cash enjoys a mean wealth increase 261j. whitworth / financial services review 27 (2018) 257-277 of 21% after 30 years, compared with a 33% increase for a 10% donor. this non-linearity in the relationship between donation level and terminal wealth exists because for higher annual giving amounts, there is a greater chance (particularly in down market years) of not having enough appreciated stocks for the full intended donation, and thereby being unable to take full advantage of the accompanying tax benefits. it is certainly possible for a donor-investor to do even better by using tax-efficient selling rules. in all but the strongest bull markets, a portfolio is likely to have at least a few losing investments. as is clear from panel b of table 1, simply realizing those capital losses and investing the tax savings has a long-term positive impact on the portfolio. as a comparison across columns shows, the investor who harvests capital losses still derives about the same incremental wealth increase from donating appreciated stock. both can be done simultaneously, since the accrued gains on the donated shares are never actually realized for tax purposes and therefore do not affect the investor’s capital loss deduction. as shown in panel c, a strategy of harvesting enough gains each year to offset any capital losses in excess of $3,000 improves portfolio performance even more. at first, it seems counterintuitive that realizing capital gains early could help. however, these gains do not incur a tax liability, as they are used only to offset losses that cannot be deducted against ordinary income in that year and would otherwise have to be carried forward to future years. while this does eliminate the possibility of deducting those losses in future years, it is quite likely that at least a few stocks in the portfolio will decline next year. for such declines to translate into deductible losses, those stocks cannot be trading too far above their cost bases—otherwise, the decline will simply turn a larger accrued gain into a smaller one. selling some stocks with gains resets the basis on that part of the portfolio, “planting seeds” for potential future loss deductions. table 1 mean (median) ratio of terminal wealth from tax-efficient investment strategies vs. buy-and-hold years invested percentage of income donated via appreciated stock 0% 5% 10% panel a: realize no losses or gains (buy and hold) 10 1.00 (1.00) 1.03 (1.03) 1.05 (1.05) 20 1.00 (1.00) 1.09 (1.08) 1.15 (1.13) 30 1.00 (1.00) 1.21 (1.16) 1.33 (1.26) 40 1.00 (1.00) 1.42 (1.27) 1.62 (1.43) 50 1.00 (1.00) 1.73 (1.41) 2.07 (1.67) panel b: realize all losses each year 10 1.01 (1.01) 1.04 (1.04) 1.06 (1.06) 20 1.04 (1.02) 1.12 (1.10) 1.18 (1.16) 30 1.08 (1.03) 1.27 (1.19) 1.39 (1.30) 40 1.13 (1.04) 1.51 (1.32) 1.71 (1.49) 50 1.22 (1.06) 1.87 (1.49) 2.20 (1.74) panel c: realize all losses and offsetting gains 10 1.03 (1.02) 1.05 (1.04) 1.06 (1.06) 20 1.11 (1.08) 1.16 (1.13) 1.19 (1.17) 30 1.26 (1.17) 1.36 (1.27) 1.42 (1.33) 40 1.48 (1.27) 1.69 (1.47) 1.79 (1.57) 50 1.78 (1.39) 2.18 (1.74) 2.36 (1.89) 262 j. whitworth / financial services review 27 (2018) 257-277 an investor who realizes offsetting gains can also benefit from donating appreciated stock, and it is better to use both strategies than merely one or the other. however, the additional benefit of donating appreciated shares is smaller when one already realizes offsetting gains, and vice versa. for example, consider an individual with a 30-year horizon who realizes only capital losses and donates 5% of her income to charity in cash (i.e., 0% in stock). as shown in panels b and c, she can increase her mean terminal wealth ratio from 1.08 to 1.26 by also realizing offsetting gains, or to 1.27 by donating appreciated stock instead of cash—an improvement of 18 or 19 percentage points relative to the buy-and-hold benchmark. however, if she further decides to implement both strategies simultaneously, her mean ratio increases to 1.36—still an improvement, but only an additional 9 or 10 percentage points. this happens because both strategies use the stocks that have increased in value most and now comprise a larger percentage of the portfolio. in some cases (particularly in bear markets), there may not be enough winning stocks to fully execute both strategies, which explains the result noted above. however, in many years the investor will be able to do both, and certainly should whenever possible. fig. 1 summarizes the results presented in the above table for a 50-year investor, clearly illustrating the long-term wealth gains available to those who donate appreciated stock. regardless of the selling rule(s) used (i.e., whether the investor harvests losses and/or gains), a 5% donor’s terminal wealth is much greater than that of a non-donor, and a 10% donor does better still (although the additional improvement from 5% to 10% is somewhat smaller). fig. 1 also illustrates that for each charitable donation level (0%, 5%, and 10%), one is clearly better off recognizing capital losses, and better still recognizing offsetting gains as well. the benefit of harvesting losses is similar regardless of how much one donates. the incremental benefit of harvesting offsetting gains is positive for all donation levels, but as previously discussed, it is somewhat smaller for stock donors than for non-donors. 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 0% 5% 10% ra � o of t er m in al w ea lt h vs . b uy -a nd -h ol d percentage of income donated via appreciated stock fig. 1. mean ratio of terminal wealth versus buy-and-hold for alternative investment strategies (50-year horizon). 263j. whitworth / financial services review 27 (2018) 257-277 3.2. effect on standard deviation of terminal wealth in addition to increasing after-tax returns, the stock donation strategy advocated here has a very desirable side effect: it reduces risk through improved diversification. if left alone over time, a handful of stocks will eventually grow to dominate a disproportionate share of a portfolio. this follows from bessembinder’s (2018) observation that long-horizon compounded returns are skewed even if single-period returns are not. while the investor is certainly better off because of the gains of these past winners, the portfolio then becomes over-dependent on their future performance, which cannot be guaranteed. regularly donating the most appreciated shares helps reduce portfolio imbalances while never actually realizing gains for tax purposes. the cash which otherwise would have been donated is then invested in the least-weighted stocks to further facilitate rebalancing. as shown in table 2, the stock donation strategy does reduce risk versus a buy-and-hold benchmark, and the more charitable one is, the greater the benefit. for example, the sd (in dollars) of terminal wealth after 40 years for an investor who realizes losses and donates 10% of his annual income via appreciated stock is only 79% of what it would be for a buy-andhold investor who donates only cash. as previously noted, one does not have to be a large donor to benefit from this. if the same investor donated only 5% of his income, his sd would be 86% of its buy-and-hold level, thereby still realizing about two-thirds the risk reduction that a 10% donor could. as expected, the longer the investment period, the greater the benefit. over relatively short periods of 10 years or less, sd is not much different from what it would be using a buy-and-hold approach. however, the improvement becomes substantial over longer periods. table 2 ratio of standard deviation of terminal wealth from tax-efficient investment strategies vs. buy-andhold years invested percentage of income donated via appreciated stock 0% 5% 10% panel a: realize no losses or gains (buy and hold) 10 1.00 0.99 0.98 20 1.00 0.95 0.94 30 1.00 0.91 0.86 40 1.00 0.86 0.80 50 1.00 0.83 0.75 panel b: realize all losses each year 10 1.01 0.99 0.99 20 1.01 0.97 0.95 30 1.01 0.92 0.87 40 1.01 0.86 0.79 50 0.99 0.83 0.72 panel c: realize all losses and offsetting gains 10 0.98 0.98 0.98 20 0.95 0.94 0.93 30 0.91 0.85 0.82 40 0.87 0.80 0.77 50 0.80 0.68 0.59 264 j. whitworth / financial services review 27 (2018) 257-277 an additional clarification is in order here. the risk reduction relative to portfolio value is actually greater than the values in table 2 might suggest, because average terminal wealth is higher with the tax-efficient strategies. even if all ratios in the table were equal to 1.00 (indicating no change in the dollar amount of the sd), that would still represent a reduction in risk as a percentage of total wealth. the fact that the ratios are actually less than 1.00 for stock donors (particularly over longer periods) is indicative of an even greater improvement in the risk-to-reward ratio. a quick comparison of panels a and b shows that realizing losses in itself, while still a good idea, does not substantially reduce risk, as tax-loss selling contributes to only limited portfolio rebalancing and does not address the handful of big winners that tend to be most responsible for portfolio imbalances. however, panel c shows that gain harvesting has a significant risk-reduction effect. an investor who normally harvests only losses and donates 5% of her income in cash can reduce her 40-year sd to 86% of its buy-and-hold level by donating stock instead, or to 87% by also harvesting “rebalancing gains.” of course, there is no reason she cannot do both. however, the additional benefit would be smaller, reducing sd to 80% of its buy-and-hold equivalent—an additional decrease of only 6 or 7 percentage points. this is because, as previously discussed, the two strategies overlap somewhat, both tending to dispose of the most appreciated shares. 3.3. portfolio concentration table 3 provides more direct evidence on the degree to which the tax-efficient strategies discussed in this paper prevent a portfolio from becoming too concentrated over time. to table 3 mean portfolio concentration index (sum of squared weights) from alternative investment strategies years invested percentage of income donated via appreciated stock 0% 5% 10% panel a: realize no losses or gains (buy and hold) 10 0.10 0.07 0.06 20 0.18 0.09 0.07 30 0.26 0.12 0.09 40 0.33 0.15 0.10 50 0.38 0.17 0.13 panel b: realize all losses each year 10 0.09 0.06 0.05 20 0.16 0.09 0.07 30 0.23 0.12 0.09 40 0.28 0.15 0.11 50 0.33 0.17 0.13 panel c: realize all losses and offsetting gains 10 0.05 0.05 0.04 20 0.08 0.06 0.05 30 0.12 0.07 0.06 40 0.17 0.09 0.07 50 0.22 0.11 0.09 265j. whitworth / financial services review 27 (2018) 257-277 measure concentration, we compute the sum of the squares of the individual weights of each stock in the portfolio at the end of the simulated investment horizon.6 in the ideal scenario of a perfectly diversified portfolio (i.e., with equal amounts invested in each of the 25 stocks), the index would be 1/25, or 0.04, whereas the most unbalanced portfolio possible (i.e., completely concentrated in just one stock) would have an index of 1.00. there certainly is a tendency for portfolios left to themselves to become more unbalanced over time. as shown in panel a, a pure buy-and-hold portfolio after 10 years will on average have a concentration index of 0.10, which is equivalent to all of the wealth being spread across just 10 of the 25 stocks. after 30 years the average index rises to 0.26, approximately equivalent to having all portfolio wealth concentrated in only four stocks. over longer time frames, this imbalance becomes worse still. fortunately, all of the tax-advantaged strategies presented improve diversification. as shown by comparing the first column of panels a and b, even just realizing losses helps some, because selling the losing stocks and then investing the proceeds equally across all of them makes at least that portion of the portfolio better balanced. far greater diversification improvements, however, are achievable either by donating appreciated stocks, harvesting some capital gains, or ideally both. as shown in panels a and b, even a 5% charitable donor can significantly slow the growth in portfolio concentration over time simply by giving appreciated stock; 10% donors can mitigate the imbalance to an even greater extent. as shown by the first column of panels b and c, the gain recognition strategy is effective in reducing portfolio concentration, even for non-donors. of course, better diversification is achievable by using all strategies simultaneously. in the best possible scenario simulated here, a 10% donor who harvests losses and offsetting gains will still have an almost perfectly balanced portfolio after 10 years of investing, with a concentration index effectively equal to the ideal value of 0.04; and although the portfolio does grow slightly more concentrated over time, he or she is still reasonably diversified after 50 years, with a concentration index of only 0.09. this compares quite favorably to a 50-year buy-and-hold investor, with the optimal strategy achieving (0.38 – 0.09)/(0.38 – 0.04) � 85% of the potential improvement in diversification. the extent to which these tax-efficient strategies keep portfolio imbalances in check is also apparent in fig. 2, which shows how much of the portfolio is invested in its most heavily weighted stock after progressively longer time horizons. as indicated by the gray dashed line, a pure buy-and-hold portfolio becomes unbalanced fairly quickly. after 10 years, one of the 25 stocks will come to make up approximately 20% of the portfolio. after 30 years, that weight grows to an average of 41%, and over longer periods it approaches and exceeds 50%. this leaves the portfolio very vulnerable to any large losses subsequently incurred by that one stock. fortunately, as the remaining lines on the graph show, the portfolio can be made much less dependent on the performance of any individual stock by harvesting some gains from the most heavily weighted stocks and/or donating some of the most appreciated shares. as expected, doing both simultaneously leads to the best possible outcome regardless of the investment horizon. as shown by the darker solid line, a 10% stock donor who realizes losses and offsetting gains will on average have just 15% invested in his or her most heavily 266 j. whitworth / financial services review 27 (2018) 257-277 weighted stock even after 50 years, which is considerably less concentrated than either of the non-donor outcomes shown.7 3.4. diversification’s contribution to portfolio return the fact that diversification reduces risk without reducing average returns is often referred to as the “only free lunch in investing.” however, booth and fama (1992) show that diversification actually improves a portfolio’s long-term compounded return—an effect that willenbrock (2011) suggests might be a “free dessert” (see also bouchey et al., 2012; erb and harvey, 2006). therefore, the increased terminal wealth from various tax-efficient strategies is attributable not only to the tax savings themselves, but also to the “diversification return” just noted. to determine how much of the wealth gain is due to diversification, we track within each iteration of the simulation the value of a portfolio which enjoys none of the incremental tax benefits of the tax-advantaged portfolio, yet is equally diversified. this non-tax-advantaged portfolio also begins with $250,000 spread equally across the same 25 stocks. like a buy-and-hold portfolio, it benefits neither from tax loss selling nor from the effective “step-up in basis” on shares donated to charity. however, each year, its total value is costlessly redistributed across the 25 stocks so that the resulting individual weights are identical to those in the tax-efficient portfolio. at the end of the investment horizon, the non-tax-advantaged portfolio is also liquidated, and taxes are paid on any gains over the original $250,000 cost basis. the difference in terminal values between this non-taxadvantaged portfolio and the more concentrated buy-and-hold portfolio is due to the “diversification return,” while the difference in value between the tax-efficient portfolio and the equally diversified non-tax-advantaged portfolio is directly attributable to tax benefits. 0.0 0.1 0.2 0.3 0.4 0.5 0 10 20 30 40 50 a ve ra ge w ei gh t of l ar ge st s to ck in p or � ol io number of years invested fig. 2. growth in weight of largest stock for alternative investment strategies. 267j. whitworth / financial services review 27 (2018) 257-277 table 4 reports how much of the increase in terminal value for each strategy (relative to a buy-and-hold portfolio) is due to tax savings and diversification, respectively. the overall results indicate that most of the additional long-term value is in fact attributable to the improved diversification these strategies create. however, there is still significant value generated by the tax savings themselves. as seen previously in table 1, donating appreciated stock improves mean and median terminal wealth, regardless of whether one realizes capital losses and/or gains. a comparison across the columns of table 4 shows that this wealth increase is due to tax savings and a diversification return, and that both effects are greater over longer horizons. this is expected since donating more stock causes more of the portfolio to be stepped-up in basis, while also reallocating more wealth from the most-weighted to the least-weighted table 4 mean (median) increase in terminal wealth vs. buy-and-hold due to tax and diversification effects years invested effect percentage of income donated via appreciated stock 0% 5% 10% panel a: realize no losses or gains (buy and hold) 10 tax savings 0.00 (0.00) 0.02 (0.02) 0.04 (0.03) diversification 0.00 (0.00) 0.01 (0.01) 0.01 (0.02) 20 tax savings 0.00 (0.00) 0.03 (0.03) 0.07 (0.05) diversification 0.00 (0.00) 0.06 (0.05) 0.08 (0.08) 30 tax savings 0.00 (0.00) 0.05 (0.04) 0.09 (0.08) diversification 0.00 (0.00) 0.16 (0.12) 0.24 (0.18) 40 tax savings 0.00 (0.00) 0.08 (0.05) 0.13 (0.09) diversification 0.00 (0.00) 0.34 (0.22) 0.49 (0.34) 50 tax savings 0.00 (0.00) 0.10 (0.06) 0.18 (0.12) diversification 0.00 (0.00) 0.63 (0.35) 0.89 (0.55) panel b: realize all losses each year 10 tax savings 0.01 (0.01) 0.03 (0.03) 0.05 (0.04) diversification 0.00 (0.00) 0.01 (0.01) 0.01 (0.02) 20 tax savings 0.02 (0.02) 0.06 (0.05) 0.09 (0.08) diversification 0.02 (0.00) 0.06 (0.05) 0.09 (0.08) 30 tax savings 0.04 (0.03) 0.10 (0.07) 0.15 (0.12) diversification 0.04 (0.00) 0.17 (0.12) 0.24 (0.18) 40 tax savings 0.04 (0.03) 0.15 (0.10) 0.21 (0.16) diversification 0.09 (0.01) 0.36 (0.22) 0.50 (0.33) 50 tax savings 0.07 (0.04) 0.21 (0.13) 0.30 (0.20) diversification 0.15 (0.02) 0.66 (0.36) 0.90 (0.54) panel c: realize all losses and offsetting gains 10 tax savings 0.01 (0.01) 0.03 (0.02) 0.04 (0.04) diversification 0.02 (0.01) 0.02 (0.02) 0.02 (0.02) 20 tax savings 0.02 (0.02) 0.06 (0.04) 0.08 (0.08) diversification 0.09 (0.06) 0.10 (0.09) 0.11 (0.09) 30 tax savings 0.04 (0.03) 0.08 (0.07) 0.12 (0.10) diversification 0.22 (0.14) 0.28 (0.20) 0.30 (0.23) 40 tax savings 0.06 (0.04) 0.13 (0.10) 0.18 (0.15) diversification 0.42 (0.23) 0.56 (0.37) 0.61 (0.42) 50 tax savings 0.09 (0.05) 0.19 (0.13) 0.26 (0.19) diversification 0.69 (0.34) 0.99 (0.61) 1.10 (0.70) 268 j. whitworth / financial services review 27 (2018) 257-277 stocks, reducing portfolio concentration. as an example, consider an investor with a 50-year horizon who realizes capital losses only. table 1 shows that such an investor who donates 10% of his income via appreciated stock enjoys a mean terminal wealth ratio of 2.20, but this ratio for a non-donor is only 1.22. these values represent total wealth increases of 120% and 22%, respectively, versus a pure buy-and-hold investor. table 4 shows that of the non-donor’s 22% wealth increase, 7% is directly due to tax loss deductions, while 15% is due to slightly better diversification. likewise, of the stock donor’s 120% wealth increase, 30% is due to tax benefits, and the remaining 90% is due to much better diversification. table 4 also reveals the tax and diversification effects simply from realizing capital losses and/or gains. comparing panel b versus panel a shows that while tax-loss selling does create a modest diversification return for non-donors, the value of loss harvesting for stock donors is almost entirely limited to the tax benefit itself. conversely, comparing panel c versus panel b shows that smith and smith’s (2008) gain harvesting strategy improves portfolio value through diversification rather than tax savings. the incremental diversification return is larger for non-donors, because investors who donate appreciated stock have already achieved much of the potential diversification improvement. the incremental tax effect from harvesting gains is very small and in some cases slightly negative. this result can be understood by considering the tax arguments both for and against the gain recognition strategy. while offsetting gains against losses does effectively reset the cost basis on those shares and can create future opportunities to deduct losses, the strategy can backfire if no shares (or very few) decline the following year, leaving the investor with little or no capital loss deduction. in such a case, it would have been better to carry at least some of the excess loss forward and deduct it the following year. it is important to note, however, that the gain harvesting strategy is still sound, as the increase in the diversification return outweighs any decrease in the tax savings component. 3.5. shortfall risk one practical risk measure is the probability of failing to earn a specified minimum return. for each of the strategy combinations simulated, table 5 reports the probability of earning a negative return and thereby ending up with less than the original investment at the end of the horizon. given the simulation parameters, a pure buy-and-hold portfolio has a 20% chance of losing money over 10 years. as expected, the risk of loss declines when investing over longer periods; however, it remains as high as 10% even after 50 years. this is understandable because, as explained previously, the portfolio becomes quite concentrated if it is never rebalanced. as shown in panel b, shortfall risk is slightly reduced when the investor merely realizes capital losses, partly because of the limited rebalancing that occurs and partly because of the tax savings that somewhat mitigate the losses. however, greater reductions in shortfall risk are achievable by harvesting some gains and/or using stock for planned charitable contributions. both help to rebalance the portfolio significantly, and donating appreciated stock erases some of the portfolio’s accrued capital gains tax liability. using all of these strategies reduces shortfall risk, but the extent of the reduction depends on the investment time frame. over relatively short periods (i.e., 10 years), most of the shortfall 269j. whitworth / financial services review 27 (2018) 257-277 risk of a buy-and-hold strategy is attributable simply to the risk that the market (and by extension, most stocks in the portfolio) will perform poorly; it is not due to excessive unsystematic risk because the portfolio is still fairly diversified. therefore, the stock donation strategy and the gain recognition strategy reduce shortfall risk only slightly (from 20% to 17%). over longer periods of 30–50 years, however, overall market returns are unlikely to be negative, but a portfolio left to itself very well could become so concentrated in a few stocks as to leave the entire portfolio vulnerable to an adverse event at one of those companies. this is where the charitable donation and gain harvesting strategies become very helpful by limiting the buildup of idiosyncratic risk in the portfolio. as shown in table 5, the risk of losing money over 30 years declines from 12% for a buy-and-hold portfolio to 5% for an investor who uses all of the tax-efficient strategies presented. over 50 years, the benefits are even more striking, as the risk of loss can be reduced from 10% to only 1%. comparing panels b and c versus panel a, we see that a non-donor can reduce shortfall risk somewhat by realizing losses, and even more so by also realizing offsetting gains. however, someone who donates appreciated stock is already very close to the minimum achievable shortfall risk for the given time horizon; that is, a stock donor could not substantially reduce this probability any further through harvesting losses and/or gains. as pointed out earlier in this paper, an investor does not have to be a large charitable donor to benefit from using appreciated stock for whatever donations he or she does make, and this is especially true when it comes to reducing shortfall risk. a comparison across columns of table 5 shows that a 5% donor enjoys almost the same shortfall risk reduction that a 10% donor would. table 6 provides more specific information on what losses a particularly unlucky investor might experience, whether due to exceptionally poor market performance or to table 5 zero-return shortfall risk from alternative investment strategies years invested percentage of income donated via appreciated stock 0% 5% 10% panel a: realize no losses or gains (buy and hold) 10 0.20 0.18 0.17 20 0.14 0.11 0.09 30 0.12 0.06 0.05 40 0.11 0.04 0.03 50 0.10 0.02 0.01 panel b: realize all losses each year 10 0.19 0.18 0.17 20 0.13 0.10 0.08 30 0.10 0.06 0.04 40 0.08 0.03 0.02 50 0.07 0.02 0.01 panel c: realize all losses and offsetting gains 10 0.18 0.17 0.17 20 0.11 0.10 0.09 30 0.07 0.05 0.05 40 0.04 0.03 0.03 50 0.03 0.02 0.01 270 j. whitworth / financial services review 27 (2018) 257-277 large losses by the portfolio’s most heavily weighted stocks. for each strategy, the table reports after-tax terminal wealth relative to the original $250,000 investment given a fifthor first-percentile scenario.8 as seen in panel a, an unlucky buy-and-hold investor can experience substantial losses. over a 30-year horizon, such an individual has a 5% chance of ending up with less than $149,207 (about 60% of the starting investment) and a 1% chance of ending up with less than $79,499 (about 32% of the starting investment). these fifthand first-percentile values are actually worse for longer horizons, reflecting the fact that portfolio concentration and risk grow over time with a pure buy-and-hold approach. while losses are still possible under the charitable donation strategy, their likelihood and potential severity is reduced, especially over longer horizons. for example, a 10% stock donor who harvests no gains or losses over 30 years would have fifthand first-percentile terminal wealth values of $250,485 and $140,841 (approximately 100% and 56% of the original investment, as shown in panel a). therefore, a 10% donor does significantly better than the non-donor when both experience the same degree of bad luck. furthermore, because the stock donation strategy greatly limits the growth in portfolio concentration over time, the donor’s fifthand first-percentile outcomes are better over longer horizons, not worse. after 50 years, the 10% donor would still have an overall gain (ending up with 178% of starting wealth) even in a fifth-percentile scenario and a relatively small loss (finishing with 85% of the original investment) with a first-percentile outcome. a comparison of panels a and b shows that in every case, an unlucky investor who harvests capital losses performs better than one who does not. this finding is intuitive because the resulting tax savings help to mitigate whatever losses have occurred. a comparison of panels b and c finds mixed results on whether harvesting gains can further table 6 fifth (first) percentile terminal wealth values relative to original investment years invested percentage of income donated via appreciated stock 0% 5% 10% panel a: realize no losses or gains (buy and hold) 10 0.62 (0.42) 0.65 (0.45) 0.68 (0.47) 20 0.60 (0.36) 0.73 (0.46) 0.79 (0.49) 30 0.60 (0.32) 0.89 (0.51) 1.00 (0.56) 40 0.59 (0.29) 1.14 (0.60) 1.32 (0.69) 50 0.59 (0.27) 1.50 (0.73) 1.78 (0.85) panel b: realize all losses each year 10 0.63 (0.43) 0.66 (0.46) 0.69 (0.48) 20 0.63 (0.38) 0.76 (0.48) 0.83 (0.53) 30 0.67 (0.36) 0.94 (0.55) 1.07 (0.63) 40 0.72 (0.35) 1.22 (0.66) 1.44 (0.79) 50 0.77 (0.34) 1.61 (0.83) 1.96 (1.02) panel c: realize all losses and offsetting gains 10 0.65 (0.45) 0.67 (0.46) 0.68 (0.48) 20 0.73 (0.44) 0.77 (0.47) 0.80 (0.50) 30 0.88 (0.47) 0.97 (0.54) 1.03 (0.58) 40 1.08 (0.53) 1.27 (0.64) 1.37 (0.72) 50 1.34 (0.60) 1.72 (0.80) 1.88 (0.91) 271j. whitworth / financial services review 27 (2018) 257-277 improve the outcomes of unlucky investors. the strategy is helpful in this respect for non-donors, as it helps maintain better balance in a portfolio that would otherwise become very concentrated and more risky over time. however, a 10% stock donor’s fifthand first-percentile wealth values are both made slightly worse by gain harvesting, even though the average investor’s results (seen previously in table 1) are made better. to understand this somewhat counterintuitive finding, note that a portfolio that has performed poorly will by definition have relatively few appreciated shares, and if most of these shares are donated to charity, the resulting portfolio will be reasonably diversified already, so that gain harvesting would not improve diversification much more. in addition, offsetting current-year losses in this manner can occasionally prevent the investor from being able to deduct capital losses in the following year(s), and the fifthand first-percentile values in panel c likely reflect scenarios where the strategy backfires in this manner. 4. impact of recent and potential future tax law changes in 2017, congress passed the tax cuts and jobs act (tcja), amending certain provisions of the tax code that affect investors and charitable donors, and without question, it will enact other tax reforms in the future. therefore, it makes sense to consider the effects of recent and potential future tax policy changes on the strategies presented in this paper. one key tcja provision approximately doubled the standard deduction while limiting or disallowing many non-charitable itemized deductions. this will result in more taxpayers taking the standard deduction and fewer itemizing. in addition, because the tcja generally reduced marginal tax rates, any itemized deductions that are taken will result in smaller tax savings. because of these factors, charitable donations are likely to decline somewhat, so that fewer individuals can enjoy the tax and diversification benefits previously noted. nevertheless, many people will continue to donate to charity. for example, higher-income taxpayers still have relatively high marginal rates and may already have enough deductible home mortgage interest and property tax payments to equal or exceed the standard deduction. these individuals still benefit from the itemized charitable deduction. many others will continue to donate for philanthropic reasons irrespective of tax incentives. burman and randolph (1994) find that the level of charitable donations is only weakly related to the associated tax savings. consistent with this result, greene and mcclelland (2001) report that about 60% of individual contributions are to religious organizations, and bradley, holden, and mcclelland (2000) find that these are much less sensitive to tax rates than contributions to other organizations. it is important to clarify that the incremental benefits of the stock donation strategy do not depend on the ordinary income tax rate, or even whether the charitable deduction can be taken at all. while these factors may influence whether and how much people donate, they do not impact the capital gains tax savings and improved portfolio diversification from making any intended contributions with stock instead of cash. however, the strategy’s value does depend on the long-term capital gains tax rate. while the tcja did not materially affect the capital gains rate, it has been modified before and likely will be again at some point in the future. if new legislation were to completely eliminate preferentially low rates on 272 j. whitworth / financial services review 27 (2018) 257-277 long-term gains (as the 1986 tax reform act did), it would negatively affect investors’ terminal wealth, but it would also increase the value of the charitable donation strategy in mitigating the impact of capital gains taxes. likewise, if capital gains taxes were reduced further, it would positively affect terminal wealth, but the tax savings from donating appreciated stock would be smaller. in 2018, some policymakers discussed the possibility of allowing investors to compute capital gains relative to higher, inflation-adjusted cost bases. such a policy would have effects similar to lowering the statutory rate, especially for shares that had been held longer. one provision that was considered but not included in the final tcja bill would have required investors who own multiple lots of a stock to use a “first-in, first-out” (fifo) method when disposing of shares, thereby eliminating the ability to specifically identify which lot(s) they wish to sell or donate. requiring investors to sell older (usually lowerbasis) lots first would generally increase their taxes, but charitable donors would be especially well-positioned to mitigate the negative effects on their portfolios by first donating shares from the older, low-basis lots, so that any subsequent stock sales could be from higher-basis lots. 5. conclusion and implications for financial planning the results of this paper strongly suggest that a stock investor who regularly contributes to charity should donate appreciated shares from his portfolio that are equal in value to whatever cash donation he was willing to make, and then use the freed-up cash to purchase more of the least-weighted stocks in the portfolio. the donated shares must have been held for more than a year and should be the investor’s most appreciated stocks. there is no disadvantage to the charity or to the donor-investor, but there are two significant advantages. first, this increases after-tax returns by allowing the investor to escape capital gains taxes on the biggest winners in the portfolio. second, regularly donating the most appreciated stocks and reallocating funds to lesser-weighted stocks counteracts the portfolio’s tendency to become increasingly concentrated over time. this improvement in diversification not only reduces risk—as evidenced by a lower standard deviation of terminal wealth and a reduced probability of losing money over the investment horizon—but also creates a substantial “diversification return.” the combined wealth increase from the tax savings and improved diversification is not trivial. depending on the amount donated and the time horizon, the investor can realistically end up with more than double the portfolio value of someone who holds the same stocks but donates only in cash. financial planners certainly should ensure that their more philanthropic clients understand why donating appreciated stock is better than donating cash. however, charitable organizations may also wish to suggest this method to their contributors. because some donors likely make decisions based on a gift’s after-tax cost, it is possible that they might choose to donate more when made aware of the additional tax savings. investors who donate appreciated stock can and should still use other known tax-efficient strategies. for example, harvesting losses is still beneficial, because the accrued gain on a donated stock is never realized for tax purposes and therefore does not reduce the investor’s 273j. whitworth / financial services review 27 (2018) 257-277 capital loss deduction. in addition, investors should still, if possible, recognize enough gains from the most heavily weighted stocks in the portfolio to offset any losses in excess of the $3,000 that can be deducted against ordinary income in a year. this helps further rebalance the portfolio, reducing risk and increasing its long-term realized return. at times, real-world constraints may limit investors’ ability to fully implement the above recommendations. for example, portfolio managers who are concerned with tracking error relative to a value-weighted index may not be able to sell or donate as much of a stock as would be optimal, if that stock is weighted heavily in the index. nevertheless, the risk and return of such portfolios can still be improved by implementing tax-efficient strategies to the extent permitted. of course, some real-world investors may be able to exploit the asymmetric tax treatment of gains and losses even more effectively—for example, by purchasing securities whose returns are negatively correlated with each other, effectively ensuring that he or she will be able to make donations from whichever shares rise in value while simultaneously deducting losses on the others. it is important to note that the conclusions in this paper are derived from a monte carlo simulation based on certain assumptions that, while reasonable, do not perfectly represent every investor’s situation. for example, to be consistent with smith and smith (2008), we assume that no additional savings are contributed to the portfolio beyond the initial $250,000. however, many investors do save regularly using a dollar-cost averaging approach, which in itself can promote portfolio diversification, perhaps lessening the incremental impact of the charitable donation strategy. we also assume that annual giving is a function only of wage income, but some donors likely consider their wealth level also, giving more or less when their portfolios have performed exceptionally well or poorly. finally, we assume that the investor realizes and pays taxes on all gains at the end of the horizon. in reality, withdrawals from the portfolio are likely to occur at multiple points in time (e.g., throughout retirement), and some of the gains may never be taxed because of the step-up in basis at death. it will be useful for future studies to model different approaches a tax-savvy investor might take to saving, charitable giving, and portfolio withdrawals. it should also be considered that while the simulated stock returns used in this study follow a joint normal distribution with constant risk and return parameters, actual stock return distributions can change over time and exhibit some degree of non-normality. it is possible that a bootstrapping approach with historical resampling may shed additional light on how well tax-efficient strategies work given how stock prices have actually behaved. while the above issues are important and should be considered in future research, this paper presents strong simulation-based evidence that investors should donate appreciate stock in lieu of cash while continuing to use the tax-loss selling and gain harvesting strategies previously documented in the literature. doing so substantially increases average returns and reduces risk. it is normally difficult to accomplish both of these objectives simultaneously, as capital market theory suggests higher returns are only available for those willing to accept more risk. while most who attempt to time the market or predict which stocks will outperform others find at best limited success, investors can work within the tax code to consistently improve performance without trying to outguess the market. given the demonstrated benefits, one certainly should do so to the maximum extent possible. 274 j. whitworth / financial services review 27 (2018) 257-277 notes 1 private foundations pay a small excise tax of 1–2% on their net investment income, which includes accrued gains on shares received from donors. however, donoradvised funds and public charities are not subject to the tax. 2 although the tax-efficient strategies in this paper can be used with dividend-paying stocks, they are most effective when all of the investor’s return is in the form of capital gains. to the extent stocks in the portfolio pay dividends (converting capital gains into current income), the investor is less able to defer gains and/or eliminate accrued tax liabilities through stock donations. 3 empirical research suggests that these values are realistic, and that investors may be able to find stocks with even higher sds and lower correlations. statman (1987) reports a 49% average individual volatility for a sample of nyse/amex-listed stocks, and the findings of campbell et al. (2001) suggest similar or higher volatilities depending on the time period examined. chan, karceski, and lakonishok (1999) find an average correlation of 0.28 between any two random stocks, and campbell et al. (2001) report similar or lower correlations. 4 if the investor owns multiple lots of the same firm’s stock (purchased in different years), shares are sold first from the lot with the highest basis-to-value ratio (i.e., the lowest percentage gain). 5 this procedure results in slightly better diversification compared with smith and smith’s (2008) method of investing the funds equally across the 10 least-weighted stocks. in practice, either method would need to consider the wash sale rule, which disallows any loss deduction on the shares just sold if identical shares are immediately repurchased. one could effectively circumvent this restriction by purchasing shares in a different company with similar risk and return characteristics, or by placing the sale proceeds in a money market fund until the 30-day wash sale period expires (after which shares of the original firm could be bought again). 6 the same method is used to compute the herfindahl-hirschman index (hhi), an indicator of how concentrated market power is within an industry (with low values indicating strong competition, and high values indicating near-monopolistic conditions where very few firms control most of the industry). fraser and jennings (2010) use a modification of the hhi to compute a “degree of diversification” measure for endowment fund portfolios. 7 although omitted from the graph for readability, a base strategy of realizing only losses results in essentially the same maximum stock weights as realizing neither losses nor gains. therefore, tax loss selling by itself does not materially affect a portfolio’s dependence on its largest stock. 8 in cases where the portfolio’s final value is below its cost basis, after-tax value is computed assuming the net loss is immediately offset against long-term gains on other property. if the investor does not have enough gains, then the excess loss is carried forward to future years. in such cases, the effective after-tax value will be somewhat higher or lower depending on the type of income against which losses are eventually deducted and when the deductions occur. 275j. whitworth / financial services review 27 (2018) 257-277 references ackerman, d., & auten, g. (2011). tax expenditures for noncash charitable contributions. national tax journal, 64, 651–688. aperio group. (2017). the double bottom line: tax-loss harvesting for the altruistic investor [pdf slides]. (available at https://www.aperiogroup.com/resource/623/node/download). bakija, j., & heim, b. t. (2011). how does charitable giving respond to incentives and income? new estimates from panel data. national tax journal, 64, 615–650. bessembinder, h. (2018). do stocks outperform treasury bills? journal of financial economics, 129, 440–457. booth, d. g., & fama, e. f. (1992). diversification returns and asset contributions. financial analysts journal, 48, 26–32. bouchey, p., nemtchinov, v., paulsen, a., & stein, d. m. (2012). volatility harvesting: why does diversifying and rebalancing create portfolio growth? journal of wealth management, 15, 26–35. boyle, p. s., loewy, d. j., reiss, j. a., & weiss, r. a. (2004). the enviable dilemma: hold, sell, or hedge highly concentrated stock? journal of wealth management, 7, 30–44. bradley, r., holden, s., & mcclelland, r. (2000). a robust estimation of the effects of taxation on charitable contributions. bureau of labor statistics. mimeo. burman, l. e., & randolph, w. c. (1994). measuring permanent responses to capital-gains tax changes in panel data. american economic review, 84, 794–809. campbell, j. y., lettau, m., malkiel, b. g., & xu, y. (2001). have individual stocks become more volatile? an empirical exploration of idiosyncratic risk. journal of finance, 56, 1–43. chan, k. c., karceski, j., & lakonishok, j. (1999). on portfolio optimization: forecasting covariance and choosing the risk model. review of financial studies, 12, 937–974. constantinides, g. m. (1983). capital market equilibrium with personal tax. econometrica, 51, 611–636. constantinides, g. m. (1984). optimal stock trading with personal taxes: implication for prices and the abnormal january return. journal of financial economics, 13, 65–89. dammon, r. m., dunn, k. b., & spatt, c. s. (1989). a reexamination of the value of tax options. review of financial studies, 2, 341–372. erb, c. b., & harvey, c. r. (2006). the strategic and tactical value of commodity futures. financial analysts journal, 62, 69–97. fama, e. f. (1970). efficient capital markets: a review of theory and empirical work. journal of finance, 25, 383–417. fama, e. f. (1991). efficient capital markets: ii. journal of finance, 46, 1575–1617. feld, a. (1999). high exposure to low-basis stock: too much of a good thing? cpa journal, 69, 60–65. fraser, s. p., & jennings, w. w. (2010). examining the use of investment policy statements. journal of wealth management, 13, 10–22. greene, p., & mcclelland, r. (2001). taxes and charitable giving. national tax journal, 54, 433–453. horan, s. m. (2005). tax-advantaged savings accounts and tax-efficient wealth accumulation. research foundation of cfa institute. malkiel, b. g. (2003). the efficient market hypothesis and its critics. journal of economic perspectives, 17, 59–82. reichenstein, w. (2007). calculating after-tax allocation is key to determining risk, returns, and asset location. journal of financial planning, 20, 44–53. shoven, j. b., & sialm, c. (2003). asset location in tax-deferred and conventional savings accounts. journal of public economics, 88, 23–38. smith, m. h., & smith, g. (2008). harvesting capital gains and losses. financial services review, 17, 309–321. statman, m. (1987). how many stocks make a diversified portfolio? journal of financial and quantitative analysis, 22, 353–363. stein, d. m., vadlamudi, h., & bouchey, p. (2008). enhancing active tax management through the realization of capital gains. journal of wealth management, 10, 9–168. 276 j. whitworth / financial services review 27 (2018) 257-277 welch, s. (2002). comparing financial and charitable techniques for disposing of low basis stock. journal of wealth management, 4, 37–46. whitworth, j., & mccormack, j. (2011). tax-efficient stock investing for charitable donors. journal of financial planning, 24, 39–44. willenbrock, s. (2011). diversification return, portfolio rebalancing, and the commodity return puzzle. financial analysts journal, 67, 42–49. 277j. whitworth / financial services review 27 (2018) 257-277 the impact of the tax cut and jobs act on ira choice for moderate income investors timothy manuela,*, lee tangedahla, kent swifta acollege of business administration, university of montana, missoula, mt 59812, usa abstract while the choice of appropriate individual retirement account (ira) type involves many factors, it is well known that differences in tax rates during the working and retirement years impact the optimal choice. lower rates of return on investment may cause an investor to prefer a traditional ira whereas higher rates of return often result in higher after-tax future wealth for the roth ira. the lower statutory tax rates for individuals resulting from the 2018 tax cut and jobs act raises the breakeven rate of return where the roth becomes preferable for many moderate income individuals who invest the amount of the deduction generated by the traditional ira. the new tax law significantly increases the breakeven rates for investors who begin investing at an older age. moderate income investors are defined as those who will be in the 22% marginal tax bracket during their working years. many investors do not invest the tax savings generated by a qualifying traditional ira. for these individuals a spreadsheet model is used to compare contribution and retirement real consumption. the traditional ira can be a better choice for many moderate income individuals who want to maximize their total lifetime real consumption rather than retirement consumption, and the compression in tax rates of the new tax law increases the number of outcomes where the traditional ira yields higher total consumption. © 2018 academy of financial services. all rights reserved. jel classification: paper classifications; g110 portfolio choice; investment decisions; d140 household saving; personal finance keywords: financial planning; retirement account; personal finance; personal savings * corresponding author. tel.: �1-406-243-2511; fax: �1-406-243-6925. e-mail address: tim.manuel@business.umt.edu (t. manuel) financial services review 27 (2018) 345-365 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. 1. introduction contributions to roth individual retirement accounts (iras) are not tax deductible, while for many moderate income investors contributions to a traditional ira are deductible. subject to certain rules the withdrawals from a roth ira are not taxed, whereas traditional ira withdrawals are taxable. the new tax cut and jobs act (tcja) generally increases the number of moderate income individuals who may prefer the traditional over the roth ira. moderate income investors are defined as those who will be in the 22% marginal tax bracket during their working years.1 when determining whether to choose a traditional ira or a roth ira conventional wisdom indicates one should consider the individual’s tax rate today versus their expected future tax rate (vanzante and fritzsch, 2013). if a person believes the future tax brackets and tax rates will be lower, or if they will have lower income in retirement, then a traditional ira may be the best choice (grossman and rose, 2012). in comparison, higher rates of return and contribution amounts lead to higher tax rates on traditional ira withdrawals and, thus, favor the roth ira. similarly, lower rates of return on investment and smaller contributions favor the traditional ira. the compression of the tax brackets, and the general reduction of the marginal tax rates in the new law, raises the breakeven rate of return beyond which an investor may achieve a higher after-tax future value with a roth ira. the breakeven rates of return are calculated before and after the new tax law. the new tax law favors the traditional ira for a broader range of returns for individuals who invest the deduction. for older investors with only 15 years to retirement the breakeven rate of return where the roth provides a greater future value is as high as 21.95% under the new tax law. most investors are not likely to achieve an average return this high and may be better off with a traditional ira. adelman and cross (2010) and beshears, choi, laibson, and madrian (2017) indicate that many people do not actually invest the deduction generated by contributions to a traditional ira, preferring instead to consume the extra income. when comparing future values of the two types of iras, the roth ira yields higher futures values than the traditional ira without the deduction invested in the majority of cases. this has led some authors such as beshears et al. (2017) to conclude that the roth provides consistently higher future savings than the traditional ira. however, choosing the roth reduces an investor’s real consumption during their working years compared with a traditional ira because the roth contributions are not deductible. even with the higher future value of a roth ira, the traditional ira could provide a higher level of lifetime consumption. monte carlo simulations were used to determine whether lifetime real consumption is maximized with a traditional or a roth ira for investors who do not invest the deduction generated by a traditional ira under various return scenarios. the traditional ira provided higher median lifetime consumption than the roth ira in the majority of our simulations for moderate income individuals who reduce the risk of their portfolios as they age. the traditional ira can be a better choice for many moderate income individuals, regardless of whether they invest the deduction from traditional ira contributions if current spending is a concern along with funding retirement. this has not generally been discussed in the literature on retirement investing.2 346 t. manuel et al. / financial services review 27 (2018) 345-365 2. ira usage according to the report on the economic well being of households in 2016 by the board of governors of the federal reserve system (2017) about 31% of households have an ira. the most common type ira is the traditional ira, and only a subset of households has a roth ira or both types. about half of households have a 401(k), and about 25% have a defined benefit plan, whereas about 28% of households have no retirement investments. the report shows that investors with roth iras are more likely to be higher income, more financially sophisticated and have utilized a financial planner than those who have a traditional ira. smith, finke, and huston (2012) find similar results. they show that investors with roth iras were much more likely to be “financially sophisticated” as measured by the extent of their equity investments and/or had consulted a financial planner. roth iras are apparently still not well understood by many households who do not use financial planners, or perhaps planners favor roth iras for their clients. for many investors the best ira choice depends on many factors such as age, expected portfolio returns over time, risk aversion, liquidity needs, and the willingness to engage in tax management strategies, particularly during retirement. for more on some of the factors that favor the traditional ira see saunders (2018). 2.1. traditional ira an ira is not an investment per se, but is a tax advantaged investment vehicle. for a traditional ira, all deductible contributions and account earnings are subject to income taxation upon disbursement. thus, an ira is a type of tax-deferred savings account. if an amount is withdrawn before age 591⁄2, the taxpayer is required to pay tax on the amount withdrawn along with a 10% penalty for the early withdrawal.3 at age 701⁄2 an investor in a traditional ira must begin making withdrawals from the ira. the withdrawals are termed required minimum distributions (rmds). rmds are calculated as the balance from the end of the previous year or life expectancy from the irs’ uniform lifetime table. the table can be found at https://www.irs.gov/pub/irs-tege/uniform_rmd_wksht.pdf. failure to make the rmd results in a 50% excise tax. additional tax details are available from the irs’s publication on iras found at https://www.irs.gov/retirement-plans/individualretirement-arrangements-iras. roth iras do not require the investor to make rmds during their lifetime. a beneficiary of a roth ira must make rmds over the beneficiary’s expected life span, but if the beneficiaries are children the required withdrawal is usually small because of a long expected life span. roth iras can, thus, be better for individuals who wish to fund a bequest to their heirs rather than provide funding for their own retirement. individuals can open a traditional ira as long as they are not older than 701⁄2. in 2018 contributions to iras are limited to the lesser of $5,500 ($6,500 if 50 or older) or their taxable compensation for the year; although our model assumes that the irs, in accordance with congressional mandates, increases the allowable contribution amount over time to keep up with inflation. the new tcja calculates the indexing of contribution amounts and tax brackets using the chained cpi-u index rather than the traditional cpi-u.4 the other main tax feature of the traditional ira is that, subject to the investor’s income and pension 347t. manuel et al. / financial services review 27 (2018) 345-365 contribution, the contributions to a traditional ira may be tax deductible in the year of the contribution. 2.2. roth ira in a roth ira the contributions are never tax deductible. however, the contributions and account earnings are not taxable when withdrawn as long as the individual has had the roth ira for five years or more and does not make withdrawals before age 591⁄2. under these conditions, roth ira distributions are not subject to income taxation. 2.3. additional tax considerations for a traditional ira contributions to a traditional ira can usually be used to reduce an individual’s modified adjusted gross income (magi). table 1 filing status and deductibility provides the 2018 taxable year table applicable for individuals that are, or are not, covered by a retirement plan at their place of employment. table 1 filing status and deductibility with a retirement plan at work (2018 tax year) panel a: filing status and deductibility with a retirement plan at work for 2018 if your filing status is . . . and your magi is. . . then you can take single or head of household $63,000 or less a full deduction up to your contribution limit more than $63,000 but less than $73,000 a partial deduction $73,000 or more no deduction married filing jointly or qualified widower $101,000 or less a full deduction up to your contribution limit more than $101,000 but less than $121,000 a partial deduction $121,000 or more no deduction married filing separately less than $10,000 a partial deduction $10,000 or more no deduction panel b: filing status and deductibility without a retirement plan at work in 2018 if your filing status is . . . and your magi is. . . then you can take single, head of household, or qualifying widower any amount a full deduction up to your contribution limit married filing jointly or separately with a spouse who is not covered by a plan at work any amount a full deduction up to your contribution limit married filing jointly with a spouse who is covered by a plan at work $189,000 or less a full deduction up to your contribution limit more than $189,000 but less than $199,000 a partial deduction $199,000 or more no deduction married filing separately with a spouse who is covered by a plan at work less than $10,000 a partial deduction $10,000 or more no deduction note: magi � modified adjusted gross income. 348 t. manuel et al. / financial services review 27 (2018) 345-365 panel a of table 1 provides deductibility of a traditional ira for various filing statuses for individuals that are covered by a retirement plan at work and panel b of table 1 provides similar information for those who do not have a retirement plan at work. withdrawals from a traditional ira in retirement can cause social security benefits to be taxable, although adelman and cross (2010) suggest that this does not usually result in a very large increase in taxes. medicare premiums may also increase with increased income resulting from a traditional ira withdrawal, but not from a roth. a traditional ira also has a valuable option component (baxendale and coppage, 2014). under existing tax rules all or part of traditional iras can be converted to roth iras (termed an ira rollover). this option is preserved under the 2018 tcja, although the ability to switch back to a traditional ira was eliminated in the new tax bill. 2.4. additional tax considerations for roth iras unlike a traditional ira, the principal invested in a roth ira can generally be withdrawn without taxation before age 591⁄2 if the funds have been invested for at least five years. not all investors can make roth contributions. if an investor makes more than $199,000 in 2018 and has a tax status of married filing jointly then the person cannot contribute to a roth ira account, although this income level increases over time. if a person is single, head of household or married filing separately the individual cannot contribute to a roth if they make more than $135,000, although this income level also increases over time.5 the contribution limits are the same for the traditional and the roth ira and the limit is cumulative for the two in a given tax year. excessive contributions to either type of ira result in a six percent tax penalty that is imposed each year until the excessive contribution and its earnings are removed. an investor must also file form 5329, or face an additional tax of up to 25% of the excise tax. see the irs explanation titled, “about form 5329, additional taxes on qualified plans (including iras) and other tax-favored accounts” found at https://www.irs.gov/forms-pubs/ about-form-5329 or the general discussion provided by reichert (2011). 2.5. math models of future values of traditional and roth iras with constant returns and the impact of the tcja examining mathematical models comparing the future value of a traditional and a roth ira can highlight the impacts of the new tax law on the choice. one cannot compare equal dollar investments in a qualifying traditional and roth ira because investments in the roth are after-tax, but are pre-tax in a traditional ira. most literature comparing the two performs an adjustment to put the two on the same after-tax basis.6 because investing x dollars in a traditional ira generates a tax deduction that can be invested, many models compare the future value at retirement (or an annuity beginning at retirement) of the traditional ira after-tax plus the future value of the extra amount invested as the result of the deduction (see for instance horan, 2002, 2003). because the value of the deduction is the tax rate at contribution times the deduction, it is easy to set up a comparison of the future values after-tax of the two alternatives. the latter amount is usually assumed to be invested 349t. manuel et al. / financial services review 27 (2018) 345-365 in a taxable account. one then compares this sum to the future value of x dollars invested in a roth ira. in particular, the traditional ira combined with a taxable investment will be preferable to the roth ira if the future value of traditional ira � future value of taxable investment � future value of roth ira or if (1 � r)n �1 � tw� � tc�1 � r�1 � tc�� n � �1 � r�n (1) where r is the constant pre-tax rate of return, n is the number of years to retirement, tc is the tax rate at contribution and tw is the tax rate upon withdrawal.7 withdrawals are assumed to begin after age 591⁄2 and at least 5 years after the funds are invested so that there are no tax penalties. sibley (2002) and various others model the future values of the two in this way. if the savings from the traditional ira are invested in an account with tax deferral benefits, such as non-dividend paying stocks or certain stock mutual funds, then the model understates the future value of the traditional ira. for details see horan (2003, 2006). higher rates of return, r, and time invested, n, favor the roth over the traditional ira (al zaman 2008; cook, meyer, and reichenstein, 2015; horan, 2003; horan and al zaman, 2009; reichenstein, 2006; reichenstein, horan, and jennings 2015; sibley 2002). funds invested in any tax deferred annuity (tda) grow at the pre-tax interest rate r while funds invested in a fully taxable account grow at the after tax rate r(1�tc). because of the difference, at higher rates of return the roth ira can generate higher retirement income than the after tax value of the combination of the traditional ira and a taxable investment even when the tax rate is lower during retirement. previous literature such as horan (2003) has shown that while it is often the case that the traditional ira combined with a taxable account will yield a higher future value than a roth ira if tw � tc, sometimes the roth is better even when the tax rate at withdrawal is lower than at the point of contribution if r is high enough. before the tcja many moderate income investors were likely to be in the 25% tax bracket during their working years and in the 15% bracket upon retirement. table 2 compares the breakeven rates of return (r) under the old and new tax laws. at r � 9% or higher the roth is preferred even if the tax rate at contribution is 25% and at withdrawal is substantially lower at 15%. before the tcja of 2018 the personal marginal tax rates started at 10%, increased to 15%, then to 25%, then to 28% and so on up to a maximum of 39.6%. under the tcja the personal marginal rates start at 10%, then progress to 12%, 22%, 24%, and so on, up to 37%. as shown in table 2 the generally lower tax rates under the new law favor the traditional ira over the roth ira for a broader range of returns earned on invested funds with a given investment horizon.8 for a 25 year investment horizon under the 2017 tax rates an investor who had a 25% tax rate during the contribution years and 15% during retirement and who earned a pre-tax rate of return r � 8% on all investments would have been better off in a traditional ira. the traditional is better because the sum of the after-tax future value per dollar invested in the traditional ira plus the after-tax future value of the taxable account (� 6.89417) is greater than the future value of the roth ira (� 6.848475). if 350 t. manuel et al. / financial services review 27 (2018) 345-365 the investor earned nine percent on all alternatives; however, then the roth generates a higher future value. panel a of table 3 contains the breakeven rates of return under the old and the new tax rates for various investment periods if the deduction generated by the traditional ira is invested and is fully taxable at tc. for instance, with a 25 year time horizon to retirement, the breakeven r under the old tax rates is 8.8024%. at all lower rates of return the traditional ira is preferable, but at rates above the breakeven the roth ira is preferred, even though the tax rate is substantially lower during retirement. with the new tax law a similar investor may face tc of 22% and tw of 12%. in this case the breakeven rate of return is increased to table 2 future value of equal after-tax investments in traditional ira plus taxable account compared with future value of roth ira fv traditional ira fv taxable account fv traditional ira � taxable account fv roth tax law tc tw r ra-t n (1�r)n(1�tw) tc(1�ra-t)n sum (1�r)n preference 2017 25% 15% 7.00% 5.25% 25 4.61332 0.89845 5.51177 5.427433 traditional 25% 15% 8.00% 6.00% 25 5.82120 1.07297 6.89417 6.848475 traditional 25% 15% 9.00% 6.75% 25 7.32962 1.27979 8.60940 8.623081 roth 25% 15% 10.00% 7.50% 25 9.20950 1.52458 10.73408 10.83471 roth 2018 22% 12% 8.00% 6.24% 25 6.02666 0.99914 7.02579 6.848475 traditional (tcja) 22% 12% 9.00% 7.02% 25 7.58831 1.19963 8.78794 8.623081 traditional 22% 12% 10.00% 7.80% 25 9.53454 1.43844 10.97298 10.83471 traditional 22% 12% 11.00% 8.58% 25 11.95521 1.72254 13.67775 13.58546 traditional 22% 12% 12.00% 9.36% 25 14.96006 2.06009 17.02014 17.00006 traditional 22% 12% 13.00% 10.14% 25 18.68288 2.46065 21.14353 21.23054 roth 22% 12% 14.00% 10.92% 25 23.28649 2.93541 26.22190 26.46192 roth 22% 12% 15.00% 11.70% 25 28.96868 3.49745 32.46613 32.91895 roth note: ira � individual retirement account; tcja � tax cut and jobs act; tc � tax rate at time of contribution; tw � tax rate upon withdrawal; r � pre-tax rate of return, assumed constant; ra-t � after tax rate of return on earnings using tc; n � number of years until retirement. table 3 breakeven return r vs. years to retirement with reinvestment preand post-tcja panel a: breakeven returns with tax deduction invested in a fully taxable account years to retirement � 15 years 25 years 30 years 35 years 40 years breakeven r old tax law 15.4637% 8.8024% 7.2425% 6.1522% 5.3472% breakeven r new tax law 21.9536% 12.2185% 10.0010% 8.4647% 7.3376% increase in breakeven r 6.4899% 3.4161% 2.7585% 2.3126% 1.9904% panel b: breakeven returns with tax deduction from traditional ira invested with an effective tax rate (ti) of 10% years to retirement � 15 years 25 years 30 years 35 years 40 years breakeven r old tax law 50.3346% 25.3536% 20.3129% 16.9441% 14.5337% breakeven r new tax law 65.5725% 31.4991% 25.0028% 20.7278% 17.7013% increase in breakeven r 15.24% 6.15% 4.69% 3.78% 3.17% note: ira � individual retirement account; tcja � tax cut and jobs act. 351t. manuel et al. / financial services review 27 (2018) 345-365 12.2185%, and the traditional ira is preferred over a broader range of returns. as time to retirement (n) increases, the breakeven rates fall under both tax regimes because higher after-tax future values are produced by the roth at longer n. the difference between the breakeven rates of return induced by the lower tax rates under the new law is thus reduced as n is increased. this implies that the new tax law should not have as large an influence on the choice of ira for younger investors. however, older investors facing the same tax rates as depicted above who are starting to invest with a shorter time to retirement such as 15 years may prefer a traditional ira for almost all investment portfolios. in this case the breakeven r where the roth provides a higher future value is 21.95% under the new tax law. most investors are unlikely to achieve this large of an average return. panel b of table 3 depicts the various breakeven rates of return under the old tax law and the new assuming that the deduction generated by the traditional ira is invested in a tax-advantaged investment.9 horan (2003, 2006) and others have shown that the ability to invest the deduction in a mutual fund with a tax deferral feature, or in a tax advantaged investment increases the future value of the traditional ira. to illustrate the impact on the breakeven rate of return the effective tax rate on the amount invested (ti) is assumed to be 10% to generate the results in panel b. the breakeven rates of return are substantially increased under both the old and the new tax rates and the traditional ira would likely provide a higher after tax future value under most return scenarios. recent work by adelman and cross (2010) and beshears et al., (2017) indicate that many people do not invest the tax savings from the traditional ira, preferring instead to consume the extra income. this possibility complicates the choice of the two types because future retirement income will vary with investor behavior.10 the fv of the roth will be greater than the fv of the traditional ira at lower breakeven rates of return if investors do not reinvest the tax savings because in this case the future value of the traditional ira per dollar is (1�r)n(1�tw) � tc, which is less than the future value of the roth ira of (1�r)n if tc/tw � (1�r)n.11 the roth will generally provide a higher future value under a broad range of investment returns, especially if n is large. table 4 provides data on the future values of the roth and traditional iras for a 20 year work period at different r when the tax savings from the traditional ira are not reinvested under the old and the new tax law.12 when an individual does not invest the tax savings from the traditional ira, the roth is preferred if (1�r)n � tc/tw. the top part of table 4 depicts rates of return where the table 4 future value of equal after-tax investments in traditional ira with tax savings not invested compared with future value of roth ira tax law tw tc r n tc/tw (1�r)n preference 2017 15% 25% 2.00% 20 1.6667 1.48595 traditional 15% 25% 3.00% 20 1.6667 1.80611 roth 15% 25% 4.00% 20 1.6667 2.19112 roth 2018 12% 22% 2.00% 20 1.8333 1.48595 traditional (tcja) 12% 22% 3.00% 20 1.8333 1.80611 traditional 12% 22% 4.00% 20 1.8333 2.19112 roth 12% 22% 5.00% 20 1.8333 2.65330 roth note: ira � individual retirement account; tcja � tax cut and jobs act. 352 t. manuel et al. / financial services review 27 (2018) 345-365 traditional or the roth ira are preferred for an investor under the old tax law who is in the 25% tax bracket during their working years, and is in the 15% tax bracket during retirement with 20 years to retirement. in this case the investor should prefer the roth ira if their pre-tax rate of return is expected to be 3.00% or higher. the breakeven return is actually 2.5870% as shown in table 5. once again the impact of the tcja is to raise the breakeven rate of return where the roth is preferable. with the new tax rates the traditional ira is still preferred at a 3.00% rate of return, but at a 4.00% return the roth again provides a higher future value assumption. the actual breakeven rate of return is 3.0771% as shown below, which is the case where (1�r)20 � 1.8333. the breakeven rates in table 5 indicate that in most return scenarios the roth is likely to provide more retirement income than the traditional ira given the low breakeven rates even with the new tax law. 3. lifetime consumption versus retirement consumption choosing the traditional ira generates more spending power in the working years than the roth on a dollar for dollar basis. varying time preference for consumption and liquidity needs may still cause an investor who does not reinvest the tax savings to prefer the traditional ira even at rates of return above the breakeven rates shown above. it is possible to model real consumption over the working and retirement years from the two alternatives. the traditional ira can be optimal with a lifetime consumption decision variable, particularly given uncertain reinvestment rates and low tolerance for risk, both of which will reduce the average return earned. consider the case for a couple who are married filing jointly, neither of whom are covered by a plan at work, who do not reinvest the tax savings generated by the deductibility of contributions to a traditional ira. choosing this filing status limits the likelihood of running afoul of income limits on deductibility of ira contributions. it is assumed that the couple will use the standard deduction under the new tcja, and then real consumption in year t during their working years after investing in a traditional ira is: real consumptiont � incomet � standard deductiont � traditional ira contributiont � taxest � standard deductiont �1 � deflator�t (2) where the deflator represents cumulative inflation so that real consumption is measured in terms of the date when the ira contributions begin. this model is drawn from bodie, kane, and marcus (2008). the traditional ira contribution reduces the couple’s taxable income and, thus, taxes owed in year t because these ira contributions are deductible. the sum of table 5 breakeven return vs. years to retirement without reinvestment years to retirement � 15 years 20 years 25 years 30 years 40 years breakeven r old tax law 3.4642% 2.5870% 2.0643% 1.7173% 1.2853% breakeven r new tax law 4.1237% 3.0771% 2.4542% 2.0410% 1.5269% difference 0.6595% 0.4900% 0.3899% 0.3237% 0.2416% 353t. manuel et al. / financial services review 27 (2018) 345-365 the first three terms in the numerator of eq. (2) is taxable income in year t. the standard deduction in time t is then added back after taxes are calculated on taxable income. ignoring social security benefits and other income during retirement, the cumulative savings from the future value of the invested ira contributions can be annuitized upon retirement. for simplicity our hypothetical couple retires at the same time. a nominal annuity amount is then used as the only source of income during retirement. real consumptiont in year t during retirement is thus: real consumptiont � nominal annuity � standard deductiont � taxest � standard deductiont �1 � deflator�t (3) if the couple chooses a roth ira instead of the traditional ira then real consumptiont during the working years is found as: real consumptiont � incomet � standard deductiont � taxest � roth ira contributiont � standard deductiont �1 � deflator�t (4) in this case the taxest are calculated from the income minus the standard deduction before considering the roth ira contribution as roth contributions are not tax deductible. real consumptiont in year t during retirement for the roth is thus: real consumptiont � nominal annuity �1 � deflator�t because no taxes are owed on the roth withdrawals. compared with the roth, the traditional ira provides extra real consumption during the working years (t � 1 to n), but at a cost of reduced consumption during the retirement years (t � n�1 to end of life) because of taxes that will have to be paid as shown below: �t�1 n tc�traditional ira contributiont� �1 � deflatort� � �t�n�1 life tw�nominal annuity � standard deductiont� �1 � deflatort� (5) the additional consumption during the working years is tc(traditional ira contributiont) per year, but choosing the traditional ira results in additional taxes in retirement on the annuity as compared with the roth. this retirement income is taxed at the rate tw. the taxable income is the nominal annuity – standard deductiont. the size of the nominal annuity is dependent on the number of years of retirement, the returns on the contributions, the number of years of contributions and the contribution amount. to have a net gain in real consumption the additional consumption during the working years (the first term above) must be larger in aggregate than the additional taxes that occur during retirement (the second term). the larger the nominal annuity the less likely the net gain will be positive, although this can be offset by a higher standard deduction if tax law allows indexing the deduction. in addition, if tw is sufficiently lower than tc the net gain can be positive. a lower tax rate in retirement may occur because the investor will be in a lower tax bracket, or because 354 t. manuel et al. / financial services review 27 (2018) 345-365 statutory tax rates are lower. higher inflation will increase the deflator more rapidly and will reduce the real value of consumption in retirement for both ira types. however, if higher inflation raises the investment rate of return, then inflation will favor the roth because higher average nominal returns generate more taxes in retirement and favor the roth in general. greater risk aversion, as indicated by reducing risk at an earlier age and, thereby, lowering the average return, favors the traditional ira because it lowers the lifetime average return on investment and, thus, potentially, the taxes paid in retirement. 3.1. the impact of the tcja on lifetime consumption for the traditional and roth ira the model can be used to demonstrate the impact of the tcja on the choice of traditional and roth ira on lifetime real consumption. in the model, tax brackets are assumed to be indexed to inflation as they have been in the past.13 investment returns are based on historical geometric average returns, correlations, and standard deviations of major asset classes such as stocks, bonds and bills using data from damodaran (2018). the model incorporates the concept of reducing portfolio risk as the investors approach retirement. all contributions cease at retirement. the model ignores other income and social security benefits. the invested amount at retirement is annuitized over the remaining expected lifetime. the model used allows an investor to choose at what age they plan to reduce risk in the portfolio by shifting more of the portfolio from stocks to treasury bonds and treasury bills. the return data and portfolio weights based on damodaran’s (2018) data are given in table 6. the asset allocation is shown above for assumed high risk, medium risk, and low risk portfolios. the choice of age at which an investor switches from a high to medium to low risk fund is a proxy for risk aversion. greater risk aversion is indicated by reducing risk at an earlier age and/or choosing a lower percentage in stocks in each portfolio. the spreadsheet model provides stochastic rates of return to determine future values based on the input data. this is substantially different than in much of the literature where many studies assume a fixed rate of return, or allow only a small amount of variation in returns.14 for simplicity the table 6 portfolio risk and return based on damodaran data stocks long term (lt) government bonds t-bills stocks lt government bonds stocks t-bills lt government bonds t-bills average return 9.65% 4.88% 3.39% � �0.028 �0.030 0.296 � 19.62% 7.72% 3.05% cov �0.000418339 �0.000178894 0.000698036 high risk 70% 30% 0% 0.0192257 medium risk 60% 40% 0% 0.0146133 portfolio variances low risk 30% 50% 20% 0.0049861 risk portfolio average return portfolio � high 8.22% 13.87% medium 7.74% 12.09% low 6.01% 7.06% 355t. manuel et al. / financial services review 27 (2018) 345-365 model does not have any excess contributions, which are rare in practice (see sibley, 2002), and there are no rmds because of the focus on lifetime real consumption rather than bequests. 3.2. impact of ira choice on consumption preand post-tcja the real consumption outcomes for a hypothetical married couple filing jointly that are not covered by a plan at work who are 30 years old, will retire at age 65, and fund their retirement to age 90 is shown below using the 2017 tax rates and rules. in 2018 the life expectancy of a 30 year old u.s. male is 71 and for a female is 75 according to countryeconomy.com. funding to age 90 should be amply sufficient to ensure an investor’s funds do not run out. the 2017 progressive tax rates and brackets are used and the brackets are indexed to inflation as per irs practice as mandated by congress. the irs normally adjusts the standard deduction for inflation as well, and it is also increased at age 65, and is higher for taxpayers that meet certain conditions.15 the couple starts with a high risk portfolio that has an average return of 8.22% with a standard deviation of 13.87%, reduces to medium risk at age 50 with an average return of 7.74% with a portfolio standard deviation of 12.09%, and reallocates to a low risk portfolio at retirement that has an average return of 6.01% and a standard deviation of 7.06%. the couple’s starting income is $80,000 between the two of them at age 30; subsequently, their income grows at six percent per year and inflation is three percent per year. the couple contributes $10,000 total to either a roth or a traditional ira each year until they retire. annual rates of return are generated from a normal distribution with the given geometric average return and standard deviation using monte carlo simulation. the cumulative savings at retirement is annuitized over the given life expectancy using the low risk portfolio return. a simulation of 500 different outcomes was created with a spreadsheet macro.16 the results are presented in table 7. each row in table 7 gives the results of 500 different trial runs with different returns in each trial. table 7 shows that using the 2017 tax rules in place before the tcja, the traditional ira provided higher lifetime consumption in 383 out of the 500 trials (77%), with the roth ira doing better in only 117 runs (23%). the median lifetime consumption from the roth was $5,059,391 whereas the median consumption for the traditional ira was $5,078,552.17 the second row replicates the results using the same inputs except that the tcja progressive tax rates and the new standard deduction (with no personal exemptions) was used. in this case the traditional ira provided higher lifetime real consumption in 361 (72%) out of the 500 trials and the roth table 7 representative simulation of traditional vs. roth ira lifetime consumption preand post-tcja no. runs traditional ira had higher lifetime consumption no. runs roth had higher lifetime consumption median lifetime consumption traditional median lifetime consumption roth difference in median lifetime consumption pre tcja choice of ira 383 (77%) 117 (23%) $5,078,552 $5,059,391 $19,161 post tcja choice of ira 361 (72%) 139 (28%) $5,049,170 $5,015,551 $33,619 note: ira � individual retirement account; tcja � tax cut and jobs act. 356 t. manuel et al. / financial services review 27 (2018) 345-365 provided more consumption in 139 trial runs (28%).18 the tcja modestly increases the median lifetime consumption, but the simulation indicates the tcja does not have a major impact on ira choice when the decision variable is lifetime consumption. higher rates of return and higher incomes will increase the number of times that the roth provides more favorable outcomes. similarly if congress decides to not increase the brackets or the standard deduction with inflation, then the roth ira will provide better outcomes in the large majority of simulation runs because the retirement tax liability is then much greater for the traditional ira. calculations for the scenario outputs shown below are provided in the appendix. in the majority of cases the traditional ira provided higher median lifetime consumption than the roth ira preand post-tcja. for moderate income individuals where current spending is a primary concern as much as funding retirement, the traditional ira can be the better choice. higher levels of starting income imply a higher tax burden in retirement and favor the roth ira. 4. conclusions the impact of the new lower rates in the tcja should make the traditional ira more attractive for moderate income investors who invest the tax deduction. the impact of the tax law is to raise the breakeven rates of return where the roth ira becomes the better alternative in terms of future value for many investors who may be in the 22% tax bracket in their working years and in the 12% tax bracket in retirement. this is likely to include a large number of investors. depending on the length of the contribution period, the breakeven rate of return is increased by between 2% to as much as 6.5% for individuals who invest the tax savings from their traditional iras. there are only small changes in breakeven for investors who do not invest the difference and are looking to maximize retirement income. individuals of modest income who are seeking to maximize their lifetime consumption, however, may still prefer the traditional ira to the roth. simulations with reasonably realistic investment strategies that reduce risk as the investor ages and incorporate the new tax brackets and standard deduction indicate that in many cases the traditional ira provides greater median lifetime consumption than the roth. this result does not hold in simulations where tax brackets are not indexed and/or when the standard deduction is not increased by inflation, so future tax rules are critical in the choice of ira. finally, investors and planners should be aware that many variables should be considered in the choice of ira, and analyzing after tax returns is not sufficient to fully inform the decision. behavioral characteristics such as whether the tax savings will be invested, the investor’s risk tolerance, their desire to fund working years’ consumption versus retirement consumption, expected market returns and the investor’s willingness and ability to engage in tax management strategies should be considered in the choice (horan, 2006). if investors (or their planners) are willing to engage in tax management strategies then having both may be optimal; nevertheless, the traditional ira remains a good choice for many investors. 357t. manuel et al. / financial services review 27 (2018) 345-365 notes 1 for the 2018 tax year a moderate income investor that is married and is filing jointly must have modified adjusted gross income (magi) over $77,401 according to the irs. 2 in reality many factors other than future value or consumption can affect the optimal ira type as discussed below. beshears et al. (2017) indicate that factors other than future value should be considered in the choice of ira type. 3 there are several “personal hardship” and other exemptions that allow an investor to withdraw money before age 591⁄2 without facing a tax penalty. nevertheless, it is generally not a good idea to withdraw funds invested for retirement unless absolutely necessary. 4 the cbo estimates the chained cpi (c-cpi-u) results in about a 0.25% lower average inflation rate than the cpi. the chained version is thought to be more accurate because it allows for substitutions to lower cost items as relative prices change and corrects for a small size sampling bias whereas the traditional cpi measure does not. for more detail see cbo’s projections of demographic and economic trends found at https://www.cbo.gov/system/files?file�2018-06/53919-2018ltbo-appendixa.pdf. 5 a phase-out of the amount contributed applies for lower income limits regardless of filing status. an investor can still create a “backdoor” roth ira if their magi is too high to directly contribute to a roth ira. the procedure is to contribute to a non-deductible ira and then roll it over to a roth ira. there is no magi limit on the rollover. if the investor has a deductible ira then part of the rollover is taxable. for more information see https://www.irs.gov/retirement-plans/retirement-plans-faqsregarding-iras-rollovers-and-roth-conversions or the discussion in harline (2014). 6 an exception is horan (2003) who compares equal pre-tax investments as well as equal after-tax investments. the preor after-tax method of comparisons are equivalent assuming that the tax deductions generated by the traditional ira are invested in the pre-tax case. thus, one need only compare an equivalent pre-tax investment amount in the traditional and roth ira and vary whether the deduction is invested. 7 this model assumes the money in both iras is withdrawn upon retirement. alternatively, n may be considered the time to when the investor begins withdrawing from the iras. this is the time to age 701⁄2 at the maximum for the traditional ira, but n could be longer for a roth ira. 8 throughout the discussion it is assumed that the changes in the tax code from the tcja will persist after 2025 when they are scheduled to sunset. this discussion centers on investors who are in the 25%/15% pre-tcja tax brackets and 22%/12% post-tcja brackets during their working years and retirement years, respectively. post-tcja this includes investors whose gross income is up to $165,000 in their working years and $77,400 during their retirement years in today’s dollars. the same income numbers for pre-tcja were $153,100 and $75,900, respectively. 9 the breakeven interest rate is found by solving the following for r: (1�r)n (1�tw)�tc[1�(r(1�ti)] n � (1�r)n where ti is set to 10% and n is varied as shown in the table. 10 it is likely that many investors do not think about the tax savings as an amount that could be invested unless a planner points this out. 358 t. manuel et al. / financial services review 27 (2018) 345-365 11 the future value component [1 � r(1�tc)] n drops out if the tax savings are not reinvested. 12 alternatively, once could compare a $5,000 qualifying contribution to a traditional ira versus $5,000(1–22%) � $3,900 in a roth ira. some investors may choose this alternative rather than investing the deduction generated by the traditional ira. the relative advantage in future value of the two ira types in this case is completely determined by whether tc is greater than or less than tw. using the $5,000 pre-tax contribution amount, the future value of the traditional ira is greater than the future value of the roth ira if $5,000(1�r)n(1�tw) � $5,000(1�tc)(1�r)n, which holds if tc � tw and not otherwise. adelman and cross (2010) make a similar point. in the literature it is more common to model the comparison of not investing the deduction as done here: $5,000(1�r)n(1�tw)�tc � $5,000(1�r)n. 13 the model incorporates the higher deduction upon retirement and the additional amount that can be contributed from age 50 onward. the model also incorporates indexed standard deductions and ira contributions. results are available from the authors. indexing deductions tends to favor the traditional ira by reducing taxes due in retirement. indexing contributions favors the roth ira because bigger contributions increase retirement income and thus taxes in retirement. the roth avoids these higher taxes in retirement. 14 the option value of the ability to convert a traditional ira to a roth ira (the rollover option) is ignored. 15 the spreadsheet incorporates the increase in deduction at age 65. 16 providing a simulation of multiple outputs is a better method to evaluate the preferable type of ira than providing a single point estimate of future value. note that the model uses realistic estimates of portfolio returns at the chosen risk levels and starts with a modest income amount. 17 the simulation contains some runs with high returns. for the roth these tend to result in higher consumption, whereas for the traditional the higher incomes that result lead to higher taxes that offset at least some of the income. thus, choosing the roth ira can result in higher average lifetime consumption, even though the median is lower. in other words choosing the traditional ira forfeits the positive skewness of returns and consumption that the roth may generate. 18 simulation results (not shown) indicate that higher returns increase the number of trials where the roth provided higher consumption, but the traditional ira still pre-dominates. however, when the brackets and/or the deductions are not indexed with inflation the roth dominates in almost all scenarios. appendix: scenario calculations 1. more on the spreadsheet model and simulation larger contribution amounts favor the roth ira in the spreadsheet outcomes, which is consistent with most of the literature that concentrates on maximizing retirement income. a 359t. manuel et al. / financial services review 27 (2018) 345-365 a pp en di x t ab le 1 si m ul at io n re su lts an d sa m pl e ca lc ul at io ns se le ct ed ye ar s pr et c ja e nt er nu m be r of tr ia ls (m ax � 1, 00 0) : 50 0 l if et im e re al co ns um pt io n a ge st ar t co nt ri bu tio n a ge m ed iu m ri sk a ge lo w ri sk a ge re tir e l if e ex pe ct an cy t ra di tio na l r ot h r un si m ul at io n 30 50 60 65 90 m in im um 4, 33 9, 99 1 4, 25 6, 48 8 st ar tin g in co m e in co m e gr ow th r at e of in fla tio n c on tr ib ut io n m ax im um 7, 17 0, 88 0 7, 64 3, 98 8 50 0 m ed ia n 4, 97 4, 77 6 4, 92 9, 16 3 $8 0, 00 0 6. 00 % 3. 00 % $1 0, 00 0 a ve ra ge 5, 07 8, 55 2 5, 05 9, 39 1 d ed uc tio n ad ju st m en t 10 0% b et te r 38 3 11 7 c ha ng e in ta x ra te af te r re t 0. 00 % 3. 00 % b ra ck et ad ju st m en t 3. 00 % w or ki ng r et ir ed l if et im e c um ul at iv e sa vi ng s 1, 78 2, 73 8 a ge st ar t co nt ri bu tio n a ge m ed iu m ri sk a ge lo w ri sk a ge re tir e l if e ex pe ct an cy in co m e ye ar s 35 25 60 l if et im e ta x ra te n om in al an nu ity 13 9, 63 0 in co m e 9, 52 9, 66 9 3, 49 0, 75 2 13 ,0 20 ,4 22 r ea l an nu ity 49 ,6 22 30 50 60 65 90 t ax es , t ra di tio na l ir a 1, 31 0, 34 2 63 ,9 81 1, 37 4, 32 3 14 .4 2% r is k a ve ra ge re tu rn st an da rd de vi at io n st ar tin g in co m e in co m e gr ow th r at e of in fla tio n c on tr ib ut io n � ag e 50 st an da rd de du ct io n an d ex em pt io n t ax es , r ot h ir a 1, 46 5, 91 5 0 1, 46 5, 91 5 15 .3 8% h ig h 8. 22 % 13 .8 7% r ea l co ns um pt io n, t ra di tio na l ir a 3, 96 6, 69 7 84 5, 51 9 4, 81 2, 21 6 b et te r m ed iu m 7. 74 % 12 .0 9% $8 0, 00 0 6. 00 % 3. 00 % $1 0, 00 0 $2 5, 20 0 r ea l co ns um pt io n, r ot h ir a 3, 88 3, 19 5 86 4, 07 9 4, 74 7, 27 4 l ow 6. 01 % 7. 06 % g ro w th in ta x ra te af te r re tir em en t 0. 00 % 3. 00 % b ra ck et ad ju st m en t 3. 00 % in co m e, co nt ri bu tio ns , ex em pt io ns t ra di tio na l ir a r ot h ir a in ve st m en t/s av in gs a ge in co m e d efl at or in de xe d ir a co nt ri bu tio ns in de xe d de du ct io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n a ve ra ge re tu rn st an da rd de vi at io n c um ul at iv e sa vi ng s 30 80 ,0 00 1. 00 00 0 10 ,0 00 25 ,2 00 44 ,8 00 15 % 5, 78 8 64 ,2 13 54 ,8 00 15 % 7, 28 8 62 ,7 13 10 ,0 00 31 84 ,8 00 1. 03 00 0 10 ,3 00 25 ,9 56 48 ,5 44 15 % 6, 32 1 66 ,1 93 58 ,8 44 15 % 7, 86 6 64 ,6 93 8. 21 9% 13 .8 66 % 21 ,4 83 32 89 ,8 88 1. 06 09 0 10 ,6 09 26 ,7 35 52 ,5 44 15 % 6, 89 2 68 ,2 31 63 ,1 53 15 % 8, 48 4 66 ,7 31 8. 21 9% 13 .8 66 % 33 ,2 35 33 95 ,2 81 1. 09 27 3 10 ,9 27 27 ,5 37 56 ,8 17 15 % 7, 50 4 70 ,3 29 67 ,7 45 15 % 9, 14 3 68 ,8 29 8. 21 9% 13 .8 66 % 48 ,8 59 34 10 0, 99 8 1. 12 55 1 11 ,2 55 28 ,3 63 61 ,3 80 15 % 8, 15 8 72 ,4 88 72 ,6 35 15 % 9, 84 6 70 ,9 88 8. 21 9% 13 .8 66 % 55 ,0 78 35 10 7, 05 8 1. 15 92 7 11 ,5 93 29 ,2 14 66 ,2 52 15 % 8, 85 7 74 ,7 09 77 ,8 44 15 % 10 ,5 96 73 ,2 09 8. 21 9% 13 .8 66 % 57 ,4 39 48 22 8, 34 7 1. 70 24 3 17 ,0 24 42 ,9 01 16 8, 42 1 25 % 27 ,5 96 10 7, 92 0 18 5, 44 6 25 % 31 ,8 52 10 5, 42 0 8. 21 9% 13 .8 66 % 45 0, 11 1 49 24 2, 04 8 1. 75 35 1 17 ,5 35 44 ,1 88 18 0, 32 5 25 % 30 ,1 37 11 0, 85 0 19 7, 86 0 25 % 34 ,5 21 10 8, 35 0 8. 21 9% 13 .8 66 % 60 8, 61 4 (c on ti nu ed on ne xt pa ge ) 360 t. manuel et al. / financial services review 27 (2018) 345-365 a pp en di x t ab le 1 (c on tin ue d) in co m e, co nt ri bu tio ns , ex em pt io ns t ra di tio na l ir a r ot h ir a in ve st m en t/s av in gs a ge in co m e d efl at or in de xe d ir a co nt ri bu tio ns in de xe d de du ct io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n a ve ra ge re tu rn st an da rd de vi at io n c um ul at iv e sa vi ng s 50 25 6, 57 1 1. 80 61 1 19 ,0 61 45 ,5 14 19 1, 99 6 25 % 32 ,6 06 11 3, 45 0 21 1, 05 7 25 % 37 ,3 72 11 0, 81 2 7. 74 2% 12 .0 89 % 78 0, 97 0 51 27 1, 96 5 1. 86 02 9 19 ,6 03 46 ,8 79 20 5, 48 3 25 % 35 ,5 16 11 6, 56 5 22 5, 08 6 25 % 40 ,4 17 11 3, 93 1 7. 74 2% 12 .0 89 % 85 9, 26 1 56 36 3, 95 1 2. 15 65 9 22 ,5 66 54 ,3 46 28 7, 03 9 25 % 53 ,3 80 13 3, 54 6 30 9, 60 5 25 % 59 ,0 22 13 0, 93 0 7. 74 2% 12 .0 89 % 1, 52 8, 33 7 57 38 5, 78 8 2. 22 12 9 23 ,2 13 55 ,9 76 30 6, 59 8 25 % 57 ,7 19 13 7, 24 3 32 9, 81 1 25 % 63 ,5 22 13 4, 63 0 7. 74 2% 12 .0 89 % 1, 58 6, 66 5 58 40 8, 93 5 2. 28 79 3 23 ,8 79 57 ,6 56 32 7, 40 0 25 % 62 ,3 51 14 1, 04 7 35 1, 27 9 28 % 68 ,3 51 13 8, 42 4 7. 74 2% 12 .0 89 % 1, 41 7, 59 8 59 43 3, 47 1 2. 35 65 7 24 ,5 66 59 ,3 85 34 9, 52 0 25 % 67 ,2 96 14 4, 96 1 37 4, 08 6 28 % 73 ,8 36 14 2, 18 5 7. 74 2% 12 .0 89 % 1, 33 5, 64 5 60 45 9, 47 9 2. 42 72 6 25 ,2 73 61 ,1 67 37 3, 04 0 28 % 72 ,6 16 14 8, 97 0 39 8, 31 2 28 % 79 ,6 93 14 6, 05 5 6. 01 3% 7. 06 1% 1, 45 4, 96 8 61 48 7, 04 8 2. 50 00 8 26 ,0 01 63 ,0 02 39 8, 04 5 28 % 78 ,6 63 15 2, 94 9 42 4, 04 6 28 % 85 ,9 43 15 0, 03 7 6. 01 3% 7. 06 1% 1, 41 8, 92 5 62 51 6, 27 1 2. 57 50 8 26 ,7 51 64 ,8 92 42 4, 62 8 28 % 85 ,1 22 15 7, 04 3 45 1, 37 9 28 % 92 ,6 13 15 4, 13 4 6. 01 3% 7. 06 1% 1, 63 1, 54 7 63 54 7, 24 7 2. 65 23 4 27 ,5 23 66 ,8 39 45 2, 88 5 28 % 92 ,0 21 16 1, 25 5 48 0, 40 8 28 % 99 ,7 28 15 8, 35 0 6. 01 3% 7. 06 1% 1, 67 4, 73 0 64 58 0, 08 2 2. 73 19 1 28 ,3 19 68 ,8 44 48 2, 91 9 28 % 99 ,3 87 16 5, 59 0 51 1, 23 8 28 % 10 7, 31 6 16 2, 68 7 6. 01 3% 7. 06 1% 1, 60 0, 10 0 65 61 4, 88 7 2. 81 38 6 29 ,1 39 77 ,9 44 50 7, 80 4 28 % 10 5, 28 0 17 0, 75 0 53 6, 94 3 28 % 11 3, 43 9 16 7, 85 1 6. 01 3% 7, 06 1% 1, 78 2, 73 8 w or ki ng 9, 52 9, 66 9 64 8, 75 9 1, 60 1, 58 8 7, 27 9, 32 1 22 % 1, 31 0, 34 2 3, 96 6, 69 7 7, 92 8, 08 1 23 % 1, 46 5, 91 5 3, 88 3, 19 5 1, 78 2, 73 8 r et ir em en t t ra di tio na l ir a r ot h ir a a ge n om in al w ith dr aw d efl at or in de xe d de du ct io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n 66 13 9, 63 0 2. 89 82 8 80 ,2 82 59 ,3 48 15 % 6, 20 0 46 ,0 38 0 48 ,1 77 67 13 9, 63 0 2. 98 52 3 82 ,6 91 56 ,9 39 15 % 5, 75 7 44 ,8 45 0 46 ,7 74 68 13 9, 63 0 3. 07 47 8 85 ,1 72 54 ,4 59 10 % 5, 44 6 43 ,6 40 0 45 ,4 11 69 13 9, 63 0 3. 16 70 3 87 ,7 27 51 ,9 03 10 % 5, 19 0 42 ,4 50 0 44 ,0 89 70 13 9, 63 0 3. 26 20 4 90 ,3 58 49 ,2 72 10 % 4, 92 7 41 ,2 94 0 42 ,8 05 71 13 9, 63 0 3. 35 99 0 93 ,0 69 46 ,5 61 10 % 4, 65 6 40 ,1 72 0 41 ,5 58 72 13 9, 63 0 3. 46 07 0 95 ,8 61 43 ,7 69 10 % 4, 37 7 39 ,0 83 0 40 ,3 47 73 13 9, 63 0 3. 56 45 2 98 ,7 37 40 ,8 93 10 % 4, 08 9 38 ,0 25 0 39 ,1 72 84 13 9, 63 0 4. 93 41 2 13 6, 67 5 2, 95 5 10 % 29 5 28 ,2 39 0 28 ,2 99 85 13 9, 63 0 5. 08 21 5 14 0, 77 6 0 10 % 0 27 ,4 75 0 27 ,4 75 86 13 9, 63 0 5. 23 46 1 14 4, 99 9 0 10 % 0 26 ,6 74 0 26 ,6 74 87 13 9, 63 0 5. 39 16 5 14 9, 34 9 0 10 % 0 25 ,8 97 0 25 ,8 97 88 13 9, 63 0 5. 55 34 0 15 3, 82 9 0 10 % 0 25 ,1 43 0 25 ,1 43 89 13 9, 63 0 5. 72 00 0 15 8, 44 4 0 10 % 0 24 ,4 11 0 24 ,4 11 90 13 9, 63 0 5. 89 16 0 16 3, 19 7 0 10 % 0 23 ,7 00 0 23 ,7 00 n ot e: ir a � in di vi du al re tir em en t ac co un t; t c ja � t ax c ut an d jo bs a ct .s el ec te d ye ar sa m pl e ca lc ul at io ns fo r on e ru n fo r th e pr et c ja sc en ar io fo r re al co ns um pt io n. 361t. manuel et al. / financial services review 27 (2018) 345-365 a pp en di x t ab le 2 si m ul at io n re su lts an d sa m pl e ca lc ul at io ns se le ct ed ye ar s po st -t c ja e nt er nu m be r of tr ia ls (m ax � 1, 00 0) : 50 0 l if et im e re al co ns um pt io n a ge st ar t co nt ri bu tio n a ge m ed iu m ri sk a ge lo w ri sk a ge re tir e l if e ex pe ct an cy t ra di tio na l r ot h r un si m ul at io n 30 50 60 65 90 m in im um 4, 37 9, 63 1 4, 30 8, 33 3 st ar tin g in co m e in co m e gr ow th r at e of in fla tio n c on tr ib ut io n m ax im um 7, 76 7, 85 0 8, 30 1, 18 3 50 0 m ed ia n 5, 04 9, 17 0 5, 01 5, 55 1 $8 0, 00 0 6. 00 % 3. 00 % $1 0, 00 0 a ve ra ge 5, 14 8, 77 1 5, 13 4, 02 9 d ed uc tio n ad ju st m en t 10 0% b et te r 36 1 13 9 c ha ng e in ta x ra te af te r re t 0. 00 % 3. 00 % b ra ck et ad ju st m en t 3. 00 % w or ki ng r et ir ed l if et im e c um ul at iv e sa vi ng s 2, 70 1, 44 9 a ge st ar t co nt ri bu tio n a ge m ed iu m ri sk a ge lo w ri sk a ge re tir e l if e ex pe ct an cy in co m e ye ar s 35 25 60 n om in al an nu ity 21 1, 58 7 in co m e 9, 52 9, 66 9 5, 28 9, 66 6 14 ,8 19 ,3 35 l if et im e ta x ra te r ea l an nu ity 75 ,1 94 30 50 60 65 90 t ax es , t ra di tio na l ir a 1, 12 9, 43 6 26 2, 25 7 1, 39 1, 69 2 14 .6 0% r is k a ve ra ge re tu rn st an da rd de vi at io n st ar tin g in co m e in co m e gr ow th r at e of in fla tio n c on tr ib ut io n � a ge 50 st an da rd de du ct io n an d ex em pt io n t ax es , r ot h ir a 1, 26 2, 64 6 0 1, 26 2, 64 6 13 .2 5% h ig h 8. 22 % 13 .8 7% r ea l co ns um pt io n, t ra di tio na l ir a 4, 05 5, 19 1 1, 24 0, 65 1 5, 29 5, 84 2 b et te r m ed iu m 7. 74 % 12 .0 9% $8 0, 00 0 6. 00 % 3. 00 % $1 0, 00 0 $2 4, 00 0 r ea l co ns um pt io n, r ot h ir a 3, 98 3, 89 3 1, 30 9, 37 1 5, 29 3, 26 4 l ow 6. 01 % 7. 06 % g ro w th in ta x ra te af te r re tir em en t 0. 00 % 3. 00 % b ra ck et ad ju st m en t 3. 00 % in co m e, co nt ri bu tio ns , ex em pt io ns t ra di tio na l ir a r ot h ir a in ve st m en t/s av in gs a ge in co m e d efl at or in de xe d ir a co nt ri bu tio ns in de xe d de du ct io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n a ve ra ge re tu rn st an da rd de vi at io n c um ul at iv e sa vi ng s 30 80 ,0 00 1. 00 00 0 10 ,0 00 24 ,0 00 46 ,0 00 12 % 5, 13 9 64 ,8 61 56 ,0 00 12 % 6, 33 9 63 ,6 61 10 ,0 00 31 84 ,8 00 1. 03 00 0 10 ,3 00 24 ,7 20 49 ,7 80 12 % 5, 58 1 66 ,9 11 60 ,0 80 12 % 6, 81 7 65 ,7 11 8. 21 9% 13 .8 66 % 21 ,4 01 32 89 ,8 88 1. 06 09 0 10 ,6 09 25 ,4 62 53 ,8 17 12 % 6, 05 4 69 ,0 22 64 ,4 26 12 % 7, 32 7 67 ,8 22 8. 21 9% 13 .8 66 % 36 ,5 05 33 95 ,2 81 1. 09 27 3 10 ,9 27 26 ,2 25 58 ,1 29 12 % 6, 55 9 71 ,1 93 69 ,0 56 12 % 7, 87 0 69 ,9 93 8. 21 9% 13 .8 66 % 42 ,5 22 34 10 0, 99 8 1. 12 55 1 11 ,2 55 27 ,0 12 62 ,7 31 12 % 7, 09 9 73 ,4 28 73 ,9 86 12 % 8, 44 9 72 ,2 28 8. 21 9% 13 .8 66 % 62 ,6 01 35 10 7, 05 8 1. 15 92 7 11 ,5 93 27 ,8 23 67 ,6 43 12 % 7, 67 5 75 ,7 28 79 ,2 35 12 % 9, 06 7 74 ,5 28 8. 21 9% 13 .8 66 % 84 ,4 92 48 22 8, 34 7 1. 70 24 3 17 ,0 24 40 ,8 58 17 0, 46 4 22 % 23 ,6 77 11 0, 22 2 18 7, 48 9 22 % 27 ,4 22 10 8, 02 2 8. 21 9% 13 .8 66 % 58 6, 72 2 49 24 2, 04 8 1. 75 35 1 17 ,5 35 42 ,0 84 18 2, 42 9 22 % 25 ,8 94 11 3, 27 0 19 9, 96 4 22 % 29 ,7 52 11 1, 07 0 8. 21 9% 13 .8 66 % 66 8, 70 2 50 25 6, 57 1 1. 80 61 1 19 ,0 61 43 ,3 47 19 4, 16 3 22 % 28 ,0 48 11 5, 97 4 21 3, 22 4 22 % 32 ,2 42 11 3, 65 2 7. 74 2% 12 .0 89 % 80 7, 05 3 (c on ti nu ed on ne xt pa ge ) 362 t. manuel et al. / financial services review 27 (2018) 345-365 a pp en di x t ab le 2 (c on tin ue d) in co m e, co nt ri bu tio ns , ex em pt io ns t ra di tio na l ir a r ot h ir a in ve st m en t/s av in gs a ge in co m e d efl at or in de xe d ir a co nt ri bu tio ns in de xe d de du ct io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n a ve ra ge re tu rn st an da rd de vi at io n c um ul at iv e sa vi ng s 51 27 1, 96 5 1. 86 02 9 19 ,6 03 44 ,6 47 20 7, 71 5 22 % 30 ,5 90 11 9, 21 4 22 7, 31 8 22 % 34 ,9 03 11 6, 89 5 7. 74 2% 12 .0 89 % 76 2, 89 5 56 36 3, 95 1 2. 15 65 9 22 ,5 66 51 ,7 58 28 9, 62 7 22 % 46 ,2 04 13 6, 87 4 31 2, 19 2 22 % 51 ,1 69 13 4, 57 2 7. 74 2% 12 .0 89 % 1, 45 5, 91 3 57 38 5, 78 8 2. 22 12 9 23 ,2 13 53 ,3 11 30 9, 26 4 22 % 49 ,9 99 14 0, 71 8 33 2, 47 7 22 % 55 ,1 06 13 8, 41 9 7. 74 2% 12 .0 89 % 1, 51 3, 68 8 58 40 8, 93 5 2. 28 79 3 23 ,8 79 54 ,9 10 33 0, 14 5 22 % 54 ,0 52 14 4, 67 4 35 4, 02 5 22 % 59 ,3 05 14 2, 37 8 7. 74 2% 12 .0 89 % 1, 67 6, 05 8 59 43 3, 47 1 2. 35 65 7 24 ,5 66 56 ,5 58 35 2, 34 8 22 % 58 ,3 79 14 8, 74 5 37 6, 91 3 22 % 63 ,7 83 14 6, 45 1 7. 74 2% 12 .0 89 % 1, 87 9, 84 9 60 45 9, 47 9 2. 42 72 6 25 ,2 73 58 ,2 54 37 5, 95 2 22 % 62 ,9 98 15 2, 93 3 40 1, 22 5 24 % 68 ,5 72 15 0, 63 7 6. 01 3% 7. 06 1% 1, 79 9, 64 9 61 48 7, 04 8 2. 50 00 8 26 ,0 01 60 ,0 02 40 1, 04 5 22 % 67 ,9 27 15 7, 24 3 42 7, 04 6 24 % 73 ,9 38 15 4, 83 9 6. 01 3% 7. 06 1% 2, 01 9, 08 0 62 51 6, 27 1 2. 57 50 8 26 ,7 51 61 ,8 02 42 7, 71 8 24 % 73 ,2 42 16 1, 65 6 45 4, 46 9 24 % 79 ,6 63 15 9, 16 3 6. 01 3% 7. 06 1% 2, 26 5, 82 0 63 54 7, 24 7 2. 65 23 4 27 ,5 23 63 ,6 56 45 6, 06 8 24 % 79 ,1 64 16 6, 10 3 48 3, 59 1 24 % 85 ,7 70 16 3, 61 2 6. 01 3% 7. 06 1% 2, 24 1, 13 5 64 58 0, 08 2 2. 73 19 1 28 ,3 19 65 ,5 66 48 6, 19 7 24 % 85 ,4 86 17 0, 67 8 51 4, 51 6 24 % 92 ,2 83 16 8, 19 0 6. 01 3% 7. 06 1% 2, 36 5, 91 8 65 61 4, 88 7 2. 81 38 6 29 ,1 39 74 ,8 49 51 0, 90 0 24 % 90 ,4 79 17 6, 01 1 54 0, 03 8 24 % 97 ,4 72 17 3, 52 5 6. 01 3% 7. 06 1% 2, 70 1, 44 9 w or ki ng 9, 52 9, 66 9 64 8, 75 9 1, 52 5, 93 9 7, 35 4, 97 1 19 % 1, 12 9, 43 6 4, 05 5, 19 1 8, 00 3, 73 1 20 % 1, 26 2, 64 6 3, 98 3, 89 3 2, 70 1, 44 9 r et ir em en t t ra di tio na l ir a r ot h ir a a ge n om in al w ith dr aw d efl at or in de xe d de du ct io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n t ax ab le in co m e m ar gi na l ta x ra te t ax es r ea l co ns um pt io n 66 21 1, 58 7 2. 89 82 8 77 ,0 94 13 4, 49 2 12 % 15 ,0 35 67 ,8 17 0 73 ,0 04 67 21 1, 58 7 2. 98 52 3 79 ,4 07 13 2, 18 0 12 % 14 ,7 24 65 ,9 46 0 70 ,8 78 68 21 1, 58 7 3. 07 47 8 81 ,7 89 12 9, 79 7 12 % 14 ,4 04 64 ,1 29 0 68 ,8 14 69 21 1, 58 7 3. 16 70 3 84 ,2 43 12 7, 34 4 12 % 14 ,0 75 62 ,3 65 0 66 ,8 09 70 21 1, 58 7 3. 26 20 4 86 ,7 70 12 4, 81 6 12 % 13 ,7 35 60 ,6 53 0 64 ,8 63 71 21 1, 58 7 3. 35 99 0 89 ,3 73 12 2, 21 3 12 % 13 ,3 85 58 ,9 90 0 62 ,9 74 72 21 1, 58 7 3. 46 07 0 92 ,0 55 11 9, 53 2 12 % 13 ,0 25 57 ,3 76 0 61 ,1 40 73 21 1, 58 7 3. 56 45 2 94 ,8 16 11 6, 77 0 12 % 12 ,6 54 55 ,8 09 0 59 ,3 59 84 21 1, 58 7 4. 93 41 2 13 1, 24 8 80 ,3 39 10 % 8, 03 4 41 ,2 54 0 42 ,8 82 85 21 1, 58 7 5. 08 21 5 13 5, 18 5 76 ,4 01 10 % 7, 64 0 40 ,1 30 0 41 ,6 33 86 21 1, 58 7 5. 23 46 1 13 9, 24 1 72 ,3 46 10 % 7, 23 5 39 ,0 39 0 40 ,4 21 87 21 1, 58 7 5. 39 16 5 14 3, 41 8 68 ,1 69 10 % 6, 81 7 37 ,9 79 0 39 ,2 43 88 21 1, 58 7 5. 55 34 0 14 7, 72 0 63 ,8 66 10 % 6, 38 7 36 ,9 50 0 38 ,1 00 89 21 1, 58 7 5. 72 00 0 15 2, 15 2 59 ,4 35 10 % 5, 94 3 35 ,9 52 0 36 ,9 91 90 21 1, 58 7 5. 89 16 0 15 6, 71 7 54 ,8 70 10 % 5, 48 7 34 ,9 82 0 35 ,9 13 n ot e: ir a � in di vi du al re tir em en t ac co un t; t c ja � t ax c ut an d jo bs a ct . sa m pl e ca lc ul at io ns fo r on e ru n of th e po st -t c ja si m ul at io n. 363t. manuel et al. / financial services review 27 (2018) 345-365 shorter time to retirement favors the traditional ira in terms of lifetime consumption. consistent with the literature, a lower tax rate during retirement than in the working years favors the traditional ira, but this situation does not always hold. at longer times to retirement, larger investment amounts, greater income and higher investment returns the roth ira performs increasingly well even if tax rates fall during retirement. at lower amounts saved the traditional ira can be better even with higher tax rates in retirement. these results are consistent with most of the prior literature and help validate the model. the model is not a complete, or even a nearly complete, depiction of which type of ira is preferable for all investors. the model is not for those who wish to maximize retirement consumption income rather than overall lifetime income. the spreadsheet model only applies for investors that are married, filing jointly, and are not covered by a plan at work. the model does not allow contributions to continue after retirement, nor ensure that rmds are made. it also does not consider estate planning, other income and the impact on taxation of social security benefits. however, the model indicates that from the perspective of lifetime consumption, the roth ira is not necessarily always the best choice. simulating a large variety of returns using actual market data also gives a richer understanding than assuming a fixed return with no risk. in appendix table 1 above taxable income � income – standard deduction – ira contributions for the traditional ira and income – standard deduction for the roth ira, respectively, for one run of the simulation. the taxes used are from the 2017 progressive tax table for married filing jointly (see www.irs.gov) with the tax rules in place before the tcja. in the contribution years real consumption � income – ira contributions – taxes, adjusted for inflation with the deflator column. annual rates of return are drawn from a normal distribution with the given geometric average return and standard deviation. the cumulative savings at retirement is annuitized over the given life expectancy. the nominal annual withdrawal during the retirement years is a constant amount as shown and real consumption declines each year because of inflation. the low amount of taxation during the retirement years for the traditional ira is an artifact of indexing the tax brackets and the standard deduction to match inflation. if the brackets or the deductions are not indexed the roth ira pre-dominates in most scenarios. the choice of ira is thus very dependent on whether the irs will continue to index deductions and tax brackets. the consumption outcomes at the top of the table are for one trial run. providing a simulation of multiple outputs rather than a single point estimate of future value is a better method to determine the preferred choice. note that the model uses realistic estimates of portfolio returns at the chosen risk levels. this provides investors better information to choose between the ira types. 2. post-tcja scenario in the following scenario the same inputs were used, but the new tcja tax brackets and standard deduction were applied. rates of return are randomly drawn so they are different from the pre-tcja scenario. nevertheless, the traditional ira still pre-dominates in terms of real consumption in 361 out of the 500 trials. see appendix table 2. 364 t. manuel et al. / financial services review 27 (2018) 345-365 references adelman, s. w., & cross, m. l. (2010). comparing a traditional ira and a roth ira. risk management and insurance review, 13, 265–277. al zaman, a. (2008). reconsidering ira and roth ira. the journal of wealth management, 11, 82–92. baxendale, s. j., & coppage, r. e. (2014). in-plan roth rollovers under the atra: a new opportunity for younger taxpayers. the cpa journal, 84, 62–63. beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2017). does front-loading taxation increase savings? evidence from roth 401(k) introductions. journal of public economics, 151, 84–95. board of governors of the federal reserve system. (2017). report on the economic well-being of u.s. households in 2016 (available at https://www.federalreserve.gov/publications/file/2016-report-economicwell-being-us-households-201705.pdf). bodie, z., kane, a., & marcus, a. j. (2008). investments (8th ed.). new york, ny: mcgraw-hill. cook, k. a., meyer, w., & reichenstein, w. (2015). tax-efficient withdrawal strategies. financial analysists journal, 71, 16–29. damodaran, a. (2018). damodaran online data (available at http://pages.stern.nyu.edu/�adamodar/ new_home_page/datafile/histretsp.html). grossmann, a., & rose, c. c. (2012). comparing the roth ira to the traditional ira: an after-tax cash flow analysis. journal of financial service professionals, 66, 55–61. harline, n. l. v. (2014). maximizing after-tax family wealth over multiple generations by using inherited 401(k) plans and inherited iras. the journal of pension planning & compliance, 40, 1–25. horan, s. m. (2002). after-tax valuation of tax-sheltered assets. financial services review, 11, 253–275. horan, s. m. (2003). choosing between tax-advantaged savings accounts: a reconciliation of standardized pretax and after-tax frameworks. financial services review, 12, 339–357. horan, s. m. (2006). optimal withdrawal strategies for retirees with multiple savings accounts. journal of financial planning, 19, 62–75. horan, s. m., & al zaman, a. (2009). iras under progressive tax regimes and income growth. financial services review, 18, 195–211. reichenstein, w. (2006). after-tax asset allocation. financial analysts journal, 62, 14–19. reichenstein, w., horan, s. m., & jennings, w. w. (2015). two key concepts for wealth management and beyond. financial analysts’ journal, 71, 70–77. reichert, c. j. (2011). iras. journal of accountancy, 212, 68–69. saunders, l. (2018, aug 18). tax report: when to ignore the crowd and shun a roth ira – wsj. dow jones institutional news (available at https://search.proquest.com/docview/2090610978). sibley, m. (2002). on the valuation of tax-advantaged retirement accounts. financial services review, 11, 233–251. smith, h., finke, m., & huston, s. (2012). the influence of financial sophistication and financial planners on roth ira ownership. journal of financial service professionals, 66, 69–81. vanzante, n. r., & fritzsch, r. b. (2013). the benefits of roth accounts: considering tax rates before and after retirement. the cpa journal, 83, 54. 365t. manuel et al. / financial services review 27 (2018) 345-365 financial literacy and its impact on the credit card debt puzzle laura c. ricaldia*, terrance k. martina, sandra j. hustonb adepartment of finance and economics, utah valley university, 800 west university parkway, ms 280, orem, ut 84058, usa bschool of financial planning, texas tech university, 1301 akron avenue, ms 41210, lubbock, tx 79406, usa abstract the credit card debt puzzle is not well understood. households exhibit inefficient behavior when they have sufficient liquid assets to pay off their credit card balance, but do not. based multinomial regression analyses of the 2016 survey of consumer finances, the study discovered that households that display this behavior are more likely to have lower financial literacy than convenience users. the findings suggest financially literate households are less likely to display irrational behavior regarding the credit card debt puzzle. © 2022 academy of financial services. all rights reserved. jel classifications: g530; j24; g510 keywords: behavioral life-cycle; credit card debt puzzle; financial literacy; 2016 survey of consumer finances 1. introduction household credit card use is widespread. comparing all debt across households, credit card debt is the debt type most extensively used by households (bricker et al., 2017). over 70% of u.s. households have a credit card (bucks, kennickell, mach, & moore, 2009). there are two main groups of credit card users: those users who pay off their balance at the end of each month, commonly called convenience users, and those who carry a balance from month to month. these households that carry a balance from month to month are known as revolving credit card users (kim & devaney, 2001). according to the 2016 *corresponding author. tel.: +1-806-787-0214. e-mail address: laura.ricaldi@uvu.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 107–124 survey of consumer finances, 44% of families carry a balance on their credit cards. since 2013, the median and mean balance both decreased by 3% balance (bricker et al., 2017). within the revolving credit card users, there is a group of users, solvent revolvers, who have enough liquid assets to pay the balance of their credit cards but choose not to pay it off. bi (2005) finds that 58% of revolving credit card users have liquid assets (liquid assets include monetary assets including checking, saving, and money market accounts and call accounts) totaling more than their credit card balance. credit cards typically charge high interest rates on a revolving balance and liquid accounts (like checking and money market accounts) accrue low (if any) after-tax interest. based on the interest rates, it is inefficient for a household to maintain a revolving balance. previous studies that have investigated this irrational behavior call this the credit card debt puzzle or the co-holding puzzle (bertaut, haliassos, & reiter, 2009; gathergood & weber, 2014; haliassos & reiter, 2005; laibson, repetto, & tobacman, 2001). many studies research the credit card debt puzzle; however, few combine human capital theory and behavioral life cycle theory. the purpose of this paper is to study the effect of financial literacy on credit card debt treatment. it is hypothesized that households that have higher financial literacy will be less likely to display this puzzling behavior. 2. literature review literature suggests a planner or doer framework is often a way to explain irrational behavior. the behavioral life cycle theory puts forward that a household is composed of two dueling selves, the planner and the doer. the planner is the forward-thinking, rational self while the doer is myopic and focused on current consumption (shefrin & thaler, 1988). the household makes decisions to satisfy both the planner and doer and may display conflicting and inefficient behaviors, such as the credit card debt puzzle. several studies attempted to explain the credit card debt puzzle, or solvent revolving credit card use, with several distinct factors such as financial human capital, precautionary savings motives, and self-control. gross and souleles (2002) find that revolving high credit card balances while simultaneously holding liquid assets stem from behavioral explanations, not lack of liquidity. 2.1. human capital the planner/doer model suggests the planner will reduce consumption in the current period by exerting a level of willpower. because willpower is costly, the planner will resort to other techniques to reduce the consumption of the doer. some of these techniques include mental accounting and rule-setting (shefrin & thaler, 1988). in the planner/doer model, the level of human capital can impact financial decisions. becker (1964) describes human capital as an individual’s stock of knowledge, health, skills, or values. it is a function of goods, services, time, and the individual’s current stock of human capital. human capital is often improved through learning, maturity, and experiences. in the realm of finance, individuals can improve their human capital by taking financial courses to improve their ability to understand and 108 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 make effective financial decisions. individuals can also improve their financial human capital through experiences like using credit cards or taking out a home mortgage. households with an elevated level of financial knowledge and experience are considered financially literate or financially sophisticated; these households can make more effective financial decisions than households with a lower level of financial literacy or financial sophistication. financial literacy gives the household the potential to improve their ability to make better financial decisions. the households with a higher level of financial sophistication tend to be aware of the consequences of their decisions. bertaut et al. (2009) suggests that financially sophisticated households would be convenience users of credit cards as well as benefit from floating and other advantages of credit cards. although, some financially sophisticated households display characteristics that are not sophisticated. for example, haliassos and reiter (2005) find that in the shopper/accountant model, the shopper is not fully financially sophisticated. the accountant/shopper model is like the planner/doer model. 2.2. mental accounting and precautionary savings motives to control the doer, the planner creates mental accounts to reduce the temptation to spend from them. households divide assets, expenditures, and income into distinct categories or mental accounts. an economist would state that these accounts are substitutable, but they are not (thaler, 1999). the household views the mental accounts, either assets or expenses, as different things, and marginal propensity to consume from these accounts are different. for example, households save money in an emergency fund account to prepare for an uncertain event. the household uses framing to earmark these accounts for difference purposes. households that save in liquid accounts for emergencies or unexpected events do not believe the assets are substitutable for other assets or expenses. when households use mental accounting for expenses, especially with credit cards, they decouple the payment from the consumption. once the bill is received, the purchase is mixed with other purchases. thaler (1999) states that it is hard for the consumer to attribute the balance to any purchase: therefore, the consumer carries a balance from month to month. because households have mental accounts, they will not view accounts used for savings as available to pay off credit card balances since these accounts are not substitutable. uncertainty and precautionary savings motives play a role in mental accounting. telyukova and wright (2008) propose that households stay solvent revolvers to maintain sufficient liquid assets for uncertain future events. bi and hanna (2006), druedahl and jørgensen (2018), and gorbachev and luengo-prado (2019) find that households will display the credit card debt puzzle when precautionary savings motives are present. 2.3. self-control although the planner creates mental accounts to control the doer, the individual must exhibit self-control. an individual displays self-control issues by either postponing action (e.g., procrastination) or by consuming immediately (e.g., no willpower to wait). households that display self-control issues either consume all their resources without saving l. c. ricaldi et al. / financial services review 30 (2022) 107–124 109 or paying debt or putting off making critical decisions. one study shows that in an accountant/shopper household, the accountant would choose not to pay off the credit card balance to impose control over the shopper (bertaut et al., 2009). by reducing the available limit on the credit card, the shopper is unable to consume more. gathergood and weber (2014) show that the likelihood of displaying the credit card debt puzzle, or co-holding, increases with self-assessed impulsiveness. other studies have identified other behavioral factors that affect solvent revolving credit card use. credit attitude and bankruptcy history are often considered when households exhibit credit card debt puzzle. first, chien and devaney (2001) find that a positive credit attitude was related to a higher credit card balance. rutherford and devaney (2009) find that those households that had a positive attitude toward credit are less likely to be convenience users. 3. theory and conceptual framework other studies evaluate how the credit card debt puzzle, and the relevant behavioral factors are related, however few focus on the intersection of financial knowledge and behavior. this study uses a combination of the behavioral life-cycle hypothesis and human capital theory to focus on how behavioral factors and human capital influence the likelihood of being a solvent credit card revolver. like the life-cycle hypothesis, the behavioral life-cycle hypothesis (shefrin & thaler, 1988) puts forward that to maximize utility, a household will shift resources in periods where the marginal utility of consumption is low to periods where the marginal utility of consumption is high. a good example of this is when households save during the working years for consumption during retirement years. unlike the traditional life-cycle hypothesis, the behavioral life-cycle hypothesis posits that households have a dual preference framework where they are both planners (long-term) and doers (short-term). the planner preference is when households make rational decisions regarding when to shift resources to maximize utility. the planner focuses on long-term decisions to maximize utility. the doer preference, or short-term preference, is when households succumb to temptation to consume in the current period. the three behavioral factors, self-control, mental accounting, and framing are what make the behavioral life-cycle hypothesis different from other life-cycle models. self-control refers to the household’s temptation to make immediate consumption decisions, rather than saving for future consumption. for the doer, immediate consumption is always a tempting alternative to future consumption. there is discomfort for the doer associated with postponing current consumption; therefore, the planner will enforce saving devices and rules of thumb to deal with self-control issues for various situations. these are types of external rules that households use to plan for future consumption. households also use internal rules, like refusing to borrow for current consumption, to maintain self-control. mental accounting refers to placing wealth into different non-substitutable accounts. the typical breakdown of mental accounts is current income, current assets, and future income (shefrin & thaler, 1988). households use mental accounting to restrict the doer from bringing future resources into the current period. the way a household frames the different mental accounts determines the temptation to spend from each account. each account has a different 110 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 level of temptation associated with spending from it. the marginal propensity to consume from the current income account is much higher than the marginal propensity to consume from the future income account. temptation plays an important role in the household’s decision to spend or save. by carrying a credit card balance, the doer is bringing consumption into the current period. the households that are solvent (i.e., households who have enough liquid assets to pay off their balance but do not) are not displaying the planner behavior but are displaying the doer behavior. in contrast, convenience users are keeping future resources in future periods by paying off the balance while still benefiting from the advantages of using a credit card. in addition to the behavioral life-cycle hypothesis, human capital theory also plays a part in solvent revolving credit card use. human capital is an individual’s knowledge, health, skills, or values and is often described as a function of goods, services, time, and the individual’s current stock of human capital. a household’s level of human capital impacts its ability to make efficient financial decisions. households with a higher level of financial human capital (i.e., financial sophistication) have the potential to improve the ability to make effective and efficient financial decisions. 4. hypotheses based on the theoretical framework, the concepts developed for this paper include human capital, mental accounting/precautionary savings motives, self-control factors, and other lifecycle control factors. the concepts serve as control factors to explain why households display puzzling behavior that is inefficient. the hypotheses are as follows: h01: higher levels of financial literacy will reduce the likelihood of revolving credit card debt even when financially solvent. h02: mental accounting behaviors will increase the likelihood of revolving credit card debt even when financially solvent. h03: self-control issues will increase the likelihood of revolving credit card debt even when financially solvent. 5. method 5.1. data and sample the data used were from the 2016 survey of consumer finances (scf), a triennial survey, which is sponsored by the federal reserve board and collected by the national organization for research at the university of chicago (board of governors of the federal reserve system, 1998-2013). the scf collects detailed information on the finances of u.s. households. the 2016 scf included 6,248 households in the public data set. the 2016 scf contains five implicates to deal with missing data. the total number of observations with all l. c. ricaldi et al. / financial services review 30 (2022) 107–124 111 five implicates is 31,240 observations. for this study, we only used the first implicate and because this study only analyzes those households that have a credit card, the final sample is limited to 4,725 observations. we created an additional subsample (n-2360) by censoring the data to only solvent revolvers and convenience users to assess credit use decisions among the most financially literate respondents (answered all financial literacy questions correctly). 5.2. dependent variables the dependent variable is constructed by categorizing credit card users into one of three categories. first, solvent revolvers are credit card users have liquid assets greater than or equal to the balance still owed on their main credit card after the last payment was made to the account. the total liquid assets are derived from the federal reserve board definition in the net worth code. from this definition, liquid assets include money market accounts, checking accounts, savings accounts, call accounts, and prepaid cards. next, insolvent revolvers are credit card users have liquid assets less than the balance still owed on their main credit card after the last payment is made to the account. last, convenience credit card users do not have an outstanding balance on their main credit card. to answer our research questions, we create an unranked three level categorical variable of credit card user, and dichotomous variables to for each credit card user type. 5.3. independent variables by concept based on the behavioral life-cycle hypothesis and human capital theory, four concepts were identified: human capital/financial literacy, mental accounting/precautionary savings motives, self-control factors, and other lifecycle control factors. independent variables operationalized these concepts. the human capital concept explains the household’s potential to make effective decisions. households that do not have a strong base in financial human capital (i.e., financially literate) tend to make suboptimal financial decisions (bertaut et al., 2009). because the purpose of this study is to evaluate the impact of financial literacy on solvent revolvers, human capital is the focus of the study. the human capital concept was measured by two independent variables: financial literacy and education. financial literacy represents a type of human capital specific to personal finance while education represents a general type of human capital. the variables that make up this concept moderate the financial decisions made by the households to not pay off their credit card balance even when the household has the financial liquidity to do so. our primary independent variable is this concept is financial literacy. respondents who participated in the 2016 scf were asked three financial literacy questions. the questions stated in the survey include: 1. suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow? 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account? 112 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 3. buying a single company’s stock usually provides a safer return than a stock mutual fund. in prior studies, researchers like lusardi and scheresberg (2013) use these same questions to construct their proxy for financial literacy. if a respondent answered all the above questions correctly, they are classified as financially literate and coded 1. if the respondent did not answer all questions correctly then they are coded 0. the second variable included in the human capital concept is the level of education for the head of household. the level of education was categorized as noncollege degree (including less than high school, high school degree, and some college) and college degree. the mental accounting concept relates to how the planner controls the doer’s consumption by using various mental accounts. past literature shows that households maintain solvent revolving due to mental accounting and the framing of the different accounts. with the behavioral life-cycle hypothesis, households do not view mental accounts as substitutable. households do not tend to use emergency savings accounts to pay off credit card balances if they are not in an emergency situation. the variables that make up this concept represent why the household will maintain liquid assets more than their credit card balance and maintain their status as a solvent revolver. mental accounting is measured by six independent variables. the first variable, having an emergency fund. if the household indicated they have savings for a subjective emergency fund, they are coded as 1 and 0 otherwise. the second variable, saving for unemployment, is constructed by combining two variables: if the household stated they had a savings motive for unemployment and if they expect their future income will decrease in comparison with prices in the next year. the variable is coded as 1 if the household had a motive to save for unemployment and 0 otherwise. the third variable, saving for illness, was constructed by combining two variables: if the household stated they had a savings motive for in case of illness or future medical expenses and if they have a poor health status. the variable is coded as 1 if the household had a motive to save for illness and 0 otherwise. the fourth variable is the household’s ability to borrow $3,000 from friends or relatives in an emergency. the variable is coded as 1 if the household was able to borrow and 0 if the household was not able to borrow from friends or relatives. the fifth variable is if either the head of household or the spouse is self-employed. the variable is coded as 1 if self-employed and 0 if not self-employed. the last variable is the proxies whether a household owns liquid accounts. if a household report owning a liquid account such as a checking account, we code that as a 1 or 0 otherwise. the self-control concept is included since households have limited time to make financial decisions. households display self-control issues regarding financial decision making. selfcontrol plays a role in the household’s susceptibility to give in to temptation to spend/consume during the current period. the self-control concept is comprised of past payment history, bankruptcy history, credit attitude, likelihood to increase spending with increased asset value, and unwillingness to decrease spending with a decrease in asset value. the first variable is the past payment history of all loans, mortgages and credit cards made during the past year. the variable is coded as 1 for those households who made payments on schedule or had no payments and 0 for those who were behind or missed payments. past literature shows that a household maintains solvent revolving due to credit attitude and bankruptcy history. the household’s credit attitude is measured by their feelings about using credit. the l. c. ricaldi et al. / financial services review 30 (2022) 107–124 113 variable is coded as positive if the household feels credit is a good idea, ambivalent if the household feels credit is good in some ways and bad in others, and negative if the household feels credit, is a bad idea. the ambivalent credit attitude group is the reference group for the regression analyses. last, bankruptcy history is coded 1 if the household has ever filed for bankruptcy, or 0 if they have never filed for bankruptcy. as a further proxy for the self-control concept, include two variables that represent spending and consumption when faced with positive or negative changes in income or financial assets. the variables, likelihood to increase spending if assets increase and unwillingness to reduce spending if assets decrease, are operationalized as dichotomous variables and coded as 1 if they report doing the action and 0 if otherwise. the lifecycle factors concept is included since households make decisions based on being in various stages of the lifecycle. the variables that make up this concept represent why the household will maintain liquid assets more than their credit card balance and maintain their status as a solvent revolver. the concept is measured using age, race, gender, marital status, income, and net worth. first, age is coded categorically as under 35, between 35 and 55, and over 55. race is separated into four categories, black, white, hispanic, and other. next, gender and marital status are included in the analysis. gender is coded as male or female and male is the reference category for the regression analyses. marital status is coded as married or not married. the married category is the reference group. next, household income is a continuous variable. income is logged to see the magnitude and its effect in each of the regressions. last, net worth is a continuous variable. household net worth is logged to see the magnitude and its effect in each of the regressions. 5.4. empirical models to answer the research questions, we use a combination of multinomial logistic regressions (mlr) and binary logistic regressions (blr). mlr is used due to the structure of the dependent variable. recall, the dependent variable is constructed by categorizing credit card users into one of three unranked categories: convenience users, solvent, and insolvent revolvers. we present four mlr models, specification a–d. in specification a, we regress the human capital concept on the dependent variable. in specification b, we include mental accounting to the previous model. we then add the self-control and estimate specification c. in the final mlr model, we include life cycle factors to all previous three concepts. the final model uses a restricted sample that includes only convenience users and solvent revolvers who are financially literate; therefore, a blr is used. the model seeks to explain why households remain solvent revolvers when they have the financial resources and financial sophistication to be convenience users. 5.5. analysis of data descriptive statistics were conducted to look at the characteristics of households. to generalize the findings back to the u.s. population, the descriptive statistics are weighted using a weight variable provided by the federal reserve (lindamood, hanna, & bi, 2007). since 114 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 the dependent variables are unranked and categorical multinomial and binary logistic regressions are used to estimate the likelihood of the dependent variables occurring given the set of independent variables. the regression analyses are not weighted (lindamood et al., 2007). 6. results 6.1. descriptive statistics since the descriptive statistics are weighted, the reported percentages, means, and standard errors represent all u.s. households. fig. 1 shows the distribution of credit user groups in the full sample. convenience users account for 46% of the sample, while insolvent and solvent revolvers make up 30% and 24%, respectively. fig. 2 presents a pie chart that shows the distribution of the sample by the number of financial literacy questions that were answered correctly. from the pie chart only 50% of the respondents in our sample were able to answer all three questions correctly. for our research, we consider respondents that answered all questions correctly as financially literate. fig. 3 separates the credit user group by financially literate or not financially literate. financially literate respondents are more likely to be convenience users and less likely to be solvent revolvers. see table 1a for frequency distributions and table 1b for mean and median of the full sample and by type of credit ser for other independent variables included in our models. 6.2. regression table 2 provides multinomial logistic regression results for four specifications of our empirical model. recall that the dependent variable is credit card user type with three levels: solvent revolver, insolvent revolver, and convenience user. convenience user is the base fig. 1. shows distribution of sample by type of credit card user type. l. c. ricaldi et al. / financial services review 30 (2022) 107–124 115 reference for the mlr models discussed. all four concepts provide some value in explaining an individual’s approach to credit card use. from specification a, we see the impact of only human capital on the estimation of the likelihood of choosing between the three types of credit card use. in comparison to convenience users, financially literate individuals 46% less likely to be solvent revolvers and 54% less likely to be in insolvent revolvers. similarly, college graduates with at least a bachelor’s degree are 58% less likely to be either solvent or insolvent revolvers. in specification b, we consider mental accounting factors in our estimation. financially literate individuals remain less likely than not financial literate individuals to be solvent and insolvent revolvers of credit card debt, 42% and 39%, respectively. turning to mental accounting factors, we see statistical significance from five of our six factors. respondents with an emergency fund, are 45% less likely to be a solvent revolver and 80% less likely to be an insolvent revolver when compared with a convenience use. we observe a similar trend among the self-employed respondents as they are 32% less likely to be a solvent revolver and 52% less likely to be an insolvent revolver relative to convenience users who are not self-employed. mixed results are observed for those who own liquid accounts, can borrow, and save for unemployment. for example, among respondents with fig. 2. presents a pie chart that shows the distribution of the sample by the number of financial literacy questions that were correct. fig. 3. separates the credit user group by financially literate or not financially literate. 116 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 the ability to borrow, we estimate a 66% increase in the likelihood of being an insolvent revolver but no statistical difference in being a solvent revolver when compared with convenience user who could not borrow from a relative. as another example, liquidity plays an interesting role on the impact of credit card user type. owning a liquid account result in a 50% increase in the likelihood of being a solvent revolver, but a 36% decrease in the likelihood of being an insolvent revolver relative to convenience users that did not own liquid accounts. table 1a frequencies of independent variables (full sample and by credit user group) full sample (%) insolvent revolver solvent revolver convenience user human capital financially literate 50.00 42.00 43.00 59.00 college degree 43.00 30.39 35.23 54.43 mental accounting/precautionary savings motives have emergency fund 58.40 30.73 59.25 72.36 saving for unemployment 3.42 3.45 3.99 3.04 saving for illness 5.62 4.80 6.08 5.74 ability to borrow from friends/relatives 27.97 40.18 26.43 22.58 self employed 14.71 11.66 15.54 15.76 have liquid accounts 63.02 46.70 70.48 66.67 self-control payment history on time/no payment 89.25 79.04 87.02 96.07 behind 10.75 20.96 12.98 3.93 credit attitude positive 27.35 27.79 30.38 25.2 ambivalent 44.38 45.74 43.19 44.45 negative 28.24 26.48 26.43 30.35 bankruptcy history 12.29 22.07 15.44 5.08 spending attitude likely increase 21.76 26 21.57 19.67 unwilling to reduce 28.58 25.67 26.02 31.81 lifecycle factors age under 35 18.47 18.45 23.12 15.21 35 to 55 36.59 45.53 40.95 29.02 over 55 44.93 35.62 35.92 55.77 gender male 76.15 69.73 75.12 80.20 female 23.85 30.27 24.88 19.80 married 53.99 48.1 50.91 59.12 race white 76.51 70.55 69.25 84.43 black 11.93 16.02 18.05 5.74 hispanic 9.19 14.59 10.80 5.28 other 4.51 2.09 4.82 5.59 note. statistics derived from weighted analysis of one implicate. source: 2016 survey of consumer finances. l. c. ricaldi et al. / financial services review 30 (2022) 107–124 117 specification c highlights the impact of human capital, mental accounting, and self-control concepts. variables in the two previous concepts maintained their statistical significance with similar effect as in specification b. being behind on payments and a record of bankruptcy all increase the likelihood of being both a solvent and insolvent revolver when compared with convenience users with opposite characteristics. respondents with delinquent accounts are more likely to be 230% and 344% more likely to be solvent revolvers and insolvent revolvers respectively in comparison to respondents reporting no delinquencies. respondents that reported bankruptcy were 203% more likely to be a solvent revolver and 284% more likely to be an insolvent revolver compared with respondents who never filed for bankruptcy. households whose assets reduced in value and refused to decrease their spending were 21% and 22% less likely to be a solvent or insolvent revolvers when compared with convenience users whose assets did not decrease. in final mlr model estimation, we include all four concepts to estimate the impact on the dependent variables. by adding lifestyle factor variables, we observe notable changes in the effect of being financially literate, could borrow, being self-employed, and who were unwilling to reduce spending even after asset values declined. as an illustration, we comment on the impact of being financially literate. households who were able to answer all the financial questions correctly were now only 18% less likely to be a solvent revolver compared with convenience users who were not financially literate. moreover, there was no statistically significant difference in being an insolvent revolver compared with convenience users who were not financially literate. of the lifestyle factors included in our analysis, we observe the following results. in comparison to respondents 34 years and younger, older respondents, 55 and over, are 59%% more likely to be solvent and 78% more likely to be insolvent revolvers when being a convenience user is an option. women are more likely than men to be insolvent revolvers than convenience users. specifically, the results show that compared with male convenience users, women are 63% more likely to be insolvent revolvers. comparing unmarried households to married households who report being convenience users, unmarried respondents are 45% less likely to be insolvent revolvers. income’s impact on the credit card user type appears only marginally statistically significant for revolving insolvent debt. a one percentage increase in income decreases the likelihood of being a solvent revolver when table 1b mean and median statistics of independent variables (full sample and by credit user group) full sample insolvent revolver solvent revolver convenience user mean net worth $912,469.96 $157,114.26 $413,347.22 $1,638,477.82 log net worth 12.19 11.2428 11.7480 12.8805 household income $123,934.20 $65,830.40 $100,734.04 $169,748.06 log hh income 11.11 10.8058 11.1073 11.2755 median net worth $ 180,180.00 $60,500.00 $126,600.00 $396,700.00 log net worth 12.32 11.42 11.99 12.96 household income $67,000.00 $51,000.00 $68,000.00 $76,000.00 log hh income 11.11 10.84 11.13 11.24 note. statistics derived from weighted analysis of one implicate. source: 2016 survey of consumer finances. 118 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 table 2 result of multinomial logistic regression results on credit user type (convenience user is the base group) variable user type a p b p c p d p intercept sr 1.09 1.24 ** 0.95 34.74 *** intercept ir 0.92 2.16 *** 1.58 *** 955.85 *** human capital financially literate sr 0.54 *** 0.58 *** 0.60 *** 0.82 ** financially literate ir 0.46 *** 0.61 *** 0.64 *** 0.94 college degree sr 0.38 *** 0.42 *** 0.46 *** 0.64 *** college degree ir 0.30 *** 0.42 *** 0.46 *** 0.77 ** mental accounting/precautionary savings motives have emergency fund sr 0.55 *** 0.62 *** 0.81 ** have emergency fund ir 0.20 *** 0.24 *** 0.34 *** saving for unemployment sr 1.57 ** 1.42 0.84 saving for unemployment ir 1.63 * 1.40 0.78 saving for illness sr 0.93 0.97 1.07 saving for illness ir 0.80 0.87 0.80 ability to borrow sr 1.09 1.06 0.88 ability to borrow ir 1.66 *** 1.59 *** 1.16 self employed sr 0.68 *** 0.67 *** 1.32 ** self employed ir 0.48 *** 0.47 *** 1.11 have liquid accounts sr 1.50 *** 1.50 *** 1.42 *** have liquid accounts ir 0.64 *** 0.64 *** 0.58 *** self-control behind on payments sr 3.30 *** 2.19 *** behind on payments ir 4.44 *** 2.97 *** positive attitude toward credit sr 1.23 ** 1.18 * positive attitude toward credit ir 1.01 1.08 negative attitude toward credit sr 0.88 0.87 negative attitude toward credit ir 0.77 ** 0.75 ** bankruptcy history sr 3.03 *** 2.19 *** bankruptcy history ir 3.84 *** 2.70 *** likely to increase spending sr 0.99 1.04 likely to increase spending ir 1.11 1.18 unwilling to reduce spending sr 0.76 ** 0.87 unwilling to reduce spending ir 0.75 ** 0.79 ** lifecycle factors age 35–55 sr 1.77 *** age 35–55 ir 2.61 *** age 55+ sr 1.13 age 55+ ir 1.30 female sr 1.08 female ir 1.65 ** not married sr 0.87 not married ir 0.50 *** black sr 2.48 *** black ir 1.54 ** hispanic sr 1.52 ** hispanic ir 1.45 ** other race sr 0.96 other race ir 0.34 *** log income sr 1.08 log income ir 0.88 * log net worth sr 0.66 *** log net worth ir 0.62 *** note. sr = solvent revolver. ir = insolvent revolver. l. c. ricaldi et al. / financial services review 30 (2022) 107–124 119 compared with a convenience user by 12%. net worth’s impact is more pronounced, as a one percentage increase in net worth estimates a 34% decrease in the likelihood of being a solvent revolver and a 38% decrease in the likelihood of being an insolvent revolver relative to convenience use of credit cards. in table 3, we present binary logistic regression results on the likelihood of being a solvent revolver compared with a convenience user. we use a restricted sample of solvent revolvers and convenience users who are financially literate. the analysis allows us to better understand the factors that affect the decision to revolve debt since solvent revolvers have the knowledge and financial ability to be convenience users. compared with noncollege graduates, college graduates are 49% less likely to be solvent revolvers. results show a similar pattern related to having an emergency fund as these households are 31% less likely to be solvent revolvers than households that do not have an emergency fund. older respondents (55 and older) are 42% less likely than respondents 34 and younger to be solvent revolvers. contrastingly, having liquid accounts, behind on payments, bankruptcy, and race all increased the likelihood of being solvent revolvers. bankruptcy filings result in 267% table 3 binary logistic results on the likelihood of being a solvent revolver variable odds ratio p intercept 23.83 *** human capital college degree 0.51 *** mental accounting/precautionary savings motives have emergency fund 0.69 ** saving for unemployment 1.01 saving for illness 0.93 ability to borrow 0.84 self employed 1.15 have liquid accounts 1.35 ** self-control behind on payments 2.14 ** positive attitude toward credit 1.22 negative attitude toward credit 0.82 bankruptcy history 3.67 *** likely to increase spending 1.21 unwilling to reduce spending 0.85 lifecycle factors age age 35–55 1.17 age 55+ 0.58 ** female 1.06 not married 0.83 race black 4.09 *** hispanic 1.72 ** other 0.83 log income 0.82 *** log net worth 0.90 *** 120 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 increase in the likelihood of be a solvent revolver. black and hispanic households are 309% and 72% more likely to be solvent revolvers, respectively, compared with white households. income and net worth are also statistically significant at reducing the likelihood of being a solvent revolver. a one percentage increase in income, decreases the likelihood of being a solvent revolver by 18%. similarly, a positive one percentage point change in net worth decreases the likelihood of being a solvent revolver by 10%. 7. discussion the research investigates the effect of financial literacy and other behavioral factors on credit card debt puzzle. there is a limited pool of empirical literature on the credit card debit puzzle. to be more specific, this study aims to evaluate solvent revolving credit card users and their financial literacy compared with insolvent revolvers and convenience users of credit cards. based on the behavioral life-cycle hypothesis and human capital theory, the results provide an extension of the current literature. human capital plays a role in the decision to display the credit card debt puzzle. hypothesis 1 suggests solvent revolvers will have a lower level of financial literacy than convenience users. regarding specific financial human capital, solvent credit card revolvers have a lower level of financial literacy than convenience users. however, solvent revolvers should be more financially literate than insolvent revolving users. the findings of this paper support previous literature that solvent revolvers are not necessarily more financially literate than insolvent revolvers but moderating factors such as having a college degree play an important role (bertaut et al., 2009; haliassos & reiter, 2005). unsurprisingly, insolvent and solvent credit card users are less likely to have college degrees compared with convenience users; as a result, households should seek as much financial knowledge as possible or invest in assistance from a financial planner or counselor when making critical financial decisions. several factors in the mental accounting concept seem to have an impact on households’ credit card debt use. hypothesis 2 suggests that solvent revolvers would be more likely to display mental accounting and have higher precautionary savings motives than convenience users. households that report having an emergency fund are less likely to revolve credit card debt. this finding is inconsistent with previous research that households that display the credit card debt puzzle are more likely to have a precautionary savings motive (bi & hanna, 2006; telyukova & wright, 2008, gorbachev & luengo-prado, 2019). self-employed households are less likely to be revolvers in general and less likely to be solvent revolvers. this finding suggests that households that are self-employed use credit cards for convenience purposes. the impact of family and friends was presented in the results. the ability to borrow increased the likelihood to be an insolvent revolver in the mlr models that included human capital, mental accounting, and self-control but not in the full model that includes lifecycle factors that differentiation from convenience users goes away. the ability to borrow allows households who may otherwise be solvent revolvers to be more like convenience users across all models. last, having liquid accounts provides us with mixed results but highlights a notable difference between the credit card user groups. households who report l. c. ricaldi et al. / financial services review 30 (2022) 107–124 121 having accounts such as checking and savings which are not earmarked for emergencies are more likely to be solvent revolvers. this finding suggests that households that save liquid assets in different accounts will choose not to pay off their credit card balance with those earmarked funds. those who use mental accounting will have more liquid accounts compared with convenience users who have the self-control not to spend from fewer accounts. the self-control factors contribute to the further understanding of the credit card debt puzzle. the variables making up this concept provide positive and negative insights in the likelihood of being a revolver of credit card debt and convenience use. in hypothesis 3, we put forward that self-control will increase the likelihood of revolving credit card debt. revolvers would be less likely to pay their bills on time, have a positive attitude toward credit, and have filed for bankruptcy in the past. first, when compared with those who are on time with their loan payments, the households that are behind are more likely to be solvent revolvers and even more so to an insolvent revolver. next, attitude toward debt provides a surprising result. we reject our hypothesis that a positive attitude toward debt would be a characteristic of solvent revolvers. however, we see evidence that a negative attitude toward credit card debt does affect the decision to be an insolvent revolver. households that have a history of bankruptcy are more likely to have revolving debt. in regard to the lifecycle factors, solvent revolvers are more likely to be younger and have a lower household wealth. first, when compared with households under age 35, those households with individuals between 35 and 55 are more likely to be solvent revolvers than a convenience user. among the most financially literate households, we see a differing effect in the over 55 as they are less likely to be solvent resolvers compared with 35 and younger. gender plays a role as married females are more likely to be insolvent revolvers of credit card debt. last, income appears only to be a statistically significant factor among the most financial literate of the sample. in this subgroup, a positive increase can decrease the likelihood of revolving credit card debt by 18%. on the other hand, wealth is a consistent factor in reducing the likelihood of revolving credit card debt. 8. implications financial planners and financial counselors have a fiduciary duty to their clients. when providing financial planning to clients, financial professionals should help and educate households on the behavioral factors that impact the client’s financial goals. a change in behavior is necessary, especially since revolving credit card users believe it is okay to spend now and pay later. since self-control factors have such an impact on solvent revolving tendencies, financial planners and counselors should help clients identify debt management issues as well as increase their awareness of the importance of making payments on time. the client would benefit from a change in behavior to maximize utility and avoid making inefficient decisions. to facilitate the education process for solvent households, there are two steps. one, financial planners and financial counselors must teach the client about behavioral biases that exist when making financial decisions. two, financial planners and counselors should educate clients about the inefficiency associated with solvent revolving. 122 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 the first step is to educate the client about behavioral biases. since behavioral biases affect most individuals, it is important to know about them to diminish some of the unfavorable effects. financial planners and counselors can help the consumer develop strategies to control inefficient behavior and implement appropriate behavior. first, the consumer can set up automatic payments toward their credit card balance. this would reduce the credit card balance while decreasing the likelihood of consumers missing or being behind on their payments. a recent study by middlewood, chin, johnson, and knoll (2018) suggest that clients who automate savings decisions typically have financial skill, but still need focused help from their advisor to set up the automated rules. next, the planner, counselor, or educator should help the client set realistic goals for paying off the debt and help the client understand their motives for saving. last, the planner could suggest reallocating high interest rate debt to a lower interest rate debt tool (greenberg & hershfield, 2019). the idea of shifting debt to lower interest rate tools is shared by another article on myopia, financial literacy, and the choice between debit and credit (ricaldi & huston, 2019). the authors suggest that households that consistently revolve a credit card balance should switch to debit cards. once the household is accustomed to budgeting techniques and self-control, they should switch to using a credit card as a convenience tool to take advantage of the rewards. the next step in helping solvent revolvers understand their inefficient behavior is to educate them as to why the behavior is inefficient. education about interest rates, savings accounts, and the tradeoff between holding a balance on a credit card and paying off the balance using savings is essential. following these steps can allow financial planners to help clients build wealth instead of exhibiting inefficient behaviors. references becker, g. s. (1964). a theoretical and empirical analysis, with special reference to education (1st ed.). new york: columbia university press for the nber. bertaut, c. c., haliassos, m., & reiter, m. (2009). credit card debt puzzles and debt revolvers for self-control. review of finance, 1–36. bi, l. (2005). the influence of uncertainty and liquidity constraints on liquid asset holdings of credit card revolvers [dissertation]. the ohio state university. available at https://etd.ohiolink.edu/apexprod/rws_etd/ send_file/send?accession=osu1127153217&disposition=inline bi, l., & hanna, s. d. (2006). do financial planners serve the interests of their clients? use of financial planners, credit card balances and liquid assets. consumer interests annual, 52, 292–314. board of governors of the federal reserve system. (1998-2013). survey of consumer finances [data files and codebooks]. available at https://www.federalreserve.gov/econres/scfindex.htm bricker, j., dettling, l. j., henriques, a., hsu, j. w., jacobs, l., moore, k. b., . . . windle, r. a. (2017). changes in us family finances from 2013 to 2016: evidence from the survey of consumer finances. federal reserve bulletin, 103, 1. bucks, b. k., kennickell, a. b., mach, t. l., & moore, k. b. (2009). changes in us family finances from 2004 to 2007: evidence from the survey of consumer finances. federal reserve bulletin, a1, 95. chien, y. w., & devaney, s. a. (2001). the effects of credit attitude and socioeconomic factors on credit card and installment debt. journal of consumer affairs, 35, 162–179. https://doi.org/10.1111/j.1745-6606.2001. tb00107.x druedahl, j., & jørgensen, c. n. (2018). precautionary borrowing and the credit card debt puzzle. quantitative economics, 9, 785–823. https://doi.org/10.3982/qe604 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 123 gathergood, j., & weber, j. (2014). self-control, financial literacy & the co-holding puzzle. journal of economic behavior & organization, 107, 455–469. https://doi.org/10.1016/j.jebo.2014.04.018 gorbachev, o., & luengo-prado, m. j. (2019). the credit card debt puzzle: the role of preferences, credit access risk, and financial literacy. the review of economics and statistics, 101, 294–309. https://doi.org/10.1162/ rest_a_00752 greenberg, a. e., & hershfield, h. e. (2019). on shifting consumers from high-interest to low-interest debt. financial planning review, 2, e1035. https://doi.org/10.1002/cfp2.1035 gross, d. b., & souleles, n. s. (2002). do liquidity constraints and interest rates matter for consumer behavior? evidence from credit card data. the quarterly journal of economics, 117(february), 149–185. https://doi. org/10.1162/003355302753399472 haliassos, m., & reiter, m. (2005). credit card debt puzzles, unpublished working paper, cfs working paper 2005/26. center for financial studies. kim, h., & devaney, s. a. (2001). the determinants of outstanding balances among credit card revolvers. financial counseling and planning, 12, 67–78. laibson, d., repetto, a., & tobacman, j. (2001). a debt puzzle. in p. aghion, r. frydman, j. stiglitz, & m. woodford (eds.), knowledge, information, and expectations in modern economics: in honor of edmund s. phelps (pp. 228–266). princeton, nj: princeton university press. lindamood, s., hanna, s. d., & bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41, 195–222. https://doi.org/10.1111/j.1745-6606.2007.00075.x lusardi, a., & scheresberg, c. d. b. (2013). financial literacy and high-cost borrowing in the united states (no. w18969). cambridge, ma: national bureau of economic research. middlewood, b. l., chin, a., johnson, h., & knoll, m. a. (2018). exploring the relationships between impatience, savings automation, and financial welfare. financial planning review, 1, e1020. ricaldi, l., & huston, s. j. (2019). the role of dual-self constructs in determining payment card choice: insights for working with credit card borrowers. journal of financial therapy, 10, 5. https://doi.org/10.4148/19449771.1177 rutherford, l. g., & devaney, s. a. (2009). utilizing the theory of planned behavior to understand convenience use of credit cards. journal of financial counseling and planning, 20, 48–63. shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609–643. https://doi.org/10.1111/j.1465-7295.1988.tb01520.x telyukova, i. a., & wright, r. (2008). a model of money and credit, with application to the credit card debt puzzle. review of economic studies, 75, 629–647. https://doi.org/10.1111/j.1467-937x.2008.00487.x thaler, r. h. (1999). mental accounting matters. journal of behavioral decision making, 12, 183–206. https:// doi.org/10.1002/(sici)1099-0771(199909)12:3<183::aid-bdm318>3.0.co;2-f 124 l. c. ricaldi et al. / financial services review 30 (2022) 107–124 financial capability across generations and technology abeba mussaa, meeghan rogersb, xu zhanga,* adepartment of economics, farmingdale state college, state university of new york, 2350 broadhollow road, farmingdale, ny 11735–1021, usa bdepartment of business management, farmingdale state college, state university of new york, 2350 broadhollow road, farmingdale, ny 11735–1021, usa abstract financial capability is critical for individuals to survive economic hardship. as the first attempt in the literature, our research explores how being technology savvy is relevant in explaining individuals’ short-term and long-term financial behavior. specifically, we use the 2018 national financial capability study (nfcs) to uncover the mixed roles of technology in personal financial management. being technology savvy was consistently associated with less desired short-term financial behavior while positively related to good long-term financial behavior after controlling for individual financial constraints and other socio-economic variables. moreover, our study demonstrates the generational disparity of being technology savvy related to financial behavior. © 2022 academy of financial services. all rights reserved. jel classifications: g5 household finance keywords: financial capability; tech-savvy; generational disparity 1. introduction financial literacy has never been more critical to people in today’s complex financial world as essential as basic reading and writing skills. financially literate participants in the labor market better understand overall economic performance and react more wisely during economically challenging times. inadequately financially literate individuals are closely associated with personal finance issues such as low savings rates, over-indebtedness, and poor financial decisions (lusardi & mitchell, 2014). the rising college tuition and costs, *corresponding author: tel.: +1-934-420-2334, fax: +1-934-420-2689. e-mail address: xu.zhang@farmingdale.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 273–296 together with unprecedented increases in student loan debt, $1.59 trillion as of july 2021, have disproportionately burdened young people more than previous generations (federal student act q2, 2009). enabled by technological innovation, the decision-making process on personal finance matters becomes even more complex when the social and economic landscapes are dramatically changed. with the evolution of internet technology and the popularity of mobile devices, how people save, spend, invest and manage cash, how individuals participate in the labor market, along with how people communicate and exchange information on personal finance matters, have experienced unprecedented transformations. in 2021, 82% of americans have used at least one type of digital payment, defined as including browserbased or in-app online purchases, in-store checkout using a mobile phone and/or qr code, and person-to-person (p2p) payments, compared with 78% in 2020 and 72% in 2016 (mckinsey digital payments consumer survey, 2021). additionally, the penetration of new forms of financing options, such as bnpl (buy now pay later), has been credited for incremental spending on certain sectors and the emergence of cryptocurrency and digital transformation in financial services have embarked opportunities and challenges on individuals’ financial investments.1 moreover, new technologies, defined as automation, digital technologies, and machine learning (dell & nestoriak, 2020), are making profound implications on the labor market outcomes such as the income distribution of workers from different occupations as well as relative return to capital compared with labor, which in turn affects the financial well-being of individuals. addressing the question of how americans feel in their financial lives requires a holistic approach with present and future-oriented views: not only whether individuals are able to meet current financial obligations and spending needs, but also how they invest and prepare to respond to financial shocks and achieve financial success in the future. as integrated into the synthesis of financial literacy concepts, financial capability, that is, how well people manage their money and control their finances, is revealed by their evolving financial behaviors and outcomes in the fast-growing digital world. plus, researchers have suggested the generational disparity of perceived value and experience in using technology (dhanapal et al., 2015; kumar & lim, 2008) and indicated age as an important factor in technology acceptance and mobile payments (liébana-cabanillas et al., 2014; phang et al., 2006) that in turn have impacts in household wealth and financial well-being. the purpose of this study is to examine the role and generational disparity of being techsavvy on short-term and long-term personal financial capability, and how the effect varies across generations. the main contributions of this study are the following: first, using the 2018 national financial capability survey (nfcs), our findings on the mixed roles of being tech-savvy in personal financial management contribute to the literature as the first study on this attempt. by using overall and individual metrics to describe the extent of being techsavvy, we find being tech-savvy is consistently associated with less desired short-term financial behavior. however, the tech-savvy variable is positively associated with good shortterm financial behavior. this could indicate that technology allows risk-tolerant individuals to behave imprudently in short-term financial matters. however, individuals may engage in more rational long-term financial behavior using technology. second, financial capability and technology savviness can vary across generations. among four generations, that is, igen/gen z (born between 1994 and 2000), millennials/ 274 a. mussa et al. / financial services review 30 (2022) 273–296 geny (born between 1993 and 1984), gen x (born between 1983 and 1964), and baby boomers (born before 1964), being equipped with good financial knowledge favors gen z and baby boomers to engage in good short-term financial behavior. more interestingly, being financially confident is positively associated with good short-term financial behavior across all generations. technology acceptance also has different implications on financial behavior across various age groups. the tech-savvy variable is found to be negative and significant across all generations in the logit regression of short-term financial behavior, meaning being tech-savvy has an adverse effect on short-term financial behavior. the marginal coefficients are highest for igen and baby boomers with �5.9% and �8.87%, respectively. this could indicate that those in the youngest generations and oldest generations are not using technology in a way that would benefit their money management in the short term. moreover, techsavvy individuals in gen z, millennials, and gen x are more likely to engage in good longterm financial behavior with marginal coefficients for being tech-savvy 7.2%, 5.9%, and 3.6%, respectively. the rest of the paper is organized as follows: section ii provides an overview of existing studies; section iii introduces the 2018 national financial capability survey (nfcs); empirical results are discussed in section iv; section v presents robustness checks, and section vi concludes with a discussion on future research. 2. literature review financial decisions have profound impacts on an individual’s financial well-being. enabled by technological innovations, the decision-making on personal finance matters becomes even more complex in a fast-changing social and economic landscape. recent literature on the impacts of digitalization on personal financial capability has focused on three areas: consumer behavior, online financial service, and digital adoptions relevant to work tasks. a line of research has examined the technological implication on financial capability through consumer’s online shopping and other consumer-credit digital transactions, such as in-game purchases, which in turn were linked to high-cost debts (carlsson et al. 2017; garrett et al. 2014). additionally, individuals are increasingly adopting online financial services such as mobile banking, fund transfers, bill payments, and cash management. hee yeo and fisher (2017) reports more frequent use of online financial service is associated with a high level of financial capability. technological advancement has not only transformed individual behavior in communication, consumption, and social interaction, but also provided a unique opportunity to complement traditional avenues to acquire financial literacy and manage personal financial matters. the federal reserve board (2016) indicates that mobile phone usage is higher among younger groups, with the same trends in smartphone usage. prior research has documented both positive and negative effects of technology adoption on personal use. using technology in social networks increases self-esteem but reduces self-control and self-regulation ability (wilcox & stephen, 2012). chan and saqib (2015) report that using online social networking services increases financial risk-taking. carlsson et al. (2017) suggest that as the use of a. mussa et al. / financial services review 30 (2022) 273–296 275 the internet and mobile use increases, policymakers need to be aware of the consequences of digital measurability. they suggest that purchases are integrated into digital games targeting children. they also examine the increase in payment forms with installment options or unsecured loans. this provides consumers with more complex financial decisions on purchasing and taking on credit. hee yeo and fisher (2017) recommend that an increased frequency of using a mobile phone for financial services is associated with a greater likelihood of having better money management skills and higher levels of financial capability. de meza et al. (2008) state that the increasing use and adoption of mobile technologies could help individuals manage their finances better. hogarth and anguelov (2004) report that families using phone banking and computer banking contribute to higher levels of financial management, with the use of computer banking having the largest impact. e-banking can be an additional tool that compliments other financial management skills rather than e-banking being the sole driver of better financial management. walsh and lim (2020) study millennials’ financial behavior and finds heavy technology adopters tend to be more likely to engage in positive financial behavior such as setting up emergency funds, retirement and investment accounts and less-desired financial behavior of overspending. in our paper, the variable tech-savvy includes both mobile and web app users. we also break down the tech-savvy variable by specific metrics in responses to survey questions: how often individuals used their phone to pay for a product or service in person, and how often individuals transfer money on a mobile phone. part of our research investigates how using technology affects different generations’ financial behavior in the short-term and long-term. mobile use for transfers is negatively and significantly associated with good short-term financial behavior of gen x, and baby boomers. those earlier generations who use technology for mobile transfers are 4.0% to 6.7% less likely to engage in good short-term financial behavior. additionally, the tech-savvy variable positively and significantly relates to the good long-term financial behavior of igen/genz, millennials/geny, and genx, while the tech-savvy variable is insignificant for the baby boomers. prior research has examined the relationship between age and technology. for example, liébana-cabanillas et al. (2014) suggest that the age of the technology user plays a significant role in trusting and using mobile payment systems. younger users are less affected by acceptance problems while older users have less trust in mobile payment systems. regarding cultural dominance, gen x falls short in the digital category (visual capitalist, 2021). shobha and kumar (2020) find that different generations have different perspectives on life, and so too is their personal financial behavior. their research focuses on the financial behaviors between generation x and generation y. they find that gen x has a high score on financial literacy, propensity to planning, and financial risk tolerance, in comparison with gen y. gen x also has more established long-term financial behavior. their study reports that gen x has aligned their investment pattern with their investment objectives, including long-term investments. kumar and lim (2008) suggest that baby boomers utilize mobile phones for a more functional and utilitarian reason in comparison with other generations. similarly, berraies et al. (2017) indicate that baby boomers put an emphasis on monetary and quality values when deciding their trust toward mobile financial use. 276 a. mussa et al. / financial services review 30 (2022) 273–296 when determining short-term and long-term financial behaviors, we follow an approach similar to wagner and walstad (2019). short-term financial behaviors involve money and credit activities that provide quick and consistent feedback. these behaviors can then be changed as a result of this feedback to attempt to avoid penalties. long-term financial behaviors require more planning with less timely feedback (wagner & walstad, 2019). these behaviors involve planning and thinking about the future. as a contribution to this literature, our paper examines how being tech-savvy is relevant in explaining shortand long-term financial behavior. additionally, we try to understand how the relevancy varies across generations. 3. data the data for this study came from the 2018 national financial capability study (nfcs). the largest component of the nfcs, the state-by-state survey, was conducted across a large and diverse sample that provided a comprehensive analysis of the financial capability of the national population as a whole. the survey was conducted online from june through october 2018 among a nationally representative sample of 27,091. the final sample used for this study is 19,725 after dropping the observations where the respondent chose “prefer not to say” or “don’t know” to the questions about financial behaviors and management. however, an answer of “prefer not to say” or “don’t know” is coded as incorrect in the case of the objective financial knowledge questions.2 the survey questionnaires were divided into 10 sessions: (1) demographics, (2) financial attitudes and behaviors, (3) banking, (4) retirement accounts, (5) government benefits, (6) home and mortgages, (7) credit cards, (8) other debts, (9) insurance, and (10) selfassessment and literacy. table 1 provides descriptive statistics on the demographic characteristics of the full sample. the sample contained about 46% male, 56% married, 26% single, 13% divorced or separated, and 5% widowed. as for education, around 16% had a high school diploma, 33% had some college education, about 23% had a college degree only, and 15% had some post-graduate education. about 74% of the sample were white, about 7% were self-employed, and 3% were looking for a job that is, unemployed. about one-third of the sample earned annual income below $75,000, and about 22% of the sample were making more than $100,000 a year. to incorporate the generational differences in the adoption and use of technology, the sample is grouped into four age groups. the youngest adult group in the sample belongs to gen z, who were born between 1994 and 2000. millennials or gen y are those born between 1993 and 1984, the youngest being 25 years old and the oldest being 34 years old as of 2018. those who were born between 1983 and 1964 belong to the gen x generation, the youngest being 35 years and the oldest being 54 years old. the oldest generation in our sample is baby boomers who were at least 55 years old as of 2018.3 the generational composition of our sample is as follows, about 7% gen z, 16% millennials or gen y, 35% gen x, and 43% baby boomers. table 1 also breaks the sample into two groups, tech-savvy respondents and non-tech– savvy respondents. a two-sample t test is used to determine if there is a significant a. mussa et al. / financial services review 30 (2022) 273–296 277 t ab le 1 d es cr ip ti v e st at is ti cs o f d em o g ra p h ic v ar ia b le s v ar ia b le s f u ll sa m p le t ec h -s av v y = 1 t ec h -s av v y = 0 t te st n = 1 9 ,7 2 5 n = 6 ,9 2 4 n = 1 2 ,8 0 1 h o : d if f = 0 m ea n s d m ea n s d m ea n s d s ta t. p -v al u e m al e 0 .4 5 9 0 .4 9 8 0 .4 8 3 0 .4 9 9 7 0 .4 4 7 0 .4 9 7 1 �4 .9 5 0 8 0 .0 0 0 a g e g ro u p /g en er at io n g en z /i g en (1 8 – 2 4 ) 0 .0 6 7 2 0 .2 5 0 0 .1 2 3 0 .3 2 8 0 0 .0 3 7 0 .1 8 9 4 �2 3 .1 5 6 2 0 .0 0 0 g en y /m il le n n ia ls (2 5 – 3 4 ) 0 .1 5 7 8 0 .3 6 5 0 .2 8 5 0 .4 5 1 4 0 .0 8 9 0 .2 8 4 7 �3 7 .2 8 4 7 0 .0 0 0 g en x (3 5 – 5 4 ) 0 .3 5 0 .3 8 4 0 .2 3 7 0 .4 2 5 5 0 .1 3 0 0 .3 3 5 9 �1 9 .5 2 1 3 0 .0 0 0 b ab y b o o m er (5 5 + ) 0 .4 2 7 8 0 .4 9 5 0 .1 8 0 0 .3 8 4 1 0 .5 6 2 0 .4 9 6 2 5 5 .6 8 8 6 0 .0 0 0 w h it e 0 .7 6 8 1 0 .4 2 2 0 .6 5 8 0 .4 7 4 4 0 .8 2 8 0 .3 7 7 7 2 7 .4 4 2 5 0 .0 0 0 h ig h sc h o o l g ra d u at es o n ly 0 .1 6 1 5 0 .3 6 8 0 .1 2 7 0 .3 3 2 6 0 .1 8 0 0 .3 8 4 5 9 .8 0 7 7 0 .0 0 0 s o m e co ll eg e ed u ca ti o n (i n cl u d in g as so ci at e) 0 .3 2 6 4 0 .4 6 8 0 .3 3 9 0 .4 7 3 2 0 .3 2 0 0 .4 6 6 5 �2 .6 5 3 2 0 .0 0 8 b ac h el o r d eg re e o n ly 0 .2 3 4 7 0 .4 2 4 0 .2 5 5 0 .4 3 5 9 0 .2 2 4 0 .4 1 6 7 �4 .9 6 9 6 0 .0 0 0 p o st -g ra d u at e ed u ca ti o n 0 .1 5 2 4 0 .3 5 9 0 .1 5 7 0 .3 6 3 7 0 .1 5 0 0 .3 5 7 1 �1 .2 7 9 0 .2 0 1 m ar ri ed 0 .5 6 1 8 0 .4 9 6 0 .5 2 9 0 .4 9 9 2 0 .5 8 0 0 .4 9 3 6 6 .9 2 5 7 0 .0 0 0 s in g le 0 .2 5 8 0 .4 3 8 0 .3 4 5 0 .4 7 5 5 0 .2 1 1 0 .4 0 7 9 �2 0 .8 3 9 7 0 .0 0 0 s ep ar at ed o r d iv o rc ed 0 .1 3 3 5 0 .3 4 0 0 .1 0 6 0 .3 0 8 2 0 .1 4 8 0 .3 5 5 5 8 .2 9 9 6 0 .0 0 0 w id o w ed o r w id o w er 0 .0 4 6 6 0 .2 1 1 0 .0 2 0 0 .1 3 9 3 0 .0 6 1 0 .2 3 9 5 1 3 .1 9 3 7 0 .0 0 0 h o u se h o ld in co m e ra n g e: l es s th an 2 5 k 0 .1 7 2 3 0 .3 7 8 0 .1 4 6 0 .3 5 3 3 0 .1 8 6 0 .3 8 9 5 7 .1 6 4 2 0 .0 0 0 $ 2 5 k – $ 5 0 k 0 .2 4 9 1 0 .4 3 3 0 .2 3 9 0 .4 2 6 5 0 .2 5 5 0 .4 3 5 7 2 .4 2 4 7 0 .0 1 5 $ 5 0 k – $ 7 5 k 0 .2 0 4 0 .4 0 3 0 .2 0 1 0 .4 0 0 8 0 .2 0 6 0 .4 0 4 2 0 .7 7 3 0 .4 4 0 $ 7 5 k – $ 1 0 0 k 0 .1 5 3 0 .3 6 0 0 .1 7 2 0 .3 7 7 0 0 .1 4 3 0 .3 5 0 0 �5 .3 3 2 8 0 .0 0 0 $ 1 0 0 k an d m o re 0 .2 2 1 4 0 .4 1 5 0 .2 4 2 0 .4 2 8 4 0 .2 1 0 0 .4 0 7 5 �5 .1 6 6 6 0 .0 0 0 u n em p lo y ed 0 .0 3 4 3 0 .1 8 2 0 .0 3 3 0 .1 7 8 1 0 .0 3 5 0 .1 8 4 2 0 .8 7 2 2 0 .3 8 3 278 a. mussa et al. / financial services review 30 (2022) 273–296 difference between the means of the two groups. the results show that the mean of the majority of the variables, listed in table 1, are significantly different. we computed two measures of financial behavior: short term and long term. the shortterm financial behavior involves money or credit management behavior that includes timely payments of bills each month, managing checking accounts to avoid overdrafts, paying off credit card balances, and making timely payments to mortgages. the long-term financial behavior is reflected by saving and investment (retirement and non-retirement) decisions that normally require long-term planning (asaad 2015; wagner & walstad, 2019). in this paper, the short-term financial behavior variable is created based on the individual and aggregate responses to the three survey questions that asked if the respondent: (1) always paid off a credit card bill in full, (2) spent less than or equal to his or her income, and (3) overdrew his or her checking account occasionally. if the respondent answered “yes” to each of the questions, a separate individual measure of short-term financial behavior is coded “1,” and otherwise “0.” the overall measure of short-term financial behavior is constructed by adding the three individual measures with an average of 2.05. a binary dependent variable of short-term financial behavior—is created and assigned “1” if the overall measure is above the average, 2.05; otherwise, “0.” that means if the respondents claimed yes to all three of these behaviors, they were assumed to have good or strong short-term financial behavior. alternatively, each of the three individual measures is used separately as a measure of shortterm financial behavior. as reported in table 2, about 79% of the respondents had never over drafted their checking account, 81% spent less than or equal to their income over the past year, and 48% always paid their credit card bill in full. about 39% of the respondents were engaged in all three behaviors. some financial decisions are complex, future-oriented, and require planning (beverly et al., 2003). an individual long-term financial behavior measure was created based on the individual and aggregate responses to the five questions that asked if the respondent: (1) had an emergency or rainy day fund; (2) had a saving account, money market account or cds; table 2 short-term and long-term financial behavior responses yes no index = 1 index = 0 n % n % short-term financial behavior always paid credit card bill in full 9,476 48.0 10,249 52.0 spent less or equal to income 16,063 81.4 3,662 18.6 not overdraft checking account 15,477 78.5 4,248 21.5 overall short-term variable 7,639 38.7 12,086 61.3 long-term financial behavior has saving account 15,672 79.5 4,053 20.5 has investments (non-retirement) 7,600 38.5 12,125 61.5 figured out retirement needs 12,544 63.6 7,181 36.4 has retirement plan 14,144 71.7 5,581 28.3 had emergency fund 10,950 51.5 8,775 44.5 overall long-term variable 11,939 60.5 7,786 39.5 a. mussa et al. / financial services review 30 (2022) 273–296 279 (3) had investments in stocks, bonds, mutual funds, or other securities that are outside of retirement accounts; (4) had ever tried to figure out their retirement needs; and (5) had any retirement plans either through an employer or not. each of the above measures is coded “1” for yes and “0” for no. the overall measure of long-term financial behavior is created by adding the five individual measures, with an average of 2.76. a binary dependent variable of long-term financial behavior is created and coded as “1” if the overall measure is 3 or more, and “0” otherwise. as indicated in table 2, about 61% of the respondents claimed at least three areas of long-term financial behavior. each response to the five questions is also used as an alternative measure of long-term financial behavior. about 80% of the respondents have a saving account, money market account, or cds; 39% of them invested either in stocks, bonds, mutual funds or other securities, not including retirement accounts; 72% had a retirement plan either from an employer or some other way, and 64% had already figured out their retirement needs. about 52% of the respondents had an emergency or rainy-day fund that would cover expenses for three months in case of sickness, job loss, economic downturn, or other emergencies. consistent with the existing literature (asaad, 2015; wagner & walstad, 2019; xiao & porto, 2017), financial literacy variables are defined separately to reflect the actual knowledge and the perceived knowledge of basic financial literacy. the objective or actual financial knowledge variable is derived from responses to questions about interest accrual, inflation, bond prices, mortgage, risk, and bond duration.4 similarly, the subjective or perceived knowledge of financial literacy, that is, financial confidence, is constructed based on the responses to three survey questions that assess how the respondents were satisfied with their personal financial condition, rate their overall financial knowledge, and their day-today financial matters.5 one of the main contributions of this paper is examining the role of technology on financial behavior. technology has dramatically changed the way people handle personal financial transactions, everything from online and mobile banking and virtual wallets to barcodebased mobile payments and cryptocurrencies. in addition, the ease of communication has allowed remote and contract workers to work on their own terms, increasing the size of the workforce in the gig economy. the emergence of the gig economy forced people to use various online applications such as instacart and doordash for grocery and meal delivery, and rideshare apps such as uber and lyft. even if there is no consensus as to who qualifies as a gig worker, the bureau of labor statistics estimated that there were 55 million gig economy workers in the nation in 2017. in our regression, we control how comfortable people are in using and adopting technology for managing their finances and defined them as tech-savvy. four items are selected from the 2018 nfcs survey and recoded as binary variable to measure tech-savvy: (1) how often respondents used their mobile phone to pay for a product or service in person at a store, gas station, or restaurant; (2) how often do respondents used mobile phone to transfer money to another person; (3) how often they used websites or apps to help with financial tasks such as budgeting, saving or credit management, such as credit karma or goodbudget; and (4) how often they took on a work assignment through a website or mobile apps such as uber. if the respondents answered “frequently” or “sometimes” for each of the above four questions, the corresponding variables—mobile use in person, mobile use for transfer, web and app for personal use, and web and app use for work—are 280 a. mussa et al. / financial services review 30 (2022) 273–296 separately coded as “1,” otherwise “0.” in addition, we constructed the overall measure of tech-savvy by horizontally adding all the above four measures, and coded as “1” if respondents were affirming at least two of the four questions. 4. empirical results logistic regressions were used to explore how the usage and adoption of technology affects personal financial behaviors. tables 3 through 6 show the estimated average marginal effects in which the dependent variable measured short-term financial behaviors, longterm financial behaviors, and individual measures of either behavior. as reported in table 3, financial knowledge and financial confidence variables are positive and significant when either one of these two variables, or both together, or the interaction term were controlled. individuals with good financial knowledge are 2.3% to 6.7% more likely to engage in good short-term financial behavior. those individuals with high confidence in their financial knowledge are 17% to 20% more likely to engage in good short-term financial behavior. this means that those who are confident in their financial knowledge have better financial behavior than those who only have financial knowledge. this is consistent with henager and cude (2016), who found that individuals with greater confidence had better financial behavior. individuals with both high confidence and knowledge are about 4.5% more likely to engage in good short-term table 3 logistic regression predicting a measure of overall short-term financial behavior short-term financial behavior variables (1) (2) (3) (4) (5) (6) financial knowledge 0.0696*** 0.0520*** 0.0243** 0.0226** 0.0229** (0.00648) (0.00633) (0.00972) (0.00969) (0.00966) financial confidence 0.200*** 0.196*** 0.170*** 0.169*** 0.168*** (0.00556) (0.00559) (0.00881) (0.00879) (0.00878) financial knowledge* confidence 0.0454*** 0.0430*** 0.0414*** (0.0121) (0.0121) (0.0121) tech-savvy �0.0640*** (0.00658) mobile use in person �0.0129* (0.00727) mobile use for transfer �0.0495*** (0.00717) web and app for personal use �0.0415*** (0.00662) web and app use for work 0.00670 (0.0104) pseudo r2 0.227 0.262 0,264 0.265 0.268 0.271 observations 19,725 19,725 19,725 19,725 19,725 19,725 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. a. mussa et al. / financial services review 30 (2022) 273–296 281 financial behaviors. as indicated in table 3, the tech-savvy variable is negatively associated with good short-term financial behavior. that means being tech-savvy becomes more of a distraction, and may lead to poor management of personal finance. when examining the individual measures of tech-savvy variables, we find mobile use in person, mobile use for transfer, and web and app for personal use are negative and significant. individuals who use their mobile phone for personal use are 1.3% less likely to engage in good short-term financial behavior, with mobile use for transfers at 5% and web and app for personal use at 4.2%. using mobile phones could trigger over spending because of pop-up ads by impulsive online purchases for convenience. table 4 below shows the result of the logistic regression when the dependent variables are individual measures of short-term financial behavior (paid credit card in full, spending less or equal to income each month, and not overdrawing checking accounts). the results of the regression, as reported in column 1, indicate individuals who paid their credit card in full tend to be more likely to have better financial knowledge, greater overall financial confidence, and use web and app for work. individuals who use the web and app for work are 9.1% more likely to pay their credit card in full each month. this could be due to the ease and usage of banking apps for timely payments. additionally, individuals paying off a credit card balance in full are less likely to be tech-savvy, use mobile for transfers, or use the web or mobile apps for personal use. for the behavior of spending less or equal to income each month, financial confidence and financial knowledge*confidence are positive and significant. tech-savvy, mobile use in person and for transfer, and web and app for personal use and work, are all significant and negative. this indicates being more tech-savvy and increased use of technology may cause individuals to poorly manage their spending. technology can make overspending easier by being able to make purchases from a phone. some apps, such as amazon, can save your information for very smooth transactions. for the last individual measure of not overdrawing checking accounts, financial knowledge, and overall financial confidence are positive and significant. tech-savvy, mobile use in person and for transfer, web, and app use for personal and work are all significant and negative. again, this indicates that those that are more likely to use technology are spending more than they have causing an overdraft in their checking account. this could be related to the impulsive shopping online from a phone or computer. financial knowledge positively and significantly explains the overall long-term financial behavior, as indicated in the six different specification/regression results as shown in table 5. individuals with good financial knowledge were 8.0% to 10.4% more likely to engage in good or strong long-term financial behavior. individuals with better than average financial confidence are 15.7% to 17.8% more likely to engage in good long-term financial behavior. the tech-savvy variable positively and significantly explains long-term financial behavior. the same is true for web and app use for personal use and for work. individuals who use the web or app for personal use and for work were 3.5% and 3%, respectively, more likely to engage in good financial behavior in the long term. these individuals could be using their phones to keep track of their finances and using apps to help save for retirement. they could be using apps to track their long-term investments. in general, technology use and easy adoption could be helpful for efficient planning and managing personal finance. 282 a. mussa et al. / financial services review 30 (2022) 273–296 t ab le 4 l o g is ti c re g re ss io n p re d ic ti n g in d iv id u al m ea su re o f sh o rt -t er m fi n an ci al b eh av io r v ar ia b le s s h o rt -t er m fi n an ci al b eh av io r p ai d cr ed it fu ll s p en d in g le ss /e q u al to in co m e n o t o v er d re w ch ec k in g ac co u n t (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) f in an ci al k n o w le d g e 0 .0 2 2 9 * * 0 .0 2 4 7 * * * �0 .0 0 5 0 6 �0 .0 0 5 3 3 0 .0 1 9 9 * * * 0 .0 1 8 9 * * * (0 .0 0 9 1 6 ) (0 .0 0 9 1 4 ) (0 .0 0 6 9 5 ) (0 .0 0 6 9 4 ) (0 .0 0 6 5 8 ) (0 .0 0 6 5 5 ) f in an ci al co n fi d en ce 0 .1 9 9 * * * 0 .1 9 4 * * * 0 .1 1 9 * * * 0 .1 2 2 * * * 0 .0 6 2 0 * * * 0 .0 6 7 3 * * * (0 .0 0 8 3 8 ) (0 .0 0 8 3 8 ) (0 .0 0 8 4 0 ) (0 .0 0 8 4 6 ) (0 .0 0 7 4 6 ) (0 .0 0 7 5 1 ) f in an ci al k n o w le d g e* c o n fi d en ce 0 .0 2 8 3 * * 0 .0 3 0 7 * * 0 .0 2 6 4 * * 0 .0 2 2 1 * 0 .0 7 4 8 * * * 0 .0 6 7 4 * * * (0 .0 1 2 0 ) (0 .0 1 2 0 ) (0 .0 1 1 6 ) (0 .0 1 1 6 ) (0 .0 1 1 2 ) (0 .0 1 1 2 ) t ec h -s av v y �0 .0 2 8 3 * * * �0 .0 4 8 3 * * * �0 .0 5 9 3 * * * (0 .0 0 6 7 2 ) (0 .0 0 5 7 6 ) (0 .0 0 5 3 9 ) m o b il e u se in p er so n 0 .0 0 6 8 1 �0 .0 1 5 0 * * �0 .0 2 0 1 * * * (0 .0 0 7 3 5 ) (0 .0 0 6 3 2 ) (0 .0 0 5 8 8 ) m o b il e u se fo r tr an sf er �0 .0 3 4 6 * * * �0 .0 3 1 2 * * * �0 .0 3 2 2 * * * (0 .0 0 7 2 5 ) (0 .0 0 6 2 4 ) (0 .0 0 5 8 5 ) w eb an d ap p fo r p er so n al u se �0 .0 4 0 1 * * * �0 .0 2 2 3 * * * �0 .0 2 3 2 * * * (0 .0 0 6 7 3 ) (0 .0 0 5 9 4 ) (0 .0 0 5 5 9 ) w eb an d ap p u se fo r w o rk 0 .0 9 0 6 * * * �0 .0 2 0 2 * * �0 .0 5 4 6 * * * (0 .0 1 0 3 ) (0 .0 0 8 3 0 ) (0 .0 0 7 5 3 ) p se u d o r 2 0 .2 6 0 0 .2 6 4 0 .0 9 8 0 .0 9 9 0 .2 4 6 0 .2 5 0 o b se rv at io n s 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 n o te . t h e m ar g in al co ef fi ci en ts o f o th er co n tr o l v ar ia b le s (g en d er , ed u ca ti o n le v el , em p lo y m en t st at u s, m ar it al st at u s, in co m e, ra ce , ri sk le v el m ea su re s, an d cr ed it sc o re ) ar e in cl u d ed in th e re g re ss io n , b u t n o t re p o rt ed h er e. s ta n d ar d er ro rs in p ar en th es es : * * * p < .0 1 , * * p < .0 5 , * p < .1 . a. mussa et al. / financial services review 30 (2022) 273–296 283 t ab le 5 l o g is ti c re g re ss io n p re d ic ti n g o v er al l m ea su re o f lo n g -t er m fi n an ci al b eh av io r v ar ia b le s l o n g -t er m fi n an ci al b eh av io r (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) f in an ci al k n o w le d g e 0 .1 0 4 * * * 0 .0 9 2 2 * * * 0 .0 7 9 9 * * * 0 .0 8 0 3 * * * 0 .0 7 9 7 * * * (0 .0 0 5 7 5 ) (0 .0 0 5 5 9 ) (0 .0 0 7 2 3 ) (0 .0 0 7 2 2 ) (0 .0 0 7 2 2 ) f in an ci al co n fi d en ce 0 .1 7 8 * * * 0 .1 7 2 * * * 0 .1 5 8 * * * 0 .1 5 8 * * * 0 .1 5 7 * * * (0 .0 0 5 2 4 ) (0 .0 0 5 2 2 ) (0 .0 0 7 2 4 ) (0 .0 0 7 2 3 ) (0 .0 0 7 2 5 ) f in an ci al k n o w le d g e* c o n fi d en ce 0 .0 2 9 8 * * * 0 .0 3 0 9 * * * 0 .0 3 3 2 * * * (0 .0 1 1 0 ) (0 .0 1 1 0 ) (0 .0 1 1 0 ) t ec h -s av v y 0 .0 2 5 8 * * * (0 .0 0 6 0 8 ) m o b il e u se in p er so n �0 .0 0 0 7 9 7 (0 .0 0 6 6 1 ) m o b il e u se fo r tr an sf er 0 .0 0 1 4 7 (0 .0 0 6 5 1 ) w eb an d ap p fo r p er so n al u se 0 .0 3 4 9 * * * (0 .0 0 6 1 1 ) w eb an d ap p u se fo r w o rk 0 .0 3 0 0 * * * (0 .0 0 9 1 6 ) p se u d o r 2 0 .3 2 1 0 .3 4 0 0 .3 4 9 0 .3 4 9 0 .3 5 0 0 .3 5 1 o b se rv at io n s 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 n o te . t h e m ar g in al co ef fi ci en ts o f o th er co n tr o l v ar ia b le s (g en d er , ed u ca ti o n le v el , em p lo y m en t st at u s, m ar it al st at u s, in co m e, ra ce , ri sk le v el m ea su re s, an d cr ed it sc o re ) ar e in cl u d ed in th e re g re ss io n , n o t re p o rt ed h er e. s ta n d ar d er ro rs in p ar en th es es : * * * p < .0 1 , * * p < .0 5 , * p < .1 . 284 a. mussa et al. / financial services review 30 (2022) 273–296 similar to short-term behaviors, we also use individual measures of long-term financial behaviors, including having an emergency fund, owning a savings account, owning a retirement account, having a non-retirement investment, and having figured out the retirement needs. the results are shown in table 6 below. the financial knowledge and financial confidence variables are positive and significant for all 10 of the specifications. the marginal effect for the interaction term, financial knowledge*confidence, is positive and significant for the first eight specifications in table 6. as indicated in columns 9 and 10, the coefficient for the interaction term is negative. however, the net overall impact of financial knowledge and financial confidence on financial behavior is still positive and significant, respectively.6 being tech-savvy is also negatively and significantly associated with behaviors of having an emergency fund. this could underscore the idea that being tech-savvy adversely affects the potential saving for a three-month emergency fund. it also confirms that individuals who are using technology more often are more likely to spend rather than save. on the other hand, the more tech-savvy participants have a higher likelihood to have retirement accounts and figure out their retirement needs. this could indicate that individuals could be using technology to establish and monitor their retirement accounts, which helps them to effectively manage their long-term financial needs. when comparing the regression results of short-term financial behavior (tables 3 and 4) to long-term financial behavior (tables 5 and 6), the observed financial knowledge is a relevant variable in explaining financial behavior. we notice that the marginal effects from the long-term regressions are larger than the corresponding marginal effects from the short-term regression. this shows that financial knowledge affects the long-term financial behavior at a much higher rate than short-term financial behavior. on the other hand, the marginal effects of financial confidence appeared to be slightly higher in the short-term models than in the long-term estimation results (table 3 vs. table 5). this could be because people receive timely feedback for the short-term financial responsibilities. for example, if people were penalized due to not paying their credit card balance in full, their perception of handling their finances would be more pronounced in short-term rather than long-term financial behavior. as shown in tables 3 and 5, the tech-savvy variable is negative and significant in the overall short-term financial behavior model, and positive and significant in the overall long-term financial behavior model. this suggests people may use technology recklessly and behave irrationally in terms of managing their short-term finances such as overspending or overriding their credit limit. on the contrary, they may engage in more rational long-term financial behavior that requires long-term planning. when looking at individual constraints and other socio-economic variables in the regression models for short-term financial behavior, we find that men have better short-term financial behavior against their counterparts, women. on average, white individuals are more likely to have better short-term financial behavior than the reference category, black. individuals with a bachelor’s or post-graduate degrees are also more likely to engage in better short-term financial behavior compared with high school graduates. married individuals are less likely to have good short-term financial behavior against unmarried counterparts. this could be due to having a heavier financial burden on a spouse and a family. all income variables are positive and significant with short-term financial behavior (earning annual average income less than 25k is the reference category). the coefficients are highest for those a. mussa et al. / financial services review 30 (2022) 273–296 285 t ab le 6 l o g is ti c re g re ss io n p re d ic ti n g in d iv id u al m ea su re o f lo n g -t er m fi n an ci al b eh av io r v ar ia b le s m ea su re s o f lo n g -t er m fi n an ci al b eh av io r e m er g en cy fu n d s av in g ac co u n t r et ir em en t ac co u n t n o n re ti re m en t in v es tm en t f ig u re d re ti re m en t (1 ) (2 ) (3 ) (4 ) (5 ) (6 ) (7 ) (8 ) (9 ) (1 0 ) f in an ci al k n o w le d g e 0 .0 1 9 6 * * 0 .0 2 0 3 * * 0 .0 3 3 6 * * * 0 .0 3 2 3 * * * 0 .0 5 5 9 * * * 0 .0 5 4 8 * * * 0 .0 8 8 1 * * * 0 .0 8 7 7 * * * 0 .1 2 4 * * * 0 .1 2 2 * * * (0 .0 0 8 0 4 ) (0 .0 0 8 0 2 ) (0 .0 0 6 7 6 ) (0 .0 0 6 7 5 ) (0 .0 0 6 8 5 ) (0 .0 0 6 8 5 ) (0 .0 0 9 8 7 ) (0 .0 0 9 8 5 ) (0 .0 0 9 6 8 ) (0 .0 0 9 5 9 ) f in an ci al co n fi d en ce 0 .2 3 9 * * * 0 .2 3 5 * * * 0 .0 8 6 0 * * * 0 .0 8 7 6 * * * 0 .0 5 0 9 * * * 0 .0 5 1 7 * * * 0 .1 4 5 * * * 0 .1 4 3 * * * 0 .0 8 2 5 * * * 0 .0 7 9 4 * * * (0 .0 0 7 4 8 ) (0 .0 0 7 4 9 ) (0 .0 0 7 6 3 ) (0 .0 0 7 6 2 ) (0 .0 0 7 2 0 ) (0 .0 0 7 2 1 ) (0 .0 0 9 4 9 ) (0 .0 0 9 5 1 ) (0 .0 1 0 1 ) (0 .0 1 0 1 ) f in an ci al k n o w le d g e* c o n fi d en ce 0 .0 3 4 0 * * * 0 .0 3 5 9 * * * 0 .0 1 2 7 0 .0 1 3 4 0 .0 3 7 4 * * * 0 .0 3 7 8 * * * 0 .0 3 3 5 * * * 0 .0 3 5 9 * * * �0 .0 9 9 0 * * * �0 .0 8 9 8 * * * (0 .0 1 1 4 ) (0 .0 1 1 4 ) (0 .0 1 1 3 ) (0 .0 1 1 2 ) (0 .0 1 0 8 ) (0 .0 1 0 8 ) (0 .0 1 2 5 ) (0 .0 1 2 5 ) (0 .0 1 3 0 ) (0 .0 1 2 9 ) t ec h -s av v y �0 .0 2 4 0 * * * 0 .0 3 2 6 * * * 0 .0 0 2 6 9 0 .0 0 2 5 0 0 .1 3 7 * * * (0 .0 0 6 3 8 ) (0 .0 0 5 8 2 ) (0 .0 0 5 7 5 ) (0 .0 0 6 8 6 ) (0 .0 0 6 7 1 ) m o b il e u se in p er so n 0 .0 0 6 6 3 �0 .0 2 0 5 * * * �0 .0 0 8 5 7 �0 .0 0 7 7 8 0 .0 2 0 3 * * * (0 .0 0 6 9 8 ) (0 .0 0 6 2 2 ) (0 .0 0 6 2 7 ) (0 .0 0 7 5 1 ) (0 .0 0 7 5 5 ) m o b il e u se fo r tr an sf er �0 .0 4 8 3 * * * 0 .0 3 9 3 * * * 0 .0 0 4 3 2 �0 .0 0 7 4 7 0 .0 8 2 7 * * * (0 .0 0 6 8 4 ) (0 .0 0 6 2 4 ) (0 .0 0 6 2 0 ) (0 .0 0 7 4 6 ) (0 .0 0 7 3 7 ) w eb an d ap p fo r p er so n al u se �0 .0 0 9 0 8 0 .0 3 3 0 * * * 0 .0 2 0 4 * * * 0 .0 0 3 5 6 0 .1 0 2 * * * (0 .0 0 6 4 2 ) (0 .0 0 5 8 3 ) (0 .0 0 5 7 9 ) (0 .0 0 6 9 2 ) (0 .0 0 6 8 9 ) w eb an d ap p u se fo r w o rk 0 .0 6 2 1 * * * �0 .0 1 5 6 * �0 .0 0 5 1 3 0 .0 6 7 6 * * * 0 .0 6 4 0 * * * (0 .0 0 9 6 9 ) (0 .0 0 8 4 6 ) (0 .0 0 8 4 8 ) (0 .0 1 0 3 ) (0 .0 1 0 3 ) p se u d o r 2 0 .2 9 7 0 .3 0 0 0 .1 9 3 0 .1 9 6 1 0 .3 2 3 0 .3 2 4 0 .2 2 6 0 .2 2 7 0 .1 3 5 0 .1 4 4 o b se rv at io n s 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 1 9 ,7 2 5 n o te . t h e m ar g in al co ef fi ci en ts o f o th er co n tr o l v ar ia b le s (g en d er , ed u ca ti o n le v el , em p lo y m en t st at u s, m ar it al st at u s, in co m e, ra ce , ri sk le v el m ea su re s, an d cr ed it sc o re ) ar e in cl u d ed in th e re g re ss io n , n o t re p o rt ed h er e. s ta n d ar d er ro rs in p ar en th es es : * * * p < .0 1 , * * p < .0 5 , * p < .1 . 286 a. mussa et al. / financial services review 30 (2022) 273–296 whose income is greater than 100k. unemployed individuals are less likely to engage in good short-term financial behavior. this intuitively makes sense as those who are unemployed are more likely overdraft their checking account and more likely to overspend. the risk behavior of an individual can also influence their financial behavior. risky individuals are also less likely to engage in good short-term behavior. this is consistent with previous research by lyons (2007, 2008) and robb (2011), which found that risky behavior is associated with poor financial behaviors such as incurring late fees, increases in interest rates, and greater borrowing costs. both an average and good credit report is positively associated with good short-term behavior. results are similar when the three individual tech-savvy measures are used as a dependent variable. for long-term financial behavior, individuals who have a bachelor’s degree or a postgraduate degree have better behavior as compared with high school graduates. all income levels are positive and significant with long-term financial behavior against the reference category—earning less than 25k. the coefficients are also highest for those whose income is greater than 100k. indicating that those individuals who earn more can save better for retirement and long-term financial needs. unemployed individuals are less likely to have good long-term financial behavior. risk-taking individuals are more likely to have better long-term financial behavior. this could be due to their risky behavior in earning high returns on long-term investments.7 there is extensive literature on the effect of technology across generations. shobha and kumar (2020) find, from a sample of india, gen x is more financially literate, and more likely to engage in long-term financial planning compared with gen y. kumar and lim (2008) indicate generation y and baby boomers substantially vary in the perceived value and loyalty decision on mobile service. additionally, “billing services were significantly related to perceived economic and emotional value for both groups” (kumar & lim, 2008, p. 577). we examine this relationship in terms of the relevance of technology on different generations’ financial behavior. table 7 reports the marginal effects by age table 7 logistic regression predicting a measure of overall short-term financial behavior by age group variables age group igen/genz millennials/geny genx baby boomers financial knowledge 0.0449* �0.00711 0.0106 0.0395*** (0.0264) (0.0196) (0.0122) (0.0151) financial confidence 0.180*** 0.227*** 0.180*** 0.199*** (0.0409) (0.0284) (0.0151) (0.0148) financial knowledge*confidence 0.0179 �0.0230 0.0334 0.00349 (0.0555) (0.0335) (0.0229) (0.0193) tech-savvy �0.0594** �0.0345** �0.0483*** �0.0887*** (0.0251) (0.0174) (0.0104) (0.0130) pseudo r2 0.176 0.173 0.204 0.257 observations 1,326 3,112 6,849 8,438 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. a. mussa et al. / financial services review 30 (2022) 273–296 287 group, with overall short-term behavior measure as the dependent variable. the age groups included are igen/genz (18 to 24), millennials/geny (25 to 34), genx (35 to 54), and baby boomers (55+). overall financial confidence is positive and significant for all age groups. financial knowledge is significantly and positively related to the short-term financial behavior of gen z and baby boomers. regardless of the age group, overall confidence toward one’s finances is more relevant in explaining one’s good short-term financial behavior. the tech-savvy variable is still negative and significant irrespective of the age group or generation. that means being tech-savvy has an adverse effect on short-term financial behavior. we notice that the marginal coefficients for the tech-savvy variable for the younger generation or gen z and the oldest generation or baby boomers are slightly higher than the other two generation groups, gen y and gen x. this could indicate that those in the youngest generation (gen z) and the oldest generation (baby boomers) are not using technology in a way that would benefit their money management in the short-term. whereas, the millennials and gen x generation are responsible for their families and they could be relatively sensitive to their spending. table 8 reports the marginal effects of the individual tech-savvy metric variables, by age group to examine this further. mobile use for transfer variables is negative and significant in explaining gen x, and baby boomers’ short-term financial behavior. those earlier generations who use technology for mobile transfers are 4.0% to 6.7% less likely to engage in good short-term financial behavior. frequent use of mobile phones for transfers can cause overspending and overdrafts if one is not aware of their current account balance. web and app for the personal variable are also negatively and significantly related to gen z, gen x, and baby boomers’ short-term table 8 logistic regression predicting a measure of overall short-term financial behavior by age group variables age group igen/genz millennials/geny genx baby boomers financial knowledge 0.0450* �0.00747 0.0110 0.0413*** (0.0264) (0.0196) (0.0122) (0.0151) financial confidence 0.177*** 0.228*** 0.183*** 0.198*** (0.0410) (0.0285) (0.0152) (0.0147) financial knowledge*confidence 0.0183 �0.0239 0.0300 0.00126 (0.0554) (0.0336) (0.0230) (0.0193) mobile use in person 0.00394 �0.00263 �0.00698 �0.0177 (0.0254) (0.0181) (0.0111) (0.0132) mobile use for transfer �0.00239 0.000917 �0.0402*** �0.0666*** (0.0258) (0.0186) (0.0110) (0.0136) web and app for personal use �0.073*** �0.00931 �0.0228** �0.0544*** (0.0249) (0.0181) (0.0107) (0.0110) web and app use for work 0.0226 �0.0248 �0.0218 �0.0358 (0.0279) (0.0216) (0.0153) (0.0247) pseudo r2 0.178 0.172 0.207 0.260 observations 1,326 3,112 6,849 8,438 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. 288 a. mussa et al. / financial services review 30 (2022) 273–296 financial behavior. these individuals are about 2.3% to 7.3% less likely to engage in good short-term financial behaviors. this web and app use for personal use could be used for shopping which could cause overspending. similarly, table 9 reports long-term financial behavior split up by age group. similar results show up for overall financial confidence. financial knowledge is positive and significant for all age groups. as previously stated, one of the main contributions of the paper is to examine the role of technology in financial behaviors. the tech-savvy variable positively and significantly explains the igen/gen z, millennials/gen y, and gen x long-term financial behavior. the tech-savvy variable is insignificant for baby boomers. that could be because baby boomers have more likely settled their retirement or savings account because they are either already retired, or closer to their retirement age. these later generations that are using technology are 3.6% to 7.3% more likely to engage in good longterm financial behaviors, consistent with findings from shobha and kumar (2020). this could indicate the effective use of technology for retirement planning and smart savings behavior. when examining the individual technology variables, in table 10, we find the variable mobile use for transfers is positively and significantly associated with the financial behavior of igen/gen z and millennials/gen y. this indicates that later generations are using mobile phones for transfers more wisely than earlier generations. later generations are typically more comfortable with technology and are using it in a positive way for their long-term financial behaviors. web and app for personal use variables are positively and significantly related to the igen/genz, millennials/gen y, and gen x long-term financial behavior. the variable web and app use for work is positively and significantly related to igen/gen z long-term financial behavior. later generations could be using technology to help with their saving and retirement planning, while earlier generations are more likely to continue to go to a brick-and-mortar location. table 9 logistic regression predicting a measure of overall long-term financial behavior by age group variables age group igen/genz millennials/geny genx baby boomers financial knowledge 0.0471 0.0694*** 0.0771*** 0.0921*** (0.0295) (0.0184) (0.0117) (0.0103) financial confidence 0.145*** 0.177*** 0.159*** 0.145*** (0.0378) (0.0241) (0.0128) (0.0110) financial knowledge*confidence 0.0476 �0.0463 0.0185 �0.000148 (0.0487) (0.0309) (0.0200) (0.0154) tech-savvy 0.0725*** 0.0585*** 0.0361*** 0.0183 (0.0249) (0.0153) (0.00987) (0.0113) pseudo r2 0.223 0.302 0.332 0.371 observations 1,326 3,106 6,849 8,438 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. a. mussa et al. / financial services review 30 (2022) 273–296 289 5. robustness check to check the stability and robustness of our results, we made some changes to the definition of our variables of interest and the method of estimation. regardless of the changes we made for robustness checks, our findings are consistent. first, we use a tighter definition based on the frequency of use of technology to redefine the aggregate measure of the tech-savvy variable, our main variable of interest, by only including the response “use frequently” to all four questions. prior logit regression results in tables 3, 4, 5, and 6 indicate being tech-savvy is negatively and significantly related to good short-term financial behavior, and positively and significantly related to good long-term financial behavior. after redefining the aggregate measure of tech-savvy variable, we then performed the logistic regressions for short-term and long-term financial behavior by age group. for short-term financial behavior, the tech-savvy variable is still negative and significant across the majority of the age groups (see appendix table t1). this confirms that the more frequent use of technology could have an adverse impact on personal finance management. when breaking down the redefined tech-savvy variable, we find that the main driver of the negatively significant effect on short-term behavior is mobile use for transfers (see appendix table t2), which indicates individuals using technology frequently tend to overtransfer or overspend and hurt their short-term financial status. similarly, looking at the regression of long-term financial behavior by age group as indicated in appendix table t3, the relationship between being tech-savvy and long-term financial behavior is still positive and significant, except for the gen z group. table t4 in the appendix shows, among the individual measures of being tech-savvy, the web and app for personal use variable is the main driver for the positive and significant effect, which demonstrates that the more frequent table 10 logistic regression predicting a measure of overall long-term financial behavior by age group variables igen/genz millennials/geny genx baby boomers financial knowledge 0.0432 0.0678*** 0.0756*** 0.0921*** (0.0293) (0.0183) (0.0117) (0.0103) financial confidence 0.136*** 0.172*** 0.156*** 0.145*** (0.0381) (0.0242) (0.0129) (0.0110) financial knowledge*confidence 0.0569 �0.0420 0.0202 �0.000144 (0.0488) (0.0309) (0.0199) (0.0154) mobile use in person 0.00183 0.000929 �0.00785 0.0187 (0.0250) (0.0161) (0.0105) (0.0115) mobile use for transfer 0.0527** 0.0418** 0.00908 �0.00765 (0.0262) (0.0163) (0.0104) (0.0117) web and app for personal use 0.0502** 0.0603*** 0.0574*** 0.0134 (0.0246) (0.0159) (0.0101) (0.00938) web and app use for work 0.0604** 0.0165 0.0207 0.0191 (0.0270) (0.0193) (0.0150) (0.0201) pseudo r2 0.229 0.306 0.335 0.371 observations 1,326 3,106 6,849 8,438 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. 290 a. mussa et al. / financial services review 30 (2022) 273–296 use of web and app among gen y, gen x, and baby boomers tends to be more likely, compared with gen z, planning for their retirement and non-retirement investments. as an additional robustness check, we adopt a stricter definition of the variable tech-savvy as individuals who affirmed at least three out of four questions on using technology in personal finance matters. as indicated in appendix table t5, the marginal effect of being techsavvy on overall short-term financial behavior is consistently negative and significant. when using individual measures of short-term financial behavior, the marginal effect of the redefined variable tech-savvy remains dominantly negative and significant on the short-term financial behavior of spending less or equal to income and not overdrawing the account, except for the behavior of paying credit in full. by the same token, as indicated in appendix table t6, the results of the logit regression on the overall long-term financial behavior demonstrate the marginal effect of the more strictly defined variable tech-savvy remains positive and significant on the long-term financial behavior. additionally, the positive marginal effect of tech-savvy stays positive across each individual measure of long-term financial behavior. to ensure tech-savvy variable affects financial behavior regardless of how the variable is defined, ordinary linear regression is reconsidered. tech-savvy variable is defined as a continuous variable, (ranging from 0 to 4, with 0 being not tech-savvy at all, and 4 being the most tech-savvy), instead of a binary variable. for interpretation purposes, the value of techsavvy is standardized in the regression. as indicated in appendix table t7, the regression results demonstrate a consistent finding: the variable tech-savvy is negatively and significantly related to short-term financial behavior across the regressions of using overall or individual measures, echoing previous results. the results in appendix table t8 show techsavvy is positively and significantly related to positive long-term financial behavior across the overall and individual measures, except for having an emergency fund. to further investigate the tangible impact of tech-savvy on individuals engaging in different levels of good short-term financial behavior, we adopt an ordered logit model. as indicated in appendix table t9 in the appendix, consistent with the earlier results, the net marginal effects show that, on average, the most tech-savvy individuals are 4.8% (0.0198– 0.0681) more likely, than non-tech–savvy individuals, to engage in adverse short-term financial behavior. similarly, on average, the most tech-savvy individuals are 1.1% (�0.0302 + 0.0196) more likely to engage in good long-term financial behavior. second, we redefined the dependent variables—both short-term and long-term behavior as an additional robustness check. while we adopt a loosened criterion to examine individuals with at least two out of three positive short-term financial behaviors, the variable tech-savvy is negatively and significantly associated with short-term financial behavior (results available upon request), which is consistent with our previous findings. additionally, the negative impact of being tech-savvy seems to be smaller (�0.045 vs.�0.064) when individuals engage in fewer good short-term financial behaviors. similarly, the binary variable of long-term financial behavior is redefined as individuals with at least four out of five good long-term financial behaviors, compared with the previous definition of at least three good long-term financial behaviors. again, the logit regression of redefined long-term financial behavior indicates the variable tech-savvy shows a positive and significant impact on the long-term financial behavior (0.0151 vs. 0.0258), which is consistent with previous findings. a. mussa et al. / financial services review 30 (2022) 273–296 291 third, we address the concern of the potential endogeneity of financial literacy as a control variable in financial behavior regression (lusardi & mitchell, 2014; van rooij et al., 2009): financial knowledge may predict financial behavior, and financial behavior may also predict financial literacy or knowledge. given the limitation of the nfcs dataset, it is difficult to do two-stage regressions with the instruments as in the previous literature (lusardi & mitchell, 2014). however, as wagner and walstad (2019), we estimated both the short-term and long-term financial behavior models with and without financial education variables. the marginal effects of our main variable of interest—tech-savvy—do not change both statistically and economically. as shown in appendix table t10, the marginal effect of being tech-savvy on the overall short-term financial behavior is consistently negative and significant before and after controlling for financial education and/or financial confidence variables. similarly, a consistently positive and significant marginal effect of being tech-savvy is found on the overall long-term financial behavior before and after controlling for financial education and/or financial confidence variables. finally, we restrict the sample when regressing short-term financial behavior into two groups, those with high/strong and low/weak long-term financial behavior. the same is done for long-term financial behavior restricted into high and low short-term financial behavior. results are consistent in the restricted regressions. tech-savvy is negatively and significantly associated with short-term financial behavior, while it is positively and significantly associated with long-term financial behavior. we also performed an ordered logit to examine the marginal effects which represent the net effects of the lowest and highest financial behavior categories. this again confirms our previous results with the net marginal effect for tech-savvy being �0.092 short-term financial behavior for the sample restricted to high long-term financial behavior and �0.005 for the sample restricted to low long-term financial behavior. when regressing long-term financial behavior, the net marginal effects of the variable tech-savvy are 0.032 and 0.0151 for the restricted samples on high short-term financial behavior and low short-term financial behavior, respectively (see appendix table t11). 6. conclusion and future discussion the study examined how being tech-savvy affects personal short-term and long-term financial capability. the tech-savvy variable was consistently negative and significant in the short-term model. however, the tech-savvy variable was positive and significant in the longterm model. this result could indicate that technology allows more risk-tolerant individuals to behave imprudently in the short-term, making it easy for individuals to overspend. however, over the long-term, individuals may engage in more rational long-term financial behavior using technology. this paper also examined how financial capability and technology savviness can vary across generations. we found being equipped with good financial knowledge favors gen z and baby boomers to engage in positive short-term financial behavior and being financially confident is positively associated with good short-term financial behavior across all generations. our variable tech-savvy was found negative and significant across all generations for 292 a. mussa et al. / financial services review 30 (2022) 273–296 short-term financial behavior. this could indicate that those in the youngest generations and oldest generations are not using technology in a way that would benefit their money management in the short-term. tech-savvy individuals in gen z, millennials, and gen x are more likely to engage in good long-term financial behavior. finally, this paper supports similar research on financial capability with updated national data. our findings echo the prior research that people with better than average actual financial knowledge or financial confidence tend to be more likely to engage in good short-term and long-term financial behavior. one of the limitations noted is related to using self-reported survey data to measure tech-savvy and financial behaviors. it is difficult to confirm whether people behave in a certain way as they respond to financial behavior-related questions. the variable techsavvy is constructed based on the general response to whether the respondents use their mobile or web app to purchase things or transfer money. it would have been more relevant if the survey asked if the respondents used their mobile phone or an app to manage their finances. plus, the nfcs data are not rich enough to specify the sources of the financial background of the respondents. in future research, we hope to collect data and focus on a college student population. looking into different demographics such as firstgeneration college students, commuters, and students who are working part-time, or full-time jobs will provide policy implications on personal financial management among young generations. notes 1 among the survey participants who used bnpl, 29% indicated they would have spent less, or not spent at all without the financing option (mckinsey digital payments consumer survey, 2021). 2 objective financial knowledge questions assess true skills and knowledge pertaining to personal financial decisions while subjective financial knowledge questions evaluate self-perception on dealing with personal financial matters. 3 the conventional age cut off in defining generation is slightly different in our sample. pew research center defines generation as follows: gen z (born after 1996), gen y (born 1981–1996), gen x (1965–1980), and baby boomers (1946–1964). 4 the responses to each of the six questions were re-coded, as a dummy variable, “1” for a correct response and “0” otherwise. the overall score was computed by adding the binary numbers attached to the six questions that ranged 0 to 6, 0 being none of the six questions were answered correctly and 6 being all were responded correctly. the binary variable of financial knowledge, then, was created by assigning “1” if the overall score is above the average score, 2.11, and “0” if not. 5 on average, respondents rated their current financial condition as 5.72 (scaled from 1, not at all satisfied, to 10, extremely satisfied), overall financial knowledge as 5.13 (scaled from 1, very low, to 7, very high), and day-to-day financial matters as 5.76 (scaled 1, strongly disagree to 10, strongly agree). then, a binary variable as a measure of composite financial confidence was created by adding all the binary responses to the a. mussa et al. / financial services review 30 (2022) 273–296 293 three questions and coded as “1” if the overall score is higher than the average score, and “0” if not (see table 3 for the detail statistics). 6 (9) financial knowledge = 0.124–0.099*0.51726 = 0.072; financial confidence = 0.0825–0.099*0.5109 = 0.0319 (10) financial knowledge = 0.122–0.0898*0.51726 = 0.076; financial confidence = 0.0794–0.0898*0.5109 = 0.034 7 the detailed estimation results for these demographic and socio-economic variables will be available upon request. appendix table a1 logistic regression predicting a measure of overall short-term financial behavior by age group variables age group igen/genz millennials/geny genx baby boomers financial knowledge 0.0422 �0.00818 0.0112 0.0377** (0.0264) (0.0196) (0.0122) (0.0151) financial confidence 0.183*** 0.225*** 0.178*** 0.198*** (0.0411) (0.0284) (0.0151) (0.0148) financial knowledge*confidence 0.0221 �0.0224 0.0343 0.00596 (0.0556) (0.0336) (0.0230) (0.0193) tech-savvy �0.0463* �0.0193 �0.0210* �0.0730*** (0.0242) (0.0169) (0.0116) (0.0154) pseudo r2 0.174 0.268 0.202 0.255 observations 1,326 3,112 6,849 8,438 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. 294 a. mussa et al. / financial services review 30 (2022) 273–296 references asaad, c. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101-117. berraies, s., ben yahia, k., & hannachi, m. (2017). identifying the effects of perceived values of mobile banking applications on customers. international journal of bank marketing, 35, 1018–1038. https://doi.org/10.1108/ ijbm-09-2016-0137 beverly, s., hilgert, m., & hogarth, j. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, july, 309–322. carlsson, h., larsson, s., svensson, l., & åström, f. (2017). consumer credit behavior in the digital context: a bibliometric analysis and literature review. journal of financial counseling and planning, 28, 76–94. https://doi.org/10.1891/1052-3073.28.1.76 chan, e., & saqib, n. (2015). online social networking increases financial risk taking. apasia-pacific advances in consumer research, 11, 105. dhanapal, s., vashu, d., & subramaniam, t. (2015). perceptions on the challenges of online purchasing: a study from ‘baby boomers’, generation ‘x’ and generation ‘y’ point of views. contadurı́a y administración, 60, 107–132. https://doi.org/10.1016/j.cya.2015.08.003 dell, k., & nestoriak, n. (2020). assessing the impact of new technologies on the labor market: key constructs, gaps, and data collection strategies for the bureau of labor statistics. available at https:// www.bls.gov/bls/congressional-reports/assessing-the-impact-of-new-technologies-on-the-labor-market.htm de meza, d., irlenbusch, b., & reyniers, d. (2008). financial capability: a behavioral economics perspective. consumer research, 69, 1–108. federal reserve board. (2016). consumers and mobile financial services. available at https://www. federalreserve.gov/econresdata/consumers-and-mobile-financial-services-report-201603.pdf federal student act q2. (2009). available at https://studentaid.gov/data-center/student/portfolio garrett, j., & rodermund, r., & anderson, n., & berkowitz, s., & robb, c. (2014). adoption of mobile payment technology by consumers. family and consumer sciences research journal, 42, 358–368. https://doi.org/ 10.1111/fcsr.12069 table a2 logistic regression predicting a measure of overall short-term financial behavior by age group variables igen/genz millennials/geny genx baby boomers financial knowledge 0.0416 �0.00805 0.0113 0.0381** (0.0265) (0.0196) (0.0122) (0.0151) financial confidence 0.174*** 0.231*** 0.182*** 0.198*** (0.0413) (0.0286) (0.0152) (0.0148) financial knowledge*confidence 0.0219 �0.0267 0.0298 0.00513 (0.0555) (0.0337) (0.0230) (0.0193) mobile use in person 0.0668* �0.000904 0.00223 �0.0136 (0.0354) (0.0233) (0.0163) (0.0245) mobile use for transfer �0.0600** �0.0285 �0.0323* �0.104*** (0.0297) (0.0221) (0.0183) (0.0325) web and app for personal use �0.0440 �0.0102 �0.0244 �0.0440** (0.0326) (0.0218) (0.0164) (0.0220) web and app use for work 0.0463 �0.0364 �0.0336 �0.100** (0.0436) (0.0289) (0.0254) (0.0510) pseudo r2 0.177 0.318 0.203 0.255 observations 1,326 3,112 6,849 8,438 note. the marginal coefficients of other control variables (gender, education level, employment status, marital status, income, race, risk level measures, and credit score) are included in the regression, not reported here. standard errors in parentheses: ***p < .01, **p < .05, *p < .1. a. mussa et al. / financial services review 30 (2022) 273–296 295 hee yeo, j., & fisher, p. (2017). mobile financial technology and consumers’ financial capability in the united states. journal of education and social policy, 7, 80–93. henager, r., & cude, b. j. (2016). financial literacy and longand short-term financial behavior in different age groups. journal of financial counseling and planning, 27, 3–19. https://doi.org/10.1891/1052-3073.27.1.3 hogarth, j., & anguelov, c. e. (2004). are families who use e-banking better financial managers? journal of financial counseling and planning, 15, 20551202–785. kumar, a., & lim, h. (2008). age differences in mobile service perceptions: comparison of generation y and baby boomers. journal of services marketing, 22, 568-577. https://doi.org/10.1108/08876040810909695 liébana-cabanillas, f., sánchez-fernández, j., & muñoz-leiva, f. (2014). antecedents of the adoption of the new mobile payment systems: the moderating effect of age. computers in human behavior, 35, 464–478. https://doi.org/10.1016/j.chb.2014.03.022 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. https://doi.org/10.1257/jel.52.1.5 lyons, a. c. (2007). credit practices and financial education needs of midwest college students. networks financial institute working paper no. 2007-wp-23. available at https://ssrn.com/abstract=1060801 lyons, a. c. (2008). risky credit card behavior of college students. in j. j. xiao (eds), handbook of consumer finance research (pp. 185-207). new york, ny: springer. mckinsey digital payments consumer survey. (2021). available at https://www.mckinsey.com/industries/ financial-services/our-insights/banking-matters/new-trends-in-us-consumer-digital-payments nfcs. (2018). the state of u.s. financial capability. the 2018 national financial capability study. finra. available at https://www.usfinancialcapability.org/downloads/nfcs_2018_report_natl_findings.pdf phang, c., sutanto, j., kankanhalli, a., li, y., tan, b., & teo, h. (2006). senior citizens’ acceptance of information systems: a study in the context of e-government services. ieee transactions on engineering management, 53, 555–569. https://doi.org/10.1109/tem.2006.883710 robb, c. a. (2011). financial knowledge and credit card behavior of college students. journal of family and economic issues, 32, 690-698. https://doi.org/10.1007/s10834-011-9259-y shobha, t. s., & sharath kumar, k. m. (2020). financial behaviors of generation x and generation y. scms journal of indian management, 17, 123–136. van rooij, m., lusardi, a., & alessie, r. (2009). financial literacy and stock market participation. ssrn electronic journal. available at https://ssrn.com/abstract=1476958. https://doi.org/10.2139/ssrn.1476958 visual capitalist. (2021). generational power index. available at https://www.visualcapitalist.com/gpi/ wagner, j., & walstad, w. b. (2019). the effects of financial education on short-term and long-term financial behaviors. journal of consumer affairs, 53, 234–259. https://doi.org/10.1111/joca.12210 walsh, b., & lim, h. n. (2020). millennials’ adoption of personal financial management (pfm) technology and financial behavior. financial planning review, 3, e1095. https://doi.org/10.1002/cfp2.1095 wilcox, k., & stephen, a. t. (2012). are close friends the enemy? online social networks, narcissism, and self-control. psycextra dataset. xiao, j. j., & porto, n. (2017). financial education and financial satisfaction. international journal of bank marketing, 35, 805–817. https://doi.org/10.1108/ijbm-01-2016-0009 296 a. mussa et al. / financial services review 30 (2022) 273–296 enumerating the value of financial advice in a competitive market – a dual structure approach and analysis† steve p. frasera,*, brian c. payneb, scott schatzlec alutgert college of business, florida gulf coast university, 10501 fgcu boulevard south, fort myers, fl 33965, usa bcollege of business administration, university of nebraska at omaha, 6708 pine street, mammel hall 228g, omaha, ne 68182, usa cmutual trust advisory group, 23190 fashion drive p203, estero, fl 333928, usa abstract this paper introduces and examines a composite, dual fee structure (cdfs) for financial planners that helps quantify the value of financial advice. our structure specifically separates financial planning (advice) fees based on total net worth (nw) from investment management (im) fees based on assets under management (aum), which are readily observable and pervasive in the marketplace. doing so facilitates quantifying the value of this financial advice in a competitive market setting. knowing the financial value of the non-im component of financial planning services can reduce perceived conflicts of interest by permitting financial planners to generate compensation for non-im planning activities in a transparent manner, whether or not the client moves investable funds to the planner. © 2021 academy of financial services. all rights reserved. jel classification: g29 keywords: financial planning; advisory fees 1. introduction the financial services industry continues to be the nexus of ongoing discussions among federal and state regulators, credentialing entities, and financial institutions, concerning *corresponding author. tel.: +1-239-590-7336. †please do not quote or distribute. e-mail address: sfraser@fgcu.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 227–245 what constitutes the fiduciary relationship between providers and clients. underlying this dialogue is the nature of the compensation paid to the agent by the client—commission, flat fee, fee-only, fee-based, or some combination of approaches (opiela 2006 and mackillop 2017). the predominant methodology for many registered investment advisors (rias) is fee-only, where agents charge a percentage on the level of clients’ assets under management (aum). many of these rias focus primarily on the function of investment management (im), in which case fees are more directly correlated with aum, but also include planning services for the broader portfolio. this is an oft used argument suggesting the interests of rias are aligned with their clients’ as both parties gain or lose value together. this model is certainly not absolute, and a common example of where this model may have a potential conflict is when a planner might recommend a client use aum to pay off a mortgage. the right recommendation might be to pay off the debt, yet using aum to do so will reduce advisor compensation in this scenario since the aum declines by the amount of the mortgage debt paid off. in contrast, many agents working at broker-dealers (bd) are paid on commission, as is also the case with those selling stocks, bonds, or insurance and annuity products. while this latter approach may indeed be a lower-cost approach for some clients, it is not necessarily clear if and whether there is a line between fees associated with completing a transaction and fees paid for financial planning or advice. the further alternative case where professionals are “dual-registered” as both rias and bds certainly does not clarify the situation (haslem 2010). no matter the compensation approach, research suggests that advisors do not always act in clients’ best interests. hoechle et al (2018), in a study of financial advisors working for banks, suggest advisors recommend transactions that are most profitable for the bank, and that independent clients performed better than advised clients. similarly, egan (2019) reports the incentives of brokers do not align with clients. perhaps most alarming, cheng and kalenkoski (2018) report that more than 20% of clients have no idea how their advisors are paid. the purpose of this research is not to solve the fiduciary debate. instead, we seek to examine a potential methodology to determine, or price, the value of financial planning.1 more specifically, we seek to separate the value of financial advice from the presumed cost of managing an investment portfolio. knowing the value of this advice in a competitive market is critical to provide greater transparency to clients and help planners understand the value of the expertise and service they provide. in this paper, we introduce and examine a composite, dual-fee structure (cdfs) for financial planners. we hypothesize that pricing standard investment management activities competitively within the market permits the creation of a reasonable pricing strategy for planning services. our structure specifically separates financial planning (advice) fees from im fees that are often subsumed in a single aum fee. this approach may reduce or perhaps limit perceived potential conflicts of interest by providing a mechanism where financial planners can be compensated for planning efforts whether or not the client moves investable funds to the planner for the planner to manage. 228 s. p. fraser et al. / financial services review 29 (2021) 227–245 2. previous research there is limited research that specifically addresses advisor fees outside of the impact of fees on investor returns. it is well documented that higher management fees are associated with lower investor returns (fama & french, 2010). fees and investor returns are naturally a zero-sum game: a dollar paid in fees is a dollar less that investors receive in returns (sec, 2014). however, there is some literature that addresses what clients might be looking for from financial planners, as well as some survey work sampling the various fee structures used by advisors of different types. we review some of the salient literature in each space next. bae and sandager (1997) examine characteristics that consumers sought from financial planners. they find that clients primarily seek advice on retirement funding, investment growth, and reducing taxes. most simply want a comprehensive review of their situation. furthermore, they find only 20% of survey respondents preferred a planner be compensated by commissions from sales exclusively. statman (2000) suggests financial planners are investor managers, and they focus too little on the value they contribute as managing investors and too much on clouding the fees they charge for those contributions. finke, huston, and winchester (2011) find that wealth is the strongest predictor in the decision to pay for financial advice. they find those with at least a college degree are more likely to hire an expert; however, those who perceive they have a better understanding of financial issues are less likely to pay for financial advice. seay et al (2017) suggests clients with different characteristics (e.g. demographic and income levels) may align themselves with different advisor compensation structures. cheng and kalenkoski (2018) survey investors to ascertain how their advisors are compensated and find 27% of clients perceive their advisors are compensated by charging a percentage of investable assets, 16% commissions, 18% some combination of fees and commissions, and 15% a flat/hourly fee. restating their startling result; more than 20% respond they had no idea how their advisors are paid. in summary, investors overwhelmingly do not know how, or how much, they are paying advisors. particularly relevant for this research, it is unlikely clients know for what they are paying—financial advice or transaction costs? lahtinen and shipe (2018) review adv data from 2009 to 2015 and find virtually all portfolio management investment companies charge a percentage of aum while financial planning services companies use fixed and hourly fees most often. the authors conclude that fee structures are not homogeneous and vary depending on the services offered by the firms. kitces (2017) suggests the average percentage of aum charged may be 1%, but that the median fee for high-net-worth clients is closer to 0.50%. furthermore, he reports the results of an inside information survey suggesting that the all-in fee, or the total fee that includes transaction costs as well as the costs of underlying products, is closer to 1.65%. in summary, it is perhaps not surprising that with the complexity of fee structures found across the financial services profession, not only do investors not know how their advisors are compensated, they surely are not likely to understand the value of financial planning relative to investment management activities. s. p. fraser et al. / financial services review 29 (2021) 227–245 229 3. methodology and data we initially outline and describe a basic dual-fee structure before we discuss how such a structure might be utilized or optimized. the purpose of this composite structure is to both value, and potentially allow an advisor to charge for, financial planning advice separately from any im fees that may be collected for managing aum (or any commissions earned associated with selling a particular product). the goal is to develop an approach and structure such that the fee schedule adequately compensates the planner for the value provided to the client and separates this advice charge from the broader aum fee. under many fee-only mechanisms where fees are charged as a percentage of aum, the assumption is the greater the amount of assets, the greater the scope of planning work. while there is likely some positive correlation between investable wealth and planning requirements, the relationship is not always so straightforward. two clients with similar levels of total wealth might have very different planning needs. here we address this challenge by distinguishing between total wealth, and investable wealth. we use a client’s net worth (nw) as a proxy for their total wealth, which we posit is a better representation for the financial planning effort associated with the client’s financial situation than simply the level of the client’s investable assets, captured as aum. in our introductory model. we consider the asset universe to include the following asset classes: stocks (s), bonds (b), real estate (re), and business ownership (proxied here by private equity, pe). the sum of these assets represents a client’s nw at time “t” as shown in eq. (1). for simplicity purposes, we assume a client’s investable assets are equal to aum, modeled here as simply containing stocks and bonds (eq. 2). the non-aum portion of nw is modeled by asset classes likely held by high net worth clients. specifically, this portion of the portfolio includes additional assets, such as a home (real estate, re) and portion of a business (private equity, pe). when we substitute eq. (2) in eq. (1), we get the final result in eq. (3). nwt ¼ st þ bt þ ret þ pet eq. (1) aumt ¼ st þ bt eq. (2) nwt ¼ aumt þ ret þ pet eq. (3) each of the nw and aum portfolios will increase or decrease based on the respective returns of the underlying asset classes (rx), where “x” represents the underlying asset class. the value of the portfolios at time “t + 1” is found by eq. (4) for aum and eq. (5) for nw. aumtþ1 ¼ st 1 þ rs,tð þ þ bt 1 þ rb,tð þ eq. (4) nwtþ 1 ¼ aumtþ1 þ ret 1 þ rre,tð þ þ pet 1 þ rpe,tð þ eq. (5) the focus of this analysis is on the fees associated with these portfolios. we first model the traditional, single, fee-only approach used by many advisors in eq. (6): 230 s. p. fraser et al. / financial services review 29 (2021) 227–245 aum feet ¼ ojv jaumj,t eq. (6) where j represents the number of tiers in the (regressive) fee tier structure and v j is the fee level in tier j. advisors using this model charge a single fee, v j, which covers services associated with their investment management function as well as some level of financial planning service. similarly, we can model the cdfs fee in component parts as follows: cdfs feet ¼ okw kimk,t þ olu lnwl,t eq. (7) where k and l are analogous to j, and w and u permit differential fee levels for the im and nw components. it is important to note the value of im in eq. (7) is equivalent to the level of aum in eq. (6). we change the identifying variable name to highlight that the im fee, w k, in eq. (7) is different from the single fee, v j, in eq. (6). while both are charged as percentage of aum, the im fee, w k, used in eq. (7) represents the investment management function only. assuming a financial planner provides the same services under either fee structure, in a competitive market it must be true that the fees are equal as depicted in eq. (8). aum feet ¼ cdfs feet eq. (8) ojv jaumj,t ¼ oj w jimj,t þ u jnwj,t � � eq. (9) substituting eqs. (6) and (7) into eq. (8) yields eq. (9). we can currently approximate v from what is observed in marketplace by fee-only planners. we can also approximate w using so-called robo advisor fees, which arguably represent the latest innovation for providing the most basic investment management functions. the resulting question is then the determination of the fee schedule for u . doing so represents the mechanism for quantifying the value of financial advice using this approach. if we allow v , w , and u to be a single weighted fee value representing a multitiered, regressive fee schedule as detailed later, then eq. (9) leads to the following for any given year t. u ¼ v � wð þaum aum þ re þ pe eq. (10) in portfolios consisting solely of stocks and bonds, eq. (10) shows the value of financial advice is intuitive, and captured simply as the difference between robo-advisor (im) fees and the (observable) aum-only fee, or u ¼ v � w . however, when other assets—such as real estate (re) or business ownership (pe)—become part of the broader portfolio, nw increases and the value of u is no longer as static and straightforward. in fact, there are as many unique solutions for u as there are unique client portfolios. additionally, as implemented later in this analysis, u can adjust over time due to the front-loaded nature of the planning function.2 s. p. fraser et al. / financial services review 29 (2021) 227–245 231 while perhaps true, it is not practical to conclude that the value of financial advice varies infinitely with each individual portfolio. an advisor would spend far too much time creating fee schedules if she tailored each one according to individual clients’ asset mixes. instead, it is appropriate to base u on some characterization of averages or expected outcomes over time. due to the countless values for the nw fee schedule u , we conduct a simulation to help us quantify one example of such an acceptable schedule. doing so in turn helps us calculate the value of financial advice in a competitive marketplace. the nw and im fees we introduce in the cdfs both follow a parallel, similar regressive structure. much like many fee-only or fee-based charges based on aum, both fees here are charged at lower marginal rates as the benchmark (i.e., aum and nw) rises. the rate charged for im increases at various breakpoints of aum. for example, the im fee might be 50 basis points (bps) on the first $250,000 of aum and decrease as aum levels rise. this rate is assumed to be charged and collected in perpetuity if aum is under the advisor’s care. doing so clearly aligns client and planner interests by benefiting both when the portfolio grows, but it also recognizes the decreasing marginal effort for the advisor as the portfolio grows larger. the nw fee is similarly structured, but instead based on the client’s net worth. we further suggest that the effort involved in financial planning is not uniform over time. for financial advisors who practice comprehensive financial planning, the initial workload involved with a new client is significantly more demanding in the first year as the planner develops a way forward for a client across all the financial planning areas (e.g., cash flows and debt service, risk and insurance, investment, tax, and estate planning). to account for this non-linear workflow, we assume the full initial financial planning fee is charged the first year at the scheduled rate. subsequent years are charged at a reduced rate; this analysis uses a rate that is one-half the bps rate used the first year in our initial illustration. this approach allows the client to pay directly for financial planning and to see how much they are paying for that support initially (and over time if they choose to do so) while remaining independent of im fees. the percentage reduction of the planning fee during subsequent time periods is just one “lever” that can be adjusted by individual planners/firms based on the level and complexity of planning services offered. this decomposition of fees also allows the planner to charge a lower, more market competitive im-only fee. our simulation sets the value of financial planning (the nw fee, or u j) as the delta between the “all-in,” single, aum model fee (v j), and the emerging robo-advisor im-only fee (w j). we evaluate the impact of these assumptions over time, and as importantly, the analysis here identifies these “levers” that firms and advisors may adjust for their specific practices. in addition to the degree of the reduction of the nw fee in subsequent years, an additional lever is the determination of the level of the individual fees (and breakpoints) charged for each component of the cdfs. table 1 shows the nw, im, and aum breakpoints and regressive fee structure, respectively, for both the dual fee cdfs (that includes nw and im) and the single fee-only (aum) structures analyzed here. the setup is basic and straightforward, yet it provides sufficient insights that generalize to fee structures with more complexity (e.g., breakpoints). the values for the im component recognize that in a competitive market there is a base cost for advisors to profitably manage investable assets. fortunately, the advent of so-called roboadvisors have revealed that these costs can be relatively low. a brief survey of multiple 232 s. p. fraser et al. / financial services review 29 (2021) 227–245 robo-advisors shows a range of fees from $0 (“free”) to almost 90 bps annually. there are many factors in play when setting these fees. for example, charles schwab inc., which offers free advisory services, mandates its recommended portfolios contain non-trivial cash positions that the firm then uses to generate “fees” from the spread between the rates paid on this cash and what the firm can charge to lend the funds. brenner and meyll (2020) suggest robo-advisors are a valid alternative for investment advice. overall, the non-scientific middle of the range for im-only fees appears to be approximately 40 bps. we subjectively assess an additional 10 bps on smaller investment portfolios (below $250,000) to recognize the lack of scale for a typical individual planner, which is supported by uhl and rohner (2018). per the regressive approach, we reduce the bps fees by 10 bps as the im portfolio grows above $2 million and has the lowest marginal im fee at 30 bps for this example. in all cases the “fee on max” value is simply the bps fee multiplied by the maximum portfolio value in that row, added to any previous fee on max value. having generated this im fee structure based on a rather objective approach in terms of market competition and economies of scale, we can now deductively generate the fee structure for the planning fee (nw component). we use this fee structure to inform us about the market value of providing financial planning and advice to clients. the fee structure shown in table 1 provides reasonably comparable overall initial fees for clients when controlling for the complexity of the planning activities. a common refrain of financial planning is recognizing that all clients are unique. it is quite possible, and even more likely, that there might be significant variation among clients’ relative values of net worth (nw) and investable assets, or aum. the size and scope of clients’ aum and nw is the focus of further analysis, a third potential “lever.” if a client’s aum represents their entire nw, say early in their career when their net worth may consist solely of assets in a 401(k) plan, the breakdown of the planning fee versus im fee may likely table 1 fee breakpoints panel a dual fee structure net worth (nw) component* plus investment management (im) component minimum maximum fee (bps) fee on max minimum maximum fee (bps) fee on max $ — $ 500,000 60 $ 3,000 $ — $ 250,000 50 $ 1,250 $ 500,000 $ 2,500,000 50 $ 13,000 $ 250,000 $ 2,000,000 40 $ 8,250 $ 2,500,000 $ — 40 $2,000,000 $ — 30 panel b fee-only (im-only) minimum maximum bps fee fee on max $ — $ 250,000 120 $ 3,000 $ 250,000 $ 2,000,000 90 $ 18,750 $ 2,000,000 $ — 60 *note the nw fee is reduced to 50% of the year 1 nw fee due to upfront planning work. this table depicts the parallel, regressive fee structures used in the simulation. panel a depicts the two components of the consolidated dual-fee structure (cdfs) and panel b depicts the single, and more widely used asset under management (aum)-based fee. s. p. fraser et al. / financial services review 29 (2021) 227–245 233 t ab le 2 t ar g et v er su s si m u la ti o n st at is ti cs (1 0 -y ea r h o ri zo n ) r et u rn s (a n n u al ) r is k (a n n u al ) c o rr el at io n s (d es ir ed /s im u la te d ) a ss et cl as s d es ir ed s im u la te d d es ir ed s im u la te d b o n d r ea l es ta te p ri v at e eq u it y s to ck 5 .6 0 % 5 .5 2 % 1 4 .3 0 % 1 3 .9 5 % 0 .0 0 0 /� 0 .0 0 4 0 .5 3 /0 .5 1 0 .7 3 /0 .7 1 b o n d 3 .1 0 % 3 .1 0 % 3 .4 2 % 3 .3 3 % �0 .1 9 /� 0 .1 8 �0 .2 3 /� 0 .2 2 r ea l es ta te 5 .8 0 % 5 .7 9 % 1 1 .0 7 % 1 0 .7 7 % 0 .4 9 /0 .4 7 p ri v at e eq u it y 8 .8 0 % 8 .7 6 % 2 0 .1 7 % 1 9 .7 2 % m o n te c ar lo ru n s 1 0 ,0 0 0 t h is ta b le sh o w s th e ta rg et an d si m u la te d v al u es fo r ri sk an d re tu rn in p u ts u se d in th e m o n te c ar lo si m u la ti o n . r es u lt s ar e sh o w n fo r al l as se t cl as se s u se d in th e si m u la ti o n as w el l as th e co rr el at io n o f th o se in p u ts . t ar g et v al u es co m e fr o m jp m o rg an 2 0 2 0 lo n g -t er m ca p it al m ar k et fo re ca st . 234 s. p. fraser et al. / financial services review 29 (2021) 227–245 differ from that of a client where the level of aum is significantly less than their nw, for instance, as they approach or have entered the retirement phase and own real estate and/or a small business interest. the ratio of nw to aum will serve as a proxy for the delta in the im fee versus financial planning fee in this analysis. in other words, as the aum decreases relative to total nw, a planning fee schedule becomes more relevant than an im fee schedule, causing the planning fee to increase and the im fee to decrease. to see these impacts, we investigate the variance of planning and im fees for three clients: a young client in the accumulation phase, a near-retirement client with some remaining work years but planning for the transition to retirement, and a retired client in the spending phase. we then add three levels of the relationship that cover this spectrum of the ratio between aum and nw: one where nw equals the level of aum, one where nw is 1.5x the amount of aum, and finally where nw is 2x the client’s aum. our quantitative approach involves simulating the dual fee structure properties over the relevant time period for each of our clients: 40 years for the young client (accumulation phase), 20 years for the client transitioning to retirement (transition phase), and 10 years for the retired client (spending phase). essentially, we construct our cdfs by first setting the im fee component comparable to robo-advisors as discussed above. we then consider a first-year planning fee that when added to the im fee, is equivalent to the typical aum fee found in the industry. we then run a simulation to examine the behavior of the component fees over various investment periods, recalling that the planning fee is adjusted downward in year two in our initial illustration. running the simulation over time requires some further assumptions: • recall a clients’ total wealth is represented by their net worth (nw). • to recognize the positive relationship between investment horizon and portfolio riskiness, the initial asset allocation of the im portion of the client’s nw is assumed to differ among these three profiles. for the im component of the portfolio we generate a generic stock/bond allocation of 80/ 20, 60/40, and 40/60 for our respective clients. • a client’s nw is comprised of aum assets (modeled with stocks and bonds and denoted by im) and other assets. these other assets are modeled to be initially split evenly between real estate (re) and private equity (pe) asset classes (representing home ownership and small business interests, respectively). • portfolios are rebalanced annually back to the original weights, but only for stocks and bonds. because real estate and private equity are less liquid and likely held longer, they are not rebalanced but instead left to “drift” with their associated returns.3 to calculate fees, we simulate returns of the four asset classes using the j.p morgan (jpm) 2020 market forecast values for returns, risk, and pairwise correlation between these asset classes. • to ensure proper return co-movement, we utilize a cholesky decomposition to generate the respective asset class returns over time. table 2 shows the target return, risk, and correlation values for the asset classes investigated here over a 10-year period, as well as the distribution of the mean values for these statistical properties over the relevant time periods using 10,000 monte carlo runs.4 again, the intent of our analysis is not to generate a debate about the specific fee levels or breakpoints selected; we use what we believe to be within the range of fees generally representative of the industry. instead, the focus of this investigation is to introduce a framework s. p. fraser et al. / financial services review 29 (2021) 227–245 235 and mechanism for truly valuing the financial advising expertise that planners provide to clients, separate from the oft-used single aum fee that likely subsumes a financial planning effort, and observe the behavior of the component fees over time. 4. results in this section we report the results of our dual fee analysis both initially, and over a longer investing period. table 3 shows the impacts of the fee breakpoints and values we depict in table 1, as a snapshot at the initial planning point in time. specifically, based on the simple u structure proposed, it quantifies the annual fees clients would pay (and planners would receive) under the cdfs (nw and im) versus the fee-only structure (aum) in those respective rows per eqs. (6) and (7). once again, while client profiles are literally limitless, we present three different profiles for illustration purposes: those in preretirement (accumulation), transition (approaching retirement soon), and spending (in retirement) phases. within each of those profiles, we further analyze three different ratios of nw to im for a total of nine potential client scenarios. as one example, client a is assumed to be in the accumulation phase with a nw of $600,000. we then further delineate this nw as follows, recalling that aum in the cdfs is designated as im: nw = im = $600,000 (column 1), nw ($600,000) = 1.5 � im ($400,000) (column 2), and nw ($600,000) = 2 � im ($300,000) (column 3). the portion of the nw portfolio not comprised of im is an evenly allocated between the real estate (re) and private equity (pe) asset classes. we contend— with support from practicing planners—that the im effort is generally the same among all of these client profiles; however, we also contend that of these three client profiles, the planning effort is likely most complex for clients a3, b3, and c3 as these clients have asset classes not composed of stocks and bonds that the planner must consider when developing a holistic financial plan. along similar lines, clients b and c have higher nw, but again we offer the possibility that their net worth can have varied compositions in terms of stocks, bonds, real estate, and private equity analogous to the ratios described for client a. within clients b and c, we again contend that the planning effort increases as the aum to nw ratio decreases. what table 3 demonstrates is that by assessing reasonable (and consistent) im fee structure or u , under a dual-fee structure and also for a fee-only (aum) structure, v , as in table 1, it is possible to come up with the financial planning fee component of the cdfs for the nw portion of the portfolio. we posit this financial planning fee serves as a proxy for the value of financial advice. specifically, the differences among the cdfs are almost negligible in relation to the all nine clients’ overall nw (see the penultimate row entitled “as % of nw” for the difference in the cdfs fee vs. fee-only fee). notably, these percentages monotonically creep upward as the sub client groups’ investable assets diminish as a fraction of net worth. that is, as nw held outside of stocks and bonds increases, so do the relative fees, at least initially. we think this is reasonable, as increased non-investment assets (assumed here to be real estate and private business interests) require additional planning efforts, all else equal, which the planner should be compensated for as part of the initial 236 s. p. fraser et al. / financial services review 29 (2021) 227–245 t ab le 3 in it ia l fe e ex am p le c li en t a (a cc u m u la ti o n ) c li en t b (t ra n si ti o n ) c li en t c (s p en d in g ) 1 2 3 4 5 6 7 8 9 n et w o rt h $ 6 0 0 k $ 6 0 0 k $ 6 0 0 k $ 2 ,4 0 0 k $ 2 ,4 0 0 k $ 2 ,4 0 0 k $ 6 ,0 0 0 k $ 6 ,0 0 0 k $ 6 ,0 0 0 k in v es ta b le as se ts $ 6 0 0 k $ 4 0 0 k $ 3 0 0 k $ 2 ,4 0 0 k $ 1 ,6 0 0 k $ 1 ,2 0 0 k $ 6 ,0 0 0 k $ 4 ,0 0 0 k $ 3 ,0 0 0 k p an el a c d f s 6 ,1 5 0 $ 5 ,3 5 0 $ 4 ,9 5 0 $ 2 1 ,9 5 0 $ 1 9 ,1 5 0 $ 1 7 ,5 5 0 $ 4 7 ,2 5 0 $ 4 1 ,2 5 0 $ 3 8 ,2 5 0 p an el b s in g le a u m 6 ,1 5 0 $ 4 ,3 5 0 $ 3 ,4 5 0 $ 2 1 ,1 5 0 $ 1 5 ,1 5 0 $ 1 1 ,5 5 0 $ 4 2 ,7 5 0 $ 3 0 ,7 5 0 $ 2 4 ,7 5 0 p an el c f ee d if fe re n ce , in it ia l $ — $ 1 ,0 0 0 $ 1 ,5 0 0 $ 8 0 0 $ 4 ,0 0 0 $ 6 ,0 0 0 $ 4 ,5 0 0 $ 1 0 ,5 0 0 $ 1 3 ,5 0 0 a s % o f n w 0 .0 0 % 0 .1 7 % 0 .2 5 % 0 .0 3 % 0 .1 7 % 0 .2 5 % 0 .0 8 % 0 .1 8 % 0 .2 3 % p la n n in g ef fo rt re q u ir ed s am e m o re s am e m o re s am e m o re t h is ta b le d ep ic ts th e in it ia l fe es fo r th re e d if fe re n t cl ie n t p ro fi le s (c li en t a , b , an d c ) ea ch w it h th re e ra ti o s o f n et w o rt h (n w ) to in v es ta b le as se ts (i a ). p an el a sh o w s th e fe es fo r th e co n so li d at ed d u al -f ee st ru ct u re (c d f s ) an d p an el b sh o w s th e fe es fo r th e si n g le , an d m o re w id el y u se d as se t u n d er m an ag em en t (a u m )b as ed fe e. p an el c h ig h li g h ts th e d if fe re n ce b et w ee n th e tw o fe e st ru ct u re s. s. p. fraser et al. / financial services review 29 (2021) 227–245 237 planning effort. we have summarized this position by labeling the relative planning effort required of the client with less investable assets (“same” – columns 1, 4, and 7) compared with the respective clients with identical net worth but less investable assets (“more” – columns 3, 6, and 9). nevertheless, these differences are only at most 25 bps of total initial net worth at the beginning of the client-advisor relationship. importantly, this decomposition of the overall fee into two components, one of which is comparable in function to the common, single fee-only aum that is pervasive today, we have deductively determined the value of the planning function that occurs above and beyond the act of solely managing investments. in doing so, we now allow planners and clients more transparency in what they are providing and buying, respectively. additionally, this approach provides planners a mechanism that can help them separate and distinguish their services such that they might address a larger number of clients’ needs. that is, planners can now price and offer more tailored services for a client who might need only im help or perhaps only broader planning assistance. under a single fee-only model with one aum fee structure, clients might think they are overpaying for planning that might not be required (columns 1, 4, and 7 in table 3). the cdfs allows planners to respond to them with the im fee schedule in table 1 that will compete with the so-called robo-advisors. additionally, if clients need significant financial planning help that is broader than investment management, planners can price it “fairly” in the market using the cdfs methodology. if a client includes an asset or asset class as part of their nw, then the advisor must plan around that asset or asset class and charge for it commensurately. however, planners cannot consider assets that they do not know to exist, so clients should not pay for this effort, nor should the advisor be expected to consider such assets in the overall plan. as many planners recognize, the financial planning effort for most clients is heavily front loaded. that is, more often than not, formulating the initial comprehensive financial plan involves collecting data about a client’s financial position, assessing a client’s goals and risk profile, analyzing a client’s financial position, and recommending potential actions to meet the client’s objectives can involve much more time and effort in the onboarding process than implementing and monitoring the plan in subsequent years. this fact motivates our next analysis, a corollary to the cdfs dual-fee structure, which is a fee feature that recognizes the lesser planning effort required in subsequent years. specifically, we next evaluate the fee impacts assuming financial planning component fee associated with the nw fee schedule is reduced by 50% after the initial year. in other words, instead the of the 60, 50, and 40 bps fee schedule shown in table 1, the fee schedule changes to 30, 25, and 20 bps of nw for every year beyond the initial year. the im fee schedule does not change, which remains consistent and competitive with the robo-advisor approach. extending this analysis over the investment lifetime of our example clients, we also aggregate the cumulative fees each client pays over their investment lifetime under the different fee approaches. table 4 contrasts the impacts of our two different fee approaches across time for the different client profiles described previously. specifically, we simulate 10,000 random time series of asset class returns and compare the average overall fees under the cdfs with the aum fee-only structure for each of our nine profiles.5 all values are in nominal dollars. panel a depicts the nw and im fees for year 1 and year 2 for the cdfs. panel b depicts the same fees for the aum model. we also aggregate the fees over the respective investment 238 s. p. fraser et al. / financial services review 29 (2021) 227–245 t ab le 4 f ee an d p o rt fo li o re su lt s o v er ti m e c li en t a (a cc u m u la ti o n ) c li en t b (t ra n si ti o n ) c li en t c (s p en d in g ) 1 2 3 4 5 6 7 8 9 n et w o rt h $ 6 0 0 k $ 6 0 0 k $ 6 0 0 k $ 2 ,4 0 0 k $ 2 ,4 0 0 k $ 2 ,4 0 0 k $ 6 ,0 0 0 k $ 6 ,0 0 0 k $ 6 ,0 0 0 k in v es ta b le as se ts $ 6 0 0 k $ 4 0 0 k $ 3 0 0 k $ 2 ,4 0 0 k $ 1 ,6 0 0 k $ 1 ,2 0 0 k $ 6 ,0 0 0 k $ 4 ,0 0 0 k $ 3 ,0 0 0 k s to ck /b o n d al lo ca ti o n 8 0 /2 0 6 0 /4 0 4 0 /6 0 in v es tm en t p er io d (y ea rs ) 4 0 2 0 1 0 p an el a d u al -f ee st ru ct u re (c d f s ) 1 ) y ea r 1 fe e, to ta l $6 ,4 33 $ 5 ,6 3 4 $5 ,2 24 $ 2 2 ,7 6 7 $ 2 0 ,0 4 2 $ 1 8 ,4 1 0 $ 4 9 ,0 1 5 $ 4 3 ,0 5 3 $ 4 0 ,0 6 3 2 ) n w fe e $ 3 ,6 5 7 $ 3 ,6 8 0 $ 3 ,6 9 2 $ 1 2 ,9 7 7 $ 1 3 ,0 7 1 $ 1 3 ,1 1 4 $ 2 8 ,0 0 9 $ 2 8 ,2 7 9 $ 2 8 ,4 1 5 3 ) im fe e $ 2 ,7 7 6 $ 1 ,9 3 4 $ 1 ,5 1 3 $ 9 ,7 9 0 $ 6 ,9 5 2 $ 5 ,2 7 7 $ 2 1 ,0 0 6 $ 1 4 ,7 5 4 $ 1 1 ,6 2 8 4 ) y ea r 2 fe e, to ta l $4 ,7 61 $ 3 ,9 2 9 $3 ,5 01 $ 1 6 ,7 1 7 $ 1 3 ,9 4 6 $ 1 2 ,2 6 4 $ 3 6 ,0 0 6 $ 2 9 ,8 3 7 $ 2 6 ,7 4 3 5 ) n w fe e $ 1 ,8 8 8 $ 1 ,9 1 6 $ 1 ,9 2 9 $ 6 ,6 6 4 $ 6 ,7 7 0 $ 6 ,8 2 0 $ 1 4 ,4 0 2 $ 1 4 ,6 9 7 $ 1 4 ,8 4 5 6 ) im fe e $ 2 ,8 7 3 $ 1 ,9 9 3 $ 1 ,5 5 2 $ 1 0 ,0 5 3 $ 7 ,1 5 7 $ 5 ,4 2 4 $ 2 1 ,6 0 3 $ 1 5 ,1 2 0 $ 1 1 ,8 7 8 7 ) l if et im e fe es $ 4 2 4 ,9 7 1 $ 3 9 5 ,9 7 2 $ 3 7 8 ,3 5 0 $ 4 5 0 ,8 6 4 $ 3 9 4 ,5 4 2 $ 3 6 2 ,2 9 3 $ 4 1 7 ,0 5 0 $ 3 5 3 ,9 3 5 $ 3 2 2 ,2 8 0 8 ) n w fe es $ 1 6 9 ,2 1 9 $ 2 2 6 ,1 1 5 $ 2 5 3 ,9 4 5 $ 1 8 4 ,0 2 4 $ 2 0 6 ,8 6 2 $ 2 1 8 ,3 6 2 $ 1 7 5 ,2 2 3 $ 1 8 6 ,1 7 7 $ 1 9 1 ,6 5 5 9 ) im fe es $ 2 5 5 ,7 5 2 $ 1 6 9 ,0 6 1 $ 1 2 3 ,6 0 9 $ 2 6 6 ,8 4 0 $ 1 8 7 ,2 8 2 $ 1 4 3 ,5 3 3 $ 2 4 1 ,8 2 7 $ 1 6 7 ,5 5 9 $ 1 3 0 ,4 2 6 1 0 ) t er m in al p o rt fo li o $ 3 ,3 3 1 ,7 8 0 $ 5 ,8 8 5 ,3 9 3 $ 7 ,1 6 8 ,0 9 3 $ 5 ,1 9 0 ,7 8 2 $ 6 ,7 2 3 ,2 0 9 $ 7 ,4 9 6 ,4 7 2 $ 8 ,4 4 3 ,1 6 6 $ 9 ,6 1 2 ,3 6 4 $ 1 0 ,1 9 7 ,0 7 9 1 1 ) t er m in al im p o rt fo li o $ 3 ,3 3 1 ,7 8 0 $ 1 ,9 8 3 ,8 9 7 $ 1 ,3 1 5 ,8 5 0 $ 5 ,1 9 0 ,7 8 2 $ 3 ,3 2 1 ,2 1 5 $ 2 ,3 9 3 ,4 8 1 $ 8 ,4 4 3 ,1 6 6 $ 5 ,5 3 7 ,7 8 0 $ 4 ,0 8 5 ,2 0 6 p an el b a u m fe e st ru ct u re (f ee -o n ly ) 1 2 ) y ea r 1 fe e, to ta l $6 ,4 53 $ 4 ,5 5 9 $3 ,6 10 $ 2 1 ,8 4 9 $ 1 5 ,8 4 9 $ 1 2 ,0 8 0 $ 4 4 ,2 8 3 $ 3 1 ,7 7 9 $ 2 5 ,5 2 6 1 3 ) y ea r 2 fe e, to ta l $6 ,6 72 $ 4 ,7 0 2 $3 ,7 14 $ 2 2 ,3 7 9 $ 1 6 ,3 4 1 $ 1 2 ,4 7 1 $ 4 5 ,5 0 6 $ 3 2 ,5 8 0 $ 2 6 ,1 1 7 1 4 ) l if et im e fe es $ 5 4 3 ,5 1 4 $ 3 8 0 ,3 6 5 $ 2 9 4 ,3 9 8 $ 5 7 0 ,6 2 1 $ 4 1 5 ,2 6 8 $ 3 2 6 ,8 4 1 $ 5 0 4 ,2 2 0 $ 3 5 8 ,0 1 7 $ 2 8 4 ,9 1 6 1 5 ) t er m in al p o rt fo li o $ 3 ,0 1 7 ,9 4 0 $ 1 ,9 6 9 ,3 8 2 $ 1 ,4 5 2 ,8 8 4 $ 5 ,0 1 0 ,1 0 7 $ 3 ,2 8 6 ,0 1 0 $ 2 ,4 4 2 ,8 2 8 $ 8 ,3 4 2 ,0 9 5 $ 5 ,5 3 5 ,0 0 0 $ 4 ,1 3 1 ,4 5 3 t h is ta b le d ep ic ts th e fe es fo r th re e d if fe re n t cl ie n t p ro fi le s (c li en t a , b , an d c ) ea ch w it h th re e ra ti o s o f n et w o rt h (n w ) to in v es ta b le as se ts (i a ). p an el a sh o w s th e fe es fo r th e co n so li d at ed d u al -f ee st ru ct u re (c d f s ) an d p an el b sh o w s th e fe es fo r th e si n g le , an d m o re w id el y u se d as se t u n d er m an ag em en t (a u m )b as ed fe e. im = in v es tm en t m an ag em en t. s. p. fraser et al. / financial services review 29 (2021) 227–245 239 period for each client group. finally, we present the terminal portfolio values for each client group, net of fees, decomposed into the nw and im components for the dual-fee structure. notably, for comparison purposes, it is appropriate to compare only the “terminal im portfolio” under the cdfs dual-fee structure with the “terminal portfolio value” under the aum fee-only structure when nw = im (columns 1, 4, and 7). this is the case because we assume a client’s nw is comprised of im (modeled with stocks and bonds) and potentially other assets. for clients represented in columns 2, 3, 5, 6, 8, and 9, these additional assets as modeled as evenly split between real estate and private equity asset classes. to the extent a client’s nw calculation includes non-appreciating real assets or property (e.g., autos, boats) instead of the real estate and private equity as we depict here, the portion of the nw portfolio not including im is likely overstated. accordingly, the level of the planning fee values in subsequent years after the first year are biased in the direction of assumed returns for real estate and private equity. this treatment highlights the importance of determining the appropriate nw for a client and whether (and how) those assets should be modeled to grow. additionally, in all other client scenarios (columns 2, 3, 5, 6, 8, and 9), the cdfs assesses fees both on nw and im, whereas the aum only fee is assessed on only the im. put another way, in the aum model, there are no fees assessed against the real estate or private equity holdings since they are not part of aum. analyzing table 4 provides opportunities for planners and clients to better grasp the potential impact of a separate financial planning and im fee, or what we call u in our model, against a single aum, or v , fee structure. by examining year 1 fees, one can see the breakout of financial planning fees and an im fee component compared with the single aum fee. recall, this illustration by design will result in the cdfs fee roughly equivalent with the single aum model fee. this is an appropriate comparison where nw = im (columns 1, 4, and 7). not surprisingly, the year 2 fees for these cases result in lower total fees due to the reduction in our illustration of the financial planning fee in subsequent years. again, these nw fees for year 2 and beyond are assessed at 50% of the initial fee schedule shown in table 2. to the extent our assumptions about the non-linear form of the holistic planning function hold, this would represent a cost savings to the client and perhaps a reduction in revenue for the planner. the timing and scope of the reduction in a planning fee are clearly variable levers included in the overarching u fee schedule that are at the disposal of planners when thinking about implementing a cdfs. the insight from the visibility of the two fee components could be a catalyst for a valuable conversation between planner and client. this distinction in fees perhaps highlights, and potentially quantifies, what some clients might perceive as the overpayment of advisors in a single aum model. alternatively, it could provide an opportunity for advisors to demonstrate their value above and beyond an im function. there are some additional observations that emerge upon further examination of table 4. while the cdfs we illustrate necessarily results in a lower year 2 fee for the cdfs, this is true only when nw is equal to im. as expected, line 1 is approximately equal to line 12 in this case, while line 4 is substantially less than line 13 for columns 1, 4, and 7. in contrast, the relationship changes when the ratio of nw to im is greater than one. when nw is greater than im—that is, the portfolio consists of more than just stocks and bonds (aum including mutual funds and etfs)—fees at the end of year 1 will no longer be 240 s. p. fraser et al. / financial services review 29 (2021) 227–245 approximately the same due to the fact the planning fee (u ) is based on the client’s nw, which includes assets not included in the im portfolio (where im and single aum fees are charged). consider the nature of the fees across the nw to im spectrum for each client a. now the total fee (revenue) in year 1 is greater under the cdfs than in the single aum model. line 1 is greater than line 12 in column 3. these observations hold across each of the other client types. the behavior of lifetime fees when comparing various clients are another result worth analyzing. some planners might consider the 50% discount to year 2 and beyond nw fees as an excessive “penalty” for still having to conduct the implementation, monitoring, and recommendation functions. again, looking at client a—as the observations hold for the others as well—we can compare the lifetime fees between the cdfs (line 7) and aum (line 14). although the cdfs fee is initially larger than the aum fee for year 1 and then smaller for year 2 as described above, over the 40-year investing lifetime of these particular clients, the overall cdfs fees outpace the aum fees. specifically, under the cdfs approach, the nw fees (line 8) increase while im fees (line 9) decrease as the initial net worth is composed more of non-investable assets and less of stocks/bonds. additionally, under our construct the annual fees are taken from im (i.e., stocks/bonds) under the cdfs model, which obviously reduces their compounded value over time, generating even lower im fees than they would if other asset classes were sold to support the fees. thus, we find that the value of financial advice, as represented by the nw fee in line 8, increases as the proportion of im decreases in an otherwise common-sized portfolio. this result is appropriate if one believes that the financial planning demands are higher for portfolios consisting of assets beyond stocks and bonds. it is also insightful that for these various clients, implementing a cdfs approach in the long-run generates comparable or even increased lifetime (nominal) fees relative to an aum only approach, despite the fact that the im are markedly different for the latter two clients in each of our client profile categories a, b, and c. in other words, accounting for all components of nw and conducting holistic financial planning based on these components can jointly generate comparable lifetime fees even when charging market competitive im-only fees and implementing a planning fee schedule that recognizes the reduced level of effort after creating the initial plan. thus, despite the almost-countless possibilities for establishing an appropriate u fee schedule, we demonstrate a simple example that could hold in a competitive market environment. to make a caveat or limitation abundantly clear, the lifetime fee levels presented in table 4 are reliant upon the various asset classes’ performance, which we simulate 10,000 times using the jpm market assumptions (see table 2). under these assumptions, the additional value of a portfolio that includes real estate and private equity becomes apparent. the difference in ending portfolio values (line 10) is stark for client a1 (stock/bond only) and client a3 (stock/bond/real estate/private equity). the former generates an ending portfolio worth $3.3 million net of fees versus $7.1 million for the latter. while we did not plan for or anticipate this result in designing this analysis, it is worthwhile noting the value of diversification is substantial under the return forecasts from jpm. additionally, our simple rebalancing assumptions for stocks and bonds coupled with the no rebalancing approach to real estate and private equity obviate some of the benefits of planning that could occur with more s. p. fraser et al. / financial services review 29 (2021) 227–245 241 sophisticated approaches (e.g., see blanchett and kaplan, 2013), for examples of these benefits, or gamma). clearly planners and firms could, and should, examine adjusting all the “levers” introduced here that are suitable for their practice: different fee levels, breakpoints, and subsequent reduction of planning or nw fee component after year 1. however, assuming similar regressive fee structures and market competition deterring any significant deviations from relatively comparable advising and im pricing, the general results from our analysis should hold for more complex fee structures. 5. implications for planners the analysis of the proposed cdfs in lieu of a single aum fee structure can prove valuable for financial advisors and clients alike. advisors only performing im functions under the single aum model are getting pressure to compress their fees today with the advent of robo-advisors. this trend has led traditional asset managers to expand their services to include more holistic and customized financial planning services, often to justify their fees. the cdfs provides an avenue for advisors to price the financial planning effort separately from the investment management function, effectively solving the problem of how to structure u fairly. such an approach provides greater transparency and could allow clients to hold advisors more accountable for planning support likely obfuscated by the single, fee-only aum models found at most ria firms (mazzoli and nicolini 2010). should advisors maintain some reduction of the planning fee component of the cdfs, planners could show that there is indeed a non-linear effort associated with the initial onboarding and financial planning function. moreover, such a dual fee structure might provide two additional benefits to planners and clients. first, it allows for the planning services to be priced, and more importantly, charged separately from any im fee allowing for a framework to charge for services—even when assets are not moved to the advisor for management, or what kitces (2013) suggests could be growing the “slice of the pic”. secondly, the cdfs allows for some level of financial planning complexity by more adequately compensating advisors when financial planning efforts are based on nw and not solely aum. a further benefit of this framework is that a cdfs could potentially reduce the conflict of interest, or at least can reduce the immediate sense of urgency faced by many advisors, to increase aum. for example, a prudent recommendation might be for a client to use investments assets to pay down a mortgage. however, a fee-only planner might be disinclined to make such a recommendation as the planner’s compensation would be reduced. in contrast, a planner using a form of the cdfs could make the recommendation knowing that any reduction in compensation due to the reduction in aum would be offset by the presence of a planning fee associated with the client’s nw. thus, the planner is not forced to choose between maintaining compensation (aum-only) or providing sound financial advice. again, the cdfs provides a mechanism to be compensated separately for financial planning, and that such an effort is based on an essential fundamental characteristic of planning, notably understanding the nw of the client. 242 s. p. fraser et al. / financial services review 29 (2021) 227–245 implementation of a cdfs is not without its challenges. first, establishing a client’s net worth, or at least the net worth used as a base for fees, is never as straightforward as it might appear. the decision to exclude certain assets, or even to value assets where there is little liquidity, is difficult.6 however, we suggest the value of these required initial and in-depth conversations with clients will only result in a stronger planner-client relationship as well as a more holistic and effective financial plan. second, the challenge of u . setting the appropriate net worth and asset breakpoint levels for each fee schedule, and the associated fees charged at each level, will need to be determined. those in this paper, while realistic, are admittedly a basic example for illustration purposes. while challenging to set them perfectly, it is essential for an advisor/firm to analyze what makes sense for their practice. it can only be beneficial for the firm to fully investigate the investment management (or aum) cost within their own firm. such an exercise will help them price their investment management operations appropriately, and such an exercise can lead to competitive—yet hopefully profitable—breakpoints and fees. the firm can then in turn move to assessing the scope and magnitude of their broader financial planning services and fees. if, and to what level, a planning fee might be reduced would also need to be accomplished. here again, this approach, and the accompanying conversation with clients, might make an advisor more attractive to potential clients. when a client does choose to move assets to the planner, the firm could charge a single, blended fee to ease in operationalizing the concept. the mere discussion of the component aspects will improve client-planner communication. for firms in which financial planning is their comparative advantage over investment management, the cdfs provides a viable option to provide and market those services as a standalone alternative. we also suggest the introduction of a cdfs can lead to new research opportunities. specifically, researchers can look to better set, and perhaps optimize, both the level and breakpoints of the cdfs schedule. these two variables are just two of the levers available to planners. so, too, more research will be needed to determine what might be the appropriate reduction in the nw component fee after the first year, or initial onboarding of a client. finally, it will be important to determine whether, and how, a cdfs might change given different client profiles (the ratio of nw to im) beyond the nine illustrated in this introductory model. we leave these efforts for further research. there is a myriad of compensation structures available for financial planners, and certainly no one compensation scheme is best for all situations. however, we posit that a composite, dual fee structure like that proposed here is worth investigating by all advisors/firms who perform holistic financial planning services over and above investment management services. such an examination would provide more transparency to clients, more appropriate pricing for services provided, and perhaps even reduce conflicts of interest. we find that under reasonable return assumptions, such a fee structure can result in comparable fees and portfolio impacts as the pervasive aum structure. all these outcomes would improve the nature of the fiduciary relationship. perhaps such an effort can improve on the awareness of the nearly 20% of clients identified by cheng and kalenkoski (2018) who have no idea what for what services or advice they are paying. the benefits of pursuing such an approach likely far outweigh the costs, for planners and clients alike. s. p. fraser et al. / financial services review 29 (2021) 227–245 243 notes 1 we use the term financial planning in lieu of financial advice to distinguish when advice might be provided incidental to the sale of a product. 2 in this analysis u reduces to u /2 for each year after the initial planning activities occur. 3 annual advisor and trading fees are paid from stock and bond (ia) asset returns. in sum, at the end of each year of analysis, either stocks or bonds are relatively overweight. im fees are paid from the overweight asset, and then the necessary remainder is sold and used to purchase the underweight asset to bring the stock-bond asset mix back to its original weighting scheme. transactions or trading fees are assessed at $9.95 � 2 for a roundtrip buy-sell transaction. the remaining portion of the nw portfolio is assumed to grow at the appropriately weighted rate of private equity and real estate. 4 these mean values become even more aligned between the simulation and target values over our 20and 40-year time periods discussed later in the study. based on the numerical similarity between the simulated and target measures (return, risk, and pairwise correlations), we are confident the simulated returns sufficiently represent the potential future returns based on the 2020 jpm capital market assumptions. 5 we depict only mean values for parsimony purposes. all distributional data are available upon request. 6 this analysis assumes planners would assess negative equity assets (e.g., a home that is underwater) as having a zero value for net worth. references bae, s. c., & sandager, j. p. (1997). what consumers look for in financial planners. journal of financial counseling and planning, 8, 9–16. blanchett, d., kaplan, p. (2013). alpha, beta, and now. . .gamma. the journal of retirement, 1, 29–45. brenner, l., & meyll, t. (2020). robo-advisors: a substitute for human financial advice. journal of behavioral and experimental finance, 25, 100275–100278. cheng, y., & kalenkosi, c. m. (2018). lost in fees; an analysis of financial planning compensation. journal of wealth management, spring, 46–54. egan, m. (2019). brokers versus retail investors: conflicting interests and dominated products. the journal of finance, 74, 1217–1260. fama, e. f., & french, k. r. (2010). luck versus skill in the cross-section of mutual fund returns. the journal of finance, 65, 1915–1947. finke, m. s., huston, s. j., & winchester, d. d. (2011). financial advice: who pays. journal of financial counseling and planning, 22, 18–26. haslem, j. a. (2010). the new reality of financial advisors and investors. the journal of investing, 19, 23–30. hoechle, d., ruenzi, s., schaub, n., & schmid, m. (2018). financial advice and bank profits. the review of financial studies, 31, 4447–4492. kitces, m. e. (2013). a ‘new normal’ look at practice growth. journal of financial planning, 26, 16. kitces, m. e. (2017). financial advisor fees comparison – all-in costs for the typical financial advisor? available at www.kitces.com lahtinen, k. d., & shipe, s. (2018). compensation of investment advisors. the journal of investing, 27, 80–86. 244 s. p. fraser et al. / financial services review 29 (2021) 227–245 mackillop, s. (2017). it’s time to reexamine your fee schedule. journal of financial planning, 30, 34. mazzoli, c., & nicolini, g. (2010). to fee or not to fee: pricing policies in financial counseling. financial services review, 19, 307–322. opiela, n. (2006). the future of fees. journal of financial planning, 19, 24–31. seay, m. c., anderson, s. g., lawson, d. r., & kim, k. t. (2017). identifying variation in client characteristics between financial planning compensation models. journal of financial planning, 30, 40–51. sec. (2014). how fees and expenses affect your investment portfolio. investor bulletin – security and exchange commission’s office of investor education, pub. no. 164. statman, m. (2000). the 93.6% question of financial advisors. the journal of investing, 9, 16–20. uhl, m. w., & rohner, p. (2018). robo-advisors versus traditional investment advisors: an unequal game. the journal of wealth management, 21, 44–50. s. p. fraser et al. / financial services review 29 (2021) 227–245 245 pii: 1057-0810(94)90020-5 financial services review, 3(2): 1.57-158 copyright 0 1994 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. a practitioner’s perspective: comments on allocation, life expectancy and shortfall” barbara s. poole “asset as i reviewed one of the articles in this issue, i thought of a client i met for the fust time on november 1, 1987, a month before his planned retirement date. until the past month, which included the market crash of october 17, this dentist had been delighted with the performance of his all-equity portfolio. however, now it was valued at about half of what he had expected, and in his view, was insufficient to produce the retirement income he needed. over the next few weeks, he and his wife met with me several times to discuss some tough choices about income and expenses. they concluded that while the children’s educations would not be disrupted, for the next two years he would continue drilling and filling and their plans to buy a retirement home would be deferred. this back up plan worked well, the investments recovered and the couple is now very happily retired. in this issue’s “asset allocation, life expectancy and shortfall,” kwok ho, moshe arye milevsky, and chris robinson (hmr) construct a model which supports this client’s 100% equity position, based on his wealth to consumption ratio and age. and ultimately, things did work out just fine; however, this particular client was lucky because he did have the option to continue working and he knew intuitively that with time his investments would recuperate. but the lives of this investor and his family had been severely disrupted by his investment position, a position to which few responsible planners will knowingly expose a client. but the model makes some very important points that should not be so quickly dismissed. hmr’s analytical model allocates a retiree’s assets between risky and risk free investments with the goal of minimizing the probability of not meeting consumption requirements. the model uses age, life expectancy, and consumption level to arrive at an ideal investment mix for the initial wealth. to see how a practitioner could benefit from hmr’s work, it may be instructive to view it in the context of constructing a financial plan for an individual at retirement age. first, let’s consider the approach of a typical planner. the planner and client discuss the individual’s investment attitudes and experience, and together they define goals and construct a list of priorities. now let’s look at the approach of a model. some models use barbara s. poole l financial services review, department of finance, university of connecticut, storrs, ct 06269 158 financial services review, 3(2) 1994 the utility curve to describe the set of risk-return trade-offs that an individual is willing to make, and seek to determine the mix of risk and return that provides the most satisfaction, or utility, to the individual. instead of incorporating a utility function into their analysis, hmr make the assumption that minimizing the probability of outliving their funds is the solution to maximizing the individual’s utility function. like the client i described earlier, while individual’s needs are usually more complicated that this, if a client were confined to expressing only one goal, this could very well be the one. retirement non-investment income can be projected and the additional investment income necessary to fund consumption is easily obtained. generally, the planner will construct the portfolio so that income producing securities generate sufficient supplemental income to cover expenses and the remainder, or some portion of it, is allocated to growth oriented securities. instead, hmr would apply their model to obtain the appropriate allocation and use income, capital gains, and principal, as necessary, to finance consumption. most clients are reluctant to use principal, and especially for young retireds (younger than 75, for instance) a typical practitioner likewise will be reluctant. some planners rely on matching income to expenses, and will make adjustments in investment risk to produce increased income. of course, this is the way that so many widows looking for increased income migrated up in risk-taking from cd’s to junk bonds as interest rates dropped in the mid 1980’s. however, under some circumstances, it is inevitable, even desirable, for a client to tap principal to pay expenses. many planners fail to recommend this responsibly by including in their plans back-up strategies such as reverse mortgages, supplemental income programs, and expense subsidies. a comprehensive plan will include projections of the social programs that would be available should funds be depleted; this planning increases the likelihood that should the situation arise, the client will be eligible for benefits when needed. as we observe the results of hmr’s model, the most striking implication is the importance that equity plays in the portfolio, even in the later years. the authors anticipated that this would be surprising to those of us who generally expect large equity holdings to pose excessive risk to most aging retirees. however, the result is not as surprising when the construction of the model is examined. it is the authors’ use of the volatility of the real return after inflation for both the treasury bills and the equity in the portfolio that drives their results. they remind us that the risk-free asset is only default risk free, not entirely risk free. but the emphasis on equity can be supported intuitively when we consider a few societal changes over the years. retirement is occurring earlier today, partly due to corporate tightening that encourages early retirement and reduces job availability. when they do retire, most individuals are healthy and look forward to an extended retirement period; in fact, the fastest growing segment of our population is the 85 + age group. so it is no surprise that retirees, especially women, need to pay more attention to growth oriented investing than they ever have before . their reminder to consider equity in older individual’s portfolios may be hmr’s most significant contribution to the practitioner. pii: 1057-0810(91)90001-f volume 1 number 1 financial services review the journal of individual financial management editor lewis mandell managing editor barbara poole associate editors financiaii~itutions: mona j. gardner illinois wesleyan university benton e. gup university of alabama david s. kidwell university of minnesota neil b. murphy virginia commonwealth university peter l. struck washington mutual savings bank financialplanning: neil g. cohen george washington university james w. jenkins brigham young university thomas waschauer san diego state university investments: jean louis heck villanova university george c. philippatos university of tennessee frank k. reilly university of notre dame g. c. uselton texas a dt m university real estate: arthur l. schwartz university of south florida reviews: phyllis s. myers virginia commonwealth university riskmanagemenk s. travis pritchett university of south carolina tax & estate pkuining: chris j. prestopino california state university chico greenwich, connecticut london, england pii: 1057-0810(92)90001-s from the editor this issue completes the second year of publication of the financial services review. at this juncture, it is useful to ask whether the new publication appears to be “successful.” the answer is mixed. from an editorial and financial standpoint we are fine. the number of quality submissions increases continually and our total number of subscribers puts us well in the black. from an academic perspective we seem to be on target. our articles are theoretically rigorous and a sizeable proportion of authors and editors are well-known in the field. survey data and calls from finance departments around the country indicate that we are regarded as a quality refereed journal in the “good” category, comparable to the financial review. finally, we must ask whether we are having an impact on the practice of individual financial management. in our “letters to the editor” section, a thoughtful reader and practitioner (with an mba) replies in the negative. “as an academic journal it may be fine, but from a practical standpoint, i did not find it very helpful.” he goes on to say that “. . . the client does not want generalities and abstractions. he wants practical, specific direction.” in creating this journal, the academy of financial services clearly opted away from the “practical.” there are hundreds of thousands of planners, brokers and bankers giving practical advice to clients, but only a few hundred scholars doing the basic research necessary to determine which advice is likely to be correct. there is a clear division of labor, but an unclear channel of communication between the two groups. a tiny handful of practitioners subscribe to the journal ( a dozen or so at best) and even our practitioner-readers are apparently dissatisfied. perhaps diffusion will occur through the textbooks to the classrooms so the next generation of practitioners will benefit from our research. we welcome suggestions. the lead article in this issue, “an index of portfolio diversification,” by walt woerheide and don persson addresses a relatively practical question for most individual investors, namely how many securities should be held in a portfolio to secure adequate diversification? while most of us are familiar with the academic studies that addressed this question, this article points out that previous studies have assumed that portfolios were evenly divided among their components. with uneven v vi financial services review, 2(2) 1993 distribution, what is the best measure of diversification? the authors examined five such measures and found that the complement of the herfindahl index was best. the second article is somewhat more theoretical than the first. “the individual investor in the market: forming a belief regarding market efficiency,” by robert m. peevy, gene c. uselton and john r. moroney asks whether high market price volatility detracts from market efficiency by causing investors to focus on price movements, per se, and away from information related to the true value of the security. while not concluding that markets are either efficient or inefficient, the authors find that the “variance bounds” literature which attempts to make the case for inefficiency on the grounds of excess price volatility, does not do so convinc ingly. the third article directly challenges advice given by many practitioners to their clients, namely that international diversification, through the purchase of shares of international mutual funds, is beneficial to investors. using recent data, the authors larry r. lang and robert m. niendorf in their article “performance and risk exposure of international mutual funds,” find that on a risk-adjusted basis, inter national funds did not outperform domestic funds. the last two articles are survey and review pieces, written to bring our readers to the state of the art in the diverse fields that comprise individual financial management. the first, “what strategies are estate planners recommending? evidence from survey data,” by chris prestopino summarizes advice from attor ney-authors who have published recent articles in the leading estate planning trade publications. our featured review article is entitled “the risks of pension plans.” written by robert w. mcleod, sharon moody and aaron phillips, this timely piece identifies and describes the risks inherent in today’s pension plans. particular attention is focussed on the roles of erisa and the pension benefit guaranty corporation in safeguarding pension assets. from the editor this issue contains volume 28 issue 4 of financial services review (fsr). this issue is our special issue on financial literacy. we thank the many authors that submitted articles for consideration for this issue. interest in this area of financial planning has been tremendous. in fact the reception was so well received, we will have an additional financial literacy issue in 2021. i would like to thank my co-editor, terrance martin at utah valley university for his wonderful insight and assistance on this special issue. i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “financial literacy, attitudes, and financial satisfaction: an assessment of credit card debt-taking behavior of australians” is coauthored by muhammad s. tahir, daniel w. richards, and abdullahi d. ahmed, all at rmit university, melbourne, victoria, australia. the authors examine the relative strength of the association of financial literacy, attitude towards balancing spending and savings, and financial satisfaction with credit card debt-taking behavior using the 2016 wave of the household, income and labour dynamics in australia (hilda) survey. they find that higher financial literacy is associated with less credit card debt. based on their results, they advise policy-makers to include components in the financial literacy curricula which encourage savings attitude to reduce problematic debttaking. the second article “can financial literacy education reduce the use of medicaid and snap?” is coauthored by abdullah al-bahrani, northern kentucky university, darshak patel, university college of dublin, ireland, and jamie weathers, western michigan university. the authors explain that policies supporting financial literacy education are promoted as a way to decrease reliance on social safety nets. the assumption is that low levels of financial literacy translate to lower economic outcomes and thus to increased dependence on social programs. the authors use the 2018 national financial capabilities study to investigate the possible relationship between high school mandated financial literacy education and social program participation and find no evidence of such a relationship. the third article, “financial literacy: profiling a successful high school outreach program” is coauthored by greg filbeck, penn state behrend, jason pettner, c.s. mckee, 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) v–vi and xin zhao, penn state behrend. during 2018-2019 the cfa society pittsburgh launched a high school financial literacy campaign resulting in significant improvements in financial behavior, subjective and objective financial knowledge, and self-esteem. in analyzing the results the authors find a disconnect between actual and perceived financial knowledge. they found that students exhibit gains in all aspects after completing the program. the subcategories with the lowest pre-survey scores or female students show the greatest improvements in the post-survey. students with lower gpas experienced greater improvement in financial behavior and objective knowledge, while higher gpa students improved more in subjective knowledge. the final article, “the association between financial risk and retirement satisfaction” coauthored by blain m. pearson and michael guillemette, both at texas tech university. the authors posit that a higher level of risky financial assets that a retiree holds may produce higher returns, resulting in utility gains. they test this hypothesis using a variable constructed measuring retirees’ ratio of risky assets to total assets (risk ratio).they examine the association between the risk ratio and retiree utility using a retirement satisfaction variable from the 1992-2014 waves of the health and retirement study. they find that increases in retirees’ risk ratio is associated positively with increases in their retirement satisfaction. thank you to those who make the journal possible, especially the referees and contributing authors. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review terrance martin co-editor, special topics issue on financial literacy vi s. michelson and t. martin / financial services review 28 (2020) v–vi pii: 1057-0810(92)90013-3 financial services review. 2(l): 21-39 copyright g, 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. i active timing decisions of equity mutual funds robert radcliffe robert brooks haim levy in this paper we examine an aspect ofprofessional investment management which has not been adequately documented and studied; the extent to which equity mutualfund managers actively adjust their portfolio’s equity risk evosure over time. estimates of a portfolio’s quarter-end beta are developed using the actual stock ho~~gs oft~~~ol~ at the quarter-end. changes in these beta estimates fivm one quarter to the next are shown to arise frtm both passive and active asset allocation. we find that active risk adjust~nt domi~~s~ssive rebalancing and that equity risk exposunz is quite variable over time. thus, individual investors who estimate the equity risk inherent in a portfolio bused on a single time series return beta might seriously misestimate the portfolio’s current equity risk. we also test whether active risk management is better characterized lis anticipatory offuture market events or reactive to past market events. i. i~ro~~~on active investment managers attempt to provide excess risk adjusted returns by a combination of judicious security selection and asset class timing. this paper deals with the extent to which equity mutual fund managers engage in active timing of their portfolio’s equity risk exposure. what sets the study apart from previous studies of investment managers is that we are able to accurately measure a portfolio’s equity risk exposure at a given point in time by observing the security holdings of the portfotio at that time. using security holdings, we calculate cross-sectional portfolio betas at each quarter-end during our sample period. previous studies have not used quarter-end security holdings but, instead, have used the time series of portfolio returns during the period studied to robert radcliffe l department of finance, university of florida, gainesville,fl 32611-2017. robert brooks l department of economics, university of alabama, tuscaloosa, al 35487-0224. hah levy l department of finance, university of florida, gainesville. 22 financial seirvices review, 2(l) 199m993 estimate the portfolio’s equity risk exposure. although most prior studies estimate a single beta for a portfolio, a few researchers have examined whether portfolio betas appear to change over time. but because their tests were based solely on a portfolio’s time series returns, they were unable to clearly document the intertemporal variabil ity in equity risk exposure within a given portfolio. this is the first study to empirically document the extent to which active asset allocation is used by profes sional equity managers.’ the basic data used in this study consist of cross-sectional betas on 94 equity mutual funds for each quarter-end during the period march, 1977 through december 1988. our measure of active asset allocation during a given quarter is equal to the difference between the actual quarter-end beta of a fund and the expected quarter end beta if the manager had not engaged in active asset allocation. our principle conclusion is that most mutual fund managers do not maintain a static level of equity risk exposure in their portfolios. instead, they engage in active market timing activities by adjusting both the percentage of the fund invested in equities and the average beta of any equities helde2 this means that single estimates of a fund’s beta which are based on a time series of past fund returns are poor estimates of current equity risk exposure and provide no information about past timing activities of the portfolio manager. this finding has important implications for individual investors. single time series regression betas (which are commonly provided in public sources of infor mation about mutual funds} provide no information about the timing activities of a portfolio manager and can seriously misrepresent the current equity risk exposure of a portfolio. when selecting one or more managers, the investor should have a good understanding of how actively the manager adjusts portfolio risk exposure. unfortunately, such information is not publicly available at present.3 we also examine whether active equity risk management is better charac terized as reactive or anticipatory. reactive decisions are defined as those which can be traced to past aggregate market returns or past portfolio returns. anticipator decisions are defined as those which are unrelated to past market or portfolio returns. the results suggest that most active asset allocation decisions are better characterized as anticipatory in nature. there is, however, some weak evidence of reactive decisions in that active changes in equity positions are statistically related to current and past aggregate stock returns. in addition, we find evidence suggesting that manager’s whose year-to-date returns have been poor relative to the s&p500 at a september quarter-end tend to incur increased equity risk during the fourth quarter of the year. the implication of this finding for individual investors is obvious. if an equity mutual fund whose year-to-date return in september is poor relative to the s&p500, the investor should obtain current information about the portfolio’s equity risk exposure to be sure that it does not exceed the investor’s tolerance for risk. active timing decisions of equily mutual funds 23 ilactive asset allocation active investment managers make two types of decisions in their attempt to achieve portfolio returns in excess of a passively managed portfolio with similar iisk: 1. decisions to maintain or change the portfolio’s percentage investment in various asset classes, and 2. within each asset class, decisions about the weights of individual securities. the first decision has historically been referred to as a timing decision. more recently, it has come to be known as tactical asset allocation. the second is known as the security selection decision. if security markets are informationally efficient, active managers will not be able to win from either decision. in fact, they would loose due to transaction costs. yet it is clear that active managers do not believe that security markets are informationally efficient and that they attempt to provide excess returns by engaging in both timing and selection activities. the extent to which managers employ timing and security selection decisions has never been clearly documented. some managers state that they make extensive use of timing in their portfolio management whereas others shun the value of timing and concentrate almost exclusively on security selection decisions. later in the paper, we present statistics on the extent to which active timing is used. knowledge of a portfolio’s current equity riskexposure is important for at least two reasons. first it provides important information to individuals who are invested in the portfolio or are considering an investment in it. second, it is potentially important when one attempts to evaluate the historical return performance of the portfolio. investment scholars have focused mainly on the second issue. jensen (1972) was the first to note errors which can arise in performance studies which rely on a sirgle time series return beta. he showed analytically that single time series betas can overstate the average equity risk exposure of a portfolio and understate the portfolio’s alpha (constant risk adjusted return) when managers are successful in their timing abilities. since jensen’s observation, many researchers have attempted to measure the extent and success of manager timing abilities. for example, kon and chen (1978) applied a switching regression model to the time series returns of 49 mutual funds and found that at least two different “beta regimes” were statistically present for most funds in their sample. they concluded that their “evidence should be regarded as a severe violation of model specification for those studies that employ o.l.s. to estimate mutual fund performance.“4 other studies of mutual fund timing also suggest that mutual fund managers do not maintain constant equity risk exposure over time. see for example, bauer, hays and upton (1987), chang and lewellan (1984). fabozzi and francis (1979). henriksson (1984). jagannathan and korajczyk (1986), kon (1983), kon and chen (1979) and lee and rahman (1990). however, none of these studies were able to calculate a portfolio’s 24 financial services review, 2(l) 1992/1993 beta at a given point in time since they relied solely on the portfolio’s time series returns. al~ough an investment manager could concep~~ly time across a large variety of asset classes, in practice they tend to specialize in only two asset classes; typically maintaining positions in money market securities and only one of a variety of other asset classes. we have chosen a sample of mutual funds which invest almost exclusively in money market securities and us. equities.’ a. measuremnt of equity risk exposure at quarter 1 the equity risk exposure of a portfolio can be measured in two different ways. the approach which has been employed in most prior studies of equity managers calculates a single estimate of a portfolio’s equity risk exposure by regressing the time series of portfolio returns on a proxy for aggregate equity market returns. the advantage of this approach is its ease. historical fund returns are available at low cost from a variety of sources. the disadvantage of the approach is that it provides only a single estimate of a portfolio’s equity risk exposure over a time period during which equity risk might have been const~tly changing. the approach used in this study calculates a portfolio’s equity exposure at each quarter-end based on the security holdings within the portfolio at that quarter-end. a market model beta is first calculated for each security held and then weighted by the percentage of the portfolio’s total market value which the security represents. defining bt as the beta (or effective equity position) of a portfolio at the end of quarter t, then: where: nt = number of securities held at the end of quarter t, bit = beta of security i at the end of quarter t, nit = number of shares of i held at end of quarter t, pit = price per share of i at end of quarter t, and tmv, = total market value of the portfolio at end of quarter t. to illustrate how the single time series beta can differ from the sequence of actual quarter-end fund betas, consider figure 1. the hormonal line shows the single beta estimate for american mutual fund (amf) based on quarterly amf returns from march, 1977 through december, 1988. the solid line which varies over time shows the actual quarter-end betas of amf using equation (1) and the security holdings of amf at each quarter-end. although the single time series beta (equal to 0.70) was close to the average of the cross-sectional portfolio betas (equal to 0.7 l), active timing decisions of z?quity mutuaz fun& 25 ” i /iigeq i 0.55 0.6 i v 0.5'( a,,,,,,,,,,,,,,l,a,~ ,,,,,,,1 ,,i -,,,,,,,,,,,,,,,i n12 7512 7812 5012 5112 ml212 8312 5412 5512 ml2 8712 82.12 quarter end figure 1. beta estimates for american mutual fund. considerable variability existed over time in the fund’s actual equity risk exposure. cross-sectional betas ranged from a high of 0.99 in march, 1980 to a low of 0.5 1 in march, 1987. in short, although betas which are based on a fund’s historical returns during a given period of time might be a reasonable estimate of the average equity risk exposure of the fund during that time period, they can be poor indicators of equity risk at given dates within the interval. b. measurement of active asset allocation in quarter t the major variable of interest in this study is a fund manager’s active asset allocation decision during quarter t, aaa,. this variable is calculated by subtracting a fund’s expected equity allocation at the end of quarter t if no active asset allocation decisions had been made during the quarter, e(bj, from the actual equity allocation at the end of the quarter, b,. aaa, = b1 e(b,) (2) in this section we discuss how the expected equity allocation is calculated. given a fund’s equity allocation at the start of a quarter, an estimate of expected fund returns during the quarter can be calculated. assuming that there are no consistent excess returns from security selection decisions during quarter t and that 26 financial services review, 2(l) 1992m!j3 all non-equity securities are treasury bills, then the return expected during quarter t would be? e(r,) = &.,rmt + (1 bdrt, where r&f refers to the return on a reasonable proxy for aggregate equities and rt is the return on treasury bills. if the manager does not make any active asset reallocations during quarter t, the expected end of quarter portfolio beta would be: jw,) = &-,i0 + rm,) i(1 + jwc))i (4) and the passive asset allocation during quarter t, paa,, would be: paa, = zz(b,) b,, (5) for example, assume that the portfolio beta at the start of a quarter is 0.56 and that during the quarter rm is equal to 10 percent and rt is 2 percent. then the expected portfolio return for the quarter would be 6.48 percent and the expected equity risk exposure, e(b,), would be 0.5785. thus, the passive asset allocation decision would be equal to +0.0185. even though the manager might not actively change the portfolio’s equity risk exposure, the portfolio beta will change passively due to relative returns on equities versus treasury bills. we measure the degree of active asset allocation, aa& by subtracting the expected portfolio beta, e(b,), from the actual quarter-end beta, bp aaa, = & j?(&) (6) an active asset allocation decision is made by a manager based on informa tion known to the manager at time t. we classify such information into two general categories: (1) info~ation which is observable in past stock market and meager returns and (2) information which is not related to past returns. if active asset allocation decisions are related to past market or manager returns, we say that the timing decision is reactive. if the decision can not be traced to past returns, we say that the decision is anticipatory. given these definitions, anticipatory decisions can be based on observation of events which occur in both current and past quarters. but as long as such info~ation is not related to historical rates of return of either the manager or the aggregate stock market, then we classify the decision as anticipa tory? past manager and equity market returns which are used in the study include:’ active timing decisions of equity mutual funds 27 spr = return on the s&p500 index during quarter t, f, = return on the fund during quarter t, xf, = excess return on the fund in quarter t, = ftsp,, csp&k = cumulative return on the s&p500 from k quarters back through the end of quarter t, = (1 + sp&(l + spr_k+,) ’ (1 + sp,), and cxf,_k = cumulative excess returns on fund from k quarters back through the end of quarter t = (1 +xf&(l +xf,_k+i)**.(l +xfj. these five variables are used to examine whether active asset allocations are related to either: (1) past aggregate stock market returns or (2) to past excess manager returns. consider first the case of aggregate stock market returns. reactive decisions based on past aggregate stock returns could arise from a variety of possible motives. for example, the manager might have a long-run portfolio beta target and desire to rebalance any passive changes in the portfolio’s equity risk exposure towards this target. if this is true we would expect to find a negative relationship between paa and aaa. in quarters in which common stock returns are large, the value of paa would be large and positive. if the manager rebalances the equity risk exposure towards a long-run target, aaa would be negative in such aquarter. another motive is often referred to as trend following. if aaa is positively related to returns on the s&p500 in the current or previous quarters, it is likely that the manager believes that current and past market returns can be extrapolated to future quarters. an alternative to trend following is referred to as a contrarian strategy. in this case the manager actively adjusts the equity risk exposure in a direction opposite to current and past market returns. if a contrarian strategy is used, aaa should be negatively related to current and past s&p500 returns. reactive asset allocation decisions could also arise if manager returns have been substantially different from returns on the aggregate stock market. managers having high relative returns might choose to reduce their equity exposure since they have already “beat the market” in the eyes of the mutual fund owners. similarly, managers with low relative returns might increase their equity exposure in the hopes of offsetting past relative returns.’ we label one variant of such reactive allocations the september hypothesis. there are two aspects to this hypothesis. first, managers are hypothesized to be mom concerned about their calendar year returns than any other yearly return (say, for example the yearly return ending in june). the manager is said to believe that investors judge the manager’s performance based primarily on yearly returns calculated as of the december quarter-end of each year. second, prior to the last quarter of the year, managers who have done poorly relative to an equity market index such as the s&p500 are hypothesized to have little to loose if they take on 28 financialslervicesreview,2(1) mjm993 extra equity risk but a lot to gain. if managers increase equity risk exposure and the equity market declines, they are not harmed since they have already been labeled as “losers” for the year-to-date. but if the equity market rises, they have a greater chance of offsetting previous poor relative returns. if this hypothesis is true, we should find that past relative returns are important determinants of aaa mainly for managers with poor relative performance and that the relationship becomes stronger for each successive quarter within a year. during any year, the strongest relationship between past relative returns and the current quarter’s aaa should occur among managers with the poorest relative performance and should occur at the end of september. the notion that active investment managers with poor historical returns have nothing to lose from increasing equity risk exposure but, instead, can only gain has been discussed in the theoretical literature. for example, see grinblatt and titman (1988). this is the first paper to empirically test the hypothesis. if aaa’s are not related to past s&p500 or manager returns, we categorize the active asset allocation decision as anticipatory. an examination of the variables which managers might consider in such anticipator decisions is beyond the scope of this study. we concentrate solely on the extend to which active asset allocations can be described as reactive or anticipatory. d. prior research the finance literature which examines timing decisions of fund managers is extensive. many of these studies were noted above. however all of this literature examines whether fund timing decisions are successful; an issue which is quite different from the subject of this paper. in this paper we do not ask whether fund managers are able to earn abnormal returns due to their active asset allocation decisions but, instead, examine the extent to which active asset allocation is used and whether such decisions are best characterized as reactive or anticipatory. a. sample description the sample was restricted to equity mutual funds for which quarter-end security holdings were available from march 3 1,1977 through december 3 1,1988 and which invested predominantly in u.s. money market securities and u.s. equities. this resulted in 94 funds with 48 quarterly ob~rvations each. quarterly fund returns were obtained from cda investment technologies, inc.. quarterly portfolio betas of equation (1) were calculated using cda spectrum tapes and the crsp daily returns tapes. for each quarter-end, the spectrum tape provides a listing of all stocks held by each mutual fund as well as the number of shares held. quarterly mutual fund betas were based on the betas of stocks held by active timing decisions of equity mutual funds 29 quarter figure 2. average fund beta by quarter. the fund. crsp tapes were used to estimate the beta on all stocks which had at least 30 monthly returns during the 60 months prior to a given quarter.io stock betas were estimated for each quarter starting with march 31,1977 and ending on december 31,1987. standard market model regressions were performed using monthly stock returns in excess of 90-day treasury bills. the s&p500 was used as the market portfolio proxy. the average beta of the 94 mutual funds is shown in figure 2 for each quarter-end in our sample period. during the late 1970’s, the equity risk exposure of the average fund was about 1 .o. however, during the early 1980’s. average fund betas declined to about 0.85 and (with some variability) remained below 1 .o through 1988. b. the extent of active asset allocation to evaluate the extent to which managers employ active asset allocation in portfolio management, we first examine data for each quarter across all managers and then examine data for each manager across all quarters. consider the two panels of figure 3. in panel a, average active asset allocation (aaa) is shown for each quarter in the sample. in panel b, the average passive asset allocation (paa) is shown. i1 vertical scales of both panels are 30 panel a. financial services review, 2(l) 19!wl!w3 -0.08 -0.1 111111111,,1,,,1,,,,11,,,,,,,,,,,,,,~~,,,1,,, 7712 7612 7912 8012 6112 8212 8312 6412 8612 8612 8712 quarter panel b. quarter figure 3. panel a: average active asset allocation. panel b: average passive asset allocation. active timing decisions of equity muhd funds 31 identical. the average aaa was negative in twenty-six quarters versus nineteen positive quarters. average quarterly aaa values ranged from +0.09 to -0.089. recall that an aaa value of +0.09 means that the equity risk exposure (fund beta) was actively increased by 0.09. during the quarter in which black monday occurred, aaa was not unusually different from other quarters. as would be expected, the black monday quarter displays the largest negative passive asset allocation. however, the important point communicated by the panels in figure 3 is that active asset allocation by mutual fund managers clearly dominates passive asset allo cation. managers do engage in active asset allocation. the degree to which managers engage in active asset allocation is even better seen in the data displayed in table 1. for example, during the quarter ended june 30, 1977, the average aaa was +0.09. but the standard deviation of aaa across all managers was 0.115. the largest active decrease in a fund’s beta during this quarter was a negative 0.118 and the largest increase was +0.683. the data in this table show clearly that active asset allocations within a single quarter can be sizeable. information about each manager over time is shown in table 2. for each manager, the average and standard deviation of their quarterly aaa variable is shown. for presentation purposes, the data are sorted by standard deviations of aaa.12 standard deviations range from 0.361 to 0.039. for comparison with figure 1, amf is fund number 3 with an average aaa of -0.0042 and a standard deviation of 0.0595. two conclusions are evident from this data. first, the extent to which managers engage in active management of portfolio risk exposure varies. second, for the typical manager, portfolio risk exposure is actively managed and can change substantially over the course of only three months. although not shown here, most of the changes in portfolio betas are due to changes in a portfolio’s stock to total assets ratio. the beta of equity securities held is much less variable; managers seemed to maintain fairly constant equity portfolio betas and altered portfolio equity risk by increasing or decreasing the percentage of assets held in equities. we fiid this comforting for two reasons. first, although we might estimate stock betas with an error since they are based on prior 5-year returns, there is virtually no error in measuring the stock to asset ratio. second, although equity betas might change by pure chance as a manager engages in stock selection activities, changes in stock to asset ratios are closely monitored by managers and made largely for the purpose of market timing.t3 c. determinants of active asset allocation decisions five regression models are shown in table 3 which use a standardized measure of active asset allocation as the dependent variable. all observations were pooled. thus, it is possible that strategies used by certain managers might offset opposite strategies used by others. our results apply to the group as a whole. earlier we defined aaa as the difference at the end of a quarter between the actual portfolio beta of a manager and the expected beta if no active allocation had occurred. because managers engage in varying degrees of active asset allocation, we standardized each 32 financial services review, 2(l) 19!ml993 table 1 active asset allocation by quarter quarter average standard deviation minimum maximum 7706 7709 :+ez 7712 0:003 7803 -0.007 7806 0.029 7809 0.020 7812 -0.059 7903 0.013 ;z -0.009 0.014 7912 0.058 8003 -0.028 %z -0.015 0.031 8012 -0.016 8103 0.003 8106 -0.039 8109 -0.061 8112 -0.007 8203 -0.089 8206 -0.007 8209 0.045 8212 8303 :iz 8306 0:015 8309 8312 z:e 8403 a.025 8406 -0.012 8409 0.041 8412 -0.030 8503 -0.042 8506 -0.007 8509 -0.020 8512 0.047 8603 0.029 8606 -0.022 8609 -0.053 8612 0.026 8703 0.026 8706 8709 ee 8712 -0:032 8803 0.014 8806 0.002 8809 -0.057 0.087 -0.373 0.205 0.115 0.102 0.091 0.126 0.172 0.147 0.124 0.083 0.097 0.074 0.127 0.115 0.121 0.112 0.142 0.117 0.109 0.123 0.126 0.184 0.138 0.120 0.162 0.195 0.165 0.175 0.166 0.111 0.089 0.147 0.089 0.091 0.111 0.110 0.134 0.104 0.092 0.121 0.084 0.114 0.101 0.103 0.160 0.107 0.083 xi.118 -0.492 -0.267 -0.659 a.879 -1.090 -0.590 a.200 -0.410 -0.190 -0.210 -0.390 a.458 -0.258 -0.443 -0.468 xi.412 a.316 -0.432 -0.925 -0.537 -0.246 4.346 -0.825 -0.735 -0.944 -0.734 -0.561 -0.248 -0.238 a.355 -0.310 -0.523 -0.430 -0.540 -0.191 -0.405 -0.522 a.178 -0.363 -0.284 -0.385 -0.516 -0.258 a.429 0.683 0.159 0.251 0.544 0.562 0.350 0.126 0.346 0.244 0.363 0.508 0.276 0.241 0.368 0.461 0.399 0.158 0.387 0.339 0.485 0.452 0.623 0.494 0.789 0.451 0.343 0.837 0.363 0.280 0.881 0.247 :.z 0:571 0.530 0.435 0.245 0.155 0.477 0.567 0.329 0.208 0.626 0.543 0.197 aaa observation. the standardized value of aaa was calculated by dividing the raw aaa measure for a given manager in a given quarter by the standard deviation of the given manager’s raw aaa over the sample perk~i.‘~ model 1 regresses the standardized aaa against events which occurred during the quarter in which aaa is observed. all independent variables are statistically active timing de&ions of equity mutual funds 33 table 2 analysis of aaa by fund sorted by standard deviation of quarterly aaa fund average quarterly standard fund avemge quarterly standard number change in aaa deviation of aaa number change in aaa deviation of aaa 62 -0.0333 0.3613 6 a.0108 0.2628 73 -0.0049 0.2572 30 0.0002 0.2421 61 -0.0063 0.2417 78 -0.0092 0.2239 59 -0.0084 0.2185 56 0.0069 0.2122 28 0.0054 0.2117 64 -0.0033 0.1993 92 -4x0133 0.1949 37 -0.0091 0.1939 74 -0.0062 0.1924 53 0.0002 0.1905 35 -0.0021 0.1684 91 -0.0095 0.1623 88 -0.0078 0.1623 63 0.0037 .01536 69 -0.0088 0.1513 29 0.0043 0.1493 19 -0.0208 0.1489 60 -0.0123 0.1485 76 -0.0064 0.1468 47 -0.0010 0.1443 2 -0.0049 0.1426 14 -0.0054 0.1423 72 0.0011 0.1414 23 -0.0003 0.1406 79 -0.0014 0.1385 46 0.0078 0.1303 24 -0.0022 0.1284 22 -0.0014 0.1274 31 0.0005 0.1253 40 -0.0032 0.1238 45 -mio99 0.1227 12 -0.0068 0.1227 13 -0.0142 0.1211 54 -0.0043 0.1204 50 ~0042 0.1112 94 -0.0093 0.1111 84 -0.0043 0.1099 36 -0.0125 0.1082 20 ~003 1 0.1055 57 0.0029 0.1034 9 -0.0035 0.0991 51 -0.0046 0.0989 77 -0.0074 0.0988 49 66 87 11 8 42 41 48 58 34 27 71 7 18 81 86 44 80 10 26 83 4 75 15 55 89 1 90 68 38 43 67 3; 93 52 21 3 16 32 65 70 82 33 85 17 25 0.0003 0.0977 -0.0099 0.0977 0.0005 0.0976 -0.0024 0.0961 0.0002 0.0961 -0.0074 0.0948 -0.0074 0.0947 -0.0122 0.0944 -0.0073 0.0932 -0.oa22 0.0932 -0.0031 0.0915 -0.0096 0.0913 -0.0056 0.0894 -0.0171 0.0893 -0.0117 0.0879 -0.0077 0.0871 -0.0104 0.0862 -0.0059 0.0854 -0.0034 0.0804 -0.0059 0.0792 -0.0065 0.0782 -0.0100 0.0779 -0.0012 0.0768 -0.0009 0.0743 -0.0050 0.0735 -0.0045 0.0734 -0.0151 0.0733 -0.0037 0.0723 -0.0123 0.0705 -0.0065 0.0702 -0.0095 0.0678 -0.0023 0.0669 -0.0061 0.0654 -0.0043 0.0653 -0.0007 0.0646 -0.0048 0.0645 -0.0077 0.0632 -0.0042 0.0595 -0.0035 0.0566 -0.0031 0.0532 -0.0070 0.0506 -0.0053 0.0481 -0.0066 0.0456 -0.0045 0.0428 -0.0055 0.0400 -0.0058 0.0393 -0.0038 0.0391 34 financial services review, 2(l) 19!w1993 table 3 active asset allocation regressions standard&ed aaa as dependent variable m&f f 2 3 4 5 intercept term 2.34 -1.89 -0.07 -0.89 -0.28 (-11.65) (-6.88) (-0.31) (-3.55) (-1.03) -4.08 4.16 -3.79 -3.31 -3.46 (-4.99) (-5.08) (-4.66) (-4.08) (-4.28) 2.21 2.23 pa-4 spt xr;, m-1 xfr-1 csp,_l se-2 cxf-1 xft-2 cm-2 sp&3 cxf;_2 csp,3 spt-4 cxft_3 x6-4 r-square (11.39) (11.51) 1.59 1.66 (4.94) (5.12) -0.46 (-2.37) -0.62 (-1.95) 0.81 (6.48) -0.83 (-4.29) 0.5 1 (2.49) -0.42 (-1.32) 0.31 (2.98) 0.47 (2.37) 0.29 (1.90) 0.31 (3.33) -0.14 (-0.63) 0.15 (1.22) -0.28 (-0.87) 3.77% 3.91% 1.92% 0.78% 0.69% nom: t-statistics are shown in parentheses. significant. the intercept terms were negative; consistent with figure 2 which shows a long-term decline in the beta of the average manager. the passive change in equity allocation which would result from relative returns on stocks and t-bills (paa) is negatively related to active equity asset allocation decisions. in fact, this is true for each of the models shown in table 3 as well as for all other models which we examined in the study. this negative relationship is consistent with managers following an active rebalancing strategy a rebalancing to a desired target equity active timing decisions of equity iuutud funds 35 allocation. quarters in which passive equity allocation increases (decreases), due to relative stock and treasury bill returns, result in active decreases (increases) in equity allocations by the managers. model 1 also suggests that, for the average fund, aaa is positively related to returns on the s&p500 as well as excess fund returns during the contemporaneous quarter. the positive sign on sp, could reflect an opinion by managers that recent market trends will persist. the positive relationship between excess returns on the manager’s portfolio and aaa suggests that, across the full sample, managers with current quarterly returns in excess of the s&p500 tend to increase their commitment to equities. similarly, managers with current quarterly returns less than the s&p500 tend to actively reduce their equity positions. but even though each of the variables in model 1 are statistically significant, the explanatory power of the model is low having an r-square of 3.77 percent. this suggests that anticipatory decisions are much more important to active asset allocation than reactive decisions. in model 2, returns from the previous quarter are included in the regression. the return on the s&p500 in quarter t 1 is negatively related to active changes in equity in quarter t. similarly, excess returns of the manager in quarter t 1 are also negatively related to active equity changes in quarter t. this negative relationship contrasts with the positive relationship in the contemporaneous quarter. the best explanation we can offer is that asset rebalancing is conducted over more than one quarter. for example, if returns on the s&p500 or excess manager returns are large, the equity commitment is reduced during the subsequent quarter. in models 3-5, the affects of quarterly returns lagged 2 through 4 quarters are examined. in none of the cases are quarterly excess manager returns significant. however, lagged s&p500 returns for two and three quarters are significant. the negative sign on the two quarter lag is similar to the one quarter lag. we have no explanation for the positive sign on the three quarter lag. d. the september hypothesis to test the september hypothesis, the sample was first split into four groups; one for each quarter-end in a year, march, june, september and december. next, the excess return of each manager was calculated from the start of the calendar year through the end of the given quarter. we refer to this year-to-date cumulative excess fund return as ycxf. finally, each of the four quarterly groups were sorted into quintiles based on the levels of ycxf. funds in rank 1 had the lowest ycxf and funds in rank 5 had the highest ycxf. illustrative regressions results are shown in table 4 using paa and ycxf as independent variables and aaa as the dependent variable. when regressions are run on the unranked groups, only paa is significant, the year-to-date cumulative excess fund return is not significant. and when regressions are run on each of the ycxfquintile rankings, ycxf is again usually insignificant. however, there is one major exception. in september, the funds in rank 1 have a statistically significant 36 financial services review, 21) 1992/19!23 table 4 tests of the september hypothesk dependent variable = aaa march paa ycxf paa ycxf ai1 -4.57 0.53 -1.14 -1.33 rank1 3.19 0.57 -1.39 1.08 rank2 1.77 -0.43 -2.70 -2.30 rank3 -0.35 -1.93 -1.40 2.20 rank4 -7.20 -0.94 0.49 0.27 rank5 -4.68 1.67 0.57 -0.77 ~ep~mber december pm ycxf paa ycxf -2.40 0.40 -4.47 0.56 -4.29 -3.54 -4.42 0.94 3.65 0.37 -4.26 -2.40 0.75 0.81 -1.80 -1.57 1.40 0.97 -2.75 1.01 -6.50 -1.12 0.22 1.17 note: entries in table are t-statistics. negative sign on ycxf. this is consistent with the hypothesis that, late in the year, managers with low returns relative to the s&p500 take on greater amounts of equity risk exposure. although this can also be seen in june and december for other low ranked ycxf groups, the significance level is smaller. iv. conclusions previous studies have found evidence that portfolio managers attempt to anticipate future relative returns on aggregate equities versus risk-free securities and alter the portfolio’s equity risk exposure to profit from the manager’s predictions. these studies were unable, however, to accurately measure portfolio risk at specific points in time since they relied solely on the time series returns of a portfolio. in this study we calculate cross-sectional portfolio betas using the security holdings of a portfolio at a given quarter-end. as such we are able to accurately measure changes in equity risk exposure. using a sample of 94 u.s. equity mutual funds over the period march 3 1,1977 through december 31, 1988, we examine the extent to which portfolio managers engage in active management of their portfolios’ equity risk exposure. active management is measured as the difference between a portfolio’s cross-sectional beta at the end of a quarter and the beta which would have been expected if no active asset allocation had been used. observation shows that cross-sectional portfolio betas can differ significantly from single time series beta estimates. the time series beta is a poor predictor of a portfolio’s current equity risk exposure and provides no information about the extent to which the manager engages in active asset allocation. the change in a manager’s beta from one quarter to the next arises from both an active and a passive asset allocation. active asset allocation is defined as the actual beta of a portfolio at a given quarter end minus the expected beta (given the prior quarter beta and relative returns on equities and risk free securities during the quarter). passive asset allocation is defined as the expected beta minus the prior active timing decisions of equity mutual funds 37 quarter’s beta. the dominant cause for changes in portfolio betas is active allocation. for example, it is not unusual to find managers who actively increase or decrease their portfolio beta by 0.5 between two quarters. and the median standard deviation of quarterly active changes in beta was 0.098. we also examine whether active asset allocation decisions are better charac terized as reactive or anticipatory. reactive decisions are said to be traceable to current and past equity returns of either the aggregate stock market or the individual manager. although most decisions appear to be best classified as anticipatory, there is some weak evidence of reactive decisions. finally, we find evidence that managers who experience poor year-to-date returns relative to the s&psoo, increase their equity risk exposure in the last quarter of the calendar year. what do these results mean to the individual investor? first, given the large variability in a portfolio’s beta which is possible over time, it is unfortunate that individual investors are not provided with information about a mutual fund’s current and past cross-sectional betas. without this information, it is very difficult to judge the extent to which a manager engages in timing activities and the portfolio’s current equity risk exposure. although this data is provided by consultants to large investors such as pension funds and endowment funds, it is not available to individual investors. second, if a fund which has poor yearly returns as of september is being considered for purchase, the investor should check to see that the manager has not recently increased equity risk beyond the investor’ risk tolerance. notes 1. the most recent study discussing problems inherent in measuring the performance of ptofes sional managers when they engage in active management of equity risk can be found in grinblatt and titman (1989). they demonstrate that informed managers who have positive timing abilities can appear to uninformed investors as having negative timing ability and larger calculated betas than actually present in the portfolio. this can happen if performance measures are calculated using the time series of the portfolio’s returns instead of examining changes over time in a portfolio’s cross-sectional beta. 2. we do not examine whether market timing activities of fund managers are successful. instead, we focus on the extent to which timing decisions cause changes in a portfolio’s equity risk exposure over time and the determinants of such decisions. 3. consultants to large portfolio owners such as pensions and endowment funds regularly track and report on the level of equity risk exposure of portfolio managers employed by their clients. unfortunately, similar information is not currently provided topubiic investors in mutual funds. 4. kon and chen (1978) p. 471. 5. occasionally the funds in our sample did have small positions (5 percent or less) in fixed income securities. we do not believe that such small positions seriously damage any results of the study. 6. this calculation, of course, assumes that no active asset reallocations ate made during quarter t. 7. we admit that this classification scheme is quite broad and that more precise knowledge of the factors which should cause specific managers to alter their equity risk exposure should be researched. however, this is the fist study to examine identifiable changes in equity risk exposure. as such, we believed that it was important to frame the issue in as fundamental a manner as possible. 38 financial services review, 2(l) 199w1993 8. 9. 10. 11. 12. 13. 14. we choose the s&p500 index for two reasons. fit, it is the most easily accessed in&x to the general public and is, thus, the most widely used index to compare equity mutual fund returns against. second, we wanted to include data extending through the end of 1988 and the crsp indexes were not available for 1988. a regression of s&p500 returns against crsp returns for the period in our sample for which both series were available resulted in an s&p500 beta of 1 .o 1 and an r-square of 99.4 percent. relative returns am calculated by substracting returns on the s&p500 from returns on the fund. this implicitly assumes that mutual fund owners believe the beta of the fund to be 1 .o. although not reported here, we also tested excess fund returns based on each fund’s actual beta, results were no different from those reported here. this could be due to a number of reasons: (1) mutual fund investors were unable to differentiate between fund beta levels, (2) they were unconcerned about beta levels and made relative fund comparisons solely against the s&p500 or (3) sample error. crsp data for 1988 was not available. stock betas calculated for december, 1987 were used for all quarters in 1988. the beta for asset holdings in excess of the cumulative value of stock holdings were assigned a beta of 0.00. in virtually all cases, non-stock holdings were money market securities. the accuracy of both aaa and paa depend on the accuracy of expected portfolio returns as calculated in equation (5). we conducted a variety of tests comparing the sequence of expected returns generated by equation (5) and actual fund returns. detailed results am available from the authors. it is our opinion that equation (5) does not impart a significant bias to the calculation of either paa or aaa. a list of funds used in this study is available from the authors. our analysis does not consider the extent to which cash inflows or outflows to the portfolio might be the cause for changes in a portfolio’s stock to asset ratio. this is an interesting issue and deserves study. however, in conversations with a number of the mutual fund managers covered in our sample, they stated that only about two percent of assets am necessary to maintain liquidity for net redemptions and that cash from net new sales is invested within a day. results were very similar when the raw aaa variables were used. references amott, robert d. and roy d. henriksson. 1989. “a disciplined approach to global asset allocation,” financial analysts journal, 45: 17-28. bauer, richard j., patrick a. hays and david e. upton. 1987. “parameter instability in mutual fund portfolios: a shifting regimestest,“quurterly journalofbusinessandeconomics,pp.5(m2. breen, william, ravi jagannathan and ahamn r. ofer. 1986. “correcting for heterocedasticity in tests of market timing ability,” journal of business, 59: 585-598. chang, eric c. and wilbur g. lewellen. 1984. “market timing and mutual fund performance,” journal of business, 57: 57-72. fabozzi, frank j. and jack c. francis. 1979. “mutual fund systematic risk for bull and bear markets: an empirical examination,” the journal of finance, 34: 1243-1250. grinblatt, mark and sheridan t&man. 1989. “portfolio performance evaluation: old issues and new insights,” review of financial studies, 2: 393421. grinblatt, mark and sheridan titman. 1988. “adverse risk incentives and the design of performance based contracts.” working paper, rodney l. white center for financial research. henriksson, roy d. and robert c. merton. 1981. “gn market timing and investment performance. ii. statistical procedures for evaluating forecasting skills,” journal of business, 54: 5 13-534 henriksson, roy d. 1984. “market timing and mutual fund performance: an empirical investiga tion,” journal of business, 57: 73-96. active timing de&ions of equity mutual fun& 39 ippolito, richard a. 1989. “efftciency with costly information: a study ofmutual fund performance. 1965-1984,” the quarterly journal of economics, 104: i-23. jagannathan, ravi and robert a. korajczyk. 1986. “assessing the market timing performance of managed portfolios,” jou~l of iciness, 59: 217-236. jensen, michael c. 1972. “optimal util~adon of market forecasts and the evaluation of investment performance,” mathematical methods in irtvestment and finance edited by szego and shell. north holland press. joehnk, michael d. 1987. asset allocation for instihrtionai portfolios. the institute of chartered financial analysts. kon, stanley j. and frank c. jen. 1978. “estimation of time-varying systematic risk and perform ance for mutual fund portfolios: an application of switching regression,” the journal of finance, 33: 457-475. kon, stanley j. and frank c. jen. 1979. ‘the investment performance of mutual funds: an empirical investi~tion of timing, selectivity, and market efficiency: journal of business, 52: 263-289. kon, stanley j. 1983. ‘the meet-timing performance of mutual fund managers,” journal of business, 56: 323-348. lee, cheng-few and shafiqur rahman. 1990. “market timing, selectivity, and mutual fund perform ance: an empirical investigation,” journal of business, 63: 261-278. metton, robert c. 1981. “on market timing and investment performance. i. an equilibrium theory of value for market forecasts,” journal of business, 54: 363-406. fsr v31 i4 masthead academy of financial services officers president tom potts baylor university executive vice president-program terrance k. martin winston-salem state university vice president-finance (interim) thanh ngo east carolina university vice president-international relations michelle cull western sydney university vice president-mktg & public relations shawn brayman smb research consulting vice president-membership cora pettipas hsbc global wealth immediate past president inga timmerman shepherd university editor, financial services review john e. grable, ph.d., cfp®, university of georgia directors jason anderson university of kansas jasmine fand massey university norah feng massey university matt goren (interim) dalton eduction wookjae heo purdue university philip gibson winthrop university barry mulholland university of akron past presidents inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 31, no. 4, 2023 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • sherman hanna, ph.d., the ohio state university • terrance martin, ph.d., winston-salem state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • stu heckman, ph.d., cfp®, texas tech university • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • thomas langdon, ph.d., roger william university, bristol, ri • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • shinae choi, ph.d., university of alabama • john clinebell, ph..d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., york university, canada • lu fan, ph.d., cfp®, university of georgia • patti fisher, ph.d., virginia tech • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph..d. stetson university, deland, fl • philip gibson, ph.d., cfp®, winthrop university • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • john nofsinger, ph.d., university of alaska anchorage • mustafa nourallah, ph.d., centre for research on economic relations • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia, support of the journal financial services review (fsr) is the official publication of the academy of financial services. it is a diamond open access journal which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peer-reviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. risk and uncertainty in style rotation timothy a. krausea,* ablack school of business, penn state behrend, 5101 jordan road, erie, pa 16563, usa abstract the chicago board options exchange (cboe vix; volatility) index has been established as a leading indicator of style returns since increases in this “fear index” lead to outperformance of “value” versus “growth” stocks. this study introduces the concept of “uncertainty” as an additional indicator of returns to value, as measured by the cboe vvix (“volatility of volatility”). this index is considered to be a proxy for “uncertainty” in the knightian sense. increases in expected volatility lead to short-term positive returns to value, while increases in uncertainty lead to negative short-term returns to value. each of these observations are especially strong during economic downturns and after decreases in the vix index. several macroeconomic indicators provide additional incremental information regarding these phenomena. © 2018 academy of financial services. all rights reserved. jel classification: g1; c1; c5 keywords: vix index; vvix index; volatility; exchange traded fund; etf; uncertainty; value versus growth; style returns 1. introduction this article examines equity financial market relationships among risk, uncertainty, and returns to value and growth stocks. specifically, the analysis examines the effectiveness of the chicago board options exchange (cboe) volatility indices as leading indicators of style returns (value vs. growth). the effectiveness of the volatility index (the cboe vix index, also known as the “fear index” in the financial press) as a leading indicator of style returns is examined by copeland and copeland (1999), who find that increases in the vix index lead to the outperformance of value-based indexes relative to growth-based indexes. boscaljon, * corresponding author. tel.: �1-814-898-6326; fax: �1-814-898-6223. e-mail address: tak25@psu.edu financial services review 27 (2018) 189-207 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. filbeck, and zhao (2011) find that these effects have diminished over time and are more recently only observable over longer return horizon periods. the theoretical underpinnings of both of these papers postulate that investors gravitate towards “value” in times of expected market turbulence (increased volatility and therefore increased risk). this supposition is first proposed by merton (1980) and french, schwert, and stambaugh (1987), who suggest a positive relationship between the market risk premium and expected future volatility that is related to the asymmetric volatility phenomenon, or, alternatively, the “leverage effect.” hence, lower beta value stocks perform better as stock prices fall in the face of expected increases in future volatility. the converse applies to higher beta growth stocks. to explore these issues further, precise definitions of risk and uncertainty are necessary. one generally accepted definition of risk is described as being exposed to a set of possible outcomes that may be observed under an objective expected probability distribution. however, many economists posit that this measure of risk does not capture the effect of “uncertainty,” which describes the effect of “unknowns” as a distinction between risk (volatility defined by an expected probability distribution of outcomes) and “unmeasurable uncertainty” (knight, 1921, p. 245) that is defined by expected outcomes over an unknown probability distribution. in contrast to the generally accepted definition of risk, uncertainty is defined by subjective assessments of possible probability distributions determined by market participants. therefore, expected volatility is time-varying and heterogeneous among market participants. first proposed by knight (1921) and supported by keynes (1921) and ellsberg (1961), this concept is often referred to as “knightian uncertainty,” suggesting a distinction between risk (volatility that can be measured using probabilities) and “unmeasurable uncertainty” (knight, 1921, p. 245). baltussen, van bekkum, and van der grient (2017)) propose an intuitive proxy for uncertainty in equity markets that is constructed from variations in the implied volatilities of single stock options that they denote as “volatility of volatility,” or “vol-of-vol.” they calculate the monthly average standard deviation of implied volatilities of at-the-money single stock options and demonstrate that “stocks with high uncertainty about risk… robustly underperform stocks with low uncertainty about risk: (p. 1). at the aggregate market level, the cboe provides a “volatility of volatility” index (vvix) that is constructed using implied volatilities of vix index options. therefore, this measure is an explicitly forwardlooking measure of the potential future distribution of returns of the s&p 500 index, reflecting the expectations of market participants, and not prior implied volatility information. further details regarding the construction of the vvix index can be found in a cboe white paper (cboe, 2012). this measure is examined recently by aboura and arisoy (2017) in a study of portfolio style returns, who find that “due to their negative uncertainty betas, uncertainty-averse investors demand extra compensation to hold small and value stocks” (p. 3217). the present analysis offers a similar explanation of the size and value anomalies in highly liquid exchange traded funds (etfs; as opposed to individual stock portfolios), and the results are obtained on a lead-lag, as opposed to a contemporaneous, basis. thus, this study may more directly reflect the future performance of “style” based investing based on investor perceptions of uncertainty. in sum, the results of the present study indicate that increases in expected volatility lead to short-term positive returns to value, but increases in uncertainty lead to negative short-term returns to value. each of these observations are 190 t.a. krause / financial services review 27 (2018) 189-207 especially strong during economic downturns and following decreases in the vix index, and these observations should spur further research into the behavior of investors that expands the traditional mean-variance framework. in support of this observation, talukdar, daigler, and parhizgari (2017), posit that “behavioral theories explain the return–volatility relation better than the fundamental theories” (p. 698). mayfield and wooten (2009) propose that investor decisions are influenced by their personality type, a further indication of potential investor heterogeneity. additionally, below, kiely, and prati (2009) demonstrate the advantages to frequent rebalancing across equity “styles” to achieve above-average performance. although the evidence from the behavioral finance literature is still being developed and contradictory results have been obtained because of differentials in the information under study (time frames, sample selection, etc.), it is not for a lack of effort. the relationships among future stock returns and proxies for uncertainty are examined relative to economic policy by brogaard and detzel (2015), who find a positive relationship between economic policy uncertainty and future excess market returns. however, ko and lee (2015) find a negative relationship between these variables using wavelet analysis on a contemporaneous basis. su, fang, and yin (2017) investigate “news-based” uncertainty that predicts future market volatility. the results of these studies indicate varying relationships between uncertainty and returns, and the conflation of contemporaneous and predictive results makes it difficult to disentangle the issue. additional studies find positive relationships between differing proxies for uncertainty and future stock returns. for example, bekaert, engstrom, and xing (2009) and bali, brown, and caglayan (2014) examine the conditional correlations of fundamental equity characteristics and macroeconomic variables, respectively. each of these studies find positive lead-lag relationships between their proxies for uncertainty “betas” and stock returns and hedge fund returns, respectively. krause (2018) finds similar results using the vvix index as a proxy for uncertainty. this study examines the effectiveness of the two cboe volatility indices as leading indicators of style returns (value vs. growth), and the results of the analysis indicate that the cboe vvix index provides significant incremental information regarding the interaction of returns, volatility, and uncertainty on a lead-lag basis. the initial analysis of the vix index relative to style returns is consistent with boscaljon, filbeck, and zhao (2011) because it finds largely insignificant short-term effects of the vix index on returns to value. however, innovations in the vvix index indicate significant negative returns to value. the inclusion of several macroeconomic factors provides additional explanatory information since the vix index indicates positive returns to value under certain conditions. the main contribution to the literature of this paper is the introduction of the additional concept of “uncertainty” into the returns to value analysis using highly liquid etfs. the availability of these products, and their recent exponential growth, provides an opportunity to examine the relation of expected volatility and uncertainty to growth and value using similar, easily tradable and low-cost instruments. the results stand in contrast to prior studies that examine msci barra and/or s&p 500 value and growth portfolios that may be costly and or difficult to implement. also, short sale constraints are virtually nonexistent for the etfs under study. the presence of short sale constraints may limit the effectiveness of other studies that examine portfolios of single stocks, as posited by shleifer and vishny (1997). therefore, the use of etf return time series’ allows for a more practical analysis of the data. the study also contributes to the 191t.a. krause / financial services review 27 (2018) 189-207 literature by examining macroeconomic variables that contribute to the explanatory power of econometric models with robust standard errors. several studies demonstrate that forward-looking implied volatility measures such as the vix index provide predictive evidence regarding future realized volatility and returns. ammann, skovmand, and verhofen (2009) find a positive relation between implied volatility and future realized volatility in single stock options, while both anderson, bollerslev, diebold, and labys (2003) and blair, poon, and taylor (2001) find a similar relationship for s&p 100 index options. these studies examine the implied volatility of equity options as an indicator of investor expectations regarding future equity volatility. the results of early studies are somewhat mixed. but most recent studies confirm a generally positive relationship between implied volatility and future realized volatility. in early work, canina and figlewski (1993) demonstrate that implied volatility of the s&p 100 index is a poor predictor of future realized volatility. additionally, jiang and tian (2005) examine the relation between past realized volatility and future realized volatility in s&p 500 index options, finding it to be a more reliable indicator than implied volatility. chan, jha, and kalimipalli (2009) find that historical volatility is not a reliable predictor of future implied volatility for s&p 500 index options. however, in more recent studies, both guo and whitelaw (2006) and bali and peng (2006) find a positive relation between the vix index and future stock returns. sarwar (2005) finds a positive relation between implied volatility and options trading volume in s&p 500 index options. demiguel, plyakha, upal, and vilkov (2013) find that implied volatility is a useful factor to consider in the selection of efficient mean-variance portfolios, since single stock implied volatility is useful in forecasts of both future volatility and returns of s&p 500 component stocks on a daily and intraday basis. similarly, giot (2005) finds a positive relationship between the vix index and future stock returns, confirming the results of copeland and copeland (1999) and boscaljon, filbeck, and zhao (2011) that are partially supported by the results in this article. an, ang, bali, and cakici (2014) finds a positive relation among in increases in call implied volatilities and future stock returns, and additional evidence on this topic is provided by bali, cakici, and chabi-yo (2015), and brous, ince, and popova (2009). these studies confirm the positive relation between expected returns and volatility that is first proposed by markowitz (1952). this study considers “risk” to be defined as the volatility of the expected return probability distribution (the proxy is the vix index), consistent with the traditional approach to mean-variance analysis. similarly, “uncertainty” is defined as the volatility of this volatility, and the proxy is the vvix index. as noted in a cboe white paper (cboe, 2012), the vvix measures the implied volatility of options on the 30-day forward price of the cboe vix index. while this index measures expected risk in the future distribution of returns, the vvix index measures variability in expectations regarding this distribution. thus, similar to the study of baltussen, van bekkum, and van der grient (2018), it is a natural proxy for uncertainty about the future distribution of returns that may not conform to a pure meanvariance framework. there remain considerable differences of opinion regarding the theoretical and empirical evidence on the relations among stock returns and uncertainty. baltussen, van bekkum, and van der grient (2018) explore this issue relative to single stocks using a proxy for uncertainty that is similar to the vvix index (“vol of vol” in single stocks) to find a negative 192 t.a. krause / financial services review 27 (2018) 189-207 relation between uncertainty and single stock returns, an indication of uncertainty-avoiding behavior. brenner and izhakian (2012) find similar results for the spdr s&p 500 etf using intraday data, although they use a slightly different measure of uncertainty. aboura and arisoy (2017) provide some theory supporting the idea that uncertainty is a significant part of the equity risk premium and should be related to expected returns. the authors also conduct an empirical study of the contemporaneous returns of single-stock portfolios sorted on various characteristics relative to the vvix index on a contemporaneous basis. the research supports the supposition that stock portfolios reflect negative “uncertainty betas” (p. 3214) that are sensitive to market capitalization. the present study sheds some new light on the topic since the relation between uncertainty and returns for the sample of etfs is examined in light of important macroeconomic variables. the distinction between volatility and uncertainty is illustrated in fig. 1. although the two measures are positively correlated (pearson correlation coefficient of 0.71 over the sample period) and generally move together, that is not always the case. as shown in the graph, uncertainty increased significantly in 2014 without a concurrent increase in volatility. additionally, the indexes diverge significantly during the financial crisis of 2007–2009 and during the european debt crisis of 2010–2012. the literature regarding the use of the volatility measures to enhance asset allocation decisions is well-documented in boscaljon, filbeck, and zhao (2011). additionally, goldwhite (2009) demonstrates that the vix index is a useful indicator for investors with different levels of risk aversion, because he examines the relationships among volatility and the returns to value and growth stocks. puttonen and seppä (2006) further document the value added by an active approach to investing in value and growth stock indexes. finally, talukdar, daigler, and parhizgari (2017) find that the vvix index is an important driver of the vix index and its relation to future stock returns. fig. 1. this graph presents values of the vix and vvix index over the past decade (in percentage). overall, the graph indicates a generally positive correlation over the sample period although there are significant deviations from this observation, especially during financial crises. 193t.a. krause / financial services review 27 (2018) 189-207 2. data sample and methodology the cboe provides daily closing levels for the vvix index beginning on june 1, 2006, so daily vix and vvix closing levels are collected from june 1, 2006 to march 31, 2017. daily total return data (including dividends) for six ishares value and growth etfs (large-, mid-, and small-cap) are obtained from bloomberg professional for the same time period. summary statistics are provided for the volatility indexes and the six style-based etfs in table 1. as is evident, each of these etfs has been in existence for over 15 years, and they are all large, liquid instruments available to easily implement style-based trading strategies, with limited short sale constraints, as opposed to the previously examined msci barra value and style indices. copeland and copeland (1999) argue that the value and growth s&p futures been have available since 1997, making them easily tradeable, but that does not obviate the fact that their analysis is based on index, not futures, data. the illiquid futures contracts based on these indices are only available for the s&p 500 (large-cap) index, and they never really caught on as liquid trading products. as of may 26, 2017, the open interest in the s&p 500 index value contracts was just twenty-six contracts while the trading volume on that day was zero. boscaljon, filbeck, and zhao (2011) also utilize cash indices for their analysis, which are not easily tradeable at a reasonable cost. for instance, there are currently 1,808 constituents in the msci u.s.a. small-cap index. however, the present analysis explicitly examines lead-lag relationships among highly liquid etfs that are easily tradeable. as seen in table 1, the relatively newly available and liquid etfs are similar in terms of total net assets and trading volume (the average daily trading volume figures in the table are the most recent three-month average as of march 31, 2017). one exception to this generalization is ivw, the large-cap growth etf, which is significantly larger than the others. table 2 provides summary statistics for the data, where the figures for the volatility indices are daily closing levels and the etf data are daily returns. over the sample time period, each of the etfs experiences similar returns, although standard deviations decline monotonically as market capitalization rises, because the larger capitalization stocks experience lower levels of volatility. a correlation matrix for daily changes in the variables is provided in table 3, and the first item of interest is the same positive relation between the vix and vvix (pearson correlation coefficient of 0.71) that is present in fig. 1. aboura and arisoy (2017) table 1 exchange traded fund (etf) descriptions, as of march 31, 2017 etf symbol net assets ($) average daily trading volume ($) annual mgmt. fee beta inception date ishares s&p small-cap 600 value etf ijs 4.69b 23.4m 0.25% 1.59 07/24/00 ishares s&p small-cap 600 growth etf ijt 4.21b 21.1m 0.25% 1.68 07/24/00 ishares s&p mid-cap 400 value etf ijj 5.62b 19.7m 0.25% 1.31 07/24/00 ishares s&p mid-cap 400 growth etf ijk 5.48b 23.0m 0.25% 1.34 07/24/00 ishares s&p 500 value etf ive 13.60b 78.8m 0.18% 1.09 05/22/00 ishares s&p 500 growth etf ivw 17.47b 96.6m 0.18% 1.02 05/22/00 data is reported by morgan stanley & co., llc and blackrock investments, llc. average daily trading volume is for the prior 30 days as of march 31, 2017. 194 t.a. krause / financial services review 27 (2018) 189-207 and talukda, daigler, and parhizgari (2017) also use the vvix index as a proxy for uncertainty in a different empirical framework using stock portfolios sorted on various characteristics and as a determinant of changes in the vix index, respectively. as noted previously, as in others, their studies examine contemporaneous rather than lead-lag relationships. in table 3, the usual negative relation is observed between the vix index and contemporaneous returns during the financial crisis, a result of the well-known asymmetric volatility phenomenon, or leverage effect (correlations approximating �0.70 for each of the etfs). additionally, for all six of the etfs, this negative relation is near �0.50 for the vvix index, suggesting that it too may provide information regarding future payoffs to value and growth in addition to the vix index. finally, with one exception, all of the correlations among etf pairs are above 0.90, suggesting, ex ante, that it may be difficult to use volatility and/or uncertainty information to forecast differential returns to style. despite this difficulty, the following empirical analysis is designed to disentangle these relationships. following copeland and copeland (1999) and boscaljon, filbeck, and zhao (2011), the first examination of the data are to model several different future return windows as a function of changes in the vix and vvix indexes. one deviation from their approach is that standard errors are now estimated with heteroskedasticityand autocorrelation-consistent (hac) errors, using the robust procedure of newey and west (1987) with five lags to represent one week of trading activity. the following robust ols equations are estimated: table 2 summary statistics variable symbol n mean (%) standard deviation skewness kurtosis min. max. volatility index vix 2,728 20.10 9.60 2.38 10.49 9.89 80.86 vol of vol index vvix 2,728 87.22 13.14 0.86 4.67 36.14 168.75 small value ijs 2,728 0.044 0.0162 �0.17 8.31 �0.12 0.09 small growth ijt 2,728 0.047 0.0015 �0.23 7.80 �0.10 0.09 mid value ijj 2,728 0.043 0.0146 �0.20 10.89 �0.11 0.11 mid growth ijk 2,728 0.046 0.0140 �0.38 8.61 �0.10 0.09 large value ive 2,728 0.034 0.0133 �0.17 11.50 �0.09 0.11 large growth ivw 2,728 0.043 0.0119 �0.14 12.61 �0.09 0.11 this table presents summary statistics for the variables under study. the volatility indexes are presented as levels while the return data is presented as daily changes. table 3 correlation matrix of volatility variables and exchange traded fund (etf) returns variable symbol vix vvix sm vl sm gr md vl md gr lg vl lg gr � volatility index vix 1.0000 � vol of vol index vvix 0.7142 1.0000 small value ijs �0.6794 �0.4381 1.0000 small growth ijt �0.7053 �0.463 0.9745 1.0000 mid value ijj �0.6996 �0.4577 0.964 0.9549 1.0000 mid growth ijk �0.7215 �0.4769 0.936 0.9616 0.9657 1.0000 large value ive �0.7238 �0.4762 0.9182 0.9026 0.9493 0.9160 1.0000 large growth ivw �0.7511 �0.5043 0.8871 0.9095 0.9243 0.9389 0.9427 1.0000 this table provides a correlation matrix of the primary variables under study in this article. 195t.a. krause / financial services review 27 (2018) 189-207 ret �valuei,t�n � growthj,t�n� � � � �1�vixt � �i, j,t (1) where ret �valuei,t�n � growthj,t�n� represents the relevant time period return (from n equals one to 60 days in discrete increments) for a long position in the value etf i (e.g., ive, the ishares s&p 500 value etf) and an equal short position in the growth etf j (e.g., ivw, the ishares s&p 500 growth etf) for each of the three size-based etf classifications. �vixt represents daily changes in the levels of the vix index on day 0. this equation is identical to the specifications of copeland and copeland (1999) and boscaljon, filbeck, and zhao (2011) for msci barra size and value indices. additionally, a second equation is estimated that adds daily changes in the vvix index as a proxy for uncertainty to determine whether it possesses further explanatory power for future returns to value: ret �valuei,t�n � growthj,t�n� � � � �1�vixt � �2�vvixt � �i, j,t (2) where �vvixt represents daily changes in the levels of the vvix index on day 0. the standard errors of these estimations are also estimated using the newey and west (1987) procedure with five lags to represent one week of trading activity. 3. empirical results 3.1. initial estimations the results of the estimations of eq. (1) for the large-cap etfs are presented in panel a of table 4. the coefficients in panel a for the large-cap etfs present the returns to value from one-day changes in the vix. these estimations report continued declines in the return-to-value predictability for the vix index over the years, since there are no significant coefficients for the large-cap etfs, and it is clear that positive returns to value are not observable for these highly liquid and efficient liquid etfs over the past decade. it seems that the “returns to value” strategies presented in copeland and copeland (1999) and boscaljon, filbeck, and zhao (2011) would likely not be profitable with these highly efficient and liquid etfs. additionally, only one of the coefficients for returns to value from the vix index (the two-day time return window) are statistically significant for the smalland midcap indexes, respectively. these coefficients are negative, which is a counterintuitive result given the expected and documented positive relation between risk and return. to further explore the returns to value from uncertainty as proxied by the vvix index, in table 5, changes in the vvix index are included in the estimations of eq. (2) as a potentially further explanatory, independent variable. in this estimation, there is one indication of the potential returns to value from volatility in conjunction with uncertainty. in panel a, for the large-cap etfs, the results for five-day returns to value are significantly positive for changes in the vix index (volatility) at the five percentage level, although some other coefficients (10and 20-day) are significant at the 10% level. additionally, the coefficients are significant and negative for changes in the vvix index (uncertainty) over fiveto 30-day time periods (the 20-day coefficient is marginally significant) at the five percentage level. the coefficients 196 t.a. krause / financial services review 27 (2018) 189-207 t ab le 4 n ew ey an d w es t (1 98 7) re gr es si on s of on eda y va lu e m in us gr ow th e xc ha ng e t ra de d fu nd (e t f) re tu rn s ag ai ns t ch an ge s in th e v ix in de x pa ne l a — l ar ge ca p v ix on ly d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 05 3 � 0. 04 6 0. 24 1 0. 15 2 0. 29 2 � 0. 18 8 0. 03 4 � 0. 08 2 0. 00 4 (� 0. 38 ) (� 0. 29 ) (1 .1 4) (0 .5 2) (0 .6 8) (� 0. 39 ) (0 .0 6) (� 0. 13 ) (0 .0 1) c on st an t � 0. 00 9 � 0. 01 8 � 0. 04 7 � 0. 09 4* � 0. 18 3* * � 0. 26 7* * � 0. 35 4* ** � 0. 44 4* ** � 0. 53 3* ** (� 1. 06 ) (� 1. 13 ) (� 1. 32 ) (� 1. 67 ) (� 2. 19 ) (� 2. 53 ) (� 2. 83 ) (� 3. 14 ) (� 3. 47 ) o bs er va tio ns 2, 72 8 2, 72 7 2, 72 4 2, 71 9 2, 70 9 2, 69 9 2, 68 9 2, 67 9 2, 66 9 a dj us te d r 2 0. 00 1 0. 00 2 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 pa ne l b — m id ca p v ix on ly d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 15 5 � 0. 34 9* * � 0. 28 6 � 0. 23 0 � 0. 32 4 � 0. 51 8 � 0. 12 5 0. 30 1 0. 15 3 (� 1. 45 5) (� 2. 61 7) (� 1. 35 6) (� 0. 91 2) (� 0. 96 6) (� 1. 17 8) (� 0. 26 6) (0 .5 42 ) (0 .2 48 ) c on st an t � 0. 00 3 � 0. 00 6 � 0. 01 9 � 0. 04 0 � 0. 07 6 � 0. 11 3 � 0. 15 3 � 0. 20 3 � 0. 25 5* (� 0. 44 1) (� 0. 52 9) (� 0. 70 6) (� 0. 89 8) (� 1. 12 2) (� 1. 32 1) (� 1. 50 6) (� 1. 77 1) (� 2. 02 7) o bs er va tio ns 2, 72 8 2, 72 7 2, 72 4 2, 71 9 2, 70 9 2, 69 9 2, 68 9 2, 67 9 2, 66 9 a dj us te d r 2 0. 00 1 0. 00 2 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 pa ne l c — sm al l ca p v ix on ly d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 10 9 � 0. 25 7* � 0. 21 0 � 0. 36 1 � 0. 51 2 � 0. 59 8 � 0. 34 4 � 0. 25 4 � 0. 26 4 (� 1. 01 9) (� 1. 99 1) (� 1. 03 1) (� 1. 53 8) (� 1. 49 4) (� 1. 41 5) (� 0. 70 0) (� 0. 45 4) (� 0. 44 3) c on st an t � 0. 00 3 � 0. 00 7 � 0. 01 9 � 0. 04 2 � 0. 08 3 � 0. 12 7 � 0. 17 1 � 0. 22 3* � 0. 28 1* (� 0. 43 6) (� 0. 56 7) (� 0. 75 4) (� 0. 99 2) (� 1. 30 4) (� 1. 54 5) (� 1. 73 2) (� 1. 98 7) (� 2. 28 3) o bs er va tio ns 2, 72 8 2, 72 7 2, 72 4 2, 71 9 2, 70 9 2, 69 9 2, 68 9 2, 67 9 2, 66 9 a dj us te d r 2 0. 00 0 0. 00 1 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 t hi s ta bl e pr es en ts th e re su lts of es tim at io ns of e q. (1 ). t he de pe nd en tv ar ia bl e is da ily re tu rn s to a po rt fo lio th at co nt ai ns a lo ng po si tio n in th e va lu e e t f an d an eq ua ls ho rt po si tio n in th e gr ow th e t f fo r ea ch fo r ea ch si ze ca te go ry .f or ea se of in te rp re ta tio n, al lo f th e co ef fic ie nt s ha ve be en m ul tip lie d by 10 0. t hu s, fo r ex am pl e, if th e co ef fic ie nt is 1. 00 ,a on e pe rc en ta ge in cr ea se in th e v ix in de x le ad s to a on e ba si s po in t re tu rn di ff er en tia l ov er th e re le va nt tim e fr am e. *s ig ni fic an t at th e 5% le ve l. ** si gn ifi ca nt at th e 1% le ve l. t he st an da rd er ro rs ar e es tim at ed us in g fiv e la gs to ac co un t fo r on e w ee k of tr ad in g ac tiv ity au to co rr el at io n an d he te ro sk ed as tic ity (n ew ey an d w es t, 19 87 ). 197t.a. krause / financial services review 27 (2018) 189-207 t ab le 5 n ew ey an d w es t (1 98 7) re gr es si on s of on eda y va lu e m in us gr ow th e xc ha ng e t ra de d fu nd (e t f) re tu rn s ag ai ns t ch an ge s in th e v ix an d v v ix in di ce s pa ne l a — l ar ge ca p v ix an d v v ix d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 01 5 0. 18 9 0. 72 5* * 0. 88 6* 1. 16 5* 0. 96 1 1. 10 7 0. 33 7 � 0. 05 8 (� 0. 07 7) (0 .8 36 ) (2 .2 00 ) (1 .7 70 ) (1 .7 04 ) (1 .1 64 ) (1 .1 86 ) (0 .3 33 ) (� 0. 05 3) � v v ix � 0. 07 5 � 0. 45 2* � 0. 93 2* * � 1. 41 7* * � 1. 68 6* � 2. 21 8* * � 2. 07 1* � 0. 80 9 0. 12 0 (� 0. 35 4) (� 1. 65 3) (� 2. 16 5) (� 2. 10 2) (� 1. 92 4) (� 1. 99 5) (� 1. 65 2) (� 0. 60 5) (0 .0 81 ) c on st an t � 0. 00 9 � 0. 01 8 � 0. 04 7 � 0. 09 4* � 0. 18 3* * � 0. 26 7* * � 0. 35 3* ** � 0. 44 4* ** � 0. 53 3* ** (� 1. 05 5) (� 1. 12 7) (� 1. 32 2) (� 1. 66 5) (� 2. 19 1) (� 2. 53 4) (� 2. 82 7) (� 3. 13 7) (� 3. 47 0) o bs er va tio ns 2, 72 8 2, 72 7 2, 72 4 2, 71 9 2, 70 9 2, 69 9 2, 68 9 2, 67 9 2, 66 9 a dj us te d r 2 0. 00 1 0. 00 2 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 pa ne l b — m id ca p v ix an d v v ix d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix 0. 04 6 � 0. 36 9 � 0. 24 0 � 0. 10 7 0. 03 2 0. 04 6 0. 45 8 0. 68 1 0. 42 6 (0 .2 45 ) (� 1. 79 5) (� 0. 84 7) (� 0. 30 3) (0 .0 66 ) (0 .0 73 ) (0 .6 77 ) (0 .8 95 ) (0 .5 15 ) � v v ix � 0. 58 5* * 0. 03 9 � 0. 08 9 � 0. 23 8 � 0. 68 7 � 1. 08 9 � 1. 12 7 � 0. 73 3 � 0. 52 7 (� 2. 60 2) (0 .1 56 ) (� 0. 26 0) (� 0. 53 2) (� 1. 09 9) (� 1. 32 1) (� 1. 27 5) (� 0. 75 6) (� 0. 46 9) c on st an t � 0. 00 7 � 0. 00 6 � 0. 01 9 � 0. 04 0 � 0. 07 6 � 0. 11 3 � 0. 15 3 � 0. 20 3 � 0. 25 5* (� 0. 56 1) (� 0. 52 9) (� 0. 70 6) (� 0. 89 7) (� 1. 12 0) (� 1. 32 0) (� 1. 50 5) (� 1. 77 0) (� 2. 02 6) o bs er va tio ns 2, 72 8 2, 72 7 2, 72 4 2, 71 9 2, 70 9 2, 69 9 2, 68 9 2, 67 9 2, 66 9 a dj us te d r 2 0. 00 1 0. 00 2 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 0. 00 0 pa ne l c — sm al l ca p v ix an d v v ix d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix 0. 04 5 0. 04 6 0. 12 6 0. 11 5 0. 10 6 0. 25 9 0. 43 4 0. 58 7 0. 23 8 (0 .2 68 ) (0 .2 45 ) (0 .4 33 ) (0 .3 15 ) (0 .2 11 ) (0 .4 15 ) (0 .6 12 ) (0 .7 67 ) (0 .2 98 ) � v v ix � 0. 29 8 � 0. 58 5* * � 0. 64 7 � 0. 91 8 � 1. 19 4 � 1. 65 5 � 1. 50 2 � 1. 62 5 � 0. 96 7 (� 1. 49 2) (� 2. 60 2) (� 1. 77 7) (� 1. 94 8) (� 1. 81 2) (� 1. 94 4) (� 1. 58 0) (� 1. 58 0) (� 0. 88 4) c on st an t � 0. 00 3 � 0. 00 7 � 0. 01 9 � 0. 04 1 � 0. 08 3 � 0. 12 7 � 0. 17 1 � 0. 22 3* � 0. 28 1* (� 0. 43 0) (� 0. 56 1) (� 0. 75 1) (� 0. 99 0) (� 1. 30 3) (� 1. 54 4) (� 1. 73 0) (� 1. 98 6) (� 2. 28 2) o bs er va tio ns 2, 72 8 2, 72 7 2, 72 4 2, 71 9 2, 70 9 2, 69 9 2, 68 9 2, 67 9 2, 66 9 a dj us te d r 2 0. 00 1 0. 00 3 0. 00 1 0. 00 1 0. 00 1 0. 00 1 0. 00 0 0. 00 0 0. 00 0 t hi s ta bl e pr es en ts th e re su lts of es tim at io ns of e q. (2 ). t he de pe nd en t va ri ab le is da ily re tu rn s to a po rt fo lio th at co nt ai ns a lo ng po si tio n in th e va lu e e t f an d an eq ua l sh or t po si tio n in th e gr ow th e t f fo r ea ch fo r ea ch si ze ca te go ry .f or ea se of in te rp re ta tio n, al lo f th e co ef fic ie nt s ha ve be en m ul tip lie d by 10 0. t hu s, fo r ex am pl e, if th e co ef fic ie nt is 1. 00 ,a on e pe rc en ta ge in cr ea se in th e v ix or v v ix in de xe s le ad s to a on e ba si s po in t re tu rn di ff er en tia l ov er th e re le va nt tim e fr am e. t he st an da rd er ro rs ar e es tim at ed us in g fiv e la gs to ac co un t fo r on e w ee k of tr ad in g ac tiv ity au to co rr el at io n an d he te ro sk ed as tic ity (n ew ey an d w es t, 19 87 ). *s ig ni fic an t at th e 5% le ve l. ** si gn ifi ca nt at th e 1% le ve l. 198 t.a. krause / financial services review 27 (2018) 189-207 in tables 4 through 9 are multiplied by 100 for ease of interpretation. specifically, for example, a one percentage daily increase in the vix index leads to a marginal 0.725 basis point increase in returns to the large-cap value etf over the following five-day period as opposed to its large-cap growth counterpart. on an annualized basis, that translates to a 3.720% excess return to the value etf over the growth etf. thus, over five-day time periods at least, investors are rewarded for investing in value stocks when expected volatility increases. this can be explained by the fact that increases in the vix index are normally accompanied by contemporaneous negative stock returns as investors demand an additional risk premium that is “repaid” over future periods more quickly for value stocks, as in merton (1980) and french, schwert, and stambaugh (1987). however, the short-term returns to value are negative and significant for several periods for the proxy for uncertainty, which is consistent with the recent empirical results of brenner and izhazian (2012) for the s&p 500 etf, and baltussen, van bekkum, and van der grient (2017) for single stocks. incrementally, for twoto forty-day periods, large-cap value-based etfs marginally underperform growth-based etfs when the vvix index increases. for example, a one percentage increase in the vvix (“uncertainty”) index leads to a 2.218 basis point return decline over a 30-day window for large-cap value versus their growth counterparts. this represents an annualized 1.879% excess return. for the five-day return window, the equivalent annual excess return is 4.807%. thus, it seems that as investors observe increased uncertainty in the marketplace, they sell growth stocks in disproportionate amounts relative to their value counterparts, and subsequently growth stocks experience higher future returns. this result is consistent with the “uncertainty-avoiding” hypothesis that is analyzed by aboura and erisoy (2017), baltussen, van bekkum, and van der grient (2017), and brenner and izhazian (2012). panels b and c of table 5 provide similar but significantly weaker results for the midcap and small-cap etfs. therefore, in the interest of brevity, the focus of the remaining discussion will center solely on the large-cap etf results, although results for the other etfs are available upon request. 3.2. additional explanatory variables the results of the previous section provide strong evidence that returns to value are driven by investor expectations regarding future risk and potential uncertainty. these results do not, however, explain the fact that overall market implied volatility and uncertainty are not constant over time. this section of the analysis examines some of the macroeconomic factors that may enhance the prior results and analyze the variations observed in the data to better explain the true drivers of the value versus growth return anomaly. to analyze the time-varying aspects of risk and uncertainty, one-day returns of the large-cap value minus growth etfs are regressed against prior day changes in the vix and vvix indexes over rolling one-year periods (252 trading days). the results of these estimations provide daily point estimates for the effect of each of the slope coefficients on future one-day returns, and they are summarized in fig. 2. notably, as seen in this graph, in recent years the coefficient values reflect the 0.71 positive correlation between the vix and vvix indexes (see table 3), although there are significant deviations, especially during the global financial crisis of 2007–2009 and the european debt crisis from 2010–2012. the 199t.a. krause / financial services review 27 (2018) 189-207 shaded areas of the graph represent time periods when the ads index of business conditions (aruoba, diebold, and scotti, 2009) is positive or negative. further implications of the differing time periods are explored in the analysis that follows. the implications of this graph are interesting because the coefficients for the proxy for uncertainty (vvix) reach their lowest levels during crisis periods, while the coefficients on the risk index (vix) reach their highest levels. thus, it seems that especially during these periods, over a one-day return horizon, investors buy value stocks that are depressed during periods of increased volatility that subsequently outperform. potential explanatory variables that may identify how investors view risk and uncertainty are provided by bali (2008) and bali and engle (2010). their analyses demonstrate that the moody’s baa-aaa corporate bond default spread and the ted spread (eurodollar over treasuries) are priced in the time-series and cross-section of equity portfolios. in their study of economic conditions and the effect on future stock returns, aruoba, diebold, and scotti (2009) postulate that their business conditions index (the aruoba-diebold-scotti, or ads index) determines “good” and “bad” states of the economy. this index is reported by the philadelphia federal reserve bank and “is designed to track real business conditions at high frequency.” its underlying (seasonally adjusted) economic indicators (weekly initial jobless claims; monthly payroll employment, industrial production, personal income less transfer payments, manufacturing and trade sales; and quarterly real gdp) blend highand lowfrequency information and stock and flow data.”1 summary statistics for these additional variables are not presented in the interest of brevity, but a correlation matrix is provided in table 6. it is clear that these variables are correlated to a certain degree with each other, but fig. 2. this figure presents rolling regression coefficients (252-day) for the one-day ahead return estimations of large-cap value minus large-cap growth exchange traded funds (etfs) using changes in vix and vvix as independent variables. the time-varying effects of risk and uncertainty, one-day returns of the large-cap value minus growth etfs are regressed against prior day changes in the vix and vvix indexes over rolling one-year periods (252 trading days). notably, as seen in this graph, in recent years the coefficient values reflect the 0.71 positive correlation between the vix and vvix indexes (see table 3), although there are significant deviations, especially during the global financial crisis of 2007–2009 and the european debt crisis from 2010–2012. shaded areas identify whether the ads business conditions index is positive or negative for that particular time frame. 200 t.a. krause / financial services review 27 (2018) 189-207 their correlation to the vix and vvix indexes are minimal; thus, they may shed further light on the relationships among risk, uncertainty, and returns to value if they are significant when included in regressions of returns to value. these variables are added to the right-hand side of eq. (2), and the results of these estimations are provided in table 7 for the large-cap etfs. the results are quite similar to the initial results of the previous section in panel a of table 5, in that the coefficient signs for the risk and uncertainty indexes are similar in direction and magnitude. also, the additional explanatory variables contribute to the model fit since all of the adjusted r2 values are higher, especially over longer time frames. the coefficients for the vix and vvix indexes are similar to the original specification. the coefficients for the ads index are not statistically significant (although they are uniformly positive and some are significant at the 10% level), but the ted spread variable is negatively related to future returns to value for the longer time frames. each of these results is consistent with future returns being positively related to business conditions (positive ads index and lower ted spread). the baa-aaa credit spread is not significant in these regressions over any time frame. table 8 provides a similar, but more informative, analysis to table 7 in that it includes the additional explanatory variables, but the two panels distinguish between “bad” and “good” states of the economy as specified by the ads index (aruoba, diebold, and scotti, 2009). in panel a for “bad” states of the economy (ads � 0), the previously reported and expected coefficients for the vix and vvix variables are observed, although they are uniformly higher (in absolute value) and stronger in significance than those observed in table 7. the ted spread coefficients remain negative and are stronger than those in table 7 as well, while the corporate bond spread remains insignificant. thus, the results of this subsample support the initial inferences made regarding the variables’ relationships in table 7. however, in panel b that contains the results for “good” states of the economy (ads � 0), the coefficients for the vix and vvix indexes are uniformly insignificant. this behavior from investors may be viewed through the lens of the “mental accounting” framework of kahneman and tversky (1984), who posit that individuals feel better about avoiding losses than they feel about making gains (“loss aversion”). in this case, where the state of the economy is “good,” investors may purchase growth stocks (or be indifferent and not sell them) even when volatility rises, giving rise to the insignificant results. in this specification, the ted spread coefficients become insignificant, while the corporate bond spread is positively related to future returns over the 20to 60-day horizons. thus, as corporate bond quality declines in a “good” state of the economy, investors prefer to invest in value rather table 6 correlation matrix of additional explanatory variables variable symbol vix vvix ads ted baa-aaa � volatility index vix 1.0000 � vol of vol index vvix 0.7142 1.0000 ads index ijs �0.0035 0.0053 1.0000 ted spread ijt 0.0157 0.0027 �0.6374 1.0000 baa-aaa spread ijj �0.0160 �0.0155 �0.8221 0.5249 1.0000 ads � aruoba-diebold-scotti. this table provides a correlation matrix of the additionally explanatory variables under study in this article. 201t.a. krause / financial services review 27 (2018) 189-207 t ab le 7 l ar ge -c ap es tim at io ns w ith ad di tio na l ex pl an at or y va ri ab le s d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 04 0 0. 15 8 0. 70 1* 0. 88 7 1. 25 6 1. 12 9 1. 37 5 0. 70 9 0. 42 1 (� 0. 21 3) (0 .7 02 ) (2 .1 64 ) (1 .7 72 ) (1 .8 39 ) (1 .3 90 ) (1 .5 20 ) (0 .7 38 ) (0 .4 14 ) � v v ix � 0. 04 6 � 0. 41 6 � 0. 89 6* � 1. 39 7* � 1. 71 2 � 2. 28 7* � 2. 17 4 � 0. 97 7 � 0. 10 0 (� 0. 22 1) (� 1. 54 2) (� 2. 11 5) (� 2. 08 9) (� 1. 94 8) (� 2. 06 0) (� 1. 77 8) (� 0. 76 3) (� 0. 07 2) a d s in de x 0. 04 5 0. 08 5 0. 19 1 0. 33 9 0. 51 0 0. 60 2 0. 61 7 0. 50 3 0. 27 7 (1 .4 48 ) (1 .4 39 ) (1 .5 65 ) (1 .7 93 ) (1 .8 11 ) (1 .8 62 ) (1 .8 80 ) (1 .5 32 ) (0 .8 45 ) t e d sp re ad 0. 01 9 0. 02 5 0. 03 8 0. 00 1 � 0. 23 9 � 0. 63 0* � 0. 95 7* * � 1. 47 3* * � 2. 14 4* * (0 .5 62 ) (0 .4 15 ) (0 .3 42 ) (0 .0 05 ) (� 0. 91 9) (� 2. 05 0) (� 2. 71 0) (� 3. 76 3) (� 4. 98 0) b aa -a aa c or p. sp re ad 0. 03 9 0. 07 8 0. 17 2 0. 30 6 0. 46 6 0. 66 8 0. 74 6 0. 79 9 0. 83 2 (0 .8 87 ) (0 .9 50 ) (1 .0 01 ) (1 .1 69 ) (1 .3 13 ) (1 .6 57 ) (1 .6 90 ) (1 .6 05 ) (1 .5 54 ) c on st an t � 0. 04 5 � 0. 08 7 � 0. 19 0 � 0. 31 4 � 0. 40 2 � 0. 48 7 � 0. 49 7 � 0. 43 5 � 0. 31 7 (� 1. 04 3) (� 1. 06 9) (� 1. 12 4) (� 1. 20 6) (� 1. 13 3) (� 1. 23 1) (� 1. 11 4) (� 0. 84 2) (� 0. 54 9) o bs er va tio ns 2, 69 9 2, 69 8 2, 69 5 2, 69 0 2, 68 0 2, 67 0 2, 66 0 2, 65 0 2, 64 0 a dj us te d r 2 0. 00 0 0. 00 2 0. 00 8 0. 01 6 0. 03 3 0. 05 2 0. 06 1 0. 07 6 0. 10 4 a d s � a ru ob ad ie bo ld -s co tti .t hi s ta bl e pr es en ts th e re su lts of es tim at io ns of e q. (2 ) w ith ad di tio na le xp la na to ry fa ct or s us ed as in de pe nd en tv ar ia bl es . t he de pe nd en t va ri ab le is da ily re tu rn s to a po rt fo lio th at co nt ai ns a lo ng po si tio n in th e va lu e e xc ha ng e t ra de d fu nd (e t f) an d an eq ua l sh or t po si tio n in th e gr ow th e t f fo r ea ch fo r ea ch si ze ca te go ry . fo r ea se of in te rp re ta tio n, al l of th e co ef fic ie nt s ha ve be en m ul tip lie d by 10 0. t hu s, fo r ex am pl e, if th e co ef fic ie nt is 1. 00 , a on e pe rc en ta ge in cr ea se in th e v ix or v v ix in de xe s le ad s to a on e ba si s po in t re tu rn di ff er en tia l ov er th e re le va nt tim e fr am e. t he st an da rd er ro rs ar e es tim at ed us in g fiv e la gs to ac co un t fo r on e w ee k of tr ad in g ac tiv ity au to co rr el at io n an d he te ro sk ed as tic ity (n ew ey an d w es t, 19 87 ). *s ig ni fic an t at th e 5% le ve l. ** si gn ifi ca nt at th e 1% le ve l. 202 t.a. krause / financial services review 27 (2018) 189-207 t ab le 8 l ar ge -c ap es tim at io ns w ith st at es of a ru ob ad ie bo ld -s co tti (a d s) pa ne l a : b ad st at e (a d s � 0) d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix 0. 03 3 0. 23 5 1. 06 3* * 1. 41 1* 2. 10 8* 2. 09 2* 2. 52 8* 1. 66 9 1. 24 2 (0 .1 51 ) (0 .8 36 ) (2 .6 83 ) (2 .2 62 ) (2 .4 76 ) (2 .1 06 ) (2 .2 87 ) (1 .4 22 ) (1 .0 18 ) � v v ix � 0. 02 6 � 0. 62 3 � 1. 23 6* � 1. 90 2* � 2. 42 5* � 3. 02 9* � 2. 96 0* � 1. 43 7 � 0. 53 6 (� 0. 10 8) (� 1. 92 3) (� 2. 37 1) (� 2. 27 3) (� 2. 19 5) (� 2. 18 2) (� 2. 00 2) (� 0. 95 3) (� 0. 34 0) t e d sp re ad 0. 00 1 � 0. 00 8 � 0. 04 0 � 0. 13 2 � 0. 44 4* � 0. 88 2* * � 1. 22 0* * � 1. 68 4* * � 2. 26 1* * (0 .0 17 ) (� 0. 15 0) (� 0. 39 3) (� 0. 95 4) (� 2. 16 7) (� 3. 52 1) (� 3. 93 3) (� 4. 86 6) (� 6. 02 9) b aa -a aa c or p. sp re ad � 0. 01 3 � 0. 02 0 � 0. 04 8 � 0. 09 4 � 0. 16 9 � 0. 11 3 � 0. 10 2 � 0. 00 3 0. 18 2 (� 0. 35 8) (� 0. 28 5) (� 0. 31 4) (� 0. 39 3) (� 0. 46 7) (� 0. 23 9) (� 0. 19 1) (� 0. 00 5) (0 .3 41 ) c on st an t 0. 00 4 0. 00 4 0. 01 8 0. 06 7 0. 23 7 0. 34 0 0. 42 5 0. 49 2 0. 55 1 (0 .0 96 ) (0 .0 53 ) (0 .1 17 ) (0 .2 72 ) (0 .6 48 ) (0 .7 08 ) (0 .7 72 ) (0 .8 26 ) (0 .9 29 ) o bs er va tio ns 2, 01 0 2, 01 0 2, 01 0 2, 01 0 2, 00 2 1, 99 2 1, 98 2 1, 97 2 1, 96 2 a dj us te d r 2 0. 00 0 0. 00 0 0. 00 3 0. 00 7 0. 02 5 0. 04 9 0. 06 4 0. 08 8 0. 12 8 pa ne l b : g oo d st at e (a d s � 0) d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 22 1 � 0. 29 5 � 0. 68 5 � 1. 14 3 � 1. 88 5 � 2. 33 2 � 2. 45 3 � 2. 04 6 � 1. 75 7 (� 0. 69 0) (� 0. 89 4) (� 1. 34 2) (� 1. 62 1) (� 1. 80 7) (� 1. 93 8) (� 1. 79 8) (� 1. 48 5) (� 1. 13 9) � v v ix � 0. 09 3 0. 49 8 0. 75 3 1. 03 2 1. 80 7 1. 44 1 1. 65 0 1. 21 6 1. 85 2 (� 0. 27 7) (1 .0 42 ) (0 .9 86 ) (0 .9 38 ) (1 .1 84 ) (0 .8 12 ) (0 .7 55 ) (0 .5 55 ) (0 .7 20 ) t e d sp re ad � 0. 02 0 � 0. 02 2 � 0. 11 4 � 0. 54 7 � 1. 25 3 � 1. 75 0 � 1. 67 6 � 2. 30 9 � 3. 57 8 (� 0. 23 7) (� 0. 14 2) (� 0. 31 4) (� 0. 85 0) (� 1. 21 8) (� 1. 34 3) (� 1. 10 7) (� 1. 39 6) (� 1. 87 1) b aa -a aa c or p. sp re ad 0. 08 8 0. 17 3 0. 36 7 0. 86 0* 1. 61 3* * 2. 16 8* * 2. 77 4* * 3. 77 7* * 4. 80 3* * (1 .4 56 ) (1 .5 62 ) (1 .5 40 ) (2 .2 62 ) (2 .8 48 ) (3 .0 39 ) (3 .3 77 ) (4 .0 89 ) (4 .5 01 ) c on st an t � 0. 07 2 � 0. 14 6 � 0. 30 3 � 0. 63 2 � 1. 17 1* � 1. 62 6* * � 2. 27 0* * � 3. 12 4* * � 3. 91 4* * (� 1. 41 6) (� 1. 56 9) (� 1. 51 2) (� 1. 94 6) (� 2. 29 3) (� 2. 59 1) (� 3. 32 1) (� 4. 24 7) (� 4. 72 3) o bs er va tio ns 68 9 68 8 68 5 68 0 67 8 67 8 67 8 67 8 67 8 a dj us te d r 2 0. 00 4 0. 00 4 0. 01 2 0. 03 3 0. 04 7 0. 05 4 0. 07 0 0. 10 5 0. 13 2 t hi s ta bl e pr es en ts th e re su lts of es tim at io ns of e q. (2 ) w ith ad di tio na l ex pl an at or y fa ct or s us ed as in de pe nd en t va ri ab le s ov er di ff er in g st at es of th e a d s bu si ne ss co nd iti on s in de x. t he de pe nd en tv ar ia bl e is da ily re tu rn s to a po rt fo lio th at co nt ai ns a lo ng po si tio n in th e va lu e e xc ha ng e t ra de d fu nd (e t f) an d an eq ua ls ho rt po si tio n in th e gr ow th e t f fo r ea ch fo r ea ch si ze ca te go ry .f or ea se of in te rp re ta tio n, al lo f th e co ef fic ie nt s ha ve be en m ul tip lie d by 10 0. t hu s, fo r ex am pl e, if th e co ef fic ie nt is 1. 00 ,a on e pe rc en ta ge in cr ea se in th e v ix or v v ix in de xe s le ad s to a on e ba si s po in t re tu rn di ff er en tia l ov er th e re le va nt tim e fr am e. t he st an da rd er ro rs ar e es tim at ed us in g fiv e la gs to ac co un t fo r on e w ee k of tr ad in g ac tiv ity au to co rr el at io n an d he te ro sk ed as tic ity (n ew ey an d w es t, 19 87 ). *s ig ni fic an t at th e 5% le ve l. ** si gn ifi ca nt at th e 1% le ve l. 203t.a. krause / financial services review 27 (2018) 189-207 than growth stocks. but overall, the results indicate that value style investors are not penalized to exposures to risk or uncertainty “good” states of the economy, although it should be noted that the number of observations for “good” states of the economy are only about one quarter of the total observations in the sample. one final analysis examines potentially “asymmetric” responses to changes in the vix index. the sample is divided into days when the change in the vix index is negative or positive, respectively. the results of this analysis are presented in table 9, and the results for prior day decreases in the vix index are presented in panel a. these results are largely consistent with the previous results in table 7 and panel a of table 8. decreases in the vix index lead to future positive returns to value while changes in the vvix index are negatively related to future returns. increases in the ted spread are related to future negative returns. however, the returns to value in panel b for the volatility indices are almost wholly insignificant. only two of the negative coefficients for the ted spread are statistically significant, and any consideration for changes in the vix and vvix indexes are fully ignored. investors do not seem to be concerned with changes in these variables on days that the vix index rises, or perhaps they simply choose not to trade in this environment, thus the new information regarding risk and uncertainty is not impounded into etf prices. however, when the vix declines, investors are rewarded for exposures to value stocks, although the effects are mitigated given a concurrent increase in the vvix index. panel b of table 9 does not provide further insight into these issues. to summarize, the most consistent results of the paper are presented in table 7 and each panel a of tables 8 and 9. in all of these estimations, one-day changes in the vix index (the proxy for risk) are positively related to returns to value, while the opposite is true for the vvix index (the proxy for uncertainty). adding macroeconomic variables to the analysis increases the models’ explanatory power, and the results are intuitive. the results are strongest during periods when the ads index is in a “good” state and when the change in the vix index is negative. 4. conclusion the results of this article examine time-varying returns to risk and uncertainty in various market states dependent on macroeconomic variables, using extremely liquid etfs. in contrast to the findings of boscaljon, filbeck, and zhao (2011), where these differences have disappeared over all but the longest time frames, the inclusion of a proxy for uncertainty and macroeconomic variables provides economically significant results. when these results are evaluated in conjunction with the vvix index, there are still returns to value in the evaluation of the vix index under certain conditions. these returns are also incrementally larger when considering information from the vvix index that provides a proxy for uncertainty. the inclusion of the vvix index and proxies for economic conditions in the analysis of the vix effect on returns to value for inexpensive and easily-traded etfs is informative, since it is straightforward and cost-effective to implement style-based trading strategies with these highly liquid securities. 204 t.a. krause / financial services review 27 (2018) 189-207 t ab le 9 l ar ge -c ap es tim at io ns w ith v ix po si tiv e/ ne ga tiv e pa ne l a : n eg at iv e ch an ge s in v ix in de x d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix 0. 27 8 0. 62 2 2. 44 2* * 2. 88 5* 4. 90 8* * 4. 27 7* 2. 99 9 1. 17 8 0. 70 7 (0 .7 33 ) (1 .1 93 ) (3 .0 12 ) (2 .2 99 ) (3 .2 30 ) (2 .3 44 ) (1 .6 00 ) (0 .5 39 ) (0 .3 02 ) � v v ix � 0. 33 9 � 0. 81 6* � 1. 42 1* � 2. 48 1* � 2. 11 3 � 2. 12 2 � 1. 96 4 0. 23 5 1. 46 0 (� 1. 25 0) (� 2. 17 3) (� 2. 12 7) (� 1. 97 1) (� 1. 28 9) (� 0. 99 0) (� 0. 93 3) (0 .1 07 ) (0 .6 33 ) a d s in de x 0. 09 7* 0. 13 1 0. 22 5 0. 40 6 0. 57 6 0. 73 5* 0. 70 3 0. 60 0 0. 42 0 (2 .4 04 ) (1 .8 75 ) (1 .5 94 ) (1 .6 19 ) (1 .5 98 ) (1 .7 40 ) (1 .5 92 ) (1 .3 67 ) (0 .9 67 ) t e d sp re ad 0. 01 8 � 0. 00 8 � 0. 04 1 � 0. 09 6 � 0. 34 5 � 0. 56 2 � 1. 12 9* � 1. 62 9* * � 2. 29 6* * (0 .4 39 ) (� 0. 13 9) (� 0. 36 9) (� 0. 49 3) (� 1. 16 9) (� 1. 42 5) (� 2. 30 3) (� 2. 89 8) (� 3. 79 8) b aa -a aa c or p. sp re ad 0. 11 6* 0. 14 1 0. 24 1 0. 44 1 0. 57 3 0. 70 9 0. 75 0 0. 81 3 0. 97 5 (2 .2 41 ) (1 .5 35 ) (1 .3 02 ) (1 .4 00 ) (1 .2 82 ) (1 .3 37 ) (1 .2 97 ) (1 .2 67 ) (1 .4 29 ) c on st an t � 0. 10 6* � 0. 11 4 � 0. 14 5 � 0. 32 6 � 0. 27 8 � 0. 35 2 � 0. 29 8 � 0. 28 0 � 0. 30 1 (� 2. 02 6) (� 1. 23 7) (� 0. 78 6) (� 1. 06 3) ( � 0. 64 2) (� 0. 68 4) (� 0. 51 0) (� 0. 41 3) (� 0. 40 2) o bs er va tio ns 1, 44 3 1, 44 2 1, 43 9 1, 43 9 1, 43 4 1, 42 8 1, 42 3 1, 41 7 1, 41 1 a dj us te d r 2 0. 00 6 0. 01 0 0. 02 0 0. 03 1 0. 05 4 0. 06 0 0. 08 3 0. 08 9 0. 11 3 pa ne l b : po si tiv e ch an ge s in v ix in de x d ay (s ) ah ea d re tu rn 1 2 5 10 20 30 40 50 60 � v ix � 0. 31 6 � 0. 35 0 � 0. 40 5 � 0. 59 7 � 0. 79 4 � 0. 20 1 � 0. 18 1 0. 27 9 0. 18 4 (� 0. 97 4) (� 1. 03 8) (� 0. 86 0) (� 0. 93 6) (� 0. 76 7) (� 0. 15 6) (� 0. 13 2) (0 .1 83 ) (0 .1 11 ) � v v ix 0. 24 1 0. 13 0 � 0. 04 9 � 0. 03 0 � 0. 67 0 � 1. 98 8 � 1. 18 5 � 1. 11 7 � 0. 78 5 (0 .8 10 ) (0 .3 66 ) (� 0. 08 9) (� 0. 03 8) (� 0. 58 6) (� 1. 55 3) (� 0. 77 1) (� 0. 66 3) (� 0. 39 6) a d s in de x � 0. 00 9 0. 03 6 0. 16 6 0. 27 3 0. 46 5 0. 47 9 0. 52 5 0. 39 3 0. 11 5 (� 0. 20 6) (0 .5 02 ) (1 .2 50 ) (1 .2 68 ) (1 .1 69 ) (1 .0 52 ) (1 .1 73 ) (0 .9 07 ) (0 .2 61 ) t e d sp re ad 0. 02 7 0. 06 4 0. 14 5 0. 13 2 � 0. 08 1 � 0. 67 1 � 0. 80 2 � 1. 35 9* * � 2. 03 2* * (0 .5 85 ) (0 .7 91 ) (1 .0 96 ) (0 .6 73 ) (� 0. 22 0) (� 1. 60 1) (� 1. 75 4) (� 2. 74 0) (� 3. 75 2) b aa -a aa c or p. sp re ad � 0. 04 1 0. 01 9 0. 11 2 0. 16 7 0. 38 2 0. 68 1 0. 81 3 0. 83 5 0. 69 9 (� 0. 62 7) (0 .1 69 ) (0 .5 89 ) (0 .5 25 ) (0 .7 73 ) (1 .2 51 ) (1 .2 80 ) (1 .1 52 ) (0 .8 55 ) c on st an t 0. 03 1 � 0. 04 2 � 0. 14 7 � 0. 19 6 � 0. 31 2 � 0. 46 1 � 0. 62 0 � 0. 55 2 � 0. 24 7 (0 .4 86 ) (� 0. 37 8) (� 0. 77 2) (� 0. 59 9) (� 0. 61 8) (� 0. 88 2) (� 0. 97 8) (� 0. 74 2) (� 0. 28 2) o bs er va tio ns 1, 24 4 1, 24 4 1, 24 4 1, 23 9 1, 23 4 1, 23 0 1, 22 5 1, 22 1 1, 21 7 a dj us te d r 2 0. 00 0 0. 00 0 0. 00 3 0. 00 6 0. 01 9 0. 04 5 0. 03 9 0. 06 2 0. 09 4 a d s � a ru ob ad ie bo ld -s co tti .t hi s ta bl e pr es en ts th e re su lts of es tim at io ns of e q. (2 ) w ith ad di tio na l ex pl an at or y fa ct or s us ed as in de pe nd en t va ri ab le s ov er di ff er in g ch an ge s in th e v ix in de x. t he de pe nd en t va ri ab le is da ily re tu rn s to a po rt fo lio th at co nt ai ns a lo ng po si tio n in th e va lu e e xc ha ng e t ra de d fu nd (e t f) an d an eq ua l sh or t po si tio n in th e gr ow th e t f fo r ea ch fo r ea ch si ze ca te go ry . fo r ea se of in te rp re ta tio n, al l of th e co ef fic ie nt s ha ve be en m ul tip lie d by 10 0. t hu s, fo r ex am pl e, if th e co ef fic ie nt is 1. 00 , a on e pe rc en ta ge in cr ea se in th e v ix or v v ix in de xe s le ad s to a on e ba si s po in t re tu rn di ff er en tia l ov er th e re le va nt tim e fr am e. t he st an da rd er ro rs ar e es tim at ed us in g fiv e la gs to ac co un t fo r on e w ee k of tr ad in g ac tiv ity au to co rr el at io n an d he te ro sk ed as tic ity (n ew ey an d w es t, 19 87 ). *s ig ni fic an t at th e 5% le ve l. ** si gn ifi ca nt at th e 1% le ve l. 205t.a. krause / financial services review 27 (2018) 189-207 note 1 source: https://www.philadelphiafed.org/research-and-data/real-time-center/businessconditions-index. references aboura, s., & arisoy, y. e. (2017). does aggregate uncertainty explain size and value anomalies? applied economics, 49, 3214–3230. ammann, m., skovmand, d., & verhofen, m. (2009). implied and realized volatility in the cross-section of equity options. international journal of theoretical and applied finance, 12, 745–765. an, b. j., ang, a., bali, t. g., & cakici, n. (2014). the joint cross section of stocks and options. the journal of finance, 69, 2279–2337. andersen, t. g., bollerslev, t., diebold, f. x., & labys, p. (2003). modeling and forecasting realized volatility. econometrica, 71(2), 579–625. aruoba, s. b., diebold, f. x., & scotti, c. (2009). real-time measurement of business conditions. journal of business & economic statistics, 27, 417–427. bali, t. g. (2008). the intertemporal relation between expected returns and risk. journal of financial economics, 87, 101–131. bali, t. g., brown, s. j., & caglayan, m. o. (2014). macroeconomic risk and hedge fund returns. journal of financial economics, 114, 1–19. bali, t. g., cakici, n., & chabi-yo, f. (2015). a new approach to measuring riskiness in the equity market: implications for the risk premium. journal of banking & finance, 57, 101–117. bali, t. g., & engle, r. f. (2010). the intertemporal capital asset pricing model with dynamic conditional correlations. journal of monetary economics, 57, 377–390. bali, t. g., & peng, l. (2006). is there a risk–return trade-off? evidence from high-frequency data. journal of applied econometrics, 21, 1169–1198. baltussen, g., van bekkum, s., & van der grient, b. (2018). unknown unknowns: vol-of-vol and the cross section of stock returns. journal of financial and quantitative analysis (forthcoming). bekaert, g., engstrom, e., & xing, y. (2009). risk, uncertainty, and asset prices. journal of financial economics, 91, 59–82. below, s. p., kiely, j. p., & prati, r. p. (2009). style index rebalancing for better diversification: lessons from broad market and equity style indexes. financial services review, 18, 231–248. blair, b. j., poon, s.-h., & taylor, s. j. (2001). modelling s&p 100 volatility: the information content of stock returns. journal of banking & finance, 25, 1665–1679. boscaljon, b., filbeck, g., & zhao, x. (2011). market timing using the vix for style rotation. financial services review, 20, 1. brenner, m., & izhakian, y. y. (2017). asset pricing and ambiguity: empirical evidence. journal of financial economics (forthcoming). brogaard, j., & detzel, a. (2015). the asset-pricing implications of government economic policy uncertainty. management science, 61, 3–18. brous, p., ince, u., & popova, i. (2010). volatility forecasting and liquidity: evidence from individual stocks. journal of derivatives & hedge funds, 16, 144–159. canina, l., & figlewski, s. (1993). the informational content of implied volatility. the review of financial studies, 6, 659–681. cboe, 2012. double the fun with cboe’s vvix index, white paper, available at http://www.cboe.com/micro/ vvix/documents/vvix-termstructure.pdf. chan, w. h., jha, r., & kalimipalli, m. (2009). the economic value of using realized volatility in forecasting future implied volatility. journal of financial research, 32, 231–259. 206 t.a. krause / financial services review 27 (2018) 189-207 copeland, m. m., & copeland, t. e. (1999). market timing: style and size rotation using the vix. financial analysts journal, 55, 73–81. demiguel, v., plyakha, y., uppal, r., & vilkov, g. (2013). improving portfolio selection using option-implied volatility and skewness. journal of financial and quantitative analysis, 48, 1813–1845. ellsberg, d. (1961). risk, ambiguity, and the savage axioms. the quarterly journal of economics, 75, 643–669. french, k. r., schwert, g. w., & stambaugh, r. f. (1987). expected stock returns and volatility. journal of financial economics, 19, 3–29. giot, p. (2005). relationships between implied volatility indices and stock index returns. journal of portfolio management, 31, 92–100. goldwhite, p. (2009). diversification and risk management: what volatility tells us. journal of investing, 18, 40. guo, h., & whitelaw, r. f. (2006). uncovering the risk–return relation in the stock market. the journal of finance, 61, 1433–1463. jiang, g. j., & tian, y. s. (2005). the model-free implied volatility and its information content. the review of financial studies, 18, 1305–1342. kahneman, d., & tversky, a. (1984). choices, values, and frames. cambridge: cambridge university press. keynes, j. m. (1921). a treatise on probability. vol. 8 of collected writings (1973 ed.). london: macmillan. knight, f. (1921). risk, uncertainty, and profit. boston, ma: houghton mifflin. ko, j.-h., & lee, c.-m. (2015). international economic policy uncertainty and stock prices: wavelet approach. economics letters, 134, 118–122. krause, t. a. (2018). hedge fund returns and uncertainty. the north american journal of economics and finance. markowitz, h. (1952). portfolio selection. the journal of finance, 7, 77–91. mayfield, c., perdue, g., & wooten, k. (2008). investment management and personality type. financial services review, 17, 219–236. merton, r. c. (1980). on estimating the expected return on the market: an exploratory investigation. journal of financial economics, 8, 323–361. newey, w. k., & west, k. d. (1987). a simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. econometrica, 55, 703–708. puttonen, v., & seppä, t. (2007). do style benchmarks differ? journal of asset management, 7, 425–428. sarwar, g. (2005). the informational role of option trading volume in equity index options markets. review of quantitative finance and accounting, 24, 159. shleifer, a., & vishny robert, w. (2012). the limits of arbitrage. the journal of finance, 52, 35–55. su, z., fang, t., & yin, l.(2017). the role of news-based implied volatility among us financial markets. economics letters, 157, 24–27. talukdar, b., daigler, r. t., & parhizgari, a. m. (2017). expanding the explanations for the return–volatility relation. journal of futures markets, 37, 689–716. 207t.a. krause / financial services review 27 (2018) 189-207 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 active asset allocation for retirement funds using the fed model john m. clinebella, douglas r. kahlb, jerry l. stevensc,* adepartment of finance, kenneth w. monfort school of business, university of northern colorado, greeley, co 80639, usa bdepartment of finance, university of akron, akron, oh 44325-4803, usa cdepartment of finance, distinguished teaching fellow, e. c. robins school of business, university of richmond, richmond, va 23173, usa abstract active asset allocation, also known as market timing, is controversial but potentially effective for individual investors and financial advisors. many studies support market timing based on the relationship between the aggregate earnings yield on equities and the intermediate treasury bond yield, known as the fed model. nevertheless, skeptics point to common flaws in these studies and challenge the validity of the fed model. in general, returns from timing models are difficult to adjust for transaction costs and tax effects from short term gains and losses. in almost all cases, there is data mining from reporting results over the same period used to build the timing model. our study addresses these concerns directly. we control the transaction costs and tax effects by focusing on funds available for retirement accounts within a vanguard fund family allowing costless monthly transfers. we use a “risk on” or “risk off” approach rather than experiment with arbitrary cutoff rules for switching funds. we first test for the time series properties of the fed model to build a prediction model and then apply it over a recent five-year holdout of period. our findings show that the switching portfolio offers attractive performance compared to either of the vanguard funds, especially with respect to enhancing upside to downside risk ratios. © 2017 academy of financial services. all rights reserved. keywords: market timing; kalman filter; active asset allocation * corresponding author. tel.: �1-804-289-8597; fax: �1-804-289-8878. e-mail address: jstevens@richmond.edu (j. l. stevens). financial services review 26 (2017) 167–180 1057-0810/17/$ – see front matter © 2017 academy of financial services. all rights reserved. 1. introduction individual investors and investment advisors make asset allocation decisions as part of an overall investment strategy. efficient market theory supports using a passive buy and hold approach linked to long run objectives and investment horizons with periodic rebalancing as objectives and horizons change. on the other hand, active asset allocation has the potential to increase portfolio returns dramatically if executed successfully. xiong, ibbotson, idzorek, and chen (2010) show that portfolio returns split evenly between asset allocation and active management when overall market moves are accounted for. shilling (1992) concludes that avoiding the 50 weakest months of the stock market would have doubled investor returns. there is a vast literature on active asset allocation, often called market timing, with few definitive conclusions. the volume of timing research is a testament to the importance of the topic to investors. in this article, we examine the investment performance of an active asset allocation model based on monthly signals from the time varying relationship between the aggregate equity earnings yield and the 10-year treasury yield, often called the fed model. we address key criticisms of prior work to build and then test timing performance of the fed model. first, we test for the appropriate time series of the fed model to define the switch points for asset allocation, rather than experiment to find points that worked in past data or simply assume the spread is a random variable. next, important deficiencies of prior studies come from a failure to adequately control for transaction costs and tax effects from gains and losses because of switches. we control for these effects by using vanguard’s s&p500 index fund (vfinx) and treasury index fund (vfitx). because transfers between these funds in the vanguard family do not have buying and selling costs for the investor, we avoid the criticism of not accounting for transaction costs from asset switches.1 changes in asset allocation result in short term realized gains and losses that lead to tax effects that many studies ignore. since the vanguard family of funds is widely used in tax deferred retirement accounts, the tax effects from realized gains and losses because of switches are irrelevant for tax-exempt retirement fund investors. the analysis is limited to retirement fund investors using a family of funds but that segment of the investing population is significant and growing. many studies report attractive performance after searching through past data with multiple models and arbitrary switch points. critics point out that a regularity in past data discovered in this way does not mean the model that works best in past periods is predictable going forward. to avoid this data mining criticism, we test for the time series properties of the fed model over long periods of time and then use out of sample data from january 2012 through december 2016 to test the investment performance of the model. we find the performance characteristics of the active asset allocation portfolio to be attractive to many investors seeking higher returns than a buy and hold equity fund with only moderately more risk than a buy and hold 10-year treasury fund. potentially more important, the investment performance results in a more favorable upside to downside volatility ratio favored by investors who fear losses more than they covet gains. we use the following organization for the article. section 2 provides a review of the literature on active asset allocation and the criticisms leveled at these studies. we follow in section 3 by reviewing the traditional use of the fed model to time the markets. we also 168 j.m. clinebell et al. / financial services review 26 (2017) 167–180 present the conceptual arguments for using the fed model and address the common problems with the way it has been tested and applied in other studies. we present the time series version of the fed model in section 4 along with the estimated parameters that define the switch points for the active asset allocation moves. portfolio performance from applying the fed model in an out-of-sample investment period of recent markets using the vanguard index funds appears in section 5. the final section contains the conclusions and suggestions for added research. 2. literature review while many studies find return enhancement from active asset allocation, other studies point out critical measurement and research design flaws affecting the findings. sharpe (1975) was one of the first to point out the difficulty of timing the market and reported that an investor would need to tell a good year from a bad year seven out of 10 times to be successful. more recently, bauer and dahlquist (2001) calculate that an investor would need to be able to switch correctly 66% of the time on a monthly basis to outperform a buy and hold strategy. early studies of timing ability outlined by pinches (1970) generated mixed results where superior performance from timing rules tended to be offset by trading costs. the literature on timing continues to grow with new and more complex trading rules based on fundamentals (feldman, jung, & klein, 2015), macroeconomic variables (breen, glosten, & jagamathan, 1989 and guido, peral, & walsh, 2011), nonfinancial indicators (krueger & kennedy, 1990), mean reversion (campbell, andrew, & mckinley, 1999), and technical indicators (lo, mamaysky, & wang, 2000). critics of these studies point to problems of transaction costs, tax effects, and data mining (aronson, 2006; asness, 2003; and sullivan, timmerman, & white, 1999). a good example of the dialogue on timing appears in the journal of portfolio management where pruitt and white (1988) provide a defense for technical analysis to beat the market while ball, kothari, and wasley (1995) present the problems of actually implementing technical trading rules in the real world. proponents of the efficient market hypothesis refute the view that the equity market could overand under-react in a predictable way to allow successful timing strategies (fama, 1998 and malkiel, 2003). the blending of psychology and finance by kahneman and tversky provides a paradigm to counterbalance the efficient market hypothesis and offers a conceptual foundation for potential market timing anomalies.2 behavioral biases and heuristics make predictably irrational outcomes in the market possible. campbell (2000), campbell, andrew, and mckinley (1999), daniel, hirshleifer, and subrahmanyam (1998), debont and thaler (1985), and shleifer (2000) are a few of the studies suggesting that overreaction to information makes the stock market deviate from fundamentals, allowing potential gains from market timing. the overreaction hypothesis, based on “follow the herd” and “regret aversion” concepts originally developed in psychology, is now part of the mainstream literature. recent overreaction in the stock market during the late 1990s and in the housing market in the 2005–2008 period reinforces the view that predictable irrationality in the market may be viable, encouraging more study of timing rules. 169j.m. clinebell et al. / financial services review 26 (2017) 167–180 3. the fed model in a speech on irrational exuberance in 1991, alan greenspan used the relationship between the aggregate equity earnings yield and the 10-year treasury yield to evaluate “abnormal” market conditions.3 when the stock market is abnormally high the earnings yield (aggregate earnings divided by the market index price) is low relative to the long term bond yield. this interpretation of the relationship between the equity yield and the treasury yield suggests a predictive model for stock market valuation commonly called the “fed” model, even though the fed never officially recognized that usage. the fed model specification appears as eq. (1) below. (e/p)t � �(y10)t (1) for each point in time (t), the measure of the equity market yield is the ratio of trailing aggregate equity market earnings to the aggregate equity market price index (e/p). the right side of eq. (1) is the 10-year treasury bond yield (y10) times a multiplier (�).4 in the strict version of the fed model, the equilibrium multiplier (�) is one, but this need not be the case. most uses of the model assume that the multiplier is a random variable, making the mean of �t the best estimate of the equilibrium relationship in eq. (1). deviations from the mean value of the multiplier represent abnormalities and mispricings. for example, bodie, kane, and marcus (noted as bkm, 2014) make the following statement in their popular investments textbook: the most popular approach to forecasting the overall stock market is the earnings multiplier approach applied at the aggregate level. (p. 429) as an illustration, if the long run average of the equity market p/e is 15 the long run average for the earnings yield (e/p) is 6.67%. if the long run average of the 10-year treasury yield is 5% the long run average of the fed model multiplier is 1.334 (0.0667 / 0.05 � 1.334).5 if the current 10-year treasury yield is 2.3%, the expected value of the e/p today is 3.068% (1.334 � 0.023). based on this logic the current market value of equities should be 32.59 times aggregate earnings (1/0.03068 � 32.59 � p/e). if the current market p/e is below 32.59, the fed model would predict undervaluation and a buy signal would be appropriate. the fed model offers a benchmark for fair value but the time series properties of the multiplier plays a crucial role. the example only holds if the multiplier is a random variable, making the long run mean the best estimate of the expected multiple in the next period. 3.1. why the fed model might work the basic argument for using the fed model as a measure of normal market relationships rests on the view that stocks and bonds are competing asset classes, prompting investors to make yield comparisons when allocating assets. when stock yields are high relative to bond yields (e/p � yt10), investors buy stocks and funds flow away from bonds to stocks. the process will bring stock prices up and equity yields down back in line with bond yields. critics point out that the earnings yield is a real return while the treasury yield is in nominal terms, suggesting that the competing asset explanation is flawed (asness, 2003). nevertheless, investors 170 j.m. clinebell et al. / financial services review 26 (2017) 167–180 may follow the fed model because of money illusion, at least in the short run. practitioners emphasize the descriptive validity of the fed model rather than its theoretical validity. another justification for using the fed model to evaluate market valuation builds on the discounted cash flow model for equity valuation. the discount rate in this valuation process is a risk free rate plus an equity risk premium. the long run treasury yield is a proxy for the risk free rate in this calculation. as the treasury yield falls the present value of equity cash flows (p) increases and the earnings yield (e/p) falls as the fed model predicts. critics point to the potential for time varying risk premiums, which are not in the fed model. a counter to this argument is that risk premiums are very difficult to predict (fernandez, aguirremalloa, & corres, 2011) and implementation of valuation theory in the real world uses a stable long run average risk premium rather than time varying forward estimates. finally, the fed model could be valid simply because practitioners use it. the market moves when the ratio of the earnings yield to the treasury yield reaches an inflection point if investors move funds in responses to that ratio, for whatever reason. 3.2. prior use of the fed model for asset allocation the original version of the fed model in eq. (1) has been tested using monthly data by koiva, pennanen, and ziemba (2005), shen (2003), and ziemba and schwartz (1991). these studies find that timing based on extreme value switch points using the traditional fed model adds value beyond a buy and hold strategy. even so, the studies are subject to various criticisms such as data mining with multiple models, failure to adjust for taxes and transaction costs, and not covering the financial crisis period in the post-2008 period. 4. time varying fed model rather than pick an arbitrary value of � as the switch point for asset allocation or use the mean based on the assumption that �t is a random variable, we test for the time series properties of the fed model. the strict version of the fed model suggests that the multiplier �t is one, but there is no prediction of how deviations will adjust back to one over time. less strict views of the fed model would allow for persistent deviations from one because of differences in equity risk premiums or changes in growth. the time series for the multiplier is subject to empirical testing. the information content of the fed multiplier (�t) may follow a random variable, random walk, or autoregressive series. eq. (2) is the constant coefficient benchmark model in our empirical analysis. (e/p)t � �(y10)t � �t (2) the specification of eq. (2) implicitly places a restriction on the multiplier coefficient (�) by assuming it is a constant. there is good reason to suspect that the multiplier (�) may change over time. for example, federal reserve intervention may artificially keep the treasury yield from adjusting when equity earnings yields are abnormally high or low. other arguments for time variation in the multiple (�) rest on the discount model approach to equity valuation if changing 171j.m. clinebell et al. / financial services review 26 (2017) 167–180 risk premiums or growth affect the relationship between equity earnings yields and treasury yields.6 kalman’s (1960) estimation procedure, known as the kalman filter, provides a flexible model for testing and estimating a model’s time varying coefficients. in our context, kalman’s time varying parameter model relaxes the assumption that � is constant and allows testing of the hypothesis that time variation occurs in the parameter (�t) versus the null hypothesis that the parameter is constant (�t). we modify eq. (2) to create a measurement equation with a time-varying coefficient as eq. (3). (e/p)t � �t(y10)t � �t (3) a state equation models the time variation of �t. in our empirical work, we test the following specification for time variation of the fed multiple (�): �t � �0 � �1�t�1 � �t (4) where the disturbance term in the state equation (�t) has a normal distribution with a mean of zero and a constant standard error (eq. 5). �t � n(0, �) (5) time variation in the fed multiplier (�t) may follow a random walk, random variable, or an autoregressive form. the constant coefficient model of eq. (2) is a special case of the model in eqs. (3) and (4). the coefficient (�) is constant if the standard error (�) from eq. (4) is zero. the coefficient (�t) follows a random walk if the standard error is not zero and the intercept coefficient is zero (�0 � 0) while the slope coefficient estimate is equal to one (�1 � 1). if the standard error � is not zero and the estimate of coefficient �1 is zero, the multiplier �t in eq. (3) is a random variable with a mean of �0. the fed multiplier �t is autoregressive if the standard error � is not zero, the intercept coefficient is not zero (�0 � 0), and the coefficient �1 is between zero and one.7 the model in eqs. (3) and (4) represents a kalman filter specification of the state equation following the cooley-prescott (1973) adaptive regression approach. estimation of the model follows a recursive maximum likelihood process. the process uses an updating method that bases the regression estimates for each period on the last period’s estimates plus data from the current period. kahl and ledolter (1983) show that a recursive kalman filter approach obtains simultaneous maximum likelihood estimates of the parameters in eqs. (3) and (4). a likelihood ratio test statistic with a �2 distribution provides a test of the null hypothesis (� � 0), which implies that the coefficient of the model (l) is constant. in general, our empirical models allow us to first test for the time series model up to the holdout period and then implement the fed model using the vanguard fund indexes and the appropriate benchmark for a switch point (�t). 5. test for time series properties of the fed multiplier we use shiller’s monthly data for aggregate earnings, s&p 500 price, and the 10-year treasury yield to first test time series models of the fed multiplier over diverse markets.8 we 172 j.m. clinebell et al. / financial services review 26 (2017) 167–180 end the test period analysis with shiller’s data in december 2011, because we later use the vanguard index fund data from january 2012 through december 2016 as our out of sample investment period. shiller computes monthly earnings data from the s&p four-quarter totals with linear interpolation to monthly figures over a very long time span starting in 1871. stock price data are monthly averages of daily closing prices. one advantage of using the shiller data source is that it is readily available and often used in academic studies, avoiding data construction as a source of different findings. the center plots in fig. 1 represent the kalman filter estimates (�t�1, t) of the fed multiplier in eqs. (3) and (4) from 1871 through 2011. a 95% confidence interval around the center plots also appears in fig. 1 with a dark arrow representing a multiplier of one, as predicted by the strict version of the fed model. the only time the 95% confidence interval does not contain the strict fed multiplier value of one occurs briefly at the end of wwii. fig. 1 illustrates time variation in the fed multiplier but the multiplier is much more stable and consistent with the strict fed model (�t�1, t � 1) before and after the period from 1912 to 1954. this period contains the great depression, wwi, wwii, and the korean war. the multiplier is stable and close to the strict fed model prediction in the modern era after 1954. table 1 provides specific tests for time variation in the multiplier and for the time series properties of the varying coefficients (multipliers). the test uses shiller’s data for the period from 1988 through 2011, leading up to our holdout investment period. the choice of the test period focused on modern markets while still allowing for a long time span. four different models of the time series for the multiplier (�t) appear in the table to include constant coefficient, random variable, autoregressive, and random walk specifications. the �2 test statistic in the second column of table 1 offers a test for time varying coefficients, based on the value of the standard error (�) in eq. (4). a statistically significant �2 value supports rejection of the hypothesis that the standard error is zero and the coefficient (�t) is fig. 1. kalman filter estimates of the fed multiplier (�t�1, t) and the 95% confidence interval (�t�1, t � 2 rmse) using shiller’s data from 1871 through 2011. notes: y the dark arrow represents a multiplier value of 1, consistent with the strict version of the fed model. y the strict form of the fed model multiplier (� � 1) is contained in the 95% confidence interval throughout the long history of the shiller data up to the investment holdout period. the rmse notation represents the root mean square error. y the vertical lines represent the period from 1912 to 1954 where the multipliers were higher and more volatile. this period contains the great depression, two world wars, and the korean war. 173j.m. clinebell et al. / financial services review 26 (2017) 167–180 a constant in eq. (3). the �2 values rest on differences in the log likelihood values of a constant coefficient specification in eq. (4) versus a time varying specification. for every alternative time varying model, the �2 statistic is highly significant, supporting rejection of the hypothesis of a constant coefficient (multiplier). both the log likelihood values and akaike (1974) goodness of fit measures in table 1 support a kalman filter model where the next month’s expected multiplier is equal to the last month’s multiplier. because the random variable model also provides a good fit, the average multiplier is also close to the best estimate. the goodness of fit results suggest that the last observed multipliers remain close to the long run mean multiplier, but the multiplier is not a constant. fig. 2 shows the estimated values of the fed multiplier using the kalman filter over the test period from 1988 through 2011. the multiplier is time varying but does not stray far from one, ranging from a high of 1.1 to a low of just over 0.85. the combined findings of fig. 1, table 1, and table 2 support a stable, but not constant multiplier in the 24 years leading up to our investment holdout period. the best estimate of the next period’s multiplier table 1 test results for alternative time varying kalman filter coefficient models using shiller’s data from january 1988 through 2011 coefficient test model �2 � 2(llconstant � llvariable) log likelihood (ll) akaike criterion constant n/a n/a n/a random variable 2,516a �5,986b 7.088c autoregressive 17,542a �13,499 15.985 random walk 2,030a �5,743b 6.801c notes: athe �2 statistic is significant at the 0.001 level, supporting rejection of the constant coefficient specifications. bthe higher log-likelihood value corresponds to the model with the best fit to the data. the random walk offers the best fit but the random variable model is a close second. cthe lowest akaike value corresponds to the model with the best fit to the data. the random walk is the best fit but the random variable is again a close second. fig. 2. test period estimates of the fed multiplier (1988 through 2011 period). 174 j.m. clinebell et al. / financial services review 26 (2017) 167–180 would be the last observed multiplier based on these findings but we would not expect it to deviate more than 0.15 from one. the caveat is that in fig. 1 we see that periods of major global disruptions, unlike anything we have seen since the late 1940s and early 1950s, appear to affect the multiplier relationship with equity and bond yields. 6. investment results in out-of-sample markets a model that offers a good fit to the data in a test period may offer good investment performance over that same period, but may not offer good predictions or investment performance in another market period. this is a key criticism of studies that find a predictable relationship over a period and then report the investment performance from only that period. we first examine the robustness of the kalman filter test findings from the long run test period in table 1 by replicating the analysis in our investment period from january 2012 through december 2016 using the vanguard mutual fund data. we then test the investment performance of the kalman filter by creating a portfolio based on switches between vanguard funds using estimated multipliers and the strict fed model. ultimately, the issue is whether the switching information from the kalman filter model translates into better investment performance than buying and holding either the vanguard equity or the bond index mutual funds. 6.1. robustness of kalman filter tests table 2 reproduces the tests of time variation in the fed model reported in table 1 but with the vanguard mutual fund data and our holdout investment period from january 2012 through december 2016. we now use the monthly holding period return relatives for each of the vanguard funds over the recent holdout period rather than shiller’s data. our findings in table 2 mirror the findings from table 1. the �2 tests support time varying coefficient models over a constant coefficient model. both the log likelihood and akaike goodness of fit measures support a random walk specification of the multiplier in the kalman filter model closely followed by a random variable. the model performs consistently over different periods and with different data sources. table 2 test results for the kalman filter time varying coefficient models using vanguard mutual fund data over the january 2012 through december 2016 out of sample investment period coefficient model �2 � 2 �llconstant � llvariable� log likelihood (ll) akaike criterion constant n/a n/a n/a random variable 2,178a �628b 4.5537c autoregressive 2,842a �960 6.9568 random walk 2,134a �606b 4.3458c notes: athe �2 statistic is significant at the 0.001 level, supporting rejection of the constant coefficient specifications. bthe higher log-likelihood value corresponds to the model with the best fit to the data. the random walk offers the best fit but the random variable model is a close second. cthe lowest akaike value corresponds to the model with the best fit to the data. the random walk offers the best fit but the random variable model is again a close second. 175j.m. clinebell et al. / financial services review 26 (2017) 167–180 6.2. investment performance results from the strict fed model our findings up to this point suggest that the fed multiplier does not exhibit a trend and deviations of the multiplier from month to month are relatively small leading up to our investment holdout period. our long run analysis of kalman filter estimates of the fed multiplier suggests that the strict version of the fed model provides a reasonable benchmark for making switching decisions. the kalman filter provided a robust prediction model for the fed multiplier. the next step is to evaluate whether using the model to switch between vanguard index funds in recent markets provides attractive investment performance. we constructed an active switching portfolio by making asset allocation shifts to the vfinx when the expected multiplier from the kalman filter model is greater than one (�t�1, t � 1). in this case, the expected aggregate equity yield, using the vanguard s&p500 index fund as the price index, exceeds the expected yield on the vfitx. this position is held until the expected multiplier is less than one (�t�1, t � 1), calling for a switch from the vanguard s&p500 index fund to the vfitx. there is potential for as many as 12 switches per year in this five-year period. however, movement of funds monthly within the vanguard family does not carry a transaction cost for trading. furthermore, any capital gains or losses would be irrelevant for retirement fund investments. table 3 summarizes the investment performance using the buy and hold vanguard funds and the active switching portfolio over the holdout investment period. the switching portfolio has an average monthly rate of return of 1.96% (26.23% effective annual rate), which is higher than the average monthly returns for either of the vanguard fund returns. the monthly standard deviation and monthly rate of return of the vanguard index treasury bond fund is the lowest, as expected. however, on a reward for risk basis, the switching portfolio has the highest ratio of monthly returns to monthly standard deviation (0.985). the switching portfolio also offers the best upside gain relative to the downside risk, providing favorable positive skewing of the return performance.9 the ratio of upside monthly standard deviation to downside monthly standard deviation is 1.29 for the switching portfolio relative to 1.113 for the vanguard index treasury bond fund and only 0.902 for the vanguard s&p 500 table 3 investment performance in the holdout period january 2012 through december 2016 for the vanguard index fund portfolios and the active switching portfolio based on the strict fed model portfolio measure vfinx portfolio vfitx portfolio active switching portfolio average monthly return .01074 .00093 0.0196a effective annual rate of return .1368 .0112 .2623b standard deviation of monthly return 0.0298 0.0102c 0.0199 average monthly return/standard deviation 0.360 0.0912 0.985d upside to downside monthly volatility .902 1.113 1.291e notes: vfinx � vanguard’s s&p500 index fund; vfitx � vanguard’s treasury index fund. athe active switching portfolio has the highest average monthly return. bthe active switching portfolio has the highest effective annual rate of return. cthe vanguard index intermediate bond fund has the lowest monthly volatility of monthly returns. dthe switching portfolio provides the best return for risk tradeoff. ethe switching portfolio provides the best upside to downside volatility. 176 j.m. clinebell et al. / financial services review 26 (2017) 167–180 index fund. for many investors, the favorable positive skewing of return performance from the switching portfolio is especially important. pfeifer (1985) demonstrates how an investor’s utility function with respect to loss aversion should be included in investment performance. prospect theory (kahneman & tversky, 1979) suggests investors fear losses more than they enjoy gains, making a simple reward to risk metric insufficient for investment performance. investors who chase winners play a losing game by buying stocks when they are high and selling when stocks are low. the switching portfolio offers a counterbalance to the tendency to chase winners. when the earnings yield (e/p) is high, the market price is low relative to earnings. rather than sell stocks in a falling market, the switching portfolio buys stocks. this contrarian aspect of investing with the fed model is an attractive feature for investors subject to recency, herding, and regret aversion biases outlined in the behavioral finance literature (nofsinger, 2013). 6.3. switches in the strict fed model fig. 3 shows the variation in the kalman filter estimates of the fed multiplier (� t�1, t) over the investment performance period from january 2012 through december of 2016. the multiplier estimate moved within a tight range from 1.05 to 0.95. each crossing of the estimate over the strict fed multiplier of one represents a portfolio switch in vanguard funds. we assumed a priori that the strict version of the fed model would be an appropriate benchmark for our switching portfolio rather than mine the data to find an optimal switch benchmark. our decision to use the strict fed model is consistent with the long run performance of the kalman filter found in the test period. we did not find that changing risk premiums, revisions of growth estimates, absence of money illusion, or other criticisms of the theoretical arguments invalidated the strict fed model interpretation. fig. 3. fed multiplier estimates (�t�1, t) using vanguard index fund data over the january 2012 through december 2016 out of sample investment period. notes: y estimation of the fed multiplier with the kalman filter uses the vanguard treasury index fund (vfitx) and the vanguard s&p 500 index fund (vfinx) data over the investment period. y the multiplier has a lack of trend with volatility around a multiplier of �t�1, t � 0 0.95 and a range of multiplier values from 1.05 to 0.95. 177j.m. clinebell et al. / financial services review 26 (2017) 167–180 7. conclusions and implications the controversy over the potential for active asset allocation to enhance investment performance is likely to continue with a debate over methods, data, return adjustments and interpretations of findings. in this article, we first focused the debate on switching funds within a family of funds to bypass concerns over transaction costs and tax effects. this focus did not reduce the relevance of the study, given the large retirement fund universe. we used a simple “risk on” and “risk off” approach with a basic switch rule defined by the strict fed model. shiller’s data allowed a very long run test period of the time varying properties of the fed model rather than assuming mean reversion or other forms of moving averages. the flexible kalman filter model offered time varying estimates of the relevant fed model multiplier in the test period. in a holdout investment period of recent markets, investment performance from switching between index funds in the vanguard family of funds resulted in attractive investment performance relative to a buy and hold of either index fund separately. the performance was especially attractive for investors seeking positively skewed investment performance consistent with prospect theory where the investor values gains less than avoidance of losses. added research could enhance the switching methods outlined here. we limited our analysis to a “risk on” and “risk off” approach, but switching at only extreme multiplier values may offer added enhancements to performance. however, such extreme value switch rules would need to be determined in out-of-sample test periods. we also know that our reported performance in the 2012 to 2016 period may not be representative of performance going forward. even so, our analysis suggests that the fed model is stable outside the 1912 to 1954 period where major global disruptions occurred. we used monthly switch points that might be excessive for many individual investors, calling for additional work on quarterly or annual switch point analysis with the kalman filter model. it is common to find changing time series patterns by changing the holding period of returns. this issue merits more attention in the context of the fed model. while we used the vanguard index funds in this analysis, it would be possible to use etfs rather than mutual funds and gain added performance. the equivalent etfs would be the vanguard s&p500 etf (voo), for the vfinx mutual fund and the vanguard intermediate-term government bond etf (vgit), for the vfitx mutual fund. notes 1 vanguard (2015) allows an investor to buy or exchange back into the same fund, in the same account, every 30 calendar days. there is no limitation for etfs. see https://personal.vanguard.com/us/whatweoffer/overview/redemptionpolicy 2 kahneman and tversky generated volumes of research that define behavioral finance. for a very readable reference to the totality of the work of kahneman and tversky, see lewis (2016). 3 the original impetus for the fed model came from the “monetary policy report to congress pursuant to the full employment and balanced growth act of 1978.” the 178 j.m. clinebell et al. / financial services review 26 (2017) 167–180 federal reserve bank never officially adopted the fed model but greenspan (2007) references the model in his writings. 4 some applications of the fed model use the spread between the earnings yield (e/p) and the treasury yield (y10) rather than the ratio. in this case the fed model is [e/p � �y10] � 0. again, when � � one the strict fed model holds. 5 the example in bodie, kane, and marcus (2014) use the spread between the equity earnings yield and the treasury bond yield in their example. 6 a manipulating of the dividend discount model shows that the equity earnings yield (e/p) is a function of the risk free rate, equity risk premium, and earnings growth rate. the fed model focuses on the relationship between (e/p) and the risk free rate, assuming that variation in the risk premium and earnings growth either remain stable or investors do not consider them in short run monthly investment decisions. 7 the estimates of the slope and intercept coefficients determine the time series model at work with the data. if both the intercept and slope are significant, both the mean and the last observation determine the next period’s multiplier. if the intercept is insignificant with a significant slope of about one, the last observation determines the next period’s multiplier. with a significant intercept and insignificant slope coefficient, the mean is the best estimate of next period’s multiplier. 8 to obtain shiller’s data see (http://www.econ.yale.edu/�shiller/data.htm). 9 see sortino (1994) for a more in-depth treatment of the use of upside to downside standard deviation relationships. references akaike, h. (1974). a new look at the statistical model identification. ieee transactions on automatic control, 19, 716–723. aronson, d. (2006). evidence-based technical analysis: applying the scientific method and statistical inference in trading signals. summerset, nj: john wiley & sons. asness, c. s. (2003). fight the fed model. journal of portfolio management, 30, 11–24. ball, r., kothari, s. p., & wasley, e. c. (1995). can we implement research on stock trading rules? journal of portfolio management, 21, 54–63. bauer, r., & dahlquist, j. (2001). market timing and roulette wheels. financial analysts journal, 57, 28–40. bodie, z., kane, a., & marcus, a. j. (2014). essentials of investments (10th ed., p. 429). new york, ny: mcgraw-hill. breen, w., glosten, l. r., & jagamathan, r. (1989). economic significance of predictable variations in stock index returns. journal of finance, 44, 1177–1189. campbell, j. y. (2000). asset pricing in the millennium. journal of finance, 55, 1515–1567. campbell, j. y., andrew, l., & mckinley, c. (1999). the econometrics of financial markets. princeton, nj: princeton university press. cooley, t., & prescott, e. (1973). tests of an adaptive regression model. review of economics and statistics. february, 248–256. daniel, k., hirshleifer, d., & subrahamanyam, a. (1998). investor psychology and investor security market underand overreaction. journal of finance, 53, 139–209. debondt, w. f. m., & thaler, r. (1985). does the stock market overreact? journal of finance, 40, 793–805. fama, e. (1998). market efficiency, long-term returns, and behavioral finance. journal of financial economics, 49, 283–306. 179j.m. clinebell et al. / financial services review 26 (2017) 167–180 feldman, t., jung, a., & klein, j. (2015). buy and hold versus timing strategies: the winner is … journal of portfolio management, 42, 110–118. fernandez, p., aguirreamalloa, j., & corres, l. (2011). u.s. market risk premium used in 2001 by professors, analysts, and companies: a survey with 5,731 answers. (available at http://srrn.com/abstract�1805852). greenspan, a. (2007). the age of turbulence: adventures in a new world. new york, ny: penguin press. guido, r., peral, j., & walsh, k. (2011). market timing under multiple economic regimes. accounting and finance, 51, 501–515. kahl, d. r., & ledolter, j. (1983). a recursive kalman filter forecasting approach. management science, 29, 1325–1333. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decisions under risk. econometrica, 47, 263–329. kalman, r. e. (1960). a new approach to linear filtering and prediction problems. journal of basic engineering, 82, 35–45. koivu, m., pennanen, t., & ziemba, w. t. (2005). cointegration analysis of the fed model. finance research letters, 2, 248–259. krueger, t. m., & kennedy, w. f. (1990). an examination of the super bowl stock market predictor. journal of finance, 45, 691–697. lewis, m. (2016). the undoing project: a friendship that changed our minds. new york, ny: w. w. norton & company, inc. lo, a. w., mamaysky, h., & wang, j. (2000). foundations of technical analysis: computational algorithms, statistical inference, and empirical implementation. the journal of finance, 55, 1705–1770. malkiel, b. g. (2003). the efficient market hypothesis and its critics. journal of economic perspectives, 17, 59–82. nofsinger, j. j. (2014). the psychology of investing (5th ed.). upper saddle river, nj: pearson education, inc. pfeifer, p. (1985). market timing and risk reduction. journal of financial and quantitative analysis, 20, 46–50. pinches, g. (1970). the random walk hypothesis and technical analysis. financial analysts journal, 26, 104–110. pruitt, s. w., & white, r. e. (1988). the crisma trading system: who says technical analysis can’t beat the market? journal of portfolio management, 14, 55–58. sharpe, w. (1975). likely gains from market timing. financial analysts journal, 31, 60–69. shen, p. (2003). market timing strategies that worked. the journal of portfolio management, 29, 57–68. shilling, g. (1992). market timing better than a buy-and-hold strategy. financial analysts journal, 48, 46–50. shleifer, a. (2000). inefficient markets: an introduction to behavioral finance. oxford, uk: oxford university press. sortino, f. a. (1994). performance measurement in a downside risk framework. journal of investing, 3, 59–65. the vanguard group. (2015). vanguard costs, fees, and minimums. (available at https://investor.vanguard.com/ etf/fees). sullivan, r., timmerman, a., & white, h. (1999). data-snooping, technical trading rule performance, and the bootstrap. journal of finance, 54, 1647–1691. xiong, j., ibbotson, r. g., idzorek, t., & chen, p. (2010). the equal importance of asset allocation and active management. financial analyst journal, 66, 22–30. ziemba, w. t., & schwarz, s. l. (1991). invest japan: the structure, performance and opportunities of japan’s stock, bond and fund markets. chicago, il: probus publishing. 180 j.m. clinebell et al. / financial services review 26 (2017) 167–180 financial services review, 32(1) i volume 32 issue 1 from the editor welcome to the latest issue of financial services review (fsr). this issue includes four fascinating and thought provoking papers. the papers comprising this issue provide an insight into the varied and multidisciplinary approaches currently being utilized by researchers in the field. the first paper was written by drs. ning wang, yiling deng, and ruohan wu. their paper looks at the way household financial decision-makers make portfolio allocation decisions in the context of the demand for life insurance. what they found is quite interesting; namely, life insurance ownership is affected by a household’s decisions to invest in cash and cash equivalents, bonds, and retirement assets. they also observed a relationship between debt repayment strategies and life insurance ownership. as noted by one of the journal’s associate editors, this study is significant in adding to the financial planning literature while at the same time contributing to the way financial service professionals provide advice. the second paper was written by drs. jason anderson, jeffrey furlong, and stuart heckman. their paper deals with a relatively under-researched topic—altruistic bequests. what makes their research so impactful is that they view bequests from a motivational and receipt perspective. they found that a variety of economic and attitudinal variables help to describe the giving and receiving of bequests. my expectation is that their work will spur on more estate planning research. the third paper, written by drs. danah jeong, benjamin hampton, and kristy archuleta, explores factors related to financial well-being. using a framework conceptualized by dr. so-hyun joo at ewha womans university, these authors found that a financial decision-maker’s demographic profile, financial status, financial behavior, and financial satisfaction were associated with reports of financial well-being during the covid-19 pandemic. furthermore, they noted that financial satisfaction mediates the relationship between subjective financial knowledge and financial wellbeing. the last paper in this issue was written by drs. stéphane chrétien and manel kammoun. this captivating paper describes a study that was designed to provide insight into investor disagreement in the performance evaluation of equity mutual funds in describing fund characteristics, the degree of active fund management, and fund flows. their work not only advances the investment and fund management literature, their study provides unique insights into the way fund investment style and net fund flows can be used to explain investor disagreements. i am going to end this editorial with some housekeeping news. since moving to the journal’s new management system, the number of submitted papers has increased. we have also noticed that the quality of submissions continues to improve. these are trends we hope to see continue into the future. my promise back to everyone who has or is considering submitting a paper to financial services review is that the associate editors and i will continue to work diligently to return decisions back to authors in a timely manner. working together, we can ensure that the journal is published on time, which is one way to improve the journal’s impact. given where the journal is at this point in time, i do want to remind those who are considering submitting a paper to carefully review the journal’s submission guidelines. the day when someone could write a paper without paying close attention to their paper’s style, formatting, editing, references, and reference list, then submitting that paper without fear of a desk rejection or a quick financial services review, 32(1) ii unreviewed return, is quicky coming to an end. as the number and quality of submissions continues to increase, competition for the few publication slots in the journal are quickly being filled. the best way to ensure that your submission gets through the review process seamlessly is to take the time to format the submission process appropriately. with this in mind, let’s go over the number one problem we continue to see with submissions. financial services review generally follows the apa manual. let’s say you typically write papers using mla or harvard or some other specialized system, but now you are wanting to submit to financial services review. one option is just to submit and see what happens. a better alternative is to go through the reference list and change the references so they follow apa. here is a quick and easy way to do just this. • go to google scholar. • copy your current reference into the search bar; here is what a reference would look like in mla: grable, john, and ruth h. lytton. "financial risk tolerance revisited: the development of a risk assessment instrument☆." financial services review 8.3 (1999): 163-181. • rather than reformat, find the links underneath the paper; these will look like this: • now, click on the icon; this is what you’ll see: • click on the apa reference, copy it, and paste it into your paper. this will work 90% of the time. when it comes to in-text references, visit the purdue university library site using apa formatting as a google search term or look at the journal’s style examples. if you do these simple things, your chances of having your paper move through the review process smoothly will be greatly improved. finally, one correction to an editorial from 2023. the letter from the editor for volume 31, issue 2/3 inadvertently omitted dr. sharon devaney as a coauthor of the best paper award for “fear and trust in financial institutions” that was written with drs. isha chawla, mia russell, and kenneth white. the editorial staff wishes to express our sincere apology for this omission. all the best, john e. grable, ph.d., cfp® pii: 1057-0810(95)90001-2 new directions karen eikrs lahey financial services review (fsr) is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines i~ivid~~ be~vi~r in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the fsr journal provides a forum for those who are interested in the individual perspective on issues in the area of banking/banking services, education in financial services, employee benefits, estate and tax planning, financial counseling, financial pianning, insurance, investments, mutual funds, nonbank financial institutions, pension and retirement planning, and real estate. while the annual meeting in october provides an opportunity to discuss and present these topics to colleagues, thejournal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, app~p~ate p~agogical papers will be accepted for publication. this represents a new area for the fsr journal. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. authors of new and revised textbooks and materials (including computer programs) are also encouraged to submit their work for review by members of the academy. the reviews will be published in a new section that will provide for rigorous analysis of published materials that are available to those teaching at universities and colleges who wish to consider them for use in their classes. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are welcome. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross-fertilization of ideas and an exchange of info~ation of interest to both ac~emicians and practitioners. thus, the new editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. the journal will also start an afs notes section. it will feature brief notes, responses to recent papers, comments, and letters to the editors. this is your opportunity as a member of the academy of financial services to give us the benefits of your thoughts in a format much briefer than a journal article. v vi financial services review 4(2) 1995 in 1997, the journal will become a quarterly publication. the addition of two more issues a year will allow publication of additional academic empirical research as well as the new features that are described above. if you have ideas for manuscripts or are interested in accepting the significant responsibility of producing a special issue of the journal, contact the new editor with your thoughts. finser_31-2_complete_issue investment literacy, overconfidence and cryptocurrency investment kyoung tae kima,*, sherman d. hannab, sunwoo t. leec adepartment of consumer sciences, university of alabama, 316-c adams hall, box 870158, tuscaloosa, al 35487, usa bdepartment of human sciences, the ohio state university, 1787 neil ave, columbus, oh 43210, usa cschool of administrative studies, york university, atkinson building, 252, 4700 keele street, toronto, on m3j1p3, canada abstract cryptocurrency has been increasingly popular with investors. using the 2018 national financial capability study investor survey, we examined the association between investment literacy and cryptocurrency investment—about 13% of investors invested in cryptocurrency directly or indirectly. results from regression analyses show that objective investment literacy was negatively while subjective literacy was positively associated with holding cryptocurrency. overconfident investors were more likely to invest in cryptocurrency, and results were robust across three overconfidence measures. this study has implications for investment advice, financial education, and research. © 2023 academy of financial services. all rights reserved. jel classification: g5; g11; g53 keywords: investment literacy; financial knowledge; overconfidence; investment; cryptocurrency 1. introduction cryptocurrency has become increasingly popular since 2008, since the invention of bitcoin by satoshi nakamoto (nakamoto, 2008). although bitcoin still dominates the cryptocurrency market, nearly 8,600 cryptocurrencies are currently trading around the world with a combined market capitalization of us$1.48 trillion as of march 2, 2020 (coin market *corresponding author: tel.: 205-348-9167, fax: 205-348-8721. e-mail address: ktkim@ches.ua.edu 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 121–132 cap, 2020). the total value of holdings in cryptocurrencies is still small relative to holdings of gold and financial investments but is being increasingly studied as a component of portfolios of wealthy investors (zhao, 2021). one indicator of the growing importance of cryptocurrencies is concern about the impact on the environment of the production of cryptocurrencies (sorkin, 2021). although investment in cryptocurrency has been rapidly growing, there are not many studies on factors related to people investing in cryptocurrency. one of the major reasons is the limited availability of appropriate survey data and questionnaires. the 2018 national financial capability study investor survey (thereafter, 2018 nfcs investor survey) released by finra investor education foundation collects a series of questions about cryptocurrency investment. this allows us a unique opportunity to identify factors associated with investing in cryptocurrency by us investors. given limited knowledge about investing in cryptocurrencies and the extreme volatility as an investment, what type of investor holds cryptocurrency investments? the purpose of this study is to investigate the association between investment literacy and cryptocurrency investment. we tested investment literacy in three ways: (1) objective literacy, (2) subjective literacy, and (3) overconfidence in investment literacy. to check the robustness of our results, we used three different indicators of overconfidence in investment literacy, measured by divergence between objective and subjective literacy. in addition to the analysis on the association between investment literacy and investing in cryptocurrency, we examined various socioeconomic factors related to investing in cryptocurrency. this study contributes to the existing literature in two ways. while there has been increasing attention to cryptocurrency investment, our study is one of first attempts to explore the association between investment literacy overconfidence and investing in cryptocurrency of us investors. further, we used an index of correct answers to investment knowledge questions, which captures one’s investment literacy more comprehensively than commonly used measures of financial literacy (lusardi & mitchell, 2014). the major findings of this study provide important insights for investment advisors, financial educators, and researchers. 2. literature review and theoretical consideration 2.1. previous studies on cryptocurrency investments while cryptocurrency is labeled “currency,” the actual adoption of cryptocurrency as a payment method is still a topic of discussion. cryptocurrency has advantages such as confidentiality, reliability of information transmission, and flexible transactions, but drawbacks such as significant volatility, lack of organized platforms, and the difficulty of projection hinder it from being integrated to the global financial market (tasca et al., 2018; titov et al., 2021). cryptocurrency has also gained a tremendous amount of attention as an investment vehicle. chuen et al. (2017) argued that due to low correlation between cryptocurrencies and traditional assets as well as the higher daily expected return of the former option, investing in 122 k. t. kim et al. / financial services review 31 (2023) 121–132 cryptocurrencies could help one diversify their portfolio risks. when examined as a form of financial asset, cryptocurrency presented quite similar dynamics to stock investments (liang et al., 2019). previous studies have documented several factors related to cryptocurrency investments. socio-demographic factors such as age, gender, and education were correlated to cryptocurrency investment (ante et al., 2022; hasso et al., 2019; henry et al., 2018). since cryptocurrency is based on a cutting-edge technological innovation, many studies have also used the technology acceptance model as a theoretical background for cryptocurrency investment. studies confirmed that the new technology’s perceived usefulness and ease of use as well as the risk associated with it affect one’s intention to utilize cryptocurrency (arias-oliva et al., 2019; bharadwaj & deka, 2021). one phenomenon that might be related to the growth of cryptocurrency investment is the bandwagon effect, or herding behavior (avital et al., 2016; bouri et al., 2019; da gama silva et al., 2019). herding behavior combined with the subsequent high volatility could pose significant risks on the financial stability of the cryptocurrency investors. thus, it calls for a cautious approach to cryptocurrency investment and emphasizes the importance of understanding whether investors selected cryptocurrency as an investment vehicle are capable of understanding complex dynamic and the potential consequences of their choices. 2.2. financial literacy and investment decisions unlike what traditional economists have assumed and approached, many studies present evidence that investors are functioning under bounded rationality (de bondt et al., 2008). the theory of bounded rationality (simon, 2000) posits that individuals have limited ability to assess given information and make optimal decisions. this is mainly due to the complexity of environments, limited mental capacity, and limited resources (ibrahim, 2009). financial literacy is a measure of individuals’ ability to understand situations and make optimal financial decisions based on their assessments (seay et al., 2017). therefore, those with higher financial literacy would behave in a more appropriate manner, and financial mistakes would be more common among the financially illiterate people. enhancing financial literacy could help reduce the probability of a major systematic financial crisis due to bounded rationality (siriopoulos, 2021). grounded within the theory of bounded rationality, previous studies have documented the positive link between financial literacy and various financial outcomes including stock ownership (kimball & shumway, 2006; van rooij et al. 2011), diversification (goetzmann & kumar, 2008; guiso & jappelli, 2008; shin et al., 2020), wealth accumulation and planning for retirement (lusardi & mitchell, 2007). cryptocurrency investment is one of the complex financial decisions which would require one’s understanding of the concept and characteristics of cryptocurrency and relatively unique procedures of investment, but there have been few studies conducted on the link between cryptocurrency and financial literacy (arias-oliva et al., 2019; zhao & zhang, 2021). financial literacy has been conceptualized as objective and subjective measures, where objective financial literacy indicates one’s actual understanding of financial concepts while subjective financial literacy refers to one’s perceived understanding of the matter k. t. kim et al. / financial services review 31 (2023) 121–132 123 (kim et al., 2020; robb et al., 2015). objective financial literacy has been found negatively related to cryptocurrency use or acceptance (zhao & zhang, 2021) while subjective financial literacy has been found positively related to (gupta et al., 2020) or did not have a statistically significant relation to cryptocurrency utilization (arias-oliva et al., 2019). based on the findings from previous studies, we constructed the following two research hypotheses on the association between investment literacy and cryptocurrency investment. hypothesis 1: objective investment literacy is negatively associated with investing in cryptocurrency. hypothesis 2: subjective investment literacy is positively associated with investing in cryptocurrency. we focused on the link between overconfidence in investment literacy and cryptocurrency investments, to provide new insights into factors related to holding cryptocurrency as an investment. investors may be influenced by herd behavior instead of making investment decisions based on the value of cryptocurrencies; therefore, it is useful to investigate the cryptocurrency investment from the perspective of bounded rationality, and we extended it to the aspect of overconfidence in this study. overconfidence bias has been described as (1) overprecision or (2) miscalibration, indicating a systematic overweighting of the accuracy of one’s own literacy (e.g., robb et al. 2015; xia et al., 2014). researchers have examined the relationship between overconfidence and irrational behaviors such as excessive trading, excessive risk taking (abreu & mendes, 2012; barber & odean, 2000), under diversification (chu et al. 2017; shin et al., 2020; xia et al., 2014) and high-cost borrowing (robb et al., 2015). some studies have raised concerns about the divergence between objective and subjective financial knowledge (e.g., kim et al., 2020; robb et al., 2015). false confidence in one’s ability to understand complex investment concepts would lead to suboptimal financial decisions. in this study, we propose the following research hypothesis on the association between investment literacy overconfidence and cryptocurrency investment. hypothesis 3: overconfidence in investment literacy is positively associated with investing in cryptocurrency. 3. method 3.1. dataset and analytic sample this study utilized data from the 2018 national financial capability study (nfcs) and the follow-up 2018 investor survey, which includes a subset of respondents who had indicated ownership of nonretirement investments. the 2018 nfcs state-by-state data collected approximately 500 observations per state plus the district of columbia, which leads to the total sample of 27,091 adults in the united states, with approximately 500 observations per state plus the district of columbia. the 2018 nfcs investor survey includes a sample drawn from individuals who indicated owning nonretirement investments. to explore the associations between investment literacy and investing in cryptocurrency, the analytic 124 k. t. kim et al. / financial services review 31 (2023) 121–132 sample started with the 2,003 individuals who completed both the 2018 nfcs and the 2018 nfcs investor survey. the final sample includes 1,819 investors, excluding missing responses. 3.2. dependent variables the 2018 nfcs investor survey includes questions on cryptocurrency. the dependent variable is a binary indicator whether the respondent has invested in cryptocurrency, based on the following question, “have you invested in cryptocurrencies, either directly or through a fund that invests in cryptocurrencies?” 3.3. investment literacy and overconfidence the 2018 nfcs survey includes questions designed to measure investment literacy covering various investment-related topics and concepts. we created an objective investment literacy index, measured as each respondent’s number of correct answers to 10 questions and ranged from 0 to 10. the topics of investment literacy cover stock, bond, bankruptcy, investment risk, investment return, municipal bond, stock margin, selling short, investment indicator, and index fund. the subjective assessment of investment knowledge was measured by the following question: “on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall knowledge about investing?” following the approach of kim et al. (2020), we tested three measures of investment literacy overconfidence: (1) overconfidence defined as having higher than the sample median for subjective literacy, but lower than the sample median for objective literacy; (2) a continuous measure of divergence between objective and subjective financial literacy; and (3) the residual from a least squares regression of objective literacy on subjective literacy. 3.4. control variables in addition to investment literacy variables, we included age, gender (male, female), marital status (married, single, separated/divorced/widow(er)), having a dependent child, race/ ethnicity (white, black, hispanic, asian/others), employment status (full-time employee, self-employed, part-time employee, homemaker, student, disabled, unemployed, retired), education (high school diploma or lower, some college, associate degree, bachelor degree, postbachelor degree), household income, investment assets in nonretirement account, and region (state of residence) as control variables. 3.5. empirical specification we used logistic regression models to investigate the effects of various factors associated with an investment in cryptocurrency, especially for the role of investment literacy and its overconfidence measured by divergence between objective and subjective literacy. both descriptive and multivariate results were weighted using the sampling weight provided by nfcs investor survey. empirical models are expressed as follows; k. t. kim et al. / financial services review 31 (2023) 121–132 125 pr cð þi = log pr cð þi 1# pr cð þi ! " # = b 0 þ b 1ili þ b 2xi þ « i where, pr cð þi is the probability of investing in cryptocurrency; b 0 is an intercept; il denotes the level of investment literacy including objective, subjective investment literacy (baseline model) and three overconfidence induce. the control variables and state of residence are denoted as xi. 4. results 4.1. descriptive results as shown in table 1, about 13% of investors invested in cryptocurrency directly or indirectly. based on the proportion of all households with nonretirement investments, we can infer that less than 4% of all households held cryptocurrency. we also tested the mean differences in investment literacy of cryptocurrency investors and noninvestors. respondents owning cryptocurrency investments had lower objective investment literacy scores and higher subjective literacy than noncryptocurrency investors. in addition, the level of overconfidence in investment literacy was higher for cryptocurrency owners than nonowners across three overconfidence indexes. socio-demographic characteristics of the analytic sample are presented in appendix. 4.2. multivariate results results of logistic regressions on cryptocurrency investment are reported in table 2. in model a (baseline), objective investment literacy was negatively, but subjective literacy was positively associated with investing in cryptocurrency. in particular, a one unit increase in objective literacy decreased the odds of investing in cryptocurrency by 10.5%; while a one table 1 descriptive results of selected variables, 2018 nfcs investor survey variables all investors investing in cryptocurrency not investing in cryptocurrency dependent variable investing in cryptocurrency 12.90% — — investment literacy objective literacy (0–10), mean (median) 4.77 (5.00) 3.96 (4.00)*** 4.89 (5.00) subjective literacy (1-7), mean (median) 4.79 (5.00) 5.55 (6.00)*** 4.68 (5.00) overconfidence in investment literacy index 1 (high subjective/low objective literacy) 28.28% 58.7%*** 23.9% index 2 (numerical difference), mean (median) 0.02 (0.00) 1.58 (3.00)*** #0.21 (0.00) index 3 (residual measure), mean (median) 0.00 (0.18) 0.86 (1.04)*** #0.13 (0.04) notes. weighted results. t test or x2 are conducted for a group comparison. significance level: *p < .05, **p < .01, ***p < .001. 126 k. t. kim et al. / financial services review 31 (2023) 121–132 t ab le 2 l o g is ti c re g re ss io n s o n cr y p to cu rr en cy in v es tm en t, 2 0 1 8 n f c s in v es to r su rv ey v ar ia b le s (a ) b as el in e m o d el (b ) o ve rc on fi de nc e in de x 1 (h ig h su bj ec ti ve , lo w ob je ct iv e) (c ) o v er co nfi d en ce in d ex 2 a (d is cr ep an cy ) (d ) o v er co n fi d en ce in d ex 3b (r es id u al ) c o ef f. s e c o ef f. s e c o ef f. s e c o ef f. s e in v es tm en t li te ra cy v ar ia b le s o b je ct iv e li te ra cy # 0 .1 1 1 4 * 0 .0 4 7 2 — — — — — — s u b je ct iv e li te ra cy 0 .5 9 1 * * * 0 .0 8 1 2 — — — — — — o v er co n fi d en ce in li te ra cy — — 1 .1 2 2 4 * ** 0 .1 9 8 2 0 .2 3 6 9 * * * 0 .0 4 2 0 .5 9 7 9 * * * 0 .0 8 0 7 a g e # 0 .0 9 1 1 * * * 0 .0 1 0 2 # 0 .0 9 1 8 * ** 0 .0 1 0 2 # 0 .0 8 8 3 * * * 0 .0 1 0 2 # 0 .0 9 1 6 * * * 0 .0 1 0 2 m al e 0 .3 8 6 8 0 .1 9 9 3 0 .3 6 6 4 0 .1 9 3 5 0 .4 9 7 5 * 0 .1 9 3 6 0 .3 7 0 7 0 .1 9 7 9 m ar it al st at u s (r ef : m ar ri ed ) s in g le # 0 .6 8 4 4 * * 0 .2 5 2 7 # 0 .7 6 2 4 * * 0 .2 4 8 2 # 0 .7 3 8 7 * * 0 .2 4 7 1 # 0 .6 8 9 8 * * 0 .2 5 2 5 s ep ar at ed /d iv o rc e/ w id o w # 0 .4 5 9 2 0 .3 4 8 6 # 0 .3 4 8 1 0 .3 4 1 7 # 0 .3 6 2 6 0 .3 4 2 1 # 0 .4 8 1 6 0 .3 4 7 1 p re se n ce o f d ep en d en t ch il d (r ef : n o ) 0 .3 5 1 9 0 .2 1 7 1 0 .3 5 0 8 0 .2 1 0 1 0 .3 5 3 9 0 .2 0 9 0 .3 6 2 7 0 .2 1 6 7 r ac e/ et h ni ci ty (r ef : w h it e) b la ck # 0 .0 2 7 3 0 .2 8 3 6 0 .0 4 2 7 0 .2 7 4 2 0 .0 8 5 5 0 .2 7 1 7 # 0 .0 1 6 1 0 .2 8 3 5 h is p an ic # 0 .3 7 8 8 0 .3 4 5 5 # 0 .3 7 9 1 0 .3 3 6 1 # 0 .3 3 5 2 0 .3 3 2 2 # 0 .3 5 4 1 0 .3 4 3 2 a si an /o th er s 0 .6 0 9 9 0 .3 1 5 7 0 .4 5 8 7 0 .3 0 7 0 .6 5 7 5 * 0 .3 0 7 6 0 .5 9 6 8 0 .3 1 5 1 e d u ca ti o n (r ef : h ig h sc h o o l d ip lo m a o r lo w er ) s o m e co ll eg e 0 .5 0 8 2 0 .2 9 0 5 0 .5 1 2 7 0 .2 8 1 3 0 .5 0 9 3 0 .2 8 0 6 0 .5 0 3 4 0 .2 9 0 6 a ss o ci at e d eg re e 0 .5 3 3 1 0 .3 5 1 1 0 .5 0 6 2 0 .3 3 7 9 0 .5 0 3 6 0 .3 4 0 1 0 .5 4 2 1 0 .3 5 1 b ac h el o r d eg re e 0 .0 9 5 7 0 .2 9 3 0 .1 6 7 7 0 .2 8 4 1 0 .1 9 4 3 0 .2 8 5 6 0 .0 7 4 3 0 .2 9 1 p o st -b ac h el o r d eg re e 0 .3 3 2 2 0 .3 5 8 6 0 .3 3 8 7 0 .3 4 8 9 0 .3 8 0 3 0 .3 5 1 1 0 .3 1 0 2 0 .3 5 6 8 e m p lo y m en t st at u s (r ef : fu ll -t im e em p lo y ee ) s el fem p lo y ed # 0 .0 0 6 4 0 .3 0 3 2 0 .0 0 6 9 0 .3 0 0 6 0 .0 2 6 6 0 .2 9 8 3 # 0 .0 1 7 3 0 .3 0 2 6 p ar tti m e w o rk er 0 .3 1 1 1 0 .3 4 0 1 0 .0 9 9 3 0 .3 3 1 0 0 .0 9 3 4 0 .3 2 8 3 0 .3 0 2 9 0 .3 4 0 4 h o m em ak er # 1 .4 0 1 7 * 0 .6 2 3 4 # 1 .4 6 1 7 * 0 .5 9 9 1 # 1 .5 8 9 8 * * 0 .5 9 5 7 # 1 .4 0 3 5 * 0 .6 2 4 8 s tu d en t # 0 .8 5 8 1 0 .4 5 5 0 # 0 .9 1 4 7 * 0 .4 3 9 9 # 0 .9 3 4 4 * 0 .4 4 7 0 # 0 .8 5 4 5 0 .4 5 4 6 d is ab le d 0 .5 0 4 8 0 .8 0 7 0 0 .1 8 7 5 0 .7 8 3 9 # 0 .0 0 0 5 0 .8 0 3 1 0 .5 2 7 3 0 .8 0 4 3 u n em p lo y ed # 1 .2 4 4 6 0 .7 3 4 2 # 1 .1 2 2 6 0 .7 1 7 2 # 1 .0 8 9 7 0 .7 0 8 4 # 1 .2 3 8 3 0 .7 3 3 2 r et ir ed # 0 .1 4 0 2 0 .4 5 5 0 # 0 .2 2 6 3 0 .4 5 2 6 # 0 .2 9 2 6 0 .4 5 4 0 # 0 .1 4 1 1 0 .4 5 4 3 in co m e (r ef : le ss th an $ 3 5 ,0 0 0 ) $ 3 5 ,0 0 0 – $ 4 9 ,9 9 9 # 0 .6 8 6 5 0 .3 6 7 6 # 0 .6 6 9 7 0 .3 6 2 9 # 0 .7 6 3 8 * 0 .3 5 7 9 # 0 .6 8 6 0 0 .3 6 7 7 $ 5 0 ,0 0 0 – $ 7 4 ,9 9 9 # 0 .3 4 4 2 0 .2 9 3 2 # 0 .4 0 9 4 0 .2 8 5 6 # 0 .4 8 9 4 0 .2 8 4 5 # 0 .3 4 8 3 0 .2 9 3 2 $ 7 5 ,0 0 0 – $ 9 9 ,9 9 9 # 0 .5 8 5 9 0 .3 3 6 9 # 0 .5 7 2 8 0 .3 2 9 2 # 0 .7 1 9 2 * 0 .3 2 4 6 # 0 .5 9 1 7 0 .3 3 7 1 (c o n ti n u ed o n n ex t p a g e) k. t. kim et al. / financial services review 31 (2023) 121–132 127 t ab le 2 (c on ti n u ed ) v ar ia b le s (a ) b as el in e m o d el (b ) o ve rc on fi de nc e in de x 1 (h ig h su bj ec ti ve , lo w ob je ct iv e) (c ) o v er co nfi d en ce in d ex 2 a (d is cr ep an cy ) (d ) o v er co n fi d en ce in d ex 3b (r es id u al ) c o ef f. s e c o ef f. s e c o ef f. s e c o ef f. s e $ 1 0 0 ,0 0 0 – $ 1 4 9 ,9 9 9 # 0 .4 3 0 7 0 .3 4 9 7 # 0 .5 9 1 2 0 .3 4 4 6 # 0 .6 1 4 7 0 .3 4 1 6 # 0 .4 4 7 3 0 .3 4 8 8 $ 1 5 0 ,0 0 0 o r m o re # 1 .2 6 9 3 * * 0 .4 8 8 4 # 1 .3 4 5 8 * * 0 .4 8 5 9 # 1 .4 3 0 6 * * 0 .4 8 5 9 # 1 .2 7 1 6 * * 0 .4 8 7 7 in v es tm en t as se ts (r ef : le ss th an $ 5 ,0 0 0 ) $ 5 ,0 0 0 – $ 2 4 ,9 9 9 # 0 .7 1 7 8 * 0 .3 1 5 4 # 0 .7 8 0 5 * 0 .3 0 6 9 # 0 .6 9 2 9 * 0 .3 0 5 4 # 0 .7 1 4 2 * 0 .3 1 5 5 $ 2 5 ,0 0 0 – $ 4 9 ,9 9 9 # 0 .1 9 1 8 0 .3 4 1 7 # 0 .1 9 0 7 0 .3 3 7 1 # 0 .0 9 3 0 0 .3 2 9 7 # 0 .1 9 5 4 0 .3 4 2 5 $ 5 0 ,0 0 0 – $ 9 9 ,9 9 9 # 0 .5 0 0 3 0 .3 2 5 1 # 0 .3 1 3 5 0 .3 1 6 1 # 0 .2 1 0 6 0 .3 1 2 7 # 0 .5 0 6 9 0 .3 2 5 0 $ 1 0 0 ,0 0 0 – $ 2 4 9 ,9 9 9 # 0 .5 7 5 1 0 .3 6 3 0 # 0 .2 8 9 7 0 .3 5 3 6 # 0 .2 6 9 8 0 .3 5 3 6 # 0 .5 8 3 5 0 .3 6 2 4 $ 2 5 0 ,0 0 0 – $ 4 9 9 ,9 9 9 0 .0 9 6 7 0 .3 7 6 9 0 .3 0 8 4 0 .3 6 7 7 0 .3 8 2 5 0 .3 6 8 3 0 .0 9 4 6 0 .3 7 6 0 $ 5 0 0 ,0 0 0 – $ 9 9 9 ,9 9 9 # 1 .3 4 5 3 * 0 .6 4 9 7 # 1 .0 0 1 9 0 .6 3 2 2 # 0 .9 7 9 0 0 .6 3 5 5 # 1 .3 8 0 9 * 0 .6 4 8 8 $ 1 ,0 0 0 ,0 0 0 o r m o re # 0 .0 8 2 2 0 .4 9 4 1 0 .4 3 6 9 0 .4 6 5 8 0 .5 1 0 9 0 .4 6 9 8 # 0 .1 3 4 9 0 .4 8 7 7 c o n st an t # 1 .7 6 5 2 * 0 .8 0 0 8 0 .3 1 7 0 .6 7 0 5 0 .4 3 1 7 0 .6 6 7 2 0 .5 8 2 1 0 .6 8 1 6 r eg io n al fi x ed ef fe ct (s ta te o f re si d en ce ) y es y es y es y es m o d el fi t c o n co rd an ce ra te 8 9 .0 % 8 7 .5 % 8 7 .7 % 8 9 .0 % n o te s. w ei g h te d re su lt s. a d is cr ep an cy b et w ee n su b je ct iv e an d o b je ct iv e fi n an ci al li te ra cy . b w e re g re ss su b je ct iv e fi n an ci al li te ra cy o n th e o b je ct iv e li te ra cy an d ta k e th e re si d u al as an o v er co n fi d en ce m ea su re . s ig n ifi ca n ce le ve l: *p < .0 5 , * *p < .0 1 , * * *p < .0 0 1 . 128 k. t. kim et al. / financial services review 31 (2023) 121–132 unit increase in subjective literacy increased the odds by 80.6%. in models b, c, and d, we found strong positive associations between overconfidence (as specified three different ways) and cryptocurrency investment. investors with overconfidence in investment literacy (model b, index 1) had 3.1 times the odds of investing in cryptocurrency as investors who were not overconfident. additionally, a one unit increase in overconfidence index 2 (model c) increased the odds of investing in cryptocurrency by 26.7%, and a one unit increase in overconfidence index 3 (model d) increased the odds of investing in cryptocurrency by 81.8%. among control variables, age was negatively associated with the likelihood of investing in cryptocurrency. single investors were less likely to invest in cryptocurrency than married couples. homemakers and students had lower likelihood of investing in cryptocurrency than full-time employee. investors with the lowest amount of income were more likely to invest in cryptocurrency than those with the highest income ($150,000 or more). investors with the lowest amount of investment assets (less than $5,000) were more likely to invest in cryptocurrency than those with the investment assets of $5,000–$24,999. 5. discussion and implications while investment in cryptocurrency has been rapidly growing, the academic research on cryptocurrency investors is still understudied. to fill this gap, this study investigated the associations between investment literacy and cryptocurrency investment of us investors. one of the notable findings is that objective investment literacy is negatively while subjective literacy and overconfidence are positively associated with the likelihood of investing in cryptocurrency. further, we found various socioeconomic characteristics related to cryptocurrency investment. in line with previous research on risky assets and stock investment, we confirmed a salient effect of investment literacy on cryptocurrency investment. given the high volatility and anomalies in the cryptocurrency market, and the short history of the market, such investments do not seem prudent, so our finding that overconfident individuals are more likely to hold them is not surprising. however, given the trend of increasing and considerable trading volume in a market, our finding provides some implications for financial practitioners and educators. many studies have documented low level of financial literacy (e.g., atkinson & messy 2012; lusardi & mitchell 2017) and significant divergence between objective and subjective financial literacy (e.g., kim et al., 2020; robb et al., 2015). use of financial advice or delegation of financial decisions could be possible alternatives to accompany improving one’s investment literacy and assessing their financial literacy level realistically. our study did not test this issue empirically, but the use of financial advice could be a complementary and supplementary source of financial and investment literacy of investors. this study has some limitations to be noted. first, given the cross-sectional design of the nfcs investor survey, this study did not address any causal inferences between investment literacy and investing in cryptocurrency. ideally, analyses that use longitudinal dataset or experimental study are needed to address the causal inference. second, the timeframe of cryptocurrency investment is not specified in the nfcs survey, for example, “have you invested . . .” also, the survey question does not ask specific type of cryptocurrency available in a marketplace. despite these limitations, this study provides a source of understanding the divergence in investment literacy and its association with cryptocurrency investment. k. t. kim et al. / financial services review 31 (2023) 121–132 129 appendix sample characteristics variables percentage age, mean (sd) 50.8 (17.6) gender male 56.7 female 43.3 marital status (ref: married) married 58.4 single 25.4 separated/divorce/widow 16.2 presence of dependent child 31.3 race/ethnicity white 76.9 black 8.3 hispanic 6.9 asian/others 7.9 education high school diploma or lower 20.4 some college 24.1 associate degree 10.8 bachelor degree 26.6 post-bachelor degree 18.1 employment status full-time employee 42.9 self-employed 9.0 part-time employee 7.7 homemaker 3.8 student 2.5 disabled 1.3 unemployed 1.9 retired 30.9 income less than $35,000 15.6 $35,000–$49,999 12.3 $50,000–$74,999 23.9 $75,000–$99,999 18.9 $100,000–$149,999 18.3 $150,000 or more 11.0 investment assets less than $5,000 12.4 $5,000–$24,999 15.1 $25,000–$49,999 10.8 $50,000–$99,999 14.9 $100,000–$249,999 17.0 $250,000–$499,999 14.2 $500,000–$999,999 8.3 $1,000,000 or more 7.3 note. weighted results. 130 k. t. kim et al. / financial services review 31 (2023) 121–132 references abreu, m., & mendes, v. (2012). information, overconfidence and trading: do the sources of information matter? journal of economic psychology, 33, 868–881. https://doi.org/10.1016/j.joep.2012.04.003 ante, l., fiedler, i., meduna, m. v., & steinmetz, f. (2022). individual cryptocurrency investors: evidence from a population survey. international journal of innovation and technology management, 19. https://doi.org/ 10.1142/s0219877022500080 arias-oliva, m., pelegrı́n-borondo, j., & matı́as-clavero, g. (2019). variables influencing cryptocurrency use: a technology acceptance model in spain. frontiers in psychology, 10, 475. https://doi.org/10.3389/ fpsyg.2019.00475 atkinson, a., & messy, f. (2012). measuring financial literacy: results of the oecd/international network on financial education (infe) pilot study. oecd working papers on finance, insurance and private pensions, no. 15. oecd publishing. https://doi.org/10.1787/5k9csfs90fr4-en avital, m., beck, r., king, j. l., rossi, m., & teigland, r. (2016). jumping on the blockchain bandwagon: lessons of the past and outlook to the future. in proceedings of the thirty seventh international conference on information systems, dublin (icis), (pp. 1–6). http://aisel.aisnet.org/cgi/viewcontent.cgi?article=1320& context=icis2016 barber, b. m., & odean, t. (2000). trading is hazardous to your wealth: the common stock investment performance of individual investors. the journal of finance, 55, 773–806. https://doi.org/10.1111/0022-1082.00226 bharadwaj, s., & deka, s. (2021). behavioural intention towards investment in cryptocurrency: an integration of rogers’ diffusion of innovation theory and the technology acceptance model. forum scientiae oeconomia, 9, 137–159. bouri, e., gupta, r., & roubaud, d. (2019). herding behaviour in cryptocurrencies. finance research letters, 29, 216–221. https://doi.org/10.1016/j.frl.2018.07.008 chu, z., wang, z., xiao, j. j., & zhang, w. (2017). financial literacy, portfolio choice and financial well-being. social indicators research, 132, 799–820. https://doi.org/10.1007/s11205-016-1309-2 chuen, d. l. k., guo, l., & wang, y. (2017). cryptocurrency: a new investment opportunity? the journal of alternative investments, 20, 16–40. https://doi.org/10.3905/jai.2018.20.3.016 coin market cap. (2020, march 2). cryptocurrency market capitalization. available at https://coinmarketcap. com/ da gama silva, p. v. j., klotzle, m. c., pinto, a. c. f., & gomes, l. l. (2019). herding behavior and contagion in the cryptocurrency market. journal of behavioral and experimental finance, 22, 41–50. https://doi.org/ 10.1016/j.jbef.2019.01.006 de bondt, w. f., muradoglu, y. g., shefrin, h., & staikouras, s. k. (2008). behavioral finance: quo vadis? journal of applied finance, 18, 7–21. goetzmann, w. n., & kumar, a. (2008). equity portfolio diversification. review of finance, 12, 433–463. https://doi.org/10.1093/rof/rfn005 guiso, l., & jappelli, t. (2008).financial literacy and portfolio diversification. european university institute working paper eco 2008/31. available at: http://cadmus.eui.eu/handle/1814/9811. gupta, s., gupta, s., mathew, m., & sama, h. r. (2021). prioritizing intentions behind investment in cryptocurrency: a fuzzy analytical framework. journal of economic studies, 48, 1442–1459. https://doi.org/10.1108/ jes-06-2020-0285 hasso, t., pelster, m., & breitmayer, b. (2019). who trades cryptocurrencies, how do they trade it, and how do they perform? evidence from brokerage accounts. journal of behavioral and experimental finance, 23, 64– 74. https://doi.org/10.1016/j.jbef.2019.04.009 henry, c. s., huynh, k. p., & nicholls, g. (2018). bitcoin awareness and usage in canada. journal of digital banking, 2, 311–337. ibrahim, m. (2009). theory of bounded rationality. public management, 91, 3–5. kim, k. t., lee, j., & hanna, s. d. (2020). the effects of financial literacy overconfidence on the mortgage delinquency of us households. journal of consumer affairs, 54, 517–540. https://doi.org/10.1111/joca.12287 k. t. kim et al. / financial services review 31 (2023) 121–132 131 kimball, m., & shumway, t. (2006). investor sophistication, and the participation, home bias, diversification, and employer stock puzzles, mimeo (unpublished manuscript), ann arbor, mi: university of michigan. liang, j., li, l., chen, w., & zeng, d. (2019). towards an understanding of cryptocurrency: a comparative analysis of cryptocurrency, foreign exchange, and stock. in 2019 ieee international conference on intelligence and security informatics (isi) (pp. 137–139). ieee. https://doi.org/10.1109/isi.2019.8823373 lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the role of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. https://doi.org/10.1016/j.jmoneco.2006.12.001 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. https://doi.org/10.1257/jel.52.1.5 lusardi, a., & mitchell, o. s. (2017). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness. quarterly journal of finance, 07, 1750008–1750031. https:// doi.org/10.1142/s2010139217500082 nakamoto, s. (2008). bitcoin: a peer-to-peer electronic cash system. available at https://bitcoin.org/bitcoin.pdf robb, c. a., babiarz, p., woodyard, a., & seay, m. c. (2015). bounded rationality and use of alternative financial services. journal of consumer affairs, 49, 407–435. https://doi.org/10.1111/joca.12071 seay, m. c., preece, g. l., & le, v. c. (2017). financial literacy and the use of interest-only mortgages. journal of financial counseling and planning, 28, 168–180. https://doi.org/10.1891/1052-3073.28.2.168 shin, s., kim, k. t., & seay, m. (2020). source of information and selection of financial instruments. journal of economic psychology, 76, 102212. https://doi.org/10.1016/j.joep.2019.102212 simon, h. a. (2000). bounded rationality in social science: today and tomorrow. mind & society, 1, 25–39. https://doi.org/10.1007/bf02512227 siriopoulos, c. (2021). financial markets are not efficient: financial literacy as an effective risk management tool. international journal of business and management research, 9, 65–73. https://doi.org/10.37391/ijbmr.090110 sorkin, a. r. (2021, march 10). it’s not just bitcoin that’s huge. so is its carbon footprint. new york times. tasca, p., hayes, a., & liu, s. (2018). the evolution of the bitcoin economy: extracting and analyzing the network of payment relationships. the journal of risk finance, 19, 94–126. https://doi.org/10.1108/jrf-032017-0059 titov, v., uandykova, m., litvishko, o., kalmykova, t., prosekov, s., & senjyu, t. (2021). cryptocurrency open innovation payment system: comparative analysis of existing cryptocurrencies. journal of open innovation: technology, market, and complexity, 7, 102. https://doi.org/10.3390/joitmc7010102 van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. https://doi.org/10.1016/j.jfineco.2011.03.006 xia, t., wang, z., & li, k. (2014). financial literacy overconfidence and stock market participation. social indicators research, 119, 1233–1245. https://doi.org/10.1007/s11205-013-0555-9 zhao, w. (2021, march 5). jpmorgan tells private wealth clients that bitcoin can be a portfolio diversifier ’if sized correctly’. available at https://www.theblockcrypto.com/post/97257/jp-morgan-bitcoin-deck-privateclient zhao, h., & zhang, l. (2021). financial literacy or investment experience: which is more influential in cryptocurrency investment? international journal of bank marketing, 39, 1208–1226. https://doi.org/10.1108/ijbm-112020-0552 132 k. t. kim et al. / financial services review 31 (2023) 121–132 assessing the relative need and demand for financial education programs: a case study of graduate students brent j. davisa,*, david p. richardsona, jason s. seligmanb atiaa institute, 8500 andrew carnegie boulevard, charlotte, nc 28262, usa binvestment company institute, 1401 h street nw, washington, dc 20005, usa abstract we measure the relative need and demand for financial education programs among graduate students at a large university system. we find that self-assessed and measured financial literacy is significantly related to interest and demand for financial education. individuals who self-report a high level of financial literacy but have low measured financial literacy are significantly less likely to be interested in financial education, while the opposite is true for financially literate individuals selfreporting a low level of financial literacy. our study adds to research showing the importance of both believed and actual financial literacy measures and has implications for financial education programs. © 2023 academy of financial services. all rights reserved. jel classifications: d14; g53 keywords: financial education; financial literacy; overconfidence 1. introduction over the last several decades, there has been a substantial increase in the investment, health, and longevity risk burdens borne by individuals. this shift has increased focus on the importance of household financial literacy in managing these risks and enhancing financial well-being. as noted by lusardi and mitchell (2014), a lack of financial literacy can have long-term impacts on households’ financial well-being. for example, not properly managing *corresponding author. tel.: +1-704-988-6271. e-mail address: brent.davis@tiaa.org (b. j. davis) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 133–150 debt or planning for retirement can have substantial impacts on financial well-being over the lifecycle. while recent research has documented the link between low financial literacy and financial fragility, many adults continue to have poor understanding of basic personal finance concepts.1 however, effectively engaging individuals in improving their financial literacy has proven challenging. in this paper, we conduct a case study on the need and demand for financial education programming by graduate students in a large university system. leveraging the differences between self-assessed and actual financial literacy levels, we find that students most in need of additional financial education are least likely to take advantage of programming. graduate students are good candidates to receive financial education programming. they are among the most receptive consumers of education, command commensurately higher lifetime earnings postgraduation, and may be more likely to have complex financial planning needs over the lifecycle. many also have prior workforce experience, which may expose them to employer-sponsored benefits and retirement savings plans. however, this prior workforce participation may make graduate students more anxious about their financial security due to low earnings and savings and likely increased debt, during graduate studies. providing financial education programming during this time can help students gain confidence in their long-term financial well-being. our interest in studying financial education programming is whether it is economically rational for individuals to invest in financial education, and relatedly, can financial education programs improve outcomes. regarding the latter, evidence increasingly suggests that financial education affects outcomes across a variety of contexts.2 with respect to financial education investment, lusardi, michaud, and mitchell (2017) discuss that investment in financial literacy should be a function of the associated expected benefits. they find that low-education low-lifetime income groups might rationally choose not to invest in obtaining financial literacy skills because of their larger relative reliance on public social welfare programs. because our target sample is of high education and higher-than-median expected lifetime income, they should rationally choose to invest in financial literacy education. we measure self-assessed and tested financial literacy for graduate students in a large public university system. similar to past studies, we find measurable weakness in the populations’ understanding of basic personal finance concepts, despite most students expressing concern regarding their own personal finances. regarding financial education engagement, we document two main findings. the first is that interest in financial education relates significantly to an individual’s self-assessed level of financial knowledge relative to their actual measured financial literacy level. individuals who estimate they have a high level of financial literacy but perform poorly on a financial literacy quiz are significantly less likely to be interested in financial education, while the opposite is true for financially literate students who self-report a low level of financial acumen. we also find that while receptivity to financial education programming increases significantly with financial literacy levels, overall engagement in financial education is low, despite nearly half of survey respondents signaling interest in programming. the students most likely to indicate interest in financial education are the ones with higher financial literacy and lower overconfidence in their financial literacy. our work adds to the small body of existing research, namely allgood and walstad 134 b. j. davis et al. / financial services review 31 (2023) 133–150 (2016) and anderson et al. (2017), that perceived financial literacy and actual financial literacy are dually important. our results indicate that employers using financial education programming need to adopt innovative engagement strategies to improve the financial literacy and well-being of those employees who need it most. 2. research design and methodology our educational protocol was developed using an early career workplace education offering from a large financial services organization.3 we targeted the financial education offering to the entire pool of graduate students with the following research protocol. at the beginning of the academic year, and ahead of inviting the graduate student population to take our initial survey, we developed a series of prompts. these consisted of 5 ! 7 inch postcards announcing the project with lighthearted financial literacy questions on a front side and the correct answer along with a description of what to expect next on the back.4 we developed three distinct cards, with the goal that students might compare the particular card they received to others and discuss them. regarding signals of validity, integrity, and quality of the effort, the cards prominently featured the collaborating university and were placed in orientation packets for incoming students by the university. the university also placed cards at each department’s reception desk. a research assistant (ra) was assigned to attend graduate student council meetings and other graduate student group meetings of various types across the university system. in these student meetings, the ra was granted five-minute slots to discuss the project and the potential benefits of participation and left cards at each meeting.5 two weeks after the cards were distributed initial invitations for the online survey were sent to all graduate students (17,819) in a large public university system. these invitations were designed to resonate with the information on the cards we had just distributed. the survey was designed to record information on students’ individual and educational characteristics, financial aspirations, personal financial concerns, self-assessed financial acumen, and a financial literacy quiz.6 once the survey was completed, we provided quiz scores to students and offered them the correct answers to missed questions. we then asked whether a respondent was interested in taking part in a financial education seminar or webinar in the near future. our survey was open for one month between mid-september and mid-october, ahead of midterm examinations. over this four-week period the student received an initial invitation and up to four reminders targeted to students who had not taken the survey nor opted out of email engagement. initial and reminder response survey engagement rates are shown in fig. 1. our email prompts to engage the survey were successful with approximately 60% of the survey sample engaged the survey following a reminder as seen with the spikes in the response rate, supporting dechausay et al.’s (2015) result that reminders can improve engagement. from the initial invite population, 2,487 students (14%) engaged the survey. to set up the second stage of our study we invited a matched, random subsample of our surveyed students to a financial education seminar or webinar, whichever they preferred. we randomly selected 1,632 students to match to invited and noninvited groups. of the 1,632 students eligible for invitation, roughly two-thirds (1,101) were invited to participate.7 our invited b. j. davis et al. / financial services review 31 (2023) 133–150 135 student group was designed to be a balanced representation of participants across several dimensions: gender, degree of concern regarding financial matters, score on the financial literacy quiz, and whether or not the student was in a quantitative field.8 invitations were also balanced across those who did or did not initially indicate interest in the program. this allowed students to change their mind if they later decided they wanted to attend. the invitation email contained a link for those wishing to sign up and clicking the link brought the student to a standard web-based submission form. in the invitation to participate, we offered two mid-day and two early evening times for either a seminar or webinar at each campus in the university system. we also offered lunch or dinner to seminar participants; something our pilot run in the previous year had revealed as being important. 3. results in this section we present results from our survey and financial education engagement protocol. we begin by examining the demographic, educational field, and financial literacy characteristics of students who engaged the survey. section 3.2 examines how financial literacy (both self-assessed and actual) relate to interest in attending a financial education seminar. in section 3.3, we use regression analyses to examine correlates of interest in financial education interest and engagement. 3.1. survey sample characteristics table 1 presents characteristics of survey respondents. the average student who participated in the survey is nearly 30 years old and has almost 10 graduate course credits. women (57%) were more likely to participate compared to men, approximately a quarter are fig. 1. survey response rates. 136 b. j. davis et al. / financial services review 31 (2023) 133–150 international students, 57% of the graduate students attend the flagship campus, and nearly one in four graduate students work as research or teaching assistants. half of the students in the sample are pursuing a master’s degree, over a quarter a doctorate, and other degree types represent less than 10% of the survey respondents. about two-thirds of survey participants have prior work experience. we categorize a student’s program or major into four mutually exclusive groups: liberal arts (e.g., liberal arts, humanities, language, music, and social sciences), stem (e.g., science, engineering, medicine, mathematics, and technology), professional (e.g., public health, public administration, education, nursing, law, and other (pre)professional programs), and business (e.g., economics, finance, business, and accounting). enrollment in a professional program represents the plurality of students surveyed (37%) followed by enrollment in stem (27%), liberal arts (20%), and business (14%) programs. table 2 displays financial education engagement numbers. of the nearly 18,000 students we sent a survey, about 14% (2,487) responded. of the representative 1,101 students to whom we sent invitations for a financial education session, 16% (176) accepted. among the accepted group, 36% (64) attended a session with 44 opting for an in-person seminar and 20 for an online webinar. acceptance rates for students who initially indicated interest (25%) were significantly higher than those who did not signal interest (8%). program attendance rates for those initially indicating interest were also higher (39% vs. 29%) but the difference is not significant. we next tabulate personal financial characteristics and engagement across program type and survey variables. table 3 panel a displays measurements of financial literacy, personal financial concerns, and education engagement for all surveyed students by program type. financial literacy quiz score is the average number of questions students answered correctly on our financial quiz (out of 12).9 financial iq is the average of students self-assessed rating of how high their financial knowledge or iq is, ranging from 1 (very low) to (7 very high). relative finiq is an individual’s relative financial knowledge calculated as the table 1 characteristics of targeted graduate student population student survey sample summary statistics mean/ proportion standard deviation obs proportion obs student characteristics degree type age 29.5 7.5 2,487 masters 0.50 2,487 credits taken 9.8 5.1 2,487 certificate 0.02 2,487 married 0.37 2,306 law 0.07 2,487 female 0.57 2,487 doctorate 0.27 2,487 international student 0.26 2,487 medical 0.08 2,487 research/teaching asst 0.24 2,487 post-doc 0.01 2,487 flagship campus 0.57 2,487 other/non-degree 0.05 2,487 prior work experience 0.68 2,288 education program type liberal arts 0.20 2,487 stem 0.27 2,487 professional 0.38 2,487 business and economics 0.14 2,487 b. j. davis et al. / financial services review 31 (2023) 133–150 137 relative difference between performance on the financial literacy quiz and one’s selfassessed financial iq level and normalized on a scale of 0 to 1. values below 0.5 represent overconfidence and values above 0.5 represent underconfidence. for example, a value of 0 indicates complete overconfidence in the self-assessment compared to their actual measured financial literacy knowledge and corresponds to a student answering 0 out of 12 financial literacy questions correctly and indicating a very high level (7) of financial iq on the selfassessment, whereas a value of 1 indicates complete underconfidence. a value of 0.5 indicates neither overnor underconfidence. our personal finance concern metric is measured on an intensity scale from 1 (no concern) to 5 (great concern) across five areas (career goals, current finances, future finances, owning a home, and retirement), with a possible range from a low of 5 to a high of 25, which we normalized from 0 to 1. for engagement metrics, we list the percentage of students who responded to the survey, indicated that they were interested in financial education, accepting an invitation for an educational session, and attended a seminar or webinar (conditional on accepting an invitation). table 3 shows the average student got 65% of the financial literacy questions correct, with business students scoring significantly higher than any of the other program groups.10 the mean self-reported financial iq was 4.7 out of 7, again with business students indicating a significantly higher level of self-reported financial knowledge compared to nonbusiness table 3 financial literacy, financial concern, and financial education engagement characteristic overall program type liberal arts stem business professional financial literacy and concern financial literacy quiz score (0–12) 7.81 7.57 7.66 8.71 7.71 financial iq, self-assessed (1–7) 4.67 4.58 4.42 5.28 4.73 relative finiq (0–1) 0.52 0.52 0.53 0.51 0.51 personal finance concern (0–1) 0.72 0.73 0.70 0.71 0.73 education engagement responded to survey 14% 14% 14% 14% 14% interested in financial education 48 47 48 46 50 accept invite j invited 16 17 15 13 17 attend financial education conditional on accepting invite 36 51 27 42 33 note. means or percents reported. table 2 financial education engagement results engagement number number invited proportion survey respondents 2,487 17,819 0.14 eligible for invite 1,101 1,632 0.67 interested in fin ed j eligible 511 1,101 0.46 accepted j invited 176 1,101 0.16 attended j accepted invite 64 176 0.36 conditional on indicating interest accepted invite 383 511 0.25 attended j accepting invite 50 128 0.39 138 b. j. davis et al. / financial services review 31 (2023) 133–150 students (5.28 vs. 4.60 for nonbusiness students). the average student was slightly underconfident in their financial knowledge. there are limited differences in the relative measure by program type; however, stem students display significantly greater underconfidence compared with the rest of the surveyed population. we find no significant differences by program for students who responded to the survey or for those who indicated interested in attending a financial education seminar or for acceptance rates. for seminar attendance, however, we find a significantly greater proportion of liberal arts students attended a seminar compared to students in other programs. table 4 shows asset and debt characteristics of the surveyed students. we hypothesize that greater participation in financial services and the incidence of debt would be positively correlated to signaling interest in financial education. and having life insurance or an investment account may signal greater interest in financial planning over the life cycle. we include the incidence of students with a checking account, savings account, investment account, or a life insurance policy. nine of 10 students surveyed have a checking account, with three in four having a savings account. the third row list the proportion of students with an investment account, which we define as having a brokerage account, an ira, or an employersponsored retirement savings plan. this is owned by a minority of students (38%), but varies significantly by program type, ranging from 22% for stem students to 50% for business students. we posit debt should be positively correlated with financial education interest. graduate students with student loan debt need to manage loan repayments in conjunction with other consumption, savings, and investment goals and in context of their postgraduate career outlook. over half of the sample has some form of student loan debt, and this varies significantly by a student’s major field. professional students were significantly more likely to have any debt, both debt from graduate and undergraduate studies and from credit cards. table 4 other student characteristics, assets and debt characteristic overall program type liberal arts stem business professional banking, assets, insurance checking account 90% 93% 88% 88% 90% savings account 74 78 71 72 75 investment account 38 38 22 50 44 life insurance 34 29 23 38 42 student debt only undergraduate 9 12 9 5 8 only graduate 19 16 22 26 17 both undergrad and grad 28 30 22 15 36 any student debt 56 57 53 46 61 other debt credit card 31 34 23 26 37 auto 20 16 15 23 24 mortgage 20 20 11 24 25 home ownership home owner 22 21 12 27 28 plan to purchase home 46 39 55 50 43 b. j. davis et al. / financial services review 31 (2023) 133–150 139 credit card debt may demonstrate greater need for financial education since it relates to household balance sheets and not to human capital acquisition or indicates the reliance on high-cost credit card debt to finance education expenses. our final category is home ownership. about one in five already own a home and roughly 50% of students are planning to purchase one in the next 10 years, which varies significantly by a student’s major field. we later control for this in our regression analysis because purchasing a home involves a substantial amount of financial planning, and we expect those planning to purchase a home to have greater interest in financial education. 3.2. interest in financial education in this section, we explore the relationship between self-assessed and measured financial literacy and graduate student initial interest in attending a financial education seminar. nearly half (48%) of the students who participated in the survey indicated they were interested in financial education. in table 5, we find differences for indicated interest across nearly all individual characteristics. while individuals who do not interact with financial institutions may gain marginally greater benefit from financial education, we find that those without checking and savings accounts, or life insurance are significantly less interested in financial education. however, on the liability side of household balance sheets, students with student loan debt or credit card debt were more likely to be interested in the educational offerings. table 5 student characteristics by financial education interest survey characteristics interested not interested sig financial literacy and concern financial literacy quiz score 8.31 7.19 *** financial iq (self-assessed) 4.59 4.82 *** relative finiq 0.55 0.48 *** personal finance concern 0.75 0.69 *** banking, assets, insurance checking account 98% 81% *** savings account 83 65 *** investment account 42 34 *** life insurance 37 31 *** student debt only undergraduate 11% 7% *** only graduate 21 18 * both undergrad and grad 32 24 *** any student debt 64 48 *** other debt credit card 37% 26% *** auto loan 23 17 *** mortgage 20 20 home ownership own home 21% 23% plan to own home 55 39 *** note. means or percents shown. * and *** indicates differences are significant at the 10% and 1% levels, respectively. 140 b. j. davis et al. / financial services review 31 (2023) 133–150 graduate students signaling interest in financial education have significantly higher financial literacy, with interested students averaging about 1.2 more correct answers compared to non-interested students. further, non-interested students are significantly more overconfident in their financial knowledge. this is also shown by the distribution of relative finiq shown in fig. 2, which displays kernel density estimates of relative finiq by indicated interest in financial education. the two distributions are significantly different (p < .01), with noninterested students having a greater estimated density in the overconfidence range (values below 0.5). this result is economically meaningful as the mean difference of 0.07 in the relative measure between interested and non-interested students equates to 1.68 fewer questions answered correctly (out of 12) on our financial quiz for the relatively more overconfident student given a fixed self-assessed score. alternatively, a 0.07 difference results in a 0.84 greater self-assessed score, given a fixed quiz score, for the relatively more overconfident student. overconfident students with low measured financial literacy would arguably benefit the most from financial education. unfortunately, table 6 and fig. 3 indicates that these students were least interested in improving their financial knowledge. in table 6, we examine the percentage of students interested in financial education by their self-assessed financial iq rating and how well they did on the financial literacy quiz, grouping the latter into four categories. the data suggest there is a strong negative correlation between overconfidence and interest in financial education.11 this finding is highlighted by fig. 3, which displays a wireframe surface plot on the data points shown in table 6. financial literacy and selfassessed financial iq are shown on the x and y axes, and the percentage of students who indicated they are interested in financial education is shown on the z (vertical) axis. here the relationship is clear: students who have high financial literacy but self-assess a low level of financial iq are the most interested in financial education (often over 70%). comparatively, students with low financial literacy but self-assess a high level of financial iq are less interested (generally less than 30% of the time). fig. 2. kernel density estimates of relative finiq by indicated interest in financial education. b. j. davis et al. / financial services review 31 (2023) 133–150 141 to understand this relationship over all possible values, fig. 4 displays predicted probabilities of financial education interest. the predicted probabilities are generated using a simple logit model estimating the likelihood that students indicated interest in financial education regressed on a student’s finiq and financial literacy quiz score, treating the two exogenous variables as categorical variables.12 the full profile of predicted probabilities is shown in appendix table a.1. fig. 4 replicates and smooths the relationship shown in fig. 3. generally, the predicted probability that a student signals interest increases significantly in one’s measured financial literacy but significantly decreases as a student’s selfassessment increases. students with a combination of high self-assessed financial iq and low actual financial literacy are predicted to be the least likely to signal interest in financial education. for example, a student answering all questions correctly but self-assesses the lowest level of financial knowledge is predicted to signal interest with a probability above 60%, which decreases to 45% for a student self-assessing the highest level of financial knowledge. by contrast, a student answering no questions correctly on the financial literacy table 6 financial education interest by measured and self-assessed financial literacy percent interested in financial education self-assessed percent of financial quiz questions correct financial iq 0–25% 26–50% 51–75% 76–100% mean n 1 (very low) 50% 80% 50% 0% 55% 20 2 29% 65% 72% 82% 63% 80 3 26% 61% 68% 70% 62% 236 4 20% 49% 61% 58% 56% 646 5 22% 30% 64% 59% 57% 481 6 6% 30% 57% 55% 51% 484 7 (very high) 22% 0% 50% 45% 42% 153 mean 20% 55% 62% 39% 48% n 221 257 974 1,035 2,100 fig. 3. surface plot of financial education interest by measured and self-assessed financial literacy. 142 b. j. davis et al. / financial services review 31 (2023) 133–150 quiz but is most confident in their financial knowledge, is predicted to signal interest with a likelihood of under 6%, increasing to only 11% for a student self-assessing the lowest level of financial knowledge. program sponsors need to be innovative when thinking about how to engage the population of students who would most benefit from financial education. 3.3. regression analysis of financial education interest and engagement we begin our regression analysis with first considering who takes the survey. table 7 uses ordinary probit regression estimating whether a student engages in the survey or not using student and program characteristics, displaying marginal effects on the coefficients and standard errors in parenthesis. because those who do not take the survey do not offer us data on their financial literacy, we only leverage the university system’s administrative data. women and student workers are significantly more likely to engage the survey. while there is no significant relationship for graduate students at the flagship campus, student workers at the flagship campus are significantly less likely to engage the survey, indicated by the interaction term in model 2. these students may be more time constrained with their studies and work duties than their counterparts. we find no significant effect for international students. students who have taken more graduate credits are significantly less likely to take the survey, but the marginal effect is small and has significant attenuation. there were no significant differences in a student’s major subject area. when examining degree type, law students were significantly less likely and doctorate students only marginally less likely to engage in the survey, highlighting the possibility of time constraints for students in terminal degree programs. table 8 estimates the likelihood that a student indicates interest in financial education using ordinary probit specifications showing marginal effects and standard errors in parenthesis. we display five specifications: model 1 uses administrative data plus financial literacy characteristics, and models 2 and 3 add asset and debt characteristics. model 4 adds fig. 4. predicted probabilities of education interest by measured and self-assessed financial literacy. b. j. davis et al. / financial services review 31 (2023) 133–150 143 t ab le 7 r eg re ss io n es ti m at es o f su rv ey en g ag em en t (1 ) (2 ) (3 ) (4 ) m ar . co ef f. s ta n d ar d er ro r m ar . co ef f. s ta n da rd er ro r m ar . co ef f. s ta n d ar d er ro r m ar . co ef f. s ta n d ar d er ro r s tu de n t ch ar ac te ri st ic s w o m an 0 .0 3 0* * * (0 .0 0 5 ) 0 .0 3 0 * * * (0 .0 0 5 ) 0 .0 3 0 * * * (0 .0 0 5) 0 .0 3 0 * * * (0 .0 0 5 ) a g e " 0 .0 0 0 (0 .0 0 0 ) 0 .0 0 0 (0 .0 0 0 ) " 0 .0 0 0 (0 .0 0 0) 0 .0 0 0 (0 .0 0 0 ) in tl st u d en t 0 .0 0 9 (0 .0 0 6 ) 0 .0 0 7 (0 .0 0 6 ) 0 .0 0 8 (0 .0 0 6) 0 .0 0 6 (0 .0 0 6 ) m ai n ca m p u s 0 .0 0 2 (0 .0 0 6 ) 0 .0 0 9 (0 .0 0 6 ) 0 .0 0 7 (0 .0 0 6) 0 .0 0 8 (0 .0 0 6 ) s tu d en t te ac h in g /r es ea rc h as st 0 .0 3 7* * * (0 .0 0 7 ) 0 .0 7 5 * * * (0 .0 1 4 ) 0 .0 7 7 * * * (0 .0 1 4) 0 .0 7 3 * * * (0 .0 1 4 ) e d u c cr ed it s " 0 .0 0 6* * * (0 .0 0 1 ) " 0 .0 0 6 * * * (0 .0 0 1 ) " 0 .0 0 6 * * * (0 .0 0 1) " 0 .0 0 5 * * * (0 .0 0 2 ) ln (e d u c cr ed it s) 0 .0 6 1* * * (0 .0 1 1 ) 0 .0 6 1 * * * (0 .0 1 1 ) 0 .0 6 0 * * * (0 .0 1 2) 0 .0 5 6 * * * (0 .0 1 2 ) m ai n c am p u s * s tu d en t a ss t " 0 .0 5 0 * * * (0 .0 1 6 ) " 0 .0 4 9 * * * (0 .0 1 6) " 0 .0 4 6 * * * (0 .0 1 6 ) p ro g ra m ty p e (b as el in e ¼ l ib er al a rt s) s t e m " 0 .0 0 2 (0 .0 0 9) p ro fe ss io n al 0 .0 0 3 (0 .0 0 8) b u si n es s 0 .0 0 7 (0 .0 1 0) p ro g ra m ty p e (b as el in e ¼ m as te rs ) d o ct o ra te " 0 .0 1 2 * (0 .0 0 7 ) l aw " 0 .0 3 3 * * * (0 .0 0 9 ) o th er /n o n -d eg re e " 0 .0 0 0 (0 .0 1 3 ) c er ti fi ca te " 0 .0 2 7 (0 .0 1 8 ) p o st -d o c " 0 .0 1 7 (0 .0 3 2 ) m ed ic al " 0 .0 1 6 (0 .0 1 1 ) o b se rv at io n s 1 7 ,7 5 0 1 7 ,7 5 0 1 7 ,7 5 0 1 7 ,7 50 n o te . p ro b it sp ec ifi ca ti o n s w it h m ar g in al co ef fi ci en ts re p o rt s an d st an d ar d er ro rs in p ar en th es is . * , * * , * * * re p re se nt s si g n ifi ca n ce at th e 1 0 % , 5 % , an d 1 % le v el s, re sp ec ti v el y . 144 b. j. davis et al. / financial services review 31 (2023) 133–150 t ab le 8 r eg re ss io n es ti m at es o f se m in ar in te re st (1 ) (2 ) (3 ) (4 ) (5 ) s tu de n t an d p ro gr am ch ar ac te ri st ic s w o m an 0 .0 2 3 (0 .0 2 2) 0 .0 2 1 (0 .0 2 3 ) 0 .0 2 1 (0 .0 2 3) 0 .0 1 9 (0 .0 2 2 ) 0 .0 0 1 (0 .0 2 2) a g e " 0 .0 0 4 (0 .0 0 2) * * * " 0 .0 0 4 (0 .0 0 2 )* * " 0 .0 0 3 (0 .0 0 2) * * " 0 .0 0 1 (0 .0 0 2 ) " 0 .0 0 1 (0 .0 0 2) in t’ l st u d en t 0 .0 3 5 (0 .0 2 7) 0 .0 2 9 (0 .0 2 8 ) 0 .0 6 5 (0 .0 3 1) * * 0 .0 5 9 (0 .0 3 1 )* 0 .0 4 2 (0 .0 3 1) m ai n ca m p u s 0 .0 0 8 (0 .0 2 4) 0 .0 0 3 (0 .0 2 4 ) 0 .0 0 9 (0 .0 2 4) 0 .0 0 5 (0 .0 2 4 ) 0 .0 0 4 (0 .0 2 5) r es ea rc h /t ea ch in g as st " 0 .0 0 1 (0 .0 2 6) " 0 .0 0 8 (0 .0 2 6 ) " 0 .0 1 0 (0 .0 2 7) " 0 .0 1 5 (0 .0 2 7 ) " 0 .0 1 7 (0 .0 2 7) c re d it s " 0 .0 0 5 (0 .0 0 2) * * " 0 .0 0 6 (0 .0 0 3 )* * " 0 .0 0 6 (0 .0 0 3) * * " 0 .0 0 6 (0 .0 0 3 )* * " 0 .0 0 6 (0 .0 0 3) * * s t e m 0 .0 4 4 (0 .0 3 3) 0 .0 4 3 (0 .0 3 3 ) 0 .0 4 7 (0 .0 3 3) 0 .0 3 8 (0 .0 3 3 ) 0 .0 3 5 (0 .0 3 4) b u si n es s 0 .0 1 4 (0 .0 3 8) 0 .0 2 1 (0 .0 3 9 ) 0 .0 2 3 (0 .0 3 9) 0 .0 1 8 (0 .0 3 9 ) 0 .0 3 6 (0 .0 3 9) p ro fe ss io n al 0 .0 5 3 (0 .0 2 9) * 0 .0 5 6 (0 .0 2 9 )* 0 .0 5 5 (0 .0 2 9) * 0 .0 5 1 (0 .0 2 9 )* 0 .0 5 4 (0 .0 2 9) * f in an ci al li te ra cy ch ar ac te ri st ic s f in an ci al li te ra cy 0 .0 4 3 (0 .0 0 4) * * * 0 .0 4 3 (0 .0 0 4 )* * * 0 .0 4 4 (0 .0 0 4) * * * 0 .0 4 4 (0 .0 0 4 )* * * s el fas se ss ed fi n an ci al iq " 0 .0 5 2 (0 .0 0 8) * * * " 0 .0 5 1 (0 .0 0 9 )* * * " 0 .0 5 0 (0 .0 0 9) * * * " 0 .0 5 0 (0 .0 0 9 )* * * r el at iv e f in iq 0 .8 5 3 (0 .0 7 2) * * * p er so n al fi na n ce co n ce rn 0 .4 5 8 (0 .0 5 5) * * * 0 .4 6 1 (0 .0 5 5 )* * * 0 .4 5 2 (0 .0 5 5) * * * 0 .4 3 1 (0 .0 5 6 )* * * 0 .4 2 6 (0 .0 5 6) * * * b an k in g, as se ts , in su ra n ce in v es tm en t o r re ti re m en t " 0 .0 1 9 (0 .0 2 6 ) " 0 .0 1 5 (0 .0 2 6) " 0 .0 0 5 (0 .0 2 7 ) 0 .0 1 0 (0 .0 2 6) s av in g s ac co u n t 0 .0 2 5 (0 .0 2 9 ) 0 .0 2 8 (0 .0 2 9) 0 .0 2 8 (0 .0 2 8 ) 0 .0 3 6 (0 .0 2 8) l if e in su ra n ce " 0 .0 2 7 (0 .0 2 5 ) " 0 .0 3 1 (0 .0 2 5) " 0 .0 1 9 (0 .0 2 5 ) " 0 .0 1 8 (0 .0 2 5) d eb t s tu d en t d eb t (u n d er g ra d o n ly ) 0 .1 0 8 (0 .0 3 8) * * * 0 .1 0 1 (0 .0 3 8 )* * * 0 .0 9 9 (0 .0 3 8) * * * s tu d en t d eb t (g ra d o n ly ) 0 .0 2 7 (0 .0 3 0) 0 .0 1 9 (0 .0 3 0 ) 0 .0 1 6 (0 .0 3 0) b o th u n d er g ra d an d g ra d d eb t 0 .0 6 0 (0 .0 2 9) * * 0 .0 5 1 (0 .0 2 9 )* 0 .0 4 3 (0 .0 2 9) c re d it c ar d " 0 .0 5 1 (0 .0 6 3) " 0 .0 5 9 (0 .0 6 3 ) " 0 .0 5 0 (0 .0 6 3) a u to 0 .0 3 2 (0 .0 2 7) 0 .0 4 2 (0 .0 2 7 ) 0 .0 4 1 (0 .0 2 7) h o m e o w n er sh ip o w n h o m e " 0 .0 6 8 (0 .0 3 7 )* " 0 .0 5 7 (0 .0 3 7) p la n to p u rc h as e h o m e 0 .0 5 3 (0 .0 2 7 )* * 0 .0 6 1 (0 .0 2 7) * * o b se rv at io n s 2 ,0 9 8 2 ,0 9 8 2 ,0 9 8 2 ,0 9 8 2 ,0 9 8 n o te . p ro b it w it h m ar g in al ef fe ct s sh o w n . s ta n d ar d er ro rs sh o w n in p ar en th es is . * , * * , * * * in d ic at es si g n ifi ca n ce at th e 1 0 % , 5 % , o r 1 % le v el . b. j. davis et al. / financial services review 31 (2023) 133–150 145 home ownership characteristics, and model 5 uses the relative finiq measure on the full specification instead of quiz score and self-assessed financial iq. beginning with model 1, younger students are significantly more likely to signal interest, but this is not robust in later specifications. professional students are significantly more likely to signal interest; however, this is only significant at the 10% level when controlling for debt characteristics. we find no significant correlation for business students.13 in all specifications, financial literacy quiz scores and self-assessed financial iq have large significant effects on the likelihood to signal interest. these effects pull in the opposite directions, as discussed in section 3.2., with predicted interest increasing in quiz score but decreasing in the self-assessed measure. an additional question answered correctly on the quiz score increases the estimated likelihood to indicate interest by 4%, while a one unit increase in the self-assessed measure decreases the likelihood by 5%. the personal finance degree of concern composite measure is significantly and positively related to indicating interest, following our hypothesis. in model 5, we use relative finiq as a regressor, instead of quiz score and the self-assess measure; and find estimated interest significantly increases (decreases) as underconfidence (overconfidence) increases. adding individual asset and banking in model 2 has no significant impact. model 3 includes student loan debt as a categorical variable (with no student loan debt as the baseline). undergraduate student loan debt has a significant impact on the likelihood to be interested in financial education, compared to those without student loan debt. although we do not find a (robust) significant effect for either graduate debt only or both debt from undergraduate and graduate school, the coefficient is positive—in the hypothesized direction. debt management is likely to become a larger concern for graduate students in the accumulation phase of their lifecycle, especially since this group delays employment income, savings, and loan repayment before (re)entering the labor force, albeit at an expected relatively higher salary. credit card debt was not a significant correlate. this may be of concern because some students may be using high-cost debt to finance part of their education and this group would benefit from financial education. when including home ownership characteristics in model 4 we find those planning to purchase a home in the next ten years are significantly more likely to signal interest, following our hypothesis. 4. discussion this paper examined financial education interest among graduate students in a large public university system. we find a strong positive and significant correlation between underconfidence (overconfidence) in self-measured financial knowledge and (lack of) interest in financial education programming. this finding suggests the need for innovative engagement strategies to identify and provide programming to individuals who would benefit the most from improving their financial literacy. the results speak to several components regarding the timing and delivery of financial education. the first is whether the timing is optimal for graduate students to engage in improving their financial literacy. because many graduate students are close to (re)entering 146 b. j. davis et al. / financial services review 31 (2023) 133–150 the workforce, they may be focused on the near-term issues of graduating, finding a job, or moving. this can make them subject to present bias through the belief that they have scant time to devote additional resources to improving their long-term financial well-being. innovative engagement strategies, such a providing lunch or tchotchkes, may nudge active participation in financial education. however, we find these nudges did not address to our main finding that those confident in their financial knowledge but have low financial literacy are significantly less interested in financial education. how should these individuals be engaged in financial education? mandatory financial education could be one response, and many states have begun to institute mandatory financial literacy programs in high school. stoddard and urban (2020), urban, schmeiser, collins, and brown (2018), and collins (2013) find some benefits to mandatory education. however, there needs to be further research in this area as there remain many open questions, including whether the education benefits persist in later life, how such programs are implemented, what is included in the content, and when in the lifecycle are they delivered, among others. with most programs continuing to rely on voluntary education efforts, often offered by employer benefit programs, designers and implementers of financial education will need to consider how to attract individuals overconfident in their knowledge of personal finances. notes 1 for research on financial literacy see yakoboski, lusardi, and hasler (2019), clark, lusardi, and mitchell (2017), lusardi and mitchell (2014), lusardi, mitchell, and curto (2014), lusardi and mitchell (2011), among others. for a broader discussion on financial literacy, financial education, and economic outcomes see hastings, madrian, and skimmyhorn (2013). 2 bernheim and garrett (2003), lusardi (2004), maki (2004), and bayer, bernheim, and scholz (2009) have studied employer-sponsored financial literacy programs and retirement preparedness. other studies have found other positive benefits to financial education programs (i.e., clark et al., 2006; skimmyhorn et al., 2016; and seligman and bose 2012). 3 in the previous year, we piloted a similar survey and education to a small group of students. while the test group gave good feedback and generally positive reviews, we adjusted both our materials and engagement strategy to improve participation. 4 this postcard campaign is consistent with findings on the positive value of prompts from dechausay, anzelone, and reardon (2015). 5 the ra had previously served as president of the graduate student council and thus was a familiar and respected source of information across the body of groups we engaged. 6 lusardi, mitchell, and curto (2014), schmeiser and seligman (2013), and knoll and houts (2012) have published work evaluating questions on measuring financial literacy using three independent methodologies. we take our financial literacy questions from this work, and consultation with financial counselors, and use a set of 12 questions from these studies. b. j. davis et al. / financial services review 31 (2023) 133–150 147 7 a control group that was roughly one-third of eligible participants was not invited so as to be able to carry forward with other research questions. 8 we defined a student as being in a quantitative field if the student’s program is in economics, business, engineering, statistics, mathematics, physics, chemistry, or computer science. 9 the financial literacy quiz covered questions on interest, inflation, and bond prices. 10 we control for degree type in our regression analysis. 11 our relative measure cannot examine whether there are differences in student interest across the entire cross product of the financial literacy score and self-assessed financial iq measure. 12 the overall model is significant (p < .01 with a x2 test). moreover, the predicted probabilities for each combination of finiq and quiz score (91 combinations) are all significant at the 5% level. 13 in separate regressions (not shown) we controlled for degree type and prior work experience; there were no significant differences. appendix table a1 predicted probabilities of indicating interest in financial education by financial literacy and selfassessed financial iq financial literacy quiz score self-assessed financial iq very low 1 2 3 4 5 6 very high 7 0 (0%) 11.1% 14.3% 13.7% 10.3% 10.5% 7.9% 5.9% 1 30.6 37.1 35.9 28.9 29.3 23.3 18.2 2 33.6 40.3 39.1 31.8 32.2 25.8 20.3 3 35.8 42.6 41.4 33.9 34.4 27.7 21.9 4 52.3 59.4 58.1 50.2 50.7 42.9 35.6 5 61.1 67.7 66.6 59.2 59.7 51.9 44.3 6 (50%) 54.1 61.2 60.0 52.1 52.6 44.8 37.3 7 65.9 72.1 71.1 64.1 64.5 57.1 49.4 8 63.0 69.4 68.3 61.1 61.5 53.9 46.2 9 64.5 70.8 69.8 62.7 63.1 55.6 47.9 10 62.0 68.6 67.5 60.1 60.6 52.9 45.2 11 62.2 68.7 67.6 60.2 60.7 53.0 45.4 12 (100%) 62.3 68.8 67.7 60.3 60.8 53.2 45.5 note. predicted probabilities from logit model as described in section 3.2, all are significant at the 5% level. 148 b. j. davis et al. / financial services review 31 (2023) 133–150 acknowledgments we thank annamaria lusardi and carrie houts for their comments on our financial literacy survey. we thank james winbush, dan rives, keatrick johnson, and others at indiana university for their engagement and contributions. we have benefited from the capable research assistance of benjamin bissette, katherine jones, and lanita rahjina. we acknowledge the financial support of indiana university in this work. any errors are our own. the views are the authors’ and do not necessarily represent the views of the tiaa, the tiaa institute, or the investment company institute. references allgood, s., & walstad, w. (2016). the effects of perceive and actual financial literacy on financial behaviors. economic inquiry, 54, 675–697. available at: https://doi.org/10.1111/ecin.12255 anderson, a., baker, f., & robinson, d. (2017). precautionary savings, retirement planning and misperceptions of financial literacy. journal of financial economics, 126, 383–398. available at: https://doi.org/10.1016/j. jfineco.2017.07.008 bayer, p., bernheim, d., & scholz, j. (2009). the effects of financial education in the workplace: evidence from a survey of employers. economic inquiry, 47, 605–624. available at: https://doi.org/10.1111/j.1465-7295.2008.00156.x bernheim, d., & garrett, d. m. (2003). the effects of financial education in the workplace: evidence from a survey of households. journal of public economics, 87, 1487–1519. available at: https://doi.org/10.1016/s0047-2727 (01)00184-0 clark, r., lusardi, a., & mitchell, o. (2017). employee financial literacy and retirement plan behavior: a case study. economic inquiry, 55, 248–259. available at: https://doi.org/10.1111/ecin.12389 clark, r. l., d’ambrosio, m. b., mcdermed, a. a., & sawant, k. (2006). retirement plans and saving decisions: the role of information and education. journal of pension economics and finance, 5, 45–67. available at: https://doi.org/10.1017/s1474747205002271 collins, m. (2013). the impacts of mandatory financial education: evidence from a randomized field study. journal of economic behavior & organization, 95, 146–158. available at: https://doi.org/10.1016/j.jebo.2012.08.011 dechausay, n., anzelone, c., & reardon, l. (2015). the power of prompts: using behavioral insights to encourage people to participate. opre report 2015-75. washington, dc: office of planning, research and evaluation, administration for children and families, u.s. department of health and human services. hastings, j., madrian, b., & skimmyhorn, w. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5, 347–373. available at: https://doi.org/10.1146/annurev-economics-082312-125807 knoll, m., & houts, c. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. journal of consumer affairs, 46, 381–410. available at: https://doi.org/ 10.1111/j.1745-6606.2012.01241.x lusardi, a. (2004). saving and the effectiveness of financial education. in: o. mitchell & s. utkus (eds.), pension design and structure: new lessons from behavioral finance (pp. 157–184). oxford: oxford university press. lusardi, a., & mitchell, o. (2011). the outlook for financial literacy. in: o. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 1–16). oxford: oxford university press. lusardi, a., & mitchell, o. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. available at: https://doi.org/10.1257/jel.52.1.5 lusardi, a., mitchell, o., & curto, v. (2014). financial literacy and financial sophistication in the older population. journal of pension economics & finance, 13, 347–366. available at: https://doi.org/10.1017/s1474747214000031 lusardi, a., michaud, p.-c., & mitchell, o. (2017). optimal financial knowledge and wealth inequality. the journal of political economy, 125, 431–477. available at: https://doi.org/10.1086/690950 maki, d. (2004). financial education and private pensions. in w. gale, j. shoven, & m. warshawsky (eds.), private pensions and public policies (pp. 126–139). washington, dc: brookings institution press. schmeiser, m., & seligman, j. (2013). using the right yardstick: assessing financial measures by way of financial well-being. journal of consumer affairs, 47, 243–262. available at: https://doi.org/10.1111/joca.12010 b. j. davis et al. / financial services review 31 (2023) 133–150 149 seligman, j., & bose, r. (2012). learning by doing: active employer sponsored retirement savings plan participation and household wealth accumulation. the quarterly review of economics and finance, 52, 162–172. available at: https://doi.org/10.1016/j.qref.2012.02.002 skimmyhorn, w., davies, e., mun, d., & mitchell, b. (2016). assessing financial education methods: principles vs. rules-of-thumb approach. the journal of economic education, 47, 193–210. available at: https://doi.org/ 10.1080/00220485.2016.1179145 stoddard, c., & urban, c. (2020). the effects of state mandated financial education on college financing behaviors. journal of money, credit and banking, 52, 747–776. available at: https://doi.org/10.1111/jmcb.12624 urban, c., schmeiser, m., collins, j., & brown, a. (2018). the effects of high school personal financial education policies on financial behaviors. economics of education review, 78, 101786. yakoboski, p., lusardi, a., & hasler, a. (2019). financial literacy in the united states and its link to financial wellness. the 2019 tiaa institute-gflec personal finance index. charlotte, nc: tiaa institute. 150 b. j. davis et al. / financial services review 31 (2023) 133–150 can financial literacy education reduce the use of medicaid and snap? abdullah al-bahrania,b, darshak patelc, jamie weathersd,* adepartment of economics and finance, northern kentucky university, highland heights, ky 41099, usa buniversity college dublin, belfield, dublin 4, ireland cdepartment of economics, university of kentucky, lexington, ky 40506, usa ddepartment of finance and commercial law, western michigan university, kalamazoo, mi 49008, usa abstract in recent decades, we have seen an increase in both the complexity of financial markets and the expectations of individual responsibility for people’s financial decision-making. policies supporting financial literacy education are promoted as a way to decrease reliance on social safety nets. the assumption is that low levels of financial literacy translate to lower economic outcomes and, thus, increased dependence on social programs. we use the 2018 national financial capabilities study to investigate the possible relationship between high school mandated financial literacy education and social program participation and find no evidence of such a relationship. © 2020 academy of financial services. all rights reserved. jel classifications: a21; g53; i26; i38 keywords: financial literacy; social programs; financial literacy education 1. introduction the relationship between financial literacy and financial outcomes has been a topic of interest in recent decades. studies examining this relationship have focused on financial behaviors such as retirement planning (lusardi, 1999; lusardi & mitchell, 2011b), savings (lusardi, 2008), student loan consumption (stoddard & urban, 2020), and the utilization of high-cost borrowing (harvey, 2019). the evidence consistently indicates that an increase in *corresponding author. tel.: +1-269-387-6056. e-mail address: jamie.weathers@wmich.edu (j. weathers) 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 303–314 financial literacy is associated with an increase in the quality of financial decision-making and, consequently, better financial and economic outcomes. this has led to the presumption that mandating financial literacy education will reduce dependence on social safety net programs. state and federal officials have proposed high school financial literacy education mandates as a policy tool to help increase financial well-being, which encompasses a range of financial behaviors, including potential reliance on social programs. as of 2017, 25 states had mandated some form of financial literacy education in high school (stoddard & urban, 2020) and this focus on providing personal finance instruction continues to evolve. in 2018, 29 states and puerto rico introduced new or modified legislation concerning financial literacy education. 1 in 2019, this number grew to 42 states plus the district of columbia and puerto rico. 2 our objective is to examine the impact of financial literacy education mandates in high school on social program participation. we are able to reduce the selection issues typical in this research because individuals required to participate in financial literacy education do not select into their programs. our methodology relies on the assumption that these individuals are required to participate in high school. this is similar to the approach taken by stoddard and urban (2020), though our approach differs in that we do not rely on state mandates to determine participation in financial literacy education, because high school requirements are possible even when the state does not mandate it. 3 we are able to identify participation in mandated financial literacy education at the individual level. we estimate social program participation rates of those required to participate in financial literacy education, and compare them to those who choose to participate, and to those who do not receive financial literacy education. we find no evidence of a relationship between mandated financial literacy courses and participation in social programs. individuals mandated to participate in financial literacy education are as likely to receive social assistance as those who choose to take a course, and those who do not participate at all. however, we do find that financial literacy levels in the top quintile are less likely to participate in social programs. our control variables are strong predictors of social program participation, and include demographic variables such as age, income, and state of residency. our findings support the results of previous studies examining earned income tax credit participation and financial knowledge (chetty, 2015; chetty, friedman, & saez, 2013). 2. financial education, literacy, and behaviors the shift toward individual responsibility in financial decision-making, coupled with the increased complexity of financial tools, has increased awareness of the role that financial literacy plays in determining optimal financial outcomes. further, an established positive relationship between financial literacy and financial behaviors has prompted increased advocacy for financial literacy education that focuses on personal finance. researchers are evaluating state mandates to measure their efficacy in changing financial behaviors. while the quality of these evaluations can suffer from endogeneity, some do indicate causal links between financial education and outcomes. 304 a. al-bahrani et al. / financial services review 28 (2020) 303–314 using a synthetic control, brown et al. (2014) find increased credit scores in states that mandate financial literacy education. similarly, using a difference-in-difference methodology, stoddard and urban (2020) find that students who graduate from high schools in states with financial literacy mandates, while no more likely to attend college, make better choices when taking out student loans and other low-cost debt. 4 harvey (2019) demonstrates that individuals residing in states that mandate financial literacy education are less likely to use alternative financial services (afs) such as check-cashing, rent-to-own financing, pawn shop services, auto title loans, tax refund anticipation loans, and payday loans. additional evidence indicates that formal financial education results in positive long-term financial behaviors as well (wagner & walstad, 2019). mandated financial literacy programs oblige individuals to receive the education ostensibly necessary to improve financial decision-making. however, it is important to differentiate the effect of financial education on financial literacy from its effect on financial behaviors. similarly, we must acknowledge the heterogeneity of financial education itself, as it can vary in quality, source, length, delivery method, scope, timing, and so forth. kaiser and menkhoff (2017) undertook a meta-analysis of 126 studies, which confirmed a strong positive impact of financial education on financial literacy, a much lesser (but also statistically significant) effect of financial education on financial behaviors, and a positive correlation between its effects on both financial literacy and financial behaviors. thus, the link from financial education to improved financial behavior appears to be mediated by financial literacy. higher returns are expected of financially literate individuals due to superior financial decisions. for example, financial literacy is positively associated with both the ownership of stocks in asset portfolios (christelis, jappelli, & padula, 2010; van rooij, lusardi, & alessie, 2011) and the selection of lower-cost funds (hastings & tejeda-ashton, 2008; hastings & mitchell, 2020; hastings, mitchell, & chyn, 2011). lusardi and tufano (2015) partnered with a market research firm to design a survey, develop their own set of financial literacy questions, and collect data from 1,000 u.s. residents via telephone in 2007. they used data concerning debt from surveys such as the health and retirement study (hrs), the rand american life panel (alp), and the survey of consumers, but no data regarding financial literacy existed at the time of their study. they find that an increase in financial literacy is inversely related to the use of high-cost debt and high fees. they attribute 30% of the fees collected by credit card companies to financial ignorance. additionally, lusardi and mitchell (2011b) find that increased financial literacy is positively correlated with deliberate long-term planning and, consequently, higher retirement wealth. overall, variations in financial literacy explain 30-40% of the inequality in retirement wealth (lusardi, michaud, & mitchell, 2017). although there is consensus on the positive impact of financial education on financial literacy, al-bahrani, weathers, and patel (2019) find variation in returns to formal financial literacy education by race, whereby white individuals exhibit significantly higher financial literacy scores than minorities, all else being equal. beyond potential curriculum bias, the scope and timing of financial education are crucial to its effectiveness. “teachable moments,” or education aimed at altering a specific financial behavior, prove more effective than comprehensive financial education (e.g., miller et al., 2015; zhan, anderson, & scott, a. al-bahrani et al. / financial services review 28 (2020) 303–314 305 t ab le 1 s u m m ar y st at is ti cs f u ll sa m p le s o ci al p ro g ra m p ar ti ci p an ts n o n -p ar ti ci p an ts t st at m ea n s ta n d ar d d ev ia ti o n m ea n s ta n d ar d d ev ia ti o n m ea n s ta n d ar d d ev ia ti o n d if fe re n ce in m ea n s s o ci al p ro g ra m p ar ti ci p at io n 0 .3 4 0 .4 7 1 .0 0 0 .0 0 0 .0 0 0 .0 0 r eq u ir ed co u rs e 0 .0 6 0 .2 4 0 .0 6 0 .2 4 0 .0 6 0 .2 4 0 .1 4 2 o p ti o n al co u rs e 0 .0 3 0 .1 7 0 .0 3 0 .1 7 0 .0 3 0 .1 7 0 .0 1 5 f in an ci al li te ra cy sc o re (% ) 4 6 .8 2 8 .3 4 0 .1 2 6 .9 5 0 .3 2 8 .5 1 6 .1 1 a g e 4 5 .6 1 7 .5 4 3 .4 1 5 .2 4 6 .7 1 8 .4 8 .3 8 6 f in a n ci a l li te ra cy (# ) 0 c o rr ec t 0 .1 2 0 .3 2 0 .1 6 0 .3 7 0 .1 0 0 .3 0 �8 .9 6 3 1 c o rr ec t 0 .1 8 0 .3 8 0 .2 1 0 .4 1 0 .1 6 0 .3 7 �5 .7 2 6 2 c o rr ec t 0 .2 4 0 .4 2 0 .2 6 0 .4 4 0 .2 2 0 .4 2 �3 .2 7 1 3 c o rr ec t 0 .2 4 0 .4 3 0 .2 3 0 .4 2 0 .2 4 0 .4 3 1 .5 9 0 4 c o rr ec t 0 .1 6 0 .3 7 0 .1 1 0 .3 1 0 .1 9 0 .3 9 9 .6 5 3 5 c o rr ec t 0 .0 6 0 .2 4 0 .0 3 0 .1 7 0 .0 8 0 .2 7 9 .2 3 5 r a ce / et h n ic it y w h it e 0 .7 1 0 .4 5 0 .6 8 0 .4 7 0 .7 3 0 .4 5 4 .4 5 7 b la ck 0 .1 2 0 .3 2 0 .1 5 0 .3 5 0 .1 0 0 .3 0 �6 .7 1 2 h is p an ic 0 .1 0 0 .3 0 0 .1 0 0 .3 1 0 .1 0 0 .3 0 �0 .6 1 4 a si an 0 .0 3 0 .1 8 0 .0 2 0 .1 5 0 .0 4 0 .1 9 3 .5 7 8 o th er 0 .0 4 0 .1 9 0 .0 4 0 .2 0 0 .0 4 0 .1 9 �1 .6 4 2 n u m b er o f d ep en d en ts n o ch il d re n 0 .4 0 0 .4 9 0 .3 3 0 .4 7 0 .4 4 0 .5 0 1 0 .5 2 c h il d re n , n o d ep en d en ts 0 .2 9 0 .4 5 0 .2 3 0 .4 2 0 .3 2 0 .4 7 8 .5 9 6 1 c h il d 0 .1 4 0 .3 5 0 .1 8 0 .3 8 0 .1 3 0 .3 3 �6 .7 6 2 2 c h il d re n 0 .0 9 0 .2 9 0 .1 4 0 .3 4 0 .0 7 0 .2 6 �1 0 .1 9 3 c h il d re n 0 .0 4 0 .2 1 0 .0 7 0 .2 6 0 .0 3 0 .1 7 �9 .5 7 3 4 o r m o re ch il d re n 0 .0 3 0 .1 6 0 .0 5 0 .2 2 0 .0 1 0 .1 2 �1 0 .7 0 in co m e b ra ck et s in co m e 0 to 1 5 k 0 .2 4 0 .4 3 0 .4 0 0 .4 9 0 .1 5 0 .3 6 �2 7 .1 3 in co m e 1 5 to 2 5 k 0 .2 2 0 .4 2 0 .2 8 0 .4 5 0 .2 0 0 .4 0 �8 .5 6 3 in co m e 2 5 to 3 5 k 0 .2 3 0 .4 2 0 .1 9 0 .3 9 0 .2 5 0 .4 4 7 .3 0 7 in co m e 3 5 to 5 0 k 0 .3 1 0 .4 6 0 .1 4 0 .3 4 0 .4 0 0 .4 9 2 5 .9 7 e d u ca ti o n le ve l < h ig h sc h o o l 0 .0 5 0 .2 2 0 .0 9 0 .2 9 0 .0 3 0 .1 8 �1 1 .9 4 h ig h sc h o o l 0 .2 7 0 .4 4 0 .2 8 0 .4 5 0 .2 7 0 .4 4 �0 .8 8 1 g e d 0 .1 2 0 .3 2 0 .1 4 0 .3 5 0 .1 0 0 .3 0 �5 .4 7 6 s o m e co ll eg e 0 .3 0 0 .4 6 0 .3 0 0 .4 6 0 .2 9 0 .4 6 �0 .3 2 0 a ss o ci at es 0 .1 0 0 .3 0 0 .0 9 0 .2 9 0 .1 0 0 .3 1 1 .8 0 8 b ac h el o rs 0 .1 3 0 .3 3 0 .0 8 0 .2 7 0 .1 5 0 .3 6 9 .9 4 3 g ra d u at e/ p ro fe ss io n al 0 .0 4 0 .1 9 0 .0 2 0 .1 5 0 .0 5 0 .2 1 5 .6 8 8 e m p lo ym en t s el f em p lo y ed 0 .0 7 0 .2 5 0 .0 7 0 .2 6 0 .0 7 0 .2 5 �0 .8 6 8 f u ll ti m e em p lo y m en t 0 .2 6 0 .4 4 0 .1 4 0 .3 5 0 .3 2 0 .4 7 1 8 .4 9 p ar t ti m e em p lo y m en t 0 .1 2 0 .3 2 0 .1 3 0 .3 3 0 .1 1 0 .3 2 �1 .4 3 4 h o m em ak er 0 .1 0 0 .3 0 0 .1 6 0 .3 6 0 .0 7 0 .2 6 �1 2 .7 2 s tu d en t 0 .0 5 0 .2 1 0 .0 3 0 .1 8 0 .0 6 0 .2 3 5 .1 4 3 d is ab le d 0 .1 0 0 .3 0 0 .2 1 0 .4 1 0 .0 5 0 .2 1 �2 4 .9 2 u n em p lo y ed 0 .0 9 0 .2 8 0 .1 2 0 .3 2 0 .0 7 0 .2 5 �8 .0 0 6 r et ir ed 0 .2 2 0 .4 1 0 .1 4 0 .3 5 0 .2 6 0 .4 4 1 1 .9 9 n 8 ,6 5 7 2 ,9 7 2 5 ,6 8 5 306 a. al-bahrani et al. / financial services review 28 (2020) 303–314 2006). further, kaiser and menkhoff (2017) report less beneficial outcomes from mandated financial education relative to optional financial education, even after controlling for teachable moments. these findings bring into question the equity of results from financial literacy program mandates (i.e., formal financial literacy education). we use non-public use data from the 2018 national financial capability study (nfcs) to examine the relationship between the use of social safety nets, financial literacy, and participation in mandatory financial education. the non-public use data set also provides useful information concerning control variables, including a continuous measure of age, racial/ethnic classification, number of dependents, income, education level, and employment status. 3. data the nfcs survey data includes the “big five” measure of financial literacy, which is a resource frequently used by researchers in this field (e.g., al-bahrani, weathers, & patel, 2019; harvey, 2019; lusardi, 2019). it also includes information as to whether participation in financial literacy education was required, giving us the ability to isolate the effect of mandated financial literacy education from financial literacy education that students have chosen to receive. further, analysis of the high school data allows us to compare three possible participation categories: no exposure to financial literacy education in high school, chosen exposure in high school, and mandatory exposure in high school. the non-public use dataset has 27,091 observations, with 500 respondents per state and oversampling in california, illinois, new york, and texas. because social program participation has an income consideration, we further restrict our sample to those making $50,000 or less per annum, and who are younger than 80 years of age. we also exclude individuals selected for financial education in college or through an employer, to mitigate measurement noise. our final sample comprises 8,657 survey respondents. the social program participants comprise 34% of this sample. 5 table 1 shows the summary statistics of the variables used in our analysis. the historical national average proportion of correct responses to the big five questions in the united states is 60% (three correct answers). in the 2018 data, the national average score is a statistically significant three percentage points lower than the 2015 data, while our subsample has an average financial literacy score of 47% (approximately two correct answers). 6 studies indicate a positive relationship between income and financial literacy; thus, we expect our restricted sample—in the lower part of the income distribution—to exhibit a lower score than the national average. six percent of our sample was required to take a financial literacy course in high school. however, financial literacy mandates are relatively new, with most states introducing requirements only after 2000 (stoddard & urban, 2020). fig. 1 shows the proportion of people required to participate in financial literacy education by birth year. required high school financial literacy courses are most commonly undertaken by those aged 18–23. this is a limitation of the data, because the mandate measure is correlated with age and, if age is correlated with participation in social programs, our results may be biased. a. al-bahrani et al. / financial services review 28 (2020) 303–314 307 we find no statistical difference in the mandated financial education rates between the social program participant and non-participant samples. however, the samples differed in almost all other categories. social program participants are more likely to have more children, lower income levels, and be disabled or unemployed. additionally, social program participants had lower financial literacy scores (see fig. 2). there is evidence that financial education provides fewer benefits to the less advantaged (e.g., fernandes et al., 2014; kaiser & menkhoff, 2017) and, more recently, a meta-analysis solely using randomized control trials (rcts), finds no difference in outcomes arising from financial education interventions for low-income individuals (kaiser et al., 2020). however, none of these studies examined social program participation as an outcome. fig. 3 shows financial literacy scores across the age distribution by financial education type. as expected, the sample required to take a financial literacy course scores highest; fig. 1. participation rate in required high school financial literacy course by year of birth. fig. 2. the average financial literacy score plot across the age distribution, comparing participants and nonparticipants in social programs. 308 a. al-bahrani et al. / financial services review 28 (2020) 303–314 second highest are those enrolled in an optional course, and the lowest scores are exhibited by those with no formal financial education. thus, there is a correlation between financial education and exhibited financial literacy. this relationship is also observed across the age distribution; however, we do see that financial literacy scores increase with age for all financial education types. while non-participants in social programs have higher financial literacy scores (see fig. 2), we find no differences in social program participation rates across the age distribution when we compare those required to complete a financial literacy course, those who chose to complete a fig. 3. illustrated plot of the mean financial literacy score across the age distribution for each of the financial literacy education options. financial literacy scores are positively related to exposure to financial literacy education. required participation in financial education is associated with higher financial literacy scores relative to optional participation. fig. 4. plot illustration of the marginals of predicted probability of participating in social programs across the age distribution for each of the financial literacy education options. a. al-bahrani et al. / financial services review 28 (2020) 303–314 309 financial literacy course, and those who did not participate at all (see fig. 4). in summary, there are many identifiable differences between social program participants and non-participants, but participation in high school financial education does not appear to be one of them. 4. methodology to identify the consequences of mandated financial literacy education on social program participation, we specify a linear probability model: yi,s = b 0 þ l 0xi þ d 0fi þ w 0li þ g s þ « i our dependent variable yi,s is the social program participation of individual i in state s. we include the demographics and predictors of social program participation (identified in table 1) in vector xi. the variables of interest are fi and li. variable fi is a vector of dummy variables that identify whether the respondent received (a) mandatory financial literacy education, (b) optional financial literacy education, or (c) no financial literacy education. variable li is a vector of dummy variables indicating the financial literacy of individual i based on his or her responses to the big five financial literacy questions. this allows us to hold financial literacy constant and separate it from the impact of mandatory financial education. previous studies measuring the impact of financial education mandates neglect to control for financial literacy and include only variable f.7 our approach allows us to isolate the impact of mandatory financial education and reduce selection bias in financial literacy education. we include a state fixed effects model to control for state-level variation in social program participation using g . chetty et al. (2013) document that participation in the earned income tax credit (eitc) varies by state and zip code and find that variation is due to differences in knowledge about social programs. however, chetty (2015) documents observed participation rates approaching eligibility rates in 2008, as more people became aware of the programs. we limit our analysis to participation in the supplemental nutrition assistance program (snap) and medicaid, because we found that eitc participation rates are equal to eligibility rates in 2018. both snap and medicaid are federal programs, but their eligibility is determined by state-level requirements. therefore, variations in program participation rates could be due to state-level differences. 5. results the results of the linear probability model are presented in table 2. although all variables identified in table 1 are included in our regressions, we report only the coefficients for the financial education types, age, race/ethnicity, and financial literacy scores for brevity. regressions (1) and (2) use our full sample, and provide results on the probability of social program participation without inclusion of financial literacy scoring (li) (left-hand 310 a. al-bahrani et al. / financial services review 28 (2020) 303–314 column of each panel) and with inclusion of financial literacy scoring (right-hand column), respectively. we find no difference in the significance of financial education in relation to participation in social programs when the purported confounder, financial literacy, is included. further, in model (2), we find that answering four and five of the big five financial literacy questions correctly is associated with a 6.0% and 8.8% decrease, respectively, in the probability of participation in social programs. therefore, policymakers’ assumption that individuals with higher financial literacy scores are less likely to participate in social programs is supported. however, this result is not causal. when we test for returns to financial education, we find no statistical evidence that those required to take a financial education course are less likely to rely on social table 2 linear probability model estimating the participation in social programs full sample sub-sample (18–25 years) (1) (2) (3) (4) did not take a financial literacy course omitted omitted took an optional financial literacy course �0.015 �0.011 �0.034 �0.030 [0.027] [0.027] [0.052] [0.052] took a required financial literacy course �0.015 �0.008 0.041 0.042 [0.019] [0.019] [0.030] [0.031] age �0.001* �0.000 0.015** 0.014** [0.000] [0.000] [0.006] [0.006] white omitted omitted black 0.052*** 0.044*** 0.020 0.018 [0.015] [0.015] [0.036] [0.037] hispanic �0.021 �0.027* 0.006 0.007 [0.016] [0.016] [0.034] [0.034] asian �0.080*** �0.084*** �0.013 �0.017 [0.027] [0.026] [0.057] [0.057] other 0.001 �0.001 �0.079 �0.077 [0.024] [0.023] [0.053] [0.053] 0 correct omitted omitted 1 correct �0.002 �0.002 [0.016] [0.037] 2 correct �0.005 0.066* [0.016] [0.035] 3 correct �0.014 �0.005 [0.016] [0.039] 4 correct �0.060*** �0.027 [0.018] [0.045] 5 correct �0.088*** 0.010 [0.023] [0.064] controls includeda yes yes yes yes state fixed effect yes yes yes yes observations 8,657 8,657 1,369 1,369 adjusted r2 0.27 0.27 0.19 0.19 *p < .10, **p < .05, ***p < .001. standard errors in brackets. a controls included, but not reported for brevity, are number of dependents, income, and employment status. a. al-bahrani et al. / financial services review 28 (2020) 303–314 311 programs. similarly, those choosing to take financial literacy education courses are as likely to participate in social programs as those who do not participate in any course. therefore, we find no evidence that mandating financial literacy education at the high school level is related to financial behavior changes in the context of social program participation. given the relatively recent growth of financial education mandates, we test models (1) and (2) on the younger population (18-25 years old) in our sample in regressions (3) and (4), as shown in table 2. we still find no statistical association of financial education courses with social program participation, but we do see significance appear on the age covariate. the probability of social program participation is associated with an approximate 1.5% increase as age increases from 18 to 25 years, whereas in the full sample—model (1)—an increase in age is associated with a decreased probability (0.1%) of social program participation. this is not a surprising result, given participants likely progress toward financial autonomy from the ages of 18 to 25. we also find a shift in significance when we include the financial literacy scores in regression (4), from a significantly smaller association with social program participation when four or five literacy questions are correctly answered to an increased association with social program participation when only two literacy questions are answered correctly. this may be a distribution effect in this age subsample since the number of our observations falls from 8,657 in models (1) and (2) to 1,369 in models (3) and (4). 6. limitations according to fernandes et al. (2014), the impact of financial literacy education on financial behaviors has been inconclusive. our research also finds no evidence supporting changes in financial behaviors via financial literacy education mandates. while those who score higher on financial literacy assessment are less likely to participate in social programs, it is important to note that their financial knowledge may be derived from places other than high school, including life experiences and informal education. the limitation of our research is that we rely on self-reported identification of class requirements. a common issue in this field of research is that there is no standard definition of financial literacy and financial literacy curriculum. additionally, the survey data we use classifies individuals as having experienced a required financial education course, a voluntary course, or no financial education course. readers should be careful when interpreting the experiment, as these were not random assignments. this research would therefore be classified as a kind of “non-equivalent control group” quasi-experiment (shadish, cook, & campbell, 2002). it is important, though, to distinguish this form of selection from “self-selection” whereby people self-select the treatment they receive. our study used several controls to address the differences between the participants in the different groups that existed before their differential treatment. however, there is a possibility that the significance of social program participation could be impacted by these pre-existing differences that may have not been accounted for by the control variables. 312 a. al-bahrani et al. / financial services review 28 (2020) 303–314 7. conclusion our research examines whether mandated financial literacy education is related to financial reliance on social programs such as medicaid and snap. we use the 2018 iteration of the nfcs data, which includes information on individual social program participation, the circumstances of respondents’ high school financial literacy education, and a measure of financial literacy via scoring on the big five questions. this data also allows us to differentiate between required financial education, optional selection of financial education, and those with no formal financial education at the high school level. we find that high financial literacy scores are inversely related to participation in social programs. scoring four or five out of five on the big five questions is associated with reduced participation in social programs. however, our results find no evidence that mandated financial literacy education is related to reduced social program participation. given the expectation that financial literacy programs would prove a worthwhile investment, we encourage more research into the curriculum design, bias, timing, scope, and delivery method of financial education. notes 1 https://www.ncsl.org/research/financial-services-and-commerce/financial-literacy2018-legislation.aspx. 2 https://www.ncsl.org/research/financial-services-and-commerce/financial-literacy2019-legislation.aspx. 3 school districts may identify their own graduation requirements beyond state minimums. 4 student debt composition is measured using data from national postsecondary student aid study (npsas)—1999, 2003, 2007, and 2011. 5 they either received medicaid and/or enrolled in snap in the prior 12 months. this data is only available as combined information and, therefore, cannot be disentangled. 6 in table 1 we show summary statistics for both the percentage calculation and the number of correct responses. 7 we replicate this approach in regressions (1) and (3) in table 2. acknowledgment we are grateful to the institute of the study of free enterprise (isfe) for their support of this research. this research received a summer research grant in 2018. a version of this paper is made available to the participants and attendees of the 2019 isfe summer research conference. references al-bahrani, a., weathers, j., & patel, d. (2019). racial differences in the returns to financial literacy education. journal of consumer affairs, 53, 572–599. a. al-bahrani et al. / financial services review 28 (2020) 303–314 313 brown, a., collins, j. m., schmeiser, m. d., & urban, c. (2014). state mandated financial education and the credit behavior of young adults. feds working paper no. 2014-68. available at ssrn: https://ssrn.com/ abstract=2498087 or http://dx.doi.org/ chetty, r. (2015). behavioral economics and public policy: a pragmatic perspective. american economic review, 105, 1–33. chetty, r., friedman, j. n., & saez, e. (2013). using differences in knowledge across neighborhoods to uncover the impacts of the eitc on earnings. american economic review, 103, 2683–2721. christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54, 18–38. fernandes, d., lynch, j. g., jr., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors.management science, 60, 1861–1883. hastings, j., & mitchell, o. s. (2020). how financial literacy and impatience shape retirement wealth and investment behaviors. journal of pension economics & finance, 19, 1–20. hastings, j., mitchell, o. s., & chyn, e. (2011). fees, framing, and financial literacy in the choice of pension manager. in o. s. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 101–115). oxford: oxford university press. hastings, j. s., & tejeda-ashton, l. (2008). financial literacy, information, and demand elasticity: survey and experimental evidence from mexico (nber working papers 14538). national bureau of economic research. harvey, m. (2019). impact of financial education mandates on younger consumers’ use of alternative financial services. journal of consumer affairs, 53, 731–769. kaiser, t., lusardi, a., menkhoff, l., & urban, c. j. (2020). financial education affects financial knowledge and downstream behaviors (nber working papers 27057). national bureau of economic research. kaiser, t., & menkhoff, l. (2017). does financial education impact financial literacy and financial behavior, and if so, when? the world bank economic review, 31, 611–630. lusardi, a. (1999). information, expectations, and savings for retirement. in h. aaron (ed.), behavioral dimensions of retirement economics (pp. 81–115). washington, dc: brookings institution press and russell sage foundation. lusardi, a. (2008). household saving behavior: the role of financial literacy, information, and financial education programs (nber working papers 13824). national bureau of economic research. lusardi, a. (2019). financial literacy and the need for financial education: evidence and implications. swiss journal of economics and statistics, 155, 1. lusardi, a., michaud, p. c., & mitchell, o. s. (2017). optimal financial knowledge and wealth inequality. the journal of political economy, 125, 431–477. lusardi, a., & mitchell, o. s. (2011b). financial literacy and planning: implications for retirement wellbeing (nber working papers 17078). national bureau of economic research. lusardi, a., & tufano, p. (2015). debt literacy, financial experiences, and overindebtedness. journal of pension economics & finance, 14, 332–368. miller, m., reichelstein, j., salas, c., & zia, b. (2015). can you help someone become financially capable? a meta-analysis of the literature. the world bank research observer, 30, 220–246. shadish, w. r., cook, t. d., & campbell, d. t. (2002). experimental and quasi-experimental designs for generalized causal inference. boston, ma: houghton-mifflin stoddard, c., & urban, c. (2020). the effects of state-mandated financial education on college financing behaviors. journal of money, credit and banking, 52, 747–776. wagner, j., & walstad, w. b. (2019). the effects of financial education on short-term and long-term financial behaviors. journal of consumer affairs, 53, 234–259. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. zhan, m., anderson, s. g., & scott, j. (2006). financial knowledge of the low-income population: effects of a financial education program. journal of sociology & social welfare, 33, 53. 314 a. al-bahrani et al. / financial services review 28 (2020) 303–314 ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review expires december 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce exams (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in ricaldi, martin, and huston, the household that has enough money to pay off the credit card balance but chooses not to is called: a. a solvent revolving user b. a convenience user c. a revolving user d. an insolvent revolving user 2. in “financial literacy and its impact on the credit card debt puzzle” by ricaldi, martin and huston, to control the “doer” in a household, the “planner” will… a. impose a spending limit on the credit card b. pay off all credit card balances each month. c. the “planner” cannot control the “doer”. d. create mental accounts to reduce the temptation to spend. 3. when comparing convenience users to solvent revolvers with the highest financial literacy scores, the authors find that: a. increases in income and net worth negatively impact the likelihood of being a solvent revolver. b. income and net worth are not statistically significant in the likelihood of being a solvent revolver. c. decreases in income and net worth negatively impact the likelihood of being a solvent revolver. d. changes in income and net worth have different impacts on the likelihood of being a solvent revolver. 4. the compensation puzzle proposed by heller, cummings, and martin describes a. the process rias undertake when determining how to charge clients b. the additional process ibds face in determining a client’s mix of products across different compensation programs c. the difficulty product manufacturers undergo when incorporating advisor compensation in new product offerings d. the differences in fee structures for ibds compared to rias 5. heller, cummings, and martin found that advisors at rias who focus more on _____ portfolios are able to outperform their ibd counterparts. a. fixed income b. equity-heavy c. balanced d. actively-managed manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: terrance k. martin, e-mail: terrance.martin@uvu.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services terrance martin woodbury school of business utah valley university 800 w. university pkwy. orem, ut 84058 (address service requested) financial services review, 32(3) 68 global perspectives on the determinants of older adults’ subjective well-being: a comprehensive longitudinal study yi liu,1 tao guo,2 and yuanshan cheng3 abstract current research has established a relationship between older adults' subjective well-being and factors extending beyond their economic status to encompass various non-monetary elements. while most studies in this domain focus on factors within a single country, our analysis utilizes international longitudinal surveys to explore older adults' well-being at both the national and global scale. this comprehensive analysis considers microand macro-level determinants of retirement well-being, revealing consistent variations in happiness levels across countries. our study specifically suggests a compelling positive relationship between age and subjective well-being within the united states. this finding presents a contrast with the negative association observed in european countries. our global analysis further indicates a positive relationship between age and subjective well-being. this study not only contributes to greater understanding of the complexities related to aging and well-being but it also provides significant implications for financial services professionals and public policymakers that aim to improve the well-being of elderly populations. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation liu, y., guo, t., & cheng, y. (2024). global perspectives on the determinants of older adults’ subjective well-being: a comprehensive longitudinal study. financial services review 32(3), 6882. introduction subjective well-being (swb) plays a crucial role in comprehending individuals' subjective evaluations of their personal satisfaction and overall quality of life. in the context of our study, subjective well-being is defined through the dimension of hedonic well-being, focusing on an individual's experiences of happiness. while 1 corresponding author (yliu@sjfc.edu). st. john fisher university, rochester, ny, usa 2 morningstar investment management llc, chicago, il, usa 3 morningstar investment management llc, chicago, il, usa financial services professionals and public policymakers have traditionally focused on providing adequate financial resources to older adults, subjective well-being at retirement has become a popular topic in the financial planning academic community. the current literature provides evidence that older adults' subjective well-being is related to not only their income and wealth level (sacks et al., 2010) but also nonhttps://creativecommons.org/licenses/by-nc/4.0/ mailto:yi.liu@cffp.edu https://creativecommons.org/licenses/by-nc/4.0/ liu et al. 69 monetary factors and demographic factors such as social interactions (deleire & kalil, 2010), age (rubio et al., 2022; lim & lee, 2021), education (fan & park, 2021), marital status (lim & lee, 2021), and similar variables. subjective well-being is an important indicator of individual and societal welfare beyond objective development indicators, such as gross domestic product (gdp) and health (dolan & white, 2007). indeed, a french government agency proposed including subjective well-being as a measure of national economic performance and social progress (steptoe et al., 2015). most studies on older adults’ subjective wellbeing only look at the determinants within one country, and few researchers have examined these determinants across the globe. findings from single-country or region-specific studies are often contradictory due to inconsistent measurements and methodologies. thus, analyzing findings holistically is challenging. furthermore, the existing literature predominantly examines countries with robust economies, including the united states, china, various european nations, and south korea (gu & wei, 2018; knight et al., 2009; lee et al., 2011; shields & wheatley, 2005). this inclination is driven by the ample availability and accessibility of data resources in these economically advanced nations (jardet & meunier, 2022). however, global macro-level studies have limitations associated with uncaptured individual characteristics, using country averages, assuming data symmetry, and small sample sizes. given the limitations of both categories of studies, international collaboration is needed to better understand microand macro-level factors globally across different cultures, economic development, and public policy. the current study fills the gap by leveraging a group of longitudinal surveys worldwide to investigate older adults’ well-being at both the withincountry and international level. we investigate older adults’ subjective well-being and its macrolevel determinants. the results have the potential to enhance public policymakers' comprehension of the subjective well-being status among older adults, shedding light on key features that distinguish the subjective well-being of older adults living in the united states from that of their counterparts around the globe. our research findings can also be used to inform policymakers and other stakeholders about the underlying reasons for country differences and which countries implement the most effective public policies and why. the findings have direct implications for public policies that involve psychological, societal, or economic interventions to improve older adults’ well-being. financial services professionals could also benefit from our findings by gaining insight into the microand macro-level determinants of subjective well-being in retirement. using information presented in this paper, financial service professionals could develop planning strategies to supplement and even complement public policies, including strategies in life planning, daily activity planning, organization of social gatherings, and volunteer activities. literature review well-being across countries: the micro level the existing literature generally explores the determinants of subjective well-being at the micro-level (i.e., country or region) and macrolevel (i.e., global). at the micro-level, nearly all academic studies on subjective well-being are based on a specific country or region. subjective well-being studies have been mainly conducted in the united states (blanchflower & oswald, 2008; luttmer, 2005; subramanian et al., 2005; yang, 2008) and european countries (clark & oswald, 1994; eren & aşıcı, 2017; gredtham & johannesson, 2001; gu & wei, 2018; hayo & seifert, 2003; oswald, 1997). in the last decade, numerous studies have been conducted in asian countries (cheah & tang, 2013; chyi & mao, 2012; peng & she, 2018; rahayu, 2016; senasu & singhapakdi, 2018) and other countries, including saudi arabia (il-khraif et al., 2019), throughout africa (kollamparambil, 2020), and latin american countries (graham & felton, 2006). in general, there is accumulating evidence that well-being is associated with age (yang, 2008), gender (alesina et al., 2004), marital status (tokuda & inoguchi, 2008), income (knight et al., 2009), education (blanchflower et al., 2004), and health condition (oswald & powdthavee, 2008). furthermore, households that reach financial services review, 32(3) 70 consensus on decision-making processes regarding savings and major life choices tend to report greater financial satisfaction compared to those lacking such agreement (gray et al., 2022). however, findings are not consistent across countries and studies. take age, for example, studies from the united states that report the age and happiness pattern as a u-shaped curve with higher levels of well-being at younger and older ages with the lowest life satisfaction in the middle ages (blanchflower & oswald, 2008; fujita & diener, 2005). researchers from the united kingdom often observe the age and happiness pattern as being ո-shaped (bartolini et al., 2013; fitzroy et al., 2014). in terms of education, studies from the united states generally indicate that well-being increases with education (blanchflower & oswald, 2004; bukenya et al., 2003). a study from latin america found that years of education increases overall well-being (graham & pettinato, 2001), while researchers from united kingdom and australia have shown a negative effect of higher education on well-being (e.g., powdthavee, 2010; shields et al., 2009). regarding marital status, numerous studies from the united states indicate that marriage is linked to an elevated sense of well-being (blanchflower & oswald, 2004; cabanas, 2016; fitzroy & nolan, 2020). conversely, research from the united kingdom suggests that, for men, marriage does not correspond to higher subjective wellbeing compared to cohabitation (perelli-harris, 2019). nonetheless, for women, marriage appears to be more advantageous on average, and the statistical disparities between marriage and cohabitation diminish (perelli-harris, 2019). in the context of german women, marriage does not show a significant difference in well-being compared to cohabitation (perelli-harris, 2019). in general, the literature shows a positive relationship between income and well-being in the united states (shields & wheatley, 2005), european countries (caporale, 2009), and latin america (graham & felton, 2006). additionally, women tend to report higher well-being (alesina et al., 2004). additionally, individuals who report good health conditions tend to be happier than those experiencing poor health conditions (shields & wheatley, 2005). well-being at global the macro level besides categorizing determinants of subjective well-being at the micro-level, the literature also includes global macro-level studies exploring the determinants of subjective well-being. a variety of entities or nonprofit organizations have released papers on global well-being. gallup is a well-known organization that has released a global well-being index. the organization for economic cooperation and development, the global happiness council, the world value survey, and the global happiness and wellbeing policy are other important organizations that have been actively contributing to measuring and monitoring global well-being. researchers have utilized these data resources intensively in the well-being literature. for example, ngamaba (2017), using the world value survey, examined the determinants of subjective well-being in representative samples of nations. ngamaba found that in the lowest 10 subjective well-being countries, health status is one of the main factors associated with subjective well-being. theoretical framework traditional utility theory (ando & modigliani, 1963) suggests that individuals derive utility and happiness through consumption and, therefore, choose to smooth out their consumption to maximize lifetime utility or happiness. this suggests that individuals with sufficient retirement income should display a flat wellbeing pattern by age, which contradicts the ushape curve observed in some empirical studies (e.g., blanchflower & oswald, 2008). the socioemotional selectivity theory (sst) was formulated by carstensen during the early 1990s as a motivational theory. it offers a structure for comprehending the evolution of individuals' objectives and incentives throughout their lives, especially while their sense of time undergoes transformations. sst suggests that individuals who see their time as finite, typically as a result of aging, tend to prioritize emotionally significant objectives, relationships, and activities over the pursuit of new knowledge or the expansion of their social connections. this theory is intricately connected to notions of well-being, particularly liu et al. 71 as individuals age. as individual age, individuals develop wisdom and select consumption, activities, and friends to enhance their happiness (carstensen et al., 2003). therefore, their ability to pick satisfying items improves. in this study, we combine this socioemotional selectivity theory with the classic utility theory as our theoretical framework. the study focuses on hedonic well-being due to the availability of the data in the surveys. we argue that a combination of theories could explain well-being. methodology data and sample the health and retirement study (hrs) is a longitudinal survey that includes participants aged 50 years of age or older. the survey is maintained by the university of michigan and supported by the national institute on aging and the social security administration. this survey started in 1992 as the first study to collect longitudinal information about both the health and economic conditions of older adults. due to the innovativeness and success of this study, the hrs has served as a model for sister studies around the world. although participants from the other countries have not been asked the same questions at the same frequency and scope, the availability of the hrs and its sister studies makes it possible to conduct multi-national analyses on social science topics. the university of southern california (usc), with the support of the nia and nih, created a gateway to global aging dataset, compiling questions from major hrs sources across the world, thus making it easier for researchers to compare these datasets. survey data is still maintained within its own country, and researchers still have to apply for access to each country’s data separately, but this dataset serves as a good starting point to know which surveys have the same questions. as of 2023, usc has harmonized survey data from 11 countries/regions, covering over one million observations. in the current study, we chose five major hrs datasets, which measure well-being across a diverse range of geographical areas. these datasets are the hrs developed for the united states, elsa developed for the united kingdom, share developed for the european union countries plus switzerland and israel, klsa for south korea, and charles for china. the chinese survey is the newest of the five (started in 2011; four waves of the survey have been conducted). all five are longitudinal surveys. measurements we utilized sas software to perform the data analysis, specifically applying logistic regressions to examine the associations between various predictors and the outcome variable, y (swb), which represents a dimension of subjective well-being. the specification of the logistic regression models is as follows: 𝑌(𝑆𝑊𝐵) = 𝑎1 + β 𝑖𝑋𝑖 + β𝑋𝑗 + ∑ β 𝑘x 𝑘 + ϵ (1) here, α serves as the intercept, β 𝑖, β 𝑗, and β 𝑘 are the coefficients representing the influence of respective predictor variables on subjective wellbeing. 𝑋𝑖 symbolizes the age variable, illustrating its effect on the facet of subjective well-being (swb). 𝑋𝑗 is erroneously also attributed to age in the initial description, suggesting a need for clarification or correction to accurately represent another dimension or interaction involving age. 𝑋𝑘 incorporates a range of control variables that are chosen to consider extra aspects that are believed to have an influence on the outcome. ϵ denotes the error term, which accounts for the variability in y (swb) that is not accounted for by the predictors in the model. every individual term inside the model represents the impact of its respective predictor on the aspect of subjective well-being, while also accounting for the potential influence of other factors included in the model. integrating socioemotional selectivity theory (sst) and utility theory, we propose a subjective well-being function, w, to explicitly include the variables available in the logistic regression models. this refined function considers both individual-level factors (model 1) and broader socio-economic and environmental variables (model 2). this function is articulated as: 𝑊 = 𝑓(𝐸, 𝑈, 𝐷, 𝑆) (2) financial services review, 32(3) 72 where e, captures emotional satisfaction, influenced by age and health, highlighting the role of health and perceived time (inversely related to age) in influencing emotional wellbeing. the utility component u represents the satisfaction derived from economic resources, such as annuitized net worth; d encompasses demographic and social variables such as education, marital status, and gender, reflecting their impact on both satisfaction and utility. the expanded component s includes socio-economic and environmental factors including a giving index, life expectancy, and gdp. subjective well-being can be evaluated using one of three approaches (steptoe et al., 2015): (a) evaluative well-being (i.e., life satisfaction), (b) hedonic well-being (i.e., feeling of happiness or sadness), or (c) eudaimonic well-being (i.e., sense of purpose and meaning in life). while each measurement is available in the u.s. hrs, surveys from other countries contain fewer measurements. the only subjective well-being measure available in all five surveys is hedonic well-being. the question asks, “much of the time during the past week, you were happy?” responses are either “yes” (coded 1) or “no” (coded 0). regarding other explanatory variables, the models included age, gender, marital status, wealth, and income as explanatory variables. these were assumed to be quality-of-life indicators. age was measured as a continuous variable. gender was a dichotomous variable that took a value of 1 if a participant was female and 0 if a participant was male. education was converted into dummy variables (i.e., less than high school, high school, and college degree and above). each dummy variable took a value of 1 if the participant had a less than a high school level of education and 0 if otherwise. income and net worth were rescaled by $1,000. health condition was a dichotomous variable. each dummy variable took a value of 1 if a participant reported a good health condition, otherwise 0. the major contribution of this study is controlling household-level demographic variables and country-level macro-economic factors simultaneously. after conducting a logistic regression analysis by country, we combined all the samples and examined the impact of both household-level demographic variables and country-level macro-economic factors on subject well-being in the same regression model. the country-level macro-economic factors in this study included a giving index, average birth rates, national tertiary education rates, average life expectancy, national average of out-of-pocket medical costs, national average of tax rates, national unemployment rates, urbanization ratio, per-capita gdp, average consumer price index (cpi) change rates (as a measure of inflation), and national co2 emission rates. results descriptive analysis the european union and united states have relatively large samples (n = 10,699 and n = 12,472, respectively). the united kingdom and south korea have similar sample sizes (n = 5,233 and n = 5,692, respectively). the sample size from china is relatively small (n = 2,977). some socio-demographic and economic characteristics were relatively similar across countries. for example, the proportion of male participants in the european union, united kingdom, south korea, and the united states were similar (46%, 46%, 43%, and 41%, respectively). china had slightly more male participants (55%). the average age of participants in the european union, united kingdom, south korea, and the united states was higher (m = 70.37, 71.15, 66.67, 72.92, and 72.2, respectively) compared to the average age of participants in china (m = 66.67). the european union, united kingdom, south korea, and the united states had a higher percentage of married participants compared to other regions, with 72%, 69%, 72%, and 59% of individuals being married, respectively. in china, all participants were married. table 1 presents further descriptive statistics. figure 1 shows differences in the level of subjective well-being by country. age was positively associated with subjective well-being in the united states, whereas age was negatively associated with subjective well-being in the european union consistent with the literature, annuitized income, net worth, educational attainment, marital status, and health condition liu et al. 73 were positively associated with subjective wellbeing globally. figure 1. the relationship between average subjective well-being and age at the country level in regard to the other demographic characteristics, guaranteed lifetime income was found to mitigate retirees’ longevity risk, which should lead to less stress and a higher level of subjective well-being. figure 2 shows the average subjective well-being by level of annuitized income across countries. consistently, subjective well-being increased with the level of annuitized income even though a difference in the level of subjective well-being by country was observed. figure 2. the relationship between average subjective well-being and annuitized income at the country level 70.00% 75.00% 80.00% 85.00% 90.00% 95.00% 60-64 65-69 70-74 75-79 80-84 85+ eu (2002) uk (2016) china (2014) korea (2016) u.s. (2016) 70.00% 75.00% 80.00% 85.00% 90.00% 95.00% 100.00% <5k 5k-10k 10k-20k 20k-40k 40k+ eu (2002) uk (2016) china (2014) korea (2016) u.s. (2016) financial services review, 32(3) 74 table 1. descriptive statistics across countries eu uk china south korea united states m in m ax m ea n m in m ax m ea n m in m ax m ea n m in m ax m ea n m in max mean age 60 104 70.37 60 90 71.15 60 90 66.67 60 102 72.92 60 107 72.2 annuitized income in $1k 0 431 2 0 300 6 0 288 5 0 7 0 0 857 7 net worth in $100k -18 111 2 -3 111 5 -8 500 2 0 7 0 -11 314 5 education high school 0 1 28% 0 1 50% 0 1 6% 0 1 26% 0 1 33% college 0 1 16% 0 1 19% 0 1 1% 0 1 8% 0 1 49% married 0 1 72% 0 1 69% 1 1 100% 0 1 72% 0 1 59% male 0 1 46% 0 1 46% 0 1 55% 0 1 43% 0 1 41% healthy 0 1 23% 0 1 39% 0 1 11% 0 1 4% 0 1 36% sample size 10,699 5,233 2,977 5,692 12,472 empirical results table 2 shows the relationship between subjective well-being and demographic characteristics in each country, estimated through a logistic regression analysis. consistent with the previous figures, age was positively associated with well-being in the united states but negatively associated with subjective well-being in the european union. specifically, when converted to odds ratios, the results indicated that, on average, participants in the united states had 8.85% higher well-being for every one-year increase in age. those in the european union reported 1.21% less well-being for every one-year increase in age. consistent with the literature, income, net worth, educational attainment, and marital status were positively associated with subjective well-being. health condition was the only variable that was significant in all countries, although in south korea there was a negative relation between reported health condition and subjective well-being. the variable inflation factor was tested; it was determined that multicollinearity was not an issue of concern in the model. liu et al. 75 table 2. logistic regression results: subjective well-being by country subjective well-being by country (model 1-1) eu (2002) uk (2016) china (2014) south korea (2016) united states(2016) age -0.0120*** 0.0070 -0.0033 0.0019 0.0218*** (0.0033) (0.0066) (0.0082) (0.0042) (0.0032) annuitized inc. in $1k 0.0038 0.0252*** 0.0182*** 0.0921 0.0010 (0.0040) (0.0086) (0.0053) (0.0613) (0.0017) net worth in $1k 0.0446*** 0.0084 -0.0023 0.0259 0.0074** (0.0099) (0.0119) (0.0023) (0.0883) (0.0035) high school 0.2241*** 0.1470 0.4662** 0.0927 0.1413* (0.0589) (0.1159) (0.2306) (0.0822) (0.0768) college 0.3091*** -0.1792 0.2225 0.00607 -0.0504 (0.0821) (0.1673) (0.5575) (0.1308) (0.0746) married 0.4095*** 0.7513*** -0.1147 0.5875*** (0.0561) (0.1105) (0.0840) (0.0587) healthy 1.1455*** 1.1145*** 0.7847*** -0.3462** 1.3091*** (0.0816) (0.1323) (0.1735) (0.1489) (0.0753) male 0.186*** 0.2697** 0.17* 0.0167 0.1977*** (0.0525) (0.1112) (0.0902) (0.0737) (0.0595) sample size 10,699 5,233 2,977 5,692 12,472 notes: significance levels: ***p < .01, **p < .05, *p < .10; standard errors are in parentheses. table 3 shows the same model with additional interaction terms among country dummies and age. model 2 was used to investigate the differences in the change in subjective well-being for a one-year increase in age between participants from the european union, china, south korea, and the united kingdom, relative to the united states the results showed that changes in subjective well-being for a one-year increase in age of participants from the european union, china, and the united kingdom were significantly lower than the difference in the change in subjective well-being of participants from the united states this suggests that subjective well-being increases more for participants from the united states than other countries/regions in the sample. model 3 in table 3 includes additional interaction terms between countries’ dummies and annuitized income. the findings indicate that the increase in subjective well-being associated with a rise in annuitized income among participants from china and the united kingdom is higher than the change observed in subjective well-being among participants from the united states. financial services review, 32(3) 76 table 3. logistic regression results: models with interactions terms subjective well-being variables model 2 model 3 estimate standard error estimate standard error age 0.020*** (0.0033) 0.006*** (0.0019) annuitized inc. in $1k 0.005** (0.0002) 0.002 (0.0002) eu x age -0.034*** (0.0045) china x age -0.021** (0.0088) south korea x age -0.002 (0.0050) uk x age -0.016** (0.0071) eu x annuitized inc 0.0060 (0.0005) china x annuitized inc 0.017*** (0.0006) south korea x annuitized inc -0.00015 (0.0060) uk x annuitized inc 0.023*** (0.0009) net worth in $100k 0.007 (0.0051) 0.008 (0.0054) high school 0.188*** (0.0370) 0.186*** (0.0369) college 0.097** (0.0456) 0.086* (0.0460) married 0.418*** (0.0350) 0.423*** (0.0350) healthy 1.099*** (0.0473) 1.111*** (0.0473) male 0.154*** (0.0309) 0.141*** (0.0309) eu dummy 2.025*** (0.3210) -0.445*** (0.0407) china dummy 0.919 (0.5944) -0.596*** (0.0657) south korea dummy -0.222 (0.3652) -0.376*** (0.0511) uk dummy 1.568*** (0.5140) 0.310*** (0.0707) sample size 37,073 37,073 notes: significance levels: ***p < .01, **p < .05, *p < .10; standard errors are in parentheses. table 4 reports logistic regression results after including macro-level data from the five countries. the household-level demographic variable coefficients were consistent with the previous analyses. at the macro level, the national post-secondary education rate was positively related to subjective well-being. life expectancy, which could be an indicator of health care services, was positively related to subjective well-being. out-of-pocket medical costs, tax rates, national average birth rates, and unemployment rates were negatively related to subjective well-being. the urban population percentage variable had a positive relation with subjective well-being. although one might expect that people living in countries with a higher gdp might report higher subjective wellbeing than those in lower gdp countries, the relationship between subjective well-being and per-capita gdp was actually negative. this finding is consistent with the literature that subjective well-being does not grow uniformly with the economy, and in some cases, it decreases (diener & oishi, 2000; easterlin, 2005). liu et al. 77 table 4. logistic regression results; the macro level parameter estimate standard error p value intercept -16.2431 2.3981 <.0001 age 0.0046 0.0019 0.0157 annuitized inc. in $1k 0.0072 0.0023 0.0014 net worth in $1k 0.0103 0.0064 0.1082 high school 0.1542 0.0380 <.0001 college -0.0238 0.0473 0.6158 married 0.4473 0.0356 <.0001 healthy 1.1114 0.0476 <.0001 male 0.1590 0.0310 <.0001 giving index -0.7060 0.4102 0.0852 national birth rate -0.0298 0.0124 0.0164 college edu rate 1.5029 0.2055 <.0001 life expectancy 0.2134 0.0301 <.0001 out-of-pocket medical -3.7743 0.5529 <.0001 national tax rate -3.2268 0.3379 <.0001 unemployment rate -2.1995 0.7963 0.0057 urban population % 0.8669 0.2602 0.0009 per-capita gdp -0.00000588 0.0000 0.0029 cpi change 43.5853 7.2885 <.0001 co2 emission 0.0000001204 0.0000 <.0001 sample size 37,073 the cpi change rate is a measure of the inflation rate that reflects the annual percentage change in the average consumer’s cost of acquiring a basket of goods and services. conventionally, higher inflation is related to lower happiness (di tella et al., 2001). inflation reduces the purchasing power of savings and retirement income. a highly inflationary situation is usually accompanied by an economic slowdown, which might lead to a reduction in subjective well-being. test results suggest the opposite. a positive relationship between the cpi change rate and happiness was observed. it might be the case that a higher inflation rate indicates a higher wage or pension increase, which could be the channel leading to retirement subjective well-being. similarly, whereas co2 emissions should be an indication of a negative environmental impact, co2 emissions can also be an indication of industrialization, which may contribute positively to well-being. discussion and conclusion subjective well-being is an important measure of overall individual well-being and one of the measures of national economic performance and social progress. although researchers have attempted to identify factors affecting well-being, to our knowledge, no study has investigated subjective well-being cross-nationally using a single dataset. the current study offers a unique opportunity to gain insight into aged populations from a national microand macro-data level. this study presents a novel contribution to the existing literature, as previous studies have not investigated the role of the association between the determinants of subjective well-being in later life at the international/global level. specifically, we examined the determinants of older adults’ subjective well-being across the european union, the united kingdom, china, south korea, and the united states. we found that the age profile of subjective well-being differs among countries. financial services review, 32(3) 78 specifically, age contributes positively to subjective well-being in the united states but negatively to subjective well-being in the european union. this finding highlights that where individuals live is an important contributor to overall satisfaction levels. from a global perspective, age is related positively to subjective well-being. consistent with the literature, this study provides supporting evidence to the hypothesis that high levels of wealth and income predict higher subjective well-being (diener et al., 1999). we also found that marital status was positively associated with subjective well-being in the european union, united kingdom, and united states, but not south korea. being alone appears to have a negative effect on subjective well-being compared to being married (diener et al., 2000). consistent with the literature, this study showed that reporting good health is an important factor in exhibiting higher subjective well-being. additionally, we found a negative relation between subjective well-being and per-capita gdp. an individual’s or a country’s overall level of well-being is not necessarily just bound by income and wealth (easterlin, 1974). socioeconomic factors tend to be equally important in describing overall satisfaction levels. subjective well-being might only differ by income level within a country but not when assessed using international data (easterlin, 1974). happiness and subjective well-being at the national level do not increase with wealth once basic needs are fulfilled (diener & oishi, 2000). studies have shown that economic growth is positively associated with happiness gains in poor countries. however, once basic needs are met, including living standards, further economic growth does not always contribute to more gains in a country’s happiness, as other factors, such as income inequality, distrust, status anxiety, and perceived conflicts, influence happiness levels (delhey & dragolov, 2014). moreover, living costs and stress in each country/region, as proxied by out-of-pocket medical costs, average tax rates, and unemployment rates, appear to also be negatively associated with subjective well-being. limitations there are some limitations associated with the current study. first, although self-report measures are the most common assessment technique used in subjective well-being research, the danger of measurement bias should be recognized. additionally, due to data limitations, less-than-ideal measures (a binary variable) for retirees’ subjective well-being were utilized, perhaps leading to inadequate accuracy and reliability in the current study. future research with more direct measures, such as likert-scale questionnaires, is needed. in addition, causation cannot be inferred from our analyses, given that we used cross-sectional data, which is not free from endogeneity resulting from omitted variables, measurement errors, and simultaneity. future studies would benefit from a longitudinal data analysis to validate the relationships between happiness and per-capita gdp, cpi change, and co2 emissions. implications subjective well-being reflects the extent to which individuals think and feel that their lives are going well (diener et al, 1999; kahneman & schwaz, 1999). consistent with the existing literature, age, health condition, and being married were found in this study to be positively associated with subjective well-being. financial service professionals should work diligently with clients and discover factors and activities that lead to high well-being. for clients with health issues, and those who are not living with a spouse or partner, financial service professionals can help clients discover their financial or emotional concerns and identify resources to address their needs. it is important to acknowledge that the level of annuitized income was, in this study, positively related to subjective well-being. financial service professionals should work with their clients and discuss the benefits and concerns of annuitizing their wealth. the findings reported here can be used to increase public policymakers’ understanding of the status of older adults’ subjective well-being, as subjective well-being is one of the measures of national economic performance and social progress. knowing that socioeconomic factors and public policy decisions influence subjective liu et al. 79 well-being, policymakers can use this to guide future policy decisions and improve the quality of life within and across countries. while continued industrialization and urbanization generally improve healthcare quality and residents’ life expectancy, which could improve residents’ subjective well-being, the marginal improvement is diminishing. residents’ subjective well-being does not always grow with the economy; in some countries, well-being decreases with gains in the economy (diener & oishi, 2000; easterlin, 2005). policies that reduce out-of-pocket medical costs and average income tax rates could improve residents’ subjective well-being. our research also offers significant insights for financial service professionals, especially when assisting immigrant clients or anyone considering relocating to another country for retirement. before making decisions about relocating, it is essential for financial service professionals to have discussions about the socioeconomic characteristics of the planned destination country or region. the economic standing and subjective well-being of individuals can be significantly influenced by factors such as the healthcare system's quality and cost, taxation rates, industrialization, and growth in urbanization. moreover, financial service professionals can improve their services by creating comprehensive strategies that not only supplement but also harmonize with public policies. these tactics may involve life planning, scheduling daily activities, and organizing social gatherings and volunteer activities. by incorporating these elements into their planning methodology, financial experts may cultivate a comprehensive and customized financial strategy for their clients, encompassing both the economic and subjective well-being aspects. references al-khraif, r. m., al-mutairi, a. n., alradihan, k. o., & salam, a. a. (2019). levels and determinants of retiree life satisfaction in saudi arabia. journal of population ageing, 12, 109-126. alesina, a., di tella, r., & macculloch, r. (2004). inequality and happiness: are europeans and americans different? journal of public economics, 88(9-10), 2009-2042. https://doi.org/10.1016/j.jpubeco.2003.07.006 ando, a., & modigliani, f. (1963). the" life cycle" hypothesis of saving: aggregate implications and tests. the american economic review, 53(1), 55-84. https://www.jstor.org/stable/1817129 bartolini, s., bilancini, e., & sarracino, f. (2013). predicting the trend of well-being in germany: how much do comparisons, adaptation and sociability matter? social indicators research, 114(2), 169-191. https://www.jstor.org/stable/24720242 blanchflower, d. g., & oswald, a. j. (2004). well-being over time in britain and the usa. journal of public economics, 88(78), 1359-1386. https://doi.org/10.1016/s00472727(02)00168-8 blanchflower, d. g., & oswald, a. j. (2008). is well-being u-shaped over the life cycle? social science & medicine, 66(8), 1733-1749. bukenya, j. o., gebremedhin, t. g., & schaeffer, p. v. (2003). analysis of rural quality of life and health: a spatial approach. economic development quarterly, 17(3), 280-293. cabanas, e. (2016). rekindling individualism, consuming emotions: constructing “psytizens” in the age of happiness. culture & psychology, 22(3), 467-480. https://doi.org/10.1177/1354067x16655459 caporale, g. m., georgellis, y., tsitsianis, n., & yin, y. p. (2009). income and happiness across europe: do reference values matter? journal of economic psychology, 30(1), 4251. https://doi.org/10.1016/j.joep.2008.06.004 carstensen, l. l., fung, h. h., & charles, s. t. (2003). socioemotional selectivity theory and the regulation of emotion in the second half of life. motivation and emotion, 27(2), 103-123. cheah, y. k., & tang, c. f. (2013). the sociodemographic determinants of self-rated happiness: the case of penang, malaysia. https://doi.org/10.1016/j.jpubeco.2003.07.006 https://www.jstor.org/stable/1817129 https://www.jstor.org/stable/24720242 https://doi.org/10.1016/s0047-2727(02)00168-8 https://doi.org/10.1016/s0047-2727(02)00168-8 https://psycnet.apa.org/doi/10.1177/1354067x16655459 https://doi.org/10.1016/j.joep.2008.06.004 financial services review, 32(3) 80 hitotsubashi journal of economics, 54(1), 1-16. chyi, h., & mao, s. (2012). the determinants of happiness of china’s elderly population. journal of happiness studies, 13(1), 167-185. https://link.springer.com/article/10.1007/s1 0902-011-9256-8 clark, a. e., & oswald, a. j. (1994). unhappiness and unemployment. the economic journal, 104(424), 648-659. https://doi.org/10.2307/2234639 deleire, t., & kalil, a. (2010). does consumption buy happiness? evidence from the united states. international review of economics, 57(2), 163-176. delhey, j., & dragolov, g. (2014). why inequality makes europeans less happy: the role of distrust, status anxiety, and perceived conflict. european sociological review, 30(2), 151-165. diener, e., gohm, c. l., suh, e., & oishi, s. (2000). similarity of the relations between marital status and subjective well-being across cultures. journal of cross-cultural psychology 31(4), 419-436. https://doi.org/10.1177/00220221000310040 diener, e., & oishi, s. (2000). money and happiness: income and subjective wellbeing across nations. in e. diener & e. m. suh (eds.), culture and subjective wellbeing (pp. 185–218). cambridge, ma: the mit press diener, e., suh, e. m., lucas, r. e., & smith, h. l. (1999). subjective well-being: three decades of progress. psychological bulletin, 125(2), 276-302. di tella, r., macculloch, r. j., & oswald, a. j. (2001). preferences over inflation and unemployment: evidence from surveys of happiness. american economic review, 91(1), 335-341. dolan, p., & white, m. p. (2007). how can measures of subjective well-being be used to inform public policy? perspectives on psychological science, 2(1), 71-85. easterlin, r. a. (1974). does economic growth improve the human lot? some empirical evidence. in nations and households in economic growth (pp. 89-125). academic press. easterlin, r. a. (2005). feeding the illusion of growth and happiness: a reply to hagerty and veenhoven. social indicators research, 74(3), 429-443. eren, k. a., & aşıcı, a. a. (2017). the determinants of happiness in turkey: evidence from city-level data. journal of happiness studies, 18(3), 647-669. fan, l., & park, n. (2021). factors mediating the association between financial socialization and well-being of young adults: testing a conceptual framework. journal of financial counseling and planning. 32(2), 202-216. fitzroy, f. r., & nolan, m. a. (2020). education, income and happiness: panel evidence for the uk. empirical economics, 58(5), 2573-2592. fitzroy, f. r., nolan, m. a., steinhardt, m. f., & ulph, d. (2014). testing the tunnel effect: comparison, age and happiness in uk and german panels. iza journal of european labor studies, 3(1), 30. https://izajoels.springeropen.com/articles/1 0.1186/2193-9012-3-24 fujita, f., & diener, e. (2005). life satisfaction set point: stability and change. journal of personality and social psychology, 88(1), 158. https://doi.org/10.1037/00223514.88.1.158 gerdtham, u. g., & johannesson, m. (2001). the relationship between happiness, health, and socio-economic factors: results based on swedish microdata. the journal of socioeconomics, 30(6), 553-557. graham, c., & felton, a. (2006). inequality and happiness: insights from latin america. the journal of economic inequality, 4(1), 107-122. https://link.springer.com/article/10.1007/s1 0888-005-9009-1 https://link.springer.com/article/10.1007/s10902-011-9256-8 https://link.springer.com/article/10.1007/s10902-011-9256-8 https://doi.org/10.1177/0022022100031004001 https://izajoels.springeropen.com/articles/10.1186/2193-9012-3-24 https://izajoels.springeropen.com/articles/10.1186/2193-9012-3-24 https://link.springer.com/article/10.1007/s10888-005-9009-1 https://link.springer.com/article/10.1007/s10888-005-9009-1 liu et al. 81 graham, c., & pettinato, s. (2001). happiness, markets, and democracy: latin america in comparative perspective. journal of happiness studies, 2(3), 237-268. gray, b. t., liu, y., & asebedo, s. d. (2022). household agreement and financial satisfaction: a bargaining perspective. applied economics letters, 29(4), 282-291. gu, y., & wei, h. l. (2018). significant indicators and determinants of happiness: evidence from a uk survey and revealed by a data-driven systems modelling approach. social sciences, 7(4), 53. hayo, b., & seifert, w. (2003). subjective economic well-being in eastern europe. journal of economic psychology, 24(3), 329-348. https://doi.org/10.1016/s01674870(02)00173-3 jardet, c., & meunier, b. (2022). nowcasting world gdp growth with high‐frequency data. journal of forecasting, 41(6), 11811200. https://onlinelibrary.wiley.com/doi/full/10. 1002/for.2858 kahneman, d., diener, e., & schwarz, n. (eds.). (1999). well-being: foundations of hedonic psychology. russell sage foundation. knight, j., song, l., & gunatilaka, r. (2009). subjective well-being and its determinants in rural china. china economic review, 20(4), 635–649. https://doi.org/10.1016/j.chieco.2008.09.003 kollamparambil, u. (2020). happiness, happiness inequality and income dynamics in south africa. journal of happiness studies, 21(1), 201-222. lee, g., lee, j., & kwon, s. (2011). use of social-networking sites and subjective wellbeing: a study in south korea. cyberpsychology, behavior, and social networking, 14(3), 151-155. https://pubmed.ncbi.nlm.nih.gov/20649450/ lim, h., & lee, j. m. (2021). retirement income sources and subjective financial wellbeing: a comparison of retirees and nonretirees. journal of financial counseling and planning, 32(3), 517-534. luttmer, e. f. (2005). neighbors as negatives: relative earnings and well-being. the quarterly journal of economics, 120(3), 963-1002. https://doi.org/10.1093/qje/120.3.963 ngamaba, k. h. (2017). determinants of subjective well-being in representative samples of nations. european journal of public health, 27(2), 377-382. https://pubmed.ncbi.nlm.nih.gov/27426951/ oswald, a. j. (1997). happiness and economic performance. the economic journal, 107(445), 1815-1831. https://doi.org/10.1111/j.14680297.1997.tb00085.x oswald, a. j., & powdthavee, n. (2008). does happiness adapt? a longitudinal study of disability with implications for economists and judges. journal of public economics, 92(5-6), 1061-1077. https://www.sciencedirect.com/science/arti cle/abs/pii/s004727270800008x peng, c., & she, p. w. (2018). raise children to fight against aging: the determinants of elderly wellbeing in today’s china. the singapore economic review, 66(6), 16131645. https://www.worldscientific.com/doi/10.11 42/s0217590818420031 perelli-harris, b., hoherz, s., lappegård, t., & evans, a. (2019). mind the “happiness” gap: the relationship between cohabitation, marriage, and subjective well-being in the united kingdom, australia, germany, and norway. demography, 56(4), 1219-1246. powdthavee, n. (2010). how much does money really matter? estimating the causal effects of income on happiness. empirical economics, 39(1), 77-92. https://link.springer.com/article/10.1007/s0 0181-009-0295-5 rahayu, t. p. (2016). the determinants of happiness in indonesia. mediterranean journal of social sciences, 7(2), 393. https://files.eric.ed.gov/fulltext/ej1346400. pdf https://onlinelibrary.wiley.com/doi/full/10.1002/for.2858 https://onlinelibrary.wiley.com/doi/full/10.1002/for.2858 https://doi.org/10.1016/j.chieco.2008.09.003 https://pubmed.ncbi.nlm.nih.gov/20649450/ https://pubmed.ncbi.nlm.nih.gov/27426951/ https://doi.org/10.1111/j.1468-0297.1997.tb00085.x https://doi.org/10.1111/j.1468-0297.1997.tb00085.x https://www.sciencedirect.com/science/article/abs/pii/s004727270800008x https://www.sciencedirect.com/science/article/abs/pii/s004727270800008x https://www.worldscientific.com/doi/10.1142/s0217590818420031 https://www.worldscientific.com/doi/10.1142/s0217590818420031 https://link.springer.com/article/10.1007/s00181-009-0295-5 https://link.springer.com/article/10.1007/s00181-009-0295-5 https://files.eric.ed.gov/fulltext/ej1346400.pdf https://files.eric.ed.gov/fulltext/ej1346400.pdf financial services review, 32(3) 82 rubio, a., rubio, a., & moreno, j. f. (2022). factors associated with financial ratios and financial well-being of hispanic households: a comparison with white households. journal of financial counseling and planning, 33(1), 126-139. sacks, d. w., stevenson, b., & wolfers, j. (2010). subjective well-being, income, economic development and growth (no. w16441). national bureau of economic research. https://www.nber.org/system/files/working _papers/w16441/w16441.pdf senasu, k., & singhapakdi, a. (2018). qualityof-life determinants of happiness in thailand: the moderating roles of mental and moral capacities. applied research in quality of life, 13(1), 59-87. https://pluma.sjfc.edu/login?url=https://ww w.proquest.com/scholarly-journals/qualitylife-determinants-happinessthailand/docview/2918664008/se2?accountid=27700 shields, m. a., & price, s. w., (2005). exploring the economic and social determinants of psychological well-being and perceived social support in england. journal royal statistical society (part 3), 513–537, https://doi.org/10.1111/j.1467985x.2005.00361.x shields, m. a., price, s. w., & wooden, m. (2009). life satisfaction and the economic and social characteristics of neighbourhoods. journal of population economics, 22(2), 421–443. https://link.springer.com/article/10.1007/s0 0148-007-0146-7 steptoe, a., deaton, a., & stone, a. a. (2015). subjective wellbeing, health, and ageing. the lancet, 385(9968), 640-648. https://pubmed.ncbi.nlm.nih.gov/25468152/ subramanian, s. v., kim, d., & kawachi, i. (2005). covariation in the socioeconomic determinants of self rated health and happiness: a multivariate multilevel analysis of individuals and communities in the usa. journal of epidemiology & community health, 59(8), 664-669. tokuda, y., & inoguchi, t. (2008). interpersonal mistrust and unhappiness among japanese people. social indicators research, 89(2), 349–360. https://link.springer.com/article/10.1007/s1 1205-007-9235-y yang, y. (2008). social inequalities in happiness in the united states, 1972 to 2004: an ageperiod-cohort analysis. american sociological review, 73(2), 204-226. https://www.nber.org/system/files/working_papers/w16441/w16441.pdf https://www.nber.org/system/files/working_papers/w16441/w16441.pdf https://pluma.sjfc.edu/login?url=https://www.proquest.com/scholarly-journals/quality-life-determinants-happiness-thailand/docview/2918664008/se-2?accountid=27700 https://pluma.sjfc.edu/login?url=https://www.proquest.com/scholarly-journals/quality-life-determinants-happiness-thailand/docview/2918664008/se-2?accountid=27700 https://pluma.sjfc.edu/login?url=https://www.proquest.com/scholarly-journals/quality-life-determinants-happiness-thailand/docview/2918664008/se-2?accountid=27700 https://pluma.sjfc.edu/login?url=https://www.proquest.com/scholarly-journals/quality-life-determinants-happiness-thailand/docview/2918664008/se-2?accountid=27700 https://pluma.sjfc.edu/login?url=https://www.proquest.com/scholarly-journals/quality-life-determinants-happiness-thailand/docview/2918664008/se-2?accountid=27700 https://doi.org/10.1111/j.1467-985x.2005.00361.x https://doi.org/10.1111/j.1467-985x.2005.00361.x https://link.springer.com/article/10.1007/s00148-007-0146-7 https://link.springer.com/article/10.1007/s00148-007-0146-7 https://pubmed.ncbi.nlm.nih.gov/25468152/ https://link.springer.com/article/10.1007/s11205-007-9235-y https://link.springer.com/article/10.1007/s11205-007-9235-y financial services review, 32(1) 63 performance evaluation disagreement: determinants and impact on fund flows stéphane chrétien1 and manel kammoun2 abstract this paper studies investor disagreement in the performance evaluation of equity mutual funds by comparing two existing approaches and estimating its relations with fund characteristics, active management level and fund flows. we find that investors disagree more about the performance of funds that have higher expense ratio and turnover, lower manager tenure and dividend yield, and that are older, smaller and part of a larger family. disagreement is also higher for funds that follow riskier investment style strategies and deviate more from their benchmarks. finally, larger disagreement leads to more net fund flows. these findings suggest that heterogeneous investors do not value funds with aggressive active trading strategies similarly, and that favorable valuations by some clienteles result in positive demands for this type of management. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation chrétien, s., & kammoun, m. (2024). performance evaluation disagreement: determinants and impact on fund flows. financial services review, 32(1), 63-94. introduction the 2021 investment company fact book reports that total worldwide assets invested in regulated open-ended funds are greater than $60 trillion and demand by investors over the past decade has resulted in more than $16 trillion in net fund flows. in consequence, “fund providers have responded to the increasing interest in funds by offering more than 125,000 regulated funds, which provide a vast array of choices for investors.” (investment company institute 2021, p. 15). one reason why the fund industry offers such a large variety of products is to cater to the multiple needs of their various clienteles. in equity portfolios, differentiation strategies have many 1 laval university, quebec city, canada 2 corresponding author (manel.kammoun@uqo.ca). université du québec en outaouais, saint-jérôme, canada dimensions. for examples, they include choices of investment style (value, growth, small-cap, large-cap, etc.), trading activity (turnover, deviations from benchmarks, etc.), risk level (from defensive to aggressive), managerial activity (stock picking, market timing, etc.), clientele-specific needs (low cost, high dividend, tax efficiency, individual versus institutional, etc.), and fund organization (small versus large family, etc.). in the u.s. alone, mutual funds are owned by more than 100 million individuals, who have different beliefs, constraints and preferences. these differences, along with the large diversity of funds, mean that investors are likely to disagree on what investments are worth the most to them, i.e., their “personal favorites” or “best https://creativecommons.org/licenses/by-nc/4.0/ mailto:manel.kammoun@uqo.ca https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(1) 64 fits”. pioneered by ferson and lin (2014), the investor disagreement approach assumes that fund evaluation differs by investors and develops ways to measure the importance of disagreement on performance evaluation. its empirical findings show a large disagreement effect for equity funds. in particular, ferson and lin find that the effect of investor heterogeneity on alpha could be as important as the well documented effects of the benchmark choice or statistical imprecision of alpha estimates. using another strategy to obtain performance bounds, chrétien and kammoun (2017) find that a positive alpha exists for some clienteles for most funds and that disagreement is large enough to change the average alpha of the fund industry from negative to positive, depending on the clienteles. given that investor disagreement is one of the current important challenges in fund evaluation (ferson, 2010), this paper aims to understand better the reasons for disagreement and its effects on behavior. specifically, we provide an in-depth characterization of disagreement in equity mutual funds by comparing both existing measurement strategies and documenting the relations between disagreement and fund characteristics, active management level and fund flows. our analysis has three steps. first, we propose a unified framework that allows for heterogeneity in beliefs and preferences, and in which existing strategies for measuring disagreement can be reconciled. although ferson and lin (2014) and chrétien and kammoun (2017) provide two different bounds on investor disagreement because they rely on different restrictions on stochastic discount factors, we derive a constraint based on the no-good-deal condition of cochrane and saá-requejo (2000) that implies their equivalence. second, we develop testable hypotheses on the signs of the relations between disagreement and variables capturing fund characteristics, active management level and fund flows. to obtain economic predictions, we exploit our theoretical results on the sources of disagreement and the findings from the literature on the link between performance and these variables. third, we empirically document the sign and significance of the relations and examine if they are consistent with our hypotheses. our tests rely mainly on the estimation of standard panel regressions. we use a sample of 2791 actively-managed open-ended u.s. equity mutual funds with returns from 1984 to 2016 to estimate disagreement with the generalized method of moments. we show that our results are similar for both disagreement measures, and are robust to various regression specifications and methodological choices. what types of funds are the most subject to disagreement? our empirical results find significant relations between disagreement and numerous fund characteristics. investors disagree more on their evaluation of funds with higher expenses, turnover, longevity, management fees and cost of bundled services, lower manager tenure, size and dividend yield, and that are part of larger fund complexes and follow riskier investment style strategies. thus, heterogeneous investors do not value similarly funds that represent somewhat risky financial products (i.e., funds that are small, with young managers and with an aggressive and costly active trading strategy) that are well supported by their organization (i.e., funds with a long existence and within a large family). what is the effect of the level of active management on disagreement? by taking active risk, managers can construct portfolios that differ greatly for their benchmarks. such relatively unique opportunities allow for greater disagreement by heterogeneous investors who price differently the part of returns not easily replicated by passive portfolios. our results confirm that future disagreement is positively related to two variables aimed to capture relevant departures from benchmarks by active managers: active share (cremers & petajisto, 2009) and asset selectivity (amihud & goyenko, 2013). average disagreement is more than twice as large for funds in the top versus bottom deciles of asset selectivity or active share. what is the impact of disagreement on the net demands for funds? high disagreement is associated with large valuation discrepancies. favorable evaluations should lead investors to large demands for funds, but unfavorable ones should lead to no demand because investors cannot sell the funds short (ferson & lin, 2014). our results support this intuition by finding a chrétien & kammoun 65 positive and statistically significant relation between future net fund flows and disagreement. a one standard deviation increase in disagreement leads to an approximate rise in net fund flows of 0.80% over the next quarter. thus, favorable evaluations by some clienteles could explain the positive demands for funds with aggressive and costly active trading strategies. this paper contributes to the growing evidence on investor heterogeneity and clientele effects in mutual funds from studies focusing on specific kinds of clienteles. for examples, these effects are related to investor monitoring and investment advice (james & karceski, 2006, bergstresser et al., 2009, evans & fahlenbrach, 2012, del guercio and reuter, 2014), taxation (ivković and weisbenner, 2009, sialm and starks, 2012, sialm and zhang, 2020), liquidity and dividend demands (nanda et al., 2000, harris et al., 2015), demographics and investor sophistication (bailey et al., 2011, evans and fahlenbrach, 2012), and behavioral biases (barber et al., 2005, bailey et al., 2011, massa and yadav, 2015, kronlund et al., 2021). many of these studies can be categorized as using a bottom up analysis, since they start from a specific clientele to examine its effects on funds. our paper follows instead a distinctive top down analysis, as we use aggregate measures of disagreement that implicitly consider multiple clienteles. while the focus of ferson and lin (2014) and chrétien and kammoun (2017) is on developing these measures and finding their implications for performance evaluation, our paper is the first to provide a comprehensive examination of the determinants and impact on flows of aggregate disagreement. this paper also adds to the literature on understanding the effects of fund characteristics and investment strategies, and the reasons why money flows into and out of funds. ippolito (1989, 1992), elton et al. (1993), gruber (1996), carhart (1997) and sirri and tufano (1998) are early works in this literature. recent articles include ferreira et al. (2012), barber et al. (2016), 3 in this setup, assuming unbiased beliefs, 𝑚𝑖 corresponds to the marginal preference of the investor. if beliefs are biased, then 𝑚𝑖 represents a modified sdf which contains an adjustment for biased beliefs. pastor et al. (2015, 2017), phillips et al. (2018), song (2020), and ben-david et al. (2022). our paper shows that investor disagreement is important to consider for these issues since it is influenced by fund characteristics and investments strategies, and it predicts increased net fund flows. we proceed as follows. first, we develop our theoretical framework for measuring disagreement. second, we discuss the relevant literature to develop hypotheses on the relations between disagreement and fund characteristics, active management level and fund flows. third, we describe the methodology and data for estimation. fourth, we present our empirical results and assess their robustness. fifth, concluding remarks are provided. a framework for investor disagreement in performance evaluation this section develops a framework for measuring maximum disagreement that encompasses two strategies available in the literature. first, we define generally the measurement of investor disagreement in performance evaluation and present the total disagreement obtained from the best and worst clientele alphas of chrétien and kammoun (2017) and from the bound on disagreement with a traditional alpha proposed by ferson and lin (2014). second, we relate both disagreement approaches to highlight their differences and obtain a condition for their equivalence. total investor disagreement measures in performance evaluation our framework uses the stochastic discount factor (sdf) approach developed by glosten and jagannathan (1994) and chen and knez (1996), extended to consider potentially biased investor beliefs, to measure the performance, or (average) alpha, such that: 𝛼𝑀𝐹,𝑖 = 𝐸[𝑚𝑖 𝑅𝑀𝐹] − 1, (1) where 𝑚𝑖 is the sdf of an investor 𝑖 interested in valuing the mutual fund with gross return 𝑅𝑀𝐹.3 bondarenko (2003) provides an analysis of the implications of biased beliefs for sdfs. he finds an equivalence relationship between preferences and beliefs, so that the same prices (or alphas in our setup) financial services review, 32(1) 66 according to ferson (2010), the sdf approach is on the most solid theoretical footing, as it does not require assumptions about utility functions or complete markets and can account for clientele effects and informed managers. most performance studies assume unbiased beliefs and a parametric asset pricing model with a representative investor to obtain a unique sdf for evaluation. the investor disagreement approach assumes incomplete markets, resulting in a multiplicity of sdfs and alphas. using the terminology of ferson and lin (2014), 𝑚𝑖 is a client-specific sdf and 𝛼𝑀𝐹,𝑖 is the corresponding client-specific alpha.4 without further assumptions, chen and knez (1996) demonstrate that there could be an infinite range of alphas in this setup. the investor disagreement approach imposes economically relevant restrictions on sdfs of all investors to obtain a restricted set and identify the fund’s most favorable alpha, �̅�𝑀𝐹, and least favorable alpha, 𝛼𝑀𝐹. from these extreme alphas, this paper defines straightforwardly a bound on total investor disagreement as: 𝐷𝐼𝑆𝑀𝐹 = �̅�𝑀𝐹 − 𝛼𝑀𝐹. (2) our first measure of investor disagreement, denoted by disck, uses the best and worst clientele alphas proposed by chrétien and kammoun (2017, 2020). their idea is to restrict the set of all investor sdfs by imposing two economic restrictions: the law-of-one-price (lop) condition (hansen & jagannathan, 1991), which assumes that investors give zero performance to passive portfolios, and the nogood-deal condition (cochrane & saá-requejo, 2000), which assumes that investors eliminate can result from either preference choices or biased beliefs (or some combination of the two). an econometrician cannot distinguish between both possibilities unless specific assumptions on beliefs and preferences are made. 4 in their rational expectations equilibrium analysis of mutual funds, berk and green (2004) and berk and van binsgergen (2015, 2017) argue that alpha is not a good measure of ability as skilled managers can extract rents from investors (in the form of fees) to bring fund value added to zero. in this paper, following ferson and lin (2014), alpha represents an investorinvestment opportunities that have too high sharpe ratios (so called good deals).5 let rk be the vector of (gross) passive portfolio returns. without loss of generality, we assume that passive portfolios include a risk-free asset with return 𝑅𝐹, which accounts for cash positions and fixes the sdf mean to a relevant value, 𝐸[𝑚𝑖] = 1/𝑅𝐹 (dahlquist & söderland, 1999). let ℎ be the maximum allowable sharpe ratio. the lop condition implies that 𝐸[𝑚𝑖 rk] = 1. the no-good-deal condition implies that 𝐸[𝑚𝑖 2] ≤ (1 + ℎ 2 ) 𝑅𝐹 2⁄ , or equivalently, 𝜎(𝑚𝑖) ≤ ℎ 𝑅𝐹⁄ . chrétien and kammoun (2017) show that these restrictions allow solutions for the best and worst clientele alphas. by taking the difference between these alphas, we obtain the disck measure: 𝐷𝐼𝑆𝐶𝐾𝑀𝐹 = 2𝑣𝐸[𝑤𝑀𝐹 2 ], (3) where 𝑣 = √ ( (1+ℎ̅2) 𝑅𝐹 2 −𝐸[𝑚𝐿𝑂𝑃 2 ]) 𝐸[𝑤𝑀𝐹 2 ] , (4) 𝑤𝑀𝐹 = 𝑅𝑀𝐹 − c′rk, (5) 𝑚𝐿𝑂𝑃 = a′rk. (6) in these equations, the parameter 𝑣 is an increasing function of the maximum allowable sharpe ratio ℎ. the replication error term 𝑤𝑀𝐹 is the residual from a linear projection of the fund return onto passive portfolio returns. the sdf 𝑚𝐿𝑂𝑃 is the minimum volatility sdf under the lop condition. it is a linear function of passive portfolio returns rk. hansen and jagannathan (1991) show that 𝐸[𝑚𝐿𝑂𝑃 2 ] = (1 + ℎ∗2) 𝑅𝐹 2⁄ , or specific evaluation or personal value added, and is thus not a general measure of skill or value added. 5 hansen and jagannathan (1991) and cochrane and saá-requejo (2000) also consider the no-arbitrage (na) condition that excludes negative sdfs by ruling out arbitrage opportunities. however, this condition imposes negligible restrictions on sdfs in the evaluation of equity funds. ahn et al. (2009) find that performance bounds under the lop and na conditions are typically wide. chrétien and kammoun (2017) show that empirical sdfs are almost always positive under the lop and no-good-deal conditions. chrétien & kammoun 67 equivalently, 𝜎(𝑚𝐿𝑂𝑃) = ℎ∗ 𝑅𝐹⁄ , where ℎ∗ is the maximum sharpe ratio obtained from the passive portfolios. thus, in the disck measure, disagreement is greater if investors are willing to allow more good deals and if fund returns are more difficult to span with passive portfolio returns. our second measure of investor disagreement, denoted by disfl, is derived from the bound on disagreement with a traditional alpha proposed by ferson and lin (2014). their approach also assumes the lop condition, but it does not impose the no-good-deal condition. instead, it assumes a restriction on the correlations that sdfs can have, specifically, |𝜌𝑚𝑖,𝜀𝑀𝐹 𝜌𝑚𝑖,𝑅∗⁄ | ≤ 1, where 𝑅∗ is the passive portfolio return that achieves the maximum sharpe ratio ℎ∗ and 𝜀𝑀𝐹 is the error term in a linear regression of excess fund return, 𝑅𝑀𝐹 − 𝑅𝐹, on excess passive portfolio returns: 𝑅𝑀𝐹 − 𝑅𝐹 = 𝑎𝑀𝐹 + b′(rk− − 𝑅𝐹1) + 𝜀𝑀𝐹. (7) with rk− being the vector of passive portfolio returns excluding the risk-free return, 𝑎𝑀𝐹 being the traditional (jensen’s) alpha that would be obtained if excess returns on the passive portfolios are the benchmark returns in a factor model, b being the vector of factor loadings, and 𝐸[𝜀𝑀𝐹] = 𝐸[ 𝜀𝑀𝐹 rk−] = 0. this restriction means that, for all sdfs, the magnitude of their correlations with the part of fund return not captured by passive portfolio returns is smaller than the magnitude of their correlations with the passive portfolio return with the maximum sharpe ratio. using the implications of the bound of ferson and lin (2014) for the maximum and minimum sdf alphas, we obtain the following disfl disagreement measure: 𝐷𝐼𝑆𝐹𝐿𝑀𝐹 = 2ℎ∗ 𝑅𝐹 𝜎(𝜀𝑀𝐹). (8) this measure indicates that maximum disagreement depends on the maximum sharpe ratio obtained from the passive portfolios, ℎ∗, and the standard deviation of the regression error term, 𝜎(𝜀𝑀𝐹). disagreement is thus greater if the 6 the only other constraint on sdf correlations in the literature is the bound on the autocorrelation of sdfs minimum volatility sdf is higher and if fund return is more difficult to explain with passive portfolio returns. relation between the disagreement measures the disck and disfl measures are general as they do not require complete markets, parametric assumptions on beliefs, and preferences or representative investors. they have different bounds on disagreement because they rely on different restrictions on sdfs. the disck measure uses an exogenous maximum allowable sharpe ratio ℎ. as discussed by chrétien and kammoun (2017), sharpe ratios have a long history in performance studies and there is guidance on ℎ as the no-good-deal restriction has been used in various contexts (e.g., ross, 1976; mackinlay, 1995; cochrane & saá-requejo, 2000; pettenuzzo et al., 2014). in contrast, the disfl measure avoids the specification of ℎ, but imposes a constraint on the correlations that sdfs can have. this restriction is difficult to interpret economically and is not made elsewhere in the literature.6 to understand better the relation between the measures, we can rewrite the disck measure in a way that is more comparable to the disfl measure by making two changes. first, the projection error term 𝑤𝑀𝐹 and the regression error term 𝜀𝑀𝐹 are similar since we include a (constant) risk-free return in the passive portfolio returns used in the projection. hence, 𝐸[𝑤𝑀𝐹 2 ] = 𝐸[𝜀𝑀𝐹 2 ] = 𝜎2(𝜀𝑀𝐹). second, since 𝐸[𝑚𝐿𝑂𝑃 2 ] = (1 + ℎ∗2) 𝑅𝐹 2⁄ , the parameter 𝑣 can be written as: 𝑣 = √ ( (1+ℎ̅2) 𝑅𝐹 2 − (1+ℎ∗2 ) 𝑅𝐹 2 ) 𝐸[𝑤𝑀𝐹 2 ] = √ (ℎ̅2−ℎ∗2) 𝑅𝐹 2 𝐸[𝑤𝑀𝐹 2 ] . (9) using these results, we can rewrite the disck measure as follows: 𝐷𝐼𝑆𝐶𝐾𝑀𝐹 = 2𝑣𝐸[𝑤𝑀𝐹 2 ] = 2√(ℎ̅2−ℎ∗2) 𝑅𝐹 𝜎(𝜀𝑀𝐹) = √(ℎ̅2−ℎ∗2) ℎ∗ 𝐷𝐼𝑆𝐹𝐿𝑀𝐹. (10) of chrétien (2012), which restricts the admissible economic time variation across two periods. financial services review, 32(1) 68 this expression clarifies the relation between the disck and disfl measures. when ℎ = √2 ℎ∗, the measures are equivalent. when ℎ > √2 ℎ∗ (ℎ < √2 ℎ∗), 𝐷𝐼𝑆𝐶𝐾𝑀𝐹 > 𝐷𝐼𝑆𝐹𝐿𝑀𝐹 (𝐷𝐼𝑆𝐶𝐾𝑀𝐹 < 𝐷𝐼𝑆𝐹𝐿𝑀𝐹). thus, the restriction on sdf correlations assumed for the disfl measure has an effect similar to assuming that the maximum allowable sharpe ratio (or maximum allowable sdf standard deviation) is 41.4% higher than the maximum sharpe ratio obtained from the passive portfolios (or the minimum sdf standard deviation). empirically, given the different methodological choices of chrétien and kammoun (2017) and ferson and lin (2014), the results will show that the disck and disfl measures are not equivalent, but are closely related, although the highest estimates can come from either measure, depending on the set of passive portfolios and period used for estimation. investor disagreement across funds: hypotheses development this paper’s main objective is to investigate the relations between investor disagreement and three types of variables: fund characteristics, active management level and fund flows. the literature offers relevant empirical findings to develop hypotheses on these relations in two ways. first, we show disagreement as a difference between upper and lower performance values. the literature on the links between performance and the variables investigated is thus relevant. second, we show that disagreement is larger for funds with returns that are more difficult to replicate. because such returns belong to managers who deviate more from their benchmarks, the literature studying active trading and managerial skills is also useful for hypotheses development. fund characteristics we consider a large number of relevant fund characteristics. this section shortly reviews the literature on these variables to formulate hypotheses on their relations with disagreement. when the literature yields mixed predictions, we rely on the best evidence or economic intuition to determine the likely relations. we discuss the most commonly used characteristics in the first subsection and some other relevant characteristics in second subsection. most common fund characteristics the sign and significance of the relations between the most commonly used characteristics and performance are often not robust across studies. expenses. prather et al. (2004) find a positive relation between expenses and abnormal fund returns, while ippolito (1989) and chen et al. (2004) show that there is no relation between the two variables. in contrast, carhart (1997) and cremers and petajisto (2009) report a negative effect of expenses on performance. turnover. some studies (ippolito, 1989, prather et al., 2004, chen et al., 2004, huang et al., 2011) find no link between turnover and performance. other studies report negative (carhart, 1997, massa & patgiri, 2009) or positive (grinblatt & titman, 1994, pastor et al., 2017) relations. age. according to some authors (prather et al., 2004, chen et al., 2004, huang et al., 2011, ferreira et al., 2012, agnesens, 2013), there is no relation between fund age and performance. however, cremers and petajisto (2009) and massa and patgiri (2009) find that fund age has a positive and statistically significant effect on performance. manager tenure. golec (1996) suggests that manager tenure is generally associated with positive excess returns. prather et al. (2004), however, find no significant relationship between performance and manager tenure. size. carhart (1997), prather et al. (2004) and phillips et al. (2018) argue that there is no relation between performance and size, measured by total net assets (tna). chen et al. (2004), ferreira et al. (2012), pastor et al. (2015) and song (2020) find instead that size tends to have a negative impact on fund returns. given these conflicting results, it is difficult to formulate clear predictions on the relations between these variables and investor disagreement. amihud and goyenko (2013) provide an helpful analysis to clarify our predictions. they study the determinants of the 𝑅2 obtained from a regression of fund returns on chrétien & kammoun 69 returns of benchmarks from a multifactor model. lower 𝑅2 suggests that a fund deviates more from the benchmarks and should correspond to higher 𝐸[𝑤𝑀𝐹 2 ] and 𝜎2(𝜀𝑀𝐹), and thus larger investor disagreement. amihud and goyenko (2013) find that funds with high expense ratio, high turnover, high age, high manager tenure and low tna typically deviate more from their benchmarks. these results suggest the following relations across funds. h1a: there is a positive relation between future disagreement and expenses. h1b: there is a positive relation between future disagreement and turnover. h1c: there is a positive relation between future disagreement and age. h1d: there is a positive relation between future disagreement and manager tenure. h1e: there is a negative relation between future disagreement and size. other relevant fund characteristics the literature identifies many fund characteristics as cross-sectional determinants of performance. we select additional variables that could be relevant given the evidence on investor heterogeneity and clientele effects discussed in the introduction. specifically, we consider fund expense components, tax burden, dividend yield, family size, factor exposures and style as additional possible determinants of investor disagreement. expense components. expenses include management fees, advertising expenses (12b-1) and the costs of bundled services. golec (1996) finds that management fees do not decrease performance, but wermers (2000) documents that they have a negative impact. ferris and chance (1987) report a negative effect of 12b-1 fees on performance. since h1a predicts a positive relation between disagreement and expenses, we expect a similar result between disagreement and management fees or the costs of bundled services. however, by providing information that can help investors in their evaluation, advertising expenses are expected to reduce disagreement. tax burden. some authors (barclay et al., 1998, gibson et al., 2000, bergstresser and poterba, 2002, ivković and weisbenner, 2009, sialm and zhang, 2020) report negative effects of taxes on fund performance. sialm and starks (2012), however, find no significant relationship between performance and taxes. we predict a positive relation between disagreement and tax burden as there is heterogeneity on the effect of taxation on investors. this is consistent with the expected positive relation between disagreement and turnover (h1b). dividend yield. harris et al. (2015) find that buying stocks before dividend payments, called juicing, reduces fund performance. but jiang and sun (2020) report a positive effect of dividend yield on performance. nanda et al. (2000) show a dividend clientele effect since liquidity demands differ across fund investors. given that dividends reduce uncertainty in returns, there should be less disagreement on funds with high dividend yield. family size. many authors (chen et al., 2004, pollet and wilson, 2008, ferreira et al., 2012, agnesens, 2013, jun et al., 2014) find a positive relation between performance and the size of the fund family. but bhojraj et al. (2012) show that this result disappears after regulatory changes (i.e., regulation fair disclosure and the global settlement) and increased scrutiny because of scandals. massa (2003) finds evidence of family driven heterogeneity among funds. in large families, product differentiation for clienteles can lead to niche funds that please or displease investors. we thus predict a positive relation between disagreement and family size. factor exposures. factor exposures consist of loadings on size (𝛽𝑆𝑀𝐵), value (𝛽𝐻𝑀𝐿) and momentum (𝛽𝑈𝑀𝐷), and are helpful to understand factor tilts. gruber (1996) and carhart (1997), among others, show that these factors are relevant in explaining fund returns. the literature offers conflicting explanations on the premium associated with these factors. value, small-cap and winner stocks could be either riskier or more neglected than growth, large-cap and loser stocks. as there is arguably more disagreement in riskier or neglected stocks than in safer or glamour stocks, we expect positive relations between disagreement and factor exposures. financial services review, 32(1) 70 styles. styles are important and highly publicized in the fund industry. barberis and shleifer (2003) discuss the interest of financial service firms to understand style preferences and there is a literature that studies the economic differences between style clienteles (bailey et al. 2011, cronqvist et al., 2015, betermier et al., 2017, chrétien & kammoun, 2019). although different styles could please different clienteles, styles associated with more aggressive investing are more difficult to evaluate. we thus expect disagreement to differ by style, with more disagreement for funds following more aggressive styles (like micro-cap funds) and less disagreement for funds following more defensive styles (like equity income funds). in summary, our hypotheses for these additional variables are as follows. h1f: there is a positive relation between future disagreement and management fees or the costs of bundled services. there is a negative relation between future disagreement and advertising expenses. h1g: there is a positive relation between future disagreement and tax burden. h1h: there is a negative relation between future disagreement and dividend yield. h1i: there is a positive relation between future disagreement and family size. h1j: there are positive relations between future disagreement and factor exposures. h1k: there is a significant relation between future disagreement and styles, with larger disagreement for funds following aggressive styles and smaller disagreement for funds following defensive styles. active management variables many studies argue that performance is significantly related to numerous active management skills (see cremers et al. (2019) for a recent review). a manager with those skills should exploit them to take active positions in his portfolio. however, taking active risk is not rewarded similarly by heterogeneous investors, leading to investor disagreement. we can thus conjecture that disagreement should be positively related to active management skills. we consider two measures of relevant departures from benchmarks to establish this relation. amihud and goyenko (2013) develop asset selectivity, a measure computed as 1 − 𝑅2, with 𝑅2 estimated by regressing fund returns on the returns of benchmarks. cremers and petajisto (2009) propose active share, a measure of the importance of the fund’s active bets, which are deviations of the fund’s stock holdings from those of its benchmark. both measures aim to capture the level of active management as their values increase when the managed portfolio deviates more from benchmarks. they aggregate into one general indicator (all possible managerial skills) since any ability should lead the manager to deviate from the benchmarks. we examine the following hypotheses for asset selectivity and active share. h2a: there is a positive relation between future disagreement and asset selectivity. h2b: there is a positive relation between future disagreement and active share. fund flows many studies (e.g., ippolito (1992), gruber (1996), chevalier & ellison (1997), and sirri & tufano (1998)) find that investors chase past performance. there is also evidence of clientele effects in the flow-performance relationship (sawicki, 2001, del guercio & tkac, 2002, huang et al., 2007, nanda et al., 2009, jun et al., 2014, barber et al., 2016, ben-david et al., 2022). this literature does not relate flows to investor disagreement directly, but ferson and lin (2014) suggest that the effect of heterogeneity on flows is positive. in general, high disagreement means that some investors have high alphas while others have low alphas, leaving the net effect of their demands unclear. however, this ambiguity is resolved when investors face trading constraints. because they cannot sell short funds, those with negative alphas cannot easily act on their evaluation. the demand by those with favorable evaluations should then generate a positive relation between disagreement and net flows. ferson and lin (2014) provide evidence that higher heterogeneity results in higher flows, but chrétien & kammoun 71 they rely on heterogeneity in electricity consumption across u.s. states to compute their admittedly “crude” proxy. this paper uses disagreement estimates to test the following hypothesis on the relation between disagreement and net flows. h3: there is a positive relation between past disagreement and net fund flows. methodology and data this section first presents the methodology for estimating the disagreement measures. then, it discusses the panel regressions used to estimate the relations with fund characteristics, active management levels and net fund flows. it also describes the data and provides summary statistics. estimation of the disagreement measures we estimate the disck and disfl measures by translating their solutions into empirical moments and applying the generalized method of moments (gmm) of hansen (1982). chen and knez (1996) and dahlquist and söderland (1999) are the first to use gmm for sdf performance evaluation. we follow closely chrétien and kammoun (2017) and ferson and lin (2014) for the implementation of their respective measure. let 𝑇 be the number of monthly observations. for the disck measure, similar to chrétien and kammoun (2017, 2020), we use the following set of moments. 1 𝑇 ∑ [(a′rkt)rkt] − 1 = 0𝑇 𝑡=1 , (11) 1 𝑇 ∑ [(𝑅𝑀𝐹𝑡 − c′rkt)rkt] = 0𝑇 𝑡=1 , (12) 1 𝑇 ∑ [(a′rkt) + 𝑣(𝑅𝑀𝐹𝑡 − c′rkt)]2 −𝑇 𝑡=1 (1+ℎ 2 ) 𝑅𝐹 2 = 0, (13) 1 𝑇 ∑ [2𝑣 × (𝑅𝑀𝐹𝑡 − c′rkt)2]𝑇 𝑡=1 − 𝐷𝐼𝑆𝐶𝐾𝑀𝐹 = 0. (14) equations (11) to (14) represent a system of 2𝐾 + 2 moments with 2𝐾 + 2 parameters, where 𝐾 is the number of passive portfolios. the 𝐾 moments in equation (11) allow for the estimation of the parameters a of the minimum volatility sdf under the lop condition, 𝑚𝐿𝑂𝑃𝑡 = a′rkt, by ensuring that it prices correctly the passive portfolio returns rkt. the 𝐾 moments in equation (12) are the orthogonality conditions between the replication error term, 𝑤𝑀𝐹𝑡 = 𝑅𝑀𝐹𝑡 − c′rkt, and the passive portfolio returns, needed to estimate the projection parameters c. the moment in equation (13) imposes the no-good-deal condition to estimate 𝑣, which is restricted to be positive. in this moment, we set ℎ = ℎ∗ + 0.5, following chrétien and kammoun (2017), who show the relevancy of this choice in the literature and empirically. 𝑅𝐹 represents a risk-free rate equivalent and is simply set to one plus the average one-month treasury bill return in our sample, which is 0.2986%. finally, using 𝑤𝑀𝐹𝑡 and 𝑣, we obtain the disagreement estimate 𝐷𝐼𝑆𝐶𝐾𝑀𝐹 with the moment specified by equation (14). for the disfl measure, we use the following set of moments. 1 𝑇 ∑ [(𝑅𝑀𝐹,𝑡 − 𝑅𝐹𝑡) − (𝑎 + b′(rk−t −𝑇 𝑡=1 𝑅𝐹𝑡))] = 0, (15) 1 𝑇 ∑ [((𝑅𝑀𝐹,𝑡 − 𝑅𝐹𝑡) − (𝑎 + b′(rk−t −𝑇 𝑡=1 𝑅𝐹𝑡))) × (rk−t − 𝑅𝐹𝑡)] = 0, (16) 1 𝑇 ∑ [(a′rkt)rkt] − 1 = 0𝑇 𝑡=1 , (17) 1 𝑇 ∑ [(a′rkt)]2 − (1+ℎ∗2 ) 𝑅𝐹 2 𝑇 𝑡=1 = 0, (18) 1 𝑇 ∑ [ 2ℎ𝑎 ∗ 𝑅𝐹 × ((𝑅𝑀𝐹,𝑡 − 𝑅𝐹𝑡) − (𝑎 + b′(rk−t −𝑇 𝑡=1 𝑅𝐹𝑡))) 2 ] − 𝐷𝐼𝑆𝐹𝐿𝑀𝐹 = 0. (19) equations (15) to (19) also represent a system of 2𝐾 + 2 moments with 2𝐾 + 2 parameters. the 𝐾 moments in equations (15) and (16) allow for the estimation of the linear regression in equation (7). equation (17) is the same as equation (11) and estimates the parameters of 𝑚𝐿𝑂𝑃𝑡 = a′rkt by ensuring that it correctly prices the passive portfolio returns. the moment in equation (18) allows the estimation of the maximum sharpe ratio obtained from the passive portfolios ℎ∗. following ferson and lin (2014), we the compute ℎ𝑎 ∗ to adjust ℎ∗ for the bias investigated financial services review, 32(1) 72 by ferson and siegel (2003).7 finally, using the regression error term 𝜀𝑀𝐹𝑡 = (𝑅𝑀𝐹,𝑡 − 𝑅𝐹𝑡) − (𝑎 + b′(rk−t − 𝑅𝐹𝑡)) and ℎ𝑎 ∗ , we obtain the disagreement estimate 𝐷𝐼𝑆𝐹𝐿𝑀𝐹 with the moment specified by equation (19). to construct a full panel useful for the panel regressions proposed in the next section, we estimate disagreement with the previous systems every quarter by using a rolling estimation window made of the previous 60 monthly observations. amihud and goyenko (2013) and others follow a similar strategy to construct their panel data. we check the robustness of the results to this choice by also using a 36-month rolling estimation window. for inferences, we use newey and west (1987) standard errors to account for the autocorrelation and heteroskedasticity in residuals. estimation of the panel regressions to examine the previously stated hypotheses, we run various panel regressions and report the coefficient estimates and their corresponding tstatistics, with standard errors clustered by time and by fund. amihud and goyenko (2013), ferson and lin (2014), doshi et al. (2015) and many others use similar methodologies. to test the hypotheses h1 on the relations between future disagreement and fund characteristics, we regress the disagreement estimates on the previously identified characteristics. to test the hypotheses h2 on the relation between future disagreement and active management variables, we add estimates of asset selectivity or active share to some of the previous regressions. because disagreement is estimated over 60 months, we use non-overlapping periods of 60 months to form a partial panel for these 7 they show that the sample maximum sharpe ratio is biased upward when the number of basis assets (k) is large relative to number of observations (t). they propose a bias correction to obtain an adjusted maximum sharpe ratio given by ℎ𝑎 ∗ = √(ℎ∗)2 (𝑇 − 𝐾 − 2) 𝑇⁄ − 𝐾 𝑇⁄ . 8 the partial panel has six points in time, so that our results are driven mainly by the cross section of funds and not the time series. following amihud and regressions, similar to amihud and goyenko (2013). specifically, we keep observations from the complete panel for which disagreement is estimated independently, starting with the last available disagreement estimates (december 2016) and moving back in time by leaps of five years. characteristics and active management variables are as of the end of the year before the beginning of the 60-month estimation period or the last available observation. hence, we match disagreement estimated for the five-year period ending in december 2016 with data for the explanatory variables in december 2011. we repeat this procedure until the sample beginning, which yields a partial panel that starts with disagreement estimated with data up to december 1991 matched with december 1986 data for the explanatory variables.8 in addition to the regression results, we provide visual representations of the relations by using a sorting procedure. specifically, to obtain further economic insights on our findings for h1 and h2, we categorize funds into deciles based on the average value of their characteristics or active management variables, and examine graphically if average disagreement varies by groups. finally, we test hypothesis h3 by regressing net fund flows on lagged values of disagreement estimates and control variables, using all available quarterly observations in the full panel. the net flow for a given quarter is the percentage growth in tna under management between the beginning and end of the quarter, adjusted for fund return. lagged disagreement estimates are the estimates for the five-year period ending in the previous quarter. control variables include three variables similar to those used by ferson and lin (2014), namely past performance, measured by the lop alpha (which uses 𝑚𝐿𝑂𝑃𝑡 for evaluation), fund return volatility and fund goyenko (2013), doshi et al. (2015) and others, we take characteristics and active management variables at the start of the period used for estimating disagreement to check if observable variables are helpful to determine the types of funds the most subject to future disagreement. our estimation strategy also follows the literature on the relation between characteristics and performance by ignoring the additional error generated from estimating the disagreement measures. chrétien & kammoun 73 return first-order autocorrelation, plus the most common fund characteristics identified earlier and the tax burden, dividend yield and family size variables. data and summary statistics mutual funds this paper uses monthly data on actively managed open-ended u.s. equity mutual funds from the crsp survivor bias free us mutual fund database, for the period from 1984 to 2016. we account for known biases in the crsp fund database. we start in 1984 because elton et al. (2001) and fama and french (2010) show that survivorship bias is problematic in prior years. to deal with back-fill and incubation biases, we eliminate observations before the fund organization date, or funds without a name, with no reported organization date or with tna inferior to $15 million in the first year of entering the database (elton et al., 2001, kacperczyk et al., 2008, and evans, 2010). to ensure that only actively managed open-ended u.s. equity funds are in our sample, we follow the selection criteria of kacperczyk et al. (2008) and chrétien and kammoun (2017).9 we obtain a final sample of 2791 funds. table 1 shows statistics for net-ofexpenses monthly fund returns. on average across funds, the mean monthly fund return (net of fees) is 0.75% and the standard deviation is 5.24%. overall, the sample of fund returns is similar to those in the literature (e.g., kacperczyk et al., 2008 and chrétien & kammoun, 2017). passive portfolios to estimate disagreement, we use three sets of passive portfolios, which allow an examination of the sensitivity of our results. the three sets include the risk-free return (rf) from crsp. our main choice follows chrétien and kammoun (2017) by using ten industry portfolios from kenneth r. french’s website. the classifications are consumer nondurables (nodur), consumer durables (durbl), manufacturing (manuf), energy (enrgy), high technology (hitec), telecommunication (telcm), shops (shops), healthcare (hlth), utilities (utils) and other sectors (others). our second choice uses the same five passive etfs as ferson and lin (2014). their tickers (underlying indices) are spy (s&p 500 index), mdy (s&p mid cap 400 index), ijr (s&p small cap 600 index), qqq (nasdaq 100 index) and iyr (dow jones u.s. real estate index), and their returns come from morningstar. our third choice is the 11 benchmark vanguard index funds proposed by berk and van binsbergen (2015). the tickers (vanguard names) are vfinx (s&p 500 index), vexmx (extended market index), naesx (small-cap index), veurx (european stock index), vpacx (pacific stock index), vviax (value index), vbinx (balanced index), veiex (emerging markets stock index), vimsx (midcap index), visgx (small-cap growth index) and visvx (small-cap value index). 9 specifically, we identify u.s. equity funds by policy codes: cs; strategic insight objective codes: agc, gmc, gri, gro, ing or scg; weisenberger objective codes: g, g-i, agg, gci, gro, ltg, mcg or scg and lipper objective codes: eiei, emn, lcce, lcge, lcve, matc, matd, math, mcce, mcge, mcve, mlce, mlge, mlve, scce, scge or scve. then, we exclude index funds identified by the lipper objective codes sp and spsp, and funds with a name that includes “index”. we also use the database variable “open to investors” to exclude funds that are not open-ended. finally, we keep the funds only if they hold between 80% and 105% in common stocks on average. financial services review, 32(1) 74 table 1. summary statistics for the mutual fund returns mean stddev min max mean 0.750 5.237 -19.963 16.388 stddev 0.306 1.551 5.779 7.766 max 2.097 16.921 0.000 89.667 99% 1.442 10.364 -4.977 41.579 95% 1.146 8.065 -12.810 32.585 90% 1.050 7.072 -14.433 27.063 75% 0.913 5.895 -16.568 18.581 median 0.768 4.917 -19.381 14.088 25% 0.625 4.314 -22.874 11.453 10% 0.434 3.870 -26.297 9.970 5% 0.287 3.491 -28.950 9.076 1% -0.114 1.565 -36.895 5.201 min -4.833 0.150 -100.000 0.493 table 1 presents summary statistics for the monthly returns on 2791 actively managed open-ended u.s. equity mutual funds from january 1984 to december 2016. it shows cross-sectional summary statistics (average (mean), standard deviation (stddev) and selected percentiles) on the distributions of the average (mean), standard deviation (stddev), minimum (min), and maximum (max) for the fund returns in percentage. for each fund, we first compute the average, standard deviation, minimum and maximum for its monthly returns. then, across the 2791 values for each of these statistics, we compute the average, standard deviation and selected percentiles. the three sets of passive portfolios are hereafter identified as 10i, etfs and vanguard. the data for 10i cover the same period as the data for mutual funds, but the data for etfs and vanguard start in 2005 and 2003, respectively, instead of 1984. table 2 presents monthly statistics. the 10i portfolios have mean returns from 0.87% (for consumer durables) to 1.17% (for consumer nondurables), with standard deviations from 3.94% (for utilities) to 6.89% (for high technology). the etfs have mean returns from 0.70% (for the nasdaq 100 index etf) to 1.06% (for the s&p mid cap 400 index etf), with standard deviations from 4.16% (for the s&p 500 index etf) to 7.28% (for the nasdaq 100 index etf). the vanguard funds have mean returns from 0.34% (for the value index) to 0.99% (for the extended market index), with standard deviations from 2.59% (for the balanced index) to 6.78% (for the emerging markets stock index). fund characteristics, active management variables, and net flows the data for fund characteristics, active management variables, net fund flows and other control variables generally come from the crsp fund database. cremers and petajisto (2009), amihud and goyenko (2013) and ferson and lin (2014) provide details on their computation. chrétien & kammoun 75 table 2. summary statistics for the passive portfolio returns passive portfolios mean stddev min max 10i nodur 1.166 4.110 -21.030 14.630 durbl 0.869 6.815 -32.630 42.630 manuf 1.069 4.932 -27.330 17.510 enrgy 1.027 5.380 -18.330 19.030 hitec 0.985 6.892 -26.010 20.780 telcm 0.997 5.076 -16.220 21.340 shpos 1.051 4.902 -28.250 13.280 hlth 1.140 4.659 -20.460 16.470 utils 0.949 3.942 -12.650 11.720 other 0.951 5.165 -23.600 16.420 etfs spy 0.814 4.164 -16.790 10.890 mdy 1.055 5.021 -21.740 14.820 iyr 0.987 6.137 -31.200 29.510 qqq 0.699 7.275 -26.410 24.980 ijr 0.956 5.403 -20.190 17.450 vanguard vfinx 0.941 4.326 -21.727 13.267 vexmx 0.985 5.170 -21.508 15.838 naesx 0.874 5.646 -32.203 18.254 veurx 0.643 5.055 -21.772 14.048 vpacx 0.842 4.233 -16.552 10.393 vviax 0.344 5.491 -18.397 20.735 vbinx 0.667 2.587 -11.593 6.978 veiex 0.664 6.781 -27.667 18.266 vimsx 0.896 5.140 -21.927 14.180 visgx 0.846 6.005 -22.149 20.996 visvx 0.856 5.483 -21.090 19.760 rf 0.299 0.231 0.000 1.000 table 2 presents summary statistics for the monthly returns on three sets of passive portfolios. it shows the average (mean), standard deviation (stddev), minimum (min), and maximum (max) for each passive portfolio. the set 10i includes ten industry portfolios (consumer nondurables (nodur), consumer durables (dur), manufacturing (manuf), energy (enrgy), high technology (hitec), telecommunication (telcm), shops (shops), healthcare (hlth), utilities (utils), and other industries (other)), with data from january 1984 to december 2016. the set etfs includes five exchange traded funds (large-cap (spy), mid-cap (mdy), small-cap (ijr), nasdaq 100 (qqq), and mortgage/real estate (iyr)), with data from january 2005 to december 2016. the set vanguard includes 11 vanguard index funds (vfinx (s&p 500 index), vexmx (extended market index), naesx (small-cap index), veurx (european stock index), vpacx (pacific stock index), vviax (value index), vbinx (balanced index), veiex (emerging markets stock index), vimsx (mid-cap index), visgx (small-cap growth index) and visvx (small-cap value index)), with data from january 2003 to december 2016. all three sets include the risk-free asset (rf) based on the one-month treasury bill. fund characteristics are defined as follows. expenses are measured by the expense ratio, the fraction of total investment that shareholders pay for the fund’s operating expenses. turnover is the minimum of aggregate sales or aggregate purchases of securities divided by the average twelve-month tna of the fund. age is the difference in years between current date and the financial services review, 32(1) 76 date the fund was first offered. manager tenure is the difference in years between the current date and the date when the current manager took control. size is given by tnas. the fund expenses components provided by crsp are management fees, advertising expenses (12b-1) and the costs of bundled services (bundled). tax burden is the weighted average of the tax rates of investors in different income brackets, where the weights correspond to the declared amounts of dividends and capital gains.10 dividend yield is the amount of annual dividends per share paid by the fund, divided by the end-of-year net asset value per share (see harris et al. (2015)). family size is the number of funds in the family to which the fund belongs in each quarter.11 factor exposures are loadings on size (𝛽𝑆𝑀𝐵), value (𝛽𝐻𝑀𝐿), and momentum (𝛽𝑈𝑀𝐷), estimated from regressions of the fund excess returns on the factors over a 60-month window. fund styles are determined as in amihud and goyienko (2013), who consider nine categories: aggressive growth (ag), equity income (ei), growth (g), long-term growth (ltg), growth and income (gi), mid-cap (mc), micro-cap (mrc), small cap (sc) and maximum capital gains (mcg). the crsp fund database provides investment objective codes from three sources: wiesenberger (from 1962 to 1993), strategic insight (from 1993 to 1998) and lipper (since 1998). we assign a fund to one style using the three sources. if no code is available for a period, we assign the style from the previous period. we exclude non-identified funds from the sample. the active management variables are defined as follows. asset selectivity is computed as 1 − 𝑅2, with 𝑅2 estimated by regressing fund returns on the returns of a set of benchmarks, which we assume to be the 10i passive portfolios in our base case. active share is obtained from morningstar and is equal to 1 2 ∑ |𝜔𝐹𝑢𝑛𝑑,𝑖 − 𝜔𝐼𝑛𝑑𝑒𝑥,𝑖|𝑁 𝑖=1 , where 𝜔𝐹𝑢𝑛𝑑,𝑖 − 𝜔𝐼𝑛𝑑𝑒𝑥,𝑖 is the deviation of the fund’s holdings in stock i from those of its main benchmark index. net fund flows are equal to [𝑇𝑁𝐴𝑀𝐹,𝑡 − 𝑇𝑁𝐴𝑀𝐹,𝑡−1𝑅𝑀𝐹,𝑡] 𝑇𝑁𝐴𝑀𝐹,𝑡−1⁄ , where 𝑇𝑁𝐴𝑀𝐹,𝑡 is the fund tna at quarter t and 𝑅𝑀𝐹,𝑡 is the quarterly fund return. other control variables in the fund flow regressions include lagged values of lop alpha, estimated using the 10i passive portfolios, fund return volatility vol, and fund return first-order autocorrelation ar. these variables are computed with a 60month rolling estimation window ending in the previous quarter. table 3 gives quarterly summary statistics for the fund characteristics, active management variables, fund flows and other control variables. the means (across funds and time) for the most common fund characteristics are 1.35% for the expense ratio, 87.22% for the turnover, 13.05 years for fund age, 4.83 years for manager tenure and $1136.94 million for fund size. fund expense components have means of 0.72% for management fees, 0.43% for advertising expenses and 0.64% for the costs of bundled services. means for tax burden, dividend yield and family size are 0.020%, 2.10% and 23.6 funds, respectively. factor exposures have means of 0.38 for size, -0.36 for value and -0.12 for momentum. means for asset selectivity, active share and net fund flows are 14.44%, 78.23% and -1.10%, respectively.12 finally, means for lop alpha, fund return volatility and fund return autocorrelation are -0.14%, 0.30% and 0.072, respectively. 10 we thank clemens sialm for providing time series of the tax rates on dividends (div), short-term capital gains (scg), and long-term capital gains (lcg). we follow sialm and zhang (2020) to compute tax burden. 11 following pollet and wilson (2008), we treat funds with the same management company name as belonging to the same family of funds. 12 active share has a maximum of 211.90%, which is unusually high. it should be below 100% unless a fund has important short selling activities and extreme equity portfolio weights (potentially a sign of derivatives positions). we investigate the data series and find that less than 1% of observations have active share values above 100%. our results are robust to the exclusion of these observations. chrétien & kammoun 77 table 3. summary statistics for the fund characteristics, active management variables, net fund flows, and other control variables mean stddev min max most common fund characteristics expenses (%) 1.350 0.970 0.000 102.440 turnover (%) 87.220 111.330 0.040 9150.000 age (years) 13.053 13.263 0.036 87.460 manager tenure (years) 4.834 5.280 0.074 61.910 fund size (in millions $) 1136.937 4191.458 0.001 109073.000 other relevant fund characteristics management fees (%) 0.720 0.270 0.000 6.670 advertising expenses (%) 0.430 0.400 0.000 1.600 bundled (%) 0.640 0.870 0.000 102.440 tax burden (%) 0.020 0.020 0.000 1.020 dividend yield (%) 2.100 0.760 1.110 4.920 family size (number of funds) 23.558 25.309 1.000 112.000 𝛽𝑆𝑀𝐵 0.378 0.491 -1.322 2.810 𝛽𝐻𝑀𝐿 -0.360 0.585 -4.169 2.624 𝛽𝑈𝑀𝐷 -0.124 0.282 -1.963 1.403 active management asset selectivity (%) 14.440 13.560 0.060 94.500 active share (%) 78.227 13.586 3.141 211.898 fund flows (%) -1.100 21.320 -101.940 3144.950 other control variables lop alpha (%) -0.140 0.390 -6.650 2.770 vol (%) 0.300 0.250 0.000 6.060 ar 0.072 0.142 -0.477 0.896 table 3 presents quarterly summary statistics for the fund characteristics, active management variables, net fund flows and other control variables, using data from january 1984 to december 2016. it shows the average (mean), standard deviation (stddev), minimum (min), and maximum (max) for each variable. the most common characteristics variables include expenses (the annual expense ratio), turnover (the minimum of aggregated sales or aggregated purchases of securities divided by the average twelve-month tna of the fund), age (the number of years since the fund was first offered), manager tenure (the number of years since the current manager took control) and fund size (the tna in millions $). other relevant characteristics variables include management fees, advertising expenses (12b-1 fees), bundled (the cost of bundled services), tax burden (the weighted average of the tax rates of investors in different income brackets, where the weights correspond to the declared amounts of dividends and capital gains), dividend yield (the annual yield of dividend payments by the fund), family size (the number of funds in the family to which the fund belongs in each quarter), and the betas 𝛽𝑆𝑀𝐵, 𝛽𝐻𝑀𝐿 and 𝛽𝑈𝑀𝐷 (the factor exposures on size, value and momentum from the carhart (1997) model). the active management variables are asset selectivity (computed as 1 − 𝑅2, with 𝑅2 estimated by regressing fund returns on the returns of a set of benchmarks) and active share (a measure of the deviations of the fund’ stock holdings from those of its main benchmark). the net fund flow variable is fund flows (the quarter-to-quarter growth in tna). other control variables are lop alpha (the performance based on the minimum volatility sdf and estimated using ten industry passive portfolios), vol (the fund return volatility) and ar (the first-order autocorrelation of fund returns). lop alpha, vol and ar are estimated with a rolling estimation window made to the previous 60 monthly observations. the unit of measure is given in parentheses. financial services review, 32(1) 78 empirical results this section presents the empirical results. first, we compare the disck and disfl disagreement estimates. then, we examine their relations with past fund characteristics or active management variables. finally, we document the impact of investor disagreement on future net fund flows. investor disagreement estimates table 4 reports cross-sectional statistics on the disagreement estimates. the first two columns show results for disagreement estimated every quarter using the 10i passive portfolios and a rolling estimation window made of the previous 60 monthly observations. the disck estimates have a mean of 0.881% and a standard deviation of 0.626%, and the disfl estimates have a mean of 0.889% and a standard deviation of 0.484%. both means are statistically different from zero. these values are similar to the total disagreement implied by the results of ferson and lin (2014) and chrétien and kammoun (2017), who use data from 1984 to 2012. the correlation between both measures is high at 0.936, which is expected given that they are closely related, as demonstrated the theoretical section.13 to assess robustness to the estimation window length, the third and fourth columns report results for disagreement estimated with a window made of the previous 36 monthly observations. the results are qualitatively similar to those using the 60-month window. furthermore, when we investigate the relations between disagreement and fund characteristics, active management level and fund flows, we find that the results are robust to this variation in the estimation window. finally, to check the sensitivity of our results to the choice of passive portfolios, the last four columns of table 4 give results for disagreement estimated using either the etfs or vanguard passive portfolios (and a rolling estimation window of 60 observations). the estimates have lower means when using the alternative sets of passive portfolios, especially when using etfs. this finding suggests that fund returns are easier to span with etfs or vanguard index funds than with industry portfolios. one likely reason for this difference is that data for etfs and vanguard funds start in 2005 and 2003, respectively, instead of 1984 for the 10i portfolios. to nullify the impact of the starting dates, we re-estimate disagreement using the common sample from 2005 to 2016. for the 10i portfolios, we find that the mean disck (disfl) estimate becomes equal to 0.578% (0.678%). the correlations between the different estimates vary from 0.276 to 0.892. hence, estimated in their common sample, disagreement values are closer, but still show important differences. 13 we show that when ℎ = √2 ℎ∗, the measures are equivalent. following chrétien and kammoun (2017), our estimation sets ℎ = ℎ∗ + 0.5 for the disck measure. hence, the disagreement estimates should be similar when ℎ∗ = 1.21. however, following ferson and lin (2014), we use ℎ𝑎 ∗ instead of ℎ∗ for the disfl measure. the bias correction in ℎ𝑎 ∗ increases with 𝐾 and decreases with 𝑇. this implies ℎ∗ > 1.21 for equivalence, with a value closer to 1.21 when 𝑇 increases or when 𝐾 decreases. in the data, the estimated ℎ∗ varies according to the set of passive portfolios and period used for estimation. this variation leads to a non-perfect correlation between the disagreement estimates, and empirical results in which the highest estimates can come from either the disck or the disfl measures. chrétien & kammoun 79 table 4. investor performance disagreement disck (10i) disfl (10i) disck (10i, 36m) disfl (10i, 36m) disck (etfs) disfl (etfs) disck (vanguard) disfl (vanguard) (10i) (10i) (10i, 36m) (10i, 36m) (etfs) (etfs) (vanguard) (vanguard) mean 0.881 0.889 0.908 0.918 0.442 0.461 0.602 0.792 stddev 0.626 0.484 0.648 0.543 0.336 0.392 0.437 0.631 (t-stat) (323.06) (356.33) (346.55) (428.57) (212.96) (203.58) (247.17) (237.12) max 8.554 8.254 9.264 11.126 7.194 10.457 9.108 12.571 99% 3.493 2.380 3.563 3.155 1.838 1.805 2.519 3.008 95% 2.028 1.723 2.117 1.853 0.901 0.982 1.281 1.753 90% 1.564 1.504 1.635 1.571 0.723 0.789 0.999 1.356 75% 1.070 0.947 1.105 0.990 0.512 0.628 0.701 0.923 median 0.708 0.735 0.727 0.737 0.365 0.382 0.499 0.697 25% 0.484 0.637 0.494 0.637 0.269 0.198 0.367 0.452 10% 0.365 0.594 0.372 0.588 0.205 0.092 0.270 0.270 5% 0.314 0.344 0.318 0.343 0.172 0.071 0.218 0.203 1% 0.236 0.318 0.239 0.314 0.08 0.041 0.114 0.100 min 0.014 0.144 0.012 0.069 0.005 0.000 0.006 0.034 table 4 shows statistics on the cross-sectional distribution of monthly performance disagreement estimates. disck is the disagreement from the best and worst clientele alphas proposed by chrétien and kammoun (2017). disfl is the disagreement from the bound with a traditional alpha proposed by ferson and lin (2014). the table provides the mean, standard deviation (stddev) and selected percentiles of the distributions of the disagreement estimates. it also reports the t-statistics (t-stat) on the significance of the mean of the disagreement estimates. in the base cases (columns ‘disck (10i)’ and ‘disfl (10i)’), we estimate disagreement every quarter using ten industry passive portfolios (10i) and a rolling estimation window made to the previous 60 monthly observations. in columns ‘disck (10i, 36m)’ and ‘disfl (10i, 36m)’, the rolling estimation window is the previous 36 monthly observations. in columns ‘disck (etfs)’, ‘disfl (etfs)’, ‘disck (vanguard)’ and ‘disfl (vanguard)’, the estimates use either the etf passive portfolios (etfs) or the vanguard index fund passive portfolios (vanguard) (and a rolling estimation window made to the previous 60 monthly observations). the data (see description in tables 1 and 2) cover the period january 1984-december 2016 when using the 10i passive portfolios, january 2005-december 2016 when using the etfs passive portfolios and january 2003-december 2016 when using the vanguard passive portfolios. all statistics are in percentage except the t-statistics. financial services review, 32(1) 80 in the rest of the analysis, we rely mainly on disagreement estimated using the 10i portfolios and an estimation window of 60 months. however, given the differences between estimates obtained from different sets of passive portfolios, we also discuss our findings when using etfs or vanguard funds. in general, the results are robust to the choice of passive portfolios. investor disagreement and fund characteristics table 5 study the relations between investor disagreement and past fund characteristics, with panels a and b focusing on the disck and disfl measures, respectively. we consider six models to investigate the impact of characteristics. for each model, we report the coefficient estimate and t-statistic associated with each included variable, along with the number of observations and r² of the regression. in all models, we find that the results are similar for both disagreement measures. our first model considers the most commonly studied determinants of performance, namely expenses, turnover, age, manager tenure and size. the second model replaces expenses with their components (management fees, advertising expenses (12b-1) and the costs of bundled services). the third model replaces turnover with tax burden as suggested by sialm and zhang (2020). the fourth model considers dividend yield, family size and factor exposures. the fifth model examines fund styles. the last model includes all characteristics except expenses and turnover. table 5 shows that performance disagreement is higher for funds with higher expenses, turnover and age. it is also generally higher for funds with younger managers, although this relation is not robust across models. positive and negative coefficients on log(fund size) and log(fund size2) indicate that investor disagreement is a concave function of fund size (in logarithm). the coefficient values suggest that there is a negative relation between disagreement and size for most funds, except the smallest ones. disagreement is higher for funds with higher management fees and costs of bundled services. there is some evidence that advertising reduces disagreement, although the negative relation is oftentimes not statistically significant. there is also evidence of a positive relation between disagreement and tax burden, although the relation loses its statistical significance once dividend yield, family size and factor exposures are included. funds with higher dividend yields have significantly lower disagreement, supporting the idea that dividends reduce uncertainty in returns. we find a positive relation between disagreement and family size, suggesting that there is family driven heterogeneity among funds that are part of a large family. however, this relation becomes insignificant once style dummies are considered. there is a positive relation between disagreement and exposure to the size factor, but the evidence is mixed for exposures to value and momentum factors. finally, as expected, disagreement is larger for funds following aggressive styles (micro-cap funds, maximum capital gain funds and small-cap funds, growth funds and mid-cap funds) and smaller for funds following defensive styles (equity income funds and growth and income funds). to ensure that these results are not specific to disagreement estimated with the 10i passive portfolios, we examine the relations with estimates using either etfs or vanguard funds. in results not included, we find that the sign and significance of the relations are mostly the same. we can report only two notable differences. first, tax burden is never statistically significant. second, the positive relation between disagreement and family size always stays statistically significant. chrétien & kammoun 81 table 5. relations between future disagreement and fund characteristics panel a. disck model (1) model (2) model (3) model (4) model (5) model (6) expenses 0.2883 (7.41) 0.3875 (14.80) 0.1679 (5.38) turnover 0.0005 (3.54) 0.0003 (1.72) 0.0002 (2.42) log(age) 0.0007 (2.02) 0.0019 (3.59) 0.0002 (0.50) 0.0033 (5.54) -0.0006 (-1.53) 0.0037 (5.47) log(manager tenure) -0.0008 (-3.93) -0.0006 (-2.54) -0.0012 (-4.96) 0.0000 (0.22) -0.0012 (-6.32) -0.0002 (-0.99) log(fund size) 0.0034 (6.80) 0.0018 (6.54) 0.0026 (6.50) 0.0031 (7.59) 0.0000 (0.16) 0.0010 (3.22) log(fund size2) -0.0007 (-5.64) -0.0004 (-5.07) -0.0004 (-4.81) -0.0006 (-6.85) 0.0000 (-0.54) -0.0002 (-2.78) management fees 0.3966 (7.87) 0.5189 (10.66) 0.2934 (5.53) advertising expenses -0.0500 (-2.02) -0.0173 (-0.63) -0.0210 (-0.83) bundled 0.2518 (2.56) 0.4439 (7.59) 0.3404 (5.55) tax burden 1.5639 (2.20) 0.4100 (1.08) 0.2311 (0.93) dividend yield -0.5014 (-9.59) -1.0102 (-12.61) log(family size) 0.0011 (4.72) 0.0003 (1.28) 𝛽𝑆𝑀𝐵 0.0044 (26.35) 0.0060 (16.38) 𝛽𝐻𝑀𝐿 -0.0005 (-0.98) 0.0011 (2.38) 𝛽𝑈𝑀𝐷 0.0009 (0.96) 0.0022 (2.48) style dummies ag 0.0064 (7.53) 0.0139 (12.20) ei 0.0043 (5.35) 0.0127 (12.10) g 0.0087 (12.18) 0.0157 (13.17) ltg 0.0071 (8.48) 0.0124 (11.92) gi 0.0048 (6.13) 0.0128 (12.34) mc 0.0080 (9.80) 0.0140 (13.60) mrc 0.0113 (11.45) 0.0141 (12.04) sc 0.0093 (11.70) 0.0132 (14.13) mcg 0.0096 (11.52) 0.0128 (12.33) n 6681 4492 5230 3492 6681 3492 r2 0.6755 0.7227 0.6706 0.7626 0.7205 0.7929 financial services review, 32(1) 82 table 5. relations between future disagreement and fund characteristics (continued) panel b. disfl model (1) model (2) model (3) model (4) model (5) model (6) expenses 0.2406 (7.11) 0.3332 (17.64) 0.1280 (5.74) turnover 0.0004 (3.45) 0.0002 (1.52) 0.0001 (1.93) log(age) 0.0016 (5.62) 0.0032 (7.25) 0.0012 (4.23) 0.0032 (5.64) 0.0006 (1.87) 0.0033 (5.46) log(manager tenure) -0.0005 (-2.65) -0.0003 (-1.26) -0.0008 (-4.10) 0.0000 (-0.19) -0.0008 (-5.30) -0.0003 (-1.49) log(fund size) 0.0032 (7.51) 0.0016 (6.68) 0.0025 (7.64) 0.0029 (7.03) 0.0001 (0.26) 0.0008 (2.81) log(fund size2) -0.0006 (-6.05) -0.0004 (-5.68) -0.0004 (-5.51) -0.0005 (-6.59) 0.0000 (-0.59) -0.0001 (-2.33) management fees 0.3459 (8.17) 0.4732 (10.18) 0.2517 (5.46) advertising expenses -0.0301 (-1.38) -0.0052 (-0.20) -0.0084 (-0.35) bundled 0.1789 (2.42) 0.3341 (6.83) 0.2299 (5.51) tax burden 1.3342 (2.23) 0.4386 (1.35) 0.2690 (1.13) dividend yield -0.4164 (-8.16) -0.9502 (-10.97) log(family size) 0.0010 (4.39) 0.0001 (0.71) 𝛽𝑆𝑀𝐵 0.0041 (24.23) 0.0060 (14.96) 𝛽𝐻𝑀𝐿 0.0003 (0.56) 0.0019 (3.63) 𝛽𝑈𝑀𝐷 -0.0024 (-2.34) -0.0009 (-0.88) style dummies ag 0.0064 (9.63) 0.0146 (12.05) ei 0.0043 (6.89) 0.0132 (11.96) g 0.0079 (14.87) 0.0163 (12.61) ltg 0.0062 (9.48) 0.0132 (12.36) gi 0.0046 (7.62) 0.0134 (12.06) mc 0.0075 (12.10) 0.0145 (13.38) mrc 0.0103 (12.90) 0.0142 (12.47) sc 0.0085 (13.81) 0.0134 (14.62) mcg 0.0079 (11.73) 0.0135 (12.64) n 6681 4492 5230 3492 6681 3492 r2 0.7458 0.7603 0.7371 0.7960 0.7814 0.8227 table 5 shows the results from the panel regressions of the disagreement estimates disck (panel a) and disfl (panel b) on lagged fund characteristics. we estimate disck and disfl every quarter using ten industry passive portfolios and a rolling estimation window made of the previous 60 monthly observations. we then use non-overlapping periods of 60 months to form partial panels for the regressions. the fund characteristics are as the end of the year before the beginning of the 60-month estimation period or the last available observation if missing. the fund characteristic variables are defined in table 3, except for the nine style dummy variables, which are ag (aggressive growth), ei (equity income), g (growth), ltg (long-term growth), gi (growth and income), mc (midcap), mrc (micro-cap), sc (small cap) and mcg (maximum capital gains). the data cover the period from january 1984 to december 2016. we present the estimated coefficients with their t-statistics in parentheses, and the number of observations (n) and r² of the regressions. chrétien & kammoun 83 figure 1 illustrates if average disagreement varies by groups by categorizing all funds into deciles based on average values of selected characteristics. positive slopes for expenses and turnover (see figures 1a and 1b) and negative slopes for manager tenure and fund size (see figures 1d and 1e) reinforce the results in table 4 and show that the impact of these characteristics can be economically large. average disagreement is at least 45% higher for funds in the top versus bottom deciles of expenses (1.23% versus 0.66%) or turnover (1.17% versus 0.71%), or for funds in the bottom versus top deciles of manager tenure (1.16% versus 0.79%) or size (1.08% versus 0.72%). while table 4 documents positive relations between disagreement and fund age, figure 1c shows u-shape relations in which younger and older funds face more disagreement than middle-aged funds. flat slopes for tax burden (figure 1f) and family size (figure 1h) are consistent with the unreliable significance of these variables in table 4, and the ones for dividend yield (figure 1g) suggest that its negative relation is significant because other variables are in the regressions. figure 1. disagreement for funds grouped by characteristics14 graph a. expenses graph b. turnover graph c. age graph d. manager tenure 14 figure 1 displays the mean monthly disck and disfl disagreement estimates for mutual funds grouped into decile portfolios according to the average value of selected fund characteristics. in graph a, funds are sorted in increasing order of their average expenses. in graph b, funds are sorted in increasing order of their average turnover. in graph c, funds are sorted in increasing order of their average age. in graph d, funds are sorted in increasing order of their average manager tenure. in graph e, funds are sorted in increasing order of their size. in graph f, funds are sorted in increasing order of their average tax burden. in graph g, funds are sorted in increasing order of their average dividend yield. in graph h, funds are sorted in increasing order of their average family size. financial services review, 32(1) 84 figure 1. disagreement for funds grouped by characteristics (continued) graph e. fund size graph f. tax burden graph g. dividend yield graph h. family size in summary, except for manager tenure, results are consistent with the hypotheses developed previously and are robust to different model specifications and sets of passive portfolios. investor disagreement is significantly related to numerous fund characteristics, which can be used to identify the types of funds that can be subject to large discrepancies in evaluation. investor disagreement and active management variables table 6 documents the relations between future investor disagreement and active management variables, controlling for fund characteristics, with panels a and b focusing on the disck and disfl measures, respectively. we consider six regression models. the first three models consider the most commonly studied determinants of performance (as in model 1 of table 5) and add asset selectivity, active share or both. the last three models replace expenses with their components (management fees, advertising expenses (12b-1) and the costs of bundled services), turnover with tax burden, and include dividend yield, family size and factor exposures (as in model 4 of table 5). chrétien & kammoun 85 table 6. relations between future disagreement and active management variables panel a. disck model (1) model (2) model (3) model (4) model (5) model (6) expenses 0.1214 (4.62) 0.0544 (2.40) 0.0503 (2.89) turnover -0.0003 (-1.50) 0.0007 (6.23) 0.0003 (2.66) log(age) 0.0022 (8.31) -0.0012 (-5.20) 0.0002 (0.62) 0.0010 (3.40) 0.0011 (3.19) 0.0006 (2.71) log(manager tenure) -0.0002 (-1.35) -0.0006 (-4.26) -0.0002 (-1.57) 0.0000 (0.05) -0.0001 (-0.65) -0.0000 (-0.20) log(fund size) 0.0006 (1.29) -0.0014 (-5.58) -0.0010 (-4.65) 0.0001 (0.46) 0.0008 (2.91) -0.0000 (-0.02) log(fund size2) -0.0001 (-1.58) 0.0003 (5.08) 0.0002 (4.03) 0.0000 (-0.16) -0.0001 (-2.49) -0.0000 (-0.13) management fees 0.0740 (1.66) 0.1986 (5.21) 0.0641 (2.77) advertising expenses -0.0215 (-1.18) 0.0025 (0.17) 0.0025 (0.20) bundled 0.2310 (4.27) 0.1485 (3.90) 0.0452 (1.78) tax burden 0.2637 (1.48) 0.0106 (0.09) 0.0464 (0.60) dividend yield -0.0817 (-2.04) -0.4310 (-13.36) -0.1430 (-5.44) log(family size) 0.0004 (2.50) 0.0002 (1.82) -0.0000 (-0.32) 𝛽𝑆𝑀𝐵 0.0027 (10.65) 0.0040 (33.15) 0.0031 (24.12) 𝛽𝐻𝑀𝐿 -0.0006 (-1.82) -0.0016 (-7.68) -0.0018 (-11.45) 𝛽𝑈𝑀𝐷 -0.0029 (-7.05) -0.0022 (-6.33) -0.0037 (-13.98) asset selectivity 0.0279 (13.42) 0.0157 (12.37) 0.0278 (9.14) 0.0159 (9.03) active share 0.0113 (22.34) 0.0069 (11.49) 0.0091 (15.52) 0.0044 (5.87) n 5360 3412 3216 3492 2661 2661 r2 0.8115 0.8730 0.8981 0.8582 0.9115 0.9307 financial services review, 32(1) 86 table 6. relations between future disagreement and active management variables (continued) panel b. disfl model (1) model (2) model (3) model (4) model (5) model (6) expenses 0.1032 (4.93) 0.0417 (2.87) 0.0402 (3.84) turnover -0.0002 (-1.42) 0.0005 (4.44) 0.0002 (1.86) log(age) 0.0029 (11.66) 0.0004 (2.02) 0.0010 (3.65) 0.0011 (3.74) 0.0012 (3.45) 0.0008 (3.05) log(manager tenure) 0.0000 (0.23) -0.0003 (-2.26) -0.0001 (-0.75) -0.0001 (-0.44) -0.0001 (-0.400) -0.0000 (-0.03) log(fund size) 0.0011 (3.14) -0.0008 (-3.75) -0.0006 (-2.76) 0.0003 (1.11) 0.0009 (3.22) 0.0003 (1.28) log(fund size2) -0.0002 (-3.43) 0.0002 (3.09) 0.0001 (2.08) 0.0000 (-0.81) -0.0002 (-2.84) -0.0001 (-1.30) management fees 0.0790 (1.97) 0.1837 (4.83) 0.0726 (2.56) advertising expenses -0.0090 (-0.49) 0.0016 (0.11) 0.0017 (0.12) bundled 0.1456 (4.48) 0.1364 (4.14) 0.0510 (2.15) tax burden 0.3090 (1.78) 0.0891 (0.74) 0.1187 (1.23) dividend yield -0.0445 (-1.37) -0.3600 (-11.02) -0.1220 (-4.06) log(family size) 0.0003 (2.26) 0.0002 (1.56) -0.0000 (-0.12) 𝛽𝑆𝑀𝐵 0.0026 (11.63) 0.0039 (29.07) 0.0031 (21.20) 𝛽𝐻𝑀𝐿 0.0001 (0.35) -0.0011 (-4.92) -0.0013 (-7.05) 𝛽𝑈𝑀𝐷 -0.0057 (-11.26) -0.0057 (-14.88) -0.0070 (-22.28) asset selectivity 0.0216 (11.77) 0.0100 (7.40) 0.0246 (8.80) 0.0131 (7.31) active share 0.0097 (22.40) 0.0074 (11.75) 0.0076 (13.37) 0.0037 (5.09) n 5360 3412 3216 3492 2661 2661 r2 0.8261 0.8713 0.8787 0.8606 0.9154 0.9257 table 6 shows the results from the panel regressions of the disagreement estimates disck (panel a) and disfl (panel b) on lagged active management variables (asset selectivity and active share, in bold) and fund characteristics. we estimate disck and disfl every quarter using ten industry passive portfolios and a rolling estimation window made of the previous 60 monthly observations. we then use non-overlapping periods of 60 months to form partial panels for the regressions. the active management variables and fund characteristics are as the end of the year before the beginning of the 60-month estimation period or the last available observation if missing. the active management variables and fund characteristics are defined in table 3. the data cover the period from january 1984 to december 2016. we present the estimated coefficients with their t-statistics in parentheses, and the number of observations (n) and r² of the regressions. chrétien & kammoun 87 we expect a positive relation between future disagreement and the level of active management, since taking active risk should not be rewarded similarly by heterogeneous investors. funds with a higher level of active management differ more from benchmarks. hence, their returns represent relatively “unique” opportunities to investors because they cannot be spanned easily by passive portfolio returns. this uniqueness allows for greater disagreement. table 6 confirms positive and statistically significant relations for both asset selectivity and active share across all models and for both disagreement measures.15 in results not included, we also find that the relations are robust to the choice of passive portfolios. figure 2 illustrates this positive relation by looking at average disagreement for decile portfolios of funds sorted by asset selectivity (figure 2a) and active share (figure 2b). average disagreement is more than twice as large for funds in the top versus bottom deciles of asset selectivity (1.69% versus 0.39% for disck; 1.50% versus 0.51% for disfl) or active share (1.35% versus 0.57% for disck; 1.28% versus 0.65% for disfl). figure 2. disagreement for funds grouped by active management variables16 graph a. asset selectivity graph b. active share investor disagreement and net fund flows table 7 examines the relation between investor disagreement and future net fund flows to understand the effect of differences in evaluation on the net demands for funds. the regressions include controls for three variables used by ferson and lin (2014) as potentially relevant to predict fund flows, namely past performance (captured by the lop alpha), fund return volatility and fund return first-order autocorrelation. models 1, 2, and 3 also consider the most common fund characteristics (as in model 1 of table 5) and add the disck estimates, the disfl estimates or both. models 4, 5 and 6 remove turnover and include tax burden, dividend yield and family size. 15 table 6 also shows that results for most characteristics are robust to the inclusion of active management variables, as they are similar to those documented in table 5. exceptions are turnover (insignificant coefficient in model 1), fund size (disagreement is a convex function of size in models 2 and 3) and the value and momentum factor exposures (significantly negative coefficients in models 4, 5 and 6). 16 figure 2 displays the mean monthly disck and disfl disagreement estimates for mutual funds grouped into decile portfolios according to the average value of their active management variables. funds are sorted in increasing order of their average asset selectivity (graph a) or their average active share (graph b). financial services review, 32(1) 88 table 7. relations between future net fund flows and disagreement model (1) model (2) model (3) model (4) model (5) model (6) lop alpha 4.371 (16.11) 4.3845 (16.18) 4.382 (16.29) 4.790 (15.01) 4.813 (15.02) 4.791 (15.02) disck 1.362 (6.80) 0.075 (0.20) 1.032 (4.70) 0.841 (2.13) disfl 1.6479 (7.65) 1.577 (4.11) 1.036 (4.87) 0.233 (0.62) expenses -0.432 (-2.86) -0.5142 (-3.39) -0.514 (-3.39) -1.480 (-7.71) -1.471 (-7.66) -1.485 (-7.72) turnover 0.002 (2.35) 0.0019 (2.23) 0.002 (2.23) log(age) 0.023 (9.09) 0.0216 (8.57) 0.022 (8.69) 0.004 (1.58) 0.004 (1.43) 0.004 (1.54) log(manager tenure) -0.003 (-1.17) -0.0033 (-1.25) -0.003 (-1.25) -0.004 (-1.40) -0.004 (-1.47) -0.004 (-1.41) log(fund size) -0.015 (-13.93) -0.0157 (-14.21) -0.016 (-14.09) -0.020 (-14.43) -0.020 (-14.48) -0.020 (-14.43) vol -1.429 (-3.28) -1.3328 (-3.28) -1.365 (-3.15) -1.156 (-2.20) -0.790 (-1.77) -1.143 (-2.18) ar 0.009 (1.64) 0.0118 (2.06) 0.012 (2.04) 0.016 (2.56) 0.017 (2.70) 0.017 (2.59) tax burden 4.135 (1.20) 4.236 (1.22) 4.147 (1.21) dividend yield 2.195 (10.73) 2.149 (10.49) 2.182 (10.69) log (family size) 0.009 (5.55) 0.009 (5.37) 0.009 (5.51) n 94255 94255 94255 76719 76719 76719 r2 0.0157 0.0159 0.0159 0.0180 0.0180 0.0180 table 7 shows the results from the panel regressions of net fund flows on lagged disagreement estimates (disck and disfl, in bold) and control variables. net fund flows are defined as the quarter-to-quarter growth in tna in excess of fund returns. every quarter, we compute net fund flows, estimate disck and disfl using ten industry passive portfolios and a rolling estimation window made of the previous 60 monthly observations, and obtain the control variables. we then form the panels for the regressions by matching fund flows in a quarter with the values for the disagreement estimates and the control variables in the previous quarter. control variables include fund characteristics (expenses, turnover, age, fund size, tax burden, dividend yield and family size) and other control variables (lop alpha, vol and ar) defined in table 3. the data cover the period from january 1984 to december 2016. we present the estimated coefficients with their t-statistics in parentheses, and the number of observations (n) and r² of the regressions. chrétien & kammoun 89 as discussed previously, we predict a positive relation between future net fund flows and disagreement. high disagreement is associated with extreme valuations for a fund. investors with highly positive alphas should have high demands for the fund, but those with negative alphas cannot sell the fund short and so should have no demand, leading to a positive relation between net fund flows and disagreement. as expected, and consistent with the findings of ferson and lin (2014), table 7 documents positive and statistically significant relations for both disagreement measures, when included individually.17 the coefficients imply economically important relations. when disagreement rises by one standard deviation, net fund flows increase by 0.85% (for disck) or 0.80% (for disfl) over the next quarter. however, when we consider both measures jointly, their significance decreases as they are strongly correlated. in results not included, we find that these relations are robust to the set of passive portfolios. conclusion the vast array of choices in the mutual fund industry is a strong indication that fund providers consider the multiple needs of investors who have different beliefs, constraints and preferences. because some funds are better “fits” for them, investors are likely to disagree with each other on their evaluation. this paper provides new insights on investor disagreement in equity mutual funds. we develop a theoretical framework that allows for heterogeneity in beliefs and preferences, and use it to highlight the similarities and differences between the strategies of ferson and lin (2014) and chrétien and kammoun (2017) for measuring disagreement. we then study the relations between disagreement and fund characteristics, active management level and fund flows. empirically, we find that funds that are the most subject to disagreement are risky financial products that appear well supported by their organizations. these funds are small with young 17 table 7 also finds that future net flows are positively related to past performance, turnover, age, fund return autocorrelation, dividend yield and family size, and negatively related to expenses, fund size and fund return volatility. hence, money tends to flow into managers. they tend to follow aggressive and costly active trading strategies, with large deviations from their benchmarks. however, they have a long existence and are part of large fund families. we also find that higher disagreement is associated with higher future net fund flows. high disagreement means that some investors have favorable evaluations, and our results are consistent with such evaluations leading to positive demands. overall, our empirical findings thus suggest that these somewhat risky financial products could have their dedicated clienteles. ferson (2010) and ferson and lin (2014) call for more research on investor disagreement and clientele effects in performance evaluation. this paper contributes to the growing evidence on investor heterogeneity and clientele effects in mutual funds by using aggregate measures of disagreement, instead of focusing on specific kinds of clienteles (like individual versus institutional investors, value versus growth investors, taxable versus tax-free investors, etc.). it also shows that disagreement is important for understanding the effects of fund characteristics and investment strategies, and the reasons why money flows into and out of funds. there are many avenues for future research. for examples, it would be insightful to develop new disagreement measures by using restrictions on heterogeneity based on our understanding of specific kinds of clienteles. also it would be important to investigate the particular roles of heterogeneity in beliefs, preferences and constraints in generating aggregate disagreement. finally, it would be interesting to adapt existing fund performance analyses to study disagreement, like whether disagreement is persistent and exploitable in an investment strategy. references ahn, d-h., cao, h. h., & chrétien, s. (2009). portfolio performance measurement: a no arbitrage bounds approach. european financial management, 15, 298−339. funds for which there is a large disagreement and that have high past performance, low expenses, high turnover, low volatility, high autocorrelation and high dividend yield, that are older and smaller, and that belong to a large family. financial services review, 32(1) 90 https://doi.org/10.1111/j.1468036x.2009.00480.x agnesens, j. (2013). a statistically robust decomposition of mutual fund performance. journal of banking and finance, 37, 3867−3877. https://doi.org/10.1016/j.jbankfin.2013.07.02 0 amihud, y., & goyenko, r. (2013). mutual fund’s r2 as predictor of performance. review of financial studies, 26, 667−694. https://doi.org/10.1093/rfs/hhs182 bailey, w., kumar, a., & ng, d. (2011). behavioral biases and mutual fund investors. journal of financial economics, 102, 1−27. https://doi.org/10.1016/j.jfineco.2011.05.002 barber, b. m., odean, t., & zheng, l. (2005). out of sight, out of mind: the effects of expenses on mutual fund flows. journal of business, 78, 2095−2119. https://doi.org/10.1016/j.iref.2014.11.019 barber, b. m., huang, x., & odean, t. (2016). which factors matter to investors? evidence from mutual fund flows. review of financial studies, 29, 2600−2642. https://doi.org/10.1093/rfs/hhw054 barberis, n., & shleifer, a. (2003). style investing. journal of financial economics, 68, 161−199. https://10.1016/s0304-405x(03)00064-3 barclay, m. j., pearson, n. d., & weisbach, m.s. (1998). open-end mutual funds and capitalgains taxes. journal of financial economics, 1998, 49, 3−43. https://doi.org/10.1016/s0304405x(98)00016-6 ben-david, i., li, j., rossi, a., & song, y. (2022). what do mutual fund investors really care about? review of financial studies, 35(4), 1723-1774. https://doi.org/10.1093/rfs/hhab081 bergstresser, d., chalmers, j., & tufano, p. (2009). assessing the costs and benefits of brokers in the mutual fund industry. review of financial studies, 22, 4129−4156. https://doi.org/10.1093/rfs/hhp022 bergstresser, d., & poterba, j. (2002). do after-tax returns affect mutual fund inflows? journal of financial economics, 63, 381−414. https://doi.org/10.1016/s0304405x(02)00066-1 berk, j. b., & green, r. c. (2004). mutual fund flows and performance in rational markets. journal of political economy, 112, 1269−1295. https://doi.org/10.1086/424739 berk, j. b., & van binsbergen, j. h. (2015). measuring skill in the mutual fund industry. journal of financial economics, 118, 1−20. https://doi.org/10.1016/j.jfineco.2015.05.002 berk, j. b., & van binsbergen, j. h. (2017). mutual funds in equilibrium. annual review of financial economics, 9, 147−167. https://doi.org/10.1146/annurev-financial110716-032454 betermier, s., calvet, l. e., & sodini, p. (2017). who are value and growth investors? journal of finance, 72, 5−46. https://doi.org/10.1111/jofi.12473 bhojraj, s., cho, y. j., & yehuda, n. (2012). mutual fund family size and mutual fund performance: the role of regulatory changes, journal of accounting research, 50, 647−684. https://10.1111/j.1475679x.2011.00436.x bondarenko, o. (2003). statistical arbitrage and security prices, review of financial studies, 16, 875−919. https://10.1093/rfs/hhg016 carhart, m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57−82. https://doi.org/10.1111/j.15406261.1997.tb03808.x chen, j., hong, h., huang, m., & kubik, j. (2004). does fund size erode mutual fund performance? the role of liquidity and organization. american economic review, 94, 1276−1302. https://10.1257/0002828043052277 chen, z., & knez, p. j. (1996). portfolio performance measurement: theory and applications. review of financial studies, 9, 511−555. https://doi.org/10.1093/rfs/9.2.511 https://doi.org/10.1111/j.1468-036x.2009.00480.x https://doi.org/10.1111/j.1468-036x.2009.00480.x https://doi.org/10.1016/j.jbankfin.2013.07.020 https://doi.org/10.1016/j.jbankfin.2013.07.020 https://doi.org/10.1093/rfs/hhs182 https://doi.org/10.1016/j.jfineco.2011.05.002 https://doi.org/10.1016/j.iref.2014.11.019 https://doi.org/10.1093/rfs/hhw054 https://10.0.3.248/s0304-405x(03)00064-3 https://doi.org/10.1016/s0304-405x(98)00016-6 https://doi.org/10.1016/s0304-405x(98)00016-6 https://doi.org/10.1093/rfs/hhab081 https://doi.org/10.1093/rfs/hhp022 https://doi.org/10.1016/s0304-405x(02)00066-1 https://doi.org/10.1016/s0304-405x(02)00066-1 https://doi.org/10.1086/424739 https://doi.org/10.1016/j.jfineco.2015.05.002 https://doi.org/10.1146/annurev-financial-110716-032454 https://doi.org/10.1146/annurev-financial-110716-032454 https://doi.org/10.1111/jofi.12473 https://10.0.4.87/j.1475-679x.2011.00436.x https://10.0.4.87/j.1475-679x.2011.00436.x https://10.0.4.69/rfs/hhg016 https://doi.org/10.1111/j.1540-6261.1997.tb03808.x https://doi.org/10.1111/j.1540-6261.1997.tb03808.x https://10.0.4.233/0002828043052277 https://doi.org/10.1093/rfs/9.2.511 chrétien & kammoun 91 chevalier, j., & ellison, g. (1997). risk taking by mutual funds as a response to incentives. journal of political economy, 105, 1167−1200. https://doi.org/10.1086/516389 chrétien, s. (2012). bounds on the autocorrelation of admissible stochastic discount factors, journal of banking and finance, 36, 1167−1200. https://doi.org/10.1016/j.jbankfin.2012.03.00 2 chrétien, s., & kammoun, m. (2017). mutual fund performance evaluation and best clienteles. journal of financial and quantitative analysis, 52, 1577−1604. https://doi.org/10.1017/s002210901700045 x chrétien, s., & kammoun, m. (2019). mutual fund styles and clientele-specific performance evaluation. international journal of economics and finance, 11, 89−116. https://10.5539/ijef.v11n12p89 chrétien, s., & kammoun, m. (2020). representative investors versus best clienteles: performance evaluation disagreement in mutual funds. ssrn working paper. https://papers.ssrn.com/sol3/papers.cfm?abst ract_id=2977656 cochrane, j., & saá-requejo, j. (2000). beyond arbitrage: good-deal asset price bounds in incomplete markets. journal of political economy, 108, 79−119. https://doi.org/10.1086/262112 cremers, m., fulkerson, j., & riley, t. (2019). challenging the conventional wisdom on active management: a review of the past 20 years of academic literature on actively managed mutual funds. financial analysts journal, 75, 8−35. https://doi.org/10.1080/0015198x.2019.1628 555 cremers, m., & petajisto, a. (2009). how active is your fund manager? a new measure that predicts performance. review of financial studies, 22, 3329−3365. https://doi.org/10.1093/rfs/hhp057 cronqvist, h., siegel, s., & yu, f. (2015). value versus growth investing: why do different investors have different styles? journal of financial economics, 117, 333−349. https://doi.org/10.1016/j.jfineco.2015.04.006 dahlquist, m., & söderlind, p. (1999). evaluating portfolio performance with stochastic discount factors, journal of business, 72, 347−383. https://doi.org/10.1086/209618 del guercio, d., & reuter, j. (2014). mutual fund performance and the incentive to generate alpha. journal of finance, 69, 1673−1704. https://doi.org/10.1111/jofi.12048 del guercio, d., & tkac, p. a. (2002). the determinants of the flow of funds of managed portfolios: mutual funds vs. pension funds. journal of financial and quantitative analysis, 37, 523−557. https://doi.org/10.2307/3595011 doshi, h., elkamhi, r., & simutin, m. (2015). managerial activeness and mutual fund performance, review of asset pricing studies, 5 (2), 156-184. https://doi.org/10.1093/rapstu/rav005 elton, e. j., gruber, m. j., & blake, c. r. (2001). a first look at the accuracy of the crsp mutual fund database and a comparison of the crsp and morningstar mutual fund databases. journal of finance, 56, 2415−2430. https://doi.org/10.1142/9789814335409_000 6 elton, e. j., gruber, m. j., das, s., & hlavka, m. (1993). efficiency with costly information: a reinterpretation of evidence from manager portfolios, review of financial studies, 6, 1−23. https://doi.org/10.1093/rfs/6.1.1 evans, r. b. (2010). mutual fund incubation. journal of finance, 65, 1581−1611. https://doi.org/10.1111/j.15406261.2010.01579.x evans, r. b., & fahlenbrach, r. (2012). institutional investors and mutual fund governance: evidence from retail-institutional fund twins. review of financial studies, 25, 3530−3571. https://doi.org/10.1093/rfs/hhs105 fama, e. f., & french, k. r. (2010). luck versus skill in the cross-section of mutual fund returns. journal of finance, 65, 1915−1948. https://doi.org/10.1086/516389 https://doi.org/10.1016/j.jbankfin.2012.03.002 https://doi.org/10.1016/j.jbankfin.2012.03.002 https://doi.org/10.1017/s002210901700045x https://doi.org/10.1017/s002210901700045x https://10.0.21.163/ijef.v11n12p89 https://doi.org/10.1086/262112 https://doi.org/10.1080/0015198x.2019.1628555 https://doi.org/10.1080/0015198x.2019.1628555 https://doi.org/10.1093/rfs/hhp057 https://doi.org/10.1016/j.jfineco.2015.04.006 https://doi.org/10.1086/209618 https://doi.org/10.1111/jofi.12048 https://doi.org/10.2307/3595011 https://doi.org/10.1093/rapstu/rav005 https://doi.org/10.1142/9789814335409_0006 https://doi.org/10.1142/9789814335409_0006 https://doi.org/10.1093/rfs/6.1.1 https://doi.org/10.1111/j.1540-6261.2010.01579.x https://doi.org/10.1111/j.1540-6261.2010.01579.x https://doi.org/10.1093/rfs/hhs105 financial services review, 32(1) 92 https://doi.org/10.1111/j.15406261.2010.01598.x ferreira, m. a., keswani, a., miguel, a. f., & ramos, s. b. (2012). the determinants of mutual fund performance: a cross-country study. review of finance, 17, 483−525. https://doi.org/10.1093/rof/rfs013 ferris, s. p., & chance, d. m. (1987). the effect of 12b-1 plans on mutual fund expense ratios: a note. journal of finance, 42, 1077−1082. https://doi.org/10.2307/2328308 ferson, w. e. (2010). investment performance evaluation, annual review of financial economics, 2, 207−234. https://doi.org/10.1146/annurev-financial120209-134007 ferson, w. e., & lin, j. (2014). alpha and performance measurement: the effect of investor heterogeneity. journal of finance, 69, 1565−1596. https://doi.org/10.1111/jofi.12165 ferson, w. e., & siegel, a.f. (2003). stochastic discount factor bounds with conditioning information. review of financial studies, 16, 567−595. https://doi.org/10.1093/rfs/hhg004 gibson, s., safieddine, a., & titman, s. (2000). taxmotivated trading and price pressure: an analysis of mutual fund holdings. journal of financial and quantitative analysis, 35, 369−386. https://doi.org/10.2307/2676209 glosten, l., & jagannathan, r. (1994). a contingent claims approach to performance evaluation. journal of empirical finance, 1, 133−66. https://doi.org/10.1016/09275398(94)90001-9 golec, j. h. (1996). the effects of mutual fund managers’ characteristics on their portfolio performance, risk, and fees. financial services review, 5, 133−148. https://doi.org/10.1016/s10570810(96)90006-2 grinblatt, m., & titman, s. (1994). a study of monthly fund returns and performance evaluation techniques. journal of financial and quantitative analysis, 29, 419−444. https://doi.org/10.2307/2331338 gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51, 783−810. https://doi.org/10.1111/j.15406261.1996.tb02707.x hansen, l. p. (1982). large sample properties of generalized method of moments estimators. econometrica, 50, 1029−1054. https://doi.org/10.2307/1912775 hansen, l. p., & jagannathan, r. (1991). implication of security market data for models of dynamic economies. journal of political economy, 99, 225−262. http://dx.doi.org/10.1086/261749 harris, l.e., hartzmark, s. m., & solomon, d. h. (2015). juicing the dividend yield: mutual funds and the demand for dividends. journal of financial economics, 116, 433−451. https://doi.org/10.1016/j.jfineco.2015.04.001 huang, j., sialm, c., & zhang, h. (2011). risk shifting and mutual fund performance. review of financial studies, 24, 2575−2616. https://doi.org/10.1093/rfs/hhr001 huang, j., wei, k. d., & yan, h. (2007). participation costs and the sensitivity of fund flows to past performance. journal of finance, 62, 1273−1311. https://doi.org/10.1111/j.15406261.2007.01236.x investment company institute. (2021). 2021 investment company fact book, 61st edition. www.icifactbook.org, ippolito. r. (1989). efficiency with costly information: a study of mutual fund performance. quarterly journal of economics, 104, 1−23. https://doi.org/10.2307/2937832 ippolito, r. (1992). consumer reaction to measures of poor quality: evidence from the mutual fund industry. journal of law and economics, 35, 45−70. http://dx.doi.org/10.1086/467244 ivković, z., & weisbenner, s. (2009). individual investors mutual fund flows. journal of financial economics, 92, 223−237. https://doi.org/10.3386/w14583 james, c., & karceski, j. (2006). investor monitoring and differences in mutual fund performance. journal of banking and finance, 30, https://doi.org/10.1111/j.1540-6261.2010.01598.x https://doi.org/10.1111/j.1540-6261.2010.01598.x https://doi.org/10.1093/rof/rfs013 https://doi.org/10.2307/2328308 https://doi.org/10.1146/annurev-financial-120209-134007 https://doi.org/10.1146/annurev-financial-120209-134007 https://doi.org/10.1111/jofi.12165 https://doi.org/10.1093/rfs/hhg004 https://doi.org/10.2307/2676209 https://doi.org/10.1016/0927-5398(94)90001-9 https://doi.org/10.1016/0927-5398(94)90001-9 https://doi.org/10.1016/s1057-0810(96)90006-2 https://doi.org/10.1016/s1057-0810(96)90006-2 https://doi.org/10.2307/2331338 https://doi.org/10.1111/j.1540-6261.1996.tb02707.x https://doi.org/10.1111/j.1540-6261.1996.tb02707.x https://doi.org/10.2307/1912775 http://dx.doi.org/10.1086/261749 https://doi.org/10.1016/j.jfineco.2015.04.001 https://doi.org/10.1093/rfs/hhr001 https://doi.org/10.1111/j.1540-6261.2007.01236.x https://doi.org/10.1111/j.1540-6261.2007.01236.x http://www.icifactbook.org/ https://doi.org/10.2307/2937832 http://dx.doi.org/10.1086/467244 https://doi.org/10.3386/w14583 chrétien & kammoun 93 2787−2808. https://doi.org/10.1016/j.jbankfin.2005.11.00 3 jiang h., & sun, z. (2020). reaching for dividends. journal of monetary economics, 115, 321−338. https://doi.org/10.1016/j.jmoneco.2019.08.00 3 jun, x., li, m., & shi, j. (2014). volatile market condition and investor clientele effects on mutual fund flow performance relationship. pacific-basin finance journal, 29, 310−334. https://doi.org/10.1016/j.pacfin.2014.05.002 kacperczyk, m., sialm, c., & zheng, l. (2008). unobserved actions of mutual funds. review of financial studies, 21, 2379−2416. https://doi.org/10.1093/rfs/hhl041 kronlund, m., pool, v. k., sialm, c., & stefanescu, i. (2021). out of sight no more? the effect of fee disclosures on 401(k) investment allocations. journal of financial economics, 141, 644−668. https://doi.org/10.1016/j.jfineco.2021.04.008 mackinlay, a. c. (1995). multifactor models do not explain deviations from the capm. journal of financial economics, 38, 3−28. https://doi.org/10.1016/0304405x(94)00808-e massa, m. (2003). how do families affect fund performance? when performance maximization is not the only game in town. journal of financial economics, 67, 249−304. https://doi.org/10.1016/s0304405x(02)00253-2 massa, m., & patgiri, r. (2009). incentives and mutual fund performance: higher performance or just higher risk taking?, review of financial studies, 22, 1777−1815. https://doi.org/10.1093/rfs/hhn023 massa, m., & yadav, v. (2015). investor sentiment and mutual fund strategies. journal of financial and quantitative analysis, 50, 699−727. https://doi.org/10.1017/s0022109015000253 nanda, v., narayanan, m. p., & warther, v. (2000). liquidity, investment ability, and mutual fund structure. journal of financial economics, 57, 417−443. https://doi.org/10.1016/s0304405x(00)00063-5 nanda, v., wang, z. j., & zheng, l. (2009). the abcs of mutual funds: on the introduction of multiple share classes. journal of financial intermediation, 18, 329−361. https://doi.org/10.1016/j.jfi.2008.12.003 newey, w. k., & west, k. d. (1987). a simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. econometrica, 55, 703−708. https://doi.org/10.2307/1913610 pastor, l., stambaugh, r., & taylor, l. (2015). scale and skill in active management, journal of financial economics, 116, 23−45. https://doi.org/10.1016/j.jfineco.2014.11.008 pastor, l., stambaugh, r., & taylor, l. (2017). do funds make more when they trade more? journal of finance, 72, 1483−1528. https://doi.org/10.1111/jofi.12509 pettenuzzo, d., timmermann, a., & valkanov, r. (2014). forecasting stock returns under economic constraints, journal of financial economics, 114, 517−553. https://doi.org/10.1016/j.jfineco.2014.07.015 phillips, b., pukthuanthong, k., & rau, p. r. (2018). size does not matter: diseconomies of scale in the mutual fund industry revisited. journal of banking and finance, 88, 357−365. https://doi.org/10.1016/j.jbankfin.2017.12.00 5 pollet, j. m., & wilson, m. (2008). how does size affect mutual fund behavior? journal of finance, 63, 2941−2969. https://doi.org/10.1111/j.15406261.2008.01417.x prather, l., bertin, w. j., & henker, t. (2004). mutual fund characteristics, managerial attributes, and fund performance. review of financial economics, 13, 305−326. https://doi.org/10.1016/j.rfe.2003.11.002 ross, s. a. (1976). the arbitrage theory of capital asset pricing. journal of economics theory, 13, 341−360. https://doi.org/10.1016/00220531(76)90046-6 https://doi.org/10.1016/j.jbankfin.2005.11.003 https://doi.org/10.1016/j.jbankfin.2005.11.003 https://doi.org/10.1016/j.jmoneco.2019.08.003 https://doi.org/10.1016/j.jmoneco.2019.08.003 https://doi.org/10.1016/j.pacfin.2014.05.002 https://doi.org/10.1093/rfs/hhl041 https://doi.org/10.1016/j.jfineco.2021.04.008 https://doi.org/10.1016/0304-405x(94)00808-e https://doi.org/10.1016/0304-405x(94)00808-e https://doi.org/10.1016/s0304-405x(02)00253-2 https://doi.org/10.1016/s0304-405x(02)00253-2 https://doi.org/10.1093/rfs/hhn023 https://doi.org/10.1017/s0022109015000253 https://doi.org/10.1016/s0304-405x(00)00063-5 https://doi.org/10.1016/s0304-405x(00)00063-5 https://doi.org/10.1016/j.jfi.2008.12.003 https://doi.org/10.2307/1913610 https://doi.org/10.1016/j.jfineco.2014.11.008 https://doi.org/10.1111/jofi.12509 https://doi.org/10.1016/j.jfineco.2014.07.015 https://doi.org/10.1016/j.jbankfin.2017.12.005 https://doi.org/10.1016/j.jbankfin.2017.12.005 https://doi.org/10.1111/j.1540-6261.2008.01417.x https://doi.org/10.1111/j.1540-6261.2008.01417.x https://doi.org/10.1016/j.rfe.2003.11.002 https://doi.org/10.1016/0022-0531(76)90046-6 https://doi.org/10.1016/0022-0531(76)90046-6 financial services review, 32(1) 94 sawicki, j. (2001). investors' differential response to managed fund performance. journal of financial research, 24, 367−384. https://doi.org/10.1111/j.14756803.2001.tb00775.x sialm, c., & starks, l. (2012). mutual fund tax clienteles, journal of finance, 67, 1397−1422. https://doi.org/j.15406261.2012.01751.x sialm, c., & zhang, h. (2020). tax‐efficient asset management: evidence from equity mutual funds. journal of finance, 75, 735−777. https://doi.org/10.1111/jofi.12843 sirri, e. r., & tufano, p. (1998). costly search and mutual fund flows. journal of finance, 53, 1589−1622. https://doi.org/10.1111/00221082.00066 song, y. (2020). the mismatch between mutual fund scale and skill. journal of finance, 75, 2555−2589. https://doi.org/10.1111/jofi.12950 wermers, r. (2000). mutual fund performance: an empirical decomposition into stock‐picking talent, style, transaction costs, and expenses. journal of finance, 55, 1655−1695. https://doi.org/10.2469/dig.v31.n1.836 https://doi.org/10.1111/j.1475-6803.2001.tb00775.x https://doi.org/10.1111/j.1475-6803.2001.tb00775.x https://doi.org/j.1540-6261.2012.01751.x https://doi.org/j.1540-6261.2012.01751.x https://doi.org/10.1111/jofi.12843 https://doi.org/10.1111/0022-1082.00066 https://doi.org/10.1111/0022-1082.00066 https://doi.org/10.1111/jofi.12950 https://doi.org/10.2469/dig.v31.n1.836 using investor utility to determine portfolio choice with reits wei fenga,*, travis l. jonesb,*, marcus t. allenc acollege of business and management, lynn university, 3601 n. military trail, boca raton, fl 33431-5507 blutgert college of business, florida gulf coast university, 10501 fgcu blvd. south, fort myers, fl 33965-6565 clutgert college of business, florida gulf coast university, 10501 fgcu blvd. south, fort myers, fl 33965-6565 abstract this article examines the decision of individual investors to allocate a portion of their existing investment portfolios to reits. it first derives the risk preferences of investors represented by their benchmark portfolios of stocks and bonds. such risk preferences are then used for portfolio decisions regarding reits. the analysis shows that investors with lower risk aversion tend to have a more substantial stock component in their benchmark portfolio and will obtain higher risk-return benefits from adding reits. in addition to the theoretical analysis, the article provides a practical solution to evaluate the benefit of investing in reits. © 2021 academy of financial services. all rights reserved. jel classification: g11 keywords: reits; target fund; efficient frontier; portfolio 1. introduction reits (i.e., equity reits) offer individual investors the ease of investing in real estate using publicly traded shares. historically, reits have provided investors dividend-based income, competitive market performance, transparency, liquidity, inflation protection, and portfolio diversification. they have been advertised as a potential candidate for investment *corresponding author. tel.: +1-239-590-7167. e-mail address: tljones@fgcu.edu (t. l. jones) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 55–66 or portfolio construction that is less correlated with stocks and bonds than other asset classes (see glascock et al., 2000) for further discussion of reits’ correlation with stocks and bonds.). many studies explore the benefits of including reits in investment portfolios, such as anderson and springer (2003), chen et al. (2005), and hudson-wilson et al. (2005). it is necessary to note that reits are generally treated as an alternative asset by many individual investors who hold the majority of their investments in traditional assets like stocks and bonds. to study the benefits of alternative or less traditional assets like reits, we need to examine the investor’s existing holdings and risk preferences. besides, the existing studies that examine the inclusion of reits in investment portfolios use mean-variance analysis but ignore the fact that that investors tend to make different choices due to their diverse risk preferences. it is essential to study the role of risk preference when determining the appropriate asset allocation to reits. this study explores the portfolio implications of risk preference. it starts by examining investors represented by their investment in stocks and bonds. their existing allocations serve as the benchmark portfolio to evaluate the benefit obtained from including reits. we derive the investor’s risk preference from his existing holdings. we further use such a riskpreference to construct a new portfolio by adding reits. we compare this new portfolio with the original benchmark to gauge the benefit of reits. this approach incorporates the role of risk preference in portfolio management. the results suggest that investors with lower risk aversion achieve more benefit from investing in reits. this approach illustrates the role of risk preference in portfolio decisions and provides an explanation of the gap between the theoretical mean-variance framework and diverse portfolio choices in practice. the article is organized as follows. first, it reviews the related studies on reits and portfolio choice. it then introduces the methodology to derive the investor risk preference from a benchmark portfolio allocation. the analysis then applies the risk preference to portfolio construction and explains the additional benefit from the addition of reits. the empirical discussion follows the methodology. the final section concludes. 2. related studies 2.1. reits as an alternative investment class in this study, a “reit” is defined as an equity reit. the data used in the analysis does not include mortgage reits or companies whose primary business is related to financing real estate. equity reits operate along with a straightforward business model. this type of company generates income by leasing space and collecting rent on its real estate and from gains from the sale of its real estate. income is then paid out to shareholders in the form of dividends. when reporting financial results, reits, like other public companies, must report earnings per share based on net income as defined by generally accepted accounting principles. reits are required to distribute at least 90% of their taxable income to shareholders annually in the form of dividends to avoid taxation at the firm level. significantly higher, on average, than other equities, the industry’s dividend yields have historically produced a 56 w. feng et al. / financial services review 29 (2021) 55–66 steady stream of income through a variety of market conditions. reits over time have demonstrated a historical track record of providing a high level of current income combined with long-term share price appreciation, inflation protection, and prudent diversification for investors across the age and investment style spectrums. they have been noted as a useful diversifier for portfolio construction (e.g., hudson-wilson et al., 2003 and chen et al., 2005). 2.2. strategies to invest in reits many studies address the benefit of including reits in a stock and bond portfolio. for example, barone (2016) shows how the inclusion of reits would improve portfolio performance when targeting minimal portfolio variance or maximum sharpe-ratio. such approaches, though often theoretically utilized, cannot explain the diverse portfolio choices in practice. the portfolio diversity stems from the varying risk preference among investors. studies such as waggle and agrrawal (2006), waggle and moon (2006), and bhuyan et al. (2014) introduce risk preferences in the decision process. they assume the utility function u ¼ m – 0.5us2, where m is the portfolio return, s is the portfolio return standard deviation, and u is the risk preference. using unconstrained optimization, they examine the benefit of including reits in stock and bond portfolios for investors with a risk preference (u ) varying from 1 to 10. however, these authors’ choice of risk preference (u ) is random, and their portfolio choice does not present an implementable solution to individual investors. instead of randomly assuming risk preference, this current article derives the risk preference of the investor from commonly used benchmark portfolios and applies such a risk preference to include reits. this study emphasizes that the benefit of investing in reits is contingent upon the allocation of stocks or bonds that the investor gives up for reits. due to the historically low correlation between stocks and bonds, portfolios can generally be constructed with these two asset classes to achieve better diversification and improved risk-adjusted performance than is available when using only one of these asset classes. the stock-bond combination has been a significant theme in investment practice. for example, target-date retirement funds (also known as lifecycle funds, see tiaa (2019) and vanguard (2019) are a popular form of mutual fund that invests in a combination of stocks and bonds. the target fund gradually shifts its asset allocation from stocks to bonds as the target date approaches, and beyond. for instance, a target-date fund intended for people retiring in 30 years might have 90% of its assets in stocks and 10% in bonds, while a fund intended for 5-year retirees may have a 50-50 mix. while the exact asset mix depends on the design from a particular fund company, the underlying rationale for the target fund is that as people get older, they tend to be more risk-averse, which leads to more conservative portfolio choices (singh, 2016; spitzer & singh, 2008). the practice of target-date funds suggests that investors choose different stock-bond allocations suitable to their risk preferences. this study originates from the above observation of target-date funds. it examines how investors, varying in risk preference as reflected in their choices in stock and bond allocation, would invest in reits. we assume investors initially are fully invested in a stock and bond portfolio and are considering adding reits. we derive a risk preference based on the w. feng et al. / financial services review 29 (2021) 55–66 57 investor’s existing (or benchmark) stock-bond allocation and apply this risk preference to construct portfolio construction among stocks, bonds, and reits. the study shows the varying benefit among investors with diverse risk preferences. being advertised as a safe alternative asset, reits bring more benefits to investors with lower risk aversion than those with higher risk aversion. 3. data and analysis the analysis uses the monthly return history from february 1990 to october 2018 of the s&p 500 total return index, the barclays long-term government bond index (lgbi), and the dow jones u.s. select reit index as proxies for an investment of stock, bond, and equity reits, respectively. it is worth noting that the s&p 500 includes 31 reits in that index.1 thus, in the analysis of this article, there is “double dipping” (i.e., an investment in the same company twice) in some reits that are included in both the s&p 500 and the dow jones u.s. select reit index. this fact has been considered by the authors and is a real-world detail faced by individual investors. an investor that desires to allocate a portion of his portfolio to reits will not likely short (or otherwise reduce exposure to) reits in the s&p 500 before adding a reit exposure. we treat an allocation in the s&p 500 as an investment to “stocks,” and an allocation in the reit index as an investment to reits. we feel that this approach is consistent with what an individual investor will likely do in practice and mimics what might happen when an investor faces a choice of mutual funds, etfs, or similar diversified portfolios in which to invest. we use the barclays long-term government bond index as a proxy for the bond component in the target-date fund. this index has been widely adopted as a benchmark for returns from u.s. treasuries. it represents the complementary investment to the equity market. besides the u.s. treasuries, target-date funds also include the investment-grade fund and can allocate a small percentage in high yield corporate bond funds. those corporate bonds feature a risk-return profile that positions between equity and treasury bonds. for the convenience of discussion, we only use the u.s. treasuries to represent the bond component. however, the major conclusions and methodology also apply to situations with additional assets or asset classes. as noted above, the analysis focuses on the equity reits, because mortgage reits represent an investment in real-estate-backed debt and not the actual real estate. the dow jones u.s. select reit index comprises publicly traded equity reits such as owners and/or operators of commercial and/or residential real estate. business excluded from this index includes specialty reits, hybrid reits, mortgage reits, home builders, and companies whose primary business is related to financing real estate. all these indices have actively been tracked with multiple index etfs, which provide a convenient and cost-effective approach to invest in a diversified manner. because this article examines long-term asset allocation, transaction cost incurred to reallocate a portfolio (while potentially important in trading) is a relatively minor issue. this treatment is more justified recently because many of the major brokerages have moved to $0 costs per trade, and 58 w. feng et al. / financial services review 29 (2021) 55–66 individual investors, especially, can reallocate many stocks, etfs, reits, and other funds for $0 commission. fig. 1 illustrates the history of the stock, bond, and reit indices. table 1 includes the statistical details over the return history for those proxy indices. the descriptive return statistics and correlation tables show that stocks and bonds tend to have a low correlation with each other, which provides a convenient approach for portfolio construction. we use a stock and bond mix along the efficient frontier to represent the investor’ existing or benchmark portfolios. investors, especially individuals, generally allocate the majority of their investable assets between stocks and bonds. even if an investor does not invest directly in the indexes used in the analysis, the results can still be applied to a diversified portfolio of stocks and bonds. from the benchmark portfolios, we introduce a utility function and derive the investor’s risk preference. such a risk preference is applied to construct a new portfolio with an allocation to stocks, bonds, and reits. the benefit from including reits is evaluated by comparing the original benchmark portfolio and the new portfolio. 3.1. investor’s risk preference to explain the role of risk preference over the portfolio choice, we model an investor’s utility function as: u ¼ m� :05us 2 (1) where m is the average portfolio return, and s is the standard deviation of portfolio returns. fig. 1. return histories from stock, bond, and reit indices. w. feng et al. / financial services review 29 (2021) 55–66 59 the utility function in equation (1) illustrates the risk-return tradeoff in a straightforward manner. other utility functions give similar results. on an iso-utility curve, the risk-return tradeoff follows: ru ¼ dm=ds ¼ u s (2) the coefficient, u , is a measure of risk aversion. higher u suggests higher risk aversion. the risk-return tradeoff along the utility curve, ru = u s , increases with higher s . the utility function in equation (1) would be latent for investors. however, the portfolio choices would ultimately be dictated by the investors’ inherent risk-preferences. we assume that the investors’ comfortable allocation reflects his best choice between stocks and bonds and achieves maximum utility. then, we can take the existing portfolio as a benchmark and calibrate the risk preference of the investor. 3.2. current benchmark portfolio fig. 2 shows the efficient frontier from stock and bond allocations. given a benchmark portfolio with a profile of (m*, s*; e.g., the 80/20 stock-bond mix) on the efficient frontier, we can show that the investor achieves maximum utility when the iso-utility curve is a tangent to the efficient frontier at point (m*, s*). this condition requires the risk-return tradeoff along the efficient frontier, denoted as rf(m*, s*), equal to tradeoff along the iso-utility curve, or ru = dm/ds . rfðm� , s �þ= ruðm� , s �þ= u �s � (3) table 1 statistics from history returns (feb. 1990 through oct. 2018) panel a return statistics asset stock bond reit monthly average 0.87% 0.64% 0.99% monthly standard 4.07% 2.80% 5.30% skewness �0.61 0.14 �0.69 kurtosis 1.35 1.54 8.23 max drawdown �0.51 �0.16 �0.68 annual average 10.50% 7.64% 11.94% annual standard 14.11% 9.70% 18.36% panel b correlation stock bond reit stock 1.00 �0.11 0.55 bond �0.11 1.00 0.05 reit 0.55 0.05 1.00 note: table 1 reports the return statistics for stock, bond, and reit indices. panel a provides the monthly return average, standard deviations, skewness, kurtosis, and maximum drawdowns. annualized return and standard deviations are also included. panel b shows the correlation between the monthly returns. 60 w. feng et al. / financial services review 29 (2021) 55–66 therefore, we can empirically estimate the risk preference as: u � = rfðm�,s �þ=s � (4) using the same reason, we then apply the derived preference (u *) to construct a portfolio among stocks, bonds, and reits. 3.3. portfolio with reits an individual investor will not (or should not, for diversification reasons) exchange all his existing stock or bond allocation into reits. it is more realistic to assume that he will only allocate a portion of assets into reits. mean-variance optimizations using expected returns on reits and other assets during the 1980s indicate allocations to real estate of 10% to 15% (ennis & burik, 1991). giliberto (1993), using a hedged reit index, finds an optimal allocation to real estate of 19%. therefore, we impose a 20% cap on the allocation to reits and explore the efficient frontier from the combination among stocks, bonds, and reits. in fig. 3, we derive the risk preference u * from the original benchmark portfolio and illustrate the new efficient frontier formed from the addition of reits. the benefit from the inclusion of reits is evaluated as the change in the utility from the benchmark to the new portfolio. du=uðmn,snþ � uðm�,s �þ (5) note that the risk parameter u = u * in the utility function is estimated using equation (4) above. fig. 2. the efficient frontier between stock (s) and bond (b). w. feng et al. / financial services review 29 (2021) 55–66 61 3.4. including reits over different benchmark portfolios we further extend the analysis to allocations along the stock-bond efficient frontier. for each benchmark portfolio, we derive the risk preference (u *). the new portfolio is constructed by adding reits and maximizing the utility of the investor with preference u *. fig. 4 illustrates both the old benchmark allocations and new portfolios along the fig. 3. portfolio choice over efficient frontiers. fig. 4. including reits over benchmark portfolios (reits capped at 20%). 62 w. feng et al. / financial services review 29 (2021) 55–66 efficient frontier curves. because we incorporate the risk preference in portfolio decision, the diversity in portfolio choices is vastly different from the result constructed using variance-minimization and sharpe ratio maximization. the change in risk-return profiles and utility reflects the varying benefits of reits over respective benchmark portfolios. fig. 5 plots the utility improvement over different benchmark allocations and the associated risk preferences (u ). panel a shows a positive relationship between the weight allocated to stocks of the benchmark and improvement in the utility. this result suggests that investors with lower risk aversion (higher benchmark allocation in stocks or lower u ) will achieve more benefit from the inclusion of reits. panel b shows the allocation among stocks, bonds, and reits in the newly formed portfolio. for the extreme case of investors with 35% benchmark stock weight, the new portfolio has a tiny (less than 5%) fig. 5. utility improvement for different risk preferences (reits capped at 20%). w. feng et al. / financial services review 29 (2021) 55–66 63 allocation to reits. on the contrary, the cases with over 55% benchmark stock weight reach the maximum cap (20%) for reits. thus, for any initial stock allocation of 55% or higher, a reit allocation of 20% is beneficial. for any initial stock allocation of 40% to 50%, a reit allocation of between 15% to 20% is helpful. table 2 shows the portfolio statistics from adding reits to a spectrum of benchmark portfolios. the results suggest that while allocating to reits, in general, improves the riskreturn profile over traditional assets, investors with lower risk aversion tend to achieve more benefit. table 2 also provides a guideline for the inclusion of reits. an investor can look directly at table 2, find the row with his current allocation between stock and bonds, and then read that row to the new weights with stocks, bonds, and reits. for example, if the benchmark portfolio is a 60/40 stock-bond split, the investor would look to the 60% stock row, look right and see an appropriate new allocation given this initial allocation would be about 45% stocks, 35% bonds, and 20% reits. the benefit of adding reits to a portfolio is inversely related to investor risk aversion. while the reits have been noted as a safe alternative asset, highly risk-averse investors (higher u or lower benchmark allocation in stocks) with less than 40% in benchmark stock allocation can only get minimum benefit from including reits. meanwhile, less risk-averse investors will receive more significant benefits from adding reits. an investor with an initial stock allocation of 55% or more will experience an increase in utility with the reits allocation quickly reaching the cap allocation of 20%. with 95% or more benchmark allocation to stocks (i.e., the least risk-averse investors), the new portfolio weights include a 20% allocation to reits with no bonds allocation. also, all benchmark stock allocations between 55% and 90% result in a new allocation of 20% in reits with a reduction in benchmark stock allocation of about 15% and a decrease in the bond allocation of about 5%. any benchmark stock allocation below 55% leads to a new portfolio weight in reits of less than the cap of 20% and varying decreases in stock and bond allocations. we also extend the analysis by raising the maximum reits allocation from 20% to 50% and 80%. these tests generate similar conclusions. 4. conclusion this study examines how reits could benefit a traditional investment approach that relies on allocation between stocks and bonds. we derive the investors’ risk preferences from their existing allocations between stocks and bonds. the risk preferences are used in the portfolio choice of adding reits. we compare the original portfolio with the new one to evaluate the benefit of reits across investors with varying risk preferences. the results show that investors with lower risk aversion obtain more benefit from the inclusion of reit into their traditional portfolio of stocks and bonds. this study highlights the role of risk preference in portfolio construction and provides an intuitive and practical approach to evaluate the value of other alternative assets. 64 w. feng et al. / financial services review 29 (2021) 55–66 t ab le 2 p o rt fo li o st at is ti cs fr o m in cl u d in g r e it s (c ap p ed at 2 0 % ) b en ch m ar k p o rt fo li o n ew p o rt fo li o r is k p ar am et er n ew p o rt fo li o w ei g h t b en ch m ar k u ti li ty n ew u ti li ty c h an g e in u ti li ty w g t. st o ck av g 1 st d 1 av g 2 st d 2 th et a s to ck b o n d r e it tu 1 tu 2 tu .c h g 3 5 .0 0 0 .0 0 7 2 1 0 .0 2 1 9 0 0 .0 0 7 2 7 0 .0 2 1 9 0 5 0 .7 0 7 6 0 0 .3 2 8 3 9 0 .6 3 7 8 4 0 .0 3 3 7 7 �0 .0 0 4 9 5 �0 .0 0 4 8 9 0 .0 0 0 0 6 4 0 .0 0 0 .0 0 7 3 3 0 .0 2 2 1 6 0 .0 0 7 4 9 0 .0 2 2 3 7 1 3 .1 0 3 8 0 0 .3 3 9 5 9 0 .5 7 2 8 7 0 .0 8 7 5 4 0 .0 0 4 1 1 0 .0 0 4 2 1 0 .0 0 0 1 0 4 5 .0 0 0 .0 0 7 4 5 0 .0 2 2 7 2 0 .0 0 7 7 0 0 .0 2 3 3 4 7 .5 2 4 0 0 0 .3 5 0 3 4 0 .5 1 0 5 0 0 .1 3 9 1 6 0 .0 0 5 5 1 0 .0 0 5 6 5 0 .0 0 0 1 4 5 0 .0 0 0 .0 0 7 5 6 0 .0 2 3 4 9 0 .0 0 7 9 0 0 .0 2 4 6 7 5 .3 8 4 2 0 0 .3 6 0 6 4 0 .4 5 0 7 4 0 .1 8 8 6 2 0 .0 0 6 0 8 0 .0 0 6 2 6 0 .0 0 0 1 9 5 5 .0 0 0 .0 0 7 6 8 0 .0 2 4 5 5 0 .0 0 8 0 4 0 .0 2 5 8 4 4 .1 2 6 7 0 0 .4 0 2 4 4 0 .3 9 7 5 6 0 .2 0 0 0 0 0 .0 0 6 4 4 0 .0 0 6 6 6 0 .0 0 0 2 3 6 0 .0 0 0 .0 0 7 8 0 0 .0 2 5 8 4 0 .0 0 8 1 7 0 .0 2 7 1 0 3 .3 4 5 4 0 0 .4 5 3 9 7 0 .3 4 6 0 3 0 .2 0 0 0 0 0 .0 0 6 6 9 0 .0 0 6 9 4 0 .0 0 0 2 5 6 5 .0 0 0 .0 0 7 9 2 0 .0 2 7 3 1 0 .0 0 8 2 8 0 .0 2 8 4 4 2 .8 1 2 8 0 0 .5 0 1 8 3 0 .2 9 8 1 7 0 .2 0 0 0 0 0 .0 0 6 8 7 0 .0 0 7 1 4 0 .0 0 0 2 7 7 0 .0 0 0 .0 0 8 0 4 0 .0 2 8 9 4 0 .0 0 8 4 0 0 .0 3 0 0 5 2 .4 2 6 6 0 0 .5 5 3 3 6 0 .2 4 6 6 4 0 .2 0 0 0 0 0 .0 0 7 0 2 0 .0 0 7 3 1 0 .0 0 0 2 8 7 5 .0 0 0 .0 0 8 1 6 0 .0 3 0 7 0 0 .0 0 8 5 3 0 .0 3 1 8 1 2 .1 3 3 5 0 0 .6 0 4 9 0 0 .1 9 5 1 0 0 .2 0 0 0 0 0 .0 0 7 1 5 0 .0 0 7 4 5 0 .0 0 0 2 9 8 0 .0 0 0 .0 0 8 2 8 0 .0 3 2 5 8 0 .0 0 8 6 4 0 .0 3 3 5 4 1 .9 0 3 7 0 0 .6 5 2 7 5 0 .1 4 7 2 5 0 .2 0 0 0 0 0 .0 0 7 2 7 0 .0 0 7 5 7 0 .0 0 0 3 0 8 5 .0 0 0 .0 0 8 4 0 0 .0 3 4 5 5 0 .0 0 8 7 6 0 .0 3 5 5 1 1 .7 1 8 5 0 0 .7 0 4 2 9 0 .0 9 5 7 1 0 .2 0 0 0 0 0 .0 0 7 3 7 0 .0 0 7 6 8 0 .0 0 0 3 0 9 0 .0 0 0 .0 0 8 5 2 0 .0 3 6 6 0 0 .0 0 8 8 8 0 .0 3 7 4 2 1 .5 6 6 2 0 0 .7 5 2 1 4 0 .0 4 7 8 6 0 .2 0 0 0 0 0 .0 0 7 4 7 0 .0 0 7 7 8 0 .0 0 0 3 1 9 5 .0 0 0 .0 0 8 6 4 0 .0 3 8 7 2 0 .0 0 8 9 9 0 .0 3 9 3 9 1 .4 3 8 7 0 0 .8 0 0 0 0 0 .0 0 0 0 0 0 .2 0 0 0 0 0 .0 0 7 5 6 0 .0 0 7 8 7 0 .0 0 0 3 1 1 0 0 .0 0 0 .0 0 8 7 4 0 .0 4 0 6 0 0 .0 0 8 9 9 0 .0 3 9 3 9 1 .3 4 3 8 0 0 .8 0 0 0 0 0 .0 0 0 0 0 0 .2 0 0 0 0 0 .0 0 7 6 3 0 .0 0 7 9 5 0 .0 0 0 3 1 n o te : in t ab le 2 , th e b en ch m ar k p o rt fo li o s ar e co n st ru ct ed b et w ee n st o ck an d b o n d at d es ig n at ed w ei g h t. t h e ri sk p re fe re n ce s o f th e in v es to rs ar e re fl ec te d b y th ei r ch o ic e o f b en ch m ar k p o rt fo li o s u si n g th e u ti li ty fu n ct io n u = m – 0 .5 u s 2 . t h e ri sk p ar am et er (u ) d er iv ed fr o m th e b en ch m ar k p o rt fo li o s is ex te n d ed to th e co n st ru ct io n o f n ew p o rt fo li o co n st ru ct ed w it h st o ck , b o n d , an d r e it s. t h e u ti li ti es fo r b o th th e ex is ti n g b en ch m ar k an d n ew p o rt fo li o s ar e ca lc u la te d fo r co m p ar is o n . w. feng et al. / financial services review 29 (2021) 55–66 65 note 1 the s&p 500 included 31 reits as of february 2020. see https://www.reit.com/ data-research/reit-indexes/reits-sp-indexes for more information. references anderson, r., & springer, t. (2003). reit selection and portfolio construction: using operating efficiency as an indicator of performance. journal of real estate portfolio management, 9, 17-28. bhuyan, r., kuhle, j., ikromov, n., & chiemeke, c. (2014). optimal portfolio allocation among reits, stocks, and long-term bonds: an empirical analysis of us financial markets. journal of mathematical finance, 4, 104-112. chen, h.-c., ho, k.-y., lu, c., & wu, c.-h. (2005). real estate investment trusts. the journal of portfolio management, 31, 46-54. ennis, r. m., & burik, p. (1991). pension fund real estate investment under a simple equilibrium pricing model. financial analysts journal, 47, 20-30. giliberto, s. m. (1993). measuring real estate returns: the hedged reit index. the journal of portfolio management, 19, 94-99. glascock, j. l., lu, c., & so, r. w. (2000). further evidence on the integration of reit, bond, and stock returns. journal of real estate finance and economics, 20, 177-194. hudson-wilson, s., gordon, j. n., fabozzi, f. j., anson, m. j. p., & giliberto, s. (2005). why real estate? the journal of portfolio management, 31, 12-21. singh, s. (2016). the evidence on target-date mutual funds. financial services review, 25, 235-262. spitzer, j. j., & singh, s. (2008). shortfall risk of target-date funds during retirement. financial services review, 17, 143-153. tiaa. (2019). a lifecycle fund can help simplify your retirement investing. available from https://www.tiaa. org/public/offer/insights/investing-101/lifecycle-funds-simplify-retirement-investing vanguard. (2019). vanguard target retirement funds. available from https://investor.vanguard.com/mutualfunds/target-retirement/#/ waggle, d., & moon, g. (2006). mean-variance analysis with reits in mixed asset portfolios – the return interval and the time period used for the estimation of inputs.managerial finance, 32, 955-968. waggle, d., & agrrawal, p. (2006). the stock-reit relationship and optimal asset allocations. journal of real estate portfolio management, 12, 209-221. 66 w. feng et al. / financial services review 29 (2021) 55–66 finser_31-2_complete_issue assessing the relative need and demand for financial education programs: a case study of graduate students brent j. davisa,*, david p. richardsona, jason s. seligmanb atiaa institute, 8500 andrew carnegie boulevard, charlotte, nc 28262, usa binvestment company institute, 1401 h street nw, washington, dc 20005, usa abstract we measure the relative need and demand for financial education programs among graduate students at a large university system. we find that self-assessed and measured financial literacy is significantly related to interest and demand for financial education. individuals who self-report a high level of financial literacy but have low measured financial literacy are significantly less likely to be interested in financial education, while the opposite is true for financially literate individuals selfreporting a low level of financial literacy. our study adds to research showing the importance of both believed and actual financial literacy measures and has implications for financial education programs. © 2023 academy of financial services. all rights reserved. jel classifications: d14; g53 keywords: financial education; financial literacy; overconfidence 1. introduction over the last several decades, there has been a substantial increase in the investment, health, and longevity risk burdens borne by individuals. this shift has increased focus on the importance of household financial literacy in managing these risks and enhancing financial well-being. as noted by lusardi and mitchell (2014), a lack of financial literacy can have long-term impacts on households’ financial well-being. for example, not properly managing *corresponding author. tel.: +1-704-988-6271. e-mail address: brent.davis@tiaa.org (b. j. davis) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 133–150 debt or planning for retirement can have substantial impacts on financial well-being over the lifecycle. while recent research has documented the link between low financial literacy and financial fragility, many adults continue to have poor understanding of basic personal finance concepts.1 however, effectively engaging individuals in improving their financial literacy has proven challenging. in this paper, we conduct a case study on the need and demand for financial education programming by graduate students in a large university system. leveraging the differences between self-assessed and actual financial literacy levels, we find that students most in need of additional financial education are least likely to take advantage of programming. graduate students are good candidates to receive financial education programming. they are among the most receptive consumers of education, command commensurately higher lifetime earnings postgraduation, and may be more likely to have complex financial planning needs over the lifecycle. many also have prior workforce experience, which may expose them to employer-sponsored benefits and retirement savings plans. however, this prior workforce participation may make graduate students more anxious about their financial security due to low earnings and savings and likely increased debt, during graduate studies. providing financial education programming during this time can help students gain confidence in their long-term financial well-being. our interest in studying financial education programming is whether it is economically rational for individuals to invest in financial education, and relatedly, can financial education programs improve outcomes. regarding the latter, evidence increasingly suggests that financial education affects outcomes across a variety of contexts.2 with respect to financial education investment, lusardi, michaud, and mitchell (2017) discuss that investment in financial literacy should be a function of the associated expected benefits. they find that low-education low-lifetime income groups might rationally choose not to invest in obtaining financial literacy skills because of their larger relative reliance on public social welfare programs. because our target sample is of high education and higher-than-median expected lifetime income, they should rationally choose to invest in financial literacy education. we measure self-assessed and tested financial literacy for graduate students in a large public university system. similar to past studies, we find measurable weakness in the populations’ understanding of basic personal finance concepts, despite most students expressing concern regarding their own personal finances. regarding financial education engagement, we document two main findings. the first is that interest in financial education relates significantly to an individual’s self-assessed level of financial knowledge relative to their actual measured financial literacy level. individuals who estimate they have a high level of financial literacy but perform poorly on a financial literacy quiz are significantly less likely to be interested in financial education, while the opposite is true for financially literate students who self-report a low level of financial acumen. we also find that while receptivity to financial education programming increases significantly with financial literacy levels, overall engagement in financial education is low, despite nearly half of survey respondents signaling interest in programming. the students most likely to indicate interest in financial education are the ones with higher financial literacy and lower overconfidence in their financial literacy. our work adds to the small body of existing research, namely allgood and walstad 134 b. j. davis et al. / financial services review 31 (2023) 133–150 (2016) and anderson et al. (2017), that perceived financial literacy and actual financial literacy are dually important. our results indicate that employers using financial education programming need to adopt innovative engagement strategies to improve the financial literacy and well-being of those employees who need it most. 2. research design and methodology our educational protocol was developed using an early career workplace education offering from a large financial services organization.3 we targeted the financial education offering to the entire pool of graduate students with the following research protocol. at the beginning of the academic year, and ahead of inviting the graduate student population to take our initial survey, we developed a series of prompts. these consisted of 5 ! 7 inch postcards announcing the project with lighthearted financial literacy questions on a front side and the correct answer along with a description of what to expect next on the back.4 we developed three distinct cards, with the goal that students might compare the particular card they received to others and discuss them. regarding signals of validity, integrity, and quality of the effort, the cards prominently featured the collaborating university and were placed in orientation packets for incoming students by the university. the university also placed cards at each department’s reception desk. a research assistant (ra) was assigned to attend graduate student council meetings and other graduate student group meetings of various types across the university system. in these student meetings, the ra was granted five-minute slots to discuss the project and the potential benefits of participation and left cards at each meeting.5 two weeks after the cards were distributed initial invitations for the online survey were sent to all graduate students (17,819) in a large public university system. these invitations were designed to resonate with the information on the cards we had just distributed. the survey was designed to record information on students’ individual and educational characteristics, financial aspirations, personal financial concerns, self-assessed financial acumen, and a financial literacy quiz.6 once the survey was completed, we provided quiz scores to students and offered them the correct answers to missed questions. we then asked whether a respondent was interested in taking part in a financial education seminar or webinar in the near future. our survey was open for one month between mid-september and mid-october, ahead of midterm examinations. over this four-week period the student received an initial invitation and up to four reminders targeted to students who had not taken the survey nor opted out of email engagement. initial and reminder response survey engagement rates are shown in fig. 1. our email prompts to engage the survey were successful with approximately 60% of the survey sample engaged the survey following a reminder as seen with the spikes in the response rate, supporting dechausay et al.’s (2015) result that reminders can improve engagement. from the initial invite population, 2,487 students (14%) engaged the survey. to set up the second stage of our study we invited a matched, random subsample of our surveyed students to a financial education seminar or webinar, whichever they preferred. we randomly selected 1,632 students to match to invited and noninvited groups. of the 1,632 students eligible for invitation, roughly two-thirds (1,101) were invited to participate.7 our invited b. j. davis et al. / financial services review 31 (2023) 133–150 135 student group was designed to be a balanced representation of participants across several dimensions: gender, degree of concern regarding financial matters, score on the financial literacy quiz, and whether or not the student was in a quantitative field.8 invitations were also balanced across those who did or did not initially indicate interest in the program. this allowed students to change their mind if they later decided they wanted to attend. the invitation email contained a link for those wishing to sign up and clicking the link brought the student to a standard web-based submission form. in the invitation to participate, we offered two mid-day and two early evening times for either a seminar or webinar at each campus in the university system. we also offered lunch or dinner to seminar participants; something our pilot run in the previous year had revealed as being important. 3. results in this section we present results from our survey and financial education engagement protocol. we begin by examining the demographic, educational field, and financial literacy characteristics of students who engaged the survey. section 3.2 examines how financial literacy (both self-assessed and actual) relate to interest in attending a financial education seminar. in section 3.3, we use regression analyses to examine correlates of interest in financial education interest and engagement. 3.1. survey sample characteristics table 1 presents characteristics of survey respondents. the average student who participated in the survey is nearly 30 years old and has almost 10 graduate course credits. women (57%) were more likely to participate compared to men, approximately a quarter are fig. 1. survey response rates. 136 b. j. davis et al. / financial services review 31 (2023) 133–150 international students, 57% of the graduate students attend the flagship campus, and nearly one in four graduate students work as research or teaching assistants. half of the students in the sample are pursuing a master’s degree, over a quarter a doctorate, and other degree types represent less than 10% of the survey respondents. about two-thirds of survey participants have prior work experience. we categorize a student’s program or major into four mutually exclusive groups: liberal arts (e.g., liberal arts, humanities, language, music, and social sciences), stem (e.g., science, engineering, medicine, mathematics, and technology), professional (e.g., public health, public administration, education, nursing, law, and other (pre)professional programs), and business (e.g., economics, finance, business, and accounting). enrollment in a professional program represents the plurality of students surveyed (37%) followed by enrollment in stem (27%), liberal arts (20%), and business (14%) programs. table 2 displays financial education engagement numbers. of the nearly 18,000 students we sent a survey, about 14% (2,487) responded. of the representative 1,101 students to whom we sent invitations for a financial education session, 16% (176) accepted. among the accepted group, 36% (64) attended a session with 44 opting for an in-person seminar and 20 for an online webinar. acceptance rates for students who initially indicated interest (25%) were significantly higher than those who did not signal interest (8%). program attendance rates for those initially indicating interest were also higher (39% vs. 29%) but the difference is not significant. we next tabulate personal financial characteristics and engagement across program type and survey variables. table 3 panel a displays measurements of financial literacy, personal financial concerns, and education engagement for all surveyed students by program type. financial literacy quiz score is the average number of questions students answered correctly on our financial quiz (out of 12).9 financial iq is the average of students self-assessed rating of how high their financial knowledge or iq is, ranging from 1 (very low) to (7 very high). relative finiq is an individual’s relative financial knowledge calculated as the table 1 characteristics of targeted graduate student population student survey sample summary statistics mean/ proportion standard deviation obs proportion obs student characteristics degree type age 29.5 7.5 2,487 masters 0.50 2,487 credits taken 9.8 5.1 2,487 certificate 0.02 2,487 married 0.37 2,306 law 0.07 2,487 female 0.57 2,487 doctorate 0.27 2,487 international student 0.26 2,487 medical 0.08 2,487 research/teaching asst 0.24 2,487 post-doc 0.01 2,487 flagship campus 0.57 2,487 other/non-degree 0.05 2,487 prior work experience 0.68 2,288 education program type liberal arts 0.20 2,487 stem 0.27 2,487 professional 0.38 2,487 business and economics 0.14 2,487 b. j. davis et al. / financial services review 31 (2023) 133–150 137 relative difference between performance on the financial literacy quiz and one’s selfassessed financial iq level and normalized on a scale of 0 to 1. values below 0.5 represent overconfidence and values above 0.5 represent underconfidence. for example, a value of 0 indicates complete overconfidence in the self-assessment compared to their actual measured financial literacy knowledge and corresponds to a student answering 0 out of 12 financial literacy questions correctly and indicating a very high level (7) of financial iq on the selfassessment, whereas a value of 1 indicates complete underconfidence. a value of 0.5 indicates neither overnor underconfidence. our personal finance concern metric is measured on an intensity scale from 1 (no concern) to 5 (great concern) across five areas (career goals, current finances, future finances, owning a home, and retirement), with a possible range from a low of 5 to a high of 25, which we normalized from 0 to 1. for engagement metrics, we list the percentage of students who responded to the survey, indicated that they were interested in financial education, accepting an invitation for an educational session, and attended a seminar or webinar (conditional on accepting an invitation). table 3 shows the average student got 65% of the financial literacy questions correct, with business students scoring significantly higher than any of the other program groups.10 the mean self-reported financial iq was 4.7 out of 7, again with business students indicating a significantly higher level of self-reported financial knowledge compared to nonbusiness table 3 financial literacy, financial concern, and financial education engagement characteristic overall program type liberal arts stem business professional financial literacy and concern financial literacy quiz score (0–12) 7.81 7.57 7.66 8.71 7.71 financial iq, self-assessed (1–7) 4.67 4.58 4.42 5.28 4.73 relative finiq (0–1) 0.52 0.52 0.53 0.51 0.51 personal finance concern (0–1) 0.72 0.73 0.70 0.71 0.73 education engagement responded to survey 14% 14% 14% 14% 14% interested in financial education 48 47 48 46 50 accept invite j invited 16 17 15 13 17 attend financial education conditional on accepting invite 36 51 27 42 33 note. means or percents reported. table 2 financial education engagement results engagement number number invited proportion survey respondents 2,487 17,819 0.14 eligible for invite 1,101 1,632 0.67 interested in fin ed j eligible 511 1,101 0.46 accepted j invited 176 1,101 0.16 attended j accepted invite 64 176 0.36 conditional on indicating interest accepted invite 383 511 0.25 attended j accepting invite 50 128 0.39 138 b. j. davis et al. / financial services review 31 (2023) 133–150 students (5.28 vs. 4.60 for nonbusiness students). the average student was slightly underconfident in their financial knowledge. there are limited differences in the relative measure by program type; however, stem students display significantly greater underconfidence compared with the rest of the surveyed population. we find no significant differences by program for students who responded to the survey or for those who indicated interested in attending a financial education seminar or for acceptance rates. for seminar attendance, however, we find a significantly greater proportion of liberal arts students attended a seminar compared to students in other programs. table 4 shows asset and debt characteristics of the surveyed students. we hypothesize that greater participation in financial services and the incidence of debt would be positively correlated to signaling interest in financial education. and having life insurance or an investment account may signal greater interest in financial planning over the life cycle. we include the incidence of students with a checking account, savings account, investment account, or a life insurance policy. nine of 10 students surveyed have a checking account, with three in four having a savings account. the third row list the proportion of students with an investment account, which we define as having a brokerage account, an ira, or an employersponsored retirement savings plan. this is owned by a minority of students (38%), but varies significantly by program type, ranging from 22% for stem students to 50% for business students. we posit debt should be positively correlated with financial education interest. graduate students with student loan debt need to manage loan repayments in conjunction with other consumption, savings, and investment goals and in context of their postgraduate career outlook. over half of the sample has some form of student loan debt, and this varies significantly by a student’s major field. professional students were significantly more likely to have any debt, both debt from graduate and undergraduate studies and from credit cards. table 4 other student characteristics, assets and debt characteristic overall program type liberal arts stem business professional banking, assets, insurance checking account 90% 93% 88% 88% 90% savings account 74 78 71 72 75 investment account 38 38 22 50 44 life insurance 34 29 23 38 42 student debt only undergraduate 9 12 9 5 8 only graduate 19 16 22 26 17 both undergrad and grad 28 30 22 15 36 any student debt 56 57 53 46 61 other debt credit card 31 34 23 26 37 auto 20 16 15 23 24 mortgage 20 20 11 24 25 home ownership home owner 22 21 12 27 28 plan to purchase home 46 39 55 50 43 b. j. davis et al. / financial services review 31 (2023) 133–150 139 credit card debt may demonstrate greater need for financial education since it relates to household balance sheets and not to human capital acquisition or indicates the reliance on high-cost credit card debt to finance education expenses. our final category is home ownership. about one in five already own a home and roughly 50% of students are planning to purchase one in the next 10 years, which varies significantly by a student’s major field. we later control for this in our regression analysis because purchasing a home involves a substantial amount of financial planning, and we expect those planning to purchase a home to have greater interest in financial education. 3.2. interest in financial education in this section, we explore the relationship between self-assessed and measured financial literacy and graduate student initial interest in attending a financial education seminar. nearly half (48%) of the students who participated in the survey indicated they were interested in financial education. in table 5, we find differences for indicated interest across nearly all individual characteristics. while individuals who do not interact with financial institutions may gain marginally greater benefit from financial education, we find that those without checking and savings accounts, or life insurance are significantly less interested in financial education. however, on the liability side of household balance sheets, students with student loan debt or credit card debt were more likely to be interested in the educational offerings. table 5 student characteristics by financial education interest survey characteristics interested not interested sig financial literacy and concern financial literacy quiz score 8.31 7.19 *** financial iq (self-assessed) 4.59 4.82 *** relative finiq 0.55 0.48 *** personal finance concern 0.75 0.69 *** banking, assets, insurance checking account 98% 81% *** savings account 83 65 *** investment account 42 34 *** life insurance 37 31 *** student debt only undergraduate 11% 7% *** only graduate 21 18 * both undergrad and grad 32 24 *** any student debt 64 48 *** other debt credit card 37% 26% *** auto loan 23 17 *** mortgage 20 20 home ownership own home 21% 23% plan to own home 55 39 *** note. means or percents shown. * and *** indicates differences are significant at the 10% and 1% levels, respectively. 140 b. j. davis et al. / financial services review 31 (2023) 133–150 graduate students signaling interest in financial education have significantly higher financial literacy, with interested students averaging about 1.2 more correct answers compared to non-interested students. further, non-interested students are significantly more overconfident in their financial knowledge. this is also shown by the distribution of relative finiq shown in fig. 2, which displays kernel density estimates of relative finiq by indicated interest in financial education. the two distributions are significantly different (p < .01), with noninterested students having a greater estimated density in the overconfidence range (values below 0.5). this result is economically meaningful as the mean difference of 0.07 in the relative measure between interested and non-interested students equates to 1.68 fewer questions answered correctly (out of 12) on our financial quiz for the relatively more overconfident student given a fixed self-assessed score. alternatively, a 0.07 difference results in a 0.84 greater self-assessed score, given a fixed quiz score, for the relatively more overconfident student. overconfident students with low measured financial literacy would arguably benefit the most from financial education. unfortunately, table 6 and fig. 3 indicates that these students were least interested in improving their financial knowledge. in table 6, we examine the percentage of students interested in financial education by their self-assessed financial iq rating and how well they did on the financial literacy quiz, grouping the latter into four categories. the data suggest there is a strong negative correlation between overconfidence and interest in financial education.11 this finding is highlighted by fig. 3, which displays a wireframe surface plot on the data points shown in table 6. financial literacy and selfassessed financial iq are shown on the x and y axes, and the percentage of students who indicated they are interested in financial education is shown on the z (vertical) axis. here the relationship is clear: students who have high financial literacy but self-assess a low level of financial iq are the most interested in financial education (often over 70%). comparatively, students with low financial literacy but self-assess a high level of financial iq are less interested (generally less than 30% of the time). fig. 2. kernel density estimates of relative finiq by indicated interest in financial education. b. j. davis et al. / financial services review 31 (2023) 133–150 141 to understand this relationship over all possible values, fig. 4 displays predicted probabilities of financial education interest. the predicted probabilities are generated using a simple logit model estimating the likelihood that students indicated interest in financial education regressed on a student’s finiq and financial literacy quiz score, treating the two exogenous variables as categorical variables.12 the full profile of predicted probabilities is shown in appendix table a.1. fig. 4 replicates and smooths the relationship shown in fig. 3. generally, the predicted probability that a student signals interest increases significantly in one’s measured financial literacy but significantly decreases as a student’s selfassessment increases. students with a combination of high self-assessed financial iq and low actual financial literacy are predicted to be the least likely to signal interest in financial education. for example, a student answering all questions correctly but self-assesses the lowest level of financial knowledge is predicted to signal interest with a probability above 60%, which decreases to 45% for a student self-assessing the highest level of financial knowledge. by contrast, a student answering no questions correctly on the financial literacy table 6 financial education interest by measured and self-assessed financial literacy percent interested in financial education self-assessed percent of financial quiz questions correct financial iq 0–25% 26–50% 51–75% 76–100% mean n 1 (very low) 50% 80% 50% 0% 55% 20 2 29% 65% 72% 82% 63% 80 3 26% 61% 68% 70% 62% 236 4 20% 49% 61% 58% 56% 646 5 22% 30% 64% 59% 57% 481 6 6% 30% 57% 55% 51% 484 7 (very high) 22% 0% 50% 45% 42% 153 mean 20% 55% 62% 39% 48% n 221 257 974 1,035 2,100 fig. 3. surface plot of financial education interest by measured and self-assessed financial literacy. 142 b. j. davis et al. / financial services review 31 (2023) 133–150 quiz but is most confident in their financial knowledge, is predicted to signal interest with a likelihood of under 6%, increasing to only 11% for a student self-assessing the lowest level of financial knowledge. program sponsors need to be innovative when thinking about how to engage the population of students who would most benefit from financial education. 3.3. regression analysis of financial education interest and engagement we begin our regression analysis with first considering who takes the survey. table 7 uses ordinary probit regression estimating whether a student engages in the survey or not using student and program characteristics, displaying marginal effects on the coefficients and standard errors in parenthesis. because those who do not take the survey do not offer us data on their financial literacy, we only leverage the university system’s administrative data. women and student workers are significantly more likely to engage the survey. while there is no significant relationship for graduate students at the flagship campus, student workers at the flagship campus are significantly less likely to engage the survey, indicated by the interaction term in model 2. these students may be more time constrained with their studies and work duties than their counterparts. we find no significant effect for international students. students who have taken more graduate credits are significantly less likely to take the survey, but the marginal effect is small and has significant attenuation. there were no significant differences in a student’s major subject area. when examining degree type, law students were significantly less likely and doctorate students only marginally less likely to engage in the survey, highlighting the possibility of time constraints for students in terminal degree programs. table 8 estimates the likelihood that a student indicates interest in financial education using ordinary probit specifications showing marginal effects and standard errors in parenthesis. we display five specifications: model 1 uses administrative data plus financial literacy characteristics, and models 2 and 3 add asset and debt characteristics. model 4 adds fig. 4. predicted probabilities of education interest by measured and self-assessed financial literacy. b. j. davis et al. / financial services review 31 (2023) 133–150 143 t ab le 7 r eg re ss io n es ti m at es o f su rv ey en g ag em en t (1 ) (2 ) (3 ) (4 ) m ar . co ef f. s ta n d ar d er ro r m ar . co ef f. s ta n da rd er ro r m ar . co ef f. s ta n d ar d er ro r m ar . co ef f. s ta n d ar d er ro r s tu de n t ch ar ac te ri st ic s w o m an 0 .0 3 0* * * (0 .0 0 5 ) 0 .0 3 0 * * * (0 .0 0 5 ) 0 .0 3 0 * * * (0 .0 0 5) 0 .0 3 0 * * * (0 .0 0 5 ) a g e " 0 .0 0 0 (0 .0 0 0 ) 0 .0 0 0 (0 .0 0 0 ) " 0 .0 0 0 (0 .0 0 0) 0 .0 0 0 (0 .0 0 0 ) in tl st u d en t 0 .0 0 9 (0 .0 0 6 ) 0 .0 0 7 (0 .0 0 6 ) 0 .0 0 8 (0 .0 0 6) 0 .0 0 6 (0 .0 0 6 ) m ai n ca m p u s 0 .0 0 2 (0 .0 0 6 ) 0 .0 0 9 (0 .0 0 6 ) 0 .0 0 7 (0 .0 0 6) 0 .0 0 8 (0 .0 0 6 ) s tu d en t te ac h in g /r es ea rc h as st 0 .0 3 7* * * (0 .0 0 7 ) 0 .0 7 5 * * * (0 .0 1 4 ) 0 .0 7 7 * * * (0 .0 1 4) 0 .0 7 3 * * * (0 .0 1 4 ) e d u c cr ed it s " 0 .0 0 6* * * (0 .0 0 1 ) " 0 .0 0 6 * * * (0 .0 0 1 ) " 0 .0 0 6 * * * (0 .0 0 1) " 0 .0 0 5 * * * (0 .0 0 2 ) ln (e d u c cr ed it s) 0 .0 6 1* * * (0 .0 1 1 ) 0 .0 6 1 * * * (0 .0 1 1 ) 0 .0 6 0 * * * (0 .0 1 2) 0 .0 5 6 * * * (0 .0 1 2 ) m ai n c am p u s * s tu d en t a ss t " 0 .0 5 0 * * * (0 .0 1 6 ) " 0 .0 4 9 * * * (0 .0 1 6) " 0 .0 4 6 * * * (0 .0 1 6 ) p ro g ra m ty p e (b as el in e ¼ l ib er al a rt s) s t e m " 0 .0 0 2 (0 .0 0 9) p ro fe ss io n al 0 .0 0 3 (0 .0 0 8) b u si n es s 0 .0 0 7 (0 .0 1 0) p ro g ra m ty p e (b as el in e ¼ m as te rs ) d o ct o ra te " 0 .0 1 2 * (0 .0 0 7 ) l aw " 0 .0 3 3 * * * (0 .0 0 9 ) o th er /n o n -d eg re e " 0 .0 0 0 (0 .0 1 3 ) c er ti fi ca te " 0 .0 2 7 (0 .0 1 8 ) p o st -d o c " 0 .0 1 7 (0 .0 3 2 ) m ed ic al " 0 .0 1 6 (0 .0 1 1 ) o b se rv at io n s 1 7 ,7 5 0 1 7 ,7 5 0 1 7 ,7 5 0 1 7 ,7 50 n o te . p ro b it sp ec ifi ca ti o n s w it h m ar g in al co ef fi ci en ts re p o rt s an d st an d ar d er ro rs in p ar en th es is . * , * * , * * * re p re se nt s si g n ifi ca n ce at th e 1 0 % , 5 % , an d 1 % le v el s, re sp ec ti v el y . 144 b. j. davis et al. / financial services review 31 (2023) 133–150 t ab le 8 r eg re ss io n es ti m at es o f se m in ar in te re st (1 ) (2 ) (3 ) (4 ) (5 ) s tu de n t an d p ro gr am ch ar ac te ri st ic s w o m an 0 .0 2 3 (0 .0 2 2) 0 .0 2 1 (0 .0 2 3 ) 0 .0 2 1 (0 .0 2 3) 0 .0 1 9 (0 .0 2 2 ) 0 .0 0 1 (0 .0 2 2) a g e " 0 .0 0 4 (0 .0 0 2) * * * " 0 .0 0 4 (0 .0 0 2 )* * " 0 .0 0 3 (0 .0 0 2) * * " 0 .0 0 1 (0 .0 0 2 ) " 0 .0 0 1 (0 .0 0 2) in t’ l st u d en t 0 .0 3 5 (0 .0 2 7) 0 .0 2 9 (0 .0 2 8 ) 0 .0 6 5 (0 .0 3 1) * * 0 .0 5 9 (0 .0 3 1 )* 0 .0 4 2 (0 .0 3 1) m ai n ca m p u s 0 .0 0 8 (0 .0 2 4) 0 .0 0 3 (0 .0 2 4 ) 0 .0 0 9 (0 .0 2 4) 0 .0 0 5 (0 .0 2 4 ) 0 .0 0 4 (0 .0 2 5) r es ea rc h /t ea ch in g as st " 0 .0 0 1 (0 .0 2 6) " 0 .0 0 8 (0 .0 2 6 ) " 0 .0 1 0 (0 .0 2 7) " 0 .0 1 5 (0 .0 2 7 ) " 0 .0 1 7 (0 .0 2 7) c re d it s " 0 .0 0 5 (0 .0 0 2) * * " 0 .0 0 6 (0 .0 0 3 )* * " 0 .0 0 6 (0 .0 0 3) * * " 0 .0 0 6 (0 .0 0 3 )* * " 0 .0 0 6 (0 .0 0 3) * * s t e m 0 .0 4 4 (0 .0 3 3) 0 .0 4 3 (0 .0 3 3 ) 0 .0 4 7 (0 .0 3 3) 0 .0 3 8 (0 .0 3 3 ) 0 .0 3 5 (0 .0 3 4) b u si n es s 0 .0 1 4 (0 .0 3 8) 0 .0 2 1 (0 .0 3 9 ) 0 .0 2 3 (0 .0 3 9) 0 .0 1 8 (0 .0 3 9 ) 0 .0 3 6 (0 .0 3 9) p ro fe ss io n al 0 .0 5 3 (0 .0 2 9) * 0 .0 5 6 (0 .0 2 9 )* 0 .0 5 5 (0 .0 2 9) * 0 .0 5 1 (0 .0 2 9 )* 0 .0 5 4 (0 .0 2 9) * f in an ci al li te ra cy ch ar ac te ri st ic s f in an ci al li te ra cy 0 .0 4 3 (0 .0 0 4) * * * 0 .0 4 3 (0 .0 0 4 )* * * 0 .0 4 4 (0 .0 0 4) * * * 0 .0 4 4 (0 .0 0 4 )* * * s el fas se ss ed fi n an ci al iq " 0 .0 5 2 (0 .0 0 8) * * * " 0 .0 5 1 (0 .0 0 9 )* * * " 0 .0 5 0 (0 .0 0 9) * * * " 0 .0 5 0 (0 .0 0 9 )* * * r el at iv e f in iq 0 .8 5 3 (0 .0 7 2) * * * p er so n al fi na n ce co n ce rn 0 .4 5 8 (0 .0 5 5) * * * 0 .4 6 1 (0 .0 5 5 )* * * 0 .4 5 2 (0 .0 5 5) * * * 0 .4 3 1 (0 .0 5 6 )* * * 0 .4 2 6 (0 .0 5 6) * * * b an k in g, as se ts , in su ra n ce in v es tm en t o r re ti re m en t " 0 .0 1 9 (0 .0 2 6 ) " 0 .0 1 5 (0 .0 2 6) " 0 .0 0 5 (0 .0 2 7 ) 0 .0 1 0 (0 .0 2 6) s av in g s ac co u n t 0 .0 2 5 (0 .0 2 9 ) 0 .0 2 8 (0 .0 2 9) 0 .0 2 8 (0 .0 2 8 ) 0 .0 3 6 (0 .0 2 8) l if e in su ra n ce " 0 .0 2 7 (0 .0 2 5 ) " 0 .0 3 1 (0 .0 2 5) " 0 .0 1 9 (0 .0 2 5 ) " 0 .0 1 8 (0 .0 2 5) d eb t s tu d en t d eb t (u n d er g ra d o n ly ) 0 .1 0 8 (0 .0 3 8) * * * 0 .1 0 1 (0 .0 3 8 )* * * 0 .0 9 9 (0 .0 3 8) * * * s tu d en t d eb t (g ra d o n ly ) 0 .0 2 7 (0 .0 3 0) 0 .0 1 9 (0 .0 3 0 ) 0 .0 1 6 (0 .0 3 0) b o th u n d er g ra d an d g ra d d eb t 0 .0 6 0 (0 .0 2 9) * * 0 .0 5 1 (0 .0 2 9 )* 0 .0 4 3 (0 .0 2 9) c re d it c ar d " 0 .0 5 1 (0 .0 6 3) " 0 .0 5 9 (0 .0 6 3 ) " 0 .0 5 0 (0 .0 6 3) a u to 0 .0 3 2 (0 .0 2 7) 0 .0 4 2 (0 .0 2 7 ) 0 .0 4 1 (0 .0 2 7) h o m e o w n er sh ip o w n h o m e " 0 .0 6 8 (0 .0 3 7 )* " 0 .0 5 7 (0 .0 3 7) p la n to p u rc h as e h o m e 0 .0 5 3 (0 .0 2 7 )* * 0 .0 6 1 (0 .0 2 7) * * o b se rv at io n s 2 ,0 9 8 2 ,0 9 8 2 ,0 9 8 2 ,0 9 8 2 ,0 9 8 n o te . p ro b it w it h m ar g in al ef fe ct s sh o w n . s ta n d ar d er ro rs sh o w n in p ar en th es is . * , * * , * * * in d ic at es si g n ifi ca n ce at th e 1 0 % , 5 % , o r 1 % le v el . b. j. davis et al. / financial services review 31 (2023) 133–150 145 home ownership characteristics, and model 5 uses the relative finiq measure on the full specification instead of quiz score and self-assessed financial iq. beginning with model 1, younger students are significantly more likely to signal interest, but this is not robust in later specifications. professional students are significantly more likely to signal interest; however, this is only significant at the 10% level when controlling for debt characteristics. we find no significant correlation for business students.13 in all specifications, financial literacy quiz scores and self-assessed financial iq have large significant effects on the likelihood to signal interest. these effects pull in the opposite directions, as discussed in section 3.2., with predicted interest increasing in quiz score but decreasing in the self-assessed measure. an additional question answered correctly on the quiz score increases the estimated likelihood to indicate interest by 4%, while a one unit increase in the self-assessed measure decreases the likelihood by 5%. the personal finance degree of concern composite measure is significantly and positively related to indicating interest, following our hypothesis. in model 5, we use relative finiq as a regressor, instead of quiz score and the self-assess measure; and find estimated interest significantly increases (decreases) as underconfidence (overconfidence) increases. adding individual asset and banking in model 2 has no significant impact. model 3 includes student loan debt as a categorical variable (with no student loan debt as the baseline). undergraduate student loan debt has a significant impact on the likelihood to be interested in financial education, compared to those without student loan debt. although we do not find a (robust) significant effect for either graduate debt only or both debt from undergraduate and graduate school, the coefficient is positive—in the hypothesized direction. debt management is likely to become a larger concern for graduate students in the accumulation phase of their lifecycle, especially since this group delays employment income, savings, and loan repayment before (re)entering the labor force, albeit at an expected relatively higher salary. credit card debt was not a significant correlate. this may be of concern because some students may be using high-cost debt to finance part of their education and this group would benefit from financial education. when including home ownership characteristics in model 4 we find those planning to purchase a home in the next ten years are significantly more likely to signal interest, following our hypothesis. 4. discussion this paper examined financial education interest among graduate students in a large public university system. we find a strong positive and significant correlation between underconfidence (overconfidence) in self-measured financial knowledge and (lack of) interest in financial education programming. this finding suggests the need for innovative engagement strategies to identify and provide programming to individuals who would benefit the most from improving their financial literacy. the results speak to several components regarding the timing and delivery of financial education. the first is whether the timing is optimal for graduate students to engage in improving their financial literacy. because many graduate students are close to (re)entering 146 b. j. davis et al. / financial services review 31 (2023) 133–150 the workforce, they may be focused on the near-term issues of graduating, finding a job, or moving. this can make them subject to present bias through the belief that they have scant time to devote additional resources to improving their long-term financial well-being. innovative engagement strategies, such a providing lunch or tchotchkes, may nudge active participation in financial education. however, we find these nudges did not address to our main finding that those confident in their financial knowledge but have low financial literacy are significantly less interested in financial education. how should these individuals be engaged in financial education? mandatory financial education could be one response, and many states have begun to institute mandatory financial literacy programs in high school. stoddard and urban (2020), urban, schmeiser, collins, and brown (2018), and collins (2013) find some benefits to mandatory education. however, there needs to be further research in this area as there remain many open questions, including whether the education benefits persist in later life, how such programs are implemented, what is included in the content, and when in the lifecycle are they delivered, among others. with most programs continuing to rely on voluntary education efforts, often offered by employer benefit programs, designers and implementers of financial education will need to consider how to attract individuals overconfident in their knowledge of personal finances. notes 1 for research on financial literacy see yakoboski, lusardi, and hasler (2019), clark, lusardi, and mitchell (2017), lusardi and mitchell (2014), lusardi, mitchell, and curto (2014), lusardi and mitchell (2011), among others. for a broader discussion on financial literacy, financial education, and economic outcomes see hastings, madrian, and skimmyhorn (2013). 2 bernheim and garrett (2003), lusardi (2004), maki (2004), and bayer, bernheim, and scholz (2009) have studied employer-sponsored financial literacy programs and retirement preparedness. other studies have found other positive benefits to financial education programs (i.e., clark et al., 2006; skimmyhorn et al., 2016; and seligman and bose 2012). 3 in the previous year, we piloted a similar survey and education to a small group of students. while the test group gave good feedback and generally positive reviews, we adjusted both our materials and engagement strategy to improve participation. 4 this postcard campaign is consistent with findings on the positive value of prompts from dechausay, anzelone, and reardon (2015). 5 the ra had previously served as president of the graduate student council and thus was a familiar and respected source of information across the body of groups we engaged. 6 lusardi, mitchell, and curto (2014), schmeiser and seligman (2013), and knoll and houts (2012) have published work evaluating questions on measuring financial literacy using three independent methodologies. we take our financial literacy questions from this work, and consultation with financial counselors, and use a set of 12 questions from these studies. b. j. davis et al. / financial services review 31 (2023) 133–150 147 7 a control group that was roughly one-third of eligible participants was not invited so as to be able to carry forward with other research questions. 8 we defined a student as being in a quantitative field if the student’s program is in economics, business, engineering, statistics, mathematics, physics, chemistry, or computer science. 9 the financial literacy quiz covered questions on interest, inflation, and bond prices. 10 we control for degree type in our regression analysis. 11 our relative measure cannot examine whether there are differences in student interest across the entire cross product of the financial literacy score and self-assessed financial iq measure. 12 the overall model is significant (p < .01 with a x2 test). moreover, the predicted probabilities for each combination of finiq and quiz score (91 combinations) are all significant at the 5% level. 13 in separate regressions (not shown) we controlled for degree type and prior work experience; there were no significant differences. appendix table a1 predicted probabilities of indicating interest in financial education by financial literacy and selfassessed financial iq financial literacy quiz score self-assessed financial iq very low 1 2 3 4 5 6 very high 7 0 (0%) 11.1% 14.3% 13.7% 10.3% 10.5% 7.9% 5.9% 1 30.6 37.1 35.9 28.9 29.3 23.3 18.2 2 33.6 40.3 39.1 31.8 32.2 25.8 20.3 3 35.8 42.6 41.4 33.9 34.4 27.7 21.9 4 52.3 59.4 58.1 50.2 50.7 42.9 35.6 5 61.1 67.7 66.6 59.2 59.7 51.9 44.3 6 (50%) 54.1 61.2 60.0 52.1 52.6 44.8 37.3 7 65.9 72.1 71.1 64.1 64.5 57.1 49.4 8 63.0 69.4 68.3 61.1 61.5 53.9 46.2 9 64.5 70.8 69.8 62.7 63.1 55.6 47.9 10 62.0 68.6 67.5 60.1 60.6 52.9 45.2 11 62.2 68.7 67.6 60.2 60.7 53.0 45.4 12 (100%) 62.3 68.8 67.7 60.3 60.8 53.2 45.5 note. predicted probabilities from logit model as described in section 3.2, all are significant at the 5% level. 148 b. j. davis et al. / financial services review 31 (2023) 133–150 acknowledgments we thank annamaria lusardi and carrie houts for their comments on our financial literacy survey. we thank james winbush, dan rives, keatrick johnson, and others at indiana university for their engagement and contributions. we have benefited from the capable research assistance of benjamin bissette, katherine jones, and lanita rahjina. we acknowledge the financial support of indiana university in this work. any errors are our own. the views are the authors’ and do not necessarily represent the views of the tiaa, the tiaa institute, or the investment company institute. references allgood, s., & walstad, w. (2016). the effects of perceive and actual financial literacy on financial behaviors. economic inquiry, 54, 675–697. available at: https://doi.org/10.1111/ecin.12255 anderson, a., baker, f., & robinson, d. (2017). precautionary savings, retirement planning and misperceptions of financial literacy. journal of financial economics, 126, 383–398. available at: https://doi.org/10.1016/j. jfineco.2017.07.008 bayer, p., bernheim, d., & scholz, j. (2009). the effects of financial education in the workplace: evidence from a survey of employers. economic inquiry, 47, 605–624. available at: https://doi.org/10.1111/j.1465-7295.2008.00156.x bernheim, d., & garrett, d. m. (2003). the effects of financial education in the workplace: evidence from a survey of households. journal of public economics, 87, 1487–1519. available at: https://doi.org/10.1016/s0047-2727 (01)00184-0 clark, r., lusardi, a., & mitchell, o. (2017). employee financial literacy and retirement plan behavior: a case study. economic inquiry, 55, 248–259. available at: https://doi.org/10.1111/ecin.12389 clark, r. l., d’ambrosio, m. b., mcdermed, a. a., & sawant, k. (2006). retirement plans and saving decisions: the role of information and education. journal of pension economics and finance, 5, 45–67. available at: https://doi.org/10.1017/s1474747205002271 collins, m. (2013). the impacts of mandatory financial education: evidence from a randomized field study. journal of economic behavior & organization, 95, 146–158. available at: https://doi.org/10.1016/j.jebo.2012.08.011 dechausay, n., anzelone, c., & reardon, l. (2015). the power of prompts: using behavioral insights to encourage people to participate. opre report 2015-75. washington, dc: office of planning, research and evaluation, administration for children and families, u.s. department of health and human services. hastings, j., madrian, b., & skimmyhorn, w. (2013). financial literacy, financial education, and economic outcomes. annual review of economics, 5, 347–373. available at: https://doi.org/10.1146/annurev-economics-082312-125807 knoll, m., & houts, c. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. journal of consumer affairs, 46, 381–410. available at: https://doi.org/ 10.1111/j.1745-6606.2012.01241.x lusardi, a. (2004). saving and the effectiveness of financial education. in: o. mitchell & s. utkus (eds.), pension design and structure: new lessons from behavioral finance (pp. 157–184). oxford: oxford university press. lusardi, a., & mitchell, o. (2011). the outlook for financial literacy. in: o. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 1–16). oxford: oxford university press. lusardi, a., & mitchell, o. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. available at: https://doi.org/10.1257/jel.52.1.5 lusardi, a., mitchell, o., & curto, v. (2014). financial literacy and financial sophistication in the older population. journal of pension economics & finance, 13, 347–366. available at: https://doi.org/10.1017/s1474747214000031 lusardi, a., michaud, p.-c., & mitchell, o. (2017). optimal financial knowledge and wealth inequality. the journal of political economy, 125, 431–477. available at: https://doi.org/10.1086/690950 maki, d. (2004). financial education and private pensions. in w. gale, j. shoven, & m. warshawsky (eds.), private pensions and public policies (pp. 126–139). washington, dc: brookings institution press. schmeiser, m., & seligman, j. (2013). using the right yardstick: assessing financial measures by way of financial well-being. journal of consumer affairs, 47, 243–262. available at: https://doi.org/10.1111/joca.12010 b. j. davis et al. / financial services review 31 (2023) 133–150 149 seligman, j., & bose, r. (2012). learning by doing: active employer sponsored retirement savings plan participation and household wealth accumulation. the quarterly review of economics and finance, 52, 162–172. available at: https://doi.org/10.1016/j.qref.2012.02.002 skimmyhorn, w., davies, e., mun, d., & mitchell, b. (2016). assessing financial education methods: principles vs. rules-of-thumb approach. the journal of economic education, 47, 193–210. available at: https://doi.org/ 10.1080/00220485.2016.1179145 stoddard, c., & urban, c. (2020). the effects of state mandated financial education on college financing behaviors. journal of money, credit and banking, 52, 747–776. available at: https://doi.org/10.1111/jmcb.12624 urban, c., schmeiser, m., collins, j., & brown, a. (2018). the effects of high school personal financial education policies on financial behaviors. economics of education review, 78, 101786. yakoboski, p., lusardi, a., & hasler, a. (2019). financial literacy in the united states and its link to financial wellness. the 2019 tiaa institute-gflec personal finance index. charlotte, nc: tiaa institute. 150 b. j. davis et al. / financial services review 31 (2023) 133–150 pii: s1057-0810(99)80015-8 financial services review, 7(l): 71-71 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. erratum the following are corrected equations for the article, “a simple and effective trading rule for individual investors,” by laurie prather and william bertin which appeared in the last issue of financial services review, volume 6, number 4, on pp. 285-294. page 289 equation 1 should be: page 290 equation 4 should be: p(n, = xln, n2, n) = n 0 n page 293 the calculation in the middle of the page should be: p(n, = 6123, 17,6) = (%1’6) = .0046_ 23 ( ) 11 distribution channel effects on advisor managed investment performance jason e. hellera*, benjamin f. cummingsb, jason martinc acoastal wealth, 1000 corporate drive, 7th floor, ft. lauderdale, fl 33334, usa bdepartment of financial planning & analytics, utah valley university, 800 west university parkway, orem, ut 84058, usa cthe american college of financial services, 605 oreland mill road, oreland, pa 19075, usa abstract this study focuses on the effects that business models have on advisor managed portfolio performance by attempting to determine if advisors at registered investment advisory (ria) firms produce higher net investment results compared with advisors employed at dually registered independent broker/dealer (ibd) firms. using data from one of the largest investment advisory platforms in the united states, we found qualified supporting evidence that advisors at rias outperformed advisors at ibds in higher-risk portfolios through the use of turnkey asset management programs and unified managed accounts. © 2022 academy of financial services. all rights reserved. jel classification: g2 keywords: cognitive load theory; registered investment advisor; independent broker/dealer; compensation puzzle; investment performance 1. introduction the efficacy of financial advice often compares investment performance to benchmark portfolios. this study segments the financial advice market into two separate distribution channels—registered investment advisors and independent broker-dealers—to determine corresponding author: tel.: +1-954-938-8800; fax: +1-954-351-2468. e-mail address: jasonheller@mycoastalwealth.com 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 145–164 if membership in a particular channel has a significant predictive relationship with advisor managed portfolio performance. other than potential philosophical differences, the main difference that separates registered investment advisory (ria) and independent broker/dealer (ibd) firms is the compensation structures employed. traditionally, ria firms derive their compensation from fee-only arrangements. in an assets-undermanagement model, the advisor charges a percentage of the client’s portfolio on a quarterly or monthly basis. in typical assets-undermanagement programs, trading securities does not generate a commission for the advisor; the value in such actions lies in the potential of the replacing security outperforming the replaced security and thus increasing the account value, in turn generating a higher dollar amount earned by the advisor. ibds employ a dual registration model that allows for fee-for-advice models as well as commission-based compensation programs. the advisor retains the sole discretion as to which model or mixture of the two they utilize. the decision ibds face about which compensation regime to pursue creates the potential for an advisor’s attention to be diverted away from their central task of investment management. due to this diversion, we seek to determine if the distribution channel affects advisor managed portfolio returns when comparing rias and ibds. the rest of the paper proceeds as follows: the literature review section provides a detailed background on the efficacy of financial advice, highlighting a gap in the literature regarding segmentation of advisor distribution channel membership. the theoretical framework relates cognitive load theory to the task of investment management. the methods and data employed for the study are then explained, and the results are presented. a discussion of the results precedes the conclusion, which includes limitations and implications of the study as well as areas for future research. 1.1. literature review numerous studies show that the majority of professional money managers do not consistently outperform passive benchmarks (del guercio, reuter, & tkac, 2010; desai & jain, 1995; gil-bazo & ruiz-verdú, 2009; jensen, 1968; malkiel, 1995). gruber (1996), french (2008), and reuter (2015) estimate that actively managed mutual funds underperform their benchmark indexes by an average of 64-67 basis points (bps) annually. other studies that show advisor recommended mutual funds underperform self-directed portfolios. karabulut (2013) found that advised investors earned lower raw and risk-adjusted returns compared with self-directed investors even before deducting advisory fees and transactions costs. bergstresser, chalmers, and tufano (2009) found that broker-sold funds had lower raw and risk-adjusted returns than direct-channel funds, even before distribution expenses were deducted. del guercio and reuter (2014) found that broker-sold actively managed mutual funds underperformed both broker-sold index funds and direct channel actively managed mutual funds. by studying the oregon university system retirement plan, chalmers and reuter (2012) found that employees who retained the services of brokers earned significantly lower after146 j. e. heller et al. / financial services review 30 (2022) 145–164 fee returns and lower risk-adjusted returns compared with those employees who were defaulted into age-based target date funds. the average fee of 0.9% was the largest reason for the underperformance. chalmers and reuter (2012) did point out that employee accounts that were self-directed also underperformed the default target date funds, but to a lesser extent than broker advised accounts, echoing the sentiment in bergstresser et al. (2009). internationally, hackethal, haliassos, and jappelli (2012) and foerster, linnainmaa, melzer, and previtero (2017) found similar results when studying german and canadian investors and advisors’ recommendations. on the other hand, kinniry, jaconetti, dijoseph, zilbering, and bennyhoff (2016) suggested that advisor-driven portfolios could outperform self-directed portfolios of clients, assuming the advisor did several tasks deemed to be too difficult, advanced, or time consuming for the novice investor. the study suggests that the so-called advisor alpha could be as high as 3.0% annually, the most valuable activity being behavioral financial coaching, which could contribute as much as 1.50% annually to a client’s portfolio return. although hackethal et al. (2012) found that the self-directed portfolios outperformed advised portfolios on average, advised accounts exhibited far greater diversification. hackethal et al. (2012) suggests that a potential reason clients pay for advice lies in the convenience of outsourcing the task rather than to outperform other alternatives. gennaioli, shleifer, and vishny (2015) put forth the concept of “money doctors” and posit that professional money managers instill confidence in the client by having a professional at the helm. this confidence reduces anxiety created by investing in risk-based assets and allows the client to invest more aggressively than they would on their own. gennaioli et al. (2015) recognize that advisors’ recommendations are costly, at times generic, and occasionally selfserving, which lead to consistent underperformance compared with passive benchmarks. while a client might earn negative market-adjusted returns after an advisor’s fees, the excess return generated compared with a counterfactual portfolio with limited risk-based assets is another measure of the value of an advisor (gennaioli et al., 2015). warshcuer and sciglimpaglia (2012) asked clients to rate the perceived value of financial planning services. making sure the client is holding a sufficiently diversified portfolio and holding investments that meet each of the client’s goals’ time horizons and cash flow needs were viewed as more important than recommending investments that beat the market averages. although advisors in general are unable to consistently outperform passive benchmarks (and in some cases self-directed portfolios), little attention has been given to determine if the efficacy of financial advice improves across different advisor business models. we seek provide insight about the differences in advisor performance based on their distribution channel. 1.2. theoretical framework cognitive load theory (clt) describes the limits of mental effort used in working memory during problem-solving (sweller, 1988). the amount of cognitive load levied on an individual engaged in a complex problem-solving exercise can be an explanatory factor in the individual’s performance (sweller, 1988). the heavier the cognitive load, the lower the expected level of performance. j. e. heller et al. / financial services review 30 (2022) 145–164 147 cognitive load is separated into three different types: intrinsic, extraneous, and germane. the first two forms of cognitive load are additive and together cannot exceed the capacity of working memory if the task is to be completed effectively (paas, renkl, & sweller, 2003). intrinsic cognitive load is the inherent level of difficulty associated with a particular task (chandler & sweller, 1991). it depends on the level of elemental interactivity in the problem-solving action (paas et al., 2003). the more interrelated the elements of the task are, the higher the intrinsic cognitive load. high elemental interactivity imposes a heavy cognitive load because each element must be processed simultaneously. in contrast, problem-solving involving large numbers of unrelated elements would not impose as heavy a cognitive load because each element could be processed individually without reference to the other elements (leppink, van gog, paas, & sweller, 2015). examples of intrinsic cognitive load for investment management include conducting due diligence and investment research and implementing portfolio decisions through trading, rebalancing, and ongoing monitoring. extraneous cognitive load, also known as ineffective cognitive load, is present when confounding variables are introduced into the problem-solving activity and interfere with its efficient completion (paas et al., 2003). these variables or processes are related to the problem-solving activity but create unnecessary and inefficient additional steps to complete the problem, which hinder performance (leppink et al., 2015). due to the additive nature of the cognitive load architecture, the presence of extraneous cognitive load is particularly important when intrinsic cognitive load is high. because cognitive load cannot exceed working memory capacity, when intrinsic cognitive load is high, there is less capacity for extraneous cognitive load (paas et al., 2003). examples of extraneous cognitive load that financial advisors may face include addressing client servicing tasks and related paperwork, developing and marketing the business, conducting administrative tasks, and engaging in professional development. germane cognitive load is used to explain any unused excess working memory capacity that can be refocused into activities that support intrinsic cognitive load (sweller, van merrienboer, & paas, 1998). the presence of germane cognitive load is desirable as it helps lessen the strain of intrinsic cognitive load and improves cognitive performance in problemsolving. since intrinsic cognitive load cannot be altered, in situations where the cognitive load level is high, reducing or eliminating extraneous cognitive load improves the overall cognitive process (leppink et al., 2015; sweller, 1988; sweller et al., 1998). the split-attention effect provides an example of the toll extraneous cognitive load can have in explaining a limitation of human information processing (chandler & sweller, 1991). extraneous cognitive load increases when a subject’s focus is split between multiple elements in a cognitive process. an additional deterrent to minimizing cognitive load is choice overload. as the choice set grows, the number of characteristics needing comparison increases and cognitive costs rise, potentially giving way to overload (greenleaf & lehmann, 1995; shugan, 1980). when choices are consequential and/or involve numerous options, the decision-making process becomes more effortful, which can lead to cognitive overload (botti & iyengar, 2006; huberman, iyengar, & jiang, 2004). 148 j. e. heller et al. / financial services review 30 (2022) 145–164 cognitive load is also more likely to be exhausted when processing more complex tasks (jacko & ward, 1996). campbell (1988) suggests that a complex task must minimally have either multiple paths, multiple outcomes, conflicting interdependence among paths, or uncertain probabilistic linkages. using this definition, portfolio management can be defined as a complex task that requires considerable cognitive resources to perform and is more likely to exhaust cognitive load. the effects on performance due to multitasking are also noteworthy. gonzàlez and mark (2005) discovered that task switching was equally created by external interruptions as well as internal self-interruptions, called discretionary switching. discretionary switching is the type most associated with tasks that require multiple related, but separate subtasks, and is most closely related to our study. like split-attention, discretionary task switching diverts cognitive resources from the primary task to a secondary or tertiary task, potentially before the completion of the primary task. czerwinski, horvitz, and wilhite (2004) found that complex tasks were more difficult to resume once interrupted. hodgetts and jones (2006) found an inverse relationship between primary task difficulty and resumption times; the more difficult or complex the primary task, the slower the resumption time once it was interrupted. gillie and broadbent (1989) found that primary task accuracy after interruptions declined as task complexity increased. jin and dabbish (2009) identified seven categories of discretionary switching. most relevant for this study is inquiry, which is switching to a secondary task to gain information that aides in completing the primary task. independent rias are typically fee-only planners whose compensation is derived either as a set fee (e.g., a flat or per hour charge for services), a percentage of the assets under management (aum), or a combination of the two. independent broker/dealers maintain a dual compensation model: (1) commission-based product placement and (2) a fee-based model similar to independent rias. with investment management as the primary task, an advisor at an ibd must first complete a secondary task and determine (i.e., inquire and engage in discretionary switching) what amount of the client’s investible net worth and/or discretionary income will be implemented through an asset-undermanagement compensation program and what amount will be implemented through a commission-based compensation program. because advisors at ibds have the additional process of determining a client’s compensation program, this adds extraneous cognitive load to the investment management task for ibd advisors, what we call, the compensation puzzle. in addition, because commission-based compensation is not impacted by subsequent returns, when ibd advisors place client assets in commission-based products, they may have lower incentives than ria advisors for their clients’ portfolios to perform well in the future. the compensation puzzle can be framed as a goal conflict between generating the highest return for clients and generating higher upfront compensation for an advisor. campbell (1988) states that the presence of goal conflict increases task complexity. thus, investment management is made more complex for ibd advisors due to the presence of the compensation puzzle. because the relationship between task complexity and performance is negative, we expect rias to perform better than ibds on the complex task of portfolio management. further, because financial planning involves an ongoing relationship, an advisor could be required to revisit this compensation puzzle multiple times, switching from primary to secondary tasks in the process. j. e. heller et al. / financial services review 30 (2022) 145–164 149 1.3. hypothesis the main hypothesis of this study states that due to the additional extraneous cognitive load levied against ibd advisors’ working memory capacity due to the presence of the compensation puzzle, net investment performance of ibds will be lower than that of independent rias. as such, we propose the following null and alternative hypotheses: h0: rias will not have significantly different net returns than ibds. h1: rias will have higher net returns compared with ibds, regardless of portfolio management approach. clt serves as the main justification for the hypothesis that rias will outperform ibds. the activity of investment portfolio creation and management is akin to a problem-solving exercise. modern portfolio theory (mpt) states that portfolios are created such that expected return is maximized for a given level of risk. each asset should be assessed based on its individual risk and return characteristics and how that asset contributes to the overall portfolio’s risk and return, emphasizing the importance of the correlations between the assets within the portfolio (markowitz, 1952). due to the high levels of elemental interactivity when engaging in portfolio construction and management, the intrinsic cognitive load placed on an advisor is high, requiring significant working memory capacity. because working memory capacity is limited, the potential addition of extraneous cognitive load from the compensation puzzle could lead to working memory capacity being exceeded and therefore, a reduced effectiveness in investment management. fig. 1 provide a graphical representation of the compensation puzzle and how it relates to the cognitive load and working memory capacity of advisors at rias and ibds. 2. method 2.1. data data were obtained through the generosity of a large, anonymous investment advisory platform. this platform provides a uniform tool that delivers advisor managed portfolios fig. 1. representation of the cognitive load and working memory capacity faced by advisors at ria and ibd firms. 150 j. e. heller et al. / financial services review 30 (2022) 145–164 (amp), unified managed accounts (uma), as well as turnkey asset management programs (tamp). they serve independent ria firms, ibds, insurance broker/dealers, banks, and trust companies. for purposes of this study, banks and trust companies were excluded due to the nature of their product and service offerings. banks and trust companies offer ancillary products and services that are outside the focus of investment management, and such offerings could influence the results. insurance broker/dealers were combined with ibds and referred to collectively as ibds because their dual registration as fee-for-service and commission-based advisors are quite similar for the two distribution channels. all advisors at the ria and ibd firms in this study receive fee-based compensation for the portfolios included in our analysis that they manage. in other words, ibds advisors in this study have decided to place their clients’ assets into fee-based models similar to those used by rias rather than to place them into commission-based products. amp are investment portfolios where the advisor maintains the responsibilities for the day-to-day investment management process, including formulating an investment strategy and asset allocation, conducing due diligence on the individual investments, implementing the strategy, and monitoring the portfolio and its component parts. amp can contain only individual securities, mutual funds, and/or exchange-traded funds (etfs). the amp data contains 1,585 records for amps for the one-year time period, 1,151 records for the threeyear time period, and 858 records for the five-year time period. tamp are investment portfolios where the day-to-day investment management process is completely handled by a third-party investment service provider. benefits of a tamp include outsourcing time-consuming activities such as investment research, portfolio allocation, and asset management tasks. a drawback of using tamps is that the originating advisor does not have direct control or input into the asset management process (kenton, 2018). the tamp contains the lowest amount of advisor responsibility for the investment management program of the three styles studied. the data contains 3,789 records for tamps for the one-year time period, 3,132 records for the three-year time period, and 2,434 records for the five-year time period. unified managed accounts (uma) are investment portfolios that act as a hybrid between amp and tamp portfolios. under a uma program, an advisor has the responsibility to create a high-level asset allocation for a portfolio as well as to conduct the due diligence on the component parts of the portfolio. the advisor is not responsible for rebalancing the portfolio like they would be in an amp; rather, these duties are handled by the investment platform. uma portfolios do not contain individual securities. instead, they contain mutual funds, etfs, tamps, and separately managed accounts (smas). while the advisor’s overall responsibility is less in the uma program compared with the amp, there are still day-today investment management responsibilities. the data contains 1,484 records for umas for the one-year time period, 1,163 records for the three-year time period, and 857 records for the five-year time period. data were provided on a firm level rather than at the account or advisor level. for each variable, the average value for each firm was provided. for example, the one, three, and five-year average returns per firm were provided for rias and ibds, for each of the three portfolio management approaches (i.e., amp, uma, and tamp) and across the risk tolerance categories that the platform uniformly employs. return data were provided for one, j. e. heller et al. / financial services review 30 (2022) 145–164 151 three, and five-year average returns for the time period ending on july 31, 2019, which means that the five-year average return data spanned august 1, 2014, to july 31, 2019. average account size, advisory fee, number of accounts, as well as number of advisors were provided as of july 31, 2019. because the data are as of a single point in time, time series analysis was not possible. firm-level data were provided to protect the identities of the individual advisors, clients, and firms that utilize the investment advisory platform as customers. the data set contains a total of 694 registered investment advisory firms and 723 independent broker/dealer firms, although many of these firms have a combination of amp, uma, and tamp portfolios. 2.2. empirical model the following ols regression model is used to test the hypothesis, if the distribution channel has a significant relation with one, three, and five-year average performance across the different risk tolerance categories regardless of the portfolio management approach (i.e., amp, uma, or tamp). the empirical model is run separately on subsamples of the data based on the portfolio management approach (amp, uma, and tamp). avg 1, 3, or 5� year return ¼ b 0 þ b 1 riað þ þ b 2 portriskð þ þ b 3 ria � portriskð þ þb 4 avgfeeð þ þ b 5in avgacctsizeð þ þ b 6 numaccountsð þ þ « (1) 2.3. dependent variables the dependent variables are the one, three, and five-year average return ending july 31, 2019. these returns are generated net of the advisor fee. accounts are included in each time frame if they have a long enough history. for example, a portfolio that has two years of return data will only be included in the analysis of one year of return data, whereas a portfolio with four years of return history will be included in the oneand three-year analyses. 2.4. independent variables the following independent variables included in the regression to determine if they impact the one, three, and five-year average rates of return of the portfolio. 2.4.1. ria this is a dichotomous variable that is positive for ria firms and zero for ibd firms. 2.4.2. portfolio risk the investment advisory platform utilizes five distinct universal risk tolerance levels. clients complete a questionnaire, which provides a risk tolerance rating. once the client’s risk tolerance rating is established, the platform will provide available tamp portfolios that 152 j. e. heller et al. / financial services review 30 (2022) 145–164 meet the client’s risk tolerance objective, account size, as well as the advisor’s licensing. for advisors who choose to employ amp or uma strategies, the client risk tolerance rating provides a risk range that the advisor must adhere to when constructing the portfolio. the platform ranks each available component investment and assigns a composite risk value. as component investments are added to the portfolio, the composite risk score for the portfolio is created and must remain within the client’s risk tolerance score to be considered compliant. the five risk tolerance categories in descending order from most conservative to most aggressive are: 1. capital preservation 2. conservative 3. moderate 4. growth 5. aggressive growth we expect that risk tolerance (manifest as portfolio volatility) and average one, three, and five-year returns will have a positive relationship. as portfolio volatility increases across the five risk tolerance categories, total net return will also increase, due to the additional equity allocations and increased risk premium. 2.4.3. ria*portfolio risk this interaction variable provides a measure of the marginal impact of increased portfolio volatility among ria firms. 2.5. control variables 2.5.1. average advisor fee this variable represents the average advisor fee (expressed as a percentage) for each portfolio, which does not represent the total cost to the client. advisor driven portfolios do not have manager fees that tamps (and umas) could have. additionally, firms charge different program fees that split revenue with the advisory platform; these fees are not included in the average advisor fee but could influence what the advisor chooses to charge. fees also tend to work on economies of scale; in other words, the larger the account, the lower the percentage fee charged. lastly, these fees are not what the advisor actually earns. each firm has a different compensation structure, and each advisor has a different payout, which could influence what the advisor chooses to charge. we expect the average advisor fee to have an inverse relation with each dependent variable. average advisor fees range from 0.000046% to 2.293% for ibds for the one-, three-, and five-year time periods. for rias, fees range from 0.0986% to 2.059% for the one-, three-, and five-year time periods. 2.5.2. ln (average account size) for all amp, uma, and tamp accounts, the average client account size is reported per firm as of july 31, 2019. average account size for ria amps ranges from $26,343 to j. e. heller et al. / financial services review 30 (2022) 145–164 153 $15,506,672. average account size for ibd amps ranges from $26,090 to $41,286,141. average account size for ria umas ranges from $27,221 to $45,829,626. average account size for ibd umas ranges from $26,259 to $4,504,614. average account size for ria tamps ranges from $25,073 to $13,600,157. lastly, average account size for ibd tamps ranges from $25,137 to $6,875,348. 2.5.3. number of accounts this variable represents the total number of accounts for each ria and ibd in each portfolio management approach. we expect the number of accounts and one, three, and fiveyear average returns to have an inverse relation with amp and uma performance. incidentally, we also expect the number of accounts and average account size to be inversely correlated. 3. results a description of the samples for the one-year time period is included in table 1. the average return for rias over one-year ranges from 3.6% for umas and 4.35% for amps, while the average return for ibds ranges from 3.25% for tamps and 4.53% for amps. average portfolio risk for rias and ibds range from 3.2 to 3.5. average account sizes by firm vary quite widely, from around $250,000 to over $800,000. the average fee charged is just under 1% across each of the models, and the number of advised accounts is considerably lower for rias than for ibds. to explore the relation between portfolio performance and business model (ria vs. ibd), we start by performing t-tests on the average returns for each of the nine models (i.e., three reporting time periods for each of the three portfolio management approaches). the results are displayed in table 2. four of the nine models had a statistically significant difference in the mean return between rias and ibds. in each of these instances, including all three tamp models, the returns of the rias were higher. before analyzing the full empirical model described previously, we performed a series of simplified regression models, as indicated in table 3. the initial model, regression #1, is a simple regression consisting of investment performance as the dependent variable and the key variable of interest, ria, as the only independent variable. regression #2 adds portfolio risk as an independent variable. regression #3 builds on the previous model by adding an interaction variable of ria and portfolio risk. finally, regression #4 incorporates all the control variables, including the average advisor fee, the natural log of average account size, and the number of accounts. each regression model was run separately on nine subsamples, one for each of the three portfolio management approaches (amp, uma, and tamp) for each of the three time periods (one-, three-, and five-year). table 4 shows the results for regression #1 for each of the portfolio management approaches in each of the three time periods. the main variable of interest (ria) is positive and significant in four of the models (one-year uma, and all three tamp time periods), consistent with the results in table 2. the other regressions do not have significant parameter estimates for rias. 154 j. e. heller et al. / financial services review 30 (2022) 145–164 t ab le 1 d es cr ip ti v e st at is ti cs fo r th e o n ey ea r sa m p le , se p ar at ed b y p o rt fo li o m an ag em en t ap p ro ac h (a m p , u m a , an d t a m p ) an d b y d is tr ib u ti o n ch an n el (r ia , ib d ) a m p u m a t a m p r ia ib d r ia ib d r ia ib d a v er ag e re tu rn 4 .3 5 % 4 .5 3 % 3 .6 0 % 3 .3 1 % 4 .1 7 % 3 .2 5 % p o rt fo li o ri sk 3 .2 3 .2 3 .5 3 .3 3 .3 3 .2 a v er ag e ac co u n t si ze $ 4 3 2 ,7 6 2 $ 3 7 7 ,1 7 6 $ 8 2 4 ,5 5 9 $ 4 3 7 ,2 6 6 $ 3 8 9 ,0 5 9 $ 2 5 0 ,6 9 3 a d v is o r fe e 0 .9 9 % 0 .9 7 % 0 .9 0 % 0 .9 4 % 0 .9 9 % 0 .9 3 % n o . o f ad v is o r ac co u n ts 7 6 2 9 0 3 9 2 7 4 5 3 2 3 2 n 4 1 3 1 ,1 7 2 7 2 4 7 6 0 1 ,5 8 8 2 ,2 0 1 j. e. heller et al. / financial services review 30 (2022) 145–164 155 table 5 shows the results using regression #2, which adds portfolio risk as an independent variable. portfolio risk is significant in all but one of the regressions. the parameter estimates for rias in each of the three tamp time periods are positive and significant. the parameter estimate for rias in the one-year uma time period is still positive but is no longer significant. once we control for portfolio risk, the parameter estimate for rias in the three-year amp and uma models are now significant and negative. table 6 shows the results using regression #3, which adds the interaction variable of ria and portfolio risk as another independent variable. in all but two of these regressions (the one-year amp and the five-year tamp), the parameter estimate for ria is now significant and negative. however, the parameter estimate for the interaction variable is also significant but positive in all but two of the regressions, suggesting the relationship between rias and investment returns includes a marginal effect dependent on the risk of the portfolio. the results seen in tables 5 and 6 suggest that the relative performance of rias may depend on the level of portfolio risk. to further explore this potential interaction, we perform t-tests on the returns of rias and ibds after separating the sample by risk category. the results are included in table 7. in the column for the capital preservation risk category, ibds have a statistically significant higher mean return than rias for the threeand fiveyear tamp time periods. at the opposite end of the risk spectrum, however, the column for the aggressive growth risk category shows statistically significant outperformance of rias table 3 progression of the regression models, indicating the independent variables that are included in each regression independent variables regression #1 regression #2 regression #3 regression #4 ria x x x x portfolio risk x x x ria*portfolio risk x x average advisor fee x ln (average account size) x number of accounts x table 2 mean one-, three-, and five-year returns for rias and ibds using amps, umas, and tamps one-year return three-year return five-year return ria amp 4.347% 6.607% 3.690% ibd amp 4.527% 6.824% 3.825% difference �0.180% �0.217% 0.501% ria uma 3.595% 6.211% 3.191% ibd uma 3.309% 6.344% 3.134% difference 0.286%* �0.133% 0.057% ria tamp 4.167% 6.272% 3.465% ibd tamp 3.247% 5.721% 2.964% difference 0.920%*** 0.551%*** 0.501%*** note. significant t-test results comparing the differences in the means are indicated with asterisks. ***p < 0.001, **p < .01, *p < .05. 156 j. e. heller et al. / financial services review 30 (2022) 145–164 over ibds by a considerable margin in five of the subsamples, including the five-year amp time period, the one-year uma time period, and all three tamp time periods. table 8 shows regression #4, which incorporates all the control variables in the empirical model. not surprisingly, advisor fees in almost every regression are significant and table 5 regressions #2 results for each of the portfolio management approaches, where one-, three-, and fiveyear returns are the dependent variables one-year return three-year return five-year return amp intercept 0.0435*** 0.0253*** 0.0177*** ria �0.0018 �0.0037* �0.0024 portfolio risk 0.0006 0.0134*** 0.0065*** r2 0.001 0.31 0.17 n 1,585 1,151 858 uma intercept 0.0270*** 0.0170*** 0.0082*** ria 0.0024 �0.0041*** �0.0010 portfolio risk 0.0019*** 0.0140*** 0.0070*** r2 0.010 0.45 0.26 n 1,484 1,163 857 tamp intercept 0.0347*** 0.0017 0.0020* ria 0.0093*** 0.0041*** 0.0039*** portfolio risk �0.0007* 0.0171*** 0.0086*** r2 0.024 0.59 0.33 n 3,789 3,132 2,434 note. portfolio risk is included in each of the regressions as an independent variable. ***p < 0.001, **p < .01, *p < .05. table 4 regression #1 results for each of the portfolio management approaches, where one-, three-, and fiveyear returns are the dependent variables one-year return three-year return five-year return amp intercept 0.0453*** 0.0682*** 0.0383*** ria �0.0018 �0.0022 �0.0014 r2 0.0005 0.0008 0.0007 n 1,585 1,151 858 uma intercept 0.0331*** 0.0634*** 0.0313*** ria 0.0029* �0.0013 0.0006 r2 0.0028 0.0007 0.0003 n 1,484 1,163 857 tamp intercept 0.0325*** 0.0572*** 0.0296*** ria 0.0092*** 0.0055*** 0.0050*** r2 0.0233 0.0073 0.0132 n 3,789 3,132 2,434 ***p < 0.001, **p < .01, *p < .05. j. e. heller et al. / financial services review 30 (2022) 145–164 157 negatively associated with returns. the natural log of average account size was positive and highly significant in all nine regressions. the interaction variable between ria and portfolio risk continues to be significant and positive in most of the regressions in table 8. for the regressions in table 8 where both the ria coefficient and the interaction variable coefficient are significant (the threeand five-year amp time periods, the one-year uma time period, and all three tamp time periods), the combined effect of rias is positive only for the higher risk categories and not for the lower risk categories. (this combined effect is calculated by using ria= 1 and risk category = 5 and multiplying by the corresponding parameter estimates.) in each of these instances, the aggressive growth risk categories shows that rias outperform ibds. in the one-year uma and all three tamp time periods, rias outperform ibds in the growth risk category as well (where risk category = 4). 4. discussion our hypothesis states that rias will outperform ibds regardless of the portfolio management approach (amp, uma, or tamp). this hypothesis was formulated based on the theoretical framework that rias expend less cognitive energy during the day-to-day activities of a practicing financial advisor due to the lack of the requirement to complete the table 6 regressions #3 results for each of the portfolio management approaches, where one-, three-, and fiveyear returns are the dependent variables one-year return three-year return five-year return amp intercept 0.0434*** 0.0278*** 0.0194*** ria �0.0016 �0.0166** �0.0108* portfolio risk 0.0006 0.0126*** 0.0060*** ria*portfolio risk �0.0001 0.0039** 0.0026* r2 0.001 0.31 0.17 n 1,585 1,151 858 uma intercept 0.0325*** 0.0185*** 0.0109*** ria �0.0098* �0.0075* �0.0081** portfolio risk 0.0002 0.0135*** 0.0062*** ria*portfolio risk 0.0036** 0.0010 0.0021* r2 0.017 0.46 0.26 n 1,484 1,163 857 tamp intercept 0.0414*** 0.0063*** 0.0051*** ria �0.0050* �0.0058** �0.0032 portfolio risk �0.0028*** 0.0157*** 0.0076*** ria*portfolio risk 0.0043*** 0.0030*** 0.0021*** r2 0.035 0.59 0.34 n 3,789 3,132 2,434 note. portfolio risk and an interaction term combining ria and portfolio risk are included in each of the regressions as independent variables. ***p < 0.001, **p < .01, *p < .05. 158 j. e. heller et al. / financial services review 30 (2022) 145–164 t ab le 7 d if fe re n ce s in m ea n re tu rn s fo r r ia s an d ib d s b y ri sk ca te g o ry d if fe re n ce s (r ia -i b d ) r is k ca te g o ry # 1 (c ap it al p re se rv at io n ) # 2 (c o n se rv at iv e) # 3 (m o d er at e) # 4 (g ro w th ) # 5 (a g g re ss iv e g ro w th ) a m p o n ey ea r re tu rn �0 .3 6 % �0 .3 0 % 0 .4 5 % �0 .5 9 % �0 .0 7 % t h re ey ea r re tu rn �0 .6 0 % �0 .8 0 % * * �0 .7 6 % * �0 .7 0 % * * 1 .1 9 % f iv ey ea r re tu rn �0 .4 3 % �0 .6 1 % �0 .4 2 % �0 .5 5 % 0 .7 1 % * * u m a o n ey ea r re tu rn �0 .9 1 % 0 .1 8 % 0 .1 7 % �0 .0 1 % 1 .0 0 % * * t h re ey ea r re tu rn 0 .1 5 % �0 .6 7 % * �0 .7 6 % * * * �0 .4 3 % * * �0 .0 7 % f iv ey ea r re tu rn �0 .1 8 % �0 .4 3 % �0 .3 0 % �0 .1 1 % 0 .3 6 % t a m p o n ey ea r re tu rn �0 .1 4 % 0 .7 5 % * * * 0 .0 5 % 0 .5 8 % * * 1 .8 1 % * * * t h re ey ea r re tu rn �0 .5 0 % * * * 0 .8 6 % * * * 0 .2 4 % �0 .1 6 % 1 .1 7 % * * * f iv ey ea r re tu rn �0 .3 3 % * 0 .4 7 % * 0 .1 5 % 0 .2 0 % 0 .7 6 % * * n o te . s ig n ifi ca n t tte st re su lt s co m p ar in g th e d if fe re n ce s in th e m ea n s ar e in d ic at ed w it h as te ri sk s. * * * p < 0 .0 0 1 , * * p < .0 1 , * p < .0 5 . j. e. heller et al. / financial services review 30 (2022) 145–164 159 compensation puzzle. the absence of this mental calculus, that ibds must perform for every client, frees working memory capacity to potentially utilize in investment management. our analysis confirmed that risk is an important aspect to consider when evaluating the relative performance of rias and ibds. overall, rias may not outperform ibds; however, when considering the risk category of the portfolio, our findings provide qualified support for our hypothesis that rias tend to outperform ibds for portfolios in higher risk categories. rias outperforming at higher risk categories can be explained through the theoretical framework and the equity risk premium, or the excess return above the risk-free rate provided to investors for taking on the additional risk of equity investments. based on data from 1928 to 2018, the geometric average annual equity risk premium is 6.11% over 3month treasury bills and 4.66% over the 10-year treasury bond (damodaran, 2019). however, to achieve this equity risk premium one must also assume increased risk. over the table 8 regressions #4 results for each of the portfolio management approaches, where one-, three-, and fiveyear returns are the dependent variables one-year return three-year return five-year return amp intercept 0.0145 �0.0150 �0.0343** ria 0.0006 �0.0158** �0.0115** portfolio risk 0.0009 0.0127*** 0.0060*** ria*portfolio risk �0.0007 0.0038** 0.0028* average advisor fee �2.5417*** �1.5082*** �1.7469*** ln (average account size) 0.0043*** 0.0046*** 0.0057*** number of accounts 0.0000001 0.0000004 0.0000005 r2 0.041 0.35 0.28 n 1,585 1,151 858 uma intercept �0.0381** �0.0241** �0.0388*** ria �0.0095* �0.0072* �0.0073* portfolio risk �0.0002 0.0132*** 0.0059*** ria*portfolio risk 0.0031** 0.0007 0.0014 average advisor fee �0.8754* �0.5646 �0.6910* ln (average account size) 0.0063*** 0.0038*** 0.0054*** number of accounts 0.0000027*** 0.0000032*** 0.0000027** r2 0.074 0.48 0.34 n 1,484 1,163 857 tamp intercept �0.0566*** �0.0478*** �0.0613*** ria �0.0076** �0.0077*** �0.0056** portfolio risk �0.0027*** 0.0156*** 0.0075*** ria*portfolio risk 0.0043*** 0.0031*** 0.0023*** average advisor fee �0.9896** �0.6072** �0.5024* ln (average account size) 0.0089*** 0.0050*** 0.0059*** number of accounts 0.0000002 0.0000004 0.0000005 r2 0.109 0.62 0.40 n 3,789 3,132 2,434 note. all the control variables are included as independent variables. ***p < 0.001, **p < .01, *p < .05. 160 j. e. heller et al. / financial services review 30 (2022) 145–164 same time period, from 1928 to 2018, the s&p 500 (including dividends) had a standard deviation of 19.58%, while the 3-month treasury bill had a standard deviation of only 3.04%, and the 10-year treasury bond had a standard deviation of 7.70%. with a higher variance of returns and a higher expected average return, equities can be considered a more difficult asset class to effectively value than fixed income. when valuing a bond, the primary concern is whether the issuing company has enough capital to honor the interest and principal repayments. although corporate profits are used to fund the capital requirements necessary to honor the covenants of a bond, the magnitude of corporate profits is not material in valuing a bond. bonds held to maturity also receive a fixed return, making valuations rather straight-forward. conversely, to properly value stocks, one must estimate future cash flows and discount those cash flows to the present. if a company does better than expected, equity shareholders could potentially be rewarded with increased dividends or improved share prices. bond holders, however, are not entitled to any additional compensation beyond the bond covenants. because equities are more difficult to value than fixed income instruments, they naturally require more cognitive load to analyze and evaluate. because advisors at rias can devote more working memory capacity toward the task of investment management (due to fewer extraneous cognitive load detractors such as the compensation puzzle), our study provides evidence that advisors at rias who focus more on equity-heavy portfolios are able to outperform their ibd counterparts. implications from these findings apply to both clients and advisors. clients with higher risk tolerance who wish to invest more in equities may be better served by employing advisors at rias rather than ibds. conversely, advisors may want to consider their competitive advantage as a financial professional. advisors at ibds, for example, may provide more benefit to clients with more conservative portfolios, while advisors at rias may have a competitive advantage on portfolios with more equity investments. 4.1. limitations we note that our study is not without limitations. for example, risk performance measures were not reported due to data limitations. while returns are key determinants of portfolio success, risk-adjusted returns would provide a more robust measurement of investment performance. in addition, individual account level data were not available, so firm level data were analyzed instead. because firms served as our unit of analysis, we made no attempt to measure the experience level of the advisors at the firms. experience could play a role in an advisor’s ability to manage investments that could influence the affect created by the business model. we also recognize that we have limited information about the advisors at the rias and ibds in our study and their clients. for example, details about the attitudes, skills, preferences, and beliefs of the advisors of the firms in our study would have enhanced our analysis. additional information about the clients of these firms would also have allowed for an analysis of potentially unobserved heterogeneity among client groups. we also recognize that the performance windows that were analyzed were rather small and at a single point in time, july 31, 2019. as such, the results of this study are heavily j. e. heller et al. / financial services review 30 (2022) 145–164 161 reliant on the capital market performance during the time periods preceding that date. in addition, the single point in time data limits the ability to analyze potential changes over time. for example, an account with a five-year track record could have seen its account size grow to the point where the advisor fee was decreased; however, the fee and size of the account were reported only as of the ending date, and changes in account sizes and fees were not observed. most importantly, we recognize that our results are correlational and do not indicate a direction of effect. although our results provide evidence that advisors at rias may perform differently than advisors at ibds because they have a different compensation motivation, it is also possible that each business model attracts different types of advisors. due to data limitations, we are not able to disentangle these possible explanations. 4.2. future research regulators continue to evaluate the role of advisor compensation in providing professional financial advice, as seen in the department of labor’s fiduciary rule and the security and exchange commission’s regulation best interest. in this discussion, one must also consider the role that business models play on portfolio performance. future research in this area that can include demographic information about advisors (e.g., education, age, gender, years in the profession, advanced designations, and disciplinary actions) would provide a better understanding of the effects business models have on advisor managed portfolio performance. in addition, analyzing performance over longer time periods, such as seven or even ten years, would provide greater insight into the long-term effects of business models on investor returns. 4.3. conclusion ample evidence both condemns professional financial advice (e.g., see bergstresser et al., 2009; del guercio et al., 2010; desai & jain, 1995; french, 2008; gil-bazo & ruiz-verdú, 2009; gruber, 1996; jensen, 1968; malkiel, 1995; reuter, 2015), and praising it (hackethal et al., 2012; gennaioli et al., 2015; kinniry et al., 2016; warshcuer & sciglimpaglia, 2012). however, the literature is scant regarding advantages or disadvantages provided to clients through the different business models available to advisors. this study sought to determine whether business models had an association with investment portfolio performance. the theoretical framework suggests that advisors at rias can eliminate the extraneous cognitive load created by the compensation puzzle and would be able to potentially redirect freed working memory capacity toward the difficult task of investment management. by virtue of having more working memory capacity to apply to the intrinsic cognitive load of investment management, we hypothesize that rias could outperform ibds regardless of the chosen portfolio management approach (amp, uma, or tamp) and regardless of risk category. our findings, however, provide qualified support of rias outperforming ibds through uma and tamp portfolios at higher risk categories but not at lower risk categories. while evidence is mixed regarding the efficacy of professional financial advice, this study provides qualified support for the hypothesis that business models have an association with 162 j. e. heller et al. / financial services review 30 (2022) 145–164 investment performance. qualified support exists in favor of rias producing higher net investment results through amp, uma, and tamp portfolios in higher risk categories when compared with ibds. references bergstresser, d., chalmers, j. m. r., & tufano, p. (2009). assessing the costs and benefits of brokers: a preliminary analysis of the mutual fund industry. review of financial studies, 22, 4129–4156. https://doi.org/ 10.1093/rfs/hhp022 botti, s., & iyengar, s. s. (2006). the dark side of choice: when choice impairs social welfare. journal of public policy & marketing, 25, 24–38. https://doi.org/10.1509/jppm.25.1.24 campbell, d. j. (1988). task complexity: a review and analysis. the academy of management review, 13, 40–52. https://doi.org/10.2307/258353 chalmers, j., & reuter, j. (2012). is conflicted investment advice better than no advice? nber working paper no. w18158. available at https://ssrn.com/abstract=2085144 chandler, p., & sweller, j. (1991). cognitive load theory and the format of instruction. cognition and instruction, 8, 293–332. https://doi.org/10.1207/s1532690xci0804_2 czerwinski, m., horvitz, e., & wilhite, s. (2004). a diary study of task switching and interruptions. in proceedings of the sigchi conference on human factors in computing systems, austria (pp. 175–182). new york, ny: association for computing machinery. https://doi.org/10.1145/985692.985715 damodaran, a. (2019). historical returns: stocks, t.bonds & t.bills with premiums [data file]. available at http://www.stern.nyu.edu/;adamodar/new_home_page/data.html del guercio, d., & reuter, j. (2014). mutual fund performance and the incentive to generate alpha. the journal of finance, 69, 1673–1704. https://doi.org/0.1111/jofi.12048 del guercio, d., reuter, j., & tkac, p. (2010). broker incentives and mutual fund market segmentation. nber working paper 16312. retrieved from https://ssrn.com/abstract=1667344 desai, h., & jain, p. c. (1995). an analysis if the recommendations of the “superstar” money managers at barron’s annual roundtable. journal of finance, 50, 1257–1273. foerster, s., linnainmaa, j., melzer, b., & previtero, a. (2017). retail financial advice: does one size fit all? journal of finance, 72, 1,441–1,482. french, k. r. (2008). presidential address: the cost of active investing. the journal of finance, 63, 1537–1573. https://doi.org/0.1111/j.1540-6261.2008.01368.x gennaioli, n., shleifer, a., & vishny, r. (2015). money doctors. the journal of finance, 70, 91–114. https:// doi.org/10.1111/jofi.12188 gil-bazo, j., & ruiz-verdú, p. (2009). the relation between price and performance in the mutual fund industry. the journal of finance, 64, 2153–2183. https://doi.org/10.1111/j.1540-6261.2009.01497.x gillie, t., & broadbent, d. (1989). what makes interruptions disruptive? a study of length, similarity, and complexity. psychological research, 50, 243–250. https://doi.org/10.1007/bf00309260 gonzàlez, v., & mark, g. (2005). managing currents of work: multitasking among multiple collaborations. in h. gellersen, k. schmidt, m. beaudouin-lafon, & m. mackay (eds.), ecscw 2005: proceedings of the ninth european conference on computer-supported cooperative work (pp.143–162). dordrecht, the netherlands: springer science. greenleaf, e. a., & lehmann, d. (1995). reasons for substantial delay in consumer decision making. journal of consumer research, 22, 186–199. https://doi.org/10.1086/209444 gruber, m. (1996). another puzzle: the growth in actively managed mutual funds. the journal of finance, 51, 783–810. https://doi.org/10.1111/j.1540-6261.1996.tb02707.x hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking & finance, 36, 509–524. https://doi.org/10.1016/j.jbankfin.2011.08.008 j. e. heller et al. / financial services review 30 (2022) 145–164 163 hodgetts, h., & jones, d. m. (2006). interruption of the tower of london task: support for a goal-activation approach. journal of experimental psychology: general, 135, 103–115. https://doi.org/10.1037/00963445.135.1.103 huberman, g., iyengar, s. s., & jiang, w. (2004). how much choice is too much: determinants of individual contributions in 401k retirement plans. in o.s. mitchell & s. utkus (eds.), pension design and structure: new lessons from behavioral finance (pp. 83–97). oxford: oxford university press. jacko, j. a., & ward, k. g. (1996). toward establishing a link between psychomotor task complexity and human information processing. computers & industrial engineering, 31, 533–536. https://doi.org/10.1016/03608352(96)00192-1 jensen, m. (1968). the performance of mutual funds in the period 1945-1964. the journal of finance, 23, 389– 416. https://doi.org/10.1111/j.1540-6261.1968.tb00815.x jin, j., & dabbish, l. a. (2009). self-interruption on the computer: a typology of discretionary task interleaving. in proceedings of the sigchi conference on human factors in computing systems, usa (pp. 1799–1808). new york, ny: association for computing machinery. https://doi.org/10.1145/1518701.1518979 karabulut, y. (2013). financial advice: an improvement for worse? available at https://ssrn.com/abstract= 1710634 kenton, w. (2018). turnkey asset management program (tamp). available at https://www.investopedia.com/ terms/t/turnkey-asset-management-program-tamp.asp kinniry, f. m., jr., jaconetti, c. m., dijoseph, m. a., zilbering, y., & bennyhoff, d. g. (2016). putting a value on your value: quantifying vanguard advisor’s alpha. vanguard research. available at https://advisors. vanguard.com/iwe/pdf/isgqvaa.pdf?cbdforcedomain=true leppink, j., van gog, t., paas, f., & sweller, j. (2015). cognitive load theory: researching and planning teaching to maximize learning. in j. cleland & s.j. durning (eds.), researching medical education (pp. 207– 218). hoboken, nj: john wiley & sons malkiel, b. g. (1995). returns from investing in equity mutual funds 1971-1991. the journal of finance, 50, 549–572. https://doi.org/10.1111/j.1540-6261.1995.tb04795.x markowitz, h. m. (1952). portfolio selection. the journal of finance, 7, 77–91. https://doi.org/10.2307/ 2975974 paas, f., renkl, a., & sweller, j. (2003). cognitive load theory and instructional design: recent developments. educational psychologist, 38, 1–4. https://doi.org/10.1207/s15326985ep3801_1 reuter, j. (2015). revisiting the performance of broker-sold mutual funds. available at https://ssrn.com/ anstract=2685375 shugan, s. m. (1980). the cost of thinking. journal of consumer research, 7, 99–111. https://doi.org/10.1086/ 208799 sweller, j. (1988). cognitive load during problem solving: effects on learning. cognitive science, 12, 257–285. https://doi.org/10.1207/s15516709cog1202_4 sweller, j., van merrienboer, j. j. g., & paas, f. (1998). cognitive architecture and instructional design. educational psychology review, 10, 251–296. https://doi.org/10.1023/a:1022193728205 warshcuer, t., & sciglimpaglia, d. (2012). the economic benefits of personal financial planning: an empirical analysis. financial services review, 21, 195–208. 164 j. e. heller et al. / financial services review 30 (2022) 145–164 finser_31-2_complete_issue financial planning time horizon and end-of-life mortality expectations zhikun liua,*, russell james iiib,*, qi sunc aemployee benefit research institute, 901 d street, s.w. suite 802 washington, dc 20024, usa bpersonal financial planning department, texas tech university, box 41210, lubbock, tx 79409-1210, usa cpacific life insurance company, 840 newport center drive, newport beach, ca 92660, usa abstract previous studies demonstrate that individuals’ financial planning time horizons significantly impact spending, saving, charitable giving, and bequest decisions. using longitudinal and cross-sectional health and retirement study data, the analyses in this paper reveal that older american adults’ financial planning horizons are strongly determined by their self-perceived life expectancy. over time, the changes in self-perceived life expectancy, marital and retirement status, health conditions, and wealth level will cause individuals to shift their financial planning horizons. the insight gained in this study helps financial planners to better understand the factors driving changes in client financial planning horizons and consequent financial decision-making. © 2023 academy of financial services. all rights reserved. jel classifications: g4; d12 keywords: behavioral finance; financial planning horizon; self-perceived life expectancy; financial decision making 1. introduction and literature review financial planning time horizon is an essential area of economic decision-making for both individuals and households (dow & jin, 2013). various studies have used financial planning horizon as an independent variable to predict various outcomes of interest (hong & hanna, 2014). *corresponding authors. tel.: +1-903-245-9598; fax: 1-202-775-6360; e-mail address: liuzhikun6@gmail.com (z. liu); tel.: +1-806-834-5130, fax: 1-806-742-5033; e-mail address: russell.james@ttu.edu (r. james) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 107–120 using the survey of consumer finances data, fisher and montalto (2010) identify that households with longer financial planning horizons are more likely to meet saving guidelines. rutherford and devaney (2009) find that households with more than five-year planning horizons are more likely to plan their consumption based on their income and use credit cards for convenience rather than revolving debt vehicles. a longer financial planning horizon also influences individuals’ risk preferences (castro-gonzález et al., 2020). for example, he and hu (2007) point out that households with longer financial planning horizons tend to hold relatively more stocks in their portfolios. when studying the medical cost risk, ayyagari and he (2017) find that when facing increased medical expenditure risk, people who have a planning horizon longer than five years are willing to take more equity exposure. liu et al. (2021) study how the financial planning horizon is associated with investors’ stock market return expectations. they conclude that having a financial planning horizon of one year or less is related to higher expectations of a 20% loss in the next year’s stock market. financial planning horizon also impacts the actual investment behavior. asebedo and browning (2020) found that a longer financial planning horizon is associated with a lower retirement portfolio withdrawal rate. munnell et al. (2001) conclude that employees with short financial planning horizons have a lower taste of saving and a smaller probability of participating in a pension plan. using the health and retirement study (hrs) data, liu and james (2017) find that individuals with longer financial planning horizons are more likely to have valid estate planning documents. liu and james (2020) find that such individuals are also more likely to make substantial gifts to charity. past studies found that the financial planning horizon is related to psychological and selfperceived financial well-being. for example, malroutu and xiao (1995) conclude that preretirees who plan to save for the next five years are more likely to perceive having adequate retirement income versus those who do not have saving plans. choung et al. (2022) found that people with major depression tend to have a shorter financial planning horizon. while most studies treat people’s financial planning horizon as a measure of time preference (khwaja, sloan, & salm, 2006). hong and hanna (2014) point out that it is unclear whether financial planning horizons reflect time preferences or situational variables. they indicate that the financial planning horizon is not measuring time preference but a situational variable because it is significantly related to the demographic variables of the respondents using the survey of consumer affairs (scf) data. they also suggest that time preference is constant, but the financial planning horizon is not constant. therefore, the financial planning horizon is not measuring time preference. however, as proposed by becker and mulligan (1997), time preference can be endogenously determined, and over time, investment in education and practice in imagining future outcomes will decrease time discounting. this is consistent with rising time preference into middle age (such as 43). additionally, time preference should be expected to change in later life based primarily upon changes in life expectancy. baranov and kohler (2018) find that individuals actively change their investment decisions based on their subjective longevity, even in a low-income environment. indeed, it would be irrational if an individual maintains a high preference for rewards paid over 30 years, as one’s life expectancy shortens with aging (moved from 40 to 108 z. liu et al. / financial services review 31 (2023) 107–120 20 to 10 to five years.) the explanation is consistent with an increasing rate of time discounting later in life, and it is the focus of the current study. some initial evidence appears inconsistent with this argument. however, if the measurement for the financial planning horizon does not extend beyond 10 or 20 years, then longevity predictions would not be expected to greatly influence this factor until the end of life. in particular, this occurs when subjective life expectancy falls below the highest financial planning measurement categories. it is this end-of-life subjective life expectancy effect that the current paper explores. the important implications of the households’ and individuals’ financial planning horizons motivate this study to investigate the different factors that affect people’s planning horizons. in particular, this study is interested in analyzing the relationship between people’s financial planning horizon and their self-perceived life expectancy. using the hrs data, both cross-sectional and longitudinal regressions demonstrate the existence of such a relationship. this paper provides robust evidence affirming the following hypothesis: individuals’ self-perceived life expectancies significantly affect their financial planning horizon. this relationship reveals that the financial planning horizon is, at least in part, a measure of rational (but not fixed) time discounting, based on life expectancy. 2. data this paper uses the hrs survey data to conduct cross-sectional and longitudinal regression analyses. variables are selected from the rand (version p) hrs data, the cross-wave tracker file, as well as the core hrs data sets from 1998 to 2018. since the hrs survey does not include direct life expectancy variables, this study constructs respondent-level, self-perceived life expectancy variables for each of the available hrs survey waves. the underlining principle for constructing such variables is described as follows: an individual’s selfperceived life expectancy is equal to her/his estimated probability of living to a target age multiplied by the difference between the target age and this individual’s current age. under this principle, the respondents’ self-perceived life expectancy variables are constructed using the formula below (the formula uses the 2018-year wave as an example; life expectancy variables are generated similarly for all the available waves): yearstolive2018 = r14liv10ð þ=100 # r14liv10a$ age2018ð þ in this formula, the variable “yearstolive(wave)” represents the respondent’s self-perceived life expectancy at the point of the survey. the variables r(wave)liv10 and r(wave) liv10a are selected from the rand hrs data. among them, r(wave)liv10 is the selfreported probability of living to a certain target age, selected according to the hrs guidelines, where 0 means “absolutely no chance,” and 100 means “absolutely certain.” for respondents whose age is less than 65 at the point of the survey, this target age is set to be 85. for respondents whose age is between 65 and 69, this target is set to be 80. for those between 70 and 74 years old, this target age is set to be 85. for those between 75 and 79, the z. liu et al. / financial services review 31 (2023) 107–120 109 target is 90. for respondents in the 80-84 age range, the target age is 95. finally, for respondents whose age is between 85 and 89, the target age is set to be 100. the variable r (wave)liv10a gives the specific age used in the questionnaire for each respondent during the survey, which ranges from 80 to 100. the summary statistics for the constructed self-perceived life expectancy variables (yearstolive(wave)) from the year 2000 to 2018 are reported in table 1. note that the rand hrs dataset does not contain the r(wave)liv10 and r(wave)liv10a variables before the 2000 wave. therefore, this summary statistics table uses the 2000-2018 wave range to demonstrate the respondents’ self-perceived life expectancies. the following figure is plotted using the 2018 wave data with individual-level weight applied. this graph indicates that the constructed life expectancy variable has an inverse relationship with the respondent’s age at the survey. this relationship is verified with ols regression: with high levels of significance, a one-year increase in respondents’ age will, on average, decrease their self-perceived life expectancy by 0.295 years. the control variables of this weighted ols regression include marital status, presence of children, years of education, wealth level, retirement status, and various health indicators. next, this study explores the relationship between the financial planning horizon variable and the constructed life expectancy variable. for the 2018 wave, two types of cross-sectional regressions (ols and ordered probit) are conducted. the summary statistics for the crosssectional regression variables are reported in the following table: 3. model becker and mulligan (1997) provide both theoretical and empirical evidence to conclude that time preference varies across individuals and wealth causes patience. trostel and taylor (2001) argue that future discounting occurs because the expected marginal utility of consumption is declining. in other words, an individual’s ability to enjoy future consumption is expected to be lower. if the financial planning horizon measures people’s time discounting preference, it should be strongly correlated with rational life expectancy changes as well as table 1 summary statistics for the imputed life expectancy variables variables number of observations mean (weighted) standard deviation (weighted) yearstolive2000 15,554 8.934495 0.0989592 yearstolive2002 14,420 8.322585 0.0827116 yearstolive2004 16,780 9.37001 0.0876633 yearstolive2006 15,467 8.93313 0.099443 yearstolive2008 14,450 9.078261 0.0825122 yearstolive2010 18,911 9.804211 0.1208808 yearstolive2012 17,925 9.176823 0.1041344 yearstolive2014 16,452 8.87597 0.1146203 yearstolive2016 12,429 9.110994 0.1268747 yearstolive2018 12,500 8.509271 0.1353772 note. respondent-level weights are applied to each wave. 110 z. liu et al. / financial services review 31 (2023) 107–120 wealth changes. the following model and the empirical analysis presented in the next sections of this paper provide both theoretical and empirical evidence for this hypothesis. in the general form of becker’s model of patience formation, a consumer is assumed to live a finite number of periods. therefore, the consumer maximizes v = o t i=0 b sð þi % fi cið þ (1) where t represents the length of the consumer’s life span, ci’s are the consumption levels at period i, the functions fi(·) map the consumptions at period i into pleasures at that period, aka utility function. future utilities are discounted based on the discount function b (·), which is less than 1. becker and mulligan define s to be the effort people make to increase their appreciation of future utility. s is determined by the time and effort spent in the appreciation of pleasures of the future, by spending on certain goods that distract one’s attention away from current pleasure, and toward future ones by schooling, saving, and so forth. in other words, s is the “patient” factor (becker & mulligan, 1997). applying the patience formation model to this paper, which analyzes the determinants of financial planning horizons, s could represent the effort and time spent by the respondents to make long-term financial plans. assuming the present values of all assets and earnings are calculated into an initial endowment of wealth a0, then the intertemporal budget constraint can be described as o t i=0 rici þ ps=a0 (2) where ri’s represent the interest factors and p stands for the price of s. the first-order conditions with respect to consumption at each period are as follows: b 0 sð þ o t i=0 i % b sð þ ! "i$1 % fi cið þ " # = k0 = f 0 0 c0ð þ (3) where l 0 denotes the marginal utility of wealth. the marginal benefit of s depends on the length of the consumer’s life (t), the level of future utilities, and the level of discount rate.1 from the first order conditions depicted by eq. (3), if an increase in life expectancy (from t to t + dt) is accompanied by an increase of lifetime earnings’ that maintains the marginal utility of wealth, l 0, then this change (dt) will increase the marginal benefit from investing in future-oriented capital – the patient factor s.2 therefore, longer lifetime, or self-perceived life expectancy increase, will directly motivate consumers to plan further into the future. hence, this model predicts a positive relationship between the consumers’ life expectancy and their financial planning horizons. 4. results to examine which factors and to what extent each of these factors drive the change in people’s financial planning horizons, this study compares the ols and ordered probit z. liu et al. / financial services review 31 (2023) 107–120 111 regression results cross-sectionally. this paper also conducts longitudinal robustness checks to examine whether intrapersonal changes in these factors cause changes in the financial planning horizon over time. table 2 reports the weighted means for the variables of interest in the 2018 wave. figure 1 demonstrates how the “years to live” variable measuring selfperceived life expectancy varies with age. table 3 reports the ordinary least squares regression results with the 2018 wave respondent level weight applied. the results in table 3 demonstrate significant relationships between the financial planning horizon variable and the independent variables such as self-perceived longevity, race, marital status, retirement status, and wealth levels. note that the age variable becomes significant after omitting the life expectancy variable from the ols regression. this finding indicates that the effect of aging on financial planning horizon is captured by the self-perceived life expectancy variable. similarly, the heart condition variable becomes marginally significant after omitting the life expectancy variable from the ols regression. this suggests that part of the effect of a heart condition diagnosis on financial planning horizon is captured by the self-perceived life expectancy variable. table 4 reports a more detailed analysis, showing the average marginal effect of the ordered probit regression for the 2018 wave data. this regression also takes the 2018 respondent-level sample weight into consideration. based on the results of both ols and the ordered probit regression, there is a strong relationship between the respondent’s self-perceived life expectancy and their financial planning horizon. table 2 summary statistics for the 2018 wave cross-sectional analysis variables mean standard deviation financial planning horizon (next few months ¼ 1; next year ¼ 2; next few years ¼ 3; next 5–10 years ¼ 4; longer than 10 years ¼ 5) 3.345433 1.171501 years to live (constructed variable measuring self-perceived life expectancy) 8.118878 5.552278 male (male ¼ 1; female ¼ 0) 0.4724504 0.4992765 age 68.72538 7.4852 black race (white ¼ 0; black ¼ 1; else ¼ 0) 0.0805977 0.2722359 other race (white ¼ 0; black ¼ 0; else ¼ 1) 0.0591926 0.2360019 married (married ¼ 1; else ¼ 0) 0.6936006 0.4610309 presence of child (have children ¼ 1; no children ¼ 0) 0.904557 0.2938471 retired (retired ¼ 1; else ¼ 0) 0.5474733 0.5326426 years of education 13.83241 2.736728 cancer (have cancer (excluding skin) ¼ 1; else ¼ 0) 0.1679238 0.3738256 heart condition (diagnosed with heart condition ¼ 1; else ¼ 0) 0.241735 0.4281654 wealth (natural logarithm of total wealth) 12.43344 1.98501 note. respondent-level sample weight of the 2018 wave hrs data is applied to the summary statistics. 112 z. liu et al. / financial services review 31 (2023) 107–120 both the ols and the ordered probit regression results show that a respondent’s financial planning horizon is strongly correlated with her or his self-perceived life expectancy, race, marital status, retirement status, and wealth level. ceteris paribus, individuals with longer self-perceived life expectancy, on average, tend to report longer financial planning horizons. this result provides direct evidence for the main hypothesis of this study and agrees with the prediction of becker and mulligan’s (1997) model of patience formation. married couples, on average, are more likely to report longer financial planning horizons. wealth level is also associated with the respondents’ financial planning horizon, indicated by both ols and the ordered probit results. everything else equal, people who have more wealth are more likely to plan relatively longer into the future. retirement status also stands out as a potential factor affecting financial planning horizon. interestingly, age is only a marginal significant determinant of people’s financial planning horizons. this result agrees with the findings of trostel and taylor (2001). the following graph (fig. 2) captures the average financial planning horizon in each age category. as one can see from the illustration above, the average financial planning horizons in different age groups do not follow a declining pattern with age increases. in other words, age is not closely related to self-reported financial planning horizon among these older american adults. dow and jin (2013) also point out that age is not a primary driver of an individual’s financial planning horizon, which agrees with the finding of this paper. however, this study does not support dow and jin’s (2013) argument, which indicates that “life expectancy is a poor predictor of financial planning horizon.” instead, robust evidence has been provided by this study supporting the hypothesis that older american adults’ self-perceived life expectancy affects their financial planning horizon positively and significantly. based on the model of patience formation from becker and mulligan (1997), the patience factor, which measures the respondents’ efforts and time spent to make long-term financial plans in this case, should decrease later in life based on life expectancy after peaking around the middle age. therefore, self-perceived life expectancy should affect financial planning time horizon at an older age when the life expectancy is closer to the financial planning horizon measurement fig. 1. average life expectancy across age spectrum. z. liu et al. / financial services review 31 (2023) 107–120 113 zones. since the respondents in the hrs data set are older american adults, one should see a significant impact of subjective longevity on the financial planning horizon. 5. longitudinal robustness check after confirming the cross-sectional relationship between the people’s financial planning horizon and their self-perceived life expectancy, the next step is to check whether this relationship persists with the same individual over time. as fig. 3 indicates, the average values of the respondent’s self-perceived life expectancies do not vary significantly across different hrs sample waves from the year 2000 to the year 2018. this figure demonstrates the consistency of our different sample waves. table 4 shows the fixed-effect longitudinal analyses and table 5 shows random-effect longitudinal analyses from the year 2000 to 2018. notice that the control variables that do not vary over time for our respondent sample (such as race, gender, years of education, etc.) are excluded from the longitudinal regression analyses. table 3 ols regression results on financial planning horizon variables coefficients coefficients (“years to live” variable omitted) years to live (constructed variable measuring self-perceived life expectancy) 0.0195*** (0.0040) male 0.0228 0.0005 (male ¼ 1; female ¼ 0) (0.0358) (0.0357) age $0.0047 $0.0115*** (0.0025) (0.0023) black race (white ¼ 0; black ¼ 1; else ¼ 0) $0.2500*** (0.0709) $0.2488** (0.0730) other race (white ¼ 0; black ¼ 0; else ¼ 1) $0.0794 (0.0744) $0.1288 (0.0725) married (married ¼ 1; else ¼ 0) 0.1264** (0.0399) 0.1437*** (0.0389) presence of child (have children ¼ 1; no children ¼ 0) $0.0752 (0.0679) $0.0634 (0.0637) retired (retired ¼ 1; else ¼ 0) 0.1063** (0.0326) 0.1005** (0.0327) years of education 0.0004 (0.0063) 0.0069 (0.0061) cancer (have cancer (excluding skin) ¼ 1; else ¼ 0) $0.0045 (0.0463) $0.0286 (0.0450) heart condition (diagnosed with heart condition ¼ 1; else ¼ 0) $0.0585 (0.0390) $0.0828* (0.0403) wealth (natural logarithm of total wealth) 0.1033*** (0.0152) 0.1083*** (0.0144) notes. the number of observations is 11,433 for the ols regression, including all the variables listed above. the number of observations is 11,826 for the ols regression without the “years to live” variable. respondentlevel ample weights are applied. standard errors are reported in parentheses. ***statistically significant at 0.1percent level. **statistically significant at 1percent level. *statistically significant at 5percent level. 114 z. liu et al. / financial services review 31 (2023) 107–120 t ab le 4 a v er ag e m ar g in al ef fe ct o f th e o rd er ed p ro b it re g re ss io n re su lt s v ar ia b le s a v er ag e m ar g in al ef fe ct s f in an ci al p la n n in g h o ri zo n n ex t fe w m o n th s n ex t y ea r n ex t fe w y ea rs n ex t 5 – 1 0 y ea rs l o n g er th an 1 0 y ea rs t h re sh o ld p ar am et er s 1 2 3 4 5 y ea rs to li v e $ 0 .0 0 33 * * * $ 0 .0 0 2 1 * * * $ 0 .0 0 2 1 * * * 0 .0 0 2 9 * ** 0 .0 0 4 5 * * * (0 .0 0 06 ) (0 .0 0 04 ) (0 .0 0 0 4 ) (0 .0 0 0 6 ) (0 .0 0 09 ) m al e $ 0 .0 0 30 $ 0 .0 0 1 9 $ 0 .0 0 1 9 0 .0 0 2 7 0 .0 0 4 2 (0 .0 0 58 ) (0 .0 0 37 ) (0 .0 0 3 6 ) (0 .0 0 5 1 ) (0 .0 0 79 ) a g e 0 .0 0 10 * 0 .0 0 0 6 * 0 .0 0 0 6 * $ 0 .0 0 0 9 * $ 0 .0 0 1 3 * (0 .0 0 04 ) (0 .0 0 03 ) (0 .0 0 0 3 ) (0 .0 0 0 3 ) (0 .0 0 06 ) b la ck ra ce 0 .0 3 35 * * 0 .0 2 1 6 * * 0 .0 2 1 0 * * $ 0 .0 2 9 9 * * $ 0 .0 4 6 2 * * (0 .0 1 11 ) (0 .0 0 74 ) (0 .0 0 7 3 ) (0 .0 1 0 1 ) (0 .0 1 57 ) o th er ra ce 0 .0 0 90 0 .0 0 5 8 0 .0 0 5 6 $ 0 .0 0 8 0 $ 0 .0 1 2 4 (0 .0 1 17 ) (0 .0 0 75 ) (0 .0 0 7 3 ) (0 .0 1 0 4 ) (0 .0 1 62 ) m ar ri ed $ 0 .0 1 83 * * $ 0 .0 1 1 8 * * $ 0 .0 1 1 4 * * 0 .0 1 6 3 * * 0 .0 2 5 2 * * (0 .0 0 63 ) (0 .0 0 41 ) (0 .0 0 4 0 ) (0 .0 0 5 7 ) (0 .0 0 86 ) p re se n ce o f ch il d 0 .0 1 27 0 .0 0 8 2 0 .0 0 8 0 $ 0 .0 1 1 3 $ 0 .0 1 7 6 0 .0 1 09 0 .0 0 6 9 0 .0 0 6 7 0 .0 0 9 7 0 .0 1 4 9 r et ir ed $ 0 .0 2 14 * * * $ 0 .0 1 3 8 * * * $ 0 .0 1 3 4 * * * 0 .0 1 9 1 * ** 0 .0 2 9 6 * * * (0 .0 0 52 ) (0 .0 0 35 ) (0 .0 0 3 5 ) (0 .0 0 4 6 ) (0 .0 0 75 ) y ea rs o f ed u ca ti o n 0 .0 0 00 0 .0 0 0 0 0 .0 0 0 0 0 .0 0 0 0 0 .0 0 0 0 (0 .0 0 10 ) (0 .0 0 07 ) (0 .0 0 0 6 ) (0 .0 0 0 9 ) (0 .0 0 14 ) c an ce r 0 .0 0 03 0 .0 0 0 2 0 .0 0 0 2 $ 0 .0 0 0 2 $ 0 .0 0 0 4 (0 .0 0 72 ) (0 .0 0 47 ) (0 .0 0 4 5 ) (0 .0 0 6 4 ) (0 .0 1 00 ) h ea rt co n di ti o n 0 .0 0 84 0 .0 0 5 4 0 .0 0 5 3 $ 0 .0 0 7 5 $ 0 .0 1 1 6 (0 .0 0 60 ) (0 .0 0 39 ) (0 .0 0 3 7 ) (0 .0 0 5 4 ) (0 .0 0 82 ) w ea lt h $ 0 .0 1 55 * * * $ 0 .0 1 0 0 * * * $ 0 .0 0 9 7 * * * 0 .0 1 3 8 * ** 0 .0 2 1 4 * * * (0 .0 0 24 ) (0 .0 0 15 ) (0 .0 0 1 4 ) (0 .0 0 2 1 ) (0 .0 0 33 ) n o te s. n u m b er o f o b se rv at io n s is 1 1 ,4 3 3 . r es p o n d en tle v el am p le w ei g h ts ar e ap p li ed . s ta n d ar d er ro rs ar e re p o rt ed in p ar en th es es . * * * s ta ti st ic al ly si g n ifi ca n t at 0 .1 p er ce n t le v el . * * s ta ti st ic al ly si g n ifi ca n t at 1 p er ce n t le v el . * s ta ti st ic al ly si g n ifi ca n t at 5 p er ce n t le v el . z. liu et al. / financial services review 31 (2023) 107–120 115 the results from the longitudinal ols regression confirm that the respondents’ financial planning horizons are strongly correlated with their self-perceived life expectancy. moreover, changes in self-perceived longevity result in changes to financial planning horizons. the marginal effects of the longitudinal ordered probit regression in table 6 confirm that, ceteris paribus, respondents with longer self-perceived life expectancy tend to have longer fig. 2. average financial planning horizon in different age categories (2018). note: the 2018 hrs respondentlevel sample weight is applied to this summary graph. fig. 3. average life expectancy summary and shift-trend (2000-2018). 116 z. liu et al. / financial services review 31 (2023) 107–120 financial planning horizons, controlling their marital and retirement status, presence of children, wealth levels, and health conditions. the longitudinal analysis results from the tables above affirm the main conclusion of this study: the respondent’s financial planning horizon is strongly correlated with rational life expectancy changes as well as wealth changes. hence, financial planning horizon is a partial measure of rational (not fixed) time discounting. overall, multivariate regression results show evidence that people’s life expectancy influences their planning horizon, even after controlling for their financial and health conditions. 6. conclusion using both cross-sectional and longitudinal analyses with the hrs data, this paper examines the potential factors that drive both differences and changes in american adults’ (age 50+) financial planning time horizons and reviews the significance of people’s financial planning horizon shifts. the empirical analysis in this study provides direct evidence to support the hypothesis that an individual’s financial planning horizon is strongly correlated with their self-perceived life expectancy, race, marital status, retirement status, and wealth level. ceteris paribus, people with longer self-perceived longevity tend to have longer financial planning horizons. over time, changes in self-perceived life expectancy, marital and retirement status, health conditions, and wealth level will cause individuals to shift their financial planning horizons. the findings of this paper can help financial planners to identify the factors, such as selfperceived longevity and wealth changes, that may cause potential shifts to the financial planning time horizon of their clients. practitioners can also benefit from this study in terms of educating their clients about the importance of their decisions on the financial planning table 5 longitudinal ols regressions on financial planning horizon variables fixed-effect random-effect years to live 0.00479*** 0.00847*** ($0.00121) ($0.000765) married $0.0614* 0.0692*** ($0.0265) ($0.0146) presence of child 0.0796 $0.0505* ($0.0715) ($0.0243) retired 0.0780*** 0.0340** ($0.0171) ($0.0125) cancer $0.0498 $0.016 ($0.0348) ($0.0195) heart condition $0.0541* $0.0519** ($0.0275) ($0.0164) wealth 0.0294*** 0.103*** ($0.00704) ($0.00329) notes. the number of observations is 49,690. ***statistically significant at 0.1percent level. **statistically significant at 1percent level. *statistically significant at 5percent level. z. liu et al. / financial services review 31 (2023) 107–120 117 horizon, which is one of the most important areas of economic decision-making for both individuals and households. the discoveries in this paper demonstrate that financial planning time horizon is not stable among older american adults and can be expected to change in later life, both rationally and empirically. therefore, practitioners should take into consideration that clients’ self-perceived life expectancy has a significant impact on their financial planning time horizon later in life. if the practitioners want to encourage longer financial planning horizon at the later life stages, non-consumption issues, such as bequest motives, may become more important. 7. future research this paper presents a new way to construct the variable that depicts the respondents’ selfperceived life expectancy using the health and retirement study data. this life expectancy variable has an inverse relationship with the respondent’s age and this inverse relationship is verified with ols regression analysis. more robust tests can be performed on this self-perceived life expectancy variable to examine whether it is a good measurement of the respondents’ longevity expectation and whether it can be used to construct an indicator for their mortality salience. once these tests are performed and approved positive, this variable can then be widely used as a measure of self-perceived longevity (or mortality salience) for studies in different financial planning fields such as bequest motives, investment choices, longterm care insurance adoptions, and estate planning decisions. table 6 marginal effects of longitudinal ordered probit regression variables average marginal effects financial planning horizon next few months next year next few years next 5-10 years longer than 10 years threshold parameters 1 2 3 4 5 years to live $0.0019*** $0.0009*** $0.0006*** 0.0014*** 0.0019*** (0.0001) (0.0001) (0.0000) (0.0001) (0.0001) married $0.0135*** $0.0062*** $0.0042*** 0.0104*** 0.0135*** (0.0027) (0.0012) (0.0009) (0.0021) (0.0027) presence of child 0.0129** 0.0059** 0.0040** $0.0100** $0.0129** (0.0045) (0.0021) (0.0014) (0.0035) (0.0045) retired $0.0083*** $0.0038*** $0.0026*** 0.0064*** 0.0083*** (0.0024) (0.0011) (0.0007) (0.0018) (0.0023) cancer 0.0031 0.0014 0.0010 $0.0024 $0.0031 (0.0036) (0.0017) (0.0011) (0.0028) (0.0036) heart condition 0.0102*** 0.0047*** 0.0032*** $0.0079*** $0.0102*** (0.0031) (0.0014) (0.0010) (0.0024) (0.0030) wealth $0.0190*** $0.0087*** $0.0059*** 0.0147*** 0.0190*** (0.0006) (0.0003) (0.0002) (0.0005) (0.0006) notes. the number of observations is 49,690. respondent-level sample weights are applied. standard errors are reported in parentheses. ***statistically significant at 0.1percent level. **statistically significant at 1percent level. *statistically significant at 5percent level. 118 z. liu et al. / financial services review 31 (2023) 107–120 literature review suggests that, everything else equal, individuals with longer financial planning time horizons tend to enjoy good implications of their financial decision making. for instance, long-term planners are more likely to meet certain saving guidelines, use credit cards for convenience instead of revolving debts, hold relatively more stocks in their portfolio, and have valid estate planning documents. future studies can further explore different behavioral strategies on how to educate clients about the importance of longer-term planning and promote this type of financial planning based on each client’s unique situation. notes 1 in the special two-period case, that is, t¼1, equation (2) can be simplified as b 0 sð þ f1 c1ð þ ! " = k0 = f 0 0 c0ð þ (4) 2 if the discount rate is unchanged, becker and mulligan (1997) suggest that consumption in the first t years would not be affected. the consumption during dt would be positive. by eq. (3), this will cause an increase of the marginal benefit from time preference investing. thus, the equilibrium accumulation of future-oriented capital is increased. references asebedo, s. d., & browning, c. m. (2020). the psychology of portfolio withdrawal rates. psychology and aging, 35, 78–90. available at: https://doi.org/10.1037/pag0000424 ayyagari, p., & he, d. (2017). the role of medical expenditure risk in portfolio allocation decisions. health economics, 26, 1447–1458. available at: https://doi.org/10.1002/hec.3437 baranov, v., & kohler, h. p. (2018). the impact of aids treatment on savings and human capital investment in malawi. american economic journal: applied economics, 10, 266–306. available at: https://doi.org/10.1257/ app.20150369 becker, g. s., & mulligan, c. b. (1997). the endogenous determination of time preference. the quarterly journal of economics, 112, 729–758. available at: https://doi.org/10.1162/003355397555334 castro-gonzález, s., fernández-lópez, s., rey-ares, l., & rodeiro-pazos, d. (2020). the influence of attitude to money on individuals’ financial well-being. social indicators research, 148, 747–764. available at: https://doi. org/10.1007/s11205-019-02219-4 choung, y., chatterjee, s., & pak, t. y. (2022). depression and financial planning horizon. journal of behavioral and experimental economics, 98, 101877. available at: https://doi.org/10.1016/j.socec.2022.101877 dow, j. p., jr., & jin, y. (2013). the determination of individual financial planning horizons. southwestern economic review, 40, 137–149. fisher, p. j., & montalto, c. p. (2010). effect of saving motives and horizon on saving behaviors. journal of economic psychology, 31, 92–105. available at: https://doi.org/10.1016/j.joep.2009.11.002 he, p., & hu, x. (2007). household investment – the horizon effect. available at ssrn. available at: https:// ssrn.com/abstract¼798431 or http://dx.doi.org/10.2139/ssrn.798431 hong, e. o., & hanna, s. d. (2014). financial planning horizon: a measure of time preference or a situational factor? journal of financial counseling and planning, 25, 184–196. khwaja, a., sloan, f., & salm, m. (2006). evidence on preferences and subjective beliefs of risk takers: the case of smokers. international journal of industrial organization, 24, 667–682. available at: https://doi.org/ 10.1016/j.ijindorg.2005.10.001 z. liu et al. / financial services review 31 (2023) 107–120 119 liu, y., guillemette, m., & gray, b. (2021). a behavioral understanding of stock market return expectations: from evaluation frequency and financial planning horizon perspective. available at ssrn 3906757. liu, z., & james, r. n. iii., (2017). behavioral and non-behavioral factors affecting will and trust ownership. available at ssrn https://ssrn.com/abstract¼3037748 liu, z., & james, r. n. iii., (2020). finding the next major donor: the relationship between financial planning horizon and charitable giving. journal of personal finance, 19, 49–65. malroutu, y. l., & xiao, j. j. (1995). perceived adequacy of retirement income. financial counseling and planning, 6, 17–23. munnell, a. h., sunden, a., & taylor, c. (2001). what determines 401(k) participation and contributions. social security bulletin, 64, 64–75. rutherford, l. g., & devaney, s. a. (2009). utilizing the theory of planned behavior to understand convenience use of credit cards. journal of financial counseling and planning, 20, 48–63. trostel, p. a., & taylor, g. a. (2001). a theory of time preference. economic inquiry, 39, 379–395. available at: https://doi.org/10.1093/ei/39.3.379 120 z. liu et al. / financial services review 31 (2023) 107–120 financial services review volume 32 number 1 (2024) volume 32, no. 1 2024 editor: john e. grable, ph.d. cfp ® university of georgia advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp®, baylor university (emeritus) martin seay, ph.d., cfp ® , kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp ® , san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia jasmine fang, ph.d., massey university, new zealand mark fedenia, ph.d., university of wisconsin stu heckman, ph.d., cfp ® , texas tech university william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea thomas langdon, ph.d., roger william university, bristol, ri terrance martin, ph.d., winston-salem state university wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc lance palmer, ph.d., cfp ® , cpa®, university of georgia abed rabbani, ph.d., cfp ® , university of missouri chris robinson, ph.d., york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida issn online 1057-0810 print 1873-5673 contents grable, john e., from the editor, i-ii. wang, ning, deng, yiling, & wu, ruohan. a dynamic analysis of the impact of household portfolio allocation decisions on the demand for life insurance. 1-28. anderson, jason, furlong, jeffery, & heckman, stuart. altruistic bequests: giving motive and receipt expectation using the 2019 survey of consumer sciences. 29-46. jeong, danah, hampton, benjamin, & archuleta, kristy l. how are you doing? financial wellbeing during covid-19. 47-62. chrétien, stéphane, & kammoun, manel. performance evaluation disagreement: determinants and impact on fund flows. 63-94. academy of financial services officers president shawn brayman smb research consulting executive vice president program michelle cull western sydney university vice president finance thanh ngo east carolina university vice president communications kirsten macdonald griffith university vice president international relations jasmine fang massey university vice president marketing & pr cora pettipas hsbc global wealth vice president membership matt goren dalton education, cerifi immediate past president tom potts baylor university editor, financial services review john e. grable, ph.d., cfp® university of georgia directors jason anderson university of kansas norah feng massey university wookjae heo purdue university thomas korankye the university of arizona barry mulholland university of akron richard stebbins university of alabama yu (yulia) chang kansas state university past presidents inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 32, no. 1, 2024 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • sherman hanna, ph.d., the ohio state university • terrance martin, ph.d., winston-salem state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • stu heckman, ph.d., cfp®, texas tech university • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • thomas langdon, ph.d., roger william university, bristol, ri • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • shinae choi, ph.d., university of alabama • john clinebell, ph..d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., york university, canada • lu fan, ph.d., cfp®, university of georgia • patti fisher, ph.d., virginia tech • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph..d. stetson university, deland, fl • philip gibson, ph.d., cfp®, winthrop university • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • john nofsinger, ph.d., university of alaska anchorage • mustafa nourallah, ph.d., centre for research on economic relations • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. it is a diamond open access journal which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peer-reviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ pii: s1057-0810(99)00005-0 book, software and website reviews money logic: financial strategies for the smart investor moshe a. milevsky, ph.d. with michael posner; stoddart publishing co., toronto, canada, 1999, isbn 0-7737-3171-7. moshe milevsky, a finance professor at york university, and michael posner, one of canada’s most respected financial journalists, have combined to produce an interesting and useful new book for the canadian investor.money logic: financial strategies for the smart investorexamines a wide range of investment decisions using the concept of probability of regret—the probability of one alternative underperforming another over the relevant time horizon—to evaluate investment risk. this is not an investments textbook. it lacks the breadth, rigor and ancillary materials needed for a course textbook.money logicis an excellent book for the canadian investing public. this easily understood book covers a wide range of investment issues including dollar cost averaging, mutual funds, international diversification and retirement planning using both fixed income and equity instruments. the book is thoroughly grounded in canadian tax law, canadian financial institutions and the canadian economic system. this makes the book far superior to most united states based personal investing books for a canadian investor, but makes the text inappropriate for most us investors. of special interest to canadian investors are the thorough discussion and analysis of the index-linked guaranteed investment certificate (ilgic) and the registered retirement savings plan (rrsp). the rrsp is covered over several chapters including both the accumulation phase and retirement alternatives. canadian investors should give this book a try. it is interesting, well written and useful. douglas r. kahl professor of finance, university of akron, college of business administration, department of finance, akron, oh 44325-4803, usa e-mail address:kahl@uakron.edu financial services review 7 (1998) 217 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00005-0 academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2021 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 29, no. 3, 2021 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2021 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. from the editor this issue contains volume 29 issue 3 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “the relationship between objective financial knowledge, financial management, and financial self-efficacy among african american students” is coauthored by kenneth white at university of georgia, narang park at texas state university, kimberly watkins at university of alabama, megan mccoy at kansas state university, and joycelyn morris at florida international university. the authors examine the factors contributing to african american college students’ financial literacy. using the national student financial wellness study and structural equation modeling, their findings suggest that for african american students, objective financial knowledge is not directly or indirectly associated with financial self-efficacy. they find that only financial management is significantly associated with increased financial self-efficacy. their findings posit that experiential learning may be effective for improving african american students’’ financial literacy. the second article “income more important than financial literacy for improving wellbeing” is coauthored by tracey west at griffith university, michelle cull at western sydney university and dianne johnson at griffith university. the authors study the impact of financial literacy on financial behaviors. they do not find that university students with higher levels of financial literacy have reduced money management stress and positive financial behavior, leading to higher levels of financial wellbeing. they do find that being older and having higher levels of income contributed most significantly and consistently to explaining better financial wellbeing. the third article, “the effect of risk literacy and visual aids on portfolio choices among professional financial planners” is coauthored by meghaan r. lurtz at university of maryland university college, michael g. kothakota at wolfbridge wealth management, stuart j. heckman at kansas state university, and kristy archuleta at university of georgia. in this article the authors explore the impact of risk literacy on the ability to understand and interpret probabilistic trade-offs. the authors use an experimental design to test financial planners’ risk literacy and their ability to select the most resilient portfolio based on whether they were given probabilistic information and a visual representation or only 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) v–vi probabilistic information. their results indicate that visual representation does help financial planners determine the appropriate choice, but risk literacy does not. the final article, “enumerating the value of financial advice in a competitive market: a dual structure approach & analysis” is coauthored by steve p. fraser at florida gulf coast university, brian c. payne at university of nebraska at omaha, and scott schatzle at mutual trust advisory group. in this article, the authors introduce and examine a composite, dual fee structure (cdfs) for financial planners that helps quantify the value of financial advice. they specifically separate financial planning (advice) fees based on total net worth (nw) from investment management (im) fees based on assets under management (aum), which are readily observable and pervasive in the marketplace. the authors state that with this knowledge, the financial value of the non-im component of financial planning services can reduce perceived conflicts of interest by permitting financial planners to generate compensation for non-im planning activities in a transparent manner, whether or not the client moves investable funds to the planner. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review s. michelson / financial services review 29 (2021) v–vi vi your mileage may vary manoj athavalea*, stephen avilaa, joseph goebela adepartment of finance and insurance, ball state university, 2000 west university avenue, muncie, in 47306, usa abstract the traditional model of retirement planning centers around the accumulation of a portfolio during the earning years followed by a drawdown from this portfolio after retirement. this drawdown is intended to support a desired retirement lifestyle, and central to a successful retirement is the sustainability of the retirement portfolio over the expected planning horizon. we define portfolio success to mean the ability of the retirement portfolio to sustain a desired retirement lifestyle over the desired planning horizon, and use simulations and logistic regressions to evaluate the impact of asset allocation, the profile of portfolio returns, the withdrawal rate, and the length of the planning horizon on portfolio success. our analysis shows that the likelihood of success is inversely related to withdrawal rate, retirement horizon, and portfolio risk increase, and directly related to portfolio return, allocation aggressiveness, and early experience. the analysis also indicates that portfolio success is highly sensitive to withdrawal rates, with conservative allocations exhibiting greater variation in portfolio outcomes and aggressive allocations providing more dependable portfolio outcomes for retirees who desire higher withdrawal rates. © 2022 academy of financial services. all rights reserved. jel classifications: g51; g53 keywords: retirement planning; portfolio success; retirement income; withdrawal rate; sequence risk 1. introduction in the process of building wealth and trying to ensure a financially secure retirement, individuals typically go through two distinct phases: an accumulation phase over the course of their career, and a postretirement drawdown or withdrawal phase. these withdrawals from *corresponding author: tel.: +1-765-285-5200; fax: +1-765-285-4314. e-mail address: athavale@bsu.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 125–143 the accumulated retirement portfolio are intended to sustain the desired lifestyle over the retired lifetime. traditionally, retirees typically relied on social security, pension benefits, and personal savings to sustain their retirement lifestyle. social security benefits in the united states are designed to only partially replace earnings that workers lose when they retire, and average only about 40% of total retirement income (biggs & springstead, 2008). further, employer-sponsored defined-benefit pension plans have been declining over the years, and are being replaced by defined-contribution pension plans. the prospective retiree has to rely on social security retirement income and their retirement portfolio comprised of the defined-contribution plan balance and personal savings. the cash flow characteristics of each source of retirement income is different. social security retirement income is similar to a defined-benefit plan with fixed inflation-indexed payments over the life of the retiree with some survivor benefits and no potential for a bequest. the traditional defined-contribution plan requires the retiree to take minimum distributions (rmd) that fluctuate every year based on the actual plan balance and the relevant life expectancy factor for that year. a retiree who lives longer than expected may face a substantially depleted plan balance and reduced cash flow towards the later retirement years, while a retiree who dies earlier than expected may leave a large bequest. elective withdrawals can be made from roth defined-contribution plans and from personal savings, and any balance remaining in either is available for a bequest. these elective withdrawals serve to supplement retirement income available from other sources. an additional consideration affecting retirement cash flows is the different tax implications of each of these sources. an important financial issue that most prospective retirees contend with, and has been extensively examined by financial planners and in the financial planning literature is: “how much can i spend each year from the retirement portfolio without completely depleting the retirement portfolio?” the most common response to this question is the “4% rule” which in its most general form, suggests that a retiree with a diversified portfolio could make inflation-adjusted annual withdrawals equal to 4% of the initial portfolio balance with a low chance of depleting the retirement portfolio over a 30-year retirement horizon. while it is popularly termed a rule, both financial planners and the academic literature understand that it is a guideline, a starting point for discussion, and that the safe withdrawal rate (swr) should be modified based on individual circumstances. this discussion suggests that the ability of the retirement portfolio to successfully sustain the desired lifestyle would depend on various factors such as asset allocation, portfolio return characteristics, the desired retirement horizon, and of course the desired withdrawal rate. while the literature suggests that retirees remain flexible to individual circumstances, there is very little guidance about the manner in which these factors impact portfolio success. all retirees do not make the same retirement portfolio choices and consequently may face very different retirement outcomes. we create plausible retirement scenarios and simulate variations in these factors to determine portfolio success or failure. we then use logistic regression to measure the impact that these factors have had in determining portfolio success. unlike previous studies, we define portfolio outcome as a binary variable and use logistic analysis to highlight the likelihood of portfolio success as a function of five retirement planning horizons, three asset allocation strategies, and five fixed real withdrawal rates to arrive 126 m. athavale et al. / financial services review 30 (2022) 125–143 at 75 plausible unique retirement scenarios. for each scenario, we use four distinct monthly return generating distributions and run 100 simulations to arrive at 30,000 portfolio outcomes and use these in a logistic regression to examine the determinants of portfolio success or failure. we show that the probability of portfolio success is inversely related to withdrawal rate, retirement horizon, and portfolio risk, and directly related to portfolio return, allocation aggressiveness, and the returns experience in the five years immediately after retirement. this paper is important because it applies an interesting methodology to a large number of retirement scenarios and portfolio outcomes to enable a retiree better understand how their unique set of circumstances and choices impact retirement success. the rest of this article is organized as follows: we begin by reviewing some of the relevant literature to provide the necessary context for the current research. we then present the empirical design of our research, data definitions, sample statistics, and the results of our regression analysis. finally, we present our conclusions, practical implications, and suggestions for future research. 2. background and literature review the traditional view of retirement is that in the early years people earn, save and invest (the “accumulation” stage) to subsequently retire and withdraw from their retirement portfolio to finance consumption over their retirement horizon (the “decumulation” or “asset distribution” stage). in the decumulation stage, individuals balance the competing goals of maintaining consumption in retirement without prematurely depleting their retirement portfolio. one of the early solutions to the problem of creating retirement cash inflows was the state retirement pension program created in 1889 by chancellor bismark of germany that provided a pension starting at age 70 when life expectancy was merely an additional two years. however, with improved working conditions and increased life expectancy, the amount of time spent in retirement today is now both longer and also a larger proportion of total life expectancy. retirement income is now needed for a few decades rather than a few years. the reduction in the number of defined-benefit plans and an increase in defined-contribution plans, transfers risk from the employer to the retirement saver. employees are now responsible for the saving decision, the asset allocation decision, and the asset withdrawal decision, while simultaneously accepting the real danger of premature retirement portfolio depletion (“portfolio failure”). the determination of a sustainable withdrawal rate that would reduce the probability of portfolio failure has been addressed in a number of prior studies generally using the overlapping periods methodology or simulation methodology. a historical analysis of the overlapping retirement experiences of individuals retiring between 1926 and 1980 led to the early consensus that a retiree with a diversified portfolio could make inflation-adjusted annual withdrawals equal to 4% of the initial portfolio balance with a low chance of depleting the retirement portfolio over a 30-year retirement horizon. this 4% rule refers to the popular withdrawal rate that originated from studies like bierwirth (1994), bengen (1994), and cooley, hubbard, and walz (1998) that were meant to dispel the notion that higher m. athavale et al. / financial services review 30 (2022) 125–143 127 withdrawal rates that matched the historic average real returns on a diversified portfolio (between 5% to 6%) were sustainable over the retirement horizon. saving for retirement is challenging, and most employees have little training upon which to draw in making the relevant decisions (benartzi & thaler, 2007, p.102). similarly, most retirees lack the skills required to manage their retirement portfolio successfully, highlighting the need for practical and easily understood solutions to generating sustainable retirement income (or, alternately, the need for an experienced and reputable financial advisor). merton (2014, p.1408) states that requiring people to save for retirement is reasonable, but expecting them to acquire the expertise necessary to make investment and withdrawal decisions is not reasonable. further, cognitive functions may decline in retirement (bonsang, adam, & perelman, 2012) and while the portfolio decisions of older investors may reflect greater knowledge about investing, their investment skill does deteriorate with age due to the adverse effects of cognitive aging (korniotis & kumar, 2011). this is where retirees may find the 4% rule to be useful, and it is certainly a reasonable and intuitive starting point in the retirement planning process. indeed, cooley, hubbard, and walz (1998, p.16) argue that individual experiences may vary due to personal behavioral traits, circumstances, and goals, and that no single rate appears appropriate for every investor. moreover, the finding of a 4% swr itself has been subject to various challenges. pfau (2010) contends that the early consensus may be an artifact of the data used in the analysis for a couple of reasons. first, the use of rolling 30-year periods emphasizes data from the middle of the period and hence introduces temporal bias in these analyses. second, from an international perspective, a 4% real withdrawal rate would have been “safe” in only four of 17 developed countries. these results are consistent with dimson, marsh and staunton (2004) who explained that the united states has had higher real returns and lower market volatility during the 1900 to 2002 period when compared with many other countries. these results are also consistent with estrada (2018, p. 62) who examined the retirement experience across 21 countries and 115 years using 11 asset allocations. using equally weighted returns to a balanced portfolio across all countries in the sample, they find that a retiree with a 30-year retirement horizon would face a 50% probability of failure with a 4.8% withdrawal rate and could only withdraw 2.6% if a 5% probability of failure was desired. further, the maximum withdrawal rate varied substantially, leading them to conclude that individuals who retired in some countries or at certain points in time had vastly different standards of living than those who retired in other countries or at other points in time. the success of any retirement portfolio certainly depends on the expected return assumptions used in the analysis. pye (2000) showed that 4% withdrawals from an equity portfolio with 8% real return and 18% standard deviation could be sustained for 35 years with an 81% chance of success, but a higher 4.5% withdrawal rate could be achieved by allocating 60% of the portfolio to treasury inflation protected securities (tips) assuming a certain 3.7% real return. finke, pfau, and blanchett (2013) used simulations to review the safe withdrawal rate in the current low-yield environment to conclude that a 30-year retirement portfolio would have a failure rate of 18% if yields revert to their historic mean in 5 years. similarly, blanchett, finke, and pfau (2014) test the sustainable withdrawal rate in a low bond-yield environment and use a drift model of bond yields to show that a 4% initial withdrawal rate has just a 50% probability of success over a 30-year retirement horizon. 128 m. athavale et al. / financial services review 30 (2022) 125–143 thus, there is some evidence that the demonstrated success of the 4% rule is partly an anomaly of historic u.s. market returns and assumptions of expected returns. the implication from these studies is that the historical asset returns used in the overlapping periods model are not suitable for forward-looking forecasts on which retirement withdrawal strategies should be based. further, asset returns experienced in the last decade appear to have disrupted the conventional thinking about the safe withdrawal rate, and athavale and goebel (2011) reinforced the notion that the 4% rule constitutes a probabilistic model and past success does not guarantee future success. in addition to returns, there is some evidence that the standard deviation of returns and the sequence of returns may impact the success of a retirement portfolio. blanchett and blanchett (2008) investigate the relative importance of portfolio return and standard deviation on portfolio success using the standard 4% withdrawal rate over a 30-year period. they find that a 1% reduction in returns is likely to result in an increase in the probability of failure that is approximately four times greater than a 1% increase in portfolio standard deviation, leading them to conclude that portfolio returns have greater impact on the probability of portfolio success compared with standard deviation. clare, seaton, smith, and thomas (2017, 2021) suggest that the sequence of returns matters in both the accumulation and decumulation stages, and show that a portfolio-timing strategy using the cyclically adjusted price-to earnings (cape) ratio can help investors mitigate sequence risk and achieve higher withdrawal rates. in addition to the overlapping periods model and the simulations model, researchers have explored the use of other sophisticated models to design withdrawal strategies. for example, milevsky and robinson (2005) use investment risk and return, mortality estimates, and spending rates in a stochastic present value framework to investigate the relationship between withdrawal rates and the probability of portfolio failure. they find that 4% withdrawals by a 65-year old retiree invested in a balanced portfolio has a 9% chance of portfolio failure, and a 3.24% withdrawal rate has a 5% chance of portfolio failure, leading them to conclude that payout ratios should be lower than generally recommended. scott, sharpe, and watson (2009) suggest a strategy that includes buying and selling 30-year european call options on the market portfolio over the planning horizon to replicate the traditional 4% withdrawals, but acknowledge that many practical issues remain to be addressed before this utility maximizing methodology can be incorporated in retirement planning. recent studies have focused on dynamic rule-based multi-asset allocation and liquidation strategies, and switching from fixed withdrawals to variable withdrawals to improve the probability of portfolio success. a case study of the retirement experience of a 1973 retiree invested in a balanced multi-asset portfolio led guyton (2004) to conclude that systematic decision rules and some restriction on subsequent inflation adjustments could allow for a 5.8% initial withdrawal rate. other examples of dynamic withdrawal strategies include adjusting the withdrawal rate based on portfolio performance and remaining life expectancy to improve retirement portfolio success and average lifetime withdrawal rates (stout & mitchell, 2006); using a multiasset portfolio with periodic adjustments to the asset mix and a “bonds first” withdrawal strategy that mitigates the higher volatility in equity returns (liu, chang, de jong, & m. athavale et al. / financial services review 30 (2022) 125–143 129 robinson, 2009); and calculating the probability of portfolio failure each year and changing the withdrawal rate based on decision rules (blanchett & frank, 2009). the importance of analyzing portfolio success rates in determining withdrawal rates has previously been emphasized by cooley, hubbard, and walz (2011). they assert that changes should be made to withdrawal rates in response to unexpected changes in financial market conditions, and use the overlapping periods methodology to present portfolio success rate tables for various combinations of withdrawal rates, portfolio compositions, and payout periods. dynamic adjustment strategies have been shown to be relevant in both the accumulation stage (estrada, 2019) and in the drawdown stage (estrada, 2020) of the retirement portfolio. specifically, estrada (2020) shows that dynamic strategies outperform a static strategy of sticking to the plan, and periodic adjustments to the withdrawal rate is superior to adjusting portfolio asset allocations. similarly, robinson and tahani (2010) treat portfolio return, longevity, and consumption as stochastic variables and use an analytical model to conclude that changing consumption to match changes in wealth could reduce the risk of portfolio failure. these dynamic withdrawal strategies do have intuitive appeal. it is logical to calibrate asset allocations and withdrawals to changing circumstances and economic realities. an unresolved question is whether retirees would have the discipline to follow decision rules and would have the flexibility to reduce consumption. these strategies are still in the early stages of their development and we need a better understanding of the manner in which relevant variables impact portfolio success (dejong & robinson, 2017). retirees may become better equipped to make these changes if they understand the implications of economic circumstances and their actions on the probability of portfolio success. our current research, therefore, is an effort to better understand the determinants of retirement portfolio success. 3. hypothesis development and empirical design each individual entering retirement has to decide about the asset allocation for their retirement portfolio, the expected retirement horizon, and the desired withdrawal rate from their retirement portfolio. in making these decisions, the retiree faces the tradeoff between maximizing consumption during retirement while minimizing the probability of prematurely exhausting the retirement portfolio. in this context it is important to understand the composition of the retirement portfolio. as previously described, the retirement portfolio may comprise some proportion of the traditional defined-contribution plan balance, the roth defined-contribution plan balance, and personal savings. each of these is taxed differently; consequently, a million dollars in a traditional retirement account is not equivalent to a million dollars in personal savings which, in turn, is not equivalent to a million dollars in a roth retirement account. generally, withdrawals from the traditional plan balance are taxable as ordinary income; the capital gains arising from assets liquidated from personal savings prior to withdrawal are taxable at a reduced capitalgains rate; and withdrawals from roth plan balances are not taxable. therefore, all references to a retirement portfolio should be on a tax-equivalent basis, and for the purpose of this 130 m. athavale et al. / financial services review 30 (2022) 125–143 research, to avoid the differential effect of taxes on retirement withdrawals, we implicitly assume that the retirement portfolio referenced here has been aggregated on an after-tax basis. it is also necessary to differentiate between the terms withdraw and consume. while the purpose of withdrawals is to finance a desired level of consumption, rmd rules may result in a withdrawal different from that necessary to finance that level of consumption. in the event rmd rules require a withdrawal greater than that needed for a desired lifestyle, we implicitly assume that the prudent retiree would reinvest the excess so as to reduce the risk of premature portfolio depletion. prior research has documented that the probability of portfolio success is impacted by withdrawal rates, asset allocation, retirement horizon, and measures of actual portfolio performance, including return, standard deviation of returns, and the sequence of returns. the retiree makes decisions about the withdrawal rate, the planning horizon, and the asset allocation for the retirement portfolio. however, the retirement portfolio will be affected by economic circumstances and chance, factors which are outside the retiree’s control, but which will nevertheless affect the actual return, standard deviation, and sequence of returns that the retirement portfolio may experience. we define a retirement portfolio to be a success if the portfolio can sustain a specified level of withdrawal over the entire retirement horizon. conversely, a portfolio that is fully consumed within the retirement horizon is a “failure.” retirees need to be cognizant of the factors that can lead to portfolio failure, and while some of these factors cannot be controlled, other factors (most commonly, the withdrawal rate) can be managed to mitigate the risk of portfolio failure. while financial planners and prior academic research encourage retirees to remain flexible with their retirement expenditures, retirees may not be aware of the impact that these variables may have on the probability of a successful retirement and may therefore be illequipped to make these decisions or respond to circumstances. we seek to identify those factors that contribute to portfolio success, and measure the impact that each of these factors will have on the probability of portfolio success. portfolio success lies at the intersection of the planning horizon, asset allocation, and withdrawal strategy (collins, lam, & stampfli, 2015, p.194), and the probability of portfolio failure (also called “ruin”) is a useful risk metric that can help retirees understand the link between their withdrawal strategy, planning horizon, and the asset allocation of their retirement portfolios (milevsky & robinson, 2005, p. 99). we use portfolio success as the dependent variable in our analysis. we model the retirement experience as a sequence of annual adjustments, with the initial retirement portfolio growing or shrinking according to the portfolio returns in the first year followed by a withdrawal at the end of the year to finance retirement expenditures. the remaining portfolio balance then grows or shrinks according to the portfolio returns in the second year followed by an inflation-indexed withdrawal, and this progression continues over the duration of the retirement. the decisions that the retiree makes about asset allocation, the retirement horizon, and the desired withdrawal rate are used as inputs in our analysis. we assume that the retiree prefers fixed real withdrawal rates as they are easy to understand and provide the retiree with constant purchasing power. annual portfolio returns and standard deviation are a function of m. athavale et al. / financial services review 30 (2022) 125–143 131 the selected asset allocation but reflect the uncertain external environment, and are determined by simulation. the simulated annual portfolio returns also determine if the retirement portfolio encounters an unfortunate sequence of negative returns during the early retirement years. the early sequence of negative returns may decimate the portfolio and significantly impair the portfolio’s ability to grow and generate income, decreasing the probability of a successful retirement (blanchett et al., 2014, p. 55). evaluation of retirement strategies involves the consideration of a large number of simulated or historical retirement periods and the subsequent estimation of their failure rate. the use of simulations in retirement planning has both proponents and opponents, and sandidge (2020) explains that that simulations are ineffective because most people lack the numeric skills needed to assess probability. collins, lam, and stampfli (2015) reviewed retirement income modeling strategies and state that the simulation methodology overcomes the limitation of relying on past returns as the basis of potential outcomes, thus allowing for a much greater range of potential outcomes. however, the inputs that drive these models need to be realistic. cooley, hubbard, and walz (2003, p. 128) find that success rates differ when using monte carlo simulation methodology compared with the overlapping periods methodology, and recommend the use of simulation methodology for the longer payout periods which are important in retirement planning. we intend to simulate a large number of scenarios representing the wide spectrum of possible retirement experiences. we then follow the previously described sequence of annual adjustments over the retirement horizon to determine the outcome of each scenario, that is, whether each of the scenarios ends in portfolio success or portfolio failure. and finally, we will model the portfolio outcome (success or failure) using a logistic regression that generally takes the form: ln p 1� p � � ¼ a þob ixi þ « where p is the probability of portfolio success, and the left-hand term is the log-odds. the explanatory variables include the retiree choice variables (withdrawal rate, asset allocation, and retirement horizon) and the chance variables (return, risk, and the early returns experience). these variables are defined in table 1. the binary logistic regression is typically preferred when modeling a dichotomous outcome variable. we intend to use a logit model because the dependent variable is a binary categorical variable, equaling one when the portfolio successfully sustains the desired withdrawal rate over the entire retirement horizon, and zero when the outcome is portfolio failure. the logistic regression allows us to identify and analyze the impact of factors that influence portfolio success in a multivariate setting. 4. data and method at the start of the retirement period, individuals can make choices about the retirement planning horizon, asset allocation, and the withdrawal rate. individuals retiring today are living longer than prior generations and spending longer periods of time in retirement. for the 132 m. athavale et al. / financial services review 30 (2022) 125–143 t ab le 1 v ar ia b le n am es an d d es cr ip ti o n s n am e d es cr ip ti o n w it h d ra w al ra te a re ti re e d et er m in ed w it h d ra w al ra te ex p re ss ed as a p er ce n ta g e o f th e in it ia l p o rt fo li o b al an ce . w it h d ra w al s fr o m th e p o rt fo li o at th is ra te , ad ju st ed fo r in fl at io n , ar e in te n d ed to fi n an ce a fi x ed le v el o f re al co n su m p ti o n o v er th e re ti re m en t h o ri zo n . w e h av e u se d fi v e w it h d ra w al ra te s in th e an al y si s (2 .5 0 % , 3 .2 5 % , 4 .0 0 % , 4 .7 5 % , an d 5 .5 0 % ). a ll o ca ti o n a re ti re e d et er m in ed p o rt fo li o al lo ca ti o n . w e h av e u se d th re e al lo ca ti o n s in th e an al y si s (c o n se rv at iv e, b al an ce d , an d a g g re ss iv e) th at d if fe r b as ed o n ex p ec te d re tu rn an d ri sk . t h is is tr ea te d as a ca te g o ri ca l v ar ia b le . h o ri zo n e x p ec te d lo n g ev it y o f th e re ti re e’ s p o rt fo li o . t h is is ex p ec te d to m at ch th e re ti re m en t h o ri zo n an d is d et er m in ed b y th e re ti re e at re ti re m en t b as ed o n ag e at re ti re m en t, li fe ex p ec ta n cy , h ea lt h , an d li fe st y le ch o ic es . w e h av e u se d fi v e h o ri zo n s in th e an al y si s (2 3 , 2 6 , 2 9 , 3 2 , an d 3 5 y ea rs ). r et u rn t h e si m p le av er ag e an n u al re al ra te o f re tu rn th at w o u ld h av e b ee n ea rn ed b y th e p o rt fo li o o v er th e re ti re m en t h o ri zo n . r et u rn is a ch an ce v ar ia b le an d is a fu n ct io n o f n o t o n ly p o rt fo li o al lo ca ti o n b u t al so se le ct io n , fe es , an d so fo rt h . t h is v ar ia b le in tr o d u ce s d if fe re n ce s in o u tc o m es b et w ee n re ti re es an d d ev ia ti o n fr o m ex p ec te d re tu rn in sm al l sa m p le s. t h is in fo rm at io n b ec o m es k n o w n at th e en d o f th e p la n n in g h o ri zo n an d ca n n o t b e u se d fo r m id te rm co u rs e co rr ec ti o n . t h is v ar ia b le is u se d in th e an al y si s as an en v ir o n m en ta l co n tr o l v ar ia b le . r is k t h e st an d ar d d ev ia ti o n o f th e an n u al re al ra te o f re tu rn , an d li k e r et u rn , is a ch an ce v ar ia b le . t h is in fo rm at io n b ec o m es k n o w n at th e en d o f th e p la n n in g h o ri zo n an d ca n n o t b e u se d fo r m id te rm co u rs e co rr ec ti o n . t h is v ar ia b le is u se d in th e an al y si s as an en v ir o n m en ta l co n tr o l v ar ia b le . e ar ly ex p er ie n ce t h e ra ti o o f th e ac tu al p o rt fo li o b al an ce at th e en d o f th e fi ft h y ea r to th e ex p ec te d b al an ce as su m in g ce rt ai n ty in re tu rn s. t h is is a ch an ce v ar ia b le th at ca p tu re s a se q u en ce o f u n fa v o ra b le re tu rn s ea rl y d u ri n g th e re ti re m en t ex p er ie n ce . w h il e th is v ar ia b le is n o t d et er m in ed b y th e re ti re e, it ca n n ev er th el es s b e u se fu l in co n si d er in g m id te rm co rr ec ti o n s. n o te . t h is ta b le d es cr ib es th e ex p la n at o ry v ar ia b le s th at ar e u se d in th is an al y si s. s o m e o f th es e v ar ia b le s ar e re ti re e ch o ic e v ar ia b le s (w it h d ra w al ra te , p o rt fo li o al lo ca ti o n , an d p la n n ed re ti re m en t h o ri zo n ) w h il e o th er s ar e ch an ce v ar ia b le s (r ea li ze d re tu rn , re al iz ed ri sk , an d se q u en ce ri sk p ro x ie d b y th e n o v el ea rl y ex p er ie n ce v ar ia b le ). m. athavale et al. / financial services review 30 (2022) 125–143 133 average american, life expectancy for males, females, and married couples is 82, 85, and 89, respectively, and the probability that one member of a married couple will live to age 95 is 18% (browning, 2016, p.51). the problem with determining the correct retirement horizon is that retirees do not know precisely when they will die. further, while in theory a retirement portfolio is meant for consumption, in practice most retirees will feel comfortable under-consuming the retirement portfolio and planning for a long retirement rather than risking portfolio failure (longevity risk). our analysis assumes that retirees will select one of five retirement planning horizons (23, 26, 29, 32, or 35 years) based on their health and lifestyle. our analysis recognizes that each retirement situation is different and this uniqueness is captured through the intentional use of a wide range of values for the independent variables, allowing the analysis to apply to many retirement scenarios. for example, assuming life expectancy of 90 years, our choice of planning horizon is wide enough to cover both the traditional age 67 retiree and with a 23-year retirement horizon and the age 55 early retiree with a 35-year retirement horizon. our analysis also assumes that a retiree will select one of three asset allocation strategies (a conservative strategy with an emphasis on fixed-income investments, with expected real return of approximately 3.1% and standard deviation of 8%, a balanced strategy with expected real return of 5.1% and standard deviation of 12%, and an aggressive strategy with an emphasis on equity investments, with expected real return of 7.1% and standard deviation of 16%) based on their personal risk tolerance. these numbers reflect the approximate averages of the mean real return and standard deviation reported in prior research referenced in this paper. finally, our analysis also assumes that a retiree will select one of five fixed real withdrawal rates (2.5%, 3.25%, 4%, 4.75%, and 5.5%) based on their consumption needs. the use of a range of annual withdrawal rates, retirement planning horizons, and stock allocations, is consistent with cooley et al. (1998). these three decisions about retirement horizon, asset allocation, and withdrawal rate result in 5 � 3 � 5 = 75 unique retirement scenarios. the next step in the analysis is to consider likely outcomes for each of these scenarios. for example, does a 2.5% withdrawal rate from a conservative portfolio sustain the retirement portfolio over a 23-year retirement horizon? while we know that our conservative portfolio will average annual returns of approximately 3.1% and have a standard deviation of 8% over long periods of time, returns in any particular year can fluctuate away from the average with a wide range of uncertain values. the standard methodology in such cases draws random returns for each year of the retirement horizon from a theoretical distribution. we should test our hypothesis by drawing random annual returns for the first year, making 2.5% withdrawals, noting the portfolio balance at the end of the year, and if the portfolio has not been depleted, continuing this exercise for a total of 23 years. drawing random annual returns to the retirement portfolio requires us to impose a priori assumptions about the functional form of the distribution of expected returns, and a standard assumption is that returns are characterized by the normal distribution. this assumption, though convenient, was empirically challenged by fama (1965) who found that the distribution of monthly stock returns belonged to a non-normal member of the stable class of distributions. subsequently, officer (1972) confirmed that the distribution of stock returns has fattails, and gray and french (1990) confirmed that the distribution of stock index returns also 134 m. athavale et al. / financial services review 30 (2022) 125–143 deviates from the normal distribution. the preponderance of empirical evidence finds that return distributions are not normally distributed (kring, rachev, höchstötter, fabozzi, & bianchi, 2009, p. 272), and have rejected the normal distribution in favor of either a skewed distribution or a fat-tailed distribution (levy & duchin, 2004 p. 48). we relax the assumption that returns follow any single distribution and use four different continuous probability distributions (beta, kumaraswamy, pert, and triangular) from which to draw random annual returns. in the absence of theoretical arguments or empirical evidence to guide our selection of the appropriate distribution, we selected four continuous distributions which allowed negative returns, allowed us to specify bounds, and displayed skewness and fat tails. the parameters of the distributions were set to ensure consistency with the desired mean and standard deviation, and reasonable boundaries were established. thus, each of the 75 previously mentioned scenarios were tested using four different returns distributions. continuing our prior example, we define a retirement portfolio to be a success if 2.5% withdrawals from a conservative portfolio could be sustained over a 23-year retirement horizon, when realized returns followed the beta distribution. in such cases we assign outcome = 1, and in cases where the retirement portfolio is prematurely depleted before the end of the retirement horizon, we assign outcome= 0. this gives us a total of 75 � 4 = 300 retirement experiences, reflecting both the choices the retiree has made and the returns uncertainty that impacts portfolio outcomes. another returns uncertainty that we considered in the analysis is that portfolio outcomes may also be impacted by the sequence of returns obtained. a sequence of large negative returns early in the retirement period may hasten portfolio depletion. sequence risk (or serial returns risk) refers to the risk of premature portfolio depletion caused by a combination of withdrawals and significant negative returns early in retirement. we proxy for sequence risk by constructing an early experience variable, defined as: early experience ¼ observed balance 5, 2:5%, 3:1%, 12%ð þ expected balance 5, 2:5%, 3:1%, 0%ð þ where, the numerator is the observed portfolio balance at the end of the fifth year after 2.5% withdrawals each year from a conservative portfolio earning 3.1% average real returns that follow the beta distribution with a standard deviation of 12%, while the denominator is what the balance would be assuming certainty in portfolio returns. a better early experience implies a greater probability of portfolio success, and we would expect a positive relation between the early experience variable and the outcome variable. the observed portfolio balance at the end of the fifth year and the value of the early experience variable at the end of the fifth year are presented in table 2. the process described above was repeated 100 times. thus, the 300 retirement experiences simulated 100 times each gives us 300 � 100 = 30,000 portfolio outcomes. another way of thinking about this is that the simulations are a way of testing the retirement scenarios to see the potential outcome of many possible trajectories and to gauge how vulnerable the scenarios are to portfolio failure. our model is consistent with pfau (2012) in that portfolio success is dependent on the interaction of withdrawal rates, capital market conditions, retirement durations, and asset allocation, and the spitzer, strieter, and singh (2007) m. athavale et al. / financial services review 30 (2022) 125–143 135 assertion that a blanket four percentage withdrawal rule may be an oversimplification of a complex set of circumstances. 5. descriptive statistics and empirical results the frequency of success (outcome= 1) or failure (outcome= 0) among the 30,000 portfolio outcomes described in the previous section is presented in table 3. portfolio success occurred 83.3% (24,989 of 30,000) of the time among the observed portfolio outcomes. the aggressive portfolio had an average success rate of 87.83% (8,783 of 10,000) while the conservative portfolio has a success rate of 76.34%. as expected, the portfolio with a 2.50% withdrawal rate had an average success rate of 99.13% (5,948 of 6,000) while the portfolio with a 5.50% withdrawal rate had a success rate of 59.08%. and finally, as expected, the portfolio with a 23-year planning horizon had an average success rate of 91.3% (5,478 of 6,000) while the portfolio with a 35-year horizon had a success rate of 75.73%. we had previously indicated that the retiree decides about portfolio allocation (based on risk tolerance), withdrawal rate (based on consumption needs), and retirement horizon (based on expected life expectancy and lifestyle choice), and we can disaggregate the observed portfolio success rates using these variables. these disaggregated observed portfolio success rates are presented in table 4. table 2 the early (returns) experience portfolio allocation withdrawal rate average portfolio balance early experience mean minimum maximum conservative 2.50% 1,033,998 100.20 47.78 170.27 3.25% 994,054 100.20 46.39 172.39 4.00% 954,110 100.21 44.89 174.71 4.75% 914,166 100.21 43.26 177.23 5.50% 874,222 100.22 41.47 179.97 balanced 2.50% 1,145,076 100.10 38.63 206.12 3.25% 1,103,597 100.10 37.27 208.77 4.00% 1,062,118 100.11 35.80 211.64 4.75% 1,020,638 100.12 34.21 214.73 5.50% 979,159 100.13 32.48 218.10 aggressive 2.50% 1,284,861 101.56 22.52 274.61 3.25% 1,241,476 101.61 21.20 278.86 4.00% 1,198,091 101.65 19.77 283.43 4.75% 1,154,706 101.70 18.25 288.35 5.50% 1,111,322 101.75 16.59 293.65 note. this table presents information about the observed portfolio balance at the end of the fifth year and the value of the early experience variable. the early experience variable is a proxy for sequence risk which refers to the risk of premature portfolio depletion caused by a combination of fixed withdrawals and significant negative returns in the early years of retirement. there are 2,000 observations in each of the 15 combinations of portfolio allocation and withdrawal rate for a total of 30,000 observations. 136 m. athavale et al. / financial services review 30 (2022) 125–143 a few observations are notable. a withdrawal rate of 2.5% can largely be sustained irrespective of portfolio allocation and retirement horizon. however, higher withdrawal rates (5.5%) over longer retirement horizons (35 years) have a success rate of 65.5% with an table 3 frequency table of observed portfolio outcomes failure success by portfolio allocation conservative 2,366 7,634 balanced 1,428 8,572 aggressive 1,217 8,783 5,011 24,989 by withdrawal rate 2.50% 52 5,948 3.25% 253 5,747 4.00% 731 5,269 4.75% 1,520 4,480 5.50% 2,455 3,545 5,011 24,989 by retirement horizon 23 years 522 5,478 26 years 750 5,250 29 years 1,004 4,996 32 years 1,279 4,721 35 years 1,456 4,544 5,011 24,989 note. n= 30,000 portfolio outcomes. this table presents the number of successes (or failures) among the 30,000 portfolio outcomes in our sample, aggregated by portfolio allocation, or withdrawal rate, or retirement horizon. these numbers suggest that conservative portfolios, higher withdrawal rates, and longer retirement planning horizons increase the chance of portfolio failure. table 4 observed portfolio success rates portfolio allocation withdrawal rate retirement horizon 23 years 26 years 29 years 32 years 35 years conservative 2.50% 100.00% 100.00% 99.50% 99.75% 99.00% 3.25% 99.75% 98.75% 96.00% 95.75% 91.25% 4.00% 96.00% 93.00% 83.25% 79.50% 71.00% 4.75% 88.25% 76.75% 59.75% 50.25% 37.25% 5.50% 66.50% 47.75% 38.75% 23.50% 17.25% balanced 2.50% 99.75% 99.50% 99.75% 98.50% 99.00% 3.25% 99.25% 97.50% 96.00% 94.50% 93.25% 4.00% 95.50% 92.75% 90.50% 84.50% 84.00% 4.75% 86.50% 83.50% 83.00% 73.50% 68.25% 5.50% 77.25% 70.75% 65.50% 56.75% 53.75% aggressive 2.50% 98.75% 99.00% 98.00% 98.50% 98.00% 3.25% 96.50% 96.00% 94.75% 94.00% 93.50% 4.00% 94.50% 92.50% 88.50% 84.50% 87.25% 4.75% 89.25% 86.00% 82.25% 77.75% 77.75% 5.50% 81.75% 78.75% 73.50% 69.00% 65.50% note. this table allows us to evaluate the simultaneous impact of portfolio allocation, withdrawal rate, and retirement horizon on portfolio success rates. portfolio success occurs in 24,989 of the 30,000 portfolio outcomes. m. athavale et al. / financial services review 30 (2022) 125–143 137 aggressive portfolio allocation, but only 53.75% with a balanced portfolio allocation and 17.25% with a conservative portfolio allocation. a withdrawal rate of 4% over a 29-year horizon has a success rate of 90.50% with a balanced portfolio allocation, 88.50% with an aggressive portfolio allocation and 83.25% with a conservative portfolio allocation. and finally, a conservative portfolio exhibits wide variations in outcomes ranging from 100% to 17.25%, while an aggressive portfolio allocation exhibits variations ranging from 99.00% to 65.50%. these results confirm that aggressive portfolios improve the probability of portfolio success for higher withdrawal rates. we next use regression analysis to analyze the impact of factors which influence portfolio success in a multivariate setting. our variable of interest is portfolio success that is a binary dependent variable. in such cases, the ordinary least squares technique can be nonconforming and the estimates of the dependent variable can go out of bounds (0, 1). the logistic regression technique is well suited to examining the relation between portfolio success and the predictor variables as it keeps the predicted values of the dependent variable within expected bounds. at the start of the retirement period, the retiree makes decisions about the withdrawal rate, the planning horizon, and the asset allocation for the retirement portfolio. the retirement portfolio will also be affected by actual returns, standard deviation, and sequence of returns that the retirement portfolio may experience. while these factors are outside the retiree’s control and will nevertheless impact portfolio outcome, the retiree may be able to observe and act on any early unfavorable sequence of returns that the portfolio may experience. these then, are the explanatory variables used in the analysis. the logistic function is used to estimate, as a function of unit changes in the independent variables, the probability that the event of interest will occur. our logistic model provides a good fit for the data if we can demonstrate an improvement over the intercept-only model, and we check this using the akaike information criterion and the schwarz criterion. we also note that the cox and snell r2 is 46%, nagelkerke’s rescaled r2 is 78%, and mcfadden’s pseudo r2 is 69%. in addition, the maximum likelihood coefficient estimates are all individually significant at 1% using the wald x2 test. direct interpretation of the logistic regression coefficients is difficult since coefficient estimates are in terms of log-odds. the estimated coefficients do not represent the marginal effects of the independent variables on the probability of portfolio success. instead of the coefficients being the rate of change in the dependent variable as the independent variable changes, a coefficient derived from a logistic regression is interpreted as the rate of change in the log-odds as the independent variable changes. exponentiating the coefficient gives us the odds ratio, which can range from 0 to infinity, and which allows for somewhat easier interpretation. the model coefficients and the odds ratio are presented in table 5. while the odds ratio is somewhat easier to interpret than the coefficients of the logistic regression, neither is as useful as the traditional marginal effect. unlike a linear regression, the marginal effect is not constant across the entire range of values of the explanatory variable, and hence the marginal effect is calculated at each observation in the dataset, and then averaged. this average marginal effect indicates expected changes in the predicted probability of portfolio success as a function of a change in an explanatory variable while keeping 138 m. athavale et al. / financial services review 30 (2022) 125–143 other covariates constant. the average marginal effect for each explanatory variable is also presented in table 5. the marginal effect of the withdrawal rate variable indicates that the probability of portfolio success changes by �0.1311 for a 1-level change in the withdrawal rate. similarly, the marginal effect of the retirement horizon variable indicates that the probability of portfolio success changes by �0.0154 for a 1-level change in the retirement horizon. early experience is a continuous variable and hence the marginal effect is defined as the partial derivative of the probability of portfolio success with respect to early experience. similarly, higher levels of return increase the probability of portfolio success while higher levels of risk decrease the probability of portfolio success, and this is consistent across all allocations. finally, an aggressive allocation changes the probability of portfolio success by 0.1142 compared with a balanced allocation, and a conservative allocation changes the probability of portfolio success by �0.0993 compared with a balanced allocation. table 5 determinants of portfolio success coefficient wald v2 odds ratio marginal effect intercept 18.4342 (0.6080) 919.2899 withdrawal rate �3.1056 (0.0552) 3161.9001 0.045 �0.1311 allocation aggressive 2.7052 (0.7069) 14.6443 14.957 0.1142 allocation conservative �2.3530 (0.5821) 16.3402 0.095 �0.0993 horizon �0.3638 (0.00870) 1746.6138 0.695 �0.0154 return � allocation conservative 1.5348 (0.0412) 1384.7595 4.641 0.0648 return � allocation balanced 1.3722 (0.0358) 1466.0887 3.944 0.0579 return � allocation aggressive 1.2288 (0.0320) 1477.4365 3.417 0.0519 risk � allocation conservative �2.648 (0.0424) 38.9240 0.767 �0.0112 risk � allocation balanced �0.3202 (0.0377) 72.1836 0.726 �0.0135 risk � allocation aggressive �0.4087 (0.0317) 166.5041 0.664 �0.0173 early experience 0.0781 (0.00178) 1916.2730 1.081 0.0033 r2 69% note. n= 30,000. standard errors are placed below the coefficient. all coefficients are significant at the 1% level. this table presents the results of the logistic regression analysis used to model the probability of portfolio success, where portfolio success is defined as the ability of a retirement portfolio to sustain a desired lifestyle over a desired retirement horizon. m. athavale et al. / financial services review 30 (2022) 125–143 139 we also analyzed the determinants of portfolio success by partitioning the data based on allocation. this facilitates validation of the full-sample results and also allows for easier interpretation of the results. the results of this analysis are presented in table 6. as before, retiree-determined variables (withdrawal rate and retirement horizon) are significant across all three values of allocation, as are the chance variables (early experience, return, and risk). the negative sign on withdrawal rate indicates a lower probability of portfolio success at higher withdrawal rates. this effect is most pronounced for the conservative portfolio allocation—the probability of portfolio success changes by �0.1962 for a 1-level change in withdrawal rate. retirees who select a conservative portfolio allocation will find that the probability of portfolio success is more sensitive to the determinants of portfolio success, as compared with selecting a conservative or aggressive allocation. 6. concluding comments this analysis has examined the impact of various retiree-determined and chance variables on the probability of portfolio success. the results of this analysis provide retirees with specific information about the impact their actions may have on the probability of portfolio success. while the chance variables, by definition cannot be controlled, retirees may be able to mitigate the risk of portfolio failure by using intermediate targets like early experience to determine the need for midterm corrections to the retirement plan. the results of the analysis confirm that portfolio success is impacted by both retiree-determined variables (allocation, withdrawal rate, and planning horizon) and chance variables (the profile of portfolio returns includes return, risk, and early experience). however, the relative impact of each of these variables differs. withdrawal rate is the most significant driver of portfolio success, and though relevant, early experience is not as significant. retirees who select a conservative portfolio allocation will find that their portfolio success is much more sensitive to the explanatory variables, when compared with retirees who select other allocations. conservative portfolio allocations also result in wide variations in portfolio success outcomes, while aggressive portfolio allocations result in relatively narrow variations table 6 determinants of portfolio success (partitioned by portfolio allocation) aggressive balanced conservative coefficient marginal effect coefficient marginal effect coefficient marginal effect intercept 14.62 16.12 24.98 withdrawal rate �2.11 �0.0814 �2.84 �0.1137 �4.62 �0.1962 horizon �0.26 �0.0101 �0.32 �0.0126 �0.55 �0.0232 early experience 0.06 0.0023 0.07 0.0030 0.11 0.0045 return 0.93 0.0360 1.26 0.0506 2.18 0.0926 risk �0.31 �0.0120 �0.29 �0.0117 �0.38 �0.0163 r2 66% 68% 75% note. n= 10,000 for each of the three partitions. all coefficients are significant at the 1% level. this table presents the results of the logistic regressions (sample partitioned by portfolio allocation) analyzing the impact of the explanatory variables on the probability of portfolio success. 140 m. athavale et al. / financial services review 30 (2022) 125–143 across different withdrawal rates. these results are consistent with the ho, milevsky, and robinson (1994) assertion that equity should have a bigger role in retirement portfolios than is recommended by most financial planners. the actual return to a particular retiree may differ from that suggested by the portfolio allocation. both return and risk are significant in the analysis suggesting that asset selection within a portfolio is important to mitigating any adverse effect that these variables would have on portfolio success. the early experience variable could serve as an early indicator of the need for midterm course corrections to the retirement plan with the withdrawal rate serving as the transmission mechanism for portfolio success. planning for success in sustaining a retirement portfolio would be incomplete without also discussing other issues which arise even in the event of portfolio success. in the event of portfolio success, by definition, there is a residual (unconsumed) portfolio balance. this reduced consumption over the retirement horizon is the prudent reality of dealing with an uncertain future in the absence of well-accepted instruments that can capture the value of a potential future surplus. another issue is the possibility of the retiree living beyond the planned retirement horizon. while any unconsumed portfolio balance may serve to mitigate longevity risk, longevity insurance is also available in the form of a single-premium deferred inflation-indexed fixed life annuity, and this could allow real consumption at the same level as that experienced during the expected retirement years. and finally, many financial planners recommend that retirees maintain a cash bucket outside their invested portfolio as part of their overall strategy. this cash bucket is intended to meet unexpected consumption needs, reduce the need to liquidate portions of the invested portfolio during market downturns, and finance any mismatch in the timing of cash flows. when it comes to retirement planning, the cost of failure is high. nevertheless, most retirees lacking the knowledge and tools, engage in wishful thinking rather than structured planning. retirement planning is complex and has inherently uncertain outcomes. this analysis is a simple and imperfect representation of the complex realities of retirement planning. although the analyses are simplistic, the results provide guidance to the manner in which various retiree-determined and chance variables impact portfolio success. the research proposed in this study is a topic of active policy debate, and may serve as a baseline for additional research and sophisticated and dynamic models for generating lifetime income for retirees, pension funds, endowments, and managed payout mutual funds. this research also has various limitations. the simulation methodology requires us to specify the unknown future distribution from which returns might obtain, and the expected parameters of that distribution. we have assumed that the retiree makes withdrawals at the end of every year, and have not considered monthly or quarterly withdrawals. we have also implicitly assumed that the retire may engage in additional retirement planning outside the invested portfolio (e.g., bequests, longevity insurance, and a cash bucket for liquidity) and that retirees desire a constant level of real consumption over the retirement horizon. and finally, our analysis examines portfolio success as a binary (success/failure) variable but does not consider the size of the bequest or the timing of the portfolio failure as measures of the extent of success or failure (estrada and kritzman, 2019). there is no “one size fits all” single right answer when it comes to addressing the various tradeoffs and interactions associated with retirement planning, and outcomes will vary based m. athavale et al. / financial services review 30 (2022) 125–143 141 on retiree choices, economic circumstances, and chance. it is nevertheless important to understand them so that a realistic initial plan and consumption target can be established, and midterm adjustments can be initiated if necessary. just be aware that your mileage may vary. references athavale, m., & goebel, j. (2011). a safer safe withdrawal rate assuming various return distributions. journal of financial planning, 24, 36–43. benartzi, s., & thaler, r. (2007). heuristics and biases in retirement savings behavior. journal of economic perspectives, 21, 81–104. https://doi.org/10.1257/jep.21.3.81 bengen, w. (1994). determining withdrawal rates using historical data. journal of financial planning, 7, 171– 180. bierwirth, l. (1994). investing for retirement: using the past to model the future. journal of financial planning, 7, 14–24. biggs, a., & springstead, g. (2008). alternate measures of replacement rates for social security benefits and retirement income. social security bulletin, 68, blanchett, d., & blanchett, b. (2008). data dependence and sustainable real withdrawal rates. journal of financial planning, 21, 70–85. blanchett, d., finke, m., & pfau, w. (2014). low bond yields and safe portfolio withdrawal rates. the journal of wealth management, 16, 55–62. https://doi.org/10.3905/jwm.2013.16.2.055 blanchett, d., & frank, l. (2009). a dynamic and adaptive approach to distribution planning and monitoring. journal of financial planning, 22, 52–66. bonsang, e., adam, s., & perelman, s. (2012). does retirement affect cognitive functioning? journal of health economics, 31, 490–501. browning, c., guo, t., cheng, y., & finke, m. (2016). spending in retirement: determining the consumption gap. journal of financial planning, 29, 42–53. clare, a., seaton, j., smith, p., & thomas, s. (2017). reducing sequence risk using trend following and the cape ratio. financial analysts journal, 73, 91–103. https://doi.org/10.2469/faj.v73.n4.5 clare, a., seaton, j., smith, p., & thomas, s. (2021). can sustainable withdrawal rates be enhanced with trend following? international journal of finance & economics, 26, 27–41. https://doi.org/10.1002/ijfe.1774 collins, p., lam, h., & stampfli, j. (2015). how risky is your retirement income risk model? financial services review, 24, 193–216. cooley, p., hubbard, c., & walz, d. (1998). retirement savings: choosing a withdrawal rate that is sustainable. american association of individual investors journal, 20, 16–21. cooley, p., hubbard, c., & walz, d. (2003). a comparative analysis of retirement portfolio success rates: simulation versus overlapping periods. financial services review, 12, 115–128. cooley, p., hubbard, c., & walz, d. (2011). portfolio success rates: where to draw the line. journal of financial planning, 24, 48–60. dejong, j., & robinson, j. (2017). determinants of retirement portfolio sustainability and their relative impacts. journal of financial planning, 30, 54–62. dimson, e., marsh, p., & staunton, m. (2004). irrational optimism. financial analysts journal, 60, 15–25. https://doi.org/10.2469/faj.v60.n1.2589 estrada, j. (2018). maximum withdrawal rates: an empirical and global perspective. the journal of retirement, 5, 57–71. https://doi.org/10.3905/jor.2018.5.3.057 estrada, j. (2019). managing to target: dynamic adjustments for accumulation strategies. journal of financial planning, 32, 46–53. estrada, j. (2020). managing to target (ii): dynamic adjustments for retirement strategies. the journal of retirement, 7, 28–38. https://doi.org/10.3905/jor.2020.1.063 142 m. athavale et al. / financial services review 30 (2022) 125–143 estrada, j., & kritzman, m. (2019). toward determining the optimal investment strategy for retirement. the journal of retirement, 7, 35–42. https://doi.org/10.3905/jor.2019.1.052 fama, e. (1965). the behavior of stock market prices. the journal of business, 38, 34–105. https://doi.org/ 10.1086/294743 finke, m., pfau, w., & blanchett, d. (2013). the 4 percent rule is not safe in a low-yield world. journal of financial planning, 26, 46–55. gray, j., & french, d. (1990). empirical comparisons of distributional models for stock index returns. journal of business finance & accounting, 17, 451–459. https://doi.org/10.1111/j.1468-5957.1990.tb01197.x guyton, j. (2004). decision rules and portfolio management for retirees: is the “safe” initial withdrawal rate too safe? journal of financial planning, 17, 54–62. ho, k., milevsky, m., & robinson, c. (1994). asset allocation, life expectancy and shortfall. financial services review, 3, 109–126. https://doi.org/10.1016/1057-0810(94)90017-5 korniotis, g., & kumar, a. (2011). do older investors make better investment decisions? review of economics and statistics, 93, 244–265. https://doi.org/20.1162/rest_a_00053 kring, s., rachev, s., höchstötter, m., fabozzi, f., & bianchi, m. (2009). multi-tail generalized elliptical distributions for asset returns. econometrics journal, 12, 272–291. https://doi.org/0.1111/j.1368-423x.2009.00290.x levy, h., & duchin, r. (2004). asset return distributions and the investment horizon. the journal of portfolio management, 30, 47–62. https://doi.org/10.3905/jpm.2004.412319 liu, q., chang, r., de jong, j., & robinson, j. (2009). reality check: the implications of applying sustainable withdrawal rate analysis to real world portfolios. financial services review, 18, 123–139. merton, r. (2014). the crisis in retirement planning. harvard business review, 92, 42–50. milevsky, m., & robinson, c. (2005). a sustainable spending rate without simulation. financial analysts journal, 61, 89–100. https://doi.org/10.2469/faj.v61.n6.2776 officer, r. (1972). the distribution of stock returns. journal of the american statistical association, 67, 807– 812. https://doi.org/10.1080/01621459.1972.10481297 pfau, w. (2010). an international perspective on safe withdrawal rates from retirement savings: the demise of the 4% rule? journal of financial planning, 23, 51–61. pfau, w. (2012). capital market expectations, asset allocation, and safe withdrawal rates. journal of financial planning, 25, 36–43. pye, g. (2000). sustainable investment withdrawals. the journal of portfolio management, 26, 73–83. https:// doi.org/10.3905/jpm.2000.319765 robinson, c., & tahani, n. (2010). sustainable retirement income for the socialite, the gardener, and the uninsured. financial services review, 19, 187–202. sandidge, j. (2020). odds are retirees don’t care about the odds. retirement management journal, 9, 37–49. scott, j., sharpe, w., & watson, j. (2009). the 4% rule at what price? journal of investment management, 7, 31–48. spitzer, j., strieter, j., & singh, s. (2007). guidelines for withdrawal rates and portfolio safety during retirement. journal of financial planning, 20, 52–59. stout, r., & mitchell, j. (2006). dynamic retirement withdrawal planning. financial services review, 15, 117– 131. m. athavale et al. / financial services review 30 (2022) 125–143 143 active vs. passive, the case of sector equity funds yuhong fana, crystal yan linb,* aschool of accounting & taxation, weber state university, ogden, ut 84408, usa bschool of business, eastern illinois university, charleston, il 61920, usa abstract this paper examines performance of 95 actively managed u.s. sector equity mutual funds from 29 fund families relative to their peer exchange-traded funds, spdr sector etfs, in the period of 2008 to 2017. our results do not show considerable evidence that actively managed sector mutual funds outperform their passive counterparties. none of the mutual fund portfolios produces a significant positive alpha through factor models or delivers a significant positive alpha against their peer etfs. when focusing on the nine oldest actively managed fidelity sector mutual funds, outperformance in the period of 1999–2010, which is reported in literature, appears to fade away during the period of 2011-2017. alpha analyses of a larger sample of 60 sector mutual funds show similar performance deterioration in the same 19-year period. the results indicate that u.s. sector equity market has become more efficient in the past decade. © 2019 academy of financial services. all rights reserved. 1. introduction are investors giving up stock selecting? is the business of picking stocks dying (tergesen & zweig, 2016)? can actively managed mutual funds still be a good choice for investors or financial planners as evidenced by lin (2014)? the interest to respond to the movement from active to passive fund strategies leads to origination of this paper, which looks into the performance of actively managed u.s. sector equity mutual funds and their peer exchangetraded funds, sector spdr etfs, in the past decade. * correspondance author: tel.: 217-581-2227; fax: 217-581-6642. e-mail address: cylin@eiu.edu (c.y. lin) 1057-0810/20/$ – see front matter © 2019 academy of financial services. all rights reserved. financial services review 28 (2020) 159–177 the debate of active versus passive investing has lasted for decades. the indexing concept was introduced to the institutional pension plan market in 1971 and to the mutual fund industry in 1976 (bogle, 2015). at that time, the indexing strategy was questioned as why investors buy a basket of both good and bad stocks and even was described as “unamerican.” the core vision, which initiated the indexing strategy and later made it successful, stands on the low-cost goal and a no-load distribution network (bogle, 2016). thus, high management and transaction costs, disappointed returns, lack of simplicity and transparency attributed to actively managed funds are commonly stated reasons that make the passive strategy a common wisdom (brown, 2016; malkiel & radisich, 2001; tergesen & zweig, 2016). one of the earliest supporters of the index strategy is burton malkiel (malkiel, 1973). at that time, there were no index funds and burton said that index funds should be available to investors. from the empirical perspective, there are numerous studies that either provide evidence to support active fund managers’ stock picking skills (e.g., lin, 2014; wermers & moskowitz, 2000) or suggest that actively managed funds underperform their benchmarks or lack of managers’ stock picking persistence (e.g., fan & addams, 2012; malkiel, 1995). when focusing on u.s. sector investing, sector equity funds started as a mainstay of the mutual fund industry in the 1930s and 1940s but lost heat in the 1950s (bogle, 2015). in the modern time, sector equity funds returned to the investment world in 1981. they became an equity fund category in addition to u.s. equity funds and international equity funds, and have been widely used by portfolio managers and financial advisors for asset allocation purpose. today, many retirement saving plans offer a broad range of investment vehicles for individual investors. sector equity funds appear on the investment menu for millions of plan participants. because companies in a certain sector/industry are exposed to similar economic/political/technological factors, sector equity funds enable investors to capture certain market opportunities through sector selecting/rotation. diversification across industries is also easily achievable through investing in sector funds not individual stocks. the invention of etfs, the vast majority of which are index funds, has propelled the tremendous growth of indexing strategy. although both mutual funds and etfs are pooled investments that represent ownership in a basket of securities, etfs can be traded in exchange markets just like individual stocks. they are also shortable, marginable, optionable, and provide great transparency, such as investors can obtain an etf’s holding list more frequently than that of a mutual fund. the federal reserve started reporting etf accounts in 1993, the year that spdr etfs were launched. the total asset value in mutual funds and etfs were $1.5 trillion and $464 million at the end of 1993, respectively, and 47% of the mutual funds were equity funds while 100% of the etfs were equity funds. at the end of 2017, the total asset value in mutual funds and etfs were $15.9 trillion and $3.4 trillion, respectively. it shows that 68% of the mutual funds and 82% of the etfs were equity funds. for the most recent ten-year period, 2008-2017, the compound annual growth rates for mutual funds and etfs are 6.0% and 18.8%, respectively. according to morningstar, the total assets for all long-term active funds and passive funds, including both open-end mutual funds and etfs, were $11.4 trillion and $6.7 trillion at the end of 2017. for the equity component, active u.s. equity net flows have been negative while passive equity net flows have been positive every year since 2006. sector equity funds have experienced the same trend. at the end of 2017, there were $424 billion of active 160 y. fan, c.y. lin / financial services review 28 (2020) 159–177 sector equity funds and $488 billion of passive sector equity funds, representing 6.3% of active equity funds and 9.1% of passive equity funds, respectively. in 2017, active sector equity funds felt the pain of $18.4 billion of fund net outflow. passive sector equity funds, on the other hand, attracted $44.2 billion of fund net inflow. this is consistent with total equity fund flows: active equity funds had a net outflow of $190.3 billion while passive equity funds had a net inflow of $468.3 billion in 2017. how do actively managed sector equity funds perform relative to their passive counterparties in the recent decade? is indexing strategy in sector equity market a wise decision for investors? do active fund managers in the sector investing category have a niche compared with the broad market-based equity investing? to answer these questions, this paper looks into performance of united states actively managed sector equity funds for the past 10-year period: 2008-2017. our study adds value to the small pool of literature on the topic of sector investing. cremers, fulkerson, and riley (2019) mention that research on active management often excludes sector funds, which means knowledge about them is limited. dellva, demaskey, and smith (2001) test fidelity sector equity funds against the broad stock market indexes, which are not investable, for the 1989-1998 time period. using nine fidelity select funds, lin (2014) shows that actively managed sector equity mutual funds provide better afterexpense returns against the broad market etfspy, which tracks the s&p 500, and their peer sector etfs. however, both studies only focus on one fund family: fidelity. other studies such as chen et al. (2018) and kaushik et al. (2014) focus fund performance on one sector, healthcare, and provide evidence that actively managed funds perform better than their peer etfs. unlike these studies, this paper expands fund scope to multi-fund families and includes not only sector funds but also more narrowly defined industry funds. kaushik, pennathur, and barnhart (2010) report sector aggregate performance results on seven sectors for the period of 1990-2005: energy, financials, healthcare, precious metals, technology, real estate, and utilities. compared with their work, our nine-sector coverage is more comprehensive. sectors that are represented in this study are: materials, energy, financials, industrials, technology, consumer staples, utilities, healthcare, and consumer discretionary. 2. sample selection our sample includes all active sector equity mutual funds that have price data available for the 10-year period 2008 through 2017 in morningstar research center database. we eliminate funds with assets under management below $100 million. for funds with the same portfolio but multiple classes, class a fund is selected to represent the portfolio. the process has generated a sample of 95 actively managed sector funds from 29 fund families. table 1 shows the distribution of the sample among mutual fund families and sectors covered. fidelity stands out in the fund family list. its funds account for 50 out of 95 sector equity funds of the sample and represent all nine sectors. fidelity has a long history of offering sector equity funds. it launched its first sector equity mutual funds on energy, healthcare, and technology in 1981. besides fidelity, franklin templeton investments has four funds in four sectors. pgim, putnam, and rydex funds each has three funds in three sectors. y. fan, c.y. lin / financial services review 28 (2020) 159–177 161 blackrock, deutsche, invesco, ivy funds, mfs, vanguard, and victory each has two funds in two sectors. columbia has two funds in one sector. each of the other sixteen fund families only has one fund in the sample. consistent with lin (2014), we use select sector spdr etfs as passive counterparties for sector equity mutual funds since they are the largest sector etf family and have the longest trading history (lin, 2014). the spdr sector etfs track the select sector indexes. each stock in the s&p 500 is allocated to one and only one select sector index. the combined companies of the select sector indexes represent all of the companies in the s&p 500. sector spdrs were launched in december 1998 with nine etfs: xlbthe materials sector; xlethe energy sector; xlfthe financial sector; xlithe industrial sector; xlkthe technology sector; xlpthe consumer staples sector; xluthe utilities sector; xlvthe healthcare sector; and xly -the consumer discretionary sector. today there are 11 sector spdr etfs. the 10th sector etf, xlrethe real estate select sector spdr, began trading in october 2015. we do not include this etf because it only covers 26months out of the 10-year sample period. the 11th etf, xlcthe table 1 sector mutual fund sample fund family number of funds in study number of sectors represented in study 1919 funds 1 1 allianz funds 1 1 berkshire 1 1 blackrock 2 2 columbia 2 1 delaware funds 1 1 deutsche 2 2 dreyfus 1 1 eaton vance 1 1 emerald 1 1 fidelity 50 9 firsthand funds 1 1 franklin templeton investments 4 4 gabelli 1 1 goldman sachs 1 1 invesco 2 2 ivy funds 2 2 john hancock 1 1 mfs 2 2 pgim 3 3 putnam 3 3 rydex funds 3 3 schwab funds 1 1 usaa 1 1 vaneck 1 1 vanguard 2 2 victory 2 2 wells fargo funds 1 1 williston 1 1 total number of fund families 29 total number of sector mutual funds 95 162 y. fan, c.y. lin / financial services review 28 (2020) 159–177 communications services select sector spdr, is also excluded because it was launched in june 2018, which is beyond our sample period. passively managed sector mutual funds are not utilized as benchmarks since they are not common. only three such funds representing two sectors have price data throughout our sample period, which are not adequate to be included to conduct a comprehensive sector fund research. 3. fund performance we compare actively managed sector mutual fund’s return to that of its passive alternative and not to the sector index itself since indexes are not investable. for the same reason, we use spy but not the actual s&p 500 index as an investable broad u.s. equity market benchmark. the monthly returns for both mutual funds and etfs are calculated as: r i, t = pi, t/pi, t-1 – 1, where pi, t is the adjusted closing price for fund i at the end of month t, and pi, t-1 is the adjusted closing price for fund i at the end of month t-1. all returns are net of expenses, commissions, and sales loads. the distribution of sample funds by sectors in table 2 shows that technology has the largest number of funds, 24. it has a representation of 15 fund families, the highest among all sectors. there are 18, 13, and 11 funds in healthcare, financials, and materials, respectively. consumer staples only has two funds in the sample. fidelity is the only fund family that covers industrials and consumer discretionary sectors. table 2 also provides a quick comparison of fund annual returns. panel a shows the top three performing sectors are technology, consumer discretionary, and healthcare with an average annual return of 15.4%, 15.4%, and 13.7%, respectively. energy sector, on the other hand, has the lowest average annual return of 2.4%. however, the sector etfs’ annual returns presented in panel b tell a somewhat different story. consumer discretionary leads with an average annual return of 15.5%. the second best is technology with an average annual return of 13.8%. healthcare and industrials are ranked the third and fourth, followed by consumer staples. consistent with panel a, energy etf generates the lowest average annual return during the sample period, 3.6%, which is higher than that of the mutual funds. concurrently, the average annual return of spy is 10.3%. compare panel a to panel b, mutual funds outperform etfs in four sectors: financials, technology, healthcare, and industrials. the annual return differences are 1.8%, 1.6%, 1.5%, and 1.4%, respectively. they largely underperform their etf counterparty in the materials sector with an annual spread of -5.0%. however, t test results show that none of the differences between mutual fund and etf mean returns is significant at the 10% level. mutual funds have a tendency of wider return ranges compared with etfs. all of the maximal returns of the mutual funds are higher than those of etfs. all of the minimal returns of the mutual funds are lower than those of the etfs except for financials. mutual funds that are actively managed are more likely to have more risk and, thus, have a wider range in returns. sector mutual funds also have a higher standard deviation in eight out of the nine sectors, which also shows that actively managed funds tend to be riskier than their counterparties. fig. 1 provides an annual return comparison of equal-weighted sector mutual fund portfolios, their peer etfs, and spy by sector/year. y. fan, c.y. lin / financial services review 28 (2020) 159–177 163 table 3 reports 10-year holding period returns for equal-weighted sector mutual fund portfolios, the nine sector etfs, and the spy. there are four mutual fund sectors outperform their etf counterparties: financials, industrials, technology, and healthcare. the return spreads over the 10-year period are 62.2%, 33.9%, 12.6%, and 34.7%, respectively. the four outperforming sectors also have the highest percentage of individual mutual funds that beat their sector etfs: 92.3%, 80.0%, 54.2%, and 50.0%, respectively. the other five sector mutual fund portfolios, however, lag. the worst performer is the energy mutual fund portfolio, which generates a loss of 16.5% compared with a gain of 12.8% earned by the energy etf. only one out of the nine energy mutual funds outperform. none of the two consumer staples funds outperforms. the average return of the sector mutual fund portfolios, 124.3%, is lower than the average etf return, 125.6%; a p-value of 0.928 shows the average mutual fund portfolio return is not statistically different from the average etf return. when using spy as the benchmark, both sector mutual fund portfolios and etfs outperform in the industrials, technology, consumer staples, healthcare, and consumer discretionary sector. the p-values show that neither the average mutual fund portfolio holding period return nor the table 2 fund annual return statistics, 2008-2017 panel a: mutual fund statistics sector no. mf no. mf family annual return mean sd max. min. materials 11 8 4.4% 30.0% 78.7% �61.4% energy 9 4 2.4% 29.5% 77.1% �63.2% financials 13 6 9.7% 22.7% 83.6% �49.9% industrials 5 1 12.9% 22.6% 50.7% �40.2% technology 24 15 15.4% 29.2% 90.3% �57.3% consumer staples 2 2 9.6% 12.9% 28.6% �23.3% utilities 9 6 6.7% 16.1% 32.9% �47.0% healthcare 18 10 13.7% 19.8% 68.6% �44.6% consumer discretionary 4 1 15.4% 22.7% 57.8% �39.3% total 95 29 panel b: etf statistics sector etf annual return t test return difference between mutual funds and etfs (p-value) mean sd max. min. materials-xlb 9.4% 25.4% 48.2% �44.1% 0.189 energy-xle 3.6% 22.0% 28.0% �38.8% 0.737 financials-xlf 7.9% 26.8% 35.5% �55.3% 0.484 industrialsxli 11.5% 22.1% 40.6% �38.9% 0.153 technology-xlk 13.8% 24.2% 51.3% �41.4% 0.541 consumer staples-xlp 10.5% 10.6% 26.3% �15.0% 0.480 utilities-xlu 7.4% 15.9% 28.7% �29.1% 0.773 healthcare-xlv 12.2% 17.6% 41.4% �23.2% 0.303 consumer discretionary-xly 15.5% 21.8% 42.7% �33.5% 0.896 spy 10.3% 19.2% 32.3% �37.0% note: spy is the spdr s&p 500 etf. 164 y. fan, c.y. lin / financial services review 28 (2020) 159–177 fig. 1. annual returns for the mutual fund portfolios, etfs, and spy 2008–2017. y. fan, c.y. lin / financial services review 28 (2020) 159–177 165 t ab le 3 1 0 -y ea r h o ld in g p er io d re tu rn (h p r ) co m p ar is o n s, 2 0 0 8 – 2 0 1 7 s ec to r n o . m f m f p o rt fo li o h p r s e t f h p r s m f b ea ts e f t n o . o f m f w it h a h ig h er h p r th an e t f m f b ea ts s p y e t f b ea ts s p y m at er ia ls 1 1 1 .0 % 8 2 .3 % n o 2 n o n o e n er g y 9 �1 6 .5 % 1 2 .8 % n o 1 n o n o f in an ci al s 1 3 1 0 5 .7 % 4 3 .5 % y es 1 2 n o n o in d u st ri al s 5 1 7 3 .1 % 1 3 9 .2 % y es 4 y es y es t ec h n o lo g y 2 4 1 9 5 .7 % 1 8 3 .1 % y es 1 3 y es y es c o n su m er st ap le s 2 1 3 3 .1 % 1 5 8 .8 % n o 0 y es y es u ti li ti es 9 6 9 .2 % 8 1 .4 % n o 1 n o n o h ea lt h ca re 1 8 2 1 4 .9 % 1 8 0 .2 % y es 9 y es y es c o n su m er d is cr et io n ar y 4 2 4 2 .2 % 2 4 9 .1 % n o 1 y es y es a v er ag e m f h p r 1 2 4 .3 % a v er ag e e t f h p r 1 2 5 .6 % s p y h p r 1 2 4 .4 % p -v al u e th e av er ag e m f h p r d if fe re n t fr o m th e av er ag e e t f h p r 0 .9 2 8 p -v al u e th e av er ag e m f h p r d if fe re n t fr o m s p y h p r 0 .9 9 7 p -v al u e th e av er ag e e t f h p r d if fe re n t fr o m s p y h p r 0 .9 6 3 n o te : s p y is th e s p d r s & p 5 0 0 e t f . 166 y. fan, c.y. lin / financial services review 28 (2020) 159–177 average etf holding period return is statistically different from the spy holding period return, which is 124.4% during the sample period. results from tables 2 and 3 indicate that in some sectors, especially in industrials, technology, and healthcare, sector mutual funds generate superior raw returns during the sample period compared with both the sector etfs and the spy. does the mutual fund’s risk adjusted performance differ from that of their peer etf? table 4 provides comparisons on both sharpe ratios and information ratios. during the sample period, financials, industrials, and healthcare mutual fund portfolios generate a higher sharpe ratio than their etf counterparties, with 92.3%, 60.0%, and 50.0% of individual mutual funds outperform, respectively. panel b of table 4 shows that in four sectors both mutual fund portfolio and etfs report a positive information ratio against spy: industrials, technology, healthcare, and consumer discretionary. among the four sectors, industrials and healthcare have a higher information ratio against spy than the etf counterparts, while technology and consumer discretionary present a lower information ratio. it is noticeable that all the ratios are less than 1, with the highest 0.771 from the consumer discretionary etf against spy. when peer etf is used as the benchmark, mutual fund portfolios in financials, industrials, technology, and healthcare sectors generate positive information ratios. however, the ratios are all lower than 0.5. one mutual fund portfolio, consumer discretionary, which generates a positive information ratio against spy, fails to do so when compared with its peer etf. in addition to risk-adjusted return measures such as sharpe ratio and information ratio, we also run regressions based on factor models. one-factor, three-factor, and four-factor models are utilized with commonly used market, small/large size, value/growth, and momentum as regression factors. according to bogle (2015), one of the situations when passive investing could go wrong is when “subsets of the equity market provide different results from the market as a whole.” simply to say, this is a benchmark selection problem. we address this issue by investigating sector equity mutual fund’s performance with two different sets of benchmarks: broad market index etf, spy, and sector equity etfs. that is, either spy or a sector etf is used as the market portfolio in the factor models. table 5 presents results from three different factor models with spy as the market portfolio. alpha estimate, p-value for alpha estimate, and adjusted r2 are reported for both mutual fund portfolios and etfs. the number of funds with a positive alpha in each sector is also reported in table 5. results from panel a, the capm model, show that none of the sector mutual fund portfolios generates a significant positive alpha, while two sector etfs, consumer staples and healthcare, do generate significant positive alphas at the 10% level. only four out of 95 individual mutual funds generate significant positive alphas: one in technology, two in healthcare, and one in consumer discretionary. most etf regressions have a higher explanation power, adjusted r2, than that of mutual funds, except for financials and utilities. similar results are presented in panel b, the three-factor model, and panel c, the four-factor model. none of the sector mutual fund portfolios generates a significant positive alpha at the 10% level. materials sector mutual fund portfolio generates a significant negative alpha. on the other hand, three sector etfs, consumer staples, healthcare, and consumer discretionary, generate significant positive alphas. the same four individual mutual funds as mentioned in the previous paragraph, produce a significant positive alpha. most etf regressions have a higher adjusted r2 than that of mutual funds, except for financials and utilities. y. fan, c.y. lin / financial services review 28 (2020) 159–177 167 t ab le 4 s h ar p e ra ti o s an d in fo rm at io n ra ti o s, 2 0 0 8 -2 0 1 7 p an el a : s h ar p e ra ti o s s ec to r m f p o rt fo li o e t f n o . m f n o . m f w it h a h ig h er ra ti o e x ce ss re tu rn s d s h ar p e ra ti o * e x ce ss re tu rn s d s h ar p e ra ti o m at er ia ls 4 .1 % 3 0 .3 % 0 .1 3 5 9 .1 % 2 5 .7 % 0 .3 5 4 1 1 2 e n er g y 2 .1 % 2 9 .9 % 0 .0 7 1 3 .3 % 2 2 .4 % 0 .1 4 8 9 1 f in an ci al s 9 .4 % 2 1 .7 % 0 .5 0 8 7 .6 % 2 7 .2 % 0 .2 8 0 1 3 1 2 in d u st ri al s 1 2 .6 % 2 2 .6 % 0 .5 5 7 1 1 .2 % 2 2 .5 % 0 .5 0 0 5 3 t ec h n o lo g y 1 5 .2 % 2 9 .8 % 0 .5 0 8 1 3 .5 % 2 4 .5 % 0 .5 5 2 2 4 3 c o n su m er st ap le s 9 .3 % 1 3 .5 % 0 .6 9 4 1 0 .2 % 1 1 .0 % 0 .9 2 4 2 0 u ti li ti es 6 .4 % 1 6 .3 % 0 .3 9 3 7 .1 % 1 6 .3 % 0 .4 3 4 9 2 h ea lt h ca re 1 3 .4 % 1 9 .6 % 0 .6 8 4 1 1 .9 % 1 7 .9 % 0 .6 6 5 1 8 9 c o n su m er d is cr et io n ar y 1 5 .1 % 2 3 .0 % 0 .6 5 6 1 5 .2 % 2 2 .2 % 0 .6 8 6 4 1 p an el b : in fo rm at io n ra ti o s s ec to r m f p o rt fo li o v er su s s p y e t f v er su s s p y m f p o rt fo li o v er su s e t f t ra ck in g er ro r s d in fo rm at io n ra ti o t ra ck in g er ro r s d in fo rm at io n ra ti o t ra ck in g er ro r s d in fo rm at io n ra ti o m at er ia ls �5 .9 % 1 9 .0 % �0 .3 1 2 �0 .9 % 1 0 .1 % �0 .0 9 2 �5 .0 % 1 1 .1 % �0 .4 5 0 e n er g y �7 .9 % 1 9 .5 % �0 .4 0 5 �6 .7 % 1 3 .1 % �0 .5 1 1 �1 .2 % 1 0 .9 % �0 .1 1 0 f in an ci al s �0 .6 % 7 .2 % �0 .0 8 4 �2 .4 % 1 0 .6 % �0 .2 2 7 1 .8 % 7 .8 % 0 .2 3 1 in d u st ri al s 2 .6 % 7 .0 % 0 .3 6 6 1 .2 % 6 .3 % 0 .1 9 5 1 .3 % 2 .7 % 0 .4 9 4 t ec h n o lo g y 5 .1 % 1 4 .2 % 0 .3 6 3 3 .5 % 9 .3 % 0 .3 7 7 1 .6 % 8 .2 % 0 .2 0 1 c o n su m er st ap le s �0 .7 % 7 .9 % �0 .0 8 5 0 .2 % 1 0 .6 % �0 .0 1 6 �0 .8 % 3 .6 % �0 .2 3 3 u ti li ti es �3 .6 % 7 .5 % �0 .4 7 8 �3 .0 % 1 3 .2 % �0 .2 2 3 �0 .6 % 6 .8 % �0 .0 9 4 h ea lt h ca re 3 .4 % 1 1 .0 % 0 .3 1 1 1 .9 % 1 0 .1 % 0 .1 8 5 1 .5 % 4 .5 % 0 .3 4 5 c o n su m e d is cr et io n ar y 5 .1 % 7 .7 % 0 .6 6 1 5 .2 % 6 .7 % 0 .7 7 1 �0 .1 % 2 .3 % �0 .0 4 3 n o te : s p y is th e s p d r s & p 5 0 0 e t f . s h ar p e ra ti o = p o rt fo li o ’s ex ce ss re tu rn /s d o f th e ex ce ss re tu rn . in fo rm at io n ra ti o = tr ac k in g er ro r/ s d o f th e tr ac k in g er ro r. t ra ck in g er ro r is th e re tu rn d if fe re n ce b et w ee n th e tw o p ai rs . 168 y. fan, c.y. lin / financial services review 28 (2020) 159–177 table 5 alphas from factor models with spy as the market portfolio, 2008–2017 panel a: one-factor model sector mf portfolio etf a estimate p-value adjusted r2 no. funds with significant positive a/no. of total funds a estimate p-value adjusted r2 materials �0.69% 0.101 0.57 0/11 �0.35% 0.170 0.81 energy �0.79% 0.128 0.47 0/9 �0.49% 0.227 0.54 financials �0.13% 0.569 0.79 0/13 �0.18% 0.782 0.42 industrials 0.13% 0.581 0.78 0/5 �0.06% 0.737 0.88 technology 0.31% 0.284 0.72 1/24 0.28% 0.136 0.83 consumer staples 0.27% 0.259 0.57 0/2 0.39% 0.057 0.59 utilities 0.05% 0.854 0.47 0/9 0.22% 0.500 0.22 healthcare 0.53% 0.148 0.43 2/18 0.39% 0.085 0.63 consumer discretionary 0.35% 0.121 0.79 1/4 0.30% 0.112 0.86 panel b: three-factor model sector mf portfolio etf a estimate p-value adjusted r2 no. funds with significant positive a/no. of total funds a estimate p-value adjusted r2 materials �0.71% 0.099 0.57 0/11 �0.35% 0.168 0.81 energy �0.78% 0.135 0.47 0/9 �0.46% 0.249 0.54 financials �0.13% 0.565 0.79 0/13 �0.32% 0.618 0.44 industrials 0.22% 0.337 0.80 0/5 �0.01% 0.942 0.89 technology 0.32% 0.265 0.72 1/24 0.27% 0.142 0.83 consumer staples 0.24% 0.319 0.57 0/2 0.34% 0.087 0.60 utilities 0.06% 0.827 0.46 0/9 0.23% 0.498 0.20 healthcare 0.56% 0.134 0.43 2/18 0.38% 0.097 0.63 consumer discretionary 0.36% 0.319 0.79 1/4 0.31% 0.098 0.86 panel c: four-factor model sector mf portfolio etf a estimate p-value adjusted r2 no. funds with significant positive a/no. of total funds a estimate p-value adjusted r2 materials �0.72% 0.086 0.59 0/11 �0.35% 0.170 0.81 energy �0.79% 0.120 0.50 0/9 �0.47% 0.236 0.55 financials �0.13% 0.564 0.79 0/13 �0.31% 0.624 0.44 industrials 0.22% 0.329 0.80 0/5 �0.01% 0.959 0.89 technology 0.32% 0.269 0.72 1/24 0.27% 0.145 0.83 consumer staples 0.24% 0.325 0.57 0/2 0.34% 0.088 0.60 utilities 0.06% 0.835 0.46 0/9 0.23% 0.500 0.20 healthcare 0.56% 0.136 0.42 2/18 0.38% 0.097 0.62 consumer discretionary 0.36% 0.107 0.79 1/4 0.31% 0.093 0.86 note: spy is the spdr s&p 500 etf. number of observations120 monthly returns. model used in panel a: ri,t � rf ,t ¼ ai þ b i rspy,t � rf ,tð þ þ « i,t.where ri,t is the return of fund/mutual fund portfolio i in month t, rf ,t is the return of one-month t-bill in month t, rspy,t is the spy return in month t, and « i,t is an error term. model used in panel b: ri,t � rf ,t ¼ ai þ b i rspy,t � rf ,tð þ þ hihmlt þ sismbt þ « i,t.where hml (high minus low) is the average return on two value portfolios minus the average return on two growth portfolios and smb (small minus big) is the average return on three small portfolios minus the average return on three big portfolios (fama & french 1993). the data is downloaded from kenneth r. french data library. model used in panel c: ri,t � rf ,t ¼ ai þ b i rspy,t � rf ,tð þ þ hihmlt þ sismbt þ mimomt þ « i,t.where mom (momentum) is the average return on the two high prior return portfolios minus the average return on the two low prior return portfolios (see carhart 1997). the data is downloaded from kenneth r. french data library. y. fan, c.y. lin / financial services review 28 (2020) 159–177 169 the r2 changes only marginally when adding more factors to the capm for both mutual fund and etf regressions. these results reject the hypothesis that sector mutual fund portfolios outperform market portfolio, spy, during the sample period. we tested robustness of our results with two equal subsample periods.1 none of the sector mutual fund portfolios generates a significant positive alpha for any of the three models in either subsample period. consumer staples and consumer discretionary etfs, however, do generate significant positive alphas in the first subsample period, 2008-2012, across three models. none of the etfs generates a significant positive alpha in the second subsample period, 2013-2017. for the one-factor model, five individual mutual funds in the first half subsample period and two mutual funds in the second half subsample period generate significant positive alphas, respectively. using spy as the market portfolio may be questionable when our goal is to investigate the performance of sector equity funds. the risk exposure of a sector fund should be much narrower than that of the spy. according to state street global advisors, each of the sectors only represents 2.6% to 26.0% of the spy companies. thus, we modify the capm by using peer sector etfs as market portfolios instead. results in table 6 are consistent with results from table 5. none of the sector mutual fund portfolios generates a significant positive alpha at the 10% level. only two individual funds outperform their peer etfs by delivering a significant positive alpha: one in materials and one in financials. most of the adjusted r2 are higher compared with results in table 5 panel a. results from factor models do not show significant evidence to advocate the active fund management strategy from the standpoint of sector equity funds. when measured by alpha, not a single sector mutual fund portfolio outperforms either spy or its peer sector etf. only two or three individual sector mutual funds, out of the sample of 95, beat their peer etfs or the broad market spy for the period of 2008-2017. 4. time trend analysis of alpha lin (2014) presents evidence that nine actively managed fidelity sector mutual funds provide higher risk adjusted returns compared with their peer etfs during the period of 19992010. however, results in this study do not show much evidence that actively managed sector funds outperform their passive counterparties. these inconsistent findings demand further research. are results in lin (2014) still valid after 2010? we address this question by analyzing sector mutual fund performance in different time periods. one of the reasons that lin (2014) focuses on the fidelity family is that fidelity was the largest mutual fund family in the united states for several decades until vanguard moved to the top in 2010. unlike vanguard who launched the first index fund, fidelity has a reputation for actively managed mutual funds. the large size of fidelity’s assets under management enables it to provide economy of scale for security research, which is critical to the key element of active management: security selection. active managers can beat their benchmark by overweighting future winners, underweighting future losers, or a combination of both besides the freedom of holding cash. focusing on only one sector, a sector mutual fund manager can potentially grow knowledge and experience in that sector and gradually achieve a 170 y. fan, c.y. lin / financial services review 28 (2020) 159–177 t ab le 6 a lp h as fr o m o n efa ct o r m o d el w it h p ee r e t f s as th e m ar k et p o rt fo li o , 2 0 0 8 -2 0 1 7 s ec to rs n o . o f fu n d s a e st im at e p -v al u e o f a b e st im at e p -v al u e o f b a d ju st ed r 2 n o . fu n d s w it h si g n ifi ca n t p o si ti v e a n o . fu n d s w it h si g n ifi ca n t n eg at iv e a m at er ia ls 1 1 �0 .3 9 % 0 .2 2 6 0 .9 6 0 .0 0 0 .9 1 1 1 e n er g y 9 �0 .2 4 % 0 .2 3 6 1 .1 6 0 .0 0 0 .8 3 0 1 f in an ci al s 1 3 0 .3 8 % 0 .2 7 3 0 .4 1 0 .0 0 0 .5 0 1 0 in d u st ri al s 5 0 .2 0 % 0 .3 0 0 0 .9 0 0 .0 0 0 .8 6 0 0 t ec h n o lo g y 2 4 0 .0 3 % 0 .8 8 9 1 .0 7 0 .0 0 0 .7 1 0 1 c o n su m er st ap le s 2 �0 .0 5 % 0 .7 8 6 1 .0 1 0 .0 0 0 .7 5 0 0 u ti li ti es 9 0 .0 8 % 0 .6 7 9 0 .8 2 0 .0 0 0 .7 3 0 1 h ea lt h ca re 1 8 0 .1 1 % 0 .6 8 9 1 .0 7 0 .0 0 0 .6 8 0 0 c o n su m er d is cr et io n ar y 4 0 .0 7 % 0 .6 1 7 0 .9 4 0 .0 0 0 .9 1 0 0 n o te : n u m b er o f o b se rv at io n s1 2 0 m o n th ly re tu rn s. m o d el u se d : r m f j,t � r f, t ¼ a j þ b j r e t f j,t � r f, t ð þþ « j,t .w h er e r m f j,t is th e re tu rn o f in d iv id u al m u tu al fu n d /m u tu al fu n d p o rt fo li o o f se ct o r j in m o n th t, r f, t is th e re tu rn o f o n em o n th t -b il l in m o n th t, an d r e t f j,t is th e re tu rn o f e t f in th e sa m e se ct o r j in m o n th t, an d « j,t is an er ro r te rm . y. fan, c.y. lin / financial services review 28 (2020) 159–177 171 superior fund performance. moreover, fidelity is the only fund family whose actively managed sector funds cover the nine examined sectors in both lin (2014) and our studies. for instance, in our sample, the second largest sector coverage is franklin templeton investments, whose funds represent only four sectors. the nine fidelity sector funds are examined in the same way as lin (2014) does and the study is continued till the year 2017. lin chooses the year 1999 as the starting sample period since these nine fidelity funds had the first full year price data in 1999. the same factor models are used in this paper to match lin’s study but with a longer sample period of 19 years. we also break the nine funds’ performance into two subsample periods: 1999 to 2010 and 2011 to 2017. the 1999 to 2010 subsample period is the full sample period of lin (2014). results from the one-factor model with peer etfs as market portfolios are presented in table 7. over the whole sample period, 1999-2017, only one fidelity sector mutual fund generates a significant positive alpha at the 10% level against peer etfs: the fidelity select materials portfolio (fsdpx). when focusing on the two subsample periods, results from panel b and c show that fidelity sector mutual funds perform better against their peer etfs in the 1999-2010 period than in the 2011-2017 period. in the first sub sample period, 19992007, four fidelity funds outperform their peer etfs: materials (fsdpx), industrials (fcyix), technology (fsptx), and consumer staples (fdfax). fund performance in the first subsample period 1999-2010 is consistent with lin (2014), which reports the same four fidelity funds outperform peer etfs using the one-factor model. interestingly, none of the funds outperforms in the second subsample period 2011-2017 measured by alpha. moreover, seven out of nine funds produce a negative alpha given that they are not statistically significant. why does performance of the nine fidelity sector mutual funds deteriorate in the second subsample period? one explanation could be that the u.s. stock market has become more efficient because of increased level of competition over time. another explanation could be attributed to correlation convergence between sector mutual funds and their etf counterparties over time. this is the “closet indexing” argument stated by petajisto (2013). closet indexing is a strategy used to describe funds that claim to be actively managed investments but wind up with portfolios not much different from their benchmarks. these portfolio managers achieve a return similar to an underlying benchmark without exactly replicating it. to investigate this time-varying dynamic of sector mutual fund performance, we calculate 36month rolling correlations for the nine fidelity funds with their peer etfs.2 there is no consistent evidence of correlation convergence over the 19-year period. for most sectors, the 36-month rolling corrections are relatively high in the period of 2008 to 2013. the deep bear market of 2007-2009 could contribute to the high correlation as in falling markets stocks tend to move together. the 2000-2002 bear market also come with a relatively high correlation for several sectors. the deteriorated performance of the nine fidelity sector mutual funds could not be attributed to the correlation convergence explanation since correlation convergence is not consistently presented. results in table 7 show that the nine fidelity funds’ outperformances from the earlier time period of 1999 to 2010 have disappeared in the later time period of 2010 to 2017. does performance from the whole group, the active sector mutual funds, exhibit the same down 172 y. fan, c.y. lin / financial services review 28 (2020) 159–177 trend as the nine fidelity funds do? we expand the above time trend study sample to all sector mutual funds from a variety of fund families to see whether the deterioration of alpha is unique to fidelity funds. sixty sector mutual funds from the sample of 95 are included in table 7 alphas of nine fidelity funds with peer etfs as market portfolios panel a: period 1999–2017 sector fund a estimate p-value a b estimate p-value b adjusted r2 materials fsdpx 0.28% 0.073 0.95 0.00 0.86 energy fsenx 0.07% 0.701 1.07 0.00 0.87 financials fidsx 0.05% 0.849 0.60 0.00 0.59 industrials fcyix 0.21% 0.151 0.99 0.00 0.84 technology fsptx 0.37% 0.116 1.17 0.00 0.83 consumer staples fdfax 0.23% 0.135 0.85 0.00 0.63 utilities fsutx 0.05% 0.849 0.72 0.00 0.40 healthcare fsphx 0.32% 0.242 0.78 0.00 0.37 consumer discretionary fscpx �0.01% 0.955 0.81 0.00 0.70 panel b: sub-period 1999–2010 sector fund a estimate p-value a b estimate p-value b adjusted r2 materials fsdpx 0.45% 0.041 0.94 0.00 0.86 energy fsenx 0.18% 0.416 1.07 0.00 0.88 financials fidsx 0.09% 0.647 0.85 0.00 0.85 industrials fcyix 0.38% 0.046 0.98 0.00 0.87 technology fsptx 0.70% 0.021 1.19 0.00 0.87 consumer dtaples fdfax 0.34% 0.094 0.83 0.00 0.62 utilities fsutx �0.08% 0.834 0.68 0.00 0.34 healthcare fsphx 0.22% 0.500 0.62 0.00 0.31 consumer discretionary fscpx �0.07% 0.770 0.77 0.00 0.70 panel c: sub-period 2011–2017 sector fund a estimate p-value a b estimate p-value b adjusted r2 materials fsdpx �0.02% 0.918 0.97 0.00 0.88 energy fsenx �0.12% 0.655 1.06 0.00 0.84 financials fidsx 0.54% 0.225 0.28 0.00 0.25 industrials fcyix �0.10% 0.672 1.01 0.00 0.77 technology fsptx �0.09% 0.830 1.09 0.00 0.58 consumer staples fdfax �0.05% 0.834 0.94 0.00 0.63 utilities fsutx 0.20% 0.415 0.80 0.00 0.64 healthcare fsphx �0.12% 0.802 1.23 0.00 0.52 consumer discretionary fscpx �0.17% 0.551 1.00 0.00 0.72 note: number of observations: panel a-228 monthly returns; panel b-144 monthly returns; panel c-84 monthly returns. the fidelity funds are: select consumer discretionary portfolio (fscpx), select consumer staples portfolio (fdfax), select energy portfolio (fsenx), select financial services portfolio (fidsx), select health care portfolio (fsphx), select industrial portfolio (fcyix), select materials portfolio (fsdpx), select technology portfolio (fsptx), and select utilities portfolio (fsutx.). model used with etf: rmfj,t � rf,t ¼ aj þ b j retfj,t � rf ,tð þ þ « j,t.where rmfj,tis the return of individual mutual fund/mutual fund portfolio of sector j in month t, rf ,tis the return of one-month t-bill in month t, and retfj,tis the return of etf in the same sector j in month t, and « j,t is an error term. y. fan, c.y. lin / financial services review 28 (2020) 159–177 173 this analysis. funds incepted after 1999 and the nine fidelity funds studied before are excluded. table 8 presents results of one factor model of equal-weighted sector mutual fund portfolios against their peer etfs in the same three time periods as in table 7, 1999-2017, 19992010, and 2011-2017. the consumer staples sector is eliminated in table 8 because the only two funds in the sector are either incepted after 1999 or belong to the nine fidelity fund group. in panel a, industrials, technology, and healthcare sector fund portfolios produce a significantly positive alpha for the period of 1999 to 2017. a total number of eleven individual funds from financials, industrials, technology, and healthcare have a significant positive alpha at the 10% level. when breaking the whole sample period into two subperiods, divergence of results appears. for the earlier subperiod 1999-2010, mutual fund portfolios in four sectors, materials, financials, technology, and healthcare outperform their peer etfs with a significant positive alpha at the 10% level. the total count of individual funds with a significant positive alpha is 18. only two sectors, utilities and consumer discretionary, do not have funds that outperform their peer etfs. when moving to the second subperiod of 2011-2017, however, only financials portfolio outperforms with a significant positive alpha. in contrast, the energy sector portfolio produces a significant negative alpha. six individual funds from financials show a significant positive alpha, whereas six funds from materials and two funds from energy sector produce a significant negative alpha. generally speaking, results from table 8 are similar to table 7, which confirms the proposition that not only the nine fidelity funds from lin (2014), but also a large number of sector funds in this study, have experienced a performance deterioration. why is the so-called smart money not as smart as investors observed a decade ago? our explanation is that the u.s. stock market has become more efficient and it is more difficult to hunt alphas even for the smart money. the game has changed as investment assets of low cost index strategies has surpassed assets of active managed funds for the first time in september 2019 (lim, 2019). the time trend analysis of alphas shown in this study provides explanations of the movement that investors shift their money from active to passive strategies. 5. conclusions this paper investigates performance of united states actively managed sector equity mutual funds and their peer spdr etfs for the period of 2008–2017. the sample includes 95 actively managed individual sector mutual funds from 29 fund families. a portfolio with multi-funds is constructed for each of the nine sectors: materials, energy, financials, industrials, technology, consumer staples, utilities, healthcare, and consumer discretionary. results show little evidence that actively managed sector equity funds outperform their passive counterparties. only three or four out of the nine sector equity fund portfolios exhibit a higher sharpe ratio or a positive information ratio for the sample period. none of the sector equity fund portfolios produces a significant positive alpha using factor models with spy as the market portfolio, even though several peer sector etfs do generate significant positive alphas. none of the sector equity fund portfolios delivers a significant positive alpha against their passive managed peer etfs through the one-factor model. 174 y. fan, c.y. lin / financial services review 28 (2020) 159–177 t ab le 8 t im e tr en d o f al p h as w it h p ee r e t f s as m ar k et p o rt fo li o s p an el a : p er io d 1 9 9 9 -2 0 1 7 s ec to rs n o . o f fu n d s (6 0 ) a e st im at e p -v al u e o f a b e st im at e p -v al u e o f b a d ju st ed r 2 n o . fu n d s w it h si g n ifi ca n t p o si ti v e a n o . fu n d s w it h si g n ifi ca n t n eg at iv e a m at er ia ls 8 0 .7 6 % 0 .2 0 2 0 .7 8 0 .0 0 0 .2 2 0 0 e n er g y 5 0 .1 0 % 0 .6 2 4 1 .1 5 0 .0 0 0 .8 5 0 0 f in an ci al s 1 1 0 .3 6 % 0 .1 0 2 0 .4 8 0 .0 0 0 .5 4 3 0 in d u st ri al s 4 0 .3 2 % 0 .0 6 4 0 .8 1 0 .0 0 0 .7 4 2 0 t ec h n o lo g y 1 4 0 .3 3 % 0 .0 7 1 1 .0 9 0 .0 0 0 .8 8 3 0 u ti li ti es 7 0 .1 6 % 0 .3 1 4 0 .7 2 0 .0 0 0 .5 0 0 0 h ea lt h ca re 8 0 .4 8 % 0 .0 6 8 0 .7 7 0 .0 0 0 .6 4 3 0 c o n su m er d is cr et io n ar y 3 0 .1 6 % 0 .3 7 2 0 .8 6 0 .0 0 0 .7 3 0 0 p an el b : s u b -p er io d 1 9 9 9 -2 0 1 0 s ec to rs n o . o f fu n d s (6 0 ) a e st im at e p -v al u e o f a b e st im at e p -v al u e o f b a d ju st ed r 2 n o . fu n d s w it h si g n ifi ca n t p o si ti v e a n o . fu n d s w it h si g n ifi ca n t n eg at iv e a m at er ia ls 8 0 .8 6 % 0 .0 2 9 0 .7 4 0 .0 0 0 .5 4 6 0 e n er g y 5 0 .4 4 % 0 .1 2 0 1 .1 4 0 .0 0 0 .8 3 1 0 f in an ci al s 1 1 0 .4 0 % 0 .0 3 6 0 .6 7 0 .0 0 0 .8 1 2 0 in d u st ri al s 4 0 .3 4 % 0 .1 5 4 0 .7 9 0 .0 0 0 .7 3 1 0 t ec h n o lo g y 1 4 0 .5 9 % 0 .0 0 9 1 .1 0 0 .0 0 0 .9 1 5 0 u ti li ti es 7 0 .1 7 % 0 .4 3 0 0 .7 2 0 .0 0 0 .6 2 0 0 h ea lt h ca re 8 0 .5 5 % 0 .1 0 0 0 .6 1 0 .0 0 0 .2 9 3 0 c o n su m er d is cr et io n ar y 3 0 .1 7 % 0 .5 3 1 0 .8 3 0 .0 0 0 .7 0 0 0 p an el c : s u b -p er io d 2 0 1 1 -2 0 1 7 s ec to rs n o . o f fu n d s (6 0 ) a e st im at e p -v al u e o f a b e st im at e p -v al u e o f b a d ju st ed r 2 n o . fu n d s w it h si g n ifi ca n t p o si ti v e a n o . fu n d s w it h si g n ifi ca n t n eg at iv e a m at er ia ls 8 0 .5 1 % 0 .7 3 3 0 .9 1 0 .0 0 0 .0 9 0 6 e n er g y 5 �0 .4 9 % 0 .0 2 6 1 .1 7 0 .0 0 0 .9 1 0 2 f in an ci al s 1 1 0 .7 3 % 0 .0 8 5 0 .2 4 0 .0 0 0 .2 1 6 0 in d u st ri al s 4 0 .1 9 % 0 .4 1 6 0 .9 1 0 .0 0 0 .7 5 0 0 t ec h n o lo g y 1 4 �0 .0 8 % 0 .8 0 2 1 .0 7 0 .0 0 0 .6 6 0 0 u ti li ti es 7 0 .1 2 % 0 .5 4 4 0 .7 3 0 .0 0 0 .6 8 0 0 h ea lt h ca re 8 �0 .2 4 % 0 .5 5 3 1 .2 2 0 .0 0 0 .5 8 0 0 c o n su m er d is cr et io n ar y 3 0 .0 0 % 0 .9 9 2 0 .9 7 0 .0 0 0 .8 3 0 0 n o te : n u m b er o f o b se rv at io n s: p an el a -2 2 8 m o n th ly re tu rn s; p an el b -1 4 4 m o n th ly re tu rn s; p an el c -8 4 m o n th ly re tu rn s. m o d el u se d : r m f j,t � r f, t ¼ a j þ b j r e t f j,t � r f, t ð þþ « j,t .w h er e r m f j,t is th e re tu rn o f in d iv id u al m u tu al fu n d /m u tu al fu n d p o rt fo li o o f se ct o r j in m o n th t, r f, ti s th e re tu rn o f o n em o n th t -b il l in m o n th t, an d r e t f j,t is th e re tu rn o f e t f in th e sa m e se ct o r j in m o n th t, an d « j,t is an er ro r te rm . y. fan, c.y. lin / financial services review 28 (2020) 159–177 175 moreover, we conduct time trend analysis of fund performance over an extended period, 1999-2017. our results indicate that outperformance fades away as time passes. this is true for both the nine fidelity sector mutual funds studied in lin (2014) and a sample of 60 sector mutual funds with data available in the 19-year period excluding the nine fidelity funds. although there are some sector mutual funds outperform in the 1999-2010 period, much less funds outperform in the later period of 2011-2017. this result holds for both equal-weighted sector mutual fund portfolios and individual sector mutual funds. this could be interpreted as the u.s. sector equity market has become more efficient in the past decade than the decade before. our results are consistent with many researches, which argue that active managers do not add value after fees and expenses on average, for example, barras, scaillet, and wermers (2010). the value of active management may lie on the concept of making market more efficient by decisions such as asset allocation and security selection as stated by jones and wermers (2011). for investors who are interested in sector investing, this study shows that actively managed mutual funds may not be a good candidate for sector allocation or rotation. instead, passive sector etfs, which are index-tracking vehicles, can provide average returns with much lower costs and simplicity. notes 1 to save space, tables for the two sub-sample period are not presented. results are available upon request. 2 to save space, the rolling 36-month correlations are not reported here. results are available upon request. references barras, l., scaillet, o., & wermers, r. (2010). false discoveries in mutual fund performance: measuring luck in estimated alphas. the journal of finance, 65, 179–216. bogle, j. c. (2015). bogle on mutual funds: new perspectives for the intelligent investor. hoboken, nj: wiley. bogle, j. c. (2016). the index mutual fund: 40 years of growth, change, and challenge. financial analysts journal, 72, 9-13. brown, s. j. (2016). why hedge funds? financial analysts journal, 72, 5-7. carhart, m. m. (1997). on persistence in mutual fund performance. the journal of finance, 52, 57-82. chen, h., estes, j., & pratt, w. (2018). investing in the healthcare sector: mutual funds or etfs. managerial finance, 44, 495-508. cremers, k. j., fulkerson, j. a., & riley, t. b. (2019). challenging the conventional wisdom on active management: a review of the past 20 years of academic literature on actively managed mutual funds. financial analysts journal, 75, 8-35. dellva, w. l., demaskey, a. l., & smith, c. a. (2001). selectivity and market timing performance of fidelity sector mutual funds. the financial review, 36, 39-54. fama, e. f., & french, k. r. (1993). common risk factors in the returns on stocks and bonds. journal of financial economics, 33, 3-53. fan, y., & addams, h. l. (2012). united states-based international mutual funds: performance and persistence. financial services review, 21, 51-61. jones, r. c., & wermers, r. (2011). active management in mostly efficient markets. financial analysts journal, 67, 29-45. 176 y. fan, c.y. lin / financial services review 28 (2020) 159–177 kaushik, a., pennathur, a., & barnhart, s. (2010). market timing and the determinants of performance of sector funds over the business cycle.managerial finance, 36, 583-602. kaushik, a., saubert, l. k., & saubert, r. w. (2014). performance and persistence of performance of healthcare mutual funds. financial services review, 23, 77-91. lin, c. y. (2014). does active management work? evidence from equity sector funds. financial services review, 23, 249-271. lim, d. (2019). index funds are the new kings of wall street. the wall street journal, 18 september, 2019. malkiel, b. g. (1973). a random walk down wall street. new york, ny: norton. malkiel, b. g. (1995). returns from investing in equity mutual funds, 1971–1991. the journal of finance, 50, 549-572. malkiel, b. g., & radisich, a. (2001). the growth of index funds and the pricing of equity securities. the journal of portfolio management, 27, 9-21. petajisto, a. (2013). active share and mutual fund performance. financial analysts journal, 69, 73-93. tergesen, a., & zweig, j. (2016). the dying business of picking stocks.wall street journal, 17 october, 2016. wermers, r., & moskowitz, t. j. (2000). mutual fund performance: an empirical decomposition into stockpicking talent, style, transactions costs, and expenses/discussion. the journal of finance, 55, 1655-1703. y. fan, c.y. lin / financial services review 28 (2020) 159–177 177 finser_31-2_complete_issue timing and frequency of financial education and positive financial behaviors lua a. v. augustina,*, terrance k. martinb aslippery rock university, 1 morrow way, slippery rock, pa, 16057, usa bwinston-salem state university, 601 s. martin luther king jr. drive, winston salem, nc 27110, usa abstract the association between timing and frequency of financial education and financial behaviors has not been studied in tandem. various studies indicate the more financial education individuals have, the better their chances of making quality financial decisions. in addition to receiving financial education, the timing and frequency of that education may be relevant to overall decision quality. early and frequent exposures are expected to have a compounding effect on positive financial behaviors. using the 2015 wave of the national financial capability study, we found that the timing and frequency of financial education are positively associated with positive financial behaviors. this is important because it indicates that the more individuals know about their finances the more likely they are to engage in actions that will have a positive association with their financial futures. © 2023 academy of financial services. all rights reserved. jel classifications: g5 keywords: financial literacy; financial education; financial behaviors 1. introduction it is becoming more important for individuals to manage their own personal finances. market changes, such as the move from defined benefit plans to defined contribution plans place more responsibility on individuals. changes in technology are making banking products more interactive and allowing individuals to direct their own savings and investing goals. individuals also face wide variation in the terms and conditions attached to credit cards, mortgages, and online accounts. this makes the selection of these products more *corresponding author. tel.: 724-738-4374, fax: 724-738-2959. e-mail address: lua.augustin@sru.edu (l. a. v. augustin) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 151–167 difficult and likely increases the chances of making a financial mistake. financial education and literacy can help individuals make better-informed decisions when selecting among alternatives. financial education lays the groundwork that can later be transformed into financial literacy skills. individuals can attain financial literacy by first getting a financial education and then learning how to apply that knowledge to their financial situation. however, getting a financial education does not automatically lead to being financially literate. financial education is simply getting exposure to financial concepts. in many cases, employers provide this information via seminars or handouts. it is up to the employees to learn the material and then apply it to their own financial situation. at the college level financial education is in the form of coursework over a college term. this is a more intensive level of applying the knowledge to real life; however, in many cases, these courses are electives and go unnoticed by students. at the public school level, there may be some basic financial education lumped in with other general economics courses (willis, 2009). again, these are usually elective courses and do not reach the entire student body. sadly, financial education is lacking in many high schools and even in some colleges. when financial education courses are available, there is a large variation in the quality of the course offerings. individuals with higher levels of financial literacy tend to make more positive financial decisions and reduce the number and severity of poor financial decisions. high financial literacy is associated with positive financial behaviors such as saving for retirement (bayer, bernheim, & scholz, 2009; sekita et al., 2011) and diversifying the portfolio (van rooij, lusardi, & alessie, 2011; vinet & zhedanov, 2010). in addition, those with high financial literacy experience greater financial well-being (taft, hosein, mehrizi, & roshan, 2013). low financial literacy is associated with negative financial behaviors such as having high credit card debt and using payday loans. the more financially literate an individual is, the better he/she can navigate complex financial products. we know that exposure to concepts affects learning and can build skills over time (heckman, 2006). we examine the role that timing and frequency of financial education play in individuals practicing positive three financial behaviors. these financial behaviors are (1) owning an ira, (2) calculating retirement needs, and (3) contributing regularly to retirement. we found that not only is the timing important, but that frequency matters as well. the earlier individuals get exposure to financial literacy the more likely they are to make good financial decisions. our results suggest that getting financial education at work is associated with all three financial behaviors. we also found evidence that getting financial education three times is positively associated with owning an ira and calculating retirement needs. 2. literature review we turn to theory to examine the role that the timing and frequency of financial education plays in making positive financial behaviors. human capital theory predicts that individuals 152 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 will invest time in acquiring knowledge when the expected return exceeds the expected time and transaction costs (becker, 1994; blundell, dearden, meghir, & sianesi, 1999). if individuals can see the benefits from gaining financial education, they are more likely to access this education. the literature shows that those with higher levels of financial literacy tend to make better financial decisions (carlin & robinson, 2012; de bassa scheresberg, 2013; disney & gathergood, 2013). a rational individual would acquire this financial knowledge early in life to compound the effects of good financial decisions. when coupled with the life cycle theory we note that acquiring financial education at an earlier age may increase the likelihood of an individual being able to maximize the utility from consumption over time. this is because they should be able to acquire and allocate resources more effectively and efficiently at an earlier point in time. we also note that the life cycle theory explains the participation in the three financial behaviors. individuals at the accumulation phase are more likely to calculate their retirement needs than those in the acquisition phase and we found support for this in our results. we also found that individuals in the distribution phase are still more likely to have calculated their retirement needs than those in the acquisition phase. financial knowledge helps with decision making because financial products vary in complexity. financial products are becoming more complex and require higher levels of financial literacy. with the shift from defined benefit plans to defined contribution plans, individuals now have more responsibility over their retirement account decisions. under the defined benefit plans, individuals are guaranteed to receive some dollar amount at retirement. the employer bore the investment risk and made the investment decisions. under the defined contribution plans these guarantees disappear. individuals now bear the full risk of the investment and must choose where to invest. there is no guaranteed dollar amount upon retirement, as the portfolio value depends on how the investments perform over time. individuals must determine how much to contribute, where to invest it, and how to make the retirement income cover their needs through the end of life. along with these decisions, they also must navigate the taxable effects of withdrawals during retirement. the problem of managing financial decisions starts even before retirement. college students are plagued with increasingly higher levels of debt (lusardi, mitchell, & curto, 2010) and show low levels of financial literacy (chen & volpe, 1998), despite increased exposure to personal finance concepts. this is even more disturbing when coupled with the fact that older generations show low literacy levels as well (chen & volpe, 1998). in 2006, the us department of the treasury launched the first national program aimed at increasing literacy levels nationwide (remund, 2010). this spurred movements such as the jump$tart coalition for financial literacy. the coalition launched a survey among high school students. it measures what they know, as well as areas of concern relating to financial products. after the introduction of the coalition, financial education scores increased slightly over time. however, financial education and financial literacy are not the same. 2.1. financial education versus financial literacy financial literacy is not to be confused with financial education. financial education is simply gaining knowledge and understanding of financial products and processes (schwab l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 153 et al., 2008). this can be done face to face or online and falls under many formats within those two platforms. the content of these courses is not well regulated and can vary widely in quality (willis, 2008). financial education courses provide the basic knowledge, but this still needs to be combined with human capital to make financial decisions. some argue that financial education courses simply do not work (willis, 2008) and early exposure to the topics is futile. financial education courses may increase consumer confidence in their ability to correctly manage their finances (bernheim et al., 2001; guiso & jappelli, 2006; willis, 2008), while doing little to affect actual capability (bernheim et al., 2001). the term “financial literacy” refers to the ability to combine knowledge and skills to effectively manage finances (huston, 2010; remund, 2010). those who are more financially literate make better financial decisions overall. they save more for retirement, have less debt, and diversify their investments (guiso & jappelli, 2006; van rooij et al., 2011). individuals who are less financially literate tend to have higher debt (gathergood, 2012; lusardi & tufano, 2009), utilize predatory lending, and have less savings (campbell, 2006; lusardi & tufano, 2009). there is mixed evidence of the association between early financial literacy classes on financial decisions. studies have shown that high school students are not financially literate (chen & volpe, 1998), and that they are unable to answer basic questions about money management and investing. other studies show the lasting association between early financial education in high school (bernheim et al., 2001; lührmann, serra-garcia, & winter, 2015) and even elementary school (batty, collins, & odders-white, 2015). in these studies, even brief exposure to financial education had effects that lasted months afterward (ning & peter, 2015). the students who received financial education not only had more positive attitudes but were also more inclined to save. they also tended to have higher credit scores (brown, collins, schmeiser, & urban, 2014) and lower delinquency rates than those who did not get financial education in school. more states are including financial education for high schools, but it is not mandatory or tested in all states. this paper examines the association between timing and frequency of exposure to financial education and positive financial behaviors. we found support for the hypotheses that timing and frequency (kaiser & menkhoff, 2017) are associated with quality financial decisions. 3. data, hypotheses, and methods 3.1. data and sample this paper uses the 2015 data from the national financial capability study (nfcs). this is a nationally representative study conducted on 27,564 american adults. it is a state-bystate survey and samples about 500 individuals per state and includes the district of columbia. the following states represent an oversampling of individuals: new york, texas, illinois, and california. the three dependent variables are explicitly captured in questions in the study. the study contains questions related to the point in time when respondents receive 154 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 financial education and how many times they received this education. these are the two primary independent variables and are explicitly recorded in the data. there are also a number of questions that capture the various independent variables of interest, such as the income level, employment status, and risk tolerance levels of respondents. we arrive at the final sample of 3,794 observations after censoring the data by removing any nonresponses to the variables of interest. the censoring process is described in more detail under the independent variables section below. 3.2. dependent variable financial behavior is based on questions that ask about specific behaviors, namely: owning an ira, calculating retirement needs, and regular contributions to retirement. we chose these variables as they represent positive financial behaviors. we code these three behaviors as dummy variables with 1 representing a true value of owning an ira, calculating retirement needs, and regular contributions to retirement, and 0 otherwise. each of these three variables are run as a separate regression. 3.3. independent variables we initialize the sample by removing any nonresponses to the main independent variables. the main independent variables consist of the following two questions: “was financial education offered by a school or college you attended, or a workplace where you were employed?” • yes, but i did not participate in the financial education offered • yes, and i did participate in the financial education • no • don’t know • prefer not to say we use only those respondents who chose “yes and participated in the financial education offered.” the follow-up question was: “when did you receive that financial education?” • in high school • college • at work • military respondents could answer “yes/no/don’t know/refuse to say” to each of the four choices above. we remove all responses of “no/don’t know/refuse to say” to capture only those who received the financial education at one of the four periods of time. we create dummy variables for each response to create the “timing” independent variable. we then combine the four responses to create values from 1 to 4, to create the “frequency” l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 155 independent variable. a respondent who scored one under frequency only received financial education at one point in time. a respondent who scored two would have received financial education at two points in time, and so on. we also removed any respondents who indicated that they had received some financial education but then answered “no” to all four options of when they received the financial education. other demographics include gender, age, race, income, education, marital status, employment status, number of children, and risk tolerance level. we code gender as binary with 1 being male. we code age into three ranges, based on the stage of life the respondent is in. acquisition encompasses ages 18–44, accumulation encompasses ages 45-64, and distribution encompasses ages 65 and above. we code race as binary with 1 representing white and 0 all other. we code income as high, moderate, and low, based on the distribution of the data. we code education as binary with 1 representing a college degree and above, and 0 representing less than a college degree. we code marital status as binary with 1 representing married and 0 representing all else. we code employment as employed, not employed, and retired. we code number of children as 1 representing the presence of children and 0 representing no children. we code risk tolerance as ranging from low to medium and ending with high-risk tolerance based on the distribution of the sample. respondents in the low-risk tolerance group are not willing to take financial risks. we create the financial literacy variable by using three commonly used financial literacy questions. the scores are combined to give a rank ranging from 0 correct to 3 correct. we code respondents who got all three questions correct as having high literacy and everyone else as not, using dummy variables. we remove any nonresponses to the demographic variables and had a final sample of 3,794 respondents. 3.4. model we estimate the model to be: financial behavior = f financial-education-timing, financial-education-frequency, demographicsð þ we use the binary probit model as follows: y#i = b 1ti þ b 2fi þ b 3gi þ b 4ai þ b 5rai þ b 6ii þ b 7edi þ b 8mi þ b 9emið þb 10ci þ b 11rti þ b 12fli þ « iþ yi = 1 if y#i > 0 0 if y#i ≥ 0 ! here y#i represents the probability of expressing one of the three types of financial behaviors. ti represents timing of financial education, fi represents frequency of financial education, gi represents gender, ai represents age, rai represents race, ii represents income, edi 156 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 represents education level, mi represents marital status, emi represents employment, ci represents number of children, rti represents risk tolerance, fli represents financial literacy, and « i represents the error term. 3.5. hypotheses we hypothesize that the timing of financial education courses will have a significant, positive association with the decision to own an ira, calculate retirement needs, and save for retirement. we also hypothesize that more frequent exposure to financial education courses will have a significant, positive relation to the three behaviors mentioned above. 4. methods we use the probit regression method to estimate the effects of the independent variables on the dependent variable. in this paper, we look at the outcomes of having early exposure to financial education classes on positive financial behaviors. we examine three main positive financial behaviors, namely owning an ira, saving for retirement, and calculating retirement needs. 5. results 5.1. univariate analysis table 1 shows the descriptive statistics for the sample. there are 3,794 observations in the final sample. when examining the positive financial behaviors, we found that about 47% of respondents own an ira and about 53% do not. about 49% calculate their retirement needs and about 49% continue to make regular contributions to retirement. timing of financial education measures when the respondents received financial education. there may be an overlap in each of the four categories since they simply answered whether they received financial education at the different points included in the survey question. we do not know the first point of exposure, simply whether respondents are exposed to financial education at one or more of the four options. about 53% received financial education in high school, increasing to about 67% during college. about 46% received financial education at work and about 9% received it in the military. frequency measures the number of times respondents received financial education. about 48% of respondents had been exposed to a financial education course twice, with only about 37% having taken more than two financial education courses. as frequency increased the proportion of respondents taking financial education courses drops, until only about 3% had taken four financial education courses. about 60% of the sample is male. about 44% of the sample are in the age range 18-44. this drops to about 36% in the age range 45-64 and closes with about 21% being age 65 and l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 157 table 1 descriptive statistics for timing and frequency sample (n 5 3,794) variable frequency financial behavior own ira no 46.92% yes 53.08% calculate retirement needs no 50.69% yes 49.31% regular contributions to retirement no 50.98% yes 49.02% timing of financial education high school no 47.47% yes 52.53% college no 33.32% yes 66.68% work no 54.51% yes 45.49% military no 90.91% yes 9.09% frequency of financial education one time 47.52% two times 33.66% three times 16.32% four times 2.50% male no 40.77% yes 59.23% life cycle age range acquisition (18–44) no 56.51% yes 43.49% accumulation (45–64) no 64.07% yes 35.93% distribution (65+) no 79.41% yes 20.59% white no 27.78% yes 72.22% income low no 78.81% yes 21.19% moderate no 50.74% yes 49.26% (continued on next page) 158 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 above. about 72% of the sample reports being white. the mean income is between $50,000 and $75,000. about 9% of the sample reports having only a high school education and about 45% reports having a college education. about 61% of the sample is married and about 26% report being single. about 61% of the sample are working, about 16% are not working, and about 22% are retired. more than half of the sample had no children. about 27% report having low risk tolerance, about 57% have moderate risk tolerance, and about 16% have highrisk tolerance. about 70% of the sample provided correct responses to all questions used to measure financial literacy. table 2 shows the distribution of positive financial behaviors. this table reports the three positive behaviors across each variable. the totals do not sum to 100% horizontally because the three behaviors are not mutually exclusive. for respondents who received financial education in high school we see that about 52% own an ira. about 52% have calculated table 1 (continued) variable frequency high no 70.45% yes 29.55% college degree no 33.39% yes 66.61% married no 39.22% yes 60.78% employment status working no 38.56% yes 61.44% not working no 83.74% yes 16.26% retired no 77.70% yes 22.30% children no 29.84% yes 70.16% risk tolerance low no 73.09% yes 26.91% moderate no 42.62% yes 57.38% high no 84.29% yes 15.71% financially literate no 29.63% yes 70.37% l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 159 table 2 distribution of positive financial behaviors (n 5 3,794) variable own ira calculate retirement needs contribute regularly to retirement timing high school yes 51.33% 51.53% 49.32% no 48.67% 48.47% 50.68% college yes 56.05% 51.23% 52.17% no 43.95% 48.77% 47.83% work yes 64.54% 53.77% 56.43% no 35.46% 46.23% 43.57% military yes 65.51% 58.55% 59.13% no 34.49% 41.45% 40.87% frequency one time yes 45.04% 43.21% 41.15% no 54.96% 56.79% 58.85% two times yes 55.60% 52.08% 53.56% no 44.40% 47.92% 46.44% three times yes 67.69% 58.80% 58.97% no 32.31% 41.20% 41.03% four times yes 76.84% 66.32% 72.63% no 23.16% 33.68% 27.37% male yes 56.34% 50.16% 52.02% no 43.66% 49.84% 47.98% age acquisition (18-44) yes 43.03% 58.36% 56.06% no 56.97% 41.64% 43.94% accumulation (45-64) yes 56.49% 56.93% 57.30% no 43.51% 43.07% 42.70% distribution (65+) yes 68.37% 16.90% 19.72% no 31.63% 83.10% 80.28% white yes 55.51% 48.07% 48.43% no 44.49% 51.93% 51.57% income low yes 21.77% 32.46% 17.66% no 78.23% 67.54% 82.34% moderate yes 55.22% 48.42% 49.81% no 44.78% 51.58% 50.19% high yes 71.99% 62.89% 70.21% no 28.01% 37.11% 29.79% (continued on next page) 160 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 retirement needs and about 50% make regular contributions to retirement. respondents who receive financial education in college show increased participation in positive financial behaviors. of the respondents who receive financial education in college, about 56% own an ira, about 51% have calculated retirement needs, and about 52% contribute regularly to retirement. of the respondents who receive financial education at work, about 65% own an ira, about 54% have calculated retirement needs, and about 53% contribute regularly to retirement. when we examine the frequency of financial education, we note an increasing trend as well. among respondents who received financial education one time, about 45% own an ira, about 43% have calculated retirement needs, and about 41% make regular contributions to retirement. among respondents who receive financial education two times we note that about 56% own an ira, about 52% have calculated retirement needs, and about 54% make regular contributions to retirement. of the respondents who receive financial education three times about 68% own an ira, about 54% have calculated retirement needs and about 59% contribute regularly to retirement. of the respondents who receive financial table 2 (continued) variable own ira calculate retirement needs contribute regularly to retirement degree (no degree) yes 59.87% 53.50% 55.76% no 40.13% 46.50% 44.24% married (not married) yes 61.84% 53.58% 57.63% no 38.16% 46.42% 42.37% employment working yes 54.10% 68.94% 66.19% no 45.90% 31.06% 33.81% not working yes 31.77% 42.79% 29.34% no 68.23% 57.21% 70.66% retired yes 65.84% 16.08% no 34.16% 100.00% 83.92% children (no children) yes 58.11% 51.54% 52.74% no 41.89% 48.46% 47.26% risk tolerance low yes 37.41% 31.83% 28.80% no 62.59% 68.17% 71.20% moderate yes 57.79% 53.28% 53.42% no 42.21% 46.72% 46.58% high yes 62.75% 64.77% 67.62% no 37.25% 35.23% 32.38% financially literate yes 56.40% 47.38% 48.46% no 43.60% 52.62% 51.54% l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 161 education four times, about 77% own an ira, about 66% have calculated retirement needs, and about 73% contribute regularly to retirement. about 56% of males own an ira, about 50% have calculated retirement needs, and about 52% make regular contributions to retirement. during the acquisition stage about 43% of respondents own an ira, about 58% have calculated retirement needs, and about 56% make regular contributions to retirement. during the asset accumulation stage, we see an increase in the proportions who own an ira and make regular contributions to retirement, but note a decrease in the proportion who have calculated retirement needs. during the distribution phase about 68% own an ira, about 17% have calculated retirement needs, and about 20% contribute regularly to retirement. the drop in retirement needs and contribution levels are consistent with life cycle theory. at this life stage individuals would be focused on asset decumulation and not accumulation. we note an increase in the proportions of respondents who report all three positive behaviors as income increases. about 22% of low-income individuals own an ira, compared to about 55% with moderate income, and about 72% with high income. about 32% of individuals with low income have calculated retirement needs, compared to about 48% with moderate income, and about 63% with high income. about 18% of individuals with low income contribute regularly to retirement, compared to about 50% with moderate income and about 70% with high income. about 60% of individuals with a college degree own an ira, about 54% have calculated retirement needs, and about 56% contribute regularly to retirement. we also note that larger proportions of working respondents engage in the three behaviors, compared to those who are not working or retired. about 54% of working individuals own an ira, compared to about 32% who are not working, and about 66% of the retired population. about 69% of the working population has calculated retirement needs and about 43% of those who are not working have also calculated retirement needs. about 58% of individuals who have children own an ira, about 52% have calculated retirement needs, and about 53% contribute regularly to retirement. about 37% of respondents with low risk tolerance own an ira, compared to about 58% with moderate risk tolerance, and about 63% with high risk tolerance. about 56% of individuals with high financial literacy own an ira, about 47% have calculated retirement needs, and about 49% regularly contribute to retirement. 5.2. multivariate analysis table 3 shows the marginal effects of the probit regressions on the positive financial behavior variables. we run three separate regressions with three different dependent variables. the main variables of interest are timing and frequency of financial education. we found that timing and frequency are statistically significant for most instances of the three positive financial behaviors. 5.3. owning an ira getting financial education at work and in the military is positively and significantly associated with owning an ira. individuals who got financial education at work and in the 162 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 t ab le 3 m ar g in al ef fe ct s o f in d ep en d en t v ar ia b le s o n p o si ti v e fi na n ci al b eh av io rs v ar ia b le o w n ir a c al cu la te re ti re m en t n ee d s c o n tr ib u te to re ti re m en t re g u la rl y m ar g . ef f. s ta n d ar d er ro r s ig . m ar g . ef f. s ta n d ar d er ro r s ig . m ar g . ef f. s ta n d ar d er ro r s ig . t im in g (h ig h sc h o o l) c o ll eg e % 0 .0 0 1 7 (0 .0 2 1 4 ) 0 .0 0 7 4 (0 .0 2 33 ) 0 .0 3 2 9 5 (0 .0 1 9 3 ) * w o rk 0 .0 6 6 2 (0 .0 2 0 5 ) * * * 0 .0 9 0 6 7 (0 .0 2 35 ) * * * 0 .0 8 6 4 6 (0 .0 1 8 7 ) * * * m il it ar y 0 .0 6 0 1 (0 .0 3 2 5 ) * 0 .0 7 1 1 5 (0 .0 3 79 ) * 0 .0 2 6 3 5 (0 .0 2 9 2 ) f re qu en cy (o n e ti m e) t w o ti m es 0 .0 1 7 7 (0 .0 2 0 4 ) 0 .0 3 2 3 6 (0 .0 2 30 ) 0 .0 2 1 5 2 (0 .0 1 8 4 ) t h re e ti m es 0 .0 7 5 4 (0 .0 3 1 5 ) * * 0 .0 8 6 9 8 (0 .0 3 61 ) * * 0 .0 2 7 6 7 (0 .0 2 8 6 ) f o u r ti m es 0 .0 8 2 2 (0 .0 6 6 8 ) 0 .0 6 3 1 6 (0 .0 8 05 ) 0 .0 7 8 4 1 (0 .0 6 2 4 ) m al e % 0 .0 1 9 4 (0 .0 1 5 4 ) % 0 .0 2 4 2 (0 .0 1 74 ) % 0 .0 0 7 5 (0 .0 1 4 0 ) a g e (a cq u is it io n ) a cc u m u la ti o n 0 .0 6 6 8 (0 .0 1 7 4 ) * * * 0 .0 5 1 3 4 (0 .0 1 83 ) * * * 0 .0 0 0 8 4 (0 .0 1 5 8 ) d is tr ib u ti o n 0 .1 8 4 8 (0 .0 2 7 4 ) * * * 0 .0 7 7 2 1 (0 .0 3 65 ) * * % 0 .1 6 8 2 (0 .0 2 4 3 ) * * * w h it e 0 .0 4 5 3 (0 .0 1 6 6 ) * * * % 0 .0 0 2 6 (0 .0 1 83 ) 0 .0 0 1 8 6 (0 .0 1 5 1 ) in co m e (l o w ) m o d er at e 0 .2 2 0 1 (0 .0 2 0 6 ) * * * 0 .0 9 9 1 2 (0 .0 2 26 ) * * * 0 .2 0 8 9 8 (0 .0 1 9 0 ) * * * h ig h 0 .3 2 5 3 (0 .0 2 3 8 ) * * * 0 .1 8 7 7 1 (0 .0 2 71 ) * * * 0 .3 1 9 2 9 (0 .0 2 1 5 ) * * * d eg re e (n o d eg re e) 0 .0 7 2 6 (0 .0 1 7 2 ) * * * 0 .0 5 6 1 8 (0 .0 1 95 ) * * * 0 .0 5 4 9 2 (0 .0 1 5 6 ) * * * m ar ri ed (n o t m ar ri ed ) 0 .0 4 0 1 (0 .0 1 7 5 ) * * 0 .0 1 6 3 4 (0 .0 1 99 ) 0 .0 6 4 9 4 (0 .0 1 5 8 ) * * * e m p lo y m en t (w o rk in g) n o t w o rk in g % 0 .0 1 7 6 (0 .0 2 2 2 ) % 0 .1 0 2 1 (0 .0 2 12 ) * * * % 0 .1 3 8 6 (0 .0 1 8 8 ) * * * r et ir ed 0 .0 5 1 7 (0 .0 2 4 4 ) * * o m it te d o m it te d o m it te d % 0 .3 2 6 1 (0 .0 2 0 0 ) * * * c h il d re n (n o ch il d re n ) 0 .0 1 4 6 (0 .0 1 7 6 ) 0 .0 7 0 4 7 (0 .0 1 93 ) * * * 0 .0 4 9 5 7 (0 .0 1 6 0 ) * * * r is k to le ra n ce (l o w ) m o d er at e 0 .1 6 3 0 (0 .0 1 7 2 ) * * * 0 .1 2 6 3 4 (0 .0 1 98 ) * * * 0 .0 9 9 3 7 (0 .0 1 5 8 ) * * * h ig h 0 .2 6 2 8 (0 .0 2 3 9 ) * * * 0 .1 5 9 8 7 (0 .0 2 68 ) * * * 0 .1 8 1 (0 .0 2 1 8 ) * * * f in an ci al ly li te ra te 0 .0 3 2 7 (0 .0 1 7 2 ) * 0 .0 0 2 9 4 (0 .0 1 88 ) 0 .0 0 5 2 7 (0 .0 1 5 8 ) n o te s. f o r o w n ir a an d co n tr ib u te re g u la rl y to re ti re m en t (n ¼ 3 ,7 9 4 ), fo r ca lc u la te re ti re m en t n ee d s (n ¼ 2 ,9 4 8 ). d en o te s si g n ifi ca n ce at th e fo ll o w in g le v el s: * p < .1 , * *p < .0 5 , * * * p < .0 1 . b ei n g re ti re d is a p er fe ct p re d ic to r fo r ca lc u la ti n g re ti re m en t n ee d s an d w as d ro p p ed fr o m th e re g re ss io n b y o u r so ft w ar e. l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 163 military were more likely to own an ira than those who got financial education in college. individuals with three exposures to financial education were more likely to own an ira than those with one exposure. we found that as age increased, individuals were more likely to own an ira. consistent with theory, we found that at both the accumulation and distribution stages, individuals were more likely to own an ira than at the acquisition stage. we found that whites were more likely to own an ira than nonwhites. we found that as income increases individuals are more likely to own an ira. individuals with a college education were more likely to own an ira than those with no degree, and married respondents were more likely to own an ira than those who were not married. retired individuals were more likely to own an ira than those still in the work force and as risk tolerance increases individuals were more likely to own an ira than those with low risk tolerance. finally, we found that financially literate individuals were more likely to own an ira than those who were not. 5.4. calculating retirement needs the sample size for this regression is 2,948. we found that respondents who get financial education at work and in the military were more likely to calculate retirement needs compared to those who get it in high school. we found a high, statistically significant positive relationship between getting financial education three times and calculating retirement needs. we found that respondents in the accumulation and distribution phases were more likely to calculate retirement needs compared to those in the acquisition phase. we found that as income increases individuals were also more likely to have calculated retirement needs. we found that individuals who are not working were less likely to calculate retirement needs than those who are working. having children increases the likelihood of calculating retirement needs. we found that both moderate and high risk tolerance levels increase the likelihood of calculating retirement needs, compared to those with low risk tolerance. 5.5. contributing regularly to retirement individuals who get financial education in college are more likely to contribute regularly to retirement than those who get a financial education in high school. we found that getting a financial education at work increases the likelihood of contributing regularly towards retirement, compared to getting a financial education in high school. we found that at the distribution stage individuals were less likely to contribute regularly to retirement than those in the acquisition stage, consistent with the life cycle theory. individuals with moderate and high incomes were more likely to contribute regularly to retirement than those with low incomes. we found that having a college degree and being married increases the likelihood of contributing regularly to retirement, compared to not having a college degree and being unmarried. we found that individuals who are retired or not working were less likely to contribute regularly to retirement than those who are working. the presence of children increases the likelihood of contributing to retirement compared to those without children. finally, as risk tolerance increases so does the likelihood of contributing to retirement. 164 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 6. discussion the study examines the association between the timing and frequency of financial education on financial behaviors. we found evidence that getting a financial education at work has the most significant association with positive financial behaviors. this indicates that getting financial education at work may highly influence the decision to own an ira, calculate retirement needs, and regularly contribute to retirement. we also found that three exposures to financial education have the most significant association with the three positive financial behaviors. this indicates that frequency of financial education may be vital to individuals making financial decisions. these results are consistent with our hypotheses that both timing and frequency are positively associated with positive financial behaviors. theoretically, we expect to see a direct relationship between financial education and financial literacy. the more financial education one receives, the more financially literate we expect that individual to be. this increased financial literacy would then have a direct association with the financial decision quality that individuals make. based on the results of the regression we fail to observe significant evidence that financial literacy has a strong association with the likelihood of engaging in the behaviors analyzed in the study. what we do note is that both timing and frequency of financial education are relevant in making quality financial decisions. these findings in combination may warrant the need for additional research on the relationship between financial education and financial literacy and how the two move in tandem to affect the decisions and outcomes experienced by individuals. one limitation of this study is that we do not know the quality of the financial education course taken by the respondents. the nature of financial education courses is quite nebulous and can range from incredibly detailed to summary overviews of financial topics. future research will examine which combinations of timing and frequency are associated with financial behaviors. another avenue for future research would be to examine the relationship between timing and frequency of financial literacy and negative financial behaviors. references ando, a., & modigliani, f. (1963). the “life cycle” hypothesis of saving: aggregate implications and tests author (s). the american economic review, 53, 55–84. attanasio, o., & weber, g. (2010). consumption and saving: models of intertemporal allocation and their implications for public policy. journal of economic literature, 48, 693–751. bae, s. c., & sandager, j. p. (1997). what consumers look for in financial planners. association for financial counseling and planning education, 8, 9–16. banks, j., blundell, r., & turner, s. (1998). is there a retirement-savings puzzle? the american economic review, 88, 769–788. barber, b. m., & odean, t. (2000). trading is hazardous to your wealth: the common stock investment performance of individual investors. the journal of finance, 55, 773–806. batty, m., collins, j. m., & odders-white, e. (2015). experimental evidence on the effects of financial education on elementary school students’ knowledge, behavior, and attitudes. journal of consumer affairs, 49, 69–96. bayer, p. j., bernheim, b. d., & scholz, j. k. (2009). the effects of financial education in the workplace: evidence from a survey of employers. economic inquiry, 47, 605–624. l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 165 becker, g. (1994). human capital revisited. journal of chemical information and modeling, 53, 1689–1699. bernheim, b. d., garrett, d. m., & maki, d. m. (2001). education and savings: the long-term effects of high school financial curriculum mandates. journal of public economics, 80, 435–465. blundell, r., dearden, l., meghir, c., & sianesi, b. (1999). human capital investment: the returns from education and the economy. fiscal studies, 20, 1–23. brown, a., collins, j. m., schmeiser, m., & urban, c. (2014). state mandated financial education and the credit behavior of young adults (fedsworking paper no. 2014-68). available at ssnr https://ssrn.com/abstract¼2498087 campbell, j. y. (2006). household finance. journal of finance, 61, 1553–1604. carlin, b. i., & robinson, d. t. (2012). what does financial literacy training teach us? the journal of economic education, 43, 235–247. chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. de bassa scheresberg, c. (2013). financial literacy and financial behavior among young adults: evidence and implications. numeracy, 6, article 5. disney, r., & gathergood, j. (2013). financial literacy and consumer credit portfolios. journal of banking & finance, 37, 2246–2254. fox, g. l., & chancey, d. (1998). sources of economic distress: individual and family outcomes. journal of family issues, 19, 725–749. de nardi, m., french, e., & jones, j. b. (2010). why do the elderly save? the role of medical expenses. journal of political economy, 118, 39–75. gathergood, j. (2012). self-control, financial literacy and consumer over-indebtedness. journal of economic psychology, 33, 590–602. gokhale, j., & kotlikoff, l. j. (2003). who gets paid to save? tax policy and the economy, 17, 111–139. guiso, l., & jappelli, t. (2006). information acquisition and portfolio performance. csef working paper, 00235, 67. hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking & finance, 36, 509–524. hamerrnesh, d. s. (1982). consumption during retirement: the missing link in the life cycle. review of economics and statistics, 66, 1–7. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 36–62. heckman, j. j. (2006). investing in disadvantaged children. social sciences, 312, 2005–2007. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 106, 309–322. howlett, e., kees, j., & kemp, e. (2008). the role of self regulation, future orientation, and financial knowledge in long term financial decisions. journal of consumer affairs, 42, 223–242. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. jappelli, t., & padula, m. (2013). investment in financial literacy and saving decisions. journal of banking & finance, 37, 2779–2792. kaiser, t., & menkhoff, l. (2017). does financial education impact financial literacy and financial behavior, and if so, when? the world bank economic review, 31, 611–630. kramer, m. m. (2012). financial literacy, cognitive ability, and financial advice-seeking. working paper. available at ssnr https://ssrn.com/abstract¼2081795 laster, d., suri, a., & vrdoljak, n. (2012). systematic withdrawal strategies for retirees. the journal of wealth management, 15, 36–49. loibl, c., & hira, t. k. (2005). self-directed financial learning and financial satisfaction. journal of financial counseling and planning, 16, 11. lührmann, m., serra-garcia, m., & winter, j. (2015). teaching teenagers in finance: does it work? journal of banking & finance, 54, 160–174. lusardi, a. (2008). financial literacy: an essential tool for informed consumer choice? working paper no. w1408. cambridge, ma: national bureau of economic research. lusardi, a., & mitchell, o. s. (2005). financial literacy and planning: implications for retirement wellbeing. michigan retirement research center research paper, no. wp, 10. available at ssnr https://ssrn.com/ abstract¼1848593 166 l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 lusardi, a., mitchell, o. s., & curto, v. (2010). financial literacy among the young: evidence and implications for consumer policy. journal of consumer affairs, 44, 358–380. lusardi, a., & tufano, p. (2015). debt literacy, financial experiences, and overindebtedness. journal of pension economics & finance, 14(4), 332–368. mcgarry, k. (1999). inter vivos transfers and intended bequests. journal of public economics, 73, 321–351. meier, s., & sprenger, c. d. (2013). discounting financial literacy: time preferences and participation in financial education program. journal of economic behavior & organization, 95, 159–174. mitchell, o. s. (1988). worker knowledge of pension provisions. journal of labor economics, 6, 21–39. ning, t., peter, p. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience. financial services review, 24, 119–137. picone, g., uribe, m., & wilson, r. m. (1998). the effect of uncertainty on the demand for medical care, health capital and wealth. journal of health economics, 17, 171–185. remund, d. l. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. rix, s. e., haider, s., hurd, m., reardon, e., & williamson, s. (2000). patterns of dissaving in retirement. washington, dc: aarp public policy institute. scholz, j. k., seshadri, a., & khitatrakun, s. (2006). are americans saving "optimally" for retirement? journal of political economy, 114, 607–643. schwab, c., iannicola, d., jr, beck, t., et al. (2008). president’s advisory council on financial literacy. washington, dc: the department of the treasury. sekita, s., horioka, c. y., kajitani, s., et al. (2011). financial literacy and retirement planning in japan. journal of pension economics and finance, 10, 637–656. sethi-iyengar, s., huberman, g., & jiang, g. (2004). how much choice is too much? contributions to 401(k) retirement plans. pension design and structure: new lessons from behavioral finance, 83, 84–87. taft, k. m., hosein, z. z., mehrizi, s. m. t., & roshan, a. (2013). the relation between financial literacy, financial wellbeing and financial concerns. international journal of business and management, 8. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. van rooij, m. c. j., lusardi, a., & alessie, r. j. m. (2012). financial literacy, retirement planning, and household wealth. the economic journal, 122, 449–478. vinet, l., & zhedanov, a. (2010). a “missing” family of classical orthogonal polynomials. quarterly journal of finance, 07, 1750008. willis, l. e. (2008). against financial literacy education. iowa law review, 94, (08-10). willis, l. e. (2009). evidence and ideology in assessing the effectiveness of financial literacy education. san diego law review, 46, 415–458. yuh, y., montalto, c. p., & hanna, s. (1998). are americans prepared for retirement? financial counseling and planning, 9, 1–13. l. a. v. augustin and t. k. martin / financial services review 31 (2023) 151–167 167 an examination of the federal employee retirement system (fers) survivor annuity benefit kevin davisa,*, steve p. fraserb, william w. jenningsc adepartment of management, usaf academy, 2354 fairchild drive, suite 6h-130, usaf academy, co 80840, usa bflorida gulf coast university, 10501 fgcu blvd south, fort myers, fl 33965, usa cdepartment of management, usaf academy, 2354 fairchild drive, suite 6h-130, usaf academy, co 80840, usa abstract the federal employees retirement system (fers) provides survivor annuity benefits for employees who forfeit a portion of their annuity as a premium. in this study, we construct a monte carlo simulation to describe the distributions and implied internal rates of return for fers annuitants who elect a joint and survivor annuity. our analysis suggests that the survivor benefit program is quite lucrative for most male retirees. in contrast, the program is less rewarding for female retirees, especially if younger than their spouse. for many female retirees, the program actually produces a negative return. retirees and planners can use our results to make more informed annuity decisions. © 2018 academy of financial services. all rights reserved. jel classification: g23; h55; j26; j38 keywords: survivor benefit plan; insurance; valuation; pensions; retirement; simulation 1. introduction unlike most defined-benefit plan studies, we focus on the value of the spousal insurance option. we explore the value of insuring the surviving spouse’s benefit under different scenarios using a cost-benefit analysis. considerable research exists on retirement annuities, as summarized in davis and fraser (2012) in their examination of the survivor benefit * corresponding author. tel.: �1-719-333-4130; fax: �1-719-333-9715. e-mail address: kevin.davis@usafa.edu (k. davis) financial services review 27 (2018) 133-146 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. program (sbp) for military annuitants. here we extend that analysis to the federal employees retirement system (fers) survivor benefit. limited academic research exists on fers, even though it has over 3 million active and retired participants (cbo, 2016). a comprehensive search for articles using databases such as abi inform and business source premier found few studies. these efforts simply describe the program (e.g., isaacs, 2010; purcell, 2009). books on the subject, like matthew and berman (2008), provide no information about the value of fers annuity insurance. we add to the literature by examining the costs and benefits of the fers survivor annuity for individuals. in the remaining sections, we review the parameters of the fers system, describe our simulation methods, and present our simulation results. after discussing the robustness and limitations of our approach, we conclude with implications for federal retirees. 1.1. federal employee retirement and the joint and survivor annuity while defined-benefit plans are becoming scarce for private sector workers, various government employees continue to earn defined-benefit retirements. federal government employees, for example, are eligible for the fers and can receive a lifetime annuity after as few as 5 years. for those under 62, fers utilizes a minimum retirement age (mra, analogous to social security’s full retirement age, fra) to determine eligibility. for most fers retirees, the amount of their annuity is based on a formula that gives retirees 1% credit for each year of service. their total percentage is then applied to their average salary during the 36-month period during which their pay was highest (high-3). for example, an employee who retired after exactly 20 years with a high-three average salary of $50,000 would be eligible for a $10,000 yearly annuity (20% of $50,000). after 5 years of service, government workers qualify for a retirement—however, if they have less than 10 years of service, they must wait until age 62 to initiate their benefits. currently, the mra is 56 if the worker has at least 10 years of service. a retirement based on mra is typically accompanied by a 5% per year reduction in the retiree’s annuity for each year short of age 62. however, with 20 years of service, a government worker can retire at age 60 without an accompanying annuity reduction; if the annuitant has 30 years of service, they can retire at age 55 without a reduction to their annuity. however, two incentives exist that encourage federal workers to continue working until age 62: 1. retirees younger than 62 are initially ineligible for cost of living adjustments; 2. new retirees older than 62, with at least 20 years, earn 1.1% credit for each year of service instead of 1%. the payments required to support fers retirement are significant. the cbo reports the federal government contributed about $26 billion to fers in 2014 (cbo, 2016). further, government contributions are expected to exceed $50 billion by 2027. senators richard burr (r-nc), tom coburn (r-ok), and saxby chambliss (r-ga) introduced legislation in 2013 to end this defined benefit pension (hicks, 2013). more recently, the trump administration has called for severe cuts to the entire fers program (burr, 2018; davidson, 2017). unless insured, retirement annuities are only paid until the retiree dies. the fers annuity insurance program offers the opportunity to insure either 25% or 50% of the retiree’s annuity. more accurately, the question is whether one should opt out of the benefit—50% joint and 134 k. davis et al. / financial services review 27 (2018) 133-146 survivor is the default option. (spousal consent is required to elect a 25% or 0% survivor benefit.) this choice to insure the benefit is the focus of our analysis. jennings and reichenstein (2001) examined retirement annuities for members of the military, outlining a valuation method while discussing the portfolio and asset allocation implications. they did not directly examine the costs and benefits of the insurance decision. davis and fraser (2012) analyzed the costs and benefits of the military survivor insurance program. this article builds on that research by examining a different federal program. we identify the relevant factors underlying the fers insurance decision, modeling and comparing the costs and benefits of the program. our research complements the work of milevsky (2006) who notes that retirees tend to focus too much on the accumulation of wealth and too little on preparing how to spend or protect wealth. 2. the fers survivor benefit annuity 2.1. general decision factors the fers insurance decision occurs at retirement. if the joint and survivor option is not selected within 30 days of retirement, a retired federal employee’s retirement income ceases upon death. if fers insurance is not declined, the retiree must choose to insure either 25% or 50% of their retirement annuity. for example, a retiree receiving $1,000 a month, who chooses to insure 50%, provides $500 each month to their surviving spouse. the premium to insure the 50% benefit is 10% of the retiree’s full annuity. the premium for the 25% benefit is 5% of the retiree’s annuity. continuing our example, a retiree with a gross retirement benefit of $1,000/month, who elects to insure 50%, will pay an insurance amount of $100/month during the insured spouse’s lifetime. this payment is pretax and reduces the retiree’s net pretax retirement from $1,000/month to $900/month. each of the factors in table 1 plays a role in the fers insurance selection. before making a final decision, retirees should consider the asset allocation implications discussed in jennings and reichenstein (2001) and the portfolio survivability implications of americks, veres, and warshawsky (2001); however, these effects do not impact our analysis. 2.2. individual factors to evaluate the insurance program, we considered the life expectancy of both the retiree and the surviving spouse. we use the same fig. 1 as in davis and fraser (2012) to illustrate our decision framework. for example, if a recently-retired couple lives together for 30 years and the spouse ultimately dies first, the couple will have paid premiums for 30 years and received no benefit. in contrast, if a retiree dies in the first month of retirement and the spouse lives for 30 years, the program would yield an extremely positive return on investment. we measure the value of fers insurance by examining the implied internal rate of return associated with the cash outflows over the life span of the retiree (the premiums or cost period—labeled “x”) with the cash inflows received by the spouse (the benefits period— labeled “z”). if z is positive, there will be benefits for the surviving spouse from fers 135k. davis et al. / financial services review 27 (2018) 133-146 annuity insurance. the likelihood that survivor benefits outweigh the costs increases with the duration of the benefit period z and decreases with increases in the duration of the retiree life span x. we model the probabilities associated with x, y, and z using actuarial tables provided by the social security administration (ssa).1 because ssa tables address averages for the total american population, it is important for planners and advisors to modify these distributions with subjective probabilities concerning heath and family circumstances. limitations notwithstanding, our analysis serves as a foundation for an informed decision. 2.3. external considerations in lieu of enrolling in the fers annuity insurance program, a retiring couple might seek to purchase an insurance policy to protect their retirement income stream (see jennings, merrell, o’malley, and payne, 2018). however, such insurance depends on the insurability of the retiree and the payoff depends on the solvency of the insurer. the fers program eliminates these concerns, reducing uncertainty for beneficiaries. in addition to the life expectancy of the retiree-beneficiary pair, there are other factors that affect the fers insurance decision, most notably inflation. fortunately, we can somewhat table 1 u.s. federal employee retirement system (fers) annuity insurance details costs 10% to insure 50% of pension; 5% to insure 25% of pension payment must start within 18 months of retirement (if it does not start immediately the retiree must pay a lump sum penalty) payment ceases when either retiree or spouse dies payment is pre-tax fers retirement is increased by cpi-w each year; hence the fers insurance payment increases by cpi-w each year (50% insurance always requires 10% of the retiree’s pension; 25% insurance always requires 5% of the retiree’s pension) benefits either 25% or 50% of the retiree’s unreduced pension amount benefit is taxable (but avoids payroll taxes) benefit is increased by cpi-w each year (see specific rules below) benefit ceases upon death of survivor specific cola rules if cpi-w ��2.0%, retiree or survivor annuity is increased by cpi-w if cpi-w �2% and cpi-w �3%, retiree or survivor annuity is increased by 2% if cpi-w ��3%, retiree or survivor pension is increased by cpi-w minus 1% fig. 1. timeline of lifespans. 136 k. davis et al. / financial services review 27 (2018) 133-146 deemphasize the role of inflation in our analysis because of the nature of the fers insurance benefit. as shown in table 1, a key attribute of fers insurance is that benefits are partially indexed to inflation. however, the link is imperfect. when inflation (cpi-w) exceeds 2%, annuities receive less than the full cpi-w increase. to account for the probabilities of retirees and their survivors getting less than a full inflation adjustment, we adjust the expected retirement cash flows. we use forecasted data from the philadelphia fed survey of professional forecasters (spf) and historical data from social security cpi-w inflation adjustments from 1975 to the present to simulate prospective fers cost of living adjustments (cola). using a historical resampling simulation that chose from past social security inflation adjustments, we found expected inflation to be 2.7%, resulting in a loss to fers retirees of approximately 0.7% each year. however, the spf forecasts 2.3% inflation for the next 10 years (as of 2017q2). we assume a 2.5% inflation estimate, halfway between the spf projection and the historical social security cola. this results in a 0.5% real loss each year for fers retirees. for our analysis of the effects of early retirement (at age 56), we use 2.3% inflation to compute the real loss in purchasing power for the first six years, until retirees reach age 62, then revert to the 2.5% and 0.5% numbers discussed above; this correctly accounts for the pre-62 absence of a cola and the post-62 partial indexation. 3. simulation model we explore the cost-benefit tradeoff for the fers survivor annuity by computing the internal rate of return associated with the annuity premium cash outflows and the benefit cash inflows received. we construct a monte carlo simulation accounting for the factors in table 1. our simulation illustrates the cost and benefit distributions—and helps describe the characteristics of those distributions. specifically, in addition to finding the implied discount rate associated with the fers, we find: table 2 distribution of spousal ages husband 20� years older 1.0% 15–19 years older 1.6% 10–14 years older 4.9% 6–9 years older 12.2% 4–5 years older 13.4% 2–3 years older 21.7% husband and wife within 1 year 32.2% wife 2–3 years older 6.4% 4–5 years older 2.9% 6–9 years older 2.6% 10–14 years older 0.9% 15–19 years older 0.2% 20� years older 0.2% based on census table fg3 (census website, 2002; per 100k marriages). the data is collected based on age as of last birthday; hence, spouses listed as 23 and 21 are considered “2–3 years” apart even though they could be less than 13 months apart. 137k. davis et al. / financial services review 27 (2018) 133-146 1. the distributions describing fers insurance premiums payments; 2. the distributions describing fers insurance benefit payments; 3. the descriptive statistics about durations of premium and benefit payments; 4. the percentage of participants who earn at least the implied discount rate. 3.1. method description as described in davis and fraser (2012), jennings and reichenstein (2003), and stoller (1992), we compute the expected future cash flows of beneficiaries. this approach does not use the projected direct cash flows paid to the beneficiary. rather, we decrement the projected cash flows for the actuarial probability that the beneficiary is alive at any given age.2 our randomly generated sample size is 500,000 to achieve stable results from one simulation run to another. because the distributions produced depend on retiree and spouse ages and genders, we focus on three scenarios: 1. a 62-year-old male retiree with a 62-year-old female spouse; 2. a 62-year-old female retiree with a 62-year-old male spouse; and 3. a 62-year-old female retiree with a 65-year-old male spouse. the first two scenarios allow a direct comparison between male and female retirees; the third scenario maps well to current american marriage realities, where the average retirement-age spouse is three years older than his spouse as seen in table 2. table 3 depicts a summary of results and includes additional scenarios to provide context for age sensitivity within the simulation. 4. results 4.1. case 1: 62-year-old male retiree and 62-year-old female spouse we first describe the number of years spouses outlive retirees; we show that distribution in fig. 2. fig. 2 illustrates that for case 1, the widow outlives the male retiree 58.9% of the time, which implies that 41.1% of those paying fers survivor insurance premiums receive no benefit.3 on average, for case 1, a widow will live 2.89 years longer than the male retiree. to construct the expected benefits from the survivor annuity, we extract from fig. 2 those instances where benefits are paid: the average spouse who outlives her spouse does so by an additional 12 years. when the 41.1% of spouses who collect no benefits are included, the average for all widows is still 6.89 years of survivor benefits. next, we use the age at which the retiree dies and the age at which the spouse dies to construct a distribution showing how many male retirees of the starting 500,000 are still paying premiums, by number of years since retirement. we also construct a distribution showing the timing of survivor annuities. fig. 3 shows these two distributions. the results of the simulation show that the average male retiree pays premiums for 15.2 years. fig. 3 also shows that the benefits from the fers survivor annuities occur much later than the costs. 138 k. davis et al. / financial services review 27 (2018) 133-146 fig. 3 compares the costs and benefits of fers annuity insurance, but the areas under each curve represent retiree or survivor counts and not dollar amounts. hence, they do not provide a sense of the relative costs and benefits. we generate the estimated costs by summing participants’ premiums paid in our distribution and aggregate estimated benefits. fig. 4 shows the simulated total real dollar value paid in by retirees each year and the total dollar amount paid out by the government. the starting point is 500,000 62-year-old male retirees who have 62-year-old female spouses. fig. 4 shows the area under the benefits curve is larger than the area described by the costs curve. since both cash flows are inflation-adjusted, we conclude there is substantial aggregate real return here. to be more specific we calculate the internal rate of return. for case 1 the computed rate is 7.4%. as an after-inflation (real) return, this is remarkable for a government-guaranteed contract. according to siegel (2014), real returns from the table 3 summary of results implied discount rate percent of annuitants at or above implied discount rate average number of years receiving survivor annuity percent chance spouse outlives retiree case 1 7.6% 30.0% 6.89 58.9% 62 y/o male retiree 62 y/o female spouse case 2 1.9% 27.3% 3.98 41.1% 62 y/o female retiree 62 y/o male spouse case 3 �0% na 3.00 34.1% 62 y/o female retiree 65 y/o male spouse additional cases case 4 10.3% 29.8% 6.29 58.6% 66 y/o male retiree 66 y/o female spouse case 5 3.6% 27.0% 3.69 41.4% 66 y/o female retiree 66 y/o male spouse case 6 4.8% 30.8% 7.71 59.3% 56 y/o male retiree 56 y/o female spouse case 7 �0% 27.3% 4.37 40.7% 56 y/o female retiree 56 y/o male spouse case 8 8.65% 30.9% 8.60 65.9% 62 y/o male retiree 59 y/o female spouse case 9 3.5% 29.2% 5.21 48.4% 62 y/o female retiree 59 y/o male spouse case 10 62 y/o male retiree 6.25% 29.4% 5.39 51.4% 65 y/o female spouse 139k. davis et al. / financial services review 27 (2018) 133-146 stock market over the last 86 years have been approximately 6.4%. in this case, the real return of the low-risk fers annuity benefit exceeds the (risky) historical real return from stocks.4 fig. 2. number of years spouse outlives retiree (500,000 simulations, each retiree was 62-year-old male with 62-year-old spouse). fig. 3. timing of benefits and premiums by number of retirees or survivors (based on a starting point of 100,000 survivor benefit plan [sbp] insured retirees; each retiree is 62-year-old male with 62-year-old spouse). 140 k. davis et al. / financial services review 27 (2018) 133-146 4.2. case 2: 62-year-old female retiree, 62-year-old male spouse fig. 5 replicates fig. 4 with the genders of the retiree and spouse reversed. in fig. 5, the difference in the areas under the two curves is much less dramatic than in fig. 4. with genders reversed, the female retiree is expected to outlive her spouse approximately 58.9% of the time, and the implied real rate of return falls to a relatively meager 1.7%.5 fig. 4. timing and amounts of federal employees retirement system (fers) insurance costs and benefits (62-year-old male retiree with 62-year-old female spouse). fig. 5. timing and amounts of federal employees retirement system (fers) insurance costs and benefits (62-year-old female retiree with 62-year-old male spouse). 141k. davis et al. / financial services review 27 (2018) 133-146 4.3 case 3: 62-year-old female retiree and 65-year-old male spouse for female retirees, the potential benefits from fers annuity further diminish when one considers the typical age difference between spouses. table 2 shows census figures for the age differences between spouses. this data, which applies to the u.s. population, implies a typical age difference just under three years. for an average couple, the male tends to be almost three years older. unsurprisingly, this has important implications. fig. 6 displays how the cash flow streams change from fig. 5. the age difference reduces the insurance premiums paid because these payments end upon the death of the older male spouse. however, the biggest change is in the decline in the expected benefit payout. as a result, the peak payouts in fig. 6 are substantially lower than in the same-age fig. 5. here, the younger female retiree is now expected to outlive her spouse approximately 65.9% of the time. for younger-spouse case 3, the computed internal rate of return is negative (�0.3%). absent personal health or trust considerations, for a 62-year-old female retiree, the typical three-year difference appears to define an important breakpoint in terms of whether female retirees should opt for fers annuity insurance. table 3 summarizes each of the three cases examined here in detail. extra cases provide additional insight. to better illustrate the relationship of spousal ages and genders on the implied discount rate, we graph six cases in fig. 7. fig. 7 shows, as expected, the younger the spouse, the higher the implied return for a 62-year-old fers retiree. this is true for both male and female workers. cases 4–7 of table 3 show that the implied discount rate increases for same-age couples older than the 62 years used in case 1. again, this is as expected. the results here contrast with the findings of davis and fraser (2012). in their analysis of the military survivor benefit, the computed returns were all positive. the primary driver for fig. 6. timing and amounts of federal employees retirement system (fers) insurance costs and benefits (62-year-old female retiree; 65 year-old-male spouse). 142 k. davis et al. / financial services review 27 (2018) 133-146 the difference in the results is the nature of the specific retirement programs, both in the timing and magnitude of cash flows. military personnel can retire and start receiving an immediate benefit after 20 years of service, even if younger than 40. this contrasts with the base case here of a 62-year old fers retiree. additionally, military retirees pay a 6% premium to receive 55% of the retirement benefit, which is both cheaper and more generous. 5. implications for retirees and planners our study has important implications for retirees and financial planners. our numeric and graphical representations help frame the fers annuity insurance decision. our finding of gender differentials is striking; while the survivor benefit is lucrative for couples with a male fers retiree, it is meaningfully less rewarding (perhaps even wealth destroying) for couples with a female retiree. if couples elect or reject the survivor benefit based on some threshold return level, our analysis has financial planning implications. the fers survivor annuity decision also has portfolio ramifications. for example, the decision to take survivor benefits directly impacts retiree insurance and asset allocation decisions. finally, our results have economic significance. the congressional budget office estimates there will be more than 1 million fers retirees by 2022, paying approximately $2 billion a year for fers annuity insurance (cbo, 2012). the analysis presented here can aid planners who must consider the myriad of payout options associated with defined-benefit plans. however, our results are a starting point. while homo economicus might rationally reject negative rates of return or rates below a market-based threshold, fers retirees who are more risk averse might accept a lower internal rate of return to avoid worst case scenarios. prospect theory (kahneman and tversky, 1979) tells us this is not an unusual reaction. we also know that the elderly have high discount rates (see huffman, mauer, and mitchell, 2016). further, some fers retirees have marital dynamics that prevent opting out (that might be viewed as a specific spousal form of risk aversion). fig. 7. sensitivity of implied discount rate by gender and age of spouse (62-year-old retiree). 143k. davis et al. / financial services review 27 (2018) 133-146 personal considerations are critical in any retirement decision. significant medical conditions can trump the general results demonstrated here. furthermore, in rare cases where the retiree has very young or disabled children, a special program covering such insurable interests could change the general situation described here. additional research is needed to construct scenarios that address subsegments of the married population. for example, in using social security tables, there is an assumption that deaths are independent. however, research suggests that spouses influence each other’s longevity. drefahl (2010) and neiman and dortmann (2010) found that married men generally live longer than single men. further, the death of the first spouse impacts the life expectancy of the surviving spouse (elwert and christakis, 2008). building these interdependencies into the insurance decision analysis would substantially improve the discussion.6 another avenue for further research on subpopulations concerns the difference in life expectancies based on socio-economic status. olshansky et al. (2012) showed that retiree life expectancy is strongly impacted by educational background and other factors. our research has important policy implications. the gender differences noted are striking. the current system defaults all retirees, male and female, into the survivor insurance program. a reasonable policy response might be to align the default-enrollment “nudge” (thaler and sunstein, 2009) with the on-average gender-specific economics. alternatively, policy makers could adjust the pricing scheme (a 10% reduction for a 50% benefit) to produce better outcomes for female retirees. notes 1 social security administration. period life table (2009). actuarial publications. see: http://www.ssa.gov/oact/stats/table4c6.html. 2 jennings and reichenstein (2003) note that the expected cash flow approach results in about 10% more value than evaluating cash flows over life expectancy. 3 the fers survivor annuity program will cover a new spouse, but we take the more conservative valuation approach of ignoring re-marriages. 4 however, because of the skewness in the benefits distribution, slightly less than 30% of fers annuity insurance participants will earn the 7.4% real return on their payments. 5 though this 1.7% real return is still generous relative to the current real return on another government-guaranteed inflation-indexed investment, treasury inflation protection securities. 6 note that health interdependencies are not amenable to universal simulation. for example, while the hazard ratio (likelihood of death) generally increases for widows and widowers (elwert and christakis, 2008; martikainen and valkonen, 1996), the size of the effect depends on the cause of death for the first spouse. many cancers, for example, have little effect on the mortality probabilities for the surviving spouse. lung cancer is an exception—however, causes like lung cancer introduce another problem—death of a spouse from lung cancer probably only “increases” the surviving spouse’s hazard ratio because both spouses were smokers. these interactions are best 144 k. davis et al. / financial services review 27 (2018) 133-146 handled on a case-by-case basis. the importance of having an advisor address death interdependencies is discussed by davis and fraser (2012). by applying the likelihoods discussed by elwert and christakis (2008), davis and fraser found that an early death of the retiree (e.g. from 50 to 60 years of age) can result in the survivor living as much as a full year less than predicted by the social security tables. acknowledgments opinions, conclusions, and recommendations expressed or implied within are solely those of the authors and do not necessarily represent the views of the usaf academy, the u. s. air force, the department of defense, or any other government agency. the authors would like to thank two anonymous reviewers for their valuable contributions. references ameriks, j., veres, r., & warshawsky, m. j. (2001). making retirement income last a lifetime. journal of financial planning, 14, 60–76. burr, j. (2018). opm is out to cut retirement benefits. (available at https://www.federaltimes.com/management/ pay-benefits/2018/05/07/opm-is-out-to-cut-retirement-benefits/) burrelli, d. f. (2011). the military survivor benefit plan: a description of its provisions. congressional research service. (available at https://digital.library.unt.edu/ark:/67531/metadc122283/m1/1/high_res_d/ rl31664_2011jan12.pdf) census bureau. (2002). married couple family groups, by presence of own children under 18, and age, earnings, education, and race and hispanic origin of both spouses. census website, table fg3. (available at http://www.census.gov/population/socdemo/hh-fam/cps2002/tabfg3.pdf) cbo. (2012). civil service retirement and disability fund, march 2012 baseline. (available at http://www. cbo.gov/publication/43067) cbo. (2016). civil service annual report. (available at https://www.opm.gov/about-us/budget-performance/ other-reports/2015-civil-service-retirement-and-disability-fund-annual-report.pdf) davidson, j. (2017). trump’s budget calls for hits on federal employee retirement programs. the washington post, may 18, 2017. davis, k. j., & fraser, s. (2012). insuring defined-benefit plan value: an examination of the survivor benefit plan (sbp) decision. financial services review, 21(4), 275–290. drefahl, s. (2010). how does the age gap between partners affect their survival? demography, 47, 313–326. elwert, f., & christakis, n. (2008). the effect of widowhood on mortality by the causes of death of both spouses. american journal of public health, 98, 2092–2098. hicks, j. (2013). gop plan would end defined-benefit pension for future feds. washington post. (available at http://www.washingtonpost.com/blogs/federal-eye/wp/2013/11/15/gop-plan-would-end-defined-benefitpension-for-future-feds/) huffman, d., maurer, r., & mitchell, o. s. (2016). time discounting and economic decision-making among the elderly. michigan retirement research center working paper 2016–347. isaacs, k. p. (2010). cost-of-living adjustments for federal civil service annuities: 94–834. congressional research service: report, 1–4. jennings, w. w., merrell, j. c., o’malley, t. c., & payne, b. c. (2018). a framework for analyzing defined benefit pension insurance: the survivor benefit plan for veterans. finacial services review, 27, xxx-xxx. jennings, w. w., & reichenstein, w. r. (2001). the value of retirement income streams: the value of military retirement. financial services review, 10, 19–35. 145k. davis et al. / financial services review 27 (2018) 133-146 jennings, w. w., & reichenstein, w. r. (2003). valuing defined-benefit plans. financial services review, 12, 179–199. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica 47(2), 263–292. martikainen, p., & volkonen, t. (1996). mortality after the death of a spouse: rates and causes of death in a large finnish cohort. american journal of public health, 86, 1087–1093. matthews, j. l., & berman, d. m. (2008). social security, medicare & government pensions (13th ed.). berkeley, ca: nolo. milevsky, m. a. (2006). the calculus of retirement income: financial models for pension annuities and life insurance. new york, ny: cambridge university press. neimann, s., dortmund, t., & james, j. (2010). honey, why don’t you see a doctor? proceedings, royal economic society annual conference. olshansky, s. j., antonucci, t., berkman, l., binstock, r. h., boersch-supan, a., cacioppo, j. t., carnes, b. a., carstensen, l. l., fried, l. p., goldman, d. p., jackson, j., kohli, m., rother, j., zheng, y., & rowe, j. (2012). differences in life expectancy due to race and educational differences are widening. health affairs, 31, 1803–1813. purcell, p. j. (2009). survivor benefits for families of civilian federal employees and retirees. journal of pension planning and compliance, 34, 69–77. siegel, j. j. (2014). stocks for the long run (5th ed.). new york, ny: mcgraw-hill. stoller, m. a. (1992). estimating the present value of pensions: why different estimators get varying results. journal of legal economics, 2, 49–61. thaler, r. h., & sunstein, c. r. (2009). nudge: improving decisions about health, wealth and happiness. new york, ny: penguin books. 146 k. davis et al. / financial services review 27 (2018) 133-146 why are women less motivated to become financially literate? jaclyn j. beierleina,*, kaleigh launsbyb, haley smith forbesb adepartment of finance, east carolina university, college of business, 3133 bate building, greenville, nc 27858, usa bhonors college and department of finance, east carolina university, 3133 bate building greenville, nc 27858, usa abstract research suggests women know and care less about personal finance than men but offers no explanations. women in our sample score lower on basic finance questions and report lower motivation to learn personal finance. men and women report higher motivation when they say finance is “very important.” however, women who expect to make decisions with a spouse report lower motivation than do women who expect to make decisions alone. men’s motivation does not vary with such expectations. our results suggest women are less motivated to become financially literate when they lack confidence and when they expect to share financial responsibilities. © 2022 academy of financial services. all rights reserved. jel classifications: d14; g530; a22 keywords: personal finance; financial literacy 1. introduction prior research indicates that women are less knowledgeable about and less interested in personal finance than men and less likely to invest in the stock market. why? do they think finance less important than men do? do they feel less confident in their abilities to understand it? are they more likely to expect that their parents or spouses will help them make financial decisions? *corresponding author: tel.: +1-252-737-1592; fax: +1-252-737-1578. e-mail address: beierleinj@ecu.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 89–105 the women in our sample score lower than the men on average on a series of questions designed to test basic financial knowledge. the women also report lower motivation levels on average. they do not rate it as less important, but they are less confident in their answers to our questions and in their ability to make future financial decisions. statistically, the women and men in our sample have similar expectations regarding who will make their future financial decisions. about 30% of women and men chose the response “i will” and about 70% of women and men chose “my spouse and i will.” no one chose either remaining response “my spouse will” or “my parents will.” however, women who report that they expect to make financial decisions with a spouse also report lower motivation to learn personal finance than do women who expect to make financial decisions alone. men’s motivation levels do not appear to vary with their expectations regarding whether they will make decisions alone or with a spouse. if our sample results are representative of the larger female population, the high expectation among women that they will have a spouse with whom to make financial decisions coupled with the dampening effect that expectation appears to have on women’s motivation to learn about finance could help explain why in study after study women tend to know less and care less about personal finance. 2. literature review much empirical evidence exists that financial literacy is relatively low, particularly among women, and that low financial literacy reduces financial well-being (see, e.g., lusardi & mitchell, 2007; mandell, 2006; the national council on economic education, 2005; tzu-chin, bartholomae, fox, & cravener, 2007; van rooij, lusardi, & alessie, 2011; xiao, serido, & shim, 2010). despite evidence that college students and college graduates are more knowledgeable than other high school graduates (mandell & klein, 2009), the average college student demonstrates relatively little financial knowledge, and the average female student is less knowledgeable than the average male student (volpe, chen, & pavlicko, 1996). chen and volpe (1998) find that female college students and students who are not business majors, are in the lower class ranks, under age 30, and have little work experience have lower levels of financial knowledge, and that students who display less financial knowledge are more likely to assign little importance to financial matters and to make incorrect financial decisions. chen and volpe (2002) find that female college students are less knowledgeable than their male counterparts on personal finance topics, that they exhibit less enthusiasm, lower confidence, and less willingness to learn about personal finance topics than male students, and that they are less likely to rank personal finance, math, economics, and science courses as important. mandell and klein (2007) focus on the importance of one’s motivation to be financially literate. using data from the national jump$tart survey of high school seniors, they find that students who believe financial difficulty results from poor decisions, that it is important to have enough money to pay bills, and that social security alone provides insufficient retirement income tend to earn higher scores on the financial literacy questions in the survey. level of aspiration is another important determinant of financial literacy in their data. students bound for a four-year college, a professional job, or a higher starting salary earned higher scores, on average, than students with lower aspirations. 90 j. j. beierlein et al. / financial services review 30 (2022) 89–105 beierlein and neverett (2013) find that female undergraduates are less likely than male undergraduates to take an elective personal finance course. tang and peter (2015) examine how young adults acquire financial knowledge and find that their financial education and experience and the financial experience of their parents are important factors. tokar asaad (2015) finds that financial confidence is also an important component of literacy. harrington and smith (2017) find that students are more interested in financial education when they perceive a higher return from financial education and when they are financially independent. in summary, the literature tells us that financial education and experience, the financial experience of one’s parents, confidence, interest, motivation, and educational and professional aspirations are important determinants of financial literacy and that, compared with men, women tend to be less knowledgeable, less confident, and less motivated to learn personal finance. yet, extant literature offers little insight into why women may be less motivated to become financially literate. this is the question we explore in our research. expectancy theory of motivation, notably vroom’s (1964) model of motivational force, indicates that motivation is determined by expectancy, instrumentality, and valence. expectancy is the belief that an action will result in an outcome. instrumentality is the belief that the outcome will lead to an appropriate reward, and also perhaps that the outcome is necessary to earn the reward. finally, valence is the importance placed on the reward. thus, in our context, one’s motivation to learn personal finance depends on how much one believes that one’s efforts to learn will improve understanding of personal finance, how much one expects to rely on one’s own knowledge to make good financial decisions, and how important one believes it is to make good financial decisions. as we are specifically interested in women’s motivation, we take note of claudia goldin’s (2006) address to the american economic association, “the quiet revolution that transformed women’s employment, education, and family.” in it, she explains that as women’s estimates of the time they would spend in the labor force increased, their investments in education and interest in more professional majors also increased. therefore, it seems reasonable to expect that when a woman believes that she will be responsible for making financial decisions, she will be more motivated to become financially literate. the literature reviewed above suggests several hypotheses to be tested: 1. the women in our sample are less knowledgeable about personal finance, less confident in their abilities related to finance, and less motivated to learn it than the men. 2. motivation to learn personal finance is positively related to one’s confidence (expectancy), whether one expects to be responsible for financial decision-making (instrumentality), and the importance one assigns to personal finance decisions (valence). 3. women are less motivated to learn personal finance because they are less confident, they think it is less important, or they expect to have less responsibility for making financial decisions. 3. methodology and data the link to an online qualtrics survey approved by the university’s institutional review board was emailed to undergraduate students at east carolina university (ecu). additionally, the link was posted on social media pages open only to students, and flyers were hung around j. j. beierlein et al. / financial services review 30 (2022) 89–105 91 campus. emails and other advertising for the survey noted that respondents would be entered in a drawing for gift cards to increase interest. the survey is composed of 64 questions. students could exit the survey at any time. we analyze only completed surveys. the entire survey and answer choices are in the appendix. the first 11 survey questions address demographics, including verifying that the respondent was an undergraduate at east carolina and asking for age, gender, class (e.g., freshman), race, employment status, parents’ income range, and whether either parent works in the financial services industry. the next seven questions address students’ awareness of the personal finance course offered at ecu, whether they had taken it or any other finance or economics courses or attended any related seminars and if so, where, and whether they planned to take the personal finance course or attend a related seminar. there are also four follow-up questions that appeared only for those who responded that they had taken the course. these address why they took the course, how it affected their opinions, whether they expected to use what they learned in the future, and the grade earned. for those who responded that they had not taken it, one follow-up question asks why not, and another asks if they plan to take it. the next question asks, “how important do you think personal finance is?” and then a free response follow-up question depending on the answer that asks “why do you find personal finance (un)important?” the next question asks, “how motivated are you to learn about personal finance?” then, there are 16 questions that ask how strongly the following had influenced the respondent’s knowledge of personal finance and their financial decisions: parents, other relatives, instructors, peers, media, online resources, own experiences, and religion. the next question asks respondents to indicate where they might go for advice in the future when they had questions regarding personal finance. then, a series of eight questions asks whether they had heard of or currently had the following financial products: a credit card, renters’ insurance, a retirement account, and other types of investment accounts. the next question asks if they planned to invest in the stock market, and why or why not. the next seven questions are financial test questions adapted from the oecd/international network on financial education pilot study (atkinson & messy, 2012).1 these questions cover division of money, time value of money, inflation, interest (two questions), risk and return, and diversification. they do not require a calculator to complete and are either multiple choice or true/false. for the multiple-choice questions involving simple math, the choice “i don’t know” is an option in addition to three numerical options. the next two questions asked how confident they were in their answers to the financial literacy questions and in their ability to make financial decisions in the future. the final question asked, “in the future, who do you expect to make your financial decisions?” some of the questions described above were of interest to us beyond our hypotheses regarding women’s motivation to become financially literate. the questions relevant to our hypotheses include several of the demographic questions, the financial test questions, the confidence, motivation, and importance questions, and the question regarding who would make their future financial decisions. the distributed survey collected 194 responses from undergraduate students at ecu. of these, 176 participants finished the survey. this is approximately 0.8% of the university’s undergraduate population. descriptive statistics are in table 1. age ranges from 18 to 23 and averages 19.49years. the majority (70%) of the respondents were female. twenty-five percent of the students identified themselves as freshmen, 24% as sophomores, 31% as juniors, and 20% as seniors. a large number of respondents (53%) indicated they were honors college students, who may have been more likely to attend to the email 92 j. j. beierlein et al. / financial services review 30 (2022) 89–105 because the survey was distributed by honors college students for their thesis research. as shown in table 2, compared with the university, our sample contains a higher proportion of females, asians, hispanics, and business majors, and a lower proportion of african americans. compared with 2019 u.s. population estimates, our sample contains a higher proportion of females and whites, and a lower proportion of african americans, asians, and hispanics. 4. analysis and results to address our first hypothesis, we consider whether the patterns found so often in previous research persist in our sample. do our female respondents exhibit less financial literacy, table 1 descriptive statistics variables n mean median standard deviation min. max. panel a: means and medians age 176 19.49 19.50 1.19 18 23 total correct 176 5.70 6.00 1.15 2 7 allied health 176 0.05 0.00 0.21 0 1 business 176 0.15 0.00 0.36 0 1 education 176 0.03 0.00 0.17 0 1 engineering 176 0.07 0.00 0.26 0 1 fine arts 176 0.06 0.00 0.23 0 1 health 176 0.19 0.00 0.39 0 1 nursing 176 0.18 0.00 0.38 0 1 arts and sciences 176 0.27 0.00 0.45 0 1 male 176 0.30 0.00 0.46 0 1 white 176 0.81 1.00 0.40 0 1 african american 176 0.07 0.00 0.26 0 1 hispanic 176 0.03 0.00 0.18 0 1 asian 176 0.04 0.00 0.20 0 1 other race 176 0.05 0.00 0.21 0 1 work during school year 176 0.56 1.00 0.50 0 1 parents in finance 176 0.22 0.00 0.42 0 1 importance of finance 176 0.79 1.00 0.41 0 1 finance education 176 0.80 1.00 0.40 0 1 honors college 176 0.53 1.00 0.50 0 1 panel b: frequencies (%) 0 1 2 3 4 5 class 176 25.00 23.86 31.25 19.89 parents’ income 176 14.77 6.25 9.66 18.75 25.57 25.00 motivation 176 17.61 48.30 34.09 note. descriptive statistics for the full sample of 176 undergraduate students who responded to a survey invitation and completed the survey. indicator variables are coded 1 if respondent selected that attribute, 0 otherwise. means and frequencies for the indicator variables represent the percentage of respondents who selected that attribute. importance of finance is coded 0 if respondents selected “important” and 1 if “very important.” no other choices were selected. finance education is coded 1 if respondents indicated any prior instruction in finance or economics, 0 otherwise. class rank is coded 1 for freshman, 2 for sophomores, 3 for juniors, and 4 for seniors. parents’ income is coded 0 if respondent chose “unsure,” 1 if <$30,000, 2 if $30,000-$59,000, 3 if $60,000$89,000, 4 if $90,000-$119,000, and 5 if >$120,000. motivation is coded 0 if respondent chose “neither motivated nor unmotivated,” 1 if “somewhat motivated,” and 2 if “very motivated.” no other choices were selected. j. j. beierlein et al. / financial services review 30 (2022) 89–105 93 confidence, and interest in personal finance than the men do? using t-tests, we compare the women’s and men’s scores on our financial literacy test questions and their reported levels of motivation to learn about personal finance. each respondent’s score is number of correct answers, which could range from 0 to 7. the respondent’s motivation level is coded as 0 for “neither motivated or discouraged,” 1 for “slightly motivated,” and 2 for “very motivated.” none of the survey respondents chose either of the remaining two choices for this question: “slightly discouraged” or “strongly discouraged.” results in table 3, panel a suggest that our female respondents are less knowledgeable about and less motivated to learn personal finance. on average, the women score about half a point lower than the men on our seven finance questions, scored at one point each. this difference is significant at less than 1%. women were less likely to correctly answer questions about the time value of money, interest, saving, diversification, and risk and return, but equally likely to correctly answer questions about division of money and the definition of inflation. when asked, “how motivated are you to learn about personal finance?”, a higher percentage of women than men said, “neither motivated or discouraged” or “slightly motivated,” and a higher percentage of men than women said, “very motivated.” women’s mean motivation values are 26% lower than men’s, on average, and this difference is significant at less than 1%. in panel b of table 3, we examine why women might be less motivated to learn finance. we ask how confident they are in their test answers and in their abilities to make future financial decisions, who they expect to make their financial decisions in the future, and how important finance is. using a likert scale from 1 to 5 to indicate confidence levels from not at all confident to very confident, respondents rated their confidence in their answers on our finance questions and their confidence in their abilities to make financial decisions in the future. in both cases, women reported significantly lower confidence levels than men. when we asked, “in the future, who do you expect to make your financial decisions?” more than 60% of male and female respondents said, “my spouse and i together” and the remainder said, “i will.” no one chose the other options, “my parents,” or “my spouse.” a slightly but insignificantly higher percentage of women than men said, “my spouse and i” while the reverse is true of “i will.” more than 70% of respondents said personal finance is “very important” and the remaining respondents said, “important.” no one chose the other options, “neither important or unimportant,” “unimportant,” or “very unimportant.” we assigned a table 2 comparison to university demographic data variables university data our sample u.s. population female 61% 70% 51% african american 14% 7% 13% asian 3% 4% 6% hispanic 2% 3% 19% white 80% 79% 60% business 9% 14% n/a note. comparison of university demographic data to that of full sample of respondents who finished the survey. shows percentages of students/respondents who selected that attribute on enrollment materials/in the survey. u.s. demographic estimates dated july 1, 2019 are available at https://www.census.gov/quickfacts/fact/ table/us/pst045219. 94 j. j. beierlein et al. / financial services review 30 (2022) 89–105 t ab le 3 d if fe re n ce s in m ea n s b y re p o rt ed g en d er 0 = fe m al e 1 = m al e n m ea n s ta n d ar d d ev ia ti o n t s ig . (t w o ta il ed ) m ea n d if fe re n ce f m s ta n d ar d er ro r d if fe re n ce p an el a : a re w o m en le ss k n o w le d g ea b le an d m o ti v at ed ? t o ta l co rr ec t 0 1 2 4 5 .5 4 1 .1 9 9 �3 .1 6 1 0 .0 0 2 �0 .5 3 7 0 .1 7 0 1 5 2 6 .0 8 0 .9 4 7 in fl at io n 0 1 2 4 0 .4 2 0 .4 9 5 �1 .9 2 1 0 .0 5 6 �0 .1 5 8 0 .0 8 2 1 5 2 0 .5 8 0 .4 9 9 in te re st 0 1 2 4 0 .9 7 0 .1 7 7 �2 .0 2 5 0 .0 4 5 �0 .0 3 2 0 .0 1 6 1 5 2 1 .0 0 0 .0 0 0 s av in g 0 1 2 4 0 .8 9 0 .3 1 8 �2 .7 2 2 0 .0 0 7 �0 .0 9 4 0 .0 3 4 1 5 2 0 .9 8 0 .1 3 9 d iv er si fi ca ti o n 0 1 2 4 0 .6 7 0 .4 7 2 �1 .9 8 7 0 .0 4 9 �0 .1 3 8 0 .0 7 0 1 5 2 0 .8 1 0 .3 9 8 r is k an d re tu rn 0 1 2 4 0 .7 7 0 .4 2 0 �4 .0 4 4 0 .0 0 0 �0 .1 8 7 0 .0 4 6 1 5 2 0 .9 6 0 .1 9 4 d iv is io n o f m o n ey 0 1 2 4 0 .9 5 0 .2 1 5 0 .2 5 4 0 .8 0 0 0 .0 0 9 0 .0 3 7 1 5 2 0 .9 4 0 .2 3 5 in fl at io n d efi n it io n 0 1 2 4 0 .8 7 0 .3 3 7 1 .0 7 7 0 .2 8 3 0 .0 6 3 0 .0 5 9 1 5 2 0 .8 1 0 .3 9 8 m o ti v at io n 0 1 2 4 1 .0 6 0 .6 7 8 �3 .2 4 7 0 .0 0 1 �0 .3 6 7 0 .1 1 3 1 5 2 1 .4 2 0 .6 9 6 (c o n ti n u ed o n n ex t p a g e) j. j. beierlein et al. / financial services review 30 (2022) 89–105 95 t ab le 3 (c o n ti n u ed ) 0 = fe m al e 1 = m al e n m ea n s ta n d ar d d ev ia ti o n t s ig . (t w o ta il ed ) m ea n d if fe re n ce f m s ta n d ar d er ro r d if fe re n ce p an el b : w h y ar e w o m en le ss k n o w le d g ea b le an d m o ti v at ed ? im p o rt an ce 0 1 2 4 0 .7 7 0 .4 2 5 �1 .1 8 7 0 .2 3 7 �0 .0 8 0 0 .0 6 7 1 5 2 0 .8 5 0 .3 6 4 c o n fi d en t an sw er s 0 1 2 4 3 .3 4 0 .8 9 2 �5 .5 6 3 0 .0 0 0 �0 .7 3 8 0 .1 3 3 1 5 2 4 .0 8 0 .7 6 3 c o n fi d en t d ec is io n s 0 1 2 4 3 .5 2 0 .8 2 1 �4 .1 7 0 0 .0 0 0 �0 .5 1 4 0 .1 2 3 1 5 2 4 .0 4 0 .7 1 3 d ec is io n s w it h sp o u se 0 1 2 4 0 .6 9 0 .4 6 3 0 .5 1 3 0 .6 0 8 0 .0 4 0 0 .0 7 7 1 5 2 0 .6 5 0 .4 8 0 d ec is io n s m y se lf 0 1 2 4 0 .3 1 0 .4 6 3 �0 .2 6 6 0 .7 9 1 �0 .0 2 0 0 .0 7 7 1 5 2 0 .3 3 0 .4 7 4 q u es ti o n s p ar en ts 0 1 2 4 0 .9 0 0 .2 9 7 1 .7 5 4 0 .0 8 1 0 .0 9 6 0 .0 5 4 1 5 2 0 .8 1 0 .3 9 8 q u es ti o n s b o o k s 0 1 2 4 0 .1 0 0 .2 9 7 �2 .3 1 3 0 .0 2 4 �0 .1 5 3 0 .0 6 6 1 5 2 0 .2 5 0 .4 3 7 s to ck m ar k et to o ri sk y 0 1 2 4 0 .1 9 0 .3 9 7 5 .4 3 3 0 .0 0 0 0 .1 9 4 0 .0 3 6 1 5 2 0 .0 0 0 .0 0 0 n o te . in d ep en d en t sa m p le tte st s re p o rt ed . t h e sa m p le is 1 7 6 u n d er g ra d u at e st u d en ts w h o re sp o n d ed to a su rv ey in v it at io n an d co m p le te d th e su rv ey . 96 j. j. beierlein et al. / financial services review 30 (2022) 89–105 value of 1 to very important and 0 to important. women’s mean importance value was slightly but not significantly below the men’s mean. additional questions we asked to explore other possible differences between men and women included where they would turn when they had questions about personal finance, whether they were currently invested in or planned to invest in the stock market, and why or why not.2 these questions are motivated by van rooij et al. (2011) and cheng, lin, and liu (2011). van rooij et al find that those with low basic financial literacy are less likely to invest in the stock market and to use newspapers, financial magazines, guides, and books, financial information on the internet, and professional financial advisers and more likely to rely on informal sources of information like family and friends than those with higher literacy. they also find that women score lower on their literacy measure and are less likely to be invested in the stock market but do not otherwise report findings by gender. similarly, cheng et al. find that women are more likely to choose a mortgage lender based on family and friends’ recommendations, while men are more likely to search for lenders offering the lowest rates, with the result that women pay significantly higher mortgage rates, on average, after controlling for other interest rate determinants. in our sample, more than 80% of men and women said they would go to their parents when they had questions about personal finance. women were slightly more likely than men to choose this response, but this difference is only significant at 10%. in contrast, men were significantly more likely than women to say they would look in books when they had questions, but only 25% of men and 10% of women chose this option. when asked about current or future investments in the stock market, 20% of women said no because it is too risky, while none of the men chose that response. this difference is significant at 1%. additional results showing no significant differences were excluded from the table for brevity. about 50% of men and women said they would look online when they have financial questions, 50% said they would consult a financial advisor, 30% said they would ask friends, and 20% said other relatives. no more than 20% of men and women said they were currently invested in stock market, and only about 50% of men and women said they planned to invest in the stock market. overall, our univariate results are consistent with prior literature and our first hypothesis. with respect to personal finance, the women in our sample appear to be less knowledgeable, less motivated to learn, less confident, and more risk averse, on average, than the men. they may be more likely to turn to parents when they have questions and less likely to turn to books than men are. yet, like the majority of men in our sample, the majority of women think finance is very important and expect to make financial decisions with their spouses. does their lower confidence and greater tendency to rely on others when making financial decisions help explain why they are less motivated to learn about personal finance? we turn to multivariate analysis to examine this question, focusing first on our second hypothesis to examine the determinants of motivation. we use ols regression and ordinal regression with the respondents’ chosen level of motivation to learn personal finance as the dependent variable.3 for measures of confidence (expectancy), we use each respondent’s rankings of how confident he or she feels about the answers to the financial knowledge questions and how confident he or she feels about having to make financial decisions in the future. both rankings are on a 1 to 5 scale to indicate “not j. j. beierlein et al. / financial services review 30 (2022) 89–105 97 at all confident” to “very confident,” respectively. for an indicator of whether one expects to be responsible for financial decision-making (instrumentality), we use each respondent’s answer to the question, “in the future, who do you expect to make your financial decisions?” because all respondents answered either “my spouse and i together” or “i will,” we set decisions with spouse to 1 for the former and 0 for the latter answer. for importance, we use an indicator set to 0 if the respondent selected important or to 1 if very important when asked to rate the importance of personal finance. the other answer choices: neither important nor unimportant, unimportant, or very unimportant were not selected by any of the respondents. we also include demographic variables that may affect one’s interest in finance. we include age and whether the respondent works during the school year because older students and those who work may see finance as more relevant to their lives than do younger students or those that do not work. we include parents’ income because it may impact how students have experienced financial behaviors such as budgeting, borrowing, and investing. we include a dummy coded 1 if the respondent reported at least one parent who works in a finance-related field as an additional indicator of the student’s exposure to finance or as an indication that the student has a close relationship with someone well-qualified to offer financial advice. finally, we include gender set to 1 if the respondent self-identified as male, 0 as female to capture any remaining gender-related traits that could be affecting motivation. the results in table 4 are consistent with hypothesis 2. the respondent’s confidence in his or her answers and how highly the respondent rates the importance of finance are significantly and positively associated with increased motivation to learn, while expecting to make decisions with a spouse significantly decreases motivation to learn. motivation is also weakly associated with age. gender is not significantly associated with motivation after controlling for other possible determinants, suggesting that the average differences between men and women that affect motivation to learn about personal finance are captured in the other independent variables. though confidence in one’s answers is a significant determinant of motivation, confidence in one’s ability to make future financial decisions is not. to examine whether the first measure of confidence is masking the influence of the second, we repeated the regression without the confidence in one’s answers variable. both the coefficient and the t-statistic of the second confidence measure increased, but the p-value was still above 10% and the adjusted r-square of the regression fell slightly. in contrast, when we ran the regression with confidence in one’s answers and without confidence in one’s ability to make future decisions, the coefficient and t-statistic of the variable and the adjusted r-square of the regression all increased. perhaps confidence in one’s answers is a better measure of expectancy because it does not require respondents to speculate on their future abilities. alternatively, confidence in one’s future decisions may have a weaker relationship with motivation because it reflects both expectancy and the likelihood of sharing the responsibility for decision-making with a spouse, which tend to have opposite effects on motivation. two potentially important questions we failed to ask in our survey are what income level respondents expected to earn in the future and whether they expected to be financially dependent on a spouse or anyone else. high (low) expected income and independence may simultaneously increase (decrease) motivation to learn finance and decrease (increase) one’s 98 j. j. beierlein et al. / financial services review 30 (2022) 89–105 intention to make decisions with someone else. we cannot rule out this alternative explanation of the negative association we find between motivation and decisions with spouse. to investigate whether the likely determinants of motivation affect women and men differently and test our third hypothesis, we split our sample into female and male subsamples and run the motivation regressions, excluding the gender independent variable, on each subsample. the results in table 5 indicate that, among our female respondents, motivation to table 4 factors associated with motivation b standard error t sig. panel a: ols (constant) �1.250 0.835 �1.497 0.136 gender 0.174 0.110 1.573 0.118 age 0.074 0.041 1.822 0.070 work �0.029 0.095 �0.307 0.759 parents’ income 0.016 0.027 0.583 0.560 parents in finance �0.116 0.115 �1.009 0.315 importance 0.612 0.114 5.368 0.000 confident answers 0.138 0.059 2.320 0.022 confident decisions 0.032 0.063 0.502 0.616 decisions with spouse �0.254 0.102 �2.498 0.013 adjusted r2 0.263 f (sig.) 7.922 (0.000) panel b: ordinal regression estimate standard error wald sig. gender = 0 �0.526 0.370 2.025 0.155 age 0.283 0.138 4.221 0.040 work = 0 0.061 0.317 0.038 0.846 parents’ income 0.066 0.092 0.509 0.475 parents in finance = 0 0.298 0.387 0.591 0.442 importance 1.976 0.410 23.201 0.000 confident answers 0.470 0.200 5.521 0.019 confident decisions 0.181 0.212 0.732 0.392 decisions with spouse = 0 0.775 0.345 5.058 0.025 cox and snell 0.295 x2 (sig.) 61.411 (0.000) note. reports results of ols and ordinal regression with motivation as the dependent variable. motivation is coded 0 if respondent chose “neither motivated nor unmotivated,” 1 if “somewhat motivated,” and 2 if “very motivated.” no other choices were selected. gender is 1 if the respondent self-identified as male, 0 as female. work is coded 1 if respondent works partor full-time during the school year, 0 otherwise. parents’ income is coded 0 if respondent chose “unsure,” 1 if <$30,000, 2 if $30,000-$59,000, 3 if $60,000-$89,000, 4 if $90,000$119,000, and 5 if >$120,000. parents in finance is coded 1 if respondent reports at least one parent who works in a finance-related field. importance of finance is coded 0 if respondent selected “important” and 1 if “very important.” no other choices were selected. confident answers is the respondent’s ranking of how confident he or she felt when answering the financial knowledge questions from 1 to indicate “not at all confident” to 5 to indicate “very confident.” confident decisions is the respondent’s ranking of how confident he or she feels about having to make financial decisions in the future on the same 1 to 5 scale. decisions with spouse is coded 1 if respondent expected to make future financial decisions with a spouse, 0 if respondent expected to make future financial decisions alone. ols coefficient estimates for binary variables estimate the impact on the dependent variable, motivation, when the response is coded 1. ordinal regression coefficient estimates for binary variables estimate the impact on the dependent variable, motivation, when the response is coded 0. j. j. beierlein et al. / financial services review 30 (2022) 89–105 99 learn finance increases weakly with age and confidence and significantly with how highly they rate the importance of finance. it decreases significantly when female respondents have at least one parent who works in a finance-related field and when they expect to make financial decisions with a spouse. in stark contrast, the only significant determinant of our male table 5 factors associated with motivation by reported gender female male b standard error t sig. b standard error t sig. panel a: ols (constant) �1.047 1.044 �1.003 0.318 �1.413 1.385 �1.020 0.313 age 0.079 0.050 1.587 0.115 0.045 0.068 0.669 0.507 work �0.109 0.115 �0.948 0.345 0.261 0.168 1.550 0.128 parents’ income 0.025 0.032 0.783 0.435 �0.037 0.051 �0.726 0.472 parents in finance �0.319 0.148 �2.159 0.033 0.201 0.173 1.159 0.253 importance 0.473 0.130 3.629 0.000 1.063 0.231 4.610 0.000 confident answers 0.125 0.066 1.895 0.061 0.098 0.143 0.689 0.495 confident decisions 0.024 0.071 0.343 0.733 0.157 0.147 1.065 0.293 decisions with spouse �0.369 0.124 �2.969 0.004 �0.096 0.170 �0.564 0.576 adjusted r2 0.208 0.359 f 5.030 4.570 significance 0.000 0.000 panel b: ordinal regression estimate standard error wald sig. estimate standard error wald sig. age 0.315 0.169 3.493 0.062 0.191 0.288 0.440 0.507 work = 0 0.388 0.386 1.008 0.315 �1.403 0.740 3.595 0.058 parents’ income 0.090 0.109 0.692 0.405 �0.201 0.232 0.753 0.385 parents in finance = 0 1.128 0.504 5.004 0.025 �1.405 0.826 2.889 0.089 importance 1.536 0.457 11.316 0.001 4.289 1.199 12.801 0.000 confident answers 0.426 0.223 3.657 0.056 0.443 0.616 0.518 0.472 confident decisions 0.139 0.237 0.343 0.558 0.721 0.656 1.209 0.272 decisions with spouse = 0 1.221 0.426 8.210 0.004 0.181 0.719 0.063 0.802 cox and snell 0.255 0.445 x2 36.549 30.583 significance 0.000 0.000 note. reports results of ols and ordinal regression with motivation as the dependent variable and sample split by reported gender. motivation is coded 0 if respondent chose “neither motivated nor unmotivated,” 1 if “somewhat motivated,” and 2 if “very motivated.” no other choices were selected. work is coded 1 if respondent works partor full-time during the school year, 0 otherwise. parents’ income is coded 0 if respondent chose “unsure”, 1 if <$30,000, 2 if $30,000-$59,000, 3 if $60,000-$89,000, 4 if $90,000-$119,000, and 5 if >$120,000. parents in finance is coded 1 if respondent reports at least one parent who works in a finance-related field. importance of finance is coded 0 if respondent selected “important” and 1 if “very important.” no other choices were selected. confident answers is the respondent’s ranking of how confident he or she felt when answering the financial knowledge questions from 1 to indicate “not at all confident” to 5 to indicate “very confident.” confident decisions is the respondent’s ranking of how confident he or she feels about having to make financial decisions in the future on the same 1 to 5 scale. decisions with spouse is coded 1 if respondent expected to make future financial decisions with a spouse, 0 if respondent expected to make future financial decisions alone. ols coefficient estimates for binary variables estimate the impact on the dependent variable, motivation, when the response is coded 1. ordinal regression coefficient estimates for binary variables estimate the impact on the dependent variable, motivation, when the response is coded 0. 100 j. j. beierlein et al. / financial services review 30 (2022) 89–105 respondents’ motivation to learn finance is how highly they rate the importance of finance. together, the results shown in tables 3 and 5 are partially consistent with hypothesis three. the t-tests indicated that the only significant difference between men and women among the hypothesized determinants of motivation is in confidence. women were less confident in their test answers and in their abilities to make future financial decisions than men were. but women did not rate personal finance as less important. women were not more likely than men to say they would make decisions with their spouse; however, only among the women did the expectation that they would make decisions with their spouse decrease their motivation to learn personal finance. thus, our results suggest that women tend to be less motivated to learn personal finance because they tend to be less confident. their motivation also decreases when they expect to rely on a spouse or a parent to help them make their decisions. 5. conclusion study after study indicates that women, with and without college educations, from high schoolers to retirees, are less financially literate, less interested in personal finance, and less confident in their financial knowledge and decision-making abilities than are men of similar education and age. other studies show that these relative weaknesses in knowledge and interest have negative impacts on women’s financial well-being, particularly with respect to how much retirement savings they accumulate and how much they pay for loans. our results are consistent with our hypotheses derived from expectancy theory of motivation and related literature. the women in our sample of undergraduate students appear to be less knowledgeable, less interested, less confident, and more risk averse, on average, than the men. motivation to learn personal finance increases with confidence and how highly one rates the importance of finance. the majority of both men and women in our sample rated personal finance as very important and said that they expect to make future financial decisions with their spouse. none of the respondents indicated that they expected their parents or spouse to make decisions for them. nevertheless, the women appear to be less motivated to learn about personal finance when they have at least one parent who works in a financerelated field or when they expect to make financial decisions with a spouse. this suggests that today’s young women still expect to rely on parents or spouses when managing household finances. we know of no other recent evidence that this rather old-fashioned notion persists. the principal limitation to our study is our low response rate and subsequently small sample size, particularly with respect to the number of male participants. women and whites are over-represented compared with the general population, and our survey participants are all college students aged 18-23. another important limitation is the potential for measurement error that comes with self-reported data. we rely on our respondents to rate how confident and motivated they are and how important personal finance is. one respondent’s definition of confidence, motivation, or importance may vary from others. similarly, how one determines the magnitudes of these attributes can also vary. notably, if the women in j. j. beierlein et al. / financial services review 30 (2022) 89–105 101 our sample tend to define these attributes differently than men do, or if they are less likely to describe themselves or their beliefs using extreme modifiers, such as very confident or very motivated, this general tendency could explain some of our results in tables 3 and 4. a third limitation is that two key questions rely on respondents’ future expectations. we ask them to rate their “confidence in their abilities to make financial decisions in the future,” and “in the future, who do you expect to make your financial decisions?” finally, we must acknowledge that our respondents are unmarried, financially dependent, college students with no children. their motivation to learn finance could vary greatly from that of others with children, spouses, financial independence, and varying levels of income and education. for example, having children may increase one’s motivation to take personal responsibility for financial decisions regardless of gender, marital status, age, or income. similarly, data in tables 4 and 5 indicate that parents’ income does not affect reported motivation to learn personal finance, but one’s own income could affect motivation and reliance on a spouse. nevertheless, every study must start somewhere, and college is the place where many people start to experience independence and financial responsibility. it may also be the easiest time of one’s life to learn about personal finance, since education and preparing for the future is a key theme of college life. furthermore, our results are consistent with prior literature and with the expectancy theory of motivation. thus, we believe our key results that one’s motivation to learn personal finance varies with how important one rates personal finance and that women tend to be less motivated to become financially literate when they are less confident in their financial knowledge and when they expect to make financial decisions with their spouse will generalize to larger, more representative samples and across education and income levels and ages. further exploration to confirm and expand on our results is warranted. notes 1. atkinson and messy (2012) had eight financial knowledge questions. we chose not to include their compound interest question because they awarded a point for the correct answer only if another question on interest calculation was answered correctly. we also changed three open-ended questions requiring simple math to multiple choice because we felt students would be more likely to answer if choices were provided. 2. the question was, “when you have questions regarding personal finance in the future, where do you think that you will go to for advice? (choose all that apply).” 3. as the motivation rating is an ordinal variable, ordinal regression is likely to be a more appropriate regression technique than ols. see norusis (2012). references atkinson, a., & messy, f. (2012). measuring financial literacy: results of the oecd/international network on financial education (infe) pilot study. oecd working papers on finance, insurance and private pensions, no. 15. paris: oecd publishing. available at https://doi.org/10.1787/5k9csfs90fr4-en 102 j. j. beierlein et al. / financial services review 30 (2022) 89–105 beierlein, j. j., & neverett, m. (2013). who takes personal finance? financial services review, 22, 151–171. chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. https://doi.org/10.1016/s1057-0810(99)80006-7 chen, h., & volpe, r. p. (2002). gender differences in personal finance literacy among college students. financial services review, 11, 289–307. cheng, p., lin, z., & liu, y. (2011). do women pay more for mortgages? journal of real estate finance and economics, 43(4). available at ssrnhttps://ssrn.com/abstract=1942424 goldin, c. (2006). the quiet revolution that transformed women’s employment, education, and family. american economic review, 96, 1–21. https://doi.org/10.1257/000282806777212350 harrington, c., & smith, w. (2017). college student interest in personal finance education. available at ssrn https://ssrn.com/abstract=2782788 lusardi, a., & mitchell, o. s. (2007). financial literacy and retirement preparedness: evidence and implications for financial education programs. business economics, 42(january), 35–44. https://doi.org/0.2145/20070104 mandell, l. (2006). financial literacy: improving education results of the 2006 national jump$tart survey. washington, dc: jumpstart coalition. mandell, l., & klein, l. s. (2007). motivation and financial literacy. financial services review, 16, 105–116. mandell, l., & klein, l. s. (2009). the impact of financial literacy education on subsequent financial behavior. journal of financial counseling & planning, 20, 15–24. norusis, m. (2012). ibm spss statistics 19 advanced statistical procedures companion, chapter 4: ordinal regression. upper saddle, nj: pearson. available at http://www.norusis.com/pdf/aspc_v13.pdf tang, n., & peter, p. c. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience. financial services review, 24, 119–137. the national council on economic education. (2005). what american teens and adults know about economics. mimeo. tokar asaad, c. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101–117. tzu-chin, m. p., bartholomae, s., fox, j. j., & cravener, g. (2007). the impact of personal finance education delivered in high school and college courses. journal of family and economic issues, 28, 265–284. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. https://doi.org/10.1016/j.jfineco.2011.03.006 volpe, r. p., chen, h., & pavlicko, j. j. (1996). personal investment literacy among college students: a survey. financial practice and education, 6, 86–94. vroom, v. h. (1964).work and motivation. new york, ny: john wiley & sons. xiao, j. j., serido, j., & shim, s. (2010). financial education, financial knowledge, and risky credit behavior of college students. working paper 2010-wp-05. networks financial institute. available at http://ssrn.com/ abstract=1709039 appendix appendix: survey the following survey questions were administered electronically via qualtrics. open-ended questions are indicated by open following the question. otherwise, answer choices are italicized following the question. 1. what is your pirate (school) id? open 2. are you an undergraduate student at east carolina university? yes, no 3. what college or school do you fall into? engineering and technology, arts and sciences, nursing, business, health and human performance, allied health sciences, education, fine arts and communication, other 4. what is your age? open 5. what is your class rank? freshman, sophomore, junior, senior j. j. beierlein et al. / financial services review 30 (2022) 89–105 103 6. what is your gender?male, female, other 7. what is your race? asian, white/caucasian, african american, hispanic, pacific islander, other 8. what is your employment status? i do not work, i work part time, i work full time. 9. what is your/your parents total household income? not sure, less than $30,000, $30,000 to $59,999, $60,000 to $89,999, $90,000 to $119,999, $120,000 or more. 10. do either of your parents work in the financial services industry? no; not anymore, but one or both used to; yes, one of my parents; yes, both of my parents. 11. is english your primary language? yes, no 12. do you know about the personal finance course (fina 1904) offered at ecu? yes, no 13. have you taken the personal finance course (fina 1904) at ecu? yes, no 14. if answer to q13 is yes, why did you choose to take the course? i thought it would be valuable to my future, i heard good reviews about the professor(s), i needed an elective course, i heard the class was easy. 15. if answer to q13 is yes, after taking the course do you feel that personal finance is more important or less important than before you took the course? less important, neither more nor less important, slightly more important, more important 16. if answer to q13 is yes, what grade did you receive in the course? a, b, c, d, f 17. if answer to q13 is yes, how valuable do you feel what you learned will be in the future? not at all valuable, somewhat invaluable, neither valuable nor invaluable, somewhat valuable, very valuable 18. if answer to q13 is no, why did you choose not to take this course? i did not think it would be worthwhile, i did not have any room in my schedule, i heard the class was hard, other. 19. if answer to q13 is no, do you plan to take the personal finance course (fina 1904) at ecu? yes, no 20. have you ever attended a personal finance seminar/informational session? yes, no 21. where have you ever studied personal finance? i have never studied personal finance, high school personal finance class, college level personal finance class, other. 22. do you plan to attend a personal finance seminar/informational session? yes, no 23. have you ever taken a finance course besides personal finance? yes, in high school; yes, in college; no 24. have you ever taken an economics course? yes, in high school; yes, in college; no 25. how important do you think personal finance is? very unimportant, unimportant, neither unimportant nor important, important, very important. 26. if answer to q24 is very unimportant or unimportant, why do you find personal finance unimportant? open 27. if answer to q24 is very important or important, why do you find personal finance important? open 28. how motivated are you to learn about personal finance? very discouraged, slightly discouraged, neither motivated nor discouraged, slightly motivated, very motivated. 29. how strongly have your parents influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 30. how strongly have other relatives (besides your parents) influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 31. how strongly have instructors influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 32. how strongly have your peers influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 33. how strongly has the media influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 34. how strongly have online resources influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 35. how strongly have your own experiences influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 36. how strongly has your religion influenced your knowledge of personal finance? 1not at all, 2, 3, 4, 5 a great deal. 37. how strongly have your parents influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 38. how strongly have your other relatives (besides your parents) influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 104 j. j. beierlein et al. / financial services review 30 (2022) 89–105 39. how strongly have your instructors influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 40. how strongly have your peers influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 41. how strongly have online resources influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 42. how strongly has the media influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 43. how strongly have your own experiences influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 44. how strongly has your religion influenced your financial decisions? 1not at all, 2, 3, 4, 5 a great deal. 45. when you have questions in regard to personal finance in the future, where do you think that you will go to for advice? (choose all that apply). friends, parents, other relatives, online, books, financial advisor. 46. for the following financial products, please indicate if you have heard of the product credit card. yes, no 47. for the following financial products, please indicate if you have heard of the product renters insurance. yes, no 48. for the following financial products, please indicate if you have heard of the retirement account. yes, no 49. for the following financial products, please indicate if you have heard of the product investment account (529 college savings plan, trust, etc.) . yes, no 50. for the following financial products, please indicate if hold this type of account credit card. yes, no 51. for the following financial products, please indicate if you hold this type of account renters insurance. yes, no 52. for the following financial products, please indicate if you hold this type of account retirement account. yes, no 53. for the following financial products, please indicate if you hold this type of account investment account (529 college savings plan, trust, etc.) . yes, no 54. do you plan to invest in the stock market? why or why not? yes, i plan to invest in the stock market in the future; yes, i a.m. currently invested in the stock market; no, the stock market is too risky; no, i will not make enough money to invest; other. 55. imagine that five brothers are given a gift of $1,000. if the brothers have to share the money equally how much does each one get? 100, 200, 250, i don’t know. 56. now imagine that the brothers have to wait for one year to get their share of the $1,000 and inflation stays at 3%. in one year’s time will they be able to buy: less than they could buy today, the same amount as they could buy today, more than they could buy today, i don’t know. 57. you lend $25 to a friend one evening and he gives you $25 back the next day. how much interest has he paid on this loan? 100%, 50%, 0%, i don’t know. 58. suppose you put $100 into a no fee savings account with a guaranteed interest rate of 2% per year. you don’t make any further payments into this account, and you don’t withdraw any money. how much would be in the account at the end of the first year once the interest payment is made? 98, 100, 102, i don’t know. 59. high inflation means that the cost of living is increasing rapidly. true, false. 60. it is usually possible to reduce the risk of investing in the stock market by buying stocks in many different companies. true, false. 61. an investment with a high return is likely to be low risk. true, false. 62. how confident are you in your answers to the above financial literacy questions? 1 not confident at all, 2, 3, 4, 5 extremely confident. 63. how confident are you in your ability to make financial decisions in your future? 1 not confident at all, 2, 3, 4, 5 extremely confident. 64. in the future, who do you expect to make your financial decisions? yourself, you and your spouse together, your parents, other j. j. beierlein et al. / financial services review 30 (2022) 89–105 105 pii: s1057-0810(99)00006-2 book reviews the wall street journal interactive editioninternet site 1. introduction with a circulation of 1,774,880,the wall street journal(or simply “the journal” as it is often referred to) is the leading daily newspaper with a business orientation in the world. it is often referred to as “the paper of record for business” (top newspapers by circulation, 1997). students enrolled in university business programs are able to get substantially reduced subscription rates by participating in the journal’s journal in education program. recently, the journal provided student subscribers free access tothe wall street journal interactive edition. the purpose of this review is to (a) provide a brief overview ofthe wall street journal interactive editionsite and (b) to examine how the resources available at this site can be integrated into basic economics and finance courses at the college level. 2. overview of the site the prospective subscriber tothe wall street journal interactive editionsite is able to obtain an overview of the site at http://interactive.wsj.com/. to actually enter the site one must be a subscriber and enter a username and password. as with the hard copy version of a daily newspaper, one generally begins viewing the site at the front page (http://interactive. wsj.com/edition/current/summaries/front.htm). the front page is structured like the hard copy version of the journal, except that it provides intraday updates of news events. the front page contains summaries of world and national news and business news. if the reader wants more detail about a particular story, links to the more detailed story are provided. the marketplace section (http://interactive.wsj.com/edition/current/summaries/marketpl. htm) contains a wide range of general business articles primarily in the areas of marketing and management. this section also contains links to regional editions of the journal. the money and investing section (http://interactive.wsj.com/edition/current/summaries/ money.htm) contains articles on investing and finance. the data bank portion of this page contains very detailed information for financial markets in the united states as well for most market economies around the world. in addition, subscribers to the journal have free access financial services review 7 (1998) 218–221 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00006-2 via links to the online version of the personal finance magazinesmart money(http:// www.smartmoney.com/) and to the online version of the financial publicationbarron’s (http://interactive.wsj.com/pages/barrons.htm). another section of the online version of the journal that would be of value to the undergraduate business student is the tech center. two features of the tech center are of value to the undergraduate business student. these are the watching the web page and the personal technology center. the watching the web page, which appears every thursday, contains the web sites (with descriptions and links) that the journal staff deems worthy of further review by the reader. the recommended sites are not limited to business topics, but rather include sites with very diverse content, including sites concerned with, among other topics, computer technology, social issues, music, health, and genealogy. the personal technology center contains articles written by walter s. mossberg. these are “how to” articles for individuals who wish to develop and improve their personal computer skills. the articles are devoid of jargon and easy to follow. the site provides easy access to an archive of previous articles by mossberg. in addition, mossberg answers computer and technology questions from readers. readers can email their questions to mossberg. the careers page of the online version of the journal contains articles that deal with various aspects of employment. undergraduate business students will find these articles helpful as they consider their career options, and in understanding some of the issues they will face in the workplace. there are two additional features available at the journal’s interactive site that undergraduate business students will find useful. the search feature allows the subscriber to conduct a topical search from articles that appeared over the past fourteen years inthe wall street journaland other major business publications. the personal journal allows the subscriber to capture an individualized collection of items that appear in the journal. the personal journal can be configured in such a way so as to automatically receive any articles on a specific topic, specific columns from the journal, and mailing lists and news alerts. in addition, the subscriber can create up to five portfolios to track stocks, bonds and mutual funds. once the portfolio is set up, the subscriber receives information on current prices (stock prices are on 20 minute delayed basis, mutual funds prices are updated daily), percent change, gain/loss information, the portfolio’s current value, and links to current news stories on any of the items contained in the portfolio. the portfolio data can be downloaded into spreadsheets such as lotus and excel. 3. integrating the site into classes it is recognized that a single internet site cannot adequately serve as the sole source of internet resources in a class. nevertheless, i believe thatthe wall street journal interactive edition can serve as an important internet resource in introductory economics and finance 219r.f. bieker / financial services review 7 (1998) 218–221 pii: s1057-0810(99)00006-2 classes. in this section, i consider how this site can be integrated into the principles of economics and personal finance classes. the first principles of economics course (principles of macroeconomics) offered at delaware state university is one of the required general education courses at the university. as such, the course must contain elements that serve to enhance the student’s writing skills and level of computer literacy. these requirements are in addition to the main objective of insuring that the student achieves a minimum level of economic literacy as specified in the common course syllabus. to attempt to further the writing skills of students in the principles of macroeconomics course, i require students to maintain an economic issues notebook as proposed by petr (1990). each week students are required to compile an article from the online version of the journal dealing with a current economic issue. each article must be annotated with the student’s comments. student comments must address the following: (1) does the article make sense from an economic perspective? (2) what are the implications of the article for the macroeconomy? and (3) what are the implications of the article for the student personally? this exercise has several objectives: (1) to have students continuously writing under the assumption that practice makes perfect; (2) to give the students an appreciation for the research process; and (3) to foster the integration of what the student learns in an academic sense with the student’s personal experience. there are both advantages and disadvantages of developing the economic issues notebook exclusively from the online journal. one advantage is that by continuous exposure to the journal over the semester, students become familiar with this major business publication. in addition, requiring students to access the online journal at least once a week serves to further the objective of enhancing the students’ level of computer literacy. currently, many students enrolled in the course have rather limited experience with the use of the personal computer and the internet. requiring them to access the journal site each week and search for the article of interest serves to enhance their computer literacy skills. the major disadvantage of limiting articles in the economic issues notebook to those from the journal is that students are limited to only one of the many available sources. the other course in which the journal site is used is my personal finance course. the features of the online journal that are utilized most intensively in this class are the online version ofsmart money, the money and investing section of the journal, and the portfolio feature of the journal. thesmart moneysite is used throughout the course since it provides easy access to supplemental information on the course topics. by reading the articles at this site students gain additional insight into course topics and appreciate that the topics addressed in the course have a high level of current interest. also, the articles at this site are more current and in many cases of more direct practical value to the student. the journal site is most heavily used during the investment portion of the course. students use the tools section of the smart money site as well as the money and investing portion of the journal to develop a portfolio. the portfolio feature of the journal is then used by the students to monitor their portfolios. 220 r.f. bieker / financial services review 7 (1998) 218–221 pii: s1057-0810(99)00006-2 4. summary and conclusions the purpose of this review was to provide an overview ofthe wall street journal interactive editioninternet site and to consider how this site can be integrated into selected undergraduate economics and finance courses. the site provides a wealth of information that can be used to enhance the principles of economics and personal finance courses. integration of the site into a course also allows the students to enhance their computer skills. however, it is important to recognize that a single internet site cannot adequately serve as the sole source of internet resources in a class. richard f. bieker* professor, delaware state university, school of management, department of accounting and finance, 1200 north dupont highway, dover, de 19901-4932, usa e-mail address:rbieker@dsc.edu references petr, j. l. (1990). student writing as a guide to student thinking. in p. saunders & w. b. walstad (eds.),the principles of economics course(pp. 127–140). new york: mcgraw-hill. top newspapers by circulation. (1997, september).editor & publisher. [online]. available: http://www. mediainfo.com. 221r.f. bieker / financial services review 7 (1998) 218–221 pii: s1057-0810(99)00006-2 pii: 1057-0810(95)90006-3 financial services review, 4(2): 109-122 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the market pricing of disability income insurance for individuals larry a. cox sandra g. gustavson individuals ’ needs for disability income insurance dominate those for life insurance, yet relativelyfew buyersandsellersenterintodisabilitycontractscomparedtolifecontracts. this phenomenon appears contradictory to the existence of a workably competitive market. this study examines the relation of disability income insurance prices to underlying contractual and insurer characteristics. our results are supportive of a competitive market scenario. we observe a strong relation between prices and elimina tion periods, which is consistent with the presence of adverse selection. our results have implications for how individuals should choose some policy and insurer characteristics, but they also suggest that buyers may need to be better informed about other pricing factors. i. introduction based on actuarially expected losses, individuals’ needs for disability income insurance (dii) dominate those for life insurance at all ages of their working lifetimes (cox, gustavson, & stam, 1991). despite such evidence, relatively few insurers offer individual dii (limra, 1989) and only a small percentage of consumers purchases it (cox, 1991). low purchase rates for dii can appear to be inconsistent with a competitive market environment, but direct scrutiny of the market for individual dii reveals many aspects of a workably competitive environment-similar to those observed in life insurance markets (pritchett & wilder, 1986). for example, no product suppliers have dominated historically with respect to market share or price leadership.’ also, barriers to entry and exit are not substantial and evidence of collusion or other anticompetitive activities is not apparent. further, the same regulatory structure applies to both dii and life insurance markets. given these similarities for both market and regulatory structures, the dii and life insurance markets larry a. cox l the university of mississippi, department of economics and finance, 301 conner hall, university, ms 38677. sandra g. gustavson l the university of georgia, faculty of risk management and insurance, 206 brooks hall, athens, ga 30602-6255. 110 financial services review 4(2) 1995 should share approximately the same universe of potential buyers and sellers. yet, the fact remains that life insurance is widely purchased while dii is not.* in a competitive environment, we expect dii prices to effectively impound both contractual and insurer attributes, in this study, we focus on the general composition and strength of such pricing relations. we use a hedonic pricing model-similar to that applied by walden (1985) for life insurance products-to determine the strength of pricing relations and to estimate the impact of various contractual provisions and insurer attributes on dii prices. our results provide evidence of strong relations for some of these underlying characteristics. throughout our analysis, we emphasize the expected impact of the elimination period clause, as supported by the theory of adverse selection. the elimination, or waiting, period contained in dii policies is equivalent to the monetary deductible found in other types of insurance products. when a loss of health occurs, the insured first must absorb a loss of income, which is directly proportional to the duration of the elimination period, before receiving insurance benetits.3 our test results specific to elimination periods, although not direct measures of adverse selection, are quite consistent with its presence in dii markets. our results provide indications about how individuals with lower risk exposures should evaluate elimination periods. the impact of non-price factors, such as service, also is demonstrated. although dii markets are shown to be reasonably efficient, our findings suggest that buyers may need to be better informed about disability definitions, preexisting conditions clauses, and insurer solvency. in the next section, we focus on how the elimination period and other devices are used by insurers to reduce problems caused by adverse selection. we discuss the primary contractual components and insurer characteristics that are likely dii price drivers in section iii. in section iv, we specify the data and model for testing price relations. we then present our results in section v and conclusions in section vi. ii. adverse selection and elimination periods studies of insurance markets often adopt akerlof s (1970) thesis that insurance buyers have informational advantages because they know more about their personal risk charac teristics than do insurers. this asymmetry of information can result in adverse selection in insurance markets. adverse selection is the tendency of persons with higher risk exposures either to be more likely to buy insurance or to buy more insurance than persons with lower risk exposures (cummins et al., 1983). empirical research has confirmed the existence of adverse selection in the markets for automobile (dahlby, 1983; puelz & snow, 1994), individual life (beliveau, 1984) and individual health (browne, 1992; browne & doerping haus, 1993) insurance. for reasons discussed below, we expect the presence of adverse selection to have a substantial impact on the design and pricing of dii contracts. dii is a form of health insurance covering indirect losses. if insured persons suffer losses of health such that they cannot perform their occupational duties, they recover a portion of their lost income. following akerlof (1970), we assume that applicants for any health insurance products, including dii, know more about their personal health on the contracting date than do insurers. insurers have less information for reasons such as errors in health assessment by medical examiners serving as agents for the insurer, personal physicians’ the market pricing of disubirity income insurance 111 sympathy for applicants, and inadequate histories of family health traits. insurers, therefore, should have less complete information than applicants regarding their proper classification as relatively high or low risks. in their seminal work, rothschild and stiglitz (1976) theorized that insurers will offer “price and quantity” menus to applicants as a method for reducing adverse selection. high-risk and low-risk applicants are likely to select different price-quantity combinations, thereby revealing their perceived level of risk to insurers. specifically, rothschild and stiglitz predict a separating equilibrium in which high-risk applicants will buy relatively more insurance at a higher price, while low-risk applicants will buy less at a lower price.4 puelz and snow (1994) expand on the framework provided by rothschild and stiglitz to show specifically how insurers respond to adverse selection. they focus primarily on the relation of deductibles to price. according to puelz and snow, insurers first screen the pool of applicants and categorize them based on observable traits that relate to risk. they then offer various combinations of prices and deductibles to applicants within each riskcategory. lower prices are offered in tandem with higher deductibles. lower-risk applicants are expected to signal their status by selecting high-deductible policies. puelz and snow find support for these expectations based on results generated from one auto insurer’s data. our observations indicate that insurers in dii markets use pricing methods similar to those found by puelz and snow for the automobile insurance industry. in dii markets, insurers first use the observable traits of age, gender, and occupation to place applicants in risk categories. then, applicants in each risk category are offered a menu of prices and contractual provisions, including a variety of elimination periods. because elimination periods essentially are deductibles and the levels selected by dii applicants serve as signals of riskiness to insurers, we expect a particularly strong relation to policy price. another way that insurers can control adverse selection is by marketing through brokers who are closer to the point of purchase and, therefore, should generate better estimates of applicants’ health risks. indeed, most of the larger dii providers use brokerage arrangements to distribute their products. unlike the life insurance market, direct retailing and mass merchandising in the dii market are not common.5 because of our data limitations, however, assessing the impact of alternative distribution systems is beyond the scope of this study. iii. characteristics affecting disability income ~surance prices a. contractual attributes of disability income insurance products 1. elimination period as explained previously, applicants’ selections of particular elimination periods signal insurers as to their unobservable health status. because insurers should heavily rely on this characteristic as an indicator of applicant risk and because longer elimination periods, by definition, reduce total benefits paid in the event of a disability, we anticipate a very strong, negative relation between price and elimination period. 112 financial services review 4(2) 1995 2. definition of disability the definition of disability is critical in determining whetherinsureds qualify as disabled for purposes of receiving claim payments. buyers’ applications for policies with particular definitions can provide signals to insurers about applicants’ perceived health. such decisions also can signal the applicants’ levels of risk aversion, however. most insurers offer either “own occupation” or “any occupation” definitions of insur ance.6 the own occupation definition is considered more liberal because insureds must lose only their ability to perform the principal duties of their present occupations to qualify for benefits. under the any occupation definition, insureds must be unable to perform the primary duties of any occupation for which they are suited by training, experience, or education. many dii policies contain “split” definitions in which the own occupation criterion applies for a specified period, such as two years, and the any occupation standard applies thereafter. because the own occupation definition is more liberal, we expect a positive relation between dii price and the time period for which this definition applies. 3. prexisting conditions most dii contracts contain a preexisting conditions clause. if a disability eventually arises out of an injury or illness incurred within a specified period-such as six months before the policy was initiated, benefits will not be paid. insurers include such a provision to prevent the adverse selection that would occur if applicants could begin receiving insurance coverage after probable health problems were revealed. because longer exclusion periods reduce expected benefit payments to insureds, we anticipate a negative relation between the length of the exclusion period and policy price. 4. residual disability benefit dii policies vary considerably with respect to the treatment of partial disability. some specify no benefits whatsoever unless insureds meet the total disability definition. others provide payment of “residual” benefits only if the partial disability follows a previous total disability. the most liberal clause pays benefits for a partial disability whether or not it is immediately preceded by a period of total disability. the residual disability benefit is stated as a percentage, usually based on insureds’ weekly work hours during their partial disability, of the total disability benefit. adding a residual disability provision to a dii contract can provide incentives for insureds to malinger. many insurers offer a generous residual benefit structure, however, to increase disabled insureds’ incentives to return to work, thereby reducing the likelihood of extended, total disabilities. generally, a residual disability benefit should raise insureds’ expectations of receiving benefits and, therefore, be positively reflected in dii prices. these prices should be further enhanced if residual disability benefits are not conditional on an immediately preceding total disability. 5. other contractual provisions other policy provisions are used to control the risk facing insurers. these include clauses addressing causes of loss (e.g., sickness and accident), benefit amount and duration, and the market pricing of disabiriry income insurance 113 participating dividends. in this study, we are able to control for causes of loss and benefit amounts and durations by analyzing only policies with identical provisions.’ the participa tion factor is perfectly proxied by the organizational form variable, which is among the insurer characteristics discussed below. b. insurer characteristics 1. organizational form the major dii insurers have either a stock or mutual (or mutual-like) organizational structure. to explain the effects of organizational structure on the corporate policies of insurers, mayers and smith (198 1) developed the managerial discretion hypothesis (mdh). they posit that mutual firms have a comparative advantage in insurance lines where less managerial discretion is required: for example, where stable actuarial tables and long-term contracts are the norm and where renewal options are relatively valuable (mayers & smith, 1981; smith, 1986). although the sensitivity of disability claims to economic changes over time may make actuarial tables somewhat less stable than those for life insurance, the dii market is characterized by relatively well-defined actuarial experience, long-term contracts, and the great importance of renewal options. if the costs of controlling insurer management are lower for mutual firms in such an environment, then consumers will discount prices for dii issued by stock insurers. stock firms can overcome this disadvantage, however, if stockholders agree to limit dividend, investment, and financing policies (mayers & smith, 1981). the evidence regarding the impact of organizational form on insurance prices is both scant and mixed. myers and pritchett (1983) observed higher premiums for the participating life insurance policies issued by mutual insurers versus those issued by stock insurers. walden (1985) did not find organizational form to be a significant factor affecting life insurance prices when other contract and insurer variables were considered, however. given the conflicting empirical evidence and the countervailing theoretical arguments of mayers and smith, we cannot predict the direction of any relation between organizational form and dii prices. 2. default risk in well-functioning markets, insureds should discount prices of dii issued by insurers with greater default risk. we expect a negative relation between dii prices and the default risk of insurers. further, we expect buyers of longer duration policies to be relatively more concerned with default risk because of the greater magnitude of their potential claims. 3. market share higgins, shughart, and tollison (hst, 1989) show that firms with larger market shares strive to acquire more market information and this leads to greater variation in both price and such non-price characteristics as product quality. hst conclude that higher market share will increase the degree to which differentiated price-quality combinations are offered to consumers. if larger firms are able to command premium prices by offering attractive non-price characteristics, as suggested by both hst and carlton (1986), the relation between market share and dii prices will be positive. 114 financial services review 4(2) 1995 iv. research design a. data our data encompass dii prices for 54 insurers at year-end 1988 collected from national underwriter company’s 1989 disability income & health insurance. these data allow us a relatively large number of observations, because after 1988 many large insurers began marketing products of the largest dii issuers via private labelling or co-marketing agree ments (conning & company, 1993, pp. 48-50). discount rates are derived from seasoned u.s. treasury bond rates as of december 30, 1988. to estimate default risk, we use insurance regulatory information system (iris) ratios published by the national association of insurance commissioners (naic; 1989). the naic uses the 12 iris ratios as an initial screen of insurer solvency. insurers with four or more ratios outside defined “usual ranges” for each ratio are reviewed more closely by regulators to determine if remedial measures must be taken. market shares are calculated from survey data published by limra (1989). b. method 1. test model the characteristics model, first introduced by lancaster (1971), allows the hedonic decomposition of prices. we employ ols regression to estimate the dii pricing relation. jones (1988) verifies that this approach is robust over nonlinear price functions and different market structures. our model follows: dp = y. + ylep + y2prex + y3def + y&id1 + y5rsid1 *rsid2 + y60rg + y,drisk + ysms + e (1) where: dp = price of the disability income insurance, ep = elimination period, def = definition of disability, prex = exclusion period for preexisting conditions, rsidi = residual disability benefit (yes = 1; no = 0), rsidz = contingency of residual disability benefit on previous total disability (no = l;yes=o), org = organizational form (mutual or fraternal = 1; stock = 0), drisk = default risk, and ms = market share. we apply the model to policies with both long (to age 65) and short (two-year) durations. 2. price proxy the proxy for dii price is the discounted present value, divided by $1,000, of annual standard-rate premiums over the duration of the policy.* the benefit amount is assumed to be $2,000 per month for a unisex-rated individual aged 35.9 unisex-rated policies are used because the larger dii providers generally offered such policies in 1988 and, as a result, the market pricing of disubil@ income insurance 115 sample size is maximized.‘0 the insurers in our sample account for approximately 40% of the dii premiums written in 1988. 3. contract options the ratio of the elimination period to the maximum benefit period serves as the elimination period (ep) estimate. the proxy for the definition of disability (def) is the period for which the own occupation definition applies divided by the maximum benefit period. the preexisting conditions proxy (prex) is the exclusion period applicable for preexisting conditions divided by the maximum benefit period. two binary variables represent the residual disability benefit. the first (rsidi) equals one if the policy includes a residual disability benefit and zero if not. the second (rszd2) equals one if the residual disability benefit is unconditional with respect to a previous total disability and zero if it is conditional. because the rszd2 variable applies only if a residual disability benefit exists initially, we include the product of the two proxies (rsidi *rsid2), which represents the interaction effect between the two, rather than the rszd2 proxy itself. consequently, we must consider the coefficients for both rsidi and rsidi*rsid2 in tandem. 4. insurer characteristics we use a binary variable to denote organizational form (org), with one indicating a mutual or fraternal structure and zero a stock firm. the default risk proxy (drzsk) is the percentage of 12 iris ratios for each insurer that fall outside the “usual range” defined by the naic. the market share proxy is the percentage of 1988 dii premiums written by each company compared to the industry total. c. limitations of the research design the screening of dii policies for homogeneous risk categories-such as those based on age, gender, and occupation-and product attributessuch as benefit levels and duration necessarily reduces sample size. response bias in the data may exist because insurers voluntarily report to national underwriter, but we cannot predict the direction of any such bias. we also cannot be certain that all the policies in our data set are representative of what each insurer actually sells. in some instances, information on a contract provision may not be published in the data source. we then assume this provision does not exist, which may not always be true. finally, we do not have data to test for effects of alternative distribution systems and claims practices, so these remain outside the scope of this study. v. empirical results and discussion table 1 contains selected statistics for the data gathered from the full sample of 44 policies providing benefits to age 65. as shown by these results, the annual premium per $100 of monthly benefit varies widely between these policies. elimination periods also vary substantially-from 30 to 180 days. over half the policies contain a 30-day elimination period, however. although we are certain that most insurers routinely offer a range of 116 financial services review 4(2) 1995 table 1 selected statistics for the full sample of policies with durations to age 65 variables premium elimination period preexisting conditions residual disability organizational form iris ratios number mean std. dev. minimum qi median q3 maximum > 0 42.04 9.36 24.70 34.50 40.94 49.40 67.19 44 48.41 32.49 30 30 30 75 180 44 57 90 0 0 0 2 2 13 .57 so 0 0 1 1 1 25 .41 so 0 0 0 1 1 18 1.02 .98 0 0 1 2 3 27 notes: premium = annual premium per $100 monthly benefit; elimination period = number of days in elimination period; preexisting conditions = number of years in period stated in preexisting conditions clause; residual disability = 1 if residual disability benefit exists, 0 otherwise; organizational form = 1 if mutual or fraternal insurer, 0 otherwise; iris ratios = number of iris ratios outside naic-prescribed range. elimination periods, the prices published for each insurer in our data source generally reflect only one, and at most two, elimination periods. the statistics for the preexisting conditions and residual disability provisions indicate that many insurers do not offer these provisions. as stated previously, however, theseoptions may be offered but not published in our data source. although not contained in table 1, the statistics for the definition of disability variable show that 40 of the 44 policies contain some form of “own occupation” definition, consequently, pricing anomalies among the few “any occupation” policies easily can produce spurious test results. the results for the iris ratios indicate that most of the insurers in the sample were quite solvent and none were targeted for regulatory attention. thus, solvency distinctions based on iris ratios are quite subtle at best for the sample insurers. table 2 contains regression results for dii policies with benefit durations to age 65. regression (1) accounts for all available policies, including those issued by insurers for which no market share is available. as expected, a strongly negative relation between price and the elimination period is apparent. no other characteristics are significant, however. we do not find this surprising in that some insurers in the full sample may be relatively inactive and their policies not competitive, even though they publish pricing information. the condition index, a measure of collinearity developed by belsley, kuh, and welsch (1980), is relatively low for all the regressions reported in table 2. greene (1990) suggests that collinearity is not a problem unless the condition index exceeds 20. we also applied the breusch-pagan test for all the reported regressions and we cannot reject homoscedasticity at the .ool level, so heteroscedasticity is not a problem in our data set. in regression (2), we test only the policies of insurers for which we have market share data. the goodness-of-fit improves, as evidenced by the increase in the coefficient of determination (adjusted r*), despite the loss of five observations. the elimination period remains significantly negative at the .ol level and the market share coefficient is significantly positive, as expected, at the .ol level. the latter finding supports the premise that insurers serving as market leaders devise attractive packages of both price and non-price-for example, quality and service-factors. to further assure that the data we are using is representative of competitive market conditions, we next analyze only the policies of insurers with a minimum market share of t a b l e 2 r eg re ss io n r es ul ts f or u ni se x pr ic in g of d is ab ili ty i nc om e in su ra nc e w ith d ur at io ns t o a ge 6 5 r eg re s si on n um be r ia t (1 ) 12 .4 * (. 80 3) (2 ) 11 .3 ’ (. 8w (3 ) 12 .1 * t.9 66 1 i~ ep e~ en t v ar ia bl es ( ex pe ct ed s ig np sl at & ti cs ep d ef p r ex r si d i *r sl d 2 o r g ir is m s a dj us te d c on di tio n (1 (+ ) (i r si d i (c om bi ne d + ) w (1 (+ i r 2 in de x o bs er ve d -4 69 .2 * -. 73 7 .2 24 -1 .2 3 1. 15 .3 56 -1 .6 3 .3 91 8. 55 44 (9 0. 52 ) (. 69 3) (9 .6 8) (. 71 5) (. 78 7) w 37 ) (3 .3 8) -4 55 .0 * -. 57 9 1. 71 -. 39 5 .7 52 -. 52 7 -1 .8 1 30 .5 % ,5 12 .9 49 39 (8 4. 9) (. 69 2) (1 0. 54 ) c. 78 6) (. 76 1) (. 62 8) (3 .5 0) (1 0. 1) -4 47 .1 ” -1 .6 1 1. 08 -2 .1 9* * 2. 50 ** -. 45 7 1. 10 27 .4 * .6 18 11 .3 6 31 (9 1. 3) (. 90 7) (1 1. 14 ) (1 .0 2) (1 .0 4) (. 63 5) (3 .7 3) (9 .8 ) n ot es : pr ic es a re d is co un te d va lu es o f an nu al p re m iu m s fo r an in di vi du al a ge u p to 3 5 ye ar s. ‘i n t = in te rc ep t; d e f = r at io o f “o w n oc cu pa tio n” d ur at io n to to ta l po lic y du ra tio n; p r e x = r at io o f du ra tio n fo r pr ee xi st in g co nd iti on s cl au se t o to ta l d ur at io n; e p = ra tio o f el im in at io n pe ri od t o to ta l d ur at io n; r si d l = 1 if r es id ua l d is ab ili ty b en ef it of fe re d, 0 o th er w is e; r si m = 1 if r es id ua l d is ab ili ty b en ef it no t c on di tio na i on pr ev io us t ot al d is ab ili ty , 0 ot he rw is e; r si d l * r si d 2 = in te ra ct io n ef fe ct ; o r g = 1 if m ut ua l o r fr at er na l in su re r, 0 i f st oc k in su re r; d r is k = n um be r of i r is ra tio s ou ts id e n a ic -p re sc ri be d ra ng e di vi de d by t ot al r at io s re qu ir ed ; m s = m ar ke t sh ar e. v al ue s in p ar en th es es a re s ta nd ar d de vi at io ns . *s ig ni fi ca nt a t t he . o l l ev el . ** si gn if ic an t a t t he . 05 le ve l. t a b l e 3 r eg re ss io n r es ul ts fo r u ni se x pr ic in g of d is ab ili ty in co m e in su ra nc e w ith t w oye ar d ur at io ns r eg re ss io in de pe nd en r v ar ia bl es ( ex pe dr ed s ig @ sr af is ri cs n ep d ef p r ex r si d i *r si d 2 o r g ir is m s n um be r ii ~ a a@ st ed c on di ri on h (+ i -) r si d i (c om bi ne d +) (3 6) (+ ) r 2 in de x o bs er ve d (1 ) 1. 09 * -5 .0 5* .0 56 .0 72 -.1 62 .1 57 .1 69 -.2 04 .3 95 16 .8 8 25 (.2 24 ) (1 .4 7) (. 17 8) (. 10 2) (. 12 6) (. 11 9) (. i 13 ) (6 .3 1) (2 ) 1 . & i* -4 .1 1* -.0 57 .0 52 -.0 85 ,2 05 -.0 94 .3 34 5. 04 ** .5 32 18 .4 3 20 .2 37 ) ( 1 .4 8) (. 19 3) (1 .2 2) (. 14 4) (. 12 6) (. 14 0) (. 71 4) (1 .8 5) (3 ) .9 28 * 4. 60 * -.0 67 .2 06 -. 41 2* * .4 86 * -.2 75 2. 11 ** 5. 92 * .i 34 18 .0 1 17 (. 20 9) (1 .2 5) (. 15 7) (. i 16 ) (. w (. w (1 .3 0) (. 86 5) (1 .5 8) n or es : pr ic es ar c di sc ou nt ed va lu es of a nn ua l pr em iu m s fo r an i nd iv id ua l ag ed 35 . fo r de fi ni tio ns of v ar ia bl es , se e t ab le 2. *s ig ni fi ca nt at t he . 01 l ev el . ** si gn if ic an t at t he . 05 l ev el . the market pricing of disabirity income insurance 119 one percent in regression (3). despite the loss of eight more observations, the adjusted r2 increases by over 10 percentage points to 61.8%. the elimination period and market share coefficients again are significant in the expected directions. both the residual disability coefficients also are significant at the .05 level, but exhibit opposite signs. to assess the combined effect for residual disability, we add the rszdi coefficient to the product of the second coefficient (rszdl*rsid2) multiplied by the rszdi value (zero or one) for each policy in the subsample. the resulting effect is indeterminate as price is positively related to the residual benefit clauses for only 15 (48%) of the insurers in the subsample. table 3 exhibits regression results for unisex-rated policies with relatively short, two-year durations. the coefficients for the elimination period and market share are significant and in the expected direction for all three regressions. results for the residual disability variables again are mixed, although the combined effect is significantly positive for the majority of firms in the reduced sample of more active dii insurers, as shown by regression (3). one abnormal result is the positive relation, significant at the .05 level, between default risk (drzsk) and price for the reduced sample examined in this regression. this may be a spurious result because our iris proxy does not greatly discriminate between all sample insurers, as we explained previously, and this subsample is very small. another possible explanation is that buyers of very short duration policies should be less sensitive to default risk than are buyers of long-term policies because they have much less at risk. we also note that, because of historically low failure rates in the life insurance industry, insureds may have been little concerned with insurer solvency risk in 1988. the failures of such visible life/health insurers as executive life, first capital, and mutual benefit have changed public perception of insurer default risk in subsequent years. vi. summary and conclusions in our analysis, we show that the individual disability income insurance (dii) market is similar in many ways to the more heavily researched life insurance market. our empirical results indicate that dii prices effectively impound several contractual and insurer charac teristics in a manner consistent with a competitive environment. even though our data set is limited and subject to some strong assumptions, our results show that elimination period provisions and insurers’ market share are strongly related to dii prices, while residual disability benefits also are significant factors in some instances. given the incentive structure of the individual dii market, we expect insurers to face adverse selection. in this environment, we expect them to offer a broad menu of deductibles to distinguish between highand low-risk applicants. our results confirm that insurers do offer a wide variety of elimination periods and that these provisions are quite important in determining dii prices. such evidence is generally supportive of a scenario in which adverse selection is present. from the individual’s perspective, our results show that the elimination period is a very important pricing factor and that low-risk buyers are given substantial price incentives to choose longer elimination periods. whether high-risk buyers should do the same is a central economic question that we cannot resolve, however. our results also indicate that individuals willing to forgo non-price attributes, such as service, can extract lower prices from smaller 120 financial services review 4(2) 1995 dii providers. finally, our findings provide a preliminary indication that disability defini tions, preexisting conditions clauses, and insurer solvency may not be fully impounded in dii prices. if subsequent investigations confirm these results, then individuals need to be better informed about these factors by financial advisors and educators. because of data limitations, we cannot provide a direct test of adverse selection in the dii market. we anticipate that further refinement of the theory of adverse selection and direct testing will provide more definitive answers as to why individuals’ purchase rates are so low. efforts exploring the impact of insurers’ claims practices and their marketing and distribution policies on adverse selection also represent potential contributions. alternative data sets also may advance our knowledge with respect to contractual and insurer characteristics that did not have a significant impact on prices in our study. new research initiatives inevitably will face difficult constraints, however, because the available public data are quite limited and this problem is likely to be exacerbated as market power becomes increasingly concentrated within a small group of insurers. acknowledgments: the authors thank helen doerpinghaus, han kang, and jorge urrutia for their helpful comments on previous versions of this paper. notes 1. according to the life insurance marketing and research association (limra; 1989), the largest provider of individual disability income insurance wrote only 12.6% of premiums written on both guaranteed renewable and noncancellable policies combined in 1988. the top four insurers accounted for only 37.2% of the market surveyed by limra. more recent information, discussed below, indicates increasing market concentration since that time. 2. the life insurance fact book (american council of life insurance, 1994) indicates that 67% of all adults own some type of life insurance, while cox, gustavson, and stam (1991) find that only 22% of u.s. earners own dii. 3. for further discussion of the elimination period and how it varies from a probationary period or time deductible, see f’luet (1992). 4. other researchers, most notably wilson (1977) and miyazaki (1977), disagree with roth schild and stiglitz’s view of a separating equilibrium. they theorize a pooling equilibrium in which both high-risk and low-risk applicants purchase insurance at the same price, such that low-risk applicants subsidize their high-risk counterparts. 5. one exception is the direct marketing of accidental disability insurance by banks to cover individuals’ monthly mortgage payments. 6. a growing minority of policies contains definitions based upon the percentage of income lost. such “loss of income” definitions can be either more or less generous than ones with own occupation definitions, depending upon the specific terms. 7. we initially examined renewability provisions, which should be of great value to insureds. al1 firms in our sample offered renewability clauses, but between-firm differences were small. we now omit them from our analysis because they provide no incremental explanatory power with regard to prices. 8. annual dii premiums are discounted by seasoned, u.s. treasury bond rates for bonds with maturity dates equal to the benefit duration. 9. no dividend schedules for the sample policies are provided by our data source. our anecdotal evidence indicates that major insurers issuing policies labeled as participating by national the market iwing of ~is~~~ income insurance 121 underwriter do not publish such schedules, although they do so for life insurance products. in fact, representatives of several large mutual insurers informed us that dividends are rarely, if ever, declared on dii policies. 10. at the time of this writing, many leading dii issuers are returning to gender-based pricing by charging females higher rates @lease, 1995). although the frequency of disability during the working lifetime has been greater historically for cohorts of females (see cox, gustavson, & stam, 1991), recent industry experience now indicates that loss severity is higher for females, too (conning & company, 1993). references akerlof, g.a. (1970). the market for “lemons”: quality uncertainty and the market mechanism. quarterly journal of economics, 84,488-500. american council of life insurance. (1994). 1994 iife insurancefact book. washington, dc: author. beliveau, b.c. (1984). theoretical and empirical aspects of implicit information in the market for life insurance. journal of risk and insurance, 51,286-307. belsley, d.a., kuh, e., & welsch, r.e. (1980). regression diagnostics. new york: john wiley and sons. blease, r.l. (1995). di writers try toend the bleeding. best’sreview (life/health edition), (february), 56-60. browne, m.j. (1992). evidence of adverse selection in the individual health insurance market. journaz of risk and insurance, 59, 13-33. browne, m.j., & doerpinghaus. hi. (1993). information asymmetries and adverse selection in the market for individual medical expense insurance. journal ofrisk and insurance, 60,300-3 12. carlton, d.w. (1986). the rigidity of prices. american economic review, 76,637-668. cox, l.a. (199 1). disability income insurance and the individual. financialservices review, i, 6 l-78. cox, l.a., gustavson, s.g., & stam, a. (199 1). disability and life insurance in the individual insurance portfolio. journal of risk and insurance, 58, 128-l 37. conning & company, (1993). non-cancelable disability insurance: decision time. hartford, ct: author. cummins, j.d., smith, b.d., vance, n.r., & vanderhei, j. (1983). risk classification in hfe insurance. boston: kluwer, nijhoff. dahlby, b.c. (1983). adverse selection and statistical discrimination: an analysis of canadian automobile insurance. journal of pubiic economics, 20, 121-i 30. flue& c. (1992). probationary periods and time-dependent deductibles in insurance markets with adverse selection. in g. dionne (ed.), contributions to insurance economics (pp. 359-375). norwell, ma: kluwer. greene, w.h. (1990). econometric analysis. new york macmillan. higgins, r.s., shughart, w.f., ii, & tolfison, r.d. (1989). price leadership with incomplete informa tion. journal of~onom~ behavior and organization, ii, 423-429. jones, l.e. (1988). the characteristics model, hedonic prices, and the clientele effect. journal of political economy, 96, 55 1-567. lancaster, k. (1971). consumer demand: a new approach. new york: columbia university press, life insurance marketing and research association (limra) international. (1989). 1988 individual health issues and in force survey. hartford, ct: author. mayers, d., br smith, c.w., jr. (1981). contractual provisions, organizational structure and conflict control in insurance markets. journal ofbusiness, 54,407-434. miyazaki, h. (1977). the rat race and internal labor markets. the bell journal of economics, 8, 394-418. 122 financial services review 4(2) 1995 myers, p.s., & pritchett, st. (1983). rate of return on differential premiums for selected participating life insurance contracts. journal ofrisk and insurance, 50,569-586. national association of insurance commissioners. (1989). insurance regulurory information system ratio results 1988. kansas city, mo: author. pritchett, s.t., & wilder, r.p. (1986). stock life insurance company profitability and workable competition, philadelphia: s.s. huebner foundation of insurance education, university of pennsylvania. puelz, r., & snow, a. (1994). evidence on adverse selection: equilibrium signaling and cross-sub sidization in the insurance market. journal of political economy, 102,236-257. rothschild, m., & stiglitz, j. (1976). equilibrium in competitive insurance markets: an essay on the economics of imperfect information. quarrerly journal of economics, 90,629-649. smith, c.w., jr. (1986). on the convergence of insurance and finance research. journal of risk and insurance, 53.693-7 17. walden, m.l. (1985). the whole life insurance policy as an options package: an empirical investiga tion. journal of risk and insurance, 52,4&58. wilson, c. (1977). a model of insurance markets with incomplete information. journal of economic theory, 16, 167-207. from the editor this issue contains volume 28 issue 1 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “the decrease in life insurance ownership: implications for financial planning” is coauthored by kyoung tae kim at university of alabama, travis p. mountain at virginia tech university, sherman d. hanna at ohio state university, and namhoon kim at korea rural economic institute. using the survey of consumer finances dataset the authors find the proportion of households owning a life insurance policy decreased from 72% in 1992 to 60% in 2016. they estimate logistic regressions on the likelihood of ownership of term and cash value life insurance. they find that changes in household characteristics accounted for the decrease in term life insurance ownership, but not for the decreases in cash value life insurance ownership. they also find a positive association between use of a financial planner and life insurance ownership. the second article “are ‘fun’ sources of windfalls destined to be spent hedonistically?” is coauthored by eugene bland at texas a&m university – corpus christi and valrie chambers at stetson university. the authors show that fun sources of income are more likely to be spent on a fun expenditures. money won on a game show would be spent more on ‘fun’ than money received from a tax rebate. they find support for rejecting the hypothesis that there is no difference in allocations for regular expenses, credit card payments, durable assets or investing in stocks, bonds and savings account (“adult” uses of funds) by source of windfall. they found significant evidence that there is a difference in investing based on the source of the windfall. people apparently spend significantly more on fun when a fun windfall is received, but that spending on fun is not limitless. additionally they find that there may be such a thing as “enough spending on fun. the third article, “a portfolio of leveraged exchange traded funds” is coauthored by william j. trainor jr., indudeep chhachhi, and christopher l. brown, all at western kentucky university. in this study, the authors demonstrate how a portfolio of leveraged exchange traded funds (letfs) outperforms a portfolio using traditional etfs while simultaneously reducing downside risk. their results are primarily a function of letfs borrowing short while the investor lends the additional wealth generated from this leverage in 1 to financial services review 28 (2020) v–vii 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. 7year treasury bonds or similar type of assets. they also present that for every 1% earned above the implied borrowing rate, a portfolio of 2x and 3x letfs outperforms a traditional portfolio by 0.41% and 0.63% respectively, they show that more than 90% of letfs outperformance is explained by the borrowing lending differential. the final article, “are multiple share class funds poorly governed?” is coauthored by jonathan handy at furman university and thomas smythe at florida gulf coast university. utilizing independent morningstar stewardship grades, the authors find that multiple share class mutual funds (ms funds) have lower quality governance. using ordered probit regressions they find that ms funds are more likely to have lower board quality ratings and managerial incentive ratings. their results show that less sophisticated investors seeking financial advice (those typically utilizing ms funds) may potentially be directed to funds that underperform and have higher costs. thanks to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. john e. grable tim kaiser sina ehsani travis l. jones michael highfield swarn chatterjee william trainor sarah reiter john clinebell shaun pfeiffer sophie shive yuliya plyakha xu sun gene stout timothy krause john clinebell laszlo sandor anders carlander lewis w. coopersmith kenneth ryack jerry stevens sonya britt robert ottr gary porter kent baker david yeske jean lown scott moore stuart heckman jonathan guyton haiwei chen shibashish chakraborty wade d. pfau jiri sindelar timothy lu tianyang wang kaustav misra daniel huerta-sanchez sandeep singh sara shirley larry frank philippe cogneau greg geisler david perkins james dow stuart heckman hanna lim jian zhou sherman hanna david hulse victoria bryant jeremy burke jeremy clark cathy faulcon bowen hem c. basnet yan liu george korniotis diego escobari harin desilva juan gallardo hal hershfield jason p. berkowitz david nanigian sharon danes vi editorial / financial services review 28 (2020) v–vii please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review viieditorial / financial services review 28 (2020) v–vii what do financial planning organizations communicate to stakeholders and consumers? an empirical narrative analysis wookjae heoa,*, narang parkb, robin henagerc, john e. grabled adepartment of consumer sciences, south dakota state university, swg 149, box 2275a, brookings, sd 57007, usa bdepartment of financial planning, housing, and consumer economics, college of family and consumer sciences, university of georgia, 205 dawson hall, 305 sanford drive, athens, ga 30602, usa cdepartment of business & economics, school of business, whitworth university, weyerhaeuser hall 310c, 300 w. hawthorne road, spokane, wa 99251, usa ddepartment of financial planning, housing, and consumer economics, college of family and consumer sciences, university of georgia, 300 dawson hall, 305 sanford drive, athens, ga 30602, usa abstract this study examined how financial organizations present relevant information to stakeholders and consumers and what differences exist between what organizations intend to deliver and what consumers and stakeholders perceive from the communication channels. using a text mining technique, text data collected from financial planning organizational websites, social media, and news reports were analyzed. the results showed that the financial planning organizations successfully address their own position in the financial planning profession; however, financial planning organizations often fail to communicate specifics about their value to consumers. the findings suggest that financial planning organizations should enhance messaging strategies to promote both the organization and the profession it uniquely serves. © 2018 academy of financial services. all rights reserved. jel classification: l3; l5; y8 keywords: text analysis; cfp board; financial planning association; napfa * corresponding author. tel.: �1-605-688-5835; fax: �1-605-688-5578. e-mail address: wookjae.heo@sdstate.edu (h. wookjae) financial services review 27 (2018) 115-131 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. 1. introduction the success of a firm or an organization is often dependent upon, among other things, consumers’ perceptions of the firm’s products or the organization’s brand identity (kotler and keller, 2006). the outward, or brand, identity of an organization is frequently influenced by the inward cultural identity of the employees and structure of the organization (daft, 2007). as an individual’s identity is often closely associated with his or her culture and related societal expectations, similarly, the culture of an organization can be observed among its stakeholders. this is particularly true for organizations that are comprised of organization members, as members generally share the same vision and philosophy. in addition, the specific identity, or culture, of an organization creates expectations of its members. the existence of a group identity unique to organizations has been demonstrated through selfcategorization theory or what is known as social identity theory (turner, hogg, oakes, reicher, and wetherell, 1987). this theory is supported by numerous empirical studies (e.g., ho and lin, 2016; kunst, thomsen, sam, and berry, 2015; sanchez and vargas, 2016; vezzali, cadamuro, versari, giovannini, and trifiletti, 2015). the theory describes the ways in which individuals identify as members of a group through shared experiences and social categorization. understanding the connection between an organization’s identity and perceptions of the organization among consumers can help an organization communicate (1) the organization’s potential role in the marketplace and (2) the way in which the organization can meet the needs of consumers and other stakeholders. the role of organizational identification is especially important in emerging industries and professions. consider the financial service marketplace. neither the financial planning association (fpa)—the largest professional membership organization for financial planning practitioners— or the certified financial planner board of standards, inc. (cfp board)—the regulatory body for those holding the certified financial planner (cfp) mark— existed before 1969. these organizations have grown to meet the needs and demands of those currently working as financial service professionals. as organizations, the fpa and the cfp board, in addition to other entities and organization operating in the field, have grown very quickly—perhaps so fast that they may not have considered the group identity of stakeholders or specifically paid attention to each organization’s culture as perceived by consumers. this project focused on the communication approaches used by the fpa, the cfp board, and the national organization of personal financial advisors (napfa) as an element of each organization’s attempt to a build group identity within stakeholder communities and cultural awareness among consumers. it is important to note that while sharing many similarities, the three organizations do serve different niches. the cfp board, for example, targets a very broad audience that includes cfp professionals, consumers, and those who are preparing to earn the cfp certification. the fpa and the napfa, as membership organizations, primarily serve the needs of paying members, with an emphasis on using a website and social networks to deliver information to members. as such, the elements of communication across each organization are likely different in purpose, leading to a broad range of messages. because websites and social networks (e.g., facebook) are publicly available, these communication platforms may reach non-stakeholder audiences. it is likely that an individual 116 w. heo et al. / financial services review 27 (2018) 115-131 outside an organization (e.g., consumers and potential stakeholders) may view the content of communications differently based or her or his own perspective. this is particularly true in the field of financial planning. much of the information consumers and potential financial planning clients and other stakeholders hear or read about financial planning comes from the news media (mitchell, gottfried, barthel, and shear, 2016). news today plays an important role in shaping consumers’ opinions. the news media has a profound influence on the way consumers obtain and process information (cho, keum, and shah, 2015). additionally, news helps frame information that can potentially lead consumers to a biased or selective viewpoint (lecheler, bos, and vilengenthart, 2015). for example, news is often framed, through the media, to evoke emotion (kim and cameron, 2011; lecheler et al., 2015; myers, nisbet, maibach, and leiserowitz, 2012). repeated news stories can help consumers form strong beliefs about emotional topics (lecheler, keer, schuck, and hänggli, 2015). in other words, consumers can be influenced by the framing of a topic presented by media sources. conceptually, this is a possibility that organizations need to manage. framing is strongly associated with confirmation bias (bazerman and moore, 2013). confirmation bias describes a situation where a consumer, once influenced by the framing of a question or topic, shows a tendency to hold onto her or his initial belief. once established, these consumer beliefs cannot be easily changed, even with the introduction of additional information. neale and baxerman (1985a, 1985b) suggested that organizations use diverse channels of communication when coping and responding to negative framing and confirmation biases held by consumers. a first step requires organizations to determine if communication efforts (as public assets). match organizational goals. this evaluation must include an assessment of the narratives presented on websites and through social media platforms. within the domain of financial planning, the process of investigating story content about financial planning is one way to evaluate how effective the fpa, the cfp board, and the napfa have been in delivering a cognizant image of financial planning, as well as in establishing a unified identity across organizations. financial planning organizations may find it helpful to obtain an understanding of how well internal messaging and external receipt of messages mesh together. this can be accomplished by analyzing organizational level communication approaches and assessing messages send via media reports, websites, and social media sites. if the messages found on financial planning organizations’ websites and social networks (i.e., messages designed for stakeholder used) are different from the content disseminated in news articles (i.e., what is being perceived by consumer external to an organization), it is possible that negative seeds of framing and confirmation bias could arise among consumers.1 the purpose of this study was to apply large data analysis techniques to compare the content of communications presented by the fpa, the cfp board, and the napfa to estimate the consistency of messaging among these organizations. this study adds to the literature by showing how a relatively popular statistical methodology from marketing can be applied to emerging fields like financial planning. for the purposes of this study, internal message content from the fpa, the cfp board, and the napfa included articles from each organization’s website(s) and social networks. external message content included news reports written by nonaffiliated members of the media as found online. 117w. heo et al. / financial services review 27 (2018) 115-131 2. methodological background researchers have been adopting and refining qualitative or narrative data analysis methods as a social science research tool for over 20 years (ignatow and mihalcea, 2017). the combination of qualitative or narrative data and quantitative analytic models are generally described as mixed methodology techniques (teddlie and tashakkori, 2008). although rarely applied to the analysis of financial planning data, narrative qualitative data have been quantitively analyzed widely in other disciplines. it is most common for researchers to collect narratives from newspapers when attempting to analyze narrative contexts with quantitative analytic methods. for instance, cerulo (1998) analyzed newspaper headlines by evaluating the sequential narrative organization between victims and perpetrators. his findings showed that narrative analysis can be used to identify patterns of organization not generally seen when traditional quantitative data analysis techniques are used. franzosi, de fazio, and vicari (2012) investigated newspaper archives and by using a mixed-methods technique to analyze newspaper accounts of racial lynching. similar to the work of cerulo et al., were able to uncover patterns in the data that were not apparent using traditional analytic techniques. sudhahar, franzosi, and cristianini (2011) evaluated new york times articles from 1987 to 2007 as a way to identify criminal victims’ profiles. a common theme among these studies is that the use of mixed-methodologies, including narrative data mining, has been and can be adopted through the social sciences to gain a deeper understanding of shared identities and cultures. this approach to data analysis—sometimes called big data analysis—allows for a more nuanced assessment of data, which often leads to a better understanding of systematic relationships within datasets (mische, 2014; roberts, 2008; tausczik and pennebaker, 2010). the basic underlying laws of text analysis come from zipf’s law and heaps’ law (ignatow and mihalcea, 2017). using zipfs’ law, it is possible to obtain an estimate of the distribution of words in a corpus. utilizing heaps’ law, it is possible to predict/generalize a specific word’s frequency in a(n) expected/given corpus. eq. (1) represents zipf’s law. eq. (2) denotes heaps’ law. f � r � k . . . (1) where, f is the frequency of a word in a corpus; r denotes a word’s rank, which is rth most frequent word in a corpus; and k means a constant value of each corpus. v � kn�, 0 � � � 1 . . . (2) where, k is a parameter that is changed by a corpus; n denotes a corpus’ word count; � is the parameter of a specific word; and v is a targeted word to predict/generalize from a corpus with n words. 2.1. research questions and methodology this research study was designed to answer two broad questions: (1) what do the organizations that represent practicing financial planners believe is important when messag118 w. heo et al. / financial services review 27 (2018) 115-131 ing in public domains, and (2) what do consumers, through a media lens, think and know about the emerging profession of financial planning and financial planners? based on these overarching questions, the following specific research questions were examined: 1. using a financial planning organization’s unique communication strategy, how does the organization present relevant information to its immediate stakeholders? 2. how do consumers and outside stakeholders view each organization and its members based on news stories? 3. what, if any, differences exist between and among what organizations are presenting through online messaging and what consumers, through the media, have identified as important within a financial planning context? two analyses were utilized in this study to answer these questions. first, this study collected and analyzed textual data from the three organization’s websites, facebook, and news articles. after collecting the textual data, a comparison of the most frequent key words used by each organization was made. second, this study utilized network visualizations to identify relationships within and among the textual data. based on the foundational algorithm developed by fruchterman and reingold (1991), this study used a visualization technique using the statistical program r (epskamp, cramer, waldorp, schmittmann, and borsboom, 2012). the relationship visualization approach created an environment in which each organization’s communication approach, through the use of words and textual images, could be compared and matched to what consumers were searching for at the time of the survey. matches, as well as mismatches, in content emerged as a primary finding in this study. a process of text mining was used to collect textual data and to analyze the following websites: (1) cfp board (http://www.cfp.net), (2) fpa (https://www.onefpa.org), and (c) napfa (https://www.napfa.org). for the fpa and the napfa websites, the public version of each organization’s website was used for analysis because the purpose of this study was to investigate the communicative information exhibited by organizations as perceived by consumers (attard and coulson, 2012; haigh and jones, 2005). given that the cfp board website did not have a consumer only link, the entire site was evaluated. text data from the three organizational websites was used to analyze how financial planning organizations position their message content to the public. specifically, introductory and explanatory paragraphs were collected from each organization’s website on february 26, 2017. however, specific linked document files and linked websites (e.g., information that came as a pdf file) were excluded from the text analysis because this type of information required software that may not have been accessible to all consumers at the time of the analysis. across the three websites a total of 47,782 words were analyzed: 29,506 words from the cfp board webpages, 10,387 words from the public fpa webpages, and 7,889 words from the public napfa webpages. the text mining technique was also applied to facebook postings. at the time of the study, each organization maintained its own facebook site. given the nature of facebook, text data obtained through the site was used as an indicator of each organization’s social network strategy. as of 2018, facebook users totaled more than 2.0 billion individuals. in addition to the large number of users, facebook allows for asynchronous communication providing access to users regardless of time and location (cava, 2014). as such, facebook is often used 119w. heo et al. / financial services review 27 (2018) 115-131 as a public relations tool. for instance, in 2013, it was reported that 15 million businesses, companies, and organizations utilized facebook to communicate with consumers and potential clients (koetsier, 2013). to analyze the facebook messages posted by each organization, all facebook messages from january 1, 2016 to february 17, 2017 were collected as of february 27, 2017. the total number of messages analyzed included: 116 messages from fpa, 62 from cfp board, and 431 messages from napfa. text from news articles pertaining to each organization were collected as a way to compare each organization’s communication approach to what has been published in the media regarding consumer demand for financial planning. text data were collected using google’s search engine and two key words: financial planning and financial services. because of various usages in the media, some key words (e.g., financial advising and financial counseling) were excluded from the analysis. thousands of articles and reports were obtained for the period january 1, 2016 through february 17, 2017. only those postings that were publicly and legally available were included in the analysis. text data sources included: bloomberg, boston business journal, business insider, cbs news, chicago tribune, cnbc, forbes, fortune, harvard business review, investment news, investopedia, national public radio, u.s.a. today, the wall street journal, washington post, and washington times. in total, 127 news articles were included in the study. these were primarily used to examine how the general public perceived financial planning and financial services in general. r-studio, an open source analytical tool, was utilized for the text mining analysis. this program was used to synthesize text data from each organization’s website, facebook, and media links. the tool was also used to create visual maps of the text data. for the analysis, unrelated and meaningless words were excluded from the study. words such as have, best, can, care, don’t, get, help, http, just, like, many, may, much, new, offers, one, only, people, print, said, says, take, use, want, and will were excluded from the analysis. 3. results it is not surprising that the three organizations expressed a different core competency (i.e., mission) in their public facing messaging. table 1 shows how the three organizations differed in terms of communicating their core uniqueness within the financial planning profession. 3.1. website analysis results figs. 1, 2, and 3 illustrate the output from the textual data mining technique. specifically, the figures display the words that were emphasized on each organization’s website. these words can be viewed as each organization’s intentional introduction. in line with social identity theory (turner et al., 1987), each organization’s identity can be explained by the text narrative presented to the public. in effect, text acts as an indicator of interactive behaviors in a social context. as shown in figs. 1, 2, and 3, the thickness of the lines represent the frequent combination of words. in other words, two keywords linked with a thick line 120 w. heo et al. / financial services review 27 (2018) 115-131 denotes a strong text association. on the other hand, a weak or no line between two keywords indicates a weak or nonexistent relationship between two keywords. on the fpa website (fig. 1), the organization emphasized the value of membership, connection among members, business success as a financial planner, and professional development. fig. 1 highlights words about local organizations and connections. for instance, “financial,” “fpa,” “planning,” “certified,” “members,” “organization,” and “media” were strongly connected on the fpa website. these words were frequently mentioned in the same context. considering that fpa emphasizes membership among professionals, these frequent word combinations appear to both appropriate and well utilized. however, some key words had weak connections on fpa’s website. for instance, “policy” stood apart from keywords like “profession,” “advocacy,” and “standards,” implying that fpa’s membership focus was overshadowing the organization’s connections to professional discussions about policy. the cfp board website (fig. 2) focused more intently on educational terms (e.g., resources, exams, and centers). this is consistent with the purpose of the organization. it is not surprising that fig. 2 emphasizes words referring to education and exams. for instance, the following educational keywords were found to be major components on the website: “cfp,” “certification,” “board,” “exam,” and “standard.” along with educational keywords, the following keywords were mentioned together: “professional,” “financial,” and “resources.” these keywords were used as cues to clearly communicate educational resources for website visitors. the keyword associations appeared consistent with the cfp board’s focus on managing all aspects related to the cfp marks. however, connections between and table 1 top 20 most frequent words used on the fpa, cfp board, and napfa websites (single words) fpa frequency cfp board frequency napfa frequency fpa 83 cfp 177 advice (se, ser, sor) 40 financial 74 certification 60 financial 26 planning 43 professional(s) 61 planning 22 profession(al) 31 financial 59 napfa 15 members 18 board 56 issues 10 cfp 17 planning 40 find 6 planner 13 exam 29 client 5 organization 12 find 25 members 5 advocacy 10 standards 23 policy 5 board 10 career 22 professional 5 certified 10 learn 22 public 5 denver 9 education 21 advanced 4 education 9 center 20 contract 4 media 9 resources 19 national 4 national 9 account 18 personal 4 policy 8 certified 15 plan 4 standards 8 create 14 retirement 4 conduct 7 become 12 tips 4 facebook 7 july 12 view 4 goals 7 news 11 apr 3 fpa � financial planning association; cfp board � certified financial planner board of standards, inc.; napfa � national organization of personal financial advisors. 121w. heo et al. / financial services review 27 (2018) 115-131 among some keywords were found to be weak. for instance, “career” was not fully integrated with other key words, such as “professional,” “exam,” and “resources.” this may indicate that the cfp board was attempting to disassociate or minimize career information as it relates to the cfp board. the napfa website was unique in its focus and messaging. the primary thrust of the site was focused on perspective concepts. because napfa’s membership consists of fee-only financial planners (i.e., those who do not accept commissions for services or for the implementation of recommendations), the extensive use of the term “financial planning” was expected, compared to educational or professional development phraseology. fig. 3 shows the key words found on the website and the linkages among the terms. as illustrated, major keywords on the website were “financial,” “advisors,” “planning,” “napfa,” “issues,” and “members.” comparing these keywords with the two other websites, napfa’s website exhibited more generalized financial planning terms. on the other hand, some key words such as “tips,” “policy,” “public,” and “contact” tend to be used in isolation, with fewer connections to other keywords. this can be interpreted to mean that the napfa site lacked content variety. in summary, it was determined that fpa’s messaging was focused on connections among financial planners. the cfp board’s messaging focused more on educating financial planners, while napfa was narrowly focused on promoting the concept of financial planning. no meaningful focus on consumers was evident in napfa’s messaging. in the case of the fpa and cfp board websites, combinations of major keywords indicated each website’s emphasis. for instance, “financial,” “fpa,” “planning,” “certified,” “members,” and “orgafig. 1. linkages among the most frequent words found on the financial planning association (fpa) website. 122 w. heo et al. / financial services review 27 (2018) 115-131 nization” were strongly linked with membership terms on the fpa website. on the cfp board website, “cfp,” “certification,” “board,” “exam,” and “standard” were used together to emphasize educational information. however, in the case of napfa, general terms (e.g., “financial,” “advisor,” and “planning”) were used on the website. 3.2. facebook analysis results evaluating the words and phrases used on each organization’s official facebook page showed that facebook was used primarily as a consumer interface. as shown in table 2, each organization used facebook to share information about financial planning in general. for instance, keywords such as “financial,” “news,” and “tips” were common across the pages. the three organizations posted messages from outside news sources, but the news sources were slightly different. in the case of fpa, news posted primarily originated from outside news sources like reuters.com, governmental notices (e.g., ssa.gov), and cnbc.com. in the case of the cfp board, postings generally came from within the organization (e.g., cfp.net and centerforfinancialplanning.org). similarly, in the case of napfa, news postings came from writers associated with the organization (e.g., napfa.org and figuide.com). figs. 4, 5, and 6 show that the use of facebook by each organization mirrored, in general terms, what was presented on each organization’s website. overall, the three organizations fig. 2. linkages among the most frequent words found on the certified financial planner board of standards, inc. (cfp board) website. 123w. heo et al. / financial services review 27 (2018) 115-131 used facebook to deliver news to consumers; however, the news was similar to the content available on each organization’s website. within the fpa’s facebook pages (fig. 4), the organization attempted to promote useful information to visitors rather than organizational membership benefits. fig. 4 highlights the types of words used to highlight the informational content of social media postings. for instance, “financial,” “google,” “read,” “help,” “survey,” “article,” and “advisors” were strongly connected throughout the fpa facebook pages. even so, the facebook keywords tend to mirror keywords found on fpa’s website. state another way, fpa’s facebook postings tended to be a professional resource rather than consumer friendly documentation. similar to the organization’s website, the cfp board facebook pages (fig. 5) focused more intently on educational terms (e.g., resources, exams, and centers). for instance, prominent educational keywords were: “cfp,” “center,” “board,” “planning,” and “learn.” it was surprising, however, that less emphasis on the cfp exam was found throughout cfp board’s facebook pages. keywords like “university,” “career,” and “academic” were disconnected from other keywords. fig. 6 shows the key words found on napfa’s facebook account and the linkages among the terms. major keywords were “financial,” “advisors,” “planning,” “napfa,” “issues,” and “members.” similar to the organization’s website, napfa’s facebook pages delivered limited communication about the value of working with a financial planner. facebook content tended to be lacking in variety. the analysis of the 126 news articles about financial planning obtained using google.com revealed what consumers, through the media, perceive when thinking about “financial fig. 3. linkages among the most frequent words found on the national organization of personal financial advisors (napfa) website. 124 w. heo et al. / financial services review 27 (2018) 115-131 planning.” as shown in table 3, the most frequent words associated with financial planning were “financial,” “advice,” “advisor(er),” “planning,” “plan,” and “retirement.” this implies that the media tends to, when responding to consumer questions and interests, relate financial planning to very specific elements associated with the financial planning process. a potential mismatch between the messaging that was sent by the three organizations and what consumers were searching for was evident. organizational messaging focused primarily on introducing each organization’s unique fit within the field of financial planning. further, the emphasis was focused on the benefits of membership or affiliation rather than on positive outcomes for consumers. as shown in fig. 7, the analysis indicated that recent news articles dealt with more practical issues related to personal finance topics. for instance, the linkages among “financial,” “service,” “money,” “plan,” and “clients” were very strong. in addition, several daily life terms, such as “life,” “income,” “savings,” and “firm,” emerged as strongly connected. very few of these words emerged as important any of the three organization’s websites or facebook offerings. these findings indicate that what was occurring in the media, which is often based on consumer demand, tells a different story from what was being communicated internally and externally among the leading financial planning organizations. 4. conclusion and discussion the purpose of the study was to address the following questions: table 2 top 20 most frequent words used on the fpa, cfp board, and napfa facebook sites (single words) fpa frequency cfp board frequency napfa frequency google 81 cfp 69 napfa 374 financial 61 financial 49 financial 127 advisor(s) (ers) 39 planning 36 advisor(s) 106 retirement 18 board 29 member(s) 96 read 16 learn 20 planning 80 fpa 14 center 12 register 70 survey 14 certification 11 conference 64 article 12 career 8 figuide 58 clients 12 consumer 7 post 55 full 12 academic 6 fiduciary 44 money 11 arc 6 rule 42 planning 11 director 6 fee only 41 time 11 keller 6 career 40 years 11 kevin 6 changes 38 research 10 professionals 6 clients 38 college 9 today 6 free 38 know 9 university 6 practice 33 learn 9 advice 6 earn 30 life 9 ceo 5 business 29 plan 9 colloquium 5 cfp 29 fpa � financial planning association; cfp board � certified financial planner board of standards, inc.; napfa � national organization of personal financial advisors. 125w. heo et al. / financial services review 27 (2018) 115-131 1. using a financial planning organization’s unique communication strategy, how does the organization present relevant information to its immediate stakeholders? 2. how do consumers and outside stakeholders view each organization and its members based on news stories? 3. what, if any, differences exist between and among what organizations are presenting through online messaging and what consumers, through the media, have identified as important within a financial planning context? a text mining methodology was used to address these questions using data from the fpa, the cfp board, and the napfa. first, it was determined that each organization used its website and facebook page to communicate a unique mission and value proposition to potential members. while some effort was taken to promote financial planning in general, the clear focus was on promoting each organization’s strengths. second, it was found that consumers, through the media, were less interested in the benefits of organization membership, or even the value added from working with a professional affiliated with one of the organizations. instead, consumers were looking for very applied information about financial planning topics and questions. third, it was apparent from the different analyses that a mismatch existed between what the organizations were messaging and what consumer and those in the media were writing about. this last point should not be inferred to have a value implication. each of the organizafig. 4. linkages among the most frequent words found on the financial planning association (fpa) facebook site. 126 w. heo et al. / financial services review 27 (2018) 115-131 tions serves a unique membership group within the financial planning profession. serving the needs of members is the primary way in which these organizations generate revenue, so it makes sense that messaging would focus on promoting membership and affiliation benefits. however, it did appear that a significant benefit was being under-messaged; namely, the value of financial planning for households and consumers. rather than being a source for information that might prompt consumers to seek out an organization’s members, consumers would find little to meet their needs on any of the three sites. stated another way, it appeared that the organizations were messaging a set of cultural values that may have been different from what those in the media and consumers were looking for. as described in social identity theory, professionals and consumers often hold diverse (and sometimes divergent) cultural aspirations, which can differ by state and region (osoba, 2009; woodard, 2011). a best organizational practice involves being attuned to the various cultures represented by stakeholders and consumers of products and services. this includes understanding and anticipating biases and conflicting interests. for instance, it can be beneficial to compare what is being presented nationally to what is appropriate at the local or regional level. additionally, outward marketing tools, such as facebook messaging, needs to match what consumers and those in the media are searching for, rather than on reproducing what is already in the media or available on an organization’s website. without this focus, consumers and those in the media will search elsewhere for information. when this happens, fig. 5. linkages among the most frequent words found on the certified financial planner board of standards, inc. (cfp board) facebook site. 127w. heo et al. / financial services review 27 (2018) 115-131 table 3 top 20 most frequent words identified in media reports key word frequency financial 1,514 advice, advisor(s), adviser(s) 1,205 planning, plan 604 retirement 579 money 401 year(s) 401 clients 338 save, savings 269 time 256 investment 237 services 229 life 182 income 170 fiduciary 166 rule 166 insurance 165 accounts 152 firms 145 fee 144 interest 144 pay 144 fig. 6. linkages among the most frequent words found on the national organization of personal financial advisors (napfa) facebook site. 128 w. heo et al. / financial services review 27 (2018) 115-131 the value proposition associated with being a member of an organization falls.2 this also provides a platform for other organizations and firms to define keywords and shape the way consumer perceive financial planning. a key takeaway from this study is that financial planning organizations should consider their messaging strategies holistically and investigate the differences among various messaging techniques. the use of text data mining methodologies can be utilized for this purpose. with today’s computing power, it is possible to compare various messaging and ongoing dialogs in a way that provides timely reports based on mixed-methods data analysis techniques. in conclusion, while each financial planning organization was successful in communicating its own position in the financial planning profession, the organizations, individually and jointly, failed to communicate specifics about their value to consumers. given the increasingly interconnected mechanisms of the internet, this approach to communication messaging may be counterproductive in terms of each organization’s goal: to increase organization membership through a clearly defined valued proposition. if an organization’s messaging fails to resonate with consumers who were likely to use the services of an organization’s members, consumers may search for services elsewhere. thus, enhancing communication in a way that meets consumers’ needs outside of an organization may create a more robust channel to promote both the organization and the profession it serves. notes 1 framing and confirmation bias occur when a general belief held among consumers about something does not match the real identity of an object. 2 the cfp board launched the public awareness campaign in april 2011 to increase consumer awareness of the cfp certification and the financial planning profession. the campaign included television, radio, and online advertising, as well as public relations via social media (cfp board). over the past few years, the board’s efforts have contributed to raising the overall brand awareness (i.e. 79% to 85%), preference fig. 7. linkages among the most frequent words found in media reports. 129w. heo et al. / financial services review 27 (2018) 115-131 (i.e. 22% to 52%), and willingness to use a financial professional (i.e. 30% to 52%); however, consumers’ understanding of the uniqueness and quality of the cfp are still only 43% and 44%, respectively (cfp board, 2015). this suggests that the disparity of the two perspectives (i.e. internal and external) described in this study may originate from misperceptions rather than unfamiliarity with financial planning professionals. references attard, a., & coulson, c. (2012). a thematic analysis of patient communication in parkinson’s disease online support group discussion forums. computers in human behavior, 28, 500–506. bazerman, m. h., & moore, d. a. (2013). judgement in managerial decision making (8th ed.). hoboken, nj: wiley & sons, inc. cava, m. (2014, february 2). gamechanger. facebook turns 10. usa today. (available at https://www. usatoday.com/story/tech/2014/02/02/facebook-turns-10-cultural-impact/5063979/) cerulo, k. a. (1998). deciphering violence: the cognitive structure of right and wrong. new york, ny: routledge. cfp board. (2015). 2015 annual brand tracking report: executive summary. (available at http://www.cfp.net/ news-events/latest-news/2015/05/19/awareness-of-cfp-certification-leaps-ahead) cho, j., keum, h., & shah, d. v. (2015). news consumers, opinion leaders, and citizen consumers: moderators of the consumption–participation link. journalism & mass communication quarterly, 92, 161–178. http:// dx.doi.org.10.1177/1077699014554766 daft, r. l. (2007). understanding the theory and design of organizations. mason, oh: thompson southwestern. epskamp, s., cramer, a. o. j., waldorp, l. j., schmittmann, v. t. d., &, borsboom, d. (2012). qgraph: network visualizations of relationships in psychometric data. journal of statistical software, 48, 1–18. http://dx. doi.org/10.18637/jss.v048.i04 franzosi, r., de fazio, g., & vicari, s. (2012). ways of measuring agency: an application of quantitative narrative analysis to lynchings in georgia (1875–1930). sociological methodology, 42, 1–42. fruchterman, t. m. j., & reingold, e. m. (1991). graph drawing by force-directed placement. software – practice & experience, 21, 1129–1164. http://dx.doi.org/10.1002/spe.4380211102 haigh, c., & jones, n. (2005). an overview of the ethics of cyber-space research and the implications for nurse educators. nurse education today, 25, 3–8. ho, t., & lin, y. (2016). the effects of virtual communities on group identity in classroom management. journal of educational computing research, 54, 3–21. http://dx.doi.org/10.1177/0735633115611645 ignatow, g., & mihalcea, r. (2017). an introduction to text mining research design, data collection, and analysis. thousand oaks, ca: sage publication, inc. kim, h. j., & cameron, g. t. (2011). emotions matter in crisis: the role of anger and sadness in the publics’ response to crisis news framing and corporate crisis response. communication research, 38, 826–855. http://dx.doi.org/10.1177/0093650210385813 koetsier, j. (2013). facebook: 15 million businesses, companies, and organizations now have a facebook page. san francisco, ca: venturebeat. kotler, p., & keller, k. l. (1988). (2006). marketing management (12th ed.). upper saddle river, nj: prentice-hall. kunst, j. r., thomsen, l., sam, d. l., & berry, j. w. (2015). ‘we are in this together’: common group identity predicts majority members’ active acculturation efforts to integrate immigrants. personality and social psychology bulletin, 41, 1438–1453. lecheler, s., bos, l., & vilengenthart, r. (2015). the mediating role of emotions: news framing effects on opinions about immigration. journalism & mass communication quarterly, 92, 812–838. http://dx.doi.org/ 10.1177/1077699015596338 130 w. heo et al. / financial services review 27 (2018) 115-131 lecheler, s., keer, m., schuck, a. r. t., & hänggli, r. (2015). the effects of repetitive news framing on political opinions over time. communication monographs, 82, 339–358. mische, a. (2014). measuring futures in action: projective grammars in the rio�20 debates. theory & society, 43, 437–646. mitchell, a., gottfried, j., barthel, m., & shearer, e. (2016). pathway to news. pew research center. (available at http://www.journalism.org/2016/07/07/pathways-to-news/) myers, t. a., nisbet, m. c., maibach, e. w., & leiserowitz, a. a. (2012). a public health frame arouses hopeful emotions about climate change: a letter. climatic change, 13, 1105–1112. http://dx.doi.org/10.1007/s10584012-0513-6 neale, m. a., & bazerman, m. h. (1985a). perspectives for understanding negotiation: viewing negotiation as a judgmental process. journal of conflict resolution, 29, 33–55. neale, m. a., & bazerman, m. h. (1985b). the effects of framing and negotiator overconfidence on bargaining behaviors and outcomes. academy of management journal, 28, 34–49. osoba, b. (2009). culture and entrepreneurial activity in the united states: a quantitative analysis. innovation: the european journal of social sciences, 22, 341–370. http://dx.doi.org/10.1080/13511610903367820 roberts, c. w. (2008). the fifth modality: on languages that shape our motivations and cultures. leiden, netherlands: brill publishers. sanchez, g. r., & vargas, e. d. (2016). taking a closer look at group identity. political research quarterly, 69, 160–174. http://dx.doi.org/10.1177/1065912915624571 sudhahar, s., franzosi, r., & cristianini, n. (2011). automating quantitative narrative analysis of news data. jmlr: workshop and conference proceedings, 17, 63–71. tausczik, y. r., & pennebaker, j. w. (2010). the psychological meaning of words: liwc and computerized text analysis methods. journal of language and social psychology, 29, 24–54. teddlie, c. b., & tashakkori, a. m. (eds.). (2008). foundations of mixed methods research: integrating quantitative and qualitative approaches in the social and behavioral sciences. thousand oaks, ca: sage publication, inc. turner, j. c., hogg, m. a., oakes, p. j., reicher, s. d., & wetherell, m. s. (1987). rediscovering the social group: a self-categorization theory. oxford, uk: basil blackwell. vezzali, l., cadamuro, a., versari, a., giovannini, d., & trifiletti, e. (2015). feeling like a group after a natural disaster: common ingroup identity and relations with outgroup victims among majority and minority young children. british journal of social psychology, 54, 519–538. woodard, c. (2011). american nations: a history of the eleven rival regional cultures of north america. new york, ny: viking. 131w. heo et al. / financial services review 27 (2018) 115-131 financial, demographic, and psychological differences between chapter 13 bankruptcy filers and non-filers scott e. kehiaiana, albert a. williamsb,*, carolyn l. birdc asouthern new hampshire university, 2462 mountain lake road, asheboro, nc 27205, usa bnova southeastern university, 3301 college avenue, ft. lauderdale, fl 33314-7796, usa cnorth carolina state university, 512 brickhaven drive, raleigh, nc 27695-7606, usa abstract this study finds financial, demographic, and psychological differences between chapter 13 filers and non-filers. financial training reduces the likelihood of filing for personal bankruptcy. males are twice as likely as females to be filers. blacks are twice as likely as whites to be filers. a single person is 38% less likely to file than a married person. homeowners are five times as likely as renters to be filers. increases in education, religious commitment, and parents’ income reduce the likelihood of filing. increases in the psychological factors, self-efficacy, locus of control, and self-control, reduce the likelihood of filing for chapter 13 bankruptcies. these results can be used to influence public policy to reduce personal bankruptcy. © 2021 academy of financial services. all rights reserved. jel classification: d12; d14 keywords: chapter 13 bankruptcy filers; non-filers; demographic differences; psychological differences; financial literacy i. introduction personal bankruptcy occurs frequently and is a significant problem in the united states. evans and bauchet (2017) state that each year, hundreds of thousands of u.s. households choose to file for bankruptcy and accept the longer-term effects that bankruptcy has on their credit reputation over the challenges of dealing with collectors and/or creditors. according corresponding author. tel.: +1-954-262-5286; fax: +1-954-262-3974; e-mail address: albewill@nova.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 67–84 to the u.s. courts, in 2019, there were 752,160 non-business bankruptcies (chapters 7, 12, and 13) in the united states, with 281,702 (38%) classified as chapter 13. during the same period, there were 13,092 non-business bankruptcies in north carolina (state for this study), with 7,960 (61%) being classified as chapter 13 (http://www.uscourts.gov/statistics-reports/ caseload-statistics-data-tables). bankruptcy can occur because of mismanagement, lifestyle, consumption patterns, economic conditions, or unpredictable misfortunes such as unexpected medical expenses and layoffs. there are different ways to deal with insolvency, including reduction of spending, consolidation of loans, obtaining consumer credit counseling, surrendering of collateral, relocation to lower costs areas, and/or filing for bankruptcy. filing for bankruptcy reduces financial stress and provides a financial lifeline. even though filing for bankruptcy is a lifeline, it is still a difficult and major financial decision with long-lived impact on a consumer’s credit profile. the motivation for this study is to find out more about the factors that correlate with this major financial decision. another motivation is the limited research done on the differences between chapter 13 bankruptcy filers and non-filers. this study expands the literature in this area. we look for relationships between demographic, psychological, and financial variables and filing for bankruptcy. we have not seen psychological variables utilized in bankruptcy filings research. in addition, we include two variables not seen in the literature. they are parents’ income, and parents’ education, representing intergenerational relationships. this study uses data from chapter 13 bankruptcy filers in north carolina and compares them to people in the state who have not filed for bankruptcy. the results from this study can be used by institutions, like the u.s. courts, to reduce the number of personal bankruptcies in the state of study or the country. for example, we find that males are twice as likely as females to be filers, so financial literacy intervention programs can be targeted more to males to reduce their likelihood of filing for bankruptcy. 2. literature review 2.1. theoretical foundation several theoretical models are used to explain personal financial behavior. miller, levin, whitaker, and xu (1998) discuss sequential and life events models that can be used to explain financial behavior. samuelson (1937) develops the discounted utility model where future cash flows are discounted and utilized to make financial decisions. efrat (1998) advocates for the adoption of a multiutility model that includes two distinct sources of utility that shape the individual’s behavior—pleasure and morality. the human ecological model explains financial decisions in the context of the ecological environment, which is separated into four systems (the microsystem immediate family and friends, the mesosystem immediate family and friends and other systems; the exosystem groups and institutions influencing microsystem; and the macrosystem inclusion of all systems) (bronfrenbrenner, 1979). deacon and firebaugh (1988) use the human ecological 68 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 model and systems theory to provide a context for understanding the goal-directed behavior of families, using inputs, throughputs, and outputs. this sequential managerial process, outlined by deacon and firebaugh, is similar to the financial planning process recommended by the certified financial planners board of standards (establish goals, gather data, analyze information, develop a plan, implement the plan, and monitor progress toward the goal; schuchardt et al., 2007). thaler and shefrin (1981) develop the theory of self-control that can also explain financial behavior. this theory suggests that individuals have personality traits to be either a planner who is concerned with lifetime utility or a doer who is focused on the present. later, shefrin and thaler (1988) proposed the behavioral life cycle hypothesis suggesting that individuals practice mental accounting, meaning that they have different propensities to save in different categories of accounts. schuchardt et al. (2007) provide the trans-theoretical model of behavior change (ttm) that can be used to explain financial behavior. the stages of change are precontemplation (no intention to change behavior), contemplation (aware of problem but not committed to changing behavior), preparation (intending to change within a month), action (changing the problem behavior by employing a variety of strategies), and maintenance (working to prevent relapse). research has shown that successful self-changers use a variety of strategies to achieve their goal. in terms of filing for bankruptcy, the theory gives some structure of what a filer may be going through, from not wanting to file, contemplating to file, filing, receiving financial training (using strategies to change behavior), and hopefully not relapsing into a second filing. we have reviewed the sequential model, the live event model, the discounted utility model, the human ecological model, the theory of self-control, and the trans-theoretical model of behavior change. these different models and theories can be used to explain financial behavior in general. we have not seen any theory that directly explains personal bankruptcy behavior. 2.3. personal bankruptcy 2.3.1. demographic and financial factors. domowitz and sartain (1999) find that medical and credit card debt are the strongest contributors to bankruptcy, with homeownership playing an important role with respect to both the decision to declare bankruptcy and the alternative choice of bankruptcy (e.g., debt consolidation). chakravarty and rhee (1999) find several factors affecting an individual’s decision to file for bankruptcy. these factors include age of the head of household; past problems with money management; the gender of a single head of household; the (un)employment status of the household head; the length of employment of the household head; (bad) health; (the lack of) medicare/medicaid protection; household income; and the dollar benefit level of filing for bankruptcy. clements, johnson, michelich, and olinsky (1999) study 60 people who file for bankruptcy in southern ohio federal district bankruptcy court. their results reveal that the reason most often cited for filing is overuse of credit for clothing, household goods, paying bills, and cash advances. they find that most filers are embarrassed about their situation and have desires to learn about setting goals and about the difference between wants and needs. s. e. kehiaian et al. / financial services review 29 (2021) 67–84 69 loibl, hira, and rupured (2006) study the difference between first-time versus repeat filers, using a sample of 489 participants from georgia. results indicate that repeat filers are more likely than first-time filers to start an emergency fund, to reduce spending, and to write a spending plan. evans and lown (2008) also study predictors of chapter 13 completion rates. the completion of a chapter 13 repayment plan is not associated with a debtor’s monthly income or expenses. the factors, never married, having dependent children, having a previous filing, and having a higher mortgage arrears, increase the likelihood of dismissal from the program. caputo (2008) study marital status and other factors associated with personal bankruptcy using data from 1986 to 2004 from the national longitudinal survey. in 2004, those who are divorced are most likely to have declared bankruptcy (16.4%), followed by those who are separated (13.9%), married with spouse present (11.2%), and never-married (7.0%). marital status is associated with likelihood of declaring bankruptcy in only six of the 14 survey years. never-married persons at the time of declared bankruptcy are less likely than married persons to declare. formerly married persons, whether divorced or separated, are more likely than married persons to declare for bankruptcy. beck, hackney, hackney, and mcpherson (2014) find that religion is the driving force behind the abnormally high level of chapter 13 filings in the southern united states. lefgren and mcintyre (2009) look at cross-state differences in bankruptcy rates. using zipcode level demographic research data, they find that the major differentiating factors include wage garnishment restrictions and the frequency of chapter 13-style bankruptcy claims. zhu (2011) studies household consumption and personal bankruptcy in delaware using data from 2003. she finds that household expenditures on durable consumption goods, such as houses and automobiles, contribute significantly to personal bankruptcy filings. also, medical conditions lead not only to personal bankruptcy filings, but to other adverse events, such as divorce and unemployment. williams, kehiaian, and bird (2017) find significant differences in financial actions between chapter 13 bankruptcy filers and non-filers. non-filers do the following significantly more often than filers: pay their bills on time; avoid living paycheck-to-paycheck; review their credit more frequently per year; pay more than the minimum on their credit card; save money to prepare for home ownership; track their expenses before budgeting; review their total financial situation; use cash for all purchases; evaluate their insurance needs; understand the true cost of credit; and have more knowledge of the components that make up their credit score. fisher (2019) finds that bankruptcy filers are middle income, more likely to be divorced, more likely to be black, more likely to be veterans, less likely to be immigrants, and more likely to have only a high school degree or some college education. filers are more likely to be employed. the bankruptcy population is aging faster than the u.s. population as a whole. individuals are likely to get divorced in the years before bankruptcy and then remarry. he also finds that income falls before bankruptcy and rises after bankruptcy. 2.3.2. psychological factors. danes, casas, and boyce (1999) study the impact of a financial planning curriculum on self-efficacy and find a significant increase in confidence in managing money after taking the curriculum. asaad (2015) finds that financial confidence is a critical component of financial literacy and financial behavior. 70 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 perry and morris (2005) find that locus of control impacts financial decisions. they find that consumers’ propensity to save, budget, and control spending depends partly on their level of perceived control over financial outcomes. miotto and parente (2015) study the level of control in brazilian households and find that different levels of control are linked to various levels of exercising discipline in their spending habits. in summary, these studies have found several demographic, and financial variables related to bankruptcy. research on psychological variables and bankruptcy are not seen in the literature. 3. data and methodology 3.1. data and sample selection primary data are collected in the middle district of north carolina (nc) and include the cities of winston-salem, greensboro, and durham. the sample has 559 participants, with 314 chapter 13 filers and 245 non-filers. the sample of chapter 13 filers is collected by one of the authors, who is an experienced trainer of chapter 13 bankruptcy filers in north carolina. he asks the filers to complete the financial literacy quiz and the questionnaire at the beginning of the training sessions. this is done voluntarily. the data on filers are collected based on convenience sampling. however, we are confident that the data represent the filers adequately. we compare the sample with the population. in terms of race, our sample of filers and non-filers has 55% whites and 40% blacks. winston-salem has 56.3% whites and 34.8% blacks; durham has 50.91% whites, and 39.46% blacks; and greensboro has 57.03% whites and 36.03% blacks (https://www. census.gov/quickfacts/fact/table/). in terms of education, both filers and non-filers have high levels of education above high school (94.6%, and 97.6%, respectively) with the state having 88.2%. as important, the bankruptcy administration enrolls filers in the financial literacy workshops based on the day of their creditors’ meetings. no other grouping criteria, like age, gender, or race, are used. the data for the non-filers are also collected by the same author. he samples a wide cross section of participants from the same area. he uses an online survey created on survey monkey. the survey is advertised at various cold stone creamery stores, and at steals and deals, an online advertising firm selling a wide assortment of goods and services, in the middle district of north carolina. this publication is distributed widely across the region studied. in addition, letters are sent to churches, and schools in the region. there are 142 non-filers who complete the survey online. the remaining non-filers include 34 from the first pentecostal church, 33 from the cathedral of faith, 20 from the neighbor’s grove school, and 17 other individuals, all from the middle district of nc. those non-filers who complete the survey receive a $15 coupon for cold stone creamery and a complimentary financial management class. the data for non-filers are collected across a wide base to replicate the population. for example, a wide cross-section of the population uses the steals and deals online advertising publication. s. e. kehiaian et al. / financial services review 29 (2021) 67–84 71 the data collected include two parts: a questionnaire and a financial literacy quiz. the questionnaire includes sections for demographic data, psychological data, and financial data. the variables in the questionnaire are measured using ordinal or categorical scales. a likert scale (1 = least to 5 =most) is used for the ordinal variables. the financial literacy quiz includes 63 multiple choice questions on a wide range of personal finance topics. some of the topics covered include financial goals, credit card usage, insurance, rule of 72, insurance, and returns on investments. the questions are basic financial literacy questions. the score received on this quiz represents the level of personal finance knowledge. both filers and non-filers take this quiz before the financial literacy class. 3.2. dependent variable the dependent variable is filer/non-filer. it is binary, with a filer coded as 1 and a non-filer coded as 0. 3.3. independent variables the independent variables include 17 demographic variables, 15 psychological variables, and 3 finance variables. these are listed below. 3.3.1. demographic variables. the 17 demographic variables are: gender, age, highest level of education, race, total years of work experience, level of career, level of personal income, income potential, marital status, number of times married, number of children, number of children living at home, type of religion, level of religious commitment, primary residence (rent or own), parents’ highest education level, and parents’ highest income level. 3.3.2. psychological variables. the 15 psychological variables are divided into four areas: self-efficacy (confidence in one’s ability to perform), locus of control, motivation; and selfcontrol. self-efficacy includes three variables, general confidence level, financial confidence level, and education confidence level. locus of control include seven variables. they are: i have had very little control over life; i have had many negative experiences with my household finances; i have very little control over my household finances; i have little control over my income level; i have no control over my savings; i have little control over my expenses; and, i believe the way i manage my money will affect my future. motivation includes three variables. they are: i plan to take more financial education courses; i plan to substantially increase my income; and i plan to increase my net worth. self-control includes two variables. they are: how important is immediate gratification to you? and how important is financial planning to you? 3.3.3. financial variables. the financial variables are utilized as control variables. they are financial knowledge, financial training, and financial work experience. 3.4. research hypotheses we want to determine if there are significant differences between chapter 13 filers and non-filers. to do this we analyze demographic, psychological, and financial factors. the hypotheses are as follows: 72 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 hypothesis 1: demographic variables are related to filing for chapter 13 bankruptcy. hypothesis 2: psychological variables are related to filing for chapter 13 bankruptcy. hypothesis 3: financial variables are related to filing for chapter 13 bankruptcy. the demographic, psychological, and financial variables are stated above. this list of variables is reduced to account for correlation. 3.5. empirical specifications 3.5.1 correlation and endogeneity analyses. correlation analysis is used to adjust the number of independent variables to address potential multicollinearity problems in the regression analysis. to test for endogeneity, we conduct a correlation analysis between the residuals from the binary logistic regression and the independent variables. no significant correlations imply no issues with endogeneity in the model. 3.5.2. binary logistic regression. binary logistic multiple regression is used to analyze the data. the dependent variable is a binary variable, representing filers and non-filers. the independent variables include the reduced number of demographic, psychological, and financial variables. the estimated regression coefficients for the independent variables are used to calculate the odds ratios, which are used to predict the likelihood of a person being a filer when there is a small change in each independent variable, holding all other variables constant. all analyses are done using the statistical software, spss 26. 4. results 4.1. descriptive results 4.1.1. descriptive statistics for demographic variables. descriptive statistics are presented for the demographic variables (table 1). filers are 43.3% males and 56.7% females, whereas non-filers are 29.3% males and 70.7% females. eighty eight percentage of filers are 35 years or older, compared with 63.4% for non-filers. filers are comprised of 51.9% whites and 43.6% blacks, whereas non-filers are comprised of 58.5% whites and 35.8% blacks. whites and blacks account for more than 90% of the sample of filers and non-filers. a high school education is the highest level of education achieved by 41.1% of filers compared with 18.7% of non-filers. nineteen percentage of filers have financial education compared with 32.9% of non-filers. those with 15 years or more of work experience account for 88.2% of filers compared with 67.1% for non-filers. for income, 22.6% of filers earn more than $45,000 per year compared with 35.8% of non-filers. eighty-four percentage of filers are married compared with 65.0% of non-filers. most filers (88.9%) have at least one child, compared with 71.5% of non-filers. sixty percentage of filers have children living at home compared with 49.2% for non-filers. the three religious groups, baptists, nondenominational, and protestants, make up more than 90% of both filers and non-filers. eighty three percentage of filers are homeowners, compared with 59% of non-filers. of all filers, 75% of them have parents with an education level of high school or less compared with 48.8% of non-filers. about 51.3% of filers have parents with income level of $30,000 or less per year compared with 34.5% of non-filers. s. e. kehiaian et al. / financial services review 29 (2021) 67–84 73 t ab le 1 d em o g ra p h ic ch ar ac te ri st ic s fo r c h ap te r 1 3 fi le rs an d n o n -fi le rs d em o g ra p h ic an d fi n an ci al v ar ia b le s f il er s (n = 3 1 4 ) (p er ce n t) n o n -fi le rs (n = 2 4 6 ) (p er ce n t) c o d es 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 . g en d er (1 = m al e, 2 = fe m al e) 4 3 .3 5 6 .7 2 9 .3 7 0 .7 2 . a g e (1 = 1 8 to 2 5 , 2 = 2 6 to 3 5 , 3 = 3 6 to 5 0 , 4 = 5 1 to 6 0 , an d 5 = 6 1 + ) 1 .3 1 1 .1 4 7 .5 2 7 .1 1 3 .1 1 9 .1 1 7 .1 3 2 .1 1 9 .9 1 1 .4 3 . h ig h es t le v el o f ed u ca ti o n (1 = m id d le sc h o o l g ra d u at e, 2 = h ig h sc h o o l g ra d u at e, 3 = so m e co lle g e, 4 = co m m u n it y co ll eg e g ra d u at e, 5 = fo u r y ea r co ll eg e g ra d u at e, 6 = m as te rs le v el g ra d u at e, an d 7 = d o ct o ra te le v el g ra d u at e) 5 .4 3 5 .7 2 9 .3 1 3 .7 1 1 .5 3 .5 1 .0 2 .4 1 6 .3 3 1 .3 1 3 .4 2 2 .4 1 1 .8 2 .4 4 . r ac e (1 = w h it e, 2 = b la ck , 3 = h is p an ic , 4 = a si an , an d 5 = o th er ) 5 1 .9 4 3 .6 2 .5 0 .3 1 .3 5 8 .5 3 5 .8 0 .8 1 .2 3 .7 5 . t o ta l y ea rs o f w o rk ex p er ie n ce (1 = 0 – 3 , 2 = 4 – 6 , 3 = 7 – 1 0 , 4 = 1 1 – 1 5 , 5 = 1 6 – 2 5 , an d 6 = 2 6 + ) 3 .2 2 .5 6 .1 2 0 .4 2 3 .2 4 4 .6 1 1 .4 8 .9 1 1 .4 1 4 .6 1 7 .1 3 5 .4 6 . l ev el o f ca re er (1 = en tr y le v el , 2 = ex p er ie n ce d w o rk er , 3 = su p er v is o r le v el , 4 = m an ag er le v el , an d 5 = ex ec u ti v e le v el ) 1 0 .8 5 1 .6 2 0 .1 1 5 .0 2 .5 1 6 .7 5 0 .0 1 3 .0 1 4 .2 6 .1 7 . l ev el o f p er so n al in co m e (1 = < $ 2 0 k , 2 = $ 2 1 – $ 3 0 k , 3 = $ 3 1 – $ 4 5 k , 4 = $ 4 6 – $ 7 5 k , 5 = $ 7 6 – $ 1 5 0 k , 6 = > $ 1 5 0 k ) 1 9 .7 2 4 .8 3 2 .8 1 8 .5 3 .5 0 .6 2 6 .0 1 9 .1 2 8 .0 1 5 .4 1 0 .2 1 .2 8 . in co m e p o te n ti al (1 = h ig h , 2 = m ed iu m , 3 = lo w , 4 = n o n e, an d 5 = re ti re d ) 7 .3 3 5 .7 3 6 .0 1 3 .7 7 .3 1 3 .8 3 8 .6 3 2 .1 8 .1 7 .3 9 . m ar it al st at u s (1 = m ar ri ed , 2 = p ar tn er , 3 = si n g le , 4 = d iv o rc ed , 5 = w id o w ed , 6 = se p ar at ed ) 6 5 .6 1 8 .2 1 1 .8 1 .0 2 .5 1 .0 5 2 .8 1 2 .2 3 0 .9 2 .0 1 .2 0 .8 1 0 . h o w m an y ti m es h av e y o u b ee n m ar ri ed (1 = o n ce , 2 = tw ic e, 3 = th re e ti m es , 4 = n ev er ) 5 7 .6 2 5 .5 5 .4 1 1 .5 4 7 .2 1 8 .3 4 .1 2 8 .9 1 1 . n u m b er o f ch il d re n (1 = n o n e, 2 = 1 , 3 = 2 , 4 = 3 , 5 = 4 , an d 6 = 5 + ) 1 1 .1 2 2 .3 3 6 .3 1 4 .3 9 .6 6 .4 2 8 .5 1 9 .1 2 4 .8 1 5 .0 7 .3 5 .3 74 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 t ab le 1 (c o n ti n u ed ) d em o g ra p h ic an d fi n an ci al v ar ia b le s f il er s (n = 3 1 4 ) (p er ce n t) n o n -fi le rs (n = 2 4 6 ) (p er ce n t) c o d es 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 d em o g ra p h ic v ar ia b le s: 1 2 . n u m b er o f ch il d re n li v in g at h o m e (1 = n o n e, 2 = 1 , 3 = 2 , 4 = 3 , 5 = 4 , an d 6 = 5 + ) 4 0 .4 3 0 .6 1 9 .1 6 .1 3 .8 5 0 .8 2 4 .0 1 6 .3 6 .5 1 .2 1 .2 1 3 . t y p e o f re li g io n (1 = p ro te st an t, 2 = c at h o li c, 3 = m o rm o n , 4 = f ri en d s, 5 = b ap ti st , 6 = n o n -d en o m in at io n , 7 = m et h o d is t, 8 = m u sl im , je w is h , l u th er an , b u d d h is t, an d d o n ’t b el ie v e 9 .2 2 .5 0 .3 0 .3 5 4 .5 2 7 .4 1 .6 4 .0 1 6 .7 4 .9 0 .4 3 .3 2 3 .2 4 7 .6 1 .2 3 .7 1 4 . l ev el o f re li g io u s co m m it m en t (1 = n o n e, 2 = lo w , 3 = so m ew h at , an d 4 = h ig h ) 7 .0 2 0 .4 2 9 .9 4 2 .7 5 .3 1 3 .8 2 5 .2 5 5 .7 1 5 . p ri m ar y re si d en ce (1 = o w n , an d 2 = re n t) 8 3 .4 1 6 .5 5 8 .5 4 1 .5 1 6 . p ar en ts ’ h ig h es t ed u ca ti o n le v el (1 = n o n e, 2 = m id d le sc h o o l g ra d , 3 = h ig h sc h o o l g ra d , 4 = so m e co ll eg e n o d eg re e, 5 = co m m u n it y co ll eg e g ra d u at e, 6 = fo u r y ea r co lle g e g ra d u at e, 7 = m as te rs le v el g ra d u at e, an d 8 = d o ct o ra te le v el g ra d u at e) 5 .7 1 9 .4 4 9 .7 1 0 .8 5 .1 6 .4 2 .5 0 .4 1 2 .6 3 5 .8 1 9 .9 8 .5 1 3 .0 8 .5 1 .2 1 7 . p ar en ts ’ h ig h es t in co m e le v el (1 = < $ 2 0 k , 2 = $ 2 1 – $ 3 0 k , 3 = $ 3 1 – $ 4 5 k , 4 = $ 4 6 – $ 7 5 k , 5 = $ 7 6 – $ 1 0 0 k , 6 = $ 1 0 0 – $ 1 5 0 , an d 7 = > $ 1 5 0 k ) 1 9 .1 3 2 .2 2 2 .0 1 4 .3 8 .3 2 .9 1 .3 1 3 .4 2 1 .1 2 1 .5 1 8 .3 1 3 .8 6 .5 4 .9 0 .4 f in an ci al v ar ia b le s: 1 . f in an ci al tr ai n in g (y es = 1 , n o = 2 ) 1 8 .8 8 1 .2 3 2 .9 6 6 .7 2 . f in an ci al w o rk ex p er ie n ce (y es = 1 , n o = 2 ) 1 1 .1 8 8 .9 2 2 .8 7 7 .2 3 . f in an ci al k n o w le d g e (t o ta l = 6 3 ) 4 0 .4 9 4 2 .3 7 n o te : t h is ta b le p ro v id es th e fr eq u en cy d is tr ib u ti o n in p er ce n ta g e o f d em o g ra p h ic ch ar ac te ri st ic s fo r c h ap te r 1 3 fi le rs an d n o n -fi le rs . a ll d at a ar e m ea su re d u si n g a l ik er t sc al e. s. e. kehiaian et al. / financial services review 29 (2021) 67–84 75 4.1.2. descriptive statistics for finance variables. for financial training, 18.8% of filers have financial training compared with 32.9% of non-filers. for financial work experience, 11.1% of filers have financial work experience compared with 22.8% of non-filers. for financial knowledge, of the 63 questions, the mean financial literacy quiz scores for the chapter 13 filers and non-filers are 40.49 and 42.37, respectively. 4.1.3. descriptive statistics for psychological characteristics. the descriptive statistics for the psychological characteristics are presented below. the percentage representing “high” or “agree” is the percentages of responses that are coded as 4 s and 5 s from the data for each variable. for self-efficacy, filers generally have lower levels of confidence than non-filers (table 2). of all filers, 53.5% of them have a high level of general confidence level compared with 63.9% of non-filers. (table 2). corresponding statistics for financial confidence and education confidence are (20% of filers, and 29.1% of non-filers), and (46% of filers and 60.1% of non-filers), respectively. for locus of control, non-filers seem to have higher levels of control than do filers (table 2). for “i have had very little control over life,” 28.4% of filers agree compared with 23.2% of non-filers. for “i have had many negative experiences with my household finances,” 35.1% of filers agree compared with 25.6% of non-filers. for motivation, the variable, “i plan on substantially increasing my income,” 46.8% of filers agree compared with 50.4% of non-filers. for “i plan to increase my net worth,” 60.2% of filers agree compared with 72.4% of non-filers. for self-control, the variable, “how important is immediate gratification to you?”; 35.7% of filers strongly agree versus 26.4% of nonfilers. for commitment, about 43% of filers have high levels of commitment compared with 55.7 for non-filers. 4.3. logistic regression results 4.3.1. correlation results. because of high correlations (r > 0.300), a reduced set of demographic, psychological, and financial variables are included in the regression analysis. the nine demographic variables included are gender, education, years of work, religious commitment, religious faith, home ownership, parents’ income, race, and marital status. the five psychological variables included are: financial confidence for self-efficacy; control over life, negative experience with finances, and money management affecting my future for locus of control; taking financial education in the future for motivation; and importance of immediate gratification for self-control. the three financial variables included are finance knowledge, finance training, and importance of financial planning. 4.3.2. endogeneity testing results. a correlation analysis is done between the residuals from the logistic regression estimation and each of the independent variables and no significant correlation is found. this implies that endogeneity is not an issue with the logistic regression model. 4.3.3. binary logistic regression results for demographic variables. for gender, the odds of a male filing for bankruptcy is 2.075 times the odds for a female filing (odds ratio [or] = 2.075, p-value = 0.004; table 3). for race, the odds ratio for a black person to file for bankruptcy is 2.264 times the odds of other races (primarily whites) filing (or= 2.264, pvalue = 0.003). for marital status, a single person filing for bankruptcy is 0.377 times the odds of a married person filing for bankruptcy (or= 0.377, p-value = 0.002). for education, 76 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 t ab le 2 p sy ch o lo g ic al ch ar ac te ri st ic s fo r c h ap te r 1 3 fi le rs an d n o n -fi le rs p sy ch o lo g ic al ch ar ac te ri st ic s f il er s (n = 3 1 4 ) (p er ce n t) n o n -fi le rs (n = 2 4 6 ) (p er ce n t) c o d es 1 = l ea st 2 3 4 5 = m o st 1 = l ea st 2 3 4 5 = m o st 1 . s el fef fi ca cy : a. g en er al co n fi d en ce le v el 5 .4 8 .6 3 2 .5 3 2 .5 2 1 .0 4 .5 4 .9 2 6 .8 4 0 .7 2 3 .2 b . f in an ci al co n fi d en ce le v el 1 3 .7 2 2 .3 4 5 .5 1 4 .0 4 .5 1 4 .2 1 8 .3 3 8 .2 2 1 .1 8 .1 c. e d u ca ti o n co n fi d en ce le v el 3 .8 8 .9 4 1 .1 3 1 .5 1 4 .6 1 .6 7 .3 3 0 .9 3 9 .8 2 0 .3 2 . l o cu s o f co n tr o l: a. i h av e h ad v er y li tt le co n tr o l o v er li fe 1 7 .5 1 8 .5 3 5 .7 1 6 .9 1 1 .5 2 9 .7 2 3 .6 2 3 .6 1 7 .1 6 .1 b . i h av e h ad m an y n eg at iv e ex p er ie n ce s w it h m y h o u se h o ld fi n an ce s 8 .0 2 0 .4 3 6 .6 1 6 .9 1 8 .2 2 0 .7 2 6 .8 2 6 .8 1 4 .2 1 1 .4 c. i h av e v er y li tt le co n tr o l o v er m y h o u se h o ld fi n an ce s 1 5 .9 2 6 .8 3 6 .9 1 3 .4 7 .0 3 3 .7 2 4 .4 2 3 .2 1 0 .2 8 .5 d . i h av e v er y li tt le co n tr o l o v er m y in co m e le v el 1 3 .1 2 4 .2 3 7 .9 1 4 .3 1 0 .5 2 7 .2 2 7 .2 2 8 .5 1 1 .0 6 .1 e. i h av e n o co n tr o l o v er m y sa v in g s 2 0 .7 2 5 .5 3 5 .0 8 .9 9 .9 4 3 .5 2 4 .0 1 7 .9 9 .3 5 .3 f. i h av e li tt le co n tr o l o v er m y ex p en se s 1 8 .5 2 2 .0 3 9 .5 1 2 .4 7 .6 4 0 .7 2 4 .4 1 8 .3 9 .8 6 .9 g . i b el ie v e th e w ay i m an ag e m y m o n ey w il l af fe ct m y fu tu re 6 .1 6 .4 1 9 .1 2 0 .4 4 8 .1 5 .7 6 .5 1 5 .4 2 4 .4 4 8 .0 3 . m o ti v at io n : a. i p la n to ta k e m o re fi n an ci al ed u ca ti o n co u rs es 1 6 .6 1 4 .6 3 1 .2 1 7 .2 2 0 .4 1 6 .7 2 0 .3 2 6 .4 1 9 .9 1 6 .7 b . i p la n o n su b st an ti al ly in cr ea sin g m y in co m e 1 1 .1 1 4 .6 2 7 .4 2 3 .2 2 3 .6 7 .3 1 3 .4 2 8 .9 2 0 .7 2 9 .7 c. i p la n to in cr ea se m y n et w o rt h 7 .6 9 .2 2 2 .9 3 0 .3 2 9 .9 4 .9 9 .8 2 4 .0 2 6 .0 3 5 .4 4 . s el fco n tr o l: a. h o w im p o rt an t is im m ed ia te g ra ti fi ca ti o n to y o u ? 1 2 .1 5 2 .2 2 8 .7 7 .0 2 1 .1 5 2 .4 1 8 .7 7 .7 b . h o w im p o rt an t is fi n an ci al p la n n in g to y o u ? 2 .2 2 8 .0 5 3 .8 1 5 .9 2 .0 3 3 .3 4 1 .5 2 3 .2 n o te : t h is ta b le p ro v id es th e fr eq u en cy d is tr ib u ti o n in p er ce n ta g e o f p sy ch o lo g ic al ch ar ac te ri st ic s fo r c h ap te r 1 3 fi le rs an d n o n -fi le rs . a ll d at a ar e m ea su re d u si n g a l ik er t sc al e. s. e. kehiaian et al. / financial services review 29 (2021) 67–84 77 t ab le 3 b in ar y lo g is ti c re g re ss io n re su lt s fo r d em o g ra p h ic d if fe re n ce s b et w ee n fi le rs an d n o n -fi le rs d em o g ra p h ic ch ar ac te ri st ic s b s e s ig . e x p (b ) d em o g ra p h ic v ar ia b le s: 1 . g en d er (f em al e = 0 , m al e = 1 ) 0 .7 3 0 0 .2 5 3 0 .0 0 4 2 .0 7 5 * * 2 . e d u ca ti o n �0 .2 4 7 0 .0 9 0 0 .0 0 6 0 .7 8 1 * * 3 . y ea rs o f w o rk 0 .2 8 0 0 .0 8 0 0 .0 0 0 1 .3 2 3 * * * 4 . l ev el o f re li g io u s co m m it m en t (1 = n o n e, 2 = lo w , 3 = so m ew h at , an d 4 = h ig h ) �0 .3 0 6 0 .1 3 2 0 .0 2 1 0 .7 3 7 * * 5 . p ri m ar y re si d en ce (0 = o w n , an d 1 = re n t) 1 .6 0 2 0 .2 7 0 0 .0 0 0 4 .9 6 3 * * * 6 . p ar en ts ’ h ig h es t in co m e le v el (1 = < $ 2 0 k , 2 = $ 2 1 – $ 3 0 k , 3 = $ 3 1 – $ 4 5 k , 4 = $ 4 6 – $ 7 5 k , 5 = $ 7 6 – $ 1 0 0 k , 6 = $ 1 0 0 – $ 1 5 0 , an d 7 = > $ 1 5 0 k ) �0 .1 4 0 0 .0 7 8 0 .0 7 2 0 .8 6 9 * 7 . r ac e (w h it e = 0 , (b en ch m ar k )) 0 .0 1 3 * * r ac e (b la ck = 1 , w h it es an d o th er ra ce s = 0 ) 0 .8 1 7 0 .2 7 8 0 .0 0 3 2 .2 6 4 * * r ac e (o th er ra ce s = 1 , w h it es an d b la ck s = 0 ) 0 .5 9 2 0 .5 3 8 0 .2 7 0 1 .8 0 8 8 . m ar it al st at u s (m ar ri ed = 0 , b en ch m ar k )) 0 .0 1 7 * * m ar it al st at u s (1 = p ar tn er s, 0 = al l o th er ca te g o ri es ) 0 .0 6 9 0 .3 1 2 0 .8 2 6 1 .0 7 1 m ar it al st at u s (1 = si n g le , 0 = al l o th er ca te g o ri es ) �0 .9 7 6 0 .3 2 3 0 .0 0 2 0 .3 7 7 * * m ar it al st at u s (1 = o th er , 0 = al l o th er ca te g o ri es ) �0 .1 5 1 0 .5 3 5 0 .7 7 8 0 .8 6 0 9 . r el ig io n (b ap ti st = 0 , (b en ch m ar k )) 0 .0 0 0 * * * r el ig io n (n o n -d en o m in at io n al fa it h s = 1 , al l o th er ca te g o ri es = 0 ) �1 .3 8 9 0 .2 6 6 0 .0 0 0 0 .2 4 9 * * * r el ig io n (p ro te st an ts = 1 , al l o th er ca te g o ri es = 0 ) �1 .8 2 6 0 .3 7 8 0 .0 0 0 0 .1 6 1 * * * r el ig io n (c at h o li c = 1 , al l o th er ca te g o ri es = 0 ) �0 .7 5 3 0 .5 8 1 0 .1 9 5 0 .4 7 1 r el ig io n (o th er = 1 , al l o th er ca te g o ri es = 0 ) �1 .3 0 3 0 .4 6 6 0 .0 0 5 0 .2 7 2 * * 78 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 t ab le 3 (c o n ti n u ed ) p sy ch o lo g ic al an d fi n an ci al ch ar ac te ri st ic s: b s e s ig . e x p (b ) p sy ch o lo g ic al v ar ia b le s: 1 . s el fef fi ca cy a. f in an ci al co n fi d en ce le v el �0 .2 3 9 0 .1 1 3 0 .0 3 3 0 .7 8 7 * * 2 . l o cu s o f co n tr o l: a. i h av e h ad v er y li tt le co n tr o l o v er li fe 0 .0 3 0 0 .0 9 5 0 .7 5 2 1 .0 3 1 b . i h av e h ad m an y n eg at iv e ex p er ie n ce s w it h m y h o u se h o ld fi n an ce s 0 .3 3 3 0 .0 9 9 0 .0 0 1 1 .3 9 5 * * c. i b el ie v e th e w ay i m an ag e m y m o n ey w il l af fe ct m y fu tu re �0 .0 2 7 0 .1 0 0 0 .7 8 4 0 .9 7 3 3 . m o ti v at io n : a. i p la n to ta k e m o re fi n an ci al ed u ca ti o n co u rs es 0 .1 2 4 0 .0 9 2 0 .1 7 9 1 .1 3 2 4 . im m ed ia te g ra ti fi ca ti o n : a. im p o rt an ce o f im m ed ia te g ra ti fi ca ti o n 0 .3 8 4 0 .1 5 3 0 .0 1 2 1 .4 6 8 * * c o n tr o l v ar ia b le s (fi n an ci al v ar ia b le s) : a. f in an ce tr ai n in g (n o = 0 , y es = 1 ) �0 .6 6 8 0 .2 6 7 0 .0 1 3 0 .5 1 3 * * b . im p o rt an ce o f fi n an ci al p la n n in g 0 .0 3 3 0 .1 6 6 0 .8 4 0 1 .0 3 4 c. f in an ci al li te ra cy q u iz �0 .0 0 2 0 .0 1 8 0 .8 9 6 0 .9 9 8 n o te : t h is ta b le p ro v id es th e re su lt s fo r th e d if fe re n ce s b et w ee n c h ap te r 1 3 fi le rs an d n o n -fi le rs fo r su b -c at eg o ri es o f p sy ch o lo g ic al ch ar ac te ri st ic s. e ac h su b -c at eg o ry fo r th e p sy ch o lo g ic al v ar ia b le s h as se v er al v ar ia b le s. a ll d at a ar e m ea su re s u si n g a l ik er t sc al e. b in ar y lo g is ti c re g re ss io n te st in g (fi le rs = 1 , n o n -fi le rs = 0 fo r th e d ep en d en t v ar ia b le ) is u se d to te st th e d if fe re n ce b et w ee n th e tw o g ro u p s. a ll p sy ch o lo g ic al , d em o g ra p h ic an d fi n an ci al v ar ia b le s ar e th e in d ep en d en t v ar ia b le s. f o r su b -c at eg o ry , se lf -e ffi ca cy , fi n an ci al co n fi d en ce is si g n ifi ca n t. f o r lo cu s o f co n tr o l, h av in g m an y n eg at iv e ex p er ie n ce s w it h fi n an ce s is si g n ifi ca n t. f o r th e su b -c at eg o ry , m o ti v at io n , th er e ar e n o si g n ifi ca n t d if fe re n ce s. im m ed ia te g ra ti fi ca ti o n is si g n ifi ca n t. t h e co n tr o l v ar ia b le , fi n an ce tr ai n in g , is si g n ifi ca n t. * * * s ig n ifi ca n t at th e 1 % le v el . * * s ig n ifi ca n t at th e 5 % le v el . * s ig n ifi ca n t at th e 1 0 % le v el . s. e. kehiaian et al. / financial services review 29 (2021) 67–84 79 the odds of a person whose education goes up by one level to file for bankruptcy is 0.781 times the odds for a person whose education has not changed (or= 0.781, p-value = 0.006). for years of work, a one-year increase in years of work increases the odds of being a filer by 1.323 times (or= 1.323, p-value = 0.000). for homeownership, the odds of a homeowner to file for bankruptcy are 4.963 times higher than the odds of a nonhomeowner to file (or= 4.963, p-value = 0.000). this is a highly significant result. for parents’ income, an increase by one level, the odds for the son or daughter to file for bankruptcy is 0.869 times the odds if the parents’ income does not increase (or= 0.869, p-value = 0.072). this intergenerational relationship that is not seen in the literature. for religious commitment, if a person’s religious commitment increases by one level, the odds of that person filing for bankruptcy is 0.737 times the odds of filing for bankruptcy with no increase (or= 0.737, pvalue = 0.021). the odds of a person from a protestant faith to file for bankruptcy is 0.161 times the odds of filing by the other faiths studied (or= 0.161, p-value = 0.000). the odds of a person from a nondenominational faith filing is 0.249 times the odds of a person from another faith filing (or= 0.249, p-value = 0.000). these results for different faiths impacting bankruptcy filing are not seen in the literature. 4.3.4. logistic regression results for psychological variables. for self-efficacy, the odds that a person with an increase in financial confidence filing is 0.787 times the odds before the increase (or= 0.787, p-value = 0.033; table 5). for locus of control, the odds of a person who has an increase in negative experiences with his household finances filing is 1.395 times the odds of that person filing before the increase in negative experiences (or= 1.395, pvalue = 0.001). for self-control, the odds of a person with an increase in the importance of immediate gratification filing is 1.5 times the odds for that person before the increase (or =1.468, p-value = 0.012). motivation is not significant. 4.3.5. logistic regression results for financial variables. the odds of a person who receives financial training to be a filer is 0.513 times the odds of someone who does not receive financial training (or= 0.513, p-value = 0.013). 5. discussion this study compares chapter 13 bankruptcy filers and non-filers using demographic, psychological, and financial variables. the significant demographic variables are gender, education, religious faiths, religious commitment, racial group, homeownership, parent’s income, and marital status. the significant psychological variables are self-efficacy, locus of control, and self-interest. the significant financial variable is financial training. 5.1. demographics for gender, we find that males are twice as likely as females to file for bankruptcy. chakravarty and rhee (1999) find that the gender of a single head of household appears to be an important predictor of bankruptcy filing. for education, we find that increasing education is likely to reduce the odds of filing for bankruptcy. fisher (2019) finds that people with lower education are more likely to file for bankruptcy. for years of work, we find that an 80 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 increase in years of work is likely to increase the odds of filing for bankruptcy. we have not seen this result in the literature for personal bankruptcy. by having more years of work, a person’s income usually increases. this can trigger the accumulation of more debt, which can lead to filing for bankruptcy. we find that homeowners are five times more likely to be filers than renters. this is a highly significant result. home ownership puts a heavy financial strain on a homeowner and, if he or she were to have a loss of employment or a major medical occurrence, the likelihood of filing for bankruptcy increases sharply. also, the chapter 13 bankruptcy law is structured such that, a person can keep the home while implementing a restructured debt repayment plan. this benefit from home ownership increases the likelihood to file for chapter 13 bankruptcy. domowitz and sartain (1999) and zhu (2011) find that homeownership impacts the decision to file for bankruptcy. for racial groups, blacks are twice as likely to file for bankruptcy than whites. this is a highly significant result. fisher (2019) finds that blacks and other minority groups are more likely to file for bankruptcy. for parents’ income, an increase in the parents’ income reduces the likelihood of the son or daughter to file for bankruptcy. we have not seen this result in the literature. there is a high correlation between a parent’s income and education. these two factors show an intergenerational positive financial effect of reducing the odds of filing for chapter 13 bankruptcy by a son or daughter. for marital status, singles are significantly less likely to be filers than married people. this is a reasonable result as married people generally have more debt (including a mortgage), more children, and more children living at home. caputo (2008) finds that single people, at the time of declared bankruptcy, are less likely than married persons to declare bankruptcy. we find than an increase in religious commitment reduces the likelihood of filing for bankruptcy. increased religious commitment can instill principles of discipline, moderation and caring, which can lead to better money management and less filing for bankruptcy. nondenominational and protestant faiths are less likely to file for bankruptcy than other faiths. according to the economist (02/2018), protestants embrace the notion that diligence and self-improvement are pleasing to god. they tend to be more disciplined with their money management, which leads to less filing for bankruptcy. this supports our finding that protestants are less likely to file for bankruptcy than other faiths. khan (2010) studies faith and finance and finds that religion has a strong impact on economic behavior. beck et al. (2014) find that religion is a driving force behind the abnormally higher chapter 13 filings compared with chapter 7 filings in the southern united states. there is a high concentration of evangelicals and fundamentalists in the southern united states. people of these faiths prefer to file for chapter 13 bankruptcy because they feel that it is their moral obligation to honor their debt commitments with a new repayment plan rather than to pass them on to creditors and society at large. 5.2. psychological factors for self-efficacy (belief that one can accomplish a goal), we find that those with more financial confidence are less likely to be filers. danes, casas and boyce (1999) and asaad s. e. kehiaian et al. / financial services review 29 (2021) 67–84 81 (2015) find that financial confidence is a critical component of financial behavior. an increase in financial confidence is likely from more financial literacy education. this will likely lead to fewer personal bankruptcies. for locus of control, an increase in negative experiences with household finances increases the odds of filing for bankruptcy. perry and morris (2005) and miotto and parente (2015) also find that locus of control impacts financial decisions. our study takes it further by showing the relationship between locus of control and the financial decision of filing for bankruptcy. for self-control, we find that an increase in the importance of immediate gratification (implying less self-control) increases the odds of filing for bankruptcy. this result is not seen in the literature. miotto and parente (2015) find that more self-control does lead to more savings and less financial defaults. this is a reasonable result as a person with high immediate gratification needs will spend more on satisfying these needs and pay less attention to good budgeting practices and future needs. 5.3. financial factors we find that financial training reduces the likelihood of filing for bankruptcy by about 50%. this result shows the importance of financial training to reduce chapter 13 bankruptcy filing. again, we have not seen this result in the bankruptcy filings literature. collins (2010) study financial literacy of lower income families and find that financial education increases long-term savings and long-term credit scores. our results show the importance of financial literacy to address the bankruptcy problem in the united states. 6. conclusion filing for bankruptcy is not an easy decision, yet thousands file each year to reduce the debt pressure in their lives. we find several demographic, psychological and financial differences between filers versus non-filers of chapter 13 bankruptcy. the significant demographic variables are gender, education, religious faiths, religious commitment, blacks, homeownership, parent’s income, and marital status. the significant psychological variables are self-efficacy, locus of control, and self-control. the significant financial variable is financial training. the variables not seen in the literature are the demographic variable, parents’ income, the psychological variables, self-efficacy, locus of control and self-control, and the financial variable, financial training. there are several implications from this study. we find a negative relationship between financial literacy and filing for bankruptcy. this implies that an increase in financial literacy is likely to reduce personal bankruptcy. we recommend that government (local, state, and federal), educational institutions, and lending institutions, especially for home purchases, provide more financial literacy education to address the personal bankruptcy situation in the society. ongoing financial literacy education, particularly for adults, may be more effective. we find that males are twice as likely as females to be filers. we recommend targeting more males with financial literacy training to reduce bankruptcy. we find that blacks are more likely to be filers than whites. we recommend that more financial literacy education be offered to 82 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 blacks to reduce bankruptcy filings. we find that an increase in education reduces the odds of filing for bankruptcy. reduced bankruptcy is another benefit from more education. we also find that an increase in religious commitment reduces the odds of filing for bankruptcy. hence, the encouragement to increase one’s religious commitment is recommended. religious leaders may be able to promote this finding. we also find that increases in parents’ income (and education) reduce the likelihood of a son or daughter filing for bankruptcy. information on the parents’ education or income may help in the risk assessment for bankruptcy for a son or daughter applying for a mortgage or other big financial commitment. that is, lending institution can ask the question, “what do your parents do for a living?” we find that a homeowner is five times more likely to file for bankruptcy than a nonhome owner. we recommend that homeowners create a bigger financial cushion for emergencies than non-homeowners. we also recommend that homeowners take more financial literacy training, especially budgeting. these actions will reduce the likelihood of filing for bankruptcy. these actions are also relevant for married people as they are more likely to file for bankruptcy than single people. there are also implications from the psychological results. we find that an increase in a person’s financial confidence reduces the odds of filing for bankruptcy. we also find that an increase in negative experiences with household finances increases the odds of filing for bankruptcy. one recommendation to address these two findings is to increase financial literacy education. with more financial knowledge, one should become more financially confident and should have fewer negative financial experiences. this should reduce personal bankruptcies. we find that filers have higher levels of immediate gratification (less self-control) compared with non-filers. we recommend financial discipline training, especially budgeting, to alleviate the situation. a limitation of this study is that the data are taken from the middle district of north carolina and hence the results may not be applicable elsewhere the results may be acceptable for places with high percentage of whites and blacks. caution is required to apply the results to other states like florida, new york, and california, where the population has a different mix of racial groups, including higher percentages of latinos, italians, caribbean blacks, african blacks, and/or asians. also, we do not have the data of a few other variables that could expand to the model. these could include debt, wealth, principal balance on mortgage, and health status. further studies can be done by looking at these demographic, psychological, and financial variables in other communities across the country or abroad. this will add more robustness to the results found in the few studies done on bankruptcy filing. in the present economic environment with unprecedented unemployment and loss of income resulting from the coronavirus pandemic, more research on personal bankruptcy is recommended. references asaad, c. t. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101. beck, j. h., hackney, d. d., hackney, j., & mcpherson, m. q. (2014). regional differences in chapter 13 filings: southern legal culture or religion? review of social economy, 72, 186-208. bronfrenbrenner, u. (1979). purpose and perspectives and basic concepts. the ecology of human development (pp. 3-28). cambridge, ma: harvard university press. s. e. kehiaian et al. / financial services review 29 (2021) 67–84 83 caputo, r. k. (2008). marital status and other correlates of personal bankruptcy, 1986–2004.marriage & family review, 44, 5-32. chakravarty, s., & rhee, e. y. (1999). factors affecting an individual’s bankruptcy filing decision. available at ssrn 164351. https://dx.doi.org/10.2139/ssrn.164351 clements, j., johnson, d., michelich, k., & olinsky, c. f. (1999). characteristics of bankruptcy filers: implications for educators. journal of family and consumer sciences, 91, 71. collins, j. m. (2010). effects of mandatory financial education on low-income clients. focus, 27, 13-18. danes, s. m., huddleston-casas, c., & boyce, l. (1999). financial planning curriculum for teens: impact evaluation. journal of financial counseling and planning, 10, 26. deacon, r. e., & firebaugh, f. m. (1988). family resource management: principles and applications (2nd ed.). boston, ma: allyn and bacon domowitz, i., & sartain, r. l. (1999). determinants of the consumer bankruptcy decision. the journal of finance, 54, 403-420. efrat, r. (1998). the moral appeal of personal bankruptcy.whittier law review, 20, 141-170. evans, d., & bauchet, j. (2017). bankruptcy determinants among us households during the peak of the great recession. consumer interests annual, 63, 1-7. evans, d. a., & lown, j. m. (2008). predictors of chapter 13 completion rates: the role of socioeconomic variables and consumer debt type. journal of family and economic issues, 29, 202-218. fisher, j. d. (2019). who files for personal bankruptcy in the united states? journal of consumer affairs, 53, 2003-2026. available at https://www.census.gov/quickfacts/fact/table/; https://www.economist.com/financeand-economics/2018/02/22/protestantism-might-be-good-for-the-wallet-after-all; http://www.uscourts.gov/ statistics-reports/caseload-statistics-data-tables khan, a. k. (2010). essays on faith and finance (unpublished doctoral dissertation). harvard business school, boston, ma. lefgren, l., & mcintyre, f. (2009). explaining the puzzle of cross-state differences in bankruptcy rates. the journal of law and economics, 52, 367-393. loibl, c., hira, t. k., & rupured, m. (2006). first time versus repeat filers: the likelihood of completing a chapter 13 bankruptcy repayment plan. journal of financial counseling and planning, 17, 23-33. miller, k. s., levin, m. l., whitaker, d. j., & xu, x. (1998). patterns of condom use among adolescents: the impact of mother-adolescent communication. american journal of public health, 88, 1542-1544. miotto, a., paula, s. c., & parente, j. (2015). antecedents and consequences of household financial management in brazilian lower-middle-class. revista de administração de empresas, 55, 50-64. perry, v. g., & morris, m. d. (2005). who is in control? the role of self-perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39, 299-313. samuelson, p. a. (1937). a note on measurement of utility. the review of economic studies, 4, 155-161. schuchardt, j., durband, d., bailey, w. c., devaney, s. a., grable, j. e., leech, i. e., . . . xiao, j. j. (2007). personal finance: an interdisciplinary profession. journal of financial counseling and planning, 18, 1-9. shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26, 609-643. thaler, r. h., & shefrin, h. m. (1981). an economic theory of self-control. journal of political economy, 89, 392-406. williams, a. a., kehiaian, s. e., & bird, c. l. (2017). differences in financial actions between chapter 13 bankruptcy filers and non-filers. the journal of applied business and economics, 19, 136-155. zhu, n. (2011). household consumption and personal bankruptcy. the journal of legal studies, 40, 1-37. 84 s. e. kehiaian et al. / financial services review 29 (2021) 67–84 pii: 1057-0810(92)90012-2 financial services review, 2(l): l-20 copyright 8 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. a multicriteria approach to mutual fund selection wade d. cook kevin j. hebner in practice, when investors select a mutual fund, they take into account a number offactors. however, the most popular approach for evaluating mutual funds employs only a single criterion, thefinds’ mean, n’sk-adjusted, rate of return (jensen’s a coefficient). in thepresent paper a multicriteria approach to mutual fund selection is presented. the multicriteria methodology allows numerous factors to be considered, for example, the standard deviation of the funds’ a’s, front and back-end loadfees, the level of diversification, quality of service, and so on. it also recognizes that individual investors possess heterogeneous attributes and preferences, and hence, allows investors to formulate different ratings (and consequently rankings) of the set of competing mutual funds. i. i~roducti~n in the decade ended december, 1989, the number of u.s. households owning mutual fund shares increased by 495 percent to 22.8 million and the total number of shareowner accounts increased to 58.2 million. furthermore, total u.s. mutual fund assets increased by 1,750 percent to $982 billion and the number of u.s. funds more than quadrupled to 2,9 18.’ the mutual funds offered vary enormously in terms of their investment objectives, types of securities held, historical returns and risk levels, load and management fees, levels of diversification, quality of service, and so on. similarly, fund shareholders vary enormously in terms of their wealth levels, rates of portfolio turnover, degrees of risk tolerance, understanding of financial markets, beliefs in the ability of mutual funds to outperform the market, their own portfolio’s level of diversification, and so on. with this impressive growth record and the increasing complexity, diversity and competitiveness of the mutual fund industry, it is important to examine the approach adopted by investors when evaluating mutual fund managers. wade d. cook l faculty of administrative studies, york university, toronto, canada, m3j lp3. kevin j. hebner l faculty of administrative studies, york university. 2 financial services rjmew, 2(l) l!i!w1993 evaluating competing money managers is an integral part of the decision making process facing individual investors who am determining which mutual fund to select or corporations who are deciding which pension fund manager to hire. in what has become the classic article on performance evaluation, jensen (1968) uses a single criterion approach to provide evidence that (on a risk-adjusted, net of m~age~nt fees basis) mutual funds as a class do not outfox the market. it is important to note however, that his purpose was “not to evaluate the funds from the standpoint of the individual investor, but only to evaluate the fund managers’ forecasting ability” (p. 404). for examining the forecasting ability of mutual fund managers’ as a class, jensen’s single criterion approach may be appropriate.’ however, if one’s objective is to evaluate competing funds from the perspective of a potential investor, a multicriteria approach is required.3 jensen’s single criterion is the “average incremental rate of return on the portfolio per unit time” (p. 394); what is now referred to as jensen’s a coefficient. jensen explicitly chose not to consider other criteria, such as a mutual fund’s consistency in earning incremental returns, or its level of diversi~cation (p. 415). furthermore, jensen omitted the fund’s front-end and back-end load fees (p. 4041, as well as the level of service offered to its investors. the objective of this paper is to provide an alternative to the single criterion approach adopted by jensen (and numerous more recent studies), so that a large number of competing mutual funds can be evaluated by an individual investor. the multic~~ria approach developed in this paper is based on a model by cook and kress (1991). this model, which was designed for multiple criteria problems with purely ordinal data, is adapted here to incorporate the cardinal data inherent in the mutual fund problem. the model explicitly recognizes that investors possess heterogeneous attributes and preferences, and hence, in general, they formulate different ratings (and con~uently rankings) of the set of vomiting mutual funds. in contrast to this implication of the multicriteriaapproach, if investors employed jensen’s single criterion approach, unanimity would exist in their fund ratings (and rankings). all investors would then select the same mutual fund, an implication which is clearly inconsistent with the large number (2,918) of funds currently sold in the u.s. to develop a multicriteria approach to mutual fund selection by an individual investor, this paper proceeds as follows: section ii presents, for illustrative purposes, a set of six criteria which could be used by the investor. section iii first introduces the general concept used for combining multiple criteria, and presents three general forms of investor preference regarding the mutual funds and the criteria against which the funds are to be judged. then, from the three general forms of investor preference, five constraints on the weights chosen by the model to rank the various funds are derived. next, the general multicriteria approach for ranking mutual funds is presented, along with a simple illustrative example. section iv provides an example of the selection procedure using data from ten mutual funds evaluated on the basis of the six suggested criteria. im~~tly, it is demons~ted that a very a multicriteria approach to mutual fund selection 3 different ranking of the ten funds is derived using this paper’s multicriteria approach compared with jensen’s single criterion approach. furthermore, it is demonstrated that, using the multicriteria approach, investors with heterogeneous attributes and preferences, do obtain different fund rankings. conclusions are presented in section 5. ilthekcriteria a set of m mutual funds i = 1,2,...,m are to be ranked according to a set of k criteria k = 1,2,...,k. for each of the m funds an observation (either cardinal or ordinal) for each of the k criteria must be provided by the investor. for illustrative purposes, assume there are six criteria: i). a, the mean, risk-adjusted rate of return, net of management fees, observed for fund i during the last t periods (i.e., jensen’s a coefficient).4 using prior rates of return to select a mutual fund could be considered a questionable practice in view of evidence that past performance contains little or no predictive value. however, allerdice and farrar (1967), friend, blume and crocket (1970), smith (1978) and woerheide (1982) provide evidence that past rates of return are positively correlated with a fund’s net sales. furthermore, survey evidence presented by lewellen, lease and schlarbaum (1977) and the investment company institute (1987) indicate that prior returns are one of the primary variables used by investors when choosing mutual funds. investors’ beliefs regarding the ability of mutual funds to outperform the market will be a primary determinant of the importance they accord to a,.. ii). ai, the standard deviation of the t a,‘~ observed for mutual fund i.’ investors who are less risk tolerant will place relatively greater importance on a,. iii). di, mutual fund i’s degree of diversification (that is, how closely the fund’s portfolio approximates the market portfolio).6 investors who are less risk tolerant or whose portfolios are less well diversified, will accord relatively more importance to dp7 iv). fi, fund i’s front-end load fee. normally fi is a decreasing function of the quantity purchased and hence, will be a relatively less important criterion for investors who are wealthier and purchase larger quantities. v). bi, fund i’s back-end load fee. normally bi is a decreasing function of the investor’s holding period and hence, will be a relatively less important criterion for investors with longer expected time horizons or equivalently, lower rates of portfolio turnover. vi). si, fund i’s service level.* investors whose understanding of financial markets is limited will generally accord relatively more importance to s,., the six examples of criteria provided above are used throughout this paper to illustrate the multicriteria approach to mutual fund selection.g financial services review, 2(l) 1992m93 iii. a model for c~~~~ mottle criteria 1. the general concept in general, it is assumed that the investor can express three forms of preference regarding the funds and the criteria against which the funds are to be judged. preferences among funds by criterion. it is assumed that the investor can rank order the m mutual funds according to each criterion, that is, for each criterion, the investor can decide which fund ranks in first place, which in second, and so on.‘* to rank order the m funds according to criterion k, define a binary (0 1) matrix ar where ak = (a$) 1 if fund i is ranked e’th on criterion k, and a$ 0 otherwise. (1) for the moment, assume that the investor is only interested in the ordinal ranking of the funds along each of the k dimensions. while this approach does not immediately take advantage of the info~ation in the cardinal data available, the cardinal data is used in the formal model presented below when choosing the weights to be applied to the rank positions (see (3a) below). to illustrate how a binary matrix is constructed, assume there are five mutual funds to be ranked according to q, the k’th criterion, with i = 1,2,-a used to index the funds. then, if a1 = 10 percent, a2 = 6 percent, a3 = 12 percent, a4 = 9 percent and a5 = 8 percent, the ordinal ranking can be represented by the matrix rank fund 1 2 3 4 5 1 0 10 00 2 000 01 ak= 3 1 0 0 0 0 4 0 0 100 5 000 10 ranking the criteria by i~rt~~e the second assumption is that the investor can rank the k criteria themselves in order of impo~nce. ‘i for convenience, let k = 1 denote the most impo~nt a multicriteria approach to mutual fund selection 5 criterion, k = 2 the next most important, and so on, with k = k denoting the least important criterion.‘* ranking the criteria by clearness criteria “clearness” is a third concern in rank ordering the m mutual funds.i3 it is an issue because the ability of the investor to distinguish between funds on the basis of, say, cli may be greater than his ability to distinguish between funds on the basis of di or si (where si may be especially difficult to quantify). to incorporate criteria clearness into the model, the investor must be able to rank the criteria from the most to the least clear. this is a ranking over and above the “criteria importance” ranking. once the three forms of investor preference are incorporated into the model, the investor’s objective is to obtain, for each mutual fund i, an overall rating, ri, which reflects the fund’s aggregate standing over the set of criteria. to accomplish this,asetofweights (c$),fork= l,..., k and 2 = l,..., m, is determined, where u$ is the level of importance or weight accorded the z’th rank position for criterion k. the rating for fund i is then k m (2) ri = 11 afpf . k-l l-1 for example, if fund #l is ranked first, fourth, third, fifth, first and third on criteria k = 1, k = 2, . . . . k = 6 respectively, then its rating would be the model’s weights must be chosen in such a manner that the three forms of preferences outlined above are adhered to. specifically, for a given criterion, a higher ranked fund should be accorded more weight than one ranked at a lower level. more important criteria should be weighted more heavily than less important criteria. similarly, clearer criteria should be weighted more heavily than less clear criteria. in the subsections to follow, these ideas are formalized, and a model for determining a set of weights and hence, a rating ri for each mutual fund i, is presented. 2. constraints on the choice of weights, of constraint #l (discriminating among criteria): once the investor has rank ordered the k criteria (from the most important criterion, k = 1, to the least important, k = k), it is necessary that the weight accorded to criterion k, of, be at least as large as that accorded to criterion k + 1, e.$+’ (for a given rank position i). letting the 6 financial services review, 2(l) 19!j2n!w3 variable v denote the minimum gap between these two weights, the constraint c$ c$+l z v must be imposed. a further consideration in discriminating among the criteria, is that the investor may wish to distinguish more strongly between some pairs than others, by specifying a vector of relative criteria importance parameters, k = (kl, ~~ ,...,kk-i). the parame ter k’ reflects the “relative’* positioning of criteria k and k + 1, or the degree to which the investor believes that criterion k is more important than criterion k + 1. for example, if the investor specifies that k~ > k’ ,..., > k~-‘, then he wishes to discriminate (relatively) strongly between criteria 1 and 2, less strongly discriminate between criteria 2 and 3, and so on, with a relatively weak distinction made between criteria k 1 and k. formally, the constraints on criteria importance are then given by 6$-u+ k+l _ vickzo for k = 1,2 ,..., k 1 and i = 1.2 ,..., m. (3) in this format it is clear that, while the parameter ~~ reflects the relative gap between the weights on criteria k and k+ 1, the product vtck provides the absolute gap between the weights. the role of v will become clearer when the full model is presented below. to illustrate what relative criteria importance parameters mean, assume that k = 4 and that, regardless of what weights are actually used to represent the 4 criteria, the investor believes that the difference between c$ and of is twice the difference between c$ and 0: (for a given rank i). if k’ = 6, then k’ = 3 or, equivalently, k’ /k’ = 2. similarly, if (c$ 0:) = 3(c.$ 0;‘) with k’ = 6 and ~~ = 3, then k~ = 1 and k2/k3 = 3. if, contrary to the above example, the investor believes that the relative positioning should reflect equal spacing, then he should choose k’ = k2 = k3= 1. constraint #2 (discriminating among rank positions) once the investor has ranked the m mutual funds according to each of the k criteria (from the best fund, i= 1, to the worst, z=m), it is necessary that the weight accorded to rank position i, of, be at least as large as that for rank position i + 1, a$+, (for a given criterion k). letting the variable p (which plays a similar role to v in constraint #l) denote the minimum gap between these two weights, the constraint c$ c.$+, r ~1 must be imposed. however, this constraint assumes that the minimum difference, u, is the same for all rank positions, and ignores the information regarding relative rank positions that exists in the cardinal data available.14 the cardinal data can be used to define a vector of relative rank difference parameters. for criterion k, define a vector ak = (hf, g ,. . . , a&_,) where ?$ (4 c$+,) / (6 &), c$ is the cardinal observation for the z’th ranked fund, 0 s hf s 1 and cei’ ?$ = 1 .15 the larger hf is, the larger is the relative difference between the z’th and z+l’th ranked security (according to criteria k). if only ordinal data is available for criterion k, then h might be chosen as jl~ = 1 / m 1 for 1 = 1,2,..., m 1. for example, if there are five mutual funds and criterion 1 refers to the ai’s, a muh’icriteria appmuch to mutual fund sebtion i withc~=.10,c~=.08,c~=.02,c~=oandc~=-.10,thenh~=.10,~=.30,h~=.10 and ai = so. note that ai, and hence u$ w: is relatively large because the difference between ci and c: is relatively large. if only ordinal observations are available for criterion i, then h! = .25 for i= l,..., 4. while the vector of relative rank difference parameters, a’, recognizes the differences among rank positions for a given criterion k, it does not reflect the relative importance of criterion k itself. for this reason, multiply each a; by a contraction factor tk = zck kj/z& ~j, which reflects the required reduction in gap size (between relative rank positions) for criterion k relative to criterion 1 (the most important criterion). formally, the constraints on rank position importance are given by 4 a:+, -p5+0 for k = 1,2 ,..., k and 2 = 1,2 ,..., m 1. (3a) 4 co!+, tyo: cl$+i) s 0 fork = 1,2 ,..., k and 2 = 1,2 ,..., m 1. (3b) similar to (2) from constraint #l, the r%f reflect the relative gap between the importance attached to consecutive rank positions, while the product yz%f provides the absolute gap. furthermore, the gaps between consecutive rank positions are bounded from below by the constraint set (3a), and from above by the constraint set (3b). constraint #3 (criteria clearness discrimination): to incorporate criteria clearness into the model, the investor must be able to rank the criteria from the most to the least clear. then the variable u in the previous set of constraints (equation (3a)), is replaced by the vector of variables @[‘i, pt2], ,..., l&m), where l,pl represents the relative clarity of the clearest criterion, p t21 the relative clarity of the second clearest criterion, and so on. to reflect differences in clarity among the k criteria, the absolute sizes of the ptkl variables are incorporated into the model. then, if there is a large difference in clarity between two consecutive criteria, [k] and [k+l], the corresponding difference, $kl _ pw+ll will be large, as will be the difference between the two corresponding weights. equation (3a) (representing constraint #2) must then be replaced by the following two sets of constraints, 63: co:+, pkl%f z 0 for k=1,2 ,..., k and i = 1,2 ,..., m 1 (4) and $kl _ $k+u x 2 0 fork= 1,2 ,..., k1, (5) where x is a variable measuring the minimum gap between consecutive clearness variables. 8 financial services review, 2(l) 19921993 constraint #4: the weights must be chosen in such a manner that the overall rating accorded to mutual fund i, r, is less than or equal to some exogenous parameter, say 1, k m (6) choosing a value of 1 is just a scaling convention; any parameter value greater than zero could be chosen and would have no impact on the overall ranking of the funds. constraint #5: finally, constraints on the minimum values of the variables v and x are added, vzz and (7) where z is a non-negative parameter. 3. ranking the mutual funds given any set of weights {of}, the investor can obtain a rating ri for each of the m mutual funds. by ranking the funds in descending order according to ri, an optimal fund is identified. effectively, by choosing a set of weights, an additive utility function r,(o) is defined, where o denotes an (m x k) matrix (e$ is the (lk) element of o) and any set of weights satisfying constraint #l-constraint #5 is feasible. rather than simply, and arbitrarily, choosing any set of weights that is feasible, an approach that has been applied in efficiency analysis (see, for example, charnes, cooper and rhodes (1978)) can be adopted. this approach, applied in the present context, determines for each mutual fund i, the set of weights {e.$) that provides the most favorable rating ri. specifically, solve the following maximiza tion problem for each fund i (where c#i denotes constraint #i),16 subject to: fork = 1,2,...,k 1 and vz ml) (2) 4 dl l”(cl$ o:+,> 5 0 for 2 = 1,2,..., m1 and vk (c&i) (3b) a multkriteriu apptvuch to mutuul fund selection for i = 1,2,...,m 1 and v k fork= 1,2,...,k1 k-1 i-1 v-ztoandx-zr0 ((~5) (7),(g) of, uk, 2 2 0 v 1,k (non-negativity) 9 (c#2,#3) (4) (c#3) (5) vw (6) a simple example: a simple example, using one criterion and two funds, illustrates how the maximization problem is solved. while the example is trivial in terms of identifying the best fund, it is useful in illustrating the geometry of the maximization procedure. let the single criterion be a, with a, = 0.30 and a2 = 0.05, so that a r, = 1, u22 = 1 (see figure 1) and h: = h = 1.0 (the k superscript is omitted throughout this example to simplify the exposition). from (l), rr = q1 and r2 = 022, where an additional subscript is introduced to denote the fund being rated.” next, observe that discriminating among criteria and criteria clearness are not issues in this example. hence, constraints #l and #3 are not relevant, and the vector ~1 in (4) can be represented by a single variable. finally, add u z z to constraint #5 and let z = 0.25. theconstraintsetforthetwomaximizationploblems,max1,,,~r,andmax(,),~r2js then oil oi2 u 2 0 (constraint #2) oil 5 1 and 0i2 s 1 (constraint #4) u l 0.25 (constraint #5) oil z 0 and oi2 z 0, (non-negativity) for i = 1 and i = 2 respectively. the constraint set for this problem is represented by the hatched region in figure 2. in solving “max r,” one obtains o;2 = 1 and o;, is any value in the closed interval [0,0.75]. for example, take a;2 = 0 and call 10 financial !services review, 2(l) 199m993 1 e’ fund #2 i fund #1 1 a,, razr figure 1. requirements space. constraint set 0, 0, p 10 o,,~,~i p i 0.25 o,, o,>o figure 2. solution space for h = 1 .o, z = 0.25. a mu&i&&&z appvuch to mutual fund s&dim 11 a12ra22 i figure 3. requimments space for h = 1 .o, z = 0.25. this point a. similarly, in solving “max r2)‘, one obtains e& = 0.75 and o;, = 1; call this point b. next, observe that constraint #4 defines two hyperplanes (one for both max r, and max r2, &la21 + g2a2* = 1, which are drawn in figure 3. note that the solution to the second maximization problem (max r2) is given by the ratio of oq / or = 1 / 1.33 = 0.75, while the solution to the first maximization problem (max r,) is given by op = 1. in the literature on efficiency measurement (see, especially, chames, et al (1978)), the line segment rp and the vertical segment em~ating downward from p, are referred to as the efficient frontier. points on the frontier are said to be efficient, while points beneath the frontier (toward the origin) are inefficient. consequently, in figure three, mutual fund #l (point p) is effkient, while fund ##2 (point #iq) is inefficient. solving tbe maximization problem for a specific mutual fund is then equivalent to de~~mg the position of the fund with respect to the effiient frontier. note that, if in this example we had set a, = a2 so that h = 0 (i.e., there is no distinction between the first and second rank positions), then constraint ##2 would have been wil oi2 z 0 for i=1,2. in this case, the constraint set would appear as 22 financial services review, z(1) 19924993 1 01 figure 4 solution space for il = 0.0, t = 1.0. shown in figure 4, and the corresponding hyperplanes would be as shown in figure 5. in this case, both mutual funds lie on the frontier, hence both are efficient. al~matively, if h= 1 .o and z is increased to 0.50, the co~es~nd~g ef~cient frontier is shown in figure 6. here, r2 = w; 2 = 0.5. note that oqior=zo.s as well. while it is difficult to construct a geometric image of higher dimension problems, it can be seen from this simple example that increasing a parameter such as z acts to decrease the likelihood of a given mutual fund being classed as figwe 5. requirements space for h = i .o, l = 1 .o. a multicriteriu approach to mutud fund selection 13 figure 6 requirements space for a = 1 .o, z = 0.5. “efficient”. the same is true, in general, if k~, h:, or h are increased. thus, as z is increased, the set r* = ( ilr7 = 1 } of efficient funds becomes progressively smaller, until z reaches its maximum (beyond this value of z, there is no feasible solution to the constraint set). the number of funds in the set r* is therefore determined, to a great extent, by the value of the parameter z. if, for example, z = 0, then v and x may both be zero, and the maximization problem enjoys tremendous flexibility in its choice of weights, {of}. in order to decrease the number of funds in the set r*, the degree of flexibility allowed must be reduced. this can be accomplished by increasing incrementally the value of the parameter z. a faster and more efficient method of accomplishing this is now presented. 14 financial services rrviiew, 2(l) 1992m93 the rn~~~ation problem at the beginning of this subsection can be modified slightly so that it identifies the largest value of z for which the set of constraints is feasible. that is, solve for the maximum value of z, subject to the set of constraints listed above, for which the set r* is not empty. k-l l-1 y-zz0 and x;-zz0 (7)@9 for k = 1,2,...,k 1 and vl for i= 1,2,...,m 1 and k% for 1 = 1,2,...,m 1 and vk fork 1,2,...,k -1 (2) gb) (41 (3 (6) at the optimum z*, constraint 114 will be binding for at least one fund i*, i.e., such a mutual fimd i* will be optimal for the investor.” section four now provides an example of how to use this approach to rate ten funds, evaluated on the basis of the six criteria suggested in section 2. iv. anexample this section demonstrates the multic~~~a approach to rating a set of mutual funds and illustrates that investors with different preferences do, in general, obtain different fund ratings. this section also demonstrates that, in general, the multicriteria approach will obtain a different set of ratings than that obtained by employing jensen’s single criterion approach. to demonstrate the multicriteria approach, two types of information must first be provided: 1) data: for each of the 10 funds, an observation (either cardinal or or~nal) for each of the 6 criteria must be provided a m&k&via appmach to mutual fund sek&n 15 (the data is employed in the u$ and lf papers). 2) parameter values: the investor must specify values for the vector of criteria importance parameters (k) and a ranking of the six criteria according to their relative clarity (recall the ptcl). for this example, ten mutual funds were randomly selected. for the required data, ai and <ji were calculated using thirty-six monthly observations from 1986-89 (obtained from the mutual fund sourcebook). information on d, fi, and bi wzls obtained directly from the selected mutual funds. quantifying si was quite difficult and was done on a fairly arbitrary basis. the two types of parameter values required would normally be supplied by the investor performing the ranking. for this example, a single ranking of the six criteria according to their relative clarity is provided. it is assumed that tr[ll= utzl = $41 = $51 > $31 = ~“1, where [l] refers to cy.+ [2] to or,, [3] to d, [4] to fil [5] to bi and [6] to si. next, three vectors of criteria importance parameters are assumed, each representing a different (hypothetical) investor’s set of preferences. this is done to, as simply as possible, demonstrate that investor’s with different preferences do, in general, obtain different fund ratings. as mentioned in the introduction, fund shareholders vary enormously in terms of at least six variables, their: wealth levels; rates of portfolio turnover; degrees of risk tolerance; understanding of financial markets; beliefs in the ability of mutual funds to outperform the market; and their own portfolio’s level of diversification. assume that the three hypothetical investors can be represented by whether they rate high (1), medium (2) or low (3) according to each of the six variables: investor #i investor #2 investor #3 wealth level 1 2 3 portfolio turnover 1 2 risk tolerance 1 2 ; understanding 1 2 3 belief in funds’ ability 1 2 3 diversification 1 2 3 consistent with the above ~presen~tion, the three vectors of criteria impo~ce parameters are assumed to be: investor #l investor #2 investor #3 rank criteria kk criteria kk criteria k’ : ii d 3 ; b 0 f 0 3f 1 s 5 :: 4 u 1 b 1 5 ;:a1 al 6 s f a 16 financial services review, 2(l) wwl993 table 1 rankings of the ten mutual funds for investors #l , #2 and #3 and for jensen’s single criterion procedure jensen ‘s single rank investor #i investor #2 investor #3 criterion procedure : 3* 1* 2 1 5* 6* 3 1 3 8 10 1 8 4 2 6 10 6 : 2 : 9 ; 7 10 9 8 10 8 9 7 4 9 9 4 5 3 4 10 7 4 7 7 * denotes a tie for first place. using the data and parameter values described above, the m~im~tion problem presented in (10) identifies the optimal mutual fund for each of the three hypothetical investors. the full ranking of the ten mutual funds for each of the three investors (using the multicriteria approach), as well as the ranking obtained by employing jensen’s single criteria approach, are presented in table 1. observe that the three hypo~etical investors obtain different fund rankiigs fromboth one another and from that obtained by using jensen’s single criterion approach. in table 1 observe that investor #l, for whom a is the most important criterion, obtains an almost identical ranking of the ten mutual funds as that provided by jensen’s single criterion procedure. also observe that, while investors #2 and #3 obtain significantly different fund rankings from that provided by jensen’s proce dure, they place a very low (fifth and sixth, respectively). four additional hypotheti cal investors are now introduced, to demonstrate that investor’s who place a as their first, second or third most important criterion, may obtain fund rankings significantly different from that provided by jensen’s procedure. the four additional vectors of criteria impo~nce parameters are assumed to be: investor #4 investor #5 investor #6 investor #7 rank criteria kk criteria kk criteria kk criteria kk 1 ii 1 ii 0 b 3 b 2 2 2 0 f 1 3 f 1 f : ii i a 1 4 1 5 : 0 ii 1 1 1 1 ii 1 ;;5 1 6 s s s s table 2 presents the full ranking of the ten mutual funds, for each of the four additional hypothetical investors. note specifically that mutual fund #3, which a multicriteria approach to mutual fund !s&xtion 17 table 2 rankings of the ten mutual funds for investors #4, #6. #6 and #7 using the multi-criteria procedure rank inveslor #4 investor #s investor #6 investor #7 1 1 1 5 2 3 5 1 3 5 3 6 4 8 6 3 5 6 8 8 6 2 2 4 7 9 4 9 8 10 9 2 9 4 10 10 10 7 7 7 5 6 : 4 8 9 2 7 10 ranks first using jensen’s single criterion procedure, ranks second for investor #4, third for investor #5 and fourth for investors #6 and #7. also, note that fund #l, which ranks second using jensen’s procedure (and places first for investors’s #l, #4 and #5), drops to second place for investor #6 and to third place for investor #7. hence it is shown how the ranking of funds change as the investors’ priorities change. further, it is demonstrated that, even when investors place a relatively highly (their first, second or third most important criterion), investors may obtain fund rankings significantly different from that provided by jensen’s procedure. v. conclusions this paper presented an approach which allows an individual investor to identify an optimal mutual fund, given his specific set of attributes and preferences. this multicriteria approach is based on the pure ordinal model of cook and kress (1991). the initial premise is that investors select mutual funds on the basis of several distinct criteria, rather than on the basis of a single criterion, such as jensen’s a coefficient (as is generally assumed in the mutual fund evaluation literature). a general procedure is developed for ranking a set of competing mutual funds, based upon the ranking of the set of funds according to each criteria and on the investor’s ranking of the criteria (in order of importance) themselves. an example, involving ten mutual funds and six criteria, demonstrated two important points. first, investors with heterogeneous attributes and preferences do, in general, obtain different fund rankings from one another. second, the fund rankings obtained using the multicriteria approach differ from those obtained using jensen’s single criterion method. also, jensen’s single criterion approach predicts that investors unanimously agree on their fund ratings (and rankings) and hence, they all select the same mutual fund. the implication of the multicriteria methodology that investors obtain different fund rankings and hence, select different funds to 18 financial services review, 2(l) 19!wl993 purchase is, however, much more consistent with the large number of mutual funds currently sold in the u.s. acknowledgment: supported under nserc grant #a8966. we would like to acknowledge helpful comments by an anonymous referee. the second author was on leave at kyoto university, japan. notes 1. in january. 1990. mutual tind assets surpassed the one trillion dollar mark, with equity funds accounting for 26 percent of the total, bond funds 31 percent, money market funds 36 percent and short-term municipal bond funds 7 percent. also, note that not all mutual fund shares are held directly by individual investors. in addition to the 5 1 .l million individual accounts, there are 7.1 million institutional accounts (a majority, 3.7 million, of these are fiduciary accounts), holding 25.4 percent of all stock, bond and income fund shares and 42.5 percent of all money market shams. statistics from the investment company institute (1990). 2. while jensen’s measure is the most widely used measure in academic empirical studies of forecasting ability, it has been the subject of numerous criticisms. for a good survey of the criticisms, see grinblatt and titman (1989). 3. treynor (1965) also employed a single criterion approach, asserting that ‘the comptehensive ness of this rating is a question for the reader to decide for himself....most readers are likely to agree, however, that at least one dimension and a critical one of the quality of the investment management is analyzed by this new method.” 4. using prior rates of return to select a mutual fund could be considered a questionable practice in view of evidence that past performance contains little or no predictive value. however, allerdice and farrar (1%7), friend, bhtme and crocket (1970). smith (1978) and woerheide (1982) provide evidence that past rates of return are positively correlated with a fund’s net sales. furthermore, survey evidence presented by lewellen, lease and schlarbaum (1977) and the investment company institute (1987) indicate that prior returns am one of the primary variables used by investors when choosing mutual funds. 5. including ui to measure a fund’s performance consistency was suggested by c. poll, managing director, micropal. he evaluates funds by observing their annual rank (by that year’s return) over, for example, each of the last five years, and then selecting the fund which has most consistently exhibited a high annual rank. 6. if the mutual fund shares are to be held within a perfectly diversified portfolio, then the fund’s level of diversification is irrelevant however, the evidence is quite strong that a very large majority of households do not hold well diversified portfolios (blume and friend (1975)). further, survey data presented by iewellen. lease and schlarbaum (1977) and the investment company institute (1987) demonstrate the importance of the fund’s level of diversification to the individual investor’s selection decision. 7. to calculate di, sum the 4uared deviations of the proportions invested in each security in the market portfolio from the cotresponding proportions in the mutual fund. since the weight of each security in the market portfolio is very small, di can be approximated by the sum of the squares of the proportions invested in each security in the mutual fund. note that 0 s di s 1, with lower values representing higher levels of diversification. for example, di = 1 for a one stock portfolio and di = 0 for the market portfolio. this measure is used by blume and friend (1975) and is a special case of the diversification measure suggested by sharpe (1972); the two measures are equivalent if all securities possess the same unsystematic or residual variance. a multicriteriu bptvuch to mutud fund selection 19 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. for a description of the various services offered by mutual funds, see investment company institute (1990; p. 28-30). for survey results on the importance investors place on different fund services, see investment company institute (1987; p. 33.34). jensen’s performance index can be viewed as a special case of the multicriteria model presented here. he p1a~e.s a weight of 1 on oi, and a zero weight on oi, di. fi, bi and si. this may be a weak ordering, containing ties if the investor views two or more funds as being equally important according to a particular criterion. this may also be a weak ordering. to illustrate, recall the six criteria presented in section two. the investor may view ai as the most important criterion followed by, in decmasing order of importance, 01. di, fi, bi and si. to illustrate what is meant by clearness, recall the six criteria presented in section ii (ai, ok di, fi, bi and si) and observe that them are measurement problems associated with each of the six criteria. regarding ai and oi, roll (1978) demonstrates that, unless the exact composition of the true market portfolio is known, a fundamental ambiguity exists when ai (and hence ui) is measured using the security market line (for a demonstration of how sensitive mutual fund rankings based upon ai am to the benchmark portfolio chosen, see lehmann and modest (1987)). uncertainty regarding the exact composition of the market portfolio also imparts an ambiguity into the measurement of di (recall that di measures how closely fund i’s composition approximates that of the market portfolio). load fees, because of their dependence on the quantity purchased and the investor’s time horizon, also lack clarity. finally, the fund’s service level is possibly the “fuzziest” or most vague of the six criteria. note that, if the investor specified that all the wf of +t were exactly equal to 1 for all 1, this would be equivalent to describing the rank positions as cardinal numbers, m,m l,..., 1. many ordinal tanking models, particularly those involving consensus derivation, do in fact use such a mechanism to weight rank positions (see, for example, cook and seiford (1978)). a large number of ways to define d exist, hence this definition should only be viewed as an example. it is, however, the simplest method of employing the cardinal data first, two types of information must be provided: 1) data: for each of the m funds, an observation (either cardinal or ordinal) for each of the k criteria must be provided (the data is employed in the a$ and ?$ parameters). 2) parameter values: the investor must specify values for the vector of criteria importance parameters (k), a ranking of the k criteria according to their relative clarity (recall the &“i) and the non-negative parameter, z. when there is substantial flexibility in the choice of weights, it is generally the case that a different set of weights will arise when solving the optimization problem for one fund than for another. in this simple example, the weights determining the maximumvalueof rt are different from those for r2. hence, for this example, a second subscript is added so that oil denotes the weight attached to rank position 1 when rating fund i. in general, however, the second subscript can be omitted because the maximization problem presented in (10) below, by solving for the optimal value of z, minimizes the degree of flexibility allowed and hence, derives an identical set of weights for all m funds. it is possible that, given z*, mom than a single fund satisfies (11). references ali, i., w. cook and m. kress. 1986. “on the minimum violations ranking of a tournament,” management science, 32: 660672. allerdice, f. and d. farrar. 1967. “‘factors that affect mutual fund growth,” journal offinancial and quantitative analysis, 2: 365-382. blume, m. and i. friend. 1975. ‘the asset structure of individual portfolios and some implications for utility functions,” journal of finance, 30: 585-603. 20 financial services review, 2(l) l!xj2/1993 charnes, a., w. cooper and e. rhodes. 1978. “measuring the efficiency of decision making units,” european journal of operational research, 2: 4294. cook, w. andm. kmss. 1991.“multiplecriteriadecisionmodelwithordinalreferencedata,”euro pears journal of operational research, 54: 191-198. cook, w. and l. &ford. 1978. “priority ranking and consensus formation,“marragement science, 24: 1721-1732. friend, i., m. blume and j. crocket. 1970. mutual funds and other institutional investors: a new perspective. new york: mcgraw-hill. grinblatt, m. and s. titman. 1989. “portfolio performance evaluation: old issues and new insights,” review of financial studies, 2: 39w21. investment company institute. 1987. mutual fund shareowners: the people behind the growth. washington, d.c. investment company institute. 1990.1990 mutual fund fact book. washington, dc: author. jensen, m. 1968. ‘the performance of mutual funds in the period 1945-1964,” journal of finance, 23: 389416. lehmann, b. and d. modest. 1987. “mutual fund performance evaluation: a comparison of benchmarks and benchmark comparisons,” journal of finance, 42: 233-265. lewellen, l., r. lease and g. schlarbaum. 1977. “some evidence on the patterns and causes of mutual fund ownership,” journal of economics and business, 30: 57-67. litzenberger, r. 1975. “discussion: the allocation of wealth to risky assets,” journal of finance, 30: 624-629. roll, r. 1978. “ambiguity when performance is measured by the securities market line,” journal of finance, 33: 1051-1069. sharpe, w. 1972. “diversification and portfolio risk,” financial analysts journal, 28: 74-80. smith, k. 1978. “is fund growth related to fund performance?,” journal of portfolio management, 4: 49-54. tteynor, j. 1965. “how to rate management of investment funds,” harvard business review, 43: 63-75. woerheide, w. 1982. “investor response to suggested criteria for the selection of mutual funds,” journal of financial and quantitative analysis, 17: 129-137. financial services review, 33(1) 179 does overspending harm retirement preparation? christina lynn,1 stuart heckman,2 michael kothakota,3 and derek lawson4 abstract this study addressed the research question of how overspending is related to retirement preparation. a commonsense answer to this question is this: overspending should negatively impact retirement preparation. however, the existing body of knowledge does not provide evidence to support or deny this assumption. the behavioral life cycle hypothesis was tested as a theoretical framework to answer this research question, providing valuable insight. three data sets were used, including the survey of household economic decisionmaking (shed), the survey of consumer sciences (scf), and the national financial capability study (nfcs), to conduct logit and ols regressions in testing the hypotheses. because the overspending measurements were only negatively related to retirement preparation in a little over half the analyses, the results point to a new cultural norm where one’s overspending behavior does not necessarily reflect one’s retirement preparation behavior. results provide support for policy actions related to tightening credit card policies, exposing a lack of awareness on overspending, providing practical approaches for avoiding overspending behavior, and the value of using multiple data sets as a robustness check. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation lynn, c., heckman, s., kothakota, m., & lawson, d. (2025). does overspending harm retirement preparation? financial services review, 33(1), 179-204. introduction does the average american family feel stretched so thin they cannot afford to contribute to their retirement account? is overspending to blame? this study examines the intricate relationship between overspending and retirement preparation using data from multiple sources, including the survey of household economics and decisionmaking (shed), the survey of consumer finances (scf), and the national financial capability study (nfcs). guided by 1 corresponding author (christinalynn0809@gmail.com). kansas state university, manhattan, ks, usa 2 kansas state university, manhattan, ks, usa 3 kansas state university, manhattan, ks, usa 4 kansas state university, manhattan, ks, usa the behavioral life cycle hypothesis (blch) (shefrin & thaler, 1988), this research explores the dynamics of overspending and its impact on retirement preparation behaviors among average working adults in the united states. advisors may feel stumped when advising clients who habitually overspend. this issue is alarming, as many households are ill-prepared for retirement (board of governors of the federal reserve system, 2021). additional challenges, such as longevity risk (lim & lee, 2021), rising https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(1) 180 inflation (bennett, 2021), and a general lack of consumer awareness (fan et al., 2022), exacerbate these concerns for advisors. while extensive research exists on retirement preparation, the specific effects of overspending remain underexplored. the assumed negative relationship between overspending and retirement readiness appears logical, yet academic evidence is scant. this study seeks to address the critical question: does overspending significantly undermine retirement preparation? background review and theoretical orientation retirement preparation retirement preparation involves actions taken throughout one's working career to ensure financial wellbeing in retirement (muratore & earl, 2010; muratore & eckert, 2004). saving for the future is a vital aspect of most households' personal finances. given that retirement planning is highly individualized, there is no universal formula. the key to successful retirement preparation lies in adequately planning for one's goals and needs (adams & rau, 2011). however, inadequate retirement savings (board of governors of the federal reserve system, 2021; oakley & kenneally, 2019), longevity risk (lim & lee, 2021), inflation risk (bennett, 2021), and a lack of consumer awareness (fan et al., 2022) indicate a broader problem in retirement readiness. this study aims to better understand why the average household is underprepared for retirement, despite the critical importance of retirement planning. overspending there is no consensus in the literature on why overspending, a self-harming financial behavior, occurs. possible explanations include spending preferences influenced by social norms and cultural trends (cynamon & fazzari, 2008), social comparison (pahlevan sharif et al., 2022; watson, 2003), materialism (kimiyagahlam et al., 2019), and budgeting behaviors (choe & kan, 2021; sui et al., 2021). other researchers highlight the concept of the "pain of payment" to explain overspending (achtziger, 2022; choe & kan, 2021; rick, 2018). "pain of payment" refers to the discomfort associated with spending money, which can deter overspending when the anticipated financial consequences are significant (choe & kan, 2021). despite these insights, little is known about the relationship between overspending and retirement preparation behavior. behavioral life cycle hypothesis the blch recognizes how people often act irrationally when it comes to money (shefrin & thaler, 1988). in contrast to the flat consumption line of the lch, the consumption line under the blch model may not be straight. consumption may be higher in pre-retirement, due to low levels of self-control leading to overspending, and then decline in later years due to deficient savings, as depicted in figure 1. a dual preference framework explains the opposing forces of the doer and the planner preferences in consumption choices. a doer preference is associated with a short-term orientation, whereas planner preferences are associated with a long-term orientation. among the three behavioral factors identified by the blch, self-control, framing, and mental accounting (shefrin & thaler, 1988), this study isolates self-control as the behavioral influence impacting consumption decisions. lynn et al. 181 figure 1. blch and lch conceptual models of consumption note: adapted from www.economicshelp.org lch model. hypothesis this study proposes the following hypothesis: h1: overspending is negatively associated with retirement preparation. methodology data sets this study utilized three separate cross-sectional datasets: the 2019 shed, the 2019 scf, and the 2018 nfcs. using these nationally representative surveys adds to the credibility of the findings because each survey offers similar yet different measures of the constructs, enhancing the study’s robustness. the samples were limited to full-time workers, assuming that individuals who were not working would not be actively preparing for retirement. this restriction was based on a respondent’s employment status in each data set, ignoring the employment status of a spouse’s working status. this restriction reduced the shed sample size from 12,173 to 6,651 respondents, the scf sample size from 5,777 to 3,361, and the nfcs sample size from 5 a data limitation of this measurement was that researchers have operationalized it both as spending less than income (e.g., borsch-supan & lusardi, 2003) as well as a retirement preparation measurement (e.g., ferdous et al., 2010; heckman & hanna, 2015; kim 27,091 to 10,800. operationalizations of each measure were based upon data set availability. overspending variables from an advisor’s perspective, overspending is straightforward to measure: it occurs when spending exceeds income. to evaluate overspending, one subjective and three objective measures are utilized, offering different perspectives and enabling a comprehensive analysis of the phenomenon. spending more than income. operationalizing overspending with spending more than income (smti) was consistent with prior research (borsch-supan & lusardi, 2003).5 if income exceeds expenditures, the household was assumed to be saving. if expenditures exceed income, the household was assumed to be overspending. the survey question used to operationalize smti was similar in all three data sets, and read as, “over the past year, would you say your household’s spending was less than, more than, or equal to your household’s income?” & hanna, 2017; yuh & hanna, 2010). because retirement preparation was not operationalized in this way in the current study, it was assumed to not cause any methodological concerns, but still warranted disclosure. http://www.economicshelp.org/ financial services review, 33(1) 182 only respondents that selected smti were treated as overspending. the 22 responses in the shed and 290 in the nfcs of “don’t know” or “prefer not to say” were treated as missing data. revolving credit card debt. experts suggest that credit cards facilitate overspending because they were easy to obtain and use, they reduce the pain of payment, and the credit limit and minimum payment due displayed on statements may act as anchors that induce higher levels of consumption gärling & ranyard, 2020. revolving credit card debt (rccd) was also considered financially damaging because of the interest and fees charged. all three surveys had similar questions to operationalize rccd. if a respondent indicated that they had a credit card and did not pay off the balance every month, they were treated as rccd. the measure had three categories: (a) credit card revolvers, (b) noncredit card revolvers, and (c) non-credit card holders. the 23 “don’t know” or “prefer not to say” observations in the shed and 190 in the nfcs were treated as missing data. alternative financial service usage. alternative financial services (afs) usage was a third measure of overspending. afs were highcost debt instruments, such as payday loans or tax refund anticipation loans, (robb et al., 2015) used to “…fund purchases of desired consumer products” (gärling & ranyard, 2020, p. 272). because afs products were egregiously financially disadvantageous to the consumer, utilizing them indicated spending beyond one’s budget. with the shed, if a respondent indicated they had used either a payday loan, auto title loan, or a pawn shop loan, or a tax refund advance in the past 12 months, they were treated as overspending. the 26 “refused to answer” observations were treated as missing data. the scf assesses whether a respondent had used a payday loan in the past year. in the nfcs, if a respondent had used either an auto title loan, payday loan, tax refund advance, pawn shop loan, or used a rent-to-own store in the past five years, they were treated as utilizing afs. the 266 responses of “don’t know” or “prefer not to say” were treated as missing data. absence of emergency fund. the absence of an emergency fund (aef) was the final binary measurement of overspending. maintaining an emergency fund was considered a critical component of managing personal finances to weather unexpected expenses or income dips (farrell et al, 2019). the shed question “have you set aside emergency or rainy day funds that would cover your expenses for 3 months in the case of sickness, job loss, economic downturn, or other emergencies?” operationalizes aef, where a “no” response indicated aef. the 17 “refused to answer” observations were treated as missing data. the scf question that captures aef reads, “if tomorrow you experienced a financial emergency that left you unable to pay all of your bills, how would you deal with it?” if a respondent answered anything other than “spend out of savings or investments,” they were treated as aef. the question, “have you set aside emergency or rainy-day funds that would cover your expenses for 3 months, in case of sickness, job loss, economic downturn, or other emergencies?” was used to operationalize aef in the nfcs, where “no” responses were treated as aef. the 406 “i don’t know” and “prefer not to say” responses were treated as missing data. retirement preparation variables retirement preparation was a challenging construct to measure objectively because individuals have unique goals and requirements that necessitate adjustments to life events. to address this challenge, four measurements were used to operationalize retirement preparation: three were objective measures and one was subjective. retirement account ownership. retirement account ownership was assumed to be a viable proxy for retirement preparation because researchers (e.g., lim & lee, 2021; oakley & kenneally, 2019; sturr et al., 2021) consider defined contribution plans to be the foundational vehicle for funding retirement income. in the shed, if a respondent indicated owning any of the following account types, they were treated as owning a retirement account: 401(k), 403(b), keogh, other defined contribution plan through an employer, pension with a defined benefit, ira, roth ira, savings outside a retirement account, lynn et al. 183 a business or real estate that will provide income in retirement, or other retirement savings. the 529 missing observations were treated as missing data. in the scf, if a respondent had any of the following types of accounts, they were treated as owning a retirement account: keoghs, pensions, retirement, or tax-deferred savings plans. in the nfcs, if a respondent owned any of the following types of accounts, they were treated as owning a retirement account: a pension plan, a thrift savings plan (tsp) or a 401(k), an ira, keogh, simplified employee pension (sep), or any other type of retirement account that they had set up themselves or through an employer. the 846 “don’t know” or “prefer not to say” responses were treated as missing data. active contribution to a retirement account. active contribution to a retirement account (acra) was reflective of active household retirement savings, and thus is assumed to be a viable operationalization of retirement preparation (sturr et al., 2021). in the scf, if anyone in the household made contributions to iras or keoghs in the previous year or contributed to a traditional pension, 401(k), 403(b), tsp, profit sharing plan, supplemental retirement annuity, cash balance plan, portable cash option plan, sep, simplified incentive match plan for employees (simple), money purchase plan, stock purchase plan (esop), 457 plan, or other retirement account, they were treated as acra. if respondents in the nfcs regularly contributed to a thrift savings plan (tsp), 401(k), or ira, they were treated as “yes” for acra. the 4,024 missing observations are treated as missing data. the shed did not measure acra. retirement account assets. retirement account assets require a large enough balance to sustain steady distributions for the duration of one’s retirement. while there was no one formula to calculate the “right” amount of retirement account assets, retirement account assets can represent retirement preparation because they are a source of someone’s future retirement income 6 in the shed, respondents were asked if they considered themselves to be retired, which was separate from the employment status measure used to apply the sample restriction. those who considered stream. in the shed, retirement account assets were captured with the question: “approximately how much money do you currently have saved for retirement?” the 1,702 missing observations were treated as missing data. in the scf, retirement account assets were measured by the combination of roth iras, roll-over iras, regular or other iras, keoghs, and employersponsored plans balances. the log of retirement account assets was used to account for skewness associated with this measure. the nfcs did not measure retirement account asset balance. perceived retirement preparation. perceived retirement preparation was measured by a subjective question capturing how well a respondent felt financially prepared for retirement. in the shed, respondents were treated as “yes” for perceived retirement preparation if they answered yes to the question, “do you think that your retirement savings plan is currently on track?” the 1,481 missing observations were treated as missing data.6 the scf question that captures perceived retirement preparation reads “how would you rate the retirement income you receive (or expect to receive) from all sources?” respondents could respond on a five-point likert type scale with a score of one indicative of the lowest level of perceived retirement preparation and five indicating the highest level of perceived retirement preparation. the nfcs question that captured perceived retirement preparation reads, “i worry about running out of money in retirement.” respondents could respond on a seven-point likert-type scale, which was reverse coded to directionally align with measure in the shed and scf, such that one represents the lowest level of perceived retirement preparation and seven represents the highest. the 211 “don’t know” and “prefer not to say” responses were treated as missing data. control variables where possible, the following socio-economic and behavioral characteristics were controlled for: age, marital status, race, gender, education themselves to be retired were not asked the perceived retirement preparation question, which contributed to the missing observations of this measure. financial services review, 33(1) 184 of a respondent, parental education, income, health, objective and subjective financial knowledge, risk tolerance. controlling for these factors which were known to be associated with retirement preparation (e.g., muratore & earl, 2010; ozgen & esiyok, 2020), help prevent their effects from being captured in the error term. analyses regressions were used in each data set to analyze the relationship between overspending (the independent variables) and retirement preparation (the dependent variables). listwise deletion was used to handle missing data in the shed and nfcs. the repeated-imputation inference (rii) method as well as bootstrapped standard errors, was used to handle missing data in the scf. figure 2. empirical model results descriptive statistics overspending descriptive statistics. table 1 displays the overspending descriptive statistics. in the shed, only about one-sixth of the sample considered themselves to be overspending, as measured by smti (16%). even fewer were overspending as measured by afs usage (5%). over half the respondents, however, were overspending as measured by rccd (52%), and almost half the sample overspent as measured by aef (47%). in the scf, very few respondents were overspending as measured by smti (4%) and afs usage (3%). over one-third of the sample was overspending as measured by rccd (38%). half the sample overspent as measured by aef (50%). in the nfcs, one-fifth the sample overspent, as measured by smti (20%). about one-third of the sample overspent, as measured by afs usage in the past five years (32%). almost half the sample overspent, as measured by aef (46%), and over a third of the sample overspent, as measured by rccd (36%). lynn et al. 185 retirement preparation descriptive statistics. retirement preparation descriptive statistics were also listed in table 1. in the shed, over four-fifths of the sample owned a retirement account (84%), but a smaller portion felt on track for retirement (52%). almost half of the sample had under $50,000 saved for retirement (46%), whereas almost a fourth of the sample had under $10,000 saved for retirement (24%). in the scf, the majority of respondents owned a retirement account (64%), with about half of respondents acra (49%). on a scale from zero to $16,400,000, the median retirement account balance was $288,783. the average respondent reported moderate perceived retirement preparation (m = 3.03, sd = 1.25), on a scale from one to five. in the nfcs, four-fifths of the sample owned a retirement account (80%) and even more indicated acra (88%). perceived retirement preparation had a mean of 3.21 on a scale of one to seven. regression results: retirement account ownership the full list of results was displayed in table 2. the shed results show those who smti had 31% lower odds of owning a retirement account (or = 0.69, p < 0.001); those who rccd had 22% lower odds of owning a retirement account (or = 0.78, p < 0.001); those with aef had 65% lower odds of owning a retirement account (or = 0.35, p < 0.001). the scf results reveal that those who rccd had 26% lower odds of owning a retirement account (or = 0.74, p < 0.001); those with aef had 30% lower odds of owning a retirement account (or = 0.70, p < 0.001). the nfcs results show that those who rccd had 33% lower odds of owning a retirement account (or = 0.67, p < 0.001), and those with aef had 53 percent lower odds of owning a retirement account (or = 0.47, p < 0.001). financial services review, 33(1) 186 table 1. descriptive statistics shed, n = 6,651 scf, n = 3,361 nfcs, n = 10,800 variable n % n m/% sd min max n m/% sd min max retirement preparation owns a retirement account 5,104 83.50% 2,135 63.53% 8,012 80.49% actively contributes to a retirement account 1,652 49.15% 5,954 87.87% retirement account assets 3,361 $2,000 $288,783 0 $16,400,000 <$24,999 1,732 35.00% $25,000-$249,999 1,902 38.43% $250,000 $1,000,000+ 1,314 26.57% perceived retirement preparation: on track 5,170 51.59% 3,361 3.03 1.25 1 5 10,589 3.21 1.93 1 7 overspending spends more than income 1,034 15.60% 123 3.67% 2,104 20.02% revolves credit card debt 3,437 51.87% 1,288 38.31% 3,766 35.50% uses afs 351 5.30% 101 3.01% 3,356 31.86% absence of emergency fund 3,143 47.37% 1,678 49.93% 4,808 46.26% note. analyses were weighted. data from the 2019 shed, 2019 scf, and 2018 nfcs. samples restricted to only those working as a paid employee. lynn et al. 187 table 2. binary logistic regression of retirement account ownership shed, n = 5,069 scf, n = 3,361 nfcs, n = 9,114 variable b se or b se or b se or intercept 0.39 0.34 1.48 -7.43*** 0.87 0.00 -1.14*** 0.3 0.32 spends more than income -0.36* 0.14 0.69 0.03 0.22 1.03 0.02 0.09 1.02 revolves credit card debt -0.25*** 0.04 0.78 -0.30*** 0.07 0.74 -0.40*** 0.06 0.67 uses afs 0.14 0.23 1.15 0.14 0.19 1.15 -0.09 0.09 0.91 absence of emergency fund -1.05*** 0.15 0.35 -0.35*** 0.07 0.70 -0.75*** 0.08 0.47 control variables age (under 35) 0.01** 0.00 1.01 35-44 0.54*** 0.16 1.71 0.28** 0.10 1.32 45-54 0.74*** 0.18 2.10 0.46*** 0.11 1.59 55-64 0.92*** 0.20 2.51 0.72*** 0.15 2.06 65 or older 1.06* 0.43 2.90 0.42 0.30 1.53 marital status (married) living with partner or never married -0.49*** 0.14 0.61 0.10 0.10 1.11 -0.32*** 0.09 0.72 separated or divorced -0.29 0.18 0.75 -0.01 0.11 0.99 -0.28* 0.13 0.75 widowed 1.53 0.96 4.63 -0.74** 0.27 0.48 race of respondent (white) -0.01 0.08 0.99 black 0.10 0.18 1.11 -0.05 0.08 0.95 hispanic -0.12 0.17 0.95 -0.45*** 0.13 0.64 asian or other -0.05 0.24 0.87 0.10 0.17 1.11 gender of respondent (male) 0.15 0.12 1.17 0.12 0.10 1.13 0.38*** 0.08 1.45 education (high school degree) no high school degree -0.79** 0.27 0.45 -0.76*** 0.20 0.47 -0.69 0.37 0.50 financial services review, 33(1) 188 some college 0.18 0.15 1.20 0.13 0.09 1.14 0.12 0.11 1.12 bachelor's degree 1.03*** 0.18 2.80 0.62*** 0.12 1.86 0.61*** 0.15 1.85 advanced 0.90*** 0.13 2.46 0.55** 0.19 1.74 parental education (high school degree) no high school degree 0.01 0.23 1.01 -0.36** 0.12 0.70 0.02 0.19 1.02 some college 0.06 0.16 1.06 -0.08 0.11 0.92 -0.04 0.11 0.96 bachelor's degree -0.14 0.17 0.87 0.03 0.10 1.03 -0.48*** 0.13 0.62 advanced degree 0.13 0.22 1.14 -0.39* 0.17 0.67 risk tolerance 0.08** 0.02 1.08 0.02 0.02 1.02 0.10*** 0.02 1.10 income (under $25,000) 0.64*** 0.07 1.90 $25,000-$49,999 0.70*** 0.18 2.02 0.67*** 0.12 1.94 $50,000-$74,999 1.00*** 0.19 2.71 1.27*** 0.13 3.56 $75,000-$99,999 0.91*** 0.22 2.46 2.00*** 0.17 7.39 $100,000-$149,999 1.43*** 0.24 4.20 2.15*** 0.18 8.63 $150,000 or more 1.40*** 0.30 4.04 2.43*** 0.27 11.36 objective financial knowledge 0.34*** 0.06 1.41 0.17*** 0.04 1.19 0.18*** 0.03 1.20 subjective financial knowledge -0.01 0.02 0.99 0.01 0.03 1.01 financial attitude 0.06** 0.02 1.06 health (excellent) fair -0.14 0.12 0.87 -0.16 0.09 0.85 poor 0.27 0.44 -1.18** 0.34 0.31 model fit statistics c-statistic 0.88 log likelihood -1,772 c-statistic 0.83 pseudo r2 0.32 pseudo r2 0.20 pseudo r2 0.24 note. analyses were weighted. scf income was logged. data from the 2019 shed, 2019 scf, and 2018 nfcs. restricted samples to only those working full-time. * p < .05; ** p < .01; *** p < .001. lynn et al. 189 retirement results: active contribution to a retirement account the scf results (table 3) indicate that those who rccd had 21% lower odds of acra (or = 0.79, p < 0.001). the nfcs results show that those who reported aef were associated with 37% lower odds of acra (or = 0.63, p < 0.001). table 3. binary logistic regression of active contributions to a retirement account scf, n = 3,361 nfcs, n = 6,351 variable b se or b se or intercept -6.16*** 0.80 0.00 0.95*** 0.44 2.57 spends more than income 0.06 0.22 1.06 -0.16 0.13 0.86 revolves credit card debt -0.24*** 0.06 0.79 -0.11 0.13 1.00 uses afs -0.02 0.18 0.98 -0.46 0.11 0.89 absence of emergency fund -0.12 0.07 0.89 -0.11*** 0.11 0.63 control variables age (under 35) -0.01* 0.00 0.99 35-44 0.08 0.14 1.09 45-54 0.00 0.14 1.00 55-64 -0.35* 0.15 0.71 65 or older -0.83** 0.24 0.44 marital status (married) living with partner or never married 0.19* 0.08 1.21 0.13 0.12 1.15 separated or divorced 0.02 0.10 1.02 0.01 0.16 1.01 widowed -0.37 0.34 0.69 race of respondent (white) 0.01 0.11 1.01 black -0.09 0.09 0.91 hispanic -0.34** 0.12 0.71 asian or other 0.28* 0.12 1.32 gender of respondent (male) 0.23* 0.09 1.26 0.03 0.10 1.03 education (high school degree) no high school degree -0.76*** 0.19 0.47 -0.73 0.58 0.48 some college -0.09 0.08 0.91 -0.07 0.17 0.93 bachelor's degree 0.26** 0.09 1.30 -0.17 0.18 0.84 advanced 0.16 0.13 1.17 -0.50* 0.21 0.60 parental education (high school degree) no high school degree -0.33** 0.10 0.72 -0.29 0.23 0.74 some college 0.10 0.10 1.11 0.26 0.16 1.30 bachelor's degree -0.04 0.08 0.96 0.07 0.15 1.07 financial services review, 33(1) 190 advanced degree 0.20 0.19 1.23 risk tolerance 0.01 0.02 1.01 0.03 0.02 1.03 income 0.55*** 0.06 1.73 $25,000-$49,999 0.60* 0.24 1.82 $50,000-$74,999 0.95*** 0.25 2.58 $75,000-$99,999 1.10*** 0.25 3.01 $100,000-$149,999 1.53*** 0.27 4.60 $150,000 or more 2.18*** 0.32 8.84 objective financial knowledge 0.11* 0.04 1.12 0.00 0.04 1.00 subjective financial knowledge -0.01 0.01 0.99 0.07 0.05 1.08 financial attitude -0.06 0.03 0.94 health (excellent) fair -0.13 0.09 0.88 poor -0.60 0.31 0.55 model fit statistics log likelihood -2,096 c-statistic 0.69 pseudo r2 0.10 pseudo r2 0.06 note. analyses were weighted. scf income was logged. data from the 2019 scf, and 2018 nfcs. restricted samples to only those working full-time. * p < .05; ** p < .01; *** p < .001. regression results: retirement account balance the shed results (table 4) reveal that rccd (b = -0.29, p < 0.001) and aef (b = -0.92, p < 0.001) had a significant negative relationship with retirement account balance. smti and afs usage did not have a significant relationship with retirement account balance. the scf results indicate that rccd (β = -0.05, p < 0.001) and aef (β = -0.08, p < 0.001) had significant negative relationships with retirement account balance. lynn et al. 191 table 4. ordinal (shed) and ols (scf) regressions of retirement account asset balance shed, n = 4,207 scf, n = 3,361 variable b se b se b β intercept -13.35*** 1.33 spends more than income -0.18 0.12 -0.09 0.39 0.00 revolves credit card debt -0.29*** 0.05 -0.64*** 0.12 -0.05 uses afs -0.27 0.28 -0.24 0.39 -0.01 absence of emergency fund -0.92*** 0.09 -0.98*** 0.15 -0.08 control variables age (under 35) 0.05*** 0.01 0.12 35-44 1.74*** 0.13 45-54 2.63*** 0.14 55-64 3.21*** 0.15 65 or older 3.39*** 0.18 marital status (married) living with partner or never married -0.21* 0.11 0.27 0.17 0.01 separated or divorced -0.30* 0.13 0.28 0.21 0.02 widowed -0.43 0.26 race of respondent (white) black -0.83*** 0.15 -0.54* 0.22 -0.03 hispanic -0.42** 0.14 -0.90** 0.28 -0.05 asian or other 0.07 0.16 0.43 0.31 0.02 gender of respondent (male) -0.20* 0.08 0.08 0.15 0.01 education (high school degree) no high school degree -0.07 0.46 -0.80** 0.31 -0.03 some college 0.01 0.13 -0.01 0.20 0.00 bachelor's degree 0.42 0.14 1.20*** 0.21 0.09 advanced 1.47*** 0.22 0.11 parental education (high school degree) no high school degree -0.33 0.17 -0.82*** 0.23 -0.05 some college -0.04 0.12 -0.13 0.21 -0.01 bachelor's degree or beyond -0.02 0.12 0.08 0.18 0.01 advanced degree -0.04 0.12 risk tolerance 0.14*** 0.02 0.13*** 0.03 0.06 income (under $25,000) 1.35*** 0.12 0.43 $25,000-$49,999 0.53 0.34 $50,000-$74,999 1.27*** 0.33 $75,000-$99,999 1.70*** 0.34 financial services review, 33(1) 192 $100,000-$149,999 2.19*** 0.34 $150,000 or more 2.87*** 0.35 objective financial knowledge 0.19*** 0.05 0.51*** 0.09 0.07 subjective financial knowledge 0.05 0.03 0.02 health (excellent) fair -0.16 0.08 -0.34 0.18 -0.03 poor -0.22 0.48 -1.46* 0.64 -0.03 model fit statistics pseudo r2 0.34 r2 0.30 adjusted r2 0.29 note. analyses were weighted. scf income was logged. data from the 2019 shed, 2019 scf, and 2018 nfcs. restricted samples to only those working full-time. * p < .05; ** p < .01; *** p < .001 regression results: perceived retirement preparation the shed results (table 5) show that those who smti had 43% lower odds of feeling on track for retirement (or = 0.57, p < 0.001); those who rccd had 23% lower odds of feeling on track for retirement (or = 0.77, p < 0.001); and aef had 65% lower odds of feeling on track for retirement (or = 0.35, p < 0.001). the scf results indicate that smti (β = -0.06, p < 0.001), rccd (β = -0.05, p < 0.001), afs usage (β = 0.04, p < 0.001), and aef (β = -0.06, p < 0.001) had a significant negative relationship with perceived retirement preparation. the nfcs results indicate that only those who utilized afs were more likely to perceive retirement preparation negatively (β = -0.03, p = 0.003).7 7 additional robustness checks were conducted, including (a) an unweighted sample, (b) multiple imputation, (c) excluding non-credit card holders, (d) an alternative sttp, and (e) and alternative measure for smti. interaction effects were tested between overspending and income with retirement preparation, due to the suspicion that the relationship between overspending and retirement preparation may have depended on the level of income. to explore the reliability of overspending and retirement preparation constructs, additive scales, and confirmatory factor analyses (cfa) were employed. while the robustness checks reveal variation between data sets and various significant interaction effects between overspending and income with retirement preparation, there is not enough evidence to cause concern over the validity of the primary analyses. the results of robustness checks are available upon request. lynn et al. 193 table 5. logistic (shed) and ols (scf and nfcs) regression of perceived retirement preparation shed, n = 4,317 scf, n = 3,361 nfcs, n = 9,524 variable b se or b se b β b se b β intercept -0.42 0.30 0.66 0.65* 0.27 6.27*** 0.16 spends more than income -0.55*** 0.12 0.57 -0.48*** 0.11 -0.06 -0.04 0.05 -0.01 revolves credit card debt -0.26*** 0.05 0.77 -0.15*** 0.02 -0.05 0.01 0.03 0.00 uses afs 0.21 0.24 1.23 -0.36** 0.12 -0.04 -0.14** 0.05 -0.03*** absence of emergency fund -1.06*** 0.09 0.35 -0.16*** 0.04 -0.06 0.02 0.04 0.01 control variables age (under 35) -0.00 0.00 0.00 35-44 0.13 0.12 1.14 -0.26*** 0.05 -0.06*** 45-54 0.09 0.13 1.09 -0.37*** 0.05 -0.08*** 55-64 0.36** 0.13 1.43 -0.34*** 0.06 -0.06*** 65 or older 0.32 0.21 1.38 -0.20 0.10 -0.02 marital status (married) living with partner or never married -0.24* 0.11 0.79 0.08* 0.04 0.03 0.05 0.05 0.01 separated or divorced -0.32 0.14 0.73 0.04 0.05 0.01 0.01 0.06 0.00 widowed 0.13 0.47 1.14 0.07 0.16 0.00 race of respondent (white) -0.04 0.04 -0.01 black -0.10 0.15 0.91 0.22*** 0.06 0.06 hispanic -0.57*** 0.15 0.99 0.03 0.08 0.01 asian or other -0.01 0.18 0.56 -0.21** 0.06 -0.04 gender of respondent (male) -0.03 0.09 0.97 -0.12** 0.04 -0.04 -0.08* 0.04 -0.02* education (high school degree) no high school degree -0.61 0.39 0.55 0.08 0.11 0.00 0.16 0.23 0.01 some college -0.02 0.13 0.98 -0.04 0.06 0.00 0.02 0.06 0.01 bachelor's degree 0.32* 0.13 1.37 0.12 0.07 0.04 0.07 0.07 0.02 financial services review, 33(1) 194 advanced degree 0.09 0.07 0.03 0.02 0.08 0.00 parental education (high school degree) no high school degree -0.05 0.20 0.95 -0.10 0.07 -0.02 0.09 0.10 0.01 some college -0.11 0.12 0.90 -0.02 0.07 0.00 0.03 0.06 0.01 bachelor's degree or beyond 0.04 0.12 1.04 0.01 0.05 0.00 0.08 0.06 0.02 advanced degree 0.01 0.13 1.01 0.11 0.07 0.02 risk tolerance 0.12*** 0.02 1.12 0.06*** 0.01 0.12 -0.01 0.01 -0.01 income 0.19*** 0.02 0.27 $25,000-$49,999 0.79*** 0.22 2.19 -0.02 0.09 -0.00 $50,000-$74,999 0.98*** 0.21 2.66 -0.07 0.09 -0.01 $75,000-$99,999 1.06*** 0.22 2.89 -0.08 0.09 -0.02 $100,000-$149,999 1.08*** 0.21 2.95 0.07 0.10 0.01 $150,000 or more 1.33*** 0.23 3.79 0.34** 0.11 0.05 objective financial knowledge 0.01 0.05 1.01 -0.01 0.03 -0.01 -0.03* 0.01 -0.02 subjective financial knowledge 0.07*** 0.01 0.11 0.05** 0.02 0.03 financial attitude -0.62*** 0.01 -0.63 health (excellent) fair -0.49*** 0.09 0.61 -0.24*** 0.05 -0.09 poor -1.03* 0.42 0.36 -0.73*** 0.19 -0.07 model fit statistics c-statistic 0.81 r2 0.16 r2 0.42 pseudo r2 0.23 adjusted r2 0.15 note. the analyses were weighted. scf income was logged. data from the 2019 shed, 2019 scf, and 2018 nfcs. samples restricted to only those working full-time. * p < .05; ** p < .01; *** p < .001. lynn et al. 195 discussion out of all the retirement preparation analyses conducted, a significant negative relationship was found in many but not all the analyses, partially supporting h1. the lack of consistency in results across the data sets temper the conclusions drawn. retirement account ownership overspending, as measured by smti, was found to have a significant negative relationship with retirement account ownership in the shed analysis, but not in the scf or nfcs analyses. rccd was significant negatively related to retirement account ownership in the shed, scf, and nfcs analyses. afs usage was not related to retirement account ownership in the shed, scf, or nfcs analysis. aef was negatively associated with retirement account ownership in both the shed, scf, and the nfcs analyses. because some overspending measures, but not all, exhibited a significant negative relationship with retirement account ownership, only partial support for h1 was noted. active contributions to a retirement account overspending, as measured by smti and afs usage, was not related to acra in the scf or nfcs analyses. rccd was found to have a significant relationship with acra in the scf analysis, but not the nfcs analysis. aef was negatively associated with acra in the nfcs analysis, but not in the scf analysis. as such, h1 was partially supported. 8 a possible explanation for only about half of the overspending and retirement preparation analyses being significant is that overspending has become normalized in american culture. a few ideas are provided here to describe what normalized overspending behavior could look like. first, respondents may not realize they are spending more than their income. second, credit cards could be considered necessary for some households to manage their cashflow, and thus does not significantly relate to retirement preparation behavior. third, afs usage could be seen by some as an acceptable financing resource to occasionally bridge cashflow gaps when needed and does not correspond to a shift in one’s retirement preparation behavior. fourth, an emergency retirement account balance overspending, as measured by smti and afs usage, were not related to a respondent’s retirement account balance in the shed analysis or the scf. rccd was negatively related to retirement account balances in the shed and scf analyses. aef was negatively associated with retirement account balances in both the shed and scf analyses. once again, h1 was only partially supported. perceived retirement preparation overspending, as measured by smti and credit card revolving, was found to have a significant negative relationship with perceived retirement preparation in the shed and scf analyses, but not the nfcs analysis. overspending, as measured by afs usage, was negatively associated with perceived retirement preparation in the scf and nfcs analyses, but not the shed analysis. overspending, as measured by aef, was negatively related to perceived retirement preparation in both the shed and scf analyses, but not the nfcs analysis. because some, but not all, overspending showed a significant negative relationship with perceived retirement preparation, h1 could only be partially supported.8 implications borrowing using credit cards is incredibly easy for the average consumer, both for those who can afford it and those who cannot. one policy implication is to make it more challenging to qualify or transact using consumer credit (gärling & ranyard, 2020). for example, lenders could fund may seem out of reach or not necessary for cash flow management by some but does not necessarily mean a household is not preparing for retirement. fifth, overspending could be occurring simultaneously while preparing for retirement. sixth, automatic enrollment and automatic contribution increases by employers could be the cause of a significant negative relationship between overspending and retirement preparation where the employer, rather than the respondent, is responsible for the respondent’s retirement preparation behavior. in sum, the analyses may reveal a new normal where overspending does not necessarily reflect a respondent’s retirement preparation behavior. financial services review, 33(1) 196 require setting credit limits with reference to a client’s respective affordability level or limit the number of credit cards that may be issued to an individual (gärling & ranyard, 2020). because credit limits and minimum payments have been found to unintentionally act as anchor points (gärling & ranyard, 2020; tescher and stone, 2022), tightening the minimum payment due formula could potentially help households get out of credit card debt faster than current standards (tescher & stone, 2022). an industry implication to help promote the use of emergency funds is to implement the automatic enrollment of employees in savings accounts linked to their 401(k)s, which may potentially be matched by employers. this is now possible due to the recent passing of the secure act 2.0 (senate finance committee, 2022). studies have shown that automatic enrollment increases participation rates (see beshears et al., 2010). advisors can also make clients aware of this new opportunity. client overspending behavior is an occasional obstacle advisors face. despite a financial planner’s best efforts, it can seem like some clients at risk of running out of money fail to realize the consequences of overspending. one way to indirectly assist individuals who overspend is to focus on augmenting self-control levels, rather than focusing on cutting back overspending. this may be particularly impactful in youth. at the practitioner level, as a value-add to clients, advisors could provide clients with young children or grandchildren educational games or activities that teach self-control in personal finances. one such example is the cashflow® game. some have suggested that potential life hacks may help increase levels of self-control when it comes to overspending. gärling and ranyard (2020) highlighted techniques originally tested by karlsson (2003) as plausible modern solutions to help bolster selfcontrol and thereby curbing overspending. additionally, financial practitioners could coach their clients to limit themselves to owning only one credit card, rather than multiple credit cards, to simplify financial decisions where borrowers often make mistakes (gärling & ranyard, 2020). alternatively, budgeting could be promoted as a tool to help prevent overspending. simplistic budgeting techniques should be promoted over complex ones (kim & choi, 2018), such as minimizing the number of categories withing a budget (kim, 2022). budgeting using mental accounting could also be encouraged, as researchers have found it may help improve savings and consumption decisions (sui et al., 2021; xiao & o’neill, 2018). xiao and o’neil (2018) also made the point that simply suggesting the practice of budgeting is not enough, because households may not know how to do so. households would benefit from advisors teaching them budgeting techniques and being shown projections of the hypothetical impact of prioritizing savings in the budget. another practitioner implication is to ensure that all clients receive education on overspending. clients may not realize that rccd or aef is a form of overspending. they may also not realize the negative implications of afs usage or smti. advisors should educate clients and prospects regarding the financial dangers of normalizing overspending. in this regard, because overspending may be a manifestation of a deeper psychological issue, it is important for advisors to be self-aware of their own limitations in providing help (lutter, 2022) . conclusion the results from this study suggest that overspending is negatively related to retirement preparation. logically, it seems obvious that overspending should be negatively related to retirement preparation. the blch was found to be a valuable theoretical framework for the study of the relationship between overspending and retirement preparation. because the overspending measurements were negatively related to retirement preparation in a little over half the analyses, it is possible that a new cultural norm where one’s overspending behavior does not necessarily mean that the person is not also preparing for retirement may be existence. this possibility is worthy of future study. references achtziger, a. (2022). overspending, debt, and poverty. current opinion in psychology, 46, 101342. lynn et al. 197 https://doi.org/10.1016/j.copsyc.2022.1013 42 adams, g. a., & rau, b. l. (2011). putting off tomorrow to do what you want today: planning for retirement. american psychologist, 66(3), 180–192. https://doi.org/10.1037/a0022131 bennett, j. n. (2021). a dollar’s worth: inflation is real. in page one economics (issue october). https://doi.org/10.5860/choice.51-1007 beshears, j., choi, j. j., laibson, d., & madrian, b. c. (2010). the impact of employer matching on savings plan participation under automatic enrollment. in david wise (ed.), research findings in the economics of aging (pp. 311–327). university of chicago press. board of governors of the federal reserve system. (2021). economic well-being of u.s. households in 2020 (5). borsch-supan, a., & lusardi, a. (2003). measuring life-cycle saving. in borschsupan, a. (ed.), life-cycle savings and public policy: a cross-national study of six countries. academic press. choe, y., & kan, c. (2021). budget depreciation: when budgeting early increases spending. journal of consumer research, 47(6), 937– 958. https://doi.org/10.1093/jcr/ucaa049 cynamon, b. z., & fazzari, s. m. (2008). household debt in the consumer age: source of growth--risk of collapse. capitalism and society, 3(2). https://doi.org/10.2202/1932-0213.1037 fan, l., stebbins, r., & kim, k. t. (2022). skint: retirement? financial hardship and retirement planning behaviors. journal of family and economic issues, 43, 354–367. https://doi.org/10.1007/s10834-021-09779z farrell, diana, fiona greig, and chenxi yu. (2019). “weathering volatility 2.0: a monthly stress test to guide savings.” jpmorgan chase institute. https://institute.jpmorganchase.com/institut e/research/household-incomespending/report-weathering-volatility-2.0. ferdous, n., pinjari, a. r., bhat, c. r., & pendyala, r. m. (2010). a comprehensive analysis of household transportation expenditures relative to other goods and services: an application to united states consumer expenditure data. transportation, 37(3), 363–390. https://doi.org/10.1007/s11116-010-9264-2 gärling, t., & ranyard, r. (2020). the psychological perspective on the antecedents and consequences of consumer borrowing. in psychological perspectives on financial decision making (pp. 267– 290). springer international publishing. https://doi.org/10.1007/978-3-030-455002_12 heckman, s. j., & hanna, s. d. (2015). individual and institutional factors related to low-income household saving behavior. journal of financial counseling and planning, 26(2), 187–199. https://doi.org/10.1891/10523073.26.2.187 karlsson, n. (2003). consumer self-control strategies: an empirical study of their structure and determinants. journal of consumer policy, 26, 23–41. kim, g. j., & hanna, s. (2017). do self-control measures affect saving behavior? journal of personal finance, 16(2), 7–20. kim, m. j. (2022). two sides of the same coin: the simultaneous effects of spending and saving needs on budget estimation. journal of financial services marketing, 0123456789. https://doi.org/10.1057/s41264-021-00132x kim, y., & choi, s. m. (2018). the effects of breadth of product categories on budgeting. advances in consumer research, 46, 440– 443. kimiyagahlam, f., safari, m., & mansori, s. (2019). influential behavioral factors on retirement planning behavior: the case of malaysia. journal of financial counseling https://doi.org/10.5860/choice.51-1007 financial services review, 33(1) 198 and planning, 30(2), 244–261. https://doi.org/10.1891/10523073.30.2.244 lim, h. n., & lee, j. m. (2021). retirement income sources and subjective financial well-being: a comparison of retirees and non-retirees. journal of financial counseling and planning, 32(3), 517–534. https://doi.org/10.1891/jfcp-19-00101 lutter, s. (2022, august 25). psychology of financial planning webinar. cfp board. https://www.cfp.net/events/2022/08/thepsychology-of-financial-planning-webinar muratore, a. m., & earl, j. k. (2010). predicting retirement preparation through the design of a new measure. australian psychologist, 45(2), 98–111. https://doi.org/10.1080/0005006090352447 1 oakley, d., & kenneally, k. (2019). retirement insecurity 2019: households’ view of the retirement crisis (issue march). ozgen, o., & esiyok, e. (2020). consumer ethics, materialism and material satisfaction: a study on turkish adolescent consumers. international journal of consumer studies, 44(1), 14–24. https://doi.org/10.1111/ijcs.12541 pahlevan sharif, s., she, l., yeoh, k. k., & naghavi, n. (2022). heavy social networking and online compulsive buying: the mediating role of financial social comparison and materialism. journal of marketing theory and practice, 30(2), 213–225. https://doi.org/10.1080/10696679.2021.190 9425 rick, s. (2018). tightwads and spendthrifts: an interdisciplinary review. financial planning review, 1(1–2), e1010. https://doi.org/10.1002/cfp2.1010 senate finance committee. (2022). secure 2.0 act of 2022. https://www.finance.senate.gov/imo/media /doc/secure percent202.0_section percent20by percent20section percent20summary percent2012-19-22 percent20final.pdf shefrin, h., & thaler, r. (1988). the behavioral life-cycle hypothesis. economic inquiry, 26(4), 609–643. sturr, t., lynn, c., & lawon, d. (2021). financial self-efficacy and retirement preparation. journal of financial planning, june, 86–98. https://doi.org/10.1002/9781119203124.ot h1 sui, l., sun, l., & geyfman, v. (2021). an assessment of the effects of mental accounting on overspending behaviour: an empirical study. international journal of consumer studies, 45(2), 221–234. https://doi.org/10.1111/ijcs.12613 tescher, j., & corey stone. (2022). revolving debt’s challenge to financial health and one way to help consumers pay it off. the brookings institution. https://www.brookings.edu/research/revolv ing-debts-challenge-to-financial-healthand-one-way-to-help-consumers-pay-it-off/ watson, j. j. (2003). the relationship of materialism to spending tendencies, saving, and debt. journal of economic psychology, 24(6), 723–739. https://doi.org/10.1016/j.joep.2003.06.001 xiao, j. j., & o’neill, b. (2018). mental accounting and behavioural hierarchy: understanding consumer budgeting behaviour. international journal of consumer studies, 42(4), 448–459. https://doi.org/10.1111/ijcs.12445 yuh, y., & hanna, s. d. (2010). which households think they save? journal of consumer affairs, 44(1), 70–97. https://doi.org/10.1111/j.17456606.2010.01158.x lynn et al. 199 appendix table a1 sample demographics of the shed sample (n = 6,651) unweighted weighted variable n m/% sd n m/% sd min max age under 35 2,039 30.66% 2,328 35.00% 35-44 1,306 19.64% 1,454 21.87% 45-54 1,268 19.06% 1,158 17.41% 55-64 1,490 22.40% 1,310 19.69% 65+ 548 8.24% 401 6.02% marital status married 3,700 55.63% 3,655 54.96% living w/ partner, single, never married 2,094 31.48% 2,241 33.69% separated/divorced 741 11.14% 651 9.79% widowed 116 1.74% 103 1.55% race white 4,495 67.58% 4,145 62.32% black 713 10.72% 797 11.98% hispanic 882 13.26% 569 8.56% other 561 8.43% 1,141 17.15% gender male 3,605 54.20% 3,426 51.52% female 3,046 45.80% 3,225 48.49% education of respondent no high school degree 159 2.39% 420 6.32% high school degree 1,244 18.70% 1,614 24.27% some college 1,990 29.92% 1,875 28.19% financial services review, 33(1) 200 bachelor’s degree or higher 3,258 48.99% 2,742 41.22% parental education (higher of mother or father's) no high school degree 544 8.52% 648 10.16% highschool degree 1,855 29.06% 1,856 29.09% some college 1,642 25.72% 1,618 25.34% bachelor’s degree 1,272 19.93% 1,208 18.93% advanced degree 1,070 16.76% 1,052 16.49% income $0-$24,999 620 9.32% 564 8.47% $25,000-$49,999 1,148 17.26% 1,088 16.35% $50,000-$74,999 1,216 18.28% 1,129 16.97% $75,000-$99,999 1,079 16.22% 998 15.01% $100,000-$149,999 1,395 20.97% 1,372 20.63% $150,000 or more 1,193 17.94% 1,501 22.56% health excellent 3,208 54.94% 3,196 57.74% fair 2,550 43.67% 2,558 43.81% poor 81 1.39% 85 1.45% objective financial knowledge 6,651 1.93 1.08 6,651 1.84 1.10 0 3 risk tolerance 6,628 4.59 2.53 6,628 4.58 2.56 0 10 note: 2019 shed. restricted sample to only those working as a paid employee. lynn et al. 201 table a2 sample demographics of the scf sample (n = 3,361) unweighted weighted variable n m/% sd n m/% sd min max age 3,361 47.46 13.40 3,361 43.96 13.02 18 90 marital status married 1,913 56.91% 1,644 48.90% living with partner or never married 924 27.49% 1,136 33.78% separated/divorced 42 13.64% 509 15.15% widowed 66 1.96% 73 2.16% race white 2,333 69.41% 2,169 64.54% black 436 12.98% 541 16.09% hispanic 362 10.78% 427 12.71% asian/other households 229 6.82% 224 6.66% gender male 2,168 64.50% 1,935 57.57% female 1,193 35.50% 1,426 42.43% education of respondent no high school degree 190 5.66% 217 6.46% high school degree 576 17.14% 673 20.01% some college 809 24.07% 960 28.56% bachelor’s degree 938 27.91% 899 26.73% advanced degree 848 25.23% 613 18.23% parental education (higher of mother or father's) no high school degree 399 11.87% 441 13.11% highschool degree 1,003 29.84% 1,059 31.50% some college 568 16.90% 623 18.53% financial services review, 33(1) 202 bachelor’s degree or beyond 1,391 41.39% 1,239 36.85% health excellent 2,819 83.88% 2,754 81.93% fair 503 14.98% 562 16.73% poor 39 1.15% 45 1.34% income (log) 3,361 $1,272,932 $13,000,000 3,361 $76,359 $557,981 0.00 $704,000,000 objective financial knowledge 3,361 2.38 0.81 3,361 2.23 0.85 0.00 3.00 subjective financial knowledge 3,361 7.42 1.99 3,361 7.11 1.98 0.00 10.00 risk tolerance 3,361 5.27 2.54 3,361 4.78 2.51 0.00 10.00 note: 2019 scf. restricted sample to only those working full-time. lynn et al. 203 table a3 sample demographics of the nfcs sample (n = 10,800) unweighted weighted variable n m/% sd n m/% sd min max age under 35 3,540 32.78% 3,757 34.79% 35-44 2,648 24.52% 2,596 24.04% 45-54 2,532 23.44% 2,452 22.70% 55-64 1,757 16.27% 1,694 15.68% 65+ 323 2.99% 301 2.79% marital status married 5,984 55.41% 5,798 53.69% single 3,499 32.40% 3,750 34.72% separated/divorced 1,169 10.82% 1,110 10.28% widowed/widower 148 1.37% 142 1.31% race white 7,737 71.64% 6,488 60.07% non-white 3,063 28.36% 4,312 39.93% gender male 5,680 52.59% 6,292 58.26% female 5,120 47.41% 4,508 41.74% education of respondent no high school degree 88 0.81% 92 0.85% high school degree 2,009 18.60% 2,281 21.12% some college 3,892 36.04% 4,203 38.92% bachelor’s degree 3,021 27.97% 2,599 24.06% advanced degree 1,790 16.57% 1,625 15.04% parental education (higher of mother or father's) no high school degree 482 4.52% 576 5.40% financial services review, 33(1) 204 highschool degree 2,941 27.59% 3,064 28.75% some college 3,064 28.75% 3,174 29.78% bachelor’s degree 2,643 24.80% 2,425 22.75% advanced degree 1,529 14.34% 1,420 13.32% income $0-$24,999 846 7.83% 966 8.94% $25,000-$49,999 2,515 23.29% 2,540 23.52% $50,000-$74,999 2,401 22.23% 2,446 22.64% $75,000-$99,999 2,010 18.61% 1,964 18.19% $100,000-$149,999 1,958 18.13% 1,877 17.38% $150,000 or more 1,070 9.91% 1,008 9.33% objective financial knowledge 10,800 3.24 1.63 10,800 3.15 1.64 0 6 subjective financial knowledge 10,591 5.22 1.27 10,591 5.23 1.29 1 7 risk tolerance 10,584 5.59 2.57 10,584 5.71 2.62 1 10 financial attitude 10,591 4.67 1.94 10,682 4.68 1.94 1 7 note: 2018 nfcs. restricted sample to only those working full time. statement of conflict of interest one of the authors of this paper (michael kothakota) worked as the head of research at the certified financial planner board of standards, inc. and was a member of its executive leadership at the time of submission. he received no compensation for his role in preparing this manuscript. academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. institutional membership to academy of financial services is $250 and includes the four annual issues of fsr. how to submit: there is a $100 submission fee payable to the academy of financial services (afs) for all submissions to fsr. submission fees should be paid online at academyfinancial.org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org. when authors pay the $100 submission fee and are not currently members, they receive their first year of afs membership at no charge. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2022 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 30, no. 1, 2022 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $250. academic subscription price is $125 and is available by joining the academy of financial services. academic subscription includes four emailed online issues of financial services review each year. hardcopies of the journal are available at additional cost. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2022 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. pii: 1057-0810(92)90002-t letter to the editor rick keamey indicated that you might want some practical feedback from a practicing financial planner on your first issue as an academic journal it may be fine, but from a practical standpoint, i did not find it very helpful. i have been in my business for 10 years and the financial field for 15 years. i hold an mba, a clu, a chfc, and numerous securities licenses including two principal’s licenses. i am a member of the iafp’s registry, and serve on its national advisory council. my firm is also a registered investment advisor with both the sec and the state of florida. all that to say that i have worked at my profession and have to deal with the practical realities, but do not presume to criticize your efforts in any regard except in that slanted respect. i have no idea of who your target audience is-only that rick said my viewpoint might be helpful to you. in that regard you are welcome to it for what it is worth. my world is one of constantly shifting laws (10 major tax acts in the last 15 years), perennially evolving financial products, volatile economic conditions, abys mally basic levels of consumer sophistication, and shrinking compensation patterns. the level of consumer sophistication is a major source of liability in my business. all the product fields are highly regulated even if enforcement and thoroughness of fee-only-delivery of advice is not. in any case, the expectations of the client are often more than can be reasonably delivered. a major concern is to convince the client that while an advisor can help quantify and analyze a risk for them, he cannot accept the investment risk for the client. law suits are rampant and the cost of defending them can bankrupt a completely honest and diligent practitioner even if he wins. yet, the client does not want generalities and abstractions. he wants practical, specific direction. unfortunately, he often bases subsequent evaluations on sudden, unforeseen current economic changes, peer pressures or contradictory input from a neighbor, financial vendor or competitor. but, back to your journal. the first article on the “game of life” is not viewed as a game by the consumer as indicated above. comments such as “clearly the family wants return after taxes” ignore the fact that the majority do not recognize the distinction, and for many of them in the minimum bracket, total return may more vii . . . nn financial services review, 2(2) 1993 often determine product selection. my comments above dictate the futile nature of the first article to me. the pension article would be straight forward, except that it ignores a number of realities. first, investment choices are frequently uninformed or simply do not do well due to changing parameters. stocks of any kind that have losses cannot be deducted inside a plan. that tends to argue for “safer” and the more income-oriented assets of the overall portfolio to be placed in the shelter. since there is not significant distinction between ordinary income and capital gains at this time, it is a stretch to assume the capital gains exclusion will come back in light of deficit problems. it is obvious to anyone that the sheltered nature of gains and current favorable treatment of retirement distributions make sheltered plans the vehicle of choice. there are also less obvious benefits, such as the creditor protection characteristics and ability to purchase permanent life insurance with before-tax dollars. the early distribution penalty is a real deterrent for those who may not yet be otherwise financially secure. the individual risk profile is a key determinant in any investment decision. i do not necessarily agree that it is in every individual employee’s best interest to make individual decisions. it increases transaction and administrative costs. in addition, the level of financial sophistication and who contributed the money are also important. the next two articles seemed to have no conclusion or hypothesis of any consequence. the conclusion of the check writing article that consumers will always respond to pricing, is certainly not a surprise to anyone. the disability article was very thorough and well-researched. the question posed failed to recognize that the individual disability product is principally a white collar, largely professional, occupation-oriented product. evaluation of claims and costs must be largely experienced rated on a historical basis. in fact, a couple of large, older carriers dominate this market because of their specialization and disability claims base. it is easy to refer to optimal product allocations, but the reason those selections are not made by formula on a daily basis is that the “optimum” is different for every individual, and what is optimal today is never optimal tomorrow. the variables are always too varied. i am afraid i always lean to the specific and transferable ideas that can be explained as simply as possible to my clients. good luck with your journal. best regards, michael m. cain financial services review, 32(2) 77 retirement planning: a moderated mediation model of cognitive beliefs, retirement planning attitude, and money availability afm jalal ahamed1 and yam b. limbu2 abstract retirement planning has been extensively studied in developed countries; however, it has received scant scholarly attention in developing nations. this study examines the role of cognitive factors associated with retirement planning intentions in the context of a developing country, focusing on financial risk tolerance and self-efficacy within the cognitive appraisal theory framework, considering the mediating role of retirement planning attitudes and the moderating impact of financial resource availability. data collected from surveys of 301 adults in dhaka, bangladesh were analyzed using a partial least squares structural equation modeling (pls-sem) approach. findings revealed that retirement planning attitudes mediate the relationship between cognitive factors and retirement planning intentions. interestingly, risk tolerance negatively impacts retirement planning intentions through attitudes, while financial self-efficacy shows a positive influence. furthermore, the availability of financial resources moderates these relationships, indicating that retirement planning attitudes significantly influence intentions when financial resource availability is low. this research contributes to the understanding of retirement planning in a developing country context, highlighting the importance of cognitive factors and financial resources. tailored retirement planning strategies should consider individual financial conditions and cognitive beliefs. the insights are valuable for policymakers and financial advisors, particularly in developing nations. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation ahamed, a. j., & limbu, y. b. (2024). retirement planning: a moderated mediation model of cognitive beliefs, retirement planning attitude, and money availability. financial services review, 32(2), 77-93. 1 corresponding author (jalal.ahamed@his.se). school of business, university of skövde, skövde, sweden 2 feliciano school of business, montclair state university, montclair, nj, usa https://creativecommons.org/licenses/by-nc/4.0/ mailto:jalal.ahamed@his.se https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(2) 78 introduction recent years have seen heightened attention given to retirement planning, particularly in developing countries. research has mostly concentrated on western and developed countries, where an aging population and a lack of retirement readiness are major concerns (hibbert et al., 2012; kung et al., 2023). these studies have underscored the criticality of understanding the motivations and practices behind individual retirement savings, particularly in light of the increasing reliance on personal savings for post-retirement life (noone et al., 2010). however, the landscape of retirement planning in developing countries presents a starkly different context, which has been relatively underexplored in the literature. developing nations, particularly in asian countries, have traditionally relied on intergenerational support systems, with children caring for elderly parents (gruijters, 2017). this cultural norm, often overshadowed by more immediate concerns such as poverty, has typically relegated retirement planning to a secondary concern. however, urbanization, increased life expectancy, and the move towards nuclear families are altering perceptions of retirement planning, even among younger demographics (macfarland et al., 2004). a report by hsbc limited (2015) on retirement concerns in 15 countries highlights this shift, with significant percentages expressing anxiety about running out of retirement funds and inadequate savings (kimiyaghalam et al., 2017). despite wide-reaching attention to retirement planning, the factors affecting attitudes and intentions towards retirement planning, particularly in developing nations, has received scant scholarly attention. numerous studies have emphasized the influence of financial literacy and knowledge on retirement planning attitudes and intentions (e.g., lusardi & mitchell, 2011; meir et al., 2016; safari et al., 2021; van rooij et al., 2012). however, little research has examined the cognitive aspects influencing these attitudes, such as financial risk tolerance and self-efficacy. moreover, retirement planning, distinguished from general saving behaviors, is complex and influenced by various demographic, psychological, and cultural factors (kimiyaghalam et al., 2017; macfarland et al., 2004; petkoska & earl, 2009). even though financial self-efficacy is a well-versed topic in personal finance, little research has explored how it impacts personal financial management behavior (goyal et al., 2022). our study aims to fill this gap by exploring how cognitive factors and financial resource availability affect retirement planning intentions in a developing country context, using bangladesh as a case study. our research enriches the literature by incorporating cognitive appraisal theory (lazarus & alfert, 1964) and examining the roles of financial risk tolerance and self-efficacy on retirement planning intention through a person’s retirement planning attitude. we investigate how these cognitive beliefs, formed through the perception and assessment of financial scenarios, influence retirement planning attitudes and intentions. we propose and test a moderated mediation model to examine the moderating effect of financial resource availability, acknowledging its significant impact on the relationship between retirement planning attitudes and intentions. this study addresses a critical gap in the literature by focusing on the context of a developing country and offers new insights into the cognitive foundations of retirement planning. it enhances understanding of the interplay between individual beliefs and financial conditions in shaping retirement planning intentions, providing valuable implications for policymakers, financial advisors, and individuals in emerging economies. literature review retirement planning intention in the realm of retirement planning, understanding the allocation of income for saving, investing, and spending during retirement is critical (kimiyaghalam et al., 2017). behavioral finance research suggests that the financial behavior of investors is shaped by their attitudes (roberts & jones, 2001). studies have identified saving attitudes as key predictors of retirement planning, with a positive orientation towards retirement linked to more extensive ahamed & limbu 79 financial planning (kimiyaghalam et al., 2017; taylor‐carter et al., 1997). much of the existing literature has centered on the impact of financial literacy and knowledge on retirement planning attitudes and intentions (lusardi & mitchell, 2011; meir et al., 2016; safari et al., 2021; van rooij et al., 2012), leaving the cognitive factors influencing these attitudes relatively unexplored. this study leverages cognitive appraisal theory (lazarus & alfert, 1964), widely used in psychology to understand how emotions are shaped by personal interpretations and evaluations of situations, to dissect the cognitive underpinnings of retirement planning. it is important to note that this theory has been applied to a variety of psychological fields, including stress research, health psychology, and emotion regulation (chen & matthews, 2003; sorić et al., 2013; yih et al., 2019). this investigation examines how financial risk tolerance and selfefficacy as cognitive dimensions affect retirement planning intentions and how retirement planning attitudes mediate this effect. moreover, we examine the moderating effect of an individual's financial availability on these relationships, providing a nuanced understanding of the interplay between cognitive factors and financial circumstances in shaping retirement planning behavior. the ensuing sections detail the study variables and a conceptual framework for this study. financial risk tolerance over the past 20 years, there has been an increase in interest among researchers, policymakers, and investment advisers in understanding risky financial decisions (grable, 2016). risk tolerance, conceptualized as an individual's propensity to engage in either riskier or more conservative investment decisions, is increasingly being recognized as a central theme in financial decision-making research. recognized as a cognitive belief, financial risk tolerance encompasses an individual's perceptions and mental processes surrounding risk, and their capacity to handle potential losses, rather than tangible behaviors or outcomes (bayar et al., 2020; grable, 2016; jacobs-lawson & hershey, 2005). following grable's (2000) definition, financial risk tolerance is described as an individual's willingness to endure uncertainty in financial decisions. an individual’s propensity for financial risktaking significantly influences their investment choices and, consequently, their overall financial behavior, including key areas such as retirement planning (bapat, 2020; garman & forgue, 2014; grable, 2016; mathew et al., 2022). financial risk tolerance is instrumental in shaping how much one allocates towards various financial safety nets like emergency and pension funds (harahap et al., 2022). the literature on general investments and retirement planning consistently indicates that risk-tolerant individuals are inclined towards high-risk assets or larger defined contribution plans, while risk-averse individuals prefer safer investments like bonds (jacobslawson & hershey, 2005; park & martin, 2022). existing research also suggests that higher risk tolerance correlates with more comprehensive planning, higher income, and inversely with age (deaves et al., 2007). however, croy et al. (2010) noted this correlation is specific to equity investment decisions, not additional contributions. in the united states, risk tolerance positively impacts retirement planning (park & martin, 2022), while in indonesia, entrepreneurs with high financial risk tolerance demonstrate more robust saving behaviors (harahap et al., 2022). contrarily, studies in india (garman & forgue, 2014; mathew et al., 2022) indicate a negative impact of risk tolerance on financial well-being, attributed to cultural factors leading to less materialistic tendencies and satisfaction with safer, lower returns. interestingly, younger indians show a trend towards riskier investments like stocks; a similar pattern was observed among financially stable bangladeshis living in urban areas (ahamed & limbu, 2018). research on financial risk tolerance primarily centers on general investment decisions, with limited focus on its impact on retirement savings plans (jacobs-lawson & hershey, 2005; park & martin, 2022). the current study seeks to bridge this gap, particularly in the context of a nonwestern developing country undergoing rapid economic and social transformations. by examining financial risk tolerance as a key cognitive belief influencing retirement planning attitudes, this research aims to enrich the financial services review, 32(2) 80 understanding of retirement planning behaviors in evolving economic landscapes. financial self-efficacy bandura's social cognitive theory, foundational in self-efficacy research (bandura, 1982, 2012), posits that an individual's self-efficacy, or belief in their capability to execute tasks and manage life's challenges, is pivotal (bandura, 2006; farrell et al., 2016). self-efficacy, a multidimensional construct encompassing beliefs about personal control and performance, significantly influences motivation and task persistence, with lower levels often leading to disengagement or reduced effort in the face of adversity (fan, 2022; furrebøe & nyhus, 2022; goyal et al., 2022). within this framework, financial self-efficacy emerges as a critical subset of general selfefficacy, significantly predicting financial behaviors (fan, 2022; goyal et al., 2022; lone & bhat, 2022). defined as the confidence in one's ability to manage personal finances effectively (fan, 2022; farrell et al., 2016), individuals with higher financial self-efficacy are more adept at controlling their finances and perceive financial challenges as opportunities rather than threats. this proactive attitude fosters achievements that further enhance financial outcomes (farrell et al., 2016). high financial self-efficacy correlates with efficient financial management and positive financial results (mathew et al., 2022). farrell et al. (2016) found it to be a strong predictor of investment and savings product ownership among women. tang et al. (2019) demonstrated its direct and indirect impact on investment decisions, mediated by thinking styles of investors. however, goyal et al. (2022) observed no significant link between financial self-efficacy and personal financial management behavior, attributing this to external factors like environmental conditions, especially in developing countries. financial self-efficacy is closely associated with emotional stability and informationprocessing capabilities, essential for effective decision-making (fan, 2022). it reflects cognitive processes such as self-assessment, goal-setting, and strategic planning. bandura (2006) emphasized the notion that individual behavior is shaped by internal experiences, environmental contexts, and perceptions. previous studies have highlighted the positive influence of financial knowledge (joo & grable, 2005; van rooij et al., 2012) and education (macfarland et al., 2004) on retirement planning. financial knowledge is also a significant determinant of financial selfefficacy and behavior (lone & bhat, 2022). consequently, in a similar line of thinking, this current study explores the relationship between behavior skills (i.e., financial self-efficacy) and attitude and intention toward retirement plans. hypothesis development mediating effect of retirement planning attitude investment decisions in retirement planning are influenced by an individual's level of risk aversion (the opposite of which is risk tolerance) towards retirement products (meir et al., 2016). however, empirical research exploring the connection between financial cognitions, such as financial risk tolerance and financial selfefficacy, and attitudes towards retirement planning remains limited. individuals with a high tolerance for financial risk typically perceive themselves as capable of managing financial challenges and opportunities, often adopting a positive financial attitude, including optimism in retirement planning (ramalho & forte, 2019). this mindset is manifested in their propensity for engaging in investments with higher risk and potential returns, indicative of a proactive stance towards retirement planning (nguyen et al., 2019; samsuri et al., 2019). individuals with a riskaverse disposition typically adopt a cautious and conservative approach to retirement planning (park & martin, 2022), whereas those who are not risk-averse are linked to positive saving behaviors and the choice of retirement-related financial products (safari et al., 2021). moreover, the association between positive attitudes towards retirement planning and the intention to actively engage in such planning is well-established (setyawan & wijaya, 2020). this relationship suggests that attitudes ahamed & limbu 81 significantly influence an individual’s commitment to long-term objectives like retirement planning. this study thus conceptualizes the mediating role of retirement planning attitude in the relationship between financial risk tolerance and retirement planning intention as pivotal. this notion is consistent with the theory suggesting that attitudes are key mediators in the cognitive processes that lead to behaviors (ajzen, 2011; białowolski et al., 2020). it serves as a vital link between an individual's tolerance for risk and their actions taken in planning for retirement. thus, the following hypothesis is advanced. h1: retirement planning attitude mediates the effect of financial risk tolerance on retirement planning intention. when a person has high financial self-efficacy, they are more confident in their ability to make financial decisions, which reflects a belief in their ability to manage financial matters effectively. by feeling capable and empowered in managing their finances, they may be more likely exhibit proactive and protective financial behavior such as budgeting, saving, and investing and a positive attitude towards finance (lown, 2011; lown et al., 2015). in other words, financial self-efficacy can shape an individual's retirement planning attitude by influencing their behavior, confidence, and emotional responses; people who believe in their financial capabilities are more likely to approach money management with a positive attitude. thus, in this study, we conceptualize the mediating role of retirement planning attitude in the link between financial self-efficacy and retirement planning intention as attitudes integrate cognitive and emotional responses, provide motivation, and ensure consistency between what people believe and how they intend to act (ajzen, 2011). accordingly, we advance the following hypothesis. h2: retirement planning attitude mediates the effect of financial selfefficacy on retirement planning intention. moderated mediation effect of money availability financial resources are one of the key factors affecting welfare in old age (herrador-alcaide et al., 2021). however, past research indicates that individuals are unprepared for retirement due to insufficient savings or limited assets (joo & grable, 2005). one of plausible explanation is that people allocate their finances based on their life stage while being limited by their available resources (safari et al., 2021). in such a scenario, the money available at a person’s current disposal is a significant condition affecting their overall retirement planning intention. typically, consumers use retirement planning to maintain a quality of life comparable to their preretirement level. previous research shows that present income and retirement planning are positively correlated (park & martin, 2022). instead of looking at a person's absolute income, we consider their available money, a more justified conceptualization for retirement planning, as an individual’s ability to spend money is determined by the amount of budget or extra funds available at the moment (badgaiyan & verma, 2015; foroughi et al., 2012). research shows that the lack of money affects a person's purchasing power, since they avoid shopping and buying if they do not have the necessary money (foroughi et al., 2012). an individual's retirement planning attitude reflects their beliefs and attitudes about retirement plans, but it doesn't necessarily determine their financial resources. when an individual has limited financial resources due to low income or high debt, their retirement planning attitude may have less influence on their retirement planning intention (mauldin et al., 2016). regardless of their attitude toward saving and planning for retirement, their retirement planning intention is affected by the availability of resources (in this case, money) (kumar et al., 2019). one's attitude towards retirement can influence their intention to save for retirement, but their actual ability to do so depends on their financial resources and circumstances. people with positive retirement attitudes may still face limitations in their ability to save for retirement if they are burdened by other financial obligations. hence, money availability can moderate the indirect effects of financial self-efficacy and financial risk tolerance on retirement planning financial services review, 32(2) 82 intentions. this conceptualization indicates a moderated mediation relationship from a statistical perspective (preacher et al., 2007). accordingly, we hypothesize the following: h3: money availability moderates the indirect effect of financial risk tolerance on retirement planning intention through retirement planning attitude. h4: money availability moderates the indirect effect of financial self-efficacy on retirement planning intention through retirement planning attitude. to better understand the hypothesized relationships, we developed a conceptual model, as shown in figure 1, which depicts the mediating role of retirement planning attitude on the direct influences of financial risk tolerance and selfefficacy on retirement planning intention. it also shows the moderating role of money availability on such mediating effects. figure 1. conceptual framework showing hypothesized relationships methodology sample and data collection a market research firm was hired to collect responses from adults living in dhaka, the capital and most populous city of bangladesh. the recruited fieldworkers distributed selfadministered survey questionnaires to respondents in key commercial areas of the city, including motijheel, gulshan, mohammadpur, banani, and uttara. the data collection period spanned from april to may 2023. out of 870 individuals approached, 329 successfully completed the survey, resulting in a response rate of 37.82%. after eliminating cases with incomplete responses, a total of 301 usable responses were retained for the analysis. table 1 provides an overview of the study population's characteristics, offering insights into their gender distribution, marital status, age range, income level, occupational sectors, and educational background. ahamed & limbu 83 table 1. demographic characteristics of respondents (n=301) variable numbers percent gender male 197 65.40 female 104 34.60 marital status married 264 87.7 unmarried 35 11.6 other 2 0.70 age median age 36 – 40 years family income median monthly family income 101,000 – 120,000 bdt (equivalent to approx. usd 913 – 1,084) occupation banker 85 28.20 governement service 96 31.90 businessman /entrepreneur 41 13.60 private service 38 12.60 other 41 13.60 education higher secondary/vocational high school 13 4.30 bachelor’s degree 3 1.0 master degree 42 14.0 doctoral degreee 17 17.0 other 226 75.10 measures we adapted four-items from joo and grable (2004) to measure financial risk tolerance (e.g., i am more comfortable putting my money in a bank account than in the stock market). all responses ranged from 1 (strongly disagree) to 7 (strongly agree). we used farrell et al.'s (2016) six-item scale to measure financial self-efficacy (e.g., i worry about running out of money in retirement). the retirement planning attitude (e.g., planning for retirement needs too much time and effort) and retirement planning intention (e.g., i participate in workshops/seminars on retirement planning) variables were measured using a sixitem scale and nine-item scale, respectively, that were adapted from previous studies (kimiyaghalam et al., 2017; macfarland et al., 2004; noone et al., 2010; petkoska & earl, 2009; van rooij et al., 2012). the measure of money availability was adapted from badgaiyan & verma (2015). data analysis method and bias checks we investigated and tested the hypothesized relationships using a two-step partial least squares (pls-sem) approach. we first evaluated the validity and adequacy of the constructs (outer measurement model), followed by testing the structural model with hypothesis testing (inner model) (hair et al., 2014). as the data were collected using a self-reported survey, multiple measures were employed to detect common method variance (cmv). firstly, we ensured the confidentiality and anonymity of all respondents. secondly, we employed harman's single-factor test as part of the exploratory factor analysis (efa) process, utilizing spss software and refraining from rotating factors. the results indicated that cmv was not a significant concern in our data, as evidenced by a single factor accounting for only 31% of the variance (hair et al., 2014). results and discussion measurement model the first step in pls-sem is the test of the measurement model by assessing convergent validity and discriminant validity (hair et al., 2014). first, we evaluated the factor loadings of the measurement items of the respective constructs and removed those that fell below the threshold of 0.60 (field, 2013). we then used the financial services review, 32(2) 84 average variance extracted (ave) values to evaluate the convergent validity of the constructs. according to hair et al. (2014), the ave values must be at least 0.50. additionally, we ensured the reliability of the constructs by using composite reliability (cr) values greater than 0.60, which exceeded the statistically acceptable threshold (shi et al., 2012). appendix 1 summarizes the validity and reliability of the constructs used in the research model. as a measure of discriminant validity, we calculated the heterotrait-monotrait ratio (htmt) of correlations (appendix 2) and applied the fornell-larcker criterion (appendix 3). we demonstrated proper discriminant validity for both indices by satisfying the recommended thresholds (hair et al., 2014). structural model and hypothesis testing once the reliability and validity of the construct have been established, the next stage in partial least squares structural equation modeling (pls-sem) involves conducting a path analysis to evaluate the proposed direct and indirect relationships within the model. additionally, this step assesses the overall quality and accuracy of the model's predictive capabilities. we adhered to the established procedures recommended by prior researchers for examining direct relationships and conditional mediation (e.g., hair et al., 2014; preacher et al., 2007). during bootstrapping, we employed the widely accepted approach of using 5,000 subsamples. in earlier versions of the smart-pls software, researchers had to perform multiple steps and run different models to investigate moderated mediation analyses. however, in the latest iteration, smart-pls 4.0, this analysis can be conducted within a single model. figure 2 presents a visual representation of the results derived from the path model, including beta coefficients and their significance. figure 2. hypothesized relationships with coefficients and their significance in terms of assessing the quality of the model, we observed that the r-squared (r2) and adjusted rsquared (adjusted r2) values for retirement planning intention were 39% and 38%, ahamed & limbu 85 respectively. these values reflect the predictive power of our model. furthermore, we employed t statistics and p values to determine the significance of the direct path coefficients, as outlined in table 2. table 2. moderated mediation analysis results showing coefficients and their significance for hypothesized relationships relationships β direct effects financial risk tolerance -> retirement planning attitude -0.28*** financial risk tolerance -> retirement planning intention -0.26*** financial self-efficacy -> retirement planning attitude 0.21*** financial self-efficacy -> retirement planning intention -0.05 retirement planning attitude -> retirement planning intention 0.58*** money availability -> retirement planning intention 0.40** specific indirect effect (mediation) (h1 and h2) risk tolerance -> retirement planning attitude -> retirement planning intention -0.16*** financial self-efficacy -> retirement planning attitude -> retirement planning intention 0.12*** conditional direct effect (moderating effect) retirement planning attitude -> retirement planning intention conditional on money availability at +1 sd 0.09 retirement planning attitude -> retirement planning intention conditional on money availability at mean 0.24*** retirement planning attitude -> retirement planning intention conditional on money availability at -1 sd 0.39*** conditional indirect effects (moderated mediation) (h3 and h4) risk tolerance -> retirement planning attitude -> retirement planning intention conditional on money availability at +1 sd -0.02 risk tolerance -> retirement planning attitude -> retirement planning intention conditional on money availability at mean -0.07*** risk tolerance -> retirement planning attitude -> retirement planning intention conditional on money availability at -1 sd -0.11*** financial self-efficacy -> retirement planning attitude -> retirement planning intention conditional on money availability at +1 sd 0.02 financial self-efficacy -> retirement planning attitude -> retirement planning intention conditional on money availability at mean 0.05** financial self-efficacy -> retirement planning attitude -> retirement planning intention conditional on money availability at -1 sd 0.08*** note: * p < .05, ** p < .01, and *** p < .001. sd = standard deviation financial services review, 32(2) 86 table 2 displays the mediating effects of retirement planning attitude. the results indicate that retirement planning attitude acts as a negative mediator in the relationship between risk tolerance and retirement planning intention (β = 0.16, p = 0.000). this outcome contradicts our initial hypothesis 1, which posited a positive effect in this regard. however, it's worth noting that retirement planning attitude serves as a significant and positive mediator in the link between financial self-efficacy and retirement planning intention, thereby lending support to hypothesis 2 (β = 0.12, p value = 0.001). in hypothesis 3, we posited that the influence of financial risk tolerance on retirement planning intention, mediated by retirement planning attitude, is moderated by money availability. to test this, we employed a three-step analytical approach. first, we examined the mediating role of retirement planning attitude in the link between financial risk tolerance and retirement planning intention, as established in hypothesis 1. subsequently, we assessed the moderating impact of money availability. finally, we analyzed the conditional indirect effect (i.e., moderated mediation). the analysis confirmed the mediating role of retirement planning attitude (outlined in hypothesis 1). moreover, the findings demonstrate that money availability negatively moderates the connection between retirement planning attitude and intention, indicated by a beta coefficient of -0.09 and a p-value of 0.001, as depicted in figure 2. to delve deeper into this moderation, we conducted a simple slopes analysis using smartpls. this analysis highlights how varying levels of money availability (below, above, and at the mean standard deviation) affect the relationship between retirement planning attitude and intention. the results of the analysis are shown in figure 3. figure 3. moderating effect of money availability on retirement planning attitude and retirement planning intention link ahamed & limbu 87 in the third step, as shown in table 2, we observed that at low (-1 sd) and mean levels of money availability, there was a moderation of the mediating effect of retirement planning attitude on the relationship between risk tolerance and retirement planning intention, with beta values of -0.11 (p = 0.000) and -0.07 (p = 0.001), respectively. however, this moderating effect was not present at high levels of money availability (+1 sd). hence, hypothesis 3 was partially supported. these results highlight the significant role money availability plays in influencing the mediation by retirement planning attitudes between risk tolerance and retirement planning intentions. at lower levels of financial resources, this mediation is more highlighted due to the crucial role attitudes play amidst financial limitations. conversely, with high money availability, the need for such mediation diminishes, as the ability to act on retirement planning intentions faces fewer financial constraints. in hypothesis 4, we proposed that the availability of financial resources would moderate the influence of financial self-efficacy on retirement planning intention, with retirement planning attitude serving as a mediator. this hypothesis was examined using a three-step analytical process similar to that employed for hypothesis 3. firstly, the mediating role of retirement planning attitude in the relationship between financial self-efficacy and retirement planning intention was confirmed in hypothesis 2. the moderating impact of money availability on the link between retirement planning attitude and intention was then illustrated in figures 2 and 3. for the third step, focusing on the moderated mediation effect, table 2 reveals how money availability influences this relationship at different levels. at low and mean levels of money availability, there was a significant moderating effect on the indirect relationship between financial self-efficacy and retirement planning intention via retirement planning attitude, with beta coefficients of 0.08 (p = 0.001) and 0.05 (p = 0.003), respectively. however, at high levels of money availability, the moderating influence of financial resources was not evident in this indirect relationship (β = 0.02, p = 0.287). these findings lead to a partial validation of hypothesis 4. this result signifies that that the impact of financial self-efficacy on retirement planning intention is heightened by the level of money availability. when resources are low or average, the confidence and positive attitude generated by financial self-efficacy play a larger role in shaping retirement planning intentions. in contrast, the capacity to save and plan for retirement is less dependent on an individual's belief in their financial capabilities, as the financial means to save and invest are readily available. as a result, even if an individual has high financial self-efficacy, it does not significantly alter the likelihood of retirement planning as the financial capability to do so is already present. conclusion and implications in this study, grounded in cognitive appraisal theory, we explored the influence of financial risk tolerance and self-efficacy on retirement planning intention, mediated by retirement planning attitude. our investigation highlights that risk tolerance and financial self-efficacy are complex factors that might be influenced by sociodemographic characteristics and personal circumstances (samsuri et al., 2019). a significant aspect of our research focused on examining how money availability moderates the effects of financial risk tolerance and selfefficacy on retirement planning intention via retirement planning attitude. our findings underscore the notion that the availability of financial resources markedly affects the mediation process of retirement planning attitudes in the relationship between cognitive beliefs and retirement planning intentions. this research emphasizes the complex connection between financial risk tolerance, retirement planning attitude, intention, and financial resources. it recognizes retirement planning attitude as a negative mediator between risk tolerance and retirement planning intention, shedding light on why individuals with high risk tolerance may have lower intentions to plan for retirement. furthermore, the study highlights that the influence of financial risk tolerance on retirement planning intention, mediated by retirement planning attitude, is contingent on the availability of money. attitude plays a substantial mediating role when money is limited or average, financial services review, 32(2) 88 but this impact lessens when financial resources are abundant. this insight is vital for developing financial education and planning programs targeted towards populations with different levels of financial resource availability. this paper contributes to the field of financial behavior by exploring the mediated processes leading to retirement planning intentions, a dimension often overlooked in previous studies, especially in the context of a developing nation. our moderated mediation model offers a nuanced understanding of retirement planning behavior. theoretically, we demonstrate how different cognitive beliefs distinctly impact retirement planning attitudes and intentions, with risk tolerance having a negative effect and financial self-efficacy a positive one. the study highlights the applicability of cognitive appraisal theory in explaining retirement planning intentions. practically, this research has significant implications for practice and policy, especially for countries like bangladesh. despite the introduction of the universal pension scheme 2023 by the bangladesh government, its uptake has been limited (the business standard, 2023). this paper suggests that the varying levels of resource availability among individuals might influence their response to such schemes. it emphasizes the need for tailored communication strategies to foster positive retirement planning attitude, a crucial factor in retirement planning intention. as the role of perceived information transparency in fostering financial self-efficacy, particularly pertinent in emerging economies, is also paramount (zia-ur-rehman et al., 2021), these insights can be beneficial for other developing countries implementing similar retirement policies. however, our study is not without limitations. the data were collected in a developing country context, and thus, the results are not generalizable, underscoring a need for future research in diverse economic and cultural settings. our findings resonate with earlier observations of misinformation or lack of information about retirement benefits (gustman & steinmeier, 2001), suggesting further exploration into the relationship between financial risk tolerance, retirement planning attitude, and intention in different demographics. additionally, the relatively young and financially stable sample in our study may not fully represent the broader population, particularly in terms of retirement planning attitudes. future research should study more varied populations to enhance the generalizability of the findings. references ahamed, a. j., & limbu, y. b. (2018). dimensions of materialism and credit card usage: an application and extension of the theory of planned behavior in bangladesh. journal of financial services marketing, 23(3), 200-209. https://doi.org/10.1057/s41264-0180058-5 ajzen, i. (2011). the theory of planned behaviour: reactions and reflections. psychology & health, 26 (9), 1113-1127. https://doi.org/10.1080/08870446.2011. 613995 badgaiyan, a. j., & verma, a. (2015). does urge to buy impulsively differ from impulsive buying behaviour? assessing the impact of situational factors. journal of retailing and consumer services, 22, 145-157. https://doi.org/10.1016/j.jretconser.2014 .10.002 bandura, a. (1982). self-efficacy mechanism in human agency. american psychologist, 37(2), 122-147. https://psycnet.apa.org/doi/10.1037/000 3-066x.37.2.122 bandura, a. (2006). guide for constructing selfefficacy scales. self-efficacy beliefs of adolescents, 5(1), 307-337. bandura, a. (2012). cultivate self‐efficacy for personal and organizational effectiveness. handbook of principles of organizational behavior: indispensable knowledge for evidence‐based management, 179-200. https://doi.org/10.1002/9781119206422. ch10 bapat, d. (2020). antecedents to responsible financial management behavior among https://doi.org/10.1057/s41264-018-0058-5 https://doi.org/10.1057/s41264-018-0058-5 https://doi.org/10.1080/08870446.2011.613995 https://doi.org/10.1080/08870446.2011.613995 https://doi.org/10.1016/j.jretconser.2014.10.002 https://doi.org/10.1016/j.jretconser.2014.10.002 https://psycnet.apa.org/doi/10.1037/0003-066x.37.2.122 https://psycnet.apa.org/doi/10.1037/0003-066x.37.2.122 https://doi.org/10.1002/9781119206422.ch10 https://doi.org/10.1002/9781119206422.ch10 ahamed & limbu 89 young adults: moderating role of financial risk tolerance. international journal of bank marketing, 38(5), 11771194. https://doi.org/10.1108/ijbm-102019-0356 bayar, y., sezgin, h. f., öztürk, ö. f., & şaşmaz, m. ü. (2020). financial literacy and financial risk tolerance of individual investors: multinomial logistic regression approach. sage open, 10(3), 2158244020945717. https://doi.org/10.1177/2158244020945 717 białowolski, p., cwynar, a., cwynar, w., & węziak‐białowolska, d. (2020). consumer debt attitudes: the role of gender, debt knowledge and skills. international journal of consumer studies, 44(3), 191-205. https://doi.org/10.1111/ijcs.12558 chen, e., & matthews, k. a. (2003). development of the cognitive appraisal and understanding of social events (cause) videos. health psychology, 22(1), 106-110. https://psycnet.apa.org/doi/10.1037/027 8-6133.22.1.106 croy, g., gerrans, p., & speelman, c. (2010). the role and relevance of domain knowledge, perceptions of planning importance, and risk tolerance in predicting savings intentions. journal of economic psychology, 31(6), 860-871. https://doi.org/10.1016/j.joep.2010.06.0 02 deaves, r., veit, e. t., bhandari, g., & cheney, j. (2007). the savings and investment decisions of planners: a cross-sectional study of college employees. financial services review, 16(2), 117-133. fan, l. (2022). consumer financial information processing: an integrated approach to examine individual differences. international journal of consumer studies, 47(2), 703-722. https://doi.org/10.1111/ijcs.12859 farrell, l., fry, t. r., & risse, l. (2016). the significance of financial self-efficacy in explaining women’s personal finance behaviour. journal of economic psychology, 54, 85-99. https://doi.org/10.1016/j.joep.2015.07.0 01 field, a. (2013). discovering statistics using ibm spss statistics. sage. foroughi, a., buang, n. a., & sadeghi, r. h. m. (2012). exploring the influence of situational factors (money & time available) on impulse buying behaviour among different ethics. international journal of fundamental psychology & social sciences, 2(2), 41-44. furrebøe, e. f., & nyhus, e. k. (2022). financial self‐,efficacy, financial literacy, and gender: a review. journal of consumer affairs, 56(2), 743-765. https://doi.org/10.1111/joca.12436 garman, e. t., & forgue, r. (2014). personal finance. cengage learning. goyal, k., kumar, s., xiao, j. j., & colombage, s. (2022). the psychological antecedents of personal financial management behavior: a meta-analysis. international journal of bank marketing, 40(7), 14131451. https://doi.org/10.1108/ijbm-022022-0088 grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625-630. https://doi.org/10.1023/a:10229943149 82 grable, j. e. (2016). financial risk tolerance. handbook of consumer finance research, 19-31. https://doi.org/10.1007/978-3-31928887-1_2 gruijters, r. j. (2017). intergenerational contact in chinese families: structural and cultural explanations. journal of marriage and family, 79(3), 758-768. https://doi.org/10.1111/jomf.12390 https://doi.org/10.1108/ijbm-10-2019-0356 https://doi.org/10.1108/ijbm-10-2019-0356 https://doi.org/10.1177/2158244020945717 https://doi.org/10.1177/2158244020945717 https://doi.org/10.1111/ijcs.12558 https://psycnet.apa.org/doi/10.1037/0278-6133.22.1.106 https://psycnet.apa.org/doi/10.1037/0278-6133.22.1.106 https://doi.org/10.1016/j.joep.2010.06.002 https://doi.org/10.1016/j.joep.2010.06.002 https://doi.org/10.1111/ijcs.12859 https://doi.org/10.1016/j.joep.2015.07.001 https://doi.org/10.1016/j.joep.2015.07.001 https://doi.org/10.1111/joca.12436 https://doi.org/10.1108/ijbm-02-2022-0088 https://doi.org/10.1108/ijbm-02-2022-0088 https://doi.org/10.1023/a:1022994314982 https://doi.org/10.1023/a:1022994314982 https://doi.org/10.1007/978-3-319-28887-1_2 https://doi.org/10.1007/978-3-319-28887-1_2 https://doi.org/10.1111/jomf.12390 financial services review, 32(2) 90 gustman, a. l., & steinmeier, t. l. (2001). imperfect knowledge, retirement and saving. in: national bureau of economic research cambridge, mass., usa. hair jr, j. f., sarstedt, m., hopkins, l., & kuppelwieser, v. g. (2014). partial least squares structural equation modeling (pls-sem): an emerging tool in business research. european business review, 26(2), 106-121. https://doi.org/10.1108/ebr-10-20130128 harahap, s., thoyib, a., sumiati, s., & djazuli, a. (2022). the impact of financial literacy on retirement planning with serial mediation of financial risk tolerance and saving behavior: evidence of medium entrepreneurs in indonesia. international journal of financial studies, 10(3), 66-81. https://doi.org/10.3390/ijfs10030066 herrador-alcaide, t. c., hernández-solís, m., & topa, g. (2021). a model for personal financial planning towards retirement. journal of business economics and management, 22(2), 482-502. https://doi.org/10.3846/jbem.2020.1397 8 hibbert, a. m., lawrence, e. r., & prakash, a. j. (2012). the role of financial education in the management of retirement savings. journal of behavioral finance, 13(4), 299-307. https://doi.org/10.1080/15427560.2012. 735727 jacobs-lawson, j. m., & hershey, d. a. (2005). influence of future time perspective, financial knowledge, and financial risk tolerance on retirement saving behaviors. financial services review, 14(4), 331344. joo, s. h., & grable, j. e. (2004). an exploratory framework of the determinants of financial satisfaction. journal of family and economic issues, 25, 25-50. joo, s.-h., & grable, j. e. (2005). employee education and the likelihood of having a retirement savings program. journal of financial counseling and planning, 16(1), 37-49. https://ssrn.com/abstract=2255072 kimiyaghalam, f., mansori, s., safari, m., & yap, s. (2017). parents' influence on retirement planning in malaysia. family and consumer sciences research journal, 45(3), 315-325. https://doi.org/10.1111/fcsr.12203 kumar, s., shukla, g. p., & sharma, r. (2019). analysis of key barriers in retirement planning: an approach based on interpretive structural modeling. journal of modelling in management, 14(4), 972986. https://doi.org/10.1108/jm2-092018-0134 kung, c. s., zhu, j., zaninotto, p., & steptoe, a. (2023). changes in retirement plans in the english older population during the covid-19 pandemic: the roles of health factors and financial insecurity. european journal of ageing, 20(1), 1-12. https://doi.org/10.1007/s10433-02300770-1 lazarus, r. s., & alfert, e. (1964). shortcircuiting of threat by experimentally altering cognitive appraisal. the journal of abnormal and social psychology, 69(2), 195-205. https://psycnet.apa.org/doi/10.1037/h00 44635 lone, u. m., & bhat, s. a. (2022). impact of financial literacy on financial well-being: a mediational role of financial selfefficacy. journal of financial services marketing, 1-16. https://doi.org/10.1057/s41264-02200183-8 lown, j. m. (2011). development and validation of a financial self-efficacy scale. journal of financial counseling and planning, 22(2), 54-63. https://ssrn.com/abstract=2006665 lown, j. m., kim, j., gutter, m. s., & hunt, a.t. (2015). self-efficacy and savings among middle and low income https://doi.org/10.1108/ebr-10-2013-0128 https://doi.org/10.1108/ebr-10-2013-0128 https://doi.org/10.3390/ijfs10030066 https://doi.org/10.3846/jbem.2020.13978 https://doi.org/10.3846/jbem.2020.13978 https://doi.org/10.1080/15427560.2012.735727 https://doi.org/10.1080/15427560.2012.735727 https://ssrn.com/abstract=2255072 https://doi.org/10.1111/fcsr.12203 https://doi.org/10.1108/jm2-09-2018-0134 https://doi.org/10.1108/jm2-09-2018-0134 https://doi.org/10.1007/s10433-023-00770-1 https://doi.org/10.1007/s10433-023-00770-1 https://psycnet.apa.org/doi/10.1037/h0044635 https://psycnet.apa.org/doi/10.1037/h0044635 https://doi.org/10.1057/s41264-022-00183-8 https://doi.org/10.1057/s41264-022-00183-8 https://ssrn.com/abstract=2006665 ahamed & limbu 91 households. journal of family and economic issues, 36, 491-502. https://doi.org/10.1007/s10834-0149419-y lusardi, a., & mitchell, o. s. (2011). financial literacy and retirement planning in the united states. journal of pension economics & finance, 10(4), 509-525. https://doi.org/10.1017/s147474721100 045x macfarland, d. m., marconi, c. d., & utkus, s. p. (2004). money attitudes and retirement plan design: one size does not fit all. pension design and structure: new lessons from behavioral finance, 97-120. mathew, v., pk, s. k., & sanjeev, m. (2022). financial well-being and its psychological determinants—an emerging country perspective. fiib business review, 23197145221121080. https://doi.org/10.1177/2319714522112 1080 mauldin, t. a., henager, r., bowen, c. f., & cheang, m. (2016). barriers and facilitators to saving behavior in low-to moderate-income households. journal of financial counseling and planning, 27(2), 231-251. https://psycnet.apa.org/doi/10.1891/105 2-3073.27.2.231 meir, a., mugerman, y., & sade, o. (2016). financial literacy and retirement planning: evidence from israel. israel economic review, 14(1), 75-95. https://ssrn.com/abstract=2889925 nguyen, l., gallery, g., & newton, c. (2019). the joint influence of financial risk perception and risk tolerance on individual investment decision‐making. accounting & finance, 59, 747-771. https://doi.org/10.1111/acfi.12295 noone, j. h., stephens, c., & alpass, f. (2010). the process of retirement planning scale (preps): development and validation. psychological assessment, 22(3), 520-531. https://psycnet.apa.org/doi/10.1037/a001 9512 park, h., & martin, w. (2022). effects of risk tolerance, financial literacy, and financial status on retirement planning. journal of financial services marketing, 27, 167176. https://doi.org/10.1057/s41264021-00123-y petkoska, j., & earl, j. k. (2009). understanding the influence of demographic and psychological variables on retirement planning. psychology and aging, 24(1), 245-251. https://psycnet.apa.org/doi/10.1037/a001 4096 preacher, k. j., rucker, d. d., & hayes, a. f. (2007). addressing moderated mediation hypotheses: theory, methods, and prescriptions. multivariate behavioral research, 42(1), 185-227. https://doi.org/10.1080/0027317070134 1316 ramalho, t. b., & forte, d. (2019). financial literacy in brazil–do knowledge and selfconfidence relate with behavior? rausp management journal, 54, 77-95. https://www.scielo.br/j/rmj/a/hvc4pk6f xgrrmn6zq38h4xg/?lang=en roberts, j. a., & jones, e. (2001). money attitudes, credit card use, and compulsive buying among american college students. journal of consumer affairs, 35(2), 213-240. https://doi.org/10.1111/j.17456606.2001.tb00111.x safari, k., njoka, c., & munkwa, m. g. (2021). financial literacy and personal retirement planning: a socioeconomic approach. journal of business and socioeconomic development, 1(2), 121-134. https://doi.org/10.1108/jbsed-04-20210052 samsuri, a., ismiyanti, f., & narsa, i. m. (2019). effects of risk tolerance and financial literacy to investment intentions. international journal of innovation, creativity and change, 10(9), 40-54. https://doi.org/10.1007/s10834-014-9419-y https://doi.org/10.1007/s10834-014-9419-y https://doi.org/10.1017/s147474721100045x https://doi.org/10.1017/s147474721100045x https://doi.org/10.1177/23197145221121080 https://doi.org/10.1177/23197145221121080 https://psycnet.apa.org/doi/10.1891/1052-3073.27.2.231 https://psycnet.apa.org/doi/10.1891/1052-3073.27.2.231 https://ssrn.com/abstract=2889925 https://doi.org/10.1111/acfi.12295 https://psycnet.apa.org/doi/10.1037/a0019512 https://psycnet.apa.org/doi/10.1037/a0019512 https://doi.org/10.1057/s41264-021-00123-y https://doi.org/10.1057/s41264-021-00123-y https://psycnet.apa.org/doi/10.1037/a0014096 https://psycnet.apa.org/doi/10.1037/a0014096 https://doi.org/10.1080/00273170701341316 https://doi.org/10.1080/00273170701341316 https://www.scielo.br/j/rmj/a/hvc4pk6fxgrrmn6zq38h4xg/?lang=en https://www.scielo.br/j/rmj/a/hvc4pk6fxgrrmn6zq38h4xg/?lang=en https://doi.org/10.1111/j.1745-6606.2001.tb00111.x https://doi.org/10.1111/j.1745-6606.2001.tb00111.x https://doi.org/10.1108/jbsed-04-2021-0052 https://doi.org/10.1108/jbsed-04-2021-0052 financial services review, 32(2) 92 https://www.ijicc.net/images/vol10iss9/1 0904_samsuri_2019_e_r.pdf setyawan hutabarat, a. s., & wijaya, c. (2020). analysis the effect of financial literacy on financial planning for retirement (case study lecturers and administrative staffs in universitas indonesia). international journal of management, 11(5), 741-750. https://ssrn.com/abstract=3631012 shi, j., mo, x., & sun, z. (2012). content validity index in scale development. zhong nan da xue xue bao. yi xue ban= journal of central south university. medical sciences, 37(2), 152-155. https://doi.org/10.3969/j.issn.16727347.2012.02.007 sorić, i., penezić, z., & burić, i. (2013). big five personality traits, cognitive appraisals and emotion regulation strategies as predictors of achievement emotions. psihologijske teme, 22(2), 325-349. https://hrcak.srce.hr/108516 taylor‐carter, m. a., cook, k., & weinberg, c. (1997). planning and expectations of the retirement experience. educational gerontology: an international quarterly, 23(3), 273-288. https://doi.org/10.1080/0360127970230 306 the business standard. (2023, october 22). why there isn't much enthusiasm over universal pension anymore. available at https://www.tbsnews.net/bangladesh/wh y-there-isnt-much-enthusiasm-overuniversal-pension-anymore-724694 . date accessed on december 21, 2023. van rooij, m. c., lusardi, a., & alessie, r. j. (2012). financial literacy, retirement planning and household wealth. the economic journal, 122(560), 449-478. https://doi.org/10.1111/j.14680297.2012.02501.x yih, j., uusberg, a., taxer, j. l., & gross, j. j. (2019). better together: a unified perspective on appraisal and emotion regulation. cognition and emotion, 33(1), 41-47. https://doi.org/10.1080/02699931.2018. 1504749 zia-ur-rehman, m., latif, k., mohsin, m., hussain, z., baig, s. a., & imtiaz, i. (2021). how perceived information transparency and psychological attitude impact on the financial well-being: mediating role of financial self-efficacy. business process management journal, 27(6),1836-1853. https://doi.org/10.1108/bpmj-12-20200530 https://www.ijicc.net/images/vol10iss9/10904_samsuri_2019_e_r.pdf https://www.ijicc.net/images/vol10iss9/10904_samsuri_2019_e_r.pdf https://ssrn.com/abstract=3631012 https://doi.org/10.3969/j.issn.1672-7347.2012.02.007 https://doi.org/10.3969/j.issn.1672-7347.2012.02.007 https://hrcak.srce.hr/108516 https://doi.org/10.1080/0360127970230306 https://doi.org/10.1080/0360127970230306 https://www.tbsnews.net/bangladesh/why-there-isnt-much-enthusiasm-over-universal-pension-anymore-724694 https://www.tbsnews.net/bangladesh/why-there-isnt-much-enthusiasm-over-universal-pension-anymore-724694 https://www.tbsnews.net/bangladesh/why-there-isnt-much-enthusiasm-over-universal-pension-anymore-724694 https://doi.org/10.1111/j.1468-0297.2012.02501.x https://doi.org/10.1111/j.1468-0297.2012.02501.x https://doi.org/10.1080/02699931.2018.1504749 https://doi.org/10.1080/02699931.2018.1504749 https://doi.org/10.1108/bpmj-12-2020-0530 https://doi.org/10.1108/bpmj-12-2020-0530 ahamed & limbu 93 appendix 1. reliability and validity of the constructs constructs cronbach's alpha composite reliability (rho_a) composite reliability (rho_c) average variance extracted (ave) retirement planning intention 0.77 0.78 0.85 0.59 retirement planning attitude 0.67 0.67 0.80 0.51 financial risk tolerance 0.80 0.80 0.91 0.83 financial selfefficacy 0.80 0.83 0.84 0.56 money availability 0.61 0.62 0.84 0.72 appendix 2. discriminant validity htmt matrix constructs retirement planning intention retirement planning attitude financial risk tolerance financial self-efficacy money availability retirement planning intention retirement planning attitude 0.63 financial risk tolerance 0.72 0.61 financial selfefficacy 0.19 0.40 0.18 money availability 0.38 0.65 0.54 0.75 appendix 3. discriminant validity fornell-larcker criterion constructs retirement planning intention retirement planning attitude financial risk tolerance financial self-efficacy money availability retirement planning intention 0.77 retirement planning attitude 0.46 0.71 financial risk tolerance -0.57 -0.45 0.91 financial selfefficacy 0.17 0.41 -0.23 0.75 money availability -0.26 -0.41 0.37 -0.56 0.85 finser_31-2_complete_issue from the editor dear esteemed readers, as i pen this letter for my final issue as the editor of the financial services review (volume 31 issues 2 & 3), i am reminded of the rich tapestry of research and knowledge we have woven together. this special double issue, marking the transition of editorship to dr. john grable, is a testament to the evolving field of personal financial planning. we begin with kristine l. beck, hsin-hui chiu, and inga timmerman’s exploration into the changing landscape of risk tolerance among young investors. their study, focusing on millennials and gen z, challenges the traditional tools used for risk assessment, suggesting a pivotal shift in how financial advisors approach risk with the next generation. zhikun liu, russell james iii, and qi sun, delve into how older american adults’ financial planning horizons are influenced by their self-perceived life expectancy. their work reveals the dynamic nature of financial planning, shaped by personal health, marital status, and wealth levels, offering crucial insights for financial planners. the intriguing world of cryptocurrency investment is examined by kyoung tae kim, sherman d. hanna, and sunwoo t. lee. their analysis of the 2018 national financial capability study investor survey reveals a complex relationship between investment literacy, overconfidence, and cryptocurrency investment, highlighting the nuances of modern investment behaviors. brent j. davis, david p. richardson, and jason s. seligman, present a compelling case study on the need and demand for financial education among graduate students. their findings on the correlation between self-assessed and measured financial literacy underscore the importance of accurate self-perception in financial education. in a collaborative effort, lua a. v. augustin and terrance k. martin, examine the impact of the timing and frequency of financial education on financial behaviors. our study, utilizing the 2015 national financial capability study, underscores the significance of early and frequent financial education in fostering positive financial behaviors. greg filbeck and xin zhao’s research offers a unique perspective on financial literacy during the pandemic. comparing pre-pandemic residential programs with virtual/hybrid programs, their study demonstrates the effectiveness of financial literacy education across different modalities, a crucial insight in these changing times. christine mcclatchey’s analysis of the home borrower’s dilemma – whether to refinance, pay additional principal, or recast – is particularly relevant in the current economic climate. 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) v–vi her discussion on the mechanics and benefits of recasting provides valuable guidance for homeowners navigating financial decisions. lastly, isha chawla, mia b. russell, and kenneth j. white, explore consumer fear and trust in financial institutions during the covid-19 pandemic. their qualitative content analysis reveals key themes that are essential for understanding the relationship between consumers and financial institutions in turbulent times. as i transition to the role of an associate editor, i reflect on the honor of contributing to this prime journal in the field of personal financial planning. i am confident that under dr. grable’s leadership, the financial services review will continue to illuminate the path forward in our ever-evolving discipline thank you for your support and engagement over the years. warm regards, dr. terrance k. martin jr. editor (outgoing), financial services review 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. vi t. k. martin / financial services review 31 (2023) v–vi pii: 1057-0810(93)90001-7 volume 3 number 1 1!293/1!294 financial services review the editor lewis mandell university of connecticut mamging editor barbaraf’oole university of connecticut associate jmitms neil g. cohen george washington university mona j. gardner illinois wesleyan university benton e. gup university of akabama jean louis heck villunova university david s. kidwell university of minnesota neil b. murphy virginia commonwealth universiry phyllis s. myers virginia commonwealth university george c. philippatos university of tennessee chris j. prestopino california state university chico s. travis pritchett university of south carolina william reichenstein baylor university frank k. reilly university of notre dame arthur l. schwartz university of south florida peter l. struck washington mutual savings bank g. c. uselton texas a & m university thomas waschauer san diego state university walt woerheide rochester institute of technology @ i jai press inc. greenwich, connecticut london, england framing the annuity as bequest protection: an experimental test ying yana, russell n. james iiib,* adepartment of personal financial planning, eastern new mexico state university, 1500 s avenue k, portales, nm 88130, usa bdepartment of personal financial planning, texas tech university, box 41210, lubbock, tx 79409-1210, usa abstract bequest motives are commonly cited as a barrier for annuitization. this paper tests how framing partial annuitization as a protection for an intended bequest against the risk of asset exhaustion due to unexpected longevity influences the desire to purchase an annuity among a sample of 2,160 participants. the results indicate that this framing argument does increase interest in purchasing an annuity. regression results demonstrate that this framing has a larger positive effect for individuals with a greater bequest motive. the relationship between annuitization framing and bequest motive demonstrated by this experiment has important practical and theoretical implications. © 2021 academy of financial services. all rights reserved. jel classifications: d1; d14; d15 keywords: annuities; estate planning; bequest motive; framing effects 1. introduction longevity risk is a risk that a person lives longer than expected and has insufficient wealth to support his or her planned consumption expenditures. a life annuity is an insurance instrument that converts wealth into a lifetime income stream. it can insure people against longevity risk and can serve as a valuable part of retirees’ investment portfolios (lockwood, 2012). however, relatively few people annuitize any of their wealth *corresponding author: tel.: +1-806-787-5931; fax: 806-742-5033. e-mail address: russell.james@ttu.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 277–291 (peijnenburg et al., 2016). franco modigliani (1986) referenced this “annuity puzzle” in his nobel prize acceptance speech and indicated that there are substantial reasons for people to annuitize at least part of their wealth. bequest motives are a common explanation for this annuity puzzle. annuity benefits will end at the death of the annuitants; therefore, they are not bequeathable wealth. economists and researchers indicate that this non-bequeathable characteristic may potentially reduce or even eliminate the desire for annuitization (bernheim, 1991; bütler & teppa, 2007; yaari, 1965). bequest motives, or proxies for such, have been measured in a variety of ways such as the presence of children, bequest expectations, or the self-reported importance of leaving a bequest. typically, researchers do find a negative association between a bequest motive and purchasing a life annuity (bernheim, 1991; bütler & teppa, 2007; yaari, 1965). this paper examines the effects of framing an annuity as bequest protection on intentions to purchase an annuity and how these effects are impacted by the bequest motive. the estimations of regression models show that framing the annuity as an inheritance protection did increase interest in using an annuity. further, it did so to a much stronger degree for people who first reported having a relatively higher bequest motive. this was true both when measured as the change in estimated numerical probability of using an annuity (multiple linear regression model) and when measuring the binary outcome of whether or not interest in an annuity increased (probit regression model). the increase in the interest in annuitization grew more strongly for those with a higher bequest motive, even when controlling for other factors such as the personal longevity estimations, education, income, age, and gender. a higher bequest motive leads to a greater impact from bequest protection framing of an annuity. 2. literature review 2.1. the annuity puzzle people face significant longevity risk in the united states (lockwood, 2012) in part due to generally increasing life expectancy (knell, 2018). according to data from the world bank, the average u.s. life expectancy for males increased from 46.3 to 79 between 1900 and 2018 and the average life expectancy for females increased from 48.3 to 84 between 1900 and 2018 (the world bank, 2019). about one-fifth of 65-years-olds will live to age 90 and beyond (lockwood, 2018). therefore, retirees should be increasingly concerned about the need to hedge longevity risk. the economic literature provides theoretical and empirical evidence that life annuities bring substantial welfare improvements to retirees (cox & lin, 2007; scott, 2008). an immediate annuity can bring a lifetime income stream for retirees and help them maximize retirement spending (scott, 2008). moreover, scott (2008) suggests partial annuitization should be an optimal strategy for typical retirees and recommends that retirees convert 1015% of wealth to a life annuity in the retirement period. however, few retirees annuitize any of their wealth. among people who are eligible for defined benefit pension plans, a form of an annuity, when the plans offer a lump-sum option, about 50% to 75% of these benefits are taken as a lump-sum (banerjee, 2013). this is true 278 y. yan and r. n. james / financial services review 29 (2021) 277–291 even if the lifetime benefit is the default option and retirees have to spend great time on complex paperwork to get the lump-sum distribution (banerjee, 2013; benartzi, previtero, & thaler, 2011; mottola & utkus, 2008). williams and james (2019) found that only 4% of retirees were receiving income from a commercial annuity other than traditional pension plans. thus, the propensity to annuitize is low in the united states (peijnenburg et al., 2016). unwillingness to annuitize is not a unique behavior for americans. for example, an analysis using japanese data also reports similarly low propensity to voluntarily annuitize wealth (purcal & piggott, 2008). this widespread resistance to annuitization is known as the “annuitization puzzle” (modigliani, 1986). in other words, annuities are underutilized. 2.2. bequest motives researchers have used various explanations for this annuitization puzzle. the low demand could result from the public pension system. retirees may have sufficiently hedged the longevity risk with annuities provided by the public pension system, such as social security (brown, 2001, 2003). others argue that the annuity market is not actuarially fair (brown & poterba, 2000; donnelly, 2015; lockwood, 2012; pecchenion & pollard, 1997), and that rates are unattractive due to the issue of adverse selection (abel, 1985; finkelstein & poterba, 2004; mitchell & mccarthy, 2002). still others argue that individuals self-insure using other assets or resources (laferrère, 2012; vidal-meliá & lejárraga-garcı́a, 2006). however, the most common explanation for the annuitization puzzle is the bequest motive. several authors conclude that the bequest motive could be a barrier that might reduce the demand for a life annuity (bernheim, 1991; friedman & warshawsky, 1990; lockwood, 2012, 2018; purcal & piggott, 2008; yaari, 1965). this is because annuitized wealth is not bequeathable (lockwood, 2018). annuity benefits end at the death of the annuitant. 2.2.1. the economics of annuities as bequest protection as yaari (1965) pointed out, although a bequest motive might prevent annuitization of all assets, it should not prevent annuitization of any assets, that is, partial annuitization. longevity risk can result in all assets being completely exhausted due to living expenses incurred during an exceptionally long life. this would result in the elimination of any intended bequest. annuitization can protect against this risk. as davidoff et al. (2005, 1589) explain, “partial annuitization can reduce the variation in the bequest.” thus, partial annuitization should be attractive to people who have a bequest motive (davidoff et al., 2005). those with a bequest motive should hold a portfolio which combines annuity wealth and bequeathable wealth to maximize the marginal utility of both bequests and consumption (yaari, 1965). although theoretically sound, this concept may be sufficiently complex such as to require explanation for a lay audience. financial advisors may be able to bridge this complexity gap between economic theory and practical advice by framing the annuity as a way to protect the heirs’ inheritance. y. yan and r. n. james / financial services review 29 (2021) 277–291 279 2.2.2. framing in behavioral finance frame dependence references circumstances in which people tend to make different choices as the result of changes in phrasing or comparison salience, even though the relevant objective facts remain the same (baker & nofsinger, 2010). in behavioral finance, examples of such effects include gain/loss framing (kahneman, knetsch, & thaler, 1991), narrow/broad framing (barberis, huang, & thaler, 2006), mental accounting (thaler, 1985), and many others (baker & nofsinger, 2010). past experiments have shown that framing can influence decisions in retirement planning in general (choi et al., 2004) and annuity decisions in particular (agnew et al., 2008; salisbury & nenkov, 2016). for example, salisbury and nenkov (2016) found that people were more interested in purchasing annuities described with life words than identical annuities described with “death” words. agnew et al. (2008) found that annuities became more attractive when alternative investments were framed negatively. after reviewing evidence of the annuity puzzle, brown (2007) recommends that future annuity research should focus more on such behavioral explanations. 2.2.3. the psychology of annuities as bequest protection another argument for the potential power of bequest protection framing for annuities comes from past research connecting annuities and personal mortality salience. terror management theory, as well as a parallel utility maximization economic model (james, 2016), predict that personal mortality salience (such as that triggered by death reminders) will generate two responses: avoidance and pursuit of lasting social impact (a.k.a., symbolic immortality; pyszczynski, greenberg & solomon, 1999). salisbury and nenkov (2016) identified that annuities, which can be thought of as a bet on one’s own mortality timing, do indeed serve as a death reminder that triggers mortality salience. the avoidance response will tend to reduce interest in annuities, simply because they are a death reminder. fitting with this response, salisbury and nenkov (2016) found that people were more interested in purchasing annuities described as paying “if the annuity holder lives up to different ages” (life framing) than in purchasing annuities described as paying “depending on the age when the annuity holder dies” (death framing; salisbury & nenkov, 2016, p. 423). however, the resistance to annuities goes beyond simple avoidance of a death reminder. the second response to mortality reminders predicted by theory is pursuit of lasting social impact (a.k.a., “symbolic immortality”). this is the idea that although a person will die, some part of them—their name, family, community, values, and so forth—will live on. a bequest motive can be the financial expression of this psychological desire. thus, annuities serve as a death reminder, increasing the motivation to leave a bequest, but simultaneously offer a financial benefit that disappears at death. confirming this idea, williams and james (2019) found that higher levels of personal mortality reminders generate a greater preference for an annuity paying lower income but with a bequest provision—a form of life insurance. such product combinations (life annuity plus survivor bequest benefits) “solve” the desire for bequests. and indeed, lockwood (2012) estimates that about three-fourths of commercial annuities owned by recent retirees have some provision that passes money to heirs after death. however, these product combinations may not be optimal for financial outcomes. the need to protect against longer than expected life span (annuity benefits) need not be 280 y. yan and r. n. james / financial services review 29 (2021) 277–291 accompanied by a need to protect against shorter than expected life span (life insurance benefits). the current experiment tests an alternative approach. rather than coupling the annuity with financial products offering actual bequest benefits, partial annuitization is simply verbally framed as a form of bequest protection. if effective, this approach would provide the opportunity to communicate the benefits of a life annuity in a way that matches underlying psychological mechanisms, but without the need to combine it with other, potentially unnecessary, financial products. 2.3. hypotheses hypothesis 1: framing an annuity as a protection for bequest goals will increase individuals’ interest in using an annuity in retirement. hypothesis 1a: framing an annuity as a protection for bequest goals will increases individuals’ interest in using an annuity in retirement to a relatively greater extent for those who were relatively more focused on leaving wealth to heirs (i.e., who had expressed a stronger bequest motive). hypothesis 1 predicts that framing partial annuitization as protecting an inheritance for heirs will increase interest in using an annuity in retirement. hypothesis 1a predicts that the effect of this framing will be more powerful on those who have expressed a relatively greater bequest motive. those who have expressed a relatively greater focus on leaving wealth to heirs should logically be particularly influenced by this framing. finding this to be true would also lend support to the idea that increased interest resulting from framing the annuity as bequest protection was arising through the proposed mechanisms, rather than, for example, simply through generalized experimenter demand (zizzo, 2010). in other words, the effect of the framing would not be due solely to respondents changing their response to match what they believe to be the answer desired by the experimenters, but would actually relate to the respondent’s underlying bequest motive. 3. data and results 3.1. sample participants were recruited from adults in the united states via amazon’s mechanical turk (mturk) platform with the headline description “university survey of opinions on retirement planning.” clicking on this generated the following description, “survey of retirement planning opinions. we are conducting an academic survey about opinions on retirement planning options. this takes around 10 minutes (includes minimum timed responses and written text requirements) and it is intended to advance research about people and their retirement planning, so please make sure you can commit the time. at the end of the survey, you will receive a unique ‘completion code’ to paste into the box below to receive credit for taking our survey.” there were 2,206 respondents who participated in the survey on october 18, 2019 with 2,160 responding to all questions used in the following analyses. the study was approved by y. yan and r. n. james / financial services review 29 (2021) 277–291 281 t ab le 1 k ey v ar ia b le s q u es ti o n te x t a n n u it y ch an ce (p re an d p o st ) im ag in e th at y o u ar e 6 5 y ea rs o ld an d b eg in n in g re ti re m en t. y o u h av e so m e re ti re m en t sa v in g s th ro u g h y o u r em p lo y er re ti re m en t p la n an d ar e d ec id in g h o w to m an ag e u si n g th at m o n ey in th e co m in g y ea rs . y o u h av e th e o p ti o n to p u t y o u r re ti re m en t sa v in g s in to an an n u it y th at w il l g iv e y o u m o n th ly p ay m en ts ea ch y ea r y o u li v e. w h en y o u p u t y o u r sa v in g s in to an an n u it y , y o u p ay a lu m p su m o f m o n ey u p fr o n t. in re tu rn fo r th at lu m p -s u m in v es tm en t, y o u re ce iv e a se ri es o f re g u la r m o n th ly p ay m en ts ea ch y ea r y o u li v e. p le as e ra te th e p er ce n ta g e li k el ih o o d th at y o u m ig h t co n si d er p u rc h as in g an an n u it y as p ar t o f y o u r re ti re m en t p la n n in g . l ik el ih o o d o f p u rc h as in g an an n u it y : [0 – 1 0 0 sl id er b ar ] a g re e fr am in g ar g u m en t (y es = 1 ) o n e b en efi t to an n u it ie s is th at th ey ca n p ro te ct an in h er it an ce fo r h ei rs . n o m at te r h o w lo n g y o u li v e, y o u w il l h av e li fe ti m e in co m e, so y o u w o n ’t b e fo rc ed to sp en d aw ay y o u r o th er in v es tm en ts o r as se ts ju st to h av e re g u la r in co m e. d o es it m ak e se n se to y o u h o w h av in g g u ar an te ed li fe ti m e in co m e fr o m an an n u it y ca n p ro v id e p ro te ct io n o f an in h er it an ce fo r y o u r h ei rs ? _ _ y es , th at m ak es se n se _ _ n o , th at d o es n ’t m ak e se n se _ _ i’ m n o t re al ly su re b eq u es t m o ti v e p re ce d in g q u es ti o n : s u p p o se y o u w er e ag e 6 5 an d h ad sa v ed $ 1 m il li o n to u se d u ri n g re ti re m en t in ad d it io n to so ci al se cu ri ty . y o u r in v es tm en ts w il l re tu rn an in fl at io n -a d ju st ed am o u n t o f 5 % (e .g ., $ 5 0 ,0 0 0 fr o m $ 1 m il li o n ) p er y ea r. w h ic h o f th e fo ll o w in g sp en d in g p la n s w o u ld y o u p re fe r? — (p ag e b re ak ) — o n a sc al e fr o m 0 to 1 0 0 , h o w m u ch d id y o u co n si d er y o u r d es ir e to le av e m o n ey to y o u r h ei rs w h en m ak in g th e sp en d in g ch o ic e o n th e la st q u es ti o n ? c o n si d er ed le av in g an in h er it an ce : [0 – 1 0 0 sl id er b ar ] m o rt al it y sa li en ce t o w h at ex te n t h av e y o u b ee n th in k in g ab o u t d ea th in th e p as t se v er al m in u te s? n ev er [1 ] v er y r ar el y [2 ] r ar el y [3 ] o cc as io n al ly [4 ] f re q u en tl y [5 ] v er y f re q u en tl y [6 ] p le as e ra te y o u r le v el o f ag re em en t w it h th e fo ll o w in g st at em en t: t h e ta sk s in th is su rv ey re m in d ed m e o f d ea th . v er y s tr o n g ly a g re e [6 ] s tr o n g ly a g re e [5 ] a g re e [4 ] d is ag re e [3 ] s tr o n g ly d is ag re e [2 ] v er y s tr o n g ly d is ag re e [1 ] t o w h at ex te n t d id th e ta sk s in th is su rv ey ev o k e th o u g h ts o f d ea th ? n ev er [1 ] v er y r ar el y [2 ] r ar el y [3 ] o cc as io n al ly [4 ] f re q u en tl y [5 ] v er y f re q u en tl y [6 ] c h an ce li v e to 1 0 0 o n a sc al e fr o m 0 % to 1 0 0 % , w h at ar e th e ch an ce s th at y o u w il l li v e to ag e 1 0 0 o r o ld er ? p ro b ab il it y o f li v in g to 1 0 0 : [0 – 1 0 0 sl id er b ar ] 282 y. yan and r. n. james / financial services review 29 (2021) 277–291 the institutional review board for human research protection program of the second author’s affiliated institution. 3.2. variables table 1 reports the question text used to generate the key variables used in the following analyses. reports of education level were converted to number of years of education as follows, nine for “less than high school diploma,” 12 for “high school diploma,” 14 for “some college or associates’ degree,” 16 for “bachelor’s degree,” and 19 for “graduate degree.” categorical reports of income were converted to $10,000 (for less than $10,000 category), category midpoints of $15,000, $25,000, $35,000, $45,000, $55,000, $65,000, $75,000, $85,000, $95,000, and $125,000, and finally $150,000 for greater than $150,000. in the following regressions, socio-demographic control variables were also included. more years of education may be associated with greater financial literacy (van rooij, lusardi, & alessie, 2011). this might increase the likelihood of understanding the bequest protection argument, but also increase the likelihood of having already considered the argument before the framing intervention. alternatively, high education may be associated with disagreement with the framing claim resulting from acceptance of an alternative investment strategy, such as replacing annuities with mutual fund investments and protecting an inheritance by accumulating sufficient wealth to provide for any income needs (pang & warshawsky, 2010). income is also associated with financial sophistication and may impact the acceptance of the framing argument (van rooij, lusardi, & alessie, 2011). the longer one expects to live, the more relevant the bequest protection argument becomes. a standard annuity involves a bet on one’s longevity. thus, a simple measurement of subjectively anticipated longevity—the chance one will live to age 100—is included. age is included as it has been associated with both financial literacy and the confidence in one’s beliefs about financial products (such as annuities) and may impact the influence of a framing argument (finke, howe, & huston, 2017). gender is table 2 descriptive statistics variable mean (sd) min max age 44.973 (16.061) 19 85 male 0.509 (0.500) 0 1 income 54.912 (34.108) 10 150 education years 15.525 (1.995) 9 19 annuity chance (pre-framing) 49.635 (27.179) 0 100 annuity chance (post-framing) 53.992 (27.685) 0 100 annuity interest increased 0.493 (0.500) 0 1 annuity interest decreased 0.231 (0.421) 0 1 agree framing argument 0.789 (0.408) 0 1 bequest motive 54.390 (34.571) 0 100 chance live to 100 31.010 (29.007) 0 100 mortality salience 11.460 (3.525) 3 18 n 2,160 y. yan and r. n. james / financial services review 29 (2021) 277–291 283 included as some research has found females to be more risk averse than males when they make financial decisions in general (jianakoplos & bernasek, 1998) and with regard to annuities in particular (agnew et al., 2008). as such, females may respond more favorably to the framing of an annuity as a hedge against a risk of total bequest depletion. table 2 reports sample means, standard deviations, minimums, and maximums. the average respondent age was 45 with 51% of respondents being male. average years of education was 15.5, and average income was $55,000 per year. 3.3. hypothesis 1: results table 1 indicates that the mean percentage likelihood of purchasing an annuity as part of retirement planning (annuity chance) increased following the bequest protection framing statement. a two-tailed paired t test finds that this difference is statistically significant at p < .0001. additionally, the percentage likelihood of purchasing an annuity as part of retirement planning increased following the bequest protection framing statement for 49.3% of respondents but decreased for only 23.1% of respondents. both findings support hypothesis 1. these two outcomes measure the change in likelihood of using an annuity in two different ways. the pre versus post comparison of probability changes measures the magnitude of the change. this is important but is also subject to the effects of outliers where very large changes take place. however, using a dummy variable for any increase in probability is not subject to these outlier effects, as all positive changes are measured as a “1” regardless of magnitude. thus, it is useful to look at both forms of outcome measurements and this will be repeated in the following regression analyses. 3.4. hypothesis 1a: results table 3 reports results for an ordinary least squares regression where the dependent variable is the percentage likelihood of purchasing an annuity as part of retirement planning table 3 ols regression with dependent variable: annuity chance (post-framing) variable parameter estimate (se) p-value intercept 10.793 (2.8989) 0.0002 annuity chance (pre-framing) 1.0902 (0.0376) <0.0001 annuity chance (pre-framing) sq �0.0026 (0.0004) <0.0001 bequest motive 0.0560 (0.0090) <0.0001 mortality salience 0.1453 (0.0869) 0.0948 male �1.9764 (0.6082) 0.0012 income �0.0160 (0.0093) 0.0867 education years �0.2583 (0.1597) 0.1060 age �0.0432 (0.0192) 0.0241 chance live to 100 0.0155 (0.0109) 0.1536 note: n = 2,160; r2 = 0.7505. ols = ordinary least squares. 284 y. yan and r. n. james / financial services review 29 (2021) 277–291 following exposure to the bequest protection framing argument. the regression controls for the percentage likelihood reported before the bequest protection framing and reflects the change (pre vs. post) in this reported percentage likelihood. in addition, the regression controls for the square of the percentage likelihood reported before the bequest protection framing. this is included because the dependent variable, annuity chance (post-framing), is constrained to be between 0 and 100. as the initial percentage, a.k.a. annuity chance (preframing), increases, the possible magnitude of an increase resulting from the framing falls, suggesting a nonlinear relationship (e.g., if the likelihood of using an annuity was 99% before the framing, there is little opportunity for increasing this likelihood). the results confirm this nonlinear relationship as both the linear and squared term are highly significant. the increase resulting from the bequest protection framing argument is greater for those who had expressed a greater bequest motive at p < .0001. this result supports the second hypothesis (1a). an additional possible relationship is that the bequest protection framing might be more powerful for those experiencing greater mortality salience. terror management theory suggests that those experiencing greater mortality salience will have an increased interest in making a lasting social impact. bequest protection framing should appeal to this increased interest by showing the potential benefits of partial annuitization for heirs. contemplation of annuities has been found in past research to generate mortality salience (salisbury & nenkov, 2016), making this potential relationship particularly relevant for the current experiment. additionally, bequest motive was measured before the annuity questions while mortality salience was measured after the annuity questions. thus, the level of experienced mortality salience, controlling for bequest motive, would be sensitive to changes generated by the annuity questions themselves. table 3 shows a weak positive relationship between mortality salience and the increase in annuity interest following the bequest protection framing intervention. however, this relationship is significant only at p < .10. table 4 reports results from a probit analysis measuring whether or not the probability of purchasing an annuity at retirement increased following the bequest protection framing argument. unlike the analysis reported in table 3 this analysis is not affected by large outliers, such as where the reported percentage changed from 0 to 100 or vice-versa. nevertheless, the coefficient for bequest motive is still positive and significant at p < .0001. this confirms table 4 probit analysis with dependent variable: annuity chance increased parameter estimate (se) pr > v2 intercept 0.0348 (0.2497) 0.8892 bequest motive 0.0037 (0.0008) <0.0001 mortality salience 0.0109 (0.0079) 0.1654 male �0.1157 (0.0553) 0.0363 income �0.0008 (0.0009) 0.3208 education years �0.0012 (0.0145) 0.9316 age �0.0061 (0.0017) 0.0004 chance live to 100 0.0006 (0.001) 0.5650 note: n = 2,160. y. yan and r. n. james / financial services review 29 (2021) 277–291 285 that the relationship identified in table 3 is not driven only by large outliers but reflects a general tendency among the respondents. thus, the likelihood that interest in an annuity will increase to some extent following the bequest protection framing also increases as bequest motive increases. this result further supports the second hypothesis (1a). the direction and significance of the relationship with age and gender remains consistent with the results reported in table 3. however, the relationship with mortality salience becomes nonsignificant (p = .1654). the primary interest in this analysis was to explore the results from introducing a bequest protection framing argument on interest in purchasing an annuity as part of retirement planning. bequest protection framing may fail to have an effect because respondents disagree with the bequest protection argument. however, it may also fail to have an effect even if respondents agree with the bequest protection argument. this could result from respondents having already considered such a justification before its introduction in the framing argument or because other unmeasured factors make such a justification irrelevant for the respondent. thus, table 5 reports results from an additional measurement of the effectiveness of this framing argument: whether respondents agree that the argument itself makes sense. this measurement of the power of the bequest protection framing argument also shows a positive association with bequest motive (p < .0001). thus, those with a higher bequest motive are more likely to agree that the bequest protection framing argument makes sense. one difference as compared with the previous outcome measurement is that mortality salience is a highly significant predictor (p < .0001) of agreeing with the bequest protection framing argument. this may be because the increased desire for making a lasting impact resulting from mortality salience may increase acceptance of the argument describing such an impact. future investigation of this relationship may be critical as previous research shows contemplating annuity purchases increases mortality salience (salisbury & nenkov, 2016). thus, a framing argument that works particularly well in the presence of mortality salience may be particularly effective in a real-world application. however, in the present experiment this relationship may arise because acceptance and understanding of the bequest protection argument actually generates mortality salience. future experimental research may be able to disentangle this relationship more precisely. table 5 probit analysis with dependent variable: agree with framing argument parameter estimate (standard error) pr > v2 intercept 0.7593 (0.2819) 0.0071 bequest motive 0.0038 (0.0009) <0.0001 mortality salience 0.0355 (0.0090) <0.0001 male �0.0054 (0.0627) 0.9311 income �0.0001 (0.0009) 0.8822 education years �0.0262 (0.0163) 0.1085 age �0.0054 (0.0019) 0.0041 chance live to 100 0.0036 (0.0011) 0.0014 note: n = 2,160. 286 y. yan and r. n. james / financial services review 29 (2021) 277–291 4. discussion this research examines bequest motives and annuitization in a new way. it examines an intervention framing an annuity as a protection for inheritance goals from longevity risk. in the presence of other assets intended to be left as a bequest, annuitization can be viewed as protecting those assets against consumption due to a person’s excessive longevity. thus, in the absence of (partial) annuitization, these assets might be completely exhausted due to living expenses incurred during an exceptionally long life. this bequest protection frame presents the annuity as a downside protection for inheritance goals. we hypothesized that this framing intervention would encourage interest in using annuities in retirement. the results are consistent with this hypotheses that framing matters. on average, introducing the bequest protection framing intervention increases interest in using an annuity in retirement. additionally, regression models showed that a greater bequest motive leads to a more positive effect of the bequest protection framing intervention on the predicted likelihood of using an annuity in retirement. thus, the impact of the intervention is likely not due exclusively to generalized experimenter demand effects. this result matches economic arguments for partial annuitization as benefitting heirs by reducing bequest volatility but suggests a necessary role for the advisors to explain this justification. this bequest protection framing may represent an additional, practical approach to addressing the annuitization puzzle (bernheim, 1991; bütler & teppa, 2007; yaari, 1965). this preliminary evidence suggests a potential method to directly address previous findings of a negative association between bequest motives and annuitization (bernheim, 1991; bütler & teppa, 2007; yaari, 1965). without special framing, this negative association is obvious. annuity wealth is not bequeathable. therefore, people who have a bequest motive may be less likely to purchase life annuities. however, bequest protection framing appears to most strongly influence those who have higher bequest motives. developing a method, such as that tested here, to address these bequest motive objections may be particularly important because the act of annuity contemplation may actually increase bequest motives. past research has shown that annuity contemplation generates mortality salience (salisbury & nenkov, 2016) and also that mortality salience generates an increased desire for lasting social impact, a.k.a., pursuit of symbolic immortality (james, 2016; pyszczynski, greenberg, & solomon, 1999). this results in the outcome, demonstrated by williams and james (2019), that increasing mortality salience increases interest in annuities that pay less income but include bequest benefits relative to those that pay more income with no such bequest benefit. however, this outcome is unlikely to be an ideal financial planning choice. the current reframing approach offers the advantage that no product changes are needed to increase the acceptability of the annuity choice. of course, the present experiment measures responses to a hypothetical question about an annuity purchase decision to be made at age 65. it does not measure actual purchase behavior. behavioral intentions can differ from actual behavior. however, because this analysis focuses on the change in behavioral intentions, the results can be instructive even in the presence of imprecise measurements or additional barriers between intentions and action. y. yan and r. n. james / financial services review 29 (2021) 277–291 287 4.1. financial planning practice implications: client conversations economic theory demonstrates that partial annuitization may be ideal even for those individuals with a strong bequest motive. partial annuitization reduces the volatility of a bequest by providing protection against reduction or exhaustion of the bequest assets due to income needs resulting from longevity risk. the current results suggest that this argument, when presented in nontechnical terms, can be persuasive to the general public. however, the argument is not uniformly attractive. it is more compelling for some people than others. in particular, the argument is more compelling for those who begin with a higher bequest motive. additionally, this framing is more effective among women and those who are younger. determining bequest motive need not be a complicated task for the advisor. some clients may bring up such concerns when annuities are discussed. in this study, bequest motives were measured by the response to the question, “how much did you consider your desire to leave money to your heirs when making the [retirement] spending choice on the last question?” similarly simple questions may also elicit the relevant motivations from clients. where such bequest motives exist, the bequest protection framing for annuities may be particularly effective. 4.2. financial planning practice implications: client annuity usage the results presented here have two practical implications for client choice of annuity products. first, the results can be helpful where annuities are an appropriate financial option for clients. the results demonstrate the effectiveness (especially for certain people) of a bequest protection framing argument. second, the results can be helpful where the typical form of annuities (annuities with survivor benefits) are less than optimal. where williams and james (2019) showed that annuities producing less income and greater bequest benefits become more attractive with increased mortality salience, this research shows that it is not necessarily required to change the product to include more bequest benefits, but it can work to simply change the product framing to highlight the otherwise unrealized bequest benefits. 4.3. annuity advantages finding a cost-free way to encourage the use of annuities—without the addition of potentially unnecessary life-insurance-like bequest benefits—can be potentially beneficial for clients. the first reason is the risk. compared with investing in the stock market, the risk of investing in annuities is relatively low and the annuity lifetime income or certain period income is guaranteed. for example, the sequence of returns risk, which could have a significant impact on retirees’ portfolios, does not arise with annuities but does occur with investing in the stock market. a monte carlo simulation shows the impact from this sequence of returns risk. the simulation assumes a retiree with an initial investment of $500,000 dollars, average returns for 288 y. yan and r. n. james / financial services review 29 (2021) 277–291 stock market of 6% (annually), a standard deviation of return of 20%, a time horizon for investment of 20 years, and the amount added to the investment portfolio of $15,000 at the end of each year. in 10,000 iterations, the ending portfolio (mean $2,192,524; median $1,633,929; standard deviation $1,966,518) varies widely from the fifth percentile outcome of $465,544 to the 95th percentile outcome of $5,722,992. if the investors suffers a “bad time” at the beginning of the investment period, that is, negative returns at the beginning of the investment period (e.g., caused by an economic crisis or other global issues), the investors might suffer a huge initial loss with regard to the investment portfolio. if retirees have such a loss in their investments, their assets might be completely exhausted, and they may not have sufficient wealth for their retirement. however, the returns for annuity investments are stable and the annuity income is typically guaranteed. for example, in the above simulation, placing 10% ($50,000) of the initial investment into a fixed twenty-year guaranteed annuity with a lower (4%) return increases the fifth percentile outcome by 30% but decreases the 95th percentile outcome by only 3%. compared with the stock and mutual fund market, the bond market is a closer alternative to annuities for retirees. however, bond investors still have to face many risks, such as credit risk (bonds may not pay as expected), inflation risk (as the inflation rate increases, the coupon income values becomes less), reinvestment risk (investors may not be able to find a similar bond investment option after existing bonds mature), and liquidity risk (bonds may be unavailable to meet short-term financial goals). thus, if retirees only invest in stock, mutual funds, and bonds, they have greater risk for meeting their short-term and long-term financial goals. having an annuity could help retirees have lifetime income or certain period income, which could be a protection for their financial goals and reduce risk (including systematic risk from overall market returns and unsystematic risk from individual investment selection). therefore, annuities could improve retirees’ welfare and could be valuable for retirees’ portfolios. 4.4. annuity disadvantages despite the previous advantages, the reality of annuity pricing may reduce the advantages compared with theoretical potential.there is adverse selection with regard to the mortality risk in the annuity market (blake, 1999). this is the risk that only individuals who think they will live longer than the average for the population based on their own medical and family histories will choose to purchase annuities. and also, conversely, those who know of a major mortality risk factor will not ever purchase an annuity (i.e., sick people do not buy annuities). however, insurance companies do not have the same access to this information with the same degree of reliability as do annuity purchasers (blake, 1999). therefore, there is asymmetric information between the annuity insurance company and prospective annuitants. insurance companies may not be able to perfectly distinguish the prospective annuitants who have heavier mortality risk (insurance companies make money) from those who have lighter mortality risk (insurance companies lose money). however, companies realize that people who voluntarily purchase an annuity have a lighter mortality risk than the typical person of the same age. thus, pricing for all prospective annuitants will reflect this reality, potentially making the annuity (in y. yan and r. n. james / financial services review 29 (2021) 277–291 289 practice) a poor investment choice for those with a typical life expectancy. once this is combined with the need for paying commissions, advertising, and profits, the theoretically advantageous annuity may become practically less attractive. references abel, a. (1985). capital accumulation and uncertain lifetimes with adverse selection. nber working paper series, 1664. cambridge, ma: national bureau of economic research. agnew, j. r., anderson, l. r., gerlach, j. r., & szykman, l. r. (2008). who chooses annuities? an experimental investigation of the role of gender, framing, and defaults. american economic review, 98, 418– 422. baker, h. k., & nofsinger, j. r. (2010). behavioral finance: an overview. in h. k. baker & j. r. nofsinger (eds.), behavioral finance: investors, corporations and markets (pp. 3–20). hoboken, nj: john wiley & sons. banerjee, s. (2013). annuity and lump-sum decisions in defined benefit plans: the role of plan rules. ebri issue brief, 381, 1–20. barberis, n., huang, m., & thaler, r. h. (2006). individual preferences, monetary gambles, and stock market participation: a case for narrow framing. american economic review, 96, 1069–1090. bernheim, b. d. (1991). how strong are bequest motives? evidence based on estimates of the demand for life insurance and annuities. journal of political economy, 99, 899–927. benartzi, s., previtero, a., & thaler, r. h. (2011). annuitization puzzles. journal of economic perspectives, 25, 143–164. blake, d. (1999). annuity markets: problems and solutions. the geneva papers on risk and insurance issues and practice, 24, 358–375. brown, j. r., & poterba, j. m. (2000). joint life annuities and annuity demand by married couples. the journal of risk and insurance, 67, 527–553. brown, j. r. (2001). private pensions, mortality risk, and the decision to annuitize. journal of public economics, 82, 29–62. brown, j. r. (2003). redistribution and insurance: mandatory annuitization with mortality heterogeneity. journal of risk & insurance, 70, 17–41. brown, j. r. (2007). rational and behavioral perspectives on the role of annuities in retirement planning. nber working paper no. w13537. cambridge, ma: national bureau of economic research. bütler, m., & teppa, f. (2007). the choice between an annuity and a lump sum: results from swiss pension funds. journal of public economics, 91, 1944–1966. cox, s. h., & lin, y. (2007). natural hedging of life and annuity mortality risks. north american actuarial journal, 11, 1–15. choi, j. j., laibson, d., madrian, b. c., & metrick, a. (2004). for better or for worse: default effects and 401 (k) savings behavior. in d. a. wise (ed.), perspectives on the economics of aging (pp. 81–126). chicago, il: university of chicago press. davidoff, t., brown, j. r., & diamond, p. a. (2005). annuities and individual welfare. american economic review, 95, 1573–1590. donnelly, c. (2015). actuarial fairness and solidarity in pooled annuity funds. astin bulletin, 45, 49–74. finke, m. s., howe, j. s., & huston, s. j. (2017). old age and the decline in financial literacy. management science, 63, 213–230. finkelstein, a., & poterba, j. (2004). adverse selection in insurance markets: policyholder evidence from the uk annuity market. journal of political economy, 112, 183–208. friedman, b., & warshawsky, m. (1990). the cost of annuities: implications for saving behavior and bequests. the quarterly journal of economics, 105, 135–154. 290 y. yan and r. n. james / financial services review 29 (2021) 277–291 james, r. n., iii., (2016). an economic model of mortality salience in personal financial decision making: applications to annuities, life insurance, charitable gifts, estate planning, conspicuous consumption, and healthcare. journal of financial therapy, 7, 62–82. jianakoplos, n. a., & bernasek, a. (1998). are women more risk averse? economic inquiry, 36, 620–630. kahneman, d., knetsch, j. l., & thaler, r. h. (1991). anomalies: the endowment effect, loss aversion, and status quo bias. journal of economic perspectives, 5, 193–206. knell, m. (2018). increasing life expectancy and ndc pension systems. journal of pension economics & finance, 17, 170–199. laferrère, a. (2012). housing wealth as self-insurance for long-term care. in financing long-term care in europe (pp. 73–90). london, uk: palgrave macmillan. lockwood, l. m. (2012). bequest motives and the annuity puzzle. review of economic dynamics, 15, 226–243. lockwood, l. m. (2018). incidental bequests and the choice to self-insure late-life risks. american economic review, 108, 2513–2550. mitchell, o. s., & mccarthy, d. (2002). estimating international adverse selection in annuities. north american actuarial journal, 6, 38–54. modigliani, f. (1986). life cycle, individual thrift, and the wealth of nations. science, 234, 704–712. mottola, g., & utkus, s. (2008). lump sum or annuity? an analysis of choice in db pension payouts. pension benefits, 17, 2–4. pang, g., & warshawsky, m. (2010). optimizing the equity-bond-annuity portfolio in retirement: the impact of uncertain health expenses. insurance: mathematics and economics, 46, 198–209. pecchenino, r., & pollard, p. (1997). the effects of annuities, bequests, and aging in an overlapping generations model of endogenous growth. the economic journal, 107, 26–46. peijnenburg, k., nijman, t., & werker, b. j. (2016). the annuity puzzle remains a puzzle. journal of economic dynamics and control, 70, 18–35. purcal, s., & piggott, j. (2008). explaining low annuity demand: an optimal portfolio application to japan. journal of risk & insurance, 75, 493–516. pyszczynski, t., greenberg, j., & solomon, s. (1999). a dual process model of defense against conscious and unconscious death-related thoughts: an extension of terror management theory. psychological review, 106, 835–845. salisbury, l. c., & nenkov, g. y. (2016). solving the annuity puzzle: the role of mortality salience in retirement savings decumulation decisions. journal of consumer psychology, 26, 417–425. scott, j. s. (2008). the longevity annuity: an annuity for everyone? financial analysts journal, 64, 40–48. thaler, r. (1985). mental accounting and consumer choice.marketing science, 4, 199–214. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101, 449–472. vidal-meliá, c., & lejárraga-garcı́a, a. (2006). demand for life annuities from married couples with a bequest motive. journal of pension economics and finance, 5, 197–229. williams, j. a., & james, r. n. , iii., (2019). bequest provision preferences in commercial annuities: an experimental test of the role of mortality salience. journal of financial counseling and planning, 30, 121–131. the world bank. (2019). life expectancy at birth, total (years). available at https://data.worldbank.org/ indicator/sp.dyn.le00.in?cid=gpd_10 yaari, m. e. (1965). uncertain lifetime, life insurance, and the theory of the consumer. the review of economic studies, 32, 137–150. zizzo, d. j. (2010). experimenter demand effects in economic experiments. experimental economics, 13, 75–98. y. yan and r. n. james / financial services review 29 (2021) 277–291 291 manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) pii: 1057-0810(95)90003-9 financial services review, 4(2): 71-80 copyright q 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. fund closings as a signal to investors: investment performance of open-end mutual funds that close to new shareholders timothy r. smaby john l. fizel this article examines the growing phenomenon of mutualfund closings by analyzing the investmentper$ormance of open-endfinds that close to new investors. we find that: (i) the average excess return (estimated by jensen’s alpha) was positive in the 24 months prior to closing, (2) the average excess return was not sign$icantly differentfrom zero in the 24 months ajier closing, and (3) the funds in the sample on average exhibited a significant decline in investment pe$ormance afrer closing. these findings suggest that thefund managers’ strategic decision to close the find in order to slow down the growth in net assets does not prevent investmentperjormancefrom declining. for the individual investor, an impendinghnd closing is a signal not to invest in the fund. it is also a signal that current shareholders consider alternative investments. i. introduction in june, 1991, janus funds decided to close its janus venture fund to new investors as of september 30, 1991. after that date, it would only accept additional investment from existing shareholders. the fund’s high returns had gained for it a reputation as a “hot” small company fund, inducing investors to pour $500 million into the fund during the previous three quarters and limiting the fund manager’s ability to specialize in small company stocks. the impending closing prompted an additional $500 million in new funds (see laderman, 1993). should investors have viewed this impending closing as a signal to jump into the fund, as it appears they did, or does a hot mutual fund that closes to new investors cool off once the euphoria is past? the ability of fund managers to maintain investment performance after a closing is of interest to both investors and researchers. as the example above suggests, investors may respond to imminent closings by pouring large dollar volumes into the fund to beat the timothyr.smaby and johnl.fiel l penn state erie, the behrend college, school of business, station road, erie, pa 16563-1400; e-mail: trslo@psuvm.psu.edu 72 financial services review 4(2) 1995 closing deadline. the volume of new investments made near closing should not, however, be seen as extraordinary. rather, it represents a continuation of rapid fund growth that precipitated the closing in the first place. for example, median net asset growth in our sample of closed funds was approximately 65% in the nine months prior to the closings and an additional 15% in the quarter in which the closing occurred. nevertheless, the flow of new funds around the closing suggests that investors believe that hot funds will continue to be hot. this assumption is consistent with the empirical evidence provided by hendricks, patel, and zeckhauser (1993) and grinblatt and titman (1992). using large samples of no-load growth-oriented funds, they find that funds which have higher-than-average risk-adjusted returns in one period are more likely to have higher-than-average risk-adjusted returns in subsequent periods. however, the wall street journal wams investors that “hot stock funds can cool after closing” (mcgough, 1993, p. cl). there are two reasons why fund performance might decline after a closing. first, if the fund is closed too late, the large growth in assets may hinder the fund manager’s ability to invest in a manner consistent with the fund’s stated investment objective. for example, the fund manager of an income fund like the lindner dividend fund cannot provide the income his shareholders want if he has too much cash chasing too few good investment opportunities. according to the manager: “i can’t invest money [in treasury bills] at 3% and continue to pay out the 7% yield that i’m trying to give my shareholders” (see laderman, 1993). the manager of the oppenheimer global biotech fund, a sector fund, had to close to “keep the fund purely biotech oriented, and to control the stampede of investors into this hot and risky field” (montgomery, 1991, p. 75). the rush of new money coming in under the deadline exacerbates the fund manager’s predicament. funds often mitigate the announcement effect by giving short notice of the impending closing and making announcements only to current fund family shareholders.’ current shareholders usually have the information to make new or additional investments, but individuals considering investing for the first time must do timely, indepth research to spot upcoming closings and then act quickly.* fund performance may also decline after closing because the compensation of most fund managers is based on the asset size of the fund.” closing a fund restricts asset growth and size-based compensation. the fund manager can be expected to respond by redirecting some attention from the closed fund to other funds he or she manages, perhaps a clone fund opened at the time the original fund was closed. although compensation may also deteriorate if a decline in performance prompts investors to sell out, closing a fund dampens the incentives for managers to maximize returns of that fund. indeed, there is evidence to suggest that a decline in performance does nor prompt investors to sell out. according to sirri and tufano (1993, p. i), “consumers base investment decisions on historical performance asymmetrically: exceptionally high performing funds reap rewards (in terms of new money inflow), but poorer performance is generally not penalized.” the conflicting hypotheses and empirical evidence leave unanswered two important questions facing the individual investor. have funds that closed generated “excess” returns prior to the closing? and if so, did they close in time to maintain that performance level after the closing? if the answer to both questions is yes, the appropriate response of an individual investor to such a closing would be to jump into the fund. if the answer to the second question figure 2. number of mutual funds reported by barron’silipper mutual fund quarterly to be closed to new investors as of the end of each quarter. is no, the closing is a signal to the investor that the fund manager believes that it is not possible to maintain the previous level of investment performance. these questions have become increasingly important as the rate of fund closings accelerates. as interest rates have dropped to xi-year lows in recent years, investors have searched for alternatives to the relatively low yields available through bank certificates of deposit and money-market mutual funds. the result has been a flood of money into open-end mutual funds that invest in equity and fixed-income securities. although mutual funds have been closing intermittently for many years (state street investment trust was closed in 1944 because it was too big and was not reopened until 19891, the recent growth in the mutual fund industry has led to a spate of new closings, figure 1 documents the growth in the number of closed funds, as reported by lipper analytical services in the quarterly mutual funds issue of burron’s (called barran’s/lipper mutual. fund quarterly), from the second quarter of 1985 (when lipper first reported funds as “closed to new investors”) to the third quarter of 1993. as of september 30, 1993, 64 funds were closed to new investors, in this paper, we examine the growing phenomenon of fund closings by comparing the excess returns of closed funds before and after the closings. our results suggest that closed 74 financialservicesreview 4(2) 1995 funds do earn positive excess returns prior to closing but that fund performance drops significantly after closing. for the individual investor, an impending fund closing is a signal not to invest in the fund. ii. methodology the analysis of the investment performance of mutual funds closed to new investors is done in three steps, first, we split the sample of fund returns into preand post-closing subsamples. second, we estimate the excess returns of each fund before and after closing. finally, we test whether the average excess returns of the funds before and after closing are significantly different from zero and whether there is any significant difference between the average preand post-closing excess returns. excess returns for each fund are estimated using the following regression equation, derived from the capital asset pricing model: (ri,r rf,,) = a; + pi(rb,r rf,,) + ei,t i=l,...,n i=-m,... 1, pre-closing, and t=l,... m, post-closing. where ri,, is the total rate of return for fund i in month t, rf, is the risk-free rate of return in month t, rb,, is the benchmark portfolio return for month t, and pi measures the risk of mutual fund i relative to the risk of the chosen benchmark portfolio. the additional return (or the excess return) on mutual fund i after adjusting for risk is represented by ai, commonly known as jensen’s alpha. an cli that is statistically significantly different than zero is evidence that mutual fund i has demonstrated a superior risk-adjusted investment performance. an analysis of the differences between ctpre and c$,,,~ is the key to determining if impending closings are signals to invest in the fund. if post-closing performance is expected to be at least as good as pre-closing performance (a,,, 2 a,,), closing should prompt investors to rush to invest in the soon-to-be closed fund. alternatively, if post-closing performance is expected to be worse than pre-closing performance (a,,, < a&, then closing should induce investors to withdraw their investment. iii. data we identified all mutual funds “closed to new investors,” as reported in barron’shpper mutual fund quarterly from the second quarter of 1985 to the third quarter of 1993. a total of 128 different fund names were reported as closed at some time during that period. after accounting for terminations, mergers, and renaming of funds, we identified 91 funds for which return data was available on the morningstar data base “ondisc” as of september 30,1993. because the ondisc data base only contains data for funds in existence as of that date, any fund that was terminated or merged into another fund subsequent to its closing is not included in our sample. in addition, because the lipper listing is more inclusive fund closings us a signul to investors 75 than the morningstar data base, some current funds noted in lipper as closed were not on the momingstar data base. whenever possible, the actual closing date was determined by contacting the investor services department of each fund.4 funds close to new investors for a number of reasons. the following list, although not exhaustive, encompasses most of our sample. 1. the fund manager closes the fund for strategic reasons, usually to slow down net asset growth, as discussed above. 2. the fund is closed pending a merger or liquidation. in one case, the fund was closed briefly pending a proxy vote of shareholders to change the investment objective. 3. in at least one case, the fund originally only issued a limited number of shares, which sold out immediately. the fund then only accepted new investment when existing shareholders sold back to the fund. in this case, the fund was never really “open.” 4. funds are sometimes closed for legal or regulatory reasons. four steadman funds were closed to new investors in 1989 on the order of the securities and exchange commission in a dispute over failure to register funds in some states where the funds were sold. two first investor funds were closed to new investors in 1990 after state officials in two states filed legal complaints against the company. as we are interested in the effect on performance of closing a fund that is growing too quickly, we only included closings that we determine to be the result of a strategic decision related to net asset growth. clearly, our sample of funds has a “survivorship effect” because we only include funds still in existence as of september 30, 1993. however, the bias this might introduce into the analysis does not affect the interpretation of our results. first, we are interested in funds that close because of above-average performance, and it is unlikely that such a fund would subs~uently be liquidate because of poor perfo~ance. second, we find that investment performance of our sample of funds drops significantly after the closing. such an effect would be exacerbated by any poorly performing funds that were subsequently liquidated. we were able to identify 17 funds that were closed pending liquidation or merger. these observations were dropped from our final sample because they do not appear on the momingst~ data base. selecting the appropriate length of time for the preand post-closing periods involves a tradeoff. because the number of mutual fund closings has increased significantly in recent years, many of the 91 funds for which data were available were closed between september 199 1 and september 1993. reliable estimation of g,,, is difficult because return data for post-closing periods is sparse for these recently closed funds. in addition, some funds closed quickly after shares were issued and had only a brief track record of pre-closing return performance. therefore, the tradeoff in selecting an appropriate sample period involves the size of the closed fund sample versus the number of observations available for each fund in the preand post-closing periods. ultimately, we identified 2.5 funds (20 equity funds and 5 hybrid fund# for which 24 months of return data were available on either side of the closing month. the closing dates range from december 1982 to august 1991. table 1 is a breakdown of the funds in the sample by inves~ent objective as reported by ~orningstar. results for two benchmark portfolios &, in equation (1)) are reported: first, the s&p 500 index for equity funds or an index consisting of 50% of the return on the lehman brothers aggregate bond index and 50% of the s&p 500 for hybrid funds; and second, the 76 financial services review 4(2) 1995 table 1 a breakdown of the 25 funds in the sample by the investment objective of each fund as categorized by morningstar type investment objective number total equity aggressive growth 1 growth 6 growth-income 4 small company 6 specialty 3 total equity 20 hybrid balanced 1 corporate high yield 3 income 1 total hybrid 5 table 2 summary statistics for the monthly average excess returns before and after closing statistics are estimated using jensen’s alpha using two types of benchmark portfolios, a market portfolio (panel a) and an objective porrfolio (panel b). the t-statistic and two-tailed significance levelfor testing whether the average alpha is different than zero in the pre-closing and post-closing are also reported. the s&p 500 index is used as the benchmark portjolio for equity funds a 5040 mix of the lehman brother aggregate bond index and s&p 500 index is usedfor ilybridfirnds. mean 0.41% -0.13% standard deviation 0.84 0.52 t-statistic 2.83*** -1.25 median 0.56% -0.08% maximum 2.34% 0.77% minimum -2.22% -1.68% panel b the average of alljimds with similar invesmtent objective is used as the benchmark portfolio for each fund. mean 0.50% 0.03% standard deviation 0.52 0.47 t-statistic 4.79*** 0.34 median 0.52% 0.00% maximum 1.34% 1.35% minimum -1.07% 4.84% note: ***significant at the 1% level. fund closings as a signal to investors 77 average return of the funds with the same investment objective. hereafter, the first bench mark is referred to as the “market” portfolio, and the second as the “peer group” portfolio. there is evidence (e.g., lehmann & modest, 1987) that mutual fund performance measures are sensitive to the benchm~k against which excess returns are estimated. we use two distinct benchmarks portfolios for each fund to demonstrate the robustness of our results to the choice of benchmark. iv. e~~rical results a. pm-closing excess returns the monthly mean excess returns as estimated by jensen’s alpha are reported in panels a and b of table 2 for the market benchmark and peer group benchmark portfolios, respectively, the mean monthly excess return in the pre-closing period is 0.47% (about 6% on an annual basis) using the market portfolio. the r-statistic of 2.83 is significantly different from zero at the 1% significance level. individual fund excess returns range from 2.34% to -2.22%. the mean monthly excess return in the pre-closing period using the peer group table 3 summary statistics for the betas of the 25 funds in the sample before and after closing statistics are estimated using two types of benchmark portfolios, a market portfolio (panel a) and an objective portfolio (panel b). the t-statistic and two-tailed significance level for testing whether the average beta is d#erent than 1 irr the pre-closing and post-closing are also reported. pp= bposl panel a the s&p 500 index is used as the benchmark portfolio for equity funds. a 50/50 mix of the lehmnn brother aggregate bond index and s&p 500 index is usedfor hybrid funds. mean 0.83 0.79 standard deviation 0.28 0.30 t-statistic -3.11*** -3.3s*** median 0.78 0.73 maximum 1.46 1.32 minimum 0.40 0.33 panel b the average of ai~~ads with similar i~ves~ent objective is used as the benchmark po~fol~o for each fund. mean 0.87 0.89 standard deviation 0.31 0.35 t-statistic -2.12** -1.60 median 0.82 0.88 maximum 1.60 1.44 mi~rn~ 0.45 0.18 notes: ***significant at the 1% level. **significant at the 5% level. 78 financial services review 4(2) 1995 benchmark portfolio is 0.50% (r-statistic of 4.79), which is also significantly different from zero at the 1% level. individual excess returns range from 1.34% to -1.07%. it appears that funds that close to new investors earn significant excess returns on average in the 24 months prior to the closing. this result is robust to the choice of benchmark portfolio. b. post-closing excess returns the mean monthly excess return in the post-closing period is -0.13% using the market benchmark and 0.03% using the peer group benchmark, neither of which is significantly different from zero. funds that close to new investors do not earn excess returns on average in the 24 months following the fund closing. c. beta estimates summary statistics of the estimated fund betas are reported in table 3, panels a and b. the funds in our sample appear to be less risky than both the s&p 500 and their own investment peer group averages, although market portfolio betas range from less than 0.5 to table 4 summary statistics for the difference in monthly average preand post-closing excess returns returns are estimated using jensen’s alpha using two types of benchmark portfolios, a market portfolio (panel a) and an objective portfolio (panel b). the t-statistic and two-tailed significance level for testing whether the average difference is different than zero are also reported. panel a the s&p 500 index is used as the benchmark portfolio for equity funds. a 50/50 mix of the lehman brother aggregate bond index and s&p 500 index is used for hybridfunds. mean -0.60% standard deviation 1.11 t-statistic -2.72*** median -0.46% maximum 2.99% minimum -4.02% panel b the average of allfunds with similar investment objective is used as the benchmark portfolio for each fund. mean difference -0.47% standard deviation 0.72 t-statistic -3.27*** median difference -0.57% maximum difference 2.42% minimum difference -1.43% note: ***significant at the 1% level. fund closings as a signal to investors 19 nearly 1 so. the mean pre-closing betas were 0.83 and 0.87 relative to the market benchmark and the peer group benchmark, respectively. the mean post-closing betas were 0.79 and 0.89, respectively. the reported t-statistics in table 3 test whether the mean beta is significantly different than 1. only the post-closing peer group benchmark mean beta is not significantly different than 1 at the 5% level. it is somewhat surprising that the funds in our sample are somewhat less risky than average, given that 10 of the funds are aggressive growth, small-company, or specialty funds. it may be that the buildup in cash of these funds prior to closing reduces the systematic risk of the portfolios. d. change in investment performance after closing in table 4, we test whether the mean post-closing alphas are different than the pre-closing alphas. using the market benchmark portfolio, post-closing monthly excess returns are 0.60% less than pre-closing excess returns, which is significantly different from zero at the 1% level. using the peer group benchmark portfolio, post-closing monthly excess returns are 0.47% less than pre-closing excess returns, also significant at the 1% level. investment performance of the funds in our sample declines significantly on average after the closing of the funds to new investors. v. conclusion the purpose of this paper was to analyze the investment performance of open-end mutual funds that closed to new investors. we report three important results based on an analysis of 25 funds that closed to new investors between 1982 and 1991: 1. the average excess return was positive in the 24 months prior to the closings. 2. the average excess return was not significantly different from zero in the 24 months after the closings. 3. the funds in the sample on average exhibited a significant decline in investment performance after the closing. this suggests that the fund manager’s strategic decision to close the fund in order to slow down the growth in net assets did not prevent investment performance from declining. this decline in performance may have occurred because the dollar flow into the fund outstripped the fund’s investment opportunities and the funds closed too late to shut off the flow of funds. a number of the funds that close invest in high-yield, specialized industry, or small-company securities. because there are a limited number of undervalued securities in these categories, a fund manager may have too many dollars chasing too few good opportu nities. this decline in performance may also be the result of the form of the fund manager’s compensation. because compensation is typically tied only to the asset size of the fund and since closing the fund limits net asset growth, incentives for fund managers to maximize returns are reduced this represents an area of future research. whatever the reason for the decline in investment performance, the individual investor should interpret an impending closing as a signal not to invest in the fund. it is also a signal that current shareholders consider alternative investments. 80 financial services review 4(2) 1995 notes 1. fund families representing many of the closed funds included in our sample provided this information. they emphasized that public announcements of closings were rare; because the goal of closing the fund was to preserve performance and not increase the asset size, funds did not want public announcements to generate additional investments. 2. atlas and zweig (1993) discuss several methods for investing in funds that have closed to new investors. for instance, buying a share of a closed fund from an existing shareholder entitles the new shareholder to make additional investments. 3. golec (1992, pp. 88-89) reports that only 29 of 476 mutual funds included in moody’s bank and finance manual (1985) used performanceor return-based compensation for fund managers. asset size of the fund was the only compensation criterion used for the remaining funds. 4. determining the actual closing date was difficult in some cases. the fund itself sometimes did not have records going back far enough. therefore, for some funds, the closing month was determined as the middle month of the quarter in which the fund was first reported closed by lipper. 5. momingstar classifies funds according to investment objective as equity, hybrid, fixed-in come (taxable-bond), or municipal bond (tax-free) funds. the hybrid category includes funds that typically invest in both equities and bonds, as well as corporate high-yield (junk bond) funds. references atlas, r., & zweig, j. (1993). getting past the bouncer. forbes, (august 30), 114. golec, j.h. (1992). empirical tests of a principal-agent model of the investor-investment advisor relationship. journal of financial and quantitative analysis, 27, 81-95. grinblatt, m., & titman, s. (1992). the persistence of mutual fund performance. journal of finance, 47,1977-1984. hendricks, d., pate], j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974-1988. journal of finance, 48,93-l 30. laderman, j.m. (1993). investor, keep out. business week, (march 29), 78. lehmann, b., & modest, d. (1987). mutual fund performance evaluations: a comparison of bench marks and benchmark comparisons. journal of finance, 42.233-265. mcgough, r. (1993). hot stock funds can cool after closing. wall street journal, (march 23). cl & c21. montgomery, l. (1991). fw’s quarterly mutual fund review. financial world, (november 12), 74-84. moody’s bunk and finance manual (1985). new york: moody’s investor services. sirri, e.r., & tufano, p. 1993. buying and selling mutual funds: flows, performance, fees and services. working paper 93-017, harvard business school, cambridge, ma. pii: 1057-0810(93)90003-9 efficient frontiers in estate planning ronald r. crabb this article explores the nature of the efficient frontier in probabilistic estate planning for 16 different estate pkms by considering as random variables ages at deatk rates of return on assets, and borrowing rates on debts. the s~a~tion considers two couples, one middle aged, the other elderly. two &&i6 ~t~ces, one for each couple, are used to record and compare the z-es&s of every s~a~t~~n. that eqmfar~‘ve data, in ~~j~~‘o~ with the co&icient of ~~~n based e~~~ntfro~der, contain use1 ~nfo~ti~n for couples who, ~o~~stent with their tie& of?+& desire to muximize the net present value of assets passitxg to their heirs. the eficient frantier is shown to be a jkaction of three fators: ~s~ptions, ages of the estate owners, and the discount raies of the heirs. became of the instabili~ shown in the eficient frontier, estate pkmners and estate owners must care&ily examine not only the estate phzns which fall on the e@s?nt frontier but also those estate pkm which fall just off that frontier. 1. intr0dum0n in the second issue of the f~~~i~~ services review (esr), crabb (1992) demon strates that traditional point estimate estate planning based on life expectancy is neither realistic nor unbiased, and that ~~pmbabi~istic estate planning permits modern portfoliu theory (mean, variance tradeoffs) to be used to select an optimal estate plan.” rather than ignoring risk by assuming that death occurs at life expectancy, probabilistic estate planning treats ages at death as a random variable. markowitz”s (1952) e-v rule is applied, and an optimal estate plan is defined as one which falls on the efficient frontier. the pm-pose of this paper is to explore the nature of that frontier. in crabb’s (1992) article, the only random variable considered was ages at death. only three afternate estate plans were considered, inves~ent returns on assets were fixed over the dnration of the anaiysis, and a single discount rate was used to compute the net present value of assets passing to heirs. those rather severe and unrealistic limita tions were necessary to demonstrate the superiority of making estate planning ronald r. crabb 0 finance and law, 5003 carlson hail, uw-whitewater, whitewater, wi 53190-1790. 2 financial servicxs review, 3(l) 1993 decisions based on a rnea~v~~ce tradeoff versus making estate planning deci sions based on the remaining life expectancy(ies) of an estate owner(s). in the real world, iras, tsas, and 4ol(k)s are used to accumulate wealth on a tax deferred basis. annual gifts are used to pass wealth to heirs to avoid estate taxes on both the gifts and growth of the gifts after receipt by the heir(s). whole life insurance trusts and term life inset trusts are colon estate planing tools. rates of return on assets may vary with, among other things, inflation expectations and the state of the economy. with the popularity of home equity loans and variable rate mortgages, borrowing rates on debts can vary as well. wills are revised as income tax, estate tax, and family situations change. although it is impossible to consider the universe of possible estate plans and to predict changes in income taxes, estate taxes, and family situations, the 16 plans chosen for analysis employ estate planning tools which estate planners commonly recommend to estate owners. consider ages at death, rates of return on assets, and borrowing rates on debts as random variables. a computer simulation, which systematically uses alternate wills, annual gifts, whole life insurance, term life insurance, leverage, and tax shelters, can generate a set of attainable estate plans (consisting of the net present values of the after-tax estates passing to the heir(s) and the standard deviations associated therewith). the shape of the efficient frontier emerges from among the set of attainable estate plans. multivariable probabilistic estate planning can probe the nature of the estate plans which fall on the efficient frontier, and the nature of those which do not. 11. &&itiodology the random death selection process is thoroughly explained on pages 144 through 147 of the second issue of the fsr, but a quick review is in order. given a mortality table and the age of an estate owner, divide the number of persons expected to be alive at some future age by the number of persons alive at the age of the estate owner today. the resulting quotients compute for that estate owner the probabilities of survival to any future age. except for very old people, the probability survival curve is shaped like a ski slope, with the probability of survival decreasing at an increasing rate through about age 80, and then decreasing at a decreasing rate to allow for the few persons who live into their early 100s. the computer chooses a random number, then interpolates the expected age at death from the survival curve. random deviations from the expected rates of return on assets and borrowing rates on debts are computed via a metrology similar to that used to compute random death ages. that is, a random number, defined on the unit internal [0 to i], can be interpolated to yield a normal distribution, just as a random number can be interpolated to yield an expected age at death. assume, for example, that a particular asset has an expected rate of return of 10% with a standard deviation of 5%. assume that the computer generates the random number 0.4544. that number falls into the interval between 0.44433 and effieiertt frontiers in es&&e pitanning 3 0.46414, and that interval is associated with a minus 0.10 deviation from the normal expectation. hence, assuming an expected rate of return of 10% with a standard deviation of 5%, the 5% standard deviation is multiplied by the -0.10 randomly chosen deviation, and results in a randomized rate of return of 9.5% [lo% + (5%)*(-0.10)] for the time period under consideration. randomly chosen numbers close to 0.500 result in small deviations from the expectation; randomly chosen numbers close to one or zero result in large deviations (positive or negative) from the expectation. the computer simulation generates a random number for each asset and for each debt for each time period until death, and then interpolates from those random numbers deviations from the expectations, resulting in random and normally distributed expected rates of retum for all assets and all debts. two types of life insurance are considered: whole life insurance and term life insurance. the whole life policy is a tiaa unisex whole life insurance policy. although the author would have preferred to be consistent and use tiaa term life insurance, tiaa ends all of their term products at age 70. hence, a commercial insurance product, with rates guaranteed for 20 years, is used for the term life insurance policy. the term product has a terminal age of 95; that is, at the end of age 94, or the coming of age 95, the policy is te~nat~, that could be catastrophic from an individual investor’s viewpoint; the premium for the $l~,~ policy in the 94th year of life is in excess of $30,000. debt (estate leverage) is, for tax purposes, assumed to be secured by a mortgage on the house. that mortgage is assumed to be a variable rate home equity mortgage, and, although in the real world payments would be made on a monthly basis, the simulation mortgage payment is made annually. hence, on an annual basis, the old balance is increased by the variable borrowing rate and is reduced by the mortgage payment. since the mortgage payment is a function of the size of the end-of-the-year balance (for computational purposes, the payment is 15% of the end-of-the-year balance), the mortgage will exist for the duration of the simulation, the tax deductibility of the interest is accounted for by reducing the amount of the payment. that is, if the gross payment on the mortgage is $16,500, and the tax savings associated with that mortgage payment is $2,800, then the net payment is $13,700 [the difference between $16,500 less $2,800]. when the mortgage is incurred, an offsetting asset account is created. see figure 1. payments on the mortgage are made from that asset account. if the after tax cost of borrowing is equal to the after tax earnings on the asset account, the transaction is, for future net worth purposes, a financial “wash.” the results of this dua1 transaction (an increase in debt accompanied by an increase in assets) are visually linked together on the computer screen. by compar ing the future mortgage balance with the future asset account balance, the individual investor can see the effect of leverage. obviously, if the after tax rate of return on the asset account is greater than the after tax borrowing rate on the mortgage, then the leverage will serve to increase the net present value of the assets passing to the 4 financial services review, 3(l) 1993 heir(s). conversely, if the reverse is true, the leveraged estate will reduce the heir(s)‘s inheritance(s). sixteen different estate plans were considered for this simulation. the first estate plan is referred to in tables 1 through 8 as plan a, and in the text as either simple will, as a (simple will), or, if recently described, as a. under this estate plan, the husband’s assets are passed to his widow when he dies, or the wife’s assets are passed to her widower when she dies, and on the death of the survivor the remaining assets are passed to their heir(s). when estate planners first work with couples, they typically encounter the simple will estate plan. often, the first step in estate planning is to show clients the value of the exemption trust will. the primary characteristic of this estate planning tool is to take advantage of the $600,000 which can be passed free of federal estate tax on the death of an individual. when the first individual dies, the heir(s) receive $600,000 (in trust), and those assets are not a part of the survivor’s gross estate when the survivor dies. as a consequence, the exemption trust will increases the amount an unleveraged estate assets debt net worth rate of return 10% $600,000 so $600,000 beginning assets $600,000 $643.200 $6895 io $739, i55 interest earned $60,000 $64,320 $68,95 i $73,916 tax effect on earnings ($16,800) ($18,010) (s 19.306) ($20,696) end of year assets $643,200 $689,5 io $739,155 $792,374 net worth $643,200 $689,5 io $739, i55 $792,374 a leveraged estate assets debt net worth rate of return 10% $700.000 6 100,000 $600,000 cost of debt 10% beginning debt (f i00,000) ($93,500) ($87,423) ($81,740) interest expense ($ i0.000) ($9,350) ($8,742) ($8,174) end of year debt (6 i 10,000) (f 102,850) (696,165) ($89,9 14) end of year payment $16,500 6 15,428 $14,425 s 13,487 tax effect ($2,800) ($2,6 18) ($2,448) ($2,289) net payment (from assets) $13,700 $12,810 f i 1,977 $i 1,198 end of year debt balance (493,500) ($87,423) ($81,740) ($76,427) beginning assets $700,000 $736,700 $776.933 $020.095 interest earned $70,000 $73,670 $77,693 $82,090 tax effect on earnings (f 19,600) ($20,628) ($2 1,754) ($22,985) net payment (e 13,700) ($12,810) (s i 1,977) ($1 1,198) end of year assets $736,700 $776,933 $820,895 s860,80 i net worth $643,200 $689,5 io $739, i55 $792,374 note: to keep net worth constant, the payment on the mortgage must be made from the asset account to which the mortgage debt was transferred. figure 1. effxient frontiers in estute planning 5 of assets passed to a couple’s heir(s). the exemption trust will is referred to as plan b in tables 1 through 8, and in the text as either the exemption trust will, as b (exemption trust will), or, if recently described, as b. once clients are familiar with the basics (the simple will and the exemption trust will), estate planners typically expose their clients to additional estate plan ning opportunities for passing assets to their heir(s): annual gifts and life insurance trusts are two of the most common tools. to increase the size of their future estates, estate planners often show their clients how to avoid current income taxes by using debt (leverage) and tax shelters. since the particular tool(s) used are often a function of the professional training of the estate planner and the way s/he is compensated for her/his time, estate plans three through 16 do not necessarily represent the order in which an estate planner would introduce her/his clients to alternate estate plans. rather, the order of the plans is related to the logic underlying the computer code used to perform the analysis. that is, after the code has examined the simple will and the exemption trust will, gifts are introduced into the estate planning process, followed by whole life insurance, then followed by term life insurance. after eight alternate estate plans (referred to as plans a through h in tables 1 through 8) were analyzed, control of the program was returned to the author. assets were reallocated to consider the use of tax shelters and/or debt. after reallocation, control of the program was returned to the computer, and the simple will estate plan was modified to include tax sheltering and estate leverage for the age 45 couple, or modified to use only estate leverage for the age 70 couple. as before, gifts, whole life, and term life insurance were sequentially introduced into the analysis to complete the simulation. note that the use of debt, tax shelter, gifts, or insurance is not random, since in the real world those would not be random events. the rates of return (or borrowing rates) are random, but the use of the tool(s) is not. given a set of death ages, a set of input data for the initial estate, and a set of randomly and normally distributed rates of return for all assets and debts over all death ages, the remaining step is to compute the net present value of the wealth which passes to the heir(s) for the 16 different estate plans. a flow chart for the simulation follows. see figure 2. a simulation-a middle aged couple initially, eight different estate plans are considered: a. b. c. d. e. f. g. h. simple will; exemption trust will; simple will and gifts; exemption trust will and gifts; simple will, gifts, and whole life; exemption trust will, gifts, and whole life; simple will, gifts, and term life; and exemption trust will, gifts, and term life. financial services revirw, 3(l) 1993 suction flow chart 1 start 1 i read the moftallty tabte into memory j i read the standard devlatlon data into memory i i read the life insurance data into memory 1 input other variables: i husband‘s age/ wtfe’s age number of trials ]olnt and separate assets annual increase/decease of those assets ages at which the tncrease/decrease stops rates of return on those assets variance of those rates of return tncome taxablllty of those assets compute, for each joint and separate asset, the annual random devlatlon from the expectation based on a standard normal curve and the variances input above set up the assets arrays to hold slmulatlon data, with the size of those arrays a function of the amount of data enput above 1 determlne radon7 death age comblnattons i the first estate plan is a simple wlli and no gifts spouse to spouse remalnder to childken) 1 i tax and distribute estate to helrs for each combination of death aqes 1 input the heir’s tlme preference for money compute and save the net present value of assets passing to heirs for each trlat f output results to screen 1 change the estate plan from a simple wjll and no gifts to an exemption trust wi11 and no glf ts i tax and dtstrlbute estate to helrs for each comblnatlon of death ages i [ input the heir’s time preference for money 1 compute and save the net present value of assets passlnq to helrs for each trlal r output results to screen i figwe 2. simulation fiow chart (confines flow chart for lnlll~lzatlon of the simulaticn simple will nd gifts estate analysis exemption trust will no gifts estate analysis efsicient frontiers in esfate planning 7 change the estate plan from a exemptlon trust will and no gifts to a simple will and gifts i tax and distribute estate to heirs for each combination of death ages 1 input the heir’s tlme preference for money compute and save the net present value of assets passing to heirs for each trial i output results to screen change the estate plan from a simple wi11 and gifts to an exemptlon trust will and gifts tax and dlstrlbute estate to helrs for each comblnatlon of death ages 1 input the heir’s time preference for money compute and save the net present value of assets passlng to helrs for each trial i output results to screen ii change the estate plan from an exemption trust will and gilts to an simple will and gifts and whole life insurance tax and dlstrlbute estate to heirs for each combination of death ages 1 simple will gifts estate analysis exemption trust will gifts estate analysis simple will gifts 1 input the helr’s time preference for money 1 rl whole life insurance compute and save the net present value of estate analysis assets passing to helrs for each trial l output results to screen change the estate plan from a simple will and gifts to an exemptlon trust will and gifts and whole life insurance exemption trust will i tax and distribute estate to helrs for each comblnatlon of death aqes gifts 1 input the heir’s tlme preference for money whole life insurance compute and save the net present value of assets passing to heirs for each trial estate analysis i output results to screen figure 2. continued. fii’mncfal services review, 3(l) 1993 change the estate plan from an exemptton trust will and cflts and wttole cfle to an sffnpie wili and sffts and term lire insurance simple will tax and utstrlbute estate to heirs for each combmatlon of death ages [ input the helr’s tlme prelerence for money term life insurance compute and save the net present value of estate analysis assets passing to heirs ior each trtal i output results to screen change the estate plan from a sfmple wtfl and etfts and term lfle to an exempt&n trust wftl and carts and term lire mwrance exmpticsn trust for each combination of death ages input the heir’s ttme preference for mone tern life insurance compute and save the not present value of estate analysis assets passing to heirs for each trlal 1 if ye5 to any of the above questtons, go to a if no to all or the above questlons, continue or output e/v analysis to the screen analyze the results of each trial by comparing ‘f-----j the net present value or that trial to the net present value of all ather trials, holdmg final analysis constant the death age comblnatlons and the randomized devlatfons for each asset/debt f mttput net present value enalysfs to the screen [ _ assume the following estate: husband-separate assets-$250,000, with a taxable 10% expected rate of return and a standard deviation af 5%; wife-separate assets-$250,000, with a taxable 7% expected rate of return and a standard deviation of 3%; and joint assets (a house) of $250,000, with anon-taxable expected rate of return of 6% and a standard deviation of 2%. assume that both husband and wife are 45 years of age. assume a 100 trial simulation. after these eight plans have been evaluated, the estate assets are reallocated, holding consumption constant. tax shelter use, to reduce current income taxes and emient frontiers in estate planning 9 to compound assets tax deferred, is considered as an asset reallocation. that is, when money is routed into a tax shelter [like a tsa, a 401(k), or an ira], the source of that money is an existing asset. consumption is held constant by reducing an existing asset and putting that money into a tax shelter. the reduction amount is adjusted for taxes; that is, if $2000 is invested in an ira account, then $1440 is removed from an asset account. if $8000 is invested in a tsa account, then $5760 is removed from an asset account. the use of debt, also holding consumption constant, has been previously explained. the next eight estate plans, referred to as i through p in tables 1 through 8 and by their names and/or letters in the text of the paper, are: i. j. k. l. m. n. 0. p. simple will, ira, and mortgage; exemption trust will, ira, and mortgage; simple will, ira, mortgage, and gifts; exemption trust will, ira, mortgage, and gifts; simple will, ira, mortgage, gifts, and whole life; exemption trust will, ira, mortgage, gifts, and whole life; simple will, ira, mortgage, gifts, and term life; and exemption trust will, ira, mortgage, gifts, and term life. for the age 4.5 couple, the results of the simulation (a total of 1600 separate estate outcomes) are summarized in tables 1 through 4. table 1, labelled “com parative analysis-male and female, both age 45various estate plans,” has an alphabetic code at the bottom which refers to the different types of estate plans. in order to read the data shown in the 16x16 matrix, note first a diagonal of bold zeros (0), sloping downward to the right. when reading across the top of the table, from left to right, the numbers lying above that zero diagonal represent the number of times a particular estate plan had a higher net present value to the heir(s) [npvh] than the estate plan shown on the left horizontal axis; when reading from top to bottom, the numbers lying below that zero diagonal represent the number of times that the estate plan shown on the left vertical axis had a lower npvh than the estate plan shown on the top horizontal axis. for example, go across the top line to estate plan c (simple will and gifts), drop down to the number 87, and go left on the horizontal axis to estate plan b (exemption trust will>. relative to the exemp tion trust will estate plan, the simple will and gifts estate plan resulted in a higher npvh in 87 of the 100 trials. reading in the other direction, go down the left edge to estate plan c, go right to the number 13, and go up to estate plan b. relative to the exemption will trust estate plan, the simple will and gifts estate plan resulted in a lower npvh in 13 of the 100 trials. the next three tables, labelled “e-v analysis . . . ? %,” where ? is equal to 4,5, or 6, summarize the results of the 1600 trials discounted at 4,5, and 6 percent. one hundred death age combinations, with randomized rates of return for each asset and borrowing rates for each debt until death for both husband and wife, evaluated 10 financial services review, 3(l) 1993 for estates plans a (simple will> through p (exemption trust will, ira, mortgage, gifts, and term life), resulted in the set of npvhs and standard deviations associated therewith. in the top left rectangle, the several estate plans are npvh ranked, listed from the highest npvh to the lowest npvh. in the top right rectangle, the several estate plans are coefficient of variation ranked, listed from the highest coefficient of variation to the lowest coefficient of variation. in the two lower rectangles, the set of possible outcomes has been reduced to the efficient set, both on a standard deviation basis and on a coefficient of variation basis. for example, in the 4% discount based table, two plans, m (simple will, ira, mortgage, gifts, and whole life) and k (simple will, ira, mortgage, and gifts) are eliminated from consideration. their means are lower than those of plan h (exemption trust will, gifts, and term life), and their standard deviations are higher. after other plans have been eliminated on the same decision criteria, 10 plans remain, and the individual investor, given his/her risk preference, could choose from among those 10. in the 4% discount based table, on a coefficient of variation basis, the 16 plans can be reduced to six plans, and the individual investor would only need to consider six alternate plans, rank order: n. exemption trust will, ira, mortgage, gifts, and whole life; p. exemption trust will, ira, mortgage, gifts, and term life; f. exemption trust will, gifts, and whole life; h. exemption trust will, gifts, and term life; 0. simple will, ira, mortgage, gifts, and term life; and g. simple will, gifts, and term life. unfortunately for the individual investor, the efficient frontier is a function of the individual investor’s discount rate, and in the 6% table, the coefficient of variation based efficient frontier is: n. exemption trust will, ira, mortgage, gifts, and whole life; l. exemption trust will, ira, mortgage, and gifts; d. exemption trust will and gifts; m. simple will, ira, mortgage, gifts, and whole life; 0. simple will, ira, mortgage, gifts, and term life; and e. simple will, gifts, and whole life. see figures 3,4, and 5 for the coefficient of variation based efficient frontiers for discount rates of 4,5, and 6%. the standard deviation based efficient frontier also changed as the individual investor’s discount rate changed. decision criteria problems for the individual investor continue. when dis counted at 4 and 5%, plan a (simple will) is on the standard deviation based efficient frontier. however, in the “comparative analysis table,” the simple will emient frontiers in estate planning 11 table 1. comparative analysis-male and female, both age 45-various estate plans a b c d e f g ii i j k l n n 0 p a 0 c d e i j k l n n 0 p 0 0 ( 5 6 1 42 1 0 0 0 0 1 17 1 61 1 23 ( 91 / 0 0 ] simple wiii exemption trust will simple will and girts exemption trust will and gifts simple wiii. gllts, and whole llle exemptlon trust wiii. gifts. and whole llle simple wiii, giits, and term llle txcmptlon lrust wi,,, gilts. and term ltle simple wiii, ira, and mortgage exemptlon trust wiii. ira. and mortgage simple will, ira. tlortgage, and guts exemption trust will, ira, rlortgage, and gilts simple wiii. ira, tlortgage, gifts. and whole !-ire exemptlon trust wiii, ira, tlortgage, gifts. and whole l_lle simple will. ira, mortgage, gifts, and term i_lfe exemptlon trust wiii, ira, mortgage. gifts. and term ~!le is shown to be inferior to all other plans under consideration. that is, while the simple will does represent the lowest point on the efficient frontier, it is so low that on a nf’vh basis the 100 npvhs of the simple wizz estate plan, when compared to the npvhs of all other plans, is always inferior. while a financial planner, who, when counselling a very risk averse investor, could recommend the simple will estate plan because it has the lowest standard deviation, that financial planner would have to explain very carefully that while the risk is low, so is the size of the 12 financial services review, 3(l) 1993 table 2. e-v analysis-male and female, both age 45-various estate plans--i% discounted at 4% discounted at 4 % net present standard net present standard coefficient value to heirs deviation value to heirs deviation of variation n t 1,585.299 $295,925 l f 1,534,620 $291,595 0.19001121 l f 1,534,620 $29 1,595 k s 1,383.048 $258,847 0.18715692 p t i ,5 17,893 $263,083 n s i ,585,299 $295,925 0.18666826 f f i ,503.692 f253,6 i i j f i, 183,677 $218,924 0.18495248 d s i ,453,o i3 $250,388 n $ i .433,727 $249,824 0.17424796 h $1,436,286 $219.805 p $ i ,5 17,893 $263,083 0.173321 i8 n s i .433,727 $249,824 c fl,315,516 $227,866 0.1732 i41 6 k t i .383,048 s258.847 d 0 i ,453,o i3 $250,388 0.1723233 0 $ i ,366,320 t204,o i6 b f i, i i 1,999 $188,194 0.16923936 e f 1,366, i95 $219,191 f $ i .503,692 $253.6 i i 0.16865887 c sl.315,516 s227.866 e s 1,366, i95 $219,191 0.16043903 g t i ,298,789 16 174,094 i s i ,003,47 i $159,056 0.15850583 j t i, 183,677 $218,924 ii f i ,436.286 $219,805 0.15303707 b fl.l11,999 f 188, i94 0 6 i .366,320 $204.016 0.14931788 i $ i ,003,47 i t 159,056 a s936,039 s127.849 0.13658512 a s936,039 s 127,849 g f i ,298,789 s 174,094 0.13404333 the c-v fronlier (npv verus standard devlatlon) n l p f d ii 0 g i a the e-v frontier (npv verus coefficient of variation) estate being passed to heir(s). [note that this problem can be eliminated if the coefficient of variation is used to select the efficient frontier.] another simulation-an elderly couple this simulation is similar to the one above, but the couple is older and wealthier. assume the following estate: husband-separate assets-$500,000, efficient frontiers in estate planning 13 table 3. e-v analysis-male and female, both age 45-various estate plans-5 % discounted at 5% discounted at 5 8 net present standard net present standard coefficient value to heirs deviation value to heirs deviation of variation n f i ,068,795 t 124,754 j $799,565 $99,456 0.12438764 l t 1,033,8 i2 $121,463 l f i ,033.a i2 $121,463 0.1 1749041 p $1.025.226 f 1 12,099 n f i .068,795 $124,754 0.11672397 f s 1 ,o 14,970 f 100,728 k $931,746 $107,250 0.1 1510648 d $979,988 $97,76 i i3 $752,250 $85,785 0.1 i403789 h $97 i ,40 i $88.728 p f i ,025,226 $i 12,099 0.10934077 tl $966,728 $96,267 c $887, i88 $90,202 0.1016718 k $931,746 f 107.250 d $979,988 t97,76 1 0.09975734 0 $923, i59 $69.305 n $966,728 $96.267 0.09958023 e $922,170 $79,525 f s i ,o i 4,970 t 100,728 0.09924234 c $887, i88 $90,202 ii 3971,401 $88,728 0.09 i34024 g $878.60 i f54,004 e $922,170 $79.525 0.0862368 i j $799.565 $99.456 i $677,506 $56.4 i2 0.0832642 1 i3 $752,250 $85,785 0 $923,159 $69,305 0.07507374 i $677,506 $56.4 12 a $633,028 $39.252 0.06200674 a $633.028 $39,252 g t878,60 i $54,004 0.06 i4659 the e-v frontier the e-v frontier (npv verus standard devlatlon) (npv verus coefficient of variation) with a taxable 10% expected rate of return and a standard deviation of 5%; wife-separate assets-$500,000, with a taxable 7% expected rate of return and a standard deviation of 3%; and joint assets (a house) of $200,000, with a non-taxable expected rate of return of 5% and an expected variance of 1%. assume that both husband and wife are age 70. assume a 100 trial simulation. after the initial eight plans (simple will through exemption trust will, gifts, and term life) have been evaluated, the estate assets are reallocated, holding 14 financial services reyiew, 3(l) 1993 table 4. e-v analysis--hflale and female, both age 45-various estate plans-6% discounted at 6% discounted at 6 % n l p f d ii n k 0 e c g j is i a net present standard alue to heirs devlation s727.526 $703,108 $699,198 $691,701 $667,283 $663,373 6658,108 $633,690 $629,780 $628.447 s604.303 $600, i20 $545,417 $5 13,903 $46 1,866 f46,8 12 $37,552 s54,27 i $37,336 $25,535 $48,524 $20,03 i $3 1,872 s 18,276 6 15,602 $25, i40 $23,263 $53‘47 1 $52,2 i5 $12,100 $432.27 i $14,193 0 j p h n f l k c g d a n 0 i e net present standard coefflclent lrlue to helrs deviation of variation $5 13,903 $545,417 $699,198 $663,373 5727,526 $69 i .70 1 $703,108 $633,690 $604,030 $600, i20 $667,283 $432,27 1 $658,108 $629,780 $461,866 $52,215 0.10160478 $53,47 i 0.0980369 i $54,27 1 0.0776 1893 $48,524 0.073 14738 $46.8 12 0.06434409 $37,336 0.05397708 $37,552 0.05340858 $3 i.872 0.05029589 $25, i40 0.04162045 $23,263 0.0387639 i $25,535 0.038267 i2 $14,193 0.03283357 $20,03 1 0.03043725 f 18,276 0.0290 1966 $12, io0 0.026 19807 $628,447 s 15,602 0.02482628 the e-v frontier (npv verus standard deviation) s727.526 s46,t3 12 6703,108 $37,552 5691,701 $37,336 $667,283 s25,535 6658,108 s20,03 1 s629.780 $18,276 $628.447 $15,602 s46 1,866 s 12,100 the e-v frontier (npv verus coefficient of variation) s727.526 s46.8 12 0.06434409 $703,108 s37,552 0.05340858 s667,283 $25,535 0.038267 12 s658,108 s20.03 1 0.03043725 $629,780 $18,276 0.02901966 $628,447 s 15,602 0.02482628 consumption constant. a $100,000 mortgage is taken out on the house. assume a 7% mortgage rate with a 5% standard deviation, and assume an offsetting investment of $100,000 with an expected taxable rate of return of 8% and with a 4% standard deviation. note that there is, subject to random deviations, positive leverage; that is, the expected rate of return on the invested assets is 1% higher than the borrowing rate. iras, tsas, and 4ol(k)s are not used since the couple is assumed to be retired. e_#tcient frontiers in estate pkiznning 15 the results of the simulation (a total of 1600 separate estate outcomes) are summarized in tables 5 through 8. in table 5, it is interesting to note that estate plan a (simple will) is no longer absolutely inferior. in two of the 100 trials, the simple will estate plan was superior to estate plan g (simple will, gifts, and term life). in table 6, the 4% discount based table, on a coefficient of variation basis, the 16 plans can only be reduced to 10 plans, and the individual investor would need to consider 10 alternate plans, rank order: n. f. p. j. b. m. e. 0. i. a. exemption trust will, mortgage, gifts, and whole life; exemption trust will, gifts, and whole life; exemption trust will, mortgage, gifts, and term life; exemption trust will and mortgage; exemption trust will; simple will, mortgage, gifts, and whole life; simple will, gifts, and mortgage; simple will, mortgage, gifts, and term life; simple will and mortgage; and simple will. at 5%, the coefficient of variation based efficient frontier is: n. exemption trust will, mortgage, gifts, and whole life; l. exemption trust will, mortgage, and gifts; d. exemption trust will and gifts; k. simple will, mortgage, and gifts; and c. simple will and gifts. at 6%, the coefficient of variation based efficient frontier is n (exemption trust will, mortgage, gifts, and whole life), and l (exemption trust wizl, mortgage, and gifts). see figures 6,7, and 8 for the coefficient of variation based efficient frontiers for discount rates of 4,5, and 6%. an analysis of the efficient frontiers in the elderly couple simulation at 6% (above paragraph), estate plans n and l define the efficient frontier; however, estate plan d (exemption trust will and gifts) is almost the equivalent of l (exemption trust will, mortgage, and gifts). it is only in the fourth digit of the coefficient of variation that l outranks d. since in estate plan l the cost of obtaining the mortgage was not considered, it is possible that d could also be on the efficient frontier if the cost of the loan were high enough such that the npvh of l dropped below $938,202 (the npvh of d). the efficient frontier would then be n, d, l. more important is the assumed 1% favorable interest rate differential between the borrowing rate and the investing rate. [note that if negative leverage were 16 financial services review, 3(l) 1993 table 5. comparative analysis-male and female, both age 70-various estate plans a i3 c d e f g tl i j k l n n 0 p a i3 c d e f g tl i j k l n n 0 p simple will exemption trust will simple wtll and gifts exemption trust wtll and girts simple will, girts, and whole llk exemption trust wiii, girts. and whole llre simple wiii. girts. and term life cxempllon trust wiii, girts. and term llre slmple will and mortgage exemptton trust wltl and mortgage sfmple will. mortgage, and gifts exemption trust will, mortgage, and gifts stmpie will, h&gage, girts, and whole lire exemption trust will. mwtgage, girts. and whole llle simple will. mortgage, girts, and term llre excmptlmi trust will. mortgage, gilts. and term life assumed, then, even ignoring the cost of obtaining the mortgage, d would be preferable to l.] ceteris paribus, the positive leverage tilted the analysis in favor of any estate plan of which the mortgage was a part. tltat is, if one compares i. (simple will and mortgage) to a (simple will); j. (exemption trust wiil and mortgage) to i3 fxemption trust will); k. (simple will, mortgage, and gifts) to c (simple will and gifts); efficient frontiers in estate planning 17 table 6. e-v analysis-male and female, both age 70-various estate plans-4 % discounted at 4% discounted at 4 % net present standard net present standard coefllclent value to heirs devlatlon value to heirs devlatlon of varlatlon n t i .x37.976 $1 io.815 n p 1 ,x7.970 silo,8l5 0.08160294 f fl.353.610 $i 10.176 f t i .353.6 io $1 io.176 0.0813942 p t1.344.510 f 105.833 l s 1,326, i49 t 105,432 0.07950238 ii s 1,340, i42 $105,524 d fl.321.781 f 104,455 0.07902595 l fl.326.149 0 105.432 i4 t 1,340, i42 5 105,524 0.0787409 i d sl.321.781 $104,455 p t i .344,5 io s 105.833 0.07671492 j f i .203,2x $92. i47 j s i .203,295 $92, i47 0.07657889 i3 $1.199.131 $9 1,539 0 f1.199.131 $91.539 0.07633778 n 11.1 is.582 $41,743 k t i .083,752 $72.239 0.06665639 e $1.1 il.526 $40.978 c t i .079,696 $71,300 0.0660371 i 0 $1.102.114 $38,619 n $1.1 15,582 $41,743 0.03741814 g f i .098.057 $30.833 e $1.1 il.526 $40.978 0 03666643 k s i ,083,752 $72,239 g s i .098,057 $38,833 0.0353652 c t i ,079,696 t71.300 0 31.102.1 i4 $38.6 i9 0.03504084 i $950.373 $28,922 i $950,373 $28.922 0.03043226 a $946,353 $27,967 a $946.353 $27,967 0.0255524 the e-v frontier (npv verus standard dewatlon) the e-v rronticr (npv verus coefflclent of varlatlon) n f 1.357.970 $i 10,815 0.08160294 f t i .353,6 io $110,l76 00813942 p t 1,344,s io $105,833 007871492 j s i .203,295 $92, i47 0.07657889 0 $1.199.131 $91,539 0.07633778 tl $1.1 15,582 $41,743 0.03741814 e t i, i i 1,526 $40,978 0.03696643 0 s1.102.114 $38.619 0.03504084 t $950,373 628,922 0 03043226 a $946,353 s27.967 0.0295524 l. (exemption trust will, mortgage, and gifts) to d (exemption trust will and gifts); m. (simple will, mortgage, gifts, and whole life) to e (simple will, gifts, and whole life); n. (exemption trust will, mortgage, gifts, and whole life) to f (exemp tion trust will, gifts, and whole life); financial services review, 3(l) 1993 table 7. e-v analysis-male and female, both age 70-various estate plans-5% discounted at 5% discounted at 5 % n f p h l d j 8 tl e 0 g k c i a net present standard alue to helrs deviation f 1,142,989 f l,i39,353 s l,i33,679 s l,i30,043 si‘l14.497 sl,l 10,861 s i ,0 12,875 s i ,009,409 s939.846 $936.473 s930.536 $927,162 sqi 1,355 s907,980 s800,718 $76,430 s76,499 s97,873 s98, i33 $40,767 $40,373 s63,020 s63, i09 s34,690 s35,238 s68,862 $69,453 s24,817 524,635 $25,653 6797,375 s26,296 the e-v frontier (npv verus standard deviatton) s i, 142,989 s76,430 s i, i 14,497 s40,767 sl,l io.861 s40,373 s 1 ,o 12,875 1663,020 s939,846 s34,690 691 1,355 f24,8 i7 s907,980 s24,635 h p g 0 f n 0 j e n l d a i k c n l d k c net present standard coefficient atue to heirs deviation of variatton s 1,130,043 $1,133,679 3927.162 s930,536 b 1 ‘i 39,353 f i (142,989 $ i ,009,409 s i ,o 12,875 5936,473 s939,846 s 1.1 14,497 s i, i io,86 i $797,375 $800,718 $9 i 1,355 s98,133 0.08684006 597,873 0.0863322 s69,453 0.07490924 s68,862 0.0740025 1 s76,499 0.067 14249 576,430 0.06686854 s63,109 0.06252074 s63,020 0.0622 1893 335,238 0.03762842 s34,690 0.0369 103 s40,767 0.03657803 s40,373 0.03634308 s26,296 0.0329782 i s25,653 0.0320375 s24,8 17 0.02723088 s907,980 s24,635 0.027 13 165 the e-v frontier fnpv verus coefficient of variation) $ i ( 142.989 s76,430 0.06686854 t i, 1 i 4,497 s40,767 0.03657883 6 i, i io,86 1 540,373 0.03634388 sq 1 1,355 s24,8 t 7 0.02723088 s907.980 s24.635 0.02713165 0. (simple will, mortgage, gifts, and term life) to g (simple will, gifts, and term life); and p. (exemption trust will, mortgage, gifts, and term life) to h (exemp tion trust will, gifts, and term life); the table 5 comparative analysis demonstrates that i to a is 100 to 0, that j to b is 100 to 0, . . . and that p to h is 100 to 0. l@icieti frontier in inmate piam& 19 table 8. e-v analysis-male and female, both age 78-various estate w% ~is~~~ed at 4% ~is~~~~ed at 4 % net present standard net present standard coefficient value to heirs deviation value to heirs deviation of varlatlon n $966,829 $93,049 g $786,728 $102,002 0.12965345 f $963,790 $93,348 0 $789,542 f 10 t ,582 0.1286594 p $960,640 t 1 19,936 h t957,60 i $120,253 0.12557735 n $957,601 s 120,253 p $960,640 $119,936 0.1248501 l $941,241 $50,893 f $963,790 $93,348 0.09685512 0 f938,202 $51,173 n s966.829 $93,049 0.09624 i42 j $856,823 $80,107 i3 $853,924 $80,395 0.09414772 t) $853.924 $80,395 j $856,823 $80, io7 0.09349306 n $795,736 $72,524 e $792,917 $72.956 0.09200963 e $792,917 $72,956 n $795,736 $72,524 0.09 i 14078 cl $789,542 $101.582 a $675.185 $61,092 0.09048187 g $786,728 t 102,002 i $677,976 $60,639 0.08944 122 k $770.149 $45,892 c $767,330 $46.245 0.06026742 c $767,330 $46.245 k $770,149 $45.892 0.05958847 i $677.976 $60,639 0 $938,202 $5 1,173 0.05454369 a $675,185 $6 1,092 t $941,24f $50,893 0.05407011 the e-v frontler the e-v frontier (npv verus standard devlatlon) (npv verus coefficient of variation) it is critical for estate planners to determine whether or not their assumptions caused a plan to fall on the efficient frontier or whether the plan fell on the frontier on its own merits. in addition to lookiig at the plans which are on the efficient frontier, it is necessary for estate planners to look at the plans which fall just off that frontier (like estate plan d), and to analyze the input data, assumptions, and other factors that may cause a plan to just fall off of the efficient frontier. 20 fecal servkes rjxview, 3(l) 1993 0.19 0.1875 0.185 0.1825 0.18 0.1775 0.175 0.1725 0.17 0.1675 0,165 coefficient 0.1625 ol 0.16 vartatfon 0.1575 0.155 0. i525 0.15 0.1475 0.145 0.1425 0.14 0.1375 0.135 0.1325 0.13 0.1275 l k n j f-l c d p f e ii 0 a 6 940 ioo0 1060 1120 1180 1240 1300 1360 1420 1480 1540 1600 npvh (.ogq) figure 3. the efficient frontier (bold letters) and ten other estate plans for age 4.5 at 4% when two or more plans result in almost similar outcomes, then sensitivity analysis should be done to determine which of the nearly identical plans is the most stable (least affected by changes in assumptions and/or random variations in rates of return or borrowing rates). in the d to l comparison, d has the advantage that it is an unleveraged estate, and, as such, is not subject to interest rate v~iations, an important consideration for an elderly couple. consider the age 45 efficient frontier discounted at 5%. the order of the plans on that frontier are: n. exemption trust will, ira, mortgage, gifts, and whole life; p. exemption trust will, ira, mortgage, gifts, and term life; f. exemption trust will, gifts, and whole life; h. exemption trust will, gifts, and term life; 0. simple will, ira, mortgage, gifts, and term life; g. simple will, gifts, and term life; makes sense. tiaa whole life insurance, given the excellent risk characteristics of the tiaa group, should outperform commercial term life insurance, at least on an emtent frontiers in estate planning 21 i’wvh basis {and possibly on an eyv basis as well); that is, there should exist an inherent bias in favor of tiaa whole life insurance over commercial term life insurance. since iras and mortgages, especially where the mortgage has an assumed favorable leverage, serve to increase the wealth of the estate owners, estate plans that involve mortgages and iras should outperform those that do not. ceteris paribus, the exemption trust will at age 45 is absolutely superior to the simple will. if one looks at the age 45, coefficient of variation, 5% discounted-efficient frontier, the four plans that fall at the top of the frontier (n, p, f and h) are rank ordered first on the type of will, second on the use of tax shelter and leverage, and third on the basis of the type of insurance. the last two plans that form the bottom of the frontier, estate plans 0 ~si~p~e wi~z, ira, mortgage, gifts, and term life) and g (simple wizl, gifts, and term life) should, on a npvh basis, be inferior to m (simple wizz, ira, mortgage, gifts, and whole life) and e (simple wizl, gifts, and whole life). while in table 1 that is true, as m outperforms 0,9 1:9, and e outperforms g, 9 1:9, on the efficient frontier m is eliminated by f (exemption trust will, gifts, and whole life) and e is eliminated by 0 (exemption trust will, ira, mortgage, gifts and term life). then, because the npvhs of g and 0 are respectively about $44,000 lower than e and m, and because their standard deviations are about $26,000 lower, g and 0 qualify for the efficient frontier, albeit they lie on the lowest portion of that frontier. see figure 4. the timing of cash flows to heir(s), coupled with varying discount rates, can affect the nature of plans falling on and lying off the efficient frontier. while the npvh rank order of the 16 plans did not change as the discount rate changed from 4% to 6%, the rank order of their coefficients of variation did. (see the top left rectangle of tables 2,3, and 4, and tables 6,7, and 8 for npvh rankings, and the top right rectangle of the same tables for coefficient of variation rankings.) plans which were on the frontier dropped off, and plans which were not on the frontier appeared. for example, consider the age 45 couple simulation. on a coefficient of variation basis discounted at 4%, plan g (simple will, gifts, and term life) is the lowest plan on the efficient frontier. while plan e (simple will, gifts, and whole life) has a higher npvh, it has a considerably higher coefficient of variation, and plan e, eliminated by plan h (exemption trust wilt, gifts, and term life), is not on the efficient frontier. at 6%, g fell off of that frontier, and the lowest plan on the efficient frontier is plan e (simple will, gifts, and whole life). at age 70, at a 5% and/or 6% discount rate, estate plans l (exemption trust will, mortgage, and gifts) and/or l and k (simple will, mortgage, and gifts) are on the efficient frontier. at 4%, neither l nor k fall on that frontier. the efficient frontier, when defined on a coefficient of variation basis, is very sensitive to the discount rate of the heir(s). the estate planner must carefully consider the ages of her/his clients when reco~ending estate planning tools. for example, consider the simple wiz~ and gifts estate plan and the exemption trust will estate plan. for the age 45 couple, financial services review, 3(l) 1993 p n. y ” w &ecien# frontiers in wate planning 23 0.1039 0.1014 0.0989 0.0964 0.0939 0.0914 0.0809 0.0864 0.0039 0.00 14 0.0739 0.0764 0.0739 0.07 14 0.0689 coefflclent 0.0664 d 0.0639 varlatlon 0.06 14 0.0589 0.0564 0.0539 0.05 14 0.0489 0.0464 0.0439 0.0414 0.0389 0.0364 0.0339 0.03 14 0.0289 0.0264 0.0239 b j p n k c g d a n 0 i f l 433 466 500 533 566 600 633 666 700 733 npvh (,ooo) figure 5. the efficient frontier (bold letters) and ten other estate plans for age 45 at 6% 87 of the 100 npvhs (table l--comparative analysis) of the simpk will and gifts estate plan were superior to the exemption trust will estate plan. however, for the age 70 couple, 84 of 100 npvhs (table 5-comparative analysis) of the exemp tion trust will estate plan were superior to the simple will and gifts estate plan. extrapolating from these results, somewhere in between the ages of 45 and 70 there exists an age where the split would be about 50-50, and the simulation results would predict that both plans would be equally likely to produce similar results for the heir(s) of the estate owner(s). [since the ~rnpfio~ trust will and gifts estate plan includes both gifts during the liietimes of the estate owners as well as a $600,000 transfer at death, the ewmption trust will and gifts estate plan is preferable to either the simple will and gifts estate plan or the exemption trust will estate plan; on a comparative analysis npvh basis (tables 1 and 5), the numbers 0. 08 i6 0. 08 06 0. 07 96 0. 07 86 0. 07 76 0. 07 66 0. 07 56 0. 07 46 0. 07 36 0. 07 26 0. 07 16 0. 07 06 0. 06 96 co ef fic ie nt 0. 06 86 o f 0. 06 76 v ar ia tio n 0. 06 66 0. 03 75 0. 03 65 0. 03 55 0. 03 45 0. 03 35 0. 05 25 0. 03 i5 0. 03 05 0. 02 95 0. 02 85 0. 02 75 0. 02 65 0. 02 55 0. 02 45 fi gu re 6 . th e e f ie nt f ro nt ie r (b ol d l et te rs ) an d si x o th er e st at e pl an s fo r a ge 7 0 at 4 % f n ck dl hp b j fi e g o a i 94 6 98 1 10 16 10 51 10 86 11 21 11 56 ii9 1 12 26 12 61 12 96 13 31 13 66 14 01 i4 36 np vh f.o o o ) 25 0.0868 0.0848 0.0828 0.0808 0.0788 0.0768 0.0748 0.0728 0.0708 0.0688 0.0668 0.0648 0 0628 0.0608 0.0588 coefficient 0 0568 of 0.0548 variation 0 0528 0.0508 0.0488 0.0468 0.0446 0.0428 0.0408 0.0388 0.0368 0.0348 0.0328 0.0308 0.0288 0.0268 0.0248 0 0228 hp go fn bj em dl al ck 800 840 880 920 960 1000 1040 1080 1120 1160 npvh woo) figure 7. the efficient frontier (bold letters) and eleven other estate pians for age 70 at 5% are 10&o for the exemption trust will and gifts estate plan relative to the other two estate plans.] finally, note that at age 45 estate plan n (exemption trust will, ira, mortgage, gifts, and whole life) was almost a comer-point solution on the table 1 comparative analysis, recording 10010 npvh results against all other estate plans except for h (exemption trust will, gifts, and term life) and p (exemption trust will, ira, mortgage, gifts, and term life) where the results were 99: 1 for n to h, and 91:9 for n to p. for both the estate planner and the estate owner, n seems to be a logical choice. however, at age 70, the estate owners and heir(s), when trying to choose between n and p, have a difficult task, as does the estate planner who has that couple as clients. on a comparative analysis basis (table 5), the n to h results were 53:47, financial services review, 3(l) 1993 y ” effieient frontiers in estate planning 27 and the n to p results were 51:49. at the 4% discount rate, both n and p were on the efficient frontier, h was eliminated by p in the fifth digit of the coefficient of variation, and n is no longer a clearly logical choice. iii. conclusions the configuration of the efficient frontier is a function of three factors: the assump tions made by the estate owners (or made for them by an estate planner), the ages of the estate owners, and the discount rate(s) applied to the transferred wealth. estate owners (and estate planners) engaged in probabilistic estate planning must be certain to address the sensitivity that this process has to those three factors, and to probe the nature of not only the plans that fall on the efficient frontier but also the nature of those plans lying close to the frontier. as people age and the probabilities of death increase, probabilistic estate planning becomes more difficult. the near corner-point solution at age 45 becomes indeterminate at age 70. unlike point estimate life expectancy estate planning, where risk is ignored and an “optimal” estate plan is easy to find, when the discount rate is allowed to vary, and when ages at death, rates of return on assets, and crowing rates on debts are treated as random variables, estate planning decisions become harder to make. while harder to make, those decisions allow the estate owner(s) and the estate planner to address the riskiness of the alternate estate plans, and provide a realism to the estate planning process. if used by an unbiased estate planner, then the probabilistic estate planning process can assist the individu~ estate owner(s) in choosing an efficient estate plan to maximize the expected net present value of wealth transferred to heir(s) consistent with the risk preferences of that (those) heir(s). crabb, ronald r. 1992. “probabilistic estate planning,” financial services review, 1, 143-i 57. markowitz, harry m. 1952. “portfolio selection,” journal of finance (march): 312,324. pii: s1057-0810(97)90002-0 financial services review, 6(4): 227-242 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. an analysis of the tradeoff between tax deferred earnings in iras and preferential capital gains terry l. crain jeffrey r. austin this paper extends prior research in evaluating the decision of whether to invest in a mutualfund either outright or through one of the three available iras: the deductible ira, the roth ira, and the nondeductible ira. we provide mathematical models for after-tax accumulations for each of the investments that are a function of return, the percentage of the return currently taxable to the investor, the time horizon of the invest ment, the capital gain tax rate, and the ordinary income tax rate. the roth ira and the deductible ira always dominate investments in the nondeductible ira or through out right investment. however, in comparing the nondeductible ira and outright investments, the outcome is dependent on the investment goals of the mutualfind and whether it generates substantial dividend distributions or capital gain distributions. mutualfunds with small dividend and capital gain distributions may accumulate larger amounts if held outright while mutualfunds that pay substantial dividends or make sub stantial capital gain distributions accumulate larger after-tax amounts when invested in a nondeductible ira. i. introduction today’s financial markets provide investors with numerous investment alternatives. not only are there many specific investments to choose from, but the u.s. income tax system adds another layer of complexity to the investment decision. with the passage of the tax payer relief act of 1997, investors now have three iras to invest in for their retirement. these are (1) deductible iras, (2) roth iras, and (3) nondeductible iras. deductible iras and roth iras are always preferable to nondeductible iras for taxpayers who qual ify for them. however, deductible iras and roth iras are both subject to adjusted gross income (agi) limitations, leaving nondeductible iras as the iras available for upper income individuals covered by another retirement plan. terry l. crain and jeffrey r. austin l school of accounting, michael f. price college of business, university of oklahoma, 307 west brooks, room 200, norman, ok 730194450; e-mail: tcrain@ou.edu. 228 financial services review 6(4) 1997 while earnings from nondeductible iras accumulate tax deferred, later distributions are taxed at ordinary income tax rates. recent increases in the ordinary income tax rate and decreases in the capital gain tax rate have raised the question of whether individuals should invest in a nondeductible ira or make an outright investment in a mutual fund. in this paper, we hold the specific investment (a mutual fund) constant and examine the amounts an investor may accumulate in after-tax dollars in a deductible ira, roth ira, nondeduct ible ira, and in an outright investment in a mutual fund. the mathematical solution for the decision between the deductible ira and the roth ira is straightforward. due to greater tax incentives for the roth ira and the deductible ira, the decision between either the of those investments and the nondeductible ira is obvious. however, the comparison between the nondeductible ira and the outright investment in a mutual fund is very com plex. therefore, the majority of our analysis is on the choice between nondeductible iras or outright investments. our equations may be modified to allow pre-tax comparisons of the alternatives. our analysis is based on a marginal income tax rate of 31% for ordinary income, the marginal tax rate in effect for 1998 for individuals likely to invest in nondeductible iras, and 20% for capital gains. we provide mathematical models for after-tax accumulations for each of the investments that are a function of return, the percentage of the return cur rently taxable to the investor, the time horizon of the investment, the capital gain tax rate, and the ordinary income tax rate. we find that deductible iras and roth iras are always preferable to nondeductible iras and outright investments in mutual funds. the choice between nondeductible iras and outright investments is not as clear cut and depends on the percent of return currently taxable and the investor’s marginal tax rate. our results show that nondeductible iras are preferred when a larger portion of the investments return is distributed as dividend or capital gain while the outright investment in a mutual fund is preferred when dividends and capital gain distributions are a smaller percentage of the total return. the remainder of the paper is organized as follows. the next section discusses the income taxation of iras. we review the ira decision literature in section iii. in section iv, we develop mathematical models to compare investments in deductible iras, roth iras, nondeductible iras, and an outright investment in a mutual fund. we then apply our model to a sample of mutual funds to determine which mutual funds would be more appro priate for a nondeductible ira or an outright investment. the final section provides our conclusions and limitations. ii. income taxation of iras individual retirement accounts (iras) were introduced in 1974 as a means of encourag ing individuals to save for retirement. in order to accomplish this objective, the income tax law allows a deduction for the initial contribution to a deductible ira and the earnings accumulate tax deferred. when the individual withdraws funds from a deductible ira, he or she has to include the distribution in ordinary income. however, to discourage individ uals from withdrawing the funds before retirement, congress added a 10% penalty on amounts withdrawn before the individual attains age 59 l/2. an analysis of the tradeofl 229 characteristic table 1 requirements for ira investments deductible ira nondeductible ira roth ira initial contribution withdrawal of earnings withdrawal of contributions to ira limits on contributor income limitation on contributor (1998) married filing jointly single deductible in computing taxable income taxable as ordinary income taxable as ordinary income income limit if covered under another retirement plan no deduction if adjusted gross income exceeds $6o,ockl and a phase out if adjusted gross income is between $50,000 and $60,000 no deduction if adjusted gross income exceeds $40,000 and a phase out if adjusted gross income is between $30,000 and $40,000 not deductible in computing taxable income taxable as ordinary income nontaxable return of investment no limitations no limitations no limitations not deductible in computing taxable income excluded from taxable income nontaxable return of investment income limit if covered under another retirement plan cannot contribute if adjusted gross income exceeds $160,000 and a phase out if adjusted gross income is between $150,000 and $160,000 cannot contribute if adjusted gross income exceeds $110,000 and a phase out if adjusted gross income is between $95,000 and $110,000 prior to 198 1, individuals who were covered by other retirement plans could not par ticipate in iras. however, in 1981 congress relaxed the participation rules to allow indi viduals who were covered by other retirement plans to also participate in iras. this relaxation of the participation rules spurred investment in iras. u.s. internal revenue service (1984) shows that contributions to iras increased over 250% from 198 1 to 1982. in 1986, congress again changed the participation rules to disallow an income tax deduction for individuals covered by other retirement plans whose income exceeded $35,000 for single taxpayers and $50,000 for married taxpayers filing joint returns. how ever, congress introduced a new type of ira, the nondeductible ira, which retains the deferral of earnings but does not allow the investor an income tax deduction. since the investor is not allowed a deduction for the investment in the nondeductible ira, only the withdrawal of earnings is subject to income taxation and a portion of each withdrawal is treated as a tax free return of capital. the taxpayer relief act of 1997 added the roth ira, a retirement account where no deduction is allowed for the contribution to the account, but all earnings are excluded from income tax. the roth ira is available to married individuals filing joint returns with adjusted gross income less than $160,000 and single individuals with adjusted gross 230 financial services review 6(4) 1997 income less than $110,000. the ira alternatives and qualifications for 1998 are shown in table 1. each taxpayer is allowed to contribute $2,000 per year to an ira. generally, taxpayers are better off contributing to deductible iras or roth iras if they qualify. however, indi viduals covered under another retirement plan who exceed the deductible ira income lev els as shown in table 1, may not contribute to a deductible ira. such persons may qualify for a roth ira, since the income limits for roth iras are substantially higher than for deductible iras. individuals covered under a retirement plan who exceed the income level for roth iras ($110,000 single or $160,000 joint), may make contributions to a nonde ductible ira. the sum of all contributions to iras for a given year cannot exceed $2,000. therefore, if an individual’s annual investment in a roth ira is limited to, say $1,500, he or she may make an additional $500 investment in a nondeductible ira. the literature that examines the ira investment decision is discussed in the next section. iii. literature review several previous studies have examined various aspects of the ira investment decision. burgess and madeo (1980) and o’neil, saftner and dillaway (1983) use simulations to examine the break-even point for a deductible ira by considering the 10% premature withdrawal penalty. burgess and madeo’s study is for the tax law in effect prior to 1981, while o’neil, saftner, and dillaway’s study is an extension to the law in effect after the 1981 tax law changes. both studies find that, for long-term investments, deductible iras perform better than non-ira investments. however, if the deductible ira is held only for a short time period, the penalty for premature withdrawal causes deductible iras to not perform as well as non-ira investments. owens and willinger (1985) also use a simulation to examine deductible ira invest ments. however, instead of a break-even analysis, they compute internal rates of return to evaluate the ira investment decision. they find that an investment in a deductible ira has a greater internal rate of return, even in the short term, as long as the penalty for premature withdrawal is not imposed. yaari and fabozzi (1985) investigate the use of growth stocks in deductible iras. while their analysis suggests that iras should be invested in non-growth stocks, they point out that iras tend to invest in growth stocks. simonds (1986) examines the situation for an investor who wishes to have a balanced portfolio of interest-earning assets and aggressive-growth mutual funds. he finds that the interest-earning assets should be in a deductible ira and the aggressive-growth funds should be held outright. however, if the investor wishes to hold only aggressive-growth mutual funds, he or she should do so in a deductible ira. the aforementioned studies all examine the issue of investing in a deductible ira. however, in 1986 the federal income tax law was changed to allow nondeductible iras for persons who do not qualify for deductible iras. randolph (1994) examines the issue of investing in a nondeductible ira and finds that the nondeductible ira (invested in a mutual fund) dominates the outright investment in a mutual fund. however, in his study no distinction was made between ordinary income tax rates and capital gain tax rates. an am&s of the tradeof 231 scholes and wolfson (1992, pp. 34-40) discuss after-tax accumulations for two types of investments. the earnings on the first investment accumulate tax deferred until maturity of the investment. once the funds are withdrawn, the earnings are taxed at the ordinary income tax rate (similar to a nondeductible ira). the earnings on the second investment are taxed annually at the capital gain tax rate. scholes and wolfson conclude that either investment may accumulate a larger amount depending on the length of time the invest ment is held and the percentage of the investment return that is taxed at the ordinary income tax rate. the primary difference between randolph’s analysis and scholes and wolfson’s analysis is that &holes and wolfson allow for capital gain to be taxed at a lower rate than is ordinary income. the taxpayer relief act of 1997 introduced the roth ira, which provides for no tax deduction but allows exclusion for all income. steuerle (1997) provides an example showing that accumulations in a deductible ira and a roth ira appear to be equal if the investor is in the same tax bracket when the funds are invested and when the funds are withdrawn. however, he points out that the same $2,000 contribution limit applies to both deductible iras and roth iras. the investor in a deductible ira owns only one minus his or her tax rate times $2,000 while the investor in the roth ira owns the full $2,000 investment. there fore, the roth ira may generate a larger after-tax accumulation than the deductible ira. in ranking the three types of ira investments, the following conclusions may be drawn. first, the tax-favored deductible ira and roth ira will accumulate larger after-tax amounts than either the nondeductible ira or the outright investment. second, the deduct ible ira will accumulate a larger after-tax amount than the roth ira if the individual is in a lower tax bracket after retirement. on the other hand, if the investor is in a higher tax bracket after retirement, the roth ira will accumulate a larger amount of after-tax funds than will the deductible ira. if the taxpayer is in the same tax bracket before and after retirement, the accumulations in the roth ira and the deductible ira will be the same. the mathematical verification is provided in section iv. an individual who qualifies for either a deductible ira or a roth ira should invest in the deductible ira or roth ira instead of investing in a nondeductible ira. however, individuals who are covered under another retirement plan, and whose agi exceed the lim its shown in table 1, cannot contribute to either a deductible ira or a roth ira. these individuals may invest in a nondeductible ira. the issue of whether to make a contribu tion to an nondeductible ira is a very challenging one. depending on the assumptions made, an investment in a nondeductible ira, which still provides tax deferral of income, might be advisable. however, when funds are withdrawn from the nondeductible ira, all income is taxed at ordinary income rates, even when the nondeductible ira is invested in mutual funds. as an alternative to the nondeductible ira, the investor might invest out right in a mutual fund. while the investor must pay tax annually on dividend income and capital gain distributions, unrecognized gains are deferred and qualify for more favorable long-term capital gain tax rates when recognized later. in discussing the effect of the 1997 reduction in the capital gain tax rates on investments in mutual funds, brush (1997) points out that the greater the difference in long-term capital gain tax rates and ordinary income tax rates, the less likely a nondeductible ira will out perform an outright investment in a mutual fund. in general, a narrowing of the difference between the ordinary income tax rates and the long-term capital gain tax rates favors an investment in a nondeductible ira while the widening of the difference between the ordi nary income tax rates and the long-term capital gain tax rates favors an outright investment. 232 financial services review 6(4) 1997 iv. development of mathematical model in this section, we develop mathematical models to compare investments in deductible iras, roth iras, nondeductible iras, and outright investment in mutual funds. for our analysis, we assume that the investor does not dispose of the investment in an ira (deduct ible, roth, or nondeductible) before he or she reaches age 59 l/2, thus there will be no pen alty for premature withdrawal. the after-tax accumulation is a function of the return, the percentage of the return that is currently taxed, the ordinary income tax rate, the capital gain tax rate, and the number of years that the investment will be held. accumulation,, = f (r, p, to, tcg, n ) (1) all variables are defined as follows: a,, = after-tax accumulation apt = pre-tax accumulation r = return (appreciation, dividends, and capital gain distributions) po = percentage of return that is a dividend distribution and is taxed at the ordinary income tax rate peg = percentage of return that is a capita gain distribution and is taxed at the capital gain tax rate t0 = ordinary income tax rate t =g = capital gain tax rate n = number of years the investment will be held a. comparison of deductible ira and roth ira an investment in a deductible ira allows the investor an income tax deduction for the investment plus deferral of income tax on the earnings. an individual who wishes to make the maximum contribution to a deductible ira has a net investment in the deductible ira of one minus his marginal tax rate times the amount of the investment. for example, a tax payer in a 3 1% income tax bracket may contribute $2,000 to a deductible ira. however, he or she would have an immediate tax savings of $620. therefore, we assume that the investor funds $2,000 of investment in the deductible ira with the $620 in tax savings and $1,380 in funds from other sources. in developing our formula we grow the investor’s net investment of $1,380 divided by one minus his or her marginal tax rate (to gross the invest ment to the $2,000 that the investor contributed) for a period of n years at a rate of r. the deductible ira grows as follows: apt=(l+r)“/( l-t,) (2) when funds are withdrawn from a deductible ira, income tax is assessed against the withdrawal. if the individual is in the same ordinary income tax bracket when the funds are withdrawn as when the investment in the deductible ira is made (to at the time of invest ment equals to when the funds are withdrawn), the after-tax accumulation in the deductible ira is a,,=(l-t,)((l+r)“l(l-t,)) (3) an analysis of the tradeoff 233 which reduces to aat= (1 + r)” (4) if to at the time of withdrawal is less than to at the time of the investment, then aat will be greater than (1 + r)“. on the other hand, if r, at the time of withdrawal is greater than t, at the time of the investment, then a,, will be less than (1 + r)“. next, we compute the accumulation in the roth ira. the initial investment in a roth is not deductible when made by the investor. however, the earnings of the roth ira are excludable from taxable income. therefore, after-tax accumulation in the ira is aqr=(l +# (5) the after-tax accumulation of the roth ira is not dependent on the investor’s income tax rate, but remains constant at (1 + t)~. however, as shown above, the change in the investor’s ordinary income tax rate affects the accumulation in the deductible ira. if the investor’s ordinary income tax rate is greater when the funds are withdrawn from the deductible ira than it was when the investment in the ira was made, the accumulation will be less than (1 + r)“. on the other hand, if the investor’s ordinary income tax rate is less when the funds are withdrawn from the deductible ira than when the in~es~ent in the ira was made, the accumulation will be greater (1 + t)~. b. comparison of nondeductible ira with deductible ira and roth ira for the nondeductible ira, all earnings are tax deferred and an investment of one for a period of n years results in a pre-tax accumulation as shown in equation (6). a,,=(1 +r)” (6) all returns are tax deferred and all income is taxed at the ordinary income tax rate (1,) when the money is withdrawn from the nondeductible ira. therefore, the after-tax accu mulation in a nondeductible ira is a ar= 1 + ((1 -to) [(l + r)n 1]} (7) for all it, > 0, the after-tax accumulation in a nondeductible ira is less than the accu mulation in either a deductible ira (equation 4) or a roth ira (equation 5). therefore, individuals who qualify for either a deductible ira or a roth ira should invest in either one before investing in a nondeductible ira. c. comparison of outright investment in mutual fund with nondeductible ira as shown in table 1, many higher income individuals do not qualify for either a deductible ira or a roth ira. these individuals may select between a nondeductible ira, which does provide tax deferral on earnings, or they may select to invest in a non tax-shel tered manner. randolph (1994) compares nondeductible iras (invested in mutual funds) and outright investments in mutual funds and finds that nondeductible iras outperform 234 financial services review 6(4) 1997 outright investments. however, randolph assumes that the ordinary income tax rate and the capital gain tax rate are equal. recent changes in the tax law have increased the spread between the ordinary tax rate and the capital gain tax rate. we extend randolph (1994) by examining the impact of differentially taxing ordinary income and long-term capital gain on the decision to invest in a mutual fund either outright or through a nondeductible ira. consider a mutual fund that invests only in growth stocks that pay no dividends and the mutual fund makes no capital gain distributions. regardless of whether the investment in the mutual fund is outright or through a nondeductible ira, the investment grows to the same amount after n years. however, when the funds are withdrawn from the mutual fund, the outright investment is afforded capital gain treatment, while the nondeductible ira is subject to the higher ordinary income tax. therefore, in this situation, the outright invest ment in the mutual fund would provide greater after-tax funds. next consider a mutual fund that pays a dividend and has no appreciation. an individ ual who makes an outright investment pays tax on the dividend currently at ordinary rates and reinvests the after-tax amount in the mutual fund. however, with a nondeductible ira the individual defers the tax and reinvests the entire dividend. therefore, the nondeductible ira grows to a larger amount after n years. due to the difference in the taxation of a non deductible ira and an outright investment in a mutual fund, the after-tax accumulations will differ. section 85 1 (b) of the internal revenue code requires a mutual fund to distribute div idends and capital gains to avoid having an income tax assessed against the mutual fund. therefore, an individual who makes an outright investment in a mutual fund must pay tax annually on his or her share of the fund’s dividend and capital gain distributions. short term capital gains, interest income, and dividend income are distributed to the investor as dividends, and taxed at the ordinary income tax rate. capital gain distributions, which may be the result of either the mutual fund manager selling off a stock which he or she no longer wishes to hold or may result from rebalancing the portfolio, are taxed to the investor as long-term capital gains. for the outright investment in the mutual fund, we assume that all dividends and cap ital gain distributions are paid out in cash. the investor pays the income tax on the divi dends at the ordinary income tax rate (t,) and on the capital gain distributions at the capital gain tax rate (t,,), and reinvests the after-tax amount in the mutual fund. after n years the accumulation (after tax on distributions but before tax on liquidations) is shown in equation (8). a,, = [cl + d rp,t, rp&gln when the mutual fund shares are sold, the investor pays tax at the long-term capital gain tax rate (tc,) on the gain. we determine the gain by subtracting the adjusted basis in the shares in the mutual fund from the sales price. the adjusted basis is composed of the initial investment (which is 1) plus the dividend distributions and capital gains distribu tions less the income tax on those distributions, as shown in equation (9). adtusted his = 1 + [(rp,)(l $)i u1 (1 + r rp,4, ~p&j”] / [rp,t, + rpcgtcg 41 + krp,>(l &)i { [ 1 (1 + r rpoto rp,,tjl / [rp,t, + vcgtcg 41 (9) an analysis of the tradeoff 235 next, the capital gain tax is determined by subtracting the adjusted basis as determined in equation (9) from the pre-capital gain tax accumulation in equation (8) and multiplying by tcg to get the capital gain tax as shown in equation (10). capital gain tax = tcg{ [(l + r> rp,t, rp,t,,l” { 1 + krp,)(l $)i { [ 1 (1 + r rp,t, rp,,t,jl / [rp,t, + rp,,t,, rl) + krp,)(l t,,>l u1 (1 + r rwo rpcgtcg~“l / [rp,t, + rpcgfcg 41 (10) the after-tax accumulation in the outright investment in the mutual fund is determined by subtracting the capital gain tax in equation (10) from the pre-capital gain tax accumula tion in equation (8). this result is shown in equation (11). aat = [( 1 + r) rp,t, rp,,t,,i” t,,{[(l + r> rp,t, rp&,l” { 1 + [(rp,)(l to>1 w (1 + r rp,t, rp,4jl/ [rpob + rp&, rl} + [(rp,,>(l @i (11 (1 + r rp,t, rpcgt,jl 1 [rp,t, + rpcgtccg rl) (11) to compare the nondeductible ira with the outright investment, we set equation (7) equal to equation (11). with this condition, the after-tax accumulations of the two altema tives will be equal. the investor may substitute his or her marginal ordinary income tax rate for to and marginal long-term capital gain tax rate for tcg. in selecting to for our analy sis, we consider that individuals who qualify for one of the more tax-favored iras (either the deductible ira or the roth ira) will not invest in a nondeductible ira. as shown in table 1, single individuals who are covered by another retirement plan qualify for a roth ira unless their adjusted gross income exceeds $110,000 ($160,000 for married individu als filing jointly). after reducing the adjusted gross income limit of $110,000 for single taxpayers ($160,000 for married taxpayers) by exemptions and either the standard deduc tion or itemized deductions, the resulting taxable income is in the 3 1% bracket. therefore, weuset,=.31. most investors will have a long-term capital gain tax rate of 20% for 1998. while the taxpayer relief act of 1997 provides for a reduction in the long-term capital gain rate to 18% for tax years beginning after december 3 1,2000, there is no certainty that this sched uled reduction will take place. therefore, we use 20% in our analysis. the reduction of the rate to 18% would lower the indifference points. the investor may select n equal to the number of years until he or she wishes to liquidate the investment and r, the return rate of the investment. we demonstrate the use of the formula for r = 12%, n = 20 years, to = 3 l%, and tcs = 20%. for a given r, n, t,, and tc,, there are two unknowns, po, which is the portion of the annual return that is a dividend and taxed at ordinary income tax rates, and peg, the portion of the annual return that is a long-term capital gain distribution and taxed at the more favor able long-term capital gain tax rates. with two unknowns we cannot solve for a unique p. and peg_ instead, we have to fix a value for either p. or peg and solve for the other variable. we select p. = .20, that is, 20% of the 12% return for the year is distributed and taxed at the ordinary income tax rates, to demonstrate our formula. substituting these values into equation (1 l), and setting equation (11) equal to equation (7), yields equation (12), a poly nomial with one unknown, p,__. 236 financial services review 6(4) 1997 6.965942 = [1.11256 .024p,j2’ .20[1.11256 .024p,j2’ {l+ .01656 { [ 1 (i. 11256 .024p,)*‘] / [.024pcg . 112561) + [.096p,] {[l -(1.11256.024p,)20]/[.024p,r-.11256]} (12) we use m&hematica, a microcomputer program by wolfram (1996), to compute the numerical solutions for peg for the polynomial equation. since the equation is of the 21st degree, there are 21 solutions for peg. however, only one of the solutions, 0.123688, is in the relevant range between zero and one. where peg 0.123688, the accumulations in the nondeductible ira and the outright investment in the mutual fund are equal. if peg > 0.123688, the nondeductible ira accumulates a larger amount and if peg < 0.123688, the outright investment accumulates a larger amount. using equation (1 l), we may compare after-tax accumulations in outright investments in mutual funds with investments in mutual funds through a nondeductible ira. in order to further examine the issue, we select annual rates of return from .ol to .20 and time periods from five years to 40 years, in five-year increments. we hold p0 constant at 0.07 (which is approximately equal to the mean p. for growth funds computed in our analysis of a sample of mutual funds in section v) and solve forpcg. as shown by the aster isks in table 2, there are no indifference points for peg for a five-year time horizon. if p. = 0.07, and the investment is held for five years, the investor will accumulate a greater amount in the outright investment than in the nondeductible ira. two additional observations can be made from the information presented in table 2. first, the higher the rate of return, the lower the indifference point. for example, given a table 2 selected returns (r) and time horizons (n) for indifference point for percentage of investment currently taxed as capital gain with ordinary income (p,) = .07 outright investment in mutual fund vs. nondeductible ira investment horizon (in years) r 5 10 15 20 25 30 .ol * * * * * * .02 * * * * * * .03 * * * * * * .04 * * * * * * .05 * * * * .905998 .705428 .06 * * * * .718072 .552278 .07 * * * .804866 .585683 .444888 .08 * * * .678566 .487860 .365908 .09 * * .877317 .581400 .412962 .305711 .10 * * .770245 .504579 .354011 .258537 .i1 * * .6833 11 .442486 .306573 .22073 1 .i2 * * .611445 .391385 .267702 .i89866 .13 * * .551139 .348694 .235359 .i64272 .14 * .910578 .449890 .3 12572 .208097 .i42761 .i5 * .839303 a45861 .281670 .i84857 .i24471 .i6 * .777226 .4 17679 .254980 .i64850 .i08758 .i7 * .7227 12 .384291 .23 1732 .i47474 .095136 .i8 * .674490 .354883 .211331 .i32265 .083229 .19 * .631559 .328810 .i93307 .i18859 .072746 .20 * .593116 .305559 177287 .i06967 .063453 now * no indifference point outright investment accumulates a greater after-tax amount. 35 40 * * * * * .92 1503 .762527 .632603 .5657 12 .463357 .437307 .353464 .347692 ,277 123 .282084 .221470 .232293 .179394 .193424 146654 .162380 .i20575 .137111 .099390 .i 16207 .08 1892 098673 .06723 1 .083787 .054790 .071012 a44117 .059945 .034870 .050276 .026788 .041763 .019668 .034216 .013352 an analysis of the tradeoff 237 table 3 selected returns (r) and time horizons (n) for indifference point for percentage of investment currently taxed as capital gain with ordinary income (p,) = .20 outright investment in mutual fund vs. nondeductible ira investment horizon (in years) i5 10 15 20 25 30 35 40 .ol * * * * * * * * .02 * * * * * * * * .03 * * * * * * .73465 i .582689 .04 * * * * * .598008 .44505 1 .332488 .05 * * * * .569444 .395644 .274580 i85887 .06 * * * .629372 .406662 .262981 .i63343 ii90679 .07 * * * .48 1958 .291975 .i69944 .085698 .024526 .08 * * .66 i742 .372629 .207222 .i01509 .028842 & .09 * * .544928 .288350 .i42324 .049343 & & io * * ,452 148 .22 i784 .091238 .008457 & & .i i * * .3768 i4 i67975 .050 i 24 & & & .i2 * .7297 io .314533 i23688 .016431 & & & .i3 * .645804 .262268 .086687 & & & & .i4 * .574192 .2 i7848 .055375 & & & & .i5 * .5 i2402 .i79685 .028587 & & & & .i6 * .458584 146587 005488 & & & & .i7 * .411321 .i 17644 & & & & & .i8 * .3695 i3 ,092 i48 & & & & & .i9 * .332289 .069543 & & & & & .20 * .298956 .049383 & & & & & noipt: * no indtfference point outright mvestment accumulates a greater after-tax amount. & no indifference point nondeductible ira accumulates a greater after-tax amount. 20-year investment horizon, a mutual fund with a seven percent annual rate of return would have to pay out more than 0.804866 of its annual return as a capital gain before the nonde ductible ira would accumulate a greater after-tax amount than the outright investment. on the other hand, the value of peg at the indifference point for a mutual fund with a 20-year investment horizon and a 20% annual rate of return decreases to 0.177287. secondly, the compounding of returns for longer time periods favors the deferral provided by a nonde ductible ira. for example, holding the return constant at 0.12, the value ofpcg at the indif ference point for an investment horizon of 15 years is 0.611445, while the value ofp,, at the indifference point for an investment horizon of forty years decreases to 0.099390. next, we select p, = 0.20 (which is approximately equal to the mean p. for growth and income funds computed in our analysis of a sample of mutual funds in section v). we compute pcb for earnings rates from 0.01 to 0.20 and for time horizons from five years to 40 years, in five-year increments. indifference points for p. = 0.20 are shown in table 3. we use an asterisk to indicate earnings rates and time horizons where no indifference point exists and the outright investment accumulates a larger after-tax amount. the indifference points for p0 = 0.20 (growth and income funds) are attained at lower values of r and n than they were for p. = 0.07 (growth funds) due to the larger tax drag on the increased dividend income. combinations of high earnings and long time horizons (e.g., r = 0.19 and n = 30) result in the nondeductible ira generating the larger after-tax accumulation. we denote cases where the nondeductible ira generates the larger after-tax accumulation by an ampersand (&) sign. 238 financial services review 6(4) 1997 v. analysis of mutual funds in this section, we examine actual returns and distributions from mutual funds to determine which mutual funds would be more suitable for an outright investment versus an invest ment through a nondeductible ira. our previous analysis suggests that, for a given r and n, an investment through a nondeductible ira would be more beneficial if the mutual fund pays out large dividend distributions or capital gain distributions. an outright investment is more advantageous if the mutual fund does not pay out large distributions (dividends or capital gains). the american association of individual investors (1997) classifies mutual funds as aggressive growth, growth, growth and income, and balanced, based on the objective of the mutual fund. aggressive growth funds generally invest in stocks that pay little or no divi dends. these funds tend to stay fully invested in stock, may use financial leverage, and carry a greater amount of risk than other mutual funds. growth mutual funds also invest primarily in growth stocks that pay little or no dividends. however, they generally do not use leverage and are less risky than aggressive growth funds. the third category of mutual funds, growth and income funds, generally invests in stocks that pay cash dividends. their objective is to provide some income and some growth for investors. finally, balanced funds consist of investments in dividend paying stocks and interest paying bonds. there fore, if the outright investment is to outperform the nondeductible ira, it will most likely occur for aggressive growth funds, and will least likely occur for balanced funds. for our sample, we use a random number generator to select ten low-load mutual funds with at least five years of data from each of the categories designated as aggressive growth funds, growth funds, growth and income funds, and balanced funds by american association of individual investors (1997). we do not select bond funds since the goal of the bond fund is to generate interest income which favors the nondeductible ira over the outright investment. we use equation (7) to determine after-tax accumulations for investments in mutual funds through a nondeductible ira and equation (11) to determine after-tax accumulations for outright investments in mutual funds. as shown in tables 2 and 3, results differ for dif ferent investment time horizons. most individuals first qualify for an ira in their early to mid twenties, when they have approximately 40 years until they begin to withdraw their funds from the ira. individuals may continue to invest in iras until the time of retire ment. therefore, we select 20 years since it is in the center of the investment horizon (from zero years to 40 years). we use five-year annual returns for each of the sample mutual funds as shown in american association of individual investors (1997) as our return (r). for each mutual fund, we estimate p0 and peg by decomposing returns into dividend distributions, capital gain distributions, and unrealized appreciation over the five-year period from 1992 through 1996. for example, assume a mutual fund has a five-year return of 12% with a dividend of one dollar, capital gains distributions of two dollars, and appreciation of seven dollars. the total return of $10 includes $1 of dividends, $2 of capital gain distributions, and $7 of unre alized appreciation. therefore,p, = $l/$lo x .12 = 0.012 andpcg = $2/$10x .12 = 0.024. in tables 4 through 7, we provide lists of the randomly-selected mutual funds. table 4 lists the aggressive growth funds, table 5 the growth funds, table 6 the growth and income funds, and table 7 the balanced funds. for each mutual fund, we compute the after tax accumulation of one dollar invested outright in the mutual fund and the after-tax accu an analysis of the tradeoff 239 table 4 accumulation of outright investment vs. nondeductible ira invested in aggressive growth funds aggressive growth funds r po pc8 outright ira fidelity select air 0.1200 0.0000 0.2707 7.4168 6.9659 transportation* fidelity select automotive* 0.1700 0.0214 0.2994 16.4035 16.2529 git equity trust special 0.0890 0.0794 1.8833 3.4111 4.1067 growth invesco emerging growth 0.1670 0.0068 0.5372 14.4655 15.4549 invesco strategic portfolio 0.0710 0.0086 0.7806 3.0808 3.0304 health science* janus twenty 0.1130 0.1941 0.5894 5.6496 6.1815 janus venture 0.1100 0.1910 0.6133 5.3836 5.8730 legg mason special 0.1440 0.0117 0.2283 11.0946 10.4811 investment trust* sit mid cap growth 0.1140 0.0067 0.7257 6.0566 6.2880 wasatch aggressive equity* 0.1270 0.0000 0.4225 8.0112 7.8492 mean 0.1225 0.0520 0.6350 8.0973 8.2484 note: * outright investment exceeds nondeductible ira table 5 accumulation of outright investment vs. nondeductible ira invested in growth funds growth funds r po acorn investment trust: 0.1760 0.0487 acorn fidelity select insurance* 0.1710 0.0076 gradison mcdonald 0.1500 0.1031 established value ia1 regional 0.1170 0.0688 ia1 value 0.1340 0.0349 scout regional 0.0980 0.1818 sentry 0.1200 0.1058 sound shore 0.1860 0.0540 twentieth century heritage 0.1260 0.0498 vanguard primecap* 0.1800 0.0440 mean 0.1458 0.0699 note: * outright investment exceeds nondeductible ira pc, 0.3938 0.3183 0.3189 0.807 1 0.6322 0.3459 0.4449 0.5308 0.4848 0.1463 0.4423 outright ira 17.0372 17.9702 16.7133 16.5276 11.2298 11.6029 6.0531 6.6183 8.3062 8.8432 4.7268 4.7861 6.7613 6.9659 18.6919 21.2291 7.5645 7.7 166 20.1269 19.2112 11.7211 12.1471 mulation of one dollar invested in a nondeductible ira. it is anticipated that, if the outright investment is to outperform the nondeductible ira, it will most likely occur for aggressive growth funds and growth funds. these funds have lower dividends that are subject to cur rent income taxation at the higher ordinary income tax rates. table 4 contains the results for the ten randomly-selected aggressive growth funds. while these funds generally have fairly low dividend distributions (mean p. = 0.0520), they often have significant capital gain distributions (mean peg = 0.6350). however, the lower tax rate on capital gain distributions does not provide as large a tax drag as the ordi nary tax on dividend distributions. the accumulation of the outright investment exceeds the accumulation in the nondeductible ira for five of the ten mutual funds. for the five 240 financial services review 6(4) 1997 table 6 accumulation of outright investment vs. nondeductible ira invested in growth and income funds growth and income funds r po dreyfus 0.0900 0.1996 fidelity convertible 0.1410 0.4200 securities harbor value 0.1370 0.2018 hotchkis & wiley equity 0.1490 0.2045 income schwab looo* 0.1450 0.1280 selected american shares 0.1420 0.0903 sit large cap growth 0.1250 0.0489 t. rowe price equity index* 0.1470 0.1621 usaa income stock 0.1260 0.4226 wpg growth & income 0.1420 0.167 1 mean 0.1344 0.2046 note: * outright investment exceeds nondeductible ira pc, 1.2332 0.3536 0.6086 0.3428 0.0000 0.648 1 0.469 i 0.0758 0.1720 0.6332 0.4536 outright ira 3.6386 4.1770 8.0365 9.9608 7.9806 9.3062 10.3054 11.4081 11.2814 10.6604 9.0479 10.1313 7.4788 7.5861 11.1446 11.0281 6.7669 7.7166 8.6940 10.1313 8.4375 9.2046 table 7 accumulation of outright investment vs. nondeductible ira invested in balanced funds balanced funds bb&k diversa dodge and cox balanced fidelity asset manager: growth greenspring maxus income strong asset allocation t. rowe price balanced value line income vanguard asset allocation vanguard wellington mean r 0.0890 0.1370 0.1440 0.1480 0.2600 0.3184 9.8867 11.2165 0.0720 0.9572 0.0963 2.5865 3.0817 0.0940 0.4596 0.5722 3.8833 4.4710 0.1140 0.3508 0.1904 5.8195 6.2880 0.0930 0.3628 0.7866 3.8326 4.3956 0.1330 0.3040 0.2384 7.8637 8.6939 0.1340 0.3390 0.1393 8.0460 8.8432 0.1 158 0.3793 0.3070 6.4763 7.0884 po 0.3177 0.2964 0.1453 pcr 0.4023 0.0822 0.2440 outright ira 3.8869 4.1067 8.7673 9.3062 10.1909 10.4811 funds where the outright investment exceeds the nondeductible ira, the mean p. is 0.0083 and the mean pcr is 0.4003. for the remaining five funds, where the nondeductible ira accumulation is greater, the mean p. is 0.0956 and the mean peg is 0.8698. in table 5, we show the accumulations for investments in ten randomly-selected growth mutual funds. the mean dividend distribution (p, = 0.0699) is slightly larger than for the aggressive growth funds. however, the mean capital gain distribution (pcg = 0.4423) is considerably smaller than for the aggressive growth funds. the accumulation in the outright investment exceeds that in the nondeductible ira in only two of the ten funds. for the two funds where the outright investment exceeds the nondeductible ira, the mean p. is 0.0258 and the mean peg is 0.2323. for the eight funds, where the nondeductible ira accumulation is greater, the mean p. is 0.0809 and the mean peg is 0.4948. next we examine ten randomly-selected growth and income mutual funds. the mean dividend distribution (p, = 0.2046) is much larger than for the aggressive growth funds or for growth funds. as shown in table 6, the outright investment exceeds the nondeductible an analysis of the tradeoff 241 ira for only two of the ten mutual funds. for the two funds where the outright investment exceeds the nondeductible ira, the mean p. is 0.1454 and the mean peg is 0.0379. for the remaining eight funds where the nondeductible ira accumulation is greater, the mean p. is 0.2194 and the meanpcg is 0.5575. finally, as shown in table 7, the accumulation in the outright investment does not exceed the accumulation in the nondeductible ira for any of the balanced mutual funds. this was expected since a large portion of the returns for balanced funds is dividend income that is taxed currently and at the ordinary income tax rates. the dividend distribu tions (mean po = 0.3793) for this group of funds is much higher than for the other three groups of funds we examined. also, the capital gain distributions (mean peg = 0.3070) remains fairly high. vi. conclusions and limitations the taxpayer relief act of 1997 created the new roth ira, adding another retirement investment alternative. we compare the roth ira to the deductible ira and conclude that where an individual remains in the same ordinary income tax bracket when the investment is withdrawn as when the investment is made, the two types of iras will accumulate like amounts. on the other hand, an individual who expects to be in a lower tax bracket after retirement will accumulate a larger after-tax amount through a deductible ira while an individual who expects his or her tax bracket to decrease after retirement will accumulate a larger after-tax amount in a roth ira. however, it should be emphasized that while both types of iras allow a maximum annual investment of $2,000, the investor in a deductible ira has only a net investment of one minus his or her marginal tax rate times the invest ment. therefore, the roth ira generally allows an individual to shelter a greater amount. we also compare the tax-favored iras (roth ira and deductible ira) to a nonde ductible ira. the tax-favored iras allow an investor to accumulate a greater after-tax amount than does the nondeductible ira. however, many high income taxpayers do not qualify for either the deductible ira or the roth ira. therefore, the nondeductible ira remains a viable investment alternative for high income taxpayers. as an alternative to a nondeductible ira, an individual may make an outright investment in a mutual fund. the analysis to compare an outright investment in a mutual fund to a nondeductible ira (also invested in a mutual fund) is more difficult. randolph (1994) demonstrates that the nonde ductible ira outperforms the outright investment. however, he makes no distinction between ordinary income and capital gains. we extend his research to show that differen tial tax rates between ordinary income and capital gains affect the performance of the investments. mutual funds having little or no income distributions accumulate larger amounts when held outright instead of in a nondeductible ira. the greater the differential in the ordinary tax rate and capital gain tax rate, the greater the likelihood that this will occur. our research is very timely because congress has recently reduced the tax rate on long-term capital gains from 28% to 20% and the rate is scheduled to be decreased to 18% in 2001. investors who want to examine the effect of the change in the long-term capital gain tax rate may do so with our model which allows an investor to change any of the parameters, such as the long-term capital gain tax rate, and compare after-tax accumula tions of the two investments. 242 financial services review 6(4) 1997 we have presented a model that aids an investor in comparing an outright investment in a mutual fund with an investment in the same mutual fund through a nondeductible ira. the model incorporates the investor’s ordinary income tax rate and long-term capital gain tax rate. however, a limitation in tax planning is that future income tax rates cannot be pre dicted with certainty. also, the investor is uncertain about future dividend distributions, capital gain distributions, and per share values of mutual funds. therefore, the investor must make his or her decision based on imperfect information. references american association of individual investors. (1997). the individual investor’s guide to no-load mutualfinds (16th ed.). chicago, il: author. brush, m. (1997). with the clinton administration’s proposal, capital gains tax relief looks like a sure thing. money daily, (july 1). burgess, r.d., & madeo, s.a. (1980). a simulation study of tax sheltered retirement plans. journal of the american taxation association, i, 34-4 1. o’neil, c.j., saftner, d.v., & dillaway, m.p. (1983). premature withdrawals from individual retire ment accounts: a break-even analysis. journal of the american taxation association, 4, 35 43. owens, r.w., & willinger, g.l. (1985). investment potential of an individual retirement account. journal of bunk research, 16, 161-168. randolph, w.l. (1994). the impact of mutual fund distributions on after-tax returns. financial ser vices review, 3, 127-141. scholes, m.s., & wolfson, m.a. (1992). taxes and business strategy: a planning approach. engle wood cliffs, nj: prentice-hall. simonds, r.r. (1986). mutual fund strategies for ira investors. the journal of portfolio manage ment, i2,40-43. steuerle, g. (1997). beware of economic proclamations on back-loaded iras. tax notes, 76, 177% 1776. u.s. internal revenue service. (1984). selected statistical series, 1970-1984. washington, dc: u.s. department of the treasury, statistics of income division. wolfram, s. (1996). the muthematica book. champaign, il: wolfram research. yaari, u., & fabozzi, f.j. (1985). why ira and keough plans should avoid growth stocks. the jour nal of financial research, 8,203-215. pii: 1057-0810(92)90015-5 financial, services review, 2( 1): 5 1-61 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reprodudion in any form reserved. nobody gains from dollar cost averaging analytical, numerical and empirical results john r. knight lewis mandell dollar cost avemging is an investment system that is widely advocated by brokerage firms and mutual funds. in its best known form, an investor seeking to put a lump sum into risky assets is counseled to invest the money over a period of time in equal installments in or&r to avoid the devastating effect of a marketfall immediately afkr a single, lump-sum investment. using graphical analysis, historical stock market returns, and monte carlo simulations, this article demonstmtes that no such benefit accrues to a dollar cost averaging stmtegy. two alternative strategies, optimal rebalancing and buy and hold achieve better performance in all three analyses. introduction “dollar cost averaging” refers to an investment methodology in which a set dollar amount is placed in risky assets at equal intervals over a holding period. a number of rationales have been offered for this system in the investment literature (e.g. dodson [ 1989]), centering largely around the advantages of avoiding a poten tially unfortunate timing of a lump sum investment and the purported benefits of buying more shares at low prices and fewer shares at high prices. constantinides [1979] has shown that the inherent rigidity of dollar cost averaging makes it inferior to a constantly rebalanced portfolio on an a priori basis. offsetting this, however, may be some of the other presumed advantages of dollar cost averaging including its simplicity, its forced savings implications, and its possible reduction in transactions costs from not having to rebalance in each period. piros [ 19861 has quantified some of the differences in transactions costs. john r. knight l department of finance, university of connecticut, ct 06268-2041. lewis mandeu l department of finance, university of connecticut. 52 financialservicesreview,2(1) 199m993 the purpose of this paper is to demonstrate analytically, numerically and empirically the lack of any advantage accruing to dollar cost averaging relative to two other types of systematic investment strategies, rebalancing and buy and hold. the paper is organized as follows. first, we provide a graphical representation of the effect on investor utility of following each of the investment policies. next, we compare certainty-equivalent returns generated by the three systematic investment strategies in a series of numerical simulations. then, using historical stock market returns, we show the lower level of utility derived from dollar cost averaging as compared with alternative strategies. in all three instances, dollar cost averaging provides the least desirable performance. dollarcost averaging in evaluating the three types of systematic investing, we assume that investors have an initial stock of wealth invested in the riskless asset.’ we assume that they know the balance between risky (call it, for simplicity a diversified portfolio approximating the s&p 500) and riskless (short t-bill) assets that will optimize their utility given expected returns on both assets, the volatility of the risky asset, and the investor’s degree of risk aversion. to illustrate the three types of systematic investment, let us assume that an investor’s optimal balance is 50-50. the investor who practices optimal rebalanc ing would invest $50 thousand in risky assets immediately and would rebalance the portfolio at the end of each period to maintain exactly half of the portfolio in risky assets. the investor who practices the buy and hold strategy would put $50 thousand into risky assets immediately but would never rebalance in subsequent periods. in contrast, the investor who practices dollar cost averaging would take the initial wealth endowment and move it into the risky asset in equal increments over the holding period. using our example, the investor would put $5 thousand per year into the risky asset in each of the ten years. graphical analyses relative to the investor who uses optimal rebalancing both the investor using dollar cost averaging and the investor who buys and holds may be seen to experience utility loss, regardless of the degree of risk aversion. figure 1 illustrates the loss of the dollar cost average user. utility curve ubd represents the optimally balanced investor whose balance is at point j on the capital market line. an investor who is dollar cost averaging to reach an optimal balance of j over the holding period, currently has a balance of k. this investor’s utility curve is represented by u dca which, since it must pass through point k on the capital market line, is lower than utility curve u,,. if we hold variance constant at axa the loss in return can be measured as ae(r). the dollar cost averager suffers lost utility in two ways. first, nobody gains j+om do&r cost averaging figure i. the optimally balanced (bal) investor compared with the dollar cost averaging (dca) investor. e(t) t-r b&h cl= figrue2. the opti~ly balanced (sal) investor compared with the buy and hold (b&h) investor. 54 financial services review, 2(l) 19!32/1993 he sacrifices the higher return he could have achieved on average with a larger proportion invested in the risky asset. second, he suffers a utility loss from being “non-optimal” in terms of his investment balance over the entire investment horizon. in figure 2, the buy and hold investor u,,, may also be seen to be out of balance, but for a very different reason. the risk premium on the risky asset will cause the expected value of the risky asset to grow more rapidly than the riskless asset, leaving the initially balanced buy and hold investor with a continually growing proportion of the risky asset. the added return moves the balance above j to l and the utility curve that intersects the capital market line at l is ubah which is below u,, the buy and hold investor starts with a utility curve coincident with that of the optimally balanced investor, but over time moves to lower curves reflecting the lost utility from bearing more than his optimal level of risk. numerical and empirical comparisons we approach the problem of comparing the three investment strategies de scribed above with numerical simulations. within the simulation framework, we examine a wide range of attitudes toward risk by varying the risk aversion parameter y in the direct utility function wu/y. in our model, following that of merton [ 19691, wealth grows over the invest ment horizon with the following price dynamics: for the riskless asset, x: where r is the riskfree rate of return. for the risky asset, y: where p > r is the mean rate of return on the risky asset, u the standard deviation of that return, and &v(l) the increment of a wiener process. wealth dynamics then are a function of 1) the stochastic evolution of stock prices and 2) the proportion invested in risky assets. and dw(t) = w(f)[(p r)a + r]dt + w(f)[aadw(t)] where a is the proportion of total wealth invested in the risky asset.* our investor begins with an initial wealth w,,. with that initial wealth, the investor can employ three strategies. nobody gains jhm lb&r cost averaging 55 1. strategy 1 optimally balanced (bal). the investor selects the optimal balance determined by his or her degree of risk aversion. in each subsequent period, the resulting amount is rebalanced to correspond to the initial optimal balance. by rebalancing his optimal proportion each period, this investor effectively buys more of the risky asset when prices have fallen and less when prices have risen. here, his rationale is not one of market timing, but rather one of maintaining a utility maximizing propor tion of his total wealth in the risky asset. 2. strategy 2 buy and hold (b&h). the investor selects the optimal balance as in strategy 1 but does not subsequently rebalance, which means that he or she will go further and further out of balance as the risky assets grow more rapidly than the riskless asset due to the higher returns on the former. this strategy creates lost utility in the gradual departure from the optimal investment proportions. as such, investors at the two ends of the risk aversion continuum would experience the smallest costs from this departure. extremely risk averse investors optimally hold a very low proportion of their total wealth in risky assets, and this small percent age results in a very slow rate of departure from the “optimum.” at the other extreme, while departure from the “optimum” is at a fast rate, the limit, 100 percent in the risky asset, is not far from “optimal” so them is little potential for loss in utility even when the limit is reached. 3. strategy 3 dollar cost averaging (dca). the investor puts a proportion of the initial wealth into risky assets each period so that the optimal balance is achieved at the end of the time horizon. this investor is under-invested in risky assets over the entire horizon and therefore loses utility from the “non-optimal” allocation of his wealth. worse still, he forgoes the excess returns on the percentage of the risky asset that should be, but is not, in his investment portfolio, and he forgoes the utility that would be derived from the additional wealth.’ for our monte carlo simulation, we use familiar parameters extracted from historical new york stock exchange data. these data indicate an average return on stocks five percent greater than the risk free rate, and a standard deviation (our proxy for riskiness) of approximately 20 percent. the investor may choose any balance of risky to total assets from 10 to 90 percent that conforms to his coefficient of relative risk aversion (crra), (1 y). for example, investors with 10 percent in risky assets have a crra of 12.5 while those with 90 percent in risky assets have a crra of 1.3889. these coefficients are implied by the proportion of total wealth held in the risky asset and are shown at the bottom of the accompanying tables. for each time horizon, we measure the investor’s expected utility of wealth as the average of the utility from the wealth provided by each of 500 draws from our simulated stock market. having calculated the expected utility from following each 56 financial services review, 2(l) 1992/1993 table 1 certainty equivalent wealth for the three strategies investment proportion in risky asset by percentage horizon 10 20 30 40 50 60 70 80 90 a. certainty equivalent wealth for strategy 1 (optimal balancing) 2 1.0061 1.0087 1.0160 1.0242 1.0235 1.0347 3 1.0086 1.0143 1.0229 1.0264 1.0314 1.0452 4 1.0089 1.0150 1.0364 1.0416 1.0427 1.0622 5 1.0141 1.0209 1.0321 1.0528 1.0825 1.0724 6 1.0141 1.0444 1.0334 1.0639 1.0504 1.1022 7 1.0155 1.0345 1.0548 1.0831 1.0598 1.1220 8 1.0180 1.0440 1.0614 1.0704 1.1054 1.0935 9 1.0221 1.0551 1.0722 1.0821 1.1019 1.1375 10 1.0278 1.0525 1.0785 1.1142 1.1191 1.1907 b cer ,tainty equivalent wc :alth for str ategy 2 (buy & hold) 2 1.0060 1.0086 1.0165 1.0239 1.0242 1.0340 3 1.0085 1.0138 1.0223 1.0269 1.0316 1.0436 4 1.0083 1.0161 1.0368 1.0403 1.0434 1.0617 5 1.0134 1.0200 1.0311 1.0526 1.0820 1.0703 6 1.0140 1.0428 1.0320 1.0613 1.0480 1.0997 7 1.0152 1.0305 1.0495 1.0824 1.0588 1.1191 8 1.0177 1.0397 1.0568 1.0671 1.1036 1.0923 9 1.0192 1.0531 1.067s 1.0809 1.0943 1.1330 10 1.0254 1.0489 1.0739 1.1116 1.1168 1.1832 1.0350 i.0405 1.0431 1.0629 1.0768 1.0793 1.0912 1.0765 1 s978 1.1018 1.0819 1.1273 1.1129 1.0991 1.1252 1.1234 1.1629 1.1409 1.1531 1.1469 1.1544 1.1696 1.2426 1.2417 1.2132 1.2455 1.1881 1.0353 1.0403 1.0433 1.0613 1.0769 1.0789 1 ky24 1.0776 1 .o!j75 1.1029 1.0818 1.1266 1.1094 1.0983 1.1245 1.1220 1.1623 i.1409 1.1530 1.1416 1.1539 1.1680 1.2390 1.2411 1.2056 1.2436 1.1864 c. certainty equivalent wealth for strategy 3 (dollar cost averaging) 2 1.0049 1.0073 1.0129 1.0198 1.0200 1.0268 1.0287 3 1.0061 1.0098 1.0169 1.0200 1.0255 1.0316 1.0476 4 1.0060 1.0120 1.0264 1.0274 1.0335 1.0479 1.0624 5 l.oo!z 1.0174 1.0226 1.0380 1.0540 1.0461 1.0758 6 l.oo!x 1.0271 1.0242 1.0406 1.0356 1.0641 1.0708 7 1.0106 1.0234 1.0322 1.0506 1.0465 1.0737 1.0818 8 1.0125 1.0269 1.0397 1.0475 1.0676 1.0658 1.0957 9 1.0133 1.0329 1.0455 1.0551 1.0611 1.0829 1.0987 10 1.0151 1.0316 1.0471 1.0706 1.0725 1.1056 1.1261 note: implied coefficient of relative risk aversion 12.5000 6.2500 4.1667 3.1250 2.5000 2.0833 1.7857 1.0332 1.0561 1.0514 1.0600 1.0692 1.0986 1.0979 1.1385 1.1452 1.0386 1.0606 1.074s 1.0954 1.0872 1.0994 1.0976 1.1391 1.1257 1.5625 1.3889 of the three compared strategies, we can solve for the certainty equivalent wealth implied by the level of utility. u(w) = w/y * w = (yu(w))"l this analysis permits the direct comparison of wealth differences among strategies as shown in table 1. as anticipated, little difference can be seen between strategies 1 and 2, indicating that the gains to rebalancing each period are small if the investor nobody gainsjiom dollnr cost averaging 57 table 2 investment proportion in risky asset by percentage horizon 10 20 30 40 50 60 70 80 90 a. certainty equivalent wealth difference: strategy 1 (optimal balancing) strategy 2 (buy & hold) 2 .oool .oool -.ooos .0002 -.ooos .ooo7 -0003 .0003 -0002 3 .oool .0005 .ooo6 -.0005 -.oool .0016 .0016 -.oool .0004 4 .0005 -.ooll -.0004 .0013 -.0007 .ooos -.ooll -.ooll mm2 : .0007 .oool .0009 .0016 .oolo .0014 .0002 .0027 .0005 0024 .0021 jo25 -.ooll .0034 .oool .0008 .0007 .0007 7 .0002 .0041 .0052 0007 .oolo .@i28 .0014 .0006 0001 8 .0004 .0043 .0046 .0033 .oql8 .0012 .oool 0053 .0005 9 .0029 .0020 .0047 .0012 .0076 .0045 0016 .0036 0006 10 0024 .0035 .0046 ii027 0023 .0075 .0077 .0020 .0017 note: implied coefficient of relative risk aversion 12.50 6.25 4.17 3.13 2.50 2.08 1.79 1.56 1.39 b. certainty equivalent return from following strategy 1 (optimal balancing) vice strategy 2 (buy & hold) 2 .00004 00005 -.00023 .00012 -00038 .00033 -.oool7 .00013 -.ooolo 3 .00002 .00017 .00021 -.00016 -.00004 .00052 .00052 -.00003 .00014 4 .00013 -.00027 -.ocolo .00033 -.00017 .00013 -00028 -.00027 .00005 5 .oool3 .00018 .00019 .00004 .ooolo .00043 -00022 .00002 .00015 6 .ooool .00027 .oco23 .00045 .00039 .ckuml .ooo57 .00013 .oooll 7 mi003 .00058 .00075 00009 ii0014 .00040 .00020 .00008 .ooool 8 .00004 00054 .00058 .00041 .00022 .oool5 .ooool 00067 .00006 9 .ooo32 00023 ii0052 .00014 .00085 .00050 .00018 .00040 .00007 10 .00024 .00035 .00046 .00027 .00023 .ooo75 .oco76 .00020 .00017 note: implied coefficient of relative risk aversion 12.50 6.25 4.17 3.13 2.50 2.08 1.79 1.56 1.39 is initially in balance. this means that continual rebalancing yields little gain over the buy and hold strategy. however, the certainty equivalent wealth of the dollar cost averager in strategy three is consistently and substantially below that of the investors using the other two strategies. table 2a shows the certainty equivalent wealth difference between strategies one and two and table 2b shows the difference in annualized (geometric mean) return over all holding periods. the difference is tiny in all cases and, in the short run, will occasionally even go negative as a result of monte carlo variation acting upon small differences. with short time horizons, the investor does not venture too far from his “optimal” allocation of wealth. table 3a shows the certainty equivalent wealth difference between strategies one and three and points up a considerable loss associated with dollar cost averaging. while the cost of dollar cost averaging varies with the proportion invested in risky assets and the number of holding periods, with 90% invested in risky assets the optimally balanced investor has half again the certainty equivalent 58 financial services review, 2(l) 1!491/1993 table 3 investment horizon 10 20 proportion in risky asset by percentage 30 40 50 60 70 80 !?o a. certainty 4uiv~~t wealth diffeuxce: strategy 1 (~~mal balancing) s&ategy 3 (doilar cost averaging) 2 .ooll .0014 .0031 .0044 .0035 .0079 .0063 .0073 3 .oa24 .0044 .0060 .0064 .0059 .0136 .0153 _.0207 4 .oo29 .0030 .0101 .0142 .0092 .0142 .o288 .0251 : .0049 .0044 .0173 0035 .0095 .0092 .0147 .0233 .0285 .0148 .0381 .0263 .0420 .o260 so219 .0298 7 .0048 .0111 .0226 so325 .0134 .0483 .0416 .0642 8 .0055 .0171 .0217 .0228 .0378 .0278 *os73 .0490 9 .0088 .0223 .0267 .027 1 0409 .0546 .0710 .1041 10 .0127 .0208 .0314 .0437 .0466 .0851 .087 1 .1003 note: implied coefficient of relative risk aversion 12.50 6.25 4.17 3.13 2.50 2.08 1.79 1.56 b. certainty equivalent return from following strategy 1 (optimal balancing) vice strategy 3 (dollar cost averaging) 9 .00081 .00056 .00072 .00147 .00201 .00156 .00215 .00219 .00175 jo196 .00394 .00450 .00506 .00315 4 .00073 .00075 00250 ii0354 .00230 .00354 .00713 5 .00098 .00069 .00189 .00293 so0564 .00521 .00515 6 .00074 ~30286 .@i152 .oo385 .00245 .00625 xi0689 7 .00069 .00158 .00319 .0@+58 .00190 .00676 .@i584 8 .00069 .00212 00268 xxi283 .00465 .00343 .00699 9 mo98 .00245 .00293 .00297 mm46 xl0592 .00765 10 xxi127 .00206 .00309 .00428 .00457 ii0820 .@i839 note: implied coefticient of relative risk aversion 12.50 6.25 4.17 3.13 2.50 2.08 1.79 .00685 .00622 .00435 *00491 .00893 .oo600 .01106 .00961 1.56 1.39 a046 .0187 .0232 .0319 .0380 .0415 .0568 .1026 .0624 1.39 .00229 .00620 do576 .00630 .00623 .00583 do693 .01092 .00608 return of the dollar cost averager. and since the buy and hold investor does nearly as well as the optimal balancer, the dollar cost averaging strategy is also substan tially inferior to the buy and hold strategy. empirical analysis the theoretical and simulation results were tested empirically by using actual monthly returns from 1962 to 1992. dividend adjusted returns of the s&p 500 and treasury bill returns were used as proxies for the risky and riskless assets respec tively. three disparate investors were chosen for the analysis: one who was very risk averse, corresponding to the investor with 10 percent in risky assets in tables l-3; a 50-50 investor of middle risk aversion; and, a 90 percent risky asset investor who represents the not very risk averse. ten year rolling holding periods were begun nobody gainsfiom do&r cost averaging table 4 emplrkd resuits for the three strategies 59 optimal balancing buy and hold dollar cost a. high degree of risk aversion (10% risky assets) averaging mean annualized return .0731 .0729 (standard deviation) (.0179) (.01&q) mean utility -270.09 -270.29 (standard deviation) (10.15) (10.29) number of periods strategy yielded 110 highest return (percentage) (45.8%) (e.2, number of periods strategy yielded 114 61 highest utility (percentage) (47.5%) (25.4%) b. moderate degree of risk aversion (50% risky assets) mean annualized return xl837 .0828 (standard deviation) (.0296) (.0313) mean utility -266.38 -267.03 (standard deviation) (14.68) (15.15) number of periods strategy yielded 87 76 highest return (percentage) (36.3%) (31.7%) number of periods strategy yielded 120 44 highest utility (percentage) (50.0%) (18.3%) c. low degree of risk aversion (90% risky assets) mean annualized return (standard deviation) (:~~ (:e) mean utility -264.82 -265.07 (standard deviation) (20.49) (20.64) number of periods strategy yielded 84 64 highest return (percentage) (35.0%) (26.7%) number of periods strategy yielded 118 highest utility (percentage) (49.2%) $16) .0713 (.01x) -27 1.02 (9.83) 48 (20.0%) 65 (27.1%) .0762 (.0234) -269.75 (11.48) 77 (32.0%) 76 (31.7%) .0805 (.~%~ -268.69 (13.07) (e.38) 93 (38.7%) each month from 1962 to 1982, giving us a total of 240 holding periods. during each holding period, the optimal rebalancer rebalanced monthly and the dollar cost averager invested an equal amount of his initial stock of wealth in risky assets in each of the 120 monthly periods. the buy and hold investor began with the correct balance and never varied his proportion. table 4 summarizes tbe empirical results. over all levels of risk aversion, the dollar cost averaging system yielded the smallest annualized return and mean utility, although differences were not signifi cant. when the success of the three strategies was measured in terms of the number of holding periods (of the total of 240) in which the method yielded the best results, dollar cost averaging also fared poorly. for the very risk averse investor, dollar cost averaging produced the highest returns in only 20 percent of the cases. this increased to 32 percent and 38.3 percent for the moderate and not very risk averse 60 financial services review, 2(l) 1992/1993 investor. when measured in terms of utility, dollar cost averaging showed similar results. the very risk averse investor would have done better in 27.1 percent of the cases as compared with 31.7 percent and 38.7 percent for those investors with moderate or low risk aversion. brokerage fiis promote dollar cost averaging primarily with two rationales. first, they argue that returns are augmented because more shares are purchased when prices are low and fewer when prices are high. second, they assert that dollar cost averaging enhances investor utility by preventing an ill-timed lump sum invest ment. our results do not support either of these contentions. it is reasonable to assume that transactions costs would vary inversely with the size and directly with the frequency of investment. therefore, incorporating such costs into the model would serve to further strengthen the case against dollar cost averaging. one might also speculate that reducing the frequency of investment would improve the performance of the optimal rebalancing strategy relative to its alternatives. conclusion using three separate methods of comparison, we have shown the lack of any advantage of dollar cost averaging relative to two alternative investment strategies. our numerical simulations and empirical evidence, in consonance with our graphical analysis, both favor the optimal rebalancing and buy and hold strategies over dollar cost averaging. optimal rebalancing and buy and hold strategies convinc ingly outperform dollar cost averaging on theoretical grounds as well as on the basis of numerical simulations. historical evidence also supports these two strate gies, though the empirical differences are not significant. our results strongly imply that the additional cost and effort associated with dollar cost averaging cannot be justified for any investor, regardless of degree of risk aversion. with the possible exception of its promoters, nobody gains from dollar cost averaging. acknowledgements: we gratefully acknowledge many beneficial conver sations with tom o’brien on the subject of dollar cost averaging. notfs 1. thus, we take care not to confuse dollar cost averaging with periodic investing where the investor puts aside a regular amount of savings each period to invest in risky securities. periodic investment of savings as funds become available is a buy and hold strategy and has the additional positive attribute of encouraging habitual savings, a characteristic not shared by dollar cost averaging. 2. merton (1969) shows that, with costless and continous portfolio rebalancing possible, the investor with constant relative risk aversion has an optimal proportion of wealth (independent of the level of wealth) invested in the risky asset. this proportion is nobody gains from dollar cost averaging 61 where y < 1 is the investor’s coefficient of relative risk aversion. we use this analytically optimal proportion as the benchmark for each utility function analyzed, but recognize that portfolio rebalancing is neither costless nor continous. 3. since risky assets, on average, grow more rapidly than riskless assets, desired end of horizon results am also difficult to achieve with dollar cost averaging. to use theexample given earlier, a dollar cost averager with $100 thousand in assets who wished to achieve a xl-50 balance at the end of 10 periods might put $5 thousand per period into risky assets. however, over the ten periods the funds invested in risky assets would be expected to increase more rapidly than the funds put into riskless assets. therefore, on average, more than 50 percent of the portfolio would be in risky assets at the end of the ten periods. references constantinides, george m. 1979. “a note on the suboptimality of dollarcost averaging as an investment policy,” journal of financial and quantitative analysis,14(2): 443-450. dodson, lori. 1989. ‘“lhree dimensions of diversification,” journal of accountancy, 167(june): 131-139. merton, robert c. 1%9. “lifetime portfolio selection under uncertainty: the continuous-time case,” the review of economics and statistics, 5 1: 247-257. piros, christopher d. 1986. “how bad is dollar cost averaging?’ unpublished paper. pii: s1057-0810(99)00016-5 from the editor karen eilers lahey this last issue of volume 7 for 1998 includes a formal thank-you to the 65 individuals who are specially listed on the reviewers pages and those associate editors who graciously give of their time by reviewing several manuscripts a year. each manuscript is reviewed by at least two reviewers and sometimes three for the initial version of the paper and all subsequent revisions until the final acceptance of the article. kenneth s. bigel’s article entitled, “the correlations of professionalization and compensation sources with the ethical development of personal investment planners” is taken from dr. bigel’s dissertation. he begins by explaining the current state of the investment planning industry and then discusses the psychology of ethical development. the results of his research indicate that there are differences in ethical development between individuals who hold the certified financial planner (cfp) mark and those who do not hold it. he also finds differences based on age and gender. i would certainly like to encourage you to read his article and consider submitting manuscripts that add to our knowledge about ethics in individual financial management. robert p. goss, president of the certified financial planner board of standards, provides comments on bigel’s paper. a copy of the the cfp licensee standards, section i—code of ethics and professional responsibility and section iii—financial planning practice standards, is published as part of the comment to inform readers of the specify standards that cfps are required to adhere to in their professional activities. the full text can be found on the cfp website at (www.cfp-board.org/licensee/eadtoc-frame.set.html). a topic of interest to academics, financial planners, and parents is how to have sufficient funds available for college tuition when children are ready to attend school. judson w. russell and robert brooks categorize the prepaid state college tuition plans that have been developed since 1987 into three group that they call contract, tuition credit, and certificate plans. they apply a surplus framework model to managing the tuition risk faced by these state plans and suggest that it can be used by individuals in allocating their assets that are designated for college costs. andy saporoschenko tests the impact of dividend reinvestment plans (drips) on the karen eilers lahey, the university of akron, college of business administration, department of finance, akron, oh 44325-4803. financial services review 7 (1998) vii–viii 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00016-5 firm’s value and efficiency. he is concerned with the issue of the cheap source of outside financing that reinvested dividends provide to firms that have this type of plan. a description of drips is provided with an analysis of the industry distribution of this type of individual investor benefit and previous research. his results indicate that there is no statistical difference in the value of firms that have drips versus those that do not have the plans. larger firms are more likely to have drips, which supports his bookkeeping rationale to explain the existence of these plans. as the population in the united states increases in age, planning for retirement housing becomes a more critical issue for all investors. karen m. gibler, george p. mochis and euehun lee provide the results of a survey of individuals who are 55 years of age or older to determine their interest in retirement community housing. they provide readers with a review of seniors housing options that are currently available and their associated costs. their survey provides insight into the thoughts of those near retirement or already retired in terms of desired characteristics of retirement communities and why they would move to that type of housing. the last article by gordon j. alexander, jonathan d. jones, and peter j. nigro reports on a survey of 2,000 randomly selected mutual fund investors conducted for the office of the comptroller of the currency and the securities and exchange commission. the results strongly suggest that efforts at improving the financial literacy of mutual fund investors are needed. this is particularly important in light of the ever increasing number of individual investors who trade in financial markets today. viii k. eilers lahey / financial services review 7 (1998) vii–viii ce 1hour general principles of financial planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. cfp board recently adopted revisions to several provisions of its ce policies, including changing the minimum number of questions for self-study assessments from 10 to 5 per full ce credit hour. therefore, financial services review ce exams will have 5 questions. ce credit for this issue expires december 31, 2022, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in “the association between financial risk and retirement satisfaction” by pearson and guillemette, which of the following arguments was made: a. as retirement approaches, individuals should increase risk in response to a decrease in human capital. b. there is not a strong theoretical rationale for the reduction of risk during retirement. c. all retirees should bear additional portfolio risk for greater return potential in the future. d. retirees’ objective risk preferences should be considered more than their subjective risk preferences. 2. in pearson and guillemette, the authors posit that: a. increasing financial risk should lead to a reduction in retirement satisfaction because of the resulting increase in income uncertainty. b. decreasing financial risk should lead to an increase in retirement satisfaction because of the resulting increase in income certainty. c. financing consumption in retirement is dependent on the retiree’s level of human capital. d. higher risk comes with the potential for higher returns, and higher returns provide retirees with more income to finance their consumption in retirement. 3. pearson and guillemette argue that if retirees reduce the ratio of stocks held relative to total financial assets: a. it may decrease retirement satisfaction. b. it may increase retirement satisfaction. c. there should be an increase in retirement satisfaction only if the decrease in the ratio of stocks is offset by an increase in the ratio of bonds. d. there should be no effect on retirement satisfaction. 4. in “financial literacy, behavioural factors, and problematic debt-taking: an empirical analysis of relationships in australia” by tahir, richards, and ahmed which factor was identified as having the strongest relationship with credit card debttaking behaviour in the paper? a. financial literacy b. financial satisfaction c. attitude towards balancing spending and savings d. gender 5. in tahir, richards, and ahmed, the authors highlighted that ________ is relatively a more stressful and problematic type of household debt? a. mortgage b. credit card debt c. investment loan d. car loan manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) pii: s1057-0810(96)90002-5 fsr referees financial services review would like to thank the following ad hoc reviewers for providing their valuable and timely contributions to volume 5: michael alderson university of missouri at st. louis james j. angel georgetown university robert angel1 north carolina a & t state university peter bacon wright state university s.g. badrinath northeastern university vickie bajtelsmit colorado state university w. scott bauman northern illinois university roger bey university of tulsa jerry boswell metropolitan state college james boyd kent state university ronald braswell florida state university leroy brooks university of south carolina kitt butler michigan state university julie cagle xavier university richard calloway university of wisconsin at lacrosse don chance virginia polytechnic institute and state university saeyoung chang arizona state university s.j. chang illinois state university anthony cherin san diego state university t. michael clauretie university of nevada james co&ran georgia state university dan cooper marist college eurico ferreira indiana state university andrew fields university of delaware donald fischer university of texas at tyler george frankfurter louisiana state university erika gilbert illinois state university joseph golec clark university benton gup university of alabama sherman hanna ohio state university kendall hill university of alabama birmingham thomas howe illinois state university marc ingram clark atlantic university michael isimbabi morgan state university marlin jensen auburn university thomas johansen fort hays state university susan jordan university of missouri at columbia george kaufman loyola university dan klein bowling green state universrty linda s. klein university of connecticut david lange auburn university at montgomery james ligon university of alabama timothy loughran university of iowa d.k. malhotra philadephia college of textiles & sciences steven mann university of south carolina kyle mattson rochester institute of technology robert mcleod university of alabama william meggison university of georgia stuart michelson eastern illinois university donald nast florida state university theron nelson university of north dakota daniel pace valparaiso university jayen pate1 university of alabama at huntsville roy patin jr. midwestern state university glenn pettengill emporia state university dev prasad university of texas at san antonio william reichenstein baylor university frank reilly university of notre dame frederick seigel university of louisville keith smith purdue university laura t. starks university of texas william templeton butler university steven thorley brigham young university david wright university of wisconsin at parkside james yoder university of south alabama alan ziobrowski lander university brigette ziobrowski augusta college vii pii: 1057-0810(92)90004-v financial services review, 2(2): 87-96 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the individual investor in the market: forming a belief regarding market efficiency robert m. peevey gene c. uselton john r. moroney do the actions of investors drive the market toward efficiency or do investors utilize fads and other information unrelated to the true value of the security to drive the market away from eficiency? investors have been forced to examine a multitude of challenges to the eficient markets hypothesis in recent years. one of the most formidable of the challenges is the “excessive market price volatility” argument. we examine this argument, as presented in the “variance bounds” literature, and conclude that, although markets may be ineficient, the “variance bounds” literature has not proved the case conclusively. i. introduction for investors who want to “beat the market,” there is a shortage neither of data nor of advice. readily available are such primary data as annual reports, loks, and 1oqs. secondary data sources include the financial press (e.g., the wall street journal and investors daily); investor’s services such as the value line investment survey, standard & poor’s, and moody’s; and computer data bases that can be obtained on floppy disks. for advice investors can turn to wull street week or any number of newsletters published by advisors who claim to know what stock to buy and/or when to buy it. can investors reasonably expect to “beat the market” by superior application of this information? many, if not most, financial economists in academia believe that individuals should simply purchase a diversified portfolio of securities (or, perhaps, a well-managed, diversified mutual fund). they insist that efforts to “beat the market” robert m. peevey l department of finance, university of houston, houston, tx, 77204-6282; gene c. uselton l department of finance, texas a&m university, college station, tx, 77843-4218; john r. moroney l department of economics, texas a&m university, college station, tx, 77843-4218. 88 financial services review, 2(2) 1993 are destined to failure. their argument proceeds from the “efficient markets hypothesis” (emh). according to the emh, the prices of securities, at any given moment, reflect all of the information available to investors. suppose that information about a common stock were available which would justify paying a higher (or lower) price than the market price. what would happen? investors, eager to profit from the pricing error, would immediately purchase (sell) the mispriced security, thus driving its price up (down) to the price justified by the information available. numerous attempts over the past 25 years to test the emh have led to refinements. the emh now comes in three different forms: the weak form, the semi-strong form, and the strong form. it is important for investors to distinguish between these three forms. the weak form of the emh holds that the current price of any security is independent of the security’s price history. that is, knowing all of the prices at which a security has traded for every trade in the security’s history cannot help investors to predict any future price. if the weak form of the emh is correct, then technical analysis can be of no value to investors. the semi-strong form of the emh incorporates the assumptions of the weak form but adds more. it holds that all public information (including the series of prices at which a security has sold) is impounded in the market price. if so, reading annual reports, loks, loqs, etc., is a nonproductive activity; security analysis can be of no value to investors. the strong form goes one step further; it holds that all informdon (including insider information) is impounded in the prices of securities. under the strong form investors cannot obtain excessive profits even if they know about a buy-out before it is announced. a belief regarding the efficient markets hypothesis dictates investment prac tices. accepting the concept implies that investors should (1) form diversified portfolios of securities meeting their risk and income objectives and (2) trade securities only when objectives change, when excess funds are available, or when funds are needed. investors who reject the emh might reasonably attempt to (1) identify and purchase underpriced securities, (2) identify and sell overpriced issues, and (3) trade securities to benefit from market timing. it has long been held that security prices change only in response to new information. new information occurs randomly since any trend or prior knowledge would make the information predictable; therefore, it would not be new information. price changes therefore must be random. however, early tests implying a random walk in stock price changes (fama, 1965) have been modified by evidence of potential long-term price reversion (fama & french, 1988) and short-term correla tions (lo & ma&inlay, 1988). investors and security analysts gathering and using information in an attempt to “beat the market” would appear to drive markets toward efficiency. however, market efficiency anomalies, such as the p/e ratio effect, january effect, etc., focus the individual investor in the market 89 on specific “pockets” of potential inefficiency and provide evidence that markets may not be driven all the way to efficiency. in general, empirical tests are inconclu sive and, in some degree, contradictory.’ a basic premise in asset valuation is that an asset is worth the present value of its future cash flows. we refer to this as the true value of the asset. if the market price of a security is not consistent with its true value, investors must be responding to something other than pertinent information concerning the asset. fads (shiller, 1989), overreaction (debondt & thaler, 1985), or speculative bubbles (west, 1988) may explain security prices. if so, markets cannot be efficient. a test which proves that market prices are not consistent with the present value of their future cash flows would prove that the emh is false and end the market efficiency question once and for all. the “variance bounds” test is promoted as such a test. some scholars (e.g., shiller, 1989) performing the test find that the volatility of market prices is too great for the value of a security to equal its true value and conclude that markets are inefficient. do investors utilize “irrelevant” information to drive markets away from efficiency? we offer no answer to this question and draw no conclusion about the efficiency of financial markets. however, an examination of the foundation of the variance bounds test will prove useful to investors who are concerned about the issue. this paper focuses on the theory which asserts that the volatility of market prices is too great to be consistent with the efficient markets hypothesis. we briefly review the theory and examine its theoretical underpinnings. our objective is to expose weaknesses in the excessive volatility argument and to show that the variance bounds tests are not valid when risk is considered. we conclude that the individual should look with skepticism upon the excessive volatility argument when forming a belief concerning market efficiency. in section ii we explain the traditional variance bounds theory. section iii examines the theory and uncovers three flaws in the original argument. our summary and conclusions are presented in section iv. ii. variance bounds theory leroy (1989) explains that very low market prices in the mid-1970s stimulated researchers to reexamine the theory of asset pricing.* they argue that asset prices should have low volatility relative to the volatility of dividends; and, further, that these volatility implications apply directly to market efficiency theory. leroy and porter (1981), shiller (1979, 1981a, 1981b, 1989), and singleton (1980) lay the foundation for what is now known as variance bounds testing. their finding, that the variance of market prices exceeds that of a proxy for fundamental value,3 introduces a new challenge to market efficiency. 90 financial services review, 2(2) 1993 the formal argument shiller (1979, 1981a, 1981b), leroy and porter (1981), and singleton (1980) introduce the rationale that the fundamental value of a security is equal to the security’s market price plus a random error. this implies that the fundamental value of security i at time t, ~,, may be linearly decomposed into an expected component (i.e., market price), j+,, and an unexpected component, xi,. pt = pit + xit (1) equation (1) implies that the fundamental value is predicted by market price and disturbances. it necessarily follows that o;=cr;+0,+2cr,, (2) the market price must be uncorrelated with the random component as a condition of market efficiency. hence the variance of the time-series of fundamental values must be equal to the sum of the variances of the price series and the series of random errors: 0; = 0; + 0, (3) since variances are non-negative, the variance of the fundamental value series sets an upper bound on the variance of the price series. 0; 2 0; (4) inequality (4) is the cornerstone of first generation variance bounds tests. leroy and porter (1981) and shiller (1979, 1981a, and 1981b) find that the variance of market prices exceeds that of the fundamental value series, thus leading them to question the validity of the efficient markets hypothesis.4 many subsequent tests have been performed and, although many of the problems of the first-generation variance bounds tests have been overcome, primarily by improved econometric methods, the underlying theory has not been changed. it is this theoretical foundation that is examined here. iii. three flaws in the traditional variance bound argument the first flaw first, the variance bounds specification in (4) is less restrictive than it seems. zero correlation between pit and xi, clearly implies the relationship expressed in (3). equation (4) follows from (3), but the converse is not necessarily true. equation (4) does not guarantee zero correlation between pir and xi,. in fact, the relation expressed the individual investor in the market 91 in (4) implies a range of correlation between pit and xit, consider (2) and note that the variance bounds specified in (4) imply that o2 >21 ct2 p and, therefore, that (5) (6) where ppr is the correlation coefficient between pit and xi,. from (6) it follows that (4) actually identifies a range of values of correlation between pit and xit (7) that is, the variance of market price exceeds the variance of fundamental value if and only if (8) states the actual rejection region for market efficiency under the traditional variance bounds test5 the second flaw the second flaw is that traditional variance bounds testing is strongly influ enced by the empirical relationship between 7c, andp,. scott (1985, (2), (4), and (5)) and easton (1985, (7)) generalize (1) as 7c, = u + bpi, + uil (9) in performing regression-based variance bounds tests. utilizing (9), the ratio of variances is, where r2 is the coefficient of determination from the regression of (9). applying the traditional variance bounds criteria for market efficiency from (5) implies that 92 financial services review, 2(2) 1993 r21b2 (11) this has the unsettling implication that the proportion of fundamental value ex plained by market price must be smaller than some specific limit, otherwise we reject market efficiency. for example, if b2 = 0.5 markets are tested to be efficient if market price explains 10% of the variation in fundamental value and are tested to be inefficient if market price explains 90% of the variation in fundamental value. this is contrary to the intuitive concept that market price and fundamental value should be closely linked in an efficient market. equation (11) presents no problem if b 2 1; however, scott (1985, table 1, p. 602) finds that b is significantly less than one; indeed, he finds b not significantly different from zero. easton (1985) finds that b varies with the level of the discount rate used in calculating the proxy for fundamental value. the estimate of b is less than unity for the lower range of discount rates. scott’s and easton’s results cast serious doubt upon the validity of the linear assumption in ( 1).6 the third flaw the third flaw in the traditional variance bounds argument is that (1) is an inequality for any state other than risk neutrality. to examine this claim, we reconstruct the relationship. the relationship between the market price of a security and its fundamental value begins with the decomposition of the actual cash distributions, acfi,, into their components: the expected dividends, ec%, and unexpected dividends, ucfi, acfi, = ecfi, + ucfi, (12) the fundamental value, 7cil, can then be expressed as (13) where cpi is the consumer price index at the time of the distribution (t+j) and the beginning of the time period (t) and rr is the real rate of return which is assumed to be the constant (e.g., 3%) part of the time varying realized return. equation (13) presents the price that a rational investor would pay for a security with (1) full knowledge of all future cash (or equivalent) distributions, (2) perfect information concerning future inflation as accounted for by the cpi, and (3) a desire to achieve a realized real rate of return of rr. note that the certain cash flow is discounted using a rate having no risk premium. since it involves an infinite stream of distributions, we use the term “fundamental value.” we combine the inflation and real return factors into a time varying realized return, rt+j, which includes no risk premium. the individual investor in the market 93 (14) equations (12), (13), and (14) can be combined if and only if the assumption of risk neutrality is imposed. assuming a risk neutral world we have acf,, c m ecf,t+j m ucf,t+j .= j=l (1 + rt+j)' c -+c j=l (1 + rt+jy j=l (1 + rt+jy (15) the left hand side (lhs) variable, acf, is a certain cash flow and must be discounted with no risk premium. the right hand side (rhs) variables, ecf and ucf, are uncertain cash flows and can be discounted with no risk premium only under the risk neutrality assumption. however, the risk neutrality assumption is not attractive. the first term on the rhs is the discounted expected dividend stream, the market price. it is contrary to basic concepts of stock valuation to assume that there is no risk premium associated with this term. the relationship is a decompo sition and both sides must be discounted by the same factor to maintain equality (shleifer & summers, 1990). if we discount the rhs by a factor containing a risk premium, the lhs, the certain cash flow, must also be discounted using a risk premium. this practice is used in variance bounds theory and the pseudo value is called the “ex post rational” price. but, this is not theoretically correct and there is no assurance that an equality is established by applying some arbitrary risk premium to the lhs. for clarity let us examine the conditions under which (15) reduces to (1). investors price risky equities by discounting expected, but uncertain, cash distribu tions by a discount factor, ki. thus, the current market price is pit = c ecfij+j j=l ( 1 + k,t+j’ (16) the discount rate, kit, includes a risk premium and varies from firm to firm and over time as implied by the subscripts. under the assumption of risk neutrality, ki,,+j = r,+j and xit = ucfi,t+j c j=l (1 + rr+j)j ’ (17) then (15) reduces to (l), which implies that the fundamental value is a linear combination of the expected information effects, pit, and the unexpected effects, 94 f’inancial services review, 2(2) 1993 xif in a risk neutral world. but, there is no reason to believe that the same relationship holds in a risk averse world. a theoretically correct relationship under risk aversion can be established utilizing assumptions normally accepted in security valuation. without loss of generality, we treat both actual and expected dividend growth rates, (gj and (g&, respectively, as constants over the period t = l,...,~. thus, ecf,, = dit( 1 + gei)j (18) and acfi,t+j = di,( 1 + g,& similarly, ki,t+j and rt+j are treated as their effective geometric average values over time and designated henceforth as ki and r,., respectively. the sum of the infinite series created by substituting (18) into (16) converges (see gordon (1962)) to: pir = da1 + gd ki _ gei where (ki g&is strictly positive for any finite market price. in a similar manner, substituting (19) into (13) (or the equivalent lhs of (15)) yields the theoretical true (or fundamental) price n, = dixl + gai) it rfhi (21) where (rrgai) is strictly positive for any finite fundamental price.’ solving for di,, in (20) and substituting into (21) yields the relationship % = pir (i+ gaj(ki gd (1 + gsj(rf gai) 1 (22) the ratio in brackets is a certainty equivalent factor between the fundamental value (the discounted value of a stream of certain real cash distributions) and the market price (the discounted value of a stream of expected but uncertain cash distribu tions). in a risk averse world the relationship is nonlinear as shown by (22). it is obvious that this relationship does not produce the variance boundary that is produced by (1). we argue that (1) is an inequality in a risk averse world; thus, the variances that are derived from the relationship cannot be interpreted as valid bounds. the individual investor in the market 95 iv. stjmmary and conclusions this study shows that the theory underlying the excessive price volatility challenge to market efficiency is flawed. we have not shown that markets are efficient; but we have given investors reason to doubt the argument for market inefficiency based on the evidence for excessive price volatility. our conclusion is consistent with a statement by fama (1990), “as always, then, the answer to the basic market rationality question must be left to the reader.” notes 1. fama (1991) provides an excellent review. 2. the dow jones industrial average reached a modem period low in 1974. 3. researchers applying ex ante discount rates to realized prices and dividends refer to this proxy for the security’s fundamental value as the ‘&ex post rational price.” see kleidon (1986a) for a discussion of this term. 4. critics (flavin (1983), kleidon (1986a,b), and marsh and merton (1986)) quickly find problems with the early analyses. new tests intended to correct these problems are developed by mankiw, romer, and shapiro (1985), scott (1985), campbell and shiller (1988a,b), and leroy and parke (1990). these second-generation tests are criticized by shea (1989). articles by leroy (1989). west (1988), and gilles and leroy (1990) present excellent reviews of the variance bounds literature. 5. leroy and porter (1981, p. 561) apparently recognize that their variance bound specification implies some range of correlation between p and x, but they do not elaborate. their orthogonality test implicitly assumes a correlation of zero. froot (1987), in an unpublished paper, spells out the correlation range more clearly. frankel and stock (1987), comparing tests of efficiency of foreign exchange markets, make a similar point. our results are derived independently and presented in a different form. 6. we abstain from lengthy review of second generation tests because (a) that task is accomplished admirably by gilles and leroy (1990) and (b) those tests are not pertinent to our examination of the model. 7. economic rationality is lost if market price or fundamental price is allowed to assume an infinite value. references campbell, john y. and shiller, robert j. 1988a. ‘the dividend-price ratio and expectations of future dividends and discount factors,” review of financial studies, 1: 195-228. -. 1988b. “stock prices, earnings, and expected dividends,“journalof finance, 43(3): 661-76. debondt, werner f.m., and richard h. thaler. 1985. “does the stock market overreact,” journal of finance, 40(3): 793-805. easton, peter d. 1985. “accounting earnings and security valuation: empirical evidence of the fundamental links,” journal of accounting research, 23 supplt.: 54-77. fama, e.f. 1970. “efficient capital markets: a review of theory and empirical work,” journal of finance, 25(2): 383-416. 96 financial services review, 2(2) 1993 -. 1990. “stock returns, expected returns, and real activity,” journal of finance, 45(4): 1089-l 108. -. 1991. “efficient capital markets: ii,” journal of finance, 46(5): 1575-1617. and k.r. french. 1983. “permanent and temporary components of stock prices,” journal if political economy, 9 l(6): 929-56, frankel, jeffrey a. and james a. stock. 1987. “regression vs. volatility tests of the efficiency of foreign exchange markets,” journul of international money and finance, 6: 49-56. froot, kenneth a. 1987. tests of excess forecast volatility in the foreign exchange and stock markets. national bureau of economic research working paper #2362. gordon, marion j. 1962. the investment, financing and valuation of the corporation, homewood, illinois: irwin. gilles, christian and stephen f. leroy. 1991. “econometric aspects of the variance-bound tests,” review of financial studies, 4(4): 753-92. kleidon, allan w. 1986a. “variance bounds tests and stock price valuation models,” journal of political economy, 94(5): 953-1001. -. 1986b. “bias in small sample tests of stock price rationality,” journal of business, 59(2): 237-61. leroy, stephen f. 1989. “eflicient capital markets and martingales,” journal of economic litera ture, 27(4): 1583-1621. and richard d. porter. 1981. ‘the present-value relation: tests based on implied variance bounds,” econometrica, 49(3): 555-74. lo, andrew w. and a. craig ma&inlay. 1988. “stock market prices do not follow random walks: evidence from a simple specification test,” review of financial studies, l( 1): 41-66. mankiw, n.g., d. romer, and m. shapiro. 1985. “an unbiased reexamination of stock price volatility,” journal of finance 40(3): 677-687. marsh, terry a. and robert c. merton. 1986. “dividend variability and variance bounds tests for the rationality of stock market prices,” american economic review, 76(3): 483-98. scott, louis 0. 1985. ‘the present value model of stock prices: regression tests and monte carlo results,” review of economics and statistics, 67: 599-605. shea, gary s. 1989. “ex-post rational price approximations and the empirical reliability of the present-value relation,” journal of applied economettics, 4: 139-59. shiller, robert j. 1979. “the volatility of long-term interest rates and expectations models of the term structure,” journal of political economy, 87(6): 1190-1219. -. 1981 a. ‘the use of volatility measures in assessing market efficiency,” journal of finance, 36(2): 291-304. -. 1981b. “do stock prices move too much to be justified by subsequent changes in dividends?" american economics review, 71(3): 421-36. -. 1989. market volatility, cambridge, massachusetts: mit press. shleifer, andrei and lawrence h. summers. 1990. ‘the noise trader approach to finance,” journal of economic perspectives, 4(2): 19-33. singleton, kenneth j. 1980. “expectations models of the term structure and implied variance bounds,” journal of political economy, 88(6): 1159-76. west, kenneth d. 1988. “bubbles, fads and stock price volatility tests: a partial evaluation,” journal of finance, 43(3): 639-60. pii: 1057-0810(95)90012-8 from the editor entering its fourth year of publication, financial services review has tenaciously held to its editorial policy of publishing only high-quality research in the field of individual financial management. the dual criteria of academic rigor and focus on the individual have had the disadvantage, from an editor’s perspective, of limiting the number of manuscripts submitted for publication. however, the scope and caliber of articles that have been published in the first seven issues of this journal have set the foundation for an academically respectable subspecialization of finance. the five articles in the current issue continue to apply the tools of modern financial analysis to decisions faced by individuals and their advisors. some decisions are complex, such as the use of junior equity interests to save taxes when passing on a family business to the next generation. others appear to be mundane, such as the purchase of u.s. savings bonds, the true value of which can only be evaluated through option pricing analysis. many individuals own stock in banks. this is due to the fact that banks make up a sizeable proportion of publically traded corporations in the united states as well as to the local or regional appeal stemming from prohibitions on nationwide banking. to a great extent, the financial condition of banks depends upon the fair market value of their assets and liabilities, figures which need be reported only as a footnote to the annual financial statement. lacking this information on an ongoing basis, investors must seek other clues about the banks’ financial condition. in their article titled “bank dividend policy as a signal of bank quality,” boldin and leggett examine the extent to which the dividend policy of bank holding companies is a signal of their quality. the argument is that banking organiza tions in poor financial condition are unlikely to distribute large cash dividends. the study found a positive relationship between bank dividends per share and bank quality rating but also found an inverse relationship between the dividend payout ratio and bank quality. banks that paid out a large proportion of earnings in the form of dividends were left with little in the way of retained earnings to build up a capital position. hamill and steinberg examine a complex but important issue in their article titled “tax savings opportunities in estate freeze transactions: an application of the black scholes model.” since much of the growth in employment in the united states comes from smaller, often family-owned businesses, there is policy interest in smoothing the ownership succes sion of these businesses from one generation to the next. however, with federal estate taxes taking as much as half the value of assets passed to descendents, it is difficult to keep a business in the family. one method of doing so involves the use of two classes of stock. the older family members will recapitalize the company, retaining voting preferred equity for themselves and passing on the common equity to the next-generation managers, this transaction has the effect of freezing the value of the corporation at the time of recapitalization for estate tax vii . . . vu1 financial services review, 4( 1) 1995 purposes and passing on an option to the younger generation to acquire the preferred shares at this frozen price after the death of their elders. if the corporation has grown in value, this call option may become quite valuable. in a series of tax law changes in 1987 and 1990, congress tried to deal with the value of these options for estate tax purposes, first assigning a minimum value of 20% of the interests held by the senior generation family members and then lowering that to just 10%. the article shows how the black-scholes option pricing model can be used to value the junior interest created by the freeze and to determine whether a freeze, with the statutory minimum value set by congress, is justified. this article may be of great value for valuation experts brought in advise the family on whether they should recapitalize. the third article, “quantifying present valuation errors” by mangier0 and mangiero, will be of interest to anyone who has learned or taught finance using the ubiquitous present value tables. while we have intuitively realized that linear interpolation was inaccurate, this article shows mathematically why it is inaccurate and estimates the degree of error. although the use of computers and finance calculators have rendered these tables obsolete, the authors point out that many existing mortgage contracts and other installment contracts that are several years old may be inaccurate as well, resulting perhaps in overcharges which may be recoverable. the fourth article, “a simplified approach to measuring bond duration” by heck, zivney, and modani, gives readers a quick way to approximate duration and convexity. this will make it less daunting for those who wish to have a quick estimate of interest rate risk when evaluating alternative investments. in their article titled “analysis of u.s. savings bonds,” potts and reichenstein find that these bonds are complex contracts which have some valuable features that are often overlooked. for example, as opposed to other bonds of comparable duration and coupon, the value of savings bonds (which can be redeemed at any time) cannot go below the purchase price plus some stated (positive) rate of interest. this put option, which is generally overlooked by financial planners, can make a savings bond a better buy at times than other bonds that lack this feature. as begun in the previous issue, managing editor barbara poole gives readers her perspective on how an article from this issue can be used in practice. finally, phyllis schiller myers presents abstracts of articles relating to individual financial management from a variety of academic journals. lewis mandell editor pii: s1057-0810(99)80002-x from the editor karen eilers lahey jenny ridge and martin young present the first of two articles in this issue on inter esting international financial investments. they examine the history and operation of a sav ings scheme in new zealand and why it appeals to individual investors in their article entitled, "innovations in savings schemes: the bonus bonds trust in new zealand." it combines elements of both bonds and lottery tickets in the united states, and is held by a third of all new zealanders. "closed-end investment companies: historic returns and investment strategies" by carolyn reichert and j. douglas timmons examine simple trading rules that individual investors could follow. they provide a review of the literature on closed-end funds and find that institutional investors have been successful in earning excess returns through complex trading strategies. their study tests a buy and hold yearly rebalancing of closed funds by individuals. judson w. russell focuses on united states exchange-listed investments as a potential vehicle for international diversification. his article, "the international diversification fal lacy of exchange-listed securities" examines closed-end country funds, american depos itory receipts (adrs), and multinational corporations (mncs). results suggest that they do not provide united states investors with expected diversification. "an analysis of personal financial literacy among college students" was initially presented at the 1997 academy of financial services meeting in hawaii. haiyang chen and ronald p. volpe survey college students to determine their level of personal financial knowledge. their results indicate that there are differences based on college major, gender, rank, age, and work experience. they argue that college students do not understand per sonal finance and that steps should be taken to correct this problem. the second international article by diana beal and michelle goyen provides an expla nation for possible motivations for investing in "ethical firms" in australia. in their article, "'putting your money where your mouth is.' a profile of ethical investors," they explain the problems facing individuals who wish to select socially responsible firms and report on a survey they conduct of shareholders in a publically owned firm called earth sanctuaries limited (esl). they compare their findings with a survey conducted by the australian stock exchange. pii: 1057-0810(91)90029-x financial services review, l(2): 131-142 copyright 0 1991 by jai press inc. issn: 1057-08 10 all rights of reproduction in any form reserved. comparing mortgages with different payment frequencies arefaine g. yohannes the biweekly-payment mortgage is an alternative to the monthly-payment mortgage. in this study, logistic regression is used to determine the influence of demographic characteristics on the choice between these two types of mortgages. of the six independent variables that were considered, only two had significant influence on mortgage choice. they were thefrequency of paydays and education. in the last few years, there have been a number of studies dealing with mortgage choice. many of these studies dealt with the choice between fixed-rate mortgages and adjustable-rate mortgages. examples of such studies include brueckner and follain (1988), chmura (1989), dhillon, shilling and sirmans (1987), gardner, kang and mills (1987), and tucker (1989). other studies examined the choice between fifteen-year and thirty-year tixed rate mortgages. examples of such studies include dhillon, shilling and sirmans (1990) and yohannes (1986). in this study, mortgages with different payment frequencies are compared. specifically, this study deals with the choice between the biweekly-payment mort gage and the monthly-payment mortgage. although the biweekly-payment mortgage (bpm) has been around for some time in other countries such as canada, it is fairly new in the united states. for this reason, it is not as widely known or as widely used as the monthly payment mortgage (mpm). also, research on the bpm is limited. yohannes (1988) evaluated fast-pay mortgages, including bpms, from the borrower’s point of view. shilling and sirmans (1987) examined the pricing of fast-pay mortgages from the lender’s point of view. in a bpm, half of the monthly payment of a mpm is paid every two weeks. that means 26 payments are made in a year. in undiscounted terms, the 26 biweekly arefaine g. yohannes l school of management, the university of michigan-dearborn, 4901 evergreen road, dearborn, mi 48128. 132 financial services review, l(2) 1991 payments are equal to 13 monthly payments. the extra monthly payment and the frequency of payments associated with the bpm reduce the term of a 30-year mortgage by about one-third. in addition, the home owner pays down the principal faster than she or he would with a mpm. on the negative side, the bpm involves larger annual payments than a mpm. furthermore, for tax purposes, the annual interest deductions associated with the bpm are smaller than those of the mpm. finally, unless the biweekly payments are made electronically, the frequency of payments would be burdensome for borrowers and lenders. the purpose of this study is to examine the influence of specific borrower characteristics on the choice between the bpm and the mpm. the borrower characteristics that were included in this study are frequency of paydays, income, net worth, education, age and marital status. the hypotheses pertaining to these factors are given in the next section. with the exception of frequency of paydays, the borrower characteristics listed in the previous paragraph were used in previous studies of mortgage choice. the studies cited above did not use the same variables. apparently, the choice of variables depended on the type of mortgage choice under study and/or the availabil ity of data. the authors do not claim that the list of variables included in this study is exhaustive or complete. however, at the time the survey form was prepared, we believed that the variables listed above would influence the mortgage payment frequency decision. to the extent that other important variables were left out, we hope that some of their effects would be captured by some of the variables that were included in the study. for example, the borrower’s expected duration of stay in the house was not included in the model and it may be argued that this variable is an important factor. to the extent that younger home owners are more mobile than other home owners, age may be considered a proxy for mobility or the expected housing tenure. themodel since the dependent variable is a categorical variable, the logistic regression model was used to determine the effects of borrower characteristics on the odds or probability of selecting a bpm. the present value framework as well as probit and the logistic regression models were used in studies of mortgage choice. chmura (1989) and yohannes (1986, 1988) used the present value framework; dhillon et al. (1986, 1989) used probit analysis and gardner et al. (1987) and tucker (1989) used logit analysis. the probit and the logistic regression models are both appropriate for models in which the dependent variable is categorical or dichotomous. however, the logis tic model was chosen for this study for the same reasons that are given in tucker (1989). first, it does not require normality of the sampling distribution. second, since the mid-range ofthe logit curve tends to be steeper than that of the probit curve, comparing mortgages with different payment frequencies 133 it is more likely to capture the effects of small differences between the characteristics of those who select the biweekly-payment mortgage and the characteristics of those who select the monthly-payment mortgage. the logistic regression model is given by equation (1). e(y) = exp(w 1 + eww) = [l + exp(-w)]-i (1) where w= e(y)1 x’s = b’s = exp = b, + b,x, + . . . + bp_,xp_, logit response function expected value of the dependent variable independent variables model parameters base of the natural logarithmic system since equation (1) is non-linear, it is transformed into a linear function that can be used to estimate the parameters of the model. specifically, if we let e( y) = p, where p is the probability that y will be one, equation (1) may be transformed into the logit response function: ln(p/l p) = w = b, + b,x, + . . . bp_,xpmi the dependent variable, y, was mortgage choice. it was coded a “ 1” if the bpm was selected; it was coded a “0” if a mpm was selected. as pointed out above, the independent variables were frequency of paydays, income, net worth, education, age and marital status. one of the hypotheses that was tested is that borrowers who are paid biweekly will tend to choose the biweekly-payment mortgage. first, there is the convenience factor and secondly, if the people who are paid biweekly select a monthly-payment mortgage, there is a risk that those people will spend the money on something else before the mortgage due date arrives. people with higher incomes and larger net worth would be expected to choose the bpm. they would be in a better position to handle the larger annual payments associated with the bpm. on the other hand, high-income borrowers may be more interested in the larger interest deductions that the monthly-payment mortgage offers. so, the relationship between income and the choice between the bpm and the mpm is not clear. the two effects could, possibly, offset each other. some authors refer to the bpm as the “yuppie mortgage,” the mortgage for young, urban professionals with high income. see debat (1986). however, a priori, it is not clear what the effect of education on the choice between the bpm and the mpm would be. to the extent that borrowers with higher education levels tend to 134 financial services review, l(2) 1991 have higher incomes and net worth than borrowers with lower education levels, borrowers with higher formal education would tend to choose the bpm. they would be in a better position to pay the larger annual payments associated with the bpm. on the other hand, because of the greater tax benefits of larger interest deductions, the mpm may be more appealing to the more educated borrowers. furthermore, the more educated borrowers may know more about alternative mortgages and investment products and they may not be impressed by the bpm as the borrowers with lower education might be. for example, a home owner can obtain a mpm and then design a prepayment plan that is equivalent to the bpm without refinancing. the mpm gives the home owner greater flexibility than does the bpm. as far as age is concerned, we would expect older borrowers to choose the bpm so they can pay off their mortgage sooner and own their homes free and clear before retirement. in addition, older borrowers may have higher incomes and larger net worth and they may be in a better position to handle the bpm. here, again the greater tax benefit associated with the mpm may lead them to choose the mpm. finally, married couples may tend to choose the bpm since they are more likely to have higher incomes than singles. they also benefit from the economies of joint living. on the other hand, the mpm may be more attractive to them than the bpm for tax reasons. with the exception of the frequency of paydays, the effects of the factors discussed above on the choice between the bpm and the mpm is not clear. some of the factors may lead to the choice of the bpm and others may lead to the choice of the mpm. it is also possible that because of the types of offsetting effects that were discussed above, some of the variables may not have any significant effect on the choice between the two mortgage instruments. with the exception of age, all the independent variables in this study were treated as categorical variables. two of the independent variables, income and net worth, had three classes each and they were each represented by two indicator variables. the income classes were under $35,000, $35,000-70,000 and over $70,000. the net worth classes were under $100,000, $loo,ooo-300,000 and over $300,000. with respect to education, there were three classes. those who completed college, those who completed high school and those who did not complete high school. since all the respondents were either high school or college graduates, the third class was dropped. the data for this study were collected from a limited regional survey of home owners and potential home owners in the northern suburbs of detroit. first, zip codes were selected randomly from a list of zip codes for the northern suburbs of detroit. second, 500 addresses were selected randomly from the lists of addresses within the zip codes selected in the first stage. third, the home owners or potential home owners were asked to supply demographic data and to indicate their choice between the bpm and mpm. to make sure that the respondents knew what the bpm was and how it differed comparing mortgages with different payment frequencies 135 table 1. summary of survey results payday frequency: biweekly or weekly: 65 monthly: 19 income: less than $35,000 per year 6 between $35,000 & $70,000 31 over $70,000 47 education: less than cdllege 10 college completed 74 net worth: less than $100,000 between $100,000 & $300,000 over $300,000 11 22 51 marital status: married 69 single, divorced or widowed 15 age: highest 86 lowest 24 average 49 standard deviation 13 from the mpm, a description of the bpm, the differences between the two types of mortgages and a numerical example were sent along with the survey form to the home owners or potential home owners. a copy of the description is attached to the manuscript. also, in order to increase the response rate, the subjects selected for the survey were offered a free evaluation of alternative mortgage prepayment plans. the evaluation involved the determination of the effects of various prepayment plans on the total interest costs and the terms of specific mortgages. of the 84 respondents, 58 % selected the bpm and 42 % selected the mpm. seventy seven percent were paid biweekly or weekly and only 23% were paid monthly. eighty-two percent were married and the ages ranged from 24 to 86, with an average of 49. most of the respondents were college educated and they were in the high income and high net worth categories. the results of the survey are summa rized in table 1. before running the logistic regression, we constructed 2 x 2 tables to see if there is any discernible association between the dependent variable and each of the independent variables, except age. the two-by-two tables are shown in table 2. the odds ratios were computed for each of the 2 x 2 tables. 136 financial services review, l(2) 1991 table 2. two by two tables mortgage type mpm bpm mortgage type mpm bpm mortgage type mpm bpm mortgage type mpm bpm odds ratios mortgage type mpm bpm odds ratios payday frequency monthly biweekly or weekly odds ratio 12 23 3.31 i 42 marital status single married odds ratio 8 27 1 .i8 i 42 education less than college college completed odds ratio 2 33 0.27 9 40 low net worth medium high high/medium 0.64 3 8 24 8 14 27 medium/law high/law 0.66 0.42 income low medium high 2 10 22 4 21 24 medium/law high/law high/medium 1.05 0.55 0.52 notes: mpm = monthly-payment mortgage bpm = biweekly-payment mortgage based on the 2 x 2 tables and the odds ratios, only two of the independent variables seemed to be strongly associated with mortgage choice. the first one was the frequency of paydays. the odds ratio was 3.3 1. home owners that were paid biweekly or weekly tended to prefer the biweekly payment mortgage. the second important independent variable was education. the odds ratio was 0.27. the group with a higher education had greater relative preference for the mpm. there seems to be some association between mortgage choice and the other variables, too, but the association does not seem to be as strong as it is with the payday frequency and education. comparing mortgages with different payment frequencies 137 the statistical analysis system (sas) was used to estimate the full model, with all the independent variables included. the estimated values for the full model are reported in table 3. for large samples, the maximum likelihood estimators are distributed approximately normal. for this reason, the usual z tests apply. at a significance level of 10 % , the coefficients of the payday frequency and income 1 (a dummy variable for the medium income category) are significantly different from zero. also the coefficient of the payday frequency has the expected sign. however, none of the coefficients is statistically significant at the 5% level of significance. thus, home owners with medium income and those who are paid weekly or biweekly seem to favor the biweekly-payment mortgage. table 3. estimated values of the full logistic regression model variable coeficient std. error p-value odds ratio intercept -0.3605 2.786 0.897 freq of pay 1.3610 0.721 0.059 income1 2.5940 1.437 0.071 income2 1.5557 1.466 0.289 age -0.0170 0.027 0.536 education -1.1828 1.136 0.298 networthl -0.6691 1.157 0.563 networth -0.8090 1.243 0.515 mar status 1.6980 0.880 0.054 3.900 13.383 4.738 0.983 0.306 0.512 0.445 5.463 model chi-square = 14.85 with 8 degree of freedom model p-value = 0.0621 payfreq = 1, if paid weekly or biweekly = 0, if paid monthly income1 low income 0 medium income 1 high income 0 education = 0, if less than college = 1, if college completed income2 0 0 1 networth 1 networth low networth 0 0 medium networth 1 0 high networth 0 1 marital status = 0, if single, divorced or widowed = 1, if married 138 financial services review, l(2) 1991 since we did not find the results of the full model satisfactory, the stepwise logistic regression procedure in sas was used to generate the final model for this study. the statistics for the final model are given in table 4. here again, the z tests indicate that the coefficients of the payday frequency and education are significantly different from zero at both the 5 % and the 10% levels of significance. in addition, these results seem to be consistent with the observations derived from the individual 2 x 2 tables. only two of the six independent variables, frequency of paydays and education, remained in the final model. also, note that the model chi-square is smaller for the final model than for the full model. furthermore, the odds ratios from the logistic regression indicate stronger relationships between the independent variables and the dependent variable than was suggested by the odds ratios from the 2 x 2 tables. finally, using a cutoff point of 58 % , the estimated logistic regression model correctly classified 68 % of the respondents. a cutoff point of 58 % was used because that was the percentage of the survey respondents who selected the bpm. conclusion this study examined the effects of various demographic factors on the choice between the bpm and the mpm. of the six factors that were considered, only two had statistically significant influence on mortgage choice. these factors were the frequency of paydays and education. the importance of the first factor is rather obvious. ceteris paribus, home owners who are paid weekly or biweekly, are more likely to choose the bpm than home owners who are paid monthly. the effect of education was, however, surprising. home owners with college table 4. estimated values of the final logistic regression model and odds ratios vuriable coefjicient std. error intercept 0.8389 0.837 freq of pay i s784 0.629 education -1.9342 0.917 model chi-square = 10.37 with 2 degree of freedom p-value = 0.0056 freq of pay = i, if paid weekly or biweekly = 0, if paid monthly education = 0, if less than college = 1, if college completed p-value 0.316 0.012 0.035 odds ratio 4.847 0.145 comparing mortgages with different payment frequencies 139 degrees tended to prefer the mpm and home owners with just high school diplomas tended to prefer the bpm. as indicated above, one possible explanation for this result is that the more educated group knows more about alternative mortgage and investment products and that they are not as impressed by the bpm as those with just high school education. for example, a home owner can obtain a mpm and then design a prepayment plan that is equivalent to the bpm without refinancing. the mpm gives the home owner greater flexibility than does the bpm. another possible explanation is that the more educated group had higher income and consequently preferred the mpm for the greater tax benefits that it offers. in other words, for the larger interest payments that may be deducted for tax purposes. the results of this study have implications both for borrowers and lenders. assuming other factors are held constant, the biweekly-payment mortgage is, obviously, more suited for home owners who are paid weekly or biweekly rather than for those who are paid monthly. from the lender’s point of view, the marketing of the biweekly-payment mortgage should focus on home owners who are paid biweekly or weekly. home owners who are paid monthly or those with college degrees seem to be relatively more interested in the monthly-payment mortgage. while this study has identified borrower characteristics that are important for the choice between the biweekly-payment mortgage and the monthly-payment mortgage, it must be pointed out that it was based on a limited regional survey. future studies may be based on national surveys. they may also include other variables that were not considered in this study. appendix dear home-owner or potential home-owner: i am preparing a study of the factors that influence the choice between monthly payment mortgages and biweekly-payment mortgages. this study will be based on data gathered from a mail survey of home-owners or potential home-owners in this area. i would appreciate it very much if you could take a moment to read this memo and complete the enclosed short form. in the last few years, some lending institutions in michigan started offering biweekly mortgages. in a biweekly mortgage, half of the monthly payment of a monthly-payment mortgage is paid every two weeks. over a period of a year, 26 biweekly payments are made and these payments are equal to 13 monthly payments of a regular monthly-payment mortgage. the extra payment associated with the biweekly mortgage reduces the term of a 30-year mortgage by about one-third. as an example, suppose you borrow $75,000 at 9.75% interest for 30 years. your monthly payment would be $644.37. if you make biweekly payments of $322.18 140 financial services review, l(2) 1991 (that is half of $644.37)) the term of the 30-year mortgage would be reduced to about 21 years. the monthly-payment mortgage involves smaller annual payments and offers larger annual interest deductions for tax purposes. on the other hand, the biweekly payment mortgage accelerates the pay-off of the mortgage and results in greater equity for the home buyer overtime. given this background information about the monthly-payment mortgage and the biweekly-payment mortgage, we would appreciate it if you could tell us which of these two types of mortgages you would choose by completing the enclosed form and returning it to us in the stamped, self-addressed envelope by february 15, 1990. as you can see, the form is completed anonymously. form i. biweekly vs. monthly-payment mortgages please check the appropriate items below. your preference 0 monthly-payment mortgage •i biweekly-payment mortgage your pay checks received 0 biweekly 0 monthly 0 other your annual cl under $35,000 0 $35,ooo-70,000 gross income 0 over $70,000 age in years education completed cl less than high school cl high school cl college net worth* cl under $100,000 0 $loo,ooo-300,000 cl over $300,000 marital status 0 married 0 single, divorced, widowed *net worth is the amount you would be left with if you sold all your assets and paid off all your debts. please return the completed form in the enclosed, self addressed, stamped envelope. comparing mortgages with different payment frequencies 141 form 2. prepayment options if you already have a mortgage, you may be able to reduce the term of your mortgage without refinancing if your lender allows partial prepayments without penalty. there are various prepayment options. you could make additional principal payments monthly, quarterly, semi-annually or annually. you could also make lump-sum payments. for a free evaluation of various prepayment options, fill the blanks below and we would be glad to prepare a customized evaluation. you may get some of the informa tion requested from your mortgage documents or your bank. item amount current loan balance remaining term of the loan in years annual interest rate (note rate) in % current principal & interest payment amount of additional principal payments you wish to make (you may fill one or more blanks) prepayment plan amount one-time, lump-sum monthly quarterly semi-annual annual note: for purposes of anonymity, please return this form in a separate envelope. acknowledgments: the author wishes to thank the campus grants commit tee of the university of michigan-dearborn for financial support. references brueckner, jan k., and james r. follain. 1988. “the rise and fall of the arm: an econometric analysis of mortgage choice,” review of economics and statistics, february: 93-102. chmura, christine. 1989. “choosing between an adjustable rate and a fixed rate mortgage,” pp. 6 8 in federal reserve bank of dallas, cross section, summer. 142 financial services review, l(2) 1991 debat, don. 1986. the mortgage manual. chicago: contemporary press. dhillon, upinder, james shilling, and c.f. sirmans. 1987. “choosing between fixed and adjustable rate mortgages,” journal of money, credit and banking, february: 260-267. dhillon, upinder, james shilling, and c.f. sirmans. 1990. “the mortgage maturity decision: the choice between 15yearand 30-year frms,” southern economicjournal, april: 1103-i 116. gardner, mona j., han bin kang, and dixie l. mills. 1987. “consumer profiles and acceptance of arm features: an application of logit regression,” journal ofreal estate research, winter: 63-74. rossi, peter, j. wright, and a. anderson. 1983. handbookofsurvey research. new york: academic press. shilling, j.d., and c.f. sirmans. 1987. “pricing fast-pay mortgages: some simulation results,” journal of financial research, spring: 25-32. tucker, michael. 1989. “adjustable rate and fixed rate mortgage choice: a logit analysis,” journal of real estate research, spring: 81-90. yohannes, a.g. 1986. “comparing mortgages with different terms, ” real estate review, fall: 93 96. yohannes, a.g. 1988. “evaluating alternative fast-pay mortgages,” journal of real estate research, fall: 23-29. pii: s1057-0810(97)90030-5 financial services review, 6(1): 27-39 issn: 1057-0810 copyright © 1997 by jai press inc. all rights of reproduction in any form reserved. short selling and trading abuses on nasdaq robert l. albert jr. timothy r. smaby h. david robison we examine the potential for short-selling trading abuses unique to nasdaq during a period when there was no up-tick rule and no effective prohibitions against "naked" short selling. we find that (a) short sellers earned significant abnormal returns on nas daq securities, but these were smaller than on nysf_/amex securities; (b) they did not destabilize markets by selling into falling markets and exacerbating price drops; and (c) nasdaq short sellers may be more susceptible than nyse/amex shorts to "short squeezes." our results cast doubt on the appropriateness of recent regulatory reforms established for nasdaq and public concern over nasdaq short-selling abuses. i . introduction short sellers attempt to profit from expected declines in stock prices by borrok, ing and sell ing shares they do not own, then buying and replacing the borrowed shares when the price drops. while it is a time-honored method of exploiting bad news about stock prices, the practice has come under increasing regulatory scrutiny as other market participants have accused short sellers of manipulating prices, increasing volatility and exacerbating price drops in declining markets. descriptions of short sellers' trading activities have been decidedly one-sided, por traying shorts as stock-bashing rumor mongers. weiss (1996), in a recent business week report, reveals common misperceptions of short sellers held by other investors. he identi fies an executive who blames his company's recent stock slide of 60% over three months on "mudslinging'" short sellers. according to weiss, many market participants view short sellers as the "assassins of corporate america." they believe small-cap stocks are espe robert l. albert jr. • assistant professor of finance, morehead state university, morehead, ky 40351; e-mail: r.albert@morehead-st.edu. timothy r. smaby • assistant professor of finance, penn state eriedthe behrend college, station road, erie, pa 16563-1400; e-mail: trsl0@psuvrn.psu.edu. h. david robison • associate professor of economics, la salle university, philadelphia, pa 19141; e-mail: robison@lasalle.edu. 28 financial services review 6(1) 1997 cially vulnerable to the trading abuses of short sellers because negative rumors can have a more significant impact on share prices of thinly capitalized firms. it is generally acknowledged by market participants that short selling fraud schemes have been more prevalent in nasdaq securities due to the previously more lax short-selling requirements for nasdaq securities. in a recent wall street journal article, power (1995), identifies an atlanta trader who was charged with a short-selling fraud scheme which cost wall street brokerage firms $13.5 million. the trader allegedly sold stock in six firms (five of which were traded on nasdaq) he did not own. he then planned to settle the trades by purchasing the stock from another firm when prices fell. when prices did not fall, the trader allegedly walked away from his obligation, leaving the brokerage houses liable for his losses. the primary motivation for this paper is to examine the potential for short-selling trad ing abuses unique to nasdaq stocks before recent restrictive regulatory changes were enacted. in order to determine the potential for trading abuses uniquely attributable to the previously more lax nasdaq short selling regulations, we also examine short-selling prac tices on the organized exchanges. of particular interest is whether short sellers earn abnor mal returns on nasdaq securities, and whether they exacerbate price drops by selling in falling markets. previous research, with one exception, has focused on the profitability of short selling on the nyse/amex organized exchanges. until 1994, nasdaq, unlike the organized exchanges, lacked an uptick mie and effec tive prohibitions against naked short selling. an uptick rule is intended to prevent short sellers from exacerbating price declines by requiring them to sell at a price greater than that of the last recorded trade. naked short selling occurs when a trader sells securities short without frrst borrowing the shares in the open market. if the short is then forced to purchase shares to deliver when the buyer requests a buy-in, there may be a short squeeze in illiquid issues. in the mid-1980' s, the national association of securities dealers (nasd) commis sioned a report by pollack (1986) and began to develop proposals on changes in short sell ing of nasdaq securities. in 1986, nasd received approval from the securities and exchange commission (sec) to require affmmative determination of the deliverability of shorted shares to buyers of those shares. this affirmative determination requirement, how ever, was often met simply by identifying borrowable stocks listed on daily fax sheets, making it possible to short sell securities without first identifying the specific location and ownership of the shares to be borrowed (despite nasd's policy disallowing this practice). discussions with nasd's general counsel's office reveal that the lack of an annotation requirement often led to problems in enforcement of the affirmative determination require ment. on july 28, 1994, the nasd filed a proposed rule change with the sec that would require members or persons associated with members to annotate the affirmative determi nation made prior to effecting a short sale. the sec approved this recommendation on september 12, 1994. the lack of an uptick rule also contributed to the chance of abusive trading practices by aggressive short sellers, according to a january 1992 report by the house committee on government operations entitled "short-selling activity in the stock market: market effects and the need for regulation (part i)." on july 29, 1994, the sec granted tempo rary 18-month approval of the nasd-requested "bid test" rule for short selling of nasd securities. the rule prohibits members from effecting short sales when the bid price is at or below the previous bid, but provides an exemption for qualified market makers. this rule became effective in september 1994. short selling and trading abuses 29 individual investors closely follow the activities of short sellers by monitoring large increases in open short-interest positions of nyse/amex and nasdaq stocks reported monthly in the wall street journal. if large increases in short interest are bearish signals, individual investors can earn abnormal returns by shorting those shares following the release of the information in the wall street journal. if large increases are bullish signals, investors can profit by buying shares. we define two related empirical questions. first, can short sellers earn abnormal returns? second, can individual investors earn abnormal returns by reacting to published reports of short-selling activity? previous studies confuse the empirical implications of these questions by misidentifying the appropriate event date. a secondary contribution of this paper is to correctly identify the appropriate event dates for these empirical questions and disentangle the two effects. the paper is organized as follows. we review the literature on the profitability of short selling and the relationship between stock returns and short interest levels in section ii. we describe our data and methods in section hi, including the estimation of size-adjusted abnormal returns and the definition of the proper event dates. we present our results in section iv, and offer some conclusions and directions for future research in section v. h. literature review there are two strands to the empirical literature on short selling. the early literature exam ined the temporal relationship between stock market returns and aggregate short interest levels. figlewski (1981) reports a significant negative relationship between levels of short interest and abnormal returns using monthly returns data from january 1973 through june 1979, which is consistent with a bearish interpretation of high levels of short interest. sen eca (1967) and kerrigan (1974) also report a significant negative relationship between the short interest ratio and s&p 500 returns. aksu and gunay (1995), however, test for the presence of unit roots and co-integration among stock prices, short interest, and average trading volume, and find no relationship between levels of short interest and stock prices. other studies reporting similar results include mayor (1968), biggs (1966), and smith (1968). more recently, researchers have used event studies to estimate the abnormal returns to individual securities for which large increases in short interest levels have been reported. choie and hwang (1994) estimate abnormal returns to heavily-shorted nyse/amex stocks by measuring the change in prices from the midpoint between monthly wall street journal publication dates (eleven trading days prior to the publication date and one day before the compilation date---the date on which the exchanges compile short interest data) to the publication date. they report that stocks with large short positions underperform the stock market in this period immediately following the compilation date. they note this is evidence of excess returns to short sellers. they also find evidence that stocks with large short positions underperform the market in the twenty trading days after publication. this suggests large increases in short interest on nyse/amex stocks are interpreted by the market as bearish signals. we refine choie and hwang's event date definitions, and extend their results to nasdaq securities. 30 financial services review 6(i) 1997 woolridge and dickinson (1994) examine the coincidental relationship between the level of short sales and monthly holding-period returns for 50 randomly-selected securities from the wall street journal monthly short-sale list for both nyse/amex and nasdaq stocks. they report aggregate short positions increase as market prices increase, and short sellers are unable to earn abnormal profits at the expense of less informed traders. woolridge and dickinson conclude high levels of short interest are neither bullish nor bearish. however, they fail to account for noninformational arbitrage trading that requires short selling but would not be expected a priori to affect prices. strategies such as going long in a merger target and simultaneously going short in the acquiring firm, or "shorting against the box," are examples of such trading. as a result, it is likely that woolridge and dickinson significantly underestimated the returns to information-based short selling. in addition, because they use the publication date as the event date, they do not accu rately estimate the abnormal returns to short sellers. in addition, their use of monthly returns may cause them to miss any intra-month return differences between the nyse/ amex and nasdaq markets. moreover, statistical tests based on the relatively small wool ridge-dickinson sample (50 nysf_,/amex and 50 nasdaq f'lrms) may lack power. we use the appropriate event date (the compilation date) and a much larger sample of nyse/amex and nasdaq firms to estimate returns to short sellers. we also drop from our sample observations that potentially represent noninformational short selling, and we use the bid-ask midpoint to compute returns for nasdaq stocks rather than closing prices. vu and caster (1987) investigate the market reaction to the release of the wall street journal short interest data between 1975 and 1983. they find significantly positive cumu lative abnormal returns in the 40-day period before the information publication date. these findings are consistent with the tendency of investors to sell short in periods of rising stock prices. vu and caster report negative, but insignificant, cumulative abnormal returns dur ing the 40-day period after the announcement, and conclude that the published level of short interest is neither bullish nor bearish. senchack and starks (1993) examine a sample of nyse/amex stocks listed in the wall street journal's monthly short interest column for which short interest positions had at least doubled from the prior month. they find a small, but statistically significant, neg ative reaction around the publication date. they conclude their'results provide weak sup port for the hypothesis that the market reaction to unusual increases in unexpected short interest is negative. hi. data and methods a) data the initial sample includes firms with the largest percentage increase in short interest from the prior month as reported in the wall street journal between january 1987 and december 1991. we discard from the sample any observation that potentially represents noninformational trading: a firm identified as a new listing, as having undergone a recent stock split, or as possibly involved in arbitrage activity. a 2-for-1 stock split, for instance, would double the reported short interest without any increase in speculative short posi tions. strategies such as going long in a merger target and simultaneously going short in the short selling and trading abuses 31 acquiring firm, or "shorting against the box," would be examples of nov_informational arbi trage trading. an observation is also deleted if the firm had been listed in the largest per centage increase in short interest category within 180 days of a prior listing. for inclusion a firm must have bid and ask prices (for the nasdaq sample) or daily returns (for the nysf_,/ amex sample) and a size decile portfolio ranking on the crsp nyse/amex and nas daq daily returns tape. these restrictions resur in a final sample of 497 nyse/amex and 531 nasdaq observations. b) abnormal returns returns for the nasdaq sample are computed using the midpoint of the bid-ask spread rather than closing prices to mitigate the problem of upward-biased returns of stocks with small prices, which is a particular concern for nasdaq stocks. returns for the nyse/ amex sample are taken from the crsp tapes. we use a standard market-adjusted event study methodology discussed in brown and warner (1980, 1985). abnormal returns are generated by subtracting the daily return for the market from the return for the individual firm. negative abnormal returns defined in this manner indicate positive abnormal returns to short sellers since they profit from price declines. our proxy for the market return is the equally-weighted average return for all firms in the same size decile as the firm in question. deciles are constructed separately for nasdaq and nyse/amex flrms. we use size adjusted returns to avoid the benchmarking problems associated with using either a pre event or post-event estimation period. test statistics are reported for the null hypothesis that mean abnormal returns or mean cumulative abnormal returns are zero. although they are technically t-statistics, they are equivalent to z-statistics in large samples such as ours, and are reported as z-statistics in the tables. the test statistics are consu'ucted by normalizing abnormal fm'n returns by an estimate of the standard deviation and summing these standardized residuals across flrms. due to the method by which the test statistics are constructed, it is possible for the mean abnormal return to be of a different sign than the corresponding test statistic. c) defining the event date large increases in the level of open short interest can mean one of two things to the individual investor. first, the increase can be interpreted as a general bearish sentiment by informed investors (to the extent short sellers are informed investors), suggesting that stock prices will fall farther. diamond and verrecchia (1987) develop a rational-expectations model consistent with this interpretation. the key testable empirical implication is that an unexpected increase in short interest is bad news and will cause stock prices to decline. according to this model, because short-selling restrictions eliminate more uninformed than informed short sales, the result is a set of short sales that contains a higher percentage of informed trades than in overall sell orders. second, the increase may represent latent buy ing pressure (because the short positions must eventually be covered) suggesting that stock prices will rise. whether short interest is interpreted as a bullish or bearish signal depends on the sub sequent returns to a short position following public release of information on short interest levels of individual stocks in the wall street yournal. whether short sellers on average earn abnormal returns depends on any excess returns to short positions after they are estab lished. the research problem is to estimate the date at which the short positions were taken. 32 financial servicf~ review 6(1) 1997 the nyse/amex and nasdaq markets compile monthly short interest positions on indi vidual stocks as of the fifteenth of each month. because settlement takes five trading days, the exchanges actually compile short interest data as of the eighth to the tenth of each month (the compilation date). (in june of 1995, after the period of our study, the markets went to three-day settlement.) the wall street journal consistently publishes short interest data for nyse/amex and nasdaq stocks ten and thirteen trading days after compilation dates, respectively. we use the compilation date as the best estimate of the date on which the short positions were estabfished and define it as day ~t=-0 in event time. previous studies have used a variety of dates between compilation and publication. the compilation date is the latest date at which large increases in short interest could have taken place and still be reported in the next wall street journal, and more closely approximates the date on which short sellers establish their positions. it is a better estimate than the publication date, which is several trading days after significant increases in short interest levels. compilation publication date date t =-20 t=0 t = +10 (nysfjamex) t ffi +13(hasdaq) t = +30 p m t ~ n m n i l ~ t i n n ~.~ (-20,0) lnto~rim (+1, publication date) pnc:t-~.nmr~il~ (+1,+30) pn~-ihd~l ie~n (publication date,+30") ion figure 1. the time-line and relevant event intervals are presented in figure i. we define the com pilation date as day t = 0 in event time. the nyse/amex and nasdaq markets compile monthly short interest positions on individual stocks as of the fifteenth of each month. because settlement takes five trading days, the exchanges actually compile short interest data as of the eighth to the tenth of each month (the compilation date). the wsj consistently publishes short interest data for nyse/ amex and nasdaq stocks ten and thriteen trading days after compilation dates, respectively. excess returns in the pre-compilation interval include any rim-up in prices prior to large increases in short interest. excess returns in the post-compilation interval estimate returns to short sellers. excess returns to investors who respond to the public announcement of large increases in short interest are estimated during the post-pubfication period. excess returns to short sellers are the sum of returns in the interim period between compilation and publication, and returns after publication. short se///ng and trad/ng abuses 33 the time-line and relevant event intervals are presented in figure 1. note that negative abnormal returns represent positive excess returns to short sellers. excess returns in the pre-compilation interval include any price movements prior to large increases in short interest. excess returns in the post-compilation interval estimate returns to short sellers. excess returns to investors who respond to the public announcement of large increases in short interest are estimated during the post-publication period. excess returns to short sell ers are the sum of returns in the interim period between compilation and publication, and returns after publication. iv. e m p i r i c a l results a) the profi tabil i ty of shor t selling mean cumulative abnormal returns (cars) of both nyse/amex and nasdaq stocks over days t = -20 to t = 30 are plotted in figure 2. stocks in both samples experience a price increase prior to the compilation date, suggesting that short sellers tend to sell into rising markets. after the compilation date the mcars of both samples begin to drift down ward, which suggests short sellers earn excess returns. 0.06 0.05 0.04 0.03 0.o2 0.01 0 -0.01 o # ~ l b d ~ | m b "-75 -,0 -5 0 5 io ,5 20 . n ~ l a q event date figure 2. the mean cumulative abnormal returns (cars) for both the nyse/amex and nas daq samples over days t = -20 to t = +30 are plotted. the compilation date is defined as t = 0 in event time. abnormal returns are generated by subtracting the daily return for the market from the return for the individual firm. negative abnormal returns defined in this manner indicate positive abnormal returns to short sellers, since they profit from price declines. our proxy for the market return is the equally-weghted average remm for all firms in the same size decile as the firm in question. deciles are constructed sepetately for nasdaq and nyse/amex firms. 34 financial services review 6(1) 1997 table 1 event day (t) mean abnormal returns z-statistic -5 +0.0008 +0.37 -4 +0.0006 +0.09 -3 +0.0029 +4.23"** -2 +0.0064 +3.84*** -1 +0.0044 +3.77"** compilation date +0.0018 + 1.30 +1 +0.0006 +1.85" +2 -0.0002 -0.81 +3 -0.0016 -1.25 +4 -0.0015 -4.37*** +5 +0.0004 +0.73 +6 -0.0012 -0.64 +7 +0.0011 +2.15"* +8 -0.0018 -1.30 +9 +0.0001 +0.70 publication.date -0.0031 -2.43"* + 11 4).0001 +0.92 +12 -0.0017 -2.16"* +13 -0.0014 -2.51"* +14 -0.0010 -1.37 + 15 -0.0003 -0.74 +16 -0.0016 -1.63 +17 -0.0001 -0.61 +18 +0.0004 +0.11 +19 -0.0003 +0.64 +20 -0.0018 -1.31 notes: "**significant at the 1% level; ** significant at the 5% level; * significant at the 10% level. mean abnormal remrns(ars) for 497 nyse/ami~ firms are repotted from event day t = -5 to t = +20. abnormal are generated by subtracting the daily return for the market from the return for the individual firm. negative abnemml returns defined in this manner indicate positive almormal retrain to short sellers, since they wofit from price declines. our proxy for the market t~m'n is the equally-weighted average retta'n for all finns in the same size dacile as the firm in question. deciles are ccex,qr~ed separately for nasdaq and nyse/amex firms. mean abnormal returns (ars) for days t = -5 to t = +20 for the nyse/amex and nasdaq samples are listed in tables 1 and 2, respectively. the nyse/amex stocks generate significant positive excess returns on t = -3, -2 -1, and +1 around the compila tion date. there are significant negative excess returns on days t = +10 (the publication date), +12 and +13. the nasdaq stocks experience significant positive excess returns short sd//ng and trad/ng abuses 35 table 2 event day (0 mean abnormal returns z.statistic -5 +0.0066 +5.00*** -4 +0.0047 +3.36*** -3 +0.0063 +6.27*** -2 +0.0037 +3.51"** -1 +0.0084 +4.93*** compilation date +0.0113 +7.41 + 1 +0.0039 +2.97*** +2 +0.0014 +0.36 +3 +0.0006 -0.57 +4 +0.0033 +1.72"** +5 +0.0010 +0.54 +6 ,0.0019 1.42 +7 +0.0009 +i.17"* +8 ,0.0028 -2.35** +9 -0.0022 -2.30"* + 10 -0.0029 + 1.73" +11 +0.0001 -0.18 + 12 -0.0008 1.98"** publication date +0.0004 1.74" +14 -0.0044 --4.65*** +15 ,0.0017 -2.25** +16 +0.0011 +1.87" +17 -0.0007 -1.07 + 18 +0.0003 + 1.03 +19 -0.0001 -0.48 +20 -0.0031 -3.64*** notes: ***significant at the 1% level; ** significant at the 5% level; * significant at the 10% level. mean almormal returns(ars) for 531 nasdaq firms are repotled from event day t= -5 tot= +20. abnormal returns are generated by subtracting the daily return for the market from the return for the individual firm. negative abnormal returns defined in this manner indicate positive abnormal returns to short sellers, since they profit from wice declines. our proxy for the market retom is the ~lually-weigined average return for all firms in the same size decile as the firm in question. decries axe conm~cted separately for nasdaq and nyse/amex firms. on days t = -5 through +1 around the compilation date, and significant negative excess returns on days t = +12, +13 (the publication date), +14 and +15. excess returns over several intervals are presented in table 3, panels a and b, for the nyse/amex and nasdaq samples, respectively. the intervals are defined in figure 1. the results confirm the evidence in figure 2 and table 2 that short sellers tend to sell into 36 financial services review 6(1) 1997 table 3a mean cumulative abnormal returns (cars) for 497 nyse/amex firms are reported for selected intervals. interval event intervals (inclusive) car z-statistic pre-compilation (-20,0) +0.0320 +6.92"** interim (+1,+9) -0.0041 -0.10 post-publication (+ i 0,+30) -0.0174 -3.29"** post-compilation (+ 1,+30) -0.0215 -3.29 °** notes: *"significant at the !% level; *" significant at the 5% level; * significant at the 10% level. almonnal returns are generated by subtracting the daily return for the market frvm the return for the individual firm. negative almofmal returns defined in this marmex indicate positive abnormal returns to short sellers, since they profit from wice declines. our proxy for the market return is the equally-weighted average return for all finns in the same size decile as the firm in question. deciles are constructed sepm-ately for nasdaq and nyse/amex firms. table 3b mean cumulative abnormal returns (cars) of 531 nasdaq ftrms are reported for selected intervals. interval event intervals (inclusive) car z-statistic pre-compilation (-20,0) +0.0408 +5.74"** interim (+1,+12) +0.0063 -0.09 post-publication (+ 13,+30) -0.0092 -2.9 i** post-compilation (+ 1+30) -0.0028 -2.31"* notes: **'significant at the 1% level; ** significant at the 5% level; * significant at the 10% level. abnormal returns are generated by subffacfing the daily return for the market from the return for the individual firm. negative abnormal returns defined in this manner indicate positive abnormal returns to short sellers, since they wofit from price declines. our proxy for the market return is the equally-weighted average retarn for all firms in the same size decile as the firm in question. decries are conslrucmd separately for nasdaq and nyse/amex firms. rising markets. mcars in the pre-compilation interval are significantly positive in the nyse/amex market (3.20%) and the nasdaq market (4.08%). these results, suggesting short sellers in both markets short stocks that have recently experienced significant price increases, are consistent with vu and caster (1987) and woolridge and dickinson (1994). excess returns in the post-publication interval are significantly negative in both mar kets (-1.74% on nyse/amex and -0.92% on nasdaq). our evidence suggests the publi cation of large increases in short interest is a bearish signal to other investors, which is consistent with choie and hwang (1994), figlewski (1981), senchack and starks (1993), seneca (1967) and kerrigan (1974). it is inconsistent with vu and caster (1987) and wool ridge and dickinson (1994). excess returns in the post-compilation interval are significantly negative in both mar kets as well, suggesting short sellers earn excess returns of 2.15% on nyse/amex and 0.28% on nasd~. note, however, that short sellers actually earned smaller returns on nas daq than nyse/amex securities. cars in the interim interval are negative in the nyse/amex sample and positive in the nasdaq sample, but neither is statistically significant. this suggests that most of the short selling and trading abuses 37 return to short sellers comes from the public release of their trading activities that reflect their private information, since most of their excess return comes after the wall street journal publication date. this evidence is in contrast to choie and hwang (1994), who fmd significant negative returns between the compilation and publication dates. b) differences between nyse/amex and nnsdaq as discussed previously, a large increase in short interest can be interpreted as a gen eral bearish sentiment by informed investors (to the extent short sellers are informed inves tors), suggesting stock prices will fall farther. our initial results support this view. however, the increase may also represent latent buying pressure (because the short posi tions must eventually be covered) suggesting stock prices will rise. the lack of effective prohibitions against naked short selling of nasdaq securities dur ing the period of this study makes the possibility of rising prices during the post-compila tion period much stronger on the over-the-counter market than on the nyse or the amex. if the short seller does not immediately deliver the shares (known as naked short selling), and the buyer's broker is directed to execute a buy-in, the short seller may be squeezed in acquiring shares for delivery. as the short seller scrambles to acquire the shares to meet delivery requirements, the share price may increase due to the added buying pressure. this benefits buyers at the expense of short sellers and is more likely to be effective for thinly traded securities. therefore, if trading in a security is thin, buyers may see an incentive to request buy-ins. this may be especially true for small, illiquid stocks traded on the nasdaq market with no prohibitions against naked short selling. evidence of short squeezes in small issues has not previously been documented in the literature. however, any post-compilation price increases for heavily-shorted nas0~aq stocks (rel ative to exchange-traded stocks) may be due to the already well-documented firm-size effect and not due to a short squeeze, because the smallest nasdaq stocks are significantly smaller than the smallest nyse/amex stocks. even when size is controlled, it may be dif ficult to attribute any differential returns of heavily-shorted nasdaq stocks and nyse/ amex stocks to a short squeeze because of the difficulty in controlling for the many struc tural differences between the two markets. to shed some light on the issue, we regress post-compilation cars of each firm on the log of its market value of equity (lmveq) separately for each market. the results of the two regressions are reported in table 4. the relationship between size and cars on the nyse/amex exchanges is significantly positive (y1 = +0.019) indicating smaller nyse/ amex firms experience larger negative (smaller positive) cars than larger firms. the relationship for nasdaq firms, however, is not statistically significant (tl = -0.000). the difference between the slope coefficients in the two markets is statistically significant. these results are consistent with a different pattern of abnormal returns for heavily shorted small nasdaq stocks relative to nyse/amex stocks. small nasdaq stocks may be more susceptible to buy-ins and short squeezes which drive up prices after large increases in short interest. this calls into question the criticisms many market participants have lev ied against the activities of short sellers. in contrast, naked short sellers may be more sus ceptible to the trading abuses of the buyers who create the short squeezes (see weiss, 1996). because naked short selling is not permitted on the nyse/amex, heavily-shorted shares there are not as susceptible to buy-ins and short squeezes, and therefore continue to experience price declines after the compilation date. 38 financial services review 6(1) 1997 table 4 regression estimates for the sample of 497 nyse/amex firms and 531 nasdaq firms coefficient nyse nasdaq difference t l °0.264 +0.001 -0.263 (-3.97"**) (+0.01) (-2.67***) ~/2 +0.019 -0.000 +0.019 (+3.66***) (-0.05) (2.35**) notes: ***significant at the 1% level; **significant at the 5% level; *significant at the 10% level. we estimate the following regression separately for each market: cai~ ffi ~! + ~lmveqj + ej where: = cumulative abmmnal return for firm j from day t = +l to t = +30; ~ j ffi iog(mar~ value of equity of firm j); t-statistics in parentheses test the null hypothesis that the parameter estimate is not significantly different than zero. the significance of the differences between the two markets of the intercept and slope estimates are estimated from a dummy variable regression model. significance levels are for two-tailed tests. it is possible that the inverse relationship between size and return for nyse/amex stocks is a result of the bid-ask bias in those stocks for which we did not adjust. recall that returns were computed from the bid-ask midpoint for nasdaq stocks only. however, we also analyzed the date using closing prices for the return series for both nasdaq and nyse/ amex. those results (which are not reported) are not significantly different from the results we report in this paper, suggesting our results are robust to the method of estimating abnormal returns. v. summary and conclusions we do not find compelling evidence that short sellers on nasdaq are the "assassins of cor porate america" portrayed in the financial press. during a period in which there was no up tick rule and no effective prohibitions against naked short selling, short sellers did on aver age earn statistically significant abnormal returns on nasdaq securities. however, these abnormal returns were actually smaller than those on nyse/amex securities, and most of the return occurred after publication of short sellers' trading positions in the wall street journal. therefore, even though short sellers in both markets earned excess returns, those returns were also available to individual investors. furthermore, it appears that short sellers did not destabilize markets by selling into falling markets and exacerbating price drops. instead, they added liquidity, selling into rising markets by shorting stocks that had expe rienced large price increases in the thirty days prior to the establishment of their positions. we do, however, document some evidence that nasdaq short sellers, unlike those who trade in nsyfjamex securities, may be more susceptible to "short squeezes" which drive up stock prices and result in substantial short-selling losses. our results cast some doubt on the appropriateness of the regulatory reforms recently established for nasdaq and the prev alent public concern over short-selling abuses in the nasdaq market. however, too many short selling and trading abuses 39 other structural differences exist between the two markets for us to draw a definitive con clusion. we intend in future research to examine the effect of recent regulatory changes on nasdaq by comparing our results in this paper on nasdaq short-selling activity before the changes to recent activity after those changes. r e f e r e n c e s as]m, c., & gunay, e. (1995). an empirical analysis of the causal relationship between short inter est and stock prices. journal of business finance and accounting, 22, 733-747. biggs, b. (1966). the short interest: a false proverb. financialanalysts journal, 22, 111-116. brown, s., & warner, j. (1980). measuring security price performance. journal of financial eco nomics, 8, 205-258. brown, s., & warner, j. (1985). using daily stock returns: the case of event studies. journal of financial economics, 14, 3-31. choie, il, & hwang, s. (1994). profitability of short selling and exploitability of short information. journal of portfolio management, 20, 33-38. diamond, d., & verrecchia, j. (1987). constraints on short selling and asset price adjustment to pri vate information. journal of financial economics, 18, 277-311. figlewski, s. (1981). the informational effects of restrictions on short sales: some empirical evi dence. journal of financial and quantitative analysis, 16, 463-476. kerrigan, t. (1974). the short interest ratio and its component parts. financialanalysts journal, 30, 45-49. mayor, t. (1968). short trading activity and the price of equities: some simulation and regression results. journal of financial and quantitative analysis, 3, 283-298. pollack, i. (1986). short-sale regulation of nasdaq securities. washington i)c: national associa tion of securities dealers. power, william, (1995, february 28). sec charges trader with free-riding in alleged short selling fraud scheme. the wall street journal, p. b-5. senchack, a. jr., & starks, l. (1993). short sale restrictions and market reaction to short-interest announcements. journal of financial and quantitative analysis, 28, 177-194. seneca, j. (1967). short interest: bullish or bearish.9 journal of finance, 22, 67-70. smith, r. (1968). short interest and stock market prices. financial analysts journal, 24, 151-154. vu, j., & caster, p. (1987). why all the interest in short interest? financialanalysts journal, 43, 76 79. weiss, gary. (1996, august 5). the secret world of short sellers. business week, pp. 62-68. woolridge, j., & dickinson, a. (1994). short selling and common stock prices. financial analysts journal, 50, 20-28. pii: 1057-0810(95)90013-6 financial services review, 4( 1): 1-8 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any fm reserved. bank dividend policy as a signal of bank quality robert boldin keith leggett this article examines whether the dividendpolicy of bank holding companies is used as a signal of their quality. the study found evidence to support the dividend signaling argument-that is, that there is a positive relationship between bank dividendsper share and bank quality rating. aaiiitionally, an inverse relationship between the dividend payout ratio and bank quality was found. therefore, both aspects of a bank holding company’s dividendpolicy yields information about the quality of a fmancial institution. i. introduction during the decade of the 198os, the u.s. banking industry experienced a slow erosion in its financial health. while recent evidence indicates a reversal in the fortunes of the banking industry, long-term secular forces suggest a continued erosion in earnings in the banking industry and a greater potential for increased risk taking (barth, brumbaugh, & litan, 1992). with the secular deterioration in the bank industry’s financial health, there was renewed emphasis on the recapitalization of the banking industry. the process has continued into the 199os, with the industry’s capitalization ratio reaching its highest level in 30 years. the primary venue available for banks to raise capital is through retained earnings. the federal deposit insurance corporation (fdic) reported that in 1993, the banking industry paid out 50.7% of its earnings in cash dividends. this is lower than the banking industry’s 80% payout rate in the first quarter 1991 but much higher than 27.5% payout rate in 1985. given this more recent liberal dividend payout ratio, especially in light of anticipated long-term decline in earnings, the ability of banks to expand their capital base must be questioned. a fundamental tension associated with bank dividend policy is thus apparent. as bank earnings are squeezed, banks must choose between maintaining stable dividend payments per share versus a constant dividend payout ratio from earnings. both policy decisions can provide information about the existing and future soundness of the bank. yet, there is some concern as to whether the financial markets incorporate this information. the financial robert boldin l finance and legal studies department, indiana university of pennsylvania, indiana, pa 157051087. keith legptt l department of economics, davis & elkins college, elkins, wv 2624 l-3996. 2 financial services review 4(l) 1995 literature is replete with studies which have examined the impact of dividends as an investor signal (eades, 1982; aharony & swary, 1980). generally, an unexpected increase in dividends is viewed as conveying positive information about the financial health of the firm while negative information about asset quality is imparted by an unexpected cut in dividends. according to keen (1983), a central tenet of bank financial management is to avoid a cut in cash dividends because a dividend cut connotes a weakening in the soundness of a bank. furthermore, bhattacharya (1979) argues that due to informational asymmetries, cuts in dividends will have a greater negative impact on shareholder’s wealth than will positive effects associated with dividend increases. while cash dividends paid provides information about the future well-being of a bank, dividend payout as a percent of earnings (i.e., dividend payout ratio) also yields valuable information. mayne (1980) points out that in the mid 1970s retained earnings constituted 56% to 76% of the net growth in bank equity; currently, they represent about 30%. one may hypothesize that changes in the dividend payout ratio can potentially affect the capital position of a depository institution, the ability of banks to meet new opportunities, and, foremost, the potential soundness of the institution. the purpose of this study, therefore, is not only to define a bank management’s dividend policy (i.e. stable cash dividends versus stable payout ratios) but also to describe the role of dividends as a signal of market quality. market quality refers to the future expected cash flow from an asset. section ii outlines the dividend signaling argument. section iii provides information on the data used, while section iv presents the analysis. the last section summarizes the research. ii. dividendsignalingargument-backgroundinformation the rationale for analyzing the dividend policy of banks stems from the fact that reported financial information by depository institutions reveals book values rather than market values, even when the market value of a bank’s assets is available to management. (financial accounting standard 107 now requires publicly held banks or bank holding companies to report fair market values of assets and liabilities in a footnote in the annual financial statement.) however, since market values are not presented on a continuous basis, deposi tors, investors, and creditors know too little about the actual net worth and risk of the institution (white, 1989). this lack of information introduces uncertainty, because deposi tors and creditors alike cannot distinguish between zombie (poorly managed) and non-zom bie (well-run) institutions, and therefore the cost of capital is raised to all institutions (kane, 1989). hence, the potential arises for bad institutions to end up making good institutions unsound (an application of gresham’s law). therefore, when sellers know the quality of their product and buyers do not, sellers of the high-quality product (well-managed banks) have an incentive to signal this information to buyers. if the seller is successful, then the strong bank has segmented the market so that it is not actually competing in the same market as poor-asset-quality institutions. given this circumstance, the cost of capital should be lower for the strong bank, thereby improving the value of the bank. thus, between annual financial statements, one vehicle available to management to disclose its quality is its dividend policy. it is reasonable to assume that banks use their dividend policy to signal their well-being rather than continuously update and reveal market information (since asset quality and bank dividend policy as a signal of bank quality 3 market value of assets are closely related). disclosure of market value by a bank is represented by a transaction cost, c. bank management must compare the cost associated with disclosing market values versus the benefits received, which is measured by the difference between the market value of a bank’s assets, a,,, and the book value of a bank’s assets, a,,. if a, -ab is greater than or equal to c, then management has an incentive to reveal market value information. however, ifa, -a, is less than c, there does not exist an incentive to reveal market values. thus, it is possible for a bank not to reveal market values even though the reported book value of the bank’s assets is less than its market value. hence, dividend policy becomes an interim vehicle to distinguish between these sound and unsound institutions, since banks, in practice, do not generally disclose market values on an ongoing basis. ideally, if a bank increases its dividends, this should send a signal that management expects superior future cash flows. a high cash dividend indicates a reduced probability of failure, and this should improve the value of shareholders’ wealth. therefore, the dividend policy provides public information to the capital markets. additionally, false signaling will be discouraged by the financial markets because this will lead to higher transaction costs since the bank’s cash flow will be insufficient to maintain its stated dividend policy. however, a potential problem of high dividend payouts from earnings may jeopardize the future safety of a bank. the dividend payout ratio should provide information about the safety of a banking entity. a managerial incentive model, developed by bar-yosef and huffman (1986), indicates that there is an inverse relationship between the dividend payout ratio and risk. as risk increases, the dividends paid as a percent of earnings will decrease. thus, more earnings are being committed to improving a bank’s capital position. however, it is suspected that in the banking industry there is a direct relationship between the dividend payout ratio and risk. additionally, since risk or quality is generally unobservable to the public, the causation of the signaling argument should run from the dividend payout ratio to risk, not vice versa as the managerial incentive model contends. therefore, the signaling argument suggests a high dividend payout reflects an increase in the level of risk being assumed by a bank. thus, both dividends paid per share and dividends as a percent of earnings provide signals to the capital markets concerning the soundness and safety of depository institutions. iii. data the data in this study represent a cross-section of 207 publicly traded bank holding companies (bhcs) as of december 1989. nine bhcs are money center financial institutions and 25 bhcs are superregionals. the remaining 173 are regional bhcs. sources for the information in this study are american banker, sheshunoff bank holding company quar terly, and moody’s banking and finance manual. the united states is divided into six banking regions: northeast (ne), southeast (se), central (c), midwest (mw), southwest (sw), and west(w). the geographic distribution of the 207 bhcs has the largest concentration in the ne at 60, followed by the c with 53, the se with 45, the w with 28, the mw with 14, and 7 bhcs in the sw. table 1 provides a summary of the key financial statistics by region. performing amultiplerange test, we attempt to explain if there is any difference between the regional means for these selected measures. at a confidence level of .05, a significant financial services review 4( 1) 1995 table 1 mean financial statistics by region-december 3 1, 1989 variables region northeast southeast central midwest southwest west all dividends per earnings per dividend market price per asset size share share payout ratio captial ratio share (oooooo1 (dollars) (dollars) tw (w (dollars) 19035.0 1.18 1.31 90.1 10.236 21.94 8846.9 0.89 2.39 31.2 11.032 22.97 6489.0 0.91 2.85 31.9 11.825 27.38 6535.6 0.98 2.68 36.6 10.358 30.80 5014.5 0.67 1.17 57.3 7.397 18.92 13617.0 0.74 2.05 36.1 12.627 27.52 11564.0 o.% 2.13 45.1 11.025 26.67 difference was discovered between the mean earnings per share for bhcs in the northeast and the central regions. additionally, dividends per share in the northeast are significantly different from dividends per share in the south, central, midwest, and southwest regions at a .0.5 level of confidence. finally, the market value capitalization ratio for the west region is significantly different from the northeast and southwest regions; and also, the southeast and central regions’ market capitalization ratio is significantly different from the southwest. the data in table 1 indicate that bhcs in the southwest had not recovered from the local economic shocks of the mid 1980s; the data further provide early evidence of the weakening real estate markets impact on bhcs in the northeast. additionally, bhcs in both regions possess higher dividend payout ratios than the industry norm. this higher dividend payout ratio may account for the lower market value capitalization position for depository institutions in those regions, and it provides initial support for the hypothesis. additionally, it is possible that management was reluctant to reduce dividends (in the face of lower earnings per share) because of the negative effects of signaling. this study utilizes the sheshunoff s presidential rating as the measure of bank quality. the presidential rating is a weighted ordinal composite cabl (capital, asset, earnings, liquidity) percentile ranking of institutions within a peer group. these percentile rankings are then converted into a letter grade rating. the transformation of the data to an ordinal measure limits any potential problems of collinearity with explanatory variables. the letter grades are a+, a, b+, b, c+, c, and not rated (nr). so, sheshunoff s rating provides a ranked ordering of publicly traded financial institutions. table 2 reports the distribution of bhcs used in this study by letter grade. to test if the sheshunoff presidential percentile rating provides a consistent measure of bank quality, a comparison was made between the market value capital-to-asset ratio (which was calculated by multiplying the market to book value ratio by the capital to asset ratio) and the sheshunoff percentile rating. it is contended that as the market value capitalization ratio increases, the bank becomes more sound. to test this notion, a spearman rank correlation analysis was performed. the correlation coefficient between the two variables was found to be .6422, which is significant at a .ool level of confidence. this indicates a strong positive association between these two variables, which are both used to measure bank soundness. so, the maximization of quality should be consistent with the goal of maximizing shareholder wealth. bank dividend policy us a signal of bank quality table 2 the distribution of bank holding companies by letter grade rating rating built of bhcs a+ 26 a 54 b+ 52 b 39 ci 17 c 12 nr 7 iv. empiricalresults the following empirical analysis differs from most studies concerning the info~ation~ content of a bank’s dividend policy in three aspects. first, the data is cross-sectiona rather than longitudinal. second, instead of treating stock prices as the dependent variable, the dependent variable is the ordered ranking generated by sheshunoff. third, this analysis is not an event study which examines the reaction of stock prices to changes in dividend policy. in this study, the key explanatory variable is dividend policy-that is, dividends per share (d) and dividends paid as a percent of earnings (dout). the signaling ~gu~nt indicates that dividends should provide the relevant information about the future earnings of the bank and hence, the quality of the bank. so, one would hypothesize a positive relationship between dividends per share (d) and quality (i.e., sheshunoff rating). in addition, dividends paid as a percent of earnings should provide information about the level of bank capital and, thereby, the safety and the soundness of the institution. given the low levels of capitalization in the banking industry as compared to other industries, there should actually be a negative relationship between dividend payout and bank rating, ceteris paribus. the larger the dividend payout ratio, the more reliant is a bhc upon using external liability financing to generate income earning assets. this indicates that less capital is available to support any given level of assets and, thus, greater risk. since the dependent variabie is ordinal, the empirical procedure used in this study is an ordered probit response model, y=px+& y is the response variable, x is a vector of explanatory variables, and e is the residual. the residual, e, is distributed with a mean of zero and a variance of one. ordered response models have been used in previous studies to examine rating schemes of bonds (ederington, 1985; farnham & cluff, 1985). other control variables included in this study are a bhc’s primary capital position (cap), banksizeas measured by total assets (apse, and asset growth (ac), the annu~i~d growth rate. additionally, since there may exist regional variations in quality, this analysis incorporates regional dummy variables. furthermore, anticipating differences among re gional (f?), superregional (sr), and money center (mc) financial institutions, dummy financial services review 4(l) 1995 table 3 results from ordered probit response model dependent variable = sheshunoff ranking (nr=cj,c=~,c+=~,b=~,b+=~,a=~,a+=~) observations = 194 variables coefjfcient srandard error : -2.60461* 0.54803* 0.7748 0.2000 dout 9.00036** 0.00018 ag 0.03089” 0.01074 asset -4mloo3* 0.00001 cap 0.60971* 0.08597 ne -0.10017 0.27 18 se 0.33380 0.2824 c 0.46249 0.2967 mw 0.6925 1 0.4652 sw -0.72583 0.5924 mc -0.03062 0.5113 sr 0.44396 0.4292 pi 1.00212** 0.4847 112 1.55377* 0.495 i p3 2.49119* 0.5157 p4 3.430.53* 0.5377 15 4.67133* 0.5800 log-~~elih~ -266.77 chi-squared 119.83* notes: *significant at the .ol level. ** significant at the .05 level. variables are employed to determine if there is any variation in ranking based upon organizational type. since 13 bhcs reported losses for 1989, these institutions are excluded from the study. the rationale for this exclusion is that the dividend payout ratio would yield meaningless info~ation. the letter grade breakdown of these excluded institutions are as follows: nr equals 4, c = 3, c+ = 4, b = 1, and b+ = 1. table 3 reports the conclusions of the ordered response model. first, the null hypothesis that there is not any regional variation in the quality of bhcs cannot be rejected at a .0.5 level of confidence. additionally, there appears to be no evidence of qualitative differences based upon being classified as a regional, superregional, or money center financial institution. with respect to the other control variables, asset growth and primary capital are positively correlated and statistically significant at a .ol level of confidence. faster growing institutions tend to be the stronger entities, while retrenchment is the norm for management at weaker institutions. additionally, given the regulatory fiasco of the early 1980s in the thrift indus~y where regulators thought that the industry could grow its way out of its difficulties, it is unlikely that regulators would allow weak or marginal institutions to grow rapidly. also, those institutions with higher capital-to-asset ratios are better able to absorb losses and, hence, exhibit safer behavior. bank dividend policy as a signal of bank @dity 7 however, size is inversely related to quality and the null hypothesis of no relationship between size and quality can be rejected at a .ol level of confidence. this may reflect the problem that large bhcs had with their asset portfolio in the late 1980s as they wrote off non~~o~ng loans to developing countries and as the commercial real estate market moved into a recession. this may also reflect that larger banks engage in a greater variety of risk-taking activities than their smaller counterparts. finally, with respect to dividend policy, the results are consistent with expectations. first, the null hypothesis that dividends per share do not reveal any information about the quality of an institution can be rejected at the .ol level of confidence. there is a positive and significant relationship between dividends per share and bank quality rating. so, the empirical results are consistent with the the dividend signaling argument. high dividends per share, ceteris paribus, signal that the bank is healthy and expects to remain healthy. second, the null hypothesis that there does not exist a relationship between bank soundness and the dividend payout ratio can be rejected at an .05 level of confidence. as anticipated, dividends as a percent of earnings and quality are inversely related; however, the coefficient is small. a potential reason for this outcome is that the banking industry is unique in that retained earnings provide significant information concerning the future capital position of the industry. thus, as the dividend payout increases, the ability of an institution to expand its capital base is diminished, as is the quality or soundness of the bank. therefore, while dividends per share is the dominant factor, it should not be conside~d in isolation. the investing public needs to weigh both aspects of a bank’s dividend policy when selecting an investment target. v. summary this study examined the relationship between bank dividend policy and bank quality rating. empirical evidence shows that bank management uses its dividend policy as a vehicle for signaling its financial health to the investing public. using an ordered probit response model, this study found that a positive relationship exists between quality and dividends per share, which is consistent with the dividend signaling argument. additionally, an inverse relation ship was found between quality and dividends as a percent of earnings. the result clearly points to the facts that retained earnings are a key source of capital for bhcs and that capital position provides information about institutional soundness or, alternatively, risk. however, of the two measures, dividend per share provides a stronger signal. this study also shows that depositors, shareholders, and creditors can acquire informa tion about the overall quality of a bhc by examining both aspects of a bhc’s dividend policy. while dividends per share are important, they may not reveal a complete picture of the bank’s financial health. references aharony, j., & swary, 1. (1980). quarterly dividends and earnings announcements and shareholders’ returns: an empirical analysis. journal @finance, 35, l-12. barth, j., brumbaugh, rd., & litan, r. (1992). the future of american banking. new york: m.e. sharpe. 8 financial services review 4( 1) 1995 bar-yosef, s., & huffman, l. (1986). the informational content of dividends: a signalling approach. journal of financial and quantitative analysis, 21.47-58. bhattacharya, s. (1979). imperfect information, dividend policy, and the “bird in the hand fallacy.” bell journal of economics, io, 47-58. eades, k. (1982). empirical evidence on dividends as a signal of fii value. journal of financial and qualitative analysis, 17,471-500. ederington. l. (1985). classification models and bond ratings. financial review, 20,237-262. farnham, p., & cluff, g. (1985). a problem of discrete choice: moody’s municipal bond ratings. journal of economics and business, 37.277-302. federal deposit insurance corporation (fdic). (1985). the fdicstatistics on banking. washington, dc: fdic. federal deposit insurance corporation. (1991, first quarter). the fdic quarterly banking profile. washington, dc: fdic. federal deposit insurance corporation. (1994, first quarter). the fdic quarterly banking profile. washington, dc: fdic. kane, e. (1989). the high cost of incompletely funding fslic shortage of explicit capital. journal of economic perspective, 3, 3 l-47. keen, h. (1983). the impact of a dividend cut announcement on bank share prices. journal ofbank research, 13, 274-28 1. mayne, l. (1980). bank dividend policy and holding company affiliation. journal of financial and qualitative analysis, 15, 469-480. white, l. (1989). the reform of federal deposit insurance. journal ofeconomic perspective, 3,1 l-30. pii: 1057-0810(95)90009-8 financial services review, 4(2): 157-162 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. abstracts of articles on individual financial management edited by phyllis schiller myers virginia commonwealth university consumption-savings behavior credit cards and household demand for monetary assets: a cross-sectional study, by john v. duca (federal reserve bank of dallas) and william c. whitesell (board of governors of the federal reserve system) this study investigates credit card holdings and the household demands for several monetary assets in a simultaneous equations framework. it exploits the detailed data on household assets, as well as demographic and preference characteristics in the 1983 survey of consumer finances. a key finding is that, consistent with theory, a higher probability of credit card ownership implies lower demand for transaction balances with no effect on small time deposit balances. journal of money, credit and banking, may 1995,27(2): 604-623. (reprinted with permission of the journal of money, credit and banking.) are rising earnings profdes a forced-saving mechanism? by david neumark this paper tests the hypothesis that rising earnings profiles are a mechanism by which individuals engage in forced saving. it does this by examining the cross-sectional relation ship between overwithholding on income tax payments-behaviour that is consistent with a preference for forced saving-and the slopes of age-earnings profiles. the forced-saving hypothesis receives support from earnings regression estimates. individuals who receive tax refunds are on earnings profiles that are significantly steeper and have significantly lower intercepts. the economic journal, january 1995, 105(428): 95-106. (reprinted with per mission of the economic journal.) the dynamics of aggregate consumption in an open economy life cycle model, by j. ermisch and p. westaway this paper examines the dynamic path of aggregate consumption resulting from the relaxation of borrowing constraints and changes in the age composition of the population. 1.58 financial services review 4(2) 1995 using simulation analysis, it indicates that, because of overlapping generations, the adjust ment to a new steady-state aggregate consumption-income ratio following the relaxation of borrowing constraints is spread over a long time and it takes the form of a near linear fall following the immediate jump in the ratio. while potentially large, age composition changes have small effects on the aggregate consumption-income ratio when plausible parameter values are assumed. scottish journal of political economy, may 1994, 41(2): 113-127. (reprinted with permission of the journal of economic literature.) investmentsandindiwdualportfoliomanagement optimal portfolio and consumption decisions in a stochastic environment with precommitment, by isaac ehrlich and william a. hamlen, jr. (state univer sity of new york, buffalo) in this paper we solve the stochastic portfolio-consumption control problem under the assumption that individuals follow precommitment strategies over finite intervals of time. this precommitment approach is an alternative to merton’s (1969) continuous-time stochas tic dynamic control problem which assumes instantaneous feedback and costless revisions of choices all along the time path. our solution to the problem is contrasted with that of merton and several contributions to the subject. we show that under precommitment individuals will tend to hold portfolios that are a function of their expected risk and return parameters, but are independent of their wealth levels and risk preferences. we also show that the intertemporal consumption growth path would be a relatively smooth function of the risk-free rate of return, time preference, and the coefficient of relative risk aversion, and independent of the portfolio’s risk parameters. the latter would influence only the initial consumption level. we derive a number of empirical implications of our analysis for both portfolio holding and consumption patterns. journal of economic dynamics and control, april 1995, 19(3): 457-480. (reprinted with permission of the north-holland publishing company .) individual decision making in an investment setting, by carl a. kogut (north east louisiana university) and owen r. phillips (university of wyoming, laramie) this paper examines individual decision making in an experimental investment setting. when individuals make payments sequentially in order to complete an investment project their decision to continue investing should, according to standard theory, be based on a comparison of the expected marginal costs and gains. past payments are sunk and should be ignored. our experimental subjects usually did not make the correct decision at the margin when positive returns were threatened. there was an overwhelming tendency to over-com mit resources. journal of economic behavior and organization, december 1994, 25(3): 459-47 1. (reprinted with permission of north-holland publishing company.) abstracts of articles on individwl f~~nc~~ ma~~ement 159 the challenges of investor communication: the case of cuc international, inc., by paul m. healy (sloan school of management, cambridge, ma) and krishna g. palepu (harvard university) we examine investor communication issues using the experience of cuc international. cuc had difficulty convincing investors that its marketing outlays were profitable invest ments, leading to stock misvaluation over an extended period. to resolve this problem, cuc adopted an accounting change and then underwent a leveraged recapitalization. sub sequently, it accelerated recap debt repayments and initiated a stock repurchase. cuc’s experience suggests that accounting reports are not always effective in investor communi cation. while financial signals are more effective, their impact is not as immediate as predicted by prior research. the cuc case suggests that investor communications is a rich area for future research. journal of financial ecorwmics, june 1995, 38(2): 11 l-140. (reprinted with permission of north-holland publishing company.) some models of the international capital market, by b. dumas the original feature of international portfolio theory is that it considers investors of different countries who look at returns differently from each other because of deviations from purchasing power parity. its purposes are to explain the different portfolio compositions of investors of different countries and to explain the structure of expected returns across securities worldwide. the author shows here that this theory fails on the first count (home equity bias) but may still have some potential on the second count. he also discusses some results of general-equilibrium models of the international economy that incorporate frictions. european eco~orn~~ levier, april 1994,38(3-4): 923-93 1. (reprinted with per~ssion of the journal of economic literature.) signaling theory and risk perception: an experimental study, by haim levy (hebrew university) and esther lazarovich-porat (university of florida and hebrew university) theoretical models provide unsatisfacto~ solutions to many issues in finance and economics. for some important issues, signaling theory bridges the gap between the theoretical solutions and the firm’s actual behavior. however, many of the signaling theories are difficult, if not impossible, to test empirically. in this paper, we provide experimental tests of one of the signaling theories. we test the hypothesis that the larger the propo~ion of ent~preneur p~icipation in a project, the higher is its stock prices, as determined in an auction. the results strongly support the hypothesis. journal ofeconomics and business, february 1995,47(1):39-56. (reprinted with permis sion of north-holland publishing company.) 160 financial services review 4(2) 1995 the impact of calls of preferred stock on common sha~holde~’ wealth, by archana hingorani, anil k. makhija and kuldeep shastri (university of pitts burgh) this paper examines the behavior of common stock prices around the announcement dates of calls of straight and convertible preferred stock. our resuhs indicate that ealis of in-the-money convertibles are associated with stock price increases. in addition, our results indicate that the negative price impact of in-the-money calls is related to the reduction in cash dividend payout induced by the conversion of preferred stock to common stock, and not related to whether the call is underwritten or not, and the “moneyness” of the convertible preferred. finally, we find that calls of straight preferred that are financed with debt result in an increase in stock price, while all other straight preferred calls have no impact on stock price. journal ofbunking and finance, december 1994,18(6): 1095-l 111. (reprinted with permission of north-holland publishing company.) forward exchange bias, hedging and the gains from international diversi fication of investment portfolios, by h. levy and k. c. lim the gains to the u.s. investor from international diversification of investment portfolios are examined for portfolio strategies that hedge and strategies that do not hedge exchange rate risk via the interbank forward market. using the sharpe performance index and stochastic dominance as performance measures, almost all the unhedged strategies outper formed the hedged strategies for 1985-88; the opposite held for 1981-84. the results are explained by the biasedness of forward rates in predicting future spot rates. journal of international money and finance, april 1994, 13(2): 159-170. (reprinted with permission of the journal of economic literature.) individual risk management and insukance decisions the effects of ~ousehoid charac~~sti~ on demand for insurance: a tobit anatysis, by vince e. showers and joyce a. shotick (bradley university) tobit analysis is used to analyze the impact of household characteristics on demand for total insurance. this approach examines the marginal change in demand for insurance, as well as the change in the probability of purchasing insurance. demand effects are dominated by the marginal impacts from existing purchasers of insurance. although income and number of earners are both positively related to the demand for insurance, the marginal effect from an increase in income is greater for single-earner households than for multi-earner house holds. also, as either family size or age increases, the marginal increase in insurance expenditure diminishes. the journal of risk and insurance, september 1994,61(3): 492 502. (reprinted with permission of the journal of risk and insurance.) abstracts of articles on individual financial management 161 real estate issues apartment rent, concessions and occupancy rates, by g. stacy sirmans (the florida state university), c. f. sirmans (university of connecticut) and john d. benjamin (the american university) this paper examines the effects of rental concessions on apartment rent and occupancy rates. using limited-information maximum likelihood estimation, equations for rent, occu pancy and concessions have a positive effect on both rent and occupancy rates. rental concessions seem to provide the landlord a means to collect higher average rent and at the same time to increase occupancy rates. the results also indicate that a negative relationship exists between rent and occupancy rates and that certain amenities, services and occupancy restrictions influence rent. the journal of real estate research, summer 1994, 9(3): 299-3 12. (reprinted with permission of the journal of real estate research.) the thrift crisis, mortgage-credit intermediation, and housing activity, by michael g. bradley, stuart a. gabriel, and mark e. wohar this paper evaluates the influence of periodic disruptions in the thrift industry on mortgage credit intermediation and housing activity. results of the analysis indicate that disruptions in the thrift industry during the early 1980s imposed real costs on the economy by reducing the amount of mortgage credit intermediation. estimation findings indicate a sizable mortgage interest rate premium stemming from the loss in thrift intermediation services during the early 1980s. those higher mortgage rates served to dampen housing demand, in turn exacerbating the cyclical decline. in contrast, empirical findings for the post-recession period of the 1980s indicates a severing of the link between mortgage interest rates and thrift provision of mortgage credit. journal of money, credit and banking, may 1995, 27(2): 476497. (reprinted with permission of the journal of money, credit and banking.) differentiated contracts, heterogeneous borrowers and the mortgage choice decision, by j. sa-aadu (university of iowa) and c. f. sirmans (university of connecticut) previous analyses of mortgage choice assume that arms are standard contracts and estimate bivariate models. we estimate a multinomial logit model that explicitly treats mortgages as differentiated contracts and provide insights on several important issues: the impact of price variables differs significantly across alternative mortgage contracts; bor rower heterogeneity, particularly mobility, influences the type of contract chosen; and borrowers respond to market conditions as expected when choosing between alternative mortgage contracts. journal of money, credit and banking, may 1995, 27(2): 498-510. (reprinted with permission of the journal of money, credit and banking.) 162 financial services review 4(2) 1995 dwellings for the severely mentally disabled and neighborhood property values: the details matter, by g. galster and y. williams this research investigates the effects of dwellings occupied exclusively by severely mentally disabled tenants on sale prices of nearby homes. hedonic price models are estimated for an exhaustive sample of single-family home sales from 1989 to first quarter 1992 in newark and mt. vernon, ohio. proximity within two blocks of rehabilitated dwellings occupied by severely mentally disabled tenants had no significant relationship with sales prices. prices of homes proximate to two small, newly constructed apartment complexes were 40 percent lower after the complexes opened, although those near three other similar apartment complexes were not. land economics, november 1994, 70(4): 466-477. (reprinted with permission of the journal of economic literature.) taxation the influence of ethical attitudes on taxpayer compliance, by phillip m. j. reekers (arizona state university), debra l. sanders (washington state univer sity, pullman) and stephen j. roark (missouri southern state college) the development of tax decision-making models has focused on economic and behav ioral factors affecting compliance. a possible explanatory factor that has been overlooked in these decision-making models is tax ethical beliefs. this study examines the influence of ethical beliefs on tax compliance decisions. specifically, the research empirically tests whether an individual’s ethical beliefs about tax compliance mediate withholding effects (overwithheld or tax due) and tax rate effects (low or high) in tax evasion decisions. the results indicate that tax ethics are highly significant in tax evasion decisions and may be a “missing variable” in decision making models. national tax journal, december 1994, xlvjj (4): 825-836. (reprinted with permission of the national tax journal.) pii: 1057-0810(91)90004-i i financial services review, 1(1):9-22 copyright @ 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. i the optimal allocation of pension fund assets: an individual’s perspective tom l. potts william reichenstein i&lis study asks which assets in individual should hold indirectly in a pension in order to maximize total portfolio return, where the totaiportfolio consists of assets held direct& outside of pensions pius the pro rata share of pension assets. yeshe answer depends upon whether the individual actively or passively manages the nonpension portion of the total portfolio. most individuals should place high-yield stocks in the pension fund. but active investors should place low-yield growth stocks in the pension, while very passive investors should place bonds in the pension fund. rei~tatement of capital gains excbon would usually make corporate bonds the optimalpens~on asset. i. introduction what is the optimal allocation of pension fund assets? the answer depends in part upon the objective function. bicksler and chen (1985), black (1980), tepper (1981), and sharpe (1976) take the perspective that it is to optimize the share price of the sponsoring corporation. others look to achieve a mean variance efficient portfolio. this study takes the broader perspective of an individu~ or employee who considers pension assets as only part of a total portfolio of pension and nonpension assets. tax structures influence the allocation of assets between pension and nonpension portions of the portfolio. the pension tax structure is almost always the preferred structure. however, penalties for withdrawing funds from pensions before retirement and limitations on the amount of funds that can be tax-deferred discourage individuals from placing all assets in pension tax structures. most individuals must decide which assets best utilize the pension tax structure. the critical decision for these individuals is whether to hold common stocks or bonds in the pension fund. tom l. potts* department of finance, hankamer school of business, baylor university, waco, tx 76798. william reichenstein* pat and thomas r. powers chair in investment management, hankamer school of business, baylor university, waco, tx 76798. 10 financial services review, l(1) 1991 the answer depends upon whether the i~d~yid~~~ actively or passively manages the nonpension portion of the total portfolio. most individuals should place stocks, especially high-yield stocks, in their pension fund. active investors should place low-yield growth stocks in their pensions, and very passive investors should place bonds in the pension fund. ii. basics of tax structures pension tax structures generally have three advantages and one potential disadvantage when compared to nonpension tax structures. the first advantage is the taxdeferral of the investment amount until withdrawal.’ the second is the tax deferral until withdrawal of investment return-interest, dividends, and capital gains. the third advantage stems from the common expectation that individuals will be in lower tax brackets upon retirement. a potential disadvantage is present due to a possible reinstatement of a capital gains exclusion; under a pension tax structure all capital gains are taxed upon withdrawal, while capital gains realized outside of a pension could benefit from the exclusion. iii. the setting we assume that at any time there is an optimal allocation of each individual’s total portfolio among debt, stocks, real estate, and other assets, and that all funds are to be held for post-retirement use. the allocation of funds varies among individuals according to risk-return preferences, desired investment horizons (reichenstein, 1986), and other factors. given this asset allocation decision and government constraints that effectively prevent all assets from being held in pensions, the question becomes which assets should be held indirectly in pension funds. the optimal strategy is defined as the one that maximizes total portfolio after-tax return. the pension should contain assets that maximize the difference between assets’after-tax returns if held in pensions and after-tax returns if held in nonpension accounts. three factors that influence the size of this after-tax returns differential include: the annual before-tax return on the asset, the proportion of the annual returns in the form of capital gains, and the speed with which capital gains are realized. the last two factors are important when one considers that unrealized capital gains do not benefit from the pension tax structure. consider, for example, a long-term investment in gold that is held passively until retirement. the tax on returns is identical whether the asset is held in a pension account or not. thus unrealized capital gains in a nonpension account have the same tax advantages as a pension asset.* the optimal allocation of pension fund assets 11 this raises the question whether debt or equity best captures the advantages of the pension tax structure. do higher returns on stocks more than offset the fact that part of their returns can be in the form of unrealized capital gains? the analysis in this paper indicates that most individuals maximize total portfolio after-tax returns by placing common stocks, especially high-yield stocks, in the pension fund. the higher returns on high-yield stocks compared to debt are expected to more than offset the fact that some of the return can be in the form of unrealized capital gains. exceptions to this rule would be individuals who either very actively or very passively manage the common stock portion of their nonpension portfolio. iv. models of pension and nonpension tax structures ferris and reichenstein (1988) present a flexible model of the after-tax value of an investment that is based on earlier work by doyle (1984). by changing parameters, the model can reflect after-tax returns on the same asset whether held in a pension or nonpension account. the asset that maximizes the difference between these annual returns is the one that benefits most from the pension tax structure. the model assumes an initial $1 investment of after-tax funds and i percent annual before-tax return on investment, with g percent of the i percent returns in the form of capital gains.3 the proportion of capital gains realized annually is p. we assume a tax rate of t in all years before withdrawal, tw in the year of withdrawal, an n-year investment horizon, and deposit and withdrawal of funds on january 1. the after-tax value of an original $1 investment in an asset held in a nonpension account, k, is: v* = (1 + r)” t,(l -p)gi[((l + r)” 1)/r] (1) where r = (1 tp)gi -i(1 t)(l g)i. the investment grows at r percent annually for ii years. the product (1 g)i is the current income (dividends or interest) of the asset, so (1 t) * (1 g)i represents the after-tax amount of current income. the capital gains portion of annual returns is gi. tax on realized annual returns is thus tpgi, and the after tax amount of realized and unrealized capital gains is (1 tp)gi. thus r is the sum of the after-tax returns in the form of current income and capital gains. the product to the right of the first minus sign in equation 1 represents taxes upon withdrawal in yearn on the accumulation of unrealized capital gains. the product gi is the capital gains portion of annual returns, and the dollar amount of each year’s unrealized capital gains is the product of (1 p)gi and 12 financial services review, l(1) 1991 the beginning-of-year investment amount. unrealized gains are $(i p)gi in year 1, and grow with the beginning-of-year investment amount at r percent annually. the expression in brackets is based on the future value of an annuity formula. average annual after-tax returns on a nonpension asset are returns on a pension asset are tax-deferred until withdrawal. the after tax value of an asset in a pension account, vp, is vp = (1 + 2)” t,[(l + z)” 11 the asset accumulates tax-free at the before-tax rate i until withdrawal in n years. the amount to the right of the minus sign represents taxes at rate tw on accumulated returns, which is the full amount of the withdrawal less the original $1 investment. average annual returns on the pension asset are (4) the tax advantage of holding an asset in a pension account over holding it in a nonpension account is negatively related to the unrealized capital gains on the asset. when all returns are in the form of capital gains (g = 1) and no gains are realized until retirement 0, = 0), there is no advantage to holding the asset in a pension; that is, equations 1 and 3 are identical for assets with p = 0 and g = 1. the difference between the returns on pension and nonpension assets (i$ rn) is positively related to the before-tax return i and the asset turnover rate p, and negatively related to the capital gains proportion g. v. analysis the optimal assets to place in the pension fund vary with the assumed asset turnover rates p in the nonpension portfolio. first we consider an investor with perhaps “typical” asset turnover rates, then investors practicing more active and more passive nonpension portfolio management strategies. average investor table 1 indicates the additional returns many investors can expect from holding assets in a pension instead of a nonpension account. the before-tax expected returns on treasury and corporate bonds are the approximate yields the optimal ahcation of pension fwd assets 13 table 1. returns differentials for “typical” investors %-fi asset i g p n=5 n = 15 n=30 treasury bond 0.087 0.00 * 0.0067 0.0114 0.0158 corporate bond 0.100 0.00 * 0.0081 0.0140 0.0191 high-yield stock 0.129 0.38 0.6 0.0098 0.0172 0.0227 average-yield stock 0.150 0.81 0.5 0.0085 0.0151 0.0195 zero-yield stock 0.163 1.00 0.4 0.0064 0.0115 0.0146 gold 0.131 1.00 0.00 0.0000 0.0000 0.0000 assumption: t = 0.28, tw = 0.23 notes: * any asset turnover rate p produces the same result. the returns differential, & r,,, is the difference between the n-year average annual after-tax returns from holding an asset in a pension tax structure versus a nonpension structure. before-tax annual returns are i, and g percent of the i percent returns is assumed to be in the form of capital gains. the annual portfolio turnover rate is p. the tax rate in the year in which funds are withdrawn from the pension is t,, and t is the tax rate in prior year. to-maturity at the time of this writing. the expected returns on high-yield stocks, average-yield stocks, and zero-yield stocks are based on a 6.3 percent market risk premium and average value line betas of 0.67 for utility stocks, 1.0 for average stocks, and 1.2 for zero-yield stocks.4 expected returns on gold are based on an average beta of 0.70 for gold stocks. investment horizons of 5, 15, and 30 years are considered to represent individuals with short-, intermediate-, and long-term investment horizons before the funds will be withdrawn from pension accounts. the planned investment horizon depends upon the time before funds will be withdrawn, which usually will exceed the years before retirement. a s-year old investor expecting to withdraw funds uniformly between ages 65 and 75 has a 15-year average investment horizon. the 28 percent tax rate in years before withdrawal reflects current marginal federal tax rates for many individuals. the 23 percent rate upon withdrawal is slightly below tax rates in prewithdrawal years and offers some tax timing benefit for all pension assets and the unrealized capital gains on nonpension assets. the proportion of returns in the form of capital gains g varies from zero, which reflects most debt instruments, to 1.0, which reflects investments in zero-yield stocks, gold, coins, art, and many other real assets.’ the capital gains proportions for high-yield and average-yield stocks are based on value line expected dividend yields on utility stocks and the average stock, respectively. 14 financial services review, l(1) 1991 table 1 uses turnover rates that may reflect those of an average investor. financial theory and empirical evidence indicate that the presence of unrealized capital gains tends to lock an investor into an asset.6 the tax structure has encouraged a slower realization of gains on assets with substantial capital gains benefits, and elimination of the 60 percent long-term capital gains exclusion will only strengthen this tendency. the structure of secondary markets for most low-yield assets-e.g., art, coins, real estate-also contributes to the negative association between g and p. the values of the asset turnover rates p in the tables reflect these infhrences. table 1 indicates for this “typical” investor that the advantages of the pension tax structure are greatest for high-yield stocks, followed by average yield stocks and corporate bonds. higher returns on high-yield and average yield stocks compared to bonds are expected to more than offset the fact that part of the returns is in the form of unrealized capital gains. the order of asset preference is not affected by the investment horizon, but the relative advantage of high-yield stocks over bonds increases with the length of the investment horizon. assets with low current yields and/or slow turnover rates such as gold, art, undeveloped real estate, and, to a lesser degree, low-yield stocks do not receive as large a benefit from the pension tax structure. the gold example confirms that passively held assets (p = 0) yielding no current income (g = 1 .o) do not benefit from the pension tax structure. individuals should keep assets with a substantial portion of return expected to accrue as unrealized gains in the nonpension portion of their portfolio. active investor we have noted that the decision on the optimal assets to place in the pension fund depends in part on the speed with which capital gains in the nonpension portfolio will be realized. consider the extreme example of an individual with very actively managed stocks and bonds in the nonpension portfolio, i.e.,p = 100 percent. the top half of table 2 shows that this individual maximizes the total portfolio expected return by holding the highest return, and thus highest risk, securities in the pension fund. high risk stocks are the optimal pension asset for this individual; in the table, these are zero-yield (e.g., growth) stocks, reflecting the negative correlation between dividend yields and beta risk.7 the bottom half of table 2 shows that an investor practicing a 70 percent annual stock turnover rate should hold stocks instead of bonds in the pension fund. the highest differential rp r, exists for the riskiest zero-yield stocks, but the differentials for all the stock categories are roughly comparable. the major conclusion to be drawn from table 2 is that active investors should place stocks, especially higher risk growth stocks, in the pension fund. the optimal allocation of pension fund assets 15 table 2. returns differentials for active investors r-r, asset i g p tl=z n = 15 n = 30 treasury bond corporate bond high-yield stock average-yield stock zero-yield stock treasury bond corporate bond high-yield stock average-yield stock zero-yield stock 0.087 * 1.0 0.0067 0.0114 0.0158 0.100 * 1.0 0.0081 0.0140 0.0191 0.129 * 1.0 0.0114 0.020 1 0.0266 0.150 * 1.0 0.0141 0.0249 0.0322 0.163 * 1.0 0.0158 0.0279 0.0357 0.087 0.00 * 0.0067 0.0114 0.0158 0.100 0.00 * 0.008 1 0.0140 0.0191 0.129 0.38 0.7 0.0102 0.0179 0.0237 0.150 0.81 0.7 0.0107 0.0191 0.0246 0.163 1 .oo 0.7 0.0112 0.0198 0.0253 notes: * any value produces the same result. the returns differential, rp il, is the difference between the n-year average annual after-tax returns from holding an asset in a pension tax structure versus a nonpension structure. before-tax annual returns are i, and g percent of the i percent returns is assumed to be in the form of capital gains. the annual portfolio turnover rate is p. the tax rate in the year in which funds are withdrawn from the pension is c,, and t is the tax rate in prior year. yaari and fabozzi (1985) are puzzled by the evidence that roughly half of equity funds under ira or keogh plans are in growth funds. this allocation is rational for active investors; the higher returns from these high risk stocks translates into larger expected tax benefits from the pension tax structure. passive investor table 3 shows that investors practicing passive investment strategies in their nonpension portfolios should keep corporate bonds or high-yield stocks in their pension fund. the top half of the table indicates that extremely passive investors (p = 0) should pl ace first corporate bonds followed by high-yield stocks and treasury bonds in the pension fund. average-yield stocks and low-yield stocks as well as gold, art, and undeveloped real estate should be held directly in the nonpension account. the lower half of the table indicates that a relatively passive investor @ = 0.3) should 1 p ace first high-yield stocks and then corporate bonds in the pension account. the & r, values for high-yield stocks and bonds are roughly comparable. this investor should place real assets and low-yield stocks in the nonpension portfolio. 16 financial services review, l(1) 1991 table 3. returns differentials for passive investors asset g-k i &? p n=_5 n = 15 n=30 treasury bond corporate bond high-yield stock average-yield stock zero-yield stock treasury bond corporate bond high-yield stock average-yield stock zero-yield stock 0.087 1.00 * 0.0067 0.0114 0.0158 0.100 1.00 * 0.0081 0.0140 0.0191 0.129 0.38 0.0 0.0076 0.0127 0.0168 0.150 0.81 0.163 1.00 0.0 0.0 0.0027 0.0000 0.087 1 .oo * 0.0067 0.100 1.00 * 0.0081 0.129 0.38 0.3 0.0093 0.150 0.81 0.3 0.0062 0.0111 0.0143 0.163 1.00 0.3 0.0049 0.0087 0.0110 0.0049 0.0000 0.0114 0.0140 0.0150 0.0063 0.0000 0.0158 0.0191 0.0198 notes: * any asset turnover ratep produces the same result. the returns differential, & il, is the difference between the n-year average annual after-tax returns from holding an asset in a pension tax structure versus a nonpension structure. before-tax annual returns are i, and g percent of the i percent returns is assumed to be in the form of capital gains. the annual portfolio turnover rate is p. the tax rate in the year in which funds are withdrawn from the pension is l, and t is the tax rate in prior year. critical turnover rates table 4 shows the asset turnover rate that equates & r, for each stock category with 4 r,, for corporate bonds by investment horizon. take the five-year investment horizon as an example. with a 19 percent annual turnover rate, the differential for high-yield stocks equals the corporate bond differential. this indicates that individuals should place corporate bonds in the pension portfolio instead of high-yield stocks only if they expect to realize less than 19 percent of the stocks’capital gains annually. the table shows that only the more passive individuals should keep corporate bonds instead of high-yield stocks in the pension fund. table 4 indicates furthermore that individuals practicing slower asset turnover rates (e.g., p < 0.5) generally will benefit more from placing bonds instead of averageor low-yield stocks in their pension fund. more active investors (p > 0.5) generally will benefit more by placing averageor low-yield stocks before corporate bonds in their pension fund. the length of the planned investment horizon has little effect on the asset that expects to benefit most from pension tax treatment. of course, the effect of a given rp r,, differential on retirement wealth increases with the planned investment horizon. this suggests that, other things the same, the tax benefits the o@iml ailocad*on of pension fundassets 17 table 4. critical asset turnover rates asset high-yield stock average-yield stock zero-yield stock fended ~nves~~~~ horizon 5 15 30 0.19 0.17 0.23 0.47 0.44 0.48 0.47 0.49 0.55 note: a critical turnover rate is the annual turnover rate in the nonpension portfolio that equates the expected tax benefits of the pension tax structure for each stock category with the benefits for corporate bonds. if the annual turnover rate on the stock portion of the nonpension portfolio is expected to be less than the critical rate, corporate bonds are expected to benefit most from the pension tax structure. at higher turnover rates, stocks are expected to benefit most from the pension tax structure. we point out are more important for individuals who are further from retirement. finally, figure 1 illustrates the sensitivity of the returns differential to the nonpension portfolio turnover rate. the turnover rate is most important for zero(or low-) yield stocks. this returns d~f~r~nti~ is zero for the extreme buy and-hold strategy, p = 0. at the other extreme, p =i 1, there are no unrealized gains and all returns are taxed annually. the high risk on low-yield stocks implies that active investors are expected to receive the largest tax advantage on low-yield stocks. t 0 0.2 0.4 0.6 0.6 1 6, turnover rate hlsh yield + ava. yield ++ zero yield figwe 1. sensitivity of returns diferentials to turnover rates. assumptions: z = 0.28, fw = 0.23, n = 30, and i and g values from table. 18 financial services review, l(1) 1991 the slopes on the averageand high-yield stocks in the graph are positive but flatter than for zero-yield stocks. the slopes reflect the lower potential for unrealized capital gains due to both higher dividends and lower returns. sensitivity to assumptions the values in the tables are based on specific tax rates. simulations not reported here, however, indicate that the investment implications prevail for a wide range of possible tax rates. the table values also assume that an average stock will produce 6.3 percent higher before-tax returns than treasury bonds, and 5 percent higher returns than corporate bonds. reducing these assumed risk premiums would improve the relative position of bonds in the pension fund. yet, there are reasons to believe that the average stock will promise more than a 5 percent risk premium over corporate bonds-the bonds discussed most prominently in the analyses. the 10 percent returns on corporate bonds assume no loss from default. if expected returns after allowance for default risk on corporate bonds equal the treasury rate, then stocks almost always will be the preferred asset in the pension fund.8 recent studies by ferris and reichenstein (1988) and yaari and fabozzi (1985) provide further evidence that high-yield stocks are often the optimal asset to place in the pension portfolio. they conclude that there exists a yield-tilted risk-return plane: high current yield assets offer higher before-tax returns for a given level of risk than low-yield assets. ferris and reichenstein argue that the capital gains tax advantages remaining after passage of the tax reform act should produce the yield-tilt. they estimate this yield-tilt to be about half as strong as before elimination of the 60 percent capital gains exclusion. yaari and fabozzi also conclude that the risk-return plane is yield-tilted, but for different reasons. if the risk-return plane is yield-tilted then the expected returns in the tables on high-yield stock, and to a lesser extent average-yield stock, are too low. expected returns on bonds are based on yields-to-maturity and already reflect this premium, if it exists. this implies that the expected benefits of placing high yield and average-yield stocks in the pension fund may be greater than the analysis in this paper suggests. vi. capital gains exclusion the table values assume the absence of a long-term capital gains exclusion. although this reflects the current tax code, there is discussion of reinstituting a capital gains exclusion. nonpension assets enjoy the benefit, but assets held the optimal allocation of pension fund assets 19 table 5. returns differentials with capital gains exclusion li$ rce asset treasury bond corporate bond high-yield stock average-yield stock zero-yield stock gold i g p 0.087 0.00 * 0.100 0.00 * 0.129 0.38 0.6 n=s n = is n = 30 0.0067 0.0114 0.0158 0.0081 0.0140 0.0191 0.0000 0.0107 0.0168 0.150 0.81 0.5 -0.0194 0.0003 0.0065 0.163 1.00 0.4 -0.0262 -0.0064 -0.0004 0.131 1.00 0.0 -0.0306 -0.0101 ~,~60 notes: * any asset turnover rate p produces the same result. the returns differential, & &d, is the difference between the n-year average annual after-tax returns from holding an asset in a pension tax structure versus a nonpension structure with 60 percent capital gains exclusion. before-tax annual returns are i, and g percent of the i percent returns is assumed to be in the form of capital gains. the annual portfolio turnover rate is p. the tax rate in the year in which funds are withdrawn from the pension is t ,.,, and t is the tax rate in prior year. in a pension do not receive this benefit. with this in mind, it is instructive to note how the exclusion affects the asset allocation decision. table 5 shows returns d~ferentials co~esponding to table 1, except that 60 percent of capital gains on nonpension assets are assumed to be tax free. addition of the capital gains exclusion produces several noteworthy changes. first, it becomes preferable to place corporate bonds instead of common stocks in the pension portfolio. second, the exclusion tends to make the asset allocation decision more important than under the present tax structure; differences among returns differentials for bonds and stocks generally are substantially larger in table 5 than under the current tax structure. third, the pension tax structure is seldom the preferred structure for real assets, low yield stocks, and average-yield stocks. figure 2 illustrates this last point. for a 1%year investment horizon, it shows that the returns differential on zero-yield stocks is almost always negative, and the differential for average-yield stocks is negative for less active investors. a negative differential exists when the tax advantages of the capital gain exclusion exceed the advantages of the pension tax structure. a comparison of the values in tables 1 and 5 suggests that there may be a substantial increase in the demand for pension funds. before the tax reform act, many assets performed best under a nonpension tax structure. under the current tax code, however, all assets we considered do at least as well, and often substantially better, if held in a pension fund.g in the absence of a capital gains exclusion, we expect substantial growth in tax-deferred annuities, 20 financial services review, l(1) 1991 rp-rold 1::: 0 0.2 0.4 0.6 0.8 1 p. turnover rate -high yield + avg. yield + zero yield figure 2. returns differentials and turnover rates with 60 percent capital gains exclusions. assumptions: t = 0.28, tw = 0.23, n = 15 and i and g values from table. vii. ~ummaryandimplicati~ns we have asked which assets an individual should hold indirectly in a pension in order to maximize total portfolio returns, where the total portfolio consists of assets held directly outside of pensions plus the pro rata share of pension assets. the answer depends upon whether the individual actively or passively manages the nonpension portion of the total portfolio. investors who practice annual common stock turnover rates of between about 20 and 50 percent can maximize the after-tax returns on the total portfolio by holding high-yield stocks, and then average-yield stocks and corporate bonds, in the pension portfolio. more active investors should place common stocks, especially low-yield growth stocks, in the pension fund. more passive investors should place high-yield stocks and corporate bonds in the pension fund, with the order of preference determined by their degree of passivity. reinstatement of a 60 percent capital gains exclusion would substantially change the optimal pension fund asset mix for most individuals by making corporate bonds the optimal pension asset. further, the asset allocation decision generally would be more important than under the current tax code. finally, in the presence of a capital gains exclusion many assets, especially real assets and low-yield stocks, prefer the nonpension tax structure. under the current tax code, almost all assets prefer the pension tax structure, and there is no regulatory limit on the amount of funds that can be the optinttd a uocation of jbnsien fund assets 21 placed in tax-deferred accounts. what then stops individuals from placing (nearly) all their assets in tax-deferred accounts? one limiting factor is the 10 percent penalty tax on early withdrawals of pension funds. a second factor may be the failure by many indi~du~s to recognize the comparative advantage of the pension tax structure in the absense of a capital gains exclusion. unless a capital gains exclusion is reinstated, we expect substantial growth in the amount of funds in tax-deferred accounts. it has been noted that pension managers are reluctant to purchase as much real estate and other real assets as their divers~cation benefits would suggest (fogler, 1984; brueggeman, chen, and thibodeau, 1984). we find in this study that this reluctance is well-justified. individuals can take advantage of the current tax structure best by directly holding art, gold, real estate, and other assets with substantial capital gains potential in the nonpension portfolio. the relative absence of real estate in pension funds probably reflects fund managers’ understanding that the major nonpension asset for most individuals is their personal residence. finally, our analysis supports the argument that individuals should be allowed to designate where funds in a defined contribution plan are invested. the optimal assets to place in the pension fund will depend upon individu~ characteristics, including individual risk-return preferences as well as individual management practices in the nonpension portfolio. there is no allocation of a defined contribution pension portfolio that will be optimal, or even approximately optimal, for all individuals. a~knowie~g~e~ts: the authors thank andrew h. chen, frank j. fabozzi, and two anonymous referees for their help, notes the portion of a withdrawal from a nondeductible ira that represents repayment of the original investment is nontaxable. this analysis ignores minor differences in the pension and nonpension tax structures. for example, the tax liability on the capital gain actually occurs at the withdrawal of funds from the pension, while it occurs with the sale of the nonpension asset. second, the nonpension tax structure probably is preferable ifan investor expects congress to reinstitute a long-term capital gain exclusion for nonpension assets. pension investments usually are made with before-tax dollars, compared to the after-tax dollars invested in nonpension assets, the advantage of investing before-tax dollars is well documented, and, as a rule, individuals should take full advantage of this tax feature. for individuals who have maximized this benefit, however, there remains a valid question of which assets should be held indirectly in pensions. the model ignores some details of the tax code. for example, currently only $3,~ of net capital losses are deductible each year. the 6.3 percent market risk premium is the approximate geometric average excess return of stocks less the treasury bond rate since the mid 1920s. see ibbotson and sinquefield, 1982. financial services review, l(1) 1991 22 5. 6. 7. 8. 9. all returns from rolling over short-term debt and buying floating rate bonds are in the form of interest, g = 0. bonds offer some capital gains benefits, but for the purposes of this analysis the assumption of g = 0 is a reasonable approximation. constantinides (1984) shows that the 60 percent long-term capital gains exclusion stimulated more rapid realization of long-term gains on volatile stocks. the remaining capital gains preferences since the tax reform act will encourage slower realization of gains. the values of p in the table reflect the fact that the larger the unrealized capital gain, the greater the tendency to lock an asset into a portfolio. see, for example, report to congress on the capital gains tax reductions of 1978. we exclude gold from the analysis because the secondary markets for gold and other low yield real assets such as art, coins, and undeveloped real estate are not conducive to rapid turnover rates. historic~ly, lower-rated bonds on average have produced a higher return net of default than highly-rated bonds. despite this evidence, some financial economists argue that a bond’s risk premium should be an unbiased estimate of the expected loss due to default. an individual’s personal residence is not one of the assets considered in this study, in part because it is illegal to hold it in a pension. notice, however, that the one-time $125,000 capital gain exclusion and deductibility of interest favor holding it outside of a pension tax structure, references bicksler, j.l., and a.h. chen. 1985. “the integration of insurance and taxes in corporate pension strategy,” journae of finance, 40: 943-955. black, f. 1980. “the tax consequences of long-run pension policy,” financial anaiysts journal, 3: 21-28. brueggeman, w. b., a. h. chen, and t.g. thibodeau. 1984. “real estate investment funds: performance and portfolio considerations,” areuea journal, 12: 333-354. constantinides, g.m. 1984. “optimal stock trading with personal taxes: implications for prices and the abnormal january return,” journal of financial economics, 18: 65-89. doyle, r.j. 1984. “iras and the capital gains tax effect,” financiuz analysts journul, 7: 417 427. ferris, k. and w. reichenstein. 1988. “a note on a tax-induced clientele effect and tax reform,” national tax journal, 41: 131-138. fogler, h. r. 1984. “20% in real estate: can theory justify it?” journal of portfolio management, 10: 6-13. ibbotson, r.g. and r.a. sinquefield. 1982. stocks, bonds, bills and zn~ution: t&e past and the future. ch~lottesville, va: financial analysts research foundation. reichenstein, w, 1986. “when stock is less risky than treasury bills.” financial analysts journal, 9: 71-75. u.s. treasury office of tax analysis. 1985. report to congress on the capital gains tax reductions of 1978. washington, dc: u.s. government printing office, september. sharpe, w. 1976. ‘corporate pension funding policy,” journal of financial economics, 10: 183-193. tepper, i. 1981. “taxation and corporate pension policy,” journal of finance, 36: i-13. yaari, u. and f.j. fabozzi. 1985. “why ira and keogh plans should avoid growth stocks,” journal of financial research, 8: 203-215. pii: s1057-0810(99)80014-6 financial services review, 7(1): 69-70 issn: 1057-0810 copyright © 1998 by jai press inc. all fights of reproduction in any form reserved. book, software and website reviews douglas r. kahl, editor university o f akron personal financial planning (2nd edition) kwok ho and chris robinson north york, ontario, canada: captus press inc.; 1996 (isbn 1-896691-18-8) reviewed by: douglas r. kahl, professor of finance, university of akron kwok ho and chris robinson's personal financial planning is an excellent textbook written for the canadian personal finance class. the text is thorough and rigorous and appropriate for upper division (junior or senior) finance students. the text is clearly designed for serious financial planning students who have some finance background and want a broad based and rigorous treatment of financial planning topics. the text would not be appropriate for a service course for non-business students having no finance back ground. the text is thoroughly grounded in canadian tax law, canadian financial institu tions and the canadian economic system. this makes the text far superior to most united states based personal finance texts for a canadian financial planning course, but makes the text inappropriate for a personal finance course in the united states. while not divided into sections, the text follows a standard format. the introductory chapters cover time value of money (including the financial calculator), goal setting, per sonal financial statements and the financial planning life cycle. the two tax chapters pro vide an excellent discussion of personal income taxes in canada and income tax planning. the fundamentals of risk management are covered in three chapters along with the basics of life, health and property insurance. credit is covered in two chapters split between con sumer credit and mortgages. the three investments chapters cover the basic principles of investments, types of investments and mutual funds. retirement and estate planning are covered in two chapters. a chapter on comprehensive financial planning pulls the book together at the end with a comprehensive long-term case analysis including a suggested solution and a section on report writing. the last chapter in the text appears to be more an appendix on stochastic models for financial planning than an integral part of the book. for those wanting to cover the stochastic models, the chapter might well be covered in sections throughout the course. the text as reviewed was in soft-cover and black-and-white. while less eye-catching than many other current texts in the field, it should also be much more cost effective for the 70 financial services review 7(i) 1998 students and provide the students with thorough, extensive, rigorous and professional cov erage of personal financial planning topics. each chapter begins with a clear well-orga nized statement of learning objectives. users of the text will find a thorough and extensive set of discussion questions and problems at the end of the chapter. many of the problems are tied to the texas instruments ba-35 financial calculator and the super rep for students software package. overall, ho and robinson have produced an excellent no-nonsense textbook for per sonal financial planning with a thoroughly canadian orientation. for anyone in the cana dian market wanting to teach a professionally oriented personal financial planning course to upper division business students, this textbook would be an excellent choice. pii: s1057-0810(97)90009-3 from the editor karen eilers lahey s. travis pritchett provides a thought provoking discussion of the projection of life insurance cash values based on historical investment experience that is helpful for both academics and financial planners. as a past president (1988-89) of the academy of finan cial services, his continued interest in our organization is very much appreciated. as edi tor, i would like to encourage other authors to provide this type of rigorous analysis of commonly used industry practices. “performance and persistence in money market fund returns” by dale l. domian and william reichenstein offers strong evidence that the one factor that is significant in money fund performance is the expense ratio. this factor leads to a persistence in fund per formance over time that makes it easy for the investor to select a fund that provides high returns. sandeep singh and william h. dresnack examine the individual investor’s choice between national municipal bond portfolios and state specific bond mutual funds in two states with high taxes, new york and california. their article entitled, “market knowl edge in managed municipal bond portfolios” utilizes previous research in this area to explain the difficulties that investors face in this type of investment. douglas r. kahl explains how he solves the problem of running a student-managed investment fund at an urban university. “the challenges and opportunities of student managed investment funds at metropolitan universities,” illustrates how important indus try financial contributions such as oak associates ltd. gift to ten higher educational insti tutions can be in helping students understand the investment process. “asset allocation and investment horizon,” by keith v. smith examines asset alloca tion recommendations by brokerage firms that appears quarterly in the wall street journal. he finds that a buy-and-hold strategy appears to produce a higher return than quarterly, semi-annual, and annual asset allocation adjustments. three very interesting books are reviewed in this issue. the first book, the millionaire next door, is written by two academics in marketing and reflects their survey data descrip tion of those who achieve millionaire status. the financial services revolution, looks at the changes in the industry during the last 20 years. lastly, harold evensky’s, wealth management, is a comprehensive guide to financial planning by a widely recognized pro fessional in the field. dede pahl has asked me to include an addendum to brian grinder’s article, “an over view of financial services resources on the internet,” which appeared in vol. 6 no. 2 of this journal. the addendum appears in the back of this issue. v finser_31-2_complete_issue financial literacy during a pandemic: does modality matter? greg filbecka,*, xin zhaob apenn state behrend, black school of business, 281 burke, erie, pa 16563, usa bpenn state behrend, black school of business, 256 burke, erie, pa 16563, usa abstract we compare the performance of financial literacy programs launched by the cfa society pittsburgh in residential settings (2017–2019) with virtual/hybrid programs during the covid-19 pandemic (2020-2021). pretest baseline knowledge assessment shows that female students scored lower on subjective and objective financial knowledge questions and self-esteem. however, the global pandemic did not impact the effectiveness of programs based on modality. students experienced a statistically significant improvement in all four assessment areas of financial literacy. the largest gains in subjective and financial knowledge center on retirement planning. objective knowledge and self-esteem improvements occur most in credit and inflation. female students experience more significant gains in subjective knowledge, objective knowledge, and self-esteem. © 2023 academy of financial services. all rights reserved. jel classifications: g53 keywords: financial literacy; survey; outreach; education; teaching modality 1. introduction the state of financial literacy and numeracy worldwide is in distress. academic studies affirm a significant lack of financial literacy exists across nearly all demographics. while financial literacy statistics are important, the implications of financial literacy and numeracy are far-reaching due to their impact on financial decisions. financial literacy measures the degree to which one understands critical financial concepts and possesses the ability and *corresponding author. tel.: +1-814-898-6549. e-mail address: mgf11@psu.edu (g. filbeck) 1057-0810/23/$ – see front matter © 2023 academy of financial services. all rights reserved. financial services review 31 (2023) 169–196 confidence to manage personal finances through appropriate, short-term decision-making and sound, long-range financial planning, while mindful of life events, and changing economic conditions. the covid-19 pandemic dominated headlines throughout 2020 and into 2021. high school teachers had to rapidly transform in-class content delivery into modalities that included synchronous and asynchronous online instruction. at-risk students were most likely to have limited access to the internet and technology to participate fully. many high schools with financial literacy components in their curriculum had to postpone such instruction to ensure coverage of the core subject matter. this study extends prior research on the effectiveness of financial literacy education by comparing the in-person effectiveness of a financial literacy campaign launched by the cfa society pittsburgh during 2017–2019 with online/hybrid instruction during 2020-2021. online/hybrid programs are defined as those that contain at least some virtual learning component to those programs that are delivered completely in an in-person environment. we conduct pretests to assess baseline knowledge for program participants before the program administration. overall, pretest results from financial literacy programs before and during the pandemic show no overall differences regarding baseline financial literacy subjective knowledge. however, we observe gender differences across pre-surveys as female students scored lower than male students on subjective and objective financial knowledge. females also exhibited lower self-esteem. following the program administration, we conduct posttests to assess gains in financial literacy due to the program. both male and female students experienced a statistically significant improvement in subjective and objective knowledge, behavior, and self-esteem. the most significant gains in subjective and financial knowledge center on retirement planning. objective knowledge and self-esteem improvements occur most in credit and inflation. female students experience greater gains in subjective and objective knowledge, and self-esteem than their male peers. students in both modalities experienced a statistically significant improvement in subjective and objective knowledge, behavior, and self-esteem. there is no statistical difference in improvements between modalities except in financial behavior. the following section outlines existing literature and our hypotheses, followed by sample and methodology, results, and conclusions. 2. literature review 2.1. financial literacy financial literacy is critical because the knowledge and skills enable the proper use, accumulation, increase, and management of incomes while directly affecting countries’ economies. in context, financial literacy refers to the knowledge of financial concepts and applications that are vitally important in everyday life (semercioglu & akcay, 2016). financial literacy programs primarily aim to understand individuals’ financial knowledge 170 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 better. semercioglu and akcay conclude that low-level financial literacy across numerous countries is associated with a lack of financial training for individuals in their traditional education experience. the united states offers an example of a financial literacy system needing improvement. only 49% of americans with a college education can answer basic questions regarding financial literacy (faulkner, 2017). faulkner believes that the observed lack of financial literacy is positively correlated because household spending in the united states has consistently ranked among the highest globally. essentially, lower financial literacy contributes to the higher spending habits of individuals. additionally, the average savings rate in the united states has hovered around 5%, while the recommended level according to the u.s. bureau of economic analysis is often double this value (faulkner, 2017). henager and cude (2016) note increased academic research focusing on financial literacy and renewed interest in financial education and related policy since the mid-2000s driven by an increase in school-based financial education programs. in addition, an increase in state mandates for financial education in high schools and the creation of entities (e.g., the financial literacy and education commission and the consumer financial protection bureau) addressing financial literacy illustrate an increase in attention to improving financial literacy across the country. henager and cude consider how financial knowledge correlates with short-term and long-term behavior segmented by age group. they discovered a positive correlation between financial literacy and short-term and long-term financial behavior. an example of short-term financial behavior is paying your monthly bills on time, while longterm behavior could be saving for retirement. de beckker, de witte, and van campenhout (2021) note that financial education does not result in better customer choices, despite demonstrating improvements in financial literacy. one purpose of implementing financial literacy programs across high schools is to increase students’ knowledge of basic financial concepts and make smarter financial decisions as an adult. brown, grigsby, van der klaauw, wen, and zafar (2016) consider the effects of exposure to financial training on early adults’ debt outcomes based on their high school’s inclusion of financial literacy training. brown et al. analyze large-scale changes in financial training exposure using a sample of young americans and their debt behaviors over the decade immediately following high school training to see if financial trading has had any positive impact on their behavior. they find that financial and quantitative education during high school has moderate implications for young adults’ financial decisions aged 19 to 29. digital financial literacy (dfl), the education of financial literacy through digital platforms, is likely to become an increasingly important aspect of education, especially in a post-covid-19 world. this transition means individuals can become more responsible for their financial planning, including retirement and other goals (morgan, huang, & trinh, 2019) via a digital learning environment. since the early 2000s, many technological advances have emerged across the globe that have impacted several countries’ structures, yet various educational systems have not adequately adapted to a digital learning environment. financial technology (fintech), using software, applications, and digital platforms to deliver financial services to consumers and businesses through digital devices such as smartphones, has become recognized as a promising tool to promote financial inclusion. inclusion g. filbeck and x. zhao / financial services review 31 (2023) 169–196 171 involves necessary access to financial products and services provided to excluded households and small firms. fintech is revolutionizing the financial services industry at a very rapid pace. views differ regarding the likely impact that fintech is expected to have on personal financial planning, well-being, and societal welfare (panos & wilson, 2020). many scholars have one question whether the implementation of fintech in the high school curriculum will have a material effect on students’ improvement of financial knowledge. they argue that financial literacy research should make financial education more effective through improved content design and delivery. the disadvantage of enabling a more user-friendly and easily accessible program online is that students might develop misconceptions of specific topics more often than if they were engaged in an in-person program. for instance, panos & wilson indicate that students had a persistent misunderstanding of the risk/reward tradeoff with derivatives, more so than any other topic due to the program’s online nature. if enough student participants form this misconception, then a possibility exists that other misconceptions related to basic financial knowledge could emerge. however, van alten, phielix, janssen, and kester (2019) find that in flipped classroom settings in which students are required to watch online videos, students achieved significantly higher assessed learning outcomes than students in traditional classrooms, assuming face-to-face class time was not reduced compared to nonflipped classrooms. their study did not have any direct evidence of differences. noetel, griffith, delaney, sanders, parker, del pozo cruz, and lonsdale (2020) find that substituting video for existing teaching methods leads to smaller improvements in student learning. in contrast, videos as a supplementary teaching tool led to stronger learning benefits. wolla (2017) reports that fdic research indicates that 80% of the states in the united states have currently adopted some form of personal finance education standard, up from 42% in 1998. while the numbers show an improvement, these states do not mandate financial education. with only 29.7% of schools offering financial education, the student’s need for financial education is not adequately satisfied. wolla, a member of the federal reserve bank of st. louis, notes that one of its resources, soar to savings, provides an online learning module that teaches essential personal finance and economics concepts. with a sample of 3,061 sets of pretest and post-test scores of students across 100 schools, wolla concludes that a statistically significant improvement exists in scores from both the overall student results and the school results. overall, the results confirm that the paired samples at both levels indicate that the soar to savings online module effectively increases financial knowledge among high school students. although traditional printed materials and in-person classroom-style workshops are most prevalent, technological advances have created online financial education opportunities in recent years (kim, russell, & schroeder, 2017). government studies, however, show that wolla’s (2017) findings are not broadly applied to all schools since most of the schools do not include online modules. while programs such as the soar to savings online module are gaining momentum in high schools, authors often omit relevant research and theoretical frameworks to develop such programs. kim et al. show an imbalance of the benefits provided by online financial literacy programs. while wolla (2017) and panos and wilson (2020) show evidence in favor of students’ increase in financial knowledge after financial literacy programs, other studies question the 172 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 methodological rigor and whether the “improvement” of student knowledge diminishes over time. kim et al. (2017) argue that wolla’s findings that programs do not adequately measure the students’ knowledge retention after the post-test was completed. a key implication described by kim et al. is that many financial education programs, such as soar to savings, lack an explicit theory to frame information delivery. this lack of framework explains why students lose retention of learned financial concepts over time. technology is integrated into everyday life, and offering online financial education may offer alternative and innovative ways to reach broader audiences. however, this process requires more than just publishing previously printed materials online. financial educators and practitioners should design effective and interactive online programs and tools that build knowledge, facilitate improved financial decision-making, and foster positive behavior change. kim et al. (2017) provide several recommendations to incorporate financial concepts into the school curriculum. first, teachers can use online tools and resources to extend in-person educational encounters by allowing for more flexibility and frequency between student and teacher. these digital resources can provide a more accessible way for students to get involved in the program. educators can also create websites that offer tools, lectures, webinars, videos, downloadable documents, activities, worksheets, and other resources for “ondemand” learning, which would be especially useful for schools that offer financial education programs as an extracurricular activity. second, standalone online educational programs should be tailored to specific financial behaviors and audiences. targeted programs (e.g., programs derived from surveys of student populations or social media campaigns for specific user groups) may be more effective than general financial education. third, multiple modalities such as blogs, online games, chatrooms, and smartphone apps could be employed by teachers to accommodate a range of various learning styles, which should increase students’ overall engagement with the program. fourth, educators should structure programs with reminders, alerts, and prompts to help students monitor their progress and keep them engaged throughout the program’s entire duration. cameron, calderwood, cox, lim, and yamaoka (2014) suggest that younger generations are poorly prepared for making potentially life-changing financial decisions. the previous research provides multiple opportunities to enhance financial education programs to increase individuals’ overall financial literacy knowledge. however, the most notable two findings are related to the modality (online/hybrid vs. in-person) and the program’s length. first, assuming the research discussed remains true, individuals enrolled in more extended financial education programs tend to retain more concepts as time passes. continual education improves students’ cognitive understanding and increases their financial attitude and application of financial concepts with everyday life. secondly, newer research conducted by panos and wilson (2020), wolla (2017), and kim et al. (2017) shows that online modules are effective in increasing students’ short-term financial literacy. although each question whether the modules have a lasting impact on their financial cognitive ability similar to what in-person education programs show. therefore, creating more extended financial education programs with the flexibility of in-person or online modules can provide the most efficient form of increasing students’ overall financial attitude, behavior, and literacy. in this study, we investigate two issues: first, we compare the effectiveness of the cfa society pittsburgh’s financial literacy program under a model of in-person instruction g. filbeck and x. zhao / financial services review 31 (2023) 169–196 173 (2017–2019) and virtual/hybrid instruction (2020-2021). second, we explore whether gender differences exist in in-person versus virtual/hybrid instruction effectiveness. we create preand post-surveys to assess four key factors of financial success: self-esteem, perceived (subjective) knowledge, behavior, and objective numeracy. the surveys reflect the work of filbeck, zhao, and pettner (2020), which were motivated by previous works by lusardi and mitchell (2008, 2011) and lusardi, mitchell, and curto (2010). the in-person instruction data are collected during 2017-2019, while the virtual/hybrid instruction data are during 2020-2021. this paper extends work done by filbeck et al. (2020) by including a teacher survey to assess the modality of the courses taught in 20202021. our hypotheses are as follows: h1: the virtual/hybrid implementation of financial literacy education will significantly differ in students’ performance compared to in-person learning. h2a: male students have better financial literacy in all four key areas before the survey in both modalities. h2b: female students will show a greater improvement in knowledge, at the statistically significant level, across both modalities than male students. based on the previous research, the first hypothesis statement should hold for the sample students in this survey. first, if the hypothesis holds true, in-person financial education programs will prove more efficient than a virtual/hybrid instruction mode. the hypothesis aligns with amagir et al.’s (2018) research, which finds that person-to-person financial education is positively associated with higher financial knowledge scores from their survey. additionally, panos and wilson (2020) find in their research that virtual/hybrid modes of financial education courses failed to teach students in certain financial topics, stating a lack of direct student-toteacher contact as the vital reason. these financial topics test students in their objective and subjective financial knowledge. our hypothesis h2a builds on prior research that has found a financial literacy gender gap. in particular, cupák, fessler, schneebaum, and silgoner (2018) find that women score lower than men on financial literacy, with a more pronounced gap in developed countries. additionally, preston and wright (2019) examine the financial literacy gap in australia. while the “human capital variables” (age and education) were not significant in explaining the gap, “labor market variables” (including sector and occupation) were significant in explaining the gap. one possible explanation for gender differences is how females respond to financial literacy questions compared to males. lusardi and mitchell (2014) indicate that females are disproportionately more likely than males to respond to a question with “i don’t know.” bucher-koenen, lusardi, alessie, and van rooij (2017) find that females tend to give themselves lower scores in self-assessed knowledge. they conclude that “i don’t know” reflects not simply the lack of knowledge but rather the lack of confidence in their possessed knowledge. riener and wagner (2017) find that girls are more likely to skip questions (“i don’t know”) than boys if they deem them difficult. they find differences cease when extrinsic rewards are provided. they argue that their findings are consistent with a stereotype threat explanation. support for a stereotype threat explanation hinges on the disappearance of the gender gap when the task’s difficulty is made less salient. the quantitative nature of 174 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 financial literacy may also offer a possible explanation for females skipping questions. saygin and atwater (2021) find that female test-takers skip significantly more questions than males in quantitative areas, while no differences exist in nonquantitative areas. overall, in-person classes would increase students’ financial knowledge and behavior, ultimately improving financial success and responsibility. our hypothesis h2b builds on the arguments that females and males would show a statistically significant difference in learning with virtual/hybrid courses than in-person, with females showing a more significant improvement in financial knowledge than males. pangestu and karnadi (2020) and gerrans and heaney (2019) support this hypothesis by concluding in their research that, since women are building from a lower financial literacy base, a more significant improvement is observed when compared to men’s improvement. 3. sample and methodology the cfa society pittsburgh has been active in financial literacy outreach for over a decade, with a financial literacy committee charged with curricular development and training. financial literacy leaders conduct act 48 training sessions in several pennsylvania locations and deliver hour-long presentations on core financial literacy concepts as requested. the cfa society pittsburgh provides participating high schools with instructional materials for a semester-long equivalent course based on the missing semester (kabala & natali 2012) and the missing second semester (2020). schools voluntarily participate and deliver their programs based on their schedule. participating teachers are supplied with powerpoint resources to accompany the book, along with a web-based portal (available through the cfa society pittsburgh website) of best practices and exercises submitted from previous participating schools. program participation is open to any high school. each school determines the nature of program delivery, which ranges from within the context of a subject class (such as math) or as a standalone unit. the teachers decide at what point in high school to offer the programs. in most cases, programs are delivered as for-credit courses. after gathering the list of enrolled teachers, we assigned each a unique class code. links for preand post-surveys were provided to the participating teachers. in the introductory email, instructors were given directions to assign each student a unique id number, allowing preand post-surveys to be matched for analysis. participating schools agreed to administer the pre-survey before any instructional delivery. post-surveys were completed within a week after completing the last instructional unit on financial literacy. the participants of surveys that study the effectiveness of financial literacy education consisted of 2,297 students (1,622 in 2017-2019 and 675 in 2020-2021) across 108 schools (79 in 2017-2019 and 29 in 2020-2021). this sample is labeled the “whole sample.” a description of each survey follows: • pre-survey: distributed to students at the beginning of the financial education program by teachers that measure students’ initial financial knowledge in four main categories: subjective knowledge, objective knowledge, behavior, and self-esteem. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 175 • post-survey: distributed to students upon completion of the financial education program by teachers that measure the same four categories from the pre-survey to analyze. different input variables are used for quantitative questions measuring objective knowledge. we also conducted a teacher survey during 2020-2021 to obtain information about the individual financial education programs, including program length, modality, and resources used to teach the material. the surveys distributed were extensions of the work conducted by filbeck, zhao, and pettner (2020). the three surveys can be found in appendixes a, b, and c. table 1 reports the descriptive statistics for the pre-survey and post-survey samples. the pre-survey sample consists of 2,297 students completing the pre-survey, while the post-survey sample includes only 1,138 students who submitted both a preand post-survey.[1] of the pre-survey sample, 2,075 (90.3%) students are in their junior or senior year; in the postsurvey sample, 1,021 (89.7%) students are in their junior or senior year. female students account for approximately 47.0% and 48.6%, respectively, in the pre-survey and post-survey samples. additionally, students who participated during 2020-2021 make up 29.4% in the pre-survey sample and 26.8% in the post-survey sample. moreover, of the pre-survey (postsurvey) sample, 94.9% (98.0%) of students are in their junior or senior years participated during 2020-2021, while the corresponding number is 88.0% (86.7%). this finding suggests a higher percentage of students who participated during 2020-2021 with online/hybrid learning are in their junior or senior years. 3.1. survey methodology the surveys are modified based on filbeck, zhao, and pettner (2020) based on the framework established by tang and baker (2016), using path models to analyze the relationship table 1 sample description grade 9th 10th 11th 12th total panel a. whole sample year 2017–2019 female 36 39 147 536 758 male 54 59 194 557 864 year 2020–2021 female 11 4 200 105 320 male 3 16 213 123 355 total 104 118 754 1321 2297 panel b. test sample year 2017–2019 female 19 26 85 264 394 male 30 36 112 261 439 year 2020–2021 female 1 1 106 52 160 male 0 4 101 40 145 total 50 67 404 617 1138 note. table 1 shows the number of students across different grade levels and favorite subjects for the whole sample (panel a) and the test sample (panel b). 176 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 between characteristics and four financial literacy measures. the model shows independent variables produce both direct and indirect effects on a dependent variable. behavior (subjective) is the total score for the questions in financial behavior (subjective financial knowledge) questions, while objective (idk) is the total scores for students who choose correct answers (“i don’t know” answers) for objective questions. we use financial behavior scores (behavior), subjective financial knowledge scores (subjective), correct answers in the objective questions (objective), and “i don’t know” answers in the objective questions (idk) as dependent variables and test the effect of the financial literacy program. detailed variable definitions can be found in appendix d. we introduce several control variables associated with our study. in all three path models, we include gender (female), grade level (sophomore, junior, and senior), favorite subject (english, math, and science), and gpa. following amagir et al. (2018), we also include learning method preferences (learning by doing [lbd], listening, discussing, and visual) as independent variables that may impact financial literacy program success. students’ preand post-survey questions have been divided into two major categories: financial behavior and financial knowledge. financial knowledge questions are further split into objective and subjective financial knowledge. the surveys consist of 21 questions: three financial behavior, six objective financial knowledge, and 12 subjective financial knowledge. subjective financial knowledge and financial behavior questions are rated on a 5-point scale ranging from 1 ¼ strongly disagree to 5 ¼ strongly agree. the three financial behavior questions are “i like to save money more than i like to spend it,” “i have a checking and/or a savings account,” and “i have conversations with my parents regarding personal finance.” subjective financial knowledge questions involve a perceived understanding of financial concepts. they include questions such as “i understand how to establish a financial plan,” or “i understand the process by which my parents/guardians make financial decisions.” five broad categories of financial literacy make up the survey questions: interest (numeracy), compound interest, inflation, and credit. however, objective financial knowledge questions are conducted with “right” or “wrong” answers. each objective financial knowledge question contains at least one wrong answer and the option to choose “i don’t know.” the questions are analyzed using two methods: willingness to answer and correctness. the first method, willingness to respond, assigns a score of 1 for an answer of “i don’t know” and 0 for any other answer. secondly, questions with the correctness method assign a score of 1 for each correct answer and 0 for any other answer. the teacher survey comprises five informational questions regarding how the instructors teach financial literacy courses in their respective high schools. of the five informative questions, two focus on the mode of instruction, while the other three focus on the program’s duration. questions regarding the mode include: “how are you teaching the material?” and “what methods did you use to teach the material?” concerning durational questions, these include: “how many total contact hours will you spend teaching financial literacy?” “how often will students receive financial literacy instruction?” and “what best describes the total length of your financial literacy instruction program.” for 2020-2021, according to the teacher’s survey results, teachers conducted their programs in an online/hybrid mode (instruction occurring in a combination of virtual and in-person formats). g. filbeck and x. zhao / financial services review 31 (2023) 169–196 177 4. results 4.1. pre-survey analysis for the t-test of the pre-survey responses, two characteristics are analyzed: student gender and the delivery mode of content. student gender is divided into subgroups (female or male), while the delivery mode is based on year: in-person (2017-2019) or virtual/hybrid (2020-2021). table 2 lists the initial results of the pre-surveys submitted by students categorized by the four areas of financial knowledge tested: subjective knowledge, objective knowledge, financial behavior, and self-esteem. we only report the total scores for each area of financial knowledge in table 2 as cronbach’s a for each area of financial literacy ranges between 0.71-0.88, which is at an acceptable level of reliability and suggests the category totals are consistent measures of each question responses in that category. the detailed responses for each question are reported in appendix e: table 2a. the average responses are compared to the individual subgroups of gender and the courses’ modality. compared to the total average response score for subjective financial knowledge questions, female students scored lower than males (statistically significant at the 1% level). students who learned virtually/hybrid in 2020-2021 show a slightly higher score than in-person learning, but the difference is not statistically significant. for financial behavior, students who partook in a virtual/hybrid format are better financially behaved (statistically significant at the 1% level), while no statistically significant differences exist based on table 2 differences based on student characteristics: pre-survey results question type average response gender modality male female in-person online/hybrid panel a. financial subjective knowledge questions total score 33.928 34.571 33.196*** 33.893 34.010 panel b. financial behavior questions total score 14.895 14.835 14.964 14.746 15.254*** panel c. objective questions (correct answers) total score 2.982 3.253 2.707*** 2.934 3.100** panel d. objective questions (“i don’t know” answers) total score 1.667 1.411 1.977*** 1.717 1.545** note. table 2 shows the differences of pre-survey student responses on financial knowledge, financial behavioral, and objective questions across different gender and gpas for the whole sample. we only report total scores for each category as cronbach’s alpha for each category ranges between 0.71-0.88, which is at an acceptable level for internal consistency. detailed responses for each question can be found in appendix: table 2a. gender is identified as a dummy variable equal to 1 if the student is female and 0 otherwise. modality is denoted as a dummy variable equal to one if the sample comes from an online/hybrid modality, the 2020-2021 sample, otherwise, the dummy variable is zero, indicating an in-person structure based on the sample from 2017 to 2019. we also conduct a t-test to test the differences in gender (i.e., male vs. female) and modality (in-person vs. online/hybrid). we use asterisks to show the significance of the t-test result. ***, **, * indicate statistical significance at 0.01, 0.05, and 0.10 level, respectively. 178 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 gender. females scored lower in correctness and self-esteem regarding objective financial knowledge, while virtually/hybrid trained students scored higher in correctness and selfesteem (statistically significant at the 1% level). this finding is consistent with our h2a in that male students have better baseline financial literacy than female students before the survey in all areas except for financial behavior. our results are consistent with riener and wagner (2017) and saygin and atwater (2021) and may support the stereotype threat explanation. this explanation would argue the gender gap vanishes if the task’s difficulty is made less salient (riener & wagner, 2017). table 3 reports pearson correlations between financial behavior and self-esteem (idk), subjective, and objective financial knowledge for both the whole sample and the two subsamples. financial behavior, objective, and subjective knowledge are significantly correlated with each other. idk is significantly negatively correlated with the other three measures of financial literacy, which indicates a positive correlation between self-esteem level and the other three measures of financial literacy. naturally, because of the setup of the self-esteem measure, we find a negative correlation ("0.76) between idk answers and objective knowledge. comparing in-person and virtually/hybrid learning modality, we do not observe significant differences in correlations between these four variables during 2017-2019 and 20202021. 4.2. post-survey analysis the results of the preand post-surveys are compared using the post-survey sample of 833 students in 2017-2019 (in-person) and 305 students in 2020-2021 (virtual/hybrid). improvement is defined in several ways. gains from subjective financial knowledge, objective financial knowledge, and financial behavior are defined as the post-survey scores minus the pre-survey response scores. to gauge financial self-esteem, we define confidence gains as a decrease in “i don’t know” responses in the post-survey minus the pre-survey. in other words, students exhibit better self-esteem when they have fewer “i don’t know” answers in the post-survey compared to the pre-survey. table 4 illustrates the t-test results by question and overall score for each of the four items measured: subjective financial knowledge, financial behavior, objective financial knowledge, and financial self-esteem. we report results of total scores of each financial literacy category in table 4 and detailed results for each question in appendix f: table 4a. questions from subjective knowledge and financial behavior are assessed based on 1-5, 1 ¼ i highly disagree and 5 ¼ i highly agree. the results in table 4 and appendix f: table 4a show improvements with nearly every question in all categories, per question and total. for example, the financial behavior question “i like to save money more than i like to spend it” had an average pre-survey score of 3.54 and a post-survey score of 3.81. the increase shows that, on average, students’ preference for the balance between saving and spending improved, improving their overall financial behavior. each of the improvements is statistically significant at the 1% level, except for one question in the financial behavior category “i have a checking and/or savings account.” given the covid-19 pandemic that occurred in g. filbeck and x. zhao / financial services review 31 (2023) 169–196 179 2020-2021, the result for this question is reasonable given the reduction of in-person banking services. the most significant improvement of student responses to questions related to subjective financial knowledge comes from understanding roth ira (a gain of 1.653) and retirement (an increase of 1.376). all subjective financial knowledge questions are statistically table 3 correlation coefficients between financial behavior and self-esteem, objective and subjective financial knowledge question type subjective behavior objective idk answers panel a. whole sample subjective corr 1.000 p-value behavior corr 0.357*** 1.000 p-value <.0001 objective corr 0.286*** 0.267*** 1.000 p-value <.0001 <.0001 idk answers corr "0.364*** "0.267*** "0.759*** 1.000 p-value <.0001 <.0001 <.0001 panel b. year 2017–2019 subjective corr 1.000 p-value behavior corr 0.349*** 1.000 p-value <.0001 objective corr 0.291*** 0.255*** 1.000 p-value <.0001 <.0001 idk answers corr "0.366*** "0.255*** "0.753*** 1.000 p-value <.0001 <.0001 <.0001 panel c. year 2020–2021 subjective corr 1.000 p-value behavior corr 0.384*** 1.000 p-value <.0001 objective corr 0.276*** 0.293*** 1.000 p-value <.0001 <.0001 idk answers corr "0.359*** "0.292*** "0.774*** 1.000 p-value <.0001 <.0001 <.0001 note. table 3 shows the correlation coefficients between financial behavior, self-esteem (idk answers), and objective and subjective financial knowledge. ***, **, * indicate statistical significance at 0.01, 0.05, and 0.10 level, respectively. 180 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 significant at the 1% level. further analyzing the financial behavior questions’ responses, the biggest gain derives from the importance of contributing to a retirement plan (an increase of 0.402). all questions are statistically significant at least the 5% level. the results link confidence to answer a question (self-esteem) and correctness (objective financial knowledge). meanwhile, the most significant improvements seen from objective financial knowledge regarding questions involving credit, the concept of an agreement to purchase a product or service with the express promise to pay for it later (a gain of 0.251), and knowledge of inflation rate (a gain of 0.200). the largest increase in correct responses between preand post-surveys comes from self-esteem questions (how often students answered “i don’t know” on the objective financial questions). these two questions also show the greatest improvement in students’ self-esteem. that is, fewer students answered “i don’t know” with those two questions than any other between the preand post-surveys. next, we test h1 and h2b with subgroup results. table 5 reports the t-test results for different subgroups. panel a reports the subgroups by gender. the statistically significant improvement across all subgroups at the 1% level for both male and female groups indicates a significant improvement after completing the financial literacy program for both males and females. t-test results on the differences of improvements between males and females suggest that female students experience more significant improvements in financial behavior, objective knowledge, and self-esteem than their male peers. this finding is consistent with gerrans and heaney (2019) and h2b. panel b of table 5 reports the subgroup results by modality (in person, 2017-2019 vs. online/hybrid, 2020-2021). the statistically significant differences between preand postsurvey in all areas of financial literacy and both modalities show that students have significant improvements after the literacy program in both modalities. these findings reject h1 and are consistent with wolla’s (2017) finding that online learning effectively increases financial knowledge among high school students. further, students’ financial behavior gains are statistically significantly higher in 2020-2021 than the gains experienced during 20172019. however, this result should be interpreted with caution as the sample has a potential attrition issue because of covid-19 and will be tested with more robustness checks. next, we run regression analysis to examine how student characteristics and other control variables affect their knowledge and behavior gains. table 6 reports the regression results. table 4 t-test results between preand post-survey question type pre post diff t-stat panel a. financial subjective knowledge questions total score 33.839 43.963 10.125 36.38*** panel b. financial behavior questions total score 14.944 16.365 1.421 15.11*** panel c. objective questions (correct answers) total score 2.970 3.979 1.009 18.55*** panel d. objective questions (“i don’t know” answers) total score 1.690 0.482 "1.209 "22.43*** note. table 4 shows the t-test results of student responses to financial behavior and knowledge questions before and after the financial literacy educational efforts for the test sample. ***, **, * indicate statistical significance at 0.01, 0.05, and 0.10 level, respectively. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 181 t ab le 5 tte st r es u lt s b et w ee n p re an d p o st -s u rv ey fo r d if fe re n t su b sa m p le s s u b sa m p le ty p e f in an ci al b eh av io r s u b je ct iv e q u es ti o n s o b je ct iv e q u es ti o n s s el f es te em p re p o st p o st p re p re p o st p o st p re p re p o st p os t p re p re p o st p o st p re p an el a . s u b sa m p le s b y g en d er m al e (1 ) 3 4 .4 6 4 4 3 .6 0 2 9 .1 3 8 * * * 1 4 .9 01 1 6 .2 70 1 .3 6 9 * * * 3 .2 1 5 4 .0 7 2 0 .8 5 7 * * * 1 .4 1 0 0 .4 2 5 " 0 .9 8 5 * * * f em al e (2 ) 3 3 .1 7 7 4 4 .3 4 5 1 1 .1 6 8 * ** 1 4 .9 89 1 6 .4 66 2 .7 1 1 * * * 3 .8 8 1 3 .8 8 1 1 .1 7 0 * * * 1 .9 8 7 0 .5 4 2 " 1 .4 4 6 * * * (1 ) " (2 ) " 2 .0 3 0 * * * " 0 .1 0 8 " 0 .3 1 3 * * * 0 .4 6 1 * * * p an el b . s u b sa m p le s b y y ea r y ea r 2 0 1 7 -2 01 9 (3 ) 3 3 .7 8 0 4 3 .6 7 7 9 .8 9 7 * * * 1 4 .7 95 1 6 .3 34 1 .5 3 9 * * * 2 .9 3 6 3 .9 4 4 1 .0 0 7 * * * 1 .7 7 0 0 .5 5 5 " 1 .2 1 5 * * * y ea r 2 0 2 0 2 0 2 1 (4 ) 3 3 .9 9 7 4 4 .7 3 9 1 0 .7 4 3 * ** 1 5 .3 49 1 6 .4 50 1 .1 0 1 * * * 3 .0 6 2 4 .0 7 5 1 .0 1 3 * * * 1 .4 7 6 0 .2 8 3 " 1 .1 9 2 * * * (3 ) " (4 ) " 0 .5 5 3 * * * 0 .4 3 8 " 0 .0 0 6 " 0 .0 2 3 n o te . t ab le 5 sh o w s th e tte st re su lt s o f st u d en t re sp o n se s to se lf -e st ee m , fi n an ci al b eh av io r, su b je ct iv e, an d o b je ct iv e q u es ti o n s b ef o re an d af te r th e fi n an ci al li te ra cy ed u ca ti o n al ef fo rt s fo r d if fe re n t su b sa m p le s. * * * , * * , * in d ic at e st at is ti ca l si g n ifi ca n ce at 0 .0 1 , 0 .0 5 , an d 0 .1 0 le v el , re sp ec ti v el y . 182 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 table 6 regression results model (1) model (2) model (3) model (4) dep. var.: diff_behav dep. var.: diff_know dep. var: diff_obj dep. var.: diff_idk coefficient t-stat coefficient t-stat coefficient t-stat coefficient t-stat panel a. ols regression results on student and school district characteristics intercept 6.740 0.46 "20.697 "0.47 16.818 1.91* "8.520 "1.01 year 2020 "0.036 "0.13 2.945 3.61*** 0.311 1.89* "0.351 "2.23** female 0.022 0.10 1.537 2.28** 0.274 2.02** "0.460 "3.53*** upperclass "0.286 "0.76 1.606 1.45 "0.071 "0.32 0.120 0.56 gpa "0.162 "1.01 0.190 0.40 "0.149 "1.55 0.280 3.03*** log (population) "1.225 "4.42*** "6.492 "7.89*** "0.822 "4.96*** 0.741 4.66*** log (household_ income) 0.611 0.42 7.633 1.78* "0.695 "0.80 0.000 0.00 poverty "0.510 "0.05 31.186 1.09 "3.677 "0.64 "0.908 "0.16 pct_college 2.788 1.89* 11.935 2.73*** 2.369 2.69*** "1.949 "2.31** model (1) model (2) model (3) model (4) dep. var.: diff_behav dep. var.: diff_know dep. var: diff_obj dep. var.: diff_idk coefficient z-stat coefficient z-stat coefficient z-stat coefficient z-stat panel b. glm regression results after controlling for fixed effect of classes, with standard error clustered at the class level intercept 3.087 4.17*** 7.549 3.71*** 1.091 2.57** "1.662 "3.18*** year 2020 "0.782 "2.02** "1.062 "0.60 "0.055 "0.18 0.189 0.66 female 0.169 1.07 2.313 3.66*** 0.397 3.09*** "0.526 "4.71*** upperclass "0.359 "1.14 1.377 0.94 0.044 0.17 "0.020 "0.08 gpa "0.238 "1.69* 0.402 0.92 "0.073 "0.76 0.158 1.85* english 0.014 0.05 "1.250 "2.20** "0.231 "2.33** 0.071 0.54 math "0.082 "0.26 "0.138 "0.15 "0.247 "1.50 0.170 1.44 science 0.119 0.33 "0.582 "0.78 "0.248 "1.97** 0.184 1.67* lbd "0.207 "1.11 0.628 1.64 0.218 2.00** "0.116 "0.86 listening "0.271 "1.17 "1.638 "1.83* "0.236 "1.50 0.300 1.94* discussing "0.125 "0.35 "1.424 "1.54 "0.101 "0.55 0.143 0.98 visual "0.196 "1.13 0.675 1.04 0.209 1.39 "0.142 "1.10 note. table 6 shows the regression results of the test sample. panel a reports the ols regression results on student and school district characteristics. panel b reports the regression results on student characteristics after controlling for fixed effect of classes, with standard errors clustered at the class level. diff_behav (diff_subj, diff_obj, diff_idk) is the difference between the student’s preand post-survey scores for the financial behavior (knowledge, objective) questions. year2020 is a dummy variable equal to 1 if it belongs to the 2020-2021 sample. female is a dummy variable equal to 1 if the student is a female student and 0 otherwise. upperclass is a dummy variable equal to 1 if the student is a junior or senior and 0 otherwise. english (math, science) is a dummy variable equal to 1 if the student’s favorite subject is english (math, science) and 0 otherwise. lbd (listening, discussing, and visual) is a dummy variable equal to 1 if the student chooses learning by doing (listening, discussing with peers, and features visual support) as a favorite instruction method, and 0 otherwise. gpa is a student’s grade point average. log(population) is the log of the population in the school district. log(household_income) is the log of the median household income in the school district. poverty is the poverty rate in the school district. pct_college is the percentage of students with parents/guardians who have attained a bachelor’s degree or higher. ***, **, * indicate statistical significance at 0.01, 0.05, and 0.10 level, respectively. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 183 panel a reports the regression results on student characteristics and school district characteristics. the dependent variable is diff_behav (diff_subj, diff_obj, diff_idk), which is the difference between the students’ preand post-study scores (post-minus pre) for the financial behavior (subjective, objective) questions. all the other variables are listed in appendix d. the results show that students who participated in virtually/hybrid format and female students gained most in all areas except financial behavior. similarly, students in school districts with a smaller population and a higher percentage of parents/guardians with bachelor’s degrees experience a statistically significant gain in financial behavior. this finding is encouraging, as stolper and walter (2017) point out that the opportunity to relate financial literacy to various demographics in the context of their spending behavior is key to program success. in financial knowledge, we find students coming from higher-income households perform better. these results echo the findings of kaiser and menkhoff (2017), who find that financial literacy programs are less effective for clients with lower incomes and from lowand lower-middle-income economies. in such environments, altering financial behaviors in debt handling may be less effective. the results in panel a may be misleading on the conclusion of modality because schools with less ability to conduct financial literacy programs during the covid pandemic may not be participating in the 2020-2021 period. students who participated in the 2020-2021 program may come from a school with better infrastructure to adapt to the challenges of online learning. therefore, we need to control for school or class differences in our regressions. next, we include other control variables such as favorite subjects and learning style and use fixed effect regressions controlling for class differences. specifically, the class fixed effects allow the class dummy variable to differ and control for the variations across classes. we also cluster standard errors at the class level. table 1 shows a higher percentage of junior or senior students who participated in the 2020-2021 program. therefore, we add upperclass (i.e., a dummy variable equal to 1 if the student is a junior or senior) as an independent variable to control for grade level in the regression. panel b of table 6 reports the regression results. we use the same dependent variables as in panel a. the results show no statistically significant improvements except for financial behavior for students who participated in virtually/hybrid learning programs. this finding is consistent with table 5 and rejects h1. a trio of findings is notable for female students. first, female students whose favorite subject is not english tend to gain more subjective financial knowledge. in addition, female students whose favorite subject is not english or science and who prefer more hands-on learning opportunities (learning by doing) gain more objective financial knowledge. finally, female students with higher gpa experience the most significant self-esteem gain. results in tables 4-6 use test samples (including preand post-survey results from students who participated in both preand post-survey) and test the improvements (i.e., differences from preto post-survey scores) in students’ financial literacy. however, this setup may be subject to potential self-selection and attrition bias in our sample. specifically, from table 1, we can see that about 52% of female students and 51% of male students completed both preand post-survey during 2017-2019, while the corresponding numbers for the 20202021 sample are 50% and 41%, respectively. in other words, in the 2017-2019 sample, there 184 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 is roughly 51% to 52% attrition for both male and female students. while for the 2020-2021 sample, there is 50% attrition for female students, but about 59% attrition for male students. given that male students have higher baseline financial literacy and less scope for financial literacy improvement, this might bias the results of comparisons between the 2017-2019 sample and the 2020-2021 sample as presented in previous tables.[2] also, the sample is subject to self-selection bias as schools with less ability to conduct financial literacy programs during the covid pandemic may not be participating in the 2020-2021 period. students who participated in the 2020-2021 program may come from a school with better infrastructure to adapt to the challenges of online learning. to address this potential self-selection and attrition bias in our sample, we pool our whole pre-survey data (from all students who participated in pre-survey) and post-survey data. we use the standard diff-in-diff model and run regression models using outcome variables (i.e., behavior, subjective, objective, and idk) as dependent variables. we report the results in table 7. in our regressions, we add post, post # female, and post # year2020 as additional control variables. post is a dummy variable that equals 1 if the data are post-survey results and 0 otherwise. post# female is an interaction term and captures the change in the outcome variable (i.e., behavior, subjective, objective, and idk) from the pre-survey period to the postsurvey period for female students, relative to the change in outcome variable for male students. similarly, post # year2020 is an interaction term that captures the change in outcome variables from the pre-survey period to the post-survey period for the 2020-2021 program, relative to the change in outcome variable for the 2017-2019 program. other control variables are defined the same as in table 6 and appendix d. we also include student fixed effects to control student characteristics differences, with standard errors clustered at the student level. the statistically significant coefficient on post in table 7 suggests that students show statistically significant improvements in all four financial literacy areas in virtually/hybrid learning and in-person programs after controlling for potential attrition bias. moreover, the statistically insignificant regression coefficient on post # year2020 in table 7 using subjective, objective, and idk as dependent variable (i.e., models 2-4) suggest that there are no statistically significant differences of improvements in students subjective knowledge, objective knowledge, and self-esteem in virtual/hybrid format compared with the in-person format. this finding is consistent with our previous results in table 6 and rejects our h1. the statistically significant coefficients female in models 2-4 also confirm our previous results that male students have a higher level of financial literacy in all areas except for financial behavior than female students, which supports our h2a. more importantly, female students experienced the most gains in all financial literacy programs except for financial behavior, as evidenced by the statistically significant interaction term post # female. this supports our h2b. 5. conclusions this study’s primary purpose is to investigate the effectiveness of high school financial literacy education programs in a virtual/hybrid environment instead of an in-person setting. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 185 t ab le 7 r eg re ss io n re su lt s: d if fe re n ce -i n -d if fe re n ce m o d el m o d el (1 ) m o d el (2 ) m o d el (3 ) m o d el (4 ) d ep . v ar .: b eh av io r d ep . v ar .: s u b je ct iv e d ep . v ar : o b je ct iv e d ep . v ar .: id k c o ef fi ci en t z -s ta t c o ef fi ci en t z -s ta t c o ef fi ci en t z -s ta t c o ef fi ci en t z -s ta t in te rc ep t 1 1 .2 0 9 1 2 7 .5 1 * * * 3 2 .7 2 5 3 2 .4 3 * * * 1 .1 5 4 5 .3 4 * * * 2 .6 9 4 1 3 .1 8 * * * y ea r2 0 2 0 0 .7 0 2 3 .3 4 * * * 0 .9 1 9 1 .6 3 0 .1 8 2 1 .4 2 " 0 .2 1 7 " 1 .7 7 * p o st 1 .5 2 5 1 0 .0 3 * * * 8 .8 5 3 2 1 .7 0 * * * 0 .7 6 8 1 0 .2 6 * * * " 0 .9 2 7 " 1 4 .2 5 * * * p o st # y ea r2 0 2 0 " 0 .4 7 2 " 2 .4 6 * * 0 .7 1 4 1 .3 6 " 0 .0 2 4 " 0 .2 1 " 0 .0 9 4 " 1 .0 1 f em al e " 0 .0 4 7 " 0 .3 4 " 1 .4 6 2 " 4 .3 8 ** * " 0 .6 5 4 " 8 .4 4 * * * 0 .6 3 6 7 .8 8 * * * p o st # f em al e 0 .0 6 4 0 .3 5 2 .1 4 6 4 .1 9 ** * 0 .3 9 1 3 .8 3 * * * " 0 .4 5 0 " 4 .6 7 * * * u p p er cl as s 0 .7 5 1 3 .8 3 * * * " 0 .0 4 1 " 0 .0 9 0 .1 6 8 1 .6 1 " 0 .1 1 8 " 1 .1 3 g p a 0 .7 6 0 9 .3 7 * * * 0 .2 0 6 0 .9 9 0 .5 1 8 1 0 .6 8 * * * " 0 .3 2 5 " 7 .0 4 * * * e n g li sh " 0 .3 7 4 " 2 .1 0 * * " 0 .1 7 7 " 0 .4 1 " 0 .3 0 9 " 3 .0 3 * * * 0 .1 8 3 1 .9 7 * * m at h " 0 .2 2 7 " 1 .5 1 0 .0 2 4 0 .0 6 " 0 .0 6 2 " 0 .7 0 0 .0 3 9 0 .5 2 s ci en ce " 0 .3 6 1 " 2 .3 6 * * " 0 .1 4 3 " 0 .3 7 0 .0 8 8 1 .0 3 " 0 .0 5 9 " 0 .7 6 l b d 0 .4 2 8 3 .2 5 * * * 0 .8 7 4 2 .6 5 ** * 0 .2 6 8 3 .7 2 * * * " 0 .0 7 4 " 1 .1 6 l is te n in g 0 .0 6 1 0 .4 6 0 .6 7 7 1 .9 9 ** 0 .0 0 6 0 .0 8 " 0 .1 0 1 " 1 .4 3 d is cu ss in g 0 .0 7 7 0 .5 5 0 .5 9 5 1 .6 9 * 0 .0 1 5 0 .1 7 " 0 .0 1 2 " 0 .1 6 v is u al 0 .2 6 1 1 .7 0 * " 0 .1 0 5 " 0 .2 7 0 .0 8 7 0 .9 5 0 .0 3 6 0 .4 2 n o te . t ab le 7 sh o w s th e re g re ss io n re su lt s u si n g a d if fin -d if f m o d el . s p ec ifi ca ll y , w e u se o u tc o m e v ar ia b le s (i .e ., b eh av io r, s u b je ct iv e, o b je ct iv e, an d id k ) as d ep en d en t v ar ia b le s w it h p o o le d p re an d p o st -s ur v ey d at a. w e re p o rt th e re g re ss io n re su lt s o n st u d en t ch ar ac te ri st ic s af te r co n tr o ll in g fo r fi x ed ef fe ct o f st u d en ts , w it h st an d ar d er ro rs cl u st er ed at th e st u d en t le v el . b eh av io r (s u b je ct iv e, o b je ct iv e, id k ) is th e st u d en t’ s su rv ey sc o re s fo r th e fi na n ci al b eh av io r (k n o w le d g e, o b je ct iv e) q u es ti o n s. y ea r2 0 2 0 is a d u m m y v ar ia b le eq u al to 1 if it b el o n g s to th e 2 0 2 0 -2 02 1 sa m p le . p o st is a d u m m y v ar ia b le eq u al to 1 if th e su rv ey sc o re is fr o m p o st -s ur v ey an d 0 o th er w is e. p o st # y ea r2 0 20 is an in te ra ct io n te rm o f p o st an d y ea r2 0 2 0 . f em al e is a d u m m y v ar ia b le eq u al to 1 if th e st u d en t is a fe m al e st u d en t an d 0 o th er w is e. p o st x f em al e is an in te ra ct io n te rm o f p o st an d f em al e. u p p er cl as s is a d u m m y v ar ia b le eq u al to 1 if th e st u d en t is a ju n io r o r se n io r an d 0 o th er w is e. e n g li sh (m at h , s ci en ce ) is a d u m m y v ar ia b le eq u al to 1 if th e st u d en t’ s fa v o ri te su b je ct is e n g li sh (m at h, sc ie n ce ) an d 0 o th er w is e. l b d (l is te n in g , d is cu ss in g , an d v is u al ) is a d u m m y v ar ia b le eq u al to 1 if th e st u d en t ch o o se s le ar n in g b y d o in g (l is te n in g , d is cu ss in g w it h p ee rs , an d fe at u re s v is u al su p p o rt ) as a fa v o ri te in st ru ct io n m et h o d , an d 0 o th er w is e. g p a is a st u d en t’ s g ra d e p o in t av er ag e. l o g (p o p u la ti o n ) is th e lo g o f th e p o p u la ti o n in th e sc h o o l d is tr ic t. l o g (h o u se h ol d _ in co m e) is th e lo g o f th e m ed ia n h o u se h o ld in co m e in th e sc h o o l d is tr ic t. p o v er ty is th e p o v er ty ra te in th e sc h o o l d is tr ic t. p ct _ c o ll eg e is th e p er ce n ta ge o f st u d en ts w it h p ar en ts /g ua rd ia n s w h o h av e at ta in ed a b ac h el o r’ s d eg re e o r h ig h er . * * * , * * , * in d ic at e st at is ti ca l si g n ifi ca n ce at 0 .0 1 , 0 .0 5 an d 0 .1 0 le v el , re sp ec ti v el y . 186 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 financial literacy improvement is measured in four areas: subjective financial knowledge, financial behavior, objective financial knowledge, and self-esteem. the pre-survey results taken by students before beginning the financial education program are initially analyzed using a t-test. overall, the results show no statistical difference in presurvey financial subjective knowledge between the two subperiods based on modality. virtually/hybrid trained students scored higher in financial behavior, correctness, and selfesteem. however, female students scored lower on subjective and objective financial questions and exhibited lower self-esteem regarding objective financial knowledge. the only area in which females did not underperform males in the pre-survey was financial behavior. to test the effectiveness of financial literacy programs, a t-test was implemented between the preand post-survey results taken after completion of the course. the t-test analyzes the four major topic areas listed above. overall, male and female students experienced a statistically significant increase in all four topics. interestingly, the one area in which no differences occurred was whether students have a checking account. given the reduction in in-person banking during the pandemic, this finding is not surprising. with the four areas studied, subjective financial knowledge gains center on retirement planning. similarly, the greatest gains in financial behavioral questions are tied to the importance of contributing to a retirement plan. objective knowledge improvements and self-esteem improvements occur most credit and inflation. while statistically, significant improvement occurred for both genders, female students experience a greater improvement in financial behavior, objective knowledge, and self-esteem than male peers. modality does not appear to impact the effectiveness of the financial literacy programs as all areas of financial literacy and in both modalities show that students have significant improvements. thus, online learning is an effective tool for increasing financial literacy among high school students. students in school districts with smaller populations and a higher percentage of parents holding bachelor’s degrees experience a statistically significant gain in financial behavior. results should be interpreted with some caution as schools not participating in the financial literacy program during the pandemic may have been in more socioeconomically challenged areas than in more urban areas. based on the analysis extracted from the surveys, statistically, significant improvements in subjective financial knowledge, financial behavior, objective financial knowledge, and financial self-esteem lead us to conclude that the cfa society pittsburgh financial literacy program successfully increases students’ chances of financial success. the analysis affirms that one area the pandemic failed to impact was the effectiveness of financial literacy programs as no statistical differences exist between improvements based on modality. these findings add to the existing literature in financial literacy, demonstrating that educational techniques including online and hybrid learning, which were often required during the covid-19 pandemic, do not diminish effectiveness when used for financial literacy programs. based on these findings, we argue that school districts that currently do not require financial literacy components within their curriculum consider making changes based on the societal ramifications of such deficiencies and the significant improvements associated with successful financial literacy programs, such as the focus of our study. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 187 notes 1 to test whether there are selection biases between students who participated in both preand post-survey and students who participated only in pre-survey, we run t-test on the differences of students’ pre-survey results between students who participated in the post-survey and students who participated in the pre-survey but not post-survey. we find no statistically significant differences in their financial behavior, subjective and objective financial knowledge, and self-esteem in the pre-survey results between these two groups. the results are available upon request. 2 we thank an anonymous referee for valuable comments and suggestions on this issue. appendix a pre-survey questions class code: student id: gender: gpa: grade: favorite subject in school: ____ english ____ math ____ social studies ____ science questions: 1. i like to save money more than i like to spend it. 2. i understand how to establish a financial plan. 3. i think financial literacy is important for my future. 4. i have a checking and/or a savings account. 5. i have conversations with my parents regarding personal finance. 6. i understand the process by which my parents/guardians make financial decisions. 7. i know how to determine the appropriate total costs associated with the colleges/universities i am interested in attending. 8. i understand the process by which loan repayments take place including the impact of interest, delinquency, and default. 9. i understand the process by which credit card charges and repayment schedules can impact the level of financial debt levels. 10. when it comes to purchasing a car, i know how to determine how much of a car i can afford. 11. i understand how to evaluate the cost-benefit analysis of training for the job i would like to perform after completing school. 12. i know what a roth ira is and how it works from a taxation standpoint. 13. i know how to create a savings plan based on the ability to estimate monthly living expenses. 14. i know how to plan financially for retirement. 188 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 learning preferences: i am able to master material when instruction includes: 1. learning by doing/manipulating objects 2. listening 3. discussing with peers 4. features visual support (e.g., powerpoint slides) objective questions: 1. is it safer to put your money into one investment or put your money into multiple investments? 2. if you invest $100 in a roth ira and earn 10% per year for 3 years, how much would it be worth at the end of three years. 3. if you use a credit card in january for a total of $300, which payment option will result in the lowest amount of overall interest paid. 4. suppose you decide to buy a bmw for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? 5. in the future, the cost of things you buy doubles and your income also doubles. how much will you be able to buy in the future in comparison to today? 6. suppose you have $30,000 in student loans. which payment option would result in the lowest amount of overall interest paid? appendix b post-survey questions class code: student id: questions: 1. i like to save money more than i like to spend it. 2. i understand how to establish a financial plan. 3. i think financial literacy is important for my future. 4. i have a checking and/or a savings account. 5. i have conversations with my parents regarding personal finance. 6. i understand the process by which my parents/guardians make financial decisions. 7. i know how to determine the appropriate total costs associated with the colleges/universities i am interested in attending. 8. i understand the process by which loan repayments take place including the impact of interest, delinquency, and default. 9. i understand the process by which credit card charges and repayment schedules can impact the level of financial debt levels. 10. when it comes to purchasing a car, i know how to determine how much of a car i can afford. 11. i understand how to evaluate the cost-benefit analysis of training for the job i would like to perform after completing school. 12. i know what a roth ira is and how it works from a taxation standpoint. 13. i know how to create a savings plan based on the ability to estimate monthly living expenses. 14. i know how to plan financially for retirement. 15. i think it is important to contribute to a retirement plan (ex. roth ira, 401k, etc.) g. filbeck and x. zhao / financial services review 31 (2023) 169–196 189 learning preferences: i am able to master material when instruction includes: 1. learning by doing/manipulating objects 2. listening 3. discussing with peers 4. features visual support (e.g., powerpoint slides) objective questions: 1. which is less risky: investing your money into one investment or multiple investments? 2. if you invest $100 in a roth ira and earn 5% per year for 3 years, how much would it be worth at the end of three years. 3. if you use a credit card in january for a total of $500, which payment option will result in the lowest amount of overall interest paid. 4. suppose you decide to buy an audi for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? 5. in the future, the cost of things you buy doubles but your income remains the same. how much will you be able to buy in the future in comparison to today? 6. suppose you have $40,000 in student debt. which payment option would result in the lowest amount of overall interest paid? appendix c teacher survey questions class code: teacher id: gender: questions: 1. how are you teaching the material? 2. how many total contact hours will you spend teaching financial literacy? 3. how often will students receive financial literacy instruction? 4. what best describes the total length of your financial literacy instruction program? 5. what methods did you use to teach the material? (select all that apply). 6. did you take a finance-related course in high school or college? 7. what subject do you teach? 8. what suggestions do you have about the program that we can improve next year? 190 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 appendix d variable definitions dependent variables: subjective total score for the subjective questions in the survey objective total score for the objective questions in the survey with correct answers behavior total score for the financial behavior questions in the survey idk total score for the objective questions in the survey with “i don’t know” answers diff_subj the difference between the students’ preand post-study scores (post minus pre) for the subjective questions. diff_obj the difference between the students’ preand post-study scores (post minus pre) for the objective questions with correct answers. diff_behav the difference between the students’ preand post-study scores (post minus pre) for the financial behavior questions. diff_idk the difference between the students’ preand post-study scores (post minus pre) for the objective questions with “i don’t know” answers. independent variables: post a dummy variable that is equal to 1 if it is post-survey and 0 otherwise. post # year2020 an interaction term of post and year2020 dummy variables. post # female an interaction term of post and female dummy variables. sample year: year2020 a dummy variable that is equal to 1 if the survey is conducted in 2020-2021. gender: female a dummy variable that is equal to 1 if the student is a female and 0 otherwise. grade level: upperclass a dummy variable that is equal to 1 if the student is in grade 11 or 12 and 0 otherwise. favorite subject: english a dummy variable that is equal to 1 if the student’s favorite is english and 0 otherwise. math a dummy variable that is equal to 1 if the student’s favorite is math and 0 otherwise. science a dummy variable that is equal to 1 if the student’s favorite is science, and 0 otherwise. favorite learning style: lbd a dummy variable that is equal to 1 if the student’s favorite learning style is learning by doing (lbd) and 0 otherwise. listening a dummy variable that is equal to 1 if the student’s favorite learning style is listening and 0 otherwise. discussing a dummy variable that is equal to 1 if the student’s favorite learning style is discussion and 0 otherwise. visual a dummy variable that is equal to 1 if the student’s favorite learning style is visualization and 0 otherwise. gpa: gpa a student’s grade point average (gpa). school district characteristics: log (population) the log of the population in the school district. log (household_income) the log of the median household income in the school district. poverty the poverty rate in the school district. pct_college the percentage of residents who have attained bachelor degree or higher. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 191 appendix e table 2a differences based on student characteristics: pre-survey results average response gender modality male female inperson online/ hybrid panel a. financial subjective knowledge questions 2. i understand how to establish a financial plan. 3.014 3.083 2.937*** 3.039 2.956* 3. i think financial literacy is important for my future. 4.379 4.361 4.399 4.365 4.412 6. i understand the process by which my parents/ guardians make financial decisions. 3.360 3.408 3.306** 3.358 3.366 7. i know how to determine the appropriate total costs associated with the colleges/universities i am interested in attending. 3.044 3.050 3.037 3.086 2.942*** 8. i understand the process by which loan repayments take place including the impact of interest, delinquency and default. 2.758 2.879 2.620*** 2.743 2.792 9. i understand the process by which credit card charges and repayment schedules can impact the level of financial debt levels. 3.425 3.430 3.419 3.428 3.418 10. when it comes to purchasing a car, i know how to determine how much of a car i can afford. 3.288 3.403 3.156*** 3.287 3.290 11. i understand how to evaluate the cost-benefit analysis of training for the job i would like to perform after completing school. 3.022 3.115 2.917*** 3.015 3.038 12. i know what a roth ira is and how it works from a taxation standpoint. 1.986 2.093 1.865*** 1.951 2.072*** 13. i know how to create a savings plan based on the ability to estimate monthly living expenses. 3.127 3.160 3.090 3.148 3.077 14. i know how to plan financially for retirement. 2.629 2.748 2.494*** 2.621 2.647 total score for financial subjective knowledge questions 33.928 34.571 33.196*** 33.893 34.010 panel b. financial behavior questions 1. i like to save money more than i like to spend it. 3.534 3.587 3.473*** 3.505 3.603** 4. i have a checking and/or a savings account. 4.258 4.238 4.280 4.381 4.003*** 5. i have conversations with my parents regarding personal finance. 3.460 3.444 3.479 3.418 3.560*** 15. i think it is important to contribute to a retirement plan (ex: roth ira, 401k, etc.) 4.085 4.079 4.092 4.083 4.089 total score for financial behavior 14.895 14.835 14.964 14.746 15.254*** panel c. objective questions (correct answers) 1. is it safer to put your money into one investment or put your money into multiple investments? 0.635 0.699 0.569*** 0.622 0.665* 2. if you invest $100 in a roth ira and earn 10% per year for 3 years, how much would it be worth at the end of three years. 0.298 0.352 0.238*** 0.291 0.315 3. if you use a credit card in january for a total of $300, which payment option will result in the lowest amount of overall interest paid. 0.501 0.533 0.469*** 0.496 0.513 (continued on next page) 192 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 table 2a (continued) average response gender modality male female inperson online/ hybrid 4. suppose you decide to buy a bmw for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? 0.439 0.500 0.375*** 0.432 0.456 5. in the future, the cost of things you buy doubles and your income also doubles. how much will you be able to buy in the future in comparison to today? 0.567 0.602 0.534*** 0.560 0.585 6. suppose you have $30,000 in student loans. which payment option would result in the lowest amount of overall interest paid? 0.543 0.567 0.521** 0.533 0.567 total score for objective questions (correct answers) 2.982 3.253 2.707*** 2.934 3.100** panel d. objective questions (“i don’t know” answers) 1. is it safer to put your money into one investment or put your money into multiple investments? 0.251 0.191 0.322*** 0.263 0.222** 2. if you invest $100 in a roth ira and earn 10% per year for 3 years, how much would it be worth at the end of three years. 0.334 0.259 0.424*** 0.345 0.307* 3. if you use a credit card in january for a total of $300, which payment option will result in the lowest amount of overall interest paid. 0.337 0.321 0.358* 0.340 0.328 4. suppose you decide to buy a bmw for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? 0.290 0.226 0.366*** 0.295 0.278 5. in the future, the cost of things you buy doubles and your income also doubles. how much will you be able to buy in the future in comparison to today? 0.176 0.161 0.194** 0.186 0.151** 6. suppose you have $30,000 in student loans. which payment option would result in the lowest amount of overall interest paid? 0.279 0.252 0.313*** 0.287 0.260 0.000 total score for objective questions (“i don’t know” answers) 1.667 1.411 1.977*** 1.717 1.545** note. table 2a shows the differences of pre-survey student responses on financial knowledge, financial behavioral, and objective questions across different gender and gpas for the whole sample. gender is identified as a dummy variable equal to 1 if the student is female and 0 otherwise. modality is denoted as a dummy variable equal to one if the sample comes from an online/hybrid modality, the 2020-2021 sample, otherwise, the dummy variable is zero, indicating an in-person structure based on the sample from 2017 to 2019. we also conduct a ttest to test the differences in gender (i.e., male vs. female) and modality (in-person vs. online/hybrid). we use asterisks to show the significance of the t-test result. ***, **, * indicate statistical significance at 0.01, 0.05 and 0.10 level, respectively. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 193 appendix f table 4a t-test results between preand post-survey pre post diff. t-stat panel a. financial subjective knowledge questions 2. i understand how to establish a financial plan. 3.033 4.029 0.996 27.66*** 3. i think financial literacy is important for my future. 4.416 4.660 0.227 9.89*** 6. i understand the process by which my parents/guardians make financial decisions. 3.397 3.888 0.491 14.42*** 7. i know how to determine the appropriate total costs associated with the colleges/universities i am interested in attending. 3.085 3.861 0.776 20.46*** 8. i understand the process by which loan repayments take place, including the impact of interest, delinquency, and default. 2.716 3.864 1.148 28.88*** 9. i understand the process by which credit card charges and repayment schedules can impact the level of financial debt levels. 3.406 4.199 0.792 20.53*** 10. when it comes to purchasing a car, i know how to determine how much of a car i can afford. 3.258 4.197 0.940 24.74*** 11. i understand how to evaluate the cost-benefit analysis of training for the job i would like to perform after completing school. 3.010 3.889 0.879 22.70*** 12. i know what a roth ira is and how it works from a taxation standpoint. 1.909 3.562 1.653 38.31*** 13. i know how to create a savings plan based on the ability to estimate monthly living expenses. 3.087 4.168 1.079 28.21*** 14. i know how to plan financially for retirement. 2.581 3.953 1.376 33.45*** total score for financial subjective knowledge questions 33.839 43.963 10.125 36.38*** panel b. financial behavior questions 1. i like to save money more than i like to spend it. 3.540 3.810 0.269 8.09*** 4. i have a checking and/or a savings account. 4.290 4.267 0.085 2.41** 5. i have conversations with my parents regarding personal finance. 3.511 3.839 0.328 9.01*** 15. i think it is important to contribute to a retirement plan (ex: roth ira, 401k, etc.) 4.064 4.467 0.402 12.02*** total score for financial behavior 14.944 16.365 1.421 15.11*** panel c. objective questions (correct answers) 1. is it safer to put your money into one investment or put your money into multiple investments? 0.634 0.691 0.057 3.21*** 2. if you invest $100 in a roth ira and earn 5% per year for 3 years, how much would it be worth at the end of three years. 0.294 0.481 0.187 10.58*** 3. if you use a credit card in january for a total of $500, which payment option will result in the lowest amount of overall interest paid. 0.489 0.740 0.251 14.90*** 4. suppose you decide to buy an audi for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? 0.438 0.578 0.140 7.65*** 5. in the future, the cost of things you buy doubles but your income remains the same. how much will you be able to buy in the future in comparison to today? 0.573 0.773 0.200 12.07*** 6. suppose you have $40,000 in student debt. which payment option will result in the lowest amount of overall interest paid? 0.542 0.716 0.174 10.05*** (continued on next page) 194 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 references amagir, a., groot, w., maassen van den brink, h., & wilschut, a. (2018). a review of financial-literacy education programs for children and adolescents. citizenship, social and economics education, 17, 56–80. https:// doi.org/10.1177/2047173417719555 brown, m., grigsby, j., van der klaauw, w., wen, j., & zafar, b. (2016). financial education and the debt behavior of the young. the review of financial studies, 29, 2490–2522. https://doi.org/10.1093/rfs/hhw006 bucher-koenen, t., lusardi, a., alessie, r., & van rooij, m. (2017). how financially literate are women? an overview and new insights. journal of consumer affairs, 51, 255–283. https://doi.org/10.1111/joca.12121 cameron, m. p., calderwood, r., cox, a., lim, s., & yamaoka, m. (2014). factors associated with financial literacy among high school students in new zealand. international review of economics education, 16, 12–21. https://doi.org/10.1016/j.iree.2014.07.006 cupák, a., fessler, p., schneebaum, a., & silgoner, m. (2018). decomposing gender gaps in financial literacy: new international evidence. economics letters, 168, 102–106. https://doi.org/10.1016/j.econlet.2018.04.004 de beckker, k., de witte, k., & van campenhout, g. (2021). the effect of financial education on students’ consumer choices: evidence from a randomized experiment. journal of economic behavior & organization, 188, 962–976. https://doi.org/10.1016/j.jebo.2021.06.022 faulkner, a. (2017). financial literacy education in the united states: exploring popular personal finance literature. journal of librarianship and information science, 49, 287–298. https://doi.org/10.1177/0961000615616106 filbeck, g., zhao, x., & pettner, j. (2020). financial literacy: profiling a successful high school outreach program. financial services review, 28, 315–340. gerrans, p., & heaney, r. (2019). the impact of undergraduate personal finance education on individual financial literacy, attitudes, and intentions. accounting & finance, 59, 177–217. https://doi.org/10.1111/acfi.12247 henager, r., & cude, b. (2016). financial literacy and long-and short-term financial behavior in different age groups. journal of financial counseling and planning, 27, 3–19. https://doi.org/10.1891/1052-3073.27.1.3 table 4a (continued) pre post diff. t-stat total score for objective questions (correct answers) 2.970 3.979 1.009 18.55*** panel d. objective questions (“i don’t know” answers) 1. is it safer to put your money into one investment or put your money into multiple investments? 0.246 0.045 "0.201 "14.87*** 2. if you invest $100 in a roth ira and earn 5% per year for 3 years, how much would it be worth at the end of three years. 0.346 0.117 "0.230 "15.03*** 3. if you use a credit card in january for a total of $500, which payment option will result in the lowest amount of overall interest paid. 0.339 0.068 "0.270 "18.50*** 4. suppose you decide to buy an audi for $50,000. if you take out an auto loan for 5 years with 5% interest, how much total will you pay per year? 0.297 0.096 "0.201 "13.15*** 5. in the future, the cost of things you buy doubles but your income remains the same. how much will you be able to buy in the future in comparison to today? 0.172 0.066 "0.106 "8.68*** 6. suppose you have $40,000 in student debt. which payment option will result in the lowest amount of overall interest paid? 0.290 0.089 "0.201 "13.96*** total score for objective questions (“i don’t know” answers) 1.690 0.482 "1.209 "22.43*** note. table 4a shows the t-test results of student responses to financial behavior and knowledge questions before and after the financial literacy educational efforts for the test sample. ***, **, * indicate statistical significance at 0.01, 0.05, and 0.10 level, respectively. g. filbeck and x. zhao / financial services review 31 (2023) 169–196 195 kabala, m., & natali, g., jr., (2012). the missing semester. harriman house: petersfield, hampshire, uk. kaiser, t., & menkhoff, l. (2017). does financial education impact financial literacy and financial behavior, and if so, when? the world bank economic review, 31, 611–630. https://doi.org/10.1093/wber/lhx018 kim, j., russell, m. b., & schroeder, a. (2017). online financial education programs: theory, research, and recommendations. journal of human sciences and extension, 5. https://doi.org/10.54718/lwde8654 lusardi, a., & mitchell, o. (2008). planning and financial literacy: how do women fare? american economic review, 98, 413–417. https://doi.org/10.1257/aer.98.2.413 lusardi, a., & mitchell, o. (2011). financial literacy and planning: implications for retirement wellbeing. in o. s. mitchell & a. lusardi (eds.), financial literacy: implications for retirement security and the financial marketplace (pp. 17–39). oxford: oxford university press. lusardi, a., & mitchell, o. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. https://doi.org/10.1257/jel.52.1.5 lusardi, a., mitchell, o., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. https://doi.org/10.1111/j.1745-6606.2010.01173.x morgan, p., huang, b., & trinh, l. q. (2019). the need to promote digital financial literacy for the digital age, in adbi and jica. in k. yoshida & a. cardini (eds.), realizing education for all in the digital age (pp. 40–47). asian development bank institute, adbi press: tokyo. natali, g. (2020). the missing second semester. harriman house: petersfield, hampshire, uk. noetel, m., griffith, s., delaney, o., sanders, t., parker, p., del pozo cruz, b., & lonsdale, c. (2020). video improves learning in higher education: a systematic review. available at https://psyarxiv.com/kynez pangestu, s., & karnadi, e. (2020). the effects of financial literacy and materialism on the savings decision of generation z indonesians. cogent business & management, 7, 1743618. https://doi.org/10.1080/23311975.2020.1743618 panos, g., & wilson, j. (2020). financial literacy and responsible finance in the fintech era: capabilities and challenges. the european journal of finance, 26, 297–301. https://doi.org/10.1080/1351847x.2020.1717569 preston, a., & wright, r. (2019). understanding the gender gap in financial literacy: evidence from australia. economic record, 95, 1–29. https://doi.org/10.1111/1475-4932.12472 riener, g., & wagner, v. (2017). shying away from demanding tasks? experimental evidence on gender differences in answering multiple-choice questions. economics of education review, 59, 43–62. https://doi.org/ 10.1016/j.econedurev.2017.06.005 saygin, p. o., & atwater, a. (2021). gender differences in leaving questions blank on high-stakes standardized tests. economics of education review, 84, 102162. https://doi.org/10.1016/j.econedurev.2021.102162 semercioglu, m., & akcay, a. (2016). high school student financial literacy according to school type. economic and social development: book of proceedings, proceedings of the 15th international scientific conference on economic and social development – human resources development, varazdin 9-10, 2016, 88–92. stopler, o., & walter, a. (2017). financial literacy, financial advice, and financial behavior. journal of business economics, 87, 581–643. https://link.springer.com/article/10.1007/s11573-017-0853-9 tang, n., & baker, a. (2016). self-esteem, financial knowledge, and financial behavior. journal of economic psychology, 54, 164–176. https://doi.org/10.1016/j.joep.2016.04.005 van alten, d. c. d., phielix, c., janssen, j., & kester, l. (2019). effects of flipping the classroom on learning outcomes and satisfaction: a meta-analysis. educational research review, 28, 100281. https://doi.org/ 10.1016/j.edurev.2019.05.003 wolla, s. (2017). evaluating the effectiveness of an online module for increasing financial literacy. social studies research and practice, 12, 154–167. https://doi.org/10.1108/ssrp-04-2017-0014 196 g. filbeck and x. zhao / financial services review 31 (2023) 169–196 financial services review, 32(2) 29 esg perceptions: investigating investor motivations and characteristics yu zhang1 abstract integrating esg factors into investment strategies is a rapidly growing trend, but less is known about how investors value these esg factors. the characteristics of investors prioritizing esg in their decisions still need to be recognized. this study uses the value–belief–norm conceptual framework to investigate the relationship between socially responsible motivation and the perceived importance of esg when making investment decisions among investors in the united states. this study also explores the correlation among financial, sociodemographic, human capital, and economic variables and the perceived importance of esg factors. analyzing data from the 2021 national financial capability study (nfcs) state-by-state and investor survey through hierarchical regressions and segmentation analyses revealed that socially responsible motivation was significantly and positively linked to the likelihood of assigning greater importance to esg factors. variables such as objective and subjective investment knowledge, investment experience years, and information dependence on financial professionals emerged as significant factors. the segment analysis, which was differentiated based on the level of socially responsible motivation, further highlighted that financial-related variables are significantly associated with the importance placed on esg factors. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation zhang, y. (2024). esg perceptions: investigating investor motivations and characteristics. financial services review, 32(2), 29-52. introduction environmental, social, and governance (esg) investing has gained significant attention recently. at the start of 2022, institutional investors, money managers, and community investment institutions that incorporate environmental, social, and governance considerations into investment decisions and portfolio selections held approximately $7.6 trillion in us-domicile assets (us sif foundation, 2022). based on the report from the finra foundation, younger investors 1 corresponding author (yuliazhang@ksu.edu). kansas state university, manhattan, ks, usa exhibited a far higher inclination than older respondents to invest for motives other than longterm profitability (lin et al., 2022). specifically, younger respondents invested for social responsibility, entertainment, and social activity reasons at twice the proportions of those aged 55 and above (lin et al., 2022). current literature conducted on esg investing has focused on the relationship between esg factors and corporate financial performance (friede et al., 2015; kim & li, 2021), risk and opportunity management, and https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(2) 30 the relative importance of each criterion among institutional investors (park & jang, 2021). limited attention has been devoted to how individual investors adopt esg, and even fewer studies have been conducted on the driving force behind investors’ perspectives on esg factors when making investment decisions. this study aims to investigate the role of socially responsible motivation and financial profile variables such as objective and subjective investment knowledge, years of investment experience, information dependence on financial professionals, and risk tolerance on the valuation of esg criteria in the investment decision process. meanwhile, this study also controls the role of sociodemographic characteristics, human capital attributes, and economic variables on investors’ perceptions of the importance of esg factors, aiming to provide a comprehensive investigation into the multifaceted dynamics shaping the integration of esg considerations into investment strategies. the findings indicate that socially responsible motivation and financial variables play significant roles. this study represents an early attempt to make a valuable contribution to the existing body of literature by exploring variables connected to the prioritization of esg factors among individual investors. the paper seeks to fill the gap in understanding the determinants of investors’ perspectives on esg in the context of their decision-making, targeting an area that remains underexplored despite its growing importance in the way portfolios are managed. practical implications for financial planners, financial institutions, and policymakers are also discussed. conceptual framework and literature review overview of esg esg is a collection of standards that are essential for responsible investors to evaluate an organization/company’s performance in terms of environmental impact, social responsibility, and governance attributes. instead of exclusively prioritizing financial factors, esg investing involves the incorporation of environmental, social, and governance considerations when making investment decisions (mottola et al., 2022). based on the international framework of esg factors, environmental factors include greenhouse gas emissions, energy use and efficiency, air and water pollution, waste management, the effect on biodiversity, and the innovation of eco-friendly new products. social factors include the protection of customer privacy and workforce development, the prevention of child and forced labor, worker and consumer safety, diversity, antidiscrimination practices, poverty’s impact on the community, ethical supply chain management, and safeguarding customer information. governance factors include conduct codes, accountability and transparency procedures, executive compensation, board composition and diversity, antibribery and corruption regulations, stakeholder involvement, and shareholder rights (european banking authority, 2021). hence, adopting esg as an investment principle reflects a philosophical stance that seeks to promote holistic and sustainable development in society. however, based on the current literature, there is no commonly agreed-upon standard for assessing which companies are esgcompliant, resulting in a lack of consistency in esg investing (plastun et al., 2019). importance of esg investing global esg-related asset holdings surpassed $30 trillion as of 2022 and are anticipated to exceed $40 trillion by 2030 (bloomberg, 2024). the current state of study on the integration of esg information by individual investors is inadequate despite the increasing interest of these investors in esg investment. a recent study pointed out that individual investors, on average, anticipate that the monetary returns on esg will be lower than those of the broader equity market over a 10year period (giglio et al., 2023). however, when adopting a broader definition of return that includes nonmonetary components, holding stocks in companies with strong esg ratings could benefit individual investors who value social responsibility in their total return and profit nonfinancially (cornell, 2021). esg investing, when considering the risk associated with holding investments with high esg ratings, can potentially serve as a hedge against unforeseen environmental legislation and the impact of climate-related events (cornell, 2021). zhang 31 potentially, investors may perceive the reduced expected returns as a well-rounded result stemming from the attractive hedging attributes of esg stocks in mitigating potential future climate disasters or their advantageous nonfinancial attributes for investors with ethical concerns (giglio et al., 2023). for example, during the covid-19 pandemic, funds with higher esg ratings exhibited a capacity to outperform other funds (pisani & russo, 2021). similarly, cerqueti et al. (2021) found that, although esg funds generate lower returns outside of crises, they exhibit superior performance compared with conventional funds in times of turmoil. lee et al. (2020) also provided evidence that investment strategies involving portfolios with high esg ratings tend to perform better than those comprising portfolios with lower esg ratings. investors are generally prepared to pay a higher fee annually for funds with an esg mandate over identical funds lacking such a mandate, indicating that investors anticipate esg investments to generate competitive returns (baker et al., 2022). investing in esg not only aligns with investors’ values but also could generate competitive profits and make a beneficial impact on the environment and society. esg factors can be indicative of a company’s financial health, esg disclosure influences the financial, operational, and market performance measures of s&p 500-listed firms in the united states in a positive way, including return on assets (roa), return on equity (roe), and tobin’s q (alareeni & hamdan, 2020). firms characterized by high financial leverage and substantial assets were found to be more likely to disclose information pertaining to corporate governance, environmental matters, and social responsibility (alareeni & hamdan, 2020). similarly, kim and li (2021) discovered that esg factors positively influence the profitability of businesses, with this effect being more pronounced for larger companies. in general, the social dimensions of esg factors have the most significant and positive influence on a company’s credit rating, whereas the environmental dimensions have the opposite effect. the demand for esg integration in financial planning is increasing. esg investment has the potential to boost investors’ satisfaction by aligning their financial goals with their own values. with the growing focus on esg management, investors actively seek companies aligning with their preferred esg criteria. for example, due to the development of new laws and regulations focused on esg criteria in various countries, many european sovereign wealth funds and pension funds now have an obligation to provide information on their esg practices in response to the growing need (park & jang, 2021). aligning investment decisions with esg factors is consistent with long-term financial planning objectives, including but not limited to fostering sustainable growth, generating ethical wealth, and contributing positively to society (giglio et al., 2023). additionally, investing in esg can potentially enhance investors’ long-term financial well-being in a volatile pandemic market (mavlutova et al., 2021). value-belief-norm theory and esg in examining the factors associated with the prioritization of esg criteria in the investment decision-making process, this study is inspired by the value–belief–norm (vbn) theory. stern et al. (1999) introduced the vbn framework, which establishes the relationships among individual values, beliefs, and ethical norms. according to this theory, deeply held personal values can have a direct impact on one’s beliefs. these beliefs shape attitudes and behaviors that are consistent with an individual’s values and beliefs. while vbn theory is traditionally applied to explain pro-environmental behavior, its principles can be extended to the domain of consumer purchasing behaviors (lópez-mosquera & sánchez, 2012), consumer innovation adoption behaviors (jansson et al., 2011), and sustainable tourism (lind et al., 2015). these applications highlight the theory’s relevance in analyzing consumer behaviors and perceptions related to social and environmental concerns, which can be applied to investment perspectives and behaviors. this study proposes that individuals whose personal values align with socially responsible investment motivations may exhibit a vbn orientation, thus fostering a sense of moral obligation that is reflected in a higher tendency to consider integrating environmental, social, and governance factors in their investment decisions. financial services review, 32(2) 32 therefore, socially responsible investment motivation can manifest in the perceived importance of these esg factors when making investment decisions. the research found that political or social preferences have an influence on investment decisions. investors’ preference to choose particular investment options is strongly associated with their investment motivation, especially in socially responsible investing. for example, delmas and blass (2010) emphasize the significance of environmental and social preferences in choosing or avoiding company investments. investors who recognize investing opportunities that align with their personal values are far more likely to invest in socially responsible investment products, thus underscoring personal value on investment preference (bauer & smeets, 2015). previous research has also shown that esg investors’ portfolio decisions are influenced by their opinions about esg returns, motives for investing in esg, and concerns about climate change (giglio et al., 2023). investors pay increasing attention to socially responsible investing (cucinelli & soana, 2023), and the majority of individual investors who invest in esg mutual funds base their investment decisions primarily on ethical considerations, highlighting the pivotal role that personal values play in shaping their financial choices (giglio et al., 2023). the study incorporates social responsibility motivation variables to identify whether investors are driven by the desire to make a difference in the world, support values they care about, and be socially responsible. this study hypothesizes that: h1: investors motivated by social responsibility are more likely to rate esg factors as more important in their investment decisions. other determinants of esg investing previous research highlights the role of past financial knowledge to individuals’ esg investing practices. international studies have identified financial knowledge as one of the most important determinants for engaging in sustainable investments (cucinelli & soana, 2023). in particular, the existing body of literature concerning socially responsible investment is predominately based on objective financial knowledge. objective financial knowledge positively and substantially influences the intent to pursue sustainable investments and a strong preference for socially responsible financial intermediaries (cucinelli & soana, 2023; kar & patro, 2024). conversely, bauer and smeets (2015) employed a single-item measure of self-assessed investment knowledge. they noted a positive correlation between subjective financial knowledge and non-socially responsible investment accounts among netherlands investors. there is a lack of research on the relationship between financial knowledge and investment in esg factors in the united states. this study is one of the initial efforts to explore the relationship between investmentspecific financial knowledge and the perceived importance of esg factors in investment decisions among investors in the united states. esg investing possesses distinct risk-return dynamics and premium attributes (cornell, 2021; pisani & russo, 2021); potentially, individuals who possess or perceive themselves to have advanced investment knowledge might demonstrate a preference for esg factors. experienced investors or those relying on financial advisors for insights will also likely stay updated on current trends and regulations, which could influence their perceptions of esg. according to a study conducted on latvian investors, the likelihood of investing in assets that adhere to esg criteria is contingent upon factors such as net income, education attainment, financial literacy, and savings and investment experience (mavlutova et al., 2021). a higher proportion of younger respondents under 35 were more inclined to invest due to their aspiration to acquire knowledge about investment (lin et al., 2022). ethical investors are predominantly female, suggesting there may be a gender disparity in favor of ethical investment practices (tippet & leung, 2001). geographical and industry-specific variations exist in esg preferences across the united states (baker et al., 2022). more specifically, higher-income regions are more likely to have an esg investment option in 401(k) plans. areas with aging and highly educated populations are more likely to offer an zhang 33 esg investment choice in 401(k) retirement plans. however, limited studies are exploring what drives the significance of ratings based on esg factors among individual investors when making investment decisions. this study aims to determine whether there is a significant relationship between individual attributes and the perceived value of esg investing among investors in the united states. building on previous research, this study acknowledges the significance of finance-related variables. therefore, this study proposed that: h2: objective and subjective financial knowledge, investment experience, reliance on advisors for financial information, and risk tolerance are significantly associated with the valuation of esg factors in investment decision-making. methodology data the data utilized in the research were obtained from the 2021 national financial capability studies (nfcs) investor survey and state-bystate survey. the nfcs is an intensive, largescale research project conducted triennially by the finra investor education foundation since 2009. the investor survey delves more deeply into investing-related topics. the 2021 investor survey comprised 2,824 respondents who participated in the 2021 state-by-state survey and disclosed holdings in nonretirement accounts, providing substantial information regarding investment motivations, investment knowledge, and investment attitudes. after removing the “don’t know” and “prefer not to say” answers for the key variables of interest, the combined data set analyzed for this study now contains 2,324 investors. key variable esg importance. the dependent variable measured the importance investors placed on esg factor. this investment preference measure was built by answering the question, “how important is esg (environmental, social, and corporate governance issues) to you when making investment decisions?” possible answers range from 1 = not at all important to 10 = extremely important. social responsibility motivation. in the investor surveys, respondents were questioned about their investment motivations. specifically, they were presented with the statement “to make a difference in the world/support values i care about/be socially responsible” and asked to assess how well this statement describes their motivation for investing. responses indicating the statement “describes somewhat” or “describes very well” their motivation were coded as 1. conversely, if a respondent selected “does not describe at all,” this response was coded as 0. this variable was also dummy-coded into three levels to capture the gradation in respondents’ motivations. financial variables financial-related variables included objective investment knowledge and subjective investment knowledge. objective investment knowledge was measured by correctly answering 11 multiplechoice questions that objectively examined the respondents’ investment concept. the questions encompass a range of investment-related topics, including comprehension of fundamental concepts related to stock and bond ownership, evaluation of risks associated with different securities, analysis of investment returns, recognition of the benefits of index funds compared with actively managed funds, understanding of margin trading and short selling, and proficiency in calculating the value of call options. “don’t know” and “prefer not to say” were coded as incorrect answers for each item. subjective investment knowledge was assessed by asking respondents to self-rate their overall understanding of investing on a 7-point likert scale, with 1 indicating a very low level of knowledge and 7 indicating a very high level of knowledge. detailed information regarding investment knowledge is provided in the appendix. financial-related variables also included investment experience categories: starting investment less than a year ago, one year to less than two years ago, two years to less than five years ago, five years to less than 10 years ago, and 10 years ago or more (reference group), risk financial services review, 32(2) 34 tolerance level, and whether investors depend on guidance from financial professionals. the investors’ risk tolerance was assessed using a 10point likert scale, with a score of 1 indicating a weak willingness to take risks and a score of 10 indicating a strong willingness to take risks when considering their financial investments. the information dependence on financial professionals was assigned a value of 1 if investors admitted relying on financial professional recommendations when determining investment opportunities and a value of 0 otherwise. other variables to investigate the variables that may have a substantial association with esg investment preference, sociodemographic variables, human capital variables, and economic variables were included as control variables. sociodemographic variables included gender, racial group, marital status, the presence of financially dependent children, employment status, and age categories. human capital variables included educational attainment categories. economic variables included high investment account balance (higher than $100,000), homeownership, and income categories. analysis the dependent variable, the perceived importance of the esg factors on investing, was assessed using an ordinal scale ranging from 1 to 10, with each value denoting a unique level of importance. the dependent variable indicates a relative symmetric distribution with a mean of 4.823; however, the kurtosis value of 2.026 suggests a platykurtic distribution, which deviates from the normal distribution kurtosis. given the significant sample size, the central limit theorem helps mitigate the impact of nonnormality. consequently, an ordinary leastsquares linear regression analysis employing robust standard errors was performed. specifically, this study employed a two-step hierarchical cumulative order logistic regression approach to analyze the data. in the first step, the model included financial-related, sociodemographic, human capital, and economic variables to establish a baseline model. the second step of the model, i.e., social responsibility motivation, was introduced as an additional variable. by employing this methodology, this research can differentiate and assess the unique influence of social responsibility motivation on investors’ perception of the extent to which esg is important. 𝐸𝑆𝐺 = 𝛽0 + 𝛽1𝑋1 + 𝛽2𝑋2 + ⋯ + 𝛽𝑘𝑋𝑘 + 𝜖 (1) 𝐸𝑆𝐺 = 𝛽0 + 𝛽1𝑋1 + 𝛽2𝑋2 + ⋯ + 𝛽𝑘𝑋𝑘 + 𝛽𝑆𝑅𝑀𝑋𝑆𝑅𝑀 + 𝜖 (2) where: 𝛽0: the intercept; 𝛽1 … 𝛽𝑘: the coefficients for the independent variables (financial-related, sociodemographic, human capital, and economic variables) in step 1; 𝛽𝑆𝑅𝑀 coefficient for the additional socially responsible motivation variable, introduced in step 2; and 𝜖: error terms to elevate understanding of the perceived significance of esg factors, this study also segmented the analysis based on the level of social responsibility motivation among investors, categorizing them as “none,” “somewhat,” and “very well” motivated. as a result of the heterogeneous nature of having socially responsible investment motivations, the relationship between variables and the importance of esg may vary. consistent and significant variables may yield crucial insights when contrasting the outcomes of two distinct groups. additional analyses were conducted on subsample groups, specifically emphasizing gender, financial experience, and dependence on information provided by financial professionals. results descriptive results table 1 provides an overview of the characteristics of the sampled investors as well as a detailed breakdown of the two groups, differentiated by the presence and absence of socially responsible motivation. the average importance rating for esg in investment zhang 35 decisions was 4.792 on a 10-point likert scale. among the 2,324 investors surveyed, 41.61% somewhat or well identified with investing motivations related to making a difference, supporting personal values, or being socially responsible. when tested on 11 objective investment knowledge questions, investors typically answered five to six questions correctly. investors also rated their investment knowledge subjectively on the higher end of the 7-point likert scale, with a mean score of 4.904. investing experience varied among sampled investors, with 68.80% having over 10 years of experience, and only 3.87% being new investors with less than a year of experience. risk tolerance, measured on a 10-point likert scale, averaged at 6.207. demographically, the majority were male (63.94%), white (80.98%), and married (66.95%). singles accounted for 18.29% of the sample. financial dependents were present in 26.03% of investors’ households. employment status showed that 53.57% were employed at the time of surveying. age distribution was broad, with the highest representation from those aged 65 and older (41.52%) and the lowest from 18to 24-year-olds (2.88%). educational attainment revealed that 49.61% had a college or postgraduate degree. economically, 59.72% had nonretirement investment accounts exceeding $100,000. homeownership was common, with 84.94% owning homes. income varied, with the highest percentage (23.92%) earning between $100,000 and $150,000 annually. exploratory research demonstrated significant differences between individuals motivated by social responsibility and those who are not across various dimensions. this was evidenced by twosample t-tests for continuous variables, highlighting disparities in investment knowledge (objective: t = 8.768***; subjective: t = 7.555***) and risk tolerance level (t = -8.179***). additionally, two-sample tests for proportions on dummy coded variables revealed significant differences in gender (z = 3.478**). regression results results of the two-step hierarchical ols regression model can be found in table 2. all variance inflation factor (vif) values are below 5, indicating no issues with multicollinearity. in model 1, financial-related and sociodemographic variables showed significant relationships with the perceived importance of esg factors. a clear distinction was demonstrated between objective investment knowledge and subjective investment knowledge concerning the rating of esg importance level. objective investment knowledge score was negatively associated with the perceived importance of esg factors when making investment decisions. on the contrary, subjective investment knowledge was positively associated with the perceived importance of esg factors. investors with shorter investing periods tended to assign more importance to esg factors than those who started investing 10 years ago or more. investors who rely on recommendations from their personal financial professionals see esg factors as more important. as risk tolerance increases, so does the perceived importance of esg factors. employed individuals, those with financial dependents, and investors aged 18 to 24 were positively associated with importance rating to esg factors than their counterparts. men, white respondents, and widowed were associated with a lower perceived importance of esg factors. in model 2, socially responsible motivation has a substantial and positive association with the importance investors place on esg factors when making investment decisions. h1 was supported. consistent with the findings of baseline model 1, an inverse relationship existed between objective investment knowledge and the degree of importance assigned to esg factors during the investment process. objective investment knowledge was negatively associated with the perceived importance of esg factors, whereas subjective investment knowledge showed a positive association. compared with those with 10 years more investment experience, individuals with less than 10 years but more than one year of investment experience showed a positive association with assigning greater importance to esg factors when making investment decisions. reliance on a financial advisor’s recommendations and higher risk tolerance were associated with greater perceived importance of esg factors. sociodemographic variables, such as males, being widowed, and aged 45 to 54, financial services review, 32(2) 36 indicated a lower perceived importance of esg factors compared with the reference group. nevertheless, no human capital or economic variables exhibited a substantial role after incorporating the socially responsible motivation. additionally, a significant disparity of r2 was observed between models 1 and 2, underscoring the necessity of including the motivation variable in the analyses. table 3 presents the outcomes of the segmentation analysis, which examines socially responsible motivation across three distinct levels. h2 received further validation. the analysis revealed that objective and subjective investment knowledge significantly influenced the degree of importance placed on esg factors, independent of the level of socially responsible motivation. in particular, among investors with some degree of socially responsible motivation, subjective investment knowledge and two to 10 years of investment experience were positively correlated with assigning greater importance to esg factors. on the other hand, investors who are highly motivated by social responsibility, objective investment knowledge, and risk tolerance stood out as key factors influencing the perceived importance of esg factors in investment decisions. consistent with the main findings, objective and subjective investment knowledge exhibited an opposite relationship with the perceived importance of esg factors among investors who were not motivated by a sense of social responsibility. additionally, investors relying on information from personal financial professionals rated esg factors as more important than those not motivated by social responsibility. male investors who were either not motivated or only partially motivated by social responsibility were less likely to place significant importance on esg factors. the perceived importance of esg factors was positively correlated with a high investment account balance and being between the ages of 18 and 24 among those who were somewhat motivated by social responsibility. robustness check the results of the subsample analysis, which considered gender, financial experience, and reliance on information from personal financial professionals, upheld the direction and significance of socially responsible motivation in the perception of the importance of esg factors. the financial-related variables largely mirrored the main findings. table 4 presents a comprehensive overview of these results. discussion and implications results discussion the findings of this study contribute to the growing body of literature on esg investing. drawing on the vbn theory proposed by stern et al. (1999), when ethical, societal, and environmental concerns resonate with investors’ personal values, those driven by a sense of social responsibility were more inclined to give greater weight to these factors when making investments. the heightened importance assigned to esg could indicate that investors believe that their investment choices should align with their personal values and address ethical, societal, and environmental concerns positively. as a result of this vbn connection, they prioritize esg concerns when making investing decisions, which may finally translate into esg investing as a norm behavior. this study utilized a two-step hierarchical regression to distinguish the significance of social responsibility motivation in the perceived importance of esg factors during investment decisions, and the findings highlight the key role of ethical motives. the motivation, which includes the aspiration to provide a beneficial impact on society, uphold individual ideals, or be socially responsible, is instrumental in shaping the preferences and choices of investors. additionally, the subgroup analysis, segmented by gender, financial experience, and reliance on information from financial professionals, offers strong evidence supporting the significance of socially responsible motivation. investors are likely driven by intrinsic motivations when it comes to investing, especially on esg factors. the results shown in tables 2 and 4 align with prior literature that highlights the importance of individual values and preferences in the process of investing selection (bauer & smeets, 2015; delmas & blass, 2010). h1 was fully supported through theoretical framework as well as analysis evidence. zhang 37 table 1. descriptive analysis whole sample socially responsible motivation (n = 2,324) yes (n = 967) no (n = 1,357) variable mean/% std. dev. mean/% std. dev. mean/% std. dev. min max esg importance rating 4.792 2.710 6.560 2.209 3.531 2.299 1 10 socially responsible motivation 41.61% does not describe at all 58.39% describes somewhat 30.59% describes very well 11.02% financial variables objective investment knowledge 5.508 2.431 4.993 2.388 5.875 2.395 0 11 subjective investment knowledge 4.904 1.298 5.142 1.235 4.734 1.315 1 7 investment experience in years less than a year 3.87% 5.17% 2.95% 1 year to less than 2 years 7.44% 11.17% 4.79% 2 years to less than 5 years 9.08% 13.13% 6.19% 5 years to less than 10 years 10.80% 13.44% 8.92% 10 years or more 68.80% 57.08% 77.16% information dependence (advisor) 70.96% 78.08% 65.88% risk tolerance 6.207 2.238 6.650 2.220 5.891 2.197 1 10 socio-demographic variables male 63.94% 59.26% 67.28% whites 80.98% 74.25% 85.78% married 66.95% 64.01% 69.05% single 18.29% 21.92% 15.70% divorced or separated 9.90% 9.20% 10.39% widowed 4.86% 4.86% 4.86% has dependents 26.03% 34.54% 19.97% employed 53.57% 65.15% 45.32% age categories age 18 to 24 2.88% 5.07% 1.33% age 25 to 34 7.31% 12.00% 3.98% age 35 to 44 13.08% 17.27% 10.10% age 45 to 54 12.87% 13.44% 12.45% age 55 to 64 22.33% 20.37% 23.73% age 65 and above 41.52% 31.85% 48.42% financial services review, 32(2) 38 table 1 continued human capital variables high school and lower 8.39% 8.27% 8.47% some college 17.25% 16.13% 18.05% college degree 49.61% 49.33% 49.82% graduate degree 24.74% 26.27% 23.66% economic variables high investment account balance 59.72% 55.53% 62.71% homeownership 84.94% 83.25% 86.15% income level $35,000 and lower 10.03% 10.65% 9.58% $35,000-$50,000 9.42% 8.79% 9.87% $50,000-$75,000 18.98% 17.89% 19.75% $75,000-$100,000 19.84% 20.48% 19.38% $100,000-$150,000 23.92% 25.13% 23.07% $150,000 and above 17.81% 17.06% 18.35% note: this table presents descriptive statistics for the entire sample alongside a detailed comparative descriptive analysis of investors categorized by their socially responsible motivation. the variable for socially responsible motivation was converted into a binary format, indicating whether motivation was present or not. in the original dataset, motivation was categorized into three levels of agreement in response to the statement, "to make a difference in the world, support values i care about, and be socially responsible." these levels were identified as "does not describe at all," "describes somewhat," and "describes very well," allowing respondents to rate how accurately this statement reflected their investment motivation. zhang 39 table 2. hierarchical ols regression results model 1 model 2 coef. robust se. t p>z coef. robust se. t p>z socially responsible motivation 2.407 0.100 23.98 *** financial variables objective investment knowledge -0.192 0.024 -8.06 *** -0.129 0.022 -6.00 *** subjective investment knowledge 0.481 0.046 10.54 *** 0.325 0.042 7.74 *** investment experience in years (ref: 10 years or more) less than a year 0.593 0.256 2.32 * 0.287 0.240 1.20 1 year to less than 2 years 1.239 0.229 5.42 *** 0.960 0.204 4.71 *** 2 years to less than 5 years 0.936 0.191 4.91 *** 0.634 0.174 3.64 *** 5 years to less than 10 years 0.527 0.175 3.00 ** 0.341 0.158 2.16 * information dependence (advisor) 0.818 0.118 6.96 *** 0.515 0.105 4.88 *** risk tolerance 0.096 0.027 3.52 *** 0.057 0.025 2.31 * socio-demographic variables male -0.701 0.111 -6.29 *** -0.479 0.099 -4.81 *** whites -0.365 0.131 -2.78 ** -0.061 0.118 -0.51 marital status (ref: single) married -0.253 0.164 -1.55 -0.157 0.142 -1.10 divorced or separated -0.096 0.218 -0.44 -0.053 0.192 -0.28 widowed -0.625 0.252 -2.49 * -0.685 0.245 -2.80 ** has dependents 0.302 0.145 2.08 * 0.185 0.127 1.46 employed 0.388 0.133 2.93 ** 0.177 0.119 1.48 age categories (ref: age 65+) age 18 to 24 0.797 0.368 2.17 * 0.492 0.333 1.48 * age 25 to 34 0.352 0.260 1.35 0.120 0.235 0.51 age 35 to 44 0.154 0.221 0.70 0.159 0.198 0.81 age 45 to 54 -0.350 0.198 -1.77 -0.308 0.176 -1.75 * age 55 to 64 -0.067 0.147 -0.46 -0.055 0.129 -0.43 human capital variables (ref: high school and lower) some college -0.054 0.214 -0.25 -0.116 0.198 -0.59 college degree 0.028 0.195 0.14 -0.060 0.181 -0.33 graduate degree 0.146 0.214 0.68 -0.036 0.196 -0.18 economic variables high investment account balance 0.060 0.121 0.49 0.119 0.109 1.09 homeownership 0.157 0.159 0.98 0.083 0.144 0.58 financial services review, 32(2) 40 table 2 continued income level (ref: $150,000+) $35,000 and lower 0.181 0.236 0.77 0.111 0.211 0.53 $35,000-$50,000 0.203 0.227 0.90 0.166 0.199 0.83 $50,000-$75,000 0.217 0.184 1.18 0.179 0.158 1.13 $75,000-$100,000 0.187 0.171 1.09 0.107 0.152 0.70 $100,000-$150,000 0.133 0.160 0.83 0.016 0.140 0.11 intercept 2.356 0.416 5.67 *** 2.207 0.377 5.86 *** f (30,2293) = 26.48 f (31,2292) = 55.51 r2 = 0.2258 r2 = 0.3860 note. this table shows the results from the two-step hierarchical ols regression analysis assessing the impact of socially responsible motivation, financial-related variables, and other factors on investors' perception of esg importance. in model 2, socially responsible motivation was binary coded, signifying its presence or absence. objective investment knowledge scores reflect correct responses out of 11 investment-specific questions, while subjective investment knowledge was scored on a 7-point scale. risk tolerance spans a 10-point scale, indicating the willingness of respondents to undertake financial risks. investment experience in years was differentiated into five levels and was dummy coded to reference those with 10 years or more of experience. information dependence (advisor) was represented as a dummy variable to indicate whether respondents rely on financial professionals for information. socio-demographic variables, human capital variables, and economic variables were dummy coded. model 1 and model 2 report the coefficients, robust standard errors, t statistics, and significance levels. the significance levels are indicated by asterisks: *p < 0.05, **p < 0.01, ***p < 0.001. zhang 41 table 3. segmentation analysis ols regression results not at all (n = 1,357) somewhat (n = 711) very well (n = 256) coef. robust se. t p>z coef. robust se. t p>z coef. robust se. t p>z financial variables objective investment knowledge -0.107 0.031 -3.47 ** -0.057 0.034 -1.67 -0.170 0.065 -2.62 ** subjective investment knowledge 0.232 0.057 4.10 *** 0.398 0.080 4.95 *** 0.294 0.157 1.88 investment experience in years (ref: 10 years or more) less than a year 0.129 0.392 0.33 0.375 0.343 1.09 0.152 0.615 0.25 1 year to 2 years 0.668 0.351 1.91 1.290 0.312 4.13 *** 0.788 0.430 1.83 2 years to 5 years 0.427 0.281 1.52 0.804 0.271 2.97 ** 0.188 0.445 0.42 5 years to 10 years 0.214 0.236 0.91 0.493 0.244 2.02 * -0.021 0.370 -0.06 information dependence (advisor) 0.568 0.137 4.14 *** 0.275 0.181 1.52 0.520 0.434 1.20 risk tolerance 0.014 0.033 0.43 0.057 0.042 1.35 0.202 0.076 2.65 ** socio-demographic variables male -0.562 0.140 -4.01 *** -0.636 0.155 -4.09 *** -0.158 0.270 -0.58 whites -0.167 0.183 -0.91 -0.095 0.174 -0.55 0.509 0.284 1.79 marital status (ref: single) married -0.319 0.201 -1.59 0.111 0.225 0.49 -0.487 0.383 -1.27 divorced or separated -0.198 0.262 -0.75 0.077 0.296 0.26 0.530 0.506 1.05 widowed -0.385 0.328 -1.17 -1.085 0.365 -2.97 ** -0.315 0.699 -0.45 has dependents 0.193 0.184 1.05 0.011 0.208 0.05 0.156 0.313 0.50 employed 0.263 0.155 1.70 -0.073 0.192 -0.38 0.107 0.456 0.23 age categories (ref: age 65+) age 18 to 24 0.224 0.685 0.33 0.814 0.491 1.66 * 0.365 0.734 0.50 age 25 to 34 0.211 0.390 0.54 0.167 0.377 0.44 0.077 0.595 0.13 age 35 to 44 -0.207 0.280 -0.74 0.327 0.335 0.98 0.616 0.601 1.03 age 45 to 54 -0.388 0.242 -1.60 -0.174 0.261 -0.67 0.079 0.656 0.12 age 55 to 64 -0.246 0.166 -1.48 0.241 0.211 1.14 0.445 0.510 0.87 human capital variables (ref: high school and lower) some college -0.358 0.266 -1.35 0.348 0.327 1.06 0.132 0.566 0.23 college degree -0.222 0.243 -0.91 0.247 0.291 0.85 -0.098 0.522 -0.19 graduate degree -0.280 0.267 -1.05 0.229 0.316 0.73 -0.056 0.548 -0.10 economic variables high investment account balance -0.239 0.151 -1.58 0.555 0.174 3.20 ** 0.478 0.291 1.64 financial services review, 32(2) 42 note: this table presents the outcomes of an ols regression analysis segmented by the levels of socially responsible motivation among investors, categorized as "not at all," "somewhat," and "very well." the coefficients, robust standard errors, t statistics, and significance levels are reported, with significance denoted by asterisks: *p < 0.05, **p < 0.01, ***p < 0.001. the model's goodness of fit is indicated by the r-squared values, whereas the f statistics signify the overall significance of each model. table 3 continued homeownership 0.031 0.209 0.15 0.072 0.229 0.31 0.016 0.406 0.04 income level (ref: $150,000+) $35,000 and lower -0.040 0.288 -0.14 0.251 0.358 0.70 -0.002 0.602 0.00 $35,000-$50,000 0.147 0.264 0.56 0.148 0.335 0.44 -0.308 0.607 -0.51 $50,000-$75,000 0.109 0.216 0.50 0.237 0.255 0.93 -0.089 0.474 -0.19 $75,000-$100,000 0.248 0.206 1.20 -0.239 0.256 -0.94 0.031 0.388 0.08 $100,000-$150,000 0.099 0.191 0.52 -0.091 0.237 -0.38 0.145 0.372 0.39 intercept 3.525 0.534 6.60 3.283 0.607 5.41 4.263 1.345 3.17 ** f (30, 1326) = 3.32 r2 = 0.0734 f (30, 680) = 4.91 r2 = 0.1632 f (30, 225) = 4.61 r2 = 0.2974 zhang 43 table 4. robustness check with segmentation analysis ols regression results gender men (n = 1,486) women (n = 838) coef. robust se. t p>z coef. robust se. t p>z socially responsible motivation 2.501 0.135 18.58 *** 2.252 0.149 15.16 *** financial variables objective investment knowledge -0.172 0.028 -6.20 *** -0.040 0.033 -1.19 subjective investment knowledge 0.313 0.053 5.86 *** 0.311 0.066 4.72 *** investment experience in years (ref: 10 years or more) less than a year 0.008 0.363 0.02 0.593 0.321 1.84 1 year to less than 2 years 1.094 0.245 4.47 *** 0.472 0.332 1.42 2 years to less than 5 years 0.558 0.220 2.54 * 0.686 0.296 2.32 * 5 years to less than 10 years 0.426 0.207 2.06 * 0.104 0.236 0.44 information dependence (advisor) 0.336 0.135 2.50 * 0.773 0.165 4.68 *** risk tolerance 0.013 0.031 0.42 0.122 0.039 3.14 ** socio-demographic variables male whites -0.055 0.155 -0.36 -0.076 0.183 -0.41 marital status (ref: single) married -0.029 0.187 -0.15 -0.410 0.219 -1.87 divorced or separated -0.212 0.270 -0.78 -0.001 0.276 0.00 widowed -1.289 0.363 -3.55 *** -0.496 0.341 -1.45 has dependents 0.261 0.169 1.55 0.013 0.201 0.07 employed 0.261 0.161 1.62 0.024 0.177 0.13 age categories (ref: age 65+) age 18 to 24 0.666 0.421 1.58 0.477 0.517 0.92 age 25 to 34 0.064 0.303 0.21 0.308 0.361 0.85 age 35 to 44 0.266 0.244 1.09 -0.029 0.349 -0.08 age 45 to 54 -0.378 0.232 -1.63 -0.076 0.274 -0.28 age 55 to 64 -0.150 0.177 -0.85 0.187 0.189 0.99 human capital variables and economic variables included intercept 2.198 0.501 4.39 *** 1.687 0.587 2.87 ** f (30, 1455) = 44.77 r2 = 0.4056 f (30, 807) = 18.52 r2 = 0.2929 financial services review, 32(2) 44 table 4 continued financial experience < 10 years (n = 725) >= 10 years (n = 1,599) coef. robust se. t p>z coef. robust se. t p>z socially responsible motivation 2.219 0.201 11.01 *** 2.450 0.116 21.03 *** financial variables objective investment knowledge -0.181 0.041 -4.38 *** -0.087 0.026 -3.31 ** subjective investment knowledge 0.537 0.073 7.35 *** 0.170 0.054 3.16 ** investment experience in years (ref: 10 years or more) less than a year 1 year to less than 2 years 2 years to less than 5 years 5 years to less than 10 years information dependence (advisor) 0.384 0.195 1.97 * 0.511 0.126 4.04 *** risk tolerance 0.080 0.046 1.73 0.043 0.029 1.49 socio-demographic variables male -0.258 0.176 -1.46 -0.583 0.120 -4.87 *** whites 0.008 0.185 0.05 -0.159 0.154 -1.03 table 4 continued marital status (ref: single) married -0.081 0.244 -0.33 -0.297 0.177 -1.68 divorced or separated 0.075 0.359 0.21 -0.247 0.231 -1.07 widowed -0.466 0.592 -0.79 -0.847 0.281 -3.02 ** has dependents 0.055 0.201 0.27 0.151 0.170 0.89 employed 0.140 0.255 0.55 0.190 0.134 1.41 age categories (ref: age 65+) age 18 to 24 0.625 0.462 1.35 -0.276 0.759 -0.36 age 25 to 34 0.042 0.372 0.11 0.445 0.517 0.86 age 35 to 44 0.219 0.355 0.62 -0.265 0.278 -0.95 age 45 to 54 -0.766 0.350 -2.19 * -0.111 0.211 -0.53 age 55 to 64 -0.172 0.332 -0.52 -0.044 0.142 -0.31 human capital variables and economic variables included intercept 2.008 0.621 3.23 *** 3.357 0.494 6.79 *** f (27, 697) = 32.12 r2 = 0.4753 f (27, 1571) = 26.37 r2 = 0.2929 zhang 45 table 4 continued information dependence from advisors yes ( n = 1,649) no ( n = 675) coef. robust se. t p>z coef. robust se. t p>z socially responsible motivation 2.346 0.118 19.96 *** 2.531 0.199 12.70 *** financial variables objective investment knowledge -0.143 0.025 -5.67 *** -0.071 0.042 -1.67 subjective investment knowledge 0.358 0.050 7.13 *** 0.230 0.081 2.83 ** investment experience in years (ref: 10 years or more) less than a year 0.528 0.288 1.83 -0.072 0.396 -0.18 1 year to less than 2 years 1.012 0.255 3.97 *** 0.852 0.341 2.50 * 2 years to less than 5 years 0.710 0.199 3.56 *** 0.440 0.341 1.29 5 years to less than 10 years 0.268 0.177 1.51 0.490 0.339 1.44 information dependence (advisor) risk tolerance 0.087 0.029 2.95 ** -0.003 0.045 -0.07 socio-demographic variables male -0.555 0.116 -4.81 *** -0.268 0.198 -1.35 whites 0.025 0.139 0.18 -0.328 0.226 -1.45 marital status (ref: single) married -0.142 0.169 -0.84 -0.134 0.270 -0.50 divorced or separated 0.076 0.231 0.33 -0.289 0.340 -0.85 widowed -0.617 0.283 -2.18 * -0.952 0.458 -2.08 * has dependents 0.267 0.153 1.75 -0.097 0.236 -0.41 employed 0.174 0.141 1.24 0.180 0.229 0.79 age categories (ref: age 65+) age 18 to 24 0.205 0.385 0.53 0.919 0.681 1.35 age 25 to 34 0.010 0.282 0.04 0.252 0.428 0.59 age 35 to 44 0.144 0.237 0.61 0.083 0.367 0.23 age 45 to 54 -0.433 0.215 -2.02 * -0.180 0.319 -0.56 age 55 to 64 -0.169 0.148 -1.14 0.227 0.259 0.87 human capital variables and economic variables included intercept 2.285 0.442 5.17 *** 3.337 0.716 4.66 *** f (30, 1618) = 44.09 r2 = 0.4031 f (30, 644) = 11.70 r2 = 0.3188 note: table 4 provides the outcomes of robustness checks through three separate segmentation analyses within ols regression, aimed at assessing the influence of socially responsible motivation, financial variables, and other factors on investors’ perceptions of esg importance. the analyses differentiate the sample by gender (men vs. women), financial experience (<10 years vs. >=10+ years), and information dependence on advisors (yes vs. no). socially responsible motivation is coded in a binary manner, indicating its presence or absence. significant levels are marked with asterisks: *p<0.05, **p<0.01, ***p<0.001. the results of this study challenge the positive correlation between objective financial knowledge and esg investing identified in earlier international research (i.e., cucinelli & soana, 2023; kar & patro, 2024). it is important to note that, unlike previous studies that focused on fundamental financial knowledge, this study emphasizes investment-specific knowledge. among the whole analytical sample, findings in the current study highlight a negative relationship between objective investment knowledge and the level of importance assigned to esg when financial services review, 32(2) 46 making an investment. on the contrary, subjective investment knowledge was found to be positively associated with the perceived importance of esg factors. the results of segment analysis in table 3 reveals that financialrelated variables, including objective or subjective investment knowledge, investment experience over time, information dependence, and risk tolerance, influence the degree of importance individuals attributed to esg factors differently among investors with varied levels of socially responsible motivation. for investors somewhat motivated by social responsibility, subjective investment knowledge played a significant positive role. conversely, objective investment knowledge demonstrated a significant negative role among investors highly motivated by social responsibility. the opposite sign might contribute to the fact that traditional investment vehicles may be appealing to investors with highly objective investment knowledge, as they possess a factual understanding of investment principles. as noted in giglio et al. (2023), individual investors generally expect lower monetary returns on esg investment; investors with high objective investment knowledge may prioritize quantitative metrics and perceive the esg factor as less directly linked to financial performance. they may place a higher value on financial performance from traditional fundamental financial analysis compared with esg factors. on the contrary, the positive relationship between subjective investment knowledge and esg importance may indicate differences in values and beliefs. investors with higher subjective investment knowledge may exhibit heightened consciousness regarding social and environmental concerns, thereby attributing more significance to esg factors. given the inherent difficulty in quantifying the nonmonetary return associated with fulfilling personal values (cornell, 2021), investors with a high degree of subjective investment knowledge may be more receptive to incorporating nontraditional factors such as esg into their investment decisions, believing that doing so will increase their total returns. it is crucial to emphasize that there is presently a scarcity of research that connects investment knowledge with individuals’ perspectives of the significance of esg aspects or their actual investment behaviors on esg. this study seeks to investigate objective and subjective investment knowledge in order to provide fundamental insights into this relatively unexplored field. further examination is warranted to explore the intricate relationship between objective and subjective investment knowledge and the subsequent manifestation of overconfidence. this field of study holds significant potential for future research. investors with more than one year but less than a decade of investing experience were found to place a higher value on esg factors than investors with more than 10 years of experience, especially those who are moderately motivated by social responsibility. young investors, especially those aged 18 to 24, compared with those aged 65 and above, were more likely to assign a higher importance to esg factors. this may arise due to an increasing number of higher education institutions incorporating sustainable development concepts into their curricula to educate students about sustainability (gigauri et al., 2022), which could raise novice investors’ awareness of the importance of sustainability and ethical considerations when making investments. individuals who rely on recommendations from financial experts on investment matters were found to be more likely to assign greater perceived value to esg factors when making investment decisions. this holds true even for investors who are not driven by a sense of social responsibility. further, investors may trust their advisors’ expertise if they depend on financial advisors for information. financial professionals who emphasize the significance of esg factors can shape investors’ perceptions. through engaging discussions, financial professionals could enhance investors’ understanding of esg issues, the advantages of esg in times of market uncertainty, and its potential to mitigate longterm risks (cerqueti et al., 2021; mavlutova et al., 2021; pisani & russo, 2021). this result should be interpreted with caution because the data set employed in this study did not provide information on whether financial advisors offer zhang 47 detailed insights into esg factors to individual investors. future studies could extend this line of research and utilize direct measurement of whether financial advisors incorporate esg considerations into their advice and how this influences investors’ perceptions and actual investment behaviors. the higher the risk tolerance, the perception of the importance of esg factors also increases. this relationship was obvious among investors who were strongly motivated by social responsibility, indicating that such investors might anticipate lower returns outside of crises (cerqueti et al., 2021; giglio et al., 2023). a higher level of risk tolerance was required to engage in esg investments. implications the findings of this study provide financial institutions and financial planners with critical insights. financial institutions could develop targeted marketing strategies based on the identified unique characteristics of investors who place a high value on esg. financial practitioners could accommodate their clients more effectively when products that emphasize esg criteria align well with client’s investment portfolios and individual situations. this calls for identifying investors with a strong commitment to social responsibility or having more than one year but less than 10 years of experience. financial planners and advisors might consider developing new strategies to meet the specific needs of different investors based on their clients’ investment motivation. additionally, financial planners and advisors need to acknowledge their crucial influence on shaping the perceived significance of esg factors among investors who may not be inherently driven by social responsibility. they should also consider providing guidance and educational resources to the broader investor community. the favorable relationship between subjective investment knowledge and the importance of esg factors indicates that confident investors would benefit from education programs regarding trending investments. investors may be able to make more informed decisions if robust reporting and transparency standards for esg factors are implemented and enforced. policymakers could establish an awareness program to enlighten investors regarding the benefits and potential risks of esg investing during various market conditions, the investment’s long-term orientation, and the potential expenses linked to esg investing. individuals with a high level of objective investment knowledge who place little weight on esg factors when making investment decisions may underestimate the significance of these factors. financial institutions might consider designing and providing a comprehensive brochure that discusses investment options in accordance with esg criteria. tailored guidelines may be more effective, given that investors with different risk tolerance levels place differing degrees of significance on esg factors. given the increasing awareness and demand for esg mutual funds, policymakers might prioritize promoting esg transparency and establishing reporting standards that require companies to report comparable and dependable information on their esg practices. this would enable investors, irrespective of their level of investment knowledge, to make more informed decisions. due to a significant correlation between information reliance on the financial profession and the perceived importance of esg factors, policymakers should consider establishing training requirements on esg issues among financial professionals to ensure that advisors are adequately equipped to provide guidance on esg investing. understanding investors’ desire for socially responsible investing and their financial circumstances can help financial professionals provide more targeted and effective investment recommendations. with the assistance of financial institutions and policymakers, investment opportunities with an esg focus could be matched with education guidelines that are more suitable for prospective investors. limitation and future studies the current study acknowledges certain limitations that must be recognized. while the vbn theory posits that personal values and beliefs can influence normative behaviors, the lack of data information on actual esg investing behaviors limits the scope of our findings. instead, this study concentrates on the significance of financial services review, 32(2) 48 esg factors in investment decisions, providing indirect insights into esg investing propensity. while the current analysis provides valuable insights into the perceived importance of esg factors, due to the cross-sectional nature of the 2021 nfcs state-by-state and investor survey data set, this study is not intended to establish causal relationships among investing motivations, investor characteristics, and the emphasis placed on esg factors in investment considerations. future studies should focus on decisions that reduce potential issues associated with reverse causality. for example, as individuals perceive esg factors as important, they might also start to see themselves as investing to make a difference, thus potentially reversing the assumed direction of impact. future studies could also extend the findings of this study by employing longitudinal data sets or experiments as potential ways to verify the influence of investor motivations and characteristics on the inclusion of investments that comply with esg criteria in their portfolios. additional research is necessary to thoroughly investigate the factors influencing investors’ investment decisions in esg options. further analysis is required to address endogeneity and reverse causality concerns effectively, thus facilitating a more comprehensive understanding of the underlying dynamics. furthermore, future research could build upon the foundation provided by this study to explore how external events or regulatory policies influence investors in choosing investments, which may contribute to a more holistic comprehension of esg investment. conclusion the primary goal of the present study was to investigate the relationship between investor attributes and the degree to which they prioritize environmental, social, and governance factors in their investment decision-making. in particular, this study validates the positive association between the motivation for socially responsible investment and the level of importance attributed to esg factors. the findings provide valuable insights into the existing body of literature by establishing a connection between the vbn theory and the financial aspect, thus enhancing the current understanding of whether certain investors prioritize esg criteria out of a selfidentified motivation to effectuate positive global change, uphold personal values, or engage in socially responsible practices. socially responsible motivation is the most robust and prominent variable linked with the perceived importance of esg factors in subgroup analyses. investors with lower objective investment knowledge, higher subjective investment knowledge, more than one year but less than a decade of investing experience, reliance on information provided by financial professionals, and higher risk tolerance levels were found to assign greater importance to esg factors during their investment decision-making process, as indicated by their responses in the hierarchical regression model. however, the significance of these financial variables differed among investors based on their varying levels of socially responsible motivation. these findings underscore the importance of considering investor profiles in understanding esg investment and contribute to the growing body of literature on sustainable investment by illustrating the potential multifaceted nature of esg determinants and offering targeted strategies for engaging different investor groups. these findings have significant implications for policymakers, financial institutions, and financial practitioners. financial practitioners should recognize an investor’s investment motives and financial profile variables to facilitate the provision of more tailored and impactful investment guidance and different targeted communication strategies. financial institutions and policymakers might develop educational programs to help investors understand complex esg criteria and make the information more accessible and understandable to fulfill the needs of investors with varied profiles. references alareeni, b. and hamdan, a. (2020). esg impact on performance of us s&p 500-listed firms. corporate governance, 20(7), 1409-1428. doi: 10.1108/cg-06-20200258 zhang 49 baker, m., egan, m. l., & sarkar, s. k. (2022). how do investors value esg? national bureau of economic research. http://www.nber.org/papers/w30708 bauer, r. and smeets, p. (2015). social identification and investment decisions. journal of economic behavior & organization, 117, 121-134. doi: 10.1016/j.jebo.2015.06.006 bloomberg. (2024, february 08). global esg assets predicted to hit $40 trillion by 2030, despite challenging environment, forecasts bloomberg intelligence. bloomberg.com. cerqueti, r., ciciretti, r., dalò, a., & nicolosi, m. (2021). esg investing: a chance to reduce systemic risk. journal of financial stability, 54, 100887. doi: 10.1016/j.jfs.2021.100887 cornell, b. (2021). esg preferences, risk and return. european financial management, 27(1), 12-19. cucinelli, d., & soana, m. g. (2023). investor preferences, financial literacy and intermediary choice towards sustainability. research in international business and finance, 102027. delmas, m. and blass, v. (2010). measuring corporate environmental performance: the trade‐offs of sustainability ratings. business strategy and the environment, 19(4), 245-260. doi: 10.1002/bse.676 european banking authority. (2021). eba report on management and supervision of esg risks for credit institutions and investment firms. https://www.eba.europa.eu/sites/default/ files/document_library/publications/rep orts/2021/1015656/eba%20report%20 on%20esg%20risks%20management% 20and%20supervision.pdf friede, g., busch, t., & bassen, a. (2015). esg and financial performance: aggregated evidence from more than 2000 empirical studies. journal of sustainable finance and investment, 5(4), 210-233. doi: 10.1080/20430795.2015.1118917 gigauri, i., vasilev, v., & mushkudiani, z. (2022). in pursuit of sustainability: towards sustainable future through education. international journal of innovative technologies in economy. giglio, s., maggiori, m., stroebel, j., tan, z., utkus, s., & xu, x. (2023). four facts about esg beliefs and investor portfolios (no. w31114). national bureau of economic research. jansson, j., marell, a., & nordlund, a. (2011). exploring consumer adoption of a high involvement eco‐innovation using value‐ belief‐norm theory. journal of consumer behaviour, 10(1), 51-60. kar, d., & patro, b. (2024). unravelling the impact of financial literacy on the intention of indian retail investors to adopt socially responsible investments: the interplay of social self-efficacy as a mediating mechanism. global business review. doi: 10.1177/09721509231223571 kim, s. and li, z. (2021). understanding the impact of esg practices in corporate finance. sustainability, 13(7), 3746. doi: 10.3390/su13073746 lee, d., fan, j., & wong, v. (2020). no more excuses! performance of esg‐integrated portfolios in australia. accounting and finance, 61(s1), 2407-2450. doi: 10.1111/acfi.12670 lin, j. t., bumcrot, c., mottola, g., valdes, o., and walsh, g. (2022). the changing landscape of investors in the united states: a report of the national financial capability study. finra investor education foundation. www.finrafoundation.org/investorreport2021 lind, h. b., nordfjærn, t., jørgensen, s. h., & rundmo, t. (2015). the value-beliefnorm theory, personal norms and sustainable travel mode choice in urban areas. journal of environmental psychology, 44, 119-125. financial services review, 32(2) 50 lópez-mosquera, n., & sánchez, m. (2012). theory of planned behavior and the value-belief-norm theory explaining willingness to pay for a suburban park. journal of environmental management, 113, 251-262. mavlutova, i., fomins, a., spilbergs, a., atstāja, d., & brizga, j. (2021). opportunities to increase financial well-being by investing in environmental, social and governance with respect to improving financial literacy under covid-19: the case of latvia. sustainability, 14(1), 339. doi: 10.3390/su14010339 park, s. r. and jang, j. y. (2021). the impact of esg management on investment decision: institutional investors' perceptions of country-specific esg criteria. international journal of financial studies, 9(3), 48. doi: 10.3390/ijfs9030048 plastun, a., makarenko, i., yelnikova, y., & makarenko, s. (2019). environmental, social and governance investment standardization: moving towards sustainable economy. environmental economics, 10(1), 12-22. doi: 10.21511/ee.10(1).2019.02 pisani, f. and russo, g. (2021). sustainable finance and covid-19: the reaction of esg funds to the 2020 crisis. sustainability, 13(23), 13253. doi: 10.3390/su132313253 stern, p. c., dietz, t., abel, t., guagnano, g. a., & kalof, l. (1999). a value-belief-norm theory of support for social movements: the case of environmentalism. human ecology review, 81-97. tippet, j. and leung, p. (2001). defining ethical investment and its demography in australia. australian accounting review, 11(25), 44-55. doi: 10.1111/j.18352561.2002.tb00389.x us sif foundation. (2022). 2022 report on us sustainable investing trends. https://www.ussif.org//files/trends/202 2/trends%202022%20executive%20su mmary.pdf zhang 51 appendix survey questions for key variables perceived importance of esg factors how important is esg (environmental, social, and corporate governance issues) to you when making investment decisions? not at all important extremely important 1 2 3 4 5 6 7 8 9 10 socially responsible motivation how well does the following describe why you invest? to make a difference in the world/support values i care about/be socially responsible. 1 = does not describe at all 2 = describe somewhat 3 = describes very well objective investment knowledge 1. if you buy a company’s stock… you own a part of the company you have lent money to the company you are liable for the company’s debts the company will return your original investment to you with interest 2. if you buy a company’s bond… you own a part of the company you have lent money to the company you are liable for the company’s debts you can vote on shareholder resolutions 3. if a company files for bankruptcy, which of the following securities is most at risk of becoming virtually worthless? the company’s preferred stock the company’s common stock the company’s bonds 4. in general, investments that are riskier tend to provide higher returns over time than investments with less risk. true false 5. the past performance of an investment is a good indicator of future results. true false 6. over the last 20 years in the us, the best average returns have been generated by: stocks bonds cds money market accounts precious metals 7. what is the main advantage that index funds have when compared to actively managed funds? index funds are generally less risky in the short term index funds generally have lower fees and expenses index funds are generally less likely to decline in value 8. which of the following best explains why many municipal bonds pay lower yields than other government bonds? municipal bonds are lower risk financial services review, 32(2) 52 there is a greater demand for municipal bonds municipal bonds can be tax-free 9. you invest $500 to buy $1,000 worth of stock on margin. the value of the stock drops by 50%. you sell it. approximately how much of your original $500 investment are you left with in the end? $500 $250 $0 10. which is the best definition of “selling short”? selling shares of a stock shortly after buying it selling shares of a stock before it has reached its peak selling shares of a stock at a loss selling borrowed shares of a stock 11. if you own a call option with a strike price of $50 on a security that is priced at $40, and the option is expiring today, which of the following is closest to the value of that option? $10 $0 -$10.00 subjective investment knowledge on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall knowledge about investing? very low 1 2 3 4 5 6 very high 7 1 2 3 4 5 6 7 pii: s1057-0810(97)90029-9 i financial services review, 6(1): 19-25 issn: 1057-08 i0 copyright © 1997 by jai press inc. all fights of reproduction in any form reserved. the congressional calendar and stock market performance reinhold p. lamb k.c. ma r. daniel pace william f. kennedy this study reports on the existence of a curious calendar effect--a relationship between stock market performance and the schedule of the u.s. congress. almost the entire advance in the market since 1897 corresponds to the periods when congress is in recess. this is an impressive result, given that congress is in recess about half as long as in sesskm. furthermore, average daily returns when congress is not meeting are almost thirteen times greater than when congress is in session. throughout the year, cumulative returns during recess are eight times that experienced while congress is in session. i. introduction there exists an interesting calendar effect regarding stock market performance. our research indicates that an overwhelming majority of positive returns in the stock market over the last century occur when the u.s. congress (house of representatives) is in recess (closed). several studies report calendar anomalies in stock market returns. for example, rozeff and kinney (1976) find that the average monthly return in january is significantly greater than the average return for the other months. gultekin and gultekin (1983) con clude that this seasonality is not limited to the u.s. market. belgium, netherlands, italy, japan, and several other countries exhibit returns in january that are larger than those dur ing the rest of the year. keim (1983), roll (1983) and reinganum (1983) find that this sea sonality is present mainly in small firms and during the first five days of january. r e i n h o l d p. lamb • the university of north carolina at charlotte, charlotte, nc; e-mail: fba00rpl@email.uncc.edu. k.c. m a * kcm asset management group, wilmette, il. r. daniel pace * the university of west florida; e-mail: dpace@uwf.edu. william f. kennedy * the university of north carolina at charlotte, charlotte, nc, e-mail: wfkenned@emailuncc.edu. 20 financial services review 6(1) 1997 a weekend effect is documented whereby the returns for monday are significantly lower than those of the other days of the week (french, 1980; keim & stambaugh, 1984; smirlock & starks, 1986). jaffe and westerfield (1985) find a similar effect in japan. rogalski (1984) decomposes the monday returns and finds that all of the weekend effect is due to the pure weekend return (close friday to open monday). the pure monday effect (monday open to monday close) is positive. ariel (1987) determines that all of the market's cumulative increase during 1963 1981 occurs around the first half of the month. the second half contributes nothing to the cumulative advance in the market. predominantly negative returns exist after the midpoint of the month. merrill (1984) reports that the market is usually up the day preceding a holiday, and the return for the day after a holiday is usually lower than that before the holiday. a notable exception is thanksgiving; both days surrounding the holiday exhibit a similar positive market performance. another interesting calendar anomaly, called the presidential cycle, is traced back to the presidency of andrew jackson. the last two years (election year and pre-election year) of the 41 administrations since 1832 account for a total net market gain of 527%. the cumulative market gains during the first two years of those administrations produce only a 74% return (hirsch, 1992). furthermore, over the past 12 presidential elections, a market (s&p 500) advance during the election year occurs 10 times (83%). the market experi ences only two declines. this behavior contrasts sharply with that of nonelection years whereby a market increase occurs only 65% of the time (maturi, 1993). an interesting study by michelson (1993) reports that investors can expect high returns when congress is in recess, and when a democratic congress and democratic president are in power. our paper expands on this curious congressional calendar effect and examines the behavior of the market during open and closed congressional sessions over the last 97 years. we find very different market performances while congress is meeting as opposed to when it is in recess. the next sections describe the data and methodology that this study employs, followed by the results and the conclusions from the findings. h. data and methodology the sample period for this study comprises 1897-1993. the dates that congress is in session (open) and in recess (closed) during this 97-year period are collected from the congressional record. since congress is either in session or in re~ss during each year, a long sample period is necessary to provide statistically robust results. data limi tations on the center for research in security prices (crsp) tape prevents extending the sample period much beyond 30 years. using the crsp data, therefore, limits our sample to about thirty open and thirty closed observations. instead, ninety-seven years of daily dow jones industrial averages (djia) are collected to form as large a sample period as possible. the open and closed session dates are compiled and then the corresponding returns and points are calculated for the djia. the djia point value for each session is the dif ference between the djia at the end of the session and the closing value on the day the congressional calendar 21 immediately preceding the beginning of the session. returns are calculated in a similar way. both arithmetic average daily and annual reumas are calculated for closed and open subsamples. the focus on the average returns is not on the return value for the closed and open sessions. rather, the interesting characteristic is the degree of differ ence between the session returns. t-tests are performed to expose significant differ ences in returns. other models are developed to provide additional insights into the behavior of the market during the congressional calendar. these models test for (a) dif ferences in return levels between open and closed periods, and (b) the relationship between session returns and the month of january. m . results: the congressional effect table 1-panel a indicates that during 1897-1993, open dates for congress total 16,387 days. closed dates are 9,950 days. thus, congress is typically open for about two thirds of each year and closed for about one-third of each year. overall, the duration congress is open is almost twice as long as the closed period. the average number of closed and open days each year is about 105 and 173, respectively. this difference is significant at the 1% level. the range of open days is 59-301; the range for the closed period is 8-241. since the general trend of the stock market is upward over the long-term, the obser vation that congress is open a substantially longer time than closed implies that returns table 1 descriptive statistics of stock market performance around congressional sessions 1897-1993 in recess (closed) in session (open) panel a: session duration # of sessions 95 95 # of days 9,950 16,387 average # of days i 104.7 172.5 minimum # of days 8 58 maximum # of days 241 301 standard deviation 58.9 61.7 panel b: session returns average daily return 2 average annual return 0 .0541% 4 .2379% 3 0.0042% 0.5532% 4 notes: ! t-value is 6.219 of a difference between closed and open. 2 t-value is 2.500 of a difference between closed and open. 3 t-value is 2.578 of a difference between closed return and zero. 4 t-value is 0.372 of a difference between open return and zero. 22 financialservicesreie 6(1) 1997 should be higher for the open subperiod. table 1-panel b presents, however, that the average daily returns and average annual returns are actually higher for the closed ses sion. the average daily return when congress is closed is 0.0541%. when congress is open, the average daily return is only 0.0042%. closed session returns are, thus, almost table 2 regression results of relationship between annual market returns and the congressional calendar 1897-1993 panel a: relationship between market return and if congress is open or closed ret m = a + 1 3 d + e where d : 0 i f congress is in recess (closed), and d = 1 if congress is in session (open) ret m = 0 . 0 4 1 0 . 0 3 9 d + 0.016 (3.323) (-2.500) f-value = 6.254 p-value = 0.012 panel b: relationship between annual market return and closed congress ret m = a + i~out + e where out = market return when congress is in recess (closed) ret = 0.447 + 1.025out + 1.545 r a (0.290) (10.990) f-value = 120:783 p-value = 0.001 panel c: relationship between annual market return and open congress ret m = a + l ~ i n + e where in = market return when congress is in session (open) ret m = 4 . 2 2 1 + 1.030in + 1.653 (2.554) (0.372) f-value = 0.104 p-value = 0.901 thirteen times greater than for when congress is in session. this difference is signifi cant at the 5% level. throughout the year, the closed return accumulates to about 4.24%; the open return is not significantly different from zero (0.5532%). the closed period produces returns almost eight times greater than experienced in open sessions. these observations provide evidence that the returns accumulated by the market may be concentrated in congressional recesses. the congressional calendar 23 to gain more insight into the relationship between market returns and the congres sional calendar, several regressions on the data are performed. table 2-panel a pre sents the results of regressing the market returns on a dummy variable representing the closed and open periods. the model indicates that the fact that congress is either closed or open has a significant impact on market returns. table 2-panel b regresses annual market returns on the closed congress returns and exhibits a significant relation ship. the intercept term is not significant. the annual return is attributable to the closed period return. table 2-panel c performs a similar test on open period returns. no significant relationship is observed between the annual returns and the open period returns. the intercept term is significant. something else (closed returns), therefore, explains the annual returns. it appears that market behavior is sensitive to the congressional calendar. the returns generated by the market during the sample period total 607%. returns while congress is closed total 538% (0.0541% * 9,950 days). while congress is open, returns accumulate to 69% (0.0042% * 16,387 days). about 89% of the total returns are, therefore, obtained while congress is closed. only 11% of the total returns are pro duced while congress is open. this is an impressive result given that congress is open almost twice as long as it is closed. a disproportionate level of returns is, thus, earned while congress is closed. that is, about 89% of the market returns are generated in only about one-third of the time. a possible explanation for this interesting observation may be traced back to the well-known calendar anomalies described at the beginning of the paper. since congress is usually in recess in early january, and abnormally large returns accumulate in january, perhaps this congressional effect is really a manifestation of the january effect. we believe, however, that our results are not driven by such seasonalities. first, our market proxy, the djia, is comprised of large firms. banz (1981), keim (1983) and reinganum (1983) show that large firms do not experience a significant january effect. the seasonal is concentrated mainly among small firms. since the djia does not include small firms, the returns we observe should bear little exposure to the january effect and, consequently, the congressional effect should not be influenced by january returns. second, we perform a regression analysis of the returns by controlling for january. table 3 presents the results. the 't ' variable, which represents january returns, is not significant. the "d" variable, which represents the congressional table 3 results for relationship between congressional effect and january 1897-1993 ret m a + ~1 d + ~2 j + e where d -0 if congress is in recess (closed) = 1 if congress is in session (open) j = 1 if january; 0 otherwise varic~le parameter estimate t-value a -0.066 -3.634 d -0.043 -2.685 j 0.043 1.586 r~-value 0.0003 0.0072 0.1127 24 financial services review 6(1) 1997 calendar, is still significant. these observations show that the congressional effect is not driven by january. iv. recent evidence of the congressional effect the dates that congress is in and out of session during 1984-1993 are isolated to observe the recent record of the market around the open and closed schedule. table 4-panel a shows an open congress for 1,652 days over the sample period and a congress in recess for 930 days. during this time, the change in the djia is +2672.74 points. table 4-panel b shows that the majority of this market rise (2346.92 points) occurs while congress is closed. the open sessions produce only 325.82 points. despite congress being in session almost 80% longer than in recess, over 87% of the market rise occurs during recess. this difference is significant at the .10 level. this result is similar to that for the entire sample period. table 4-panel c indicates an advance in the market in 73% of the periods in which congress is closed, and about 60% of the time it is open. this is not surprising, given that the general trend of the market is to advance. although the market moves upward a similar number of times in both sessions, the size of the movements is significantly different. as indicated in table 1, open returns are very small; closed returns are about thirteen times larger. the results are even more striking for the most recent six-year subperiod. since 1988, the change in the djia is +1966.34 points; 1870.75 of which occur while congress is closed. the market rise is only 95.59 points in the open period. over 95% of the rise" in the market between 1988 and 1993 occurs while congress is in recess. this difference is sig rdficant at the .05 level. furthermore, between 1988-1993, 83% of the sessions that con gress is closed accompany an increase in the djia. in only 49% of the sessions in which congress is meeting does the djia advance. perhaps the investing community is becoming even more skeptical when congress is in session. table 4 relationship between the performance of the djia and the congressional calendar 1984-1993 panel a: number of business days in the congressional calendar in session {open) 1984-1993 1,652 1988-1993 1,000 930 568 panel b: change in djia points in session/open) 1984-19931 +325.82 +2346.92 1988-19932 + 95.59 +1870.75 panel c: number of periods with a gain in the djia in session (ooeq) 1984-1993 37/62 (60%) 1988-1993 17/35 (49%) 45/62 (73%) 29/35 (83%) notes: i t-value is 1.701 of a difference between closed and open. 2 t-value is 2.196 of a difference between closed and open. the congressional calendar 25 v. conclusions the findings above provide evidence for the existence of a curious calendar effect: the rela tionship between stock market performance and whether or not the u.s. congress is in ses sion. the results indicate that almost the entire market rise since 1897 corresponds to the periods when congress is closed. an open congress sees only a small market rise. this behavior is amazing given that congress is open almost twice as long as it is closed. per haps this observation is due to the uncertainty generated while congress is debating policy, regulatory and procedural issues. the outcome of the various bills and items remains largely unresolved until passage of the pending legislation. it is, therefore, very difficult to predict the ramifications of possible congressional decisions until the final votes. on the other hand, when congress is in recess, no bills and regulatory matters are being formally debated or formulated. perhaps the market enjoys the temporary certainty exhibited by the absence of congressional decisions, and responds with positive movements. references ariel, r.a. (1987). a monthly effect in stock returns. journal of financial economics, 18(1), 161 174. banz, r.w. (1981). the relationship between return and market value of common stocks. journal of financial economics, 9(1), 3-18. french, k.r. (1980). stock returns and the weekend effect. journal of financial economics, 8(1), 55-70. gultekin, m.n., & gultekin, n.b. (1983). stock market seasonality: international evidence. journal of financial economics, 12(4), 469-482. hirsch, y. (1992). the stock trader's almanac, 25th ed. new york: penguin books. jaffe, j., & westerfield, r. (1985). the week-end effect in common stock returns: the international evidenc. journal of finance, 40(2), 433-454. keim, d.r. (1983). size-related anomalies and stock return seasonality: further empirical evidence. journal of financial economics, 12(1), 13-32. keim, d.r., & stambaugh, r.f. (1984). a further investigation of the weekend effect in stock returns. journal of finance, 39(3), 819-835. maturi, r.j. (1993). divining the dew. chicago, il: probus publishing. merrill, a.a. (1984). behavior ofprices on wall street. chappaqua, n'y: the analysis press. michelson, s. (1993). using congressional sessions to predict the stock market. journal of business and economic perspectives, 9, 89-99. reinganum, m.c. (1983). the anomalous stock market behavior of small firms in january: empiri cal tests for tax-loss effects. journal of financial economics, 12(1), 89-104. rogalski, rj. (1984). new findings regarding day of the week returns over trading and nontrading periods: a note. journal of finance, 39(5), 1603-1614. roll, r. (1983). vas ist das? journal of portfolio management, 9(2), 18-28. rozeff, m.s., & kinney, w.r., jr. (1976). capital market seasonality: the case of stock returns. journal of financial economics, 3(4), 379-402. smirlock, m., & starks, l. (1986). day-of-the-week and inwaday effects in stock returns. journal of financial economics, 17(1), 197-210. conceptualizing financial advice in australia: the impact of business models and external stakeholders on client’s best interest practice d. w. richardsa,*, e. f. mortona aschool of accounting, rmit university, melbourne australia abstract many individuals entrust financial advisors to navigate through important financial decisions, yet extant research and persistent scandals bring the effectiveness of advice into question. in examining the australian financial advice sector, we determine when financial advice can be provided in a client’s best interest and formulate a model differentiating types of financial advice and their relationship to best interest practice. we extend this model to form an integrated framework considering external stakeholders that encourage, and business models which prevent, best interest practice. our examination reveals some business models prioritize financial institution interests whilst thwarting external stakeholders encouraging best interest practice. © 2019 academy of financial services. all rights reserved. 1. introduction because of an aging population and increased longevity, the responsibility of being financially stable in retirement has moved from the state to the individual and is illustrated by a decrease in state funded pensions with an increase in mandatory personal retirement accounts (holzmann, 2013). creating a sound financial plan is difficult when considering the complexity of ones’ life and the myriad of financial products available. professionals, such as the financial advisor1, have a fiduciary duty—or a duty to act in a client’s best interest— *corresponding author: tel.: +61 3 9925 5935; fax: +61 3 9925 5624. e-mail address: daniel.richards@rmit.edu.au (d. w. richards) 1057-0810/20/$ – see front matter © 2019 academy of financial services. all rights reserved. financial services review 28 (2020) 133–158 when they provide services to navigate through financial decisions and in doing so mitigate or avoid risking conflicts of interest. yet, the financial advisors providing advice are employed or remunerated by financial institutions that have differing and arguably conflicting interests; and, in particular, financial interests in attaining consumers of the products they create. as such, financial advisors are inherently caught between acting in a client’s best interest and their own interests, or the interests of financial institutions. this research investigates the extent to which financial advice in australia can be provided in a client’s best interest. we address this from a contextual perspective, where we identify the impediments and enablers of financial advisors achieving client’s best interest practice. this research investigates dimensions that distinguish types of financial advice, the business models used in providing financial advice and external stakeholders in the australian financial services industry that may induce or mitigate prioritizing a client’s best interest. in order to address these topics, we conducted a qualitative study encompassing content analysis, in-depth interviews and triangulation. firstly, the content analysis included analyzing australian government inquiries into financial advice (parliamentary joint committee on corporations and financial services (pjccfs) inquiries). then in-depth semi-structured qualitative interviews were conducted with financial service professionals to test the validity of our findings emerging from the content analysis. finally, results were then triangulated with an australian royal commission into misconduct in the banking, superannuation and financial services industry (henceforth referred to as ‘royal commission’). using these data, we conceptualize financial advice, and its purported conflict, in relation to acting in a client’s best interest. the contributions of this article are threefold. firstly, we develop a general model of best interest practice in financial advice, showing the extent to which a client’s best interest can be prioritized is due to the orientation of advice (product advice or personal advice) and alignment of the advisor (integrated with or independent from product providers). as such, this model conceptualizes the differing forms of financial advice and clarifies which type of advice can prioritize a client’s best interest. the second contribution addresses the dearth of understanding through the inclusion of the contextual environment that influence financial advisors. extant research is largely international and shows presence of conflicts (foerster, linnainmaa, melzer, & previtero, 2017; hackethal, haliassos, & jappelli, 2012; hoechle, ruenzi, schaub, & schmid, 2018), but simplifies explanation of contextual issues of financial advisors work to remuneration (angelova & regner 2013). while remuneration influences how financial advisors operate, this article argues that an understanding of the australian business models in which financial advisers operate in, as well as external stakeholders that impose numerous levels of regulation and influence upon the financial advisor, are necessary. we find that within the australian context, business models and external stakeholders can enable or inhibit the ability of a financial advisor to act in a client’s best interest. the third contribution of this article is the articulation of the interaction between the nature of advice provided, the business models and external stakeholders. this is explained through the development of a holistic and integrated framework. the framework illustrates how initiatives by external stakeholders aimed at achieving client’s best interest practice are impeded by certain business models. as such, this framework offers more meaningful 134 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 insights and implications for policymakers, financial planners, and clients by conceptualizing where initiatives should be targeted to effectively influence changes in the financial advice industry in australia. therefore, this article offers timely clarity on current australian initiatives that are thwarted by structural barriers and, therefore, are at risk of becoming blunt tools of change. the remainder of the article is structured as follows. next is a review of the literature on financial advice and client’s best interest. after this, the research objectives and methodology are outlined. next, the findings are outlined in three sections: a general model of a client’s best interest in financial advice, three business models in financial advice and the external stakeholders influencing best interest practice. finally, this analysis is brought together through an integrated framework, followed by concluding comments. 2. literature review 2.1. conceptualizing the “financial advisor” in respect of financial advice 2.1.1. the scope of the “financial advisor” our article purposely uses the holistic term of “financial advisor” to encompass the many occupations and professions that offer financial advice. for example, the following occupations and professions have an element of financial advice in their work: financial planner, investment advisor, wealth advisor, tax advisor, stockbroker, financial product sales agent, mortgage broker, real estate agent, accountant, or lawyer. from this perspective, advice can be considered financial advice if “it is intended to influence a person or persons in making a decision about a particular financial product” (australian securities & investment commission (asic) 2016: 5). our holistic perspective can be contrasted with a narrower focus, such as on the financial planner alone within the international setting. warschauer (2002: 204), considered financial planning in particular, noting that “one can be a financial adviser or consultant or give financial advice without being a financial planner, but one is not practicing as a financial planner without the elements of the definition and the process intact.” our perspective focuses on the outcome of the provision of financial advice, rather than the provider process, and as such captures numerous providers of financial advice. in that sense, both a lawyer and a stockbroker are providing financial advice, but the advice they provide may differ greatly. in doing so, we can “capture the diversity and complexity of financial planning engagement” (heckman, seay, kim, & letkiewicz, 2016: 442) more broadly. 2.1.2. client’s best interest an overarching element of financial advice is that clients seek advice to aid in making financial decisions because of an inability to undertake these decisions wholly by themselves. prioritizing a client’s best interest is an element of financial advice because the client is entrusting some aspects of decision making to the financial advisor (inderst & ottaviani, 2012b). we use the terminology client’s best interest, but this is akin to fiduciary duty in d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 135 financial services (boatright, 2000), with the terms used interchangeably in academic literature (angel & mccabe, 2013) and financial advice practice (maley, 2018). boatright (2000: 202) states “fiduciary duties are the duties of a fiduciary to act in that other person’s interest without gaining any material benefit except with the knowledge and consent of that person.” with this in mind, a client’s best interest duty is obliged in financial advice but has some limitations. firstly, this duty obscures the type of advice being provided and whether a client’s best interest obligation is enacted. some advice, such as tv commercials of financial products, by their nature, are presenting information. yet, this information provider cannot aim to be acting in a client’s best interest as the client’s interest is unknown. however, such information is used to influence financial decisions and should be considered financial advice. secondly, there is a growing amount of research that suggests that a client’s best interest duty is not being met by financial advisors. however, this reflects an international context and because of local regulatory environments poses issues of generalizability. 2.2. an international perspective on conflicts of interest examining the international setting, extant research reveals conflicting effectiveness of financial advice. relevant research shows that many individuals entrust financial advisors (or financial planners) to help navigate through financial decision making (e.g., within the italian context: calcagno, giofre, & urzi-brancati, 2017). however, inconsistency can be seen in the effectiveness of financial advisors and their advice. foerster et al. (2017) find that on-going commission-based fees charged by financial advisors reduce investment performance for investors in canada. in germany, hackethal et al. (2012) find that portfolio performance of investors who use bank financial advisors is inferior to independent financial advisors. whilst in switzerland, hoechle et al. (2018) find that advised clients incur worse returns than independent clients and that advised transactions focus on mutual funds which are more profitable for the advisor’s employer. these findings can be compared with other jurisidictions, where research advocates the benefits of financial advice in creating less bias, improved returns and more diversified portfolios (kramer, 2012; shapira & venezia, 2001; von gaudecker, 2015). further international research reveals that organizational structure (within a firm) and the local regulatory environment are important in considering the effectiveness of financial advice. within the u.s. setting, bigel (2000) finds that a longer tenure as a financial planner is positively related to a decrease in ethical reasoning ability, yet individual incentives (compensation via commission v. fee for service) are not. mazzoli and nicolini (2010) compare different compensation structures across the financial advice industries in italy and the united states. they find that more opaque pricing strategies occur when financial advisors are tied to financial institutions and product providers. independent financial advisors are more likely to have transparent fee structures and adopt a fee for service style of compensation. an argument presented to explain why a client’s best interest is not being met is that material benefits (commissions) provided to financial advisors negate this obligation. for example, inderst and ottaviani (2009, 2012a) use economic models of financial advice to show the inherent conflicts of interest when incentives from product providers are high. similarly, 136 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 angelova and regner (2013) show that truth telling decreases in the presence of commissions. however, material benefits encapsulate some, but not all, of the contextual issues which impede financial advice from achieving client’s best interest practice. šindelá�r and budinský (2017) find that commissions do not apply homogenously in the czech republic. in particular, they note that organizational structures need to be considered as the relationship between commissions and quality of financial advice only hold in flat-business structure models (šindelá�r & budinský, 2017). international literature also suggests disclosure is not an effective method for negating conflicts of interest in financial advice (chater, huck, & inderst, 2010; rubin, 2015), because trust binds clients to their financial advisors (gennaioli, shleifer, & vishny, 2015). that is, the trust that a client has in their advisors will allow clients to use financial advice when conflicts of interest are disclosed. this is consistent with empirical research in italy, which shows there is a high level of trust reported by clients of advisors and if a client’s trust is higher, they are more inclined to delegate decisions to financial advisors (calcagno et al., 2017). calcagno et al. (2017) find that 74% of clients that hold risky assets do not attempt to influence their financial advisors. out of the 26% investors that do influence decisions, they do so by monitoring (19%), obtaining second opinions (3.4%), or both (3.3%). these results indicate that most clients trust their advisor’s financial advice and that they will not enforce a strict client’s best interest practice on financial advisors. within this international setting, there is a variety of regulatory frameworks and contrasting contextual factors. therefore, this limits the potential generalizability to the australian context. moreover, within the australian context there is minimal empirical research considering the ethical behavior of financial advisors (cull & bowyer, 2017). 2.3. the australian financial advice industry the australian financial advice industry has experienced rapid growth since 2002, driven by social, cultural, institutional, political and economic factors (cull, 2009). following a government inquiry in 1996 (called thewallis inquiry), australia adopted an authorized representative governance model to regulate this growing industry (mcinnes, 2020). under this model, an australian financial services licensee (henceforth called a ‘licensee’) can authorize a financial advisor to give advice on their behalf and also takes responsibility to ensure that their financial advisors comply with relevant legislation. the effectiveness of this model is under question as extreme market conditions, numerous scandals and conflicts of interests have caused havoc in australia’s financial advice industry (north, 2015). mcinnes (2020) researched the legitimacy of the authorized representative governance model and highlights that problems of independence of financial advisors from product providers have persisted since 2009. mcinnes surveyed financial advisors in australia finding that financial advisors did not report having independence as an authorized representative. this is a critical issue given the importance of financial advisors within the australian context. clients often trust financial advisors to navigate complexities within the financial environment, especially when many consumers have low levels of financial literacy (brimble & murphy, 2012). hunt, brimble, and freudenberg (2011) find that trust in the relationship between clients and financial planners is critical for relationship quality. a lack of d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 137 independence can break this trust and, unsurprisingly, australia has seen an increase in regulation and scrutiny to increase trust in the industry (cull & sloan, 2016). recently, cull and bowyer (2017) find that clients rank ethical values as more important than technical competence in a financial advisor. 2.4. client’s best interest in australia looking specifically at the regulatory context, at the most basic level, financial advice is governed by the laws of equity as well as common law, imposing duties on a financial advisor creating an obligation to act in a client’s best interest. common law places a duty of care on professionals in contract and tort. glover (2002) argues that when professionals possess special expertise, they must render services to a client according to ‘prevailing community standards’ because they a liable in contract or in tort. as the financial advisor is providing expertise to clients, they are liable to ensure advice is provided at community standards. there is also legal precedent of a fiduciary requirement more generally in financial services.2 the community standards can be ascertained from professional association’s exhortations and legal requirements which stress a client’s best interest standard to resolve and avoid conflicts of interests (glover, 2002; batten & pearson, 2013). the two major professional bodies are the financial planning association of australia (fpa) and the association of financial advisers (afa) whose code of ethics include principles such as client first (fpa, 2013) and best interests (afa, 2018), respectively. in addition, all financial advisors are governed by a new code of ethics, established by the financial adviser standards and ethics authority (fasea), which includes the ethical behaviour. “you must act with integrity and in the best interests of each of your clients” (fasea, 2019). moreover, this code has legislative backing via section 921e of the corporations act 2001 (cth) (‘corporations act’), requiring all financial advisors to comply with the code. a client’s best interest obligations was also legislated in part 7.7a corporations act, following the future of financial advice (fofa) reforms in 2012 (north, 2015). the reforms cover best interest obligations (division 2); charging of ongoing fees (division 3); conflicted remuneration (division 4); and banned remuneration (division 5). for example, section 961b(1) states that “the provider must act in the best interests of the client in relation to the advice” and s961b(2) details a “safe harboring” process to show satisfaction of this duty. moreover, s961j specifies that priority must be given to the clients’ interest where there is a conflict and the advisor may face civil penalties if this duty is contravened (batten & pearson, 2013). in summary, extant literature suggests that local organizational structures and regulatory environments are important factors influencing whether financial advice is provided in a client’s best interest. within the australian setting, there is a developed financial advice industry that has undergone significant regulatory reform, yet pervasive misconduct has continued (the royal commission offers the most recent summary). despite this, the australian setting is infrequently researched or theorized (exceptions being cull, 2009; mcinnes 2020; north, 2015) and relevantly, despite having established client’s best interest obligations for financial advisors, impediments in meeting this duty continue to persist. this research, therefore, 138 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 investigates this setting holistically by researching when financial advice can be provided in a client’s best interest. it also researches the business models and wider environment, including key external stakeholders, that can enable and/or impede the providing of financial advice in a client’s best interest. in doing so, the research aims to reveal the stakeholders that encourage, and the business models which prevent, client’s best interest practice. 3. methodology we undertake an exploratory and mixed methodology that utilises both document analysis and interviews to achieve the research aims (morse, 2003). the exploratory qualitative methodology was adopted so that theory could be generated from data. the research process was carried out in three phases. the first phase involved content analysis of archival data, including regulatory documents on financial advice, to ascertain how financial advice was regulated and governed in australia. this data was pjccfs inquiries into financial advice (see appendix 1, table 1). the data were coded using nvivo 11 into three broad categories: (1) professional challenges facing financial advice, (2) state regulation and other control, and (3) the professional community. using these data, theories on the regulatory factors influencing how financial advice is governed were developed so that they could be explored through the second phase, indepth semi-structured qualitative interviews. the interviews were conducted with eight financial advisors and one governance director (refer to appendix 1, table 2) to gain insight into the contextual factors of financial advice in practice. as sandy and dumay (2011: 255) note, “the benefit of the research interview lies in its unique ability to uncover the private and sometimes incommunicable social world of the interviewee, to gain insight into alternative assumptions and ways of seeing.” this allowed us to gather further information outside of the archival data, including interpretations and perceptions of pertinent participants in the financial advice industry. as such, this enabled us to augment the identified business models and influences developed in phase one of the project. comparable to cull and bowyer (2017), interviewees were selected randomly from financial advisors in victoria, australia, verified by asic and contactable through public registers of financial advisors, including the yellow pages, the fpa and afa websites.3 further participants were then contacted through referrals from other participants. as such, the cohort represents financial advisors currently practicing in the field and excludes advisors not currently authorized by a financial service license holder to practice. the interviews were conducted in person and the average length was 54 minutes (mean), ranging between 33 and 79 minutes. ethical approval was obtained before interviews being conduct and participant anonymity has been maintained. in these interviews, open-ended questions on the contextual factors surrounding financial advisors were posed so that respondents could explain how they perceived financial advice was being influenced by various parties (see appendix 1, table 3). business models of financial advice were presented in the interview to verify their accuracy and changes were made iteratively. this process of beginning with open-ended questions, followed by targeted d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 139 questions can be described as direct content analysis (hsieh & shannon, 2005). here, the goal of the interview process is centered around validating or extending the theoretical framework or theory proffered (hsieh & shannon, 2005). in particular, the findings from the interviews support and extend the theories developed from the document review process. we use this process to verify a general model of a client’s best interest in financial advice, business models, and contextual factors that can explain client’s best interest practice. in the final phase of this research, triangulation was undertaken using the royal commission conducted in australia (interim report published in 2018 and final report in 2019). these reports offered new insight into how financial advice is regulated and governed in australia and, therefore, offered a third source of data for this research project. we reviewed the royal commission reports and reconciled their key findings with the findings of this research project. triangulation strengthens the validity of the research findings and, therefore, enabled the refinement of the models and framework presented in this article. 4. findings 4.1. two dimensions in a client’s best interest in financial advice the analysis of the government inquiries into financial advice revealed two dimensions underpinning the giving of financial advice in a client’s best interest. these were orientation and alignment. 4.1.1. orientation and alignment an important distinction made in defining financial advice is between personal advice and general advice. this is summed up succinctly by this quote: the corporations act 2001 defines ‘personal advice’ in section 766b(3) as financial product advice given or directed to a person (including by electronic means) in circumstances where: the person giving the advice has considered one or more of the client’s objectives, financial situation and needs . . . “general advice” is defined in section 766b(4) as financial product advice that is not “personal advice.” (report 3: 18) here the differentiator between personal and general financial advice is the information that the advice is based on. we refer to this as orientation and it is the extent to which a financial advisor is basing their advice on product characteristics or a client’s characteristics. financial advice which focuses on the product is conveying the essential information about that product to possible users of the product. the basis of the information given is on the technical details of the product. financial advice that focuses on the client’s characteristics takes into consideration a client’s financial situation, objectives and risk tolerance to formulate the advice given. that is, the basis of the advice provided is taken from the client’s situation and it matches products or a strategy to their situation. we conjecture that orientation is considered a continuum, where variability exists between these two positions as financial advice can be more productor more client-focused. 140 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 a second differentiator that came under scrutiny in many government inquiries was whose interests are being served. this can be illustrated by the following quote: where an adviser is employed by, or aligned with and acts on behalf of, a principal who manufactures or sells financial products, the adviser’s interests (and the principal’s) will be advanced by persuading a client to acquire one of the principal’s products. (report 2: 18) therefore, in one extreme a financial advisor will serve the interests of product providers and in the other extreme a financial advisor is independent of product providers. we use the term alignment to represent differences in the relationship between a financial advisor and a product provider. an integral aspect, but not the sole aspect, of alignment, is the remuneration model used to conduct financial advice. this perspective is noted in report 2 (:190): although the most obvious conflicts of interest affecting the provision of financial advice are the conflicts between an adviser’s duty and his or her financial interests, they are not the only conflicts. other conflicts can also arise from the associations or relationships between a financial adviser and the issuer of financial products. at one end of the alignment continuum, the financial advisor is fully aligned with product providers and the product provider is the financial advisors’ sole source of income. in this situation, a financial advisor does not charge a client for advice but receives their income from products providers. the other extreme is where the financial advisor is independent of product providers and their income is generated from charging fees to clients, even at a flat rate. variability exists between these two points where a financial advisor will earn income from both product providers and clients or may charge fees depending on the value of assets under management. furthermore, another influence on alignment is the business model (reviewed in section 4.2) used to provide financial advice, as this engenders a close alignment with, or independence from, financial product providers. 4.1.2. a general model of a client’s best interest in financial advice in the government inquiries these two dimensions of financial advice are conflated as being one dimension. for example, report 3 proposes to rename general financial advice as sales or product information to resolve misleading representation and conflicts of interest. however, our research finds that structural issues run deep, as illustrated by the following quote: the industry is still structurally broken because . . . the industry was born out of product providers wanting to distribute their product, (and) financial planners were merely just . . . a part of the distribution arm. (participant 2) an improved model is needed to conceptualize how financial advice is provided in a client’s best interest. we propose the dimensions should be intersected to construct a general model of a client’s best interest in financial advice. fig. 1 presents a graphic form of this model. these intersecting dimensions lead to four categories of financial advice we term: brokers, agents, advisors, and best interest advisors (bias). using this model, it is possible to classify occupations that provide financial advice and the extent to which they can achieve client’s best interest practice. d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 141 brokers refer to those financial professions where alignment is independent of product providers and the information provided is specifically product advice. an example of this is a stockbroker. a stockbroker provides information about stocks to clients; however, is independent of the stocks being recommended. brokers will have remuneration models that are separate from product providers, such as charging commissions to clients for each transaction. one of the impediments to stockbrokers achieving best interest practices is its orientation towards product (stock) based information. agents refer to those providing financial advice that is product based and connected to the product provider. an example is a real estate agent selling investment opportunities. these agents give specific information on a product (real estate) that is not based on a client’s circumstances. the agent receives remuneration from the product provider, often when a transaction has been completed. a real estate agent is far from achieving client’s best interest practice as their alignment is specifically with the product provider (seller of the property) and orientation is focused on the product. an antithesis example in real estate is a buyer’s advocate who focuses specifically on their clients’ needs and has no orientation towards a product provider, because they will source multiple products. advisors are those offering personal advice to clients and are aligned with product providers. an example is a financial planner working for a large financial institution. in this position, they offer personal advice to clients as they ascertain a client’s situation, objectives, and risk profile to create their advice. however, the financial advisor is being remunerated and is environmentally constrained by the financial institution. thus, their recommendations will be mostly products within the range of products that their financial institution controls. advisors have high conflicts of interest as they must serve the product provider’s interests and attend to a client’s needs. lastly, the bias, like a broker, is independent of product providers but offers personal advice matched to a client’s needs. an example of a bia would be an independent financial planner. the independent financial planner offers client centered financial advice and is independent of product providers, as they receive no remuneration from, and are not environmentally constrained by, product providers. instead, the financial planner may charge a feefor-service model where they charge clients based on the quantity of services rendered or a flat fee for being a client of the advisor. fig. 1. a general model for providing financial advice in a client’s best interest. 142 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 a limitation worth noting with this model is that it shows when a client’s best interest can be prioritized but does not mean they necessarily will achieve best interest practice. this occurs because conflicts of interest still exist for bias, who are independent of product providers and provide client-based advice. for example, consider a situation where a financial advisor provides advice using a fee-for-service model. this advisor will gain material benefits by charging fees and providing the smallest amount of service possible, whilst using the time saved to attract new additional clients. this issue was revealed in report 1 and report 2 as “fees for no service,” where ongoing advice fees are being charged with no advice being given to the client. these reports concluded that poor advice often given to clients has ultimately led to those clients being left worse off than if they had been given proper advice. 4.2. business models of financial advice international research highlights the importance of organizational structure for financial advice, but a key theme emanating from our research is that the business models of financial advice was important for providing financial advice in a client’s best interest. yet, these models differ from one advisor to the next, because of the arrangement of three major parties involved. these parties are: the licensee (the organization or person licensed by asic to provide financial advice services to the public); the financial product provider (the organization that makes the product a client uses); and the financial advisor (the person who recommends the product to the client). as previously indicated, australia operates an authorized representative model, with the licensee able to authorize a financial advisor to give advice on their behalf whilst taking responsibility to ensure compliance with relevant legislation. an important aspect of these models is that the work of a financial advisor is largely governed by the licensee because of several factors. firstly, the licensee sets the terms under which financial advice will be given. that is, they create the contract between the financial advisor and their clients, referred to as the financial services guide (fsg). an fsg is a general document provided at the commencement of an advice relationship, which must outline the kinds of financial services and products the licensee is authorized to provide, as well as any remuneration, commission and other benefits that may be received by the providing entity as a result of advice being offered and any potential conflicts of interest. (report 5: 3) secondly, a licensee will also issue the financial advisor with an approved product list (apl) that outlines the financial products that a financial advisor can recommend. participant 6 describes this as: again, another subtle way that providers will try to encourage advisors to use their products and their platforms over others will be . . . apls; (these) can be quite restrictive. (participant 6) finally, the licensee will stipulate professional development requirements, suggest financial services software and conduct audits of the financial advisor practices. thus, when assessing the extent to which financial advice is provided in a client’s best interest, it is important to consider the organizational context of the licensee, financial advisor, and financial product providers. d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 143 the main three organizations involved with providing financial advisory services can be arranged into three different business models of financial advice, based on the relationship between parties. these models were shown, amended and confirmed in the interviews. fig. 2 is a graphical illustration of these three models. 4.2.1. model 1: all groups are the same organization model 1 is a business model in which the financial advisor, financial product provider and licensee are the same organization. this occurs when a large financial institution, such as a bank, creates products and obtains a license to give advice and employs financial advisors. in this model, the alignment of financial advisors is close with product providers as the remuneration is being provided by that organization. there is also difficultly in this business model distinguishing between the provision of general (product-orientated) advice and personal (client-orientated) advice. although a financial advisor will provide personal advice, there may be several other employees (e.g., such as bank tellers) who provide general advice about the products the organization offers. 4.2.2. model 2: all groups are different organizations in model 2 all parties work in separate organizations. in this model, the financial advisor will engage the services of a licensee (often referred to as a dealer group) to obtain authorization to give financial advice. from a legal perspective, the licensee will work as a separate organization to the financial product provider. however, the relationship between the fig. 2. the three business models of financial advice: licensees, financial product providers, and financial advisors. 144 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 financial product provider and licensee varies significantly in practice. at one end of the continuum, the licensee is simply a subsidiary of the financial product provider. consider this quote from a financial planner working in model 2: there’s two types of product providers. there’s, i call them “internal” and “external”. so, with my business being licensed through [company name withheld; licensee] any [company name withheld; product provider] product i consider it internal, because [licensee] are owned by [product provider], anything else is external. (participant 7) in practical terms, model 2 resembles model 1, with independence occurring only in a legal sense and not as separate structures. in contrast, at the other end of the continuum, the relationship between the financial product provider and the licensee is independent. consider this quote from a financial planner in model 2: you might use product a, b, or c, . . . you have a relationship with the product provider, you need to know about it and these types of things, but there’s no contractual obligation in any way, shape or form to use that product. (participant 5) from a client’s point of view, the relationship between the financial product provider and licensee is not obvious and it is difficult to ascertain the alignment of financial advisors with product providers. clients may not be aware of this structural relationship until the product provider is disclosed in documentation and may not appreciate that nature of this relationship. nonetheless, business model 2 is orientated towards providing personal advice rather than product advice. it is important to note that there are pressures on licensees to align with product providers and not be independent. this is noted by this quote: it is difficult to run a profitable dealer group (licensee) in financial planning. so you do often need to create some sort of deals, that is where the pressure comes . . . it is very hard for independent dealer groups . . . where the only revenue they would be getting is from advisor only . . . you have to come up with arrangements, that can add to the bottom line or they can pass onto advisors for better fees or whatever. (participant 2) 4.2.3. licensee and advisor are the same organization the final model is model 3, which involves a financial advisor’s organization, or the financial advisor themselves, obtaining their own financial services license and then forming a relationship with one or many financial product providers to recommend products. consider this quote from a financial planner: i think independence is really important. so we have our own license and we have from the start. . .. i think being independent and not being incentivized to recommend a particular product or strategy to a client is most important. (participant 1) model 3 can be seen as the most independent model of financial advice because the licensee is not aligned with any financial product provider, but the licensee is aligned with the financial advisors. it is common in this model that the financial advisor will operate in a feefor-service model to remove remuneration from product providers and reduce conflicts of d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 145 interest. this model also specializes in giving personal advice rather product advice. however, in contrast with the first two models, model 3 involves more costs for the financial advisor as they incur the cost of obtaining a license. it also contains the highest amount of upfront expense for a client whose fee will be paying for the financial advisor to comply with regulations. in summary, of the three business models, model 3 will most likely enable best interest practice to occur as obtaining an individual license removes the alignment of a financial advisor from a product provider. however, this model is the least common: approximately 85% of financial advisers are associated with a product manufacturer, so that many advisers effectively act as a product pipeline. of the remainder, the vast majority receive commissions from product manufacturers and so have incentives to sell products. (report 6: 70) despite this, calls to individually license financial advisors have not been endorsed by parliamentary enquiries into financial advice. the organizational context and the licensee in particular can be considered the primary influence over the financial advisor’s financial advice due to having the most power to influence practices. however, in addition to the organizational context, there is a secondary influence stemming from external stakeholders that shape financial advice practices. these stakeholders seek to create or influence a client’s best interest practice in financial advice; however, experience corresponding impediments created by the particular business models. 4.3. external stakeholders advocating best interest practice wider contextual influences beyond the employing financial institution and product providers can impact how financial advice is provided. whilst there will be multiple, cultural and institutional elements which impact financial advice, in this review, we focus the search towards those people, organizations or government that aim to implement client’s best interest practice on financial advisors. analysis of the interview data narrowed this search down to clients, professional bodies, ombudsman, education providers, and the government (fig. 3). the stakeholders and their attempts to promote client’s best interest practice are now delineated. fig. 3. external stakeholders seeking to influence financial advice. 146 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 4.3.1. the client an essential element of financial advice is the relationship that the advisor has with their clients. this relationship is a two-way interaction where advisors will provide services to clients but also clients will have the ability to influence and control their advisor. thus, the client can be seen as a control mechanism to enforce best interest practice on a financial advisor. historically, the australian context of financial advice has placed a large emphasis on the client influencing financial advisors because of the disclosure practices legislated. some of the disclosure requirements of financial advisors include reporting clients’ information about their remuneration (including commissions), conflicts of interest with product providers and, more recently, an optin requirement for ongoing fee arrangements. a disclosure practice in financial advice assumes that this process will control the actions of financial advisors and financial institutions. it could be that clients will choose not to participate in financial advice when conflicts of interest are disclosed and/or the necessity to disclose such practices will stop conflicts from arising in the first place. however, when asked about client’s reading of the fsg, a disclosure document provided to clients, one financial planner noted: i don’t know if a client’s ever read it. i know well, from myself personally, if you go for a financial product through your insurance and these types of things, i’m sure i’ve never read them either and i work in the industry. (participant 5) furthermore, government inquires have found that disclosure practices have not been effective in stopping a persistence of scandals in the financial services industry. consider these quotes: . . . [t]here was a broadly held view that disclosure had been ineffective in managing conflicts of interest, necessitating other more robust measures . . . (report 6: 111) the reforms recognize that current disclosure requirements are not, on their own, sufficient to fully inform consumers. (report 2: 105) this suggests that clients will not enforce a strict client’s best interest practice on financial advisors, irrespective of disclosure requirements. moreover, the business models create impediments to the control mechanism proffered via the client-advisor relationship. with many clients being unaware of the true nature of the business model, as discussed in section 4.2, and the majority of business models not positioned optimally for best interest practice (bias), clients may inherently face barriers to meaningfully influence their financial advisor. in doing so, client influence can be directed mainly towards perceptions of best interest practice. 4.3.2. professional bodies another influence on the practice of financial advisors is the influence of professional bodies. professional bodies require financial advisors to ascertain certain levels of education and experience before being admitted to the body and, thereafter, undertake training and assessments to become accredited. despite the intention of professional bodies to influence financial advisor practices, and enforce client’s best interest practice, the ability to achieve this is thwarted. the reason for d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 147 this is the organizational context; the business models previously outlined. in this organizational context, direct power over a financial advisor is given to the licensee as they authorize the advisor to give financial advice. in other words, a licensee awards the ability to practice to a financial advisor and sets the conditions under which a financial advisor practice. a professional body’s code of conduct is a secondary influence, which can only be invoked once the licensee conditions are met. this order of priority is epitomized by the fact that a financial advisor did not have to be a member of a professional body to practice. just as there is no requirement for individual financial advisers to be registered by asic, there is also no requirement for advisers to be members of any particular industry association . . . both the fpa and the afa seek to advance the cause of financial advisers generally. each seeks to promote the creation and growth of financial planning and advice as a profession. both the fpa and afa now have processes and systems for discipling members. but the evidence before the commission did not show that either the fpa or the afa currently plays any significant role in maintaining or enforcing proper standards of conduct by financial advisers. (report 1: 208) moreover, there are many professional bodies within the financial services industry all competing for membership. until the shift of authority is moved away from licensees to professional bodies, their ability to enforce best interest practice on financial advisors will remain limited and influence perceptions of best interest practice. 4.3.3. the ombudsman a requirement for licensees in australia is that they must provide clients with both an internal and external complaint service. the term ombudsman is used here to refer to the external services bodies that offer external complaint services. reports 1 and 2 notes that these bodies amalgamated into a single external dispute resolution service; the australian financial complaints authority (afca). when asked if the ombudsman service influences practices in financial advice one respondent noted: no i don’t think it does. i don’t think advisors think too much about what may come of the advice they are giving . . . i think most day to day advisors, they say that the complaint schemes are there. it’s part of the industry and i think the advisors don’t think on it too much. (participant 2) whilst the ombudsman service could enforce best interest practice on financial advisors, its ability to influence financial advisors is limited because of its reactive nature. rather than being a proactive influence on advice, they adjudicate on any customer complaints only after the issues are raised. this reduces the effectiveness of an ombudsman service to enforce best interest practice on financial advice. it is possible though, that a decision made by an ombudsman could guide financial advice as they serve as examples of malpractice and the ombudsman could conduct outreach initiatives to influence financial advice. at the moment, such practices do not regularly occur in australia. the ombudsman service can also have an indirect influence over financial advisors as the advisors will require professional indemnity insurance, which is typically purchased by the licensee. if a financial advisor is involved with monetary resolutions from an ombudsman service, then the professional indemnity insurers will be involved to cover these costs. as a result, premiums for professional indemnity will increase following resolutions and the 148 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 licensee may review practices, including impediments to best interest practice, to reduce professional indemnity insurance premiums. however, this review of practices to increase a client’s best interest practice again relies on the licensee enforcing it. if a licensee is closely aligned with financial product providers, then it will not change practices to enforce a prioritization of a client’s best interest because of this alignment. thus, the ability for an ombudsman service to enforce client’s best interest practice is impeded by the license business model. 4.3.4. education providers education providers influence how financial advice is given by providing initial education and also providing ongoing professional development. a strong argument presented in the government enquiry data is that prerequisites to give financial advice, set at a diploma level, are too low and that one factor to improve financial advice should be to raise the qualification level. for example: lifting the qualifications of financial advisers and the standards of advice provided to consumers and investors becomes just one important defense mechanism to help reduce the risk of failure in the broader system. (report 3: 15) thus, the education level has been raised with new entrants to financial advice requiring a degree from 2019, whilst existing advisors have until january 1, 2024 to meet the new requirements (fasea, 2018). going forward universities will play more of a role in influencing financial advisors by educating future advisors on best practice in the workforce. education does present an ability to influence financial advisors towards client’s best interest practice. through rigorous ethics training, education providers could create the framework and discourse for considering a client’s best interest in financial advice. the ability to implement has previously been limited for two reasons. firstly, licensees could stipulate some of the cpd that a financial advisor undertakes and thus may not include courses on ethics and conducting advice in a client’s best interest. secondly, education does not fundamentally change the alignment between a financial advisor and a product provider that causes breaches in a client’s best interest practice. this relationship is controlled by the licensee who authorizes the financial advisor and controls the terms under which the financial advisor operates. thus, whilst education can offer an opportunity to improve client’s best interest practice of financial advisors, it has been thwarted by the context in which a financial advisor operates. 4.3.5. government in addition, the government also has two organizations it uses to enact legislation: asic and more recently, fasea. asic controls the financial services environment in australia and is responsible for enacting legislation by creating regulatory guides that govern financial advice. thus, it yields substantial influence over both licensee and a direct influence over financial advisors via a register it maintains (asic, 2017). the register provides the australian public with a list of all past and present financial advisors who are or have been allowed to practice in australia. the information on public display includes the advisor’s: name, unique identifier, status (current or ceased), licensee, qualifications, affiliations, products they can advise on, appointments, and any disciplinary actions. d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 149 asic can change an advisor’s status following disciplinary action where their right to provide financial advice is revoked. however, report 1 finds asic has a fragmented and ineffective disciplinary system for financial advisors and explains the reason for this where it states: asic said that, as a regulator, its role is to oversee advisers’ compliance with the law and not to supervise or monitor their work. it said that primary responsibility for discipline lies with licensees, who are responsible under the law for the conduct of their advisers. (report 1: 209) report 1 also noted that licensees were not sufficiently sharing information about advisors with asic. some research participants also noted issues with asic sharing information with them: so it’s really frustrating when you deal with asic. to be honest . . . beyond . . . lodging the required forms i have no involvement. we’ve not had an asic review done. i’d welcome it because . . . you get some great feedback and know when you’re on the right path. (participant 4) the second control mechanism by which the government directly influences financial advice is fasea formed in 2017. this authority comprises of members from industry, professional bodies and universities. its primary role is to implement changes to the corporations act in 2017 which raise the education, training, and ethical standards of financial advisers (fasea, 2018). fasea will govern via indirect and direct methods. indirect governance occurs through accreditation of universities financial advisory degrees to ensure financial advisors receive the education needed to improve the professionalism of the financial advice industry. direct governance occurs through enforcing a national exam, professional year, continuous professional development and a code of ethics to ensure the competency of financial advisors. thus, the government, also can influence financial advice through legislation, asic and, more recently, fasea. despite legislation being passed, a prevalence of misconduct still occurs in financial services. why does misconduct persist despite new legislation? we contend that a reason is that the relationship between financial advisors, licensees, and product providers remains unchanged. this relationship ensures that efforts by asic and fasea to control financial advisors come secondary to the relationship between a licensee and financial advisor. that is, even though a register, a degree, a national exam and a code of ethics are enforced on financial advisors, their ability to practice and the conditions under which they practice are set by the licensee. thus, the ability for government to create best interest practice in financial advice will be thwarted until the mechanism is stopped. to articulate the interaction between the nature of advice provided and the environmental considerations, this article now extends the broad applicability of the general model of best interest in financial advice to the specific context of australia, conceptualizing the environmental factors that enable (inhibit) client’s best interest practice in financial advice. 5. discussion 5.1. an integrated framework the findings of this research are used to create a general model of providing financial advice in a client’s best interests using two continuums: orientation and alignment. this 150 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 model shows that the conditions under which financial advice is most likely to prioritize a client’s best interest practice—and, therefore, more likely to be effective and beneficial to investors—is when orientation is client centered and alignment is independent from product providers. when these dimensions are aligned, they are labelled as best interest advisors (bias) in fig. 1. this model offers theoretical clarity to financial advice and policy implications on who should be held to a client’s best interest obligation. a second finding is that an understanding of the business model is required in order to determine the ability of a financial advisor to align with the bia classification. as described, financial advice in australia can occur via three different business models due to the licensing structure legislated and are outlined in fig. 2: • model 1 depicts financial advisors as being closely aligned with product providers because the licensee, product provider and financial advisor all work for the same company. as such, this model reduces the capability to prioritize a client’s best interest. • model 2 depicts all parties as independent and enables financial advisors to have independence from product providers, but the extent to which this happens is dependent on how the licensee operates. in one extreme the licensee could be a subsidiary of the product provider offering no independence for financial advisors and in the other extreme the licensee could be entirely independent of product providers. importantly, this relationship is difficult for a client to understand, even if disclosed to them. • model 3 depicts a business model where a financial advisor or his/her company, obtains its own license and forms relationships with one or more financial product providers. this last business model offers the most opportunity for prioritizing a client’s best interest, as obtaining an individual license removes alignment of a financial advisor from a product provider. however, this business model remains the rarest form of business model. whilst most of the business models (apart from model 3) reduce the ability of financial advisors to achieve best interest practice in financial advice, external stakeholders aim to encourage client’s best interest practice by financial advisors: fig. 3. however, despite these influences, the same business models inhibit their influences. these business models often enable licensees to have primary power over financial advisors, creating a barrier for stakeholder efforts to have an effect. any aims become secondary to the business model and, therefore, any initiatives lead to a perceived client’s best interest practice or inherent trust that are not conceptually aligned with bias, and as such do not necessarily result in effective financial advice to investors. layering these organizational and regulatory factors to create an integrated framework, enables the broad applicability of the general model of best interest in financial advice to be extended to consider factors that are enabling (inhibiting) in achieving clients best interest practice (fig. 4). inherent within the financial advice context is a narrowing of opportunity for client’s best interest practice to be achieved because of the business models in operation. 6. conclusion best interest practice is of manifest importance in financial advice, particularly as personal financial stability is increasingly the responsibility of the individual rather d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 151 than the state. we contend that in considering client’s best interest practice in australia, it is not only the individual financial advisor that is important, but also the influence of the contextual environment that the advisor operates within. often, the business models in place are conceptually positioned to prevent or impede financial advisors from becoming bias. as per the general model of a client’s best interest in financial advice, these models are less able to enable the orientation (client-centered) and alignment (independent from product providers) that is most likely to prioritize client’s best interest practice. as such, advice is positioned to be less effective and beneficial to investors. this article offers important implications for policymakers, financial advisors and clients. this integrated framework can enable more targeted initiatives towards the nature of the advice provided in the defined context. initiatives to promote a client’s best practice should be targeted at advice with an orientation towards clients and alignment independent from product providers (bia). other types of advice (agents, brokers, and advisers), should not be expected to achieve client’s best interest practice. creating clarity in roles identified in this framework would inform on the nature of the advice provided, enable more effective financial advice, and lead to better outcomes for investors. however, in doing so, the related business models need to be considered to ensure financial advisors are enabled to achieve client’s best interest practice rather than the perception of best interest practice. as such, this research continues the conversation towards business models and the risk of stakeholder initiatives functioning as a blunt tool for change. understanding that initiatives towards bias are substantially thwarted by certain business models (model 1 and model 2), suggests that for progress to occur, the change should be directed towards overcoming the barriers that business models create. fig. 4. the integrated framework. note: the dark grey shaded areas reflect enabled influences working towards the prioritizing a client’s best interest, whilst the light grey areas represent inhibited influences. 152 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 future research opportunities should consider how the business structures and the external regulatory context in other jurisdictions interact with the general model of best interest in financial advice, both in terms of determining the existence and pervasiveness of relative barriers. such examination would offer potential resolutions to impediments identified in australia or vice versa. considering the substantial controversy within the australian context, it is paramount that we further the conversation over best interest practice and progress financial advisors towards the status of bias. 6.1. international application of findings this research focuses on australia but there is global applicability of our findings because conflicts of interest in financial advice occur in switzerland (hoechle et al., 2018), germany (hackethal et al., 2012), united states (egan, matvos, & seru, 2019), and canada (foerster et al., 2017). the model presented in fig. 1 can be applied to many contexts and is useful in differentiating types of financial advice that are or are not provided in a client’s best interest. the business models outlined in fig. 2 are australian specific due to the licensing regime legislated there. the applicability of these business models to other contexts will depend on how financial advisors are authorized to practice in those contexts. comparing contexts in terms of how financial advice is governed would be a worthwhile pursuit because such research could ascertain strengths and weaknesses of current governance systems. finally, the various external stakeholders outlined in fig. 3 will be prevalent in other countries. however, the influence of each stakeholder over financial advisor will vary from context to context depending on their authority and the interplay between the relative organizational context and regulatory framework. acknowledgment the authors would like to thank professor chris robinson, associate professor eva tsahuridu and two reviewers for their comments on this paper. we also acknowledge the feedback provided from the presentations at york university, australian business ethics network conference (melbourne), rmit university and personal finance and investment symposium (melbourne). notes 1 here the term ‘financial advisor’ encompasses numerous consultant categories (financial advisors, financial planners, brokers etc.). the encapsulation of what a financial advisor is, is detailed in a later section. 2 see justice jacobsen’s interpretation draws from finn (1989: 46-47, cited in australian securities and investments commission v citigroup global markets australia pty limited (acn 113 114 832) (no. 4) [2007] fca 963 at 274). 3 these websites are www.yellowpages.com.au, https://fpa.com.au/; and www.afa. asn.au/find-afa-financial-adviser, respectively. d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 153 appendix 1 overview of data used for analysis table 1 parliamentary joint committee on corporations and financial services pjccfs inquiry reports and royal commission reports into misconduct in the banking, superannuation and financial services industry reports reviewed for this research report number citation submissions pages report name 1 hayne (2019) 10,323 530 final report royal commission into misconduct in the banking, superannuation and financial services industry volume 1 2 hayne (2018) 375 interim report royal commission into misconduct in the banking, superannuation and financial services industry volume 1 3 pjccfs (2014) 39 103 inquiry into proposals to lift the professional, ethical and education standards in the financial services industry 2014 4 pjccfs (2012b) 77 204 inquiry into the collapse of trio capital 5 pjccfs (2012a) 69 220 corporations amendment (future of financial advice) bill 2011 and corporations amendment (further future of financial advice measures) bill 2011 6 pjccfs (2009) 407 209 inquiry into financial products and services in australia 7 pjccfs (2003) 49 86 inquiry into the disclosure of commissions on risk products table 2 overview of the in-depth interviews interview number duration (hh:mm:ss) experience in financial services position company size 1 00:53:24 12 years financial planner large corporation 1000+ employees 2 01:12:05 29 years financial ombudsman representative/financial planner boutique 0–5 employees 3 00:51:55 17 years financial planner/partner medium firm 150+ employees 4 01:09:01 16 years governance director medium firm 150+ employees 5 01:03:39 10 years business owner/financial planner boutique 0–5 employees 6 01:03:41 5 years financial planner large corporation 50,000+ employees 7 00:38:25 17 years business owner/financial planner small firm 5–50 employees 8 00:46:36 28 years business owner/financial planner boutique 0–5 employees 9 00:33:43 14 years financial planner boutique 0–5 employees 154 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 t ab le 3 s em i st ru ct u re d in te rv ie w q u es ti o n s in tr o d u ct io n – e x p la in p u rp o se o f in te rv ie w h o w lo n g h av e y o u b ee n w o rk in g as a fi n an ci al ad v is o r/ p la n n er ? w h at p ro d u ct s d o y o u ad v is e o n ? in y o u r o p in io n , w h at ar e th e q u al it ie s th at m ak e a g o o d fi n an ci al ad v is o r/ p la n n er ? w h at d o y o u u n d er st an d b y “g o v er n an ce ” o f th e fi n an ci al ad v ic e in d u st ry ? g o v er n an ce : o rg an iz at io n al le v el h o w h av e y o u o r th e o rg an iz at io n y o u w o rk fo r ad o p te d th es e n ew g o v er n an ce re fo rm s at th e o rg an iz ati o n al le v el ? w h at m ec h an is m s ar e th er e in p la ce to en su re th es e re fo rm s ar e co m p li ed w it h ? g o v er n an ce : l ic en se (d ea le r g ro u p if se p ar at e fr o m th e o rg an iz at io n ) c an y o u o u tl in e y o u r li ce n si n g (d ea le r g ro u p ) ar ra n g em en t? w h at ar e y o u r v ie w s o n th e cu rr en t m et h o d o f li ce n si n g fo r fi n an ci al p la n n er s/ ad v is o rs ? if th ey ar e o r ar e n o t ef fe ct iv e, p le as e ex p la in w h y ? w h at d o y o u b el ie v e w o u ld im p ro v e th e li ce n se (d ea le r g ro u p ) m et h o d o f g o v er n an ce in fi n an ci al ad v ic e? g o v er n an ce : f in an ci al p ro d u ct p ro v id er s (i f se p ar at e fr o m th e o rg an iz at io n ) is y o u r fi n an ci al p ro d u ct p ro v id er an d li ce n se h o ld er th e sa m e in st it u ti o n ? r el at ed ? o r in d ep en d en t? in w h at w ay d o fi n an ci al p ro d u ct p ro v id er s in fl u en ce th e w ay y o u g iv e fi n an ci al ad v ic e? in w h at w ay is th e in fl u en ce o f fi n an ci al p ro d u ct p ro v id er s p o si ti v e fo r th e fi n an ci al ad v ic e in d u st ry ? w h at ab o u t n eg at iv e im p ac ts ? w h at ar e y o u r v ie w s o n th e cu rr en t m et h o d b y w h ic h fi n an ci al p ro d u ct p ro v id er s ar e g o v er n ed ? if th ey ar e o r ar e n o t ef fe ct iv e, p le as e ex p la in w h y ? g o v er n an ce : e d u ca ti o n an d p ro fe ss io n al b o d ie s w h at ar e y o u r v ie w s o n th e re q u ir em en t fo r p ro fe ss io n al q u al ifi ca ti o n s fo r fi n an ci al p la n n er s/ ad v is o rs ? if th ey ar e n o t ef fe ct iv e, p le as e ex p la in w h y ? a re y o u a m em b er o f a p ro fe ss io n al b o d y ? w h y o r w h y n o t? w h at ar e y o u r v ie w s o n th e re q u ir em en ts fr o m y o u r p ro fe ss io n al b o d y ? if th ey ar e o r ar e n o t ef fe ct iv e, p le as e ex p la in w h y ? g o v er n an ce : c li en ts w h at in te rn et /e x te rn al cl ie n t co m p la in t se rv ic es d o y o u h av e? w h at ar e y o u r v ie w s o n th e m et h o d fo r re so lv in g d is p u te s w it h cl ie n ts ? if th ey ar e/ ar e n o t ef fe ct iv e, p le as e ex p la in w h y ? d o th e cl ie n t co m p la in t se rv ic es in fl u en ce h o w y o u g iv e fi n an ci al ad v ic e? o v er al l: g o v er n an ce ef fe ct iv en es s w h at , in y o u r v ie w , ar e th e re fo rm s th at h av e h ad th e m o st p o si ti v e im p ac t o n y o u r p ra ct ic e? w h at ar e th e re fo rm s th at h av e h ad th e m o st n eg at iv e im p ac t o n y o u r p ra ct ic e? d o y o u b el ie v e th at th e re fo rm s o v er al l ar e b en efi ci al fo r th e fi n an ci al ad v ic e in d u st ry ? w h y o r w h y n o t? w h at ar e y o u r v ie w s o n th e cu rr en t p en al ti es fo r n o n -c o m p li an ce o f li ce n si n g , p ro fe ss io n al q u al ifi ca ti o n s? if th ey n o t ef fe ct iv e, p le as e ex p la in w h y ? o th er to p ic s o th er q u es ti o n s m ay b e as k ed as a re su lt o f th e an sw er s g iv en to q u es ti o n s ab o v e. d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 155 references american finance association. (2018). afa code of conduct. (available at https://www.afa.asn.au/afa-codeconduct). angel, j. j., & mccabe, d. (2013). ethical standards for stockbrokers: fiduciary or suitability? journal of business ethics, 115, 183–193. angelova, v., & regner, t. (2013). do voluntary payments to advisors improve the quality of financial advice? an experimental deception game. journal of economic behavior & organization, 93, 205-218. asic. (2016). licensing: financial product advice and dealing. brisbane, australia: australian securities & investment commission. (available at https://download.asic.gov.au/media/3889417/rg36-published-8-june-2016.pdf). asic. (2017). financial advisers register. brisbane, australia: australian securities & investment commission. (available at from https://www.moneysmart.gov.au/investing/financial-advice/financial-advisers-register). australian securities and investments commission v citigroup global markets australia pty limited (acn 113 114 832) (no. 4) [2007] fca 963. batten, r., & pearson, g. (2013). financial advice in australia: principles to proscription; managing to banning. st. john’s law review, 87, 511-559. bigel, k. (2000). the ethical orientation of financial planners who are engaged in investment activities: a comparison of united states practitioners based on professionalization and compensation sources. journal of business ethics, 28, 323-337. boatright, j. r. (2000). conflicts of interest in financial services. business and society review, 105, 201-219. brimble, m., & murphy, b. (2012). past, present, and future: the role of tertiary education in supporting the development of the financial planning profession. journal of business ethics education, 9, 105-124. calcagno, r., giofre, m., & urzi-brancati, m. c. (2017). to trust is good, but to control is better: how investors discipline financial advisors’ activity. journal of economic behavior & organization, 140, 287-316. chater, n., huck, s., & inderst, r. (2010). consumer decision-making in retail investment services: a behavioural economics perspective: final report. retrieved apr 1, 2019, from https://ec.europa.eu/info/ sites/info/files/retail_investment_services_2010_en.pdf. corporations act 2001 (cth). (2001). corporations act chapter 7 act no. 50. cull, m. (2009). the rise of the financial planning industry. australasian accounting business & finance journal, 3, 26-37. cull, m., & bowyer, d. (2017). ethics in financial planning: myth, fact or rhetoric paradox? e journal of social & behavioural research in business, 8, 56-69. cull, m., & sloan, t. (2016). characteristics of trust in personal financial planning. financial planning research journal, 2, 12-35. egan, m., matvos, g., & seru, a. (2019). the market for financial adviser misconduct. journal of political economy, 127, 233-295. fasea. (2018). fasea standards summary. fasea. (available at https://www.fasea.gov.au/wp-content/ uploads/2018/11/standards-summary.pdf). fasea. (2019). financial planners and advisers code of ethics 2019. fasea. (available at https://www. legislation.gov.au/details/f2019l00117). foerster, s., linnainmaa, j. t., melzer, b. t., & previtero, a. (2017). retail financial advice: does one size fit all? the journal of finance, 72, 1441-1482. fpa. (2013). the pillars of our profession: code of professional practice. financial planning association of australia ltd. (available at https://fpa.com.au/wp-content/uploads/2015/09/fpa_codeofpractice_july2013.pdf). gennaioli, n., shleifer, a., & vishny, r. (2015). money doctors. the journal of finance, 70, 91-114. glover, j. (2002). conflicts of interest, conflicts of duty and the information professional. adelaide law review, 23, 215-242. hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking & finance, 36, 509-524. 156 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 hayne, k. (2018). interim report royal commission into misconduct in the banking, superannuation and financial services industry volume 1. (available at https://financialservices.royalcommission.gov.au/ documents/interim-report/interim-report-volume-1.pdf). hayne, k. (2019). final report royal commission into misconduct in the banking, superannuation and financial services industry volume 1. commonwealth of australia. (available at https://www.royalcommission.gov.au/ sites/default/files/2019-02/fsrc-volume-1-final-report.pdf). heckman, s. j., seay, m. c., kim, k. t., & letkiewicz, j. c. (2016). household use of financial planners: measurement considerations for researchers. financial services review, 25, 427-446. hoechle, d., ruenzi, s., schaub, n., & schmid, m. (2018). financial advice and bank profits. the review of financial studies, 31, 4447-4492. holzmann, r. (2013). global pension systems and their reform: worldwide drivers, trends and challenges. international social security review, 66, 1-29. hsieh, h.-f., & shannon, s. e. (2005). three approaches to qualitative content analysis. qualitative health research, 15, 1277-1288. hunt, k., brimble, m., & freudenberg, b. (2011). determinants of client-professional relationship quality in the financial planning setting. australasian accounting, business and finance journal, 5, 69-100. inderst, r., & ottaviani, m. (2009). misselling through agents. american economic review, 99, 883-908. inderst, r., & ottaviani, m. (2012a). competition through commissions and kickbacks. american economic review, 102, 780-809. inderst, r., & ottaviani, m. (2012b). financial advice. journal of economic literature, 50, 494-512. kramer, m. m. (2012). financial advice and individual investor portfolio performance. financial management, 41, 395-428. maley, d. (2018). are advisors acting in clients’ best interests? the tussle over terminology. the globe and mail. (available at https://www.theglobeandmail.com/investing/globe-advisor/advisor-news/articleare-advisors-acting-in-clients-best-interests-the-tussle-over/). mazzoli, c., & nicolini, g. (2010). to fee or not to fee: pricing policies in financial counseling. financial services review, 19, 307-322. mcinnes, a. (2020). the regulation of financial planning in australia: current practice, issues and empirical analysis (first edition). abingdon, oxon: routledge. morse, j. m. (2003). principles of mixed methods and multimethods research design. in a. tashakkori & c. teddlie (eds.), handbook of mixed methods in social & behavioural research. thousands oak, ca: sage. north, g. (2015). the future of financial advice reforms: will they achieve their long term objectives?. competition & consumer law journal, 22, 197-217. pjccfs. (2003). inquiry into the disclosure of commissions on risk products. canberra: commonwealth of australia (available at https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_and_ financial_services/completed_inquiries/2002-04/dcrp/index). pjccfs. (2009). inquiry into financial products and services in australia. canberra: commonwealth of australia. (available at https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_ and_financial_services/completed_inquiries/2008-10/fps/index). pjccfs. (2012a). corporations amendment (future of financial advice) bill 2011 and corporations amendment (further future of financial advice measures) bill 2011. canberra: commonwealth of australia. (available at https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_ and_financial_services/completed_inquiries/2010-13/future_fin_advice/index). pjccfs. (2012b). inquiry into the collapse of trio capital. canberra: commonwealth of australia. (available at https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_and_financial_services/ completed_inquiries/2010-13/trio/index) pjccfs. (2014). inquiry into proposals to lift the professional, ethical and education standards in the financial services industry. canberra: commonwealth of australia. (available at https://www.aph.gov.au/parliamentary_ business/committees/joint/corporations_and_financial_services/financial_adviser_qualifications) rubin, g. d. (2015). advisers and the fiduciary duty debate. business and society review, 120, 519-548. d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 157 sandy, q. q., & dumay, j. (2011). the qualitative research interview. qualitative research in accounting & management, 8, 238-264. shapira, z., & venezia, i. (2001). patterns of behavior of professionally managed and independent investors. journal of banking & finance, 25, 1573-1587. šindelá�r, j., & budinský, p. (2017). evaluating the relationship between ifa remuneration and advice quality: an empirical study. financial services review, 26, 367-386. von gaudecker, h. m. (2015). how does household portfolio diversification vary with financial literacy and financial advice? the journal of finance, 70, 489-507. warschauer, t. (2002). the role of universities in the development of the personal financial planning profession. financial services review, 11, 201-216 158 d. w. richards, e. f. morton / financial services review 28 (2020) 133–158 pii: 1057-0810(91)90008-m financial services review, 1(1):61-78 copyright @ 1991 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. disability income insurance and the individual larry a. cox in this article, both institutional and research literature pertaining to disability income insurance issues is reviewed from the perspective of the individual. primary issues include the nature of the risk, the quantity of purchase decision, the choice of product, and the selection of supplier. an ultimate objective is to provide the reader with the foundation necessary to initiate future research efforts focusing on disability-related loss exposures. potential areas for future research are suggested. disability income (di) insurance may be the basic risk management tool most frequently ignored by individuals. despite wide acknowledgement of individual disability needs by authors of finance, insurance, and employee benefits textbooks, u.s. earners generally have not purchased the insurance necessary to counter the risk exposures implicit with disability. researchers similarly have tended to ignore di insurance issues, with the exception of studies limited to specific public programs. in this study, both institutional and research literature on di insurance issues is reviewed from the perspective of managing risk to the individual. the primary objective is to provide the reader with the foundation necessary to initiate future research efforts focusing on the many and varied di issues not adequately addressed to date. in the next section, the loss exposure facing the individual is discussed. subsequent sections examine the determination of insurance needs, sources of protection, and the selection of product and supplier. conclusions and potential research issues are summarized in the final section. thedisability loss exposure the event of disability can cause several types of economic loss for the individual. rejda (1984, pp. 199-202) suggests that loss categories include earned income losses, abnormally large medical expenses, loss of employee benefits, larry a. cox l department of insurance, legal studies, and real estate, college of business administration, the university of georgia, athens, ga 30602. 62 financial services review, l(1) 1991 and other expenses triggered by the disability of the earner, such as home nursing care or additional child care expenses. implicit in rejda’s discussion are the differences between the loss exposure caused by disablement versus that caused by death. with the event of death, the loss of the earner’s income is certain and survivors’ needs easily can be capitalized, given reasonable assumptions about the growth of the family’s needs, inflation, and investment rates.’ in the event of disability, however, the amount of loss is much less certain. the need for income replacement will depend upon the continuation of the individual’s disability (quantified by actuaries for large cohorts of individuals as “continuance factors”). thus, the disability loss exposure is more a bayesian problem. given that the loss occurs, the severity of the loss is contingent upon both (i) the individual’s ability to recover or be rehabilitated and (2) whether such recovery or rehabilitation is complete or partial. potential “additional” losses, such as rehabilitation, physical therapy, nursing, or child care expenses, pose a difficult planning issue generally not addressed by insurers. insurance to indemnify the individual for such losses, which often are excluded from medical expense insurance plans, generally is not part of the package of options offered by public agencies or private insurers. rather, di insurance benefits simply are tied to the individual’s income, either contractually or, implicitly, via the maximum monthly benefit allowed by the insurer’s underwriters. research literature on the loss exposures posed by the additional expenses caused by disability is virtually non-existent and represents an opportunity for interested researchers. because existing di insurance contracts primarily address the problem of earned income replacement, most research focuses upon the income replacement problem, as does this review. in the next section, the quantification of individual needs for di insurance and sources of supply are discussed. the quantity decision alternative risk management techniques before a purchase of any di insurance is considered, the individual must determine the extent to which disability loss exposures can be managed by implementing the other basic risk management techniques of (1) avoidance, (2) retention, (3) loss prevention, (4) loss control, and (5) combination. using the first technique, the individual can avoid many hazards, such as risky hobbies or occupations, thereby reducing his or her disability loss exposure. numerous potential perils cannot reasonably be avoided, however. an extreme implementation of the second method, retention of the entire risk, is not affordable to all but the very wealthy because of the possibility of permanent dkability income insurance and the individual 63 disability. partial retention, via a deductible included in a di insurance policy, is possible, however. for di insurance, the deductible is manifested in the elimination, or waiting, period provision. loss prevention entails reduction of the probability of a loss event. carefully managed behavior, e.g., taking proper safety measures at work and home, can reduce, but not completely eliminate the likelihood of a disability related loss. loss control techniques, professional physical therapy for example, may reduce the severity of loss if the disability occurs, but may have minimal impact in many cases and cannot reduce the probability of loss incidence. finally, combining loss exposures, e.g., marrying another earner or having a large family, may reduce the impact of disability-related loss of income.’ unlike an institutional venture, however, the individual will not be able to combine enough exposures to provide adequate diversification. in addition, the combination of exposures within a single family still may be subject to a common disaster, such as the disability of several family members caused by a single, severe automobile accident. the previous discussion indicates that alternative personal risk management methods may be used to reduce disability-related loss exposures and, possibly, the need for di insurance. the degree to which such alternative methods may substitute for an individual’s insurance requires further modelling and empirical examination, however. because the individual cannot completely control his or her environment or diversify adequately his or her source(s) of earned income, he or she normally is forced to consider the final risk management method, i.e., transfer of the loss exposure to a public or private institution via an insurance contract. determining the extent of di insurance usage is the difficult question discussed next. the magnitude of the need mehr and gustavson (mg) (1987, pp. 426-428) provide a simple procedure for estimating di insurance needs for the individual and his or her family. the method involves annual estimation summarized as follows: where: n = gross disability income insurance needs; e = total family expenses; x = extra expenses caused by the earner’s disability; i = post-tax family income after the earner’s disabiity; s = social security disability income (ssdi) benefits; and t = time period (year). 64 financial services review, l(1) 1991 the individu~ would purchase di insurance paying a monthly benefit in the amount indicated for the current year, iv,, less existing financial assets, but the benefit purchased would be adjusted over time following the needs estimation procedure contained in equation 1. after-tax estimates are used because di insurance policies generally provide tax-exempt benefits as long as the indi~du~ or his or her employer have not taken an income tax deduction for the premiums paid previously. mg do not discount the stream of annual estimates to the present, as would be the case for a life insurance needs estimate, because potential adverse selection problems, especially those created by moral hazard, prevent insurers from offering di insurance benefits in a lump sum. using a case example, mg demonstrate the potenti~ly great variability of disability-related family needs over time and note that benefits based on an earner’s present income may prove inadequate to meet the family’s future needs, even when a cost-of-living rider is attached to the policy. cox, gustavson, and stam (cgs) (1990) introduce an expected value model to show that, from the individual’s perspective, the need for income replacement caused by long-term disability (ltd) normally dominates the comparable need caused by death. although their model discounts future cash flows in order to compare the actuarial expected values of ltd versus life insurance objectives for individuals, cgs caution that the model is not applicable for personal planning purposes. they call for better planning models that impound loss frequency information and account for insurance portfolio effects. while generating preliminary evidence that individual needs for ltd insurance may dominate those for life insurance, cgs note that less than 22 percent of the public is covered by some form of long-term disability (ltd) insurance other than social security. price (1986, p. 8) indicates that nearly 60 percent of u.s. wage and salary workers have some form of short-term disability (std) income plan. in contrast, over 80 percent of u.s. households own life insurance (stemnock, 1988). the possibly inadequate response by individu~s to the demonstrated need for ltd insurance must concern the researcher interested in individual financial management because of the potentially catastrophic loss exposure accepted by the vast majority of u.s. earners. other problems observed in the u.s. marketplace include perceived product complexity, the paucity of standardized pricing, non-coordination of some private and public benefit programs, and availability problems for individu~s in certain occupational classes. the public interest, as well as intellectual curiosity, demands that both individual behavior and the informational efficiency of di insurance markets be studied carefully. next, the literature concerning the availability and efficacy of alternative sources of di insurance protection are discussed. disability income znsurwwe and the individual 65 sources of disability income protection shortterm disability benefits the majority of u.s. earners have some form of short-term disability (std) protection, defined here as coverage during the first six months of disability. private and governmental group plans provide 93 percent of std coverage for the non-occupationally disabled, while private individual plans account for only 7 percent (price, 1986). when studying either individual behavior or disability markets, the researcher should have a working knowledge of the sources and extent of std benefits that may accrue to individuals before the need for ltd income arises. a discussion of std income protection plans follows. social security. in the u.s., the old-age, survivors and disability insurance (oasdi) program, popularly referenced as social security, provides substantial disability income benefits to eligible individuals. the basic definition of disability requires that the earner be unable to perform any substantial gainful activity that exists in the national economy. substantial gainful activity now is defined as an activity generating more than $300 per month (schwartz and grundmann, 1989). the disability generally must be expected either to result in death or to last for a continuous period of not less than 12 months. if the disabled earner meets both oasdi eligibility requirements and the oasdi definition of disability, benefit payments may commence, but only after a five-month elimination period. disability benefits are determined as if the worker had reached normal retirement age and were eligible for retirement benefits, although the benefit formula differs somewhat from that used to calculate retirement benefits. workers’ compensation benefits may be fully or partially offset against the estimated oasdi payment if the total from both plans exceeds 80 percent of the earner’s average earnings prior to the disability occurrence. considering the lengthy waiting period and the required disablement expectation of one year, oasdi does not serve as a substantial source of std protection for the individual. workers’ compensation. the purpose of workers’ compensation laws, effective in all 50 states, is to bypass common law pertaining to negligence torts when a worker suffers an occupationally related injury or sickness. instead of suing his or her employer, the worker merely files a workers’compensation claim and, upon approval, the employer or its insurer must pay state-mandated disability income benefits, medical care expenses, and death benefits. in turn, the worker’s ability to sue for negligence is legally limited, as are the maximum benefits payable for disability income or death. 66 financial services review, l(1) 1991 in 1987, 87 percent of the u.s. labor force was covered under workers’ compensation laws (nelson, 1989). most states require that the disabled earner receive two thirds of his or her weekly earnings up to a maximum limit. the maximum weekly benefit ranged from $175 in georgia to $1,094 in alaska in mid-1988. typical elimination periods are three to seven days for disability income benefits, although these periods usually are covered on a retroactive basis if the disability continues for a minimum, specified period of time, generally from one to six weeks. the providers of workers’ compensation benefits include private insurers, employer self-insurance plans, and state-run insurance funds. four states require exclusive use of the state insurance fund, while four others require use of either the state fund or self-insurance. the remaining states allow use of private insurance as an alternative to either the state plan, if one exists, or self insurance. nelson (1988a) estimates that private insurers provide approximately 59 percent of workers’ compensation benefits in the u.s., while state and self insurance funds supply slightly over 20 percent each. state-mandated temporary disability plans. california, hawaii, new jersey, new york, rhode island, and puerto rico require temporary disability insurance (tdi), or cash sickness, plans for most earners. the plans are intended to partially compensate earners for income losses caused by either non occupational disability or maternity. although tdi is mandated only in the six jurisdictions, approximately one fourth of the u.s. commercial and industrial labor force is employed in these jurisdictions and, therefore, is covered by a compulsory tdi plan (kerns, 1989). tdi laws generally define disability as the earner’s inability to perform his or her regular or customary work, although new jersey and new york use a somewhat stricter definition. tdi plans usually provide at least half of the earner’s income up to a maximum benefit for a limited period of time. in 1988, the maximum weekly benefit varied from $104 in puerto rico to $252 in rhode island (kerns, 1989). benefit duration ranged from 26 to 39 weeks, with waiting periods of between three and seven days. most jurisdictions do not require payment of tdi benefits if workers’ compensation benefits are applicable, although california requires payment of tdi benefits to the extent they exceed workers’ compensation payments. the providers of tdi benefits encompass state-run plans, employer operated self-insurance funds, private insurers, and fraternal funds. the types of providers operating in each state often are determined by statutory law. at the extremes are rhode island, which mandates tdi only through the state run fund, and hawaii, which requires tdi coverage be provided only by private insurers or employer self-insurance. other jurisdictions allow tdi coverage to be provided via state and/or private plans (social security administration, 1990, p. 31 i). disability income insurance and the individual 67 private insurance. two types of private std benefit plans are dominant in the u.s. sick leave plans generally offer full replacement of earnings with no waiting periods and are funded by operating earnings of the employer. the earner often receives an allocation of five to 15 days of sick leave per year, but can accumulate unused days for the future. the second major source of std benefits is sickness and accident insurance offered via private insurers or employer self-insurance funds. sickness and accident insurance typically replaces two thirds of the earner’s weekly income after a waiting period of three to seven days. the duration of benefits usually is limited to between three and six months. sick leave plans provide a relatively higher level of benefits, approximately 76 percent of std income losses for plan beneficiaries in 1983. in contrast, price (1986) shows that the combination of individual and group sickness and accident insurance plans have provided only 35 to 38 percent of lost earnings for plan beneficiaries each year since 1970. houff and wiatrowski (hw) (1989) survey std plans available to employees of (1) medium and large private firms and (2) state and local governments. among employees of private firms, 46 percent have sick leave only, 24 percent have sickness and accident insurance only, 25 percent have both, and a mere 6 percent have no std coverage. for state government employees, 83 percent have sick leave only, 1 percent have sickness and accident insurance, 14 percent have both, and only 3 percent have none. for private employers, hw estimate that total std benefits from both sick leave and sickness and accident insurance plans would range from 53 to 65 percent of pre-disability income, depending upon the employee’s years of service. the duration of benefits is limited to between 110 and 134 work days, again depending upon years of service. plans funded with sickness and accident insurance only normally provide longer benefit durations, however. although nearly 60 percent of all u.s. earners are covered by some form of std plan, less than 48 percent of workers in the 45 non-tdi states have any such protection. most std benefits are provided through employer group plans, rather than through individual insurance policies. sick leave plans tend to provide higher income replacement ratios, but sickness and accident insurance is likely to pay benefits for a longer duration in the event of a relatively more severe std. because std plans are of limited duration, however, potentially catastrophic losses must be addressed by ltd plans, which are discussed subsequently. longterm disability benefits social security. for disabled individuals meeting the stringent definition of disability stated previously, social security disability income (ssdi) benefits provide an important source of income. packard (1987) finds that ssdi 68 financial services review, l(1) 1991 provides approximately two thirds of total income for unmarried, ssdi recipients and 40 percent of total income for the families of married recipients. despite this financial support, grad (1989) shows that fully 30 percent of the recipients remain near or below the poverty level. based upon the most recent census bureau data, the median household income of ssdi beneficiaries is only $850 per month, while net worth is $18,884 and only $2,635 if the family home is excluded. during the past decade, the probability of receiving ssdi benefits has declined for the disabled u.s. earner. after a substantial expansion of benefits in the 1960s and early 1970s reno and price (1985) and lando, farley, and brown (lfb) (1982) document the measures taken by the u.s. government to restrict ssdi payments. these incremental restrictions include (1) limiting income replacement ratios for some types of workers, (2) closely administering eligibility rules, and (3) lowering acceptance rates on applications for benefits. regarding the latter, lfb show that acceptance rates for initial applications for ssdi benefits declined from over 50 percent in 1966 and 1967 to less than 22 percent in the early 1980~.~ several studies analyze data from surveys of ssdi applicants and provide insight into the severity of disability losses for these applicants. hennessey and dykacz (hd) (1989) use a maximum likelihood weibull function to predict the ultimate fate of ssdi beneficiaries. independent variables included in the model are primary cause of disability, education, occupation, wealth, sex, race, and age. the results indicate that, after initial entitlement, only 11 percent will recover prior to retirement age 65,36 percent will die, and 53 percent will remain disabled until retirement. the mean duration of ssdi benefits to qualifying beneficiaries exceeds nine years. dykacz and hennessey (dh) (1989) then analyze the subset of ssdi benefici aries who recover. a two part weibull function is employed, using the predictive factors from hd. the authors predict that 43 percent of “recovered” beneficiaries will return to ssdi rolls within five years, 5 percent will die, and 52 percent will remain employed until retirement age, defined as age 62 in the study. the studies by hd and dh suggest that very few ssdi recipients will return to the job market and that nearly half of those recovering will again become disabled. the empirical analysis contained in these two studies is based upon surveys of beneficiaries conducted in the 197os, well prior to the stringent enforcement of eligibility rules and disability definitions that generally are affiliated with the reagan administration. recovery rates are likely to be even lower for current ssdi beneficiaries. halpern and hausman (hh) (1986) develop a model of individual choice for ssdi benefits and apply it to the previously cited survey data. a log likelihood function is implemented to test the impact of administrative policies regarding acceptance rates and benefit levels on the individual’s propensity to leave the labor force and apply for benefits. hh find that probable acceptance disability income insurance and the zndividud 69 rates do not have a large effect on the decision to apply, while benefit levels have a somewhat greater effect. nevertheless, the effects of benefit levels on the application decision are not as strong as previous research indicated. the type of disabling illness or injury, e.g., breathing difficulty or blindness, is found to have a very substantial effect on the decision to apply, however. bound (1989) addresses the issue of whether historically increasing ssdi benefits has created a disincentive for individuals to remain in the labor market. he focuses on the applicants rejected for ssdi benefits as a natural control group for comparison with ssdi beneficiaries. applying logit analysis, bound finds rates of labor force non-participation similar to those for ssdi beneficiaries. most of the rejected applicants report substantial health limitations. bound concludes that his direct evidence indicates that ssdi benefits do not generate disincentives to work. several research efforts provide information on the attitudes and behavior of different demographic cohorts vis-a-vis ssdi benefits. greenblum and bye (gb) [ 19871 analyze survey data from the social security administration and find that belief in the importance of work does not decline after individuals begin receiving ssdi benefits. gb conclude that vocational rehabilitation and incentive programs promoting a return to work are very important. mudrick (1988) uses the same data to explore the relationship between the demographic characteristics, attitudes, and social roles of individuals with their disability status. logit regression analysis indicates that pain and fatigue are strongly related to disability. the linkage of self-esteem and work is inversely related to disability claims. molho (1989) examines the relationship between demographic factors and the british equivalent of ssdi claims. using both ols and logit regression, he finds that the non-economic factors of age and medical history are the primary predictors of claims. traditional economic factors, including regional economic status and housing environment, have only a secondary influence. an overview of the three previously cited studies supports the preference of most disabled individuals to return to work and, hence, the necessity of rehabilitation efforts that may remove many individuals from ssdi rolls. workers’ compensation. all workers’ compensation plans cover occupationally related ltd claims and most states require benefits to be paid for the entire life of a permanently disabled worker. some states impose limits on benefit duration, however, with west virginia mandating the shortest limit of 208 weeks.’ a substantial majority of states place no cap on the total amount of disability income benefits paid to a worker, but a significant minority do impose such limits, ranging from $45,000 in maryland to $198,590 in new mexico during 1989. seventeen states require that any social security disability benefits and/ or unemployment benefits received by the disabled earner be used as offsets 70 financial services review, l(1) 1991 against workers’ compensation disability income benefits (u.s. chamber of commerce, 1989, pp. 18-20). federal law generally requires social security benefits to be supplemental to state workers’compensation benefits, but excepts these 17 states, which passed “reverse” offset laws prior to february 1981. devol(l986) investigates the extent to which workers’ compensation plans can replace the pre-disability income of the average worker in a high-benefit state (massachusetts) and a low-benefit state (georgia). she constructs expected value models that incorporate such factors as workers’ compensation claims experience, wage inflation and life cycles, employee benefit losses, and typical worklife experience. her results indicate that, on an after-tax basis, workers’ compensation benefits replace over 85 percent of the worker’s income for disabilities lasting one year. for longer-term disabilities, income replacement ratios drop to as low as 40 percent of after-tax income. if the state of residence does not use a social security offset, the income replacement ratios are raised significantly. workers’ compensation plans pay both std and ltd benefits. nelson (1988b) reports that approximately 75 percent of all workers’ compensation claims are for std, yet the 25 percent pertaining to ltd claims account for 75 percent of the benefits paid by these plans. despite the large proportion of benefits paid by workers’ compensation plans for ltd claims, this source may represent a relatively small factor for disabled individuals in the u.s. packard (1987) reports that, among beneficiaries of social security disability payments, less than 3 percent of total income is supplied by workers’ compensation plans. these data do underestimate somewhat the impact of workers’ compensation because of the social security benefit offsets implemented in 17 states, however. hansen, macavoy, and smith (hms) (1989) provide a broad view of programs, including workers’ compensation, that deliver benefits to earners suffering from occupationally related disease or disability. hms categorize the various types of programs and supply a general analysis of how major attributes of each program affect claims, administrative costs, compensation errors, and worker safety from the perspective of the public policymaker. while public ltd benefit programs are demonstrably important to the average earner in the u.s., private sources may be of greater importance to upper-middle and upper class earners. the researcher must be aware of how public programs may affect both private markets and individual decisions, however. table 1 summarizes some major provisions of public programs applicable for u.s. citizens employed in the states of new york and texas. as shown in the table, individuals employed in these two diverse states are subject to very different “safety nets” of publicly funded disability protection that apply even before private plans are considered. in the next section, private ltd plans are examined. 71 public disability protection plans available to individuals on january 1, 1989 plan state provisions workers compensation new york texas temporary new york texas social security all 2/ 3 x compensation weekly maximum: $300 maximum duration: life elimination period: 7 days social security offset 213 x compensation weekly maximum: $238 maximum duration: 401 weeks elimination period: i days no social security offset l/ 2 x compensation weekly maximum: $170 maximum duration: 26 weeks elimination period: 7 days covers non-occupational disability only no plan benefits based on compensation and family size monthly maximum: $ t ,090/ individual $1$35/family maximum duration: age 65 l elimination period: 5 months l workers compensation offset sources: u.s. chamber of commerce (i989), hay/ huggins (1990). private sources. the preponderance of ltd insurance contracts in the u.s. are provided to individuals through employer group plans. data from the 1988 update: source book of wealth insurance data coupled with general population data indicates that approximately 15 percent of the working population is covered by private group ltd plans, while about 7 percent own private individual policies. the u.s. bureau of labor statistics (bls) (1989) provides detailed information on employer group ltd coverage for employees of large and medium size firms. approximately 42 percent of these employees are covered by group ltd plans. white-collar workers are more than twice as likely to receive coverage as their blue-collar counterparts, although 38 percent of the latter are eligible to receive immediate disability pension benefits. typical benefit levels for group ltd plans are set at 50 to 60 percent of pre-disability earnings, subject to a monthly maximum that ranges between $2,500 and $5,000. for 63 percent of group plan participants, offsets for ssdi 72 financial services review, l(1) 1991 and workers’ compensation benefits apply when total disability benefits exceed 70 to 75 percent of pre-disability earnings. an additional 7 percent are subject to even lower total benefit ceilings. the normal waiting period for group ltd plans is six months. blostin, burke and lovejoy (bbl) (1988) compare a prior version of the bls study with an analysis of group benefits available to employees of state and local governments. for the years studied, bbl find that only 31 percent of the government employees received group ltd coverage as opposed to 48 percent of corporate employees. government workers were twice as likely as corporate workers (80 percent versus 40 percent) to have access to immediate disability benefits from their retirement plans, however. virtually all u.s. earners have the option of buying individual ltd contracts from private insurers, although relatively few do so. edmonston and scott (es) (1987) supply an insightful consumer survey of individuals who have purchased individual ltd policies. the response data encompass buyers of 6,489 noncancellable policies from 27 insurers. demographic analysis by es reveals that males account for 78 percent of the ltd policies purchased and 84 percent of premiums paid. the median age of buyers is 35 for males and 36 for females. purchasers’ incomes are approximately four times that of the average u.s. citizen, with median incomes of $54,700 for males and $28,600 for females. professionals and executives are the dominant purchasers of individual ltd, generating 86 percent of total premiums derived from male buyers and 76 percent from female buyers. in the next section, information about private di insurance products and suppliers that is necessary for the individual to make a rational decision is discussed. the product and supplier decisions product selection much of the literature on di insurance products represents primarily a description of these relatively complex contracts, e.g., morris (1986) and soule (1984). in a similar fashion, lyons (1987) describes the income tax treatments of premiums paid and benefits received for various types of disability plans that apply to individual employees and small business owners. the edmonston and scott (es) (1987) study supplies information about the contractual options of ltd policies typically selected by individuals. among their findings, es observe that waiting periods vary widely, but a plurality of policies contain 60 to 90 day waiting periods. benefit durations are typically quite long-term, with two thirds providing benefits ranging from 10 years to payments until age 65. over half of the purchasers select an “own occupation” definition, which means that benefits will be received if a disabled policyholder disability income insurance and the individual 73 cannot perform at least one significant duty of his or her current occupation. more than 50 percent of the policies provide proportional benefits if the policyholder can perform his or her normal duties on a part-time, but not full time, basis. cox and gustavson (cg) (1990) explore individual ltd insurance prices for a cross-sectional sample of 54 insurers in 1988. cg show preliminary evidence of price dispersion exceeding that for other types of insurance and consumer products. regression analysis indicates that the elimination period is significantly related to price for all the types of policies tested. other variables having a significant impact on price for some types of ltd policies include the definition of disability, the residual (or partial) disability provision, and the organizational form of the insurer. although surveys of di insurance attributes and prices are available to the public (blease and pallay, 1990), the present literature offers little to guide the individual in making an optimal purchase decision. for instance, while a number of studies investigate comparative price indices for life insurance, e.g., schleef (1989), no such information has been developed for the relatively more complex di insurance products. effort also should be devoted to the optimal selection of contractual options. the individual should be fully informed of trade-offs between the various contractual options, and between these options and financial assets or alternative risk management techniques. for example, the elimination period is itself an alternative risk management technique (risk retention). the choice of elimination period should not be made in isolation, but only after considering (1) other risk management methods, (2) other contractual options, such as benefit duration or the residual disability provision, and (3) available liquid assets. virtually no modelling has been provided to help the individual make the optimal decision, however. next, the selection of the product supplier is considered. insurer selection a healthy stream of literature focuses upon the financial solvency of insurers, e.g., ambrose and seward (1988) and barniv and hershbarger (1990). studies of this genre attempt to identify econometric techniques and variables that will provide the most reliable predictions of insolvency. researchers typically test their proposed models against the insurance regulatory information system (iris) promulgated by the national association of insurance commissioners (naic). with the iris system, the naic establishes “acceptable” ranges for 12 financial ratios. life and health insurers with four or more ratios outside the acceptable range are targeted for special attention by the naic. the results of the iris tests can be replicated by researchers and are available to individuals and their advisors (belth, 1990a). as shown 14 financial services review, l(1) 1991 by barniv and hershbarger, however, the predictive capacity of the iris system is limited and multiple classification techniques may be advisable. the standard source of information regarding both the financial and operational status of insurers has been best s life and health reports published by the a.m. best company. within the last few years competing firms, most notably standard and poor’s, moody’s, and duff and phelps, have begun evaluating and rating insurers. belth (1990b) documents substantial inconsistencies between best’s ratings and those of the competition, citing increasing liberalization of best guidelines over time as a primary cause. while the individual now has access to more information on insurers via the rating services, an additional source of noise has been introduced and additional research is likely on the impact of this development. in addition to financial risk, the individual should assess the underwriting risk and efficiency of insurers, if possible. little guidance based upon research can be offered, however. one reason may be the obfuscating statutory accounting practices of insurers. for instance, although overall ratios for underwriting losses to premiums and underwriting expenses to premiums are available for total accident and health insurance lines written by the life and health insurer, information specific to disability lines only is not (zucconi, 1987, pp. 17618 1, 226-227; doligalski, 1990). past records of claims service and rehabilitation intervention also are likely to vary widely across insurers. for instance, one insurer recently revealed a new group plan with the announced goal of early intervention and rehabilitation resulting in “a quick return to work, accompanied by a smooth flow of financial support” (koco, 1990). unless further disclosure is provided by insurers or regulators, the individual (or employer) will have no way of evaluating such marketing claims. for the most part, research into di insurance has been descriptive in nature and focused upon social programs and public policy. the many empirical studies addressing ssdi and workers’compensation benefits indicate that these programs are likely to be of limited value to middle and upper class individuals. the research pertaining to private di insurance markets also has been largely descriptive, but does provide some insight into the buying behavior of different demographic groups and the pricing of individual products. the available evidence only raises more questions about the limited acceptance of individual di insurance plans by the buying public and the extent to which essential information is available to individuals. concluding remarks are provided in the following section. conclusions the preponderance of research efforts pertaining to disability income protection focuses upon social insurance and public policy. much of this ohability income insurance and the zndividual 75 research has been conducted or sponsored by u.s. government agencies, which may explain the dominance of a public policy perspective. as with many other areas of individual financial management, optimal risk management practices applicable to the disability loss exposure have not been adequately modelled. potential research topics may be found in the categories of disability insurance markets, individual planning, and insurance portfolio formation. a number of questions regarding disability insurance markets need further investigation. for example, why is disability insurance so widely recognized as a fundamental need by financial planners and authors, yet relatively little insurance is purchased, especially to meet the ltd exposure? are individuals really exhibiting behavior that is close to optimal? are disability insurance markets efficient in terms of information dissemination? if not, how can informational constraints best be relaxed or removed? can comparative cost indices be constructed that will be adequate and comprehensible for implementation by consumers? many individual planning questions may interest the prospective researcher. how can disability income needs be better modelled in financial planning programs? can “additional” disability-related costs, such as nursing care or child care, be reasonably predicted and included in individual needs models? can optimal packages of contractual options be modelled? what trade offs may the individual reasonably consider between di insurance, alternative risk management techniques, liquid financial assets, and disability pension benefits? supplier attributes also deserve closer examination. how can the individual best assess the underwriting leverage and efficiency of prospective suppliers? how can claims service and rehabilitation services be evaluated? once the preceding issues are investigated, the scope of disability income research can be expanded. for instance, can optimal portfolios containing disability, medical expense, life, and other types of insurance be developed using models comparable to those described in modern portfolio theory? can optimal insurance and investment portfolios be reconciled and solutions devised that can be reasonably implemented by individuals? can integrated packages of disability income insurance be produced to replace the current agglomeration of programs available from public and private suppliers? the paucity of rigorous research in di insurance provides a unique opportunity to the innovative researcher. major objectives of this review have been (1) to provide the prospective researcher with a foundation for understanding essential disability insurance issues and the underlying market structure and (2) to stimulate interest in some of the previously discussed research topics related to disability income insurance, an area largely ignored in the academic research literature of economics, finance, and insurance. 76 financial services review, l(1) 1991 4. 5. 6. notes for examples of models capitalizing survivor needs caused by the death of an earner, see belth (1964), rose and mehr (1980), and gustavson (1982). for example, goldsmith [ 19831 offers preliminary empirical evidence that individuals may substitute a spouse’s human capital for the purchase of life insurance. the law determining primacy between oasdi disability benefits and state-mandated workers compensation plans is explained in the subsequent section on long-term workers compensation benefits. official social security administration [ 1990, p. 2601 acceptance rates are higher than those estimated by lfb, rising from a low of 26.5 percent in 1982 to the 33-34 percent range in the mid-1980s. acceptances spiked upward to 44 percent in 1988, a year in which appiications fell by 23 percent. in response to recent budgetary problems, the administration has used such strategies as simply suspending all ssdi hearings [pear, 19901 and relying upon a single, unreliable diagnostic test to determine whether applicants should receive benefits [lambert, 19901. for a by-state synopsis of workers compensation requirements for benefits paid, see u.s. chamber of commerce [ 19891. for a discussion of how typical offsets are structured in group ltd plans, see hill [ 1987, pp. 18-191. references ambrose, jan m., and j. allen seward. 1988. “best’s rating, financial ratios and prior probabilities in insolvency prediction,” journal of risk and insurance, s(2): 229-244. barniv, ran, and robert a. hershbarger. 1990. “classifying financial distress in the life insurance industry,” journal of risk and insurance, 57( 1): 110-136. belth, joseph m. 1964. “dynamic life insurance programming,” journal of risk and insurance, 31(4): 539-556. belth, joseph m. 1990. “a watch list of insurance companies based on the naic’s 1990 iris ratios,” the insurance forum, 17(9): 143-151. belth, joseph m. 1990. “the dilemma of the a.m. best company,” 71re rnsurance forum, 17( 12): 177-182. blease, roger l., and gary s. pallay. 1990. “noncance~able disability income survey,” best*s review [life/health edition], 90( 10): 101-105. blostin, allan p., thomas p. burke, and lora m. lovejoy. 1988. “disability and insurance plans in the public and private sectors,” monthiy labor review, 11 1( 12): 9-17. bound, john. 1989. “the health and earnings of rejected disability insurance applicants,” american economic review, 79(3): 482-503. bureau of labor statistics. 1989. employee benefits in medium and large firms, 1988, bulletin 2336. washington, dc: u.s. government printing office. cox, larry a., and sandra g. gustavson. 1990. “impact of disability insurance options for individuals,” paper presented at the annual meeting of the academy of financial services, october 24, 1990. cox, larry a., sandra g. gustavson, and antonie stam. 1991. “disability and life insurance in the indi~dual insurance portfolio,” journai of risk and insurance, 5x( 1): 128-l 37. devol, karen r. 1986. income replacement for long-term disability: the role of workers’ compensation and ssdi. washington, dc: workers compensation research institute. d&ability income insurance and the individual 77 doligalski, david w. (ed.) 1990. national underwriter profiles: 1990 health insurers. cincinnati, oh: the national underwriter company. dykacz, janice m., and john c. hennessey. 1989. “postrecovery experience of disabled worker beneficiaries,” social security bulletin, 52(9): 42-66. edmonston, b., and p.s. scott. 1987. 7’he 1986 disability income buyer. hartford, ct: life insurance marketing and research association. goldsmith, art. 1983. “household life cycle protection: human capital vs. life insurance,” journal of risk and insurance, 50(3): 473486. grad, susan. 1989. “income and assets of social security beneficiaries by type of benefit,” social security bulletin, 52(l): 2-10. greenblum, joseph, and barry bye. 1987. “work values of disabled beneficiaries,” social security bulletin, 50(4): 67-74. gustavson, sandra g. 1982. “flexible income programming: comment,” journal of risk and insurance, 49(2): 290-296. halpern, janice, and jerry a. hausman. 1986. “choice under uncertainty: model of applications for the social security disability insurance program,” journal of public economics, 31: 131-161. hansen, ronald w., paul w. macavoy, and clifford w. smith, jr. 1989. “compensation alternatives for occupational disease and disability,” 77re journal of risk and insurance, 55(2): 252-274. hay/ huggins reference guide. 1990. philadelphia: hay/ huggins company. hennessey, john c., and janice m. dykacz. 1989. “projected outcomes and length of time in disability insurance program,” social security bulletin, 52(9): 2-41. hill, diane b. 1987. “employer-sponsored long-term disability insurance,” monthly labor review, 110(7): 16-22. houff, james n., and william j. wiatrowski. 1989. “analyzing short-term disability benefits,” monthly labor review, 112(6): 3-9. kerns, wilmer l. 1989. “social security programs in the united states: temporary disability insurance or cash sickness insurance,” social security bulletin, 52(7): 37-41. koco, linda. 1990. “cigna managed di ‘unites’ short and long-term covers,” national underwriter, 94(22): 5, 10. lambert, wade. 1990. “federal agency improperly denied claims,” wall street journal, june 28, 1990, p. a16. lando, mordechai e., alice v. farley, and mary a. brown. 1982. “recent trends in the social security disability insurance program,” social security bulletin, 45(8): 3-14. lyons, james r. 1987. “impact of taxes on disability income insurance plans,” journal ofthe american society of clu and chfc, 41( 1): 56-63. mehr, robert i., and sandra g. gustavson. 1987. lzfe insurance: theory and practice, fourth edition. plano, tx: business publications. miller, john h. 1978. “disability insurance, an assessment of its social value,” clu journal, 32(4): 12-24. molho, ian. 1989. “a disaggregate model of flows onto invalidity benefit,“appzied economics, 21: 237-250. morris, jr, william h. 1986. disability insurance. lexington, ky: lexington house. mudrick, nancy r. 1988. “predictors of disability among midlife men and women: differences by severity of impairment,” journal of community health, 13(2): 70-84. nelson, jr., william j. 1988a. “workers’ compensation: coverage, benefits, and costs, 1985,” social security bulletin, 5 i( 1): 4-9. nelson, jr., william j. 1988b. “workers’ compensation: 1980-84 benchmark revisions,” social security bulletin, 51(7): 4-21. financial services review, l(1) 1991 nelson, jr., william j. 1989. “social security programs in the united states: workers’ compensation,” social security bulletin, 52(7): 27-36. packard, michael d. 1987. “income of new disabled-worker beneficiaries and their families: findings from the new beneficiary survey,” social security bulletin, 50(3): s-23. pear, robert. 1990. “u.s. is suspending hearings on new disability payments,” new york times, august 1, 1990, pp. al, b6. price, daniel n. 1986. “cash benefits for short-term sickness: thirty-five years of data, 1948 83,” social security bulletin, 49(5): 5-19. reno, virginia p., and daniel n. price. 1985. “relationship between the retirement, disability, and unemployment insurance programs: the u.s. experience,” social security bulletin, 48(5): 24-37. rejda, george e. 1984. social insurance and economic security, 2nd edition. englewood cliffs, nj: prentice-hall. rose, terry, and robert i. mehr. 1980. “flexible income programming,” journal of risk and insurance, 48( 1): 44-60. schwartz, david, and herman grundmann. 1989. “social security programs in the united states: old-age, survivors, and disability insurance,” social security bulletin, 52(7): 6-19. schleef, harold. 1989. “whole life cost comparisons based upon the year of required protection,” journal of risk and insurance, 56(l): 83-103. social security administration. 1990. social security bulletin: annual statistical supplement, 1989. washington, dc: government printing office, 1990. soule, charles e. 1984. disability income insurance: the unique risk. homewood, il: dow jones-irwin. stemnock, suzanne k. (ed.) 1988. 1988 life insurance fact book. washington, dc: american council of life insurance. u.s. chamber of commerce. 1989. i989 analysis of workers compensation laws. washington, dc: u.s. chamber of commerce. zucconi, paul j. 1987. accounting in lfe and health insurance companies. atlanta, ga: life office management association. the decrease in life insurance ownership: implications for financial planning kyoung tae kima, travis p. mountainb, sherman d. hannac,*, namhoon kimd adepartment of consumer sciences, 312 adams hall, box 870158, university of alabama, tuscaloosa, al 35487, usa bdepartment of agricultural and applied economics, virginia tech university, 250 drillfield drive, blacksburg, va 24061, usa cdepartment of human sciences, ohio state university, 1787 neil avenue, columbus, oh 43210, usa dkorea rural economic institute, 601 bitgaram-ro, naju-si. jeollanam-do, 58321, south korea abstract based on our analyses of survey of consumer finances datasets, the proportion of households owning a life insurance policy decreased from 72% in 1992 to 60% in 2016. we estimated logistic regressions on the likelihood of ownership of any, term, and cash value life insurance. we conclude that changes in household characteristics accounted for the decrease in term life insurance ownership, but not for the decreases in any and in cash value life insurance ownership. we also found a positive association between use of a financial planner and life insurance ownership. we discuss implications for financial planning. © 2020 academy of financial services. all rights reserved. jel classification: d12; d14; g22 keywords: life insurance; declining insurance demand; survey of consumer finances 1. introduction life insurance is an important component of risk management and insurance planning, and is also important in other financial planning topics, including employee benefits, income tax planning, and estate planning. “as late as 1960, life insurance was the * corresponding author. tel.: �1-614-292-4584; fax: �1-614-292-4339. e-mail address: hanna.1@osu.edu (s.d. hanna) financial services review 28 (2020) 1-16 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. substance of financial planning. . . ” (brandon and welch, 2009, p. 2). however, the importance of life insurance seems to be decreasing in terms of ownership rates, and the relative number of life insurance agents (scism, 2016). also, the certified financial planner board (2015) list of principal knowledge topics weight for risk management and insurance for the cfp exam, has decreased from 14% (hanna, et al., 2011) to 12% (certified financial planner board, 2015). a wall street journal article (scism, 2016) suggested that because of consolidation and decreasing ownership rates, the life insurance agent may be going the way of the dinosaur. cordell, finke, and lemoine (2007) noted the long-term trend of decreasing ownership of cash value life insurance between 1995 and 2004, and retzloff (2010) reported that life insurance ownership in the united states was at a 50 year low. however, life insurance is still a vital safety net for u.s. households. the premature death of a wage-earner is one of the more serious financial risks a household faces. income replacement and burial expenses are the top two reasons households own life insurance (durham, 2015). the main purpose of our research was to ascertain factors related to the decline of life insurance ownership rates of u.s. households. we analyzed a combination of 1992 to 2016 survey of consumer finances (scf) datasets, and found that ownership of life insurance decreased until 2013 and then stayed about the same in 2016. we conducted logistic regression analyses of the pooled dataset to analyze the extent to which changes in the composition of u.s. households might have contributed to the decreases in any, term, and cash value life insurance. further, we conducted additional analyses to investigate the role of financial planner use on life insurance ownership. our study contributes to some implications for financial planners by providing insights into life insurance ownership trends. 2. literature review 2.1. overview of life insurance without life insurance, most families would need to reduce their current standard of living in the event of the death of a spouse or partner (auerbach and kotlikoff, 1991; bernheim, carman, gokhale, and kotlikoff, 2003; bernheim, forni, gokhale, and kotlikoff, 2003). cash value and term life insurance are the two main categories of life insurance. in addition to the insurance aspect of a cash value life policy, it also provides a savings component and is intended to last the insured’s “whole” life. term policies are pure insurance that are associated with a set number of years ranging from one to 40 with 20 years being a common term contract length. if the insured dies during this contract period, the beneficiary receives the face amount of the policy. if one dies after the contract expires, no benefits are paid to the beneficiary. employer-provided group life insurance is one of the most common types of employer benefits, and in 2017, 60% of employers provided it, with 73% of eligible employees participating in group life insurance (greenwald and fronstin, 2019). 2 k. tae kim et al. / financial services review 28 (2020) 1-16 2.2. previous studies on life insurance ownership mulholland, finke, and huston (2016) examined the declining trend in cash value life insurance ownership rates using the 1992 to 2010 waves of the scf. the proportion of cash value life policies relative to term life policies dramatically decreased over this period. they found a substitution effect, as those who owned term life insurance were less likely to own cash value life insurance. however, frees and sun (2010) found a complementary effect, as those who owned term life insurance were more likely to own cash value life insurance. mulholland et al. (2016) and glazer (2007) noted that in addition to the income protection features of term life insurance, cash value life insurance policies have been marketed as tax-advantaged investments. however, with the introduction of more tax-advantaged savings instruments such as roth iras, along with the introduction of tax-advantaged savings plans for college costs, many households now have attractive alternatives to cash value life insurance for important financial goals. increases in federal income tax marginal rates could potentially decrease the demand for cash value life insurance. mulholland et al. (2016) noted that cash value life insurance has had an important role in estate planning tools, especially when more households were potentially subject to the federal estate tax. the increases in the exemption amounts for the federal estate tax have generally reduced the number of households potentially subject to the tax after 2004. mulholland et al. (2016) also noted that the cost of term life insurance has dropped substantially since the introduction of internet marketing and price comparisons, but the cost of cash value life insurance has not dropped as much. heo, grable, and chatterjee (2013) examined the 2004 and 2008 national longitudinal survey of youth 1979 cohort that consisted of respondents aged 43–51. heo et al. (2013) noted that there was a net decrease of one percentage in life insurance ownership over this period. guillemette, hussein, phillips, and martin (2015) analyzed the 1992 through 2010 scf datasets, finding a decreasing ownership trend and examined if household size affected life insurance ownership differently for minority households. compared with white households, as household size increases, the likelihood of life insurance ownership decreases for both black and hispanic households. additionally, they found that being employed, age, education, presence of children, homeownership, and being married to be positively associated with having life insurance. relative to 1992, all subsequent survey years were significant and negative with the exception of 1995 that was not significantly different than 1992. they also found that self-employed households were less likely to own life insurance compared with those working for an employer, and hispanic households were less likely to own life insurance than otherwise comparable white households, while black households were more likely to own life insurance than white households. gutter and hatcher (2008) examined the demand for life insurance using the 2004 scf, with an objective of determining factors related to white and black household life insurance ownership differences. gutter and hatcher (2008) found age to be positively correlated to life insurance ownership. homeownership was positively related to life insurance ownership while the likelihood of life insurance ownership for those with children was not different from those without children. mountain (2015), using the 2013 scf, examined life insurance ownership for working coupled households aged 30–64. mountain (2015) estimated that the 3k. tae kim et al. / financial services review 28 (2020) 1-16 median proportion of insurable human wealth insured with life insurance was only 28%. age and education were positively associated with life insurance ownership. households with a black respondent were more likely, and those with hispanic and asian/other respondents were less likely, to own life insurance than households with a white respondent. married couples were more likely to own life insurance than partnered couples. liebenberg, carson, and dumm (2012) explored the determinants of life insurance demand for both term and whole life policies in a dynamic analysis where the authors used 1983–1989 scf panel datasets in a cragg model. they found that new parents were more likely to own life insurance than those who were not new parents, and households who had started a new job were more likely to increase their holdings of term insurance coverage than those who had not. bernheim, carman et al. (2003) examined financial vulnerability and life insurance coverage at all stages of the life cycle. the authors proposed two explanations on why a household with financial vulnerabilities might not be protected by life insurance: (1) young households purchased long term life insurance contracts but failed to update them as life events change; (2) actual needs had no effect on purchase. finke, huston, and waller (2009) used data from the 2004 scf and compared the difference of life insurance purchases between people who were advised by either financial planners or brokers (dealers) to those who were not. they found that households that used financial planners and that used broker/dealers shared similar demographic characteristics. more important, the estimation results showed that people who relied primarily on financial planners were more likely to purchase adequate life insurance holdings, while the use of broker/dealers had no impact on levels of life insurance. finally, households who were wealthier, self-employed, and home-owning tended to purchase more adequate life insurance holdings. baek and devaney (2005) developed a model of term and whole life insurance ownership and amount of coverage, and tested the effects of human capital, bequest motives, and risk variables, as well as other household characteristics. they found that ownership of term life insurance decreased with age, was higher for middle income households than for low income households, and higher for homeowners than for renters, but education was not significantly related to term life insurance ownership. they also found that ownership of cash value life insurance was highest for those over the age of 65, and there was also a positive relationship with age controlling for other characteristics including life expectancy. cash value life insurance ownership was also negatively related to having excellent health, positively related to the level of liquid assets, and those in the highest income tax bracket were more likely to own than those in the lowest bracket. ownership was not significantly related to attitudes about leaving a bequest. using finke et al. (2009) as framework, scott and gilliam (2014) focused on baby boomer life insurance adequacy before and after the 2008 financial crisis. the authors used 2004 and 2010 scf datasets to represent before and after the 2008 financial crisis. they found that the use of a financial planner was higher among household with adequate life insurance. in addition, having a financial planner acted as a positive predictor of life insurance adequacy in 2004, while it had no impact on life insurance adequacy in 2010, after the 2008 financial crisis. mountain (2015) found that households who consulted a financial planner or broker were more likely to own life insurance and also, given ownership, to have higher face value 4 k. tae kim et al. / financial services review 28 (2020) 1-16 amounts, however, he did not find the financial planner or broker variable to be significant when exploring the proportion of insurable human wealth that was insured by life insurance. 2.3. overview of past research on life insurance we reviewed selected research on factors related to life insurance ownership and adequacy. while previous researchers have found a number of household characteristics to be related to life insurance ownership, none of the studies have had a focus on identification of factors related to changes in life insurance ownership over many years. we estimated models of life insurance ownership with a combined dataset of 1992 through 2016 scf datasets; thus, using more recent data than previous research, some of which included datasets only through 2010. by estimating a time trend controlling for household characteristics, we investigated the question of whether the decrease in life insurance ownership has been primarily because of changes in household characteristics, or to some other factors, such as the increasing availability of alternative tax advantaged investment options such as 401k accounts (mulholland et al., 2016). 3. methods 3.1. data and sample selection we used a pooled dataset from nine waves of the survey of consumer finances (scf), from 1992 to 2016. the federal reserve board (frb) has released the scf cross-sectional dataset triennially since 1983. hanna, kim, and lindamood (2018) provide detailed information about using scf datasets. for more information about specific variables, see board of governors of the federal reserve system (2014, 2017). the total sample size of the pooled dataset is 44,634. there were 883 cases where life insurance ownership was missing (shadow variables with values over 90, see hanna et al., 2018), and those cases were excluded from our analyses. the final analytic sample was 43,751. 3.2. measurement of variables 3.2.1. dependent variable the scf has life insurance ownership variables for both term life insurance and cash value life insurance. for our empirical analysis as well as the life insurance ownership trend, we use three separate binary dependent variables: whether the respondent or any family member has, term, cash value, or any life insurance policy (i.e., term or cash value). the actual questions are presented in appendix 1. 3.2.2. independent variables following the existing literature on life insurance ownership, the set of independent variables included survey year, age of the household head, age squared, marital status (married, single male, single female, and partnered), education of the household head (less 5k. tae kim et al. / financial services review 28 (2020) 1-16 than high school, high school, some college, bachelor degree, and postbachelor degree), race/ethnicity (white, black, hispanic, and asian/others), presence of a child under 18, current income compared with normal income (low, normal, and high), log of household income, home ownership, employment status of the household head (salary worker, selfemployment, retired, and not working) and health status of a household head (excellent, good, fair, and poor). in the scf, race/ethnicity is of the respondent (lindamood, hanna, and bi, 2007), but for convenience, we refer to the race/ethnicity of the household, for example, white households. the scf has had a question about whether a household reported using a financial planner for information on (1) saving and investment decisions and/or (2) borrowing and credit decisions since the 1998 scf. we used an indicator for comprehensive financial planner use, defined as a household using a financial planner for saving/investment and/or borrowing/ credit issues (elmerick, montalto, and fox, 2002). 3.2.3. analysis given the binary dependent variables of life insurance ownership, logistic regression models are utilized to analyze factors related to the ownership of any life insurance, term life insurance, and cash value life insurance. because of the nature of the data and our analysis, we are not able to attribute specific causal relationships for life insurance ownership. for descriptive purposes, means tests are also employed. we used the repeated-imputation inference (rii) method with all of the five implicates in each scf dataset, which provides an estimate of variances more closely representing the true variances than estimates obtained by only one implicate (hanna et al., 2018; lindamood et al., 2007). in order to provide conservative hypothesis tests (shin and hanna, 2017), we did not weight the multivariate analyses. 4. results 4.1. descriptive results fig. 1 displays the historical downward trend in life insurance ownership from the combined 1992–2016 scf datasets. the proportion of households owning some type of life insurance peaked at 72% in 1992 and reached a low of 60% in 2013, with the 2016 rate about the same as 2013. the proportion of households owning a term life insurance policy had a generally downward trend, starting at 54% in 1992, and decreasing to 48% by 2013, and the rate remained about the same in 2016. the proportion of households owning a cash value life insurance policy decreased from 36% in 1992 to 20% in 2013, and the 2016 rate was about the same as 2013. table 1 presents the logistic regression results of the association between survey year and the likelihood of three different types of life insurance ownership, not controlling for other household characteristics. there were significant negative relationships between the survey year and the likelihood of ownership of any life insurance, term life insurance, and cash value life insurance. 6 k. tae kim et al. / financial services review 28 (2020) 1-16 4.2. rii means test the characteristics of our analytic sample are presented in table 2. most respondents were white, homeowners, did not have a child under 18, had current income about the same as normal, salary workers, completed education beyond the high school level and had good or excellent health status. results of means tests are also included to show descriptive patterns of any, term, and cash value life insurance ownership by selected household characteristics. in the discussion below, we report patterns for ownership of any life insurance. the life insurance ownership rate was 46% for households with a respondent under 30, 71–72% for those with a respondent 40 to 49 and 50 to 59, and 63% for those with a respondent 70 and over. white households had a higher rate of life insurance ownership (69%) than each of the other racial/ethnic groups, and hispanics had the lowest rate, 37%. homeowners had a higher ownership (75%) than renters (46%). the proportion of life insurance ownership was higher for married couples (77%) than for other groups. households with a dependent child under age 18 had a higher life insurance ownership rate (68%) than those without a child (64%). households with unusually high fig. 1. trend in life insurance ownership rates, any type, term, and cash value, 1992–2016 scf weighted results. n � 43,751. table 1 logistic regressions on ownership of any, term, and cash value life insurance by survey year only, 1992–2016 scf any life insurance term life insurance cash value life insurance coefficient standard error coefficient standard error coefficient standard error survey year �0.0270*** 0.0030 �0.0072*** 0.0028 �0.0394*** 0.0030 intercept 54.9622 6.0089 14.3913 5.5264 78.1899 6.0701 mean concordance 50.1% 45.8% 53.1% note: *p � 0.05, **p � 0.01, ***p � 0.001. unweighted analysis with rii technique. n � 43,751. 7k. tae kim et al. / financial services review 28 (2020) 1-16 current income were more likely to own life insurance (71%) than those with lower current income (54%) and those with current income about the same as normal (67%). households with the head working for a salary had a higher rate of life insurance ownership (71%) than those in other types of employment status, and households with the head not working had only a 34% life insurance ownership rate. the proportion of life insurance ownership was highest for households with a head having a postbachelor degree (77%), compared with 75% for households with a bachelor degree, 67% for households with some college, 63% for those with a high school degree, and only 46% for those less than high school degree. households with a head with excellent health had the highest proportion of life insurance, 70%, while those with poor health had the lowest proportion, 50%. table 2 proportion of ownership of any, term, and cash value life, by household characteristics, 1992–2016 scf variable category distribution (%) any life insurance term life insurance cash value life insurance rate of insurance ownership (%) significance level rate of insurance ownership (%) significance level rate of insurance ownership (%) significance level age of head less than 30 12.8 45.5 reference 37.7 reference 12.7 reference between 30 and 39 19.0 63.7 �0.001 54.4 �0.001 20.0 �0.001 between 40 and 49 20.5 71.1 �0.001 60.1 �0.001 23.6 �0.001 between 50 and 59 18.0 71.5 �0.001 58.4 �0.001 28.0 �0.001 between 60 and 69 14.0 70.2 �0.001 49.6 �0.001 33.5 �0.001 70 and over 15.7 62.9 �0.001 37.3 �0.001 32.7 �0.001 race/ethnicity white 73.2 68.7 reference 53.4 reference 27.6 reference black 13.5 66.5 �0.001 50.5 �0.001 24.7 �0.001 hispanic 9.2 36.9 �0.001 31.8 �0.001 8.6 �0.001 asian/other 4.1 59.0 �0.001 47.8 �0.001 18.8 �0.001 homeowner no 33.7 45.5 reference 36.6 reference 13.7 reference yes 66.3 75.0 �0.001 58.0 �0.001 30.9 �0.001 marital status married 50.3 77.1 reference 61.4 reference 31.6 reference single male 14.9 51.7 �0.001 38.9 �0.001 18.2 �0.001 single female 27.3 53.7 �0.001 39.9 �0.001 19.5 �0.001 partner 7.5 52.2 �0.001 43.2 �0.001 16.2 �0.001 the presence of child under 18 no 66.2 63.8 reference 47.4 reference 26.1 reference yes 33.8 67.6 �0.001 57.5 �0.001 23.3 �0.001 current income relative to normal normal 73.1 67.1 reference 52.3 reference 26.1 reference income higher 8.5 71.4 �0.001 56.9 �0.001 28.4 �0.001 income lower 18.4 54.0 �0.001 42.2 �0.001 19.8 �0.001 employment status of head salary worker 58.0 70.7 reference 59.4 reference 23.6 reference self-employment 10.9 65.4 �0.001 48.8 �0.001 31.5 �0.001 not working 5.7 33.8 �0.001 25.5 �0.001 12.5 �0.001 retired 25.3 59.0 �0.001 37.6 �0.001 28.9 �0.001 education of household head less than high school 14.6 45.6 reference 33.5 reference 16.3 reference high school degree 30.7 63.1 �0.001 47.7 �0.001 24.6 �0.001 some college 23.0 66.6 �0.001 52.2 �0.001 25.4 �0.001 bachelor degree 18.6 74.7 �0.001 61.0 �0.001 28.7 �0.001 post-bachelor degree 11.5 76.7 �0.001 61.6 �0.001 32.2 �0.001 health status of head excellent health 27.1 70.2 reference 56.6 reference 26.8 reference good health 47.3 66.9 �0.001 52.9 �0.001 25.6 �0.001 fair health 19.4 58.4 �0.001 42.8 �0.001 23.3 �0.001 poor health 6.2 49.7 �0.001 34.9 �0.001 20.1 �0.001 note: author analyses of combined 1992–2016 scf datasets, n � 43,751. rii technique is used for significance tests. the reference category used in the means test is indicated in bold face. significance test is for mean difference from reference category for each variable. 8 k. tae kim et al. / financial services review 28 (2020) 1-16 4.3. multivariate results table 3 shows the logistic regression results of the likelihood of three different types of life insurance ownership from the 1992–2016 scf dataset, controlling for household characteristics. there was a negative relationship between the survey year and the likelihood of life insurance ownership for any life insurance and cash value life insurance, but the relationship was not significant for term life insurance ownership. the magnitudes of the coefficients of the time trend for ownership of any life insurance, and also for ownership of cash value life insurance were very similar between the results in table 3 and the results in table 1, suggesting that the negative trends in ownership and in cash value life insurance were not because of changes in household characteristics. however, the effect of the time trend on term life insurance ownership was not significant in table 3, and the magnitude of table 3 logistic regression analysis of likelihood of ownership of any, term, and cash value life, 1992–2016 scf any life insurance term life insurance cash value life insurance coefficient standard error coefficient standard error coefficient standard error survey year �0.0277*** 0.0034 �0.0039 0.0030 �0.0422*** 0.0033 age of head 0.0776*** 0.0093 0.0838*** 0.0092 0.0510*** 0.0108 age squared (/10000) �6.3726*** 0.8658 �8.6443*** 0.8716 �2.6687** 0.9693 race/ethnicity (reference: white) black 0.5309*** 0.0862 0.2519** 0.0793 0.3662*** 0.0903 hispanic �0.9927*** 0.0972 �0.7450*** 0.0957 �0.8456*** 0.1414 asian/other �0.3491** 0.1244 �0.2491* 0.1155 �0.2701* 0.1330 homeowner 0.7555*** 0.0636 0.5241*** 0.0614 0.5487*** 0.0738 marital status (reference: married) single male �0.7504*** 0.0793 �0.5418*** 0.0754 �0.4549*** 0.0864 single female �0.7767*** 0.0689 �0.5186*** 0.0651 �0.5665*** 0.0753 partnered �0.6230*** 0.1021 �0.4268*** 0.0970 �0.3280** 0.1205 the presence of child under 18 0.1236 0.0657 0.1713** 0.0582 0.0739 0.0646 current income relative to normal (reference: normal) income higher 0.0294 0.0898 �0.0272 0.0762 0.1150 0.0799 income lower �0.3024*** 0.0671 �0.2439*** 0.0628 �0.1013 0.0724 log of household income 0.0547*** 0.0138 0.0218 0.0129 0.0451** 0.0147 employment status of head (reference: salary worker) self-employment �0.5130*** 0.0736 �0.6081*** 0.0624 0.2854*** 0.0646 not working �1.1669*** 0.1182 �1.0809*** 0.1190 �0.3563* 0.1546 retired �0.7338*** 0.0883 �0.6588*** 0.0793 �0.0608 0.0854 education of head (reference: less than high school) high school degree 0.3858*** 0.0842 0.1921* 0.0828 0.4210*** 0.1008 some college 0.5439*** 0.0901 0.3173*** 0.0870 0.4984*** 0.1050 bachelor degree 0.7273*** 0.0951 0.4433*** 0.0895 0.6248*** 0.1055 post-bachelor degree 0.5862*** 0.1015 0.4375*** 0.0942 0.5248*** 0.1092 health status of head (reference: excellent health) good health 0.0215 0.0615 0.0278 0.0537 0.0007 0.0574 fair health �0.0697 0.0813 �0.0677 0.0747 �0.0392 0.0831 poor health �0.2140 0.1256 �0.0945 0.1239 �0.2482 0.1422 intercept 53.3805 6.7225 5.6963 6.0027 80.5698 6.5469 mean concordance 75.6% 70.7% 72.0% note: *p � 0.05, **p � 0.01, ***p � 0.001. unweighted analysis with rii technique. n � 43,751. 9k. tae kim et al. / financial services review 28 (2020) 1-16 the effect was much less than the magnitude in table 1, suggesting that the decrease in term life insurance might have been because of changes in household characteristics from 1992 to 2016. both age and age squared were significantly related to the likelihood of insurance ownership in each regression. the combined effect of the age variables implies that at the 2016 mean values of other independent variables, the likelihood of owning any life insurance increased with age until age 61, then decreased with age; ownership of term life insurance increased with age until age 48, then decreased with age; and ownership of cash value life insurance increased with age for all ages in the sample. fig. 2 shows the calculated relationship between age of the head and the likelihood of ownership of each type of life insurance, based on the coefficients in table 3 and assuming the 2016 mean values of other independent variables. this pattern is plausibly related to the cost of life insurance, which increases substantially with age. households who purchased 20-year term policies in their 40s and 50s would have expiring policies in their 60s and 70s. at this point in life the cost of the insurance may outweigh the benefits, particularly if the household has relatively little debt and has adequate assets to offset the loss of income resulting from the death of a spouse or significant other. on the other hand, given estate planning goals and reduced opportunities for tax sheltered retirement accounts after age 70, the estimated increase in ownership with age of cash value ownership is plausible, since household income is assumed to remain at the overall 2016 sample mean. for the remaining independent variables in table 3, we discuss only the effects for ownership of any life insurance, though for many of the variables, the effects were similar for ownership of term life insurance and of cash value life insurance. for racial/ethnic patterns based on the regressions in table 3, assuming other household characteristics had the mean levels for 2016, black households had the highest predicted likelihood of owning life insurance, over 73%, compared with 61% for white households, 53% for asian/other households, and only 36% for hispanic households. there is no theoretical reason for there fig. 2. effect of age on the proportion of households owning any, term, and cash value life insurance, assuming other household characteristics have 2016 mean levels. estimates based on logistic regressions in table 3. 10 k. tae kim et al. / financial services review 28 (2020) 1-16 to be differences in life insurance ownership between difference race/ethnic groups, if other household characteristics are equal. the reason for these differences may be related to differences in marketing practices of insurance companies. hispanic and asian households consist of a higher proportion of immigrant households, so it is possible that insurance companies have not effectively found a way to market to these individuals (mountain, 2015), especially if life insurance is a product that is not well established in their country of origin. hispanic and asian households may also rely more heavily on their extended family network than both black and white households, something that is not measured in the scf. for the relationship between marital status and life insurance ownership, assuming other household characteristics had the mean levels for 2016, married households had the highest predicted likelihood of owning life insurance, almost 70%, compared with 56% for partner households, 50% for single male households, and 49% for single female households. surviving individuals in a coupled household are likely to be financially burdened if a spouse or partner were to die. life insurance can help fill this burden. in general, single headed households are not financially dependent on another wage-earner and have little need for life insurance, because there would be no surviving spouse, though as nam and hanna (2019) discussed, there might still be a need in terms of making sure the person with custody of dependent children would have adequate resources. the difference between married and partnered coupled households may be because of the married households having a longer-term time horizon than otherwise similar partnered households. presence of a dependent child under age 18 was not significantly related to ownership of any life insurance or of cash value life insurance, but was positively related to ownership of term life insurance. this is reasonable, because dependent children make the need for support for them paramount. household income was positively associated with the likelihood of owning any life or of owning cash value life insurance. for the relationship between income shocks and life insurance ownership, assuming other household characteristics had the mean levels for 2016, households with current income below normal had the lowest predicted likelihood of owning life insurance, 54%, compared with 64% for households with higher than normal income and 61% for households with normal income. this may be because of household budget constraints. households may accurately perceive the likelihood of death in the current year as very low, and, thus, temporarily terminate a life insurance policy if household income falls below normal income. conversely, when household income is above normal income, households may feel less of a constraint on the household budget and use this extra income on life insurance, something they otherwise were not willing to own. current income was positively related to the likelihood of owning any life insurance and to the likelihood of owning cash value life insurance, but not to the likelihood of owning term life insurance. given that employment status is controlled, the lack of significance is reasonable, as an employer group life insurance is something that is available for many employees without having to make a purchase decision. for the relationship between employment status of the head and life insurance ownership, assuming other household characteristics had the mean levels for 2016, households headed 11k. tae kim et al. / financial services review 28 (2020) 1-16 by an employee had the highest predicted likelihood of owning life insurance, 68%, presumably partly because of the availability of group life insurance from many employers. group insurance may be provided for free or at reduced rate as an employer benefit. additionally, group life insurance removes the barrier of having to have a medical exam to be insured, something that is common in the individual insurance market. this barrier may prevent individuals from seeking life insurance, even if they would qualify for the policy upon completion of the exam. the predicted life insurance ownership rates were only 51% for households with a self-employed head, 49% for households of a retired head, and 40% for households with heads not working but not of retirement age. for the relationship between the education of the head and life insurance ownership, assuming other household characteristics had the mean levels for 2016, households headed by somebody with a postbachelor degree had a predicted life insurance ownership rate of 65%, and those with a bachelor degree had a predicted rate of 68%, compared with 62% for those with some college but no degree, 57% for those whose highest education is a high school degree, and 44% for those with no high school degree. higher education levels are likely related to more future oriented thinking (yuh and hanna, 2010), which would be consistent with the life insurance ownership patterns shown in fig. 10. for the relationship between homeownership status and life insurance ownership, assuming other household characteristics had the mean levels for 2016, homeowners had a predicted ownership rate of 67%, compared with a predicted rate of 47% for renters. homeowners with a mortgage may need to have life insurance to make sure the loan is paid off if a primary earner dies. homeownership may also be a signal of financial stability. typically, homeowners are required to make a substantial down payment to purchase a home, which, for most households requires substantial financial planning. therefore, homeowners may be more likely to both afford and plan for life insurance. 4.4. the association between use of financial planner and life insurance ownership table 4 shows the logistic regression results of the association between the use of comprehensive financial planner and the likelihood of different types of life insurance table 4 logistic regression analysis, the association between comprehensive financial planner use and the likelihood of life insurance ownership, 1998–2016 scf any life insurance term life insurance cash value life insurance coefficient standard error coefficient standard error coefficient standard error use of comprehensive financial planner 0.3727*** 0.0425 0.2117*** 0.0356 0.2989*** 0.0364 survey year �0.0230*** 0.0021 0.0169 0.0020 �0.0398*** 0.0022 intercept 44.0071*** 4.2934 �5.6204 3.9183 75.8085*** 4.3208 other control variables yes yes yes mean concordance 75.9% 71.3% 71.7% note: *p � 0.05, **p � 0.01, ***p � 0.001. unweighted analysis with rii technique. n � 35,667. control variables include the independent variables shown in table 3. 12 k. tae kim et al. / financial services review 28 (2020) 1-16 ownership, controlling for the other independent variables in table 3. the scf did not have a specific code for use of financial planners before 1998, so we used a pooled dataset from the 1998 –2016 scf for this additional analysis. the utilization of a comprehensive financial planner was positively associated with ownership of any, term, and cash value life insurance policies. households who used a comprehensive financial planner had 45.2% higher odds of owning any life insurance than those without using a financial planner. similarly, the use of comprehensive financial planner increased the odds of owning term life and whole life insurance policy by 24% and 35%, respectively. controlling for use of a financial planner did not substantially change the effect of the survey year compared with the results in table 3, as the survey year had a negative effect for any life insurance and for cash value life insurance, but the effect was not significantly different from zero for term life insurance. 5. discussion and implication the effect of the time trend in the logistic regression means that even if the characteristics of households in the united states had not changed between 1992 and 2016, there would have been a substantial decrease in ownership of any life insurance and of cash value life insurance during that period. however, based on the logistic regression in table 3, ownership of term life insurance might not have decreased if household characteristics had not changed. the trends for some household characteristics related to lower likelihood of life insurance ownership (hispanic and asian/other households, unmarried households) might have contributed to lower ownership during the 1992 to 2016 period, but the trends for education and age might have contributed to increased ownership. overall, changes in household characteristics during the periods played a relatively small role in the decreases in ownership of any and of cash value life insurance, partly because most household characteristics changed slowly. for instance, the proportion of households with married couples decreased from 53.9% in 1992 to 46.8% in 2016, and the proportion of household heads who were employees (vs. self-employed or retired or not working) changed from 55% in 1992 to 56% in 2016. insuring human wealth is an appropriate life insurance objective (chen, ibbotson, milevsky, and zhu, 2006) but even at the current ownership rates, the amount of life insurance is not enough to replace lost human wealth for many households (mountain, 2015). while the cash-value policy decline may reflect rational decision-making because of changes in tax laws and the increase in tax advantaged investment options, these patterns could change with changes in the federal income tax and with any substantial changes in the federal estate tax. durham (2015) found that households often overestimate the cost of life insurance and do so by drastic amounts. when households were asked to provide what they thought a $250,000 20-year term policy would cost, one-quarter overestimated the cost by 625% and one half overestimated the cost by 250%. meanwhile, life insurance advice from a financial planner is something that families place a low value on (warschauer and sciglimpaglia, 2012), making life insurance planning something a family is more likely 13k. tae kim et al. / financial services review 28 (2020) 1-16 to do on its own. while we found higher likelihood of life insurance ownership for households who have a financial planner, we cannot be certain if the financial planner’s advice led to life insurance ownership or if households who have life insurance were more likely to seek advice from a financial planner. future research might use methods to correct for the selection effect, as was done by kim, pak, shin, and hanna (2018) for analyses of the effect of financial planner use on holding a retirement savings goal. it is of little surprise that life insurance ownership trends are decreasing when households place little importance on life insurance planning while also dramatically overestimating its cost. if financial planners were to view term life insurance as a loss leader product, or a breakeven product, instead of a profit generator, both the financial planner and client might be better off. as a gateway product, term insurance can help establish trust and rapport with clients. planners could start by asking the client how much they think a $100,000 20-year term life policy would cost them, and see how that compares with reality. once the client has the peace of mind that his or family will be financially taken care of upon an untimely death, relatively inexpensively, transition to other planning areas can ensue. this approach would likely benefit financial planners who are following or not following a fiduciary standard, recommendations of term life insurance without direct financial benefit may diminish apprehension some clients might otherwise feel may have when presented with the advice to purchasing life insurance. for clients with more complicated financial needs where cashvalue life insurance is appropriate, it may be a product that is brought up later, rather earlier, in the financial plan. if the ownership rates continue to decline, more and more families will face financial hardship and economic distress if a spouse or partner dies prematurely. in terms of the potential benefits of financial planning advice based on normative economic models, risk management is an important component of the value of advice (hanna and lindamood, 2010), and for many families, life insurance still should be a salient component of a financial plan. this is surprising as there are several studies have found that a substantial proportion of households still remain underinsured. financial planners and educators can better inform their clients and the population as to the merits of life insurance. additionally, alternative products, adjustments to the current life insurance products, or additional saving will need to fill the void of the disappearing life insurance. given our conclusion that the decreases in ownership of any life insurance and in ownership of cash value life insurance were not because of changes in household characteristics between 1992 and 2016, it is plausible that future changes in federal income tax and estate tax policies could lead to future increases in life insurance ownership. for instance, if a proposal by senator bernie sanders to decrease the estate tax exemption from $11.4 million in 2019 to $3.5 million (2009 level) were to be implemented, many wealthy households would likely make substantial changes in their estate planning (shenkman, 2019), and cash value life insurance might provide one type of strategy. further, proposed increases in federal income tax rates might also make cash value life insurance more attractive. even if no near-term tax legislation is passed, most individual income tax changes in the tax cuts and jobs act of 2017 are set to sunset january 1, 2026. the increases in income tax rates might nudge households to reconsider cash-value life insurance policies. 14 k. tae kim et al. / financial services review 28 (2020) 1-16 references auerbach, a. j., & kotlikoff, l. (1991). life insurance inadequacy-evidence from a sample of older widows (nber working papers 3765). national bureau of economic research. baek, e., & devaney, s. a. (2005). human capital, bequest motives, risk, and the purchase of life insurance. journal of personal finance, 4, 62–84. bernheim, b. d., carman, k. g., gokhale, j., & kotlikoff, l. j. (2003). are life insurance holdings related to financial vulnerabilities? economic inquiry, 41, 531–554. bernheim, b. d., forni, l., gokhale, j., & kotlikoff, l. j. (2003). the mismatch between life insurance holdings and financial vulnerabilities: evidence from the health and retirement study. american economic review, 93, 354–365. board of governors of the federal reserve system. (2014). codebook for the 2013 survey of consumer finances. (available at http://www.federalreserve.gov/econresdata/scf/files/codebk2013.txt) board of governors of the federal reserve system. (2017). codebook for the 2016 survey of consumer finances. (available at https://www.federalreserve.gov/econres/files/codebk2016.txt) brandon, jr., e. d., & welch, h. o. (2009). the history of financial planning: the transformation of financial services. hoboken, nj: john wiley & sons. certified financial planner board. (2015). 2015 principal knowledge topics. (available at https://www.cfp.net/ become-a-cfp-professional/2015-job-task-analysis/2015-principal-knowledge-topics) chen, p., ibbotson, r. g., milevsky, m. a., & zhu, k. x. (2006). human capital, asset allocation, and life insurance. financial analysts journal, 62, 97–109. cordell, d. m., finke, m., & lemoine, c. (2007). the death of life? the demand for cash-value life insurance has slipped in recent years, say the authors, much of it due to bad publicity. journal of financial planning, 20, 42–44. durham, a. (2015). 2015 insurance barometer study. life happens. (available at https://www.lifehappens.org/ industry-resources/2015barometer/) elmerick, s. a., montalto, c. p., & fox, j. j. (2002). use of financial planners by u.s. households. financial services review, 11, 217–231. appendix 1 life insurance variables in the scf variable description x4001 do you (or anyone in your family living here) have any life insurance? please include individual and group policies, but not accident insurance. 1. *yes 5. *no x4002 the two major types of life insurance are term and cash-value policies. term policies pay a benefit if the insured person dies, but otherwise have no value. they are often provided through an employer or union, but may also be bought by individuals. cash-value policies also pay a death benefit, but differ in that they build up a value as premiums are paid. are any of your (family’s) policies term insurance? 1. *yes 5. *no 0. inap. (no life insurance: x4001∧ � 1) x4004 do you have any policies that build up a cash value or that you can borrow on? these are sometimes called ‘whole life’, “straight life,” or “universal life” policies. 1. *yes 5. *no 0. inap. (no life insurance: x4001∧ � 1) 15k. tae kim et al. / financial services review 28 (2020) 1-16 federal reserve bank. (2016). codebook for 2016 survey of consumer finances. washington, dc: federal reserve bank. finke, m. s., huston, s. j., & waller, w. (2009). do contracts impact comprehensive financial advice? (available at http://dx.doi.org/10.2139/ssrn.1429807) frees, e. w., & sun, y. (2010). household life insurance demand: a multivariate two-part model. north american actuarial journal, 14, 338–354. glazer, d. a. (2007). the effect of commission versus non-commission benefits on customer value: the case of life insurance policy performance. financial services review, 16, 135–153. greenwald, l., & fronstin, p. (2019). the state of employee benefits: findings from the 2018 health and workplace benefits survey. employee benefit research institute no. 470. (available at https://www.ebri.org/ docs/default-source/ebri-issue-brief/ebri_ib_470_wbs2-10jan19.pdf?sfvrsn�c5db3e2f_10) guillemette, m., hussein, m., phillips, g. m., & martin, t. (2014). minority household size and the life insurance purchase decision. financial services review, 24, 37–51. gutter, m. s., & hatcher, c. b. (2008). racial differences in the demand for life insurance. journal of risk and insurance, 75, 677–689. hanna, s. d., & lindamood, s. (2010). quantifying the economic benefits of personal financial planning. financial services review, 19, 111–127. hanna, s. d., ji, h., lee, j., son, j., letkiewicz, j., lim, h., & zhang, l. (2011). content analysis of financial services review. financial services review, 20 (3), 237–251. hanna, s. d., kim, k. t., & lindamood, s. (2018). behind the numbers: understanding the survey of consumer finances. journal of financial counseling and planning, 29, 410–418. heo, w., grable, j. e., & chatterjee, s. (2013). life insurance consumption as a function of wealth change. financial services review, 22, 389–404. kim, k. t., pak, t. k., shin, s. h., & hanna, s. d. (2018). the relationship between financial planner use and holding a retirement saving goal: a propensity score matching analysis. financial planning review, 1, e1008. https://doi.org/10.1002/cfp2.1008 liebenberg, a. p., carson, j. m., & dumm, r. e. (2012). a dynamic analysis of the demand for life insurance. journal of risk and insurance, 79, 619–644. lindamood, s., hanna, s. d., & bi, l. (2007). using the survey of consumer finances: some methodological considerations and issues. journal of consumer affairs, 41, 195–222. mountain, t. p. (2015). life insurance and financial vulnerability. ph.d. dissertation, the ohio state university. mulholland, b., finke, m., & huston, s. (2016). understanding the shift in demand for cash value life insurance. risk management and insurance review, 19, 7–36. nam, y., & hanna, s. d. (2019). the effects of risk aversion on life insurance ownership of single-parent households. applied economics letters, 26, 1285–1288. http://dx.doi.org/10.1080/13504851.2018.1546044 retzloff, c. d. (2010). household trends in u.s. life insurance ownership. limra, report, november, 1. scism, l. (2016, feb 25). death of the insurance salesman at metlife. the wall street journal. scott, j. k., & gilliam, j. (2014). boomers’ life insurance adequacy pre & post the 2008 financial crisis. financial services review, 23, 287–304. shenkman, m. (2019, march 21). sanders estate tax proposal: estate planning steps to take now. (available at https://www.forbes.com/sites/martinshenkman/2019/03/21/sanders-estate-tax-proposal-estate-planning-steps-to-take-now/#16bae6299f5e) shin, s. h., & hanna, s. d. (2017). accounting for complex sample designs in analyses of the survey of consumer finances. journal of consumer affairs, 51, 255–283. warschauer, t. a. (2012). the economic benefits of personal financial planning: an empirical analysis. financial services review, 21, 195–208. yuh, y., & hanna, s. d. (2010). which households think they save? journal of consumer affairs, 44, 70–97. 16 k. tae kim et al. / financial services review 28 (2020) 1-16 career and education choice as central elements of longterm financial planning inga timmermana,*, nikanor volkovb afinance, financial planning and insurance, david nazarian college of business and economics, california state university, northridge, northridge, ca 91330, usa bstetson-hatcher school of business, mercer university, 3001 mercer university dr, atlanta ga, 30341, usa abstract career and education choice play a central role in individual’s long-term financial planning. in defining individual’s overall wealth, we not only consider the financial wealth (net worth), but also incorporate individual’s human capital (expressed as the present value of future earnings) as a component of the wealth function. we demonstrate that human capital accounts for the majority of an individual’s wealth portfolio in most cases for most of individual’s life. by including human capital in the traditional theoretical portfolio choice framework, we show that the choice of career and education level has a significant effect on the sharpe ratio of an individual’s overall wealth portfolio. we conclude by providing several examples of the computations of the present values of future average earnings streams for individuals of various education levels from several different occupations. the calculations are meant to illustrate how an individual can perform a simple npv-like analysis when in the process of career planning. © 2020 academy of financial services. all rights reserved. 1. introduction and motivation while the conventional approach to measuring individual wealth is through an estimation of the dollar value of financial assets (cash, stocks, bonds, etc.) and real assets (real estate, private business, etc.), such a view deemphasizes human capital.1 yet, human capital rather than financial capital is responsible for the bulk of the wealth for most individuals during many stages of their lives, especially during their younger years. using a small sample from *corresponding author. tel.: +1-818-677-4615; fax: +1-818-677-6079. e-mail address: inga.chira@csun.edu, www.csun.edu/ 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 179–200 the mid-1980s u.s. survey of consumer finances, lee and hanna (1995) find that financial assets represent only two percentage of the total wealth of most households.2 furthermore, the value of individual’s financial and real asset portfolio often reflect the human capital value possessed by the individual in the past because there exists a strong relationship between the value of human capital and an individual’s long-term financial wealth. for example, budria, diaz-gimenez, quadrini and rios-rull (2002) and dıaz-gimenez, quadrini, and rios-rull (1997) show high correlations between the level of individual household’s earnings (defined as wages and salaries from labor) and household wealth. we build on the existing literature on the relationship between education, career choice, and the individual’s lifelong wealth profile by proposing a new approach that incorporates the value of human capital in the wealth function. despite the financial importance of the topic for an average individual, there is not much research in financial planning that focused on the optimization of lifetime earnings. this is understandable given that financial planners do not have much input into the career paths or education levels of their clients. traditionally, by the time one becomes a financial planner’s client, they are already in the asset (financial and real) accumulation or decumulation phase. at that point, the goal of the financial planner is to maximize the value of the existing assets and/or cash flows. in general, as people get older, the value of their financial and real assets increases, while human capital decreases, which results in a gradual shift of the individual’s planning focus from earning maximization to efficient investments in financial and real assets that, in time, would provide sufficient cash flows in retirement. earnings maximization does not imply that an individual will always seek the highest paid career (job). rather, it implies that, everything else held constant, an individual will choose a career (job) that provides greater financial remuneration. given the apparent shift in the financial planning industry3 such as the attempt to work with younger and less affluent clients and the emergence of a new type of planner (one who seeks to work with the younger population), planners who compete for the business of younger clients and who strive to help clients through the entire life span should first prioritize on development of their clients’ human capital along with value-maximization of clients’ financial capital. the value of human capital is primarily driven by an individual’s skills, which, in turn, results from (1) education level, (2) career choice, and (3) experience. thus, examining the impact of education and career choice should be an integral step in maximizing the individual’s total wealth. financial planners can help their clients decide what to study and where to study.4 by developing a framework that will incorporate human capital, the biggest asset an average younger client can bring to a planner’s attention, into the conversation of overall wealth maximization, advisors can help their clients achieve financial goals more effectively and more efficiently. a holistic view of an individual’s wealth portfolio as a combination of human capital as well as financial and real assets shows that, for most of an average individual’s working life, human capital is responsible for most or nearly all of the individual’s wealth. as one ages, the proportion of human capital in the value of overall wealth drops while the value of financial and real assets in the portfolio grows (see fig. 1). the objective of long-term financial 180 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 planning is to ensure that, at some point in time, the return from the financial and real asset portfolio fully substitutes for the need to produce financial return on the possessed human capital (retirement). there exists a clear and, arguably, nonlinear relationship between the value of human capital and the long-term value of one’s financial and real asset portfolio. in this study we develop a model that incorporates the value of human capital into the framework of overall financial wealth for an average individual. by looking at total wealth, including human capital, we aim to expand the current real and financial asset-centric view on investments. we show that, despite one’s preferences for academic majors and career paths, one can still make optimal financial decisions within the chosen career path by examining the lifetime earnings across several subdimensions within the career of interest. our contribution to the current literature is in presenting a model that can be used to maximize the individual’s wealth portfolio through an emphasis on the financial implications of a given education level and career path. the model also demonstrates that inclusion of human capital in the individual’s portfolio choice problem maximizes the value of the sharpe ratio, thereby optimizing the relationship between the expected return and the risk of the wealth portfolio. additionally, we provide several useful examples of the computation of the present value of the lifelong earnings for different education-occupation combinations. these calculations serve as a guide in making education and career choice decisions. finally, we calculate the net present value of the difference between the average age-earnings profiles in given education with various education levels beyond high school and with high school education. the net present value represents a maximum expenditure one should be willing incur for education beyond high school if her goal is to increase her overall wealth as the result of achieving the given level of education. in other words, the net present value represents the maximum tuition one should pay for the given education level beyond high school. the sample calculations are performed by occupation and are meant to aid in establishing an optimal education level for a given occupation.5 fig. 1. wealth profile. i. timmerman, n. volkov / financial services review 28 (2020) 179–200 181 the study proceeds as follows: section 2 provides a brief overview of related literature; section 3 discusses the proposed model of total wealth; section 4 illustrates the application of the proposed model to portfolio theory; section 5 provides several applied examples of the impact of education and career choice on the overall wealth; and section 6 presents the implications and limitations of the model. 2. literature review 2.1. human capital and the financial planning profession despite very limited academic research in financial planning on the topic of incorporating human capital into a traditional investment model, some practitioners are attempting to broach the topic with their clients. this is not surprising, as the decision to pursue a specific career path, change careers, or pursue advanced education within a career track could add more return to one’s overall investment portfolio and lifetime earnings than many of the typical decisions on which financial planners spend much of their time. even when the intention is present, the lack of easily accessible resources and, at times, lack of training and knowledge about the subject makes it difficult for financial planners to provide advice as to the optimal amount of education and career choice to their clients. some financial planners have identified this as a problem and are trying to incorporate the human capital aspect into their planning practices. for example, haubrich (2013) argues that the client’s most important asset is the human capital equity and that career choice is a new asset class that needs to be factored into the financial planning process. as such, clients are encouraged to keep separate working capital funds for education, career sabbaticals and changing jobs. this idea, however, has not gained significant traction and planners who focus on it are hard to find. while the incorporation of human capital in financial planning is relatively new, the broader literature in economics, management and finance addresses the question of human capital optimization. in the “people equity” framework, schiemann (2006) looks at talent optimization and the factors that influence it. caballe and santos (1993) explore the importance of human capital in the context of growth by using a generalized lucas-uzawa model of endogenous growth that includes both physical and human capital, concluding that human capital is at the core of economic growth. 2.2. the value of education in an academic setting, milevsky has focused on the role of human capital on the personal balance sheet. the investment in human capital may depend on several factors, one of which, according to hanna et al. (2001) is the individual’s risk tolerance. other factors include income, net worth, work history and specific job skills. despite all these controls, education is still seen as a primary factor in attainment of human capital maximization. although generalized theoretical models that incorporate human capital are hard to find, the importance and central place of human capital in planning out one’s financial life has been discussed in specific contexts in the past. for example, bridges, de’armond, and dean 182 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 (2013) examine the human capital of divorced women, concluding that women who do not invest in their human capital through education and work experience end up being negatively exposed to financial shocks. similarly, britt-lutter et al. (2018) apply the human capital framework to argue for increased levels of education (with higher knowledge and positive behavior) and as a result, higher well-being in college students. as noted earlier, the path to maximization of human and, therefore, total capital is through maximization of the return from one’s career choice at an optimal level of education. there is a significant body of existing research about the economic benefit of education. increased investment in education has several economic benefits, among which the most directly measured is increased earning capacity (morgan & david, 1963). in addition to overall higher earnings, more education also translates into steadier jobs.6 studies on the differential in pay between different levels of education have been conducted since the 1960s. over time, the gap in pay and arguments in favor of education shifted but the central theme is still the same: more education translates into more lifetime earnings. in the later part of the 1990s and in developed countries, the gap between more and less educated workers widened. a canadian study shows that this trend resulted in higher demand and higher income for more educated individuals, and increased the returns on investment in education (chung, 2006). in the early 2000s, though, the highest job growth has shifted to skilled labor such as gas extraction, construction, and real estate. this, however, may be more of a consequence of a strong job market in general than a sign of decline in the economic payoff of education. the value of education is also dependent on the condition of the economy. chakrabarti, jiang, and nober (2019) look at the graduates of four year colleges during boom and bust conditions and conclude that on average, students who enter college in bad economic times are on average going to have earnings about 10% lower than the students who enter college in boom times, but the earnings differential is dependent on the choice of major. again, this points to the idea that human capital cannot be viewed in a uniform way and that financial planners can enhance their services by advising on career and education choices to their clients. on average, most studies conclude that increased education results in higher earnings independent of career choice. while it is easy to identify outliers who, despite the lack of a degree, became financially successful, our goal is to study average individuals and the education and career choices they face with limited information availability. brown et al. (2012) estimated that the average return of a college education over high school is $300,000. however, the degree, the choice of college and the occupation choice add significant variability to this number. 2.3. the value of career choice in an early article on the topic, theoretically, riley (1976) addresses the marginal cost of education across different individuals in the context of productivity. if the marginal earnings increase is positive, individuals will stay in school and acquire additional education. by presenting a human capital model of education screening, he argues that there exists a weak equilibrium at specific wages. he concludes that there is a specific wage at which individuals are hired based on their education level and paid based on their productivity. i. timmerman, n. volkov / financial services review 28 (2020) 179–200 183 oreopoulos and petronijevic (2013) argue that, to make the best decision about the investment in college education, students need to estimate the cost of schooling, their major and their anticipated occupation. some of the newest research shows that the relationship between income inequality and number of years spent on education is still positive; additional years of schooling result in higher income, on average, and that this relationship is stronger in emerging and developing countries (imf, 2017). overall, it is hard to argue with the fact that, on average, the decision to continue education results in higher economic returns as well as lower unemployment rates.7 a 2011 pew research center study conducted by taylor et al. (2011) finds that 86% of respondents believe their college experience was a good investment. adults who graduated from a fouryear college earned on average $20k more per year than adults who did not receive a degree beyond high school. according to leonhardt (2014), in 2013 the gap in pay reached a new high: americans with a college degree made 98% more per hour than people with no degrees. this was five percentage more than five years earlier and 35% more since 1980. furthermore, some studies demonstrate that the level of education has a significant impact on the work life expectancy of an individual. as such, skoog, ciecka, and krueger (2011 and 2019) show that individuals with higher education levels consistently have longer work life expectancies, everything else held constant. for example, a 25-year-old female who is initially active in the labor market and holds a high school diploma has a work life expectancy of 28.68 years, a female with the same characteristics, but with a bachelor’s degree has a 33.20year work life expectancy (data obtained from skoog et al., 2011). the work life expectancy increases by another 2.7 years for a master’s degree and jumps by almost three more years if this individual is to have a terminal degree. these differences in the work life expectancy inevitably have a very significant impact on long-term financial planning as they increase the lifelong earning potential and decrease the need for additional retirement funding. the importance of increased work life expectancy is particularly important in the context of increasing longevity.8 as bajtelsmit and wang (2018) find, the “top third of households by longevity need approximately 20% more retirement wealth” than average longevity households. however, as already mentioned, education alone is not sufficient. career choice is also essential in lifelong wealth maximization. the major, career path and subsequent specialization within that path influence the economic benefits of the decision to pursue advanced degrees. a recent study from the georgetown university center on education and the workforce examines how the payoff is related to the field of study. it concludes that some majors, for example engineering, have a strong financial advantage that results in an extra $1.1 million in lifetime income. similar conclusions were reached about graduate school. compared with the research on education choices, the research on the optimal financial payout of career choice is sparse. kyrychenko (2008) uses a comprehensive mean-variance model to incorporate non-financial aspects of one’s life into the optimal asset allocation model. by examining factors such as human capital, the author derives optimal portfolios for workers in different industries and locations. bernardo et al. (2017) examine the payoff for different university majors in chile, concluding that, although some majors are more desirable than others, overall higher education is still an attractive option. similar results were found by mun and wong (1999) in singapore. although such studies are generally 184 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 informative, they are hard to implement for the average student as they are limited and circumstantial to the sample environment. in a more generalized setting, davies and guppy (1997) use a longitudinal study in the united states to examine the characteristics of students who pursue financially lucrative careers. they find that, on average, males are more likely to enter fields with financially higher expected outcomes than females, but socio-economic factors do not affect this decision net of other background factors. roksa and levey (2010) sum up the state of research on the relationship between education and income by stating that “while income inequality between college graduates is documented, inequality in occupational status remains largely unexplored.” in this article we combine the literature on the optimal education level and optimal career choice by developing a theoretical model that incorporates the two decisions. our goal is to present a model that would help an individual to arrive at the best combination of education and career choice to maximize her overall wealth portfolio, a portfolio that includes human capital along with financial and real capital it is important to mention that there are a number of additional factors that might influence lifetime earnings such as change in jobs, continuous education required of specific degrees, the stability of the job, mortality risk and longevity risk. although these factors are all important in determining the optimal lifetime earnings of a specific individual, our goal is to look at the data on average. the scope of this paper is to introduce a model that optimizes the level of education given a chosen career path. 3. the wealth model for simplicity, assume that individuals do not possess any financial or real assets at time zero. time zero is assumed to be the time when an individual first enters the labor market. denote overall individual’s total wealth at time t aswt, 9 therefore: wt ¼ et þ it where et is the present value of the sum of the future earnings (wages and salaries from labor) of an individual and it is the current value of the investment portfolio of an individual. et can be expressed as: et = o t t=1 et 1þ gtþ1ð þt 1þ rtð þt ! where et is the earnings in a given year t with the first year being the first year of individual’s employment, et is the function of education, occupation and experience and can be presented in a following form: et ¼ f st,ot,xtð þ where s is the level of individual’s education; o is the individual’s occupation; x is the individual’s level of experience, or i. timmerman, n. volkov / financial services review 28 (2020) 179–200 185 e1 . . . et 0 @ 1 a ¼ s1 o1 x1 . . . . . . . . . st ot xt 0 @ 1 a t is the remaining time that the individual has in the labor force (earnings span); g is the expected rate of growth of earnings that is due to gained experience and inflation adjustment; and r is the rate of return on assets, which need not be constant over time. the rate r therefore can be presented as follows: rt = warf þ w et –að þrsoa where wa ¼ a et is the weight of a risk-free wage and w (et – a) ¼ 1 � wa is the weight of the wage that stems from the educational and occupational choice an individual makes. rf is the risk-free rate of return and rsoa is the individual-specific required rate of return on the investment in education and career choice, which can also vary with age. the assumption of the model is that the individual can secure a full-time position (40 hours a week) with wage a, which is a risk-free wage. this wage does not require any training or experience and, in an empirical setting, can be assumed as the minimum wage. because wage a is assumed to be risk free, it is discounted at the risk-free rate rf. note that the earnings growth rate g does not need to be constant over time and can take both positive and negative values; g also incorporates both experience related earnings growth and inflation adjustment. the current value of the individual’s investment portfolio, it, can be expressed as: it ¼ o n t¼1 b tet 1þ ið þn where b t is a percentage of earnings et that an individual contributes to her portfolio of financial and real assets in time t; n is the number of periods between the first time the assets have been allocated to the investment portfolio and time t; and i is the average annual investment return achieved during period n. note that b does not need to be constant over time and can take both positive and negative values. a negative b implies a voluntary withdraw of funds from the investment portfolio. fig. 1 provides an illustration of a typical age-wealth profile of an individual. for the purposes of this illustration, the initial value of the investment portfolio equals 0, the worklife expectancy of the individual is assumed to precede the individual’s life expectancy and it is assumed that the individual does not return to employment after retirement. the illustration in the figure also assumes that the value of the investment portfolio at the time of an individual’s life expectancy is depleted to 0 (no inheritance is left to the heirs from the investment portfolio). both the rate of contributions to the retirement portfolio and the rate of return on the retirement portfolio are assumed to be constant. fig. 1 presents the gradual substitution effect of the value of the human capital portfolio in the total wealth profile by the value of investments. what is notable is that the value of human capital is overwhelmingly the main component of the overall wealth of this 186 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 individual for the majority of her work life expectancy (see total wealth portfolio weights in fig. 2). while the magnitude of the allocation of capital to the investment portfolio (retirement savings) would affect the weights of human and investment capital in the overall wealth portfolio, human capital will still dominate in the weights for the majority of one’s life. 4. application of the wealth model to portfolio theory the traditional mean-variance model proposed by markowitz (1952) suggests that a rational investor should choose a portfolio that provides the highest possible return. this investor is also risk averse; therefore, she chooses an investment with the lowest possible risk. the assumption of risk-free lending and borrowing leads to the conclusion that a rational risk averse investor should invest in an optimal portfolio and adjust for her level of risk aversion by either borrowing to increase the investment into the optimal portfolio or lending at the risk-free rate to reduce the overall portfolio risk. sharpe (1966) introduced the widely accepted sharpe ratio that presents a relationship between the excess return on a portfolio of assets and the risk of such portfolio. it is now regarded as conventional wisdom that the objective of a portfolio manager is to maximize the sharpe ratio of the managed portfolio. fig. 1 provides an illustration of a typical age-wealth profile of an individual. for the purposes of this illustration, the initial value of the investment portfolio equals 0, the worklife expectancy of the individual is assumed to precede the individual’s life expectancy and it is assumed that the individual does not return to employment following retirement. the illustration in the figure also assumes that the value of the investment portfolio at the time of an individual’s life expectancy is depleted to 0 (no inheritance is left to the heirs from the investment portfolio). in the illustration, both the rate of contributions to the retirement portfolio and the rate of return on the retirement portfolio are assumed to be constant; they do not need to be constant in the theoretical model and are clearly not constant in practice. fig. 2. portfolio weights. i. timmerman, n. volkov / financial services review 28 (2020) 179–200 187 champagne and kurmann (2013) demonstrate that while the volatility of wages has increased substantially between the 1950s and the 2000s, it is still very low relative to any other risky assets available to investors. according to their findings, the volatility of wages is below 1.02%. furthermore, because wage changes do not behave identically to stock returns, one can also view volatility of wages as the probability of temporary unemployment. given the relatively low historic unemployment rates and availability of alternative employment options in the labor markets, the wage volatility viewed through the prism of unemployment is also rather low. as we illustrate further in this study, the expected return of making a certain education or occupational decision often provides a rather significant return. thus, inclusion of human capital in the calculation of the optimal portfolio results in a significantly higher (and a more complete) sharpe ratio (see fig. 3). this relationship is particularly evident for individuals who are in the first half of their careers, as the present value of human capital (et) profile for most occupations is concave and eventually reaches zero at the time of permanent retirement for all individuals. the depletion of the human capital in one’s overall wealth portfolio is exemplified by the orange arrow in fig. 3 as the individual ages, the weight of the human capital in the optimal portfolio drops gradually only to reach zero (retirement), resulting in a clockwise rotation of the efficient frontier of the individual (green curve on fig. 3) and thus a gradual decrease in the slope of individual’s wealth portfolio sharpe ratio. at the point when the weight of human capital in the wealth portfolio reaches zero, the efficient frontier from fig. 3 is made up of the bond and the stock investment (assuming that the investment assets that an individual possesses consist of an investment into bonds and stocks) only (blue line on fig. 3). we argue that, similarly to a portfolio manager, an individual should be concerned with maximizing the sharpe ratio of her individual overall wealth portfolio, which would in turn fig. 3. human capital as part of investment portfolio. where b stands for bonds, s is stock, and hc for human capital. 188 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 maximize the value of the portfolio (wt) at all times. since the central component of wt is et, which is a function of age, education, and occupation and it is largely a function of et, one’s optimal return stems from an investment in human capital, particularly so at younger ages. in practical terms, this means choosing the optimal level of education for a given career is imperative to achieving such optimality. 5. applied illustrations in this section we provide several examples intended to demonstrate the principle discussed above. we are not suggesting that everyone should become a lawyer, doctor or any other highly paid profession, but rather our examples serve to demonstrate how an individual’s career and education level choice, even within a specific area of interest, can have a significant impact on the overall wealthwt maximization. to provide meaningful results, we use the bureau of labor and statistics american community survey (acs) for the recent period of 2012 through 2016. we identify individuals who are working full-time by including only the observations for individuals who indicated that they were employed at least 50weeks in the last 12months and that worked at least 35 hours per week.10 we then perform quintile regressions of the following form: wagei ¼ a þ b 1ageþ b 2age 2 þ ei where wage is the reported wage of the individual and age is her reported age. the wage variable is adjusted for inflation to reflect 2019 dollars. the adjustment is based on the actual inflation rate for the years 2013 through 2017 and the congressional budget office’s outlook for inflation for years 2018 and 2019.11 age is the age of the individual at the time of the acs data. the age2 term is added to control for the nonlinearity of earnings over the life span of an individual. we run these regressions by the standard occupational classification (soc) code and by specific education level simultaneously. we output the fit of the model (median) by age, given a specific level of education. the result is an age earnings profile (eap) for an individual who possesses a specific education level and is employed in a specific occupation as defined by soc.12 we assume the earliest an individual can secure full-time employment is age 18 and allow one year of no employment for individuals who reported that they have one or fewer years of college experience, two years of no employment for individuals who reported one or more years of college, two years for those who reported that they possess an associate’s degree, four years for those with bachelor’s degrees, six years for master’s degree holders, and nine years for individuals who responded that they have either professional or doctorate degrees (the fitted earnings for these individuals are set to zero until the assumed graduation and commencement of employment). we then compute the present value of the future earnings of these individuals (soc – education sorted age earnings profiles [aep]) on the 18th birthday using a five percentage discount rate. since age earnings profiles are all expressed in today’s dollars, the five percentage discount rate implies a real discount rate applied to future earnings.13 for illustration purposes, below are two examples of the present value of i. timmerman, n. volkov / financial services review 28 (2020) 179–200 189 the generated aeps: (1) for all sales related occupations and (2) for all business-related occupations, as defined by soc.14 5.1. sales and related occupations table 1 displays the present values of future earnings of individuals in a specific occupation within the sales and related occupations soc category at different levels of education. the absence of a number implies that there were not enough observations to output an age-earnings profile for a specific soc-education combination.15 it is notable that education levels of one or fewer years of college, one or more years of college experience, and an associate’s degree result in almost identical earnings for the vast majority of occupations within this category. it is also remarkable that bachelor’s degrees result in a very significant increase in the present value of earnings for all occupations reported in table 1. as can be seen in table 1, attainment of a bachelor’s degree or higher results in more income over the working lifetime of an individual in most of the subcategories of the sales profession. in table 2 we compute the difference between the present value of the earnings of an individual with a high school diploma and an education level that is greater than a high school diploma. the motivation for this calculation is that a high school diploma generally does not require a financial investment by the individual while any higher level of education does require a financial investment (tuition, opportunity cost of not working full-time while working on a degree). thus, the difference in the present value of the age earnings profiles of a higher than high-school educated individual and an individual who has a high school diploma represents a maximum cost that one can pay for a given level of education to have the same financial outcome as if she would have stayed with a high school education, holding the occupation constant (i.e., net present value of obtaining the given level of education). note that the maximum education cost is also computed at age 18, a presumable age at which the education level choice is being made. a negative number in table 2 implies that an individual is better off by not investing in her education beyond high school. the numbers for any given education level are the maximum cost one should incur for all degrees above the high school degree, implying that a dollar figure displayed in the master’s column includes the cost of both bachelor’s and master’s degrees. thus, an individual who wants to pursue a career of a cashier has up to $35,435 dollars (in present value terms at age 18) to complete her bachelor’s degree to end up as well or better off than if she would have earned only a high school education; at the same time, she has only $8,668 to complete both bachelor’s and master’s degrees if she wants to remain indifferent to just a high school diploma. thus, an individual who wants to become a cashier should only earn a bachelor’s degree at a program that costs less than $35,435 over the entire course of study; furthermore, this individual should not consider a master’s degree regardless of its cost as the bachelor’s degree results in higher present value of future earnings than does the master’s degree for this occupation. parts salespersons appear to maximize their wealth by earning only a high school education and not investing in additional schooling. retail sales persons do not benefit from a degree beyond a bachelor’s. these results are not surprising as cashier and parts sales person positions do not normally require specialized 190 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 t ab le 1 s al es an d re la te d o cc u p at io n s p re se n t v al u e o f th e ag eea rn in g s p ro fi le s 0 -1 2 1 2 th n o d ip lo m a h s g e d < 1 y ea r co ll eg e > 5 1 y ea r co ll eg e a ss o ci at es b ac h el o rs m as te rs p ro fe ss io n al d o ct o ra te f ir st -l in e su p er v is o rs o f re ta il sa le s w o rk er s $ 5 2 8 ,5 3 5 $ 5 5 7 ,8 8 4 $ 5 8 8 ,1 0 9 $ 5 6 7 ,6 4 1 $ 6 2 0 ,6 0 1 $ 6 1 2 ,8 1 5 $ 6 1 0 ,1 5 9 $ 7 1 9 ,2 7 7 $ 7 8 5 ,2 2 4 $ 8 5 8 ,0 1 1 $ 1 ,2 2 3 ,9 0 2 f ir st -l in e su p er v is o rs o f n o n -r et ai l sa le s w o rk er s $ 6 5 5 ,5 0 9 $ 7 1 8 ,8 1 2 $ 7 2 8 ,1 5 6 $ 7 4 1 ,5 4 8 $ 7 6 9 ,2 3 5 $ 7 7 2 ,4 8 8 $ 7 9 5 ,7 3 9 $ 1 ,0 4 8 ,2 2 2 $ 1 ,2 2 2 ,2 6 2 $ 1 ,1 6 9 ,0 5 5 $ 1 ,3 1 3 ,4 0 9 c as h ie rs $ 3 6 9 ,7 3 1 $ 3 7 4 ,1 4 3 $ 3 9 7 ,0 9 2 $ 3 8 8 ,3 8 3 $ 3 9 3 ,6 6 3 $ 3 8 2 ,0 2 0 $ 3 8 5 ,7 5 3 $ 4 3 2 ,5 2 7 $ 4 0 5 ,7 5 9 — — c o u n te r an d re n ta l cl er k s $ 4 1 7 ,7 9 0 — $ 4 8 0 ,2 2 6 — $ 5 4 5 ,7 7 3 $ 5 2 8 ,1 5 5 $ 5 7 3 ,9 5 7 $ 6 1 6 ,7 5 2 — — — p ar ts sa le sp er so n s $ 4 9 6 ,0 8 4 $ 5 3 2 ,8 4 3 $ 5 8 4 ,2 9 1 $ 5 6 9 ,8 6 4 $ 5 8 3 ,7 8 5 $ 5 4 5 ,8 2 7 $ 5 6 9 ,0 5 0 $ 5 4 2 ,4 0 3 — — — r et ai l sa le sp er so n s $ 4 4 6 ,7 9 5 $ 4 8 3 ,8 0 2 $ 5 1 5 ,1 7 9 $ 5 1 5 ,4 6 9 $ 5 4 8 ,3 1 9 $ 5 4 5 ,6 8 7 $ 5 4 5 ,5 5 6 $ 6 8 9 ,7 8 1 $ 6 8 0 ,4 7 3 $ 5 0 5 ,3 3 3 $ 5 4 8 ,9 6 6 a d v er ti si n g sa le s ag en ts — — $ 7 1 1 ,3 1 9 — $ 6 6 8 ,9 2 5 $ 7 0 0 ,5 1 6 $ 7 0 5 ,7 7 0 $ 1 ,0 2 3 ,9 3 3 $ 1 ,0 3 3 ,0 3 4 — — in su ra n ce sa le s ag en ts $ 5 8 1 ,0 2 4 — $ 6 0 9 ,7 8 8 $ 6 2 0 ,0 4 9 $ 6 5 2 ,6 9 1 $ 6 3 0 ,2 6 4 $ 6 2 4 ,0 8 2 $ 8 9 6 ,7 5 3 $ 9 5 4 ,1 9 9 $ 1 ,0 0 4 ,2 9 3 — s ec u ri ti es , co m m o d it ie s, an d fi n se rv sa le s ag en ts — — $ 7 1 9 ,4 1 3 $ 7 6 5 ,8 8 8 $ 7 4 3 ,3 6 2 $ 7 5 9 ,4 7 9 $ 7 6 5 ,8 9 4 $ 1 ,3 8 0 ,6 2 5 $ 1 ,8 4 3 ,8 1 2 $ 2 ,0 3 3 ,2 2 6 $ 1 ,9 0 1 ,3 6 1 t ra v el ag en ts — — $ 6 2 3 ,9 0 0 — $ 6 1 6 ,0 2 5 $ 5 6 1 ,3 5 9 $ 5 3 9 ,2 1 2 $ 6 1 2 ,7 4 9 $ 7 8 5 ,7 4 0 — — s al es re p re se n ta ti v es , se rv ic es , al l o th er $ 6 9 8 ,2 3 3 $ 6 1 6 ,5 4 4 $ 7 0 8 ,0 7 4 $ 7 1 8 ,6 0 5 $ 7 6 4 ,3 3 9 $ 7 6 8 ,6 0 4 $ 7 5 8 ,0 2 5 $ 1 ,1 0 3 ,9 3 9 $ 1 ,2 5 6 ,9 2 1 $ 9 7 7 ,6 0 4 — s al es re p re se n ta ti v es , w h sl an d m n fg $ 6 5 6 ,7 7 3 $ 6 7 0 ,1 3 7 $ 7 3 5 ,3 4 6 $ 7 1 4 ,1 2 9 $ 7 6 8 ,3 9 6 $ 7 6 7 ,6 0 5 $ 7 8 0 ,2 4 3 $ 1 ,0 9 5 ,9 1 1 $ 1 ,2 0 7 ,5 9 1 $ 1 ,0 0 9 ,3 1 6 $ 1 ,1 7 7 ,1 9 3 m o d el s, d em o n st ra to rs , an d p ro d u ct p ro m o te rs — — $ 4 5 0 ,1 3 5 — — $ 5 4 2 ,3 8 6 — $ 8 4 8 ,0 9 1 — — — r ea l es ta te b ro k er s an d sa le s ag en ts $ 6 6 5 ,1 0 3 — $ 6 4 2 ,6 3 6 $ 6 1 7 ,2 6 8 $ 6 7 3 ,2 7 6 $ 7 0 0 ,1 3 6 $ 6 8 3 ,6 8 0 $ 9 3 0 ,8 4 2 $ 9 8 9 ,2 5 9 $ 9 0 1 ,4 9 4 $ 9 4 9 ,7 6 7 s al es en g in ee rs — — — — — $ 1 ,2 9 0 ,0 3 7 $ 1 ,0 6 1 ,2 4 1 $ 1 ,3 7 6 ,6 6 7 $ 1 ,4 2 4 ,4 2 3 — — t el em ar k et er s — — $ 4 7 5 ,5 4 9 — $ 4 5 3 ,5 0 3 $ 4 2 2 ,1 1 2 $ 4 8 3 ,5 9 9 $ 5 5 9 ,8 6 3 — — — d o o rto -d o o r sa le s w rk s, n ew s an d st r v en d o rs $ 4 7 4 ,5 6 6 — $ 5 0 3 ,6 3 2 — $ 4 8 6 ,9 0 2 $ 5 4 7 ,1 7 5 $ 5 4 2 ,2 4 0 $ 5 8 2 ,9 6 4 — — — s al es an d re la te d w o rk er s, al l o th er $ 5 2 0 ,1 3 2 — $ 6 0 9 ,6 4 2 $ 7 2 6 ,4 6 1 $ 6 9 4 ,8 1 0 $ 6 8 5 ,7 0 9 $ 6 7 1 ,7 8 0 $ 9 9 4 ,3 3 3 $ 1 ,0 9 8 ,4 9 0 — — i. timmerman, n. volkov / financial services review 28 (2020) 179–200 191 t ab le 2 m ax im u m ed u ca ti o n co st at ag e 1 8 fo r sa le s re la te d o cc u p at io n s < 1 y ea r co ll eg e > = 1 y ea r co ll eg e a ss o ci at es b ac h el o rs m as te rs p ro fe ss io n al d o ct o ra te f ir st -l in e su p er v is o rs o f re ta il sa le s w o rk er s $ 3 2 ,4 9 2 $ 2 4 ,7 0 5 $ 2 2 ,0 4 9 $ 1 3 1 ,1 6 8 $ 1 9 7 ,1 1 5 $ 2 6 9 ,9 0 2 $ 6 3 5 ,7 9 3 f ir st -l in e su p er v is o rs o f n o n -r et ai l sa le s w o rk er s $ 4 1 ,0 7 9 $ 4 4 ,3 3 2 $ 6 7 ,5 8 3 $ 3 2 0 ,0 6 6 $ 4 9 4 ,1 0 6 $ 4 4 0 ,8 9 9 $ 5 8 5 ,2 5 3 c as h ie rs ($ 3 ,4 2 9 ) ($ 1 5 ,0 7 2 ) ($ 1 1 ,3 3 9 ) $ 3 5 ,4 3 5 $ 8 ,6 6 8 — — c o u n te r an d re n ta l cl er k s $ 6 5 ,5 4 7 $ 4 7 ,9 2 9 $ 9 3 ,7 3 0 $ 1 3 6 ,5 2 5 — — — p ar ts sa le sp er so n s ($ 5 0 5 ) ($ 3 8 ,4 6 4 ) ($ 1 5 ,2 4 0 ) ($ 4 1 ,8 8 8 ) — — — r et ai l sa le sp er so n s $ 3 3 ,1 4 0 $ 3 0 ,5 0 8 $ 3 0 ,3 7 8 $ 1 7 4 ,6 0 2 $ 1 6 5 ,2 9 4 ($ 9 ,8 4 6 ) $ 3 3 ,7 8 7 a d v er ti si n g sa le s ag en ts ($ 4 2 ,3 9 3 ) ($ 1 0 ,8 0 3 ) ($ 5 ,5 4 9 ) $ 3 1 2 ,6 1 4 $ 3 2 1 ,7 1 5 — — in su ra n ce sa le s ag en ts $ 4 2 ,9 0 3 $ 2 0 ,4 7 6 $ 1 4 ,2 9 4 $ 2 8 6 ,9 6 4 $ 3 4 4 ,4 1 1 $ 3 9 4 ,5 0 4 — s ec u ri ti es , co m m o d it ie s, an d fi n se rv sa le s ag en ts $ 2 3 ,9 4 8 $ 4 0 ,0 6 5 $ 4 6 ,4 8 1 $ 6 6 1 ,2 1 2 $ 1 ,1 2 4 ,3 9 8 $ 1 ,3 1 3 ,8 1 2 $ 1 ,1 8 1 ,9 4 8 t ra v el ag en ts ($ 7 ,8 7 5 ) ($ 6 2 ,5 4 0 ) ($ 8 4 ,6 8 8 ) ($ 1 1 ,1 5 0 ) $ 1 6 1 ,8 4 0 — — s al es re p re se n ta ti v es , se rv ic es , al l o th er $ 5 6 ,2 6 5 $ 6 0 ,5 3 0 $ 4 9 ,9 5 1 $ 3 9 5 ,8 6 5 $ 5 4 8 ,8 4 7 $ 2 6 9 ,5 3 0 — s al es re p re se n ta ti v es , w h sl an d m n fg $ 3 3 ,0 5 1 $ 3 2 ,2 5 9 $ 4 4 ,8 9 7 $ 3 6 0 ,5 6 5 $ 4 7 2 ,2 4 5 $ 2 7 3 ,9 7 0 $ 4 4 1 ,8 4 8 m o d el s, d em o n st ra to rs , an d p ro d u ct p ro m o te rs — $ 9 2 ,2 5 1 — $ 3 9 7 ,9 5 6 — — — r ea l es ta te b ro k er s an d sa le s ag en ts $ 3 0 ,6 4 1 $ 5 7 ,5 0 0 $ 4 1 ,0 4 5 $ 2 8 8 ,2 0 6 $ 3 4 6 ,6 2 3 $ 2 5 8 ,8 5 9 $ 3 0 7 ,1 3 1 s al es en g in ee rs — — — — — — — t el em ar k et er s ($ 2 2 ,0 4 6 ) ($ 5 3 ,4 3 7 ) $ 8 ,0 4 9 $ 8 4 ,3 1 3 — — — d o o rto -d o o r sa le s w rk s, n ew s an d st r v en d o rs ($ 1 6 ,7 3 0 ) $ 4 3 ,5 4 3 $ 3 8 ,6 0 8 $ 7 9 ,3 3 2 — — — s al es an d re la te d w o rk er s, al l o th er $ 8 5 ,1 6 9 $ 7 6 ,0 6 7 $ 6 2 ,1 3 9 $ 3 8 4 ,6 9 2 $ 4 8 8 ,8 4 9 — — 192 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 skills beyond a basic level. as a result, they should not be rewarded for the extra level of education pursued and the cost incurred for obtaining such education. by comparison, insurance agents can spend up to $286,964 on bachelor’s degrees and nearly additional $57,000 on their master’s degrees to be better off than insurance agents with only a high school diploma. securities, commodities, and financial services sales agents can spend up to $1,124,398 on their bachelor’s and master’s degree and still be financially better off than someone who decided not to go to college. the master’s degree appears very beneficial in this position as it adds over $450,000 in value. a professional degree adds almost another $200,000 in value to an individual involved in this occupation, while a ph.d. degree appears to be less valuable than a master’s degree in present value terms. as expected, the more specialized the skills required, the more the pursuit of additional education pays off. the appeal of presenting the information this way is that is gives consumers a way to evaluate whether pursuing a specific level of education or major is worth the overall investment. 5.2. business and financial occupations table 3 displays the present value of the future earnings for all business-related occupations as defined by soc. similar to the previous tables, table 3 presents the present value of future earnings for all business-related occupations using a five percentage real discount rate. as with the sales related occupations, there is a noticeable homogeneity of values across individuals with less than a high school diploma. there is also a very noticeable jump in the present value of earnings at the bachelor’s degree level for all occupations without exceptions. the marginal benefit of a degree beyond a bachelor’s degree though varies. as such, agents and business managers of artists, claims adjusters, event planners, appraisers and financial examiners appear to be better off by earning only a bachelor’s degree and not considering a master’s degree. on the contrary, logisticians, credit analysts, tax preparers, and market research analysts appear to benefit the most from a master’s degree. furthermore, when looking at the human resource workers, a marginal benefit from pursuing a bachelor’s degree is observed to the magnitude of about $164,000 in present value terms over one’s working lifetime. a master’s degree brings, on average, an additional $63,500 above and beyond a bachelor’s degree (the difference in the present value of the bachelor’s and master’s level lifetime earnings). by comparing these numbers to those reported in table 4, one can see the maximum amount that the level of education beyond high school is worth on average. an individual who wants to become a human resource worker should not spend more than $164,000 on his bachelor’s degree. his maximum expenditure on both the bachelor’s and the master’s degree should not exceed $228,000. it is relatively safe to conclude that finding educational programs to fit these costs is relatively simple, and most people would be better off by pursuing a higher degree in the field of human resources. by comparison, when looking at the lifelong earnings for meeting and convention or event planners, we conclude the opposite. on average, an event planner is better off finishing a bachelor’s degree (as long as the total cost does not exceed $78,000) and not pursuing further education since a maximum combined expenditure on both a bachelor’s and a master’s i. timmerman, n. volkov / financial services review 28 (2020) 179–200 193 t ab le 3 b u si n es s re la te d o cc u p at io n s p re se n t v al u e o f th e ag eea rn in g s p ro fi le s 0 – 1 2 1 2 th n o d ip lo m a h s g e d < 1 y ea r co ll eg e > 5 1 y ea r co ll eg e a ss o ci at es b ac h el o rs m as te rs p ro fe ss io n al d o ct o ra te a g en ts an d b u s m n g rs o f ar ti st s, p er fo rm er s, an d at h le te s — — $ 6 9 6 ,6 6 7 — — $ 6 8 2 ,4 3 2 — $ 9 5 9 ,9 1 1 $ 8 7 7 ,6 6 5 — — b u y er s an d p u rc h as in g ag en ts , fa rm p ro d u ct s — — $ 6 5 7 ,2 1 8 — — — — $ 9 8 4 ,8 3 0 — — — w h o le sa le an d re ta il b u y er s, ex ce p t fa rm p ro d u ct s $ 4 9 5 ,7 0 0 $ 5 4 4 ,9 0 9 $ 6 2 5 ,2 6 6 $ 6 9 4 ,3 5 0 $ 6 3 8 ,1 2 8 $ 5 9 1 ,2 0 5 $ 6 0 9 ,2 5 5 $ 7 9 8 ,1 7 5 $ 8 0 8 ,3 0 7 — — p u rc h ag en ts , ex ce p t w h ls , re ta il , an d fa rm p ro d u ct s $ 7 2 6 ,0 1 2 — $ 6 8 9 ,4 5 2 — $ 6 7 2 ,3 6 6 $ 6 8 7 ,5 8 0 $ 7 2 3 ,7 4 4 $ 8 2 8 ,5 7 7 $ 9 3 5 ,2 5 2 $ 8 6 1 ,3 4 2 — c la im s ad ju st er s, ap p ra is er s, ex am in er s, an d in v es ti g at o rs — — $ 6 8 4 ,7 5 7 $ 6 4 9 ,7 6 6 $ 7 0 1 ,0 3 2 $ 6 7 0 ,2 2 9 $ 6 9 3 ,6 6 1 $ 8 0 9 ,1 9 1 $ 7 9 8 ,2 1 4 $ 8 2 9 ,4 9 4 — c o m p li an ce o ffi ce rs — — $ 7 6 0 ,6 4 5 $ 8 2 0 ,8 3 7 $ 6 7 8 ,8 7 3 $ 7 6 6 ,0 9 7 $ 7 3 3 ,5 1 5 $ 9 1 5 ,7 2 2 $ 9 9 8 ,4 1 3 $ 1 ,1 3 1 ,9 8 7 $ 1 ,0 8 7 ,3 7 4 c o st es ti m at o rs $ 8 2 7 ,2 7 6 — $ 8 3 1 ,9 3 4 $ 7 3 1 ,9 0 4 $ 8 0 8 ,0 9 2 $ 8 2 0 ,4 7 5 $ 7 8 4 ,4 3 6 $ 9 4 3 ,3 1 3 $ 9 7 1 ,8 3 0 — — h u m an re so u rc es w o rk er s $ 7 0 1 ,3 0 3 $ 7 7 6 ,5 6 1 $ 7 1 6 ,0 6 4 $ 7 4 5 ,8 3 7 $ 7 3 7 ,3 6 8 $ 7 3 9 ,4 5 1 $ 7 1 2 ,7 1 4 $ 8 8 0 ,5 7 3 $ 9 4 4 ,0 6 5 $ 9 5 7 ,3 5 4 $ 1 ,0 3 9 ,1 1 0 c o m p en sa ti o n , b en efi ts , an d jo b an al y si s sp ec ia li st s — — $ 7 1 4 ,3 9 0 — $ 6 6 4 ,8 8 8 $ 6 6 4 ,5 7 9 $ 6 6 2 ,9 8 6 $ 7 7 2 ,6 3 4 $ 8 7 7 ,7 5 6 — — t ra in in g an d d ev el o p m en t sp ec ia li st s — — $ 7 2 8 ,6 3 6 $ 6 1 1 ,9 3 6 $ 7 4 0 ,1 2 2 $ 7 2 5 ,1 9 3 $ 7 4 0 ,3 7 8 $ 8 1 2 ,8 5 6 $ 8 8 5 ,4 4 9 — — l o g is ti ci an s — — $ 7 0 5 ,2 0 0 $ 5 9 1 ,7 6 1 $ 6 8 6 ,4 4 1 $ 7 3 3 ,1 0 9 $ 7 3 9 ,1 0 7 $ 8 5 7 ,9 2 3 $ 1 ,0 2 8 ,8 7 8 — — m an ag em en t an al y st s $ 9 4 1 ,4 2 6 — $ 8 7 7 ,9 0 1 $ 1 ,0 5 0 ,8 5 9 $ 8 9 1 ,4 9 2 $ 8 7 0 ,2 6 2 $ 8 4 8 ,9 1 6 $ 1 ,1 4 9 ,7 9 3 $ 1 ,2 7 8 ,4 2 3 $ 1 ,2 2 1 ,6 7 4 $ 1 ,2 5 7 ,6 5 8 m ee ti n g , co n v en ti o n , an d ev en t p la n n er s — — $ 6 7 9 ,6 0 7 — $ 6 7 3 ,1 4 4 $ 6 9 5 ,7 0 5 $ 6 4 6 ,1 9 8 $ 7 5 8 ,4 6 4 $ 7 0 0 ,7 4 2 — — f u n d ra is er s — — — — — $ 6 6 1 ,7 4 0 $ 6 7 1 ,0 9 9 $ 8 6 8 ,5 4 0 $ 8 8 6 ,6 3 0 $ 8 8 3 ,9 0 4 — m ar k et re se ar ch an al y st s an d m ar k et in g sp ec ia li st s — — $ 7 7 7 ,0 3 5 — $ 7 2 8 ,2 7 7 $ 7 9 8 ,8 6 1 $ 8 0 1 ,9 4 1 $ 1 ,0 4 7 ,6 9 9 $ 1 ,2 2 1 ,1 0 9 — — b u si n es s o p er at io n s sp ec ia li st s, al l o th er $ 7 1 2 ,4 5 0 — $ 6 9 7 ,1 1 7 $ 6 5 1 ,0 9 4 $ 6 8 8 ,4 5 6 $ 7 1 6 ,6 8 4 $ 6 9 7 ,7 7 1 $ 8 8 5 ,6 0 2 $ 1 ,0 2 3 ,4 5 3 $ 1 ,0 2 5 ,6 3 9 $ 1 ,1 0 2 ,0 3 4 (c o n ti n u ed o n n ex t p a g e) 194 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 t ab le 3 (c o n ti n u ed ) 0 – 1 2 1 2 th n o d ip lo m a h s g e d < 1 y ea r co ll eg e > 5 1 y ea r co ll eg e a ss o ci at es b ac h el o rs m as te rs p ro fe ss io n al d o ct o ra te a cc o u n ta n ts an d au d it o rs — — $ 6 4 6 ,5 5 0 $ 6 7 7 ,3 2 7 $ 6 4 1 ,6 4 3 $ 6 4 7 ,9 0 7 $ 6 3 7 ,7 9 2 $ 9 1 2 ,7 4 5 $ 1 ,0 4 3 ,8 5 1 $ 9 8 0 ,9 1 5 $ 8 6 7 ,0 0 8 a p p ra is er s an d as se ss o rs o f re al es ta te — — $ 6 0 0 ,7 8 0 — $ 7 9 3 ,5 6 6 $ 6 6 0 ,1 8 7 $ 6 3 9 ,9 3 6 $ 8 3 3 ,9 4 8 $ 8 2 0 ,0 1 4 — — b u d g et an al y st s — — $ 9 7 6 ,2 0 8 — $ 9 6 4 ,7 2 1 $ 8 8 0 ,0 6 7 $ 8 6 0 ,4 2 3 $ 9 9 9 ,6 0 9 $ 1 ,0 5 5 ,7 2 8 — — c re d it an al y st s — — $ 6 5 4 ,1 2 9 — $ 7 2 6 ,4 7 6 $ 7 0 7 ,0 0 7 $ 8 2 9 ,2 5 0 $ 1 ,0 3 0 ,9 6 5 — — f in an ci al an al y st s — — $ 1 ,0 2 0 ,4 2 5 — $ 8 3 6 ,9 5 3 $ 8 3 5 ,6 2 1 $ 8 3 9 ,2 6 3 $ 1 ,1 6 5 ,9 5 2 $ 1 ,3 0 0 ,2 1 9 $ 1 ,4 5 5 ,5 5 1 $ 1 ,8 0 4 ,0 3 3 p er so n al fi n an ci al ad v is o rs — — $ 7 4 5 ,8 7 0 — $ 7 4 3 ,4 9 1 $ 8 5 9 ,1 1 9 $ 8 1 2 ,9 3 7 $ 1 ,2 6 2 ,4 2 1 $ 1 ,3 6 4 ,0 8 1 $ 1 ,4 6 8 ,0 0 4 $ 1 ,4 0 1 ,7 2 9 in su ra n ce u n d er w ri te rs — — $ 7 4 6 ,8 6 6 — $ 7 2 9 ,8 6 7 $ 7 3 4 ,3 4 9 $ 7 2 3 ,7 9 7 $ 9 5 0 ,7 0 9 $ 9 8 6 ,8 6 9 — — f in an ci al ex am in er s — — — — — — — $ 1 ,1 8 6 ,3 0 0 $ 1 ,1 3 2 ,0 3 9 — — c re d it co u n se lo rs an d lo an o ffi ce rs $ 7 2 1 ,2 0 0 — $ 7 0 0 ,1 0 5 $ 7 1 6 ,8 2 9 $ 7 2 3 ,4 7 4 $ 7 1 3 ,8 9 1 $ 7 0 1 ,6 0 6 $ 8 9 2 ,5 7 7 $ 1 ,0 1 6 ,1 8 9 $ 8 7 6 ,8 0 2 — t ax ex am in er s an d co l le ct o rs , an d re v en u e ag en ts — — $ 6 6 9 ,3 9 1 -— $ 6 4 9 ,8 7 2 $ 6 4 6 ,6 1 7 $ 6 5 8 ,0 0 1 $ 7 9 6 ,2 7 7 $ 8 7 0 ,2 6 9 — — t ax p re p ar er s — — $ 6 4 0 ,1 6 8 — $ 6 4 3 ,8 9 7 $ 6 3 1 ,8 0 8 $ 5 5 8 ,9 6 9 $ 8 6 5 ,3 0 2 $ 1 ,2 7 8 ,5 4 5 $ 1 ,5 0 6 ,9 0 3 — f in an ci al sp ec ia li st s, al l o th er — — $ 7 1 5 ,9 3 7 — $ 6 6 7 ,5 3 6 $ 6 7 6 ,1 4 2 $ 6 3 1 ,9 8 2 $ 9 1 6 ,8 1 4 $ 1 ,2 1 8 ,7 8 8 — — i. timmerman, n. volkov / financial services review 28 (2020) 179–200 195 t ab le 4 m ax im u m ed u ca ti o n co st at ag e 1 8 fo r b u si n es s re la te d o cc u p at io n s < 1 y ea r co ll eg e > 5 1 y ea r co ll eg e a ss o ci at es b ac h el o rs m as te rs p ro fe ss io n al d o ct o ra te a g en ts an d b u s m n g rs o f ar ti st s, p er fo rm er s, an d at h le te s — ($ 1 4 ,2 3 5 ) — $ 2 6 3 ,2 4 4 $ 1 8 0 ,9 9 7 — — b u y er s an d p u rc h as in g ag en ts , fa rm p ro d u ct s — — — $ 3 2 7 ,6 1 2 — — — w h o le sa le an d re ta il b u y er s, ex ce p t fa rm p ro d u ct s $ 1 2 ,8 6 2 ($ 3 4 ,0 6 0 ) ($ 1 6 ,0 1 1 ) $ 1 7 2 ,9 0 9 $ 1 8 3 ,0 4 1 — — p u rc h ag en ts , ex ce p t w h ls , re ta il , an d fa rm p ro d u ct s ($ 1 7 ,0 8 6 ) ($ 1 ,8 7 2 ) $ 3 4 ,2 9 3 $ 1 3 9 ,1 2 5 $ 2 4 5 ,8 0 0 $ 1 7 1 ,8 9 0 — c la im s ad ju st er s, ap p ra is er s, ex am in er s, an d in v es ti g at o rs $ 1 6 ,2 7 5 ($ 1 4 ,5 2 8 ) $ 8 ,9 0 4 $ 1 2 4 ,4 3 4 $ 1 1 3 ,4 5 6 $ 1 4 4 ,7 3 6 — c o m p li an ce o ffi ce rs ($ 8 1 ,7 7 2 ) $ 5 ,4 5 2 ($ 2 7 ,1 3 0 ) $ 1 5 5 ,0 7 7 $ 2 3 7 ,7 6 8 $ 3 7 1 ,3 4 2 $ 3 2 6 ,7 2 9 c o st es ti m at o rs ($ 2 3 ,8 4 2 ) ($ 1 1 ,4 5 9 ) ($ 4 7 ,4 9 8 ) $ 1 1 1 ,3 7 9 $ 1 3 9 ,8 9 6 — — h u m an re so u rc es w o rk er s $ 2 1 ,3 0 4 $ 2 3 ,3 8 7 ($ 3 ,3 5 0 ) $ 1 6 4 ,5 0 9 $ 2 2 8 ,0 0 1 $ 2 4 1 ,2 9 0 $ 3 2 3 ,0 4 6 c o m p en sa ti o n , b en efi ts , an d jo b an al y si s sp ec ia li st s ($ 4 9 ,5 0 2 ) ($ 4 9 ,8 1 1 ) ($ 5 1 ,4 0 3 ) $ 5 8 ,2 4 4 $ 1 6 3 ,3 6 7 — — t ra in in g an d d ev el o p m en t sp ec ia li st s $ 1 1 ,4 8 6 ($ 3 ,4 4 3 ) $ 1 1 ,7 4 2 $ 8 4 ,2 2 0 $ 1 5 6 ,8 1 3 — — l o g is ti ci an s ($ 1 8 ,7 6 0 ) $ 2 7 ,9 0 9 $ 3 3 ,9 0 7 $ 1 5 2 ,7 2 3 $ 3 2 3 ,6 7 8 — — m an ag em en t an al y st s $ 1 3 ,5 9 1 ($ 7 ,6 4 0 ) ($ 2 8 ,9 8 6 ) $ 2 7 1 ,8 9 2 $ 4 0 0 ,5 2 1 $ 3 4 3 ,7 7 3 $ 3 7 9 ,7 5 7 m ee ti n g , co n v en ti o n , an d ev en t p la n n er s ($ 6 ,4 6 3 ) $ 1 6 ,0 9 9 ($ 3 3 ,4 0 9 ) $ 7 8 ,8 5 8 $ 2 1 ,1 3 6 — — f u n d ra is er s — — — — — — — m ar k et re se ar ch an al y st s an d m ar k et in g sp ec ia li st s ($ 4 8 ,7 5 8 ) $ 2 1 ,8 2 6 $ 2 4 ,9 0 6 $ 2 7 0 ,6 6 4 $ 4 4 4 ,0 7 4 — — b u si n es s o p er at io n s sp ec ia li st s, al l o th er ($ 8 ,6 6 1 ) $ 1 9 ,5 6 7 $ 6 5 4 $ 1 8 8 ,4 8 5 $ 3 2 6 ,3 3 6 $ 3 2 8 ,5 2 1 $ 4 0 4 ,9 1 7 a cc o u n ta n ts an d au d it o rs ($ 4 ,9 0 7 ) $ 1 ,3 5 7 ($ 8 ,7 5 8 ) $ 2 6 6 ,1 9 5 $ 3 9 7 ,3 0 2 $ 3 3 4 ,3 6 5 $ 2 2 0 ,4 5 8 a p p ra is er s an d as se ss o rs o f re al es ta te $ 1 9 2 ,7 8 6 $ 5 9 ,4 0 7 $ 3 9 ,1 5 6 $ 2 3 3 ,1 6 8 $ 2 1 9 ,2 3 4 — — b u d g et an al y st s ($ 1 1 ,4 8 7 ) ($ 9 6 ,1 4 1 ) ($ 1 1 5 ,7 8 5 ) $ 2 3 ,4 0 1 $ 7 9 ,5 2 0 — — c re d it an al y st s — $ 7 2 ,3 4 7 $ 5 2 ,8 7 8 $ 1 7 5 ,1 2 1 $ 3 7 6 ,8 3 5 — — f in an ci al an al y st s ($ 1 8 3 ,4 7 2 ) ($ 1 8 4 ,8 0 4 ) ($ 1 8 1 ,1 6 2 ) $ 1 4 5 ,5 2 8 $ 2 7 9 ,7 9 4 $ 4 3 5 ,1 2 7 $ 7 8 3 ,6 0 8 p er so n al fi n an ci al ad v is o rs ($ 2 ,3 7 9 ) $ 1 1 3 ,2 5 0 $ 6 7 ,0 6 7 $ 5 1 6 ,5 5 1 $ 6 1 8 ,2 1 1 $ 7 2 2 ,1 3 4 $ 6 5 5 ,8 6 0 in su ra n ce u n d er w ri te rs ($ 1 6 ,9 9 9 ) ($ 1 2 ,5 1 7 ) ($ 2 3 ,0 6 9 ) $ 2 0 3 ,8 4 3 $ 2 4 0 ,0 0 3 — — f in an ci al ex am in er s — — — — — — — c re d it co u n se lo rs an d lo an o ffi ce rs $ 2 3 ,3 6 9 $ 1 3 ,7 8 5 $ 1 ,5 0 1 $ 1 9 2 ,4 7 1 $ 3 1 6 ,0 8 3 $ 1 7 6 ,6 9 6 — t ax ex am in er s an d co ll ec to rs , an d re v en u e ag en ts ($ 1 9 ,5 1 8 ) ($ 2 2 ,7 7 4 ) ($ 1 1 ,3 9 0 ) $ 1 2 6 ,8 8 6 $ 2 0 0 ,8 7 8 — — t ax p re p ar er s $ 3 ,7 2 9 ($ 8 ,3 6 0 ) ($ 8 1 ,1 9 9 ) $ 2 2 5 ,1 3 5 $ 6 3 8 ,3 7 7 $ 8 6 6 ,7 3 5 — f in an ci al sp ec ia li st s, al l o th er ($ 4 8 ,4 0 1 ) ($ 3 9 ,7 9 5 ) ($ 8 3 ,9 5 5 ) $ 2 0 0 ,8 7 7 $ 5 0 2 ,8 5 2 — — 196 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 degree that is sound from a financial standpoint is only $21,000. similarly, several other occupations see diminishing returns from obtaining professional and doctorate degrees. a very noticeable and important observation that relates to the analysis of both sales and business-related occupational groups is that dropping out of college at any time or getting an associate degree is not a value adding proposition for any of the occupations examined. 6. implications, limitations, and conclusion the different estimates in lifetime earnings given a specific education will most likely change as real wages change/increase for future workers (morgan and david, 1963). it is also possible that the differential will change as more/less students decide to attend college. however, as miller (1960) points out in a very early study, despite the increased rate of college graduates the differential effect persists. newer research confirms this inference. another assumption of examining lifelong earnings is that the benefit derived from a college education will likely be similar for the future generations as it was for the people who are already working. some articles in popular press argue that the gap in pay in the united states has been decreasing and education will no longer have an advantage at one point in the future. these declines are normally juxtaposed against the increased cost and increased student loan debt of college graduates. as mentioned earlier, when the economy is strong the economic advantage of advanced education tends to be lower. the analysis needs to be performed over the entire career span to be meaningful. in this article, we combine the benefits of additional schooling and specific career paths available. this is a valuable tool for financial planners who aim to work with younger, career-changing, or education-pursuing clients. they can now frame the decision of obtaining more education in a more specific career path by analyzing the benefits of any additional degree or career option. by applying the theoretical model to specific occupations and levels of education, financial planners can talk to their clients about the lifetime benefit of obtaining such education and the maximum costs to obtain a degree. as mentioned in the introduction, we understand the limitations of applying a theoretical model for a practitioner, and, as such, our next step is to develop a practical tool that financial planners can use in their work with clients. in the interim, the results from our study can be used to help frame the education planning conversations that financial planners have with their clients. specifically, the results can be used to enhance financial planner and client dialog around the value of higher education and the financial outcomes associated with investing in a particular degree program. notes 1 stiglitz (2015) provides the following definitions of wealth and capital: “wealth and capital are two distinct concepts; the former reflects control over resources, the latter is a key input into production processes.” while we acknowledge the subtle differences in the definitions, we often use the words interchangeably as human capital i. timmerman, n. volkov / financial services review 28 (2020) 179–200 197 element of overall individual’s wealth represents both control over resources and an input into the production process. 2 the data is based on a random sample of 3,824 us households and a supplemental sample of 438 higher income households that are more representative of the upper tail of wealth distribution. 3 see, for example, https://thefinancialbrand.com/71459/millennial-wealth-managementbanking-digital-cx-trends/, https://www.financial-planning.com/opinion/the-richestadvisors-are-targeting-millennials-now or https://www.thinkadvisor.com/2017/07/18/ these-millennial-advisors-are-killing-it-with-youn/?slreturn=20181013183259. 4 the center on education and the workforce at georgetown university has recently developed a ranking system of colleges in the united states by the return on investment in education (see https://cew.georgetown.edu/cew-reports/collegeroi/). while such a tool can certainly help in selecting a college to attend, it does not provide insight into the career and appropriate (from a financial standpoint) education level selection. 5 to further explore age-earnings profiles for different education-occupation combinations, one can refer to www.alignme.app. 6 based on the data provided by the bureau of labor and statistics (www.bls.gov), there is a consistent inverse relation between the level of education and the unemployment rates in the united states. this inverse relation spans back for decades. 7 “education and unemployment,” washington post, february 27, 2012 (www. washingtonpost.com/politics/education-and-unemployment/2012/02/27/giqarnmzer_ graphic.html); and caralee adams, “new study tracks lifetime income based on college major,” college bound(education week blog), march 24, 2011 (http:// blogs.edweek.org/edweek/college_bound/2011/05/new_study_tracks_lifetime_ income_based_on_college_major.html?qs=lifetime+college+earning 8 while the average life expectancy of the u.s. population dropped ever so slightly over the last three years (mainly due to increasing suicide rates and the opioid consumption), the general longevity trend is upward. 9 note that the model ignores personal consumption and assumes that there are no social welfare programs that, in some circumstances, may substitute or outweigh the value of individual’s human capital. 10 the only exceptions are teachers and librarians, who are assumed to be working full time if they indicate that they worked 35 or more weeks in the last 12 months. 11 an update to the budget and economic outlook: 2018 to 2028, table a-1, the congress of the united states, congressional budget office, august 2018. http:// www.cbo.gov/ 12 one can run an unlimited number of age-earnings profiles by going to www. alignme.app. 13 while seemingly arbitrary, the choice of a five percentage real rate of return is predicated by the fact that it roughly proxies for the investment returns in a diversified market portfolio, which, depending on the methodology and the time frame used is shown to generate a return that is about 5 to 7 percent above the rate of inflation. 198 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 14 note that the calculations presented below assume that all individuals retire the age of 67, which is the age of full eligibility for social security benefits. thus, the general finding of a positive relation between the level of education and the work life expectancy is omitted in this calculation. if the work life expectancy differentials were to be incorporated into the calculations, the positive effects of additional education on the present value of the aeps would be even more pronounced. 15 we require a minimum of 75 observations to run an aep using our quintile regression method. references bajtelsmit, v. l., & wang, t. (2018). household financial planning strategies for managing longevity risk. financial planning review, 1, e1007. bernardo, l., meller, p., & valdés, g. (2017). life-cycle valuation of different university majors. case study of chile. interciencia, 42, 380. bridges, k., de’armond, d. a., & dean, l. r. (2013). importance of human capital in recovery from divorce for women. journal of financial planning, 26, 40-47. britt-lutter, s., moore, c., gllardo, j., & scott, a. (2018). appling human capital framework to college student’s financial well-being. the association for financial counseling and planning education proceedings, pp. 6677 (available at http://afcpe. org/uploads/symposium/symposium_2018/2018_afcpe_proceedings_final. pdf). brown, j., fang, c., & gomes, f. (2012). risk and return on education. nber working paper no.18300. (available at https://www.nber.org/papers/w18300). budria rodriguez, s., diaz-gimenez, j., quadrini, v., & rios-rull, j. (2002). updated facts on the us distributions of earnings, income and wealth. federal reserve bank of minneapolis quarterly review, 26, 2-35. caballe, j., & santos, m. (1993). on endogenous growth with physical and human capital. journal of political economy, 101, 1042-1067. chakrabarti, r., jiang, m., & nober, w. (2019). did the value of a college degree decline during the great recession? federal reserve bank of new york liberty street economics. (available at https:// libertystreeteconomics.newyorkfed.org/2019/07/did-the-value-of-a-college-degree-decline-during-the-greatrecession.html). champagne, j., & kurmann, a. (2013). the great increase in relative wage volatility in the united states. journal of monetary economics, 60, 166-183. chung, l. (2006). education and earnings. perspectives on labor and income, 7, 28-35. davies, s., & guppy, n. (1997). fields of study, college selectivity, and student inequalities in higher education. social forces, 75, 1417-1438. dıaz-gimenez, j., quadrini, v., & rios-rull, j.-v. (1997). dimensions of inequality: facts on the u.s. distribution of earnings, income and wealth. quarterly review, 21, 3-21. hanna, s., gutter, m., & fan, j. (2001). a measure of risk tolerance based on economic theory. journal of financial counseling and planning, 12, 53-60. haubrich, m. (2013). career asset management: where wealth is created. journal of financial planning, 26, 22-24. kyrychenko, v. (2008). optimal asset allocation in the presence of nonfinancial assets. financial services review, 17, 69-86. lee, h.-k., & hanna, s. (1995). investment portfolios and human wealth. journal of financial counseling and planning, 26, 129-152. leonhardt, d. (may 27, 2014). is college worth it? clearly, new data say. the new york times. (available from https://www.nytimes.com/2014/05/27/upshot/is-college-worth-it-clearly-new-data-say.html) markowitz, h. m. (1952). portfolio selection. the journal of finance, 7, 77-91. i. timmerman, n. volkov / financial services review 28 (2020) 179–200 199 miller, h. p. (1960). annual and lifetime income in relation to education 1939–59, american economic review, 50, 962-986. morgan, j., & david, m. (1963). education and income. the quarterly journal of economics, 77, 423-437. mun, h. t., & wong, c. s. (1999). rates of return to education in singapore. education economics, 7, 235-252. oreopoulos, p., & petronijevic, u. (2013). making college worth it: a review of the returns to higher education. the future of children, 23, 41-65. pew research center. (2011). is college worth it? (available at http://www.pewsocialtrends.org/2011/05/15/iscollege-worth-it/) riley, j. (1976). information, screening and human capital. the american economic review, 66, 254-260. roksa, j., & levey, t. (2010). what can you do with that degree? college major and occupational status of college graduates over time. social forces, 89, 389-415. schiemann, w. (2006). people equity: a new paradigm for measuring and managing human capital. human resource planning, 29, 1-10. sharpe, w. f. (1966). mutual fund performance. the journal of business, 39, 119-138. skoog, g., ciecka, j., & krueger, k. (2011). the markov process model of labor force activity: extended tables of central tendency, shape, percentile points, and bootstrap standard errors. journal of forensic economics, 22, 165-229. skoog, g., ciecka, j., & krueger, k. (2019). the markov process model of labor force activity 2012-2017: extended tables of central tendency, shape, percentile points, and bootstrap standard errors. journal of forensic economics, 28, 15-108. stiglitz, j. (2015, summer). inequality, wealth, and capital. new york, ny: queries, columbia business school. taylor, p., parker, k., fry, r., cohn, d., wang, w., velasco, g., & dockterman, d. (2011). is college worth it? pew social and demographic trends. washington, dc. (available at https://www.pewresearch.org/wpcontent/uploads/sites/3/2011/05/higher-ed-report.pdf) 200 i. timmerman, n. volkov / financial services review 28 (2020) 179–200 pii: 1057-0810(94)90015-9 from the editor as baby boomers gradually replace their parents as members of the retired population, the ratio of retired to working persons will increase. and when we add to this ever increasing longevity, something has to give. americans must either retire later, or retire with a diminished proportion of aggregate personal income, or both. changes in both real after-tax social security benefits and qualified pension plans reflect this new reality. the net value of social security retirement benefits has been diminished by gradual increases in the age needed for full benefits, increased taxation of payments, and increased premiums on workers. the institution of two-tier systems by many employers has affected qualified pension plans by lowering future benefits for today’s younger workers. in addition, there has been a trend to replace generous defined benefit plans with less generous defined contribution plans, which also shifts investment risk from the employer to the employee. the net impact of these changes is to increase the burden on the individual to supplement his or her retirement income with personal savings and investments. the burden is twofold: to achieve a comfortable retirement, most people must begin voluntary retire ment savings at an early age and continue to fund them throughout the life cycle. this is not easy for those with young families or children in college. the second part of the burden is to invest the funds intelligently. while much of the advertising aimed at retirement dollars focuses on tactical decisions relating to things such as asset timing or sector rotation, the more basic and more difficult strategic decisions are often ignored. yet it is only the strategic decisions that can be supported by accepted methods of financial analysis. in this issue of financial services review, we present four sophisticated papers which give sound advice to those who will make retirement investment decisions. the first, “an optimization model for scheduling withdrawals from tax-deferred retirement accounts” by cliff ragsdale, andrew seila, and philip little, addresses the very real problems faced by those who have saved for their retirements through tax-deferred accounts. while some professional advisors use various “rules of thumb” for making withdrawals, the problems are complicated by five important variables. 1) if you begin withdrawal too early, you must pay a premature distributions tax; 2) if you take out too little (after age 70 $5) you must pay a minimum distribution tax; 3) if you take out too much in any year (over $150,000), you are subject to the excess distributions tax; and 4) if you leave a gross estate of more than $600,000, it is subject to estate taxes. to further complicate the matter, 5) you may have two or more tax-deferred retirement accounts from multiple jobs or from a combination of 401(k)s, 403(b)s, iras and other qualified plans. “rules of thumb” will just not cut it in an analysis of this complexity. to help retirees and their advisors make these complex decisions, the authors present a mathematical programming model. while the model itself is necessarily complex, it can v vi financial services review, 3(2) 1994 be used with the ubiquitous personal computer and its complexity is transparent to the user who needs only answer a few simple questions. the second paper addresses the basic question of asset allocation between risk and risk-free assets for the retiree. entitled “asset allocation, life expectancy and shortfall,” by kwok ho, moshe arye milevsky, and chris robinson, this academy of financial services award-winning paper shows the asset allocation that will minimize the probability that a retired person will be unable to consume at a certain desired level over his or her expected lifetime. the paper exemplifies the financial tools that are available to assist in true strategic asset allocation. the third paper, entitled ‘the impact of mutual fund distributions on after-tax returns” by william randolph was particularly interesting to me. the paper focuses on the often-large effect of mutual fund portfolio turnover on after-tax returns. since realized capital gains of mutual funds are distributed to fund owners, if funds are held in non-tax-sheltered form, funds with high turnover rates impose a high tax burden on owners. this becomes particularly significant in the case of growth funds which have a huge proportion of expected return in the form of capital gains. it implies that growth funds held in non-tax-sheltered form may be selected on the basis of low turnover to minimize the effect of taxation. the paper was especially significant to me since i was about to publish a basic financial planning textbook in which i implied that non-deductible iras offered few advantages to the investor. clearly i had focused my analysis on fixed income assets and ignored the beneficial effects of holding growth funds with high turnover rates in non-deductible but tax-deferred form. this error was cleared up before shipping the manuscript to the publisher thanks to the author of this timely (for me) piece. the final paper deals with a somewhat different problem of retirement investing, the complex new provisions found in many corporate bonds. as investors approach retirement, many find that their degree of risk aversion increases with respect to their investment portfolio since they are becoming more dependent on their non-human capital. one common response is to increase the proportion of fixed income assets held in the portfolio. corporate bonds of well-known companies are often used for this purpose because they generally have good credit ratings and yields which are higher than those from the government, and generally good credit ratings. however, as joseph fields, david kidwell, and linda klein discuss in the paper entitled “coupon resets versus poison puts: a valuation of event risk provisions in corporate debt,” corporate bonds are subject to event risk or loss in value due to corporate restructurings such as taking on huge amounts of new debt. to protect themselves from these potentially harmful events, bondholders have insisted on “covenants” or provisions that will protect them from the financial effect of restructurings or other actions of management that may adversely affect bond values. one such provision is the “poison put” which forces the company to repurchase the bonds at face value if management engages in specific activities that are harmful to bondholders. a second provision is the “coupon reset” which mandates that the bond’s coupon payment be increased if the rating falls. the authors find that investors value coupon resets but not poison puts. however, the overall value of this article is greater than just this empirical finding in that it points out and explains the increasing complexity in formulating a retirement investment portfolio. this places additional responsibility on the shoulders of the investor and his or her advisor or broker to read and understand the indenture provisions carefully to insure that particular bonds are suitable vehicles for a particular retirement investment portfolio. from the editor vii in this issue, we add a regular feature, “a practitioner’s perspective”. the objective of this column is to discuss one of our academic papers and the implications that the research has for the individual and the financial practitioner. in addition to her academic work, our managing editor, barbara s. poole, has a financial planning practice and has earned the certified financial planner (cfp), chartered financial consultant (chfc), and other professional designations. as the financial services review finishes its third year of publication, managing editor barbara poole and i wish to extend our sincere thanks to our editorial board and reviewers for the help that they have provided in launching this journal. when you publish the first journal in an emerging field, you have the added responsibility of “developing” your authors and their papers. for this reason, barb and i welcome telephone calls or drafts from prospective authors and will often take a promising but unsuitable piece and give detailed suggestions of how it could be reworked into a paper suitable for publication. further on in the process, our editors (who are often the reviewers) also contribute to this process of development. in a letter that accompanied the final draft of a paper in this issue, the authors wrote: “i would like to thank you and the reviewer for your help on the paper. the review was the most comprehensive and perceptive one i have experienced. the reviewer’s comments added immeasurably to the final result.” pii: 1057-0810(93)90007-d i financial services review, 3(l): 75-81 copyright q 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. i an empirical analysis of the use of money orders, the payment system of the poor kenneth n. daniels neil b. murphy dennis m. o’toole although money orders hove been available in the united states since the civil war, until the mid 1970’s and the failure of united states ~avigatioa company and ~m’versa~ money orders, there had been little analysis of the money order ma&et. this s&y e~i~ca~~y investigates the dete~inants of money order usage by ~use~l~. the results of the st&, which utilizes two large national samples, indicate that money orders are clearly an inferior good which have a high probability of being purchased by a low income, young, ethnic minority. i. introduction money orders have been available in the united states since the civil war but actually originated in france in the form of postal money orders in 1627. yet until the mid 197os, there had been very little analysis of the money order market in the united states. it took the failure of united states navigation company and univer sal money orders in 1977 to focus attention on the money order industry. this failure left 250,000 customers holding over $15 million in dishonored money orders. horvitz and harper (1980) estimated that in 1977 nearly a billion money orders were sold with a total value of over $40 billion dollars. assuming no growth in the number of money order sales since 1977, and adjusting only for price changes, a total value of over $87 billion were sold in 199 1. despite this long history and the significant size of the market, money orders remain a forgotten part of the payment system for researchers. the lack of research information on money orders is due to several factors. kenneth n. daniels l dep~ent of finance and insurance, v~rgj~a commonweai~ u~versity, richmond, va 23284~4ooo. neil b. murphy l department of finance and insurance, virginia commonwealth university, richmond, va 23284-4000. dennis m. o’toole l department of economics, virginia commonwealth university, richmond, va 23284-4000. 76 ~anc~ls~rvices~v~w,3~1) 1993 1. most states do not require money order companies to submit data related to volume of money order sales. 2. information on money orders is difficult to compile because of the fragmented nature of the market. little, if any, financial data are available on the vast number of small issuers in the market such as liquor stores and small grocery stores. often, no record exists of even how many such firms there are. 3. money order information is often combined with travelers’ checks or certified checks and is difficult to separate. 4. many licensed issuers do not file public financial statements because they are privately held and have less than 500 sh~eholde~. the purpose of this study is to empirically investigate the determinants of usage of money orders by households. in section ii, the background and the nature of the money order industry are presented. section iii contains a review of infor mation on the users of money orders found in a previous study. the data source and the empirical results are discussed in section iv. section v is the summary and conclusion. ii. backgr~u~on~ money o~er~~us~y a money order, like a check, is a means of transmitting money from one party to another. more specifically, a money order is an order for the payment, usually to a third party, of a sum of money specified by the purchaser. for example, a person could go to the local post office and for a small fee plus one hundred dollars, purchase a one hundred doliar money order to pay his utility bill. the purchaser receives the money order from the post office, writes the name of the utility on the money order, and sends it to the utility. thus, the money order serves the same purpose as a check. money orders are included in the ml definition of the money supply. bank money orders are considered a part of the “demand deposit” component of m 1 and thrift money orders are included in the “other checkable deposit” of ml. money orders issued by commercial issuers such as american express are included in the “travelers checks” component. thus, the exact component of ml depends on the agency issuing the money order and varies among “demand deposit,” “other checkable deposit,” and “travelers checks” components. the structure of the money order industry is focused on the money order issuer.’ the issuer is the one that is ultimately responsible for the financial obligation created by the sale of the money order. the money order industry has three major types of issuers. the three types are as follows: 1. postal money orders 2. depository ~stitu~on money orders 3. commercial money orders postal money orders are obligations of a federal government agency and are an implicit obligation of the federal government. money orders issued by depository insti~tions (not by a subsidiary of a depository insti~tion holding company) are insured by the federal deposit insurance corporation because the money order is a direct obligation of the depository institution and considered deposits. commercial issuers of money orders are a very diverse group. in addition to the major issuers such as american express which have a large national presence, there are a number of very small firms with only one or two offices in one geographic area. most state regulations require commercial issuers to buy a surety bond to provide holders of their money orders with a guarantee that the instrument will be covered if the firm fails. however, the value of the surety bond is a small fraction of the firm’s outstanding money orders on any given day. money orders are sold through a system of agents. an agent is a party that directly contracts with the issuer to sell money orders. for example, american express (the issuer) may contract with a convenience store chain (the agent) to sell money orders. the terms of the contract depend on the desirability of the agent and the degree of competition in the market. based on these two factors, several items in the contract are negotiable. three of these items are as follows: 1. the remittance time 2. the rate scale 3. the division of revenue fees the re~tt~ce time is the number of days that may pass before the agent must send the funds from money order sales to the issuer. the longer the remittance time the longer the agent has to use the float from money order sales. the rate scale is generally uniform for all agents. however, if an agent feels the rates are too high, the issuer and the agent can negotiate a special rate structure. the third factor that can be negotiated between the issuer and the agent is the division of revenue fees. the more desirable the agent and the better the market, the larger the share of the proceeds the issuer is willing to give to the agent. the primary regulators of money order sales are the individual states. most of the regulation is focused on commercial issuers and agents. there are three main types state regulations. these are as follows: 1. licensing requirements 2. net worth requirements 3. posting of surety bond or securities in general, any person may obtain a license to engage in the money order business provided that certain legislative requirements are met. in most states, the main requirement is for the department charged with administering the law to investigate the financial condition and responsibility of the applicant. most states 78 financialservicesreview,3(1) 1993 require that an issuer of money orders must rn~nt~n a ~imum net worth. the net worth requirement ranges from a low of $5,000 to a high of one million dollars. a number of states require a licensee of money orders to post some kind of security as evidence of good faith and to protect the public if the licensee defaults. the failure of universal money order and united states navigation, both subsidiaries of intemation~ express, in 1977 was estimated to have left approxim~ely 4~,~ individuals with about $20 million worth of dishonored money orders. the surety bond provisions proved to be much too small in most states. state regulation of the money order industry has resulted in little uniformity in requirements or regulation of money order firms. thus, very little attention has been given to the public policy implications of the payments system used mostly by the poor. iii. t~usersof money orders one previous research study that addresses the question of who uses money orders was conducted by pierce (1977). the data used for that study were the claim forms submitted to the california banking department by the holders of dishonored money orders from the failed universal money order company. each of the claim forms contained the person’s name, address, number of defaulted money orders held, and the amount of the claim. pierce found that, relative to the general population, money order users have lower incomes, have less formal education, are older, and are more likely to be in a minority group. the author concluded that the sample group preferred money orders to checking accounts and used money orders as substitutes for checks for two main reasons. first, most of the sample group did not fee1 they made enough income to justify a checking account. secondly, a number of those interviewed felt that filling out the check stubs and worrying about insufficient funds was “too much of a hassle”. however, pierce’s study was not based upon scientific sampling procedures and did not apply multivariate techniques to the analysis of the data. iv. datasowrceandempiricalf~esults in 1984 and 1986, the board of governors of the federal reserve systemcontracted with the survey research center of the university of michigan to conduct in each year a survey of currency and transaction account usage. between may and august of 1984, a total of 1,946 telephone interviews was obtained from a randomly selected sample of 2,500 families residing in the united states. for the 1986 sample, the survey of currency and transaction account usage was conducted as part of the monthly survey of consumer attitudes of the survey research center of the university of michigan. in this case, telephone interviews are conducted by select ing telephone numbers from a cluster sample of residential numbers. in the 1986 money ordirs 79 survey, there was a response rate of 75% with a total of 658 interviews completed. a discussion of the results of these surveys, along with more detailed information regarding the sampling methods, is presented in avery, elliehausen, kennickell, and spindt (1986, 1987). in this study, household usage of money order users is examined. unfortunately, the 1984 survey didn’t ask households about their money order usage. however, the 1984 survey does provide information on the total dollars of money orders purchased by the household in the previous month. for the 1984 sample a dummy variable, morduse2, was created that equals one if the house hold purchased money orders in the previous month and equals zero if otherwise. the 1986 survey does provide information on money order usage with the binary variable, morduse, which equals one if the household uses money orders, and equals zero if otherwise. however, to check the validity of the dummy variable, morduse2, created in the 1984 sample, the same procedure was applied to the 1986 sample. using the 1986 sample, the logit regressions revealed the same results with morduse or morduse2 as the dependent variable. the basic model the model for money order use was developed as the demand for a financial asset. in this case important independent variables are a scale variable and several demographic variables. the natural logarithm of annual household income for the preyious year is the scale variable. demographic variables that are of interest are race, education, and gender. a logit regression technique is used to estimate the extent to which the independent variables affect the probability that a household is a money order user. table 1. logit results dependent variable is mordusez 1984 survev variable coe@cienr chi-square p-value constant 4.81’ 19.32 0.00 lnage -o.90* 24.33 0.00 race 1.14* 50.91 0.00 lnannlinc -0.36; 17.84 0.00 model chi-square = 128.11. note: * indicates that the &i-square statistic is significant at the 5-percent level. wage = natural logarithm of age of respondent in years morduse = use of money orders by household, equal to 1 if money orders are used, 0 for all others. morduse2 = dummy variable created for 1984 sample, equal to 1 if household purchased money orders in the previous month, 0 for all others. race = is tbe respondent a minority, equal to 1 if black, hispanic, or asian descent, 0 for all others. lnannlinc = natural logarithm of household annual income for the previous year. 80 financial services review, 3(l) 1993 table 2. logit results dependent variable is morduse 1986 survey variable constant lnage race lnannlinc coelgicient 4.93. -1.14; 1.31* -0.31* chi-square 6.29 12.29 19.25 4.07 p-value 0.01 0.00 0.00 0.04 model chi-square = 40.54* note: * indicates that the &i-square statistic is significant at the 5-percent level. lnage = natural logarithm of age of respondent in years morduse = use of money orders by household, equal to 1 if money orders are used, 0 for all others. morduse2 = dummy variable created for 1984 sample, equal to 1 if household purchased money orders in the previous month, 0 for all others. race = is the respondent a minority, equal to 1 if black, hispanic, or asian descent, 0 for all others. lnannlinc = natural logarithm of household annual income for the previous year. a4ordzjse was the dependent variable for the 1986 logit regression analysis, and a4orlxjse2 was the dependent variable for the 1984 logit regression analysis. empirical results the results of the logit regressions are shown in tables 1 and 2. for both years, the variables that are statistically significant are lnage, lnannlznc, and race as indicated by their respective chi-square statistic. a likelihood-ratio test was also performed to test the significance of each coefficient independently, and the results coincided with the chi-square test. the significant variables have the expected sign on their respective coefficients. this indicates that the probability of money order use is less likely with households that are older, have higher incomes, and are white. money order use exhibits the behavior of an inferior good, as indicated by the significant negative coefficient on the scale variable, confirming the findings of pierce (1977). being black or hispanic increases the probability of money order use. moreover, the results are remarkably consistent from one sample to the next. v. summary and conclusions money orders are an important and sizable portion of the united states payments system. the results of this study, utilizing two large national samples, indicate that money orders are clearly an inferior good which have a higher probability of being purchased by a low income, young ethnic minority. money orders 81 note 1. for additional discussion of the money order industry, see gonzalez and caron (1977). rjwerences avery, robert b., gregory e. elliehausen, arthur b. kennickell, and paul a. spindt. 1986. “the use of cash and transaction accounts by american families.” federal reserve bulletin, 87-108. avery, robert b., gregory e. elliehausen, arthur b. kennickell, and paul a. spindt. 1987. “changes in the use of transaction accounts and cash from 1984 to 1986,” federal reserve bulletin, 179-196. bucher, jeffrey. m., and gail ruffin-cruz. 1977. “federal approaches to regulation of money orders and traveler’s checks,” symposium on money orders and traveler’s checks, san francisco: california state banking department. carrig, james f., and joseph w. russell. 1977. “regulation of issuers of money orders andtraveler’s checks,” symposium on money orders and traveler’s checks, san francisco: california state banking department. gonzalez, gisela a., and susan t. caron. 1977. ‘the structure of the money order industry,” fdic working paper no. 77-5, washington, dc.: federal deposit insurance corporation. horvitz, paul m., and charles p. harper. 1980. ‘the regulation of the money order industry,” financial management, 9, 13-20. penzer, michael l. 1977. “the nature and size of money order and traveler’s check markets in california and the nation,” symposium on money orders and traveler’s checks, san francisco: california state banking department. pierce, james l., 1977, ‘the users of money orders,” symposium on money orders and travelers checks, san francisco: california state banking department. pii: 1057-0810(93)90002-8 from the editor as the financial services review enters its third year of publication, the quantity and quality of submissions continues to increase. this was anticipated for two reasons. first, as the scope and standards of the journal have become more widely known, inappropriate pieces are less frequently sent, enabling our editors to offer faster turnaround and better service. second, a mature journal offers less risk to established authors and encourages them to send their best work in the area of individual financial management to us. in creating the financial services review, the academy of financial services hoped, in part, to encourage a group of scholars to focus on the finances of the individual. a glance at the table of contents of this issue offers some evidence that such a group may be forming. each of the articles was authored or co-authored by a person who has previously written for this journal and/or who has won the academy’s best paper award. in the case of jessie xiaojing fan, yu-chun regina chang, and sherman hanna, co-authors of “real income growth and optimal credit use,” this group has won the best paper award twice in successive years. while taking pride in its ability to present multiple pieces by the same authors, the financial services review also points out that many of the co-authors of articles presented in this issue have not appeared here before. in addition to its desire to showcase the work of experienced scholars in this field, the journal also attempts to develop new authors and authors who are new to the field. for this reason, we encourage calls to the editors to discuss ideas as well as informal submissions seeking comments and direction. we also encourage our readers to direct appropri ate research of their colleagues to our attention. in the current issue, ronald r. crabb continues his work in the area of probabilistic estate planning to the formation of efficient frontiers consisting of the most commonly-employed estate planning tools. in the article entitled “efficient frontiers in estate planning,” professor crabb uses numeric simulation of the age of death and borrowing and lending rates to see which combination of estate-plan ning tools is likely to yield the best results. the technique that he develops can be used by sophisticated estate planners and should become the norm among practitio ners. professors steven p. rich and william reichenstein explain how an individual investor can use market timing to boost portfolio returns. in their article entitled “market timing for the individual investor: using the predictability of long-horizon v vi financial services review, 3(l) 1993 stock returns to enhance portfolio performance,” they present evidence that measures of the market risk premium which are readily available to the individual investor show strong ability to time the market. they also offer an economic justification for their results which is not inconsistent with efficient markets. consumer borrowing is often justified under the assumption that real income will grow and the relative burden of payback will be diminished. in today’s economic era, however, growth in real income is far from assured and people who borrow may end up paying back their debts with a larger piece of their income. this is the problem investigated by jessie xiaojing fan, yu-chun regina chang, and sherman hanna in their article entitled “real income growth and optimal credit use.” they approach it theoretically through a two period model of consumption and also analyze the impact of real income growth with numerical analysis. income uncertainty crops up again in the question of whether well-heeled individuals should hold private-activity tax-exempt bonds in their portfolios if such bonds may be subject to the alternative minimum tax. in her article entitled ‘the individual’s tax-exempt bond portfolio decision under income uncertainty,” professor amy puelz shows that most risk-averse investors will maximize the expected utility of their (uncertain) after-tax income by holding a sizeable propor tion of private activity bonds. the last article is unique in that it addresses the interests of a different group of individuals, using different techniques than the other articles in this issue. “an empirical analysis of the use of money orders, the payment system of the poor,” by professors kenneth n. daniels, neil b. murphy and dennis o’toole uses federal reserve surveys carried out by the university of michigan’s survey research center to find out who uses money orders. the results indicate that money orders are an important part of the payments system of the united states but that they are an inferior good, used primarily by those with lower incomes. -lewis mandell from the editor this issue contains volume 30 issue 3 of the financial services review (fsr). i would like to personally thank the board and academy of financial services members for their support. i would also like to thank you, our authors, and readers for your continued patience. although we are about three months off our publication timeline, we are excited to bring you this issue. the lead article “the effect of racial/ethnic differences on the financial obligations ratio of renters?” is coauthored by congrong ouyang at kansas state university and sherman hanna at ohio state university. the purpose of this research was to investigate the effects of racial/ethnic status on the ratio of financial obligations payments to income among u.s. renter households. based on the ols regression results, the author found that households with black, hispanic, and asian respondents had higher financial obligations ratios than otherwise similar households with a white respondent. the authors conclude that while discrimination could be a factor in higher ratios for the groups other than whites, immigrant status and other factors plausibly are related. the second article examines whether gender moderates the perceptions of industry corruption. “impact of consumer perceptions of industry corruption on the choice to engage a financial advisor: does gender matter?” this article is authored by danielle winchester, roland leak, and nicole mccoy at north carolina agricultural and technical state university, willie a. deese college of business and economics. their study explores the intersection of financial advisors’ gender and consumers’ industry corruption perceptions on the likelihood of using a female advisor as females are perceived as more trustworthy and less prone to corruption than their male counterparts. analyses reveal individuals prefer female advisors when corruption is low, but these preferences wane as corruption perceptions heighten. their results suggests the interpersonal characteristics of females being more trustworthy and ethical do not carry as much weight for consumers when they perceive the industry as corrupt. coauthored by richard stebbins and kyoung tae kim at the university of alabama and martin saey at kansas state university, “financial professionals and financial well-being: evidence from the national financial well-being survey” examined the association between financial professional use and financial well-being using the 2016 national financial wellbeing survey. the authors tested financial well-being across various sources of financial advice such as financial professionals, family, employer, community, financial institution, and government. using logistic regression results they showed that those who received advice from financial professionals had higher levels of financial well-being than those who did not receive 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) v–vi advice from a financial professional. moreover, further analyses highlighted that those who received financial advice from any source also showed that the use of a financial professional had a stronger positive association with financial well-being. the final article of this issue highlights work from lua augustin at slippery rock university and terrance martin at winston-salem state university. in their paper “financial literacy and financial planning use,” the authors evaluate the relationship between financial literacy and the level of financial planning behavior that individuals display. using 2012 and 2014 waves of the national longitudinal survey of youth (nlsy79), they found that higher levels of financial literacy are positively associated with more ‘comprehensive’ levels of financial planning behaviors. additionally, the results of the analysis indicates that individuals at the highest level of financial literacy are more likely to choose ‘comprehensive’ planning over doing nothing at all. in addition, those with higher levels of financial literacy are also more likely to display some level of financial planning behavior compared to doing nothing at all. i want to thank everyone who makes the fsr production possible, including our authors, referees, and readers. if you have a manuscript and looking for a publication outlet, please consider submitting it to the financial services review. the journal welcomes articles on the many areas of personal financial planning. the editorial team remains committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. yours sincerely, terrance k. martin jr. editor financial services review vi t. k. martin / financial services review 30 (2022) v–vi pii: s1057-0810(97)90005-6 i i financial services review, 6(4): 271-284 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. i i the optimal choice of index-linked gics: some canadian evidence moshe arye milevsky sharon kim indexed linked guaranteed investment certificates (ilgics), also known as equity linked term deposits, have become quite popular over the last few years. from the con sumer’s point of view, there are two basic categories of indexed linked gig, the capped ilgic and the participating ilgic. this paper compares the relative value and appeal of the two ilgic products from the perspective of the individual investor, using valuation techniques from option pricing theory. we compute the value per pre mium dollar invested in an ilgic. our main conclusions are that ilgics become less attractive the longer the time horizon of the investor. in addition, participating ilglcs are preferable to capped ilgics for short-term maturities and vice versa for longer maturities. a detailed analysis of two particular canadian products is provided as an application of the basic concepts. we conclude by demonstrating that values per pre mium dollar range from 94% 98%. corresponding to a 2%~6% total expense ratio on index linked gics. i. introduction with the general decline in the level of interest rates, indexed linked guaranteed invest ment certificates (ilgics) have become extremely popular over the past few years. every one of the six (major) canadian banks offers a variant on this product. an ilgic (in the u.s., an equity enhanced certificate of deposit or equity indexed annuity) is a retail sav ings vehicle, similar to a term deposit, with an interest rate that is linked to a diversified stock index. upon maturity of the product, the total is determined based on the perfor mance of the underlying index over the pre-specified term. within the family of products available to the consumer, there are two basic structures of ilgics, namely the capped ilgic (c-ilgic) and the participating ilgic (p-ilgic). the c-ilgic allows the holder to participate in any upward movement in the underlying stock index, up to a pre-specified annual percentage cap. the p-ilgic allows the holder to moshe arye milevsky and sharon kim l schulich school of business, york university, toronto, ontario, m3j lp3, canada; e-mail: milevsky@yorku.ca. 272 financial services review 6(4) 1997 participate in a fixed fraction of any upward movement in the underlying stock index, with no pre-specified cap. both c-ilgics and p-ilgics have an implicit guarantee of principal plus minimal interest. there are, however, some products available that are both capped and have partial participation. we will not address them in this paper, since they constitute a very minor part of the market. here is an example: bank c offers a 5-year ilgic whose total return at maturity will be the greater of 5% and the net return (excluding dividends) on the sp500 index, up to a maximum of 60%. this product is a c-ilgic where the floor is 5% and the cap is 60%. likewise, bank p offers a 5-year ilgic whose total return at maturity will be the greater of 3% and 80% of the net return (excluding dividends) on the sp500 index. this product is a p-ilgic where the floor is 3% and the participation rate is 80%. at maturity, if the sp500 increases by 40% over the five year term, bank c will allow full participation and credit the account interest in the same amount. on the other hand, bank p will only credit the account 80% of the 40% increase, resulting in a total return of 32%. in a similar fashion, if the sp500 increases by 100% over the five year term, bank c will cap the return at 60%, while bank p will credit the account 80% of the 100% increase, resulting in a total return of 80%. figure 1 is a graphical representation of the payoff structure from both types of ilgics. the horizontal axis is the value of the underlying index at maturity, denoted by 5,. the vertical axis represents the total payoff from the ilgic. the c-ilgic has a maximum possible payoff of exp(c7), where c is the continuously compounded annualized cap rate and t is the maturity of the contract. the minimum guaranteed payoff from the c-ilgic is exp(g,t), where g, is the continuously compounded annualized floor. in contrast, the p ilgic has no maximum payoff, but the slope of the payoff diagram is reduced in propor figure 1. ilgic payoffs the optimal choice of index-linked gzcs 273 tion to p, which is the participation rate. once again, the minimum interest is exp(gj), where gp denotes the continuously compounded annualized floor on the p-ilgic. note that although the caps and floors are applicable to the entire term of the ilgic, the figures are presented on an annualized basis in order to facilitate the comparison between products with different maturities. with hindsight it is quite obvious which of the two basic products “was” better. in practice, however; it is quite difficult to choose the right one ex-ante. this study compares the relative value and appeal of the two basic ilgic products from the perspective of the individual investor, zn addition we provide a methodology for extracting the implicit expense ratios of the two types of ilgic. using concepts from option pricing theory, we are able to value both types of ilgic by decomposing their payoff into a zero-coupon bond and a suitably parameterized collec tion of call options on the underlying stock index. these values apply regardless of the risk preferences of the individual investor. the intuition behind this result follows from arbi trage valuation and is described at length in the body of the paper. a detailed analysis of two particular bank products is provided as an application of the basic concepts. the remainder of this study is organized as follows: section ii conducts a brief litera ture review of ilgics and then describes the particular ilgic products available in can ada; section iii develops the pricing relationships for the two basic categories of ilgics; section iv presents a numerical example of the relative valuation of the two most popular bank products, thus extracting the embedded expense ratios; and section v concludes the paper. ii. literature review in an introductory study geared to the individual investor, cohn and edleson (1993) exam ine the basic features of ilgics. they discuss the various products available in the u.s. market at the time and identify the embedded zero-coupon bond plus call option structure. they do not, however, focus on the relative benefits of capped vs. participating gics, nor do they attempt to identify the best products available in the canadian market. there are a few business press articles on the subject of canadian-based ilgics, see for example bell (1997) and croft (1997). however, there is little, if any, rigorous academic research on the specifics of the canadian market. baubonis, gastineau, and purcell (1993) provide a qual itative description, from an institutional perspective, of the embedded derivative securities in ilgics, but avoid any discussion of pricing or relative value. brooks (1996) provides a framework for comparing certificates of deposit that vary in their features. in particular, he uses a derivative pricing methodology to compute the value of guaranteed interest rate floors and caps. our analysis is a natural continuation of this line of research by focusing on equity-enhanced, as opposed to interest rate-enhanced, products. we are thus able to facilitate comparison of vastly different products, using the black and scholes (1973) and merton (1973) framework, as suggested by brooks (1996). as mentioned in the introduction, purchasing an ilgic can be viewed as an alterna tive to investing in the underlying stock index and using the dividends from the portfolio to purchase put options which protect the investment against a market decline beyond a t a b l e 1 in de xlin ke d g ic s y p b an k t er m t ax t re at m en t d es cr ip ti on n at io na l b an k of c an ad a 5 ye ar s a va ila bl e on ly as a r eg is te re d r et ir em en t c la ss if ie d as c -i l g ic t or on to d om in io n b an k 3 ye ar s g ig p lu s o r 5 ye ar s l au re nt ia n b an k 3 ye ar s o r 5 ye ar s b an k of n ov a sc ot ia 2 ye ar s c an ad ia n im pe ri al b an k 3 ye ar s of c om m er ce c an ad ia n or in de x fu nd 5 ye ar s sa vi ng s pl an (r r sp ) if h el d ou ts id e th e r r sp , it ca n be s ol d ba ck to t he b an k as r eg ul ar in te re st in co m e w ith no ac cr ua l t he m in im um gu ar an te ed in te re st is r ec ei ve d ye ar s an d ca lc ul at ed fo r in co m e ta x pu rp os es as i nt er es t in co m e if h el d ou ts id e an r r sp , ga in s ar e ca lc ul at ed as i nt er es t in co m e if h el d ou ts id e an r r sp , ga in s ar e ca lc ul at ed as i nt er es t in co m e in t he y ea r in w hi ch it is r ec ei ve d l in ke d to t s e 3 5 g, = 3 % f or e nt ir e sye ar te rm c = 60 % f or e nt ir e sye ar te rm t he t se 35 cl os in g le ve l on t he i i th o f ea ch m on th of th e fi na l ye ar is u se d to c al cu la te th e fin al in de x le ve l c la ss if ie d as p -i l g ic as i t us es s om e av er ag in g l in ke d to t s e io 0 g, , = 0. 25 % pe r an nu m fo r th e 3ye ar pr od uc t gp = 2 .2 5% pe r an nu m fo r th e 5ye ar pr od uc t in b ot h ca se s, in te re st is r ec ei ve d at t he e nd o f th e en tir e te rm m ea n cl os in gs of t he t se lo o on t he s am e da y of ea ch m on th of t he f in al ye ar ar e us ed t o ca lc ul at e th e fi na l in de x le ve l. c la ss if ie d as a p -i l g ic l in ke d to t he t s e 3 5 g, = 1% p er a nn um fo r bo th te rm s, w hi ch is r ec ei ve d ye ar ly an d la te r de du ct ed if t he i nd ex ’s fi na l yi el d is gr ea te r th an 1% p .a . p = 75 % f or t he 3 -y ea r pr od uc t p = 10 0% f or t he 5 -y ea r pr od uc t t he i nd ex ’s fi na l le ve l is c al cu la te d as t he m ea n of t he cl os in gs on t he f in al 3 da ys pr io r to m at ur ity c la ss if ie d as a c -i l g ic l in ke d to t he t s e 35 c = 2 0% fo r th e en tir e 2 ye ar te rm t he f in al le ve l of t he i nd ex is th e cl os e le ve l of t he t s e 3 5 on t he m at ur ity da te c la ss if ie d as p -i l g ic as i t us es s om e av er ag in g l in ke d to t he t s e 3 5 g= o % t he t s e 3 5’ s fi na l le ve l is c al cu la te d as t he m ea n cl os in gs o f th e in de x in t he l as t 11 m on th s of t he t er m the optimal choice of index-linked gics 215 certain level. this strategy can also be implemented in the form of portfolio insurance with an arbitrary floor. before continuing, we note an issue related to ilgics that we will not discuss in detail, but is still relevant. from a utility based perspective, there is some question as to why an individual investor would want to insure a portfolio beyond a certain floor. indeed, it appears that most individual investors exhibit decreasing, or at least constant, relative risk aversion (rra). see recent empirical work by schooley and worden (1996) for evi dence and references. however, benninga and blume (1985), brennan and solanki (1981), and many others show that even in markets that are highly incomplete, rational investors with decreasing relative risk aversion will avoid portfolio insurance strategies. their research questions the need for products with arbitrary guarantees in a fully rational market place. in plain english, why are investors exhibiting preferences that are discontinuous at the guaranteed floor? nevertheless, these products do exists in the market place, and the remainder of this paper will attempt to shed light on their relative value. we are thus able to answer the ques tion: “if i do want to buy an ilgic, which gives me the best value?’ 1. details of the canadian market the canadian banking industry is very concentrated and dominated by six national banks and two or three deposit-taking trust companies and regional credit unions. within the last two years, every one of these institutions has started to offer and promote a variant of the index linked gic. table 1 provides a brief summary of the details of the most popular ilgics in canada together with a description of their embedded options. the same table also identifies the tax consequences of purchasing the product outside of a self-directed pension plan. (in canada they are called registered retirement saving plans similar to iras and 401(k)s in the united states.) we also note that some of the products have embedded averaging fea tures which reduce the final payoff from the ilgic compared to the performance of the market. these products would fall under the broad category of p-ilgics since their final payoff is proportionally reduced. the actual decomposition would involve asian options which would be bounded from above by the corresponding vanilla options. our numerical examples, however, focus on products offered by laurentian bank and the bank of nova scotia that do not contain this added complication. iii. pricing relationship this section will present a formula for the market value of an ilgic, as a percent ofpur, in a black and scholes (1973) and met-ton (1973)framework. by percent ofpur we mean the ratio of the theoretical price to the initial investment in the ilgic. thus, a value equal to one denotes a fairly priced ilgic, a value less than one denotes an unfairly priced ilgic, and a value greater than one denotes a super-fairly priced ilgic. a super-fairly priced ilgic will admit arbitrage and is thus very unlikely to exist. our main goal is to calculate values per premium dollar (vpds) for the various prod ucts available in the canadian market place. we achieve this by computing the risk-neutral 276 financial services review 6(4) 1997 expected payoff from the c-ilgic and p-ilgic and then discounting by the appropriate risk-free rate to obtain the present value. 1. c-ilgic an investor places an amount, denoted by i, in a c-ilgic. the payoff from the c ilgic, at maturity, is: i lexp{ct}, if st exp{ ct} i so st i> st so if exp{gct) i: 5 exp{ct} so (1) i zexp{ct}, if st s_ s e-v{gcti 0 where c > 0 is the annualized cap rate, g, 2 0 is the annualized guaranteed minimal interest floor, t is the time to maturity, so is the initial level of the stock index and st is the (sto chastic) value of the stock index at maturity. by definition we assume that c > g, for the ilgic to make sense. from a qualitative perspective, if the one-plus total return from the stock market index, s&so over the time period [oj, is less than the minimal guarantee exp{g,t}, the payoff from the c-ilgic is simply i exp{g,t}. on the other hand, if the one plus total return from the stock market index, s&su over the time period [o,z’j, is greater than the minimal guarantee exp{g,t}, but less than the total cap exp{ct}, the payoff from the c-ilgic is i (st/so). finally, if the one-plus total return from the stock market index, s&so over the time period [0, t], is greater than the total cap exp{ct}, the payoff from the c-ilgic is (capped at) i exp{ct}. the three branches of the payoff structure can be com bined using the max[ .] operator. see the appendix for details. 2. p-ilgic upon investing i in the participating ilgic, the payoff is: if ~ewg,,ti, if (2) where p > 0 is the participation rate, gp 2 0 is the annual guaranteed minimal interest floor, t is the time to maturity, su is the initial level of the stock index and st is the (stochastic) value of the stock index at maturity. intuitively, if one-plus p% of the total return from the stock market over the time period [0, t], is less than the minimal guarantee exp{gpt}, the payoff from the c-ilgic is simply i exp{gpt}. on the other hand, if the one-plus p% of the total return from the stock market, over the time period [0, z’j, is greater than the mini the optimal choice of index-finked gics 217 ma1 guarantee exp{gpt}, the payoff from the c-ilgic is z st ( 1 1 so p + i, which is simply the original investment plus the proportional participation. once again, the two branches of the payoff structure can be combined using the max[.] operator. see the appendix for details. 3. valuation of c-ilgic we refer the reader to the appendix where we derive the fair value of the c-ilgics using the standard black and scholes (1973) and merton (1973) methodology for pricing contingent claims. in this section we simply reproduce the valuation formulas. given a set of product parameters (c, g,), where c is the cap rate and g, is the guaran teed floor, and capital market parameters (r, q, o), and maturity t, the value per premium dollar (vpd) of the c-ilgic is: v,(c,g,) = exp{(g, r)t) + expt-qtll\r(bl) exp((g, ww2) exp{-qt}n(bs) + exp((c g, gvv74) (3) where n(.) denotes the cumulative density function of the normal distribution, r is the risk free interest rate in the market, q is the dividend yield on the underlying stock index, (t is the volatility of the underlying stock index and b, = r-4-gc 0 6, = b, = r-q-c-gc b, = r-q-c-g= 0 (t (4) notice that equation (3) does not involve the variable so, since all that is relevant is the per cent increase. we explicitly parametrize v, by the two important variables (c) and (g,). obviously, the actual value will depend on the implicit parameters o, q, r, t as well. 4. valuation of p-ilgic in the same manner, the vpd of the p-ilgic is: vp(p, gp> = e-v {(g, $“i + pedqw(a 1) pew((g, 6‘ww where p is the participation rate and g, is the guaranteed floor and (6) r-4-g al = ( ci “+4 jt; 1 a2= ( r-q-g, 0 0 -jfi 5. analysis (7) a few stylized facts emerge from our analysis. from a qualitative point of view, we can make the following statement without need for much algebra. 278 financial services review 6(4) 1997 for a given floor g, no cap is better than partial participation. likewise for a given floor g, total participation is better than a cap, no matter how large the cap. also, can be verified by plugging c = m, p = 1, g, = g, = g into equation (3) and (6), respectively. figure 2 displays the vpd for both the c-ilgics and p-ilgics as a function of the maturity time horizon t, for an arbitrary set of capital market and product parameters. the first point that is evident from the picture is that both curves decrease as a function of time. this means that the consumer gets less, per dollar invested, the longer the maturity of the ilgic. the intuition behind this result is simple. the price of the zero coupon bond, used to fund the minimal interest, decreases at a faster rate than the price of the call option increase, which is used to fund the equity participation. therefore, all else being equal the longer the time horizon, the more the investor looses. this result can be rigorously obtained by taking the derivatives of equation (3) and (6) with respect to the time horizon, t, and showing that it is negative. of course, the rate at which the vpds decline, or the figure 2. value per premium dollar (vpd) of index linked gic. capped-ilgic vs. participating ilgic. cap = 9.12% p.a., participation = 75% no minimal interest guarantee risk free interest rate = 4.3%, volatility = 12%. dividend yield = 2.0% the optimal choice of index-linked gics 279 slope, will depend on the capital market parameters (r, q, (t) vis a’ vis the magnitude of the cap and the participation rate. another interesting fact to note about figure 2 is that for maturity horizons less than approximately 3 years, the participating ilgic is worth more than the capped ilgic. at around 3 years they are both worth the same, and then for longer horizons the situation is reversed and the c-ilgic is preferable to the p-ilgic. this is not a spurious result of the input parameters, but rather a fundamental insight into the structure of c-ilgics vs. p ilgics. basically, as in all option pricing models, it all comes down to probabilities. for short time horizons, the c-ilgic is worth less than the p-ilgic because the cap is a strong constraint on the performance of the product. indeed, it is quite likely that the market will earn more than c percent annualized, and the gains will be truncated. on the other hand, as the maturity of the product is extended, the probability that the cap is constraining, on an annualized basis, decreases and thus the relative value of the c-ilgic increases. in other words, it is more likely that the sp500 will increase by more than 30% in one year compared to 30% annualized over 5 years. therefore, the cap becomes less of an issue as time increases. of course, the fact remains that the absolute value of the c-ilgic declines with maturity. finally, one should not lose sight of the main conclusion that the values per premium dollar on available products are uniformly less than one, which means these products are unfairly priced for the consumer. (the anomaly in the short horizon for the c-ilgic and p-ilgic is an artifact of the algebra and simply reflects the fact that no institution is likely to offer this kind of c-ilgic with maturity less than three months and a p-ilgic with maturity less than 1.6 years.) in sum, a vpd of x% implies that (1 -x)% of the initial investment is paying for com missions, transaction costs and bank profits. consequently, if the individual investor can obtain the same exact payoff structure, using basic options, but incur transaction costs that are less than (1 x)%, it is cheaper to replicate the product than to purchase it from the bank. this will obviously depend on the size of the investment vis a’ vis the transaction costs incurred in self replication. from a practical point of view, the bank as an intermedi ary is likely to take advantage of its economies of scale in the market and incur lower trans action costs when purchasing the necessary derivatives. iv. numerical example: b.n.s vs. l.b in this section, we compare the ilgic offered by the bank of nova scotia (bns), a c ilgic, to that offered by laurentian bank (lb), a p-ilgic. as mentioned in section ii and table 2 value per premium dollar of the bank of nova scotia 3 year capped-ilgic as a function of risk-free interest rate (r) and volatility (0). r\o -1% vol. +i% -1% 99.224 99.3844 99.5153 int. 97.0285 97.1939 97.329 +l% 95.5944 95.6792 95.7432 note: the middle number is the point estimate for the vpd based on the market parameters provided by bloomberg on november i i. 1997. 280 financial services review 6(4) 1997 table 3 value per premium dollar of the laurentian bank 3 year participating ilgic as a function of risk-free interest rate (r) and volatility (cr). no -1% vol. +i% -1% 94.7062 95.1545 95.6082 int. 92.7801 93.6980 93.6260 +l% 90.9344 91.3008 91.6861 nc&-: the middle number is the point estimate for the vpd based on the market parameters provided by bloomberg on november 11, 1997. figure 3. value per premium dollar of index linked gic. bank of nova scotia (c-ilgic) vs. laurentian bank (p-ilgic) valuation date: november 1 i, 1997 data source: bloombergs displayed in table 1, the bns product is a 2-year c-ilgic linked to the toronto stock exchange’s index of the 35 largest stocks as measured by market capitalization. (referred to as the tse35 index.) the bns product does not offer a minimum guaranteed interest rate and its payoff is capped at 20% over the 2 years. in contrast, although lb’s p-ilgic is linked to the tse 35 index, it offers a 1% minimum guaranteed interest per annum. we will examine the 3-year product as it is an example of a product with partial (75%) participation. using the pricing equations presented in section iii, we insert the relevant parameters and solve for the vpd of a c-ilgic and p-ilgic. we value the bns product recalling equation (3), where the relevant product parameters are: c = in j1.2 = 0.09 12, g, = 0, and t = 2. we arbitrarily take our relative valuation date to be november 11, 1997 which results in a dividend yield of q = 0.015, a 2-year volatihty of 0 = 10.2% and a 2-yr risk-free the optimal choice of index-linked gics 281 interest rate of r = 4.64%. all of which were provided by bloomberg’s information sys tems which is the de-facto standard for market values in the derivatives trading industry. the resulting vpd for the bns product is 97.19%, which translates into an implicit ‘cost’ of 2.81%. we conducted a sensitivity analysis of our vpd’s by perturbing the market volatility and interest rate figures by 1% around the point estimate provided by bloomberg. table 2 is a summary of those results. continuing, equation (6) can then be used to value lb’s 3-year product, where the parameters are: p = 0.75, gp = in(1 .ol) = 0.0995 and t = 3. once again, we take our rela tive valuation date to be november 11, 1997 which, in this case, results in a dividend yield of q = 0.015, a 3-year volatility of cr = 9.42%and a 3-year risk-free interest rate of r = 4.89%. the resulting vpd is 93.69%, which translates into an implicit ‘cost’ of 6.31%. once again, we conducted a sensitivity analysis of our vpd’s by perturbing the mar ket volatility and interest rate figures by 1% around the point estimate provided by bloomberg. table 3 is a summary of those results. based on a vpd analysis, it therefore appears that the 2-year c-ilgic offered by the bank of nova scotia is more valuable than the 3-year p-ilgic offered by laurentian bank, by approximately 4.50%. figure 3 is a graphical illustration of the vpd for both products as a function of time horizon, which an average implied volatility of 10% and incorporating the complete term structure of interest rates. v. conclusion every one of the six major canadian banks offers a variant of the indexed linked guaran teed investment certificates product. an ilgic is a retail savings vehicle, similar to a term deposit, with an interest rate that is linked to a diversified stock index. upon maturity of the product, the total return will be determined based on the performance of the underlying index over the pre-specified term. there are two basic structures of indexed linked gics.the capped ilgic allows the holder to participate in any upward movement in the underlying stock index, up to a pre-specified annual percentage cap. the p-ilgic allows the holder to participate in a fixed fraction of any upward movement in the underlying stock index, with no pre-specified cap. both products have an implicit guarantee of princi pal plus minimal interest. this study compares the relative value and appeal of the two basic ilgic products from the perspective of the individual investor. our valuation technique can be applied to any ilgic. from an empirical point of view, our main conclusions are that ilgics becomes less attractive, the longer the time horizon of the investor. in addition, partici pating ilgics are preferable to capped ilgics for short term maturities and vice versa for longer maturities. a detailed analysis of two particular canadian products was pro vided as an application of the basic concepts. we concluded by demonstrating that vpds range from 95% 98%, corresponding to a 3% 5% expense ratio on index linked gics. 282 financial services review 6(4) 1997 acknowledgment: this project was supported by a grant from the york university research authority. the authors would like to acknowledge helpful suggestions and comments from aron gottesman, karen lahey, the editor, and two anonymous reviewers. appendix derivation of valuation formulas with some algebra, we can transform the payoff from the capped-ilgic, in equa tion (l), to: 1 [ st -max --exp{(c-g,)t},o so 11 (8) or zexp{gct) + i [max[st-soexp{g,t},o] max[~t-so~xp~(c-g,)t~,oll (9) given a state-contingent payoff structure, we compute the no-arbitrage price for the ilgic using risk neutral pricing. zexp{(g, r)tl + 1 bs(so,soexp(g,t),t,q,r,o) -1 ws,,s,exp-l(c g,)tl,t,q,r,o) (10) where r is the risk free rate (by definition greater than the minimal guarantee gk) and bs(s, x, t, q, r, o) is the black and scholes price of a call option, defined by: bs(s, x, t, q, r, (t) = s exp{-qt}n(dl) -x exp{-rt}n{dz) (11) where in g + d, = [i ( ojt d, = ojt plugging back into equation (lo), simplifying, and then dividing by the initial investment i, we obtain the c-ilgic value as a percent ofpar, equal to: v,(c, g,) = exp{(g, r)t] + exp{--qt]nbl) exp((g, r)tw(w -exp{-qt}n(b$ + exp((c g, r)tw(b4) (12) the optimal choice of index-linked gics 283 where, +;)fi, 6, = (‘-;-“‘-;)fi (13) 6, = r-q-c-g, b, = r-q-c-g, (t d (14) next, let us examine the participating ilgic. with some algebra, we can transform the payoff from the p-ilgic, in equation (2), to: iexp(g,t) + ipmax st s 0 which simplifies to: exp{gpt170 1 (15) zexp{g,t) + zemax[s,-soexp{gpt},o]. so (16) once again, we compute the no-arbitrage price for the ilgic using risk neutral pric ing to obtain: lexp{(g, r)t1 + 1 bs(so,soexp(g,t},t,q,r,o) (17) and, finally we obtain the p-ilgic value as a percent ofpar, equal to: vp@, gp) = e&(g, r)t) + p exp{-qtw(al) -p exp((gp r)tw(a2) (18) where r-9-g a, = ( pp+; j?;, (r 1 a2 = ( r-q-g, 0 d ?)jt: (19) references baubonis, c., gastineau, g., & purcell, d. (1993). the banker’s guide to equity-linked certificates of deposit. the journal ofderivatives, (winter), 87-95. bell, a. (1997). investors buy into market-linked gics. the globe and mail, (april), c2. benninga, s., & blume, m. (1985). on the optimality of portfolio insurance. the jourmd offinance, #o(5), 1341-1352. black, f., & scholes, s. (1973). the pricing of options and corporate liabilities. the journal of polit ical economy, 81,637~653. brennan, m., & solanki, r. (1981). optimal portfolio insurance. journal of financial and quantita tive analysis, 16(3), 279-300. 284 financial services review 6(4) 1997 brooks, r. (1996). computing yields on enhanced cds. financial services review, 5(l), 3 l-42. cohn, j., & edelson, m. (1993). banking on the market: equity-linked cds. aaii journal, (march), 1 l-15. croft, r. (1997). use options to replicate indexed-linked gics. investment executive, (may), 39. merton, r.c. (1973). theory of rational option pricing. bell journal of economics and management science, 4, 141-183. schooley, d., & worden, d. (1996). risk aversion measures: comparing attitudes and asset alloca tion. financial services review, 5(2), 87-99. describing investor profiles: a test of the associations among financial knowledge, confidence, and help and information sources abed rabbania,*, john e. grableb, ann woodyardb, zheying yaoa adepartment of personal financial planning, university of missouri, 239 stanley hall, columbia, mo 65211, usa bdepartment of financial planning, housing and consumer economics, university of georgia, 205 dawson hall, athens, ga 30602, usa abstract faced with multiple asset choices for use when developing a household portfolio, investors often turn to various sources of help and information for help before making an investment decision. this study used a large, nationally drawn dataset of individuals who own financial assets to explore the relationships between and among types of investments owned, knowledge characteristics, investor confidence, and help and information sources. general profiles of investors emerged from the analyses that can be used by financial planners to better understand the unique profiles of those who own certain types of investments. investors who exhibited over-confidence in their financial knowledge were more likely to hold annuities, cash value life insurance, and commodities. it was also determined that financial planners play an important role in promoting diversification and mitigating portfolio risk. jel classification: d14; d81; d9; g11; g41 keywords: knowledge confidence; investments; decision-making; financial planning 1. introduction investors have access to various asset alternatives that can be used when developing a portfolio at the household level. obvious investment choices include stocks (equities), bonds * corresponding author. tel.: �1-573-882-9187; fax: �1-573-884-8389. e-mail address: rabbania@missouri.edu financial services review 27 (2018) 209-230 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. (fixed-income securities), mutual funds, and exchange traded funds (etfs). other asset alternatives include annuities, cash value life insurance, commodities, and real estate. the financial planning literature offers robust normative and descriptive explanations of the household characteristics associated with the ownership of certain investments. for example, those who own stocks tend to be wealthier, better educated, more risk tolerant, and more knowledgeable about the financial markets (kuzniak, rabbani, heo, ruiz-menjivar, and grable, 2015; wang and hanna, 2007). the characteristics of those who prefer bonds include general financial risk aversion, being older/retired, and having a fixed or limited time horizon (gibson, 2000). the literature indicates that there is considerable heterogeneity in investor profiles, yet the underlying sources of these differences are not well understood. there are a number of possible explanations for heterogeneity in investor profiles, such as variability in risk tolerance (see barsky, juster, kimball, and shapiro, 1997; kimball, sahm, & shapiro, 2008), barriers to entry that vary across households (vissing-jorgensen, 2002), and differences in the riskiness or magnitude of a household’s non-financial wealth, for example, human capital (guiso, jappelli, and terlizzese, 1996; heaton and lucas, 2000; vissingjorgensen, 2002). although, previous empirical research has found some evidence for each of these explanations, much of the observed variation in household characteristics is still unexplained. in addition, it is less well known, or described in the literature, if household characteristics are shared across investment types. one might imagine that the profile of a stock investor differs significantly from that of an annuity investor, or that those who own bonds are more likely to also own cash value life insurance. determining whether such assumptions are true and what may be the underlying explanation for differences (or similarities) is important in the context of financial planning. specifically, financial planners need insights into the profile of investors who are more or less likely to hold specific types of investment products. understanding how an investor’s profile relates to her or his investment preferences and choices is one way to more precisely match portfolio development needs to a client’s unique situation. over the past 20 years, two distinct lines of research have taken place to better understand the general profile and specific characteristics of investors. these studies have practical significance to the providers of financial services. one line of inquiry has revolved around understanding how preferences, attitudes, and behavioral and cognitive biases influence investment choices. much of this research has centered on the concept of investor confidence. consider the work of barber and odean (2001). they found that over-confident investors (typically men) trade excessively, usually to the detriment of portfolio diversification and performance. the notion of confidence encompasses a variety of variables, including knowledge, expectations, and preferences. as such, being able to identify states of underor over-confidence has emerged as a powerful financial planning tool to describe investor profiles. the hypothesis underlying the over-confidence model is that investors often over-estimate the value of private financial knowledge. this insight has led to the second line of inquiry into investor profiles: the role of information search and help-seeking behavior. existing help-seeking research has resulted in the development of several frameworks used to describe the profile of those who are more likely to use the services of a professional financial 210 a. rabbani et al. / financial services review 27 (2018) 209-230 advisor compared with other sources of help, including family, friends, colleagues, and the media. the outgrowth of this type of research has resulted in documentation that financial professionals add value by helping clients better implement recommendations that reduce tax liabilities (i.e., gamma) (blanchett and kaplan, 2013) and decrease wealth volatility (i.e., zeta) (grable and chatterjee, 2014). the source of help may partially explain the types of assets owned by investors. financial planners, as well as other financial advisors, have a vested interest in understanding how the source of financial help an investor uses to guide financial decision making can influence the preference for and use of specific investment products. it is reasonable to expect that the help-seeking behavior of investors moderates the effect of confidence on the preference for specific investment assets. curiously, very little work has been conducted to determine the synergistic roles of confidence and help and information sources as factors associated with an investor’s profile. the purpose of this study was to add to the existing literature in three ways: first, to develop profiles of investors across several asset classes; second, to estimate the association between overand under-confidence and asset ownership, and third, to describe the association between help seeking characteristics and ownership of specific investment assets. it was hypothesized that source of help might moderate the effect of being overor under-confident in relation to owning certain types of investments. the results from this study provide insights into understanding the heterogeneity among investors who hold various investments in their portfolios. this information can be used to more effectively anticipate the type of approach a financial planner ought to take when working with clients in relation to investment strategies and recommendations. 2. review of literature expected utility theory of choice under uncertainty assumes that investors are (1) rational, (2) able to deal with complex choices, (3) risk-averse, and (4) wealth-maximizing. utility theory predicts that a rational investor will select a portfolio that maximizes expected returns while minimizing risk. that is, at a given expected return, a rational investor will prefer a low-risk asset over a high-risk asset. however, this prediction does not discriminate between investors based on exhibited confidence. odean (1998) found that over-confident investors hold riskier portfolios than other investors with the same degree of risk aversion. standard economic theory also assumes that investors have complete information about the market. however, verrecchia (1982) showed that investors who face a higher cost of acquiring information, and those who are risk averse, acquire less information, often because these investors intend to invest less in stocks. this means that more informed investors are likely to invest in riskier assets, whereas less informed investors are likely to invest in less risky assets. the existing evidence suggests that over-confidence and excessive information search lead to suboptimal asset allocation choices. the following discussion provides an overview of investor confidence and help and information sources in the context of investment choices and outcomes. research on investor financial attitudes, demographics, and socioeconomic factors is also reviewed. 211a. rabbani et al. / financial services review 27 (2018) 209-230 2.1. over-confidence nearly all models of over-confidence are based on the premise that some investors exhibit a degree of over-confidence when making investment decisions. over-confident investors tend to trade more and take more risk than other investors, including those who exhibit under-confidence. odean (1998) showed that over-confident investors are more likely to obtain lower overall utility, often over reacting to new information. unlike other investors, over-confident investors overestimate the precision of their knowledge (odean, 1998). when viewed holistically, over-confident investors often to exhibit behaviors that negatively impact investment performance (barber and odean, 2001; bruine de bruin, parker, and fischhoff, 2007; statman, thorley, and vorkink, 2006). in addition to the observation that over-confidence leads to substandard portfolio performance, it is also known that overconfidence is associated with less diversification and higher portfolio turnover. gervais and odean (2001) found that over-confident traders trade too aggressively, which is often accompanied by increasing trading volume. moreover, over-confident traders behave suboptimally, eventually lowering their expected profits. additionally, higher trading volumes, at both the individual and market level, have been linked with over-confidence. bruine de bruin et al. (2007, 2012) found a relationship between better decision-making skills and fewer negative outcomes. a key variable in their study was confidence. bruine de bruin et al. (2007, 2012) reported a significant correlation between maximizing decision processes, partly by reducing over-confidence, and obtaining better investment outcomes. over-confidence has been shown to be associated with the gender of an investor. barber and odean (2001) reported that men trade more often and, therefore, perform less well than women. biais, hilton, mazurier, and pouget (2005) reached similar conclusions in an experimental setting where they found that problematic trading performance was related to over-confidence. over-confidence is known to be associated with other types of financial behavior as well. consumers who display over-confidence with regard to financial knowledge are more likely to engage in problematic credit behaviors (allgood and walstad, 2013; robb, babiarz, woodyard, and seay, 2015; woodyard, robb, babiarz, and jung, 2017), although exceptions have been noted in the literature. for example, investors with more confidence in their financial ability than actual knowledge may be more prone to pay off credit cards, which is a desirable outcome, but this tendency diminishes with age (allgood and walstad, 2013). robb and his associates (robb et al., 2015) noted that alternative financial services, such as payday loans, title loans, and rent-to-own arrangements, were more likely to be used by those (1) exhibiting an over-confidence condition, (2) having below median levels of objective financial knowledge, and (3) holding higher levels of subjective financial knowledge or confidence. over-confidence can be viewed as a “false-positive” condition in the context of financial decision making. woodyard et al. (2017) found that over-confident decision makers are less likely to engage in positive financial behaviors, such as having an emergency savings plan and always paying off a credit card in full at the end of the month. woodyard et al.’s work reinforced earlier findings from the literature on the connection between financial knowledge and positive financial behaviors (e.g., robb and woodyard, 2011). 212 a. rabbani et al. / financial services review 27 (2018) 209-230 in summary, the current literature shows that over-confidence often leads to substandard portfolio performance, less diversification, and higher portfolio turnover. moreover, overconfident investors trade more and take more risk. however, little is known about the association between dimensions of confidence and investment asset holdings. based on the review of literature, it is reasonable to hypothesize that over-confident investors are likely to hold more risky assets. the current study evaluates this possibility by testing the following hypothesis: h1: there is a direct relationship between confidence level and investment asset holdings. 2.2. help and information search behavior investors are generally counseled to search for additional information when doing so will increase expected utility (grossman and stiglitz, 1980). over-confident investors take such advice to extremes. they tend to hold unrealistic beliefs about their knowledge and expend too many resources (e.g., time and money) on investment information (odean, 1998), typically as a means to confirm previously established opinions. barber and odean (2001) found support for the prediction that investors who possess too much information may often trade speculatively. the concept of household level help and information seeking has been extensively examined in the financial planning literature over the past two decades. grable and joo (1999) were among the first to propose a framework of help-seeking behavior in personal financial planning. the grable and joo framework was based on a help-seeking for health care model originally described by suchman (1966). the grable and joo framework described six-stages of help seeking based on the hypothesis that consumers conduct internal cost/benefit analyses when deciding to seek help and where to seek help. grable and joo originally focused on a simple decision to either seek help or not to seek help. later, grable and joo (2001) expanded the model to examine the choice of seeking help from a financial professional or a non-professional. those who indicated the use of financial planners, financial counselors, insurance agents, or stockbrokers as their primary help provider were classified as seekers of professional help, while those who indicated the use of friends, family, or work colleagues as their primary help provider were classified as seekers of non-professional help. kwon (2004) identified four types of help and information sources based on whether the source is personal or impersonal, and whether the source is marketer-controlled, such as a salesperson with potential personal gains to be made from a transaction or non-personal, such as family, friends, and other neutral sources (e.g., news media). kwon noted that friends and family were the most frequently used source of help and information, which matched what grable and joo (2001) reported. chang (2005) found that income was a major determinant of financial help seeking, in that higher-income households were more likely to seek professional advice while lower-income households relied upon social networks, such as family or co-workers. closely related to the concept of help seeking is the concept of information search. information search tends to be broader than help seeking. whereas models of help seeking 213a. rabbani et al. / financial services review 27 (2018) 209-230 are typically designed to provide insights into the mechanisms associated with the choice and use of help providers, information search behavior includes elements related to activities designed to improve knowledge. the term “information intermediary” can refer to the media, to advertising, to a search engine, or any other non-human source of information that is not originally produced by the provider (rose, 1999). the use of information intermediaries helps consumers save time when making decisions or when acquiring new/additional information. it is not necessary for an information intermediary to provide direct help. lee and cho (2005) evaluated the use of information intermediaries based on value perceptions. they concluded that value can be attached to differing characteristics. the model developed by lee and cho did not account for the actual cost of intermediary services (e.g., using an internet search vs. consulting a financial advisor), but the model did help lee and cho conclude that a large portion of the public indicate an unwillingness to pay for financial advice. in their model, younger consumers reported a higher value for intermediary sources, perhaps reflecting a lack of confidence or knowledge among younger consumers. lee and cho also found higher levels of education were associated with enhanced perceptions of value from information intermediaries. lee and cho reported that perceived expertise was negatively associated with perceived value, implying that consumer information search activities may be indicative of a defensive posture related to confidence in the financial markets. the literature provides some evidence that there may be parallels in the search for health information with that of financial information search behavior (grable and joo, 1999; woodyard and grable, 2014). chen and feeley (2014) found that the relationship between numeracy and self-efficacy in health behavior was mediated by information seeking behavior through multiple channels, including mass media, professional providers, friends, and family members. numeracy, which is often considered a component of financial knowledge (robb and woodyard, 2011), has also been found to be related to risk tolerance (sages and grable, 2010) and other factors that are known to be related to financial literacy (lusardi, 2012) and help seeking. in this regard, hung and yoong (2010) found that solicited, or sought, advice was more effective in changing consumers’ retirement plan behaviors than unsolicited advice. in summary, the existing literature shows that some investors, particularly those who exhibit over-confidence, often make aggressive investment decisions and trade speculatively. however, little is known about the way in which help and information sources relate to investment choices. investment holdings likely vary based on the type of help and information source used by an investor. the current study evaluates this possibility by testing the following hypothesis: h2: there is a direct relationship between help and information source and investment asset holdings. 2.3. other factors associated with investment asset holdings based on the literature, some demographic factors are known to be associated with asset allocation choices. stock holdings are generally thought to proportionally increase with age (ameriks and zeldes, 2004). additionally, the likelihood of under-diversification is greater among younger, less-educated, and low-income investors (goetzmann and kumar, 2008), 214 a. rabbani et al. / financial services review 27 (2018) 209-230 whereas higher income tends to be associated with holding more diversified portfolios (anderson, 2013; calvet, campbell, and sodini, 2007; roche, tompaidis, and yang, 2013). it is worth noting that some research shows that housing crowds out stockholding (cocco, 2005; hu, 2005). additionally, business owners often exhibit risky asset ownership tendencies (faig and shum, 2002; xiao, 1996). gender differences in financial knowledge, satisfaction, and asset choices have been well documented in the literature. woodyard and robb (2012), for example, found a consistent gender gap across most age groups with regards to objective financial knowledge, although less so for subjective knowledge, financial satisfaction, and participation in actual financial behavior. studies of college students (e.g., chen and volpe, 2002) and older americans (e.g., lusardi and mitchell, 2007) have also indicated a disparity in financial literacy by gender. household specialization has been offered as an explanation for these differences (fonseca, mullen, zamarro, and zissimopoulos, 2012). demographic and socioeconomic characteristics are also thought to be associated with a wide range of decision making outcomes. for example, age (younger), gender (male), education (more), and wealth (more) are known to be associated with financial risk tolerance generally and investment decision making specifically (barber and odean, 2001; duasa and yusof, 2013). those with a higher risk tolerance have been shown to be more likely to seek professional financial help (joo and grable, 2001). hanna (2011) reinforced this finding by noting that those who are financially less tolerant of risk are less likely to seek financial advice from professionals. as highlighted in this review, the identification of household characteristics associated with investor confidence, help seeking, and ownership of investments is robust, with each topic area having a well-defined understanding of important factors. however, a gap exists in the literature. specifically, little is known about the profile of investors across asset categories or whether characteristics such as confidence, help seeking preferences, and attitudes are the same or different among those who hold various investment assets. what is known is that older investors tend to hold more fixed-income assets compared with others, and that, overall, risk averse investors also hold less risky assets. while these types of descriptions are quite useful, they lack the depth to be able to identify nuances between and among investors. more specifically, the literature is relatively silent when it comes to describing how investor confidence and the use of sources of help jointly describe patterns of asset ownership. while it is generally acknowledged that an investor’s confidence level, source of decision making help, and risk tolerance are associated with behavioral outcomes, few studies have tested the moderating effects of these constructs. the literature suggests that those who exhibit over-confidence may seek more and varied information, which can then result in different investment asset ownership decisions. models of over-confidence and help-seeking emphasize that help-seeking behavior can have a moderating role on the relationship between confidence level and investment asset holdings, whereas the risk tolerance literature provides a basis for suggesting that an investor’s willingness to take financial risk may also play a moderating role.1 if the type of help and information source moderates the relationship between confidence level and investment asset holdings, then the strength of the relationship between confidence and investment asset holding should differ based on the type of help and 215a. rabbani et al. / financial services review 27 (2018) 209-230 information source. for example, the relation between confidence and investment asset holdings might be stronger for investors who receive help and information from a professional financial advisor and less strong or non-existent for investors who receive help and information from television. the current study evaluates this possibility by testing the following hypotheses: h3: the relationship between confidence level and investment asset holdings is moderated by help and information source. h4: the relationship between confidence level and investment asset holdings is moderated by risk tolerance. 3. methodology data for this study were obtained from the 2015 national financial capability study (nfcs), which was funded by the finra investor education foundation. the sample was comprised of adults over the age of 18 who also completed the 2015 state-by-state nfcs survey and indicated that they had investments outside of retirement accounts. this delimitation decision was made because investing outside of a retirement account (e.g., taxable brokerage account) involves a different set of decisions on the part of the investor. for example, investors in non-retirement accounts generally must choose between and among asset choices, whereas retirement plan choices are usually predetermined by a plan administrator (and often limited to mutual fund investments). for the purposes of this study, the sample was also delimited to those who were the primary or shared decision-maker regarding investments for the household. the sample for the survey was a subset of the larger state-by-state survey and selected at random from three online panels. these panels were designed to validate the inclusion of respondents based on current demographic characteristics. the survey was self-administered using a website link in july 2015. respondents were not told that the investor survey and the state-by-state survey were related. findings from the survey were weighted to approximate the investor population in terms of age and education, based on the 2015 nfcs state-by-state survey. no additional weighting was used to account for non-response bias. given missing data, the final sample size used in the study included 1,987 individuals. 3.1. dependent variables the dependent measure was a binary variable that indicated whether investors currently owned any of the following investments in non-retirement accounts: (1) individual stocks; (2) individual bonds; (3) mutual funds; (4) exchange traded funds; (5) fixed, indexed, or variable annuities; (6) whole life insurance or similar investment products; (7) commodities or futures; and (8) other investments such as reits, options, private placements, or structured notes. answers were coded 1 � yes, 0 � no for each investment holding. it was possible for a respondent to indicate owning more than one type of asset; however, everyone in the sample owned at least one of the eight investment types. 216 a. rabbani et al. / financial services review 27 (2018) 209-230 3.2. independent variables the independent variables included confidence level, financial risk tolerance, demographic characteristics, financial education, student loans, financial stress, and financial self-efficacy. the following discussion describes each independent variable in greater detail. 3.2.1. confidence level a multiple step process was used to calculate a respondent’s level of financial confidence. first, each respondent’s objective investment knowledge was estimated using 10 questions (see appendix) developed by the finra investor education foundation. the sum of correct answers to these questions served as an indicator of respondents’ objective financial knowledge. the mean objective financial knowledge score was 4.66 (sd � 2.23). next, a measure of each respondent’s subjective evaluation of her or his financial knowledge was evaluated with the following question: “on a scale of 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall knowledge about investing?” the mean score was 4.86 (sd � 1.39). objective financial knowledge scores were then used to predict subjective knowledge scores in an ordinary least square (ols) regression as follows: subjective knowledge score � b0 � b1 objective knowledge score the regression model was statistically significant (f(1, 1986) � 125.22, p � 0.001). the predicted unstandardized value (i.e., predicted knowledge score) was saved for each respondent. the predicted mean score, across the sample, was 4.85 (sd � 0.34). next, the confidence level of each respondent was estimated by subtracting the predicted subjective financial knowledge score from his or her self-reported subjective financial knowledge score. in this study, positive scores indicated knowledge over-confidence, with negative scores suggesting knowledge under-confidence. as expected, the mean confidence score was 0.00 (termed congruent in this study), with a standard deviation of 1.35. the minimum score was �4.21, while the maximum score was 2.86, indicating a general level of under-confidence among respondents. confidence level scores were then recoded dichotomously by classifying anyone whose knowledge confidence score was one standard deviation or higher above the mean as being over-confident (coded 1, otherwise 0). those whose score was one standard deviation or lower below the mean were categorized as under-confident (coded 1, otherwise 0). approximately 68% of respondents were classified as being congruent in their knowledge assessment, with 16% classified as over-confident and 16% categorized as under-confident. 3.2.2. help and information sources four help and information source categories were created. the categories were based on responses to the following question: “which of the following information sources do you use when making an investment decision?” nine possible answers were provided: (1) stockbro217a. rabbani et al. / financial services review 27 (2018) 209-230 kers; (2) financial advisors; (3) information from the company the respondent was investing in (e.g., annual reports, company websites); (4) information from brokerage firms, mutual fund companies, or other financial services companies (e.g., analyst reports, brochures, newsletters, seminars, and websites); (5) industry regulators (e.g., finra, sec, and state securities regulators); (6) “media,” which included television, radio, newspapers, magazines, online news sources, and financial information websites; (7) investment clubs or investor membership organizations; (8) employers; and (9) friends, colleagues, or family members. it was possible for a respondent to indicate the use of more than one help provider. these sources of help and information were re-categorized. the first help-seeking category was called “actively sought help at a cost.” this included the use of (1) stockbrokers and (2) financial advisors. the second help-seeking category was termed “actively sought help for free.” this category included (1) information from the company the respondent was investing in (e.g., annual reports, company websites); (2) information from brokerage firms, mutual fund companies, or other financial services companies (e.g., analyst reports, brochures, newsletters, seminars, and websites); and (3) industry regulators (e.g., finra, sec, and state securities regulators). the third category was called the media, which included television, radio, newspapers, magazines, online news sources, and financial information websites. the fourth help-seeking category was called “networks.” this category included (1) investment clubs or investor membership organizations; (2) employers; and (3) friends, colleagues, or family members. 3.2.3. financial risk tolerance financial risk tolerance was measured with one item that was adapted from the survey of consumer finances: “which of the following statements comes closest to describing the amount of financial risk that you are willing to take when you save or make investments?” four options were provided: 1 � take substantial financial risks expecting to earn substantial returns; 2 � take above average financial risks expecting to earn above average returns; 3 � take average financial risks expecting to earn average returns; and 4 � not willing to take any financial risks. given the coding of responses, a dichotomous variable was created with 1 � willing to take some risk (i.e., answers 1, 2, and 3 combined), otherwise 0. 3.2.4. confidence in the security markets two confidence items regarding the security markets were included in the analysis. the first asked, “how confident are you that u.s. financial markets offer good long-term opportunities for investors?” the second asked, “how confident are you that u.s. financial markets are fair to all investors?” a 10-point scale was used to record answers to both items, with 1 � not at all confident and 10 � extremely confident. the mean and standard deviation score for each item was 7.05 (sd � 2.03) and 5.81 (sd � 2.47), respectively. 3.2.5. control variables based on the review of literature, several additional control variables were used in the analysis, including gender, age, ethnicity, education, income level, and value of investments. a summary of the descriptive statistics for the independent variables is shown in table 1. 218 a. rabbani et al. / financial services review 27 (2018) 209-230 table 1 descriptive statistics for the independent variables variables distribution mean (sd) gender male 55.1% female 45.0% age 18 to 34 years 16.2% 35 to 54 years 31.6% 55� years 52.3% ethnicity white 80.3% non-white 19.7% education some college or less 39.0% college (bachelor’s) or more 61.0% income � $50,000 21.0% $50,000 to $100,000 44.7% $100,000� 34.4% value of investments less than $2,000 5.1% $2,000 to less than $5,000 4.3% $5,000 to less than $10,000 5.8% $10,000 to less than $25,000 7.3% $25,000 to less than $50,000 8.1% $50,000 to less than $100,000 15.0% $100,000 to less than $250,000 19.9% $250,000 to less than $500,000 15.9% $500,000 to less than $1,000,000 10.5% $1,000,000 or more 8.0% risk tolerance some risk 90.5% no risk 9.5% long-term market confidence m � 7.05 (sd � 2.03) confidence in fairness of markets m � 5.81 (sd � 2.47) over-confident yes 16.0% no 84.0% under-confident yes 15.6% no 84.4% active fee help (ac) yes 67.9% no 32.2% active free help (af) yes 85.1% no 14.9% media help yes 47.8% no 52.2% network help yes 57.4% no 42.6% 219a. rabbani et al. / financial services review 27 (2018) 209-230 3.2.6. interaction terms finally, interaction terms were created to account for possible moderation effects between overand under-confidence and financial risk tolerance and help and information sources. each variable was centered before creating interaction terms. 3.3. data analysis method eight logistic regression models were estimated. the outcome variable in each model was whether a respondent owned an investment asset (i.e., individual stocks, individual bonds, mutual funds, etfs, annuities, cash value life insurance, commodities, or other investment assets). the operationalized models were estimated using the following formula: p � �ea�bdxd�bvxv�brxr�bcxc�boxo�buxu�bhxh�bixi� �1 � ea�bdxd�bvxv�brxr�bcxc�boxo�buxu�bhxh�bixi� (1) where: p � the probability of asset ownership, e � the base of natural logarithms, a � the equation constant, bs � beta coefficients xs � set of independent variables (d � demographic variables, v � value of investments, r � risk tolerance, c � market confidence, o � over-confidence, u � under-confidence, h � help-seeking sources, i � interactions) 4. results table 2 provides a summary of results from the eight logistic regression analyses. each of the models was statistically significant, with the amount of explained variation in the outcome variables ranging from approximately 11% to 23%. the one consistent factor associated with investment asset ownership across the models was the value of a respondent’s investments. in each case, the relationship was positive. this result was not surprising given who was in the sample; namely, individuals who indicated some level of asset ownership. a weak endogeneity issue may have been present within the sample. to be included in the study, individuals already exhibited asset ownership tendencies. it is reasonable to assume, based on the results from table 2, that there was a direct association between owning investment assets and holding more wealth in such assets. given the purposes of this study, the directional causality issue was not deemed to be a significant problem. the endogeneity issue indicates what is already known: a key element associated with the profile of those who own specific types of investment assets is a greater likelihood of holding more wealth in investments. instead, this study provides more nuanced indicators of investment asset ownership. results from each regression are discussed below. 220 a. rabbani et al. / financial services review 27 (2018) 209-230 t ab le 2 l og is tic re gr es si on an al ys es sh ow in g pa tte rn s of in ve st m en t as se t ow ne rs hi p v ar ia bl es st oc ks b on ds fu nd s e t fs a nn ui tie s c v lif e c om m od iti es o th er b e xp (b ) b e xp (b ) b e xp (b ) b e xp (b ) b e xp (b ) b e xp (b ) b e xp (b ) b e xp (b ) g en de r (1 � m , 2 � f) � .0 1 .9 9 .3 4* * 1. 41 .0 3 1. 03 � .0 5 .9 5 � .0 1 .9 9 � .1 5 .8 6 � .2 0 .8 2 � .1 9 .8 3 a ge � .0 4 .9 6 � .3 7* * .6 9 � .2 2* .8 0 � .4 1* * .6 6 .0 2 1. 02 � .0 3 .9 7 � .6 0* * .5 5 � .0 6 .9 4 e th ni ci ty (1 � w , 2 � n w ) .0 6 1. 06 .1 1 1. 12 � .1 9 .8 3 .1 9 1. 21 .1 8 1. 20 .1 9 1. 21 .2 7 1. 31 .2 7 1. 31 e du ca tio n (1 � so m e co lle ge or le ss , 2 � co lle ge or m or e) .0 0 1. 00 � .1 7 .8 5 .1 9 1. 21 .1 9 1. 21 � .2 6* .7 7 � .2 8* .7 6 .0 9 1. 10 � .0 3 .9 7 in co m e .1 4* * 1. 15 � .0 5 .9 6 � .0 5 .9 5 � .0 2 .9 8 � .2 0* .8 2 .1 5* 1. 16 � .1 2 .8 9 � .1 5 .8 6 v al ue of in ve st m en ts .0 8* * 1. 09 .2 6* * 1. 30 .3 2* * 1. 38 .1 9* * 1. 21 .2 7* * 1. 31 .1 0* * 1. 10 .1 8* * 1. 19 .2 3* * 1. 26 r is k to le ra nc e (1 � so m e ri sk , 0 � no ri sk ) .7 5* * 2. 11 � .0 8 .9 2 .5 9* * 1. 81 4. 36 77 .9 2 � .2 1 .8 1 � .2 0 .8 2 .2 2 1. 25 .2 9 1. 33 l on gte rm m ar ke t co nfi de nc e .2 2 1. 24 .0 4 1. 04 � .1 0* .9 1 .0 0 1. 00 .0 6 1. 06 .1 0* 1. 11 .0 5 1. 05 � .1 0 .9 1 c on fid en ce in fa ir ne ss of m ar ke ts .0 4 1. 04 .1 3 1. 14 .1 8 1. 19 .3 2* * 1. 37 � .2 1* .8 2 � .1 7* .8 4 .2 2 1. 25 .4 3* * 1. 54 o ve rco nfi de nt � .1 9 .8 3 .2 5 1. 29 � .4 8* .6 2 .0 1 1. 01 .6 4* * 1. 89 .5 4* * 1. 71 .5 9* 1. 80 � .1 0 .9 0 u nd er -c on fid en t .6 4 1. 89 .7 5 2. 11 � .1 0 .9 0 � 17 .2 0 .0 0 � .2 8 .7 6 � .3 9 .6 8 � .7 8 .4 6 .1 5 1. 16 a ct iv e fe e he lp � .0 7 .9 3 .6 6* * 1. 94 .6 3* * 1. 87 .3 1* 1. 37 .9 9 2. 70 .5 4* * 1. 71 .5 0* 1. 66 .0 5 1. 06 a ct iv e fr ee he lp .2 3 1. 26 .4 4* 1. 56 .2 4 1. 27 � .1 4 .8 7 � .1 8 .8 4 .3 4 1. 41 � .1 2 .8 9 � .1 2 .8 9 m ed ia he lp .3 2* 1. 38 .2 3* 1. 26 � .1 0 .9 1 .3 2* 1. 38 � .0 3 .9 7 .0 4 1. 04 .2 6 1. 29 .3 4* 1. 40 n et w or k he lp .0 8 1. 08 .1 9 1. 21 � .0 9 .9 1 � .1 6 .8 5 .3 9* * 1. 48 .4 4 1. 56 .4 5* 1. 57 .0 1 1. 01 r is kx o ve rc on f .6 9 1. 98 � .1 6 .8 5 � .0 6 .9 4 .4 7 1. 60 .2 0 1. 22 � .3 6 .7 0 .4 5 1. 57 .0 4 1. 04 r is kx u nd er c on f .0 5 1. 05 � .1 0 .9 1 .2 0 1. 22 17 .2 1 2. 16 � .2 5 .7 8 .0 0 1. 00 .7 8 2. 17 � .2 5 .7 8 a fe ex o ve rc on f 1. 17 ** 3. 23 .1 5 1. 16 .2 4 1. 27 � .1 1 .8 9 � .4 5 .6 4 .1 2 1. 12 .2 7 1. 31 .1 8 1. 19 a fr ee xo ve rc on f .4 4 1. 55 .3 2 1. 37 � .0 6 .9 4 � .3 1 .7 4 1. 09 * 2. 97 .3 4 1. 41 � .0 9 .9 2 .5 2 1. 69 m ed ia xo ve rc on f .2 1 1. 24 .2 2 1. 25 .3 5 1. 41 .3 6 1. 44 .2 5 1. 28 .8 4* * 2. 32 .4 0 1. 50 1. 20 ** 3. 31 n et w or ks xo ve rc on f � .8 8* .4 2 .8 3* 2. 30 .0 6 1. 06 1. 02 ** 2. 76 .5 0 1. 65 .7 8* 2. 19 .1 9 1. 21 � .2 9 .7 5 a fe ex u nd er c on f � .6 3 .5 3 � .2 2 .8 0 � .3 4 .7 1 � .5 7 .5 7 .0 6 1. 06 � .5 3 .5 9 � 2. 36 ** .0 9 .2 5 1. 28 a fr ee xu nd er c on f .6 6 1. 93 .2 4 1. 27 � .3 6 .7 0 .1 5 1. 16 .3 8 1. 47 � .8 7* .4 2 .8 1 2. 26 � .1 4 .8 7 m ed ia xu nd er c on f � .0 4 .9 6 .1 1 1. 11 � .4 7 .6 3 � .0 8 .9 2 � .0 9 .9 1 .4 7 1. 59 � .3 0 .7 4 .2 5 1. 28 n et w or ks xu nd er c on f � .9 7* .3 8 � .1 7 .8 5 .4 1 1. 51 � .7 6 .4 7 � .1 6 .8 5 .1 0 1. 11 1. 20 3. 32 .2 1 1. 24 c on st an t � 2. 34 ** .1 0 � 3. 73 ** .0 2 � 1. 84 ** .1 6 � 5. 37 ** .0 1 � 1. 94 ** .1 4 � 1. 47 ** .2 3 � 4. 35 ** .0 1 � 4. 70 ** .0 1 e t fs � ex ch an ge tr ad ed fu nd s; e xp � ex pe ri m en t. m od el st at is tic s— st oc ks : � 2 � 14 4. 28 ,p � .0 01 ;n ag el ke rk e r 2 � 0. 11 ;b on ds :� 2 � 27 0. 30 ,p � .0 01 ;n ag el ke rk e r 2 � .1 9; fu nd s: � 2 � 28 8. 81 ,p � .0 01 ; n ag el ke rk e r 2 � .2 1; e t fs : � 2 � 26 3. 42 ,p � .0 01 ; n ag el ke rk e r 2 � 0. 21 ; a nn ui tie s: � 2 � 25 3. 34 ,p � .0 01 ; n ag el ke rk e r 2 � .1 8; c as h v al ue l if e in su ra nc e: � 2 � 17 0. 90 , p � .0 01 ; n ag el ke rk e r 2 � .1 2; c om m od iti es : � 2 � 22 8. 96 , p � .0 01 ; n ag el ke rk e r 2 � .2 3; o th er : � 2 � 14 5. 93 , p � .0 01 ; n ag el ke rk e r 2 � .1 4. *p � .0 5, ** p � .0 1. 221a. rabbani et al. / financial services review 27 (2018) 209-230 4.1. individual stocks three confidence interactions were noted. those who were over-confident and sought help from a stockbroker or financial advisor were more likely to report owning stocks. on the other hand, over-confident respondents who sought help from networks were less likely to own stocks. confidence in one’s own financial knowledge was not directly associated with individual stock ownership. instead, confidence moderated by a help provider was related to holding stock. income, the value of investments, risk tolerance, and seeking help from the media were found to be positively associated with individual stock ownership. 4.2. individual bonds one interaction was noted. over-confident respondents who obtained help and information from networks were more likely to report owning individual bonds. the results indicate that sources of help are significant factors associated with bond ownership. the value of investments, seeking help from a stockbroker or financial advisor, seeking active free help, and obtaining information from the media were found to increase the likelihood of owning individual bonds. age was negatively related to the likelihood of owning bonds. 4.3. mutual funds no interactions were noted. the findings suggest that mutual fund ownership offers less appeal to investors who are over-confident in their own knowledge and confident in the future of the investment markets. the value of investments, risk tolerance, and working with a stockbroker or financial advisor were found to increase the likelihood of owning mutual funds. age, having greater confidence in the markets, and being over-confident were associated with a decreased likelihood of owning mutual funds. it may be that mutual fund investors select other investment assets unless otherwise counseled by a stockbroker or financial advisor. 4.4. exchange traded funds a positive interaction was noted between being over-confident and seeking help from a network. those fitting this profile were 2.76 times more likely to report owning etfs. similar to individual stock ownership, the profile of etf investors included younger confident investors. sources of help also increased the likelihood of owning etfs. the value of investments, greater confidence in the fairness of the markets, and seeking help from a stockbroker or financial advisor were found to be related to an increased likelihood of owning etfs. the relationship was negative for age. 4.5. annuities a positive interaction between being over-confident and seeking help from free sources was found to be associated with an increased likelihood of annuity ownership. results indicate that the profile of annuity investors is nuanced, with higher income and better 222 a. rabbani et al. / financial services review 27 (2018) 209-230 educated investors shying away from annuities, and over-confident investors purchasing annuities. the value of investments, being over-confident, and seeking help from networks was found to be associated with an increased likelihood of owning annuities. having at least a college level of education, more income, and greater confidence in the markets was negatively related to annuity ownership. 4.6. cash value life insurance three interactions were noted. being over-confident and seeking help from the media and networks were related to an increased likelihood of insurance ownership, whereas being under-confident and seeking help from free sources was negatively associated with insurance holdings. these findings indicate that confidence and sources of help can be used to describe the profile of life insurance investors. as a sold product, it was not surprising that working with a financial professional was positively associated with annuity ownership. income, the value of investments, long-term market confidence, being over-confident, and seeking help from stockbrokers and financial advisors were positively associated with the likelihood of owning cash value life insurance. a negative association was noted between education and confidence in the fairness of the markets and cash value life insurance holdings. 4.7. commodities and futures a negative interaction between being under-confident and seeking help from a stockbroker or financial advisor was noted. findings suggest that those who invest in commodities and futures tend to be younger over-confident individuals who work with paid financial help providers. the value of investments, being over-confident, seeking help from stockbrokers and financial advisors, and seeking help from networks were found to be positively associated with an increased likelihood of commodity and futures holdings. the relationship between age and owning commodities and futures was negative. 4.8. other investments a positive interaction between being over-confident and seeking help from the media was noted. those fitting this profile were 3.31 times more likely to report holding other investment assets. these findings indicate that measures of confidence and sources of help describe, in part, the profile of investors in other assets. the value of investments, being confident in the fairness of the markets, and seeking help from the media were also associated with an increased likelihood of holding other assets, including reits. 5. discussion table 3 summarizes the findings from the eight models. partial support was found for the first hypothesis that stated there is a direct relationship between confidence level and investment asset holdings. those who were over-confident held more annuities, cash value 223a. rabbani et al. / financial services review 27 (2018) 209-230 life insurance, and commodities. however, over-confident investors held fewer mutual fund assets. in general, support was found for the second hypothesis that stated there is a direct relationship between help and information source and investment asset holdings. investors who used active fee help were more likely to owner bonds, mutual funds, etfs, life insurance, and commodities. bond ownership was associated with the use of free help sources. those who relied on the media held more stocks, bonds, and commodities, whereas those who used networks owned more annuities and commodities. some support was also found for the hypothesis that posited a moderation effect between confidence level and help and information source, with those who (1) used network information sources and (2) exhibited over-confidence owning more bonds, etfs, and cash value life insurance but few stocks. among other associations, over-confident investors who used active fee help providers owned more stock. no support was found for the hypothesis that stated the relationship between confidence level and investment asset holdings is moderated by risk tolerance. as noted above, it was not surprising that the value of investments exhibited a strong positive association with investment asset ownership. this relationship can be viewed in a number of ways. for example, it takes financial resources to purchase investments outside a retirement plan. as such, it makes sense that those with more investment assets would hold a wider variety of investments. also, this finding highlights the relationship between holding wealth in investments and diversification. those with more wealth, either out of prudence or through serendipity, tend to be better diversified than those with fewer investable resources. demographic factors were less consistent in describing the profile of investors. age was negatively associated with securities ownership across four categories of assets: bonds, mutual funds, etfs, and commodities. education and income were negatively related to owning annuities. the education effect was negative for cash value life insurance, whereas income was positively associated with owning life insurance. it may be that cash value life table 3 variables associated with the likelihood of owning different investment assets stocks bonds funds etfs annuities cv life commodities other age � � � � education � � income � � � value of investments � � � � � � � � risk tolerance � � long-term market confidence � � confidence in fairness of markets � � � � over-confidence (knowledge) � � � � active fee help � � � � � active free help � media help � � � network help � � afeexoverconf � afreexoverconf � mediaxoverconf � � networksxoverconf � � � � afeexunderconf � afreexunderconf � networksxunderconf � 224 a. rabbani et al. / financial services review 27 (2018) 209-230 insurance is used by high income households as a tax-deferral tool. it is also possible that cash value life insurance is being used, as designed, less as an investment and more as a way to hedge the unanticipated loss of income resulting from death for high income earners. surprisingly, risk tolerance was found to be associated with the probability of security ownership in only two cases: stocks and mutual funds. the relationship was positive, suggesting that those who were willing to take some financial risk were more likely to own these assets. risk tolerance was not, however, associated with the ownership of other assets. confidence in the markets was generally not a significant variable in the models; however, those who were confident in the markets were less likely to own mutual funds. on the other hand, those who reported confidence in the markets were more likely to own cash value life insurance. confidence in the fairness of the markets was shown to have a greater association with the likelihood of owning investments. the relationship with owing etfs and other investments was positive. the association with owning annuities and cash value life insurance was negative. one explanation for this finding is that those with greater confidence in the fairness of the markets might value the security offered by annuities and life insurance less than the returns offered by etfs and other investments, including reits. the findings related to financial knowledge confidence were mixed. being underconfident was not significantly associated with ownership in any of the investments. some over-confidence effects were noted. those who were over-confident were less likely to hold mutual funds. over-confident investors were more likely, on the other hand, to own annuities, cash value life insurance, and commodities. as stand-alone variables, underand over-confidence showed less importance in the models. the most noteworthy findings related to confidence were in relation to interactions with help-seeking behavior. in fact, sources of investment help were dominant in describing the profile of investors. for five out of the eight investments (i.e., individual bonds, mutual funds, etfs, cash value life insurance, and commodities), seeking help from a stockbroker and other investment advisor was positively related to the likelihood of owning these investments. this indicates that among investors, these types of investments are a “sold” product, meaning that in addition to other factors, it takes a paid help provider to promote the use of some investments, although it is also possible that those who own these types of investments seek out help from professionals to help manage the investments. actively sought free help was found to be associated with only one investment asset: individual bonds. a positive association with seeking help from the media and the likelihood of owning stocks, bonds, and other investments was noted. advice from networks was positively related to the probability of owning annuities and commodities. the most noteworthy results were related to the interactions between and among knowledge confidence and source of investment help. consider the situation with individual stock ownership. by itself, being over-confident was not associated with the probability of owning stock. however, when over-confident investors worked with a stockbroker or other financial advisor, the likelihood of owning stocks increased. this hints at the possibility that compared with those with a congruent self-assessment of financial knowledge, over-confident investors are more likely to own individual stocks after working with a financial professional. over-confidence was also positively associated with seeking help from free sources (i.e., annuities), the media (i.e., cash value life insurance and other investments), and networks 225a. rabbani et al. / financial services review 27 (2018) 209-230 (i.e., individual bonds, etfs, and cash value life insurance). recall that the network category included (1) investment clubs or investor membership organizations, (2) employers, and (3) friends, colleagues, or family members. these help providers may serve as a source of confirmation, whereas free and media sources of help may act as promoters of certain investments and behaviors (e.g., purchasing life insurance and other investments, such as hard assets, including real estate and precious metals). only one negative interaction was noted in relation to being over-confident: obtaining financial help from networks decreased the likelihood of owning individual stocks. network help providers may dampen demand for stocks among those who are over-confident. rather than investing in individual stocks, those who were over-confident and working with network help providers were more likely to own other types of risk-bearing investments, such as etfs. whereas no direct under-confidence relationships were found in the models, three interactions were present. those who were under-confident and working with a stockbroker or financial advisor were less likely to own commodities. under-confident investors who obtained free advice were less likely to hold cash value life insurance. finally, the likelihood of owning individual stocks was lower for under-confident investors working with network help providers. 6. implications for financial professionals in an ideal world, all financial professionals will seek to improve the financial health of clients by providing cost-effective holistic financial plans that meet client goals and promote long-term financial well-being. effective communication with clients is essential within the financial planning process (grable and goetz, 2017), but so is the ability to recognize the unique characteristics and profile of clients that can be used to anticipate investment preferences. consider the issue of financial risk tolerance. although generally thought to be a primary determinant of the type of investments held by investors, in this study, financial risk tolerance was only associated with the ownership of stocks and mutual funds. other household characteristics were more important in describing an investor’s profile. investors who exhibited over-confidence in their financial knowledge were more likely to hold annuities, cash value life insurance, and commodities. over-confident respondents were also less likely to hold mutual funds. even so, sources of help emerged as a variable of even more relevance. active fee help providers, which includes financial planners, was found to be important in describing who held bonds, mutual funds, etfs, cash value life insurance, and commodities. in general, the use of financial planners, and other fee-for-service help providers, appears to promote diversification among investments. the use of media help providers, on the other hand, may create a mindset that focuses on individual holdings of stocks, bonds, and real estate. the powerful role of help providers was exemplified by the way in which investment ownership was moderated by the type of help provider used by an investor. those who worked with a fee-for-service help provider exhibited stock ownership. this hints at the positive way in which paid help providers can moderate a client’s tendencies. in other words, over-confident clients who seek help from financial planners may benefit from advice to 226 a. rabbani et al. / financial services review 27 (2018) 209-230 diversify investments across asset classes. moderation effects with various types of advice platforms suggests a certain “coachability” associated with over-confidence in investing. the association between network help providers and over-confident investors provides additional evidence regarding the power of help seeking. network help providers significantly moderated the effect of being over-confident, but not necessarily in a good way. over-confident respondents who used the help of networks were less likely to own stocks, but more likely to own bonds, etfs, and cash value life insurance. findings from this study related to under-confident investors indicates a need for additional research. while noting that all respondents in this study reported owning investments outside of retirement plans, under-confidence may indicate a reluctance to invest in assets classes where there is uncertainty of outcomes or unfamiliarity with asset characteristics. the lack of a moderation effect with media-based advice and asset class ownership may point to a reluctance to take such sources of advice seriously. on the other hand, fee-for-service and free help providers may be providing under-confident clients valuable advice. while being under-confident was not directly associated with investment ownership, when moderated by these help providers, under-confident respondents were less likely to report holding cash value life insurance or commodities, indicating that a relationship with a financial professional can potentially be beneficial in helping a client manage portfolio risk. because over-confident and under-confident clients are encountered on a regular basis, it behooves financial planners to be able to identify signs of confidence bias and take steps to counteract such biases. a financial planner can increase the value of services offered to a client by steering the over-confident into less risky and lower cost asset classes and by guiding the under-confident into asset classes that fit a more aggressive risk profile but may not be familiar or initially comfortable. finally, as with most studies that are exploratory, the results from this study should be evaluated in the context of potential limitations. for example, data for the study were obtained from a pre-existing dataset. some of the questions asked in the survey may suffer from validity issues. consider the risk-tolerance measure. as noted by grable and schumm (2010), the scf risk item that is available in the dataset suffers from validity and reliability deficiencies. future studies ought to consider a more robust measure of client risk tolerance/ aversion. additionally, the questions used to measure financial knowledge, while generally accepted as valid, may have been too trivial as measures of investment knowledge, which could have skewed the estimate of knowledge confidence. additionally, future studies should attempt to identify the causality of investment ownership. even in the context of these and other limitations, the results from this study are noteworthy in providing insights into who is more or less likely to hold specific types of investment products. this study shows how understanding an investor’s profile can be used to more precisely match portfolio development needs to a client’s unique situation. note 1 a moderator is defined as a variable that affects the direction and/or strength of the relationship between an independent variable and a dependent variable (baron and kenny, 1986). 227a. rabbani et al. / financial services review 27 (2018) 209-230 references allgood, s., & walstad, w. b. (2013). financial literacy and credit card behaviors: a cross-sectional analysis by age. numeracy, 6, article 3. http://dx.doi.org/10.5038/1936-4660.6.2.3 ameriks, j., & zeldes, s. p. (2004). how do household portfolio shares vary with age. working paper, columbia university. anderson, a. (2013). trading and under-diversification. review of finance, 17, 1699–1741. barber, b. m., & odean, t. (2001). boys will be boys: gender, overconfidence, and common stock investment. the quarterly journal of economics, 116, 261–292. baron, r. m., & kenny, d. a. (1986). the moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. journal of personality & social psychology, 51, 1173. barsky, r. b., juster, f. t., kimball, m. s., & shapiro, m. d. (1997). preference parameters and behavioral heterogeneity: an experimental approach in the health and retirement study. the quarterly journal of economics, 112, 537–579. biais, b., hilton, d., mazurier, k., & pouget, s. (2005). judgemental overconfidence, selfmonitoring, and trading performance in an experimental financial market. the review of economic studies, 72, 287–312. appendix measures of objective financial knowledge 1. in general, investments that are riskier tend to provide higher returns over time than investments with less risk. true 76% false 14% don’t know 9% 2. if you buy a company’s stock: (1) you own a part of the company 73% (2)you have lent money to the company 14% (3) you are liable for the company’s debts 1% (4) the company will return your original investment to you with interest 6% (5) don’t know 5% 3. if you buy a company’s bond: (1) you own a part of the company 10% (2)you have lent money to the company 65% (3) you are liable for the company’s debts 4% (4) you can vote on shareholder resolutions 5% (5) don’t know 16% 4. over the last 20 years in the united states, the best average returns have been generated by: (1) stocks 55% (2) bonds 8% (3) cds 2% (4) money market accounts 6% (5) precious metals 8% (6) don’t know 19% 5. if a company files for bankruptcy, which of the following securities is most at risk of becoming virtually worthless? (1) the company’s preferred stock 13% (2) the company’s common stock 53% (3) the company’s bonds 15% (4) don’t know 19% 6. which of the following best explains why many municipal bonds pay lower yields than other government bonds? (1) municipal bonds are lower risk 32% (2) there is a greater demand for municipal bonds 10% (3) municipal bonds can be tax-free 34% (4) don’t know 23% 7. what has been the approximate average annual return of the s&p 500 stock index over the past 20 years (not adjusted for inflation)? (1) �10% 0% (2) �5% 2%, (3) 5% 25%, (4) 10% 26% (5) 15% 7%, (6) 20% 4%, (7) don’t know 36% 8. you invest $500 to buy $1,000 worth of stock on margin. the value of the stock drops by 50%. you sell it. approximately how much of your original $500 investment are you left with in the end? (1) $500 21% (2) $250 35% (3) $0 23% (4) don’t know 20% 9. which is the best definition of “selling short?” (1) selling shares of a stock shortly after buying it 11% (2) selling shares of a stock before it has reached its peak 20% (3) selling shares of a stock at a loss 26% (4) selling borrowed shares of a stock 21% (5) don’t know 22% 10. which of the following best explains the distinction between nominal returns and real returns? (1) nominal returns are pre-tax returns; real returns are after-tax returns 12% (2) nominal returns are what an investment is expected to earn; real returns are what an investment actually earns 21% (3) nominal returns are not adjusted for inflation; real returns are adjusted for inflation 12% (4) nominal returns are not adjusted for fees and expenses; real returns are adjusted for fees and expenses 9% (5) don’t know 44% the bold text representing the correct answer and the percentage figure indicating the proportion of respondents who selected a particular answer. 228 a. rabbani et al. / financial services review 27 (2018) 209-230 blanchett, d., & kaplan, p. (2013). alpha, beta, and now … gamma. the journal of retirement, 1, 29–45. bruine de bruin, w., parker, a. m., & fischhoff, b. (2007). individual differences in adult decision-making competence. journal of personality and social psychology, 92, 938–957. bruine de bruin, w., parker, a. m., & fischhoff, b. (2012). explaining adult age differences in decision–making competence. journal of behavioral decision making, 25, 352–360. calvet, l. e., campbell, j. y., & sodini, p. (2007). down or out: assessing the welfare costs of household investment mistakes. journal of political economy, 115, 707–747. chang, m. l. (2005). with a little help from my friends (and my financial planner). social forces, 83, 1469–1498. https://doi.org/10.1353/sof.2005.0061 chen, y., & feeley, t. h. (2014). social support, social strain, loneliness, and well-being among older adults: an analysis of the health and retirement study. journal of social and personal relationships, 31, 141–161. https://doi.org/10.1177/0265407513488728 chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–307. (available at https://search.proquest.com/docview/212012102). cocco, j. f. (2005). portfolio choice in the presence of housing. review of financial studies, 18, 535–567. https://doi.org/10.1093/rfs/hhi006 duasa, j., & yusof, s. a. (2013). determinants of risk tolerance on financial asset ownership: a case of malaysia. international journal of business and society, 14, 1–16. (available at https://search.proquest.com/docview/ 1365272777). faig, m., & shum, p. (2002). portfolio choice in the presence of personal illiquid projects. the journal of finance, 57, 303–328. fonseca, r., mullen, k. j., zamarro, g., & zissimopoulos, j. (2012). what explains the gender gap in financial literacy? the role of household decision making. journal of consumer affairs, 46, 90–106. https://doi.org/ 10.1111/j.1745-6606.2011.01221.x gervais, s., & odean, t. (2001). learning to be overconfident. the review of financial studies, 14, 1–27. gibson, r. c. (2000). asset allocation: balancing financial risk (3rd ed.). new york, ny: mcgraw-hill. goetzmann, w. n., & kumar, a. (2008). equity portfolio diversification. review of finance, 12, 433–463. https://doi.org/10.1093/rof/rfn005 grable, j. e., & chatterjee, s. (2014). reducing wealth volatility: the value of financial advice as measured by zeta. journal of financial planning, 27, 45–51. grable, j. e., & goetz, j. (2017). communication essentials for financial planners: strategies and techniques. hoboken, nj: wiley. grable, j. e., & joo, s. h. (1999). financial help-seeking behavior: theory and implications. journal of financial counseling and planning, 10, 14–25. (available at https://search.proquest.com/docview/135587007). grable, j. e., & joo, s. h. (2001). a further examination of financial help-seeking behavior. journal of financial counseling and planning, 12, 55–73. (available at https://search.proquest.com/docview/1364688790). grable, j. e., & schumm, w. (2010). an estimate of the reliability of the survey of consumer finances risk-tolerance question. journal of personal finance, 9, 117–131. grossman, s. j., & stiglitz, j. e. (1980). on the impossibility of informationally efficient markets. the american economic review, 70, 393–408. guiso, l., jappelli, t., & terlizzese, d. (1996). income risk, borrowing constraints, and portfolio choice. the american economic review, 86, 158–172. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. (available at ssrn https://ssrn.com/abstract�1953400). heaton, j., & lucas, d. (2000). portfolio choice in the presence of background risk. the economic journal, 110, 1–26. hu, x. (2005). portfolio choices for homeowners. journal of urban economics, 58, 114–136. https://doi.org/ 10.1016/j.jue.2005.02.002 hung, a., & yoong, j. (2010). asking for help: survey and experimental evidence on financial advice and behavior change. rand working paper series wr-714-1. http://dx.doi.org/10.2139/ssrn.1532993 229a. rabbani et al. / financial services review 27 (2018) 209-230 joo, s. h., & grable, j. e. (2001). factors associated with seeking and using professional retirementplanning help. family and consumer sciences research journal, 30, 37– 63. http://doi.org/10.1177/ 1077727x01301002 kimball, m. s., sahm, c. r., & shapiro, m. d. (2008). imputing risk tolerance from survey responses. journal of the american statistical association, 103, 1028–1038. kuzniak, s., rabbani, a., heo, w., ruiz-menjivar, j., & grable, j. e. (2015). the grable and lytton risktolerance scale: a 15-year retrospective. financial services review, 24, 177–192. (available at http://proxyremote.galib.uga.edu:80/login?url�https://search.proquest.com/docview/1685862633). kwon, j. (2004). clustering users of multiple sources of information for saving and investment. journal of personal finance, 3, 33–48. lee, j., & cho, j. (2005). consumers’ use of information intermediaries and the impact on their information search behavior in the financial market. journal of consumer affairs, 39, 95–120. https://doi.org/10.1111/ j.1745-6606.2005.00005.x lusardi, a. (2012). numeracy, financial literacy, and financial decision-making. numeracy, 5, article 2. http://dx.doi.org/10.5038/1936-4660.5.1.2 lusardi, a., & mitchell, o. s. (2007). financial literacy and retirement preparedness: evidence and implications for financial education. business economics, 42, 35–44. https://doi.org/10.2145/20070104 odean, t. (1998). volume, volatility, price, and profit when all traders are above average. the journal of finance, 53, 1887–1934. robb, c. a., babiarz, p., woodyard, a., & seay, m. c. (2015). bounded rationality and use of alternative financial services. journal of consumer affairs, 49, 407–435. https://doi.org/10.1111/joca.12071 robb, c., & woodyard, a. s. (2011). financial knowledge and best practice behavior. financial planning and counseling, 22, 36–46. (available at ssrn https://ssrn.com/abstract�2061308). rose, f. (1999). the economics, concept, and design of information intermediaries: a theoretic approach. heidelberg, ny: physica-verlag. roche, h., tompaidis, s., & yang, c. (2013). why does junior put all his eggs in one basket? a potential rational explanation for holding concentrated portfolios. journal of financial economics, 109, 775–796. https:// doi.org/10.1016/j.jfineco.2013.03.016 sages, r. a., & grable, j. e. (2010). financial numeracy, net worth, and financial management skills: client characteristics that differ based on financial risk tolerance. journal of financial service professionals, 64, 57–65. statman, m., thorley, s., & vorkink, k. (2006). investor over-confidence and trading volume. the review of financial studies, 19, 1531–1565. https://doi.org/10.1093/rfs/hhj032 suchman, e. a.(1966). health orientation and medical care. american journal of public health and the nations health, 56, 97–105. verrecchia, r. e. (1982). information acquisition in a noisy rational expectations economy. econometrica: journal of the econometric society, 50, 1415–1430. vissing-jorgensen, a. (2002). towards an explanation of household portfolio choice heterogeneity: nonfinancial income and participation cost structures (no. w8884). nber macroeconomics annual, 2003. wang, c., & hanna, s. d. (2007). the risk tolerance and stock ownership of business owning households, financial counseling and planning, 18, 3–18. woodyard, a., & grable, j. (2014). doing good and feeling well: exploring the relationship between charitable activity and perceived personal wellness. voluntas: international journal of voluntary and nonprofit organizations, 25, 905–928. https://doi.org/10.1007/s11266-013-9382-5 woodyard, a., & robb, c. (2012). financial knowledge and the gender gap. journal of financial therapy, 3, 1–16. https://doi.org/10.4148/jft.v3i1.1453 woodyard, a. s., robb, c., babiarz, p., & jung, j. y. (2017). knowledge and practice: implications for cash and credit management behaviors. family and consumer sciences research journal, 45, 300–314. https://doi.org/ 10.111/fcsr.12202 xiao, j. j. (1996). effects of family income and life cycle stages on financial asset ownership. journal of financial counseling and planning, 7, 21–30. 230 a. rabbani et al. / financial services review 27 (2018) 209-230 pii: s1057-0810(99)80003-1 financial services review, 7(2): 73-81 copyright © 1998 by jai press inc. issn: 1057-0810 all tights of reproduction in any form reserved. innovations in savings schemes: the bonus bonds trust in new zealand jenny ridge and martin young bonus bonds were introduced into new zealand in march 1970 based on the premium bond savings scheme operating in the united kingdom. a bonus bond is a unit in the bonus bonds trust with all income from the trust's conservative investments being bal loted for on a monthly basis by the bond homers. this paper looks at the history and operation of the bonus bonds savings scheme in new zealand and examines why it has such wide appeal for new zealanders. the paper also addresses the following aspects of the scheme. performance of the bonds; their level of success as a savings vehicle; factors contributing to their relative success over time; the risk profile of the bond hold ers. the major findings of the paper are that investors in bonus bonds have little appreciation of real values and they view the bonds as an appropriate investment for risk averse investors. i. introduction bonus bonds are a savings vehicle which were introduced to the new zealand public in march 1970. this paper outlines the history of bonus bonds in new zealand and examines their operation and the reasons for their wide appeal to the new zealand public. a bonus bond is a unit in the bonus bonds trust and the return earned by the bonus bonds trust is balloted for by the bond holders in the trust. the principle deposited in the bonus bonds trust is invested in low risk assets. this paper examines the interesting risk profile of the bond holders, being very risk averse with their principle contributions in nominal terms yet risk takers with their income component. ii. defining a bonus bond bonus bonds are not an investment in the traditional sense. instead of being paid the inter est earned by the fund, the bond holder has a chance of winning between $20 and $300,000 jenny ridge and martin young ° department of finance and property studies, massey university, private bag 11222, palmerston north, new zealand; phone: (06) 356 9099, fax: (06) 350 5651. 74 financial services review 7(2 ) 1998 a month in prize money. chances of winning are currently one in 13,000 per bond with approximately 1.6 billion units currently on issue. section 28 of the finance act 1991 under which the bonus bonds trust operates requires that no unit is to have a better chance of winning a prize than one in 9,600. the level of prizes is set by the interest the funds of the trust earn in their highly conservative portfolio, being over $6 million for the month of june 1997. some new zealanders treat bonus bonds as a serious investment on which they hope to win a steady number of small prizes with the excitement of possibly winning a large prize. as is normally the case for managed funds operating in new zealand, tax on the bonus bond fund's income is deducted prior to the allocation, through ballot, of the prizes and is therefore tax-free in the hands of the winners. the tax rate currently applied is the top tax rate in new zealand of 33% making the fund a tax inefficient investment for those bond holders on lower personal tax rates. as prizes are tax-free in investor's hands they do not have to be declared on tax returns. one substantial benefit for investors is that there appears to be no fluctuation in the capital value of each bonus bond from the viewpoint of the investor. while the fund is subject to some interest rate risk the mandate under which it operates allows for some flexibility in adjusting the prize payouts to ensure an asset back ing of $1 per bond is maintained. also the duration of the fund is kept short to minimise capital gains and losses through interest rate movements. each bonus bond on issue has a cost of $1 with a minimum purchase being twenty $1 units making it a very affordable fund to invest in, but expensive to manage with a manage ment fee of 1.33% of gross assets. the fact that the manager of the bonus bonds trust has a virtual monopoly on this type of product in new zealand may also be a contributing fac tor to this high management fee which was 15% of total income for 1996. a trustee fee is also payable at the rate of 0.3%. no entry or exit fees are charged and bonus bonds can be redeemed within 7 working days. no commission is payable to agents with the exception of new zealand post, the new zealand postal service company, which is paid a 0.3% fee for the applications it pro cesses. this unique product is part of new zealand life. one in three new zealanders are holders of the bonds of which over 90% hold less than $2,000 worth. bonus bonds are competing with term deposits and other fixed interest investments at one end of the risk spectrum and popular gambling games at the other, of which there are many operating legally in new zealand. the term deposit market tends to be very compet itive in new zealand with all major banks competing for market share. as an indication of this short-term bank interest rates on personal deposits tend to be very close to the whole sale 90 day bank bill rates. gambling in new zealand has always been a very popular past time with a number of legal gambling options being available to willing participants. in recent times the number of these has increased significantly which could be seen as com petition for bonus bonds. bonus bonds are different from other forms of gambling, how ever, as when you redeem your bonds your original investment is repaid. a. a history of bonus bonds bonus bonds were introduced to the new zealand public under section 129a of the post office act 1959. the scheme was based on the premium bond scheme in operation in the united kingdom at that time. premium bonds went on sale in britain in 1956 and found instant success with total sales having reached £6.6 billion by 1997. comparing innovations in savings schemes 75 bonus bonds with premium bonds, premium bonds have a face value of £1 each with any £1 bond having a one in 19,000 chance of winning a prize. the major prize for premium bonds is the £1 million jackpot which was introduced in november 1993. premium bonds are subject to a maximum holding of £20,000, unlike bonus bonds, on which there is no maximum holding. this limit is fixed by regulation and may be changed from time to time. premium bonds can be bought in multiples of ten £ 1 units with a minimum purchase of one hundred units. the only exception to this is where the purchase is made under an automatic prize reinvestment mandate. the £ 1 million jackpot prize is guaranteed each month and the numbers of each of the other prizes varies according to the total prize pool. a person with a maximum holding of £20,000 of premium bonds would, on average, expect to win an average of 13 prizes a year. in new zealand too, bonus bonds found instant success. at the time of launching it had been anticipated that $1 million worth of bonus bonds might be sold in the first month whereas the actual figure was $10.9 million. bonus bonds were introduced by the finance minister of the day, sir robert muldoon, in conjunction with national development bonds as part of a scheme designed to direct savings towards so called "productive" gov ernment projects. the money raised by bonus bonds was used to invest in new zealand infrastructure such as roads, airports and schools. new zealanders were encouraged to save by investing in bonus bonds, an investment which guaranteed the nominal value of the principal but appealed to the risk taking nature of new zealanders by having the interest earnings balloted for. until 1987 the bonus bonds trust was managed by the new zealand post office sav ings bank on behalf of the new zealand government. the post office became a corpora tion in 1987 and post-bank came into being. post-bank continued to manage bonus bonds until it was sold to the australia and new zealand banking group on october 1, 1990. at the time of the post-bank sale to the anz banking group, the parties agreed that the bonus bond scheme would also transfer. the bonus bond fund was transferred to a unit trust scheme named the bonus bonds trust under the bonus bonds trust deed on sep tember 17, 1990. management and ownership of the scheme was then transferred to the anz trust with unit holders becoming reliant on the trustee and manager of the fund to protect their interests rather than the new zealand government. until october 1990 the face value of deposits represented liabilities of the government. from this date until octo ber 1992 only pre-october 1990 bonds had the guarantee with this being removed in octo ber 1992. all units are now unsecured although a substantial percentage of deposits are invested in government securities. b. operat ion of the bonus bonds trust as previously stated income on the bonus bonds trust is not distributed proportion ally but is placed in a pool and distributed as prizes. prize winning serial numbers are selected at random by elsie the bonus bonds random number selector. for the year end ing september 30, 1996 the bonus bonds trust funds were invested as follows. investment securiues 1996 1995 1994 $000 $000 $000 nz government securities 899,835 1,097,134 1,072,673 local body securities 87,752 47,738 65,852 redeemable preference shares 9,875 76 financial services review 7(2) 1998 debentures/state owned enterprise securities 28,240 14,994 10,196 bank deposits 95 1,901 2,336 corporate bonds 317,246 2 9 6 , 0 0 2 268,437 discounted securities 260,389 140 ,528 283,706 totals 1,593,357 1,608,172 1,703,230 source: notes to the financial statements 1996. bonus bonds trust. the trust has a new zealand-only investment policy but does not impose any limits on the proportion of assets that may be invested in any one company. although equities, futures, options, swaps contracts, insurance and underwriting contracts are permitted it has been the policy of the manager to invest in low risk debt securities. in 1997 new zealand government stock made up 56% of the portfolio. the manager does not have to inform unit holders of any changes made to the bonus bonds trust investments but the manager, along with the trustees could be held personally responsible should the asset backing of each bond drop significantly below the $1 level. the prize pool equals the return to the trust fund after allowing for trustee's and manager 's fees, allocations to reserves and tax ation. the pool is divided into prizes as for prize draw no. 324, april 8, 1997. prize amount first prize $300,000 second prize $100,000 third prize $50,000 112 prizes of $5,000 280 prizes of $1,000 565 prizes of $500 3,176 prizes of $100 34,800 prizes of $50 123,040 prizes of $20 each bonus bond has a serial number and each bonus bond has the chance to win more than one prize in any one prize draw. the pool allocation has remained unaltered since march 28, 1995. the ballot results are available on the second tuesday of each month and are published in the press. a letter is also sent to all winners. the full prize draw list is available at all branches of anz bank, post bank and post shops by the friday of the week following the draw. any prize money not claimed within six months is trans ferred to the surplus account. i f a prize is still not claimed within 25 years it is paid to the crown. no interest is paid on unclaimed prizes and the substantial interest accumulated on these belongs to the anz bank. currently more than $1.4 million of prizes are unclaimed. in the year to 30, september 1996 the average after tax payout was 4.7%. c. performance of the bonus bond trust the performance success of the bonus bonds trust can be considered from the view point of the managers of the trust; growth of the fund, and also from the viewpoint of the unit holders; returns, real and nominal. when analysing this performance it is important to take note of the economic climate in new zealand over the period under consideration. throughout the seventies inflation in new zealand was running at very high levels and interest rates in general were not compensating investors for the inflationary impact on innovations in savings schemes 77 their deposits. this situation was reasonably common in the western world at that time leading to periods of negative real interest rates. in 1984 inflation came down very sharply, though not through market forces but on account of the imposition of a wage-price freeze. this freeze was lifted in late 1984 and deregulation of the economy began. inflation rose again through to 1987 before being brought under control by tight monetary policy imple mented by new zealand's central bank. in 1989 the central bank in new zealand was given the sole economic objective of maintaining stable prices. stable prices in new zealand are currently defined as being an inflation rate between zero and 3% (reserve bank of new zealand, 1992). high short-term real interest rates are the main weapon of the central bank to achieve this goal. this approach to controlling inflation has been largely successful since the early 1990s. it should be noted at this point that it is difficult to make comparisons between the pre-1984 and post-1984 performances of bonus bonds as the economic changes which occurred in new zealand from 1984 on were very substantial. since march 1971 the bonus bonds fund has grown from $20.8 million to $1.63 bil lion. the periods which have shown the highest real growth rates were the initial five years when the concept was clearly gaining support and the periods of lowest relative inflation being 1984 and 1988 to 1997. the same generally holds for nominal growth'rates, both of table 1 the growth of the bonus bonds trust (nominal and real) % increase in total deposits deposits % increase in date $ million (nominal) deposits (real) march 1971 20.8 march 1972 33.1 59.13 46.84 march 1973 55.8 68.58 59.13 march 1974 79.6 42.65 29.34 march 1975 100.3 26.00 11.29 march 1976 128. i 27.71 8.95 march 1977 138.3 7.96 -5.01 march 1978 167.4 21.04 5.61 march 1979 196.4 17.32 6.27 march 1980 215.7 9.82 -7.22 march 1981 239.8 i 1.17 -3.51 march 1982 287.6 19.93 3.55 march 1983 339.7 18. i 1 4.85 march 1984 431.7 27.08 22.79 march 1985 486.1 12.60 -0.70 march 1986 494.5 1.72 -9.91 march 1987 541.4 9.48 -7.45 march 1988 641.6 18.50 8.77 march 1989 760.3 18.50 13.94 march 1990 900.9 18.49 10.71 march 1991 1064.0 18.10 12.07 march 1992 1327.0 24.71 24.27 march 1993 1448.0 8.36 7.44 march 1994 1592.0 9.05 7.86 march 1995 1687.0 5.63 1.86 march 1996 1659.0 -1.66 -3.56 march 1997 1630.0 -1.75 -2.83 source: new zealand key statistics, 1970-1997 and new zealand official yearbook, 1970-1997. 78 f i n a n c i a l s e r v i c e s r e v i e w 7(2) 1998 table 2 returns to the bonus bond trust payout real value 90 day 5 year rate real investment treasury govt date nominal return % dollar stock bond march 1970 1.0000 march 1971 4.00 -6.30 0.9415 5.14 5.29 march 1972 4.00 -4.37 0.9035 4.99 5.24 march 1973 4.00 1.94 0.8870 4.42 5.18 march 1974 4.00 -6.28 0.8365 4.20 5.41 march 1975 4.00 -9.22 0.7683 4.44 5.66 march 1976 4.00 13.22 0.6817 5.51 7.44 march 1977 4.16 -9.51 0.6247 7.00 10.00 march 1978 5.17 -9.44 0.5733 7.50 9.75 march 1979 5.76 -4.63 0.5492 10.60 10.00 march 1980 5.70 12.68 0.4904 11.25 13.00 march 1981 5.50 -9.73 0.4490 11.25 12.50 march 1982 5.70 10,12 0.4097 12.00 14.00 march 1983 7.20 -5.45 0.3899 7.80 10.30 march 1984 6.90 3,41 0.4027 7.80 11.00 march 1985 7.40 -6.00 0.3814 17.73 12.33 march 1986 7.70 -5.23 0.3638 15.00 12.47 march 1987 7.50 -1.08 0.3306 17.45 12.20 march 1988 8.00 -1.44 0.3262 10.71 8.99 march 1989 8.00 4.00 0,3387 8.92 8.83 march 1990 8.00 0.98 0.3418 9.17 8.09 march 1991 8.00 2.62 0.3503 8.00 7.62 march 1992 6.00 5.08 0.3700 4.91 5.76 march 1993 5.50 4.50 0.3868 4.84 4.85 march 1994 3.50 4.20 0.4030 3.64 4.01 march 1995 3.50 -0.05 0.4010 6.31 5.71 march 1996 4.20 2.30 0.4102 5.94 5.49 march 1997 4.70 3.60 0.4250 5.07 5.22 note: all returns are on an after tax basis. source," new zealand key statistics, 1970-1997 and new zealand official yearbook, 1970-1997. which are shown in table 1. table 2 shows the annual nominal payout rates to the fund from march 1971 through to march 1997 together with the level of real returns to the fund and the real value of the investment dollar adjusted for the payout level. comparative rates of returns on new zealand 90 day treasury stock and 5 year government stock are also shown. these rates of return show that the government achieved its objective of raising cheap capital for economic development over the period that it had control of the fund. despite the fund's private management in recent times it has still failed to return at a com petitive rate in most years but the highly positively skewed distribution of returns that the fund displays would appear to compensate investors for this. on account of the fact that the fund's investments are all in interest bearing deposits the drop in interest rates, which accompanies a drop in inflation, leads to drops in the nom inal payout rate. these drops in the nominal payout rate had no apparent effect on growth rates for the fund, however, at least until just recently. in fact it would appear that the lower nominal bank deposit rates, which occur with lower inflation, act as a strong incentive for people to increase deposits in bonus bonds. innovations in savings schemes 79 0 100000 ~ ~ 0 ~ ~ ~ 0 ~0~_(0 ! 1 1 1 1 1 d l l l l l l ~ l ~ r t m b0rd t.l~ 0 loooo0 ~i~o ~ mmmm figure 1. probability and returns to bonus bonds per monthly draw holdings of one unit to ten thousand units the real return on bonus bonds has been very poor over most of the period in which the fund has been operating. the nominal payout rate was initially set at 4% and stayed there until 1977 from where it rose steadily, 1984 excluded, until it reached a peak of 8% in the late eighties. since then it dropped steadily to a level of 3.5% before rising again to its current level of 4.7%. after allowing for inflation expected real returns to the fund have been negative in eighteen of the twenty-eight years in which it has been operating. from inception until 1988 (excluding 1984) the real value of the investment dollar, adjusted for the payout rate, declined steadily to under 33 cents. despite this very poor performance, however, bonus bonds have retained a large degree of popularity among the investing public. d. the risk profile of the bonus bond holders one of the key features of bonus bonds is the public perception, well justified in nom inal terms, that the principal amount is very secure to the point of being almost risk free. as current sales literature states, "if the fund managers have one philosophy, it is the abso lute need for security of investment." currently 56% of the fund is invested in government securities effectively giving a large degree of government guarantee. it can be argued that investors view their principal 80 financial services review 7(2 ) 1998 as being invested in a zero return risk free asset. the compensation for forgoing at least the risk free rate of return on the principal is the high risk expected return which, while posi tive, is balloted for and clearly lies in the risk taking segment of the risk spectrum. for nominal values then, it can be argued that investors in bonus bonds view the total invest ment as belonging to the risk averse segment of the risk spectrum. if the positive real returns that have been experienced lately continue the same may well apply in real terms as well. this apparent conflict between risk averse and risk taking behaviour has been noted many times in the literature going back to the friedman-savage hypothesis which postu lated that individuals have both risk averse and risk taking segments within their utility functions (elton & gruber, 1995). this hypothesis came about from the observation that individuals will both buy insurance and gamble at the same time. more recently levy and sarnat (1984) reported a finding that investors have a liking for positive skewness within their investment returns, being prepared to forgo a small part of their expected return for a chance of a very high return. bonus bonds appear to be able to exploit this characteristic of human behavior in a very effective manner. the risk nature of this investment can be further explained in figure 1. because the principal is justifiably regarded as being secure the distribution of returns per month is highly positively skewed with the downside risk being no more than the going rate of inter est on deposits. that is, for most bond holders, the return per month will be zero with no downside risk. as the number of bonds held by an individual investor rises the chances of at least one prize per month increases. if a bond holder holds just one hundred units the probability of no bond winning a prize is 98.99%. in the case of a bond holder holding ten thousand units the probability of no bond winning a prize drops to 36.33% with a 27.94% chance of winning a $20 prize and a 7.9% chance of winning a $50 prize. the average holding in bonus bonds is around one thousand units for which the probability of no bond winning a prize at the monthly draw is 90%. this shows that the average investor in bonus bonds is prepared to forgo the normal return on deposits on approximately $1,000 for a very small chance of a high return with no risk to capital so long as they receive some small reward about one month in ten. iii. conclusion the investment mix of retaining the nominal value of one's capital with a high degree of certainty and having the chance of a large win through balloting for the interest earned on the capital has attracted many depositors to the bonus bonds trust operating in new zealand, making it new zealand's largest managed fund. currently in new zealand only the australian and new zealand banking group, through their ownership of post-bank, are able to operate such a fund, although there are schemes operating at banks where a lower interest rate is paid on deposits so that prizes can be won by lucky depositors. poor real returns, often negative, have done little to discourage investment in this savings vehicle. while the high level of security of the capital may well lead investors to view the bonds as being appropriate for a risk averse investor, authorities could view such bonds which promise a zero return on capital with near certainty for most holders as being close to gambling and therefore undesirable. whatever view the relevant authorities might hold, innovations in savings schemes 81 one cannot argue against the proven attractiveness o f such schemes and other countries might well consider introducing similar products. similar schemes could be set up where only part of the return to the fund is balloted for which might well prove more attractive to investors than the new zealand scheme. which ever way such a scheme was set up, three components appear necessary to facilitate its success. the worst outcome for a bond holder must be a small or zero return with near certainty, the distribution of returns must be highly posit ively skewed, and some small reg ular reward should be received. references department of statistics, new zealand, key statistics, 1970-199z author. department of statistics, new zealand official yearbook, 1970-1997. author. elton, e. j., & gruber, m. j. (1995). modem portfolio theory and investment analysis (5th ed.). new york: wiley. levy, h., & sarnat, m. (1984). portfolio and investment selection: theory and practice. prentice hall international. reserve bank of new zealand. (1992). monetary policy and the new zealand financial system (3rd ed.). author. academy of financial services officers president inga timmerman california state university, northridge president-elect executive vice president-program terrance k. martin utah valley university vice president-communications colleen tokar asaad baldwin wallace university vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations shawn brayman planplus global immediate past president janine sam shepherd university editor, financial services review stuart michelson stetson university directors charles chaffin cfp board of standards lu fan university of missouri barry mulholland university of akron tom potts baylor university laura ricaldi utah valley university past presidents janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial. org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2020 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 29, no. 1, 2021 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academyfinancial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2021 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. advisor compensation: which clients know and how do they pay? somer g. andersona, martin c. seayb,*, kyoung tae kimc, derek r. lawsond amaryville university, 650 maryville university drive, st. louis, mo, 63141 bkansas state university, 318 justin hall, manhattan ks, 66506 cuniversity of alabama, 312 adams hall, tuscaloosa al, 35487 dkansas state university, 303 justin hall, manhattan ks, 66506 abstract using the 2015 national financial capability study investor survey, this study uses agency theory to inform an exploration of characteristics associated with knowing how one’s financial advisor/broker is compensated. this study further examines how individuals who do know the compensation method choose between financial advisors with different compensation models. proposed changes in the financial advising regulatory landscape, as well as the pending changes to the cfp board standards of professional conduct, brings greater emphasis on understanding consumers’ advisor compensation preferences. primary results indicate that clients who place importance on fees, that are more knowledgeable about diversification, and that perform background checks are more likely to know compensation methods. a follow-up analysis reveals distinct differences between individuals that used each model, but also provide some mixed results, indicating that clients may not fully understand the compensation paid to their advisors. discussion and implications related to these results are provided. © 2018 academy of financial services. all rights reserved. jel classification: d10; g20; g41 keywords: agency theory; compensation models; financial advisor; investor survey; national financial capability study * corresponding author. tel.: �1-785-532-1486; fax: �1-785-532-5505. e-mail address: mseay@ksu.edu financial services review 27 (2018) 231-255 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. 1. introduction given the shift away from defined benefit (db) plans to defined contribution (dc) plans, u. s. workers have been tasked with an increased responsibility to plan and prepare for their retirement. this task requires individuals to undergo complex financial decisions that involve significant uncertainty, yet individuals often have difficulty making complex decisions under uncertainly (tversky and kahneman, 1974). as such, the shift in retirement funding responsibility has helped to ignite the growth of the financial planning profession (seay, kim, and heckman, 2016). a significant body of research exists investigating the relationship between financial literacy, help-seeking behavior (calcagno and monticone, 2015; collins, 2012; molton, loibl, samak, and collins, 2013), and seeking professional advice (finke, huston, and winchester, 2011; hanna, 2011; lachance and tang, 2012; seay et al., 2016). however, little research has examined how consumers choose between financial advisors, and more specifically, how consumers choose between advisors with different compensation methods. the certified financial planner board of standards (cfp board) provides three categories of compensation models under which cfp certificants can operate: (a) commission only, (b) fee-only, and (c) combined commission-and-fee. while advisors compensated by commission are paid based on specific product recommendations, fees are typically charged based on assets under management (aum), through an annual retainer, or through hourly or project based planning charges. further, commission-based advisors have historically been held to a suitability standard, simply requiring that a product be suitable for a client given their financial situation. on the other hand, fee advisors, operating in the registered investment advisor (ria) space, have typically been held to a fiduciary standard that requires recommendations be in the client’s best interest. the differences in how compensation has been derived, and more specifically different standards of care, have often been used to differentiate these advisors. notably, commission-and-fee advisors operate in the hybrid space, operating as a fiduciary when providing advice through an ria and under the suitability standard when operating through a broker-dealer. pending changes in the financial planning regulatory landscape brings greater emphasis on understanding consumers’ advisor compensation preferences. while the department of labor’s (dol) conflict of interest rule (federal register, 2017) was vacated by the courts in june of 2018, many financial services firms made structural changes in how their financial professionals engage with clients. further, much of the spirit of the dol rule is now found in the security and exchange commission’s proposed rule regulation best interest (reg bi). the current reg bi proposal would require financial advisors previously held to the suitability standard to act in the best interest of their customers at the time a recommendation is being made, without placing the advisors, or their firms, interests ahead of the client. within the nonregulatory space, cfp professionals will be required to act as fiduciaries to their clients when providing financial advice beginning october 1, 2019. this is a clear expansion of the previous standard of conduct, which only required cfp professionals to act in a fiduciary capacity when providing financial planning services. financial advice is defined very broadly, stated as any client communication that “would reasonably be viewed as a suggestion that the client take or refrain from taking a particular course of action with respect to” an array of financial planning areas including, but not limited to, anytime the cfp 232 m. seay / financial services review 27 (2018) 231-255 professional has discretion of client assets, development and/or implementation of a financial plan, and investment advice and strategies (cfp board, 2017, p. 16–17). in light of these pending changes, most, if not all, cfp professionals will engage in fiduciary work with their clients. similarly, under the proposed reg bi the standards of care between all financial professionals, although significantly different, will continue to converge around putting the client’s interest first. this change would, in many ways, level the financial planning client services market from a client care perspective, making it imperative to gain an understanding of what factors clients consider when they differentiate between advisors that use differing financial planning compensation models. given this backdrop, the current study primarily explores two central research questions. first, what are the characteristics of clients that know their advisors’ compensation method? secondly, of those individuals who did know how their financial advisor was compensated, what are the differences in clients that use commission-only, fee-only, and combined commission-and-fee advisors? for empirical analyses, this study used two linked datasets, the 2015 national financial capability study (nfcs) and its supplementary dataset, the 2015 nfcs investor survey. this study provides insights into the client characteristics that are associated with knowledge of financial advisor compensation and the use of a certain type of financial advisor based on their compensation method. furthermore, it provides financial advisors insights into their clients’ characteristics and why they choose to utilize their services. 2. literature review the use of professional financial advisors in the united states has been growing steadily over the last several decades. the percentage of u.s. households that reported using a financial advisor grew from 21% in 1998 (hanna, 2011) to 29% in 2012 (seay et al., 2016). previous research has established an understanding of who seeks financial advice and why they do so, and the value that households receive from it. agency theory provides a lens for informing hypotheses and directional expectations related to an individual knowing or not knowing their financial advisor’s compensation method. 2.1. who seeks financial advice? the characteristics of financial professionals that consumers are looking for has been examined in previous studies. bae and sandager (1997) found that consumers were predominantly interested in using a financial advisor for services related to retirement planning, tax planning, and investment management. furthermore, bae and sandager (1997) found that consumers with both the desire for the personal assurance and confidence that a trusted financial professional can provide and with lower levels of financial knowledge sought a financial professional. the authors also reviewed what was important to consumers in selecting their financial professional, showing that confidentiality, objectivity, competence, honesty, and the ability to communicate were very important factors to prospective clients. prospective clients found access to the financial professional and compensation transparency 233m. seay / financial services review 27 (2018) 231-255 to be less important. finally, education and certification were found as critical components towards selection of a financial professional, with 92% of clients reporting that a financial professional with the cfp designation is important (bae and sandager, 1997). much research has also examined the characteristics of consumers that sought financial advice. individuals that sought financial help from a financial professional typically had higher levels of financial risk tolerance (grable and joo, 2001; hanna, 2011; joo and grable, 2001; marsden, zick, and mayer, 2011; robb, babiarz, and woodyard, 2012), financial satisfaction (grable and joo, 2001), wealth, educational attainment (finke et al., 2011; hanna, 2011; lachance and tang, 2012), and income (hanna, 2011). these individuals were also older (finke et al., 2011; grable and joo, 2001; marsden et al., 2011) than their counterparts and displayed more positive financial behaviors (e.g., regularly saved, eliminated credit card debt, or planned for big purchases; grable and joo, 2001; joo and grable, 2001). another important positive predictor in financial help-seeking is an individual’s willingness to trust (lachance and tang, 2012; martin, finke, and gibson, 2014). finally, individuals with more proactive attitudes and women were more likely to seek financial help from a professional (joo and grable, 2001). 2.2. the value of financial advice some of the best evidence of the value of financial advice based on real client data were provided by marsden et al. (2011). they found that meeting with a financial advisor encouraged more prudent retirement planning activities. specifically, individuals were more likely to have established long-term financial goals, calculated financial needs in retirement, higher levels of more supplemental savings, diversified retirement portfolios, saved more regularly, and spent more time learning about financial topics. measurable long-term impacts of meeting with an advisor were increased confidence in the client’s ability to retire and larger emergency fund savings. montmarquette and viennot-briot (2015) found that having a financial advisor for a minimum of four years had a positive and significant impact on client wellbeing. after controlling for many variables (socio-economic, demographic, and attitudinal variables), individuals using a financial advisor had significantly higher levels of financial assets. this increase in asset levels was not purely because of investment related advice, as evidence suggested that clients receiving financial advice created better savings disciplines. blanchett and kaplan (2013) explored the value of financial advice through the concept of gamma. they evaluated the advantages of using a financial advisor as evidenced through: (1) appropriate asset allocation strategies; (2) creating dynamic withdrawal strategies; (3) understanding the appropriate use of guaranteed income strategies (e.g., annuities); (4) choosing between tax-efficient strategies; and (5) building portfolios that account for risks faced by retirees. although theoretical and not driven by actual client data, improved financial decision making associated with financial advisors generated increased returns of approximately 1.6% annually within the context of retirement planning for retirees. grable and chatterjee (2014) expanded upon this research to introduce zeta, a function of gamma and alpha (portfolio return beyond expected return within a portfolio), which 234 m. seay / financial services review 27 (2018) 231-255 focuses on the value of advice in reducing wealth volatility during turbulent markets. they found that clients who met with a financial advisor before the great recession experienced significantly less negative returns and wealth volatility, adjusted for the risk taken within the portfolio. 2.3. understanding compensation method previous literature provides little information in regard to which clients would be expected to know how they are paying their financial advisors. in response to this gap, one study, the 2014 advisor impact economics of loyalty study conducted by nielsen, collected information about knowledge of financial professional compensation methods from a sample of 1,229 respondents who worked with a financial professional and possessed investable assets of at least $50,000 (raskie, herbison, and martin, 2017). descriptive results from that study suggested that 68% of respondents knew how their financial professional was compensated. however, not all respondents that knew the method being used to pay their financial professional indicated that they knew the amount which the professional was being compensated (raskie et al., 2017). in fact, the results suggested that the more commission that was involved with the compensation method, the less likely it was that the respondent knew how much they were paying their financial professional. specifically, 82% of respondents who were paying fee-only, 79% who were working with a commission-and-fee professional, and 67% of those who were paying commission-only indicated that they knew how much their professionals were being compensated for their services. raskie et al. (2017) used the nielsen study to investigate the relationship between knowledge of compensation method and knowledge of the financial professional’s investment process and found evidence of a link between the two. compared with those who did not know the compensation method, those who used a fee-only professional had the highest odds of knowing the investment process; with those using fee-and-commission coming in next highest, and those using commission-only having the lowest odds of knowing the investment process. they also investigated the association between compensation method and having an understanding of the investment process and found that client understanding of the investment process used by the financial service professional was affected by the method of compensation. consumers who knew how their advisor was compensated had a better understanding of the professional’s investment process than those who did not. use of the commission-only method was not found to be a significant predictor of understanding the investment process; however, using the fee-only and commission-and-fee methods were associated with higher levels of understanding as compared with not knowing the compensation method at all. although sparse, previous literature (raskie et al., 2017) suggests that there is a relationship between some individual client characteristics and the likelihood of knowing the compensation method used to pay a financial professional. agency theory can provide a lens that can be used to explore this relationship. clients and financial advisors have an agency relationship where the consumer (client) compensates the agent (financial advisor) to perform services on the consumer’s behalf (jensen and meckling, 1976; raskie et al., 2017). because the financial advisor is presumably more informed about the services that are 235m. seay / financial services review 27 (2018) 231-255 delegated by the client, information asymmetry exists between the financial advisor and the client (jensen and meckling, 1976). information asymmetry creates agency costs that can be incurred by both the client and the advisor in the form of monitoring and bonding costs. the client can incur monitoring costs when he or she feels they must observe the behavior of the advisor to make certain the advisor is protecting the client’s best interests. the advisor may incur bonding costs, such as obtaining professional designations and licenses, to signal to the client that they have the knowledge and expertise to perform the services in the client’s best interests. there should be an inverse relationship between monitoring and bonding costs where the client can reduce monitoring costs when the advisor has signaled an investment in bonding costs. in this vein, clients who are sensitive to agency costs would be expected to understand the method they were using to compensate their financial advisors. given the gap in research related to knowledge of how financial professionals are compensated, the primary purpose of the current study is to explore the relationship between client characteristics and knowledge of financial advisor compensation method at the multivariate level. specifically, the associations between the following client characteristics: (1) client socio-demographic profiles, (2) financial characteristics, (3) financial knowledge, (4) attitudes toward advisor preference, and (5) compensation; and the dependent variable, knowing or not knowing the advisor compensation method, were investigated. informed by agency theory, the following hypotheses were explored in the primary analysis. h1: financial knowledge will be positively associated with knowing the advisor compensation method. the financial knowledge block of characteristics explored in this study includes objective knowledge about compound interest, inflation, bonds, mortgages, and diversification. additionally, this block includes indicators of both subjective financial knowledge and confidence in applying that knowledge. it is expected that there would be less information asymmetry in the relationship between clients with high financial knowledge and their advisors. therefore, those with higher financial knowledge may better understand the advisor compensation methods that are available and should choose the method that best fits their desired level of agency costs. h2: attitudes about advisor preferences will be positively associated with knowing the advisor compensation method. agreement with the importance of using an advisor to free up time, prevent losses, and improve investment performance were the attitudes investigated in the advisor preference block of characteristics. additionally, this block includes whether or not the client performed a background or registration check, and the degree of importance of an advisor’s professional designation. it is expected that clients who had strong advisor preferences would be sensitive to agency costs and would seek to understand the compensation method used by the advisor so they could gauge the level of bonding costs incurred by the advisor and adjust monitoring costs based on those preferences. for instance, agency costs would appear to be very important to a client who went to the trouble to perform a background check on a financial advisor. therefore, it would be expected that this client would also seek to understand the compensation method they were using to pay their advisor. 236 m. seay / financial services review 27 (2018) 231-255 h3: compensation beliefs will be positively associated with knowing the advisor compensation method. the compensation beliefs investigated in this study related to agreement with the impact on services received from an advisor who received commission on trading activity or received selling incentives. additionally, the degree of importance that a respondent placed on advisor fees was also included in this block of characteristics. it is expected that clients who had strong beliefs in this area would be more likely to know how their financial advisors were paid. 2.4. client preferences in compensation models beyond understanding how an advisor is paid, previous literature has provided evidence that there is a relationship between some client characteristics and the compensation method those clients used to pay their advisors (raskie et al., 2017). however, more research is required to understand how clients choose between compensation models. a first step in understanding how client preferences influence advisor selection was provided by seay et al. (2017), who analyzed bivariate statistics of clients that used each compensation model. results indicated that commission-only clients were mostly older (65 and over), relatively financially knowledgeable and confident in their knowledge about finances, and were comfortable with their decision-making as it related to investments. these clients reported placing less importance on professional designations and using a financial advisor to free up time. similarly, they were not as concerned about fraud, were less prone to believe that commissions influenced advice received, and believed that the fees they paid were reasonable. commission-and-fee clients were relatively younger (specifically, age 18 –34), more racially diverse and were full-time workers (seay et al., 2017). they reported the highest levels of concern about fraud and were the most likely group of clients to perform a background check on their advisor, yet they believed the fees they paid were reasonable. these clients had the lowest financial knowledge but reported the highest subjective knowledge of all client groups. they were confident in their investment decision-making, were the most optimistic about the future prospects of the financial markets, and used an advisor to free up time. lastly, fee-only clients were the largest group of clients, had high incomes, were financially knowledgeable but were not confident in their investment decision-making and also reported the lowest subjective financial knowledge of all client groups (seay et al., 2017). fee-only clients were interested in working with an advisor with a professional designation, wanted to delegate decisions to their advisor, and utilized an advisor to free up their time. finally, fee-only clients had strong beliefs about commissions influencing advice received. given the basis for compensation model preference provided by seay et al. (2017), a secondary analysis within the current study explored the association between client characteristics and the use of either a commission-only, fee-only, or combined commission-and-fee financial advisor for those individuals who did report knowing the compensation model used to pay their financial advisor. given the heterogeneity of services provided within each 237m. seay / financial services review 27 (2018) 231-255 compensation method, it is difficult to make consistent a priori hypotheses related to characteristics associated with the use of each method. consequently, this secondary analysis is purely exploratory in nature, and there are no directional expectations for which clients might be using which method. 3. method 3.1. data and sample selection this study utilized data from the 2015 nfcs, a nationally representative survey of american adults administered between june and october 2015, and the follow-up 2015 nfcs investor survey that was administered in july 2015 to a subset of respondents who had indicated ownership of nonretirement investments. the larger 2015 nfcs state-by-state data were collected from an online survey of 27,564 adults in the united states, with approximately 500 observations per state plus the district of columbia. the 2015 study included an oversampling for large states, resulting in 1,000 respondents each from california, illinois, new york, and texas. respondents were recruited from established online panels and selected using a nonprobability quota sampling methodology with criteria based on income, gender, ethnicity, age, and education level. to explore the associations between client attributes and financial advisor compensation models, the analytic sample was drawn from the 2,000 individuals who indicated owning nonretirement investments and completed the 2015 nfcs investor survey. from this subset, the sample was then limited to 955 respondents who indicated having a specific broker or advisor. observations with missing data for any variables in the regression model were listwise deleted, leaving an analytic sample of 750 for the primary analysis. of this 750, 142 respondents indicated that they did not know the compensation method used to pay their advisor; while 608 did know the compensation method. the secondary analysis used only the sample of 608 individuals who knew their compensation method. to obtain further details on this dataset, see mottola and kieffer (2017). 3.2. measurement of variables the dependent variable, compensation method, was based on the following question from the 2015 nfcs investor survey: “which of the following types of fees do you pay for your nonretirement investment accounts” with the response choices of: (a) a commission on trades, (b) a percentage of the total value of assets managed, and (c) a fixed monthly or annual fee. no information related to hourly planning fees was collected. responses were organized into four mutually exclusive categories: commission-only (n � 143) if the respondent selected only option a; fee-only (n � 291) if the respondent selected option b and/or c but not option a, commission-and-fee (n � 174) if the respondent selected option a along with another option, and do not know (n � 142) if the respondent indicated that they did not know the compensation method. for the primary analysis, we created a binary indicator of knowing compensation method and coded it as 1 if the compensation method was located in the 238 m. seay / financial services review 27 (2018) 231-255 commission-only, fee-only, or commission-and-fee categories and 0 if the compensation method was located in the do not know category. the secondary analysis, “compensation method” was coded as 0 for commission-only, 1 for fee-only, and 2 for commission-and-fee. independent variables were organized into five blocks of client characteristics: (a) sociodemographic, (2) financial, (3) basic financial knowledge, (4) advisor preference, and (5) compensation beliefs. these variables and their measurements are discussed below. additional measurement information is provided in the appendix. 3.2.1. socio-demographic characteristics the nfcs includes a variety of demographic and socio-economic variables. however, because of sample size limitations, some groups were condensed to ensure appropriate cell sizes for the multinomial logit analysis. socio-demographic variables included in the analysis include age, gender, race, education, marital status, and employment status. 3.2.2. financial characteristics the 2015 nfcs and 2015 nfcs investor survey include both objective and subjective assessments of an individual’s financial situation. objective assessments include income and approximate value of all investments in nonretirement accounts. subjective measures include respondents’ self-reported risk tolerance measured on a 10-point likert-type scale, with higher scores indicating greater willingness to take risks. 3.2.3. financial knowledge the 2015 nfcs asked five questions about the fundamental concepts of personal finance drawn from a 5-question scale created and used by lusardi and mitchell (2009, 2011). for the purposes of this analysis, binary variables were created to signify whether the respondent correctly answered each question related to compound interest, inflation, bonds, mortgages, and diversification. additionally, respondents were asked to assess their subjective financial knowledge and confidence in financial ability according to a 7-point likert-type scale, with higher scores indicating greater self-assessed financial knowledge and confidence in financial ability. 3.2.4. advisor preference characteristics the nfcs 2015 investor survey provided several measures related to financial advisor use and preference. respondents were asked to assess the importance of each of the following reasons for using a financial advisor: to free up my time, to help avoid losses, and to improve investment performance. for this analysis, responses were coded as binary variables with very important � 1, otherwise coded as 0. respondents were also asked about the importance of financial advisor designations (very important � 1, otherwise � 0) and performance of background checks (yes � 1, otherwise � 0). 3.2.5. advisor compensation characteristics the 2015 nfcs investor survey asked a question that measured the respondents’ attitudes toward sales-based compensation structures. responses were coded as categorical variables 239m. seay / financial services review 27 (2018) 231-255 with the following levels: would not affect at all, would affect somewhat, and would affect a great deal. additionally, the survey contained a question that measured respondents’ agreement with fee importance when opening nonretirement investment accounts, assessed on a 10-point likert-type scale. 3.3. empirical models a binary logit was used to analyze the effect of the selected predictor variables on the likelihood of knowing advisor compensation method. the model for the probabilities of clients knowing the compensation method is: prob �yi � j�xi� � pij � exp� x�i�j� 1 � � k�1 j exp� x�i�k� , j � 0, 1. we estimate probabilities for being in the j�1 type for clients with characteristics, xi for the group indicator j, 0 � do not know compensation method and 1 � know compensation method. to analyze the effect of the selected predictor variables on the likelihood of utilizing the three compensation methods, a multinomial logit model was used. the model for the probabilities of clients using each compensation is: prob �yi � j�xi� � pij � exp� x�i�j� 1 � � k�1 j exp� x�i�k� , j � 0, 1, 2. we estimate probabilities for being in the j�1 type for clients with characteristics, xi. for the group indicator j, 0 � commission-only, 1 � fee-only, and 2 � commission-and-fee. multicollinearity among the predictor variables was tested using variance inflation factor (vif) scores. although the nfcs provides weighting information for the full sample, weights are not provided for the subset represented by the nfcs investor survey. consequently, results were not weighted for the purpose of this research. 4. results 4.1. descriptive results tables 1 through 5 display the descriptive statistics of financial planning clients overall and by compensation model. 4.1.1. socio-demographic characteristics table 1 contains the descriptive statistics for socio-demographic characteristics. the majority of the overall sample respondents were age 55 or older (60%); however, the ages of the respondents using the commission-and-fee compensation method were more evenly split between under 55 (52%) and over 55 (48%). overall, the sample respondents 240 m. seay / financial services review 27 (2018) 231-255 were male (54%), white (83%), had at least some college education (92%), and married (71%). there were some demographic differences noted between the respondents who did not know their compensation and the rest of the sample groups. there were more women (62%), there was more diversity (22% non-white), and more of these respondents indicated that they had no college education (15%). most of the overall sample respondents were either employed full-time (39%) or retired (34%); however, the majority of the respondents using the commission-and-fee compensation method were employed full-time (59%). respondents that knew their compensation method also had a higher rate of being employed full time (41%) than respondents who did not know their compensation method (29%). 4.1.2. financial characteristics table 2 displays the descriptive statistics for financial characteristics. most participants reported a household income of $50,000 or more (85%) and a total nonretirement account value of less than $500,000 (75%). participants in the sample reported relatively high risk tolerance (m � 6.19 out of 10); however, respondents using the commission-and-fee method had the highest risk tolerance (m � 6.78 out of 10) compared with respondents using the table 1 socio-demographic characteristics, 2015 national financial capability study (nfcs) investor survey compensation model full sample (n � 750) do not know compensation (n � 142) know compensation (n � 608) commission only (n � 143) fee only (n � 291) commissionand-fee (n � 174) age 18 to 34 11.20% 11.97% 11.02% 6.29% 9.97% 16.67% 35 to 44 12.00% 9.15% 12.66% 12.59% 11.68% 14.37% 45 to 54 16.80% 14.79% 17.27% 17.48% 15.12% 20.69% 55 to 64 25.07% 25.35% 25.00% 23.08% 25.09% 26.44% 65 and older 34.93% 38.73% 34.05% 40.56% 38.14% 21.84% gender male 54.13% 38.03% 57.89% 61.54% 52.58% 63.79% female 45.87% 61.97% 42.11% 38.46% 47.42% 36.21% race white 82.80% 78.17% 83.88% 89.51% 83.85% 79.31% non-white 17.20% 21.83% 16.12% 10.49% 16.15% 20.69% education high school or less 8.00% 14.79% 6.41% 5.59% 6.87% 6.32% some college 24.40% 26.76% 23.85% 20.28% 24.05% 26.44% college 67.60% 58.45% 69.74% 74.13% 69.07% 67.24% marital status married 70.67% 68.31% 71.22% 70.63% 72.85% 68.97% not married 29.33% 31.69% 28.78% 29.37% 27.15% 31.03% employment status full-time 39.07% 28.87% 41.45% 34.27% 34.71% 58.62% self-employed 10.00% 6.34% 10.86% 13.29% 10.31% 9.77% retired 33.87% 36.62% 33.22% 36.36% 38.83% 21.26% other 17.07% 28.17% 14.47% 16.08% 16.15% 10.34% 241m. seay / financial services review 27 (2018) 231-255 commission-only or fee-only methods (m � 6.24 and 6.12, respectively). moreover, respondents who did not know their compensation had the lowest risk tolerance (m � 5.56). 4.1.3. financial knowledge characteristics descriptive statistics for basic financial knowledge characteristics are located in table 3. most respondents in the sample correctly answered objective financial knowledge questions related to compound interest (91%), inflation (83%), bonds (55%), mortgages (91%), and diversification (79%); however, respondents using the commission-and-fee method had a table 2 financial characteristics, 2015 national financial capability study (nfcs) investor survey compensation model full sample (n � 750) do not know compensation (n � 142) know compensation (n � 608) commission only (n � 143) fee only (n � 291) commissionand-fee (n � 174) household income less than $35,000 7.60% 11.27% 6.74% 11.19% 6.87% 2.87% $35,000 to $49,999 7.07% 9.86% 6.41% 6.29% 6.53% 6.32% $50,000 to $74,999 24.27% 28.87% 23.19% 24.48% 22.34% 23.56% $75,000 to $99,999 21.07% 20.42% 21.22% 20.28% 19.93% 24.14% $100,000 or more 40.00% 29.58% 42.43% 37.76% 44.33% 43.10% total value of non-retirement accounts $0 to $49,999 20.93% 27.46% 19.41% 24.48% 19.24% 15.52% $50,000 to $99,999 13.07% 12.17% 12.17% 10.49% 11.68% 14.37% $100,000 to $249,999 20.53% 19.90% 19.90% 16.08% 20.62% 21.84% $250,000 to $499,999 20.13% 21.22% 21.22% 19.58% 21.65% 21.84% $500,000 to $999,999 13.73% 9.15% 14.80% 13.99% 15.46% 14.37% $1,000,000 or more 11.60% 7.75% 12.50% 15.38% 11.34% 12.07% risk tolerance 6.19 5.56 6.34 6.24 6.12 6.78 mean (sd) (2.05) (2.14) (2.01) (2.01) (0.20) (2.00) table 3 basic financial knowledge, 2015 national financial capability study (nfcs) investor survey compensation model full sample (n � 750) do not know compensation (n � 142) know compensation (n � 608) commission only (n � 143) fee only (n � 291) commissionand-fee (n � 174) financial knowledge (correct %) compound interest 91.07% 85.21% 92.43% 90.91% 93.81% 91.38% inflation 83.33% 82.39% 83.55% 90.91% 84.19% 76.44% bond 55.20% 41.55% 58.39% 66.43% 55.67% 56.32% mortgage 91.07% 86.62% 92.11% 95.10% 91.41% 90.80% diversification 79.33% 64.79% 82.73% 84.62% 85.22% 77.01% subjective financial knowledge, mean (sd) 5.83 5.51 6.00 5.96 5.87 5.90 (0.86) (0.99) (0.81) (0.79) (0.83) (0.82) confidence in financial ability, mean (sd) 6.45 6.25 6.50 6.47 6.59 6.39 (0.92) (1.19) (0.84) (0.92) (0.71) (0.94) 242 m. seay / financial services review 27 (2018) 231-255 lower percentage of correct answers on each question compared with respondents using the other two compensation methods. further, respondents who did not know their compensation also had a lower percentage of correct answers on each question compared with those who did know their compensation. in terms of self-assessed financial knowledge measures, respondents rated themselves at the higher-end of the subjective financial knowledge and confidence in financial ability scales (m � 5.83 and m � 6.45 out of 7, respectively). however, those paying commission-and-fee reported lower confidence in financial ability (m � 6.39) than those using the other two methods. those who did not know their compensation reported the lowest subjective financial knowledge (m � 5.51) and confidence of all groups (m � 6.25). 4.1.4. advisor preference characteristics table 4 contains descriptive statistics related to advisor preference characteristics. most respondents found it very important to use their advisor to help them prevent losses (79%) and improve investment performance (84%). only 30% of respondents found it very important to use an advisor to help them free up time, while 35% of those who did not know their compensation and only 20% of respondents using the commission-only method found this factor to be very important. concerning advisor selection, 59% of the respondents felt that it was very important for their financial advisor to have possession of a professional designation; but 63% of those who did not know their compensation method and only 50% of respondents using the commission-only method found professional designations to be very important. only 25% of the overall sample respondents reported that they had performed a background, registration, or license check on a financial advisor; however, 39% of respondents using the commission-and-fee method and only 8% of those who did not know their compensation had done so. table 4 advisor preference, 2015 national financial capability study (nfcs) investor survey compensation model full sample (n � 750) do not know compensation (n � 142) know compensation (n � 608) commission only (n � 143) fee only (n � 291) commissionand-fee (n � 174) use advisor to free up time very important 29.60% 34.51% 28.45% 19.58% 30.93% 31.61% not very important 70.40% 65.49% 71.55% 80.42% 69.07% 68.39% use advisor to prevent losses very important 79.47% 81.69% 78.95% 80.42% 81.44% 73.56% not very important 20.53% 18.31% 21.05% 19.58% 18.56% 26.44% use advisor to improve investment performance very important 84.40% 84.51% 84.38% 81.12% 86.94% 82.76% not very important 15.60% 15.49% 15.63% 18.88% 13.06% 17.24% performed advisor background/ registration check 25.07% 7.75% 29.11% 23.08% 26.46% 38.51% importance of professional designation very important 59.20% 62.68% 58.39% 49.65% 62.20% 59.20% not very important 40.80% 37.32% 41.61% 50.35% 37.80% 40.80% 243m. seay / financial services review 27 (2018) 231-255 4.1.5. advisor compensation characteristics in regard to the dependent variables, the majority of the sample (81%) indicated that they knew the compensation method. of those that did know their compensation method, almost half of the respondents reported paying compensation on a fee-only basis (48%), with the rest of the sample reporting compensation based on commission only (23%) or both commission and fees (29%). table 5 displays descriptive statistics related to advisor compensation characteristics. respondents overwhelmingly felt that advisor compensation based on sales factors such as commissions on trading activity and selling incentives would at least somewhat affect advice received. in terms of fee importance, respondents felt that fees and pricing structure related to nonretirement accounts were important factors when opening an account (m � 7.9 out of 10), but those who did not know their compensation method found it to be less important (m � 7.2). 4.2 regression results—primary analysis to analyze the effect of the selected predictor variables on the likelihood of knowing the compensation method, a binary logit model was used. results from the binary logistic regression analysis are presented in table 6. in regard to model fit, pseudo r2 was 0.195. variance inflation factors ranged from 1.1 to 4.3, indicating that multicollinearity among the predictor variables is not a concern in this analysis. holding all else equal, variables from each block of characteristics except for the financial characteristics block were found to be significant predictors of knowing the compensation method (p � 0.05). hypothesis one expected a positive relationship between financial knowledge and knowing the advisor compensation method; however, knowledge of diversification was the only variable in the financial knowledge block that was significantly associated with knowing the compensation method. specifically, those answering the question correctly had 2.5 times higher odds of knowing the compensation method than those who answered the question table 5 compensation beliefs, 2015 national financial capability study (nfcs) investor survey compensation model full sample (n � 750) do not know compensation (n � 142) know compensation (n � 608) commission only (n � 143) fee only (n � 291) commissionand-fee (n � 174) commission on trading activity impacts advice received would not affect at all 21.60% 16.90% 22.70% 32.17% 19.24% 20.69% would affect somewhat 42.40% 47.89% 41.12% 41.26% 39.18% 44.25% would affect a great deal 36.00% 35.21% 36.18% 26.57% 41.58% 35.06% selling incentive impacts advice received would not affect at all 14.13% 11.97% 14.64% 16.08% 15.12% 12.64% would affect somewhat 38.53% 46.48% 36.68% 36.36% 34.71% 40.23% would affect a great deal 47.33% 41.55% 48.68% 47.55% 50.17% 47.13% fee importance, mean (sd) 7.89 7.15 8.07 8.06 7.99 8.20 (1.93) (2.33) (1.78) (1.62) (1.88) (1.72) 244 m. seay / financial services review 27 (2018) 231-255 table 6 binary logistic regression results estimating probability of knowing compensation method (n � 750) characteristic b se b odds ratio p value intercept �.3281 1.210 0.007** socio-demographic characteristics age (ref group: 65 and older) 18 to 34 0.114 0.472 1.121 0.809 35 to 44 0.450 0.470 1.569 0.338 45 to 54 0.412 0.401 1.510 0.304 55 to 64 0.175 0.300 1.192 0.559 gender (ref group: female) male 0.570 0.227 1.767 0.012* race (ref group: non-white) white 0.421 0.289 1.524 0.145 education (ref group: college) high school or less �0.780 0.351 0.459 0.027* some college �0.002 0.261 0.999 0.996 marital status (ref group: not married) married �0.182 0.254 0.834 0.474 employment status (ref group: full-time) self-employed 0.280 0.445 1.322 0.531 retired �0.196 0.345 0.822 0.576 other �0.623 0.312 0.534 0.044* financial characteristics household income (ref group: $100,000 or more) less than $35,000 �0.452 0.482 0.637 0.349 $35,000 to $49,999 �0.561 0.447 0.571 0.210 $50,000 to $74,999 �0.234 0.305 0.790 0.439 $75,000 to $99,999 �0.210 0.312 0.811 0.503 total value of non-retirement accounts (ref group: $1,000,000 or more) $0 to $49,999 0.145 0.504 1.156 0.774 $50,000 to $99,999 �0.378 0.494 0.685 0.444 $100,000 to $249,999 �0.029 0.456 0.971 0.949 $250,000 to $499,999 0.334 0.463 1.399 0.467 $500,000 to $999,999 0.192 0.486 1.211 0.694 risk tolerance 0.033 0.058 1.034 0.563 basic financial knowledge compound interest 0.603 0.352 1.827 0.087 inflation �0.158 0.326 0.854 0.628 bond 0.274 0.233 1.316 0.239 mortgage 0.147 0.374 1.158 0.695 diversification 0.934 0.273 2.545 0.001** subjective financial knowledge 0.210 0.139 1.234 0.131 confidence in financial ability 0.058 0.118 1.060 0.622 advisor preference very important to use advisor to free up time �0.279 0.244 0.756 0.251 very important to use advisor to prevent losses �0.050 0.313 0.951 0.874 very important to use advisor to improve investment performance 0.106 0.333 1.111 0.751 performed advisor background/registration check 1.392 0.372 4.021 0.000** professional designation very important �0.286 0.231 0.751 0.214 compensation beliefs commission on trading activity impacts advice received (ref group: would not affect at all) would affect somewhat �0.232 0.376 0.793 0.537 would affect a great deal �0.385 0.416 0.680 0.355 selling incentive impacts advice received (ref group: would not affect at all) would affect somewhat �0.486 0.443 0.615 0.273 would affect a great deal �0.231 0.469 0.793 0.621 fee importance 0.243 0.056 1.275 �.0001*** pseudo r2 0.195 †p � .10, *p � .05, **p � .01, ***p � .001. 245m. seay / financial services review 27 (2018) 231-255 incorrectly. answering questions correctly that were related to compound interest, inflation, bonds, and mortgages was not associated with knowledge of the compensation method. regarding the subjective financial knowledge variables, self-assessed financial knowledge and confidence in financial ability, neither client characteristic was associated with knowing the compensation method. attitudes about advisor preferences were expected to be positively related to knowing the compensation method in hypothesis two. only one variable in this block of characteristics was significantly related to knowing the compensation method. specifically, those who had performed an advisor background or registration check on their financial advisor had four times higher odds of knowing the compensation method compared with those who did not perform these checks. clients who indicated that it was very important to use a financial advisor to free up time, prevent losses, and improve investment performance and those who found professional designations to be very important did not have significantly different odds of knowing the compensation method than those who did not find these variables to be very important. in hypothesis 3, it was expected that compensation beliefs would be positively associated with knowing the advisor compensation method. one item in the compensation beliefs block was associated with knowing the compensation method. specifically, an increase in the belief that fees were important when opening nonretirement investment accounts was associated with 28% higher odds of knowing the compensation method. believing that commission on trading activity or selling incentives would impact advice received was not significantly associated with knowledge of the compensation method. several variables in the socio-demographic block of characteristics were associated with knowing the compensation method. men had 1.8 times higher odds of knowing the compensation method in comparison to women. education was also a significant predictor of knowing the compensation method. compared with having a college education, those with an education of high school or less had lower odds of knowing the compensation method; however, there was no significant difference between completing a college education and only having some college. in terms of employment status, there were no significant differences between clients who classified themselves as full-time versus those who considered themselves self-employed or retired. however, clients with an employment classification of “other” had significantly lower odds of knowing the compensation method. age, race, marital status, household income, the total value of nonretirement accounts, and risk tolerance were not found to be significant predictors of knowing the compensation method. 4.3. regression results—secondary analysis as a follow-up analysis, a multinomial logit model was used to analyze the effect of the selected predictor variables on the likelihood of utilizing the three compensation methods using only the sample of respondents who reported knowing the compensation method used to pay their advisor. results from the multivariate logistic regression analysis are presented in table 7. in regard to model fit, pseudo r2 was 0.12. holding all else equal, variables from all five blocks of client characteristics were found to be significant predictors of compen246 m. seay / financial services review 27 (2018) 231-255 t ab le 7 m ul tin om ia l lo gi st ic re gr es si on re su lts es tim at in g pr ob ab ili ty of us in g co m pe ns at io n m od el (n � 60 8) c om pe ns at io n m od el fe eon ly vs . co m m is si on -o nl y c om m is si on -a nd -f ee vs . co m m is si on -o nl y c om m is si on -a nd -f ee vs . fe eon ly b se b o dd s ra tio p va lu e b se b o dd s ra tio p va lu e b se b o dd s ra tio p va lu e in te rc ep t 1. 50 4 1. 49 7 0. 31 5 2. 40 0 1. 63 1 0. 14 1 0. 89 6 1. 36 4 0. 51 1 so ci ode m og ra ph ic ch ar ac te ri st ic s a ge (r ef gr ou p: 65 an d ol de r) 18 to 34 0. 77 5 0. 54 2 2. 17 0 0. 15 3 1. 22 8 0. 58 8 3. 41 3 0. 03 7* 0. 45 3 0. 45 9 1. 57 3 0. 32 4 35 to 44 (0 .1 20 ) 0. 45 4 0. 88 7 0. 79 1 0. 01 5 0. 50 4 1. 01 5 0. 97 6 0. 13 5 0. 43 6 1. 14 5 0. 75 6 45 to 54 (0 .0 78 ) 0. 40 3 0. 92 5 0. 84 6 0. 08 8 0. 44 9 1. 09 2 0. 84 4 0. 16 7 0. 39 0 1. 18 1 0. 66 9 55 to 64 0. 15 1 0. 31 4 1. 16 3 0. 63 0 0. 39 7 0. 36 4 1. 48 7 0. 27 6 0. 24 6 0. 31 2 1. 27 8 0. 43 1 g en de r (r ef gr ou p: fe m al e) m al e (0 .4 00 ) 0. 23 9 0. 67 0 0. 09 3† 0. 03 4 0. 26 8 1. 03 5 0. 89 9 0. 43 4 0. 22 0 1. 54 4 0. 04 9* r ac e (r ef gr ou p: n on -w hi te ) w hi te (0 .5 59 ) 0. 36 2 0. 57 2 0. 12 3 (0 .7 02 ) 0. 38 2 0. 49 6 0. 06 6† (0 .1 43 ) 0. 28 1 0. 86 7 0. 61 0 e du ca tio n (r ef gr ou p: c ol le ge ) h ig h sc ho ol or le ss 0. 26 8 0. 49 0 1. 30 7 0. 58 5 0. 41 6 0. 54 6 1. 51 6 0. 44 6 0. 14 9 0. 44 7 1. 16 0 0. 73 9 so m e co lle ge 0. 31 8 0. 29 3 1. 37 4 0. 27 7 0. 64 9 0. 32 4 1. 91 3 0. 04 5* 0. 33 1 0. 26 0 1. 39 2 0. 20 3 m ar ita l st at us (r ef gr ou p: n ot m ar ri ed ) m ar ri ed (0 .1 92 ) 0. 27 7 0. 82 5 0. 48 7 (0 .3 98 ) 0. 30 3 0. 67 2 0. 18 9 (0 .2 06 ) 0. 24 9 0. 81 4 0. 40 9 e m pl oy m en t st at us (r ef gr ou p: fu lltim e) se lf -e m pl oy ed (0 .2 67 ) 0. 39 4 0. 76 6 0. 49 8 (0 .7 61 ) 0. 42 9 0. 46 7 0. 07 6† (0 .4 94 ) 0. 37 0 0. 61 0 0. 18 2 r et ir ed 0. 40 4 0. 36 6 1. 49 7 0. 27 0 (0 .5 75 ) 0. 41 3 0. 56 3 0. 16 4 (0 .9 78 ) 0. 35 1 0. 37 6 0. 00 5* * o th er 0. 00 3 0. 35 8 1. 00 3 0. 99 3 (0 .7 60 ) 0. 41 3 0. 46 8 0. 06 6† (0 .7 63 ) 0. 34 5 0. 46 7 0. 02 7* fi na nc ia l ch ar ac te ri st ic s h ou se ho ld in co m e (r ef gr ou p: $1 00 ,0 00 or m or e) l es s th an $3 5, 00 0 (1 .2 41 ) 0. 51 9 0. 28 9 0. 01 7* (1 .9 26 ) 0. 69 3 0. 14 6 0. 00 5* * (0 .6 85 ) 0. 62 9 0. 50 4 0. 27 6 $3 5, 00 0 to $4 9, 99 9 (0 .8 78 ) 0. 53 1 0. 41 6 0. 09 8† (0 .6 44 ) 0. 59 5 0. 52 5 0. 27 9 0. 23 4 0. 49 1 1. 26 3 0. 63 4 $5 0, 00 0 to $7 4, 99 9 (0 .8 94 ) 0. 32 8 0. 40 9 0. 00 6* * (0 .6 07 ) 0. 36 5 0. 54 5 0. 09 6† 0. 28 7 0. 30 0 1. 33 3 0. 33 8 $7 5, 00 0 to $9 9, 99 9 (0 .6 51 ) 0. 32 5 0. 52 2 0. 04 5* (0 .3 36 ) 0. 35 5 0. 71 5 0. 34 4 0. 31 5 0. 28 8 1. 37 1 0. 27 4 t ot al va lu e of no nre tir em en t ac co un ts (r ef gr ou p: $1 ,0 00 ,0 00 or m or e) $0 to $4 9, 99 9 0. 28 2 0. 46 9 1. 32 6 0. 54 8 (0 .7 02 ) 0. 54 3 0. 49 6 0. 19 6 (0 .9 84 ) 0. 47 7 0. 37 4 0. 03 9* $5 0, 00 0 to $9 9, 99 9 0. 46 5 0. 49 1 1. 59 2 0. 34 3 0. 05 2 0. 54 4 1. 05 3 0. 92 4 (0 .4 13 ) 0. 46 0 0. 66 1 0. 36 9 $1 00 ,0 00 to $2 49 ,9 99 0. 78 7 0. 44 6 2. 19 6 0. 07 8† 0. 52 9 0. 49 6 1. 69 7 0. 28 6 (0 .2 58 ) 0. 41 6 0. 77 3 0. 53 5 $2 50 ,0 00 to $4 99 ,9 99 0. 58 6 0. 41 3 1. 79 7 0. 15 6 0. 39 4 0. 46 1 1. 48 3 0. 39 3 (0 .1 92 ) 0. 39 7 0. 82 5 0. 62 8 $5 00 ,0 00 to $9 99 ,9 99 0. 70 3 0. 42 9 2. 01 9 0. 10 1 0. 44 5 0. 47 7 1. 56 0 0. 35 1 (0 .2 58 ) 0. 41 3 0. 77 3 0. 53 2 r is k to le ra nc e 0. 01 0 0. 06 1 1. 01 0 0. 87 1 0. 07 6 0. 07 1 1. 07 9 0. 28 1 0. 06 6 0. 05 9 1. 06 8 0. 26 6 (c on ti nu ed on ne xt pa ge ) 247m. seay / financial services review 27 (2018) 231-255 t ab le 7 (c on tin ue d) c om pe ns at io n m od el fe eon ly vs . co m m is si on -o nl y c om m is si on -a nd -f ee vs . co m m is si on -o nl y c om m is si on -a nd -f ee vs . fe eon ly b se b o dd s ra tio p va lu e b se b o dd s ra tio p va lu e b se b o dd s ra tio p va lu e b as ic fin an ci al kn ow le dg e c om po un d in te re st 1. 00 2 0. 45 2 2. 72 3 0. 02 7* 0. 84 7 0. 49 8 2. 33 2 0. 08 9† (0 .1 55 ) 0. 42 5 0. 85 7 0. 71 6 in fla tio n (1 .0 34 ) 0. 40 1 0. 35 6 0. 01 0* (1 .1 51 ) 0. 42 1 0. 31 6 0. 00 6* * (0 .1 17 ) 0. 30 3 0. 89 0 0. 69 9 b on d (0 .3 92 ) 0. 24 7 0. 67 6 0. 11 3 (0 .1 77 ) 0. 27 4 0. 83 8 0. 51 8 0. 21 5 0. 22 3 1. 23 9 0. 33 5 m or tg ag e (0 .3 47 ) 0. 49 5 0. 70 7 0. 48 4 0. 01 5 0. 54 2 1. 01 5 0. 97 8 0. 36 2 0. 39 0 1. 43 6 0. 35 3 d iv er si fic at io n 0. 30 7 0. 34 7 1. 35 9 0. 37 7 (0 .1 45 ) 0. 36 9 0. 86 5 0. 69 2 (0 .4 52 ) 0. 29 5 0. 63 6 0. 12 6 su bj ec tiv e fin an ci al kn ow le dg e (0 .2 27 ) 0. 16 0 0. 75 8 0. 08 3† (0 .2 10 ) 0. 17 6 0. 81 0 0. 23 3 0. 06 7 0. 14 8 1. 06 9 0. 65 2 c on fid en ce in fin an ci al ab ili ty 0. 30 7 0. 15 3 1. 35 9 0. 04 5* 0. 01 8 0. 15 2 1. 01 8 0. 90 4 (0 .2 86 ) 0. 14 0 0. 74 9 0. 03 9* a dv is or pr ef er en ce v er y im po rt an t to us e ad vi so r to fr ee up tim e 0. 61 8 0. 28 2 1. 85 5 0. 02 9* 0. 65 4 0. 31 0 1. 92 3 0. 03 5* 0. 03 6 0. 24 1 1. 03 7 0. 88 1 v er y im po rt an t to us e ad vi so r to pr ev en t lo ss es (0 .3 61 ) 0. 32 5 0. 69 7 0. 26 6 (0 .8 11 ) 0. 34 5 0. 44 4 0. 01 9* (0 .4 50 ) 0. 28 2 0. 63 8 0. 11 1 v er y im po rt an t to us e ad vi so r to im pr ov e in ve st m en t pe rf or m an ce 0. 25 3 0. 34 6 1. 28 7 0. 46 6 (0 .0 23 ) 0. 37 3 0. 97 8 0. 95 2 (0 .2 75 ) 0. 32 4 0. 75 9 0. 39 6 pe rf or m ed ad vi so r ba ck gr ou nd / re gi st ra tio n ch ec k 0. 18 1 0. 28 3 1. 19 9 0. 52 1 0. 30 6 0. 30 3 1. 35 8 0. 31 3 0. 12 5 0. 24 6 1. 13 3 0. 61 2 pr of es si on al de si gn at io n ve ry im po rt an t 0. 48 5 0. 24 7 1. 62 5 0. 05 0* 0. 35 6 0. 27 9 1. 42 8 0. 20 1 (0 .1 29 ) 0. 23 4 0. 87 9 0. 58 1 c om pe ns at io n be lie fs c om m is si on on tr ad in g ac tiv ity im pa ct s ad vi ce re ce iv ed (r ef gr ou p: w ou ld no t af fe ct at al l) w ou ld af fe ct so m ew ha t 1. 01 4 0. 36 1 2. 75 6 0. 00 5* * 0. 73 2 0. 39 1 2. 08 0 0. 06 1† (0 .2 82 ) 0. 37 1 0. 75 5 0. 44 8 w ou ld af fe ct a gr ea t de al 1. 85 0 0. 40 4 6. 35 9 � .0 00 1* ** 1. 17 8 0. 43 8 3. 24 8 0. 00 7* * (0 .6 72 ) 0. 40 5 0. 51 1 0. 09 7† se lli ng in ce nt iv e im pa ct s ad vi ce re ce iv ed (r ef gr ou p: w ou ld no t af fe ct at al l) w ou ld af fe ct so m ew ha t (0 .8 06 ) 0. 42 8 0. 44 7 0. 05 9† (0 .1 53 ) 0. 48 0 0. 85 8 0. 75 1 0. 65 4 0. 43 1 1. 92 3 0. 13 0 w ou ld af fe ct a gr ea t de al (1 .1 55 ) 0. 44 7 0. 31 5 0. 01 0* (0 .4 74 ) 0. 50 0 0. 62 2 0. 34 3 0. 68 0 0. 45 4 1. 97 4 0. 13 4 fe e im po rt an ce (0 .1 14 ) 0. 06 9 0. 89 2 0. 09 8† (0 .0 35 ) 0. 08 0 0. 96 5 0. 65 9 0. 07 9 0. 06 7 1. 08 2 0. 23 5 ps eu do r 2 0. 12 0 †p � .1 0, *p � .0 5, ** p � .0 1, ** *p � .0 01 . 248 m. seay / financial services review 27 (2018) 231-255 sation method (p � 0.05). significant results from each multinomial logit model are detailed below. 4.3.1. fee-only compared with commission-only compensation method elements of financial knowledge were found to be associated with using the fee-only versus the commission-only compensation method, although there were confounding results. first, those who correctly answered the compound interest question had 2.7 times higher odds of using the fee-only than the commission-only method, and a one-unit increase in self-assessed confidence in financial ability was associated with a 36% increase in the odds of using the fee-only method. however, those who correctly answered the inflation question had lower odds of using the fee-only than the commission-only method. these divergent results were surprising, and remained present in a variety of different model specifications. advisor preference and compensation beliefs were also associated with advisor compensation method. as related to advisor preferences, those who found it very important to use an advisor to free up time and those who found a professional designation to be very important had 86% and 63% higher odds, respectively, of using the fee-only method than the commission-only method. however, confounding results were found related to compensation beliefs. while respondents who believed commissions affect advice received were more likely to use a fee-only advisor, respondents who believed selling incentives impacted advice received were less likely to use a fee-only advisor as compared with a commission-only advisor. lastly, household income was found to be associated with advisor choice, as higher income households were more likely to use a fee-only advisor. when compared with households with income of $100,000 or more, those making less than $35,000, between $50,000 and $74,999, and between $75,000 and $99,999 had lower odds of using the fee-only compensation method. none of the variables in the socio-demographics block of characteristics were significant predictors of compensation method in this model (p � 0.05). 4.3.2. commission-and-fee compared with commission-only compensation method one element of financial knowledge was associated with compensation method in this model. respondents who correctly answered the inflation question were less likely to use the commission-and-fee method than the commission-only method. two characteristics related to advisor preference were significant in this model. respondents who found it very important to use an advisor to free up time had 92% higher odds of using the commissionand-fee method; and those who found it very important to use an advisor to prevent losses had lower odds of using the commission-and-fee method. related to compensation beliefs, compared with respondents who felt that advisors receipt of commission on trading activity would not impact advice received, those who thought advice would be a great deal impacted had higher odds of using the commission-and-fee method compared with the commissiononly method. in terms of socio-demographic and financial characteristics, age, education, and household income were associated with compensation method in this model. compared with those aged 65 and older, respondents in the youngest age group of 18 to 34 had 3.4 times higher odds 249m. seay / financial services review 27 (2018) 231-255 of using the commission-and-fee method compared with the commission-only method. compared to those with a college degree, respondents with only some college had 91% higher odds of using the commission-and-fee than the commission-only method. households making less than $35,000 annually had lower odds of using the commission-and-fee than the commission-only method. 4.3.3. commission-and-fee compared with fee-only compensation method one characteristic of financial knowledge was associated with compensation method in this model. a one-unit increase in self-assessed confidence in financial ability was associated with a decrease in the odds of using the commission-and-fee method compared with the fee-only method. in terms of financial characteristics, the total value of nonretirement accounts was also a significant predictor of compensation method in this model. compared with respondents who reported having $1,000,000 or more in nonretirement accounts, those who reported having between $0 and $49,999 were less likely to use the commission-and-fee method than the fee-only method. two elements of socio-demographic characteristics, employment status and gender, were found to be associated with compensation method in this model. compared with respondents who were working full-time, those who identified as being retired or other had lower odds of using the commission-and-fee compensation method than the fee-only method. compared to females, male respondents had higher odds of using the commission-and-fee compensation method than the fee-only method. 5. discussion the primary purpose of this paper was to investigate the characteristics of individuals that are associated with knowing the compensation method used to pay their financial advisor. of those respondents who did know how their financial advisors were compensated, a secondary analysis explored the characteristics between clients who used the commission-only, feeonly, and combined commission-and-fee methods to pay their advisors. using data from the 2015 nfcs investor study, this study is one of the first attempts to explore whether clients knew how they paid their advisors, and how clients make choices between compensation models at the multivariate level. although significant variation was found between clients who knew and did not know how their advisors were compensated, and between compensation models for those who did know how their advisor was paid, some results were somewhat contradictory in nature. given that this study represents the first use of the nfcs investor survey to evaluate advisor compensation methods at the multivariate level, an evaluation of the sample quality in relation to previous research is important. sample descriptive results were largely consistent with previous research that suggested clients who seek financial help from financial professionals are older, have higher levels of income, are better educated, and have higher risk tolerance (finke et al., 2011; grable and joo, 2001; hanna, 2011; marsden et al., 2011). however, most of the clients in this sample were male, despite previous research indicating women were more likely to seek financial help from a professional (joo and grable, 2001). similarly, incongruences were found with bae and sandager’s (1997) suggestion that 250 m. seay / financial services review 27 (2018) 231-255 consumers sought a financial professional because they had lower levels of financial knowledge, as respondents in this sample were very knowledgeable about compound interest, inflation, mortgages, and diversification. however, this pattern is consistent with more recent research indicating that financial knowledge, specifically understanding diversification and mortgages, is positively related with the use of a financial advisor (calcagno and monticone, 2015; seay et al., 2016). regarding subjective financial knowledge, previous research found that clients were most likely to pay for financial advice if they believed themselves to be less financially knowledgeable (finke et al., 2011). however, clients in the sample assessed their own financial knowledge very highly on average. in support of bae and sandager’s (1997) suggestion that certification was a critical component when clients were selecting a financial professional, most of the financial planning clients in this sample found certification to be very important. regarding advisor compensation, previous research that found compensation transparency to be less important to prospective clients (bae and sandager, 1997) was not supported as the clients in this sample found fees and pricing structure to be important factors when opening nonretirement accounts. overall, it appears that the nfcs investor survey provides a reasonable sample of individuals that use financial advisors, but it likely is not perfectly generalizable. moving to the primary analysis, some evidence was generated to support all three hypotheses related to financial knowledge, advisor preference, and compensation beliefs blocks. hypothesis 1 proposed that higher financial knowledge would be associated with knowing the compensation method. this hypothesis was supported for those who correctly answered the diversification question, but it was not supported for the other objective financial knowledge questions nor the two subjective questions related to confidence and ability. this limited result indicates that one’s level of financial knowledge may not inform the level of agency costs that a client is willing to incur in the client/financial advisor relationship, or that it may be a domain specific relationship. specifically, diversification, as compared with the other knowledge questions, may be measuring a higher level of investment sophistication that is more closely tied to scrutiny of investment professionals. hypothesis 2 proposed that clients with strong advisor preferences would be more likely to know the compensation method used to pay their advisors. this hypothesis was supported for those who had performed a background or registration check on an advisor. this finding suggests that clients who are sensitive to agency costs will seek to understand how their advisors are being paid. however, there was no significant relationship found between the level of importance placed on professional designations and knowing the compensation method; which indicates that either these designations are not functioning as a signal of an investment in bonding costs or that professional designations are not a strong enough signal of knowledge and expertise to motivate clients to adjust their monitoring costs. preferences about advisor use (to free up time, to prevent losses, and to improve investment performance) were not associated with knowing the compensation method, indicating that clients who outsourced financial decisions for these specific reasons were not taking agency costs into account when choosing their financial advisor. finally, hypothesis 3 posited that strong beliefs about compensation would be associated with knowing the compensation method used to pay the advisor. results did provide evidence of a positive relationship between finding fees to be important and knowing how 251m. seay / financial services review 27 (2018) 231-255 the advisor was compensated, suggesting that clients who were sensitive to agency costs would be more likely to gain an understanding of how their advisors were being compensated. however, there were no relationships revealed between beliefs about commissions and sales incentives and knowing compensation method. lastly, a follow-up exploratory analysis was conducted to determine characteristics associated with each advisor compensation model among those that were aware. instead of a clear picture, results related to financial knowledge were confounding. clients who correctly answered the compound interest question were less likely to use a commission-only advisor as compared with the other methods; but those who answered the inflation question correctly preferred the commission-only method above commission-and-fee. in terms of self-assessed financial knowledge, the more knowledgeable that a client felt about finances, the less likely they were to use the fee-only method as compared with the commission only method. however, this pattern did not align with confidence in financial ability, as higher confidence was associated with a preference for the fee-only method as compared with the commission-and-fee method. taken together, it is hard to get a clear understanding of how knowledge informs compensation model choice. it should be noted that these results may be due to the wide variation in services that are provided by different advisors operating within each compensation model. this could be an indication that there is something going on beyond compensation model, potentially quality of the financial advisor, which is driving the advisor choice and confounding the analysis. however, these results could also indicate that there is confusion in the market about what each of these fee-models entails. more consistent patterns were found around advisor use preferences. respondents who found it very important to use an advisor to free up time overwhelmingly preferred either method over the commission-only method. alternatively, clients who found it very important to use an advisor to prevent losses preferred the commission-only method, but only when compared with the commission-and-fee method. regarding professional designations, clients who thought it was very important that a financial advisor have a professional designation preferred the fee-only to the commission-only method; but there was no significant preference between the commission-and-fee and fee-only methods. these preferences seem to reveal client beliefs that service level is related to compensation structure. the results echo the patterns revealed in seay et al. (2017), and may indicate that advisors using a fee-only advisor are looking to outsource financial decision making, while those using commission based advisors are looking more for a partner to facilitate their financial decisions, as opposed to offload on to. 5.1. limitations several limitations exist in this study. first, the primary analysis was limited to 750 clients who reported owning nonretirement investments, with the follow-up analysis further limited to 608 individuals who knew the type of compensation model they were using to pay their financial advisor. this sample restriction limits the type of client that is analyzed, and results may not be generalizable to the entire population of financial planning clients in the united states. additionally, the analyses were based on self-reported survey data and perhaps respondents were confused about the compensation method they were using for their financial advisor. it is also 252 m. seay / financial services review 27 (2018) 231-255 possible that respondents were confused about the difference between commissions and incentives, given the confounding results in that area. finally, the nfcs 2015 investor survey did not collect information about compensation based on hourly fees, so either clients who use this type of compensation structure are missing from this analysis or those who use an hourly fee model chose one of the three models available to them in the survey. 6. conclusion using data from the 2015 nfcs investor survey, this study explored the associations between client socio-demographic characteristics, financial and financial knowledge characteristics, advisor preferences, and compensation beliefs and the likelihood of knowing the compensation method they used to pay their financial advisor/broker. a secondary analysis explored the relationship between these client characteristics and the use of either a commission-only, fee-only, or commission-and-fee compensation model for their financial advisor. primary results revealed patterns between client characteristics and knowledge of how their advisors were compensated. specifically, clients with knowledge about diversification; those who took the time and initiative to perform a background or registration check on their advisor/broker, and those who found fees to be important had higher odds of knowing how their financial advisors were being compensated. these results indicate that clients with a higher level of investment sophistication and those who are sensitive to agency costs will seek to understand how their financial advisors are compensated. in understanding this relationship, it is important to point out that 19% of respondents did not know how their advisor was paid. this is consistent with previous literature, and highlights the limited information that many clients gather, or retain, in assessing the use of a financial advisor. for clients who did know the compensation method, results revealed some relationships between client characteristics and preferred compensation models. however, the results also provided evidence of some misunderstanding related to advisor compensation. overall, there are clear differences in the services clients are expecting based on the type of compensation method used, and given clear regulatory trends leveling the standard of care between financial services professionals, it is important to continue to gain an understanding of what services are important to which clients, and how clients will be able to differentiate between financial advisors in the future. 253m. seay / financial services review 27 (2018) 231-255 references bae, s. c., & sandager, j. p. (1997). what consumer look for in financial planners. journal of financial counseling and planning, 8, 9–16. blanchett, d., & kaplan, p. (2013). alpha, beta, and now . . . gamma. the journal of retirement, 1, 29–45. calcagno, r., & monticone, c. (2015). financial literacy and the demand for financial advice. journal of banking & finance, 50, 363–380. certified financial planning board of standards (cfp board). (2017). code of ethics and standards of conduct. (available at https://www.cfp.net/about-cfp-board/proposed-standards). collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307. federal register. (2017). rules and regulations, 82(66), 16902–16918. (available at https://www.gpo.gov/fdsys/ pkg/fr-2017-04-07/pdf/2017-06914.pdf). finke, m. s., huston, s. j., & winchester, d. d. (2011). financial advice: who pays. journal of financial counseling and planning, 22, 18–26. appendix 1 description of key variables in the 2015 national financial capability study (nfcs) variable description compensation method “which of the following types of fees do you pay for your non-retirement investment account(s)?” non-retirement account value “not including retirement accounts, have any investments in stocks, bonds, mutual funds, or other securities?” objective financial knowledge compound interest “suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow?” inflation “imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account?” bonds “if interest rates rise, what will typically happen to bond prices?” mortgage “a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less.” diversification “buying a single company’s stock usually provides a safer return than a stock mutual fund.” subjective financial knowledge “on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall financial knowledge?” confidence in financial ability “i am good at dealing with day-to-day financial matters, such as checking accounts, credit and debit cards, and tracking expenses.” reason for advisor use “below are some reasons that people might use a financial adviser. how important is each of the following to you, personally?” importance of designation “if you were looking for a financial adviser, how important would the person’s professional designations or certifications be in your decision to work with that person?” performed advisor registration/ background check “have you ever checked with a state or federal regulator regarding the background, registration, or license of a financial professional?” sales based compensation impacts advice received “how much do you think each of the following would affect the advice that a financial adviser gives to you?” fee importance “how important to you were the fees and pricing structure when opening your non-retirement investment accounts?” 254 m. seay / financial services review 27 (2018) 231-255 grable, j. e., & chatterjee, s. (2014). reducing wealth volatility: the value of financial advice as measured by zeta. journal of financial planning, 27, 45–51. grable, j. e., & joo. s. h. (2001). a further examination of financial help-seeking behavior. journal of financial counseling and planning, 12, 55–73. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10, 36–62. jensen, m. c., & meckling, w. h. (1976). theory of the firm: managerial behavior, agency costs and ownership structure. journal of financial economics, 3, 305–360. joo, s. h., & grable, j. e. (2001). factors associated with seeking and using professional retirement-planning help. family and consumer sciences research journal, 30, 37–63. https://doi.org/10.1177/1077727x01301002 lachance, m. e., & tang, n. (2012). financial advice and trust. financial services review, 21, 209–226. lusardi, a., & mitchell, o. s. (2009). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness (no. w15350). national bureau of economic research. lusardi, a., & mitchell, o. s. (2011). financial literacy and planning: implications for retirement wellbeing (no. w17078). national bureau of economic research. marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32, 625–643. http://dx.doi.org/10.1007/s10834-011-9258z martin, jr., t. k., finke, m., & gibson, p. (2014). race, trust and retirement decisions. journal of personal finance, 13, 62–71. montmarquette, c., & viennot-briot, n. (2015). the value of financial advice. annals of economics and finance, 16, 69–94. mottola, g. r., & kieffer, c. n. (2017). understanding and using data from the national financial capability study. family and consumer sciences research journal, 46, 31–39. moulton, s., loibl, c., samak, a., & collins, j. (2013). borrowing capacity and financial decisions of low-to-moderate income first-time homebuyers. journal of consumer affairs, 47, 375–403. raskie, s., herbison, g., & martin, j. (2017). do designations attract a more knowledgeable consumer? journal of financial service professionals, 71, 80–96. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals’ advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21, 291–305. seay, m. c., anderson, s. g., lawson, d. r., & kim, k. t. (2017). identifying variation in client characteristics between financial planning compensation models. journal of financial planning, 30, 40–51. seay, m. c., kim, k. t., & heckman, s. j. (2016). exploring the demand for retirement planning advice: the role of financial literacy. financial services review, 25, 1–20. tversky, a., & kahneman, d. (1974). judgment under uncertainty: heuristics and biases. science, 185, 1124–1131. 255m. seay / financial services review 27 (2018) 231-255 pii: s1057-0810(97)90028-7 financial services review, 6(i): 1-17 issn: 1057-0810 copyright © 1997 by jai press inc. all rights of reproduction in any form reserved. the impact of inflation on roe, growth and stock prices frank k. reilly using the constant growth dividend discount model (ddm), it can be shown that the critical factor which determines whether common stocks will be able to be an inflation hedge is the growth rate o f dividends. in turn, the growth of dividends is mainly impacted by the aggregate return on equity (roe). using the dupont formula, it is clear that the main variable that drives the aggregate roe in an inflationary environ ment is the profit margin. following from this background, this article updates and extends an earlier analysis that involves an analysis of roe and its components for the 40-year period 1956-1995. the analysis demonstrates that the aggregate roe is currently at about the same level as in the 1960's, but the components have changed, that is, there has been a decline in total asset turnover and profit margin, but a significant increase in financial leverage that has compensated for the declines in turnover and profit margin. it is further shown that there have been periods ofhigh and low inflation since 1956, and the negative impact of inflation of the implied growth rate is confirmed, which helps explain why investigators find consistent empirical results that common stocks are poor inflation hedges. i. introduction an article by fuller and petty (f&p, 1981) using data through 1979 considered the ability of common stocks to provide a hedge against inflation. using the constant growth dividend discount model (ddm), it was shown that the critical factor which determined whether common stocks would provide the higher rates of return required of an inflation hedge is the growth rate of dividends (g). in turn, it is known from babcock (1970) that g is a func tion of the retention rate and the return on equity (roe). finally, they employed the dupont formula which shows that roe is composed of: (a) the net profit margin (pm), co) the total asset turnover (tat), and (c) a financial leverage variable equal to assets/equity (lev). it is shown that the critical variable that constrained the increase in roe (and, therefore, the increase in g) was the profit margin. having identified the profit margin as the culprit, the authors conclude the article on an optimistic note because they contend that management frank k. reflly • bernard j. hank professor of business administration, university of notre dame, notre dame, in 46556. 2 financial services review 6(1) 1997 as of 1980 had recognized the importance of maintaining the profit margin. the concluding section of the article assumes stability for the profit margin and considers the effect of sev eral scenarios where stocks sell at alternative price to book value ratios (p/bv). this paper extends this interesting analysis in several ways. first, we include the 16 years since the original presentation, second we expand the analysis of the profit margin, and finally, we analyze the roe using a five part analysis suggested by cohen, zinbarg and zeikel (1987). the five part analysis of roe provides additional insights into what has caused the changes in roe during the past 19 years and indicates what might transpire in the years ahead with and without changes in the rate of inflation. the first section contains a brief review of the dividend discount model (ddm) and its implications related to inflation and stock prices. the second section considers the dupont formula breakdown and analyzes changes in the three components for the fortune 500 s&p 400 since 1956 with an emphasis on the period since 1979. in the third section, we discuss the five part breakdown of roe and examine these five components for the s&p 400 during the period 1977-1995 (the data were not available prior to 1977). section four contains an analysis of the relationships among inflation, rates of renan on stock, roe, and the components of roe. in section five, we consider what happens to stock returns, roe components and earnings growth during periods of high and low rates of inflation. we con clude with a discussion of the outlook for roe and growth in an environment of low and high rates of inflation. ii. the dividend discount model (ddm) readers are familiar with the reduced form of the dividend discount model as it would apply to the aggregate stock market: 1 where p = the price of stocks; d 1 p = , (1) k g d~ = expected dividend in period 1 [d o (l+g)]; k = the required rate of return on stocks; and g = the expected growth rate of dividends for common stocks. using this model, it is possible to consider: (a) what will happen if expectations change regarding the rate of inflation, and (b) what must happen if common stocks are to be a complete hedge against inflation. a hedge is a transaction intended to safeguard against loss on another investment. a hedge against inflation, then, is the acquisition of an asset that will safeguard against an increase in the general price level. in the case of common stocks, the expected rate of return should increase in line with the required rate of return that includes the expected rate of infla tion. the point is, we know that the required rate of return (ki) is determined by a real risk free rate, the expected rate of inflation, and a risk premium. 2 therefore, given a change in the expected rate of inflation, there will be an increase in k for all risky assets including com mon stocks. given a change in k, the crucial question becomes: what will happen to the value the impact of inflation 3 of the asset to ensure that the investor will receive the higher nominal required rate of return (k)? one way to view this is to transform the ddm valuation model as follows: if then d 1 ,p= k g d i k = + g . (2) p given this specification, if there is an increase in the expected rate of inflation and nothing happens to the expected growth rate of dividends of finns (i.e., there is no change in g or d i ), we can see that stock prices must decline--that is, the p must decline until there is an increase in the d1/p term to compensate for the increase in the required return. this see nario is similar to a bond where the price of the bond must necessarily adjust to increase the expected yield because there is no change in the expected cash flows (interest payments). clearly, during such a period of price adjustment, the investor who owns stocks (or bonds) will experience large negative returns. another possibility is that the growth rate of dividends (g) will increase by approxi mately the increase in the rate of inflation. if this occurs, stock prices will not change, because the spread between k and g will change only slightly. the expected return on stocks (k) will increase because of the increase in g and an increase in the dividend yield because of an increase in d 1 which equals do(l + g). therefore, in this scenario the inves tor's expected rate of return (k) has increased in line with a change in expected inflation, and common stocks will be a complete inflation hedge. the point is, g must increase by almost the change in the expected rate o f inflation if common stocks are to be a complete inflation hedge without stock prices declining. such an increase in g is the implicit assump tion made by observers who contend that common stocks should be an inflation hedge. for example, when jahnke (1975, p. 74) employs the dividend model to explain changes in stock prices, he states, "thus common stocks should serve as a hedge against inflation to the extent that changes in the rate of inflation are mirrored in the dividend growth rate." hi. a secular analysis of roe and its dupont components as noted, the crucial factor affecting the growth of dividends is the aggregate roe. in turn, we can analyze roe using the traditional dupont formulation which includes three com ponents as follows: net income sales net income total assets x x ( 3 ) equity total assets sales equity as noted by fuller and petry, these ratios are not available for the standard & poor's series prior to 1977. alternatively, they are available for the fortune 500 industrial series which 18.0 financial services review 6(i) 1997 18.0 1 7 . 0 1 6 . 0 1 5 . 0 1 4 . 0 ~ 13.0 12.0 11.0 io.o 9.0 8.o v v //\ i / / v ' ' ' : : : : : : : : i i i i t . . . . : : ; : : . . . . . . . . . . (f~ n tj') lo ~ po ~ td @y n i n td od fo ~-. i ~ qo ~ io ~d (~ years figure l . time series plot of return on equity (roe) for fortune 500 s&p 400: 1956-1995. 1 7 . 0 1 6 . 0 1 5 . 0 1 4 . 0 1 3 . 0 1 2 . 0 1 1 . 0 1 0 . 0 9 . 0 8 .0 is correlated very highly to the s&p 400 industrial series. table 1 contains the dataand the ratios for the 40-year period 1956-1995 wherein the data are for the fortune 500 during the period 1956-1976 and for the s&p 400 for the period 1977-1995. figure 1 shows that the roe ranged from about 10 to 17 percent, but began at about 13 percent and was at 13 per cent in 1993 prior to a spurt to about 17 percent in 1994 and 1995. the significant question is: "how did the three components of roe contribute to this performance?" we know that 9.0 , 9.0 8.o 7 . 0 6 . 0 < 0 n, 5.o 4.o 3 . o \ v 6 .0 7 . 0 6 . 0 5 . 0 4 . 0 3 . 0 figure 2. time series plot of return on assets (roa) for fortune 500 s & p 400: 1956-1995. 2 . 0 , : : : : : : : : : : : : : i i i : : : ,' : : ; : : *. " : : : i i i i i i i i 7, .0 io lid lid io ~-* i~. go lid 40 gd c/) o~ cd cll o~ 1~ o) old o) oi ci~ c~ cd cl) o'j years 1.3 5 1.4 " k . . / 1.2 1.1 1.0 0.9 the impact of inflation 1.3 1.2 1.1 1.0 0.9 0.8 . : : .~ : : '. : : '. : : : ; i ,~ ~ [ [ ~ [ i i i i i i i i i 1 i i i i i i i 0.8 years figure 3. time series plot of total asset turnover (tat) for forame 500s&p 400:1956-1995. the product of the first two ratios [total asset turnover (tat) and net profit margin (pm)], equals return on assets (roa). the plot in figure 2 shows that the roa has generally declined from about 8 percent in 1956 to 5 percent in 1994-1995 after a trough of 2.9 per cent in 1991. the fact is, both tat and pm contributed to the overall decline. as shown in table 1 nominal and real rates of return for the s&p; 500 return on equity and dupont components for fortune 500 and s&p 400:1956-1995 s~p 5oo ~n~ ady total u.s. lnflas&p 500 total asset profit return on t. assets to return on return tion rate % tr turnover* margin* assets* equip* equity* 1956 6.56 2.86 3.59 1.25 6.6 8.3 1.59 13. lo 1957 -10.78 3.02 -13.40 1.27 6.2 7.9 1.57 12.30 1958 43.36 1.76 40.88 1.15 5.4 6.2 1.53 9.50 1959 11.96 1.50 10.30 1.17 6.1 7.1 1.55 11.00 1960 0.47 1.48 -0.99 1.16 5.7 6.6 1.53 10. i0 1961 26.89 0.67 26.04 1.12 5.6 6.2 1.54 9.60 1962 -8.73 1.22 -9.83 1.16 5.9 6.8 1.55 10.60 1963 22.80 1.65 20.81 1.17 6.1 7.1 1.56 11.10 1964 16.48 1.19 15.11 1.19 6.5 7.7 1.58 12.10 1965 12.45 1.92 10.33 1.18 6.7 8.0 1.63 13.00 (continued) 6 financial services review 6(1) 1997 table 1 (continued) s&p 500 lnfi adj total u.s. lnflas&p 500 return tion rate % tr total asset profit return on t. assets to return on turnover* margin* assets* equity* equity* 1966 -10.06 3.35 -12.98 1.18 6.6 7.8 1.69 13.20 1967 23.98 3.04 20.32 1.13 6.0 6.8 1.74 11.80 1968 11.06 4.73 6.05 1.12 5.6 6.7 1.82 12.20 1969 -8.51 6.11 -13.77 1.11 5.5 6.1 1.87 11.50 1970 4.01 5.49 -1.41 1.07 4.7 5.0 1.90 9.50 1971 14.31 3.36 10.60 1.10 4.7 5.1 1.90 9.80 1972 18.98 3.41 15.05 1.15 5.0 5.7 1.90 10.90 1973 -14.66 8.80 -21.56 1.20 5.8 7.0 1.96 13.70 1974 -26.47 12.20 -34.47 1.33 5.2 6.9 2.03 14.10 1975 37.20 7.01 28.21 1.29 4.9 6.5 2.23 14.43 1976 23.84 4.81 18.16 1.32 5.1 6.7 2.01 13.50 1977 -7.18 6.77 -13.07 1.27 5.1 6.5 2.08 13.47 1978 6.56 9.03 -2.26 1.27 5.2 6.6 2.15 14.11 1979 18.44 13.31 4.53 1.30 5.6 7.2 2.20 15.90 1980 32.16 12.40 17.81 1.31 4.9 6.5 2.23 14.43 1981 -4.91 8.94 -12.71 1.28 4.9 6.2 2.25 13.97 1982 21.41 3.87 16.88 1.17 4.0 4.6 2.31 10.73 1983 22.51 3.80 18.03 1.15 4.4 5.1 2.28 11.63 1984 6.27 3.95 2.22 1.27 4.6 5.8 2.39 13.86 1985 32.16 3.77 27.36 1.15 3.8 4.4 2.54 11.21 1986 18.47 1.13 17.15 1.07 3.7 4.0 2.58 10.33 1987 5.23 4.41 0.79 1.08 4.7 5.1 2.62 13.35 1988 16.81 4.42 11.87 0.98 5.5 5.3 3.03 16.19 1989 31.49 4.65 25.65 0.97 5.1 4.9 3.17 15.56 1990 -3.17 6.11 -8.75 0.97 4.4 4.2 3.26 13.74 1991 30.55 3.06 26.67 0.94 3.1 2.9 3.24 9.39 1992 7.67 2.90 4.64 0.95 3.4 3.2 3.55 11.45 1993 9.99 2.75 7.05 0.92 3.8 3.5 3.76 13.25 1994 1.31 2.68 1.33 0.95 5.3 5.0 3.40 17.18 1995 37.43 2.54 34.03 0.96 5.3 5.1 3.32 16.95 notes: = the ratios for the period 1956-1976 are for the fortune 500 industrials: for the period 197% 1995 for the s&p 400. the impact of inflation 7.0 7 7.0 6.5 6.0 5.s 5.0 q. 4.s &o &s 3.0 l / 2.~ i , i l i i t ,i i i i , i i i t t , , , , i , , , , , , i i i i , , , ,. i cb at ~ ~id ~ (it (d (it = all (11 d y e a r s figure 4. time series plot of profit margin (pm) for fortune 500 s&p 400: 1956-1995. 6.5 6.0 5.s v 5.0 4.5 4.o 3.5 3.0 2.5 i n (ira figure 3, the tat ratio started at about 1.25, peaked during the period 1974-1980 (when the f&p article was published) and subsequently declined to 0.96 in 1995. notably, this decline was not envisioned by f&p. specifically they note that one should expect apositive bias for the tat ratio during periods of inflation because sales should be affected by infla 4.0 t ",-. 4.0 &o+ 4~3.o :> lli 2,5, .j 2.5 2.o + ~ ;2.o l . s+ ~ ~ l .s 1.0 1.0 (1~ (1~ . ~ gi~ (l l m ~ (d cli (d (n (it y e a r s figure 5. time series plot of leverage (lev) for fortune 500 s&p 4130: 1956-1995. 8 financial services review 6(1) 1997 tion (i.e., sales are in current dollars), while most asset values are based on historical costs which are not adjusted for higher prices. as shown in figure 4, the pm also declined from 6.6 percent in 1956 to a low of 3.1 percent in 1991 followed by a recovery to 5.3 percent in 1994-1995. again, this ratio should be constant or possibly increase during inflation because sales (revenue) are in cur rent dollars, while many costs such as depreciation do not fully reflect the inflation. as of 1980, f&p felt that management recognized the problem regarding declining profit mar gins and was responding to it. unfortunately, the results in figure 4, for the subsequent period 1979-1993 do not support this expectation. specifically, the pm declined sharply as expected during the 1981-1982 recession, recovered to over 5.5 percent in 1988, but declined to 3 percent in 1991 and finished at 5.3 percent in 1995. clearly, the overall sec ular decline shown in figure 4 extended through 1993. we know from the prior equations that roa times the leverage ratio (t. assets/ equity) equals roe. given the overall declines in tat, pm and roa, and an roe that did not experience a secular decline, it is obvious that there must have been a signi f icant increase in the financial leverage multiplier (lev) that offset the decline in roa. the graph of the financial leverage ratio in figure 5 clearly confirms that financial leverage more than doubled from 1.59 in 1956 to a peak of 3.76 in 1993 prior to finishing at 3.32 in 1995. clearly, the overall increase in this leverage ratio offset the decline in roa. in summary, these results show that the u.s. economy maintained and even increased its roe over the period 1956-1995, but the values for the three dupont components have changed. specifically, we have l ower operating efficiency as reflected by tat, and lower profit margins, but these declines have been offset by substantially h igher financial lever age which implies higher financial risk. iv. extended dupont analysis of roe as noted, beyond the three component dupont analysis, some authors have suggested a more extended breakdown that provides insights into the effect of interest expense and the tax rate on the roe. the presentation in table 2 begins with an operating profit margin (ebit/sales) and multiplies by the tat to derive ebit/t. assets (operating return on total assets). to consider the income effect of financial leverage, you subtract: int. exp./ t.assets. this interest expense ratio reflects the effect of fixed income debt on the balance sheet, but also the level of interest rates. the result is nbt/t. assets. the next column reflects the balance sheet leverage or the financial leverage multiplier (similar to the basic dupont analysis). the product of these ratios generates a before tax return on equity (nbt/ equity). the next column is referred to as the "after-tax retention rate" and reflects the impact of changing tax rates over time (1 (tax/nbt)). multiplying by this ratio generates roe (net inc/equity). in summary, the breakdown is as follows: e b i t sa les int . exp. t. asse t s ( tax x . x x ~,1 i = r o e sales t. a s s e t s t. asse t s equ i t y nb't) (4) this analysis can only begin in 1977 when these data are available for the s&p 400 series. the time series plot of roe for 1977-1995 is contained in figure 6. the operating profit the impact of lnflation 18.0 9 18.0 17.0 16.0 15.0 ~ 14.0 ~" 13.o 1zo 11.0 10.0 9.0 11.5 11.0 ~__1 o's ~ 10.0 i- ] 9.s 9.0 8.5 i l i i i i i i i i i i i i i i " i years figure 6. time series plot of return on equity (roe) for s&p 400: 1977-1995. 17.0 1$.0 15.0 14.0 13.0 12.0 11.0 10.0 9.0 margin (figure 7) generally declined from 11.5 percent to a low of about 8 percent prior to a recovery in 1994, 1995. as noted previously, the tat (figure 8) declined steadily from about 1.30 to 0.96. the debt effect in terms of interest expense (figure 9) increased steadily from 1.84 to 3.35 in 1989 (the peak in financial leverage) and then declined to 2.03 in 1995 mainly due to lower interest rates but also a decline in balance sheet debt. this reduction in debt on the balance sheet was not adequately reflected in the t. assets/equity ratio (figure 10) because of reduced equity due to numerous share repurchase programs (i.e., total com mon equity for the s&p 400 was 188.25 in 1990 and 188.88 in 1994 implying that 1z0 1z0 8.0 j 11.5 11.0 10.5 10.0 9.5 9.0 8.5 8.0 i ~ i~ i~, m • eo 00 ~ ~ 10 i l l ~ o~ ; cd ot ~/~ years figure 7. time series plot of ebit/sales for s&p 400: 1977-1995. 10 financial services review 6(1) 1997 t .4 1.4 1.3 1.2 i- 1.11 0.9 0.8 i i i i l i i i i i ! ! i i i i i ~ ~ ¢b o ~" . t~ g t n w ~ ~ o e4 m years figure 8. time series plot of total asset turnover (tat) for s&p 400: 197%1995. 1.3 1.2 1.1 1.0 0.9 i 0.8 increases in retained earnings were offset by share repurchases). overall, the financial leverage multiplier increased from 2.08 to a peak of 3.76 in 1993 and 3.32 in 1995. finally, the after-tax retention ratio (figure 11) indicated a very accommodative tax policy. specifically, the after-tax retention rate went from about 50 percent in 1977 to a peak of over 64 percent in 1994. this reduction in the effective tax rate (from 50 percent to 36 percent) was a significant factor contributing to the peak roe in 1994. in summary, this extended analysis is consistent with the traditional dupont analysis wherein it reflects the decline in tat, some weakness in the operating profit margin, and it highlights the increase in financial leverage on the income statement as well as the bal 3.5 -, -~ 3.s 3.3 ' 3.3 3.1 3.1 o) 2.9 :" 2.9 k,.. i11 2.7 2.7 ¢) 2.s . 2.5 <. i-" 2.3 2.3 2.1 2.1 1 .s t .9 z 1.7~ ~'1.7 1.s 1.s years figure 9. time series plot of int. exp./t. assets for s&p 400: 197%1995. th¢ in~e t af lnflatian 11 4.0 4.0 3.s 0 u.j 3 . 0 u) i ' u.i 2.s c~ u) i-2.o 3.5 3.0 2.5 2.0 1.s ¢1~ ob ¢~ ¢~ ¢d o~ ¢d od o~ ¢d ¢d oi clip a ¢m ~ ~ m ¢d years figure 10. time series plot of t. assets/equity for s&p 400: 1977-1995. 1.s table 2 extended dupont analysis of return on equity s&p 400 industrial: 1977-1995 ebit/ sales~. ebit/r. int. e a ' p j nbt/ ,7' assets/ nbt/ ! tax net. lncj net. lncj year sales assets assets 7". assets t.assets equity equity n b t equity sales 1977 1978 1979 1980 1981 11.52 1.27 14.65 1.84 12.80 11.55 1.27 14.64 1.94 12.70 11.93 1.30 15.51 2.04 13.47 10.92 1.31 14.35 2.39 11.96 10.80 1.28 13.81 2.78 11.02 2.08 26.58 50.69 13.47 5.11 2.15 27.25 51.79 14.11 5.19 2.20 29.58 53.74 15.90 5.57 2.23 26.67 54.11 14.43 4.92 2.25 24.77 56.38 13.97 4.86 1982 1983 1984 1985 1986 9.70 1.17 11.39 2.90 8.50 10.33 1.15 11.92 2.63 9.29 10.26 1.27 13.08 2.74 10.34 9.59 1.15 11.02 2.67 8.35 9.10 1.07 9.71 2.68 7.03 2.31 19.63 54.66 10.73 3.95 2.28 21.19 54.91 11.63 4.42 2.39 24.69 56.14 13.86 4.55 2.54 21.25 52.76 11.21 3.84 2.58 1 8 . 1 6 56.91 10.33 3.75 1987 1988 1989 1990 1991 10.29 1.08 i 1.14 2.55 8.61 11.62 0.98 11.37 3.02 8.36 11.48 0.97 11.09 3.35 7.74 10.48 0.97 10.14 3.28 6.86 8.34 0.94 7.88 3.02 4.86 2.62 22.54 59.23 13.35 4.71 3.03 25.32 63.93 16.19 5.46 3.17 2 4 . 5 1 63.47 15.56 5.08 3.26 22.39 61.39 13.74 4.35 3.24 1 5 . 7 6 59.59 9.39 1992 1993 1994 1995 8.08 0.95 7.69 2.56 5.12 8.48 0.92 7.78 2.23 5.55 10.49 0.95 9.94 1.93 8.01 10.60 0.96 10.14 2.03 8.12 3.55 1 8 . 2 0 62.91 11.45 3.39 3.76 20.89 63.43 13.25 3.84 3.40 27.22 64.32 17.18 5.33 3.32 26.97 63.79 16.95 5.33 12 financial services review 6(1) 1997 ance sheet. the most significant new insight provided is the very positive effect of a decline in the tax rate during this period. beyond an analysis of the secular trend for roe and its components, because we are interested in the effect of inflation on roe, growth and the rate of return on common stock, it is important to analyze the specific relationship of inflation to stock returns, roe and the components of roe. v. the relationship among inflation, stock returns, roe and its components table 3 contains the correlation matrix among inflation, common stock returns, roe, and the components of roe for the 40 year period 1956-1995. the correlations among infla tion and the other variables for the period 1956-1995 provide results generally consistent with past results. the negative relationship (-0.24) between inflation and stock returns is very consistent with almost all prior studies (jaffe & mandelker, 1976; fama, 1981) which imply that common stock have been a poor inflation hedge. the positive correlation between inflation and tat is consistent with the earlier discussion of a positive bias because sales are impacted by inflation, while the historical cost of assets are not. the negative correlation between inflation and the pm is consistent with past results, even though there is a tendency toward positive results. these results imply that fn'ms are typi cally not able to pass cost increases along to customers. leverage has almost no correla tion with inflation because it displayed a constant increase doing a period when inflation was fairly volatile, the relationship between inflation and roa was also very small reflecting the positive relationship with tat and the negative correlation with pm. the significant positive correlation between inflation and roe is somewhat surprising because of the mixed results with the components. 66 ~ r 66 64 62 ill i . z m lu i ' 56 i z u u j 4 : ! s4 ijj iv" s2 50 years t a x figure 11. time series plot of 1 ~ for s&p 400: 1977-1995. 64 62 60 s8 56 54 52 50 the impact of inflation 13 table 3 correlation matrix of inflation, equity rates of return and components of the return on equity inflation stock trt* tat pm lev roa roe a. annual 1956-1995 inflation j stock -0.24 trt* tat 0.47 -0.19 pm -0.10 -0.26 0.37 lev 0.06 0.15 -0.69 roa 0.13 -0.29 0.71 roe 0.45 -0.15 -0.01 m -0.67 r 0.91 -0.80 0.19 0.45 0.13 note: * total rate of return on the s&p 500 as computed by ibbotson associates. in summary, inflation had a negative relationship with stock returns and profit margins although the profit margin had a significant relationship with roa and roe. the point is, inflation had a negative relationship with the profit margin which will, in turn, reduce roe and expected growth. this implies that there should be a poor relationship between inflation and expected growth which is examined in the next section that considers periods of high and low inflation. vi. analysis of high and low inflation periods in the f&p article (1981) the authors examined the differential results during two sepa rate periods including one of low inflation (1956-1967) and a subsequent period of high inflation (1968-1979) and showed that the operating and stock return results were very different. this analysis can be extended to the recent period by expanding the sec ond period to include two additional years of high inflation (1968-1981) and by adding the recent period of relatively low inflation (1982-1995). figure 12 contains a time series plot of inflation and the three periods are identified. the average results for the series in table 1 during these three periods are in table 4. notably, the nominal returns for the s&p 500 are clearly higher during the periods of low inflation (11 and 17 per cent) compared to the high inflation period (7.5 percent). the real difference was in inflation-adjusted returns where the returns on stock during the period of high inflation was basically zero compared to real returns of 9 and 13 percent during the periods of low inflation. the effect on roe components was not as significant--the tat moved up and down with inflation, the profit margin declined during the period of high infla tion, but continued down during the subsequent period of low inflation. as a result, roa showed a steady decline. as noted earlier, the leverage ratio increased steadily, which drove the roe higher. 14 financial services review 6(1) 1997 table 4 time period averages for stock returns roe components, and nominal and real earnings growth u.s. /n/7 ad/ s& p % inflation s&p 500 annual growth rate total % price % total nominal real return return return tat pm roa lev roe earnings earnings 1956-1967 11.28 1.97 9.18 1.18 6.12 7.20 1.59 11.45 4.40 2.46 (12 yrs) 1968-1~81 (14 yrs) 7.51 7.60 0.08 1.22 5.12 6 . 2 8 2 . 0 2 12.75 8.11 0.52 1982-1995 17.01 3.57 13.02 1.04 436 4.52 2_96 13.20 5.34 1.80 (14) yrs) the effect of inflation on the growth rate of nominal earnings showed a positive impact going from about 5 percent in 1958-1967 to almost 9 percent in 1968-1981, and down to about 5 percent in 1982-1995. notably, there was a higher rate of growth for real earnings during the periods of low inflation, that is, about 3 percent growth during the peri ods of low inflation versus 1 percent growth during the high inflation period. a) inflation and growth the results in tables 3 and 4 clearly indicate that the ultimate effect of inflation on stock returns is negative. we expected this because we envisioned using the ddm that the 14.0 14.o 12.0 10.0 ~. 8.0 .• $.0 m .4.0 2.o f j.i low inflation + * * * + l o w inflation -~--* --* v " inflation + + ~ : : , . , , : : : : : ' , , . . . . . . . . . . . ° g ad lid ~ ~ y e a r s ~ 0 . * , , | t t l t | t l g g g figure 12. time series plot of inflation: 1956-1995. 12.0 10.0 8.0 6.0 4.0 2.0 ' 0 . 0 g l l g l l the impact of inflation 15 12.0 10.0 8.0 , . o ~.1 4.0 2.0 0.0 i " ' ' ~ ¢ ' ~ ~ ~ ~°1 a i 12.0 ilo.o ~ ~ 8.0 , $.0 i 4.0 ~ 2.o 0.0 years figure 13. time series plot of 4 year moving averages of inflation and implied growth rate for s&p 400: 1956-1995. growth rate of dividends (g) would generally not be able to adjust for changes in the required return (k) caused by changes in inflation. this implies an important direct compar ison between g and inflation. figure 13 contains such a comparison between a time series plot of four year moving averages for the two series. a moving average is used to help smooth two fairly volatile series, with the understanding that the four year averages should be plotted at its center point (the value for the four year period 1956-1959 is recorded at the end of 1957). the two lines in figure 13 demonstrate why stocks have done so much better during periods of low inflation. specifically, during the two periods of low inflation that prevailed at the beginning and end of the period, the implied growth rate (equal to roe times the retention rate) was substantially larger than the rate of inflation which implies a decline in the k-g spread and an increase in stock prices. in contrast, during the period of high infla tion the two rates were at best equal, and during several years (1979-1983), the inflation rate exceeded the growth rate which implies an increase in the k-g spread. an example of the ultimate positive effect of low inflation is 1995 when the aggregate roe and the implied growth rate for the s&p 400 increased while the inflation rate declined. the result was a return on stocks of over 37 percent. vii. summary and conclusion the purpose of this paper is to extend and expand the analysis of the relationship between inflation and stock returns by examining the effect of inflation on the factors that affect roe and ultimately the growth of earnings and dividends. following a brief review of the ddm and what needs to occur for stocks to be an inflation hedge, it was demonstrated that the critical variable was what happened to roe, which was determined by what happened to the dupont components and especially the profit margin during periods of inflation. an analysis of the secular trend over 40 years showed an overall decline in tat and the pm, 16 financial services review 6(1) 1997 with a secular increase in the financial leverage multiplier as an offset. the extended dupont analysis for the recent period generally confirmed the long-run results, but also showed the positive effect of a lower effective tax rate during the recent 10 year period. the correlation analysis confirmed prior results which showed a negative relationship between stock returns and inflation (stocks are a poor inflation hedge) and between profit margins and inflation which helps explain the stock return results. an analysis of stock returns and roe results during periods of relatively low inflation (1956-1967 and 1982 1995) versus a period of high inflation (1968-1981) confirms these results because real stock returns were significantly higher during the periods of low inflation and there was clearly a higher growth rate of real earnings during periods of low inflation. finally, the superior returns on stocks during periods of low inflation can be explained by the direct comparison of inflation and the implied growth rate of earnings. specifically, during peri ods of low inflation the implied growth rate of earnings generally exceeds inflation, while during periods of high inflation, the implied growth rate of earnings is equal to or less than the rate of inflation. in conclusion, the analysis in this paper confirms and extends the prior analysis which explains why inflation is detrimental to stock returns and demonstrates that this has contin ued through 1995. it is noteworthy that the u.s. stock market has generally benefited from relatively low inflation since 1982 and especially during 1995 when the implied growth rate increased while the inflation rate was declining, resulting in very high stock returns. it appears that 1996 will likewise be a good year due to a high roe which implies continued growth and low inflation. although it is always difficult to project the future, these results should help investors understand the importance of concentrating on inflationary expectations and the relationship between inflation, roe, and growth. there is obviously a large difference in stock returns during periods of high versus low inflation or during years when the rate of inflation expe riences a large increase or decline and these results help explain the reason for this difference. acknowledmnent: the author acknowledges the research assistance of darren seidel, sumner weymouth, maria vivero, and kevin poppink. notes 1. foradetailedderivationofthisreducedformmodel, see reilly and brown, 1997,chapter 13. 2. for a discussion of this, see reilly and brown, 1997, chapter 1. references babcock, guilford c. (1970). the concept of sustainable growth. financial analysts journal, 26(3 may-june), 236-242. cohen, jerome b., zinbarg, edward d., & zeikel, arthur. (1987). investment analysis and portfolio management, 5th ed. homewood, il: richard d. irwin. fama, eugene f. (1981). stock returns, real activity, inflation and money. american economic review, 71(4 september). the impact of inflation 17 fuller, russell j., & perry, glenn h. (1981). inflation, return on equity, and stock prices. the jour nal of portfolio management 7(4 summer), 19-25. jaffe, jeffrey f., & mandelker, gershon. (1976). the 'fisher effect' for risky assets: an empirical analysis. journal of finance, 31(2 may), 447-458. jahnke, william w. (1975). what's behind stock prices? financial analysts journal, 31(5 septem ber-october), 69-76. reilly, frank k., & brown, keith c. (1997). investment analysis and portfolio management, 5th ed. fort worth, tx: the dryden press. from the editor this issue contains volume 29 issue 1 of financial services review (fsr). i would like to thank the board and members of the academy of financial services for their continued support. i continue to work in broadening the scope of articles, while still focusing on individual financial management and personal financial planning. i encourage authors to reach out when discussing implications of their findings in a more comprehensive way. as such, all articles in the journal more appropriately relate to financial planning issues. the lead article “optimism, overconfidence, and insurance decisions” is coauthored by jennifer coatsa and vickie bajtelsmit, both at colorado state university. the authors present experimental evidence regarding overconfidence, optimism and insurance decisions. they distinguish between an individual’s optimism bias and overconfidence bias, a contribution particularly important for understanding insurance decisions related to risks beyond the purchaser’s control. their results show that optimistic participants incur a higher total cost of risk and are more likely to underinsure than non-optimistic participants, even when purchasing insurance maximizing expected payoffs. they also find that overconfidence does not significantly affect the decision to insure, participants with higher overall overconfidence show larger differences in insurance behavior when the risk of loss arises from their own mistakes. the second article “the impact of using financial technology on positive financial behaviors” is coauthored by qianwen bi, utah valley university, lukas r. dean, utah valley university, tao guo, william paterson university, and xu sun, utah valley university, the authors use the 2013 survey of consumer finances data to explore the impact of financial technologies on households’ positive financial behaviors. the authors find that only planning technologies (e.g. direct deposit and computer software) are positively related to households’ engagement in positive financial behaviors. they also find that the impact of transaction technologies (e.g. using atm card, credit card, phone banking, and computer banking) is negative. the third article, “using investor utility to determine portfolio choice with reits” is coauthored by wei feng, lynn university, travis l. jones, florida gulf coast university, and marcus t. allen, florida gulf coast university. the authors examine the decision of individual investors to allocate a portion of their existing investment portfolios to reits. they derive the risk preferences of investors represented by their benchmark portfolios of stocks and bonds and then use the risk preferences to determine portfolio decisions regarding reits. their analysis shows that investors with lower risk aversion tend to have a more 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) v–vii substantial stock component in their benchmark porfolio and will obtain higher risk-return benefits from adding reits. the final article, “demographic and psychological differences between chapter 13 bankruptcy filers and non-filers” is coauthored by scott e. kehiaian, southern new hampshire university, albert a. williams, nova southeastern university, and carolyn l. bird, north carolina state university. in this article the authors find financial, demographic, and psychological differences between chapter 13 filers and non-filers. they also show that financial training reduces the likelihood of filing for personal bankruptcy and males are twice as likely as females to be filers. a single person is less likely to file than a married person and homeowners are more likely than renters to be filers. increases in education, religious commitment, and parents’ income reduce the likelihood of filing. increases in the psychological factors, self-efficacy, locus of control, and self-control, reduce the likelihood of filing for chapter 13 bankruptcies. thank you to those who make the journal possible, especially the referees and contributing authors. over the past year, the following reviewers provided excellent reviews of the articles you enjoyed within the pages of financial services review. i would like to send a special thank you to the many reviewers that have significantly contributed to the quality of our journal by providing timely and thorough reviews of the submissions to our journal. fenaba addo lisa fiksenbaum thomas krueger jinfei sheng stephen agnew john gathergood marie-eve lachance william skimmyhorn abdullah al-bahrani john grable kyre lahtinen christina stoddard arthur allen adam greenberg andre liebenberg ning tang somer anderson john grigsby hanna lim sharon tennyson anders anderson michael guillemette ana luiza paraboni ruilin tian nikolaos artavanis andreas hackethal annamaria lusardi colleen tokar asaad kremena bachmann nathan harness john lynch jr sami vahamaa vickie bajtelsmit christine harrington zdravko marjanovic neal van zutphen bhanu balasubramnian stuart heckman terrance martin bruce vanstone mirco balatti robin henager greene camilla mazzoli christian walkshäusl michael batty robert henderson stephan meier william walstad levon blue thorsten hens steffen meyer tom warschauer paola bongini hal hershfield brianna middlewood jamie weathers peter brady russell james young park jaya wen sonya britt-lutter thomas jansson darshak patel melissa wilmarth j. michael collins jing jian xio cliff robb danielle winchester brenda cude jesse jurgenson david robinson jianzhong (andrew) zhang lucy delgadillo mary kabaci chris robinson terry zhang catherine d’hondt charlene kalenkoski john salter xin (jessica) zhao donna dudney elizabeth kiss marta serra-garcia timothy zimmer yaman erzurumlu vladimir kotomin daniel fernandes marc kramer fred fernatt vi s. michelson / financial services review 29 (2021) v–vii please consider submission to the financial services review and rely on the style information provided to ease readability and streamline the review process. the journal welcomes articles over the range of areas that comprise personal financial planning. while fsr articles are certainly diverse in terms of topic, data, and method, they are focused in terms of motivation. fsr exists to produce research that addresses issues that matter to individuals. i remain committed to the goal of making financial services review the best academic journal in individual financial management and personal financial planning. best regards, stuart michelson editor financial services review s. michelson / financial services review 29 (2021) v–vii vii pii: 1057-0810(95)90014-4 financial servlces review, 4( 1): 9-22 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. tax savings opportunities in estate freeze transactions: an application of the black scholes model james r. hamill joel s. stemberg transfer tax valuation rules for interests in family businesses include a minimum value to reflect the option value of “junior” equity interests transferred tofamily members. we examine the option features of the junior interest and use the black&holes model to identify situations in which an estate planner could structure a plan of ownership succession that results in undervaluation of the transferred interest. the black-scholes model may also be used to identify situations in which a lifetime ownership transfer should be avoided because of the minimum value rule. i. introduction it is well documented that organizations will structure their affairs to minimize costs, both explicit and implicit, and will also restructure in response to changing costs (wilson, 1991). prior to the 1987 tax act, it was common for closely held business organizations to engage in costly restructuring of ownership (estate valuation “freezes”) to minimize future estate taxes resulting from the death of the senior generation owner. in response to perceived abuses, congress, in 1987, enacted statutory rules that generally eliminated the tax advan tages of such restructurings. however, the 1987 changes raised concerns that family-con trolled businesses could be forced to admit outside owners to satisfy an estate tax liability and that uncertainty regarding the application of the rules could also severely restrict commonly used ownership forms that were not adopted to save taxes. in response to growing criticism, in 1990 congress retroactively repealed the 1987 changes and replaced them with a new set of rules. the 1990 provisions attempt to recognize the call option features of junior equity interests created by estate freeze restructurings. the option value is explicitly recognized by use of a single minimum value that applies to all transferred junior interests. the purpose of this article is to show situations where the minimum value rule is most likely to misstate the james r. hamill l anderson schools of management, university of new mexico, albuquerque, new mexico 87131. joel s. sternberg l department of finance, university of arizona, tucson, arizona 85721. 10 financial services review 4(l) 1995 option value of the transferred interest. we use the option pricing model developed by black and scholes (1973) to identify fact patterns in which a business recapitalization may still minimize transfer taxes because of improper valuation of the option. limitations on the use of the black-scholes model for this purpose will also be addressed. ii. description of estate freeze transactions a significant portion of the net worth of the owner of a closely held business is the value of the business itself,’ and taxes imposed on the transfer of ownership to younger generation family members can force the sale of business assets or equity interests to satisfy the tax liability. to mini~ze the effects of transfer taxes, it has been common to incur costs to recapitalize the business ownership. typically, a family corporation will issue two classes of equity, with voting preferred retained by the older generation and nonvoting common transferred to the younger generation.* the terms of the retained interest are intended to “freeze” its value at an amount approximating the value of the entity at the date of recapitalization. this would typically involve a stated r~emption value for the preferred issue that would fix the amount that the estate would receive for the interest. rights attached to the retained interest would support an appraisal assigning a large value to the retained interest and a corresponding small value to the transferred interest, although there often was no intent that the rights would ever be exercised3 the estate freeze was designed to serve two purposes. first, there would be little or no gift tax attributable to the transfer to the younger generation, because the interest was assumed to have little value! second, the estate tax value of the retained interest would be deterministic, because all future increases in the value of the corporation would be captured by the transferred interest. in 1987, congress adopted valuation rules applicable to freeze transactions that received critical commentary from business owners and tax experts. the joint committee on taxation (jct) prepared a discussion draft describing problems with the 1987 rules and suggesting that rights retained by the senior generation be valued at zero unless the rights must be exercised within a specified time and at a specified amount5 the jct noted that the interest transferred to the younger generation contained elements of a call option and that the call feature is undervalued by taxpayers. to reflect the option value, the jct recommended that a minimum value of 20% of the total value of the interests held by the senior generation family member be assigned to the junior interest, the 1990 tax act adopted a 10% minimum value rule to explicitly reflect the option feature of the transferred interest.6 the remainder of this article evaluates the basis for treating the transferred interest as possessing the characteristics of acall option and analyzes how the option feature could be valued. the statutory minimum value rule is compared to use of the black-scholes option pricing model (opm), and implications for the structure of future freeze transactions are discussed. iii. call option features of junior equity interests at the death of the senior generation family member following an estate freeze recapitaliza tion, the retained preferred stock would be redeemed by the corporation,’ “bootstrapping” the junior equity interest holders into complete ownership of the firm. if the value of the firm tax savings opportunities in estate freeze transactions 11 declines below the stated redemption price, the junior equity owners cannot be forced to contribute to capital the excess of the redemption price over the value of firm assets to fund the redemption (as equity owners, they can simply walk away from a corporation with assets valued at less than the redemption claim of the preferred stock interest). thus, the junior equity owners have the right, but not the obligation, to acquire full ownership by payment of a fixed sum at a deter~nable time. in this way, their ownership interest resembles that of a call option.’ the use of a minimum value may not be appropriate to value the call feature of the junior interest. while attractive for the policy reasons of objectivity and simplicity, the minimum value rule may lead to an undervaluation of the option feature of the junior interest. identifying the situations where the transferred interest may be undervalued can be of assistance to estate planners structuring freeze transactions. we believe that the opm may be used to develop a representative value of the call feature of the junior interest. use of opm to value the call option feature of junior equity interests relies on the ability to reasonably estimate the model parameters in relation to an estate freeze recapitalization. the following form of the opm was used in this article: where c = sn(dt) keen n(&) d, = ln(s/ke-“) / 04+ l/20* d2=d,-cd (1) c = value of call option s = company value at the date of the r~apit~ization n = normal density function k = strike or exercise price cr = annual logarithmic standard deviation of company returns t = time to expiration of option right r = logarithmic risk-free interest rate a call option on an asset represents a long position in that asset, together with riskless borrowing because the option holder may earn interest on the strike price until he exercises the option. n(q) is the probability that the option will finish in-the-money and therefore be exercised, and this probability increases as the underlying asset value increases. the final term in equation (1) is the product of the strike price and the probability of the option being exercised, discounted, and represents the effective amount being borrowed at any point in time. n(q) is the fractional shares of stock represented by the option at any time and is also known as the hedge ratio or delta of the option. as the stock price rises, changes in the option value more closely match changes in the value of the stock and n(d,) increases.’ the first term in equation (1) is the dollar amount of stock represented by the option at any point in time. the subsequent sections discuss how the parameters of the opm could be estimated in a freeze transaction. we then provide an example of the use of opm to identify situations where a freeze transaction leads to the most favorable or unfavorable tax law valuation. future estate freezes may be structured to take advantage of situations where the ~nimum value rule undervalues the call feature of the transferred interest. 12 financial services review 4( 1) 1995 a. value of the asset subject to the option right in an estate freeze transaction, the junior equity owners can acquire ownership of the entity by redeeming the senior equity interest. although valuation of a closely held business is difficult, firm value must otherwise be determined at the time of the recapitalization transaction, and opm would add no complexity.” b. exercise price of the option right the exercise price of the option right is the redemption price of the retained interest of the senior generation family member. typically, this amount will approximate the value of the firm at the date of recapitalization. in any event, the amount is generally readily determinable by reference to the stated terms of the redemption agreement for the retained interest. c. risk-free rate of return the value of the call option is, in part, a function of the interest rate which enters the second term, ke-“, of the opm. this reflects the fact that the holder of a european call option will not exercise the option, nor pay the strike price, until expiration of the option right. therefore, the higher (lower) is r, the lower (higher) will be the present value of the strike price and the higher (lower) will be the value of the call option. the yield on a u.s. treasury security is typically used as a surrogate for the risk-free rate of return. a treasury yield for an instrument with a term comparable to that of the redemption agreement could be used as the risk-free rate. d. variance of asset returns and time to maturity of the option the value of an option is an increasing function of both annual return variance and time. the greater is the annual return variance or the time to maturity of the option, the greater will be the variance of the asset return distribution, o&, over the life of the option. since a call option is a contingent claim, losses can never exceed the amount paid to acquire the option, regardless of the actual return distribution from the asset. positive returns, however, occur whenever the value of the asset exceeds the sum of the exercise price and the amount paid to acquire the option. thus, the option holder benefits from the entire right tail of the distribution of returns, while his losses from returns in the left tail of the distribution are truncated because they are limited to the amount paid to acquire the option. as time to maturity lengthens, the option holder can also collect interest on money that would otherwise have been used to exercise the option right. the value of a closely held firm, at any point in time, is not supported by equity traded on an active exchange. firms may rely on use of an appraisal to determine value. however, because appraisals can be very costly, many closely held businesses planning for succession within the family define value by reference to financial statements (that is, book value). book value and market value are not the same, but the purpose of our analysis is to identify situations in which the transfer tax value of the firm is too low. book value is accepted for purposes of the transfer tax and is often used by family businesses. if book value is used in a buy-sell agreement, then the variance of the value may be readily determined. if value is to be set by appraisal, the firm must have an appraisal conducted at the time of the initial tax savings opportunities in estate freeze transactions 13 recapitalization to support the values assigned to the retained and transferred interest. if the appraiser uses capitalization of earnings, then earnings variance may be an appropriate proxy for annual return variance. the same may be said for use of cash flow valuation or any other approach favored by a competent appraiser. the insights and analysis used by the appraiser at the time of the recapitalization may be used to develop a return variance proxy. the results of use of the opm may also be readily analyzed for sensitivity to different specifications of the return variance, as shown in our example. if the senior equity interest must be redeemed by a certain date, clearly that represents the maturity period. generally, the redemption of the senior interest is expected to occur upon the death of the holder (since the recapitalization is intended to freeze the value of the senior interest for estate tax purposes). in the analysis presented in this article, we use the life expectancy of the senior generation family member as the maturity period. since the life expectancy (which could also be for joint lives) is the longest period to maturity, the value of the call feature and the taxable transfer at the time the junior equity interest is created will be maximized for a given dividend payout. the maturity assumption could readily be altered for a more specific fact pattern. iv. a comparisonof opm andthestatljtoryminimljmvalue in this section, we compare the value assigned to a junior equity interest under the new tax valuation rules with that obtained using the black-scholes model. the black-scholes model will value only the call feature of the junior interest, while the tax law requires valuation of the entire interest. however, in situations where the 10% minimum value rule of the tax law applies, the junior interest is valued strictly as a call option and opm would provide a relevant comparison.’ ’ a comparison of opm and current tax law valuation is of interest in two situations. first, if the tax law value of the junior interest is determined to be less than 10% of the value of all interests, the minimum value rule is binding and the interest is valued strictly as a call option. opm can then be used to determine whether the option value is appropriate (either too high or too low) given the terms of the instrument. the second situation occurs if the tax law value is not less than lo%, the minimum value rule is nonbinding, and the interest is valued using new tax law valuation principles without regard to the option feature. opm can then be used to determine whether the option value exceeds the value of the interest as determined under tax law principles. since the option value is only part of the total value of the junior interest, an opm value in excess of tax law value would suggest that the true value may be substantially understated by the tax law. as discussed later, this analysis is limited by the appropriateness of the use of opm to value the junior interest. the tax law values the junior interest by subtracting the value of the retained (senior) interest from the total value of the firm.12 the senior interest is valued by discounting required cash flows at a rate equal to 120% of the “applicable federal rate” (afr) as determined under section 1274 of the internal revenue code. the afr is a treasury security yield for an instrument of similar maturity. the only cash flows that are discounted are those that, under the terms of the instrument, must be paid with certainty (see irc section 2701 (a)). we will consider three different situations: 14 financial services review 4(l) 1995 a preferred stock issue that provides for a cumulative dividend yield equal to 120% of the air, and a redemption value equal to the date-of-recapitalization value of the firm, with redemption required at the death of the senior interest holder. a preferred stock issue that provides for a cumulative dividend yield equal to the afr, and a redemption value equal to the date-of-recapitalization value of the firm, with redemption required at the death of the senior interest holder. a preferred stock issue that provides for no required payment of a dividend, and a redemption value equal to the date-of-recapitalization value of the firm, with redemption required at the death of the senior interest holder. table 1 value of call option feature of junior equity interest using opm, expressed as a percentage of firm value (dividend payout = 120% of afr) value determined using opm with log-volatility of: entrepreneur tax law value 20% 30% 40% life expectancy i%) (%) c%) c%) 1 10.0 6.6 10.2 13.8 2 10.0 8.2 12.9 17.4 3 10.0 8.9 14.1 19.1 4 10.0 9.1 14.5 19.8 5 10.0 9.1 14.6 19.9 6 10.0 8.9 14.4 19.6 7 10.0 8.6 14.0 19.1 8 10.0 8.3 13.5 18.4 9 10.0 7.9 12.9 17.6 10 10.0 7.5 12.2 16.7 11 10.0 7.0 11.6 15.8 12 10.0 6.6 10.9 14.9 13 10.0 6.2 10.2 14.0 14 10.0 5.8 9.6 13.1 15 10.0 5.4 9.0 12.2 16 10.0 5.0 8.4 11.4 17 10.0 4.7 7.8 10.6 18 10.0 4.3 7.3 9.9 19 10.0 4.0 6.7 9.2 20 10.0 3.7 6.3 8.5 21 10.0 3.4 5.8 7.9 22 10.0 3.2 5.4 7.3 23 10.0 2.9 5.0 6.7 24 10.0 2.7 4.6 6.2 25 10.0 2.5 4.2 5.7 26 10.0 2.3 3.9 5.3 27 10.0 2.1 3.6 4.9 28 10.0 1.9 3.3 4.5 29 10.0 1.8 3.1 4.1 30 10.0 1.6 2.8 3.8 31 10.0 1.5 2.6 3.5 32 10.0 1.4 2.4 3.2 tax savings opportunities in estate freeze transactions 1.5 in generalized form, and assuming a constant dollar payout, the tax law would compute the value of the junior interest as: firmval [annualdiv ( 1 (1 +r)-” / r) + firmval / (1 +r)‘] (2) where: fzrmval = date-of-recapitalization value of the firm annualdiv = annual dividend required to be paid r = discount rate to be applied to cash flows n = term of analysis, equal to the entrepreneur’s life expectancy at the date of the recapitalization. in no event can the junior interest be valued at less than .10 firmval. the bracketed terms in equation (2) represent the present value of the dividend stream received by the senior generation and the present value of the redemption payment. equation (2) holds for the fact pattern assumed in this analysis. for more complex recapitalization terms, the value of the junior interest would be modified, although the value would not be difficult to model. the analysis that follows assumes that 120% of the afr is 10%. thus, r = .lo and the afr = .0833. the dividend yield will be 10% of firm value, 8.33% of firm value, and zero, in the respective situations under review. the analysis will examine life-expectancy from l-32 years with log volatility in the blacwscholes model set at 20%, 30%, and 40%. a. dividend yield equal to 120% of the afr when the dividend yield equals the discount rate used for tax law valuation, and the redemption value of the senior interest equals the initial value of the firm, the value of the junior interest is zero using the subtraction method. (the senior interest is equivalent to a 4 os01 ’ ” ” ” ’ ” ” ” ’ ” ” ” ’ ” ” ” ’ ” ’ 1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132 life expectancy of entrepreneur _._ 20% volatility + 30% volatility _,_ 40% volatility figure 1. value of option feature of junior equity interest (10% dividend, 8.33% afr) 16 financial services review 4( 1) 1995 table 2 value of call option feature of junior equity interest using opm, expressed as a percentage of firm value (dividend payout = 100% of afr) value determined using opm with log-volatility of: entrepreneur life tar hw value 20% 30% 40% expectancy i%i (%i (%/o) f%) 1 10.0 7.4 11.1 14.7 2 10.0 9.8 14.5 19.2 3 10.0 11.2 16.5 21.7 4 10.0 12.1 17.7 23.2 5 10.0 12.7 18.5 24.1 6 10.0 13.1 18.9 24.5 7 10.0 13.4 19.2 24.7 3 10.0 13.5 19.2 24.6 9 10.0 13.6 19.2 24.4 10 10.2 13.7 19.1 24.1 11 10.9 13.7 18.9 23.8 12 11.3 13.7 18.7 23.4 13 11.8 13.7 18.5 23.0 14 12.3 13.6 18.3 22.5 1.5 12.7 13.6 18.1 22.1 16 13.0 13.6 17.9 21.7 17 13.4 13.6 17.6 21.3 18 13.7 13.6 17.4 20.9 19 13.9 13.6 17.2 20.5 20 14.2 13.6 17.1 20.2 21 14.4 13.6 16.9 19.8 22 14.6 13.6 16.8 19.5 23 14.8 13.7 16.6 19.2 24 15.0 13.7 16.5 19.0 25 15.2 13.8 16.4 18.7 26 15.3 13.8 16.3 18.5 27 15.5 13.9 16.2 18.3 28 15.6 13.9 16.2 18.1 29 15.7 14.0 16.1 17.9 30 15.8 14.1 16.1 17.8 31 15.8 14.2 16.0 17.7 32 15.9 14.3 16.0 17.5 bond with a coupon equal to the required market return and a r~emption value equal to par.) however, the tax law requires that the value of the junior interest be equal to 10% of firm value, to reflect the call feature of the interest. as shown in table 1, the option value of the junior interest determined using opm varies from 1.4% to 19.9% of fitm value, depending on volatility and entrepreneur life expectancy. when log volatility is low (20%),13 the opm v~uation is below 10% for all time periods. the tax law then values the call feature of the junior interest at an amount in excess of the opm value, with the difference generally increasing with the life expectancy of the entrepreneur (see figure 1). the valuation difference ranges from less than 10% to in excess of 600% of the value determined using opm. when log volatility is 30%, the opm call value is in excess of 10% for life expectancies less than app~ximately 13.3 years, and less than 10% otherwise. depending on the term, tax savings ~ppor~ni~es in estate freeze ~rans~ct~~s 17 1 2 3 4 5 6 7 8 91011121311151617181920212223252526272829303132 life expectancy of entrepreneur 20% volatility + 30% volatility + 40% volatility figure 2. value of option feature of junior equity interest (8.33% dividend, 8.33% afr) the minimum value rule can lead to valuations less than opm by more than 30% (for shorter life expectancies) and valuations higher than opm by more than 300% (for longer life expectancies). log volatility of 40% produces similar results, with tax law undervaluations relative to opm magnified and overvaluations reduced. the crossover point, where the minimum value rule and opm produce similar values, occurs at 13.3 years with 30% volatility and 17.9 years with 40% volatility. figure 1 shows the call value as a function of entrepreneur life expectancy for each level of volatility. b. dividend yield equal to the afx when the dividend yield equals the afr, the value of the junior interest as determined by tax law is always positive, because the cash flows generated by the senior interest are less than the required market rate-of-return-that is, the senior interest is valued at a discount from firm value and the subtraction method results in a positive value assigned to the junior interest. for an entrepreneur life expectancy of less than 10 years, the value of the junior interest as determined by tax law is less than 10% of the firm value. the minimum value rule, therefore, is binding only for life expectancies less than 10 years and is nonbinding otherwise. in contrast, as shown in table 2, the opm value of the call feature of the junior interest exceeds 10% for all situations except 20% volatility with life expectancy of two years or less. for life expectancies of less than 10 years (when the ~nimum value rule is bindings, the tax law value is less than the opm call value of the junior interest, except as noted in the preceding sentence, and the difference increases with increasing volatility. figure 2 shows the call value as a function of entrepreneur life expectancy for each level of volatility. 18 financial services review 4( 1) 1995 c. dividend yield equal to zero if the dividend yield on the senior interest is equal to zero, the value of the senior interest is equal to the present value of the stated redemption amount, paid at the death of the entrepreneur. the subtraction method assigns a relatively large value to the junior interest, and the minimum 10% value is binding only for a life expectancy of one year.14 thus, for tax purposes, the value of the junior interest is determined without regard to the minimum value rule. as shown in table 3, the call value of the junior interest for the zero dividend case, as determined using opm, exceeds 10% for all life expectancies and all volatilities. further, the tax law valuation rules approximate the call value of the junior interest as determined using opm, particularly for longer life expectancies. figure 3 shows the call value as a function of entrepreneur life expectancy for each level of volatility. table 3 value of call option feature of junior equity interest using opm, expressed as a percentage of firm value (dividend payout = zero) value determined using opm with log-volatility of: entrepreneur life expectancy 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 tar law value 20% (o/4) (%) 10.0 12.1 17.4 19.4 24.9 25.8 31.7 31.4 37.9 36.6 43.6 41.4 48.7 45.8 53.3 49.9 57.6 53.6 61.4 57.1 65.0 60.3 68.1 63.3 71.0 66.0 73.7 68.6 76.1 71.0 78.2 73.1 80.2 75.2 82.0 77.0 83.6 78.8 85.1 80.4 86.5 81.9 87.7 83.2 88.8 84.5 89.8 85.7 90.8 86.8 91.6 87.8 92.4 88.7 93.1 89.6 93.7 90.4 94.3 91.1 94.8 91.8 95.3 92.4 30% (%j 15.7 24.0 30.9 36.8 42.0 46.7 51.0 54.9 58.4 61.7 64.7 67.4 69.9 72.2 74.4 76.3 78.1 79.8 81.4 82.8 84.1 85.3 86.4 87.4 88.4 89.3 90.1 90.8 91.5 92.2 92.8 93.3 f%) 19.4 28.8 36.3 42.5 47.9 52.7 57.0 60.7 64.2 67.2 70.0 72.6 74.9 77.0 78.9 80.6 82.2 83.7 85.0 86.2 87.3 88.4 89.3 90.2 91.0 91.7 92.4 93.0 93.5 94.0 94.5 95.0 tax savings op~or~nities in estate freeze transa~~~s 19 tj 3 90% s 80% 2 8 70% ?? 60% 5 8 50% iis 40% % 30% g 20% 0” =: 10% s 0% t #if,,, t ,,i 61 it i i i i ii #iii iiii f 3 1 1 2 3 4 5 6 7 8 9 10111213141516171819202l2223242526272829303132 life expectancy of entrepreneur _ 20% volatility _+_ 30% volatility + 40% volatility figure 3. value of option feature of junior equity interest (0% dividend, 8.33% afr) d. analysis of results in figures 1 and 2, there are two principal factors affecting the value of the option over time. on the one hand, all of the option values are enhanced by the increase in variance that occurs as the time to expiration increases, ceteris paribus. on the other hand, as time increases, more dividends are paid on the senior equity interest, reducing the value of the underlying asset and thereby the option right. for the 10% dividend case (figure l), at first the variance has the stronger effect, leading to increasing option values, as time passes, the negative effect of the dividend payout increases in importance, leading to a steady decrease in the value of the option towards zero. in the 8.33% dividend case (figure 2), a point is reached where the dividend and variance effects essentially cancel, and the option value levels at approximately 15% of firm value for all volatilities. in the no dividend case (figure 3), the dividend effect is nonexistent and the value of the option increases with time for all volatilities. since an option can never be worth more than its underlying asset without creating arbitrage opportunities, the option value is bounded at 100% of company value and the lines in figure 3 converge with time. the 1990 tax act explicitly requires that the call value of the interest be recognized through a statutory minimum value rule. the minimum value rule generally will apply only when the retained interest provides for fixed, cumulative dividend payments, and then only for certain payment terms. payment terms that result in a binding minimum value rule should be avoided if the opm suggests the junior interest option feature is worth less than 10% of fkm value. for example, in the analysis conducted in this article, with tog volatility of 20%, it is not advisable to pay a dividend equal to the discount rate used for tax law valuation. in such a case, the statutory minimum value rule is binding and opm suggests that the call value will be overstated, with the overstatement generally increasing with the life expectancy 20 financial services review 4( 1) 1995 e 70% t! c60% s sj 50% 8 40% 1 2 3 4 5 6 7 8 9 101112~314151617181920212223242526272829303132 life expectancy of entrepreneur _,_ 0% dividend + 8.33% dividend + 10% dividend figure 4. implied volatility for 10% option value of the entr~reneur. if a smaller dividend is paid and the entrepreneur’s life expectancy is relatively short, the minimum value rule is binding but the call feature of the junior interest is undervalued when opm is used as a benchmark. as volatility increases, the opm value of the call feature can exceed the minimum value if the dividend yield equals the assumed discount rate and the life expectancy of the entrepreneur is short, but this is reversed for longer life expectancies. the informed estate planner has an incentive to suggest terms of the senior interest to take advantage of any valuation distortions created by tax law. figure 4 plots volatiiity of the firm on the y-axis and entrepreneur life expectancy on the x-axis with all points reflecting the volatility required to value the junior interest at 10% using opm. when no dividend is paid on the senior equity interest, the underlying asset value is assumed to steadily increase and the option right becomes more valuable with time. in fact, with maturities of four years or longer there is no volatility low enough to result in a 10% (opm) option value. as the dividend payout grows, the decrease in the value of the underlying asset reduces the value of the option. at a 10% payout, figure 4 illustrates that the option value is less than 10% regardless of the volatility level when the time to maturity exceeds 2 1 years. thus, dividend payments to the senior generation that approximate the valuation discount rate (120% of the afr) shoufd be avoided if the objective is to minimize transfer taxes. this is because the minimum value rule will be binding, although the call value of the junior interest is likely to be less than the 10% value assigned by law. figure 4 is sensitive to the dividend payout assumptions used in the first part of this analysis. however, in conjunction with figures 1-3, it shows that the option value of the junior equity interest is sensitive to the interaction of time, volatility, and dividend payout. tax savings opportunities in estate freeze transactions 21 e. limitations of this analysis there are two important limitations of the analysis shown in this article. first, a valuation expert hired in connection with the estate freeze transaction must be able to reasonably estimate opm parameters. as discussed above, volatility of firm value would be the most difficult parameter to estimate. however, as shown in figure 4, it is possible to determine the volatility implied by a specified option value. if the implied volatility appears unreason able, then the tax law option value is also likely to be unreasonable. also, upper and lower boundaries could be specified, and the resulting option values could be examined for sensitivity to volatility misspecifications. the second limitation is whether it is appropriate to use the black-scholes model to value the option feature of the junior interest. the junior interest clearly has features of a call option, but the features do not literally meet all of the assumptions of opm. for example, if life expectancy is viewed as the maturity date, the junior interest “option” is not strictly european because the senior generation owner can certainly die before “maturity” and cause an early exercise. however, the analysis is intended to be used for planning purposes only and can incorporate the effects of an exercise at some term other than life expectancy. a sensitivity analysis, similar to that shown in figures l-3, can be useful in determining the effects of misspecification of a model parameter. v. conclusion this article compares the use of the statutory minimum value rule, intended to capture at least the option value of a junior equity interest, with the opm value of the option feature. the results suggest that estate freeze transactions can still lead to transfer tax savings when the tax law understates the call value of the transferred interest. conversely, estate freeze terms that result in a binding minimum value rule should be avoided if the call value is less than the statutory 10%. the results are limited by the ability of the valuation expert to estimate opm parameters and the applicability of the black-scholes model to the junior interest created in the freeze. notes 1. prior research has recognized that an entrepreneur may not be diversified to the extent suggested by portfolio theory (leland & pyle, 1977). 2. it is assumed in this analysis that the older generation owns 100% of the corporation prior to the recapitalization. the costs of restructuring this ownership can be both monetary and nonmone tary. 3. these rights could include rights to dividend payments, to liquidate the entity and receive assets, to put the interest to the entity, and to convert the frozen interest into an interest with appreciation potential, among others. because a third party would pay for these rights, an appraisal could include the value of the rights. 4. the value of the transferred interest has been determined by subtracting the value of the retained interest from the total value of the enterprise. by assigning a redemption value to the preferred stock equal to the date-of-recapitalization value of the entity and attaching valuable rights to the preferred, it was argued that the preferred value approximated the total value of the entity. the subtraction method would then assign very little value to the transferred interest. enforcement 22 financial services review 4(l) 1995 difficulties arose because of the inherently factual nature of a valuation question and the lack of any statutory guidance. 5. the joint committee on taxation is a standing committee of congress that advises both houses on fiscal and administrative issues associated with tax legislation. it consists of members of the tax writing committees of congress and professional staff with backgrounds in law, economics, and accounting. 6. irc section 2701 (a)(4). the minimum value rule was initially set at 20% under h.r. 5425 (august 1,1990), the first proposed amendment to the estate freeze rules following the joint committee discussion draft. the senate eliminated the minimum value standard in s. 3 1 i3 (september 26, 1990), but it later resurfaced at 10% in s. 3209 (october 13, 1990). the final bill was approved by both houses of congress on october 27, 1990, and retained the standard at 10%. 7. or earlier, if the agreement so provides. 8. technically, the junior equity owners do have an enforceable obligation to redeem the senior interest. the obligation arises from the capacity as equity owners of the corporation and is limited to the value of firm assets, resulting in an obligation to exercise the option when it finishes in-the-money. since this is rational behavior, the obligation does not change our conclusion. 9. at one extreme, if there is absolute certainty that the option will be exercised, the value of the option will move dollar-for-dollar with the underlying stock. n(d,) will take on a value of 1 in this situation because the option holder will inevitably acquire the stock by exercise of the option. at the opposite extreme, if there is no chance of the option being exercised, the option value is not affected by changes in the stock price and n(d,) equals zero. io. the “subtraction” method requires that the value of the retained interest be subtracted from the total value to determine the value of the transferred interest. 1 i. recall that the minimum value rule exists to recognize the call value only. 12. for ease of exposition, we are again assuming that the senior generation initially owns 100% of the firm. the analysis could readily be extended to other situations. 13. one standard deviation unit of 20% log volatility means that the value of the fii in one year has the same probability of increasing by 20% (a factor of 1.2 in log form) as decreasing by 20%. that is, pr. [vulue(r,) = vulue(t,) x i.21 = pr.[vu/ue(r,) = va/ue(ro) / 1.21. 14. of course, this is a function of the discount rate. in general, the minimum value rule would be binding only when firmval [firmval / (l+r)“] < .10 firmval. as r decreases (increases), the minimum value is binding for longer (shorter) life expectancies. references black, f., & scholes, m. (1973). the pricing of options and corporate liabilities. journnl of polirical economy, bi(may-june), 637-654. joint committee on taxation. (1990, april 20). federal transfer fax consequences of estate freezes. (jcs-134’0). leland, h., & pyle, d. (1977). information asymmetries, financial structure, and financial intermedia tion. journal of finance, 32(may), 371-385. merton, r.c. (1973). theory of rational option pricing. bell journal of economics and management science, i(spring), 141-183. irs may upstage hill as source of big tax news in ‘91, fba told. (1991). tax notes, (march ll), 1044-1045. wilson, p. (1991). future research directions in taxation. the journal of rhe american taxation associafion, 13(fall), 64-73. pii: 1057-0810(92)90014-4 financial services review, 2(l): 41-49 copyri&t 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. long-run returns on stock and bond portfolios: implications for retirement planning kirt c. butler dale l. domian this paper presents asset returns over long aping periods in a form useful for retirement planning. time diversi~cat~n, heretofore analyzed for lump-sum investments, still serves to reduce the risk of stock investments whenfunds are accumulatedmonth by month, we consider investments in five stock and bond asset classes as well as various asset allocation strategies. probability distributions are computed for retirement wealth over a range of investment horizons. 1. i~roruc~on time diversification plays a key role in retirement planning decisions. because of time diversification, the risk of meeting a particular wealth objective is reduced when risky assets with high expected returns are held over long periods of time. individuals face investment accumulation periods of up to forty years early in their careers, so information about the long-run returns and risks of various asset classes is essential. existing research on time diversification must be modified for retirement planning. levy (1978), reichenstein (1986), and butler and domian (1991) assume “lump-sum” investments at the beginning of the holding period, with no other funds added except for reinvested portfolio earnings. however, retirement investments typically have monthly or yearly contributions which increase portfolio size t~oughout the holding period. thus, returns in early years are not as important to retirement wealth as returns in later years. kirt c. butler l associate professor of finance, department of finance and insurance, the eli broad graduate school of management, michigan state university, east lansing, mi 48824. dale l. domian l assistant professor of finance, department of finance and insurance, ‘the eli broad graduate school of management, michigan state univet&y, east lansing, mi 48824. 42 financial services review, 2(l) 1wm993 most defined contribution pension plans give the asset allocation choice to the individual participant. for example, many college and university employees allocate retirement contributions between the cref common stock fund and tiaa which has assets including bonds and mortgages. additional choices are available through iras, 401(k) plans, and other tax-sheltered arrangements. investors making peri odic contributions to their retirement portfolios need to know the impact of their asset choice on ending wealth. we use observed capital market history and an empirical resampling procedure to estimate retirement wealth distributions over various investment horizons. these distributions are inflation-adjusted so that retirement income can be stated in terms of today’s purchasing power. the wealth distributions also reveal the likelihood of one asset class outperforming another over various holding periods. along with stating total wealth at the beginning of retirement, we also calculate monthly payments from lifetime annuities so that investors can plan for their retirement incomes and lifestyles. the next section reviews the existing time diversification literature. section iii describes the resampling procedure for estimating wealth distributions. section iv presents and interprets the results from the estimations. concluding remarks are made in section v. ii. previous research levy (1978) demonstrates effects of time diversification by tabulating histori cal returns over various holding periods. he finds that common stocks outperformed treasury bills in every 25year holding period over 1926-1977. by using one-year overlaps, levy constructs twenty-eight 25year periods from the 52 years. however, as levy observes, overlapping holding periods are not independent.’ if common stocks had risen by loo0 percent in january 1952, this amazing return would be included in all but three of levy’s 25-year periods. distributions based on overlap ping periods thus misrepresent the true 25-year return distributions. a way out of this problem is to use the observed history of stock and bond returns in a mote creative way. one approach is to assume stock and bond returns are normally distributed and then derive the time diversification impact of longer holding periods. reichenstein (1986) uses this method to show the manner in which portfolio risk depends on the holding period. a simulation model can also prove helpful in time diversification research. this is an especially attractive alternative for researchers convinced of fundamental differences between past and future real return distributions. for example, leibowitz and langetieg (1989) construct models assuming a 4 percent stock risk premium over bonds rather than the 7 percent common stock/treasury bond premium ob served since 1926. a lower stock/bond risk premium decreases the relative advan tage of stocks over bonds for both short and long holding periods. the disadvantage long-run returns on stock and bond portfdios 43 of this technique is that simulation models are often ad hoc and reflect the creator’s preconceptions regarding the relevant forces driving asset returns. we follow the empirical resampling procedure of butler and domian (199 1). in contrast to the simulation model of leibowitz and langetieg, this approach uses only the observed history of stock and bond returns. it requires two conditions for the empirical distribution: 1) random draws must he independent over time, and 2) the moments of the distribution must not change over tin~.~*~ a full description of this procedure is presented in the next section. iii.theresamplingprocedure the standard data source in the time diversification literature is the ibbotson monthly return indices which date back to 1926 (see ibbotson 1991). we use ibbotson’s real (inflationadjusted) returns for five indices: common stocks, small stocks, corporate bonds, treasury bonds, and treasury bills. our study covers 65 years from 1926 through 1990, a time span of 780 months. table 1 presents summary statistics of the time series. an empirical resampling procedure is used to construct the distribution of retirement plan returns. we first explain the technique with an illustrative example of a 30-year accumulation period, and then go on to show how this can be modified for other investment horizons. suppose an individual intends to make monthly retirement contributions for 30 years. at the end of this accumulation period, the total accumulated savings amount is used to purchase a single-premium lifetime annuity with monthly payments. for each dollar of monthly investment during the accumulation period, we want to identify the accumulated wealth and the size of the monthly annuity payment. based on observed capital market history, probability distributions are deter mined as follows: table1 the historkal return series common small corpomte treasury treasury statistic stocks stocks bollds bonds bills mtxll .0072 .0105 .0019 do14 .0004 st. dev. .0589 xl902 .0211 .023 1 .0057 autocorrelation .loll .1557 .i992 .0891 .5368 cross correlations small stocks x494 carp bonds .2383 .1895 t-bonds .1932 .1143 .8481 t-bills .0804 .o262 .3289 .3323 note: real monthly time series for 19264990 from ibbotson associates. 44 financlql services review, 2(l) 1992/1993 1. randomly select one of the 780 months. record the observed real return for this month for each asset class. 2. do this random selection 360 times with replacement to construct 30 years of monthly returns. 3. compound and accumulate the returns assuming additional real monthly investments of the same dollar amount. 4. do this procedure 10,000 times to generate wealth distributions from the observed history of real monthly returns. 5. divide the ending totals by an annuity factor to get the monthly payments th~ughout retirement. this procedure can easily be modified. first, the accumulation period in step 2 can be shortened or lengthened. the next section shows results for 10,20,30, and 40-year accumulation periods. second, various asset allocation strategies can be considered, using different mixes of the asset classes. examples are given in the next section. third, step 3 can be modified to allow changes in the monthly ~ves~ent amount, perhaps as real earnings increase in later years of life. finally, the annuity factor in step 5 can be changed to reflect different rates of return. iv. results distributions of ending wealth (i.e., at the end of the accumulation period) are shown in table 2. selected percentiles are given for 10, 20, 30, and 40-year accumulation periods. for each of five asset classes, the values are ending wealth per dollar of real monthly contribution. consider the illustrative 30-year accumulation period discussed in the previous section. at the median (fiftieth percentile), a one-dollar monthly investment in common stocks totals $1,194 after 30 years. small stocks produce more, $1,680, while corporate bonds, treasury bonds, and treasury bills provide substantially less. because these are medians, there are equal chances of getting either more or less than these amounts, based on observed history. the distributions for common stocks and small stocks are skewed to the right so that their means are substantially higher than their medians.4 this probabilistic interpretation can be carried a step further to make relative comparisons between stocks and bonds. for the 30-year accumulation period, the common stock distribution crosses the corporate bond distribution between the 3rd and 4th percentiles. this implies a less than four percent chance of ending up with lower wealth from diversified common stock investments than from corporate bonds over a 30-year accumulation period. the chance is even smaller for common stocks versus treasury bonds, since these distributions cross at the second percentile. with the 20-year accumulation period, there is still just a small risk that common stocks will underperform bonds. it is only when the accumulation period is reduced to 10 years that the risk is more substantial. iiere the common stock ;pofad %u~p~oqaqj~noq~nojqlspuoqlua~j,lad~~ pwy3o~swa3jad~1 uaarkllaqsaq3)!ms lciwopuejq~!q~o!lo~~odt!o~ jo!jadns s! ajn~x!mploq-pw-dnqos_os")eq)smoqs (0661) uoslanunz?g ‘ajot.ujaqwy *suoyodojd jaylo aql wojj ueyl xy.u puoq/y~o~s os_os tz way qqoam jawj% s! ajaq3 (c alqvl u! alyuawd q,jy ayl aas) salyua3 -jad qwal pue puo=s aql uaamlaa 'poyad uo!yqnwnme jealc-oc e japyuos os~e an22 alqtzl wojj suuuqo:,puoqpwy3o~s wa3jadpajpunqaug syz'o~o uoytq -ajjo3-ssoj3~~q~uou.i~ aheq qxqm‘spuoq wz?jodjo3pw sy3oo)s uouiuio3jo sayin aayz~uasardajaan(~smoqs ~a1q~~.pax!ruor?aqsassei~lasseaq,uaa~~aqsuo!~~1ar -jod-ssoj~moi iciaayqaj = ajaqiuaqm(sa~pua3jad~saqz?~qaqiivqqoamsaseaj3ap mq)uopnqrjis~paqijosa~~iua~jadisamo~aqiit!qy~amsasear~ur kkw?~sparj!sjaa!p v wssi?~:, lasse ajow jo 0~1 %uoure pap!a!p /cip.zydll ajv suoynq~~uo3 ‘pw$sui 't&m s!ql aa&?qaq sjolsaau! ma3 ‘jaaamoh *povad uo!~l?~nuin33&? aj!lua aql jaao sse13 passe a@.qs t! 01 apw.u suoynqyuo3 uo paseq a~v sajnq 2 a1qvl aql *afiya3jadqiggl aqiitz spuoqaajodro3 sso~3sy3ois ~0~1~03 put! ‘apyxa3jad ~21 aql ii? uoynqys!p puoq hmseaj~ aqi sassoj3 uoynq!jwp zo9 09s ees 80s olp zep lop 06s ele ose plz z9z esz ppz zez oei 911 ezi ozi 911 zii'i so8 os9 8zs 96s 60l ik 8pp 8le 96z 66e ize llz ipz 661 991 et1 621 lii zoi 6ez'i lo6 zpl 019 999 lsl 985 96p ezp pee sip 8ee 96z 6sz 91z 891 spi eei izi lo1 lz6‘zs 8sp’i i ewp tlr1 099 pp!tpi pzs‘e 089‘ i zsl e8z sse‘e zez'i 8e9 ose 191 l9s p6z 061 6211 9l esz'ei s6 018‘p sl zis‘z os '=aop we‘1 sz 285 s l69‘p s6 sso‘z sl v61‘1 os sil sz e9e s sw i 56 z8l sl sis os spe sz zoz s ‘=a oe ‘=a oz spe s6 8zz sl eli os ‘=a 01 eei sz 26 s 46 financial services renew, 2(l) l!mul!w3 table 3 probability distributions of ending weakb for stock/bond mixes asset allocation of contributions to each asset stock: 100% 75% 50% 25% accum. period qile bonds: 25% 50% 75% io@% 10 yrs. 20 yrs. 30 yrs. 40 yrs. 5 92 loo 108 113 116 25 133 134 133 130 120 50 173 165 155 144 123 75 228 204 180 161 126 95 345 277 227 189 130 5 202 25 345 50 515 75 782 95 1,458 5 363 25 715 50 1,194 75 2,055 95 4,697 5 582 25 1,346 50 2512 75 4,810 95 13,253 227 242 344 327 462 405 637 511 1,042 729 243 299 z 523 232 244 253 262 274 3% 413 397 670 610 526 989 808 638 1,491 1,080 785 2,820 1,688 1,061 350 373 e 432 665 671 615 470 1230 1,053 845 508 1,980 1,488 1,058 533 3,237 2,094 1,350 560 6,901 3.588 1,923 602 nore: ending wealth per dollar of real monthly contribution over n-year (n = io, 20, 30, 40) accumulation periods. contributions are allocated between common stock and corporate bonds. asset allocations between other security classes have less interesting results. common stocks and small stocks have a .85 correlation, so there is little risk reduction benefit through diversification. because of this relatively high correlation, wealth at each percentile of the logon stock/small stock dist~bution is approxi mately equal to the weighted average of wealth at the same percentiles of the single-asset distributions. a similar result is obtained for mixtures of stock and treasury bills, despite their near-zero correlation. real monthly t-bill returns have stayed near zero for most of the 1926-1990 period. the small standard deviation in the treasury bill time series results in a near-linear increase in both expected return and standard deviation as greater proportions of wealth are placed in risky stock. many brokerage services and financial planners advocate investment in com mon stock early in one’s career with a gradual shift to bonds as retirement ap proaches, the ~tirement-we~th effects of several versions of this strategy are illustrated in table 4 for a 30-year accumulation period. long-run returns on stock and rod porifdbs 47 table 4 chmging the stock-bond mix asset years of contributions to each asset stock: l-10 i-15 i-20 i-25 i-30 %ile bo&: 11-30 16-30 21-30 24-30 5 393 391 383 373 363 25 564 60s 647 678 715 50 765 867 983 1.087 1,194 75 1,083 1,304 1345 1,787 2,055 95 1,970 2$36 3,185 3,912 4,697 notes: ending wealth per dollar of real monthly contribution over 30-year accumulation periods. the asset allocation is initially 100% common stock. in later years, new contributions are invested in corporate bonds and stock accumulations are gradually shifted to bonds. the portion transferred at the end of year n is i/(30 -n + 1). consider the first column of the wealth dist~b~tions in table 4. one-dollar real monthly con~butions are invested in common stock during the first 10 years. beginning in year 1 i, new contributions are put into corporate bonds, and at the end of year 11,1/2oth of the stock total is moved to bonds. the transfer proportion rises to l/l9 in year 12, l/18 in year 13, and so on until year 30 when all remaining stock holdings are redeemed. as before, figures in the table show percentiles of ending wealth. other columns in table 4 extend the stock contribution period to 15,20,25, and 30 years, with the 30-year column taken from table 2. the wealth distributions cross around the eighth percentile. this means that there is only an eight percent chance of ending up with less wealth in an all-stock strategy than in a strategy which shifts to bonds earlier in the accumulation period. shifting to bonds early in the accumulation period also gives up much of the upside potential of the all-stock strategy. table 5 illustrates the effect of changing the retirement age when investment contributions begin at age 30. all contributions are invested in common stock. the figures show monthly annuity payments from a lifetime annuity following retire ment at age 60,65, or 70. calculations are based on mo~lity data for men7 in the 1990 life i~~r~~ fact book and a monthly discount rate of .0019, the average real monthly corporate bond return over 1926-1990.8 monthly annuity payments increase dramatically as the retirement age is increased. this is due to the combined influence of two factors: the higher ending wealth from a longer accumulation period, and a shorter remaining lifetime over which to receive annuity payments. the figures in tables 2 through 5 assume the monthly contribution amount remains constant throughout the accumulation periods. adding a growth rate to the con~ibutions increases ending wealth and annuity amounts. for example, consider financial services review, 2(l) 19!w1993 table 5 lifetime annuity payments for different retirement ages retirement age percentile 60 65 70 5 1.76 2.71 3.97 25 3.47 5.58 9.17 so 5.80 9.92 17.12 15 9.99 18.07 32.78 9s 22.82 44.28 90.33 nofex annual payments from a lifetime annuity per dollar of real monthly contribution. contributions begin at age 30 and continue until retirement at ages 60.65. or 70. investments are in common stock throughout the accumulation period. the 1,194 median ending wealth for common stock contributions over 30 years (table 2). the median increases to 1,419 with two percent annual growth in real con~ibutions, and climbs to 2,031 with 5 percent growth. the co~sponding corporate bond totals increase from 496 to 65 1 and then to 1,024. however, growth has little effect on crossing points of the distributions. for example, 30-year common stock and corporate bond distributions still cross around the 4th percentile. the relative shapes and orderings of the distributions presented in table 2 are retained. v. conclusion the return distributions and the resampling procedure of this paper can be used by both individuals and financial planners. probability distributions help determine the size of monthly retirement contributions which are needed to meet a long-term goal. the resampling procedure can adapt results to a particular investor’s invest ment needs. since all results are in real (inffation-adjusted) terms, retirement incomes are conveniently stated in terms of today’s purchasing power. the most striking result is the large advantage held by common stock invest ments over bonds for longer accumulation periods. as the accumulation period lengthens, stock investments become increasingly less “risky” than bonds. a given monthly investment in stocks can offer a reasonable chance of meeting a long-term investment goal, while the same amount invested in bonds generally has a smaller chance of meeting that goal. based on capital market history, we find that over a 30-year accumulation period there is less than a four percent chance that colon stocks will underperform corporate bonds. despite the short-term risks, the stock market is the best bet for most long-term retirement investments. acknowledgments: we thank two anonymous referees and participants at the 1990 academy of financial services annual meetings for helpful comments. lang-run returns on stock and bond potifolios 49 5. 6. 7. 8. notes this point is emphasized by mcenally (1985). the true moments of the distribution are unknown. the sample mean and variance over 1926-l 990 am only estimates of these parameters. butler and domian (1991) show that small positive autocormlations in the historical stock and bond series have little impact on return distributions produced by the resampling methodology. the means of the common stock end-of-period wealth distributions are 190 (10 years), 636 (20 years), 1,705 (30 years), and 4,205 (40 years). small stock distributions have means of 239 (10 years), 1,086 (20 years), 4,020 (30 years), and 14,082 (40 years). the cross-correlation between common stocks and treasury bonds is 0.19. resulting in distri butions which are very similar to those in table 3. samuelson (1990) considers a lump-sum portfolio without additional monthly contributions. the amounts held in each asset class must be rebalanced to maintain the so-50 mix. lower mortality for women produces lower monthly annuity amounts. corporate bonds comprise over 40 percent of the assets of u. s. life insurance companies (see the 1990 life insurance fact book). references butler, kirt c. and dale l. domian. 1991. “risk, diversification, and the investment horizon,” tk journal of portfolio management, 17(3): 41-117. ibbotson, roger g. 1991. stocks, bonds, bills and injlation: 1991 yearbook. r. g. ibbotson associates. leibowitz, martin l. and terence c. langetieg. 1989. “shortfall risk and the asset allocation decision: a simulation analysis of stock and bond risk profiles,” the journal of porrfolio management, 16(l): 61-68. levy, robert a. 1978. “stocks, bonds, bills, and inflation over 52 years,” the journal of portfolio management, 4(4): 18-19. mcenally, richard w. 1985. “time diversification: surest route to lower risk?,” the journal of portfolio management, 1 l(4): 24-26. 19w life insurance fact book. washington, d.c.: american council of life insurance. reichenstein, william. 1986. “when stock is less risky than treasury bills,” financial analysts journal, 42(6): 71-75. samuelson, paul a. 1990. “asset allocation could be dangerous to your health,” the journal of portfolio management, 16(3): 5-8. financial services review, 33(1) 1 the impact of the covid-19 income shock on debt management: a mediation analysis congrong ouyang,1 thomas crandall,2 and swarn chatterjee3 abstract this study uses the 2021 wave of the finra national financial capability study dataset to examine the association between large and unexpected income drops experienced by individuals during the covid-19 pandemic and an individual’s use of stimulus checks to settle debt obligations. this study also examines the mediating role of individuals’ perceived lack of financial control in the association between the large drop in income and the use of stimulus checks for debt payments. the results reveal that over one-third of households allocated their stimulus checks towards debt payments. notably, individuals experiencing a large and unexpected drop in income had 5.5% higher odds of using pandemic stimulus checks for debt management. moreover, this relationship was significantly mediated by an individual’s perceived lack of financial control. the findings from this study shed light on the complex associations between experiencing an unexpected and large income reduction, perception of financial control, and debt management decisions of individuals. the significant role of perceived financial control in describing individuals’ debt management decisions found in this study suggests that perceived control is not just a reflection of a household’s financial situation but also a determinant of their financial decision-making in times of crisis. furthermore, the results underscore the critical role of stimulus checks and other financial assistance in mitigating the economic impacts of the pandemic on american households. findings from this study contribute to a deeper understanding of financial decision-making processes during periods of economic uncertainty and offer implications for future economic policies and financial literacy programs. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation ouyang, c., crandall, t., & chatterjee, s. (2025). the impact of the covid-19 income shock on debt management: a mediation analysis. financial services review, 33(1), 1-27. introduction the worldwide outbreak of covid-19 infections in early 2020 led to an acute economic crisis (borio, 2020). notably, about 22 million workers 1 corresponding author (congrong@ksu.edu). kansas state university, manhattan, ks, usa 2 kansas state university, manhattan, ks, usa 3 university of georgia, athens, ga, usa filed for unemployment in the united states. during the four weeks from march 2020 to april 2020 (armantier et al., 2021). over 27% of u.s. households reported experiencing financial https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(1) 2 hardship, and over 30% missed debt payments (liu et al., 2021). the u.s. government responded to the economic crisis by introducing the 2020 coronavirus aid, relief, and economic security act (cares act, 2020), signed into law on march 27th, 2020, which provided a stimulus check of up to $1,200 per adult and $500 per child for most americans in the united states earning less than $99,000 (or $198,000 for joint tax filers) (treasury.gov). by march 2021, following two additional rounds of stimulus, pandemic-related disbursements to american households tallied $803 billion (pandemic oversight, 2022). consumer debt obligations of americans had been steadily rising long before the covid-19 pandemic (berger & houle, 2019; fan & chatterjee, 2017; ouyang & hanna, 2022). the total consumer debt in the united states increased from below $10 trillion in 2005 to more than $16 trillion mid-way through 2022, $11.6 trillion of which was outstanding mortgage debt (federal reserve bank of new york, 2022). past literature has shown varied results for the consequences of debt on people’s well-being and decision-making (amromin & smith, 2003; argento et al., 2015; butrica et al., 2010; tharayil & walstad, 2022); however, differences in debt payment classifications, the specific consumer debt variables analyzed, and the various research model structures employed in these previous studies , have likely influenced these results. in these studies, financial shocks, such as unexpected and significant income drops or sudden surges in expenses, have been associated with individuals’ debt management decisions. these differences in results and hints of association with decisions around debt service found in the extant literature underscore the need for new and continued research in this area. with debt loads higher than ever, consumers were particularly susceptible to the economic instability brought forward by the covid-19 pandemic. income shocks can directly result in debt management challenges and adversely affect the psychological well-being of households (netemeyer et al., 2018; shevlin et al., 2020). over the long term, studies have found that experiencing debt hardships (i.e., delinquency on debt payments and credit constraints) can increase mortality risk and affect the psychological well-being of both older women and men (marshall & tucker-seeley, 2018; tucker-seeley et al., 2009). this effect may be magnified for pre-retiree adults (ages 50-65), given their low financial capability and inadequate retirement savings (gillers et al., 2018; lusardi & de bassa scheresberg, 2016). the experience of personal financial hardship resulting from the inability to meet financial obligations, such as paying bills or servicing debts, has consistently been associated with adverse effects on households’ physical and psychological well-being (drentea & lavrakas, 2000; sweet, 2021). moreover, the inability to meet debt obligations has been associated with helplessness or a perception of loss of control over one’s financial situation when confronted with the impossibility of fulfilling financial obligations during times of financial hardship (lea et al., 1995). however, other studies have shown that when individuals experience uncertainty regarding their financial situation, they are driven to make decisions aimed at gaining clarity and regaining a perception of control over their finances (whitson & galinsky, 2008). due to differences in household financial resources, individuals may have distinct approaches to utilizing stimulus checks, including smoothing out their consumption patterns and managing debt obligations, especially when encountering unexpected economic shocks. consequently, people who experienced financial challenges during the covid-19 pandemic may have been inclined to utilize their available financial resources, including stimulus checks, to pay outstanding debt obligations. this urgency to pay off one’s debt obligations arises from the potentially severe consequences of the inability to meet such obligations, which may be salient in peoples’ minds. additionally, settling a portion of their debt obligations can contribute to regaining a perception of control over their financial situation. there is currently very little information available in the extant literature on the behavioral aspects of peoples’ financial decision-making during the covid-19 pandemic (xu & yao, 2022; yue et al., 2020). our study fills this important gap in the literature by examining the ouyang et al. 3 association between financial hardship and the perception of being controlled by one’s financial situation when managing debt payments amid the pandemic. furthermore, our study provides insight into the impact of stimulus payments in mitigating the debt management challenges experienced by households. literature review influence of perception of control on financial behavior and debt repayment prior research has generally converged on the consideration that individuals who perceive themselves to be in control over their life situations exhibit positive financial behaviors toward savings and budgeting (cobb-clark et al., 2016; perry & morris, 2005), debt management (caputo, 2012), investing (salamanca et al., 2020), and retirement planning (foltice & ilcin, 2019). conversely, individuals who feel a lack of control over their financial situation are likely to make cognitive errors in their financial decisions that may further exacerbate their financial situations (lindgren, 1980). a norwegian study on immigrants found that experiencing adverse economic situations and experiencing a perceived lack of control over their life situations were associated with adverse psychological outcomes (dalgard et al., 2006). additionally, zhang (1989) found that people with fewer financial resources were more likely to develop a plan for using their available financial resources. another study found that the debt burden of college students was associated with a student’s perception of lack of financial control (dryden et al., 2023). similarly, goel and rastogi (2023) found that individuals who exhibited an external locus of control, or perceived that they did not have control over their life situations, were associated with lower creditworthiness. similarly, kamleithner et al. (2013) found that over-indebtedness was associated with greater financial stress, adverse money management behaviors, and a lower perceived control over one’s financial situation. debt management decisions during income shocks income shocks pose significant challenges to households, often forcing them to reassess and modify how they manage financial obligations (colarieti et al., 2024). while income volatility in the u.s. has generally trended down in the united states since the late 1990s, it remains that volatility is negatively associated with income stability, except for those at the top of the income distribution curve (guvenen et al., 2022). additionally, despite the overall trend lower, the united states has faced three communal economic shocks in the last twenty-five years: the ‘dot.com’ bust (2000-2002), the global financial crisis (2007-2009), and the covid-19 pandemic (2020-2021). income shocks lead to increased income volatility, particularly during recessionary environments (guvenen et al., 2022; peetz & robson, 2021). broad societal trends, including the emergence of the gig economy, a population that faces higher undesired income volatility and little safety net (peetz & robson, 2021), may serve to exaggerate these trends in the future. households respond to income shocks in various ways. some households may respond by increasing their usage of debt facilities or drawing from savings, including those savings earmarked for retirement, to supplant at least a portion of the income that was lost (colarieti et al., 2024; ghilarducci et al., 2019). some households are more likely to use available resources, such as stimulus checks, to manage debt when they feel a loss of control over their financial situation (coibion et al., 2020). others may change spending habits to control the cost side of the ledger (colarieti et al., 2024), with more stringent cost controls put in place by those with higher debt levels (baker, 2015). while colarieti et al. (2024) ascribe behavioral components to this cohort, baker(2017) research on the 2013 united states federal government shutdown suggests that a household’s decision to reduce spending may simply be a function of their access to debt facilities. (colarieti et al., 2024; kamakura & du, 2012). financial literacy may be a mitigating factor in how or whether one uses debt during times of financial uncertainty (lusardi & mitchell, 2013).the diversity in responses to financial shocks underscores the importance of tailored financial education and interventions that enhance financial resilience. by understanding the nuanced behaviors that characterize different demographic groups under financial services review, 33(1) 4 financial stress, policymakers and financial advisors can better support households in strengthening their financial positions against future economic uncertainties. households debt during covid-19 pandemic consumer over-indebtedness has been a longstanding issue in the u.s., and debt has now “become a persistent feature of household balance sheets at every stage of life" (lusardi & tufano, 2015, p. 360). the issue grew during the covid-19 pandemic when households faced severe and unexpected negative shocks to household income and expenses. overindebtedness is a vital social issue that can result in extreme financial difficulties for the household, including bankruptcy, foreclosure, being denied credit, and future employment (bricker & thompson, 2016). when economic conditions deteriorate, like they did during the great recession (the period 2007 – 2009), decreased employment opportunities and incomes can strain family finances and lead to a rise in missed debt payments and defaults (ouyang & hanna, 2022). however, researchers found that delinquency rates fell during the covid-19 pandemic despite a historic rise in unemployment (dettling & lambie-hanson, 2021). pandemic-related stimulus checks received from the federal government in 2021 may have provided cushions that households used to make debt payments. stimulus use economic impact payments (eip) were provided during the covid-19 pandemic to assist u.s. households with their financial needs and stimulate the economy (garrison et al., 2022). according to the household pulse survey in june 2020, most households that received or expected to receive a stimulus payment planned to use the payment for expenses, while less than a quarter of households planned to use stimulus checks to pay down debt (garner et al., 2020). the federal reserve bank of new york reported that about 34.5% of households used their first round of eip stimulus payments on debt payments, 37.4% of households used their second round of stimulus payments on debt payments, and 33.7% of households used their third round of stimulus payments on debt payments (armantier et al., 2022). further, armantier et al. reported that households with lower than college degrees, from lower income thresholds, and that had negative shocks on income and/or employment during the pandemic, utilized stimulus checks for debt payments (2020). among pre-retirees (ages 50 to 65), more than 45% of households reported using stimulus checks to pay down debt (liu et al., 2021). economic stimulus payments had been used before and were not new to consumers. in 2008, u.s. households received over $120 billion from the economic stimulus act of 2008 for financial assistance and to help with recovery from the great recession (amadeo, 2020; parker et al., 2013). based on research results, 48% of consumers used these stimulus payments for debt payments, 32% allocated it toward saving, and less than 20% spent the funds (shapiro & slemrod, 2009). homeowners spent more of their stimulus payments than did renters, especially on debt payments. (parker et al., 2013). broda and parker (2008) discovered that although a substantial majority of households used the 2008 stimulus payments for debt repayment or savings, households with lower income were nearly twice as inclined to expend the extra funds. the consumer financial protection bureau reported an eight percentage-point decline between june 2019 and june 2020 in the proportion of people indicating a high perceived lack of financial control (fulford et al., 2021). the first of the economic impact payments was signed into law and disbursed during this window, though the authors stop short from claiming the drop is related to the stimulus. however, (kleimeier et al., 2023) claim that only economically weak households perceived a gain in control over their finances as a result of receiving government stimulus checks during covid-19. while the literature has shown the significant usage of stimulus checks to make debt payments, limited studies have analyzed the impact of households’ perceived lack of control of their finances on their financial decision-making or the effect of government-supplied stimulus payments ouyang et al. 5 on an individual’s perception of control. moreover, the findings from previous literature underscore the need for more nuanced empirical models that account for variability in income and the psychological dimensions of financial decision-making, especially during periods of economic uncertainty. theory and conceptual model extended life cycle savings theory the original life cycle model (modigliani & brumberg, 1954), which assumed certainty about future income, has evolved into the extended life cycle model to incorporate the realistic element of income uncertainty (shefrin & thaler, 1988). this model suggests that unpredictability in future income patterns significantly influences household savings and net worth accumulation (yuh & hanna, 2010). the extended life cycle certainty equivalence model, which further extends the life cycle hypothesis, posits that individuals who face greater income uncertainty should ideally save more and borrow less. empirical evidence supports this theory. browning and lusardi (1996) found a correlation between transitory income uncertainty and increased savings from current income, leading to reduced borrowing. our study extends these findings to the context of the covid-19 pandemic, a period marked by heightened income uncertainty. the pandemic's impact on employment status and income stability underscores the relevance of these theories in contemporary economic scenarios, especially for individuals who experience more variable income patterns. according to browning and lusardi (1996), such individuals should ideally borrow less than their more stable counterparts, reflecting the need for precautionary savings. however, the normative life cycle saving (lcs) model, which assumes that individuals are well-informed and fully rational (browning & crossley, 2001; modigliani & brumberg, 1954), may not fully account for the bounded rationality of real-world scenarios. ibrahim and alqaydi (2013) argued that individuals face challenges due to the complexities associated with financial decisionmaking, limited financial capability, and constraints in time and resources. this perspective is crucial in understanding peoples’ responses to income shocks, as we expected to observe in our study. additionally, a perception of losing control over one’s financial situation could arise in cases where an individual’s available financial resources were insufficient to offset the financial shock and uncertainty endured during the covid-19 pandemic (lea et al., 1995). past research has shown that feeling a sense of control over one’s circumstances reduces stress and increases resilience during periods of uncertainty (bordia et al., 2004; glass et al., 1973; polizzi et al., 2020). when people perceive a loss of control over their situation, the resulting stress can motivate them to make decisions aimed at altering their current circumstances to gain a greater perception of control over their situation (brehm, 1966; whitson & galinksy, 2008). during pandemic, when economic uncertainty is high and income is potentially reduced, many households received stimulus checks as a form of government support (cares act, 2020). according to life cycle savings theory, these households are likely to use these funds to pay down debt on maintaining balanced consumption over their lifetime (modigliani & brumberg, 1954; yuh & hanna, 2010). by reducing debt, households decrease their future interest obligations and potential financial distress from perceived financial control, which could disrupt their consumption smoothing (netemeyer et al., 2018). essentially, paying off debt during a downturn allows households to adjust their financial strategies to ensure more stable consumption in the future, minimizing the impact of current income shocks (lea et al., 1995). furthermore, life cycle savings theory suggests that by reducing liabilities, households are effectively increasing their net savings, which aligns with the theory’s principle of preparing for later stages of life when income will decrease due to retirement (modigliani & brumberg, 1954; yuh & hanna, 2010). in the context of a pandemic, this behavior reflects a precautionary saving motive, where households prioritize financial stability and resilience in anticipation of prolonged economic uncertainty or additional negative income shocks (lusardi, 1998; deaton financial services review, 33(1) 6 2005). thus, the use of stimulus checks to pay off debt during a pandemic can be seen as a strategic alignment with the life cycle savings theory, aiming to secure financial stability and smooth consumption over the uncertain period ahead (deaton, 2005). many individuals experienced income uncertainty and financial hardship during the covid-19 pandemic (bierman et al., 2021; kim et al., 2019; yue et al., 2020), and it is probable these individuals were inclined to tap into their financial resources, including stimulus checks, to settle their debt-related obligations. this decision to pay one’s debt is motivated by a desire to reduce financial uncertainty and to regain a greater perception of control over one’s financial situation (baum et al., 1986; lea et al., 1995; whitson & galinsky, 2008). hence, as shown in figure 1, we also expect that a perception of lack of control would mediate the association between experiencing financial hardship and the use of stimulus checks to pay off debt. our study contributes to understanding how perceived financial control and negative income shocks during the covid-19 pandemic influenced household debt management decisions. figure 1. conceptual framework of this study methods dataset and analytical sample this study used the 2021 national financial capability study (nfcs) state-by-state dataset collected and published by the financial industry regulatory authority (finra) foundation. the nfcs survey has been administered every three years since its first collection in 2009 (finrafoundation.org). information on the diverse characteristics of respondents, including their financial knowledge, behaviors, and attitudes, is collected. the 2021 nfcs was collected between june and october 2021. the focal variables asked about events during the covid-19 pandemic. a total of 20,218 respondents were included in the original sample, which was reduced to 18,790 respondents who received pandemic-related stimulus payments. respondents who did not receive the stimulus check were removed from the study. national sampling weights available in the 2021 nfcs datasets were used. the sample weights used were representative of the respondents’ age, gender, ethnicity, education, and census division based on the american community survey (lin et al., 2022). endogenous (dependent) variables the major dependent variable, “use stimulus checks for debt payments,” was defined based on two questions from the nfcs survey (table 1). the first question was whether one received stimulus checks from the federal government. respondents were asked, “did you receive a pandemic-related stimulus payment from the federal government in 2021?” the possible responses to this question were (1) yes, (2) no, (98) don’t know, and (99) prefer not to say. the variable was coded as 1 if yes and as 0 if otherwise. only respondents who reported “yesto receiving stimulus checks, were captured in this study. a follow-up question was presented to respondents who answered (1) yes to the question of “did you receive a pandemic-related stimulus payment from the federal government in 2021?” these respondents were subsequently asked, “what did you use the money for?” the variable was coded as 1 if “paid down debt” was ouyang et al. 7 selected as one of their uses of stimulus checks; the variable was coded as 0 if this was not the case. observations with ‘don’t know’ and ‘prefer not to say’ responses were dropped from analyses. exogenous (independent) variables periods of income shock have been associated with increased savings in past literature (abdelrahman & oliveira, 2023; cox et al., 2020; immordino et al., 2022). hence, income shock was included as a control variable for the empirical analyses of this study. negative income shock, or income drop during covid-19, was assessed based on the question (table 1), “in the past 12 months [have you / has your household] experienced a large drop in income which you did not expect?” possible options were (1) yes, (2) no/not applicable, (98) don’t know and (99) prefer not to say. a binary indicator was created and ‘don’t know’ and ‘prefer not to say’ responses were dropped from the analyses. mediator previous studies have found that a household’s money attitude, encompassing aspects such as money anxiety and money confidence, mediates their financial behaviors (forbes & kara, 2010; gasiorowska, 2014; heo et al., 2016; hayes, 2013; tang, et al., 2005). respondents’ attitudes toward their household finances were measured based on a survey question about self-assessed financial control which reads, “my finances control my life,” with answers ranging from 1 to 5. a higher score indicates a respondent perceived a greater lack of control over their finances (see table 1). control variables sociodemographic variables including age, gender, race/ethnicity, number of financial dependents, education, employment status, marital status of the respondent, homeownership, and household income were controlled because of their association with financial decision making in past literature (grable et al., 2009; grable et al., 2023; ouyang & hanna, 2022). other household factors related to the presence of an emergency fund, financial strains, objective and subjective financial knowledge 4 , and subjective credit record because of their expected association with financial behaviors and outcomes (fan & chatterjee, 2017; grable & palmer, 2022; lee et al., 2023; liu et al., 2021). 4 subjective financial knowledge responses ranged from 1-7: “on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall financial knowledge?” (nfcs, 2021). objective financial knowledge was computed based on respondents’ correct answers to seven financial knowledge related questions. the scores ranged from 0 (incorrect answers to all seven questions) to 7 (correct answers to all seven financial knowledge questions) (nfcs, 2021). the financial knowledge questions are described in appendix a). financial services review, 33(1) 8 table 1. overview and descriptions of key variables used in the empirical analysis variables description exogenous (independent) variables income drop answers based on question: “in the past 12 months, [have you / has your household] experienced a large drop in income which you did not expect?” yes was coded as 1, otherwise 0. mediator variable perceived lack of financial control self-reported scale from 1 –5 based on the statement: “my finances control my life” was recorded and normalized. where “1” means to as “never”, “2” means “rarely”, “3” means “sometimes”, “4” means “often”, “5” means “always”. endogenous (dependent) variables use stimulus checks for debt payments “what did you use the money for?” households chose “2. paid down debt” were recorded as 1, otherwise 0. methodology a weighted logistic regression, as shown in equation (1), is employed to examine the sociodemographic and financial characteristics correlated with using stimulus checks for debt payments. 𝑃(𝑌) = 𝑒𝛼+𝛽.𝑋𝑖 1+ 𝑒𝛼+𝛽.𝑋𝑖 (1) where y takes the value 1 if a respondent selected yes to having utilized stimulus checks to make debt payments, y takes the value 0 otherwise. 𝑋𝑖 denotes the sociodemographic and financial characteristics (seen in table 3) controlled for in the study. to further analyze households’ utilization of stimulus checks, we conducted a mediation analysis to investigate whether an individual’s perceived lack of financial control mediated the association between income shock during covid-19 and the utilization of stimulus checks for debt repayment. for this purpose, we employed a potential outcomes framework (pof), a fundamental tool in causal inference, which facilitates the estimation of causal effects by considering both observed and unobserved potential outcomes. this framework allows for the systematic examination of the values of outcomes that would prevail under varying conditions, such as the presence or absence of a treatment (rubin, 2005; imai et al., 2011). the pof allows the study to estimate what would happen to the debt management behavior of individuals under different scenarios: receiving a stimulus check versus not receiving one, and experiencing an income shock versus not experiencing one. this framework helps in drawing causal inferences about the effect of these variables on debt management decisions. by employing a mediation analysis within the pof, the study quantifies how much of the effect of the income shock on debt management is direct and how much is mediated through changes in perceived financial control. this is crucial for understanding not just the direct impact of economic shocks but also the psychological pathways through which these shocks influence financial behaviors (imai et al., 2011). the framework allows for the decomposition of the total effect of the income shock on the use of stimulus checks for debt payments into direct effects (impact of income shock directly on the outcome) and indirect effects (impact mediated through changes in perceived financial control). our analysis integrates causal mediation models to evaluate the impact of an intervention on specific outcomes, acknowledging that this impact may be direct or indirect through an intermediary variable identified as a mediator. the potential outcomes framework provides the flexibility needed for the analysis, allowing us to account for potential interactions between the mediator and the treatment. consequently, we did ouyang et al. 9 not presuppose a uniform effect of the mediator on the outcome across treated and untreated groups. we utilized the mediate package in stata (version 18) for the mediation analyses, applying it to both the full analytical sample and conceptual model groups. this involved the decomposition of the total effect of the treatment on the outcome into direct and indirect effects in two distinct ways, aligning with our research question. we defined these effects in a model-free manner, enabling the selection of an estimation method best suited to our data. notably, classical approaches and causal mediation analyses via the potentialoutcomes framework converge to similar results when conducting linear regressions for endogenous variables and the mediators (pattanayak, et al., 2011; stata, 2023). results descriptive results in table 2, we present the descriptive statistics of the variables used in the study. among the respondents who received stimulus checks during the covid-19 pandemic, 33.54% reported using these funds for debt payments. the average score for households' perceived lack of financial control was 2.96, within a range of 1 to 5. the responses varied, with 'sometimes' being the most common (30.25%), followed by 'rarely' (22.85%), 'often' (17.22%), 'always' (15.14%), and 'never' (14.54%). additionally, about 25.76% of the respondents reported a substantial income drop during the pandemic. regarding control variables, the demographic composition of the sample included about 54% females and 74% non-hispanic whites. living arrangements and employment status were also notable, with 58% living with a partner and approximately 39% employed full-time. financial security was a key concern, as almost 56% of respondents had emergency funds, and 20.5% experienced layoffs during the pandemic. homeownership and credit quality were also reported, with nearly 60% being homeowners and about 45% having a 'very good' credit record. finally, the average objective financial knowledge score of the sample was 3.52 (on a scale of 1 to 7), whereas the average subjective financial knowledge was 5.12 (on a scale of 1 to 7), indicating varied levels of financial literacy among the respondents. weak correlation was observed between objective and subjective financial knowledge (0.24) justifying controlling for both objective and subjective financial knowledge in our empirical analyses. this is reported in appendix b. financial services review, 33(1) 10 table 2. summary statistics key variables % stimulus checks for debt payments 33.54 perceived lack of financial control 1 never 14.54 2 rarely 22.85 3 – sometimes 30.25 4 – often 17.22 5 always 15.14 income drop yes no 25.76 74.33 control variables age 18-24 11.10 25-34 17.32 35-44 16.83 45-54 17.03 55-64 17.45 65+ 20.29 gender male 45.97 female 54.03 ethnicity non-hispanic white 73.98 non-white 26.02 marital status non-partner 41.99 partner 58.01 employment employed 38.55 part-time 8.70 self-employed 7.90 homemaker 6.71 full-time student 2.80 unemployed 8.10 disabled 5.65 retired 21.59 emergency funds 55.89 laid off during pandemic 20.54 have money left never 11.25 rarely 18.00 sometimes 25.93 ouyang et al. 11 often 19.42 always 25.40 homeownership 59.88 financial strain 31.94 subjective credit record very bad 4.15 bad 12.34 average 17.98 good 20.55 very good 44.97 note. n=18,790; mean of objective financial knowledge = 3.52, subjective financial knowledge =5.12. multivariate results table 3 presents the outcomes from the logistic regression analysis, which focused on discerning significant patterns among households utilizing stimulus checks for debt repayments. the analysis revealed that with all other variables controlled, households that reported an income drop during the pandemic had about 1.12 times the odds of using received stimulus checks for debt payments compared to households that reported no income drop. furthermore, the degree of perceived lack of financial control emerged as a robust predictor in this context. relative to households reporting no perceived lack of financial control, those categorizing their lack of financial control as 'rarely' demonstrated 34% higher odds of using stimulus checks for debt repayments. the odds escalated progressively with higher levels of perceived lack of control: households reporting 'sometimes' showed 60.3% higher odds, 'often' corresponded to 75% higher odds, and those indicating 'always' exhibited 91.2% higher odds of using their stimulus checks for debt payments. these findings underscore the influential role of perceived lack of financial control in the decision-making processes regarding the use of stimulus funds for debt management. controlling for other variables, middle-aged households demonstrated a greater propensity for using stimulus proceeds for debt service than their younger counterparts. specifically, households aged between 25 and 34 were found to have 42% higher odds of using their stimulus checks for debt payments than those under 24 years of age. this trend continued with older age groups: households aged 35 to 44 had 29.30% increased odds, and those aged 45 to 54 showed 17.30% higher odds compared to the under-24 cohort. gender differences were also notable. women had 8.30% higher odds of utilizing their stimulus checks for debt payments compared to men. ethnicity further influenced this pattern, with non-white households having 15.40% higher odds of using these funds for debt payments than white households. employment status and income levels also played a role. compared to employed households, households comprised of self-employed individuals, part-time workers, homemakers, and full-time students were less likely to allocate stimulus checks towards debt repayments. additionally, households with an income level above $15,000, particularly those in the middle to upper-middle-class bracket, exhibited higher odds of using their stimulus checks for this purpose compared to households with an income level of less than $15,000. after controlling all other factors, the presence of emergency funds significantly influenced the use of stimulus checks for debt repayments. households with such funds exhibited 1.55 times the odds of using their stimulus checks for debt payments compared to those without emergency funds. additionally, homeownership was a notable predictor; homeowners had 1.67 times the odds of using stimulus checks for debt payments than renters. financial services review, 33(1) 12 we also found a positive association between households' financial knowledge, both subjective and objective, and the use of stimulus checks for debt payments. households experiencing financial strains were 1.27 times the odds of using these checks for debt payments than those without such strains. credit history emerged as another influential factor. households with a better credit record, compared to those with a very poor credit record, showed a positive association with the use of received stimulus checks for debt payments, with odds ranging from 1.27 times to 2.40 times. furthermore, the extent of remaining funds after essential expenses also affected this usage. households reporting 'rarely' having money left exhibited 16.00% higher odds of using their stimulus checks for debt payments than those with no money left. interestingly, households that 'often' or 'sometimes' had funds remaining were less likely to use their stimulus checks for debt payments, with 18.00% and 37.00% lower odds, respectively. table 3. logistic regression of whether households used stimulus checks for debt payments use stimulus checks for debt payments odds ratio s.e. exogenous (independent) variables income drop yes 1.12** .05 lack of perceived fin control rarely 1.34*** .08 sometimes 1.60*** .10 often 1.75*** .12 always 1.91*** .14 control variables age group 25-34 1.42*** .10 35-44 1.29** .10 45-54 1.17* .09 55-64 1.10 .09 65+ 1.05 .09 gender female 1.08* .04 ethnicity non-white 1.15*** .05 family status live with partner/spouse 1.13** .04 employment self-employed .73*** .05 work part-time .84** .05 homemaker .76*** .05 full-time student .76* .10 permanently sick .96 .08 unemployed .72*** .05 retired .77*** .05 education some college 1.10* .05 ouyang et al. 13 associate's degree 1.15* .07 bachelor's degree 1.09 .05 post graduate degree .96 .06 income levels $15,000 to $25,000 1.25** .09 $25,000 to $35,000 1.40*** .10 $35,000 to $50,000 1.34*** .09 $50,000 to $75,000 1.45*** .11 $75,000 to $100,000 1.42*** .12 $100,000 to $150,000 1.35** .12 $150,000 to $200,000 1.24 .15 $200,000 to $300,000 1.11 .22 $300,000 or more 1.24 .32 laid off during pandemic yes 1.09 0.05 emergency fund yes 1.55*** .07 homeownership yes 1.67*** .04 objective financial knowledge 1.02* .01 subjective financial knowledge 1.03* .01 financial strain yes 1.27*** .05 subjective credit record bad 1.30*** .12 average 1.94*** .17 good 2.40*** .22 very good 1.79*** .17 have money left rarely 1.16* .07 sometimes 1.08 .07 often .82** .06 always .63*** .05 n 18,790 note. standard errors in parentheses, * p<.05 ** p<.01 *** p<.001; financial strain includes unpaid medical bills, late payments on student loans, late on mortgage payments, and being late on credit card payments; never perceived lack of financial control, employed, high school, age group of younger than 25 years, annual income less than $150k, no subjective/objective financial knowledge, very bad credit record, no money left are the omitted categories. mediation results for the full sample table 4 presents the mediation results for the full sample. the total effect indicates that for households who received stimulus checks, the probability of using the stimulus checks for debt payments increased by 0.05 points on the probability scale compared to households that did not report an income drop. the indirect effect indicates the indirect association between households’ income drop financial services review, 33(1) 14 and utilization of stimulus checks to pay down debt when mediated by lack of financial control. the direct effect captures the direct association between households’ income drop and the likelihood of using stimulus checks to pay down debt when controlling for other factors. the results from the full mediation model, as shown in table 4, indicate that income drop had a total (coef = 0.05***; odds = 1.24), direct (coef = 0.04***; odds = 1.18), and indirect association (coef = 0.01***; odds = 1.06) through lack of financial control, on the utilization of stimulus checks for paying down debt. table 4. full mediation model results mediator: perceived lack of financial control treatment te de ie income drop 0.05*** odds = 1.24 0.04*** odds = 1.18 0.01** odds = 1.06 note: outcome equation includes treatment and mediator interaction. coefficients in parentheses, * p<.05 ** p<.01 *** p<.001; n = 18,790. table 5 presents the mediation results for the partial model, where we focused on the mediation effects among respondents’ reported income drop—perceived lack of financial control— utilized stimulus checks for debt payments (figure 1). the results indicate that income drop had a total (coef = 0.11***; odds = 1.59), direct (coef = 0.08***; odds = 1.41), and indirect association (coef = 0.03***; odds = 1.13) through lack of financial control, on the utilization of stimulus checks for paying down debt. therefore, based on the partial model, the probability of using the stimulus checks for debt payments increased by 10% on the probability scale compared to households that did not report an income drop. table 5. partial mediation model results mediator: perceived lack of financial control treatment te de ie income drop 0.11*** odds = 1.59 0.08*** odds = 1.41 0.03** odds = 1.13 note: outcome equation includes treatment mediator interaction; coefficients in parentheses, * p<.05 ** p<.01 *** p<.001; n = 18,790; multicollinearity analysis was also done for the study. the variable inflation factors (vif) for the variables are reported in appendix c. no multi-collinearity was found in the analyses of this study. discussion this study examined how a perceived lack of financial control influences the relationship between unexpected negative income shocks and the use of stimulus checks for debt payments during the covid-19 pandemic. the findings revealed a significant association between reported income drop and the use of stimulus checks for debt payments, supporting the study’s proposed mediation model. regardless of other factors, a household's perceived lack of financial control was a strong predictor of using stimulus checks for debt payments. the greater the perceived lack of control, the more likely households were to allocate stimulus checks towards debts. this suggests that policy measures aimed at increasing financial stability should not only provide temporary financial assistance but also support efforts to enhance financial literacy and management skills among the population. this research validates earlier findings that a lack of financial control mediates individuals’ ouyang et al. 15 willingness to alleviate their financial situation and improve their financial wellness by settling their debt obligations when faced with sudden income shock or economic uncertainty (gasiorowska, 2014; tang et al., 2005). we found that households experiencing an income drop were 1.055 times more likely to use stimulus checks for debt payments than those without such income shocks, especially when influenced by perceived financial insecurity. further, with other factors controlled, households that experienced an income drop exhibited 13.10% higher odds of using stimulus checks for debt payments, mediated by their perceived lack of financial control. however, it should be noted that income drop was a self-reported measure, reflecting the presence of an income shock during the pandemic but not the specific magnitude. the decision to use stimulus checks for debt payments represented a direct financial response to the pandemic. the perceived lack of financial control was gauged through self-assessment rather than their actual financial management abilities. this study suggests that perceived financial control—or the lack thereof—plays a mediating role between negative income shocks and the decision to pay off debt using financial resources, including using covid-19 stimulus checks. it has been suggested by others that the decision to pay one’s debt is motivated by a desire to reduce financial uncertainty and to regain a greater perception of control over their financial situation (e.g., baum et al., 1986; lea et al., 1995; whitson & galinsky, 2008). furthermore, the findings from this study underscore the importance of considering psychological factors, such as individuals’ perceptions of control and helplessness, when examining their debt management behaviors. the findings from this study also underscore the usefulness of considering self-assessment of individuals’ ability to be in control of their finances when examining the financial behavior of individuals who experience income shock or economic uncertainty. moreover, while the findings from this study indicate that providing households with stimulus checks during the covid-19 pandemic was a timely intervention that helped some people reduce their debt obligations, findings also underscore the importance of educating people to build sufficient emergency savings to sustain a household through periods of economic uncertainty. emergency savings may play a protective role as a financial resource during periods of income uncertainty, in addition to being an important factor in the psychological well-being of people. this is because people can use emergency savings to meet their financial obligations, and the decision to reduce their debt obligations may provide them with a sense of control over their financial situation during periods of economic uncertainty. based on the findings from this study, outstanding debt was likely a salient factor for individuals who faced income uncertainty and were unable to meet their financial debt obligations. interestingly, the results from this study indicate that traditionally financially underserved socio-demographic groups such as women and non-white households were more likely to use stimulus checks to reduce their outstanding debt obligations. additionally, the results from this study suggest that people with higher levels of financial resources may be able to cope better during periods of income uncertainty than people without adequate emergency funds. and as a corollary, it should be noted that carrying a lower debt burden can protect people from experiencing adverse outcomes when experiencing financial hardship; and having adequate emergency savings, while being able to maintain lower financial debt burdens, may ultimately help households stay out of poverty during periods of income uncertainty. individuals’ debt burdens can have deleterious consequences on their well-being, and more broadly, the consequences of carrying debt obligations may have a cascading negative effect in the community during periods of economic uncertainty. the significance of financial knowledge variables also underscores the need for developing greater financial capability within the population as previous studies have indicated that financial capability can play an important role and contribute to the overall financial resiliency of the population (klapper & lusardi, 2020; lusardi & mitchell, 2013). financial services review, 33(1) 16 policy implications from this study are significant. first, it suggests that economic policies, such as the provision of stimulus checks, play a critical role in supporting households during downturns by providing them with the necessary liquidity to manage debts. however, the effectiveness of such measures could be enhanced by simultaneous initiatives aimed at improving financial literacy and control. programs designed to increase financial awareness and planning capabilities could help individuals make more informed decisions about how to use such financial aids effectively. additionally, our findings suggest a need for policies that are tailored to the psychological impacts of financial stress. since perceived lack of financial control is a significant stressor that influences financial behavior, interventions that address both the financial and psychological needs of individuals during crises are crucial. this might include counseling services or workshops on financial management during the rollout of stimulus measures. in conclusion, while stimulus payments serve as essential tools for immediate economic relief, their long-term efficacy could be significantly enhanced by policies that also address financial education, financial advice and psychological resilience. this multidimensional approach could form a cornerstone of future economic crisis interventions, ensuring that households are not only financially supported but also empowered to manage their financial health proactively during and after economic shocks. limitations while providing valuable insights, this study has some limitations. primarily, its cross-sectional design restricts its ability to capture the dynamic nature of financial attitudes and behaviors over time. given the ever-evolving economic environment, understanding how an income drop and the perceived lack of financial control evolve and interact over time is crucial. future research could employ longitudinal datasets to delve deeper into these dynamics, potentially uncovering causal relationships and offering a more nuanced understanding of how these factors influence households' financial decisions as economic conditions fluctuate. moreover, the study's focus could be broadened in subsequent research to encompass a more diverse range of factors that contribute to financial well-being. investigating various dimensions of money attitudes beyond the scope of perceived financial control (or lack thereof), would enrich the field’s understanding of the mechanisms at play in the relationship between changes in household wealth and overall financial health. additionally, incorporating variables such as personality traits, community backgrounds, and other financial benchmarks could offer a more holistic view of the factors that shape households' financial decision-making processes. furthermore, the study suggests potential linkages between perceived lack of financial control, negative income shocks, and debt management. however, there are additional relationships that warrant exploration in future studies. aspects such as financial literacy, credit management skills, and the role of financial education (lusardi & mitchell, 2013; xiao et al., 2022), could provide further insights. understanding these connections could be instrumental in developing more effective strategies for improving financial well-being and resilience among households. in conclusion, while this study contributes significantly to our understanding of financial behavior during economic downturns, it also opens avenues for more comprehensive future research. such research could ultimately aid in the development of targeted policies and interventions to enhance financial stability and preparedness among households facing economic uncertainties. conclusion the labor market disruptions caused by the covid-19 pandemic led to rapid and significant increases in the unemployment rate, mirroring other historical labor market shocks like the great recession. however, our study extends beyond merely recognizing these economic upheavals. it delves into the complex relationship between negative income shocks, perceived lack of financial control, and the subsequent use of stimulus checks for debt payments. our findings indicate a notable positive correlation between a decrease in income and stronger perceived lack of ouyang et al. 17 financial control, which in turn correlates with a higher likelihood of using stimulus checks to manage debt obligations. this study highlights a critical aspect of financial behavior during times of economic distress. when households allocate a substantial portion of their resources to financial commitments, their capacity to handle unforeseen economic shocks diminishes significantly. this situation was evident during the covid-19 pandemic, where many americans found themselves unprepared for such abrupt and severe financial challenges. our results suggest the necessity for improved financial resilience among the population. by fostering greater financial capability and promoting strategies for financial management, households can be better equipped to handle future economic uncertainties and shocks. furthermore, these findings have broader implications for policymaking. they underscore the importance of measures aimed at enhancing financial capability and providing tools for effective financial planning. such initiatives could play a pivotal role in mitigating the adverse effects of future labor market shocks. in conclusion, this study not only sheds light on the behavioral responses to the recent pandemic but also offers valuable insights for preparing more resilient financial strategies to withstand future economic challenges. references abdelrahman, h., & oliveira, l. e. (2023). the rise and fall of pandemic excess savings. frbsf economic letter, 11. amadeo, k. (2021, october 24). stimulus checks: 2020 compared to 2009 and 2008. the balance. https://www.thebalance.com/stimuluschecks-3305750. amromin, g., & smith, p. (2003). what explains early withdrawals from retirement accounts? evidence from a panel of taxpayers. national tax journal, 56(3), 595-612. argento, r., bryant, v. l., & sabelhaus, j. (2015). early withdrawals from retirement accounts during the great recession. contemporary economic policy, 33(1), 1-16. armantier, o., goldman, l., koşar, g., lu, j., pomerantz, r., & van der klaauw, w. (2021). how have households used their stimulus payments and how would they spend the next? liberty street economics. https://libertystreeteconomics.newyorkfe d.org/2021/04/an-update-on-howhouseholds-are-using-stimulus-checks/ baker, s. r. (2015). debt and the consumption response to household income shocks (vol. 2541142). ssrn. baker, s. r., & yannelis, c. (2017). income changes and consumption: evidence from the 2013 federal government shutdown. review of economic dynamics, 23, 99-124. baum, a., fleming, r., & reddy, d. m. (1986). unemployment stress: loss of control, reactance and learned helplessness. social science & medicine, 22(5), 509-516. berger, l. m., & houle, j. n. (2019). rising household debt and children’s socioemotional well-being trajectories. demography, 56(4), 12731301. bierman, a., upenieks, l., glavin, p., & schieman, s. (2021). accumulation of economic hardship and health during the covid-19 pandemic: social causation or selection?. social science & medicine, 275, 113774. bordia, p., hobman, e., jones, e., gallois, c., & callan, v. j. (2004). uncertainty during organizational change: types, consequences, and management strategies. journal of business and psychology, 18, 507-532. borio, c. (2020). the covid-19 economic crisis: dangerously unique. business economics, 55(4), 181-190. brehm, j. w. (1966). a theory of psychological reactance. academic press. https://www.thebalance.com/stimulus-checks-3305750 https://www.thebalance.com/stimulus-checks-3305750 https://libertystreeteconomics.newyorkfed.org/2021/04/an-update-on-how-households-are-using-stimulus-checks/ https://libertystreeteconomics.newyorkfed.org/2021/04/an-update-on-how-households-are-using-stimulus-checks/ https://libertystreeteconomics.newyorkfed.org/2021/04/an-update-on-how-households-are-using-stimulus-checks/ financial services review, 33(1) 18 bricker, j., & thompson, j. (2016). does education loan debt influence household financial distress? an assessment using the 2007–2009 survey of consumer finances panel. contemporary economic policy, 34(4), 660-677. broda, c. & parker, j. (2008). the impact of the 2008 tax rebates on consumer spending: preliminary evidence. northwestern university. http://citeseerx.ist.psu.edu/viewdoc/dow nload?doi=10.1.1.188.9572&rep=rep1& type=pdf browning, m., & crossley, t. f. (2001). the lifecycle model of consumption and saving. journal of economic perspectives, 15(3), 3-22. browning, m., & lusardi, a. (1996). household saving: micro theories and micro facts. journal of economic literature, 34(4), 1797-1855. butrica, b. a., murphy, d. p., & zedlewski, s. r. (2010). how many struggle to get by in retirement? the gerontologist, 50(4), 482-494. caputo, r. k. (2012). patterns and predictors of debt: a panel study, 1985-2008. journal of sociology and social welfare, 39, 7. cobb-clark, d. a., kassenboehmer, s. c., & sinning, m. g. (2016). locus of control and savings. journal of banking & finance, 73, 113-130. coibion, o., gorodnichenko, y., & weber, m. (2020). how did u.s. consumers use their stimulus payments? (nber working paper no. 27693). national bureau of economic research. http://www.nber.org/papers/w27693 colarieti, r., mei, p., & stantcheva, s. (2024). the how and why of household reactions to income shocks (working paper no. w32191). national bureau of economic research. doi 10.3386/w32191 cox, n., ganong, p., noel, p., vavra, j., wong, a., farrell, d., ... & deadman, e. (2020). initial impacts of the pandemic on consumer behavior: evidence from linked income, spending, and savings data. brookings papers on economic activity, 2020(2), 35-82. dalgard, o. s., thapa, s. b., hauff, e., mccubbin, m., & syed, h. r. (2006). immigration, lack of control and psychological distress: findings from the oslo health study. scandinavian journal of psychology, 47(6), 551-558. deaton, a. (2005). franco modigliani and the life cycle theory of consumption. https://dx.doi.org/10.2139/ssrn.686475 dettling, l. j. & lambie-hanson, l. (2021). why is the default rate so low? how economic conditions and public policies have shaped mortgage and auto delinquencies during the covid-19 pandemic. feds notes no. 2021-03-042 https://doi.org/10.17016/23807172.2854. drentea, p., & lavrakas, p. j. (2000). over the limit: the association among health, race and debt. social science & medicine, 50(4), 517-529. dryden, r. p., campbell, a. v., perry, r. p., hamm, j. m., chipperfield, j. g., parker, p. c., & leboe-mcgowan, l. (2023). assisting students in debt to overcome academic setbacks with a cognitivereframing motivation intervention. journal of educational psychology, 115(1), 55. fan, l., & chatterjee, s. (2017). borrowing decisions of households: an examination of the information search process. journal of financial counseling and planning, 28(1), 95-106. https://doi.org/10.1891/10523073.28.1.95 federal reserve bank of new york. (2022, november). quarterly report on household debt and credit (q3 2022). https://www.newyorkfed.org/medialibra ry/interactives/householdcredit/data/pdf/ hhdc_2022q3 http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.188.9572&rep=rep1&type=pdf http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.188.9572&rep=rep1&type=pdf http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.188.9572&rep=rep1&type=pdf http://www.nber.org/papers/w27693 https://dx.doi.org/10.2139/ssrn.686475 https://doi.org/10.17016/2380-7172.2854 https://doi.org/10.17016/2380-7172.2854 https://doi.org/10.17016/2380-7172.2854 https://doi.org/10.1891/1052-3073.28.1.95 https://doi.org/10.1891/1052-3073.28.1.95 https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/hhdc_2022q3 https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/hhdc_2022q3 https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/hhdc_2022q3 https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/hhdc_2022q3 ouyang et al. 19 foltice, b., & ilcin, p. (2019). the influences of locus of control, debt, and framing on retirement contributions. journal of advanced studies in finance, 10(1), 3444. forbes, j., & kara, s. m. (2010). confidence mediates how investment knowledge influences investing self-efficacy. journal of economic psychology, 31(3), 435-443. https://doi.org/10.1016/j.joep.2010.01.0 12 fulford, s., rush, m., & wilson, e. (2021). changes in consumer financial status during the early months of the pandemic: evidence from the second wave of the making ends meet survey. consumer financial protection bureau office of research reports series, (21-2). garner, t. i., safir, a. & schild, j. (2020, june). receipt and use of stimulus payments in the time of the covid-19 pandemic. u.s. bureau of labor statistics. https://www.bls.gov/opub/btn/volume9/pdf/receipt-and-use-of-stimuluspayments-in-the-time-of-the-covid-19pandemic.pdf. garrison, s.t., rampold, s.d., vasquez, k., gillen, m. & baker, l.m. (2022). parents' employment, income, and finances before and during the covid19 pandemic. journal of consumer affairs. 56(1), 276-291. gasiorowska, a. (2014). the relationship between objective and subjective wealth is moderated by financial control and mediated by money anxiety. journal of economic psychology, 43, 64-74. http://dx.doi.org/10.1016/j.joep.2014.04. 007 ghilarducci, t., radpour, s., & webb, a. (2019, spring). new evidence on the effect of economic shocks on retirement plan withdrawals. the journal of retirement. gillers, h., tergesen, a., & scism, l. (2018). a generation of americans is entering old age the least prepared in decades. the wall street journal, 22. glass, d. c., singer, j. e., leonard, h. s., krantz, d., cohen, s., & cummings, h. (1973). perceived control of aversive stimulation and the reduction of stress responses. journal of personality, 41(4), 577-595. goel, a., & rastogi, s. (2023). understanding the impact of borrowers' behavioural and psychological traits on credit default: review and conceptual model. review of behavioral finance, 15(2), 205-223. grable, j. e., park, j. y., & joo, s. h. (2009). explaining financial management behavior for koreans living in the united states. journal of consumer affairs, 43(1), 80-107. grable, j., kwak, e. j., & archuleta, k. (2023). distrust of banks among the unbanked and banked. international journal of bank marketing, 41(6), 1498-1520. grable, j. e., & palmer, l. (2022). introduction to personal finance: beginning your financial journey. john wiley & sons. guvenen, f., pistaferri, l., & violante, g. l. (2022). global trends in income inequality and income dynamics: new insights from grid. quantitative economics, 13(4), 1321-1360. hayes, a. f. (2013). introduction to mediation, moderation, and conditional process analysis. a regression-based approach. new york, ny: the guilford press. heo, w., nobre, l. h. n., grable, j. e., & ruizmenjivar, j. (2016). what role does financial risk tolerance play in mediating investing behavior? journal of financial service professionals, 70(5). ibrahim, m. e., & alqaydi, f. r. (2013). financial literacy, personal financial attitude, and forms of personal debt among residents of the uae. international journal of economics and finance, 5(7), 126-138. https://doi.org/10.1016/j.joep.2010.01.012 https://doi.org/10.1016/j.joep.2010.01.012 https://www.bls.gov/opub/btn/volume-9/pdf/receipt-and-use-of-stimulus-payments-in-the-time-of-the-covid-19-pandemic.pdf https://www.bls.gov/opub/btn/volume-9/pdf/receipt-and-use-of-stimulus-payments-in-the-time-of-the-covid-19-pandemic.pdf https://www.bls.gov/opub/btn/volume-9/pdf/receipt-and-use-of-stimulus-payments-in-the-time-of-the-covid-19-pandemic.pdf https://www.bls.gov/opub/btn/volume-9/pdf/receipt-and-use-of-stimulus-payments-in-the-time-of-the-covid-19-pandemic.pdf http://dx.doi.org/10.1016/j.joep.2014.04.007 http://dx.doi.org/10.1016/j.joep.2014.04.007 financial services review, 33(1) 20 imai, k., jo, b., & stuart, e. a. (2011). commentary: using potential outcomes to understand causal mediation analysis. multivariate behavioral research, 46(5), 842-854. https://doi.org/10.1080/00273171.2011. 606743 immordino, g., jappelli, t., oliviero, t., & zazzaro, a. (2022). fear of covid‐19 contagion and consumption: evidence from a survey of italian households. health economics, 31(3), 496-507. kamakura, w. a. & du, r. y. 2012. how economic contractions and expansions affect expenditure patterns. journal of consumer research 39(2), 229-247. kim, k. t., lee, j., & hanna, s. d. (2019). the effects of financial literacy overconfidence on the mortgage delinquency of u.s. households. journal of consumer affairs, 54(2), 517–540. klapper, l., & lusardi, a. (2020). financial literacy and financial resilience: evidence from around the world. financial management, 49(3), 589-614. klapper, l., lusardi, a., & panos, g. a. (2013). financial literacy and its consequences: evidence from russia during the financial crisis. journal of banking & finance, 37(10), 3904-3923. kleimeier, s., hoffmann, a. o., broihanne, m. h., plotkina, d., & göritz, a. s. (2023). determinants of individuals’ objective and subjective financial fragility during the covid-19 pandemic. journal of banking & finance, 153, 106881. lea, s. e., webley, p., & walker, c. m. (1995). psychological factors in consumer debt: money management, economic socialization, and credit use. journal of economic psychology, 16(4), 681-701. lin, j.t., mottola, g., & lusardi, a. (july, 2022). financial capability in the united states: highlights from the finra foundation national financial capability study. https://finrafoundation.org/sites/finrafou ndation/files/nfcs-report-fifthedition-july-2022.pdf lindgren, h. c. (1980). great expectations: the psychology of money. los atos: william kaufmann inc. liu, y., zhang, y., & chatterjee, s. (2021): financial hardship and depression experienced by pre-retirees during the covid-19 pandemic: the mitigating role of stimulus payments. applied economics letters, 30(3), 391-396 doi: 10.1080/13504851.2021.1989364 lusardi, a. (1998). on the importance of the precautionary saving motive. the american economic review, 88(2), 449453. lusardi, a., & de bassa scheresberg, c. (2016). americans’ troubling financial capabilities: a profile of pre-retirees. public policy & aging report, 26(1), 2329. lusardi, a., & mitchell, o.s. (2013). debt and debt management among older adults. draft michigan retirement research center, ann arbor, mi. lusardi, a., & tufano, p. (2015). debt literacy, financial experiences, and overindebtedness. journal of pension economics and finance, 14(4), 332-368. doi:10.1017/s1474747215000232. marshall, g. l., & tucker-seeley, r. (2018). the association between hardship and selfrated health: does the choice of indicator matter?. annals of epidemiology, 28(7), 462-467. modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. post-keynesian economics, 1, 388-436. netemeyer, r. g., warmath, d., fernandes, d., & lynch jr, j. g. (2018). how am i doing? perceived financial well-being, its potential antecedents, and its relation to overall well-being. journal of consumer research, 45(1), 68-89. ouyang et al. 21 ouyang, c., & hanna, s.d. (2022). the effect of racial/ethnic differences on the financial obligations ratio of renters. financial services review. 30(2022), 165–177. pandemic oversight https://www.pandemicoversight.gov/dat a-interactive-tools/data-stories/threerounds-stimulus-checks-see-how-manywent-out-and-how-much parker, j. a., souleles, n. s., johnson, d. s., & mcclelland, r. (2013). consumer spending and the economic stimulus payments of 2008. american economic review, 103(6), 2530-2553. pattanayak, c. w., rubin, d. b., & zell, e. r. (2011). propensity score methods for creating covariate balance in observational studies. revista española de cardiología (english edition), 64(10), 897-903. peetz, j., & robson, j. (2021). living gig to gig and paycheque to paycheque: how income volatility affects financial decisions. centre for economic policy research. perry, v. g., & morris, m. d. (2005). who is in control? the role of self‐perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39(2), 299-313. polizzi, c., lynn, s. j., & perry, a. (2020). stress and coping in the time of covid-19: pathways to resilience and recovery. clinical neuropsychiatry, 17(2), 59. rubin, d. b. (2005). causal inference using potential outcomes: design, modeling, decisions. journal of the american statistical association, 100(469), 322331. https://doi.org/10.1198/0162145040000 01880 salamanca, n., de grip, a., fouarge, d., & montizaan, r. (2020). locus of control and investment in risky assets. journal of economic behavior & organization, 177, 548-568. shapiro, m.d. & slemrod, j. (2009) did the 2008 tax rebates stimulate spending? american economic review, 99(2), 374– 379. https://doi.org/10.1257/aer.99.2.374 shefrin, h. m., & thaler, r. h. (1988). the behavioral life-cycle hypothesis. economic inquiry 26(4). 609-643. https://doi.org/10.1111/j.14657295.1988.tb01520.x shevlin, m., mcbride, o., murphy, j., miller, j. g., hartman, t. k., levita, l., ... & bentall, r. p. (2020). anxiety, depression, traumatic stress and covid19-related anxiety in the uk general population during the covid-19 pandemic. bjpsych open, 6(6), e125. stata. causal mediation analysis. in stata press of mediate. retrieved dec 10, 2023, from stata.com. sweet, e. (2021). debt-related financial hardship and health. health education & behavior, 48(6), 885-891. tang, t. l. p., luna-arocas, r., & sutarso, t. (2005). from income to pay satisfaction: the love of money and pay equity comparison as mediators and culture (the united states and spain) and gender as moderators. management research: journal of the iberoamerican academy of management, 3(1), 7–26. tharayil, a., & walstad, w. b. (2022). financial literacy and the early withdrawal of funds from retirement accounts. journal of financial counseling and planning, 33(1), 44-55. tucker-seeley, r. d., li, y., subramanian, s. v., & sorensen, g. (2009). financial hardship and mortality among older adults using the 1996–2004 health and retirement study. annals of epidemiology, 19(12), 850-857. united states congress. (2020). coronavirus aid, relief, and economic security act, pub. l. no. 116-136, 134 stat. 281. https://www.congress.gov/bill/116thcongress/house-bill/748/text https://www.pandemicoversight.gov/data-interactive-tools/data-stories/three-rounds-stimulus-checks-see-how-many-went-out-and-how-much https://www.pandemicoversight.gov/data-interactive-tools/data-stories/three-rounds-stimulus-checks-see-how-many-went-out-and-how-much https://www.pandemicoversight.gov/data-interactive-tools/data-stories/three-rounds-stimulus-checks-see-how-many-went-out-and-how-much https://www.pandemicoversight.gov/data-interactive-tools/data-stories/three-rounds-stimulus-checks-see-how-many-went-out-and-how-much https://doi.org/10.1257/aer.99.2.374 https://doi.org/10.1257/aer.99.2.374 https://doi.org/10.1111/j.1465-7295.1988.tb01520.x https://doi.org/10.1111/j.1465-7295.1988.tb01520.x financial services review, 33(1) 22 whitson, j. a., & galinsky, a. d. (2008). lacking control increases illusory pattern perception. science, 322(5898), 115-117. xia, t., wang, z., & li, k. (2014). financial literacy overconfidence and stock market participation. social indicators research, 119(3), 1233–1245. https://doi.org/10.1007/s11205-0130555-9 xiao, j. j., & porto, n. (2022). financial capability and wellbeing of vulnerable consumers. journal of consumer affairs, 56(2), 1004-1018. xiao, j.j., chen, c., chen, f. (2014). consumer financial capability and financial satisfaction. social indicators research, 118(1), 415–432. xu, y., & yao, r. (2022). u.s. household financial vulnerability: prediction analyses in the covid-19 pandemic. journal of financial counseling and planning, 33(2), 228-242. yue, p., gizem korkmaz, a., & zhou, h. (2020). household financial decision making amidst the covid-19 pandemic. emerging markets finance and trade, 56(10), 2363-2377. yuh, y. & hanna, s.d. (2010). which households think they save? journal of consumer affairs, 44, 70–97. zhang, k. (1989). factors affecting financial resources management behaviors. the university of arizona. https://doi.org/10.1007/s11205-013-0555-9 https://doi.org/10.1007/s11205-013-0555-9 https://doi.org/10.1007/s11205-013-0555-9 https://doi.org/10.1007/s11205-013-0555-9 ouyang et al. 23 appendix a: details of independent variables details of selected independent variables are presented in this section. a.1 age age was categorized based on the respondent’s response as 25-34, 35-44, 45-54, 55-64, and 65+. the code was based on the following question: a3a)5 what is your age? a.2 gender gender was categorized based on the respondent’s responses based on the following question: a50a)4 [builder: punch gender from q.a50; if q.a50 = 3, randomly assign to male or female male .............................................................................................................................. 1 female .......................................................................................................................... 2 a.3 ethnicity ethnicity was categorized as non-hispanic white and non-white based on the following question: a4)7 which of the following best describes your race or ethnicity? select all that apply. [code 99 exclusive][builder: note punch 7 is not in order] [m] white or caucasian ....................................................................................................... 1 black or african-american ........................................................................................... 2 hispanic or latino/a ...................................................................................................... 3 asian ............................................................................................................................. 4 native hawaiian or other pacific islander .................................................................... 7 american indian or alaska native ................................................................................ 5 other ............................................................................................................................. 6 a.3 education i categorize the respondent’s highest level of education as less than college, some college, associate degree, bachelor’s degree, and post-graduate degree. the code was based on following questions: a5)12,13 what was the highest level of education that you completed? did not complete high school ...................................................................................... 1 high school graduate – regular high school diploma ................................................... 2 high school graduate – ged or alternative credential ................................................. 3 some college, no degree ............................................................................................... 4 associate’s degree ......................................................................................................... 5 bachelor’s degree .......................................................................................................... 6 post graduate degree ..................................................................................................... 7 prefer not to say .......................................................................................................... 99 a.4 employment employment status of respondents were categorized based on the following question: a9) which of the following best describes your current employment or work status? self-employed ............................................................................................................... 1 work full-time for an employer [if q.am21 = 1 insert: or the military] ................ 2 work part-time for an employer [if q.am21 = 1 insert: or the military] ............... 3 homemaker ................................................................................................................... 4 financial services review, 33(1) 24 full-time student ........................................................................................................... 5 permanently sick, disabled, or unable to work ............................................................. 6 unemployed or temporarily laid off ............................................................................. 7 retired ........................................................................................................................... 8 a.5 income levels respondents’ income was categorized based on the following question: a8)15 what is [if q.a7a = 3 insert: your approximate annual income/ if q.a7a = 1, 2 insert: your household’s approximate annual income], including wages, tips, investment income, public assistance, income from retirement plans, etc.? would you say it is… less than $15,000 ......................................................................................................... 1 at least $15,000 but less than $25,000 ......................................................................... 2 at least $25,000 but less than $35,000 ......................................................................... 3 at least $35,000 but less than $50,000 ......................................................................... 4 at least $50,000 but less than $75,000 ......................................................................... 5 at least $75,000 but less than $100,000 ....................................................................... 6 at least $100,000 but less than $150,000 ..................................................................... 7 at least $150,000 but less than $200,000 ..................................................................... 8 at least $200,000 but less than $300,000 ..................................................................... 9 $300,000 or more ........................................................................................................ 10 a.6 emergency funds whether respondents have emergency fund or not was based on the following question: j5) have you set aside emergency or rainy day funds that would cover your expenses for 3 months, in case of sickness, job loss, economic downturn, or other emergencies? yes ................................................................................................................................. 1 no ................................................................................................................................. 2 a.7 laid off during pandemic whether respondents got laid off during pandemic was based on the following question: j52)27 as a result of the pandemic, were you laid off or furloughed at any time in 2020 or 2021? yes ................................................................................................................................. 1 no/not applicable ......................................................................................................... 2 a.8 subjective financial knowledge score the subjective financial literacy score is based on the following survey questions: m4) on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall financial knowledge? a.9 objective financial knowledge score the objective financial literacy score is based on the number of correct answers based on a series of financial knowledge questions in the 2021 nfcs. the survey questions are as follows: m6) suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow? more than $102 ............................................................................................................. 1 exactly $102 ................................................................................................................. 2 less than $102 .............................................................................................................. 3 # m7) imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. ouyang et al. 25 after 1 year, how much would you be able to buy with the money in this account? more than today ............................................................................................................ 1 exactly the same ........................................................................................................... 2 less than today ............................................................................................................. 3 # m8) if interest rates rise, what will typically happen to bond prices? they will rise ................................................................................................................ 1 they will fall ................................................................................................................. 2 they will stay the same ................................................................................................ 3 there is no relationship between bond prices and the interest rate ............................... 4 m9) a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. true ............................................................................................................................... 1 false .............................................................................................................................. 2 # m10) buying a single company’s stock usually provides a safer return than a stock mutual fund. true ............................................................................................................................... 1 false .............................................................................................................................. 2 m31)65 suppose you owe $1,000 on a loan and the interest rate you are charged is 20% per year compounded annually. if you didn’t pay anything off, at this interest rate, how many years would it take for the amount you owe to double? less than 2 years ........................................................................................................... 1 at least 2 years but less than 5 years ............................................................................ 2 at least 5 years but less than 10 years .......................................................................... 3 at least 10 years ............................................................................................................ 4 # m50)66 which of the following indicates the highest probability of getting a particular disease? [randomize punches 1-3] there is a one-in-twenty chance of getting the disease ................................................ 1 2% of the population will get the disease ..................................................................... 2 25 out of every 1,000 people will get the disease ......................................................... 3 a.10 money left whether respondents have money left was reported based on the following question: j42_1) i have money left over at the end of the month (1-5). a.11 homeownership homeownership was recorded based on the following question: ea_1)43 do you [if q.a7a = 1 insert: or your spouse/ if q.a7a = 2 insert: or your partner] currently own your home? yes ................................................................................................................................. 1 no ................................................................................................................................. 2 a.12 financial strain respondents’ financial strain was based on all of their unpaid medical bills, being late on student loan payments (g35), late on mortgage payments (e15), and being late on credit card payments (f2_4): g35)52 how many times have you been late with a student loan payment in the past 12 months? (if you have more than one student loan, please consider them all.) never, payments are not due on my loans at this time .................................................. 1 never, i have been repaying on time each month ......................................................... 2 once .............................................................................................................................. 3 more than once ............................................................................................................. 4 financial services review, 33(1) 26 e15)47 how many times have you been late with your mortgage payments in the past 12 months? (if you have more than one mortgage on your home(s), please consider them all.) never ............................................................................................................................. 1 once .............................................................................................................................. 2 more than once ............................................................................................................. 3 f2) in the past 12 months, which of the following describes your experience with credit cards? (select an answer for each) f2_4) in some months, i was charged a late fee for late payment a.13 subjective credit record respondents’ self-assessed credit record was coded based on the following question: j32)28 how would you rate your current credit record? very bad ........................................................................................................................ 1 bad ................................................................................................................................ 2 about average ............................................................................................................... 3 good ............................................................................................................................. 4 very good ..................................................................................................................... 5 appendix b: correlation table for objective and subjective financial knowledge correlation table: objective financial knowledge and subjective financial knowledge correlation obj. fin knowledge sub. fin knowledge objective financial knowledge 1 subj. financial knowledge 0.248 1 ouyang et al. 27 appendix c: variance inflation factors appendix table c. variance inflation factor (vif) for the independent variables exogenous (independent) variables vif 1/vif income drop yes 1.46 .685 perceived financial control rarely 2.13 .469 sometimes 2.64 .379 often 2.32 .431 always 2.38 .420 control variables age group 25-34 3.08 .325 35-44 3.22 .311 45-54 3.37 .297 55-64 3.68 .272 65+ 5.51 .181 gender female 1.16 .864 ethnicity non-white 1.08 .922 family status live with partner/spouse 1.37 .729 employment self-employed 1.17 .855 work part-time 1.21 .825 homemaker 1.24 .809 full-time student 1.12 .890 permanently sick 1.35 .743 unemployed 1.30 .768 retired 2.80 .357 education some college 1.58 .634 associate's degree 1.37 .732 bachelor's degree 1.83 .547 post graduate degree 1.54 .648 income levels $15,000 to $25,000 1.98 .505 $25,000 to $35,000 2.18 .459 $35,000 to $50,000 2.71 .368 $50,000 to $75,000 3.47 .288 $75,000 to $100,000 3.15 .318 $100,000 to $150,000 3.19 .314 $150,000 to $200,000 1.75 .572 $200,000 to $300,000 1.19 .839 $300,000 or more 1.09 .913 laid off during pandemic yes 1.33 .751 emergency fund yes 1.36 .571 homeownership yes 1.47 .679 objective financial knowledge 1.36 .737 subjective financial knowledge 1.25 .801 financial strain yes 1.50 .665 subjective credit record bad 3.62 .276 average 4.57 .219 good 5.21 .192 very good 8.61 .116 have money left rarely 2.36 .423 sometimes 3.22 .310 often 3.20 .313 always 4.00 .250 average vif 2.43 pii: s1057-0810(97)90024-x financial services review, 6(2): 141-1.50 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. conversions of mutual savings institutions: do initial returns from these ipos provide investors with windfall profits? julie a.b. cagle gary e. porter we examine initial returns offidly underwritten ipos of converting thrifts for evidence that managers and depositors of conversion-related offers earn significantly greater returns than investors in ipos of otherfinancial institutions. regulators have suggested that new guidelines for conversion from a mutual to a stock thrift are designed to curb “windfall profits” earned by insiders investing in conversion-related ipos. while there are reports of average initial returns of more than 20% for conversion-related ipos, our results suggest that investors earn average initial returns of about 7%, which is not significantly different than returns from ipos of other thrifts and commercial banks, i. introduction conversions of mutual savings and loans to stock organizations are accompanied by initial public offers (ipos) of stock which infuse the institutions with equity capital. average ini tial returns of 24% for 1992 and 29% for 1993 have been reported for conversion-related ipos (barth, brumbaugh, & kleidon, 1994), which is several times the average initial return of 7% for conversions during the mid-1980s (alli, yau, & yung, 1994). the office of thrift supervision (ots) and the federal deposit insurance corporation (fdic) adopted new guidelines for conversions, effective january 1, 1995, intended to reduce the ability of bank executives to earn excessive profits from the conversion process. in partic ular, regulators suggest that insider’s set the offer price of the ipo too low or take a dispro portionate number of shares. the new rules include a requirement that long-term depositors be given first chance to buy the stock of a converting institution and ban the use of “running proxies,” which are obtained from depositors when accounts are opened and julie a.b. cagle l assistant professor of finance, college of business administration, xavier university, cincinnati, oh 45207; e-mail: cagle@xavier.xu.edu. gary e. porter l assistant professor of finance, college of business administration, university of central florida, orlando, fl 32816; e-mail: gary.porter@bus.ucf.edu. 142 financial services review 6(2) 1997 used by managers without authorization. stock option plans or other nonqualified stock benefit plans are also prohibited from being implemented within one year of conversion, unless the plan is fully disclosed in the offering materials. the ability to capture abnormal returns in conversion-related initial offers is of partic ular significance to individual investors because it contrasts sharply with their experience in the typical ipo market. individual investors are usually at a big disadvantage in the ipo market because lucrative offers are preserved for institutional investors, with individual investors receiving a disproportionate share of losing offers (ibbotson, sindelar, 62 ritter, 1994). conversion-related ipos give priority to depositors in the mutual thrift, so individual investors, as depositors, cannot be closed-out of the offer in favor of institutional investors. this study examines the following hypotheses: h,: ipos of converting institutions are not priced significantly different from ipos of other financial institutions. ha: ipos of converting institutions are significantly more underpriced than ipos of other financial institutions. the results of the analysis will also provide important information to depositors and regulators. if depositors have the ability to earn abnormal returns by investing in conver sion-related ipos, it might affect their willingness to approve the conversion plan, as well their decision to invest in the subsequent ipo. the evidence will also provide information for regulators regarding the pricing of the conversion-related ipo as a means of wealth transfer to insiders. the following section contains a discussion of the agency issues associated with con verting mutual organizations to stock organizations and a summary of empirical studies related to conversions. the data and methodology used to analyze ipo underpricing is pre sented, followed by the results of the study. finally, a summary of the study and implica tions for conversion investors are discussed. ii. overview of conversion studies the increased regulatory scrutiny associated with conversion reflects the uncertainty asso ciated with effective control of mutual organizations. first, mutual organizations are insu lated from the market for corporate control because there is no marketable ownership claim. also, unlike publicly-owned institutions, outside board members are generally cho sen by internal managers. as a result, boards of directors may play less of a monitoring role in mutual organizations than they do in public corporations. finally, owners of mutual organizations (the depositors) regularly give insiders perpetual proxies that have limited disclosure requirements. for these reasons, management may have effective control of the mutual organization (kreider, 1972). as a result, managers of mutual organizations may be more effective at, and/or more likely to, expropriate wealth from other claims holders by underpricing conversions. however, conversion to a stock thrift will also subject the man agers of such firms to the market for corporate control at some point in the future, although regulatory restrictions offer protection from hostile takeovers for 3 to 5 years post-conver sion (cordell, macdonald, & wohar, 1993). if conversions are motivated by expropria conversions of mutual savings instihrtions 143 tion, management must believe that the wealth transferred by conversion to a stock company is greater than that would accrue in the future if the firm operated as a mutual organization and management maintained effective control. in addition to temporary insulation from the market for corporate control, conversion related ipos differ in several other ways from typical ipos. the underpricing of an ipo associated with a conversion can take place in the absence of asymmetric information, that is, even if outsiders have full information about the risks of the issuing bank. this under pricing can occur because the thrift’s initial shareholders receive pro-rata claims to the pro ceeds of the sale of stock since no founding owners exist to claim a portion of the net proceeds or the initial net worth as in a typical ipo (masulis, 1987). as a result, some underpricing is predicted as long as the thrift has positive pre-conversion value, and the proceeds are not expected to be invested in unprofitable ventures. conversion-related ipos are also less likely to experience potential conflicts of inter est between insiders and outside investors because insiders and outsiders will buy shares at the same price, though insiders must hold shares for one year after conversion (maksimo vie & unal, 1993). depositors may subscribe to a maximum of 5% of the issue, whereas collective management holdings have a maximum of 15%-25% of the issue (aharony, falk, & linn, 1996). aharony, et al. report average subscription rates of 39% for regular depositors and 5% for management, while dunham (1985) suggests managers and direc tors purchase an average of 20% of all conversion shares. empirical research has addressed possible motivations for conversion of savings and loans, including expropriation as asserted by regulators. masulis (1987) provides evi dence that conversions from the period 1976-1983 were motivated by efficiency rather than expropriation. the results suggest an increase in both management turnover and access to capital. the conversion-related ipos were found to generate average initial returns of 5.61%. similar studies of conversions in the 1980s do not find evidence consis tent with efficiency gains, nor are the results suggestive of expropriation. simons (1992) compares 55 converted new england savings banks to mutual savings banks over the period 1983-1990, and her results suggest that converted institutions have a greater pro pensity to assume risk, while no increase in profitability is found. cordell, et al. (1993) find similar trends. simon’s analysis of insider ownership indicates management and directors are more likely to increase their stake in thrifts as the risk of the assets increase. however, the moti vation for management to increase risk while reducing job stability is not addressed. while managers may benefit from increased returns associated with increased risk through stock ownership, they simultaneously put their human capital at risk. regulator and media reports on conversions suggest that the change to a stock organi zation may provide excessive compensation to managers and trustees through ipo sub scription or stock compensation. while the range for the value of the offering is determined by independent appraisal, the stock is generally offered within a range of 15% above or below appraised value as decided by management and approved by regulators (simons, 1992). because the conversion price tends to be below market value, managers and trustees as subscribers to the ipo have an opportunity to profit from the conversion. this suggests that the greater the difference between the conversion price and perceived market value, the more likely management is to seek conversion. if conversion reflects an attempt by management to expropriate wealth by using their ability to manipulate the offer price, all 144 financial services review 6(2) 1997 else equal, underpricing of converting institutions’ ipos will be greater than underpricing of ipos for other savings institutions. maksimovic and unal (1993) study the post-offer price performance of 287 thrifts converting during 1980-1988. on average, the converted thrift’s ipos are offered at 5% below their initial market value. the choice of offer size by converting institutions was examined to determine the extent to which management and outsider interests are aligned. management can choose between minimizing issue value, thereby lowering their purchase price for a set number of shares, or maximizing issue value, thereby raising the amount of capital and financial slack available to consume perquisites. the results indicate manager and investor interests are aligned regarding the choice of issue size; that is, issue size and underpricing are positively related. the influence of the participation of depositors and management in the offering is examined by aharony, et al. (1996). in a study of 100 conversions between january 1984 and june 1987, they find that following the conversion the value of the converted institu tion is positively related to depositors’ subscriptions. for small levels of managerial sub scription, they find a negative relationship between managerial ownership and the value of the converted institution, but this relationship turns less negative as managerial subscrip tion increases. the average level of underpricing for firms in the sample was 6.5%. they also report that the majority of offerings are generally purchased by outside investors, rather than by depositors or employees of the converting institutions. these studies of ipos of converting institutions provide evidence that they are consis tently underpriced. however, underpricing of ipos is a well-documented phenomenon out side of converting institutions or financial institutions. some theoretical works suggest that the underpricing of ipos is associated with asymmetric information and investors’ con cerns that the decision to issue equity is an attempt to expropriate wealth from outsiders (ibbotson, et al., 1994). empirical studies have found evidence that the underpricing for ipos of financial institutions is related to proxies for asymmetric information. offer size (megginson & weiss, 1991), age of the firm (muscarella & vetsuypens, 1987; barry & brown, 1984; megginson & weiss, 1991) underwriter reputation (carter & manaster, 1990; logue, 1973; mcdonald & fisher, 1972) and the volatility of post-offer returns (ritter, 1984) have all been associated with ipo underpricing. if insiders profit by setting the initial offer price too low, ipos associated with conver sion should be more underpriced than ipos of other financial institutions with similar lev els of asymmetric information. alli, yau, and yung (1994) report that conversion related ipos are priced similarly to ipos of nonfinancial institutions, but conversion-related ipos are significantly more underpriced than ipos of other financial institutions. the average underpricing reported was 7% for converting thrifts and 5% for other financial institutions. the other financial institutions include banks, bank holding companies (bhcs), and other thrifts. they argue that the lower level of underpricing for ipos of other financial institu tions reflects reduced uncertainty as the result of regulation. they further suggest that this regulation effect is offset for converting institutions by a one-time economic gain that may result from conversion. unlike the alli, yau, and yung (1994) study, this study does not include bhcs. bhcs differ from banks in several ways which may influence uncertainty or asymmetric informa tion surrounding initial offerings by these institutions. these include the ability of bhcs to “downstream” borrowed capital as equity for subsidiary banks and engage in nonbanking activities via nonbank subsidiaries. evidence provided by polonchek, slovin, and sushka conversions of mutual savings institutions 145 (1989) suggests that capital standards reduce the underpricing of seasoned bank stocks rel ative to those of nonfinancial firms because leverage restrictions of banks inject noise into the negative information conveyed by the announcement to issue equity. similarly, the ability of a bank subsidiary to receive equity capital from its parent may affect the informa tion conveyed by the announcement of a security issue. also, if regulations reduce the level of asymmetric information, the restrictions of the bhc act and its amendments may have an effect. iii. data and methodology our sample is chosen from the list of fully underwritten depository institution equity ipos reported in the directory of corporate financing, published by investment dealer digest, for the period january 1982 through december 1994. conversion status was determined using the “history” section of each institution in moody’s bank and finance manual. thrift conversions were verified using a list of approved conversions from the ots and the lexis/nexis database. the type of each depository institution was determined using both moody’s bank and finance manual and ward’s business directory. the financial institu tions include savings and loans, savings banks, and commercial banks (sic 6021, 6022, 6035, and 6036 at the time of the initial offer). ideally, our test of the influence of the conversion process on pricing of ipos requires a sample of converting thrifts and a control sample of ipos of other thrifts, that is, thrifts that did not operate previously as a mutual. asymmetric information in convert ing thrifts should be the same as that of other thrifts, unless the conversion process influ ences the dissemination of information. however, due to the small number of initial offers by other thrifts, we expand the control sample to include ipos of commercial banks. commercial banks are similar to thrifts in that they are subject to similar regula tory restrictions and requirements, though regulatory authority may differ. both institu tions also have similar sources of funds, though concentration of their loan portfolios differs. it is not clear how differences in the asset structures of the two types of institu tions would affect the market’s ability to value initial offers of stock. the sample of other financial institutions includes commercial banks, as well as thrifts that did not operate previously as a mutual. our analysis of the initial returns employs two econometric models. model 1, below, uses an indicator variable (convrsn) to distinguish offers of converting thrifts. model 1: ri = (3, + b, conversni + 82 lofsizei + b3 agei + b, reputei + b, std, + ei (1) where ri = (closing price on the first day of trading)/(ofser price) -1, for firm i. conversn = 1, if the ipo is offered by a converting thrift, and 0, otherwise. lofsize = the natural log of the product of offer price and number of shares issued. age = the number of years from establishment of the firm until the ipo. 146 financial services review 6(2) 1997 repute = 1, if the lead underwriter is among the top 15 underwriters for the year of offer in terms of market value, and 0, otherwise. std = the standard deviation of daily returns for the first 20 trading days following the initial trading day (days l-20). offer size, age of the institution, investment banker’s reputation, and post-offer vola tility of returns have been shown to be associated with the initial return of ipos. we use the natural log of the dollar amount of the initial equity offering (lofsize) as reported in the directory ofcorporate financing. the variable age is the difference between the year of the ipo and the year of the institution’s establishment, reported in moody’s bunk and finance manual. investment banker reputation (repute) is measured according to the ordinal scale developed in carter and manaster (1990). their procedure uses the under writer listing position in the tombstones to rank investment bankers. our methodology is a variation of the adaptation of the carter and manaster scale presented in megginson and weiss (1991). using the directory of corporate financing, the top 15 underwriters, in terms of total market value of offerings, were identified for each of the thirteen sample years. for each year, the top 15 underwriters are deemed to be of high reputation, while all others are not. the lead investment banker for each ipo and the offer price and issue size were collected from the directory of corporate financing. since the directory of corpo rate financing discontinued reporting the lead underwriter in 1991, moody’s bank and finance manual was used to obtain this information for subsequent years. volatility of the post-offer returns is captured by the variable std. std is the standard deviation of the returns for the 20 days following the first day of trading, days l-20, where day 0 is the first day of trading. the closing price on the first day of trading and the subsequent 20 days of returns were obtained from the crsp data tapes. the unadjusted percentage change between the offer price and the closing price on the first day of trading is used to analyze the degree of under pricing of the ipos. beatty and ritter (1986) show that a market adjustment of the ipo ini tial return is less than 0.1% and has an insignificant impact on the calculation of the degree of underpricing. model 2 accounts for the high correlation between the variables for reputation and offer size by eliminating the offer size variable. estimation of the two models is repeated for the broader control sample of commercial banks and thrifts. iv. results table 1 provides summary statistics for the sample of ipos. converting thrift institutions are substantially older than the banks at the time of the initial offering. commercial banks are the largest firms in the sample in terms of average total assets, while thrifts are larger when comparing medians. all subsamples of depository institutions experience average underpricing that is different from zero at the 1% level of significance. the level of under pricing for financial institutions is similar to that reported in previous empirical studies (aharony, et al., 1996; alli, et al., 1994; maksimovic & unal, 1993). the mean return for the converting thrift ipos in the sample, 7.0%, is not significantly different from the mean return of ipos of other financial institutions, 5.7% (t = 0.5 1). conversions of mutual savings institutions 147 table 1 summary statistics for sample of 152 initial public offerings of common equity by commercial banks, and state and federal chartered thrifts made between 1982 and 1994 thrifts and commercial banks n = 15.2 thrifts n = 119 commercictl converting banks thrifts n = 33 n = 107 commercial banks and other thrifts n = 45 age (ye=) mean 38 44 15 41 17 median 33 50 4 53 4 assets (millions) mean $1,940 $1,181 $4,681 $1,157 $3,805 median $424 $486 $274 $486 $215 offer size (millions) mean $20.3 i $21.02 $17.71 $21.96 $16.38 median $9.00 $9.53 $6.60 $9.53 $7.10 under pricing (day = 0) mean* 6.6% 6.9% 5.4% 7.0% 5.7% (7.47)* (7.10)* (2.62)* (7.02)* (3.10)% median 2.2% 2.3% 1.5% 2.6% 1.8% notes: t-test for underpricing is ho: mean of (closing price first day of tradmg offer price) / offer pnce = 0. * t-statistic for test: mean = 0, in parentheses. * significant at 1% level, one-tail test. the difference m mean returns between thrift and commercial bank ipos = 1.5%. t-statistic = 0.49. the difference m mean returns between convening thrift ipos and ipos of other financial insfitufmn = 1.3%. t-statistic = 0.51. table 2 provides results of tests of differences in underpricing between converting thrifts and other financial institutions when offer size, age of the firm, underwriter reputa tion, and post-offer standard deviation of returns are included as explanatory variables. panel a provides results for thrift institutions only. converting thrift ipos are found not to be priced differently from ipos of other thrifts (conversn = -0.001, t = -0.17). the proxy for reputation is significant at the 10% level (repute = -0.032, t = 1.50), indicating that ipos of thrifts underwritten by bankers with greater market share provide lower initial returns. also, the variable for post-issue volatility is significant at the 1% level (std = 3.44, t = 3.32), suggesting that higher initial returns are associated with relatively greater risk. model 2 in panel a does not include lofsize due to the positive correlation between the size of the equity offer and the reputation variable (correlation coefficient = 0.5 1, p-value = 0.0001). the reputation variable is significant at the 1% level (repute = -0.048, t = 2.56). the omission of lofsize does not affect the other results. panel b contains the results of our analysis comparing initial returns of converting thrifts to those of other thrifts and commercial banks. a test of residuals reveals the pres ence of heteroskedasticity in model 1. in the presence of heteroskedasticity, the coeffi cient estimates are accurate, but the test statistics are not. a correction procedure suggested by white (1980) adjusts the test statistic by producing a consistent estimate of the covariance matrix. the asymptotic t-statistics obtained from this method are reported. 148 financial services review 6(2) 1997 table 2 panel a. thrifts regression results for a sample of 119 thrift ipos from 1982 through 1994, where initial return is the dependent variable conversn repute (n a&r2 intercept (ii = 107) lofxze age =60) std prub f model 1 0.049 -0.001 -0.013 0.000 -0.032 3.44 0.1037 (i .33) (-0.17) (-1.43) (0.22) (-lso)* * * (3.32)* 0.0037 model 2 0.029 -0.007 0.000 -0.048 3.51 0.0959 (0.85) (-0.21) (0.04) (-2.56)* (3.37)* 0.0037 panel b. banks and thrifts regression results for a sample of 152 thrift and commercial bank ipos from 1982 through 1994, where initial return is the dependent variable converse repute adj r2 intercept {rz = 107) lofsize age fn = 69) std prob>f ~ ..--..... model 1 0.043 0.0001 -0.017 o.octll -0.010 3.30 0.1001 (1.74) * * * (0.004) (1.42) (.036) (0.39) (3.29) * 0.0010 model 2 0.012 0.001 0.000 -0.029 3.41 0.0826 (0.58) (0.06) (0.12) (-1.62) * * * (3.87) * 0.0022 niritps: * sqnificant at 1% level. one-tail test. ’ * * significant at 10% level, one-tail cat. rerun = (price at close of first day of tr~din~/~fer pricei -1 0. crsp provides the closing price nn day r or the average of the bid/ask prices convrsn = i, if the thrift ipo was preceded by the c~nvetxon of the thrift from mutual ownership to stock ownership, 0 otherwise. lofsize = the natural log of the product of crtyrr prices and number of shares offered. age = year of 1po year of e%ablishment. repute = i, if the lead underwriter is among: the top i5 in marker value listed m the directory of corporate fmance for year of issue, 0 otherwise. std is the stindard deviation of the first 20 returns relative to the day of trading. regression models were tested for heternskeda&tlcrty. panel 8, model 1 exhibits heteroskedarticify. aymp totic i-stafst~cs are reported besed on white (1980). the adjustment rcsulfs in test statstics whzh are posttive. regardless of rhc sign of the coefficient. the results in panel b are consistent with those in panel a. conversion is not found to affect initial returns, while underwriter reputation and post-offer volatility are both associ ated with underpricing. after controlling for factors associated with asymmetric information, we find no evi dence that conversion significantly affects ipo unde~~cing. this suggests that the act of conversion does not significantly influence the initial returns of ipos of financial institu tions. this result contrasts with alli, yau, and yung (1994) who report conversion-related ipos are significantly more underpriced than ipos of other financial institutions. vi. conclusion the results of the study indicate that initial stock offers of converting thrifts are not priced differently than ipos of other financial institutions. from 1982 through 1994 fully under written, conversion-related ipos were underpriced, on average, approximately 7%. the evidence is inconsistent with claims that abnormal or windfall gains accrue to investors in conversions of mutual savings institutions 149 conversion-related ipos. thus, there is no support for the assertion that managers of con verting financial institutions gain at the expense of depositors by setting ipo offer prices too low. conversion-related ipos do provide the opportunity for depositors to participate in an ipo without competition from institutional investors. the results suggest that, in evaluating conversion-related ipos, depositors and regulators should focus on issues other than offer price. acknowledgments: we wish to thank cyndi mcdonald, james wansley, charlene sullivan, and the anonymous referees for their suggestions. thanks also to john callahan for his assistance. references aharony, j., falk, h., & lin, c. (1996). changes in ownership structure and the value of the firm: the case for mutual-to-stock converting thrift institutions. journal of corporate finance, 2(3), 301-316. alli, k., yau, j., & yung, k. (1994). the underpricing of ipos of financial institutions. journal of business finance and accounting, 21, 1023-1030. barth, j.r., brumbaugh, b., jr., & kleidon, a.w. (1994). “windfall” gains in mutual-to-stock con versions of thrift institutions? challenge, july-august, 43-49. barry, c.b., & brown, s.j. (1984). differential information and the small firm effect. journal of financial economics, 12(2), 283-294. beatty, r.p., & ritter, j.r. (1986). investment banking, reputation, and the underpricing of initial public offerings. journal of financial economics, 15,213-232. carter, r., & manaster, s. (1990). initial public offerings and underwriter reputation. journal of finance, 45, 1045-1067. cordell, l., macdonald, g., & wohar, m. (1993). corporate ownership and the thrift crisis. journal of law and economics, 36,719-755. dunham, c. (1985). mutual to stock conversion by thrifts: implications for soundness. new england economic review, january/february, 3 l-45. ibbotson, r.g., sindelar, j.l., & ritter, j.r. (1994). the market’s problems with the pricing of ini tial public offerings. journal of applied corporate finance, 6,66-74. kreider, g.p. (1972). who owns the mutt&? proposals for reform of membership rights in mutual insurance and banking companies. university of cincinnati law review, 41(2), 275-295. logue, d. (1973). on the pricing of unseasoned equity issues: 1965-1969. journal of financial and quantitative analysis, 8, 91102. maksimovic, v., & unal, h. (1993). issue size choice and the underpricing in thrift mutual-to-stock conversions. journal of finance, 158, 1659-1691. masulis, r.w. (1987). changes in ownership structure. journal of financial economics, l&29-59. mcdonald, j., & fisher, a. (1972). new issues stock price behavior. journal of finance, 27, 97 102. megginson, w., & weiss, k. (1991). venture capitalist certification of initial public offerings. jour nal of finance, 46, 879-903. muscarella, c.j., & vetsuypens, m.r. (1989). a simple test of baron’s model of ipo underpricing. journal of financial economics, 24(l), 125-136. polonchek, j., slovin, m., & sushka, m. (1989). valuation effects of commercial bank securities offerings. journal of banking and finance, 13,443-461. ritter, j.r. (1984). the hot issue market of 1980. journal of business, .57(2), 215-240. 150 financial services review 6(2) 1997 simons, k. (1992). mutual to stock conversions by mutual savings banks: where has all the money gone? new england economic review, march-april, 4553. white. (1980). a heteroskedasticity-consistent covariance matrix estimator and a direct test for het eroskedasticity. econometrica, 48,817-838. pii: 1057-0810(95)90004-7 financial services review, 4(2): 81-95 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. commission-motivated trading patterns of brokers across the production month earl d. benson david s. rystrom greg t. smersh the intramonthpattem of broker commission earnings is examinedfor a sample of one hundred brokers from a national brokerage firm. it is hypothesized that the structure of broker commissions leads to distortions in trading. the evidence shows that in the last five days of the production month, more than one-fourth of the brokers earned a significantly higher proportion of their monthly commissions than would be expected if trading were uniform across the month. this suggests that the structure of the commission system may lead some brokers to encourage individual investors to unnecessarily trade securities near the end of the production month to boost their commission income. i. introduction the trading practices of individual brokers have not been investigated in the academic finance literature. an abundance of articles has been written on the pricing efficiency (or external efficiency) of financial markets and many articles have also focused on the operational efficiency (or internal efficiency) of these markets, by examining broker com missions and dealer spreads. however, little has been written about the trading behavior of brokers and dealers, and how their behavior may independently influence the trading patterns and the costs of trading for individual investors. it is possible that the trading behavior of brokers may have a significant impact on the welfare of the clients for whom they make trades. the trading behavior could additionally affect the profitability of the firms for whom the brokers are employed and the trading efficiency (or “inside” efficiency) of the financial markets in which they operate. the purpose of the paper is to investigate whether individual brokers have an intramonth pattern of stock trading that is influenced by the structure of the commission system of earl d. benson and david s. rystrom l depaitment of finance, marketing, and decision sciences, western washington university, bellingham, wa 98225. greg t. smersh l department of finance, insurance, and real estate, college of business administration, university of florida, gainesville, fl 32611. 82 financial services review 4(2) 1995 compensation. to do this, we first examine the intramonth trading activity of individual brokers of a national brokerage firm, and second compare this to trading volume in themarket as a whole. for the firm examined in this study, the empirical evidence suggests that more than one-fourth of the individual brokers studied do exhibit an unusual trading pattern that seems to be influenced by the commission structure. this leads to the questions of whether the trading activities of these brokers are always in the clients’ best interests and whether the trading may sometimes be motivated primarily by a broker’s desire to earn commission income. the latter would suggest that these brokers at times may be encouraging their clients to engage in unnecessary trading. for the market as a whole, the evidence suggests that trading patterns are not uniform across the month, whether looking at total nyse trading volume or odd-lot trading volume. we then examine the intramonth, broker-motivated trading relative to the identified intra month pattern in nyse trading. when this is done, about one-fourth of the brokers continue to exhibit a pattern of commission-induced trading. section ii examines the motivations for individuals to trade securities and presents the results of previous research on trading volume. the methodology and data are explained in section iii. the empirical findings for a sample of brokers from a national brokerage firm are included in section iv, and an analysis of nyse total volume and nyse odd-lot trading volume is in section v. section vi builds on the empirical findings of section iv by comparing broker trading with nyse trading volume. section vii provides a summary and the implications of this research. ii. the motivation for trading and trading volume the focus in this paper is on within-month stock trading. to gain a better understanding of individual trading and trading volume, we first review previous research regarding the motivations for trading securities and the empirical evidence on trading volume. this review illuminates what type of distribution of securities trading across a given month may be expected. several authors have attempted to identify the motives for securities trading. treynor and wagner (1990) classify traders according to four potential motives: value-based traders, information-based traders, pseudo-information-based traders, and liquidity-based traders. admati and pfleiderer (1988) develop a model of transacting that classifies traders as being either informed traders, discretionary liquidity traders, or nondiscretionary liquidity traders. lakonishok and smidt (1986) discuss tax-motivated traders. these models have been used to attempt to explain calendar price anomalies. for instance, the admati and pfleiderer model may help to explain intraday trading patterns. foster and viswanathan (1990) used a model that includes informed traders and liquidity traders to help explain interday trading. the lakonishok and smidt article focuses on turn-of-the-year effects. these articles have been primarily concerned with identifying trading motivations that might create unusual patterns of returns. in a market where the flow of information and the liquidity needs of investors are randomly distributed, the volume of buy and sell orders should not be concentrated around any particular time within the month because buy and sell opportunities present themselves randomly across time. in such a world, the intramonth trading volume across the entire commission-motivated trading patterns of brokers 83 market may be expected to be uniformly distributed across the month. likewise, the dollar volume of an individual broker’s trades also may be expected to conform roughly to a uniform distribution. on the other hand, if the information flows or liquidity needs are concentrated in a pattern that regularly recurs across months, one would expect something other than a uniform distribution of intramonth trading. thompson, olsen, and dietrich (1987) show that season alities exist in the flow of information. in their study, information flows varied across days of the week and across months of the year, but they did not examine intramonth information flows. however, their results suggest that some seasonalities in the flow of information across the month could exist as well. if information flows are not uniform across the month, the pattern of intramonth trading volume would not be expected to be uniform. in this case, the pattern of individual broker trades across days of the month may be expected to be similar to the intramonth pattern of trading in general.’ studies of trading volume have focused on the relationship between equity returns and trading volume. for example, karpoff (1987) reports that previous studies suggest that a rise in trading volume is associated with large absolute price changes, that the correlation between volume and price changes is positive, and that volume is higher when prices increase than when prices decrease. gallant, rossi, and tauchen (1992) confirm the above findings and find that large price movements (both positive and negative) are followed by high volume. pettengill and jordan (1988) find that there are seasonal patterns in trading volume similar to return seasonalities. further, they report that their tests suggest that there is a strong positive relationship between trading volume and return and that causality tests indicate that higher volume leads to higher returns (especially for smaller firms). little work has been done regarding trading volume alone, other than the construction of models to explain volume, such as karpoff (1986). in the literature regarding seasonal return anomalies, some work has been done to test if there are any systematic return patterns associated with the months of the year. several authors (e.g., jones, pearce, & wilson, 1987; roll, 1983) report a turn-of-the-year (or january) effect. ritter (1988) provides an explanation of the turn-of-the-year effect for small stocks. ariel (1987, 1988) reports that within-month returns are positive for the first two weeks of the month and negative for the second two weeks. lakonishok and smidt (1988) find a strong turn-of-the-month effect for equities around the end of each month. pettengill and jordan (1988) find higher returns in the first and last few trading days of each month. they, also, find higher returns in the first two weeks of the month and negative returns in the third week. they say the negative third week returns are related to the expiration of options. none of the above models or empirical work have suggested that brokers’ motivations and activities have an effect on trading volume. to a great extent, the models have been formed to attempt to explain well-documented empirical return anomalies. our hypothesis, in contrast, identifies the broker commission system as a motivating source of broker trading activities within each month and tests if individual broker trading follows this pattern. iii. methodology we hypothesize that trading activity may be biased; that is, some brokers may have an uncharacteristically large portion of their trades concentrated within a few days of the month. a4 financial services review 4(2) 1995 this trading pattern may occur because of the incentive com~nsation system used in most brokerage firms. brokers are paid on a monthly commission basis; as the end of the “production month” nears, they are under pressure to meet commission goals that they have personally set or that have been set for them by superiors.’ this may lead some brokers to attempt to increase their level of trading activity near the end of the production month, particularly those whose co~ission production is lagging far behind expectations (or quotas). most national brokerage firms pay their individual brokers once a month; this pay is based largely, if not entirely, on commissions generated. all commission trades that “settle” by the last day of the month are normally used to determine the commission income paid to the broker in the following month.3 a significant portion of the commission dollars earned by most brokers comes from stock trades. because stock trades allow five trading days for settlement, only those commissions from stock trades that occur on or before five trading days prior to the last day of the month contribute to the next paycheck to be received. brokers who are under pressure to meet their monthly commission goals must trade stocks on or before thenext-to-last friday in agiven month ifthey wish to boost that month’scommission production from stock trading. commissions earned on other trades that must be settled more promptly may be generated during the final week of the month and still boost the following month’s paycheck. our hypothesis is that this commission payment system leads some individual brokers to a pattern of trading in which a high proportion of their stock trading for clients occurs in the few days on or before the end of the production month. while many brokers are very diligent about serving their customers’ needs on a day-to-day basis t~oughout the month, some may not be as diligent. the less diligent brokers may tend to be somewhat lazy in finding new accounts and reviewing existing clients’ portfolios during the first few weeks of a production month but will suddenly become very active on behalf of their clients during the final week of the production month in order to generate “sufficient” commission revenue.4 further, this tendency is much more likely to show up with brokers who service accounts for individual investors as opposed to institutions. institutions rely much less on brokers’ advice, whereas, many individuals trade primarily on the advice of their brokers. the focus here is on stock trading, rather than all trading, because the pattern of trading we are hypothesizing is more likely to be evident in the trading of stocks for which a broker may feed “hot tips” to clients to generate trading. the trading of other assets such as bonds or mutual funds is less likely to exhibit this pattern because their trading is normally not as lucrative to brokers and is not done as much on the basis of “hot tips.” to test this hypothesis, we obtained daily brokerage commission data on the stock trades for a random sample of brokers from one of the nation’s leading brokerage firms. the data include the daily brokerage commission production earned on stock transactions for each of 100 randomly selected brokers5 the data cover a two-year period from the last week of august 1990 through august 1992. in addition, daily data were collected on nyse total trading volume and on odd-lot purchases and sales for the same two-year period using barron’s and the standard & poor’s daily stock price record. the data are grouped into 24 “production months” over the two-year period. for each month, day 1 is identified as being the last trading day of the production month (i.e., five trading days before the last trading day of the calendar month). day 2 is the trading day before day 1; day 3 is two trading days before day 1, and so forth back to day 19. this provided us with a 19-day period for the analysis, day 1 back through day 19, (most months had more than 19 days in a production month, but the inclusion of more days in some months would make the data analysis particularly difficult when averaging acrossdaysof the month,) the hypothesis is tested in two parts. in section iv, we assume that info~ation flows arc random across the month and that one may expect the intramonth pattern of broker trades to be uniform. in section v, we assume that there may be seasonalities in information flows, that these seasonalities are reflected in nyse trading volume, and that we may expect the pattern of broker trading to be similar to the pattern reflected in nyse trading. iv. ~~idual broker trading activity to perform the empirical tests, we first calculate the proportion of total commissions that each of the 100 brokers in the sample earned in each of the 19 days of each “production month.” for each day of the month, the co~issions earned by each individual broker are expressed as a proportion of that broker’s total commissions. for example, pij,k is equal to the proportion for broker i on dayj during month k: pjj,k = dcijk/tcik, (1) where dcek is the daily co~ission for broker i on day j during month k, and rc,, is the total commission for broker i during month k. next, the average proportion for each of the nineteen days @ij is the average for broker i on day j) is calculated for each broker by averaging across the 24 months of data, using: these averages appear in columns 2 through 8 of panel a of table 1 for seven of the 100 brokers in the sample. (these seven brokers are the ones who, on average, earned the highest proportions of their monthly commissions in the last week of the production month.) the final column in panel a provides the combined data for all loo brokers in the sample by showing the average proportion of monthly com~ssions earned across the sampl6 if trading is uniformly distributed across the 19-day commission month, one may expect about 1119th (or 5.26%) of the brokers’ total 19-day commissions to be earned each trading day. thus, the null h~othesis in this section is that the pro~~ions are equal to 5.26%. the alternative hypothesis is that near the end of the commission month, these proportions are greater than 5.26%, as brokers try to meet their commission goals. a r-test is used to test if the numbers in panel a of table 1 are significantly different from 5.26%. for example, we test whether the 19.25% for broker 1 on trading day 1 is signi~cantly different from 5.26%. the sample standard deviation, across the 24 day is for the commission proportions of broker 1, is 23.82%. the f-value for this test is 2.88, where t is found by: t = (19.25 5.26)/(23.821fi) = 2.88. (3) using a one-tailed test, this t-value suggests that the commission proportion of 19.25% is signi~cantly different from the hypothesize value of 5.26 at the 1% level of confidence. financial services review 4(2) 1995 table 1 percentage of total monthly commissions earned by individual brokers during the last week of the production month numbers in the columns are the average percentage of total monthly commissions earned by that broker on the trading day (panel a) or the group of trading days (panel b) listed in the first column. trading day i is the last day of the production month, trading day 2 is the next-to-last day. and so forth. average of trading loo day broker i broker 2 broker 3 broker 4 broker 5 broker 6 broker 7 brokers panel a: single trading days 1 19.25* 8.06 17.32* 14.39* 1.97 11.36** 6.94 6.16* 2 6.36 12.45** 6.60 10.05** 8.06 5.52 7.93 5.72** 3 9.36 9.59 4.60 7.68 19.50** 1024 6.62 5.66 4 10.64** 4.95 7.04 5.76 8.15 5.88 9.76 5.52 5 3.73 11.88 9.61** 4.80 4.01 7.34 9.16** 5.55 panel b: grouped trading days last few trading days 2 days 25.61* 20.51 23.93* 24.44* 10.03 16.87* 14.87 11.88* 3 days 34.96* 30.10** 28.52* 32.12* 29.53** 27.72* 21.49 17.54* 4 days 45.60* 35.05** 35.57* 37.88** 37.68** 33.60* 31.25** 23.06* 5 days 49.33* 46.93** 45.18* 42.69* 41.69** 40.94* 40.41** 28.69* notes: * significant at the i % level, using a one-tailed test. ** significant at the 5% level, using a one-tailed test. overall, for broker 1, the day 1 and day 4 proportions are significantly higher than the expected values. each of the other brokers shown in panel a of table 1 has daily averages that are higher than expected. similar calculations can be made using the average proportions across all 100 brokers, as shown in the final column of panel a. the table shows that an average of 6.16% of brokers’ commission income was earned on day 1. this 6.16% for day 1 is significantly different (at the 1% level) from the hypothesized value 5.26%, with a r-value of 3.42. the day 2 value of 5.72% is also significantly higher for the firm (but at the 5% level of significance). panel b of table 1 presents data by grouping the commission proportions into the last x trading days of the commission month. this data is used to test if brokers earned a higher than expected portion of their commissions in the last x days of the commission month. the null hypothesis is that for the last two, three, four, and live days of the commission month, the broker or firm should earn two-, three-, four-, and five-nineteenths, respectively, of total 19-day commissions. these hypothesized proportions are 10.53%, 15.79%, 21.05%, and 26.32%, respectively. the alternative hypothesis is that they tend to earn more than these proportions. the evidence shows that these brokers did earn a significantly higher proportion of their commission income in the last week of the production month. for example, panel b of table commiwion-motivated trading patterns of brokers 87 table 2 number of brokers categorized by proportion of commissions in the last week of the production month this table shows the number of brokers out of the total sample of 100 who fall into each category in the first column. the first column lists. by spercentage-point intervals, the possible percentage of commissions that each of the sample brokers could have earned, on average over the sample period, during the last week of the 19day production month proportion of commissions brokers wirh significantly higher earned in the last week entire sample of 100 brokers trading 15 to 19.99% 1 20 to 24.99% 14 25 to 29.99% 40 30 to 34.99% 25 9 35 to 39.99% 13 13 40 to 44.99% 4 4 45 to 49.99% 3 -j total 100 29 1 shows that broker 1 earned, on average, 25.61%, 34.96%, 45.60%, and 49.33% of the total 19-day commission income during the last 2, 3, 4, and 5 days, respectively, of the production month. these are all significantly higher than their expected values, with t-values of 2.88, 3.09, 3.85, and 4.69, respectively. further, most of the last x trading day averages for brokers 2 through 7 are significantly higher than the expected values, with all the five-day averages being statistically significant. finally, the 100 broker averages for the last 2, 3, 4, and 5 trading days (from the final column) are 11.88%, 17.54%, 23.06%, and 28.69%, respectively, all of which are significantly higher (at the 1% level) than the expected values. table 2 provides summary data on all 100 brokers in the sample for the last 5 days of the production month. it shows that 29 of the 100 brokers had a “significantly” higher percentage of commission earnings in the last week compared to the 26.32% expected value. (seventy-three of the 100 brokers had an average percentage higher than 26.32%, but only 29 were “significantly” higher in a statistical sense.) the evidence from tables 1 and 2 is strong enough that we may conclude that many brokers earned significantly higher proportions of their commission income in the last few days before the end of the production month. further, the brokers as a group earned higher than expected commissions during this period. overall, more than one-fourth of the brokers examined show a tendency toward a higher level of trading in the last week of the production month. seven of the brokers (those shown in table 1) earned more than 40% of their commission income, on average, in this last week. figure 1 shows graphically the average intramonth trading of brokers 1 through 6 from table 1, one broker earned nearly 50% of total monthly commission income in this five-day period. the data for the average of the 100 brokers show that, on average, the brokers earned nearly 29% of total 19-day commis sions during the last five trading days. 88 financial services review 4(2) i995 broker 1 daily commission earnings (versus the expected 5.26316%) % 20 .c_ e i w” 15.______---_-----____-____---- 0-n * * 4’ j ’ * ” i : “’ :q 19 'if '15 '13 'i? ' cj 1 ti a 3 trading days (1 = last day of month) broker 2 daily commission earnings (versusihe expected 5.26316%) !j& 14 7 j 12 .____________________-_-_ ; io-----___---------__-_-_ _ ‘z ,g a._-_-_ _______________ -- i __--___________-_--_-- ; 4.__ 6-;$px *._ . -. ..i ___ _i_ _ -1-1 .*. --_--- 5 2 .-__--__ __----_---_ ---_-- 2 0 iq if 15 15 11 trading days (1 = last day of month) broker 3 daily commission earnings (versus the expected 5.26316%) g 18 '; 16 ______----_---_____--__----~ z 14 .e 12 .e 10 e 0 53 g: $0 t _-____-___----_____------- / +_----_----____-----__--____,_ t __-__-------------_----__ _ n i 19 '17 '15 '13 11 g 7 5 3 i trading days (1 = last day of month) figure i. daily commission earnings of brokers 1 through 6 commission-motivated trading patterns of brokers 89 broker 4 daily commission earnings (versus the expected 5.26316%) 19 'if '15 '13 'l? 3 trading days (1 q last day of month) broker 5 daily commission earnings (versus the expected 5.26316%) % 20 .r trading days (1 = last day of month) broker 6 daily commission earnings (versus the expected 5.26316%) u) 12 trading days (1 = last day of month) figure 1. (continued) 90 financialservicesreview 4(2) 1995 v. nyse total volume and nyse odd-lot trading volume we now turn to the general stock market to see if there are any market-wide patterns of intramonth trading and, in particular, to see if trading increases toward the end of the production month. both nyse trading volume and odd-lot trading volume are examined for nonuniform tendencies. in the market in general, we must examine trading volume rather than commission income as was done for individual brokers-since commission data is not available for the market. total nyse trading volume is examined first. our focus is on the intramonth pattern of this trading. second, intramonth trading by individuals is examined by looking at odd-lot trading-aproxy for individual trading. lakonishok and maberly (1990) suggest that odd-lot trading may serve as a proxy for individual trading since few institutions engage in odd-lot trading. odd-lot “purchases” should be particularly instructive because institutions generally do not purchase odd lots. however, odd-lot “sales” may not be as good a guide to individual trading activity because of a significant volume of institutional sales. though institutions do not purchase odd lots, they must sell odd lots that they acquire through stock splits, stock dividends, and conversions of convertible securities. the market-wide data are presented in table 3. the nyse volume data suggest that the volume of trading is approximately even across the month, but there does appear to be a tendency for volume to be higher near the end of the commission month. statistical tests suggest that approximately 1/19th of the trading volume occurs on each of the 19 trading days in the commission month. for example, panel a shows that only the day 3 proportion of 5.66% is significantly greater than the expected 5.26% (or 1/19th) at the 1% level of significance, having a t-value of 2.84. (the day 5 value of 5.55 is significantly different only at about the 10% level of significance.) none of the other daily nyse volume proportions are significantly different from 5.26%. however, looking at the last x trading days data in panel b of table 3, all the percentages are above those of a uniform distribution (10.53%, 15.79%, 21.05%, and 26.32%), and two of the four are statistically significant at the 5% level. in the last week of each production month, 27.33% of the nyse monthly trading volume took place. this proportion is significantly higher than the 26.32% expected value (the r-value of the test is 1.92). overall, the nyse volume data provide evidence the general trading volume is higher than expected near the end of the production month. this higher level of trading could be due to increased broker activity near month-end, to higher information flows near month-end, or to some other unexplained factor. the nyse odd-lot trading data also provide evidence of uneven intramonth trading. panel a of table 3 shows that the proportion of daily odd-lot purchases and sales is greater than 5.26% for each of the last five days of the production month. however, only two of these are significantly different from 5.26%. looking at the last five trading day proportions in panel b, both the odd-lot purchase and odd-lot sale proportions are significantly higher (at the 1% and 5% levels of significance, respectively) than the expected value of 26.32. the odd-lot purchase proportions are somewhat higher than those for odd-lot sales and support the suggestion that odd-lot purchases better reflect the level of individual trading than do odd-lot sales. the odd-lot purchase results shown in table 3 also support the hypothesis that higher trading around the end of the production month may have occurred in brokerage firms other than the single firm examined in this paper. commission-motivated trading patterns of brokers 91 table 3 percentage of total monthly trading in the stock market during the last week of the production month numbers in the columns are the average percentage of total monthly trading in that market on the trading day (panel a) or the group of trading &ys (panel b) listed in the first column. trading day 1 is the last a’ay of the production month, trading by 2 is the next-to-last trading day, and so forth. trading day nyse volume odd-lot purchases odd-lot sales panel a: single trading days 1 5.39 5.53 5.35 2 5.48 5.44 5.49 3 5.66* 5.74% 5.50 4 5.24 5.64 5.32 5 5.55 5.52 5.51 panel b: grouped trading days last few trading days 2 days 10.86 10.97 10.84 3 days 16.53** 16.72* 16.34 4 days 21.78 22.36* 21.65 5 days 27.33** 27.88* 27.16** notes: *significant at the 1% level, using a one-tailed test. ** significant at the 5% level, using a one-tailed test. vi. broker trading activity versus nyse trading volume the data in table 3 indicate a slight increase in general market trading at the end of the production month. one explanation for this intramonth seasonality in overall trading could be that it is caused by the commission-motivated trading we have identified. this possibility seems unlikely, however, since our data reflect transactions made by individual customers of brokerage firms, which are only a small proportion of overall trading. (trading by individuals for 1993 was less than 20% of total nyse volume.) another explanation for intramonth seasonality of overall volume might be that there is a pattern of intramonth seasonality in information, although, as discussed above, we know of no published research that substantiates this conjecture. there may also be other exoge nous, unknown factors that contribute to the volume seasonality. regardless of the underly ing cause, we will now treat the seasonality in the overall volume as the null hypothesis against which to test our observed pattern of broker trading. if we find that the observed pattern of broker trading is significantly different from the pattern of overall trading, we can conclude that the broker trading pattern is not caused by the exogenous factors affecting overall trading, such as possible information seasonality. therefore, instead of using 1/19th (or 5.26) as the expected value (as was done in section iii), we now use the actual nyse volume percentage for that day, as shown in the second column of panel a in table 3. (we use the nyse trading volume data as a reflection of the trading patterns expected of brokers, rather than the odd-lot trading percentages, because the nyse trading may better reflect the many factors that affect securities trading patterns.) the f-test shown in equation (3) (for broker 1 on trading day 1) would become: t = (19.25 5.39) / (23.82 / fi) = 2.85, 92 financial services review 4(2) 1995 where 5.39 is the nyse day 1 trading volume proportion. to test the significance of the proportion 6.16 for the average of 100 brokers for trading day 1 in table 1, we can use the test: t = (6.16 5.39) / (1.29 / a) = 2.93, rather than: r=(6.16-5.26)/(1.29/m)=3.42. the new tests of significance for the 100 brokers for the last 2,3,4, and 5 trading days would be: i = (11.88 10.86)/( 1.93/a) = 2.60, t = (17.54 16.53)/(2.78/m) = 1.78, t = (23.06 -21.78)/(3.41/a) = 1.84, and t = (28.69 27.33)/(4.15/m) = 1.61, respectively. these t-values are significant at the l%, 5%, 4%, and 7% levels, respectively. overall, the statistical significance of the values in table 1 remains very high even when the nyse volume percentages are used as the expected values in the t-tests. while 29 brokers had significantly higher trading during the final week, as shown in table 2, the number falls to 24 when the nyse trading volume percentages are used as the expected values. thus, nearly one-fourth of the entire sample of brokers still shows significantly higher trading in the last week of the production month when this more restrictive test is used.’ therefore, even when it is assumed that the general level of trading is not uniform across the month-as a result of uneven information flows or other unexplained factors-commission-motivated trading seems to be taking place. vii. sukfmry and implications this paper provides evidence that some individual brokers who work for a national brokerage firm earn an unusually high proportion of their stock commission income in the last week of the production month. the last five days of trading are characterized by an unexpectedly high volume of trading by several of the individual brokers and for the average of all brokers examined. this evidence suggests that a large number of the brokers in our sample may be engaging in “commission-motivated trading” in order to boost their commis sion income near the end of the production month. one is led to wonder how widespread this practice may be. in our limited sample, we find that about one-fourth of the brokers traded stocks for their clients more actively than expected during the last week of the production month. seven of the 100 brokers showed particularly active trading, with more than 40% of their monthly commissions being earned in the last week of trading. the broker behavior that has been observed in this paper by looking at a single national brokerage firm may be very widespread. because most national brokerage firms have a commission payment system that is similar to that of the firm in this study, other brokers may be behaving very much like the 29 brokers shown in table 2. if about one-fourth of the brokers in one brokerage firm are encouraging individual investors to engage in what may be viewed as “unnecessary” trading near the end of the pr~uction month, how many more brokers across the nation are engaging in the same practices as they try to meet their commission quotas? our discussions with brokers across the country suggest that this is widespread and is a natural occurrence, given the commission system and the tendency of many individuals to procrastinate.’ because we have looked at a sample from only one national brokerage firm, it is impossible for us to draw general conclusions. however, the implications of the study are clear, and only further research can suggest how widespread this pattern of commission motivated trading may be. the findings in this paper have implications for brokerage firms, their clients, and financial markets. for brokerage firms, these trading patterns could affect their profitability to the extent that rather than having a relatively smooth flow of orders across the month, they have a more uneven flow. the brokerage firms may also be concerned about those ciients who may be pressured into unnecessary trades by brokers who arc trying to meet their quotas at the end of the month. in light of the findings, brokerage firms may wish to change the frequency of their evaluation of broker commission production from monthly to weekly or biweekly. the implications are clear for individual investors (and the clients of other providers of financial services where the sales force is paid on a commission basis). the closer it gets to the end of the production month, the more wary investors should be about the advice they receive from their broker. while the broker trading patterns identified in this paper cannot be defined as “churning,” the behavior leads to similar results-that is, a high level of unnecessary trading done to generate higher commissions for brokers.’ the clients of brokers who have this tendency may be un~owing ~ctims.” for financial markets, this research may also be instructive. many financial products are sold using a commission-paid sales force. if this system of pay distorts the trading pattern of financial assets or leads to more costly trading, it has implications for the internal efficiency of the financial markets and how these markets may best serve the needs of both the suppliers and demanders of capital. the regulators of these markets may have a particular interest in “protecting” unwary investors from unnecessary trading and in lessening distor tions in the pattern of trading. notes 1. it was suggested to the authors that a large segment of wage earners may be paid at the end of the month and that the flow of funds into institutions such as pension funds may be higher at month-end. while this may he true, funds do not immediately flow through the hands of investors directly into the stock market. investors temporarily place funds into checking accounts, money funds, and other idle balances, and then move the funds into stocks at varying intervals over the coming month(s). further, the date of pay from one employer to another is not unifo~ across the month, with a variety of pay schedules being foliowed. the flow of investable funds to institutional investors may be more lumpy across the month, but we know of no regular monthly trading patterns that arise from this flow. 94 financlal services review 4(2) 1995 2. we define a “production month” to be the period of time during which commissions earned will contribute toward the next paycheck. for stock trades, the production month typically ends one week prior to the last trading day of the month. 3. some firms use the last friday in the month rather than the last day of the month as the date that trades must “settle” in order to have the commissions included in the next paycheck. 4. the production of sufficient commission revenue is necessary for brokers to earn their expected income level and for most of them to keep their jobs. a brokerage fii’s regional or national office generally defines what level of commission production is “sufficient.” those brokers who repeatedly fall short of this “sufficient” level will lose their jobs. 5. the data provided by the brokerage firm includes stock trades for a random sample of about 125 brokers. the sample was reduced to 100 brokers after taking out those brokers who traded stock infrequently or did not trade stock for more than a few months during the sample period. 6. an alternative way to construct the empirical tests in this paper is to look at average broker dollar commissions across days of the month compared to the average daily dollar commission, rather than looking at the percentage of commissions earned each day compared to some expected percentage. we did not use this alternative method because the potential for bias is greater when using actual dollars due to the fact that dollar commissions of brokers could be trending up or down over time. by stating each day’s commission as a percentage of that month’s total commission, we eliminate any influence from the other 23 months in the sample. empirical tests were conducted using dollar commissions rather than percentages. these tests confirmed the results reported in tables 1 and 2 by showing that about one-fourth of the brokers earned significantly higher commissions in the last days of the production month compared to the average daily dollar commission that they earned across the month. 7. in addition to the individual broker comparisons, the odd-lot trading averages in table 3 may be compared to the nyse volume percentages. when this is done, the odd-lot trading averages are not significantly different from the nyse percentages. 8. a preliminary study conducted on 1986 and 1987 data from a single office of a national firm showed similar results to those reported here. that study showed that four of eight brokers traded more heavily than expected in the last week of the production month. one broker (who was reported to us by a colleague as engaging in a “feeding frenzy” in the last two days of production month) earned 30% of monthly commissions in the last two days and 53% of commissions in the last week of the production month, on average, over the two-year period. 9. “churning,” according to the sec and court cases, is the excessive trading in a particular customer’s account (usually a discretionary account) for the purpose of generating commission income. 10. instead of churning a particular client’s account, this paper suggests many brokers are “stirring up” unnecessary activity in many different accounts near the end of the production month. (one dictionary defines “churning” as a “violent stirring.“) perhaps we need to invent a term for this behavior to distinguish between intensive (one client) and extensive (multi-client) “churning.” references admati, a. & pfleiderer, p. (1988). a theory of intraday patterns: volume and price variability. the review of financial studies, i, 3-40. ariel, r. (1987). a monthly effect in stock returns. journal of financial economics, 18, 161-74. ariel, r. (1988). evidence on intra-month seasonality in stock returns. in e. dimson (ed.), stock market anomalies (pp. 117-138). cambridge, uk: cambridge university press. foster, f., & viswanathan, s. (1990). a theory of interday variations in volume, variance, and trading costs in securities markets. the review of financial studies, 3,593-624. commission-motivated trading patterns of brokers 95 gallant, a., rossi, p., & tauchen, g. (1992). stock prices and volume. the review offinancialdtudies, 5.199-242. jones, c., pearce, d., & wilson, j. (1987). can tax-loss selling explain the january effect? a note. the journal of finance, 42,453-461. karpoff, j. (1986). a theory of trading volume. the journal offinance, 41, 1069-1087. karpoff, j. (1987). the relation between price changes and trading volume: a survey. journal of financial and quantitative analysis, 22, 109-l 26. lakonishok, j., & maberly, e. (1990). the weekend effect: trading patterns of individual and institutional investors. the journal of finance, 45.231-243. lakonishok, j., & smidt, s. (1986). volume for winners and losers: taxation and other motives for stock trading. the journal of finance, 41,951-974. lakonishok, j., & smidt, s. (1988). are seasonal anomalies real? a ninety-year perspective. the review of financial studies, 1,403-425. pettengill, g., & jordan, b. (1988). a comprehensive examination of volume effects and seasonality in daily security returns. the journal of financial research, 11,57-70. ritter, j. (1988). the buying and selling behavior of individual investors at the turn of the year. the journal of finance, 43,701-717. roll, r. (1983). vas ist das? the turn-of-the-year effect and the return premia of small firms. journal of portfolio management, 9(2), 18-28. thompson, r., olsen, c., & dietrich, r. (1987). attributes of news about firms: an analysis of firm-specific news reported in the wall street journal index. journal ofaccounting research, 25,245-274. treynor, j., & wagner, w. (1990). implementation of portfolio building: execution. in j. maginn & d. tuttle (eds.), managing investment portfolios: a dynamic process (chap. 12, pp. 12/1 12/50). boston, ma: warren, gorham and lamont. pii: s1057-0810(97)90031-7 financial services review, 6(1): 41-52 issn: 1057-0810 copyright © 1997 by jai press inc, all rights of reproduction in any form xeserved. financial planning and college saving recommendations: let's set things straight thomas h. eyssell continuing increases in the cost o f higher education, along with an ever-changing financial aid environment, suggest that financial planning is more important than ever for those seeking to send a child to college. one commonly used aid to the financial planner is the "college savings table, " which is ubiquitous in both the popular press, as well as the literature geared toward financial planners. although they are intended to simplify the planning process, some of these tables may lead to misallocation o f family resources. the tables generally purport to show how much money parents must save monthly to fund four years of college at a specified future date. we demonstrate that these tables often employ flawed methodology. upon correction (and given reasonable assumptions about investor behavior, growth rates in tuition costs, and investment yields), the monthly savings necessary to fund a given level of college expenses can be substantially less than those reported. additionally, published tables typically provide the planner with a limited range o f investment yield assumptions, suggesting a narrow range of porrfolio possibilities. we provide a series of tables which allow the financial planner to estimate required savings for various combinations of investment yields and tuition growth rates. "there are 135 billionaires on the [forbes 400] list. the others have kids in college."- entertainment weekly, october 11, 1996. i. introduction a primary concern of those planning family finances is the ability to fund a college education. recent articles in both the popular and financial press, as well as in the aca demic literature, suggest that it is more important than ever to put in place an appropri ate financial plan. the confluence of rising tuition costs, adverse changes in the amount and nature of student financial aid, and the perception of an increasing inability dr. thomas el eysseil • school of business, university of missouri-st. louis, 8001 natural bridge road, st. louis, mo 63121; e-mail: cl803@umslvma.umsledu. 42 financial services review 6(1) 1997 of college graduates to find suitable employment undoubtedly cause many to question their willingness or ability to provide a college education for their children. unfortunately, financial planners may unwittingly exacerbate this problem by relying on published tables to guide the family saving/investment process. these tables typically purport to show how much money parents must save monthly to fund four years of college at some future date, and which invariably suggest values that would place a significant financial swain on the average family's budget. we demonstrate that these tables often employ flawed methodology. upon correction (and given reasonable assumptions about investor behavior, growth rates in tuition costs, and investment yields), it is shown that the periodic savings necessary to fund a given level of college expenses can be substantially less than those reported. h. the citrrent state of educational economics the field of educational economics is complex and largely beyond the scope of this paper; however, we identify three variables which highlight the importance of careful financial planning for the aspiring student--continued increases in tuition and fees, changes in the nature and availability of financial aid, and uncertainty about the returns to educational expenditure. given the amount of press the topic has received, it is, perhaps, not surprising that the rate of increase in student-borne college costs continues to outstrip that of the consumer price index (cpi). a recent issue of the wall street journal noted that "[t]he average tuition at a private university's four-year program.., in 1993-1994 was 10 times" that for the 1964-1965 school year. and, according to the college board, college tuition and fees for the 1995-1996 academic year "climbed about 5% for the fourth year in a row." (the wall street journal, september 29, 1996, p. a2) further, recent history suggests that little relief is in sight. between 1975 and 1995, the annual increase in the higher education price index (hepi) has been as low as 2.9 percent (1995) and as high as 10.7 percent (1981), with an average annual increase of 5.8 percent. in short, the rate of increase in tuition costs as measured by hepi has risen substantially faster than the overall inflation rate over the last two decades. the implication for the financial planner is clear: rising education costs cannot be ignored, and must be carefully considered in any calculation. the financial aid picture is less clear. on one hand, the college board reports that total amount of financial aid distributed to students reached $50 billion in 1995, a real increase of approximately 4 percent over the previous year. however, a number of factors bear adversely on this picture. first, the dominant form of financial aid is increasingly shifting from outright grants to interest-bearing loans. fenske and barberini suggest that this may reflect the fact that "[1loan programs require fewer federal dollars than do grant programs to provide the same dollar support to students" (1994, p. 502). (the full impact of this shift on the saving pro cess is open to debate; feldstein, 1992, suggests that scholarship rules may provide a dis incentive to those planning to save for higher education.) second, as repayment burdens have increased, so have the level of defaults, as well as the length of the repayment schedules. the higher education amendments act of 1986 has had the effect of lengthening the maximum amortization period on many stu financial planning and college savings 43 dent loans from 10 years to 25 years. and while this may make monthly payments easier to handle for many, it also suggests that the former college student will continue to be paying on college debt well into middle age. it should undoubtedly trouble the financial planner that few have attempted to evaluate the "impact of one generation not being able to retire its own student loan indebtedness before its sons and daughters enter college" (fenske & barberini, 1994, p. 503). a final variable which suggests the need for careful financial planning is the uncer tainty about the returns to investment in education. traditional cost-benefit analysis sug gests that one planning a substantial investment of resources must weigh the outlay against the expected returns. and uncertainty about the magnitudes of those returns increases the importance of careful planning of the investment. while it is traditionally assumed that the substantial financial (and nonfinancial) ben efits accrue to more highly educated individuals, the combination of accelerating tuition costs and shrinking job markets in some fields has contributed to the perception that the expense of higher education may not be "worth it." while it is beyond the scope of this paper to tackle this complex issue, we do note two important factors. first, the (financial) returns to education are unlikely to be constant over time. colin and hughes (1994) esti mate internal rates of return (irrs) on college expenditures over the period 1969-1985 and conclude that the returns on investment declined from the late 1960s through the mid 1970s, increased temporarily, then levelled off in the mid-1980s. it should be noted that some have argued that the mere possession of a degree per se is valuable. hecker (1992) provides evidence on this perceived value of a college degree in terms of employability and earnings potential. the works of arrow (1973), heywood (1994), spence (1974), and wolpin (1977) suggest the existence of a different type of "sheepskin effect"--the mere possession of a college degree may serve as a signal of one's future productivity to potential employers. in any event, one could reasonably infer that the prudent planner will attempt to relate the cost of education to its expected benefits, and incorporate those values into the financial plan being prepared. hi. current practice given the complexities of this financial planning problem, as well as the fact that it impacts millions of people each year, it is no surprise that "one size fits all" tables designed to help one plan for meeting college expenses are ubiquitous. among other places, they have appeared in such publications as money magazine, usa today, various newspapers, reports distributed by brokerage finns, at various sites on the world wide web, as well as in professional publications whose audiences include professional financial planners. tables 1 and 2 are examples prepared by the american institute of certified public accountants (aicpa) and the college board, respectively. consider table 1. assuming one has a five year-old child, college costs $10,000 today and is expected to increase at 7 percent annually for the foreseeable future, the total cost of four years of college is predicted to be $106,996, according to column two. (this value rep resents the sum, at matriculation, of the four future values of the assumed initial cost. that is, $106,996 = $10,000(1.07) 13 + $10,000(1.07) 14 + $10,000(1.07)15 + $10,000(1.07)16.) columns three, four, and five indicate the deposits necessary to accumulate this amount 44 financial services review 6(1) 1997 table 1 projected cost and required funding for four years of college projected fouryear required funding child's age college cost single payment annual monthly 1 $140,250 $37,905 $3,848 $323 2 131,074 38,259 4,002 336 3 122,499 38,617 4,177 352 4 114,485 38,978 4,378 369 5 106,996 39,342 4,609 389 6 99,996 39,710 4,876 413 7 93,454 40,081 5,199 441 8 87,340 40,456 5,582 474 9 81,627 40,834 6,052 515 10 76,286 41,215 6,641 566 11 71"296 41,600 7,398 632 12 66,632 41,989 8,410 719 13 62,272 42,382 9,828 842 14 58,199 42,778 11,959 1,026 15 54,391 43,177 15,513 1,333 16 50,833 43,581 22,629 1,947 17 47,505 43,988 43,988 3,791 source: aicpa personal financial planning manual. note: the exhibit assumes a 7% inflation rate for college costs, an 8% after-tax return on investments, and current college costs of $10,000 annually. given (a) an eight percent after-tax return on funds invested, and (b) investment frequencies of one lump sum, annual payments, or monthly payments. according to the table, one would need to set aside $39,342 today to accumulate the necessary funds, or make thirteen annual deposits of $4,609 (assuming deposits begin one year from today), or make 156 ( = 13 x 12) monthly deposits of $389 (assuming deposits begin one month from today). sim ilarly, the table distributed by the college board compares the expected costs and required savings between public and private institutions (table 2). given (a) the child's current age, co) the current cost of four years of school, and (c) estimates of the growth rate of college costs and investment yields, the table indicates the monthly savings necessary to finance the child's college education. this table assumes that college costs will rise at a 6 percent annual rate and that invested funds will yield 8 percent per annum. column 2 indicates the cost of four years of college from three to seventeen years from today; that is, it is useful for the child who will enter college at age eighteen and who is now as old as fifteen or as young as one. the columnar values indicate that the cost of four years of college at a public university will be $65,231 by the time a child who is five years old today enters school. (by extrapolation, we find that those who prepared this table used $6,991 as the assumed current cost of one year at a public university.) financial planning and college sm,ings 45 table 2 monthly saving required to fund a college education publicf college required monthly private college amount to save age of child (4 years) saving (4 years) monthly 1 $82,353 $191 $180,444 $418 2 77,691 201 170,230 440 3 73,294 212 160,594 464 4 69,145 224 151,504 492 5 65,231 239 142,928 524 6 61,539 256 134,834 561 7 58,056 276 127,206 604 8 54,769 299 120,005 656 9 51,669 328 113,213 719 10 48,745 364 106,804 798 i 1 45,985 410 100,759 899 12 43,382 471 95,055 1,033 13 40,927 557 89,675 1,220 14 38,610 685 84,599 1,501 15 36,425 899 79,810 1,969 source: college board. note: the exhibit assumes a 6% inflation rate for college costs, an 8% after-tax return on investments, and cunent college costs of $6,991 annually. funding the indicated $65,231 expenditure will, according to the table, require monthly savings of $239, assuming that one can invest to earn 8 percent per annum. the penalty for failing to plan ahead is apparent when one considers the indicated monthly savings require ment for a child who is now fifteen years old and for which no funds have previously been set aside: $899 per month. few family budgets can absorb this expenditure without strain. a closer examination of table 2 reveals some rather questionable (but commonly employed) implicit assumptions: (a) that parents desire to have the entire amount needed for four years of schooling available at the time the child first enters school, co) that no fur ther deposits/'mveslments will be made while the child is in school; and (c) that if the entire amount is on hand at the be~nning of the freshman year, funds not used immediately will earn no yield over the remaining four years. (table 1 employs similar assumptions.) it will be shown that the combined effect of these unrealistic assumptions is to substantially over state the required savings figures. iv. setting things straight each of the three assumptions above is restrictive; taken together, they can artificially inflate the required monthly savings figure by a sizable amount. consider the following 46 financial services review 6(1) 1997 adjustments. first, we suggest that what is actually needed at the time of matriculation is enough money to meet the costs of the coming year, rather than for the entire college expe rience. (for the sake of simplicity, we have assumed annual tuition payments in our calcu lations. adjusting for semi-annual or quarterly tuition payments does not materially affect the conclusions of this paper.) at the second enrollment date, what is needed is enough money to cover the cost of the second year, and so on. the implied requirement to accumulate a lump sum upon the entry date inflates the required period saving/investment figure. second, by assuming that no deposits or investments will be made while the child is in school, thirty-six monthly savings opportunities are eliminated, which further increases the required monthly savings. finally, financial theory suggests that rational investors will act to ensure that unused funds will be invested to earn the going rate of return on investments consistent with the risk tolerance of the investor. the assumption that excess funds are held idle is inconsistent with investor rationality and adds to the magnitude of the required savings figures. we suggest that a more realistic scenario is as follows. assume that equal monthly deposits will be made beginning in one month and continuing until the child's last tuition payment is due; that is, just before s/he enters the fourth year of college. in other words, savings/investment inflows continue during the first three years of college, while cash out flows occur at the time of first enrollment, then one, two, and three years thereafter. the assumption that parents continue to make deposits while the child is in school is debatable, given that other college-related expenses often arise during this period. to the extent, how ever, that additional funding is necessary, we suggest that one is better off attempting to incorporate this into the planning process (perhaps by factoring up the "initial college cost" figure), and planning for the necessary deposits accordingly. additionally, empirical evi dence on family spending and savings patterns indicates that families are increasingly likely to fund college in a "pay-as-you-go" manner (baum, 1994). assume also that any funds not used immediately will continue to earn the assumed investment rate. under these more realistic assumptions, (and retaining the 6 and 8 percent inflation and investment rates, respectively, for comparison purposes) the required monthly savings figures fall dramatically. for example, the five-year old's parents need only make monthly deposits of $190 rather than $239 (a reduction of approximately one fifth), and the parents of the 15-year-old need only make monthly payments of $445--less than half the amount indicated in table 2. in other words, the assumptions made to con struct the table cause the indicated saving values to be substantially overstated. v. generalizing the model of course, we are still operating under the twin assumptions of 8 percent after-tax yields and a 6 percent average annual growth in college expenses. clearly, these assumptions are unlikely to be universally appropriate for all planners, all portfolios, and all educational institutions. as noted earlier, annual increases in the higher education price index (hepi) have varied widely over the recent past. the returns on financial assets have also been volatile over the last two decades. since 1975 the annual returns on long-term u.s. treasury bonds have ranged from-7.8 percent financial planning and college savings 47 to 40.4 percent. and annual returns on common stocks (as proxied by the standard and poors 500 index) have ranged from approximately -25 percent to 45 percent over the same period. (ibbotson & sinquefield, 1995) if nothing else, the historical variation in financial asset returns underscores both the need for understanding the risk and return characteristics of different asset classes, as well the need to construct a college savings portfolio consistent with one's funding requirements, risk tolerance, and overall financial situation. in sum, what is needed is a computational method which gives one the ability to take his/her own estimates of current college costs, the growth rate in those costs, and the expected rate of return on invested funds, and use these parameters to obtain a monthly savings figure most appropriate to the situation at hand. our approach follows a two-step procedure. first, we construct a table of compound ing factors which allows us to obtain an estimate of future college costs given a child's cur rent age and the expected annual rate of increase in college costs. we then generate a series of tables which indicate the required monthly savings figures under various age, inflation, and investment return scenarios. the values in table 3 are compounding factors with which one can easily determine the total nominal cost of four years of college for children currently aged 0 to 18 years, assuming annual cost increases ranging from 0 to 10 percent. to use the table, simply multiply the current annual cost at the target institution by the factor corresponding to the child's current age and estimated inflation rate. for example, assume that mr. and mrs. jones, the parents of a 5-year-old and a 15-year-old, know that anystate u. currently costs $6,991 annually. further, they believe that costs at this institu tion will increase at a 6 percent annual rate for the foreseeable future. consistent with the previous example, the table indicates that, at the time of enrollment, the cost of four years of school for the 15-year-old will total $36,425 (= $6,991 x 5.2102). similarly, the cost of the younger child's education will total $65,231 ($6,991 x 9.3307). the factors in table 3 are obtained by summing future value interest factors appropri ate for each tuition date. that is, each table factor equals " ~ + 3 fvlf~, r "c where tau = 18 minus the child's current age. as expected, the computed values mirror those found in table 2. however, the real utility of this table is that one can determine immediately, and with a minimum of compu tation, analogous values for several realistic combinations of current cost, tuition inflation rate, and student age. in the second step we provide the means by which one can determine the monthly sav ings requirement, given more realistic assumptions about savings behavior, and a wide range of possible investment yields and growth rates of college costs. consider the planner attempting to construct portfolios for clients with varying levels of risk tolerance. those with little tolerance may feel more comfortable investing in inter mediate or long-term certificates of deposit. those with somewhat greater willingness and ability to tolerate risk may prefer portfolios which are weighted heavily to medium-grade corporate bonds and/or common stocks. the values in tables 4 through 6 below are pay ment factors associated with expected investment yields of four, eight, and twelve percent, 48 financial servicf.~ review 6(1) 1996 o l~ i ii!ii financial plam6ng and college ~ 49 table 4 payment factors for various yield and growth rate scenarios assumed investment yield = 4% tuition growth rates year5 age until entry 2% 4% 6% 8% 10% 18 0 0.114612 0.11797 0.121415 0.124978 0.128571 17 1 0.085905 0.090155 0.094572 0.099161 0.103926 16 2 0.068672 0.073483 0.078565 0.083931 0.089593 15 3 0.057175 0.06238 0.067978 0.073991 0.080445 14 4 0.048957 0 . 0 5 4 4 6 1 0.060489 0.067082 0.074284 13 5 0.042788 0.048532 0.05494 0.062078 0.070015 12 6 0.037985 0.043928 0.050685 0.058351 0.067031 11 7 0.034137 0.040253 0.047338 0.055526 0.064966 10 8 0.030986 0.037254 0.044653 0.053364 0.063594 9 9 0.028356 0.03476 0.042465 0.051707 0.06276 8 10 0.026127 0.032656 0.040662 0.050446 0.062363 7 11 0.024214 0.030858 0.039162 0.049501 0.062329 6 12 0.022552 0.029304 0.037905 0.048817 0.062606 5 13 0.021096 0.027949 0.036848 0.048351 0.063156 4 14 0.019809 0.026758 0.035956 0.048071 0.063953 3 15 0.018662 0.025704 0.035203 0.047953 0.064977 2 16 0.017634 0.024764 0.034569 0.047976 0.066213 1 17 0.016707 0.023922 0.034035 0.048127 0.07651 0 18 0.015866 0.023163 0.033589 0.048393 0.069284 respectively, which many would consider reasonable expected returns for the portfolios described above. to compute a required monthly savings figure, multiply the current annual cost at the target institution by the factor in the appropriate table. for example, table 5 contains fac tors for an 8 percent investment yield, along with college cost growth rates ranging from 2 to 10 percent. assume that college costs are expected to increase at a 6 percent rate over the foresee able future. the indicated monthly savings requirement for our hypothetical 5-year-old wishing to attend a public university thirteen years from today is $189.52 (= .027109 x $6,991). the analogous value for the 15-year-old from our previous example is sa.a.5.28 (= $6,991 x .063694). these values confirm those noted previously. finally, it should be noted that since the monthly savings requirement is a function of the relatively easily obtained current annual cost figure, obtaining the monthly savings fig ure is not particularly difficult. additionally, the percentage difference between the base case value and the revised value is not affected by the magnitude of the current cost figure. 50 financial services review 6(1) 1997 table 5 payment factors for various yield and growth rate scenarios assnnie.d investment yield = 8% tuition growth rates years age until entry 2% 4% 6% 8% 10% 18 0 0.114826 0.118076 0.121409 0.124826 0.128328 17 1 0.084253 0.088336 0.092576 0.096977 0.101545 16 2 0.065906 0.070455 0.075257 0.080322 0.085663 15 3 0.053675 0.058505 0.063694 0.069263 0.075236 14 4 0.044939 0.049943 0.055418 0.061401 0.067931 13 5 0.038388 0.0435 0.049197 0.055537 0.062581 12 6 0.033296 0.038469 0.044344 0.051003 0.058537 11 7 0.029225 0.034428 0.040449 0.047401 0.055409 10 8 0.025898 0.031106 0.037249 0.~.a.a. 75 0.052952 9 9 0.023129 0.028325 0 . 0 3 4 5 7 1 0.042056 0.050999 8 10 0.020789 0.025959 0.032293 0.040026 0.049437 7 11 0.018789 0.023921 0.03033 0.038302 0.048183 6 12 0.017059 0.022145 0.028618 0.036821 0.047179 5 13 0.01555 0.020582 0.027109 0.035539 0.046378 4 14 0.014223 0.019195 0.025768 0 .034 .419 0.045748 3 15 0.013048 0.017955 0.024567 0.033433 0.045261 2 16 0.012002 0.016839 0.023483 0.032561 0.04.4897 1 17 0.011065 0.015828 0.022499 0.031784 0.044638 0 18 0.010222 0.014909 0.021599 0.03109 0.04.~a. 71 vi. further applications of the model it is relatively straightforward to show that the reported savings figures in table 2 are over stated for every age category. we have already seen that the modified monthly savings requirements are substantially lower than the two "base case" values. but what about the remaining values? returning to table 2, the revised monthly savings amounts range from $157 for the child who is currently one-year-old (versus the $191 base figurema 17.6 per cent decrease), to the $445 (versus $899) for the fifteen-year-old noted earlier. we believe that the tables presented in this paper will be useful in the portfolio plan ning process. given that historical yields on different financial instruments vary in relation to their risk, another interesting application of tables 4 through 6 is to use them to compare the monthly savings necessary to accomplish our goal, given the yields on different invest ment vehicles. at the time of this writing, passbook savings accounts yield between 4 and 5 percent, investment-grade intermediate-term corporate bonds are priced to yield between 7 and 8 percent, and the long-term return on the average common stock is just over 12 per financial planning and college savings 51 table 6 payment factors for various yield and growth rate scenarios assumed investment yield 12% tuition growth rates years age until entry 2% 4% 6% 8% 10% 18 0 0.11513 0 . 1 1 8 2 7 5 0.121499 0.124803 0.128188 17 1 0.08267 0 . 0 8 6 5 4 9 0.090617 0.094838 0.099214 16 2 0.063179 0 . 0 6 7 4 7 5 0.072006 0.076781 0.081812 15 3 0.050263 0 . 0 5 4 7 3 3 0.059532 0.064677 0.070191 14 4 0.041082 0 . 0 4 5 6 1 3 0.050566 0.055973 0.06187 13 5 0.034238 0 . 0 3 8 7 6 0.043795 0.049393 0.055607 12 6 0.028955 0 . 0 3 3 4 2 1 0.038489 0.044228 0.050714 11 7 0.024765 0 . 0 2 9 1 4 5 0 . 0 3 4 2 1 0.040053 0.046778 10 8 0.021371 0 . 0 2 5 6 4 5 0.03068 0.036598 0.043534 9 9 0.018576 0 . 0 2 2 7 2 8 0.027714 0.033683 0.040809 8 10 0.016242 0 . 0 2 0 2 6 2 0.025182 0.031183 0.03848 7 11 0.014271 0 . 0 1 8 1 5 2 0.022993 0 . 0 2 9 0 1 0.036461 6 12 0.01259 0 . 0 1 6 3 2 8 0 . 0 2 1 0 8 1 0.027099 0.03469 5 13 0.011145 0 . 0 1 4 7 3 8 0.019394 0.0254 0.033118 4 14 0.009895 0 . 0 1 3 3 4 1 0.017894 0.023878 0.031709 3 15 0.008807 0~012107 0.016551 0.022503 0.030436 2 16 0.007856 0 . 0 1 1 0 1 1 0 . 0 1 5 3 4 1 0.021252 0.029277 1 17 0.00702 0 . 0 1 0 0 3 2 0.014247 0.020108 0.028214 0 18 0.006283 0 . 0 0 9 1 5 5 0.013251 0.019055 0.027232 cent. as such, we note that tables 4, 5, and 6 have been prepared assuming investment returns of 4, 8, and 12 percent, respectively. the effect of portfolio choice on the required monthly savings figure is illustrated in the following example. assume that mr. and ms. risk-averse are more concerned with safety than with yield on their invested funds, and that they construct a low-risk portfolio expected to return 4 percent per annum. mr. and ms. takachance, on the other hand, invest in a diver sifted portfolio of common stocks and hope to earn an average annually compounded return approximating 12 percent. assuming a child aged five, current annual costs of $6,991, and 6 percent tuition inflation rate, the investors more concerned with safety set aside $257 (= $6,991 x .036848) per month. the parents who have invested in the riskier portfolio need only set aside $136 (= $6,991 x .019394) per month. in short, financial planners will note that it becomes a simple matter to demonstrate to clients the effect of one's risk tolerance on the monthly savings requirements. 52 financial servic'f~ review 6(1) 1997 vil conclusion the goals of this paper are twofold. first, by replacing the commonly employed "one-size fits-all" tables for college saving with those impounding several possible assumptions about inflation rates and investment yields, we seek to add greater precision and flexibility to the financial planning process. additionally, by providing a simple two-step algorithm for determining the required saving amount, we hope to retain the simplicity of the tables currently in use. given a single, easily obtainable input parameter (the current annual cost of college), we provide the means to estimate the total cost for a range of inflation assump tions, regardless of the child's current age. further, we modify the unrealistic assumptions often made about how people save for college to better reflect economic behavior. then we supply tables with which to compute the monthly savings required to fund the education, given today's annual cost and the estimated investment yield. these values are signifi cantly lower than those appearing in many published tables. from the viewpoint of the financial planner, the information provided in this paper should serve to make the planning process easier and more accurate by reducing the monthly outlays required and by spelling out the assumptions embedded in traditional presentations. references arrow, ic (1973). higher education as a filter. journal of public economics, 2, 193-216. banm~ s. (1994). access, choice, and the middle class. journal of studentfinancialaid, 24, 17-25. college for financial planning. (1996). study guide---financial planning process and insurance. national endowment for financial education press. cohn, e., & hughes, w. (1994). a benefit-cost analysis of investment in college education in the united states: 1969-1985. economics of education review, 13, 109-123. fenske, r., & batberini, p. (1995). "financial aid to students." encyclopedia of educational research. new york: macmillan publishing. feldstein, m. (1992). college scholarship rules and private saving. working paper 4032. cam bridge, ma: national bureau of economic research. hecker, d. (1992) reconcifing conflicting data on jobs for college graduates. monthly labor review, july, 3-12. heywood, j. (1994) how widespread are sheepskin returns to education in the u.s.? economics of education review, 13, 227-234. ibbotson associates. (1995). stocks, bonds, bills, & infiation--1995 yearbook. chicago, il: author. research associates. (1995). inflation measures for schools, colleges, & librariesd1995 update. washington, dc: author. spence, m. (1974). market signalling. cambridge, ma: harvard press. wolpin, k. (1977). education and screening. american economic review, 67, 949-958. pii: s1057-0810(96)90028-1 financial services review, 5( 1): 7 l-g 1 issn: 1057-0810 copyright (4 1996 by jai press inc. all rights of reproduction in any form reserved. an emerging partnership: afs and the cfp board dede pahl the past 20 years have witnessed growth in personalfinancial planning-as a profes sion and as a separate body of knowledge worthy of attention by the academic community. the certified financial planner board of standards is a private, nonprobt, regulatory organization that has emerged as the group bringing these new practitioners and the personalfinance academics together in the quest towards an independent body of knowledge developed exclusively for personal financial planning. the cfp board has funded its own research, through job analyses and market surveys and academic research through grants and monetary awards for seminal articles and research papers. i. introduction over the last 5 to 10 years on campus or in print, you may have noticed an increase in the following terms: financial planner, financial planning, cfp. fellow professors, students, or deans may be inquiring about this newly emerging profession of personal financial plan ning and how it relates to the academic world of finance. the term “emerging” is used because 20 years ago there was no such category, although many professionals (lawyers, accountants, insurance specialists, financial advisors) were being asked by clients to move out of their “‘professional box” to provide assistance in other related areas. finance, partic ularly, moved out of the theoretical and corporate worlds into the individual investor’s realm. deregulation, the baby boom generation, and the proliferation of 401(k) plans have spawned the consumer’s need for education, coordination, and planning. while the practi cal side of financial planning has accelerated through the process of professional certifica tion, the research and academic pursuit is poised to catch up under a new partnership for growth. with over 30,000 cfp licensees currently, the financial planning profession is obviously growing rapidly and is in need of resources (studies, research, applications) for their practices. how can each planner do a better job of planning for each client? how can they know if the information they receive is useful, reliable, valid? how can they use cur rent research in new ways? are theories only that-where is the application? the afs board of directors is extremely interested in bringing the academicians and the practitio ners closer together so that some of these questions can be answered. let me start by intro ducing you to financial planners, the cfp board, and the process of cff certification. dede p&l l director of certification, cfpboard, 1660 lincoln street, suite 3050, denver, co 80264. 12 financial services review 5( 1) 1996 ii. w~tispersonalfinancl~lplanning? there is no precise definition of personal financial planning. however, it is generally accepted to be aprocess for determining how a person can meet specific life goals through the proper management of his or her financial resources. personal goals might include a comfortable retirement, buying a home, saving for a child’s education, or starting a small business. to help a client reach specific goals, a personal financial planner might address a host of interrelated issues such as budgeting and saving, tax planning, investments, and insurance, or focus on only one area of a person’s financial life. regardless of the number or extent of issues involved, the process of financial planning usually involves the follow ing six key steps: 1. establishing the client-planner relationship. 2. gathering client data, including the client’s financial goals objectives. 3. identifying and evaluating client’s financial status. 4. developing and presenting a financial plan. 5. implementing the financial plan. 6. monitoring the plan and implementations. financial planning is a process, not a product. while the practice of financial planning can be performed in conjunction with providing financial products, it can also be performed as a distinct and separate function from the practice of any other profession or occupation. although financial planning often involves the use of financial products, it is not, when properly practiced, a process driven by the sale of insurance products or securities. iii. the origins of the cfp credential the designations, certified financial planner and cfp, were first used in 1972 by the col lege for financial planning, an educational distance-learning organization located in den ver, co. several financial services professionals around the united states had been trading information over the years in order to help clients with a broad outlook on financial goals and strategies (investments, insurance, tax, retirement, etc.). this group met in chicago on december 12, 1969 to form two organizations: the international association for financial planning (an industry association) and the college for financial planning (the educational arm). the college’s goal was to provide a student with the education (the equivalent of 15 18 college semester credits) and the designation. uniquely, the college for financial plan ning also registered the designations as trademarks with the u.s. patent and trademark office. when a candidate received the designation, he or she was actually earning the right (as in “license”) to use the trademark. this also means that the trademark use can be taken away, unlike an educational designation. in 1982, the american college created the char tered financial consultant (chfc) designation followed by the american institute of cer tified public accountants personal financial planning division’s personal financial specialist (pfs) designation in 1986. an emerging partnership 13 while the young financial planning profession initially served the well-to-do, other consumers came to recognize its benefits, especially in light of an explosion in the number of financial products being offered, a more complex, global economy, more complicated tax laws, and the need for more personal involvement in retirement planning. due to the lack of federal or state regulatory oversight, by the mid-1980s, an estimated 10040,000 individuals (depending on who was doing the counting) were calling themselves financial planners. the vast majority of these individuals had no formal education in financial plan ning and were using the term merely to sell traditional and not-so-traditional financial products, a practice that eventually led to large consumer losses due to tax law changes and a subsequent plunge in public confidence in the financial planning profession. iv. who is the cfp board? as a result, the cfp board (originally called the international board of standards and prac tices for certified financial planners or ibcfp) was founded in july 1985 by the college for financial planning and the institute of certified financial planners (a membership group) as an independent, standards-setting and certifying board to act in the public inter est by maintaining the quality of the cfp and certified financial planner marks. with tire establishment of the cfp board as an independent certifying body and the transfer of the ownership of the cfp and certified financial planner marks, the emphasis shifted from the cfp license as an educational credential to one of professional certifica tion. professional certification connotes competency, occupational experience, adherence to standards of practice. this type of certification can be mandatory (as in the state licens ing of medical doctors, architects, engineers, nurses, etc.) or voluntary (as in certified trust officers, chartered financial analysts). the growth of the financial planning profes sion during the early 1980s had created a demand for more financial planning training and education and had moved the need for the separate maintenance and surveillance of profes sional standards and self-discipline. because a planner can provide services without the cfp credential, we can only demonstrate this growth through the increasing numbers of those who are certified. in 1986, the cfp board monitored 12,000 certified planners. by the end of 1995, we had grown 263%. according to a 1995 cfp board survey, the top reasons financial professionals cited for wanting the designation were: enhanced credibility, increased knowledge and exper tise, and improved professional standing among colleagues. since june 1986 when the cfp board acquired ownership of the marks as well as all responsibility for tire testing and certification of cfp licensees, the cfp board adopted a code of ethics and standards of practice, disciplinary rules and procedures. other ele ments of the certification process quickly followed: the registration of financial planning programs of educational institutions other than the college for financial planning (now numbering more that 75 programs around the united states, listed in table 1). continuing education requirements, and the delineation of experience requirements and annual licens ing procedures. the cfp board accelerated the monitoring of cfp licensees in the international arena in 1990 with the signing of a license and affiliation agreement with the iafp of australia and now six countries are members of the international cfp council (united states, aus financial services review 5( 1) 1996 table 1 cfp board-registered programs: summary sheet graduate degree program georgia state university golden gate university loyola university, new orleans northeastern university nova southeastern university san diego state university texas tech university university of dallas university of miami widener university undergraduate degree programs baker university barry university baylor university california state university-chico california state university-fullerton campbell university delta state university fort hays state university kansas state university mankato state university mercyhurst college metropolitan state college mississippi state university purdue university san diego state university southern arkansas university southwest texas state university texas tech university university of alabama university of northern colorado university of wisconsin western kentucky university wright state university certtjicate programs baker college barry university boston university college for financial planning* fairleigh dickinson university florida institute of technology florida state university georgia southern university golden gate university lehman college long island university loyola university-chicago manchester community college marist college medaille college merrimack college metropolitan state college new york university nova southeastern university oakland university ogelthorpe university old dominion university pace university rockland county community college rollins college southwest texas state university st. john’s university the american college university of california university of central florida university of houston university of miami university of north texas university of tennessee associate degree programs waukesha county technical college tralia, japan, united kingdom, canada, and new zealand). the international cfp coun cil is an international assembly of financial planning bodies that exists to promote the professionalism of individuals and organizations offering personal financial planning ser vices and to insure that such services are offered in an ethical and competent manner throughout the global community. to join the international cfp council, organizations must make a commitment to adopt the cfp certification program, symbolized by the trademarks cfp and certified financial planner. the council is charged with establishing these marks as the globally recognized standards of excellence for professional financial planners. the number of member-financial planners outside the united states totals more than 15,000 individuals, which brings the total worldwide to approximately 50,000 cfp licensees. an emqging parfnersh@ 75 the cff board is governed by a board of governors, 19 volunteers reflecting the practice of financial planning: sole practitioners, wall street, consumer advocates, aca demia. the board of governors oversees the work of four sub-boards. the board of examiners is responsible for the examination and education requirements. the board of ethics is responsible for the ethics enforcement and disciplinary procedures. the intema tional cfp council was described above. the board of practice standards has just been formed to begin the process of establishing standards of practice based on the job analy sis. the cfp board’s staff headquarters is located in denver, co, and currently 30 staff members administer the preand post-certification policies put into place by the govem ing board. v. certification requirements to be authorized to use the trademarks cfp and certified financial planner, individuals must meet requirements in four areas: education, examination, experience, and ethics. these requirements, which are based on a 1987 job analysis study of practicing financial planners (which was updated in 1994), assure the public that persons who have been autho rized to use the cfp marks have met rigorous professional standards. a job or practice analysis is the foundation of any bona fide certification program. it is a systematic collec tion of data that describes the responsibilities required of a professional and the skills and knowledge needed to perform the job. the cfp board, whose appointed officers and board members are planning practitioners, determined from the job analyses that the combination of the following four requirements were necessary for attesting to the competency of the mark holder. a. jmucatlon before applying for the cfp certification examination, a candidate must first com plete academic coverage of a financial planning curriculum through one of the three edu cational paths: completion of a financial planning education program that has registered with the cfp board, completion of the curriculum through transcript review of upper divi sion coursework taken at an accredited u.s. college or university; or holding certain finan cial planning related designations or degrees (cpa, chfc, clu, cfa, phd in business economics, doctor of business administration, or licensed attorney). b. cfp board-registered personal financial planning programs the first educational path is through the completion of a course of study in financial planning offered by an educational program registered with the cfp board. a current list of these programs is found in table 1. each registered program provides complete cover age of the topics determined by the cfp board to constitute the core curriculum for per sonal financial planning practitioners. each individual educational institution will present these topics under various course names and titles, such as risk management, facial analysis, estate planning. some programs provide a “capstone” course that integrates all areas of financial planning, based on case analysis; others will include a review course in preparation for the certification examination. typically, the course work can be com 76 financial services review 5( 1) 1996 pleted in 18 to 24 months. successful completion of a registered program automatically satisfies the education component of the cfp certification process. program directors at any cfp board-registered program have the authorization to waive students out of courses if the program director deems the student has already satisfied the topic coverage of that course. there is no time limit between the completion of a registered program and sitting for the certification examination. c. transcript review if candidates completed upper-division course work at an accredited u.s. college or university that covers some or all of the topics, then they may have the transcript(s) reviewed by the cfp board of standards. the nonrefundable fee for this review is $100. “upper-division” is defined as courses at the junior, senior, or graduate level. if the review confirms that all of the necessary topics have been covered, then the candidate may apply to sit for the certification examination. however, if the review confirms that some of the topics have not been covered, then the candidate will receive a letter outlining the missing topic areas. the cfp board will not pre-determine the status of any course. if the transcript review confirms topic coverage is lacking, the candidate may then (a) complete the educa tion requirement at any accredited college or university and resubmit this course work to the cfp transcript review process with the understanding that the course work may still not be accepted as adequate topic coverage. or (b) complete the requirement at a cfp board registered program (see table 1) with the understanding that this additional course work will be accepted by the cfp board without further review. d. challenge status certain degrees and professional credentials will automatically allow a candidate to sit for the certification examination without having to verify educational topic coverage. the board-recognized degrees and credentials (certified public accountant, chartered finan cial consultant, chartered financial analyst, phd in business or economics, doctor of business administration, and licensed attorney) with 3 years ofjhznciul planning-related work experience have been approved by the cfp board as meeting the educational require ment. if candidates are challenging the exam on the basis of a cpa designation or licensed attorney, they must provide a letter of good standing from the applicable licensing board. (an inactive cpa license will be accepted.) if the challenge is on the basis of the chfc or cfa designation, evidence of good standing from the membership organization, or its equivalent, must be provided. e. curriculum registration one of the cfp board’s principal assignments is to evaluate applications for financial planning curriculum registration. the board of examiners, together with the administrator of certification services and the director of certification, operate as an objective and unbi ased review body. its members judge degree or certificate programs based on submitted applications and supportive materials, and requested written or oral testimony, to deter mine compliance with the board-approved registration criteria an approved educational guidelines. the purpose of these standards is to promote the overall high quality of finan an emerg& partnersh& 77 cial planning educational programs. emphasis should be placed on substantial rather than form-related compliance. financial planning programs are registered for 3 years. after initial registration is granted, every 3 years all registered programs must submit a complete application to renew the registration. financial planning programs change over time and the criteria may be periodically amended. this registration is renewed and programs reviewed in their entirety every 3 years. an annual report is also required to demonstrate continued compliance with the registration criteria and amendments. in addition, the reports provide the cfe board with national information and yearly data on financial planning programs. deficiencies noted in an annual update could lead to a review of the institution’s registration. f. examination once candidates have successfully completed the education requirement, they are eli gible to apply for the cfe certification examination, which serves as the uniform financial planner competency examination. the certification examination is a n-hour, 2-&y examination designed to assess a candidate’s ability to apply the financial planning educa tion to financial planning situations in an integrated format. topics on the examination are based on the financial planning job analysis (see below) in accordance with procedures that ensure its job relatedness and content validity. the format of the exam is all multiple choice questions (approximately 320 questions per exam); three case questions are also included-a three or four page fact pattern with 20 multiple-choice questions. g. experience cfp candidates must also show evidence of having worked in a personal financial planning-related position for compensation. a summary of the work experience, signed by a supervisor, is reviewed by the cfr board’s registrar, pursuant to guidelines developed by the governing board. the experience requirement is designed to provide the public with the assurance that the candidate understands the counseling nature of personal financial planning. three years of financial planning-related experience is the basis of the work experience requirement for candidates with a baccalaureate degree. h. ethics prior to certification, cfp candidates must complete an ethics statement which requires the disclosure of past or pending litigation or agency proceedings relating to their professional conduct. for example, the candidates and all current licensees disclose: (a) extensive professional background; (b) conflict of interest that might affect the profes sional relationship, objectivity, or independence; and (c) any past or pending litigation or agency proceedings and to acknowledge the right of the cfe board to enforce its code of ethics and professional responsibility. as a private entity, the cfp board does not have the governmental authority to prohibit an individual financial planner from practicing. however, if a cfp licensee fails to meet the high standards of the code of ethics, appro priate disciplinary action may be taken. the severest of those actions is the individual’s loss of the right to use the cfe marks. revocation of the license or other disciplinary actions are disseminated to the media and other regulatory bodies, like the state securities and insurance departments. disclosure forms are reviewed by the cfp board’s legal staff 78 financial services review 5( 1) 1996 and any questionable disclosure is investigated before the individual receives the right to use the cfp marks. vi. job analysis ctb/mcgraw-hill completed a comprehensive job analysis study for the practice of cer tified financial planner licensees on june 30,1994. the study was an important component of continuing to ensure that the certification examination reflects the requirements of prac tice of personal financial planning. the study included input from a diverse sample of prac ticing cfp licensees and updated the job analysis published in 1990. the study was conducted for the cfp board by ctb/mcgraw-hill staff members with extensive experi ence in job analysis and certification program development. to be licensed to use the cfp board’s certification marks, candidates must meet the experience, ethics, education, and examination standards established by the cfp board. the examination is a critical portion of the certification process and representative of the standards of practice. the examination must reflect the requirements to practice as a cfp licensee and, therefore, should indicate the impact, if any, of current trends, financial and regulatory agency mandates, market requirements, and various practice settings on the tasks required to perform the job of a practicing cfp licensee. to reflect these requirements, an accurate assessment of current practice must be con ducted at regular intervals. the cfp board anticipated that the new job analysis would pro vide an update to the previous analysis performed in 1987, without revealing radical changes in practice. the results of the new job analysis study enable the cfp board to re evaluate the content outline for the certification examination for currency of practice and job relatedness and to re-evaluate the emphasis on certain tasks required for competent, professional practice. in keeping with recognized standards in testing, the job analysis process included information about current job practice provided by content experts and a large-scale, national survey of practitioners. prior to the development of the survey, ctbimcgraw-hill conducted a review of materials on the cfp licensee program, the results of the previous job analysis, the code of ethics, sample courses, and other related materials. approxi mately, 30 telephone interviews were conducted to obtain additional information for the survey instrument. the interviews were conducted with cfp licensees who were familiar with large-scale changes, demands, innovations, and trends in current practice. participants were solicited by the cfp board. next, a job analysis update committee was selected by the cfp board from out standing cfp licensees representing a variety of practice settings. the committee mem bers included a diversity of content expertise, geographic distribution, practice settings, and demographics of the profession. the committee’s leadership was crucial to the success of the project. with guidance from ctb, the committee finalized the list of tasks that describe the scope of practice, created the survey instrument, and identified a sampling plan for the survey. the survey was distributed to 2,600 cfp licensees. participants were given detailed instructions and asked to rate each of the 154 tasks described to indicate the frequency of the performance of the task and the criticality of the task to competent perfor mance (a measure used to define tasks that have serious consequences). survey respon an emerging partnership 79 dents were also give the opportunity to write in additional comments regarding the structure and content of the survey task list. demographic questions included in the survey also enabled the cfp board to collect background information of value toward the analysis of the results. the surveys were collected, processed, and analyzed by ctb, and the results presented to the job analysis update committee for final interpretation and recommenda tions. the results indicated a high importance placed on the financial planning process with particular emphasis on tasks associated with gathering client data (especially determining client goals and objectives) and tasks associated with analyzing and evaluating client finan cial status. respondents considered of next importance the tasks included in developing and presenting the financial plan and establishing the client-planner relationship. areas of knowledge which were rated highest included: tax planning, investments, retirement issues, and the time value of money and principles of compounding. it is notable that no additional changes to the job analysis task list were necessary as a result of information obtained from the respondents. based on the results of the job analysis, the cff board has proceeded to update the certification examination to reflect the revised content outline. this will be accomplished by a correlation of the job analysis results to new test specifications which reflect the estate planning investment retirement risk tax planning planning management planning f&we 1. job analysis licensee practice areas 80 financial services review 5( 1) 1996 results. accordingly, item development, examination construction, and final examination forms will proceed to reflect the new job analysis as this interpretation is integrated into the process through the test specifications. changes in the content outline and test specifica tions will be indicated on material sent to prospective candidates as the process is com pleted. this revision, scheduled to be completed by january, 1997, reinforces the role of the cfp board to ensure that the examination process reflects the practice and to assure the continuance of the high standards of the cff’ board’s certification marks. based on certain demographic information captured during the job analysis, tbe fol lowing summarizes some of the key benchmark findings. l sixty percent of respondents identified investment planning as an area of exper tise, and 54% selected retirement planning. investment vehicles are rated the most important area of their practice by survey respondents, while retirement and tax planning are gaining prominence. the consumer emphasis has shifted from seeking advice simply on investments to also seeking advice on retirement and tax issues. see figure 1. l forty-nine percent of survey respondents were associated with a financial plan ning firm, 14% with a securities brokerage firm, 12% with an accounting firm, and 10% with an insurance company. 1994 results show fewer cfp licensees were working in financial planning firms than in 1987; more were being employed in the financial, legal, and education industries. l ffty-three percent of the cfp licensees polled charged a combination of com missions and fees. the majority of survey participants are charging clients a combination of feeand commission-based fees, rather than fee-only or commis sion-only charges. figures for respondents providing fee-only and commission only services dropped from 21% and 64% in 1987 to 19% and 19% respectively. l increasing from 14% in 1987 to 23% in 1994, the number of women becoming cfp board licensed financial planners almost doubled in tbe last 7 years. see table 2 for additional findings from the study. table 2 how cfp licensees rate the importance of their financial tasks i. determine client’s personal and financial goals, needs and proprieties. ii. obtain information from client through interview/questionnaire about financial resources and obligations. iii. determine client’s risk tolerance level. iv. determine the client’s financial status by analyzing and evaluating current investments. v. determine the client’s financial status by analyzing risk tolerance. vi. determine client’s time horizons. vii. determine the client’s financial status by analyzing and evaluating current investment strategies and policies. viii. assess client’s values, attitudes, and expectations. ix. present and review the plan with the client. x. determine the client’s financial status by analyzing and evaluating current assets, liabilities, cash flow, debt management. an emerging partnership 81 vii. future research needs of the cfp board obviously, the cornerstone of our profession is the emphasis on education+f the planner before and after certification and of the clients who become “better” clients as they become more educated. the cfp board is poised to assist both groups. we currently provide grant assistance for research on financial planning subjects. each year, the board of examiners, on behalf of the cfp board, reviews grants submitted by february 1. five grants of $5,000 each are available for research into the integration of the major areas of planning: insur ance, income tax, retirement planning, investments, and estate planning. half of the grant is paid upon initial acceptance, and the balance is paid if the work is completed (in publish able form) within 18 months. to date, the cfp board has awarded four grants, two of which have been completed. the cfp board also regularly recognizes outstanding articles in the “nontraditional” financial planning arena (“nontraditionai” meaning not necessarily based on academic research) that add quality to the practice of fmancial planning. five monetary awards are available each year, and a $1,000 “best paper on personal financial planning” is available to presenters at the annual afs conference. further, the cfp board works closely with more than 75 colleges and universities that have registered their financial planning curricula with us as covering the topics currently tested for certification. the rapid increase in interest by the academic community has fur thered the interest of students, practitioners, and consumers, all of whom are seeking more information on personal financial planning subjects. finally, tire growth in interest from the international community is testimony to the growth of the profession. the six current members of the international cfp council (united states, australia, japan, united kingdom, canada, and new zealand) are all inter ested in expanding the body of financial planning knowledge. of particular concern is find ing the common areas of study and a network of information and research. for more information on any of these topics, please feel free to contact the cfp board dede pahl, director of certification, cfp board, 1660 lincoln street, suite 3050, denver, co 80264. m. sumhlary the past 20 years have witnessed growth in personal financial planning-as a profession and as a separate body of knowledge worthy of attention by the academic community. the certified financial planner board of standards is a private nonprofit, regulatory organiza tion that has emerged as the group bringing these new practitioners and the personal finance academics together in the quest towards an independent body of knowledge devel oped exclusively for personal financial planning. the cfp board has funded its own research, through job analyses and market surveys, and academic research through grants and monetary awards for seminal articles and research papers. financial planning practitio ners are eager for new approaches to old planning dilemmas, for easy-to-understand appli cations of past academic theories, and for a partnership with the academic community to further the information available to themselves as planners and to their clients. manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: ‘‘smith (1992) reports that’’ or ‘‘this problem has been studied previously (ho, milevsky, & robinson, 1999).’’ the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) prsrt std u.s. postage p a i d easton, md permit no. 114 mobile bank applications: loyalty of young bank customers mustafa nourallaha,*, christer strandberga, peter öhmana adepartment of economics, geography, law and tourism, and centre for research on economic relations, mid sweden university, se-851 70 sundsvall, sweden abstract the purpose of this study is to investigate how young bank customers (ybcs) perceive the relationships between several antecedents (i.e., usability, responsiveness, customer satisfaction, and reliability) and loyalty in the context of mobile bank applications (mbas). an electronic questionnaire was sent to 500 ybcs in sweden, 146 of whom completed it. confirmatory factor analysis was used to test the measurement model, and structural equation modeling was used to test the hypotheses. the results indicate that usability is indirectly related to loyalty through responsiveness and customer satisfaction. the study contributes to the literature by developing a usability–loyalty model of ybcs using mbas. © 2021 academy of financial services. all rights reserved. jel classification: m keywords: usability; customer satisfaction; loyalty; mobile bank application; young bank customers 1. introduction studies in the literature on loyalty in the financial services context have indicated that while bank customers in general are loyal (e.g., strandberg, wahlberg, & öhman, 2015), young bank customers (ybcs) are often not (nicoletti, 2017). ybcs are twice as likely to change banks as are older bank customers (accenture, 2015). although ybcs represent an important customer category for traditional banks (foscht, maloles, schloffer, chia, & sinha, 2010), members of this group show a tendency to use financial services provided by fintech companies. gomber, kauffman, parker, and weber (2018) state that ybcs seem to *corresponding author. tel: +46 10 142 79 75; fax: +46 10 142 85 10. e-mail address: mustafa.nourallah@miun.se (m. nourallah) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 147–167 prefer financial services provided by google, amazon, apple, or paypal, that is, fintech companies, rather than by traditional banks. it has also been emphasized that “fintech companies offer new products and solutions which fulfill customers’ needs that have previously not or not sufficiently been addressed by incumbent financial service providers [e.g., traditional banks].” (gomber, koch, & siering, 2017, p. 540). the ambition of fintech companies to provide one-fifth of financial services by 2020 (gimpel, rau, & röglinger, 2018) has led to competition between these companies and traditional banks, and ybcs seem to be the target of both these groups. traditional banks have developed mobile bank applications (mbas), that is, an advanced type of mobile banking. mbas allow customers connected to the internet to conduct various financial tasks, such as checking account balances, transferring money, and paying bills (malaquias & hwang, 2019). simultaneously, the fintech companies have promoted mobile-only banks (mobs), a recent innovation that offers financial services to customers connected to the internet solely via mobile applications (nourallah, strandberg, & öhman, 2019). one place where this rivalry between fintech companies and banks is found is sweden. in 2018, the first mob, n26 launched mobile financial services, and in 2019, another mob, lunar way announced that every month roughly 6,000 new customers subscribed to their services (lundell, 2019). the literature identifies several research gaps concerning the various antecedents of loyalty applicable to mbas based on the perceptions of ybcs. in their review of the loyalty literature, kandampully, zhang, and bilgihan (2015) highlight the mobile loyalty of the young generation as a future research area. in another review of the mobile banking literature, tam and oliveira (2017, p. 1060) state that “knowing the determinants of the postadoption phase, and keeping customers loyal to m-banking are the emerging issues that should be considered in future research.” larsson and viitaoja (2017) emphasize the need to investigate how usability affects loyalty. another research area was summarized by chakraborty and sengupta (2013), who discuss the need to study the relationship between customer satisfaction and loyalty in the mba context. iberahim, taufik, adzmir, and saharuddin (2016) investigate reliability and responsiveness in the automated teller machine (atm) context, and suggest considering these concepts in other contexts as well. addressing these research gaps, the current study investigates how ybcs perceive the relationships between a number of antecedents (i.e., usability, responsiveness, customer satisfaction, and reliability) and loyalty in the context of mbas. the structure of the rest of the article is as follows: the next section presents the frame of reference; section three concerns methodological issues; section four presents the results; and section five concludes the article. 2. frame of reference and hypothesis development 2.1. the context of the study 2.1.1. mobile financial services shaikh and karjaluoto (2019) argue that mobile financial services can be divided into mobile banking, mobile payments, and mobile money. the first two types of mobile financial 148 m. nourallah et al. / financial services review 29 (2021) 147–167 services are found in more inclusive financial systems, for example, in sweden, and are conducted with more inclusive customer segments, that is, customers who have access to banking services. the third service represents a relationship between a mobile money solution, such as m-pesa, and nonbank customers. this service is common in less inclusive financial systems, for example, in sub-saharan africa (demirguc-kunt, klapper, singer, ansar, & hess, 2018), and in less inclusive customer segments, that is, customers who face difficulties (e.g., long distance) in accessing banking services. shaikh and karjaluoto (2019) examine the landscape of mobile financial services, giving insight into the types of relationships between more or less inclusive financial systems and more or less inclusive customer segments. however, they do not differentiate between the types of financial institutions that offer mobile financial services, that is, traditional banks and fintech companies. shaikh and karjaluoto (2019) use mobile banking to refer to various types of mobile financial services, including mobile banking provided by traditional banks. because mobile banking does not represent a homogeneous type, it can be divided into services provided by wireless application protocol (wap), short message service (sms), and mbas. it is worth noting that wap and sms banking represent earlier versions of mobile banking in which bank customers access their bank accounts via either a mobile internet browser or sms. these rudimentary types of mobile banking prompted remarkable customer aversion. for example, during the 2003–2006 period, 15 german banks stopped offering such services to customers due to lack of use (scornavacca & hoehle, 2007).1 in south korea, only 4% of online customers adopted these earlier versions of mobile banking in that period (lee, park, chung, and blakeney, 2012). moreover, “in 2003 . . . less than 1% of banking transactions in taiwan were conducted through mobile handsets” (luarn & lin, 2005, p. 874). similar situations existed in finland (suoranta & mattila, 2004), china (laforet & li, 2005), and the united states (mallat, rossi, & tuunainen, 2004). earlier versions of mobile banking were not as widespread as expected (koenig-lewis, palmer, & moll, 2010; mohammadi, 2015; shaikh & karjaluoto, 2015). mobile banking system limitations, such as tiny screens and keypads and slower transaction speeds, caused this aversion (laukkanen, 2007; lee & chung, 2009). however, since 2007—after the first iphone was launched (shaikh & karjaluoto, 2019)—the situation changed dramatically and mbas have become a basic means of conducting daily financial transactions such as checking balances, transferring money, and paying bills (liébana-cabanillas, alonso-dos-santos, soto-fuentes, & valderrama-palma, 2017; tan & lau, 2016). this change likely emerged due to greater accessibility to the internet (lu, tzeng, cheng, & hsu, 2015), advanced generations of smartphones (shaikh & karjaluoto, 2015), and the development of application technology (sun, wang, & wang, 2015). 2.1.2. young bank customers young customers are more enthusiastic about using their mobile phones than are members of other age groups (yeh, wang, & yieh, 2016), and they have advanced skills in dealing with various technological financial platforms (killins, 2017). they also spend m. nourallah et al. / financial services review 29 (2021) 147–167 149 significant amounts of time using these platforms (kaur & medury, 2011). it is worth noting that reaching ybcs is a top priority for banks (tan & lau, 2016). recent studies recommend investigating bank customers, such as ybcs, who possess limited financial information (aydin & akben selcuk, 2019). moreover, ybcs will seek home mortgages and other financial services in the near future, so it is important for banks to secure loyal ybcs, given that fintech companies will be the main providers of financial services (gimpel et al., 2018) and that these companies can satisfy customers in other and possibly better ways than can traditional banks (gomber et al., 2017). ybcs can contribute to increased bank profits in terms of immediate profits, future profitability, market share, and diverse profitable relationships (foscht et al., 2010). the literature reveals that different terms have been used interchangeably to refer to ybcs: the millennial generation (e.g., tan & lau, 2016), the young generation (e.g., koenig-lewis et al., 2010), and generation y (killins, 2017). also, previous studies have used different age groups when investigating ybcs. calisir and gumussoy (2008) use the 18–26-year age range, sum chau and ngai (2010) 16–29 years, and akturan and tezcan (2012) 16–25 years. in this study, ybcs are bank customers aged 18–29 years, that is, the interval from first being considered “adult” in sweden to the highest year considered in the three studies mentioned above. 2.2. conceptual framework 2.2.1. central concepts electronic financial services refer to accessing a bank account via computers and/or mobile financial services (shaikh & karjaluoto, 2019). in this context, studies have addressed responsiveness and reliability (broderick & vachirapornpuk, 2002), customer satisfaction (sampaio, ladeira, & santini, 2017), and loyalty (larsson & viitaoja, 2017). overall, studies report that customer satisfaction and loyalty are the most important factors delivering a good experience (berraies, yahia, & hannachi, 2017), while reliability is identified as a necessary risk-related factor in technology-based financial services (hanafizadeh, behboudi, koshksaray, & tabar, 2014). in a similar vein, sindwani and goel (2015) argue that responsiveness is an important concept in the electronic financial services context. a number of previous studies have addressed usability-related issues (e.g., mohammadi, 2015). the international organization for standardization (ios, 1998) defines usability as “the extent to which a product can be used by specified users to achieve specified goals.” kang, lee, and lee (2012) state that mba usability likely concerns mobile interface and navigation issues. casaló, flavian, and guinalı́u (2007, 2008) and flavian, guinalı́u, and gurrea (2006) find that in the banking industry, the essence of usability is represented by ease of understanding, observed content, simplicity, speed, ease of site navigation, and user control. from an electronic financial services perspective, customer satisfaction is created by meeting customer expectations regarding financial issues (amin, 2016), while loyalty is seen as a dichotomy between attitude and behavior. attitudinal loyalty includes “a degree of 150 m. nourallah et al. / financial services review 29 (2021) 147–167 dispositional commitment, in terms of some unique value associated with the brand” (lin & wang, 2006, p. 272), and behavioral loyalty refers to a customer’s repurchase behavior, due to their liking for particular financial services (amin, 2016). it is worth mentioning that most previous studies of service quality have used the servqual instrument (parasuraman, zeithaml, & berry, 1988), which consists of five dimensions: tangibles, reliability, responsiveness, assurance, and empathy. the current study excludes three of these dimensions: tangibles, assurance, and empathy. parasuraman et al. (1998, p. 23) state that tangibles are “physical facilities, equipment, and appearance of personnel,” assurance is the “knowledge and courtesy of employees and their ability to inspire trust and confidence,” and empathy is “caring, individualized attention the firm provides its customers.” it can be argued that these dimensions are related to the customer–employee relationship dimension, which is not part of the mba context. hence, the current study only uses the responsiveness and reliability dimensions of service quality, since mbas have evolved in an environment in which the nature of mobile–human interaction differs from personal interaction (oliveira, thomas, baptista, & campos, 2016), and because ybcs do not prefer personal connections when accessing banking services (carlander, gamble, gärling, hauff, johansson, & holmen, 2018). in this regard, responsiveness is the willingness to help consumers and provide prompt service (parasuraman et al., 1988), and in terms of mbas, it has two components: service speed and technology (iberahim et al., 2016). reliability is defined as “the ability to perform the promised service dependably and accurately” (parasuraman et al., 1988, p. 23). the current study adopts these definitions. 2.2.2. the research model and hypotheses the research model is presented in fig. 1 as can be seen, the literature suggests that usability is related to responsiveness, customer satisfaction, and reliability, as indicated by h1, h2, and h3. subsequently, responsiveness and reliability are related to customer satisfaction, as indicated by h4 and h5. finally, these three concepts are related to loyalty, as indicated by h6, h7, and h8. the eight hypotheses are developed below. in the banking sector, usability will likely enhance speed, ease site navigation, and increase user control (casaló et al., 2007, 2008; flavian et al., 2006). usability can offer various benefits to customers (calisir & gumussoy, 2008), such as the ability to get banking help in various critical situations (gumussoy, 2016), to access a user-friendly system (hussien & aziz, 2013), and to use a variety of communication channels (laukkanen, 2007). offering a high level of usability will likely lead to good responsiveness (raza, jawaid, & hassan, 2015). accordingly, this study proposes the following hypothesis in the mba context: hypothesis 1 (h1): the higher the usability, the higher the responsiveness is likely to be. generally, usability can lead to a pleasant user experience (nielsen, 1994), affect customer expectations (bhattacherjee, 2001), and ensure customer satisfaction. theoretical arguments and empirical results have emphasized the importance of usability for customer decisions to use certain technological applications (e.g., hoehle & venkatesh, 2015). in the online banking context, empirical results indicate that usability can significantly affect m. nourallah et al. / financial services review 29 (2021) 147–167 151 customer satisfaction (casaló et al., 2008; flavian et al., 2006; hussien & aziz, 2013). similarly, thakur (2014), when studying mbas in india, finds that usability affects customer satisfaction. the following hypothesis is accordingly formulated: hypothesis 2 (h2): the higher the usability, the higher the customer satisfaction is likely to be. usability is also considered a key factor in e-business success (lee & kozar, 2012), and high usability ensures fewer difficulties in using a certain system (davis, 1989), promotes ease of use of that system (nielsen, 1994), and reduces possible errors (sanchez-franco & rondan-cataluña, 2010). in contrast, low usability generates payment-related issues (flavian et al., 2006). since high mba usability ensures trustworthy financial services in terms of transferring money, obtaining account information, and paying bills (mohammadi, 2015), it can be argued that usability drives reliability (cf. benlian & hess, 2011). hence, it is hypothesized that: hypothesis 3 (h3): the higher the usability, the higher the reliability is likely to be. responsiveness is the ability to provide help and instant services to customers, that is, provide fast replies regarding their bank accounts (raza et al., 2015), in turn increasing the customer satisfaction (iberahim et al., 2016). previous studies in the banking sector have presented contrasting observations about this relationship. while raza et al. (2015) and saleem, zahra, ahmad, and ismail (2016) state that there is a significant relationship between responsiveness and customer satisfaction, other studies (kassim & asiah abdullah, 2010; munusamy, chelliah, & mun, 2010) report contrary results. based on theoretical assumptions and more recent empirical studies, the current study suggests the following hypothesis: hypothesis 4 (h4): the higher the responsiveness, the higher the customer satisfaction is likely to be. bauer, falk, and hammerschmidt (2006) conclude that reliability is the most critical factor driving customer satisfaction. in investigating reliability in mobile payment services, arvidsson (2014) finds that consumers highly rate the importance of reliability. similarly, calisir and gumussoy (2008) emphasize the role of reliability in banking, and raza et al. (2015) demonstrate that reliability has a considerable effect on customer satisfaction. munusamy et al. (2010) investigate this relationship in the banking sector in malaysia and fig. 1. research model. 152 m. nourallah et al. / financial services review 29 (2021) 147–167 report no significant relationship, and wen and hilmi (2011) find the same lack of relationship in another malaysian study. overall, the results reported by most of the above studies lead to the following hypothesis: hypothesis 5 (h5): the higher the reliability, the higher the customer satisfaction is likely to be. quick responses to customer questions are seen as a factor leading to customer loyalty (srinivasan, anderson, & ponnavolu, 2002). loyalty can be ensured by offering a variety of communication channels (verhoef & donkers, 2005) and by providing embedded ways to ask for help (awwad & awad neimat, 2010). previous studies report that responsiveness could well affect loyalty (marimon, yaya, & casadesus fa, 2012; moorthy, chee, yi, ying, woen, & wei, 2017). in a study of mobile commerce, lin (2012) finds a significant relationship between responsiveness and loyalty. this leads to the following hypothesis: hypothesis 6 (h6): the higher the responsiveness, the higher the loyalty is likely to be. customer satisfaction is an important issue for any company (santouridis & trivellas, 2010), and banks are no exception. it explains post-purchase perceived performance (fornell, 1992) and ensures customer retention and profitability (strandberg, wahlberg, & öhman, 2012). previous studies report a strong relationship between customer satisfaction and loyalty (lin & wang, 2006; liébana-cabanillas et al., 2017; thakur, 2014). fornell (1992, p. 7) describes this relationship as follows: “loyal customers are not necessarily satisfied customers, but satisfied customers tend to be loyal customers.” the current study emphasizes this relationship, and formulates the following hypothesis: hypothesis 7 (h7): the higher the customer satisfaction, the higher the loyalty is likely to be. reliability enhances the ability of mbas to perform the promised customer services dependably and accurately (jun & palacios, 2016). previous studies have investigated reliability as a dimension of service quality, and empirical results support the relationship between reliability and loyalty (e.g., karatepe, 2011). other studies of reliability have reached similar conclusions. ho and lee (2007) suggest that reliability is a crucial factor for retaining customers, and moorthy et al. (2017) conclude that reliability is significantly and positively related to loyalty. however, in mobile retailing, lin (2012) finds no relationship between reliability and loyalty. in a similar vein, zhou, lu, and wang (2010) suggest that reliability might not be as important for ybcs as for older bank customers. nevertheless, the following hypothesis is based on most previous research: hypothesis 8 (h8): the higher the reliability, the higher the loyalty is likely to be. 3. method 3.1. measure development the items in the preliminary questionnaire were adopted from previous studies to ensure content validity (see the appendix). usability was measured by items (usa 1–5) from m. nourallah et al. / financial services review 29 (2021) 147–167 153 casaló et al. (2008). responsiveness items (res 1–3) and reliability items (rel 1–3) were adopted from lin (2013). customer satisfaction items (sat 1–2) were adopted from aydin and özer (2005) and yoon (2010), and loyalty items (loy 1–2) from chaudhuri and holbrook (2001) and wirtz, mattila, and lwin (2007). two focus group interviews were conducted with four and five ybcs, respectively. all participants belonged to the target age group, that is, 18–29 years, and had at least one year’s experience of mba use in sweden. the focus group interviews contributed to the detailed improvement of some items in the preliminary questionnaire. back translation was conducted to ensure that the items had good consistency (cf. brislin, 1970), and certain language-related revisions were made as a result. the last step was to send the preliminary questionnaire to two experienced ybcs to check for readability, and minor revisions were made based on their feedback. the final questionnaire was based on a seven-point likert scale ranging from 1= strongly disagree to 7 = strongly agree. the background variables included were age, gender, and mba experience. 3.2. sample, data collection, and data analysis procedures the final questionnaire was sent in electronic form to 500 students at a university in the midsweden region in late 2018. these students studied business administration, political science, or sociology, were aged 18–29years, and differed in the mba usage duration and the number of mbas used. in addition, the students were diverse in terms of socioeconomic class, gender, and cultural background. the main criterion for selecting these students was use of mbas for at least one year, which requires a swedish bank account. sampling university students enabled the current study to avoid limitations related to a sample associated with a single bank (e.g., strandberg et al., 2012), because the present respondents were customers of several banks. harm to participants, confidentiality of information provided, confidentiality of collected data, and data-storage issues were among the ethical concerns of the current study, and certain processes were used to address these concerns and the general limitations associated with questionnaires (cf. grinyer, 2009). approval to send out the questionnaire was obtained from responsible persons at the university program and course levels. brief information about the study was presented to the students, including advising that completing the questionnaire was voluntary and that financial information would not be gathered for the study. the anonymity of responses was ensured by using online software complying with the eu’s general data protection regulation. initially, 129 completed questionnaires were received; after two reminders, the total number of completed questionnaires increased to 146, that is, a response rate of 29.2%. following the suggestion of pohlmann (2004), an analysis was conducted comparing the results of those responding before and after the first reminder; no notable differences were found between these two groups. descriptive statistics, sample adequacy, and common method bias tests were calculated. in a further step, confirmatory factor analysis (cfa) was utilized to test how well the observed variables represent the latent variables (cf. hair, black, babin, & anderson, 2014). the current study used cfa to delete unnecessary items and refine the measurement model; it was also used to address reliability and validity issues. subsequently, structural equation 154 m. nourallah et al. / financial services review 29 (2021) 147–167 modeling (sem) was used to test the research model and the hypotheses. both cfa and sem were performed using lisrel 9.30. 4. results 4.1. descriptive statistics the characteristics of the sample are presented in the appendix. most participants were 18–23 years of age, and the sample was fairly equally distributed in terms of gender. only a small percentage of participants used more than three mbas. regarding usage experience, a large majority had two or more years of mba experience, and almost half the participants perceived themselves as highly experienced. 4.2. sample adequacy and common method bias exploratory factor analysis was used to assess sample adequacy and common method bias. kaiser-meyer-olkin (kmo; cf. sharma, 1996) and harmon’s single-factor tests (cf. podsakoff, mackenzie, lee, & podsakoff, 2003) were conducted. the kmo value was 0.809 (kmo >0.8), indicating that the sample adequacy is good. harmon’s single-factor test showed that there was no maximum variance explained by a single factor 4.3. measurement model the initial results of the measurement model, that is, the model that contains all the items included in the research model, did not meet the suggested thresholds. to refine the model, the suggestions of the modification indices in lisrel 9.30 were applied (cf. jöreskog & sörbom, 1993). this resulted in three factors (i.e., usa 4, usa 5, and rel 3) being eliminated. the results of the final measurement model with standardized factor loadings and t-values are presented in fig. 2 observed variables are represented by rectangles; the standardized factor loading values are indicated before the slashes and the tvalues after the slashes. the final measurement model shows that x2 = 55.07 (with 44 degrees of freedom). the x2/df ratio equals 1.25 (x2/ df > 2), which is considered a good fit (cf. jöreskog, olsson, & wallentin, 2016). the root mean square error of approximation (rmsea) is 0.0428 (rmsea > 0.8), which indicates good fit (cf. bagozzi & yi, 1988). the results of the goodness of fit index, normed fit index, non-normed fit index, and comparative fit index were all >0.9, which is the recommended threshold (cf. jöreskog et al., 2016). table 1 shows that the overall fit indices of the measurement model meet the recommended values. cfa was used to measure the reliability, convergent validity, and discriminant validity of the measurement model. the current study uses two tests to assess reliability: (1) squared multiple correlations (smc), that is, the degree to which the observed variable’s variance is m. nourallah et al. / financial services review 29 (2021) 147–167 155 explained by a latent variable, and (2) composite reliability (cr), to assess the internal consistency (hair et al., 2014). cr was calculated from the squared sum of factor loadings (li) for each latent variable and for the sum of the error variance terms for the latent variables, as shown in eq. (1). table 2 shows that the smcs of all observed variables are higher than 0.5, except for usa 1 and res 1, which are below the cutoff value. table 2 indicates that the cr values are above 0.6 for all latent variables. fig. 2. the results of the measurement model with standardized factor loadings and t-values. x2 = 55.07, p-value = 0.12243, rmsea = 0.042; rmsea = root mean square error of approximation. 156 m. nourallah et al. / financial services review 29 (2021) 147–167 cr ¼ o n i¼1 li � �2 o n i¼1 li � �2 þ o n i¼1 ei � � (1) to assess convergent validity, this study used the average variance extracted (ave), standardized factor loadings, and t-values. ave was computed from the mean variance of the item loadings on a latent variable, as shown in eq. (2): ave ¼ o n i¼1 li 2 n (2) in table 2, the computations indicate that ave is above 0.5 and that all standardized factor loadings exceed 0.6 (cf. hair et al., 2014). all the t-values are significant. to assess discriminant validity, a confidence interval of 62 standard errors around the standardized correlations between latent variables was calculated based on lisrel output (cf. hansen, samuelsen, & sallis, 2013). the calculations indicated that the confidence interval was within the acceptable range, that is, not more than 1 or less than –1. the measurement purification was confirmed by the good results of assessing the goodness of fit (cf. jöreskog et al., 2016) and by the reliability and validity of the variables (cf. fornell & larcker, 1981). overall, it can be assumed that the reliability (cf. bagozzi & yi, 1988; hair et al., 2014), convergent validity, and discriminant validity are good (cf. fornell & larcker, 1981). 4.4. testing the research model sem was performed using lisrel 9.30 (using maximum likelihood and covariance matrices) to test whether the empirical data support the research model. all fit indices correspond to the recommended values (cf. jöreskog et al., 2016). the calculations indicate that there are five significant relationships (p< .01), while three hypotheses are not supported. table 1 the fit indices of the measurement model fit indices result x2/df 1.25 root mean square error of approximation (rmsea) 0.042 goodness of fit index (gfi) 0.942 normed fit index (nfi) 0.943 non-normed fit index (nnfi) 0.978 comparative fit index (cfi) 0.986 m. nourallah et al. / financial services review 29 (2021) 147–167 157 table 3 presents the standardized loadings, t-values, hypothesis outcomes, and fit indices of the model. the structural model shows that usability is directly related to responsiveness (in line with h1) and customer satisfaction (in line with h2), and that responsiveness and customer satisfaction are directly related to loyalty (in line with h6 and h7, respectively). in this sense, an indirect relationship appears between usability and loyalty (see fig. 3). it should be mentioned that there is a direct relationship between usability and reliability (in line with h3), but not between reliability and loyalty (in contrast to h8). in contrast to h4, there is no relationship between responsiveness and customer satisfaction, and in contrast to h5, there is no relationship between reliability and customer satisfaction. table 2 standardized factor loading, t-value, smc, ave, and cr for the measurement model latent variables observed variables standardized factor loading tvalue smc ave cr usability usa 1 0.69 9.09 0.48 0.64 0.84 usa 2 0.88 12.84 0.78 usa 3 0.82 11.54 0.68 responsiveness res 1 0.66 8.20 0.44 0.74 0.85 res 2 0.74 9.43 0.55 res 3 0.78 10.15 0.62 customer satisfaction sat 1 0.84 11.14 0.72 0.69 0.82 sat 2 0.82 10.68 0.63 reliability rel 1 0.92 11.47 0.84 0.53 0.77 rel 2 0.80 9.89 0.64 loyalty loy 1 0.72 8.22 0.51 0.52 0.68 loy 2 0.72 8.26 0.53 note: smc = squared multiple correlations; ave = average variance extended; cr = composite reliability. table 3 structural model results hypothesis standardized loading t-value outcome h1 usability ! responsiveness 0.70 6.26* supported h2 usability ! customer satisfaction 0.49 3.22* supported h3 usability ! reliability 0.58 6.49* supported h4 responsiveness ! customer satisfaction 0.13 0.98 not supported h5 reliability ! customer satisfaction 0.16 1.57 not supported h6 responsiveness ! loyalty 0.44 3.27* supported h7 customer satisfaction ! loyalty 0.37 2.77* supported h8 reliability ! loyalty 0.37 0.28 not supported x2/df = 1.21, rmsea= 0.038, p= .15, gfi = 0.942, nfi = 0.933, nnfi = 0.982, cfi = 0.987 note: rmsea = root mean square error of approximation; gfi = goodness of fit index; nfi = normed fit index; nnfi = non-normed fit index; cfi = comparative fit index. *p-value < 0.01. 158 m. nourallah et al. / financial services review 29 (2021) 147–167 5. discussion 5.1. conclusion this study investigates how ybcs perceive the relationships between the usability, responsiveness, customer satisfaction, and reliability antecedents and loyalty in the context of mbas. based on the empirical results, it can be argued that usability has a direct relationship with responsiveness, customer satisfaction, and reliability and an indirect relationship with loyalty via responsiveness and customer satisfaction. the finding that usability has an indirect relationship with loyalty through customer satisfaction is in line with the findings of casaló et al. (2008) and flavian et al. (2006). responsiveness is significantly related to loyalty, but not to customer satisfaction. the latter finding was not as hypothesized, but is in line with the findings of kassim and asiah abdullah (2010) and munusamy et al. (2010). these results draw attention to the argument of fornell (1992, p. 7) that “loyal customers are not necessarily satisfied customers.” as a consequence, the lack of relationship between responsiveness and customer satisfaction might cause mbas to lose ybcs in the long term. several previous studies (arvidsson, 2014; bauer et al., 2006; calisir & gumussoy, 2008; raza et al., 2015) find a significant relationship between reliability and customer satisfaction. however, as in the studies of munusamy et al. (2010) and wen and hilmi (2011), our empirical results do not significantly support this relationship, calling into question whether reliable mbas can increase the satisfaction of ybcs. this lack of relationship could be attributable to ybcs’ perceptions of reliability in the mba context, and to ybcs’ search for more than just a reliable mba, for example, a usable one. previous studies have stressed that reliability might not be as important for ybcs as for older bank customers (zhou at al., 2010). this also corresponds to our finding that no significant relationship exists between reliability and loyalty, which is in line with the conclusion of lin (2012), who find the same lack of relationship in mobile retailing. it is also possible that the two dimensions of servqual used here, that is, responsiveness and reliability, have different roles regarding ybc experience of mbas. it was no surprise to find a significant relationship between customer satisfaction and loyalty in the mba context. the more satisfied ybcs are, the more loyal they could be. it is claimed that ybcs, who will likely be significant for the future of financial services, prefer fintech companies (gomber et al., 2018) and tend to change banks more than any other age group (accenture, 2015). the exclusive offering of financial services in traditional banks will likely weaken due to the attempts of fintech companies to offer and promote improved fig. 3. usability–loyalty model of young bank customers (ybcs) on mobile bank applications (mbas). m. nourallah et al. / financial services review 29 (2021) 147–167 159 services (nicoletti, 2017). therefore, cultivating the loyalty of ybcs is considered a top priority for traditional banks. 5.2. theoretical and practical implications the present findings have theoretical as well as practical implications. compared with previous studies in the banking sector that highlighted service quality as a key to customer satisfaction and loyalty (e.g., kassim & asiah abdullah, 2010; santouridis & trivellas, 2010), we argue that in today’s digital world, usability can also ensure satisfied and loyal customers. in other words, it can be concluded that the higher the usability a certain mba offers, the more satisfied ybcs will be. this is of interest because our empirical findings regarding ybcs’ perceptions in the mba context demonstrate that both responsiveness and customer satisfaction are directly related to loyalty. usability is a relatively unstudied phenomenon in the financial services context, but the rise of fintech has drawn attention to investigations of usability-related issues. previous studies have acknowledged both ease of use and usefulness (mohammadi, 2015), and financial services studies such as those of casaló et al. (2007, 2008) and flavian et al. (2006) suggest that ease of use is likely the proper way to articulate usability. the current study is in line with this suggestion, because ease of use was found to represent a single usability construct. that banks have made the largest information technology (it) investments across all industries (puschmann, 2017) means that it-related costs represent a significant percentage of bank expenditures. the results of the current study might be used to prioritize such it investments, especially those related to usability, because three usability attributes were found to be particularly important: it should be (1) easy to use the mba the first time, (2) easy to find information, and (3) easy to navigate the mba. mobile application technology represents a promising opportunity for traditional banks and fintech companies, because a mobile application can be an independent financial service provider, for example, an mob. this is unique compared with the earlier versions of mobile banking. traditional banks need to take this development into consideration because ybcs can easily move to other types of financial service providers. 5.3. limitations and future research some limitations of this study can be seen as potential areas for future research. the study was conducted in a specific country and the number of responses was limited. it is therefore suggested that cross-cultural studies be conducted in the future, as cultural differences represent a crucial factor in the banking sector, and that more respondents be included. additional studies are also important because of the general limitations of questionnaire research (cf. sharma & sidhu, 2001), including social desirability bias when data are selfreported and the risk of measuring respondents’ recalled rather than “lived” perceptions. another suggestion is accordingly to conduct “big data” studies focusing on text conversations in social media. 160 m. nourallah et al. / financial services review 29 (2021) 147–167 raza et al. (2015) find that reliability is related to customer satisfaction, which was not supported by the current study. a suggested area for future studies would accordingly be to explore additional aspects of reliability in the mba context. it is also recommended that further studies should cover related issues such as privacy and security. notes 1 including four large private banks: deutsche bank, dresdner bank, hypoverein bank, and commerzbank. 2 since some respondents use more than one mba, the original questionnaire asked those to answer part ii based on their experience on the main mba they use. m. nourallah et al. / financial services review 29 (2021) 147–167 161 the appendix: the final questionnaire part i. background including demographic variables variables frequency (number) frequency (%) cumulative (%) age 18–23 years 105 71.9% 71.9% 24–29 years 41 28.1% 100% total 146 100% gender male 63 43.1% 43.1% female 81 55.5% 98.6% prefer not to say 2 1.4% 100% total 146 100% how many mbas do you use? 1 55 37.7% 37.7% 2 42 28.7% 66.4% 3 40 27.4% 93.8% 4 or more 9 6.2% 100% total 146 100% how long have you usedmbas? 1 year–under 2 years 14 9.6% 9.6% 2 years–under 3 years 33 22. 6% 32.2% 3 years–under 4 years 35 23.9% 56.1% 4 years or more 64 43.9% 100% total 146 100% part ii. usability, responsiveness, customer satisfaction, reliability, and loyalty2 usability usa 1 it was easy to use the mba when i used it for the first time. casaló et al. (2008) usa 2 it is easy to find the information i need from the mba. usa 3 it is easy to navigate in the mba. usa 4 it is easy to carry out transactions in the mba. usa 5 transactions can be carried out quickly in the mba. responsiveness res 1 the mba responds quickly to my questions. lin (2013) and malaquias and hwang (2019) res 2 the different communication channels in the mba help me to solve my problems. res 3 the mba provides opportunities to ask for help. customer satisfaction sat 1 the mba always meets my expectations. aydin and özer (2005)sat 2 i am very pleased with the mba. reliability rel 1 it is reliable to transfer money in the mba. lin (2013) rel 2 i can trust that the account information in the mba is correct. rel 3 it is reliable to pay bills in the mba. loyalty loy 1 i am committed to the mba. chaudhuri and holbrook (2001) loy 2 i carry out all my banking transactions via the mba. wirtz et al. (2007) note: mba = mobile bank applications. 162 m. nourallah et al. / financial services review 29 (2021) 147–167 acknowledgment a preliminary version of this study was presented at the 10th nordic workshop on relationship dynamics – nord2018, karlstad, sweden, september 19–21, 2018, under the title: “mobile bank applications: an empirical study of young bank customers.” the authors would like to thank the conference participants for their valuable comments. also, the authors would like to thank the two anonymous reviewers for their helpful comments. references accenture. (2015). banking shaped by the customer. (available at https://www.accenture.com/us-en/;/media/ accenture/conversion-assets/microsites/documents17/accenture-2015-north-america-consumer-banking-survey.pdf). akturan, u., & tezcan, n. (2012). mobile banking adoption of the youth market: perceptions and intentions. marketing intelligence & planning, 30, 444-459. amin, m. (2016). internet banking service quality and its implication on e-customer satisfaction and e-customer loyalty. international journal of bank marketing, 34, 280-306. arvidsson, n. (2014). consumer attitudes on mobile payment services–results from a proof of concept test. international journal of bank marketing, 32, 150-170. aydin, a. e., & akben selcuk, e. (2019). an investigation of financial literacy, money ethics and time preferences among college students: a structural equation model. international journal of bank marketing, 37, 880900. aydin, s., & özer, g. (2005). national customer satisfaction indices: an implementation in the turkish mobile telephone market.marketing intelligence & planning, 23, 486-504. awwad, m., & awad neimat, b. (2010). factors affecting switching behavior of mobile service users: the case of jordan. journal of economic and administrative sciences, 26, 27-51. bagozzi, r. p., & yi, y. (1988). on the evaluation of structural equation models. journal of the academy of marketing science, 16, 74-94. bauer, h. h., falk, t., & hammerschmidt, m. (2006). etransqual: a transaction process-based approach for capturing service quality in online shopping. journal of business research, 59, 866-875. benlian, a., & hess, t. (2011). the signaling role of it features in influencing trust and participation in online communities. international journal of electronic commerce, 15, 7-56. berraies, s., ben yahia, k., & hannachi, m. (2017). identifying the effects of perceived values of mobile banking applications on customers. international journal of bank marketing, 35, 1018-1038. bhattacherjee, a. (2001). understanding information systems continuance: an expectation-confirmation model. mis quarterly, 25, 351-370. brislin, r. w. (1970). back-translation for cross-cultural research. journal of cross-cultural psychology, 1, 185-216. broderick, a. j., & vachirapornpuk, s. (2002). service quality in internet banking: the importance of customer role.marketing intelligence & planning, 20, 327-335. calisir, f., & gumussoy, c. a. (2008). internet banking versus other banking channels: young consumers’ view. international journal of information management, 28, 215-221. carlander, a., gamble, a., garling, t., hauff, j. c., johansson, l., & holmen, m. (2018). the role of perceived quality of personal service in influencing trust and satisfaction with banks. financial services review, 27, 83-98. casaló, l. v., flavián, c., & guinalı́u, m. (2007). the role of security, privacy, usability and reputation in the development of online banking. online information review, 31, 583-603. casaló, l. v., flavián, c., & guinalı́u, m. (2008). the role of perceived usability, reputation, satisfaction and consumer familiarity on the website loyalty formation process. computers in human behavior, 24, 325-345. chakraborty, s., & sengupta, k. (2013). an exploratory study on determinants of customer satisfaction of leading mobile network providers – case of kolkata. india. journal of advances in management research, 10, 279-298. m. nourallah et al. / financial services review 29 (2021) 147–167 163 chaudhuri, a., & holbrook, m. b. (2001). the chain of effects from brand trust and brand affect to brand performance: the role of brand loyalty. journal of marketing, 65, 81-93. davis, f. d. (1989). perceived usefulness, perceived ease of use, and user acceptance of information technology. mis quarterly, 13, 319-340. demirguc-kunt, a., klapper, l., singer, d., ansar, s., & hess, j. (2018). the global findex database 2017: measuring financial inclusion and the fintech revolution. the world bank. (available at https://elibrary. worldbank.org/doi/abs/10.1596/978-1-4648-1259-0). flavian, c., guinalı́u, m., & gurrea, r. (2006). the role played by perceived usability, satisfaction and consumer trust on website loyalty. information and management, 43, 1-14. fornell, c. (1992). a national customer satisfaction barometer: the swedish experience. journal of marketing, 56, 6-21. fornell, c., & larcker, d. f. (1981). evaluating structural equation models with unobservable variables and measurement error. journal of marketing research, 18, 39-50. foscht, t., maloles, c., schloffer, j., chia, s. l., & “jay” sinha, i. (2010). banking on the youth: the case for finer segmentation of the youth market. young consumers, 11, 264-276. gimpel, h., rau, d., & röglinger, m. (2018). understanding fintech start-ups–a taxonomy of consumer-oriented service offerings. electronic markets, 28, 245-264. gomber, p., koch, j.-a., siering, m., & siering, m. (2017). digital finance and fintech: current research and future research directions. journal of business economics, 87, 537-580. gomber, p., kauffman, r. j., parker, c., & weber, b. w. (2018). on the fintech revolution: interpreting the forces of innovation, disruption, and transformation in financial services. journal of management information systems, 35, 220-265. grinyer, a. (2009). the anonymity of research participants: assumptions, ethics, and practicalities. pan-pacific management review, 12, 49-58. gumussoy, c. a. (2016). computers in human behavior usability guideline for banking software design. computers in human behavior, 62, 277-285. hair, j. f., black, w. c., babin, b. j., & anderson, r. e. (2014).multivariate data analysis.malaysia: person. hanafizadeh, p., behboudi, m., koshksaray, a. a., jalilvand, m., & tabar, s. (2014). mobile-banking adoption by iranian bank clients. telematics and informatics, 31, 62-78. hansen, h., m. samuelsen, b., & sallis, j. e. (2013). the moderating effects of need for cognition on drivers of customer loyalty. european journal of marketing, 47, 1157-1176. ho, c.-i., & lee, y.-l. (2007). the development of an e-travel service quality scale. tourism management, 28, 1434-1449. hoehle, h., & venkatesh, v. (2015). mobile application usability: conceptualization and instrument development.mis quarterly, 39, 435-471. hussien, m. i., & aziz, r. a. e. (2013). investigating e-banking service quality in one of egypt’s banks: a stakeholder analysis. the tqm journal, 25, 557-576. iberahim, h., mohd taufik, n. k., mohd adzmir, a. s., & saharuddin, h. (2016). customer satisfaction on reliability and responsiveness of self service technology for retail banking services. procedia economics and finance, 37, 13-20. international organization for standardization. (1998). ergonomic requirements for office work with visual display terminals – part 11: guidance on usability. (document no. 9241–11). jöreskog, k. g., & sörbom, d. (1993). lisrel 8: structural equation modeling with the simplis command language. chicago, il: scientific software international. jöreskog, k. g., olsson, u. h., & wallentin, f. y. (2016). multivariate analysis with lisrel. basel, switzerland: springer. jun, m., & palacios, s. (2016). examining the key dimensions of mobile banking service quality: an exploratory study. international journal of bank marketing, 34, 307-326. kandampully, j., zhang, t. (christina), & bilgihan, a. (2015). customer loyalty: a review and future directions with a special focus on the hospitality industry. international journal of contemporary hospitality management, 27, 379-414. 164 m. nourallah et al. / financial services review 29 (2021) 147–167 kang, h., lee, m. j., & lee, j. k. (2012). are you still with us? a study of the post-adoption determinants of sustained use of mobile-banking services. journal of organizational computing and electronic commerce, 22, 132-159. karatepe, o. m. (2011). service quality, customer satisfaction and loyalty: the moderating role of gender. journal of business economics and management, 12, 278-300. kassim, n., & asiah abdullah, n. (2010). the effect of perceived service quality dimensions on customer satisfaction, trust, and loyalty in e-commerce settings: a cross cultural analysis. asia pacific journal of marketing and logistics, 22, 351-371. kaur, a., & medury, y. (2011). impact of the internet on teenagers’ influence on family purchases. young consumers, 12, 27-38. killins, r. n. (2017). the financial literacy of generation y and the influence that personality traits have on financial knowledge: evidence from canada. financial services review, 26, 143-165. koenig-lewis, n., palmer, a., & moll, a. (2010). predicting young consumers’ take up of mobile banking services.marketing intelligence and planning, 28, 410-432. laforet, s., & li, x. (2005). consumers’ attitudes towards online and mobile banking in china. international journal of bank marketing, 23, 362-380. larsson, a., & viitaoja, y. (2017). building customer loyalty in digital banking. international journal of bank marketing, 35, 858-877. laukkanen, t. (2007). internet vs mobile banking: comparing customer value perceptions. business process management journal, 13, 788-797. lee, k. c., & chung, n. (2009). understanding factors affecting trust in and satisfaction with mobile banking in korea: a modified delone and mclean’s model perspective. interacting with computers, 21, 385-392. lee, y., & kozar, k. a. (2012). understanding of website usability: specifying and measuring constructs and their relationships. decision support systems, 52, 450-463. lee, y. k., park, j. h., chung, n., & blakeney, a. (2012). a unified perspective on the factors influencing usage intention toward mobile financial services. journal of business research, 65, 1590-1599. liébana-cabanillas, f., alonso-dos-santos, m., soto-fuentes, y., & valderrama-palma, v. a. (2017). unobserved heterogeneity and the importance of customer loyalty in mobile banking. technology analysis & strategic management, 29, 1015-1032. lin, h.-h. (2012). the effect of multi-channel service quality on mobile customer loyalty in an online-and-mobile retail context. the service industries journal, 32, 1865-1882. lin, h. f. (2013). determining the relative importance of mobile banking quality factors. computer standards & interfaces, 35, 195-204. lin, h. h., & wang, y. s. (2006). an examination of the determinants of customer loyalty in mobile commerce contexts. information & management, 43, 271-282. lu, m. t., tzeng, g. h., cheng, h., & hsu, c. c. (2015). exploring mobile banking services for user behavior in intention adoption: using new hybrid madm model. service business, 9, 541-565. luarn, p., & lin, h. h. (2005). toward an understanding of the behavioral intention to use mobile banking. computers in human behavior, 21, 873-891. lundell, s. (2019). lunar way startar bank i sverige – ska erbjuda både bolån och kreditkort. (available at https://www.breakit.se/artikel/21392/lunar-way-startar-bank-i-sverige-ska-erbjuda-bade-bolan-och-kreditkort). malaquias, r. f., & hwang, y. (2019). mobile banking use: a comparative study with brazilian and u.s. participants. international journal of information management, 44, 132-140. mallat, n., rossi, m., & tuunainen, v. k. (2004). mobile banking services. communications of the acm, 47, 42-46. marimon, f., petnji yaya, l. h., & casadesus fa, m. (2012). impact of e-quality and service recovery on loyalty: a study of e-banking in spain. total quality management & business excellence, 23, 769-787. mohammadi, h. (2015). a study of mobile banking usage in iran. international journal of bank marketing, 33, 733-759. moorthy, k., chee, l. e., yi, o. c., ying, o. s., woen, o. y., & wei, t. m. (2017). customer loyalty to newly opened cafés and restaurants in malaysia. journal of foodservice business research, 20, 525-541. m. nourallah et al. / financial services review 29 (2021) 147–167 165 munusamy, j., chelliah, s., & mun, h. w. (2010). service quality delivery and its impact on customer satisfaction in the banking sector in malaysia. international journal of innovation, management and technology, 1, 398-404. nicoletti, b. (2017). future of fintech. basingstoke, uk: palgrave macmillan. nielsen, j. (1994). usability engineering. san francisco, ca: ap professional. nourallah, m., strandberg, c., & öhman, p. (2019). understanding the relationship between trust and satisfaction on mobile bank application. in 3rd international conference on e-commerce, e-business and e-government, in lyon, france. june 18-21, 2019. oliveira, t., thomas, m., baptista, g., & campos, f. (2016). mobile payment: understanding the determinants of customer adoption and intention to recommend the technology. computers in human behavior, 61, 404414. parasuraman, a., berry, l. l., & zeithaml, v. a. (1988). servqual: a multi item scale for measuring consumer perception of service. journal of retailing, 64, 12-40. podsakoff, p. m., mackenzie, s. b., lee, j. y., & podsakoff, n. p. (2003). common method biases in behavioral research: a critical review of the literature and recommended remedies. the journal of applied psychology, 88, 879-903. pohlmann, j. t. (2004). use and interpretation of factor analysis in the journal of educational research: 19922002. the journal of educational research, 98, 14-23. puschmann, t. (2017). fintech. business & information systems engineering, 59, 69-76. raza, a. s., jawaid, s. t., & hassan, a. (2015). internet banking and customer satisfaction in pakistan. qualitative research in financial markets, 7, 24-36. saleem, m. a., zahra, s., ahmad, r., & ismail, h. (2016). predictors of customer loyalty in the pakistani banking industry: a moderated-mediation study. international journal of bank marketing, 34, 411-430. sampaio, c. h., ladeira, w. j., & santini, f. d. o. (2017). apps for mobile banking and customer satisfaction: a cross-cultural study. international journal of bank marketing, 35, 1133-1151. sanchez-franco, m. j., & rondan-cataluña, f. j. (2010). virtual travel communities and customer loyalty: customer purchase involvement and web site design. electronic commerce research and applications, 9, 171-182. santouridis, i., & trivellas, p. (2010). investigating the impact of service quality and customer satisfaction on customer loyalty in mobile telephony in greece. the tqm journal, 22, 330-343. scornavacca, e., & hoehle, h. (2007). mobile banking in germany: a strategic perspective. international journal of electronic finance, 1, 304-320. shaikh, a. a., & karjaluoto, h. (2015). mobile banking adoption: a literature review. telematics and informatics, 32, 129-142. shaikh, a. a., & karjaluoto, h. (2019). mobile financial services. in a. a. shaikh & h. karjaluoto (eds.), marketing and mobile financial services: a global perspective on digital banking consumer behaviour (pp. 1-26). new york, ny: routledge. sharma, s. (1996). applied multivariate techniques. new york, ny: wiley. sharma, d. s., & sidhu, j. (2001). professionalism vs commercialism: the association between non-audit services (nas) and audit independence. journal of business finance & accounting, 28, 563-594. sindwani, r., & goel, m. (2015). the impact of technology based self service banking dimensions on customer satisfaction. international journal of business information systems strategies, 4, 1-13. srinivasan, s. s., anderson, r., & ponnavolu, k. (2002). customer loyalty in e-commerce: an exploration of its antecedents and consequences. journal of retailing, 78, 41-50. strandberg, c., wahlberg, o., & öhman, p. (2012). challenges in serving the mass affluent segment: bank customer perceptions of service quality.managing service quality: an international journal, 22, 359-385. strandberg, c., wahlberg, o., & öhman, p. (2015). effects of commitment on intentional loyalty at the personto-person and person-to-firm levels. journal of financial services marketing, 20, 191-207. suoranta, m., & mattila, m. (2004). mobile banking and consumer behaviour: new insights into the diffusion pattern. journal of financial services marketing, 8, 354-366. 166 m. nourallah et al. / financial services review 29 (2021) 147–167 sum chau, v., & ngai, l. w. l. c. (2010). the youth market for internet banking services: perceptions, attitude and behaviour. journal of services marketing, 24, 42-60. sun, y., wang, m., & wang, n. (2015). technology leadership, brand equity, and customer loyalty in mobile banking: moderating role of need for uniqueness. digit 2015 proceedings, 14. tam, c., & oliveira, t. (2017). literature review of mobile banking and individual performance. international journal of bank marketing, 35, 1044-1067. tan, e., & lau, j. l. (2016). behavioural intention to adopt mobile banking among the millennial generation. young consumers, 17, 18-31. thakur, r. (2014). what keeps mobile banking customers loyal? international journal of bank marketing, 32, 628-646. verhoef, p. c., & donkers, b. (2005). the effect of acquisition channels on customer loyalty and cross-buying. journal of interactive marketing, 19, 31-43. wen, c. h., & hilmi, m. f. (2011). exploring service quality, customer satisfaction and customer loyalty in the malaysian mobile telecommunication industry. in 2011 ieee colloquium on humanities, science and engineering in penang, malaysia wirtz, j., mattila, a. s., & lwin, m. o. (2007). how effective are loyalty reward programs in driving share of wallet? journal of service research, 9, 327-334. yeh, c. h., wang, y. s., & yieh, k. (2016). predicting smartphone brand loyalty: consumer value and consumer-brand identification perspectives. international journal of information management, 36, 245-257. yoon, c. (2010). antecedents of customer satisfaction with online banking in china: the effects of experience. computers in human behavior, 26, 1296-1304. zhou, t., lu, y., & wang, b. (2010). integrating ttf and utaut to explain mobile banking user adoption. computers in human behavior, 26, 760-767. m. nourallah et al. / financial services review 29 (2021) 147–167 167 pii: s1057-0810(97)90025-1 financial services review, 6(2): 151-154 copyright 0 1997 by jai f’ress inc. issn: 1057-0810 all rights of reproduction in any form reserved. book, software and web site reviews the financial services and financial institutions j. kimball dietrich upper saddle river, nj: prentice hall; 1996 (isbn o-02-329545-7) reviewed by: james marchand, professor of finance, westminster college, salt lake city, ut this text represents one of the first texts for a course on financial services industry. this is becoming more frequent in finance and financial service and/or planning majors. there is much good to say about the text. however, like most first editions, there is much room for change and improvement. dietrich’s text is one of the most comprehensive texts recently published. it evens the width and breadth of depository and nondepository intermediaries using accounting, eco nomics, and descriptive statistics as necessary, along with the modem analytic and concep tual tools of finance. while the book has these virtues, it should not overwhelm most undergraduate students. it is written with a level and style that should make it easily read able for both undergraduates and mba’s. the level of the content is probably appropriate for 85-90% of typical undergradu ates and mba’s, i find the level above the rose, kolari, and frazer (1993) text. how ever, the level, especially analytically, is below recent entrant texts such as anthony saunders’ (1997) financial institutions management, gardner and mills’ (1993) manag ing financial institutions, or kolb and rodriquez’s (1996) financial institutions. the text also has the more traditional balancing between the older bread and butter financial institutions management topics (asset, liability, liquidity management, regulation, etc.) and the new management techniques including risk measurement and management, derivatives and off balance sheet activities. however, the more modem topics are rele gated to the end of the text. it is also notable that after writing a 777-page text, the author has had the fortitude to write the instructors manual. this manual is well done and should lighten the instructors burden. its usefulness is indicated by the fact that the question or problem is stated first before giving the answer to end-of-chapter questions and problems. hence, the instructor does not need to take the text to class. the manual also offers many good suggestions for teaching, class organization, and short projects or homework. most of my concerns for improvement center on pedagogy. terminology is important to understanding most disciplines including finance. unfortunately, the text does not help the student learn terminology section by section or chapter by chapter. there is no glossary to conclude the text. like most recent core course texts, each chapter needs to emphasize terminology with a beginning or end-of-chapter listing of important terms, text highlight 1.52 financial services review 6(2) 1997 ing, or marginal notes. the end-of-chapter discussion questions and problems help the stu dent review the basic ideas of the chapter. however, most only require that the student regurgitate or paraphrase the text. the text would be better and more widely usable if addi tional challenging, “thinking” questions were added. in this vain for chapters which teach aspects and application of formal analysis, there should be more problems requiring calcu lation and describing or thinking about the meaning of the calculations. some of the analytic figures should also be improved by adding numerical scales to the axis and coordinate points on the functional lines. using grids would support these sug gested additions. lastly, multiple colors or shades of black/gray would enhance all of the previously suggested graphical changes. finally, there is a concern that arises with many textbooks. many of the tables of descriptive statistics are dated. as this textbook is long and evolved over a considerable period of time, it is easy to understand how this might occur. however, in this world of modem technology and computer information retrieval, it would seem possible and rela tively low cost to have more updated data, in the instructors manual if nothing else. all in all if you are teaching a fmancial services course, there is enough good here to give dietrich’s book a try, particularly if you want comprehensiveness and readability. if some of my pedagogical suggestions are adopted for the second edition, you may be on the way to using a classic text. it would not hurt to have it around as a reference book as well. references gardner, m., & mills, d. (1993). ~uffuging~nanciul i~tjt~tions, 3rd ed. dallas, tx: dryden press. kolb, r., & rodriquez, r. (1996). financiue institutions. cambridge, ma: blackwell. rose, p., kolari, j., & fraser, d. (1993). financial institutions understanding and managingfinan cial services, 4th ed. homewood, il: irwin. saunders, a. (1997). fhancial institutions management: a modern perspective, 2nd ed. chicago, il: irwin. american risk and insurance association (aria) web site reviewed by: j. tim query, department of risk management and insurance, university of georgia the american risk and insurance association (aria) is the premier professional associ ation of insurance scholars and other thoughtful insurance and risk management profes sionals. as stated in its home page (http:nwww.aria.org), aria emphasizes research relevant to the operational concerns and functions of insurance professionals, and pro vides resources, info~tion and support on important insurance issues. goals also include the expansion and improvement of academic inst~ction to students of risk man agement and insurance. several topics of interest to association members, as well as non-members, are acces sible on the home page. for those active in the job market, a listing of positions available is updated frequently, along with the posted resumes of candidates seeking positions. announcements about aria officers and board members, grant programs, and calls for papers are obtainable at this site, as well as info~ation on aria’s annual meeting. the book, software and web site reviews 153 annual meeting information is somewhat dated in that the program for the 1996 meeting is listed, but no information on the upcoming 1997 meeting had been posted as of mid april. an on-line application form is available for those interested in becoming a member of the american risk and insurance association. the web site is especially useful for those conducting research in the areas of risk man agement and insurance. information can be accessed on such research periodicals as the journal of risk and insurance, geneva papers on risk and insurance theory, and a new academic journal sponsored by aria, entitled risk management and insurance review. other research-oriented resources available through aria’s home page include a risk and insurance working paper archive, risk and insurance database search engines, and a connection to the home page of the risk theory society. on the teaching side, links to the home pages of leading collegiate risk management and insurance programs are available, an ambitious and very useful work-in-process is the risk and insurance teaching archive. this offering includes lecture notes, risk manage ment case studies, and other resources available to facilitate the delivery of high quality risk management and insurance-related courses. for example, dr. bill rabel, senior vice president of education at the life office management association (loma), has prepared, “a manual on how to manage an effective collegiate program in risk management and insurance.” this paper is available in its entirety and can be downloaded by interested par ties. while some of the suggestions in the paper are specific to the area of risk manage ment and insurance, many ideas are germane for other business disciplines as well. the american risk and insurance association home page also provides links to dif ferent risk and insurance webs. these include such insurance industry associations as the life insurance management association, society of actuaries, insurance information institute, and the chartered property casualty underwriters (cpcu) society. this aria-sponsored world wide web site is provided by insweb corporation. while there is little in the area of graphics, the home page of the american risk and insur ance association is a “must” for those seeking information on the academic side of risk management and insurance. the mutual fund investor’s centertm web site reviewed by: douglas r. kahl, professor of finance, the university of akron the mutual fund education alliance, established in 1971, is a national trade association of direct marketed mutual funds. the alliance sponsors the mutual fund investor’s cen ter (http://www.mfea.com) to provide facts and educational information to mutual fund investors. this free site provides a wide range of information and services to mutual fund investors. the site has information about more than 40 mutual fund companies including major shareholder services, a listing of the company’s different funds, and basic company infor mation. there are links to the company’s home page for 30 of the companies. included in the list of covered mutual fund companies are such well known families as the vanguard group, the dreyfus family of funds, and fidelity investments, as well as many less well known families of funds. among the more useful information services is a daily prices sec tion, which allows the user access to the daily fund prices on nearly 1000 direct marketed, no-load or low-load mutual funds. 154 financial services review 6(2) 1997 the site also provides educational articles on such useful topics as the basics of mutual fund investing, risk vs. reward, tax considerations, and how to read a prospec tus. the well written and easy to understand articles provide coverage at a basic introduc tory level. given the level of the presentation and the relatively short length of the articles, the coverage is both thorough and accurate. a special focus section, “women and invest ing: what every woman needs to know,” is devoted entirely to women’s issues. this sec tion includes biographies and commentary by women investment professionals, a woman’s investor kit, and investor survey with results of a few recent studies. only the newsmak erstm section of the site seemed too overly promotional. while some of the commentaries were interesting, i found this section to be less useful than other parts of the site. a service called “fund quicklisttm” provides a system for sorting and identifying funds by such criteria as fund company, sales charges, investment categories, fees and min imum initial investment. as with the daily fund prices, this service covers nearly 1000 direct marketed mutual funds. the menu driven search system is fast, easy to use, and appears to provide accurate screening of the funds. the site also provides a fairly simplistic test for investor personality and risk tolerance. this is followed by a basic table relating investment objectives to types of funds, fund risk, and expected returns. finally, a selection of model portfolios is provided. the model port folios are much less aggressive than i would usually recommend for any of the model sit uations. however, a conservative bias is probably to be expected from an organization like the alliance and in such a general format as a web site. in general, i found this site to be very well done, easy to use, and informative. the site will provide a wealth of basic information for novices in mutual fund investing. in addition, the daily prices and quicklist services would be very useful to most mutual fund investors. the site also contains a useful set of links to the home pages of many of the more popular mutual fund families. ce 1-hour general principles of financial planning, risk and insurance planning, and estate planning afs and fpa members can earn ce credits through financial services review. go to fpajournal.org. to receive one hour of continuing education credit allotted for this exam, you must answer four out of five questions correctly. cfp board recently adopted revisions to several provisions of its ce policies, including changing the minimum number of questions for self-study assessments from 10 to 5 per full ce credit hour. therefore, financial services review ce exams will have 5 questions. ce credit for this issue expires may 31, 2023, subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online only—paper continuing education will not be processed. go to fpajournal.org to take current and past ce (free to afs and fpa members). you may use this page for reference. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. in “the potential benefit of liability management” by liu and blanchett, which of the following factors are positively related to u.s. households carrying debt: a. number of household children. b. home ownership. c. education level. d. all of the above. 2. liu and blanchett concluded that households with longer financial planning horizons are: a. more likely to have debts. b. much less likely to have debts. c. in the similar debt situation compared to the households with short financial planning horizons. d. none of the above. 3. according to liu and blanchett, bad debts are not only expensive, but they may also: a. negatively influence the borrowers’ credit scores. b. hinder their financial and retirement goals. c. cause stress and health issues. d. all of the above. 4. according to “encouraging living will completion using social norms and family benefit” by hussein and james, which of the following is a potential advantage of advanced planning such as with living wills and durable powers of attorney for healthcare? a. achieve personal and family financial goals. b. ensure that patients’ preferences for medical treatment is followed. c. can limit the financial impact of this end-of-life medical care. d. all the above. 5. in hussein and james, which of the following added statements resulted in the greatest increase in intentions to complete a living will document? a. a living will can relieve family members of difficult decisions (family benefit only). b. many people like to have a living will (social norms only). c. many people like to have a living will because it can relieve family members of difficult decisions (family benefit and social norms combined). d. the living will is only used at the end of life if a person cannot be cured (terminally ill) or is permanently unconscious (end of life only) call for papers & proposals virtual conference due june 1, 2021 the academy of financial services 35th annual meeting september 21-22, 2021, virtual conference the academy of financial services will hold its annual conference in conjunction with the fpa be annual conference. this year, we will have a virtual conference. $199 for academics and practitioners and $99 for students. � the afs conference will feature speakers, symposia, and several special sessions. among them, we will introduce a new panel session for phd students, highlighting how to best navigate the job market. � with the generous support of our sponsors, the academy has awarded several best paper awards during past meetings and we anticipate continuing best paper awards in 2021. � we will continue with our emerging scholar award to a current graduate student for promising research work on a paper or poster presented at the conference. � in addition, in 2021, we will continue with the program directors track. the goal is to allow program directors to present and discuss program issues and best practices in a panel environment, such as “working with your university’s foundation”, “capstone course cases: what’s the right content?”, “understanding career paths and student fit” and “developing a passionate program in a box: scholarships, competitions, student organizations”. we welcome other panel topics deemed beneficial to program directors. submission information: research papers and abstracts covering all aspects of individual financial management and education are sought for inclusion in the program. papers in the areas of estate planning, insurance, tax accounting aspects of financial planning, investments, and retirement planning are encouraged. proposals for panel discussions and tutorials devoted to current issues in individual financial management or the practice of financial planning will also be considered for inclusion in the program. several sessions will be registered for continuing education (ce) credit with the cfp® board. please be advised that you must be an afs member in good standing when you. register for the conference. membership has always been required to register for and attend the afs conference. for further information: � go to the afs website at academyfinancial.org that will be frequently updated. � for content questions contact program chair, terrance k. martin jr at terrance.martin@uvu.edu important information � submit proposals here: http://proposalspace.com/calls/d/1306 � submissions are due june 1st � the review period ends on july 15th, 2021 with the selection period and formulation of the agenda estimated to be completed by august 15th, 2021. notice of acceptance as an oral session is targeted for august 31st, 2021. � note that the terms and conditions of this call-for are outlined in the online submission form. � only accepted presentations are included in the subsequent proceedings, which are posted on the afs website. thus, the proceedings publication is refereed in order to accommodate the rules of the american association of intercollegiate schools of business-international (aacsb) on table 2-1 (intellectual contributions). manuscript submissions and style (1) papers must be in english. (2) papers for publication should be sent to the editor: professor stuart michelson, e-mail: smichels@stetson.edu. electronic (email) submission of manuscripts is encouraged, and procedures are discussed below. there is a $100 submission fee payable to the academy of financial services (afs) if at least one of the authors is a member of afs. submission fees should be paid online at academy financial org. if none of the authors is a member of afs, please complete an online membership application form, which can be downloaded at http://academyfinancial.org, and pay online ($225 total; $125 for a one-year membership and $100 submission fee). submission of a paper will be held to imply that it contains original unpublished work and is not being considered for publication elsewhere. the editor does not accept responsibility for damage or loss of papers submitted. upon acceptance of an article, author(s) transfer copyright of the article to the academy of financial services. this transfer will ensure the widest possible dissemination. (3) submission of papers: authors should submit their papers electronically as an e-mail attachment to the editor at smichels@stetson.edu. please send the paper in word format. do not sent pdfs. ensure that the letter ‘l’ and digit ‘1’, and also the letter ‘o’ and digit ‘0’ are used properly, and format your article (tabs, indents, etc.) consistently. do not allow your word processor to introduce word breaks and do not use a justified layout. please adhere strictly to the general instructions below on style, arrangement and, in particular, the reference style of the journal. (4) manuscripts should be double spaced, with one-inch margins, and printed on one side of the paper only. all pages should be numbered consecutively, starting with the title page. titles and subtitles should be short. references, tables, and legends for the figures should be printed on separate pages. (5) the first page of the manuscript, the title page, must contain the following information: (i) the title; (ii) the name(s), title, institutional affiliation(s), address, telephone number, fax number and e-mail addresses of all the author(s) with a clear indication of which is the corresponding author; (iii) at least one classification code according to the classification system for journal articles as used by the journal of economic literature, which can be found at http://www.aeaweb.org/journal/elclasjn.html; in addition, up to five key words should be supplied. (6) information on grants received can be given in a footnote on the title page. (7) the abstract, consisting of no more than 100 words, should appear alone on page 2, titled, abstract. (8) footnotes should be kept to a minimum and should only contain material that is not essential to the understanding of the article. as a rule of thumb, have one or less footnote, on average, per two pages of text. (9) displayed formulae should be numbered consecutively throughout the manuscript as (1), (2), etc. against the right-hand margin of the page. in cases where the derivation of formulae has been abbreviated, it is of great help to the referees if the full derivation can be presented on a separate sheet (not to be published). (10) the financial services review journal (fsr) follows the apa publication manual, 6th edition, style. however, consistent with the current trend followed by other publications in the area of finance, the journal has a very strong preference for articles that are written in the present tense throughout. references to publications should be as follows: “smith (1992) reports that” or “this problem has been studied previously (ho, milevsky, & robinson, 1999).” the author should make sure that there is a strict one-to-one correspondence between the names and years in the text and those on the reference list. the list of references should appear at the end of the main text (after any appendices, but before tables and legends for figures). it should be double spaced and listed in alphabetical order by author’s name. references should appear as follows: books: hawawini, g. & swary, i. (1990). mergers and acquisitions in the u.s. banking industry: evidence from the capital markets. amsterdam: north holland. chapter in a book: brunner, k. & meltzer, a. h. (1990). money supply. in: b. m. friedman & f. h. hahn (eds.), handbook of monetary economics (vol. 1, pp. 357-396). amsterdam: north holland. periodicals: ang, j. s. & fatemi, a. m. (1997). personal bankruptcy costs: their relevance and some estimates. financial services review, 6, 77-96. note that journal titles should not be abbreviated. (11) illustrations will be reproduced photographically from originals supplied by the author; they will not be redrawn by the publisher. please provide all illustrations in quadruplicate (one high-contrast original and three photocopies). care should be taken that lettering and symbols are of a comparable size. the illustrations should not be inserted in the text, and should be marked on the back with figure number, title of paper, and author’s name. all graphs and diagrams should be referred to as figures, and should be numbered consecutively in the text in arabic numerals. illustration for papers submitted as electronic manuscripts should be in traditional form. the journal is not printed in color, so all graphs and illustrations should be in black and white. (12) tables should be numbered consecutively in the text in arabic numerals and printed on separate sheets. any manuscript which does not conform to the above instructions will be returned for the necessary revision before publication. page proofs will be sent to the corresponding author. proofs should be corrected carefully; the responsibility for detecting errors lies with the author. corrections should be restricted to instances in which the proof is at variance with the manuscript. extensive alterations will be charged. reprints of your article are available at cost if they are ordered when the proof is returned. financial services review (issn: 1057-0810) academy of financial services stuart michelson stetson university school of business 421 n. woodland blvd. unit 8398 deland, fl 32723 (address service requested) continuing education credits afs and fpa members can earn ce credits through financial services review. visit https://academyfinancial.org/ce-quizs-for-fsr to learn more about earning continuing education credits. to receive continuing education credit allotted for this exam, you must answer four out of five questions correctly. ce credit for this issue of financial services review is open for a limited time and is subject to any changes dictated by cfp board. afs and fpa offer financial services review ce online-only; paper continuing education will not be processed. please allow 2-3 weeks for credit to be processed and reported to cfp board. 1. of the following investors, who is most likely to increase their ownership of life insurance? a. an investor who has increased their ownership of stocks. b. an investor who has increased their ownership of bonds. c. an investor who has decreased their ownership of bonds. d. an investor who believes an economic recession is imminent. 2. jacob is a financial planner who works with emerging high net worth clients. jacob is interested in segmenting his clients into those who are likely and unlikely to give a bequest. after reading the anderson et al. paper, jacob knows that those who are most likely to give a bequest a. are also more likely to have received an inheritance. b. are married. c. less likely to make charitable donations. d. exhibit a low level of risk tolerance. 3. when variables such as financial satisfaction are accounted for, who is the most likely to report a high level of financial well-being? a. younger individuals. b. being female. c. being employed on a full-time basis. d. holding a college degree or above level of education. 4. when asked, chris will tell you that compared to others, her level of financial knowledge is quite high. given what jeong et al. found in their research study, chris’s financial planner can anticipate that chris i. is financially satisfied. ii. will be more likely to report a high level of well-being. iii. is financially dissatisfied. a. only i b. only iii c. both ii and iii d. both i and ii https://academyfinancial.org/ce-quizs-for-fsr 5. rarely do investors and their financial planners agree about the future performance of mutual funds. research by chrétien and kammoun shows that the larger the disagreement, the greater net fund flows. which of the following fund characteristics is associated with the greatest level of investor disagreement? a. high fund manager turnover. b. the use of a balanced, versus aggressive, portfolio strategy. c. lower manager tenure. d. a fund from a large fund family. pii: 1057-0810(94)90021-3 financial services review, 3(2): 159-165 copyright q 1994 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. abstracts of articles on individual financial management edited by phyllis schiller myers virginia commonwealth university consumption-savings behavior non-marketable assets and households’ portfolio choices: a case study of italy, claudio giraldi and rony hamaui (banca commerciale italiana, economic research and planning department, milan, italy) and nicola rossi (university of modena, italy) this paper analyzes the determinants of aggregate households’ portfolio choices conditional on those components of wealth which can be regarded as highly or totally illiquid because of the existence of substantial transaction costs and/or institutional constraints on agents behavior. the empirical analysis broadly confirms the importance of the above-men tioned institutional aspects in the allocation of financial wealth and provides the means for addressing some interesting questions regarding the behavior of portfolio holders in financial markets. in particular, it is shown that the presence of non-tradable assets changes the risk premia on marketable assets. hence, in cases such as the italian one, where the state is both the issuer of government securities and the collector of social security funds, the positive correlation (due to a common default risk) between returns on government securities and on non-marketable assets could lead to a higher cost of public debt financing. journal of banking and finance, december 1993, 17(6): 1171-l 190. (reprinted with permission of the north-holland publishing company.) mental accounting and outcome contiguity in consumer-borrowing decisions, by d. eric hirst (university of texas at austin), edward j. joyce (university of minnesota) and michael s. schadewald (university of wisconsin at milwaukee). research indicates that decisions are affected by how outcomes are framed. mental accounting is a type of framing in which individuals are hypothesized to form psychological accounts for the costs and benefits of outcomes. prior research has focused on mental accounting’s consequences rather than its determinants. thus, little is known about the processes that underlie mental accounting. this study investigates the role that temporal contiguity (the co-occurrence of multiple outcomes) plays in mental accounting for consumer-borrowing decisions. thaler’s (1985) extension of kahneman and tversky’s (1979) prospect theory was used to predict that consumers will prefer to finance purchases 160 financial services review, 3(2) 1994 of goods with loans whose terms correspond with the life of the good. the results of four experiments involving 13 1 mba students provide support for this prediction. the present study adds to our knowledge of mental accounting by examining the effect of temporal contiguity in the domain of multi-period costs and benefits. it also adds to the consumer behavior literature by examining an important factor affecting debt utilization. organiza tional behavior and human decision processes, april 1994,58: 136-152. (reprinted with permission of organizational behavior and human decision processes.) estate planning and distribution government intervention as a bequest substitute, by michel strawczynski (hebrew university of jerusalem). a subsequent generations model is used in order to characterize consumption alloca tion under the future generation’s income uncertainty. altruistic concerns towards future generations give rise to ‘precautionary bequests’ which act as a hedge on risk. it is shown that given a first-order correlation between mean future income and the present generation’s income, government can provide a pareto improvement through a tax-transfer policy with universal participation. this policy acts as a substitute for precautionary bequests. distribu tional aspects of governmental tax-transfer policy are also discussed. journal of public economics, march 1994,53(3): 477-495. (reprinted with permission of journal of public economics.) financial services delivery and products brokerage commission schedules, by michael j. brennan (university of california, los angeles, and the london business school) and tarun chordia (university of california, los angeles). it is generally optimal for risk-sharing reasons to base a charge for information on the signal realization. when this is not possible, a charge based on the amount of trading, a brokerage commission, may be a good alternative. the optimal brokerage commission schedule is derived for a risk-neutral information seller faced with risk-averse purchasers who may differ in their risk aversion. revenues from the brokerage commission are compared with those from a fixed charge for information and the optimal mutual fund management fee. the journal of finance, september 1993,48(4): 1379-1402. (reprinted with permission of the journal offinance.) implications for financial planning is investing for the long term theory or just mumbo-jumbo?, by p.l. bernstein. in academic models and in wall street, investment “for the long run” presumes the existence of a long-run trend. the notions of “undervaluation” and “overvaluation” implies a “regression to the mean.” the long run in the popular view is a process that smooths the bumps and captures the main trend. in reality, the long run is a complex, ambiguous, and abstracts of articles on individual financial management 161 elusive concept in which volatility as well as liquidity matter. volatility is unimportant only if liquidity is unimportant. if liquidity is important, in the long run, one cannot presume that a volatile investment will regress to the mean. journal of post keynesian economics, spring 1993, 15(3): 387-393. (reprinted with permission of economic literature.) insurance decision and individual risk management the inefficiency of private constant annuities, by tadashi yagi (nagoya university) and yasuyuki nishigaki (yokkaichi university). this article investigates why individuals save to finance consumption during retire ment, and we focus this investigation on the inefficiency of constant annuities. to thoroughly examine this problem, we first derive the demand function for the annuities in the case where the capital market is imperfect and the annuities are constrained to be constant throughout the retirement period. with these constraints, we then show that the individual holds assets not only in the form of actuarial notes but also in the form of monetary wealth. the journal of risk and insurance, september 1993,60(3): 385412. (reprinted with permission of the journal of risk and insurance.) consumer information and decisions to switch insurers, harris schlesinger (university of alabama) and j.-matthias graf von der schulenburg (university of hanover). this article examines the interaction of various factors in an individual’s decision to switch insurers. in particular, expectations about insurer quality attributes as well as search costs and switching costs are modeled as affecting the consumer’s switching decision. data from a 1983 survey of 2,004 german individuals are used to determine consumers’ impres sions about the quality and price of their auto insurance policies. the empirical analysis shows how consumer informedness plays a key role in the switching decision. factors affecting consumer informedness and the sources of consumer information also are exam ined. the journal of risk and insurance, december 1993,60(4): 591-615. (reprinted with the permission of the journal of risk and insurance.) investment performance investment performance over bull and bear markets: fabozzi and francis revisited, by john m. clinebell (university of northern colorado), jan r. squires (southwest missouri state university) and jerry l. stevens (university of richmond). fabozzi and francis present strong evidence in a widely referenced paper that alpha and beta measures do not change over bull and bear markets. test procedures used by fabozzi and francis are replicated in this paper for new samples and for more current periods. unlike the findings of fabozzi and francis, stocks exhibiting statistically different beta measures over bull and bear markets are significantly greater than predicted by chance. this result was especially strong for the 1972 through 1977 market period. quarterly journal of business and economics, autumn 1993,32(4): 14-25. (reprinted by permission of quarterly journal of business and economics.) 162 financial services review, 3(2) 1994 aftermarket support and underpricing of initial public offerings, by paul h. schultz (ohio state university) and mir a. zaman (university of northern iowa). we study the aftermarket for 72 initial public offerings (ipos) using comprehensive trade and quote-change data from every market maker for the first three days of trading. underwriters quote higher bid prices than other market makers for issues that commence trading at or below the offer price. underwriters repurchase large quantities of stock in the aftermarket without risk by overselling the issue by the amount of the overallotment option. if the ipo is hot, the overallotment option is exercised. if not, the short position is covered with aftermarket selling. we discuss several reasons for underwriter support. journal of financial economics, 1994 35: 199-219. (reprinted with permission of the north-holland publishing company.) relative mean-variance effbziency of a given portfolio: an application to mutual fund performance, by shafiqur rahman (portland state university). this research article uses two performance measures based on mean-variance effi ciency to evaluate the performance of managed portfolios. they do not require the specifi cation of the market portfolio and the risk-free rate. as a result, the measurement error attendant in the proxy is resolved. our empirical results show that for the test period, approximately half of the sampled funds performed better than a naive buy-and-hold strategy based on these measures. when compared to other measures, results indicate that only these measures are independent of their risk proxy and as such are unbiased measures, results indicate that only these measures are independent of their risk proxy and as such are unbiased measures of investment performance. the quarterly review of economics and finance, spring 1994,34( 1): 13-24. (reprinted with permission of the quarterly review of econom ics and finance.) investment selection and individual portfolio management portfolio performance of the sdr and reserve currencies: tests using the arch methodology, m. g. papaioannou and t. temel in evaluating their foreign exchange exposure, international investors often compare actual portfolios with those calculated under the assumption that the variability of returns on various currency assets is time invariant. this paper uses autoregressive conditional heteroskedastic models to test that assumption. for major reserve currencies, including the sdr, the authors find evidence that the variances of returns do vary over time and that autoregressive conditional heteroskedastic models that specify changing variances are superior to models that assume constant variance. by incorrectly assuming a constant variability of returns, the error introduced is smaller with the sdr than with any other national currency. international monetary fund staff papers, september 1993, 40(3): 663-679. (reprinted with permission of the journal of economic literature.) abstracts of articles on individual financial management 163 art as an investment: the market for modern prints, james e. pesando (university of toronto) repeat sales of modern prints at auction are used to estimate a semiannual index of prices for the period 1977-1992. as in other studies of art as an investment, prints do not compare favorably to traditional financial assets. there is substantial noise in auction prices, but little or no support for the proposition that some artists command higher prices in certain countries or that masterpieces outperform the market. one puzzle is the continuing tendency for prices realized at certain auction houses to exceed those realized at others: notable, at sotheby’s relative to christie’s in new york. the american economic review, december 1993, 83(5): 1075-1089. (reprinted with permission of the american economic associa tion.) downside risk and investment choice, by k.s. maurice tsi (indiana university), jamshed uppal (catholic university of america) and mark a. white (university of virginia). this paper develops an optimal investment strategy for individuals concerned with avoiding the possibility of realizing returns below a predetermined target level within a prescribed period of time. assuming a brownian motion process, a model is developed which allows computation of the exact probability of failure. the algorithm and associated comparative statics with respect to the mean and standard deviation of returns, target return, time horizon, and risk-free rate of return are likely to have many useful practical applications. the financial review, november 1993,28(4): 585-605. (reprinted with permission of the financial review.) intertemporal asset pricing under knightian uncertainty, by larry g. epstein (university of toronto) and tan wang (university of waterloo). in conformity with the savage model of decision-making, modern asset pricing theory assumes that agents’ beliefs about the likelihoods of future states of the world may be represented by a probability measure. as a result, no meaningful distinction is allowed between risk, where probabilities are available to guide choice, and uncertainty, where information is too imprecise to be summarized adequately by probabilities. in contrast, knight and keynes emphasized the distinction between risk and uncertainty and argued that uncertainty is more common in economic decision-making. moreover, the savage model is contradicted by evidence, such as the ellsberg paradox, that people prefer to act on known rather than unknown or vague probabilities. this paper provides a formal model of asset price determination in which knightian uncertainty plays a role. specifically, we extend the lucas (1978) general equilibrium pure exchange economy by suitably generalizing the representation of beliefs along the lines suggested by gilboa and schmeidler. two principal results are the proof of existence of equilibrium and the characterization of equilibrium prices by an “euler inequality.” a noteworthy feature of the model is that uncertainty may lead to equilibria that are indeterminate, that is, there may exist a continuum of equilibria for given fundamentals. that leaves the determination of a particular equilibrium price process to “animal spirits” and sizable volatility may result. finally, it is argued that empirical investigation of our model is potentially fruitful. econometrica, march 1994, 62(3): 283 322. (reprinted with permission of econometrica.) 164 financial services review, 3(2) 1994 real estate investment probabilistic valuation models and income tax asymmetries with an application to the analysis of passive loss restrictions, by david c. ling (university of florida). this paper develops a modified version of the standard real estate discounted cash flow valuation model that allows the analyst to specify probability distributions, rather than point estimates, on general inflation and rental income for each year of the expected holding period. explicit modeling of rental income uncertainty is especially critical in the presence of restrictions that cause the tax treatment of income-producing property to be asymmetric. this is demonstrated by an analysis of the restrictions on passive activity losses (pal) that were introduced with the passage of the tax reform act of 1986. the probability model results indicate that rational investors can pay significantly more for an income property as a result of the elimination of pal restrictions than would be indicated by the results produced by a standard discounted cash flow “point estimate” model. this is due to the reduction in the skewness of the after-tax returns generated by the property when pal restrictions are eliminated, and the decline in the required equity discount rate that results from the decreased variance of the after-tax returns. the journal of real estate research, spring 1993, 8(2): 205-220. (reprinted with permission of the journal of real estate research.) factors influencing capitalization rates, by brent w. ambrose (the university of wisconsin-milwaukee) and hugh 0. nourse (the university of georgia). this study examines the variations in quarterly mean “capitalization rates” for com mercial and industrial investment properties. by explaining the variations in the capitaliza tion rate, we hope to expand the research in explaining variations in the overall return to property. this study differs from other research on portfolio capitalization rates because we separately analyze the rates by property type. the results show that using “averaged” capitalization rates across property types eliminates important information. we use the band of investment approach to develop a theoretical model explaining the capitalization rate and test this model using both seemingly unrelated regression (sur) and cross-sectional/time series regression (panel data). the journal of real estate research, spring 1993, 8(2): 221-237. (reprinted with permission of the journal of real estate research.) neighborhood racial transition and housing returns: a portfolio approach, by michael devaney (southeast missouri state university) and william b. rayburn (university of mississippi). a portfolio test was used to determine whether neighborhood racial transition impacts the risk/return of housing. we conclude that the diversification advantages of residential real estate were adversely affected by racial transition and that hedge efficiency was positively related to neighborhood income and the proportion of owner occupancy. this result applied only to neighborhoods with a high rate of racial transition and was independent of neighbor hood racial composition. the journal of real estate research, spring 1993,8(2): 239-252. (reprinted with permission of the journal of real estate research.) abstracts of articles on individual financial management 165 risk attitudes of investors increased risk aversion and risky investment, by jiarong fu (university of iowa). this article investigates the relationship between the size of investment in a risky asset and the degree of risk aversion. the necessary and sufficient conditions are established that permit the prediction of whether agents with differing degrees of risk aversion will increase or decrease investment in the risky asset. it shows, in particular, that when the marginal return to investment decreases (increases) with an improvement in the state of nature, greater risk aversion will induce higher (lower) investment. the journal of risk and insurance, september 1993, 60(3): 494-501. (reprinted with permission of the journaf of risk and insurance.) pii: s1057-0810(99)00017-7 fsr referees financial services review would like to thank the following ad hoc reviewers for providing their valuable and timely contributions to volume 7. jack aber boston university troy a. adair hofstra university robert l. albert jr. moorehead state university seth c. anderson university of north florida noyan arsan state university of west georgia robert balik western michigan university steven balsam temple university joel barber florida international university earl benson western washington university richard bieker delaware state university robert boldin indiana university of pennsylvania daniel j. borgia florida gulf coast university jerry boswell metropolitan state college of denver ben branch university of massachusetts betty brewer north carolina a&t state university robert brooks the university of alabama donald brown arkansas state university stewart l. brown florida state university william t. chittenden northern illinois university conrad ciccotello penn state university frank l. clark eastern illinois university elizabeth cooperman university of colorado at denver larry a. cox university of mississippi dale domian memorial university of new foundland dean dudley eastern illinois university david ely san diego state university dan french new mexico state university delbert c. goff appalachian state university joseph h. golec clark university robert goss board of certified financial planners sue greninger university of texas brian grinder eastern washington university brett hammond tiaa cref vickie hampton university of texas sean m. hennessey university of pei lorene hiris long island university thomas s. howe illinois state university jann c. howell iowa state university riaz hussain university of scranton thomas c. johansen fort hays state university dimitri kapelianis university of witwatersrand surendra kaushik pace university arthur keown virginia polytechnical institute engin kucukkaya university of south florida larry lockwood texas christian university larry lynch roanoke college john a. macdonald clarkson university craig j. mccann kpmg peat marwick krish menon boston university jeffrey m. mercer northern illinois university moshe milevsky york university joel morse university of baltimore hiam a. mozes fordham university jim musumeci southern illinois university darius palia columbia university christopher m. robinson york university thomas r. robinson university of miami michael schellenger university of wisconsin at oskosh thomas schneeweis university of massachusetts nejat seyhun university of michigan michael k. shaub hillsdale university myron b. slovin louisiana state university stanley d. smith university of central florida kenneth b. schwartz boston college william k. templeton butler university alex triantis university of maryland lenos trigeorgis university of chicago oscar a. varela university of new orleans anand m. vijh university of iowa ronald p. volpe youngstown state university mahmoud wahab university of hartford m. mark walker university of mississippi shelly webb xavier university of cincinnati debra worden george fox college pii: s1057-0810(97)90033-0 financial services review, 6(1): 69--74 copyright © 1997 by jai press inc. i issn: 1057-0810 all rights o(f reproduction in any form reserved. i book, software, and web site reviews douglas kahl, editor university of akron personal financial planning, 7th edition. lawrence j. gitman and michael d. joehnk ft. worththe dryden press; 1996 (656 pp.) reviewed by: sue greninger the seventh edition of personal financial planning is user-friendly for undergraduates and features an integrative, dynamic approach. an introductory profile for each chapter pre sents a person or family, depicted in an attractive photograph, who is facing a real-life finan cial decision or problem. undergraduate students assisting in this review said these vignettes helped focus their attention on the relevancy of the chapter's content to their lives. although one student lamented the replacement of the "facts and fantasies" questions at the beginning of chapters by "boring learning objectives, she agreed that the objectives provided a helpful overview of the chapter's purpose. six learning objectives at the beginning of each chapter, however, seemed a tad too standardized for a text seeking to emphasize diversity and change. these two themes--diversity and change---are highlighted quite effectively in the new "financial shock" boxes. the student reviewers felt the situations portrayed in these boxes contributed greatly to their motivation to study the other personal finance content in the text. other positive features are the updated information appearing in the "smart money" and "issues in money management" boxes. these add current interest and are not too numerous to be disruptive to the reader. the streamlined "getting a handle on your financial future" and the integrative case study that appear at chapter ends are also realistic application-level extensions of the subject matter. a very worthwhile change is the interactivity introduced within chapters by the "concept check" questions. given the proclivity of undergraduates to skip over end-of-chapter review questions, this addition should facilitate greater student involvement and learning. part 1, "foundations in financial planning," includes new information on career plan ning, employee benefits, and record-keeping. although additions have been made to the section on psychology and money, there is still a need for greater depth of this topic, as well as on the interpersonal/communication aspects of family financial management. combin ing chapters 2 and 3 from the former edition into a single chapter, "your financial state ments and plans," works well since this allows planning concepts to be discussed before specific financial statements are introduced. one short-coming of the new chapter is the disconnect that occurs between forming goals and making them operational via time value calculations. following the presentation of time value problems, these concepts are not uti 70 financial services review 6(i) 1997 lized for establishing the $4,500 amount included for savings and investments in the weaver's budget on page 84. this is unfortunate since this example provides an excellent opportunity to show how the time value calculations can be used to develop budget esti mates. the recommendation to start the budgeting process with take-home rather than gross pay limits individuals from seeing their entire financial picture and assumes that withholdings accurately match taxes owed, which is often not the case. although the tax chapter is now better organized, a more detailed explanation is greatly needed about flex ible spending accounts than the brief one presented in the "smart money" box on page 129. this is particularly true since these accounts are recommended for those "'just starting out" and '%hirty-something" on page 137. there is also a need for more information about taxes on self employment income since so many young people are entrepreneurs today, not to mention students who unknowingly hire on as "contract laborers." the truth-in-savings regulations description on part 2, "managing basic assets," is most welcome; however, their limitations need to be addressed, especially the exemption of brokerage firms in required disclosure of the annual percentage yield. the inclusion of rent-or-own and lease-or-own information in the housing and automobile chapter is useful. although the organization of this chapter is improved, the sections on home financing and rental options appear later in the chapter than is logical. part 3, "'managing credit," includes rebate and prepaid features on credit cards, as well as updated information on stu dent loan and credit fraud. efforts to tighten up the two credit chapters are appreciated; however, detailed annual percentage rate formulas and calculations might be better suited for an appendix. part 4, "'managing insurance needs," includes good comparative cost data for various types of policies. the comprehensive attention to managed health care plans is most rele vant given their market prevalence and rapid growth. the long-term care insurance discus sion is quite comprehensive, but it may not grab the attention of young students as much as more mature ones. wording utilized in this section seems to downplay the importance of disability income insurance which seems unwise since this coverage is often overlooked by consumers. the role of state regulation in property insurance should be discussed since this causes diversity across the country. it might be helpful to acknowledge that financial plan ners often use less conservative assumptions about rates of return and inflation than the authors when calculating life insurance needs. part 5, "managing investments," presents broadened coverage of mid-and small-cap stock groups as well as detailed information on brokerage commission and transactions pro cesses. the first chapter describes stocks and bonds in general, followed by a second chap ter on the specifics of securities transactions. another approach might be to have separate chapters for stocks and for bonds with pertinent information, such as how to interpret the respective listings, incorporated with the appropriate security types. this would allow com missions to be factored out when rates of return are discussed in chapter 11. the consoli dation of the mutual fund chapter allows a much more comprehensive approach to this important vehicle, especially with regard to the various loads and fees charged to investors. it would be useful to mention the tax aspects of mutual funds in more detail, particularly the lack of investor control over capital gains distributions and their taxation. the omission of other investment vehicles, real estate, derivative, etc., does not seem problematic for an introductory personal finance course given the difficulty in covering all topics in one semes ter. these investments could be included in an appendix for interested students to review. book, software, and web site reviews 71 the updated information on the importance of living wills and powers of attorney in part 6, "retirement and estate planning," is most welcome. a few more examples of how the different trusts in exhibit 15.4 can minimize estate taxes would be helpful. the fact that many people die intestate is well-covered in this section; however, the use of "sweetheart" wills without appropriate trusts needs to be mentioned. the ancillaries to this text include a student workbook, a bound package of blank forms, software that can be run only on ibm-compatible computers, and a comprehensive instructor's manual and test bank. the outlines in the student workbook are more useful than the topic outlines in the instructor's manual. helpful case studies, problems, and vocabulary exercises are included in the student workbook; however, answers to the "'con cept check" questions are unfortunately not included in the workbook for the students to self-check. these answers are included in the instructor's manual along with key concepts, supplemental class project ideas, and answers to discussion questions, case problems, and the integrative case study. a weakness of the instructor's manual appears to be the test bank which offers objective questions that are generally less challenging than the subject matter merits. in conclusion, gitman and joehnk have produced another stellar edition of personal financial planning. it is evident that these authors have listened to the suggestions of their audiences and incorporated many positive changes into their seventh edition. the merrill lynch web site (www.ml.com) reviewed by: john grable, cfp, doctoral student, virginia tech internet users face two annoying problems when using web sites sponsored by the major brokerage firms, namely, slow information retrieval and advertising overload. the first problem, accessing information quickly, is directly related to the user's computer capabilities and inversely related to the number of bells, whistles, and graphics the web site offers. the second problem, advertising overload, is becoming so common that users are forced to browse through product information in order to find useful data. many internet users have concluded that the need for timely and accurate information retrieval often out weighs the benefits provided by fancy graphics and excessive advertising. after subjecting several brokerage firm-sponsored web sites to extensive review based on their ease of use, speed on information retrieval, breadth of information provided, and usefulness of data to practitioners, researchers, investors, educators, and students, the web site sponsored by merrill lynch (www.ml.com) emerged as one deserving further review and use by readers of financial services review. the merrill lynch home page downloads very quickly offering users the choice of five linking pages. financial professionals and investors will find the "financial news & research" ink an easy way to obtain stock quotes (provided on a 20-minute delay during trading hours), market updates, research publications, and legislative updates. the ability to get a glimpse of recent research reports on topics like equities, fixed income strategies, and market commentaries (provided three times daily) are enough to make this web site interesting. the "washington watch" link offers a unique view of legislation that may impact investors, both individuals and institutions. the investor learning center link is an excellent source of basic investing terms and concepts that will interest educators, researchers, and students. obviously merrill lynch pii: s1057-0810(97)90007-x financial services review, 6(4): 295-296 copyright 0 1997 by jai press inc. issn: 1057-0810 al1 rights of reproduction in any form reserved. book, software, and web site reviews risk management and insurance (7th edition) s. travis pritchett, joan t. schmit, helen i. doerpinghaus, and james l. athearn st. paul, mn: west publishing company; 1996 (isbn o-3 14-06427-3) reviewed by: ryan b. lee and j. tim query, department of risk management & insur ance, university of georgia there are a number of principles textbooks on the market, each with its own style of pre sentation. risk management and insurance by pritchett et. al. is a textbook whose principal readership is students enrolled in a principles of risk management and insurance course. as the authors point out, this book is operational, and thus is also concerned with what the non-student user needs to know and what they need to do. as a principles textbook, it cov ers a wide myriad of topics related to risk management and insurance, while at the same time providing depth in an effective manner. the text is arranged in five parts with 26 chapters, and seven appendices. consumer applications at the end of each chapter provide real-world applications of the material just covered. a number of these applications are designed to capture the interest of students by using examples they can relate to now. further, these applications will be useful to all stu dents whether or not they pursue employment in the risk management industry. some changes made to the 7th edition include an ethical dilemmas case or question relating to the chapter material. the conflicts raised between competing interests are designed to stimulate discussion of complex situations that do not have exact answers. expanded coverage of business risk and insurance, personal risks, employee benefits, and general liability has been added to this edition as well. electronic spreadsheets are used in several places to illustrate material. part i is titled “fundamentals of risk management and insurance,” and contains chap ters on risk, risk management, insurance, fundamental doctrines affecting insurance contracts, and structure and analysis of insurance contracts. an appendix in chapter 2 provides a useful introduction into cash flow analysis. useful tables and figures are inter spersed throughout the chapters to provide summary information, beneficial statistics, and more details regarding coverages, exclusions, etc. part ii covers “property and liability risks,” and has chapters discussing the prop erty exposure, the liability risk, managing home risks, managing automobile risks, business property insurance, workers’ and unemployment compensation, and business liability insurance. the organization of the homeowners and automobile sections were well done, and should provide students with easy access to specific inquiries regarding those policies. 296 financial services review 6(4) 1997 part iii covers the “personal risks.” in depth coverage of life insurance, including a thorough discussion of comparing costs associated with different life insurance options is included. a well-researched history of the social security system is provided including a timely discussion of its future. employee benefits is thoroughly covered and provides the student with excellent insight as to the rationale behind providing such beneflts on a group basis. the last chapters of part iii provide detailed case studies of both ~nancial and retire ment planning to help solidify the concepts introduced in the preceding chapters. part iv entitled “insurance markets and regulation” begins with a superb summariza tion of the development of insur~ce. the consumer applications of this section are par ticularly of note as they contain helpful information for all students. further on, this section contains advice on how to become a better consumer of insurance thereby increasing the value of a principles class from the students’ perspective. the cases provided at the end of each chapter are well-developed tests of a student’s ability to apply their knowledge practically. further, to the course instructor, these cases provide an excellent skeleton for essay-style questions. special mention should be made of the appendices in this textbook. they include examples of various insurance policies, which is very effective in connecting the discus sion of an insurance policy to the actual contract. there is also an appendix with the com misioners 1980 standard ordinary mortality table, and a handy list of addresses and telephone numbers of every state insurance commissioner. overall this text is an excellent resource of info~ation and would serve a student well during and beyond their coursework. it contains much more information than one could ever hope to cover in a beginning course, however it does give an instructor many options as to what topics to cover while still offering the level of coverage necessary to serve the students’ needs. pii: s1057-0810(97)90023-8 financial services review, 6(2): 125-140 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. an overview of financial services resources on the internet brian grinder individual financial decision making is a process requiring a great deal of informa tion, most of which can be found on the internet if one knows where to look. this paper is a guide to recent financial service developments on the internet that allows academics, students, professionals, and consumers to find information, interact with others, and conduct financial transactions online. it also offers an extensive list oj jmancial service internet sites, and provides a glimpse into the future offinancial ser vices on the internet. i. introduction the explosive growth of the internet has overwhelmed us with information. although information that was difficult if not impossible to obtain only a few years ago is readily accessible, the danger of information overload has never been greater. effective discipline specific navigation tools must be designed in order to avoid the pitfalls that accompany an abundance of information. the purpose of this paper is to develop a general guide to finan cial services information on the internet for academics, students, and practitioners. the paper is organized as follows. the next section reviews the literature on the impact of the internet on finance and finance education. this is followed by a discussion in section iii of the internet’s impact on various financial services categories. an extensive guide to financial service internet resources is provided in section iv. section v looks at the future of financial services on the internet. ii. literature review academics and professionals alike have shifted away from the idea that the internet is just another technological fad to the realization that this medium will fundamentally change the brian grinder l assistant professor of finance, department of management, eastern washington university, 668 n. riverpoint, ms #3, spokane, wa 99202; e-mail: bgrinder@ewu.edu. 126 financial services review 6(2) 1997 way business and education are conducted. financial practice and education’s inclusion of three articles on finance and the internet in its fall/winter 1996 edition bears witness to this shift in thought. in the first article, ray (1996) furnishes an introduction to the internet for novices. he provides a short history of the internet and develops a concise guide to information, communication, financial services, and financial education on the internet. after a brief discussion of internet basics, the second article by pettijohn (1996) provides a comprehensive guide to general finance resources on the internet. the third article by herbst (1996) discusses the use of the internet by financial engineers. financial services review has also recognized the importance of financial information on the internet and initiated a regular web site review feature. kahl(l996) began this fea ture with a review of the london international financial futures and options exchange web site. subsequent reviews have discussed the web sites of the association for invest ment management and research (smaby, 1996), the new york stock exchange, the chi cago board options exchange, the financial management association (eyssell, 1996), merrill lynch (grable, 1997), and charter media (albert, 1997). finance instructors are increasingly relying on the internet as a source of data for stu dent projects. this is evidenced in smith’s (1996) discussion of a term project that requires students to use data from the security and exchange commission’s edgar site and in st. pierre’s (1996) review of a proprietary on-line database from bridge information systems that can be used to provide financial data for student projects. this increased reliance on the internet for data is only the beginning of a radical transformation of higher education (updegrove, 1996). iii. a guide to financial services on the internet the internet allows us to find information with the click of a mouse, receive instantaneous responses to inquiries, gain access to multiple information sources, and conduct financial transactions online. in short, it has transformed nearly every aspect of business. financial services have readily adapted to this new environment by providing information and devel oping innovative online services. what follows is a brief discussion of the information and services available for each of the major financial services categories. a. general searches search engines such as alta vista, excite, hotbot, infoseek, lycos, and webcrawler provide good starting points for any internet exploration. however, the best search engine for finding financial services information is yahoo! because of its easy-to-use search fea ture and its extensive list of financial services links. yahoo! makes up for the fact that it is not the most powerful search engine available by providing search options to the user that are both convenient and useful. b. investments there are probably more web sites devoted to investments than any other financial ser vice. it is possible to obtain a great deal of timely news and information about stocks, bonds, mutual funds, and investment strategies from the internet. a number of sites pro resources on the internet 127 vide stock and index quotes. while real time quotes are generally available for a fee, 15-20 minute delayed quotes are available for free on many sites. annual reports, prospectuses, and sec documents for a host of corporations are available to download from various sites including edgar, shareholder direct, and corporate financials online. on-line trading is available on sites such as e*trade. communicating with other investors is possible through a number of methods. chat servers allow individual investors to get together in cyberspace to discuss investment top ics. these servers allow participants to interact with one another in real time. another form of interaction is the investment discussion forum. participants in discussion forums can post questions, answers, or observations on a message board. other participants can then read and, if they choose, respond to anything posted on the board. usenet newsgroups are similar to discussion forums except they use a newsgroup server whereas discussion forums are available on world wide web sites. a list of investment newsgroups is avail able in the appendix. individual corporate web sites are another important source of investor information. for example, intel’s web site provides on-line access to the company’s 1996 annual report, press releases, sec documents, investor news, current quotes for intel stock, facts and faqs about intel, and earnings release reports. many corporate web sites are begin ning to provide information in audio or video format. c. mutual funds online mutual fund information includes prospectuses, price quotes, screening tools, news, and investment advice. information about specific fund families can often be found at the fund’s own web site. for instance fidelity’s site contains descriptions of the various funds offered by fidelity, daily fund prices, and information for individuals and institutions on mutual fund investing strategies. one can also use sites such as schwab’s mutual fund onesource online to find information on a number of mutual fund families. another good source for mutual fund information is momingstar.net which provides morningstar mutual fund reports as well as portfolio tracking, news, and financial planning strategies. d. banking/banking services banking web sites offer a great deal of useful material for academics and profession als. the articles available at the alex sheshunoff management services site are a good example of the information that can be obtained online. these articles deal with timely top ics such as electronic banking, strategic planning, and corporate culture. the federal deposit insurance corporation site also offers a rich database of statistics and information about the banking industry. because of its dynamic nature, the internet provides academics and professionals alike the opportunity to keep abreast of developments in the banking industry. for instance, the latest trends in credit cards are tracked at the bank rate monitor and ram research web sites. information about online banking is also available on sites such as huntington banks on the web and the moneypage. 128 financial services review 6(2) 1997 e. planning the internet has information, advice, and tools for financial planning, pension and retirement planning, and estate planning. the institute of certified financial planners site is a good starting point for students interested in financial planning as a career. here they can learn about the process of earning the certified financial planner (cfp) license and access feature articles from the journal of financial planning. pension resources available at the pension research council site include a newsletter and listings of publications and working papers. another useful site is provided by the pension benefit guaranty corporation (pbgc) which maintains a database of opinion let ters from the office of the general counsel of the pbgc. these letters “advise the public of its views of the meaning of the provisions of title iv of the employee retirement income security act.” one of the most impressive sites for pension researchers is main tained by the pensions institute at the university of london. this site offers working papers, data, and links to other pension sites on the internet. there are even computer codes available for download that calculate everything from internal rates of return to aggregate claim distributions. estate planning information is available on an number of sites including the national network of estate planning attorneys. many law offices also maintain web sites dedicated to estate planning. f. tax planning one of the best resources on the world wide web for tax planners is the web site of the internal revenue service. this site provides a frequently asked questions page which contains information on everything from alternative filing methods to the depreciation of assets. irs forms and publications are also available online. state tax information can be found on the taxweb site which maintains a tax page for all 50 states. g. real estate the american real estate society (ares) web page is one of the most useful resources available for real estate. the society’s four real estate journals all have pages on the ares web site which provide submission information, past issue contents, special issue contents, and calls for papers. there is also an extensive list of links to real estate organizations, real estate investment advisors, consultants, commercial real estate informa tion, and journals. the department of housing and urban development’s homes and communities web site provides useful information for individuals, professionals, and academics. hud posts consumer alerts, housing information, and community involvement ideas for individuals. academics and professional researchers can access hud user databases for information on assisted housing, market rate housing, and public housing. access is also available to mapping software, hud handbooks and directives, hud forms, and the hud phone book. h. insurance the world wide web offers insurance information and services for both consumers and insurance professionals. sites such as insweb allow consumers to obtain online quotes resources on the internet 129 for a number of different insurance products. it is even possible to obtain quotes from a specific agent. the site also provides insurance news and financial planning resources. pro fessionals can benefit from online educational resources, regulation information, and access to potential customers. i. financial counseling a number of counseling services have developed web sites that provide financial information, debt reduction strategies, and opportunities for counseling. the consumer credit counseling service web site, for instance, provides a number of publications and resources that can help consumers learn to budget, save money, and get out of debt. j. employee benefits employee benefits sites include news and information. government information about employee benefits is available at many of these sites. of course, it is also possible to directly access government web sites, such as social security online, that offer employee benefits information. the webenroze site is an online open enrollment system that should make changing employee benefits much easier once implementation is complete. webenroll is a good example of how innovators are employing new technology to streamline and simplify pro cesses that traditionally take an enormous amount of time and effort to perform. k. education in financial services the impact of the internet on education has been tremendous. it is now commonplace for instructors to post syllabi, lecture notes, and supplementary materials in electronic form on a web server. the technology promises to revolutionize the educational process as we know it. a wide range of educational opportunities are available online. sites such as the financial pipeline offer internet users the opportunity to learn about financial planning and investments on an informal basis. those looking for a formal educational experience can check out sites such as the national endowment for financial education, which offers a distance education program that leads to a master of science degree with an emphasis in financial planning. distance education on the internet is still in the early stages of develop ment, but it has already provided hundreds if not thousands of individuals unprecedented educational opportunities. the mission of the academy of financial services (afs) explicitly encourages the development of financial services curricula at the university level. financial services review, the official journal of the afs, helps to fulfill this goal by publishing appropriate pedagogical papers from time to time. continuing education for professionals is also important to the afs. there is currently an announcement on the afs web site that encourages members to contribute financial planning case studies to the institute of certi fied financial planners. 130 financial services review 6(2) 1997 iv. f~ancial services sites the appendix below is designed to provide readers with a sample of what is currently available at financial service sites on the internet. the format of the appendix follows the design of choi (1988) and pettijjohn (1996) by providing the name of the information ser vice, the info~ation provider, a brief annotation, and the internet address of the site. be forewarned that internet addresses can and do change. if a web site changes its address, the information provider will often leave a link to the new address at the old site. if this is not the case, one can always use a search engine to find the new address. v. the future the pace at which new applications and technologies for the internet are developed is mind boggling. terms such as java, javascript, and dynamic html were not a part of our lan guage a few years ago. however, these tools promise to make the internet more interactive and useful. for financial service providers this means an increasing amount of business will be carried out in cyberspace. for consumers of financial services, the future promises custom tailored information, efficient service, and lower prices. a number of problems such as bandwidth and security are not yet resolved, but none of the current problems with the internet is so severe that progress is threatened. while academics can look forward to new teaching tools and increased access to important databases, one of the most profound changes will take place in academic jour nals. currently, journal web sites tend to be viewed as secondary to the printed journal. however, as editors begin to focus on the internet as the primary method of disseminating scholarly information, many innovative methods for scholarly interaction will be &vel oped. the journal of interactive bedim in ~~~c~s~~~ (jime) is a good example of what can be expected in the near future. jime is a scholarly journal with an editorial board and a review process. however, nearly everything from the initial submission of a paper to final publication happens electronically. papers, which can be submitted via ftp, are reviewed by two reviewers and one or more members of the editorial board. papers accepted in princi ple are then posted on the jime web site (h~p:/lwww-jime.open.ac.u~jime~ for a one month period of open review. anyone interested in the topic may read the paper and com ment on it. pertinent comments from the open review process are included with the final document. once a paper is accepted for publication, it receives a “published” tag. vi. conclusion the internet has created many opportunities for academics, professionals, and consumers of financial services. this study has endeavored to familiarize the reader with the informa tion and services that are currently available. hopefully, this information will peak the reader’s interest, encourage further exploration, and perhaps spark innovative ways for financial services to use the internet. resources on the internet 131 appendix a guide to financial services sites information service information provider inform&on typo url of home page general corporate finance network financenter national financial services network quicken financial network yahoo! investments corporate financials online dbc online edgar database e*trade the great stocks project story street partners, financenter, inc. national financial services network intuit inc yahoo! inc corporate financials online, inc data broadcasting corporation securities and exchange commission e*trade group, inc. the great stocks project the high-tech mathew investor ingram provides a variety of information resources for the financial services industry including news and investment information. the corpfinetsoo offers links to both company and information web sites. there are also helpful guides for financial service providers wishing to do business on the internet. information is available on a variety of personal finance issues ranging from budgeting to bonds. the site boasts a number of online interactive financial calculators. this site is divided into personal finance and commercial finance. the resource libraries contain extensive lists of links to financial resources and information on the internet. offers information on investments, retirement, taxes, banking, insurance, and more. stock/mutual fund quotes are available as well as life insurance quotes. one of the best search engines available. the business and economy page offers links to banking, commercial financial services, financing, mutual funds, and real estate. offers news and investor information for public corporations. the site also tracks the latest sec tilings. dbc offers quotes and charts for stocks, bonds and mutual funds. financial news and a stock chat server are also available. an electronic database of corporate sec filings. a two-year window of electronically filed documents is maintained. several special purpose search features have recently been added to the site. one of the first online trading services, e*trade is a good example of the internet’s potential to transform the brokerage industry. the intent of the great stocks project is to pool the intellectual capital of the individual investor. subscribers can post messages to one of three e mail lists. the greatstocks list focuses on fundamental analysis, the daytrades list focuses on technical analysis, and the marketwise list deals with general investing issues. a narrative format interspersed with hypertext links guides you to many useful investment sites. http://www. corpfinet.com/ http:// www.linancenter. corn//œ http://www.nfsn.com/ http://www.qfn.coml http:// www.yahoo.coml http:// www.cfonews.com/ http://www.dbc.com/ http://www.sec.gov/ edaux/searches.htm http:// www.etrade.coml http:// www.geocities.com/ wallstreet/ l/ index.html http:// www.want2know. com/invest/invest.htm 132 financial serviceis review 6(2) 1997 hoover’s oniine! intel invest-o-rama! investorsedge investorweb microsoft investor the motley fool naic online shareholder direct stockmaster stock smart newsgroups intei corporation douglas gerlach investors edge the stone age consulting company microsoft, inc. the motley fool national association of investors corporation direct report corp. marketplace. net inc. stock smart offers free company capsules for over 10,oflu corporations. the capsules include a description of the company, traditional mail addresses, corporate web site urls, and links to other information sources such as dbc, edgar, and stockmaster. a good example of the financial information available on individual corporate web sites. download intel’s annual reports, quarteriy reports, and sec documents. news, stock quotes, investment research, investor education. a great site for the individual investor. of particular interest is the financial services marketplace which offers information on a state by-state basis on everything from insurance to electronic payment systems. includes market iuformation, stock quotes, a portfolio tracker, and research . at-a-glance summaries provide recent information about individual corporations, analyst’s recommendations, and charts. investor web features links to newsletters, stock quotes, private equity isolation, and corporate web sites offers a portfolio tracker, stock quotes, and historical price data that can be downloaded for use in a spreadsheet. company news, and investment articles are also available. the fools provide a forum for individual investor online interaction. message boards provide investor views on securities. educational materials, investment strategies explanations, quotes, and news are also available. this site is dedicated to assisting investment clubs, but there are many tools, such as the “company info~ation and investing ideas” page, that are useful to the individual investor. offers information to shareholders via the internet, e-mail, traditional mail, telephone or fax. many corporations arc using the service as a replacement for mailing out financial material to shareholders. one of the first sites to offer stock and mutual fund quotes and charts. stockmaster now offers information on indexes. a rich source of data; stocksmart has one of the few portfolio trackers able to track stocks, mutual funds, and bonds. quotes, and graphs are also available. the time lag on market index quotes is the shortest available for a free service. http:// www.hoovers.comi http:// www.investorama. coml http:// server1 .imet.com/ scripts/ sqlgate.exe?www+ homepage http:// www.investo~eb. comidefaultasp http:// investor.msn.com http://www.better investing.orgf http:// www.shareholder. corn//œ http:// wwwstockmaster. corn//œ http:/1 wwwl stocksmart. corm news:misc.invest news:misc.invest. funds news:misc.invest. futures news:misc.invest. stocks news:misc.invest. technical resources on the znternet 133 mutual funds fidelity investments the fund library investorguide webfinance, mutual funds inc. momingstat the mutual fund cafe mutual funds interactive the mutual fund investor’s center networth schwab mutual fund onesource online fidelity investments the fund library inc momingstar, inc. wechsler & partners. inc. brill editorial services, inc. the mutual fund education alliance galt technologies inc. charles schwab & co., inc the site provides information on the fidelity family of mutual funds. prospectuses and shareholder reports are available online. although the fund library is intended to provide information about canadian mutual funds, there is a great deal of useful information here on mutual funds in general including discussion forums, and investment tools. the site consists primarily of links to other mutual fund sites. categories include all-purpose sites, performance data and ratings, screening, how to choose a mutual fund, and learning. momingstar reports for open-ended funds are available. closed-end fund reports will be available soon. the site also includes information on stocks, portfolio tracking, and investor education. the mutual fund cafe uses a clever format to dispense news, trend information, information on legal issues, and information on accounting issues. includes resource links, expert advice, fund manager profiles, news, quotes, charts, and investor education. the site also provides a chat server for discussing mutual funds online. provides links to mutual fund web sites, educational resources, daily prices, and a screening procedure for selecting funds. the “map out a plan” page helps investors focus on their reasons for investing. online tools for assessing an investor’s risk tolerance level are also provided. networth’s fund atlas offers a directory of mutual funds, market commentaries, net asset values, recent and historical fund prices, and momingstar profiles. in addition, the “equities center” offers a great deal of information on common stock. onessource offers profiles of mutual fund companies, performance rankings, and online trading. banking/banking services alex sheshunoff management services inc. alex there is a useful list of bank resources for sheshunoff community bankers available here as well as a management number of articles by alex sheshunoff on services inc. electronic banking, when to sell a bank, and bank valuation. banxquoter banxquote inc. this site has general information and rate information on mortgages, credit cards, vehicle loans, student loans, jumbo money markets &cds, and savings accounts. there is also a reference library with a wide variety of information available on a number of financial services topics. http:// www.fidelity.com http:// www.fundlib.com/ http:// www.investorguide.c om/mutualfunds.htm http:// www.momingstar net/ http:// www.mfcafe.com/ http:// www.fundsinteractive .com/index.shtml http:// www.mfea.coml mfeaindx.html http:// networth.galt.com/ http:// www.schwab.coml schwabnow/ snlibrary/snlib014/ sn014.html http:// www.ashesh.coml http:llwww.banx.coml 134 bank rate bank rate monitor: monitor inc credit crossroads webnation federal deposit insurance corporation federal deposit insurance corporation huntington banks on the web institute of canadian bankers (icb) huntington bancshares inc institute of canadian bankers the money page moneypage inc. mybank directory fitech, inc. ram research group ram research group financial services review 6(2) 1997 monitors rates on credit cards, savings and checking accounts, and many other bank products. the site also provides banking news, an online payment calculator, and useful banking tips for consumers. useful information about credit cards, credit reports, divorce and your credit, credit repair, bankruptcy, and credit scoring. the data bank contains the summary of deposirs report, fdic statistical publications (online), and a report on bank underwriting practices. banking news and consumer news are also available. the “laws and regulations” page contains a searchable law data base. the “public information” page contains press releases and other information about the fdic. learn about online banking with huntington’s online banking demo. a number of online financial calculators are available. icb is dedicated to helping financial service professionals stay current. to that end, the institute offers a number of courses and seminars. many of these are available through distance education. self-tests and quizzes are available on the web site. the moneypage site maintains a monthly list of the top ten banks in cyberspace, and offers feature articles on crucial issues in banking. resource pages are available for banks, s&ls & credit unions; mortgage banking; investment banking; technology; brokerage firms; consumers; electronic money; u.s. legislative banking committees; news & reference: and regulatory compliance. locate bank web sites in all 50 states and around the world. everything you ever wanted to know about credit cards. find the best rates, monitor the industry, or download data. financial planning/pension and retirement planning/estate planning asec online american savings education council crash course in michael t. wills & trusts palermo iafp interactive international association for financial planning online brochures dealing with retirement planning, financial goals and objectives are available. an online form is available for requesting an asec speaker. described as a “concise and practical guide to what everyone should know about the law of wills and trusts before an estate plan is designed.” iafp offers consumer services information on the various aspects of financial planning. professional services include links to iafp’s official publication financial planning magazine, the iafp code of ethics, governmental affairs information, and links to other financial planning resources. http:// www.bankrate.com http://amdream.com/ credit/cr_index.htm http://www.fdic.gov/ http:// www.huntington. corn/ http://www.icb.org/ http:// www.moneypage. corn//œ http:// www.mybank.com/ http:// wwwmmresearch. co& http://www.asec.org/ http:// www.mtpaletmo. coml http://www.iafp.org/ resources on the znternet 135 institute of certified financial planners layne t. rushforth’s estate planning articles moneyadvisor national network of estate planning attorneys nafep pension pension benefit benefit guaranty guaranty corporation corporation the pension institute the university of london the pension research council robert clofine’s estate planning page tax planning the internal revenue service national national association of association tax of tax practitioners practitioners institute of certified financial planners layne t. rushforth complete financial services national network of estate planning attorneys national association of financial and estate planning the pension research council of the wharton school of the university of pennsylvania robert clofine the internal revenue service the consumer services page offers advice on selecting a financial planner. the professional information page describes the process of becoming a certified financial planner. the site also hosts the journal of financial planning home page. the site features articles on basic estate planning, and advanced estate planning by layne t. rushforth, a fellow with the american college of trust & estate council. moneyadvisor is primarily a directory to other online financial services. online referrals to finance experts are available as well as several online financial calculators. the “keys to wealth ” page offers information on estate planning, retirement planning, insurance strategies, and education funding. the last will and testament of jacqueline kennedy onassis is available on the site. the nafep site includes information on asset protection and estate planning, wealth planning and tax tips, capital gains tax deferral, and home mortgages. offers opinion letters, pbgc regulations, interest rate updates, and a pension search directory. the pension institute focuses entirely on pension research. a mailing list keeps subscribers apprised of the institute’s activities. a number of working papers are also available online. the prc newsletter is available online. publications and working papers lists are available, and may be ordered via fax, phone, or traditional mail. topics covered include the 1997 social security increase, iras, estate planning, tax planning, insurance. and medicaid. the ultimate source of federal tax information offers tax resources for individual and businesses. tax regulations written in plain english and irs bulletins are available as well as online forms and publications. the site offers information on tax workshops, new tax issues, and membership in the natp. featured articles from the tar practitioners journal, the 104oreport monthly newsletter, and the taxtips client newsletters are available online. http://www.icfp.org/ http:// coyote.accessnv. com/rushfott/ planning.htmk+top http:// www.moneyadvisor.c od http:// www.netplanning. corn/ http:// www.nafep.com/ http://www.pbgc.gov/ http:// www.econ.bbk.ac.uk/ pi/ http:// prc.wharton.upenn.ed ulprc/prc.html http://home.ptd.net/ -clofine/ http:// www.irs.ustreas.gov/ http:// www.natptax.comj 136 the national tax association tax analysts taxcast tax management, inc. tax web tax world real estate the american real estate society the american real estate and urban economics association (areuea) coldwell banker online countrywide mortgages fannie mae arizona state university college of business tax analysts ernst and young, llp bureau of national affairs, inc. webtech consulting, inc. thomas c. omer, univ. of illinois at chicago the american real estate society the american real estate and urban economics association coldwell banker corporation and lnterealty corp. countrywide credit industries, inc. federal national mortgage association financial services review 6(2) 1997 the nta site supplies summaries of articles from the national tax journal. there is also an extensive list of tax related sites. offers continuously updated tax news, tax discussion groups, a tax clinic, and a guide to 1996 tax changes. the tax calendar provides notification of government events as well as tax meetings and seminars. taxcast provides news for tax analysts, a public forum for tax professionals, a tax site of the month, and an extensive list of tax links. the knowledge center provides access to ernst and young publications. the education page offers a great deal of information on continuing education opportunities. tax news, cpe and cle educational opportunities, and a good list of tax links arc available at this site. provides links to tax information for all fifty states. links to discussion groups, tax publishers, professional organizations, discussion groups, mailing lists, general tax gateways, and tax articles. tax world offers a list of tax links, a list of tax courses online, 1996 tax forms, and tax policy forum. this is a comprehensive site with a great list of links to real estate organizations, investment advisors, consultants, commercial real estate, and journals. areuea provides abstracts from real &tare economics, a list of real estate links, and a membership directory. provides tips on buying and financing a home. search the coldwell banker data base to find houses for sale anywhere in the nations. screening by price, size, style, and street location is possible. online mortgage calculators are also available. offers online applications for mortgages, loan information and services, and information for real estate professionals. information about fannie mae and the housing finance system. prospective home buyers will find a great deal of useful information on purchasing a home. http:// www.cob.asu.edulnta http://www.tax.org/ http:// www.taxcast.com/ http://www.bna.com/ tm/ http:// www.taxweb.coml http:// omer.actg.uic.edu/ http:// www.aresnet.org/ http:// www.areuea.org/ http:// www.coldwellbanker. coml http:// www.countrywide.co ml http:// www.fanniemae. corn resources on the internet 137 hud’s homes and communities interactive home buying on web the international real estate digest r.e.infonet insurance the american insurance association ariaweb a. m. best insurance information exchange (11x) insurance information institute (iii) insurancenews network insweb of housing and urban development maxsol, inc’s ired.com, inc. r.e.infonet the american insurance association the american risk and insurance association a. m. best company insurance information exchange, l.l.c. the insurance information institute insurance news network, llc strategic concepts corp. hud offers housing information and consumer alerts for individuals. however, real estate professionals will find a vast amount of useful information here as well. the online library includes legal information, fair housing guidelines, housing reports, and public policy information. a comprehensive guide to buying a home. use the fifteen-step process to work through the tangled web of home buying. online mortgage calculators allow comparison of fixed rates to adjustable rates, determination of amortization schedules, and mortgage qualifications. ired features news, editorials, financial resources links, real estate directories, and a monthly top ten list of real estate web sites. provides information for buyers and sellers of homes. a nationwide referral network is available for finding real estate agents. links to member companies, state insurance commissioners, and related areas are available. other features include an online order form for purchasing association publications, and aia press releases. ariaweb hosts the journal of risk and insurance, and the risk theory society home pages as well as a jobs database for individuals with phds in risk and insurance. links are provided to research, teaching, and other risk and insurance sites. the site primarily provides product and price information. however, there is an explanation of best’s rating system and a sample company report available. 11x offers news, industry links, and an insurance e mail directory. consumer information includes news, consumer alerts, and an iii catalog of brochures. media information includes facts about various types of natural catastrophes. news and information about auto insurance, home insurance, life insurance, annuities, insurance ratings, and state insurance are available. get quotes online for many different insurance products. quotes are also available directly from the agent of your choice (not available for all states). other features include news, an agent locator, faqs, and financial planning resources. there is also a page for insurance professionals as well as a career center. http://www.hud.gov/ http:// www.maxsol.com/ homes/calcsm.htm http://www.ired.com/ http:// www.reinfonet.com http://www.aiadc.org/ http://www.aria.org/ aria.html http:// www.ambest.com/ http://www.iix.com/ http://www.iii.org/ http:// www.insure.com/ http:// www.insweb.com/ 138 is0 national association of insurance commissioners rightquote rims on the web riskweb rmis-web society of actuaries insurance services office, inc. national association of insurance commission ers coverdell & company the risk and insurance management society, inc. james r. garven infotech consulting inc. society of actuaries financial counseling consumer consumer credit credit counseling counseling service service credit credit counselors counselors corporation debt debt counselors of counselors of america america metropolitan metropolitan financial financial management management national the national association of association personal of personal financial financial advisors advisors employee benefits benefitslink david rhett baker financial services review 6(2) 1997 is0 studies and analysis are available as well as links to other insurance sites. features include links to financial data and reporting, the journal of insurance regulation, regulator information, law, and securities valuation. rightquote provides educational courses, mortgage calculators, life insurance quotes, and articles. offers news, government affairs information, and educational development materials. professional risk managers can network with one another on the rims public bulletin board. the site also hosts risk management magazine’s web site. one of the first web sites dedicated to risk and insurance, riskweb offers an e-mail discussion forum that deals with risk and insurance issues. there are also a number of links to other risk and insurance web sites. self-described as “the internet resource for risk management information systems,” rmis-web provides reviews of vendors and systems, articles from several leading risk management publications, and a directory of systems software vendors and service providers, soa publications, actuarial exam information, continuing education information, and online research reports are available. features include a money manager planner, a fritter finder (for those of us who can never figure out where our money goes), educational services, and budgeting strategies. provides help for getting out of debt. an online application form is available. offers a debt discussion forum, and the debt eliminator program which allows individuals to make a single monthly payment for all debts. offers counseling, and debt restructuring help. this site offers a service for locating a fee only financial planner, and links to other financial planning resources on the web. a comprehensive site with information on everything from actuarial assumptions to workers’ compensation laws. there is also a useful reference library of official u.s. government documents. http://www.iso.com/ http://www.naic.org/ http:// www.rightquote. coml http://www.rims.org/ http:// www.riskweb.com/ http://rmisweb.com/ http://www.soa.org/ http:// www.cccsedu.org/ home.html http:// web.creditcounselors. comlccc/ http:// www.dca.org http://www.debt help.com/ http://www.napfa.org/ http:// www.benefitslink corn/ resources on the internet 139 employease, inc. employee benefit research institute employee benefit education world wide web site social security online webenroll employease, inc. employee benefit research institute international foundation of employee benefit plans the social security admin. serious data company llc education in financial services afcpe the association for financial counseling and planning education academy of academy of financial financial services (afs) services national national endowment for endowment financial for financial education education the financial pipeline financial pipeline inc. emplyease offers career opportunities links, human resource guides, and an employee benefits primer. there are also forums available dealing with human resources, insurance and benefits, and using the internet. the industry phone book lists addresses and phone numbers for brokers, insurance providers, and government institutions. educational materials include briefings, policy forums, and congressional testimony by ebri researchers. ongoing research projects on social security reform, and retirement confidence are available online. industry news, job postings, information on employee benefit seminars and conferences, and a list of links to other benefits sites are available. there are a number of online publications available including the so&l security handbook. online forms, laws and regulations, medicare information, and links to other sites of interest are available. a web based enrollment system that should help reduce the time and effort of open enrollment. the system is not yet complete. however, a demo is available. afcpe provides continuing education opportunities for financial counseling and planning professionals. the site provides information about afcpe’s professional certification programs. the finuncial counseling und pluming journal and the afcpe new.slerter are also available online. afs is clearly dedicated to improving financial services education. the site provides a convenient online membership directory and a list of links to important financial service sites. for professionals, the site has information on the endowment’s master of science degree with an emphasis in financial planning. there are a number of other course descriptions on the site. for nonprofessionals, there are several articles available on financial planning as well as tutorials dealing with money management, financial planning and life insurance. this is a great site for financial novices because it clearly explains a number of financial concepts. the consumer education section offers advice on credit, financial planning, leasing, and investing in gics. real audio interviews with finance professionals are also available. finpipe also provides reviews of recent finance books. http:// www.eease.coml http://www.ebri.org/ http://www.ifebp.org/ http://www.ssa.gov/ http://webenroll.com/ http://www.hec.ohio state.edu/hanna/afcpe/ index.htm http://www.umsl.edu/ -eyssell/afs/ afspage.html http:// www.insweb.com/ carriers/nefe/ default.htm http:// www.finpipe.com/ 140 financial services review 6(2) 1997 references albert, r.l. (1997). charter media’s briefing.com web site. financial services review, 6,72-73 choi, f.d.s. (1988). international data sources for empirical research in financial management. financial management, 17, 80-98. eyssell, t. (1996). the new york stock exchange (nyse), the chicago board options exchange (cboe), and the financial management association web sites, financial services review, 5, 150-151. grable, .i. (1997). the merrill lynch web site (wwwmlcom). financial services review, 6,71-72. herbst, a.f. (1996). the ways in which the financial engineer can use the intemet. financial pruc ticeandeducation, 6, 111-121. kahl, d.r. (1996). liffe on the intemet. financial services review, 5, 85-86. pettijohn, j.b. (1996). a guide to locating financial information on the intemet. financial practice and education, 6, 102-l 10. ray, r. (1996). an introduction to finance on the intemet. financial practice and education, 6, 95-101. smaby, t.r. (1996). association for investment management and research (aimr) web site. financial services review, 5, 149-150. smith, sd. (1996). using edgar on the intemet to teach finance and business courses. journal of financial education, 22,76-78. st. pierre, e.f. (1996). using bridge information systems in the classroom. journal of financial education, 22,79-83. updegrove, d. (1996). revolution @ alma mater.edu: the intemet and higher education. change: the magazine of higher learning, 28,41-44. risk and reward of fractionally leveraged etfs in a stock/bond portfolio james dilellioa,* adepartment of decision sciences, information systems and strategy, pepperdine graziadio business school, 24255 pacific coast highway, malibu, ca 90263, usa abstract this article investigates using 1.25x leveraged stock and bond exchange-traded funds (etfs) as an asset allocation strategy. performance is analyzed by replicating funds from 1989 to 2017, including all relevant costs. conditions for excess returns are derived analytically and confirmed empirically. simulations are conducted to assess performance under a variety of market conditions, and demonstrate opportunities for excess returns over unlevered funds in a 60/40 stock/ bond allocation. these results are accomplished with a small reduction in the sharpe ratio and no need to access margin. we conclude that this asset allocation strategy could be well suited for investors more interested in total returns during upward trending markets. © 2018 academy of financial services. all rights reserved. jel classification: g11; c53 keywords: leveraged etfs; simulation; geometric brownian motion; bootstrapping 1. introduction individual use of financial leverage is fairly common by individual investors. many first-time home buyers use a mortgage where the buyer provides 20% of the closing price of a home, and finance the remaining 80%. the use of leverage in the stock market is also common, where most brokers offer margin accounts to support individual investors purchasing securities above their cash balance, shorting a stock, or investing in financial derivatives. * corresponding author. tel.: �1-714-403-0085; fax: �1-714-832-3654. e-mail address: james.dilellio@pepperdine.edu financial services review 27 (2018) 413-432 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. these margin accounts can amplify gains and losses. however, retirement accounts do not permit this use of leverage, because of the need to provide collateral and their tax exempt status.1 nevertheless, individual investors and financial planners seeking to boost returns in retirement accounts can invest in leveraged funds. because of the effect of daily rebalancing, most leveraged exchange-traded funds (etfs) have largely remained in the hands of short-term traders. many academic articles have discussed the technicalities behind why leveraged etfs are not intended for passive buyand-hold investors. cheng and madhavan (2009) and guedj et al. (2010) show that, because of the path dependency of leveraged etfs, longer holding periods reduce the value of the leverage. lu et al. (2012) show that, for holding periods up to one month, a 2x leveraged etf produces approximately twice the return of its underlying. however, for longer holding periods, the ability of the 2x fund providing twice the return of its underlying diminishes. somewhat surprisingly, some authors indicate that there are potential opportunities. in trainor and carroll (2013), the term “decay” is defined when the difference between a leveraged etf’s return and the leverage factor multiplied by underlying index is negative. this article showed that low volatility coupled with a significant upward price trend can substantially offset decay. avellaneda and zhang (2010) show that with a dynamic hedging strategy, it is possible for active traders to manage leveraged etfs over longer time horizons. however, the complexity of employing such an active strategy is likely beyond the skill of many individual investors. dilellio et al. (2014) find potential portfolio diversification exists with inverse and leveraged etfs used as an alternative asset class in a long-term passive investment strategy. diversification benefits were shown to exist, dependent on the behavior of equity and debt markets, and were shown to be generalizable to a variety of core stock and bond funds often utilized by individual investors and financial planners. trainor and baryla (2008) point out that, for individual investors interested in leverage, the cost of obtaining it via a leveraged etf can often be less than a typical margin account. barnhorst and copcozza (2010) discuss investors taking “volatility risk,” and show how a 2x leveraged fund can under (over) perform its underlying benchmark because of higher (lower) price volatility. giese (2010) summarizes the positive benefits of holding leveraged funds in bullish markets, but recognizes that these benefits can be offset by increased volatility. given the general downward trend in volatility in the equity markets since the 2008 financial crises, there may be future opportunities for individual investors to take this so-called volatility risk. as noted above, the speed at which the leveraged etf can diverge from its underlying index is proportional to the amount of leverage. therefore, individual investors may have interest in using etfs with smaller amounts of leverage than the 2x and 3x etfs that have dominated the leveraged etf marketplace. to this end, this article investigates whether less leverage, at 1.25x, provides a viable investment opportunity for a passively managed portfolio of stocks and bonds without taking on excessive risks. this article also contributes to work by ott and zimmer (2016), where theoretical and empirical results support the use of fractional leverage on return maximizing a portfolio of risky and risk free assets. while their work determined optimal leverage at 1.77–1.85, we chose to focus on a value of 1.25, which was the level available in the financial markets at the time this paper was written. a summary of current stock and bond etfs offering 1.25x leverage appear in table 1 below, along with their unleveraged counterparts. 414 j. dilellio / financial services review 27 (2018) 413-432 table 1 illustrates that the assets under management for the 1.25x leveraged etf tracking the stock and bond indices are extremely small, compared with the unleveraged etfs with the same underlying index, and have only existed for a few years. similar to other leveraged etfs, the 1.25x leveraged etfs also have a higher expense ratio. however, despite the 1.25x leveraged etf’s median daily share volume being hundreds of times smaller than the unleveraged version, the average spread is less than ten times larger, thanks to the liquidity of underlying index. this low volume and low spread outcome differs from what occurs for the majority of etfs, and as shown in agrrawal and clark (2009), who show low volume more often leads to exponentially higher spreads. consequently, transaction costs of rebalancing these leveraged etfs may not significantly reduce returns, so will be investigated in more detail in this article. the 1.25x leveraged stock and bond etfs also exhibit larger premium/discounts. therefore, it is possible that during times of market stress, these etfs may be trading in the lower or higher price range, producing a difference in holding period performance. for purposes of the analysis that follows, we assume the net effect of this premium/discount to be negligible. as a final note about 1.25x leveraged funds, there are currently two mutual funds available to individual investors that offer this fractional leverage, but on a monthly rather than daily basis. the target index for these mutual funds is the nasdaq-100 total return index, and tickers dxnlx and dxnsx provide positive and inverse leverage. in the analysis that follows, we include the bull 1.25 monthly fund as part of a sensitivity analysis to highlight the effect of a monthly versus daily return matching 125% of the underlying. 2. data source and replication technique as shown and discussed in the previous section, there is very limited historical data available on fractionally leveraged etfs. so, to study long-term investment performance under a variety of market conditions, we replicated returns of two equity index etfs in table 1 using the s&p 500 total return index (bloomberg index code spxt). the first fund was an unleveraged etf, and the second was a 1.25x leveraged etf. because we are also table 1 1.25x leveraged stock and bond exchange-traded funds (etfs) and their unleveraged counterparts* etf symbol underlying index assets under management expense ratio median daily share volume average spread premium/discount median (range) issue date pplc s&p500 $81.11m 0.34% 25,223 0.04% �0.03% (�2.28% to 2.25%) january 7, 2015 ivv† $144.83b 0.04% 3,870,076 0.01% 0.00% (�0.11% to 0.13%) may 15, 2000 pptb barclay’s us aggregate bond $25.05m 0.34% 15,100 0.04% 0.01% (�0.39% to 0.17%) february 15,2018 agg $53.62b 0.05% 3,627,118 0.01% 0.07% (�0.16% to 0.26%) september 22, 2003 note: *obtained from www.etf.com in january 2018. †ivv was selected as one of three exchange-traded funds (etfs) that track the s&p 500 index, and is issued by blackrock. the other two, voo and spy, issued by vanguard and state street global advisors, have similar expense ratios and spreads, and could have also been chosen. the results that follow are generalizable to using any of these s&p 500 index etfs. 415j. dilellio / financial services review 27 (2018) 413-432 interested in a mixed stock/bond portfolio, we similarly replicated returns for a 1.25x leveraged and unleveraged bond etf, which are based on the barclay’s aggregate total return bond index (bloomberg index code lbustruu). we will refer to these funds simply as our leveraged or unleveraged “stock etf” and “bond etf” for the remainder of the paper. because we will be replicating the effect of constant daily leverage, we obtained our stock and bond index data on a daily basis. the total return data from bloomberg provided full years of daily returns from 1989 through 2017. we began by replicating the unleveraged stock and bond etf daily returns. let ri be the ith day’s total return (including dividends) of the underlying stock or bond index. to construct the unleveraged etf, we impose an expense ratio e, so that the yearly return r can be expressed as r ��1 � � i�1 n �1 � ri � e n� (1) where �i � 1 n represents the product over each ith day of a year with n total days. typically, n � 252, corresponding to an average of 21 trading days per month. we similarly constructed leveraged stock and bond etf daily returns. let r̃i be the leveraged etf daily return (including dividends) so that, before imposing any expenses or fees, r̃i � f * ri (2) where f is the leverage factor. for a 1.25x leveraged etf, we set f � 1.25. then, the annual return r̃ of the leveraged etf is r̃ ��1 � � i�1 n �1 � r̃i � ẽ n � �libori � x n � � f � 1�� , (3) where ẽ is the expense ratio of the leveraged fund, libori is the annual rate for the 1-month libor,2 and x is an additional financing cost imposed on the etf provider to borrow the money and leverage the underlying index’s assets. note that, with the exception of the potentially different expense ratios, eq. (3) reduces to (1) when f 3 1. table 2 lists values that will be used for the analysis that follows. it is also important to note that no attempt to model divergence from the daily benchmark will be made, which can occur when such a fund trades at either a discount or premium. we also assume that these investments are held in a tax deferred or exempt retirement account. lastly, we assume that slippage from incomplete baskets and managing leverage has a negligible long-term effect on returns. in the case of table 2 parameters to replicate unleveraged and leveraged etfs expense ratio, unleveraged fund (stock, bond) leverage factor expense ratio, leveraged fund (stock, bond) additional financing cost e f ẽ x 0.04%, 0.05% 1.25 0.32%, 0.32% 0.25% note: etf � exchange-traded funds. 416 j. dilellio / financial services review 27 (2018) 413-432 pplc, the fund holds ivv (ishares s&p 500 index etf) instead of individual securities included in the underlying index. we believe that ivv is well known to track the s&p 500 index very well. leverage is obtained by the fund investing in total return swaps on the s&p 500. data provided by direxion investments on pplc show that from the fund’s inception at the end of 2016, the tracking error because of managing this leverage was less than 0.001% on an annualized basis. we can now replicate annual returns and volatility for unleveraged and leveraged stock and bond etfs. annual return data and volatility for the stock funds appear in table a1, while table a2 contains annual return and volatility of the bond funds appear in table a2 of the appendix. the growth of a $1 investment in the 1.25x leveraged and underlying index etf constructed with these assumptions appear in the left pane of fig. 1. the right pane of fig. 1 provides a reference to the effect that borrowing costs have on limiting return growth, where higher libor rates reduce the total return for the leveraged stock and bond funds. the excess return, or the positive difference between the 1.25x stock etf and unleveraged stock etf, grows during bull markets, then shrinks during market downturns. the longest run of growing excess returns began in 2009, helped by two conditions. first, there has not been a significant stock market downturn since 2009, and volatility has generally trended down. additionally, the excess returns have benefited by the near 0% libor rates following the 2008 financial crisis. in the next section, we derive the conditions of annual returns and volatility of the underlying etf that produce excess returns, and compare it to these replicated returns. 2.1. derivation of return and volatility conditions that produce excess returns it is well understood that excess returns of a leveraged etf can be amplified by lower volatility, but returns on the underlying return are also important. here, we consider the annual return expectations for a leveraged etf, and compare it to the unleveraged return, net of fees and financing costs. leveraged etf annual expected returns can be expressed as fig. 1. (left pane) the 29 years of growth of a $1 investment in replicated 1.25x leveraged and underlying stock and bond etfs (right pane) 1-month libor annual returns over same periods. 417j. dilellio / financial services review 27 (2018) 413-432 e�r̃� � f � � f2 2 �2 (4) as shown in lu et al. (2012), where � is the rate of return and � is volatility. thus, the ability to generate excess returns occurs when r̃ � r. in terms of continuously compounded expected returns, we solve when e�r̃� � e�r� (5) for a given year. including expenses implies that returns are reduced accordingly. thus, �3 � � ẽ � �libori � x�� f � 1� (6) for the leveraged etf and �3 � � e (7) for the unleveraged etf. then, the inequality to produce excess returns becomes f �� � ẽ � �libori � x�� f � 1�� � f2 2 �2 � � � e � 1 2 � 2. (8) solving for � yields the quadratic relationship between volatility and returns of the underlying etf to produce excess returns. � � � f2 � 1� 2� f � 1� �2 � f �ẽ � �libori � x�� f � 1�� � e f � 1 . (9) the first term on the right hand side of eq. (9) contains the quadratic coefficient, and the second is a constant. because there is little correlation between the libor rate and volatility,3 the average libor rate from 1989 to 2017 of 3.57% was used that, along with the values in table 2 for the stock etf, simplify eq. (9) to � � 0.062 � 1.125�2. fig. 2 is a scatter plot of the annual returns and volatility from 1989 to 2017 of the unleveraged stock etf, where the caption for each point shows the excess return. the solid line represents eq. (9), so as expected, positive call-outs occur above the quadratic line and negative ones are below it. the data used to produce figs. 2 can be found in the appendix. similar results occur for leveraged bond funds, where lower volatility and higher returns often produce higher excess returns, as shown in fig. 3. however, there is less sensitivity to higher volatility, shown by a flatter quadratic. also, for eq. (9) to apply in fig. 3, we restricted annual returns to appear from 2009 to 2016, when libor rates were nearly constant, and averaged 0.27%. note that from eq. (9) that for every 1% increase in the annual libor rate, the quadratic in fig. 3 increases by 1.25%. 2.2. relationship of excess returns and underlying risk adjusted returns from the previous section, we established that for sufficiently large returns and sufficiently small volatility, excess returns occur. that is, excess returns obtained from using 418 j. dilellio / financial services review 27 (2018) 413-432 leverage is not only the function of upward trending markets, but the market’s associated volatility. in this section, we quantify this strong relationship from our dataset using a risk-adjusted return measure. we consider a simple linear regression of where the dependent variable is excess returns and the independent variable is the sharpe ratio of the underlying fig. 2. excess returns and their relationship to underlying volatility and returns, stock etf, 1989–2017. fig. 3. excess returns and their relationship to underlying volatility and returns, bond etf, 2009–2016. 419j. dilellio / financial services review 27 (2018) 413-432 etf. we hypothesize that annual excess returns, r̃ � r, increase with higher underlying sharpe ratio. formally, we express this relationship as a linear function of sharpe ratio �, as described in sharpe (1994), as r̃ � r � �0 � �1� � � (10) where �0 is the intercept, �1 is the slope, and � is a normally distributed random variable. we define the sharpe ratio in eq. (10) as � � r � rf � , (11) where r is the annual return of the underlying etf, rf is the annualized risk free rate obtained from the 30-day libor rate, and � is annualized volatility. fig. 4 shows this relationship for stocks, while fig. 5 shows a similar result for bonds based on their replicated annual values from 1989 to 2017, where the call outs in these figures show the year. the values of r2 at 82% for stocks and 97% for bonds produced in figs. 4 and 5 supports our hypothesis that the underlying sharpe ratio has a direct and strong correlation to excess returns before leverage. the coefficients and their statistical significance appear in table 3. all coefficients are statistically significant at either the 0.05 or 0.001 level. from these results, we can confirm that periods of upward trending prices with lower volatility will produce the highest excess returns. conversely, periods of flat prices and high volatility will produce lower or negative excess returns. fig. 4. excess returns from 1.25x leveraged stock etf versus its underlying etf, net of fees and expenses, 1989–2017. 420 j. dilellio / financial services review 27 (2018) 413-432 3. research methods we next investigate how a 1.25x stock and bond etfs perform in a periodically rebalanced portfolio. simulation methods are used to determine a distribution of 10 years of future price paths of the underlying assets. from these paths, we determine a distribution of annualized returns and sharpe ratios for each portfolio. we selected 10,000 trials, because they produced the half-width of a 95% confidence interval for returns which is less than 0.001.4 all simulation models were completed in excel using vba macros, and are available upon request. both of our sampling methods assume that price changes are independent of one another, so follow a markov or memoryless process, as suggested by fama (1965a, 1965b). our first approach is to simulate prices from a geometric brownian motion process. our second simulation approach is often termed “bootstrapping,” because it samples from a historical distribution. two excellent references for this methodology are davison and hinkley (1993) and efron and tibshirani (1993). fig. 5. excess returns from 1.25x leveraged bond etf versus its underlying etf, net of fees and expenses, 1989–2017. table 3 regression coefficients and p-values for coefficients appearing in figures 3 and 4, indicating statistical significance between the underlying sharpe ratio and excess returns n � 29 (1989–2017) coefficient p-value �1 (stocks) 0.0332 �0.001 �0 (stocks) �0.0091 �0.05 �1 (bonds) 0.0094 �0.001 �0 (bonds) �0.0035 �0.001 421j. dilellio / financial services review 27 (2018) 413-432 3.1. simulation of leveraged and unleveraged stock and bond prices with a gbm process the first model is a monte carlo simulation that assumes prices follow a geometric brownian motion (gbm) process. here, a future asset price at time (st�1) is found as st�1 � ste ��� �2 2 ��t���t� (12) where st is the current price, � is the return rate, � is the volatility or standard deviation of returns, �t is the time increment between the current and future price, and � is a random number obtained from the standard normal distribution. the use of a gbm process has a long history of modeling prices of financial assets. the gbm process generates future prices that are lognormally distributed with variance growing over time, assumes returns are normally distributed, and allows the modeler to either calibrate the gbm process parameters (� and �) from market data or set them to some future expectation. the downside of using gbm process to model security prices is that, because it assumes returns are normally distributed, it assumes a very small likelihood of extreme (greater than a few standard deviations) returns. in practice, it is well known that “fat tails” of returns exist, meaning the probability of more extreme returns is not modeled well by the gbm process. please see the seminal work by fama (1965a, 1965b) for a discussion of this behavior in commons stock. nevertheless, modeling stock and bond etf prices using a gbm process is fairly straightforward, and can be used to illustrate portfolio performance under a variety of parameters selected. using a gbm process, 10-year investment returns were evaluated for a variety of rebalancing strategies. drift and volatility were estimated from the 29 year history from 1989 to 2017, and appear in table 4 for stocks, bonds, and the risk-free asset. the simulation also included correlation between stock and bond returns set to �9.5%, as observed over this time period.5 correlations between bonds and risk-free rate and stocks and the risk-free rate were negligible over this time period, so this correlation was not included in the simulated results. additional parameters for transaction costs were also included. the bid-ask spread for stock and bonds was set to 0.01%, and for leveraged stock and bonds to 0.04%. trading commissions were set to $4.95 per trade for an account with $200,000 at the start of each 10 year simulation. results for the baseline gbm simulation appear in table 5. as expected, the mean annualized return for the unleveraged stock holdings are 9.5%, which is less than the continuous return parameter of 10.6% assumed in table 5. this effect is aptly described in winston (2008), because over time, volatility creates a drag on the growth rate of a stock modeled by a gbm process. however, a little surprisingly, average annualized unleveraged bond returns are slightly higher. we attribute this to the much smaller volatility of the bond table 4 continuous return and volatility parameters in geometric brownian motion (gbm) simulation stock bond risk-free asset � � � � � � 10.6% 17.8% 5.9% 3.9% 3.2% 0.2% 422 j. dilellio / financial services review 27 (2018) 413-432 markets effectively eliminating volatility drag, as well as the effect of continuously versus annually compounded rates causing a small perturbation on the numerical results.6 table 5 also shows the average annualized returns for the leveraged stock and bond etfs, where excess stock annual returns are 0.8% and excess bond annual returns are 0.3%. reviewing the mean sharpe ratios of the stock and bond etfs, we see that sharpe ratios are modestly reduced by the 1.25x stock and bond etfs. we attribute this to financing costs exceeding the risk-free rate, higher expense ratio of leveraged funds, and the path dependent effect of daily leverage. the effect of leverage on sharpe ratios is also consistent with mulvey et al. (2007). table 5 also displays the performance of the 60/40 leveraged stock/bond portfolio that is periodically rebalanced. we observe that excess returns are approximately 0.8% annually, and does not appear to depend on the rebalancing interval, which is beneficial to individual investors who may prefer to rebalance once a year. unfortunately, risk-adjusted return measured by sharpe ratio reduced by about 5% when rebalanced annually or more. table 5 also shows that rebalancing on a quarterly basis provided the optimal sharpe ratio, indicating that the effect of spreads and trading commissions associated with portfolio turnover for a 1.25x leveraged stock and bond etf do not significantly cause a drag on risk-adjusted performance. with our baseline results established in table 5, we next investigated the sensitivity of these results to the volatility of the stock etf. to this end, we kept all other parameters constant, and considered high and low volatility values. as expected, decreasing volatility increases excess returns. table 6 shows the results when stock volatility is decreased to 10% and increased to 25% with an annual rebalancing policy to 60/40 stock/bonds. decreasing the volatility to 10% increases excess return to 1.0%, while increasing the volatility to 25% reduces excess return to 0.4%. sharpe ratios tell a similar story, where lower volatility table 5 averaged annualized returns and sharpe ratios for 10,000 trial gmb simulation 10,000 trial gbm mean annualized return mean sharpe ratio unleveraged leveraged unleveraged leveraged 100% stock etf 9.5% 10.3% 0.118 0.114 100% bond etf 6.0% 6.3% 0.202 0.183 60/40 stock bond portfolios annual 8.5% 9.3% 0.145 0.138 quarterly 8.5% 9.3% 0.148 0.140 monthly 8.4% 9.2% 0.145 0.138 note: etf � exchange-traded funds; gbm � geometric brownian motion. table 6 averaged annualized returns and sharpe ratios for above and below average long-term volatility for an annually rebalanced portfolio of 60/40 stock and bonds annual rebalancing policy sensitivity to volatility mean annualized return mean sharpe ratio unleveraged leveraged unleveraged leveraged � (baseline from table 5) 8.5% 9.3% 0.145 0.138 � (low volatility) 8.9% 9.9% 0.260 0.247 � (high volatility) 7.9% 8.3% 0.101 0.096 423j. dilellio / financial services review 27 (2018) 413-432 increases the sharpe ratio of both the leveraged and unleveraged portfolios. however, the 5% reduction in sharpe ratio because of leveraged shows little, if any, significant change because of changing volatility. we can conclude that the reduction in sharpe ratio because of using the 1.25x stock and bond etfs is insensitive to the underlying stock etf volatility, when returns are kept constant. the last row in table 6 provides for an interesting extension to the monthly 1.25x mutual fund dxnlx mentioned previously. using data obtained from nasdaq global indices,7 237 monthly returns were available from may 31, 1999 through december 31, 2018. the mean annualized return over this period was 9.3%, and mean annualized volatility was 24.3%. thus, the sensitivity to higher volatility shown in table 6 can be representative of a daily leveraged stock fund that had a higher volatility than the s&p 500, like the nasdaq-100. next, we investigated how the simulation results would change in table 5 change if, instead of 1.25 daily leverage, we used 1.25 monthly leverage. table 7 shows this effect, with all other model parameters set to their baseline values from tables 2 and 4, to isolate the effect of monthly versus daily leverage on absolute and risk-adjusted returns. the results in table 7 suggest that choosing a fund that has monthly versus daily 1.25 leverage changes stock and bond fund returns by a negligible amount. there is also a small, yet nearly insignificant change in sharpe ratio in the third decimal place. this very small affect yields no change to the 60/40 leveraged portfolios, regardless of rebalancing period. we can conclude from these results that an investor can be indifferent regarding choosing a broad-based stock or bond fund with monthly versus daily leverage, assuming expenses are kept constant. 3.2. gbm limitations the gbm process generally models daily returns of stocks and bonds well. however, exceptions occur. the histograms in fig. 6 shows daily returns from 1989 to 2017 for our replicated stock and bond etfs. in the top pane of fig. 6, the normal distribution, labeled “normal, �” appears to miss the peak of the observed stock market returns. reducing � in the normal distribution to 75% of its original value appears to improve the fit for observations near the center of the distribution, but aggravates the already poor fit of return extremes, such as the daily returns from october of 2008. the normal distribution of bond returns exhibits table 7 averaged annualized returns and sharpe ratios for 10,000 trial gbm simulation (monthly vs. daily leverage) 10,000 trial gbm mean annualized return mean sharpe ratio leveraged (monthly) leveraged (daily) leveraged (monthly) leveraged (daily) 100% stock etf 10.2% 10.3% 0.113 0.114 100% bond etf 6.3% 6.3% 0.179 0.183 60/40 stock bond portfolios annual 9.2% 9.3% 0.138 0.138 quarterly 9.3% 9.3% 0.140 0.140 monthly 9.2% 9.2% 0.138 0.138 note: etf � exchange-traded funds; gbm � geometric brownian motion. 424 j. dilellio / financial services review 27 (2018) 413-432 similar, but less pronounced differences between observed daily returns and the theoretical expectations from a normal distribution. thus, the volatility assumption made in a gbm simulation may not fit the return distribution well. because we know that excess returns are sensitive to volatility, an alternate simulation approach termed “bootstrapping” is used to evaluate the performance of leveraged etfs under actual market conditions. the bootstrapping approach also benefits from imposing the risk-free rates observed during particular market periods, as opposed to modeling the evolution of the risk-free asset as a gbm process. 3.3. bootstrapping using 10-year historical periods because the return distribution shown above suffers from some modeling issues, we also investigate leveraged etf performance using a bootstrapping model. applying the methodfig. 6. daily return distribution of stock (top pane) and bond market (bottom pane) indices. two normal distributions are also shown, with volatility estimates using historical returns from 1989 to 2017. reducing the volatility appears to provide a slightly improved fit near the center of the distribution, but worsens the fit in the distribution tails. 425j. dilellio / financial services review 27 (2018) 413-432 ology proposed in dilellio et al. (2014), we generate returns by randomly sampling empirical return histories observed from market data. by using this alternative simulation approach, the “fat tails” that occurred over the empirical return histories can be properly included in a portfolio performance evaluation, as well as the higher peaks near the center of the distribution of stock and bond returns. the following steps are adapted from the methodology found in the appendix of dilellio et al. (2014). a key input to any bootstrapping simulation is the time period selected. to select two meaningful periods representing extremes in stock market returns and volatility, we obtained 51 years of daily returns for the s&p 500 total return index from the center for research in security prices (crsp). next, we determined 10-year simple moving average annual returns and volatilities, and divided them by their corresponding 51 year averages of 11.1% for annual returns and 15.1% for annual volatility, respectively. thus, values above one represent above average returns/volatilities, and values less than one represent below average returns/volatilities. the results appear in fig. 7, where the call outs designate the final year in each 10-year moving average. to select two extremes, we selected the 10 years ending in 2000 and 2010. the first period is from 1991 to 2000, when above average returns and below average volatility of stocks occurred. this first period corresponds to the equity bull market often associated with the tech bubble. the second period is from 2001 to 2010, when below average stock returns occurred with higher than average volatility. this second period includes the tech bubble correction as well as the financial crisis of 2008. these periods also benefit from being non-overlapping, so using them eliminates the possibility of temporal correlation. it should be noted that there are some other interesting 10-year periods that may be worth investigating. for example, the 10 years ending in 1990 and 1979. unfortunately, our bloomberg data set for daily stock and bond indices does not extend this far back, as noted previously. also, the 1-month libor rates were not available on a daily basis before the mid-1980s. tables 8 and 9 below highlight how bootstrapping from these unique 10-year periods affect excess returns. in table 8, the period of 1991–2000 produced excess returns from the leveraged stock(bond) etf of 2.5% (0.3%). these returns then translate to excess returns for a 60/40 stock bond portfolio, with a value of approximately 1.7% annually when rebalanced annually, quarterly, or monthly. thus, given the small spreads and low turnover of the 1.25x step simulation process 1 assign a numerical index of 1 to n for each of the daily returns observed in a historical 10-year period. assuming 252 trading days in a year yields n � 2,520. 2 for each day, determine the cumulative return up to and including the previous 21 days, so that the accumulated returns from these days are representative of the distribution of monthly returns. 3 generate 120 random numbers ranging from 1 to n. 4 select returns from the distribution of representative monthly returns determined in step 2 using the random numbers found in step 3, and generate 10 years of monthly returns. use the same set of random numbers for each trial to select monthly returns from each of investment category (stock, bonds, and risk-free asset), ensuring that the historical correlation among assets is preserved. 5 repeat for 10,000 trials, collecting annualized return and sharpe ratio for each trial. 6 determine mean annualized return and sharpe ratios found from the 10,000 trials generated in step 5. 426 j. dilellio / financial services review 27 (2018) 413-432 portfolio, rebalancing frequency does not affect excess returns. again, these excess returns are because of this period containing historically lower volatilities and higher returns, as predicted by trainor and carrol (2013). unfortunately, the favorable excess returns do not translate to an increase in the mean sharpe ratio, which is similar to results from the gbm simulation approach. from table 8, we see that the stock etf sharpe ratio is reduced by 3% fig. 7. the 10-year moving average of stock market returns and volatilities divided by 51 year average. values above one represent above average returns/volatilities, and values less than one represent below average returns/volatilities. table 8 bootstrapping simulated results from 1991 to 2000, where 60/40 portfolio excess returns were approximately 1.7% annually; sharpe ratio reduced by 4% 10,000 trial bootstrapping 1991–2000 mean annualized return mean sharpe ratio unleveraged leveraged unleveraged leveraged 100% stock etf 17.8% 20.3% 0.269 0.260 100% bond etf 8.0% 8.3% 0.194 0.175 60/40 stock bond portfolios annual 14.0% 15.7% 0.283 0.272 quarterly 13.9% 15.6% 0.284 0.273 monthly 13.9% 15.6% 0.285 0.274 note: etf � exchange-traded funds. 427j. dilellio / financial services review 27 (2018) 413-432 and bond etf sharpe ratio is reduced by 10% with 1.25x leverage. the 60/40 stock/bond portfolios see the sharpe ratio reductions of 4%, regardless of the rebalancing frequency. table 9 shows a very different story, and summarizes the performance by sampling from the 2000–2010 return history with its high volatility and low returns of the stock etf. here, excess returns for the leveraged stock fund are no longer positive, with a value of �1.3%. fortunately, the leveraged bond etf fared better, with a 0.5% annual excess return. however, this positive excess return from the leveraged bond etf could not be entirely offset by the leveraged stock loss, so the 1.25x leveraged 60/40 stock bond investment had negative excess returns of 0.4% annually under a variety of rebalancing policies ranging from annually to quarterly. results for the mean sharpe ratio are more adversely affected when the volatility is high and returns are low. leverage in the stock etf reduces the averaged sharpe ratio by 70%, while the bond etf reduction is a more modest 8% reduction. the stock etf sharpe ratio reduction is extreme, because the average daily return of the leveraged stock etf narrowly exceeded the risk-free rate in 2001–2010. consequently, the leveraged 60/40 portfolio reduces the sharpe ratio by 23%, regardless of rebalancing frequency. 4. conclusions this article investigated the use of 1.25x leveraged stock and bond exchange-traded funds (etfs) as a stock and bond asset allocation strategy. analytical conditions of return and volatility leading to excess returns over unleveraged funds is derived and agree favorably with empirical findings from 1989 to 2017, including all relevant costs. we also show the relationship of excess returns and the underlying’s sharpe ratio, demonstrating high levels of statistical significance. we use a gbm process to simulate asset prices to determine a first approximation of future portfolio returns. we find that at 60/40 leveraged stock/bond fund generates excess returns of 0.8% annually (net of fees and expenses) but reduces the sharpe ratio by 5% because of the financing exceeding the risk-free rate, higher volatility, and higher expense ratios of leveraged funds. we also show that changing leverage from daily to monthly has a negligible effect on the simulation results. table 9 bootstrapping simulated results from 2001 to 2010, where 60/40 portfolio excess returns were approximately �0.4% annually; sharpe ratio reduced by 23% 10,000 trial bootstrapping 2000–2010 mean annualized return mean sharpe ratio unleveraged leveraged unleveraged leveraged 100% stock etf 1.4% 0.1% 0.010 0.003 100% bond etf 5.8% 6.3% 0.238 0.218 60/40 stock bond portfolios annual 3.6% 3.2% 0.043 0.033 quarterly 3.5% 3.1% 0.043 0.033 monthly 3.5% 3.1% 0.043 0.034 note: etf � exchange-traded funds. 428 j. dilellio / financial services review 27 (2018) 413-432 by using a bootstrapping simulation approach, non-overlapping periods are also evaluated using specific market conditions. we show the effect the 1.25x leverage can produce excess returns for a 60/40 portfolio of stocks and bonds of 1.7% (�0.4%) annually in higher return/lower volatility (lower return/higher volatility) equity markets, relative to their nonleveraged counterparts, and net of fees and expenses. the resulting sharpe ratios are reduced between 4 and 23%, depending on the behavior of the stock and bond market. we can conclude from these findings that some leverage, such as the 1.25x factor assumed here, can have long term net return benefits to a 60/40 stock/bond portfolio under upward trending and lower volatility conditions of the underlying asset. additionally, while some individual investors may be unwilling to reduce their sharpe ratio, the modest reduction found here may be acceptable to individual investors seeking to increase their returns and not able or willing to access margin. we conclude that this investment strategy could be well suited for investors more interested in total returns during upward trending markets. 5. future work there are several additional avenues for future work in this area. perhaps the most obvious question is related to the best amount of fractional leverage. this determination would likely have a behavioral component, since additional leverage will likely magnify the marginal reduction in sharpe ratio, which at some point an investor may not be willing to consider. one could also expand this work to include other equity asset classes to enhance diversification benefits and reduce portfolio risk. these additional asset classes could include small caps, international, commodities, and real estate investment trusts. lastly, we could evaluate how fractionally leveraged etfs may exist on the efficient frontier, and develop optimal portfolio strategies based on these new insights. notes 1 https://www.fool.com/knowledge-center/can-you-trade-on-margin-in-an-ira.aspx. 2 https://fred.stlouisfed.org/series/usd1mtd156n. 3 from 1989 to 2017, the correlation between annual volatility of the s&p500 and the annual libor rate was �8.8%, suggesting no significant relationship between these two variables. 4 the primary driver on sample size was the underlying asset’s volatility. 5 in the article by johnson et al. (2013), the correlation between stocks and treasury bonds is shown to change over time because of macroeconomic conditions. however, we did not attempt to include an econometric model like the one developed by these authors, instead using the long-term correlation. 6 in this example, the continuous rate for bonds was 5.9% and volatility 3.9%, so that exp(5.9% 0.5*3.9%2) � 1.060, implying an annually compounded rate was 6%, as shown in table 5. 7 https://www.quandl.com/data/nasdaqomx/xndx-nasdaq-100-total-return-xndx. 429j. dilellio / financial services review 27 (2018) 413-432 acknowledgments financial support for this project was made possible by direxion funds and pepperdine graziadio business school and the julian virtue professorship. the author wishes to thank andy o’rourke for his insights into the financing aspects of leveraged etfs, and his sharing of operational data from direxion investments to quantify the effect of slippage on replicated returns, and consequently tracking error caused by incomplete baskets and managing leverage. the author also appreciates the excellent feedback from dr. darrol stanley on an earlier version of this manuscript, and proofreading support from barb drayer. additionally, two blind reviewers provided excellent recommendations on enhancements to this manuscript, which substantially added to the quality and relevance of this work. i also wish to thank the participants at the 2018 academy of financial services annual meeting in chicago on several suggestions for extensions and future work. lastly, although every attempt was made to eliminate errors from this paper, any remaining issues are solely the responsibility of the author. table a1 annualized returns and volatility of unleveraged and leveraged stock etf year n s&p 500 index annual return s&p 500 index annual volatility s&p 500 etf annual return 1.25x annual volatility 1.25x annual return excess return 1989 252 31.68% 13.03% 31.63% 16.28% 36.934% 5.30% 1990 253 �3.10% 16.10% �3.14% 20.13% �6.58% �3.43% 1991 253 30.47% 14.32% 30.41% 17.90% 36.41% 6.00% 1992 254 7.62% 9.73% 7.58% 12.16% 8.01% 0.43% 1993 253 10.08% 8.63% 10.03% 10.79% 11.30% 1.26% 1994 252 1.32% 9.82% 1.28% 12.28% �0.01% �1.29% 1995 252 37.58% 7.83% 37.52% 9.78% 46.09% 8.57% 1996 254 22.96% 11.82% 22.91% 14.78% 26.95% 4.04% 1997 253 33.36% 18.16% 33.31% 22.70% 40.02% 6.71% 1998 252 28.58% 20.29% 28.53% 25.37% 33.64% 5.11% 1999 252 21.04% 18.08% 20.99% 22.60% 24.19% 3.20% 2000 252 �9.10% 22.22% �9.14% 27.77% �13.66% �4.52% 2001 248 �11.89% 21.38% �11.92% 26.72% �16.36% �4.44% 2002 252 �22.10% 26.04% �22.13% 32.55% �28.18% �6.04% 2003 252 28.68% 17.07% 28.63% 21.34% 35.51% 6.87% 2004 252 10.88% 11.10% 10.84% 13.87% 12.71% 1.87% 2005 252 4.91% 10.29% 4.87% 12.86% 4.71% �0.16% 2006 251 15.79% 10.02% 15.75% 12.53% 17.97% 2.22% 2007 251 5.49% 15.96% 5.45% 19.95% 4.70% �0.75% 2008 253 �37.00% 41.05% �37.02% 51.31% �45.91% �8.89% 2009 252 26.46% 27.28% 26.41% 34.09% 31.95% 5.54% 2010 252 15.06% 18.06% 15.02% 22.58% 18.03% 3.02% 2011 252 2.11% 23.29% 2.07% 29.11% 1.33% �0.74% 2012 250 16.00% 12.70% 15.96% 15.88% 19.56% 3.60% 2013 252 32.39% 11.07% 32.34% 13.84% 41.12% 8.79% 2014 252 13.69% 11.38% 13.64% 14.23% 16.66% 3.02% 2015 252 1.38% 15.49% 1.34% 19.37% 0.91% �0.43% 2016 252 11.96% 13.10% 11.92% 16.37% 14.28% 2.36% 2017 251 21.83% 6.67% 21.78% 8.33% 27.07% 5.28% note: etf � exchange-traded funds. 430 j. dilellio / financial services review 27 (2018) 413-432 references agrrawal, p., & clark, j. m. (2009). determinants of etf liquidity in the secondary market: a five-factor ranking algorithm. institutional investor journals, 2007, 59–66. avellaneda, m., & zhang, s. (2010). path-dependence of leveraged etf returns. siam journal of financial mathematics, 1, 568–603. barnhorst, b. c., & cocozza, c. r. (2010). inverse and leveraged etfs: considering the alternatives. journal of financial planning, 24, 44–49. cheng, m., & madhavan, a. (2009). the dynamics of leveraged and inverse exchange traded funds. journal of investment management, 7, 43–62. davison, a. c., & hinkley, d. v. (1993). bootstrap methods and their application. new york, ny: cambridge series in statistical and probabilistic mathematics. dilellio, j., hesse, r., & stanley, d. (2014). portfolio performance with inverse and leveraged etfs. financial services review, 23, 123–149. efron, b., & tibshirani, r. (1993). an introduction to the bootstrap. boca raton, fl: chapman and hall/crc. table a2 annualized returns and volatility of unleveraged and leveraged bond etf year n aggregate bond index annual return aggregate bond index annual volatility aggregate bond etf annual return 1.25x annual volatility 1.25x annual return excess return 1989 250 14.53% 3.99% 14.47% 4.98% 15.32% 0.84% 1990 251 8.96% 4.26% 8.91% 5.32% 8.61% �0.30% 1991 248 16.00% 3.47% 15.95% 4.34% 18.17% 2.22% 1992 251 7.40% 3.73% 7.35% 4.66% 7.88% 0.54% 1993 249 9.75% 3.23% 9.69% 4.04% 11.01% 1.31% 1994 249 �2.92% 5.13% �2.97% 6.41% �5.09% �2.13% 1995 248 18.47% 4.04% 18.42% 5.05% 21.31% 2.89% 1996 249 3.63% 4.74% 3.58% 5.92% 2.73% �0.85% 1997 250 9.65% 3.34% 9.60% 4.18% 10.20% 0.60% 1998 250 8.69% 3.60% 8.63% 4.50% 9.01% 0.38% 1999 250 �0.82% 3.94% �0.87% 4.93% �2.70% �1.83% 2000 251 11.63% 3.54% 11.57% 4.43% 12.47% 0.90% 2001 247 8.44% 4.46% 8.39% 5.57% 9.16% 0.77% 2002 251 10.25% 3.84% 10.20% 4.79% 12.03% 1.83% 2003 250 4.10% 4.27% 4.05% 5.33% 4.41% 0.36% 2004 251 4.34% 3.89% 4.29% 4.86% 4.63% 0.35% 2005 250 2.43% 2.89% 2.378% 3.61% 1.78% �0.60% 2006 250 4.33% 2.78% 4.28% 3.47% 3.71% �0.57% 2007 250 6.97% 3.51% 6.91% 4.38% 6.95% 0.03% 2008 251 5.24% 5.91% 5.19% 7.39% 5.42% 0.24% 2009 250 5.93% 4.49% 5.88% 5.61% 6.94% 1.06% 2010 252 6.54% 3.72% 6.49% 4.64% 7.73% 1.24% 2011 250 7.84% 3.85% 7.79% 4.81% 9.39% 1.60% 2012 250 4.22% 2.50% 4.16% 3.12% 4.82% 0.66% 2013 250 �2.02% 3.30% �2.07% 4.13% �2.96% �0.88% 2014 250 5.97% 2.66% 5.91% 3.32% 7.05% 1.14% 2015 251 0.55% 3.78% 0.50% 4.72% 0.23% �0.27% 2016 250 2.65% 3.19% 2.60% 3.99% 2.78% 0.19% 2017 250 3.54% 2.80% 3.49% 3.50% 3.75% 0.26% note: etf � exchange-traded funds. 431j. dilellio / financial services review 27 (2018) 413-432 fama, e. (1965a). random walks in stock-market prices. selected papers no. 16, booth school of economics, university of chicago, p. 7. fama, e. (1965b) the behavior of stock-market prices. journal of business, 64, 34–105. giese, g. (2010). on the risk-return profile of leveraged and inverse etfs. journal of asset management, 11, 219–228. guedj, i., guohua, l., & mccann, c. (2010). leveraged and inverse etfs, holding periods, and investment shortfalls. journal of index investing, 1, 45–57. johnson, n., vaik, v., pedersen, n., & sapra, s. (2013). the stock-bond correlation. pimco quantitative research. https://nl.pimco.com/en-nl/insights/viewpoints/quantitative-research-and-analytics/the-stock-bondcorrelation/, p. 1–12. lu, l., wang, j., & zhang, ge. (2012). long term performance of leveraged etfs. financial services review, 21, 63–80. mulvey, j. m., ural, c., & zhang, z. (2007). improving performance for long-term investors: wide diversification, leverage, and overlay strategies. quantitative finance, 7, 175–187. ott, r. a., & zimmer, t. e. (2016). determining the return maximizing portfolio leverage and its limitations. financial services review, 25, 415–425. sharpe, w. f. (1994). the sharpe ratio. journal of portfolio management, 21, 49–59. trainor, w. j., & baryla, e. a. (2008). leveraged etfs: a risky double that doesn’t multiply by two. journal of financial planning, 21, 48–55. trainor, w. j., & carroll, m. g. (2013). forecasting holding periods for leveraged etfs using decay thresholds: theory and applications. journal of financial studies and research, 2013. article id 715425, 12 pages. winston, w. (2008). financial models 2: using simulation and optimization (2nd ed.). ithaca, ny: palisade. 432 j. dilellio / financial services review 27 (2018) 413-432 pii: s1057-0810(99)80004-3 financial services review. 7(2): 83-93 copyright © 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. closed-end investment companies: historic returns and investment strategies carolyn reichert and j. douglas t i m m o n s studies conducted in the past have identified inefficiencies in the market for closed end investment company ( ceic) shares. in addition, studies have demonstrated the potential for trading strategies to exploit these inefficiencies. the purpose of this paper is to investigate the possibility of achieving excess returns through the utilization of rel atively simple strategies not requiring continuous monitoring of discount(s) or frequent trading. our investigation demonstrates that realizing excess returns through the use of simple mechanical trading strategies will not be possible. i. i n t r o d u c t i o n the purpose of this study is to test for closed-end investment company (ceic) market inefficiency, and therefore the possibility of excess profits. prior studies have developed complex trading strategies that outperformed the market. our concern is whether small investors can expect to exceed market returns using uncomplicated techniques to select funds. the plan to be tested is much like the "dow dogs" strategy of buying the five dow jones industrial stocks that have the highest yield. investors who bought these dow stocks at the beginning of each year, without consideration of any other factors, have fared well (u.s. news & world report, 1996). should ceic investors who buy discounted funds at the beginning of the year, and rebalance their investments after one-year holding periods expect to profit? this study evaluates four simple ceic trading strategies. using one year holding peri ods, the plans select funds trading at either maximum discounts or higher than average dis counts. the appeal of these investment strategies is their simplicity. they provide a longer holding period and fewer transactions than previous studies. the initial investment is smaller, and short selling is not required. this makes the strategies particularly appealing to small investors, who are active in the ceic market. these investors lack the time and resources to follow the more complex systems developed in earlier studies. in fact, any investor could easily duplicate these strategies. carolyn reichert ° southwest texas state university. j. douglas timmons ° department of economics and finance, p.o. box 27, middle tennessee state university, muffreesboro, tn 37132. 84 financial services review 7(2) 1998 ii. l i t e r a t u r e r e view premiums and discounts on ceic shares are determined by comparing the market price of the shares to the net asset value (nav) of the fund' s holdings on a per share basis. if the fund's market price is less than the nav, then the fund sells at a discount; if it is greater than the nav, it trades at a premium. when the discount or premium exceeds a value that can be adequately explained by market forces (malkiel, 1977; pratt, 1966), the departure from nav is deemed excessive and suggests market inefficiency in setting fund prices. thus, the size of the discount or premium is a key consideration in selecting ceic shares. most closed-end funds sell at a discount, and this discount has persisted over time (pontiff, 1994). this paper focuses on profitable trading strategies, rather than explaining the exist ence of the discount. several researchers have devised trading strategies that attempt to select profitable funds based on the size of the discount or premium. thompson (1978) reveals that funds trading at a premium are bad investments. he creates portfolios that incorporate all funds trading at a discount. he finds significant performance that is consistent across benchmark portfolios, even after adjusting the returns for tax receipt distributions and reinvestment options. richards, fraser, and groth (1980) focus on two strategies: buy and sell trading points and filter rules. their strategies involve buying ceic shares trading at a discount and shorting ceic shares trading at a premium. using weekly rebalancing, all of their strategies outperform the s&p 500. they find that the extreme buy and sell points and larg est filters produce the highest returns. anderson (1986) extends their study to investigate three different time periods. he demonstrates that only the buy and sell points provide excess returns for all three periods. brauer (1988) constructs a frequency distribution based on the potential for open ending to get a cutoff for inclusion in the portfolio. his portfolios outperform the s&p 500. pontiff (1994) separates the funds into groups based on the size of the premium. he finds that premium funds are bad investments. he shows that funds with 20% discounts have expected twelve-month returns that are 6% greater return than non-discounted funds. prior studies of closed-end funds have not explicitly considered transaction costs, although pontiff (1994) does state that the round trip transaction costs needed to eliminate the abnormal returns in his study are 8.25% for buying securities and 3.13% for shorting them. pesaran and timmerman (1994) use .5% and 1% rates in their paper on stock market trading with transaction costs. this produces round trip transaction costs of 1% to 2% per ceic share. taxes are another important consideration. morris and scanlon (1996), malk iel (1977), and kim (1994) examine the impact of taxes in explaining the discount. they use a variety of tax rates including 25%, 31% and split rates (31% on dividends and 28% on capital gains). they find that taxes are an important factor in explaining the fund's dis count. small investors are important in the closed-end fund market. palomino (1996) hypoth esizes that small (or noise) traders may earn higher expected utility than rational investors (by deviating from the nash equilibrium strategy) because they create additional market volatility. thus, rational investors are reluctant to trade in small markets, and noise trader risk persists. some ceic funds are thinly traded, and small investors have a greater impact in these markets. the initial public offering (ipo) evidence supports the idea that small investors are active in the closed-end fund market. an examination of closed-end fund ipos reveals no abnormal performance in the first two days of trading (barry & jennings, closed.end investment companies 85 1993). subsequently, the price declines sharply as large traders sell to small "noise" traders (barry & peavy, 1990; weiss-hanley, lee, & seguin, 1996). the primary goal of any trading strategy is to outperform the market portfolio on a risk-adjusted basis. brickley and schallheim (1985), brauer (1984, 1988), thompson (1978) and pontiff (1994) all use market and risk-adjusted returns (market model). how ever, there is some evidence that the market model may be inadequate for examining ceic funds. thompson (1978) observes that the two-parameter market model does not describe the return generating process for ceic funds. brickley and schallheim (1985) find evi dence that the market model may be inadequate if new (uncertain) information is likely to occur in the marketplace. barry and peavy (1990) discover that the ipos of ceic funds have low betas in the first trading months due to extensive price stabilization. beta increases as funds season in the after market. pontiff (1994) finds that beta increases as fund premiums increase. if new information about the ceic fund is contained in the pre mium, this could bias the market model results. instead of using the market model, rich ards, fraser, and groth (1980) and anderson (1986) compare the return and standard deviation of the proposed strategies to the s&p 500. this avoids some of the market model 's problems. closed-end funds that convert to open-end could bias the results. if the fund liquidates or open-ends, the value of the fund's shares should return to the nav. brickley and schall heim (1985) and brauer (1984, 1988) present evidence that funds with higher than average discounts are more likely to open-end. after the announcement date, the discount narrows. investors can earn abnormal returns if they buy the fund (even after the announcement) and hold it until it is open-ended or liquidated. iii . strategies and data previous studies use strategies with complex investing rules, weekly trading and rebalanc ing, large initial investments, or short selling. in contrast, small investors typically want simple trading rules with a minimum of rebalancing. they lack the time for frequent trad ing and the money for large initial investments. small investors are often encouraged to buy assets and hold them for long periods of time to minimize transaction costs. risk aver sion and limited investment knowledge deter them from shorting shares. this paper tests four simple trading strategies to determine if small investors can earn abnormal returns with minimal effort. they are variations of buy and hold strategies with annual rebalanc ing. this allows the investor to minimize transaction and monitoring costs. they require five or fewer funds, reducing the investment size, and they do not involve shorting ceic shares. four investment strategies are tested: (1) single most discounted ceic; (2) single most relatively discounted ceic; (3) five most discounted ceics; and (4) five most relatively discounted ceics. for the first strategy, the investor selects the fund trading at the largest percentage discount from the nav. this is the most discounted ceic. these shares are purchased and held for one year. the process is repeated at the end of each year. if the year end calculations show that the most discounted issue has changed, then the original shares are sold and the new most discounted shares are purchased, using all of the funds released by the sale of the old issue. 86 financial services review 7(2) 1998 t h e s e c o n d s t ra tegy i nvo lves a s imi la r p rocess , excep t tha t e ach i s sue ' s d i s coun t is c o m p a r e d to its f i ve -yea r r u n n i n g average . t h e issue tha t is m o s t d i s c o u n t e d c o m p a r e d to its a v e r a g e is the m o s t relatively d i s c o u n t e d ceic . w h e n f ive years o f h i s to ry are unava i l able , the r u n n i n g ave r age is ba sed on ava i l ab le data. t h e th i rd and four th s t ra tegies are mul t ip le fund s t ra tegies tha t i nvo l ve se lec t ing the f ive m o s t d i s c o u n t e d or r e la t ive ly dis c o u n t e d funds . w h e n us ing mul t ip le funds , the i nves to r ho lds an equa l a m o u n t o f all f ive funds in a por t fol io . c a r r y o v e r funds are he ld w i t h o u t t r ansac t ions . i f a f und is rep laced , the p roceeds f r o m the sale o f tha t f und are e v e n l y d i s t r ibu ted in to the n e w funds p laced in the por t fol io . t a b l e 1 f -tes t for equa l v a r i a n c e s fund name classification period of inclusion adams express diversified 1971 1995 advance investors diversified 1974-1976 america south africa (asa) ltd. specialized 1971 1995 american utility shares specialized 1973-1978 carriers-general diversified 1971 1981 central fund of canada international 1988-1990 central securities corp. specialized 1986-1995 cypress fund specialized 1987-1990 dominick fund, inc., the diversified 197 l1974 drexel utility shares specialized 1973-198 l duff & phelps sel. utility specialized 1988-1990 emerging medical specialized 1986-1987 first australia international 1987-1995 general american investors diversified 1971-1994 germany fund international 1987-1995 griesedieck co., the diversified 1971 1975 highland capital corp. specialized 1976-1982 international holdings diversified 1971 1975 japan fund, inc., the international 1971-1987 keystone inc. specialized 1973-1977 korea fund international 1986-1995 lehman corp., the diversified 1971 1990 madison fund, inc. diversified 1971 1982 malaysia fund international 1988-1990 mexico fund international 1987-1990 national aviation corp. specialized 1971 1979 nautilus fund diversified 1980-1985 niagara share corp. diversified 197 l1990 petroleum & resources corporation specialized 1971-1995 pilgrim regional specialized 1987-1995 precious metals specialized 1975-1983 providence investors, inc. diversified 1971-1976 reit income fund specialized 1973-1980 s-g securities, inc. specialized 1974-1979 source diversified 1975-1995 surveyor fund specialized 1971 1973 thai fund international 1989-1990 tri-continental corp. diversified 1971 1995 u.s. & foreign securities corp. diversified 1971-1984 value line development specialized 1976-1979 zweig fund diversified 1987-1995 closed.end investment companies 87 if funds that subsequently liquidate or open-end have larger discounts than those that do not, our strategies would tend to pick these funds. if we find positive abnormal returns, it could be due to the influence of these funds on the results. since the focus of the paper is making abnormal profits using a simple trading rule, this would support the use of our rules rather than detract from them. the data set consists of funds from the weisenberger investment company survey (1965-1990) defined as "specialized," "international" or "diversified." specialized funds consist predominantly of securities from one industry or group of closely related industries, or stocks of a specific type (i.e., letter stock). international funds invest chiefly in the secu rities of foreign companies. diversified funds hold a well-diversified portfolio of invest ments. due to changes in reporting by weisenberger, the s&p nyse stock reports (1995 1996) are used for the 1991-1995 data. the final database includes 41 funds: 17 specialized, 8 international, and 16 diversi fied. a complete listing of the funds originally included in the database is provided in table 1. the data spans december 31, 1965, to december 31, 1995. five years of data are needed to construct the relatively most discounted strategies, thus all trading strategies begin at year-end 1970 and end at year-end 1995. the data collected for each fund includes classification, year-end price and discount, and distributions during the year. capital gains distributions are included as distributions. the year-end price is the actual transaction price. for informational purposes, results are provided for four "pooled" portfolios: (1) all specialized funds; (2) all diversified funds; (3) all international funds; and (4) all funds (aggregate). since these portfolios are for comparative purposes and assume rebalancing each year, they are assessed transaction costs annually at the full value of the portfolio. this is necessary to avoid complex distribution decision processes that, in the end, would appear arbitrary and would cloud the results of the strategies above. for comparative purposes the s&p 500 index is also presented. data for the s&p 500 index comes from the s&p security price index record and the s&p stock market ency clopedia. the investor starts at year-end 1970 by purchasing the s&p 500 index, which is held for the entire sample period. taxes and transaction costs are applied to the index (on both capital gains and dividends). annualized yields on three month t-bills are used to estimate the risk-free rate. this information is obtained from business statistics, (1963 91), and the statistical abstract of the united states (1996). iv. m e t h o d o l o g y annual holding period returns are computed as follows: where rl po pl d1 (pi p0) + dl (1) r1 = p0 = annual return = price at end of previous year = price at end of current year = distributions made during the current year 88 f i n a n c i a l s e r v i c e s r e v i e w 7(2) 1998 two types of costs are considered in computing the holding period returns: transaction costs and taxes. each time a fund is bought or sold, transaction costs are assessed. for sim plicity, a flat fee is applied to the entire amount purchased or sold by the investor. when purchasing a fund, the investor pays an additional percentage b to the broker, and p0(1 + b) is used in place of p0. when selling a fund, the investor loses percentage b to the broker, and pi(1 b) is used in place o f p 1. two rates are tested: .5% for low transaction costs and 1% for high transaction costs. taxes are more problematic. the data spans a longer time period than previous studies of taxes and ceic funds, and individual tax rates vary considerably over the study period. in order to evaluate trading strategies rather than tax rate changes, this study uses a fiat tax rate of 40%. the flat rate allows consideration of the tax impact from trades without adding variability from changing tax rates. the 40% rate accommodates the wide range of tax rates on dividends and capital gains over the 1965 to 1995 time period (other tax rates are also tested, with no significant difference in the results). when a fund is sold or liquidated, capital gains (losses) are realized. for convenience, capital losses are assumed to offset other capital gains (so the loss results in tax savings). brokerage fees are included in assessing the taxable amount of the gain (loss). two differ ent methods are used to compute capital gains. the actual method uses the price change over the fund's entire holding period. since the tax is only paid when the fund is sold, there could be complications with using the actual gain in computing annual returns. annual returns examine the annual change in price, but the taxable amount is based on a gain earned over multiple years. to remedy this problem, the approximate method uses the fund's price change over the last year to compute the taxable gain. since both methods pro duce similar results, only the approximate method is reported. all distributions by the fund are adjusted for taxes. for informational purposes, before-tax returns with no transaction costs are also computed. the annual returns are used to calculate both geometric and arithmetic means. for sta tistical analysis requiring an estimate of the mean, the arithmetic mean is used. the sample standard deviation is calculated for each strategy to estimate overall risk. to assess market table 2 f-test for equal variances after-tax returns before-tax returns 1% transaction no transaction costs costs strategy computed f-value computed f-value most discounted 1.67 3.89* most relatively discounted 3.46* 5.32" 5 most discounted 1.38 2.25** 5 most relative discounted 1.60 2.13** specialized 0.79 2.12** diversified 0.52 1.39 international 1.88*** 5.07* aggregate 0.58 1.58 notes: *significant at the 1% level (critical f-value = 2.66). **significant at the 5% level (critical f-value = 1.98). ***significant at the 10% level (critical f-value = 1.70). t a b l e 3 su m m ar y st at is tic s fo r a ft er -t ax r et ur ns w ith 1 % t ra ns ac tio n c os ts r~ ,ig st ra te gy t ra di ng s tr at eg ie s: m os t d is co un te d m os t r el at iv el y d is co un te d 5 m os t d is co un te d 5 m os t r el at iv e d is co un te d po ol ed p or tf ol io s: sp ec ia li ze d d iv er si fi ed in te rn at io na l a gg re ga te c om pa ri so n: s& p 50 0 t -b il ls st an da rd a ve ra ge m in im um m ax im um d ev ia tio n b et a 13 .8 7% -1 1. 85 % 78 .8 8% 0. 19 6 0. 61 14 .5 4% -2 7. 47 % 77 .0 4% 0. 28 2 0. 79 12 .8 4% -2 0. 81 % 47 .7 4% 0. 17 8 0. 95 13 .2 3% -2 0. 69 % 70 .1 6% 0. 19 2 0. 82 7. 57 % 1 9. 29 % 39 .2 0% o . 1 34 0. 54 7. 09 % -1 3. 45 % 26 .0 8% 0. 10 9 0. 60 11 .7 2% -2 1. 89 % 58 .9 2% 0. 20 8 0. 82 8. 15 % -1 5. 50 % 32 .5 6% o .l l6 0. 63 10 .8 5% -2 6. 50 % 34 .0 0% 0. 15 2 1. o 0 2. 79 % 1. 21 % 5. 63 % 0. 01 1 c oe ffi ci en t of v ar ia tio n 1. 41 1. 94 1. 39 1. 45 1. 77 1. 54 1. 77 1. 42 1. 40 0. 38 sh ar pe in de x 1. 33 0. 98 1. 33 1. 28 0. 84 0. 93 1. 01 1. 09 1. 00 g eo m et ri c m ea n 12 .4 0% 11 .3 4% 11 .4 7% 11 .7 8% 6. 78 % 6. 57 % 9. 95 % 7. 56 % 9. 78 % 2. 78 % tw o sa m pl e tst at is tic 0. 60 93 0. 57 66 0. 42 45 0. 48 68 -0 .8 10 2 -1 .0 06 9 0. 16 94 -0 .7 07 2 ill o o t a b l e 4 s u m m ar y s ta ti st ic s fo r b ef o re -t ax r et u rn s w it h n o t ra n sa ct io n c o st s st ra te gy t ra di ng s tr at eg ie s: m os t d is co un te d m os t r el at iv el y d is co un te d 5 m os t d is co un te d 5 m os t r el at iv e d is co un te d po ol ed p or tf ol io s: sp ec ia li ze d d iv er si fi ed in te rn at io na l a gg re ga te c om pa ri so n: s & p 5 00 t -b il ls st an da rd a ve ra ge m in im um m ax im um d ev ia tio n b et a 21 .7 0% -1 8. 16 % 13 6. 04 % 0. 30 9 0. 92 22 .4 1% -2 4. 24 % 98 .6 7% 0. 36 2 0. 94 20 .9 7% -1 8. 80 % 67 .9 2% 0. 23 5 1. 14 21 .2 8% 1 8. 65 % 81 .8 9% 0. 22 9 0. 90 14 .8 1% -3 0. 83 % 68 .6 1% 0. 22 8 0. 88 13 .9 7% -2 0. 95 % 46 .2 5% 0. 18 5 0. 98 21 .8 5% -3 5. 31 % 10 2. 13 % 0. 35 3 1. 24 15 .7 8% -2 4. 42 % 57 .2 9% 0. 19 7 1. 01 13 .0 4% -2 4. 35 % 36 .4 1% 0. 15 7 1. 00 6. 97 % 3. 02 % 14 .0 8% 0. 02 7 c oe ff ic ie nt o f v ar ia tio n 1. 43 1. 61 1. 12 1. 08 1. 54 1. 33 1. 62 1. 25 1. 20 0. 38 sh ar pe i nd ex 1. 54 1. 38 1. 92 2. 02 1. 11 1. 22 1. 36 1. 44 1. 00 g eo m et ri c m ea n 18 .5 4% 17 .6 0% 18 .8 0% 19 .3 4% 12 .6 5% 12 .5 3% 17 .1 1% 14 .1 6% 11 .9 3% 6. 94 % t w o sa m pl e t st at is ti c 1. 24 9 1. 18 9 1. 40 4* 1. 48 6* 0. 31 9 0. 19 0 1. 14 0 0. 54 5 z > z < n ot e: si gm fi ca m at th e 10 % le ve l. < co closed.end investment companies 91 risk, each strategy's beta is computed. the coefficient of variation and the sharpe index allow investors to analyze return in conjunction with risk. investors want a small coeffi cient of variation since it indicates a smaller standard deviation for a given return. in con trast, they want a large sharpe index because it means a better yield with respect to the standard deviation. the sharpe index focuses on the risk premium, while the coefficient of variation considers the total return. for each fund, the sharpe index is divided by the s&p 500's sharpe index to produce a relative index value. it is important to determine if the proposed strategies outperform the market on a risk adjusted basis. the market model requires either a pre-event estimation period for deter mining beta or co-estimating beta and abnormal returns through the construction of portfo lios. data limitations prevent either of these techniques from being used in this paper. in addition, the market model may be inadequate for examining ceic funds. consequently, this paper compares the return and standard deviation of the proposed strategies to the s&p 500. a two-sample t-test is performed to determine if the returns for the proposed strategies are significantly different from the s&p 500. v. results before performing the two-sample t-test, an f-test is used to determine if the s&p 500 and the test strategy have equal variances. if the variances are unequal, it is necessary to adjust the test statistic to reflect this fact. the results are presented in table 2. because transaction costs of .5% and 1% produced virtually identical results, only the statistics for the higher transaction costs are reported in tables 2 and 3. most of the after-tax returns have equal variances to the s&p 500. however, the before-tax returns do not. when the two-sample t test for equal means are re-run assuming unequal variances, similar results occur. only the equal variance results are reported. table 3 contains summary statistics for the after-tax returns using transaction costs of 1%. the four trading strategies have higher arithmetic and geometric means than either the s&p 500 or the pooled portfolios. all four also have higher standard deviations than the s&p 500. since the tested strategies require annual rebalancing while the s&p buy and hold strategy does not, taxes and transaction costs are incurred more frequently. this could confound the after-tax measurement of beta. the coefficient of variation and sharpe index provide mixed results. three of the trading strategies have a higher sharpe index than the s&p 500. since a higher sharpe index is desirable, this supports the use of these trading strategies. with 1% transaction costs, the pooled portfolios have a higher coefficient of variation than the s&p 500. since investors want a lower coefficient of variation, this result does not support the use of pooled portfolios. the two-sample t-test determines if a trading strategy or pooled portfolio outperforms the s&p 500 on a risk-adjusted basis. none of the t-statistics are significant, even at the 10% level. this suggests that none of the simple trading strategies outperform the s&p 500. the summary statistics for the before-tax returns (no transaction costs) are presented in table 4. both the trading strategies and the pooled portfolios have higher arithmetic and geometric means than the s&p 500. in addition, the geometric means of the trading strate gies exceeds that of the pooled portfolios. all of the trading strategies have a higher mini mum and maximum than the s&p 500. the s&p 500 has a lower standard deviation than 92 financial services review 7(2) 1998 any of the tested strategies or pooled portfolios. betas for the trading strategies range from .90 to 1.14. most of the results for the coefficient of variation are mixed, although the pooled portfolios do have higher coefficients of variation than the s&p 500. this suggests pooled portfolios should not be used. both the trading strategies and the pooled portfolios have sharpe indexes in excess of 1.0. this suggests the use of neither of these investing methods. thus, the coefficient of variation and the sharpe index provide conflicting evi dence on the use of pooled portfolios. the two-sample t-test is significant at the 10% level for only two of the tested strategies. none of the t-statistics for the pooled portfolios are significant. although two of the trading strategies marginally outperform the s&p 500 on a risk-adjusted basis, taxes and transaction costs erode away these marginal benefits. vi. c o n c l u s i o n the most obvious differences between this study and others are the temporal distribution of the data and the consideration of taxes and transaction costs. the strategies tested in this paper are simple and relatively inexpensive. investors need to purchase at most five funds to implement the trading plan. short selling is not used, and rebalancing occurs annually. the purpose of the study is to determine if a simple trading strategy can yield superior per formance without substantially increasing risk. the results do not support the use of a simple mechanical strategy. even when margin ally significant before-tax returns are available, transaction costs and taxes erode the ben efits. excess returns are not possible for investors lacking the time or resources to actively trade in the marketplace. small investors following simple trading rules with a minimum of rebalancing are unlikely to earn the abnormal returns documented in earlier studies. this should serve as a warning to investors lured by the promise of excess returns from ceic funds selling at discounts. it is important for small investors to be aware of the need for additional monitoring, more frequent trading, larger initial investments, or short selling if they want to use ceic funds to outperform the market. investors wanting to avoid these complications should consider alternative investments. references anderson, s. c. (1986). closed-end funds versus market efficiency. journal of portfolio manage ment, fall, 63-65. barry, c. b., & jennings, r. h. (1993). the opening price performance of initial public offerings of common stock. financial management, 22(1), 54--63. barry, c. b., & peavy, j. w., iii. (1990). risk characteristics of closed-end stock fund ipos. journal of financial services research, 4( 1 ), 65-76. brauer, g. a. (1984). 'open-ending' closed-end funds. journal of financial economics, •3(4), 491 507. brauer, g. a. (1988). closed-end fund shares' abnormal returns and the information content of dis counts and premiums. journal of finance, 43(1), 113-127. brickley, j., & schallheim, j. s. (1985). lifting the lid on closed-end investment companies: a case of abnormal returns. journal of financial and quantitative analysis, 20(1 ), 107-118. closed-end investment companies 93 federal reserve board (1975, 1979, 1982, 1985, 1988, and 1991). interest rates: money and capital markets (statistical table). federal reserve bulletin. kim, c.-s. (1994). investor tax-trading opportunities and discounts on closed-end mutual funds. journal of financial research, 17(1), 65-75. malkiel, b. g. (1977). the valuation of closed-end investment company shares. journal of finance, june, 847--859. morris, m. h., & scanlon, k. (1996). a reexamination of the tax motivation for closed-end dis counts. journal of accounting, auditing, and finance, 11(2), 323-332. palomino, f. (1996). noise trading in small markets. journal of finance, 5•(4), 1537-1550. pesaran, m. h., & timmermann, a. (1994). forecasting stock returns: an examination of stock mar ket trading in the presence of transaction costs. journal of forecasting, •3(4), 335-367. pontiff, j. (1994). closed-end fund premia and returns: implications for financial market equilibria. journal of financial economics, 3 7(3), 341-370. pratt, e. j. (1966). myths associated with closed-end investment company discounts. financial ana lysts journal, july-august. richards, r. m., fraser, d. r., & groth, j. c. (1980). winning strategies for closed-end funds. jour nal of portfolio management, fall, 50-55. standard & poor's. (1995-1996). s&p nyse stock reports. standard & poor's. (1994). annual range, and close of daily indexes (statistical table, p. 1). s&p security price index record. standard & poor's. (1994). yield of the s&p daily stock price indexes (statistical table, p. 127). s&p security price index record. standard & poor's. (1995-1996). market measures (statistical table). s&p stock market encyclope dia, 17(1), 3. standard & poor's. (1996). market measures (statistical table). s&p stock market encyclopedia, 18(1), 3. thompson, r. (1978). the information content of discounts and premiums on closed-end funds. journal of financial economics, 6, 151-186. u.s. department of commerce. (1992). money and interest rates (statistical table). business statis tics 1963-91 (27th ed.). u.s. department of commerce. (1996). money market interest rates and mortgage rates: 1980 to 1995 (statistical table, p. 520). statistical abstract of the united states 1996, ( 116th ed.). u.s. news & world report. (1996). going for the gold, 121(2), 71. weisenberger, a. (1965-1990). statistical survey of closed-end investment companies (statistical table). investment company survey, part iv. weiss-hanley, k., lee, c. m. c., & seguin, p. j. (1996). the marketing of closed-end fund ipos: evidence from transactions data. journal of financial intermediation, 5(2), 127-159. pii: 1057-0810(92)90016-6 fl[nancial services review, 2( 1): 63-71 copyright 8 1993 by jai press inc. issn: 1~7-0810 all rights of reproduction in any form resewed abstracts of articles on individual financial management edited by phyllis schiller myers virginia commonwealth university charitable giving voluntary cont~butio~ to united charities, by marc bilodeau (dqartement d’ l?conomique, universite de sherbrooke, quebec, canada). if individuals are free to direct their gifts to any charity, why would they contribute instead to an institution like the united way that may not disburse their donations as they would have themselves? it is shown that ‘contributing only to the united fund’ can be a subgame perfect equilibrium, if the fund plays after everyone else and is able to offset direct con~ibutions. however, donations to a united fund may be lower than direct con~butions would have been, so that an optimal grants policy for the united fund would involve trading off a less desirable mix of services for higher total contributions. jourrrui of public fiommics, june 1992, 48( 1): 199-233. (reprinted with permission of north-holland publishing company). consumption-savings decision the decline in saving: evidence from household surveys, by b. bosworth, g. burless and j. sabelhaus. using the microeconomic survey data covering 1962-89, this paper examines the recent decline of private saving in three countries the united states, canada, and japan. there is no evidence that shifts in the demographic composition of the population between highand low-saving groups has contributed to the recent decline. the most interesting result is that saving has fallen within nearly all groups examined, although in the united states it fell most among those over forty-five. 64 financial services review, 2(l) 1992mw3 nor can capital gains explain the pattern of saving decline. saving fell among families with limited assets as well as among households enjoying large capital gains. brookings papers on economic activity, 1992, no. 1, pp. 183-24 1. (re printed with permission of the journal of economic literature). the consumption of stockholders and nonstockholders, by n.g. mankiw (harvard university) and s.p. zeldes (university of pennsylvania). only one-fourth of u.s. families own stock. this paper examines whether the consumption of stockholders differs from the consumption of nonstockholders and, if so, whether these differences help explain the empirical failures of the consump tion-based capm. household panel data are used to construct time series on the consumption of each group. the results indicate that the consumption of stockhold ers is more volatile and more highly correlated with the excess return on the stock market. these differences help explain the size of the equity premium, although they do not fully resolve the equity premium puzzle. journal of financial economics, march 1991, 17(l): 97-l 12. (reprinted with permission of north-holland publish ing company). saving and liquidity constraints, by a. deaton. this paper is concerned with the theory of saving when consumers are not permitted to borrow, and with the ability of such a theory to account for some of the stylized facts of saving behavior. the models presented in the paper seem to account for important aspects of reality that are not explained by traditional life-cycle models. econometrica, september 1991, 59(5): 1221-1248. (reprinted with per mission of the journal of economic literature). do individuals optimize in intertemporal consumption/savings decisions? a liberal method to encourage savings, by y .k. ng. rational consumption/saving choice requires rough knowledge regarding how much a dollar saved now could be compounded to become after various numbers of years. an indicative questionnaire reveals that most people grossly underestimate the significance of compound interest. when told of the right figures, they indicate their willingness to save much more. a simple model of intertemporal optimization under plausible parameters (4-5 percent real rate of return) prescribe many times higher consumption at older ages than at younger ages, implying very high rates of savings when young. journal of economic behavior and organization, january 1992, 17( 1): 101-l 14. (reprinted with permission of north holland publishing company). abstracts of articles on individual financial management how strong are bequest motives? evidence based on estimates of the den&l for life insurance and annuities, by b. douglas bem heim (princeton university). this paper presents new empirical evidence in support of the view that a significant fraction of total savings is motivated by the desire to leave bequests. specifically, i find that social security annuity benefits significantly raise life insurance holdings and depress private annuity holdings among elderly individuals. these patterns indicate that the typical household would choose to maintain a positive fraction of its resources in bequeathable forms, even if insurance markets were perfect. evidence on the relationship between insurance purchases and total resources reinforces this conclusion. journal of political economy, october 199 1, 99(5): 899-927. (reprinted with permission of the university of chicago press). insurance decisions and indmdual risk management income tax deductions for losses as insurance, by louis kaplow (harvard university). the federal income tax allows deductions for some categories of personal losses, notably for casualty losses (such as destruction of one’s home or car) and medical expenses above a threshold. the latter, even with lower marginal rates and further restrictions brought about by the 1986 tax reform, involves more than a $3 billion annual revenue loss (office of management and budget, 1990). deductions like these act as partial insurance: individuals receive a tax benefit equal to their marginal rate multiplied by the magnitude of their loss. however, this form of insurance is unnecessary when private insurance is available. moreover, as will be emphasized here, these deductions have a perverse effect because they am allowed only for the uninsured portion of losses. this induces individuals to be less protected against risk in the aggregate than if the implicit insurance provided by the tax system were unavailable; if the tax rate is sufficiently high, individuals would forgo insurance coverage altogether. it will be demonstrated that a tax system with no deductions for personal losses pareto dominates the current system. these conclusions are demonstrated in section i using a simple model in which risk-averse individuals may purchase actuarially fair insurance against loss and individual behavior does not affect the risk of loss (no moral hazard). section ii considers the applicability of the results when one allows for moral hazard, admin istrative costs, and other imperfections. it also discusses the applicability of tradi 66 financial services review, 2(l) 1992fl993 tional notions of tax equity. american economic review, september 1992,82(4): 1013-1017. (reprinted with permission of the american economic review). inexment selection and indmjmjal portfolio management stock prices and the dissemination of analysts’ recommendations, by m.d. beneish (duke university). this article investigates alternative explanations for the significant stock-price reaction to analysts’ information reported in the “heard on the street” column of the wall street journal. the observed market reaction persists after eliminating firms with confounding releases and firms for which analysts’ reports are issued immediately prior to publication. the evidence indicates that the column is not usually a secondary dissemination. first, stock prices adjust prior to publication when recommendations are reported on a single firm. second, analysts have incen tives to release information to the column before disseminating it to their clients. overall, the evidence suggests that the “heard on the street” column gathers information, forms a consensus, and provides it to investors. journal of business, 64(3): 393-416. (reprinted with permission of the university of chicago press). makroiikonomisches umfeld und opthnaler anlagemix. eine ana lyse filr die schweiz unter spezieller berilcksichtigung von immobi lienanlagen (with english summary), by w. burger and p. meier. in this paper, the stability of an optimal asset mix of bonds, shares, and real estate is analyzed using different time frames and indicators for real estate returns. diverging portfolio structures are discussed in light of different macroeconomic environments. an attempt is made to evaluate efficiency losses caused by asset allocation strategies based on nominal returns if-as is often argued-the investor’s prime concern lies with the preservation of purchasing power. schweizerische zeitschrift fir volkswirtschaft und statistiwswiss journal of economics and sta tistics, june 1991, 127(3): 51 l-523. (reprinted with permission of the joumal of economic literature). towards an equilibrium model of the mutual funds industry, by j. dermine, d.j. neven and j.f. thisse. the authors consider an industry in which mutual funds can form portfolios at lower cost than individual investors. investors can gather their own portfolio from abstracts of articles on individual fi~~ancial management 67 primary securities and/or shares of mutual funds. in this context, the authors model competition between mutual funds as a noncooperative game in which funds select their portfolios. they show that a small number of funds suffices to ensure a pareto superior equilibrium. journal of bunking finance, june 1991, 15(3): 485499. (reprinted with permission of the journal of economic literature). portfolio choice and risk, by j. encarnacion, jr. risk and risk aversion are interpreted in terms of a lexicographic model of portfolio choice where the telser criterion of expected value is maximized if the cramer-roy safety first criterion is satisfied, whence less risk aversion or more wealth implies a riskier portfolio and a higher return per dollar, and the asset demand for money-if held in the portfoliois less at a higher rate of interest. journal of economicbehaviorandorgunization, december 1991,16(3): 347-353. (reprinted with permission of the journal of economic literature). die kommunikationsfunktion der finanzmiirkte (with english sum mary), by m. hellwig. this paper compares different approaches to modeling communication in financial markets with asymmetric information. it appears that market outcomes do not depend on whether outsiders infer the insiders’ information directly from their behavior or indirectly from prices. the difference between selfselection models a la leland-pyle (1977) and market models a la grossman-stiglitz (1980) hinges on whether “insiders” behave as oligopolists or as price takers. price taking cannot be justified strategically unless communication is perturbed by “noise.” there is, thus, an inherent conflict between the requirements of information efficiency and alloca tion efficiency of the market. schweizerische zeitschrij? ftir volkswirtschujl und stutistiwswiss journal of economics and statistics, september 1991, 127(3): 35 l-364. (reprinted with permission of the journal of economic literature). tests of inflation and industry portfolio stock returns, by k. c. john wei (indiana university) and k. matthew wong (st. john’s university). this study examines the relationship between stock returns and inflation across 19 industry sectors during the preand post-world war ii periods. the results in the postwar period suggest that the proxy hypothesis can explain the spurious negative relationship between stock returns and expected inflation in all industries, but not the negative relationship between stock returns and unexpected inflation in non-natural-resource industries. we find little evidence supporting the nominal 68 financial services review, 2(l) 1992(3!493 contracting hypothesis. also, the expected inflation/returns sensitivity appears to be positively related to the level of real assets, but negatively related to debt ratio during the postwar period. journal of economics and business, february 1992, 44( 1): 77-94. (reprinted with permission of north-holland publishing company). ein interuationales kapitahnarktmodelt mit unterschiedlich infor mierten aniegeru (with english summary), by t. gehrig. international portfolio diversification seems rather limited. while there may be a long list of potential mutually compatible explanations, this paper provides another often neglected but important factor: investors* information structure. it is argued that, in general, the asymmetry of information renders foreign stock holdings relatively mote risky than domestic equity. accordingly, investors* portfolio exhibit a bias toward domestic stock. schweizerische zeitschriftflr vokswirtschaft und statistik/swiss journal of economics and statistics, september 1991, 127(3): 617-629. (reprinted with permission of the journal of economic literature). internationai asset pricing and equity market risk, by tc. chiang. this paper presents a model to examine the behavioral relationship between the excess returns of foreign exchange and the variables that measure risk factor. the test results of four major currencies support the hypothesis that the excess exchange returns are related to the relative risks of the two national equity markets. the evidence validates the existence of a risk premium in foreign exchange markets. journal of international money and finance, september 1992, lo(3): 349-364. (reprinted with permission of the .toumul of economic ~terature). evoiution in dynamic linkages across daily national stock indexes, by p.d. koch and t.w. koch. this article investigates how dynamic linkages among the daily rates of return of eight national stock indexes that evolved since 1972. a dynamic simultaneous equations model is estimated to describe the contemporaneous and lead/lag mlation ships across national equity markets over three different years: 1972, 1980, and 1987. results reveal growing market interdependence within the same geographical region over time. while there are many significant intermarket relationships within the same twenty-fog-hour period, there are few signi~cant lagged responses across markets beyond twenty-four hours. this suggests a high degree of inter national market efficiency. furthermore, japan’s market influence has been growing to rival that of the united states. journal of znternationaz money and abs&acts of artbks on lndividuai financial munagement 69 finaracq june 1991,10(2): 23 1-25 1, (repented with permission of the ~o~~a~ 0fecooumic ~terat~re~_ real esx+ate financing and investing prix et rendements immobiliers: construction d%uiices pour les an&s 1980 (with english summary), by c. zimmermarm. an index of real estate prices based on the obviation of the stock prices of estate mutual fimds is computed for the years 1980 to 1989. with i.9 percent a year, the prices have not risen as much as many thought. this index is compared to the price of land and to the costs of residential construction, allowing to confirm that the rent market is constrained. the average earning from real estate (6.5 percent) is slightly higher than the theoretical earning from capm (5.05 percent), which may augur future price decreases. ~chweizer~c~e ~itsc~r~~ fir vo~~irts~~~ und statistiwswiss journal of economics and statistics, december 199 1, 127(4): 73% 753. (reprinted with permission of the journal of economic literature). fiied versus variabk rate financing: the influence of borrower9 lender, and market characteristics, by lawrence g. goldberg and andrea j. heusen (university of miami). previous research has analyzed the problem faced by borrowers who must choose between fixed rate and variable rate loans when each loan carries different cost and riskcharacteristics and the borrowers face various income and employment prospects. in addition, the existing literature contains theoretical and empirical studies of how lenders react when given the ability to offer both fixed and variable rate financing. this article unifies the two strands of reseamh to develop and test a model of the equilib~um pro~~on of variable rate lending. results indicate that factors related to borrower, lender, and market characteristics are significant deter minants of the equilibrium proportion of variable rate credit originated. journal of financial services research, may 1992, 1: 49-60. (reprinted with permission of kluwer academic ~blishers). the refstionship between the demand and supply of home financ ing and neighborhood characteristics: an empirical study of mort gage racy, by george j. benston (emory university~ and dan horsky (universi~ of rochester). 70 financial services review, 2(l) 1992j1993 in this study the hypothesis that central cities are “redlined” was tested with data gathered from three separate central city and matching city-suburban areas. potentially unmet mortgage demand was estimated from interviews of people who tried unsuccessfully to sell their homes. a hedonic price function revealed that these owners did not “overprice” their homes. however, they did not attribute their inability to sell to potential buyers’ problems with financial institutions. a logit analysis of data provided by home buyers showed that individuals bought in the areas they preferred. we also found that central city and city-suburban home buyers got the type of mortgage they wanted at market terms and home owners in both areas had no di~cul~ in fm~c~g home repairs. joumal offs sentices rtseurch r;ebntary 1992,5(3): 235-260. (reprinted with permission of kluwer academic publishers). a general equilibrium model of housing, taxes, aud portfolio choice, by james berkovec (federal reserve board) and don fullerton (university of virginia and carnegie mellon university). we describe a model in which rental and owner housing are risky assets, tenure choice is endogenous, and each household is constrained to consume the same amount of owner housing that it has in its investment portfolio. at each iteration in the search for an equilibrium, we determine the new taxable income for each of 3,578 households (from the survey of consumer finances), and we use statutory schedules to find the marginal rate and tax paid. equilibrium net rates of return are major dete~~ants of the amount of owner housing, but a logit model indicates that demographic factors are the main determinants of ownership rates. in our simula tion, taxes on owner housing would raise welfare not only by reallocating capital but also by the government’s taking part of the risk from individual properties and diversifying it away. measures to disallow property tax or mortgage interest deduc tions do not help share this risk. simulations of the 1986 tax reform indicate a small shift from rental to owner housing and welfare gains from reallocating risk. journuf o~~~zitic~1 ~co~~y, april 1992,100(2): 390429. (repined with permission of the university of chicago press). statistical methodology generalized bankruptcy models applied to predicting consumer credit behavior, by darral g. clarke and james b. mcdonald (b~gham young university). although it is well known that many financial variables do not behave according to the assumptions of commonly used statistical methodologies, it has abstracts of articles on individual financial management 71 been generally assumed that the robustness of these methodologies assures ade quately accurate results. recent work in the econometrics literature has explored the use of estimation methodologies that are appropriate for variables that do not have symmetric or normal distributions. this article explores the application of the egb2 family of distributions to the problem of predicting bankruptcy and offering consumer credit. the distributions of the forecasts of good and bankrupt accounts violate many of the assumptions of the multiple discriminant analysis methodology used in most commercial bankruptcy forecasting models. the importance of the criterion used to classify good and bankrupt accounts is also considered. since there is a positive payoff to the correct classification of good accounts, minimizing the number or expected costs of misclassifications does not maximize the expected economic return. the theoretical development is demonstrated on a data base of more than 16,000 observations of demographic and credit variables of bankrupt and solvent households. the performance of the various classification models is tested on a holdout sample of more than 11,000 observations. although the generalized distri bution is found to be statistically superior to discriminant analysis, the managerial importance of this superiority is seen to vary with the characteristics of the bank ruptcy forecasting problem. joumul of economics and business, february 1992, 44( 1): 47-62. (reprinted with permission of north holland publishing company). pii: 1057-0810(92)90009-2 authors index volume 2,1992/1993 brooks, robert, see radcliffe, robert butler, kirt c. and dale l. domain, “long-run returns on stock and bond portfolios: implications for re tirement planning,” 2( 1): 4149 cook, wade d., see hebner, kevin j. domain, dale l., see butler, kirt c. hebner, kevin j., and wade d. cook, “a multicriteria approach to mutual fund selection,” 2( 1): l-20 knight, john r., and lewis mandell, “nobody gains from dollar cost averaging: analytical, numerical, and empirical results,” 2( 1): 5 141 lang, larry r., and robert m. niendorf, “performance and risk exposure of international mutual funds,” 2(2): 97-l 10 levy, hiam, see radcliffe, robert mandell, lewis, see knight, john r. mcleod, robert w.. sharon moody, and aaron phillips, ‘the risks of pension plans;’ 2(2): 131-156 moody, sharon, see mcleod, robert w. moroney, john r., see peevey, robert m. niendorf, robert m., see lang, larry r. peevey, robert m., g. c. uselton, and john r. moroney, ‘the individual in vestor in the market: forming a belief regarding market efficiency,” 2(2): 87-96 persson, don, see woerheide, walt phillips, aaron, see mcleod, robert w. prestopino, chris, “what strategies are estate planners recommending? evidence from survey data,” 2(2): 111-130 radcliffe, robert, haim levy, and robert brooks, “active timing decisions of equity mutual funds,” 2(l): 21-39 uselton, g. c., see peevey, robert m. woerheide, walt, and don persson, “an index of portfolio diversification,” 2(2): 73-85 articletitles “active timing decisions of equity mu tual funds,” robert radcliffe, haim levy, and robert brooks, 2( 1): 21-39 171 172 financial services review, 2(2) 1993 “index of portfolio diversification, an,” walt woerheide and don persson, 2(2): 73-85 “individual investor in the market: form ing a belief regarding market efti ciency, the” robert m. peevey, g. c. uselton, and john r. moroney, 2(2): 87-96 “long-run returns on stock and bond portfolios: implications for retire ment planning,” kirt c. butler and dale l. domain, 2( 1): 41-49 “multicriteria approach to mutual fund selection, a” kevin j. hebner, and wade d. cook, 2( 1): l-20 “nobody gains from dollar cost averag ing: analytical, numerical, and em pirical results,” john r. knight and lewis mandell, 2( 1): 51-61 “performance and risk exposure of inter national mutual funds,” larry r. lang and robert m. niendorf, 2(2): 97-110 “risks of pension plans, the,” robert w. mcleod, sharon moody, and aaron phillips, 2(2): 131-156 “what strategies are estate planners rec ommending?-evidence from sur vey data,” chris prestopino, 2(2): 111-130 the perfect withdrawal amount over the historical record e. dante suareza,* adepartment of finance and decision sciences, trinity university school of business, one trinity place, san antonio, tx 78212, usa abstract what has been the perfect withdrawal amount (pwa) from retirement savings accounts in longterm historical data? the pwa is that which, if taken out in the first year of retirement and used again every year adjusted by inflation, leaves exactly the desired final balance on the account. we present the formula for obtaining this measure and evaluate the values it has taken in the past under varying combinations of the relevant parameters. we find that safety-minded investors should enter retirement with a higher stock allocation than what is currently used in most investment funds designed to provide income during retirement. © 2020 academy of financial services. all rights reserved. jel classification: j26; d14; g11 keywords: retirement income; perfect withdrawal amount; sequencing risk; sustainable withdrawal rate; distribution strategies 1. introduction a considerable body of research has explored the question of what is the safe level of withdrawals from a savings account that must provide income during retirement. a frequent objective of these studies is to examine the historical record to determine the withdrawal figure that has “rarely” depleted the funds before the end of the retirement period. however, if the withdrawal amount chosen does not exhaust the funds in the account, it means that it left a positive balance at the end. if the retiree had not planned to leave a bequest or posthumous gift—or if the final balance is larger than what she intended to leave behind—this is money *corresponding author. tel.: +1-210-999-7860; fax: +1-210-999-8134. e-mail address: esuarez@trinity.edu 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 96–132 that she would rather have consumed herself. in these cases, it can be said that the amount withdrawn was not perfect; it should have been higher. here we use results presented in suarez, suarez, and walz (ssw; 2015) that show that for any given series of return rates on a portfolio, there is one and only one constant withdrawal amount that will take the starting balance to the desired ending balance in a given span of time. if this time span is the number of years for which the retiree wishes to plan, then this is the perfect withdrawal amount (pwa) for her: it’s constant (no ups-and-downs in the income stream) and it attains the final goal exactly and just in time (no portfolio “failure,” or inheritance shortfall, or overaccumulation at the end). if she withdraws more than her pwa, she will run out of funds—or at least leave behind less money than she intended. if she withdraws less, she will leave money on the table. it is important to stress that the pwa is a measure inherent in any sequence of return rates. a sequence of returns has a pwa just as it has a mean, a standard deviation, or any other conventional statistical measure, and by calculating these values we are not endorsing any particular viewpoint on retirement withdrawals. rather, we intend for this paper to be of the same nature as ibbotson and sinquefield’s seminal 1976 article (ibbotson, r. g., & sinquefield, r. a., 1976), where long-run historical rates of return were first presented. we argue that the retirement withdrawal question can be understood as the search for what the pwa will turn out to be (eventually) for each retiree’s portfolio and objectives. and a first step in this direction should be a revision of the values that this measure has taken, historically, for different values of the relevant parameters. in this sense, the results presented here can significantly assist the decision-making process in retirement. we also refer to another result presented in ssw, namely that if the retiree is interested in using up her retirement savings in stages—for example, withdrawing a certain amount for a number of periods and then a different, smaller amount as social security benefits start to come in—the “perfect” set of figures is a function of the “regular” pwa that is reported here, which is a constant amount throughout the retirement period (bridges & choudury, 2007). the numbers presented here can thus provide guidelines for many different circumstances and cases. we present results for the pwa over the period 1926-2014. using data from ibbotson sbbi classic yearbook (morningstar, inc., 2015), we start by computing pwas for a basecase scenario of all-stock portfolios and 30-year long retirement periods, where the investor’s goal is to exhaust the account’s funds completely. the retirement periods are rolled forward one month at a time, with the first one beginning in january 1926 and ending in december 1955, and the last one running from january 1985 through december 2014. this one-month-shift approach produces a series of 708 figures that, although not all independent, provide a bird’s-eye view of how this particular pwa has evolved over time. we also use a separation result obtained in ssw to look at how the pwa is affected by the sequential order in the portfolio’s returns. this is known as sequencing risk, and we examine it by computing the values for the sequencing factor—a measure developed in that paper for this type of risk—over the same period. we then compute the values corresponding to other asset allocations for the portfolio. by increasing the bond allocation in 10 percentage-point increments—and decreasing the corresponding allocation to stocks—the relation between asset allocation and historical pwas is explored. e.d. suarez / financial services review 28 (2020) 96–132 97 next, we look at how our results change when the retiree’s goal is not to exhaust her savings entirely, but to leave a portion of her balance as bequest or final gift. this is perhaps a more relevant case, especially since the protected balance can also be considered an emergency reserve for unforeseen expenditures and, in this sense, it might be a desirable feature in most retirement plans. this interpretation of a nonzero final balance goal as a safety device suggests an important question that, to our knowledge, has not been discussed before: if the retiree wants to “play it safe,” should she withdraw the amount that rarely exhausts the balance in her account, or the amount that frequently produces a final balance that can cover some unexpected outlays? are these approaches equivalent? the results in this section provide insight for this discussion. we close with the pwa results for other horizon lengths. by changing the duration of the retirement period from 30 years to 25, 35, and 40 years we obtain a ballpark estimate of the level of risk incurred by using a specific endpoint in retirement plans, when in reality this closing date is of course unknown. taken together, the numbers presented in this paper draw the outline of a surface that can be used for decision-making. the true perfect withdrawal amount awaiting in each retiree’s future cannot be known in advance, but by perusing the cross-sections of this object we can make well-educated guesses of what it may end up being in each case. the aim of this paper is to recruit the assistance of financial history to make this guesswork more robust. 2. previous research the research program on optimal retirement withdrawals began in 1994, with the release of two seminal papers. the first one, bierwirth (1994), actually computes what we are now calling pwas, although he obtains them by trial-and-error. bierwirth does not seek the withdrawal amount that exhausts the starting balance, but rather the figure that leaves the account balance unchanged in nominal terms after 25years (even though the withdrawal amount is kept constant in real terms). however, the stated goal of bierwirth’s work—and its lasting contribution—is to use historical data instead of simplistic assumptions to evaluate any withdrawal strategy. this key insight is taken up by bengen (1994) to develop the concept of “portfolio longevity,” which he uses to arrive at the conclusion that “a first-year withdrawal of 4%, followed by inflation-adjusted withdrawals in subsequent years, should be safe.” this “4% rule” was later reinforced by a series of papers by cooley et al. (1998, 1999, 2003, 2011). these studies used a more exhaustive scope, and varying conditions and assumptions, to confirm that this strategy provides a high probability (85%–95%) that the retiree will not run out of funds ahead of time. the unifying feature of these “first-generation” approaches is that the retiree is asked to select a specific withdrawal amount at the beginning of a very long period, and then she is supposed to take out that same amount, in either nominal or real terms, each and every year. this motivated attempts to develop “adaptive” rules, where the withdrawal amount is modified in some particular way depending on the circumstances faced in every period. in this direction, the field has veritably exploded in recent years (mitchell & blanchett, 2011). currently there are, for example, rules that modify the withdrawal amount when the rate of return is negative (guyton & klinger, 2006), or when it deviates too much from an 98 e.d. suarez / financial services review 28 (2020) 96–132 average value (frank et al., 2011), or when the withdrawal amount/rate becomes higher than a predefined number (pye, 2000; zolt, 2013). there are rules that change the amount every five years to meet withdrawal rate targets (spitzer, 2008), and rules that change it every year depending on the length of the remaining horizon (blanchett & frank, 2009). some approaches incorporate mortality tables to the analysis, including the investor’s age in the calculation of the withdrawal amount (stout & mitchell, 2006), or computing the expected present value of all future withdrawals (mitchell, 2011; stout, 2008). furthermore, there are even some approaches that forgo the stability criterion altogether, letting the withdrawal amount bear the brunt of the variability in the rates of return. waring and siegel (2015), for example, propose a rule in which “the investor will receive 30 years of payments of varying size, each appropriate to the then-current value of the portfolio and the interest rate . . . the only risk is to the size of each payment, which will vary . . . with investment results.” although these procedures may have some merits, we think that most retirees prefer a stable flow of income year after year. in a sense, most of these methods are just different ways of “hunting” for one single, elusive number: the constant withdrawal amount that can be used year after year, over a certain period in the future, without causing neither early depletion nor overaccumulation. this paper provides the historical series for that number. 3. the perfect withdrawal amount the perfect withdrawal amount formula can be thought of as a generalization of the payment formula for a fixed-rate loan. in this case, however, the rate is not fixed but changes in every period, so the simplification allowed by geometric series does not occur. also, here we are talking about payments from an investment account, instead of payments into a loan that must be amortized. finally, the generalization includes a final balance term that must be non-negative, but not necessarily zero. as shown in ssw, this constant value is given by: w ¼ ks qn i¼1 1þ rið þ � ke � � =on i¼1 qn j¼1 1þ rjð þ (1) where w is the yearly withdrawal amount, ks is the starting balance, ri is the rate of return in year i, n is the planning horizon in years, and ke is the ending balance sought at the end of these n years. this equation shows that a period of time with a large cumulative return is not necessarily associated with a high pwa, because the sequential order by which the total return comes about is also relevant. this effect is seen more clearly by rewriting eq. (1) as: w ¼ rnks � keð þ sn (2) where rn ¼ qn i¼1 1þ rjð þ is the return generated by the assets in the portfolio over the entire period, and [sn ¼ on i¼1 qn j¼i 1þ rjð þ]�1 can be interpreted as an “adjustment” applied to this total return to take into account the specific sequence of period-by-period return rates that produced it. the expression sn is what we call sequencing factor. e.d. suarez / financial services review 28 (2020) 96–132 99 ordering matters, of course, because for an investment account that is constantly drawn down—as is the case in the retirement stage—getting high rates at the beginning and low rates later on is better than the other way around, because if the high rates come early they will impact a larger balance. a closer inspection of the sequencing factor formula shows how this feature is captured by our measure: sn ¼ ½on i¼1 qn i¼1 1þ rjð þ��1 ¼ ½ 1þ r1ð þ 1þ r2ð þ 1þ r3ð þ . . . 1þ rnð þþ 1þ r2ð þ 1þ r3ð þ . . . 1þ rnð þþ 1þ r3ð þ 1þ r4ð þ . . . 1þ rnð þþ . . .þ 1þ rn�1ð þ 1þ rnð þþ 1þ rnð þ��1 (3) we note that the rates that come later in the sequence (higher index number) appear more times in the summation than the rates that come sooner. because sn is the reciprocal of the summation, this means that if we take a set of return rates (unordered) and arrange it into a sequence (ordered), we can increase the value of sn by swapping any pair of rates so that the larger of the two figures comes before the lower one1. therefore, as it should, the value of sn is higher when the sequence is “better.” the next step in the pwa approach is just to realize that eq. (1) defines a stochastic variable. that is, since the rate of return is a random variable, the pwa measure is a random variable as well. in that sense, the results presented here can be considered as snapshots of a stochastic process that could have given us other values as observations. the purpose of this paper is to examine those values—the ones that actually came up and became observable— to derive insights for the management of retirement accounts. 4. the data for all the calculations in this paper we use data from ibbotson sbbi classic yearbook, 2015 edition. specifically, we use the series “large-capitalization stocks: total returns” as the monthly return on stocks, and “intermediate-term government bonds: total returns” as the monthly return on bonds. all the rates of return are deflated using cpi-u to obtain real-term yields. because we use real-term returns throughout, the withdrawal amounts that we obtain are constant in terms of the purchasing power they provide (the initial withdrawal amount is adjusted for inflation in every period) and the desired ending balance is attained in real terms as well. to provide perspective for the results of this study, it is instructive to review these return figures briefly. the geometric average of the real monthly rate of return for stocks over the entire period spanning from january 1926 through december 2014 (1,068months) was 0.56% (6.98% annualized), with a minimum of �29.26% (in september 1931) and a maximum of 42.56% (in april 1933). the arithmetic mean was 0.71%, and the standard deviation was 5.47 percentage points. in the case of bonds, the geometric average of the real monthly return over the 89-year period was 0.18% (2.24% annualized). the minimum value was �7.71% (february 1980), and the maximum was 10.74% (april 1980). the mean for this series was 0.19%, and the standard deviation was 1.39 percentage points. 100 e.d. suarez / financial services review 28 (2020) 96–132 we now ask what the real rate of return has been, for these two financial instruments, over 30-year periods in the historical record. this 30-year horizon is what we use below as our “standard scenario,” which in turn corresponds to the set-up that is most frequently used in the literature on optimal withdrawals. fig. 1 shows these 30-year rates, expressed as annual rates, and we notice a major upward trend in the long-term real returns for bonds. however, although no more recent data can be included because 30-year periods starting after december 1984 have not yet ended (at least not in our dataset), using 20or 15-year periods we see that this series will probably end up hovering around 3% annual, or even lower, when the next decade of observations becomes available. in any case, these figures are well known; we show them here just as a reference for the analysis of the next sections. 5. results for the standard scenario the scenario we use as our “base case” is an investor that enters her retirement period with a portfolio fully invested in stocks, and whose aim is to deplete her balance entirely over a period of 30 years. in other words, our investor profile in this section is a person of approximately 65 years of age who wants her money to last until she’s 95 and has no heirs, or otherwise has no desire to leave behind any inheritance or gift. when there is no bequest sought, ke in eq. (2) is zero and the expression simplifies to: w ¼ rnsnks (4) fig. 1. real rates of return for stocks and bonds, over 30-year periods beginning on the indicated date. shown are yearly geometric averages over the 30-year periods. e.d. suarez / financial services review 28 (2020) 96–132 101 before moving on to the actual results, we make one last modification to express our perfect withdrawal amount as a perfect withdrawal rate (pwr), relative to the starting balance in the account: w=ks ¼ rnsn (5) fig. 2 presents the pwr for the standard scenario as a monthly variable, assigning to each month the pwr value corresponding to the 30-year period beginning on that month. this computation assumes that withdrawals are made only once per year, on the very first day and adjusted by the previous year’s inflation, whereas the returns are credited to the account on the last day of the year (or, equivalently, at the end of every month). as mentioned before, under this convention for dating the data the pwr cannot be computed beyond december 1984 because the 30-year periods beginning after that date are not yet complete in our data set. the first thing one notices in fig. 2 is the wide range of variation, and how these up-anddowns do not follow closely the changes in the rate of return for stocks (see fig. 1). to see how this can come about, and to further pin down the pwa concept, fig. 5 shows the explicit sequence of debits and credits on the account in two retirement periods, beginning fig. 2. perfect withdrawal rate for 100%-stocks portfolio over 30-year period with zero final balance. for the indicated retirement start dates, these withdrawal rates (applied to the starting balance and updated by inflation every year) would have exhausted the available funds exactly after 30-years. 102 e.d. suarez / financial services review 28 (2020) 96–132 in june 1949 and february 1969, respectively, chosen to illustrate some of the points we make below. another salient feature of fig. 2 is that the 4% rule would seem overly conservative, as the chart rarely pierces the 4% level. however, in fact, the “90% safe” withdrawal rate is only slightly higher, at 4.27%. that is, out of the 708 observations in this series, 70 were lower than 4.27%. fig. 4 shows the frequencies by pwr range. still, the fact that the pwr has been higher than 4% so frequently—and higher by so much, with more than half of the periods registering over 7%—raises the question of what would have happened, in the past, to the final balance of an investor who followed the 4% rule. under the pwa framework this is easy to answer, as we can rewrite eq. (2) to solve for the final balance2: ke ¼ rnks � w=sn (6) fig. 3. all-stocks portfolio with $1 m starting balance, to be depleted over 30 years, begins retirement at two different dates. an all-stocks portfolio of $1 million that entered retirement in june 1949 would have been exhausted after 30 annual withdrawals of $139,929 (plus inflation), but this same portfolio would have provided only $39,612 (plus inflation) per year over 30 years if it had entered retirement in february 1969—even though the average annual real return over both 30-year periods was practically the same (7.2%). e.d. suarez / financial services review 28 (2020) 96–132 103 if we express the final balance as a fraction of the starting balance and define the withdrawal rate r=w/ks, we get: ke=ks ¼ rn � r=sn (7) fig. 5 presents the results of eq. (7), over 30-year periods, for an all-stocks portfolio using r = 4%—the 4% rule. this chart brings forward a point that was made in ssw, namely, that choosing a withdrawal amount that provides a high safety level is tantamount to accepting a high probability of ending up with an account balance much larger than what was intended. we find that in 81% of the periods, following the 4% rule would have left the investor with more money—in real terms—than what she had at the beginning (instead of exhausting the balance). in almost two-thirds of the historical instances the final balance would have been more than twice the starting balance. an in fully one-third of the cases the account would have closed with five times or more the original money! since overaccumulation means that the retiree’s standard of living was unnecessarily restricted, the fact that it has happened so frequently under the 4% rule (and that it is so large) is in itself a significant drawback for this common recommendation. additionally, when we consider that retirement savings accounts are rarely the only source of income for retirees3—so that even if withdrawals exhaust the balance, there is usually another source of funds—a likely culprit for this excess balance comes to the fore: using 90% as the chosen safety level is probably an inadequate requirement for an adaptive plan. fig. 4. frequency distribution of pwr in standard scenario. in the standard scenario the portfolio is invested entirely in stocks and is completely depleted (no bequest) after 30 years. the first decile cutoff is at 4.27%, so this is the “90%-safe” withdrawal rate in this case. 104 e.d. suarez / financial services review 28 (2020) 96–132 returning to fig. 2, we see basically three “times” when the pwr dipped below the 4% level. first, there are the periods starting in december 1928 through november 1929, or in march 1930 through june 1930. anybody who started retirement in these specific months, with stocks as their savings vehicle, would have met the brunt of the great depression at a very early stage of their retirement period, which is when the retiree’s financial situation is most vulnerable. in the first period mentioned, for example, they suffered the huge crashes of october and november 1929 (-19.7% and �12.5%, respectively, in real terms) during their very first year of retirement. everybody in this group bore the losses of june and september 1930 (-15.8%, �13.3%) and also took, only one year later, the largest one-month drop in the history of the stock market (-29.3%, september 1931). furthermore, a few months after that, stocks fell an additional 43.2% in the three-month period of march through may 1932. all this would seem obvious if not for the fact that the pwr turned out to be much higher (in the 6%–8% range) for people who retired only a few months earlier. indeed, for people who retired in july 1927 (only 17months before the first of our ill-fated groups) the pwr was 6.2%. if the retirement period had started in august 1931—taking the “mother of all crashes” head on, and then going through the ordeal of the spring of ‘32—the pwr would still have been 5.9%. and if the retirement period had started only a little after this turmoil, in july 1932, the retiree would have obtained the highest pwr ever recorded—an eye-popping 15.3%—even though the awful returns of 1937 would have hit her while still in the first stage of the retirement period. fig. 5. final balance for 100%-stocks portfolio under 4% rule after 30 years, relative to starting balance. the 4% rule would have led to large (possibly unwanted) final balances in the historical record. e.d. suarez / financial services review 28 (2020) 96–132 105 the examination of this first segment in the pwr series throws into sharp relief the complex relationship between rates of return and optimal withdrawal rates. although this volatile behavior could be attributed to the general market turbulence of the time, the sensitivity of the results to what might seem like minor disturbances is still remarkable. in particular, we see how market rebounds sometimes manage to “save” the portfolio, but they have to happen at a specific time to produce a meaningful impact—sequencing is crucial. the sequencing effect is crucial because it determines the perfect withdrawal amount. we stress the point that a set of return rates that might be considered “good”—in the sense that upon compounding it produces a high total return for the entire period—will nonetheless lead to a low pwa if the particular ordering of the set is unfavorable (from low returns to high returns, in general). in particular, a “bad” ordering will be associated with a relatively low pwa, and this cannot be changed by adjusting the withdrawal amount from year to year as it would only zig-zag around, hover about or glide towards an inevitably low value.4 next in our analysis of very low pwr periods is the six-month stretch beginning in october 1965. this instance of “bad times to start retirement” is puzzling because not much happened market-wise in that period. rather, it would seem that it was just a few instances of negative monthly returns (march, may, august 1966) that, though none terribly bad, combined with a generally lackluster market in the first few years to produce a very deleterious effect. the most recent period with a very low pwr is the one spanning from august 1967 to june 1969. in this 23-month period, 13months had a pwr of less than 4%, with the remaining 10months barely managing to surpass this benchmark. as in the previous case, this instance is hard to understand except as a case of peculiar timing. it would seem that a series of early punches to the investment account (october 1967, jan-feb 1968) combined with the bear market of 1969-1970 to create the kind of “perfect storm” that hurts new retirees very much. it would seem appropriate to also discuss periods when the pwr was particularly high, but by now it is clear that the relationship between rates of return and pwrs is too complex to allow any definite conclusions. for example, the lowest 30-year real rate of return on record for the all-stocks portfolio corresponds to the period starting in october 1955 (4.1% annualized), but the pwr for the cohort that began retirement on that month is 6.2%—not particularly low. the so-called kennedy slide of 1962 produced very low returns at the beginning of these retirement periods, yet the pwr ended up in the 4.5%–5.5% range—higher than at other, less remarkable times. although the results for 30-year periods starting just before the crash of 1987 are not yet available, fig. 2 shows that retiring at the end of 1984— only three years before that big hit—nonetheless provided an extremely high pwr (10.7%). the complicated relationship between rates of return and pwrs is illustrated in fig. 6. panel (a) superimposes the rate of return for stocks shown in fig. 1 with the pwr presented in fig. 2. panel (b) shows the corresponding scatter plot, where a linear regression between these two variables attains a coefficient of determination of just 14%. as can be seen, it is hard to establish a clear-cut connection between these two variables over the historical record. we notice, for example, a long stretch of time spanning from late 1946 until mid-1954 when the pwr was much higher than what the return rates at the time would seem to warrant. and the opposite holds true from mid-1967 to mid-1972: high return 106 e.d. suarez / financial services review 28 (2020) 96–132 fig. 6. real rate of return (annualized) and standard-scenario pwr. panel (a) compares the all-stocks average rate of return over 30-year periods with the corresponding pwr over the historical record. panel (b) plots the same-date values of these two variables against each other. the x’s in panel (b) mark jun ‘49 and feb ‘69, the two dates used to construct fig. 3. e.d. suarez / financial services review 28 (2020) 96–132 107 rates, but low pwrs. as shown in fig. 3, the annualized real rate of return for the 30-year period beginning in june 1949 was 7.2% and the pwr was 14.0%. for the period beginning in february 1969 the rate of return was also 7.2%, but the pwr was 4.0%—10 percentage points lower!5 the period starting in august 1952 obtained an average rate of return of just 4.6%, yet managed to attain a pwr of 10.0%. the period starting in january 1933 got a much higher rate of return, at 10.1%, but ended up with the same pwr of 10.0%! it is now that the explanatory power of eq. (5) becomes truly useful, as all these puzzling cases become just specific instances of one single relationship. this expression tells us concisely and unambiguously that the gist lies in the sequencing part of the equation, and it also tells us how to explore this further. therefore, even if a specific sequence of return rates will, in all likelihood, never be repeated, the problem of forecasting the pwr in any particular case is no longer the problem of forecasting the whole series of rates. rather, as the derivations presented here show, what the investor needs to do is to obtain an estimate of the average annual return for the portfolio she holds (for which there is abundant literature that can help her), and then address the problem of the sequencing factor value that she must use to adjust it. the next section focuses on how can the historical evidence shed light on this latter problem. 6. the sequencing factor over the historical record the sequencing factor can be interpreted as an “adjustment” that has to be applied to the rate of return to express the effect of the specific sequence of results that produced that total return. it can be “good” or “bad”; in general, a series of returns that goes from high to low is better than one that goes from low to high. the periods discussed at the end of the previous section show that the magnitude of the sequencing impact can be downright stunning. eq. (3) shows the explicit formula for the sequencing factor which, as discussed in ssw, is not a proxy measure. it comes directly from an analytic dissection of the problem and it is a measure of orientation, in the same sense that variance is a measure of dispersion—and the orientation of return rates (going up, going down, up a little then down a lot, etc.) is the crucial element that the adjustment factor should capture. fig. 7 displays the values that this formula has taken over the historical record in the standard scenario, with the corresponding frequency distribution shown in fig. 8. the first salient feature in fig. 7 is that the sequencing factor is low (unfavorable ordering) and relatively stable for a very long time at the beginning of the chart. for the first 173months (almost 14 and a half years!) the sequencing factor stays inside the lowest two brackets of the frequency distribution in fig. 8, not surpassing the 0.0075 mark until june 1940. the reason for this is, of course, the general instability that affected equity markets during most of the 1930s—the great depression. this “shaky ground” situation meant that if an investment portfolio entered the withdrawal stage at any time during this period, it didn’t take long for it to reach one or more months when the stock market crashed violently. for the 30-year periods beginning during 1926, 1927, or 1928, the crash of october 1929 came along very quickly, hurting the sequencing factor irreparably even though the average rate of return ended up in the 6%–8% 108 e.d. suarez / financial services review 28 (2020) 96–132 range. the periods beginning shortly after black tuesday did not fare much better, sequence-wise, since “just around the corner” they would meet terrible results in june 1930, september 1931, and the march-april-may stretch of 1932. later periods in this era would be similarly crippled by major downturns in september 1937, march 1938, and may 1940. fig. 7. sequencing factor for 100%-stocks portfolio over 30-year periods. the sequencing factor is a measure of the orientation of an ordered set of numbers. if it is large it means that the set goes from high to low, instead of the other way around—which is good, if the numbers in the set are rates of return. fig. 8. frequency distribution for sequencing factor in standard scenario. the sequencing factor is a measure of orientation in return rates (“going up” vs. “going down”). in general, a low value of the sequencing factor means that the sequence of returns went mostly downwards, which in turn is correspondingly worse for the retiree. e.d. suarez / financial services review 28 (2020) 96–132 109 therefore, we could say that in these years the sequencing factor was low because the return rates could not sustain a normal level for long, instead crashing constantly. the cumulative rate of return over 30 years was very acceptable, but the initial years were bad and the damage was done. next in fig. 7 we come across a stretch of rising values culminating in august 1952, when the sequencing factor reached its maximum value of 0.0261. indeed, the sequencing factor was over 0.02 for 17months—between september 1951 and january 1953—which produced pwrs around 10% even though the average return rate was quite low. and none of these 17months produced spectacularly high returns (the maximum was 5.7% in november 1952); rather, it was a period when returns followed more or less an increasing pattern: from slightly negative in its first two months to moderately strong at the end of 1952. the return sequences then managed to stay clear of major crashes for a number of years, but otherwise these moderately favorable starts were all it took for the sequencing factor to attain very high values. it would seem that a favorable return sequence is not so much one that starts with a bang, but one that doesn’t get its legs cut off too quickly. finally, we comment on a period of nearly two years—from january 1969 to november 1970—when the sequencing factor again dropped below 0.005. this case would seem to be different from the great depression period discussed above, because now the low values of the sequencing factor were not caused by “crashes” in the early stages of the 30-year periods. rather, this was a somewhat peculiar time that registered negative (and positive, but small) returns in the first months, and then rose to strong returns in the second half of 1970. this unfavorable ordering caused the pwr to hover in the 4%–6% range, even though the average return was between 7% and 9% for these periods. 7. results for other portfolio asset mixes we now examine the pwr results for portfolio mixes that include bonds as well as stocks. we retain the parameters from the standard scenario regarding horizon length (30 years) and bequest goal (zero), but now allow bonds to take up a constant, positive fraction of the asset mix. we identify each asset mix by its percentage of stocks/bonds—as in 90/10 for 90% stocks and 10% bonds—and consider different mixes by varying the proportions in 10-point steps. thus, we examine the 90/10 portfolio, the 80/20, the 70/30, and so on, all the way to the 0/100 (100% bonds) portfolio. to maintain the same asset mix in the portfolio at all times, we assume that a rebalancing of asset classes takes place every month. that is, after the account obtains the return for each asset class at the end of every month, the investor buys and sells each instrument as needed to restore the corresponding proportions to the asset mix. we clarify this assumption because our set-up would also be consistent with an annual rebalancing scheme (withdrawals are made only once a year), and the two rebalancing methods are not equivalent. we chose monthly rebalancing because our retirement horizons are being rolled over month-by-month, and this procedure facilitates comparisons across periods that overlap. it seems to us that both methods produce very similar results, but in any case, the choice is arbitrary. 110 e.d. suarez / financial services review 28 (2020) 96–132 we begin by adding the minimal amount of bonds to the mix (10%), and present the pwr results for this 90/10 portfolio in fig. 9 the results for the all-stocks portfolio are also shown for comparison. fig. 10 presents the differential between the two series, as well as the variation in the actual withdrawal amount resulting from the change of mix. variation charts similar to fig. 10 could be produced for all the alternate model specifications presented in this paper; we present only this one instance, to convey the idea of the type of analysis that can be conducted. these charts showcase the enhanced clarity provided by the pwr approach. an examination of figs. 9 and 10 may lead us to conclude that the all-stocks portfolio is the best alternative, as the instances when the 90/10 mix provides better results are less frequent and less significant. by changing the all-stocks mix to 90/10, the pwr improves in 173 of the periods considered (24% of the 708 total), but in 104 of these instances the increase is less than onetenth of a point—presumably negligible. the 69 periods where the pwr increased by more than 0.1 points represent just 10% of the total, and are all in the great depression era. also, these increases never surpass 0.42 points, whereas the pwr decreases resulting from the change reach as high as 1.43 points. however, the periods that were benefited the most by the change in the mix also had a low pwr in the standard scenario, so the impact on the perfect withdrawal amount is much more significant. the pwa variation is a more appropriate measure because it presents the situation as the retiree perceives it: her retirement income becoming x% higher or lower because of the asset mix change. this is the dashed line in fig. 10 which, for example, at its highest point informs us that if a person entering retirement in september 1929 had switched fig. 9. pwr for all-stocks portfolio versus 90/10 portfolio. when the asset mix is changed from all-stocks to 90% stocks and 10% bonds, the pwr changes accordingly. e.d. suarez / financial services review 28 (2020) 96–132 111 10% of her retirement funds to bonds instead of stocks, her annual income over the next 30 years would have been 12.5% higher. this is a significant impact but, overall, the chart shows unfavorable pwa variations, so the case against including bonds seems to stand—at least in the marginal proportion examined so far. what happens if we allot a larger proportion to bonds? for the sake of clarity in the charts, the next results are presented in sets of three portfolio mixes: 100/0 versus 70/30 versus 50/50, and 100/0 versus 30/70 versus 0/100. the summary results for the entire set of 11 mixes (from 100/0 to 0/100, in 10-point increments) are given later in this section. now our previous conclusion that bonds should not be included in the portfolio becomes less clear-cut, but the arguments in favor and against are easily grasped from our charts. if we want to make the withdrawal profile more stable, the 30/70 mix seems appealing (see fig. 12). if what the investor wants is to minimize the probability of a catastrophic outcome, 50/50 looks like the way to go (fig. 11). the grid shown in fig. 13 is particularly useful for this kind of decisions. from the fig. we see that the 50/50 mix is the one that has breached the 4% pwr level the least number of times, so this could be considered the ideal portfolio for a safety-minded investor. indeed, regarding the 4% rule it is interesting to note that in none of the 708 historical periods covered here has it been the case that the pwr was over 4% for the 100/0 portfolio, but less than this benchmark for the 50/50 portfolio. however, the opposite has indeed happened, as in 32 historical periods the 100/0 pwr was less than 4% but the 50/50 pwr was more than 4%. therefore, an investor who is satisfied with a 4% withdrawal rate should fig. 10. change in pwr and variation in pwa when switching from all-stocks to 90/10 mix. pwr differential (how many points is the withdrawal rate affected) is in percentage points, solid line, left axis. pwa variation (how large is the modification in the withdrawal amount) is in percentage, dashed line, right axis. 112 e.d. suarez / financial services review 28 (2020) 96–132 never hold the all-stocks portfolio: 50/50 is better. by the same token, the idea that an investor who cares a lot about safety should hold more bonds, seems unwise. in particular, if the funds’ owner is overly concerned that her withdrawal amount might crash below some subsistence level, she should avoid asset mixes such as 20/80, 10/90, or 0/100.6 however, another interpretation of “safe” is that the retiree should pick a withdrawal amount that has a high probability of being sustainable. under this version of safety, the task of selecting a portfolio mix then consists of finding the proportions that maximize, say, the 90%–safe rate. we have included these values at the end of fig. 13, where we find that the 70/30 mix is now ideal because the upper bound of the first decile in this pwr distribution is almost 4.5%. indeed, we can say that by selecting an appropriate asset mix, the 4% rule’s withdrawal levels can be improved by over 10% without incurring additional risk. using the upper bound of the first decile as the watermark for separating safe from risky is, of course, an arbitrary criterion. especially if, as discussed in ssw, the procedure for selecting the withdrawal amount calls for periodic reassessments—or if, as argued above, the retiree has other, reliable sources of income—the retiree may be willing to take a less conservative approach. after all, under a periodic revision scheme the retiree’s main concern is not so much running out of funds, but having to decrease the withdrawal amount significantly in the future. be it for those reasons, or simply because the retiree has a higher tolerance for risk, it is useful to review the figures corresponding to other “safety levels” in the historical record. fig. 14 presents these milestone values by asset mix. fig. 11. pwr for 100/0 (all-stocks) portfolio versus 70/30 and 50/50 portfolios. e.d. suarez / financial services review 28 (2020) 96–132 113 8. results for bequest goals other than zero in this section we restore the asset mix from the standard scenario (100% stocks) to look at the behavior of the pwa when the “perfection” of the measure does not mean that it will fig. 12. pwr for 100/0 (all-stocks) portfolio versus 30/70 and 0/100 (all-bonds) portfolios. fig. 13. frequency distribution for historical pwr, by portfolio mix, for 30-year periods and zero bequest goal. over the historical record, the pwr was lower than 4% in the least number of periods using the 50/50 mix. but the 90%-safe rate was highest when the 70/30 mix was used. 114 e.d. suarez / financial services review 28 (2020) 96–132 exhaust the account balance in a given number of periods but, rather, that it will make that balance reach a certain figure by the end of the retirement period. as in the previous sections, we assume a horizon length of 30 years and all figures are in real terms, so the bequest goal is attained in the sense that this is the purchasing power that the ending balance provides. this investigation is important not only because many prospective retirees are actually interested in bequeathing or giving away some of their funds, but also because of the possibility of introducing positive-valued bequest goals as safety devices in retirement plans. that is, if the investor is concerned about getting low returns, or that perhaps a large unforeseen expenditure may arise, she could protect against these contingencies by planning as if she intended to leave an inheritance. this maneuver provides the funds’ owner with three adjustment levers to react to new developments: the withdrawal amount, the safety level, and the bequest goal. thus, for example, in a scenario of poor portfolio performance the retiree can either reduce her withdrawals, accept a lower safety level, or reduce her bequest target/protection balance—or combine these adjustments to spread the impact. before presenting our results, we discuss a change in the role that the sequencing factor plays when the bequest goal is nonzero. the zero final balance specification used in the fig. 14. cutoff points of the first five deciles in historical pwr distributions, by asset mix. if the retiree wants to be 90%-safe she should use the 70/30 mix, but for lower safety levels (higher risk tolerance) she should add more stocks. e.d. suarez / financial services review 28 (2020) 96–132 115 standard scenario produces the very succinct expression shown in eq. (5), but the corresponding equation for the general case is: w=ks ¼ rnsn � sn ke=ksð þ (8) this complete version reveals a peculiar feature of the relationship between the pwr and the sequencing factor. although a favorable ordering in the sequence of return rates will provide a higher pwr, it will also make the result more sensitive to any change in the bequest goal (ke/ks). as we did with the asset mix, we change the bequest goal parameter in 10-point increments from zero to 100%, going from the case where the account’s funds are exhausted completely (no bequest) to the case where the balance on the account is unchanged in real terms at the end of the planning horizon (100% bequest). the intermediate points correspond to bequeathing 10% of the starting balance, 20%, 30%, and so forth, always in constant purchasing power terms. we start by comparing, in fig. 15, the pwr for the standard scenario against the results for 10% and 30% bequest goals. we will argue that the main thing to see in fig. 15 is, ironically, that there is not much to see in it. that is, the three variables shown are so close to each other that they “fudge” the chart, making it difficult to distinguish one line from the others. what this means is that the reduction in the pwr necessary to protect or salvage a non-negligible fraction of the fig. 15. pwr for standard scenario (no bequest) versus 10% bequest goal and 30% bequest goal. the adjustment in the withdrawal rate needed to accommodate significant bequest intentions is surprisingly small (see also fig. 16). 116 e.d. suarez / financial services review 28 (2020) 96–132 investment balance is surprisingly small. one would think, for example, that if a person who enters retirement with $1 million in her account wants to leave a $300,000 estate, she would have to reduce her withdrawal amount very significantly with respect to what she could take out if she didn’t want to leave any funds at all. however, the evidence in the historical record indicates that the pwr differential in this case (all-stocks portfolio over 30-year periods) has never been larger than 0.8% points, it has been as low as 0.1 points, and its median value was 0.3 points—that’s all she would have to shave off. again, it may be more relevant to discuss the required variation in the pwa, that is, how much of her annual income would the retiree have to sacrifice to reach a given bequest goal. our results for the 30% bequest goal say that the most she would have had to cut back was 8.9%, and that in some periods the necessary decrease in her withdrawal amount was just 0.1%. the median pwa variation was �3.7%. however, it would be wrong to conclude that by committing to reductions as small as these the retiree would immediately have 30% of the balance available for emergencies. the figures presented here leave the desired remainder in the account, yes, but at the end of the planning horizon, and they depend crucially on the length of this timespan. emergencies and unforeseen expenditures can arise at any time, and if they come up in the first few years of retirement we would be “cashing our bequest early” to deal with them. this would render the results in this section inapplicable by, in effect, shortening the retirement horizon. still, by following these guidelines an investor will come progressively closer to having a freedisposition fund available—if she manages to navigate a number of years in retirement without major mishaps. the next chart presents the results for bequest goals of 50% and 100% of the starting balance. the results for all the bequest-goal levels considered are presented at the end of this section. with these two bequest-goal levels we may be moving into unrealistic territory, as perhaps very few people will want to leave behind 50% of their starting balance (let alone 100%) when they could have used it for consumption during retirement. but 50% is psychologically appealing (the “half for me, half for the kids” approach), whereas there is something very realistic about the 100% bequest goal: endowment funds. an institutional endowment fund is, essentially, a retirement investment account that must last forever. the “infinitely-lived retiree” (the institution) is interested in withdrawal strategies that provide a relatively stable flow of funds, but leave the principal value unchanged in the long run, so that the withdrawing process can go on indefinitely. this is similar to leaving a bequest of 100% after a number of years, so we comment on the results for this specification at some length below. as with the smaller bequest goal levels shown in fig. 15, in fig. 16 we find it remarkable that the lines in the chart are not further apart, even though we are now dealing with very ambitious bequest intentions. to leave behind a full 50% of the starting balance as departure gift, there have been periods when the retiree only had to decrease her withdrawal rate by 0.2 percentage points. for the period beginning in june 1935, for example, the pwr for zero bequest was 8.0% and the pwr for 50% bequest was 7.8%. of course, in some other periods the half-and-half approach has been a much more costly endeavor. for the period beginning in august 1952 (a peculiar period that we have already singled out) the pwr with no e.d. suarez / financial services review 28 (2020) 96–132 117 bequest was 10.0%, while for the 50% bequest goal the pwr was 8.7%—a 1.3-point difference. the median reduction in the pwr, when going from zero bequest to 50% bequest, was just 0.4 points. the variation in the pwa, when moving from a zero-bequest goal to a 50% bequest goal, was also rather small. the maximum decrease in the withdrawal amount was 14.9%, and the minimum was 1.8%. the median variation was �6.2%, which means that, by reducing the amount withdrawn by only that much, the retiree would have ended up with a final balance equal to 50% or more of the starting balance in one half of the historical periods covered here—instead of zero dollars at the end. as for the 100% bequest level, the necessary decrease in the pwr to achieve this goal ranged from 0.4 to 2.6 points with respect to the standard scenario, with a median of �0.9 points. the pwa variation, in turn, was as large as �29.8% in some periods, and as small as �3.5% in others. not surprisingly, these extreme values correspond to retirement periods for which the average rate of return was also close to the endpoints of its own range. the smallest required decrease in the withdrawal amount—to leave the real balance unchanged after 30 years instead of exhausting it—occurs in the period beginning in june 1932 (11.8% average real return); the largest, in october 1955 (4.1% average real return, lowest ever, see fig. 1). the median pwa variation was �12.4%. it could even be argued that the inclusion of a high bequest goal should be the hallmark of a safe retirement plan. the 4% rule, for example, is considered a safe strategy because fig. 16. pwr for standard scenario (no bequest) versus 50% bequest and 100% bequest. 118 e.d. suarez / financial services review 28 (2020) 96–132 it lowers the probability of “failing,” but it defines failure as running out of funds altogether! even in the presence of other sources of income, would it not be more prudent to plan so that the worst-case scenario—or at least the most unlikely ones—still provides means for sustenance, or some amount of supplementary funds? a promising idea would be to develop retirement plans that, at the outset, use a very high bequest goal—perhaps 100%. this feature would provide the plan with an “embedded safety” that would support the selection of an otherwise risky withdrawal rate—a 50%–safe level could even be used. then, if future developments indicate that our case is in the unfavorable range, the bequest goal can be adjusted gradually until a safe situation is reestablished. it should be noted that if we select the median value from the correct pwr distribution, in one half of the cases the withdrawal amount would be revised up, not down.7 with respect to endowment funds, the manager’s problem is to determine the maximum amount that can be withdrawn consistently each year without jeopardizing the long-term viability of the fund. this is equivalent, to some extent, to the problem of sustaining a long retirement period as comfortably as possible and yet bequeath a large portion (close to 100%) of the starting balance in the account.8 for example, an all-stocks fund manager that would like to protect the entirety of the fund’s assets, but who considers acceptable to end up with a 10% decrease in the balance after a 30-year period, may use the values shown in fig. 17 as a reasonable range for the withdrawal rate. from fig. 17 we see that a withdrawal rate between 3.5% and 4.1% would seem to be appropriate in this case. if our manager takes out more than 4.1%, the “probability” that the fund’s balance after 30 years will be lower than 90% of the starting balance would be in excess of 20%—too much risk, perhaps. on the other hand, if she were to withdraw fig. 17. milestone safety levels in historical pwr distributions with 100% bequest goal and 90% bequest goal. possible range of withdrawal rates for balance protection. portfolios are all-stocks and horizon length is 30 years. e.d. suarez / financial services review 28 (2020) 96–132 119 less than 3.5%, the likelihood that the balance in the distant future will be larger than the current balance would be more than 90%, which seems overly conservative. because the actual pwr is likely to be higher than these values—at least under our interpretation of past frequency as probability—perhaps the ideal procedure would be to “hunt down” the correct withdrawal rate. instead of taking out the same amount (in real terms) every year, the fund administrator would use this range of rates to see if they translate into withdrawal amounts with the same purchasing power as what was taken out in the previous year. in case of significant discrepancy, any adjustment would be pondered against the benefits of a stable income flow. it is interesting to note that a more simplified approach to the endowment fund problem would identify the average real return as the 100%–bequest pwr. however, these two concepts are not the same and, in general, do not coincide. they would be strictly equal only if the rate of return was the same in every period, but otherwise the sequencing effect makes them differ. fig. 18 superimposes these two variables, and we see that the difference has been very substantial. fig. 19 displays the relative frequency of the perfect withdrawal rates according to the range of bequest goals. finally, our discussion about how it is not so difficult to bequeath a large portion of the retirement funds is just the flip side of what we mentioned in a previous section: if the retiree undershoots her pwr (takes out less than the perfect amount) even by just a fraction of a point, she can end up with huge unwanted balances at the end. these realizations can lead us to conclude that the whole business of determining an optimal withdrawal amount is too fig. 18. comparison of average real return versus perfect withdrawal rate for all-stocks portfolio with 100% bequest and 30-year horizon. although it might be argued that the applicable withdrawal rate for endowment funds is simply the average real return on the investment account, in the historical record the correct rate has turned out to be a very different figure. 120 e.d. suarez / financial services review 28 (2020) 96–132 fragile, subject to some kind of “butterfly effect.” however, for us they only underscore the need to develop algorithms, instead of rules of thumb, to navigate this stage of life with some degree of stability. we think that should be the research program for this field and, although we do not present any proposals here, ssw can provide sketches of what a pwabased algorithm would look like. we close this section with the pwr frequency distribution under different bequest goals, computed using a 100%–stocks portfolio and with a planning horizon of 30 years. 9. results for other planning horizon lengths the last item in our research list is to see what happens to the pwr when the length of the planning horizon changes. as in the previous two sections, we reset the values of the other two parameters in eq. (1) to what they are in the standard scenario. thus, in this section the rates of return ri correspond to an all-stocks portfolio and the ending balance ke is set at zero, but we use different values for n. exploring the impact of horizon length changes is important for at least two reasons. first, the horizon length used when putting together a retirement plan will be either an educated guess of the person’s remaining life span (based on age, life expectancy data, health outlook, family history, etc.) or an intentional “overshoot” figure to accommodate the possibility of an unusually long life. if she is using the realistic figure, she wants to know what happens if she misses; in particular, she would like to know how large would the fig. 19. frequency distribution for historical pwr, by bequest goal level, for all-stocks portfolio and 30-year horizon. for any given period in the historical record, the reduction in the withdrawal rate required to accommodate a positive final balance is equal to the sequencing factor for that period times the bequest goal sought. this means that periods that registered a high sequencing factor (good for the retiree) are also more sensitive to the introduction of a bequest goal. e.d. suarez / financial services review 28 (2020) 96–132 121 adjustments be if she ends up living longer than she thought. if she is “padding up” the horizon length, she wants to know the price she is paying for doing this, that is, how much is the withdrawal amount lowered by this arbitrary extension of the planning horizon. the second reason is that the retiree will probably want to change this figure along the way. perhaps a new medical breakthrough will benefit her, increasing her life expectancy. or if she adopts healthier habits—exercise, better nutrition, quitting smoking—what happens then? how bad would it be if she suddenly needed to make her funds last a longer number of years? and we must also be ready for the opposite case, when health developments are such that a reduction in the planning horizon seems to be in order. indeed, the number of years remaining in the horizon is a different sort of uncertainty that needs to be addressed, perhaps stochastically. in this paper we are treating it as a purely subjective choice, but even a change in the perception of the numbers of years still ahead for the retiree (or a change in the current age of the retiree under consideration) is a drastic modification of the estimation problem. even though the return rates going forward remain of course the same, a change in the planning horizon is a change in our choice of cutoff point for that future sequence. we start by extending the planning horizon to 35 and 40 years. fig. 20 presents the values that the pwr has taken, over the historical record, for these alternate specifications. once again, as in the case of bequest goals, we find that the decrease in the pwr produced by the extension of the planning horizon is remarkably small. for the 40-year specification, the difference in the pwr with respect to the standard scenario is at most 1.5 fig. 20. pwr for standard scenario (30 years) versus 35-year long and 40-year long horizons. all cases computed using 100%-stocks and zero bequest (total exhaustion). the adjustments required to accommodate longer lifespans are smaller than might be expected. 122 e.d. suarez / financial services review 28 (2020) 96–132 percentage points, and sometimes it is as low as 0.1 points. the median differential is �0.4 points. if we use the pwa variation as measure, the results remain striking. in the historical record, the cutback to the withdrawal amount required to make the funds last 40 years instead of 30 has never been larger than 11.4%. as for the minimum, an investor who entered retirement in april 1932 with $1 million could have taken out $94,143 for 30 years, or $92,634 for 40 years—a variation of just 3.5%! the median value for the pwa variation is �6.3%, which means that if a retiree decreases her withdrawal from the outset by just 6.3%, she will have a 50% chance of being able to cover 10 more years.9 we believe that a majority of investors and retirees would think that an increase of one-third in the number of periods that must be covered with the funds available would have a much larger impact on the amount of money they can withdraw. it is certainly possible that some 35or 40-year periods had very high returns in the last five or 10 years, so the horizon extension managed to carry its own weight, so to speak, and did not affect the withdrawal amount very much. but this would be a rare occurrence, and fig. 20 shows that a small decrease in the pwr is the norm, not the exception. rather, this feature seems to be embedded in the intrinsic dynamics of the drawdown process. this is an important result, as it seems that prospective retirees perhaps do not have to worry so much about underestimating their life expectancy: there will be plenty of time to react and, in all likelihood, the reaction will not have to be severe. it is important to stress, however, that for this result to hold the horizon extension needs to be acknowledged and applied in the early stages of the retirement period. thus, for example, if at the very beginning of retirement, the investor is unsure whether to use a 30-year long or a 35-year long planning length, it is useful to know that the respective pwa distributions are not very different. however, this is not the same as traversing, say, one half of a 30-year retirement horizon and then decide at that point that we would like to cover five more years. in that case we would need to craft a 20-year plan—what we need to cover from that point on—to substitute for a 15-year plan—the remaining horizon according to the original plan. this is a different situation, although it can be assessed in a similar fashion. to provide an idea of what these “midterm” adjustments might entail, we present summary results for 20-, 15-, 10-, and 5-year long horizons later in this section. although the literature in this field has settled on 30 years as the typical horizon length for planning purposes, there are reasons to consider it excessively long. the full retirement age to receive social security benefits is currently 66 years and, according to data compiled by the social security administration (ssa), an american man reaching age 65 today can expect to live, on average, until age 84.3. an american woman turning 65 today can expect to live, on average, until age 86.6. only about one out of every four 65-year-olds today will live past age 90, and one out of 10 will live past age 95.10 this means that the standard scenario has a large additional safety margin, as this horizon length will actually be required by just 10% of all retirees. our position here is that it is advisable to use a withdrawal amount with a good chance of being sustainable, but this must be taken from a pwa distribution with a realistic horizon length. otherwise, the sustainability of the chosen amount would be incorrectly estimated, and the true safety level will likely be much higher than what was intended. under this e.d. suarez / financial services review 28 (2020) 96–132 123 view, the ssa’s life expectancy data indicate that 25 years would be more appropriate than 30 years for first-approximation calculations. fig. 21 presents the pwr results under this horizon length,11 keeping in mind that about 25% of all retirees—particularly women—will need to apply corrections along the way to cover additional years. we see that the pwa for 30-year periods has been between 2.9% and 13.1% lower than for the 25-year periods beginning on the same month, with a median of �5.3%. this means that an intervention in the early years of a 25-year plan, aiming to turn it into a 30-year plan instead, should call for reductions of the withdrawal amount in the 3% to 10% range to realistically expect to cover the additional longevity. the relatively small size of these adjustments supports our notion that it is a good idea to build retirement plans using 25-year horizons, and then operate on these plans using prudent withdrawal strategies. the 90%-safe rate in 25-year horizons with no bequest is maximal in the 70/30 mix, reaching 4.80%. the 50%-safe rate in 25-year horizons with 100% bequest— the version of conservative approach that we advocate in this paper—is highest in the allstocks allocation, at 6.6%. per life expectancy data, this timespan should be appropriate for most retirees and, if this is not the case, then the conservative withdrawal program will probably provide sufficient leverage to cover the extra years. furthermore, even if the perfect storm materializes in the form of a longer-than-usual life span and a terrible sequence of returns, the option of curtailing the withdrawal amount does not look catastrophic because the required reduction is probably not so large. as for the timing of the actual adjustments to be made, we advocate in favor of the procedure proposed in ssw. as time moves forward—and either the planning horizon shortens or the retiree decides to choose an endpoint farther into the future—the withdrawal amount fig. 21. pwr for standard scenario (30 years) versus 25-year planning horizon. the 25-year planning horizon is more realistic for people retiring at the age of 66. 124 e.d. suarez / financial services review 28 (2020) 96–132 under consideration must be assessed in terms of the percentage it represents of the account balance at that point (i.e., that is, the implied withdrawal rate). the withdrawal amount is to be modified when the location of this withdrawal rate in the corresponding pwr distribution falls outside the confidence range (probability interval) with which the retiree feels more comfortable: if the confidence level becomes too high, it is advisable to increase the fig. 22. frequency distribution for historical pwr, by horizon length, for 100%-stocks portfolio and zero bequest goal. panel (a) shows bins in integer percentage points. panel (b) shows cutoff points for first five deciles. fig. 23. pwr for standard scenario (30 years) versus 20-, 15-, 10-, and 5-year planning horizon. these shorter timeframes may be used for midcourse adjustments, once inside the retirement stage. e.d. suarez / financial services review 28 (2020) 96–132 125 withdrawal amount; if it gets too low, the amount withdrawn should be reduced. the magnitude of the adjustments, in turn, may conceivably be based on the results presented in this paper. thus, for example, if the investment account is all-stocks, the retiree is trying to use all her funds for consumption, and the appropriate planning horizon length for the current year is deemed to be, say, 15 years, then the third column on the right side of fig. 24 is applicable. there we find that she should take out at least 6.1% of her balance, lest she takes on too much risk of running up an excess balance, and perhaps no more than 10.5% of the balance—unless she is not troubled with being more than 50% likely of having to reduce her withdrawal amount in the future. this is what the evidence in the historical record tells us. fig. 22 summarizes the main features of the historical pwr distributions for the four different horizon lengths that we discussed in detail in this section. fig. 23 presents the corresponding values for shorter horizons, and fig. 24 shows the historical evolution of the pwr for these shorter timeframes.12 as for how does the sequencing factor change when the length of the planning horizon is modified, an algebraic analysis indicates that it should vary inversely with the horizon length parameter. first, in general, a longer horizon should lead to lower withdrawal rates simply because the starting balance would need to cover a larger number of periods (a longer life span). second, the total compounded return generated by the investment portfolio over the entire period should become larger as the planning horizon lengthens —again, in general—because there will now be more periods over which the return rates will compound to provide the total return. finally, from the decomposition of the pwr fig. 24. same as fig. 22, but with shorter planning horizons. 126 e.d. suarez / financial services review 28 (2020) 96–132 provided by eq. (5) (w/ks = rnsn) we see that a lower withdrawal rate (left-hand side of the equation) may be reconciled with a larger total return (rn, in the right-hand side of the equation), only if the corresponding sequencing factor decreases, because sn is the sole remaining term in the right-hand side of the equation. fig. 25. sequencing factor for 100%-stocks portfolio over different horizon lengths. in general, longer timeframes should be associated with lower values for the sequencing factor but, although true in general in the historical record, we see that this has not always been the case. fig. 26. historical 90%-safe pwr for zero bequest goal, as asset mix and horizon length vary. this is a particular 3-day section of the historical pwr hypercube; similar sections can be produced for other safety levels or bequest goals. also, any of these two control parameters may be swapped for one of the three variables in the axes of this chart, so as to explore different situations (see, for comparison, the work of frank et al. (2011)). e.d. suarez / financial services review 28 (2020) 96–132 127 fig. 25 shows the actual values of the sequencing factor, over the historical record, for the four horizon lengths that we examined more closely in this section. it is interesting to note that, the arguments of the previous paragraph notwithstanding, there are several periods in the historical record where the sequencing factor for shorter horizons was lower than the factors for longer horizons beginning on the same date. in general, however, the arguments presented above are confirmed by our results. it would also be useful to understand the pwr interplay between horizon lengths and asset mixes in the historical record, as this would provide guidelines for the construction of glide path strategies and for the management of funds that are to provide income during retirement. of course, all the algorithms and rules currently used by serious practices have a strong statistical foundation, but it would still be instructive to know what these programs look like when placed under the lens of historical experience. we offer the results shown in fig. 26 as a point of entry for this discussion. this chart shows that for the 40, 35, and 30-year horizon lengths, the highest 90%safe rate is attained with the 70/30 mix, whereas for the 25-year horizon this rate is slightly higher with the 60/40 mix. from the point of view of mutual funds designed to provide income during retirement, the 25-year and 30-year timespans can be considered reasonable or even short, and even so the “very safe” withdrawal rate is highest at equity allocations that are larger than what is normally used in these funds.13 also, a strategy that switches the portfolio mix away from stocks when a certain accumulation goal is attained (basu et al., 2011) may also be called into question, because presumably the retiree’s goal is not to reach the payout phase with a large balance in her account, but to make a series of large withdrawals from it. keeping a high stocks allocation can help her achieve this, even after the phase transition point is passed. 10. conclusions the examination of historical data using the pwa approach yields useful results for the management of retirement investment accounts. the historical record would recommend the use of relatively stock-heavy portfolios, with about 70% of total funds allocated to equity. even with horizons of just 25 years—that would be well inside the retirement stage according to the assumptions of most models—the maximization of the 90%-safe withdrawal rate calls for stock allocations of at least 60%. for retirement plans that intend to exhaust the available funds after 30 years—the most frequent setup in optimal withdrawal studies—portfolios containing 70% stocks and 30% bonds produced a pwr higher than 4.5% in 90% of the historical periods covered here, and in none of the 708 periods examined did it drop below 3.9%. in our opinion, this calls for a reexamination of the 4% rule because “safe” withdrawal amounts can be increased by more than 10% through a judicious selection of asset allocation. the 4% rule would seem to be more appropriate for managing endowment funds, instead. we argue, however, that safe withdrawal strategies should include a bequest goal— even if the retiree has no such intentions—for at least three reasons. the first is that 128 e.d. suarez / financial services review 28 (2020) 96–132 retirement plans should strive to avoid undesirable outcomes, but running entirely out of funds is too drastic as worst-case scenario. the second reason is that the retiree needs to have a recourse in case that unexpected large outlays arise for any reason. and the third is that the presence of a bequest goal in the plan creates an additional adjustment lever for midcourse corrections, should these become necessary in case of poor portfolio performance. under this interpretation, safety would be provided by the bequest goal embedded in the plan, not by the location of the withdrawal amount in a histogram. for an all-stocks portfolio with a 100% bequest goal and a 30-year planning horizon, the pwr historical series produced a median of 6.4%. this could then be used as a safe withdrawal guideline, with the bequest goal as an adjustment lever that will have to be lowered in about 50% of all cases. if an adjustment via the bequest goal becomes insufficient and the withdrawal amount has to be reduced, our results suggest that the impact on the pwa for the remaining horizon length will not be very large, if the correction is applied in the first five to 10 years. indeed, the sensitivity of pwa distributions to changes in parameter values—be it the horizon length or the bequest goal level—is surprisingly small. upon changing the planning horizon from 30 to 35 years, the median decrease in the zero-bequest pwr for all-stocks portfolios was just 0.3 percentage points. if the horizon is extended to 40 years, the median pwr decrease was only slightly higher at 0.4 points. this would indicate that the adjustments demanded by longer lifespans are not very large, if detected and implemented in the early stages of retirement. in turn, if the horizon length is kept at 30 years but the bequest goal is increased to 50%—one-half of the starting balance’s purchasing power remaining in the account at the end—the median pwr decrease for all-stocks portfolios was 0.4 points. compared with the zero-bequest case, a reduction of just 6.2% in the amount withdrawn would be sufficient to achieve (or surpass) the 50% bequest goal in one-half of the historical periods examined. researchers are already applying the pwa methodology with some profit, with clare et al. (2016a) using it to implement a trend-following overlay that improves the 90%-safe rate by 2.2 percentage points in 20-year horizons. in another paper, these authors employ the pwa concept to discuss ways to address the problem of sequence of returns in decumulation portfolios (clare et al., 2016b). in turn, estrada (2018) takes the methodology to an international context and computes pwas for 21 different countries. the pwa framework was conceived as a toolkit for constructing withdrawal strategies. the results presented here should be considered guidelines, provided by financial history, to assist in the design and calibration of these strategies. notes 1 this characteristic of non-commutativeness (changing the order in the sequence does change the result) lies at the heart of the sequencing factor formula. it is thus e.d. suarez / financial services review 28 (2020) 96–132 129 fundamentally different from the “volatility drag” formula (½ s2), and we thank an anonymous reviewer for comments that prompted this clarification. 2 this expression is valid only when it produces non-negative values. if it is negative it must have happened that the withdrawal amount was larger than the available balance at some point, which is either not allowed or implies a change in the rate of return (to a rate of interest). 3 bridges and choudhury (2007) report that 89% of the population aged 65 and older receive social security benefits. in turn, coile and milligan (2009) present data from the health and retirement study that show that for households where the older member of the couple is in the 65-69 age bracket, 39.1% of their financial assets are not in their iras, nor as stocks, bonds, or bills, but in checking, savings, and money market accounts. 4 we thank an anonymous referee for bringing to our attention the need to establish this point clearly. 5 this example, borne out of actual historical data, illustrates the extent to which sequencing can be crucial indeed. the return rates in both periods produced the same total return, yet the corresponding pwas were very different solely by the effect of sequencing. if the corresponding retirees had made periodic adjustments to the withdrawal amount it would not have made any difference: the one retiring in ’49 would have fared well, the one retiring in ’69 would have done poorly. 6 this is perhaps a very relevant observation. studies of actual asset allocation choices have found more than 55% of people of retirement age having no equity at all in their ira portfolios (waggle and englis, 2000). 7 the median for the 100% bequest all-stocks pwr with 30-year horizons over the historical record is 6.4%. with the caveats mentioned regarding mid-course corrections, this could then be called the “6.4% rule”. 8 strictly speaking, the problem should be posed as the limit convergence of eq. (1) with ks = ke as n increases. we use the results for n=30 as an approximation, with the next section discussing their sensitivity to extensions of this horizon. we also simplify by ignoring additional future contributions. 9 the number of pwr figures that can be produced with our dataset is 648 for the 35year horizon length and 588 for the 40-year length. this is less than the 708 data points we had in the standard scenario, so the validity of these findings for inference purposes is correspondingly weaker. 10 social security administration (2016). 11 for the 25-year horizon length, the number of pwr figures that can be produced with our dataset is 768. the dashed line in fig. 21 (25-year long horizon) is thus longer than the solid line (30-year long horizon). 12 the number of data points for these horizon lengths is 828 (20 years), 888 (15 years), 948 (10 years), and 1,008 (5 years). so the lines in fig. 24 become longer as the planning horizon shortens. 13 john hancock investments’ retirement living portfolios enter the retirement stage with a 50% stock allocation. fidelity investments’ income replacement funds offers information that implies allocations of 53%, 34%, and 13% to stocks, bonds, and cash 130 e.d. suarez / financial services review 28 (2020) 96–132 in year 1 of retirement. vanguard’s managed payout fund invests 57.5% in stocks, 17.9% in bonds, and 24.6% in “other”. charles schwab’s monthly income fundenhanced payout allocates between 10% and 40% to equities, and the monthly income fund-maximum payout uses 0% to 25% as its equity allocation range. references basu, a. k., byrne, a., & drew, m. e. (2011). dynamic lifecycle strategies for target date retirement funds. the journal of portfolio management, 37, 83-96 bengen, w. p. (1994). determining withdrawal rates using historical data. journal of financial planning, 7, 171-180. bierwirth, l. (1994). investing for retirement: using the past to model the future. journal of financial planning, 7, 14-24. blanchett, d. m., & frank, l. r. (2009). a dynamic and adaptive approach to distribution planning and monitoring. journal of financial planning, 22, 52-66. bridges, b., & choudhury, s. (2007). social security benefits as a retirement resource for u.s. near-retirees. review of income and wealth, 53, 538-567. charles schwab investment management, inc. (2015). schwab monthly income funds: three diversified income solutions (available at http://ims.schwab.wallst.com/repository/?doc=schwabmonthlyincomefundssalessheet). clare, a., seaton, j., smith, p. n., & thomas, s. (2016a). reducing sequence risk using trend following investment strategies and the cape. discussion papers in economics. university of york, 16, 11. clare, a., seaton, j., smith, p. n., & thomas, s. (2016b). the trend is our friend: risk parity, momentum and trend following in global asset allocation. journal of behavioral and experimental finance, 9, 63-80. coile, c., & milligan, k. (2009). how household portfolios evolve after retirement: the effect of aging and health shocks. review of income and wealth, 55, 226-248. cooley, p. l., hubbard, c. m., & walz, d. t. (1998). retirement savings: choosing a withdrawal rate that is sustainable. journal of the american association of individual investors, 20, 16-21. cooley, p. l., hubbard, c. m., & walz, d. t. (1999). sustainable withdrawal rates from your retirement portfolio. financial counseling and planning, 10, 39-47. cooley, p. l., hubbard, c. m., & walz, d. t. (2003). a comparative analysis of retirement portfolio success rates: simulation versus overlapping periods. financial services review, 12, 115-129. cooley, p. l., hubbard, c. m., & walz, d. t. (2011). portfolio success rates: where to draw the line. journal of financial planning, 24, 48-60. estrada, j. (2018). maximum withdrawal rates: an empirical and global perspective. the journal of retirement, 5, 57-71. fidelity investments, inc. (2015). fidelity income replacement funds: at-a-glance (available at http://personal. fidelity.com/myfidelity/insidefidelity/newscenter/mediadocs/firf_at_a_ glance.pdf). frank, l. r., mitchell, j. b., & blanchett, d. (2011). probability-of-failure-based decision rules to manage sequence risk in retirement. journal of financial planning, 24, 44-53. guyton, j. t., & klinger, w. j. (2006). decision rules and maximum initial withdrawal rates. journal of financial planning, 19, 49-57. ibbotson, r. g., & sinquefield, r. a. (1976). stocks, bonds, bills, and inflation: year-by-year historical returns (1926-1974). the journal of business, 49, 11-47. ibbotson, s. b. b. i. (2015). classic yearbook:market results for 1926-2014. john hancock funds, llc. (2016). asset allocation: redefining diversification in your portfolio (available at http://www.jhinvestments.com/cms/downloadableitems/educationandguidance/p_piaabr-rl.pdf). mitchell, j. b. (2011). retirement withdrawals: preventive reductions and risk management. financial services review, 20, 45-59. e.d. suarez / financial services review 28 (2020) 96–132 131 morningstar inc. (2015). 2015 ibbotson stocks, bonds, bills, and inflation (sbbi) classic yearbook. chicago, il: morningstar inc. pye, g. b. (2000). sustainable investment withdrawals. the journal of portfolio management, 26, 73-83. social security administration. (2016). calculators: life expectancy (available at https://www.ssa.gov/planners/ lifeexpectancy.html). spitzer, j. j. (2008). retirement withdrawals: an analysis of the benefits of periodic ‘midcourse’ adjustments. financial services review, 17, 17-29. stout, r. g. (2008). stochastic optimization of retirement portfolio asset allocations and withdrawals. financial services review, 17, 1-15. stout, r. g., & mitchell, j. b. (2006). dynamic retirement withdrawal planning. financial services review, 15, 117-131. suarez, e. d., suarez, a., & walz, d. t. (2015). the perfect withdrawal amount: a methodology for creating retirement account distribution strategies. financial services review, 24, 331-357. the vanguard group, inc. (2016). vanguard managed payout fund (available at https://personal.vanguard.com/ us/funds/snapshot?fundid=1498&fundintext= int). waggle, d., & englis, b. (2000). asset allocation decisions in retirement accounts: an all-or-nothing proposition?. financial services review, 9, 79-92. waring, m. b., & siegel, l. b. (2015). the only spending rule article you will ever need. financial analysts journal, 71, 91-107. zolt, d. m. (2013). achieving a higher safe withdrawal rate with the target percentage adjustment. journal of financial planning, 26, 51-59. 132 e.d. suarez / financial services review 28 (2020) 96–132 academy of financial services officers president janine sam shepherd university president-elect inga timmerman california state university, northridge executive vice president-program terrance k. martin utah valley university vice president-communications david nanigian california state university, fullerton vice president-finance thomas p. langdon roger williams university vice president-international relations philip gibson winthrop university vice president-mktg & public relations chris browning texas tech university immediate past president swarn chatterjee university of georgia editor, financial services review stuart michelson stetson university directors shawn brayman planplus global charles chaffin cfp board of standards victoria javine university of alabama colleen tokar-asaad baldwin wallace university frances lawrence university of missouri tom potts baylor university robert moreschi virginia military institute past presidents swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie halil kiymaz, 2010-11 rollins college of business david lange, 2009-10 auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university published in collaboration with the financial planning association financial services review is the journal of the academy of financial services, published in collaboration with the financial planning association. membership dues of $125 to the academy include a one-year subscription to the journal. financial planning association members receive digital access to the current volume/issue of the journal. how to submit: membership in afs ($125) is required to submit an article to financial services review. join afs at academyfinancial.org. a submission fee of $100 per article should be paid at: https://academyoffinancialservices.wildapricot.org/submit-an-article. submit your article electronically as an email attachment in word format only (no pdfs please) to the editor stuart michelson at smichels@stetson.edu. should a manuscript revision be invited, no additional fees will be required. style information for the manuscripts can be found on the inside back cover of this journal. copyright © 2020 academy of financial services. all rights of reproduction in any form reserved. financial services review the journal of individual financial management vol. 28, no. 1, 2020 editor stuart michelson, stetson university associate editors benefits and retirement planning vickie bajtelsmit colorado state university stephen m. horan cfa institute walter woerheide the american college estate planning anne wenger san diego state university giovanni fernandez stetson university investments robert brooks university of alabama john clinebell university of northern colorado james dilellio pepperdine university dale domian york university jim gilkeson university of central florida william jennings united states air force academy david nanigian csu fullerton insurance larry cox university of mississippi financial planning swarn chatterjee university of georgia sherman hanna ohio state university patti fisher virginia tech university wade d. pfau the american college john salter texas tech university financial institutions stanley d. smith university of central florida investor psychology and counseling john nofsinger washington state university meir statman santa clara university financial literacy ning tang san diego state university international lawrence rose massey university education jerry stevens university of richmond financial planning profession tom warschauer san diego state university co-published by the academy of financial services and the financial planning association the editor of financial services review wishes to thank the stetson university, school of business, for its continuing financial and intellectual support of the journal. aims and scope: financial services review is the official publication of the academy of financial services. the purpose of this refereed academic journal is to encourage rigorous empirical research that examines individual behavior in terms of financial planning and services. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial issues. the journal provides a forum for those who are interested in the individual perspective on issues in the areas of financial services, employee benefits, estate and tax planning, financial counseling, financial planning, insurance, investments, mutual funds, pension and retirement planning, and real estate. publication information. financial services review is co-published quarterly by the academy of financial services, and the financial planning association. institutional subscription price is $100. personal subscription price is $125 and is available by joining the academy of financial services. further information on this journal and the academy of financial services is available from the website, http://www.academyfinancial.org. postmaster and subscribers should send change of address notices to stuart michelson, academy of financial services, stetson university, school of business, 421 n. woodland blvd., unit 8398, deland, fl 32723. editorial office: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email address: smichels@stetson.edu. web address: www.academy financial.org. advertising information. those interested in advertising in the journal should contact stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. email address: smichels@stetson.edu, (386) 822-7376. printed in the usa © 2020 academy of financial services. all rights reserved. this journal and the individual contributions contained in it are protected under copyright by the academy of financial services, and the following terms and conditions apply to their use: photocopying single photocopies of single articles may be made for personal use as allowed by national copyright laws. in addition, the academy of financial services hereby permits educators and educational institutions the right to make photocopies for non-profit educational classroom use. permission of the academy is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. permissions may be sought directly from the editor, stuart michelson. contact information: stuart michelson, school of business, stetson university, 421 n. woodland blvd., unit 8398, deland, fl 32723. phone: (386) 822-7376. email: smichels@stetson.edu. derivative works subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their institutions. permission of the academy is required for resale or distribution outside the institution. permission of the academy is required for all other derivative works, including compilations and translations. electronic storage or usage permission of the academy is required to store or use electronically any material contained in this journal, including any article or part of an article. except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permission of the academy. address permissions requests to the editor at the address noted above. notice no responsibility is assumed by the academy for any injury and/or damage to persons or property as a matter of products liability, negligence, or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. the editors, the academy, and our sponsors assume no responsibility for the views expressed by our authors. financial capability: literacy, behavior, and distress janine k. scotta,*, nghia nguyen vub, yuanshan chengc, philip gibsonc ashepherd university, college of business administration, shepherdstown, wv 25443, usa bschool of economics & finance, massey university, palmerston north, 4442, new zealand cdepartment of finance, winthrop university, college of business and administration, rock hill, sc 29733, usa abstract we inspect the influence of individual financial knowledge and financial behavior on the probability of experiencing financial distress. using the 2015 national financial capability study, we examine three measures of financial distress related to bill payment, retirement saving, and being late with a mortgage payment. financial literacy and financial behavior indices are constructed using questions from the survey that pertain to financial knowledge (ranging in complexity) and financial decision-making. in addition to the influence of socioeconomic factors, the conclusion suggests that financial literacy and positive behavior reduces financial distress stemming from simple financial matters. however, the opposite is observed for more complex financial decisions. © 2018 academy of financial services. all rights reserved. jel classification: d10; d14; d19 keywords: financial distress; financial behavior; literacy 1. introduction the recognition of financial literacy has been increasing remarkably as a crucial skill for individuals in the modern world. moreover, financial literacy has gained the notice of many stakeholders, including academic scholars and legislators. given the uncertainty of welfare systems in many developed countries, financial risks and responsibilities are being deflected to government institutions, such as in the form of health insurance and retirement saving * corresponding author. tel.: �1-304-876-5281. e-mail address: jscott@shepherd.edu (j.k. scott) financial services review 27 (2018) 391-411 1057-0810/18/$ – see front matter © 2018 academy of financial services. all rights reserved. programs or initiatives. meanwhile, the increasing complexity of financial products serves as a barrier for many individuals to effectively evaluate and compare financial products as they try to engage with financial markets. a detrimental factor associated with hindering financial capability is the enlargement of credit and mortgage segments, where it is apparent that access and subsequent informed behavior does not go hand in hand. for example, many find it challenging to afford daily expenses because of risky financial decisions such as taking on “too much house” in the form of unaffordable mortgages, signing contracts with terms and conditions not fully understood, and spending far more than earning, impeding “rainy day” savings. many researchers suggest the application of financial knowledge can be seen in consistent saving, retirement planning, and building investments. while studies have examined the correlation of financial literacy and financial behavior (fernandes, lynch, and netemeyer 2014), little is known on how financial literacy and financial knowledge impact financial distress. this paper aims to widen this area of research by introducing analysis on a (united states) national scale. specifically, we analyze whether financial literacy affects making sound financial decisions; thus, reducing financial distress. however, to effectively manage personal finances, there should be a certain level of rationality with similar behavior. therefore, financial behavior is a determinant that will be considered to assess the financial performance of an individual in different financial circumstances. the 2015 national financial capability study is useful for examining causes of financial distress; looking at the capability of individuals to meet basic and more complex financial needs, as well as the skills and knowledge necessary to make wise financial decisions. we use a set of questions from the survey to construct a financial literacy index, a financial behavior index, and to assess three aspects of financial distress. this includes difficulty in paying bills, worrying about retirement saving and being late with a mortgage payment. we find that financial literacy positively contributes to the prevention of financial distress. nonetheless, financial behavior emerges as having more of an impact on experiencing financial distress, more so than financial literacy. the results are statistically significant for all three measures. this study is structured as follows: in the following section, we review the existing literature on financial literacy, financial behavior, and financial distress. the third section of the paper describes the data used and the socioeconomic characteristics of the sample. the model and methodology are presented in the fourth section, where the construction of the measures for financial literacy, financial behavior, and over-indebtedness are explained in more detail. section 5 presents the data analysis, and section 6 summarizes and concludes. 2. literature review 2.1. financial literacy: concept and measurement in general, literacy refers to an individual’s capability of reading and writing (zarcadoolas, pleasant, and greer 2006). the standard definition of literacy consists of understanding (i.e., recognition of vocabulary and arithmetic operations) and using forms of documents. 392 j.k. scott et al. / financial services review 27 (2018) 391-411 huston (2009) has also conceptualized financial literacy as including two dimensions— understanding (knowledge of personal finance) and use (application of personal finance). remund (2010) defines financial literacy as a measure of the degree to which an individual understands personal finance—familiarity with economic concepts and having the ability/ skills and confidence to manage personal finances. this should lead to both proper short-term decision-making and long-range financial planning, considering life cycle events and uncertain economic circumstances. based on the definition approved by the president’s advisory council on financial literacy (pacfl, 2009), hung et al. (2009, p. 12) explains financial literacy as: “knowledge of basic economic and financial concepts, as well as the ability to use that knowledge and other financial skills to manage financial resources effectively for a lifetime of financial well-being.” it is evident that financial literacy extends beyond mere financial knowledge. atkinson and messy (2011, p. 4) looks at financial literacy via the field of financial education as “a combination of awareness, knowledge, skills, attitude, and behaviors necessary to make sound financial decisions and ultimately achieve individual financial well-being.” allgood and walstad (2016) also summarize the effects of financial literacy on financial behaviors of credit cards, investments, loans, insurance, and financial advice. financial capability is a complicated and broader concept not merely about understanding, but also behaviors and actions. hence, the term “financial capability” was adopted by several countries such as the united kingdom, canada, and the united states, which include three main components: (1) knowledge, (2) skills, and (3) confidence and attitudes (kempson, collard, and moore, 2005). both terms—financial literacy and financial capability—cover decision-making, practical skills, and behavior as well as knowledge and understanding (o’connell, 2007). it should be noted, that financial capability also considers the relevance of outside institutions and regulation. it calls for individuals to not only develop financial knowledge and skills, but also gain access to financial instruments and institutions (johnson and sherraden, 2007). in attempts to measure financial literacy, huston (2010) converts theoretical concepts into measurable criteria by employing both performance tests and self-reported tests. moreover, van rooij, lusardi, and alessie (2007) construct a financial literacy index (using five questions) and a progressive financial literacy index (using 11 items) based on a factor analysis. moore (2003) adopted an approach using 12 financial-management questions, covering budget management, credit, savings, investment, mortgages, and a broad category of other financial topics. in 2007, lusardi and mitchell introduced a measurement method using three “simple” questions related to the compounding of interest rates, inflation, and risk diversification. in an earlier piece of research, hilgert, hogarth, and beverly (2003) introduced a 28-questionnaire about credit, saving, budget management, investment, and mortgages along with an 18-question quiz to test individuals’ financial behavior. it should be noted that several countries have collected financial data using academic approaches to date (table 1). unfortunately, divergence in content and methodologies discourages international comparison which could be useful, despite the availability of data. therefore, the oecd and its international network for financial education (infe) call for developing and applying a financial literacy questionnaire by atkinson and messy (2011, 393j.k. scott et al. / financial services review 27 (2018) 391-411 2012) in 14 countries. the three component-questionnaire includes knowledge, behavior, and attitude. there is an overall financial literacy score, which is comprised of the sum of the scores for knowledge, behavior, and attitude. questions about interest, risk, and return, and inflation are used to assess respondents’ financial awareness. financial behavior is observed through issues related to saving, budget management, borrowing, and choosing financial products. attitude towards money and future planning are also used to determine financial position. most financial literacy research generally focuses on four main points: (1) measuring the level of financial literacy, (2) valuing the effects of financial literacy on financial behavior table 1 panel a: national financial capability survey across countries countries financial capability survey the united kingdom levels of financial capability in the united kingdom: results of a baseline survey (fsa, 2006) the united states national financial capability study (finra, 2009, 2015) new zealand financial knowledge survey (anz-retirement commission, 2009) ireland financial capability: new evidence for ireland (keeney and o’donnell, 2009) canada understanding financial capability in canada: analysis of the canadian financial capability survey (mckay, 2011) the netherlands financial literacy and retirement planning in the netherlands (van rooij et al., 2009) portugal survey on the financial literacy of the portuguese population (2010) (banco de portugal, 2011) panel b: list of dependent variables variables description male 1 if male, 0 if female age 25–54 1 if in group age 25–44 and 45–54, 0 if otherwise age 55� 1 if in group age 55–64 and 65�, 0 if otherwise swd 1 if single, widowed, or divorced, 0 if otherwise college 1 if the individual has a college education (college graduate or post-graduate education) and 0 otherwise (did not complete high school, high school graduate or has some college) unemployed 1 if the individual is unemployed and 0 otherwise inactive 1 if the individual is inactive (full-time student, homemaker, permanently sick, disabled, or unable to work) and 0 if otherwise mortgage 1 if have any mortgage, 0 if otherwise white 1 if white, 0 if non-white south 1 if south, 0 if otherwise drop in income 1 if experience a large drop in income in the past 12 months, 0 if otherwise married 1 if married, 0 if otherwise children 1 if at least 1 dependent child, 0 if no dependent child inc1 1 if the annual household income is at least $25,000 but less than $50,000 and 0 otherwise inc2 1 if the annual household income is at least $50,000 but less than $100,000 and 0 otherwise inc3 1 if the annual household income is above $100,000 and 0 otherwise self-employed 1 if the individual is self-employed and 0 otherwise retired 1 if the individual is retired and 0 otherwise fli financial knowledge index fbi financial behavior index 394 j.k. scott et al. / financial services review 27 (2018) 391-411 and attitude, (3) assessing unique characteristics that widely influence financial literacy levels, and (4) evaluating financial literacy programs (altintas, 2011). however, there is a noticeable difference in measuring financial literacy through objective tests and examining respondents’ self-assessment and their perception of financial issues. indeed, past research demonstrates a discrepancy between what individuals believe they know and what they do know, with the self-assessment often more positive than actual understanding (asaad, 2015). 2.2. financial distress: concept and causes financial distress in the context of corporate finance shows that in some circumstances, a company can avoid moving from financial distress to bankruptcy. however, it requires a lot of effort and financial support. often, financial distress can come with costs, such as fees paid to lawyers or the charges of additional interest for late payments. if financial distress cannot be fully improved or at least minimized, bankruptcy becomes unavoidable. when it comes to individual financial distress, it can be viewed as not saving enough for retirement, accruing excessive liabilities, and not making use of financial innovation (campbell, 2006; lusardi and mitchell, 2007). ware (2015) cites personal financial distress as a danger to households. prawitz, kim, and garman (2006) describe financial distress as a response, such as a mental or physical discomfort pertaining to general financial well-being. this definition includes perceptions about one’s capacity to manage wealth, for example, paying bills, repaying debts, and maintaining the basic needs and wants of life. similarly, garman et al. (2004) introduce financial distress as intense physical or mental tension that includes concerns and worries about financial situations. financial distress can endure for a short period, or it can become a constant state. stressful events that can lead to financial distress include receiving overdue notifications from creditors and collection agencies, issuing checks without having sufficient funds, being late with bill payments, and feeling depressed about being unprepared for major life events such as retirement. worrying about financial problems affects many other aspects of a person’s life, such as health, productivity, and relationships. despite the severe adverse spill-over effects of financial distress, measuring the constructs of perceived financial distress and financial well-being can be viewed as a worthwhile goal of many researchers. with adequate education and consulting interventions, it may be possible to measure whether financial lives are positively changed (garman et al., 2004). with that endeavor in mind, a team of researchers developed the incharge financial distress/financial well-being (ifdfw) scale, an eight-item self-reported subjective measure of financial distress/financial well-being (garman et al., 2005; prawitz et al., 2006). the ifdfw scale measure respondents’ feelings about their financial situation on a continuum, from overwhelming financial distress, the lowest level of financial well-being, to no financial distress or the highest level of financial well-being. saving goals are useful clues when it comes to assessing financial behavior. browning and lusardi (1996) suggest eight saving goals, namely: precautionary needs, foresight, calculation, improvement, independence, enterprise, pride, and avarice. more recently, lee and hanna (2015) examine associations between saving goals and saving behavior from a perspective of maslow’s hierarchy. using the 1998–2007 surveys of consumer finance datasets, they find that the retirement/security goal was the most frequently mentioned, and 395j.k. scott et al. / financial services review 27 (2018) 391-411 the self-actualization goal was the least cited as the reason for saving behavior. in other words, besides providing for the necessities of life and paying off debt, not saving enough for retirement is a key contributing factor to financial distress. demographic factors also play a role in explaining the cause of financial distress. from previous studies, it has been noted that having dependent children, being separated or divorced, earning a low income, and/or having no job, increases the likelihood of financial distress. financial distress has also been linked to gender, with men being less likely to have debt—and to age, with younger people being more at risk (orc, 2015) because they are less reluctant to use credit to finance expenses (disney et al., 2008). moreover, living in a region dominated by poverty such as in the southern states of the united states, is also a factor contributing to financial difficulties (fram, miller-cribbs, and van horn, 2007). after examining and comparing the national financial capability study, the incharge financial distress/financial well-being (ifdfw), and maslow’s hierarchy as it relates to personal finance, three measures of financial distress are assessed in this piece of research: (1) difficulty in paying bills, (2) worrying about retirement saving, and (3) being late with a mortgage payment. given the lack of a single widely accepted measurement of financial distress, the three measurement approaches are vital in developing our study. 2.3. financial literacy and individual financial decisions in addition to investigating the relation between financial literacy and financial behavior, researchers have also looked at evidence of correlation and causality between knowledge and practice in personal finance. chen and volpe (2002) showed that the enthusiasm for a personal finance course varied by gender. women were more enthusiastic about english and humanities and men favored mathematics and science. moore (2003) concludes that individuals with a lower level of financial literacy are expected to make poor financial decisions, such as taking on costly mortgages, because of inadequate understanding of basic financial concepts (like that of compound interest rates). the gap in financial knowledge between lenders and borrowers also lends credence to borrowers’ mortgage experiences, that is, taking out loans with more uncertain or less beneficial contractual terms (campbell, 2006). using data from the uk financial capability survey, mccarthy (2011) examines the relation between financial distress and financial literacy. alongside personal traits, mccarthy (2011) finds that people with higher levels of financial literacy are less likely to experience financial distress. in summary, while several papers examine different aspects of financial distress and the impact on behavior, in addition to financial literacy on financial outcomes, prior research has not studied the effect of financial literacy and behavior while focusing specifically on financial distress—be it mild or extreme. therefore, this piece of research, which uses a nationally representative dataset from finra (2015), aims to provide insight into the fundamental causes of financial distress. 396 j.k. scott et al. / financial services review 27 (2018) 391-411 3. data we use the 2015 national financial capability study (nfcs), a state-by-state online survey of 27,564 (2015) respondents. survey respondents are age 18 years or older, with approximately 500 interviewed in each of the 50 states plus the district of columbia. the survey is dedicated to delving deeper into individual financial capability. most survey questions focus on eight financial themes. the first section includes habits and attitudes in managing the family budget as well as willingness to accept risk, household spending, saving for a “rainy day,” amassing savings for retirement or college education, and whether a significant drop in income was experienced in the last year. the second section refers to the use of financial advice related to debt, saving, investing, insurance, and tax planning. the third part is devoted mainly to banks and financial issues. the fourth section focuses on retirement accounts and pensions. the fifth section primarily surveys homeownership, specifically monthly mortgage payments, and any experience with debt or foreclosure. the sixth section focuses on the use of credit cards, and the seventh addresses consumer loans. the eighth section covers insurance. in this paper, several sections of the survey are used to investigate our research question. the survey also includes a group of questions designed to probe respondents’ financial knowledge. the study includes a set of socioeconomic questions about gender, age, race, living region, education, marital status, living arrangements, income, employment status, the number of dependent children, and the most knowledgeable household member when it comes to financial concepts. the purpose of this paper is to identify the main factors that might drive an individual into financial distress. financial literacy is an input that contributes to better financial behavior and this should reduce the probability of falling into a circumstance of financial distress. our empirical specification, detailed in the next section of the paper, has both financial literacy and financial behavior as two different explanatory variables. we consider three measures of financial distress in this research: having trouble in servicing lifestyle needs, worrying about retirement, and being late with mortgage payments. the perception of financial knowledge and behavior in this paper is consistent with the definition of financial literacy proposed by hung et al. (2009). 3.1. financial distress measure consistent with the definitions of financial distress reviewed, we examine difficulty in paying bills, worrying about retirement saving and being late with a mortgage payment. experiencing financial distress is based on the responses to the following questions: 1. in a typical month, how difficult is it for you to cover your expenses and pay all your bills? (a) very difficult*; (b) somewhat difficult*; (c) not at all difficult; (d) don’t know; (e) prefer not to say. 2. how strongly do you agree or disagree with the following statements—i worry about running out of money in retirement? (a) disagree 1–3; (b) neutral 4; (c) agree 5–7; (d) don’t know; (e) prefer not to say. 397j.k. scott et al. / financial services review 27 (2018) 391-411 3. how many times have you been late with your mortgage payments in the last 2 years? (a) never; (b) once*; (c) more than once*; (d) don’t know; (e) prefer not to say. table 2 shows the descriptive statistics for the financial literacy questions and financial distress questions. as shown in panel b, close to 50% of the sample indicated having either a difficult or somewhat difficult time paying their bills on time. when respondents were asked to state their concern about running of money in retirement on a seven-point likert scale, ranging from one, strongly disagree to seven, strongly agree, over 25% selected strongly agree. overall close to 56% of respondents indicated that they are concerned about retirement funding. when asked about making mortgage payments on time, 31% responded that they had not been late on a mortgage payment in the last two years. however, this question does have a very high missing rate (63.6%). among those who answered the questions, 84% of them did not miss any mortgage payment. there are, however, around 14% of those reported that they either missed one or more mortgage payments. table 2 panel a: financial literacy questions variable name correct incorrect don’t know n prob n prob n prob interest rate 21,234 77.04 3,492 12.67 2,838 10.30 inflation 17,101 62.04 5,311 19.27 5,152 18.69 bond 8,199 29.75 9,154 33.21 10,211 37.04 mortgage 2,1399 77.63 2,132 7.73 4,033 14.63 risk 13,412 48.66 2,714 9.85 11,438 41.50 panel b: financial distress questions variable name answers n prob bill paying don’t know 380 1.38 not at all difficult 13,622 49.42 prefer not to say 188 0.68 somewhat difficult 10,485 38.04 very difficult 2,889 10.48 retirement concern strongly disagree-1 2,404 8.72 2 1,944 7.05 3 2,079 7.54 neutral 4,991 18.11 5 4,160 15.09 6 4,208 15.27 strong agree-7 7,079 25.68 don’t know 560 2.03 prefer not to say 139 0.50 mortgage late payment don’t know 131 0.48 more than once 780 2.83 never 8,430 30.58 once 652 2.37 prefer not to say 41 0.15 missing 17,530 63.60 (continued on next page) 398 j.k. scott et al. / financial services review 27 (2018) 391-411 3.2. financial literacy measure we use a set of five financial literacy questions from the nfcs to evaluate financial knowledge and to construct a financial literacy measure (finra, 2015). atkinson and messy (2012) used this approach with the inclusion of questions such as: 1. suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow: (a) more than $102*; (b) exactly $102; (c) less than $102; (d) don’t know; (e) prefer not to say. 2. imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after one year, how much would you be able to buy with the money table 2 (continued) panel c: financial behavior questions spending control don’tknow 797 2.89 prefer not to say 146 0.53 spending equal to income 10,427 37.83 spending less than income 11,358 41.21 spending more than income 4,836 17.54 overdraw checking don’t know 172 0.62 no 20,585 74.68 prefer not to say 92 0.33 yes 4,554 16.52 missing 2,161 7.84 calculate retirement don’t know 750 2.72 no 11,843 42.97 prefer not to say 199 0.72 yes 8,977 32.57 missing 5,795 21.02 emergency fund don’t know 789 2.86 no 13,316 48.31 prefer not to say 341 1.24 yes 13,118 47.59 pay card in full don’t know 189 0.69 no 9,826 35.65 prefer not to say 113 0.41 yes 1,1647 42.25 missing 5,789 21 payday loan 1 time 925 3.36 2 times 710 2.58 3 times 565 2.05 4 or more times 852 3.09 don’t know 203 0.74 never 24,193 87.77 prefer not to say 116 0.42 health insurance cover don’t know 260 0.94 no 2,726 9.89 prefer not to say 130 0.47 yes 24,448 88.7 399j.k. scott et al. / financial services review 27 (2018) 391-411 in the account? (a) more than today; (b) exactly the same; (c) less than today*; (d) don’t know; (e) prefer not to say. 3. if interest rates rise, what will typically happen to bond prices? (a) they will rise; (b) they will fall*; (c) they will remain the same; (d) there is no relationship between bond prices and the interest rate; (e) don’t know; (f) prefer not to say. 4. a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. (a) true*; (b) false; (c) don’t know; (d) prefer not to say. 5. buying a single company’s stock usually provides a safer return than a stock mutual fund. (a) true; (b) false*; (c) don’t know; (d) prefer not to say. table 2, panel a shows that most respondents provided the correct answer to the interest rate question (77%), inflation question (62%), and mortgage question (77%). however, the percentage of correct answers decreased when respondents were tested on their knowledge regarding the impact of inflation on purchasing power. the worst performance is found in responses to the bond price question, where 29.75% of respondents failed to select the correct choice, and 37% admitted not having any clue about the answer; regarding the risk question, 41.5% of respondents admit to not knowing the correct answer. from the financial literacy quiz, a financial literacy index is formed (fli) that reflects the percentage of questions answered correctly. the “don’t know” and “prefer not to say” choices were categorized as wrong answers. the fli is assigned to distinct values of 0, 0.2, 0.4, 0.6, 0.8, and 1. 3.1. financial behavior measure previous studies addressed mainly the effect of financial literacy on financial behavior. the goal of this study is to examine the effect of financial literacy on financial wellbeing/ distress. thus, it is necessary to control for financial behavior in this study to isolate its effect on financial wellbeing/distress. this study follows atkinson and messy (2011) to create a financial behavior index. to measure financial behavior, we focus on seven questions from the survey that touch on budget management, savings, credit, insurance, and financial advice. this approach resembles one adopted by atkinson and messy (2011), as shown below: 1. over the past year, would you say your household’s spending was less than, more than, or about equal to your household’s income? (…) (a) spending less than income*; (b) spending more than income; (c) spending about equal to income*; (d) don’t know; (e) prefer not to say. 2. do you or your spouse/partner overdraw your checking account occasionally? (a) yes; (b) no*; (d) don’t know; (e) prefer not to say. 3. have you ever tried to figure out how much you need to save for retirement? (non-retired respondent) or before you retired, did you try to figure out how much you need to save for retirement? (retired respondent). (a) yes*; (b) no; (d) don’t know; (e) prefer not to say. 400 j.k. scott et al. / financial services review 27 (2018) 391-411 4. have you set aside emergency or rainy-day funds that would cover your expenses for three months, in the case of sickness, job loss, economic downturn, or other emergencies: (a) yes*; b) no; (d) don’t know; (e) prefer not to say. 5. in the past 12 months, which of the following describes your experience with credit cards? i always paid my credit cards in full: (a) yes*; (b) no; (d) don’t know; (e) prefer not to say. 6. please indicate if in the past five years you have taken out a short-term “payday” loan? (a) yes; (b) no*; (d) don’t know; (e) prefer not to say. 7. are you covered by health insurance? (a) yes*; (b) no; (d) don’t know; (e) prefer not to say. in table 2, panel c, reports the descriptive statistics for the financial behavior questions. eighteen percent of respondents indicated that spending in the past year exceeded income (question 1), 17% reported overdrawing their checking account occasionally (question 2). planning is critical for retirement preparedness or making provisions to buffer against adverse shocks. close to half of respondents (43%) who answered the third question have not tried to determine how much they should put into a retirement saving account. additionally, 49% of respondents have not set aside an emergency or a rainy day fund (question 4). concerning credit behavior, 36% of our survey respondents do not pay their credit card balance fully, which translates into increased interest costs over time (question 5). alongside that result, when it comes to alternative forms of borrowing, such as taking on a payday loan, 11% of respondents use this type of high-cost lending method (question 6). finally, regarding insurance coverage, 10% of respondents report not being covered by a health insurance plan (question 7). based on the questions above, the fbi (financial behavior index) is constructed by scoring the respondents’ answers. in question (1), the response for “spending less than income” is equal to (‘2’), for “spending about equal to income,” (‘1’), and for “spending more than income,“ (‘0’). for questions (3) to (7), except question (6), a “yes” is scored as (‘1’), and a “no” is scored as (‘0’). for questions (2) and (6) a “no” takes on a value of (‘1’), and a “yes” on a value of (‘0’). all the valid scores from the fbi are summed and then divided by 7. for those that are missing, the authors equated that to not practicing positive financial behavior. the authors separately analyzed the findings by dropping those with missing financial behavior questions; the results are very similar. table 3 provides a list of each dependent and independent variable that is used in this study. the whole sample is comprised of 27,564 respondents. we also present a cross tabulation of our independent variables using our financial distress measurements. we find that 45% of the respondents are male, 55� years old (36%), white (excluding hispanics; 72%), living in the southern region of the united states (33%). most respondents are married (54%) without dependent children (63%). almost half of those surveyed (49%) work for an employer, and more than 34% have an annual income that ranges between $50,000 and $100,000, and 21% stated that they recently experienced a decline in income. as expected, those who had lower income or recently experienced a decrease in income made up a greater portion of respondents who indicated that they recently had difficulty paying bills. those 401j.k. scott et al. / financial services review 27 (2018) 391-411 who demonstrated a lower level of financial literacy and scored lower on the financial behavior index scale, also had a difficult time paying bills. 4. model and methodology 4.1. multivariate analysis to investigate the relation between financial literacy, financial behavior, and financial distress, we use a multinomial logistic regression model. our model takes into consideration respondents’ socioeconomic characteristics, in addition to experiencing an unexpectedly large drop in income. the dependent variable is the likelihood of a respondent being financially distressed, taking on a value of (‘1’) (yi � 1) if the respondent: (1) has difficulty paying bills; (2) is table 3 descriptive statistics (control variables) name value n prob mean of financial distress dummy bill pay retirement mortgage male 12,293 44.60 0.424 0.344 0.053 age 35–54 9,613 34.88 0.521 0.310 0.072 age 55 9,888 35.87 0.381 0.321 0.023 white 19,836 71.96 0.462 0.331 0.048 south 9,027 32.75 0.51 0.334 0.058 dependent children 10,066 36.52 0.561 0.313 0.098 married 15,007 54.44 0.42 0.329 0.065 sdw 4,449 16.14 0.554 0.312 0.036 income 50–100k 9,395 34.08 0.40 0.335 0.070 income � 100k 4,981 18.07 0.227 0.322 0.049 unemployed 1,556 5.65 0.701 0.308 0.040 self employed 1,985 7.20 0.506 0.323 0.065 inactive 5,068 18.39 0.605 0.341 0.044 drop in income 5,870 21.30 0.778 0.240 0.123 fli 0 1,886 6.84 0.575 0.367 0.048 fli 0.2 3,163 11.48 0.604 0.329 0.079 fli 0.4 5,048 18.31 0.595 0.326 0.078 fli 0.6 6,286 22.81 0.516 0.329 0.049 fli 0.8 6,677 24.22 0.419 0.338 0.039 fli 1 4,504 16.34 0.296 0.321 0.029 fbi 0 32 0.12 0.938 0.344 0.156 fbi 0.14 399 1.45 0.875 0.268 0.155 fbi 0.29 1,287 4.67 0.830 0.293 0.124 fbi 0.43 2,781 10.09 0.724 0.306 0.099 fbi 0.57 3,406 12.36 0.583 0.320 0.097 fbi 0.71 3,155 11.45 0.426 0.352 0.061 fbi 0.86 2,996 10.87 0.209 0.381 0.017 fbi 1 2,027 7.35 0.096 0.344 0.008 fbi na 11,481 41.65 0.502 0.328 0.029 note: fli � financial knowledge index; fbi � financial behavior index; sdw � single, divorced, or widowed. 402 j.k. scott et al. / financial services review 27 (2018) 391-411 uncertain about retirement saving, or (3) has been late with a mortgage payment in last two years, and (‘0’) otherwise. the variables considered (described in table 2) include gender, age, race, region, having children, marital status, education, income level, and employment status. the incidence of a significant drop in income is also considered. we expect that the probability of financial distress increases if respondents are: female, younger, have children, are divorced or separated, within the lower income bracket, are unemployed, and have a mortgage. as for financial literacy and financial behavior, we expect both to have a decreasing effect on the probability of financial distress. 5. data analysis in table 4, panel a, we explore the relation between our fli and the potential for financial distress because of the failure to pay bills on time. our results show that a one unit increase in the fli is associated with a statistically significant increase in the log odds of selecting the choices of “somewhat difficult” or “not at all difficult,” relative to those who experienced a very difficult time paying their bills. in other words, those who demonstrated a higher level of financial knowledge, while controlling for financial behavior, experienced less financial distress. to our knowledge, this is the first study to highlight this relation. previous studies primarily focused on the association between financial behavior and financial literacy. using the planned behavior theory suggest that financial knowledge can help to enforce a sense of control and reduce financial distress. similar results are found when we explore the connection between our fbi and paying bills on time. with regards to our control variables, younger respondents, age 35–54 are significantly more likely to fall into a state of financial distress. similar results were found among those who had dependents, were unemployed, self-employed, and experienced a decline in income. in table 4, panel b, we provide details on how our fli and fbi measures relate to financial distress resulting from the inability to make mortgage payments on time. compared with those who have never been late making a mortgage payment, a one-unit increase in the fli and fbi was associated with a decrease in the log odds of being late one or more times on a mortgage payment. in table 4, panel c, we shift our attention to focus on how financial literacy and financial behavior impact the concerns that people have about retirement. here we find unexpected but interesting results. as financial knowledge increases, as indicated by our fli, there is an increase in the log odds of being concerned about running out of money in retirement. similar results are also found when looking at our fbi. to test the robustness of the results, an additional analysis was conducted by separating the groups into different ages. we expect respondents in different age groups to have varying levels of concerns about retirement. the results shown in table 5 are consistent with table 4 panel c. when measured by something as simple as paying bills on time, more financial knowledge and positive financial behavior was associated with lower financial distress. the results do not persist for a more complex measure of financial distress (i.e., worrying about retirement). retirement planning is complex, and it involves many assumptions. therefore, regardless of the level of financial 403j.k. scott et al. / financial services review 27 (2018) 391-411 t ab le 4 pa ne l a : b ill pa yi ng (r ef : ve ry di ffi cu lt) so m ew ha t di ffi cu lt n ot at al l di ffi cu lt d on ’t kn ow pr ef er no t to sa y (i nt er ce pt ) 0. 80 3* ** � 1. 82 5* ** � 2. 40 7* ** � 3. 80 8* ** (0 .1 28 ) (0 .1 43 ) (0 .3 8) (0 .6 24 ) m al e � 0. 13 7* * 0. 08 7 0. 04 2 � 0. 33 2 (0 .0 67 ) (0 .0 72 ) (0 .2 02 ) (0 .3 28 ) a ge 35 –5 4 � 0. 19 2* ** � 0. 18 6* * � 0. 20 3 0. 52 7 (0 .0 74 ) (0 .0 81 ) (0 .2 31 ) (0 .4 00 ) a ge 55 � 0. 09 4 � 0. 06 6 0. 07 5 0. 64 8 (0 .1 05 ) (0 .1 12 ) (0 .3 04 ) (0 .4 90 ) w hi te � 0. 03 7 � 0. 08 6 � 0. 07 � 0. 14 5 (0 .0 69 ) (0 .0 75 ) (0 .2 09 ) (0 .3 37 ) so ut h � 0. 02 � 0. 07 6 � 0. 32 6 � 0. 49 6 (0 .0 67 ) (0 .0 73 ) (0 .2 14 ) (0 .3 53 ) d ep en de nt ch ild re n � 0. 21 6* ** � 0. 63 1* ** � 0. 75 9* ** � 1. 02 2* ** (0 .0 73 ) (0 .0 79 ) (0 .2 36 ) (0 .3 66 ) m ar ri ed 0. 06 4 0. 21 5* * � 0. 33 6 � 0. 20 2 (0 .0 83 ) (0 .0 91 ) (0 .2 55 ) (0 .4 09 ) sd w � 0. 04 3 0. 08 9 � 0. 12 4 0. 13 (0 .1 08 ) (0 .1 20 ) (0 .3 43 ) (0 .4 94 ) in co m e 50 –1 00 k 0. 37 9* ** 0. 89 1* ** 0. 86 0* ** 1. 27 1* ** (0 .0 76 ) (0 .0 82 ) (0 .2 30 ) (0 .3 72 ) in co m e � 10 0k 0. 42 8* ** 1. 45 0* ** 1. 45 9* ** 1. 89 8* ** (0 .1 22 ) (0 .1 25 ) (0 .3 12 ) (0 .4 75 ) u ne m pl oy ed � 0. 90 4* ** � 1. 18 8* ** 0. 58 4* * 0. 74 6 (0 .1 16 ) (0 .1 40 ) (0 .2 96 ) (0 .4 65 ) se lf em pl oy ed � 0. 36 6* ** � 0. 41 5* ** � 0. 14 8 � 0. 09 5 (0 .1 05 ) (0 .1 14 ) (0 .3 74 ) (0 .5 55 ) in ac tiv e � 0. 18 0* * � 0. 16 2* 0. 46 9* * 0. 54 1 (0 .0 81 ) (0 .0 89 ) (0 .2 37 ) (0 .3 80 ) d ro p in in co m e � 1. 13 9* ** � 2. 22 3* ** � 1. 86 1* ** � 2. 20 2* ** (0 .0 64 ) (0 .0 74 ) (0 .2 47 ) (0 .4 47 ) fl i 0. 65 2* ** 0. 30 8* * � 2. 10 0* ** � 0. 58 3 (0 .1 21 ) (0 .1 31 ) (0 .3 61 ) (0 .5 58 ) fb i 2. 22 8* ** 6. 69 9* ** 3. 27 7* ** 1. 99 1* ** (0 .1 69 ) (0 .1 86 ) (0 .5 00 ) (0 .7 69 ) n 16 ,0 83 (c on ti nu ed on ne xt pa ge ) 404 j.k. scott et al. / financial services review 27 (2018) 391-411 t ab le 4 (c on tin ue d) pa ne l b : r et ir em en t co nc er n (r ef : st ro ng ly di sa gr ee –1 ) 2 3 4 5 6 st ro ng ag re e7 d on ’t kn ow pr ef er no t to sa y (i nt er ce pt ) � 1. 13 6* ** � 0. 35 6* 2. 01 6* ** 1. 30 3* ** 1. 77 2* ** 2. 78 3* ** 1. 82 4* ** 0. 95 1 (0 .2 20 ) (0 .2 01 ) (0 .1 67 ) (0 .1 69 ) (0 .1 67 ) (0 .1 61 ) (0 .2 82 ) (1 .1 30 ) m al e � 0. 06 5 � 0. 10 4 � 0. 13 5* � 0. 20 9* ** � 0. 29 0* ** � 0. 45 5* ** � 0. 46 8* ** � 0. 50 0 (0 .0 94 ) (0 .0 88 ) (0 .0 77 ) (0 .0 77 ) (0 .0 77 ) (0 .0 74 ) (0 .1 55 ) (0 .6 32 ) a ge 35 –5 4 � 0. 25 6* * � 0. 24 7* * � 0. 1 � 0. 06 9 0. 21 5* * 0. 45 8* ** � 0. 05 7 2. 94 7* ** (0 .1 18 ) (0 .1 09 ) (0 .0 95 ) (0 .0 94 ) (0 .0 94 ) (0 .0 92 ) (0 .1 74 ) (0 .8 94 ) a ge 55 � 0. 32 5* * � 0. 54 3* ** � 0. 37 0* ** � 0. 63 0* ** � 0. 22 9* * 0. 00 4 � 0. 77 2* ** 1. 50 9 (0 .1 35 ) (0 .1 29 ) (0 .1 13 ) (0 .1 13 ) (0 .1 14 ) (0 .1 10 ) (0 .2 49 ) (1 .1 23 ) w hi te 0. 30 7* ** 0. 19 6* * 0. 18 8* * 0. 33 7* ** 0. 35 4* ** 0. 24 4* ** 0. 29 4* � 0. 50 5 (0 .1 03 ) (0 .0 95 ) (0 .0 82 ) (0 .0 82 ) (0 .0 82 ) (0 .0 79 ) (0 .1 57 ) (0 .5 61 ) so ut h � 0. 08 7 � 0. 01 5 � 0. 00 9 � 0. 09 8 � 0. 12 5 � 0. 08 7 � 0. 03 1 � 0. 85 8 (0 .0 97 ) (0 .0 91 ) (0 .0 79 ) (0 .0 79 ) (0 .0 79 ) (0 .0 76 ) (0 .1 53 ) (0 .6 25 ) d ep en de nt ch ild re n � 0. 22 5* * � 0. 16 3* � 0. 27 7* ** � 0. 16 2* 0. 02 7 0. 02 8 � 0. 45 7* ** � 1. 16 3* (0 .1 04 ) (0 .0 99 ) (0 .0 87 ) (0 .0 86 ) (0 .0 86 ) (0 .0 83 ) (0 .1 73 ) (0 .6 67 ) m ar ri ed 0. 22 4* 0. 35 9* ** 0. 32 4* ** 0. 29 6* ** 0. 18 2* 0. 29 6* ** � 0. 17 2 0. 02 3 (0 .1 24 ) (0 .1 17 ) (0 .1 01 ) (0 .1 00 ) (0 .1 00 ) (0 .0 97 ) (0 .1 91 ) (0 .6 77 ) sd w 0. 19 7 0. 11 6 0. 13 2 0. 03 9 � 0. 24 0* 0. 20 3 � 0. 15 2 � 0. 04 8 (0 .1 73 ) (0 .1 66 ) (0 .1 40 ) (0 .1 41 ) (0 .1 42 ) (0 .1 33 ) (0 .2 66 ) (0 .7 37 ) in co m e 50 –1 00 k 0. 20 6* 0. 03 � 0. 07 1 0. 13 1 0. 12 2 � 0. 11 6 � 0. 30 3* � 5. 39 2 (0 .1 23 ) (0 .1 13 ) (0 .0 96 ) (0 .0 97 ) (0 .0 96 ) (0 .0 93 ) (0 .1 79 ) (5 .8 68 ) in co m e � 10 0k 0. 11 1 � 0. 22 5* � 0. 47 6* ** � 0. 32 7* ** � 0. 40 4* ** � 0. 75 8* ** � 0. 88 1* ** � 0. 85 1 (0 .1 36 ) (0 .1 27 ) (0 .1 11 ) (0 .1 10 ) (0 .1 11 ) (0 .1 08 ) (0 .2 65 ) (1 .1 82 ) u ne m pl oy ed � 1. 19 7* ** � 0. 46 2* * � 0. 33 3* � 0. 63 4* ** � 0. 59 2* ** � 0. 36 3* * � 0. 08 4 � 0. 88 (0 .3 29 ) (0 .2 35 ) (0 .1 87 ) (0 .1 95 ) (0 .1 89 ) (0 .1 75 ) (0 .3 03 ) (1 .2 09 ) se lf em pl oy ed � 0. 41 6* ** � 0. 49 6* ** � 0. 42 2* ** � 0. 48 4* ** � 0. 53 8* ** � 0. 46 2* ** � 0. 46 9 � 2. 17 9 (0 .1 41 ) (0 .1 38 ) (0 .1 18 ) (0 .1 17 ) (0 .1 17 ) (0 .1 12 ) (0 .2 90 ) (2 .3 35 ) in ac tiv e � 0. 25 0* � 0. 27 0* * � 0. 09 7 � 0. 34 9* ** � 0. 47 3* ** � 0. 45 8* ** 0. 10 7 0. 52 9 (0 .1 32 ) (0 .1 22 ) (0 .1 03 ) (0 .1 04 ) (0 .1 05 ) (0 .1 00 ) (0 .1 76 ) (0 .5 89 ) d ro p in in co m e 0. 02 7 0. 05 2 0. 09 6 0. 39 1* ** 0. 77 7* ** 1. 25 6* ** 0. 30 7 0. 89 4 (0 .1 46 ) (0 .1 34 ) (0 .1 13 ) (0 .1 11 ) (0 .1 08 ) (0 .1 04 ) (0 .1 90 ) (0 .5 95 ) fl i 0. 83 7* ** 0. 92 5* ** 0. 03 1 0. 62 4* ** 0. 43 1* ** 0. 14 6 � 1. 35 3* ** � 3. 67 9* ** (0 .1 85 ) (0 .1 73 ) (0 .1 46 ) (0 .1 46 ) (0 .1 45 ) (0 .1 40 ) (0 .2 76 ) (1 .1 83 ) fb i 0. 65 6* ** 0. 12 9 � 1. 33 1* ** � 0. 90 5* ** � 1. 69 9* ** � 2. 66 9* ** � 3. 02 5* ** � 8. 75 6* ** (0 .2 40 ) (0 .2 22 ) (0 .1 90 ) (0 .1 89 ) (0 .1 88 ) (0 .1 83 ) (0 .3 61 ) (1 .7 56 ) n 16 ,0 83 (c on ti nu ed on ne xt pa ge ) 405j.k. scott et al. / financial services review 27 (2018) 391-411 t ab le 4 (c on tin ue d) pa ne l c : m or tg ag e la te pa ym en t (r ef : n ev er ) o nc e m or e th an o nc e d on ’t kn ow pr ef er no t to sa y (i nt er ce pt ) 0. 39 8* 1. 26 8* ** � 0. 28 � 0. 43 3 (0 .2 35 ) (0 .2 40 ) (0 .5 62 ) (0 .9 75 ) m al e 0. 24 4* * 0. 36 4* ** 0. 10 6 � 0. 43 4 (0 .1 02 ) (0 .1 04 ) (0 .2 71 ) (0 .5 47 ) a ge 35 –5 4 � 0. 52 0* ** � 0. 50 7* ** 0. 06 1 0. 00 8 (0 .1 11 ) (0 .1 13 ) (0 .2 98 ) (0 .5 65 ) a ge 55 � 0. 59 1* ** � 0. 53 2* ** � 0. 36 3 0. 36 2 (0 .1 76 ) (0 .1 78 ) (0 .4 68 ) (0 .7 51 ) w hi te � 0. 29 8* ** � 0. 45 4* ** � 0. 15 7 � 1. 38 4* ** (0 .1 08 ) (0 .1 08 ) (0 .2 84 ) (0 .4 88 ) so ut h 0. 10 6 0. 17 3* 0. 18 8 � 0. 26 2 (0 .1 04 ) (0 .1 04 ) (0 .2 66 ) (0 .5 09 ) d ep en de nt ch ild re n 0. 57 1* ** 0. 63 4* ** � 0. 36 � 0. 02 9 (0 .1 19 ) (0 .1 23 ) (0 .2 91 ) (0 .5 36 ) m ar ri ed � 0. 43 3* ** � 0. 14 9 � 0. 48 1 � 0. 53 (0 .1 39 ) (0 .1 48 ) (0 .3 43 ) (0 .6 17 ) sd w � 0. 56 7* ** � 0. 18 4 � 0. 33 6 � 0. 22 8 (0 .2 17 ) (0 .2 14 ) (0 .4 95 ) (0 .8 05 ) in co m e 50 –1 00 k � 0. 01 3 0. 13 2 0. 01 7 � 0. 26 6 (0 .1 23 ) (0 .1 2) 3 (0 .3 02 ) (0 .5 35 ) in co m e � 10 0k � 0. 20 6 � 0. 04 7 � 0. 19 7 � 0. 79 2 (0 .1 55 ) (0 .1 58 ) (0 .4 24 ) (0 .8 55 ) u ne m pl oy ed � 0. 13 1 0. 10 4 � 0. 06 9 0. 21 2 (0 .2 85 ) (0 .2 69 ) (0 .6 34 ) (1 .0 86 ) se lf em pl oy ed � 0. 24 5 0. 02 9 � 0. 33 1 � 0. 25 7 (0 .1 86 ) (0 .1 75 ) (0 .5 35 ) (1 .0 61 ) in ac tiv e � 0. 57 2* ** � 0. 47 3* ** � 0. 05 9 0. 14 2 (0 .1 54 ) (0 .1 50 ) (0 .3 39 ) (0 .5 73 ) d ro p in in co m e 1. 34 5* ** 1. 62 5* ** 1. 15 4* ** 0. 29 6 (0 .1 01 ) (0 .1 01 ) (0 .2 62 ) (0 .5 38 ) fl i � 1. 35 4* ** � 1. 59 7* ** � 2. 81 4* ** � 2. 95 9* ** (0 .1 90 ) (0 .1 92 ) (0 .4 94 ) (0 .9 13 ) fb i � 3. 00 3* ** � 5. 37 3* ** � 4. 01 6* ** � 3. 64 7* ** (0 .2 60 ) (0 .2 80 ) (0 .6 86 ) (1 .2 45 ) n 7, 36 5 n ot e: ** *p � 0. 01 ,* *p � 0. 05 ,* p � 0. 10 .r ob us t st an da rd er ro rs in pa re nt he se s. fl i � fi na nc ia l k no w le dg e in de x; fb i � fi na nc ia l b eh av io r in de x; sd w � si ng le , di vo rc ed , or w id ow ed . 406 j.k. scott et al. / financial services review 27 (2018) 391-411 t ab le 5 pa ne l a : a ge le ss th an 55 gr ea te r th an 35 2 3 4 5 6 7 d on ’t kn ow pr ef er no t to sa y (i nt er ce pt ) � 1. 62 9* ** � 1. 22 3* ** 1. 69 6* ** 0. 90 5* ** 1. 60 7* ** 2. 94 8* ** 1. 41 7* ** 1. 76 0* (0 .3 70 ) (0 .3 39 ) (0 .2 79 ) (0 .2 77 ) (0 .2 72 ) (0 .2 61 ) (0 .5 01 ) (1 .0 24 ) m al e 0. 16 9 0. 06 8 0. 12 7 � 0. 06 4 � 0. 17 1 � 0. 38 1* ** � 0. 12 4 � 0. 32 (0 .1 52 ) (0 .1 42 ) (0 .1 24 ) (0 .1 21 ) (0 .1 20 ) (0 .1 16 ) (0 .2 53 ) (0 .5 97 ) w hi te 0. 35 2* * 0. 31 5* * 0. 18 4 0. 45 6* ** 0. 39 8* ** 0. 41 2* ** 0. 66 1* * 0. 42 1 (0 .1 62 ) (0 .1 50 ) (0 .1 29 ) (0 .1 27 ) (0 .1 26 ) (0 .1 21 ) (0 .2 80 ) (0 .6 16 ) so ut h � 0. 06 0. 01 8 � 0. 11 4 � 0. 01 8 � 0. 02 3 � 0. 06 3 � 0. 06 6 � 0. 77 9 (0 .1 56 ) (0 .1 44 ) (0 .1 27 ) (0 .1 24 ) (0 .1 23 ) (0 .1 19 ) (0 .2 53 ) (0 .6 67 ) d ep en de nt ch ild re n � 0. 13 5 0. 11 3 � 0. 25 9* � 0. 07 0. 03 8 � 0. 08 7 � 0. 31 6 � 0. 61 5 (0 .1 60 ) (0 .1 52 ) (0 .1 33 ) (0 .1 30 ) (0 .1 30 ) (0 .1 25 ) (0 .2 69 ) (0 .6 07 ) m ar ri ed 0. 28 2 0. 39 1* * 0. 37 9* * 0. 33 8* * 0. 33 2* * 0. 37 1* * � 0. 3 0. 17 (0 .2 07 ) (0 .1 98 ) (0 .1 66 ) (0 .1 63 ) (0 .1 62 ) (0 .1 57 ) (0 .3 20 ) (0 .7 15 ) sd w 0. 36 2 0. 17 1 � 0. 03 4 � 0. 15 8 � 0. 23 5 0. 13 � 0. 27 4 0. 00 4 (0 .2 59 ) (0 .2 51 ) (0 .2 13 ) (0 .2 11 ) (0 .2 09 ) (0 .1 96 ) (0 .3 77 ) (0 .7 71 ) in co m e 50 –1 00 k 0. 14 6 0. 01 7 � 0. 01 6 0. 16 7 0. 12 2 � 0. 05 7 � 0. 05 � 1. 36 7* (0 .2 16 ) (0 .1 96 ) (0 .1 65 ) (0 .1 64 ) (0 .1 61 ) (0 .1 54 ) (0 .3 03 ) (0 .8 31 ) in co m e � 10 0k 0. 02 1 � 0. 28 9 � 0. 53 4* ** � 0. 38 0* * � 0. 53 8* ** � 0. 78 6* ** � 0. 34 1 � 0. 84 8 (0 .2 30 ) (0 .2 11 ) (0 .1 82 ) (0 .1 79 ) (0 .1 77 ) (0 .1 71 ) (0 .3 83 ) (0 .9 07 ) u ne m pl oy ed � 1. 03 4* * � 0. 57 3 � 0. 37 8 � 0. 64 3* * � 0. 58 3* � 0. 12 1 0. 36 8 � 0. 49 9 (0 .5 21 ) (0 .4 06 ) (0 .3 15 ) (0 .3 20 ) (0 .3 06 ) (0 .2 79 ) (0 .4 63 ) (1 .1 11 ) se lf em pl oy ed 0. 00 6 � 0. 45 7* � 0. 19 4 � 0. 18 7 � 0. 42 6* * � 0. 27 � 1. 01 9 � 0. 59 3 (0 .2 29 ) (0 .2 40 ) (0 .2 00 ) (0 .1 92 ) (0 .1 95 ) (0 .1 85 ) (0 .6 23 ) (1 .0 74 ) in ac tiv e � 0. 49 9* * � 0. 09 8 0. 07 7 � 0. 39 8* * � 0. 45 7* ** � 0. 30 0* 0. 40 7 � 0. 38 8 (0 .2 50 ) (0 .2 07 ) (0 .1 76 ) (0 .1 79 ) (0 .1 76 ) (0 .1 67 ) (0 .3 02 ) (0 .7 11 ) d ro p in in co m e � 0. 04 7 0. 03 7 � 0. 08 9 0. 14 3 0. 45 9* ** 0. 99 9* ** 0. 03 5 0. 55 6 (0 .2 32 ) (0 .2 08 ) (0 .1 79 ) (0 .1 73 ) (0 .1 67 ) (0 .1 60 ) (0 .3 27 ) (0 .6 24 ) fl i 1. 01 4* ** 1. 12 6* ** 0. 24 9 0. 97 6* ** 0. 81 9* ** 0. 41 4* � 1. 28 6* ** � 1. 03 9 (0 .2 94 ) (0 .2 72 ) (0 .2 28 ) (0 .2 26 ) (0 .2 22 ) (0 .2 13 ) (0 .4 36 ) (1 .0 21 ) fb i 0. 53 4 0. 29 3 � 1. 43 8* ** � 1. 09 8* ** � 1. 74 0* ** � 2. 74 2* ** � 3. 60 6* ** � 6. 91 7* ** (0 .3 74 ) (0 .3 45 ) (0 .2 97 ) (0 .2 91 ) (0 .2 88 ) (0 .2 80 ) (0 .6 00 ) (1 .5 43 ) n 7, 11 9 (c on ti nu ed on ne xt pa ge ) 407j.k. scott et al. / financial services review 27 (2018) 391-411 t ab le 5 (c on tin ue d) pa ne l b : g re at er th an 55 2 3 4 5 6 7 d on ’t kn ow pr ef er no t to sa y (i nt er ce pt ) � 1. 87 8* ** 0. 08 7 2. 83 3* ** 0. 96 3* * 2. 52 9* ** 4. 05 4* ** 2. 03 7* * � 2. 44 (0 .5 46 ) (0 .5 04 ) (0 .4 12 ) (0 .4 37 ) (0 .4 21 ) (0 .4 02 ) (0 .8 85 ) (5 .3 81 ) m al e � 0. 43 8* * � 0. 53 3* ** � 0. 45 3* ** � 0. 52 5* ** � 0. 67 1* ** � 0. 69 2* ** � 0. 43 4 � 0. 16 (0 .1 75 ) (0 .1 75 ) (0 .1 54 ) (0 .1 56 ) (0 .1 56 ) (0 .1 52 ) (0 .4 20 ) (1 .2 49 ) w hi te 0. 43 1* * 0. 37 4* 0. 30 7* 0. 68 5* ** 0. 73 0* ** 0. 44 2* * 0. 34 5 � 0. 66 2 (0 .2 14 ) (0 .2 10 ) (0 .1 76 ) (0 .1 91 ) (0 .1 89 ) (0 .1 74 ) (0 .4 70 ) (1 .1 13 ) so ut h 0. 03 7 � 0. 17 2 0. 05 5 � 0. 11 9 � 0. 23 5 � 0. 15 5 � 0. 10 5 � 0. 92 1 (0 .1 74 ) (0 .1 79 ) (0 .1 53 ) (0 .1 58 ) (0 .1 59 ) (0 .1 53 ) (0 .4 15 ) (1 .2 57 ) d ep en de nt ch ild re n � 0. 18 2 0. 03 8 � 0. 03 9 � 0. 01 2 0. 27 9 0. 37 4* * � 0. 4 � 4. 40 8 (0 .2 11 ) (0 .2 07 ) (0 .1 83 ) (0 .1 87 ) (0 .1 83 ) (0 .1 76 ) (0 .5 67 ) (7 .6 39 ) m ar ri ed 0. 49 5* 0. 55 5* * 0. 25 2 0. 36 3 0. 06 6 0. 32 1 0. 53 7 3. 89 7 (0 .2 79 ) (0 .2 73 ) (0 .2 25 ) (0 .2 31 ) (0 .2 23 ) (0 .2 21 ) (0 .5 84 ) (5 .0 71 ) sd w 0. 32 9 0. 16 3 0. 16 5 0. 17 5 � 0. 38 9 0. 15 1 0. 32 3 3. 27 7 (0 .3 09 ) (0 .3 04 ) (0 .2 45 ) (0 .2 54 ) (0 .2 49 ) (0 .2 39 ) (0 .5 94 ) (5 .1 31 ) in co m e 50 –1 00 k 0. 28 4 0. 03 6 � 0. 08 4 � 0. 07 7 � 0. 04 � 0. 39 8* * � 1. 34 9* * � 4. 65 9 (0 .2 47 ) (0 .2 33 ) (0 .1 97 ) (0 .2 01 ) (0 .1 99 ) (0 .1 92 ) (0 .5 54 ) (8 .3 14 ) in co m e � 10 0k 0. 22 1 � 0. 22 2 � 0. 30 1 � 0. 56 7* * � 0. 61 4* ** � 0. 78 4* ** � 1. 48 8* * 0. 15 (0 .2 64 ) (0 .2 56 ) (0 .2 21 ) (0 .2 26 ) (0 .2 27 ) (0 .2 19 ) (0 .7 07 ) (1 .4 26 ) u ne m pl oy ed � 1. 23 2* � 0. 59 8 � 0. 25 8 � 0. 50 7 � 0. 52 9 � 0. 16 4 � 0. 94 9 � 2. 02 8 (0 .6 64 ) (0 .5 01 ) (0 .3 76 ) (0 .3 95 ) (0 .3 83 ) (0 .3 52 ) (1 .0 95 ) (8 .2 76 ) se lf em pl oy ed � 0. 39 9* � 0. 35 4* � 0. 38 6* * � 0. 46 8* * � 0. 43 2* * � 0. 41 6* * 0. 03 3 � 2. 62 8 (0 .2 13 ) (0 .2 11 ) (0 .1 89 ) (0 .1 91 ) (0 .1 90 ) (0 .1 86 ) (0 .5 46 ) (7 .9 17 ) in ac tiv e � 0. 22 7 � 0. 87 8* ** � 0. 34 9 � 0. 59 9* ** � 0. 90 6* ** � 0. 69 4* ** � 0. 14 7 1. 50 0 (0 .2 62 ) (0 .2 83 ) (0 .2 18 ) (0 .2 30 ) (0 .2 32 ) (0 .2 16 ) (0 .4 75 ) (1 .3 34 ) d ro p in in co m e � 0. 17 4 � 0. 33 4 � 0. 02 2 0. 45 1* 0. 73 2* ** 1. 33 6* ** 0. 04 5 0. 91 4 (0 .3 05 ) (0 .3 05 ) (0 .2 42 ) (0 .2 37 ) (0 .2 30 ) (0 .2 18 ) (0 .5 81 ) (1 .2 79 ) fl i 1. 00 0* ** 1. 07 4* ** � 0. 24 1 0. 72 0* * 0. 58 3* 0. 33 � 1. 71 3* * � 4. 36 0* * (0 .3 69 ) (0 .3 66 ) (0 .2 98 ) (0 .3 14 ) (0 .3 11 ) (0 .2 96 ) (0 .7 24 ) (2 .2 13 ) fb i 0. 71 2 � 1. 21 3* * � 2. 52 1* ** � 1. 37 6* ** � 2. 90 1* ** � 4. 53 1* ** � 4. 11 8* ** � 3. 59 7 (0 .5 19 ) (0 .4 81 ) (0 .4 13 ) (0 .4 27 ) (0 .4 18 ) (0 .4 05 ) (0 .9 88 ) (2 .7 89 ) n 3, 58 3 n ot e: ** *p � 0. 01 ,* *p � 0. 05 ,* p � 0. 10 .f l i � fi na nc ia l k no w le dg e in de x; fb i � fi na nc ia l b eh av io r in de x; sd w � si ng le ,d iv or ce d, or w id ow ed . 408 j.k. scott et al. / financial services review 27 (2018) 391-411 knowledge that someone might possess, they will experience some anxiety or distress associated with retirement planning. we find these results to be fascinating. it aligns with the perceived planned behavior theory that knowledge can help individuals perceive a higher level of control. however, retirement planning is very complicated and requires time, understanding of financial markets and products, time value of money, and some level of professional guidance. retirement can be overwhelming to nonprofessionals even when they have some knowledge on the topic. furthermore, many individuals, especially baby boomers, grew up seeing their parents or guardians pay bills. however, planning for retirement might not have been a topic discussed at home, as 30 years ago many families had pensions. we find our results noteworthy. financial knowledge and behavior reduce financial distress when it comes to paying bills. however, regardless of financial knowledge, many individuals have concerns about retirement preparedness. this is understandable as we cannot predict the future; there is a natural sense of uncertainty. 6. conclusion this study contributes to the literature by exploring whether financial literacy has any association with financial distress, even after controlling for financial behavior. it is quite common to assume that good (bad) financial behavior will be associated with less (more) financial distress or concerns. the results are as expected. as mentioned in previous studies, financial literacy and financial knowledge can be distinct concepts, and higher financial literacy does not necessarily translate into better financial behavior; implementation in real life can be quite complex (e.g., retirement planning). however, when the decision is not very complicated, we see that the impact of financial literacy can be larger and easier to observe (e.g., understanding that late payments are not good for credit cards and mortgages). more specifically, using data from the 2015 national financial capability study, and the survey questionnaire designed by atkinson and messy (2011), we use three measures of financial distress to develop financial literacy and financial behavior indices. these measures include having difficulty paying bills, worrying about retirement funding, and being late with mortgage payments. from our results, individuals with dependent children have a higher probability of suffering financial distress. respondents within the age 55� range are less likely to suffer from financial difficulties. married individuals are also less likely to become financially distressed. practicing wise financial behaviors such as saving for emergency funds, having health insurance, and paying credit card balances in full, highly reduces the probability of experiencing financial distress. individuals with higher levels of financial literacy, who engage in positive financial behavior such as paying bills on time, are less likely to suffer from financial distress. however, regardless of the level of financial knowledge and behavior, evidence of financial distress remains for retirement planning. arguably, this can be attributed to the complexity and uncertainty associated with planning and being prepared for retirement. highlighting the limitations of our study is worthwhile. in some cases, financial decisionmaking or behavior is also affected by innate human capital characteristics, for example, iq and time preference, alongside parental and peer influence. if the dataset allowed for the 409j.k. scott et al. / financial services review 27 (2018) 391-411 inclusion of those variables, we might see different results. also, while we include income and a reduction in income as control variables, we did not control for net worth as the nfcs does not allow for the operationalization of this variable. therefore, future research would benefit from including the variables mentioned above when studying this topic. additionally, policy implications from our research are evident; particularly, strategies and programs aimed at promoting financial literacy (including behavior) and at reducing financial distress. specifically, when designing these programs, it may be worthwhile to focus not only on shaping personal financial knowledge but also on the use of knowledge to manage financial resources and expectations effectively. for example, past financial behavior may dictate the level of education required (in moderation) as well as behavioral strategies that might be more impactful. our study, therefore, further prompts investigation and continued development of “just-in-time” personal finance programs to help individuals who are more prone to fall into financial distress, especially during key wealth-building life stages such as retirement planning. references altintas, k. m. (2011). the dynamics of financial literacy within the framework of personal finance: an analysis among turkish university students. african journal of business management, 5, 10483–10491. allgood, s., & walstad, w. b. (2016). the effects of perceived and actual financial literacy on financial behaviors. economic inquiry, 54, 675–697. anz-retirement commission. (2009). financial knowledge survey (available at http://www.financialliteracy. org.nz/sites/default/files/webfm/research/r-fksurvey-2009.pdf). asaad, c. t. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101–117. atkinson, a., & messy, f. a. (2011). assessing financial literacy in 12 countries: an oecd/infe international pilot exercise. journal of pension economics & finance, 10, 657–665. atkinson, a., & messy, f. a. (2012). measuring financial literacy: results of the oecd. paris: oecd publishing. banco de portugal. (2011). survey on the financial literacy of the portuguese population (2010) – final report. lisbon, portugal. browning, m., & lusardi, a. (1996). household saving: micro theories and micro facts. journal of economic literature, 34, 1797–1855. campbell, j. (2006) household finance. journal of finance, 61, 1553–1604. chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11, 289–308. disney, r., bridges, s., & gathergood, j. (2008). drivers of over-indebtedness. centre for policy evaluation, university of nottingham. fernandes, d., lynch jr., j. g., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors. management science, 60, 1861–1883. finra. (2009). national financial capability study (available at http://www.usfinancialcapability.org/ results.php?region�us). finra. (2015). national financial capability study (available at http://www.usfinancialcapability.org/ results.php?region�us). fram, m. s., miller-cribbs, j. e., & van horn, l. (2007). poverty, race, and the contexts of achievement: examining educational experiences of children in the us south. social work, 52, 309–319. fsa. (2006). levels of financial capability in the uk: results of a baseline survey. financial services authority, university of bristol. 410 j.k. scott et al. / financial services review 27 (2018) 391-411 garman, e. t., sorhaindo, b., bailey, w., kim, j., & xiao, j. (2004). financially distressed credit counseling clients and the incharge financial distress/financial well-being scale. in j. fox (ed.), proceedings of the eastern regional family economics and resource management association conference (pp. 71–81). tampa, fl. garman, e. t., & sorhaindo, b. (2005). delphi study of experts’ rankings of personal finance concepts important in the development of the incharge financial distress/financial well-being scale. consumer interests annual, 51, 184–194. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309. hung, a. a., parker, a. m., & yoong, j. k. (2009). defining and measuring financial literacy. rand labor and population working paper wr-708 (pp. 1–27). santa monica, ca: rand corporation (available at https:// www.rand.org/pubs/working_papers/wr708.html). huston, s. j. (2009). measuring financial literacy. journal of consumer affairs, 44, 296–316. johnson, e., & sherraden, m. s. (2007). from financial literacy to financial capability among youth. journal of society & social welfare, 34, 119. lee, j. m., & hanna, s. d. (2015). savings goals and saving behavior from a perspective of maslow’s hierarchy of needs. journal of financial counseling and planning, 26, 129–147. lusardi, a., & mitchell, o. s. (2007). financial literacy and retirement preparedness: evidence and implications for financial education. business economics, 42, 35–44. mccarthy, y. (2011). behavioral characteristics and financial distress. ecb working paper 1303, european central bank, frankfurt, germany (available at http://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp1303.pdf). mckay, s. (2011). understanding financial capability in canada: analysis of the canadian financial capability survey. research paper prepared for task force on financial literacy (available at http://publications. gc.ca/collections/collection_2011/fin/f2-213-2011-eng.pdf). moore, d. l. (2003). survey of financial literacy in washington state: knowledge, behavior, attitudes, and experiences. olympia, wa: washington state dept. of financial institutions. keeney, m. j., & o’donnell, n. (2009). financial capability: new evidence for ireland (no. 1/rt/09). dublin, ireland: central bank of ireland. o’connell, a. (2007). how effective is financial education? policy quarterly, 3, 40–46. orc international. (2015). stress awareness month survey report. new york, ny (available at https:// www.cfp.net/news-events/research-facts-figures/cfp-board-research/consumer-surveys). prawitz, a., kim, j., & garman, e. t. (2006). financial distress: definition, effects and measurement. consumer interests annual, 52, 489–496. president’s advisory council on financial literacy. (2009). 2008 annual report to the president of the united states of america (available at http://www.treas.gov). remund, d. l. (2010). financial literacy explicated: the case for a clearer definition in an increasingly complex economy. journal of consumer affairs, 44, 276–295. van rooij, m. c., lusardi, a., & alessie, r. j. (2011). financial literacy and retirement planning in the netherlands. journal of economic psychology, 32, 593–608. ware, s. j. (2015). debt, poverty, and personal financial distress. american bankruptcy law journal, 89, 493. zarcadoolas, c., pleasant, a., & greer, d. s. (2009). advancing health literacy: a framework for understanding and action. san francisco, ca: john wiley & sons. 411j.k. scott et al. / financial services review 27 (2018) 391-411 pii: 1057-0810(95)90011-x volume 4 number 1 1995 financial services review the journal of individual financial management editor lewis mandell university of connecticut managing editor barbarapoole university of connecticut associate editors larry a. cox university of georgia mona j. gardner illinois wesleyan university benton e. gup university of alabama jean louis heck villanova university david s. kidwell university of minnesota robert w. mcleod university of alabama neil b. murphy virginia commonwealth university phyllis s. myers virginia commonwealth university george c. philippatos university of tennessee chris j. prestopino california state university chico s. travis pritchett universiry of south carolina william reichenstein baylor university frank k. reilly university of notre dame arthur l. schwartz university of south florida peter l. struck washington mutual savings bank walt woerheide rochester institute of technology jai press inc. greenwich, connecticut london, engtbnd pii: s1057-0810(97)90010-x financial services review, 6(3): 155-167 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. incorporating historical investment performance in projecting life insurance cash values s. travis pritchett the projection of current investment experience in life insurance sales illustrations dur ing the historically high interest rate environment of the 1980s in the u. s. has led to consumer dissatisfaction and lawsuits against life insurers and agents. as interest rates fell after 1985, insurers were unable to credit returns near the maximums illustrated earlier. new regulations still allow projections at essentially the latest current invest ment returns, along with showing guaranteed and intermediate values. the question raised in this article is: can history help the financial planner determine a range of credible investment return assumptions for projecting cash values? conclusions are based on the results of basing projections on historical investment experience. i. introduction in recent years, u.s. life insurers have used sales illustrations to emphasize growth in cash values and death benefits. rapid growth was fueled in the 1980s primarily by the com pounding of relatively high current investment returns (i.e., essentially what an insurer was earning on a certain portfolio at the time of the sale.) current experience continues to be extrapolated over the potential life of policies, often up to the insured’s age 100. the tre mendous effect of compound interest at relatively high rates over extensive periods is well known in financial sectors but may not be appreciated by non-financially-sophisticated consumers. in the context of cash value projections, insufficient attention has been given by actuaries, agents, and insurer management to what history has to say about the volatility of market returns and the resulting likelihood of high current returns being maintained throughout the potential life of policies. consequently, financial planners must be prepared to give objective advice on sales illustrations when they include unrealistic projections for their clients. in other instances, the planner may be the person presenting the original pro jection, for example, on a no-load life insurance product. (while mortality, expenses, and lapse rates also affect projections, this article is limited to the influence of investment returns alone.) s. travis pritchett l w. frank hipp professor of insurance, college of business administration, university of south carolina, columbia, sc, 29208; e-mail: pritch@darla.badm.sc.edu. 156 financial services review 6(3) 1998 the national association of insurance commissioners (naic) recently promulgated a model life insurance sales illustration bill that, in the states where enacted into law, can be expected to make illustrations more sustainable (i.e., “self-supporting” in the language of the model) and understandable. yet, essentially the latest current investment returns, including those developed through the investment generation method, can still be illus trated. the concern of this author is that in periods of relatively high investment returns cur rent returns will be above even reasonably optimistic projections of long-run performance. between the end of war ii and 1985 when non-equity investment returns followed an upwardly sloping trend, insurer performance exceeded consumer expectations created dur ing typical sales interviews. in the post1985 environment of lower returns on bonds, mort gages and other non-equity investments, it has become important to seek new ways to set client expectations that will result in long-run consumer satisfaction rather than the dissat isfaction being expressed in the 1990s. the illustration of returns consistent with historical experience deserves consideration. the question raised in this article is: is it feasible for a financial planner to use history in determining a range of credible investment return assumptions for life insurance sales illus trations? the scope of the article is limited to fixed-dollar products such as traditional par ticipating whole life insurance and interest-sensitive policies with non-guaranteed elements. after providing more background, a method of incorporating historical investment returns into the projection of a range of future returns is described. the remainder of the article presents the results of using historical returns to project ranges of future rates of return over 50 years. the article ends with conclusions. ii. background around 1980, life insurers were illustrating long-term universal life performance with investment assumptions as high as 14 percent, the current crediting rate for the insurers involved. the only other illustrated values, in some illustrations, were the much lower guarantees. also, it was not uncommon to illustrate relatively low level premiums that required the continuation of double digit investment returns to keep the contract in force to age 100. the ex-post result has been contracts that, with 1990s performance, will terminate before the insured reaches age 70 or so unless substantial additional premiums are paid. the author contends that the problem is caused primarily by two industry practices. the major problem is the use of short-term assumptions (i.e., “current investment returns”) in long-term projections. the “investment generation method” further exaggerates the overstatements that can result from current investment assumptions. the second problem is the “black box” technique of marketing participating cash-value life insurance. the con sumer is presented an illustration without any disclosure of underlying assumptions. thus, assumptions that are unreasonable for the long run are not easily detectable even by finan cially sophisticated consumers. at least, universal life reveals gross investment returns. of course, gross returns differ substantially from net returns (pritchett, 1998). a current investment return can be defined generally as the annual rate of investment return being realized currently by an insurer. insurers have flexibility in determining the rate of return in several important ways. for example, the base of the ratio varies among insurers (e.g., interest-bearing liabilities plus surplus versus total admitted assets); perfor historical investment performance 157 mance of the entire general account portfolio is utilized by some compared to performance of a segregated portion of the portfolio being used by others; investment expenses are allo cated differently; realized capital gains and losses are added to investment income by some each year while others use various smoothing techniques; forecasted trends in investment performance, if acknowledged at all, are recognized with different degrees of optimism or pessimism; and what competitors are paying is matched to different degrees (miller, 1996). the traditional approach in determining policyowner dividends for participating busi ness is to determine the current net rate of return for the entire general account. a general account being an undivided investment account that supports an insurer’s guaranteed, fixed-dollar products such as ordinary life insurance, universal life insurance, and fixed dollar annuities. then, this “average portfolio rate” is used, along with other actuarial fac tors, to calculate the current dividend scale. the current dividend scale is used to determine the maximum values shown in sales illustrations. insurers usually use their best judgement in setting a dividend scale that can remain unchanged for several years. the length of time over which the average reflects historical returns is influenced by asset maturities, portfolio turnover, policy lapse rates, growth in sales and reserves, and other factors. for a typical insurer the average is likely to be weighted by returns on relatively recently purchased assets. yet, several years of history are recognized. thirty seven percent of life insurers responding to a recent survey reported using the investment generation method (also called the “investment year” and “new money” meth ods), or some modification thereof, in determining dividend or other crediting rates for at least some plans. the predominant use is with universal and other interest-sensitive con tracts. the method segregates policies into different cells (i.e., generations) associated with assets purchased during specific years (i.e., generations). a different rate of return is cred ited on funds associated with each year assets are generated by a policy. portfolio turnover and new purchases cause a policy’s return associated with a particular year to change over time. the major problem is that new contracts may be accompanied by sales illustrations where the maximum values reflect investment returns for new investments alone. histori cal returns, for anything more than a few months or a year, may have been ignored in the past. essentially peak returns were compounded over the entire potential life of a policy. new regulation and guidelines requiring consideration of approximately two years of experience represent little change. relative to the portfolio method, the investment generation method allows higher illustrated returns when returns are following an upward trend. likewise, the portfolio method tends to be favored when rates trend downward. with both the portfolio and invest ment generation methods in the 1980s agents of some insurers were allowed to adjust assumptions, sometimes illustrating returns greater than current insurer experience (miller, 1996). the life insurance illustrations model regulation passed by the naic in december 1995 changes, but only modestly in the author’s opinion, the way in which current invest ment returns are used in illustrations. for example, the “currently payable scale” is defined as “. .a scale of non-guaranteed elements in effect for a policy form as of the preparation date of the illustration or declared to become effective within the next ninety-five (95) days.” further, the: 158 financial services review 6(3) 1998 disciplined current scale means a scale of non-guaranteed elements constituting a limit on illustrations currently being illustrated by an insurer that is reasonably based on actual recent historical experience, as certified annually by an illustration actuary des ignated by the insurer. further guidance in determining the disciplined current scale as contained in standards established by the actuarial standards board may be relied upon if the standards: (1) are consistent with all provisions of this regulation; (2) limit a disciplined current scale to reflect only actions that have already been taken or events that have already occurred; (3) do not permit a disciplined current scale to include any projected trends of improvements in experience or any assumed improvements in expe rience beyond the illustration date.. .(emphasis added) (naic, 1996, p. 582-2). also, an illustrated scale means a scale of non-guaranteed elements currently being illustrated that is not more favorable to the policy owner than the lesser of: (1) the disciplined cur rent scale or (2) the currently payable scale (naic, 1996, p. 582-2). any interest rate shown “shall not be greater than the earned interest rate underlying the disciplined current scale” (naic, 1996, p. 582-2). a statement similar to the following is placed in a narrative summary: this illustration assumes that the currently illustrated nonguaranteed elements will con tinue unchanged for all years shown. this is not likely to occur, and actual results may be more or less favorable than those shown (naic, 1996, p. 582-7). the policyowner also signs a statement regarding the non-guaranteed nature of certain elements. values are illustrated using (1) the insurer’s illustrated scale, (2) guarantees, and (3) at 50 percent of illustrated dividends or the average of guaranteed and illustrated rates. the guidelines developed by the actuarial standards board essentially repeat the def initions quoted above and add the following one: recent historical experience: mortality, persistency, interest, and expense experience on an experience factor class that is current, determinable, and credible (actuarial stan dards board, 1995, p. 2). further standards regarding investment returns specify that recent historical experi ence refers to rates being determined “on an entirely retrospective basis considering only assets to be supporting the block” (actuarial standards board, 1995, p. 7). when an insurer has inadequate experience, it can use appropriate experience and trends from other similar classes of its business, experience of another company, or experience from other sources, preferably in that order. (the securities and exchange commission recently started allowing mutual funds to show historical performance that pre-dates their formation as mutual funds. funds are being allowed to publish the records of older “substantially similar” investment mechanisms (e.g., a separate account) that con vert into mutual funds (mcgough, 1997). improving trends in returns cannot be pro jected to continue improving, however, significantly deteriorating trends or expected deteriorating trends, ” . ..between the date of the historical experience and the effective date of the scale underlying the illustration.. .,” should be recognized (actuarial stan dards board, 1995, p. 101). the assumed investment rate is to be fixed for all illus trated durations. either the portfolio average approach or the investment generation approach is acceptable. the recognition of realized and unrealized capital gains are not historical investment performance 159 2.0% i ,,,i, 194; i9ii i9’631 i i966 ’ is& i / 1957 1981 & : i 1 / i i / / / / i. 1993 1999 years s0ukty various issues of the life insurance fact book. figure 1. before-tax rate of investment income for u.s. life insurers: 1945-1994 i bs constrained other than by needing to be consistent with the practice of each insurer (actuarial standards board, 1995, pp. 7-10). while the model illustration bill and accompanying actuarial standards vastly improve overall illustration practices, the author is of the opinion that regulations still allow placing too much emphasis on the extrapolation of current investment performance as produced by recent trends, whether up, down, or level. allegedly, the recent experience requirement may only result in looking at approximately two years of past experience. if this material izes, insurer practice will continue to be in sharp contrast to practices in the u.s. securities industry that relies primarily on the disclosure of historical performance over various peri ods of time. any prospective illustrations of securities use the same investment assump tions for competing instruments. the absurdity of creating a current investment assumption by only looking back approximately two years when returns seem high, and then projecting this assumption into the future for all illustrated durations is illustrated in figure 1. the rates of return plotted there use aggregates for the universe of u.s. life insurer general accounts as reported by the american council of life insurance (acli). had the figure gone back farther in time, the return for 1930 would have been 5.05 percent with a decreasing trend between than and the low of 2.88 percent shown for 1947 (acli, 1994, p. 74). in the fig ure, rates of 9.65 percent and 9.87 percent for 1984 and 1985, respectively, form the basis of a “current” hypothetical early 1986 projection (i.e., based on portfolio average rates for the industry.) the dashed line representing current returns is obviously well above subse quent industry performance to date. the use of investment generation rates for 1984 and 1985 would have produced projections that were even more inconsistent with long-run historical experience. 160 financial services review 6(3) 1998 iii. research methods returns are forecasted for a hypothetical general account that uses an average portfo lio rate of total returns for dividend allocations and other crediting (e.g., for a universal life contract) in projections. the simulation employs current market data in addition to histor ical performance for specific asset classes. total returns include all capital gains and losses, and, therefore, increase the variability associated with the hypothetical general account forecast, relative to the net income returns used in practice by most insurers. (a group of u.s. life insurers is experimenting with total returns as the prime measure of investment performance for general accounts, viewing this as a potentially better measure of economic performance than the current net income approach (burgess, et. al., 1994). the simulations apply the arithmetic mean and standard deviation of the annual data to project expected compound annual returns over a 50-year time horizon (i.e., an assumed insurance policy holding period), using a lognormal distribution. the lognormal distribu tion has a random variable whose logarithm is normally distributed. its use is consistent with the positive distribution of historical market returns for most assets and the inability of an insurer to lose more than its investment (i.e., returns cannot fall below negative 100 percent). because the distribution is specified by the historical mean and standard devia tion the necessary parameters are practically available. the theoretical distribution is pos itively skewed, and skewness increases as variance increases (for further discussion of the lognormal distribution see aitchison & brown, 1957, and shimizu, 1988). by separating inputs into a historical risk premium and a current risk-free rate, pro jected future risk/return relationships reflect what the market itself is currently forecasting. (current market conditions are recognized through the current risk free rate that may differ substantially from the historical risk-free rate for an asset class (lucas, 1995). the result is emphasis on current returns as a starting point in the projection of future mean returns. in addition to using the mean and standard deviation as inputs, a forecast for a portfolio requires estimation of the correlation of each asset in the portfolio with every other. the first step in using the lognormal mode is calculation of the expected value (m) and standard deviation (s) of the natural logarithm (in) of the portfolio’s return relative. the arithmetic mean return (p) and standard deviation (cr) of the portfolio are used as follows (using historical data. .., 1995): m = ln(l+l,t)g ( > and s = /ln(l+ (57) ranges of simulated rates of return and the statistical confidence in the ranges are cal culated by using z-scores to specify percentiles in the forecast. thus, a 68 percent confi dence level deviates one standard deviation on each side of the mean return. likewise, the 5th and 95’h percentiles would be the lower and upper bounds providing a 90 percent con fidence level for future returns. the positive skewness of the lognormal distribution results in the mean (or expected value) being greater than the median. further, one standard devi hktorical investment performance 161 ation on each side of the mean will differ from exactly 34 percent because the probabilities depend on the parameters of the distribution (using historical data.. ., 1995). the projections in this study were calculated by the portfolio strategist program developed and distributed by ibbotson associates inc., chicago, illinois. the program pro duces efficient portfolios using the markowitz technique. it also projects ranges of future returns. only the projection part of the program is used in this study. the projections can be either for optimal or non-optimal portfolios. the forecasts in this article are for non optimal portfolios specified by the author. the hypothetical general account portfolio is constructed for this study by weighting certain assets classes (e.g., lehman brothers corporate bond index and the standard & poor’s 500 average) as reflected in ibbotson’s optimizer inputs and using them as proxies for segments of an insurer’s general account. industry averages as of the end of 1993 are used to determine portfolio weights for the hypothetical general account. this is done by building the hypothetical portfolio from series of long-term government bonds, long-term corporate bonds, treasury bills, and so forth. when an entire history (such as that for large company common stocks beginning in 1926) is relevant to projecting future returns, statistical parameters are based on the entire available history of returns. other asset classes use shorter histories of monthly returns in the calculation of annualized expected values and standard deviations. for example bond and other fixed income markets experienced a structural change in the 1970s due to the u.s. federal reserve system trying to manage the money supply. consequently, data for shorter historical periods are used for these portions of a portfolio. an implicit assumption of the methodology is that future investment results can be forecasted by historical performance. the movement of financial markets is influenced by many factors, including economics, inflation, monetary and fiscal policy, politics, and glo bal factors. the future will not be exactly like the past and the difference may not be fore seeable. consequently, a mathematical model cannot forecast, for example, future structural changes in financial markets. john maynard keynes expressed this by saying that human behavior is not “homogeneous through time.” yet the past is probably the best guide we have to the future, especially in projecting the range of returns over the long run. iv. results at the end of 1993, general accounts for the universe of u.s. life insurers consisted of the following major asset classes and weights: (1) corporate bonds, 41.74 percent; (2) govem ment securities, 25.17 percent; (3) mortgages, 14.76 percent; (4) policy loans, 5.09 percent; (5) corporate stocks, 5.07 percent; (6) real estate, 2.85 percent; and (7) miscellaneous assets, 5.32 percent. an approximation of the history for this “1993 industry portfolio” was created with total market returns for the proxies shown in table 1. (the ibbotson associ ates portfolio strategist software allows the specifications of one-, five-, or 20-year holding periods for investments. in the absence of alternatives between five and 20 years, a 20-year period was specified for the simulations in the current study. the holding period influences the riskless rate that is used in separating the total return for a class of assets into riskless and risk premia components. the current yield for a zero coupon bond with a maturity matching the chosen holding period becomes the riskless rate.) 162 financial services review 6(3) 1998 table 1. proxies for asset segments held by u.s. life insurers at the end of 1993 portfolio segment (s) pldxy weight % corporate bonds lehman brothers corporate bonds 41.7 government securities lehman brothers government bonds 25.2 mortgages & real estate ibbotson assoc. business real estate 17.6 corporate stocks standard & poor’s 500 5.1 policy loans 90.day treasury bills 5.1 miscellaneous assets 30-day treasury bills 5.3 note: for each proxy the measure of performance is total annual returns. historical holding periods for the proxies are: lehman brothers corporate and government bonds, 197351994; businew real estate. 1978-1994; s & p soo, 1926-1994; 90. day treasury bills, 1978-1994: and 30-day treasury bills, 1969-1994 statistics for u.s. life insurers at the end of 1993 show that less than one percent of corporate securities (bonds and stocks) were invested in corporations of foreign countries. private placements accounted for 23.1 percent of corporate bond holdings. all but 2.6 per cent of the corporate bond holdings were classified as high or medium grade-naic classes 1,2, and 3. average distributions by maturity have been shortening in recent years with only 34.4 percent of corporate bonds held at the end of 1993 maturing in over 10 years (17.2 percent in more than 20 years). government securities are of longer average maturity with 53.5 percent maturing in over ten years and 32 percent maturing in over 20 years. for eign government and international agency securities accounted for 9.5 percent, and state and local issues made up 3.8 percent of all government securities. the remainder were u.s. federal agency and treasury issues. mortgages were distributed as follows: 91.7 percent commercial, 4.1 percent one-to four family homes, and 4.1 percent farm mortgages. properties located in the u.s. accounted for 98 percent of all mortgages. in the absence of a proxy for mortgages, they were combined with directly-owned real estate. the small portion of directly-owned real estate consists primarily of large apartment complexes, shopping centers, multi-purpose office buildings, and home and regional offices. common stocks accounted for 95.8 percent of all life insurer stock investments. this statistic applies to common stocks for general and separate accounts combined. only 17.8 table 2. distributions of projected compound annual returns based on proxies of asset classes held in general accounts of u.s. life insurers at the end of 1993 return percentiles i year 5 years period of in-force status iv years 20 years 30 years 40 year.7 50 years 95th per 18.84 12.13 10.60 9.53 9.06 8.78 8.59 75th per 11.70 9.07 8.45 8.02 7.83 7.72 7.64 65th per 9.65 8.17 7.82 7.58 7.47 7.40 7.36 50th per 7.20 7.03 7.01 7.00 6.99 6.99 6.99 35th per 4.38 5.81 6.15 6.40 6.50 6.57 6.61 25th per 2.47 4.94 5.54 5.96 6.15 6.26 6.33 5th per -3.69 2.07 3.49 4.50 4.95 5.22 5.41 note: 0 computed using data from lbbotson associates. porrfolio stmrrgist software, 1996. chicago. ii. used with permission. all rights reserved. his&acal investment performance 163 compound annual return f%) l&ml time i 9sth percentifs 75th percentlfff -85th percentile averngc veius 35th i”erc8ntifo 25th percentile, 5th percent&e souycc: 0 computed using data from ibbotson associstes, portfxio sfmregisr software. 1996, chicago. il. used with pcrmis sion. all rights reserved figure 2. distribution of projected compound annual returns based on proxies of asset classes held in the general accounts of u.s. life insurers at the end of 1993 percent of total common stock holdings were held in the general accuunts of concern in this study. because outstanding policy loans reduce cash values and death benefits, they are highly secure. miscellaneous assets consist primarily of due and deferred premiums (14.8 percent of this asset class), investment income due and accrued (20.5 percent), cash (0.5 percent), and real estate joint ventures (acli, 1994). the use of treasury bills as proxies 164 financial services review 6(3) 1998 for the latter two categories is further justified by the fact that 2.5 percent of securities in the “government bond” category were short-term treasury and short-term federal agency issues being held for liquidity. the proxies do not perfectly represent the universe of general account investments or even those of a specific insurer: actual maturities and grades of bonds differ somewhat from those reflected in the lehman brothers indexes; the typical universal life insurer may hold shorter average maturities than the universe of insurers represented in acli data; net returns on policy loans may exceed those on 90-day treasury bills now that some policy loan rates are tied to an index of seasoned corporate bonds and some others are set at 8 per cent; a small percentage of mortgage investments is not business related; the portion of mortgages without participation provisions may be more similar to bonds than to the com mercial real estate proxy; and so forth. the purpose in this study is to demonstrate a tech nique for projecting returns rather than perfectly reproducing the history of actual industry holdings. thus, some weakness in the proxies is of little concern. the preferred data inputs for a particular insurer using the technique would come from its own portfolio. projections for the hypothetical general account have a 50th percentile (median) current return of 7.20 percent and a standard deviation of 6.85. by the fifth year the median value decreases to 7.03 percent and levels off at 4.99 percent sometime between years 20 and 30 as current returns (1994 in this simulation) have less influ ence on projected rates over time. at the 90 percent confidence level, possible first year returns range between -3.69 percent and 18.84 percent, while fifth year returns range between 2.07 percent and 12.13 percent. the fifth percentile returns of -3.69 and 2.07 percent should be of concern to insurers, however, consumers would be protected from actual performance below contractual guaranteed minimum returns (e.g., 4.5 per cent). for a 50-year policy holding period and a 90 percent confidence level, the median of 6.99 percent is bordered by 5.41 and 8.59 percent. these results are shown in table 2 and figure 2. the “trumpet graph’ (see figure 2) illustrates the rather rapid convergence of the per centile projections towards the median return over time. at any selected confidence level, the relationship between the maximum long-run return and the median return declines over time. contrarily, the relationship between the lower value and the median is positive over time. vi. discussion and conclusions the u.s. life insurance industry has for many years provided sales illustrations to prospec tive buyers that project maximum cash values using current experience with respect to investment returns. the majority of insurers use an average portfolio rate of return as a crediting rate (minus a spread for profit) or in dividend scale dete~inations. conse quently, for insurers with a relatively long average maturity for securities, the current return may reasonably represent historical experience except in periods of extreme abnor mal returns. in times, such as the 198os, with several years of historically high investment returns, however, even the average portfolio rate can substantially exceed those returns that are likely to be attainable in the long run. the environment for this situation is enhanced by strong sales producing large amounts of reserves for relatively new contracts, lapse rates historical investment performance 165 reducing the importance of older reserves, and turnover of existing investments. the pos sibility of current returns being out of synchronization with achievable long-run returns is greatly heightened by the use of the investment generation method of crediting interest rates. yet new model legislation and actuarial standards fail to effectively solve this prob lem of potential misrepresentation. thus, financial planners will continue to be challenged to bring reason to sales illustrations. a question of concern relevant to a projection like that presented above and its possi ble use in projecting cash values is: what would be the proper time period to identify the and range mid-point? one approach is for the planner to ask the client how long he or she expects to hold the policy being considered, ignoring the possibility of premature death. a conservative approach would specify the difference between an advanced age such as the client’s life expectancy or age 100 and the client’s age at the time of illustration (e.g., age 80 minus illustration age 30 equals a 50-year projection). given the projection in this study, the issue becomes academic for periods beyond 20 years or so because over 90 percent of the reduction in maximum values occurs by the 20fh year. one could argue that the prob lems in making forecasts of investment returns are such that projected values at some dis tant point (i.e., beyond 20 years) lack sufficient validity. further, most cash value policies mature through surrender or death long before life expectancy for the typical consumer. recent life insurance marketing research association data show that the expected life of a whole life policy in the u.s. varies between eight and one-half years when issued at age 25 and 12.9 years for age 55 issues (potasky et al., 1992). another issue surrounding any practical use of projections based on the methodology employed here concerns the proper statistical confidence level at which to identify the range of investment rates of return to use in illustrating cash values. in this article the pri mary discussion has been about the median long-run return and 90 percent confidence level rates. on the one hand, one can argue that there is only a 10 percent probability that long run rates of return will be at either the maximum or minimum levels of the 90 percent con fidence level. thus, these rates are “highly unlikely” based on history. on the other hand, such extremes are possible and might be of interest to risk-taking and risk-averse clients. a conservative advisor, however, might want to select a lower confidence level, believing that the recognition of extreme levels of variation goes beyond what is proper. a conservative approach is to argue that projections based on compound returns greater than the average return during a past period believed to reasonably represent the future is not credible. the problem with this approach is its reliance on only one moment of the distribution of historical returns. the method used in this study improves on the his torical average approach by also recognizing the standard deviation of historical returns and confidence levels around the mean. to the extent the theoretical distributions employed in the current study of investment projections fit reality, they demonstrate that long-run returns converge over time to a relatively narrow range around the projected median return. current returns during periods of historically high returns are likely to be well beyond what seem achievable in the long run. thus, serious concern exists about the ethics and credibility of industry practices. the ethical financial planner will want to pro vide a perspective for evaluation of the guaranteed, current, and intermediate projections found in the typical current sales illustration. possible future cash values are related to risk. the consumer is better informed if mea sures of risk, as well as expected returns, are communicated in any financial transaction. a major value of the method illustrated in this study is the ability to present levels of pessi 166 financial services review 6(3) 1998 mistic and optimistic returns around the median in a risk context. for example, the likeli hood of the long-run cash value reaching the level illustrated at the maximum value for the 90 percent confidence level is an optimistic 10 percent, given a two-tailed test. the con sumer who is a risk taker could dream about this high level of return, hopefully, still being fully informed that any return in this vicinity is unlikely to occur. with current life insurer marketing practices, the typical prospective client is given little information about the like lihood of illustrated returns, even when they have virtually no likelihood of materializing in the long run. current returns are described simply as not likely to occur and actual results may be more or less favorable than those shown. the illustration of perhaps the guaranteed return, the projected median return for a portfolio like that backing the product, a pessimis tic projection between the guaranteed and median rates, and one or two optimistic projec tions above the median, with accompanying confidence levels would provide helpful information to consumers. in any effort to improve communications, it is important to avoid overloading the consumer with information. while no scientific methodology can exactly pinpoint future performance because of unknown future factors, perhaps it is reasonable to hypothesize that long-run returns will be similar to the distribution of past returns for selected periods of history. any acceptance of this hypothesis, assuming normal or lognormal distributions, will result in a broad range of future short-run returns. however, the range of possible returns fitting a given level of statistical confidence will narrow as the investment horizon increases. it is projected rates of return for the long run that should be used as compound rates of return in projections of life insurance cash values. references acli, (1994). life insurance fact book. washington: american council of life insurance. actuarial standards board (1995). compliance with the naic life insurance illustrations model reg ulation. actuarial standard of practice no. 24, dot. no. 050, p. 2. aitchison, j., & brown, j. a. c. (1957). the lognomtal distribution: with special reference to its uses in economics cambridge: cambridge university press. bock, b. j., faucett, j., & cole, l. n. (1992). life insurance sales illustrations, society of actuaries report, vol. 18, no. 3 (pp. 1317-18). burgess, r. w., et. al. (1994). measuring insurance company investment performance. society of actuaries record, vol. 20, no. 2. crow, e. l., & shim& k. (1988). lognormal distributions: theory and applications. new york: marcel dekker. lewis, a. l., kassouf, s. t., brehm, r. d., & johnston, j. (1980). the ibbotson-sinquefield simula tion made easy. journal of business, 53(2), 205-214. lucas, l. (1995). building better inputs: a guide to the methodology and value added of ibbotson’s optimizer inputs. unpublished paper. mcgough, r. (1997, may 13). now, funds may choose their past. the wall street journal, pp. cl, c25. miller, w. n. (1996). special report: how companies are answering the iq. journal of the american socie@ of clu & chfc, 50(2), 82-89. naic, (1995) life insurance illustrations model regulation. potasky, s., feinberg, m. j. , katcher, m. r., & querfeld, e. b. (1992). individual life policy update. sociew of actuaries record, vol. 18, no. 4b, p. 2043. historical investment performance 167 pritchett, s. t. (1998). life insurance sales illustrations: perspectives on problems and corrective options. alternative approaches to issues in insurance regulation, kansas city: national association of insurance commissioners. using historical data in optimization and forecasting. (1995) stocks, bonds, bills, and inflation: 1995 yearbook, (pp. 159-73). chicago: ibbotson. pii: s1057-0810(97)90036-6 book, software, and web site reviews 73 quote.com (http://www.quote.com), they also provide sector ratings and current price quotes within sectors. the site offers intraday updates on newsworthy stocks and analyses of both the debt and equity markets. also through quote.com, briefing.com provides quotes on individual securities (and on several market indices) which include pricing information along with fundamentals such as p/e ratios dividend yields and the 52-week price ranges. these 15 minutes delayed quotes are updated constantly throughout the trading day. members can also create a portfolio of up to 25 stocks in which they can monitor daffy updates of prices and volume. in addition, the portfolio feature will track gains and losses on all individual securities and the entire portfolio based on inputted purchase price. both intraday antl his torical price charts are also available. stock briefs on over 8,300 companies are also available. these briefs include a short business summary, key financial ratios, share-related information, short interest informa tion, institutional and insider ownership, historical growth rates in sales, eps and divi dends, and historical quarterly figures for revenue and eps. these briefs are updated with current data each monday. the monthly subscription fee for briefing.corn is $6.95 and includes access to quote.com. individual investors who trade through e-trade (the electronic brokerage f'mn) have free access to briefing.com. the money book of personal finance richard eisenberg and the editors of money magazine new york, ny: warner books inc.; 1996 (isbn 0-446-51981-2) reviewed by: douglas r. kahl, professor of finance, university of akron the money book of personal finance is a popular personal finance guide available in almost any bookstore not a personal finance textbook. however, i found that it works very well as the primary textbook in a personal finance course for nonbusiness majors. the book was a collaborative effort by at least thirteen writers and editors at money magazine. the clear, concise and authoritative writing style does not assume any prior background in finance or business. since it was not designed to be a course textbook, it lacks the usual classroom sup port package. there is no study guide, no instructor's manual, no chapter end problems and no package of overheads. for this reason, i would not recommend the money book of personal finance as a primary text when the instructor is teaching the course for the first time. for the instructor who teaches a personal finance service course for nonbusiness majors on a regular basis, this text could be an interesting and refreshing change. the book was very popular with my students who were about equally split between traditional and nontraditional backgrounds. both the traditional and nontradi tional students gave considerable credence to the book because of the perceived practi cal knowledge and applied expertise of the authors. the students also noted, with considerable approval, that the price was about one-fourth that of a standard text. they found the lower cost especially appropriate for a personal finance course. the book covers the usual personal finance topics in four major sections: getting started, reaching your financial goals, investing your money and your family 74 financial services review 6.(1) 1997 finances. the book starts to draw the reader's interest very quickly, starting with the ques tion "how are you doing?" and providing a wide range of comparisons of assets, wealth and income based on age, education and family status. this is quickly followed by a dis cussion of budgeting, basics and financial advisors. the second, and largest section, pro vides an up-to-date, accurate and concise discussion of the usual range of personal finance topics from savings to debt and career planning to estate planning. the third section con sists of a knowledgeable and practical discussion of investing. the final section on family finances covers three important topics not often covered in traditional personal finance texts. one chapter discusses the differences between men and women in income, savings, investing and money management combined with advise on making the partnership work. another chapter provides an excellent discussion and some good advise on educating chil dren about money. the final chapter provides discussion and advise on an often delicate and difficult topic now facing large numbers of nontraditional students (and faculty), your parents and money. the book ends with an appendix containing a concise list of govern ment benefits, what you need to know about your benefits and whom to contact. while this book may not meet every instructor's needs for a service course in personal finance, it does provide an accurate, up-to-date, cost effective and readable alternative to a standard personal finance text for the instructor willing to exchange a little extra effort for some additional flexibility and credibility. pii: 1057-0810(95)90018-7 financial services review, 4(l): 57-60 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. a practitioner’s perspective: comments on “analysis of u.s. savings bonds” barbara s. poole in this issue, potts and reichenstein describe features of u.s. savings bonds and discuss their evaluation. the article can serve as a reference for busy practitioners who may not follow changes in the savings bond market closely, especially those who tend to rely on product related education for updates on financial instruments. savings bonds remain an important investment for individuals; on march 31, 1994, individuals held $174.9 million in u.s. savings bonds, a 7% increase from one year prior (williams, 1994). further, as table 1, “demographic characteristics of individual savings bond holders,” indicates, the holding of savings bonds is not exclusive to any particular age group or income level (federal reserve bulletin, 1994). i. potential changes the recently (november 1994) passed general agreement on tariffs and trade (gatt) included provisions that gave thetreasury department full authority over interest rate setting on u.s. savings bonds. the agreement removes the4% guarantee previously set by congress for bonds held for less than five years and enables the treasury to index the savings bond rate to that of the six-month t-bill. peter hollenbach, a spokesman for the bureau of public debt, indicated in december 1994 that changes will not be retroactive and “an announcement of any change is some months away” (lazzareschi, 1994b). while details are not finalized, changes expected are: l pegging the rate for savings bonds held for less than live years to 85% of the six-month t-bill rate, eliminating the 4% rate guarantee. bonds held for 5 years or more will continue to earn 85% of the average rate paid by treasury notes with five years left to maturity, with a minimum of 4%. l changing interest accrual from a monthly to a semiannual basis. those individuals who sell just before the six-month or one-year anniversary will forego the partial period of interest. barbara s. poole l managing editor, financial services review, department of finance, university of connecticut, storrs, ct 06269. 58 financialservicesreview 4(i) 1995 to those unfamiliar with legislative customs, inclusion of these changes in the appar ently unrelated gatt seems unusual. but because new legislation passed by congress must be “deficit neutral,” gatt needed a revenue enhancer. the congressional budget office’s projection of $120 million in savings over five years as a result of the interest changes provided that revenue source needed for gatt’s passage (lazzareschi, 1994a). the treasury has wanted to make these changes since the early 199os, when the 4% fixed rate was so favorable that investors, as the potts and reichenstein article suggests, used the bonds as short-term investments. according to hollenbach, the treasury wanted toreturn bonds “to their original purpose: to buy and hold for the longer term, instead of using them as money market mutual funds whenever [the] fixed rate looks more attractive.” (boston herald, 1994) the removal of the interest rate minimum, as well as the change in interest accrual from monthly to semiannually, will limit these opportunities. however, realizing the savings from the removal of the 4% rate floor would require an extended period of low interest rates. even with the low interest rates during the early 1990s the variable rate never fell below the 4% floor and “temporary dips would be moderated in any event since the return you get at redemption is an average of the six month variable rates” (smith &wiener, 1995). the $120 million savings estimate was based on the april 1994 assumption that treasury bill rates would be 4.3% for 1995 and 4.6% thereafter. in early 1995, when six-month t-bills paid 6.81%, an immediate implementation of the changes would have obligated the treasury to pay 5.79% rather than 4% on savings bonds cashed in early. with revised interest rate assumptions, the changes are estimated to cost the treasury about $360 million over five years in additional interest payments (kristof, 1995). ii. implicationsandstrategiesfortheindividual since changes are expected in spring 1995, those individuals who are concerned about long periods of low interest rates should lock in the minimum rate guarantee by buying before the changes are implemented. likewise, liquidity-minded individuals who want the option to redeem the bond in less than six months should purchase before the change eliminates monthly interest crediting. several other features of the bond are worth noting. first, the tax deferral differs from the deferral associated with qualified pension plans where individuals must take a required minimum distribution by the april 1 following the year of attaining age 7ot/2. in contrast, savings bonds’ interest can be deferred for up to 30 years or until redemption, regardless of the holder’s age. this could be helpful for older individuals looking for tax deferral; additionally, compared to the qualified plan, savings bonds afford individuals more liquidity and more control over when they choose to redeem the bonds and pay tax on interest. planners also should not neglect the estate planning benefits of savings bond invest ments. a bond purchaser can request that the bonds include an inscription naming a beneficiary “pod,” to be paid on death. then, at the owner’s death, the bonds are excluded from the probate process; the named beneficiary can redeem them or have the bonds reissued in the beneficiary’s own name (doyle, 1995). when the owner dies and there is no beneficiary listed, the survivor must forward appropriate forms, along with a death certiti cate, to the federal reserve bank (doyle, 1994). a ~a&titioner~s perspective 59 there are reports that the 1993 tax law set thresholds lower than congress intended for tax exemptions on savings bonds used for educational funding; there is speculation that congress may pass a technical corrections bil to make changes to the law. this could be good news for high income couples because when bonds are used for educational funding, the 199s full tax exemption on interest applies to couples with adjusted gross incomes (agi) up to $63,450, and a partial exemption is available for couples with agis up to $93,450 (klott, 1995). while impending changes may limit the savings bonds’ appeal as a short-term vehicle, their tax and estate treatment, along with safety and freedom from transactions costs, continues to make savings bonds a reasonable conservative investment vehicle. potts’ and reichenstein’s work contributes to our understanding of these bonds’ valu ation. appendix percentage of families holding u.s. savings bonds in the year 1992 an families 22.7 income (in 1992 dollars] less than $10,000 6.6 $1 o,ooo-24,999 13.3 $25,000-49,999 27.9 $5o,cco-99,999 39.5 $100,000 and over 32.1 age of head of househoid (in years) less than 3.5 35-44 45-54 55-64 65-74 75 and over 22.8 29.4 25.4 21.4 14.1 14.5 source: federal reserve b~i~e~jn (october, 1994). references changes in family finances from 1989 to 1992: evidence from the survey of ~onsum~r~nances. (1994, october). federal reserve bull&z, 80( lo), p. x61. doyle, w. (1994, june 29). savings bonds from vietnam era are worth keeping. houston chronicle, p. 3. doyle, w. (1995, january 9). no need to report stock gift on tax return. st. petersburg times, p. 4. finance. (1994, december 13). bosston herald, p. 35. klott, g. (1995, january 29). last time for paying quarterly household worker tax. fresno bee, p. fl. kristof, k.m. (1995, january 17). passing the buck in ‘95: four things legislators could resolve to do to make america more prosperous. chicago tribune, p. 7. lazzareschi, c. (1994, december 25). money talk: how to claim ex-spouse’s benefits. los angeles times, part d, p. 5. 60 financial services review 4(l) 1995 potts, t.l., & reichenstein, w. (1995). analysis of u.s. savings bonds. financial services review, 4(l), 41-56. a question of money: inheriting bonds. (1994, march). consumer reports 59(3), p. 199. smith, a.k., & wiener, l. (1995, january 23). early bets for ‘95. u. s. news and worldreport 118(3), p. 62. williams, s. (1994, may 15). savings bonds attracting interest. hartjhrd courunt, p. cl. pii: s1057-0810(99)00022-0 the correlations of professionalization and compensation sources with the ethical development of personal investment planners kenneth s. bigela,* aadjunct associate professor, new york university, 200 west 15th street, #5–a, new york, ny 10011-6539, usa k. bigel & co., director (new york city-private investment management firm) abstract recent controversies concerning personal financial planners, and investment planners in particular, have centered around two issues: competence and ethics. this paper has focused on the ethical development of a sample of investment planners in connection with the correlative roles of professionalization and compensation sources. certified financial planner designees were found to manifest higher ethical development scores than non-certified financial planner designees. fee-based planners manifested no significantly different ethical development scores than their counterparts. other demographic variables were also studied. © 1999 elsevier science inc. all rights reserved. keywords:financial planning, ethical development, professionalization, compensation sources 1. introduction there has been much controversy in connection with both the competence and the ethics of personal financial planners. this study examined the ethical development of personal financial planners who specialize in investments — “investment planners,” a group, which has been the subject of particular ethical controversy. except for examining legal and regulatory complaints, no gauge of ethical behavior currently exists. any examination of legal and regulatory complaints and rulings would be * corresponding author. e-mail address:kenneth.bigel@nyu.edu (k.s. bigel) financial services review 7 (1998) 223–236 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00022-0 limited to a small and unrepresentative subset of the relevant population. measuring ethical development is a viable alternative and its measurement is important because ethical development has been found to be linked to behavior (rest & narvaez, 1994). “ethical development” refers to a psychometric taxonomy defined by rest (1979). numerous studies of professionals’ ethical development (alternatively labeled “moral judgement”) have been conducted in connection with accounting, dentistry, journalism, teaching, veterinary medicine, and other professions (armstrong, 1984; arnold & ponemon, 1987, 1991; loeb, 1971; mccarthy, 1993; ponemon, 1990; ponemon & gabhart, 1990, 1993; self et al., 1993; self et al., 1994; westbrook, 1994). this researcher has found no ethical development studies in connection with investment planners. 2. the investment planning industry the sec, under the “securities exchange act of 1934,” also allows for the creation of self-regulatory organizations (sros), which are industry groups that govern its own members. all stock exchanges, as well as the national association of securities dealers (nasd), may be characterized as sros. the over-the-counter (otc) stock market is governed by the nasd. further, virtually every investment banking firm in the united states is a member of the nasd and is thus subject to its regulation (johnson & mclaughlin, 1997). representatives of these firms are usually referred to as “brokers” and the brokers may also position themselves as personal financial planners. the domain of the broker provides the second means by which an investment planner may choose to enter the field for compensation (the first being the ria/iar route noted above). requirements for attaining either broker/representative or ria status differ. in order to become a ria, one must merely register (i.e., file an application and pay a fee) with the sec or a state securities commission, indicate his personal history, including any criminal records. no competency examination is required. the ria must also maintain books and records and is, of course, subject to the law. in contrast, representatives must pass a nasd-administered examination in order to broker securities (nasd, 1998, membership and registration rules, rule #1030). the most common nasd examination is the “general securities representative” (series 7) examination, which allows registered representatives/brokers to effect transactions in all securities instruments; other examinations, including the iar examination, tend to be less comprehensive. in addition to passing a written examination, the broker must also be affiliated for at least four months with a nasd member broker/dealer firm. the stock exchanges, as sros, also may confer broker status limited to activities with the conferring exchange. it is noted, however, that the nasd administers the various “series” examinations and it is these registration examinations which are most common and universally accepted. normally, brokers promoting fee-based investment management services will also acquire the iar status, while independent advisers (i.e., those not affiliated with a broker/dealer firm) will obtain the ria status. both brokers and iars are considered “representatives” operationally in this study. further, all iars in this study also bore general securities representative (series 7) status. 224 k.s. bigel / financial services review 7 (1998) 223–236 the status of the investment planner also affects the compensation method he/she may demand. an ria will charge a client a fee for investment advisory services rendered, whereas a broker/representative will earn a commission for a transactional service, often but not always as the result of having rendered some investment advice. the ria may not execute the advice on behalf of the client (although he may direct its execution), whereas the broker is compensated based on execution only. iars, in turn, promote a service, whereby a ria is recommended and the iar will share in the fees generated from this multiple relationship. some investment planners offer clients the choice of either a fee or a commission compensation structure. other planners may use a “commission versus fee offset” arrangement, whereby any transactional commissions paid, may serve to reduce fees charged for services. thus, for example, if the client is to be charged 1% of the $1 million dollars he/she has invested with the investment planner/adviser, and the account has also generated $1,000 in commissions, the fees will be reduced from $10,000 (1% of $1 million) to $9,000. investment planners who provide either choice are deemed “combination” planners in this study. combination planners must be both brokers and registered investment advisers. partly in response to controversy concerning compensation arrangement, a membership organization has been formed called the national association of personal financial advisers (napfa), whose membership consists only of fee-only planners. incidentally and in general, the rendering of generic investment advice, which is incidental to the performance of other professional services (i.e., including tax accounting, estate law, and more remote activities as teaching and journalism), does not fall under the rubric of compensable investment advice. these other professionals may be exempted from the legal and regulatory requirements mentioned. thus, a teacher who, in the classroom, suggests that one invest in stocks for the long term is not acting in the capacity of investment advisor, but as a teacher. the law is usually interpreted as meaning that if you do not get paid directly for the specific activity of investment advising and do not present yourself as such, you are not acting as an investment adviser. such other professionals have not been considered in this study. apart from differences in compensation source, some financial planners carry professional designations, including particularly the certified financial planner (cfp) mark. attainment of the cfp designation requires the completion of a rigorous two-year study program, including a professional ethics component. the candidate must affirm fealty to the “code of ethics and professional responsibility,” which is promulgated by the cfp board of standards. additional requisites include passing an examination, having related work experience, and acquiring ongoing educational credits. the cfp designation is the leading credential for investment planners. indeed, the international association for financial planning (iafp) has recently announced that it is officially recommending that all financial planners (i.e., including investment specialists) obtain the cfp designation (robaton, 1997). large firms are also encouraging employees to obtain the cfp mark in order to “foster a public image of acting in the best interests of the consumer” (hansard, 1999, p. 3). 225k.s. bigel / financial services review 7 (1998) 223–236 3. psychology of ethical development this study’s dependent variable is a known measure of one’s level of ethical development, which is called the “principled score” or “p-score” (rest, 1979). according to the psychology of ethical development, varying rationales may be employed in arriving at ethical decisions, which may be articulated in terms of a progressive, multi-stage, developmental taxonomy (kohlberg, 1984; rest, 1979). in this taxonomy, the individual is said to first emphasize his egoistic need for self-preservation, disregarding any other considerations. developing further, he/she may then start paying heed to the demands of other individuals and groups, eventually acquiring more socially determined thinking. in the end, some reach the highest stages of ethical development, which represent “principled” thinking and which are akin conceptually to philosophic teleology and deontology. “teleology” refers to the assessment of the net benefits to be derived from an action in comparison to the demerits of the means employed in attaining the ends. it is thus a kind of ethical cost/benefit analysis and is often alternately referred to as “consequentialism” or “utilitarianism.” utilitarianism is usually associated with jeremy bentham and john stuart mills, english philosophers who lived in the eighteenth and nineteenth centuries. “deontology” is more intuitive, requiring the absolute adherence to the “golden rule” (i.e., respecting other people as ends in themselves and not ever as means to achieve some end). certain actions are judged to always be either “right” or “wrong” independent of any results produced. deontology is associated with immanuel kant, a prussian philosopher who lived in the eighteenth century. individuals whose thinking is dominated by teleological and deontological processes are considered most highly developed morally. ethical development will differ among individuals and may indicate not only how people think, but how they may actually respond to particular ethical dilemmas. “principled” individuals (i.e., those with the highest ethical development scores) are more . . . . . . likely to behave consistently with their own principle-based decisions—they’ll carry through and do what they think is right. . . . less likely to cheat . . . more likely to help someone in need, and more likely to blow the whistle on misconduct. (trevino & nelson, 1995, p. 92) ethical development may be assessed by examination. the moral judgment interview (mji), the medium by which kohlberg (1981) assessed one’s ethical development, is intended to estimate the highest ethical development stage a subject has achieved. rest (1979), using kohlberg’s taxonomy, created the defining issues test (dit) to gauge ethical development or moral judgement. the dit is a written instrument, which is scored by algorithm, as compared to the mji, which uses the interview. the dit measures ethical development on a continuous scale, indicating the extent to which one thinks in terms of the highest ethical principles. as the dit measures cognitive skills, it may be more useful than kohlberg’s mji in the study of the relationship between moral reasoning and actual behavior. moreover, “. . . there is a persistent statistically significant relationship” (rest & narvaez, 1994, p. 22) between ethical development and behavior. the cognitive-behavioral link is established through rest’s “four-component model” (rest & narvaez, 1994, pp. 22–25), 226 k.s. bigel / financial services review 7 (1998) 223–236 which first requires that the subject recognize that a moral dilemma is present, while also recognizing that there may be other influences on behavior. rest’s four components of morality are ethical sensitivity, reasoning, motivation, and character (rest & narvaez, 1994). it is the second component (known as “component ii”) that we are concerned with herein—the development of ethical reasoning or judgment; the second component informs ethical choices. thus, ethical development scores (i.e., rest’s “p-score”) assess the framework which people bring to solving moral problems; the higher one’s p-score, the more the subject arrives at moral judgments in a manner akin to ethical philosophers (i.e., in terms of teleology and deontology). the dit is an instrument, which has been widely used both over many years and across the study of numerous professions. the dit has been used extensively since the 1970s. currently, the number of studies using the dit totals over 1,000; the total of subjects taking the dit numbers in the hundreds of thousands; the dit has been used in over 40 countries; and the published literature on the test is extensive, with about 150 new studies each year. (rest & narvaez, 1994, p. 13) the dit combines all of the following elements, which make its use appropriate for this study. the dit... 1. is indicative of a taxonomy of ethical development, particularly as it may be used to reflect latter stage thinking, which is most appropriate for a study of adult-professionals; 2. addresses a cognitive process, which is purely moral in content; 3. is a written test, which is easily administered by mail-survey; and 4. has been tested extensively over time. 4. methodology this study examined self-proclaimed “financial planners” who were also brokers and/or registered investment advisers; such parties are termed “investment planners” in this study. membership in the international association for financial planning (iafp) qualified the practitioner as a self-proclaimed financial planner and served as this study’s sample frame. in order to effect the study, a stratified, random sample was drawn from a list of domestic members of the iafp. the iafp, whose membership (at the time the sample was drawn) exceeded 16,000 professionals, is the largest financial planning organization and has a substantial investment planner constituency. the sample was later demarcated both by the subjects’ attainment of the cfp designation—or lack thereof—and by the means by which the investment planner chose to be compensated. a mail survey containing both rest’s (short-form of the) defining issues test (dit) and a researcher-designed demographic questionnaire was implemented. the mailing consisted 227k.s. bigel / financial services review 7 (1998) 223–236 of 424 questionnaires; however only 196 or 46.2% of the questionnaires were returned with usable responses, suggesting the possibility of non-response bias. (respondents who neither indicated that they were “representatives” nor registered investment advisers were classified as non-usable responses; such respondents were minimal.) p-scores were calculated for all usable responses. next, a test of non-response bias was conducted. the difference between the p-score means of the early and late respondent groups was minimal (table 1), especially when viewed relative to their standard deviations. a two-tailedt-test for significance of independent means was conducted; the results failed to attain statistical significance,t (194)5 .27, p . .05. given this test, it would appear that there is no significant non-response bias in this study. this conclusion is based on oppenheim’s hypothesis (oppenheim, 1966), which asserts that non-respondents’ data will more closely resemble late respondents’ than earlier ones. further, the relative proportions of each of the variables in this study versus the sample frame were similar. table 1 descriptive p-score data n mean standard deviation cfp designee 99 38.01 16.98 non-cfp 71 33.62 13.69 total 170 fee-only 29 38.51 16.78 comm. & combin. 141 35.70 15.61 total 170 commissions 30 36.45 15.83 fee-only & comb. 140 36.12 15.85 total 170 h3: low educ. and no prof. cred. 66 32.83 14.99 h3: hi educ. or advc’d prof. cred. 104 38.30 15.99 total 170 h4: “low age” (under 40) 34 36.57 16.26 h4: “mid age” (40–59) 113 36.46 14.74 h4: “late age” (601) 22 34.70 20.59 total 169 h5: new in profession (0–4.9 yrs.) 36 38.98 15.34 h5: experienced (5–10 yrs.) 34 39.31 16.85 h5: established (101 yrs.) 99 33.94 15.40 total 169 h6: males 140 35.33 16.03 h6: females 30 40.11 14.26 total 170 228 k.s. bigel / financial services review 7 (1998) 223–236 5. hypotheses concerning primary variables this study focused on two key investment planning issues: compensation sources and professionalization (i.e., attainment of the cfp designation). these issues served as the study’s “primary” independent variables. compensation sources may be problematic in the investment planner-client relationship. a commonly reported abuse is “churning” or excessive turnover of holdings for the sole purpose of generating commissions to enrich the financial planner (loss & seligman, 1989; spiro & schroeder, 1995). a similar abuse is the excessive trading brokers effect near months’ ends in order to generate sufficient compensation for the entire month (brown, 1996). another is unsuitable investment recommendations, particularly where high-commission, proprietary investment products are concerned. the compensation matter is not generally thought to be as problematic in the fee-based and combination venues because of the absence or modification of transactional compensation. fee-only planners are thought to have less motivation than commission-based planners for putting their own interests ahead of their clients’; this is because the fee-based planner does not have the temptation to engage in unnecessary trading in order to generate income for himself. the fee-only planner is paid for his advice only, thereby reducing or eliminating conflicts of interest and moral hazard. an investment planner who registers relatively highly on an ethical development scale, may therefore gravitate to media of compensation that present less moral dilemmas. (it is noted that the moral development literature does not yet address this hypothesis.) this is not to say that the fee-based or combination investment planner is free from conflicts of interest or ethical dilemmas. by charging ongoing money management fees, the financial planner can create an annuity, or perpetual income stream for him/herself, which may be more costly in the long-run for the client. this compares to the alternative of having effected a one-time commissionable transaction that nonetheless satisfies all the client’s present financial objectives. “for too many planners, assets under management means assets under a fee-collection system” (abram, 1996, p. 83). the financial planning industry has been marked by the increasing professionalization of its ranks since the advent of the certified financial planner (cfp) designation in 1973. throughout history, “profession” has been defined as an occupation, which requires training and specialized study. in addition to examinations and other criteria that go along with admission to a professional society’s ranks, including cfp licensure, come numerous behavioral and ethical expectations. there is “the belief and social expectation that the possession of special knowledge and skills carries with it the mandate for its moral and ethical use” (self & baldwin, 1994, p. 147). professionals who are holders of certificates must both abide by the relevant codes of ethics and adhere to professional standards (king & chironna, 1989, p. 29). indeed, an important reason that a professional organization is formed is to create a specialized code of ethics that circumscribes the profession and its community (durkheim, 1958). the cfp designation is also well recognized. in a study concerning the recognition of professional designations, the certified public accountant (cpa) designation was most recognized, while the certified financial planner (cfp) and certified life underwriter 229k.s. bigel / financial services review 7 (1998) 223–236 (clu) were next most recognized (nelson & nelson, 1997). cfp recognition was also noted as being in ascendance, while clu recognition appeared to be in decline. this study, however, was not concerned with public accounting (cpa) or insurance (clu) respectively, but with investment planning, by and large the domain of cfp designees. although no studies were found in connection with the clu designation, a plethora of studies have been done concerning the moral judgement of cpas (armstrong, 1984; arnold & ponemon, 1987, 1991; jones & ponemon, 1993; loeb, 1971; ponemon, 1990; ponemon & gabhart, 1990, 1993; shaub, 1994). however, the moral dilemmas faced by the cpa, who until very recently did not offer commission products, are different than those facing the investment planner, who is also faced with the vagaries of the financial markets and clients’ perceptions thereof. given the absence to date of research concerning the ethical development of investment planners, prior research in connection with accountants is of some interest. in a paper concerning accounting students who studied the american institute of certified public accountant’s (aicpa) ethical code, fulmer and cargile (1987) found that the students had greater awareness of ethical issues as compared with those who had not studied the code. moreover, sellers (1979) found that such students perceived ethical issues in a manner similar to that indicated in the aicpa ethical code, whereas students who had not studied the code had a more dissimilar perception. similarly, the cfp certification examination requires that the cfp candidate be knowledgeable of the cfp board of standards’ code of ethics and professional responsibility. once being awarded the cfp designation, the cfp licensee must subscribe to the code and take continuing education courses in cfp ethics for a minimum of two credit hours (i.e., four classroom hours) every two years in order to maintain the cfp status. in summary, the respective correlations of compensation sources and professionalization with the investment planner’s ethical development constituted the focus of this research. it is expected that fee-based planners and certified financial planner designees will score higher on rest’s moral judgement scale (i.e., p-score) than both commission-based and combination planners, and non-cfp practitioners. the following articulates these directional hypotheses in operational form: h1: the certified financial planner licensee will manifest higher ethical development than the non-licensee, as defined by rest’s ethical development score, which is called “p-score.” h2: the fee-based investment planner will manifest higher ethical development than both the combination and commission-based planners respectively. 6. hypotheses concerning demographic variables this study also examined certain demographic correlates in connection with the investment planner’s ethical development/moral judgement: age, educational achievement, career tenure, and gender. due to the fact that the dit is a cognitive-developmental instrument, it stands to reason that age and education should be examined. this is based on the supposition 230 k.s. bigel / financial services review 7 (1998) 223–236 that people change over time and that the change is as predicted by developmental theory— with increased age and education come higher moral judgment scores (rest & narvaez, 1994). rest found that many studies “reveal that increased education is, in fact, generally associated with higher levels of moral judgment” (rest & narvaez, 1994, p. 28). moreover, years of formal education were found to be a far greater predictor of ethical reasoning and moral development than chronological age per se. “the evidence at hand suggests that adults in general do not show much advance beyond that accounted for by their level of education” (rest, 1979, p. 113). although rest’s taxonomy represents a chronological hierarchy, . . . “better” does not mean that a higher stage subject has more raw intelligence (brain power) . . . higher stages are said to be better conceptual tools for making sense out of the world and deriving guides for decision making. (rest & narvaez, 1994, p. 16) education levels were indicated by academic degrees and/or professional designations—or lack thereof. professional credentials are often acquired after first having completed some formal education, usually a bachelor’s degree. such credentials include certifications such as the cfp mark, cpa (certified public accountant), chfc (chartered financial consultant), and the like. indeed, one may not sit for the cfp certification examination if he has not completed abachelor’s degree. accordingly, “advanced educational achievement,” herein, was taken to mean the attainment of either advanced academic degrees (i.e., beyond a bachelors degree) and/or relevant professional designations—other than the cfp designation. career tenure may be positively correlated with higher ethical development levels, as may be expected in a developmental schema. while the literature does not address this variable in connection with investment planners, another rationale for this hypothesis was that over the course of an investment planner’s long career, production pressures wane thereby reducing ethical conflicts. gilligan (1977) made the point that gender plays an important role in ethical thinking. in response, rest asserted that both genders may nevertheless be compared on the same developmental scale, i.e., using the dit’s p-score (rest, 1986). in a study of veterinary medicine, it was found that “females are no less, and probably more, effective in the use of justice (emphasis added) for resolving moral conflicts” (self & olivarez, 1993, p. 70). (“justice” and “justice reasoning” are terms, which are used in connection with the higher stage philosophic/psychological principles employed in the kohlberg/rest taxonomy.) a study utilizing numerous versions of kohlberg’s test found that males and females earned equivalent scores (walker, 1984). additionally, a meta-analysis of 56 gender studies covering the use of the dit (thoma, 1984 as cited in rest, 1986) indicated that women scored higher than males. it has also been found that both female accounting students and female cpas had higher dit scores than males at equivalent levels (shaub, 1989, 1994). female accounting students were found to be less tolerant of academic dishonesty and less involved in related misconduct than their male counterparts (ameen et al., 1996). elsewhere, it has been found that “at every educational level, females scored slightly higher than males” (rest & narvaez, 1994, p. 14). 231k.s. bigel / financial services review 7 (1998) 223–236 the following directional hypotheses were formulated in connection with the aforementioned demographic variables: h3: investment planners with higher educational achievement, including either academic degrees and/or (non-cfp) advanced professional credentials, will manifest higher ethical development than those with less educational achievement. h4: ethical development increases with age. h5: ethical development increases with career tenure. h6: women score higher in ethical development than men. 7. results and analysis one-tailedt-tests for significance of independent means were conducted for each of the hypotheses. in connection with h1, and as expected, the mean p-score is statistically significantly greater for the cfp designee cohort than for the non-designee cohort (table 2). in connection with h2, three subgroups were involved: commission-only, fee-only, and combination planners. accordingly, a one-way analysis of variance (anova) was conducted initially in order to ascertain whether the groups differed significantly among themselves regarding the dependent variable, the p-score. the results indicate no statistical significance,f (169) 5 .42, p . .05. follow-upt-tests were also conducted in order to describe and more closely examine the subgroups. first, the fee-based planner was compared with the group consisting of both the commission-only and combination planners. although the mean p-score for the fee-based planner exceeded that for the commission and combination group (table 1) as postulated, the results failed to attain statistical significance (table 2). second, the commission-only planner was compared with the group consisting of both the table 2 statistical summary relative to hypotheses tests p-score finding mean difference of p-score t or f p h1 cfp. non-cfp 4.40 t 5 1.8 .037* h2 fee-only. comm. & combination 2.81 t 5 .87 .19 h2 fee-only & combo, commiss-only .33 t 5 .10 .46 h2 fee-only. commission-only 2.06 t 5 .49 .31 h3 adv ed. &/or prof. cred.. low ed. &/or no prof. cred. 5.47 t 5 2.2 .014* h4 low age . mid age . late age .11 and 1.76 f 5 .12 .89 h5 new , experienced. established .33 and 5.37 f 5 2.26 .11 h5 experienced. established 5.37 t 5 1.7 .04* h5 new & experienced. established 5.20 t 5 2.13 .035* (2-tailed) h6 female. male 4.78 t 5 1.5 .067 *p , .05 232 k.s. bigel / financial services review 7 (1998) 223–236 fee-only and combination planners. the commission-only planner, having greater incentive to generate commissions or to churn the client’s account, is thus presumed to be exposed to greater “moral hazard” (milgrom & roberts, 1997) and may be expected to manifest lower ethical development than the other cohorts. those who manifest higher ethical development may be repelled by the ethical conflict inherent in the commission business and therefore choose a fee-only or combination compensation arrangement. surprisingly, the p-score was slightly lower for the group consisting of both the fee-only and the combination (i.e., fee plus commission) financial planners, in comparison to the cohort consisting of commission-only brokers (table 1). the results, however, failed to attain statistical significance (table 2). this odd finding may, in part, be explained by the ambiguity of the “combination planner” variable. in the field, some “combination” planners may offer clients the choice of one or another compensation schedule, i.e. the choice of a feeor commission-based relationship. other combination planners may offer a “commission versus fee offset” plan, whereby the client’s fees will be reduced to the extent that the planner derived commission revenues from transactions, which were executed on the client’s behalf. some may offer either choice. this study did not attempt to elicit such fine differences among the combination planner subgroup. further, it is not clear how the various combination planner subgroups may compare, in terms of ethical development, to the fee-only and commission-only planners. therefore, a third test was conducted, in which the fee-only planners were compared to the commission-only planner cohort. as expected, the fee-only planner manifested a higher p-score than the commission-only planner (table 1); however, the results failed to attain statistical significance (table 2). in order to test the h3, subjects were divided into two subgroups: 1. those having neither any professional credentials (e.g., cpa, cfp, clu, etc.), nor an academic degree beyond the bachelors and 2. those having either a masters degree or more, or an advanced professional credential—other than the cfp designation. as expected, those with either advanced professional credentials and/or higher formal education reflected greater mean p-scores (table 1). the results were statistically significant, (table 2). further, among all cfp designees surveyed, 68% had “high educational achievement.” the association between cfp professionalization and educational achievement was statistically significant,x2 (1, n 5 170)5 4.217,p 5 .040. a comparison of the combination of the professionalization and educational achievement variables revealed that investment planners who both carry the cfp designation and had high educational achievement exhibited the highest mean p-scores as compared to other possible combinations (table 3). as p-scores were hypothesized as increasing with both professionalization and educational achievement as separate variables, a one-tailed test for differences in independent means of cfp designees alone based on educational achievement was conducted; the mean p-score difference attained statistical significance, (table 3). a test for statistical significance of non-cfp designees based on educational achievement failed to attain statistical significance. in order to test the h4, subjects were grouped into three cohorts: (1) “low age,” i.e., those who are under 40; (2) “mid age,” i.e., those between 40 and 59; and (3) “late age,” i.e., those 60 and over. contrary to expectation, the mean p-score declined as age increased. the 233k.s. bigel / financial services review 7 (1998) 223–236 decline in mean p-score from low age to mid age was marginal (table 1), while the decline was greater to late age. the results were not statistically significant (table 2). in order to test the h5, subjects were divided into three cohorts: (1) “new in profession,” i.e., those with less than 5 years’ experience; (2) “experienced,” i.e. those with 5–10 years’ experience; and (3) “established,” i.e., those with more than 10 years’ experience. as expected, the mean p-score increased somewhat from new to experienced (table 1), while contrary to expectation, the mean declined markedly thereafter to established. however, the results were not statistically significant, (table 2). a follow-up test did uncover a statistically significant difference between experienced planners and established planners (table 2); again, this finding is contrary to the hypothesized direction. the new and experienced cohorts were combined and compared to the established cohort. in combining the groups, the number of new and experienced subjects was closer in total quantity to the number of established subjects (table 1). as the career tenure finding above had not been in the expected direction, a two-tailed test for differences of independent means was conducted; the mean difference in p-scores was found to be significant (table 2). a possible explanation for established planners’ manifesting lower p-scores is a “survivor effect” whereby planners with relatively higher moral reasoning abilities decide to leave the business after a time as a result of the disturbing and ever-present moral conflicts that are inherent in the practice of investment planning. alternatively, the more ethical investment planner may simply fail in the business after a time. in this connection and as indicated earlier, the cfp professional is perhaps closest characteristically to the cpa. ponemon (1990) studied the ethical development of accountants and auditors. he found that junior level cpas tended to acknowledge rules of conduct, while managers and partners (who, of course, have more career tenure) were more concerned about such pressing issues as litigation and their firm’s profits in connection with auditor role conflicts. in connection with the h6, the mean p-score was greater for females than for males (table 1), as expected. the finding, however, only approached statistical significance (table 2). it table 3 combined p-score statistics for professionalization (cfp vs. non-cfp) and educational achievement n mean std. dev. mean p-score ranking cfp high ed. ach. 67 40.00 16.62 1 low ed. ach. 32 33.85 17.23 3 non-cfp high ed. ach. 37 35.23 14.52 2 low ed. ach. 34 31.86 12.72 4 comparison of p-score data for cfp designees based on educational achievement p-score finding mean difference of p-score t p cfp 6.15 1.7 .046* *p , .05 234 k.s. bigel / financial services review 7 (1998) 223–236 is noted that there are only 30 women in this cell, therefore, the statistical power of this test is low. on a final note, armstrong (1984) found that the ethical development of accountants tended to be lower than for college students and the general population of college-educated adults. the mean p-scores for cpas (37.6) in armstrong’s study and cfp licensees (38.01) in this study were similar. it appeared therefore that cfp designees also manifest lower ethical development than the general population. 8. conclusions this research implied that if an investor is primarily concerned about the ethics, in contrast to the competence, of a prospective investment planner, the client should look for certain characteristics in the planner. this client’s ideal investment planner would be a cfp designee with a high level of education (i.e., he would have at least a master’s degree and/or other professional credentials); he would also have less than ten years tenure as an investment planner and be female. still, the investor should expect his investment planner to be lower in ethical development than the general, college-educated population.caveat emptor! references abram, p. j. (1996). to fee or not to fee—that is not the question. in r. n. veres (ed.),the next agenda(p. 25). london, uk: millennium press ltd. ameen, e. c., guffey, d. m., & mcmillan, j. j. (1996). gender differences in determining the ethical sensitivity of future accounting professionals.the journal of business ethicsmay, 591–597. armstrong, m. b. (1984).internalization of the professional ethic by certified public accountants: a multidimensional scaling approach. unpublished doctoral dissertation, university of southern california. arnold, d., & ponemon, l. (1987). moral judgement perspective for the various auditing judgement issues. proceedings of the british accounting associationapril, 47–57. arnold, d., & ponemon, l. (1991). internal auditors’ perceptions of whistle-blowing and the influence of moral reasoning: an experiment.auditing: a journal of practice and theoryfall, 1–115. brown, s. (1996). churning: excessive trading in retail securities accounts.financial services review5(1), 43–56. fulmer, w. e., & cargile, b. r. (1987). ethical perceptions of accounting students: does exposure to a code of professional ethics help?issues in accounting education 23,207–217. gilligan, c. (1977) in a different voice: women’s conceptions of the self and of morality.harvard educational review47, 481–517. hansard, s. (1999). big firms seeking to make their mark through cfps.investment news. april 19. johnson c. j., & mclaughlin j. (1997).corporate finance and the securities laws. new york: aspen law and business. jones, s., & ponemon, l. (1993). a comment on multidimensional analysis of selected ethical issues in accounting.the accounting reviewapril, 411–416. king, a., & chironna, j. (1989). fraudulent financial reporting: raise the ante.management accountingmarch, 27–29. kohlberg, l. (1981).the meaning and measurement of moral development.worcester, am: clark university press. kohlberg, l. (1984).psychology of moral development. san francisco, ca: harper & row. 235k.s. bigel / financial services review 7 (1998) 223–236 loeb, s. e. (1971). a survey of ethical behavior in the accounting profession.journal of accounting research 9(3), 287–306. loss, l., & seligman, j. (1989).securities regulation(vol. viii, 3rd ed.). boston, ma: little, brown, & co. mccarthy, i. n. (1993).professional ethics code conflict situations: ethical and value orientation of collegiate accounting students. unpublished doctoral dissertation, new york university. milgrom, p., & roberts, j. (1997). moral hazard and performance incentives.economics, management, and organization. nasd. (1998). nasd manual [online]. available: http://www.nasdr.com. nelson, s. l., & nelson, t. r. (1997). the use of professional designations in the real estate industry.financial services review 6(2), 109–124. oppenheim, a. n. (1966).questionnaire design and attitude measurement. new york: basic books. ponemon, l. (1990). ethical judgements in accounting: a cognitive-developmental perspective.critical perspectives on accounting 12,191–215. ponemon, l., & gabhart, d. (1990). auditor independence judgements: a cognitive developmental model and experimental evidence.contemporary accounting research 7,227–251. ponemon, l., & gabhart, d. (1993).ethical reasoning in accounting and auditing. vancouver, canada: canadian general accountants’ research foundation. rest, j. r. (1979).development in judging moral issues. minneapolis: university of minnesota press. rest, j. r. (1986).moral development: advances in research and theory. new york: praeger. rest, j. r., & narvaez, d. (1994).moral development in the professions: psychology and applied ethics. hillsdale, nj: erlbaum. robaton, a. (1997). no. 1 group puts stamp on cfp mark.investment newsoctober 6, p. 6. self, d. j., & baldwin, d. c. (1994). moral reasoning in medicine. in j. r. rest & d. narvaez (eds.),moral development in the professions. hillsdale, nj: erlbaum. self, d. j., baldwin, jr., d. c., wolinsky, f. d., & shadduck, j. a. (1993). further exploration of the relationship between veterinary medical education and moral development.journal of veterinary medical education 20(3), 140–147. self, d. j., & olivarez, m. (1993). the influence of gender on conflicts of interest in the allocation of limited critical care resources: justice vs. care.journal of critical care 8(1), 64–74. self, d. j., olivarez, m., & baldwin, jr., d. c. (1994). moral reasoning in veterinary medicine. in j. r. rest & d. narvaez (eds.),moral development in the professions(pp. 147–162). hillsdale, nj: erlbaum. sellers, j. h. (1979). ethical perceptions of accounting students.collected papers of the american association’s annual meetingaugust, 635–644. shaub, m. (1989).an empirical examination of the determinants of auditors’ ethical sensitivity. unpublished doctoral dissertation. texas technological university-lubbock. shaub, m. (1994). an analysis of factors affecting the cognitive moral development of auditors and auditing students.journal of accounting educationspring, 1–24. spiro, l. n., & schroeder, m. (1995). can you trust your broker?businessweekfebruary 20. trevino, l. k., & nelson, k. a. (1995).managing business ethics: straight talk about how to do it right. new york: wiley. walker, l. j. (1984). sex difference in the development of moral reasoning: a critical review.child development 55, 677–691. westbrook, t. (1994).tracking the moral development of journalists: a look at them and their work(j. r. rest & d. narvaez, eds.). hillsdale, nj: erlbaum. 236 k.s. bigel / financial services review 7 (1998) 223–236 pii: 1057-0810(95)90002-0 from the editor lewis mandell this issue of financial services review completes the fourth year of publication. it is also my last edition as editor. the new editor is professor karen lahey of the university of akron. the change in editorship coincides with the change in my own career since i have assumed the role of dean of college of business administration at marquette university in milwau kee, wisconsin. when financial services review was started by the academy of financial services, the primary goal was to have an outlet for the dissemination of high quality research in the area of individual financial management. given the proliferation of new journals, cutbacks in library funding, and professional uncertainty of the “seriousness” of individual financial management, the success of this endeavor was by no means assured. indeed, this journal has suffered through most of its existence from a shortage of articles deemed by our editors to be of sufficiently high quality. but now, eight issues after our beginning, we appear to have achieved the respect of authors as well as the academic administrators who judge the quality of their work. the current issue consists of six articles representing three component areas of individ ual financial management. reflecting the interests of both authors and readers, half are in investing while two relate to consumer credit and one relates to disability income insurance. the lead article, “fund closings as a signal to investors: investment performance of open-end mutual funds that close to new shareholders,” was written by professors smaby and fizel. they find that funds that close to new shareholders tend to do so following a period of greater-than-average returns. the announcement effect of an impending closing tends to result in extraordinary sales of fund shares, but the performance after the fund closes tends to be far less than when the fund was open. the bottom line for individual investors is to be aware of fund closings because they may be nothing more than a sales device to take advantage of a past period of unusually good returns. a second article, by professors benson, rystrom, and smersh, examines “commission motivated trading patterns of brokers across the production month.” perhaps no surprise to readers who find that unsolicited calls from securities brokers tend to come at the end of the month, the authors found that a quarter of brokers in their sample earned a significantly higher portion of monthly commissions in the last five days of the month than one would expect if their customers had traded in a uniform fashion across the month. the authors are vii financial services review 4(2) 1995 concerned that broker-induced trading at the end of the month may contribute more to the welfare of brokers than their clients. a third investment-related article was written by professor knight and myself and is titled “optimal holding period for assets that must be liquidated: a certainty equivalent wealth approach.” this article examines a commonly asked question in a sophisticated manner and comes up with a counterintuitive answer. the question is: “if i put money into growth stocks for my child’s education in 15 years or my retirement in 20 years, how many years before that event do i move the assets into a more secure form?” by using a certainty equivalent wealth approach that looks at the utility function of investors across the risk preference spectrum, the answer comes up “zero.” unless an investor’s risk preference changes, utility is maximized by holding an appropriately balanced portfolio for the entire holding period. professors cox and gustavson examine “the market pricing of disability income insurance for individuals.” they find that disability income insurance appears to be sold in a competitive market and they find a strong relationship between prices and elimination periods supports the presence of adverse selection. the results suggest that buyers may need to be better informed about some of the pricing factors of disability income insurance. why are many credit card holders seemingly insensitive to interest rates charged on the cards? this phenomenon is examined by professors sullivan and worden in their article “credit cards and the option to default.” one possible reason relates to the unsecured nature of the credit card and the benefits of defaulting on such cards prior to declaring personal bankruptcy. they estimate the value of the option to default on unsecured credit card contracts and find that it is significantly affected by state and federal laws governing the collection practices of creditors and bankruptcy. in those states where it is easier to default, the default option is utilized more frequently. they also find that card holders who use their credit lines intensively before default are more likely to make a rational default decision which maximizes their benefits. the final article, by professors devaney and lytton, is titled “household insolvency: a review of household debt repayment, delinquency, and bankruptcy.” this article is one in an occasional series of review pieces designed to bring readers of financial services review up to date in the various components of their field. the paper explores issues related to the meaning and measurement of insolvency within the household. it reviews, among other things, predictive models and financial ratios as techniques for identifying insolvent households. at the end of the journal, as always, professor phyllis myers presents “abstracts of articles on individual financial management.” this feature of financial services review is designed to keep our readers informed of other related articles in scholarly journals. i would like to thank those who helped make this journal a success. the list begins with my managing editor, barbara poole, who also steps down with this issue to finish her dissertation. in addition, the editors, most of who have been with us since the inception, have been largely responsible for our success. of course, the reviewers, who have worked with the editors, have contributed a great deal to maintaining our high level of scholarship. thanks also go to the academy of financial services, which has lent solid support to the journal since its conception, and to the publisher, jai press, pii: s1057-0810(97)90006-8 financial services review, 6(4): 285-294 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. a simple and effective trading rule for individual investors laurie prather william j. bertin this study advocates the use of a simple trading strategy that is shown to outperform a passive buy and hold the market strategy. based on prior studies that find long-term stock price responses to economic news, the strategy utilizes the announcement of dis count rate changes to predict (and profit from) market movements. statistical testing indicates that the strategy correctly predicts market movements. furthermore, the results indicate that the proposed trading strategy produces higher risk adjusted returns than a buy and hold strategy. thus individual investors may reasonably expect to profit by following this easy to implement trading strategy. i. introduction many individual investors have a basic knowledge of financial economics in terms of understanding the inverse relationship between interest rates and security prices. yet these same investors may also be aware of efficiency issues regarding financial markets, which suggest that information is quickly reflected in security prices (thus negating any profit opportunities from trading based on interest rate changes). astute investors may feel frus trated in that they understand price and interest rate behavior but are unable to capitalize on their knowledge. the results of this study suggest that individual investors may in fact be able to realize above market returns and lower their risk exposure by following a trading strategy based on movements in the discount rate. the purpose of this study is to provide individual investors with a simple trading rule based on fundamental economic principles. using historical data, the investment strategy tested here is shown to be successful in predicting market trends and provides a method for reducing risk without sacrificing return. while some market strategists may seek greater sophistication and detailed development in their models, a key advantage of the successful strategy proposed here is its simplicity. the main limitation of this strategy is that historical market responses to discount rate changes may not continue, which could undermine its laurie prather and william j. berth l school of banking and finance, the university of new south wales, sydney 2052 australia. 286 financial services review 6(4) 1997 future effectiveness. the historical data used for testing the trading rule, however, encom passes a very long time period, and the rationale underlying the strategy is based on funda mental economic principles. the following discussion highlights the key features of the trading rule. the investment strategy examined here presumes that discount rate changes provide signals to enter and exit the stock market, providing the basis for a trading rule that relies on discount rate change announcements. the simple trading rule entails buying the market on an initial discount rate cut, remaining fully invested through any subsequent cuts and selling the market on an initial discount rate increase (and remaining out of the market through any subsequent increases). when not fully invested in the stock market, investors hold short-term treasury instruments (t-bills), which are held until a rate cut signal is received. this simple strategy may be easily executed by all types of investors ranging from the novice to the experienced. as with any decision rule, those attempting to benefit from the proposed strategy must follow it consistently in terms of both the buy and sell indicators. ii. literature review many economic studies consider stock price reaction to monetary policy changes. waud (1970) first documented this relationship by finding announcement effects where rate decreases (increases) are associated with positive (negative) stock market reaction. addi tionally, smirlock and yawitz (1985), hafer (1986), and hardouvelis (1987) analyze stock price changes surrounding discount rate announcements. they also cite evidence of a sig nificant announcement effect. in particular, they find that stock prices react swiftly to dis count rate changes suggesting that financial markets deem this type of information important. while most studies provide consistent evidence that monetary policy changes (and specifically discount rate changes) impact stock returns, some disagreement exists con cerning the speed of the adjustment process. examining economic news events pearce and roley (1985) find that the stock price response to new information may continue beyond the announcement day. focusing strictly on discount rate announcements, jensen and johnson (1995) analyze the long-term impact of rate changes on stock indices. they find that average stock returns are higher (lower) in those periods following discount rate decreases (increases). although their findings do not suggest causality between stock returns and interest rates, the mere notion of long-term market movements being associated with discount rate changes may provide a basis for a profitable investment strategy. the success of the proposed strategy demonstrates forecasting ability, yet a number of widely cited studies including treynor and mazuy (1966), jensen (1968), kon and jen (1978) and henriksson ( 1984) have generally concluded that mutual fund managers are not able to capitalize on directional changes in the stock market. reviewing these and other subsequent studies, ippolito (1993) argues that mutual fund performance net of expenses is comparable to that of the market and that competent funds clearly outperform the market while inept ones do not. competent investing may result from the use of sound economic reasoning with the possibility of beating the market also providing a basis for the proposed discount rate trading strategy. a simple and effective trading rule 281 a trading rule based on discount rate changes may be justified by the fact that these changes represent clear and unequivocal signals that the federal reserve (the fed) is com mitted to changing monetary conditions and business activity. a cut in rates suggests improved future cash flows to businesses, signaling an expected upward trend in the stock market, and is thus a buy indicator. by contrast, a rate increase reflects reduced future cash flows, indicating an expected down market trend, thus providing a sell signal. individual investors following the basic trading strategy are, therefore, abiding by fundamental prin ciples of financial economics and attempting to act in a manner consistent with those stud ies that find a long-term market response to economic news. to validate the trading rule, we use historical data to test and report the results of the investors’ strategy and compare it to a passive buy and hold strategy. the analysis of the proposed strategy does not consider transaction costs; however, as noted later in the paper, trading associated with the strategy is relatively infrequent, and the inclusion of these costs would not significantly alter the results. the future efficacy of the investors’ strategy will depend upon the continuation of documented long-term market reactions associated with discount rate changes. furthermore, the strategy must be consistently followed, since only partial adherence to it may result in foregone opportunities. iii. data and methodology discount rate changes and their impact on the stock market are analyzed over the period from april 7, 1933 to may 17, 1994. the dates for discount rate changes are published in the federal reserve bulletin, but we assume that investors gather information daily from the wall street journal (wsj) and then trade on the date that the discount rate change is published in the wsj. thus the investors do not capture the announcement day (day = 0) effects. in addition, the wsj publication date occasionally lags (by one day) the fed announcement date in those cases where the information is revealed before or during trad ing hours. thus the returns from the investors’ trading strategy are conservatively stated, if the discount rate changes impact the market as expected (i.e., rate cuts boost the market, while rate increases depress it). when discrepancies exist among the twelve federal reserve banks enacting a rate change, the initiating bank’s action published in the wsj is used as the definitive information source. table 1 contains the dates of the fed announced discount rate changes that initiated trading activity, the discount rate immediately preceeding a change, the new discount rate, the trading strategy implication and the dow jones industrial average (djia) on the trad ing date. the investors’ trading strategy is evident from table 1. for example, the trading strategy goes into effect on april 7, 1933 when the discount rate is cut from 3.5% to 3.0%. this initial cut, when reported in the wsj, causes investors to buy into the market. subse quent cuts in the 1930s and 1940s leave the investors’ strategy unchanged (in-market) for nearly fifteen years until january 10, 1948 when the fed raises the discount rate to 1.25% from 1.0%. investors then sell out of the market and hold treasury bills for the next six years (19481954). these holdings would be to maturity and include rollovers until febru ary 5, 1954 when the t-bills are sold, and the investor buys back into the market. the most recent signal examined here is a sell (the market), which occurred on may 17, 1994 when the fed raised the discount rate from 3.0% to 3.5%. 288 financial services review 6(4) 1997 table 1 dates of the federal reserve discount rate changes that initiate trading activity, discount rates before and after changes, trading signal and dow jones industrial average associated with discount rate changes. date of rate discount rate discount rate change reversals prior to change afer change s&d djia o&07/33 3.50 3.00 buy 58.78 01/10/48 1.00 1.25 sell 180.20 02mt54 2.00 3.75 buy 293.97 04/t 5155 i .50 1.75 sell 423.57 i i/15/57 3.50 3.00 buy 439.35 09/12/58 1.75 2.00 sell 519.43 06/10/60 4.00 3.50 buy 654.88 07117163 3.00 3.50 sell 699.72 04io7f67 4.50 4.00 buy 853.34 1 i i2of67 4.00 4.50 sell 857.78 11/13/70 6.00 5.75 buy 759.79 07/16/71 4.75 5.00 sell 888.51 1 l/19/71 5.00 4.15 buy 810.67 01115173 4.50 5.00 sell 1025.59 i 2mv74 8.00 7.75 buy 579.94 08/? 1177 5.25 5.75 sell 858.89 05/30180 13.05 12.00 buy 850.85 oqf26180 10.00 11.00 sell 940.10 1 l/02/81 34.00 13.00 buy 866.82 04toq/84 8.50 9.00 sell 1133.90 i l/21/84 9.00 8.50 buy 1220.30 09111187 5.50 6.00 sell 2549.27 iulqlqo 7.00 6.50 buy 2626.73 05117194 3.00 3.50 sell 3720.61 over the 62-year period covered in this study, 109 discount rate changes (54 rate reductions and 55 rate increases) have been implemented by the fed. following the dis count rate change trading rule, only 24 of the 109 changes have signalled either a buy or sell tr~sa~~on. from table 1, it follows that in-market periods constitute approximately 34 years of the 62-year period analyzed. treasury bills are held for the balance of the period with realized returns ranging between zero and 14.7% over the 62year period. the market proxy used for testing the investors’ trading strategy is the djia, given its availability on a daily basis dating back to 1933. the risk-free instrument held in out-mar ket periods is the one year treasury bill, which is easily liquidated when a rate cut (market buy signal) occurs. returns from the proposed trading activity are calculated i) using the holding period capital appreciation and dividend yield on the djia for the in-market peri ods and 2) the t-bill rate (appropriately adjusted for those occasions when the investor must sell the bills before matu~ty to reenter the market) for out-market periods. in addition to examining the entire 62-year period, we also analyze 12 five-year subperiods from 1935 to 1994. the returns from the investors’ trading strategy are compared to those of the buy and hold strategy, which consists of entering the market on april 7, 1933 and remaining fully invested through may 17, 1994. in addition to a direct comparison of aggregate returns, further analysis considers the differenti~ risks associated with the two strategies. recog a simple and eflective trading rule 289 nizing that investors who follow the proposed trading strategy hold risk-free securities for considerable periods of time during the 62-year study period, the risk of the investors’ strategy is substantially lower than that of the buy and hold. thus, the expected returns of the investors’ trading strategy should reflect the reduction in risk and be adjusted accord ingly. the capm is employed to make these risk adjustments and is expressed as follows: (1) where e(rpt) = the average rate of return on a portfolio resulting from the particular strategy over the specified time period, ‘ft = the risk free rate over that same period, ‘mt = the market return over that same period, bp = the portfolio beta. implementation of the investors’ strategy results in approximately 34 in-market years out of 62 total years (i.e., in the market 55% of the time). with an average risk free rate over the period of 3.87%, an average risk premium on the market of 8.23% (ibbotson & asso ciates, 1995) and a market beta equal to 1 .o, the capm required return for the investors’ trading strategy is 8.40% (from equation 1). this required return is well below the expected return for the buy and hold strategy of 12.10% (from equation 1). both strategies’ actual or realized returns may be evaluated in terms of their capm excess returns where the excess returns from each strategy are calculated by netting out the capm required returns from the actual returns. further risk analyses and comparisons of the buy and hold returns and the investors’ strategy returns utilize treynor’s measure, which we calculate for both strategies. trey nor’s measure, t, represents the return premia per unit of risk and is defined as follows: where rpt, rand & are defined above. if the value of t for a given portfolio is greater than the value of t for the market in equi librium, the portfolio lies above the security market line (i.e., is undervalued). finally, in addition to the above analysis, we employ a nonparametric analysis from merton (198 1) and henriksson and merton (198 1) to test the investor’s trading strategy proposed in this paper. according to the basic model, a forecaster predicts when stocks (treasury bills) will outperform treasury bills (stocks). the specific test examines a null hypothesis that discount rate changes do not provide information that leads to correctly forecasting stock market trends and is denoted as follows: ffo:plw+pzw= 19 where pl(t) = the conditional probability of predicting an up market when an up mar ket occurs and, ~z(t) = the conditional probability of predicting a down market when a down market occurs. 290 financial services review 6(4) 1997 henriksson and met-ton use the hypergeometric distribution (equation 4) to calculate the probability that a given outcome from a sample comes from a population that satisfies the null hypothesis (no forecasting ability). pn, = x(\n,n,, n) = n 0 n (4) where n, = n2 = n = nl = n2 = n = the number of observations where treasury bill returns are greater than market returns, the number of observations where market returns are greater than treasury bill returns, total number of observations (nl + n2), number of successful predictions given treasury bill returns greater than market returns, number of unsuccessful predictions given market returns greater than trea sury bill returns, number of forecasts predicting treasury bill returns greater than market returns. under the henriksson and merton framework, the conditional probability of a forecast does not depend on the magnitude of subsequent realized returns. this gives the henriks son and merton analysis a significant advantage in that it does not require the specification of a particular equilibrium model of returns as do the capm and the treynor analyses pre sented above. for the strategy proposed in this paper, the null hypothesis is that changes in the dis count rate do not forecast the direction of the stock market. rejection of the null hypothesis would be evidence of a successful strategy. the hypergeometric distribution equation is used to calculate the probability of obtaining at least the observed number of correct down market period forecasts under the null hypothesis (no forecasting ability) given the total number of observations, the total number of up-market period observations and the number of forecasted down-market periods. iii. results table 2 provides descriptive information on the investors’ trading strategy including the average one-year treasury bill rate, the number of years, as well as the percentage of time in the market, and an approximate beta for the overall study period and the 12 subperiods. the beta measurement for the investors’ portfolio simply reflects the percentage of the par ticular investment horizon in the market and represents a close approximation for the regression beta estimates. these beta measurements illustrate that the investors’ strategy is less risky than the buy and hold strategy with investors in the market only 55 % of the time. a simple and effective trading rule 291 table 3 reports the annual returns, capm required returns, excess returns and the treynor measures for the buy and hold strategy and the investors’ trading strategy over the entire period and the 12 subperiods. from a long term perspective (1933 to 1994), the aver age annual return from the trading strategy is approximately the same as the buy and hold strategy (11.72% for the trading strategy and 12.10% for the buy and hold). analyses of several subperiods reveal that in seven out of the 12 subperiod cases the investors’ strategy returns are equal to or exceed the buy and hold returns. after considering risk through use of the treynor measures, the trading strategy outperforms the buy and hold strategy in all subperiods except one (the early 1960s). using the wilcoxon sign rank test (one-tail) we can reject the null hypothesis that the two sample populations have identical probability distributions in favor of the alternative hypothesis that the distribution of treynor measures for the trading strategy is shifted to the right of that for the buy and hold population at the 99.5% significance level. these findings of the strategy’s superiority are consistent with the studies of pearce and roley (1985) and jensen and johnson (1995) who report a long term stock price response from certain types of economic news. consistent with a priori expectations, the excess return for the buy and hold strategy for the entire study period is zero. alternatively, table 3 reveals that the excess returns for the buy and hold strategy are negative in all but five of the subperiods examined. the diver gence of the excess returns from zero in the subperiod analyses of the buy and hold strategy arises from the use of an average market risk premium of 8.23%. this average risk pre mium is based on historical information from 1933 through 1994. alternatively, the risk adjusted excess return for the trading strategy is positive for the overall period (3.32%) and negative in only three of the 12 subperiods (the early 1940s and the 1960s). based on the wilcoxon sign rank test, we can again reject the null hypothesis that the population median table 2 descriptive information on the investors’ trading strategy this table contains the average risk-free rate, the number of years in the market, the percentage of time and an approximate beta for each time period. average number of years percentage of the approximate time period risk-free rate in the market period in the market beta entire period 1933-1994 3.87% 34.09 55% 0.55 five year subperiods: 1935-1939 0.14 5.00 100% 1 .oo 1940-1944 0.20 5.00 100% 1.00 1945-1949 0.62 3.04 61% 0.61 1950-1954 1.42 0.66 13% 0.13 1955-1959 2.34 1.08 23% 0.23 1960-1964 2.82 3.04 61% 0.61 19651969 4.94 0.63 13% 0.13 19701974 5.92 1.92 38% 0.38 1975-1979 6.72 2.67 53% 0.53 1980-1984 il.00 2.92 58% 0.58 1985-1989 6.82 2.71 54% 0.54 iwo1994’ 4.99 3.50 79% 0.79 nofe: ‘this period is 4.38 years: based on a buy signal december 19, 1990 and a sell signal may 17. 1994 which ends the study period. t a b l e 3 r is kfr ee r at es , r et ur ns a nd t re yn or m ea su re s t hi s ta bl e co nt ai ns a ct ua l, c a p m r eq ui re d an d ex ce ss r et ur ns a nd th e t re yn or m ea su re s fo r th e bu y an d ho ld a nd th e in ve st or s’ t ra di ng s tr at eg ie s fo r ea ch t im e pe ri od ( al l r et ur n fi gu re s ar e av er ag e an nu al iz ed r et ur ns ). b u y an d h o l d s t r a t e g y w is c o u n t r a t e c h a n g e st r a t e g y t im e a ct ua l r eq ui re d e xc es s t re yn or a ct ua l r eq ui re d e xc es s t re yn or p er io d r et ur n r et ur n( c a p m ) r et ur n m ea su re r et ur n r et um (c a p m ) r et ur n m ea su re e nt ir e pe ri od 19 33 -1 99 4 12 .1 0% 12 .1 0% 0. 00 % 8. 23 i 1 .7 2% 8. 40 % 3. 32 % 14 .2 7 fi ve y ea r su bp er id s: 19 35 -1 93 9 12 .1 8 8. 37 3. 81 12 .0 4 12 .1 8 8. 37 3. 81 12 .0 4 19 40 -1 94 4 5. 60 8. 52 -2 .9 2 5. 40 5. 60 8. 52 -2 .9 2 5. 40 19 45 -1 94 9 11 .0 2 8. 85 2. 17 10 .4 0 7. 77 5. 62 2. 51 11 .7 2 19 50 -1 95 4 21 .6 1 9. 65 11 .9 6 20 .1 9 8. 40 2. 49 5. 91 53 .6 9 19 55 -1 95 9 16 .0 3 10 .5 7 5. 46 13 .6 9 9. 07 4. 23 4. 84 29 .2 6 19 60 -1 96 4 9. 75 11 .0 5 -1 .3 0 6. 93 4. 79 7. 84 -3 .0 5 3. 23 19 65 1 96 9 2. 14 13 .1 7 -1 0. 43 -2 .2 0 4. 17 6. 01 -1 .2 4 -1 .3 1 19 70 .1 97 4 0. 10 14 .1 5 -1 4. 05 -5 .8 2 13 .9 2 9. 05 4. 87 21 .0 5 19 75 -1 97 9 13 .5 3 14 .9 5 -1 .4 2 6. 81 14 .5 0 11 .0 8 3. 42 14 .6 8 19 80 -1 98 4 14 .1 1 19 .2 3 -5 .1 2 3. 11 16 .8 8 15 .7 7 1. 11 10 .1 4 19 85 -1 98 9 22 .6 7 15 .0 5 7. 62 15 .8 5 22 .1 3 11 .2 6 10 .8 7 29 .3 5 19 90 1 99 4* 11 .2 6 13 .2 2 -1 .9 6 6. 27 13 .0 1 il .4 9 1. 52 10 .1 5 n o ic * fo r 1 99 4 th e se ll si gn al is o n m ay 17 , t hu s th e re tu rn i s ad ju st ed to r ef le ct th e pa rt ia l ye ar ( .3 8 ye ar s) . a simple and effective trading rule 293 of excess returns for the trading strategy is equal to zero in favor of the alternative hypoth esis that the population median of excess returns is shifted to the right (i.e., greater than zero) at the 97.5% confidence level. alternatively, the null hypothesis of a population median equal to zero cannot be rejected for the buy and hold strategy at any meaningful significance level. implementation of the trading strategy triggered 24 trades (23 intervals for market trend forecasts). over the entire period analyzed there were 12 up-market and 11 down market periods forecasted. actual occurrences consisted of 17 up-market periods and 6 down-market periods. thus, based on the conditional probabilities and the hypergeometric distribution equation, we calculate the probability of predicting 6 of 6 down-market trends if the null hypothesis is true as follows: pi(t) + p2(t) = 12/17 + 616 = 1.7059 p(n,) = 61(23, 17,6) = (3(1,” 6) = .oo46 based on these calculations we can reject the null hypothesis of no forecasting ability (pi(t) + p*(t) = 1) at a confidence level of 99.5%. the overall results suggest that investors may not only profit, but also reduce risk exposure by following the proposed strategy. the recommendation, of course, is subject to the caveat that the historical market responses to discount rate changes continue in essen tially the same direction and magnitude in the future and that investors consistently follow the strategy. additionally, the full benefits of the proposed strategy are realized when trad ing costs, consisting of taxes, commissions and other fees, are zero (i.e., retirement accounts that trade costlessly and on a tax deferred basis). even when trading costs are con sidered they do not significantly reduce the returns or alter the basic comparisons reported in table 3. for example, if round trip transaction costs of one percent were applicable, the returns from the dr strategy would be reduced by approximately 0.39% per year on aver age. this reflects the fact that trading associated with the proposed strategy is relatively infrequent (only 24 trades over 62 years). iv. summary and conclusions this paper examines a trading strategy for individual investors based on fed announce ments of discount rate changes. the strategy keeps the investor in the market (out of the market) during periods of declining (rising) interest rates. during periods of rising interest rates, the investor holds treasury bills. this strategy is evaluated relative to a buy and hold (the market) strategy. consistent with previous studies indicating a long-term relationship between stock prices and economic news, the results of this study suggest that it may be possible for indi vidual investors to capitalize on discount rate changes. in particular, the results indicate 294 financial services review 6(4) 1997 that the investors’ trading strategy produces higher risk adjusted returns than a buy and hold strategy. these findings are especially interesting in that many individuals invest in mutua1 funds, and yet most fund managers do not, on average, beat the market. the strat egy advocated in this study should be attractive to individual investors since it eliminates the need for professional management, security analysis and/or selection, but still provides investors with above market returns and beiow market risk. ackrowl~grn~n~ we would like to acknowledge the helpful comments of two anony mous reviewers and karen lahey, the editor. references hafer, r.w. (1986). the response of stock prices to changes in weekly money and the discount rate. frb st. louis review, 68(3), 5-14. hardouvelis, ga. (1987). macroeconomic information and stock prices. journal of economics and btrsiness, 39(2), 13 l-140. henriksson, r. (1984). market timing and mutual fund performance: an empirical investigation. jmrnal of business, 57,73-96. henriks~n, rd., & merton, r.( 198 1). on market timing and investment pe~o~ance. ii. statistical procedures for evaluating forecasting skill. journal ofbusiness, 54(4), 513-533. ibbotson & associates. (1996). stocks, bonds, bills and inflation 1995 yearbook. chicago: ippolito, r. (1993). on studies of mutual fund performance, 1962-1991. financial analysts journal, ( jan/feb), 42-50. jensen, m. (1968). the performance of the mutual funds in the period 195464. journal of finance, 23(may), 384-4 16. jensen, g.r., & johnson, r. (1995). discount rate changes and security returns in the u.s., 1962 199 1. journal of banking and finunce, 19,79-95. kon, s.j., & jen, f.c. (1978). the investment performance of mutual funds: an empirical investiga tion of timing, selectivity. and market efficiency. journal of business, 52(2), 263-289. merton, r. (198 1). on market timing and investment performance. i. an equilibrium theory of value for market forecasts. journal o~b~s~ne.~s, 54(3), 363-406. pearce, d., & roley, v. (1985). stock prices and economic news. journal of business, 58( 1), 49-67. smirlock, m., yawitz, j. (1985). asset returns, discount rate changes, and market efficiency. journal of finance, 40(4), 1141-i 158. treynor, j. & mazuy, k. (1966). can mutual funds outguess the market? hazard bminess review, (julylaug). 131-136. waud, r., (1970). public inte~retation of federal reserve discount rate changes: evidence on the ~nouncement effect. econo~etricu, 38(2), 23 l-250 pii: 1057-0810(95)90007-1 financial services review, 4(2): 123-136 copyright q 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. credit cards and the option to default a. charlene sullivan debra drecnik worden the value of the option to default on unsecured credit contracts is estimated and found to be signij?cantly impacted by state and federal laws governing creditors’ collection practices and bankruptcy. the data suggest that the expected value of the option to default influences debtors’ choices in default and is correlated with their use of their credit cards before default. cardholders who use their lines of credit very intensely before default are signtjiiantly more likely to make choices in default which allow them to realize a greater benefit from default. furthermore, these results offer a possible explanation for consumers’ seeming insensitivity to interest rates charged on revolving lines of credit. i. introduction in the household credit arena, two recent phenomena reflecting changes in debtor behavior attracted policy-level attention. the first is the dramatic growth in unsecured credit outstanding associated with the use of third-party credit cards. revolving credit outstanding increased from about 18% of total consumer installment credit in 1980 to 35% in 1993. during that period, consumer credit outstanding increased from about $250 billion to $770 billion. while third-party credit cards had been available from the middle to late 196os, in 1970 only 16% of u.s. households had a bank card. by 1989,54% of households had at least one bank card. this dramatic increase in the ownership and use of unsecured revolving credit lines has been attributed to multiple factors, including the elimination of most state usury ceilings, the convenience of small amounts of immediately accessible consumer credit, and a surge in the number of households in the prime debt-use phase of the life cycle (canner & luckett, 1992). the public policy concern was focused on the pricing of revolving credit, specifically, concerns about whether the market for third party credit cards was sufficiently competitive (ausebel, 1991; government accounting office, 1994). the fair credit and charge card disclosure act, effective in 1988, was designed to strengthen mandatory disclosure for credit cards, reducing consumers’ search and switching costs. an analysis of a. charlene sulliian l purdue university, 217 kraunert center, west lafayette, in 47907. debra drecnik worden l george fox college, 414 n. meridian street, newberg, or 97132. i24 financial services review 4(2) 1995 the average profitability of credit card banks between 1988 and 1993 showed that spreads and profits on credit card operations have remained high relative to other bank products. shaffer (1994) concluded that the act had not increased the degree of price competition in the market for credit cards. the second phenomenon is the surge in the annual number of non-business estates filmg for protection from creditors in bankruptcy court-from about 200,000 in 1980 to approxi mately l,ooo,ooo in 1993. historical trends in personal bankruptcy show that periods of rapid increase coincide with periods of economic contraction. the sharp increase in the incidence of personal bankruptcy during the economic expansion period that started in november 1982 and extended into 1990 was, thus, inconsistent with historical patterns (luckett, 1988; canner & luckett, 1991). explanations offered for the latter development include a change in the federal bankruptcy code that reduced the private cost of bankruptcy. one of the aims of the bankruptcy reform act of 1978 was to establish a federal standard for the value of personal assets exempted from liquidation under chapter 7 of the u.s. bankruptcy code. the act also expanded eligibility for chapter 13 of the code to practically all non-business debtors.’ a more recent reform, the bankruptcy reform act of 1994, further amended the bankruptcy code to make chapter 13 a more attractive and accessible choice in bankruptcy. canner and luckett (199 1) suggested that the 1978 reform made bankruptcy a more attractive option for troubled debtors relative to outright default or informal renego tiation with creditors. they argued that the “rise in bankruptcy may be reflecting to a significant degree a shift in the way households choose to respond to debt problems rather than an increase in debt problems per se.“the public policy concern was whether bankruptcy relief had become “too easy,” creating a high social cost which, in turn, would increase the cost of and reduce the availability of unsecured household credit (credit research center, 1982). this paper investigates a dimension of the possible relationship between the two phenomena. specifically, the question addressed is whether there is evidence that consumers who use credit cards aggressively and then default prefer filing bankruptcy to other methods of responding to debt problems. others have alluded to this relationship. analyzing petition data, demowitz claimed “an increase in credit card debt relative to income is one of the strongest contributors to the probability of filing for bankruptcy” (1992, p. 16). offering insight into consumer behavior, pozdena used option theory to conclude that cardholders have an incentive to take risks in their credit-use behavior when using credit cards. pozdena argued that “the high default potential of credit card loans means that their value depends more on the implicit default option than on the prevailing time value of money” (1991, p. 17). furthermore, less creditworthy households may self-select and use credit cards more intensely because of the valuable default option, made more so by bankruptcy reform. this potential for moral hazard, if it exists, creates an interesting risk management/pricing problem for credit card portfolio managers. the next section of this paper discusses the conditions that determine the value of the default option. the third section summarizes related empirical literature. section iv presents summary statistics of the data analyzed in the paper, and section v presents empirical evidence of the relationships between debtors’ behavior before and after default and the value of their option to default. the last section considers the implications of these results. credit cards and the option to default 125 ii. the option to default a debtor with an unsecured line of credit has the ability, at any time, to default on the contract. the debtor has an option to “buy” the outstanding credit obligation from the creditor at a cost equal to the amount of cash the debtor voluntarily repays or that the creditor is able to collect through various legally defined collection remedies. these remedies include actions against the assets or future income of the cardholder. in addition to any cash voluntarily forfeited or recovered through the creditor’s collection actions, the debtor bears other costs of default such as damage to his or her credit rating and reduced access to unsecured credit in the future. as credit records are considered not only in the extension of future credit but also in applications to rent a home, buy insurance, or seek employment, these costs should be nontrivial in the debtor’s decision process. the difference between the amount of credit outstanding on the defaulted line of credit and the present value of all expected cash and reputation costs of default is the value of the default option held by the debtor for each unsecured credit line. a. creditor’s rights and the debtor’s choice a priori, the expected value of the debtor’s option to default on an unsecured credit obligation is a function of the collection efforts the creditor is expected to pursue. these efforts are, in turn, a function of the manner in which the debtor chooses to default. it is the debtor’s choice to default outright or to file for bankruptcy. furthermore, it is the debtor’s choice to file under chapter 7 or chapter 13 of the u.s. bankruptcy code. once the debtor has made these choices, then state and federal laws defining creditors’ legal access to the debtor’s assets or future income, and the actions in which the creditor chooses to engage, ultimately determine the value of the default option. b. default and wage garnishment debtors in trouble typically default outright on their unsecured credit obligations after an extended period of delinquency, an action which is recognized by the formal charging off of the account by the creditor. following charge off, the creditor may engage in various activities to collect the account balance, including the legal attachment of the debtor’s future earnings through garnishment. in garnishment, the creditor applies for and receives court approval to attach a portion of the debtor’s paycheck to retire the debt. the availability of garnishment and the amount of wages that may be garnished are conditions established by federal and state law and vary somewhat across states. the consumer credit protection act is a federal law that provides guidelines restricting the extent to which a debtor’s wages may be garnished to satisfy consumer credit claims. some states have rules that limit the garnishment of wages further, even to the point that it is prohibited for the repayment of consumer credit claims.2 garnishment is a powerful remedy and, when used, may substan tially reduce the expected value of the debtor’s default option. c. bankruptcy-chapter 7 versus chapter 13 instead of defaulting and risking wage garnishment, individuals may petition the bankruptcy court for protection from creditors’ collection actions and for access to processes for having unsecured debts reorganized or legally discharged. in the event that the debtor 126 financial services review 4(2) 1995 files bankruptcy, all collection efforts including, garnishment are automatically stayed by court rule. most types of unsecured debt are eligible for discharge in bankruptcy, with the exception of federal tax obligations, debts incurred for the payment of federal tax obligations, child support payments, and debts incurred by fraud or in anticipation of bankruptcy. in particular, credit card charges for luxury goods and services or cash advances in excess of $1,000 and incurred within 60 days of the bankruptcy filing are not dischargable. the bankruptcy code provides that unsecured debts not satisfied after the payout of the proceeds from liquidation of nonexempt assets in chapter 7 may be discharged. unsecured debts not covered by a threeto five-year repayment plan based on the debtor’s anticipated disposable income will be discharged in chapter 13. the extent to which the debtor’s assets are accessible to satisfy unsecured creditors’ claims in bankruptcy is determined by the federal bankruptcy law, unless a state passes overriding legislation to uniquely define allowed exemptions. more than half of the states (35) have passed overriding legislation, with some having defined asset exemption allow ances that are less generous than that provided by the u.s. bankruptcy code.3 holding other things constant, the expected value of the default option would be highest in those states where a high value of personal and household assets is exempt from liquidation. in a chapter 13 case, the bankruptcy code provides that the amount repaid to unsecured creditors under the proposed repayment plan must be at least as great as would be possible if the debtor’s nonexempt assets were liquidated in chapter 7. this equivalency rule theoretically makes the debtor with nonexempt assets indifferent between chapter 7 and chapter 13, holding other conditions constant. traditionally, more than 70% of non-business petitions are filed under chapter 7. as a reflection of the low liquidation value of most household assets and liberal asset exemption allowances, most petitions filed under chapter 7 are “no asset” cases. consequently, the unsecured creditors for the typical petitioner will recover nothing and the debtor gains the value of total discharge of accumulated unsecured debts. iii. literature review the behavioral implications for debtors and creditors of the value of the default option as it is defined by rules for garnishment and bankruptcy have been investigated in studies of the cross-state variability in bankruptcy rates. variations in legal conditions that influence the expected value of the option to default are expected to be systematically related to variations in the probability that debtors would exercise the default option. this would hold unless, in anticipation of debtors’ opportunistic actions, creditors make offsetting adjust ments in credit standards. apilado, dauten, and smith (1978) published the first study that related cross-state variation in bankruptcy rates (non-business petitions per 100,000 popu lation) and legal conditions for garnishment and bankruptcy. using data from 1963 through 1974, they found that the relationship between the level of asset exemptions and bankruptcy rates was nonlinear. the rate was low in states with low asset exemptions as well as in states with high asset exemptions. states restricting or prohibiting garnishment had a significantly lower bankruptcy rate than did states that used the federal guidelines for garnishment. peterson and aoki (1984) and shephard (1984) found that the level of allowed exemption was not a critical factor for explaining cross-state variation in the rate of personal credit car& and the option to default 127 bankruptcy immediately after the enactment of the bankruptcy reform act of 1978. but in states where garnishment was prohibited, the bankruptcy rates were significantly lower both before and after the change in the federal bankruptcy code than was the case in states that allowed garnishment. shiers and williamson (1987) found that the bankruptcy rate was significantly higher in states with low asset exemptions. they attributed this to the possibility that creditors took more credit risk in those states where the value of the default option was low. they found no relationship between garnishment restrictions and the bankruptcy rate. however, their result could be attributable to a misspecification error in that the authors claimed that all states permitted garnishment to some extent and, therefore, did not include a variable for states that actually prohibited garnishment for purposes of recovering consumer credit claims. sullivan and worden (1991) found that between 1981 and 1988, the bankruptcy rate was significantly lower in states that restricted garnishment more than the consumer credit protection act. furthermore, they demonstrated that while the rate of bankruptcy was relatively low in some states with low asset exemptions, it was high in some low-exemption states. in those states, an above-average percentage of petitioners opted for chapter 13. this decision allowed them to protect their assets from liquidation. theoretically, however, this choice should not have produced a higher valued option because of the equivalency rule. in sum, the literature supports the hypothesis that the value of the option to default is influenced by regulations governing garnishment and bankruptcy and, correspondingly, influences debtor and creditor behavior. however, none of these studies included an assessment of consumer default that did not involve a bankruptcy filing. in this analysis, we study the impact of those conditions that influence the default option value on debtor’s response to debt problems. furthermore, the relationship between card use before charge-off and the value, ex post, of the debtor’s option to default on unsecured claims is investigated. specifically, we investigate the existence of moral hazard that was suggested by pozdena’s work. iv. descriptive statistics a. the data to evaluate the relationships between regulations and cardholder behavior before and after default, complete account histories of charged-off accounts from a national portfolio of credit card accounts were needed. a very large regional bank provided such a data base. the data analyzed are made up of account histories of a national sample of active bank credit card accounts in the portfolio of a single midwestern bank and a subsample of the portfolio which includes the entire population of accounts that had been open prior to 1990 and charged off in 1991. the data set included the complete account histories for both groups of accounts for 1990 and 199 1. because information about any activity in the account is retained even after charge-off, the data set enabled the reconstruction of all action in the charged-off accounts one year before and up to one year after charge-off. 128 financial services review 4(2) 1995 the samples included no gold or premium cards and no accounts specifically targeted at students. the portfolio sample included 3,094 active standard credit card accounts selected randomly from the portfolio. while these cardholders were residents of all states, with the exception of vermont, the majority (82%) resided in the midwest, followed by 13% from the south. the charge-off sample included 5,112 standard accounts, with 65% of cardholders residents of the midwest, followed by 24% from the south. many of the accounts in the portfolio sample and the charge-off data set were added to the bank’s portfolio through acquisition of entire portfolios of other financial institutions. thus, the two samples are representative of the populations of cardholders served by the banks involved in the transactions. while these distributions of accounts may not be representative of the u.s. population of cardholders, valid inferences may be drawn with regard to the potential relationships between variation in legal conditions and measures of cardholder behaviors. b. charge-off experience in the portfolio the bank was observing its normal charge-off and collection procedures in 1990 and 199 1, thus allowing observations based on those data to be representative of default behavior across time. the lender typically classified an account as “charged off’ after a standard period of time had elapsed with no payments having been received, when fraudulent card use was reported, or when a notice was received that a petition for bankruptcy had been filed by the cardholder or that the card holder had died. in the random sample of active standard accounts, 1.9% of accounts were charged off in 1991; excluding death and fraud as reasons for charge-off reduced that figure to 1.4% of active accounts. (all charge-off statistics in this analysis exclude those accounts charged off due to fraud or the death of the principal cardholder.) the initial charge-off action for 67.6% of the charged-off accounts was attributed to severe delinquency. about 9% of those accounts tiled for bankruptcy protection after the creditor started collection efforts. ultimately, 61.4% of accounts charged off in 199 1 were attributed to severe delinquency; 32.4% were attributed to a chapter 7 filing under the bankruptcy code, and 6.2% were bankruptcies filed under chapter 13. c. variability in legal conditions within the portfolio, the creditor was exposed to considerable state-by-state variability in garnishment rules and asset exemption allowances. in states that restricted garnishment more severely than the federal rule and where asset exemptions were set at or above the federal level, a pro debtor legal framework for default was said to have existed for purposes of the present analysis. in those states, creditors’ access to cash recovery is limited, making the expected value of the debtor’s default option comparatively high. about 40.6% of the charge-off accounts in the portfolio involved cardholders residing in pro debtor states (see table 1). about 36.3% of charge-offs in the portfolio originated in states with strict exemption allowances and the federal allowance for garnishment. these were characterized as pro creditor states, where the option to default would have comparatively less value, holding other things constant. about 23% of charge-off cases occurred in states that followed the federal guideline for garnishment but had high exemption allowances, characterized as mixed-message states in this study. this distribution was similar to that of the portfolio sample, as shown in table 1. credit cards and the option to default 129 table 1 accounts opened prior to 1990 total sample charge off sample legal conditions n = 3519 n = 5112 pro debtor 42.1% 40.6% mixed message 23.8 23.1 pro creditor 33.2 36.3 d. legal restrictions and charge-off rate the results of research cited earlier suggest that lenders may adjust their credit standards to manage imbalances in risk and expected return that may be created by legal conditions that influence the value of the debtors’ default option. the data in table 2 show that in this portfolio, the charge-off rate, or the number of charged-off accounts as a percentage of total active accounts, does not vary significantly across legal conditions. neither variation in garnishment rules nor in asset exemption allowances is statistically associated with variation in charge-off rates. in other words, legal conditions that were expected to have an influence on the borrowers’ incentive to default have been apparently offset by adjustments in credit underwriting standards, eliminating significant cross state variation in charge-off experience. however, the charge-off rates for the whole portfolio do not reveal the whole story concerning the value of the default option and its effect on the behavior of debtors. that story is revealed by examining defaulted debtors’ choices before and during charge-off. we hypothesize that debtors make choices in default that will maximize the expected value of their option to default. furthermore, we hypothesize that debtors’ pre-default card use behavior is correlated with the expected value of their default option; that is, cardholders who use their cards aggressively for credit make choices in default that produce a higher gross benefit, measured as the percentage of the credit line ultimately discharged. table 2 total sample all accounts charged off in 1991 default choice by year-end 1991 charge-off sample bankruptcy legal conditions pro debtor mixed message pro creditor overall charge off rate (%) 1.69% 0.68 1.52 x= = 3.7 1.4% delinqrtency 65.3% 59.5 58.3 61.4% ch. 7 30.0% 34.9 33.5 32.4% ch. 13 4.7% 5.6 8.2 x= = 35.s* 6.2% 130 financial services review 4(2) 1995 v. the results: debtor choices and the value of the default option a. legal regimes and debtors’ choices overall, the debtors charged off in 1991 were almost twice as likely to default outright as file bankruptcy (61.4% versus 38.6%) (see table 2). debtors in states providing greater protection of income and assets (pro debtor regime) were significantly more likely to default outright than was the case for debtors in states characterized as pro creditor. in pro creditor states, debtors in default were more likely to seek the protection of the bankruptcy process than was the case in pro debtor states (41.7% vs. 34.7%). a relatively high proportion of petitioners for bankruptcy filed under chapter 13 in pro creditor states (garnishment allowed; low asset exemption). a possible explanation is that under conditions where income may be garnished, debtors are more likely to file bankruptcy to protect future income, irrespective of exemption allowances. this result is inconsistent with the assumption that bankruptcy would be a lower-valued choice in states with strict restrictions on asset exemption allow ances. however, it does suggest that the bankruptcy code provides more protection for future income relative to the federal guidelines for garnishment-a factor that would be expected to be associated with an increase in the use of bankruptcy. a similar result is discussed by sullivan and worden (1991), who found that in some states with low asset exemptions, the incidence of chapter 13 was abnormally high, causing the overall rate of bankruptcy to be high in those states. in general, these data revealing the distribution of debtors’ choices suggest that debtors do make choices in default that are consistent with maximization of the expected value of their default option, given the legal conditions affecting their choices. this conclusion is further supported by the results (shown below) of an estimation of the value of the default option. b. debtors’ choice and the gross value of the default option an estimate of the value of the default option was calculated for those accounts that were charged off during the first six months of 1991 (table 3). this time frame was used to allow for the inclusion of repayments made in response to collection activities following charge off. this group of debtors had between six and 11 months to make payment on the charged-off account, either voluntarily or as a result of payments outlined in a chapter 13 repayment schedule. the creditor’s loss, after subsequent payments, was calculated as a percent of the credit line charged off for each account. what the creditor lost was estimated as being equal to what the debtor gained, which was the value of the option to default. the option to default had its highest value for chapter 7 bankruptcies, with an average value of 93.6% of the credit line, followed by an average value of 88.04% under chapter 13. for the delinquency charge off, the default option value was an average of 80.2% of the credit line. debtors realized the highest average gain from default in pro debtor states, even though debtors in those states were more likely to choose outright default than bankruptcy. however, the option value for each of the three default choices was highest in the pro debtor states, confirming the appropriateness of our label for those states. in the pro creditor states, debtors were more likely to use the courts for protection in default. however, on average, debtors in the pro creditor states did not benefit as much proportionately as those in the other credit cards and the option to default 131 table 3 mean percent credit line charged off between january-june 1991 (net of repayments by year-end 1991) bankruprcy legal conditions pro debtor mixed message pro creditor overall mean sample size delinquency 83.21% 82.72% 75.21% f = 7.13* 80.20% 1,691 chapter 7 91.29% 91.73% 91.16% f = 5.v 93.60% 846 chapter 13 88.44% 71.95% 87.34% f= 1.23 88.04% 152 total 87.7% 85.6% 81.5% 2,695 note: *significant at the 95% level of confidence. states. the relationship between cardholders’ behavior before default and the value of their choices in default is considered in section vi. c. credit card use and charge-off experience as pozdena suggested, cardholders who use their cards more intensively, or credit more aggressively, may be doing so because of the value of their option to default. in our initial investigation of this behavioral effect of the default option, the total portfolio sample was segmented in terms of various descriptors of card use behavior in 1990, the year before accounts were charged off. specifically, accounts were segmented in terms of size of credit line, the number of months that a balance had been revolved, the portion of the total credit line that had been used on average during the year (based on the average daily balance during the year), and whether or not the account had used the cash advance feature. because interest is charged from the date of transaction for a cash advance, this is also an indication of the intensity with which the card was used for credit. univariate analysis indicates that these descriptors of card use are significantly associ ated with the probability of charge-off (see table 4). the probability of charge-off was table 4 card use characteristics and the charge-off rate, total portfolio sample n = 3096 credit line charge-off rate revolving balance in 1990 charge-of/rate isl,ooo 8.9% 5 9 months 0.1% $l,ooq-$3,000 2.0 > 9 months 3.9 > $3,000 0.5 x2 = 53.4+ x2 = 71.8* percent credit line used in 1990 575% > 75% cfwrge-off rate 0.2% 14.6 x2 = 339.4 used cash advance in 1990 no yes charge-offrate 0.9% 4.5 x* = 33.0* 132 financial services review 4(2) 199.5 table 5 logit analysis of the default choice dependent variable: probability of choosing bankruptcy over outright default. independent variable estimated coefficient p-value legal condition prodebtor -0.286* .c002 procreditor 0.062 .4234 credit limit 0.0002* .0001 account age 0.006* .ockjl aggressive 0.530* boo1 constant -1.295* .oool variable mean value ,406 ,363 $2,187 61 months 253 notes: n = 5,103 (with 38.5% filing bankruptcy). model &i-squared = 200.8; significant at the 95% level of confidence. dramatically higher for accounts with credit limits less than $1,000 versus accounts with larger credit limits. the probability of charge-off was significantly higher for accounts that revolved their balances for more than nine months of the year relative to those that were used less intensely as a source of credit. the probability of charge-off was also significantly higher among those accounts in the sample for which, on average, more than 75% of the available line was used during the year relative to the rest of the sample of accounts. finally, accounts using the cash advance feature of the credit card had a significantly higher probability of charge-off. in sum, cardholders who use their credit cards aggressively for credit are distinctively different from other cardholders in terms of the probability that they will go into charge-off. in the final section, we use multivariate analysis to assess the extent to which these measures of card use are related to the value of the choices the cardholders make in default. d. card use and the default option a logit regression is used to estimate the likelihood of the debtor’s various choices in default. in logit analysis, the dependent variable in the regression model reflects a binary choice on the part of the debtor. in the first analysis, the debtor’s choice is to default outright or to file bankruptcy (table 5). if the choice to file bankruptcy is made, then the debtor must decide between filing under chapter 7 or chapter 13 of the bankruptcy code. the latter is the binary choice examined in the second analysis (table 6). given information on those factors which are hypothesized to affect debtors’ choices, the logit regressions estimate the likelihood that a debtor will file bankruptcy rather than simply default and, further, what type of bankruptcy. it is assumed that the debtor’s decisions are characterized by a logistic distribution, and the maximum likelihood estimates of the regression coefficients yield an estimated probability derived from the cumulative logistic distribution function. the logistic procedure calculates the p-value as a measure of the significance of each estimated regression coefficient. the overall model fit is measured by the -2 log-likelihood statistic. this statistic has a chi-squared distribution under the null hypothesis that all explanatory variables in the model are insignificant. the explanatory variables used in the estimations are defined as the legal condition present in the debtor’s credit cards and the option to default 133 table 6 logit analysis of the bankruptcy choice for debtors who chose bankruptcy dependent variable = probability of choosing chapter 13 over chapter 7. independent variable estimated coefficient p-value legal condition prodebtor -0.024* .8883 procreditor 0.444 .0065 credit limit -0.00003 .3052 account age 0.0005 .8436 aggressive -0.281* .0432 constant -1.704* .0001 variable mean value ,365 .392 $2,370 64 months ,323 notes: n = 1,967 (with 16.0% filing under chapter 13). model &i-squared = 17.6; significant at the 95% level of confidence. state, account characteristics, and card-use behavior. to measure the legal condition, two dummy variables are included. prodebtor takes the value of 1 for states that have higher restrictions on wage garnishment than the federal rule and where asset exemptions were set at or above the federal level, and 0 otherwise. the variable procrediror takes the value of 1 for states with strict or low asset exemption allowances and the federal allowance for wage garnishment, and 0 otherwise. the mixed message states in this study will remain in the constant. another dummy variable was created to measure aggressive card-use behavior in the year prior to charge-off. the variable aggressive takes the value 1 for those accounts that were revolved more than nine months in that year, kept an average balance of more than 75% of the credit line, and used the cash advance feature at least once. otherwise, aggressive takes the value 0. the continuous variable, credit limit, is the dollar value of the account’s credit line. it is expected that this account characteristic is positively correlated with the creditworthiness of the cardholder. analysis of these accounts indicates that the size of the credit line is significantly positively correlated with the age of the principal cardholder.4 the continuous variable account age measures the length of time the account has been opened, another measure of creditworthiness. e. outright default versus bankruptcy the estimated logit model performs well in determining the probability that a debtor will choose bankruptcy over outright default. in particular, account characteristics are significantly related to the debtor’s default choice, holding legal conditions constant (table 5). cardholders with larger credit lines and those with more mature accounts were signifi cantly more likely to file bankruptcy relative to other cardholders. furthermore, cardholders who revolved a large percentage of their line for most of the year and also used cash advances were significantly more likely to file bankruptcy than default outright. this result reinforces the idea that aggressive card users chose the default option which yields the highest value. as was the case in the univariate analysis, debtors in default in pro debtor states were significantly less likely to choose bankruptcy than were debtors in the mixed message states. 134 financial services review 4(2) 1995 however, in contrast to the results of the univariate analysis, debtors in the pro creditor states were not significantly more likely to choose bankruptcy, holding account characteristics and aggressive card use behavior constant. for further insight on the impact of legal conditions on debtors’ choices, holding card usage characteristics constant, we duplicated the analysis in table 5, disaggregating the legal conditions into separate garnishment and asset exemption variables. the results (not shown) indicate that debtors’ choices between outright default and bankruptcy were not systemati cally impacted by the level of asset exemptions. however, debtors were significantly more likely to default outright rather than tile bankruptcy in states where garnishment was prohibited or severely restricted. these results are consistent with those of several other analyses of the significance of the rules governing garnishment in consumers’ default behaviors. f. chapter 7 versus chapter 13 the estimated logit model also performed well in determining the probability that a debtor would choose chapter 13 over chapter 7, once the bankruptcy option had been selected (table 6). the results suggest that aggressive cardholders-those who revolved large balances for a large portion of the year-were significantly less likely to choose chapter 13 than the rest of the sample, holding all else constant. in other words, the more aggressive cardholders made choices in default that maximized the value of their option to default. in a separate analysis, the estimated measure of the default option value was significantly higher for the aggressive cardholders relative to the rest of the sample of defaulted accounts. debtors in states that followed the federal guidelines for garnishment and had low exemption allowances (pro creditor) were more likely to choose chapter 13 than debtors in the mixed message states, holding the use variables constant. when the legal conditions were separated, holding the use variables constant, the garnishment variables were insignificant in the choice between chapter 7 and chapter 13. but, the asset exemption variable was a significant influence in that choice. debtors in states with low exemption allowances were significantly more likely to file under chapter 13. vi. suiwimry and implications questions of the causal relationship between credit card use and the incidence of bankruptcy have arisen as the two phenomenon have grown concurrently during the last 15 years. this study used account data from a national credit card portfolio to demonstrate that among cardholders, estimates of the value of the option to default are significantly higher for bankruptcy than outright default. furthermore, holding legal conditions constant, card holders who use their credit cards aggressively for credit are significantly different from other cardholders in terms of the choices they make in default. aggressive credit-card users realize a higher value on average when they exercise their option to default than do other cardholders because they are significantly more likely to file bankruptcy and to file under chapter 7. holding card-use characteristics constant, there is a significant relationship between garnishment restrictions and debtors’ choices to default outright or file bankruptcy. however, variation in the level of asset exemption allowances is not a significant influence in that decision. once the decision to file bankruptcy is made, holding card use characteristics credit cards and the option to default 135 constant, the level of asset exemption allowances has a significant influence on the debtor’s choice between chapter 7 and chapter 13. an intriguing result of this analysis is that cardholders who use cards aggressively tend to make choices in default that maximize the value of their option to default. pozdena argued that some households may “self-select,” intensely using the credit feature of their cards and paying large interest charges with the understanding that the value of their option to default reduces the expected cost of their credit use. this moral hazard aspect of credit cards may provide another clue to the puzzle related to the high average level of interest rates for credit cards and consumers’ lack of sensitivity to those rates. the results of this analysis suggest that the consumers who are most likely to be paying the high interest rates on credit cards may discount the importance of that rate and focus instead on maximizing the value of their option to default. notes 1. luckett (1988) provides a detailed discussion of the bankruptcy code. 2. the federal standard for weekly wages exempt from garnishment (established in 1970) is the greater of 75% of the disposable earnings of the debtor or 30 times the minimum hourly wage in effect. in the late 1980s those states with more restrictive limits on the garnishment of wages included: ak, al, ct, de, hi, ia, id, il, ks, la, ma, md, me, mn, mo, ne, nh, nj, nm, ny, or, ri, vt, wv, and wi. wage garnishment was prohibited for the repayment of consumer credit or if the earnings were necessary for family support in fl, nc, nd, pa, sc, sd, and tx. 3. by the late 1980s those states that chose to “opt out” of the federal exemption allowance and establish a lower level of exempt assets included al, ar, de, ga, ky, md, ne, nh, oh, ok, sc, tn, and va. 4. the data set included the age of the principal cardholder for only about 50% of the accounts, so this cardholder characteristic could not be included in the final analysis. references apilado, v.p., dauten, j.j., & smith, d.e. (1978). personal bankruptcies. journal of legal studies, 7(june), 371-392. ausebel, l.m. (199 1) the failure of competition in the credit card market. american economic review, 31( 1, march), 50-8 1. canner, g.b., 8z luckett, ca. (1992). developments in the pricing of credit card services. federal reserve bulletin, 78(9), 652-666. canner, g.b., & luckett, c.a. (1991). payment of household debts. federal reserve bulletin, 77(4), 2 18-229. credit research center. (1982). consumers’ righr ro bankruptcy: origins and e&cts. west lafayette, in: credit research center, purdue university. demowitz, i. (1992). determinants of the chapter 7 consumer bankruptcy decision at the household level. unpublished manuscript, economics department, northwestern university. general accounting office. (1994, april). u.s. credit card industry: competitive developments need to be closely monitored. report ggd-94-23. washington, dc: gao. pozdena, r.j. (1991). solving the credit card mystery. wall street journal (december 26), 17. luckett, ca. (1988). personal bankruptcies. federal reserve bulletin, 74(9), 591-603. 136 financial services review 4(2) 1995 peterson, r.l., & aoki, k. (1984). bankruptcy filings before and after implementation of the bankruptcy reform law. journal of economics and business, 36(february), 95-105. shaffer, s. (1994). evidence of monopoly power among credit card banks. working paper 94-16, federal reserve bank of philadelphia. shepard, l. (1984). personal failures and the bankruptcy reform act of 1978. journal of law and economics, 27(cktober) 419-438. shiers, a.f., & williamson, d.p. (1987). nonbusiness bankruptcies and the law: some empirical results. the journal of consumer affairs, zl(winter), 277-292. sullivan, a.c., & worden, d.d. (1990). rehabilitation or liquidation: consumers’ choices in bank ruptcy. the journal of consumeraffairs, zi(summer), 69-88. sullivan, a.c., & worden, d.d. (1991). the law and consumer demand for debt relief under the bankruptcy code. unpublished manuscript, krannert school of management, purdue univer sity. selecting a social security age to balance consumption and risk barry r. cobb, jeffrey s. smith* virginia military institute, department of economics and business, lexington, va 24450, usa abstract this article uses monte carlo simulation to determine the maximum consumption given retirement at age 62, initial wealth, risk tolerance, and social security take decision. coile et al. (2002) argue for a delay, because the payment increases seven percent for each year. focusing on maximizing the expected present value of benefits may be misguided. this article shows that, conditional on retirement at age 62, initial consumption is always maximized by taking social security no later than age 63; it also results in the highest simulated ending wealth at death, and the lowest amount of simulated time (if any) living on just social security. © 2020 academy of financial services. all rights reserved. keywords: social security; consumption; monte carlo simulation 1. introduction everyone is faced with several major questions in life, such as: which college to attend, who to choose as a spouse, when to have children, and of course, when to take social security? unlike many of life’s other choices, the decision about when to claim social security is irrevocable; therefore, it is one that fills people with much uncertainty. this article will attempt to guide future retirees in their choices of when to begin receiving social security benefits and how much to consume throughout their retirement. building upon the insights provided by alderson and betker (2017), this article uses a monte carlo simulation model to show that the decisions regarding when to claim social security and how much to consume during retirement are ultimately questions of risk tolerance. to be more specific, *corresponding author. tel.: +1-540-464-7542; fax: +1-540-464-7005. e-mail address: smithjs@vmi.edu (j. s. smith) 1057-0810/20/$ – see front matter © 2020 academy of financial services. all rights reserved. financial services review 28 (2020) 201–221 “what level of confidence does a person require as it pertains to the possibility of exhausting their wealth before they die?” the answer to this question affects the level of consumption that a person can pursue in their retirement years, how much wealth they possess when they die, and how long they may have to live at a minimal consumption level if their savings is exhausted. alderson and betker (2017), in keeping with modigliani’s (1966) theory of consumption smoothing, presume that individuals want to maintain constant real consumption across their lifetime. they asked the question, “does a person that postpones taking social security get adequately compensated for doing so?” to answer this question, they calculated the probability of exhausting tax-deferred savings given a person’s retirement age and the age they decide to take social security. this article addresses a fundamentally different, but similar question: “what is the maximum amount of real consumption a person can choose, given their risk tolerance and the age at which they take social security?” monte carlo simulation helps to reveal the maximum amount of consumption that is consistent with a person’s risk tolerance. the questions posed by alderson and betker (2017) and this article are different than previous research. heretofore, most research that focused on when to claim social security did so with the intent to maximize the expected present value of the stream of benefits. coile, diamond, gruber, and jousten (2002) use simulations to show the optimal delay for both individuals and one-earner married couples using both the expected present value under financial calculations and a utility maximization model. they find that it is optimal to postpone, with delay times that ranged from months to years, in a large number of circumstances. docking, fortin, and michelson (2012) solve for the optimal retirement age given gender and race for single individuals. they find the decision is invariant to race and gender; individuals should either retire at age 64 if they retire early, or they should retire at age 67. similarly, shoven and slavov (2014a) calculate expected present value of benefits for a number of different situations and find that individuals who expected to live according to the average mortality tables should delay taking social security past their full retirement age unless the real interest rate is four percent (or higher). shoven and slavov (2014b) show that the benefit from delaying social security has risen over time, with someone born in 1951 (close to the baseline for this article) gaining 12.7% if they delay claiming until age 69. there is lots of evidence to suggest delaying receipt of social security; the popular press would summarize by saying, “wait as long as you can.” munnell and chen (2015) report that, in 2013, after adjusting for growing changes in the size of social security cohorts, 36% of men who claim social security benefits are aged 62. shoven, slavov, and wise (2018) conduct a survey to assess people’s attitudes pertaining to their decision to claim social security. seventy seven percent are happy with their decision to claim at age 62, while 90% are happy with their decision to claim at the full retirement age. the main reason cited for claiming before the full retirement age is a need for the money, and 23% of respondents state they use savings to finance the gap between retirement expenditures and social security income (61% stated they relied upon an employer-sponsored pension). this article aims to guide the decision regarding when to collect social security, as well as how much to spend during retirement, given an individual’s risk tolerance for failure. here it is important to stress the definition of failure. failure is defined as 202 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 fully exhausting the initial stock of wealth such that the individual must live only on their social security benefits. according to a report by the u.s. government accountability office (gao) (2019), 48% of households aged 55 and over have no retirement savings, but may have a defined benefit plan. twenty nine percent of households have neither. the gao gathered this information from the 2016 survey of consumer finances. thus, in this model, failure is defined in accordance with the state of wealth accumulation that is consistent with almost a third of american households, and possibly as many as 50% of households. 2. the model the basic decision is a function of gender, wealth, annual level of consumption, the age at which an individual begins collecting social security, the portfolio’s asset allocation, and the annual return an asset earns. gender is not a factor in the model; results are simulated for males and females separately. thus, gender affects the probability of death. a retiree’s wealth, wt, is modeled in year t as wt ¼ wt�1 þ pt � ctð þ � 1þ aet þ 1� að þbtð þ (1) where pt is the social security payment in year t; ct is the consumption in year t; a is the percentage invested in stocks, with stock (equity) and bond returns denoted by et and bt, respectively. fig. 1 depicts the structure of the simulation model utilized to evaluate the retirement portfolio that includes social security payments. the elements of the model and the components of (eq. 1) are described in the remainder of this section. fig. 1. structure of the retirement portfolio model. b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 203 2.1. stochastic assumptions the retiree divides wealth among investments in stocks and bonds. total returns from the s&p 500 are used to model stock returns, and total returns on 10-year treasury bonds represent the bond returns. both series range from 1928 to 2018 and were retrieved from professor aswath damodaran’s website (damodaran, 2019). the article uses this site due to the long history and the singular source. these data span the great depression, several wars including world war ii, and several extremely volatile periods for stocks. additionally, accurate total return series for bonds are hard to access, especially across such a long history. professor damodaran calculates the price return for a 10-year treasury bond and adds that to the coupon received for the 10-year treasury bond. therefore, these data give us average returns for both series and good correlation between the two. random variables are established by creating probability distributions for annual stock and bond returns, and these returns are correlated within each year. stock returns in each year are modeled as normally distributed with a mean of 11.5% and a standard deviation of 19.6%. bond returns are normally distributed with a mean of 5.2 percent and a standard deviation of 7.7 percent. returns are uncorrelated across years, but the stock and bond returns within each year are modeled with a correlation coefficient of –0.0276. alderson and betker (2017) use a bootstrapping approach and randomly sample an actual observation of investment returns to calculate the simulated returns for a given year (e.g., for year one their simulation might randomly select the historical return from 2010, in which case their model will use the total return for stock and bonds from 2010). in contrast, the portfolio model in this study defines the distributional parameters and then uses monte carlo simulation to sample both stock and bond total returns. inflation is estimated using the difference between the yield to maturities for treasury inflation-protected securities (tips) and the yield to maturity for similarly dated treasury securities matched for maturity dates from 2019 through 2029, with the smallest one-year difference being 1.705 percentage points and the largest being 1.959 percentage points. these inflation estimates from 2019 through 2029 are used to develop a normally distributed variable with a mean (1.88 percent) and standard deviation (0.277 percent) that follows from the projected inflation rate across these 11 years. correlation between inflation and bond and stock returns is estimated by calculating the correlation between the consumer price index (cpi) and bond and stock returns across the entire time period of the sample. this inflation rate is used to maintain constant real consumption for an individual, while also adjusting the level of social security benefits that will be paid in the future. on each simulation trial, a random number of years (l) until death is selected. this random variable is created using current social security life span assumptions. this is accomplished by running 10,000 simulation trials from the actuarial life table (social security administration, 2018) and then compiling results for conditional remaining life span given that an individual (male or female, as appropriate) obtains age 62. at the age of 62, a female is projected to live 2.82 years longer than a male. the average male life expectancy in the model is 19.44 years, which would suggest that a male that retires at age 62 will live to be approximately 81.44 years old, while a female will live to be approximately 84.26 years old after an average life expectancy of 22.26 years. the oldest person in any simulation was 204 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 108. these numbers are in line with, but longer than, the cohort life expectancy published by the social security administration in table 5.a5 of the 2019 annual report of the board of trustees of the old age and survivors insurance and federal disability insurance trust funds (2019). using calendar year 1955 (one year before the latest birth year of 1956), males are expected to live an additional 13.1 years past 65 (78.1 years), while females are expected to live 16.7 years past 65 (81.7 years). 2.2. social security benefit payments the retirement portfolio model assumes that an individual retires at the earliest opportunity to claim social security, which is currently age 62. unlike in alderson and betker (2017), who use the online social security benefits estimator, the model calculates the social security benefit using the process in appendix d from the annual statistical supplement to the social security bulletin for 2018, released may 2019 (social security administration, 2019). the model assumes that an individual who begins working at age 18 and works until age 62 earns at least the maximum amount of income that is subject to social security tax in the highest 35 years of their working life. the initial benefit sa is then estimated to coincide with the age, a, at which they start to collect social security. the benefit is adjusted for stochastic inflation, it, in each year so that the social security payment pt in each year is determined as pt ¼ r � sa � yt�1 u¼1 1 þ iuð þ (2) where r is an indicator variable that is equal to 1 if t<a–61 and 0 otherwise. for simplicity, it is assumed that someone turning 62 will retire and begin collecting benefits at some point in the following month; however, someone choosing their benefits at the age of 70 will have stopped working eight years earlier. this monthly retirement benefit is based on income earned through the end of 2018. individuals experience a reduction in the full benefit for taking their social security benefit before they reach their maximum retirement age. for example, the person that turns 70 (born in 1948) would experience a 25% reduction to their monthly benefit, while a person that turns 62 (born in 1956) would experience a 26.67% reduction. thus, the monthly benefit ranges from sa ¼ $2,150 for a 62-year old to sa ¼ $3,565 for a 70-year old. an individual would need to earn $128,400 in 2018 to earn the maximum income that is subject to social security taxation; as of tax filing year 2017, the latest statistics released by the irs (2019), only 17.1% of households earned more than $100,000 in adjusted gross income. purcell (2018) shows that the mean and median wage for males between 25 and 59 years old, in 2014 dollars, are $64,200 and $45,000, respectively. much like the irs data, a male earning above $100,000 would put them somewhere in the upper teens in terms of percentile earners. to increase applicability, the model is also used for an individual that only earned half of the maximum income ($64,200 in 2018) that is subject to social security taxation. these payments range from sa ¼ $1,552 for a 62-year old or sa ¼ $2,538 for a 70-year old. b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 205 similar to alderson and betker (2017), the model considers accumulated initial wealth (w0) in $100,000 increments up to $1,000,000; thus, the results apply to retirees at a wide range of savings levels. because individuals are assumed to retire and stop saving at age 62, waiting longer to claim social security does not directly, positively affect your accumulated wealth in the model, though it would add to the possible social security benefit. 2.3. decision variables the retiree makes two choices in the model: (1) the age, a, to begin social security payments, and (2) the percentage of the portfolio allocated to stocks, a. consumption, c1, in the first year of retirement is maximized subject to these choices. dimmock, wang, and yang (2019) derive, using values attributed to the u.s. stock market, how modern portfolio theory can justify a 60/40 split between stocks and bonds, typically used to represent endowments. this model, similar to alderson and betker (2017), sets a ¼ 60% for stocks and 1–a ¼ 40% for bonds, acknowledging that this allocation might expose individuals to an unacceptable level of beta risk as they age. someone in their seventies may be comfortable with a 25% chance of outliving their wealth, but uncomfortable with the possibility of losing significant wealth in any given year. since 1928, there have been 24 years with negative returns, and the average negative return is (13.7%). an individual with a 60% allocation is risking almost 10% of their portfolio in any given year based on historical data. to account for potential risk in the 60/40 allocation, this article also considers the “rule of thumb” portfolio (rot), where an individual allocates 120 minus their current age as the percentage invested in stocks. one other portfolio allocation considered is the “bridge fund” portfolio. an investor who decides to delay taking their social security benefit until after age 62 may want to limit the exposure of some portion of their wealth to the risk from bonds only. the portion of wealth for the bridge fund is calculated as the present value of a growing annuity, where the annuity is the annual payment to fund the foregone social security and the growth is equal to the expected inflation rate. this present value is invested only in bonds and is “used” to fund consumption until the individual begins drawing social security. at that time, any funds remaining in the bridge fund are then invested at the allocation rate for the portfolio, either 60/40 or the rot. it should be noted that it is also possible that the bridge fund will be exhausted before the individual begins to draw social security, as bond returns are stochastic. consumption adapts throughout the investor’s life to stay constant in real terms because the initial consumption, c1, adjusts with the stochastic inflation rate (it) each year. thus, consumption in year t is defined as ct ¼ c1 � yt�1 u¼1 i þ iuð þ (3) for year 2 and beyond. based on these choices, the model simulates a distribution of results for the retiree’s ending wealth at death, and the average years the retiree lives after exhausting all savings. 206 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 2.4. solution process the objective of this study is to advise a potential retiree on a consumption amount that is consistent with their risk tolerance. the investor chooses as a requirement of the portfolio that “failure” only occurs a set percentage of the time – the maximum failure rate. failure in this context means that savings are exhausted and the individual lives on social security payments alone. in contrast, alderson and betker (2017) consider a number of consumption strategies: (1) a retiree who chooses to spend 1/n of their accumulated wealth, where n is the median number of years remaining in their life, plus the projected social security benefit for the age at which they take the benefit, all adjusted for inflation; (2) a retiree who follows a similar strategy but sets n as the number of remaining years until they are 100 years old; (3) a retiree that purchases longevity insurance; and (4) a retiree that sets aside the equivalent amount spent on longevity insurance, excluding it from accumulated wealth when calculating consumption (i.e., somewhat analogous to the current model’s bridge fund). it is important to emphasize the difference between the approach in this study and the one that alderson and betker (2017) pursue. the previous study calculates the probability of failure given the age at which you select social security and the method of consumption an individual pursues. alderson and betker (2017) realize this on page 43 when they note that “the spending strategy with the lowest failure probability was not necessarily the “optimal” strategy . . . . any given retiree might well prefer a higher failure rate when balanced against higher consumption today.” given this admission, this model considers failure rates of five percent and 25% for both males and females, and determines the highest level of annual consumption that meets those failure rates, conditional on the age at which an individual takes social security and their initial level of wealth. this allows the individual to select, within some parameters, their comfort level, their consumption level, and the age at which they take social security. 3. results the case of selecting sustainable consumption at an acceptable level of risk is considered. for a 62-year old with tax-deferred savings of between $100,000 and $1,000,000, the retirement portfolio is simulated for beginning social security ages of 62 through 70 in one-year increments for male individuals and female individuals. we do not pursue joint decisions for married couples due to the number of simplifying assumptions required. the goal is to maximize current beginning consumption, while still meeting the failure threshold. charnes (2012) finds that much of the reduction in the standard error of estimates in monte carlo simulation models occurs when utilizing at least 2,000 simulation trials. in keeping with this convention, the model uses 2,000 simulation trials for each combination to identify the combinations that achieve the maximum consumption without exceeding the maximum allowable rate of failure for the portfolio, and the model simulates results for approximately 130 possible consumption levels for each social security age and initial wealth before there was no improvement in consumption. failure is defined as the percentage of time the retirement b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 207 portfolio is exhausted before death, thus resulting in the investor living strictly on social security income for the remainder of his or her lifetime. the solution process is applied to scenarios for each combination of assumptions as depicted in fig. 2, which also lists the tables at the end of the article where numerical results are reported. maximum consumption for males and females when considering a five percent failure rate is determined for the 60/40 and rot portfolios assuming an individual earned 100% of the maximum income for social security. each of these scenarios is also considered in combination with an initial bridge fund. results of these simulations are reported for all initial wealth levels in tables 1 through 4. results corresponding to those in tables 1 through 4 for a 25% required failure rate are only reported for an individual (male or female) that has $500,000 in initial wealth. full results are omitted because they do not differ meaningfully in consumption, ending wealth, or the unconditional number of years a person may fall short with the results in tables 1 through 4; however, these results are available upon request. similarly, when considering an individual that earned 50% of the maximum income for social security, we show the 60/40 allocation, both with and without the bridge fund, for males and females with $500,000 in initial wealth in the summary tables, but only for a five percent required failure rate. 3.1. results for maximum social security income without bridge fund table 1 considers a male who has earned the maximum amount of income subject to social security taxation throughout his working life. given a five percent chance of failure, table 1 displays the age at which an individual takes social security and the initial wealth they possess when they retire. keep in mind the assumption that everyone retires at age 62. the table reports the maximum initial consumption values, the median ending wealth at fig. 2. scenarios considered in the simulation results. 208 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 t ab le 1 m al e, m ax im u m so ci al se cu ri ty in co m e, fi v e p er ce n t fa il u re , 6 0 /4 0 sp li t, n o b ri d g e fu n d in it ia l w ea lt h s o ci al s ec u ri ty ag e 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 7 0 1 0 0 k c o n su m p ti o n 3 0 ,0 0 0 3 1 ,0 0 0 3 1 ,0 0 0 3 0 ,0 0 0 2 3 ,0 0 0 1 8 ,0 0 0 1 5 ,0 0 0 1 3 ,0 0 0 1 1 ,0 0 0 e n d w ea lt h 2 2 7 ,3 4 6 1 6 8 ,6 6 8 1 2 8 ,0 0 1 1 2 0 ,7 7 3 5 7 8 ,4 7 5 8 6 6 ,4 1 2 1 ,0 0 8 ,8 6 4 1 ,0 9 4 ,9 2 9 1 ,1 8 0 ,9 1 8 y ea rs sh o rt 0 .0 6 0 .0 5 0 .0 2 0 .0 3 0 .0 5 0 .0 3 0 .0 3 0 .0 4 0 .0 2 2 0 0 k c o n su m p ti o n 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 5 ,0 0 0 3 1 ,0 0 0 2 6 ,0 0 0 2 3 ,0 0 0 e n d w ea lt h 3 4 0 ,1 9 6 3 4 4 ,0 6 5 3 0 2 ,6 8 8 2 2 7 ,7 1 7 1 9 6 ,1 7 3 2 1 6 ,4 0 8 4 3 7 ,4 5 4 7 4 6 ,7 1 0 9 1 8 ,1 3 5 y ea rs sh o rt 0 .2 7 0 .1 3 0 .1 0 0 .2 0 0 .2 3 0 .0 2 0 .0 5 0 .0 4 0 .0 4 3 0 0 k c o n su m p ti o n 4 1 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 1 ,0 0 0 4 1 ,0 0 0 4 1 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 3 5 ,0 0 0 e n d w ea lt h 5 1 6 ,6 2 8 4 5 8 ,5 9 7 4 1 6 ,5 0 6 4 0 1 ,7 5 2 3 7 2 ,7 7 2 3 2 2 ,2 1 5 3 2 6 ,9 2 1 3 0 1 ,6 0 5 6 0 2 ,8 3 4 y ea rs sh o rt 0 .2 6 0 .2 8 0 .2 7 0 .2 0 0 .1 9 0 .2 9 0 .3 0 0 .7 3 0 .0 4 4 0 0 k c o n su m p ti o n 4 6 ,0 0 0 4 7 ,0 0 0 4 7 ,0 0 0 4 6 ,0 0 0 4 6 ,0 0 0 4 6 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 e n d w ea lt h 6 9 0 ,7 7 9 6 3 3 ,9 0 6 5 9 2 ,5 4 7 5 7 7 ,2 1 0 5 4 4 ,9 8 3 4 9 9 ,2 9 1 4 9 5 ,9 8 1 4 6 1 ,9 0 5 4 3 8 ,4 2 5 y ea rs sh o rt 0 .2 5 0 .2 7 0 .2 6 0 .2 1 0 .1 9 0 .2 3 0 .2 2 0 .3 0 0 .4 9 5 0 0 k c o n su m p ti o n 5 1 ,0 0 0 5 2 ,0 0 0 5 2 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 8 6 6 ,1 0 0 8 1 1 ,8 5 8 7 6 7 ,9 1 0 7 5 1 ,0 8 7 7 2 2 ,9 4 4 6 7 2 ,6 6 4 6 1 1 ,8 1 6 5 7 6 ,2 7 6 6 0 0 ,6 9 7 y ea rs sh o rt 0 .2 5 0 .2 6 0 .2 6 0 .2 1 0 .2 0 0 .2 2 0 .3 0 0 .3 6 0 .3 3 6 0 0 k c o n su m p ti o n 5 6 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 e n d w ea lt h 1 ,0 4 2 ,0 8 2 9 8 5 ,7 3 8 9 4 2 ,3 8 5 9 2 4 ,8 7 3 8 4 3 ,9 5 5 8 4 5 ,7 7 7 7 8 5 ,7 1 3 7 5 2 ,7 4 7 7 0 8 ,8 4 9 y ea rs sh o rt 0 .2 5 0 .2 6 0 .2 5 0 .2 2 0 .2 8 0 .2 2 0 .2 8 0 .3 0 0 .3 8 7 0 0 k c o n su m p ti o n 6 1 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 1 ,0 0 0 6 1 ,0 0 0 6 1 ,0 0 0 e n d w ea lt h 1 ,2 1 6 ,5 3 1 1 ,1 5 9 ,4 9 0 1 ,1 1 6 ,6 6 3 1 ,0 5 0 ,5 5 6 1 ,0 1 5 ,4 0 8 9 6 8 ,8 6 4 9 6 0 ,3 1 5 9 2 5 ,1 0 9 8 8 5 ,7 3 9 y ea rs sh o rt 0 .2 4 0 .2 5 0 .2 5 0 .2 8 0 .2 7 0 .2 9 0 .2 6 0 .2 8 0 .3 2 8 0 0 k c o n su m p ti o n 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 6 ,0 0 0 6 6 ,0 0 0 6 6 ,0 0 0 e n d w ea lt h 1 ,3 4 0 ,5 3 0 1 ,3 3 5 ,4 2 4 1 ,2 9 1 ,8 2 8 1 ,2 2 4 ,8 1 9 1 ,1 9 2 ,4 4 3 1 ,1 4 4 ,4 7 5 1 ,1 3 6 ,5 8 6 1 ,0 9 9 ,6 3 0 1 ,0 6 0 ,3 2 7 y ea rs sh o rt 0 .2 9 0 .2 5 0 .2 5 0 .2 8 0 .2 6 0 .2 8 0 .2 5 0 .2 6 0 .2 9 9 0 0 k c o n su m p ti o n 7 2 ,0 0 0 7 2 ,0 0 0 7 3 ,0 0 0 7 2 ,0 0 0 7 2 ,0 0 0 7 2 ,0 0 0 7 1 ,0 0 0 7 1 ,0 0 0 7 1 ,0 0 0 e n d w ea lt h 1 ,5 1 2 ,5 0 9 5 ,1 5 0 ,4 9 3 1 ,4 1 4 ,8 4 3 1 ,3 9 6 ,5 3 6 1 ,3 7 0 ,8 5 7 1 ,3 1 8 ,5 3 7 1 ,3 1 0 ,6 1 1 1 ,2 8 0 ,4 9 7 1 ,2 3 7 ,4 4 7 y ea rs sh o rt 0 .2 9 0 .2 5 0 .2 9 0 .2 7 0 .2 6 0 .2 7 0 .2 4 0 .2 6 0 .2 7 1 ,0 0 0 k c o n su m p ti o n 7 7 ,0 0 0 7 8 ,0 0 0 7 8 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 6 ,0 0 0 e n d w ea lt h 1 ,6 9 2 ,1 8 2 1 ,6 3 1 ,5 2 2 1 ,5 9 1 ,1 0 2 1 ,5 6 8 ,3 5 0 1 ,5 4 9 ,1 0 3 1 ,4 8 9 ,1 8 9 1 ,4 4 1 ,1 3 0 1 ,4 0 2 ,7 2 2 1 ,4 1 3 ,9 5 2 y ea rs sh o rt 0 .2 8 0 .2 9 0 .2 9 0 .2 7 0 .2 6 0 .2 6 0 .3 0 0 .3 1 0 .2 6 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 209 t ab le 2 f em al e, m ax im u m so ci al se cu ri ty in co m e, fi v e p er ce n t fa il u re , 6 0 /4 0 sp li t, n o b ri d g e fu n d in it ia l w ea lt h s o ci al s ec u ri ty ag e 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 7 0 1 0 0 k c o n su m p ti o n 3 0 ,0 0 0 3 1 ,0 0 0 3 1 ,0 0 0 3 0 ,0 0 0 2 3 ,0 0 0 1 8 ,0 0 0 1 5 ,0 0 0 1 3 ,0 0 0 1 1 ,0 0 0 e n d w ea lt h 2 5 8 ,0 2 8 1 9 3 ,8 1 4 1 5 3 ,7 0 3 1 5 8 ,5 5 7 7 9 4 ,2 2 8 1 ,2 0 4 ,1 5 4 1 ,4 0 5 ,1 4 5 1 ,5 5 9 ,6 4 7 1 ,7 0 6 ,1 4 7 y ea rs sh o rt 0 .1 2 0 .1 0 0 .0 3 0 .0 3 0 .0 5 0 .0 3 0 .0 3 0 .0 4 0 .0 2 2 0 0 k c o n su m p ti o n 3 5 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 5 ,0 0 0 3 6 ,0 0 0 3 5 ,0 0 0 3 1 ,0 0 0 2 6 ,0 0 0 2 3 ,0 0 0 e n d w ea lt h 4 5 6 ,4 0 3 3 9 0 ,2 0 7 3 4 4 ,7 0 1 3 3 7 ,3 1 9 2 2 9 ,8 2 3 2 7 4 ,4 1 1 5 9 8 ,7 1 5 1 ,0 3 4 ,7 4 1 1 ,2 9 4 ,8 2 8 y ea rs sh o rt 0 .2 6 0 .2 7 0 .2 1 0 .0 8 0 .3 8 0 .0 2 0 .0 5 0 .0 3 0 .0 3 3 0 0 k c o n su m p ti o n 3 9 ,0 0 0 4 0 ,0 0 0 4 1 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 3 9 ,0 0 0 3 5 ,0 0 0 e n d w ea lt h 7 2 7 ,8 2 3 6 6 1 ,1 6 7 5 4 3 ,3 0 6 5 3 3 ,9 1 3 5 1 0 ,3 3 7 4 5 5 ,4 3 8 3 9 1 ,3 7 5 4 6 6 ,7 8 3 8 1 2 ,7 6 6 y ea rs sh o rt 0 .1 8 0 .1 7 0 .3 0 0 .1 9 0 .1 4 0 .1 9 0 .4 2 0 .0 4 0 .0 5 4 0 0 k c o n su m p ti o n 4 4 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 4 ,0 0 0 e n d w ea lt h 9 2 9 ,7 9 2 8 6 3 ,2 3 2 8 2 0 ,6 7 4 7 3 0 ,3 4 3 7 0 3 ,1 0 0 6 4 3 ,7 6 8 5 7 9 ,2 1 2 5 5 7 ,3 0 0 6 1 2 ,4 2 6 y ea rs sh o rt 0 .2 4 0 .2 4 0 .2 1 0 .2 7 0 .2 2 0 .2 4 0 .3 5 0 .4 3 0 .3 9 5 0 0 k c o n su m p ti o n 4 9 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 1 ,1 2 6 ,6 1 1 1 ,0 5 9 ,7 2 3 1 ,0 1 8 ,6 0 9 9 3 1 ,7 0 1 9 0 1 ,2 2 6 8 4 7 ,4 3 4 7 7 5 ,4 4 8 7 4 1 ,8 8 2 7 0 7 ,9 2 7 y ea rs sh o rt 0 .2 8 0 .2 8 0 .2 6 0 .3 1 0 .2 7 0 .2 8 0 .3 5 0 .3 8 0 .4 8 6 0 0 k c o n su m p ti o n 5 4 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 e n d w ea lt h 1 ,3 2 4 ,2 9 1 1 ,2 5 8 ,9 1 9 1 ,2 1 1 ,5 3 6 1 ,2 0 3 ,6 3 8 1 ,0 9 9 ,2 5 8 1 ,0 4 8 ,1 2 3 1 ,0 5 3 ,9 8 0 9 3 1 ,0 5 3 9 8 4 ,2 3 1 y ea rs sh o rt 0 .3 1 0 .3 1 0 .3 0 0 .2 5 0 .3 0 0 .3 1 0 .2 5 0 .3 7 0 .2 9 7 0 0 k c o n su m p ti o n 5 8 ,0 0 0 5 9 ,0 0 0 6 0 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 e n d w ea lt h 1 ,5 9 7 ,4 1 1 1 ,5 3 1 ,6 6 0 1 ,4 0 7 ,7 5 6 1 ,4 0 0 ,9 0 9 1 ,3 7 2 ,6 8 0 1 ,3 2 0 ,6 8 5 1 ,2 4 7 ,5 1 9 1 ,2 1 1 ,2 4 1 1 ,1 7 3 ,6 0 6 y ea rs sh o rt 0 .2 6 0 .2 6 0 .3 2 0 .2 8 0 .2 5 0 .2 5 0 .2 8 0 .2 8 0 .3 0 8 0 0 k c o n su m p ti o n 6 3 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 e n d w ea lt h 1 ,7 9 6 ,2 4 0 1 ,7 2 8 ,6 1 6 1 ,6 9 2 ,5 3 6 1 ,5 9 7 ,5 3 7 1 ,5 7 2 ,8 9 5 1 ,5 1 9 ,8 5 3 1 ,4 4 5 ,5 3 7 1 ,4 0 3 ,8 3 3 1 ,3 5 9 ,9 8 0 y ea rs sh o rt 0 .2 8 0 .2 9 0 .2 7 0 .3 0 0 .2 8 0 .2 8 0 .3 0 0 .3 0 0 .3 1 9 0 0 k c o n su m p ti o n 6 8 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 e n d w ea lt h 1 ,9 9 2 ,8 4 4 1 ,9 2 7 ,8 9 4 1 ,8 8 1 ,1 1 8 1 ,7 9 2 ,6 2 1 1 ,7 6 5 ,7 2 5 1 ,7 1 1 ,8 3 7 1 ,6 5 2 ,6 1 6 1 ,6 0 4 ,5 4 5 1 ,5 5 6 ,3 1 2 y ea rs sh o rt 0 .3 0 0 .3 0 0 .2 9 0 .3 2 0 .3 0 0 .3 0 0 .3 2 0 .3 1 0 .3 2 1 ,0 0 0 k c o n su m p ti o n 7 3 ,0 0 0 7 4 ,0 0 0 7 4 ,0 0 0 7 3 ,0 0 0 7 4 ,0 0 0 7 4 ,0 0 0 7 3 ,0 0 0 7 4 ,0 0 0 7 4 ,0 0 0 e n d w ea lt h 2 ,1 9 6 ,6 8 8 2 ,1 3 1 ,3 2 2 2 ,0 7 9 ,9 4 8 2 ,0 7 5 ,7 3 9 1 ,9 6 8 ,9 6 0 1 ,9 1 4 ,3 1 5 1 ,9 1 9 ,2 3 7 1 ,8 0 1 ,4 7 3 1 ,7 6 0 ,4 1 5 y ea rs sh o rt 0 .3 2 0 .3 2 0 .3 1 0 .2 8 0 .3 2 0 .3 2 0 .2 7 0 .3 3 0 .3 3 210 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 t ab le 3 m al e, m ax im u m so ci al se cu ri ty in co m e, fi v e p er ce n t fa il u re , 6 0 /4 0 sp li t, b ri d g e fu n d in it ia l w ea lt h s o ci al s ec u ri ty ag e 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 7 0 1 0 0 k c o n su m p ti o n 3 0 ,0 0 0 3 1 ,0 0 0 3 1 ,0 0 0 3 1 ,0 0 0 3 0 ,0 0 0 3 2 ,0 0 0 3 2 ,0 0 0 3 4 ,0 0 0 3 6 ,0 0 0 e n d w ea lt h 2 2 7 ,3 4 6 1 6 8 ,6 6 8 1 2 1 ,0 2 0 3 3 ,0 7 6 1 4 8 ,0 8 4 1 3 5 ,2 0 8 1 9 5 ,6 4 8 2 0 4 ,5 3 5 2 1 3 ,8 7 5 y ea rs sh o rt 0 .0 6 0 .0 5 0 .0 1 0 .3 7 0 .0 0 0 .0 5 0 .0 1 0 .0 1 0 .0 1 2 0 0 k c o n su m p ti o n 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 5 ,0 0 0 3 4 ,0 0 0 3 6 ,0 0 0 e n d w ea lt h 3 4 0 ,1 9 6 3 4 4 ,0 6 5 2 9 3 ,6 6 9 2 0 4 ,7 3 9 1 5 4 ,5 7 8 7 8 ,5 6 4 9 3 ,2 8 7 2 0 4 ,5 3 5 2 1 3 ,8 7 5 y ea rs sh o rt 0 .2 7 0 .1 3 0 .1 0 0 .1 6 0 .0 9 0 .3 5 0 .0 1 0 .0 0 0 .0 0 3 0 0 k c o n su m p ti o n 4 1 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 1 ,0 0 0 4 1 ,0 0 0 4 1 ,0 0 0 4 1 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 e n d w ea lt h 5 1 6 ,6 2 8 4 5 8 ,5 9 7 4 0 8 ,4 8 7 3 8 0 ,3 4 5 3 3 0 ,9 8 3 2 4 9 ,9 8 7 1 5 6 ,3 3 6 1 4 2 ,3 1 2 1 1 7 ,9 2 8 y ea rs sh o rt 0 .2 6 0 .2 8 0 .2 7 0 .1 9 0 .1 5 0 .1 7 0 .3 5 0 .3 1 0 .0 4 4 0 0 k c o n su m p ti o n 4 6 ,0 0 0 4 7 ,0 0 0 4 7 ,0 0 0 4 6 ,0 0 0 4 6 ,0 0 0 4 6 ,0 0 0 4 6 ,0 0 0 4 6 ,0 0 0 4 5 ,0 0 0 e n d w ea lt h 6 9 0 ,7 7 9 6 3 3 ,9 0 6 5 8 3 ,2 2 2 5 5 5 ,1 0 5 5 0 6 ,8 4 4 4 2 7 ,9 2 4 3 3 1 ,6 3 4 2 4 9 ,0 3 4 2 1 7 ,2 6 9 y ea rs sh o rt 0 .2 5 0 .2 7 0 .2 6 0 .2 0 0 .1 7 0 .1 8 0 .2 6 0 .3 3 0 .2 1 5 0 0 k c o n su m p ti o n 5 1 ,0 0 0 5 2 ,0 0 0 5 2 ,0 0 0 5 1 ,0 0 0 5 2 ,0 0 0 5 2 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 e n d w ea lt h 8 6 6 ,1 0 0 8 1 1 ,8 5 8 7 6 1 ,9 5 9 7 3 4 ,0 8 0 6 3 0 ,5 1 5 5 6 0 ,1 3 8 5 2 8 ,8 3 7 4 5 1 ,5 3 1 3 7 0 ,8 2 9 y ea rs sh o rt 0 .2 5 0 .2 6 0 .2 5 0 .2 1 0 .2 8 0 .3 0 0 .2 2 0 .2 1 0 .2 5 6 0 0 k c o n su m p ti o n 5 6 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 e n d w ea lt h 1 ,0 4 2 ,0 8 2 9 8 5 ,7 3 8 9 3 5 ,6 9 0 8 4 6 ,6 8 6 8 0 1 ,4 8 8 7 7 8 ,3 5 6 6 8 4 ,1 0 0 5 9 9 ,7 4 2 5 0 1 ,2 7 4 y ea rs sh o rt 0 .2 5 0 .2 6 0 .2 5 0 .3 0 0 .2 8 0 .2 0 0 .2 4 0 .2 4 0 .2 7 7 0 0 k c o n su m p ti o n 6 1 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 2 ,0 0 0 6 1 ,0 0 0 6 1 ,0 0 0 6 1 ,0 0 0 e n d w ea lt h 1 ,2 1 6 ,5 3 1 1 ,1 5 9 ,4 9 0 1 ,1 1 0 ,3 4 9 1 ,0 2 7 ,4 6 1 9 7 5 ,1 2 1 9 0 1 ,8 2 6 8 5 8 ,6 3 0 7 7 4 ,2 8 5 6 7 6 ,3 7 8 y ea rs sh o rt 0 .2 4 0 .2 5 0 .2 5 0 .2 8 0 .2 7 0 .2 8 0 .2 4 0 .2 4 0 .2 5 8 0 0 k c o n su m p ti o n 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 7 ,0 0 0 6 6 ,0 0 0 6 6 ,0 0 0 6 6 ,0 0 0 e n d w ea lt h 1 ,3 4 0 ,5 3 0 1 ,3 3 5 ,4 2 4 1 ,2 8 5 ,4 7 1 1 ,2 0 1 ,6 0 1 1 ,1 5 2 ,4 5 8 1 ,0 7 7 ,0 2 4 1 ,0 3 3 ,2 7 8 9 5 4 ,1 8 6 8 5 2 ,8 3 7 y ea rs sh o rt 0 .2 9 0 .2 5 0 .2 5 0 .2 7 0 .2 6 0 .2 7 0 .2 4 0 .2 3 0 .2 4 9 0 0 k c o n su m p ti o n 7 2 ,0 0 0 7 2 ,0 0 0 7 3 ,0 0 0 7 2 ,0 0 0 7 2 ,0 0 0 7 2 ,0 0 0 7 1 ,0 0 0 7 1 ,0 0 0 7 1 ,0 0 0 e n d w ea lt h 1 ,5 1 2 ,5 0 9 5 ,1 5 0 ,4 9 3 1 ,4 0 6 ,1 6 6 1 ,3 7 8 ,3 8 6 1 ,3 2 9 ,2 5 2 1 ,2 5 2 ,1 3 4 1 ,2 0 8 ,6 3 3 1 ,1 2 6 ,5 7 6 1 ,0 2 6 ,8 3 5 y ea rs sh o rt 0 .2 9 0 .2 5 0 .2 9 0 .2 7 0 .2 6 0 .2 7 0 .2 4 0 .2 3 0 .2 4 1 ,0 0 0 k c o n su m p ti o n 7 7 ,0 0 0 7 8 ,0 0 0 7 8 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 7 ,0 0 0 7 6 ,0 0 0 7 6 ,0 0 0 e n d w ea lt h 1 ,6 9 2 ,1 8 2 1 ,6 3 1 ,5 2 2 1 ,5 8 3 ,0 9 6 1 ,5 5 3 ,4 9 7 1 ,5 0 3 ,7 5 9 1 ,4 2 5 ,5 8 7 1 ,3 3 6 ,0 4 2 1 ,3 0 1 ,8 8 4 1 ,2 0 2 ,7 7 1 y ea rs sh o rt 0 .2 8 0 .2 9 0 .2 9 0 .2 6 0 .2 6 0 .2 6 0 .2 9 0 .2 3 0 .2 4 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 211 t ab le 4 f em al e, m ax im u m so ci al se cu ri ty in co m e, fi v e p er ce n t fa il u re , 6 0 /4 0 sp li t, b ri d g e fu n d in it ia l w ea lt h s o ci al s ec u ri ty ag e 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 7 0 1 0 0 k c o n su m p ti o n 3 0 ,0 0 0 3 1 ,0 0 0 3 1 ,0 0 0 3 1 ,0 0 0 3 0 ,0 0 0 3 2 ,0 0 0 3 2 ,0 0 0 3 4 ,0 0 0 3 6 ,0 0 0 e n d w ea lt h 2 5 8 ,0 2 8 1 9 3 ,8 1 4 1 4 4 ,2 1 2 3 6 ,0 6 7 2 0 7 ,3 6 7 1 9 1 ,2 4 0 2 8 3 ,2 1 1 2 9 9 ,4 3 9 3 1 5 ,6 5 9 y ea rs sh o rt 0 .1 2 0 .1 0 0 .0 2 0 .5 0 0 .0 0 0 .0 5 0 .0 0 0 .0 0 0 .0 1 2 0 0 k c o n su m p ti o n 3 5 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 5 ,0 0 0 3 6 ,0 0 0 3 6 ,0 0 0 3 5 ,0 0 0 3 4 ,0 0 0 3 6 ,0 0 0 e n d w ea lt h 4 5 6 ,4 0 3 3 9 0 ,2 0 7 3 3 7 ,1 7 8 3 1 0 ,6 5 7 1 7 8 ,4 6 3 9 3 ,9 3 7 1 2 9 ,0 7 7 2 9 9 ,4 3 9 3 1 5 ,6 5 9 y ea rs sh o rt 0 .2 6 0 .2 7 0 .2 0 0 .0 6 0 .1 8 0 .5 2 0 .0 1 0 .0 0 0 .0 1 3 0 0 k c o n su m p ti o n 3 9 ,0 0 0 4 0 ,0 0 0 4 1 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 4 0 ,0 0 0 e n d w ea lt h 7 2 7 ,8 2 3 6 6 1 ,1 6 7 5 3 3 ,5 6 7 5 0 6 ,8 7 1 4 5 5 ,0 6 5 3 7 0 ,0 9 8 2 5 7 ,1 0 0 1 7 8 ,0 8 7 1 6 8 ,4 9 1 y ea rs sh o rt 0 .1 8 0 .1 7 0 .3 0 0 .1 8 0 .1 1 0 .0 9 0 .1 5 0 .3 7 0 .0 4 4 0 0 k c o n su m p ti o n 4 4 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 4 5 ,0 0 0 e n d w ea lt h 9 2 9 ,7 9 2 8 6 3 ,2 3 2 8 1 0 ,6 7 4 7 0 6 ,1 5 1 6 4 9 ,7 6 1 5 6 2 ,3 6 9 4 5 1 ,1 3 2 3 6 3 ,6 4 3 2 5 4 ,6 3 4 y ea rs sh o rt 0 .2 4 0 .2 4 0 .2 1 0 .2 6 0 .1 9 0 .1 8 0 .2 2 0 .1 8 0 .2 8 5 0 0 k c o n su m p ti o n 4 9 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 1 ,1 2 6 ,6 1 1 1 ,0 5 9 ,7 2 4 1 ,0 0 5 ,8 1 8 9 0 3 ,4 3 3 8 5 1 ,5 5 6 7 5 7 ,1 5 1 6 5 0 ,7 7 0 5 5 8 ,5 1 8 4 4 6 ,6 3 7 y ea rs sh o rt 0 .2 7 8 9 6 0 .2 8 2 6 0 .2 6 2 8 1 0 .3 0 7 4 9 0 .2 5 9 3 7 5 0 .2 5 4 5 2 0 .2 9 1 1 8 0 .2 5 3 6 0 .2 4 7 5 6 0 0 k c o n su m p ti o n 5 4 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 e n d w ea lt h 1 ,3 2 4 ,2 9 1 1 ,2 5 8 ,9 1 9 1 ,2 0 2 ,5 5 2 1 ,1 7 8 ,7 4 9 1 ,0 5 1 ,7 6 8 9 6 0 ,0 0 1 9 2 1 ,5 1 8 7 5 8 ,0 2 3 6 4 3 ,3 9 2 y ea rs sh o rt 0 .3 1 0 .3 1 0 .3 0 0 .2 4 0 .3 0 0 .3 0 0 .2 1 0 .3 1 0 .3 0 7 0 0 k c o n su m p ti o n 5 8 ,0 0 0 5 9 ,0 0 0 6 0 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 e n d w ea lt h 1 ,5 9 7 ,4 1 1 1 ,5 3 1 ,6 6 0 1 ,4 0 6 ,0 6 4 1 ,3 7 4 ,8 6 5 1 ,3 2 4 ,5 7 7 1 ,2 2 9 ,6 6 9 1 ,1 2 3 ,6 5 0 1 ,0 2 9 ,7 6 8 9 1 7 ,1 7 0 y ea rs sh o rt 0 .2 6 0 .2 6 0 .3 2 0 .2 8 0 .2 4 0 .2 4 0 .2 5 0 .2 2 0 .2 1 8 0 0 k c o n su m p ti o n 6 3 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 6 4 ,0 0 0 e n d w ea lt h 1 ,7 9 6 ,2 4 0 1 ,7 2 8 ,6 1 6 1 ,6 7 6 ,5 9 9 1 ,5 7 2 ,4 1 6 1 ,5 2 3 ,0 2 9 1 ,4 3 6 ,5 4 5 1 ,3 2 4 ,9 8 1 1 ,2 2 8 ,2 0 7 1 ,1 1 4 ,9 8 6 y ea rs sh o rt 0 .2 8 0 .2 9 0 .2 8 0 .3 0 0 .2 7 0 .2 7 0 .2 9 0 .2 7 0 .2 5 9 0 0 k c o n su m p ti o n 6 8 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 6 8 ,0 0 0 6 9 ,0 0 0 6 9 ,0 0 0 e n d w ea lt h 1 ,9 9 2 ,8 4 4 1 ,9 2 7 ,8 9 4 1 ,8 7 0 ,2 3 8 1 ,7 7 8 ,3 0 3 1 ,7 1 9 ,1 2 5 1 ,6 3 2 ,2 1 7 1 ,5 8 9 ,5 3 6 1 ,4 3 1 ,6 5 3 1 ,3 1 5 ,2 4 6 y ea rs sh o rt 0 .3 0 0 .3 0 0 .2 9 0 .3 2 0 .3 0 0 .3 0 0 .2 4 0 .3 0 0 .2 9 1 ,0 0 0 k c o n su m p ti o n 7 3 ,0 0 0 7 4 ,0 0 0 7 4 ,0 0 0 7 3 ,0 0 0 7 3 ,0 0 0 7 3 ,0 0 0 7 3 ,0 0 0 7 3 ,0 0 0 7 4 ,0 0 0 e n d w ea lt h 2 ,1 9 6 ,6 8 8 2 ,1 3 1 ,3 2 2 2 ,0 7 3 ,8 4 0 2 ,0 4 7 ,2 9 2 1 ,9 9 2 ,8 9 4 1 ,9 0 6 ,9 6 9 1 ,7 9 8 ,1 2 5 1 ,6 9 7 ,7 2 8 1 ,5 1 9 ,9 2 4 y ea rs sh o rt 0 .3 2 0 .3 2 0 .3 1 0 .2 8 0 .2 6 0 .2 6 0 .2 7 0 .2 5 0 .3 1 212 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 death associated with that decision, and the average length of time that someone who does fail would need to live on their social security benefits alone. consider the specific example of a 62-year old with $500,000 in tax-deferred savings. an initial consumption value of $52,000 is the maximum initial payout that allows the male investor to meet the five percent failure rate on average, and this occurs when starting social security payments as early as age 63. fifty-two thousand dollars is only the initial consumption at age 63; this amount is indexed by inflation in each subsequent year. while maximum initial consumption is the same when beginning social security payments at either age 63 or 64, the amount “left on the table” in median ending wealth differs between these two possible alternatives, with the male investor ending with about $44,000 in additional wealth when payments begin at age 63 as opposed to age 64. acknowledging that median ending wealth is a tertiary goal, it should be viewed as a tiebreaker when considering similar consumption levels and rates of failure. shoven, slavov, and wise (2018) identify reasons why an individual might claim immediately that generally center around the desire to retire coupled with a lack of liquidity. the results in table 1 suggest that it may be a consumption-maximizing decision. table 2 displays the results for a female investor. much like males, females maximize their consumption by claiming as early as age 63. starting social security payments at age 63 also maximizes median ending wealth relative to all other claiming ages with the same level of consumption. these results differ from traditional research regarding when to take social security benefits. many previous studies are concerned with ensuring the individual (or couple) maximizes the expected present value of their social security benefits. coile, diamond, gruber, and jousten (2002) answer this question and find that a male that retires at age 62 with $200,000 in initial wealth should wait three years to earn almost $11,000 (in 1992 dollars, which would be just over $20,000 in today’s dollars using the change in cpi over this time period) in expected present value. table 1 above shows that a male with $200,000 that waits three years to claim social security instead of claiming at age 62, assuming a five percent risk of failure, would be able to consume the same amount, $36,000, but they would die with about $112,000 less in median ending wealth. consider that dushi, iams, and trenkamp (2017) use data from the redesigned cps survey and find that half of all people aged 65 or older use social security for at least 50% of their family income and almost 25% of the population used social security for 90% of their income (the percentages are slightly lower for males only, which would be the most directly comparable to this scenario). they compare this to spending from assets (public and private pensions as well as iras), which only represent 21% of income. table 1 confirms that delaying social security does not increase consumption; it only lowers ending median wealth. thus, delaying social security to collect an additional $11,000 in expected present value is not a utility-maximizing decision. considering the risk-averse individual, the tables also identify the implications of the five percent failure rate, as they provide the mean amount of time a person would need to live on just their social security payment. this is calculated by taking the conditional number of years short times the probability of actually falling short. the individuals (male and female) possessing $500,000 in initial wealth that retire at age 62, begin consuming $52,000 (male) or $50,000 (female) a year, and begin drawing social security at age 63 will fail, or draw b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 213 down their initial wealth, five percent of the time. on average, the individual spends just over one-fourth of the last year of their life living solely on their social security payment, and is most likely to die with assets totaling either $811,858 (male) or $1,059,723 (female). specifically, for the male retiree, there is a 95% probability that wealth is not exhausted before death; whereas on the five percent of trials that failure occurs, this person spends an average of 5.2 years living strictly on social security (for an average years short of 5.2 times five percent, or 0.26 years). for this retiree, the first year spent living on social security payments typically occurs at age 84 years. 3.2. results for maximum social security income with bridge fund how does the possibility of contributing to the bridge fund change the results? table 3 reports the results for a male that earns 100% of the social security income requirement, has a risk tolerance level that supports a five percent failure rate, and maintains a 60/40 portfolio allocation. notice that consumption guidance is virtually unchanged, with maximum consumption still occurring at age 63 for initial wealth levels of $300,000 and above. ending wealth is slightly lower when using the bridge fund when social security payments are delayed until age 64 or beyond. this pattern is repeated for female investors as displayed in table 4. when establishing a bridge fund, females also maximize consumption at age 63 ($50,000) with no decrease in ending median wealth. the considerable wealth remaining at death, whether or not an individual establishes a bridge fund, is primarily due to precautionary consumption to ensure that their savings is only exhausted five percent of the time. the fact that consumption and ending median wealth do not change by establishing the bridge fund suggests that individuals could conceivably consider less risk and still achieve their goals. this pattern does not hold as the time between retirement and social security claiming increase. a male with $500,000 in initial wealth who retires at age 62 and claims social security at age 70 will have $230,000 less in median ending wealth, while a female under the same assumption will have $261,000 less. the bridge fund supports the same or slightly higher consumption, but the lower portfolio return while the bridge fund is active generates significantly less wealth. 3.3.results for additional scenarios to illustrate the effect of varying portfolio allocations and failure requirements, a male (table 5) or a female (table 6) that has $500,000 in initial wealth is considered. for these investors, maximum consumption is determined in the following scenarios: (1) bridge fund and no bridge fund versions of the rot portfolio allocation at the five percent level of failure; (2) bridge fund and no bridge fund versions of the 60/40 allocation at the 25% level of failure; (3) bridge fund and no bridge fund versions of the rot allocation at the 25% level of failure; and finally (4) bridge fund and no bridge fund versions of the 60/40 allocation, assuming the investor earns only 50% of the maximum social security income across their lifetime, with a five percent probability of failure. these results allow a comparison with the same individual, the retiree with $500,000 in initial wealth, across a wide variety of differing 214 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 t ab le 5 m al e, 5 0 0 k in it ia l w ea lt h , m u lt ip le as su m p ti o n co m p ar is o n s o ci al s ec u ri ty ag e 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 7 0 n o b ri d g e, fi v e p er ce n t, 6 0 /4 0 c o n su m p ti o n 5 1 ,0 0 0 5 2 ,0 0 0 5 2 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 8 6 6 ,1 0 0 8 1 1 ,8 5 8 7 6 7 ,9 1 0 7 5 1 ,0 8 7 7 2 2 ,9 4 4 6 7 2 ,6 6 4 6 1 1 ,8 1 6 5 7 6 ,2 7 6 6 0 0 ,6 9 7 y ea rs sh o rt 0 .2 5 0 .2 6 0 .2 6 0 .2 1 0 .2 0 0 .2 2 0 .3 0 0 .3 6 0 .3 3 b ri d g e, fi v e p er ce n t, 6 0 /4 0 c o n su m p ti o n 5 1 ,0 0 0 5 2 ,0 0 0 5 2 ,0 0 0 5 1 ,0 0 0 5 2 ,0 0 0 5 2 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 e n d w ea lt h 8 6 6 ,1 0 0 8 1 1 ,8 5 8 7 6 1 ,9 5 9 7 3 4 ,0 8 0 6 3 0 ,5 1 5 5 6 0 ,1 3 8 5 2 8 ,8 3 7 4 5 1 ,5 3 1 3 7 0 ,8 2 9 y ea rs sh o rt 0 .2 5 0 .2 6 0 .2 5 0 .2 1 0 .2 8 0 .3 0 0 .2 2 0 .2 1 0 .2 5 n o b ri d g e, fi v e p er ce n t, r o t c o n su m p ti o n 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 7 6 9 ,0 9 7 7 7 0 ,2 1 5 7 3 2 ,4 9 9 6 6 5 ,8 6 0 6 4 0 ,1 0 8 5 9 4 ,0 5 3 5 4 5 ,7 5 0 5 2 1 ,2 4 7 5 3 8 ,8 2 4 y ea rs sh o rt 0 .2 6 0 .1 9 0 .1 8 0 .2 2 0 .2 1 0 .2 3 0 .3 1 0 .3 5 0 .3 0 b ri d g e, fi v e p er ce n t, r o t c o n su m p ti o n 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 5 1 ,0 0 0 e n d w ea lt h 7 6 9 ,0 9 7 7 7 0 ,2 1 5 7 2 6 ,6 5 1 6 4 6 ,9 1 5 6 0 1 ,4 0 8 5 3 1 ,3 5 4 4 4 8 ,2 2 9 3 7 3 ,9 1 7 2 8 7 ,5 8 2 y ea rs sh o rt 0 .2 6 0 .1 9 0 .1 8 0 .2 2 0 .1 9 0 .2 0 0 .2 5 0 .2 6 0 .3 2 n o b ri d g e, 2 5 % 6 0 /4 0 c o n su m p ti o n 6 0 ,0 0 0 6 1 ,0 0 0 6 0 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 8 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 e n d w ea lt h 4 3 7 ,7 7 3 3 9 1 ,1 3 7 3 8 7 ,7 4 8 3 7 0 ,3 0 9 3 4 0 ,9 3 6 3 4 0 ,9 2 7 3 3 6 ,9 2 9 3 0 4 ,4 8 6 3 2 1 ,2 2 4 y ea rs sh o rt 1 .7 1 1 .8 9 1 .8 0 1 .7 5 1 .8 7 1 .8 1 1 .8 4 2 .1 5 2 .2 0 b ri d g e, 2 5 % , 6 0 /4 0 c o n su m p ti o n 6 0 ,0 0 0 6 1 ,0 0 0 6 0 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 8 ,0 0 0 5 6 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 e n d w ea lt h 4 3 7 ,7 7 3 3 9 1 ,1 3 7 3 8 0 ,9 6 3 3 4 8 ,4 9 7 3 0 1 ,3 4 7 2 7 2 ,3 1 4 2 6 9 ,7 1 5 2 3 6 ,4 8 7 1 8 8 ,8 8 3 y ea rs sh o rt 1 .7 1 1 .8 9 1 .7 5 1 .7 9 1 .9 4 1 .9 3 1 .6 1 1 .5 4 1 .5 8 n o b ri d g e, 2 5 % , r o t c o n su m p ti o n 6 0 ,0 0 0 6 0 ,0 0 0 6 0 ,0 0 0 5 9 ,0 0 0 5 8 ,0 0 0 5 8 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 e n d w ea lt h 3 7 8 ,0 1 2 3 7 1 ,7 7 8 3 3 2 ,6 8 4 3 1 9 ,6 4 4 3 3 2 ,5 0 5 2 9 0 ,1 8 4 2 8 8 ,2 6 9 3 0 9 ,5 7 8 2 7 7 ,5 1 8 y ea rs sh o rt 1 .8 4 1 .7 5 1 .8 7 1 .8 8 1 .6 6 1 .9 0 1 .9 1 1 .8 1 2 .2 5 b ri d g e, 2 5 % , r o t c o n su m p ti o n 6 0 ,0 0 0 6 0 ,0 0 0 5 9 ,0 0 0 5 8 ,0 0 0 5 8 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 e n d w ea lt h 3 7 8 ,0 1 2 3 7 1 ,7 7 8 3 6 8 ,1 9 4 3 3 8 ,4 1 6 2 9 2 ,9 3 0 2 6 9 ,2 5 9 2 3 4 ,5 0 0 2 0 5 ,6 2 2 1 6 5 ,5 5 3 y ea rs sh o rt 1 .8 4 1 .7 5 1 .5 9 1 .6 1 1 .7 3 1 .6 7 1 .7 3 1 .6 4 1 .6 8 n o b ri d g e, fi v e p er ce n t, 6 0 /4 0 , 5 0 % c o n su m p ti o n 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 3 ,0 0 0 e n d w ea lt h 8 5 8 ,9 2 6 8 5 8 ,4 6 0 8 2 4 ,6 3 6 7 6 9 ,5 3 2 7 4 8 ,3 2 3 7 0 6 ,9 5 6 6 6 7 ,1 2 0 6 4 2 ,9 8 6 6 5 7 ,6 6 5 y ea rs sh o rt 0 .2 6 0 .2 1 0 .2 1 0 .2 4 0 .2 3 0 .2 5 0 .3 0 0 .3 3 0 .2 6 b ri d g e, fi v e p er ce n t, 6 0 /4 0 , 5 0 % c o n su m p ti o n 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 4 ,0 0 0 4 3 ,0 0 0 e n d w ea lt h 8 5 8 ,9 2 6 8 5 8 ,4 6 0 8 2 0 ,3 8 8 7 5 4 ,1 1 0 7 1 9 ,0 0 7 6 6 1 ,3 2 0 5 9 2 ,7 0 8 5 2 8 ,9 3 8 5 1 7 ,6 2 4 y ea rs sh o rt 0 .2 6 0 .2 1 0 .2 1 0 .2 4 0 .2 3 0 .2 4 0 .2 8 0 .3 0 0 .1 9 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 215 t ab le 6 f em al e, 5 0 0 k in it ia l w ea lt h , m u lt ip le as su m p ti o n co m p ar is o n s o ci al s ec u ri ty ag e 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 7 0 n o b ri d g e, fi v e p er ce n t, 6 0 /4 0 c o n su m p ti o n 4 9 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 1 ,1 2 6 ,6 1 1 1 ,0 5 9 ,7 2 3 1 ,0 1 8 ,6 0 9 9 3 1 ,7 0 1 9 0 1 ,2 2 6 8 4 7 ,4 3 4 7 7 5 ,4 4 8 7 4 1 ,8 8 2 7 0 7 ,9 2 7 y ea rs sh o rt 0 .2 8 0 .2 8 0 .2 6 0 .3 1 0 .2 7 0 .2 8 0 .3 5 0 .3 8 0 .4 8 b ri d g e, fi v e p er ce n t, 6 0 /4 0 c o n su m p ti o n 4 9 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 1 ,1 2 6 ,6 1 1 1 ,0 5 9 ,7 2 4 1 ,0 0 5 ,8 1 8 9 0 3 ,4 3 3 8 5 1 ,5 5 6 7 5 7 ,1 5 1 6 5 0 ,7 7 0 5 5 8 ,5 1 8 4 4 6 ,6 3 7 y ea rs sh o rt 0 .2 8 0 .2 8 0 .2 6 0 .3 1 0 .2 6 0 .2 5 0 .2 9 0 .2 5 0 .2 5 n o b ri d g e, fi v e p er ce n t, r o t c o n su m p ti o n 4 9 ,0 0 0 5 0 ,0 0 0 5 1 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 4 9 ,0 0 0 4 9 ,0 0 0 e n d w ea lt h 9 5 6 ,2 9 9 8 8 6 ,0 8 2 8 0 4 ,8 6 7 7 9 4 ,7 4 2 7 7 3 ,0 7 1 7 0 2 ,7 2 0 6 6 3 ,8 5 5 7 4 3 ,1 8 8 6 9 6 ,4 7 6 y ea rs sh o rt 0 .2 5 0 .3 1 0 .3 1 0 .3 2 0 .3 4 0 .3 6 0 .3 6 0 .3 1 0 .3 7 b ri d g e, fi v e p er ce n t, r o t c o n su m p ti o n 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 4 9 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 5 0 ,0 0 0 e n d w ea lt h 8 8 3 ,0 2 1 8 7 6 ,1 3 3 8 5 9 ,4 2 2 8 2 7 ,0 6 1 7 1 7 ,7 5 3 6 4 3 ,9 7 5 5 6 3 ,0 4 2 4 7 4 ,5 6 6 3 7 4 ,9 5 8 y ea rs sh o rt 0 .3 2 0 .2 8 0 .2 7 0 .1 9 0 .2 5 0 .3 0 0 .2 8 0 .2 7 0 .2 8 n o b ri d g e, 2 5 % , 6 0 /4 0 c o n su m p ti o n 5 8 ,0 0 0 5 9 ,0 0 0 5 9 ,0 0 0 5 8 ,0 0 0 5 8 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 5 ,0 0 0 e n d w ea lt h 5 2 4 ,2 6 0 4 6 0 ,7 3 7 4 3 0 ,5 3 2 4 1 2 ,7 5 0 3 6 3 ,4 5 3 4 0 3 ,0 3 6 3 9 4 ,1 7 4 3 4 5 ,5 2 6 3 8 4 ,7 5 8 y ea rs sh o rt 2 .0 1 2 .1 8 2 .2 8 2 .1 3 2 .3 0 1 .9 1 2 .1 8 2 .5 2 2 .4 5 b ri d g e, 2 5 % , 6 0 /4 0 c o n su m p ti o n 5 8 ,0 0 0 5 9 ,0 0 0 5 8 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 5 3 ,0 0 0 e n d w ea lt h 5 2 4 ,2 6 0 4 6 0 ,7 3 7 4 6 9 ,1 3 1 4 3 5 ,0 9 7 3 7 6 ,5 6 8 3 5 3 ,7 3 7 3 1 2 ,1 4 8 2 8 2 ,6 4 9 2 3 6 ,2 1 8 y ea rs sh o rt 2 .0 1 2 .1 8 1 .9 8 1 .9 7 2 .0 8 1 .9 6 1 .9 6 1 .7 6 1 .6 5 n o b ri d g e, 2 5 % , r o t c o n su m p ti o n 5 7 ,0 0 0 5 8 ,0 0 0 5 8 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 5 5 ,0 0 0 e n d w ea lt h 4 8 0 ,7 4 5 4 2 3 ,7 4 3 3 8 6 ,1 0 2 3 7 3 ,4 6 4 3 4 8 ,5 3 0 3 7 0 ,1 0 1 3 7 3 ,1 9 7 3 4 4 ,4 5 8 3 2 0 ,3 7 4 y ea rs sh o rt 1 .8 4 2 .0 0 2 .1 0 2 .0 6 2 .1 3 1 .9 7 1 .9 0 2 .1 1 2 .5 1 b ri d g e, 2 5 % , r o t c o n su m p ti o n 5 7 ,0 0 0 5 8 ,0 0 0 5 7 ,0 0 0 5 7 ,0 0 0 5 6 ,0 0 0 5 6 ,0 0 0 5 5 ,0 0 0 5 4 ,0 0 0 5 3 ,0 0 0 e n d w ea lt h 4 8 0 ,7 4 4 4 2 3 ,7 3 4 4 3 1 ,6 0 7 3 5 1 ,9 0 1 3 6 0 ,3 0 6 2 9 1 ,4 0 0 2 6 0 ,4 9 1 2 3 6 ,5 0 9 2 0 1 ,1 3 6 y ea rs sh o rt 1 .8 4 2 .0 0 1 .7 7 2 .1 0 1 .8 1 2 .0 9 2 .0 8 1 .8 6 1 .7 3 n o b ri d g e, fi v e p er ce n t, 6 0 /4 0 5 0 % c o n su m p ti o n 4 2 ,0 0 0 4 2 ,0 0 0 4 3 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 e n d w ea lt h 1 ,1 1 1 ,0 8 4 1 ,1 1 8 ,5 1 1 1 ,0 0 6 ,3 2 6 1 ,0 2 3 ,6 7 3 9 9 8 ,7 7 7 9 5 9 ,9 3 4 9 0 3 ,4 0 9 8 7 2 ,8 2 8 8 4 5 ,2 6 6 y ea rs sh o rt 0 .3 0 0 .2 3 0 .3 2 0 .2 5 0 .2 2 0 .2 3 0 .2 1 0 .2 6 0 .2 9 b ri d g e, fi v e p er ce n t, 6 0 /4 0 , 5 0 % c o n su m p ti o n 4 2 ,0 0 0 4 2 ,0 0 0 4 3 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 4 2 ,0 0 0 e n d w ea lt h 1 ,1 1 1 ,0 8 4 1 ,1 1 8 ,5 1 1 1 ,0 0 5 ,3 2 6 1 ,0 0 3 ,9 2 5 9 6 7 ,1 2 2 9 0 0 ,3 5 1 8 1 6 ,6 9 5 7 4 3 ,9 1 1 6 6 3 ,1 3 5 y ea rs sh o rt 0 .3 0 0 .2 3 0 .3 2 0 .2 5 0 .2 1 0 .2 1 0 .2 3 0 .2 1 0 .1 9 216 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 scenarios. some scenarios introduce more risk of failure, such as the 25% maximum failure rate, while others exhibit more risk-averse behavior (rot allocation). in all cases for both males and females, consumption is maximized no later than age 63, and sometimes as early as age 62, with no loss of ending median wealth and no significant increase in the number of years short. two examples illustrate the effect of changing these assumptions. a female that accepts a 25% probability of failure who does not establish a bridge portfolio and maintains a 60/40 allocation can maximize initial consumption at $59,000 by claiming social security at age 63. she will still have a median ending wealth of $460,737, and she would accept the possibility of living an average of 2.18 years on just her social security payments (there is no shortage with 75% probability, with an average of 8.72 years short on the trials where failure occurs). as noted earlier (u.s. government accountability office, 2019), at least 29% of households live entirely on social security throughout retirement, and there are likely a much larger percentage without personal savings that live entirely on social security plus defined benefit pension plans. the same male could consume $61,000 and would possess $391,000 in ending median wealth, with less time on social security if he fails (1.9 years; there is no shortage with 75% probability, with an average of 7.6 years short on the trials where failure occurs). the other example includes a male and female with $500,000 in initial wealth that earn just 50% of the maximum social security income. assuming a five percent probability of failure and a 60/40 allocation, a male would maximize consumption of $44,000 by claiming at their retirement age of 62. similarly, a female would maximize consumption of $43,000 also by claiming at age 64. 4. conclusions when examining the results of the study, some general trends emerge. once past the lower levels of wealth accumulation (starting at $300,000), delaying receipt of social security payments beyond age 63, for both male and female, only decreases median ending wealth at death with effectively the same level of consumption in nearly all scenarios, and it has virtually no effect on the average years spent living on social security if you fail. the conclusion is that an individual that retires at age 62 should draw social security almost as soon as they retire. because of this, when an individual takes social security has far less importance in terms of how much they can consume as compared with how much wealth they possess when they retire. at a five percent chance of failure, each additional $100,000 means $5,000 in additional initial annual consumption for both males and females. the reason for this must be emphasized. if an individual retires and delays taking social security, then they obviously make sizable withdrawals from their initial wealth to fund their current consumption. unlike alderson and betker (2017), who assume a constant level of consumption equal to the individual’s initial wealth divided by their life expectancy, which is then added to their social security benefit, this article solves for the maximum amount of consumption that only exhausts the initial wealth level consistent with the chosen probability of failure. using alderson and betker’s formula, an individual that retires at age 62 with $500,000 in initial wealth would have initial consumption equal to approximately b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 217 $50,948. this model has a consumption level that roughly matches alderson and betker’s consumption level assuming five percent failure; however, consumption is equal to $61,000 at a 25% level of failure. at lower levels of income, the previous conclusion about when an individual should take their social security payment is magnified. an individual that retires at age 62 that has not accumulated much wealth must take their social security within the first few years of eligibility or else they must be willing to constrain their consumption to very small levels. for example, a male that has $100,000 in wealth that maximized their social security contributions that waits to take social security until age 67 can only safely consume $18,000 a year if they accept a five percent chance of failure and do not establish a bridge fund. this is consistent with the evidence provided by shoven, slavov, and wise (2018) that indicates a primary reason for starting early social security payments is the need to consume more than the income earned from a retirement nest egg. this individual would be able to consume considerably more by establishing the bridge fund ($32,000), albeit with over $730,000 less in median ending wealth. while the individual fails five percent of the time, the average number of years that the male would fall short is very low, because consumption is severely restricted. adopting a more conservative allocation strategy such that the percentage invested in stocks is 120 minus an individual’s current age did not meaningfully affect the results. initial consumption was either the same or less under the “rule of thumb” portfolio across both five percent and 25% failure probabilities, and there was no discernable pattern in ending wealth or unconditional years short. for any level of initial wealth beyond $300,000, the age social security payments begin does not meaningfully change the maximum level of consumption at the five percent level of failure; at the 25% level of failure, it only changes consumption by $5,000. beyond an initial wealth level of $300,000, a $100,000 increase in your initial wealth level changes initial consumption by as much, if not more, than taking social security at the earliest opportunity. this suggests that advisors must continue to emphasize that it is as much about how much an individual chooses not to spend before retirement as it is about how much an individual makes in any given year. the included tables help determine the amount of consumption a person can afford given their wealth and the tolerance for risk. while it is unlikely that an individual would completely maximize their social security income, using either table 5 (males) or table 6 (females) can help guide an individual. once an individual has determined which scenario applies by matching their earnings profile to either of the two scenarios, the maximum income or 50% of maximum income, they can determine levels of consumption associated with their choice of establishing a bridge fund, their asset allocation, and their accepted probability of failure. the respective table demonstrates the resulting levels of wealth and amount of time associated with failure. thus, an individual can see the difference. for example, assuming maximum income earnings is the baseline, an individual that has $500,000 in retirement assets with no bridge fund and a rot allocation can take social security at 63 and consume, in real terms, $51,000 a year, assuming a five percent level of failure; this individual will likely die with almost $770,000 dollars in wealth. if this does not match the individual’s needs (or wants) for current consumption compared with ending wealth, they 218 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 can choose to accept a higher level of risk (25%) and take social security at age 62 and begin consuming $60,000 a year. the results suggest they will most likely still have $438,000 in wealth at death. comparing the tradeoffs between consumption, bridge fund, ending wealth, and possible years of failing will help form opinions about the relative risk aversion for each individual. simulations also show that the standard 60/40 endowment allocation provides results that are as good or better than a model that adjusts allocation as a client ages, when it comes to initial level of real consumption and ending wealth. there is a slight increase in the amount of time the average person that fails will live only on social security, but it is measured in days not month or years (at the five percentage level). this suggests, though it warrants further investigation, that it is the flow of funds into and out of the investment account, as opposed to making fine tuning adjustments to the asset allocation of the portfolio, that matter for retirement consumption. likewise, a male investor that believes an average death at 81 is too young could always make their choices from the respective female table. doing so would add approximately 2.75 years, or an additional 10% of years beyond the statistical death table for males. limitations there are some limitations that must be identified. first and foremost, this article assumes that an individual retires at age 62. as reported by munnell and chen (2015), the current trend is to work longer. also, the model uses social security actuarial tables, randomly sampling the age of death; however, many financial planners allow for gradients of life past the standard statistical tables. for instance, current financial planning software allows for a 10%, 20%, and 30% longer life span than projected by the social security tables, given a certain age. for this model, the average life expectancy is 19.47 years; extending this by 30% would add an additional 5.84 years of social security benefits, as well as consumption. we also only consider single individuals, although a reasonable way to follow the advice provided in this article would be to pool the couple’s initial wealth and use guidance from the female results to select an appropriate consumption level. for each age and wealth combination, approximately 130 simulations of various levels of initial consumption were attempted, each with 2,000 trials, before the highest consumption value where the failure requirement was met was assumed to be the optimal solution. there is no guarantee the solution is optimal, but all results presented are consistent with experiments where more simulation trials were used for comparison. additionally, the selected returns for either stocks or bonds in actuality may not resemble the returns used in the simulations, or that the long-run correlation used in this simulation will hold going forward. the model in this article assumes constant, real consumption. both social security payments and consumption are indexed by inflation and the market’s estimate of inflation may not be realized. some retirees may wish to front-load their consumption while they are healthy enough to enjoy it. additionally, some retirees will still have some expenses such as remaining house payments, and so forth, in the first few years of retirement. at a certain point, as mobility begins to decline and adverse health conditions start to increase, it is b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 219 reasonable to expect spending to become less on discretionary items such as travel and more on nondiscretionary, such as health care. thus, a non-linear consumption pattern may be more realistic and is worthy of further research. a retiree who anticipates this type of spending pattern could potentially increase the maximum failure rate (say from five percent to 25%) because social security payments would remain for the last few years of the person’s life in the event savings is exhausted. finally, this research does not account for the effect of taxes. geisler and hulse (2018) look at the implications associated with portfolio drawdown strategies when including social security benefits pertaining to efforts to extend the life of the portfolio. they show that drawdown strategies do matter and can impact the life of the portfolio. however, as their research shows, assumptions must be made regarding the split between tax-advantaged accounts and those that are fully taxable, marital status, and deduction status (itemized vs. standard). acknowledgments we thank three anonymous reviewers for comments and suggestions which greatly improved the article. references alderson, m. j., & betker, b. l. (2017). does the benefit of deferring social security offset the opportunity cost to do so? journal of financial planning, 30, 38-47. charnes, j. m. (2012). financial modeling with oracle crystal ball and excel. hoboken, nj: john wiley and sons. coile, c., diamond, p., gruber, j., & jousten, a. (2002). delays in claiming social security benefits. journal of public economics, 84, 357-385. damodaran, a. (2019). damodaran online. http://www.stern.nyu.edu/;adamodar/pc/datasets/histretsp.xls (accessed january 17, 2019). dimmock, s. g., wang, n., & yang, j. (2019). the endowment model and modern portfolio theory. nber working paper 25559. docking d. s., fortin, r., & michelson, s. (2012). the influence of gender and race on the social security early retirement decision for single individuals. journal of economics and economic education research, 13, 87104. dushi, i., iams, h. m., & trenkamp, b. (2017). the importance of social security benefits to the income of the aged population. social security bulletin, 77, 1-12. geisler, g., & hulse, d. s. (2018). the effects of social security benefits and rmds on tax-efficient withdrawal strategies. journal of financial planning, 31, 36-47. irs. (2019). soi tax stats individual statistical tables by size of adjusted gross income. https://www.irs. gov/statistics/soi-tax-stats-individual-statistical-tables-by-size-of-adjusted-gross-income (accessed january 17, 2019). modigliani, f. (1966). the life cycle hypothesis of saving, the demand for wealth and the supply of capital. social research, 33, 160-217. munnell, a. h., & chen, a. (2015). trends in social security claiming. center for retirement research issue brief, may, 15-18, 1-5. purcell, p. j. (2018). trends in men’s wages, 1981–2014. social security bulletin, 78, 1-29. shoven, j. b., & slavov, s. n. (2014a). does it pay to delay social security? journal of pension economics and finance, 13, 121-144. shoven, j. b., & slavov, s. n. (2014b). recent changes in the gains from delaying social security. journal of financial planning, 27, 32-41. 220 b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 shoven, j. b., slavov, s. n., & wise, d. a. (2018). understanding the social security claiming decisions using survey evidence. journal of financial planning, 31, 35-47. social security administration. (2018). actuarial life table. baltimore, md: social security administration. retrieved from www.ssa.gov/oact/stats/table4c6.html (accessed march 1, 2019). social security administration. (2019). ssa publication no. 13-11700. (available at www.ssa.gov/policy/docs/ statcomps/supplement/2018.supplement18.pdf). the 2019 annual report of the board of trustees of the federal old-age and survivors insurance and federal disability insurance trust funds. (2019). (available at https://www.ssa.gov/oact/tr/2019/). u.s. government accountability office. (2019). gao-19-442r. retirement security: most households approaching retirement have low savings, an update. (available at https://www.gao.gov/assets/700/ 697898.pdf). b.r. cobb, j.s. smith / financial services review 28 (2020) 201–221 221 pii: s1057-0810(97)90012-3 financial services review, 6(3): 185-196 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. market knowledge in managed bond portfolios sandeep singh and william h. dresnack municipal in this paper, we analyze the investment performance of two types of open-end munici pal bond mutual finds: first, mutual funds that have the objective of generating income free from federal income taxes; second, funds that have the objective of generating income free from not only federal but also a particular state’s income taxes. our results suggest the following: first, municipal bond fund monthly returns either mirror or lag the lehman brothers municipal bond index returns on a risk-adjusted basis; second, when state taxes are significant, such as in california and new york, the investor ben efits from investing in state-specific municipal bond funds. i. introduction because of the tax-free nature of their associated coupon payments, municipal bonds are particularly attractive investment vehicles for high-income individuals. in recent years, municipal bond mutual funds have become investment vehicles of choice for municipal bond investors for several reasons. first, due to information inefficiencies, limited trading, and high spreads in the municipal bond market, transferring the monitoring and trading decisions to a full time manager is appealing. second, even for relatively small dollar investments, such mutual funds offer greater diversification than would be available through investment in individual bonds. and finally, the absence of indexed municipal bond funds points to the supposed superior portfolio management skills and consequent higher returns implicitly promised by the managers of these funds. considerable research has been directed towards the study of managed stock portfo lios (see shukla and trzcinka (1992)). despite the extensive investment in managed bond portfolios (estimated at $684 billion of the total $2.16 trillion invested in mutual funds as of the end of 1994), scant attention has been paid to the study of bond mutual funds. in this paper, we examine the performance of three categories of open-end municipal bond funds. sandeep singh and william h. dresnack l department of business administration and economics, state university of new york, college at brockport, brockport, ny 14420; e-mail: ssingh@acsprl.acs.brockport.edu. 186 financial services review 6(3) 1998 we analyze the benefits of buying into a municipal bond mutual fund vis-a’-vis a self directed or unmanaged portfolio of municipal bonds or an index portfolio. based on the results, we find, as with many equity fund studies, strong support in favor of indexing. as a group, all types of municipal bond mutual funds examined in this study are unable to out perform the index on a risk-adjusted basis. following this introduction, section ii provides a review of the literature. in section iii, the data and methodology of the study are discussed. in section iv, the empirical results are provided. finally, in section v, a summary and implications of the findings are presented and concluding remarks are made. ii. review of the literature although most municipal bond coupon payments are free from federal income taxes, to be free from a particular state’s income tax requires that such bonds meet pre-defined criteria. this requirement assumes special significance in states where such taxes are particularly punitive, such as new york, oregon, and california. over the past several years, a substan tial market for bonds that are free from both federal and state income taxes has developed. in an effort to appeal to this segment of the market, investment companies sponsor funds whose objective is to generate income free from both federal and state taxes, for example, the california municipal bond fund. the tremendous popularity of state municipal bond funds is evidenced by the fact that at the end of 1994 there were 703 such funds compared to only 258 at the end of 1989 (mutual fund fact book, 1995). in a study of bond mutual funds, gudikunst and mccarthy (1992) find that bond mutual fund performance closely mirrors that of stock mutual funds; i.e., after adjustment for fees and expenses, the performance of bond mutual funds closely matches that of the index on a risk-adjusted basis. they further observe that fund size, yield spreads, and asset growth of the funds affect excess returns. blake, elton, and gruber (1993) find that bond funds under-perform relevant indexes. fund expenses are also found to be inversely related to fund performance. after accounting for survivorship bias they find some evidence of persistence of performance. one of the two popular approaches to active bond portfolio management involves the forecasting of interest rates. the second requires a manager to identify bonds that s/he con siders mispriced and to build a portfolio of such bonds. eventually managers with superior selectivity should be able to outperform, on a risk-adjusted basis, a passive index or a group of peers. this ability to identify mispriced bonds can have two possible sources: the supe rior analytical skills of the fund manager, and the acquisition of superior information regarding the issue/issuer. many qualitative issues, especially those relating to the dissemination and analysis of financial information, are unique to the municipal bond market. for example, issues such as the timeliness with which information is reported by municipalities, factors that influ ence rating changes in the municipal bond market, and influences of geographic market segmentation on the cost of municipal borrowing have been the subject of previous research. dwyer and wilson (1989) found that the timeliness of municipal reporting is influ enced by incentives for management signaling and is related to the nature of the report market knowledge 187 message; that is, good news or bad news. the assessment of the value of a municipal bond is further complicated by the role of the political underpinnings of the issuer and socio-eco nomic influences. ingram, raman, and wilson (1989) suggest that bond prices may not be free from the unobserved political and economic factors which influence the accounting numbers of municipal issuers. such variables not only have to be assessed initially, but must also be continuously monitored by the manager, a task made difficult by the informa tion issues previously discussed. benson, marks, and raman (1991) find differential levels of compliance with gaap, and the degree of compliance, to have a significant impact on net interest cost, because municipal bond issuers are not subject to the securities & exchange commission (sec) requirement of compliance with generally accepted accounting principles (gaap). the sec, in an effort to improve the quality of information in the municipal bond market, adopted rule 15~2-12 in june 1989. the rule requires greater and more accurate disclosure on the part of dealers and underwriters of new municipal bond issues. through this rule, the sec also imposed the responsibility of “reasonable basis of belief’ upon dealers and underwriters in key representations made in the disclosure statements. despite the rule, in a 1993 article in the government finance review, arthur levitt, the chairman of the sec, expressed his concern over “abusive practices” in the municipal securities market. he remarked that more information regarding the issuers, transactions, and the market was essential to the reform of the municipal bond market. bond portfolio managers who attempt to identify undervalued bonds often focus their analyses on the default risk associated with a bond. the principal intent of such managers is to identify in advance bonds whose real default risk is lower than is widely agreed upon in the marketplace. some indication regarding the default risk associated with a municipal bond issue can be inferred from its rating. the pricing of municipal bonds is further com plicated by the impact of split ratings on yields and investor anticipation of rating changes, both of which have been the subject of much discussion. moreover, many municipal bond issues are not rated at all. reeve and hartwell (1988), in a study of non-rated municipal bonds, find non-rated bonds do not necessarily indicate poor or high credit risk. additional variables must be considered, such as the size of the municipality issuing these bonds, along with the decision of the municipality to rate or not to rate. in a study of municipal bond rating changes, lewis, patton, and green (1988) demon strate uniformity in terms of statistical models and humans in their ability to predict bond rating changes. marquette and wilson (1992), in a study of rating decisions of seasoned municipal bond issues, show that rating decisions that can be predicted from publicly avail able information can be anticipated two years in advance. interestingly, however, they find that changes that are not predictable from publicly available information cannot be antici pated by the market. eshan and wilson (1992) find a geographical (regional versus national) segmentation influence on the interest costs of primary issues of municipal bonds and financial disclosure by the municipalities. in this paper, we examine the performance of the following three categories of man aged municipal bond funds: (1) open-end mutual funds whose objective is to generate income free from federal income taxes, referred to as “national funds”; (2) similar funds whose objective is to generate income free from both federal and new york state income taxes, referred to as “new york funds”; and (3) similar funds whose objective is to gener ate income free from federal and california state income taxes, referred to as “california funds.” we examine the benefits of buying into a municipal bond mutual fund vis-a’-vis a 188 ftnancial services review 6(3) 1998 self-directed or an unmanaged portfolio of municipal bonds or an index portfolio. we also compare the performance of state-specific mutual funds to funds that are managed with the objective of generating income that is free from federal income taxes only. iii. data and methodology the total returns and traditional risk-adjusted returns generated by the previously men tioned groups of funds are analyzed. the monthly return on the lehman brothers munici pal bond index (lbmbi or “the index”) is used as a proxy for market return. comparisons of performance are also conducted between national and new york funds and national and california funds. the total returns of the index are used as a benchmark. based on comparative performance, we attempt to identify the return premium, if any, to state municipal bond fund investors. an additional analysis based on excess return per unit of total risk is also utilized. this methodology is free from the bias in results that might be caused by the choice of an inappropriate benchmark. a) risk adjustment the proper measure of a bond portfolio’s risk is a subject of debate. while a convinc ing argument can be made to regard a bond portfolio’s duration as a measure of its risk, duration-based risk measurement methodologies are not without their shortcomings. sim ple duration-based measures are computed on the assumption of parallel shifts in the yield curve and are a time-decaying measure of risk. moreover, the utility of duration as a mea sure of risk is further compromised if the portfolio of bonds has a significant exposure to non-option-free bonds. ilmanen (1992) shows that in the 1980s duration explained 80 to 90 percent of the cross-sectional variation in government bond returns. however, for cor porate bonds, ilmanen, mcguire, and warga (1994) find that duration is an incomplete measure of risk and its utility declines as the credit quality of the bond declines. the other popular measure of the risk of a portfolio is derived from regression-based index models. here, the risk of a portfolio is measured by beta. with regard to bond port folio management, this methodology also has weaknesses. aside from the traditional criti cisms of estimation errors and appropriateness of benchmarks, any sampling of time series data assumes that the sample period is representative of the norm. furthermore, index model-based performance evaluation measures assume stationary returns, stationary port folio return variance, and stationary portfolio composition. in this study, we opted for the use of the index model and the widely-used statistics of sharpe’s measure, treynor’s measure, and jensen’s alpha for use as risk-adjusted perfor mance measures for three primary reasons: 1. the performance of mutual funds, especially in the popular press, is almost always judged against an index (e.g. the eafe or s&p 500). 2. bonds often form only part of the total investment wealth of an individual. the per formance of other asset classes, stocks for instance, is judged relative to a benchmark index and it would be confusing to the individual investor to use a duration-based measure which is unrelated to the measure of risk of other asset classes. moreover, duration is a difficult concept for the individual investor to comprehend. market knowledge 189 3. many other papers on bond portfolio performance evaluation, gudikunst and mccarthy (1992) and blake, elton, and gruber (1993), for example, have used index based models in determining the risk of the portfolio. by using the index model we hope to add to this body of knowledge. one of the criticisms advanced in the evaluation of risk-adjusted returns of a managed portfolio relative to an index is that the index is usually uninvestible and does not reflect passive management costs. to address this issue we reduced the monthly return on the unmanaged index by 20 basis points per year (see sharmin mossavar-rahmani (1987)) to reflect a uniform management fee for the index. when comparing the performance of state specific mutual funds to the lbmbi, the return on the index was further reduced by the highest prevailing marginal income tax rate for the year in question in the state. the monthly return on the lbmbi was first reduced by 20 basis points when using it as a bench mark for national funds (lbmbi adj). it was then reduced by 20 basis points plus the highest marginal income tax rate for the state in that particular calendar year (lbmbi adjl & lbmbi adj2). the income tax rate used for analyzing california funds was 11%; for new york funds, it ranged from a high of 8% for 1988 to a low of 7.5% for 1995. in performance measurement studies, an inappropriate benchmark choice likely dis torts results and conclusions (see lehman and modest [1987] and grinblatt and titman [1989, 19931). to test the validity of our benchmark-based, risk-adjusted performance measures and results, we conducted non-parametric, benchmark-independent, tests of per formance comparison. we used chi-square tests of independence to evaluate the number of funds having a sharpe’s measure greater than the median of the mixed sample. b) sample selection the morningstar mutual funds ondisc database of july 1995 was the source for total monthly returns. we selected mutual funds which had the objective of investing in bonds whose coupon payments were free from federal income taxes and state of california and new york state income taxes, respectively, and that were in existence on january 1, 1988. this gave us a sample of 153 national, 36 new york, and 42 california funds. from the preceding sample, funds with available returns for all ninety months for the period from january, 1988 to june, 1995 were selected. this provided a sample of 152 national, 36 new york, and 42 california funds. only one mutual fund (mackenzie limited term municipal a) was lost to attrition. the sample is, therefore, relatively free from survivor ship bias for the period of the study. monthly returns on 90-day u.s. treasury bills were obtained from lbbotson associ ates, and monthly returns on the lbmbi were provided by the lehman brothers fixed income research department. the lbmbi consists of approximately 25,000 municipal bond issues with a market value totaling $400 billion. inclusion in the index requires an issue to have a credit rating of baa3 or better, an outstanding par value of at least $3 mil lion, and the issue must have been part of a deal of at least $50 million. the bonds must also have been issued within the past five years and must have at least one year remaining until maturity. the index includes general obligation, revenue, insured, and pre-funded bonds; the index excludes bonds subject to alternative minimum tax, taxable municipal bonds, and bonds with floating or zero coupons. statistics on the index are calculated at mid-month and at each month-end (a guide to the lehman brothers family of fixed income indices, 190 financial services review 6(3) 1998 1994). we know of no indexed municipal bond mutual fund currently available to individ ual investors. cross-sectional data for the chi-square tests of independence were obtained from the morningstar mutual funds ondisc database of july 1995. iv. empirical results a) descriptive statistics and total returns demographic, total return information, and other descriptive statistics on the lbmbi and the three sets of mutual funds appear in table 1. as the table shows, all groups of funds have mean monthly returns lower than that of the index, while both new york and califor nia funds have a higher total risk, as measured by the standard deviation of their monthly returns. a majority of funds in each group carried greater total risk than that of the index, while only a small percentage provided a comparatively greater return as compensation. b) risk-adjusted performance using the 90 monthly returns for each fund and the monthly return on the 90-day u.s. treasury bill, three risk-adjusted measures of performance were calculated. the total monthly returns on the portfolio were used to calculate the standard deviation used in table 1 fund demographics amd returns (total and risk adjusted) (all returns are based on monthly observations) lbmbi national new york california number of funds 152 36 42 mean fund size ($ millions) 756.08 507.39 723.20 average maturity (years) 16.06 18.63 19.16 mean expense ratio 0.84% 0.86% 0.82% average manager tenure (years) 6.93 7.08 6.57 mean return awi& 0.64%* 0.66%* 0.66%% std. deviation of returns (mean) 135% 1.34% 1.55% 1.59% funds with average return > mean lbmbi ret 33 (21.71%) 5 (13.89%) 6 (14.28%) funds with std. dev. > std. dev. of lbmbi 85 (55.92%) 31(86.11%) 37 (88.09%) mean sharpe’s ratio 0.18 0.13** 0.14** 0.13** std. deviation of sharpe’s ratio 0.06 0.03 0.02 mean treynor’s ratio tm% 0.19% 0.19% 0.19% std. deviation of treynor’s ratio 0.10 0.05 0.04 funds with sharpe’s ratio > lbmbi 20 (13.16%) 2 (5.55%) 2 (4.76%) funds with treynor’s ratio > lbmbi 26 (17.10%) 3 (8.33%) 3 (7.14%) mean alpha -0.04%*** -0.06%*** -0.07%*** std. deviation of alpha 0.05 0.05 0.05 mean beta 0.94 1.09 i.12 std. deviation of beta 27% 13% 18% average r2 89.46% 91.09% 91.38% note: * significantly different from mean lbmbi return (a = .05, two tailed test). **significantly different from mean lbmbi sharpe’s ratio (a = .05. two tailed test). *** h,: at 0 rejected at 5% level of significance. conclude (y < 0 market knowledge 191 deriving sharpe’s measure. the ninety monthly excess returns of all mutual funds were then regressed individually against the monthly excess return on the index. summary sta tistics are presented in table 1. all groups of funds had a sharpe’s ratio significantly different from the mean sharpe’s ratio of the lbmbi (a = .05, two-tailed test). a majority of the funds had a return per unit of total risk lower than that of the index. using treynor’s ratio as a measure of return per unit of systematic risk, it was again found that, as a group, the funds were unable to gener ate excess returns per unit of systematic risk superior to that of the index. based on the 90 month beta, alphas were calculated for each fund. for each set of funds the hypothesis that 01 z 0 was tested. the evidence was overwhelmingly against managed portfolios. for all funds the preceding hypothesis was rejected at a 5% significance level, indicating negative alphas for all groups of funds. stated alternatively, most funds had returns below the level that would be justified by the market risk of their portfolios. c) comparative performance in order to determine the total and risk-adjusted return benefits to state-specific mutual fund investors, a performance comparison between national and state-specific funds was then conducted. on a pre-tax basis, the total return and the related standard deviation of new york funds were significantly different from national funds. for california funds, the total return was not statistically different from national funds at a 5% level. results of these tests appear in table 2. national funds had a systematic risk different from the new york funds. in relation to risk-adjusted performance measures at a 5% level, for the three risk-adjusted measures examined, the difference was not statistically significant. the difference in the systematic risk between the national and california funds, as measured by beta, was statistically sig nificant. for risk-adjusted performance measures, the difference was statistically signifi cant for jensen’s alpha only. table 2 comparative statistics (all returns are based on monthly observations) total return standard sharpe ‘s treynor’s jensen ‘s deviation ratio ratio alpha beta mean (national new york) standard error (national new york) t-statistic (national new york) mean (national california) standard error (national california) t-statistic (national california) -0.022% -0.216% -0.003 -0.005% 0.028% -0.15 0.011% 0.043% 0.007 0.011% 0.010% 0.03 -2.00* -4.99* -0.46 -0.46 1.65 -4.94* -0.017% -0.245% 0.003 0.003% 0.028% a.18 0.010% 0.048% 0.006 0.017% o.ow% 0.035 -1 .i4 -5.05* 0.41 0.19 3.21* -5.126* note: * significant at the level of 0.05. two tailed test. 192 financial services review 6(3) 1998 although the systematic risk of the state-specific funds was higher than that of the national funds, the investors were compensated through higher returns. we found no dif ference between the funds in the risk-adjusted measures of sharpe’s measure and trey nor’s measure, and we found the jensen’s alpha of state-specific funds to be the same as, or higher than, that of national funds. it can therefore be concluded that if the holders of a national fund must pay a certain portion of their return in state taxes, thereby reducing the average return, it will be beneficial to invest in state-specific mutual funds rather than mutual funds subject to state income taxes. it is also be possible that state-specific funds may be subject to greater non-systematic risk such as budgetary problems of a particular state, but at this stage we do not find any statistical evidence to support this hypothesis. we would also like to point out that our sample of state-specific municipal bond mutual funds included funds relating to new york and california only. although these funds comprise a major portion of the state-specific municipal bond fund market, the results should be interpreted with this fact in mind. d) benchmark independent performance comparison the choice of an inappropriate benchmark might distort conclusions regarding risk adjusted performance measures. to verify the results obtained in the earlier sections, the hypothesis of independence of sharpe’s measure and fund type of the two samples was tested using the chi-square test of expected frequencies. this test is independent of bench mark influence. in the case of new york and national funds, the hypothesis of independence of fund objectives and sharpe’s measure could not be rejected. in the case of california and national funds, the hypothesis of independence of fund objectives and sharpe’s measure was rejected at a 5% level of significance. the implication is that a fund’s sharpe’s mea sure is dependent upon fund type, i.e., national or california funds. e) performance comparison (adjusted lbmbi) when analyzing performance of the state-specific mutual funds, the monthly returns of the index were adjusted downward to reflect the cost of indexing and the effect of the highest tax rate on the returns. total return and risk-adjusted performance measures after reducing the index returns for management fees and income taxes appear in table 3. the results reinforce the case for indexing. with the exception of california funds, other groups of funds were unable to generate statistically significant superior risk-adjusted returns. this is particularly compelling considering the reduction in index returns to reflect the costs of indexing and state income taxes. when compared to a management-fee-adjusted index, national funds as a group were unable to generate superior post-fee, risk-adjusted returns. the national funds had a lower systematic risk but had a significantly lower trey nor’s measure than the index and a mean jensen’s alpha less than zero. new york funds had a higher systematic risk but no risk-adjusted measure was statis tically different from those of the index. california funds as a group had a higher system atic risk than the index, and the mean treynor’s measure was higher than that of the index. california funds as a group indicated a statistically significant non-negative alpha at a five percent significance level. market knowledge 193 table 3 total and risk-adjusted returns (benchmark adjusted for management fee and state taxes) (all returns are based on monthly observations) national new york california (lbmbi alu) (lbmbi adji) (lbmbi adj2) number of funds 152 mean return 0.64%* std. deviation of returns (mean) 0.0134 funds with average return > mean lbmbi ret 57 (37.50%) funds with std. dev. > std. dev. of lbmbi mean sharpe’s ratio std. deviation of shave’s ratio mean treynor’s ratio std. deviation of treynor’s ratio funds with sharpe’s ratio > lbmbi funds with treynor’s ratio > lbmbi mean alpha std. deviation of alpha funds with alpha > 0 mean beta std. deviation of beta average r2 87 (57.24%) 0.13* 0.06 0.19%* 0.097 33 (21.71%) 48 (31.58%) -0.02%** 0.05 48 (31.58%) 0.95 27% 89.45% 36 0.66%* 0.0155 32 (88.89%) 35 (97.22%) 0.14 0.03 0.18% 0.045 17 (47.22%) 20 (55.55%) 0.0055% 0.05 20 (55.55%) 1.19 14.5% 9 1.08% 42 0.66%% 0.0159 37 (88.07%) 38 (90.47%) 0.13 0.02 0.17%* 0.038 24 (57.14%) 31 (73.81%) 0.016%*** 0.045 32 (71.11%) 1.26 26% 91.28% now: * significantly different from mean lbmbi measure (a = .05, two tailed test). ** h,: a 2 0 rejected at 5% level of significance. conclude a < 0. *** significantly equal to or greater than 0 (a = ,051) f) fund variables and performance previous studies of bond funds have shown that fund-specific factors, such as expense ratio and asset growth, influence fund performance. we tested the hypothesis of indepen dence of fund’s sharpe ratio and average maturity, fund size, expense ratio, manager ten ure, and turnover rate, for all three set of funds. the results of the tests for national funds are presented in table 4 below. for national funds we were unable to reject the hypothesis of independence of fund performance, as measured by sharpe’s ratio, and the fund-specific variables of expense ratio, manager tenure and portfolio turnover. in other words, the hypothesis that the sharpe’s measure was independent of these fund variables could not be rejected at a 5% level of significance. however, for national funds, the hypothesis of independence of the sharpe’s measure and average maturity and fund’s asset size was rejected at a 5% level. in other words, the number of funds in the top 50% of average maturity was disproportion ately greater than expected, with the inference that funds with greater average maturity tended to have higher than the median sharpe ratio. similar results were indicated for asset size. larger funds that invested in municipal bonds nationally tended to have a sharpe’s ratio higher than the median. identical tests were conducted with the same fund-specific variables for new york and california funds. for all the fund-specific variables, namely average maturity, asset size, expense ratio, manager tenure, and turnover, the hypothesis of the independence of those variables and sharpe’s measure could not be rejected for all the fund specific variables at 194 financial services review 6(3) 1998 table 4 contingency table for national funds (independence of fund variables and performance) funds with shame’s measure z= median -c = median total average maturity top 50% bottom 50% total chi square = 5.51789* asset size top 50% bottom 50% total chi square = 10.5263* expense ratio top 50% bottom 50% total chi square = 0.9473 manager tenure top 50% bottom 50% total chi square = 1.6842 turn over top 50% bottom 50% total chi square = 3.7895 45 31 76 31 45 76 76 76 152 48 28 76 28 48 76 76 76 152 35 41 76 41 35 76 76 76 152 42 34 76 34 42 76 76 76 152 44 32 76 32 44 76 76 76 152 note: (critical value at a = .05 and i degree of freedom is 3.841) a five percent significance. that is, the sharpe’s measure was independent of these specific fund variables for both groups of funds. v. summary and conclusions as concluded in earlier studies (gudikunst and mccarthy [1992]; and blake, elton and gruber [ 19931) of other types of bond funds, municipal bond fund returns either mirror or lag municipal bond index returns on a risk-adjusted basis. this study provides further sup port in favor of indexing. mutual fund companies may find the idea of offering indexed municipal bond funds worth pursuing. from an individual investor’s point of view, perfor mance across actively managed municipal bond funds is similar. for an investor with no forecasting capabilities, if an opportunity to invest in an indexed municipal bond fund becomes available, such a fund should be the investor’s first choice. s/he can expect per formance that will match or surpass the returns of actively managed municipal bond port folios on a risk-adjusted basis. market knowledge 195 when state income taxes are considered significant, for example, in states like new york and california, it is beneficial for the individual investor to buy into state-specific municipal bond funds. we find that the total risk of state specific mutual funds for new york and california is consistently higher than that of the national funds. the owners of state specific municipal bond funds are subject to higher systematic (market) risk, though it is accompanied by a higher excess return. we further find that national funds are unable to overcome the return disadvantage created by the imposition of state income taxes. thus, when the choice for the investor is to invest in one of the two sets of funds, it is beneficial, on a risk-adjusted basis, for the investor to allocate his or her investments to funds that are free from both federal and the investor’s state income tax. the preceding assertion can only be made, however, for residents of california and new york, since our sample of state-spe cific bond funds includes mutual funds pertaining to these states only. references benson, e. d., marks, b. r., & raman, k. k., (1991). the effect of voluntary gaap compliance and financial disclosure on governmental borrowing costs; professional adaptation. journal of accounting, auditing and finance, 6, 303-324. blake, c. r., elton, e. j., & gruber, m. j. (1993). the performance of bond mutual funds. journal of business, 66,37 l-403. brown, s., goetzmann w., ibbotson r., & ross, s. (1992). survivorship bias in performance studies. review of financial studies, 5,553-580. brown, s., goetzmann w., ibbotson r., & ross, s. (1995). performance persistence. journal of finance, 50,679-698. cornell, b., & green, k. (1991). the investment performance of low-grade bond funds. journal of finance, 46, 29-48. dwyer, p. d., & wilson e. (1989). an empirical investigation of the factors affecting the timeliness of reporting by municipalities. journal of accounting c? public policy, 8, 29-55. ehsan, f., & wilson, e. (1992). market segmentation and the association between municipal finan cial disclosure and net interest costs. accounting review, 67,480-495. gudikunst, a., & mccarthy j. (1992). determinants of bond mutual fund performance. journal of fixed income, 2,95-101. grinblatt m., & titman s. (1989). mutual fund performance: an analysis of quarterly portfolio hold ings. journal of business, 69, 393-416. grinblatt m., & titman s. (1993). performance measurement without benchmark: an examination of mutual fund returns. journal of business, 66,47--68. ilmanen, a., mcguire d., & warga a. (1994). the value of duration as a risk measure for corporate debt. journal of fixed income, 4,70-76. ilmanen, a. (1992). how well does duration measure interest rate risk? journal of fixed income, i, 43-5 1. investment company institute. (1995). mutual fund 1995 fact book, 35th edition. washington, dc: 101-l 10. ingram, r. w., raman, k. k., & wilson, e. r. (1989). the information in governmental accounting reports: a contemporaneous price reaction approach. accounting review, 64,250-268. jensen, m. (1968). the performance of mutual funds in the period 1945-1964. journal of finance, 23,318-416. lehman, b. n., & modest, d. n. (1987). mutual fund performance valuation: a comparison of benchmark and benchmark comparisons. journal of finance, 42,233-265. 196 financial services review 6(3) 1998 levitt, a. (1993). the state of the municipal securities market. government finance review, 9, 33 35. lewis, b. l., patton, j. m., & green, s. l. (1988). the effect of information choice and information use on analysts’ prediction on municipal bond rating changes. accounting review, 63, 270 282. marquette, r. p., & wilson, e. r. (1992). the case for mandatory municipal disclosure: do seasoned municipal bond yields impound publicly available information? journal ofaccounting & pub lic policy, ii, 181-206. mossavar-rahamani, s. (1987). understanding and evaluating index fund management. in fabozzi, f. j., & galicki, t. d. (eds.), advances in bond analysis and portfolio strategies (pp. 433 437). illinois: porbus publishing. reeve, j. m., & herring, h. c. (1986). an examination of non rated municipal bonds. journal of eco nomics and business, 38,65-76. shukla, r. k., & trzcinka, c. (1992). performance evaluation of managed portfolios. financial mar kets, institutions & instruments. cambridge, ma: blackwell publishers. pii: 1057-0810(92)90003-u financial services review, 2(2): 73-85 copyright q 1993 by jai press inc. all rights of reproduction in any fom reserved. an index of portfolio diversification walt woerheide don persson a recurring question in the literature concerning diversification is what is the minimum number of securities required to achieve adequate diversification. the problem is that studies on this topic assume equally distributed holdings. in reality, portfolios are not evenly divided. the purpose of this paper is to evaluate the ability offive different measures of diversification to provide meaningful information about the degree of diversi’cation of an unevenly distributed stock portfolio. the complement of the herjindahl index was found to be the best of the jive measures and its explanatory power was deemed to be adequate for general use. i. introduction a critical question for any investor is how many securities does one have to hold in order to achieve adequate diversification. complete diversification would be achieved if one held a share of the “market” portfolio, defined as the portfolio of all assets. adequate diversification is achieved when the variability of one’s portfolio is not significantly different than that of the market portfolio. the two classic studies which define the “minimum” portfolio size to be adequately diversified are evans and archer (e&a) (1968) and fisher and lorie (f&l) (1970). the word “minimum” is used in the sense that diversification beyond this size has little economic value in terms of risk reduction and may contain significant costs in terms of transaction fees and monitoring activity. a significant problem with these two studies and the rest of the literature that looks at the topic of portfolio size and diversification is that, as a practical matter, they do not actually answer the simple question of whether a specific portfolio is adequately diversified. virtually the entire literature on the question of portfolio size and diversification is based on portfolios that are evenly distributed. in reality, walt woerheide l rochester institute of technology, college of business, rochester, new york 14623-0887; don persson l ac rochester division, general motors, flint, mi. financial services review, 2(2) 1993 it would be incredibly rare for an investor to have his wealth evenly divided among securities. thus, all of our knowledge about adequate diversification is irrelevant unless we could tell investors whether an unevenly distributed portfolio of 15 securities is as adequately diversified as, say, an evenly distributed portfolio of 10 securities. furthermore, even if an investor held an evenly distributed portfolio at a point in time, such a portfolio would become unbalanced for two reasons. one reason is that prices of different securities change at different rates. another reason is the changes in portfolio composition that result from the addition and removal of cash (an event that occurs often in practice, and rarely in empirical research!). what investors need is a spot measure of diversification based on nothing more than the distribution of the weights representing the proportion of the portfolio invested in each security. this paper shows that an index of portfolio diversification can be constructed, which is as good in indicating the degree of diversification of an unevenly distributed portfolio as is the number of securities in indicating the degree of diversification of an evenly distributed portfolio. such an index could be used by individuals, financial advisors, and others to evaluate the degree of diver sification of any stock portfolio. section ii provides a review of the literature on the topic of portfolio size and diversification. section iii looks at five measures of diversification used in the industrial organization literature. section iv tests the quality of these measures in distinguishing portfolio diversification and recommends the complement of the herfindahl for future application. section v presents standards for diversification using the recommended index, and section vi provides a summary along with limitations and extensions. ii. hterature review portfolio size as indicated above, the two classic studies on the topic of portfolio size and diversification are those by e&a and f&l.’ as this research is modelled directly on that of e&a, we would like to provide a rather detailed description of their results. e&a compute the mean of the standard deviations for 60 equally weighted portfo lios at each size level ranging from 1 to 40 securities. they then regressed the mean portfolio standard deviations against the inverse of the number of securities in the portfolio and obtained the following estimate: y = .08625/n + .1191 where y = mean portfolio standard deviation, and n = number of securities in the portfolio. an index of portfolio diversifiation 75 e&a concluded that a ten-security portfolio provided adequate diversification. a graph of their results, or a variation of it, is presented in nearly all corporate finance and investments textbooks today. f&l measure risk by examining various measures of dispersion for wealth ratios over various time periods for portfolios of sizes 1, 2, 8, 16, 32, and 128 securities. for our purposes, their most significant result probably is the observation that approximately 80 percent of the achievable reduction in dispersion can be attained by holding eight stocks (the reductions range from 65 to 91 percent). in a follow-up study to e&a, upson, jessup, and matsumoto (1975) looked at the standard deviation of the standard deviations, and concluded that portfolio managers should diversify among more than 16 stocks, and that diversifying among even 30 or more stocks can be worthwhile in terms of risk reduction. statman (1987) argues that a well-diversified portfolio must include at least 30 to 40 stocks. statman’s analysis is based on the assumption that all investors have the opportunity to buy no-load index funds, and thus the cost of adding assets combined with the risk reduction benefits of adding these assets must be compared to the cost and risk of portfolios that combine the risk-free asset with an index fund. a variation on statman’s study by shanker (1989) shows that the conclusions about portfolio size are dependent on the size of the benchmark portfolio used for comparison and the assumed size of transaction fees. smaller benchmark portfolios suggest smaller optimal portfolio sizes, and smaller transaction fees imply larger optimal portfolios. a follow-up study by murphy (1991) questions the validity of the numbers used by statman, and concludes that portfolios of the size suggested by e&a and f&l may in fact provide the minimum necessary degree of diversification. much of the literature on portfolio size examines what happens to the standard deviation function in the e&a study if various conditions are placed on the types of stocks in the portfolio. in one of the most cited studies, solnik (1974) shows that more efficient diversification is possible when one considers foreign securities, particularly if one hedges for exchange rate risk. the greater efficiency in diversi fication is demonstrated by the result that e&a’s standard deviation curve declines at a faster rate and to a lower level when foreign securities are added to the stock population. wagner and lau (1971) show that far fewer stocks are necessary to achieve a specific level of diversification when the portfolio consists of stocks rated highly by the standard & poor stock guide than those rated poorly. klemkosky and martin (1975) show that diversification can be more readily achieved with low-beta stocks than with high-beta stocks. martin and klemkosky (1976) show that diversification can be more readily achieved when stock classifications are considered. their stock classifications included growth stocks, cyclical stocks, stable stocks, and oil stocks. all of the above studies are empirical. there are some theoretical studies that have shed light on the topic of portfolio size and diversification, goldsmith (1976) shows that not only do transaction fees limit the size of the number of securities in a portfolio, but they will also cause the optimal number of securities to hold in a 76 financial services review, 2(2) 1993 portfolio a function of an investor’s initial wealth. conine and tamarkin (1981) show that investor preference for positive skewness combined with other assump tions of perfect capital market may severely restrict the number of securities held by an individual even without transaction fees. measures of diversification most of the literature makes no statements about how to measure the degree of diversification. as shown above, the most common measure of diversification is simply to count the number of securities in the portfolio. however, this naive measure has meaning only when the portfolio is evenly invested across all holdings, and the point of this paper is to define a more effective measure when security holdings are not evenly distributed. the one other method suggested in the literature for measuring the degree of diversification of a portfolio is to examine the correlation coefficient between the rates of return on that portfolio and the rates of return on a surrogate for the market portfolio (see sharpe and alexander (1990, p. 654)). this measure obviously requires data from a period of several years and poses a serious estimation problem any time the portfolio’s composition has changed. iii. defining new measures of diversification in this section of the paper, we define various indices of diversification that could be used at a point in time. to dream up such measures would be haphazard and arbitrary. therefore, the authors have elected to examine measures of diversification that have been used in the industrial organization literature on concentration? a thorough review of the literature revealed five indices that would be plausible to measure portfolio diversification. in all cases, there is no profound reason for the use of a particular functional form, other than it is mathematically different and may empirically work better than other forms. the first one is the complement of the herfindahl index, perhaps the most widely used measure of economic concentration.3 thus, our first proposed index is di(l)=l-hi=l-i wi’ i=l where di = diversification index, hi = herfindahl index. an index of portflio diversifxation 77 wi = the proportion of portfolio market value invested in security i (in decimal form), and n = the number of securities in the portfolio. our use of the complement of this index is for the stylistic purpose of altering the index value so that zero represents a portfolio with absolutely no diversification (a one security portfolio) and 1 .o would represent the ultimate in diversification.4 to facilitate the reader developing a feel for these various indices, index values for nine sample portfolios are shown in table ia. (the compositions of the various portfolios are listed in table 1b.) the first portfolio (portfolio a) contains only one security and thus has absolutely no diversification. the ninth portfolio (portfolio i) contains 100 evenly distributed securities. note that for our first index the portfolio with only one security has an index value of zero, and the portfolio of 100 securities has an index value of .99. the second index we consider is the complement of one originated by rosenbluth (1961) and described in marfels (1971). in this index, security holdings are ranked in descending order by size with the i-th firm receiving rank i.5 the index value is di(2) = 1 l/(2 x i i(wi) 1). i=l note that in table 1a these first two indices and the next one to be introduced provide identical values for portfolios g, h, and i. it is easily shown that all three indices are mathematically equivalent to the term [l l/n] when security holdings are evenly held. the third index is defined by marfels (197 1) as the “exponential of the entropy measure” and is computed as n di(3) = 1 n wiawi. i=2 a fourth measure of diversification is the complement of one offered by horvath (1972) and named by him the “comprehensive concentration index.” it is n di(4) = 1 w, c we [ 1 + (1 wi)] i=2 where w, = the largest single portfolio holding. the index value for a single-security portfolio (portfolio a) is zero, and for an evenly weighted portfolio of 100 securities 78 i?inancialservicesrevijzw,2(2) 1993 table 1a. values of diversification indices for various portfolios diversification index a b c d e f g h i di (1): herfindahl 0 a4 .61 .70 .76 .79 .88 .90 .99 di (2): rosenbluth 0 .40 .57 .67 .73 .77 .88 .90 .99 di (3): exp. of entropy 0 .47 .64 .72 .77 .81 .88 .90 .99 di (4): cc1 0 .15 .26 .36 .43 .49 .67 .73 .97 di (5): entropy 0 .64 1.01 1.28 1.49 1.66 2.08 2.30 4.60 security no. 1 2 3 4 5 6 table 1b. composition of portfolios used in table 1a % distribution in portjolio a b c d e f 100% 66 2/3% 50% 40% 33 l/3% 28.6% 33 l/3 33 l/3 30 26 2l3 23.8 16 213 20 20 19.0 10 13 l/3 14.3 62f3 9.5 4.8 portfolio g: 8 securities, evenly distributed, (12 l/2% each). portfolio h: 10 securities, evenly distributed, (10% each). portfolio i: 100 securities, evenly distributed (1% each). (portfolio i) is .97. this and the next index are the only indices we are considering that, for evenly distributed portfolios, do not reduce to the value of [ 1 l/n]. our final index is the entropy measure. it is defined in hart (197 1) as di(5) = 2 wi ln(wi). i=l where in = natural logarithm. the entropy measure is distinct from the others in that it is not constrained between zero and one. note that in table la, the index values for all the sample portfolios except a and b are greater than one. iv. evaluatingthequalityofthefiveindicfs in order to evaluate the quality of the indices proposed in section iii, we empirically examine the relationship between the standard deviation of returns of randomly selected portfolios and the respective indices of diversification. the quality of each index is measured by the closeness of the fit (in regression terms) between portfolio risk (i.e., the standard deviation of returns) and the index number. the methodology an index of portfolio diversifhtion 79 is the same as that used by e&a, except that the distribution of securities in our portfolio is based on randomly determined weights, rather than evenly distributed weights. a total of 1,740 portfolios are examined. there are 60 portfolios which contain two securities each, 60 with three each, and so on up to 60 portfolios which contain 30 securities each.6 to compute the standard deviation for each portfolio we obtained monthly rates of return data from the 1985 crsp tapes.7 we used the rate of return series which included dividends. we selected only those companies with monthly rates of return covering the entire period from december 1965 to december 1985. this provided us with a sample size of 483 companies.’ we computed the monthly value relative for each portfolio as n ri = c wi ri,k k=l where ri,k = the value relative for security k during month i (the value relative is the price at the end of the month plus any dividend paid during the month, divided by the price at the start of the month). next, the monthly mean value relative for each portfolio was computed as m rp = exp(( 11 m) x c 1ogeri). i=l where m = 240 months. finally, the portfolio standard deviation was computed as m sd, = [( l/(m 1)) x c (log& 10g,~i)2]‘, i=l it should be noted that implicit in these formulations is the assumption that each portfolio is reweighted at the start of each month to the original distribution of that portfolio.9 in order to ascertain the weights of the securities within each portfolio, we generated one random number for each security in the portfolio, summed the random numbers, and set the weights equal to the ratio of each random number to the sum. so if we were constructing a two-security portfolio and our two random numbers were 43 and 82, then our weights were. 344 (= 43/125) and. 656 (= 82/125).” once the weights were computed, we then calculated the five index values for each portfolio. clearly, portfolios with the same number of securities were unlikely to produce the same index values. the means for each of the five indices for portfolios of selected sizes are shown in table 2. although the first four indices are theoretically 80 financial services review, 2(2) 1993 table 2. average index values for selected portfolio sizes number of securities in portfolio 2 3 4 5 6 7 8 9 10 12 14 16 18 20 25 30 i .384 .575 .676 .741 .781 .814 .836 .852 .868 .889 .904 .917 .926 .934 .947 .956 diver$ication index 2 3 4 5 ,357 .421 .124 ~561 ,547 .608 .238 .943 .654 .703 .335 1.221 .720 .762 .415 1.441 .761 .801 .475 1.619 .795 .830 .527 1.772 .820 .a51 .570 1.905 .838 .866 .600 2.012 .853 .879 .633 2.115 .877 .899 .680 2.296 .a94 .913 ,718 2.447 .907 .924 .749 2.581 .918 .933 .774 2.704 .926 ,940 .795 2.813 ,941 .952 .832 3.028 .950 ,960 .856 3.212 bounded by zero and one, note that indices 1,2, and 3 jump to the upper end of the interval fairly quickly. for these three indices, portfolios with five or more securi ties have index values, on average, between .720 and .999. the point is that nearly three-fourths of the scale available to measure diversification has been used up in measuring the diversification of portfolios with one to five securities. this leaves approximately one-fourth of the scale to measure differences in diversification of portfolios with more than five securities. this imbalance in the use of the scale appears to be a deficiency overcome by the last two indices. in fact, the fifth index has the advantage of being open-ended on the upper end. as the purpose of our research is to show that a diversification index could measure the degree of diversification of an unevenly distributed portfolio as well as the number of securities could measure the degree of diversification of an evenly distributed portfolio, we must first show how well the latter measures the degree of diversification. a standard can be established by repeating the e&a study, changing only the range of portfolio sizes. that is, we compute portfolio standard deviations for 60 sets of portfolios, where each set of portfolios range in size from two to 30 securities and each portfolio is equally weighted. we then compute the average portfolio standard deviation for each size portfolio and regress these average portfolio standard deviations against the inverse of the number of securities in the portfolio. the results of this regression are shown as regression 1 in table 3. the regression statistics are essentially the same as those obtained by e&a, the differ ence being that they used ten years of semiannual rates of return data and portfolios an index of portfolio diversijication 81 table 3. standard deviations regressed against index values constant coeficient root sample regression # (t-statistic) (t-statistic) adjusted r2 f-ratio mse size 1 (l/x;equal weights, avgs.) 4.063 6.132 ,980 1,398.91 .090 29 (170.597) (37.403) 2 (l/x;equal weights) 4.063 6.133 549 2,123.56 567 1740 (210.141) (46.082) 3-di( 1) 9.880 -5.765 .548 2,110.69 ,688 1740 (89.917) (-45.942) 4-di(2) 9.590 -5.505 ,544 2,077.2 .691 1740 (91.791) (45.576) 5-di(3) 10.337 -6.213 ,544 2.075.8 .691 1740 (85.637) (-45.561) 6-di(4) 7.445 -3.783 .511 1,822.o .715 1740 (119.534) (42.685) 7-di(5) 7.306 -1.011 ,479 1,602.3 .739 1740 (116.133) (-40.029) note: this table is based on 60 repetitions where each repetition has portfolios containing 2 to 30 securities. the data for each firm is based on 240 monthly rates of return. of one to 40 securities, whereas we use 20 years of monthly data and portfolios of two to 30 securities. the adjusted r-square is .980 and the sample size is 29. although comparable to that reported by e&a, regression 1 is not yet a valid standard for evaluating our indices. the reason is that our diversification indices are continuous variables, but the portfolio sizes used in the e&a study are discrete variables. furthermore, the standard deviations used as the dependent variable in regression 1 are arithmetic averages covering 60 observations. a valid comparison with our indices could be made if we rerun the e&a regression using the individual portfolio data rather than portfolio averages. thus, the sample size increases from 29 to 1,740. naturally, the regression coefficients are the same, but there is a substantial decline in the adjusted r-square to s49 as reported in regression 2 in table 3. so the relevant question becomes, can any of the diversification indices for the unevenly distributed portfolios produce an adjusted r-square as good as the r2 in regression 2. regressions 3 through 7 reported in table 3 are regressions of the portfolio standard deviations against the diversification indices for each of the portfolios. di(l) (the complement of the herfindahl), provides virtually the same explanatory power as the e&a study with an adjusted r-square of s48. di(2) and di(3) appear as good as di( 1) with adjusted r-squares of .544 for each. di(4) and di(5) are clearly weaker indices with adjusted r-squares of .5 11 and .479.” based on the results in table 3, we conclude that at least three indices exist which explain the degree of diversification of unevenly distributed portfolios as well as portfolio size does for evenly distributed portfolios. although we believe any of the three could be used to measure portfolio diversification, we recommend the complement of the herfindahl (di(l)) because we find it to be the best known of 82 f'inancialservicesreview,2(2) 1993 the various measures and it is a simpler mathematical computation than the other two. v. standardsofdiversification the use of the classical works to define a standard in this section, we seek to clarify the explicit or implicit criteria used by e&a and f&l to determine the minimum number of securities in an evenly-weighted portfolio necessary to achieve adequate diversification. we then will apply these same criteria to ascertain the minimum value of index di(l) that is necessary to achieve adequate diversification. as stated in section ii, e&a conclude that a ten-security portfolio provides adequate diversification. for one and 10 security portfolios, the mean standard deviations using their equation would be .20535 and .127725. the reported standard deviation for the portfolio containing all the securities in the population (the surrogate market portfolio) was . 1166. thus, e&a imply that a reasonably diversi fied portfolio size provides a reduction in unsystematic risk of 87.5 percent ((.20535 -. 127725)/(.20535 .1166) = .875). f&l reported that approximately 80 percent of the achievable reduction in dispersion can be attained by holding eight stocks. this 80 percent is quite close to the 87.5 percent reduction in unsystematic risk indicated by e&a’s study. if we accept the criterion for the minimum size portfolio to achieve reasonable diversification is one that on average reduces unsystematic risk by 80 to 87.5 percent, we can compute the diversification index value that corresponds to each of these numbers. the mean standard deviation for all 483 securities in our sample is 8.77 percent and the standard deviation for the portfolio consisting of all the securities (our surrogate for the market portfolio) is 4.01 percent. an 80 percent reduction in unsystematic risk would imply a portfolio with a standard deviation of 4.96 percent (4.01 + .20 x (8.77 4.01)). similarly, an 87.5 percent reduction in unsystematic risk would imply a portfolio with a standard deviation of 4.60 percent (4.01 + .125 x (8.77 4.01)). if we use these values as the dependent variable in conjunction with the coefficients from regression 3, we obtain index values of .85 and .91.12 based on the explicit and implicit criteria used by e&a and f&l, we offer as an observation that index values of less than .85 would imply a portfolio was probably not adequately diversified. portfolios with index values greater than .91 probably are adequately diversified. applying the index without a standard it is our hope that the diversification index identified in this paper becomes a standard component of brokerage account statements. if the cutoff values for diversification are not universally or even popularly agreed to, then it is doubtful an index of portfolio diversifeation 83 our index will come into common use. there is an alternative way to provide investors with the information provided by our index, without actually citing the index itself. the alternative is that in lieu of the index value, investors could be told the diversification equivalent of their portfolio if their portfolio were evenly distributed.i3 for example, we know that an evenly distributed portfolio of eight securities produces an index value of .88. therefore, if the index for a portfolio rounded off to .88, the investor could be told simply that his diversification is equivalent to an evenly distributed portfolio of eight securities and the investor could decide for himself if this was sufficient diversification. vi. sijmmary and conclusions, limitations and extensions the purpose of this paper has been to explore the two questions of 1) whether a diversification index could measure the degree of diversification of an unevenly distributed portfolio as well as the number of securities does when the holdings are evenly distributed, and 2) if such an index exists, what index value represents a reasonably diversified portfolio. variations of five measures of diversification that are frequently used in the industrial organization literature on industry diversification were identified and tested. three of these five indices were found to be satisfactory measures of diversification. the complement of the herfindahl index was recommended as the measure of diversification for unevenly distributed portfolios based on its simplicity and wide-spread usage. based on the explicit and implicit criteria used by e&a and f&l, we concluded that index values of less than .85 imply that a portfolio was probably not adequately diversified. portfolios with index values greater than .91 were probably adequately diversified. we also indicated that the index could be used to define an evenly distributed portfolio equivalency. although this research does not provide information for the important question of whether a portfolio lies on the efficient frontier, it does provide information important to all investors. there are obvious deficiencies to the index as proposed. the index and standards presented herein are for portfolios consisting entirely of common stock. future work will need to focus on the impact of such non-stock holdings as options, futures contracts, bonds, treasury securities, money market instruments, and vari ous types of mutual funds. in addition, the impact of buying on margin will also have to be considered. nonetheless, we think the index proposed here is a good first step to providing a critical tool for investors, but we acknowledge many more steps are necessary for practical application. acknowledgments: the authors wish to acknowledge the helpful sugges tions of david bumie, and the business faculty at the university of michigan flint, on an earlier draft, as well as the suggestions of an “anonymous” referee for this journal. 84 financial services review, 2(2) 1993 notes 1. for an excellent review some of the literature on diversification, the reader is referred to alexander and francis (1986), pp. 193-202. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. related to concentration ratios are inequality measures. we considered measures of inequality such as the lorenz-curve-based gini index, but found them to be internally inconsistent for our purposes and therefore do not include them here. see, e.g., polakoff, et al., (1981, p. 684); or lovett, (1988, pp. 197-202). the use of the complement to distinguish diversification indices from concentration indices is common in the industrial organization literature. see acar, et al.(unpublished). hart (1971, p. 77) shows that the rosenbluth index is a modified version of the gini coefficient. hall and tideman (1967) have also proposed the same measure as rosenbluth. the decision to use a maximum portfolio size of 30 securities is somewhat arbitrary. our early work on this paper used a maximum portfolio size of 40 securities. the 10 additional securities added at least a day to the time it took our computer program to run. more importantly, the 10 additional portfolios had the effect of making the empirical tests of our results appear stronger than they really were because our indices clearly indicate there is no significant difference in the diversification of 30and 40-security portfolios. at the time the data was collected for this research, more recent tapes were available but they included the year 1987. we wanted to avoid any potential problems associated with the dramatic market movements of that year. although we believe our results are independent of the time period used, we are doing follow-up tests with a later time period. this method of selection obviously creates a survivorship bias. as we are not aware of any evidence that nonsurvivor firms have variances and covariances different from survivor firms, it is not felt that this bias is of any significance. these formulas are the same as those used by e&a except that we have adjusted them to allow for unequal weights. e&a also implicitly reweight their portfolios each month to the original (even) distribution. to generate the random numbers, we used the randomize command in basica. the seed number was one for the first portfolio, and was augmented by one for each successive set of weights computed. at the suggestion of an earlier referee, we twice reran the last five regressions. first we added as a second independent variable the number of securities in each the inverse of the number of securities. the change in adjusted r ? ortfolio, and then we added ‘s was negligible. specifically, let 4.96 = 9.88 -5.76 * di(l), then di(l) = .85. also, let 4.60= 9.88 5.76 * di(l), then di(i) = .91. our thanks to lew mandell for this suggestion. references acar, william, pankaj bhatnagar, kizhekepat sankaran, and kenneth d. gartrell. “calibrating measures of concentration and diversification for the ‘distribution effect’.” unpublished. alexander, gordon j., jack clark francis. 1986. portfolio analysis, third edition, prentice-hall. conine, t. e., and m. j. tamarkin. 198 1. “on diversification given asymmetry in returns,” joumuf of finance, 36: 1143-l 155. evans, john, and stephen archer. 1968. “diversification and the reduction of dispersion: an empirical analysis,” journal of finance, xxiv: 761-769. fisher, lawrence, and james h. lorie. 1970. “some studies of variability of returns on investments in common stocks,” journal of business, 43: 99-l 17 an index of portfolio diversification 85 goldsmith, david. 1976. “transactions costs and the theory of portfolio selection, ” journal of finance, 31: 1127-1139. hall, m., and n. tideman. 1967. “measures of concentration,” journal of the american statistical association, 62: 162-168. hart, p.e. 1971. “entropy and other measures of concentration,” journal of the royal statistical society, 134: 73-85. horvath, janos. 1972. “absolute and relative measures of diversification reconsidered: a com ment,” kyklos, 24: 841-843. klernkosky, robert c., and john d. martin. 1975. “the effect of market risk on portfolio diversifi cation,” journal of finance, xxx: 147-154. lovett, william a. 1988. banking and financial institutions law in a nutshell, second edition, west publishing co. marfels, christian. 1971. “absolute and relative measures of concentration reconsidered,” kyklos, 24: 753-766. martin, john d., and robert c. klemkosky. 1976. “the effect of homogeneous stock groupings on portfolio risk,” journal of financial and quantitative analysis, 12: 181-195. murphy, j. austin. 1991. “evaluating diversification adequacy with different asset risk estimates,” new york economic review, xxi: 50-55. polakoff, murray, and thomas a. durkin (eds.). 1981. financial institutions and markets, second edition, houghton mifflin. rosenbluth, g. 1961. “address to ‘round-table-gesprach uber messung der industriellen konzen tration”‘, die konzentration in der wirtschafi, edited by f. neumark, schriften des vereins fur socialpolitik, n.s., 22: 391-394. shanker, latha. “benchmark portfolios, transactions costs and the number of stocks in a diversified portfolio,” presented at the 1989 meeting of the midwest finance association, cincinnati, ohio. sharpe, william f., gordon j. alexander. 1990. investments, fourth edition, prentice-hall. solnik, bruno. 1974. “why not diversify internationally rather than domestically?’ financial analysts journal, 30: 48-54. statman, meir. 1987. “how many stocks make a diversified portfolio?’ journal of financial and quantitative analysis, 22: 353-363. upson, roger b., paul f. jessup, and keishiro matsumoto. 1975. “portfolio diversification strate gies,” financial analysts journal, 3 1: 86-88. wagner, w.h., and sc. lau. 1971. ‘the effect of diversification on risk,” financial analysts journal, 27: 48-53. pii: s1057-0810(97)90003-2 financial services review, 6(4): 243-256 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. an analysis of nondeductible ira contributions and roth ira conversions stephen m. horan jeffrey h. peterson robert mcleod on average investors have an income replacement rate of 64% of their pre-retirement income, which in many cases results in a lower tax rate in retirement. we analyze the impact of declining withdrawal tax rates on the choice between taxable mutual fund investments and nondeductible iras. the relative attractiveness of the taxable mutual fund option declines significantly when withdrawal tax rates decline. converting exist ing iras to roth iras is generally beneficial for investors who remain in the same tax bracket upon withdrawal. for short (long) time horizons and low (high) expected returns, the marginal value of conversion in i998 is greater (less) than the marginal value of optimal conversion. for investors dropping into the 15% tax bracket, conver sion is generally not beneficial unless the conversion is done optimally, the time horizon is long, and the expected return is high. investors in the 15% tax bracket should convert existing ira assets. i. introduction disagreement exists about the true tax advantages of traditional iras. some have sug gested that traditional iras have tax advantages only if the investor’s marginal tax rate upon withdrawal is lower than the investor’s marginal tax rate upon contribution (e.g., kai ser, 1990). others have argued that the deferral of tax payments on investment returns over time yields a tax advantage to iras even if current and terminal marginal tax rates do not differ (see bodie & merton, 1998). the significance of this tax-deferment value has been called into question however. crain and austin (1997) analyze the tradeoff between tax deferment value in iras and the preferential treatment of capital gains in a taxable invest ment enacted in the taxpayer relief act of 1997 (hereafter, the act) for mutual fund inves tors in the 31% tax bracket. in this paper, we seek to extend their analysis in two ways. stephen m. horan and jeffrey h. peterson l school of business, st. bonaventure university, st. bonaventure, new york 14778; e-mail: shoran@sbu.edu, or e-mail: peterson@sbu.edu. robert mcleod l department of economics, finance and legal studies, university of alabama, tuscaloosa, alabama 35487-0224; e-mail: rmcleod@cba.ua.edu. 244 financial services review 6(4) 1997 first, we allow investors to drop into lower tax brackets upon withdrawal of retirement assets. this possibility has a significant effect on an investor’s choice between nondeduct ible ira contributions and taxable mutual fund investments. second, we analyze the option to convert existing ira assets to roth ira assets. according to the panel study of income dynamics (psid) as interpreted by bern heim, skinner and wienberg (1997), on average, retirement income is about 64% of pre retirement income, suggesting that marginal tax rates for retirees are likely to fall over their investment horizon. many of these individuals face the decision of making nondeductible ira contributions (in which earnings will be taxed upon withdrawal as ordinary income) or taxable mutual fund investments (in which a portion of earnings will be taxed each year at a lower capital gains tax rate). we compare the after-tax future values of nondeductible ira contributions and taxable mutual fund investments for investors whose marginal tax rate falls from 31% during the term of investment to 28% upon withdrawal. meanwhile, the act has also given taxpayers in lower tax brackets the option to convert traditional iras to roth iras, a retirement account in which contributions are not deductible, but all earnings are excluded from tax. although the conversion permits tax-free earnings into the future, it requires all deductible contributions and earnings being converted from the tradi tional ira to be added to taxable income, creating an opportunity cost in that the tax paid is not able to accumulate earnings. the act also creates a one-time incentive to convert (or a one-time tax revenue advance) by allowing those who convert in the 1998 tax year to spread the additional taxable income over four years. we analyze the conversion option both with and without the 1998 conversion incentive. the following section reviews the literature on ira investing. section iii analyzes the decision between nondeductible ira and taxable mutual fund investments for tax rates that decline upon retirement using the model in crain and austin (1997). we analyze the roth ira conversion decision in section iv. the final section concludes and offers avenues for further investigation. ii. ira literature and methodology there exists a debate over whether tax-deferred savings programs actually increase savings rates. some estimate that iras have had little, if any, impact on savings (e.g., gale & scholz, 1994). others report evidence that iras have stimulated savings (e.g., hubbard & skinner, 1995, 1996). perhaps the ambiguity over whether iras increase savings rates stems from the complexity of the tax code, which creates ambiguity over whether investors will have greater after-tax accumulations in taxable or tax deferred accounts. for example, o’neil, saftner, and dillaway (1983) examine the impact of the 10% premature with drawal penalty. they find that the penalty outweighs the tax deferment for short invest ment horizons, but not for long horizons. in fact, yaari and fabozzi (1985) find that the indifference point may be as short as two years. ragdsdale, seila, and little (1993, 1994) incorporate many tax code complexities in a mathematical programming model for optimal withdrawal policies from tax-deferred retirement accounts based on investment returns, life expectancy, and beneficiary designa tions. they demonstrate that their model significantly improves upon results of proposed hueristic rules (see saftner & fink, 1990, and sage, 1988), which can produce large finan ira contributions and conversions 245 cial losses in certain circumstances. ragdsdale, seila, and little highlight the usefulness of mathematical models for optimal decision making. similarly, crain and austin (1997) develop a mathematical model to analyze the choice between taxable investments, deductible iras, nondeductible iras and roth iras. making a distinction between ordinary income tax rates and capital gain tax rates, they build on the work of randolf (1994) who examines similar issues with mutual funds that make periodic taxable distributions. aithough randolf makes no distinction between ordi nary and capital gains tax rates, he demonstrates that mutual funds with high turnover and dist~butions (such as some aggressive growth funds) should be in iras, while mutual funds with low turnover and distributions (such as index funds) should be placed in taxable accounts when both tax deferred and taxable savings accounts are used. cram and austin establish that when investors expect to be in a lower (higher) tax bracket upon withdrawai than upon contribution, deductible iras accumulate more (less) than roth iras. they also find that roth iras accumulate more than nondeductible traditional iras. crain and austin examine the break-even points for the percent of return distributed as capital gains which make investors indifferent between nandeductible ira investments or a taxable investment in the same mutual fund. they recognize that a trade-off exists between deferring a tax liability or accepting a lower capital gains tax sooner. their analy sis is limited to investors facing a 3 1% marginal tax rate throughout the investment horizon and upon withdrawal. our initial analysis is similar except that we analyze the impact of investors facing lower tax rates upon withdrawal. investors facing 28% tax rates are not forced to choose between nondeductible iras and taxable mutual fund investments since deductible iras and roth iras are always preferable (see crain & austin) and the phase out limits for roth ira are well into the 3 1% tax bracket. hence, investors in the 28% tax bracket should not make nondeductible ira contributions at all. however, many investors in the 3 1% tax bracket that must choose between nondeductible iras and taxable mutual fund investments are likely to fall into lower tax brackets upon retirement. this scenario is the setting for the first part of our analysis. the act also permits taxpayers to convert an existing ira to a roth ira. any pre-tax contributions and tax-deferred earnings at that point are considered a taxable distribution, although the 10% penalty for early wi~drawal will not apply. eligibility for conversion is limited to taxpayers-married or single-whose agi is under $l~,~. those converting in 1998 can spread the taxable income evenly over four years. we examine the value of conversion for taxpayers in the 15 and 28% tax brackets. generally, individuals facing a 3 1% marginal rate are not eligible for conversion. the analysis in this paper permits tax rates upon contribution and withdrawal to vary. the after-tax accumulations of converting in 1998 and post1998 are also compared, illustrating that the advantage to converting in 1998 is significant. iii. changing tax rates a. the foundations comparing future values of taxable investments and nondeductible ira invest ments requires establishing the formulas governing their after-tax accumulations. for 246 financial services review 6(4) 1997 a nondeductible ira contribution, all returns are taxed at the ordinary rate upon with drawal, while the nondeductible contribution is excluded from withdrawal tax. hence, the after-tax future value of a dollar invested in a nondeductible ira for n years is f”nira = 1 + [( 1 + t)n 1]( 1 t,) (1) where t,, = the ordinary marginal tax rate on income upon withdrawal; r = the expected rate of return on the investment; and n = the number of years until withdrawal of the contribution. equation (1) will hold assuming the investor does not take a 10% penalty for with drawal before age 59 l/2. it is important to note that even when ira withdrawals com mence in the near future, the investment horizon is not necessarily short. for example, suppose a 55year old investor will start making withdrawals at age 60. suppose fur ther, that other assets are already in place such that withdrawals would begin at age 60, whether or not the considered investment was made. in this case, n is not equal to five years. rather, n is equal to the time until the marginal ira withdrawal is made. if mar ginal withdrawals (as a result of the investment made at age 55) begin at age 70, n = 15 years. mutual funds are required to distribute interest income, dividend income, and short term capital gains to fund shareholders on a prorata basis as dividends, which are taxable to the investor at the ordinary income tax rate. long-term capital gains realized by selling appreciated stock are distributed to shareholders on a prorata basis as well and are taxable to investors at the long-term capital gains rate. crain and austin show that for a taxable mutual fund investment, a dollar invested for n years has a before-withdrawal tax future value of where toi = the intermediate marginal tax rate on ordinary income over the term of the investment: t cg = the intermediate marginal tax rate on capital gains over the term of the invest ment; poi = the percent of annual return distributed to shareholders as ordinary income; and pc8 = the percent of annual return distributed to shareholders as capital gains. a capital gain is also recognized when mutual fund shares are sold. the tax is based on the before-withdrawal tax future value in equation (2) less the adjusted basis, which is composed of the initial investment plus dividend and capital gain distributions (on which tax has already been paid) less the income tax on those distributions. the after-withdrawal tax future value of a taxable mutual fund investment is ira contributions and conversions fvmo = (1 + r rpoitoi rp,t,,)” -tcg (1 + r rpoitoi rpcgt,g)n 1 +p,i( l t,i) ( 1 + r rpoitoi rpcgt,g)n 1 (r rpoitoirpcgtcg) -rp,i( l ‘oi) (i + r rpoitoi rpcgtcgf 1 (r rp,jt,i rpcgtcg) 247 (3) crain and austin (1997) present a more thorough development of equation (3). the term inside the brackets represents the before-withdrawal tax accumulation less the adjusted basis. there exists a tradeoff between the taxable mutual fund investment and the same investment in a nondeductible ira. investors may be willing to forgo the tax deferral offered by iras in exchange for paying the lower capital gains tax of 20% established by the act. in addition, since capital gains are not distributed until the fund sells appreci ated stock, the taxable mutual fund investment has an inherent, albeit partial, tax deferral element. b. taxable mutual funds vs. nondeductible iras for falling tax rates our objective in this section is to determine the percent of capital gains distribu tion that makes investors indifferent between a taxable mutual fund investment and the same investment in a nondeductible ira. this exercise provides guidance to investors in deciding between the two options under different conditions. our analysis high lights the impact of declining marginal tax rates during retirement years. since roth iras are always preferable to either taxable investments or nondeductible iras (see crain & austin) and the phase-out limits for roth ira are well into the 3 1% tax bracket, investors facing a 28% tax rate should always choose deductible iras or roth iras. investors with 31% marginal tax rates may have to choose between nonde ductible iras and taxable investments. however, it is likely that investors’ tax rates will decline during retirement. allowing the tax rate upon withdrawal to decline to 28% greatly affects an investor’s decision. as the withdrawal tax rate (t,) decreases, the advantage of paying the lower 20% capital gains tax early is reduced. hence, the relative attractiveness of the taxable mutual fund is reduced and the nondeductible ira becomes relatively more attractive than when the withdrawal tax rate remains con stant. the proportion of return distributed as ordinary income (poi) is initially set at .07, the average for growth funds reported in crain and austin (1997). further, the tax rate on ordinary income during the term of the investment (toi) is set at 31%, and the tax rate on capital gains over the term of the investment (t,,) is set at 20%. for a given return (r) and time horizon (n), we set equation (1) equal to equation (3) and solve for the percent of return distributed as capital gain (pc& subject to the constraint that 0 < peg < 1. alge braically, 248 flnancial services review 6(4) 1997 1 +[(l +r)“-l](l-t,)=(l +r-rpo~toj-rpcstcg)n -t cg (1 + r rpoitoi rp,gtcg)n 1 -rp,i( 1 ‘oil (1 + r rpoitoi rpcgtcg)’ 1 (r rp,if,; rpcgtcg) -rp,j(l t,j) ( 1 + r rpoitoi rp,gtcg)n 1 (rrpoitoi rpcgtcg) (4) s.t. 0 <peg < 1. table 1 displays the results for various expected returns and investment horizons. the approximate average peg for growth funds are in bold print to highlight the indifference point for an average growth fund. the table is interpreted as follows. for 20-year invest ment horizons and mutual funds with an 8% expected return, investors will accumulate the same after-tax future value in a taxable account and a nondeductible ira if the fund dis table 1 typical taxable growth mutual fund investment vs. nondeductible ira indifference points of percentage of investment return taxed as capital gain (peg) for selected returns (r) and time horizons (n) with the percentage of return distributed as ordinary income @oi) = .07 investment horizon in years (n) r 5 10 15 20 25 30 35 40 1% * * * * * * * * 2% * * * * * * * * 3% * * * * * 0.886 0.722 0.601 4% * * * * 0.788 0.614 0.492 0.403 5% * * * 0.804 0.591 0.453 0.357 0.286 6% * * 0.946 0.639 0.462 0.347 0.268 0.210 7% * * 0.785 0.522 0.371 0.274 0.207 0.158 8% * * 0.665 0.435 0.303 0.219 0.161 0.120 9% * * 0.572 0.368 0.252 0.178 0.127 0.090 10% * 0.894 0.498 0.315 0.211 0.145 0.100 0.068 11% * 0.799 0.438 0.272 0.178 0.119 0.079 0.050 12% * 0.719 0.389 0.237 0.152 0.098 0.06 1 0.035 13% * 0.653 0.347 0.207 0.129 0.080 0.047 0.023 14% * 0.596 0.312 0.183 0.110 0.065 0.035 0.023 15% * 0.547 0.282 0.161 0.094 0.053 0.025 0.004 16% * 0.504 0.255 0.143 0.08 1 0.042 0.016 ** 17% * 0.466 0.232 0.127 0.069 0.032 0.008 ** 18% * 0.433 0.212 0.113 0.058 0.024 0.001 ** 19% * 0.403 0.194 0.100 0.049 0.017 ** ** 20% * 0.377 0.178 0.089 0.041 0.011 ** ** notes: *taxable mutual fund investment yields a greater after-tax future value than the nondeductible ira. **nondeductible ira yields a greater after-tax future value than the taxable mutual fund investment. bold figures indicate approximate average capital gain distribution as a percent of return for the growth mutual funds in grain and austin (1998). ira contributions and conversions 249 table 2 typical taxable growth and income mutual fund investment vs. nondeductible ira indifference points of percentage of investment return taxed as capital gain qcg) for selected returns (i) and time horizons (n) with the percentage of return distributed as ordinary income (p,i) = .20 investment horizon in years (n) r 5 io i5 20 25 30 35 40 1% * * * * * * * * 2% * * * * * 0.938 0.739 0.591 3% * * * 0.970 0.68 1 0.493 0.361 0.264 4% * * * 0.630 0.414 0.274 0.176 0.103 5% * * 0.726 0.428 0.256 0.144 0.066 0.009 6% * * 0.543 0.294 0.151 0.059 ** ** 7% * 0.867 0.412 0.200 0.077 ** ** ** 8% * 0.714 0.315 0.129 0.023 ** ** ** 9% * 0.596 0.240 0,075 ** ** ** ** 10% * 0.501 0.181 0.032 ** ** ** ** 11% * 0.424 0.132 ** ** ** ** ** 12% * 0.360 0.092 ** ** ** ** ** 13% * 0.306 0.059 ** ** ** ** ** 14% * 0.260 0.030 ** ** ** ** ** 15% 0.991 0.220 0.006 ** ** ** ** ** 16% 0.911 0,185 ** ** ** ** ** ** 17% 0.841 0,155 ** ** ** ** ** ** 18% 0.778 0,128 ** ** ** ** ** ** 19% 0.722 0,104 ** ** ** ** ** ** 20% 0.672 0.083 ** ** ** ** ** ** notes: *taxable mutual fund investment yields a greater after-tax future value than the nondeductible ira. **nondeductible ira yields a greater after-tax future value than the taxable mutual fund investment. bold figures indicate approximate average capital gain distribution as a percent of return for the growth and income mutual funds in grain and austin (1997). table 3 instances in which the after-tax future value of a taxable investment in 10 randomly selected mutual funds from crain and austin (1997) exceeds the after-tax future value in a nondeductible ira under varying terminal tax rates t, = 31% t,, = 28% aggressive growth 5 2 growth 2 i growth and income 2 i balanced 0 0 tributes 43.5% of its return as capital gains. if the fund distributes less as capital gains, the nondeductible ira is preferred. if the fund distributes more as capital gains, the taxable investment is preferred. the results differ significantly from those reported in crain and austin (1997). for example, given an investment horizon of 20 years and a tax rate of 28% upon withdrawal, an investor requires an expected return of 8% to be indifferent with the typical growth 250 financial services review 6(4) 1997 fund. crain and austin (1997) report an 11% expected return is required when the with drawal tax rate remains at 3 1%. alternatively, given an expected return of 1 l%, a 15year investment horizon is required to be indifferent. crain and austin (1997) report that a 20 year investment horizon is required with a 3 1% withdrawal tax rate. hence, investors that anticipate a lower tax rate during withdrawal will favor a wider range of growth mutual funds in nondeductible iras. since growth funds in the crain and austin sample have a mean return of 14.58%, growth fund investors with horizons of at least ten or fifteen years should choose nondeductible iras, on average. the results for growth and income funds having an average pai = .20 are similar. according to table 2, given an investment horizon of 15 years and a tax rate of 28% upon withdrawal, an investor requires an expected return of 7% to be indifferent with the typical growth and income fund. crain and austin (1997) report that a 10% return is required when the withdrawal tax rate remains at 31%. table 2 shows that, given an expected return of lo%, a lo-year investment horizon is required to be indifferent. crain and austin (1997) report that a 15-year investment horizon is required with a 3 1% withdrawal tax rate. hence, investors that anticipate a lower tax rate during withdrawal will favor a wider range of growth and income mutual funds in nondeductible iras. since growth and income funds in the crain and austin sample have a mean return of 13&l%, growth and income fund investors with horizons of at least ten years should choose nondeductible iras, on average. using the fund-specific data reported by crain and austin, we determine the instances in which the after-tax future value of a taxable investment in ten randomly selected mutual funds exceeds the after-tax future value in a nondeductible ira assuming a 20-year invest ment horizon and the fund’s specific five-year mean return, mean poi and peg. these results are reported in table 3. for ten aggressive growth funds, crain and austin find that five would accumulate a larger amount as a taxable investment after 20 years if the withdrawal tax rate remains at 31%. if the withdrawal tax rate declines to 28%, the instances fall to two. similarly, for growth funds, only one (instead of two) would have accumulated larger amounts in taxable accounts if the withdrawal tax rate drops to 28%. the same is true for growth and income funds. no balanced funds would have yielded larger amounts as tax able investments under either scenario. for investors that fall into the 15% tax bracket, the nondeductible ira will always yield a higher after-tax future value. since the capital gains tax rate is 20%, there exists no tradeoff between paying a lower capital gain tax early or a higher withdrawal tax later. iv. converting roth iras the new roth ira allows all future earnings and withdrawals to be free from tax. the tax free nature of withdrawals associated with the roth ira is contingent upon the account having been establish for more than five years and the account holder being 59’/2 years old, permanently disabled, purchasing a first home, or dead upon withdrawal. the contribution, however, is taxable at the time of contribution. as such, the future value of an after-initial tax dollar invested in a roth ira is zz?a contributions and conversions fvroth = $l(l t,)(l + r)” 251 (5) where to = the initial marginal tax rate upon contribution, the future value of a dollar in an existing ira account is fvira = $l( 1 + ry( 1 t,). (6) investors with adjusted gross incomes of no more than $100,000 may convert existing traditional iras to roth iras. all deductible contributions and earnings increase taxable income. for the 1998 tax year only, the amount converted is taxed as income evenly over four years. if the tax liability is paid from the ira assets in the year of conversion then investors should convert when (1 t,)( 1 + t)n = (l-m)(l+i)n >l. (7) hence, investors should convert when their withdrawal tax rate is expected to be less than their current tax rate. when the initial tax rate is greater than or equal to the terminal tax rate, it is beneficial to convert existing traditional iras to roth iras. equation (5) assumes that all assets being converted are deductible contribu tions and earnings and, hence, subject to tax. we use this assumption throughout the analysis because there is no way to know what contributions were not deduct ible when made. paying the tax liability from the ira assets is sub-optimal, however, because the tax liability triggered by the conversion can be paid with dollars that would not qualify for tax deferment. for example, make the simplifying assumption that the converted amount is taxed entirely in the year of conversion at the initial tax rate, to. the conversion tax can be deducted from the assets being converted or paid directly by the investor. paying the tax liability out of the assets being converted decreases the principal in the new roth ira. in this case, the future value of the converted roth ira is simply (1 to)( 1 + r)“. alternatively, the tax liability can be paid from assets that would not qualify for tax deferred status, leaving the principal in the new roth ira unchanged from the traditional ira. this technique has the effect of lowering the opportunity cost associated with pay ing the conversion tax. in this case, the future value of a converted ira dollar is equal to the future value of the new roth ira dollar less the after-tax future value of conversion tax, to, or fvrorhconv = (1 + dn tjfvmal (8) where fvma is the after-tax future value of a taxable mutual fund investment from equation (3). the first term represents the future value of a dollar in the new roth ira. the second term represents the lost future value of the to dollars used to pay the con version tax. the latter tax-payment method is the preferred method of conversion and 252 financial services review 6(4) 1997 is assumed in our subsequent analysis. substituting equation (3) into equation (8), we have i (1 + rrpj,i rpcgtcgjn 1 (1 + r rpoitoi rpcgtj 1 fvrorhconv = (1 ++tu (1 + r rpoitoi rp&_j 1 (9) -t,g +-po& 1 t,i) (irp’oitoi rpcgt,j -rpoi( 1 foj) (1 + r rpoitoi rpcgtc,)” 1 (r rpoitoi rpcgt,j hence, the conversion should be made when fvrorhconv = fvira (i+ rln t,[fvtx,l , 1 (1 t,)( 1 + r)n (10) since the act permits the tax liability of iras converted in 1998 to be paid over four years, the value of converting is higher in 1998. to reflect this conversion incentive, we adjust the second term in the numerator to be the present value of a four-year annuity with payments equal to one-fourth of the second term. the discount rate is the required return table 4 after-tax future values of converted roth iras divided by after-tax future values of traditional iras when the tax rate on ordinary income is 28% investment horizon in years (n) r 5 io 15 20 2.5 30 35 40 1.070* 1.086 1.100 1.112 1.123 1.133 1.142 1.151 6% 1.021** 1.039 1.055 1.069 1.082 1.093 1.104 1.114 1.052*** 1.052 1.052 1.052 1.052 1.052 1.052 1.052 1.089 1.108 1.124 1.137 1.149 1.160 1.170 1.179 8% 1.027 1.050 1.068 1.085 1.099 1.112 1.125 1.136 1.067 1.067 1.067 1.067 1.067 1.067 1.067 1.067 1.107 1.127 1.144 1.159 1.171 1.183 1.193 1.203 10% 1.033 1.059 1.080 1.098 1.114 1.129 1.142 1.155 1.081 1.081 1.081 1.081 1.081 1.081 1.081 1.081 1.122 1.144 1.162 1.177 1.190 1.202 1.213 1.223 12% 1.038 1.067 1.090 1.110 1.127 1.143 1.158 1.171 1.094 1.094 1.094 1.094 1.094 1.094 1.094 1.094 1.137 1.160 1.178 1.193 1.207 1.219 1.230 1.241 14% 1.043 1.074 1.099 1.121 1.139 1.156 1.171 1.186 1.106 1.106 1.106 1.106 1.106 1.106 1.106 1.106 1.150 1.174 1.192 1.208 1.222 1.234 i.245 i.256 16% 1.047 1.081 1.108 1.130 1.150 1.168 1.184 1.199 1.117 1.117 1.117 1.117 1.117 1.117 1.117 1.117 notes: *top figures indicate optimally converted iras in 1998. **middle figures indicate optimally converted iras after 1998. ***bottom figures indicate sub-optimally converted iras in 1998 ira contributions and conversions 253 table 5 after-tax future values of converted roth iras divided by after-tax future values of traditional iras when the tax rate on ordinary income drops from 28% to 15% upon withdrawal investment horizon in years (n) r 5 10 is 20 2.5 30 35 40 0.907* 0.920 0.932 0.942 0.95 1 0.960 0.967 0.975 6% 0x65** 0.880 0.894 0.906 0.916 0.926 0.935 0.943 0.891*** 0.89 i 0.891 0.89 1 0.891 0.89 1 0.891 0.891 0.923 0.938 0.952 0.963 0.973 0.983 0.991 0.999 8% 0.870 0.889 0.905 0.919 0.93 1 0.942 0.953 0.962 0.904 0.904 0.904 0.904 0.904 0.904 0.904 0.904 0.937 0.955 0.969 0.98 1 0.992 1.002 i.011 1.019 10% 0.875 0.897 0.915 0.930 0.944 0.956 0.968 0.978 0.915 0.915 0.915 0.915 0.915 0.915 0.915 0.915 0.95 1 0.969 0.984 0.997 1.008 1.018 1.028 1.036 12% 0.879 0.904 0.924 0.940 0.955 0.968 0.98 1 0.992 0.926 0.926 0.926 0.926 0.926 0.926 0.926 0.926 0.963 0.982 0.998 1.01 i 1.022 1.033 1.042 1.051 14% 0.883 0.910 0.93 1 0.949 0.965 0.979 0.992 1.004 0.937 0.937 0.937 0.937 0.937 0.937 0.937 0.937 0.974 0.994 1.010 1.023 1.035 1.045 1.055 i .0&l 16% 0.887 0.916 0.938 0.957 0.974 0.989 1.003 1.015 0.946 0.946 0.946 0.946 0.946 0.946 0.946 0.946 noes: *top figures indicate optimally converted iras in 1998. **middle figures indicate optimally converted iras after 1998. ***bottom figures indicate sub-optimally converted iras in 1998. on the investment. table 4 shows the value of conversion whenthe tax rate on ordinary income remains at a constant 28%. for each return and investment horizon, three ratios are provided. the top figure is the value (relative to an unconverted traditional ira) of an opti mally converted ira in 1998 (i.e., tax liability paid with non-ira assets). the middle fig ure is the relative value of a sub-optimally converted ira in 1998. the bottom figure is the value of an optimally converted ira after 1998. the calculations are made using the aver age poi and peg for growth funds reported by crain and austin earlier. the ratios should be interpreted as the value of conversion relative to nonconversion. for example, given a 20-year investment horizon and 10% return, an optimally converted ira will have a 15.9% greater value than a nonconverted ira. several observations can be made from the trends in table 4. first, for investors who are likely to remain in the 28% tax bracket, conversion is almost always advisable. second, the value of converting from a traditional ira to a roth ira increases as the investment horizon, n, and rate of return, r, increase. for short time horizons and low expected returns, the marginal value of conver sion in 1998 is greater than the marginal value of optimal conversion. the effect reverses itself for higher returns and longer time horizons. in other words, for long time horizons or high returns, it is more important to optimally convert than it is to convert sub-optimally in 1998. table 5 illustrates the effect of investors falling into the 15% tax bracket upon retire ment. the difference is significant. for investors dropping into the 15% tax rate conversion is generally not beneficial unless three conditions are met: 1) the conversion is done opti 254 financial services review 6(4) 1997 table 6 after-tax future values of converted roth iras divided by after-tax future values of traditional iras when the tax rate on ordinary income increases from 15% to 28% upon withdrawal investment horizon in years (n) r 5 io 15 20 2.5 30 35 40 1.218* 1.227 1.234 1.241 1.246 1.252 1.257 1.261 6% 1.192** 1.202 1.210 1.218 1.224 1.231 1.236 1.241 1.208*** 1.208 1.208 1.208 1.208 1.208 1.208 1.208 1.228 1.238 1.247 1.254 1.260 1.266 1.272 1.277 8% 1.195 1.207 1.217 1.226 1.234 1.241 1.247 1.253 1.216 1.216 1.216 1.216 1.216 1.216 1.216 1.216 1.238 i.249 i.258 1.266 1.272 1.278 1.284 1.289 10% 1.198 1.212 1.223 1.233 1.242 1.250 1.257 1.263 1.224 1.224 1.224 1.224 1.224 1.224 1.224 1.224 1.246 1.258 1.267 1.215 1.283 1.289 1.295 1.300 12% 1.201 1.216 1.229 1.240 i.249 1.257 1.265 1.272 1.231 1.231 1.231 1.231 1.231 1.231 1.231 1.231 i.254 1.266 1.276 1.284 1.291 1.298 1.304 1.310 14% 1.203 1.220 1.234 1.245 1.255 1.264 1.212 1.280 i .237 i .237 1.237 1.231 1.237 1.237 1.237 1.237 1.261 1.274 1.284 1.292 1.299 1.306 1.312 1.318 16% 1.206 1.224 1.238 1.250 1.261 1.210 1.279 1.287 1.243 1.243 1.243 1.243 1.243 i .243 1.243 1.243 notes: *top figures indicate optimally converted iras in 1998. **middle figures indicate optimally converted iras after 1998. ***bottom figures indicate sub-optimally converted iras in 1998. mally, 2) the time horizon is long, and 3) the expected return is high. otherwise, many of the same trends are present. for example, the value of conversion increases with the time horizon and return. for short time horizons and low expected returns, the marginal value of optimal conversion is less than the marginal value of conversion in 1998. for long time horizons or high returns, it is more important to convert optimally than it is to convert sub optimally in 1998. investors with 3 1% tax rates generally do not have the option to convert since the agi limit is $100,000. according to table 6, however, investors in the 15% tax bracket who expect to be in the 28% bracket upon withdrawal should almost always convert traditional iras to roth iras. again, the same trends prevail. v. conclusion according to the panel study of income dynamics (psid) as interpreted by bemheim, skinner, and wienberg (1997), on average, retirement income is about 64% of pre-retire ment income, suggesting that marginal tax rates for retirees are likely to fall over their investment horizon. in this paper, we analyze the effect of declining withdrawal tax rates on an investor’s choice between taxable mutual fund investments and similar investments in a nondeductible ira. we assume capital gains are taxed at 20%, the new rate established by the act and find that as the withdrawal tax rate decreases, the advantage of paying the ira contributions and conversions 255 lower 20% capital gains tax early is reduced. hence, the relative attractiveness of the tax able mutual fund is reduced and the nondeductible ira becomes relatively more attractive than when the withdrawal tax rate remains constant. investors that anticipate a lower tax rate during withdrawal will favor a wider range of mutual funds in nondeductible iras rather than taxable accounts. with an expected return of approximately 15%, fund inves tors with horizons of at least ten or fifteen years should choose nondeductible iras. for investors that fall into the 15% tax bracket, the nondeductible ira will always yield a higher after-tax future value. regarding conversion of traditional iras to roth iras, we find it is optimal for inves tors to pay the conversion tax with dollars that would not qualify for tax deferment rather than using ira assets. investors who are likely to remain in the 28% tax bracket should almost always convert. also, the value of converting from a traditional ira to a roth ira increases as the investment horizon and rate of return increase. for short time horizons and low expected returns, the marginal value of conversion in 1998 is greater than the marginal value of optimal conversion. for long time horizons or high returns, it is more important to convert optimally than it is to convert sub-optimally in 1998. for investors dropping into the 15% tax bracket, conversion is generally not beneficial unless the conversion is done optimally, the time horizon is long, and the expected return is high. young investors already in the 15% tax bracket who expect tax rates to rise would almost always find con version beneficial. the analysis present here is limited in that it ignores the uncertainty associated with future tax rates, nor does it account for the premature 10% withdrawal penalty. further investigations can also focus on options that ira investors receive and give up. by using an ira account, the investor gives up the option to sell an asset to realize a capital tax loss. future research can focus on these issues. nonetheless, this paper serves as a useful guide for individual retirement planning and for making decisions under the new tax laws. references bernheim, b.d., skinner, j.s., & wienberg, s. (1997). what accounts for the variation in retirement wealth among u.s. households? unpublished mimeo, stanford university. bodie, z., & merton, r.c. (1998). finance (preliminary ed.). englewood cliffs, nj: prentice-hall, inc. crain, t.l., & austin, j.r. (1997). an analysis of the tradeoff between tax deferred earnings in iras and preferential capital gains. financial services review, 6(4), 227-242. gale, g., & scholz, j.k. (1994). iras and household saving. american economic review, 84(5), 1233-1260. hubbard, g.r., & skinner, j.s. (1995). the effectiveness of savings incentives. mimeo, columbia university, november. hubbard, g.r., & skinner, j.s. (1996). assessing the effectiveness of saving incentives. journal of economic perspectives, 10(4), 73-90. kaiser, r.w. (1990). individual investors, managing investment porrfolios: a dynamic process (2nd ed.). new york, ny: warren, gorham & lamont, inc. o’neil, c.j., saftner, d.v., & dillaway, m.p. (1983). premature withdrawals from individual retire ment accounts: a break-even analysis. journal of american taxation association 4, 35-43. randolf, w. l. (1994). the impact of mutual fund distributions on after-tax returns. financial ser vices review, 3(2), 127-141. 256 financial services review 6(4) 1997 ragsdale, ct., seila, f.s., & little, p.l. (1993). optimizing distributions from tax-deferred retire ment accounts. personal finanical planning, 5(3), 20-28. ragsdale, c.t., seila, f.s., & little, p.l. (1994). an optimization model for scheduling withdrawals from tax-deferred retirement accounts. financial services review, j(2), 93-108. saftner, d., & fink, p. (1990). timing withdrawals from retirement accounts can increase tax sav ings. taxation for accountants, 44, 172-178. sage, j.a. (1988). selection of qualified retirement plan distribution options reflection tra ‘86. taxes, (april), 301-310. yaari, u., & fabozzi, f.j. (1985). why ira and keogh plans should avoid growth stocks. journal of financial research, 8(3), 203-215. pii: s1057-0810(96)90001-3 from the editor karen eilers lahey great news! the publication of this issue marks the final semi-annual issue of the financial ser vices review (fsr). the long awaited quarterly publication of the journal starts with the next issue which will published in january, 1997. as i am sure you are aware, the move to a quarterly journal means more articles can be published in a timely fashion and more manuscript submissions are required. for volume 5, the acceptance rate is approximately 15 percent of the manuscripts that have been submitted and reviewed through august, 1996. in recognition of the hard work of the members of the editorial board and the ad hoc reviewers, i wish to formally thank each of you. the editorial board members are recog nized in each issue of the journal. the 67 ad hoc reviewers who have made a significant contribution to the selection of the final articles are recognized in this issue. if you are seeking additional information on the fsr journal, please consult our new website. the address is: http://www.uakron.edu/cba/fsr/fsr.html a request for more volunteers seems appropriate with the move to a quarterly journal. if you are willing to be an ad hoc reviewer send me a message by e-mail (klahey@uak ron.edu) or by “snail mail.” please indicate the topic areas that you feel most competent to review and we will certainly accept your offer to work as soon as we receive a manuscript that matches your expertise. doug kahl, who is in charge of the new column on book, software and web site reviews, is actively looking for individuals who are willing to review textbooks and websites. this issue of the fsr journal contains four articles on very different topic areas. the first article is entitled, “risk aversion measures: comparing attitudes and asset alloca tion” and is written by diane schooley and debra drecnick worden. they examine the very fundamental question of household’s willingness to accept risk in their selection of investments. they find that portfolio allocations are reliable indicators of attitudes about risk and that individuals understand the differences in the risk of alternative investments. their article also considers individual socioeconomic factors and goals in the selection of assets. i would like to encourage other authors to submit manuscripts that address these important issues of the individual’s understanding of risk and asset selection. the topics are important in light of the need for increased savings for retirement and the demand that individuals select appropriate investments for their 401, 403(b), ira, and keogh plans. manuscripts on pension and retirement planning would be most pertinent. v vi ftnancial services review 5(2) 1996 the second article is entitled, “a simulation approach to the choice between fixed and adjustable rate mortgages” by william k. templeton, robert s. main, and j. b. orris. this represents the first real estate article that has been published in the financial sewices review journal and i would like to encourage other authors to submit real estate papers on issues of concern to individual investors. templeton, main, and orris presented the initial draft of this paper at the annual meet ing of the academy of financial services held in new york city in october, 1995. they provide borrowers with a model on which to base their choice of mortgage types by exam ining the impact of changes in interest rates, the spread between the two types of mort gages, the teaser discount rate, points, refinancing costs, and the cost and payments for the planned life of the mortgage. jill vihtelic’s article is entitled, “personal finance: an alternative approach to teaching under~aduate finance.” she has presented her ideas on this subject at various academic meetings. an extensive review of the literature is employed to argue for a change in the tirst course to be taught in the finance curriculum from corporate finance to personal financial management. her recommendation is based on education and learning theory. i would encourage those who share vihtelic’s viewpoint or have a different vie~oint to submit either a letter to the editor with your opinions or a formal manuscript which rig orously argues for your recommendation. the declining enrollment in some colleges of business, the change in the type of students enrolled, and a different set of expectations by the american assembly of collegiate schools of business suggests the need for a lively debate on changes in the c~culum offered in ~der~aduate and graduate finance. the final article in this issue is entitled, “the effects of mutual fund managers’ char acteristics on their portfolio performance, risk and fees” and is written by joseph h. golec. he tests mutual fund managers’ characteristics in an effort to explain fund perfor mance, risk, and fees. he incorporates the relevant mutual fund academic literature in dis cussing his results and making specific recommendations for individual investors. lastly, i would like to encourage you to read timothy r. smaby’s review of the asso ciation for investment management and research (aimr) web site and thomas eysell’s reviews of the new york stock exchange, the chicago board of options exchange and the financial m~agement association websites. ho~~lly, you will be able to use the infor mation that is provided to incorporate these sites into your research and teaching activities. pii: s1057-0810(99)00030-x gender differences in defined contribution pension decisions vickie l. bajtelsmita, alexandra bernasekb, nancy a. jianakoplosb adepartment of finance and real estate, colorado state university, fort collins, co 80523, usa bdepartment of economics, colorado state university, fort collins, co 80523, usa abstract this paper considers gender differences in allocation of household wealth to defined contribution pensions. using data from the 1989survey of consumer finances, we estimate the coefficient of relative risk aversion based on the allocation of wealth into defined contribution pensions. unlike previous studies, we consider the problem in the context of the household’s overall portfolio. we find that women exhibit greater relative risk aversion in their allocation of wealth into defined contribution pension assets. © 1999 elsevier science inc. all rights reserved. jel classification: j16; g11; g23 keywords:risk aversion; pension; gender differences 1. introduction trends in employer provision of private pensions show that defined contribution pension plans are becoming increasingly prevalent (gustman and steinmeier, 1992). the united states department of labor (1997) reports that the number of workers covered by defined benefit plans decreased by 1.7 million between 1975 and 1993 whereas the number covered by defined contribution plans increased nearly 300% over that same period to 36.4 million in 1993. bajtelsmit and vanderhei (1997) summarize recent trends in pensions and conclude that an increasing percentage of new plans also require participants to make their own * corresponding author. tel.:11-970-491-0610; fax:11-970-491-7665. e-mail address:vickieba@lamar.colostate.edu (v.l. bajtelsmit) financial services review 8 (1999) 1–10 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00030-x investment decisions. although investment choice is arguably a source of empowerment, increasing evidence that individuals tend to choose conservative investment strategies for self-directed accounts raises questions of whether today’s workforce will retire with significantly lower pension benefits than those that would have resulted from investment by professional pension managers. thus this trend has potentially serious implications for the retirement income adequacy of american workers. a related problem that has received less attention is the impact of this trend on women in retirement. previous studies have found that women display greater risk aversion in a wide variety of activities, including smoking, seat belt usage, and financial decision making (hersch, 1996; jianakoplos and bernasek, 1998; bajtelsmit and bernasek, 1996). there is also evidence that women tend to exhibit greater risk aversion within their defined contribution plans (dcps) (bajtelsmit and vanderhei, 1997; hinz, mccarthy and turner, 1997; sunden and surette, 1998). women’s greater longevity implies that, even with the same investment strategy and pension investment accumulation as men, retirement wealth must support a longer period of retirement. thus, all else, equal, consumption in retirement will be lower for women. furthermore, although the pension gap between women and men has closed more rapidly than the earnings gap, women still tend to have lower coverage, lower participation rates, and lower contribution rates (magenheim, 1993). this paper differs from previous studies of gender differences in pension decisions by investigating pension allocations within the larger context of the household’s overall portfolio balance. the studies by bajtelsmit and vanderhei (1997) and hinz, et al. (1997) rely on data from a single employer/provider. many other studies of pension allocation, such as those conducted by pension plan providers (e.g., goodfellow and schieber, 1997), are flawed in that the researchers lack complete information on the household. for example, a conservative pension portfolio might be part of an overall portfolio that includes investment in a risky privately-held business or a small stock mutual fund. a similar problem arises when the researcher considers individual investment portfolios in isolation because many households may pool their resources. one spouse may have a conservative pension portfolio when the other spouse has their pension invested in risky assets such that the overall household portfolio is more diversified than it would appear by considering each in isolation. jianakoplos and bernasek (1998), investigating gender differences in financial risk taking, find that demographic and household variables are highly significant determinants of risky decision making. the paper proceeds as follows. section 2 explains the theoretical models underlying our estimation of relative risk aversion. in section 3 we explain the empirical methodology and provide the results of the estimations. finally, conclusions and policy implications are given in section 4. 2. measurement of risk aversion under expected utility theory, the dollar amount and proportion of risky assets in an investor’s portfolio are a function of wealth and individual risk preferences. the relationship between risk preferences and wealth was further developed by pratt (1964) and arrow 2 v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 (1971), who define measures of absolute risk aversion (change in dollar allocation to risky assets as wealth increases) and relative risk aversion (change in portfolio allocation to risky assets as wealth increases). based on empirical and experimental studies of individual decision-making under risk, there is now general consensus thatabsoluterisk aversion declines with wealth. that is, individuals will invest a higher dollar amount in risky assets as their wealth increases. the characteristics of individualrelative risk aversion are not as clear and seem to exhibit systematic differences by some characteristics such as age and income. the impact of gender on relative risk aversion was considered in jianakoplos and bernasek (1998) who find evidence that single women exhibit greater relative risk aversion than single men do. the empirical estimation of risk aversion in pension allocations undertaken in this paper follows the methodology developed by friend and blume (1975). this methodology for measuring relative risk aversion is also employed by siegal and hoban (1982,1991), morin and suarez (1983), ballante and saba (1986), riley and chow (1992), jianakoplos and bernasek (1998), and schooley and worden (1996). they define an individual’s division between risky and riskfree assets in their portfolio, in the absence of taxes, according to the following: ak 5 e~rm 2 r f! sm 2 p 1 ck (1) whereak is the proportion of investork’s net worth that is placed in risky assets,e(rm 2 rf) is the expected difference between the return on the market portfolio of risky assets (rm) and the return on the risk free asset (rm), sm 2 the variance of the returns on the market portfolio of risky assets,ck is pratt’s measure of relative risk aversion(ck 5 [2u“(wkt)/u’(wkt)]wkt)” andwkt is investork’s wealth in period t. we ignore the effects of taxes in the estimation of the model based on previous research which shows that taxes do not have a significantly affect on the results (bellante and saba, 1986). the model assumes that financial assets are infinitely divisible and can be traded with zero transactions costs. these assumptions are problematic when applied to housing wealth and human capital. for that reason, some studies estimate the value of wealth with and without the value of housing (friend and blume, 1975). to include human capital in net wealth, a reformulation of eq. (1) is required takes into account the dependence ofak on the covariance between the return on the market portfolio (rm) and the return on human wealth (rh). this results in the following specification for allocation to risky assets: ak 5 e~rm 2 r f! sm 2 p 1 ck~1 2 hk! 2 hk ~1 2 hk! p bh,m (2) wherehk is the ratio of investor k’s human wealth to net wealth, andbh,m is the ratio of the covariance ofrm andrh to sm 2 . this equation can be simplified however, by making use of the findings by liberman (1980) and fama and schwert (1977) thatbh,m is zero. then eq. (2) becomes: 3v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 ak 5 e~rm 2 r f! sm 2 p 1 ck~1 2 hk! (3) for the analysis in this paper, eq. (3) is rewritten to focus on the proportion of riskypension assets in an individual’s portfolio as follows: ak1 5 2ak2 1 e~rm 2 r f! sm 2 p 1 ck~1 2 hk! (4) whereak1 is the proportion of net worth of investor k in riskypensionassets andak2 is the proportion of net worth of investor k inother risky assets. eq. (4) forms the basis for estimating the coefficient of relative risk aversion when we consider an individual’s investment in risky pension assets. 3. empirical model and estimation 3.1. data this study employs data from the 1989survey of consumer finances(scf89) to examine whether defined contribution pension allocation decisions differ by gender. the scf89 includes interview data collected from a sample of 2, 277 households, chosen to be representative of households in the contiguous 48 states of the united states. because one of the principal goals of the survey was to obtain estimates of household wealth, which is a highly skewed variable, an additional oversampling of 866 wealthy households was undertaken, selected from tax records to represent wealthier households. the combined sample of 3,143 households reported on the composition of their balance sheets, employment status, income, pensions, and other economic and demographic information. kennickell and shack–marquez (1992) provide more details on the survey methods. according to several studies (curtin, et al., 1989; juster and kuester, 1991; and starr–mccluer, 1996) thesurveys of consumer financesare the best available source of individual household wealth data collected in the united states. the scf89 respondents reported balances in defined contribution pensions (dcps) for both themselves and spouse/partners, if any. because the focus of this study is on pension balances of individuals, rather than of households, data for respondents and spouses are included separately in the sample. consequently, pension allocation behavior is examined for a sample of 5, 287 individuals—the 3, 143 respondents and an additional 2, 144 spouse/ partners. in the cases where there are two individuals from the same household, they have the same household wealth, because there is no basis for dividing their joint assets. because of the oversampling of the wealthy households, all summary statistics reported in this paper are sample weighted. in addition, the multiple imputation procedure employed on the public-use data tapes by the federal reserve system to handle missing data is used here. 4 v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 3.2. empirical methodology based on eq. (4), that is the theoretical basis for the allocation of an individual’s portfolio into risky pension assets, the estimating equation takes the following specification: alpha5 b1 1 b2ln wealth1 b3age1 b4age2 1 b5age-s 1 b6age2-s1 b7educ-12 1 b8black 1 b9kids 1 b10single1 b11humanj 1 b12human-s 1 b13risky1 b14homeowner1 b15otherpens 1 b16otherpens-s1 b17lambda1 m, (5) where alpha is the ratio of individual holdings of dollar balances in dcps to total householdwealth (wealth) that includes both riskfree and risky assets. riskfree assets include dollar balances in checking, savings, and money market accounts, certificates of deposit, u.s. savings bonds, ira balances invested in certificates of deposit or bank accounts, and the cash value of life insurance less policy loans outstanding. risky assets are the sum of: balances in iras not invested in bank deposits, stock holdings less margin loans outstanding, bonds, trust assets, the net value of real estate other than residential housing, the net value of businesses owned, the net value of other miscellaneous assets (e.g., precious metal, futures contracts, art work, etc.) reported by the household, and balances in dcps. to control for demographic differences, the specification includes:age andage2, the individual’s age in years and it’s square;age-sandage2-s, age and age squared of the spouse, if any;educ-12, a dummy variable that equals one if the individual has completed more than 12 years of schooling;black, a dummy variable that equals one if the individual’s race is reported as black;kids, the number of children living in the household; and single, a dummy variable equal to one if the individual’s marital status is not reported to be married or living with a partner. a number of variables are included to capture other aspects of the individual’s portfolio. riskyis the ratio of other risky assets, excluding assets held in dcps, towealth(the empirical estimate ofak) in eq. (4). other risky assets are all household risky assets except those included inalphaij . human andhuman-sare the ratios of the individual’s and spouse’s (if any) human capital to wealth, respectively. human capital for each individual is calculated as the present value of the stream of future earnings, assuming that current wages, salaries, and/or self-employment earnings grow at a constant rate until retirement. retirement is assumed at age 65 for those 65 or younger. if the individual is still working and between the ages 65 and 69, current earnings are assumed to continue for four more years; if between 70 and 74, to continue for three more years; if between 75 and 79, to continue for two more years; and if over 79, for one more year, consistent with friend and blume (1975). the discount rate is assumed to be 2%, which approximates the long run growth in real gdp. although this measure of human capital is obviously an approximation, analysis by thornton, rodgers, and brookshire (1997) suggests that the assumption of a constant growth rate 5v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 of earnings, rather than the traditional inverted u-shaped pattern, is supported by longitudinal data. given the ambiguity between investment and consumption aspects of residential housing, rather than including it as a risky asset, we have included a dummy variable,homeowner that equals one if the respondent owns a house.otherpensand otherpens-sare dummy variables that equal one if the individual or spouse, respectively, reports entitlement to defined benefit pensions. following previous research, the sample is limited to those individuals withwealthj greater than $1, 000. furthermore, the nature of the question under consideration requires that the sample be limited to those individuals with defined contribution pensions. if sample selection were random, ols would be an appropriate estimation technique. however, if the sample selection is not random, ols yields inconsistent results (greene, 1993). to correct for possible sample selection bias, heckman’s (1979) two-step procedure is employed. a probit regression of the factors that influence inclusion in the sample is estimated over all observations jointly with eq. (5). the specification of the probit equation is: hasdc5 g1 1 g2age1 g3union 1 g4bigfirm 1 g5prof1 g6sales 1 g7crafts1 g8labor1 g9farm1 y, (6) wherehasdc is a dummy variable that equals one if the individual has a defined contribution pension and haswealthgreater than $1, 000. the explanatory variables include: age, union, a dummy variable that equals one if the individual belongs to a union, bigfirm, a dummy variable that equals one if the individual works for a firm employing over 500 people, and a set of five occupational dummies (prof, sales, crafts, labor, andfarm), where the excluded occupational category is service jobs. an estimate of the inverse mill’s ratio is calculated from this equation and included in eq. (5) as the variable lambda. 3.3. results because the sample used for analysis in this paper is comprised of only those survey participants who report having dcps and wealth greater than $1, 000, table 1 provides a comparison of variable means for the dcp subset and the rest of the scf89 sample. individuals in this subset are, on average, more likely to be younger, to be union members, to be employed by a large firm, to have higher education, and to have more children. comparison of male and female variable means shows differences that would be expected. on average, women with dcps allocate a smaller proportion of wealth to their pensions (27%) compared to men (35%). the results of the first stage probit estimation, that are available from the authors upon request, indicate that the survey subset is significantly different from the overall scf89 survey sample. the inclusion of thelambdavariable in the estimation procedure therefore serves to ensure that the other coefficients are consistent. the results of the full information maximum likelihood estimations of eq. (2) are shown separately for men and women in table 2. to test whether the factors determining the proportion of wealth allocated to dcps differ by gender, all of the variables in eq. (2) were 6 v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 interacted with a dummy variablefemaleand the equation was re-estimated for the sample including both men and women. a likelihood ratio test of the null hypothesis that all the female interaction terms equal zero is rejected, indicating that the factors determining women’s allocation of wealth to dcps are different from men’s. the test statistic was 220.66 distributedx2 with 16 degrees of freedom. the critical value at the 1% level of significance is 31.99. thus, the null hypothesis is rejected. coefficients that differ significantly between the men’s and women’s equations are indicated in table 2. significant gender differences exist for age (both respondent’s and spouse’s), education, number of children, marital status, human capital (both respondent’s and spouse’s), allocation to other risky assets, and having claims on other pensions. the coefficient oflnwealthis an estimate of relative risk aversion (ck) up to a positive multiplicative constant. a positive (negative) coefficient indicates decreasing (increasing) relative risk aversion. for men, the coefficient is significantly positive, indicating decreasing relative risk aversion. the estimated coefficient for women is negative, but not significantly different from zero, indicating constant relative risk aversion. thus women exhibit greater relative risk aversion than men in the allocation of wealth to dcps, holding other factors constant. to illustrate the impact on individual portfolios, consider the following application of these estimation results. if initial wealth for both men and women is assumed to be $116, 000, the mean wealth of the women in the sample, and all other variables are held table 1 variables means defined contribution pension no defined contribution pension women men women men alpha 0.27 0.35 — — wealth 116, 044 212, 933 134, 763 148, 497 age 40.5 41.6 48.4 37.4 age-sa 41.1 39.8 48.5 46.5 educ-12 0.58 0.68 0.35 0.41 black 0.10 0.06 0.13 0.08 kids 0.99 1.21 0.95 0.92 single 0.35 0.15 0.33 0.20 human 32.33 45.8 194.76 81.51 human-sa 32.76 12.30 192.90 99.36 risky 0.36 0.34 0.41 0.42 homeowner 0.74 0.78 0.62 0.63 other pension 0.41 0.42 0.14 0.25 other pension-sa 0.43 0.20 0.27 0.17 union 0.26 0.25 0.07 0.16 big firm 0.66 0.60 0.16 0.23 professional 0.39 0.46 0.13 0.18 sales 0.49 0.18 0.20 0.13 crafts 0.01 0.19 0.01 0.15 labor 0.06 0.11 0.05 0.12 farm — 0.01 0.01 0.04 n 319 573 2, 489 1, 906 a means are calculates only for individuals with a spouse present. 7v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 constant at the sample means for men and women separately, the proportion of wealth held in dcps is predicted to be 27.1 and 35.5% for women and men, respectively. if wealth is now increased to $213, 000, the mean wealth of men in the sample, the proportion of wealth held in dcps is predicted to decrease slightly to 26.3% for women and to increase only slightly to 35.9% for men. the gender difference in relative risk aversion explains only part of the difference in allocation of wealth to dcps. for example, an increase in number of children is estimated to increase allocations to dcps for both men and women, but by a small proportion for women. holding everything else constant, single men are estimated to allocate more to dcps compared to married men, but single women are estimated to allocate significantly less than married women. when either the individual or spouse has access to pensions other than dcps, men are estimated to increase relative allocations to dcps, but women are estimated to reduce their dcp allocations. one way to assess the relative importance of the factors that contribute to the lower allocation of wealth to dcps by women is to consider the following example. if the sample mean characteristics for women are applied to the coefficients of the men’s equation, the allocation to dcps is predicted to be 33.8% of wealth compared to the sample average of 27%. this would bring the female allocation to dcps very close to the mean for men which table 2 dependent variable: proportion of wealth held in defined-contribution pensions women men independent variables coef se coef se ln wealth 20.004††† 0.004 0.007*** 0.002 age 0.021***†† 0.006 0.005 0.003 age2 20.000*** 0.000 20.000 0.000 age-s 20.027***††† 0.006 0.006** 0.002 age2-s 0.000***††† 0.000 20.000** 0.000 educ-12 20.015††† 0.011 20.054*** 0.010 black 20.004 0.018 0.002 0.018 kids 0.007† 0.005 0.017*** 0.003 single 20.634***††† 0.146 0.168*** 0.050 human 20.000††† 0.000 0.000*** 0.000 human-s 20.000 0.000 20.000 0.000 risky 20.376***††† 0.018 20.561*** 0.015 homeowner 20.027** 0.014 20.013 0.011 other pension 20.014††† 0.011 0.024*** 0.008 other pension-s 20.025**†† 0.013 0.011 0.010 constant 0.650***††† 0.106 0.161** 0.072 lambda 20.028** 0.013 0.022** 0.010 n:probituregression 2, 808 319 2, 479 573 log likelihood1 23140 24707 * ,** ,*** significantly different from zero at the 10%, 5%, and 1% levels, respectively. †,††,†††significantly different from male coefficient at the 10%, 5%, and 1% level, respectively. 1 the likelihood ratio statistic for the women’s equation was 2, 929.96 distributed x2 with 23 degrees of freedom. the equation is significant at the 1% level. the likelihood ratio statistic for the men’s equation was 2, 469.51 distributed x2 with 24 degrees of freedom. the equation is significant at the 1% level. 8 v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 is 35%. on the other hand, if women were assumed to have the same characteristics as the sample average for men, and these mean values were applied to the estimated coefficients of the women’s equation, the predicted proportion of wealth in dcps would increase to only 27.9%. thus, the gender differences in the allocation of wealth to dcps are largely attributable to differences in male versus female behavior. this is a reflection of the differences in the estimated regression coefficients, rather than differences in the characteristics of women versus men as summarized in the variable means. 4. conclusions and policy implications although several previous studies have examined the issue of gender differences in investing, this study specifically considers the allocation of wealth to pensions and improves on earlier studies by including important socio-economic and demographic explanatory variables. the results of this analysis demonstrate that there are significant gender differences in allocation of wealth into defined contribution pensions. this study improves on earlier research by analyzing pension decisions within the broader context of the household portfolio. this conclusion has important implications for public policy, particularly in light of recent demographic and legislative trends. although pension coverage rates for women have improved substantially in the last two decades as the number of women in the workforce has increased, this study indicates that women allocate a smaller proportion of their total wealth to these retirement vehicles. at the same time, social security replacement ratios are lower than they have been in the past and most new pensions require self-direction of pension account allocations. given evidence that women tend to be very risk averse with respect to the pension allocation decision, it is likely that women will retire with significantly lower pension wealth than their male counterparts. furthermore, this smaller wealth will have to be spread over a longer retirement due to greater average longevity. references arrow, k. j. (1971).essays in the theory of risk bearing. chicago: markham. bajtelsmit, v. l., & bernasek a. (1996).why do women invest differently than men? finan counsel plan, 7, 1–10. bajtelsmit, v. l., & vanderhei, j. a. (1997). risk aversion and retirement income adequacy. in. mitchell, o. s. (ed.) positioning pensions for the year 2 000(pp. 45–66). philadelphia: university of pennsylvania press. bellante, d. & saba r. p. (1986). human capital and life-cycle effects on risk aversion.j finan res, 9(spring), 41–51. curtin, r., thomas, f., & morgan, j. (1989). survey of estimates of wealth: an assessment of quality. in lipsey, r. & tice, h. s. (eds.).the measurement of saving, investment, and wealth(pp. 473–548). chicago: university of chicago press. fama, e. f. & schwert, g. w. (1977). human capital and capital market equilibrium.j finan econ, 4, 95–125. friend, i. & blume, m. e. (1975). the demand for risky assets.am econ rev, 65(december), 900–922. gilbert, r. f. (1994). estimates of earnings growth rates based on earnings profiles.j legal econ, (summer), 1–17. 9v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 goodfellow, g. p., & schieber, s. j. (1997). investment of assets in self-directed retirement plans. in mitchell, o. (ed.)positioning pensions for the year 2000(pp. 67–90). philadelphia: university of pennsylvania press. gustman, a. l. & steinmeier, t. (1992). the stampede toward defined contribution plans: fact or fiction?indus relations, 31(2), 361–369. greene, w. h. (1993).econometric analysis(2nd. ed). new york: macmillan publishing company. heckman, j. (1979). sample selection bias as a specification error.econometrica 47(january), 153–161. hersch, j. (1996). smoking, seat belts, and other risky consumer decisions: differences by gender and race. managerial decision economics, 17(5), 471–481. hinz, r. p., mccarthy, d. d., & turner, j. a. (1997). are women conservative investors? gender differences in participant-directed pension investments. in mitchell, o. (ed.)positioning pensions for the year 2000(pp. 91–103). philadelphia: university of pennsylvania press. jianakoplos, n., & bernasek, a. (1998). are women more risk averse?economic inquiry, 36(4), 620–630. juster, f. t., & kuester, k a. (1991). differences in the measurement of wealth, wealth inequality, and wealth composition obtained from alternative u.s. wealth surveys.rev income wealth, 37(march),33–62. kennickell, a., & shack-marquez, j. (1992). changes in family finances from 1983 through 1989: evidence from the survey of consumer finances.fed reserve bull, (january), 1–18. liberman, j. (1980). human capital and financial capital market.j bus, 53, 165–191. magenheim, e. b. (1993). gender related patterns in pensions: a review of the literature, “u.s. department of labor paper no. b9324129. washington dc: u.s. govt. printing office. morin, r. a. & suarez, f. (1983). risk aversion revisited.j finan, 37(september), 1201–1216. pratt, j. (1964). risk aversion in the small and the large.econometrica, 32(january/april), 122–136. riley, w. b., & chow, k. v. (1992). asset allocation and individual risk aversion.finan anal j, 48(nov./dec.), 32–37. schooley, d. k., & worden, d. d. (1996). risk aversion measures: comparing attitudes and asset allocation. finan serv rev, 5(2), 87–100. siegel, f. w., & hoban, j. p. (1982). relative risk aversion revisited.rev econ stats 64(august), 481–487. siegel, f. w., & hoban, j. p. (1991). measuring risk aversion: allocation, leverage, and accumulation.j finan res. 14(spring), 27–35. starr–mccluer, m. (1996). health insurance and precautionary savings.am econ rev 86(1), 285–295. sunden, a. e., & surette, b. j. (1998). gender differences in the allocation of assets in retirement savings plans. am econ rev, 88(2), 207–211. survey of consumer finances. (1989). federal reserve. thornton, r. j., rodgers, j. d., & brookshire, m. l. (1997). on interpretation of age-earnings profiles.j labor res, 18(2), 351–365. tiaa-cref. (1994). replacement ratio projections in defined contribution retirement plans: time, salary growth, investment return, and real income.res dialog 41(september), 1–6. u.s. department of labor. (1997).report on the american work force.washington, dc: u.s. government printing office. 10 v.l. bajtelsmit et al. / financial services review 8 (1999) 1–10 pii: s1057-0810(96)90023-2 financial services review, 5(l): 1-12 copyright 0 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. equity fund size and growth: implications for performance and selection conrad s. ciccotello c. terry grant should individuals choose the largest or smallest equity jimis for investment? this study explores the relationship of equity find size to pe$ormance. historical returns of largejiouis are found to be superior to their smaller peers. yesterday’s bestpet$orming funds tend to become to&y’s largestjkds as individuals invest heavily in response to the communications about the fund’s past success. but the findings suggest that, once large, equity jimak do not outperform their peers. especially for jima!s in aggressive growth objectives, the advantages of being small appear to outweigh the disadvantages. for individual investors wtih aggressive growth objectives, a strategy of investing in smallerjiouis may thus be wealth maximizing. i. introduction equity fund size and growth are becoming increasingly critical considerations in the choice of funds. this paper documents that successful funds grow dramatically as investors react to favorable past performance. but as a fund grows, can it continue to outperform its peers? the findings in this paper suggest to individual investors that smaller funds, especially those within the more aggressive investment objectives, tend to outperform larger funds. advice on choosing funds is more relevant than ever for individuals who rely heavily on mutual funds to accumulate wealth. both the mutual fund industry and the number of funds available are growing rapidly. the popularity of fund investing stems from the ben efits of diversification. in addition, professional mutual fund management can generally trade stocks at favorable prices. ippolito (1989, 1993, p. 42) concludes that mutual funds are “sufficiently successful in finding and implementing new information to offset their expenses.” to argue that the mutual fund industry is informationally efficient in the aggregate, however, ignores individual fund characteristics such as asset size. funds can be enor conrad s. ciccotello l 2354 fairchild wve suite 6h-94, department of management, united states air force academy, co 80840-5701. c. terry grant l school of professional accountancy, university of southern mississippi, 730 east beach blvd., long beach, ms 39560. 2 financial services review 5( 1) 1996 mous; as of early 1996, fidelity magellan fund has over $40b in assets. at the same time, many other open-end funds have assets less than $lom. do large funds have systematic advantages over small funds that permit superior performance? or, should individual investors rely on smaller funds for superior returns? to examine this issue, this paper stud ies the relationship of mutual fund performance and asset size. in particular, this research extends the work done in previous studies by focusing on the implications of fund size for individuals faced with selecting funds for investment. the issue of asset size and fund performance has been examined by other researchers. grinblatt and titman (1989) study the performance of mutual fund portfolios over the period from 1975 to 1984. mutual funds are ranked by asset size and then divided into quintiles. using 1975 asset size, the authors find some evidence of abnormal performance in gross returns over a io-year period (19751984) in the smaller asset-size quintiles, espe cially in the more aggressive fund objectives. net of expenses, however, the returns of funds in smaller quintiles are not different from the returns of funds in larger quintiles. droms and walker (1994) find no relationship between fund size and performance in their study of international mutual funds. droms and walker (1994) use fund size as one of a number of explanatory variables for fund performance. they find that the coefficient for fund size is generally insignificantly different from zero for both unadjusted and risk adjusted returns. from these studies, the general inference is that asset size is a poor predictor of a fund’s future performance. but does this conclusion mean that a fund’s asset size is mean ingless? a fund’s current size may be related to its past performance. if investors rely on funds’ past performance to infer how the fund will do in the future, then mutual fund his torical performance will be highly relevant to investors’ fund selections. funds that have performed well, especially in the recent past, may thus attract significant investor attention. hendricks, patel, and zeckhauser (1993) document that mutual fund managers can have “hot hands” that allow them to outperform their peers in the short run. grinblatt and titman (1992) find that mutual funds can enjoy positive abnormal performance for periods up to five years of time. patel, zeckhauser, and hendricks (1992) find that investors pour money into funds that have performed well recently, based on favorable press coverage and the positive advertisements made by the fund to the investing public. based on evidence from surveys of individual investors who had recently made mutual fund purchases, goet zmann, greenwald, and huberman (1992) and capon, fitzsimons, and prince (1992) observe that recent investment performance is a critical input in fund selection. the result is that successful funds can change greatly in size over a relatively short period of time. one example is twentieth century’s ultra fund. according to morningstar, this aggres sive growth stock fund grew in assets from $458.3m in 1990 to $9.85b in 1994. once attracted by a fund’s recent success, new individual investors may benefit or be hurt by a fund’s large size. large funds have several structural and institutional advantages over small funds. for example, big funds can spread fixed overhead expenses, such as rent or salaries of administrative personnel, over a larger asset base. this economy of scale advantage could lead to large funds outperforming small funds after adjustment for risk. this is because additional expenses for overhead do not contribute directly to obtaining information that allows managers to execute trades at favorable prices (see ciccotello & grant, 1996). influential managers of big funds can obtain positions in lucrative investment oppor tunities not available to other market participants. for example, smith (1994) reports that equily fund size and growth 3 managers at fidelity investments routinely are allotted shares in oversubscribed initial public offerings. glosten and harris (1988) suggest that managers of larger funds may be able to execute trades at more favorable spreads, given their powerful market position and large volume trading. together, these institutional and cost advantages should lead to large funds outperforming small funds. but big funds also present management challenges. continuing to find worthwhile investment opportunities as the fund grows may strain the capabilities of even a top man ager or management team. phalon (1994) outlines the circumstances associated with the closing of the highly-successful sogen international fund. how much strain may depend upon the objective of the fund being managed. if more aggressive funds invest in smaller firms, then rapid asset growth may be more difficult to manage. large asset size also reduces a fund’s “nimbleness.” the manager becomes less able to quickly move in or out of positions without attracting a great deal of attention. managing the fund becomes like “maneuvering a battleship in a bathtub.” beating the market indexes becomes difficult as the fund itself grows to become a market proxy. these challenges support the assertion that funds can grow too big, and that optimal performance occurs in smaller funds. this paper contributes to the existing research by offering individual investors insight regarding the relationship of fund performance and fund size. the findings suggest that, smaller funds, especially those within the more aggressive objectives, offer the best poten tial for superior returns. to illustrate these findings, the paper proceeds as follows: section ii describes the data and hypotheses. section iii presents the results and discussion. section iv concludes and summarizes. ii. data and hypotheses this study uses equity mutual fund data taken from several sources. they include alex ander steele’s mutual fund database, wiesenberger, and momingstar. the paper’s initial sample is contained in steele and consists of 182 aggressive growth (ag) funds, 248 long-term growth (ltg) funds, and 196 growth and income (gi) funds.’ mutual fund annual returns and descriptive statistics are evaluated over the period from 1982 through 1992. descriptive data for the equity funds is summarized in table 1. for this study, funds are classified by investment objective. the average 3-year beta* for the funds in each objec tive is shown to illustrate that as the fund risk objective moves from growth and income through long-term growth to aggressive growth, beta increases monotonically, as expected. table 1 also displays the differences in average fund expenses, turnover, and size among these three investment objectives. many of the significant differences observed in table 1 are not surprising. larger funds tend to have lower expenses, lower turnover ratios, and be older, on average. using the sample from the steele database, this paper first tests whether large funds have greater historical returns than their smaller counterparts. the hypothesis is as fol lows: hl: loge funds should have greater historical returns than small funds. historical returns are relevant to individual investors as they are the basis for adver tisement of performance.3 to test hl, funds are ranked within investment objectives on the 4 financial services review 5( 1) 1996 table 1 descriptive comparison of fund variables (whole sample) mean sd sample size aggressive growth assets beta expenses turnover years incorporated long-term growth assets beta expenses turnover years incorporated growth & income assets beta expenses turnover years incorporated 325.96 622.32 182 1.12 .28 182 1.43 .53 181 118.oii 216.98 167 11.75 10.21 180 544.68 1567.12 248 1.04 .22 248 1.28 .57 248 78.30 80.24 221 14.60 14.56 244 676.40 1674.40 196 .91 .17 196 1.09 .53 195 59.25 95.63 175 18.90 20.03 194 notes: this table contains mean values and standard deviations of fund variables. included are tests for significant differences in the variables. each of these individual parameters for each objective is significantly different from its counterparts at the 5% level. the only exception is the difference between the average asset size of the long-term growth and growth & income objectives. basis of 1992 asset size. each set of ranked funds is then divided into quartiles and both 5 year (1987-1992) and lo-year historical (1982-1992) quartile returns within each invest ment objective are examined for significant differences.4 large funds have become large for two reasons: (a) the value of the stocks in the fund has increased; and/or (b) the additions to the fund by investors have outstripped redemptions. the reasons for growth are comple mentary. net additions to the fund have probably occurred because of the fund’s superior historical performance. to examine the patterns of fund growth and past performance, the asset growth rates of the best and worst performers are studied for the 1982-1992 period. as patel, zeckhauser, and hendricks (1992) have observed, funds that have performed well attract large amounts of new capital from investors. but new investors entering the fund do not earn the fund’s historical returns. these investors care about what will happen after they invest. to examine this issue, the paper tests whether asset size can predict future returns. the hypothesis is as follows: h2: small funds should have greater future returns than large funds. to test h2, funds are ranked on the basis of historical (1982) asset size and divided into quartiles. both 5-year (1982-1987) and io-year (1982-1992) future returns for each quartile are then examined for significant differences. funds are also ranked on 1987 asset size and 5-year (1987-1992) returns are evaluated. these tests are similar to those con ducted by grinblatt and titman (1989). if small fund advantages such as nimbleness and ready investment opportunities dominate, then the smaller quartiles should outperform the larger. such a finding would support h2. on the other hand, if economies of scale and other equity fund size and growth 5 institutional advantages dominate, then the future performance of larger funds should dom inate that of smaller funds within an investment objective. iii. results and discussion a. historical returns table 2 separates funds within each risk objective into the smallest and largest quar tiles based on 1992 asset size. all funds with 5 (10) years of performance data are included in the 5-year (lo-year) results. the large-fund quartiles have superior historical returns, as predicted by hl. significant differences are found maimy in tbe io-year historical returns. ten-year superior returns make attractive advertisement testimonials for funds seeking additional capital. these findings are consistent with the argument that successful funds tend to grow rapidly in size while poorly performing funds do not attract additional cash from investors at as rapid a rate. to further examine the relationship between fund growth and historical performance, the best performing and worst performing funds in each investment objective over the past 10 years are identified. this process starts with finding the 15 best lo-year historical per formers and the corresponding 15 worst performers in each investment category. these groups performance and asset growth are compared in table 3. table 3a shows the top and bottom performers’ mean growth in each investment objective. the fund growth in the aggressive growth objective is a typical pattern. in that category, tire mean percentage growth in assets for the top performers over the decade is 2622.6%. the corresponding mean lo-year asset growth for the bottom performers is 527.4%. a fund’s asset value can table 2 comparison of mean performance values-ranked on 1992 asset size quartile 1 (smallest-size) vs. quartile 4 (largest-size) quartile i quartile 4 significance aggresdve growth 5-year return % n= 134 lo-year return 46 n=62 long-term growth s-year return % n= 166 io-year return % n=97 growth & income 5-year return % n= 137 io-year return % n=82 113.4 131.3 .070 216.0 297.1 .051 104.7 117.1 .086 229.4 325.1 .ooo 94.0 99.2 .233 243.0 301.9 ml1 norex this table camps the mean total returns of the quartiles of futtds with the smallest current asset size to the mean total returns of the qwtiles of foods with the largest cumttt asset size. all retums are in percentage terms; assets are in $m. nis the number of funds with return data for each of the performance periods. p-vahtes arc reported for f-tests of equality between qwtile returns. all performance periods end on december 31,1992. financial services review 5( 1) 1996 table 3a asset growth in the lo-year best and worst performers top 15 botrom 15 performers performers aggressive growth io-year return 1982 assets ($m) 1992 assets ($m) percent growth (raw) percent growth (new) long-term growth io-year return 1982 assets ($m) 1992 assets ($m) percent growth (raw) percent growth (new) growth 8~ income lo-year return 1982 assets ($m) 1992 assets (%m) percent growth (raw) percent growth (new) 372.5 65.0 1094.3 2622.6 2308.1 412.1 146.2 3124.8 2645.2 2233.0 350.6 299.3 3190.3 2378.5 2027.9 170.9 .ooo 181.6 ,157 444.8 ,149 527.4 ,033 356.5 .05 1 164.8 .ooo 116.9 ,524 401.9 ,075 863.5 ,085 698.7 ,141 201.2 .ooo 103.2 ,201 220.4 ,025 253.3 .009 52.0 ,013 p-value notes: table 3a compares the mean asset size and growth for the best performing funds over the period 1982-1992 with the asset growth of the worst performing funds over that period. fifteen funds are chosen from each investment objective. percentage growth (raw) considers the total growth in fund size over the decade. percentage growth (new) removes the fund’s io-year performance to examine the change in size based on new additions. p-values are for significant differences in i-tests. table 3b lists the actual funds from each objective. grow from either the appreciation of assets in the fund or net cash inflows from investors. the extreme differences in these growth rates cannot be explained by differences in invest ment performance. after removing the effects of lo-year fund performance from the sam ples, the differences in growth rates are still highly statistically significant, as a comparison of the percent growth (new) figures demonstrates. these findings are consistent with the research of patek zeckhauser, and hendricks (1992), who find that successful funds receive large amounts of new investment capital. they are also in line with the results from surveys of individual investors done by capon, fitzsimons, and prince (1992) and goetzmann, greenwald, and huberman (1992) showing that recent fund performance is a key element in the fund selection decision. b. future returns using current asset size and historical returns does not address the issue of whether asset size can be used by individual investors to predict future returns. to examine this issue, all funds with 5 (10) years of performance as of the end of 1987 (1992) are ranked on the basis of end-of-year 1982 asset size. funds are also ranked on 1987 asset size and 5 year (1987-1992) performance is shown. all 5and lo-year future returns quartiles are contained in table 4. table 4 illustrates that investing in smaller asset size funds does not lead to superior future returns in the growth & income objective. in this objective, lo-year returns are equity fund size and growth 7 table 3b asset growth in the lo-year best and worst performers top 15 performers: bottom i5 pegormers: aggressive growth long-term growth growth & income putnam otc emerging growth aim constellation acorn twentieth century ultra sit new beginning growth putnam voyager /a neuherger manhattan keystone america-omega special portfolios-stocks twentieth century growth fortis growth fpa capital quest for value fund inc saiomon brothers opportunities delaware trend cgm capital development fidelity magellan fidelity destiny plan 1 new york venture aim weingarten fidelity contrafund ids ,new dimensions guardian park avenue sequoia steinroe special phoenix growth berger one hundred janus eltim trusts fortis fiduciary mutual shares mutual qualified merrill lynch phoenix-a selected american shares windsor fpa paramount washington mutual fundamental investors dodge & cox stock mutual benefit investment company of america lexington corporate leaders vanguard index trust-500 john hancock sovereign federated stock trust lord abbett developing growth security ultra vanguard explorer steinroe capital opportunities usaa aggressive growth price new horizons ids progressive oppenheimer target prudential gwth opportunities (b) dfa us 9-10 small company ids discovery fund keystone s-4 value line leveraged growth investors research scudder development american investors growth value line special situations merrill lynch special value-a oppenheimer american growth mfs capital development american national growth usaa growth eaton vance special equity security action msb a-c pace (a) safeco growth keystone s-3 united vanguard security investment national industries philadelphia gateway index plus keystone s-l tne growth opportunities value line provident mutual transamerica growth/income (a) financial industrial trustees commingled usa a-c growth & income a-c comstoclda united retirement shares winthrop focus growth/income n&?: see table 3a notes. nearly the same across all size quartiles. so far, these results are consistent with those of grinblatt and than (1989). these results are not consistent with h2, that small funds would outperform large funds within a given risk objective in future periods. 8 financial services revlew 5( 1) 1996 table 4 comparison of mean performance values ranked on 1982/1987 asset size quartile 1 (smallest-size) vs. quartile 4 (largest-size) significance of return quartile i 2 3 4 comparisons p-value 1982-1987 s-year return % 1982 assets ($m) n 1987-1992 .5-year return % 1987 assets ($m) n 1982-1992 io-year return % 1982 assets ($m) n 1982-1987 5-year return % 1982 assets ($m) n 1987-1992 5-year return % 1987 assets ($m) n 1982-1992 io-year return % 1982 assets ($m) n 1982-1987 5-year return % 1982 assets ($m) n 89.8 76.4 64.9 54.0 10.2 29.3 62.7 240.5 (13) (13) (14) (14) 151.5 135.2 96.5 113.1 11.8 45.7 114.0 450.8 (29) (29) (30) (30) 298.1 260.6 283.8 222.3 10.2 29.3 62.7 240.5 (13) (13) (14) (14) 82.5 79.3 86.5 82.4 16.3 55.2 140.1 477.6 (24) (24) (24) (25) 131.2 108.9 102.2 104.3 18.4 79.0 204.8 944.6 (37) (37) (37) (38) 277.1 264.8 278.4 282.9 16.3 55.2 140.1 477.6 (24) (24) (24) (25) 99.7 96.0 89.3 98.6 11.5 51.5 157.5 777.0 (20) (20) (20) (21) ql vs 42 (.388) qlvsq3* (.081) ql vs q4 ** (.015) 42 vs q3 (.400) 42 vs q4 (.108) 43-44 (.370) ql vs 42 (.227) ql vs q3*** (.ooo) ql vs q4*** (.004) 42 vs q3*** (.ooo) 42 vs q4** (.038) q3 vs q4* (.060) ql vs 42 (.199) ql vs q3 (.690) ql vs q4** (.025) q2 vs 43 (.462) q2vsq4 (.169) q3 vs q4* (.080) ql vs 42 (.682) ql vs q3 (.631) ql vs 44 (.990) 42 vs q3 (.434) 42 vs q4 (.718) ql vs q2** (.012) ql vs q3*** (-000) ql vs q4*** (.ooo) q2 vs q3 (.402) q2vsq4 (.522) q3 vs q4 (.725) ql vs 42 (.537) ql vs 43 (.958) ql vs 44 (.773) q2 vs 43 (.615) 42~~44 (448) q3 vs q4 (.866) ql vs 42 (.654) ql vsq3 (.225) ql vs 44 (.890) q2 vs 43 (.427) q2vsq4 (.723) 43 vs q4 (.232) (continued) equity fund sixe and growth 9 quartile table 4 (continued) significance of return 1 2 3 4 comparisons p-value 1987-1992 5-year return % 1987 assets ($m) n 1982-1992 io-year return 46 1982 assets ($m) n 105.0 92.5 91.3 93.4 ql vs 42 (.172) 16.0 66.3 207.7 1399.3 ql vs 43 (9134) (29) (30) (30) (30) qlvs44 (.203) 42 vs 43 (.775) 42~~44 (.854) q3vsq4 (.612) 278.3 268.7 267.7 283.0 ql vs 42 (.567) 11.5 51.5 157.5 777.0 ql vs 43 (.555) (20) (20) (20) (21) qlvsq4 (.731) 42 vs 43 (.959) q2vsq4 (.379) q3vsq4 (.380) notes: *, **, and *** connote significance at the 10%. 5%. aad 1% levels, respectively. this table compares the mean total returns of the quardlts of funds with the smallest asset size to the mean total returns of the goattiles of funds with the largest asset size. for tbe 1982 size rankings. both the s-year (1982-1987) and the io year (1982-1992) rctoms am examined. for the 1987 size rankings, the s-year (1987-1992) returns are examined. all returns am in percentage terms; assets are in sm. n is the number of funds in each quartile. p-values are reported for t-tests of equality between quartile rctams. but in support of h2, the results do show strong indications that smaller funds outper form their larger counterparts in the aggressive growth category. superiority is evident in both the 5and lo-year returns. there is also some evidence of small fund superiority in the long-term growth objective. unlike grinblatt and titman (1989), these significant dif ferences are net of expenses. in further support of the superiority of small funds in this objective is the calculation of the increased wealth an investor could achieve by rebalanc ing her portfolio every 5 years to include only the smallest quartile of funds at that time. table 5 illustrates the results of starting in 1982 with an investment of $10,000 in each quartile of each investment objective, equally spread among the funds in each quartile. the portfolio is then reinvested in 1987 in the funds in each quartile at that time. as of 1992, this portfolio would have been worth $47,735 had it been invested in the smallest fund-size table5 results of a $10,000 investment in 1982 in each size quartile. quartile 1 (smallest-size) vs. quartile 4 (largest-size) quartile 1 2 3 ’ 4 aggressive growth $47.735 $41,489 $32,403 $32,817 long-term growth growth & income $42,194 $37,456 $37,710 $37,264 $40.939 $37,730 $36,213 $38,409 nores: assume an original investment portfolio on december 3 1.1982 of $10,000 in each quartile of each investment objective, equally weighted among the funds in cash quartile. after 5 years. the portfolio is reinvested to be equally weighted amongst the funds that arc in that same size quartile at that time (1987). at the end of 1992, the investments in the respec tive quartiles would be worth the following. these calculations am based on the rctum percentages presented in table 4. 10 financial services review 5( 1) 1996 quartile of the aggressive growth objective versus $32,817 for an investment in the larg est fund-size quartile of that objective. the findings support the argument that more aggressive funds have a smaller optimal size than less aggressive funds. managers of aggressive funds must often invest in smaller, lesser-known fms to achieve their objectives. finding these “diamonds in the rough” may become more difficult as the fund grows. a random sample of 20 of the funds in each investment category supports the assertion that aggressive growth funds invest in smaller, less-known firms. using morningstar, the median market value of the stocks that each of the 20 funds holds as an investment is determined. next the median of these individual fund medians is computed for each investment objective. for the aggressive growth objective, the median market value of stocks in the funds is $908m, for the long-term growth objective, $3.68b, and for growth and income objective, $6.88b. superior per formance in these smaller, aggressive growth funds thus attracts cash inflows thus making the achievement of superior returns in the future harder to achieve. asset growth is easier to manage in less aggressive fund objectives. additional cash contributed by investors may be easier to employ as these less aggressive funds generally invest in larger, more studied fms. having more cash may permit them to take larger stakes in these firms or add to their portfolio without a great deal of additional research. more aggressive funds, on the other hand, may find growth more of a research challenge as they are investing in smaller, less known firms. regarding fund nimbleness, managers with aggressive objectives may have to engage in more trading to stay on top of changing technologies and trends in emerging industries. table 1 shows that turnover is positively correlated to risk objective. for aggressive funds, growing larger impedes a manager’s ability to move quickly without attracting attention. in sum, the evidence suggests that current size offers some insight into future returns, but only with regard to the more aggressive funds. for investors with aggressive objec tives, smaller funds offer the better potential for superior returns. this is a new finding pre viously undocumented in the literature. grinblatt and titman (1989) find superior future returns in smaller aggressive growth funds, but not net of expenses. the returns here are net of expenses. for less risky fund objectives, the findings parallel those of grinblatt and tit man (1989) and droms and walker (1994). iv. summary and conclusions this paper builds upon existing research to provide a framework for individual investor consideration of fund size in the selection of mutual funds. superior historical returns are largely found in today’s largest funds. funds grow based mainly on new investments made in response to favorable communications about superior performance. to attempt to use fund size to predict future returns, however, is not useful unless investors have aggressive growth objectives. these investors should choose smaller funds. this is an intuitive result, as a flood of new cash presents more problems than opportunities for managers with aggressive growth objectives. for funds in growth and income objectives, on the other hand, the evidence supports the argument that there is no systematic relationship between fund size and future performance. equity fund size and growth 11 unfortunately, for individual investors seeking superior returns, investing in the funds with the best historical performance is not the answer. yesterday’s best performers are today’s largest funds. to achieve superior returns, the individual investor must find the “next magellan.” expecting the current-day “magellans” to systematically outperform their peers appears to he unrealistic. acknowledgments: the authors thank session participants in the 1995 academy of financial services meeting in new york, ny, w. taylor, t. lynch, j. k. dunlevy, k. lahey (the editor), and two anonymous referees for comments on earlier drafts. all remaining errors and omissions are our responsibility. notes 1. sreele and montingsrur are commercially available, computer-based sources of mutual fund information. both include annual returns, which are the basic returns periods studied. from the original steele database, a total of 26 funds (10 in ag, 8 in ltg, and 8 in gi) are closed (as of december 31,1992) to new investors. this paper focuses on size and return, and does not drop these funds from the sample or examine them separately. despite their closure, they still have an asset size and existing investors are generally permitted to continue to contribute (and withdraw) cash from the fund. several of the “closed funds,” such as acorn, also reopened periodically to new investors dur ing the return periods studied. this paper does not remove any merged funds in the database from the sample. the resulting survivorship bias has previously been found to be very small by grinblatt and titman (1989). 2. individual betas for each fund are computed for the most recent 3-year period using the standard and poors 500 index as the proxy for the market. 3. one common adage given to individual investors is to “pick a fund with a good track record.” fund management realizes this. it is common to see advertisements similar to: ‘the xyz fund is ranked number one in its category over the period from january 1, 19xx to december 31, 19xx.” 4. the market model is not explicitly used for risk adjustment here. as ippolito (1993) has observed, the inconsistent results of prior mutual fund studies are mainly due to the selection of the market portfolio. as roll (1977) argues, the true market portfolio cannot be determined. the paper relies on fund risk objective to classify funds. within each risk objective, three-year betas are tested for each quartile to assess whether risk bias is driving results. none of the quartile betas are signifi cantly different. see ciccotello and grant (1996). when samples have sizes not equally divisible by four, the “extra” funds are assigned to the largest asset-size quartile fist. references capon, n., fitzsimons, g., & prince, r. (1992). an individual level analysis of the mutual fund investment decision. working paper, columbia university, new york, ny. ciccotello, c.s., & grant, c.t. (1996). information pricing: the evidence from equity mutual funds. financial review, 31,365380. drams, w.g., & walker, d.a. (1994). investment performance of international mutual funds. juur nal of financial research, 42, l-14. 12 financial services review 5( 1) 1996 glosten, l., & harris, l. (1988). estimating the components of the bid-ask spread. joumul offinan economics, 21, goetzmann, w., b., & g. (1992). response to fund perfor working paper, university, new ny. grinblatt, & titman, (1989). mutual performance: an of quarterly holdings. journal business, 62, grinblatt, m., titman, s. the persistence mutual fund journal of 47, 1977-1984. d., patel, & zeckhauser, (1993). hot in mutual short-run persistence relative performance, joumul of 48,93-130. ippolito, (1989). efficiency costly information: study of fund performance. journal of 104, l-23. r.a. (1993). studies of fund performance, financial analysts jour nal, (jan./feb.), 42-50. patel, j., zeckhauser, r., & hendricks, d. (1992). investment flows and performance: evidence from mutual funds, cross-border investments, and new issues. in r. sato, r. levich, & r. ramachandran, e?ds., japan and internationalfinancial markets: analytical and empirical perspectives. cambridge, ma: cambridge university press. phalon, r. (1994, march 28). no room at the inn. forbes, p. 140. roll, r. (1977). a critique of the asset pricing theory’s tests; part i: on past and potential testability of theory. journal of financial economics, 4, 129-176. smith, g. (1994, october 10). inside fidelity: how the fund giant’s stock picking machine works. business week, p. 88. financial services review, 32(3) 1 cognitive ability and stock investment among chinese middle-aged and older population shan lei1 and lu fan2 abstract this study uses the wave 2 (2013), wave 3 (2015), and wave 4 (2018) data released by the harmonized china health and retirement longitudinal study (charls) to examine the relationship between cognitive ability and stock investment in the middle-aged and older populations of china. this study evaluates the relationship between the subjective and objective aspects of cognitive ability and the stock ownership and holdings in financial investments over time. we further compare the relationships of subsamples (older adults vs. middle-aged adults). the findings in this study provide implications for policymakers and financial professionals, as well as investors. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation lei, s., & fan, l. (2024). cognitive ability and stock investment among chinese middle-aged and older population. financial services review 32(3), 1-19. introduction understanding the relationship between cognitive ability and investment decision-making is an important research topic, particularly given the prolonged lifespan and increasingly complex market dynamics faced by households today. cognitive decline, including difficulty recalling and being challenged by numerical problems, is an inevitable aspect of aging (harada et al., 2013). research shows that cognitive ability typically peaks at the age of 30 and gradually diminishes thereafter (uscf, 2022). while a minor cognitive decline may not affect an individual’s management of their daily financial matters such as paying bills, it might impede their efficiency in making more complicated financial decisions such as portfolio choices and security selections 1 corresponding author (sxlei@salisbury.edu). salisbury university, salisbury, maryland, usa 2 university of georgia, athens, georgia, usa (agarwal et al., 2013; gorlick, 2010; starnes, 2019), ultimately influencing the quality of life. there is evidence supporting the positive role of cognitive ability in investment decisions and behaviors in western cultures (e.g., in the united states, europe, and the united kingdom). research shows that individuals with superior cognitive abilities have a higher propensity to invest in stocks, own more diversified portfolios, and achieve better investment performance (korniotis & kumar, 2010). conversely, individuals with weakened cognitive abilities are more likely to hold underperforming portfolios and less wealth. aging individuals experiencing cognitive declines are particularly susceptible to making problematic investment decisions. https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(3) 2 however, little empirical research has been conducted regarding this relationship, especially within an asian context (i.e., chinese middleaged and older populations). according to data released by the world health organization (who), china is experiencing one of the most rapid growth rates in aging populations globally. “the population of people over 60 in china is projected to reach 28% by 2040” (who, 2022, p. n.a.). stock investment has become a significant financial instrument for individual investors in china since the establishment of security markets in china in the 1990s (lu et al., 2020). today, stocks and mutual funds are considered to be the most risky financial assets owned by chinese households, excluding real estate (chen et al., 2020; gan et al., 2013; liao et al., 2010). given the severe impact that a cognitive decline can exert on financial decision-making, especially among those in the oldest age cohorts, and the important role of stock investment in chinese households, it behooves researchers, financial advisors, and policymakers to gain a deeper understanding of the association between cognitive ability and stock investment decisionmaking processes among middle-aged and older adults in china. the findings of this study provide insights and implications for individual investors, financial professionals, and policymakers in china. this study distinguishes itself from prior research in the following aspects. first, it incorporates both subjective and objective measures of cognitive ability, thus providing a more comprehensive perspective. second, this study is one of the few published studies investigating the relationship between cognitive ability and stock investment among the chinese middle-aged and older population. previous studies in this realm have predominantly focused on decision-makers in north america, europe, and australia, leaving a gap in understanding regarding the relationship between cognitive ability and stock investment decisions among chinese middle-aged and older adults. this study addresses this gap in the literature. literature review cognitive ability cognitive ability is a multifaceted concept referring to a person’s “ability to perform the mental processes required in a variety of tasks” (mazzonna & peracchi, 2018, p. 3). among the subdomains associated with cognitive ability, the mental skills and abilities related to orientation, memory, reasoning, concentration, and the executive function have been closely examined, as reported in the literature (e.g., christelis et al., 2010; richards et al., 2004). another dichotomous categorization of cognitive ability includes crystallized intelligence and fluid intelligence (li et al., 2015). researchers have reported finding that people with strong cognitive ability have better memories. memory and numeracy are two important components of cognitive ability and have been found to be associated with financial behaviors and decisions. memory, as an important dimension of cognitive ability, tends to decline with age (e.g., mazzonna & peracchi, 2018; salthouse, 1996). memory loss, which frequently occurs among older individuals, is highly associated with decreases in financial literacy, capability, and confidence, which in turn, impair financial decision-making (gamble et al., 2015). severe memory loss is known to be associated with out-of-pocket medical expenditures, which may influence the overall wealth condition of a household. some researchers have found that cognitive ability is related to household credit card and home equity loan decisions (agarwal & mazumder, 2013), risk preference (dohmen et al., 2010; frederick, 2005), along with retirement saving behavior (banks & oldfield, 2007). moreover, numeracy has also been studied as a significant component of cognitive ability. numeracy is known to be associated with retirement savings and investment portfolio decisions (banks & oldfield, 2007); borrowing, savings, and tax decisions (huhmann & mcquitty, 2009); and wealth accumulation, risk perception, and time preference (estrada-mejia et al., 2016). cognitive ability and financial decisions/stock investment previous studies have explored the relationship between cognitive ability and the financial decisions and behaviors of individuals and families, as well as other economic outcomes in general. for example, pak and babiarz (2018) lei & fan 3 investigated the association between cognitive ability and risky asset holding among older american adults. they found that a lower cognitive ability was associated with lower participation in the stock market (i.e., stocks out of total financial net worth and the probability of holding stocks); however, in their study, instrumental variable models did not support a causal inference. likewise, mazzonna and peracchi (2018) noted that an unawareness of cognitive decline may result in negative financial consequences such as incurring financial losses and making poor financial decisions. further, fan and lim (2022) noted that cognitive ability was also significantly associated with financial advice-seeking behavior and advice-source preferences, with older adults who exhibited better cognitive ability (i.e., better memories and numeracy) being less likely to seek financial advice from family members or social networks. cognitive abilities, which decline with age, also play an important role in describing the investment decisions of individual investors (korniotis & kumar, 2010). using data from the health and retirement study (hrs), cheung and yilmazer (2019) reported that memory issues in older american adults diagnosed with dementia or alzheimer’s disease made them less likely to hold risky assets. they also noted a negative association with the number of risky assets in their investment portfolios and household financial net worth, through the mediator of cognitive ability. christelis et al. (2010) used a european survey of older individuals that measured three dimensions of cognitive ability: (a) numeracy (or mathematical skills), (b) verbal fluency, and (c) recall skills. they found that participation in the stock market, either through direct investing in stocks or indirect investing through mutual funds or retirement accounts, was strongly and positively associated with cognitive abilities, after controlling for bequest motives, health status, social interactions, and socioeconomic characteristics. the association was not as strong for bond holding among older adults living in european countries. christelis et al. argued that cognitive ability is correlated with the ability to process financial data, with those who exhibit higher cognitive ability using lower costs to process complex financial data related to the stock market. the strong association between cognitive ability and asset allocation choices has been confirmed in other studies. using the 2006–2008 waves of the hrs, browning and finke (2015) observed a significant relationship between retirees’ cognitive ability and stock reallocation decisions during recessions. in their study, they focused on fluid cognitive functioning proxied by working memory and the numeracy of respondents, controlling for risk tolerance and sentiment during a recession. they found that those with low cognitive ability were more likely to respond to sentiment effects, and thus were more likely to allocate away from stocks, compared to those with a better working memory and numeracy. kim et al. (2012) used the 2004 hrs to determine that a strong relationship exists between cognitive ability and stock ownership among older americans. other factors relating to stock investment in addition to cognitive ability, the current literature also documents other factors associated with stock investing. prior research has widely documented the positive role of investors’ wealth and income in stock investment decision-making (e.g., browning & finke, 2015; cheung & yilmazer, 2019; pak & babiarz, 2018). holding one’s cognitive score and other controls constant, being married, having a college degree, holding life insurance coverage, and having a higher net worth are known to be positively associated with holding stock among older american adults (pak & babiarz, 2018). similarly, with fluid cognitive ability (i.e., word recall and numeracy) and other factors controlled, browning and finke (2015) found that an investor’s planning horizon, educational attainment, wealth, and race (being white) were positively associated with stock allocation decisions. additionally, in a study by cheung and yilmazer (2019), it was determined that household net worth and income were positively associated with not only the ownership of risky assets such as stocks but also the proportion of such assets in a household’s overall portfolio. in their study, age was negatively associated with the proportion of risky assets after controlling cognitive ability and other financial services review, 32(3) 4 factors. some research has also determined the role of gender in investment behavior (holden & tilahun, 2022), with men being more likely to hold stocks and other risky investment assets. conceptual framework and hypotheses according decision theory, “decision making is one of the basic cognitive processes of human behaviors by which a preferred option or a course of actions is chosen from among a set of alternatives based on certain criteria” (wang & ruhe, 2007, p. 11). within the realm of simple decision-making, which encompasses intuitive and empirical analyses, lies the fundamental concepts of basic and core cognitive processes. heuristic and rational decision-making strategies can help in real-world decisions (hastie, 2001; wang & ruhe, 2007). the current study focuses on stock investment choices, which, in essence, involves a “repetitive application of the fundamental cognitive process” (wang & ruhe, 2007, p. 1). as frederick (2005) stated, the relationship between cognitive ability and decision-making is too important to be ignored. decision theory provides a solid theoretical background for examining this relationship. as noted above, prior studies have highlighted the role of cognitive ability in investment decision-making. in the majority of these studies, a lower cognitive ability was associated with poor financial decisions in general (mazzonna & peracchi, 2018), along with less financial net worth (cheung & yilmazer, 2019), lower participation in the stock market (pak & babiarz, 2018), and a lower probability of holding risky asset portfolios (cheung & yilmazer, 2019; kim et al., 2012). this study extends the literature by focusing on a specific market segment: middle-aged and older adults. this is an important population to evaluate given the negative relationship between age and cognitive ability. in this regard, the scaffolding theory of aging and cognition (stac) and its revision (park & reuter-lorenz, 2009; reuterlorenz & park, 2014) are relevant. the original stac explains the potential variations in cognitive ability by age from the perspective of biological and neurophysiological factors. the scat shows that aging may trigger neural challenges (i.e., structural changes in the brain) and functional deterioration (i.e., maladaptive age-related brain activity, including decreased memory), which may reduce cognitive functioning levels. the revised stac (stac-r) expands the understanding of aging and cognitive function to incorporate a life-course perspective, which emphasizes the experience accumulated over a person’s life. the stac-r proposes that two factors have impacts on brain functions: (a) an individual’s lifespan and (b) experiences during their life course. based on this conceptual framework, and the theoretical and empirical literature regarding cognitive ability and stock investment, this study examined the following hypotheses: h1: cognitive ability is positively associated with stock ownership among chinese middle-aged and older adults. h2: cognitive ability is positively associated with holding stock shares in an investor’s portfolio among chinese middle-aged and older adults. methodology data this study used the wave 2 (2013), wave 3, (2015), and wave 4 (2018) datasets, which are the ones most recently released by the harmonized china health and retirement longitudinal study (charls), to examine the association between cognitive ability and stock investment among older chinese people. the charls is a national longitudinal survey conducted by the national school of development (china center for economic research) at peking university. the harmonized longitudinal data are a user-friendly version, and are part of the gateway to global aging data, which compare similar surveys of older populations in different countries. the charls provides a large array of information on the financial situations faced by individuals aged 45 years and older. the charls also provides some information about individuals’ expectations and preferences, such as their self-reported health status. information regarding demographic characteristics such as age, gender, education, and marital status was also collected in this survey. the first wave of the charls was lei & fan 5 conducted in 2011. respondents were followed every two years. the financial variable questions were significantly changed after wave 1. thus, in this study, only data from the harmonized wave 2, wave 3, and wave 4 surveys were evaluated. the final sample size used in this study was 36,661 observations. variables stock ownership and shares. the main dependent variable in this study was stock ownership, which was defined as whether or not a household owned stocks in their portfolio in the form of directly held stocks or stock mutual funds. stock ownership was used as a dummy variable equal to 1 (otherwise 0) for respondents whose stock amount was positive. another main dependent variable was stock shares, which was defined as the percentage of the stock amount in financial assets. cognitive ability. following prior studies (e.g., browning and finke (2015), who used the hrs dataset; christelis et al. (2010), who used the share dataset; and yu et al. (2021), who used the charls dataset), objective cognitive ability was measured using total recall words and math test scores, which are more relevant to financial decisions. in particular, total recall words (ranging from 0 to 20) was measured based on the total amount of words a respondent could immediately recall correctly from a 10-word list, together with the number of words the respondent could recall correctly from a 10-word list after a delay spent answering other survey questions. respondents were required to complete a math test, which “asks the individual to subtract 7 from the prior number, beginning with 100 for five trials” (harmonized charls documentation, 2021, p. 211). math scores (ranging from 0 to 5) measured the number of correct subtractions in the serial 7’s test across five trials. in addition, subjective cognitive ability was assessed using self-reported doctor’s diagnosis of a memory problem (see fritsch et al., 2014), with 1 indicating a respondent being told they had a memory-related disease and 0 otherwise. further, self-reported memory was also used to assess subjective cognitive ability, which was measured using a scale ranging from 1 for excellent to 5 for poor (harmonized charls documentation, 2021, p. 204). control variables. based on the literature review, the following controlled variables were included in the multivariate tests: income, total financial net worth, homeownership, vehicle ownership, age, gender, marital status, education, selfreported health status, whether a respondent had a dependent child, self-employment, retired, living in an urban region, and having private health insurance coverage. the operationalization of these variables is shown in table 1. models a descriptive analysis was conducted to show the sample characteristics overall and by subsample. given that the charls dataset used in this study is a longitudinal dataset, repeated responses about stock investments were collected from the same respondents. thus, a generalized estimating equations (gees) methodology, introduced by liang and zeger (1986), was used to determine the relationship between cognitive ability and stock investment over time. a gee analysis estimates the relationship between changes in an outcome variable of a subject and the covariates while allowing for the possible correlations between repeated measures of the variables over time for the same individuals in longitudinal datasets. thus, it provides more accurate estimates (liang & zeger, 1986; smith & smith, 2006). this method has been widely used in previous research using longitudinal datasets (see filer & golbe, 2003; ghisletta & spini, 2004; hamza et al., 2021; klos et al., 2005; rechner & dalton, 1991). based on this method, a logistic regression model was specified as the link function to evaluate the relationship between cognitive ability and stock ownership, and an ols regression model was specified as the link function to examine the relationship between cognitive ability and stock shares, while considering the within-subject correlation. financial services review, 32(2) 6 table 1. summary of the variables used in the empirical model variable name variable type measure type variable description stock ownership dependent variable dichotomous yes=1, defined as the situation that respondents’ stock amount was positive; no=0 [reference] stock share dependent variable continuous defined as the percentage of stock amount in financial assets self-reported memory independent variable 5-level categorical scale from 1 for excellent to 5 for poor total recall words independent variable continuous varies between 0 and 20 math test scores independent variable continuous varies between 0 and 5 doctor diagnosed memory-related disease independent variable dichotomous yes=1; no=0 household income independent variable continuous included as log-transformed variable due to its non-normal distribution household total financial net worth independent variable continuous included as log-transformed variable due to its non-normal distribution homeownership independent variable dichotomous non-homeowners=0 [reference], homeowner=1); vehicle ownership independent variable dichotomous non-vehicle owners=0 [reference], vehicle owners=1); age independent variable continuous gender independent variable 2-level categorical male [reference], female marital status independent variable 2-level categorical not married [reference], married education independent variable 3-level categorical less than lower secondary education less [reference], upper secondary & vocational training, and tertiary education. have a dependent child independent variable dichotomous yes=1; no=0 [reference] self employed independent variable dichotomous yes=1; no=0 [reference] retired independent variable dichotomous yes=1; no=0 [reference] living in the urban region independent variable dichotomous yes=1; no=0 [reference] have private health insurance coverage independent variable dichotomous yes=1; no=0 [reference] self-reported health status independent variable 5-level categorical poor [reference], fair, good, very good and excellent financial services review, 32(2) 7 to examine the relationship between cognitive ability and household stock ownership, the following logistic model was estimated: 𝑃(𝑌𝑖𝑡 = 1|𝐶𝑖𝑡, 𝑋𝑖𝑡) = exp⁡(𝛽′𝐶𝑖𝑡+𝛾 ′𝑋𝑖𝑡+𝜀𝑖𝑡) 1+exp⁡(𝛽′𝐶𝑖𝑡+𝛾 ′𝑋𝑖𝑡+𝜀𝑖𝑡) , (1) where 𝑌𝑖𝑡 is the stock ownership of household 𝑖 at time 𝑡; 𝐶𝑖𝑡 is a vector measuring the cognitive ability of the household’s financial respondent, including their self-reported memory, total recall words, math test scores, and doctor diagnosed memory-related disease; and 𝑋𝑖𝑡 captures other control variables related to household stock ownership. all the variables are defined in table 1. to study the relationship between cognitive ability and stock shares as a proportion of financial assets, the following ols model was estimated: 𝑌𝑖𝑡 ∗ = ⁡𝛽𝐶𝑖𝑡 + 𝛾𝑋𝑖𝑡 + 𝜀𝑖𝑡 , 𝜀𝑖𝑡~𝑁(0, 𝜎 2), (2) where 𝑌𝑖𝑡 ∗ denotes the stock shares, depending on the same cognitive ability variables and other controls as described for model (1). results descriptive statistics the analysis began by examining the descriptive statistics overall and by subsamples of respondents with and without stock investments in the 2018 wave. on average, only 2% of the respondents owned stocks in their portfolios. stock owners allocated nearly 40% of their financial assets to stocks (37.7%). across the sample, many respondents reported themselves to have poor (31.8%) or fair (54%) memory. the mean total recall word score was 7.3 out of 20, while the mean math score was 3.6 out of 5. slightly more than 2% of the respondents reported having been told by their doctor that they had a memory-related disease. overall, the mean household income was rmb 37,417, whereas the median household income was rmb 15,500. the mean total household nonhousing financial net worth was rmb 17,486. the median total household non-housing financial net worth was rmb 2,000. on average, 56% of respondents were homeowners, and nearly 63% of the respondents reported owning a vehicle. on average, nearly 50% of the respondents were female and over 40% had dependent children (40.4%). a majority of the respondents were married (85.7%) and had less than a lower secondary level of education (88%). over one-third of the respondents lived in an urban region (37.6%), with few being selfemployed (6.3%) or having private health insurance coverage (3.5%). over half of the respondents reported a good health status (51.9%). table 2 presents the characteristics of the sample. financial services review, 32(2) 8 table 2. demographic profile: cognitive ability and stock ownership—wave 4 variables all non-stockowners (98.0%) stockowners (2.0%) cognitive ability self-reported memory excellent* 0.8 0.7 1.5 very good** 5.3 5.3 7.8 good*** 8.2 8.0 17.1 fair*** 54.0 53.8 60.4 poor*** 31.8 32.2 13.2 total recall words*** mean: 7.3 median: 7 mean: 7.2 median: 7.0 mean: 10.5 median: 11.0 math test scores*** mean: 3.6 median: 4 mean: 3.6 median: 4 mean: 4.4 median: 5.0 doctor diagnosed memory-related disease 2.3 2.3 2.5 financial situations household income*** mean: 37,416.9 median: 15,500.0 mean: 35,576.6 median: 14,600.0 mean: 138,375.3 median: 87,000.0 household total financial wealth*** mean: 17,485.6 median: 2,000.0 mean: 9,364.9 median: 2,000.0 mean: 439,474.9 median: 157,265.1 homeowners*** 56.0 55.7 70.9 vehicle owners** 62.7 62.6 68.0 demographics age*** mean: 59.3 median: 59.0 mean: 59.4 median: 59.0 mean: 56.5 median: 55.0 female 47.8 47.8 49.0 married*** 85.7 88.9 43.5 educ: less than lower secondary education*** 88.0 9.8 37.7 educ: upper secondary & vocational training*** 10.3 1.4 18.8 educ: tertiary education*** 1.7 15.3 84.7 having a dependent child 40.4 40.4 40.7 lei & fan 9 note: weighted mean values are reported for household income (in chinese yuan), household total non-housing final wealth (in chinese yuan), total recall words, math test scores, and age. for other variables, weighted mean percentages are reported. percentages may not sum to 100% due to rounding. a chi-square test was employed for categorical independent variables while t-test was used for continuous independent variables. *** p < 0.001, ** p < 0.01, * p < 0.05 table 2 continued self-employed 6.3 6.3 6.1 retired*** 33.1 32.7 53.1 living in the urban region*** 37.6 36.6 89.5 having private health insurance coverage*** 3.5 3.2 16.7 health expectation: excellent 11.4 11.4 10.9 health expectation: very good*** 13.1 12.9 25.3 health expectation: good 51.9 51.8 55.4 health expectation: fair*** 18.4 18.6 6.9 health expectation: poor*** 5.2 5.3 1.5 financial services review, 32(2) 10 it was noted that more stock owners reported having good, very good, or excellent memory, while more non-stock owners reported having poor or fair memory. among stock owners, the mean total recall words and math scores were 10.5 and 4.4, respectively, while these measures were 7.2 and 3.6 for non-stock owners, respectively. for stock owners, the mean household income was rmb 138,375, whereas the mean total household non-housing financial net worth was rmb 439,475. for non-stock owners, the mean household income was rmb 35,577. the mean total household non-housing financial net worth was rmb 9,365. more stock owners were homeowners (70.9% vs. 55.7%) and vehicle owners (68.0% vs. 62.6%). the proportion of stock owners (84.7%) with a tertiary level of education was significantly higher than that among non-stock owners (5.3%). additionally, more stock owners considered their health status to be very good (25.3%), while more non-stock owners reported their health status as only poor or fair (18.6% for fair and 5.3% for poor). multivariate analyses the results listed in table 3 provides confirmation of the first hypothesis: cognitive ability was positively associated with stock ownership. specifically, respondents with better objective cognitive ability were more likely to own stocks. the results showed that respondents who recalled more total words (at a p < 10% significance level) and had a higher score on the math test showed an increased probability of owning stocks (odds ratio = 1.034 and 1.147, respectively). the results showed a significant positive relationship between doctor-diagnosed memory problems and stock ownership, which indicates that those who reported being diagnosed with memory problems were more likely to invest in stocks. however, the results did not find a significant relationship between self-reported memory status and stock ownership. consistent with much of the prior literature (e.g., cheung & yilmazer, 2019), wealth and income were found to be positively associated with stock ownership. it was also noted that homeowners had a nearly 50% higher propensity to invest in stocks, compared to non-homeowners. having private health insurance coverage was also found to be a significantly positive contributor to a respondent’s degree of stock ownership. in addition, findings confirmed the results from prior studies (e.g., browning & finke, 2015) that those with higher education levels had a higher probability of investing in stocks. for example, respondents in this study with upper secondary and vocational training were approximately 2.8 times as likely to invest in stocks, compared to respondents with less than a lower secondary level of education. further, their self-reported health status in general bore a positive relation to the propensity to invest in stocks. for example, compared to respondents who claimed to be in poor health, those who reported having excellent health were 1.6 times more likely to own stocks in their portfolios. findings also showed that respondents who were married, lived in cities, and were retirees were more likely to own stocks (odds ratio = 1.147, 6.675, and 2.417, respectively). lei & fan 11 table 3. multivariate analysis—cognitive ability, stock ownership, and stock shares stockownership stock share coef odds ratio coef st. err. independent variables cognition self-reported memory (ref: poor) excellent -0.189 0.828 0.002 0.007 very good 0.042 1.043 -0.001 0.002 good 0.082 1.085 0.002 0.003 fair 0.135 1.144 0.000 0.001 total recall words 0.034* 1.034 0.000 0.000 math scores 0.137*** 1.147 0.001 0.000 having doctor diagnosed memory-related disease (ref: no) 0.726* 2.067 0.003 0.004 financial situations household income (log) 3.157** 23.491 0.007 0.005 household total financial wealth (log) 13.836*** 1.021*106 0.003* 0.002 homeownership (ref: non-owner) 0.375* 1.455 0.003** 0.001 vehicle ownership (ref: non-owner) -0.004 0.996 -0.001 0.002 demographics age 0.0329 1.033 0.001* 0.001 age2 -0.0004 1.000 0.000* 0.000 female -0.103 0.902 -0.002 0.001 married 1.257*** 3.516 0.005** 0.002 educ: (ref: less than lower secondary education) upper secondary & vocational training 1.019*** 2.770 0.017*** 0.003 tertiary education 1.842*** 6.307 0.077*** 0.013 having a dependent child -0.098 0.907 0.001 0.001 self-employed (ref: no) -0.070 0.932 -0.001 0.002 retired (ref: no) 0.883*** 2.417 0.009 0.002 living in an urban region 1.898*** 6.675 0.010*** 0.001 having private health insurance coverage 0.939*** 2.558 0.021*** 0.006 self-reported health status: (ref: poor) excellent 0.955* 2.598 0.005 0.003 very good 1.200** 3.321 0.008** 0.003 good 0.841* 2.318 0.003 0.002 fair 0.175 1.191 0.000 0.002 intercept -278.835 0.000 -0.196* 0.081 note: *** p < 0.001, ** p < 0.01, * p < 0.05 the results listed in table 3 also highlight that those with greater math acumen (i.e., having a higher score on the math test) held more stocks in their financial portfolios (p < 0.08). to be specific, for every one point increase in the math score earned by a respondent, the weight of the stock holdings in their portfolios increased by 0.1%. additionally, the results provide evidence that older adults with a greater financial net worth held more shares of stock in their portfolios (at p < 0.10). further, it was noted that homeowners invested 0.3% more in stocks in their portfolios than non-homeowners. respondents with private medical insurance coverage were found to have 2.1% more shares of stock than those who did not. financial services review, 32(3) 12 the results also highlight the roles played by other personal characteristics in describing the weight of stocks held in financial portfolios. for example, a one-year increase in age was found to be related to a 0.12% increase in the weight of stock holdings, but with a decreasing rate. further, compared to those who had less than a lower secondary level of education, those who reported having upper secondary and vocational training were found to invest 1.7% more of their portfolio in stocks, whereas those who reported having a tertiary level of education were shown to invest 7.7% more in stocks. subsample results previous research has suggested that people are more likely to reduce their stock holdings in alignment with declines in cognitive ability. following christelis et al. (2010), we split the sample in this study by age based on whether a respondent was younger (age 45–59, middle-aged adults) or older adults (age 60 and older), which is the normal retirement age for most people in china. age 60 is also the acceptable cut-off used to define older adults (han et al., 2020). an analysis of the subsamples allowed for an assessment of potential differences in cognitive ability and their relationship to stock ownership between middle-aged and older adults. the results listed in table 4 illustrate how the math score measure was a significant positive factor related to stock ownership among older investors. those with doctor-diagnosed memory problems were more likely to also hold stocks. this finding contradicts what some have reported in the literature. it was found that middle-aged investors who could recall more total words had a higher probability of owning stocks and were shown to own more stocks in their portfolios. a comparison of respondents’ personal characteristics related to stock investment provided some interesting insights. for example, in the older adult subsample, those with an excellent health status were found to be almost three times as likely to own stocks, compared with those who reported being in poor health. married older investors were observed to have a higher probability of investing in stocks (odds = 3.767), while living with dependents was found to contribute negatively to stock ownership (odds = 0.555). these factors were not significant in the middle-aged subsample group. interestingly, homeownership was positively associated with stock ownership in the middle-aged subsample, which provides some evidence that homeownership is an important wealth indicator (beracha et al., 2017) among middle-aged investors. it was also found that middle-aged investors with private health insurance coverage had a higher propensity to invest in stocks, as well as a higher weight of stock shares in their portfolios. this might be explained by the cushion effect and resilience provided by private health insurance coverage (jain & garg, 2022). an analysis of data from the subsamples also provides some insight regarding the shares of stock investment in financial assets over time. first, it was noted that health status was a vital factor related to stock share ownership among the older-adults subsample. those who reported having a very good health status reported having 1.1% more weight of stocks in their portfolios compared to the those in poor health. additionally, consistent with the association between homeownership and stock ownership, for the middle-aged-adults subsample, compared to non-homeowners, homeowners were found to have 0.49% more weight in stocks in their portfolios. lei & fan 13 table 4. multivariate analysis—subsamples—cognitive ability, stock ownership, and stock shares older adults, aged 60 and older middle-aged adults, aged 45 to 59 stock ownership stock shares stock ownership stock shares coef odds ratio coef st. err. coef odds ratio coef st. err. independent variables cognition self-reported memory (ref: poor) excellent -1.064 0.345 0.000 0.008 -0.017 0.983 0.001 0.010 very good 0.045 1.046 -0.003 0.004 -0.028 0.973 0.001 0.004 good 0.465 1.592 0.007 0.004 -0.329 0.720 -0.002 0.003 fair 0.193 1.213 0.003 0.002 0.032 1.032 -0.002 0.002 total recall words 0.005 1.005 0.000 0.000 0.055* 1.057 0.0004* 0.000 math scores 0.211** 1.235 0.001 0.000 0.055 1.056 0.000 0.001 having doctor diagnosed memory-related disease (ref: no) 0.664* 1.942 0.002 0.004 0.638 1.892 -0.003 0.002 financial situations household income (log) 2.484*** 11.987 0.098 0.078 5.275** 195.430 0.003 0.003 household total financial wealth (log) 22.012*** 3.629*109 0.058 0.070 9.512** 1.352*104 0.002 0.001 homeownership (ref: nonowner) 0.019 1.019 0.000 0.002 0.789*** 2.202 0.005*** 0.001 vehicle ownership (ref: non-owner) 0.207 1.230 0.001 0.002 -0.115 0.891 -0.003 0.002 demographics age 0.205 1.228 0.004 0.003 -0.118 0.889 -0.001 0.002 age2 -0.002 0.998 0.000 0.000 0.001 1.001 0.000 0.000 female -0.045 0.956 -0.002 0.002 -0.181 0.835 -0.002 0.002 married 1.326** 3.767 0.004* 0.002 0.901 2.462 0.004 0.003 educ: (ref: less than lower secondary education) upper secondary & vocational training 0.632* 1.881 0.014** 0.005 1.349*** 3.853 0.015*** 0.003 tertiary education 0.893* 2.443 0.062** 0.021 2.617*** 13.690 0.091*** 0.018 financial services review, 32(3) 14 table 4 (continued) stock ownership stock shares stock ownership stock shares coef. odds ratio coef. st. err. coef. odds ratio coef. st. err. having a dependent child -0.590* 0.555 -0.005*** 0.002 0.285 1.330 0.005** 0.002 self-employed (ref: no) -0.749 0.473 -0.007*** 0.002 -0.058 0.943 0.001 0.003 retired (ref: no) 1.449*** 4.257 0.008*** 0.002 0.555 1.742 0.009*** 0.003 living in an urban region 2.202*** 9.042 0.008*** 0.002 1.641*** 5.159 0.009*** 0.002 having private health insurance coverage 0.582 1.789 0.026 0.015 1.021*** 2.775 0.018** 0.006 self-reported health status: (ref: poor) excellent 1.073* 2.924 0.006 0.004 0.888 2.430 0.004 0.004 very good 1.334** 3.796 0.011* 0.005 1.012 2.751 0.005 0.004 good 0.843 2.324 0.004 0.003 0.859 2.361 0.002 0.004 fair 0.511 1.667 0.003 0.003 -0.498 0.608 -0.004 0.004 intercept -407.271 0.000 -2.476 1.559 -236.063 0.000 -0.070 0.058 note: *** p < 0.001, ** p < 0.01, * p < 0.05 discussion and implications this study was designed to investigate the relationship between cognitive ability and stock investment using a longitudinal dataset of middle-aged and older chinese households. overall, the findings confirm the positive role of cognitive ability in the context of stock investment decisions. specifically, this study contributes to the existing literature by examining multiple dimensions of cognitive ability among chinese middle-aged and older investors. the study confirms h1, as evidenced by identifying the positive roles of math skills and word recall ability in stock ownership. furthermore, it highlights that math skill holds greater significance and magnitude compared to word recall ability in describing stock investment decisions. we did not find any significant results to support h2 in the full sample. however, a positive relationship between word recall ability and stock shares in the portfolios of the middleaged adults subsample was observed, partially supporting h2. the results from this study can be used to assist policymakers and financial service institutions to depict a profile of middle-aged and older stock investors in china. they share some similar personal characteristics, including being wealthy, younger (age 45–59), married, and well-educated. compared to non-stock owners, stock investors live in urban regions and report relatively better health status. this information is crucial for financial institutions in identifying target clients. it is also useful for professionals who work with middle-aged and older clients in china. as the results from this study show, engaging clients in ways that uncover aspects of numeracy and cognitive decline can help a financial service professional provide relevant advice with appropriate strategies. for example, the results indicate that individuals with better math skills are more likely to invest in stocks and increase the stock holdings of their portfolios. since these types of clients understand math better, it is reasonable to argue that they are more comfortable using data when making financial decisions. the use of charts, graphs, and other visual aids might appeal to these types of clients. additionally, this study shows that objective cognitive ability contributes positively to stock investment decisions. however, this study did not find evidence supporting the role of subjective lei & fan 15 self-reported memory problems in stock investment decisions. moreover, a positive relationship between self-reported doctordiagnosed memory problems and stock ownership was noted. these findings create educational opportunities for financial institutions, financial service professionals, and educators. on the one hand, the findings indicate the usefulness of objective measures of cognitive ability. financial service professionals should rely more on objective measures of cognitive ability when collecting information about a client’s situation. on the other hand, reflecting on an increased probability of stock investment for those who reported being told by doctors that they had memory-related problems might indicate a point of caution when providing advice and education. those with memory loss may not understand the risk and return characteristics of stocks, and thus could make allocation decisions that put them at risk of financial loss and distress. it is possible that some cognitively impaired individuals make decisions based on the fear of losing future opportunities rather than basing decisions on a comprehensive analytical investment strategy. it is also possible that because of the chinese culture’s emphasis on wealth accumulation, these investors, despite receiving medical advice regarding memory issues, might actively pursue investment avenues such as stocks to augment their wealth and ensure future provisions. conversely, chinese cultural norms often regard cognition decline as a natural aspect of aging, which leads to insufficient attention being given to this concern (dai et al., 2013; han et al., 2020). this might cause investors to disregard medical recommendations and advice about memory loss. this possibility creates opportunities for financial institutions, working with policymakers, to design investment programs that help older individuals make more informed investing decisions that align with longterm financial goals. for example, financial institutions and policymakers could work together to develop heuristic tools to help clients overcome cognitive biases during the investment decision-making process (otuteye & siddiquee, 2015). further, the multivariate analyses showed how different factors can be associated with stock investment decisions across subsamples of older (i.e., 60 years of age and older) and younger investors (i.e., aged 45 to 59 years) in china. for the younger subgroup, total word recall skill was found to be a significant contributor to stock investment decisions, while doctor-diagnosed reported memory problems and math scores were two significant factors related to stock ownership for those in the older subgroup. financial institutions and educators may want to apply these findings when developing assessments and strategic interventions. rather than relying on one common assessment tool across client populations, findings from this study suggest the need for a more nuanced approach, where different aspects of cognitive ability for older and younger subgroups can be used for better outcomes. as mentioned earlier, a positive relationship between doctor-diagnosed memory problems and an increased possibility of stock investment raises the need for further investigation. this is particularly important for the older subgroup because of their shorter investment horizon and potential need to rely on financial assets to maintain their retirement life quality, as well as an increased possibility of potentially diminishing cognitive ability. although findings from this study cannot be used to indicate a causal pathway from cognitive decline to increased probability of stock ownership, these results do nonetheless suggest that it is essential to build a structured financial planning process to ensure that a client’s requests are handled appropriately (starnes, 2019). this study also showed that unique factors such as being married, living with dependent children, and exhibiting general good or better health status provide significant insights into the stock investment decisions of investors in the older subgroup. findings also highlight the positive role of homeownership in stock investment among the younger subgroup. it is possible that owning a home provides the capacity to take on more financial risk. it is also possible that homeownership provides a pathway to portfolio diversification. it is therefore important for financial service professionals to endeavor to familiarize themselves with a client's overall financial services review, 32(3) 16 health status and patterns of asset ownership, given the significance of these findings. finally, this study adds to the current literature by showing important associations between various characteristics of chinese middle-aged and older investors and their preferences for holding stocks in their portfolios. for example, consistent with prior research, investors with higher educational achievements were found to be more likely to hold stocks in their portfolios. this confirms the significance of literacy programs, including those at the general education level, in prompting increased participation in the stock market. we also found that people living in urban regions had a higher propensity for owning stocks. this provides policymakers with valuable insight into the disparities regarding access to stock market resources or a comprehension of stock investments in rural regions of china. although this study used a longitudinal dataset to examine the relationship between cognition and stock investment, it was limited to incorporating only three waves of data because of dataset constraints. consequently, caution is warranted in interpreting the results as causal inferences because of the potential endogeneity of the measured cognitive ability. furthermore, the findings were constrained by the lack of detailed information regarding cognitive ability within the dataset. notably, the reliance on the respondents' self-reported memory-related disease diagnoses introduced uncertainty regarding accuracy, and crucial details such as disease types, severity, treatment, and outcomes remained undisclosed. moving forward, future research directions could entail revisiting this topic with expanded data collection efforts. this would allow for a more comprehensive examination of the role of cognitive ability in portfolio choice across a broader time span and through more sophisticated analytical approaches. references agarwal, s., & mazumder, b. (2013). cognitive abilities and household financial decision making. american economic journal: applied economics, 5(1), 193–207. https://doi.org/10.1257/app.5.1.193 banks, j., & oldfield, z. (2007). understanding pensions: cognitive function, numerical ability and retirement saving. fiscal studies, 28(2), 143–170. https://doi.org/10.1111/j.14755890.2007.00052.x beracha, e., skiba, a., & johnson, k. h. (2017). housing ownership decision making in the framework of household portfolio choice. journal of real estate research, 39(2), 263-289. browning, c., & finke, m. (2015). cognitive ability and the stock reallocations of retirees during the great recession. journal of consumer affairs, 49(2), 356375. https://doi.org/10.1111/joca.12065 cheung, c. h., & yilmazer, t. (2019). wealth management while dealing with memory loss. journal of family and economic issues, 40(3), 470-485. https://doi.org/10.1007/s10834-01909610-w chen, f., hsu, c. l., lin, a. j., & li, h. (2020). holding risky financial assets and subjective wellbeing: empirical evidence from china. the north american journal of economics and finance, 54, 101142. christelis, d., jappelli, t., & padula, m. (2010). cognitive abilities and portfolio choice. european economic review, 54(1), 1838. https://doi.org/10.1016/j.euroecorev.200 9.04.001 dai, b., mao, z., mei, j., levkoff, s., wang, h., pacheco, m., & wu, b. (2013). caregivers in china: knowledge of mild cognitive impairment. plos one, 8(1), e53928. dohmen, t., falk, a., huffman, d., & sunde, u. (2010). are risk aversion and impatience https://doi.org/10.1111/joca.12065 https://doi.org/10.1016/j.euroecorev.2009.04.001 https://doi.org/10.1016/j.euroecorev.2009.04.001 lei & fan 17 related to cognitive ability? american economic review, 100(3), 1238–1260. estrada-mejia, c., de vries, m., & zeelenberg, m. (2016). numeracy and wealth. journal of economic psychology, 54, 53-63. fan, l., & lim, h. (2022). cognitive abilities and seeking financial advice: differences in advice sources. journal of financial counseling and planning, 33(1), 97-114. doi: 10.1891/jfcp-1900079 filer, r. k., & golbe, d. l. (2003). debt, operating margin, and investment in workplace safety. the journal of industrial economics, 51(3), 359-381. finke, m. s., howe, j. s., & huston, s. j. (2017). old age and the decline in financial literacy. management science, 63(1), 213-230. https://doi.org/10.1287/mnsc.2015.2293 fisher, g. g., hassan, h., faul, j. d., rodgers, w. l., & weir, d. r. (2017). health and retirement study: imputation of cognitive functioning measures: 19922014 data description (final release version). survey research center, university of michigan. http://hrsonline.isr.umich.edu fritsch, t., mcclendon, m. j., wallendal, m. s., hyde, t. f., & larsen, j. d. (2014). prevalence and cognitive bases of subjective memory complaints in older adults: evidence from a community sample. journal of neurodegenerative diseases, 2014. frederick, s. (2005). cognitive reflection and decision making. journal of economic perspectives, 19(4), 25-42. gamble, k. j., boyle, p. a., yu, l., & bennett, d. a. (2015). aging and financial decision making. management science, 61(11), 2603–2610. https://doi.org/10.1287/mnsc.2014.2010 gan, l. (2013). financial market participation. in l. gan, z. yin, n. jia, s. xu, s. ma, & l. l. zheng (eds.), data you need to know about china: research report of china household finance survey (pp. 95–99). berlin: verlag. ghisletta, p., & spini, d. (2004). an introduction to generalized estimating equations and an application to assess selectivity effects in a longitudinal study on very old individuals. journal of educational and behavioral statistics, 29(4), 421-437. gorlick, a. (january 28, 2010). stanford report. stanford study shows older investors prone to mental misfires while playing the market. retrieved from https://news.stanford.edu/news/2010/jan uary25/older-investing-study012810.html hamza, c. a., ewing, l., heath, n. l., & goldstein, a. l. (2021). when social isolation is nothing new: a longitudinal study on psychological distress during covid-19 among university students with and without preexisting mental health concerns. canadian psychology/psychologie canadienne, 62(1), 20. han, y., he, y., lyu, j., yu, c., bian, m., & lee, l. (2020). aging in china: perspectives on public health. global health journal, 4(1), 11-17. harada, c. n., love, m. c. n., & triebel, k. l. (2013). normal cognitive aging. clinics in geriatric medicine, 29(4), 737-752. harmonized charls documentation. (2021). retrieved from https://charls.pku.edu.cn/en/team/codeb ook.pdf hastie, r. (2001). problems for judgment and decision making. annual review of psychology, 52(1), 653-683. huhmann, b. a., & mcquitty, s. (2009). a model of consumer financial numeracy. international journal of bank marketing, 27(4), 270–293. https://doi.org/10.1287/mnsc.2015.2293 https://doi.org/10.1287/mnsc.2014.2010 https://news.stanford.edu/news/2010/january25/older-investing-study-012810.html https://news.stanford.edu/news/2010/january25/older-investing-study-012810.html https://news.stanford.edu/news/2010/january25/older-investing-study-012810.html financial services review, 32(3) 18 https://doi.org/10.1108/0265232091096 8359 holden, s. t., & tilahun, m. (2022). are risk preferences explaining gender differences in investment behavior?. journal of behavioral and experimental economics, 101, 101949. jain, m & garg, a. (june 14, 2022). financial health-centric insurance for catalyzing customer wellbeing and business growth. impact capital for development. retrieved from https://www.uncdf.org/article/7780/fina ncial-health-centric-insurance-forcatalyzing-customer-wellbeing-andbusiness-growth kim, e. j., hanna, s. d., chatterjee, s., & lindamood, s. (2012). who among the elderly owns stocks? the role of cognitive ability and bequest motive. journal of family and economic issues, 33(3), 338-352. https://doi.org/10.1007/s10834-0129295-2 klapper, l., lusardi, a., & van oudheusden, p. (2015). financial literacy around the world. world bank. washington dc: world bank. https://media.hotnews.ro/media_server1/ document-2015-11-19-20605811-0finlit-paper-16-f2-singles.pdf klos, a., weber, e. u., & weber, m. (2005). investment decisions and time horizon: risk perception and risk behavior in repeated gambles. management science, 51(12), 1777-1790. korniotis, g. m., & kumar, a. (2010). cognitive abilities and financial decisions. in h. k., baker and nofsinger, j. r. (eds.), behavioral finance: investors, corporations, and markets (pp. 559576). john wiley & sons, inc. li, y., gao, j., enkavi, a. z., zaval, l., weber, e. u., & johnson, e. j. (2015). sound credit scores and financial decisions despite cognitive aging. proceedings of the national academy of sciences, 112(1), 65–69. https://doi.org/10.1073/pnas.141357011 2 liang, k. y., & zeger, s. l. (1986). longitudinal data analysis using generalized linear models. biometrika, 73(1), 13-22. liao, l., huang, n., & yao, r. (2010). family finances in urban china: evidence from a national survey. journal of family and economic issues, 31(3), 259–279. lu, x., zhang, y., zhang, y., & wang, l. (2020). can investment advisors promote rational investment? evidence from micro-data in china. economic modelling, 86, 251-263. mazzonna, f., & peracchi, f. (2018). selfassessed cognitive ability and financial networth: are people aware of their cognitive decline? eief working paper no. 1808. einaudi institute for economics and finance. http://www.eief.it/eief/images/wp_18.8. pdf ofstedal, m. b., mcclain, c. a., & couper, m. p. (2021). measuring cognition in a multi‐ mode context. advances in longitudinal survey methodology, 250-271. otuteye, e., & siddiquee, m. (2015). overcoming cognitive biases: a heuristic for making value investing decisions. journal of behavioral finance, 16(2), 140-149. pak, t. y., & babiarz, p. (2018). does cognitive aging affect portfolio choice?. journal of economic psychology, 66, 1-12. https://doi.org/10.1016/j.joep.2018.03.0 01 park, d. c., & reuter-lorenz, p. (2009). the adaptive brain: aging and neurocognitive scaffolding. annual review of psychology, 60, 173. rechner, p. l., & dalton, d. r. (1991). ceo duality and organizational performance: https://media.hotnews.ro/media_server1/document-2015-11-19-20605811-0-finlit-paper-16-f2-singles.pdf https://media.hotnews.ro/media_server1/document-2015-11-19-20605811-0-finlit-paper-16-f2-singles.pdf https://media.hotnews.ro/media_server1/document-2015-11-19-20605811-0-finlit-paper-16-f2-singles.pdf https://doi.org/10.1016/j.joep.2018.03.001 https://doi.org/10.1016/j.joep.2018.03.001 lei & fan 19 a longitudinal analysis. strategic management journal, 12(2), 155-160. reuter-lorenz, p. a., & park, d. c. (2014). how does it stac up? revisiting the scaffolding theory of aging and cognition. neuropsychology review, 24(3), 355370. richards, m., shipley, b., fuhrer, r., & wadsworth, m. e. (2004). cognitive ability in childhood and cognitive decline in mid-life: longitudinal birth cohort study. british medical journal, 328, 552–557. https://doi.org/10.1136/bmj.37972.5138 19.ee starnes, s. (september 30. 2019). protecting and guiding clients with diminishing mental capacity. retrieved from https://www.kitces.com/blog/stevestarnes-diminished-capacity-dementiasenior-safe-act-finra-rule-2165-4512trusted-contact-person/ salthouse, t. a. (1996). general and specific speed mediation of adult age differences in memory. the journals of gerontology series b: psychological sciences and social sciences, 51(1), 30–42. smith, t., & smith, b. (2006). proc genmod with gee to analyze correlated outcomes data using sas. san diego (ca): department of defense center for deployment health research, naval health research center. university of california san francisco. cf. healthy aging. retrieved from https://memory.ucsf.edu/symptoms/healt hy-aging. wang, y., & ruhe, g. (2007). the cognitive process of decision making. international journal of cognitive informatics and natural intelligence (ijcini), 1(2), 73-85. world health organization (2022). ageing and health in china. retrieved from https://www.who.int/china/healthtopics/ageing yu, b., steptoe, a., chen, y., & jia, x. (2021). social isolation, rather than loneliness, is associated with cognitive decline in older adults: the china health and retirement longitudinal study. psychological medicine, 51(14), 2414-2421. https://doi.org/10.1136/bmj.37972.513819.ee https://doi.org/10.1136/bmj.37972.513819.ee https://memory.ucsf.edu/symptoms/healthy-aging https://memory.ucsf.edu/symptoms/healthy-aging https://www.who.int/china/health-topics/ageing https://www.who.int/china/health-topics/ageing pii: s1057-0810(97)90013-5 financial services review, 6(3): 197-200 issn: 1057-0810 copyright 0 1998 by jai press inc. all rights of reproduction in any form reserved. the challenges and opportunities of student-managed investment funds at metropolitan universities douglas r. kahl the number of student-managed investmentfinds has grown rapidly in recent years. in the four decades since the first student-managed investment fimd was established at gannon university, the number of such&nds has grown at a rate of less than one per year to thirty-four in 1993. however, that rate of growth has changed dramatically in recent years. oak associates ltd., the akron, ohio based investment-management com pany, has funded ten student-managed investment finds since january 1996. the fund established by oak associates ltd. at the university of akron provides students with the opportunity to learn about investing real money on a real-time basis. the oak grant provides significant educational opportunities at the university and some real chal lenges in the organization and management of the finds. i. introduction a number of recent papers discuss the history, development, challenges and opportunities for student-managed investment funds at colleges and universities. lawrence (1994) pro vides an excellent discussion of the history and development of student-managed invest ment funds since the first such fund was established at gannon university in 1952 through the thirty-four student managed funds operating as of june 1993. at that time, the funds ranged in size from six thousand dollars to more than eight million dollars. johnson, alex ander and allen (1996) provide a thorough presentation of alternative decision making environments. block and french (1991) provide an excellent discussion of the process for starting and operating a student-managed investment fund. the student-managed investment fund provides some interesting challenges in its organization and operation. the university of akron, like many large metropolitan univer sities, has a diverse student body. non-traditional students are common, with the family, community and work commitments usually faced by these students. most of our students work at least part-time, and many of our students, especially at the graduate level, are employed full time and take classes in the evening. with classes scheduled from 7 am douglas r. kahl l professor of finance, the university of akron, akron, oh 443254803. 198 financial services review 6(3) 1998 until 11 pm, attending meetings of a student organization on a regular and consistent basis is problematic for many students and subject to changes in work or family commitments. to maximize the opportunity for student participation in the educational benefits of the grant, the fund is managed as a part of the portfolio management classes in the department of finance. ii. course structure the student-managed investment fund was established with an initial grant of $50,000 from oak associates ltd. an undergraduate portfolio management class manages the fund in the fall semester followed by an mba portfolio management class in the spring semester. all students are required to have a one-semester investments course as a pre-requisite to the portfolio management class. the students therefore begin the portfolio management course with a solid background in investments. the class is broken into teams of three to four stu dents. the first assignment is for each team to prepare and present an evaluation of the man agement of the fund by previous classes. each team then develops a major written proposal for the future investment of the fund and makes a formal presentation of their proposal to the rest of the class. the class, acting as an investment committee, then selects one team to implement their investment proposal for the semester. the major written proposal selected for implementation by the class provides a historical record of the investment philosophy, strategy, goals and logic of previous fund managers. developing and presenting this proposal for investing the oak grant funds constitutes about 50 percent of the course. twenty percent of the course grade depends on student peer evaluation of both the team’s proposal and the student’s activities as a team member. the oak grant gives the students practical experience in portfolio construction, management and evaluation by managing real money on a real-time basis. the nature of the oak grant requires that the students be completely responsible for security selection and portfolio construction. this encourages the development of the stu dents’ creativity and innovation in problem solving in a group setting and in a very com plex real world environment. the oak grant and the university of akron board of trustees resolution governing the use of the funds place only the most general limitations on the students’ management of the funds and restrict the instructor’s participation to a general teaching role in the course. iii. educational opportunities the objectives of the initial grant are to allow students to learn about investing on a real-time basis with real money and to have performance evaluated both on an absolute basis and relative to (currently ten) other institutions with similar grants. in addition, man aging the grant funds facilitates and encourages a number of other important educational outcomes and experiences. the students develop interpersonal skills such as teamwork, leadership, negotiation and an appreciation of cultural diversity through working in diverse teams to develop and present an investment proposal. teams frequently consist of both sexes, several races and student-managed investment funds 199 several countries of origin. members of the team often have very different and strongly held views on such topics as market efficiency, fundamental and technical analysis and international diversification. the experience of culturally diverse teams working through differences of viewpoint and belief to arrive at a unified proposal is an interesting and invaluable educational experience. the students’ oral and written communications skills are developed through both a minor project-the evaluation of previous management via a short written report and a short group presentation-and a major project, the investment proposal consisting of a sub stantial written report and group presentation of the proposal. this proposal will be evalu ated by both the students’ peers and instructor in determining the course grade, but only the students’ peers decide which proposal will actually be implemented. the students’ analytical skills are challenged and developed in several ways. the basic problem of constructing an investment portfolio in the real world is a very complex prob lem requiring extensive data analysis. the students must select from more than ten thou sand different stock, bond, money market and investment company investments for the funds. in the end, they will probably select about one out of every thousand potential investment vehicles. computer skills are developed through use of the s & p pc plus cd-rom database (which is networked through the computer lab) to obtain and screen historical financial data on potential investments and through use of the internet to obtain current information on a wide variety of subjects. internet applications range from obtain ing current stock quotes through on-line sources and current financial information from sources such as the security exchange commission’s edgar database to obtaining cur rent economic information from such sources as the st. louis federal reserve bank’s internet site. this project gives the students substantial experiential learning in the man agement of both the latest technology and the information obtained through that technol ogy. in the end, the results of the students’ analysis are applied in the real world and have real monetary consequences. the students’ global/international orientation and awareness are developed in several ways. by the very nature of the problem, international diversification of the investment portfolio is an integral part of the problem. this requires a knowledge and awareness of all of the major financial markets in the world. in addition, the students often have the chance to develop an increased global/international orientation by working in a team with students from another country or culture. the students will ultimately have the experience of eval uating proposals developed and presented by students from another country or culture. managing the oak grant funds allows integration across functional areas in the col lege of business. for approximately one week of the course, a member of the tax account ing faculty will teach about recent changes in tax law, which affect portfolio management. for a similar time period, a member of the marketing faculty will teach about selling finan cial services since the student’s investment proposal and presentation really constitutes an exercise in selling investment management skills to the rest of the class. a member of the business law faculty will provide an extensive exposure to legal and ethical standards and responsibilities in the course. the legal and ethical topics covered range from very general ethical standards to very specific legal and ethical standards applicable to securities man agement and trading. the topics are emphasized through extensive discussion of current real world examples such as the stock trading profits from possible insider trading by a local medical researcher. 200 financial services review 6(3) 1998 finally, the oak grants provide a unique form of competition among the ten colleges and universities that have currently received grants from oak associates ltd. the college with the highest returns on its investments each year receives an additional $15,000. the second and third highest receives an additional $10,000 or $5,000 respectively. this inter collegiate competition adds another layer of reality not found in most other student-man aged investment funds. the real world of investment management is a very competitive business. iv. conclusion a student-managed investment fund can provide students with experience managing real money on a real-time basis with real evaluation. in a properly structured course, a stu dent-managed investment fund can provide the framework for a wide range of other valuable educational experiences and outcomes. students can learn to work in culturally diverse teams to reach specific objectives. students can sharpen and find realistic application for their ana lytic and computer skills. students can learn to integrate knowledge and skills across aca demic boundaries. students can use their oral and written communication skills in attempting to market their investment management skills. competing with other teams and other col leges and universities in the investment performance competition will encourage an entre preneurial spirit that will be of great value to the students in their business careers. references block, s. b., & french, d. w. (1991). the student-managed investment fund: a special opportunity in learning. financial practice and education, i( i), 35-40. johnson, d. w., alexander, j. f., & allen, g. h. (1996). student-managed investment funds: a com parison of alternative decision-making environments. financial practice and education, 6(l), 97-101. lawrence, e. c. (1994). financial innovation: the case of student investment funds at united states universities. financial practice and education, 4(l), 47-53. pii: s1057-0810(97)90011-1 financial services review, 6(3): 169-183 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. performance and persistence in money market fund returns dale l. domian and william reichenstein we study the factors affecting the cross section of net returns of money market mutual finds from 1990 through 1994, and the persistence of relative returns across years. we find that the expense ratio is the most important factor in explaining differences in net returns. government-only funds produce slightly lower returns than other-taxable funds, and economies of scale are evident only for the smaller funds. money funds’ rel ative returns show strong persistence. most funds maintain stable expense ratios, so low-cost funds produce consistently high relative returns. i. introduction asset allocation models make recommendations about how an individual or family should allocate financial assets among stocks, bonds, and cash. many individuals invest through mutual funds. hence, it is important to evaluate fund performance. numerous studies examine stock mutual funds, beginning with jensen’s (1968) seminal paper on the perfor mance of the industry. in more recent years, hendricks, patel, and zeckhauser (1993), goetzmann and ibbotson (1994), malkiel(1995), and carhart (1997) investigate the per sistence of stock funds’ relative returns. a few studies, such as clements (1991), cornell and green (1991), and blake, elton, and gruber (1993) look at returns on bond funds. to the best of our knowledge, our paper is the first academic study to examine the factors affecting the cross section of net returns on money market mutual funds and the persistence of their relative returns across years. two factors explain 87 percent of the variance of five-year net returns across larger-than-$300 million money funds: the expense ratio and portfolio type. portfolio type distinguishes funds that can only invest in u.s. government securities (hereafter, go funds) from other-taxable (ot) money funds. not surprisingly, go funds produce slightly lower net returns. the expense ratio, however, is the dominant factor explaining dale l. domian l memorial university of newfoundland, faculty of business administration, st. john’s, newfoundland, alb 3x5 canada. william reichenstein l baylor university, department of finance, insurance, and real estate, p.o. box 98004, waco, tx 76798-8004. 170 financial services review 6(3) 1998 differences in net returns. we conclude that economies of scale exist until fund size reaches $300 million. for funds larger than $300 million, we conclude that all go funds are essentially commodities and all ot money funds are essentially commodities. that is, (1) they produce similar gross returns, (2) differences in net returns are driven almost exclusively by differences in expenses, and (3) expenses are a dead-weight loss to inves tors. in addition, we demonstrate that money funds’ relative returns show strong persis tence. most funds maintain stable expense ratios. a few funds have unstable expenses because management decides to temporarily absorb some or all expenses. funds with consistently low expense ratios produce consistently high relative returns. we conclude that individual investors can confidently predict relative returns on money funds. in fact, for 1990 through 1993 every go fund and every ot fund that produced a net return in the top five in one year produced a net return in the top quintile in the next year. ii. literature review money market funds made their debut in the early 1970s. industry growth in the early years was most pronounced when market interest rates rose above the regulation q limits on bank deposit interest rates. the development of the industry is examined by cook and duf field (1979a, b). one of the first research questions explored by academics was determining whether money fund managers exhibit the ability to predict interest rates. predictions can be inferred from changes in the average maturity of fund assets, since it would be profitable to shorten fund maturity prior to interest rate increases and to lengthen maturity prior to interest rate decreases. ferri and oberhelman (1981) present evidence that money fund managers were successful in making the appropriate maturity adjustments over the period 1975-1980. kane and marks (1987) demonstrate that significant gains would be obtain able from accurate forecasts. however, only a few funds in their study showed statisti cally significant forecast ability over the period 1978-1981, and the realized gains were small. other aspects of money market funds are considered by three studies from the early 1980s. rosen and katz (1983) look at the role of money funds in households’ asset alloca tion decisions. hubbard (1983) finds that fund deposits are not a substitute for money in traditional money demand models. lyon (1984) looks at the possibility of arbitrage profits due to money funds’ use of amortized cost valuation. recent studies have returned to the question of evaluating managers’ predictive abil ity. total money fund assets did not exceed $100 billion until 1982. so, tests using later data may show results that differ from studies based on earlier years when the industry was much smaller. domian (1992) concludes that fund managers showed no aggregate predic tive ability over the period 1982-1990, and in addition, maturity seemed to be adjusted solely in response to past interest rate changes. degennaro and domian (1996) develop models of fund managers’ maturity decisions, and note that transaction costs preclude managers from acting aggressively on their beliefs. pefformance and persistence 171 iii. the money fund industry and the data a) the industry money market mutual funds can be classified into three portfolio types and four fund types. the portfolio types are: (1) government-only (go) funds, which invest only in trea sury securities, agency securities, and repurchase agreements collateralized by treasury bills, (2) other-taxable (ot) funds, which invest in commercial paper, floating rate notes, u.s. government obligations and other assets, and (3) municipal money funds. in this study, we examine the returns on go and ot money funds. we do not study returns on municipal money funds because differences in their returns should reflect differences in tax codes across states. go and ot funds can be further separated into four fund types. general purpose funds (gp) and stockbroker-affiliated general purpose funds (sb) are open to any investor. spe cial purpose funds (sp) are only open to individuals who have a particular affiliation, such as customers of a certain bank. gp, sb, and sp funds are also called retail funds. institu tions-only funds (in) are designed for institutional investors-bank trust departments, pen sions, corporations, and the like. go funds are distinct from other taxable funds due to differences in tax status and, per haps, perceived credit risk. the interest that funds pass through from treasury bills and other direct u.s. government obligations is tax exempt at the state and local levels. interest income is not state-exempt on government agency debt or repurchase agreements collater alized with treasury bills. the portion of go income that is state-exempt varies from 100 percent to 0 percent, and averages about 38 percent. by comparison, the average ot fund holds 1 percent of assets in treasury bills. despite recent budget impasses, investors prob ably consider treasury and agency securities to be less risky than first-tier commercial paper. except for the go versus ot distinction, money funds have little ability to distinguish their portfolios from those of their competitors. in particular, regulations today largely pre clude distinctions based on maturity, default risk, and the use of derivatives. since june 199 1, the securities and exchange commission has restricted the average maturity to 90 days or less. ibc/donoghue does not provide annual data on average maturity. so, we can not rule out the possibility that average maturity influences average net returns among money funds. however, the fact that regulations restrict average maturity to a narrow range suggests that its influence on average return is, at most, minor. since june 1991, the sec has also restricted second-tier (i.e., less-than-highest grade) commercial paper at ot funds to 5 percent of the portfolio. in november 1995, only four funds held any second-tier paper (ibudonoghue’s money fund report, 1995) and only two, both relatively small, held more than 1 percent. moreover, in the few defaults to date the parent companies of the money funds absorbed the loss to avoid “breaking the buck’ that is, allowing net asset value to fall below $1 .oo per share. the 1995 edition of the zbc/ donoghue’s money fund report says, “no investor has ever lost money in a modem money fund.” we did not find data on the use of derivatives by money funds between 1990 and 1994. on at least three occasions between 1991 and 1993, the sec expressed in writing its con cerns about investments by some money funds in derivative securities, including inverse floaters and capped floating and variable rate instruments. in private conversations, a 172 financial services review 6(3) 1998 money manager told us he did not think such use was widespread. moreover, the orange county bankruptcy, which produced derivatives-related losses primarily at municipal money funds, was due to fraud. it was not due to money managers’ deliberate exposure to derivatives. finally, as with other defaults, most losses were absorbed by the fund’s parent company or adviser. consequently, we do not suspect fund managers’ deliberate decisions to use derivatives would explain much of the cross section of fund returns. since money funds have limited ability to compete based on differences in their port folios, they primarily compete based on differences in expenses. some funds, especially new ones, temporarily absorb all or a portion of expenses. they hope to establish a strong record of relative returns and thus attract investors before raising expenses. another method of reducing expenses is to impose a large minimal initial investment. it costs far less to service one $1 million account than five hundred $2,000 accounts. thus, the van guard federal portfolio fund has a 0.32 percent expense ratio and a $3,000 initial minimum investment, while vanguard’s u.s. treasury money market portfolio charges 0.15 percent but requires a $50,000 minimum initial investment. money funds can also compete by deciding whether or not they want to offer checking services. almost all retail funds offer limited checking. for example, they may allow unlimited check writing for checks of at least $500. of course, the cost of checking is reflected in the funds’ expenses. institutions-only funds also require large initial investments and have low expense ratios. the 1990 through 1995 issues of ibudonoghue’s money fund directory state, “[ins] are considered by many to be the most active and aggressive money funds, often outperforming general purpose funds.” we will examine whether ins outperform retail funds after adjusting for expenses or whether their net return advantage is due solely to their low expense ratios. b) data data for this study come from 1990 through 1995 issues of zbudonoghue’s money fund directory. the primary sample consists of all go and all other-taxable money funds with complete records of net returns and expense ratios for 1990 through 1994 and end-of-year net assets beginning 1989. the definitions of the variables used in the analysis are: nri, = expit = gurit = goi = ini = gp, = spi = sbi = net return for fund i for year t; the expense ratio for fund i for year t; nr,, + expit, grossed-up return for fund i for year t; a dummy variable denoting a government-only fund, go = 1 for a go fund and 0 otherwise; a dummy variable denoting an institutions-only fund, in = 1 for an in fund and 0 otherwise; a dummy variable denoting a general purpose fund, gp = 1 for a gp fund and 0 otherwise; a dummy variable denoting a special purpose fund, sp = 1 for a sp fund and 0 otherwise; a dummy variable denoting a stock-broker fund, sb = 1 for a sb fund and 0 otherwise; performance and persistence 173 table 1 average returns and expenses of money funds in 1994 avg. net return avg. expenses grossed-up total assets sample % % return (gur) % ogur (in $loom) all money funds all 3.70 0.65 4.35 0.13 3,436 go 3.67 0.61 4.28 0.11 1,065 ot 3.73 0.68 4.40 0.11 2,372 go, gp 3.54 0.73 4.27 0.12 354 go, sb 3.56 0.68 4.23 0.08 90 go, sp 3.70 0.58 4.28 0.13 47 go, in 3.93 0.37 4.30 0.10 574 ot, gp 3.64 0.75 4.39 0.12 1,138 ot, sb 3.68 0.72 4.40 0.07 591 ot, sp 3.73 0.7 1 4.43 0.12 173 ot, in 4.06 0.37 4.43 0.10 469 larger money funds (>$300 million) all 3.78 0.58 4.36 0.12 3,271 go 3.75 0.53 4.28 0.11 991 ot 3.80 0.61 4.41 0.09 2,280 go, gp 3.61 0.66 4.21 0.11 301 go, sb 3.56 0.68 4.23 0.09 90 go, sp 3.80 0.53 4.32 0.12 40 go, in 3.97 0.35 4.31 0.10 559 ot, gp 3.70 0.7 1 4.41 0.10 1,060 ot, sb 3.69 0.7 1 4.40 0.07 591 ot, sp 3.83 0.59 4.42 0.11 165 ot, in 4.06 0.37 4.43 0.09 464 obs 321 131 190 73 9 9 40 122 17 17 34 197 79 118 35 8 6 30 60 16 13 29 notes: the averages are equal-weighted across money funds. the portfolio types are government-only (go) funds, which invest only in u.s. government secunties. and other taxable (ot) funds, which invest in commercial paper. bank cds. government obligations and other assets. the fund types are general purpose (gp), stockbroker affiliated (sb). p 3 e&l purpose (sp). and institutions-only (in) funds. the last three columns denote the standard deviation of grossed-up return, total assets (in $100 million), and the number of funds or observations. cki = a dummy variable indicating that checking privileges exist on the fund, ck = 1 if checking privileges exist and 0 otherwise; and sizei, = the natural log of net assets (in millions) for fund i at the start of year t. table 1 presents the 1994 average net returns, average expense ratios, average grossed-up returns and standard deviation of grossed-up returns for samples of all funds and funds exceeding $300 million in net assets. the same patterns of returns and expenses occur in other years. as expected, other-taxable funds produce larger grossed-up returns than go funds. for example, in the larger-than-$300 million sample, ot funds produced a 13 basis point higher grossed-up return than go funds, which can be separated into an 8 basis point higher expense ratio and a 5 basis point higher net return. it is not clear from table 1 whether institutions-only funds produce higher gross returns than general purpose gp funds as suggested in several issues of zbudono ghue’s money fund directory. in funds produce larger net returns but this advantage may be due entirely to their lower expense ratios. for example, in the all-funds sample, go/in funds (i.e., institutions-only funds that hold only government obligations) enjoy 174 financial services review 6(3) 1998 a 39 basis point net return advantage compared to go/gp funds (3.93% vs. 3.54%), but 36 basis points are due to lower expenses and 3 basis points to higher grossed-up returns. in other such comparisons, in funds enjoy a 2 to 4 basis point grossed-up return advantage compared to gp funds. we cannot state with certainty whether the advantage is small but real or due to chance alone. we return to this issue in the regres sion analysis. finally, the sample may suffer from a survivorship bias. poorly performing stock funds typically merge into more successful funds, thereby burying the bad stock fund’s record. therefore, the record of the average surviving stock fund is an upward biased mea sure of the average fund’s performance. the evidence from this study suggests that manag ers have little ability to raise or lower the money fund’s gross return. so, survivorship bias is not likely to be a major problem. iv. methodology and results a) the determinants of net returns: economies of scale we begin the analysis by examining the role of net asset size in fund returns. table 2 presents regressions of net returns on expense, size, and go for 1990-1994 and each year. again, size denotes the natural log of beginning net assets in millions. table 2 determinants of net returns expense’ all money funds 1990-1994 -1.001 1990 -0.985 1991 -0.943 1992 -0.955 1993 -0.956 1994 -0.996 funds > $300 million 1990-1994 -1.015 1990 1.036 1991 -0.938 1992 -0.926 1993 -0.962 1994 -0.992 funds < $300 million s. e. size se. go s.e. r’ 0.028 0.020” 0.004 -0.131a 0.012 0.87 0.026 0.014a 0.005 -0.196” 0.012 0.87 0.05 1 0.035” 0.007 -0.197” 0.018 0.71 0.040 0.03 1” 0.007 4.078” 0.018 0.71 0.029 0.02 i a 0.004 -0.080” 0.014 0.82 0.033 0.0 1 5a 0.004 -0.122= 0.013 0.85 0.031 0.006 0.006 -0.13ga 0.014 0.87 0.030 0.007 0.006 -0.177= 0.015 0.88 0.055 0.022 0.012 -0.205” 0.022 0.67 0.041 0.007 0.010 -0.107” 0.021 0.67 0.030 0.005 0.008 -0.094” 0.015 0.82 0.032 0.005 0.007 -0.1 29a 0.015 0.83 1990-1994 -0.993 0.044 0.030” 0.009 4.127” 0.017 0.86 1990 -0.949 0.037 0.018 0.015 -0.216” 0.018 0.85 1991 -0.96 1 0.088 0.057” 0.017 -0.190” 0.03 1 0.68 1992 -1.016 0.069 0.060” 0.016 -0.034 0.032 0.70 1993 -0.970 0.048 0.049” 0.011 -o.066a 0.026 0.79 1994 -0.947 0.057 0.035a 0.011 a.1 18” 0.022 0.82 notes: a significantly different from zero at the 5 percent level. ’ the expense coefficient is indistinguishable from -i .o in every regression. performance and persistence 175 for the sample of all funds, the size coefficient is always positive and significant at the 5 percent level, the level of significance used throughout this study. however, separate analysis implies that positive economies of scale only exist until fund size reaches about $300 million. among funds larger than $300 million, the size coefficient is never signif icant. among funds smaller than $300 million, size is significant in the 1990-1994 regression and all but one of the individual year regressions. moreover, separate analyses of go funds and ot funds are consistent with the story that positive economies of scale exist until the fund reaches $300 million in net assets. we do not want to imply that $300 million is the only acceptable point of separation. we considered regressions that separate small and large funds at $100 million increments through $500 million. economies of scale clearly exist below $300 million and clearly do not exist above $500 million. in the 1990-1994 regression for funds exceeding $300 mil lion, the size coefficient of 0.006 (with a t-statistic of 1 .o) suggests that weak economies of scale may continue to exist among larger funds. however, the size coefficient falls to 0.001 when the sample is limited to funds above $500 million. in summary, we conclude that positive economies of scale exist in the management of money funds until net assets reach $300 million, but they do not exist in the management of larger money funds. in contrast, ciccotello and grant (1996) conclude that a larger fund size tends to reduce the performance of aggressive growth equity funds. in the next section, we examine whether the commodity view can explain the cross-sectional variance of net returns among larger money funds. one of the predictions of this view is that net returns are driven almost exclusively by expense ratios. thus we want to exclude the smaller funds whose returns may be affected by size. the larger funds contain over 95 percent of money fund assets. so, they represent the bulk of the industry. b) the commodity view the financial press and academicians often view all money market funds of a given portfolio type (i.e., go or ot) as indistinguishable commodities. for example, bogle (1994) says “gross returns are virtually identical within each type of money fund and...net returns are therefore driven almost exclusively by fund costs” (p. 131). we label this the commodity view of money market funds. the driving principle behind the commodity view is that money markets are extremely efficient. therefore, financial markets set gross mutual fund returns; managers cannot enhance gross returns. furthermore, cross-sectional differences in funds’ net returns are driven almost exclusively by differences in their expense ratios and portfolio types. the strong version of the commodity view makes the following statistical predictions. in cross-sectional regressions of net returns on predictor variables, 1. go, the dummy variable denoting a government-only fund, should be negative; 2. the expense coefficient should be -1 .o; 3. the coefficient of determination should be large; 4. no variable besides the expense ratio and go should prove significant; and 5. the grossed-up return should be essentially the same for all funds of the same portfolio type. statistically, the standard deviation of the cross-sectional distribu tion of grossed-up returns should be small. 176 financial services review 6(3) 1998 year table 3 tests of the commodity view of money funds expense (s. e.) go (se.) other coef (s.e.) r2 199&1994 1990 1991 1992 1993 1994 -1.019 (0.03 1) -0.139a (0.014) -1.012 (0.033) 4). 140” (0.014) -1.013 (0.03 1) -0. 140a (0.014) -1.022 (0.030) -0.137a (0.014) -0.986 (0.039) -0.144” (0.014) -1.028 (0.037) -0. 139a (0.014) -1.042 (0.030) 4k179a (0.016) -1.031 (0.037) -0.1 80a (0.016) -1.035 (0.030) -0.180a (0.016) -1.042 (0.030) -0.179= (0.016) -0.984 (0.044) -0.1 88a (0.016) -1.043 (0.04 1) -0.1 79a (0.016) -0.956 (0.055) -0.209a -0.982 (0.058) -0.210= -0.943 (0.056) -0.209a -0.956 (0.054) -0.209a -0.950 (0.063) a210a -0.990 (0.059) -0.209a -0.934 (0.040) -0.109a -0.927 (0.041) -0.108a -0.927 (0.04 1) -0.109a -0.935 (0.040) -0. 108a -0.900b (0.047) -o.lloa -0.956 (0.046) -0.109a -0.969 (0.029) -0.095a -0.964 (0.034) -0.095= -0.966 (0.029) -0.095a -0.970 (0.029) -0.095a -0.952 (0.04 1) -0.096a -0.985 (0.039) -0.09y -0.996 (0.033) -0.130a -0.967 (0.038) -0.1 3oa -0.986 (0.033) -0.130” -0.995 (0.033) -0. 129a -0.919 (0.044) -0.132a -0.969 (0.042) -0.130= (0.022) (0.022) (0.022) (0.022) (0.022) (0.022) (0.021) (0.021) (0.021) (0.021) (0.021) (0.021) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) (0.015) gp sb sp in ck gp sb sp in ck gp sb sp in ck gp sb sp in ck gp sb sp in ck gp sb sp in ck 0.87 -0.008 (0.015) 0.86 -0.027 (0.017) 0.87 0.024 (0.027) 0.87 0.027 (0.018) 0.87 0.007 (0.016) 0.86 0.88 -0.013 (0.016) 0.88 -0.032 (0.019) 0.88 0.003 (0.023) 0.88 0.050a (0.021) 0.89 0.001 (0.018) 0.88 0.67 0.028 (0.023) 0.67 -0.062a (0.029) 0.67 0.005 (0.046) 0.66 0.005 (0.026) 0.66 0.031 (0.024) 0.67 0.67 -0.008 (0.022) 0.67 -0.029 (0.034) 0.67 0.013 (0.037) 0.67 0.028 (0.024) 0.67 0.021 (0.024) 0.67 0.82 -0.005 (0.017) 0.82 -0.014 (0.018) 0.82 0.009 (0.022) 0.82 0.014 (0.019) 0.82 0.014 (0.019) 0.82 0.83 -0.027 (0.016) 0.83 -0.032 (0.017) 0.83 0.018 (0.026) 0.83 0.058a (0.021) 0.84 -0.022 (0.019) 0.83 notes: this table shows regressions of net returns on expense, go, and, sometimes, another variable for 199t 1994 and each year. the sample consists of all funds with $300 million of net assets at the beginning of the period. the independent variables are the expense ratio (expense) and dummy variables denoting a govem mat-only fund (go), a general purpose fund (gp), a special purpose fund (sp). stockbroker fund (se), an institutions-only fund (in), and a fund offering checking privileges (ck). there are 167 observations for 1990 and 199& 1994, i98 for i99 i, 203 for 1992, i98 for 1993, and i97 for 1994. a significantly different from zero at the 5 percent level. b significantly different from -1.0 at the 5 percent level. 4a. the weak version of the commodity view retains predictions 1, 2, 3, and 5, but replaces 4 with, adding another variable to the regression of net returns on expense and go will not appreciably raise the adjusted coefficient of determination. perj%ormance and persistence 177 table 3 presents, for the larger-than-$300 million sample, a summary of regressions examining the determinants of net returns for 1990-1994 and each year. as predicted by both forms of the commodity view, the go coefficient is negative and significant in each regression. also as predicted, the expense ratio’s coefficient is indistinguishable from -1 .o in 35 of 36 regressions. these results imply that expenses are a dead-weight loss to investors. the third prediction says the adjusted coefficient of determination (henceforth, coeffi cient of determination or r2) should be “large”. in regressions on go and expense, it ranges from 0.67 to 0.88 for individual years and is 0.87 for 199c1994. are these values “large”? we believe they are. by comparison, malkiel (1995) finds that the average expense ratio can explain only 5 1 percent of the variance of lo-year average net returns on a sample of stock funds. moreover, when malkiel excludes two outliers and funds with expenses greater than 2.5 percent, the average expense ratio explains only 5 percent of the variance of io-year returns. in contrast, the expense ratio explains at least 67 percent of the variance of one-year net returns and 87 percent of the variance offive-year net returns on money funds. figure 1 graphs 1994 net returns and expense ratios for ot money funds; graphs for go funds in 1994 and for go and ot funds in other years are essentially the same. we believe this picture supports the claim that differences in net returns are largely attributable to differences in expense ratios. for example, the lowest net return among funds with expenses of 0.50 percent or below exceeds the highest net return among funds with expenses above 0.70 percent. 4.50 4.00 / 3.25 i i i 3.00 i 0.00 0.25 0.50 0.75 1.00 1.25 1.50 expense ratio (%) figure 1. 1994 returns and expenses 118 ot (other-taxable) funds with assets above $300 million 178 financial services review 6(3) 1998 the strong version of the commodity view says, within a portfolio type, net returns are driven exclusively by fund costs. the weak form says net returns are driven almost exclu sively by costs. in the regressions, the strong form receives mixed support and the weak form receives strong support. the five-year regressions support the strong version. in particular, the variables denot ing fund type or the existence of checking-i.e., gf, sb, sp, in, and ck-prove insignif icant. however, in individual year regressions, the in slope is positive in all years and significant in 1990 and 1994. so, we cannot rule out the claim that the average in fund beats its competitors by perhaps 3 basis points after adjusting for expenses. to repeat, even this ciaim receives mixed support. we find no evidence supporting the supposition made annually in the zbudonoghue’s money fund directory that aggressive management allows managers to substantially raise the gross returns at in funds. clearly, in funds’ net return advantage is due mostly, if not entirely, to their lower expenses. the weak form of the commodity view receives consistent support. adding gp, sb, sp, in, or ck raises the r* by at most 0.01. so, within a portfolio type, net returns are driven almost exclusively by fund expenses. the commodity view says that gross returns are essentially similar across funds with the same portfolio type. table 1 provides support for this position. among the larger money funds (in the bottom half of the table), the standard deviation of grossed-up returns for go funds is 0.11 percent. the standard deviation for ot funds is 0.09 percent. so, about two-thirds of go funds will earn a gross return within 0.11 percentage points of the average for its portfolio type, and 95 percent will earn within 0.22 percentage points of the average. by contrast, differences in funds’ expense ratios often exceed 0.60 percentage points. it is not surprising, therefore, that investors can reliably predict relative fund returns based on prior knowledge of their expense ratios. the next section supports this position. c) can we predict a winning fund? in the last section, we saw that the cross section of funds’ one-year expense ratios explains at least 67 percent of the cross section of one-year net returns. with few excep tions, a fund’s expense ratio is stable from year to year. consequently, winning money market funds tend to repeat. the spearman rank correlations between net return rankings in year t and t + 1 average 0.89 for go funds and 0.90 for ot funds (t = 1990 through 1993). they imply that rank ings of funds by net return change little from one year to the next. another method of testing persistence in returns is to calculate the percent of winning funds that repeat as winners, where a winning fund is defined as any fund that earns a net return above the median. table 4 presents the number of winning money funds that repeat in later years. in 1990,65 go funds had net returns above the median. of the 65,56 funds produced winning returns in 1991. moreover, 49 of the 56 funds repeated as winners in 1992,45 of the 49 repeated in 1993, and 44 repeated in 1994. on average, 87 percent of go winners repeated the next year, 77 percent repeated the next two years, 71 percent repeated the next three years, and 68 percent repeated the next four years. the results were almost identical for ot funds. clearly, relative returns show persistence. since relative returns can be easily predicted, individual investors should not settle for a slightly above median return. they should try to pick a top-returning fund. table 5 shows how the top five performers in one year ranked the next year. consider the go funds first. performance and persistence 179 year t table 4 winning money funds that repeat f t+l t+2 r+3 t+4 government-only (go) funds 1990 65 1991 63 1992 63 1993 65 other-taxable (ot) funds 1990 94 1991 94 1992 94 1993 95 56 55 54 58 87% 76 82 86 80 86% 49 45 44 49 46 50 77% 71% 68% 70 67 60 77 67 73 78% 71% 65% notes: a winning fund is one with an above-median return. among go funds, 65 produced winning net returns in 1990. of the 65,56 repeated as winners in 19??1,49 repeated as winners in 1991 and 1992.45 repeated in 1991 through 1993, and 44 repeated in 1991 through 1994. in some initial years (i. e., year i), several funds earned the median return, so less than half the funds were winners. of the top five funds, every fund finished in the top quintile the next year, 85 percent fin ished in the top decile, and 75 percent finished in the top 10. for the ot funds, every top-five fund finished in the top quintile the next year, 86 percent finished in the top decile, and 64 percent finished in the top 10. past performance allows individuals to be virtually guaranteed of choosing a top quintile fund and almost certain of choosing a top decile fund, especially if they exclude funds that have temporarily lowered expenses. table 6 presents the simulated returns from strategies of buying the money funds with the best prior-year net returns. the returns assume that on january 1st of 1991 through table 5 do the elite repeat?: next-year rankings of the top-five money funds rank in: 1990 1991 1991 i992 1992 1993 1993 1994 government-only funds 1 2 3 4 5 total funds = 13 1 other-taxable funds it 1t 3 4 5t 5t total funds = 190 8 3 7 11 2 4 1 2 3 2 3 19t 3 6 14 4 19 10 5t 17t 11 5t 22t 22 2 1 4 18 2 4t 3 13t 20 1 6t 1 2 8t 2 3 37t 8t 4t 1 16t 4t 11 1 14 2 2 3 1 4t 5t 4t 5t note: “t” denotes tie. 180 financial services review 6(3) 1998 table 6 simulated average returns from buying money funds with best single-year returns 1991-1994 simulated return annual alpha government-only funds buy top fund prior year buy top 5 funds prior year buy and hold top fund in 1990 buy and hold top 5 funds in 1990 other-taxable funds buy top fund prior year” buy top 5 funds prior year buy and hold top fund in 1990a buy and hold top fund in 1990” buy and hold top 5 funds in 1990 4.19% 4.26% 3.94% 4.21% 4.42% 4.39% 4.09% 4.41% 4.33% 0.22% 0.29% -0.03% 0.24% 0.40% 0.37% 0.07% 0.39% 0.31% nores: alpha denotes the additional average annual return on the simulated portfolios compared to the (asset-weighted) average return on money funds of the same portfolio type. atwo funds tied for the top return among other-taxable funds in 1990. for “buy top fund prior year,” we assume an equal investment in both funds for 1991. 1994 an individual ranks all go and all ot money funds based on their net returns over the preceding year. we examine four simulated strategies. in the first, he buys the top ranked fund in the prior year and adjusts each january 1st. in the second, he buys an equal-weighted portfolio of the top-five funds and adjusts each january 1st. in the last two strategies, he either buys and holds the top ranked fund in 1990 or buys and holds an .equal-weighted portfolio of the top five funds in 1990. the strategies’ alphas indicate the additional average return compared to the investor who earned the (asset-weighted) aver age return among funds with the same portfolio type. the strategy of buying the top go fund from the prior year produced a 4.19 percent average return, or 0.22 percent higher than the return received by the average investor. buying an equal-weighted portfolio of the top-five funds produced an average return of 4.26 percent and an alpha of 0.29 percent. buying and holding the top go fund in 1990 produced a -0.03 percent alpha. manage ment of the benham government agency fund, the top fund in 1990, temporarily absorbed a portion of the expenses. as the expense ratio was raised from 0.28 percent in 1990 to 0.45 percent in 1991 and 0.50 percent in 1992, the fund’s ranking fell. in contrast, the second through fifth ranked funds in 1990 maintained consistently low expense ratios (0.20 percent or lower) and consistently high rankings (never worse than 14th). so, the strategy of buying and holding the top five funds produced a positive alpha of 0.24 percent. this discussion highlights the need to beware of high rankings based on temporarily reduced expenses. money funds, especially new ones, sometimes waive or partially waive expenses in their marketing campaigns to attract investors. for ot funds, the results for the four simulated strategies are similar. the strategies of buying the top fund and the top five funds from the prior year produced alphas of 0.40 per cent and 0.37 percent. two funds tied for the top honors in 1990. one produced mediocre returns and the other produced good returns for the next four years. the dreyfus world wide dollar money market fund produced an alpha of 0.07 percent. it raised its expense performance and persistence 181 table 7 simulated average returns from buying money funds with best multi-year returns 1992-1994 simulated return annual alpha government-only funds buy top 5 funds in 1990-1991 3.61% 0.25% buy fop 10 funds in 1990-1991 3.60% 0.24% other taxable funds buy top 5 funds in 1990-1991 3.67% 0.29% buy top 10 funds in 1990-1991 3.71% 0.33% note: alpha denotes the additional average annual return on the simulated portfolios com pared to the (asset-weighted) average return on money funds of the same ponfolio type. ratio in steps from 0.20 percent in 1990 to 0.77 percent in 1994 and its ranking fell from tied for first to 1 62nd. the expenses of the alger money market portfolio ranged from 0.06 percent in 1990 to 0.41 percent in 1993 when it ranked 22nd, its lowest ranking for the five years. it produced an alpha of 0.39 percent. the third through fifth ranked funds in 1990 maintained consistently low expense ratios (0.20 percent or lower) and consistently high rankings (tied for 19’h or better). the alphas from low-cost money funds are relatively small, but they are predictable and entail no additional risk. selecting a low-cost money fund, one that is not temporarily absorbing expenses, may produce a 0.2 percent to 0.4 percent return advantage. this advantage is large when expressed as a percent of interest income. the interest income from a fund yielding 4 percent is 8.1 percent larger than the interest income from a fund yielding 3.7 percent. a low-cost strategy reliably produces a positive alpha. the evidence produced thus far reveals that money funds’ one-year relative returns can be predicted. table 7 shows that their multi-year relative returns can also be predicted. the top-five go funds based on 1990-1991 returns earned, on average, 0.25 percent more per year than the (asset weighted) average go fund in 1992-1994. the top-five ot funds earned 0.29 percent more than the average ot fund. the results to this point indicate that one-year and multi-year relative returns are highly predictable. to better appreciate this conclusion, let us contrast it with the conclu sion from similar studies on stock mutual funds. rankings of one-year stock returns are considered virtually useless. for example, bogle (1994) picks the top 20 broad-based stock funds in each year from 1982 through 1991 and observes their average rank in the next year. it is 284 out of 681. in another study, malkiel (1995) examines the reliability of one-year stock-fund rankings and concludes, “all in all, the simulation results do not give one confidence that investors can consistently beat the market by buying [last year’s win ning stock] funds.” in addition, malkiel finds that even long-horizon return records are of little benefit when choosing a stock fund. the 20 funds with the highest returns from 1970 through 1980 produce an average rank of 118 out of 260 in the next ten years. we end this section with some advice for individual investors on determining the best money market funds. table 8 lists larger-than-$300 million retail funds with the highest returns in 1994, the final year of our study. for more current information, investors can look at weekly listings of 30-day yields-that is, net returns-in barron ‘s. 182 financial services review 6(3) 1998 table 8 top-ten money funds for 1994 fund name net return (96) expense ratio (%) fund type government-only funds vanguard mmr, federal portfolio 4.03 0.32 gp kemper mmf, government securities portfolio 3.97 0.47 gp fidelity mmt, retirement government mmp 3.94 0.42 sp hanover government mmf 3.87 0.60 sp fidelity u.s. government reserves 3.85 0.50 gp the one group u.s. treasury securities mmf 3.85 0.60 sp pacific american fund, u.s. treasury portfolio 3.84 0.43 sp ust master government money fund 3.83 0.50 gp norwest u.s. government fund 3.82 0.50 gp the rodney square fund, u.s. government portfolio 3.82 0.53 gp other taxable funds fidelity spartan mmf the one group prime mmnfiduciary vanguard mmr, prime portfolio fidelity mmt, retirement mmp westcore mmf usaa mmf kemper mmf, money market portfolio strong mmf fidelity cash reserves victory financial reserves portfolio 4.14 0.45 gp 4.09 0.58 sp 4.08 0.32 gp 4.08 0.42 sp 4.06 0.30 gp 4.06 0.46 gp 3.99 0.52 gp 3.99 0.70 gp 3.96 0.58 gp 3.96 0.57 sp v. summary and conclusions we studied the determinants of money market fund returns from 1990 through 1994 and the persistence of funds’ relative returns across years. among money funds larger than $300 million, two factors explain the cross-sectional variance of net returns: the expense ratio and a variable distinguishing u.s. government-only (go) funds from other-taxable (ot) money funds. go funds produced lower net returns than ot funds, everything else the same. the fund expense ratio is the dominant factor explaining the variance of net returns. in fact, the evidence supports the commodity view of money funds, which says, after adjusting for portfolio type (i.e., go or ot), net returns are driven almost exclu sively by fund expenses. we could find no other factor that had a statistically signifi cantly impact on 1990-1994 average net returns. among larger-than-$300 million funds of a given portfolio type, net returns are driven exclusively or almost exclusively by expenses. money funds’ relative returns show strong persistence. most funds maintain stable expense ratios. so, funds with low costs or high returns in one year usually produce strong relative returns the next year. of the five go funds with the highest returns in year t (t = 1990 through 1993), 85 percent produced top decile returns the next year. of the top-five ot funds in one year, 86 percent produced top decile returns the next year. every top-five go and ot fund finished in the top quintile the next year. we conclude that an individual or family can easily pick a top-performing money fund, especially if one excludes funds that have temporarily lowered expenses. performance and persistence 183 acknowledgment: we thank fritz curtis of amr investment services for valuable comments and insights into the money fund industry, burton g. malkiel for valuable com ments, and ibudonoghue’s for supplying the data. references blake, c. r., elton, e. j., & gruber, m. j. (1993). the performance of bond mutual funds. journal of business, 66, 37 l-403. bogle, j. c. (1994). bogle on mutualfinds: new perspectives for the intelligent investor. burr ridge, il: irwin. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52,57-82. ciccotello, c. s., & grant, c. t. (1996). equity fund size and growth: implications for performance and selection. financial services review, 5, 1-12. clements, j. (1991, april 4). in picking bond fund, expense factor remains the key. wall street jour nal, cl. cook, t. q., & duffield, j. g. (1979a). average costs of money market mutual funds. federal reserve bank of richmond economic review, 32-39. cook, t. q., & duffield, j. g. (1979b). money market mutual funds: a reaction to government reg ulations or a lasting financial innovation? federal reserve bank of richmond economic review, 15-3 1. cornell, b., & green, k. (1991). the investment performance of low-grade bond funds. journal of finance, 46, 29-48. degennaro, r. p., & domian, d. l. (1996). market efficiency and money market fund portfolio man agers: beliefs versus reality. financiae review, 31,453474. domian, d. l. (1992). money market mutual fund maturity and interest rates. journal of money, credit and banking, 24,519-527. ferri, m. g., & oberhelman, h. d. (1981). a study of the management of money market mutual funds: 19751980. financial management, 10,24-29. goetzmann, w. n., & ibbotson, r. (1994). do winners repeat? patterns in mutual fund behavior. journal of porrfolio management, 20,9-l 8. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974-1988. journal offinance, 48,93-130. hubbard, c. m. (1983). money market funds, money supply, and monetary control: a note. journal offinance, 38, 1305-1310. ibcldonoghue’s money fund directory. (1990). ibudonoghue’s money fund directory. (1991). ibudonoghue’s money fund directory. (1992). ibudonoghue’s money fund directory. (1993). ibcldonoghue’s money fund directory. (1994). ibudonoghue’s money fund directory. (1995). ibc’s money fund report. (1995, november 24). jensen, m. c. (1968). the performance of mutual funds in the period 1945-1964. journal offinance, 50,549-57 1. kane, a., & marks, s. g. (1987). the rocking horse analyst. journal ofportfolio management, 13, 32-37. lyon, a. b. (1984). money market funds and shareholder dilution. journal of finance, 39, 101 l 1020. malkiel, b. g. (1995). returns from investing in equity mutual funds 1971 to 1991. journal of finance, 50,549-57 1. rosen, k. t., & katz, l. (1983). money market mutual funds: an experiment in ad hoc deregulation: a note. journal of finance, 38, 101 l-1017. pii: s1057-0810(96)90004-9 financial services review, 5(2): 101-117 copyright 0 1996 by jai press inc. issn: 1057-08 10 all rights of reproduction in any form reserved. a simulation approach to the choice between fixed and adjustable rate mortgages william k. templeton, robert s. main, and j. b. orris this study uses a simulation approach to model the choice between a fixed rate mort gage (frm} and an adjustable rate mortgage (arm). our simulations help assess the risks and benefits of choosing an arm rather than a frm. we represent the risk of the arm with distributions of present value cost differentials for a variety of mortgage life periods. we provide insight on thejinancial planning aspect by modeling the impact of mortgage rate changes on the size of payments for arms. the simulations yield non intuitive results that may lead to better decision making by borrowers. mortgage borrowers appear to have a difficult time evaluating the costs and risks associ ated with the choice between a fixed rate mortgage (frm) and an adjustable rate mortgage (arm). anecdotal evidence suggests they often consider only the worst case interest rate scenario, misunderstand the effect of the time value of money and the expected life of the mortgage in making this choice, and often worry more about the uncertain prospect of large monthly payments than the effective costs of the competing mortgages. this study offers an approach by which borrowers may more effectively evaluate the arm-frm choice. it provides a simulation model that yields information on a number of cost and risk factors of demonstrated importance to borrowers. the simulation output allows a borrower to view probability distributions of present value cost differentials between the arm and the frm for a variety of mortgage life periods. it also provides a distribution of the breakeven period, the number of years for which the arm maintains its present value cost advantage from the initiation of the loan. finally, the simulation permits insight into the financial planning aspect of the choice by modeling the impact of mortgage rate changes on the size of payments for the arm. a borrower may then compare this uncertain payment to the frm payment. william k. templeton, robert s. main, and j. b. orris l butler university, 4600 sunset avenue, indianapolis, in 46208. 102 financial services review s(2) 1996 i. literature review thus far, the academic literature largely has focused on discovering which variables signif icantly influence actual borrower choice (e.g., see brueckner & follain, 1988; dhillon, shilling, dz sirmans, 1987; goldberg & heuson, 1992; o’brien & wong, 1990; phillips & vanderhoff, 1991; tucker, 1989). these studies generally show that pricing variables are the most important determinants of choice. these factors include the initial mortgage rate differential between arms and frms and discount point differenti~s. normally, an arm offers a lower initial mortgage rate compared to a frm. arms entail a higher degree of risk with regard to cost, and introduce uncertainty into the financial planning process. bor rowers ought to require a lower initial rate as compensation. thus, the proportion of bor rowers choosing an arm has been found to be positively related to the size of the initial rate differential. five of the six studies cited above have found that the rate level of the frm is signif icantly positively associated with the probability of choosing an arm, celerisparibus. the explanation for this may be that borrowers take advantage of a lower initial rate on the arm in order to qualify for a larger mortgage. an alternative explanation is that when frm rates are high, borrowers assume that rates will decline and that the arm will be the lower cost alternative. at lower frm rates borrowers may prefer the lower risk of the frm, or they may assume that rates will increase which would cause the arm to be a higher cost alternative. table 1 contains historical data on average contract interest rates for both frms and arms. it also shows the percent of loans of each variety. table 1 con firms that higher frm rates generally are associated with a higher percentage of arm loans. sprecher and willman (1993) have used historical data to determine if borrowers could have achieved an effectively lower mortgage rate using an arm during a period (1987 to 1991) in which rates began relatively low and then increased. they collected a sample of arm and frm loans initiated at the beginning of this period and compared the table 1 annual averages for mortgage interest rates for existing homes, 1982-1994. excludes refinancing loans and federally underwritten loans year contract interest rate (percent): percent of number of loans with: fixed rate a&stable rate fixed rate adjustable rate 1982 14.8 14.7 61 39 1983 12.6 11.9 59 41 1984 12.1 11.6 36 64 198.5 11.9 10.5 50 50 1986 10.1 9.1 69 31 1987 9.5 8.2 56 44 1988 10.1 8.2 76 24 1989 10.2 9.2 63 31 1990 10.4 9.2 73 21 1991 9.7 8.2 78 22 1992 8.5 6.5 79 21 1993 7.5 5.7 so 20 1994 8.2 6.4 61 39 source: statistical abstract of the united states, various years. fined and adjustable rate mortgages 103 effective rates of the arms for this period to the rates on the frms. they found the arm borrowers paid an effective rate that was 150 basis points lower than the frm borrowers paid. in the present study, both of these issues surface. in the first reported simulation, rates were relatively low, but the arm offered a large initial rate advantage. in the second simulation, rates were relatively high and the initial rate advantage of the arm was much narrower. a survey conducted by lino (1992) looked into borrower perception of mortgage costs and risks. he found that borrowers did not properly incorporate expected cost into their decision making. he surveyed recent mortgage borrowers and found that many were unable to correctly determine which type of mortgage had the higher expected present value cost based on the borrowers’ own expectations for interest rate changes and the life of the mort gage. for example, borrowers who expected interest rates to fall and who expected to reside in their homes for less than 8 years should have viewed the arm as the lower cost alterna tive. lino found that 30 percent of these actual borrowers thought the relative cost of frms and arms was the same under these conditions. those borrowers who expected rates to increase and who planned a residency of more than 12 years should have viewed the frm as the lower cost alternative. lino found 10 percent perceived the frm as the higher cost alternative and another 19 percent thought there was no difference in cost under these expectations. thus, the ability to predict interest rate movements and length of residency is often beside the point for borrowers. a substantial minority are unable to select the low-er cost alternative even given these factors. lino also has shown that the financial planning aspect of the mortgage payment commitment and the perceived risk of arms are quite important to borrowers. indeed, these factors were significant in a logit analysis predicting mortgage choice, while cost variables were not significant. borrowers may be well served by a method offering insight into the risk and planning aspects of arms. yohannes (1991) and tucker (1991) have offered frameworks for helping individuals better make the arm-frm choice. yohannes (1991) has suggested using interest rate changes implied by forward rates to construct an expected breakeven number of years before the present value cost of an arm exceeds that of an frm. this expected breakeven period, along with a worst case scenario breakeven period, offers some guidance in making the choice. this approach is limited in that it offers just the two cases and concerns itself only with the present value cost breakeven period. tucker (199 1) has shown that simulation can reveal important information related to the expected costs of mortgages to borrowers. he has simulated interest rate changes to demonstrate the present value cost differential for arms and frms assuming a variety of opportunity discount rates. his analysis has shown that, based on loan data from a connect icut savings and loan institution and market conditions in 1985 to 1989, arms were often the lower cost alternative. this result held particularly true for borrowers who anticipated a shorter life for their mortgage and had higher oppo~unity cost discount rates. borrowers with a 4 percent discount rate (the lowest rate simulated in tucker’s study) would have found the arm to be the lower expected cost alternative if the mortgage life had been 18 years or less. borrowers with an 8 percent discount rate would have found that the arm dominated the frm in terms of expected present value cost for any mortgage life up to 30 years. we know, however, that borrowers are concerned with factors in addition to expected present value cost. both risk and financial planning considerations enter into the decision. 104 financial services review 5(2) 1996 the present study extends the simulation approach to making the arm-frm choice by including additional factors important to borrowers. in this study, we compare a thirty year annually adjusted arm to a thirty year frm. the simulation results provide informa tion on the present value cost differentials, the breakeven period, and the payment size. they provide both expected values and distributions of these values that would allow bor rowers to consider both cost and risk and make a more informed choice. even with a more complete understanding of cost and risk, borrower choice would likely depend on individ ual risk preferences. ii. ~tho~ology the present value cost of a mortgage is the sum of the discount points, the present value of the payments, and the present value of the payoff balance. the discount points and the interest portion of the payment are taken on an after-tax basis. mathematically, the present value cost can be expressed as i t epmt, ft. (q)l pvcost=m~(p)~(l-txj)+ t] 1 b (1 + ‘j +(1) t= 1 (1 + qt where m p 4 t pm4 4 and = the mortgage amount = the discount points rate = the personal tax rate of the borrower i = the assumed life of the mortgage in years = the mortgage payment at year t = the interest portion of the payment in dollars at year t = the oppo~unity cost discount rate for borrower i = the payoff balance at year t. a model was constructed in excel spreadsheet software to compare the present value costs of an arm and frm over a thirty year term using equation 1. the model was simu lated using the monte carlo technique in @risk simulation software. table 2 shows a sample input section of the model. a large, midwestem thrift institution provided loan terms data for the period 1989 through 1992. these items include discount points and the initial mortgage rates for both adjustable and fixed rate mortgages, as well as the selection of an index and margins over the index rate for the arm. thus, the simulations in this study were performed using real loan terms as parameters. simulating interest rate changes over the thirty year term of a mortgage is the engine that drives these simulations. while the loan parameters can be taken from actual contracts, parameters related to the rate changes must be selected by the modeler. for the most part, we use average historical values in setting these parameters. first, we collected data on the one year constant maturity yield of u.s. treasury securities as reported on a weekly basis by the federal reserve. from these treasury data, we calculated a standard deviation of annual index rate changes of 2.52 percentage points. see figure 1 for a dis~bution of annual changes in this index from november 1979 to december 199 1. this represents the fixed and adjustable rate mortgages 105 table 2 low interest rate environment simulation parameters parameters term loan amount discount points index initial index value standard deviation of index maximum value of index minimum value of index margin over index initial composite rate initial adjustable rate initial fixed rate after-tax discount rate marginal tax rate annual rate change cap life time rate change cap adjusfable rare 30 years $100,000 .500% 1 year treasury 3.16% 2.52% 11.00% 3.00% 2.75% 5.910% 5.375% 4.000% 28% 2.00% 6.00% fixed rate 30 years $100,000 1.000% 7.875% 4.000% 28% period from just after the major shift in fed policy regarding managing interest rates through the year just prior to when the first simulation begins. the annual changes appear to be approximately normally distributed. we simply observe the first year index rate. the index rate for the second year is modeled by multiplying a randomly selected z-score from a normal distribution times the standard deviation of annual changes and adding this to the index rate value for the previous year. subsequent index rates are modeled in the same way. an index rate floor of 3% and a ceiling of 11% were selected to keep the simulated index rate within a reasonable range. this method of simulating interest rates follows tucker (1991) in setting the expected index value for the subsequent period equal to the present index value. in short, it assumes no reversion toward a mean index value. we have added mean reversion parameters to our model and run again the simulations reported in this study. the effects on the results are not markedly different and the conclusions we draw would not be affected. because mean reversion is a controversial issue (e.g., chan, et al., 1992 find insignificant results when testing for it in one month treasury yields) and it adds complexity to the model, we report results without it. however, mean reversion can easily be added to the model and a user who expected mean reversion could include it. figure 2 shows a sample thirty year series of simulated index rate values constructed using the technique just described. it also shows an actual series of the index rate for comparison’s sake. for simplicity, the payments are treated as annual rather than monthly. the payment for the frm is the constant annuity mount that completely amortizes the loan over the thirty year life of the mortgage. the payment for the arm is recalculated annually. a com posite rate is determined for each period by adding the specified margin to the simulated value of the index rate. the composite rate is subjected to annual and lifetime caps speci fied in the arm contract to arrive at the simulated mortgage rate for each period. the sim ulated mo~gage rate, the remaining balance, and the rem~ning number of years in the mortgage determine the adjusted arm payment. the opportunity discount rate is the assumed after-tax return the borrower could earn if the difference between the two payment amounts were invested. we use an after-tax 106 financial services review 5(2) 1996 standard deviation = 2.52% annual changes figure 1. distribution of annual changes in one year constant maturity treasury rates, january 1979 to december 199 1. 0 5 10 15 20 25 30 year i1 year constant maturity treasuw yield -simulated value 1 figure 2. an actual thirty year series of the one year constant maturity treasury rate (observed in the first week of each year from 1963 through 1992) compared to a sample iteration of the simulated rate. fixed and adjustable rate mortgages 107 opportunity discount rate of 4% for all the simulations described in this study. an altema tive way to think about the opportunity discount rate is that it is the borrower’s next best borrowing rate. under that view the rate would likely be higher. the marginal tax rate used in these simulations is 28%, a rate typical of higher income borrowers. for each iteration, the spreadsheet collects the following items: l the difference in present value cost between the frm and the arm using equa tion 1 to calculate each. we calculate the difference in present value cost assum ing termination of the mortgage in each of years 1 through 30. l the breakeven period-the number of years for which the arm maintains its present value cost advantage from the initiation of the loan. typically, an arm begins with a lower mortgage rate so that termination of the mortgage in the early years will result in a present value cost advantage for the arm. this advantage often disappears as the life of the mortgage lengthens and the arm mortgage rate has the opportunity to move higher than the frm, although it is possible for the arm to have a present value cost advantage for the entire term of the mortgage. because of present value principles, the years of initial cost advantage for the arm may extend several years after the arm mortgage rate has risen above the frm rate. l the thirty different payment amounts for the arm and the payment for the frm. each simulation consists of one thousand iterations. expected values and a distribu tion of values for the collected items are then compiled. iii. low interest rate environme~ simulation month end of september 1992 represented a time of relatively low interest rates. we per formed this first simulation using the parameters shown in table 2. we define the differ ence between the initial composite rate and the initial mortgage rate offered by the lender as the teaser discount. it is customary for lenders to entice borrowers to choose an arm by offering an initial mortgage rate below the composite rate for the first year. at the first annual adjustment, the mortgage rate would increase even if the index rate did not increase, because the teaser discount disappears. figures 3,4, and 5 show results of the low interest rate environment simulation. in fig ure 3, the heavy black line shows the expected present value cost difference (m-arm) for mortgage lives ranging from 1 to 30 years. when the line appears above the ho~zontal axis the cost advantage is with the arm. when it goes below the axis the frm has the cost advantage. this line resembles the results shown in figure 2 of tucker’s (1991, p. 455) study. our figure 3 contains more information, however. it also shows the expected present value cost difference plus or minus one standard deviation and the range that includes ninety percent of the present value cost differentials. from figure 3 we see that if an individual with a 4% discount rate were to hold a mortgage just 6 years, the arm would have an expected present value cost advantage of approximately $5,000. the distribution around that expected advantage, though, suggests there could be a cost disadvantage to the 108 financial services review s(2) 1996 arm for that mortgage life in roughly one-sixth of the cases. at the extremes, the arm could have a cost advantage of roughly $11,000 or the ffw could have a cost advantage of roughly $6,000. the iterations produce widely disparate results, and the expected cost is just one component of a thorough analysis. borrowers must compare these expected bene fits to the risks inherent in the arm. an inference that may be drawn from the expected cost information is that the arm is advantageous as long as the expected mortgage life is 14 years or less. in figure 4, we show the distribution of this initial cost advantage to the arm. of the 1,000 iterations, the arm present value cost dominated the frm cost for the entire thirty year term over 35 percent of the time. on the other hand, the initial cost advantage to the arm would have lasted only 3 years about 8 percent of the time. clearly, the expected value of 14 years means little given the distribution of these results. one must be careful in interpreting fig 50,000 40,000 30,000 20,000 tl q z 0 10,000 i” z e $ 0 e! a -10,000 -20,000 -30,000 ~0,000 4 arm advantage life of mortgage figure 3. low interest rate environment cost comparison. the heavy black line indicates the expected present value cost differential (~~-a~) of $100,000 mortgages at a 4 percent discount rate. the lines marked with triangles represent the expected present value cost differential plus and minus one standard deviation. ninety percent of the simulation runs resulted in present value cost differentials between the lines marked with squares. fixed and adjustable rate mortgages 109 ure 4. we capture only the years of initial cost advantage. depending upon how interest rates behave in a particular iteration, the cost advantage to the arm may disappear quickly but return later in the life of the mortgage. thus, even in some of the cases in which the frm gains a cost advantage early on, the arm could regain the advantage in later years. here we capture only the number of years the initial advantage of the arm is maintained. figure 5 shows the distribution of required annual payments over the thirty year term of the mortgage. the heavy black line shows an expected arm payment that increases at a decreasing rate and levels off at about $9,570. the frm payment is shown as a horizon tal dashed line at $8,778. this figure allows one to view the expected time path of the arm payment amount as well as the possible extremes and their likelihood. it also allows one to compare these aspects to the frm payment. these results show that the payment advan tage of the arm is likely to last about 5 years. this payment advantage period is much shorter than the cost advantage period because of present value principles. from the stand point of maximizing wealth, borrowers should choose the mortgage that will have the lower expected cost. however, some borrowers may be risking insolvency with higher mortgage payments. these borrowers would probably focus on the payment aspect rather than the present value cost difference. year figure 4. low interest rate environment-number of years for which the arm maintains its present value cost advantage from the initiation of the loan. for each year, the bar height shows the fraction of cases for which the arm’s initial present value cost advantage continues for exactly that many years. thus, the bar height at year 30 indicates the fraction of cases in which the cost of the arm dominates the cost of the frm for the entire life of the mortgage. the assumed discount rate is 4 percent. 110 financial services review 5(2) 1996 8,000 f = 0” 6,000 figure 5. low interest rate environment-expected annual mortgage payment. the heavy black line represents the expected payment for the arm. the lines marked with triangles represent the expected payment plus and minus one standard deviation. ninety percent of the simulation runs resulted in annual payments between the lines marked by squares. the horizontal, heavy, dashed line represents the frm payment. iv. high interest rate environment simulation month end march 1989 represented a time of relatively high interest rates. the parameters for this simulation appear in table 3. figure 6 shows the present value cost differential results of this simulation. it tells a much different story than the low interest rate simula tion. the arm offers an expected present value cost advantage throughout the thirty year term. however, this simulation is artificial in that it assumes no refinancing opportunities for the high rate frm. a borrower in these circumstances can benefit from falling rates in two ways: select the arm, or select the frm and refinance if rates fall significantly. we incorporate this second possibility by simulating a ten year treasury index rate in the same manner as the one year treasury index rate and assuming a constant margin between the ten year rate and frm rates. the simulation of each index is driven by the same random fixed and at&&able rate mortgages 111 draw from a normal distribution, thus maintaining a connection between the short term and long term rates. both indexes will move in the same direction, but by amounts proportional to their historical standard deviations of annual changes. in addition, we enforce different floors and ceilings on the simulated index rates. table 3 also includes parameters necessary to allow the refinancing option. if the simulated frm rate drops 2.5 percentage points below the initial frm and there are at least 15 years left in the term of the mortgage, we assume refinancing. for example, if in a particular iteration the simulated frm rate drops to 8.5% with 20 years left on the mortgage, we add the discount points cost and an addi tional $1,000 in refinancing costs to the balance of the mortgage and recalculate the pay ment to amortize the new balance over the remaining 20 years. this refinancing rule approximates rules of thumb used by borrowers. including this option in the high interest rate environment simulation causes this frm to dominate this arm in terms of expected present value cost. figures 7,8, and 9 show the results of this simulation. now the expected present value cost advantage to the frm occurs by the end of the third year and increases throughout the life of the mortgage. the teaser discount was just over 2 percentage points in this case. the loss of the teaser dis count after the first year explains the short duration of the arm cost advantage. falling rates would be adv~tageous for the arm borrower, but because of the re~n~cing option frm borrowers can also benefit. since the frm has both a lower expected present value cost and less risk, risk averse borrowers should prefer the frm in choosing between these particular loan terms. figure 8 shows the distribution of breakeven periods. of the 1,000 iterations, the arm present value cost dominated the frm cost for just one year in fifty nine percent of the cases. dominance of the arm over the frm for the entire thirty year term in this simulation is rare indeed. figure 9 shows the expected path of the arm pay table 3 high interest rate environment simulation parameters parameters ~~us~~l~ rate fixed rate term 30 years 30 years loan amount $100,000 $100,000 discount points 1.250% 1.000% index 1 year treasury 10 year treasury initial index value 9.78% 9.20%* standard deviation of index 2.77% 2.21%* maximum value of index 11.00% 12.00%* minimum value of index 3.00% 5.25%* margin over index 2.75% 2.050%* initial composite rate 12.53% initial adjustable rate 10.50% initial fixed rate 11.25% after-tax discount rate 4.00% 4.00% marginal tax rate 28% 28% annual rate change cap 2.00% life time rate change cap 6.00% fixed rate refinancing trigger 2.50%* minimum required remaining term 15 years* for frm refinancing nufe~r these waxlard deviations of annual changes in index values were based on the period november 1979 to december i986. the standard deviation of changes in the one year rate is slightly higher than that used for the first simulation. the estimation period for the standard deviation for the tint simulation was november 1979 to december 1991. fixed rate parametera marked with an asterisk are only in the simu lation that allows the refinancing option for the frm. 112 financial services review 5(2) 1996 50,000 40,ooa 30,000 20,000 $ = 0” 10,000 f 3 e z 0 p! a -10,000 -20,000 -30,000 -40,000 i , - arm advantage / ..jgi life of mortgage i figure 6. high interest rate environment cost comparison. the heavy black line indicates the expected present value cost differential (em-a~) of $l~,~o mortgages at a 4 percent discount rate. the lines marked with triangles represent the expected present value cost differential plus and minus one standard deviation. ninety percent of the simulation runs resulted in present value cost differentials between the lines marked with squares. ment versus the expected path of the frh4 payment. borrowers who focus on this aspect would also be better served by choosing the frm given these loan terms. v. riscussio~ comparing the loan terms of arms and frms can be difficult. borrowers must consider the impacts of multiple variables on total present value cost, payment size, and risk. future interest rate movements, the initial spread between the arm and frm rates, the teaser dis fixed and adjustable rate mortgages 113 50,000 40,000 30,000 20,000 g z n 10,000 s 3 e 2 0 2? n -10,000 -20,000 -30,000 -40,000 life of mortgage figure 7. high interest rate environment cost comparison assuming refinancing option. the heavy black line indicates the expected present value cost differential (ffcm-arm) of $100,000 mortgages at a 4 percent discount rate. the lines marked with triangles represent the expected present value cost differential plus and minus one standard deviation. ninety percent of the simulation runs resulted in present value cost differentials between the lines marked with squares. count, discount points, refinancing costs, and the planned life of the mortgage all influence this choice. this paper has examined the arm-frm choice in both relatively high and low inter est rate environments and found results that challenge some of the advice commonly given to borrowers. for example, is an arm the low cost choice when mortgage rates are high? in the case of the lender terms used in this study, the initial spread between arm and frm rates in the high interest rate environment was only 75 basis points. in fact, without the first year teaser discount, the initial arm mortgage rate would have exceeded the frm rate. though index rates may decline, any present value cost advantage to the arm will depend on how long it takes for this decline to occur and how long rates remain at lower levels. on the other hand, the frm would offer protection against further rises in interest rates, would not be affected by the disappearance of a first year teaser discount, and can be refinanced 114 financial services review s(2) 1996 figure 8. high interest rate environment assuming refinancing option-number of years for which the arh4 maintains its present value cost advantage from the initiation of the loan. for each year, the bar height shows the fraction of cases for which the arm’s initial present value cost advantage continues for exactly that many years. thus, the bar height at year 30 indicates the fraction of cases in which the cost of the arm dominates the cost of the frm for the entire life of the mortgage. the assumed discount rate is 4 percent. if rates fall substantially. tucker’s (1991) simulation results suggest near dominance of the arh4 in terms of expected present value cost during the 1985 to 1989 time period, a time of relatively high rates. his results may have been influenced by his neglect of the frm refinancing option. when rates are relatively high, it seems that borrowers still would be better served by focusing on rate differentials. figure 10 shows data from the federal housing board’s monthly survey of major lenders. both the average frm rate and the average initial rate differential are depicted. the differential is not consistently small when the frm rate is high or vice versa. thus, we cannot generalize our results to suggest the frm is always preferred when rates are high. in addition, the teaser discount is part of the differential and tends to mask the difference between the fixed rate and the initial composite rate. when frm rates are relatively low, borrowers may see the frm as the most attractive alternative. however, our results suggest that a larger initial spread between an arm and frm in these circumstances can result in an expected cost advantage that extends for quite a few years. the payment advantage may extend for quite a while as well. for the loans compared here, borrowers whose expected mortgage life was relatively short should have found the arm terms advantageous. these results are consistent with the findings of spre cher and willman (1993) who determined that arm borrowers had an interest rate advan fixed and adjustable rate mortgages 115 14,000 8,000 !i 3 n 8,000 arm payment i ! frm payment figure 9. high interest rate environment assuming refinancing option-expected annual mortgage payment. the solid line represents the expected payment for the arm. the dashed line represents the expected frm payment. it declines because refinancing of the frm to a lower rate is likely. tage over frm borrowers during a four year period in which rates rose. borrowers must weigh these cost advantages against the increased risk associated with the arm. to summarize, although analysis yields no hard and fast rules about when to select an arm rather than a frm, larger initial rate differentials should cause borrowers to favor the arm, particularly if the differential does not merely reflect the teaser discount. borrowers should attempt to identify the size of the teaser to avoid being deceived by an attractive rate differential that will not last past the first year. borrowers who intend a short mortgage life should also tend to favor arms. arms tend to exhibit expected cost and payment advan tages for short lived mortgages. we suggest borrowers probably should avoid selecting a mortgage based on the level of fixed rates alone. in the simulations reported here, the initial rate differenti~s and the teaser discount were more impo~ant factors in dete~ining the relative cost. the opportunity to refinance a frm significantly alters its expected cost and relative attractiveness compared to arms. borrowers ought to keep this option in mind. 116 financial services review 5(2) 1996 q q t$ q $8 6 q q zfi 8 iz is p p c s 5 5 5 5 % z $ p 6 p p i fiied rate -(fixed rate adjustable rate)] figure 10. average frm rates as reported by the federal housing finance board from its monthly national survey of major lenders. also shown is the average initial rate differential between frms and arms from the same source. although we can make some statements regarding each of these factors and their influence on the choice, the value of the simulation approach is that it considers these aspects simul taneously and provides an overview of the choice. comparisons ought to be done for each pair of competing loan terms. a borrower who can select an appropriate discount rate, esti mate the mortgage life, set a tolerance for increases in payments, and determine his own risk aversion level ought to be able to make a better decision using simulation results as described in this study. though this study has compared just two types of mortgages, the thirty year frm and the one year arm, the technique can easily be extended to the more exotic types of mort gages. arms that have a fixed rate for the first five or seven years and adjust annually after that are becoming more popular. other arms adjust their rates every six months or every three years. the model presented here may be used to compare two mortgages of any variety. fixed and adjust&e rate mortgages vi. conclusion 117 this paper has presented and demonstrated a simulation model that shows the impact of future interest rate movements, the initial spread between the arm and frh4 rates, the teaser discount, discount points, refinancing costs, and the planned life of the mortgage on cost and payments. the model also provides borrowers a clearer picture of the risks involved. taken together, the simulation results would allow borrowers to make a more informed choice. typical borrowers would not have the necessary expertise to build and interpret the simulations reported here. financial advisors, such as mortgage brokers or thrift officers ought to be able to construct such models, however. simulation outputs could greatly aid their clients in comparing loan terms. references @risk (computer software), version 3.5. (1996). palisade corporation. brueckner, jan k., & follain, james r. (1988). the rise and fall of the arm: an econome~c anal ysis of mortgage choice. the review of economics and statistics, io, 93-102. chan, k.c., karolyi, g. andrew, longstaff, francis a., & sanders, anthony b. (1992). an empiri cal comparison of alternative models of the short-term interest rate. the journal of finance, 18,1209-1227. dhillon, upinder s., shilling, james d., bi sirmans, cf. (1987). choosing between fixed and adjustable rate mortgages. journal of money, credit, and banking, 19,260-267. excel (computer software), version 7.0. (1995). microsoft corporation. goldberg, lawrence g., & heuson, andrea j. (1992). fixed versus variable rate financing: the influence of borrower, lender and market characteristics. the journal of financial services research, 649-60. lino, mark. (1992). factors affecting borrower choice between fixed and adjustable rate mortgages. the journal of consumer affairs, 26,262-273. news. various issues 1985 through 1995. washington, dc: federal housing finance board. o’brien, a., maureen, & wong, shee q. (1990). the choice between fixed and variable rate mort gages: evidence from national data. ame~can 3usiness review, 8,49-54. phillips, richard a., & vanderhoff, james. (1991). adjustableversus fixed-rate mortgage choice: the role of initial rate discounts. the journal ofreal estate research, 639-51. sprecher, c. ronald, & willman, elliott s. (1993). arms versus frms: which is better? the real estate finance journal, 8,78-83. tucker, michael. (1989). adjustable-rate and fixed-rate mortgage choice: a logit analysis. the journal of real estate research, 4,81-91. tucker, michael. (1991). comparing present value cost differentials between fixedand adjustable rate loans: a mortgage simulation. the financial review, 26,447-458. u.s. bureau of the census. (various years). statistical abstract of the united states. washington, dc. yohannes, arefaine g. (1991). frm or arm: simplifying the choice. the reai estate appraiser, 57,21-24. pii: s1057-0810(99)00025-6 managing college tuition inflation using a surplus framework methodology judson w. russell,a robert brooks, ph.d., cfab,* avice president, industry research, global corporate & investment banking, bank of america, nc1-005-1102, charlotte, nc 28255, usa bsouthtrust professor of financial management, department of economics, finance and legal studies, university of alabama, 200 alston hall, box 870224, tuscaloosa, al 35487, usa abstract this paper explores prepaid college tuition plans and develops a methodology for managing college tuition inflation. a surplus framework methodology is derived that employs various securities and incorporates both the assets and liabilities associated with prepaid college tuition plans. although we present a methodology for plan management, the approach is applicable for individuals who manage their own college investment accounts. an interesting result of this analysis is that u.s. treasury inflation-indexed securities should be included in the asset allocation decision for college tuition inflation management. by incorporating both the assets and liabilities from the plans into a surplus framework methodology, this paper provides a new portfolio management tool for plan administrators and individuals. our assertion is that better managed plans offer more college financing alternatives for individuals. © 1999 elsevier science inc. all rights reserved. keywords:portfolio choice, computational techniques, investment policy 1. introduction college tuition is a subject of concern for families who are saving for their children’s education. while consumer prices have been increasing at a relatively modest rate over the past few years, college tuition has been increasing much more rapidly. during the period * corresponding author. tel.:11-205-348-8987; fax:11-205-348-0590. e-mail address:rbrooks@cba.ua.edu (r. brooks) financial services review 7 (1998) 257–271 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00025-6 1980–1995, tuition at public universities increased 234%, while median household income rose 82%, and the consumer price index (cpi) increased 74%. increases in grant aid have not kept up with tuition inflation. students, and their parents, have absorbed a large portion of the tuition increase through loans. in 1980, the average student loan was $518. by 1995 the average loan was $2,417, an increase of 367%. with tuition rising faster than household income, an affordability gap has developed. in an effort to reduce the anxiety of meeting college tuition needs, many states have adopted prepaid college tuition plans or variations on the plans. a synopsis of state plans is included in table 1 and fig. 1. in general terms, prepaid tuition plans allow investors to pay for college at approximately today’s tuition levels for college attendance at a later date. these plans should help mitigate the uncertainty associated with tuition inflation for the individual investor. the uncertainty, or risk, associated with tuition inflation is shifted from individual investors to states offering the plans or to other guarantors, depending on the structure of the state’s plan. the state’s ability to manage this risk is essential to the viability of this tuition-financing alternative. without proper management skills and methods the plans become actuarially unsound and the risk may be transferred back to the individual investor. the state’s risk stems from the possibility that plan income from premiums and investment activities will not keep pace with tuition inflation. two state programs, michigan and wyoming, have experienced the effects of this risk. michigan’s plan was suspended when it was deemed actuarially unsound. it has since been reinstated, subject to liquidation if it once again becomes actuarially unsound. wyoming’s plan has also experienced actuarial difficulties. the tuition outlays exceeded the plan’s principal and accumulated interest. the state suspended the program, but it honored its obligation and subsidized the shortfall in the contracts. this paper presents several concepts, but each directly impacts tuition risk-shifting or hedging programs that ultimately affect the individual. the essential elements of prepaid college tuition plans are discussed and a methodology for managing tuition inflation exposure is developed. a relatively new class of securities is introduced that may be beneficial to either states, that are managing prepaid tuition plans, or individuals, who may wish to hedge tuition inflation risk on their own. this paper makes three significant contributions to the existing literature: (1) prepaid college tuition plans are still in their infancy. this paper explains the mechanics of these plans and discusses the burden of tuition inflation; (2) surplus framework management is also a relatively new concept. this portfolio management concept has primarily been applied to defined benefits plans. this paper presents a new direction for using this methodology; and (3) u.s. treasury inflation-indexed securities represent a new asset class. this paper provides a brief background on the new securities and, more importantly, provides a practical application of the securities in a risk management, surplus framework context for prepaid college tuition plan management and tuition inflation hedging. in section 2, the background for prepaid college tuition plans is discussed. currently, available plans are presented and are contrasted with other financing alternatives. section 3 illustrates the transfer of the tuition inflation burden from the individual to the state and the extant surplus framework literature is reviewed. the surplus framework model is developed in section 4. this innovative portfolio management tool is further discussed by means of an example in section 5. section 6 provides concluding comments regarding the use of u.s. 258 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 table 1 college plans by state state program name program year initiated alabama prepaid affordable college tuition (pact) prepaid plan 1990 alaska university of alaska advance college tuition payment plan prepaid plan 1991 arizona arizona family college savings program savings plan 1998 arkansas arkansas college savings bond obligation bond program 1991 california golden state scholarship trust college savings program savings plan 1998 colorado colorado prepaid tuition fund prepaid plan 1997 connecticut connecticut higher education trust (chet) savings plan 1997 delaware delaware college investment plan savings plan 1998 florida florida prepaid college program prepaid plan 1988 georgia help outstanding pupils educationally (hope) scholarship program scholarship 1993 illinois illinois prepaid tuition plan prepaid plan 1998 indiana indiana family college savings plan and save indiana savings plan 1996 iowa college savings iowa savings plan 1998 kentucky kentucky educational savings plan trust savings plan 1990 louisiana louisiana student tuition assistance & revenue trust program savings plan 1997 maryland maryland prepaid college trust prepaid plan 1998 massachusetts u-plan the massachusetts college savings program prepaid plan 1995 michigan michigan education trust prepaid plan 1988 mississippi mississippi prepaid affordable college tuition program (mpact) prepaid plan 1997 missouri missouri higher education savings program savings plan 1998 nevada nevada prepaid tuition program prepaid plan 1997 new hampshire new hampshire savings plan: unique college investing plan savings plan 1998 new jersey new jersey better educational savings trust savings plan 1997 new york new york state college choice tuition savings program savings plan 1997 north carolina college vision fund savings plan 1996 ohio ohio prepaid tuition program prepaid plan 1989 oklahoma oklahoma college savings plan savings plan 1998 pennsylvania pennsylvania tuition account program prepaid plan 1993 rhode island rhode island higher education savings trust savings plan 1998 south carolina south carolina tuition prepayment program prepaid plan 1997 tennessee baccalaureate education system trust (best) prepaid plan 1996 texas texas prepaid higher education tuition program (tomorrow) prepaid plan 1996 utah utah educational savings plan trust savings plan 1996 vermont vermont higher education savings plan savings plan 1999 virginia virginia prepaid education program (vpep) prepaid plan 1996 washington guaranteed education tuition prepaid plan 1997 west virginia west virginia prepaid college plan prepaid plan 1998 wisconsin edvest wisconsin (wisconsin education investment program) prepaid plan 1997 wyoming advance payment of higher education costs program* prepaid plan 1987 sources: college savings plan network and national association of state treasurers (1998). * the program was suspended in 1995 due to lack of participation. all contracts sold during the program’s lifetime are being honored by the state of wyoming. 259j.w. russell, r. brooks / financial services review 7 (1998) 257–271 treasury inflation-indexed securities to assist plan administrators or individuals in the management of tuition inflation. 2. prepaid college tuition plans during the past few years, prepaid college tuition plans have been gaining in popularity. michigan was the first state to initiate a plan, and florida’s is now the largest. the basic idea behind the plan is to allow residents to purchase college tuition for subsequent use at roughly today’s price. in this manner, the individual will not be exposed to tuition inflation. states offer the plans to residents and allow for a variety of options and differing degrees of fig. 1. state college tuition plans. sources: college savings plan network and national association of state treasures (1998). 260 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 flexibility. particular plans are portable and allow the recipient to attend any accredited college in the country while some allow for partial payment at private colleges. several states allow the plans to be transferred to other members of the family. there are severe restrictions with a few of the plans. for instance, suppose a child decides not to attend college or does not qualify academically. some of the money that has been invested may be forfeited. other plans will refund the invested money, either without interest or with a very modest amount of interest. individuals find it appealing to eliminate the uncertainty associated with future college tuition prices. using the prepaid plans, tuition inflation risk is passed on to states or others parties that ultimately guarantee the plans. figure 2 illustrates a reason for the plans’ popularity. this figure shows that tuition prices have increased at a much higher rate than the general level of prices. families saving for college realize that tuition costs are escalating and that the prepaid college tuition plans provide more certainty regarding long-term planning. families can essentially pay for college at today’s prices and receive an assurance from the state that tuition will be covered when the child is ready to attend college. a few states have even discussed a prepaid room and board feature to allow for even more certainty regarding the financial obligation of the family. fowler (1998) examines public school inflation in detail. as noted in table 1 and fig. 1, some states have opted for savings plans rather than prepaid tuition plans. college saving alternatives vary from tax-advantaged bonds, typically fig. 2. college tuition and cpi 19835 1. source: dri: standard & poors. 261j.w. russell, r. brooks / financial services review 7 (1998) 257–271 state zero-coupon bonds, to college savings accounts. unlike prepaid tuition plans, taxadvantaged bonds do not require the funds to be spent on college expenses. to encourage college savings, some states offer a bonus with the tax-advantaged bonds as an incentive to use the funds for higher education at redemption. the zero-coupon bonds are sold at a discount with the difference between purchase price and face value representing the interest on the bonds. the interest is exempt from federal taxes and, for purchasers residing in the issuing state, from state taxes. college savings accounts allow individuals to invest in funds that are managed by trusts. as an example, the savings plan trust, offered in kentucky, provides a guaranteed minimum interest rate of 4%, but the earnings vary depending on the manager’s selection of investment vehicles. prepaid tuition plans shift tuition inflation away from the individual while college savings accounts do not shift tuition inflation to other parties. in fact, if the accounts return the guaranteed 4% and tuition continues to increase at historic levels, individuals will notice a significant shortfall between their college savings accounts and the tuition liability they will face. prepaid tuition plans come in three main varieties. there are contract, tuition credit, and certificate plans. contract plans allow the purchaser to enter an agreement for a predetermined amount of education, with the cost calculated as the current tuition level. for instance, if the current cost of four years of tuition is $16,000, an individual may purchase four years of college tuition today and receive a contractual guarantee for four years of tuition in the future. the risk associated with tuition inflation is completely shifted from the individual to the guarantor. tuition credit plans are similar to contract plans in that they allow the purchaser to obtain prepaid education. they differ in that the purchaser starts an account and continues to make periodic deposits to purchase prepaid units of education. certificate plans allow individuals to purchase certificates that are redeemable for a percentage of tuition and mandatory fees. the state commits to pay face value plus annually compounded interest of cpi plus some percent. for instance, massachusetts’ plan allows for interest to compound at cpi plus 2%. if tuition increases faster than cpi plus 2% theschoolsabsorb the loss. in each of these plans the risk of tuition inflation is shifted from the individual to either states, schools, or third-party plan managers. the introductions of prepaid college tuition plans assist students and parents, but risks have been introduced that guarantors must either hedge or absorb. 3. surplus framework a relatively new method for viewing the asset allocation decision within portfolio management is the surplus framework. this framework has become popular for defined benefit plan managers. the objective of the plan manager is to maximize the surplus, or the difference between the assets and liabilities. not only are the defined benefit plan managers concerned with asset returns, they are also acutely aware of the plan’s liabilities. the prepaid college tuition plan manager (administrator) faces a similar dilemma. the liabilities are a key concern. the liabilities faced by the prepaid college tuition plan administrator are the tuition outlays that are the responsibility of the guarantor. 262 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 much of the recent literature regarding the inclusion of liabilities in portfolio management optimization stems from the work of sharpe and tint (1990). in their study, the authors show that investors may maximize either their asset return or their surplus. traditionally, fund managers have chosen to maximize asset returns following standard modern portfolio theory. however, situations do arise in which maximizing the surplus is both prudent and appropriate such as pension funds. ezra (1991) presents a discussion of liability modeling within the context of defined benefit plans. a key characteristic of defined benefit plans is that pension benefits are defined independent of the value of the plan’s assets. the benefits are based on formulas that incorporate employee earnings and, typically, length of service. in essence, the corporation has issued benefit debt to the plan participants. this debt (liability) can be serviced on either a pay-as-you-go basis or a funded basis. a funded basis requires that funds be set aside each year to match the present value of the debt accrued that year. the plan manager needs to be cognizant of the plan’s liabilities and not simply focus on asset growth. in his paper, ezra concludes that the sponsor is better advised to focus on surplus rather than on assets alone when determining asset allocation for defined benefit plans. fong (1991) incorporates portfolio theory into the surplus management problem. leibowitz et al. (1992) present an analysis in which both asset-only performance and surplus control are considered in asset allocation decisions. the authors’ “dual-shortfall” approach demonstrates an overlay of the two constraints that were identified by sharpe and tint (1990). macbeth et al. (1994) reiterate the importance of including the surplus framework when considering asset allocations for defined benefit plans. the authors conclude that a meaningful evaluation of the true risks and rewards for defined benefit plans lies not on a mean-variance analysis of asset returns, but by focusing on future sponsor contributions instead. peskin (1997) finds that defined benefit plan investing has evolved from an asset-only methodology to an asset/liability framework. the author concludes that corporations can reduce the present value of future contributions to the plans by more than 20% by incorporating a surplus framework methodology. further, peskin finds that focusing on asset-only return maximization leads to financially risky and costly asset allocation decisions. one of the key components of the author’s study is the discussion on synchronizing assets and liabilities. based on fig. 1, the reader can see that there are roughly an equivalent number of states offering prepaid plans as savings plans. reasons offered for the popularity of the college savings plans is that they are less administratively difficult and less risky than prepaid plans which involve tuition inflation risk-shifting. in the next section, a model is developed that builds upon the surplus management literature for defined benefit plans. this model is extended to prepaid college tuition plans, but can be employed by individuals who choose to manage their own college investment account. 4. surplus framework model sharpe and tint (1990) present a model for surplus optimization in defined benefit plans. their model is based upon the goal of maximizing the surplus for a plan in the year to come. in this section, this model will be expanded and presented with applications for prepaid 263j.w. russell, r. brooks / financial services review 7 (1998) 257–271 college tuition plans. the basic foundation for the model is the measure of surplus, defined as assets minus liabilities. sharpe and tint (1990) further define a variable, k, that captures the importance attached to the liability by the fund manager. this variable can obtain values within the range of zero to one, where zero implies that liabilities are ignored, as in asset-only optimizations, and one implies that a full surplus optimization is utilized. by incorporating this variable, k, the surplus model becomes: s1 5 a0 1 a 1 «a 2 k(l0 1 l 1 «l) (1) where s1 5 next year’s surplus a0 5 plan assets in period 0 a 5 net cash for plan. this value is positive if there is a net increase in participation or negative if there is a net outflow of funds, e.g. the child decides to forego college. «a 5 return on plan assets during period 0 to 1 k 5 importance attached to liability l0 5 plan liabilities in period 0 l 5 present value of net additions/reductions in liabilities during period «l 5 return on liabilities due to underlying market factors, i.e. tuition inflation, rather than an addition or paydown in the plan’s liabilities. the liability incurred by the plan,l, is the present value of the expected future tuition payment. this value may be determined by projecting tuition into the future and then discounting the future value by an appropriate discount factor. the present values of the assets,a, and liabilities,l may differ due to the recognition of the liability. in this study, it is assumed that the fund recognizes a liability,l, that is equal to the inflow of assets,a. the present value of the future tuition outlay is equal to the prepaid tuition that is being purchased today. we can relax this assumption without materially impacting the results. assuming thata 5 kl and expressing eqn. 1 relative to today’s asset value we have the following: s1/a0 5 [a0 1 «a]/a0 – k[l0 1 «l]/a0 (2) in order to provide additions to the surplus, the plan would need for the asset investments to exceed the liability return, or«a . «l. in other words, the funds invested need to provide a higher return than the change in the liability. the change in the liability is directly impacted by tuition inflation. to maximize the plan surplus, s1, the manager must make asset allocation decisions that are inclusive of the liability of the plan. equation 2 can now be expressed using the more familiar notation of returns. for instance, the term [a0 1 «a]/a0, next year’s asset value over this year’s value, equals 11 ra. multiplying the final term by the constant l0 / l0, we may express the liability using the notation of returns, i.e., [l0 1 «l]/l0 5 1 1 rl. s1/a0 5 1 1 ra – k[l0/a0][1 1 rl] (3) equation 3 can be rearranged into certain components and uncertain components: 264 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 s1/a0 5 [1 – k(l0/a0)] 1 [ra – k(l0/a0)rl] (4) the first bracketed term on the right-hand side of eqn. 4 is certain, while the second term is uncertain. the prepaid college tuition plan manager, or individual managing his/her own college investment account, is concerned with maximizing next year’s surplus by making appropriate asset allocation choices today. the manager is concerned with the effects of the asset allocation decision on the second term on the right-hand side of eqn. 4, that is, the uncertain component. for convenience this term can be defined asm: m § [ra – k(l0/a0)rl] (5) equation 6 summarizes the objective of surplus optimization using a one-period model, t 5 1, and the goal of utility maximization where we assume a quadratic utility function: umax§ e[m] – var[m]/h (6) where: e[m] denotes the expected return of the uncertain component,m var[m] denotes the variance of the uncertain component,m h denotes the fund manager’s, or individual’s, risk tolerance with the substitution ofm into the expected return component of eqn. 6: e[m] 5 e[ra] 2 k(l0/a0)e[rl] (7) notice that eqn. 7 reduces to e[ra] when the importance of the liability is zero, k5 0. although the manager is concerned with liabilities, the asset allocation decision only involves the first term on the right hand side of eqn. 7. therefore, the utility maximization objective is consistent with traditional asset-only return management with regard to expected return. the departure from traditional asset-only return management, and the primary contribution of the surplus framework management comes from the treatment of the variance term in eqn. 6. the variance term can be expanded, as shown in eqn. 8 below: var[m] 5 var[ra] – 2k[l0/a0]cov[ra,rl] 1 k2[l0 2/a0 2]var[rl] (8) where var[ra] denotes the variance of asset returns cov[ra,rl] denotes the covariance of asset and liability returns var[rl] denotes the variance of liability returns traditional asset-only management methodology focused on the first term on the righthand side of eqn. 8. the last term on the right-hand side of eqn. 8 involves constants and variation in the returns on liabilities. this term is not influenced by asset allocation decisions. the second term on the right-hand side of eqn. 8 is the key contributing factor in a surplus framework. this is the only term that allows for a departure from traditional asset-only portfolio management. combining eqns. 7 and 8: umax§ e[ra] 2 var[ra]/h 1 2k/h[l0/a0]cov[ra,rl] (9) 265j.w. russell, r. brooks / financial services review 7 (1998) 257–271 the goal of the prepaid college tuition plan manager is to choose a utility maximizing asset allocation. the first two terms on the right-hand side of eqn. 9 are equivalent to the traditional portfolio maximization goal, that is, to maximize asset return subject to asset return variation. in other words, maximize the risk-adjusted return on assets. these two terms are identified as the expected return and risk penalty, respectively, where the risk penalty is defined as the variance of asset returns divided by the manager’s risk tolerance. without diverting into an exposition on utility theory, risk tolerance can be viewed as the reciprocal of the absolute risk aversion parameter, as developed by pratt (1964) and others. following risk aversion parameter values suggested by friend and blume (1970) and grossman and shiller (1981), this study assumes that the degree of risk aversion for the representative investor ranges from 2.0 to 4.0. this range of values is consistent with studies that take into account a full range of available assets. the following example will illustrate the risk-adjusted expected return using traditional asset-only portfolio management techniques. assume that the expected return from a particular asset allocation is 12%, the standard deviation of returns associated with the asset allocation is 10%, and the absolute risk aversion parameter is 2.0. the risk-adjusted expected return is identified below: expected return – [variance/risk tolerance], or 12%2 [(10%)2/ (1/2.0)]5 10%5 risk-adjusted expected return given the risk aversion measure, this fund manager would be indifferent between the portfolio and 10%, with certainty. this study expands the utility concept to include the final term in eqn. 9. given that the asset and liability have a positive relationship, cov[ra,rl] 0, the final term in eqn. 9 will increase utility. this final term is referred to as the liability hedging credit. to complete the example above, assume that this liability hedging credit is 2%. the risk-adjusted expected return is shown below: expected return – [variance/risk tolerance]1 liability hedging credit, or 12% 2 [(10%)2 / (1/2.0)]1 2%5 12%5 risk-adjusted expected return by including this credit, the fund manager is indifferent between the portfolio asset mix and a portfolio offering 12% with certainty, but with no hedging capabilities against liability fluctuations. the implications of the surplus framework model are that managers may choose a lower expected return or greater asset risk in order to increase the hedging capabilities of the portfolio against increases in liability values. in order to optimize the portfolio, from a surplus framework, it makes sense to choose assets that have a positive covariance with the plan’s liabilities. the liabilities associated with prepaid college tuition plans are driven by tuition inflation. a plan manager should include assets with a positive covariance with tuition inflation. u.s. treasury inflation-indexed securities are expected to have a positive relationship with tuition inflation. the degree of correlation may not be perfect; that is, the correlation coefficient may not be 1.0, but the two measures are expected to display positive correlation given the positive correlation between cpi and tuition inflation. this positive relationship is 266 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 apparent in fig. 2 and is calculated to be 0.87 using the college board’s college inflation index as a proxy for tuition inflation (chicago board of trade, 1997). in the next section, this study develops a comprehensive example using u.s. treasury inflation-indexed securities to enhance the liability hedging credit component inherent in the surplus framework model. 5. example of surplus framework model the trading history of u.s. treasury inflation-indexed securities is relatively limited. the first note was issued in january of 1997. an important feature of this security is that both the interest payments and principal amount are adjusted for inflation. the principal amount changes, or accretes, based on a predetermined lag and the reporting of cpi to reflect the present price environment. the interest payments are based upon the accreted principal amount. for instance, if the accreted principal is $1,200 and the note has a 3 3/8% coupon. the semi-annual interest payment is calculated as $1,200*(.03375/2)5 $20.25. the accreted principal amount changes to reflect the cumulative inflation effect since the security was issued by multiplying the original par value, $1,000, by the index ratio. this section explores the feasibility of including the inflation-indexed treasuries in asset allocation decisions for prepaid college tuition plans. the following section will introduce the framework for conducting an empirical investigation after sufficient data has been generated. the key element of the surplus framework methodology is the covariance term that provides the liability hedging credit. since college tuition is an annual variable, one observation per year, and the new inflation-indexed treasuries have a one-year history, an empirical investigation is not possible. for this study, a comprehensive example will be set forth that will allow managers to adopt a surplus framework in the management of prepaid college tuition plans. the necessary input variables are assumed below: ● assets today, a0 5 $1,000,000 ● liabilities today, l0 5 $1,000,000 ● new funds,a 5 $250,000 ● new liabilities generated,l 5 $250,000 ● college tuition inflation5 8% ● portfolio is limited to investments in s&p 500, u.s. government bonds, investment grade u.s. corporate bonds, and u.s. treasury inflation-indexed securities (also referred to as u.s. treasury inflation-protected securities or tips). ● using the college inflation index (cinf) as the tuition inflation variable and cpi as a proxy for tips, we calculate correlation coefficients for each asset with this liability. the results are included in table 2. given the information above, an optimal surplus framework model can be developed that not only considers portfolio return and standard deviation, but also includes the liability hedging credit. the portfolio return and standard deviation are measured using standard notation. the 267j.w. russell, r. brooks / financial services review 7 (1998) 257–271 return is an asset allocation-weighted measure. the portfolio return is the sum of the weighted returns of the individual assets. the standard deviation is measured as the weighted sum of the standard deviations with the corresponding weighted covariance terms, as in traditional portfolio theory. the surplus framework departs from the traditional portfolio concept in that the liability hedging credit (lhc) is added to the total utility. the following formulas summarize the relationship: portfolio expected return5 e(rp) 5 o i51 n xiri, for asset classes i,. . ., n (10) portfolio variance5 sp 2 5 o i51 n xi 2si 2 1 o i51 n o j51 n xixj covij , for all i,j } i þ j (11) portfolio standard deviation5 sp 5 îportfolio variance (12) liability hedging credit5 lhci 5 2 h l0 a0 covri,rl (13) table 2 example for surplus framework model asseti source returni std. dev.i rri,rl covri,rl lhci govt. bond 6% 8% 0.09 1.8 0.072 corp. bond 7% 12% 0.07 2.1 0.084 s&p 500 12% 20% 20.25 212.5 20.5 tips 6% 6% 0.87 13.05 0.522 the portfolio is limited to investments in u.s. government bonds (govt. bond), investment grade u.s. corporate bonds (corp. bond), s&p 500, and u.s. treasury inflation-indexed securities (also referred to as u.s. treasury inflation-protected securities or tips). return, standard deviation, and correlation coefficient measures are historical averages. the correlation coefficient the relationship between asset returns and the college inflation index (cinf). cpi served as a proxy for tips in the correlation coefficient calculation. rri,rl 5 correlation coefficient of returns for asset i with the plan liability covri,rl 5 covariance of returns for asset i with the plan liability liability hedging credit5 lhci 5 2 h l0 a0 vovri,rl liability hedging credit5 lhcportfolio 5 o i51 n xilhci the table above is based upon the following assumptions: ● standard deviation for the liability is 2.5 percent. this is the long-term historic volatility for the college inflation index. ● initial assets equal initial liabilities, i.e., l0/a0 5 1.0. ● risk tolerance equals 50, i.e., risk aversion equals 2. ● tips real returns are converted to nominal values for comparison purposes. 268 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 liability hedging credit5 lhcportfolio 5 o i51 n xilhci (14) utility 5 portfolio expected return2 risk penalty1 liability hedging credit (15) utility 5 e(rp) 2 (sp 2/h) 1 lhcportfolio (16) the goal of the college tuition plan manager is to maximize eqn. 16. this is a departure from the traditional portfolio format, whereby the manager maximized eqn. 10 subject to eqn. 12. using the formulas above and an iterative search for appropriate asset allocations, a simple model is developed to aid in this critical decision analysis. with the information from table 2, an asset allocation decision was performed. by maximizing the utility, as defined in eqn. 16, the optimal asset allocation is presented in table 3. the prepaid college tuition plan manager should allocate approximately 4.7% to government bonds, 19.8% to investment grade u.s. corporate bonds, 34.8% to the s&p 500, and 40.7% to u.s. tips. these allocations are driven by historic return and risk, the table 3 asset allocation results from surplus framework example the results below are based on the values in table 2 and were generated using the surplus framework model and an optimization program. note that the optimal asset allocations include a significant portion of the portfolio being invested in tips. although the return for tips is relatively modest the correlation with the liability makes this security attractive. the implication from this allocation is that managers may choose assets with low returns for their portfolios if the degree of correlation with the liability is strong enough to enhance the liability hedging credit. asseti allocation govt. bond 4.73% corp. bond 19.82% s&p 500 34.75% tips 40.69% portfolio expected return 8.28% std. dev. 7.76% lhc 0.06 utility 7.14% the portfolio is limited to investments in u.s. government bonds (govt. bond), investment grade u.s. corporate bonds (corp. bond), s&p 500, and u.s. treasury inflation-indexed securities (also referred to as u.s. treasury inflation-protected securities or tips). expected return and standard deviation on the portfolio are calculated in the usual manner. liability hedging credit5 lhci 5 2 h l0 a0 covri,rl liability hedging credit5 lhcportfolio 5 o i51 n xilhci utility 5 portfolio expected return2 risk penalty1 liability hedging credit utility 5 e(rp) 2 (sp 2/h) 1 lhcportfolio 269j.w. russell, r. brooks / financial services review 7 (1998) 257–271 correlation with the college inflation index, and the risk tolerance of the manager or investor. we allowed the key parameters; liability volatility, asset correlation with tuition inflation, and degree of risk aversion to change and performed sensitivity analyses for these key inputs. the results are included in table 4. the key observation for the surplus framework is that a plan manager should consider the risk and return merits of the assets, not on a stand-alone basis, but within the context of the liabilities generated by the plan. to the extent that a relatively low yielding asset has a high degree of positive correlation with the liability, it might warrant consideration from an asset allocation standpoint. the interpretation of the utility value is: a plan manager, with the risk tolerance and asset choices assumed in this example, would be indifferent between receiving the utility value, 7.1%, and an asset mix that provided 7.1% with no uncertainty and no ability to serve as a hedge against fluctuations in liability values. the plan manager can increase his/her risk-adjusted expected return by the value of the liability hedging credit as a result. 6. conclusion individuals have readily accepted prepaid tuition plans and the plans have received strong support in congress. given the rapidly escalating costs involved with attending college, individuals are searching for an alternative to indebtedness. prepaid tuition plans allow table 4 sensitivity analysis of model parameter values the sensitivity analysis allows the keys variables of liability volatility, risk aversion, and correlation between assets and liabilities to vary. in the tables below the liability volatility and risk aversion are chosen and the correlation coefficient,r, is allowed to vary. in each table an optimal allocation is presented for the four securities (indices) using the surplus framework model. note that in all of the allocation decisions, except one, tips should be included as a component of the portfolio. asseti rri,rl allocation rri,rl allocation rri,rl allocation liability standard deviation5 6%, risk aversion5 4 govt. bond .1 0% .1 13% .1 0% corp. bond .1 0% .1 27% .1 10% s&p 500 .2 38% .1 43% .5 61% tips .95 62% .1 17% .5 29% liability standard deviation5 20%, risk aversion5 2 govt. bond .1 0% .1 13% .1 0% corp. bond .1 0% .1 29% .1 0% s&p 500 .2 35% .1 46% .5 75% tips .95 65% .1 13% .5 25% liability standard deviation5 20%, risk aversion5 4 govt. bond .1 0% .1 13% .1 0% corp. bond .1 0% .1 34% .1 0% s&p 500 .2 27% .1 52% .5 100% tips .95 73% .1 2% .5 0 270 j.w. russell, r. brooks / financial services review 7 (1998) 257–271 individuals to purchase college tuition today for use at a later time. this reduces the dependency that has been formed between parents, students, and student loans. this paper provides an overview of the plans and suggests a framework for managing the tuition risk that has been shifted from individuals to states or other guarantors. although the focus of the surplus framework model is for prepaid tuition plan management the methodology may easily be implemented by individuals who manage their own college investment accounts. additionally, we suggest that u.s. treasury inflation-indexed securities be included in an asset allocation decision for tuition inflation management. these securities offer new alternatives for inflation-averse investors, corporations, and municipalities. this paper has shown that individuals and states may employ the new securities to facilitate in the management of tuition inflation risk. acknowledgment the author gratefully acknowledges the helpful comments of james ligon and robert mcleod. references chicago board of trade. (1997).inflation-indexed treasury note & bond futures and options: the reference and applications guide. chicago, il: chicago board of trade. ezra, d. d. (1991). asset allocation by surplus optimization.financial analyst journal(jan.–feb.), 51–57. fong, h. g. (1991). utilizing concepts of modern portfolio theory in an asset/liability management context.icfa continuing education(december), 14–18. fowler, w. j., jr. (1998).measuring inflation in public school costs. working paper series, paper no. 97-43. u.s. department of education, office of educational research and improvement. friend, i., & blume, m. (1970). measurement of portfolio performance under uncertainty.american economic review 60, 561–575. grossman, s. j., & shiller, r. j. (1981). the determinants of the variability of stock market prices.american economic review 71, 222–227. leibowitz, m. l., kogelman, s., & bader, l. n. (1992). asset performance and surplus control: a dual-shortfall approach.journal of portfolio managementvol, 28–37. macbeth, j. d., emanuel, d. c., & heatter, c. e. (1994). an investment strategy for defined benefit plans. financial analyst journal(may–june), 34–41. national association of state treasurers. (1998).special report on state college savings plans. peskin, m. w. (1997). asset allocation and funding policy for corporate-sponsored defined-benefit pension plans. journal of portfolio management(winter), 66–73. pratt, j. w. (1964). risk aversion in the small and in the large.econometrica 32, 122–136. sharpe, w. f., & tint, l. g. (1990). liabilities — a new approach.journal of portfolio management 16, 5–10. u.s. general accounting office. (1996). higher education: tuition increasing faster than household income and public colleges’ costs, report to congressional requesters. gao/hehs-96-154, august. 271j.w. russell, r. brooks / financial services review 7 (1998) 257–271 pii: s1057-0810(97)90019-6 from the editor karen eilers lahey the lead article by james s. ang and ah m. fatemi is entitled “personal bankruptcy costs: their relevance and some estimates.” numerous theoretical and empirical research articles have appeared over the years to explain corporate bankruptcy and its costs, and the popular press frequently cite bankruptcy court statistics to indicate that an ever increasing number of individuals elect to declare personal bankruptcy. however, there have not been any articles that seek to provide a theoretical basis for personal bankruptcy. ang and fatemi develop the first such theory and provide testable hypothesis of bank ruptcy costs. “an application of fuzzy set theory to the individual investor problem” by manuel tarrazo is based on a paper he presented at the 1996 academy of financial services. the article is a tutorial on fuzzy set theory and its applicability to asset allocation in terms of: (a) investor age, (b) type of assets that includes savings accounts, corporate bonds, common stock, and real estate, and (c) asset properties of liquidity, income, appreciation and safety. susan logan nelson and theron r. nelson report on a series of surveys that they have conducted to determine the public’s recognition of professional designations. their article entitled, “the use of professional designations in the real estate industry” contains a dis cussion of the goals of professional certification and the process of obtaining it. they note that the most recognized professional designation is the cpa, and that the cfp appears to be gaining increased recognition during the period of the study from 1991 to 1996. “an overview of financial services resources on the internet” by brian grinder pro vides a review of the academic literature on the impact of the internet on education and examples of current sites that can be used by financial service specialties. he constructs a stand-alone table that can be used as a convenient source of annotated information on spe cific websites. the table is organized by the following categories: (a) investments, (b) mutual funds, (c) banking/banking services, (d) financial planning/pension and retirement planning/estate planning, (e) tax planning, (f) real estate, (g) insurance, (h) financial coun seling, (i) employee benefits, and (i) education in financial services. julie a.b. cagle and gary e. porter examine the potential profit for depositors of investing in ipos. their article entitled, “conversions of mutual savings institutions: do initial returns from these ipos provide investors with windfall profits?’ finds that the initial stock offers are not priced differently than other financial institutions. the regular column, book, software and web site reviews, reviews a textbook and two websites. the textbook written by j. kimball dietrich is entitled the financial v vi financial services review 6( 1) 1997 services and financial institutions and is reviewed by james marchand. j. tim query examines the american risk and insurance association (aria) website and douglas r. kahl looks at the mutual fund investor’s center site. pii: s1057-0810(99)00035-9 investor partitioning of the components of value in corporate earnings announcements john a. macdonalda,*, david m. smithb adepartment of economics and finance, school of business, clarkson university, potsdam, ny 13699, usa bfaculty of finance, school of business, university at albany, suny, albany, ny 12222, usa abstract this study provides new insight into the market’s allocation of dividend-related and capital gains-based returns on common stock around earnings announcement surprises. to the extent that investors’ cash flow forecasts are revised as earnings surprises occur, americus trust prime and score returns reflect changes in respective future dividends and capital gains. about 70% of the value gain from positive surprises accrues in the capital gains (score) value adjustment, with expected dividends (primes) reflecting the remaining 30%. the relative proportion is greater in magnitude at the announcement of fiscal fourth quarter results when dividend changes are more likely to follow the quarter earnings announcement. © 1999 elsevier science inc. all rights reserved. jel classification:g120 keywords:asset pricing; primes; scores; earnings announcements 1. introduction it has long been known that corporate earnings surprises significantly affect returns to investors. researchers have examined the impact of quarterly and annual earnings announcements on stock prices, trading volume, average transaction size, option volatility, and option open interest, among other issues. a nearly universal conclusion in the literature is that earnings announcements have significant information content, and thus are important to investors in estimating future cash flows. the purpose of this study is to measure the impact of earnings surprises on the expected capital gains and dividend returns to common stockholders. americus trust securities provide a novel means by which to partition the components of return and to analyze the relative contribution of each component to total return. * corresponding author. tel.:11-315-268-3870; fax:11-315-268-3810. e-mail address:macdonaj@clarkson.edu (j.a. macdonald) financial services review 8 (1999) 87–99 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00035-9 the americus trust program, a derivative security innovation of the early 1980s, provided a means of splitting the return of dividend paying firms into its two basic sources. for about five years, stockholders in 27 large companies had the opportunity to tender their shares to the americus shareowner service corporation in exchange for a new security known as a unit. the certificate for each unit was perforated and divisible into aprime(prescribed right to income and maximum equity) and ascore(special claim on residual equity), each of which traded separately. primes gave investors the right to the firm’s quarterly dividend income plus a fixed termination value, on or before a stated termination date. a prime closely resembled a corporate bond, especially when the stock price exceeded the termination value. scores represented the right to a stock’s capital gains, that for americus trusts are the excess of the underlying stock’s price over the termination value. scores, that shared the primes’ termination dates, were much like long-term call options. up to 5% of each firm’s outstanding stock could be exchanged for americus trust securities. after 5% of a firm’s shares were tendered the trust closed, and the primes, scores, and units were obtainable only in the secondary market. tax rulings by the irs have discouraged the formation of any new trusts identical to the original series. the last remaining primes and scores expired in august 1992; however, the ephemeral existence of the trusts does not diminish their relevance to the financial markets, for several reasons. first, the popularity of long-term call options has increased greatly, as the major exchanges have listed leaps, options with expiration periods of up to 39 months, on over 300 individual stocks and about a dozen indexes. second, a new generation of similar trusts may emerge. any new americus trusts will, of course, have to comply with the irs ruling that doomed the first generation. the new securities may even be issued by the respective firms themselves, involve more than 5% of the stock, and thus enhance their importance relative to the original trusts. third, primes and scores provide a means of examining financial issues that would otherwise be more difficult or impossible to address. a case in point is research by venkatesh (1991), that shows how americus trust securities can be used to analyze dividend capture trading. finally, dual-purpose funds currently exist for investors that allow investors to select current income return or long-term return from funds of stocks and bonds. the present study uses americus trust securities to decompose the stock market’s response to earnings surprises into its constituent parts. the division of returns to investors attributable to expected dividends versus stock price changes is found to be roughly 30 versus 70%. analysis of these proportions shows them to be remarkably constant across a number of dimensions, including corporate payout policies, earnings estimation models, and abnormal return measures. 2. literature review and testable hypotheses 2.1. americus trust securities the finance literature has investigated primes and scores from a variety of perspectives. jarrow and o’hara (1989) examine the possible mispricing of primes and scores relative to the underlying stocks. after controlling for transaction costs, they find that the underlying 88 j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 stock consistently sold at a discount relative to americus trust instruments. the authors conclude that this disparity arises from regulatory constraints on short sales of common stock, with trading in primes and scores possibly serving as an attractive hedging alternative to short selling. thus, the effect that jarrow and o’hara observe may represent demand for americus trust securities driven by a paucity of hedging vehicles for firms’ stock. despite the pricing disparities noted, transaction costs are found to be large enough to prevent systematic gains from arbitrage. in another comparison of stocks and their americus trust counterparts, deshpande and jog (1989) compute average betas for stocks (1.07), primes (.45), and scores (5.06). the unusually high volatility for the scores probably reflects the fact that most scores are near the money (termination value) during the measurement period. of course, were a stock price well above the termination value, the score’s beta would more closely reflect that of the stock. venkatesh (1991) uses primes to document the existence of substantial dividend capture trading. he hypothesizes that because trading costs are greater for primes than for stocks, cum-ex trading is likely to manifest itself in the form of increased volume in stocks rather than primes. the empirical results of venkatesh suggest that the ex-dividend period trading volume of securities be related to their transaction costs. canina (1999) looks specifically at primes’ and scores’ price reaction to dividend change announcements. she reports a significant, positive 1.44% 3-day excess returns for scores and a significant, positive .89% 3-day excess returns for primes. without distinguishing scores for their moneyness, she concludes that dividend increases are viewed mostly as a long-term change to the earnings ability of the firm. 2.2. earnings announcements the informational value of earnings announcements has long been demonstrated. for instance, rendleman et al. (1982) confirm that the market rewards firms revealing positive earnings surprises, and penalizes those announcing lower-than-expected earnings. aharony and swary (1980) measure the announcement date earnings response coefficients (ercs) to be about1.52%, and 21 day (days210 through110) abnormal returns to be around 2.1%, suggesting that there is a strong information content to earnings announcements. as a further test of the information hypothesis, cready (1988) focuses on changes in the size of the average transaction around annual and quarterly earnings announcements. wealthy investors are assumed to have more information than other market participants do, and they also tend to trade in larger blocks. he discovers a strong positive relationship between mean unexpected transaction size and absolute price changes in the announcement week, again consistent with the hypothesis that earnings announcements are important information events. the generally accepted explanation for such findings is that the market uses earnings revelations to re-evaluate a firm’s expected future cash flows, and investor behavior and stock prices adjust accordingly. it is not clear, however, whether the stock price adjustment results from changes in the market’s expectations of future dividends, capital gains, or both. kane et al. (1984), for example, report a positive relationship between earnings and dividend changes. in the sample of kane et al., 60% of earnings changes are accompanied 89j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 by dividend changes in the same direction. consequently, if an earnings increase is announced, the value of a company’s prime (the dividend-bearing portion of an americus trust unit) should rise. such earnings information might not have immediate cash flow implications for prime holders if related dividend declarations precede rather than follow earnings announcements; however, in aharony and swary’s sample of 3399 earnings announcements, associated dividend declarationsfollow earnings revelations 77% of the time. thus, it is expected that an unanticipated earnings change would have a direct impact on the prime’s returns. further, because stock prices are observed to change significantly in response to quarterly earnings announcements, it is expected that the score’s price would be affected similarly. the association between earnings and stock prices is further explored by beaver et al. (1997), who show that the earnings/stock price relationship can be modeled using a simultaneous equations approach. their analysis, however, is hampered by uncertainty over which instrumental variables should be used in the necessary multi-stage regressions. additional empirical work on ercs by kallapur (1994) and sivakumar and waymire (1993) shows that the market may view the level of a firm’s dividend to be a sign of the credibility it can assign to earnings announcements. kallapur finds ercs to be positively related to the announcing firm’s dividend payout ratio. sivakumar and waymire examine 51 firms between 1905 and 1910, when earnings announcements were discretionary. they observe significant positive returns for dividend-paying firms, but not for non-dividendpaying firms. dividend increases are associated with greater stock price movements than are earnings increases, suggesting that the market considers earnings to be more credible if accompanied by dividend changes. hayn (1995) shows that the ercs for profits and losses are asymmetric. the coefficient is smaller for losses than for profits, suggesting that the practice of pooling profits and losses has the effect of driving down computed ercs. 2.3. time sensitivity of results lang (1991) shows that the response of ipo stock prices to earnings surprises tends to vary across time, depending on the level of uncertainty surrounding the firm’s earnings. the data set in the present study is composed of firms that are substantially different from lang’s ipo sample. nonetheless, lang’s results raise the question of whether the proportional reactions of primes and scores are constant over time. the time to expiration of each trust at the date of an earnings surprise is used as the measure of time sensitivity. the more dividend payments a firm has remaining before trust expiration, the more chances exist for managers to revise the amount. on the other hand, additional time also means that the score has the opportunity to rise further into the money, providing enhanced capital appreciation. the net effect will be determined by the relative discount rates that investors apply to dividend income versus capital gains, an unobservable factor. 2.4. moneyness scores share several important characteristics with call options. the termination value for each trust serves as ade factoexercise price. the concept of moneyness (the extent to which 90 j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 a score is inor out-of-the-money) is particularly important to this study. the moneyness for the score of firm j on day t is calculated as moneynessjt 5 stock pricejt 2 termination valuej termination valuej (1) recall that the prime’s value includes all dividends plus an amount up to a termination value. if the stock price isbelow the termination value at the time of a value-increasing, positive surprise, then the prime’s return will reflect both the increased expected future dividends plus the capital appreciation toward the termination claim. such observations are likely to introduce error into the present analysis, because in theory the prime will reflect a pure dividend stream and the score a pure capital gain return only if the score is in or at the money. hence, only scores that are in the money during the entire parameter estimation and event period are included in the final sample. some noise doubtless remains, because the prime can never represent aperfectdividend stream. even when the stock price is above the termination value, some of the prime’s return during earnings surprises will still reflect whether the prime holder’s receipt of the termination claim is made more certain or less certain as a result of the earnings news. the safety effect for primes associated with in-the-money scores has an analogy elsewhere in corporate finance. jin (1992) finds that even though the promised payments remain fixed for corporate bonds, the returns for such securities tend to be highly correlated with the sign of corporate earnings surprises. in the present study, the upshot of this safety effect is to overstate somewhat the dividend income component of return. the proportional return from primes should thus be viewed as an upper bound dividend return. even with out-of-the-money scores eliminated, there remains a high degree of crosssectional variability in moneyness among surviving observations. the prime’s safety effect will become smaller when a trust’s score is deep in-the-money, thus causing the score’s proportional return to be higher relative to the prime. a countervailing effect could arise from the propensity of stock prices toward mean reversion, as discussed by de bondt and thaler (1985, 1987). in order for a score to be deep in the money, the stock price must have risen significantly since the trust’s inception. mean reversion behavior would suggest that a price decline is more likely for stocks in the sample than for stocks whose price had not risen; however, de bondt and thaler note that the phenomenon is far less striking for prior winners than for losers. consequently, mean reverting tendencies are expected to have only a minor impact on in-the-money scores, and by extension on the division of returns between prime and score. 2.5. fiscal quarter effects jones and bublitz (1990) address the question of higher-than-expected errors in earnings forecasts found to persist in fiscal fourth quarters. non-recurring items are much more likely to be included in that quarter’s reported income than in other quarters’. accordingly, the market’s average reaction to unexpected fourth quarter earnings is somewhat muted. such a finding raises the question of whether the present study’s results might be driven by its fourth 91j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 quarter observations. if fourth quarter earnings after extraordinary items are more variable than those of other quarters, it follows that permanent dividend changes should not be as strongly tied to earnings in that quarter. thus, one should expect a greater relative return from scores versus primes around fourth quarter earnings announcements. 3. data and methodology the primary data for this study cover 26 americus trusts. the original at&t trust, established in 1983, is eliminated from the sample due to a paucity of historical earnings information for the newly restructured firm. gte’s trust is limited to the sample period before its acquisition of contel in march 1991. there are 459 earnings announcements in the initial sample, representing the time period december 1, 1985 (when only the exxon trust existed) through july 31, 1992 (when only the american express trust remained). the present study’s requirement that scores be in-the-money results in the loss of 323 observations, meaning that nine trust firms are not represented at all in the sample. quarterly primary earnings per share before extraordinary items are obtained from the compustat database from the first quarter of 1972 through the expiration of each trust. the earnings variable used is item number q19 (epspxq) in compustat pc-plus. for each observation, the analysis requires that there be at least one dividend declaration remaining between the earnings announcement date and the expiration date of the trust. the wall street journal indexis used to identify earnings announcement dates and to check for potentially confounding events over a two-day announcement period. events such as dividend declarations, merger news, and securities offerings are considered to be potential contaminants, that result in the deletion of 37 observations. the final sample consists of the 99 surviving data points. panel a of table 1 shows the time distribution of the sample. most observations are found after 1989, due to the fact that several firms’ stocks did not rise above the termination value until late in the sample period. to evaluate the market’s response to earnings surprises, cumulative abnormal returns (cars) are examined for the underlying stock and its corresponding prime and score from the americus trusts. abnormal returns are generated as the difference between a security’s actual daily return and its expected return. the expected return is the return for the crsp equally weighted index. the market adjusted returns methodology is used in lieu of the market model approach, because option betas can be quite unstable even when the underlying stock’s beta is constant (see cox and rubinstein, 1988, p. 190). day21 is defined as the public earnings announcement date, with the publication of the news inthe wall street journal taking place on day 0 in many cases. significant cars would indicate whether the market sees earnings surprises as having long-term effects reflecting increased expected dividend streams (significant cars would be found for stocks and primes), capital appreciation/depreciation effects (significant cars for the stocks and scores only), or implications for cash flows from dividends as well as price changes (significant responses in all return series). the market’s reaction to an event is expected to reflect the sign and magnitude of the earnings surprise. following kane et al. (1984), schachter (1988), and easton and zmijewski 92 j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 (1989), percentage earnings announcement surprises are estimated using the methodology of foster (1977). first, for each firm j, one year changes in earnings are regressed on the one quarter lag of the change in one year’s earnings, as noted in eq. (2): [epsj,t 2 epsj,t24] 5 aj 1 bj [epsj,t21 2 epsj,t25] 1 ej,t, (2) where epsj,t is the earnings per share before extraordinary items for firms in quarter t, and ej,t is a random error term assumed to be normally distributed. using the parameters estimated in eq. (2), expected earnings are computed with eq. (3), as follows: [epsj,t11 2 epsj,t23] 5 aj 1 bj [epsj,t 2 epsj,t24]. (3) the percentage earnings surprise is measured as the difference between the actual earnings in any period and the expected earnings for that period divided by the absolute value of expected earnings. table 1 sample description company 1986 1987 1988 1989 1990 1991 1992 total panel a: chronological distribution of earnings announcements american home products 0 0 0 2 4 0 0 6 amoco 0 0 0 0 2 0 0 2 bristol myers squibb 0 0 0 0 2 1 0 3 chevron 0 0 0 0 1 0 0 1 coca cola 0 0 0 2 4 0 1 7 dow chemical 0 0 4 4 0 0 0 8 dupont 0 0 0 0 0 2 0 2 exxon 1 4 4 4 3 0 0 16 ford motor co. 0 0 1 1 0 0 0 2 general electric 0 0 0 0 1 0 0 1 gte 0 0 0 4 4 0 0 8 johnson & johnson 0 0 0 0 2 1 1 4 mobil 0 0 0 0 2 1 1 4 merck 0 0 0 3 4 3 0 10 philip morris 0 0 0 3 4 3 3 13 procter & gamble 0 0 0 2 4 4 1 11 union pacific 0 0 0 0 0 0 1 1 total 1 4 9 25 37 15 8 99 variablea mean median sd panel b: other sample characteristics time to expiration (years) 1.93 years 1.93 years 1.04 years moneyness 40.53% 29.66% 41.37% earnings surprise foster model 10.09% 5.50% 35.16% naive model 13.73% 16.48% 39.80% long-term dividend payout ratio 56.17% 54.53% 3.61% a moneyness is the percentage premium over termination value. the earnings surprises are the percentage deviation from expected earnings using foster’s model and a naive model. the long-term dividend payout ratio is computed as the average (from 1983 through the date of each earnings surprise) of the firm’s quarterly dividend divided by the prior quarter’s earnings before extraordinary items. 93j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 for comparison purposes, and following aharony and swary (1980) and damodaran (1989), a naive earnings estimation model also is used. canina (1999) also relied on a naı¨ve expectations model for dividends. her findings are congruent with our own. expected earnings are thus defined as the earnings from the corresponding quarter in the fiscal year: e(epsj,t) 5 epsj,t24, (4) and surprises are computed as for the foster model. panel b of table 1 provides summary statistics for the data set, including information about time to expiration of the trusts, moneyness of the scores, the size of the earnings surprises, and dividend payout ratios for each firm. many of the scores are far in the money, with the median stock price exceeding the termination value by almost 30%. the mean time to expiration is about two years, so each trust has about eight dividend payments remaining. other issues remain. the fact that the score holders realize the greatest price impact reflects the high elasticity of this option-like security; however, the low price of most scores relative to the termination value makes thedollar return impact much smaller than the percentage return would suggest. for example, on october 12, 1989, the closing market prices for primes and scores of coca cola corp. were $44.70 and $25.10, respectively. the next day, the firm announced a positive earnings surprise of 11.0% relative to the expectation generated by foster’s model and 25.9% based on a naive model. after adjustment for market factors, the prime’s abnormal return in the two days in the earnings announcement period is 3.5% and the score’s abnormal return is 21.3%. dollar returns to coca cola investors, however, are $1.58 for the prime and $5.35 for the score. the difference in the dollar returns is obviously much smaller than the six fold difference in percentage terms. an analysis of the proportional price-based abnormal returns to stockholders must include an adjustment that considers the relative values of the prime and score. hence, the dollar returns to prime holders, for example, are best represented by multiplying the prime’s car by the proportion of the total stock value the prime constitutes. formally, acarpr 5 carpr p vpr vpr 1 vsc (5) where vpr and vsc are market closing prices of the two securities the day before the earnings announcement is made. the score’s acar is computed analogously. to address the hypotheses developed earlier with regard to time to expiration, moneyness, fourth quarter dividend announcements, and the magnitude of the surprise, the following multiple regression model is used to analyze cross-sectionally the division of returns between primes and scores: ratiom 5 a0j 1 a1j(ttexp)1 a2j(money) 1 a3j(qtr4) 1 a4j~surprise! 1 «j. (6) ratiom is the ratio of the proportional prime return divided by the proportional return for the score, with returns generated by the market-adjusted return method. another regressand, 94 j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 ratioc, is computed identically, but is derived from the comparison period abnormal returns approach. ttexp is the time to expiration of each trust (in years), and money is the percentage by which each firm’s stock price exceeds the trust’s termination value. qtr4 is a dummy variable assigned a value of 1 if the earnings result is from the fiscal fourth quarter and 0 otherwise, and surprise controls for the magnitude of the positive or negative earnings surprise relative to the value predicted by foster’s model or the naive model. 4. results 4.1. general findings observations in the entire population of earnings announcements are classified as positive or negative surprises, and then the security returns are examined. table 2 indicates that for the screened sample, the prices of stocks, primes, and scores all react strongly to announcements of positive earnings surprises using a naı¨ve model. negative earnings surprises are associated with statistically insignificant abnormal returns to all three securities. when classified using foster’s (1977) specification, positive surprises induce statistically significant two-day abnormal returns for all three security types. thus, earnings surprises clearly alter the market’s perception of the sign and magnitude of all future cash flows, from both dividends and capital gains. empirical results in table 2 show that the response to earnings surprises is relatively consistent across the two earnings expectation models. as the samples overlap significantly, this observation may say as much about the classification consistency (i.e., discrimination between positive versus negative surprises) of the models themselves as it does the robusttable 2 abnormal returns around earnings announcements for americus trust firmsa security sign of earnings surprise positive negative car (%) t-statistic n car (%) t-statistic n foster’s (1977) classification model stock 0.32b 2.83*** 73 20.13 20.24 26 prime 0.16 1.46* 73 0.40 0.42 26 score 1.41 2.69*** 73 21.63 20.20 26 naive classification model stock 0.26 1.58* 76 20.05 20.09 23 prime 0.11 0.99 76 0.51 0.56 23 score 1.20 2.27** 76 21.36 20.17 23 a this table contains two-day cumulative abnormal returns (cars) for stocks, primes, and scores around the time of earnings announcements. the abnormal return technique used is the market-adjusted returns approach. b all abnormal returns are for days21 and 0. * statistically significant at .10 level, one-tailed test. ** statistically significant at .05 level, one-tailed test. *** statistically significant at .01 level, one-tailed test. 95j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 ness of the observed proportions. the cars and adjusted cars reported in tables 2 and 3 are based on the pre-announcement market values of the primes and scores. the combined value of the two securities is within arbitrage bounds of equaling the stock’s value (see jarrow and o’hara, 1989). as table 3 indicates, at the time of positive surprises, prime holders receive roughly 30% of the total return whereas score holders reap about 70%. these findings are quite robust across other earnings estimation models and abnormal return metrics. for the firms in this sample, then, positive earnings surprises induce significant changes in investor expectations of future capital gains, and a much smaller revision in perception of the firm’s future dividend policy. this evidences dividend stickiness, that means that dividends are linked to earnings somewhat loosely and with a time lag. 4.2. fiscal quarter effects table 4 shows the results for the regression model using both market-adjusted and comparison period returns. the ols coefficients in panel a indicate that the proportional contributions of primes and scores to the total return around positive earnings surprises do not differ significantly across in-the-moneyness and expiration states. likewise, the magnitude of earnings surprises does not seem to affect the division of returns. table 3 adjusted cumulative abnormal returns for share equivalent investment in americus trust securitiesa security sign of earnings surprise positive negative acar (%) proportion n acar (%) proportion n naive classification model prime 0.0864 0.2913 73 0.7358 0.3397 26 score 0.3865 0.7087 73 20.3100 0.6603 26 foster’s (1977) classification model prime 0.1209 0.2886 71 0.3047 0.3804 28 score 0.4249 0.7164 71 20.3720 0.6196 28 a adjusted cumulative abnormal returns for primes (acarpr) are two-day cumulative abnormal returns to primeholders times [vpr/(vpr 1 vsc)], where the vpr and vsc are the closing prices on the day before an earnings announcement. for scores, acarsc equals market model abnormal returns to scoreholders times [vsc/(vpr 1 vsc)]. proportion for primes is the acarpr/(acarpr 1 acarsc). proportion for scores is defined analogously. the following example attempts to clarify the concept of acars and proportional return: prime score stock market value $10 $20 $30 car 10% 7% 8% acar 3.33% 4.67% 8.00% proportional return 41.67% 58.33% 100.00% viewing common stock as a portfolio of a prime and score, the adjusted car for the prime is 10% times the prime’s weight in the portfolio (one-third). the score’s acar of 4.67% equals the 7% return to the score times its portfolio weight (two-thirds). 96 j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 the only statistically significant coefficient is on the dummy variable for the fiscal quarter. the coefficient’s sign is positive, that is unexpected. one possible explanation stems from the casual observation that firms tend to make dividend changes immediately after or concurrent with announcements of fourth quarter earnings results. typically, fourth quarterinduced changes ought to be reflected in dividends that are actually paid in the first fiscal quarter. an examination of compustat quarterly dividend data from 1983 through 1992 strongly confirms this suspicion. in the 1983 through 1992 period there are 291,281 quarterly dividend changes (initiations, omissions, increases, and decreases) among compustat’s active firms. relative to the previous fourth quarter’s dollar payout, dividends changed for 14.6% of firms in fiscal first quarters. on average, 12.2% of firms changed their payments during each of the other fiscal quarters. the wilcoxon rank sum statistic shows the two percentages to be significantly different at the .01 level (one-tailed test). with dividend changes being most likely to follow fourth quarter earnings realizations, it is not surprising to find that the prices of primes are affected disproportionately when year-end earnings are announced. 5. conclusions americus trust primes and scores represent unique vehicles by which to view the components of return. this study provides new insight into the market’s allocation of table 4 analysis of adjusted cumulative abnormal returnsa intercept ttexp money qtr4 surprise f adj-r2 panel a: foster’s model surprises dependent variable: ratiom 22.2064 0.5834 1.1841 1.7700 20.9278 0.92 0.00 (21.48) (1.07) (0.89) (1.57) (20.67) dependent variable: ratioc 20.7945 0.1976 0.2844 1.4648 20.8139 1.56 0.02 (20.95) (0.648) (0.382) (2.323)* (21.044) panel b: naive model surprises dependent variable: ratiom 22.3284 0.6411 1.2922 1.6537 20.7333 0.89 0.00 (21.50) (1.08) (0.95) (1.47) (20.55) dependent variable: ratioc 20.9144 0.2561 0.3884 1.3597 20.6817 1.49 0.0197 (21.05) (0.769) (0.508) (2.151)* (2.907) a the dependent variable in each of the following regressions is the ratio of proportional abnormal returns associated with americus trust primes and the proportional returns for scores. ratiom is for market-adjusted returns and ratioc is for comparison period abnormal returns. ttexp is the time to expiration of each trust (in years); money is the degree to which each firm’s stock price exceeds the trust’s termination value (in percent); qtr4 is a dummy variable taking on a value of 1 if the earnings result is from the fiscal fourth quarter and 0 otherwise; surprise is the magnitude of the positive or negative earnings surprise relative to the value predicted by foster’s [13] model or a naive model. below each parameter estimate, the t-statistic is given in parentheses. * statistically significant at the .05 level. 97j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 dividend-related and capital gains-based returns on common stock at the arrival of new earnings information. to the extent that investors’ cash flow forecasts are revised as earnings surprises occur, americus trust prime and score returns reflect changes in respective future dividends and capital gains. cumulative abnormal returns to both primes and scores are found to be positive around positive earnings surprises, though no significant reaction is detected around negative earnings events. it seems, then, that investors adjust their expectations of both current payout and capital gains when there is good news. for the large capitalization firms in the americus trust program, about 70% of the value gain from the surprise accrues in the capital gains (score) value adjustment; with expected dividends (primes) reflecting the remaining 30%. these relative proportions are found to be robust across multiple earnings estimation models, over time, and across score ator in-the-moneyness levels. this is in line with canina (1999), although her implied 61.8% impact to the long term valuation for dividend announcements does not control for announcement quarter timing or moneyness of the scores. recall that the ratio of prime and score returns for fourth quarter earnings surprises is significantly higher than those in other quarters. this study’s results contribute to the corporate earnings literature by showing the precise nature of investors’ cash flow revisions around earnings surprise announcements. one limitation is that the findings may be less relevant to smaller growth firms, as the present sample necessarily consists of large companies. also, because it does not represent a perfect dividend-related return, the prime’s relative return (30%) should be viewed as anupper bound proportional dividend contribution. still, the results for the study are remarkably robust to methodology used. research such as this can provide designers of new financial instruments with a better understanding of the behavior and relative pricing characteristics of the securities being engineered. these results raise interesting implications for portfolio management. for investors and financial planners, these results suggest that allocation of investment dollars to stocks may be best done by also considering the variance of forecast error. that is, firms that frequently have earnings surprises, may better be represented in the portfolio around announcement time by derivatives than the stock itself. this would be especially true in the case of building or investing in a dual purpose fund. also, these findings suggest a more efficient strategy in capturing earnings announcement surprises may be through the use of leaps than listed short-term options. examination of these extensions is left to future research. acknowledgments the authors wish to thank the editor, two anonymous reviewers and linda canina for constructive suggestions and discussion. the authors also thank nusret caxici and ashok vora for valuable comments on previous drafts. patti wendt provided excellent research assistance. 98 j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 references aharony, j., & swary, i. (1980). quarterly dividend and earnings announcements and stockholders’ returns: an empirical analysis.journal of finance, 35(1), 1–12. beaver, w., mcanally, m. l., & stinson, c. h. (1997). the information content of earnings and prices: a simultaneous equations approach.journal of accounting and economics, 23(1), 53–81. canina, l. (1999). the market’s perception of the information conveyed by dividend announcements.journal of multinational financial management, 9(1), 1–13. cox, j. c., & rubinstein, m. (1988).options markets.upper saddle river, nj: prentice hall. cready, w. m. (1988). information value and investor wealth: the case of earnings announcements.journal of accounting research, 26(1), 1–27. damodaran, a. (1989). the weekend effect in information releases: a study of earnings and dividend announcements.review of financial studies, 2(4), 607–623. de bondt, w. f. m. (1987). further evidence on investor overreaction and stock market seasonality.journal of finance, 42(3), 557–581. de bondt, w. f. m., & thaler, r. (1985). does the stock market overreact?journal of finance, 40(4), 793–805. deshpande, s., & jog, v. (1989). primes and scores.aaii journal, xi(7), 15–19. easton, p. d., & zmijewski, m. e. (1989). cross-sectional variation in the stock market response to accounting earnings announcements.journal of accounting and economics, 11(2–3), 117–141. foster, g. (1977). quarterly accounting data: time series properties and predictive ability results.accounting review, 52(1), 1–20. hayn, c. (1995). the information content of losses.journal of accounting and economics, 20(2), 125–153. jarrow, r. a., & o’hara, m. (1989). primes and scores: an essay on market imperfections.journal of finance, 44(5), 1263–1287. jin, j.-d. (1992). the relationship between accounting earnings and bond returns.journal of accounting and public policy, 11(3), 245–267. jones, c. p., & bublitz, b. (1990). market reactions to the information content of earnings over alternative quarters.journal of accounting, auditing, and finance, 5(4), 549–566. kallapur, s. (1994). dividend payout ratios as determinants of earnings response coefficients: a test of the free cash flow theory.journal of accounting and economics, 17(3), 359–375. kane, a., lee, y. k., & marcus, a. (1984). earnings and dividend announcements: is there a corroboration effect?journal of finance, 39(4), 1091–1099. lang, m. (1991). time-varying stock price response to earnings induced by uncertainty about the time series process of earnings.journal of accounting research, 29(2), 229–257. rendleman, r. j., jones, c. p., & latane´, h. a. (1982). empirical anomalies based on unexpected earnings and the importance of risk adjustment.journal of financial economics, 10(3), 269–287. schachter, b. (1988). open interest in stock options around quarterly earnings announcements.journal of accounting research, 26(2), 353–372. sivakumar, k. n., & waymire, g. (1993). the information content of earnings in a discretionary reporting environment: evidence from nyse industrials, 1905–10.journal of accounting research, 31(1), 62–91. venkatesh, p. c. (1991). trading costs and ex-day behavior: an examination of primes and stocks.financial management, 20(3), 84–95. 99j.a. macdonald, d.m. smith / financial services review 8 (1999) 87–99 pii: s1057-0810(97)90004-4 financial services review, 6(4): 257-269 copyright 0 1997 by jai press inc. mn: 1057-0810 all rights of reproduction in any form reserved. performance of mutual funds before and after closing to new investors herman mar&cyan i&ton0 liano this study examines the decision to close mutual funds to new investors due to the growth of the funds’ assets. the evidence indicates that funds perform better three years prior to closing to new investors than they do afterwards. furthermore, the evi dence indicates that the closed funds outperform the control port$olios of funds with similar investment objectives and asset size during the oneand three-year periods prior to closing. however, there is no significant difference in the pe$ormance of closed finds and their matched control porcolios during the oneand three-year peri ods after closing. although the primary reason given for closing the finds is the desire to maintain performance in the face of growing assets, the strategy does not appear successful in accomplishing this objective. i. introduction since treynor (1965), sharpe (1966), and jensen (1968), numerous studies have been con ducted to examine the persistence of mutual fund performance. the results remain ambig uous. lehmann and modest (1987), grinblatt and titman (1992), hendricks, patek and zeckhauser (1993), goetzmann and ibbotson (1994) brown and goetzmann (1995) elton, gruber, and blake (1996), and gruber (1996) conclude that past performance is a good indicator of future performance. to the contrary, jensen (1968) malkiel (1995), kahn and rudd (1995), cat-hart (1997), and phelps and detzel(1997) find no evidence of persistence in mutual fund performance. an important issue regarding the performance of mutual funds remains unexamined. the issue concerns the customary practice in the mutual fund industry to close funds to new investors for various reasons, some financial, and some organizational. the decision to close to new investors may be permanent or temporary, depending on the reason for the original decision. frequently, the decision to close a fund to new investors is related to the herman manakyan l western kentucky university, department of accounting and finance, 1 big red way, bowling green, ky 42101; e-mail: herman.manakyan@wku.edu. kartono liano l mississippi state university, department of finance and economics, p.o. box 9580, mississippi state, ms 39762-9580; e-mail: kliano@cobilan.msstate.edu. 258 financial services review 6(4) 1997 fund’s asset size. the most common explanation given for closing funds to new investors is the difficulty in effectively managing a large portfolio to earn superior returns for the investors. presumably, once a portfolio reaches a critical threshold of size, it becomes “too bulky” to manage effectively, and it is difficult to find attractive investments to which new funds can be diverted without sacrificing performance. if this explanation is valid, mutual funds that close to new investors should be able to at the very least maintain performance after the decision to close. however, the views on closing mutual funds to new investors are far from unanimous in the mutual fund industry. while some investment companies follow this practice regu larly, there seems to be no consistency in how large a fund is allowed to grow prior to clos ing. other investment companies eschew this practice completely, allowing the funds to grow without limit, while continuing to post enviable investment results for their clients. an interesting example is fidelity magellan fund, the largest of all mutual funds, which was allowed to grow in excess of $60 billion in assets and for many years posted impres sive rates of return far exceeding the returns generated by smaller, more nimble funds. magellan was finally closed to new investors effective september 30, 1997. the existing literature on mutual funds provides no empirical evidence on the impact of this widespread practice on mutual fund performance. the purpose of this study is to empirically examine whether the decision to close mutual funds to new investors in order to limit the growth of assets is justified by the subsequent performance of the funds relative to both their own historical performance and the performance of their peers. ii. methodology and data in order to examine the impact of closing mutual funds to new investors, it was first neces sary to identify a sample of mutual funds that are closed to new investors, along with the date of and the stated reason for closure. the january 1996 version of the morningstar on disc data base was screened to identify all funds with purchase constraint codes of l (closed to all investors) and c (closed to new investors). the resulting 128 funds were then contacted to inquire about the date on which they stopped accepting new investors and the reason given for closing. in order to be able to examine the performance of the funds for a period of at least one year on either side of the closing date, it was necessary to limit the study to funds that closed between 1978 and 1994. of the 63 funds that closed to new investors during this time period, 38 indicated that the decision to close was related to the fund’s asset size, while 25 funds expressed other motives for closing, the most frequent of which was a change in the fund’s fee structure. the sample was reduced further due to some of the funds having closed and re-opened to new investors within the analysis win dow, resulting in overlapping periods, and due to lack of data. the final sample consisted of 27 funds that allowed an analysis of a one-year window around the closing date, and 12 of those funds had sufficient data for the analysis of a three-year window around the clos ing date. the funds included in the analysis are identified in table 1, along with their investment objectives and month of closing to new investors. the same source was used to form matching control portfolios for each fund in the sample. the control portfolios consisted of the five funds within the same investment objective with net assets closest to the sample fund, which were open to new investors and performance of mutual funds 259 had returns data available for the relevant analysis window. the monthly returns of the sample funds and the matching control portfolios for the one (three) year(s) surrounding the date of closing were then extracted from mornzngstar. the three-month t-bill returns were extracted from the federal reserve bueletin, and the monthly returns of the s&p 500 index were extracted from the security price index record. the impact of the decision to close the funds to new investors was examined by com paring mean raw returns, as well as widely used risk adjusted portfolio performance mea sures for the l-year and 3-year periods prior to and following the closing of the fund to new investors. sharpe’s index was used to measure performance adjusted by total risk, and treynor’s index was used to measure performance adjusted by systematic risk. jensen’s alpha was included as an additional indexed performance measure. fabozzi, francis, and lee (1980) demonstrate that the jensen performance measure is robust when returns are measured monthly. in addition, fabozzi and francis (1979) conclude that jensen’s alpha is not influenced by bull and bear markets. however, barber (1994) finds that stock mutual funds’ future returns are not related to the historical beta of the fund. since the sample includes fixed income and foreign funds, the 01, l3, and treynor’s index for these funds should be interpreted with caution, as the s&p 500 index was used as the market return proxy for all sample funds for consistency. symbol table 1 sample of mutual funds closed to new investments fund name objective closing month acinx acrnx argfx babex barix cpgrx cpsfx idfdx javlx javtx krfbx lomcx mmhyx mnscx montx mpscx msiqx nefgx pjigx popax popcx sequx sksex ssrsc stcsx vwndx van kampen am cap ltd mat b acorn international acorn ariel growth babson enterprise baird adjustable rate income chesapeake growth comstock partners strategy 0 idex 3 janus twenty janus venture kemper retirement ii cgm capital development mfs municipal high-income a montgomery small cap monetta mas small cap value morgan stanley instl intl equity new england growth a piper jaffray instl govt pimco adv opportunity a pimco adv opportunity c sequoia skyline special equities state st research sm cap grc strong common stock vanguard windsor adi. rate mtg. 5193 foreign _ 2194 small company 7190 small company 4190 small company ii92 adj. rate mtg. 12194 growth 12194 mult-asst glbl 7192 growth 6190 growth l/93 small company 9191 balanced 3192 growth 11186 muni nat’1 6185 small company 3192 small company 3193 small company 9194 foreign 6193 growth l/92 gvt mortgage 6194 aggr growth 12192 aggr growth 12192 growth 12182 small company 12192 small company 1 l/94 small company 3193 growth-income 5/85 260 financial services review 6(4) 1997 in addition to the raw return (unadjusted for risk), each of the three risk-adjusted per formance measures were employed to conduct three types of comparison for a total of twelve comparisons: (1) oneand three-year performance of each closed fund relative to itself (prevs. post-closing). (2) oneand three-year pre-closing performance of each closed fund relative to the control portfolio. (3) oneand three-year post-closing performance of each closed fund relative to the control portfolio. due to the small sample size, the non-parametric wilcoxon rank sum test, which is more powerful for small samples of unknown distribution (gibbons, 1985, p. 193), is used for these pairwise comparisons. iii. results table 2 presents the performance measures (the raw return, sharpe’s index, jensen’s alpha, and treynor’s index), as well as beta, and r* values for the 27 funds during the 12 months before and after the closing of a fund to new investors. as expected, the equity funds in the sample have beta values near one, and fixed income and foreign funds have lower betas. the mean returns during the 12 months before closing are higher than the mean returns during the 12 months after closing for 16 of the 27 funds. the average monthly raw return 12 months prior to closing is 1.63% and is significantly different from zero at the i % level. in comparison, the average monthly raw return 12 months after closing is 1.17% and is sig nificantly different from zero at the 1% level. on the average, mutual funds generate an additional annual return of 5.64% before closing to new investors than after closing the funds ([l + (0.016268-0.011684)]‘* 1). the non-parametric wilcoxon rank sum test reveals that the raw return during the year before closing is not significantly different from the raw return during the year after closing of the funds. furthermore, the average jensen’s alpha before closing is positive and signif icantly different from zero at the 1% level, implying that before closing to new investors, the funds beat the market portfolio, as measured by the s&p 500 index. however, the jensen’s alpha after closing the funds is not significantly different from zero, suggesting that the funds fail to beat the market portfolio after closing to new investors. based on raw returns, the sharpe’s index, jensen’s alpha, and treynor’s index, 16, 11, 19, and 16 of the 27 funds, respectively, performed better during the 12 months preceding the closing of the fund than the 12 subsequent months. however, the non-parametric wil coxon rank sum test indicates that the performance of the sample funds in the 12 months preceding their closing is not statistically superior at the usual significance levels relative to their performance after closing to new investors. the results are similar when comparing performance in the 36 months preceding and following the closing of the funds for the 12 funds for which the three-year comparison was possible. the mean returns are higher during the three years before closing than the three years after closing for 10 of the 12 funds. the average monthly raw returns during the three t a b l e 2 t he i y ea r pe rf or m an ce su m m ar y b ef or e an d a ft er c lo si ng b ef or e a ft er i r aw sh ar pe je ns en t re yn or b et a r 2 r aw sh ar pe je ns en t re yn or b et a r 2 4 a c ft x 0. 20 75 -0 .2 49 4 -0 .0 48 6 1. 52 44 -0 .0 38 0 0. 11 36 -0 .0 28 3 -0 .8 04 1 -0 .2 87 2* ** -2 .5 10 6 0. 12 07 ** * 0. 54 72 a c in x 3. 71 42 1. 21 06 3. 01 44 ** * 4. 20 63 0. 82 25 0. 29 i 2 -1 a 43 3 -0 .5 25 3 -1 .4 41 5 -2 .4 93 8 0. 57 84 ** 0. 39 52 2 a c r n x i . 06 50 0. 09 69 0. 05 36 0. 48 73 0. 82 39 ** * 0. 81 70 0. 60 50 0. 00 93 -0 .3 09 9 0. 04 86 1. 30 20 ** * 0. 90 74 a r g fx 0. 94 58 0. 07 25 -0 .2 57 3 0. 29 78 0. 89 97 ** * 0. 90 31 0. 96 83 0. 05 10 -0 .3 75 9 0. 29 78 1. 27 58 ** * 0. 78 44 [ b a b e x 3. 10 75 0. 62 34 i . 34 57 3. 25 1 8 0. 81 47 ** * 0. 76 52 1. 47 67 .0 38 66 0. 89 85 i . 32 39 0. 89 94 ** 0. 34 81 b a r ix -0 .4 32 5 -0 .5 55 i -0 .8 59 7 4. 13 12 -0 .1 87 8 0. i 63 6 i . 37 08 0. 52 15 0. 31 61 3. 10 49 0. 29 07 0. 06 18 5 c pg r x i . 26 33 0. 16 39 i . 48 94 0. 72 17 i . 27 48 ** 0. 48 1 2 2. 39 50 0. 31 00 1 .3 56 5 i. 18 41 i . 62 73 0. 14 91 b c ps fx 0. 68 50 0. 23 79 0. 26 i 2 3. 01 86 0. 10 36 0. i o 76 0. 72 67 0. 3 i3 2 0. 54 97 -i . 30 26 -0 .3 61 9 0. 16 35 id f d x 1. 67 17 0. 16 88 0. 50 03 0. 85 14 1. 18 06 ** * 0. 80 99 0. 83 42 0. 04 64 0. 44 97 0. 23 52 1. 18 08 ** * 0. 92 43 ja v l x 0. 20 50 -0 .0 28 i -0 .1 82 6 -0 .0 85 8 1. 01 03 ** 0. 49 56 0. 54 67 0. 09 20 -0 .3 41 8 0. 24 96 1. 17 10 ** 0. 47 85 ja v t x 3. 11 92 0. 66 16 1. 47 71 ** 2. 90 75 0. 89 17 ** * 0. 75 25 0. 59 00 0. 08 02 0. 00 42 0. 37 87 0. 70 48 ** * 0. 74 47 k r fb x i . 90 67 0. 40 76 i. 01 91 2. 05 24 0. 72 15 ** * 0. 74 53 i. 14 75 0. 51 06 0. 57 i 6 i . 96 96 0. 44 47 0. 22 62 l o m c x 3. 54 75 0. 44 16 i. 11 12 2. 69 95 i. 1 14 8* ** 0. 72 69 0. 52 1 7 0. 00 23 i. 10 71 0. 02 22 i . 37 97 ** * 0. 81 42 m m h y x i . 44 33 0. 40 33 0. 68 57 -9 3. 56 59 -0 .0 07 2 0. 00 03 1. 31 67 0. 68 00 0. 66 63 19 .7 88 7 0. 03 69 0. 10 15 7 m n sc x 4. 29 25 0. 61 65 3. 07 78 ** 3. 13 68 1. 23 27 ** * 0. 72 90 0. 94 75 0. 13 82 -0 .1 86 0 0. 53 65 i . 25 97 0. 22 43 m o n t x -0 .5 oo o -0 .2 44 7 -0 .9 91 i -1 .2 48 1 0. 62 45 0. 15 70 0. 44 17 0. 05 50 0. 51 i 4 0. 18 58 0. 97 47 ** * 0. 57 19 m ps c x i . 07 00 0. 23 74 0. 81 1 9 0. 96 84 0. 79 30 ** 0. 43 84 i . 46 67 0. 29 92 -0 .2 76 7 1. 18 99 0. 83 88 0. 29 88 m si q x 0. 86 58 0. 15 73 0. 76 24 -i .6 62 3 -0 .3 64 1 0. 03 77 2. 38 17 0. 48 18 2. 54 35 ** 1. 74 39 i . 20 38 ** 0. 48 62 (c on ti nu ed ) t a b l e 2 (c on ti nu ed ) b ef or e a ft er r aw sh ar pe je ns en t re yn or b et a r 2 r aw sh ar pe je ns en t re yn or b et a r 2 n e fg x 3. 97 08 0. 59 49 1. 53 47 ** 2. 84 15 1. 23 62 ** * 0. 90 62 0. 04 42 -0 .0 88 5 -0 .5 95 7 -0 .2 22 0 1. 08 92 ** * 0. 64 74 pj ig x 1 .7 90 0 -0 .4 64 5 -2 .0 00 4 -4 .2 14 5 0. 48 99 0. 06 87 0. 97 25 0. 19 38 -0 .4 86 6 0. 67 03 0. 78 03 ** * 0. 52 32 po pa x 3. 06 75 0. 43 69 1. 70 22 2. 42 52 1. 14 21 ** 0. 46 49 2. 78 08 0. 50 24 2. 03 72 1. 69 84 1. 48 75 0. 25 58 po pc x 3. 00 50 0. 42 64 1. 63 53 2. 36 08 1. 14 68 ** 0. 46 72 2. 72 00 0. 49 1 8 1. 97 50 1. 65 27 1. 49 18 0. 25 89 se q u x 2. 06 50 0. 36 98 1. 14 24 ** * 2. 20 46 0. 51 26 ** * 0. 87 18 2. 05 00 0. 64 80 1. 02 75 3. 00 77 0. 43 61 ** 0. 39 42 sk se x 3. 23 83 0. 66 87 2. 43 39 5. 42 48 0. 54 2 1 0. 21 74 1. 74 92 0. 70 33 1. 18 47 ** 1. 58 55 0. 94 27 ** * 0. 57 58 ss r sc -1 .1 30 8 -0 .3 1 83 1. 27 46 1 .6 60 5 0. 87 84 0. 29 56 1. 31 92 0. 16 17 1 .7 76 8 0. 63 69 1. 33 10 0. 14 61 st c sx 1. 27 75 0. 29 44 0. 63 92 0. 94 26 1. 05 88 ** 0. 39 69 0. 91 25 0. 21 56 0. 98 26 0. 66 8 1 0. 97 57 ** * 0. 67 82 v w n d x 2. 04 17 0. 32 60 1. 03 05 ** 1. 50 94 0. 83 35 ** * 0. 86 94 2. 33 41 0. 55 79 0. 30 80 2. 07 60 0. 83 93 ** * 0. 91 26 a ve ra ge 1. 62 68 ** * 0. 25 02 ** * 0. 74 50 ** * -2 .0 17 5 0. 72 41 ** * 0. 48 49 ** * 1. 16 84 ** * 0. 22 35 ** * 0. 28 51 1. 39 73 0. 90 00 ** * 0. 46 42 ** * v ar ia nc e 2. 57 7 1 0. 15 25 1. 46 80 33 9. 06 08 0. 21 23 0. 09 45 0. 80 00 0. 11 98 1. 06 65 15 .2 97 7 0. 24 1 1 0. 07 43 n o te s: w ilc o xo n z -s ta ti st ic fo r r aw r et u rn s: 1. 29 75 (0 .1 94 5) w ilc o xo n z -s ta ti st ic fo r je n se n ’s a lp h a: i .7 81 9 (0 .0 74 8) w ilc o xo n z -s ta ti st ic fo r s h ar p e’ s in d ex 0. 25 95 (0 .7 95 2) w ilc o xo n z -s ta ti st ic fo r t re yn o r’ s in d ex : 1. 46 19 (0 .1 43 8) p -v al u es ar e in p ar en th es es ** *s ig n ifi ca n tly d if fe re n t fr o m ze ro at t h e 1% le ve l ** s ig n iti ca n tly d if fe re n t fr o m ze ro at t h e 5% le ve l perfornmnce of mutual funds 263 years before and the three years after closing are 1.63% and 1.09%, respectively, and are significantly different from zero at the 1% level.’ on the average, mutual funds earn an additional annual return of 6.57% per year for a three-year period before closing to new investors than after closing the funds. the non parametric wilcoxon rank sum test reveals that the raw return during the three years before closing is significantly different from the raw return during the three years after closing at the 5% level. the jensen’s alpha before and after closing is positive and significantly dif ferent from zero at the 1% and 5% level, respectively, indicating that the funds beat the market portfolio before and after closing to new investors. based on the raw returns, jensen’s alpha and treynor’s index, 10, eight, and nine of the 12 funds, respectively, per formed better prior to closing. the wilcoxon test indicates that the difference in the perfor mance of the funds before and after closing is statistically significant at the 5% level. the comparison of sharpe’s index is inconclusive, as seven of the 12 funds perform better prior to closing, but the test statistic is not significant. based on the comparison of performance measures before and after the closing of a fund to new investors, there is no evidence to indicate that closing of the fund resulted in improved performance. to the contrary, fund performance seems to have deteriorated after closing to new investors. these findings are consistent with the observations of damato and jereski (1997) who compare the one-year return of 13 equity funds with more than $500 million in assets that are closed to new investors since 1989 with the s&p 500 index, and conclude that closing a fund does not guarantee performance. to further examine the impact of the practice of closing funds to new investors as the assets of the fund grow, the performance of each fund in the sample was compared to a matched control portfolio of funds. the control portfolios consisted of the five funds within the same investment objective with asset size closest to the sample fund and having sufficient monthly returns data to allow a oneor three-year comparison. in table 3, the performance measures of the 27 funds in the sample during the 12 month period prior to closing of the funds are compared to the performance measures of the matching control portfolios during the same time frame. using the raw return, sharpe’s index, jensen’s alpha, and treynor’s index, the results indicate that 19 of the 27 funds out performed their matched portfolios during the year preceding the closing of the funds. however, the wilcoxon test suggests that the performance of the sample funds in the 12 months preceding their closing is not statistically different at the usual significance levels from the performance of control portfolios. when the analysis is extended to the 36 months before the closing of the funds, the results are stronger. of the 12 funds in this sample, 11 outperformed their control portfolios based on the raw return, sharpe’s index, and treynor’s index and 10 outperformed their control portfolios based on the jensen’s alpha. the wilcoxon rank sum test is significant at the 1% level for the treynor’s index and at the 5% level for the other performance mea sures, providing strong evidence that the funds in the sample were superior in performance to funds with similar investment objectives and size. in contrast, when the sample funds are compared to their control portfolios after clos ing to new investors, their performance does not compare as favorably to the control port folios. table 4 indicates that during the 12 months following their closing to new investments, only 11 of the 27 sample funds outperformed their matched control portfolios, measured by the sharpe’s index. the jensen’s alpha measure indicates 12 of the 27 sample funds outperformed their matched portfolios. in addition, the raw return and the treynor’s t a b l e 3 t he ly ea r pe rf or m an ce c om pa ri so n w ith m at ch ed po rt fo lio s b ef or e c lo si ng r aw a c f t x sa m pl e fu nd s m at ch ed p or rf ol io s sh l lr pe je ns en t re yn or b et a r 2 r aw sh ar pe je n se n t re yn or b et a r 2 0. 20 75 -0 .2 49 4 -0 .0 48 6 1. 52 44 -0 .0 38 0 0. 11 36 0. 44 67 0. 56 7 1 0. 19 92 ** -2 .4 81 1 -0 .0 73 0 0. 22 38 a c in x 3. 71 42 a c r n x 1. 06 50 a r g f x 0. 94 58 b a b e x 3. 10 75 b a r ix -0 .4 32 5 c p g r x 1. 26 33 c p s f x 0. 68 50 id f d x 1. 67 17 ja v l x 0. 20 50 ja v t x 3. 11 92 k r f b x 1. 90 67 l o m c x 3. 54 75 m m h y x 1. 44 33 m n s c x 4. 29 25 m o n t x -0 .5 00 0 m p s c x 1. 07 00 m s iq x 0. 86 58 n e f g x 3. 97 08 p ji g x 1. 79 00 1. 21 06 0. 09 69 0. 07 25 0. 62 34 -0 .5 55 1 0. 16 39 0. 23 79 0. 16 88 -0 .0 28 1 0. 66 16 0. 40 76 0. 44 16 0. 40 33 0. 61 65 -0 .2 44 7 0. 23 74 0. 15 73 0. 59 49 -0 .4 64 5 3. 01 44 ** * 4. 20 63 0. 82 25 0. 29 12 3. 32 33 0. 05 36 0. 48 73 0. 82 39 ** * 0. 81 70 1. 72 25 -0 .2 57 3 0. 29 78 0. 89 97 ** * 0. 90 31 1. 34 17 1. 34 57 3. 25 18 0. 81 47 ** * 0. 76 52 3. 06 67 -0 .8 59 7 4. 13 12 -0 .1 87 8 0. 16 36 0. 02 42 1. 48 94 0. 72 17 1. 27 48 ** 0. 48 12 0. 15 17 0. 26 12 3. 01 86 0. 10 36 0. 10 76 0. 55 25 0. 50 03 0. 85 14 1. 18 06 ** * 0. 80 99 1. 33 25 -0 .1 82 6 -0 .0 85 8 1. 01 03 ** 0. 49 56 1. 22 08 1. 47 71 ** 2. 90 75 0. 89 17 ** * 0. 75 25 2. 61 25 1. 01 91 2. 05 24 0. 72 15 ** * 0. 74 53 1. 13 08 1. 11 12 2. 69 95 1. 11 48 ** * 0. 72 69 2. 46 17 0. 68 57 -9 3. 56 59 -0 .0 07 2 0. 00 03 1. 28 67 3. 07 78 ** 3. 13 68 1. 23 27 ** * 0. 72 90 2. 85 58 -0 .9 91 1 -1 .2 48 1 0. 62 45 0. 15 70 0. 40 92 0. 81 19 0. 96 84 0. 79 30 ** 0. 43 84 0. 26 00 0. 76 24 1. 66 23 -0 .3 64 1 0. 03 77 0. 21 17 1. 53 47 ** 2. 84 15 1. 23 62 ** * 0. 90 62 2. 51 75 -2 .0 04 -4 .2 14 5 0. 48 99 0. 06 87 0. 07 08 0. 78 38 2. 46 44 ** 2. 74 87 1. 11 65 0. 28 61 0. 20 02 0. 60 76 0. 99 09 1. 06 87 ** * 0. 84 30 0. 13 06 0. 00 19 0. 58 55 1. 13 37 ** * 0. 75 74 0. 54 54 1. 10 59 2. 62 08 0. 99 53 ** * 0. 89 75 -1 .5 25 1 -0 .3 02 8* ** -8 .6 97 0 0. 03 67 0. 23 12 -0 .0 70 7 0. 18 39 -0 .2 27 9 0. 84 09 ** * 0. 89 72 0. 12 18 0. 02 24 0. 56 84 0. 31 71 ** * 0. 77 24 0. 13 16 0. 21 52 0. 63 18 1. 05 40 ** * 0. 89 35 0. 49 25 0. 87 35 1. 58 59 0. 58 59 ** 0. 44 41 0. 35 87 0. 28 5 1 1. 44 91 1. 43 95 ** * 0. 89 27 0. 25 80 0. 31 03 1. 14 32 0. 61 66 ** * 0. 95 41 0. 38 38 0. 32 43 2. 04 80 0. 93 92 ** * 0. 95 77 0. 41 80 0. 42 97 7. 10 83 0. 07 32 0. 04 81 0. 50 69 1. 80 90 ** 2. 50 42 0. 97 03 ** * 0. 76 89 0. 02 67 -0 .2 70 5 0. 10 98 1. 18 11 0. 24 16 -0 .0 11 2 0. 01 59 -0 .0 40 1 1. 04 63 ** * 0. 56 84 -0 .0 18 6 0. 01 56 0. 32 78 -0 .1 49 5 0. 01 34 0. 43 65 0. 42 49 2. 01 60 1. 02 14 ** * 0. 97 03 -0 .2 53 4 -0 .1 73 8 -0 .8 89 1 0. 22 94 ** 0. 44 95 p o p a x 3. 06 75 0. 43 69 1. 70 22 2. 42 52 p o p c x 3. 00 50 0. 42 64 1. 63 53 2. 36 08 s e q u x 2. 06 50 0. 36 98 1. 14 24 ** * 2. 20 46 s k s e x 3. 23 83 0. 66 87 2. 43 39 5. 42 48 s s r s c -1 .1 30 8 -0 .3 1 83 -1 .2 74 6 -1 .6 60 5 s t c s x 1. 27 75 0. 29 44 0. 63 92 0. 94 26 v w n d x 2. 04 17 0. 32 60 1. 03 05 ** 1. 50 94 a ve ra ge 1. 62 68 ** * 0. 25 02 ** * 0. 74 50 ** * -2 .0 17 5 v ar ia n ce 2. 57 71 0. 15 25 1. 46 80 33 9. 06 08 n ot es : w ilc ox on z -s ta tis tic fo r r aw r et ur ns : 0. 92 56 (0 .3 54 7) w ilc ox on z -s ta tis tic fo r s ha rp e’ s in de x 0. 72 66 (0 .4 67 5) pva lu es ar e in p ar en th es es ** *s ig ni fi ca nt ly di ff er en t fr om z er o at t he i % l ev el ** si gn if ic an tly di ff er en t fr om z er o at t he 5 % l ev el 1. 14 21 ** 0. 46 49 1. 04 58 0. 18 19 -0 .0 37 6 1. 14 68 ** 0. 46 72 1. 51 75 0. 22 42 0. 11 44 0. 51 26 ** * 0. 87 18 1. 65 08 0. 13 21 0. 73 87 ** 0. 54 21 0. 21 74 1. 63 92 0. 24 18 0. 36 63 0. 87 84 0. 29 56 0. 29 50 -0 .0 08 7 0. 17 89 1. 05 88 ** 0. 39 69 0. 92 58 0. 16 55 0. 25 2 1 0. 83 35 ** * 0. 86 94 1. 44 33 0. 18 76 0. 43 80 ** * 0. 72 41 ** * 0. 48 49 ** * 1. 31 54 ** * 0. 17 06 ** 0. 38 90 ** * 0. 21 23 0. 09 45 0. 95 09 0. 16 78 0. 36 04 w ilc ox on z -s ta tis tic fo r je ns en ’s a lp ha : 1. 18 65 ( 0. 06 93 ) w ilc ox on z -s ta tis tic fo r t re yn or ’s in de x: 1. 41 86 (0 .1 56 0) 0. 89 00 0. 84 06 ** * 1. 03 15 1. 18 25 ** * 0. 75 67 0. 94 60 ** * 1. 28 59 1. 04 32 ** * -0 .0 32 3 1. 01 04 ** * 0. 55 56 1. 16 34 ** 0. 81 26 0. 81 19 ** * 0. 71 86 0. 79 41 ** * 6. 20 45 k 18 91 0. 59 74 % 0. 67 61 9 0. 96 69 0. 50 56 0. 57 71 i 4 0. 36 25 0. 98 89 : 0. 62 17 ** * i 0. 09 34 f? , 9 & r aw a c f t x t a b l e 4 t he 1 -y ea r pe rf or m an ce c om pa ri so n w ith m at ch ed po rt fo lio s a ft er c lo si ng sa m pl e f un ds m at ch ed p or tf ol io s sh ar pe je ns en t re yn or b et a r 2 r aw s h ar pe je ns en t re yn or b et a r 2 -0 .0 28 3 -0 .8 04 1 -0 .2 87 2* ** -2 .5 10 6 0. 12 07 ** * 0. 54 72 0. 02 75 -0 .5 4 18 -0 .2 38 4 -3 .6 55 6 0. 06 76 0. 11 09 a c in x 1 .0 43 3 -0 .5 25 3 -1 .4 41 5 -2 .4 93 8 a c r n x 0. 60 50 0. 00 93 -0 .3 09 9 0. 04 86 a r g f x 0. 96 83 0. 05 10 -0 .3 75 9 0. 29 78 b a b e x 1. 47 67 0. 38 66 0. 89 85 1. 32 39 b a r ix 1. 37 08 0. 52 15 0. 31 61 3. 10 49 c p g r x 2. 39 50 0. 31 00 1 .3 56 5 1. 18 41 c p s f x 0. 72 67 0. 31 32 0. 54 97 -1 .3 02 6 id f d x 0. 83 41 0. 04 64 0. 44 97 0. 23 52 ja v l x 0. 54 67 0. 09 20 -0 .3 41 8 0. 24 96 ja v t x 0. 59 00 0. 08 02 0. 00 42 0. 37 87 k r f b x 1. 14 75 0. 5 10 6 0. 57 16 1. 96 96 l o m c x 0. 52 17 0. 00 23 1. 10 71 0. 02 22 m m h y x 1. 13 67 0. 68 00 0. 66 63 19 .7 88 7 m n s c x 0. 94 75 0. 13 82 -0 .1 86 0 0. 53 65 m o n t x 0. 44 17 0. 05 50 0. 51 14 0. 18 58 m p s c x 1. 46 67 0. 29 92 -0 .2 76 7 1. 18 99 m s iq x 2. 38 17 0. 48 18 2. 54 35 ** 1. 74 39 0. 57 84 ** 0. 39 52 1. 30 20 ** * 0. 90 74 1. 27 58 ** * 0. 78 44 0. 89 94 ** 0. 34 8 1 0. 29 07 0. 06 18 1. 62 73 0. 14 91 -0 .3 61 9 0. 16 35 1. 18 08 ** * 0. 92 43 1. 17 10 ** 0. 47 85 0. 70 48 ** * 0. 74 47 0. 44 47 0. 22 62 1. 37 97 ** * 0. 81 42 0. 03 69 0. 01 57 1. 25 97 0. 22 43 0. 97 47 ** * 0. 57 19 0. 83 88 0. 29 88 1. 20 38 ** 0. 48 62 -0 .8 52 5 -0 .3 80 1 1. 25 03 1 .6 00 9 0. 78 17 ** * 0. 50 40 1. 67 33 0. 12 57 0. 63 88 0. 65 81 1. 71 94 ** * 0. 91 65 1. 54 92 0. 11 96 0. 09 78 0. 65 96 1. 45 66 ** * 0. 88 04 0. 59 83 0. 08 68 -0 .0 20 1 0. 30 52 1. 02 34 0. 32 86 0. 54 33 0. 32 58 0. 04 25 4. 63 89 0. 01 62 0. 01 17 2. 25 33 0. 82 86 -0 .1 05 9 1. 90 47 0. 93 72 ** 0. 41 12 0. 91 58 0. 58 19 0. 63 10 4. 82 65 0. 13 69 0. 04 11 0. 64 67 0. 01 39 0. 27 88 0. 06 97 1. 29 51 ** * 0. 94 91 1. 11 08 0. 41 70 0. 37 62 0. 96 56 0. 88 70 ** * 0. 65 74 0. 63 83 0. 06 82 -0 .0 13 2 0. 35 78 0. 88 11 ** * 0. 60 34 1. 18 58 0. 68 25 0. 47 89 ** i . 43 70 0. 63 61 ** * 0. 76 01 -0 .1 95 8 -0 .0 79 2 0. 09 17 -0 .6 88 3 0. 99 79 ** * 0. 97 86 1. 26 25 0. 38 65 0. 24 94 2. 74 28 0. 24 65 0. 25 64 0. 77 33 0. 13 92 -0 .1 12 6 0. 55 87 0. 89 79 0. 20 92 0. 98 42 0. 19 05 1. 12 20 0. 61 55 1. 17 56 ** * 0. 62 49 2. 64 67 0. 62 95 0. 91 42 2. 61 90 0. 83 17 0. 27 23 1. 78 50 0. 30 92 2. 09 80 ** 0. 93 14 1. 61 34 ** * 0. 70 23 n e fg x 0. 04 42 -0 .0 88 5 -0 .5 95 7 -0 .2 22 0 pj ig x 0. 97 25 0. 19 38 -0 .4 86 6 0. 67 03 po pa x 2. 78 08 0. 50 24 2. 03 72 1. 69 84 po pc x 2. 72 00 0. 49 18 1. 97 50 1. 65 27 se q u x 2. 05 00 0. 64 80 1. 02 75 3. 00 77 sk se x 1. 74 92 0. 70 33 1. 18 47 ** 1. 58 55 ss r sc 1. 31 92 0. 16 17 1 .7 76 8 0. 63 69 st c sx 0. 91 25 0. 21 56 0. 98 26 0. 66 8 1 v w n d x 2. 33 41 0. 55 79 0. 30 80 2. 07 60 a ve ra ge 1. 16 84 ** * 0. 22 35 ** * 0. 28 51 1. 39 73 * v ar ia nc e 0. 80 00 0. 11 98 1. 06 65 15 .2 97 7 n ot es : w ilc ox on z -s ta tis tic fo r r aw r et ur ns : -0 . i s. 57 (0 .8 76 3) w ilc ox on z -s ta tis tic fo r s ha rp e’ s in de x -0 .5 53 6 (0 .5 79 9) pva lu es ar e in p ar en th es es ** *s ig ni fi ca nt ly di ff er en t fr om z er o at t he 1 % l ev el ** si gn if ic an tly di ff er en t fr om z er o at t he 5 % l ev el 1. 08 92 ** * 0. 64 74 1. 10 17 0. 32 78 0. 47 55 0. 78 03 ** * 0. 52 32 0. 81 83 0. 37 97 0. 02 89 1. 48 75 0. 25 58 1. 38 25 0. 32 68 0. 74 24 1. 49 18 0. 25 89 1. 13 25 0. 26 07 0. 42 1 1 0. 43 61 ** 0. 39 42 1. 71 00 0. 28 63 0. 27 57 0. 94 27 ** * 0. 57 58 1. 19 83 0. 26 74 0. 54 01 1. 33 10 0. 14 61 3. 28 33 0. 82 30 0. 84 48 0. 97 57 ** * 0. 67 82 0. 97 50 0. 23 04 1. 07 85 ** 0. 83 93 ** * 0. 91 26 2. 52 25 0. 60 84 0. 42 04 ** 0. 90 00 ** * 0. 46 42 ** * 1. 17 30 ** * 0. 27 46 ** * 0. 37 43 ** * 0. 24 11 0. 07 43 0. 75 91 0. 10 00 0. 34 89 w ilc ox on z -s ta tis tic fo r je ns en ’s a lp ha : -0 .5 01 7 (0 .6 15 9) w ilc ox on z -s ta tis tic fo r t re yn or ’s in de x: -0 .6 40 1 (0 .5 22 1) 0. 77 91 1. 04 71 ** * i . 40 38 0. 26 28 ** 0. 96 18 1. 17 27 ** 0. 63 19 1. 38 95 ** 0. 90 99 1. 06 80 ** * 0. 76 87 1. 22 77 ** 2. 81 88 0. 99 76 0. 66 50 1. 07 43 ** * 2. 18 47 0. 88 37 ** * 1. 05 44 ** * 0. 91 57 ** * 2. 82 07 0. 20 12 0. 71 89 p 0. 47 14 s 0. 33 76 0. 49 77 0. 83 15 [ 4 0. 35 39 0. 19 32 0. 78 17 2 0. 97 98 h 0. 53 27 ** * g 0. 08 94 f 8 268 financial services review 6(4) 1997 index are higher for 13 of the 27 sample funds over the matched portfolios. however, the wilcoxon test indicates no statistical difference in performance between the sample funds and the control portfolios. when the analysis is extended to three years around the closing date, the results are similar. based on the treynor’s index only six of 12 funds outperformed their control port folios after closing to new investors. using the raw return, sharpe’s index, and jensen’s alpha, seven of 12 funds outperformed their peers in the 36 months following the decision to close to new investors. the test statistic indicates no significant differences in perfor mance between the sample funds and the control portfolios. thus, while the sample funds significantly outperformed their peers prior to closing, the performance does not seem to persist after closing to new investors. consequently, the empirical evidence does not sup port the wisdom of closing funds to new investors for reasons related to asset size. iv. implications of the study mutual funds are the most popular investment outlets for individual investors. as such, it is important to develop a thorough understanding of factors that influence the performance of mutual funds. while many facets of mutual fund performance have been examined in the finance literature, the common practice of closing funds to new investors as funds grow has not received any attention in the literature. fund managers often justify the decision to close funds to new investors based on the desire to maintain superior performance. the premise underlying this line of reasoning is that due to limited suitable investment opportunities, substantial amounts of assets or con tinued substantial cash inflows make it difficult to maintain strong performance. however, the empirical evidence presented here contradicts the premise that superior performance can be maintained by restricting access to the funds, or by capping the asset size. the evi dence indicates that based on the raw return, sharpe’s index, jensen’s alpha, and treynor’s index, funds performance declines after closing to new investors. the conclusion is the same when the sample funds are compared to control portfolios matched in size and invest ment objective. thus, maintaining performance does not seem to be an appropriate reason for closing funds to new investors. the practice of closing funds to new investors continues to be popular. while the ini tial sample of this study was the 128 funds closed to investors as of january 1996, morn ingstar data indicates there were 141 funds closed to new investors as of january 1998. however, the actual number of funds closed during the interim is significantly higher, as many funds close to new investors for short periods, and subsequently re-open. thus, there are a number of funds that were closed in january 1996, that are now open to the public, as well as a large number of funds that have closed to new investors since that time. of the 27 funds that were used in the one-year comparisons, 17 continue to be closed to new invest ments as of january 1998, and three are no longer available in mornzngstar. similarly, of the 12 funds with data available to facilitate the three-year analysis, eight remain closed to new investments and one is no longer available in morningstar. given the persis tence of this practice, it is possible that fund managers have other reasons for closing funds that are unrelated to the performance of the fund. performunce of mutual funds 269 notes 1. tables detailing the results of the three-year comparisons are available from the authors upon request. references barber, j.r. (1994). mutual fund risk measurement and future returns. quarterly journal ofbusiness and economics, 33(winter), 55-64. brown, s.j., & goetzmann, w.n. (1995). performance persistence. journal of finance, 50,679-698. carhart, m.m. (1997). on persistence in mutual fund performance. journal of finance, 53, 57-82. damato, k., & jereski, l. (1997). fidelity’s 34-day offer to jump in to magellan fund probably should be ignored, many advisers assert. the wall street journal, august 28, c 1. elton, e.j., gruber, m.j., & blake, c.r. (1996). the persistence of risk-adjusted mutual fund perfor mance. journal of business, 69, 133157. fabozzi, f.j., francis, j. c., & lee, c.f. (1980). generalized functional form for mutual fund returns. journal of financial and quantitative analysis, 15, 1107-l 120. fabozzi, f.j., & francis, j.c. (1979). mutual fund systematic risk for bull and bear markets: an empirical examination. journal of finance, 34, 12341250. gibbons, j.d. ( 1985). nonparumetric methods for quantitative unulysis. columbus, oh: american sciences press, inc. goetzmann, w.n., & ibbotson, r.g. (1994). do winners repeat? patterns in mutual fund return behavior. journal of portfolio management, l&winter), 918. grinblatt, m., & titman, s. (1992). the persistence of mutual fund performance. journal offinance, 47, 1977-1984. gruber, m.j. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51, 783-810. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974-1988. journal of finance, 48, 93-130. jensen, m.c. (1968). the performance of mutual funds in the period 1945-1964. journal offinance, 23, 389-416. kahn, r.n., & rudd, a. (1995) does historical performance predict future performance?. financial analysts journal, sl(november-december), 43-52. lehmann, b.n., & modest, d.m. (1987). mutual fund performance evaluation: a comparison of benchmarks and benchmark comparisons. journal of finance, 42, 233-265. malkiel, b.g. (1995). returns from investing in equity mutual funds 1971 to 1991. journal of finance, 50, 549-572. phelps, s., & detzel, l. (1997). the nonpersistence of mutual fund performance. quarterly journal of business and economics, 36(spring), 55-67. sharpe, w.f. (1966). mutual fund performance. journal of business, 39(special supplement), 119 138. treynor, j.l. (1965). how to rate management of investment funds. harvard business review, 43(january_february), 63-75. pii: 1057-0810(94)90019-1 financial services review, 3(2): 143-1.56 copyright 0 1994 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. coupon resets versus poison puts: the valuation of event risk provisions in corporate debt joseph a. fields david s. kidwell linda s. klein this paper examines the valuation of the two major types of event risk indenture provisions, poison puts and coupon resets, on the debt of industrial companies. in contrast with earlier work by crabbe (1991), wefind thatprotectionprovided by poison put type of covenants is not valued by investors. the inclusion of coupon reset provisions, however, lowers the yield spread of new issued industrial bonds by 32 basis points. the yields on bonds with low credit quality ratings are reduced by including coupon reset provisions in the bond indentures. i. introduction during the 1980’s event risk became a topic of concern among bondholders, analysts, and investment bankers. event risk refers to management actions or other events that increase the firms’ leverage or otherwise increase the risk of a company. the record number of takeovers and corporate restructurings produced numerous instances of substantial capital losses for bondholders. event risk has continued to be headline news in the 1990’s, as evidenced by marriott’s plan to restructure into two companies, one burdened with almost all of the existing debt. as a result, moody’s downgraded marriott’s bonds to junk, and the price of the bond plunged from $1100 to $800 (mitchell, 1992). in order to protect themselves from potential capital losses due to event risk, bond holders sought protective covenants that would limit the financial effect of restructurings and other adverse managerial actions. metropolitan life insurance company, for example, declared that it would no longer purchase debt instruments without covenants that would protect it against event risk (winkler & herman, 1988). a recent article by crabbe (1991) examined the pricing of new bonds that contained poison put provisions, which provide joseph a. fields l department of finance, university of connecticut, storrs, ct 06269; david s. kidwell l school of business, university of minnesota, minneapolis, mn 55455; linda s. klein l department of finance, university of connecticut, storrs, ct 06269. 144 financial services review, 3(2) 1994 event risk protection. he concluded that put provisions significantly reduced the yield on new debt after controlling for differences in bond and issuer characteristics. however, event risk covenants are not standard. the strength of protection and the value of poison put provisions depend upon the characteristics of the provision and the economic environment for restructuring. this study refines previous work by examining the differences in market valuation for the two major types of event risk protective covenants poison puts and the more recently developed coupon resets. poison puts are indenture provisions that require the issuing firm to repurchase debt at face value if a named event occurs. coupon resets do not mandate that the bond issue be repurchased, instead they require that the bond interest rate be reset if there is a lowering in the moody’s or standard and poor’s credit rating. in addition, the study examines the impact of the rjr/nabisco takeover on the pricing of event risk covenants. the rjr/nabisco takeover is the largest ever and illustrates that even the largest companies are subject to takeover attempts. to many, it is considered a watershed event with respect to event risk. this study extends crabbe’s sample period, november 1988 to december 1989, to include data from 1986 through 1990. this doubles the available sample of bonds with event risk protection, allowing us to examine the pricing of put provisions in bonds issued both preand post-rjr/nabisco. the sample also allows us to examine coupon reset provisions which are not available to crabbe. additionally, this sample provides for an examination of more diverse covenants including both very strong and very weak event risk covenants that are not found in crabbe’s original sample. unlike crabbe, the results of this paper indicate that poison put provisions are not valued by investors. in fact, our results indicate that, before the rjr/nabisco takeover, bonds with poison put provisions sold at penalty yields. bonds with coupon reset provisions, which appeared after the rjr/nabisco takeover, however, sold at yields below similar bonds without event risk protection. additional hypotheses tested in the paper are: (1) whether the value of event risk protection changes as the strength of the protection provided by the covenants change and (2) whether there is a relationship between credit quality and event risk protection. the issue of the event covenant’s value has particular implications for the individual investor as monitoring for the individual is a significant cost factor. often the individual uses an intermediation device to avoid such costs. the existence of complex bond covenants limits the ability of individuals to monitor and evaluate management’s actions with regard to bond holders’ wealth (smith & warner, 1979). this classic agency/monitoring cost problem (jensen & meckling, 1976) is mitigated by effective put provisions. under such arrangements, management must maximize the wealth of shareholders subject to a constraint that limits the ability of shareholders to profit at the expense of bondholders. the paper is organized as follows. section ii presents background of event risks covenants in corporate debt. section iii discusses the factors that are expected to affect the valuation of event risk covenants. section iv develops the hypotheses to be tested. section v describes the sample and data. section vi contains the methods used to test the hypotheses. section vii presents the findings and the final section reports the conclusions. coupon resets versus poison puts 145 ii. background information on protective covenants bondholder protection the stockholder bondholder conflict has been studied by numerous authors, most notably fama and miller (1972). this area of research points out that the manager’s primary responsibility is to maximize shareholder’s wealth. in a study of contractual relationships with bondholders, lehn and poulsen (199 1) find that courts generally rule that the company’s board of directors does not have the responsibility of protecting bondholders. bondholder protection against adverse actions of the corporation is limited to provisions of the bond indenture. smith and warner (1979) examine the indenture provisions of corporate bonds and find a wide variety of clauses regarding the rights and protection of bondholders. depending on the design of the individual indenture provisions, the bondholder may have substantial or inconsequential protection against wealth transfers. unless bond indentures have protective covenants, managers may increase shareholder wealth at the expense of bondholders by increasing the risk level of existing bonds in one of four ways: 1) increasing dividends to shareholders, beyond the level expected by bondholders; 2) selling additional bonds at the same or higher priority than existing bonds; 3) substituting risky assets for low risk assets; and 4) underinvesting in new projects where the payoff would be available to bond holders. when determining the value of protection provided by the different types of event risk provisions, one must consider the terms of the covenant, the probability that the event risk covenant will be triggered, and the financial and operating characteristics of the firm. another consideration is the potential for extreme financial burden from reissuing bonds for firms that have undergone a credit deterioration. bond covenants offer wide variation in the type and quality of event risk protection. apart from the triggering requirement, the primary differences between the two major types of protection, coupon resets and put provisions, are the type of triggering events that activate the provision. some coupon resets require only a change in rating or a change in rating below investment quality to become active. poison put provisions, however, require both the occurrence of a named event and a decline in bond rating to trigger the covenant. coupon resets generally provide high levels of protection, while poison puts tend to offer low levels of protection. event risk ratings in order to provide bondholders with more information about event risk protection, standard and poor’s (s&p) has developed a system to rank the quality of the protection provided by the covenants. this system provides five possible rankings with el indicating the highest quality protection against event risk and e5 indicating the weakest protection. all bonds with an el event risk ranking are coupon reset type of event risk protection. these bonds include indenture provisions that require the firm to reset the coupon rate following a change in the firms bond credit rating. the cause of the downgrade is immaterial, so that all types of events are covered by these indenture provisions. the comprehensive 146 financial services review, 3(2) 1994 table 1. bond coupon reset provisions for enron corporation 9.5 percent credit sensitive notes due 2001 s&p credit rating investment grade aaa aa+ to aa a+ to a bbb+ to bbb noninvestment grade bb+ bb bb moody’s credit rating aai3 aal to aa al to a3 baal to baa3 bal ba2 ba3 applicable rate for coupon reset 9.20% 9.30% 9.40% 9.50% 12.00% 12.50% 13.00% interest rate difserential above aaa rating (in basis points) 10 20 30 280 330 380 source: prospectus supplement for enron corporation, may 3 1, 1998. scope of this type of coupon reset provides protection against ordinary credit erosion resulting from fundamental business deterioration as well as from leveraged buyouts or other events. however, the el ranking does not provide protection against all risk. bonds are still subject to both liquidity and default risk. the el protective covenant includes a schedule of coupon interest rates for each level of bond rating, and the triggering event is any change in the bond rating. the interest rate reset is largest as the bond rating crosses from investment to speculative grade. for example, enron corp., 9.5 percent credit sensitive notes due in 2001 were issued in june 24, 1989 with a s&p credit rating of bbband a moody’s rating of baa3. the indenture provision provides that if the bond rating changes, the coupon rate for the bond will also change according to a set schedule. table 1 shows that interest rate penalty yields are quite small when a credit declines within the investment grades. however, a substantial penalty applies when credit quality drops below investment grade. most bonds that qualify for an e2 ranking also have a coupon reset structure.’ bonds with e2 rankings require a designated event in addition to a rating downgrade to speculative grade before triggering the protection. if the bond already has a speculative grade credit rating, a downgrade of one full rating category will trigger the coupon reset. bondholders must absorb any loss from credit deterioration within the range of investment quality ratings. the scope of the designated events for e2 bonds is broad and may include: acquisition of a specified percentage of voting control, change in majority of the board of directors, merger, consolidation, asset transfer, specified types of acquisitions, and large special dividends or stock repurchases. the specified thresholds for these events are sufficiently low so that they provide protection against major changes in credit quality, including changes brought about by a series of transactions over a multi-year period. the e3 bonds generally have poison put type structures and are protected against many of the same events as e2 bonds.2 the thresholds for triggering the e3 covenants, however, are high enough that the bond can fall below investment grade as a result of an event or series of events that will not trigger the protection. the e3 ranking bonds generally include put at par provisions triggered by designated events in conjunction with a rating downgrade from investment grade to speculative grade by moody’s and/or s&p. some e3 bonds are also coupon resets versus poison puts 147 triggered if the rating is withdrawn or if a speculative bond is downgraded one full rating category. bondholders must absorb any loss from credit deterioration within investment grade. most e3 bonds require that rating changes occur a minimum of 90 days before or after the designated event to trigger the protective covenant. the lack of a designated event covering acquisitions by the issuers may keep a bond from qualifying at a higher ranking. most e4 ranked bonds are similar in structure to those found in the e3 category. these bonds allow the holder to put the issue at par with the occurrence of designated events and a rating downgrade below investment grade. the event thresholds allowed under an e4 bond, however, are generally higher than those of e3. thus, the company can undertake actions which cause a deterioration in credit quality to speculative grade without triggering the put. often the protection is not triggered unless the bond has a speculative rating by both moody’s and s&p. in some cases, such as ual and coastal corporation, the put is triggered by a change in control along with a rating downgrade. bond’s with e5 rankings provide very little or no protection to bondholders. these bonds allow a put at par in the event of a change in control that is not approved by a majority of continuing directors. generally, board approval is ultimately granted even in the case of initially hostile takeovers; therefore, the probability of triggering this put is low. many credit damaging events, such as major recapitalizations, special dividends and debt financed acquisitions are not covered. armstrong world industries, for example, issued an e5 ranked bond for which the put protection requires the directors to first pass a resolution stating that the relevant event risk indenture article applies to the debentures. thus, there is no assurance that the protective provisions will be activated. hi. the poison put option and factors which influence its value theoretical development the valuation of options that provide protection to bondholders when there is a violation of an indenture covenant has been studied by black and cox (1976) and mason and bhattacharya (1981). bicksler and chen (1992) extend previous work on protective covenants by developing a theoretical model to specifically examine the valuation of bonds with event risk provisions. they argue that the value of a bond with an event risk provision has well defined upper and lower boundaries. if no event occurs and the firm remains solvent, the value of the bond is simply the set of cash flows promised at the inception of the security. if a specified event occurs, the value of the security equals the repurchase price specified in poison put or the capitalized value of the bond with a reset interest rate. if the company has insufficient assets to fully redeem the bond issue, the value of the security equals the value of the firm. given that there is some probability of triggering the provision and a positive firm value, the addition of an effective event risk provision may increase the issue price of the bond. the rjr/nabisco effect because expectations play a major role in the valuation process, a large unexpected event is likely to change substantially investor perceptions of the probability that unprotected bondholders will suffer capital losses. such an event occurred in october of 1988 when 148 financial services review, 3(2) 1994 rjr/nabisco management announced a leveraged buyout offer and a few weeks later kohlberg, kravis, and roberts (kkr) launched a successful $25 billion takeover of rjr/nabiscwthe largest takeover ever. during the takeover, the value of the firm’s debt declined by $800 million or almost 40 percent.3 in many ways the rjr/nabisco takeover is a watershed event because of the magnitude of bondholder losses even for the largest of companies. the size of the transaction and the fact that management completed a major financing program only months before they announced their intention to take the company private, appeared to have increased bondholder concern about event risk. for the five months following the takeover, the average weekly volume of new investment grade corporate bonds fell from $555 to $255 million.4 also, following the takeover, stronger poison put as well as coupon reset provisions were developed and companies increased the use of event risk protection in new bond issues.’ iv. hypotheses the value of a poison put provision depends in part on expectations about the environment for corporate control when the bond is issued. we posit that prior to the rjr/nabisco takeover, investors’ fear of event risk was lower. based on the s&p event risk rankings, the pre-rjr/nabisco event risk covenants provided only limited protection. following the rjr/nabisco takeover, the frequency of event risk provisions increased and the quality of the protection provided by the provisions improved. the first hypothesis tests whether investors’ perception of event risk changed follow ing the rjr/nabisco takeover. if investors perceive that there is a high probability of triggering the event risk covenant and that the covenant provides adequate protection, then the inclusion of the event risk covenant should reduce the cost of borrowing. if investors believe that the probability of the triggering event is low or that the event risk provision provides ineffective protection, the event risk covenant should not affect the bond’s yield. finally, if the inclusion of a protective covenant provides a signal that the potential for event risk is high and the protection to bondholders is low, investors will demand a higher yield for bonds containing these indentures. we posit that the economic environment for takeovers should affect the market for takeover protection. therefore, following the rjr/nabisco takeover, companies providing effective protection against event risk may issue new debt at lower yields. the second hypothesis tests whether the value of an event risk provision is related to the type of event risk covenant. specifically, coupon resets provide stronger protection, and issues with stronger event risk protection should provide greater value for the bondholder. therefore, bonds with coupon reset covenants should have lower yields than bonds with poison put covenants. the third hypothesis tests whether the value of a poison put provision is related to the quality of the event risk protection provided in the bond indenture. poison put provisions contain a wide range of protection and, again, stronger protection should reflect greater value for the bondholders. therefore, bonds with higher poison put rankings should have lower yields than bonds with lower poison put rankings. the fourth hypothesis tests whether the value of the event risk covenant is greater for bonds with lower credit ratings. the impact of achange in the bond quality rating is relatively small as long as the bond maintains an investment grade rating. coupon reset bonds have a coupon resets versus poison puts 149 larger reset provision when the bond changes from investment to speculative grade. puttable bonds have triggering events that require the reduction of bond ratings to less than investment grade and the occurrence of a designated event. because lower rated bonds are more likely to fall below investment grade, an event risk provision may be more valuable than a provision on a similar bond with a high credit rating. v. description of the sample and the data to examine the value of event risk protection, a sample of 65 industrial bonds and notes with event risk covenants is identified from s&p creditweek. in addition, a control sample of 163 corporate bonds, not identified with event risk covenants by s&p creditweek, is identified from moody’s bond record. the total sample consists of 228 bonds issued between january of 1987 and june of 1990 for which full data was available. utility, financial institutions, and government bonds are excluded from the sample, as regulation can affect the probability of restructuring. also, to better match the characteristics of the two samples, the sample excludes bonds with maturities less than five years, issue sizes less than $45 million or greater than $350 million, and s&p default ratings ranging above aaand below bb+. also excluded from the sample are mortgage bonds, zero coupon bonds, and bonds for which complete data are not available. firm specific information is obtained from the s&p bond guide, moody’s bond record, and disclosure. yields on treasury securities are collected from the federal reserve statistical release h.15 (519): selected interest rates. table 2 reports the descriptive statistics for the total sample as well as for coupon reset, poison put, and control subsamples. the yield spread over treasury is calculated by subtract ing the yield on a treasury security from that of an industrial bond at issue, each with the same maturity and sale date. we find that the yield off treasury for coupon reset bonds and poison put bonds is slightly higher than for the control sample. the size of coupon reset bond issues is larger than that of the poison puts or control sample. compared with the control sample, poison put bonds are more likely to have a sinking fund provision and less likely to have a call provision. the coupon reset bonds, however, tend not to have either sinking funds or call provisions. the coupon reset bonds tend to have the longest term to maturity, although both types of event risk bonds have longer terms to maturity than the control sample. companies issuing bonds with coupon resets tend to be larger than those offering poison puts, but companies issuing bonds with event risk provisions tend to be smaller than fums in the control sample. it is useful to highlight several differences between subsamples in order to understand the pattern of event risk protection used by firms. transportation companies have a greater representation in each of the event risk protection subsamples, while companies involved in the trade classification tend not to issue event risk protection. the differences between the event risk sample and the control sample indicate that companies in sectors most vulnerable to event risk are more likely to include poison put protection than companies in other sectors.6 the coupon reset sample is fairly evenly distributed across all represented credit quality categories. the poison put sample, however, is heavily concentrated in the a and bbb credit quality categories and does not contain any low rated (bb) bonds. the control sample contains bonds of each credit quality, but the majority of the bonds are rated a. table 3 illustrates the effect of the rjr/nabisco takeover on the development of protective covenants. prior to the takeover all protective covenants were rated either u or 150 financial services review, 3(2) 1994 table 2. characteristics of new debt issues by event risk protection from january 1987 to june 1990 coupon reset poison put characteristics total sample sample sample control sample size 228 10 55 163 yield spread over treasury (mean %)” 1.20 1.33 1.40 1.13 size of issue ($ millions) 162.7 196.5 158.0 163.9 presence of sinking fund (8) 7.9 0.0 12.7 6.7 presence of call provision (%) 38.2 0.0 29.1 43.6 time to maturity (years) 13.2 19.7 15.0 12.3 pii’s total assets ($ billions) 11.2 6.8 4.1 13.9 yield on a 10 t-bond (%) year 8.5 8.4 8.4 8.5 industry of issuer (%)b manufacturing 53.9 40.0 63.6 51.5 transportation 6.6 20.0 18.2 1.8 trade 13.6 10.0 3.6 17.2 other 25.9 30.0 14.5 29.4 standard & poor’s credit rating (%)’ aaa 0.0 0.0 0.0 0.0 aa 10.5 20.0 7.3 11.0 a 53.1 40.0 40.0 58.3 bbb 33.3 20.0 52.7 27.6 bb 3.1 20.0 0.0 3.1 s&p event protection ranking (%) el 2.6 60.0 0.0 0.0 e2 1.8 30.0 1.8 0.0 e3 15.8 10.0 60.0 0.0 i?4 3.5 0.0 14.5 0.0 e5 4.8 0.0 20.0 0.0 notes: a we subtract the yield on the treasury security with the same maturity and sold on the same day to adjust for term-structure effects. b we exclude utilities, financial institutions, and government issues to reduce the impact of regulation on the sample. ’ we limit the control sample to bonds with s&p ratings between aaand bb+ so that the control sample rellects the poison put sample. e5 and, thus, provided investors with low event risk protection. after the takeover, 82.5 percent of the puts were rated e3 or better. ‘ii-m, it appears that after the rjr/nabisco takeover, the design and quality of event risk covenants changed, providing bonds with better event risk protection than previously available. table 3. type of event risk protection relative to the rjr/nabisco takeover type of event risk pre-rjrhvabisco takeover post-rjr/nabisco takeover total protection number percent number percent number percent high (el)a 0 0 6 9.5 6 7.3 med (e2 and e3) 0 0 46 73.0 46 56.1 low (e4 and es) 19 100 11 17.5 30 36.6 total 19 100 63 100 82 100 note: ’ all el ranked bonds are coupon resets coupon resets versus poison puts 151 vi. empirical model and tests in this section we examine empirically the effect of event risk protection on yield spreads of newly issued bonds. the empirical model to test our hypotheses we first develop a model to explain the yield spread over treasury. previous studies suggest that new issue yields depend on the industry of the issuer (nd), the bond rating of the issue (rate), the presence of call provisions (call), presence of sinking fund provisions (sink), the years to maturity (yrmat), the size of the company (asset), and the level of interest rates at the time of sale (7’yzq (see blackwell & kidwell, 1988; ederington, 1976; and sorensen, 1979 for discussions of the determinants of interest cost for corporate bonds). these variables are used as control variables (controls) in the ols regression models testing the effect of poison put provisions. for the control model, we estimate the yield spread over treasury (spd) using the full sample as follows: + f&asset+ &size+ f$jyld+q. (1) the value of coupon reset and poison put provisions is estimated while controlling for the rjr/nabisco takeover with the model: spd=controls+pfi~r+ plor~~~~+p,,put+p,2~~~~~~+~i. (2) we estimate the relation between the yield spread and the degree of event risk protection provided in the bond with the model: spd=controls+~~~jr+~~~eset+~,~med+~~~~~ +p,slow'rjr+ei. (3) we estimate the relation between yield spread, the inclusion of a coupon reset provision, and the degree of credit quality with the model: + p17rese7"bbb+f3,8reset"bb+ei, (4) where: spd = yield spread of the issue over treasury of the same maturity. znd = zero or one variables for industry, where transportation (tran), trade (trade) and miscellaneous classification (mix) equal one and manufacturing is the reference group. rate = zero or one variables for bond ratings, where a, bbb, and bb are all equal to one, and aa is the reference group. sink = one if the issue has a sinking fund, and zero if it does not. call = one if the issue is callable, and zero if it is not. yrmat = the natural logarithm of the issue’s maturity in years. asset = the dollar size of the company’s total assets in billions. 152 financial services review, 3(2) 1994 szze = the dollar size of the bond issue. tyld = the average daily interest rate on ten-year and longer u.s. treasury bonds on the date of the bonds issue. reset = one if the issue has a coupon reset provision, and zero if it does not. put = one if the issue has a poison put provision, and zero if it does not. rjr = one if the bond is issued following the takeover, and zero if preced ing. med = one if the bond’s poison put ranking is e2 or e3, and zero otherwise. low = one if the bond’s poison put ranking is e4 or e5, and zero otherwise. discussion of the variables because most of the model’s control variables have been used in previous studies, see equation (l), the discussion here is limited to the variables that test our hypotheses. equation (2) adds the rjr variable to test whether the rjr/nabisco takeover affected the yield spread on new issue bonds. we believe that the rjr/nabisco takeover caused a basic shift in the premium investors’ demand for bearing event risk when purchasing new bond issues. thus, we expect b9 > 0. the price differential between coupon reset and poison put provisions is tested by including the resetand putvariables. because put provisions were issued before and after the rjr/nabisco takeover, the interactive variable pljt*rjr is included in the model. as stronger event risk covenants were issued following rjr, we expect pi* > pii. additionally, as s&p ranks coupon resets as providing stronger protection than poison puts we expect pi0 < (pii + pi& equation (3) analyzes the price differential between different poison put covenants. to test this effect, variables representing the strength of the poison put protection are examined (reset, med and low, with rjr*lowrepresenting the post rjr low covenant ranking). in addition, rjr is included to control for changes in the environment for corporate control. if the coefficients on the test variables decrease as the put provisions become stronger, investors demand higher yields for bonds with lower event risk protection. thus, we expect (pm + pis) > pi3 > pia. equation (4) analyzes the effect of a bond’s credit rating on pricing coupon reset provisions. to test this effect, three interactive variables are analyzed (reset*a, re set*bbb, and reset*bb). the cross product variables test how the value of event risk protection changes with the credit quality rating. if coupon reset provisions provide greater protection for lower quality bonds, then the relation between yield and the cross products will decrease with bond quality. thus we expect pi6 > pi7 > pis. vii. findings table 4 presents the empirical results. equation (1) contains only the control variables. for brevity, the discussion of the control variables is kept to a minimum. the model explains 45 percent of the interissue variations in the yield spread. the coefficient on tran is positive and significant at the five percent level, indicating that the yield for transportation companies is greater than that for manufacturing company issues. although not all the credit rating coefficients are significant, they do increase monotonically as the credit quality declines. t a b l e 4 s q rd im ry l ea st -s qu ar es r eg re si on s t es ti ng t he e ff w t of e ve nt r is k p ro te ct io n on i ni ti al y ie ld o ve r t re as ur y (s p d ) fo r b an ds i ss ue d fr om j an ua ry 1 98 7 to j un e 19 90 3 ~~ ~~ ~~ vm m p a e : po tr a n tr a d e m ls c a b b b b b si n k c4 u yr m a t a ss et sx ze ty ld $j t.2 9 0. 27 0. 15 0. 05 0. 19 0. 60 lk z* 0. 15 0. 05 0. 20 a l.0 1 0. 01 -0 .1 3 2. 61 ” 2. 08 * 1. 55 0. 75 1. 77 5. 26 * 1. 22 0. 72 3. 52 * -1 .5 6 2. 40 * $ -2 .0 4* & 1. 10 0. 20 0. 21 0. 10 0. 18 0. 55 217 % z 0. 11 0. 15 -0 .0 1 0. 01 -0 .1 2 e i. 88 i. 55 2. if i. 45 k u 4. 97 * f0 .8 7* . 1. 59 2* 35 * -1 .8 3 2. 40 * -1 .8 6 8 t e st va r ia 3l es r jr r am pu t pe er a d jr 2 0. 45 0. 26 3, 44 1 2; : 1 * ;:i : l * -5 .3 8 0. 49 -2 .3 4* eq ua fim s po tr a n tr a d e n is c a b b b b b si n k cz a ll yr kf a t a ss et sm ty ld ?z ? 3 1. 13 0. 22 0. 20 0. 09 0. 17 0, 54 2. 16 0. 15 0. 11 0. 16 -0 .m 0. 01 -0 .1 2 tva b 1. 93 1. 67 2. 15 * 1. 32 1. 71 4. 82 * 10 ,8 7* 1. 13 1. 58 2. 38 * -1 .8 3 2. 3p -1 .9 2 e qu at io n 4 b et a i. 08 0. 20 0. 13 z ! 0. 17 0. 58 2. 70 0. 19 @ 08 -0 .0 1 0. 01 -1 .1 1 tv & +3 1. 96 1. 63 1. 46 1. 69 5. 36 ” 12 .5 4” 1. 55 * 1. 30 r2 . * -2 ,w * 3. 68 * -1 .8 3* t e sr v 5l r ia b l. e s eq ua tio ns e qu ai ia n 3 b et a tv ah e f% pa tic m 4 l3 et a tv al ue r jr r es ht m el ) lo w to w *r jr r es et *a a r h .e t* a ~s ~~ 3b b r es e, l’* b b a d jr 2 !z : -0 .3 6 0. 13 0. 40 -0 .1 8 0. 50 * -2 .2 4* 0. 57 2. 55 * -0 .8 2 !z % q z 1 ,. -0 .2 3 0. 25 -0 .3 1 -2 .2 2 0. 55 -0 .7 4 1. 10 -1 .0 4 -6 *2 0* n ot e: *d if fe re nt fr om m a at th e 5 pe rc en t l ev el . 154 financial services review, 3(z) 1994 the estimated coefficients on size, yrmat and tyld have the expected sign and are significant. the estimated coefficients on the remaining control variables are not significant. equation (2) tests whether investors’ perception of event risk changes after the rjr/nabisco takeover. in addition, the regression tests the difference in the value of event risk protection before and after the takeover. the estimated coefficient on rjr is positive and significant and indicates that bond prices are negatively affected by the rjr/nabisco takeover. following the takeover, new bond issues sold at yields 26 basis points more than similar bonds issued prior to the takeover. this suggests that the rjr/nabisco takeover caused a shift in investor perceptions of event risk. more than likely, this occurred because investors became more aware of the magnitude of losses that bondholders can suffer because of event risk. the variables reset and put examine the event risk protection provided by the two major types of event risk covenants. before the rjr takeover, bonds with put provisions (put) had yields of 37 basis points more than similar bonds without event risk protection. the positive and significant coefficient on put suggests that the presence of a put provision signaled that these bonds had higher expected event risk and/or that these poison puts provided bondholders with inadequate event risk protection. we know from table 3, however, that all bonds in the sample during this period were rated e4 and e5 and, thus, had low levels of event risk protection. after the rjr/nabisco takeover, bonds with reset provisions (reset) had yields of 35 basis points less than similar bonds without event risk protection. these bonds were highly rated, the majority of which were rated as el, the highest level of event risk protection available. for poison put bonds sold after the rjr/nabisco takeover (put*rjr), the coefficient is negative and significant. these bonds sold for only one basis point (pt, + 13t2) less than similar bonds without event risk protection. this reflects the changing composition of puttable bonds following the rjr/nabisco takeover. after the takeover, stronger poison put provisions were used. equation (3) examines how bond yields change with event risk quality rankings. the findings indicate an inverse relationship; that is, as event risk quality declines bond yields increase monotonically. reset bonds, which are all highly ranked, sell for 36 basis points less than similar bonds without event risk protection. poison put bonds with medium event risk protection (med) sell at yields which are not significantly different from similar bonds without event risk protection. poison put bonds with low event risk protection (low), have yields of 40 basis points more than similar bonds without event risk protection, over the sample period; the cross product variable low*rjr is not statistically significant. equation (4) estimates the value of reset type covenants for bonds with different credit ratings. the coefficients on the cross product variables tend to decrease as the credit rating of the bonds decline. however, only the cross-product for bonds with reset type event risk protection and low credit quality (reset*bb) is statistically significant. this suggests that the value of event risk protection is greater for bonds with lower credit ratings. viii. summary this paper examines the valuation of the two major types of event risk indenture provisions on corporate debt-poison puts and coupon resets. overall, we find that investors place a positive value on coupon resets-they reduce yields on bonds by 35 basis points-and place coupon resets versus poison puts 155 virtually no value on poison puts. we also find that the rjr/nabisco takeover was a watershed event with respect to the design, the pricing, and the frequency of event risk protection in corporate debt. more specifically, before the rjr/nabisco takeover, bonds with poison put provisions had low event risk protection, and these bonds sold for penalty yields compared to similar bonds without event risk protection. after the takeover, bonds had stronger event risk protection, and bonds with put provisions sold for about the same yields as bonds without protection. bonds with the more recently developed coupon reset provisions sell for less than unprotected bonds. in addition, we find that the quality of the event risk protection, as measured by event risk rankings, significantly affects the yield on new bonds. bonds with high event risk rankings sell at lower yields than similar bonds without event risk protection, bonds with medium event risk rankings do not have their yields lowered, and bonds with low event risk rankings sell at penalty yields. finally, holding event risk rankings constant, we find that the value of event risk protection is greater for bonds with low credit ratings. this research indicates that the market values high quality puts, all else held constant. this is a significant factor for individual investors who do not have the resources for monitoring bondholder stock holder conflicts. with a high quality put provision a bond holder may reduce the effort in monitoring. this reduction in monitoring costs, assuming economies of scale in monitoring, differentially advantages small versus large investors. acknowledgments we are grateful for financial support from the carlson school of management, the corporate affiliates program at the university of connecticut and the society for fixed income research. notes 1. only one poison put type of provision in our sample has been ranked e2. 2. united technologies issued an e3 rated coupon reset bond that is triggered by a designated event and a rating downgrade of one full category by both s&p and moody’s, or any downgrade that results in speculative grade ratings from both agencies. 3. see frank (1989). 4. see winkler and white (1989). 5. based on our sample no coupon resets or poison put covenants ranked higher than b4 were issued prior to the rjr/nabisco takeover. 6. salomon brothers has argued that industrial, transportation, and natural gas pipeline companies as industrial sectors most vulnerable to event risk (d’amico, 1990). references black, f., & cox, j.c. (1976). valuing corporate securities: some effects of bond indenture provisions. journal of finance, 31(2), 351-367. 156 financial services review, 3(2) 1994 blackwell, d.w., & kidwell, d.s. (1988). an investigation between public sales and private placements of debt. journal of financial economics, 22 (december), 253-278. bicksler, j.l., & chen, a.h. (1992). pricing corporate debt with event risk provisions. international review of financial analysis, i, 5 i-64. crabbe, l. (1991). event risk: an analysis of losses to bondholders and ‘super poison put’ bond covenants. journal of finance, 46 (2), 689-706. d’ amico, e. (1990, april 2). what do bondholders really want? the event risk debate enters a new phase. investment dealers digest, 16-20. ederington, l.h. (1976). negotiated versus competitive underwritings of corporate bonds. journal of finance, 31 (l), 17-28. fama, e.f., & miller, m. (1972). the theory offinance. holt, rinehart and winston, new york. frank, d. (1989). pity the poor old retail bondholder. the banker (february), 37-38. jensen, m.c., & meckling, w.h. (1976). theory of the firm: managerial behavior, agency costs and ownership structure. journal of financial economics, 3(4), 305-360. lehn, k., & poulsen, a. (1991). contractual resolution of bondholder-stockholder conflicts in leveraged buyouts. journal of law and economics, 645-673. mason, s.p., & bhattacharya, s. (1981). risky debt, jump process, and safety covenants. journal of financial economics, 9(3), 281-307. mitchell, c. (1992, october 7). marriott plan enrages holders of its bonds. wall street journal, pp. 117-162. smith, c.w., jr., & warner, j.b. (1979). on financial contracting: an analysis of bond covenants. journal of finance, 7(2), 863-870. sorensen, e.h. (1979). the impact of underwriting method and bidder competition upon corporate bond interest costs. journal of finance, 34(2), 863-870. winkler, m., & herman, t. (1988, october 25). takeover fears rack corporate bonds: investors avoid industrial issues. wall street journal, p. c 1. winkler, m., & white, j.a. (1989, march 22). shock still clouds blue-chip corporate bond market. wall street journal, p. cl. pii: s1057-0810(00)00039-1 editorial from the special issue editor the fourth issue of vol. 8,toward a theory of instruction: teaching individual financial management, centers on education theory and practice. this marks the first time for a special edition of financial services review, and it has been my pleasure to serve as the guest editor. what makes this issue special is its personal focus; the authors reveal their thoughts, practices, and results of teaching individual financial management. the first two articles relate to general curriculum design. the lead article is entitled “an integrative approach to using student investment clubs and student investment funds in the finance curriculum.” authors brian grinder, dan w. cooper, and michael britt argue that experience in investment clubs and managing investment funds can effectively ground student learning and increase finance comprehension. in their article ellie fogarty and herbert mayo describe the senior thesis in finance required at the college of new jersey. designed as a “spire,” it addresses the issue of course sequence and presents an interesting way to allow finance students to gain specialized knowledge not offered within the formal curriculum. the next two articles use research techniques to address student learning results. in “student learning style and educational outcomes: evidence from a family financial management course,” jonathan fox and suzanne bartholomae test for a relationship between student learning style and success in an individual financial management course. in “does education affect how well students forecast the market?” john c. alexander, jr. and robert b. mcelreath find that education improves the forecasting ability of students. the next several articles directly relate to teaching personal financial management, from introductory to more advanced topics. “an integrated model for financial planning,” written by natalie chieffe and ganas k. rakes, provides a framework for structuring the financial planning course and student learning. ronald r. crabb shares his teaching methods and gives a how-to guide in “cash flow: a quick and easy way to learn personal finance.” thomas h. payne and j. howard finch discuss the importance of having students truly understand a model’s assumptions, inputs, sensitivity to error, and practical limitations in “effective teaching and use of the constant growth dividend discount model.” in his article “learning by doing: offering a university practicum in personal financial planning,” tom eyssell financial services review 8 (1999) ix–x 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00039-1 describes the practicum at his school as providing valuable and rewarding experiential learning opportunities for students. the final two articles of the special issue on education deal with web-based teaching resources. “the internet in the personal finance course,” by walt woerheide, explores the extent and implications of current and future internet integration in personal finance textbooks. finally, stuart michelson and stanley smith share how-to create and teach with personal web pages in “applications of www technology in teaching finance.” on a personal note, i give my special thanks and encouragement to the all of the authors who submitted manuscripts (37 in total) for this special issue. i look forward to their continued participation in the dialogue on teaching individual financial management. thanks also to the long list of academy of financial services colleagues who gave so unselfishly of their time to review and rereview submissions. jill lynn vihtelic x j.l. vihtelic / financial services review 8 (1999) ix–x pii: 1057-0810(95)90008-x financial services review, 4(2): 137-156 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. household insolvency: a review of household debt repayment, delinquency, and bankruptcy sharon a. devaney ruth h. lytton this review paper explores the issues related to the meaning and measurement of insolvency within the domain of householdfinances. conceptual and empirical evidence to explain the onset of insolvency is reviewed. predictive models andfinancial ratios are presentedas techniquesfor identifying insolvent households. implicationsformonitoring of solvency by households and responses to insolvency are presented, i. introduction the term insolvency is more frequently associated with a business entity than a household. the concepts of solvency and its opposite, insolvency, can be simply defined as having either a positive or negative net worth. in the equiry sense, insolvency refers to the failure to submit the timely repayment of debts as they mature. this situation can result in an increase in liabilities and a reduction in the equity in assets held. in the bankruptcy sense, insolvency means that net assets at fair market value are less than liabilities, which can necessitate the liquidation of assets through a court-ordered bankruptcy process (becker, 1992). factors that contribute to insolvency and the implications for the household vary significantly; we know that consumers are taking on more debt and more consumers can be characterized as insolvent. in fact, many policy observers have referred to the 1980s as a “decade of debt.” between 1979 and 1982, the growth in consumer installment credit outstanding was less than an annual compound rate of 6% per year. from the end of this recessionary period until 1987, consumer installment credit outstanding grew at an average compound annual rate of 18% (avery, elliehausen, & kennickell, 1987; luckett & august, 1985). it is estimated that sharon a. devaney l consumer sciences and retailing department, purdue university, room 216 matthews hall, west lafayette, in 47907. ruth h. lytton l department of housing, interior design and resource management, virginia polytechnic and state university, blacksburg, va 24061. 138 financial services review 4(2) 1995 personal debt increased by 79% between 1980 and 1990 while real personal assets rose by only 36% (federal reserve, 199 1, pp. 19-24). between 1981 and 1990, the percentage of disposable personal income consumed by consumer installment credit payments grew from 14.5% to 18.5% (courtless, 1993). in 1992, the ratio of consumer installment credit payments to disposable personal income fell to a seven-year low of 16.7%, indicating that consumers had begun to reduce their indebtedness to more manageable levels (scheld, 1993). however, scheld suggests that the apparent improvement in the total debt service ratio was due to interest reductions during the recession, not a decline in the level of total household debt. further, consumers have been replacing traditional credit card debt and personal loans with less expensive home equity loans and lines of credit which are classified as home mortgage debt. in addition, leases now account for 25% of new car purchases (cunniff, 1995). both of these consumer practices create debt obligations that are unrepresented in government credit statistics. cunniff (1995) reports that consumer installment debt increased a record $118 billion during 1994. this amount represents 17.8% ofpersonal disposable income and an even larger estimated 20% if amounts committed to both auto leases and home equity credit lines were included. it is estimated that total household borrowing for consumer debt and mortgages accounts for more than 90% of after-tax income. in comparison, total household debt represented about 70% of after-tax income in 1980. while the 1980s were the decade of debt, the 1990s have been described as the decade of repayment (hughes, 1991). and, if consumer borrowing trends continue, the 1990s may well represent a decade of “repayment” for too many consumers. economic conditions include a slower pace of income growth, flat housing prices, and sluggish economic growth. fiscal policies in many states and localities have caused significant tax increases. but for the creditors of the increasing numbers of consumers who have declared bankruptcy, the 1990s may not be the decade for repayment. the number of consumer bankruptcy filings almost tripled between 1985 and 1991, when there were 943,987 bankruptcies (bhandari & weiss, 1993). although 1992 represented a record year with 977,478 filings, the rate of increase slowed to 6.4%, compared to the 22.5% increase from 1990 to 1991 (singletary, 1993). of the 1992 filings, 92.6% were personal bankruptcies. filings decreased in 1993 for the first time since 1984. the increased bankruptcy filing rate has generated concern among creditors, legislators, and regulators responsible for laws governing the bankruptcy process. but the study of household insolvency goes beyond concerns over increasing debt levels and bankruptcy filings. insolvency, like debt, has “carrying charges” of direct and indirect costs. for consumers who are delinquent, late fees and other collection costs simply add to liabilities that are already not being reduced. although the intent of a bankruptcy filing is to give people a “fresh start,” filing fees, loss of assets, inability to use credit for a period of time, and the stigma of having declared bankruptcy cannot be ignored. these represent direct costs for the use and abuse of credit. there are also indirect personal, emotional, and psychological costs which are beyond the scope of this analysis. for businesses, the cost of doing business is affected. insolvent households directly impact the “write-off rate” and the profit margin. for example, an american banking association survey revealed that 25% to 30% of all bank consumer credit losses resulted from bankruptcy (“easing borrowers,” 1990). according to the credit union national association (“bankruptcy reform,” 1995), credit unions lost $710 million in 1993, a loss approximately 150% higher than nine years earlier. businesses are replacing in-house review household insolvency: debt repayment, de~nqueney, nnd buukr~~y 139 of customer applications with sophisticated customer screening services provided by credit bureaus to protect against potentially delinquent or fraudulent accounts. furthermore, these costs for losses and operation are indirectly passed on to consumers. this represents an increase in cost for those who pay on time as well as for those who do not. a further concern is the fraudulent or abusive use of bankruptcy filings by some consumers. finally, there is potentially a much larger cost to society when the economy is fueled through consumer spending on credit with little or no associated savings to provide capital for future investment in sustained economic growth. as a percentage of national output, savings has steadily declined from 12.3% in 1950 to approximately 2.4% in the 1990s (cunniff, 1995b). in summary, the nuances of applying and interpreting the concept of insolvency to a household financial situation should be of concern to financial professionals and to policy makers in business and government, as well as to individual consumers. the remainder of this article provides additional background on insolvency and bankruptcy then addresses the following questions: l what theoretical or empirical evidence is available to explain insolvency? * what predictive models are available for use by businesses for customer screening and by financial professionals who assist consumers? l what are the implications for credit grantors, financial professionals, and consumers? ii. background the condition of insolvency is contrary to the accounting concept of “a going concern” or the idea of a business functioning for an indefinite future. rlost households are not viewed as “going concerns,” although planning to provide for later retirement or an estate for future generations are reasons to save or realize a positive net worth. furthermore, while an insolvent business might cease to exist, dissolution of a household due to insolvency is not a viable alternative. although children grow up and leave home and parents divorce, the family, or an individual, must continue to function as an economic and financial unit. insolvency is often associated with bankmptcy, although insolvency, in the literal sense of a negative net worth, is not a r~uirement for bankruptcy filing. historically, the terms insolvency and bankruptcy represented different bodies of english law and different attitudes toward creditors. bankruptcy was an involuntary procedure designed to protect creditors through the confiscation and equal division of the debtor’s property among the creditors. debtors often faced imprisonment, at the expense of creditors. the second body of law purported to protect debtors, who would volunt~ily declare insolvency, give all their property to the court, and be discharged from debtors’ prison. debtors were still liable for the payment of their debts. the word bankruptcy is derived from the latin words for “bench” and “break” (luckett, 1988). the literal meaning of bankrupt is broken bench. under roman law, creditors would physically break the debtor’s workbench after gathering together and dividing up the debtor’s assets. the broken workbench served as both a punishment and a wring to other debtors. satisfaction of the claims of creditors and punishment of the debtors were the objectives of the early law. bankrupt persons were deprived of their civil rights. other societies required bankrupts to dress in distinctive garb. in 1705, english law provided for remission of the 140 financial services review 4(2) 1995 debts of ban~upts. the purpose was not a humane gesture to give the unkept a new start but rather a counter to the concealment of assets by debtors. almost 300 years later, a new start for debtors and the availability of debtors’ assets for meeting the needs of creditors and debtors remain a concern. individuals usually file under chapter 7, straight bankruptcy, or chapter 13, the wage earner plan. chapter 11 is available for consumers with unsecured or secured debts that exceed the limits set for chapter 13, and chapter 12 allows family farmers with regular income to restructure their debt while remaining in operation. there are no restrictions on the number of chapter 11, 12, or 13 filings per household, and these chapters allow debtors to protect more assets. a chapter 7 filing discharges most debts; however, a person cannot tile bankruptcy again for six years. under straight bankruptcy, people are allowed to keep a smalt equity in their homes, an inexpensive automobile, and limited personal property. state or federal laws govern what can be kept. some debts such as education loans, fines, alimony, child support, and income taxes may never be excused. under a chapter 13 filing (wage earner), the person is allowed to keep all assets and is protected from creditors while debts are being repaid according to the court-approved plan. the time period is usually three years. chapter 7 may be used by both business and nonbusiness petitioners while chapter 13 is limited to nonbusiness petitioners. iii. characteristics associated with debt repayment, delinquency and bankruptcy: empirical evidence the study of consumer insolvency is not a well-developed science. no theory has emerged to explain, or predict, the onset of insolvency. although fragmented and not without its limitations, research on related topics such as consumer debt repayment, delinquency, and bankruptcy offers some empirical evidence to explain the incidence of consumer insolvency. a. repayment of consumer debt the lack of delinquency in repaying debt is an important indicator of the quality of credit. the 1983 survey of consumer finances (scf) was one of the first surveys to provide information obtained from borrowers about their debt repayment behavior. prior to the 1983 scf, most info~ation about delinquent debt repayment came from lenders. approximately 22% of the 3,824 respondents of the 1983 scf reported that they had made late payments or missed payments at some time during the previous year. using univariate and bivariate analysis, sullivan and fisher (1988) showed that the incidence of slow or missed payments in the 1983 scf decreased as income increased and as the age of the borrower increased. the risk of payment difficulty was very high (37%) for the lowest income group, below $lo,o~, and very low among the highest income group (7%), $50,~0 and over. as the household head aged, generally the probability of payment difficulty declined. further, regardless of the level of income, respondents with no or very low liquid asset balances had an above-average tendency for payment difficulty, renters were almost twice as likely to report having had debt payment difficulties as were homeowners. the analyses showed that missed or slow payments were more likely to occur among nonwhites or hispanics, in households with less-educated heads, and in household insolvency: debt repayment, delinquency, and bankruptcy 141 households with higher ratios of mortgage or consumer debt payments to income. those who had obtained credit from finance companies, stores, or dealers were substantially more likely to be late or behind than those who borrowed from banks, credit unions, or savings and loan associations. canner and luckett (1990) pointed out a limitation of the sullivan and fisher study that is, the respondents had already been screened by lenders, who had weeded out those at greatest risk of default. therefore, the results showed factors associated with missed payments given the credit standards prevailing in the marketplace but not which factors were associated with risk of default. further, the study did not consider interrelationships among the variables. subsequently, canner and luckett (1990) used multivariate analysis to study the data. a logistic regression was performed to estimate the probability that a borrower would be late or delinquent holding the values of the other variables constant. they found that the variable for prior credit history-whether the person had been previously rejected for credit-had the greatest statistical significance as a predictor of late or missed payments. other important variables were age and amount of liquid assets relative to debt. just as sullivan and fisher found a strong positive relationship for age to repayment of debt, and for the ratio of liquid assets to consumer debt and timely repayment of debt, so did canner and luckett. further, households with more children had a greater probability of missed or late payments. in contrast to sullivan and fisher’s findings, canner and luckett did not find a significant relationship between income, education, or housing tenure and late or missed payments. canner and luckett suggested that creditors may have done a competent job of accounting for income in the loan approval process. the federal reserve board surveyed 1,534 families as part of the survey of consumer attitudes (canner & luckett, 1991) to obtain information on consumer debt. eighty-five percent of all households had an outstanding debt obligation or access to a line of credit. among all households, 45% had only consumer credit debt; 3% had only home mortgage debt, and 38% had both outstanding mortgage and consumer debt. about 14% of households with debt reported that they were late for at least one of the scheduled debt payments. payments that were frequently reported as being late were vehicle loans, other types of non-credit-card installment debt, and credit card debt. persons who were more likely to have payment problems were renters, divorced or separated persons, and those with the highest debt-service burdens. households headed by people under age 35 were nearly four times as likely to report payment problems as were those headed by an individual at least 55 years of age. canner and luckett (1991) found that 9% of all indebted households fell behind more than 30 days on one or more of their debt obligations in the year prior to the survey. roughly 3% of all debtors fell more than three payments behind within a 1zmonth period. of the families experiencing payment problems, 55% indicated that they became overextended; 24% either lost their jobs, were not working, or had experienced a cutback in the number of hours worked, and 6% experienced medical-related problems. nearly 40% of those having debt repayment problems reported that they paid delinquent bills the following month or “when they were able.” others with debt problems reduced spending, took second jobs, worked longer hours, sold items to raise money, or borrowed from friends or relatives. financial services review 4(2) 1995 b. default on auto loans the possibility that the relationship of default risk to personal attributes such as occupation and employment might vary when size of down payment varied was investigated by peterson and peterson (1981). the researchers postulated that interactions could exist among borrower characteristics, loan terms, and default risk. using data from the federal reserve system on commercial bank lending to consumers for auto loans from 1965 to 1970, peterson and peterson showed that changing the amount of down payment altered the risk of default. they found that default rates fell substantially if down payments were as high as 20% of the loan. peterson and peterson extended the study to examine default rates by occupational group and size of down payment. professionals had a much lower default rate than other groups and drivers or laborers had the highest default rate. however, when cash down payments were over 30% of the auto loan, the borrower’s occupation was not significantly related to default rates. then, peterson and peterson compared age to the size of the down payment and found that when down payments exceeded 30%, default rates wereconsiderably reduced for all borrowers but default was still twice as high for younger as compared to older borrowers. peterson and peterson concluded that creditors should consider the size of the down payment when evaluating credit risk. however, they cautioned that the down payment needed to represent voluntary saving by the borrower, not borrowing from other sources. livingstone and lunt (1992) looked at the growth of personal debt in the united kingdom, which has experienced similar growth of debt as in the united states. in 1981, the ratio of outstanding debt to annual household disposable income was 8%; in 1988, it was 14%. livingstone and lunt believed that a possible explanation for the increase in personal debt would involve psychological, social, and economic determinants. analysis of personal debt and debt repayment of a sample of predominantly lower-middle-class/upper-working class respondents in 1989 produced results which in several respects contradict those previously reported. debt repayment was not significantly predicted by so&demographic variables such as social class, age, or the number of dependent children; however, the amount of disposable income was an important predictor of the amount of regular debt repayment. the more income people had, the more likely they were to make regular payments on debt. the livingstone and lunt study found that those who repaid a greater amount were more concerned with personal achievement and self-direction. the findings showed that having a positive attitude toward credit was a predictor of repayment of more debt. also, those who repaid more acknowledged that their use of credit was a result of external uncontrollable needs and not caused by their own greed or simply the convenience of credit. the respondents viewed debt repayment as a budgeting strategy. a limitation of this study is that the analysis focused on the amount of debt repayment, not on late payment or default. c. propensity for insolvency the effect of age, income, and marital status on the propensity for insolvency was analyzed using data from the 1983 and 1986 surveys of consumer finance (devaney & hanna, 1994). insolvency was defined as having a net worth less than one month’s income. analysis showed that age of the household head had a negative relationship with insolvency. income had a strong negative effect with on the propensity for insolvency. in the first time period (1983), married couples had lower predicted insolvency rates than other household household insolvency: debt repayment, delinquency, and bankruptcy 143 types, but in 1986, the relationship between marital status and insolvency was unclear. household size, education, and race were not significant variables related to insolvency in either year. d. bankruptcy according to hira (1992), empirical evidence to support a philosophy of bankruptcy is fragmented and incomplete because it has been too broadly focused. nevertheless, re searchers tend to formulate generalizations based on the empirical evidence which is available. limitations of early empirical studies included the size, location, and response rate of the sample. more recently, researchers have used data from large, nationally representative samples. often studies assume that bankruptcy is associated with a fault and this assumption may introduce bias into the conclusions (hira, 1992). this section reviews several empirical studies on bankruptcy. researchers have sought to identify causes of bankruptcy by studying the financial and demographic characteristics of those who claim bankruptcy. studies in the 1960s in michi gan and utah found that most bankrupts were employed in lower-paying jobs in unskilled or semi-skilled manual labor. however, most were employed when they filed forbankruptcy. similarly, two surveys in the 1980s showed that about 80% of bankrupts were employed when they filed and that, of those who were employed, most were in blue-collar jobs. in one of the studies, 20% of the families had two incomes (sullivan, warren, & westbrook, 1989). in research using aggregate data on consumer bankruptcies from 1945 to 1981, shepard (1984) found a positive relationship between divorce rates and the ratio of consumer installment and non-installment debt to income and bankruptcy rates. also, nonwhites were positively associated to the rate of chapter 7-straight-bankruptcies. he showed a negative relationship between residential wealth and bankruptcy rates. overall, credit debt accounted for about 80% of the increment in bankruptcy filings during the period investigated. sullivan et al. (1989) collected data on 1,529 families who filed for bankruptcy in illinois, pennsylvania, and texas in 1981 with the intent of identifying characteristics of bankruptcy filers and causes of filing for bankruptcy. they found that bankrupts earned about a third less income than the average earner and that their households were larger (3.4 members compared to 2.7 members for the general population). while mortgage debt for the bankrupts was about average, consumer debt was excessive when compared to the general population ($10,800 compared to $2,400). mean unsecured debt amounted to $15,500 for each bankrupt debtor; this was about equal to the average bankrupt’s annual income. most bankruptcy filers were unemployed during some part of the two years prior to filing for bankruptcy. twenty percent of bankrupt debtors were currently or formerly involved in entrepreneurial businesses. moving a lot was characteristic of chapter 7 bankruptcy filers (sullivan et al., 1989, p. 246). using data from the 1983 scf, sullivan et al. compared the net worth of respondents from the general population with that of the bankrupt debtors. allowing for differences in method of calculation, sullivan et al. concluded that “about one-third of the general population has a net worth of less than $5,000, while 84% of the debtors are worth less than that amount” (p. 7 1). median net worth for the bankrupt debtors and the general population equalled $8,100 and $24,600, respectively. not all debtors were insolvent, as 16% reported a positive net worth greater than $5,000; homeowners represented 94% of this group, 144 financial services review 4(2) 1995 sullivan et al. found that single women who filed for bankruptcy had incomes very similar to single women not in bankruptcy ($10,600 for bankrupts and $14,100 for those not in bankruptcy). what distinguished women in bankruptcy from others was the lack of supplemental income received by other women. women in the general population received about $4,200, or 30% of their total family income, from other sources while single women who filed for bankruptcy received an average of only $500 annually from other sources. “these data suggest that supplemental income, such as alimony or a child’s income, may represent the difference for many women between staying out of bankruptcy and going in” (sullivan et al., 1989, p. 156). sullivan et al. found that married couples in financial trouble were one-income families in far greater proportion than one-income families in the general population, during the year prior to the bankruptcy filing, only about one-third of the wives in bankruptcy were employed, compared with almost two-thirds of the wives in the general population. of the wives in bankruptcy who were employed, many worked only part time. sullivan et al. concluded that the bankruptcy data portray the increased risk faced by lower-income families that do not follow the national trend toward two incomes (p. 157). an investigation of the relationship of medical debt burden indicated that medical debts did not play a central role in most consumer bankruptcies. while medical-debt-to-income ratios (constructed by the researchers) did not vary significantly by state or district, they varied by chapter. those in chapter 7 bankruptcy owed more medical debt than those in chapter 13; the mean medical debt/income ratio of 0.25 for chapter 7 was significantly higher than the 0.09 of those in chapter 13. interestingly, the joint filers’ mean medical debt/income ratio was 0.12; the mean of single-filing males was 0.24, and that of single-filing females, 0.43 (p. 171). the impact of medical debt on single-filing women was much greater because they have lower incomes available to pay their debts. sullivan et al. speculated that women were more likely to be employed in retail trade and personal services and to be without medical coverage. information about the filer’s availability of fringe benefits is not asked by the court system. examination of the credit card debt held by the tilers in the study revealed that almost one-third owed credit card debt equal to or greater than three months’ gross income, a debt-to-income ratio of 0.25. nearly 13% of the debtors owed more than a half-year’s income in credit card debt. yet, only two percent of the respondents met the researchers’ criteria for credit card abuse: high credit card debt/income ratio, high proportion of unsecured debt in credit cards, and in the top 15% of the absolute amount of credit card debt carried into bankruptcy (sullivan et al., 1989, p. 185). the debtors with the worst credit card/income ratios were more often debtors with low job tenure or income swings. hira (1992) compared american and canadian bankrupts’ attitudes and satisfaction with the bankruptcy process. data collected in manitoba (canada) and iowa in 1988 showed that the demographic profile of filers was remarkably similar. in general, the filers were young, male, and married and had children. the largest proportion of debtors in both countries not only borrowed from banks and retail stores but also owed large sums to the two sources. a majority of the debtors believed that bankruptcy should be used only as a last resort. also, a majority from each country said that filing for bankruptcy provided relief from debt, saved them from creditors’ actions, and improved their family living conditions. about half the debtors in both iowa and manitoba indicated that their bankruptcy was caused by too much borrowing and that the final decision to file was made because creditors had household insolvency: debt repayment, delinquency, and bankruptcy 145 started collection efforts. a larger proportion of canadians than americans learned the importance of setting up a budget, saving regularly, not using credit cards, and paying by cash only. hira cautions against drawing conclusions from the results because the response rate was low and only one province in canada and one u.s. state were included in the survey. in summary, sullivan et al. found that consumers in bankruptcy looked like other americans in the workplace but had very different financial circumstances. debtors tended to earn less and owe more than other americans. families with serious debt problems had about one-third less income than average earners. bankrupts may have experienced unem ployment problems within the two years prior to tiling bankruptcy. about one-fifth of bankrupts had been involved in entrepreneurial businesses. while mortgage debt was similar to other families, families in serious economic trouble had excessive consumer debt. household size tended to be larger for families with debt problems. single women were shown to be especially vulnerable because of their low incomes and possible lack of fringe benefits. the research described here provides insight into the complexity of analyzing late payment behavior, default, and bankruptcy. although the conceptual review has suggested that attitudinal and life-style factors as well as financial and life events impact repayment of debt, factors measured were limited to easily observed and readily quantifiable items. but, the studies show that individuals and families who are most at risk are more likely to become insolvent. these at-risk individuals and families include persons who are younger, nonwhite, divorced or separated, renters, single-earner low-income families, those with high debt service levels, and those with little or no liquid assets. also, sullivan et al. (1989) found that single women who received little or no supplemental income such as child support or alimony and single women with high levels of medical debt to income were particularly vulnerable to bankruptcy. the studies are not detailed enough to pinpoint the causes of insolvency but the underlying current is apparent: one or more related demographic charac teristics in conjunction with a life event or attitudinal factor. iv. predictive models of insolvency although the previous studies are useful in describing the financial statuses of house holds, a shortcoming of these studies is that they are not prescriptive. while they tend to pinpoint the problems that appear to be related to insolvency, they neglect to provide businesses or consumers with the necessary information to estimate the incidence of insolvency. consumer credit, and more recently commercial credit and mortgage lending, rely on credit scoring systems to determine creditworthiness for establishing accounts as well as later strategies for account management. selected financial ratios represent just one information source considered by some scoring schemes. recently, financial educators, counselors, and planners have advocated the use of financial ratios as a useful tool to help consumers monitor financial progress and anticipate problems. these predictive models and the assumptions upon which they are based are now reviewed. a. credit scoring: avoiding the insolvent the origins of credit scoring are traced to the 1940s (durand, 1941), but it was not until the 1960s that there was widespread interest in the development and use of credit scoring 146 financial services review 4(2) 1995 systems for establishing creditworthiness among applicants. an increasingly competitive market, a rapidly expanding and mobile consumer market, and later concerns over fairness and discrimination in lending practices increased demand for automated credit evaluation systems. more recently, other factors, such as economies of scale associated with automation, the availability of management for other tasks, and a rapidly changing marketplace charac terized by inflation and recession, have contributed to the expansion of the technology. originally, credit scoring referred to the use of statistical methods for predicting the likelihood of default by comparing key applicant characteristics with a known profile to classify credit applicants, on the basis of the score generated, as “good” or “bad” risks. “application scoring” later evolved into “behavioral scoring” for tracking and predicting the performance of individual accounts on an ongoing basis. although prediction of delinquency and bankruptcy are major concerns of behavioral scoring, other issues include account attrition, account collection efforts, changes in account credit limits, account reissue periods, and other account promotional/marketing decisions (pellegrino, 1988; radding, 1992; rosenberg & gleit, 1994). neural networks, or expert systems technology, offers yet another advancement over more traditional statistical procedures by predicting multiple, as opposed to binary, outcomes (jenson, 1992; jost, 1993). scoring systems can be developed in-house or in conjunction with a major vendor (e.g., american management systems, fair-isaac, etc.); services can be purchased from the four major credit bureau companies, or in-house and credit bureau services can be overlapped to yield additional information (robins, 1992). although a comprehensive review of the development of scoring models (see capon, 1982; makowski, 1985; rosenberg & gleit, 1994) and the statistical procedures (see chhikara, 1989; collins & green, 1982; grablowsky & talley, 1981; rosenberg & gleit, 1994) on which they are based are beyond the scope of this article, two issues are of primary concern in the context of consumer insolvency. the first considers the use of these tools for managing risk and profitability, while the second considers the characteristics used by the system for predicting consumer outcomes that could result in insolvency. initially, credit scoring systems offered the benefits of efficiency, consistency, and relative accuracy in determining the risk associated with a potential credit applicant. but rapid advances in technology and the ability to accumulate and use consumer profile data have extended the use of these tools from risk management to profitability management. singularly and in combination, the variety of consumer profile reports (e.g., credit bureau reports, behavior scoring reports, chargeoff and/or bankruptcy prediction reports, attrition reports, prescreening profiles, recovery scoring) allow a creditor to manage a portfolio not only to predict not only the “good’ and “bad” accounts but also to more accurately identify the “optimal crossover point,” or the point at which costs and losses associated with bad accounts will exceed profits from additional “good” accounts (mccorkell, 1994). in this environment, creditors can employ “lifetime value analysis” (irvin, 1994) or “adaptive control models” (marshall, 1992) to facilitate customer management and analysis as opposed to account management and analysis. these tools enable managers to achieve optimal profitability while reducing chargeoff rates, collection costs, and the high costs of attracting new accounts in an increasingly competitive marketplace. in other words, creditors can optimize the mix of consumer types with unique use and repayment strategies to maximize revenues generated versus the cost of obtaining and maintaining accounts. table 1 selected consumer characteristics considered in credit scoring family status~i~g a~~~~~n~ marital status own/rent dwelling postal code telephone length of time at current residence number of dependents age of automobiles) automobile balance(s) co-applicant information. if any employment occupation employment status length of time with employer education personal information age gender geodemographic information financial history (may duplicate some credit bureau irlfbrmarion) income debt ratios ratio of regular expenditures to income monthly income less committed payments total monthly credit payment expenses credit references (account types: bankcard, fmance company, etc.) bank references (account types: checking, saving, both, etc.) largest previous amount of debt other loan ~~~ent~level of indebtedness amount of loan purpose of loan number of monthly payments delinquency during performance period reviewed account activity during performance period reviewed account balance during performance period reviewed amount past due on account returned checks on account age of account(s) credit bureau information (my duplicate somefirtancial history) credit payment experience past due balance(s) derogatory information per tradeline derogatory information from public records number of tradelines type(s) of tradelines age of the oldest/newest tradeline inquiries creditor info~ation needs to support this sophisticate and indepth analysis of con sumers is based on the historical five cs of credit evaluation--character, capacity, capital, conditions, and collateral. but the sophisticated systems exceed what was historically a creditor’s personal knowledge of the debtor’s situation to a study of over 100,000 consumers 148 financial services review 4(2) 1995 randomly selected from a national sample for initial analysis of over 350 sets of charac teristics (gothe, 1990) or a behavior scoring model built on up to 200 variables (robins, 1993). in a review of characteristics consistently considered in credit scoring schemes interna tionally, friedland (1993) suggested a framework comprised of five categories of predictors: family status~iving a~~gements, employment, personal info~ation, financial history, and credit bureau info~ation. this framework was used, as shown in table 1, to group common predictor characteristics included in u.s. credit and behavioral scoring schemes (apilado, warner, & dauten, 1974; boyes, hoffman, & low, 1989; gothe, 1990; jenson, 1992; long & mcconnell, 1977; lyons, 1993; makowski, 1985; overstreet & kemp, 1986; robins, 1993; rosenberg & gleit, 1994). a detailed tisting of the predictors and weightings incorporated into a credit scoring system is proprietary information, and likely specific to the individual industry and geo graphic region. however, a review of table 1 suggests that decisions are limited to quantifiable data available through the application data, internal auditing of account records, or credit bureau files. financial ratios representing debt and expenditure relationships have been an acknowledged part of the review process for consumer and mortgage lending. the mortgage industry relies heavily on the use of financial ratios in the approval process and has considered the use of subsequent ratio analyses of mortgage holders to identify, and avert, potential payment problems (harney, 1994). research also suggests that among a sample of financially distressed homeowners, most could not subsequently meet the qualification ratios established at the time of purchase (lytton & parrott, 1994; o’neill, lytton, & parrott, 1995). however, as described below, it has been only in the last decade that financial educators and other financial professionals have used financial ratio calculations when assessing financial well-being. b. financial ratios: educating to prevent insolvency financial ratios emerged in the business wodd in the early part of the 20th century but their first formal use occurred during the 1920s (horrigan, 1978). the first serious empirical tests of financial ratios were conducted during the 1930s. these studies were overlooked for almost two decades. during the post-world war ii period, financial ratios were either severely criticized or just plain ignored. however, by the early 196os, there was a renewed interest in financial ratios. a study using financial ratios in the 1930s and several later studies were concerned with business failure (altman, 1968, 1971). failing firms exhibited significantly different ratio measurements than businesses which were successful, and one of the later studies provided evidence of the use of financial ratios for prediction. although, historical accounts specifi cally cite the use of ratios in predicting bankruptcy, ratios measuring profitability, liquidity, and solvency have prevailed as significant indicators of progress over time and as standards for comparison of similar companies within an industry (byrne, 1992; brand& danos, & brasseaux, 1989; chen & shimerada, 198 1; ketz, doogar, & jensen, 1990; lawder, 1989; pressel, 1991). the primary function of ratios should be to act as indicators or redflags--to point to areas of acceptable or unacceptable results or conditions. the key to ratio analysis lies not in the values which are calculated but in the significance of the relationships being studied. household insolvency: debt repayment, delinquency, and bankruptcy 149 table 2 financial ratios liquidity liquid assets/monthly expenses liquid and financial assets/monthly expenses debt liquid assets/total debt liquid and financial assets/debt liquid assets/non-mortgage debt liquid assets/one-year debt payment liquid and financial assets/one-year debt payment inflation protection tangible and equity assets/fixed dollar assets derivatives of net worth tangible and equity less home/net worth non-mortgage debt/net worth total debt/net worth* (change to debt/asset) liquid assets/net worth* (change to rota1 assets) liquid and financial assets/net worth* (change to /total assets) tangible and equity assets/net worth* (change to /total assets) tangible assets/net worth* (change to rotal assets) income-generating assets/net worth* (change to notal assets) now: *with change suggested by prather. sources: griffith (1985); prather (1987). in a seminal work, griffith (1985) noted that the analysis of personal financial state ments seemed “undeveloped” and titled his proposal for the use of 16 financial ratios “a modest beginning” (p. 123). the 16 original ratios are shown in table 2. griffith stated that financial ratio analysis could be used by individuals and families: (1) as a measure of change in financial progress over time, (2) as an objective measure of analysis of family finances, and (3) as a tool for financial professionals to make recommendations to families (griffith, 1985). later, mason and griffith (1988) discussed the application of financial ratios to personal financial statements by professionals such as bankers, life insurance brokers, certified public accountants, attorneys, and financial planners. in each setting, the financial ratio could be used to help determine the financial well-being of the client in regard to one or more of the following areas: consumption, investment, and the use of credit. following an analysis of 22 personal finance and financial planning texts, mason and griffith (1988) noted the lack of a theoretical framework for using certain data when analyzing a client’s financial situation. they stated: despite the absence of sound theory, the authors believe it is still useful to develop ratios . . . empirical research is needed to test these ratios, and those ratios that are good predictors of financial problems and performance should be retained (p. 73). prather (1990) used data from the 1983 survey of consumer finances (scf) to examine the financial ratios suggested by griffith. following statistical analysis, prather suggested: (1) instituting household norms for each of the ratios, and (2) that the divisor of five ratios be changed from net worth to total assets. prather stated: “relating a part to the whole would 150 financial services review 4(2) 1995 table 3 financial ratios liquidity net consumption expenditures/disposable income liquid assets/net consumption expenditures total housing expenses/disposable income debt servke consumer debt repayments/disposable income annual consumer and mortgage debt repayments/annual disposable income gross annual debt repayments/gross annual income meeting financial goals total household assets/total household liabilities annual total savings/annual disposable income investment assets/net worth source: lytton, garman, & porter (1991). provide a ratio value which is more intuitively meaningful” (1990, p. 66). changes to the original ratios are annotated in table 2. although garman and forgue (199 1) recommended eight ratios that measured liquidity, debt burden, and progress toward meeting financial goals, they stated: since standards for these ratios do not exist, it is best to subjectively evaluate each ratio in light of the peculiarities of each individual and family circumstance, considering such factors as stage in the life cycle, marital status, income, and financial goals (garman & forgue, 1991, p. 92). not satisfied with the current research and literature on ratios, lytton, garman, and porter (1991) presented a list of ratios (shown in table 3) and applied them to a case study. when available, “widely accepted” guidelines for interpretation were suggested. lytton et al. noted that “recommendations for change should not be made on the basis of one ratio. instead, it is imperative that these nine ratios be calculated and the combined effects of the results considered in an interrelated manner” (p. 21). iwuagwu (1989) analyzed ratios from a different perspective--as predictors of per ceived household financial security. using data from the wisconsin basic needs survey which was collected in 1981 and 1982, iwuagwu found several ratios that were statistically significant predictors of perceived financial security (shown in table 4). five of the seven ratios that were used had been identified by prather as being the “most useful” of the original 16. prather (1990) and iwuagwu (1989) acknowledged that a smaller number of ratios was more useful and each reviewer recommended the use of similar ratios. lytton et al. (1991) assumed that families were knowledgeable about their amount of disposable income and suggested that income should be a reference point for many of the ratios. in a descriptive analysis of household financial status in the 1980s devaney (1993) compared the propor tions of households meeting financial ratio guidelines, as cited in two personal finance textbooks. almost 10% of households in 1986 were technically insolvent with an asset/h ability ratio less than 1 .o (i.e., debts were greater than assets) and 40% did not have access to a standard emergency fund measure (liquid assets equal to three months of disposable household insolvency: debt repayment, delinquency, and bankruptcy 151 table 4 financial ratios liquidity liquid assets/monthly expenditures* debt liquid assets/consumer debt* consumer debt/gross income liquid assetsr’otal debt liquid assets/short-term debt plus 1 year of other debt monthly debt payment/monthly gross income inflation protection inflationary assets/total assets* note: *predictor of household’s perceived financial security. income). also, several ratios indicated that the level of household debt compared to income increased between 1983 and 1986. in another study, again using survey of consumer finance data for 1983 and 1986 and negative or zero net worth as a dependent variable, devaney (1994) showed that comparing the value of a financial ratio to a cutoff or guideline was a statistically significant predictor of household insolvency three years later. when the outcome of two statistical procedures logistic regression and a classification tree-were compared, the most likely predictors of insolvency were the liquidity ratio and the assets-to-liability ratio, respectively. however, gross annual (non-mortgage) debt compared to disposable income was the second most likely predictor of insolvency for each of the two methods. financial ratios that were tested with a cutoff or guideline are shown with the appropriate value for the guideline in table 5. a study comparing ratio measurements in three time periods with the same families has provided further evidence of the usefulness of ratios (fanslow, 1994). six financial ratios were calculated using data collected during interviews with 84 household money managers in 1982,1986, and 1991. the ratios suggested some financial concerns for the households. fewer families had adequate liquid assets to meet three months of household expenses compared to in 1986. similarly, slightly fewer met the criterion of saving 5% of annual table 5 financial ratios with guideline solvency total assets/total liabilities >1* liquidity liquid assets/disposable income >.25 * debt service annual shelter costs~otal income < .28 consumer debt paymentsldisposable income < .15 gross annual debt payments/disposable income < .30 * note: *if guideline is not met, ratio predicts a propensity for insolvency. source: devaney (1994). 152 flnancial services review 4(2) 1995 table 6 financial ratios: iowa longitudinal study housing expenses/net income i .30-a0 financial assets/net worth 2.25 expenditures/net income il.00 savings/net income 2.05 liquid assets/expenditures > .25 consumer debt service 1.10 source: fanslow (1994). take-home pay. consumer debt load had increased for more of the families when compared to the two previous years (see table 6). the rather lengthy discussion of credit scoring and financial ratios for consumers has revealed a similar history of development for each of the predictive models. each model is quantifiable so that both consumers and the industry can have information readily available for use. the creditors have a responsibility to inform educators which ratios have been most useful as predictors in their modeling (cambridge seminar, 1988; mierzwinski, 1995) and then educators have a responsibility for informing the public on the use and interpretation of those ratios. v. i~~plicati~ns the management of the credit use/payment behavior relationship needs to be a win-win situation for all the parties involved. each part of the relationship needs to work to support the other. insolvency is a breakdown in the relationship. the credit community has to be judicious in the use of credit. consumers have to be judicious in the acceptance and use of available consumer credit. educators and other professionals who offer advice, counseling, and education to consumers can mediate in this relationship. consequently, this paper has implications for all these parties: for the credit industry: l consider attitudinal or other life-style variables for inclusion in credit scoring schemes; l realize that while profit is necessary to sustain the industry, the need exists to continually improve credit scoring systems but not at the expense of losing consum ers to the bankruptcy process; l use predictive ability to identify potential problems by monitoring accounts and offering trained professionals to assist consumers to adjust their spending patterns; l guard consumer privacy to avoid abuse of the information storage and retrieval possibilities associated with sophisticated scoring schemes; and l accept responsibility for providing credit information and education to the public. for consumers: l learn to monitor their financial situation and to notify creditors when problems arise, as opposed to the common practice of creditor avoidance; household insolvency: debt repayment, delinquency, and bankruptcy 153 l accept responsibility to prepare for the unexpected by learning and practicing strategies for efficient money management such as developing an emergency fund, notifying creditors, and curbing spending to allow for the purchase of insurance, savings for emergencies, and accomplishment of other goals; and l recognize the fundamental need for planned spending and managed cash flow. the calculation and tracking of net worth, income, and expenses-and selected financial ratios-provide needed information for monitoring the individual consequences of insolvency. for financial professionals/educators: l recognize that financial stability may be short-lived (e.g, decline in housing values, downsizing of corporations, etc.) to insure that families are prepared for contingen cies through plans for risk management and other savings; and l educate clients, regardless of income level, about appropriate use and abuse of credit, in particular, the use of financial ratios to identify potential problems. for educators/researchers: l develop theory to explain the body of literature and research on insolvency; l include random effects of unexpected events and the macroeconomic environment that is, the error term in the statistical equation; l work with creditors to learn what they are including in predictive models so that consumers can be better informed; l collect data and use more sophisticated statistical procedures to support comprehen sive study of factors contributing to insolvency; and l continue research and development on the use of financial ratios to guide consumers in assessing their current financial status, in making comparisons to past records and time periods, and in making decisions about the use of financial resources in the future. vi. conclusions in conclusion, the nuances of applying and interpreting the concept of insolvency in a household financial domain is complex and multidimensional. what is unclear is the relationship between the manifest and latent characteristics that contribute to the onset, and severity, of insolvency. the randomness of life events, which can represent either latent or manifest effects, can never be predicted or controlled for. according to one author, some creditor grantors consider the life of the credit obligation as a proxy for these unforeseen life events (wagner, reichert, & cho, 1983). another author acknowledges that the combination of macroeconomic factors and the increasingly precarious nature of the microeconomic environment of too many households have increased the complexity of bankruptcy and chargeoff prediction (“who will go bankrupt ‘7” 1992). but both creditors and consumers can reduce the impact of the “random error.” consumers can monitor spending and attempt to prepare for contingencies, while creditors can use their tools to monitor and help consumers avoid overextension. although issues of privacy and fairness are paramount, the 154 financial services review 4(2) 1995 benefits of remedial, not punitive, interventions which reduce the incidence of insolvency should not be discounted. learning more about insolvency is equally important to the credit community, consum ers, and professionals who serve consumers. continued theory development and research, both public and proprietary, offers avenues for protecting all parties, including the larger economy. but, judiciousness on the part of creditors and debtors is critical to the continued success of the relationship. insolvency, in the equity sense of failing to repay debts in a timely manner, represents a deterioration of the relationship. in the bankruptcy sense, insolvency represents the failure of this relationship. references altman, e.i. (1968). financial ratios, discriminant analysis, and the prediction of corporate bankruptcy. the journal of finance, 23(4), 589-609. altman, e.i. (1971). corporate bankruptcy in america, lexington, ma: heath lexington books. apilado, v.p., warner, d.c.. & dauten, j.j. (1974). evaluative techniques in consumer finance-ex perimental results and policy implications for financial instimtions. jour~ai of financial and quantitative analysis, (march), 275-283. avery, r.b., elliehausen, ge., & kennickell, a.b. (1987). changes in consumer installment debt: evidence from the 1983 and 1986 surveys of consumer finances. federal reserve bulletin, 73(10), 761-778. bankruptcy reform: truth or consequences. f 1995, spring). everybody’s money, 3#(4), 19-21. becker, n.d. (1992). becker cpa review course, lecture 7. ventura, ca: becker. bhandari, j.s., & weiss, l.a. (1993). the increasing bankruptcy filing rate: an historical analysis. the american bankruptcy law journal, 67(winter), l-15. boyes, w.j., hoffman, d.l., & low, s.a. (1989). an econometric analysis of the bank credit scoring problem. journal of econometrics, 40,3-14. brandt, l., danos, j.r., & brasseaux, j.h. (1989). financial statement analysis: benefits and pitfalls. the practical accourt~t, 22(s), 48-70. byrne, p.m. (1992). select financially healthy suppliers. transportation and distribution, 33( 1). 21-23. canner, g.b., & luckett, c.a. (1990). consumer debt repayment woes: insights from a household survey. journal of retail bunking, 12( 1, spring), 55-62. canner, g.b., (4% luckett, c.a. (1991). payment of household debts. federal reserve bulletin, 77(april), 218-229. capon, n. (1982). credit scoring systems: a critical analysis. journalofmarketing, 46(spring), 82-91. chandler, g.g., & parker, l.e. (1989). predictive value of credit bureau reports. journal of retail banking, ll(4, winter), 41-54. chen, k.h., & shimerada, t.a. (1981). an empirical analysis of useful financial ratios. financial management, fqspring), 5 l-60. chhikara, r.k. (1989). the state of the art in credit evaluation. american journal of agricultural economics, 71(5), 1138-l 144. collins, r.a., & green, r.d. (1982). statistical methods for bankruptcy forecasting. journal of economics and business, 34,349-354. courtless, j. (1993). trends in consumer credit. family economics review, 6(3), 8-17. cut-miff, j. (1995). nobody’s saving a penny to earn a penny anymore. the roanoke times and world news, (february 5), b7-b8. household insolvency: debt repayment, delinquency, and bankruptcy 155 devaney, s.a. (1993). changes in household financial ratios between 1983 and 1986: were american households improving their financial status? financial counseling and planning, 4, 3 l-46. devaney, s.a. (1994). the usefulness of financial ratios as predictors of household insolvency: two perspectives. financial counseling and planning, 5,5-24. devaney, s.a., & hanna, s. 1994. the effect of marital status, income, age and other variables on insolvency in the u.s.a. journal of consumer studies and home economics, l&293-303. durand, d. (1941). risk elements in consumer instullmentfinancing. study #8. new york: national bureau of economic research. easing borrowers off the road to bankruptcy. (1990, february). aba bunking journal, 82(february), 42-44. fanslow, a. (1994). financial ratios as a tool for assessing the financial health of iowa’s households. in m. bretz (ed.), the link. ames: iowa state university. federal reserve system, board of governors. (1991, september). balance sheets of the u.s. economy: 1945-1990. washington, dc: author. friedland, m. (1993). in any language, credit grantors say “yes!” to scoring. credit work& 82(1, september/october), 13-16. garman, e.t., & forgue, r.e. (1991). personaljnance. 3rd edition. boston, ma: houghton mifflin company. gothe, p. (1990). credit bureau point scoring sheds light on shades of gray. credit world, 78(5, june), 26-29. grablowsky, b.j., & talley, w.k. (1981). probit and discriminant functions for classifying credit applications: a comparison. journal of economics and business, 32(3), 254-261. griffith, r. (1985). personal financial statement analysis: a modest beginning. in v. langrehr (ed.), proceedings, third annual conference of the association of financial counseling and plun ning education (pp. 123-l 31). iowa state university. ames, ia: afcpe. harney, k. (1994). credit-tracking service can spot consumer loan payment difficulties. the wash ington post, (february 5), f3-f4. hira, t.k. (1992). the rehabilitative aspects of consumer bankruptcy procedures. proceedings of the association for financial counseling and planning education, io, 120-l 38. horrigan, j.o. (ed.). (1978). financial ratio analysis: an historical perspective. new york: amo press. hughes, j.w. (1991). understanding the squeezed consumer. american demographics, 13(july), 44-49. irvin, s. (1994). using lifetime value analysis for selecting new customers. credit world, 82(3, january/february), 37-39. iwuagwu, u.s. (1989). financial ratios and perceived household financial security. unpublished m.a. thesis, university of wisconsin, madison. jenson, h. (1992). using neural networks for credit scoring. management finance, 18(6), 15-26. jost, a. (1993). neural networks: a logical progression in credit and marketing decision systems. credit world, 81(4, march/april), 26-33. ketz, j.e., doogar, r.k., & jensen, d.e. (1990). a cross-industry analysis offinancial ratios. new york: quorum books. lawder, k.e. (1989). ratios 101: back to the basics of financial analysis. business credit, 91(6), 28-30. lease instead of buy? (1995). consumer reports, (april), 268-269. livingstone, sm., & lunt, p.k. (1992). predicting personal debt and debt repayment: psychological, social, and economic determinants. journal of economic psychology, 13, 11 l-134. long, m.s., & mcconnell, j.j. (1977). credit scoring system aids second mortgage makers. mortgage banker, 37,29-34. luckett, c.a. (1988). personal bankruptcies. federal reserve bulletin, 74(9), 591-603. 156 financial services review 4(2) 1995 luckett, ca., & august, j.d. (1985). the growth of consumer debt. federal reserve bulletin, 71(6), 3891102. lyons, f. (1993). using segmentation analysis: identifying customer segments for scoring model development. credit world, 81(5, may/june), 36-37. lytton, r.h., garman, e.t., & porter, n.m. (1991). how to use financial ratios when advising clients. financial counseling and planning, 2,3-23. lytton, r.h., & parrott, k.r. (1994). could virginians qualify for a mortgage today? in a.a. hertzler, m.j. sporakowski, & n.l. templeman (eds.), virginia’s families in a changing world blacksburg, va: vp1 & su college of human resources. makowski, p. (1985). credit scoring branches out. credit world, 74( 1, november-december), 30-37. the management of risk. (1988). banking world, 6( lo), 34-36. marshall, j. (1992). zap-the computer lops off another bad risk. united states banker, 102(8, august), 49-50. mason, j.w., & griffith, r. (1988). new ratios for analyzing and interpreting personal financial statements. journal of the institute of certified financial planners, (spring), 71-87. mccorkell, p.l. (1994). regulations of credit data. credit world, 82(3, january-february), 12-14. mierzwinski, e. (1995). creditors shouldn’t be averse to helping consumers. credit wortd, 83(3, january-february), 29-31. o’neill, b.m., lytton, r.h., & parrott, k.r. (1995). living on the edge: characteristics and practices of overextended homeowners. in s. hanna (ed.), family economics and resource management biennial. washington, dc: family economics and resource management division, american association for family and consumer sciences. overstreet, g.a., & kemp, r.s., jr. (1986). managerial control in credit scoring systems. journal of retail banking, 8( 1 & 2), 79-86. pellegrino, m.k. (1988). the evolution of scoring in the 1980s. credit world, 77(2, novemberide cember), 26-29. peterson, cm., & peterson, r.l. (1981). downpayments, borrower characteristics, and defaults. journal of retail banking, 3( 1, march), 1-6. prather, c.g. (1990). the ratio analysis technique applied to personal financial statements: develop ment of household norms. financial counseling and planning, 1.53-69. pressel, j.r. (1991). tools of the trade: financial statements. business credit, 93(2), 12-14. radding, a. (1992). credit scoring’s new frontier. bank management, 68(september), 57-62. robins, g. (1992). scoring: retailers are finding new uses for the technique. stores, (september), 45. robins, g. (1993). credit scoring: can retailers benefit from neural networks? stores, 75(april), 34-35. rosenberg, e., & gleit, a. (1994). quantitative methods in credit management: a survey. operations research, #2(4, july-august), 589-613. scheld, k.a. (1993). debt in the 1990s. chicago fed letter, the federal reserve bank of chicago, 69(may). shepard, l. (1984). accounting for the rise in consumer bankruptcy rates in the united states: a preliminary analysis of aggregate data (1945-198 1). the journal of consumer affairs, 18(2), 213-229. singletary, m. (1993). bankruptcy filings hit record high. washington post, (january 12). cl. sullivan, c., & fisher, r.m. (1988). consumer credit delinquency risk: characteristics of consumers who fall behind. journal of retail banking, lo(3, fall), 53-64. sullivan, t.a., warren, e., & westbrook, j.l. (1989). as we forgive our debtors: bankruptcy and consumer credit in america. new york: oxford university press. wagner, g.m., reichart, a.k., & cho, c. (1983). credit scoring, part ii: conceptual issues in credit scoring models. the credit world, 71(6, july-august), 22-28. who will go bankrupt? (1992). u.s. banker, 102(9), 47-5 1. pii: s1057-0810(00)00050-0 authors aiello, scott, “international index funds and the investment portfolio,” 8(1): 27–36 alexander jr., john c., “does education affect how well students forecast the market?,” 8(4): 253–260 anderson, randy i., “an analysis of affinity programs: the case of real estate brokerage participation,” 8(3): 183–197 badu, yaw a., “an empirical analysis of differences in black and white asset and liability combinations,” 8(3): 129–147 bajtelsmit, vickie l., “gender differences in defined contribution pension decisions,” 8(1): 1–10 bartholomae, suzanne, “student learning style and educational outcomes: evidence from a family financial management course,” 8(4): 235–251 bernasek, alexandra, “gender differences in defined contribution pension decisions,” 8(1): 1–10 britt, michael, “an integrative approach to using student investment clubs and student investment funds in the finance curriculum,” 8(4): 211–221 brooks, robert, “municipal bonds: a contingent claims perspective,” 8(2): 71–86 buetow, gerald w., “international mutual fund returns and federal reserve policy,” 8(3): 199–210 chieffe, natalie, “an integrated model for financial planning,” 8(4): 261–268 chieffe, natalie, “international index funds and the investment portfolio,” 8(1): 27–36 cooper, dan w., “an integrative approach to using student investment clubs and student investment funds in the finance curriculum,” 8(4): 211–221 crabb, ronald r., “cash flow: a quick and easy way to learn personal finance,” 8(4): 269–282 daniels, kenneth n., “an empirical analysis of differences in black and white asset and liability combinations,” 8(3): 129–147 elder, harold w., “does retirement planning affect the level of retirement satisfaction?,” 8(2): 117– 127 eyssell, thomas h., “learning by doing: offering a university practicum in personal financial planning,” 8(4): 293–303 filbeck, greg, “family friendly firms: does it pay to care?,” 8(1): 47–60 finch, j. howard, “effective teaching and use of the constant growth dividend model,” 8(4): 283–291 fogarty, ellie, “undergraduate research: the senior thesis in finance,” 8(4): 223–234 fox, jonathan j., “racial differences in investor decision making,” 8(3): 149–162 fox, jonathan, “student learning style and educational outcomes: evidence from a family financial management course,” 8(4): 235–251 gold, steven c., “computerized stock screening rules for portfolio selection,” 8(2): 61–70 grable, john, “financial risk tolerance revisited: the development of a risk adjusted instrument,” 8(3): 163–181 grinder, brian, “an integrative approach to using student investment clubs and student investment funds in the finance curriculum,” 8(4): 211–221 gutter, michael s., “racial differences in investor decision making,” 8(3): 149–162 hanna, michael, “a nineties perspective on international diversification,” 8(1): 37–46 329 financial services review the journal of individual financial management index volume 8, 1999 ho, kwok, “international equity diversification and shortfall risk,” 8(1): 11–26 jensen, gerald r., “international mutual fund returns and federal reserve policy,” 8(3): 199–210 jianakoplos, nancy a., “gender differences in defined contribution pension decisions,” 8(1): 1–10 johnson, robert r., “international mutual fund returns and federal reserve policy,” 8(3): 199–210 lebowitz, paul, “computerized stock screening rules for portfolio selection,” 8(2): 61–70 lewis, danielle, “an analysis of affinity programs: the case of real estate brokerage participation,” 8(3): 183–197 luft, carl f., “hedging individual mortgage risk,” 8(2): 101–116 lytton, ruth h., “financial risk tolerance revisited: the development of a risk adjusted instrument,” 8(3): 163–181 mayo, herbert, “undergraduate research: the senior thesis in finance,” 8(4): 223–234 mccormack, joseph p., “a nineties perspective on international diversification,” 8(1): 37–46 mcdonald, john a., “investor partitioning of the components of value in corporate earnings announcements,” 8(2): 87–100 mcelreath, robert b., “does education affect how well students forecast the market?,” 8(4): 253–260 michelson, stuart, “applications of www technology in teaching finance,” 8(4): 319–328 milevsky, moshe arye, “international equity diversification and shortfall risk,” 8(1): 11–26 montalto, catherine p., “racial differences in investor decision making,” 8(3): 149–162 payne, thomas h., “effective teaching and use of the constant growth dividend model,” 8(4): 283–291 perdue, grady, “a nineties perspective on international diversification,” 8(1): 37–46 preece, dianna c., “family friendly firms: does it pay to care?,” 8(1): 47–60 rakes, ganas k., “an integrated model for financial planning,” 8(4): 261–268 robinson, chris, “international equity diversification and shortfall risk,” 8(1): 11–26 rudolph, patricia m., “does retirement planning affect the level of retirement satisfaction?,” 8(2): 117–127 salandro, daniel p., “an empirical analysis of differences in black and white asset and liability combinations,” 8(3): 129–147 smith, stanley d., “applications of www technology in teaching finance,” 8(4): 319–328 smith, david m., “investor partitioning of the components of value in corporate earnings announcements,” 8(2): 87–100 woerheide, walt, “the internet in the personal finance course,” 8(4): 305–317 zivney, terry l., “hedging individual mortgage risk,” 8(2): 101–116 zumpano, leonard v., “an analysis of affinity programs: the case of real estate brokerage participation,” 8(3): 183–197 titles “family friendly firms: does it pay to care?,” dianna c. preece, greg filbeck, 8(1): 47–60 “gender differences in defined contribution pension decisions,” vickie l. bajtelsmit, alexandra bernasek, nancy a. jianakoplos, 8(1): 1–10 “international equity diversification and shortfall risk,” kwok ho, moshe arye milevsky, chris robinson, 8(1): 11–26 “international index funds and the investment portfolio,” scott aiello, natalie chieffe, 8(1): 27–36 “a nineties perspective on international diversification,” michaelhanna, joseph p. mccormack, grady perdue, 8(1): 37–46 “computerized stock screening rules for portfolio selection,” steven c. gold, paul lebowitz, 8(2): 61–70 “municipal bonds: a contingent claims perspective,” robert brooks, 8(2): 71–86 “investor partitioning of the components of value in 330 index / financial services review 8 (1999) 329–331 corporate earnings announcements,” john a. mcdonald, david m. smith, 8(2): 87–100 “hedging individual mortgage risk,” terry l. zivney, carl f. luft, 8(2): 101–116 “does retirement planning affect the level of retirement satisfaction?,” harold w. elder, patricia m. rudolph, 8(2): 117–127 “an empirical analysis of differences in black and white asset and liability combinations,” yaw a. badu, kenneth n. daniels, daniel p. salandro, 8(3): 129–147 “racial differences in investor decision making,” michael s. gutter, jonathan j. fox, catherine p. montalto, 8(3): 149–162 “financial risk tolerance revisited: the development of a risk adjusted instrument,” john grable, ruth h. lytton, 8(3): 163–181 “an analysis of affinity programs: the case of real estate brokerage participation,” danielle lewis, randy i. anderson, leonard v. zumpano, 8(3): 183–197 “international mutual fund returns and federal reserve policy,” robert r. johnson, gerald w. buetow, gerald r. jensen, 8(3): 199–210 “an integrative approach to using student investment clubs and student investment funds in the finance curriculum,” brian grinder, dan w. cooper, michael britt, 8(4): 211–221 “undergraduate research: the senior thesis in finance,” ellie fogarty, herbert mayo, 8(4): 223– 234 “student learning style and educational outcomes: evidence from a family financial management course,” jonathatn fox, suzanne bartholomae, 8(4): 235–251 “does education affect how well students forecast the market?,” john c. alexander jr., robert b. mcelreath, 8(4): 253–260 “an integrated model for financial planning,” natalie chieffe, ganas k. rakes, 8(4): 261–268 “cash flow: a quick and easy way to learn personal finance,” ronald r. crabb, 8(4): 269–282 “effective teaching and use of the constant growth dividend model,” thomas h. payne, j. howard finch, 8(4): 283–291 “learning by doing: offering a university practicum in personal financial planning,” thomas h. eyssell, 8(4): 293–303 “the internet in the personal finance course,” walt woerheide, 8(4): 305–317 “applications of www technology in teaching finance,” stuart michelson, stanley d. smith, 8(4): 319–328 331index / financial services review 8 (1999) 329–331 pii: s1057-0810(99)00033-5 municipal bonds: a contingent claims perspective robert brooks* southtrust professor of financial management, department of economics, finance and legal studies, the university of alabama, 200 alston hall, box 870224, tuscaloosa, al 35487, usa abstract the purpose of this paper is to provide an overview of the municipal bond market with an emphasis on the numerous embedded contingent claims. embedded contingent claims include the standard call features, sinking funds, the advance refunding option, the synthetic advance refunding option, the credit risk option (default risk), marketability, and the numerous tax-related events. municipal bond investors must carefully assess the relative value of these contingent claims before investing in municipal bonds. also, due to unique risk premiums within the municipal bond market, it is important to carefully structure the municipal bond holdings, paying particular attention to duration, within the context of an overall financial plan. there appears to be a benefit to lengthening the duration of the municipal bond portion of the portfolio. © 1999 elsevier science inc. all rights reserved. jel classification:g28; e43; e62 keywords:municipal bonds; embedded derivatives; tax risk 1. introduction of the bond markets tracked by the bond markets association ($12.9 trillion in all), the relative size of the u.s. municipal bond market at the end of 1998 was 11.30%. u.s. treasuries comprised 26.0%, government agency mortgage-backed securities 15.6%, corporate bonds 18.5%, u.s. federal government agencies 8.3%, money market securities * corresponding author. tel.:11-205-348-8987; fax:11-205-348-0590. e-mail address:rbrooks@cba.ua.edu (r. brooks) financial services review 8 (1999) 71–85 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00033-5 15.3%, and asset-backed securities 4.9%. the municipal bond market had $1.46 trillion outstanding at the end of 1998, and hence it is relatively large (www.bondmarkets.com). according to the bond market association, municipal bonds were held in the following proportions at the end of 1998: 30.5% households, 16.6% mutual funds, 13.2% money market funds, 4.3% closed-end funds, 7.2% trusts, 7.2% commercial banks, 13.4% insurance companies, and 7.6% other. it is interesting to note that based on irs estimates, 30% of municipal interest reported was from taxpayers with less than $50, 000 of adjusted gross income (www.bondmarkets.com and feenberg and poterba, 1992). 1.1. municipal bond market maturity preferences the propensity for tax-exempt issuers to prefer long-term bonds seems different from other debt markets. for example, fig. 1 illustrates the percentage of short-term bonds issued in the municipal market. clearly there is a strong preference for longer-term fixed rate bonds. for example, the average maturity of u.s. treasury securities is estimated to be 5 years and 4 months as of september 30, 1997 (www.treas.gov/treasury/financial/domfin/debtprof.htm). hence there seems to be a greater preference for longer-term debt by municipal governments than by the federal government. kidwell and koch (1983) noted that municipal “. . . borrowers can use long-term debt for current operations only by constitutional amendment or public referendum . . . ” making it difficult for municipalities to substitute between long and short-term debt. municipal bond investors also have preferred maturities and these preferences are reflected in observed yields. these preferences, as revealed in observed market yields, are fig. 1. percentage of municipal debt issued as short-term 72 r. brooks / financial services review 8 (1999) 71–85 different between the taxable and tax-exempt bond markets. for example, the bond market association (1998) reported that for the first quarter of 1998, of the $75.9 billion of newly issued municipal debt, 94% have a long maturity. however, it is known that municipal bond investors have a strong preference for shorter maturities. for example, mcentee (1998) reported “. . . a gap exists between supply and demand in tax-free money market, that hold about $175 billion in assets. short-term notes offer about $45 billion of eligible investments yearly and vrdns (variable rate demand notes) trigger about $30 billion of debt annually, leaving sizable room for synthetic derivative products to grow.” in other words, derivative securities are creating the additional supply of synthetic short-term municipal debt demanded by the market. the supply of municipal debt is thought to be relatively interest-inelastic and the equilibrium marginal tax rate will be a function primarily of investor demand, which may vary by maturity. therefore changes in the quantity of investor funds pursuing municipal bonds will influence the relative pricing of municipal bonds. rosenbloom (1976) documented when insurance company demand for municipal bonds declined, municipal yields rose relative to the taxable market. rosenbloom (1976) observed that a rise in the ratio of tax-exempt to taxable yields “. . . has been further aggravated by a decline in the demand for municipals by fire and casualty insurance companies in the first quarter of 1975, because of a low level of industry profits.” prior to the 1980s commercial banks were dominant holders of short-term debt and hence the ratio of tax-exempt to taxable yield for the shorter maturities was related to the corporate tax rate. this relationship does not hold for longer maturities and some conclude other investors must demand those municipal bonds. poterba (1986) and metcalf (1992) as well as others found empirical support for the yield ratio being a function of the corporate tax rate. empirical evidence supporting the strong preference of tax-exempt issuers for longermaturity bonds is found in the yield to maturities of municipal bonds when compared to constant maturity treasuries of similar maturities. recall that the tax-exempt rate is technically equal to the taxable rate times one minus the tax rate. using the yields on taxable and tax-exempt bonds, we can compute the implied tax rate with the well-known relationship yte 5 (1 2 t) yt where yt is the yield on the taxable rate, yte is the yield on the tax-exempt rate, and t is the implied tax rate. in this case, we solve for the tax rate. the implied marginal tax rate (imtrt) from these two interest rates is: imtrt 5 1.02 yte,t yt,t (1) where the subscript t denotes a point in calendar time. 1.2. the role of taxes if the tax-exempt yield curve is steeper than the taxable yield curve, then the implied tax rate will be lower for longer maturities. fig. 2 provides an illustration of the historical pattern of implied marginal tax rates by bond maturity. the taxable rate is based on u.s. treasuries and the tax-exempt rate is based on aaa credit, general obligation municipal bonds. notice 73r. brooks / financial services review 8 (1999) 71–85 that the average implied tax rate was almost always lower for longer maturity bonds. hence the tax-exempt yield curve is almost always steeper. fig. 2 suggests that longer maturity municipal bonds offer relatively higher yields. therefore in a bond portfolio, perhaps longer maturity bonds should be tax exempt and shorter maturity bonds should be taxable. this relationship has been studied for decades, for example, yawitz (1978) studied this relationship in detail over 20 years ago. green (1993) presented a model for the relationship between the taxable and tax-exempt markets based on arbitrage activities within the taxable bond market. he asserts that taxable investors can synthetically create zero coupon bonds that result in no tax liability until expiration, at which time there is a taxable capital gain. this arbitrage activity will influence the appropriate tax-exempt yield and hence the tax-exempt to taxable yields ratio. empirical evidence in green (1993) and chalmers (1998) found that green’s model is not rejected. based on continuous compounding and $1 par, the present value based on the tax-exempt rate (pte) can be expressed as: pte 5 $1e2ytet (2) where t denotes time to maturity in years and yte is the tax-exempt yield to maturity. the present value based on the taxable rate can be expressed two ways as: pt 5 $1e2ytt 5 @$1 2 t~$1 2 pt!#e2ytet (3) where the second expression is the after-tax cash flow discounted at the tax-exempt rate. the implied marginal tax rate is given in equation (1), and substituting eq. (1) into eq. (3) and solving for the implied marginal tax rate, we have: fig. 2. implied marginal tax rate by maturity 74 r. brooks / financial services review 8 (1999) 71–85 imtrt 5 2 ln@$1 2 t~$1 2 e2ytt!# ytt (4) assuming a constant tax rate (t), the imtrt declines as the maturity is increased. the ability to defer taxes has the effect of lowering the implied marginal tax rate. table 1 illustrates the influence of tax deferral on the implied marginal tax rate given in eq. (4). we see that tax deferral is equivalent to lowering the implied marginal tax rate. for example, at a 10% taxable yield to maturity and a 40% tax bracket, deferring taxes for 10 years lowers the implied tax bracket to 29.15%. it remains somewhat unclear how effectively high tax entities can defer taxes on investments. however, the ability to defer taxes influences the implied tax rate as well as the relative attractiveness of tax-exempt bonds. 1.3. short maturity municipal market we turn now to examine the short maturity municipal market. most variable rate municipal issues are aaa insured making these issues relatively homogeneous. the reference floating interest rate is called the bma rate or technically the bma municipal swap index or the bond market association municipal swap index. (on bloomberg, for example, type munipsa). the bma rate (or bma index) is an index of rates based on high grade, 7 day, tax exempt, variable rate demand obligations (vrdos) that is reset each wednesday. the index is constructed by municipal market data, a thomson financial services company. for a vrdo issue to be considered for the index, it must be weekly reset effective on wednesday, not subjected to alternative minimum tax, have at least $10 million outstanding, have the highest short-term rating (vmig1 by moody’s or a-11 by s & p500), pay interest on a monthly basis, and be calculated on an actual/actual day count basis. issues included in the index are screened for outliers and participating remarketing agents cannot represent more than 15% of the index. the bma index includes roughly 250 issues in any given week (www.bondmarkets.com). fig. 3 provides weekly bma index values and 3-month libor table 1 implied marginal tax rate based on tax deferral and continuous compoundinga maturity yt 5 5% yt 5 10% yt 5 15% yt 5 20% yt 5 25% 0 40.00% 40.00% 40.00% 40.00% 40.00% 1 39.40% 38.81% 38.22% 37.64% 37.06% 2 38.81% 37.64% 36.48% 35.35% 34.25% 3 38.22% 36.48% 34.80% 33.17% 31.61% 4 37.64% 35.35% 33.17% 31.10% 29.15% 5 37.06% 34.25% 31.61% 29.15% 26.88% 10 34.25% 29.15% 24.81% 21.22% 18.30% 15 31.61% 24.81% 19.69% 15.94% 13.21% 20 29.15% 21.22% 15.94% 12.47% 10.13% 30 24.81% 15.94% 11.19% 8.49% 6.81% infinite 0.00% 0.00% 0.00% 0.00% 0.00% a we assume a 40% marginal tax rate and markets are in equilibrium. 75r. brooks / financial services review 8 (1999) 71–85 from july 5, 1989 through december 30, 1998. the bma index is quite volatile and the seasonal trends are clearly evident, especially at year-end. fig. 3 also includes the 3-month libor rate that is a popular rate upon which the taxable interest rate swaps are based and is a very close proxy for negotiated bank certificate of deposit rates. three month libor exhibits much less short-term variability and is less seasonal. although other proxies for the taxable rate are possible, such as the u.s. treasury rates, we use libor due to its highly liquid swap market and because it represents banks’ marginal cost of funds. fig. 4 presents the implied marginal tax rate (imtr) based on the 90-day average of bma index and the initial 3-month libor. we use the 90-day average for two reasons. first, we need to compare the 90-day libor to a comparable term. second, bma swaps are settled based on 90-day average bma. although there is variability in the imtr, it does seem related to current corporate and individual tax rates. the ratio has also been more stable in the recent past. 2. enumeration of embedded derivatives in this section we identify and discuss several embedded derivatives in most municipal bonds. embedded contingent claims include the standard call features, sinking funds, the advance refunding option, the synthetic advance refunding option, the credit risk option (default risk), the numerous tax-related events, and marketability. municipal bond investors should carefully assess the relative value of these contingent claims before investing in municipal bonds. fig. 3. the bma and 3 month libor july 5, 1989 through december 30, 1998 76 r. brooks / financial services review 8 (1999) 71–85 2.1. the call option most municipal bonds are callable after 10 years. the value of the embedded call makes the yields higher because the bond investor is short the option. however, municipal bonds are different in many ways. municipal bonds have lower coupons, hence higher durations, implying more interest rate risk. van horne (1987) observes that higher interest rate risk must be compensated with a higher yield that would explain the higher tax-exempt to taxable yield ratio. mallman (1981) and stock (1985) also address this issue. leibowitz (1981) and sorensen (1983 suggested that different interest rate risk levels account, in part, for the higher tax-exempt to taxable yield ratio. fortune (1991) observed that “. . . the exposure of municipal bonds to capital gains taxes when market discounts emerge will place municipal bonds at a disadvantage in periods when bond prices are more volatile.” chalmers (1998) presented strong evidence that embedded call options do not completely explain the higher tax-exempt to taxable yield ratio for longer maturities. thus, there seems to be significant other factors to consider with municipal bonds. 2.2. sinking funds sinking fund provisions within municipal debt require the retirement of a portion of the debt usually on an annual basis. sinking funds typically contain a whole set of embedded options that are inter-related. kalotay and williams (1992) provide a detailed review of the sinking fund options. for example, some bond indentures include a sinking fund acceleration option giving the fig. 4. tax rates and implied marginal tax rates 77r. brooks / financial services review 8 (1999) 71–85 issuer the option to accelerate the sinking fund retirement. this option is equivalent to a series of partial european-style options that are usually noncumulative. sinking funds typically include a delivery option that gives the issuer the opportunity to satisfy the requirements of the sinking fund provisions with open market purchases of the bonds. this option is more valuable when rates rise because the bonds can be retired at an open market price below par. there may also be a designation option that allows issuer to apply bonds the firm has purchased in the open market to future sinking fund requirements. if the firm optimally exercises this option, it will tend to reduce the value of the remaining outstanding bonds. there is a very subtle ownership option related to sinking funds. if the bonds are held in large blocks, this limits the ability of the firm to conduct outright purchases without paying a premium. the ownership option reduces the value of the delivery option. when purchasing municipal bonds with sinking funds, it is important to understand the terms of the sinking fund. 2.3. the advance refunding option because municipal bonds are tax-exempt, one would expect the coupon rate on the municipal bonds to be lower than the yields on u.s. treasury securities, even when overall interest rates have fallen since the municipal bonds were issued. due to tax laws prohibiting arbitrage transactions, municipal issuers are permitted only one (typically) advance refunding opportunity. municipal bonds are advance refunded by issuing more tax-exempt bonds and using the proceeds to purchase u.s. treasury securities to defease the original issue. hence, after the advance refunding, the original municipal bonds are still outstanding but they are backed by u.s. treasury securities and therefore the municipality does not have to make the subsequent coupon payments from their operating cash flow. therefore, the original bonds receive essentially a credit enhancement to aaa. the benefit of an advance refunding is that it permits the municipality to monetize the embedded call option. typically the municipality receives a cash payment from the advance refunding process and it is somewhat erroneously called “present value savings.” it is erroneous because an asset was sold, specifically the embedded call options. municipal finance directors are under increasing pressure to manage the municipality’s interest rate risk efficiently. often they are solicited to conduct a refinancing of their existing bonds. it is generally accepted that the marginal benefits derived from a municipal bond refunding should exceed the marginal costs. the marginal cost is the extinguishing of the embedded options and the marginal benefit is the present value savings. a callable bond can be represented as a non-callable bond less the value of the call option. because the issuer has the right to call the bonds, the bond investor is short this call option. mathematically, the relationship can be expressed as pt c 5 pt nc 2 cot (5) wherept c denotes the callable bond,pt nc the non-callable bond with identical maturity and coupon, andcot denotes the embedded call option premium. as illustrated for corporate bonds in emery and lewellen (1990) and jordan and jorgensen (1998), one can illustrate the 78 r. brooks / financial services review 8 (1999) 71–85 refunding decision using hypothetical balance sheets. table 2 illustrates in over-simplified terms, a municipality with debt and “equity.” in all the subsequent tables, the dollar amounts are given in current market value. because interest rates have fallen, the bonds are trading well above par. however, the call feature prevents the price from rising too far. if we split the callable bond into its component pieces we can represent the balance sheet as given in table 3. the embedded call option on the bond is worth say $23 million and the 25 year fixed rate bond is worth say $333 million. the explicit recognition of the option as an asset helps one understand the appropriate method for assessing the refunding decision. suppose an investment bank suggests doing an advance refunding on the fixed rate debt. also assume the advance refunding produced proceeds of $15 million. to simplify the analysis, we assume the new bonds will be 25 year fixed rate bonds so they will have the same market value as the old bonds without the embedded option. the balance sheet is illustrated in table 4. clearly the firm has increased its cash position but decreased its equity. advance refundings are popular with municipalities and bondholders do not complain because their original bonds get a credit enhancement (recall the original debt’s rating goes to aaa because they are backed by u.s. treasury bonds). advance refunding also generally resolves the uncertainty about when the municipal bonds will be called. the advance refunded municipal bonds will be called on the first call date. table 2 balance sheet of a municipality (in market value and in millions) assets liabilities and equity cash $400 floating debt $100 other assets $600 fixed debt $310a other liability $90 equity $500 total assets $1,000 total liability and equity $1,000 a suppose originally issued 30-year tax-exempt bonds, callable in 10 years at par, $300 million par value, and carry a 5.3% coupon. five years have elapsed and interest rates have fallen. table 3 typical balance sheet of a municipality (in market value) explicitly accounting for embedded option assets liabilities and equity cash $400 floating debt $100 option $23 fixed debt $333 other assets $600 other liability $90 equity $500 total assets $1,023 total liability and equity $1,023 79r. brooks / financial services review 8 (1999) 71–85 2.4. the synthetic advance refunding option the refunding bonds, the bonds used to advance refund the original bonds, could be synthetically advance refunded even though they cannot be advance refunded. if a forward refunding transaction is conducted, then the embedded call option will be exercised on the first call date. this actually should increase the value of the refunding bonds because the time value of the call option is eliminated. alternatively the embedded call option could be sold directly (at least from an economic viewpoint). suppose the decision to call the bonds is essentially sold to an investment banking firm. in this case, the investment-banking firm will likely more efficiently make the call decision lowering the value of the refunding bonds. for example, the municipality might decide not to call the bonds if they are trading at 101 percent of par whereas the investmentbanking firm might efficiently call the bonds. 2.5. credit risk (default option) municipal bonds have higher default risk than u.s. treasury securities, and hence the tax-exempt to taxable yield ratio is higher. table 5 reports the actual municipal default experience since 1940. according to the bond market association, “the default rate for municipal securities is extraordinarily low. of the 403, 152 issues sold since 1940, only 0.5% or 2, 020 issues have had either a technical or actual default. this figure also includes issues table 4 typical balance sheet of a municipality (in market value) advance refunding at 5% savings (issue noncallable debt) assets liabilities and equity cash $415 floating debt $100 other assets $600 fixed debt $333 other liability $90 equity $492 total assets $1,015 total liability and equity $1,015 table 5 municipal default experiencea period number of defaulted issues total number of long-term issues default rate 1940–49 79 40,907 0.20% 1950–59 112 74,592 0.2 1960–69 294 79,941 0.4 1970–79 202 77,620 0.3 1980–94 1,333 130,092 1 total 2,020 403,152 0.5 a source: the bond market association, www.bondmarkets.com. 80 r. brooks / financial services review 8 (1999) 71–85 that defaulted but where investors ultimately received either full or partial payment of principal and interest. it is worth noting that the default rate of corporate bonds is substantially higher, in excess of 2%” (www.bondmarkets.com). however, there are reasons to be concerned about default risk for municipal bonds. many times it is difficult to understand political motivations. also municipal assets are difficult to seize when there is default and financial statements of municipalities are generally lacking in quality. by examining refunded municipal bonds, chalmers (1998) concluded the high ratio of tax-exempt to taxable yields for longer maturities is not based on municipal default risk. however, numerous prior studies came to the opposite conclusion, for example, trzcinka (1982). however, other prior studies, such as buser and hess (1986), failed to find evidence of default risk playing a significant role. 2.6. marketability it is generally accepted that marketability has market value and hence more liquid bonds trade at lower yields. there are clearly differences in liquidity between municipal bonds and u.s. treasury bonds. for example, the bond market association (1997) reported that the average daily trading volume for municipal bonds was $8.6 billion with u.s. treasuries $242.2 billion average daily volume. thus the ability to alter the general marketability of the municipal bonds will influence its value. wilson and stewart (1990) provide evidence that marketability could also be improved with better reporting practices. 3. tax-related events holding municipal bonds subjects investors to unique risks related to unanticipated changes in tax policies. in this section, we survey these risks and their influence on the relative attractiveness of owning municipal bonds. essentially, governments hold an option on tax rates and tax policies. tax policy risk should be fully reflected in the tax-exempt bonds because tax policy will directly influence the after-tax value of the future cash flows. lowering the marginal tax rates will result in an increase in the yields on tax exempt bonds, an increase in the ratio of tax-exempt to taxable yields, and the current market value of longer maturity municipal bonds will fall. fig. 5 illustrates the highest marginal tax rates for both individuals and corporations during most of this century. the main insight from this figure is the fact that tax policy is quite volatile. there are numerous issues encompassed in tax policy risk. we enumerate a few here. first, municipal bonds are a hedge against federal tax policy risk. an increase in federal income tax rate implies lower municipal bond yields. these lower yields imply capital gains on municipal bond investments at a time when an investor is experiencing losses from higher taxes. second, congress may entirely eliminate tax-exemption status of all municipal bonds. the u.s. supreme court in the case of south carolina versus baker, established that the u.s. government has the right to rescind the tax-exemption feature of municipal bonds, and more 81r. brooks / financial services review 8 (1999) 71–85 broadly, the u. s. government has the constitutional right to tax interest on municipal bonds. poterba (1989) and chalmers (1998) provide more details. evidence that the tax-exemption status of all municipal bonds could be eliminated is seen in the following historical account by ott and meltzer (1963). interest income from securities issued by state and local governments was specifically exempted from the federal income tax under section 103(a)(1) of the first income tax act passed pursuant to the sixteenth amendment. virtually every secretary of the treasury since its passage has favored removing the exemption feature . . . congress has some six times defeated proposals to remove the exemption . . . (p. 1) also, according to ott and meltzer (1963), there were 114 resolutions from 1920 to 1943 to repeal the tax exemption of municipal interest. third, congress may change various features of the federal tax code, such as lowering the marginal tax bracket, adjusting the alternative minimum tax (amt), broadening the scope of individual retirement accounts (iras) and so forth. for example, arak and guentner (1983) noted, of the possible explanations for these phenomena (ratio of tax-exempt to taxable bond yields being too high), we find that neither liquidity considerations, nor the behavior of banks and property and casualty companies, nor the recession can explain the entire rise. a key additional factor appears to be the economic recovery act of 1981 (erta), implying that the reduced benefits of tax exemption to issuers will persist. . . . “(erta) reduced marginal tax rates and widened the availability of income tax deferral through iras, keoghs, and other plans.” (guentner (1983), p. 145). fama (1977) examined the role of municipal bonds in commercial bank portfolios. fig. 5. highest marginal tax rates 82 r. brooks / financial services review 8 (1999) 71–85 chalmers (1998) regressed the implied tax rate on maturity and observes “[i]t is also notable that the slope has become less negative in the period following the tax reform act of 1986. furthermore, during economic downturns and times of tax law uncertainty it appears that implied tax rates are lower for all maturities and the differences across maturities are less pronounced.” (p. 298–99). metcalf (1992) also examined tax policy influence on municipal bond supply. recent news related to this aspect include the house of representatives vote on july 17, 1998 to sunset the current tax code effective december 31, 2002. clearly, the future of tax policy is uncertain. fourth, due to unforeseen issues to the municipal bond investors, the particular municipal bonds are declared fully taxable by the irs due to fraud or error in compliance. although the municipal bond investor bears the risk of an unfavorable irs ruling, most municipal issuers will choose to settle with the irs to maintain their integrity within the municipal bond investor community. fifth, changes in laws about who can own municipal bonds. for example, there are restrictions on companies with regard to the quantity of municipal bonds they can own. for example, chalmers (1998) observed, “. . . the tax code continues to allow all nonfinancial u.s. corporations to hold up to 2% of their assets in tax-exempt bonds and simultaneously deduct the interest on attributed debt from their taxable income.” (p. 284–5). sixth, investors have some flexibility as to when tax liabilities are incurred. this flexibility could be viewed as a tax timing option. seventh, u.s. treasury bonds are exempt from state and local taxes. most studies of the taxable versus tax-exempt market use only u.s. treasury bonds. this state tax exemption may slightly bias the reported results that compare municipal bonds with u.s. treasury bonds. eighth, tax treatments for taxable and tax-exempt bonds are not symmetric with regard to capital gains and losses. municipal bonds’ coupon payments and original issue discount (oid) bonds are not subject to federal taxes, but capital gains are subject to federal taxes. premium municipal bonds must amortize the bond’s basis, but the amortized premium cannot be taken as an expense for tax purposes. the amortization is for the purpose of computing capital gains. premium taxable bonds may be amortized and taken as an annual tax-deductible loss over the life of the bond. on the subject of asymmetric treatment of capital gains, sorensen [1983] stated that “. . . the tax treatment of investor income is a major contributory factor to changes in the municipal yield curve shape.” (p. 61). leibowitz (1981) developed a comprehensive interest rate model and concludes that tax-exempt bonds are more volatile in general and discount tax-exempt bonds are more volatile than premium tax-exempt bonds due to these asymmetric tax treatments. also we note that almost all states exempt their own bonds from state taxes. johnson and johnson (1996) review in detail various tax issues related to municipal bonds. ninth, changes in expected future taxes influence the tax-exempt to taxable yield ratio. poterba (1989) observed that tax news explains some changes in the yield ratio and he estimated forward tax rates. recently, koch and stock (1997) observed that the yield ratio is a function of individual tax rates but not corporate tax rates (they asserted that corporate taxes actually have the opposite effect). fortune (1991) also examined this yield ratio and 83r. brooks / financial services review 8 (1999) 71–85 documented an inverse relationship with forward tax rates and concluded that “. . . anticipated future income tax rates will have a strong impact on municipal bond yields” (p. 35). tenth, state and local tax issues influence the observed relationship between tax-exempt and taxable markets. chalmers (1998) reported that, “ . . . anecdotal evidence in green (1993) implies that while state tax differences induce small parallel shifts in municipal yield curves, state taxes do not affect the slope of the municipal term structure” (p. 291). severn and stewart (1992) focused on the role of state taxes in the u.s. treasury market and concluded that the break-even state tax rate was low. also, in many states, a municipal bond issued in another state is fully taxable with regard to the state income tax. this influences the relative attractiveness of various states’ municipal debt. singh and dresnack (1998) provide a comparative analysis of the performance of tax-exempt state funds and other tax-exempt funds. 4. summary the purpose of this paper was to review the major issues related to investing in municipal bonds. also, the numerous embedded contingent claims related to municipal bonds were discussed. municipal bonds have the same types of embedded contingent claims as corporate bonds as well as many others. the tax-exempt term structure of interest rates is much steeper than the taxable term structure suggesting various duration-related strategies when managing a personal portfolio that contains both taxable and tax-exempt bonds. there has been a historical benefit to lengthening the duration of a municipal bond portfolio, due perhaps to a tax risk premium. acknowledgment the author gratefully acknowledges the assistance of bob jacobus, kent boyd, and david yan. tax rate data was provided by william d. samson. references arak, m., & guentner, k. (1983). the market for tax-exempt issues: why are the yields so high? national tax journal, 36(2), 145–161. buser, s. a., & hess, p. j. (1986). empirical determinants of the relative yields on taxable and tax-exempt securities.journal of financial economics, 17(2), 335–355. chalmers, j. m. r. (1998). default risk cannot explain the muni puzzle: evidence from municipal bonds that are secured by u.s. treasury obligations.the review of financial studies, 11(2), 281–308. emery, d. r., & lewellen, w. g. (1990). shareholder gains from callable-bond refundings.managerial and decision economics, 11(1), 57–63. fama, e. f. (1977). a pricing model for the municipal bond market. unpublished working paper: university of chicago. feenberg, d. r., & poterba, j. m. (1992). which households own municipal bonds? evidence from tax returns. national tax journal, 45(4), 93–104. 84 r. brooks / financial services review 8 (1999) 71–85 fortune, p. (1991). the municipal bond market, part i: politics, taxes, and yields. federal reserve bank of boston.new england economic review,(sept./oct), 13–36. green, r. c. (1993). a simple model of the taxable and tax-exempt yield curves.the review of financial studies, 6(2), 233–264. johnson, v. e., & johnson, l. m. (1996). tax aspects of municipal bonds.journal of accountancy, 181(4), 51–58. jordan, b. d., & jorgensen, r. d. (1998). measuring the benefit of a bond refunding.financial practice and education, 8(1), 29–36. kalotay, a. j., & williams, g. o. (1992). the valuation and management of bonds with sinking fund provisions. financial analysts journal, 48(2), 59–67. kidwell, d. s., & koch, t. w. (1983). market segmentation and the term structure of municipal yields.journal of money, credit, and banking, 15(1), 40–55. koch, t. w., & stock, d. (1997). an analysis of implied tax rates on long-term taxable and tax-exempt bonds. journal of business research, 38(2), 171–176. kochin, l. a., & parks, r. w. (1988). was the tax-exempt bond market inefficient or were future expected tax rates negative?journal of finance, 43(4), 913–931. leibowitz, m. l. (1981). volatility in tax-exempt bonds: a theoretical model.financial analysts journal, 37(6), 31–52. mallman, t. l. (1981). volatility in municipal bonds: estimating and using volatility factors.financial analysts journal, 37(6), 54–59. mcentee, c. (1998, july 27). s & p 500reports secondary market volume soars in period of low vrdn supply. the bond buyer, p. 31. metcalf, g. e. (1992). the role of federal taxation in the supply of municipal bonds: evidence from municipal governments.national tax journal, 45(4), 57–70. ott, d. j., & meltzer, a. h. (1963).federal tax treatment of state and local securities.washington, dc: the brookings institution. poterba, j. m. (1986). explaining the yield spread between taxable and tax-exempt bonds: the role of expected tax policy. in h. s. rosen (ed.),studies in state and local public finance.chicago, il: the university of chicago. poterba, j. m. (1989).tax reform and the market for tax-exempt debt.working paper no. 2900. cambridge, ma: national bureau of economic research. rosenbloom, r. h. (1976). a review of the municipal bond market. federal reserve bank of richmond. economic review, 62(2), 10–19. severn, a. k., & stewart, w. j. (1992). the corporate-treasury yield spread and state taxes.journal of economics and business, 44(2), 161–166. singh, s., & dresnack, w. h. (1997). market knowledge in managed municipal bond portfolios.financial services review, 6(3), 185–196. sorensen, e. h. (1983). who puts the slope in the municipal yield curve? the journal of portfolio management, 9(4), 61–65. stock, d. (1985). price volatility of municipal discount bonds.journal of finance research, 8(1), 1–13. the bond market association. (1997).research quarterly.new york: the bond market association (http:// www.bondmarkets.com). the bond market association (1998).research quarterly.new york: the bond market association (http:// www.bondmarkets.com). trzcinka, c. (1982). the pricing of tax-exempt bonds and the miller hypothesis.the journal of finance, 37(4), 907–923. van horne, j. c. (1987). call risk and municipal bonds.journal of portfolio management, 13(2), 53–57. wilson, e. r., & stewart, j. p. (1990). municipal financial reporting and competition among underwriters for new issues of general obligation bonds.contemporary accounting research, 6(2), 573–592. yawitz, j. b. (1978). risk premia on municipal bonds.journal of financial and quantitative analysis, 13(3), 475–485. 85r. brooks / financial services review 8 (1999) 71–85 pii: s1057-0810(97)90037-8 financial services review, 6(1): 75 issn: 1057-0810 ..... i copyright © 1997 by jai press inc. f reproduction in any form reserved. all rights of rep erratum the following page of references should have appeared as p. 148 in the article entitled "the effects of mutual fund managers' characteristics on their portfolio performance, risk and fees," by joseph h. golec, which was published in financial services review, vol. 5, no. 2, 1996. we apolo gize for any inconvenience this oversight may have caused. elton, edwin j., gruber, martin j., das, sanjiv, & hlavka, matthew. (1993). efficiency with costly infor mation: a reinterpretation of evidence from managed portfolios. review of financial studies, 6, 1-22. fama, eugene f. (1980). agency problems and the theory of the firm. journal of political economy, 88(april), 288-307. friend, irwin, & blume, marshall. (1970). measurement of portfolio performance under uncertainty. american economic review, 60(september), 561-575. gibbons, robert, & murphy, kevin j. (1992). optimal incentive contracts in the presence of career con cerns: theory and evidence. journal of political economy, i00, 468-505. goetzmarm, william n., & ibbotson, roger g. (1994). do winners repeat? patterns in mutual fund perfor mance. journal of portfolio management, 20(winter), 9-17. golec, joseph h. (1988). do mutual fund managers who use incentive compensation outperform those who don't? financial analysts journal, 44(november/december), 75-78. golec, joseph h. (1992). empirical tests of a principal-agent model of the investor-investment advisor relationship. journal of financial and quantitative analysis, 27(march), 81-95. grinblatt, mark, & titman, sheridan. (1989). mutual fund performance: an analysis of quarterly portfolio holdings. journal of business, 62, 393--416. grinblatt, mark, & titman, sheridan. (1992). the persistence of mutual fund performance. journal of finance, 47(december), 1977-1984. grinblatt, mark, & titman, sheridan. (1994). a study of monthly mutual fund returns and performance evaluation techniques. journal of financial and quantitative analysis, 29(september), 419-444. hendricks, darryll, pate1, jayendu, & zeckhauser, richard. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974-1988. journal of finance, 48(march), 93-130. holmstrom, bengt, & ricart i costa, joan. (1986). managerial incentives and capital management. quar terly journal of economics, november, 835-860. ippolito, richard a. (1989). efficiency with costly information: a study of mutual fund performance, 1965-1984. quarterly journal of economics, 104(february), 1-23. jensen, michael c. (1968). the performance of mutual funds in the period 1945-1964. journal of finance, 23(may), 389-416. lambert, richard. (1986). executive effort and selection of risky projects. rand journal of economics, 17(spring), 77-88. lee, cheng-few, & rahman, shafiqur. (1990). market timing, selectivity, and mutual fund performance: an empirical investigation. journal of business, 63, 261-278. levy, haim. (1978). equilibrium in an imperfect market: a constraint on the number of securities in the portfofio. american economic review, 68(september), 643-658. mcgough, robert. (1993a, june 29). mutual fund old-timers figure a penny saved is a penny saved. wall street journal, p. c 1. mcgough, robert. (1993b, march 23). hot stock funds can cool after closing. wall street journal, p. c1. mincer, jacob. (1973). schooling, age, and earnings. new york: columbia university press. starks, laura t. (1987). performance'incentive fees: an agency theoretic approach. journal of financial and quantitative analysis, 22(march), 17-32. sharpe, william f. (1966). mutual fund performance. journal of business, 39, 119-138. tinic, seha m., & west, richard r. (1986). risk, return, and equilibrium: a revisit. journal of political economy, 94, 126-147. topel, robert. (1991). specific capital, mobility and wages: wages rise with job seniority. journal of political economy, 99, 145-176. trueman, brett. (1988). a theory of noise trading in securities markets. journal of finance, 43(march), 83-96. financial services review, 32(4) 51 social determinants of health and desirable financial behaviors: the mediation effect of financial knowledge jia qi,1 yu zhang,2 and sheri worthy3 abstract the social determinants of health (sdh) include several conditions in a person’s environment, including economic stability, educational access and quality, health care access and quality, neighborhood and built environment, and social and community context. in this study, we investigated how sdh is related to desirable financial behaviors while considering financial knowledge as a mediation variable, using data collected in 2021 during the covid-19 pandemic. social cognitive theory (sct) was used to develop the theoretical framework. the results of this study indicate that economic stability, educational access and quality, health care access and quality, and social and community context are positively associated with desirable financial behaviors. further, financial knowledge plays a significant role in mediating the relationships between sdh and desirable financial behaviors. this paper provides implications on the effect of sdh on financial behaviors, underscoring the relevance of applying the sct framework to examine the interplay between social factors and financial behaviors. the discussion and implications section of this paper provides strategic direction to enhancing financial behaviors through the improvement of sdh and financial knowledge. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation qi, j., zhang, y., & worthy, s. (2024). social determinants of health and desirable financial behaviors: the mediation effect of financial knowledge. financial services review, 32(4), 5075. introduction the worldwide covid-19 pandemic struck the united states in march 2020, resulting in health inequalities and disparities raised by financial burdens for economically disadvantaged households (shek, 2021). covid-19 affected people’s physical and mental health for shortand long-term periods (mueller et al., 2021). while much of the extant literature focuses on the 1 corresponding author (jiaqi@uga.edu). university of georgia, athens, ga, usa 2 kansas state university, manhattan, ks, usa 3 mississippi state university, starkville, ms, usa impacts of covid-19 on the traditional criteria of health, this research broadens the scope by examining consumers’ social factors of health, which are defined as social determinants of health (sdh) by the world health organization (who). these factors, which include economic stability, educational access and quality, health care access and quality, neighborhood and built environment, and social and community context, are expected https://creativecommons.org/licenses/by-nc/4.0/ mailto:jiaqi@uga.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(4) 52 to better describe an individual’s well-being during the pandemic (office of disease prevention and health promotion, 2023). covid-19 caused negative economic and financial consequences at the family and personal level (anand et al., 2020; shek, 2021; yuesti et al., 2020). the pandemic contributed to family poverty and a reduction in financial capital (buheji et al., 2020; shek, 2021). the scope of personal or family-level finances includes savings, spending, borrowing, and planning (xiao, 2016). for many consumers, it was challenging to maintain desirable financial behaviors and make reasonable financial decisions during the financial crisis resulting from the pandemic. for families that had financial difficulties before the pandemic, their situation worsened (cantor & landry, 2020). however, families that had sufficient liquid savings and were making stable debt and saving payments before the pandemic found it easier to survive and recover from the resulting financial shocks (fox & bartholomae, 2020). households experiencing unstable economic resources due to unemployment, however, were more likely to experience financial adversities (fox & bartholomae, 2020). overall, 42% of u.s. households experienced economic instability during the pandemic (parker et al., 2020). all these issues are closely related to the domain of sdh. for this research, we investigated how the social determinants of health relate to desirable financial behaviors, while considering the mediation effect of financial knowledge, which is expected to help consumers with financial adversities face financial shocks (yuesti et al., 2020). implications from the findings include approaches consumers can use to stabilize their finances in preparing for future unexpected financial shocks. this information will help consumers make more appropriate financial decisions and engage in desirable financial behaviors even during a pandemic so that they will quickly recover from financial shocks. the results of this study indicate that economic stability, educational access, healthcare access, and social context are significantly associated with desirable financial behaviors. these social factors are also associated with financial knowledge, which has a significant mediation effect on the relationship between the sdh and desirable financial behaviors. results indicate that consumers with more financial knowledge are more likely to demonstrate desirable financial behaviors. therefore, before and during the pandemic, consumers might consider stabilizing their economic resources, maintaining or creating better education and healthcare access, enhancing their social connections, and gaining more personal financial knowledge to improve financial behaviors and financial decisions. the results also provide valuable implications for policymakers and financial advisors. before or during a future pandemic or other crisis, policymakers could implement timely approaches to prevent income drops, increase the accessibility of education and healthcare, and create more opportunities for consumers to improve social connections. while working with clients, financial advisors can assess their clients’ five components of sdh. based on the information obtained, financial advisors can create personalized plans and targeted strategies to assist their clients in enhancing financial behaviors. literature review and hypotheses social determinants of health the traditional criterion for assessing health conditions is to use self-reported health status or reports of access to health care and/or health insurance. however, over the past few decades, researchers have defined more comprehensive measures for determining health status. these measures increase the scope of health by taking social factors into consideration. this body of evidence does not deny the importance of perceived health, health insurance, or access to health care in affecting health but instead illustrates that these are not the only determinants of health outcomes (braveman & gottlieb, 2014). the social factors included in the sdh domain are more related to the conditions experienced by people daily; for example, their unemployment, education, and community engagement status (commission on social determinants of health, 2008). https://www.ncbi.nlm.nih.gov/pmc/articles/pmc8152921/#ref-13 https://onlinelibrary.wiley.com/doi/full/10.1002/cfp2.1103#cfp21103-bib-0003 https://onlinelibrary.wiley.com/doi/full/10.1002/cfp2.1103#cfp21103-bib-0022 qi et al. 53 the world health organization (who) defines the social determinants of health (sdh) as “the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life” (commission on social determinants of health, 2008). in general, sdh represents nonmedical factors associated with health outcomes. the u.s. department of health and human services has divided sdh into five domains: (a) economic stability; (b) educational access and quality; (c) healthcare access and quality; (d) neighborhood and built environment; and (e) social and community context (office of disease prevention and health promotion, 2023). in one study using the five social determinants of health categories, medical care was found to explain only 10% to 15% of preventable deaths in the united states (mcginnis et al., 2002), while another study indicated that 40% of mortality is caused by behavioral issues that can be prevented (olsen et al., 2010). several previous articles have confirmed the notion that health outcomes are strongly associated with social factors, including employment, income, and environment (braveman et al., 2011; mackenbach, 1996; mcginnis et al., 2002). research by holtlunstad (2022) introduced a systemic framework to justify the prioritization of social connection in sdh. recent studies have demonstrated the presence of specific social factors in determining health outcomes. social frailty (ragusa et al., 2022) and social isolation (naito et al., 2023) are known to be significant factors increasing the risk of allcause mortality, where social frailty is characterized as the lack of social resources, social activities, and self-management abilities (ma et al., 2018), and social isolation is defined as the state of lacking social relationships, which encompass social contacts, social resources, and participation in social or religious activities (umberson & karas, 2010). additionally, education is known to be associated with health outcomes (bautista et al., 2021; gonzález et al., 2022), with half of all deaths of working adults being accounted for by factors related to lower education (jemal et al., 2008). economic stability based on the healthy people 2020 framework on sdh, economic stability is thought to be influenced by various factors, including employment and work environment, food and housing instability, and poverty rates (office of disease prevention and health promotion, 2023). employment status has also been recognized as a significant component of economic stability (hergenrather et al., 2015). employment and reemployment are positively associated with better physical health, whereas unemployment and job loss are associated with worse physical health (hergenrather et al., 2015). individuals with different levels of employment engage in financial behaviors differently, which typically has a positive influence on individuals’ long-term financial health (rostamkalaei et al., 2022). in times of economic difficulties, such as a pandemic, fluctuations in income may serve as an indicator of economic stability. an individual’s financial behaviors are influenced by their expectation of income volatility (sherraden et al., 2000). however, during periods of economic difficulties, individuals who encounter fluctuations in their income and a lack of sufficient information about their future income prospects may confront challenges in managing their savings and consumption levels. this can lead to less effective coping mechanisms and hinder their ability to exhibit desirable financial behaviors. literature has also linked financial knowledge to economic stability. previous studies show that consumers with more financial knowledge are more likely to save and invest (lusardi & mitchell, 2014). they are also more likely to manage their finances effectively (thomas, 2019). abdullah et al. (2023) indicated that financial literacy is positively associated with financial stability for low-income households. educational access and quality human capital is a crucial investment that individuals make to increase their potential earnings and build wealth (wolla & sullivan, 2017). well-educated individuals are more likely to report better financial conditions (wolla & sullivan, 2017). a higher level of attained financial services review, 32(4) 54 education is not only associated with increased earning potential; individuals with a higher level of education also demonstrate a preference for proactive engagement in preventative health measures (fletcher & frisvold, 2009). this propensity toward health-conscious practices highlights the comprehensive benefits of education, which go beyond monetary gains. individuals with higher educational attainment are more likely to have a solid foundation in numeracy and literacy (richards et al., 2009), which is essential for comprehending complex financial products. these foundational skills, frequently refined through extended schooling, enable individuals to make informed financial decisions (wolla & sullivan, 2017). research has also established a positive association between education level and favorable outcomes such as financial market participation, investment income, and retirement income (cole et al., 2014). even an additional year of education has been found to positively affect credit rating scores, borrowing choices, and credit behaviors, substantially decreasing the likelihood of individuals declaring bankruptcy or experiencing foreclosure (cole et al., 2014). additionally, education is known to be positively associated with financial knowledge (rothwell & wu, 2019; walstad et al., 2010; xiao & o’neill, 2016). healthcare access and quality the availability of healthcare services can impact the financial burdens associated with medical expenses and the demand for health-related financial products and services (social policy institute, 2022). health insurance serves a dual purpose of protecting individuals in times of health-related challenges and generating positive impacts beyond the realm of health. having medical debts in collections is often seen as a sign of broader financial distress (batty et al., 2022), which suggests that having sufficient health insurance coverage has a positive impact on various financial behaviors. the literature has documented the influence of insurance ownership on several financial behaviors, such as mortgage decisions (houle & keene, 2014), ownership of risky assets (li et al., 2021), retirement planning (rogowski & karoly, 2000), and saving behaviors (costa‐font & vilaplana-prieto, 2017; laffargue & padieu, 2016). for example, changes in health status increase the risk of mortgage default and foreclosure, which can be partially mediated by health insurance (houle & keene, 2014). the availability of health insurance with better insurance terms is also associated with households’ investment portfolio allocation to risky assets (li et al., 2021). holding health insurance has been shown to have a significant impact on retirement behaviors. specifically, the ability to acquire retiree health benefits increases the probability of retirement (rogowski & karoly, 2000). similarly, the provision of public funding for long-term care has also been shown to be associated with a reduction in household savings, as individuals rely on this support to meet their long-term care needs (costa-font & vilaplanaprieto, 2017). consumers with private medical insurance, on the other hand, have greater savings and are more likely to actively save for future healthcare costs against future uncertainty (guariglia & rossi, 2002). consumers who report higher educational attainment, with more subjective knowledge, are more likely to hold health insurance coverage (abdel‐ghany & wang, 2001; dewar, 1998). o’connor and kabadayi (2020) also suggested that health insurance literacy is positively associated with subjective and objective financial knowledge. neighborhood and built environment a well-constructed environment can mitigate health and safety issues, thus enhancing overall quality of life (office of disease prevention and health promotion, 2023). government safety nets for assistance are generally designed to support eligible people in need. government assistance takes various forms, including medicaid, the earned income tax credit (eitc), and the supplemental nutrition assistance program (snap). this assistance has been found to promote financial satisfaction (lee et al., 2023). economic downturns and natural disasters contribute significantly to the rise in enrollment in government aid programs (gruber & sommers, 2020). according to lee et al. (2023), government safety net programs are often implemented as temporary solutions, which may qi et al. 55 not transform into accumulated savings and asset ownership. however, one study found that individuals who received stimulus checks during the covid-19 pandemic exhibited a greater propensity to save as the amount increased (liu et al., 2023). the literature shows that there is a positive association between receiving government assistance and an individual’s level of financial satisfaction (lee et al., 2023), which in turn may potentially result in positive financial behaviors, especially for low-income families. assistance from the government can support low-income households in increasing savings, reducing outstanding medical debts, preventing new delinquencies, and enhancing credit ratings (brevoort et al., 2017). placing confidence in the government’s ability to offer aid in times of crisis might motivate individuals to adopt proactive actions, thereby fostering the development of positive financial practices and bolstering their sense of safety. someone’s neighborhood and built environment is also thought to be associated with financial literacy or financial knowledge. li et al. (2022) suggested the built environment could impact a consumers’ financial literacy through exposure to knowledge acquisition opportunities. barrafrem et al. (2021) noted that trust in the government is related to financial literacy and education. public trust in government, however, tends to be associated with higher financial knowledge (niţoi & pochea, 2024). social and community context individuals’ well-being can be influenced by various factors, including the conduct of their peers, societal standards, and the availability of community resources (office of disease prevention and health promotion, 2023). those who possess robust social networks and support systems generally exhibit improved financial behavior, resulting in enhanced financial wellbeing (lebaron & kelley, 2020). recent studies on family socialization show that constructive familial and social interactions foster desirable financial behaviors. this interaction includes verbal and nonverbal communication between parents and their children, among other family members, and within romantic partnerships (lebaron & kelley, 2020). social networks can serve as valuable resources for obtaining knowledge and guidance on financial matters. engaging in social interactions provides valuable learning opportunities, which may promote the development of positive financial behaviors (lebaron & kelley, 2020). for example, college students’ savings and budgeting behaviors have been shown to be positively associated with financial social learning opportunities (gutter et al., 2010). the influence of co-workers also plays a crucial role in determining an individual’s ability to save (jamal et al., 2015). strong social relationships can thus serve as a protective barrier, enabling individuals to maintain positive behavior despite facing challenging situations. social context is also associated with financial literacy. bongomin et al. (2020) noted that financial literacy and social networks are positively related (i.e., social networks and relationships can act as conduits for consumers to obtain access to and gain financial knowledge) (reagans & mcevily, 2003; uzzi, 2018). based on the reviews of each sdh factor from above, we propose a positive relationship between sdh and desirable financial behavior and a positive association between sdh and financial knowledge, such that: h1: social determinants of health are positively associated with desirable financial behaviors. h2: social determinants of health are positively associated with financial knowledge. financial knowledge and financial behavior the oxford dictionary (n.d.) defines “knowledge” as “facts, information, and skills acquired by a person through experience or education; the theoretical or practical understanding of a subject.” in this sense, financial knowledge can be thought of as a basic understanding of financial concepts and the practical use of that knowledge (delgadillo & law, 2019). previous studies demonstrate that individuals must know financial concepts to help motivate behaviors in managing taxes (baumeister et al., 2003). tang and baker (2016) used financial knowledge as one of the antecedents of financial behavior and financial services review, 32(4) 56 concluded that objective financial knowledge is associated with financial behavior. therefore, in this study (and more broadly), objective financial knowledge is expected to determine financial decisions. past literature provides empirical evidence on the positive relationship between objective financial knowledge and financial behavior (courchane & zorn, 2005; robb & woodyard, 2011; tang & baker, 2016). chen and volpe (1998) used a college student sample to identify the relationship between financial knowledge and financial decisions, concluding that students with more financial knowledge are more likely to report desirable financial behaviors such as keeping financial records. lusardi and mitchell (2007) indicated that individuals with more financial knowledge tend to exhibit better retirement preparedness. further, mitchell and lusardi (2022) showed that financial knowledge helped improve financial decisions and well-being during the covid-19 pandemic. another study focusing on older people indicated that individuals with greater financial knowledge tend to better manage their finances (kim et al., 2018). therefore, in this study, financial knowledge is expected to be positively associated with desirable financial behavior, such that: h3: financial knowledge and desirable financial behavior are positively related. theoretical framework social cognitive theory (sct) was utilized as the conceptional framework of this research. sct was developed from social learning theory (slt), which explains the importance of environmental and cognitive factors in affecting human learning activities (bandura & walters, 1977). bandura (1986) revised slt to sct and posited that behavioral outcomes occur in a social context with interactions among the environmental, personal, and behavioral factors. another feature of sct is the concentration on personal social experiences while interacting with personal traits to shape individual behavior (bandura, 1991). in this sense, sdh can be fitted as an environmental factor that explains the nonmedical part of health outcomes, including economic stability, educational access and quality, health care access and quality, neighborhood and built environment, and social and community context (office of disease prevention and health promotion, 2023). since financial behavior refers to practices, including cash, credit, and saving behaviors (xiao, 2008), desirable financial behavior was used as the behavioral factor in the model. financial knowledge is a personal factor that indicates how well an individual understands financial concepts and uses the knowledge (delgadillo & law, 2019). the theoretical framework, developed by adapting sct, is shown in figure 1. method survey design the processes of survey development, data collection, and data documentation were completed as part of usda agricultural experiment station research. this work was supported by the usda national institute of food and agriculture, hatch-multistate project 1017241, and cooperating universities as an output of the project officially known as nc2172: “behavioral economics and the intersection of healthcare and financial decision making across the lifespan.” researchers are the project investigators in their respective states, which include california, colorado, florida, georgia, illinois, indiana, iowa, kansas, maryland, mississippi, missouri, north dakota, ohio, pennsylvania, rhode island, south dakota, texas, utah, virginia, and washington. data were collected during the covid-19 pandemic from may 24, 2021 to june 18, 2021 by qualtrics research services. qualtrics was contracted to collect at least 2,000 completed cases. to confirm the identity of respondents, qualtrics sample partners verified each respondent’s address, demographic information, and email address upon registration. qualtrics also checked every ip address and applied unique digital fingerprinting techniques to ensure validity and exclude duplication. to ensure reliability, sample blend was replicated across multiple projects. quotas were used while collecting the data. within each group, the following additional sampling parameters were used: qi et al. 57 (a) gender: approximately 49.5% male, 50.5% female; (b) region: approximately 16% northeast, 21% midwest, 41% south, 22% west; and (c) race: approximately 60.1% nonhispanic white, 13.4% non-hispanic black, 18.5% hispanic, 8% other. to be in the study, respondents needed to be between ages 18 and 65 and live in the united states. respondents received an incentive based on the completion of the survey. figure 1. theoretical framework dataset in the final dataset, 2,318 respondents completed all the survey questions. ip addresses were deleted before the data were analyzed. the survey included detailed information about respondents’ sociodemographic characteristics, healthcare access, government assistance during the pandemic, the social connections of respondents, and questions related to objective financial knowledge from the national financial capability study (nfcs). descriptive statistics are shown in table 1 and table 2. financial services review, 32(4) 58 table 1. description of the variables used in the study variables mean/percentage min max type desirable financial behavior 41.33% 0 1 binary social determinants of health economic stability 32.79% 0 1 binary education 2.77 0 5 continuous healthcare 82.75% 0 1 binary environment 24.16% 0 1 binary social context 2.22 0 5 continuous financial knowledge 2.26 0 6 continuous variables dependent variable the dependent variable of interest in this study was desirable financial behavior. according to xiao et al. (2014), desirable financial behavior includes numerous factors related to savings, retirement, and mortgage use. in this paper, desirable financial behavior was measured based on the respondents’ answers about whether they engaged in desirable financial behaviors in the past 12 months (i.e., during the covid-19 pandemic). as shown in table 2, three desirable financial behaviors were used to construct the variable: (a) “added additional money to a retirement saving account (yes = 10.91%; no = 89.09%)”; (b) “paid additional money toward mortgage principle (yes = 6.73%; no = 93.27%)”; and (c) “saved additional money (yes = 31.10%; no = 68.90%).” as shown in table 1, desirable financial behavior (yes = 41.33%; no = 58.67%) was coded as 1 if a respondent engaged in at least one of the three desirable financial behaviors in the past 12 months and coded as 0 if a respondent did not engage in any of these behaviors. desirable financial behavior was coded as a binary variable. as shown in table 2, the statistics for desirable financial behavior showed that only 0.95% of respondents engaged in all three desirable financial behaviors, whereas 5.52% of respondents engaged in two. therefore, a new variable was created where respondents who engaged in at least one of the three desirable financial behaviors were coded 1, otherwise 0. qi et al. 59 table 2. descriptive statistics variables mean/% sd freq. added additional money to a retirement saving account yes 10.91% 253 no 89.09% 2,065 paid additional money toward mortgage principle yes 6.73% 156 no 93.27% 2,162 saved additional money yes 31.10% 721 no 68.90% 1,597 the number of engaged desirable financial behavior none 58.67% 1,360 one 34.86% 808 two 5.52% 128 three 0.95% 22 employment working 73.42% 1702 not working 26.58% 616 educational attainment high school or less (ref) 27.69% 642 some college 21.92% 508 associate’s degree 9.92% 230 bachelor’s degree 24.33% 564 graduate or professional degree 16.13% 374 health insurance yes 82.57% 1,914 no 17.43% 404 favorable environment yes 24.16% 560 no 75.84% 1,758 age 36.46 8.745 gender male 50.52% 1,171 female 48.27% 1,119 other 1.21% 28 race non-hispanic white 59.36% 1,376 other 40.64% 942 health condition 0.00% financial services review, 32(4) 60 excellent 24.81% 575 very good 28.73% 666 good 28.86% 669 fair 13.76% 319 poor (ref) 3.41% 79 marital status married 47.45% 1,100 other 52.55% 1,218 household size 2.97 1.396 income less than $30,000 (ref) 32.53% 754 $30,000 to $99,999 43.74% 1,014 $100,000 to $200,000 20.19% 468 more than $200,000 3.54% 82 n = 2,318 independent variables the key independent variable used in this research was sdh. the five categories of sdh were economic stability, educational access, health care access, neighborhood and built environment, and social and community context. as shown in table 1, economic stability (yes = 32.79%; no = 67.21%) was measured based on whether a respondent’s income was higher, lower, or about the same in 2020 compared with 2019 before the pandemic. economic stability was coded as 1 if income was about the same; otherwise, it was coded as 0. educational access was indicated by respondents’ educational attainment. a higher education value indicates that a respondent had better access to education. table 1 shows that the mean score (m = 2.77) of education ranged from 0 = “less than high school” to 5 = “graduate or professional degree.” healthcare access was measured by whether a respondent was currently covered by health insurance. as shown in table 1, 82.57% of respondents were covered by health insurance. the neighborhood and built environment variable was created based on this question: “how much do you worry about your ability to rely on a government safety net for financial assistance during times of crisis?” the environment factor (yes = 24.16%; no = 75.84%) was coded as 1 when a respondent reported, “don’t worry at all,” indicating they were living in a safe and supportive environment. the variable was coded as 0 if a respondent answered, “i worry some” or “i worry a lot.” the social and community context variable was constructed based on answers ranging from “strongly disagree” to “strongly agree” to the following questions: (a) i frequently pass on social events at work due to my financial situation. (b) my financial situation frequently interferes with my relationship with coworkers/colleagues. (c) i find it difficult to talk about money with my spouse/significant other. (d) i frequently avoid attending family events because of my financial situation. (e) my financial situation frequently interferes with my family relationship. each of the questions was coded as 1 if a respondent “strongly disagreed” or “disagreed”; otherwise, it was coded as 0. the social and community context variable was then estimated as the sum of the values of the five single variables. as shown in table 1, social and qi et al. 61 community context scores ranged from 0 to 5 (m = 2.22). a lower score means a respondent was experiencing a less-than-optimal social context; a higher score indicates a respondent had excellent social connections. the cronbach’s alpha for the summed scale was 0.86. financial knowledge was used as a mediator in this research. the mediation effect of financial knowledge on the relationship between sdh and desirable financial behavior was evaluated in this study. the financial knowledge variable was constructed based on answers to six household finance-related questions from the national financial capability study (nfcs). as shown in table 1, financial knowledge (m = 2.26) was coded so that scores ranged from 0 to 6. respondent demographics were used as control variables in this research. the variables included age, gender, health situation, household size, marital status, and income. the descriptive statistics for the variables are shown in table 2. the average age of respondents was 36 years. males comprised 50.52% of the sample, females were 48.27%, and 1.21% were other gendered respondents. the percentage of non-hispanic white respondents was 59.36%, and 45.45% of respondents were married, with an average household size of three. also, in this study, 32.53% of respondents reported having an income less than $30,000, whereas 3.54% of respondents had an income greater than $200,000. among the respondents, 73.42% were working, and 26.58% were not working. as shown in table 2, 27.69% of respondents had an education of high school or less, 21.92% had completed some college, 9.92% had an associate degree, 24.33% had earned a bachelor’s degree, and 16.13% had completed a graduate or professional degree. additionally, respondents were asked about their self-reported physical health. the responses were excellent (24.81%), very good (28.73%), good (28.86%), fair (13.76%), and poor (3.41%). empirical models this study used causal steps to access mediation (baron & kenny, 1986; mackinnon et al., 2007). since the dependent variable was categorical and the mediator was treated as a continuous variable, the techniques of ordinary least squares (ols) regression and probit regression were used to evaluate the mediation effect of financial knowledge (mackinnon et al., 2007). first, a probit regression model was used to determine model relationships. the probit regression was formulated as follows: φ−1(𝑝) = 𝛽0 + 𝛽1𝑆 + 𝛽2𝐶 + 𝜀 (1) where p = p (desirable financial behavior = 1/s, c), s is the combination of the five factors of social determinants, c is the vector denoting control variables, and ε is the error term. a second estimation (equation 2) was used to validate the associations between sdh and financial knowledge: 𝑓 = 𝛽0 + 𝛽1𝑆 + 𝛽2𝐶 + 𝜀 (2) where f is financial knowledge, s is the combination of the significant factors of social determinants according to the result of the last step, c is the vector denoting control variables, and ε is the error term. equation 3 was used to examine the relationship between financial knowledge and desirable financial behavior: φ−1(𝑝) = 𝛽0 + 𝛽1𝑓 + 𝛽2𝐶 + 𝜀 (3) where p = p (desirable financial behavior = 1/s, c), f is financial knowledge, c is the vector denoting control variables, and ε is the error term. the direct and indirect effects of sdh on desirable financial behavior were investigated after all the above relationships were validated. the product of coefficients method was used to estimate the direct and indirect effects (alwin & hauser, 1975; statacorp, 2023). the estimated total effect (te), natural indirect effect (nie), and natural direct effect (nde) were estimated in stata using a causal mediation model (statacorp, 2023). results the probit regression results are presented in table 3. the model indicated that economic stability (beta = 0.209; p < 0.001), educational access (beta = 0.052; p < 0.05), healthcare access (beta = 0.214; p < 0.01), and social context (beta = 0.098; p < 0.001) had significant positive relationships with desirable financial behaviors; however, the built environment was not financial services review, 32(4) 62 significant. compared with the reference group whose annual income was lower than $30,000, those who earned between $100,000 and $200,000 (beta = 0.210; p < 0.05) were more likely to report desirable financial behaviors during the pandemic. table 3. probit regression for the social determinants of health on desirable financial behavior (n = 2,318) coefficient std. err. p > z economic stability 0.209*** 0.058 0.000 education 0.052* 0.022 0.021 healthcare 0.214** 0.074 0.004 environment 0.033 0.065 0.609 social context 0.098*** 0.014 0.000 age 0.004 0.003 0.264 female -0.083 0.059 0.162 non-hispanic white 0.008 0.059 0.894 excellent -0.011 0.160 0.946 very good 0.032 0.156 0.838 good 0.081 0.152 0.597 fair -0.042 0.159 0.794 hh size 0.022 0.021 0.281 married -0.041 0.067 0.542 in $30k to $100k 0.058 0.068 0.392 in $100k to $200k 0.210* 0.101 0.039 more than $200k 0.013 0.163 0.937 employment status 0.092 0.069 0.182 note: unweighted; significance: *p < .05, **p < .01, ***p < .001 the significant social determinants of health factors were used to test hypothesis 2. results from the ols regression on financial knowledge (table 4) indicated that economic stability (beta = -0.134; p < 0.001) was negatively associated with financial knowledge. educational access (beta = 0.120; p < 0.001), healthcare access (beta = 0.210; p < 0.01), and social context (beta = 0.076; p < 0.001) were positively associated with financial knowledge. age (beta = 0.021; p < 0.001) was positively associated with financial knowledge, whereas female (beta = -0.302; p < 0.001) was negatively associated with financial knowledge. compared with the reference group, whose income was less than $30,000, respondents who made between $30,000 and $100,000 and between $100,000 and $200,000 were more likely to exhibit higher financial knowledge. the results shown in table 4 provide support for the second hypothesis. qi et al. 63 table 4. ols regression for economic stability, educational access, healthcare access, and social and community context on financial knowledge (n = 2,318) coefficient std. err. p > z economic stability -0.314*** 0.060 0.000 education 0.120*** 0.023 0.000 healthcare 0.210** 0.074 0.005 social 0.076*** 0.014 0.000 age 0.021*** 0.003 0.000 female -0.302*** 0.061 0.000 non-hispanic white 0.016 0.060 0.794 excellent -0.048 0.161 0.768 very good 0.309 0.158 0.050 good 0.107 0.154 0.486 fair 0.086 0.161 0.591 hh size -0.038 0.021 0.070 married -0.081 0.068 0.236 in $30k to $100k 0.340*** 0.070 0.000 in $100k to $200k 0.234* 0.105 0.026 more than $200k 0.041 0.169 0.807 employment status 0.062 0.070 0.374 note: unweighted; significance: *p < .05, **p < .01, ***p < .001 the financial knowledge probit regression results are shown in table 5. the findings indicate that financial knowledge (beta = 0.0489; p < 0.013) was positively associated with desirable financial behavior. compared with respondents whose income was less than $30,000, those who made between $30,000 and $100,000 and between $100,000 and $200,000 were more likely to engage in better financial behavior. results provide support for the third hypothesis. financial services review, 32(4) 64 table 5. probit regression for financial knowledge on desirable financial behavior (n = 2,318) coefficient std. err. p > z financial knowledge 0.048* 0.020 0.015 age 0.005 0.003 0.113 female -0.041 0.059 0.483 non-hispanic white 0.008 0.058 0.892 excellent 0.017 0.157 0.913 very good 0.096 0.154 0.531 good 0.117 0.150 0.438 fair -0.061 0.158 0.697 hh size 0.021 0.020 0.310 married -0.050 0.064 0.440 in $30k to $100k 0.127 0.067 0.058 in $100k to $200k 0.322*** 0.095 0.001 more than $200k 0.139 0.158 0.380 employment status 0.078 0.067 0.240 note: unweighted; significance: *p < .05, **p < .01, ***p < .001 the results of tests of the direct and indirect effects of sdh on desirable financial behavior are displayed in tables 6 through 9. as shown in table 6, financial knowledge significantly mediated the relationship between economic stability and desirable financial behavior. the total treatment effect of economic stability on desirable financial behavior was 0.101 (p < 0.05), meaning that if every respondent in the sample had stable economic resources, the probability of reporting a higher desirable financial behavior would increase by 0.101 points on the probability scale compared with respondents who had unstable income payments. the natural indirect effect of financial knowledge was -0.010 (p < 0.05), whereas the natural direct effect of economic stability was 0.111 (p < 0.001). this situation is referred to as “inconsistent mediation” because the mediated effect had an opposite sign than the direct effect in the model (mackinnon et al., 2000; 2007). although the indirect pathway had a negative association with desirable financial behavior, the effect was not strong enough to offset the positive effect of the direct pathway (i.e., financial knowledge only explained 10% (0.01/0.101) of the total treatment effect for economic stability on desirable financial behavior). the positive treatment effect on the outcome was almost completely due to the direct effect of economic stability. qi et al. 65 table 6. the mediation effect of economic stability on the relationship between educational access and desirable financial behavior (n = 2,318) causal mediation analysis outcome model: probit mediator model: linear mediator variable: financial knowledge treatment type: binary coefficient std. err. p > z nie economic stability -0.010* 0.004 0.017 nde economic stability 0.111*** 0.022 0.000 te economic stability 0.101*** 0.022 0.000 note: unweighted; nie: natural indirect effect, nde: natural direct effect, te: total effect; significance: *p < .05, **p < .01, ***p < .001 as shown in table 7, financial knowledge was a significant mediator in the relationship between educational access and desirable financial behavior. the total treatment effect was 0.072 (p < 0.01), which indicates that if every respondent in the sample had better educational access, the probability of engaging in desirable financial behavior would increase by 0.072 points compared with respondents having worse educational access. of the total treatment effect, a 0.02 increase in the probability of exhibiting desirable financial behaviors was due to the natural indirect effect of financial knowledge (p < 0.01); further, a 0.052% increase in the probability of showing desirable financial behaviors was due to the natural direct effect of educational access (p < 0.05). financial knowledge explained 27.8% of the total treatment effect for educational access on desirable financial behavior. financial services review, 32(4) 66 table 7. the mediation effect of financial knowledge on the relationship between educational access and desirable financial behavior (n = 2,318) causal mediation analysis outcome model: probit mediator model: linear mediator variable: financial knowledge treatment type: binary coefficient std. err. p > z nie economic stability 0.020** 0.006 0.001 nde economic stability 0.052* 0.023 0.021 te economic stability 0.072** 0.022 0.001 note: unweighted; nie: natural indirect effect, nde: natural direct effect, te: total effect; significance: *p < .05, **p < .01, ***p < .001 the effect of healthcare access on desirable financial behavior when mediated by financial knowledge is shown in table 8. financial knowledge significantly mediated the relationship between healthcare access and desirable financial behavior. the total treatment effect of educational access on desirable financial behavior was 0.103 (p < 0.001), meaning that if every respondent had better healthcare access, the probability of a higher desirable financial behavior would increase by 0.103 points compared with respondents with worse healthcare access. an 0.009 increase in the probability of exhibiting desirable financial behaviors was due to the natural indirect effect of financial knowledge (p < 0.05). an 0.094 increase in the probability of showing desirable financial behaviors was due to the natural direct effect of healthcare access (p < 0.001). financial knowledge explained 8.7% of the total treatment effect for healthcare access on desirable financial behavior. qi et al. 67 table 8. the mediation effect of financial knowledge on the relationship between healthcare access and desirable financial behavior (n = 2,318) causal mediation analysis outcome model: probit mediator model: linear mediator variable: financial knowledge treatment type: binary coefficient robust std. err. p > z nie healthcare access 0.009* 0.003 0.021 nde healthcare access 0.094*** 0.027 0.000 te healthcare access 0.103*** 0.026 0.000 note: unweighted; nie: natural indirect effect, nde: natural direct effect, te: total effect; significance: *p < .05, **p < .01, ***p < .001 table 9 presents the mediation effect of financial knowledge on the relationship between social context and desirable financial behavior. financial knowledge significantly mediated the relationship between social context and desirable financial behavior. the total treatment effect of social context on desirable financial behavior was 0.212 (p < 0.001), which suggests that if every respondent had an excellent social and community context, the probability of a higher desirable financial behavior would increase by 0.212 points compared with respondents who earned a zero score on social and community context. a 0.022 increase in the probability of reporting desirable financial behaviors was due to the natural indirect effect of financial knowledge (p < 0.01); further, a 0.192 increase in the probability of exhibiting desirable financial behaviors was due to the natural direct effect of social context (p < 0.001). financial knowledge explained 9.4% of the total treatment effect for social context on desirable financial behavior. discussion our study’s findings align with expectations from sct (bandura, 1986). in the context of the covid-19 pandemic, this study introduced and developed sct as the theoretical model using an empirical dataset to explain the connections between sdh and financial behaviors. while much of the past literature has concentrated on the analysis of physical health during the covid-19 pandemic, this research fills a research gap by introducing consumers’ social determinants of health and investigating its connections to financial behaviors. the findings confirm the importance of economic stability, educational access, healthcare access, and social context in describing consumers’ financial behaviors. financial services review, 32(4) 68 table 9. the mediation effect of financial knowledge on the relationship between social context and desirable financial behavior (n = 2,318) causal mediation analysis outcome model: probit mediator model: linear mediator variable: financial knowledge treatment type: binary coefficient robust std. err. p > z nie social context (5 vs 0) 0.020** 0.007 0.007 nde social context (5 vs 0) 0.192*** 0.029 0.000 te social context (5 vs 0) 0.212*** 0.028 0.000 note: unweighted; nie: natural indirect effect, nde: natural direct effect, te: total effect; significance: *p < .05, **p < .01, ***p < .001 four of the five sdhs were significant: (a) economic stability, (b) educational access, (c) healthcare access, and (d) social/community context. a strong positive association between economic stability and desirable financial behavior was observed. this suggests that consumers with stable economic resources are more likely to engage in desirable financial behaviors during times of crisis and thus find it easier to survive and recover from financial adversities. this finding confirms the importance of stable economic payments in preventing financial crises (fox & bartholomae, 2020). educational access was also positively associated with desirable financial behaviors, which means that consumers with better accessibility to education were more likely, during the pandemic, to make less problematic financial decisions. this result confirms the role of education in stabilizing financial situations. policymakers should consider increasing consumers’ overall access to education to prevent and prepare for future financial shocks. improving healthcare access is another vital way to help consumers make desirable financial decisions. as shown in this study, healthcare access was positively associated with desirable financial behavior. it is recommended that consumers should have health insurance in place to prepare for a future pandemic and to minimize stress-related financial difficulties resulting from unexpected medical and healthcare costs. additionally, social context was found to be positively associated with desirable financial behaviors, meaning consumers reporting healthy social relationships with family, friends, and coworkers are more likely to engage in desirable financial behaviors. it may be worthwhile for consumers to build up social networks as a way to receive support during financial shocks. the neighborhood and built environment variable was not significant. it is possible that the way the variable was measured explains the nonsignificant result. one question was used. the item asked about how worried respondents felt qi et al. 69 about relying on a government safety net for financial assistance during times of crisis. this question may be more of a measure of trust in government than a measure of healthy and safe neighborhoods and environments. this possibility is worthy of future research. findings also indicate that financial knowledge partially mediates the relationships between the sdh and desirable financial behaviors. the results of the mediation effect tests of financial knowledge explained the 10%, 8.7%, and 9.4% of the total treatment effect for economic stability, healthcare access, and social context, respectively. based on these findings, to improve their overall financial behavior, consumers should be encouraged to focus on increasing their economic stability, healthcare access, and social context. consumers should also try to acquire more financial knowledge to improve their financial behavior because financial knowledge plays a significant mediation role in explaining desirable behaviors. moreover, financial knowledge explained 27.8% of the total treatment effect for educational access on desirable financial behavior. this result indicates that enhancing financial education is one way to help consumers manage their finances wisely during an unexpected pandemic or crisis. it was also found that consumers with more attained education, healthcare access, and a better social context are more likely to report more financial knowledge. this aligns with expectations and conclusions from the literature. however, in this study, there was an unexpected relationship between economic stability and financial knowledge. surprisingly, the association between economic stability and financial knowledge was negative. financial worries and perceived stress could manifest in individuals experiencing financial strain, such as that caused by a job loss (frank et al., 2014; ryu & fan, 2023). those with stable incomes may perceive a lesser sense of urgency to improve their financial knowledge level in times of uncertainty relative to their counterparts with an unstable income stream. consequently, it is possible that individuals who have encountered fluctuations in their income may find it necessary to allocate time and resources toward acquiring financial knowledge and helping them thrive in difficult situations. furthermore, if a consumer’s salary or wages remain stable, they may be able to save more or pay more on their mortgage; however, it is possible there is no perceived need to improve financial knowledge as long as they are behaving in a way that benefits them in the long run. conclusions and implications this research considered social factors of health during the covid-19 pandemic by investigating the connections between sdh and financial behavior while considering the mediating role of financial knowledge. previous research indicates that sdh affects “a wide range of health, functioning, and quality-of-life outcomes and risks” (office of disease prevention and health promotion, 2023). this study shows that sdh is also associated with financial behaviors and knowledge. the results confirm the significance of improving consumers’ overall financial behaviors by improving economic stability, educational access, healthcare access, social context, and financial education. the findings of this study can be used to provide guidance to consumers when preparing for unforeseen financial shocks and to help them recover quickly from pandemic adversities. this study offers several theoretical contributions. one of the major theoretical contributions relates to the validity of sct and using self-reported data from a u.s. sample during the pandemic. the results confirmed the significance of sdh in explaining desirable financial behaviors and the mediation effect of financial knowledge. this confirmation provides evidence for sct as the theoretical framework to establish the associations between sdh and financial behavior. additionally, the mediation effect of financial knowledge on the relationship between sdh and desirable financial behavior was validated in this study. the existence of the mediation effect indicates that sdh can affect desirable financial behaviors directly and indirectly. the findings have practical implications for consumers, financial service practitioners, and policymakers. to prepare for unexpected financial shocks and build up financial resilience, consumers can enhance their financial situation(s) by improving their economic stability. stabilizing financial services review, 32(4) 70 employment and income resources increases economic stability and helps to ensure that consumers will have more liquid savings to manage their finances during a financial crisis. at the same time, consumers could improve their financial behaviors by attaining more education since advanced educational degrees and financial knowledge were shown to be directly associated with financial behaviors in this study. further, it is vital for consumers to have health insurance in place and funded so the expensive cost of healthcare does not significantly impact their finances during a pandemic or other crisis. moreover, consumers should build social capital and maintain healthy social connections as a way to manage their financial behavior during times of crisis. the results also have implications for financial service practitioners (e.g., financial counselors and financial planners). while working with clients to help improve their financial situations, financial service practitioners might consider taking some time to understand the conditions related to their clients’ social determinants of health. for example, suppose a client does not have reliable economic resources. when clients need help with their budgeting, financial service practitioners typically obtain clients’ financial information to identify whether they have stable income resources so that they can save regularly. financial service practitioners can go further and advise their client to look for employment that can provide stable income to prepare for any unforeseen financial crisis. financial service practitioners could also provide suggestions on setting up emergency fund accounts, which could be another avenue for clients to ensure economic support during a financial crisis. when clients need help with retirement planning, financial service practitioners could advise clients to maintain stable economic resources, so they have funds available to invest in their retirement accounts. clients should also be advised to acquire financial knowledge to improve retirement product choices. each client will have their own unique situation, so financial service practitioners need to personalize their recommendations based on the client’s social factors of health. educating clients about finances could be another effective approach in assisting clients to improve their financial behaviors. financial education will be of long-term benefit to clients. additionally, policies to improve any social determinant of health can have profound effects on individuals, families, and communities. the stimulus checks issued by the federal government during the covid-19 pandemic are a noteworthy example of helping clients recover and thrive during financial shocks. therefore, any policies that can provide income resources or prevent an income drop during a pandemic will be beneficial for clients in improving their financial behavior. lastly, policymakers should focus on improving sdh in neighborhoods to help improve the overall financial behavior of residents. communities can work on economic development and provide job opportunities for community members in preparation for future unexpected crises to raise and stabilize incomes. government agencies might consider designing and constructing walkable neighborhoods or public transportation to decrease the cost of living for consumers. local governments could also provide affordable housing in safe neighborhoods for low-income consumers and consumers with financial difficulties to help them navigate financial shocks. government agencies should also implement programs to improve language and literacy skills, including financial literacy skills, to increase consumers’ financial knowledge. limitations and future research a limitation of this study is related to the sample collection process. the sampling process did not strictly follow the original quota distributions. a more representative study could be done in the future that draws from a larger population of respondents that adheres to a predefined quota distribution, thus making the dataset nationally representative. this will allow researchers to better examine the associations between the social determinants of health and financial behaviors. another related limitation of this study is that the dataset was cross-sectional. a future study using a longitudinal or panel dataset is necessary to investigate if the relationships between sdh and financial behaviors remain consistent over time. similarly, since the dataset qi et al. 71 of this study was collected during the covid-19 pandemic, it would be interesting to test the associations after the pandemic with a representative sample. this would allow researchers to understand whether the implications for consumers, financial service practitioners, and policymakers are valid during normal financial situations. finally, as noted earlier, research is needed to investigate how the neighborhood and built environment factor is related to financial behavior. in this study, the neighborhood and built environment variable was the only sdh factor not significantly associated with desirable financial behaviors. as noted in the methods section, neighborhood and built environment was measured using one question about how worried respondents were about relying on government safety nets during times of crisis. this question may be measuring trust in government rather than neighborhood and built environment. future studies should use revised questions related to the neighborhood and built environment so that the relationship between this factor and desirable financial behaviors can be better tested. references abdel‐ghany, m., & wang, m. q. (2001). factors associated with different degrees of health insurance coverage. family and consumer sciences research journal, 29(3), 252-264. https://doi.org/10.1177/1077727x01293 003 abdullah, a., ismail, w. a. a. z. w., muhammad, m. z., ikhwan, m., aziz, m. n. z., kamarudin, m. k., & shamma, s. h. (2023). financial literacy on personal finance stability among low-income communities. advanced international journal of business, entrepreneurship and smes, 5(18), 311-325. https://doi.org/10.35631/aijbes.51802 5 alwin, d. f., & hauser, r. m. (1975). the decomposition of effects in path analysis. american sociological review, 37-47. https://www.jstor.org/stable/2094445 anand, s., mishra, k., verma, v., & taruna, t. (2020). financial literacy as a mediator of personal financial health during covid-19: a structural equation modelling approach. emerald open research, 2. https://doi.org/10.35241/emeraldopenres .13735.2 bandura, a. (1986). social foundations of thought and action. englewood cliffs, nj, 1986(23-28). bandura, a. (1991). social cognitive theory of self-regulation. organizational behavior and human decision processes, 50(2), 248-287. https://doi.org/10.1016/07495978(91)90022l bandura, a., & walters, r. h. (1977). social learning theory (vol. 1). prentice hall: englewood cliffs. baron, r. m., & kenny, d. a. (1986). the moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. journal of personality and social psychology, 51(6), 1173. https://doi.org/10.1037/00223514.51.6.1173 barrafrem, k., tinghög, g., & västfjäll, d. (2021). trust in the government increases financial well-being and general well-being during covid19. journal of behavioral and experimental finance, 31, 100514. https://doi.org/10.1016/j.jbef.2021.1005 14 batty, m., gibbs, c., & ippolito, b. (2022). health insurance, medical debt, and financial well-being. health economics, 31(5), 689–728. https://doi.org/10.1002/hec.4472 https://doi.org/10.35241/emeraldopenres.13735.2 https://doi.org/10.35241/emeraldopenres.13735.2 https://doi.org/10.1016/j.jbef.2021.100514 https://doi.org/10.1016/j.jbef.2021.100514 https://doi.org/10.1002/hec.4472 financial services review, 32(4) 72 baumeister, r. f., campbell, j. d., krueger, j. i., & vohs, k. d. (2003). does high selfesteem cause better performance, interpersonal success, happiness, or healthier lifestyles?. psychological science in the public interest, 4(1), 1-44. https://doi.org/10.1111/15291006.01431 bautista, m. a., gonzález, f., martínez, l. r., muñoz, p., & prem, m. (2021). does higher education reduce mortality? evidence from a natural experiment. documentos de trabajo. https://ideas.repec.org/p/col/000092/018 486.html bongomin, g. o. o., ntayi, j. m., & malinga, c. a. (2020). analyzing the relationship between financial literacy and financial inclusion by microfinance banks in developing countries: social network theoretical approach. international journal of sociology and social policy, 40(11/12), 1257-1277. https://doi.org/10.1108/ijssp-12-20190262 braveman p., egerter s., barclay c. (2011) what shapes health-related behaviors? the role of social factors. robert wood johnson foundation. https://nmpha.wildapricot.org/resources /documents/rwjf%20issue%20brief% 20-%20the%20role%20of%20social% 20factors.pdf braveman, p., & gottlieb, l. (2014). the social determinants of health: it’s time to consider the causes of the causes. public health reports, 129(1), 19-31. https://doi.org/10.1177/0033354914129 1s20 brevoort, k., grodzicki, d., & hackmann, m. (2017). medicaid and financial health. national bureau of economic research. http://www.nber.org/papers/w24002 buheji, m., da costa cunha, k., beka, g., mavric, b., de souza, y. l., da costa silva, s. s., … & yein, t. c. (2020). the extent of covid-19 pandemic socio-economic impact on global poverty. a global integrative multidisciplinary review. american journal of economics, 10(4), 213-224. https://doi.org/10.5923/j.economics.202 01004.02. cantor, g., & landry, s. (2020). how are the most vulnerable households navigating the financial impact of covid-19. prosperity now. https://prosperitynow.org/resources/how -are-most-vulnerable-householdsnavigating-financial-impact-covid-19 chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7(2), 107-128. https://doi.org/10.1016/s10570810(99)80006-7 cole, s., paulson, a., & shastry, g. k. (2014). smart money? the effect of education on financial outcomes. the review of financial studies, 27(7), 2022-2051. commission on social determinants of health. (2008). closing the gap in a generation: health equity through action on the social determinants of health: final report of the commission on social determinants of health. world health organization. https://apps.who.int/iris/bitstream/handl e/10665/43943/97892?sequence=1 costa‐font, j. and vilaplana-prieto, c. (2017). does the expansion of public long‐term care funding affect saving behaviour? fiscal studies, 38(3), 417-443. https://doi.org/10.1111/j.14755890.2017.12139 https://doi.org/10.1111/1529-1006.01431 https://doi.org/10.1111/1529-1006.01431 https://ideas.repec.org/p/col/000092/018486.html https://ideas.repec.org/p/col/000092/018486.html https://doi.org/10.1108/ijssp-12-2019-0262 https://doi.org/10.1108/ijssp-12-2019-0262 https://nmpha.wildapricot.org/resources/documents/rwjf%20issue%20brief%20-%20the%20role%20of%20social%20factors.pdf https://nmpha.wildapricot.org/resources/documents/rwjf%20issue%20brief%20-%20the%20role%20of%20social%20factors.pdf https://nmpha.wildapricot.org/resources/documents/rwjf%20issue%20brief%20-%20the%20role%20of%20social%20factors.pdf https://nmpha.wildapricot.org/resources/documents/rwjf%20issue%20brief%20-%20the%20role%20of%20social%20factors.pdf https://doi.org/10.1177/00333549141291s20 https://doi.org/10.1177/00333549141291s20 http://www.nber.org/papers/w24002 https://doi.org/10.5923/j.economics.20201004.02 https://doi.org/10.5923/j.economics.20201004.02 https://prosperitynow.org/resources/how-are-most-vulnerable-households-navigating-financial-impact-covid-19 https://prosperitynow.org/resources/how-are-most-vulnerable-households-navigating-financial-impact-covid-19 https://prosperitynow.org/resources/how-are-most-vulnerable-households-navigating-financial-impact-covid-19 https://apps.who.int/iris/bitstream/handle/10665/43943/97892?sequence=1 https://apps.who.int/iris/bitstream/handle/10665/43943/97892?sequence=1 https://doi.org/10.1111/j.1475-5890.2017.12139 https://doi.org/10.1111/j.1475-5890.2017.12139 qi et al. 73 courchane, m., & zorn, p. (2005). consumer literacy and creditworthiness. proceedings, federal reserve bank of chicago, 3. https://www.researchgate.net/profile/ma rshacourchane/publication/5042561_consu mer_literacy_and_creditworthiness/links /547dfd7f0cf241bf4b5b9c1b/consumerliteracy-and-creditworthiness.pdf delgadillo, l. m., & law, r. h. (2019). personal finance competencies in the practice of financial counseling. financial counseling, 17-30. https://doi.org/10.1007/978-3-31972586-4_2 dewar, d. m. (1998). do those with more formal education have better health insurance opportunities?. economics of education review, 17(3), 267-277. https://doi.org/10.1016/s02727757(97)00034-4 fletcher, j. m., & frisvold, d. e. (2009). higher education and health investments: does more schooling affect preventive health care use? journal of human capital, 3(2), 144-176. fox, j., & bartholomae, s. (2020). household finances, financial planning, and covid‐19. financial planning review, 3(4), e1103. https://doi.org/10.1002/cfp2.1103 frank, c., davis, c. g., & elgar, f. j. (2014). financial strain, social capital, and perceived health during economic recession: a longitudinal survey in rural canada. anxiety, stress, and coping, 27(4), 422-438. https://doi.org/10.1080/10615806.2013. 864389 gonzález, f., martinez, l., munoz, p., & prem, m (2022). higher education and mortality: evidence from an authoritarian college contraction. http://dx.doi.org/10.2139/ssrn.3717849 gruber, j., & sommers, b. d. (2020). paying for medicaid—state budgets and the case for expansion in the time of coronavirus. new england journal of medicine, 382(24), 2280-2282. https://doi.org/10.1056/nejmp2007124 guariglia, a. and rossi, m. (2002). private medical insurance and saving: evidence from the british household panel survey. ssrn electronic journal. https://doi.org/10.2139/ssrn.320881 gutter, m., garrison, s., & çopur, z. (2010). social learning opportunities and the financial behaviors of college students. family and consumer sciences research journal, 38(4), 387-404. https://doi.org/10.1111/j.15523934.2010.00034.x hergenrather, k. c., zeglin, r. j., mcguirekuletz, m., & rhodes, s. d. (2015). employment as a social determinant of health: a systematic review of longitudinal studies exploring the relationship between employment status and physical health. rehabilitation research, policy, and education, 29(1), 2-26. https://doi.org/10.1891/21686653.29.1.2 holt-lunstad, j. (2022). social connection as a public health issue: the evidence and a systemic framework for prioritizing the “social” in social determinants of health. annual review of public health, 43, 193-213. https://doi.org/10.1146/annurevpublhealth-052020-110732 houle, j.n., & keene, d.e. (2014). getting sick and falling behind: health and the risk of mortgage default and home foreclosure. journal of epidemiology and community health, 69, 382 – 387. https://doi.org/10.1136/jech-2014204637 https://doi.org/10.1016/s0272-7757(97)00034-4 https://doi.org/10.1016/s0272-7757(97)00034-4 https://doi.org/10.1002/cfp2.1103 http://dx.doi.org/10.2139/ssrn.3717849 https://doi.org/10.2139/ssrn.320881 financial services review, 32(4) 74 jamal, a. a. a., ramlan, w. k., karim, m. a., & osman, z. (2015). the effects of social influence and financial literacy on savings behavior: a study on students of higher learning institutions in kota kinabalu, sabah. international journal of business and social science, 6(11), 110-119. https://ijbssnet.com/journals/vol_6_no_ 11_1_november_2015/12.pdf jemal, a., thun, m. j., ward, e. e., henley, s. j., cokkinides, v. e., & murray, t. e. (2008). mortality from leading causes by education and race in the united states, 2001. american journal of preventive medicine, 34(1), 1-8. https://doi.org/10.1016/j.amepre.2007.0 9.017 kim, h. h., maurer, r., & mitchell, o. s. (2018). cognitive ability, financial literacy, and the demand for financial advice at older ages: findings from the health and retirement study. wharton pension research council working papers. 22. https://core.ac.uk/download/pdf/219380 756.pdf laffargue, j. and padieu, y. (2016). insurance of household risks and the rebalancing of the chinese economy: health insurance, health expenses and household savings. pacific economic review, 21(3), 381412. https://doi.org/10.1111/14680106.12179 lebaron, a. and kelley, h. (2020). financial socialization: a decade in review. journal of family and economic issues, 42(s1), 195-206. https://doi.org/10.1007/s10834-02009736-2 lee, y. g., hales, e., & kelley, h. h. (2023). financial behaviors, government assistance, and financial satisfaction. social indicators research, 166(1), 85–103. https://doi.org/10.1007/s11205-02203051-z li, y., mutchler, j. e., miller, e. a., xiao, j. j., & tucker-seeley, r. (2022). space, context, and human capital: a micro– macro perspective on the social environment and financial literacy in later life. population research and policy review, 41(3), 1385-1404. https://doi.org/10.1007/s11113-02109695-y li, z., liu, l., shi, j. & sui, y. (2021). health insurance, risk attitudes, and household financial behavior. health economics, 30, 1239-1246. https://doi.org/10.1002/hec.4235 liu, y., zhang, y., & chatterjee, s. (2023). financial hardship and depression experienced by pre-retirees during the covid-19 pandemic: the mitigating role of stimulus payments. applied economics letters, 30(3), 391-396. lusardi, a., & mitchell, o. s. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54(1), 205-224. https://doi.org/10.1016/j.jmoneco.2006. 12.001 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52(1), 5-44. https://doi.org/10.1257/jel.52.1.5 ma, l., sun, f., & tang, z. (2018). social frailty is associated with physical functioning, cognition, and depression, and predicts mortality. the journal of nutrition, health and aging, 22, 989-995. https://doi.org/10.1007/s12603-0181054-0 mackenbach, j. (1996). the contribution of medical care to mortality decline: mckeown revisited. journal of clinical epidemiology, 49(11), 1207-1213. https://doi.org/10.1016/s08954356(96)00200-4 https://doi.org/10.1016/j.amepre.2007.09.017 https://doi.org/10.1016/j.amepre.2007.09.017 https://doi.org/10.1111/1468-0106.12179 https://doi.org/10.1111/1468-0106.12179 https://doi.org/10.1007/s10834-020-09736-2 https://doi.org/10.1007/s10834-020-09736-2 https://doi.org/10.1007/s11205-022-03051-z https://doi.org/10.1007/s11205-022-03051-z https://doi.org/10.1002/hec.4235 https://doi.org/10.1016/j.jmoneco.2006.12.001 https://doi.org/10.1016/j.jmoneco.2006.12.001 https://doi.org/10.1257/jel.52.1.5 https://doi.org/10.1007/s12603-018-1054-0 https://doi.org/10.1007/s12603-018-1054-0 https://doi.org/10.1016/s0895-4356(96)00200-4 https://doi.org/10.1016/s0895-4356(96)00200-4 qi et al. 75 mackinnon, d. p., fairchild, a. j., & fritz, m. s. (2007). mediation analysis. annual review psychology, 58, 593-614. https://doi.org/10.1146/annurev.psych.5 8.110405.085542 mackinnon, d. p., krull, j. l., & lockwood, c. m. (2000). equivalence of the mediation, confounding and suppression effect. prevention science, 1, 173-181. https://doi.org/10.1023/a:10265950113 71 mcginnis, j. m., williams-russo, p., & knickman, j. r. (2002). the case for more active policy attention to health promotion. health affairs, 21(2), 78-93. https://doi.org/10.1377/hlthaff.21.2.78 mitchell, o. s., & lusardi, a. (2022). financial literacy and financial behavior at older ages. wharton pension research council working paper, (2022-01). http://dx.doi.org/10.2139/ssrn.4006687 mueller, m., suzuki, e., paolantonio, g., hewlett, e., & james, c. (2021). the health impact of covid-19. oece library. https://www.oecdilibrary.org/sites/b0118faeen/index.html?itemid=/content/compone nt/b0118fae-en naito, r., mckee, m., leong, d., bangdiwala, s., rangarajan, s., islam, s., & yusuf, s. (2023). social isolation as a risk factor for all-cause mortality: systematic review and meta-analysis of cohort studies. plos one, 18(1). https://doi.org/10.1371/journal.pone.028 0308 niţoi, m., & pochea, m. m. (2024). trust in the central bank, financial literacy, and personal beliefs. journal of international money and finance, 103066. https://doi.org/10.1016/j.jimonfin.2024. 103066 o’connor, g. e., & kabadayi, s. (2020). examining antecedents of health insurance literacy: the role of locus of control, cognitive style, and financial knowledge. journal of consumer affairs, 54(1), 227-260. https://doi.org/10.1111/joca.12266 office of disease prevention and health promotion. (2023). u.s. department of health and human services. https://health.gov/healthypeople/objectiv es-and-data/social-determinants-health olsen, l., saunders, r. s., & yong, p. l. (eds.). (2010). the healthcare imperative: lowering costs and improving outcomes: workshop series summary. https://doi.org/10.17226/12750 oxford dictionary. (n.d.). knowledge. https://www.oxforddictionaries.com/ parker, k., horowitz, j. m., & brown, a. (2020). about half of lower-income americans report household job or wage loss due to covid-19. policy commons. https://policycommons.net/artifacts/616 246/about-half-of-lower-incomeamericans-report-household-job-orwage-loss-due-to-covid-19/1596872/ ragusa, f. s., veronese, n., smith, l., koyanagi, a., dominguez, l. j., & barbagallo, m. (2022). social frailty increases the risk of all-cause mortality: a longitudinal analysis of the english longitudinal study of ageing. experimental gerontology, 111901. https://doi.org/10.1016/j.exger.2022.111 901 reagans, r., & mcevily, b. (2003). network structure and knowledge transfer: the effects of cohesion and range. administrative science quarterly, 48(2), 240-267. https://doi.org/10.2307/3556658 https://doi.org/10.1146/annurev.psych.58.110405.085542 https://doi.org/10.1146/annurev.psych.58.110405.085542 https://doi.org/10.1023/a:1026595011371 https://doi.org/10.1023/a:1026595011371 https://doi.org/10.1377/hlthaff.21.2.78 https://dx.doi.org/10.2139/ssrn.4006687 https://www.oecd-ilibrary.org/sites/b0118fae-en/index.html?itemid=/content/component/b0118fae-en https://www.oecd-ilibrary.org/sites/b0118fae-en/index.html?itemid=/content/component/b0118fae-en https://www.oecd-ilibrary.org/sites/b0118fae-en/index.html?itemid=/content/component/b0118fae-en https://www.oecd-ilibrary.org/sites/b0118fae-en/index.html?itemid=/content/component/b0118fae-en https://doi.org/10.1371/journal.pone.0280308 https://doi.org/10.1371/journal.pone.0280308 https://doi.org/10.1016/j.jimonfin.2024.103066 https://doi.org/10.1016/j.jimonfin.2024.103066 https://doi.org/10.1111/joca.12266 https://health.gov/healthypeople/objectives-and-data/social-determinants-health https://health.gov/healthypeople/objectives-and-data/social-determinants-health https://doi.org/10.17226/12750 https://policycommons.net/artifacts/616246/about-half-of-lower-income-americans-report-household-job-or-wage-loss-due-to-covid-19/1596872/ https://policycommons.net/artifacts/616246/about-half-of-lower-income-americans-report-household-job-or-wage-loss-due-to-covid-19/1596872/ https://policycommons.net/artifacts/616246/about-half-of-lower-income-americans-report-household-job-or-wage-loss-due-to-covid-19/1596872/ https://policycommons.net/artifacts/616246/about-half-of-lower-income-americans-report-household-job-or-wage-loss-due-to-covid-19/1596872/ https://doi.org/10.1016/j.exger.2022.111901 https://doi.org/10.1016/j.exger.2022.111901 https://doi.org/10.2307/3556658 financial services review, 32(4) 76 richards, m., power, c., & sacker, a. (2009). paths to literacy and numeracy problems: evidence from two british birth cohorts. journal of epidemiology and community health, 63(3), 239-244. https://doi.org/10.1136/jech.2007.06492 3 robb, c. a., & woodyard, a. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning, 22(1). https://ssrn.com/abstract=2061308 rogowski, j., & karoly, l. (2000). health insurance and retirement behavior: evidence from the health and retirement survey. journal of health economics, 19(4), 529–539. https://doi.org/10.1016/s01676296(00)00038-2 rostamkalaei, a., nitani, m., & riding, a. (2022). self-employment, financial knowledge, and retirement planning. journal of small business management, 60(1), 63-92. rothwell, d. w., & wu, s. (2019). exploring the relationship between financial education and financial knowledge and efficacy: evidence from the canadian financial capability survey. journal of consumer affairs, 53(4), 1725-1747. https://doi.org/10.1111/joca.12259 ryu, s., & fan, l. (2023). the relationship between financial worries and psychological distress among u.s. adults. journal of family and economic issues, 44(1), 16–33. https://doi.org/10.1007/s10834-02209820-9 shek, d. t. (2021). covid-19 and quality of life: twelve reflections. applied research in quality of life, 16, 1-11. https://doi.org/10.1007/s11482-02009898-z sherraden, m., johnson, l., clancy, m., beverly, s, schreiner, m., zhan, m. & curley, j. (2000). saving patterns in ida programs. washington university in st. louis: center for social development. social policy institute. (2022, september). how do health care costs impact household finances and access to care? https://socialpolicyinstitute.wustl.edu/ho w-healthcare-costs-impact-householdfinances-and-access-to-care/ statacorp. (2023). stata 18 causal inference and treatment-effects estimation reference manual [press release]. https://www.stata.com/manuals/causal.p df tang, n., & baker, a. (2016). self-esteem, financial knowledge and financial behavior. journal of economic psychology, 54, 164-176. https://doi.org/10.1016/j.joep.2016.04.0 05 thomas, m. g. (2019). the impact of social capital, financial knowledge, skill, and attitudes on financial well-being across federal poverty level status (doctoral dissertation, university of georgia). https://getd.libs.uga.edu/pdfs/thomas_mi chael_g_201908_phd.pdf umberson, d., & karas montez, j. (2010). social relationships and health: a flashpoint for health policy. journal of health and social behavior, 51(1), s54-s66. https://doi.org/10.1177/0022146510383 501 uzzi, b. (2018). social structure and competition in interfirm networks: the paradox of embeddedness. in the sociology of economic life (pp. 213-241). routledge. walstad, w. b., rebeck, k., & macdonald, r. a. (2010). the effects of financial education on the financial knowledge of high school students. journal of consumer affairs, 44(2), 336-357. https://doi.org/10.1111/j.17456606.2010.01172.x. https://doi.org/10.1136/jech.2007.064923 https://doi.org/10.1136/jech.2007.064923 https://ssrn.com/abstract=2061308 https://doi.org/10.1016/s0167-6296(00)00038-2 https://doi.org/10.1016/s0167-6296(00)00038-2 https://doi.org/10.1111/joca.12259 https://doi.org/10.1007/s10834-022-09820-9 https://doi.org/10.1007/s10834-022-09820-9 https://doi.org/10.1007/s11482-020-09898-z https://doi.org/10.1007/s11482-020-09898-z https://socialpolicyinstitute.wustl.edu/how-healthcare-costs-impact-household-finances-and-access-to-care/ https://socialpolicyinstitute.wustl.edu/how-healthcare-costs-impact-household-finances-and-access-to-care/ https://socialpolicyinstitute.wustl.edu/how-healthcare-costs-impact-household-finances-and-access-to-care/ https://www.stata.com/manuals/causal.pdf https://www.stata.com/manuals/causal.pdf https://doi.org/10.1016/j.joep.2016.04.005 https://doi.org/10.1016/j.joep.2016.04.005 https://doi.org/10.1177/0022146510383501 https://doi.org/10.1177/0022146510383501 qi et al. 77 wolla, s. a., & sullivan, j. (2017). education, income, and wealth. page one economics. https://www.stlouisfed.org/education/pa ge-one-economics-classroomedition/education-income-and-wealth xiao, j. j. (2008). applying behavior theories to financial behavior. in handbook of consumer finance research (pp. 69-81). new york, ny: springer new york. https://doi.org/10.1007/978-0-38775734-6_5 xiao, j. j. (2016). consumer financial capability and wellbeing. handbook of consumer finance research, 3-17. https://doi.org/10.1007/978-3-31928887-1_1 xiao, j. j., & o'neill, b. (2016). consumer financial education and financial capability. international journal of consumer studies, 40(6), 712-721. https://doi.org/10.1111/ijcs.12285 xiao, j.j., chen, c. & chen, f. consumer financial capability and financial satisfaction. social indicators research, 118, 415–432 (2014). https://doi.org/10.1007/s11205-0130414-8 yuesti, a., rustiarini, n. w., & suryandari, n. n. a. (2020). financial literacy in the covid-19 pandemic: pressure conditions in indonesia. entrepreneurship and sustainability issues, 8(1), 884-898. http://eprints.unmas.ac.id/id/eprint/1065 https://doi.org/10.1007/978-0-387-75734-6_5 https://doi.org/10.1007/978-0-387-75734-6_5 https://doi.org/10.1007/978-3-319-28887-1_1 https://doi.org/10.1007/978-3-319-28887-1_1 https://doi.org/10.1111/ijcs.12285 http://eprints.unmas.ac.id/id/eprint/1065 pii: s1057-0810(99)80005-5 financial services review, 7(2): 95-106 copyright © 1998 by jai press inc. issn: 1057-0810 all fights of reproduction in any form reserved. the international diversification fallacy of exchange-listed securities judson w. russell this paper reviews various investment vehicles and examines their international diver sification potential. the primary focus is on the ability of u.s. exchange-listed investments such as closed-end country funds, american depository receipts (adrs), and multinational corporations (mncs) to provide a diversification effect similar to direct investment in foreign equity. the results show that the u.s. exchange-listed secu rities included in this study behave more like the host exchange than their home exchange. this result suggests that these u.s. exchange-listed securities, on average, do not perform an international diversification role for u.s. investors. i. i n t r o d u c t i o n beginning in the 1980s and throughout the 1990s individual investors have been urged to select international securities as a portion of their investment portfolio. even the relatively risky practice of investing in emerging markets has been viewed, by some, as a sound investment strategy for individuals. while the virtues of international investing are readily extolled in portfolio diversification literature the vehicles to facilitate the investment strat egy need to be addressed. this paper investigates the individual investor's benefits gained by investing in u.s. exchange-listed, "international" securities. specifically, this article tests whether exchange-listed securities such as american depository receipts (adrs), closed-end country funds, and multinational corporations (mncs) behave more like the exchange where they are listed or the market that they purportedly represent. two different approaches exist for investors seeking international portfolio diversifi cation. one approach is to invest directly in the foreign securities. this involves purchasing the securities directly from the exchange where the security is listed. the alternative approach is to indirectly invest in the international security. many investors wishing to acquire international securities have preferred to employ this more familiar, indirect method. this pursuit has led to tremendous growth in country specific investment compa nies (country funds), adrs, and ultimately synthetically created diversification through judson w. russell • sr. research associate, bank of america, 121 west trade street, 1 lth floor, charlotte, nc 28255; phone: (704) 388-1007; fax: (704) 386-9982; e-mail: judson.w.russell @ nationsbank.com. 96 financial services review 7(2) 1998 stock index futures. this paper reviews the international diversification effects from investing in u.s. exchange-listed securities. section ii presents the various investment vehicles available for international diversi fication. section iii contains the data and methodology used in this study. section iv is a synopsis of the results from the model. the final section is a concluding remark regarding investment in u.s. exchange-listed securities. ii. investment vehicles for inter national diversification although this paper focuses on u.s. exchange-listed investment vehicles as a means of international diversification, there are many other vehicles available. the most apparent way to achieve international diversification would be through direct foreign investment. many of the barriers to capital flows have been eliminated through the past few decades and investors are becoming more aware of the "global market." although investors mostly agree about the benefits of international investing, few invest directly. french and poterba (1991) find that there is a very strong home-country bias with investors. this bias cannot be explained with capital controls, tax burdens, or transactions costs. additionally, foreign-exchange risk from investing abroad can be mitigated through forwards or futures contracts on foreign currency. the ability for individual investors to hedge through currency contracts discounts the currency exposure risk. it would appear as though investors choose to invest domestically for the same reasons that consumers borrow from local banks, convenience and familiarity. apart from institutions, most investors appear reluctant to invest directly in a foreign country. mutual funds have become wildly popular with u.s. investors. approximately one in three americans owns shares in a mutual fund. to satisfy the demands of investors, spe cialized funds have developed. one way for investors to participate in overseas markets is through international index funds. peters (1988) suggests that investing internationally through an index fund will provide a hedge against many potential pitfalls. in addition to the potential for impressive returns, indexing allows for low trading costs due to its passive management style. indexing also assists in international stock selection that can prove challenging to some fund managers. buying shares in an international index fund appears to be a low-cost method to achieve diversification. in addition to the numerous mutual funds there are several closed-end (country/region specific) funds. these funds provide a convenient package of securities that could be dupli cated by investing in the foreign stock market directly. the funds are listed on national stock markets and trade as if they were domestic stocks. closed-end funds would appear to be a useful vehicle for a diversification-minded investor. bailey and lim (1992) investi gate the claim that closed-end funds provide international diversification. their study includes 20 country funds listed on the new york and american stock exchanges. they find that country fund returns often resemble domestic u.s. stock returns more than returns from foreign stock portfolios. chang, eun, and kolodny (1995) provide similar results in their study of 15 closed-end country funds. they find closed-end country funds tend to have significantly higher u.s. market betas and somewhat lower local market betas. chang, eun, and kolodny suggest that this tends to reduce the effectiveness of closed-end international diversification fallacy 97 country funds in providing global risk diversification for u.s. investors. barry, peavy, and rodriguez (1997) take a comprehensive look at emerging stock markets and country funds. they find that the country funds listed on the u.s. exchanges are more highly correlated with the s&p 500 than the returns for their respective benchmarks. another vehicle available to investors who wish to add "international" securities to their portfolios is the american depository receipt (adr). adrs are usually issued by united states banks, called depositories, that certify the deposit of a specific number of for eign shares with the bank's overseas branch. these shares are held as long as the adrs remain outstanding. adrs are listed on a national stock exchange or trade over-the counter and can be bought or sold as easily as a domestic security. officer and hoffmeister (1987) find that adrs help simplify investment procedures, reduce costs, and provide diversification for investors. they find that there are no currency exchange risks and that investors can reduce their risk exposure by 20-25% when as few as four adrs are combined with four domestic securities. wahab and khandwala (1993) find similar results as officer and hoffmeister, but they suggest that the optimal number of adrs to hold is at least seven. webb, officer, and boyd (1995) agree that adrs allow investors to circumvent cur rency exchange and other transactions costs, but they suggest that adrs behave very sim ilarly to the domestic market where they trade. their study consists of 74 adrs from 15 countries and estimates the relationship between the adrs and the u.s. equity market with a leading and lagging model. their study indicates that the international diversification benefits from investing in adrs are minimal. this finding is consistent with the findings from several closed-end country funds studies. both of these securities may trade as exchange-listed securities. stock index futures have been suggested as a synthetic international diversification approach. foreign stock index futures have grown rapidly in the last few years. these futures are now traded in at least fifteen countries. stock index futures offer greater liquid ity and lower transaction costs when compared to the cash markets. gastineau, arimura, belkin, clausen, kyokuta, higashino, and mitchinson (1988) estimate that for a large port folio with substantial market impact, the round-trip transaction costs are about 87 basis points (bp) and 9 bp in the u.s. cash and futures markets, respectively. jorion and roisen berg (1993) find that through the use of stock index futures they are able to replicate inter national equity indices. their five country synthetic portfolio was found to be highly correlated with the msci world stock index, a global benchmark. there have been differing views on the usefulness of mncs as a means for interna tional diversification. a mnc is simply a portfolio of internationally diversified cash flows. these cash flows may exhibit low correlation with one another, depending upon the economic cycle in each country. thus, by investing in a mnc an investor should theoreti cally achieve international diversification by acquiring foreign, as well as domestic cash flows. there have been numerous studies regarding the shareholder effects from corporate international diversification. i ll data and m e t h o d o l o g y this paper analyzes the diversification effect of u.s. exchange-listed securities. the data consists of 20 randomly selected closed-end country funds, adrs, and mncs, as well as 98 financial services review 7(2) 1998 table 1 securi t ies used in study country funds first australia fund austria fund brazil fund chile fund france growth fund germany fund india fund* indonesia fund italy fund korea fund malaysia fund mexico fund portugal fund* singapore fund asa, ltd (south africa) spain fund swiss helvetia taiwan fund thai fund united kingdom fund mnc colgate-palmolive dupont merck & company dow chemical general motors trw inc. ibm coca-cola union carbide corp ralston-purina hj heinz 3m ford motor company eastman kodak company general electric company caterpillar inc. deere & company proctor & gamble company johnson & johnson kimberly-clark corp. note: *insufficient data for inclusion in study adrs national australia bank (australia) broken hill properties (australia) barclays plc (uk) shell transport (uk) british airways (uk) bass plc (uk) hanson (uk) banco central hispano (spain) empresa nacional elec. (spain) cia telecom (chile) novo-nordisk a/s (denmark) philippines long distance (philippines) norsk hydro a/s (norway) philips electronics (netherlands) unilever nv (netherlands) kyocera corporation (japan) bank of tokyo (japan) honda motor company (japan) hitachi ltd (japan) benetton group* domestic sierra pacific resources southern company bally entertainment minnesota power & light reliance group quick & reilly group student loan marketing a alex brown inc. fingerhut companies urs corp. russell corp. kansas city southern lands' end inc. cincinnati bell inc. westvaco corp. toll brothers inc. panenergy corp. central louisiana electric inter-regional financial ku energy corp. 20 "pure ly domes t i c " f i rms resul t ing in the 80 f i rms listed in table 1. pure ly domest ic f i rms are def ined as those companies that r ece ive the vast major i ty o f the sales and cash f lows f rom domes t ic sources. all o f the securit ies in table 1 are l isted on the n e w york stock exchange . w e e k l y returns were calcula ted ove r a f ive-year per iod f rom january 1991 to d e c e m b e r 1995 amount ing to 260 observat ions per security. the return data were compi l ed f rom the bloomberg pages. the market chosen as the u.s. benchmark was the n e w york stock exchange c o m p o s i t e index. international diversification fallacy 99 the sample chosen in this study differs from that used in previous work (bailey & lim, 1992; jacquillat & solnik, 1978; wahab & khandwala, 1993) in several respects: (1) weekly data, as opposed to monthly, is used; (2) the time period investigated is more recent; (3) the securities are all new york stock exchange listed; (4) a combined data set including country funds, adrs, mncs, and purely domestic firms is analyzed. twenty securities from each category above were chosen, resulting in 20,800 weekly observations. during the period under investigation the majority of adrs traded in the over-the-counter market and relatively few country funds traded on the new york stock exchange. since the focus of this study is on new york stock exchange-listed securities a smaller set of securities is available from which to choose. a natural extension of this anal ysis would be to compare exchange-listed and over-the-counter "international" securities. despite the smaller set of securities available for this study the sample size compares favorably with previous published studies. barry, peavy, and rodriguez (1997) and akdogan (1995) suggest that u.s. and emerging market foreign indices are less positively correlated today than they were in the recent past. akdogan finds that there is integration in world capital markets, but that much of this integration takes place regionally. espitia and santamaria (1994) conclude that a high level of correlation exists between daily return time series for all the european mar kets. byers and peel (1993) suggest that there is no convincing evidence that international stock markets were cointegrated in the period from 1979-1989. this study presents the correlation coefficients for various indices with the new york composite index in the tables, but it does not attempt to test a cointegration hypothesis. although it is possible that the results reflect the integration of global markets during the sample period, 1991-1995, cointegration analysis is not explicitly employed. the model employed in this paper is similar to the johnson, schneeweis, dinning (1993) model. the weekly return for dollar-denominated securities (indices) was derived from: r = in (pt / p t 1) * 100 (1) where r = weekly return pt = the price of the security at time t, in dollars pti = the price of the security at time t 1, in dollars the dividend component of total return has been eliminated. the dividend yields across the categories of investment vehicles were very similar. this treatment of dividends is comparable to other studies such as officer and hoffmeister (1987). the weekly returns for foreign currency denominated indices were derived from: r = [in (pt / pt 1) in (s t i s t_ 1)] * 100 (2) where r = weekly return pt = the price of the security at time t, in foreign currency pt l = the price of the security at time t 1, in foreign currency s t = the spot exchange rate against the dollar at time t 100 financial services review 7(2) 1998 st 1 = the spot exchange rate against the dollar at t ime t 1 after calculating the return on all securities and indices in dollar terms this study pre sents the regression analysis. to isolate the pure diversification effect in this study the for eign index returns, in dollars, were first regressed on the new york composite index return. the residuals from these regressions were the portion of the foreign index return not explained, or related to, the new york composite index. the two-factor model employed to illustrate each securi ty 's diversification effect was: ri, t = ~i + ~r,i gr, t + ~n,i gn,t + ~i,t (3) where ri, t rr, t gtl, t = dollar-denominated return on security i at t ime t = residual from a regression of the dollar-denominated foreign stock index for security i on the new york composite index returns = dollar-denominated return on the new york composite index. using this model the correlation effect between foreign indices and the new york composite index was removed, i.e., the returns are orthogonal. this approach isolates the diversification benefits for the individual security. this method is used to compare the security return with its home index "pure" return and the new york composite index. pure return is defined as the return unexplained by the new york composite index or the true return of the index if held in isolation from u.s. market influence. to substantiate the results from the orthogonal, multifactor model above, simple regressions were also analyzed of the form: ri, t = o~i + ~f i r f t + o~i,t (4) where gi, t r:, and = dollar-denominated return on security i at time t = dollar-denominated return on the foreign stock index corresponding to secu rity i ' s "home" country ri, t = o~i + ~n,i rn, t + ~i,t (5) where ri, t gn, f = dollar-denominated return on security i at t ime t = dol lar-denominated return on the new york composite index. this study is interested in testing the null hypothesis that the beta coefficient on the independent variable is one versus the alternative that the beta coefficient is not one, i.e., ho: ~fi = 1 and h0: ~n,i = 1 hi: ~f, i# l a n d h l : ~ n , i * l t a b l e 2 c ou nt ry f un d r et ur n r es ul ts , j an ua ry 1 99 1d ec em be r 1 99 5, w ee kl y d at a u .s . d ol la r m ul ti fa ct or b et a si ng le fa ct or b et a w it h n ew c or re la ti on w it h b et a w it h h om e yo rk c om po si te c oe ff ic ie nt c ou nt ry f un ds h om e in de x in de x in de x w / i nd ex fi rs t a us tr al ia f un d 0. 33 2 (6. 60 )* 0. 45 0 (5. 57 )* 0. 69 2 (2. 23 )* * 0. 31 3 a us tr ia f un d 0. 00 8 (69 .5 9) * 0. 01 3 (73 .2 1) * 0. 82 3 (0. 62 ) 0. 13 0 b ra zi l 0. 37 9 (18 .5 5) * 0. 38 8 (18 .6 8) * -0 .9 83 (0. 62 ) -0 .2 11 c hi le f un d 0. 74 3 (3. 13 )* 0. 77 8 (2. 58 )* l. l lo (0 .5 9) 0. 08 1 fr an ce g ro w th f un d 0. 44 6 (7. 45 )* 0. 51 0 (6. 81 )* 0. 73 0 (1. 81 )* * 0. 30 3 g er m an y fu nd 0. 48 1 (6. 12 )* 0. 58 9 (4. 85 )* 1. 04 0 (0 .2 5) 0. 27 5 in do ne si a fu nd 0. 61 2 (3. 07 )* 0. 60 4 (3. 11 )* 0. 65 3 (1. 24 ) -0 .0 25 it al y fu nd 0. 51 7 (5. 55 )* 0. 52 1 (5. 53 )* 0. 26 7 (2. 58 ) 0. 09 2 k or ea f un d 0. 24 0 (13 .5 9) * 0. 24 8 (13 .3 7) * 0. 51 9 (2. 25 )* * 0. 06 5 m al ay si a fu nd 0. 60 7 (4. 34 )* 0. 70 0 (3. 18 )* 1. 30 0 ( 1. 62 )* ** 0. 18 1 m ex ic o fu nd 0. 04 5 (62 .4 5) * 0. 05 2 (59 .0 7) * 1. 39 0 ( 1. 54 )* ** 0. 08 2 si ng ap or e fu nd 1. 13 0 (0 .4 0) 1. 16 0 (0 .5 1 ) 0. 62 7 (0. 83 ) 0. 28 6 a s a l td 0. 60 8 (4. 27 )* 0. 60 0 (4. 35 )* -0 .1 12 (6. 85 )* 0. 09 3 sp ai n fu nd 0. 43 3 (6. 53 )* 0. 53 7 (5. 52 )* 0. 94 8 (0. 31 ) 0. 33 0 sw is s h el ve ti a 0. 22 8 (8. 74 )* 0. 32 4 (7. 84 )* 0. 67 2 (2. 23 )* * 0. 30 2 t ai w an f un d 0. 47 0 (8. 54 )* 0. 48 8 (8. 09 )* 0. 81 3 (0. 93 ) 0. 08 1 t ha i fu nd 0. 56 9 (6. 43 )* 0. 62 9 (5. 43 )* 1. 09 0 (0 .5 3) 0. 19 3 u ni te d k in gd om f un d 0. 23 1 (10 .7 9) * 0. 34 1 (9. 22 )* 0. 91 9 (0. 57 ) 0. 29 5 n ot es : tst at is tic s i n pa re nt he se s *s ig ni fic an tly di ff er en t f ro m o ne a t a = 0 .0 1 le ve l ** si gn if ic an tly di ff er en t f ro m on e at c t = 0 .0 5 le ve l ** *s ig ni fi ca nt ly di ff er en t f ro m o ne a t a = 0 .1 0 le ve l 102 financial services review 7(2) 1998 this study investigates the hypothesis of whether these exchange-listed securities appear to mimic their home country index or the new york composite index more closely. a beta of one would indicate that the security moves exactly with the market on average. a beta of one with the foreign index would indicate significant diversification possibilities are present through trading in these exchange-listed securities. a beta of one with the new york composite index would indicate that the security behaves as the host index and inter national diversification is limited to the covariance between the u.s. and foreign index since the two indices are not orthogonal in practice. the level of significance of the slope coefficients is investigated by performing t-tests. small t-values indicate that the slope coefficient is not significantly different from one, i.e., fall to reject the null hypothesis that the beta is equal to one. this would signal that the security returns are similar to the index returns. large values of the t-statistic indicate that the slope coefficient is significantly different from one, i.e., reject the null hypothesis that the beta is equal to one. that is, the security does not respond closely to the index. iv. results table 2 shows the results for the closed-end funds in this study. the beta coefficients of the funds included in this study suggest that the funds behave more like the new york com posite index than their respective indices overall. the only country fund that demonstrated return patterns indicative of the home country was the singapore fund. in each case, both the multifactor and single factor models produced results leading to the same decision in regards to the hypothesis. of the 18 funds in this study for which there were sufficient data, 13 lead to a conclusion that the return patterns are not statistically different from the returns to the new york composite index, i.e., the beta is relatively close to one with the market at standard levels of confidence. the results are robust and suggest that closed-end fund returns, from funds listed on the new york stock exchange, are significantly linked to the u.s. market. similar to bailey and lim (1992), the findings from this study cast doubts upon the ability of closed-end funds to provide a diversification effect similar to investing in the for eign market. since a closed-end fund can trade at a discount or premium to its net asset value (nav) the host market, new york stock exchange, appears to have a significant impact on fund returns, as it would a purely domestic common stock. the relatively low correlation coefficients for returns on the new york composite index and the foreign market indices indicate that diversification benefits may be present by combining securities from these foreign markets with u.s. securities. closed-end funds appear to lack this diversification benefit for u.s. investors. barry, peavy, and rodriguez (1997) find that 18 of their 20 emerging market country funds were more highly correlated with the s&p 500 than with their respective bench marks. they conclude that emerging market closed-end country funds did not provide as substantial diversification benefits as would have direct investments in the securities underlying the country funds. table 3 shows the adr results. many of the adrs included in this study have a beta coefficient that more strongly mimics the new york composite index. the adrs trade on t a b l e 3 a d r r et ur n r es ul ts , ja nu ar y 19 91 -d ec em be r 19 95 , w ee kl y d at a m al ti fa ct or b et a si ng le fa ct or b et a a d r w it h h om e in de x w it h h om e in de x n at io na l a us tr al ia b k 0. 36 5 (6. 38 )* 0. 43 3 (5. 96 )* b ro ke n h ill p ro pe rt ie s 0. 63 6 (4. 29 )* 0. 71 2 (3. 52 )* b ar cl ay s pl c 0. 12 4 (10 .1 0) * 0. 22 8 (9. 1 i) * sh el l t ra ns po rt -0 .0 56 (20 .0 4) * 0. 03 4 ( 1 8. 18 )* b ri ti sh a ir w ay s 0. 07 3 (9. 48 )* 0. 18 6 (8. 49 )* b as s pi c 0. 03 7 (11 .8 7) * 0. 07 4 (11 .7 6) * h an so n 0. 08 8 (14 .1 3) * 0. 21 6 (11 .8 0) * b an co c en tr al h is p. 0. 52 1 (6. 85 )* 0. 56 1 (6. 61 )* e m pr es a n ac io na l e l. 0. 82 0 (2. 96 )* 0. 83 9 (2. 80 )* c ia t el ec om 0. 78 5 (2. 68 )* 0. 80 9 (2. 33 )* n ov on or di sk a /s 0. 48 5 (10 .1 7) * 0. 50 0 (10 .0 8) * ph il ip pi ne s l on g d is t 0. 74 0 (3. 79 )* 0. 76 2 (3. 31 )* n or sk h yd ro a /s 0. 06 0 (39 .4 8) * 0. 07 4 (36 .7 9) * ph il ip s e le ct ro ni cs 1. 21 0 (1 .4 7) ** * 1. 25 4 (1 .8 9) ** u ni le ve r n v 0. 76 5 (2. 45 )* 0. 83 3 ( 1 .8 4) ** k yo ce ra c or p 0. 01 4 (85 .1 3) * 0. 01 6 (81 .8 0) * b an k of t ok yo 0. 00 5 (77 .9 0) * 0. 00 6 (81 .1 8) * h on da m ot or c om p. 0. 02 0 (80 .5 3) * 0. 02 2 (77 .8 0) * h it ac hi l td 0. 01 7 (10 2. 21 )* 0. 02 0 (96 .5 3) * n ot es : tst at is tic s i n pa re nt he se s *s ig ni fic an tly di ff er en t f ro m on e at a f fi 0 .0 1 le ve l ** si gn if ic an tly di ff er en t f ro m on e at a = 0 .0 5 le ve l ** *s ig ni fi ca nt ly di ff er en t f ro m on e at a = 0 .1 0 le ve l u .s . d ol la r c or re la tio n b et a w it h n ew c oe ff ic ie nt yo rk c om po si te i nd ex w / i nd ex -0 .4 75 (10 .9 0) * 0. 31 3 -0 .6 31 (14 .1 1) * 0. 31 3 0. 81 1 (1. 10 ) 0. 29 5 0. 60 2 (3. 79 )* 0. 29 5 0. 83 5 (0. 85 ) 0. 29 5 0. 28 0 (4. 42 )* 0. 29 5 0. 95 2 (0. 38 ) 0. 29 5 0. 60 4 (2. 89 )* 0. 33 0. 68 2 (2. 67 )* 0. 33 0. 82 3 (0. 98 ) 0. 08 i 0. 38 5 (5. 54 )* 0. 21 2 1. 15 0 (0 .7 2) 0. 06 7 1. 03 0 (0 .1 7) 0. 10 1 0. 73 5 (1. 43 )* ** 0. 34 4 0. 63 2 (2. 98 )* 0. 34 4 0. 57 3 (2. 31 )* * 0. 05 0. 55 (2. 25 )* * 0. 05 0. 75 1 ( 1 .2 8) 0. 05 0. 71 9 (1. 83 )* * 0. 05 104 financial services review 7(2) 1998 t a b l e 4 mnc return results, january 1991-december 1995, weekly data beta with new york mnc eafe beta all beta composite index colgate-palmolive 0.090 (-9.10)* 0.131 (-5.31)* 1.04 (0.31) dupont 0.063 (-9.82)* 0.114 (-5.67)* 1.24 (1.96)** merck & co 0.004 (-4.98)* 0.085 (-2.26)** 0.923 (-0.25) dow chemical 0.081 (-8.74)* 0.151 (-4.89)* 0.954 (-0.34) general motors 0.020 (-6.86)* 0.051 (-3.91)* 1.43 (2.29)** trw inc 0.014 (-9.86)* 0.019 (-6.20)* 1.04 (0.32) ibm 0.114 (-6.37)* 0.244 (-3.35)* 0.837 (-0.92) coca-cola 0.075 (-9.74)* 0.14 (-5.53)* 0.974 (-0.21) union carbide corp -0.043 (-8.00)* -0.063 (-4.89)* 1.34 (2.01)** ralston-purina -0.151 (-10.60)* -0.261 (-7.10)* 0.696 (-2.18)** hj heinz -0.136 (-11.44)* -0.225 (-7.57)* 0.761 (-1.88)** 3m 0.005 (-13.93)* 0.025 (-7.80)* 0.948 (-0.52) ford motor company 0.029 (-7.03)* 0.036 (-4.28)* 1.45 (2.52)* eastman kodak co. -0.023 (-8.90)* -0.023 (-5.34)* 0.698 (-2.04)** general electric co. -0.063 (-14.00)* -0.099 (-8.77)* 1.27 (2.76)* caterpillar inc 0.058 (-6.98)* 0.092 (-4.15)* 1.12 (0.69) deere & company -0.017 (-8.97)* -0.03 (-5.15)* 1.3 (1.97)** proctor & gamble co. -0.157 (13.12)* -0.244 (-8.62)* 1.09 (0.79) johnson & johnson -0.270 (1.18) -0.006 (-6.71)* 1.07 (0.5 i) kimberly-clark corp -0.099 (-i 1.43)* -0.17 (-7.43)* 1.02 (0.16) notes: t-statistics in parentheses *significantly different from one at a = 0.01 level **significantly different from one at 0t = 0.05 level ***significantly different from one at ct = 0.10 level the new york stock exchange and appear to resemble the composite index more closely than their respective home indices. only philips electronics offers a chance to obtain results similar to the foreign market. the hypothesis that the beta coefficient is equal to one with the new york composite index cannot be rejected for most of the adrs in this study. this suggests a relatively strong link to the u.s. index. the adrs from the united kingdom and japan display an especially robust relationship with the new york composite index and very little relation with their respective home indices. the results indicate that adrs may not be good tools for international diversification. this confirms the findings of webb, officer, and boyd (1995). there appears to be minimal international diversification benefits from investing in either closed-end country funds or adrs. rather, these securities typically display a sig nificant relationship with the new york composite index. table 4 shows the results from the mnc returns. the mncs were regressed using the morgan stanley europe, asia, and far east (eafe) index and the morgan stanley all countries (all) index as the home index. each of these indices was presented in dollar returns. the mnc beta coefficients are very indicative of the new york composite index. of the 20 mncs in this study, 13 display beta coefficients that lead to the conclusion that they behave as the new york composite index while none of the mncs display "multina tional" diversification benefits. there is little indication to suggest that any international diversification benefit is present when investing in mncs. international diversification fallacy 105 t a b l e 5 domestic firm return results, january 1991-december 1995, weekly data beta with new york composite firm index sierra pacific resources 0.452 southern company 0.504 bally entertainment 1.18 minnesota power & light 0.467 reliance group 1.36 quick & reilly group 2.07 student loan marketing a 1.35 alex brown inc 2.38 fingerhut companies 1.31 urs corporation 0.62 russell corporation 0.924 kansas city southern 1.23 lands' end inc 0.756 cincinnati bell inc 0.66 westvaco corp 1.19 toll brothers inc 2.02 panenergy corp 1.19 central louisiana electric 0.524 inter-regional financial 1.71 ku energy corp 0.533 this supports the findings o f jacquillat and solnik (1978). they find that mncs dis play a high level of correlation with their respective stock market index. an american investor wishing to purchase a u.s. mnc to gain international diversification would merely be buying a stock that virtually mimics the u.s. stock market. this investor would receive much more diversification by purchasing shares of a foreign mnc, such as a bel gian mnc. this study confirms the ineffectiveness of the home-based mnc to deliver substantial diversification benefits. table 5 shows the results from the purely domestic firms. the overall beta coefficient is 1.12. the beta coefficients from the purely domestic f irm returns are indeed more varied than the mncs. the range (difference between the highest and the lowest) of betas for the domestic firm is 1.93, while the range for mncs is 0.75. however, overall the purely domestic f irm is very similar to the mnc based upon the average beta coefficient with the new york composi te index. it is possible that the results may be driven by firm size or the nature of the domestic f i rm's industries. this analysis indicates that domestic firms, on average, behave as the domestic market and that mncs also behave more like the domestic market than world indices. this provides further evidence that the mnc does not provide significant international diversification benefits. v. c o n c l u s i o n there are many investment vehicles available for the individual u.s. investor seeking international diversification. this paper discusses the most common methods and some of 106 financial services review 7(2) 1998 the voluminous literature on international diversification. global markets have become more integrated over the past few decades. this integration has led many u.s. investors to demand foreign securities to aid in portfolio diversification. there is still debate regarding the use of u.s. exchange-listed securities as a means for portfolio diversification. the results from this paper and other studies indicate that it appears as though these securities are more indicative of the exchange where they trade rather than the index of the country where the cash flows may be generated. this paper sug gests that exchange-traded "international" securities are not the ideal vehicles for diversi fication. references akdogan, h. (1995). the integration of international capital markets: theory and empirical evi dence. aldershot and brookfield, vt. bailey, w., & lim, j. (1992). evaluating the diversification benefits of the new country funds. jour nal of portfolio management, spring, 75-80. barry, c. b., peavy, j. w., iii. ,& rodriguez, m. (1997). emerging stock markets: risks, return, and performance. research foundation of the institute of chartered financial analysts mimeo graph, charlottesville, va. the bloomberg, bloomberg l.p. byers, j. d., & peel, d. a. (1993). some evidence on the interdependence of national stock markets and the gains from international portfolio diversification. applied financial economics, 3, 239-242. chang, e., eun, c. s., & kolodny, r. (1995). international diversification through closed-end coun try funds. journal of banking and finance, 19, 1237-1263. espitia, m., & santamaria, r. (1994). international diversification among the capital markets of the eec. applied financial economics, 4, 1-10. french, k., & poterba, j. (1991). investor diversification and international equity markets. american economic review, may, 222-226. gastineau, g. l., arimura, m. m., belkin, m. j., clausen, e. a., kyokuta, t., higashino, t., & mitchinson, c. m. (1988). japanese stock index futures-structure and applications. new york: salomon brothers inc. jacquillat, b., & solnik, b. (1978). multinationals are poor tools for diversification. journal of port folio management, winter, 8-12. johnson, g., schneeweis, t., & dinning, w. (1993). closed-end country funds: exchange rate and investment risk. financial analysts journal, november-december, 75-82. jorion, p., & roisenberg, l. (1993). synthetic international diversification. journal of portfolio management, 19(2), 65-74. officer, d., & hoffmeister, j. r. (1987). adrs: a substitute for the real thing? journal of portfolio management, winter, 61-65. peters, e. (1988). the case for international index funds. pension world, april, 26-28. wahab, m., & khandwala, a. (1993). why not diversify internationally with adrs? journal of portfolio management, winter, 75-82. webb, s., officer, d., & boyd, b. (1995). an examination of international equity markets using american depository receipts (adrs). journal of business finance & accounting, april, 415 430. the effect of racial/ethnic differences on the financial obligations ratio of renters congrong ouyanga,*, sherman d. hannab adepartment of personal financial planning, college of health and human sciences, kansas state university, 343m justin hall, 1324 lovers lane, manhattan, ks 66506-1401, usa bdepartment of human sciences, ohio state university, 115a campbell hall, 1787 neil avenue, columbus, oh 43210, usa abstract the purpose of this research was to investigate the effects of racial/ethnic status on the ratio of financial obligations payments to income among u.s. renter households. the proportion of homeowner households with a ratio over 40% has decreased since 2007, but the proportion of renter households with a ratio over 40% increased until 2013 and remained high in 2016. in 2016, 13% of homeowner households and 40% of renter households had a ratio over 40%. previous research on the financial obligations burden used an arbitrary dummy variable for whether households had a high financial obligations ratio, but we used ordinary least squares (ols) regressions on the natural log of the ratio. for renters, based on the ols regression, households with black, hispanic, and asian respondents had higher financial obligations ratios than otherwise similar households with a white respondent. controlling for other variables, hispanic households had a ratio about 26% higher than white households, asian households had a ratio 16% higher than white households, and black households had a ratio 10% higher than white households. while discrimination could be a factor in higher ratios for the groups other than whites, immigrant status and other factors plausibly are related. © 2022 academy of financial services. all rights reserved. keywords: financial literacy; financial obligations; survey of consumer finances 1. introduction debt and other financial obligations have been a concern for many years, and the consequences of having a high proportion of one’s household budget committed to debt and other *corresponding author: tel.: 785-532-1480; fax: 785-532-5505. e-mail address: congrong@ksu.edu 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 165–177 obligations could be seen in dramatic fashion in the great recession that started in december 2007 (nber, 2010). some people have concluded that excessive credit use was a contributing factor to the financial crisis leading up to the recession. household financial obligations increased drastically in united states by the early 2000s, both in aggregate levels and proportionally to household earnings (campbell, 2006; debelle, 2004; dynan & kohn, 2006). the amount of the growth in household borrowing raised concerns about the sustainability of household finance and the probable consequences for personal finance system. andriotis, brown, and shifflett (2019) concluded that americans needed to go deep in debt for a middle-class lifestyle. 2. literature review 2.1. personal finance guidelines for debt and housing payments since the 1980s, scholars with a focus on personal finance have discussed using household financial ratio guidelines to help with personal financial burden management and evaluate whether households had appropriate financial behaviors (griffith, 1985; johnson & widdows, 1985). hanna, yuh, and chatterjee (2012a) reviewed some past recommendations and noted that a starting point for some guidelines was housing affordability, with the idea that housing costs should be no more than 30% of income. the u.s. government has had subsidy programs that tried to help low-income households pay no more than 30% of income for rent (mckenna & hills, 1982, p. 24). lytton, garman, and porter (1991) streamlined debt ratio guidelines and proposed that a safe debt limit for the consumer debt-service ratio should be 10% or less, and 16-20% should be considered that the household is fully stretched. they also discussed the debt-service ratio, defined as annual consumer and mortgage debt repayment divided by annual income, and suggested that a value of less than 30% for the ratio should be considered safe. greninger, hampton, kitt, and achacoso (1996) surveyed a sample of financial educators and financial planners for opinions on a variety of financial ratio benchmarks in household portfolios and found a median response for the debt service ratio as 35% for a reasonable limit and 45% as a danger point. a similar recommendation was summarized by devaney (2000), who suggested that the total debt payment to income ratio should not exceed 30-35% when using gross income. 2.2. personal finance debt payment guidelines for renters one limitation of the personal finance guidelines related to debt is that they do not apply appropriately to renters, because the guidelines treat consumer debt payments as independent of rent payments. dynan et al. (2003) proposed use of a broader measure, financial obligations, which included obligations other than loan payments, rent, and vehicle lease payments. dynan, johnson, and pence (2003) and johnson (2005) presented analyses primarily based on aggregate data. hanna et al. (2012a) were the first authors to provide analyses of household data in terms of the financial obligations ratio. our analyses (fig. 1) show that since 2007, the proportion of renters with financial obligations burdens over 40% of 166 c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 income increased in 2013 to 40% of renters, higher than at any year from 1992 to 2010, and remained at about the same level in 2016. 2.3. consequences of not maintaining financial obligations payments turunen and hiilamo (2014) reviewed 33 studies on the relationship between heavy debt loads and health and concluded that there could be serious health consequences. for renter households, having high financial obligations may also prevent accumulation of savings needed to buy a home. for renters, a consequence of not keeping up with rent payments could be eviction. there might be serious problems for households with limited liquid assets and access to credit. badger and bui (2018) reported on research estimating that there are millions of evictions in the united states, with some resulting in homelessness. 2.4. theoretical basis of financial obligations guidelines moon, yuh, and hanna (2002) noted that none of the personal finance guidelines suggested by financial educators had been developed based on rigorous economic analyses. hanna et al. (2012a) discussed the challenges of rigorously developing guidelines, because modeling the nonmonetary consequences of default could be complex. carrying heavy financial obligations might increase the risk to households, as a decrease in income might lead to loan defaults, including mortgage, car loans, and education loans, or evictions from rental housing. trying to follow appropriate financial obligation or income ratio guideline can allow households to possibly maintain a more efficient consumption level fig. 1. proportion of all households, of homeowners, and of renters, with financial obligations burden over 40% of income, 1992-2016. note: based on weighted analyses of survey of consumer finance (scf) datasets. c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 167 over their life cycle (garman & forgue, 2015; lytton, garman, & porter, 1991). analysis of financial obligations or income ratios can be useful in understanding households’ financial health and can also be used as a guideline to assist households with appropriate financial management plans (bae, hanna, & lindamood, 1993). 2.5. racial/ethnic discrimination racial/ethnic discrimination has been extensively studied, mostly for lending and employment. there is a history of discrimination against blacks, hispanics, and asians in the united states (e.g., see discussion in kim, hanna, & lee, 2022; park, 2022). phelps (1972) proposed that when there is imperfect information between lenders and borrowers, the lenders would tend to use some observable signal (e.g., age, gender, and race/ethnic group) to distinguish between high risk and low risk customers. lenders may charge higher interest rates and/or extend less credit to minority borrowers, due to the higher average default risk (park, 2022). this practice is known as statistical discrimination, and is illegal, though might take place. research on whether lenders currently discriminate has produced mixed results. munnell et al. (1996) concluded there was discrimination against blacks in mortgage lending, but others (e.g., baek & cho, 2021) have not found differences in loan denials if credit history and other factors are controlled. given the competitive nature of lending and national markets, as well as the substantial amount of information available to lenders, it is possible that lending is not discriminatory, though this is still a matter of political controversy. rental markets are basically local in the united states, and information is more limited, so individual landlords might engage in taste or statistical discrimination. rental discrimination may happen in various forms, such as taste or statistical discrimination. it may be related to the housing supply in the market, and taste discrimination is unlikely to persist in competitive markets (martin & hill, 2000). the fair housing amendments act of 1988 (fha) prohibits housing discrimination based on “race, color, sex, religion, disability, family status, and national origin” (fair housing act of 1988, sec. 800, 1988). yet, there continues to be evidence of housing discrimination in the united states and legal actions remain rare (carpusor & loges, 2006; flage, 2018; hanson & hawley, 2011). many scholars have addressed discrimination in specific rental housing markets (carpusor & loges, 2006; hanson & hawley, 2011). carpusor and loges (2006) conducted a field experiment in los angeles area and found that rental applicants with either african american or muslim sounding names received significantly fewer positive responses than applicants with white-sounding name. ewens, tomlin, and wang (2014) studied a broader range of cities than carpusor and loges (2006) and provided additional the information to landlords, including occupation information and smoking preference, and so forth. they found that african american home-seekers received nine responses for every 10 a white home-seeker receives, and that including positive information does not affect the response rate difference between races. hanson and hawley (2011) tested racial discrimination in the rental housing market among 10 big cities. they used matched-pair audits conducted via e-mail for rental units advertised on-line. they classified their sample by both white names and african names, and social class. the authors concluded that discrimination occurred against african american names, however, when the content of the email messages implied the home-seekers 168 c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 had high social class, discrimination effects were not found. racial discrimination was more severe in neighborhoods that were near “tipping points” in racial composition, and for units that were part of a larger building. if landlords discriminate in rental markets, it is possible that they will deny applications, extend security deposit, or charge higher rental rates. the impact of discrimination in rental markets could increase the financial obligations ratio of rental households. 2.6. purpose of this research many issues are still unanswered related to household financial burdens. dynan et al. (2003) and johnson (2005) proposed a definition of household financial obligations payments, including rent, auto leases, homeowners’ insurance and property taxes in addition to household debt payments but their estimates were based on aggregate household data. hanna, yuh, and chatterjee (2012b) analyzed renter and homeowner households separately and found that the proportion of renter households with financial obligations over 40% of income was much higher than the proportion for homeowner households. hanna et al. (2012a, 2012b) did not include was payments for alimony and for child support. there are legal consequences for failure to keep up with these payments, so we added them to our definition of financial obligations. the divorce rate in the united states has been rising during the past two decades (greenwood 2012; stevenson & wolfers 2007), and kennedy and ruggles (2014) estimated that almost 50% of ever-married people had been divorced or separated in their late 50s. while for some households, alimony and child support payments are important, we found that only 6% of renter households report such payments, so our results overall were not substantially different from the hanna et al. (2012a, 2012b) results. hanna et al. (2012b) found that the effects of household characteristics on having a heavy burden differed between renter and homeowner households. additionally, hanna et al. (2012a, 2012b) included household characteristics plausibly related to the extended lifecycle model but did not include some other variables plausibly related to bounded rationality. the 2016 survey of consumer finance (scf) is the first scf dataset to include financial literacy measures, so we added a financial literacy score to the household characteristics analyzed by hanna et al. (2012a, 2012b). the primary research question for this study is: for renter households, what is the effect of racial/ethnic status on the financial obligations over income ratio, controlling for other household characteristics? 3. theoretical background 3.1. extended life cycle savings model 3.1.1. consumption smoothing with certainty the life cycle savings model is the classic normative model to analyze households’ consumption and saving behaviors (deaton 2005; modigliani & brumberg 1954). this framework c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 169 assumes that individuals plan an optimal consumption path to maximize overall expected lifetime utility. the central tenet of the normative life cycle model is that a consumer will attempt to have constant marginal utility of consumptions over time (yuh & hanna, 2010). with typical income patterns over the life cycle, households should borrow more in the early life stages and consume more than their income, spend less than income when household income is high, and dissave from accumulated assets during the retirement stage (browning & crossley 2001; browning & lusardi 1996). household borrowing decisions will depend on the level of current income, but also on expected future income levels (yuh & hanna, 2010). if an individual expects to have increased income, borrowing should be positive when he/she is young, and the proportion of borrowing should decrease as the household gets older. it may be rational to have negative net worth at some stages to smooth consumption (chen & finke, 1996). consumers who are certain of large increases in real income will rationally borrow more than consumers who expect constant or declining real income (hanna, fan, & chang 1995). 3.1.2. consumption smoothing with uncertain incomes the original life cycle model had the simplistic assumption of certainty about income, so the extended life cycle model was developed for greater realism, and uncertainty about future income patterns affects households’ savings or net worth accumulation (yuh & hanna, 2010). if uncertainty of households’ future income is included in the model, the extended life cycle certainty equivalence model implies that people should save more or borrow less with greater income uncertainty. browning and lusardi (1996) showed that higher income uncertainty was correlated with larger savings out of current income, and therefore, less borrowing. hubbard, skinner, and zeldes (1995) suggested that social insurance might serve as a cushion, reducing the need for low-income households to save out of current income. the main idea of the extended life cycle model is that a household should smooth consumption levels over its lifetime, although nonzero interest rates and discounting of the utility of future consumption might complicate the model. with all issues considered, the optimal consumption path will not be constant due to those complicating factors. xiao, ford, and kim (2011) discussed the basic life cycle saving model and extensions, such as the precautionary borrowings model, as economic theories that could be used to prescribe and/or explain household behavior. in the context of the life cycle saving model, saving, and borrowing behavior are logically connected, as implicit in the life cycle saving model are periods of dissaving when current income is low, and when a household has low levels of assets, dissaving must be accomplished by borrowing. 4. method 4.1. data the dataset for this study was the 2016 survey of consumer finances (scf). the scf is a triennial interview survey of u.s. families sponsored by the board of governors of the 170 c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 federal reserve system with the cooperation of the u.s. department of the treasury (bricker et al., 2017). hanna, kim, and lindamood (2018) provide overviews of methodological issues related to use of scf datasets. 4.2. dependent variable as the focus is on whether households have a heavy financial obligations ratio, it is important to consider the distribution of the ratio. table 1 shows the mean levels and quantiles of the ratio for all households and renter households in the 2016 scf. the mean level of the ratio was 418% for renters, and the maximum levels were extremely high, for instance, 266,872% for renters. fig. 2 shows the cumulative distributions of the financial obligations burden for renters, and obviously the distributions are very skewed. we followed hanna et al. (2012a, 2012b) in defining financial obligations to include rent, vehicle leases, debt payments, and real estate taxes on the household’s residence. additionally, we included households’ alimony and child support payments in in financial obligations. according to hanna et al (2012a), rent payments include monthly rent on homes, site or farm/ranch, and lease payments include all monthly lease payments on vehicle. households’ debt payments include the total of monthly payments on all types of loans, such as credit cards, home mortgages, lines of credit, home improvement loans, land contracts, other residential property, vehicle loans, student loans, installment loans, margin loans, loans with insurance policies, pension loans, and other loans. we followed hanna et al. (2012a) in assuming that monthly credit card payments were 2.5% of the credit card balance. fig. 3 shows the proportions of each component of financial obligations among renters. on average, rent payments comprised 70% of financial obligations for renters, followed by loan payments. table 1 distribution of financial obligations ratio for all households and for renters, 2016 scf quantile renters mean 418% maximum 266,872% 99th percentile 240% 95th percentile 104% 90th percentile 77% 75th percentile 52% median 34% 25th percentile 23% 10th percentile 15% 5th percentile 10% 1st percentile 0% minimum 0% proportion with financial obligations >40% of income 40% weighted % 36% unweighted n 2,072 note: scf = survey of consumer finance. percentages are weighted. c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 171 4.3. independent variables independent variables likely to be related to consumption were selected, mostly following the model used by hanna et al. (2012a, 2012b). the explanatory variables include the age, education, and racial/ethnic self-identification of the respondent, household health (head fig. 2. cumulative distribution of financial obligations ratio for renters. note: based on weighted analyses of renter households in the 2016 survey of consumer finance (scf) dataset. fig. 3. components of financial obligations of renters. note: weighted analysis of 2016 survey of consumer finance (scf). 172 c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 and/or spouse/partner in poor/fair health vs. good/excellent health). in addition to the variables used by hanna et al., we added attitudinal variables: whether the household had education loan payments, whether the household had alimony/child support payments, risk tolerance, expectations for the u.s. economy, transitory income shocks (household income higher or lower than normal income), and expectations about whether the household’s income would increase faster or slower than inflation. we also used the financial knowledge questions in the scf, with one point for each correct answer, so the variable ranged in value from 0 to 3. we controlled for the natural log of income, setting it equal to log(0.01) for income equal to zero. 4.4. ordinary least squares regression to obtain more insights than the investigation of factors related to an arbitrary threshold, whether the financial obligations ratio is over 40%, we used an ordinary least squares (ols) regression on the ratio on the natural log of the ratio. we tried an ols regression on the actual ratio, but the estimated effects were extremely large and mostly insignificant. using the log of the ratio reduced the influence of outlier values of the ratio. 5. results 5.1. descriptive results of financial obligations to income ratio and medians of ratios by household characteristics to obtain better insights into patterns of household behaviors on financial obligations ratio, we analyzed the median financial obligations ratio by renters. table 2 presents the median level of the ratio by categories of independent variables. the median ratio ranged from 31% for households with a white respondent to almost 39% for households with a hispanic household. there was no particular pattern for the median ratio by age for renters, but the ratio for those with a graduate degree, 28%, was much lower than the ratio for those lacking a high school degree, 41%. the median ratio for renters who got all three financial knowledge questions correct was 31%, compared with 40% for those who got all the questions wrong. we do not discuss all the patterns that did not show substantial differences in median values of the ratio by subgroups, but the median ratio decreased substantially as income increased. table 2 distribution of median financial obligations ratio by racial/ethnic category, 2016 scf characteristics distribution of renters (%) median of the financial obligations ratio (%) racial/ethnic identification white 52.62 31.39 black 24.43 36.65 hispanic 17.07 38.57 asian/other 5.88 36.18 note: scf = survey of consumer finance. weighted analyses of 2016 scf. c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 173 5.2. ols regression results on log of ratios for renters the ols regression on the log of financial obligations ratio for renters is displayed in table 3. for dummy variables, the percentage effect on the actual ratio can be computed as exp(coefficient)-1. all three racial/ethnic identity variables had significant positive effects, with estimated effects of 10 percentage points for black, 26 percentage points for hispanic, table 3 ols regressions on log of financial obligations ratio for renters, 2016 scf renter households variable coefficient p se racial/ethnic identification of the respondent (reference = white) black 0.0994 0.0285 0.0449 hispanic 0.2283 0.0000 0.0518 asian/other 0.1513 0.0471 0.0751 log (income) �0.4826 0.0000 0.0149 age of respondent �0.0041 0.5115 0.0062 age squared/10,000 0.4083 0.5286 0.6372 planning horizon (years) �0.0020 0.6522 0.0045 financial literacy score (0–3) �0.0010 0.9621 0.0205 have education loan 0.0762 0.0831 0.0433 credit constrained 0.1460 0.0012 0.0439 have alimony or child support payments 0.4297 0.0000 0.0732 have a child under 18 0.0631 0.1483 0.0433 risk tolerance (0–10 scale) �0.0012 0.8569 0.0065 employment status (reference = employee) self-employed �0.0371 0.5201 0.0574 not in labor force �0.0533 0.4059 0.0635 retired �0.1382 0.0255 0.0616 expectation for the economy (reference = same) expect better 0.0439 0.2768 0.0402 expect worse �0.0326 0.4784 0.0455 respondent education (reference = high school degree) education < high school degree 0.0572 0.3551 0.0617 some college 0.1680 0.0004 0.0470 bachelor’s degree 0.2603 0.0000 0.0580 post-bachelor’s degree 0.2730 0.0003 0.0738 household composition (reference = married couple) single male �0.2102 0.0004 0.0584 single female �0.1030 0.0532 0.0526 unmarried couple �0.0065 0.9133 0.0594 poor or fair health �0.0754 0.0687 0.0404 expectation for household income (reference = increase faster than prices) sure same �0.0334 0.5569 0.0566 sure less �0.0566 0.4235 0.0692 unsure �0.0403 0.4604 0.0538 income compared with a normal year (reference = about the same) higher than normal �0.1219 0.0554 0.0630 lower than normal 0.1160 0.0118 0.0456 intercept 3.8216 <0.0000 0.2166 note: scf = survey of consumer finance; ols = ordinary least squares. rii procedures, unweighted analyses, 2016 scf. significant p-values are given in bold emphasis to indicate significant results from analysis. 174 c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 and 16 percentage points for asian/other (table 4). the effect of the log of income was negative and implies a large negative decrease in the financial obligations ratio as income increases. age, planning horizon, risk tolerance, financial literacy, expectations for the economy, expectations for the household’s income, and health did not have significant effects. the effect of the dummy variable for having an education loan was marginally significant (two-tail p value of 0.08). the dummy variables for being credit constrained, for having alimony or child support payments had significant positive effects. the retired variable had a significant negative effect compared with an employee household. three of the education variables were significant. the results implied that the ratio increased with education. 6. discussion and conclusions for renters, based on the ols regression, black, hispanic, and asian renter households had higher financial obligations ratios than otherwise similar white renter households. the positive effects of being credit constrained for renters suggested that credit constrained households already had high financial obligations, rather than having been arbitrarily denied credit and other opportunities such as rental units. given that rent comprises 70% of the financial obligations of renters, if discrimination is important in explaining racial/ethnic differences in the financial obligations ratio, landlords might have a more important role than lenders. rental discrimination may play a role in rental markets. if landlords deny applications, extend security deposit, or charge higher rental rates, minority renter households could obtain higher financial obligations than white renter households. the result that hispanic and asian households have higher ratios than otherwise similar black households (table 3) may be related to the high proportion of immigrants in the first two groups, perhaps leading to less familiarity with rental markets and less ability to obtain more affordable housing. there might also be geographic differences that we could not control due to the suppression of geographic information in the scf (hanna et al., 2012a). while we did find racial/ethnic differences in the financial obligations of renters, it is not clear whether discrimination is the cause of these differences. further research is needed to ascertain causes of the differences. table 4 effects of racial/ethnic status on financial obligations ratio for renters variable % effect reference category = white black 10.45% hispanic 25.65% asian/other 16.33% note: based on regression results in table 3. effect = exp(coefficient)-1. c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 175 references andriotis, a., brown, k., & shifflett, s. (2019, august 1). families go deep in debt to stay in the middle class. wall street journal. badger, e., & bui, q. (2018, april 8). in 83 million eviction records, a sweeping and intimate new look at housing in america. new york times. bae, m., hanna, s., & lindamood, s. (1993). patterns of overspending in u.s. households. financial counseling and planning, 4, 11–30. baek, h. y., & cho, d. d. (2021). are there racial differences in loan approvals? applied economics letters, 28(4), 260–263. https://doi.org/10.1080/13504851.2020.1751047 bricker, j., dettling, l. j., henriques, a., hsu, j. w., jacobs, l., moore, k. b., . . . &windle, r. a. (2017). changes in us family finances from 2013 to 2016: evidence from the survey of consumer finances. federal reserve bulletin, 103, 1–42. available at: https://web.archive.org/web/20190428144503id_/https://www.federalreserve. gov/publications/files/scf17.pdf browning, m., & crossley, t. f. (2001). the life-cycle model of consumption and saving. journal of economic perspectives, 15, 3–22. https://doi.org/10.1257/jep.15.3.3 browning, m., & lusardi, a. (1996). household saving: micro theories and micro facts. journal of economic literature, 34, 1797–1855. campbell, j. y. (2006). household finance. the journal of finance, 61, 1553–1604. https://doi.org/0.1111/ j.1540-6261.2006.00883.x canner, g. g., & luckett, c. a. (1991). payment of household debt. federal reserve bulletin. april: 19–30. available at: https://fraser.stlouisfed.org/title/federal-reserve-bulletin-62/april-1991-20827/payment-householddebts-113310 carpusor, g. a., & loges, e. w. (2006). rental discrimination and ethnicity in names. journal of applied social psychology, 36, 934–952. https://doi.org/10.1111/j.0021-9029.2006.00050.x chen, p., & finke, m. (1996). negative net worth and the life cycle hypothesis. financial counseling and planning, 7, 87–96. deaton, a. (2005). franco modigliani and the life cycle theory of consumption. ssrn working paper 686475. debelle, g. (2004). macroeconomic implications of rising household debt. bis working paper no 153. available at ssrn: https://ssrn.com/abstract=786385 devaney, s. a. (2000). using financial ratios. in garmane. t.xiaoj. j., & brunsonb. (eds.), the mathematics of personal financial planning (pp. 147–161). houston, tx: thompson international learning. dynan, k., johnson, k., & pence, k, & the board of governors of the federal reserve system. (2003). recent changes to a measure of u.s. household debt service. federal reserve bulletin, 89, 417–426. https://doi.org/ 10.17016/bulletin.2003.89-10 dynan, k., & kohn, l. d. (2006). the rise in u.s. household indebtedness: causes and consequences. feds working paper 2007-37. available at http://www.federalreserve.gov/pubs/feds/2007/200737/200737abs.html ewens, m., tomlin, b., & wang, l. c. (2014). statistical discrimination or prejudice? a large sample field experiment. the review of economics and statistics, 96, 119–134. https://doi.org/10.1162/rest_a_00365 fair housing act of 1988, sec. 800. [42 u.s.c. 3601 note]. available at https://www.justice.gov/crt/fair-housing-act-2 flage, a. (2018). ethnic and gender discrimination in the rental housing market: evidence from a meta-analysis of correspondence tests, 2006-2017. journal of housing economics, 41, 251–273. https://doi.org/10.1016/j. jhe.2018.07.003 garman, t. e., & forgue, r. e. (2015). personal finance (12th ed.). boston, ma: houghton mifflin company. greenwood, l. j. (2012). parent-child relationships in the context of a mid-to latelife parental divorce. journal of divorce & remarriage, 53, 1–17. https://doi.org/10.1080/10502556.2012.635959 greninger, s. a., hampton, v. l., kitt, k. k., & achacoso, j. a. (1996). ratios and benchmarks for measuring the financial well-being of families and individuals. financial services review, 5, 57–70. https://doi.org/ 10.1016/s1057-0810(96)90027-x griffith, r. (1985). personal financial statement analysis: a modest beginning. in g. langrehr (ed.), proceedings, third annual conference of the association of financial counseling and planning educators (vol. 3, pp. 123– 131). afcpe ames, ia. 176 c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 hanna, s. d., fan, j. x., & chang, y. r. (1995). optimal life cycle savings. financial counseling and planning, 6, 1–15. hanna, s. d., kim, k. t., & lindamood, s. (2018). behind the numbers: understanding the survey of consumer finances. journal of financial counseling and planning, 29, 410–418. https://doi.org/10.1891/1052-3073.29.2.410 hanna, s. d., yuh, y., & chatterjee, s. (2012a). the increasing financial obligations burden of u.s. households: who is affected? international journal of consumer studies, 36, 588–594. https://doi.org/10.1111/j.14706431.2012.01125.x hanna, s. d., yuh, y., & chatterjee, s. (2012b). were consumer mistakes related to changes in financial obligations burdens? consumer interests annual. hanson, a., & hawley, z. (2011). do landlords discriminate in the rental housing market? evidence from an internet field experiment in u.s. cities. journal of urban economics, 70, 99–114. https://doi.org/10.1016/j. jue.2011.02.003 hubbard, r. g., skinner, j., & zeldes, p. s. (1995). precautionary saving and social insurance. journal of political economy, 103, 360–399. https://doi.org/10.1086/261987 johnson, k. w., & the board of governors of the federal reserve system. (2005). recent developments in the credit card market and the financial obligations ratio. federal reserve bulletin, 91, 473–486. https://doi.org/ 10.17016/bulletin.2005.91-4 johnson, d., & widdows, r. (1985). emergency fund levels of household. in k. schnittgrund (ed.), proceedings of the 31st annual conference of the american council on consumer interests (pp. 235–241). columbia, mo: american council on consumer interests kennedy, s., & ruggles, s. (2014). breaking up is hard to count: the rise of divorce in the united states, 19802010. demography, 51, 587–598. https://doi.org/10.1007/s13524-013-0270-9 kim, k. t., hanna, s. d., & lee, j. (2022). the great expansion of 2019 and the delinquency rates of blacks and hispanics. applied economics letters, 29, 1313–1318. https://doi.org/10.1080/13504851.2021.1927957 lytton, h. r., garman, t. e., & porter, n. (1991). how to use financial ratios when advising clients. financial counseling and planning, 2, 2–23. martin, r., & hill, r. (2000). loan performance and race. economic inquiry, 38(1), 136–150. https://doi.org/ 10.1111/j.1465-7295.2000.tb00009.x mckenna, w. f., & hills, c. a. (1982). the report of the president’s commission on housing. president’s commission on housing. modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. in k. kurihara (eds.), post keynesian economics (pp. 388–446). new brunswick, nj: rutgers university press moon, s. j., yuh, y., & hanna, s. d. (2002). financial ratio analysis of korean households. family and consumer sciences research journal, 30, 496–525. https://doi.org/10.1177/1077727x02030004004 munnell, a., browne, l., mceneaney, j., & tootell, g. (1996). mortgage lending in boston: interpreting hmda data. american economic review, 86(1), 25–53. nber. (2010). business cycle dating committee, national bureau of economic research. national bureau of economic research, december 2007. park, k. a. (2022, february 7). a comparison of mortgage denial and default rates by race, ethnicity, and gender. u.s. department of housing and urban development. https://doi.org/10.2139/ssrn.4030908 phelps, e. s. (1972). the statistical theory of racism and sexism. american economic review, 62(4), 659–61. stevenson, b., & wolfers, j. (2007). marriage and divorce: changes and their driving forces. journal of economic perspectives, 21, 27–52. https://doi.org/10.1257/jep.21.2.27 turunen, e., & hiilamo, h. (2014). health effects of indebtedness: a systematic review. bmc public health, 14, 489. xiao, j. j., ford, w. m., & kim, j. (2011). consumer financial behavior: an interdisciplinary review of selected theories and research. family and consumer sciences research journal, 39, 399–414. https://doi.org/10.1111/ j.1552-3934.2011.02078.x yuh, y., & hanna, s. d. (2010). which households think they save? journal of consumer affairs, 44, 70–97. https://doi.org/10.1111/j.1745-6606.2010.01158.x c. ouyang and s. d. hanna / financial services review 30 (2022) 165–177 177 pii: s1057-0810(99)00043-8 international mutual fund returns and federal reserve policy robert r. johnsona,*, gerald w. buetowb, gerald r. jensenc aassociation for investment management & research, 560 ray c. hunt drive, p.o. box 3668, charlottesville, va 22903-0668, usa bbfrc services, 1705 owensville drive, charlottesville, va 22901, usa cnorthern illinois university, school of business, dekalb, il 60115-2854, usa received 26 april 1999; received in revised form 15 october 1999; accepted 8 november 1999 abstract this study examines the performance of international mutual fund indexes across alternative federal reserve monetary policy environments. the results suggest that the benefits touted by advocates of international diversification may be less than previous studies indicate. specifically, during restrictive us monetary policy periods, international mutual fund indexes provide lower excess returns than domestic counterparts. additionally, the correlations between international mutual funds and domestic mutual funds are higher during restrictive monetary policy periods. this evidence may represent a partial explanation for the home country bias exhibited by us-based individual and institutional investors. © 2000 elsevier science inc. all rights reserved. jel classification:e44; e52; g15 keywords:federal reserve; international diversification; discount rate; monetary policy 1. introduction academic research has touted the benefits of international portfolio diversification arguing that us investors could improve their risk-return profiles by purchasing international equi* corresponding author. tel.:11-804-951-5255; fax:11-804-951-5200. e-mail address:rrj@aimr.org (r.r. johnson) financial services review 8 (1999) 199–210 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00043-8 ties. however, home asset bias, or the tendency of investors to over invest in assets based in their home country, is particularly pronounced in us markets. for many years, market watchers have recognized that monetary conditions, and the federal reserve in particular, have a significant influence on security returns. academic research (summarized in section 2) shows that stock and bond returns are significantly higher in periods characterized by an expansive monetary policy than security returns during restrictive monetary periods. the influence of the federal reserve is not limited to domestic stock and bond returns, however, as recent research has shown that us central bank policy is also associated with foreign market return patterns. the objective of this study is to examine the performance of international mutual fund indexes across alternative federal reserve monetary policy environments. the focus is on the potential diversification benefits of investing in international mutual funds. the results suggest that the benefits touted by advocates of international diversification may be less than previous studies indicate and may help explain home asset bias. the results also suggest that the greatest diversification benefit for us investors is achieved by investing in diversified asia/pacific funds. however, in the time period studied, this diversification benefit is achieved at the cost of lower returns. the remainder of this study is organized as follows: section 2 reviews the literature on international diversification and the influence of monetary policy on security returns; section 3 describes the methodology of the study and data employed; section 4 reviews the results of our analysis; and section 5 concludes the paper. 2. literature review 2.1. international diversification proponents of international diversification claim that the volatility of domestic markets can be somewhat offset by investing in foreign markets. aiello & chieffe (1999) conclude that investment into international mutual funds may offer significant diversification benefits. still, as summarized by shawky, kuenzel, & mikhail (1997), the benefits of international portfolio diversification continue to be a controversial topic in the financial literature. opponents contend that international diversification has no economic rationale (see sinquefeld, 1996). us institutional investors seem to be slowly embracing the concept of international diversification as the trend has been toward greater commitment of funds internationally. from 1991 to 1996, foreign securities held by tax-exempt us pension funds more than doubled. however, the potential for additional international investment is enormous. gorman (1998) reports that with respect to the equity portfolio, the cross-border commitment of the typical us pension plan is less than half that of the typical non-us pension plan. almost 90% of us assets are still invested domestically, which far exceeds the domestic holdings in most countries with developed pension systems. according to melton (1996), the united kingdom and hong kong pension funds have 28.0% and 56.8% of their assets in overseas equities and bonds, respectively. 200 r.r. johnson et al. / financial services review 8 (1999) 199–210 the potential benefits of international diversification have resulted in increased individual investor interest in international mutual funds. as reported by bers (1998), in 1984 only 13 international mutual funds existed. by 1995, this number had increased to 335. according to the investment company institute (ici), as of december 1996, the average individual investor in the us held a position similar to us pension plans. gorman (1998), notes that of the $2637 billion in reported mutual fund holdings in the us, only $321 billion, or 12%, is invested abroad (principally in stocks). although this commitment is relatively small in percentage terms, as with institutional investors, the trend is increasing. the costs of international diversification and the ability of a domestic stock portfolio to hedge domestic inflation risk are cited as primary reasons for home asset bias. tesar & werner (1995) find that the high turnover on foreign equity investments relative to turnover on domestic equity markets suggests that transactions costs alone are an unlikely explanation for home asset bias. cooper & kaplanis (1994) examine each of these explanations and find that, at best, each is only a partial explanation for the home asset bias. previous research investigated the effectiveness of various forms of investment in obtaining international diversification for us investors. bailey & lim (1992), chang, eun, & kolodny (1995), and barry, peavy, & rodriguez (1997) find that country funds listed on us exchanges are more highly correlated with us markets than the returns from their respective benchmarks. russell (1998) concludes that us exchange listed securities such as american depository receipts, closed-end country funds, and multinational corporations behave more like the host exchange than their home exchange. this result suggests that these us exchange-listed securities, on average, do not perform an effective international diversification role for us investors. several recent studies provide evidence that the benefits of international diversification may be substantially less than suggested by early academic researchers. in a study of the us and its g-7 partners, hanna, mccormick, & perdue (1999) find that markets do not move in opposite directions with enough frequency to justify the assertion that foreign gains will compensate for domestic losses. ho, milevsky, & robinson (1999) conclude that international diversification provides a substantial benefit to canadian investors by reducing shortfall risk, but does not benefit american investors materially. the benefits of international diversification rest largely on the correlation structure of international market returns. as discussed by conover, jensen, & johnson (1999), an underlying factor that significantly influences the benefits of international diversification is the stability of cross-country correlations over time. the consistency of the co-movements between international stock market indexes is examined in several studies. longin & solnik (1995) and solnik, boucrelle, & lefur (1996) find that the correlation structure has fluctuated widely over time, yet has only risen slightly during the 30-year time period examined. another interesting finding of longin and solnik and confirmed by shawky, kuenzel, & mikhail (1997) is that the correlations seem to be higher in times of high market volatility. erb, harvey, & viskanta (1994) report that international equity correlations are higher during recessions than during expansions, confirming these findings. bookstaber (1997) concludes that markets are not normal and that diversification benefits are greatly mitigated when the investor needs them most. 201r.r. johnson et al. / financial services review 8 (1999) 199–210 2.2. federal reserve monetary policy and capital market returns short-term reactions to changes in federal reserve monetary policy have been empirically documented in both the us and foreign markets. among others, jensen & johnson (1993) find evidence of monetary policy “announcement effects” in the stock market. furthermore, the short-term effects are not limited to us markets, as johnson & jensen (1993) report that us monetary policy changes are also associated with reactions in foreign equity markets. recent evidence suggests that monetary conditions are also related to long-term performance patterns in security markets. prather & bertin (1997) present a simple trading strategy for individual investors based on federal reserve announcements of discount rate changes. conover, jensen, & johnson (1999) provide evidence indicating that international stock markets also exhibit patterns that are linked to monetary policy changes. the authors demonstrate that the patterns are in the same direction as us market patterns and, for several countries, of comparable size to those documented in the us stock market. 3. methodology and data 3.1. mutual fund indexes the international mutual fund data used in this analysis is from morningstar and the indexes chosen are the morningstar aggregates that have a majority of assets invested in non-us securities. in addition, to allow comparability across indexes, the index had to be in existence at the end of 1976. the five non-us indexes examined in this paper are 1) diversified emerging markets, 2) foreign stock, 3) multi-asset global, 4) diversified asia/pacific, and 5) world stock index. for comparative purposes, two morningstar us mutual fund indexes, us balanced and us growth & income, are also examined. although the investment objective stated in a fund’s prospectus may or may not reflect how the fund actually invests, morningstar assigns each mutual fund to a category based on the underlying securities in its portfolio. categories are assigned by morningstar based on the average of the past three years’ portfolio holdings. each index is a simple arithmetic average of all mutual funds in the category in existence at that time. table 1 provides a detailed description of each of the indexes examined in this analysis. 3.2. defining federal reserve monetary policy defining federal reserve monetary policy is controversial. numerous approaches to “fed-watching” involve monitoring money supply, interest rates, bank reserves, open market operations, and combinations of monetary variables. however, operationally, many of these strategies are not conducive to the vast majority of investors, as the methods are quite complex and require a constant monitoring of economic variables (see jones, 1989 for a detailed description of various approaches). the simple and unambiguous binary definition of fed policy utilized in this analysis is consistent with conover, jensen, & johnson (1999), 202 r.r. johnson et al. / financial services review 8 (1999) 199–210 jensen, mercer, & johnson (1996), booth & booth (1997) and prather & bertin (1997) and employs the federal reserve discount rate. textbook discussions of monetary policy treat the discount rate as one of the federal reserve’s three principal policy tools, the others being reserve requirements and open market operations. technically, the discount rate is the rate at which member institutions can borrow reserves from the federal reserve. in practice, however, discount rate changes are often interpreted as signals of the future course of monetary policy. previous researchers note that discount rate changes occur only at substantial intervals, they represent a rather discontinuous instrument of monetary policy, and are established by a public body having special information and competence to judge whether changes in bank credit and money is consistent with the economy’s cash needs. thus, discount rate changes may be viewed as precursors of future fed monetary policy. we classify the monetary environment as either expansive or restrictive based on the most recent discount rate change. the monetary environment remains the same until the discount rate is changed in the opposite direction, because the central bank is assumed to be operating under the same general policy until the discount rate change is reversed. the period following a decrease in the discount rate is classified as expansive. further discount rate decreases do not affect the classification of the monetary environment. likewise, restrictive monetary environments begin when the discount rate first increases and end when the table 1 description of the morningstar international indexesa morningstar index description diversified emerging markets an equity fund with at least 50% of stocks invested in emerging markets. foreign an international equity fund having no more than 10% of stocks invested in the united states. multi-asset global funds in this objective seek total returns by investing in varying combinations of equities, fixed income securities, and other asset classes. these funds may invest a significant portion of assets in securities of foreign issuers. diversified asia/pacific an equity fund with at least 65% of stocks invested in pacific countries with at least an additional 10% of stocks invested in japan. world an international fund having more than 10% of stocks invested in the united states. us balanced seek both income and capital appreciation by investing in a generally fixed combination of stocks and bonds. these funds generally hold a minimum of 25% of their assets in fixed-income securities at all times. us growth & income growth of capital and current income are nearequal objectives. investments are typically selected for both appreciation potential and dividend-paying ability. a source: morningstar mutual fund 500, 1998–1999 edition, chicago, morningstar. 203r.r. johnson et al. / financial services review 8 (1999) 199–210 discount rate is decreased. jensen, mercer, & johnson (1996) employ this classification scheme and report empirical evidence demonstrating that the levels of, as well as changes in, the federal funds premium, monetary aggregates, and reserve aggregates differ significantly across the defined environments, thus supporting the view that this classification technique effectively differentiates monetary conditions. although this classification technique has been effectively employed to differentiate fundamentally different monetary conditions, the procedure is not advocated as the best technique for identifying minor changes in the stringency of monetary policy. a more refined approach that adjusts more frequently would be required to accomplish this task (see thorbecke, 1997 and patelis, 1997 for examples). use of such measures, however, requires more frequent trading, more subjective evaluation, and a more sophisticated investor. 3.3. time period examined in the nearly 22-year period covered in this study (the period is 2 months short of 22 years), the federal reserve changed the discount rate 55 times: 26 increases and 29 decreases. in this period, however, there are only eleven “rate-change series,” effectively representing what we believe are fundamental changes in the monetary environment (see table 2 for a list of these dates). we examine monthly return data from december 1976 through september 1998. consistent with previous research, we do not include the month in which the federal reserve changed from an expansive to a restrictive monetary policy or from a restrictive to an expansive policy. these months are omitted for two reasons. first, our objective is to focus on the long-term relationship between monetary conditions and security returns, and thus, we eliminate any announcement-period effect. second, the return associated with months that mark the initiation of a new monetary environment would include both expansive and restrictive days. a total of 252 months are in the sample: 144 months in expansive periods and 108 months in restrictive periods. table 2 discount rate change series series increasing (i) or decreasing (d) first rate change in series rate changes in series monthly observations in series 1 d 12/09/74 0 8a 2 i 08/30/77 14 32 3 d 05/29/80 3 3 4 i 09/26/80 4 13 5 d 11/02/81 9 28 6 i 04/09/84 1 6 7 d 11/23/84 7 33 8 i 09/04/87 3 38 9 d 12/18/90 7 24 10 i 05/17/94 4 19 11 d 01/31/96 3 32 a beginning with december of 1976. 204 r.r. johnson et al. / financial services review 8 (1999) 199–210 3.4. methodology summary statistics are computed for the six international and two domestic indexes included in the study. in this context, a monthly excess return is defined as the monthly index return minus the monthly risk-free rate. for the riskless rate of return, we use the monthly t-bill return as compiled by ibbotson (1999). sharpe ratios are computed for each of the mutual fund indexes for the total sample period, expansive monetary policy periods, and restrictive monetary policy periods. the sharpe ratio(s) measures the average excess return per unit of total risk. the ratio is calculated as follows: s5 [(r i 2 rf)/sp] where, ri is the average rate of return on the mutual fund index, and rf is the average rate of return on the risk free asset, andsp is the standard deviation of the mutual fund index returns. differences in the distributions of monthly excess returns across monetary policy environments are also gauged by analyzing skewness and excess kurtosis. stuart & ord (1987) find that in large samples of normally distributed data, the sample skewness and sample kurtosis estimates are normally distributed with means 0 and 3 and variances 6/n and 24/n, respectively (wheren 5 the number of observations). because 3 is the kurtosis of the normal distribution, sampleexcess kurtosisis defined to be sample kurtosis less 3. campbell, lo, & mackinlay (1997) find that sample estimates of skewness for daily us stock returns tend to be negative for stock indexes but close to zero or positive for individual stocks. sample estimates of excess kurtosis for daily us stock returns are large and positive for both indexes and individual stocks, indicating that returns have more mass in the tail areas than would be predicted by a normal distribution. sample statistics for monthly returns show that these are generally less leptokurtic than daily returns. 4. results 4.1. returns and distributions of returns table 3 presents summary statistics for the first four moments of monthly excess returns of mutual fund indexes and sharpe ratios. the patterns of mean monthly excess returns are identical for each of the seven mutual fund indexes examined. for each index, the mean excess return during expansive periods is higher than the mean excess return during restrictive monetary policy periods. as expected, the sharpe ratios indicate that the return per unit of risk during expansive periods is much greater than the return per unit of risk during restrictive periods for each of the indexes examined. for the total sample period, the highest sharpe ratios were realized for the two us indexes, whereas the lowest sharpe ratios were realized by the diversified emerging market and diversified asia/pacific stock indexes. during expansive monetary policy periods, the highest sharpe ratio was realized by the multi-asset global index, whereas the lowest sharpe ratio was realized by the diversified emerging markets index. in 205r.r. johnson et al. / financial services review 8 (1999) 199–210 restrictive periods, the two us indexes and the world stock index realized the only positive sharpe ratios. as shown in table 3, the distribution of excess monthly returns for each of the seven mutual fund indexes examined all exhibit negative skewness with the exception of multiasset global index during expansive periods. each of the other 20 excess returns distributions exhibit negative skewness. with the exceptions of the diversified emerging market and diversified asia/pacific stock indexes, the restrictive period distributions for the indexes examined are more negatively skewed during restrictive periods than during expansive periods. in addition, with the exceptions of the us balanced and multi-asset global indexes during expansive periods and the diversified asia/pacific stock index during both expansive and restrictive periods, each of the excess returns distributions examined exhibits statistically significant negative skewness. table 3 also shows that each of the excess return distributions examined exhibit statistically significant excess kurtosis. this means that the distributions have more mass in the tail areas than would be predicted by a normal distribution. therefore, investors in these funds table 3 first four moments of monthly excess returns of mutual fund indexes and sharpe ratios index mean % st. dev. % skewness excess kurtosis sharpe ratio us indexes us balanced total 0.4550 2.8011 20.65016* 2.3987* 0.1624 expansive 0.7485 2.5568 20.08779 1.3324* 0.2928 restrictive 0.0273 3.0966 21.25359* 2.5995* 0.0089 us growth & income total 0.6014 3.8394 20.84386* 3.5475* 0.1566 expansive 0.9660 3.5282 20.43631* 2.0961* 0.2738 restrictive 0.0884 4.2917 21.14309* 3.4793* 0.0206 international indexes world total 0.6025 4.1841 21.18929* 4.2607* 0.1440 expansive 1.0290 3.6920 20.77808* 2.4189* 0.2787 restrictive 0.0111 4.8053 21.37309* 3.9742* 0.0023 div. emerging markets total 0.2995 5.7242 21.17958* 4.1321* 0.0523 expansive 0.5317 5.4097 21.30914* 5.5130* 0.0983 restrictive 20.0676 6.2520 21.00430* 2.6243* 20.0108 foreign total 0.4838 4.0568 21.04231* 4.0158* 0.1193 expansive 0.9777 3.7638 20.83259* 2.1898* 0.2598 restrictive 20.2197 4.4192 21.67300* 5.6920* 20.0497 multi-asset global total 0.4331 2.9166 21.12913* 5.5812* 0.1485 expansive 0.7902 2.4834 0.12967 1.1254* 0.3182 restrictive 20.1189 3.4027 21.54121* 5.4171* 20.0349 diversified asia/pacific total 0.4036 5.4497 20.35636* 1.0417* 0.0741 expansive 0.7097 5.2129 20.33147 0.9290* 0.1361 restrictive 20.2145 5.7949 20.30478 1.0145* 20.0370 * significant at the 5% level. 206 r.r. johnson et al. / financial services review 8 (1999) 199–210 would have a greater probability of realizing larger positive or negative excess returns than predicted using a normal distribution. a pattern in excess kurtosis is also evident. with the exception of the diversified emerging markets index, each index exhibits greater excess kurtosis in restrictive periods than expansive periods. thus, the leptokurtic nature of the distributions is less pronounced in expansive periods as compared to restrictive periods. interpreting only the first two moments suggested that the restrictive periods offer both lower returns and higher variability. although this conclusion is valid, the skewness and kurtosis results imply that the restrictive period is even more risky than this initial analysis suggested. during a restrictive monetary environment the probability of realizing a large negative return is much larger than that suggested by the normal distribution. additionally, the probability of realizing a large negative return during a restrictive environment is larger than during an expansive environment. these findings further distinguish the behavior of security returns during the two monetary policy environments and have important investment implications. consistent with previous researchers, the results presented in table 3 show that the security returns are not normal; they also suggest that the investment environment is far more risky than the normal distribution would indicate. the investor would be wise to adopt a tactical asset allocation strategy moving assets away from those categories that are most negatively affected by the monetary policy environment. as shown in table 4, using a simplet-test for differences in means assuming unequal variances, the differences in mean excess returns during expansive and restrictive monetary environments are statistically significant at the 5% level for five of the seven indexes examined. the magnitudes of the differences are not only statistically significant, but are economically significant. on an annualized basis, the excess return differences range from a high of 14.37% for the foreign stock index to a low of 7.19% for the diversified emerging market index. curiously, with the exception of the world stock index, all of the international table 4 tests for differences in returns and risk between expansive and restrictive monetary policy periods index difference in mean monthly excess returns (t-statistic) difference in standard deviation of mean monthly excess returns (f-statistic) us indexes us balanced 0.7212 20.5398 (1.97)* (0.68)* us growth & income 0.8776 20.7635 (1.73)* (0.68)* international indexes world 1.0178 21.1133 (1.83)* (0.59)* diversified emerging markets 0.5993 20.8423 (0.80) (0.75) foreign 1.1974 20.6554 (2.27)* (0.73)* multi-asset global 0.9091 20.9193 (2.35)* (0.54)* diversified asia/pacific 0.9242 20.5820 (1.31) (0.81) * significant at the 5% level. 207r.r. johnson et al. / financial services review 8 (1999) 199–210 indexes have negative excess returns during restrictive monetary policy periods, whereas each of the two domestic indexes have positive excess returns during restrictive periods. this indicates that investors would, on average, improve their absolute performance by simply investing in t-bills instead of diversified emerging market, foreign stock, multi-asset global, and diversified asia/pacific stock funds during restrictive monetary policy periods. patterns in the variability of excess returns are also apparent. for each of the seven indexes examined, the excess returns during restrictive periods are more variable than excess returns in expansive periods. as indicated in table 4, the results of thef-tests indicate that the risk differences are statistically significant for all indexes except the diversified emerging markets and diversified asia/pacific stock indexes. one cannot conclude that the extra returns realized during expansive periods are the result of compensation for bearing extra risk. as one might anticipate, the standard deviation of excess returns for the diversified emerging market index is the highest, while the standard deviation of excess returns for the us balanced index is lowest of the indexes examined. thus, periods of decreasing discount rates are associated with both higher returns and lower variability of returns for each of the seven classes of mutual funds examined. 4.2. the correlation structure of returns table 5 presents the correlations between us balanced funds and the other indexes of mutual funds examined in this study. balanced funds are chosen to serve as representative of the average us individual investor’s holdings. these funds contain both stocks and bonds and are often utilized by individuals in retirement plans. the extremely high correlation (0.972 overall) between the indexes for us balanced funds and us growth & index funds indicates that the correlation results would be very similar if us growth & index funds were used as the representative class of funds. the correlations between us balanced funds and the five international funds range from a low of 0.415 (with diversified asia/pacific funds) to a high of 0.883 (with world stock funds). it is not surprising that the correlation between us balanced funds and world stock funds is the highest, as world stock funds are defined as international funds having more than 10% of stocks invested in the united states. overall, the correlations are higher than the typical correlations reported in studies between country indices. this implies that investors table 5 us balanced fund and international mutual fund correlations correlation of us balanced fund with: expansive returns correlation restrictive returns correlation total correlation us growth & income 0.967 0.974 0.972 international indexes world 0.863 0.903 0.883 diversified emerging markets 0.746 0.859 0.796 foreign 0.593 0.745 0.678 multi-asset global 0.797 0.885 0.844 diversified asia/pacific 0.369 0.440 0.415 208 r.r. johnson et al. / financial services review 8 (1999) 199–210 who target international mutual funds may not be realizing the diversification benefits assumed by examining the results of previous academic studies. furthermore, as shown in table 5, the correlations during restrictive monetary policy periods are higher than the correlations during expansive monetary policy periods for each of the indexes examined. if a portfolio is formed based on average correlations, which implicitly assumes symmetry, the performance of the investment could be worse than expected in restrictive monetary policy environments because the correlations increase. similar to the conclusions of previous researchers, portfolios need to be constructed on the basis of expected correlation rather than past averages. our results suggest that one of the factors investors should use in forecasting the expected correlation structure of security returns is the monetary environment. the results show that the diversification benefits are reduced (higher correlations) when we need them most (during a restrictive environment). these findings further suggest that the investor adopt a tactical asset allocation strategy that reduces these effects during restrictive environments. 5. conclusions the results of this study provide evidence of the relative inability of international mutual funds to allow investors to realize the anticipated level of diversification benefits across federal reserve monetary policy environments. specifically, during restrictive monetary periods, international mutual funds indexes provide lower excess returns than domestic counterparts. additionally, the correlations between international mutual funds and domestic mutual funds are higher during restrictive monetary policy periods. the ineffective diversification provided by international mutual funds is consistent with other forms of international investment (i.e., country funds, and american depository receipts), which have been shown previously to offer limited diversification benefits for us investors. this evidence may represent a partial explanation for the home country bias exhibited by us-based individual and institutional investors. references aiello, s. & chieffe, n. (1999). international index funds and the investment portfolio.financial services review, 8(1), 27–36. bailey, w. & lim, j. (1992). evaluating the diversification benefits of the new country funds.journal of portfolio management, 18(3), 75–80. barry, c. b., peavy, j. w., iii, & rodriguez, m. (1997).emerging stock markets: risks, return, and performance. charlottesville, va: research foundation of the institute of chartered financial analysts. bers, m. k. (1998). causal relations among stock returns, inflation: persistence of international mutual fund performance.global finance journal, 9, 225–240. bookstaber, r. (1997). global risk management: are we missing the point?the journal of portfolio management, 23(3), 102–107. booth, j. r. & booth, l. c. (1997). economic factors, monetary policy, and expected returns on stocks and bonds, economic review: federal reserve bank of san francisco, 32–42. 209r.r. johnson et al. / financial services review 8 (1999) 199–210 campbell, j. y., lo, a. w., & mackinlay, a. c. (1997).the econometrics of financial markets. princeton, nj: university press. chang, e., eun, c. s., & kolodny, r. (1995). international diversification through closed-end country funds. journal of banking and finance, 19, 1237–1263. conover, c. m., jensen, g. r., & johnson, r. r. (1999a). monetary conditions and international investing. financial analysts journal, 55(4), 38–48. cooper, i. & kaplanis, e. (1994). home equity bias in equity portfolio, inflation hedging, and international capital market equilibrium,review of financial studies, 7, 45–60. erb, c. b., harvey, c. r., & viskanta, t. e. (1994). forecasting international equity correlations.financial analysts journal, 50(6), 39–45. gorman, s. a. (1998).the international equity commitment. charlottesville, va: research foundation of the institute of chartered financial analysts. hanna, m. e., mccormack, j. p., & perdue, g. (1999). a nineties perspective of international diversification. financial services review, 8(1), 37–46. ho, k., milvesky, m. a., & robinson, c. (1999). international equity diversification and shortfall risk.financial services review, 8(1), 11–26. ibbotson associates. (1999).stocks, bonds, bills and inflation: 1999 yearbook. chicago: ibbotson associates. jensen, g. r. & johnson, r. r. (1993). an examination of stock price reactions to discount rate changes under alternative monetary policy regimes.quarterly journal of business and economics, 32, 26–51. jensen, g. r., mercer, j., & johnson, r. r. (1996). business conditions, monetary policy, and expected security returns.journal of financial economics, 40, 213–237. johnson, r. r. & jensen, g. r. (1993). the reaction of foreign stock markets to u.s. discount rate changes. international review of economics and finance, 2, 181–193. jones, d. m. (1989).fed watching and interest rate projections. new york: new york institute of finance. longin, f. & solnik, b. (1995). is the correlation of international equity returns constant: 1970–1990?journal of international money and finance, 14, 3–26. melton, p. (1996).the investor’s guide to going global with equities. london: pitman publishing. morningstar. (1998–1999).morningstar mutual fund 500. chicago: morningstar. patelis, a. (1997). stock return predictability and the role of the monetary sector.journal of finance, 52, 1951–1972. prather, l. & bertin, w. (1997). a simple and effective trading rule for individual investors.financial services review, 6, 285–294. russell, j. w. (1998). the international diversification fallacy of exchange-listed securities.financial services review, 7, 95–106. shawky, h. a., kuenzel, r., & mikhail, a. d. (1997). international portfolio diversification: a synthesis and update.journal of international financial markets, institutions and money, 7, 303–327. sinquefield, r. (1996). where are the gains from international diversification?financial analysts journal, 52(1), 8–14. solnik, b., boucrelle, c., & lefur, y. (1996). international market correlation and volatility.financial analysts journal, 52(5), 17–34. stuart, a. & ord, k. (1987).kendall’s advanced theory of statistics(vol. 1). new york: oxford university press. tesar, l. l. & werner, i. m. (1995). home bias and high turnover.journal of international money and finance, 14, 467–492. thorbecke, w. (1997). on stock market returns and monetary policy.journal of finance, 52, 635–654. 210 r.r. johnson et al. / financial services review 8 (1999) 199–210 pii: s1057-0810(97)90015-9 financial services review, 6(3): 221-224 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. book, software, and web site reviews douglas kahl, editor university of akron the millionaire next door. thomas j. stanley and william d. danko marietta, ga: longstreet press, inc.; 1996 reviewed by: douglas r. kahl, professor of finance, the university of akron. thomas j. stanley and william d. danko surveyed 3000 heads of households expected to have high net worth. of the 1,115 respondents, 385 had household net worth of at least $1 ,ooo,ooo. they present the results of the eight page surveys in a 258-page book, the mil lionaire next door. based on the survey, in depth interviews and prior research stanley and danko paint a detailed portrait of the american millionaire. they will tell you what the millionaire is likely to drive; how much he will spend for a suit, watch and shoes; how long and how often he has been married; and, perhaps what is most important, how he became a millionaire. stanley and danko’s portrait may not surprise most experienced financial planners, but will come as a shock for many of those who want to sell goods and services to the wealthy, who want to become wealthy, or who simply want to appear to be wealthy. the results presented in the book are thoroughly discussed and are enhanced by many examples and illustrations. while the covers of the book may be a little farther apart than is abso lutely necessary, the book is well worth reading for the insights it provides into the modem american millionaire and his lifestyle. the financial services revolution. edited by clifford e. kirsch irwin publishing; 1997. reviewed by: j. tim query, dept. of risk management & insurance, university of georgia. the financial services revolution examines the important changes that have taken place in the last two decades in the financial services industry. its intention is to bridge the inter ests of policy makers, practitioners, and academics. to that end, submissions are included from practitioners, academics, and officials from regulatory agencies. editor clifford e. pii: s1057-0810(97)90032-9 financial servic~ review, 6(i): 53--67 issn:1057-0810 i copyright © 1997 by jai press inc. [ all rights of reproduction in any form reserved. i adverse selection, search costs and sticky credit card rates o. felix ayadi several scholars of financial economics observed that during the 1980s, market interest rates declined continuously with little or no impact on credit card rates. recently, mey ercord (1994), sinkey and nash (1993), and sullivan and worden (1995) recorded sigmficant changes in the credit card market intffcating an increased level of competi tion. this study represents an attempt to determine the sensitivity of credit card rates to the costs of funds in the u.$. economy. the evidence from the johansen cointegration test confirms that credit card rates and cost of funds possess a long-run equilibrium relationship with one another. furthermore, the results of the error correction models are indicative of a sluggish rate at which credit card interest rates adjust to the costs of funds. between 1982 and 1994, credit card rates adjust to changes in the cost o f funds at about 15 percent per quarter. these results represent anecdotal evidence for the validity of adverse selection, search and switch costs explanations that have been dis cussed in the financial contracting literature. i. introduction credit is the present use of future income. credit cards are probably the most convenient form of credit of all competing financial assets. they are both payment and credit devices. some holders of credit cards use them in lieu of cash for a variety of transactions while many holders use them to gain access to a revolving line of preset credit (pozdena, 1991). it is no surprise that a myriad of consumers applaud the use of credit cards primarily for its credit feature rather than its convenience (mandeu, 1990). thus, the use of credit cards has become more widespread in the past decade because they are easy to obtain, convenient to use, safer than cash and above all, they serve as a source of credit. the pricing of bank credit cards in the united states was not a concern until after 1979 when interest rates became more volatile (shay, 1987). prior to this time, interest rates were relatively stable through usury regulations. berlin and mester (1989), rose (1990), ausubel (1991), mester (1993) and calem (1992) observed that during the 1980s, market o. felix ayadi * associate professor of finance, school of business and economics, fayetteville state university, fayetteville, nc 28301; e-mail: fayadi@sbe 1.uncfsu.edu. 54 financial services review 6(1) 1997 interest rates declined continuously with little or no impact on credit card rates. the nonre sponsiveness of credit card rates to market rates led to several proposals for reinstatement of usury ceilings. this did not materialize. however, in late 1988, congress passed the fair credit and charge card disclosure act requiting banks and card issuers to provide more information to consumers. the intent of this legislation was to reduce search and switch costs associated with holding credit cards (sullivan & worden, 1995). recently, meyercord (1994), canner and luckett (1992), carries and slifer (1991), federal reserve board (1994), sinkey and nash (1993) and sullivan and worden (1995) recorded several dramatic changes in the credit card market leading to increased competi tion. meyercord observes the entrance of nonbank credit card_ issuers into the industry and that card issuers have become increasingly sophisticated in targeting their clientele. thus, card issuers are now more aggressive than ever before in attracting potential cardholders. all of these together are indicative of a more conducive environment for credit card rates to reflect market rates of interest. this paper examines the problem of sticky credit card rates and the associated search/ switch costs and adverse selection explanations. it applies johansen cointegration test and error correction model (ecm) to quarterly time series data on credit card rates and cost of funds from 1982 through 1994. this paper proceeds as follows. the second section exam ines the current market structure for credit cards in the united states, and the third focuses on the pricing of bank credit cards. the fourth section discusses the data and methodology while the fifth section presents the research results. finally, section six contains commen tary on the results. ii. market structure for bank credit cards to the current generation of consumers, credit has become a way of life. the previous gen erations detested debt and took great pride in seldom borrowing for personal reasons, because borrowing is assumed to signify character weakness. but in today's world, credit seems to be a sign of strong character. shepherdson (1991) and worthington (1995) note that the use of credit has become commonplace, particularly in the united states and the united kingdom more than in other countries. along with the change in attitude, consum ers have become more susceptible to financial problems. many are using as much as 40 to 50 percent of their income to repay outstanding loans (sbepherdson, 1991). the credit card market is quite broad and relatively unconcentrated. for example, the top ten card issuers control only forty percent of the credit card market (sinkey, 1992). the nilson report, a credit card newsletter, estimates the number of credit card issuers in the united states in 1991 at about 5,000. these issuers make a profit of 2.5 percent a year, after taxes, on all charges. moreover, sullivan and worden (1995) and sinkey and nash (1993) note that banks realize more revenue from credit cards operation compared to other ser vices. moreover, a typical cardholder in the u.s. owes visa and mastercard $400 in early 1980s, $1,096 in 1993 and $1,750 in 1994. according to ausubel (1991) and pozdena (1991), americans charge more than $200 billion a year on their credit cards. fewer than 20 percent of households are without any credit cards. prior to the marquette (national bank vs. first of omaha) decision of the u.s. supreme court of 1978, credit card issuers were subject to state usury laws. the court deci sion led to the practical elimination of price regulations. historically, credit card business sticky credit card rates 55 was not subject to the strict regulations that were applicable to banking business. thus, credit card issuers operated free of interstate banking and branch banking restrictions. the situation changed in the 1980s when several attempts were made to reintroduce usury laws at the federal level (canner and fergus, 1987). the motivation was the perception by many americans that credit card rates were insensitive to the cost of funds. this public opinion led congress to pass the fair credit and charge card disclosure act in late 1988. as noted earlier, sullivan and worden observe that the act was designed to reduce search and switch costs by requiting banks to disclose information on the costs of their services. the operators in today's credit card market are more aggressive than ever before by offering packages that include: more favorable interest rates, waivers of annual fees rebates of various types, co-branding, and accounts consolidation (ayadi and onashile, 1994). the introduction of no fee with low aprs but no interest free periods and the intensification of competition among card issuers, have led to an increasingly fragmented market. co-branding is the process in which a financial institution unites with a company to offer a credit card with unique incentives tied to the use of the card. co-branded cards have gained momentum because they offer discounts on cars, merchandise or airline tickets in proportion to the amount a cardholder charges. examples are general motors' gm card issued through household international, shell oil and chemical bank, and apple computer parmership with citicorp. a few years ago, credit card issuers began a competitive strategy of persuading cardholders to consolidate their credit balances. for example, chase manhattan bank recently began to offer the chase reward consolidator in which holders of reward cards can transfer balances on their other cards to chase. in the process, they pay a lower interest rate, and keep their rewards. finally, pozdena (1991) notes that although credit card issuers do offer lower rates to more carefully selected consumer segments, this gesture is consistent with risk management but, does not prove that the credit card market is as competitive as it should be. hi. the pricing of bank credit cards rose (1985) documented several models for pricing business and consumer loans (includ ing credit card loans) by banks. the lenders would charge a rate that incorporates risk as well as a reasonable level of profit. on the other hand, the loan rate should be low enough to accommodate the customer who must think of how to repay the loan. the models docu mented by rose are: cost-plus loan-pricing model, price-leadership model, cost-benefit loan pricing, customer profitability analysis and present value loan pricing. more impor tantly, lown and peristiani (1996) examine the loan pricing behavior of commercial banks and report that the average consumer borrowing rate is highly eon'elated with treasury rates. thus, one expects that in a deregulated f'mancial system, the cost-plus loan-pricing proposed by rose has practical usefulness. with the cost-plus loan-pricing method, the annual percentage rate (apr) of interest is defined as: ratet ~0costt + fjlgt + 62 rt + ~3i-lt + et (1) according to rose (1985), the loan interest rate or apr (rater) is the sum of the marginal cost of raising loanable funds (costt), nonfunds bank operating costs (kt), esti 56 financial services review 6(1) 1997 mated margin to protect the bank against default risk (rt) and the bank's desired profit mar gin (l'lt). in a comprehensive analysis of the u.s. credit card market, ausubel (1991) acknowledges the apparent stickiness in credit card rates. he argues that the cost of funds should be the primary determinant of the marginal cost of lending through credit cards. the intuition behind his model is consistent with the aforementioned rose's pricing model. consequently, one expects a relatively high level of correlation between the interest rate charged on credit cards and the bank's cost of funds. ausubel reports that between 1982 and 1989, the volatility of credit card rates is less than one-fifth that of treasury bills rate. the same sentiment has been expressed by demuth (1986), pozdena (1991), sinkey (1992), mester (1993), and duca and whiteseh (1995). ausubel (1991) argues that the primary determinant of credit card rates is the cost of funds to card issuers. he proceeds to test the responsiveness of credit card rates to changes in the cost of funds, defining the cost of funds as the quarterly one-year treasury-bill yield plus 75 basis points and concludes that credit card rates are sticky. he concludes that the credit card market fails to pass the tests of a competitive model if one focnsses on price responsiveness to costs and zero excess profits. furthermore, the author suggests the pres ence of adverse selection problem, search and/or switch costs and consumer irrationality in the credit card market. the credit card industry is susceptible to both search and switch costs. the costs include information cost of discovering which banks are offering lower interest rates, cost in time, effort, and emotional energy in filling out an application for a new card and the time lag between applying and receiving one. the fact that the card fee is usually billed on an annual basis, so that if one switches bank at the wrong time, one forgoes some money. there is also the perception that one acquires a better credit rating or a higher credit limit by holding the same bank card for a longer time. calem and mester (1992, 1995) analyze the dilemma facing undisciplined card users as follows. they do not intend to borrow on their cards, but always find themselves in debt. once in debt, they believe the situation will be short-lived and consequently are not moti vated to search. there are other cardholders who are bad credit risks because they carry substantial amounts on their credit cards and thus have no choice but to hold onto their credit cards. for this group of cardholders, it difficult if not impossible to switch from one credit card to another because of the negative effect of existing outstanding balances. while credit card consumers undoubtedly face some positive level of search and switch costs, there remains an empirical question as to whether the actual search/switch costs are of sufficient magnitude to justify what is observed. the adverse selection argument implies that interest rate should not be used as an instrument for competition since it becomes much more difficult for credit card issuers to compete away profits. thus, adverse selection helps to explain the observed extraordinary profits reported by most of the card issuers. mester (1993) models a consumer credit market in which banks use collateral to screen borrowers. in her model credit card rates stickiness is shown to be consistent with rational behavior on the part of issuers. mester's model incorporates asymmetric information between consumers and banks, regarding consumers" future incomes. the model offers an explanation why low-risk consumers select collateralized loans while high-risk consumers select credit card loan. more importantly, the recent movements in rates and the move of creditworthy customers to collateralized loans is said to be consistent with rational behavior. sticky credit card rates 57 iv. data and methodology a) data the data used in this study are the credit card rates (rate,) and treasury bill rates for february, may, august and november each year from august 1982 through august 1994 as reported in the federal reserve bulletin. in order to define a proxy for cost of funds (costt), i added 75 basis points to the respective one-year treasury bill rates. calem and mester (1995) and stavins (1996) report results indicating the appro priateness of the treasury bill rate as a proxy for cost of funds. more importantly, ausubel's (1991) results suggest that credit-card-backed securities have yields in the vicinity of 75 basis points above those of treasury seeurities with comparable maturi ties between 1987 and 1989. b) cointegrafion test and error correction model the two major econometric tests used in this study are cointegration test and error cor rection model (ecm). prior to the appfication of these techniques, the stationarity charac teristics of the time series data are established by applying the augmented dickey-fuller (adf), phillips-perron (p-p), and the kwiatkowski-phillips-schmidt-shin (kpss) tests. these three tests are applied to determine the consistency of the results therefrom. two of the aforementioned unit root tests (adf and p-p) are performed on the autoregressive equation 2. the lag selection is done by applying the modified akaike information crite rion (maic). the cointegration method assumes that ff the two variables rate t and cost t contain some stochastic trend, each can be described as an integrated variable. furthermore, if a linear combination, (say, rate t txcostt,) is stationary, the two variables are said to be cointegrated. as noted earlier, the adf, p-p, and kpss tests are used to establish the sta tionarity characteristics of each data series prior to the application of the cointegration method. the adf and p-p tests are performed on the following equation: q att -~ t + pz t_ l + ~ ~i~kzti + et (2) i = 1 where a ( l l ) e.g. ztffi(1-l)z, f z , z t _ 1 zt = series under consideration; t = time trend; and q = the lag is chosen such that it is large enough to ensure that e t is white noise. in equation 2 above, engle and yoo (1987) suggest the use of modified akaike infor. marion criterion (maic) in order to choose the optimal q. the maic is defined as: maic(k) -n (ssr/n) + 2k (3) 58 financial services review 6(1) 1997 where n = number of observations to which the model is fitted; ssr = the sum of squared residuals from an ols regression on equation 2; k = q+ 1 (number of parameters in equation 2); and maic(.) = modified akaike information criterion. the appropriate order of the model is determined by computing equation 2 over selected grids of values of q and choosing a q which minimizes maic(k). once the optimal lag (q) of equation 2 is selected, an ordinary least squares regression is applied to equation 2 in order to determine the order of integration of each time series. this is a necessary pre-test of the data series. the test statistic for adf test is the ratio of p to its calculated standard error obtained from the least squares regression. the decision rule is to reject the null hypothesis if p is significantly negative. the statistic derived here does not have a t-distribution, but dickey and fuller (1979) provide a table of significance levels. recently, mackinnon (1991) produces a replication of the underlying critical val ues which has wider applicability than those of dickey and fuller. in this case the hypoth esis of a unit root is rejected if the t-statistic lies to the left of the relevant mackinnon critical value. in addition to the adf test, the phillips-perron (p-p) test is applied to test the null hypothesis that each time series has a unit root. the specification of the p-p test is the same as the adf test except that the former adjusts for error autocorrelation. the third test is recently developed by kwiatkowski, phillips, sehmidt and shin (1992) [hereafter, kpss] to determine the stationarity property of any time series. this new test is more innovative in the way it makes allowance for error autocorrelation. more importantly, unlike other tra ditional tests, the null hypothesis under the kpss test is that the time series under exami nation is stationary. the test statistic is defined as: w h e r e kpss " n 2 f2~s, t= i (4) s2(l) n = number of observations; and s t = cumulative sum of the residuals ( ,~et ) from a regression of series on a constant s 2 ( l ~ n 1 n 2 l n = ~,e t + ~, ( 1 j / ( l + 1)) x etet_j ~ j t= 1 j = 1 t = j + 1 the null hypothesis of stationarity is rejected if the calculated kpss statistic exceeds the critical value. to test for cointegration, one needs to run the following 'cointegration regression': r a t e t -o~ -i~ c o s t t _ 1 + gtt (5 ) the null hypothesis that the residuals are integrated is then tested using the aug mented dickey-fuller statistic. if it is shown that rate t and cost t are cointegrated, the short-run dynamics can be described by an error correction model (ecm). this is known sticky credit card rates 59 as granger representation theorem. johansen developed a maximum likelihood estimator within a multivariate cointegration model. this procedure calculates and tests the number of cointegrating vectors in an ols setup. according to maddala (1992), the granger representation theorem implies that rate t and costt_ i may be considered to be generated by error correction models of the form: ~ t l t pl wtol + lagged (arate t, acostt.1) + el, t (6) acostt-1 = p2 wt-1 + lagged (arate r, acost t. 1) + e2, t (7) where wt = rater ot ~ costt_ 1 in the equations above, at least one of pl and p2 is nonzero and £1t and e2t are white noise errors. as noted above, the statistical notion of cointegration of two time series reflects a the oretical long-run equilibrium relationship between them. taylor (1988) observes that if economic theory suggests a long-run equilibrium relationship between rate t and cost t, then a linear combination of the two series will not only be stationary, but their cointegra tion is a necessary condition for them to have a stable long-run relationship. this suggests that the existence of a stable long-run relationship between two integrated variables means that they are also cointegrated. given that two time series are cointegrated, the ecms are appropriate when the depen dent variable is known a priori to exhibit short-term changes in response to changes in the independent variable (dun:, 1993). according to dun" (p. 165), "error correction models presume that there exists an equilibrium state in which the levels of the time series are typ ically located vis h vis one another." thus, if a shock disturbs the "moving equilibrium" by forcing the series to diverge from their long-run relationship, the equilibrium error is cor rected by pushing the relationship to a new level that captures the long-run equilibrium state. the error correction model is estimated within the vector autoregression framework in line with engle and granger (1987). the process of determining the long-run equilibrium relationship between credit card rates and cost of funds proceeds as follows. the first necessary pre-test is to establish the stationarity property of each time series noting that a time series that is integrated of order d (denoted i(d)), should be differenced d times in order to make it stationary. this study applies the adf, p-p and kpss tests at this stage. in the second stage, i apply johansen cointegration test to determine the long run equilibrium relationship between credit card rates and cost of funds. the final part applies the vector error correction estimates to study the adjustment process when the two cointegrated variables deviate from their long-run equilibrium relationship. v. empirical results a statistical summary of the sample data indicates that between 1982 and 1994 credit card issuers charge an average annual rate of interest of 18.02 percent while the average 60 financial services review 6(1) 1997 notes: 2 0 15. 10. 5 0 \ "-, . . . ii g l l l l l l l a l l j i l j l j l l l l l ~ * l l l * i l j l * l a ~ l l l i ~ l l l l i j l i 83 84 85 86 87 88 89 90 91 92 93 94 "war i l l rate ~,f'ers to credit card me costp mf'ers to one-ye.ar treasury bill rate plus 0.75% figure 1. credit card rate and the cost of funds. cost of funds in the economy was 7.43 percent. in terms of variability, credit card rates are less volatile with a standard deviation of 0.732 percent compared to 2.047 percent for cost of funds. moreover, the distribution of credit card rates is skewed to the left while one-year treasury bill rate's distribution is about normal. the two series are shown in fig ure 1. the characterization in this section is meant to describe the credit card market con ditions contemporaneous with the sample period rather than an effort to depict the current market conditions. table 1 reports the results of stationarity tests performed on credit card rate series and cost of funds. as indicated earlier, the modified akaike information criterion (maic) was employed in the selection of an appropriate lag (q) for each series. the adf, p-p and kpss tests are applied to level of the series (rate and cost) as well as the first difference of the series (arate and acost). the null hypothesis of unit root in rate and cost could not be rejected under the adf and p-p tests. however, one could not reject the null when the time series are differenced once (arate and acost). on the other hand, the kpss test could not reject the null hypothesis of sta tionarity for arate and acost. thus, the results from the adf, p-p and kpss tests show that each time series is integrated of order one. these results are robust even with the inclusion of a time trend in the respective test equations. this means that each series should be differenced once in order to achieve stationarity. since both series are $t/cky cre.d/t card rata 61 t a b l e 1 unit root test on rate and cost unit root test performed on rate arate cost acost p .0.011 -0.465 .0 .030 -0.621 adf staff stic -0.398 -3 .060" 1.275 -4.511" adf(trend) 1.719 -3 .019" -2.214 -4.462* p-p statistic 1.266 -3.606* -2.367 -5.053* p-p(trend) 1.266 -3.641" -2.397 -6.181" kpss statistic 0.524* 0.198 0.856* 0.271 kpss(trend) 0 .281" 0.055 0 . 5 2 1 " 0.102 notes: the unit root test is based on the following equation: q az t = ~,t + p z t_ 1 + ~ ~iat-ti + et i = 1 the null hypothesis of a unit root is based on the coefficient, p under adf and p-p tests. * i n d ~ ! ~ significance at the 5 pcavent level. adf statistic is the augmented dickey-fuller test slatistic. adf(trend) is the adf statistic from the equation above with a time lnmd. p-p statistic is the phiilips-p~,ou test statistic. p-~trend) is the p-p statistic from the equation above with a time trend. kpss statistic is based on the residuals from regression of variable under consideration on a constant term only. the critical value of this statistic at the 5 percent level is 0.463. kpss(trend) in based on the residuuls from mgn~ssioo of variable under consideration oo n constant and time utnd. the critical value of this statistic at the 5 percent level is 0.146. similar in terms of their stationarity properties, it is conceivable to think that they are cointegrated based on engle and granger (1987) representation theorem. the results of the johansen cointegrafion test are reported in table 2. the likelihood ratio test of the maximum eigenvalue indicates only one cointegrating equation between credit card rates and cost of funds. the implication is that credit card rates and the cost of funds have an equilibrium condition that keeps them in proportion to each other in the long-ran. the results here indicate that between 1982 and 1994, credit card rates have a consistent long-rim eqnilibrium relationship with the cost of funds. these results are con sistent with those of brito and hartley (1995) who report that credit card interest rate and the 6-month cd rate are cointegrated. the estimated long-run or cointegrating relationship is of the form rate t = 14.703 + 0.448 costt_ 1. this relationship implies that the long-run spread between ratet and costt_ 1 is 14.703 percent less 0.552 costt_ 1. the implication of this is that the long-run spread increases as the cost of funds falls and decreases as the cost of funds rises. the preceding results indicate that credit card rates and cost of funds are cointe grated. therefore, an error correction model can be fitted to determine the short-term changes in the response of credit card rates to changes in the cost of funds. in table 3, the coefficient of the cointegrating equation (cointeql) measures the rate at which dise quilibria in the long-term relationship are corrected. in other words, it is the speed of 62 financial services review 6(1 ) 1997 table 2 johansen cointegration results zero c .~ at most one c.e. eigenvalue likelihood ratio normalized cointergrafing equation: rater 0.253 0.098 18.207" 4.761 14.282 0.504cost t_ 1 notes: *rejection of the null hypothesis at the 5 percent level. c.e. = cointegrating equation. likelihood ratio test = 1 cointegrating equation at the 5% level. adjustment of any disequilibrium towards a long-run equilibrium state. thus, a number of-0.146 reported in table 3 (column 2) means that eq!filibrium errors are corrected in the short-run at the rate of about 15 percent per quarter. the negative sign associated with the error correction term assures that both credit card rates and cost of funds con verge to their long-run equilibrium, following a response to innovation shocks. the evi dence here indicates that credit card rates adjust relatively slowly to equilibrium errors. based on these results, the single-equation error correction model of credit card rates is of the form: arate t = 0.044 + 0.321arate t_ 1 0.146 (ratet 1 0"448costt2 14.703 ) + e t (8) the reslxicted vector autoregression (var) approach is used here in an effort to cor rectly specify the lag lengths for the short-term effects (johansen, 1988). from the equation above, it is obvious that there is no short term dynamics between rate and cost, because the coefficients of acostt_ 2 and acostt. 3 are not statistically significant (see table 3). moreover, the error correction mechanism also shows that changes in the credit card rate is influenced by previous changes. the granger causality test results also confirm that the treasury bill rate (cost) has a strong predictive power for explaining credit card rates (rate). incidentally, the credit card rate does not granger-cause the cost of funds as defined in this study. an interesting part of this result is the strong exogeneity of the cost of funds. the f-test of the lagged credit card rates suggests that cost of funds is weakly exogenous. the statis tical insi,,l~aificance of the coefficient of the error-correction equation reported in table 3 (column 3) is also evidence of a weakly exogenous cost of funds. an exogeneity test is carried out following kwan and kwok (1995). an f-statistic of lagged credit card rates of 3.023 confirms the weak exogeneity of the cost of funds. in addition to this, an f-statistic of 3.734 from the pairwise granger causality test leads one to reject the hypothesis that the cost of funds does not granger cause credit card rates. moreover, an f-statistic of 0.437 could not reject the hypothesis that credit card rate does not granger cause cost .of funds. kwan and kwok note that strong exogeneity includes weak exogeneity and granger noncausality (p. 1161). thus, one can conclude that the cost of funds is strongly exogenous. the econometric implication of an exogenous cost sticky credit card rates 63 table 3 vector error correction estimates cointegrating equation cointeq l coefficients ratet. 1 costt. 2 constant 1.000 -0.448 (0.093) [-4.801]* 14.703 error correction aratet. 1 acostt. 2 cointeq i -0.146 -0.149 (0.043) (0.266) [-3.378]* [-0.561] 0.321 0.757 aratet-l (0.147) (0.908) [2.177]* [0.833] -0.028 1.589 aratet-2 (0.141) (0.871) [-0.195] [-1.825] -0.026 -0.082 acostt-2 (0.031) (0.192) [-0.842] [0.429] -0.014 -0.106 acostt-3 (0.028) (0.169) [-0.512] [0.062] constant -0.044 -0.130 (0.019) (0.120) [-2.268 }* [1.087] adj. r-squared 0.395 -0.006 notes: *indicates significance at the 5 percent level. of funds is that there is no feedback effect from credit card rate to cost of funds. the relationship between the variables is unidirectional with credit card rate being the dependent variable and the cost of funds as the independent variable. the dynamic relationships between cost of funds and credit card rates can best be appreciated by examining the impulse response functions. impulse responses represent the dynamic response of the level of endogenous variable to innovations in disturbance terms. table 4 reports the dynamic responses of the level of each variable to one standard deviation innovation shock in equations 6 and 7. a credit card rate innovation shock will increase the movement in level of the credit card rate over several quarters. however, a cost of funds innovation shock will make credit card rate to rise and later makes a down ward movement toward its equilibrium level. on the other hand, the cost of funds responds only to its own innovation shocks. its response to innovation shocks from credit card rates occurs only within four quarters after which the cost of funds returns to its long-run equilibrium level. 64 financial servicf.3 review 6(1) 1997 table 4 impulse response to one standard deviation innovation period response of credit card rate to: credit card rate innovation shock cost of funds innovation shock 1 0.098 0.000 2 0.121 0.023 3 0.119 0.060 4 0.108 0.105 5 0.094 0.148 6 0.077 0.185 7 0.061 0.213 period response of cost of funds to: credit card rate innovation shock cost of funds innovation shock 1 0.149 0.584 2 0.231 0.671 3 0.081 0.675 4 0.009 0.695 5 -0.001 0.702 6 -0.002 0.685 7 -0.002 0.661 ordering: rate, cost of funds vi. commentary meyercord (1994), sinkey and nash (1993), ritzer (1995), stavins (1996) and lown and peristiani (1996) document evidence suggesting that significant changes are taking place in the credit card market. more specifically, there is increased participation in the market by nonbank card issuers, the overall cost of funds and interest rates ate declining; credit losses are rising; and credit card issuers are becoming increasingly sophisticated in targeting right value propositions. all of these imply that the credit market might not be insensitive to the cost of funds afterall. in view of the aforementioned, this study examines the relationship between credit card rates and the costs of ftmds from 1982 through 1994. the evidence from the johansen cointegration test confirms that credit card rates and cost of funds possess a long-run equi librium relationship with one another. furthermore, the error correction models are indic ative of the slow rate at which credit card rates adjust to the cost of funds. between 1982 and 1994, credit card rates adjust to the cost of funds at about 15 percent per quarter. more importantly, the granger causality results suggest that the one-year treasury bill rate has a strong predictive power in explaining credit card rate. sticky credit card rates 65 in a recent study, lown and peristiani (1996) report that after accounting for funding costs, the premium charged on consumer loans by low-capitalized banks is highly signifi cant. stavins (1996) also finds a high correlation coefficient between apr and the fraction of overdue loans suggesting that high-apr plans are associated with high rates of delin quency. these results represent indirect evidence for the validity of the cost-plus loan pric ing model proposed by rose (1985). given rose's model, the cost of funds may not be the most important determinant of credit card interest rate as ausubel (1991) argues. stavins' results imply that risk factor is another major determinant of credit card rates. therefore, it is logical for the credit card rates to respond sluggishly to the costs of funds as reported in this study. the results of this study are consistent with the financial contracting literature on asymmetric information and the associated adverse selection consequences as discussed in berlin and loeys (1988) and best and zhang (1993). in the credit card market, calem and mester (1995) argue that credit card issuers face adverse selection from the search behavior of cardholders and also the cost associated with switching. the presence of asymmetric information between banks and consumers in respect of consumers' probability of default based on their future incomes leads banks to use collateral to screen loan applicants. in general, high-risk customers will prefer credit card loans to collateralized bank loans. on the other hand, low-risk consumers prefer collateralized loans. according to mester (1993), when the cost of funds goes down, low-risk consumers would naturally abandon the credit card market and use the collateralized loan market. the credit card market would be dominated by high-risk consumers. therefore, credit card issuers are not motivated to drop rates in response to a decline in the cost of funds. the implication of this phenomenon, according to mester, is that the spread between the rate of interest on credit card and the cost of funds will be positively correlated with the default rate. this argument represents the position of the proponents of adverse selection as an explanation for nonresponse of card rate to the cost of funds. the results reported by sullivan and worden (1995) suggest that the estimated value of cardholders' option to default is significantly higher for bankruptcy than outfight default. this situation is partic ularly captivating in light of the loan pricing equation 1. any decline in the cost of funds leads to an increase in default risk. consequently, the overall loan rate (credit card rate) is relatively inflexible to changes in the cost of funds. calem and mester (1992) have also written about borrowers' switching costs. once a consumer carries a substantial amount of debt on a credit card, it is difficult for the con sumer to transfer from one issuer to another. the rationale is that the new issuer would con sider the existing debt as a negative strike against the consumer. calem and mester observed that the switching cost argument is consistent with nonprice competition in the credit card market as we currently witness. there is a pool of consumers who are reluctant to search for the card that offers the best rate of interest. the undisciplined card users do not always plan to borrow but find themselves in debt because of inability to settle their card bills. because this group of card users do not plan to exploit the credit feature of their cards, they are not motivated to search. they always believe that their peculiar situation will be short-lived. this popula tion of cardholders also includes those who refuse to search because of the fear that such an action would reflect negatively on their credit reports. in view of the aforementioned, the large pool of potential high-risk card holders moti vates credit card issuers to charge rates that reflect maximum risk. with reference to equa 66 financial services review 6(1) 1997 tion 1, one can observe that as the cost of funds (costt. 1) goes down, the risk factor (r t) increases because the low risk cardholders seek collateralized loans. the pool of cardhold ers now consists of mostly the high-risk customers. thus, the downward pressure on credit card rates from a decrease in the cost of funds is offset by an upward pressure from an increase in the riskiness of the credit card issuer's portfolio. all other things being equal, the overall impact of these two forces on credit card rates will be insignificant. this study shows that credit card rates and credit card funding costs are cointegrated and that credit card rates adjust to changes in funding costs in a sluggish fashion. these results are only applicable to the 1982-1994 period. however, the results explain why past empirical work has failed to identify a significant relationship between credit card interest rates and costs of funds, and they offer empirical support for alternative explanations why credit card rates appear invariant to funding costs in the financial economics literature. moreover, within rose's cost-plus loan-pricing framework, the credit card market can indeed be characterized as a competitive market. the difficulty in measuring the competi tive nature of the market is related to a set of specific institutional characteristics at work in the marketplace. finally, the presence of asymmetric information, adverse selection due to search and switch costs can obscure the relationship between credit card interest rates and the costs of funds. acknowledgment : this research was supported by a research grant from fayetteville state university's faculty development committee. the author is grateful for this. the author also appreciates helpful comments from the editor of this journal and the anony mous reviewers. references ayadi, o.f., & onashile, t. (1994). taking your plastic card to the limit. working paper, fay etteviue state university. ausubel, l. (1991). the failure of competition in the credit card market. american economic review, march, 50-81. berlin, m., & mester, l.j. (1989). credit card rates and consumer search. working paper no. 88-1/ r, federal reserve bank of philadelphia. berlin, m., & loeys, j. (1988). bond covenants and delegated monitoring. journal of finance, 43, 397-412. best, r., & zhang, h. (1993). alternative information sources and the information content of bank loans. journal of finance, 48, 1507-1522. brito, d.l., & hartley, p.r. (1995). consumer rationality and credit cards. journal of political economy, 103(2), 400-433. calem, p.s. (1992). the strange behavior of the credit card market. frbp business review, janu ary-february, 3-14. calem, p.s., & mester, l.j. (1992). search, switching costs, and the stickiness of credit card interest rates. working paper no. 92-24, federal reserve bank of philadelphia. calem, p.s., & mester, l.j. (1995). consumer behavior and the stickiness of credit-card interest rates. american economic review, december, 1327-1336. canner, glenn r., & fergus, james t. (1987). the economic effects of proposed ceilings on credit card interest rates. federal reserve bulletin, january, 1-13. canner, g.b., & luckett, c.a. (1992). developments in the pricing of credit card services. federal reserve bulletin, 78(9), 652-666. sticky credit card rates 67 cames, w.s., & slifer, s.d. (1991). the atlas of economic indicators. new york: harpcrbusiness. demuth, c.c. (1986). the case against credit card interest regulation. yale journal of regulation, 3(201), 201-242. dickey, d.a., & fuller, w.a. (1979). dislribution of the estimators for autoregressive time series with a unit root. journal of the american statistical association, 74, 427-431. duca, j.v., & whitesell, w.c. (1995). credit cards and money demand: a cross sectional study. journal of money, credit and banking, 27(2 may), 604-623. duff, r.h. (1993 ). what moves policy sentiment7 american political science review, 87(1 ), 158-170. engle, r.f., & granger, c.w.j. (1987). cointegration and error-correction: representation, estima tion and testing. econometrica, 55, 251-276. engle, r.f., & yoo, b.s. (1987). forecasting and testing in co-integrated systems. journal of econometrics, 35, 143-159. federal reserve board. (1994). recent developments in credit card pricing. federal reserve bulle t~n, apr/l, 296-301. johansen, s. (1988). statistical and hypothefis testing of cointegration vectors. journal of economic dynamics and control, 12, 231-254. kwan, a.c.c., & kwok, b. (1995). exogeneity and the export-led growth hypothesis: the case of clfin& southern economic journal, 61(4), 1158-1166. kwiatkowski, d., phillip, p.c.b., schmidt, p., & shin, y. (1992). testing the null hypothesis of sta tionarity against the alternative of a unit root. journal of econometrics, 54, 159-178. lown, c., & peristiani, s. (1996). the behavior of consumer loan rates during the 1990 credit slow down. journal of banking and finance, 20, 1673-1694. mackinnon, j.g. (1991). critical values for cointegration tests. in eagle, r.f., & granger, c.w.j. (eds.), long-run economic relationships (pp. 267-276). oxford university press. maddala, g.s. (1992). introduction to econometrics, 2rid ed. englewood cliffs, nj: prentice hall. mandell, lewis. (1990). the credit card industry: a history. boston: twayne publishers. mester, loretta j. (1993). why are credit card rates sticky? working paper no. 93-16, federal reserve bank of philadelphia. meyercord, a. (1994). recent trends in the profitability of credit card banks. federal reserve bank of new york quarterly review, 19(2 summer-fall), 107-111. pozdena, randall j. (1991). solving the mystery of high credit card rates. frbsf weekly letter, november 29, 91-42. ritzer, g. (1995). expressing america: a critique of the global credit card society. thousand oaks, ca: sage publications. rose, p.s. (1985). loan pricing in a volatile economy. canadian banker, 92(5 october), 44-49. rose, sanford. (1990). the coming revolution in credit cards. american banker, april 4, 4-15. shay, r.p. (1987). bank credit card pricing: is the market working? journal of retail banking, 9(1), 26-32. shepherdson, nancy. (1991). credit card america. american heritage, november, 125-132. sinkey, joseph f. (1992). commercial bank financial management. new york: macmillan. sinkey, j.f., & nash, r.c. (1993). assessing the riskiness and profitability of credit-card banks. journal of financial services research, 127-150. stavins, j. (1996). can demand elasticities explain sticky credit card rates? new england economic review, july/august, 43-54. sullivan, a.c., & worden, d.d. (1995). credit cards and the option to default. financial services review, 4(2), 123-136. taylor, m.p. (1988). an empirical examination of the long-run purchasing power parity using coin tegration techniques. applied economics, 20, 1369-1381. worthington, s. (1995). the cashless society. international journal of retail & distribution man agement, 23(7), 31-40. pii: s1057-0810(97)90034-2 book, software, and web site reviews 71 the updated information on the importance of living wills and powers of attorney in part 6, "retirement and estate planning," is most welcome. a few more examples of how the different trusts in exhibit 15.4 can minimize estate taxes would be helpful. the fact that many people die intestate is well-covered in this section; however, the use of "sweetheart" wills without appropriate trusts needs to be mentioned. the ancillaries to this text include a student workbook, a bound package of blank forms, software that can be run only on ibm-compatible computers, and a comprehensive instructor's manual and test bank. the outlines in the student workbook are more useful than the topic outlines in the instructor's manual. helpful case studies, problems, and vocabulary exercises are included in the student workbook; however, answers to the "'con cept check" questions are unfortunately not included in the workbook for the students to self-check. these answers are included in the instructor's manual along with key concepts, supplemental class project ideas, and answers to discussion questions, case problems, and the integrative case study. a weakness of the instructor's manual appears to be the test bank which offers objective questions that are generally less challenging than the subject matter merits. in conclusion, gitman and joehnk have produced another stellar edition of personal financial planning. it is evident that these authors have listened to the suggestions of their audiences and incorporated many positive changes into their seventh edition. the merrill lynch web site (www.ml.com) reviewed by: john grable, cfp, doctoral student, virginia tech internet users face two annoying problems when using web sites sponsored by the major brokerage firms, namely, slow information retrieval and advertising overload. the first problem, accessing information quickly, is directly related to the user's computer capabilities and inversely related to the number of bells, whistles, and graphics the web site offers. the second problem, advertising overload, is becoming so common that users are forced to browse through product information in order to find useful data. many internet users have concluded that the need for timely and accurate information retrieval often out weighs the benefits provided by fancy graphics and excessive advertising. after subjecting several brokerage firm-sponsored web sites to extensive review based on their ease of use, speed on information retrieval, breadth of information provided, and usefulness of data to practitioners, researchers, investors, educators, and students, the web site sponsored by merrill lynch (www.ml.com) emerged as one deserving further review and use by readers of financial services review. the merrill lynch home page downloads very quickly offering users the choice of five linking pages. financial professionals and investors will find the "financial news & research" ink an easy way to obtain stock quotes (provided on a 20-minute delay during trading hours), market updates, research publications, and legislative updates. the ability to get a glimpse of recent research reports on topics like equities, fixed income strategies, and market commentaries (provided three times daily) are enough to make this web site interesting. the "washington watch" link offers a unique view of legislation that may impact investors, both individuals and institutions. the investor learning center link is an excellent source of basic investing terms and concepts that will interest educators, researchers, and students. obviously merrill lynch 72 financial services review 6(1) 1997 makes a pitch for its services throughout this link, but users can scan recommended per sonal budgets, balance sheets, and financial strategies easily and anonymously. the "per sonal finance center" link helps users understand financial issues facing most american individuals and families and how certain financial planning techniques can be used to meet financial objectives. users interested in finding objective information on financial plan ning topics like savings, college expenses, credit usage, insurance, mortgages, taxes, retire ment, mutual funds, and making a financial plan will love this link. family and consumer economists will also find a useful discussion and analysis of the individual financial life cycle presented here. a "business planning" link is available for those interested in learning more about maximizing business cash flows, providing competitive employee benefits, obtaining financing, and institutional investing. another feature that is unique to this web site is a key work search that allows users to query the merrill lynch database for interesting facts, fig ures, and research. researchers, educators, and students will fred this feature an invaluable source for data. compared to other brokerage sponsored web sites, merrill lynch's web page lacks graphical excitement (you can, however, download audio transcripts of each linked page), but for financial professionals, investors, researchers, educators, and students in need of solid information in a timely manner, this site fills the bill perfectly. another useful feature to note is that merrill lynch provides this service without obligation, cost, or a lot of self promotion, and in today's web world that's saying a lot. charter media's briefing.com web site reviewed by: robert l. albert jr., assistant professor of finance, morehead state university individual investors can now find a multitude of web sites which provide a wide array of relevant and timely financial market information. many sites are devoted to security prices and many others are devoted to fundamental characteristics of the firms. charter media's briefing.corn (http://www.brief'mg.com) is one site which provides security price information and fundamentals along with a real time commentary on the financial markets. briefing.corn is not a site burdened with many slow-loading graphics, so users can get to different levels of information quickly. briefing.corn provides comprehensive reports of economic, industry, and company news giving investors access to the relevant information necessary for a top-down approach to investing. charter media's research staff is comprised of former senior man agers and analysts from standard and poor's mms international who provide insightful interpretations of braking economic and industry developments. at the macroeconomic level, briefing.corn provides updates of the major economic variables along with forecasts of future economic activity. in their political brief, they provide a daily commentary on those political issues which are likely to have an impact on the financial markets. in addition, briefing.corn provides updated fed briefs which given an overview of fed policy and possible interest rate moves. briefing.corn's industry reviews provide a timely analysis of the major factors affect ing specific industries and highlight timely stocks within those industries. in alliance with retirement income beliefs and financial advice seeking behaviors alejandro murguı́aa, wade d. pfaub,* amclean asset management, 1900 gallows road, suite 350, tysons, va 22182, usa bmclean asset management, the american college of financial services, 1900 gallows road, suite 350, tysons, va 22182, usa abstract this investigation identifies and validates a series of salient behavioral finance and psychological constructs that influence retirement income planning. we show how these scales relate to each other as well as retirement income concerns and investment behaviors. we also describe how four investment personas can be linked with the advisor usefulness and retirement income self-efficacy scales to successfully identify preferred financial implementation methods. this can assist individuals in more readily recognizing their relative strengths and weaknesses when implementing a retirement income strategy, and financial professionals can present advice in a manner that addresses a client’s concerns and preferred implementation. © 2022 academy of financial services. all rights reserved. jel classifications: d14 keywords: retirement income; self-efficacy; financial advisor perceived usefulness; behavioral finance 1. introduction while it is generally accepted that irrational behaviors influence general financial decisions, research detailing how various psychological constructs affect financial behaviors specific to retirement income planning has lagged. moreover, advances from behavioral finance and psychology have progressed in a parallel manner reflecting the different academic departments that help to identify these constructs. the interplay between behavioral finance, *corresponding author: tel.: +1-610-526-1569; fax: +1-703-556-8628. e-mail address: wade@retirementresearcher.com 1057-0810/22/$ – see front matter © 2022 academy of financial services. all rights reserved. financial services review 30 (2022) 1–29 psychology, and retirement planning is one that should be examined to better identify and promote successful financial behaviors while mitigating the negative ones. after we briefly address how various behavioral finance and psychological constructs have been linked to general financial behaviors, we detail the development of specific selfefficacy, financial bias, numeracy, and advisor usefulness scales. while these are generally accepted social science constructs, these scales were created with an increased level of specificity to retirement income. we assess their ability to reliably quantify these factors. we further examine how these factors associate with each other, retirement income concerns, investment behaviors, overall retirement income outlook, and the use of an advisor. this presents a further indication of construct and criterion validity for the scales (devellis, 2017). the results identify four financial implementation personas in relation to their retirement income self-efficacy and perceptions about the usefulness of financial advisors. these personas link to various behavioral finance constructs, retirement income concerns, investment behaviors, and retirement outlooks. this provides a framework to identify individual preferences for receiving financial advice and avenues that maximize those preferences. it also addresses the potential areas of strengths and weaknesses for the different types of retirees. altogether, this supports greater retirement income planning success. 2. literature review we provide a brief review of the constructs that will be included in our investigation. first, tversky and kahneman (1974) identified that individuals have two approaches to their decision-making process. one is predicated on a quick and simple heuristic approach via mental shortcuts, and another is deeply analytical and measured. while many mental shortcuts are very adaptive for everyday living, they are frequently maladaptive when making personal financial decisions. additionally, loss aversion is a significant behavioral finance construct that permeates throughout the personal finance field. it is not only the idea that losses are more psychologically impactful than gains, but that individuals evaluate these gains and losses relative to a reference point (kahneman & tversky, 1979). our tendency to rely on heuristics for financial decisions and the effect of loss aversion leads to many wellknown financial biases including hindsight, recency, survivorship, affinity, gambler’s fallacy, and the endowment effect. other behavioral finance constructs have started to indicate potential avenues for insight. inertia has been advanced as a foundational contributor to explaining financial behavior (gal, 2006). the concept of inertia indicates that individuals tend to maintain the status-quo. to change this baseline for the status quo, there must be an improvement, not just a substitution, to the current situation that makes the effort for change worthwhile. this push past indifference also requires that alternatives be posed as clear choices. the key dynamic for differentiation between loss aversion and inertia is the push and pull between action versus inaction (e.g., inertia) as opposed to the psychological valence of a gain versus a loss (e.g., loss aversion). gal (2006) posits that inertia is a more impactful construct than loss aversion as a keystone principle in behavioral finance. 2 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 numeracy is the ability to comprehend numerical concepts such as probabilities and other mathematical procedures. in the field of personal finance, it is frequently referred to as financial literacy or risk literacy. one’s level of financial literacy has been identified as having a major impact into financial decision making (lusardi & mitchell, 2014). financial literacy has also been associated with an inability to understand the impact of portfolio volatility on investment returns (newall, 2016). low financial literacy is a pervasive global observation (kell, 2014) and unfortunately only 57% of americans made a passing grade in a standard financial literacy test (zumbrum, 2015). numeracy studies within the medical field has also found that older adults experienced difficulty using numerical information to compare medicare health plans (hibbard et al., 2001). numeracy is an important personal consideration when developing a retirement income plan. agarwal and mazumder (2013) point out that one’s financial choices may not be optimal due to low proficiency or a general avoidance of mathematical concepts. while numeracy is the objective measure of mathematical competence, few studies have investigated the differences between numeracy and perceived numeracy. balasubramnian and sargent (2020) found that discrepancies between perceived and objective financial literacy lead to weaker financial decisions. they conclude that the gap between the two is of great importance when investigating consumer financial behavior. in addition, perceived financial literacy can be as or more important than actual financial literacy in influencing financial behaviors (allgood & walstad, 2016). unfortunately, this overestimation gap is frequently observed in individuals scoring lowest on numeracy tasks. kruger and dunning (1999) observe that this occurs due to the double burden of lacking the general capacity to make sound choices and that this incompetence hinders the metacognitive ability to realize it. this construct is widely recognized as the dunning-kruger effect. while a meta-analysis of over 201 studies indicated that increasing financial knowledge has little impact on financial behaviors (fernandes, lynch, & netemeyer, 2014), kruger and dunning (1999) found that addressing these blind spots (overestimation gap) by improving individual skills helped participants recognize their shortcomings more effectively. self-efficacy is a psychological construct espoused by bandura’s social cognitive theory (bandura, 1977). it is the conviction of how well one can successfully execute a specific course of action. it differs from general confidence because it represents an individual’s perception of competently achieving more localized tasks. hence, a high degree of self-efficacy in one domain does not imply a high degree in another area. due to this domain specificity, it is important to measure self-efficacy in the field of study (bandura, 1997). lown (2011) created a general financial self-efficacy scale to measure the behavioral aspects of personal financial management. the scale was positively associated with a high level of confidence to manage money. asebedo and seay (2018) observed financial self-efficacy to be positively related to savings behavior after accounting for various demographic variables. in addition, asebedo and browning (2020) found portfolio withdrawal rates to be associated with financial self-efficacy. overall, financial self-efficacy is an important construct in helping promote effective retirement income planning behaviors. many studies try to identify the additive benefits that a financial professional provides to one’s overall investment return. blanchett and kaplan (2013) quantify how advisors have positively impacted retirement income decisions via improved financial planning decisions. vanguard has also identified various value-added factors that an advisor can potentially a. murgua and w. pfau / financial services review 30 (2022) 1–29 3 provide to improve an individual’s financial standing (kinniry et al., 2015). while these studies identify the benefits of an advisory relationship, a lack of fee transparency and general mistrust may lead to an underutilization of financial professionals. while there is a lack of availability for financial advice across the general population, among those who have access to advice, there potentially remains skepticism about an advisor’s overall value. therefore, assessing one’s belief about the cost effectiveness of a financial advisor may indicate who is most likely to use one. the five-stage model for advice seeking behavior posited by grable and joo (1999) theorizes that individuals assess the benefits and cost of engaging in a range of retirement planning activities (marsden, zick, & mayer, 2011). with the increasing popularity of various self-directed methods and business models to implement a retirement income plan, such as automated investment strategies and hourly financial planners, matching one’s financial implementation approach (using or not using a financial advisor) to appropriate methods can positively influence retirement outcomes. 3. method the methodology includes several steps. first, we discuss scale construction. we utilize exploratory factor analysis with a varimax rotation across the four scales to assess if they present as valid and meaningful constructs. cronbach’s a further analyzes internal reliability and helps determine the final question set for each scale. additionally, the pearson correlation coefficient reviews retest reliability among the scales. these steps indicate the degree of content and construct validity for the scales (devellis, 2017). for the second part of this investigation, we examine bivariate correlations between the scales to further assess construct validity and measure criterion validity. furthermore, we assess multivariate relationships between these constructs against various dependent variables that relate to investment behaviors, perceived retirement risks, overall retirement outlook, and advice implementation to determine predictive validity (devellis, 2017). this analysis is conducted using ordinary least squares regressions for continuous dependent variables and logistic regression for binary ones; all coefficients in the regression analyses are standardized. we create a financial implementation matrix centered on advisor usefulness and retirement income self efficacy scales to indicate how to identify preferred financial planning implementation methods. all results are computed using sas. 3.1. constructing the scales after reviewing various sources related to our constructs surrounding retirement income, we created 157 questions to be tested. these questions were reviewed by roughly 350 volunteers to provide initial feedback. these volunteers included a mix of financial professionals and individuals who are active readers of retirementresearcher.com. this website largely focuses on retirement income planning topics. these volunteers provided feedback and suggestions about the questions in terms of their quality, clarity, and conciseness. the focus at this stage was on content validity. 4 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 this feedback helped reduce the number of questions to less than 90, which were then provided to the participants of this study. participants were recruited as a convenience sample of 1,478 individuals from the same source of retirementresearcher.com readership. participants were asked to complete an online questionnaire. they were given 14 days to complete the survey during the month of july 2019. as an incentive for participation, we offered them a retake of the final questionnaire once the analysis was completed. we also provided reports of their results. while total participants peaked at 1,478, the number of specific completed surveys by topic varies because participants could drop out at any point, and some had left before all topics had been introduced. as well, answering a question about net worth was optional and reduced the number of respondents available to use in the regression analysis. for retest purposes, the finalized survey was administered on march 27, 2020, and again on september 10, 2020, roughly six months apart. it should be noted that due to the coronavirus disease 2019 (covid-19) pandemic, this period experienced pronounced market volatility and most likely great personal uncertainty for the respondents. while the sample is not intended to be random and reflective of the population at large, it is indicative of individuals for whom retirement income is a salient personal topic. many of those taking the survey can be viewed as taking an active interest in retirement planning topics and being more knowledgeable about retirement income than the average layperson. 3.2. descriptive statistics and exploratory factor analysis with 1,478 total participants, a power analysis indicated that the sample size was well above the number of participants needed to test our hypotheses with the exploratory factor analysis even as some participants did not complete every iteration of the study. in addition, we captured other information such as age and net worth to control for these additional variables during the subsequent regression analysis. descriptive participant data are provided in table 1. we were able to attract participants in which retirement is a relevant life milestone. for example, 61% (n = 845) of participants were in between 59 and 70 years old. males represent 77% of responses (n = 1,143) and females 23% (n = 335). 86% of the respondents are married (n = 1,270) and 14% (n = 208) are single. while an optional question, 50% (n = 372) of the 740 respondents report a net worth of $1-3 million dollars. descriptive statistics for all scales are presented in table 2. these scales are divided between psychological and behavior finance constructs, retirement income concerns, investment behaviors, and retirement outlook. the table presents means and standard deviations and a range of scores. higher scores within the range represent the degree of strength for the construct being measured. for financial biases, retirement income self-efficacy, and advisor usefulness, we will also refer to subsequent tables with exploratory factor analysis to determine whether these scales reflect our distinctly hypothesized constructs, cronbach’s a to determine internal scale consistency, and pearson correlation retest scores to indicate temporal consistency (devellis, 2017). a. murgua and w. pfau / financial services review 30 (2022) 1–29 5 table 1 demographic information respondents n % total participants men 1,143 77% women 335 23% age classes younger than 40 64 4% 40-46 64 4% 47-52 97 7% 53-58 267 18% 59-64 460 31% 65-70 385 26% 71-76 131 9% above 76 8 1% marital status spouse/partner 1,270 86% single 208 14% net worth range less than $500k 46 6% $500k $1m 91 12% $1m $2m 210 28% $2m $3m 162 22% $3m $4m 85 11% $4m $5m 48 6% greater than $5m 98 13% table 2 descriptive information for all scales variable n mean standard deviation minimum maximum psychological and behavioral finance constructs financial biases 1,058 2.13 0.48 1 3.75 inertia 1,002 1.27 0.40 1 2.5 retirement income self-efficacy 1,142 4.39 1.03 1 6 advisor usefulness 969 3.32 1.41 1 6 numeracy 1,025 5.25 1.64 0 8 numeracy self-awareness 1,025 0.18 0.27 �0.75 0.96 portfolio loss aversion 1,020 0.38 0.26 0 1 retirement income concerns lifestyle 1,209 3.90 0.94 1 6 longevity 1,175 2.63 1.30 1 6 liquidity 1,174 3.91 1.03 1 6 investment behaviors dividend agnosticism 1,008 4.10 1.29 1 6 non-forecasting versus forecasting 1,157 4.67 1.07 1 6 retirement outlook nest egg satisfaction 1,016 3.64 1.08 1 5 retirement income plan anxiety 1,016 2.21 0.73 1 4 6 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 3.3. financial biases and inertia we created statements reflecting various financial biases. specifically, we focused on hindsight, gambler’s fallacy, affinity, survivorship bias, herd mentality, endowment, and recency heuristics. we presented statements and asked the participant to select the degree to which each of the following statements best represents their opinion. the options were presented as a four-item likert scale ranging from strongly disagree (score of 1) to strongly agree (score of 4). for example, one potential question would read: “i have participated in popular investment strategies because i did not want to miss out on the opportunity.” the average financial bias score of 2.13. the scale midpoint is 2.5. participants perceive themselves with slightly below average levels of financial biases. we recognize the potential for a social desirability bias with this question set and discuss it in the results section. we measure inertia by asking how long it took a participant to act after acknowledging a need for an investment or financial planning adjustment. categorized responses for both questions ranged from 1 to 5: (1) less than three months; (2) three to six months; (3) six to 12 months; (4) one to two years; (5) more than two years. the mean inertia score of 1.27 indicates that it takes our participants, on average, a little more than three months to address any needed adjustments to their investments and plans. we include inertia in our exploratory factor analysis of our financial biases item set since inertia may be influenced by other heuristics. results in table 3 indicate that the different financial biases largely present as one overall factor with an eigenvalue of 7.24. while we expected the various financial biases to present as separate and distinct constructs, the data indicate that the varied biases manifest as a singular construct of overall financial heuristics. hindsight, affinity, gamblers fallacy, and survivorship biases present with the highest factor loadings in the first column of table 3. while herd mentality questions are present in the second factor construct, the items detailing this bias has significant cross factor loadings with the first and third factors as well. inertia did not cross load with the first factor structure and loads separately as its own distinct factor with an eigenvalue of 1.45. the data indicate that the financial biases present as a single generalized level of psychological noise. the higher the score on their financial bias scale, the greater the tendency to rely on heuristics for financial decisions. it is also interesting that our two-item inertia checklist loaded as a separate factor and did not combine within the first factor structure along with the majority of the financial bias items. this supports inertia as a separate construct from the other financial biases. the final financial bias item set is presented in bold. the financial bias scale has a cronbach’s a score of .82. with respect to retest, the pearson correlation coefficient is 0.80 (p < .0001). this supports excellent internal reliability and very good consistency between the time periods measured, especially during the emotionally charged time period measured as a result of the covid pandemic. 3.4. retirement income self-efficacy while there are self-efficacy scales that measure financial attitudes (lown, 2011), there is not one addressing the retirement income planning domain. lown’s (2011) six-item scale is a. murgua and w. pfau / financial services review 30 (2022) 1–29 7 limited to one-item detailing retirement income (i.e., i worry about running out of money in retirement.). our retirement income self-efficacy scale further expands the retirement income theme. we model our retirement income self-efficacy scale after badura’s guide for constructing self-efficacy scales (bandura, 2006). participants were shown statements detailing potential hurdles when implementing a retirement income plan. they were then asked to rate how certain they were in believing they could overcome each situation described. they rated themselves on a six-point likert scale ranging from strongly disagree to strongly agree. the mean self-efficacy score of 4.39 suggests that most participants felt fairly confident about their ability to implement their retirement income plan. this was table 3 exploratory factor analysis, cronbach’s a, and pearson correlation coefficient for financial bias scale factor loadings for all items** factors eigenvalues 1 7.24 2 1.55 3 1.45 4 1.38 5 1.18 6 1.09 7 1.05 8 1.04 financial biases cronbach’s coefficient a* pearson correlation coefficient* 0.82 0.8 hindsight 2 0.67 hindsight 1 0.66 hindsight 3 0.64 gambler's fallacy 1 0.63 affinity 4 0.61 gambler's fallacy 4 0.60 survivorship bias 2 0.60 survivorship bias 3 0.60 survivorship bias 1 0.56 herd mentality 4 0.56 survivorship bias 4 0.55 gambler’s fallacy 2 0.55 affinity 1 0.54 endowment 1 0.54 affinity 3 0.53 herd mentality 1 0.51 herd mentality 3 0.51 0.43 gambler’s fallacy 3 0.43 0.43 endowment 3 0.43 recency 3 0.40 endowment 4 recency 2 herd mentality 2 0.41 affinity 2 hindsight 4 0.48 recency 4 0.44 recency 1 inertia*** 0.65 inertia*** 0.42 0.62 endowment 2 note. *cronbach’s a and pearson correlation coefficients selected for the items in bold. they represent the final questions for each scale. **only factor loadings greater or equal to 0.40 are presented. ***used separately for the inertia checklist. 8 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 expected since the sample included many participants who have a personal interest in the domain of retirement income planning. results from the factor analysis in table 4 indicate that many questions sort within the first two factors with eigenvalues of 6.80 and 5.06. the dimension in the first column identifies selfefficacy with statements that addressed both the need to deal with the emotional influence and overall competence of organizing a retirement income plan. an example includes how well someone can resist the temptation to try new solutions due to the fear that their plan is not good enough. the second factor structure introduces how well one can deal with personal issues that may naturally arise as they implement their plan. for example, asking how well someone can deal with an increasing lack of interest in financial matters as they get older reflects this. statements about dealing with one’s eventual cognitive decline also loaded on this second column. many items touched on all these issues and loaded on both factor structures. because both of these factors capture an overall sense of retirement income self-efficacy, we chose questions for the final item set that successfully loaded on both columns. the items in bold represent the final retirement income self-efficacy items for the scale. the final question set presents a cronbach’s a score of .91. with respect to retest, the pearson table 4 exploratory factor analysis, cronbach’s a, and pearson correlation coefficient for retirement income self-efficacy scale factor loadings for all items** factors eigenvalues 1 6.80 2 5.06 self-efficacy items cronbach’s coefficient a* pearson correlation coefficient* 0.91 0.71 6 0.81 5 0.79 3 0.72 9 0.70 20 0.70 11 0.69 0.45 18 0.66 0.54 15 0.62 17 0.61 8 0.60 0.52 2 0.59 0.46 4 0.58 0.47 10 0.58 0.56 14 0.56 0.55 19 0.49 0.64 12 0.41 0.63 1 16 0.74 13 0.75 7 0.83 note. *cronbach’s a and pearson correlation coefficients selected for the items in bold. they represent the final questions for each scale. **only factor loadings greater or equal to 0.40 are presented. a. murgua and w. pfau / financial services review 30 (2022) 1–29 9 correlation coefficient is 0.71 (p < .0001). this supports excellent internal reliability and very good consistency between the time periods measured. in addition, as compared with lown’s financial self-efficacy scale, our retirement income scale reports higher levels of reliability (0.91 vs. 0.76) and comparable factor loading scores. lown’s financial self-efficacy scale does not report retest scores among their sample of 726 university employees (lown, 2011). 3.5. advisor usefulness while studies have identified the benefits of working with a financial advisor, a general skepticism remains toward the true benefit of financial advice. ex ante beliefs of perceived traits and stereotypes are difficult to overcome (bargh, chen, & burrows, 1996). working with an advisor is related to various planning and behavior activities. these include goal setting, retirement needs planning, investment diversification, retirement account optimization, reserves or contingency funds, behavioral guidance, and increased retirement confidence (marsden et al., 2011). hence, a scale measuring the cost effectiveness of these activities with an advisor may be valuable in determining the potential implementation options that retirees are likely choose. for this scale, participants identifying as financial advisors were removed from the data set to avoid any potential conflicts in their answers about the perceived utility of a financial professional. for each entry, we present opposing statements based on perceived advisor usefulness. we focus on both an advisor’s role and their cost effectiveness. items were presented via a semantic differential method. one statement is on the left and the other on the right. participants are asked to identify from a six-point scale, situated between both statements, which statement they relate with the most. a sample entry may read: 1. i can readily achieve my financial goals without the assistance of a financial advisor. 2. a financial advisor can readily help me achieve my financial goals. 3. statement 1 0 0 0 0 0 0 statement 2 in this example, picking the last circle would reflect a score of 6 and indicates a strong identification with perceiving high advisor usefulness. the average score of 3.32 indicates a slightly below average usefulness score for financial advisors (3.5 is the midpoint). the results in table 5 specify a three-factor structure for advisor usefulness with eigenvalues of 8.24, 3.14, and 2.42 across our proposed questions. the items within the first factor structure provide a description of what an advisor does from a holistic planning perspective. the second factor structure provides the additional component of how an advisor can potentially keep a retiree from making costly mistakes. the third factor structure identifies statements that present the advisor as a superfluous intermediary who is becoming increasingly irrelevant in today’s environment. the final item set for the scale, in bold, are taken from the first factor structure because it is the most dominant factor structure and best represents our intended focus of addressing a more complete purview of how an advisor may add value within a client relationship. this factor also reflects the various planning activities that result from working with an advisor (marsden et al., 2011). our advisor usefulness scale has a cronbach’s a score of 0.96, and a pearson correlation coefficient of 10 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 0.65 (p < .0001). this supports excellent internal reliability and adequate consistency between the time periods measured. 3.6. numeracy and numeracy self-awareness (dunning kruger) we based our numeracy scale on the weller et al. (2013) numeracy scale. while we maintained the mathematical integrity of each question, we reframed certain questions to reflect a retirement income context. participants were asked to answer eight questions largely detailing a general understanding of probabilities. our participants average test score is 66% (5.25/8). as a general measure of perceived numeracy, we asked participants how many of the questions did they think they answered correctly. on average, participants overestimate their scores by 18 percentage points. 3.7. portfolio loss aversion to measure a general sense of portfolio loss aversion, we presented respondents with an equal probability gamble between a positive and negative portfolio outcome. after directions were presented, the first question read: please state whether you would accept the table 5 exploratory factor analysis, cronbach’s a, and pearson correlation coefficient for advisor usefulness scale factor loadings for all items** factors eigenvalues 1 8.24 2 3.14 3 2.42 advisor usefulness cronbach’s coefficient a* pearson correlation coefficient* 0.96 0.65 11 0.87 16 0.87 10 0.86 8 0.84 7 0.84 9 0.84 13 0.84 15 0.81 12 0.72 4 0.69 0.46 18 0.68 0.47 14 0.54 0.52 19 0.42 0.45 5 0.42 0.64 6 0.77 20 0.56 17 0.68 3 0.80 2 0.58 0.47 1 0.59 note. *cronbach’s a and pearson correlation coefficients selected for the items in bold. they represent the final questions for each scale. **only factor loadings greater or equal to 0.40 are presented. a. murgua and w. pfau / financial services review 30 (2022) 1–29 11 following options? a 50-50 gamble of your portfolio losing 11% or gaining 35%. as an example; if you had a $1,000,000 portfolio, would you take a 50-50 gamble of your investment portfolio losing $110,000 or gaining $350,000? questions with a decreasing gain to loss ratio are presented until the respondent responds “no.” the first question presented here represents a 3.18 gain to loss ratio (35% gain vs. 11% loss), and each subsequent question reduced the spread between the gain to loss ratio by roughly 20%. accepting a lower gain to loss ratio reflects increasing levels of risk tolerance. scores were computed by dividing the number of questions completed by the total number of available questions. as an example, a score of 0.17 would indicate that only the most conservative question was answered “yes” (0.17 = 1/6). a low score indicates greater loss aversion. the average score was 0.38 indicating that the average gain to loss multiplier was 2.34. this means that the participants need, on average, a gain of 2.34 times the amount of the potential loss to engage in an equal probability bet. 3.8. retirement income concerns while the scales above represent our key psychological and behavioral finance variables for analysis, the following variables will also help address how these constructs influence retirement income concerns, investment behaviors, and retirement outlook. first, we attempt to quantify retirement income goals by measuring the level of concern a respondent feels about achieving a retirement objective. we classify three distinct concerns; longevity, lifestyle, and liquidity objectives. scales for these retirement concerns were presented via sematic differential with a six-point scale. a high score indicates a greater level of concern. 3.8.1. longevity longevity objectives are centered around addressing the main risk of retirement: outliving your money. most examples center on financial independence and knowing that you can pay your basic expenses and not be a burden to others. the average longevity score of 2.63 indicates an overall lower level of longevity concern across our participants. 3.8.2. lifestyle lifestyle objectives focus on maintaining your desired standard of living and enjoying your retirement with more discretionary spending. these goals require you to maximize your spending power. this aspect of retirement planning is about maintaining or improving your current lifestyle, rather than living too conservatively throughout retirement. the average lifestyle score of 3.90 indicates that achieving lifestyle objectives is an above average concern. 3.8.3. liquidity liquidity objectives involve maintaining enough reserves for unexpected contingencies. maintaining enough liquidity is especially important for family emergencies, home repairs, and an unexpected death or illness. the average score of 3.91 also indicates that avoiding unforeseen disruptions to a plan is an above average concern. 12 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 3.9. dividend agnosticism we also consider two different investment behavior scales. while dividend stocks produce income for their shareholders, from an economic perspective, no value is created or destroyed from issuing dividends. the structure of capital is irrelevant. with dividend payouts, the capital just moves from one theoretical pocket to the other. regardless of this, a significant number of retirees favor a dividend matching approach to retirement income. the industry further facilitates this focus with various “investing for yield” financial products aimed at retirees. we developed a six-item dividend agnosticism scale with a semantic differential format. a low score indicates a preference for dividend producing stocks and a high score indicates an indifference for dividend stocks. the average score of 4.1 indicates that most respondents are somewhat agnostic to dividend producing stocks for retirement income. 3.10. non-forecasting versus forecasting investment approach second, while there are many investment approaches, we wanted to single out the degree to which investors favor a forecasting or non-forecasting investment approach. a forecasting approach usually anticipates either a general market or individual stock movement or both. this is frequently referred to as an active approach to investing. a non-forecasting approach accepts market prices as a best estimate of price. this is usually identified as a passive investment approach. we created a five-item forecasting versus non-forecasting scale (nf) to capture this construct. in the nf scale, forecasting is the low score, and non-forecasting is the high score. the average score of 4.67 indicates that most participants exhibit a non-forecasting approach within their investment strategy. 3.11. nest egg satisfaction and retirement income plan anxiety to measure retirement outlook, our participants were presented with the following statements: “i’m where i thought i would be with my retirement nest egg” and “i feel anxious about my retirement income strategy” they were asked to rate their nest egg satisfaction statement on a five-point likert scale and their retirement income anxiety question on a four-point likert scale. responses ranged from strongly disagree (low score) to strongly agree (high score). participants average score of 3.64 for nest egg satisfaction indicates a somewhat above average perception of their retirement nest egg. furthermore, an average score of 2.21 for retirement income plan anxiety indicates a below average apprehension about their retirement income strategy. 4. results results in table 6 display bivariate correlations between our newly developed scales, retirement income concerns, and investment behaviors. correlations between these scales a. murgua and w. pfau / financial services review 30 (2022) 1–29 13 t ab le 6 c o rr el at io n ta b le r et ir em en t in co m e se lf -e ffi ca cy f in an ci al b ia se s n u m er ac y n u m er ac y se lf -a w ar en es s a d v is o r u se fu ln es s in er ti a p o rt fo li o lo ss av er si o n r et ir em en t in co m e se lf -e ff ic ac y f in an ci al b ia se s �0 .2 5 * * * * — n u m er ac y 0 .0 8 * �0 .2 4 * * * * — n u m er ac y se lf -a w ar en es s 0 .0 2 0 .1 3 * * * * �0 .4 7 * * * * — a d v is o r u se fu ln es s �0 .4 3 * * * * 0 .1 9 * * * * �0 .1 2 * * * �0 .0 3 — in er ti a �0 .1 2 * * * * 0 .1 1 * * * 0 .0 �0 .0 6 0 .1 0 * * — p o rt fo li o lo ss av er si o n 0 .1 1 * * * �0 .1 0 * * * 0 .1 1 * * * 0 .0 1 �0 .0 9 * * �0 .0 8 * — l o n g ev it y co n ce rn �0 .4 4 * * * * 0 .1 8 * * * * �0 .0 4 �0 .0 2 0 .2 1 * * * * 0 .1 0 * * �0 .0 6 l iq u id it y co n ce rn �0 .1 5 * * * * 0 .1 0 * * 0 .0 5 �0 .1 1 * * * 0 .1 3 * * * 0 .0 8 * �0 .0 8 * * l if es ty le co n ce rn 0 .1 5 * * * * �0 .0 5 0 .0 9 * * �0 .0 1 �0 .0 5 �0 .0 1 0 .1 8 * * * * d iv id en d ag n o st ic 0 .2 3 * * * * �0 .3 7 * * * * 0 .2 1 * * * * �0 .0 9 * * �0 .2 0 * * * * �0 .0 1 0 .0 8 * n o n -f o re ca st in g v er su s fo re ca st in g 0 .3 4 * * * * �0 .3 5 * * * * 0 .1 1 * * * �0 .0 9 * * �0 .1 1 * * * �0 .0 1 0 .0 8 * * * p < .0 5 * * p < .0 1 * * * p < .0 0 1 * * * * p < .0 0 0 1 14 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 indicate very good levels of validity as evidenced by the convergent and discriminant relationships in expected directions. 4.1. psychological and behavioral finance and scales high levels for the retirement income self-efficacy scale are negatively related to financial biases (r = �0.25, p < .0001), inertia (r = �0.12, p < .0001), and perceived advisor usefulness (r = �0.43, p < .0001) scores. self-efficacy is positively related to numeracy (r = 0.08, p < .02) and loss aversion tolerance (r = 0.10, p < .0001). financial biases are negatively related to numeracy (r = �0.24, p < .0001) and loss aversion tolerance (r = �0.10, p < .0001) and positively relate to a lack of numeracy awareness (r = 0.13, p < .0001), inertia (r = 0.11, p < .0001), and perceived usefulness of an advisor (r = 0.12, p < .0001). high numeracy is related to a lower perceived advisor usefulness (r = �0.12, p < .0001) but a higher loss aversion tolerance (r = 0.11, p < .0001). there is no observable relationship between numeracy and inertia. however, higher inertia levels are positively related to advisor usefulness (r = 0.10, p < .0001) and greater levels of loss aversion (r = �0.08, p < .05). all variables with significant associations are in the expected direction. self-reported scales, especially those that potentially paint an unfavorable impression of a respondent, carry the risk of a social desirability bias in their responses. it is interesting to note that our self-reported financial bias score is also associated in the expected direction with more objective measures of numeracy, lack of numeracy awareness, and inertia. in addition, socially desirable traits like self-efficacy are also associated with numeracy in the expected direction. with regards to the numeracy and numeracy awareness, our results support a dunningkruger effect (kruger & dunning, 1999). there is a negative relationship between one’s level of numeracy and perceived numeracy (r = �0.47, p < .0001). the worst one performs on numeracy, the better one thought they performed. furthermore, fig. 1 separates numeracy scores and perceived numeracy scores by quartiles. participants in the first quartile (worst 25%) overestimate their score by 88% while participants the fourth quartile underestimate their score by 7%. the lower the quartile positioning, the greater the overestimation. in contrast, the higher scoring quartile for numeracy slightly underestimate their skill. this underestimation among top quartile performers is also observed in the original investigation of the dunning-kruger effect (kruger & dunning, 1999). 4.2. retirement income concerns longevity concerns are associated with higher levels of financial biases (r = 0.18, p < .0001), inertia (r = 0.10, p < .01), the perceived usefulness of an advisor (r = 0.21, p < .0001), and significantly lower levels of self-efficacy (r = �0.44, p < .0001). while not as strong, liquidity concerns also reflect similar directional relationships to financial biases (r = 0.10, p < .01), inertia (r = 0.08, p < .02), advisor usefulness (r = 0.13, p < .0001), and selfefficacy (r = �0.15, p < .0001). additionally, higher liquidity concerns are associated with a. murgua and w. pfau / financial services review 30 (2022) 1–29 15 lower loss aversion tolerance (r = �0.08, p < .01). lifestyle concerns exhibit significant positive associations with self-efficacy (r = 0.15, p < .0001), numeracy (r = 0.09, p < .01), and loss aversion tolerance (r = 0.18, p < .0001). results across our scales indicate a hierarchy of retirement income concerns. at a more basic level, the presence of increased financial biases, lower numeracy scores, and lower levels of self-efficacy relate to a greater concern for achieving essential spending needs and accommodating unexpected emergencies during retirement. in addition, higher levels of liquidity concerns reflect greater loss aversion. there may also be a recognition among those with higher longevity and liquidity concerns that a financial professional can help them overcome their personal hurdles and retirement concerns. in contrast, achieving more discretionary spending goals signals a shift beyond the threshold of longevity and liquidity concerns. while achieving essential spending needs are a requirement for any successful retirement income plan, participants with high levels of lifestyle concerns are not overly anxious about the ability to fulfill their longevity needs (r = �0.27, p < .0001). these individuals usually desire spending increases to achieve more lifestyle objectives in retirement. this generally requires a greater exposure to market volatility that is associated with greater levels of self-efficacy, numeracy, and loss aversion tolerance. higher levels of numeracy self-awareness are related to the need to account for unexpected contingencies. fig. 1. numeracy score versus expected score 16 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 4.3. investment behaviors 4.3.1 dividend agnostic an indifference toward dividend stocks (high score) for retirement income is associated with greater levels of retirement income self-efficacy (r = 0.23, p < .0001), numeracy (r = 0.20, p < .0001), and loss aversion tolerance (r = 0.08, p < .05). in contrast, a focus on dividends for retirement income (low dividend agnostic score) is related to higher levels of financial biases (r = �0.37, p < .0001), perceived advisor usefulness (r = �0.20, p < .0001), and lower levels of numeracy self-awareness (r = �0.09, p < .01). those with a higher understanding of numerical concepts exhibit the belief that there is no true economic benefit by focusing on dividend producing stocks for a retirement income plan. additionally, investors most susceptible to financial biases and lacking numeracy self-awareness may be emphasizing the benefits of dividend stocks for retirement income while minimizing the risk of focusing on a concentrated selection of stocks. a non-dividend focused approach and a nonforecasting investment style (r = 0.42, p < .0001) also supports this association. 4.3.2. non-forecasting versus forecasting investment strategy a high score on the non-forecasting versus forecasting scale indicates a preference for a non-forecasting (passive) investment approach. a low score indicates a preference for forecasting (active). a non-forecasting investing approach is associated with higher levels of retirement income self-efficacy (r = 0.34, p < .0001), numeracy (r = 0.11, p < .001), and to a lesser extent loss aversion tolerance (r = 0.08, p < .01). participants identifying with a forecasting approach exhibit a higher susceptibility to financial biases (r = �0.35, p < .0001) and low numeracy self-awareness (r = �0.09, p < .01). while the results are not intended to convey the benefits of one investment approach over another, results indicate that high self-efficacy, numeracy, and portfolio loss aversion tolerance are more salient for participants espousing a greater affinity for a non-forecasting investment approach. while these individuals may have the ability and domain confidence to attempt to identify market mispricing, they may also recognize the inherent difficulty and the potential for chance outcomes in these endeavors. instead, these participants exhibit a preference to efficiently capture general market returns. conversely, participants supporting market forecasting preferences are more vulnerable to financial biases; perhaps from the repeated exposures to the vagaries of unpredictable stock market movements. coupled with a lack of numeracy self-awareness and lower retirement self-efficacy, these participants are more likely to seek professional guidance for help (r = 0.11, p < .001). overall, correlations indicate that our scales for retirement income self-efficacy, financial bias, numeracy, advisor usefulness, inertia, and portfolio loss aversion present with criterion validity as all relationships were in the expected direction. in addition, these constructs have significant implications in how they are related to each other, retirement income concerns, and investment behaviors that are very impactful to retirement income success. in the next section, we will assess how these variables are related to retirement outlooks and advisor implementation. a. murgua and w. pfau / financial services review 30 (2022) 1–29 17 table 7 regression analysis of psychological and behavioral finance constructs with retirement outlooks nest egg satisfaction retirement income strategy anxiousness sample 577 577 f value 13.15 12.12 global f pr > f **** **** r2 0.20 0.19 intercept estimate 0.12 0.11 standard error 0.09 0.10 prob 0.19 0.24 self-efficacy estimate 0.37 �0.29 standard error 0.04 0.04 probability **** **** financial biases estimate �0.01 0.20 standard error 0.04 0.04 probability 0.70 **** inertia estimate �0.01 0.09 standard error 0.04 0.04 probability 0.86 * numeracy estimate �0.03 0.03 standard error 0.05 0.05 probability 0.45 0.50 numeracy awareness estimate 0.01 �0.02 standard error 0.05 0.05 probability 0.91 0.68 portfolio loss aversion estimate �0.03 0.02 standard error 0.04 0.04 probability 0.45 0.69 advisor usefulness estimate �0.07 0.02 standard error 0.04 0.04 probability 0.10 0.64 gender estimate �0.15 �0.14 standard error 0.10 0.10 probability 0.15 0.18 marital status estimate 0.06 �0.09 standard error 0.12 0.12 probability 0.59 0.47 age estimate 0.05 �0.17 standard error 0.04 0.04 probability 0.18 **** (continued on next page) 18 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 4.3.3. retirement outlook table 7 indicates the characteristics associated with the degree to which participants felt that their retirement nest eggs were on track with their expectations and whether they felt anxious about their retirement income strategy. ordinary least squares regressions with standardized coefficients are used to assess these independent variables with our newly created scales and participant demographic variables, which include gender, marital status, age, and net worth. we find a significant relationship with retirement income self-efficacy (estimate = 0.37, p < .0001) and net worth (estimate = 0.17, p < .0001) with nest egg satisfaction. the results point out the importance of self-efficacy over other psychological constructs relating to a proxy for retirement satisfaction. moreover, while net worth is a significant indicator, self-efficacy is more influential (estimates = 0.37 vs. 0.17). with regard to retirement anxiousness, retirement income self-efficacy (estimate = �0.29, p < .0001), levels of financial biases (estimate = 0.20, p < .0001), age (estimate = �0.17, p < .0001), and inertia (estimate = 0.09, p < .02) are significant contributors. while self-efficacy is in the expected opposite direction than its relationship to nest egg satisfaction, it is again the largest contributor in the model. lower self-efficacy suggests a greater the level of retirement anxiousness while holding all other variables constant. results also indicate that a greater susceptibility to financial heuristics may lead to more anxiety regarding one’s retirement income success. additionally, an inability to execute financial tasks in a timely manner also leads to greater levels of retirement anxiousness. although net worth is related to nest egg satisfaction, it is not significantly related to retirement anxiousness. age is the only significant demographic variable. our younger age cohorts, largely consisting of 40to late 50-year-olds exhibited more anxiety about their retirement income strategy than those who are already near or into their retirement. these results continue to provide support for the validity of our scales and their impact on retirement income outlooks. results also indicate when controlling for demographic variables, these psychological constructs continue to exert a significant influence on one’s retirement outlook. retirement income self-efficacy is the only variable that significantly relates to both nest egg satisfaction and retirement income strategy anxiousness. maladaptive investment behaviors such as increased financial biases and inertia also relate to increased table 7 (continued) nest egg satisfaction retirement income strategy anxiousness net worth estimate 0.17 0.00 standard error 0.04 0.04 probability **** 0.90 gender. female is the reference variable marital status. married is the reference variable *p = < 0.05 **p = < 0.01 ***p = < 0.001 ****p = < 0.0001 a. murgua and w. pfau / financial services review 30 (2022) 1–29 19 levels of retirement income anxiousness. while numeracy, numeracy awareness, portfolio loss aversion tolerance, and perceived advisor usefulness are associated with various investment behaviors and retirement risks, these associations do not manifest when assessing retirement income outlooks at the multivariate level. 4.3.4. financial implementation style while we have assessed how behavioral finance and psychological factors affect retirement income beliefs, investment behaviors, and retirement outlook, we want to further analyze their potential influence on financial implementation methods. we utilize logistic regression to examine the degree of influence of our behavioral finance and psychological factors on whether participants are currently in a financial advisory relationship. results in table 8 indicate that perceived advisor usefulness (estimate = 0.99, odds ratio [or] = 2.72, p < .0001) is the only significant predictor variable in the model. holding all other variables constant, for every unit increase in our perceived advisor usefulness scale, the odds of being in an advisory relationship increase by a factor of 2.72 times. results provide strong support for perceived advisor usefulness as the key indicator of advisor utilization. a higher perceived usefulness of an advisor means that one is more likely to engage in such a relationship. the perceived level of advisor usefulness can help identify what implementation avenues certain individuals respond to best and tailor approaches to those preferences. the significant results for perceived advisor usefulness do not remove the potential impact of reverse causality (i.e., endogeneity) when choosing to utilize an advisor. however, the inclusion of control variables (i.e., net worth, etc.) in the model to adequately capture their potential influence over the use of an advisor, helps reduce the effects of endogeneity (rosenbaum & rubin, 1984). we will discuss this factor further in our conclusions section. being able to reliably identify who is most likely to implement a retirement income plan with the assistance of a financial advisor is a significant step forward. instead of trying to convince skeptical individuals of using an advisor, a better approach may be to identify their preferences for receiving financial advice and provide avenues that maximize those preferences. by facilitating this approach, individuals may be more likely to engage in behaviors that ultimately lead to retirement income success. because advisor usefulness is the main determinant for directly working with an advisor and retirement income self-efficacy is a very strong variable throughout this investigation for identifying retirement income beliefs, risks, and investment behaviors, we establish a financial implementation matrix with these factors to help us identify how an investor prefers to implement financial tasks. by placing perceived advisor usefulness on the vertical axis and retirement income self-efficacy on the horizontal axis, we can identify four investor personas that can be aligned with preferred financial implementation approaches. fig. 2 presents the financial implementation matrix and corresponding personas. the top left quadrant identifies someone who is below average on perceived financial self-efficacy and high on perceived advisor usefulness. as a result, this person is more likely to have an advisor take the lead role in guiding their retirement plan. a profile score in this quadrant would be indicative of a delegator persona. 20 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 table 8 logistic regression analysis of psychological and behavioral finance constructs with advisory implementation in a current advisory relationship sample 577 wald test (x2) 72.39 pr > x2 <0.0001 c-statistic 0.75 rescaled r2 0.20 intercept estimate �1.17 standard error 0.17 probability <0.0001 wald 95% confidence interval limits self-efficacy odds ratio 1.24 0.99 1.56 estimate 0.22 standard error 0.12 probability 0.06 financial biases odds ratio 1.13 0.92 1.40 estimate 0.12 standard error 0.11 probability 0.24 inertia odds ratio 1.04 0.86 1.25 estimate 0.03 standard error 0.09 probability 0.72 numeracy odds ratio 1.06 0.83 1.36 estimate 0.06 standard error 0.12 probability 0.62 numeracy awareness odds ratio 1.15 0.91 1.47 estimate 0.14 standard error 0.12 probability 0.25 portfolio loss aversion odds ratio 1.15 0.93 1.43 estimate 0.14 standard error 0.11 probability 0.20 advisor usefulness odds ratio 2.72 2.13 3.46 estimate 1.00 standard error 0.12 probability **** gender odds ratio 1.19 0.69 2.04 estimate 0.09 standard error 0.14 (continued on next page) a. murgua and w. pfau / financial services review 30 (2022) 1–29 21 the top right quadrant identifies someone who is high on both self-efficacy and advisor usefulness. this describes someone who feels very confident about their own ability but also appreciates the value of an advisor. this persona enjoys contributing as an active partner with a financial professional. a profile score in this quadrant indicates that they are most likely a collaborator. table 8 (continued) in a current advisory relationship probability 0.53 marital status odds ratio 0.74 0.38 1.44 estimate �0.15 standard error 0.17 probability 0.37 age odds ratio 0.92 0.75 1.13 estimate �0.08 standard error 0.10 probability 0.44 net worth odds ratio 1.03 0.84 1.25 estimate 0.03 standard error 0.10 probability 0.79 gender. female is the reference variable marital status. married is the reference variable *p = < 0.05 **p = < 0.01 ***p = < 0.001 ****p = < 0.0001 fig. 2. financial implementation matrix 22 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 the bottom right quadrant is indicative of someone who is high on financial self-efficacy and low on perceived advisor usefulness. these individuals are confident about their aptitude to create and implement a retirement income plan and do not feel engaging an advisor for assistance is cost effective. this quadrant most likely reflects self-directed investor personas. the bottom left quadrant identifies someone who is below average on perceived financial self-efficacy and is also low on perceived advisor usefulness. while those in this quadrant do not value an ongoing advisory relationship, their low self-efficacy score leaves open the possibility of seeking specialized guidance for complex financial decisions. individuals here may seek a second opinion or a one-time consultation plan with an advisor as they continue to implement their strategy. a profile score in this quadrant relates to a validator persona. table 9 identifies the frequency breakdown between our four implementation personas and indicates whether these individuals are in a current advisory relationship. we do not include participants identifying as financial professionals in this analysis. delegators (n = 332) and self-directed investors (n = 331) each represent 34% of our sample participants (n = 965). collaborators (n = 179) and validators (n = 123) represent 19% and 13%, respectively. both delegators and collaborators (36.8% and 46.4%) are significantly more likely to have a current advisory relationship than validators and self-directed investors (13% and 10.9%). this is to be expected because the advisor usefulness score represents the vertical axis of the implementation matrix. this matrix also provides potential insight into the type of financial service model each persona may best identify with. for example, table 10 provides logistic regression results of the various behavioral finance, psychological, retirement concerns, and demographic variables described in this investigation for each persona type. because the implementation matrix is based on perceived advisor usefulness and self-efficacy, we did not include these variables in the analysis. we also remove numeracy awareness from this analysis due to its insignificant findings in the previous multivariate analyses, its high association with numeracy, and the inclusion of other retirement income concern variables. while delegator personas are naturally characterized by low self-efficacy and high advisor usefulness, they exhibit higher levels of longevity concerns (estimate = 0.43, or = 1.54, p < .001), more anxiety towards their retirement income strategy (estimate = 0.29, or = 1.34, p < .01), and higher levels of financial biases (estimate = 0.30, or = 1.34, p < .01). these potential headwinds may be why a delegator persona may be more willing to outsource more financially driven tasks to professionals. within this type of advisory relationship, an advisor can help address these concerns and biases with a financial plan that focuses on retirement income success and frames the investment process into a more goals-driven outcome. client meetings with specific themes that address current events within a behavioral finance framework or that contextualize the investment experience may help bring awareness to financial biases. the global wald chi-square for collaborators and the various independent variables are not significant (wald x2 19.14, p < .09). collaborators do not reliably exhibit higher or lower levels of the various factors. however, post hoc bonferroni (dunn) t tests reveal collaborators have higher net worth levels than delegators and validator personas (t = 2.65, p < .05 for both). hence, coupled with higher levels of self-efficacy and perceived advisor usefulness, the desire for collaboration may be due to the realization of the added a. murgua and w. pfau / financial services review 30 (2022) 1–29 23 t ab le 9 f re q u en cy o f im p le m en ta ti o n b y in v es to r p er so n as d el eg at o rs c o ll ab o ra to rs in ad v is o r re la ti o n sh ip y es n o y es n o f re q u en cy 1 2 2 2 1 0 8 3 9 6 p er ce n t “y es ” 3 6 .8 6 3 .3 4 6 .4 5 3 .6 v al id at o rs s el fd ir ec te d in ad v is o r re la ti o n sh ip y es n o y es n o f re q u en cy 1 6 1 0 7 3 6 2 9 5 p er ce n t “y es ” 1 3 .0 8 7 .0 1 0 .9 8 9 .1 t o ta l sa m p le 9 6 5 24 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 table 10 logistic regression analysis of psychological and behavioral finance constructs and investor type delegator collaborator self-directed validator sample 576 576 576 576 wald test (x2) 67.47 19.14 73.85 27.99 pr > x2 **** 0.09 **** ** c-statistic 0.73 0.63 0.74 0.69 rescaled r2 0.19 0.06 0.21 0.10 intercept estimate �0.88 �1.44 �0.87 �1.96 standard error 0.16 0.17 0.17 0.20 probability <0.0001 <0.0001 <0.0001 <0.0001 lifestyle concern odds ratio 1.02 0.93 1.07 0.89 estimate 0.02 �0.07 0.07 �0.11 standard error 0.11 0.12 0.10 0.14 probability 0.86 0.54 0.53 0.41 longevity concern odds ratio 1.54 0.85 0.53 1.37 estimate 0.43 �0.17 �0.64 0.31 standard error 0.12 0.14 0.14 0.15 probability *** 0.25 **** * liquidity concern odds ratio 1.18 0.98 0.95 0.86 estimate 0.17 �0.03 �0.05 �0.15 standard error 0.11 0.13 0.11 0.15 probability 0.14 0.84 0.65 0.33 retirement income anxiety odds ratio 1.34 0.78 0.82 1.30 estimate 0.29 �0.25 �0.20 0.26 standard error 0.11 0.13 0.11 0.15 probability ** * 0.06 0.08 financial biases odds ratio 1.34 1.13 0.79 0.83 estimate 0.30 0.12 �0.24 �0.19 standard error 0.11 0.12 0.10 0.14 probability ** 0.31 * 0.17 inertia odds ratio 1.14 1.05 0.75 1.09 estimate 0.13 0.05 �0.29 0.08 standard error 0.09 0.11 0.10 0.12 probability 0.15 0.62 ** 0.48 numeracy odds ratio 0.99 1.00 1.12 0.86 estimate �0.01 0.00 0.11 �0.15 standard error 0.10 0.12 0.10 0.14 probability 0.90 0.98 0.29 0.28 portfolio loss aversion tolerance odds ratio 0.96 0.97 1.21 0.75 estimate �0.05 �0.03 0.19 �0.29 standard error 0.11 0.12 0.10 0.15 probability 0.67 0.78 0.07 * gender odds ratio 0.73 1.14 1.00 1.35 (continued on next page) a. murgua and w. pfau / financial services review 30 (2022) 1–29 25 complexities that arise from greater amounts of wealth. while a delegator may appreciate an advisor taking the lead, a collaborator may prefer situations that allow for active input in developing and implementing a retirement income strategy. high levels of communication and providing sound reasoning behind the decision-making process is most appropriate with this persona. in contrast to a delegator that may just want to know the proverbial “time,” the collaborator may also want to know “how the clock is made.” diagonally across from delegators, on the implementation matrix, are self-directed investors. while they exhibit a high degree of self-efficacy and low perceived advisor usefulness, selfdirected investors also exhibit a significant negative relationship with longevity concerns (estimate = �0.64, or = 0.53, p < .0001), financial biases (estimate = �0.24, or = 0.79, p< .05), and inertia (estimate = �0.29, or = 0.75, p < .01). longevity concerns and degree of financial biases are in the opposite direction of delegators. in addition, while anxiety regarding their retirement income strategy is not significant at the p < .05 level of analysis (estimate = �0.20, or = 0.82, p < .06), it was also trending in the opposite direction as delegators. while a self-directed investor is less likely to utilize an advisor, there are various approaches that can engage and help them with a successful retirement income plan. because this persona is actively involved in their retirement income plan and has high levels of self-efficacy, it is important that they have access to unbiased educational materials that convey the practical application of retirement income strategies. with the rise of financial table 10 (continued) delegator collaborator self-directed validator estimate �0.15 0.06 0.00 0.15 standard error 0.12 0.16 0.13 0.17 probability 0.22 0.68 0.99 0.38 marital status odds ratio 0.70 1.63 0.62 1.84 estimate �0.18 0.25 �0.24 0.31 standard error 0.16 0.17 0.16 0.19 probability 0.27 0.16 0.13 0.10 age odds ratio 1.19 0.95 0.89 0.92 estimate 0.17 �0.05 �0.12 �0.08 standard error 0.11 0.13 0.11 0.14 probability 0.12 0.69 0.27 0.57 net worth odds ratio 0.99 1.38 0.90 0.81 estimate �0.01 0.32 �0.10 �0.21 standard error 0.10 0.11 0.10 0.14 probability 0.90 ** 0.31 0.14 gender. female is the reference variable marital status. married is the reference variable *p = < 0.05 **p = < 0.01 ***p = < 0.001 ****p = < 0.0001 26 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 technologies, many automated advisory offerings are readily available for investments. in addition, online financial planning offers could also address the specific retirement income problems facing investors. lastly, validators exhibit a significant positive relationship to longevity concerns (estimate = 0.31, or = 1.37, p < .05), and a negative relationship to portfolio loss aversion tolerance (estimate = �0.29, or = 0.75, p < .05). while they do not view an ongoing advisor relationship as cost effective, their low levels of self-efficacy may lead them to seek the reassurance from an advisor in the form of a second opinion regarding the above-mentioned concerns. to provide the appropriate assistance for these individuals, advisors may need to give serious consideration to expanding their service offering to include a planning services as a stand-alone offer and not bundled with asset management. ultimately, the financial implementation matrix is an effective way to determine what type of approach can best assist with implementing a retirement income plan. this is a more optimal approach than attempting to convince every investor that they should engage in an ongoing advisory relationship with asset management as its primary revenue source. entry level offerings may also provide a stepping-stone into higher level service models as individuals learn firsthand about the complexities of developing and implementing a retirement income plan. 5. conclusion we quantify retirement income self-efficacy, financial biases, numeracy, numeracy selfawareness, inertia, and perceived advisor usefulness and create scales for these constructs specific to retirement income. in addition, we show how these scales relate to each other and more traditional measures, such as loss aversion, to further provide significant levels of construct validity. we find these scales to significantly relate to retirement income concerns such as longevity, lifestyle, and liquidity. moreover, investment behaviors such as a preference for dividend stocks and investment approach are related to these constructs. one’s retirement nest egg satisfaction and anxiety levels towards their retirement income strategy is also shown to be related to many of these factors. overall, results indicate significant levels of criterion validity for the newly developed scales. and finally, we create four investment personas with our advisor usefulness and retirement income self-efficacy scales to successfully identify preferred financial implementation methods. the implications for this investigation affect both individuals and financial professionals. individuals can readily recognize their relative strengths and weaknesses when implementing a retirement income strategy. by clearly pointing out potential weak spots, an individual can take the necessary steps to fill in the needed gaps. for example, individuals low in numeracy can strive to attain a higher level of competence for retirement income success or can seek professional help. if they are lacking perceived self-awareness regarding their numeracy, this may serve as a wake-up call for them to temper their perceived expertise and to embrace a more receptive attitude towards individuals with more experience and knowledge in the subject matter. a high financial bias score may indicate the areas in which individuals are more susceptible to maladaptive behaviors. this information can also help them a. murgua and w. pfau / financial services review 30 (2022) 1–29 27 focus on their strengths to successfully implement a retirement income plan. for example, an individual who may have an average numeracy score but is very self-aware of this may be more willing to embrace advice from third parties in a productive manner. for a financial professional, the ability to present advice in a manner that resonates with an individual is paramount to a successful relationship. more importantly, it will help them follow through with their retirement income plan. for example, assessing individual levels of self-efficacy, numeracy, numeracy self-awareness, financial biases, inertia, and perceived advisor usefulness will help an advisor better understand their client and the relationship dynamic that is most likely to be more engaging. having greater levels of insight into whether a client understands what is presented to them and how likely they will be candid about their comprehension is beneficial to assuring adherence to a plan. if an advisor has a sense of a client’s numeracy and self-awareness, they can tailor investment presentations and recommendations in a productive manner. if an advisor knows how a client may be interpreting the investment landscape and current events, the advisor may be able to reach out and discuss these issues before the client potentially infers conclusions that are suboptimal to their plan. if an advisor has a sense of how timely a client implements advice, then the advisor can present next steps for a plan in a more digestible manner. and finally, an advisor can identify a relationship dynamic that will increase client satisfaction. a delegator and a collaborator need a different cadence with their advisor for the relationship to be productive. in addition, a validator and self-directed investor may need different options that do not require an ongoing professional relationship. while this investigation significantly enhances our understanding of retirement income beliefs and retirement outcomes, no research is without limitations. because our convenience sample largely consists of individuals interested and well-read in retirement income and possess greater levels of net worth than the larger population, further testing should be considered with a more diverse population. in addition, to address the potential issue of reverse causality for the advisor usefulness score, future investigations can address this beyond the use of control variables in the investigation, as we have, by utilizing a propensity score (rosenbaum & rubin, 1984). the propensity score methodology adjusts for this potential bias. ultimately, individuals and advisors should recognize how these factors affect the successful implementation of a retirement income plan. one can then be more aware of how to focus on individual strengths and weaknesses. future studies should also continue to explore how various investor personas and implementation strategies can be individualized with greater levels of specificity to productively engage in a successful retirement plan implementation. references agarwal, s., & mazumder, b. (2013). cognitive abilities and household financial decision making. american economic journal: applied economics, 5, 193–207. available at https://www.aeaweb.org/ articles?id= 10.1257/app.5.1.193 allgood, s., & walstad, w. (2016). the effects of perceived and actual financial literacy on financial behaviors. economic inquiry, 54, 675–697. 28 a. murguı́a and w. pfau / financial services review 30 (2022) 1–29 asebedo, s., & seay, m. (2018). financial self-efficacy and the saving behavior of older pre-retirees. journal of financial counseling and planning, 29, 357–368. asebedo, s. d., & browning, c. m. (2020). the psychology of portfolio withdrawal rates. psychology and aging, 35, 78–90. balasubramnian, b., & sargent, c. s. (2020). impact of inflated perceptions of financial literacy on financial decision making. journal of economic psychology, 80, 102306. bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84, 191–215. bandura, a. (1997). self-efficacy: the exercise of control. new york, ny: w.h. freeman and company. bandura, a. (2006). guide for constructing self-efficacy scales (revised). self-efficacy beliefs of adolescents, 5, 307–337. blanchett, d., kaplan, p. (2013). alpha, beta, and now. . . gamma. the journal of retirement, 1, 29–45. bargh, j. a., chen, m., & burrows, l. (1996). automaticity of social behavior: direct effects of trait construct and stereotype activation on action. journal of personality and social psychology, 71, 230–244. devellis, r. f. (2017). scale development theory and application. thousand oaks, ca: sage publication. fernandes, d., lynch, j. g. , jr., , & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors.management science, 60, 1861–1883. gal, d. (2006). a psychological law of inertia and the illusion of loss aversion. judgment and decision making, 1(july), 23–32. grable, j., & joo, s.-h. (1999). financial help-seeking behavior: theory and implications. journal of financial counseling and planning, 10(1), 14–25. hibbard, j. h., slovic, p., peters, e., finucane, m. l., & tusler, m. (2001). is the informed-choice policy approach appropriate for medicare beneficiaries? health affairs (project hope), 20, 199–203. kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47, 263–291. kell, p. (2014). early financial literacy education key to informed financial decisions. governance directions, 66, 685–688. available at https://search.informit.com.au/documentsummary;dn=791221674677151;res= ielapa kinniry, f. m., jaconetti, c. m., chin, p., polanco, f., & zilbering, y. (2015). putting a value on your value: quantifying vanguard adviser’s alpha. valley forge, pa: the vanguard group. kruger, j., & dunning, d. (1999). unskilled and unaware of it: how difficulties in recognizing one’s incompetence lead to inflated self-assessments. journal of personality and social psychology, 77, 1121–1134. lown, j. m. (2011). development and validation of a financial self-efficacy scale. journal of financial counseling and planning, 22, 54–63. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52, 5–44. marsden, m., zick, c., & mayer, r. (2011). the value of seeking financial advice. journal of family and economic issues, 32, 625–643. newall, p. w. s. (2016). downside financial risk is misunderstood. judgment and decision making, 11(5), 416– 423. available at https://econpapers.repec.org/repec: jdm:journl:v:11:y:2016:i:5:p:416-423 rosenbaum, p. r., & rubin, d. b. (1984). reducing bias in observational studies using subclassification on the propensity score. journal of the american statistical association, 79, 516–524. tversky, a., & kahneman, d. (1974). judgment under uncertainty: heuristics and biases. science, 185, 1124– 1131. weller, j. a., dieckmann, n. f., tusler, m., mertz, c. k., burns, w. j., & peters, e. (2013). development and testing of an abbreviated numeracy scale: a rasch analysis approach. journal of behavioral decision making, 26, 198–212. zumbrum, j. (2015 november 18). the u.s. may be the world’s richest country, but it ranks 14th in financial literacy. the wall street journal. available at http://blogs.wsj.com/economics/2015/11/18/theus-may-be-the-worlds-richest-country-but -it-ranks-14th-in-financial-literacy/ a. murgua and w. pfau / financial services review 30 (2022) 1–29 29 pii: s1057-0810(01)00070-1 from the editor retirement planning is becoming increasingly important as the baby boomers approach the age when they are eligible to retire and must deal with the consequences of their prior investment decisions. three of the six articles in this issue deal with various aspects of pension plans and ratios to be used in planning activities. kevin w. sigrist and stewart l. brown explore the issues involved in designing a state retirement plan in florida for their article entitled, “design considerations for large public sector defined contribution plans”. they identify the major types of public retirement plans, issues that face the trustees who oversee the implementation of a plan, legal issues, model legislation for these types of plans, the administrative and advisory costs of the plans, and the tradeoff involved in institutional accounts versus mutual funds. this article is of particular interest to all of those individuals who expect a pension from a state entity as well as those who are responsible for designing these plans. many educators invest at least some of their retirement funds in tiaa/cref funds that are available through 403(b) plans. edward m. miller and larry j. prather examine the possibility of predictable patterns in tiaa/cref funds that allow for an opportunity to use a trading rule that they devise that will beat a buy and hold approach. their work is interesting in light of the announcement by cref that restricts the number of trades that can be made by participants. specific ratios that financial planners can use in helping their clients to plan for retirement are provided in a study by sue alexander greninger, vickie l. hampton, karrol a. kitt, and susan jacquet. they report on a survey that seeks consensus on appropriate assumptions to use in retirement savings projections and specific steps to be taken as the individual nears retirement age. bond funds are frequently recommend for those nearing retirement as an alternative to stocks. james philpot, douglas hearth, and james rimbey examine the performance of managers of bond funds in an article entitled, “performance persistence and management skill in non-conventional bond mutual funds”. they find that managers of high-yield bond funds are able to outperform on a short-term basis. they also provide evidence that portfolio turnover is inversely related to risk-adjusted performance. individual investors are often advised to add international funds to their portfolio to reduce their overall risk. matthew o’connor and edward a. downe explore the impact of webs financial services review 9 (2000) v–vi 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00070-1 on closed-end country funds. they caution investors to be wary of discounts for these types of funds. the last article by jeanne m. hogarth and jinkook lee looks at the search pattern that individuals follow in selecting a mortgage. this article is unique in as much as it explores factors that are not normally study in the residential financing literature. k. e. lahey vi editorial / financial services review 9 (2000) v–vi pii: 1057-0810(93)90008-e financial services review, 3( 1): 83-92 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any fom reserved. abstracts of articles on individual financial management edited by phyllis schiller myers virginia commonwealth university consumption-savings behavior habit dynamics and wealth accumulation fluctuations, john g. powell (university of otago, dunedin, new zealand). this paper explores the possibility that consumption habits are directly af fected by investors’ wealth histories as unanticipated wealth losses alter perceived financial well-being and force investors to re-evaluate their minimum consumption habit requirements. the response of consumption habits to unexpected investment return changes is assumed to depend upon individual differences in need for cognition and tendency towards risk taking, and these differences lead to sharply contrasting investment strategies which alter the wealth accumulation process. investors who do not monitor their consumption habits during good times find that they must quickly move out of risky assets during downturns in order to ensure that they can continue to meet their minimum consumption needs. this dynamic trading portfolio insurance behavior is inherently risky in an economic sense, so these myopic investors must compensate by reducing the risk exposures they choose, thus resulting in a lower rate of wealth accumulation and a utility loss. investors are better off if they cut back habitual consumption during recessions in order to preserve financial well-being and maintain wealth accumulation rates. journal ofeconomic psychology, june 1993, 14(2): 267-283. (reprinted with permission of north-hol land publishing company.) patterns of overspending in u.s. households, by mikyeong bae, sherman hanna, and suzanne lindamood an original analysis of the bls consumer expenditure survey shows that almost 40% of u.s. households spent more than their income in 1990. multivariate logistic regression indicates that income level is the most important factor related to whether a household overspends. more educated consumers are likely to over 84 financial services review, 3(l) 1993 spend than are less educated consumers, when income and other factors are control led. financial counseling and planning, 1993,4: 1 l-30. (reprinted with permis sion of financial counseling and planning.) women, men, and money styles, by melvin prince (fordham uni versity). the character of gender differences in money styles is examined. on average, money style item scores for young adult males and females are found to be consistently disparate. males and females are both likely to see money as closely linked with esteem and power, but males are more prone to feel involved and competent in money handling, and take risks to amass wealth. females have a greater sense of envy and deprivation with respect to money as a means of obtaining things and experiences that they can enjoy in the present. journal of economic psychology, march 1993, 14( 1): 175-182. (reprinted with mission of non-holl~d ~blish~g company.) estate planning and distribution housing costs and bequest motives, by y. nakagami and a. m. pereira the standard user cost of housing capital used in the literature has ignored household bequest motives. in this paper, the authors generalize the standard approach to incorporate bequest motives and inheritance tax distortions between physical and financial assets. the authors show that if there is a favorable treatment of housing assets over financial assets then the standard user cost of housing capital overestimates the actual user cost by not accounting for the bequest-related benefits of home ownership. journal of urban economics, january 1993, 33(l): 68-75. (reprinted with permission of the journal of economic literature.) ethics assessing some determinant effects of ethical consulting behav ior: the case of personal and professional values, by jeff allen and duane davis university of central florida). a random sample of 207 national business consultants is employed to test the effects of individual values and professional ethics on consulting behavior. the results suggest that the individual values held by consultants are positively corre lated with professional ethics, but are negatively correlated with consulting behav ior. moreover, there appears to be no significant relationship between the professional ethics of consultants and business consulting behavior. findings and issues regarding the effectiveness of codes of ethics and implications for both the provider and recipient of professional consulting services are discussed. ~uu~l of business ethics, june 1993, 12(6): 449-458. (reprinted with permission of the journal of business ethics.) implkationsf~rfinan~l~l planning economic well-being of disabled elderly living in the community, by marlene s. stum, jean w. bauer, and paula j. delaney. this study focuses on understanding the economic well-being of a growing subgroup of elderly, noninstitution~i~d elderly facing risks of health problems and financial dependency. the combination of predisposing characteristics and re sources that best explain differences in economic well-being of elderly was exam ined using a sample from the 1984 national long term care survey of elderly with functional limitations living in the community (n=ssoo). significant differences in economic well-being were related to age, marital status, gender, race, income sources, and education. almost one-third (32.6%) were poor and 41% had income to-needs ratios between 1.0 and 2.0 suggesting economic vulnerability. resource variables added the most explanation for differences in economic well-being, especially sources of income beyond social security. the results show the impor tance of having retirement income sources beyond social security for preventing poverty in retirement. fi~nci~~ counseling and plff~ing, 1993, 4: 199-216. (reprinted with permission of financial counseling and planning.) individual financial management behavior expectation of future financial condition: are men and women different? by vi&i schram fitzsimmons and satomi wakita. expectation of future financial condition can be a powerful motivator of an individual’s financial management. this study found no difference in male and female financial managers’ expectation of future financial condition. there were some differences in the determinants of expectation. for both sexes, positive relationships were found between expectation of financial condition in five years and (a) general locus of control, (b) financial locus of control, and (c) perception of present financial condition compared to five years ago. age was inversely related. for females, education and satisfaction with level of consumption also were related positively. implications for financial counselors and educators are given. finunciae cozmseling and planning, 1993,4: 165-180. (reprinted with permission of finun cial counseling and planning.) ~~c~lservi~~v~w,3(1) 1993 a reexamination of the investment performance of junk bonds, by terry l. zivney (ball state university), william j, bertin (university of tennessee, chattanooga) and khalil m. torabzadeh (radford univer sity). this paper reexamines the investment performance of junk bonds relative to that of corporate bonds in general. bond ratings are highly correlated with promised yield to marty for bonds of all quality classes, but are not strongly correlated with realized risk-adjusted return. the previously reported superior performance of junk bonds is shown to be a result of confusing promised yield with realized performance. those investors who value results more highly than promises will need more information than the bond rating and promised yield to maturity. quarterly journal ~~~~i~es~ and economics, spring 1993,32(l): x4-93. (reprinted with permission of quarterly journal of 3winess and economics.) corporate spin-offs and closed-end funds in a sag-preference framework, by jacques a. schnabel (wilfrid laurier university, wa terloo, ontario, canada). miller’s clientele argument for market value subadditivity as an explanation for stockholder wealth gains in corporate spin-offs and discounts from the net asset value of closed-end fund shares is examined in the context of a simple state-prefer ence model. it is shown that binding short sales constraints induce value subaddi tivity and thus render miller’s clientele argument valid. this is true regardless of whether or not divergence of opinion among investors or site-de~ndent utility fictions exist. in the absence of binding short sales cons~aints, value additivity prevails and miller’s clientele argument is not viable. although personal taxes are not considered in the model developed in this paper, it is shown that tax-timing options reinforce the existence of value subadditivity. the ~~~~~c~u~ review, august 1992, 2’7(3): 391409. (reprinted with permission of z%e fi nancial review.) dividend yields and stock returns: evidence of time variation between bull and bear markets, by michael j. gombola @exe1 university~ and feng-ying l. liu (rider college), this study documents persistent shifts in the ~lationship between stock returns and dividend yields over bull and bear markets. the shift in this relationship abstracts of articles on individual financial management 87 appears as a separate effect, distinct from the january effect, after controlling for firm size and systematic risk. after controlling for these factors dividend yield is positively related to return during bear markets but negatively related to return during bull markets. this time-varying relationship between dividend yield and stock return helps to explain the anomalous results of earlier studies. the financial review, august 1993,28(3): 303-327. (reprinted with permission of the financial review.) private benefits from block ownership and discounts on closed end funds, by michael j. barclay (university of rochester), clifford g. holdemess (boston college) and jeffrey pontiff (university of wash ington). the greater the managerial stock ownership in closed-end funds, the larger the discounts to net asset value. the average discount for funds with blockholders is 14%, whereas the average discount for funds without blockholders is only 4%. this relation is robust over time and to various model specifications that control for other factors that affect discounts. we argue that blockholders receive private benefits that do not accrue to other shareholders and that they veto open-ending proposals to preserve these benefits. we support this argument by documenting a range of potential private benefits received by blockholders in closed-end funds. jountal of financial economics, june 1993, 33(3): 263-291. (reprinted with permission of north-holland publishing company.) the effects of the insider trading sanctions act of 1984: the case of seasoned equity offerings, by thomas h. eyssell and james p. rebum (university of missouri, st. louis). previous empirical research indicates that corporate insiders tend to increase (decrease) their shareholdings before events that increase (decrease) firm value. more recent evidence suggests, however, that passage of the insider trading sanctions act of 1984 (itsa) may have deterred this behavior. our results indicate that before passage of the itsa, insiders exploited their access to nonpublic information by selling shares before the announcement of equity issues. however, after passage of the itsa insiders no longer displayed this behavior. we conclude the itsa has a deterrent effect, which is more heavily concentrated on insiders at the highest level of the firm who are most visible to regulators and other market participants. the journal of financial research, summer 1993, 16(2): 161-170. (reprinted with permission of the journal of financial research.) financial services review, 3(l) 1993 the ethical investor: exploring dimensions of investment behav ior, by paul anand and christopher j. cowton (university of oxford, oxford, uk) finance theory conventionally focuses on risk and return as the factors relevant to the construction of investment portfolios. but there is evidence of a growing number of investors who wish to incorporate moral or social concerns in their decision-making. using principal components analysis, this paper attempts to infer possible ‘non-financial’ dimensions of utility functions by considering the prefer ences of 125 ‘ethical investors’. journal ofeconomic psychology, june 1993,14(2): 377-385. (reprinted with permission of north-holland publishing company.) twenty-five years of tax law changes and investor response, by david lynn skinner (ashland university). ex-dividend day research detects dividend-clientele effects that the tax-clientele hypothesis attributes to personal taxation. in this study i exam ine all 10 personal tax changes between 1963 and 1988 and find little support for the tax-clientele hypothesis. few tax changes are accompanied by significant changes in the ex-day ratio, and more than half are opposite the direction predicted. in particular, the largest tax changes, in 1982 and 1987, fail to support the tax-cli entele hypothesis. the results are consistent with some unknown, non-tax-induced clientele effect(s). the journal offinancial research, spring 1993, 16(l): 61-70. (reprinted with permission of the journal offinancial research.) investment decisions and the theory of planned behavior, by robert east (kingston polytechnic, kingston, uk) three studies of application for shares in privatized british industries are reported. study 1 was on the regional electricity companies, study 2 was on the electricity generating companies and study 3 was on the second tranche of shares in british telecom. the studies applied ajzen’s (1991) theory of planned behavior. in each case the application for shares was accurately predicted by measured intention. intention was in turn explained by attitude, subjective norm, perceived control and past behavior. at a more specific level the research demonstrated the strong influence of friends and relatives and the importance of easy access to funds as well as the financial criteria of profit and security of investment. the research was used to test aspects of planned behavior theory and two matters were considered in detail. the first matter was the conditions under which a measure of perceived control improved on the prediction of behavior obtained from intention alone. the second matter was the association between product sum variables and globally measured variables; in the theory these are equated but rather abstracts of articles on individual financial management 89 low correlations are often found. journal of economic psychology, june 1993,14(2): 337-375. (reprinted with permission of north-holland publishing company.) “investor sentiment” and the closed-end fund puzzle: a 7 percent solution by greggory a. brauer lee, shleifer, and thaler (1991) argue that the “irrational noise trader” model of delong, shleifer, summers, and waldmann (1990) “...is consistent with the published evidence on closed-end fund prices... ” however, lee, shleifer, and thaler provide no indication of how much of the variability of a closed-end fund’s discounts and premiums is due to such “investor sentiment.” using the signal extraction technique of french and roll (1986) to measure noise, this article estimates that on average only 7% of the variance of a standardized measure of weekly changes in discounts and premiums can be attributed to noise-trading activity. “investor senti ment,” therefore, seems to account for very little of a closed-end fund’s discount and premium variability over time. journal of financial services research, september 1993,7(3): 199-216. (reprinted with permission of kluwer academic publishers.) the hidden costs of stock market liquidity, by amar bhide (har vard university). the seemingly unrelated problems of stock market liquidity and manager stockholder contracting are closely intertwined. active stockholders who reduce agency costs by providing internal monitoring also reduce stock liquidity by creating information asymmetry problems. conversely, stock liquidity discourages internal monitoring by reducing the costs of ‘exit’ of unhappy stockholders. the u.s. has exceptionally many actively-traded firms with widely-diffused stockhold ing because public policy has favored stock market liquidity over active investing. and, the benefits of stock market liquidity must be weighed against the costs of impaired corporate governance. journal of financial economics, august 1993, 34( 1): 31-5 1. (reprinted with permission of north-holland publishing company.) modelsforindividualfinanclalmanagement an exploratory study for a model of personal financial manage ment style, by kathy prochaska-cue. this article explores: (a) the initial exploratory development of a model of personal financial management style drawing from the work of mckenney and keen (1974); deacon and firebaugh (1988); gross, crandall, and knoll (1980); and rettig (1987), and (b) the initial development of an instrument to be used in measuring the proposed model of personal financial management style. responses 90 f'inancialservicesreview,3(1) 1993 of 128 adults were used to test the scales for the hypothesized instrument for validity and reliability using factor analysis. results included the identification of 14 possible items for the analyzing scale, and eight possible items for the holistic scale. future research is needed for further development of both the hypothesized model of personal financial management style and of the scales for the hypothesized instrument to measure personal financial management style. financial counseling and pluming, 1993,4: 111-134. (reprinted with permission of financial counsel ing and planning.) mutualjtijndperformance the performance of bond mutual funds, by christopher r. blake (fordham university), edwin j. elton and martin j. gruber (new york university). using linear and nonlinear models, we examine two samples of bond funds: one sample designed to eliminate survivorship bias, and a second much larger sample. overall and for subcategories of bond funds, we find that bond funds underperform relevant indexes post-expenses. our results are robust across a wide choice of models. we find that, on average, a percentage-point increase in expenses leads to a percentage-point decrease in performance. the nonlinear model weights closely match actual composition weights. we find no evidence of predictability using past performance to predict future performance for our unbiased sample. journal ofbusiness, july 1993,66(3): 371-403. (reprinted with permission of the university of chicago press.) the investment performance of u.s. equity pension fund manag ers: an empirical investigation, by t. daniel coggin (virginia retire ment system), frank j. fabozzi (fabozzi and associates), and shatiqur rahman (portland state university). this paper presents an empirical examination of the selectivity and market timing performance of a sample of u.s. equity pension fund managers. regardless of the choice of benchmark portfolio or estimation model, the average selectivity measure is positive and the average timing measure is negative. however both selectivity and timing appear to be somewhat sensitive to the choice of a benchmark when managers are classified by investment style. me&analysis revealed some real abstracts of artkles on individual financial management 91 variation around the mean values for each measure. the 80% probability intervals for selectivity revealed that the best managers produced substantial risk-adjusted excess returns. we also found a negative correlation between selectivity and timing, but we argue that the observed negative correlation in our data is largely an artifact of negatively correlated sampling errors for the two estimates. the journal of finance, july 1993,48(3): 1039-1055. (reprinted with permission of the journal of finance). do more risk-averse investors have lower net worth and in come? by thomas h. mcinish (memphis state university), sridhar n. ramaswami (iowa state university), and rajendra k. srivastava (uni versity of texas, austin). this study examines the relationship between attitude toward risk and both net worth and income. the data are obtained from a financial diary kept by a national, scientifically selected sample of more than 3000 households. results show that both net worth and income are negatively related to risk aversion. the financial review, february 1993, 28(l): 91-106. (reprinted with permission of the financial re view .) taxation tax schedule changes and discount bond prices, by ricardo j. rodriguez (university of miami). research on the impact of marginal tax changes on bondholder wealth focuses on changes along a given tax schedule. in this paper the valuation consequences of changing the tax schedule are analyzed. although previous researchers show that the price of all discount bonds fall if the marginal ordinary income tax rate increases along a tax schedule, it is found that this result holds only under specific conditions when the tax schedule changes. various comparative statistics are discussed. the journal of financial research, fall 1991,14(3): 249-254. (reprinted with permis sion of the journal of financial research.) 92 financial services review, 3(l) 1993 an ex~ination of tax portioner decisions: the role of preparer sanctions and framing effects associate-d with client condition, by kaye j. newberry, philip m.j. reekers and robert w. wyndelts (ari zona state university). tax practitioners play an impo~ant role in the vohmtary compliance system. not only do professionally prepared returns account for a significant percentage of the tax returns filed in the u.s., but empirical evidence suggests that practitioners help taxpayers lower their tax liabilities by taking advantage of ambiguous features of the tax law. this study investigates whether selected factors influence the decisions made by professional tax preparers. the subjects include 107 experienced tax practitioners who are certified public accountants. the results reflect that there is a significantly greater likelihood that the tax practitioners would sign tax returns containing a large deduction associated with an ambiguous tax issue if (1) the signing decision is made in relation to an existing client (a loss decision frame in the context of the study) or (2) tax preparer penalties are co~unicated with high enforcement intent. jou~z ofeconomic psychozugy, june 1993, 14(2): 439-452. (reprinted with permission of north-holland publishing company.) the relationship between objective financial knowledge, financial management, and financial self-efficacy among african american students kenneth whitea,*, narang parkb, kimberly watkinsc, megan mccoyd, joycelyn morrise adepartment of financial planning, housing, and consumer economics, university of georgia, dawson hall, athens, ga 30602, usa bschool of family and consumer sciences, texas state university, 601 university drive, san marcos, tx 78666, usa cdepartment of consumer sciences, university of alabama, 313 adams hall, box 870158, tuscaloosa, al 35487, usa ddepartment of personal financial planning, kansas state university, 1324 lovers lane, 343m justin hall, manhattan, ks 66502, usa edepartment of teaching and learning, florida international university, 11200 sw 8th street, zeb, miami, fl 33199, usa abstract research consistently shows the positive associations of objective financial knowledge, management, and self-efficacy on college students’ financial literacy. however, there is a need for a more nuanced examination of the factors contributing to african american college students’ financial literacy. using the national student financial wellness study and structural equation modeling, findings suggest that for african american students, objective financial knowledge is not directly or indirectly associated with financial self-efficacy. only financial management is significantly associated with increased financial self-efficacy. these findings indicate that experiential learning may be effective for improving african american students’ financial literacy. © 2021 academy of financial services. all rights reserved. jel classification: d10; d14; g53 keywords: african american; financial literacy; financial knowledge; financial management; financial selfefficacy *corresponding author. tel.: +1-706-542-4879; fax: 706-542-4397. e-mail: kjwhite@uga.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 169–185 1. introduction financial literacy is essential to increase sound financial decision-making and to reduce financial distress (asaad, 2015; huston, 2010; scott, vu, cheng, & gibson, 2018). greater financial literacy leads to healthy behaviors such as budgeting, saving for emergencies, and investing for goals such as retirement (henager & cude, 2016). given the critical nature of these skills, one might assume that researchers understand how financial literacy develops. however, research is limited by conflicting definitions and unclear paths to its development (willis, 2008). what researchers do understand thus far is that both knowledge and confidence are critical components to developing financial literacy. objective financial knowledge is a necessary component of financial literacy. however, to be a financially literate individual, one must also possess self-efficacy, or confidence, in his or her abilities to perform financial tasks (bandura, 2006). individuals with higher levels of objective financial knowledge and financial self-efficacy are associated with sound financial management leading to increased financial well-being (robb & woodyard, 2011). the link between objective financial knowledge and financial self-efficacy as integral components of financial literacy is evident. so, why has it proven so difficult to teach people these skills? basic financial education classes should increase knowledge (mielitz, macdonald, & lurtz, 2018) and literacy (al-bahrani, weathers, & patel, 2019). however, well-intentioned interventions may fall short for african american students. for example, al-bahrani et al. (2019) reported the returns on financial education are higher for whites than persons of color (poc). huston (2010) concluded that a one-size-fits-all approach to personal financial education is not effective and should instead be tailored to different demographics. if not tailored, financial education in its current form could increase the objective financial knowledge gap between whites and poc. although the personal finance literature has extensively documented the effects of objective financial knowledge and financial self-efficacy for college students, the literature is lacking when specifically addressing the roles of objective financial knowledge, financial management, and financial self-efficacy regarding african american college students’ financial literacy (alhenawi & elkhal, 2013; lown, 2011). whereas many studies use financial literacy as an input variable to explore its influence on financial behavior, the purpose of this article is to explore differences in the paths to financial literacy for african american college students. we seek to examine three paths: (1) the path from objective financial knowledge to financial self-efficacy, (2) the path from objective financial knowledge to financial management, and (3) the path from financial management to financial self-efficacy. our contribution is a comparison of african american college students to their peers in regards to objective financial knowledge, financial self-efficacy, and financial management experiences. ultimately the study seeks to understand how african american college students become financially literate individuals. the article is organized as follows: section 2 introduces the conceptual framework and hypotheses; section 3 describes the data and measures; section 4 presents the results; and section 5 is the discussion, and section 6 concludes the article. 170 k. white et al. / financial services review 29 (2021) 169–185 2. conceptual framework and hypotheses 2.1. the huston framework of financial literacy when defining what it means to be a financially literate individual, huston (2010) presents two components, knowledge (objective financial knowledge) and application (financial management ability and financial self-efficacy). huston (2010) carefully distinguishes objective financial knowledge as an integral component of financial literacy, but not equivalent to financial literacy. objective financial knowledge is acquired through both formal financial education and and/or the experience of using financial concepts to manage one’s personal finances (financial management experience) (bapat, 2019; huston, 2010). huston (2010) goes on to explain that application is a combination of ability and confidence to apply knowledge (e.g., financial self-efficacy; bandura, 1997). according to huston (2010), financially literate individuals know information (objective financial knowledge) and apply it appropriately (financial management) and confidently (financial self-efficacy). financial self-efficacy is a necessary component of literacy; the question remains if financial self-efficacy is built through knowledge or experience. please see fig. 1 for a visual depiction of this operationalization of financial literacy. 2.1.1. financial knowledge the huston (2010) framework has been used in other studies to highlight the importance of the link between objective financial knowledge and management (alhenawi & elkhal, 2013; asaad, 2015; henager & cude, 2016; seay, kim, & heckman, 2016; seay, preece, & le, 2017). research consistently shows that there is a positive relationship between objective financial knowledge and financial outcomes. as individuals’ objective financial knowledge increases, they are less likely to report over-indebtedness (varum & kolyban, 2014) and experience financial distress associated with paying bills (scott, vu, cheng, & gibson, fig. 1. operationalizing financial literacy based on huston (2010). k. white et al. / financial services review 29 (2021) 169–185 171 2018). conversely, they are more likely to report higher monthly income (varum & kolyban, 2014) and are more willing to take on investment risks (chung & park, 2015). examining the role of objective financial knowledge for african american students is critical for understanding the factors associated with this group’s financial literacy. research notes that african americans feel less financially knowledgeable than their white peers (deenanath, danes, & jang, 2019; o’connor, 2019) and are often less financially literate when compared with whites (killins, 2017). mimura et al. (2015) conclude that parental influence was a significant factor in the objective financial knowledge and financial practices of college students who are first generation, poc, immigrant or children of immigrants. factors such as a mother’s and father’s highest level of education are also positively correlated with a student’s objective financial knowledge (chambers, asarta, & farley-ripple, 2019). given these factors, it is imperative to understand how objective financial knowledge develops and affects financial literacy among african american students. 2.1.2. financial management sound financial management encompasses budgeting, saving, tracking, and spending money over time while taking into account future needs, risks, managing credit, and understanding long-term financial planning concepts such as tax, insurance, investing, retirement, and estate planning needs (bapat, 2019; henry, weber, & yarbrough, 2001; spuhlera & dew, 2019). positive relationships exist between sound financial management and accumulated savings; negative associations exist between sound financial management and accumulated consumer debt (spuhlera & dew, 2019). bapat (2019) found that objective financial knowledge, measured through the ability to answer personal finance related questions, is positively associated with financial management. additionally, sound financial management is related to reductions in financial stress and increases in financial peace of mind and wellbeing (spuhlera & dew, 2019). when it comes to financial management, african american students may be at a disadvantage. research has demonstrated that african american students tend to graduate with more debt than their white peers and face more challenges in the management of credit card debt. reality education and assets partnership (reap) found that 55% of african american students who have student loans graduate with a debt burden that is nearly twice fig. 2. conceptual framework: the relationship between objective financial knowledge, financial management, and financial self-efficacy. 172 k. white et al. / financial services review 29 (2021) 169–185 that of white graduates (dorrance & mcdaniel, 2009). it is important to explore if these financial management issues are an artifact of lacking objective financial knowledge or other issues related to socioeconomic factors. 2.1.3. financial self-efficacy self-efficacy is an individual’s sense of confidence in their ability to perform a certain skill or task to obtain specific outcomes (bandura, 1977, 2006). financial self-efficacy then is one’s confidence in their financial decision-making and management ability (bandura, 1977; farrell, fry, & risse, 2016). according to social learning theory, self-efficacy spurs individuals to confront difficult tasks, and success in these difficult tasks then expands the individual’s self-efficacy (bandura, 1994). applied to financial literacy research, there is a bidirectional relationship between self-efficacy and management behaviors. the more a person experiences the financial management behavior and succeeds (or learns from mistakes) the higher their self-efficacy will become. thus, despite many studies using financial self-efficacy as an outcome variable, there is a strong argument that can be made for it to be a predictor variable. higher financial self-efficacy is linked to more productive financial behaviors and greater well-being (amatucci & crawley, 2011; farrell et al., 2016). research provides support that financial self-efficacy is important to saving behavior and increases in net worth (asebedo & seay, 2018). students who are more confident in their financial decision-making and feel more knowledgeable engage in more healthy financial management (deenanath, danes, & jang, 2019). although financial self-efficacy is an important component of financial literacy, when self-efficacy is not grounded in appropriate knowledge it can result in overconfidence (mccoy et al., 2019). overconfidence can be harmful and result in higher risk taking, less investment diversification, excessive borrowing, and may be a deterrent to seeking professional financial advice (angrisani & casanova, 2019; atlas et al., 2019; hauff & nilsson, 2020; kim, lee, & hanna, 2020; merkle, 2017). previous studies on college students note financial self-efficacy to be related to positive outcomes like lower stress (heckman, lim, & montalto, 2014; lim, heckman, letkiewicz, & montalto, 2014), higher subjective well-being, and negatively associated with credit hour reductions (robb, 2017). although race was included in two of these studies on college students (heckman et al., 2014; lim et al., 2014), to the authors’ knowledge, no study has directly explored how race impacts financial self-efficacy among college students. 2.1.4. importance of examining race/ethnicity despite the lack of explicit studies on how different racial or ethnic groups vary in financial self-efficacy, there has been research that may suggest differences in financial self-efficacy by race/ethnicity. first, racial differences in rates of financial self-efficacy may stem from the differences in rates of mathematical self-efficacy. alliman-brissett & turner (2010) found that perceived racism negatively impacted mathematical self-efficacy (i.e., the k. white et al. / financial services review 29 (2021) 169–185 173 self-efficacy related to math-related tasks and to pursue math careers) in african american youth. it is not a large leap to imagine that self-efficacy in math would be highly correlated to one’s self-efficacy with money (almenberg & widmark, 2011; grohmann, kouwenberg, & menkhoff, 2015; skagerlund, lind, strömbäck, tinghög, & västfjäll, 2018; jayaraman, jambunathan, & counselman, 2018). second, oliver and shapiro (2013) describe how the “racialization of state policy” has led to a long legacy of wealth differences that have reinforced racial inequalities in the united states (p. 39). socialization messages within african american families around how much wealth one can achieve may be shaped by these inequalities. generations of barriers that prevented equal access to the means of generating wealth may limit african american’s financial self-efficacy. finally, an important component of financial self-efficacy is the vicarious experiences of financial management (bandura, 1997). bandura (1997) suggests that we develop self-efficacy through experiencing the behavior itself. research shows that the african american community is disproportionately underbanked or unbanked (breitbach & walstad, 2014). potentially, this has led to african american’s having fewer financial management experiences resulting in lower levels of financial self-efficacy. there remains a large racial/ethnic wealth and income gap in the united states that may impact differences in financial literacy between african american and non-african american students (hamilton & darity, 2017). differences in financial literacy are not an inherent trait of being one race or ethnicity per se, but instead are associated with financial experiences. financial literacy then is tied to wealth or lack of wealth. financial behaviors, such as paying bills on time, investing and saving for retirement, are limited for individuals with few financial resources to manage (hamilton & darity, 2017; hamilton et al., 2015). in fact, if household income were equal, african american families would have a slightly higher savings rate than white families (hamilton & darity, 2017). although the literature may hint at differences between african american and nonafrican american students, there is not enough empirical data to presuppose a directional association between the components of huston’s financial literacy framework and race. therefore, the three hypotheses in this study were derived from the conceptual framework (huston, 2010) and the extant literature to examine the relationships between objective financial knowledge, financial management, and financial self-efficacy (see fig. 2). hypothesis 1: african american and non-african american students’ objective financial knowledge is positively associated with their financial self-efficacy. hypothesis 2a: financial management is a mediating factor in the relationship between objective financial knowledge and financial self-efficacy among african american and nonafrican american students. hypothesis 2b: african american and non-african american students’ financial management is positively associated with their financial self-efficacy. based on these hypotheses, the conceptual model (fig. 1) is used to explore the relationships of the components of financial literacy and compare the relationships by race. 174 k. white et al. / financial services review 29 (2021) 169–185 3. method 3.1. data and sample selection this study relies upon data collected at the ohio state university through the 2014 national student financial wellness study (nsfws).1 this survey is administered to undergraduate students (n = 18,795) from 52 participating two and four-year public institutions and four-year private higher education institutions across the united states. to learn more about the methodology of the study on collegiate financial wellness (scfw), please see montalto, phillips, mcdaniel, and baker (2019). questions relating to financial attitudes, financial management, and objective financial knowledge capture a picture of the overall financial literacy and wellness of undergraduate students in the united states. the original data sample size is 18,792. the sample includes 965 african american students and 13,697 non-african american students, excluding missing answers. to balance sample size, we drew random samples from the existing pools. first, we extracted several sets of sample combinations that are 350 samples for each group because structural equation modeling requires at least 350 samples to ensure the significant factor loadings for latent constructs (hair, black, babin, &anderson, 2014). next, the sample size of each group was reduced after eliminating observations with missing answers. however, due to limitations with the data, we encountered difficulty extracting the same number of sample sizes for each group. we chose samples with a smaller difference in size between the two groups. the total sample size was 860, 394 for african american and 466 for non-african american. 3.2. variables our model is constructed with three latent variables: financial knowledge, financial management, and financial self-efficacy. a latent variable is an underlying concept that cannot be observed or directly measured. instead, it can be inferred by several indicators, or observed variables. the following section describes how we manifested the latent constructs with a set of indicator variables. 3.2.1. financial self-efficacy the outcome variable, financial self-efficacy, is a latent variable constructed by two observed indicators: confidence in finances and confidence in money management. confidence in finances is measured by the item: “i am confident that i can manage my finances.” respondents are asked to indicate 1 = strongly disagree, 2 = disagree, 3 = agree, or 4 = strongly agree. confidence in money management is a variable measured by the item: “i manage my money well.” respondents are also given four options to indicate 1= strongly disagree, 2 = disagree, 3 = agree, or 4 = strongly agree. k. white et al. / financial services review 29 (2021) 169–185 175 3.2.2. objective financial knowledge objective financial knowledge is a latent variable consisting of three observed variables. each variable is measured by a question testing respondents’ objective financial knowledge (see appendix a). the three questions have been widely used in prior studies (al-bahrani et al., 2019; o’connor, 2019; robb & woodyard, 2011; seay, preece, & le, 2017). the variables are coded as 1 if a respondent provides the correct answer, or 0 otherwise. 3.2.3. financial management financial management is a latent variable manifested by three indicators. the financial management variable represents respondents’ normative financial behavior. the indicator variables are budgeting, tracking spending, and tracking transactions. the variables are measured by the following items: “i have a weekly or monthly budget that i follow (budgeting),” “i track my spending to stay within my budget (tracking spending),” and “i track all debit card transactions/checks to balance my account (tracking transactions).” respondents are given four options to choose 1 = never, 2 = sometimes, 3 = frequently, or 4 = always. 3.3. covariates this study includes eight control variables: age, gender, employment, annual income, marital status, having child(ren), first generation status, and financial education. these socio-demographic variables are known to be associated with financial management and financial self-efficacy (al-bahrani et al., 2019; alhenawi & elkhal, 2013; chambers, asarta, & farley-ripple, 2019; harrington & smith, 2016; henager & cude, 2016; tang & peter, 2015; varum & kolyban, 2014; wagner, 2019). age is a numeric value. gender is coded as 1 = female, or 0 = male. employment is coded as a dummy variable indicating 1 = employed and 0 = not employed. the “employed” category includes respondents who are working full-time, part-time, or self-employed. annual income is also coded dichotomously: 1 = above median income, or 0 = median income or less. because there is no direct measurement for marital status, we use a proxy variable measured by an item asking, “are you financially responsible for a spouse/partner?” having child(ren) is coded as 1 if respondents have any financially dependent child or children, or 0 otherwise. first generation status is an indicator of parents’ education. it defines whether either parent completed a bachelor’s degree. first generation is coded as 1 = first generation student, or 0 = not first generation student. finally, financial education indicates whether respondents attended any personal finance classes/workshops when they were in high school or in college. the variable is coded dichotomously (1 = yes, 0 = no). 3.4. model analysis the analytic procedure is as follows. first, we specify the model to be tested. the variables are selected from the available information within the data. the relationships among 176 k. white et al. / financial services review 29 (2021) 169–185 the variables are hypothesized based on previous literature. similar to bapat (2019), we employed structural equation modeling (sem) to test the hypothesized relationship among the variables of interest. sem is a statistical technique that conducts confirmatory factor analysis and path analysis simultaneously. confirmatory factor analysis (cfa) examines whether a latent variable is conceptually well represented by its observed indicators. the strength of relationship between each indicator variable and a latent variable is assessed by factor loading. factor loadings greater than 0.5 are generally considered as high; however, 0.3 and 0.4 are minimally acceptable if the sample size is 350 or larger (hair et al., 2014; kline, 1994). path analysis identifies the significance of hypothesized relationships among the latent variables. path analysis helps understand the sequential associations of the variables. as structural equation modeling takes advantage of both latent variable analysis (i.e., cfa) and path analysis, it is useful when testing a conceptual model as a whole. while traditional regressions are powerful when analyzing the marginal effect of an independent variable (when holding other variables constant) on a dependent variable, they have limited ability in examining the sequential relationships among the relevant variables. also, structural equation modeling minimizes the measurement errors (bollen, 1989). the second hypothesis of this study is to examine the mediating role of financial management. thus, we chose structural equation modeling as our analytic method because it is the most appropriate method to analyze our research questions. to compare the results of african american to the results of non-african american students, we applied a group comparison option when conducting sem. the path coefficients are standardized for comparison. the model was tested using stata 15. 4. results 4.1. descriptive characteristics of respondents table 1 shows the demographic characteristics of the sample. the average age of african american students (27.02) is higher than that of non-african american students (23.68). the sample is skewed to female for both african american students (74%) and non-african american students (72%). for employment status, the percentage of working full-time, part-time or self-employed is higher for african american students (73%) compared with non-african american (69%). these results may indicate a necessity for african american students to provide more of their own financial support than their non-african american counterparts. both african american students and non-african american students show a low rate of having a spouse or partner; however, more african american students (27%) are financially responsible for a child or children than their non-african american counterparts (14%). the descriptive statistics indicate more african american students (57%) than non-african american students (43%) are the first generation attending college. financial education experience of african american students and non-african american is 49% and 42%, respectively. regarding objective k. white et al. / financial services review 29 (2021) 169–185 177 financial knowledge, the percentage of correct answers for each question is higher in non-african american students. with regards to financial management, the occurrence of three management behaviors (i.e., have budget, track spending, or track account balance) are similar between african american students and non-african american students. overall, financial self-efficacy is slightly higher for non-african american students. 4.2. structural equation modeling results figs. 3 and 4 are the results of structural equation modeling. latent variables are graphically expressed by ovals and observed variables are visually represented by rectangles. the set of covariates (e.g., age, gender, employment, annual income, marital status, having child table 1 descriptive statistics of respondents (n = 860) african american (n = 394) non-african american (n = 466) variables % mean (se) % mean (se) age 27.02 (10.89) 23.68 (7.03) gender male 26.40 27.90 female 73.60 72.10 employment status employed 72.59 69.31 not employed 27.41 30.69 annual income median or less 52.03 61.59 above median 47.97 38.41 marital status having spouse/partner 12.94 11.80 no spouse/partner 87.06 88.20 having child(ren) yes 26.90 13.95 no 73.10 86.05 first generation yes 57.36 43.13 no 42.64 56.87 financial education yes 48.73 42.27 no 51.27 57.73 objective financial knowledge financial knowledge1 (correct) 48.22 61.59 financial knowledge2 (correct) 73.10 78.76 financial knowledge3 (correct) 72.59 82.62 financial management budgeting 2.68 (0.94) 2.58 (0.97) tracking spending 2.93 (0.88) 2.94 (0.94) tracking transactions 3.15 (0.91) 3.09 (1.05) financial self-efficacy confidence in finances 3.08 (0.75) 3.09 (0.67) confidence in money management 2.91 (0.72) 3.04 (0.66) 178 k. white et al. / financial services review 29 (2021) 169–185 (ren), first generation status, or financial education) are included in the pathways to control for the background characteristics. in sem, the goodness of the model is evaluated by several fit indices, including the x2 statistic, the root mean square error of approximation (rmsea), standardized root mean square residual (srmr), and comparative fit index (cfi). kline (2005) recommended at least these four indices should be reported. the model x2 test should be statistically insignificant if the model fit is good; however, this rule does not apply to a large sample because the x2 statistic is sensitive to sample size (schermelleh-engel, moosbrugger, & müller, 2003). generally, the model is considered to be good if rmsea is lower than 0.06, srmr is less than 0.08, and cfi is higher than 0.90 (hu & bentler, 1999; loehlin, 2004). the tested model statistics are x2 = 250.38 (df = 126, p < .001), rmsea = 0.048, srmr = 0.038, cfi = .920. based on the thresholds of goodness-of-fit indices, the model shows a good fit. fig. 3 shows the results for the african american students. when evaluating construct validity, all indicators are significantly loaded on each corresponding latent factor. it means the three latent variables (objective financial knowledge, financial management, and financial self-efficacy) are properly represented by its observed variables. the fig. 3. results from the structural equation modeling of african american college students with standardized coefficients. note. control variables are age, gender, employment, annual income, marital status having child(ren), first generation, financial education. *p < .05, **p < .01, ***p < .001. fig. 4. results from the structural equation modeling of non-african american college students with standardized coefficients. note. control variables are age, gender, employment, annual income, marital status having child(ren), first generation, financial education. *p < .05, **p < .01, ***p < .001. k. white et al. / financial services review 29 (2021) 169–185 179 results also demonstrate the significance of the hypothesized relationship among the latent constructs. first, the effect of objective financial knowledge on financial management is not significant (b = 0.042, p = .636). this indicates, for african american students, objective financial knowledge contributes little in forming positive financial management. second, objective financial knowledge is also not significantly linked to financial self-efficacy (b = 0.152, p = .069). this means objective financial knowledge does not suffice to build financial self-efficacy for african american students. third, financial management is greatly and positively associated with financial self-efficacy (b = 0.580, p < .001). in short, only financial management is effective to establish financial self-efficacy for african american students. fig. 4 shows the results for non-african american students. all factor loadings for latent constructs are fairly high and significant at the level of .001. the associations among the latent variables are found to be significant. first, objective financial knowledge is positively associated with financial management (b = 0.158, p < .05) and financial self-efficacy (b = 0.153, p < .05). this indicates, for non-african american students, higher objective financial knowledge is linked to more positive financial management behavior and higher financial self-efficacy. second, financial management is positively related to financial selfefficacy (b = 0.326, p < .001). 5. discussion this study contributes to the literature by examining how objective financial knowledge, financial management, and financial self-efficacy are associated with african american college students’ financial literacy. the summary of results are found in table 2. results show that objective financial knowledge is positively associated with financial self-efficacy supporting hypothesis 1 among the non-african american sample. however, when examining the link between objective financial knowledge and financial efficacy for only african american students, this association does not exist. similarly, for the non-african american sample, there is a significant indirect relationship from objective financial knowledge to financial self-efficacy mediated by financial management. both objective financial knowledge and financial management are positively related to their financial self-efficacy. this provides support for both hypothesis 2a and hypothesis 2b. however, this relationship again differs for the african american students in the study. there is only a significant link from ‘financial management’ to ‘financial self-efficacy’. table 2 summary of results hypothesis african american non-african american h1 objective financial knowledge ! financial management not supported supported h2a objective financial knowledge ! financial self-efficacy not supported supported h2b financial management! financial self-efficacy supported supported 180 k. white et al. / financial services review 29 (2021) 169–185 these results suggest that financial management experience is important for african american students to attain financial self-efficacy. the findings of this study appear to suggest that simply having objective financial knowledge is not as powerful a tool as providing experiences to counteract this difference in african american communities. it is important for african american students to manager finances to acquire the ability and confidence to become financially literate individuals. 6. conclusions the results of this study show it is important to not generalize findings of studies to all populations. more research is needed to explore between group differences to ensure that best practices are truly best for all groups. for instance, our study shows that an essential ingredient in financial education programs is a focus on building financial self-efficacy. for african american students, providing opportunities for learning through financial management experiences must be an important component of financial education (henager & cude, 2016; robb & woodyard, 2011). thus, financial education programs focusing on experiential learning may enhance financial literacy among african american students and ultimately mitigate financial problems that individuals and families face (huston, 2010). this is consistent with tang and peter (2015) who suggest that hands-on experience and application-orientated financial education is effective in improving the acquisition of objective financial knowledge. established university-based, peer-mentoring or counseling programs (e.g., the aspire clinic at the university of georgia and powercat financial at kansas state university) are invaluable resources to increase the self-efficacy of african american students through financial management experiences. providing financial education in its current form may increase the objective financial knowledge gap between white and african american college students. although financial literacy is associated with positive financial outcomes, we must acknowledge that financial literacy alone does not influence positive or negative financial outcomes. factors such as culture, family, economic, and institutional policies may result in individuals engaging in ineffective personal financial behaviors that disrupt their financial well-being (huston, 2010). this study highlights the importance of offering a variety of financial education instructional methods to make targeted financial outcomes accessible to a broader and more diverse range of students. it is important to note some of the study’s limitations. due to the cross-sectional nature of the data, causations cannot be concluded from these results. future research should focus on the use of longitudinal data or experimental designs to find direct causes. in addition, our findings suggest the need for experiential learning, yet an experimental design is needed to formalize the best practice suggestion. furthermore, more research is needed to explore the relationship between objective financial knowledge, financial self-efficacy, and financial management experiences in african americans to provide more insights in the underlying factors that cause k. white et al. / financial services review 29 (2021) 169–185 181 different results across racial groups. current and future financial education curricula should focus not only on how individuals perform financial management behaviors but also how to help increase a person’s financial self-efficacy as this is an essential component of one’s ability to be financially literate. as aforementioned, there is a bidirectional relationship between financial self-efficacy and financial management behaviors. in the current study, we tested how objective financial knowledge, financial management, and financial self-efficacy, are associated with college students’ financial literacy across racial groups. further research is needed to explore any potential for reverse causality in how we tested the relationship between these variables. despite these limitations, financial practitioners and educators can use these findings to create culturally responsive financial education programs to help increase financial self-efficacy and close the objective financial knowledge and literacy gap between african americans and other races/ethnicities. note 1 in 2014, when the data were originally collected and released for use, it was titled nsfws. the name was later changed to scfw for subsequent data collection beginning in 2017. references al-bahrani, a., weathers, j., & patel, d. (2019). racial differences in the returns to financial literacy education. journal of consumer affairs, 53, 572–599. alhenawi, y., & elkhal, k. (2013). financial literacy of us households: knowledge vs. long-term financial planning. financial services review, 22, 211–245. alliman-brissett, a. e., & turner, s. l. (2010). racism, parent support, and math-based career interests, efficacy, and outcome expectations among african american adolescents. journal of black psychology, 36, 197–225. appendix a financial knowledge items financial knowledge questions scale (1) imagine that the interest rate on your savings account is 1% per year and inflation is 2% per year. after 1 year, would you be able to buy more than today, exactly the same as today, or less than today with the money in this account? 1 = more than today 2 = exactly the same as today 3 = less than today 4 = don’t know (2) suppose you have $100 in a savings account and the interest rate was 2% per year. after 5 years, how much would you have in the account if you left the money to grow? 1 = more than $102 2 = exactly $102 3 = less than $102 4 = don’t know (3) suppose you borrowed $5,000 to help cover college expenses for the coming year. you can choose to repay this loan over 10 years, 20 years, or 30 years. which of these repayment options will cost you the least amount of money over the length of the repayment period? 1 = 10-year repayment option 2 = 20-year repayment option 3 = 30-year repayment option 4 = don’t know 182 k. white et al. / financial services review 29 (2021) 169–185 almenberg, j., & widmark, o. (2011). numeracy, financial literacy and participation in asset nmarkets. available at ssrn. amatucci, f. m., & crawley, d. c. (2011). financial self-efficacy among women entrepreneurs. international journal of gender and entrepreneurship, 3, 23–37. angrisani, m., & casanova, m. (2019). what you think you know can hurt you: under/over confidence in financial knowledge and preparedness for retirement. journal of pension economics and finance, 1–16. asaad, c. t. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101–118. asebedo, s., & seay, m. (2018). financial self-efficacy beliefs and the saving behavior of older pre-retirees. journal of financial counseling and planning, 29, 357–368. atlas, s. a., lu, j., micu, p. d., & porto, n. (2019). financial knowledge, confidence, credit use, and financial satisfaction. journal of financial counseling and planning, 30, 175–190. bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84, 191–215. bandura, a. (1994). self-efficacy. in v. s. ramachaudran (ed.), encyclopedia of human behavior (vol. 4, pp. 71–81). san diego, ca: academic press. bandura, a. (1997). self-efficacy: the exercise of control.w h freeman/times books/ henry holt & co. bandura, a. (2006).guide for constructing self-efficacy scales. in t. urdan & f. pajares (eds.), self-efficacy beliefs of adolescents (pp. 107–123). charlotte, nc: information age publishing. bapat, d. (2019). exploring antecedents to financial management behavior for young adults. journal of financial counseling and planning, 30, 44–55. bollen, k. a. (1989). a new incremental fit index for general structural equation models. sociological methods & research, 17, 303–316. breitbach, e., & walstad, w. b. (2014). financial literacy and bank affiliation: results for the unbanked, underbanked, and fully banked, perspectives on economic education research, 9, 21–35. chambers, r. g., asarta, c. j., & farley-ripple, e. n. (2019). gender, parental characteristics, and objective financial knowledge of high school students: evidence from multicountry data. journal of financial counseling and planning, 30, 97–109. chung, y., & park, y. (2015). do financial networks matter in retirement investment decisions? evidence from generation yers. financial services review, 24, 77–100. deenanath, v., danes, s. m., & jang, j. (2019). purposive and unintentional family financial socialization, subjective financial knowledge, and financial behavior of high school students. journal of financial counseling and planning, 30, 83–96. dorrance, j., & mcdaniel, m. (2009). the reality education and assets partnership: making the case for student financial management programs at minority serving institutions. chapel hill, nc: center for community capital. available at https://communitycapital.unc.edu/files/2009/12/reap.pdf farrell, l., fry, t. r., & risse, l. (2016). the significance of financial self-efficacy in explaining women’s personal finance behaviour. journal of economic psychology, 54, 85–99. grohmann, a., kouwenberg, r., & menkhoff, l. (2015). childhood roots of financial literacy, diw discussion papers, no. 1504. berlin: deutsches institut für wirtschaftsforschung (diw). hair, j. f., black, w. c., babin, b. j., & anderson, r. e. (2014). multivariate data analysis (7th edition). london, uk: pearson. hamilton, d., & darity, w. a. , jr., (2017). the political economy of education, financial literacy, and the racial wealth gap. federal reserve bank of st. louis review, 99, 59–76. hamilton, d., darity jr, w., price, a. e., sridharan, v., & tippett, r. (2015). umbrellas don’t make it rain: why studying and working hard isn’t enough for black americans. new york, ny: the new school. harrington, c., & smith, w. (2016). college student interest in personal finance education. financial services review, 25, 351–372. hauff, j. c., & nilsson, j. (2020). determinants of indebtedness among young adults: impacts of lender guidelines, explicit information and financial (over) confidence. international journal of consumer studies, 44, 89–98. k. white et al. / financial services review 29 (2021) 169–185 183 heckman, s., lim, h., & montalto, c. (2014). factors related to financial stress among college students. journal of financial therapy, 5, 19–39. henager, r., & cude, b. j. (2016). financial literacy and long-and short-term financial behavior in different age groups. journal of financial counseling and planning, 27, 3–19. henry, r. a., weber, j. g., & yarbrough, d. (2001). money management practices of college students. college student journal, 35, 244–249. hu, l. t., & bentler, p. m. (1999). cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. structural equation modeling: a multidisciplinary journal, 6, 1–55. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. jayaraman, j. d., jambunathan, s., & counselman, k. (2018). the connection between financial literacy and numeracy: a case study from india. numeracy, 11, 5. killins, r. n. (2017). the financial literacy of generation y and the influence that personality traits have on financial knowledge: evidence from canada. financial services review, 26, 143–165. kim, k. t., lee, j., & hanna, s. d. (2020). the effects of financial literacy overconfidence on the mortgage delinquency of us households. journal of consumer affairs, 54, 517–540. kline, p. (1994). an easy guide to factor analysis. london, uk: routledge. kline, r. b. (2005). principles and practice of structural equation modeling. new york, ny: guilford press. lim, h., heckman, s., montalto, c. p., & letkiewicz, j. (2014). financial stress, self-efficacy, and financial help-seeking behavior of college students. journal of financial counseling and planning, 25, 148–160. loehlin, j. c. (2004). latent variable models: an introduction to factor, path, and structural equation analysis (4th ed.). mahwah, nj: lawrence erlbaum. lown, j. m. (2011). development and validation of a financial self-efficacy scale. journal of financial counseling and planning, 22, 54–63. mccoy, m., white, k. j., & love, k. (2019). investigating the financial overconfidence of student-athletes. sport, business management: an international journal, 9, 381–398. merkle, c. (2017). financial overconfidence over time: foresight, hindsight, and insight of investors. journal of banking & finance, 84, 68–87. mielitz, k. k. s., macdonald, m., & lurtz, m. (2018). financial literacy education in a work release program for an incarcerated sample. journal of financial counseling and planning, 29, 316–327. mimura, y., koonce, j., plunkett, s. w., & pleskus, l. (2015). financial information source, knowledge, and practices of college students from diverse backgrounds. journal of financial counseling and planning, 26, 63–78. montalto, c. p., phillips, e. l., mcdaniel, a., & baker, a. r. (2019). college student financial wellness: student loans and beyond. journal of family and economic issues, 40, 3–21. oliver, m., & shapiro, t. (2013). black wealth/white wealth: a new perspective on racial inequality. new york, ny: routledge robb, c. a., & woodyard, a. s. (2011). financial knowledge and best practice behavior. journal of financial counseling and planning, 22, 60–70. robb, c. a. (2017). college student financial stress: are the kids alright? journal of family and economic issues, 38, 514–527. scfw. (2017). study on collegiate financial wellness: 2017 key findings report. the ohio state university, columbus, oh. available at https://cssl.osu.edu/posts/632320bc-704d-4eef-8bcb-87c83019f2e9/documents/ 2017-scfw-key-findings-report.pdf schermelleh-engel, k., moosbrugger, h., & müller, h. (2003). evaluating the fit of structural equation models: tests of significance and descriptive goodness-of-fit measures. methods of psychological research monline, 8, 23–74. scott, j. k., vu, n. n., cheng, y., & gibson, p. (2018). financial capability: literacy, behavior, and distress. financial services review, 27, 391–411. seay, m., kim, k. t., & heckman, s. (2016). exploring the demand for retirement planning advice: the role of financial literacy. financial services review, 25, 331–350. seay, m. c., preece, g. l., & le, v. c. (2017). financial literacy and the use of interest-only mortgages. journal of financial counseling and planning, 28, 168–180. 184 k. white et al. / financial services review 29 (2021) 169–185 shapiro, t., meschede, t., & osoro, s. (2013). the roots of the widening racial wealth gap: explaining the black-white economic divide. [policy brief]. institute on assets and social policy. available at https://heller. brandeis.edu/iasp/pdfs/racial-wealth-equity/racial-wealth-gap/roots-widening-racial-wealth-gap.pdf skagerlund, k., lind, t., strömbäck, c., tinghög, g., & västfjäll, d. (2018). financial literacy and the role of numeracy–how individuals’ attitude and affinity with numbers influence financial literacy. journal of behavioral and experimental economics, 74, 18–25. spuhlera, b. k., & dew, j. (2019). sound financial management and happiness: economic pressure and relationship satisfaction as mediators. journal of financial counseling and planning, 30, 157–174. tang, n., & peter, p. c. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience. financial services review, 24, 119–138. varum, c., & kolyban, a. (2014). wealth and credit compliance: does economic literacy matter? financial services review, 23, 325–339. wagner, j. (2019). financial education and financial literacy by income and education groups. journal of financial counseling and planning, 30, 132–141. willis, l. (2008). against financial literacy education. paper 208. philadelphia, pa: university of pennsylvania law school. available at https://heinonline.org/hol/landingpage?handle=hein.journals/ ilr94&div=7&id=&page= k. white et al. / financial services review 29 (2021) 169–185 185 pii: s1057-0810(99)00034-7 from the editor karen eilers lahey this issue of volume 8 has three articles that examine various aspects of individual portfolio decisions. a fourth article offers an unusual alternative to shielding potential homeowners from interest rate risk and the fifth tests the impact of planning on the individual’s satisfaction with retirement. the steven c. gold and paul lebowitz article entitled, “computerized stock screening rules for portfolio selection,” uses a commercial program that screens stocks to determine if portfolios can be constructed that earn excess returns. the overwhelming amount of information available on the internet suggests the need for this article and future ones that test the information sources that are being accessed by individuals. “municipal bonds: a contingent claims perspective,” by robert brooks, provides an overview of a market that has received remarkably little academic attention. he provides a tutorial on the embedded contingent claims on these types of bonds and cautions investors on factors to consider in including this type of asset in their portfolio. the third article by john macdonald and david m. smith, entitled “investor partitioning of the components of value in corporate earnings announcements” uses a unique data base to test the dividend-related and capital gains-based returns for large corporate stocks based on newly available earnings information. their results suggest that the majority of the value gained from positive announcements is reflected in capital gains rather than dividends. in addition to all of the normal risks of purchasing a home, individuals must also deal with potential changes in interest rates for permanent mortgages. terry l. zivney and carl f. luft provide a simulation for hedging this risk in their article entitled, “hedging individual mortgage risk.” harold w. elder and patricia m. rudolph in an article entitled, “does retirement planning affect the level of retirement satisfaction” present a data base that should be of interest to those researching retirement issues. the authors provide a review of the literature in this area from several disciplines and rigorously examine the impact of thinking about and attending planning sessions on individual’s satisfaction with retirement. financial services review 8 (1999) v 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00034-7 improving collegiate financial literacy via financial education seminars jonathan handya, beth pontarib, thomas smythec, suzy summersd,* adepartment of finance, western kentucky university, 1906 college heights boulevard, bowling green, ky 42101, usa bdepartment of psychology, furman university, 3300 poinsett highway, greenville, sc 29607, usa cdepartment of economics and finance, florida gulf coast university, 10501 fgcu boulevard south, fort myers, fl 33965, usa ddepartment of business and accounting, furman university, 3300 poinsett highway, greenville, sc 29607, usa abstract this paper describes a personal finance program (pfp) developed at a private liberal arts university aimed to improve financial literacy. we provide a program overview, with details about recruiting, program structure, and curriculum. using a multivariate framework, we examine program effectiveness at improving students’ financial knowledge and confidence in their financial future. our findings demonstrate that financial knowledge and confidence improve. additionally, women (minorities) narrow their financial knowledge and confidence gaps when compared with men (caucasians) and the control group. finally, follow-up analyses show that increases in confidence appear justified (vs. misplaced) in that they are calibrated to increases in knowledge. © 2021 academy of financial services. all rights reserved. jel classifications: g0; g5; i3 keywords: financial literacy; education; retirement planning; personal finance 1. introduction as this paper was being written, the coronavirus disease 2019 (covid-19) pandemic was ravaging the globe, bringing with it not only illness but significant economic uncertainty. in *corresponding author: tel.: +1-864-294-3313 email address: suzy.summers@furman.edu 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 315–341 the united states, state and local governments closed businesses deemed “non-essential.” world-wide responses were similar, as countries placed travel restrictions on their citizens and enacted various economic restrictions (https://www.aljazeera.com/news/2020/03/ coronavirus-travel-restrictions-border-shutdowns-country-200318091505922.html). the closures brought massive unemployment and income loss. while much about the pandemic was and remains uncertain, one consistent issue was and is apparent: most individuals and families were unprepared for such an economic shock. one likely explanation for the lack of preparedness is low financial literacy levels, which has been documented in a variety of studies (lusardi & mitchell, 2014). it is imperative that tools be developed to improve personal financial literacy and that these tools be implemented quickly and in as many forums as possible. although this study began well before the covid-19 pandemic, the pandemic serves as an important reminder that financial literacy is essential for a highly functioning society. each day, individuals and households make financial decisions that have greater and furtherreaching consequences. households must manage their consumer credit levels (currently over $4 trillion in the united states, see https://www.federalreserve.gov/releases/g19/ current/ as of april 2020), including credit cards, student loans, car payments, and other miscellaneous debt finances, as well as mortgage debt. households must manage savings and spending decisions and decide what investment types should be utilized. finally, households must plan for, set aside money for, and make investment decisions about retirement, a responsibility unheard of a generation ago. therefore, while households are advantaged by increased access to financial markets, they must first have the financial knowledge necessary to interact with such markets or risk potentially drastic financial consequences.1 additionally, individuals must have confidence in their ability to make these decisions effectively and manage their financial future. for example, research (farrell, fry, & risse, 2016) shows that women’s confidence in their financial capabilities predicted positive financial behaviors (i.e., they were more likely to have financial products related to saving and investing funds). at the same time, other research suggests that if confidence is overblown and not commensurate with knowledge, this could lead to poor decision making and engaging in costly behaviors (krueger & dunning, 1999; toker asad, 2015). unfortunately, existing research shows that generally, people lack financial knowledge and confidence. for example, the 2004 health and retirement study (hrs) finds that among older individuals, financial knowledge is quite low, with only 34.3% of respondents correctly answering a series of three questions regarding financial calculations, and only 70% answering two of the three questions correctly (lusardi & mitchell, 2011). these same three questions have been used in other studies gauging financial knowledge (e.g., the rand american life panel; lusardi & mitchell, 2011, 2017; lusardi, mitchell, & curto, 2010). all of these studies report findings similar to the hrs study: financial knowledge in the united states is quite low. for an exhaustive literature review documenting low financial knowledge levels, see lusardi and mitchell (2014). thus, in our research, we presume that financial knowledge is low and that it needs to improve. as such, the current study explores the effectiveness of a financial education program for graduating seniors at a liberal arts university intended to improve participant financial literacy. the program was developed because the institution had limited offerings of a for-credit personal finance 316 j. handy et al. / financial services review 29 (2021) 315–341 course. on average, there was annual capacity for approximately 25 students in the for-credit course, but it was common for the course to be oversubscribed by 100 or more students. the personal finance program (pfp) was developed to address the gap in offering this kind of course and provide a controlled opportunity to assess its effectiveness in increasing financial literacy. we define financial literacy as “people’s ability to process economic information and make informed decisions about financial planning, wealth accumulation, debt, and pensions” (lusardi & mitchell, 2014). this definition encompasses two critical dimensions: knowledge and behavior. most research focuses on knowledge or attempts to infer behavior from data such as the survey of consumer finances. or, previous research associates higher knowledge with “better” behaviors (e.g., fox, bartholomae, & lee, 2005; henager & cude, 2016; lusardi, 2008), but due to the correlational nature of that research, the conclusions are inferences at best. to our knowledge, there is no work directly examining programs aimed at improving knowledge and any subsequent behavioral changes that follow. the current study addresses the first step in increasing financial literacy. that is, we measure baseline levels of financial knowledge for a treatment group and a control group, then expose the treatment group to the pfp, and then measure financial knowledge again for both groups. we also measure changes in confidence level. as described above, there are competing predictions about the potential benefits of increasing confidence along with knowledge. nonetheless, research seems to agree upon the notion that increased confidence is a benefit, as long as it is matched with increased knowledge. indeed, toker asad (2015) summarized, that “financial literacy initiatives should focus not only on factual knowledge, but on helping individuals achieve a healthy dose of confidence.” thus, we ask participants how confident they are in managing their financial future (not how confident they are in their answers or knowledge) and can assess if this increases alongside actual financial knowledge. to summarize, the current research gauges the pfp’s efficacy in improving financial knowledge and increasing confidence in the ability to manage one’s financial future. using multivariate regression techniques, our analysis finds that program participants improve both financial knowledge and confidence about their financial future, after controlling for other factors that might influence these outcomes for college students. the results suggest that, although pfp and control group participants do not differ in knowledge at baseline, pfp participants show significantly higher financial knowledge levels from the preto the post-test, as well higher financial knowledge than control subjects after completing the pfp. the findings for participants’ confidence in their ability to manage their financial future are similar; participants are more confident about their financial outlook compared with before the pfp, and the control group shows no change in confidence over time. on the surface, the finding about improved confidence may seem suspect, given that people are often overconfident about their financial future. however, additional analysis demonstrates that seminar participants not only increase their confidence, but that this increase is calibrated with increases in knowledge level, a result not found for the control group. a particularly encouraging set of findings relate to participant gender and minority classification. according to pre-pfp survey results, and consistent with prior work (lusardi & mitchell, 2014 and others), individuals who are female or represent a minority group scored significantly lower on the financial knowledge test. however, participation in the pfp program largely eliminates the knowledge gap between women and men and between j. handy et al. / financial services review 29 (2021) 315–341 317 minorities and whites. additionally, women and minorities significantly improve their confidence in managing their financial future after completing the pfp. these findings suggest that this program (1) is a valid tool for improving financial knowledge that others should consider employing and (2) provides evidence that a follow-up study linking increased knowledge and confidence more directly to financial behaviors is worthwhile. as such, the pfp series is the first step in a long-term research program in which students will be surveyed after graduation to ascertain whether students participating in the pfp demonstrate better financial behaviors than students who do not. that research is currently underway. the remainder of this paper is organized as follows: section 1 overviews the pfp, including the recruitment process, seminar structures, and curriculum. section 2 reviews the literature regarding financial knowledge and literacy. section 3 details the survey instruments used in the study. section 4 discusses hypothesis development, introduces the empirical model and provides a preliminary analysis of data. section 5 empirically evaluates the effectiveness of the pfp in a multivariate framework. section 6 provides conclusions. 2. section 1: pfp overview and research method the pfp was developed and has been taught by a team of faculty for seniors at a private liberal arts college over the 2017-2019 period.2 the program consists of six individual sessions (the curriculum is described below) and is strictly voluntary, with students attending as many or as few sessions as they wish. programs are normally offered during the fall and spring semesters, one night per week for six weeks in a face-to-face format. senior-class student participants are solicited at the beginning of each semester via an email outlining the program content and voluntary time commitment. additionally, an email is sent to the parents of seniors detailing the program. only seniors are recruited for the program because they are about to face the challenges of real-world financial choices. anecdotally, program developers found that seniors begin to take financial issues more seriously as they realize that they will soon be in the “real world.” indeed, students generally report that they sign-up for the program because they believe learning the information is important (i.e., a question on the seminar participant presurvey asks if students are attending because “my parents ‘gently’ encouraged me,” “i heard about it from past participants,” or “i recognize that i need to know about these topics.” almost 75% report that they participate because they recognize that they need to know about the topics). students have approximately two weeks to register and several reminder emails are sent out as the deadline approaches. from the seniors not participating in the pfp, a control sample is recruited. before the first seminar session, both control and treatment groups are introduced to the logistics and purpose of the research portion of the program, and are asked to complete two on-line surveys (for students in the pfp, these occur before and after they take the course). the control group receives a survey approximately six weeks after the initial survey. in addition to demographic information, a series of 20 questions is used to assess financial knowledge levels. when participants indicate they want to participate, they complete the presurvey online. 318 j. handy et al. / financial services review 29 (2021) 315–341 each pfp session has a lecture format, but students are encouraged to ask questions. although there is some curriculum variance across program offerings each semester, it is minimal, as the same slides are used during each program offering to maintain curriculum consistency. the curriculum covers budgeting, credit management, risk and return, mutual funds, retirement planning, and risk management (insurance). these topics were chosen based on their relevance to this particular age group and their relationship to the four areas identified by huston (2010) as being critical to financial literacy. a summary of each module is provided in appendix. the underlying theme across each module is to develop the students’ knowledge necessary to maximize their “net worth” or their “wealth” over their lifetime. this message is stressed at the beginning and end of each session. additionally, faculty discuss how each topic relates to net worth. each session begins with a review of the previous week’s session and time to ask questions about prior material. sessions end with a summary of the current week’s information, along with more time to ask questions. shortly after the last session, students (participants and control students) are asked to voluntarily complete the postsurvey containing approximately 50 questions, including the same questions in the presurvey measuring financial knowledge. thus, the change in score on this knowledge survey determines the efficacy of the instructional program. participants and control students receive $10 in compensation for each survey they complete. 3. section 2: literature review the focus of the current study is to test whether students participating in the pfp improve what many prior studies refer to as financial literacy—what we refer to as financial knowledge. huston (2010) demonstrates that in prior research, terms such as financial knowledge, financial education, and financial literacy have been used interchangeably. beginning with the national endowment for financial education summit in 2005, a consensus began to develop that financial literacy has two distinct elements: a financial knowledge component and the measurement of actions/behavior based on that financial knowledge.3 in other words, to truly evaluate financial literacy, one must evaluate knowledge and behavior. furthermore, most academic literature measures financial behavior or knowledge independently from one another, or infers a relationship between financial knowledge and current financial behaviors (e.g., see allgood & walstad, 2013; gerardi, goette, & meir, 2013; henager & cude, 2016). as discussed previously, we distinguish between financial knowledge and financial literacy in this study, focusing on financial knowledge. huston and others note that research on financial knowledge inconsistently represents and measures knowledge. huston (2010) identifies four content areas that have been and should be addressed with financial education: money basics, borrowing, investing, and insurance. nonetheless, he reports that 35% of studies address only one content area, 40% address two or three areas, but only 25% address all four content areas. the pfp specifically addresses all four content areas leading to a more robust evaluation of financial knowledge. similarly, researchers (association for financial counseling and planning education, 2006; bosshardt & walstad, 2014; huston, 2010; knoll & houts, 2012) lament the fact that studies measuring financial knowledge often use widely different questions to measure it. huston (2010) finds that researchers use three to 45 questions, with most studies using three to five, largely j. handy et al. / financial services review 29 (2021) 315–341 319 centered on money basics and investing, to measure knowledge. knoll and houts (2012) address this issue and developed a knowledge test of 20 questions from past research that is psychometrically validated using item response theory and covers each content area identified in huston (2010).4 we use those same questions for our knowledge test. 4. section 3: survey construction in addition to the knowledge questions from knoll and houts (2012), the pre-seminar survey instrument captures demographic information. for the student, we capture gender, race, age, college major, whether the student is a student-athlete, as well as the sources/weights of how they pay for college: grants, loans, parent, or self. in addition to the student demographic data, we collect data about parents. for example, the students provide information about household income, parent educational attainment, and parent profession. the demographic information is not repeated on the post-seminar instrument. one of the unique benefits of our study design is that the survey poses questions that address a variety of psychological characteristics that may impact whether students choose to participate in the program and that may influence the student’s propensity to develop their financial knowledge. equally important, we ask a series of questions targeted at ascertaining the students’ confidence in their post-college financial future, confidence in each of the topics to be covered in the pfp, whether or not the student has experienced any health or financial stress during the period before the seminar, and what the student’s overall perception of how satisfied they are with their lives. the questions related to the psychological characteristics and confidence are repeated in the postsurvey instrument. our interest in student confidence is intentional. while some prior work indicates that individuals have misplaced confidence in their financial knowledge (e.g., toker asad, 2015), our focus is on whether students are confident about their ability to manage their financial future, not confidence in their responses to knowledge questions. we examine whether the pfp has an impact on this type of student confidence.5 to the extent that the statement “knowledge is power” is true, then the pfp may lead to an improvement in students’ perception that they can manage their financial future. our research design allows us a unique way to assess whether confidence is justified or overblown. if there is an increase in financial confidence without an improvement in financial knowledge, we would be reluctant to classify the program as a success. however, if we see an increase in confidence with a corresponding improvement in financial knowledge, we will be more emboldened to consider this stage of the research successful, warranting further analysis of behaviors. 5. section 4: hypothesis development, empirical model, and summary statistics 5.1. section 4.1: hypotheses the primary purpose of the pfp program is to improve college student financial literacy as measured through an evaluative survey and ultimately behaviors. thus, the foundational 320 j. handy et al. / financial services review 29 (2021) 315–341 hypothesis is that the program will improve financial knowledge for our participants (treatment) and that no similar increase would be found for the control group (future research will address changes in behaviors after the program.) specifically: hypothesis 1: after participating in the pfp, participants (treatment group) will score higher on the financial knowledge portion of the post-survey than they did on the pre-survey and non-participants (control group) will show no improvement from the preto post-test. prior research demonstrates that women have lower financial knowledge levels than men (e.g., lusardi & mitchell, 2007; lusardi & tufano, 2009; lusardi et al., 2010). a program that effectively improves financial literacy should do so for all. as such, hypothesis 2 states: hypothesis 2: female pfp participants will score no differently than male participants on the financial knowledge portion of the post-survey following completion of the pfp, that is, assuming women will score lower on the pre-survey (as has been found in previous research), with exposure to the seminar, female participants will close this knowledge gap with their male pfp peers. as with gender, prior research shows that minorities exhibit lower levels of knowledge than their peers (e.g., lusardi & mitchell, 2007, 2011). a successful program should help close or eliminate this gap if effective. as such, hypothesis 3 states: hypothesis 3: minority pfp participants will score no differently than white peers on the financial knowledge portion of the post-survey following completion of the pfp program, that is, assuming minority participants will score lower on the pre-survey (as has been found in previous research), with exposure to the seminar, minority participants will close the knowledge gap with theirwhite peers. pfp is available only to senior undergraduate students. the transition from college into the next phase of life, traditionally either graduate school or into a job environment, requires a significant change in one’s level of independent financial decision-making. those students who perceive themselves to be less prepared for such a transition are likely to face increased anxiety and lower confidence regarding their financial future, which could increase the likelihood of students failing to act on important financial decisions. if pfp effectively improves financial knowledge, then we would predict that confidence in the ability to manage one’s financial future may also increase. note that this confidence metric does not gauge a student’s confidence in how they answered the financial knowledge questions, but the student’s confidence in their ability to manage their financial future after college. the survey question used to gauge their confidence is “i consider myself to be _____________________ about managing my personal finances after college,” with five responses ranging from “not at all confident” to “extremely confident.” based on responses to this question from the pre and post pfp surveys, hypothesis 4 states: hypothesis 4: pfp participants (treatment) will express higher levels of confidence regarding their future financial outlooks in the post-survey relative to the pre-survey, and nonparticipants (control) will see no such change. j. handy et al. / financial services review 29 (2021) 315–341 321 5.2. section 4.2: empirical model (see table 2 for a summary of variable definitions) a significant strength of this research relative to prior work is the pretest, post-test design with a control group comparison. thus, the authors utilize a multivariate difference-in-difference regression to analyze the program’s effectiveness. the model is as follows: finlitvar ¼ aþ b 1postpfpdummyþ b 2treatment þ b 3 treatment � pstð þ þ b 4gender þ b 5 gender � postð þ þ b 6minority þ b 7 minority � postð þ þ ob jcharacteristicsjþ « i (1) finlitvar represents the dependent variable of interest, which is either knowscore, the number of correct financial knowledge questions answered on the student survey, or confidence, a student’s confidence level regarding their future financial outlook. post is a dummy variable that equals 1, reflecting survey results submitted after pfp completion and 0 otherwise. both the treatment group and the control group submit the survey results shortly after the completion of the pfp. treatment is a dummy variable equal to 1 if the survey results are submitted by a student who did participate in the pfp program and 0 otherwise. gender represents a student’s self-reported gender. though students were provided options of “transgender,” “other,” and “prefer not to answer,” all participants selected either male or female; thus, gender is a dummy variable equal to 1 if female and 0 otherwise. minority is a dummy variable equal to 1 if the student self-identified as belonging to a minority race and 0 otherwise. characteristics represents a series of control variables with the following description. gpa is a student’s grade point average based on a traditional 4.0 scale at the end of the students’ junior year. these are official gpa’s provided by the university, not self-reported. when knowscore is the dependent variable, we expect gpa to be positive. if higher grades promote self-confidence generally, then gpa may be positively correlated with a student’s confidence in their ability to manage their financial future. alternatively, stronger academic students may have a greater appreciation for the significance of these issues that manifests itself in lower confidence. loanperc is the student’s self-reported level of student loan debt to total college cost, and selfperc is the amount of education costs being covered by the student, but not in the form of student loans or paid by a parent, both stated as a percentage of total college costs. each of these measures accounts for financial factors that might be correlated with knowledge and confidence. we argue that students who finance more of their education with debt will have lower knowledge scores and lower levels of confidence in their financial future. we expect the opposite (a positive relationship) with the dependent variables for selfperc. income is a categorical variable ranging from one to six, representing ranges of family income, with a minimum category of less than $25,000 and a maximum category of more than $150,000. while prior studies (lusardi & tufano, 2009) show respondents with lower incomes have lower levels of financial knowledge, our inclusion of income is a proxy measure since it is not the student’s income being measured. however, if lower household incomes mean financial knowledge is not transferred, then the students from these 322 j. handy et al. / financial services review 29 (2021) 315–341 households may also exhibit lower scores. we expect students from higher earning families to have more confidence in their financial future. ecnbusacc is a dummy variable equal to 1 if the student is majoring in either economics, business administration, or accounting (the university does not offer the full range of business degrees) and 0 otherwise. we expect students from these majors to have higher levels of knowledge and confidence, in part due to their exposure to the seminar contents within their disciplines. the next control variable is unique. students were asked “how much do you think you will need to retire?” upon initial data examination, a nontrivial number of students chose not to answer the question.6 after significant thought, the authors hypothesized that the students that did not answer the question may be particularly at risk in terms of their financial knowledge and confidence. noretans is a dummy variable equal to 1 if a student did not answer the survey question about the amount needed for retirement and 0 otherwise. we expect an inverse relationship between noretans and both dependent variables, knowscore and confidence. finserv is a dummy variable equal to 1 if a student responded that either (or both) parent (s) work in the financial services industry. these parents may be more likely to pass on knowledge and develop confidence in their students relative to parents in other professions, so we expect a positive relationship to the dependent variables. finally, needcog stands for the psychological construct “need for cognition,” which captures individual differences in how much people engage in and enjoy thinking, measured by an 18-item scale (see cacioppo & petty, 1982). example items include, “i find satisfaction in deliberating hard and for long hours,” and “i only think as hard as i have to – reverse-scored.” each question uses a 5-point likert scale for answers. so, needcog can range from 0 to 90. unlike lusardi et al. (2010) that control for cognitive abilities, our measure captures a student’s propensity to be open to learning and deep thinking, factors pertinent to participating in an educational program. we expect students with a higher need for cognition to actively seek knowledge generally and that includes financial knowledge. as such, we expect a positive relationship between needcog and knowscore. the predicted relationship to confidence is less clear. if a student’s need for cognition leads them to seek out information about financial issues, that could lead to more confidence. conversely, students with a higher need for cognition may recognize the weighty nature of financial topics in a way that leaves them less confident. in addition to the variables just discussed, the following variables are used when confidence is the dependent variable. health is the sum of the responses (on a likert 5-point scale) to two questions regarding student health, one that asks if the student has experienced significant emotional or mental health issues in the past month and the other that asks whether the student has experienced significant financial distress in the past month. students that have experienced high levels of stress may also report less confidence in their ability to manage their financial future. lifesat is the satisfaction with life scale (diener, emmons, larsen, & griffin, 1985) that measures subjective global well-being and is a composite measure of five questions, each of which has responses on a 7-point likert scale. students with low satisfaction levels are expected to have lower levels of confidence in their financial future. for our primary analysis, we estimate eq. (1) using ordinary least squares with a difference-in-difference model to help mitigate the inherent endogeniety based on a student’s choice to participate in the pfp. roberts and whited (2013) recommend this approach when j. handy et al. / financial services review 29 (2021) 315–341 323 the sample has both a treatment and control group, but also when there is a pretreatment and post-treatment measurement of the treatment effect (participation in the pfp), which exactly describes our research design. that is, both treatment and control participants have pre-pfp knowledge scores (confidence scores) and post-pfp knowledge (confidence scores), thereby doubling the number of observations from 161 (158) to 322 (316). with this approach that is able to control for baseline measures of knowledge and confidence, any findings associated with the treatment effect can be unbiasedly attributed to the treatment. while we use the least squares methodology as our primary statistical technique, we also estimate eq. (1) for the dependent variable confidence using the ordered probit technique. because confidence is measured as a value from 1 to 5, the assumption of linear change between values implied by the least squares technique may not hold. to account for the possibility of this nonlinear movement between values, the ordered probit technique is employed as a robustness check. as noted below, the results demonstrate that the variables of interest have the same statistical and directional effects using both techniques; however, we choose to discuss our results in the context of the least squares technique due to its more straight-forward interpretation. 5.3. section 4.3: preliminary data analysis 5.3.1. section 4.3.1. data analysis sample construction the first delivery of the pfp occurred in fall 2017 and was delivered successively in the spring 2018, fall 2018, spring 2019, and fall 2019 semesters. while over 500 students (program participants and control respondents) initiated a pre-seminar survey over the course of the five semesters, only 161 (158) completed both surveys with all of the required data needed to estimate eq. (1) above for knowscore (confidence). in total, for the semesters identified above, there are 10, 2, 17 (15), 12, and 50 (49) control respondents respectively for knowscore (confidence) and 18, 11, 15, 5, and 21 pfp participants, respectively, from the same semesters. as is evident from the usable responses in each semester, statistical analysis required the aggregation of respondents across semesters. however, we do control for any semester differences that might be present by including a dummy variable representing each semester except fall 2017. in no case are the semester dummies statistically significant. we use these samples in the analysis discussed in tables 2-6. the exception to this is the results in table 1, which analyzes the characteristics that predict pfp participation, a subject to which we now turn. 5.3.2. section 4.3.2.: predicting pfp participation while our primary interest is in the impact of the pfp on financial knowledge and confidence, a unique characteristic of this program is that students self-select into the program, and it is completely voluntary. as such, we first examine a logit model to determine if there are student characteristics that predict participation. the dependent variable has a value of 1 if the student participates in the pfp and 0 otherwise. the variables used are discussed above. for this analysis, we use all observations (428 total, 222 control respondents, and 206 seminar participants) that have the necessary data to estimate the logit model from pre-seminar survey data. we believe the results may be specific to the population of students served; therefore, we draw no generalized conclusions from the analysis. however, the results are of 324 j. handy et al. / financial services review 29 (2021) 315–341 table 1 logit model predicting pfp participation variables (1) treatment knowscore �0.0649 (0.0550) gpa �0.157 (0.288) confidence �0.697*** (0.139) gender 0.620** (0.304) ecnbusacc 0.687 (0.369) loanperc �0.287 (0.823) selfperc 1.017 (2.141) minority �0.803 (0.447) income �0.141 (0.0965) needcog 0.0143 (0.0115) finserv 0.351 (0.319) noretans �1.059*** (0.396) recent emotional/mental distress �0.244** (0.1000) recent financial distress �0.0988 (0.124) lifesat 0.0603** (0.0236) observations 428 this table presents results from estimating a logit regression to predict student participation in personal finance program (pfp). the model takes the form: logit(p(y = 1jx1,. . .xk) = [exp(b 0 + b 1x1 + . . . + b nxn)]/[1 + exp(b 0 + b 1x1 + . . . + b nxn)] the dependent variable (y) is treatment equal to 1 if the student participates in pfp and 0 otherwise. knowscore is a student’s score on the 20 question knowledge test before the seminar. gpa is the student’s gpa at the end of the junior year, on a 4-point scale. confidence is how the student perceives their confidence in their financial future after college and ranges from 1 (not at all) to 5 (extremely). gender is a dummy variable equal to 1 if the student is a female and 0 otherwise. ecnbusacc is a dummy equal to 1 if the student is an economics, business, or accounting major and 0 otherwise. loanperc and selfperc are the percentage of student college cost paid for with debt or by the student.minority is a dummy equal to 1 if the student identified as a minority and 0 otherwise. income is a variable ranging from 1 (low) to 6 (high) capturing the student’s household income. needcog is a composite measure ranging from 0-90 that captures a student’s propensity to engage cognitive activities. finserv is a dummy equal to 1 if one or both of a student’s parents work in the financial services industry and 0 otherwise. noretans is a dummy equal to 1 if the student did not answer the question “how much do you think you will need to retire?” and 0 otherwise. recent emotional/mental distress identifies whether the student has experienced significant emotional or mental distress in the previous 30 days and an estimate of the severity, ranging from 1 (low) to 5 (high). recent financial distress identifies whether the student has experienced significant financial distress in the previous 30 days and an estimate of the severity, ranging from 1 (low) to 5 (high). lifesat is a composite measure ranging from 5-25 that expresses a student’s satisfaction with their life at the time of the pre-survey. note: standard errors in parentheses. ***p < 0.01, **p < 0.05. j. handy et al. / financial services review 29 (2021) 315–341 325 interest because they may help in recruiting efforts for underrepresented or vulnerable groups in the future. the results are presented in table 1. neither a student’s pre-seminar financial knowledge score or gpa are correlated with students participating in the pfp. thus, students who have more financial knowledge, or who generally perform well academically, are no more inclined to participate in the program than those with less knowledge or academic proclivity. however, students who are more confident in their ability to manage their financial future are significantly less likely to participate in the pfp. the coefficient for confidence is negative and statistically significant at better than 1%. one interpretation is students with low levels of confidence in making future financial decisions make a wise choice to attend the pfp. alternatively, those that choose not to participate in the seminar may have a false sense of confidence (e.g., toker asad, 2015). our measure does not allow us to distinguish between these two alternatives. of course, attending the pfp and benefitting from it are not the same, but at least those students who table 2 summary descriptive statistics variables (1) matched sample (2) treatment only (3) control only pre knowledgescore 14.23 (2.473) 14.19 (2.342) 14.26 (2.581) post knowledgescore 14.832*** (2.916) 15.343*** (2.513) 14.44 (3.149) pre confidence in future 2.684 (1.004) 2.357 (0.885) 2,943 (1.021) post confidence in future 3.184*** (0.843) 3.357*** (0.638) 3.045 (0.958) gpa 3.439 (0.427) 3.462 (0.435) 3.422 (0.423) gender 0.696 (0.462) 0.757 (0.432) 0.648 (0.480) loanperc 0.073 (0.149) 0.053 (0.117) 0.088 (0.168) selfperc 0.022 (0.076)) 0.02 (0.072) 0.024 (0.080) minority 0.0683 (0.253) 0.0714 (0.259) 0.0659 (0.250) income 4.783 (1.298) 4.786 (1.250) 4.780 (1.340) ecnbusacc 0.236 (0.426) 0.243 (0.432) 0.231 (0.424) noretans 0.0683 (0.253) 0.0714 (0.259) 0.0659 (0.250) finserv 0.168 (0.375) 0.157 (0.367) 0.176 (0.383) needcog 62.35 (11.02) 61.76 (11.26) 62.81 (10.88) proportion of sample control 0.565 (0.497) observations 161 70 91 this table presents summary descriptive statistics for variables in the knowscore and confidence analysis. knowscore is a student’s score on the 20 question knowledge test. confidence is the reported measure of a student’s confidence in their ability to manage their financial future ranging from 1-5. pre and post identify either the pre-seminar or post-seminar measurement. all of the control variables listed below are captured in the presurvey and are means for all pfp participants and control respondents that completed both surveys (column 1), the treatment group separately that completed both surveys (column 2), and the control group separately that completed both surveys (column 3). gpa is the student’s gpa at the end of the junior year, on a 4-point scale. gender is a dummy variable equal to 1 if the student is a female and 0 otherwise. loanperc and selfperc are the percentage of student college cost paid for with debt or by the student. minority is a dummy equal to 1 if the student identified as a minority and 0 otherwise. income is a variable ranging from 1 (low) to 6 (high) capturing the student’s household income. ecnbusacc is a dummy equal to 1 if the student is an economics, business, or accounting major and 0 otherwise. noretans is a dummy equal to 1 if the student did not answer the question “how much do you think you will need to retire?” and 0 otherwise. finserv is a dummy equal to 1 if one or both of a student’s parents work in the financial services industry and 0 otherwise. needcog is a composite measure ranging from 0-90 that captures a student’s propensity to engage cognitive activities. proportion of sample control is the percentage of non-participants in the matched sample. ***p < 0.01, **p < 0.05. 326 j. handy et al. / financial services review 29 (2021) 315–341 are less confident take actions to address their concerns. women are significantly (at the 5% level) more likely to participate in the program. this finding is encouraging given that women have historically shown lower levels of financial knowledge than men, a result we confirm with pre-pfp data. table 3 financial knowledgescore and pfp participation variables (1) (2) (3) (4) post pfp dummy 0.572** (0.250) 1.175*** (0.346) �0.0401 (0.456) 0.700*** (0.252) gpa 1.093*** (0.360) 1.092*** (0.358) 1.093*** (0.359) 1.097*** (0.356) gender �0.967*** (0.287) �0.968*** (0.288) �1.407*** (0.364) �0.964*** (0.287) loanperc �0.222 (0.871) �0.223 (0.873) �0.222 (0.866) �0.217 (0.857) selfperc �0.0371 (2.012) �0.0543 (1.951) �0.0381 (1.999) 0.0280 (2.012) minority �2.519*** (0.786) �2.521*** (0.766) �2.519*** (0.783) �1.512** (0.684) income 0.172 (0.105) 0.172 (0.105) 0.172 (0.105) 0.172 (0.105) ecnbusacc 2.083*** (0.279) 2.084*** (0.277) 2.083*** (0.279) 2.083*** (0.278) noretans �0.990 (0.717) �0.956 (0.707) �0.988 (0.710) �1.12 (0.665) finserv 0.127 (0.346) 0.126 (0.344) 0.127 (0.342) 0.131 (0.349) needcog 0.0456*** (0.0116) 0.0456*** (0.0115) 0.0456*** (0.0114) 0.0456*** (0.0116) treatment 0.554 (0.284) 0.0233 (0.337) 0.554 (0.285) 0.55 (0.282) treat*post 1.062** (0.494) gender*post 0.880 (0.543) minority*post �1.995 (1.279) observations 322 322 322 322 r2 0.361 0.370 0.367 0.369 this table presents results from estimating the following equation finlitvar ¼ a þ b 1post pfp dummy þ b 2treatment þ b 3 treatment� postð þ þ b 4gender þb 5 gender�postð þ þ b 6minority þ b 7 minority�postð þ þ ob j characteristicsj þ « i: where knowscore is the dependent variable. knowscore is a student’s score on the 20 question knowledge test. for each respondent in our sample there are two observations, one representing the pre-seminar survey responses and the other the post-seminar responses. post pfp dummy equals 1 if the observation represents the post personal finance program (pfp) knowledge score observation for a student and 0 otherwise. gpa is the student’s gpa at the end of the junior year, on a 4-point scale. gender is a dummy variable equal to 1 if the student is a female and 0 otherwise. loanperc and selfperc are the percentage of student college cost paid for with debt or by the student. minority is a dummy equal to 1 if the student identified as a minority and 0 otherwise. income is a variable ranging from 1 (low) to 6 (high) capturing the student’s household income. ecnbusacc is a dummy equal to 1 if the student is an economics, business, or accounting major and 0 otherwise. noretans is a dummy equal to 1 if the student did not answer the question “how much do you think you will need to retire?” and 0 otherwise. finserv is a dummy equal to 1 if one or both of a student’s parents work in the financial services industry and 0 otherwise. needcog is a composite measure ranging from 0-90 that captures a student’s propensity to engage cognitive activities. treatment is equal to 1 if the student participates in pfp and 0 otherwise. treat*post, gender*post, and minority*post are interaction terms between treatment, gender, minority, and post pfp dummy. note: robust standard errors in parentheses. ***p < 0.01, **p < 0.05. j. handy et al. / financial services review 29 (2021) 315–341 327 student major has no impact on the choice to participate in the pfp. neither the proportion of debt or self-funding by the student predicts participation. there is no evidence that minorities are more or less likely to take the pfp. household income (income), need for cognition (needcog), and having one or both parents that work in the financial services industry (finserv) are not correlated with student attendance. students who did not answer the question about how much they thought they would need to retire (noretans) are significantly (at the 1% level) less likely to participate in the program. if, as the authors suspect, the students who meet these criteria are less knowledgeable, yet are more confident in their ability to manage their financial future, then it presents a challenge to identify these students and encourage participation. while experiencing a recent table 4 financial knowledgescore and pfp participation variables (1) (2) (3) post pfp dummy 0.113 (0.352) 0.510** (0.257) 0.117 (0.351) gpa 1.091*** (0.359) 1.133*** (0.358) 1.126*** (0.357) gender �1.088*** (0.337) �0.973*** (0.286) �1.118*** (0.335) loanperc �0.166 (0.865) �0.153 (0.870) �0.0933 (0.862) selfperc �0.184 (1.962) �0.107 (2.045) �0.271 (1.998) minority �2.509*** (0.771) �3.008*** (0.902) �2.939*** (0.897) income 0.169 (0.105) 0.172 (0.104) 0.168 (0.104) ecnbusacc 2.078*** (0.277) 2.095*** (0.279) 2.088*** (0.277) noretans �0.935 (0.705) �0.926 (0.743) �0.879 (0.731) finserv 0.135 (0.342) 0.0928 (0.347) 0.107 (0.342) needcog 0.0461*** (0.0115) 0.0445*** (0.0116) 0.0453*** (0.0115) treatment 0.0370 (0.339) 0.479 (0.278) 0.0406 (0.342) fem*treat*post 1.196** (0.530) 1.090** (0.532) male*treat*post 0.648 (0.629) 0.427 (0.630) minority*treat*post 2.137** (1.007) 1.889 (0.997) observations 322 322 322 r2 0.372 0.368 0.377 this table presents results from estimating eq. (1) where knowscore is the dependent variable. knowscore is a student’s score on the 20 question knowledge test. for each respondent in our sample there are two observations, one representing the pre-seminar survey responses and the other the post-seminar responses. post pfp dummy equals 1 if the observation represents the post personal finance program (pfp) knowledge score observation for a student and 0 otherwise. gpa is the student’s gpa at the end of the junior year, on a 4-point scale. gender is a dummy variable equal to 1 if the student is a female and 0 otherwise. loanperc and selfperc are the percentage of student college cost paid for with debt or by the student. minority is a dummy equal to 1 if the student identified as a minority and 0 otherwise. income is a variable ranging from 1 (low) to 6 (high) capturing the student’s household income. ecnbusacc is a dummy equal to 1 if the student is an economics, business, or accounting major and 0 otherwise. noretans is a dummy equal to 1 if the student did not answer the question “how much do you think you will need to retire?” and 0 otherwise. finserv is a dummy equal to 1 if one or both of a student’s parents work in the financial services industry and 0 otherwise. needcog is a composite measure ranging from 0-90 that captures a student’s propensity to engage cognitive activities. treatment is equal to 1 if the student participates in pfp and 0 otherwise. fem*treat*post, male*treat*post, and minority*treat*post are equal to 1 if the observation represents a female (male, minority), is from the treatment group, and represents the post-seminar survey. note: robust standard errors in parentheses. ***p < 0.01, **p < 0.05. 328 j. handy et al. / financial services review 29 (2021) 315–341 financially distressing event is not predictive of participation, students that have experienced a more significant emotional or mental challenge are less likely to participate in the course (at the 5% level). finally, students who are more satisfied with their lives at the time of the seminar are more likely to take the pfp (at the 5% level). these findings support the idea table 5 financial confidence and pfp participation variables (1) (2) (3) (4) post pfp dummy 0.582*** (0.132) 1.267*** (0.184) 0.289 (0.235) 0.535*** (0.133) gpa �0.417*** (0.160) �0.415*** (0.160) �0.412*** (0.160) �0.427*** (0.161) gender �0.569*** (0.171) �0.608*** (0.173) �0.785*** (0.223) �0.568*** (0.171) loanperc �0.723 (0.369) �0.749** (0.377) �0.724** (0.368) �0.723** (0.369) selfperc 0.997 (0.833) 1.009 (0.794) 0.994 (0.801) 0.972 (0.842) minority 0.653 (0.372) 0.634 (0.376) 0.641 (0.375) 0.327 (0.516) knowscore 0.0933*** (0.0275) 0.0819*** (0.0283) 0.0894*** (0.0280) 0.0980*** (0.0283) income 0.0385 (0.0495) 0.0426 (0.0497) 0.0385 (0.0492) 0.0366 (0.0497) ecnbusacc 0.474*** (0.163) 0.526*** (0.167) 0.485*** (0.164) 0.465*** (0.164) noretans �0.257 (0.432) �0.244 (0.448) �0.264 (0.437) �0.203 (0.451) finserv 0.218 (0.183) 0.228 (0.184) 0.220 (0.183) 0.216 (0.182) needcog 0.0200*** (0.00621) 0.0215*** (0.00620) 0.0204*** (0.00619) 0.0199*** (0.00621) health �0.0645 (0.0430) �0.0627 (0.0434) �0.0682 (0.0426) �0.0687 (0.0429) lifesat 0.0148 (0.0155) 0.0166 (0.0154) 0.0140 (0.0155) 0.0152 (0.0154) treatment �0.241 (0.138) �0.821*** (0.197) �0.239 (0.139) �0.244 (0.137) treat*post 1.168*** (0.251) gender*post 0.427 (0.274) minority*post 0.685 (0.634) observations 316 316 316 316 this table presents results from estimating eq. (1) where confidence is the dependent variable. confidence is how the student perceives their confidence in their financial future after college and ranges from 1 (not at all) to 5 (extremely). for each respondent in our sample there are two observations, one representing the pre-seminar survey response and the other the post-seminar response. gpa is the student’s gpa at the end of the junior year, on a 4-point scale. post pfp dummy equals 1 if the observation represents the post personal finance program (pfp) confidence observation for a student and 0 otherwise. treatment is equal to 1 if the student participates in pfp and 0 otherwise. knowscore is a student’s score on the 20 question knowledge test (pre or post). gender is a dummy variable equal to 1 if the student is a female and 0 otherwise. minority is a dummy equal to 1 if the student identified as a minority and 0 otherwise. income is a variable ranging from 1 (low) to 6 (high) capturing the student’s household income. ecnbusacc is a dummy equal to 1 if the student is an economics, business, or accounting major and 0 otherwise. needcog is a composite measure ranging from 0-90 that captures a student’s propensity to engage cognitive activities. loanperc and selfperc are the percentage of student college cost paid for with debt or by the student. finserv is a dummy equal to 1 if one or both of a student’s parents work in the financial services industry and 0 otherwise. health is the sum of the values to the questions of whether the student has experienced significant emotional or mental distress in the previous 30 days and an estimate of the severity and whether the student has experienced significant financial distress in the previous 30 days and an estimate of the severity, with the value ranging from 2 (low) to 10 (high). lifesat is a composite measure ranging from 5-25 that expresses a student’s satisfaction with their life at the time of the pre-survey. noretans is a dummy equal to 1 if the student did not answer the question “how much do you think you will need to retire?” and 0 otherwise. treat*post, gender*post, and minority*post are interaction terms between treatment, gender, minority, and post pfp dummy. note: robust standard errors in parentheses. ***p < 0.01, **p < 0.05. j. handy et al. / financial services review 29 (2021) 315–341 329 that one’s state of mind, including perceived stress and overall well-being predict self-efficacy, or the belief that one can set a goal and enact the necessary behavior to achieve that goal (bandura, 1977). enrolling in and completing the pfp required students to respond to the recruiting email, attend sessions, and complete a preand post-seminar survey; these table 6 financial confidence and pfp participation variables (1) (2) (3) post pfp dummy 0.0989 (0.179) 0.549*** (0.131) 0.101 (0.180) gpa �0.416*** (0.160) �0.388** (0.161) �0.394** (0.160) gender �0.590*** (0.204) �0.585*** (0.171) �0.613*** (0.203) loanperc �0.759** (0.378) �0.679 (0.369) �0.717 (0.376) selfperc 1.030 (0.800) 0.955 (0.848) 0.983 (0.810) minority 0.633 (0.376) 0.332 (0.412) 0.396 (0.428) knowscore 0.0822*** (0.0283) 0.0859*** (0.0280) 0.0771*** (0.0287) income 0.0431 (0.0498) 0.0391 (0.0496) 0.0430 (0.0497) ecnbusacc 0.526*** (0.167) 0.503*** (0.166) 0.545*** (0.169) noretans �0.248 (0.447) �0.219 (0.445) �0.218 (0.457) finserv 0.227 (0.184) 0.197 (0.182) 0.212 (0.183) needcog 0.0214*** (0.00622) 0.0200*** (0.00626) 0.0213*** (0.00625) health �0.0625 (0.0435) �0.0733 (0.0433) �0.0693 (0.0437) lifesat 0.0165 (0.0154) 0.0145 (0.0154) 0.0163 (0.0154) treatment �0.823*** (0.197) �0.291** (0.139) �0.826*** (0.197) fem*treat*post 1.148*** (0.263) 1.097*** (0.266) male*treat*post 1.231*** (0.344) 1.121*** (0.326) minority*treat*post 1.464** (0.610) 1.101 (0.618) observations 316 316 316 this table presents results from estimating eq. (1) where confidence is the dependent variable. confidence is how the student perceives their confidence in their financial future after college and ranges from 1 (not at all) to 5 (extremely). for each respondent in our sample there are two observations, one representing the pre-seminar survey response and the other the post-seminar response. gpa is the student’s gpa at the end of the junior year, on a 4-point scale. post pfp dummy equals 1 if the observation represents the post personal finance program (pfp) confidence observation for a student and 0 otherwise. treatment is equal to 1 if the student participates in pfp and 0 otherwise. knowscore is a student’s score on the 20 question knowledge test (pre or post). gender is a dummy variable equal to 1 if the student is a female and 0 otherwise. minority is a dummy equal to 1 if the student identified as a minority and 0 otherwise. income is a variable ranging from 1 (low) to 6 (high) capturing the student’s household income. ecnbusacc is a dummy equal to 1 if the student is an economics, business, or accounting major and 0 otherwise. needcog is a composite measure ranging from 0-90 that captures a student’s propensity to engage cognitive activities. loanperc and selfperc are the percentage of student college cost paid for with debt or by the student. finserv is a dummy equal to 1 if one or both of a student’s parents work in the financial services industry and 0 otherwise. health is the sum of the values to the questions of whether the student has experienced significant emotional or mental distress in the previous 30 days and an estimate of the severity and whether the student has experienced significant financial distress in the previous 30 days and an estimate of the severity, with the value ranging from 2 (low) to 10 (high). lifesat is a composite measure ranging from 5-25 that expresses a student’s satisfaction with their life at the time of the pre-survey. noretans is a dummy equal to 1 if the student did not answer the question “how much do you think you will need to retire?” and 0 otherwise. fem*treat*post, male*treat*post, and minority*treat*post are equal to 1 if the observation represents a female (male, minority), is from the treatment group, and represents the post-seminar survey. note: robust standard errors in parentheses. ***p < 0.01, **p < 0.05. 330 j. handy et al. / financial services review 29 (2021) 315–341 steps require a level of persistence and engagement encouraged by feeling positive about one’s life and not feeling overwhelmed by stressful life events. to summarize, students who participate in the pfp are less confident in their financial future, experience less stress and have a higher level of life satisfaction. most importantly, the participants and the control group did not differ in pretest knowledge level, or academic ability as measured by gpa, as well as their interest in thinking (need for cognition), or whether their parents work in the finance industry. 5.3.3. section 4.3.3.: summary descriptive statistics table 2 presents summary descriptive statistics of financial knowledge and control variables for the sample. the sample is broken down into three groups. column (1) reflects participants and control respondents who fully completed both the pre and post surveys. columns (2) and (3) split the survey sample into those participating (treatment) in the pfp and those not participating (control) in the pfp, respectively, and that completed both surveys. a substantial number of students (over 400) over the course of the three years of the study completed only the presurvey as was described above. 5.4. participant demographics (only assessed in the pretest) results in table 2 show that seminar participants are (1) reasonably high academic achievers, having an average gpa of 3.4, (2) mostly women, and have (3) minimal student loans and minimal self-funding. relatively few minorities participated in the study, although the authors suspect that is due, in part, to university demographics. the average study participant is from a relatively wealthy household; income averages approximately 4.7, which is a categorical ranking corresponding with an average family income between $100,000 and $150,000. just under 25% of study participants were majoring in economics, business, or accounting. approximately 7% of presurvey responders failed to answer the retirement question and approximately 16% of study participants have a family member in the financial services industry. regarding need for cognition, on a 90-point scale with 90 suggesting a substantially high need for cognition, study participants show a slightly higher than average need for cognition at around 62. 5.5. knowledge and confidence scores pre-pfp completion survey results of financial knowledge scores show that all respondents correctly answer 14.23 questions out of 20 (see column 1 in table 2). however, the control group average is 14.26, while that for the treatment group is 14.19 (columns 2 and 3 in table 2), which is not statistically different. however, after pfp completion, knowledge scores improve significantly for the pfp participants, a change from 14.19 to 15.34—column (2), compared with no statistical change from 14.26 to 14.44 (column 3) for the control group. additionally, financial confidence metrics also improve in a statistically significant j. handy et al. / financial services review 29 (2021) 315–341 331 way for pfp participants (from 2.357 to 3.357—based on a 5-point scale), while there is no change for the control group, where metrics remain largely unchanged (2.943 to 3.045). equally important, whereas the control (2.943) group’s pre-pfp confidence is statistically higher than the treatment (2.348) group’s, the treatment (3.362) group’s post-pfp confidence is statistically higher than the control (3.045) group. in summary, the univariate statistics support hypotheses 1 and 4 (hypothesis 2 and 3 are not tested here). pfp participants improve both their financial knowledge and their confidence in their financial future while the control group does not. the improvement in confidence is particularly compelling. before going through the pfp, participants have a lower level of confidence than the control group. however, after completing the program, not only does the treatment group improve their confidence, but it exceeds that of the control group. combined, the univariate results for knowscore and confidence are promising, and the improvement in confidence seems to be well placed given the improvement in knowledge scores and the lower level of pretest confidence reported than the control group who chose not to participate in the program. 6. section 5: multivariate results 6.1. section 5.1.: knowscore the overall interpretation of the univariate results suggests that pfp participation improves both financial knowledge and confidence in the ability to manage one’s financial future; the finding that the control group shows no such improvement strengthens the conclusion that these improvements are a product of the program and not due to time or a practice effect (e.g., taking the knowledge test twice). we now more robustly scrutinize our findings using a multivariate framework using eq. (1) from above. column (1) of table 3 presents the baseline model results for our matched sample. in table 2, there are 161 students (70 treatment and 91 control) who fully responded to both surveys. the sample size here is 322, representing an observation for the preand post-survey responses. before addressing the hypotheses directly, we briefly discuss the control variables. neither funding their education with more debt nor more self-funding influences a student’s preor postknowledge score. additionally, students not answering the question regarding the amount needed for retirement, or having parents who work in the financial services industry has any impact on preor post-student knowledge scores. however, students with higher gpa’s have higher preand postknowledge scores, as do students who are economics, business, or accounting majors, and students with a higher need for cognition. the results for the control variables clearly demonstrate that the factors influencing a student’s financial knowledge are multifaceted and not solely linked to typical demographic information; thus, it is important to control for their influence on our results. the primary variables of interest initially to test our hypotheses are post pfp dummy, treatment, gender, and minority. the results show that post-seminar (post pfp dummy) knowledge scores are statistically higher (at the 5% level) than pre-pfp scores. the 332 j. handy et al. / financial services review 29 (2021) 315–341 coefficient estimate for treatment is positive but not significant. in combination, these results seem to demonstrate that those participating in the pfp improve knowledge more than the control group do not. however, the results for the post pfp dummy and treatment variables, while in the direction desired, do not completely provide support for hypothesis 1 because members of the control group are included in the post pfp sample, and the treatment variable includes both preand post-results for the treatment group. column (1) also shows that women have statistically lower scores than men at better than the 1% level, and minorities have statistically lower scores than their caucasian peers. while the results for our gender and minority variables are consistent with prior research, as presented, they do not adequately address hypotheses 2 and 3. currently, the gender result tells us only that women have lower scores than do men, but does not distinguish between treatment and control, nor between preand post-scores. the same is true for the minority result. column (2) more directly addresses hypothesis 1 by introducing the interaction term treat*post, which clearly identifies observations from the post seminar treatment group. the coefficient estimate on the interaction term is positive and statistically significant at the 5% level, indicating the treatment group improves its scores relative to the control group. the result provides strong support for hypothesis 1. we now turn our attention to addressing hypotheses 2 and 3 more directly. in columns (1) and (2), gender and minority are negative and significant at the 1% level; however, the models do not allow us to distinguish between the pre and post-pfp scores. in column (3), we add the additional interaction term gender*post. the results are similar to those in column (2), but gender*post is not statistically different from zero, which would not seem to support hypothesis 2. in column (4), the interaction term minority*post is not different from zero, suggesting a lack of support for hypothesis 3. one problem with the gender (minority) interaction is that while it segregates males from females (minorities from caucasians) and preand post-scores, it includes both treatment and control females (minorities). to more directly test hypothesis 2 and 3, we introduce three terms, fem*treat*post,male*treat*post, andminority*treat*post that are equal to 1 if the observation represents a female (male, minority), in the treatment group, and corresponds to the post-pfp knowledge score and 0 otherwise. the results are in columns (1 and 2) of table 4. while all other results remain consistent, we find that females and minorities in the treatment group significantly narrow the knowledge gap, as fem*treat*post and minority*treat*post are positive and statistically significant at the 5% level. for women, when combined with the insignificant coefficient estimates on treatment and male*treat*post, we can conclude that females close the gap with males if they participated in the seminar, a finding providing support for hypothesis 2. the results in column (2) are similarly encouraging. minority*treat*post is positive and statistically significant, indicating that minorities close the knowledge gap after participating in the pfp. this result is consistent with hypothesis 3. the results in column (3) show that the results are robust for women when including all of the terms in the model. in contrast, minority*treat*post is no longer statistically significant at traditional levels. in the full model presented in column (3), hypothesis (2) is definitively supported. j. handy et al. / financial services review 29 (2021) 315–341 333 6.2. section 5.2: confidence we now turn our attention to examining the impact of the pfp on a student’s confidence in their ability to manage their financial future. it is important to note again that this is not confidence in whether one answered the financial knowledge questions correctly, but rather one’s confidence regarding their future financial outlook. earlier, we reported that in a univariate framework, pfp participants had significantly lower confidence than their control group peers before the seminar but that this completely reversed in the post-seminar surveys. we now determine the robustness of those results using eq. (1). we augment eq. (1) with knowscore, health, and lifesat as additional control variables with an expectation that students with higher knowledge scores who are more satisfied with their lives currently will be more confident, but students who have recently experienced a higher degree of emotional/ mental and financial distress will have lower levels of confidence. the results for estimating eq. (1) with confidence as the dependent variable are presented in tables 5 and 6. we follow a similar pattern of presentation used for knowscore. while not of primary interest, a brief discussion of control variables is valuable as the analysis of confidence in a multivariate framework is new to the literature. regardless of the model, students who are economics, business, or accounting majors are statistically more confident in their financial future, an expected outcome. additionally, students with higher knowledge scores and a higher need for cognition have statistically higher levels of confidence in their financial future. the result for knowledge score is “comforting” in that it suggests that the confidence may not be misplaced. the result for needcog suggests that students who seek out cognitive activities feel they can acquire the skills and information necessary to be successful in their personal financial lives. students with more of their education financed with debt generally have lower levels of confidence in their future. this finding is expected but indicates that in addition to the financial burden that student debt brings, it has a significant impact on how students perceive they can handle their financial future and likely stress levels. finally, students with higher gpa’s have statistically lower levels of confidence, an unexpected finding. we now turn to testing hypothesis 4. in column (1), we present a base model that indicates post-seminar (post pfp dummy) confidence levels are statistically higher, but that pfp participants (treatment) do not have statistically different levels of confidence when compared with the control group. these findings do not support hypothesis 4; however, as was the case in column (1) of table 3, the post pfp dummy captures both control and treatment respondents, and the treatment variable captures both preand post-survey results. to better gauge the effect of the pfp on confidence, we introduce the interaction term treat*post in column (2) to more precisely isolate the post-treatment group. the coefficient estimate for treat*post is positive and statistically significant at the 1% level, while treatment is negative and significant, confirming that students in the treatment group have significantly lower levels of confidence before the seminar. this confirms the univariate results reported earlier and supports hypothesis 4. so, seminar participants not only overcome their lack of confidence in their financial future relative to the control group, but their participation leads to significantly higher levels of confidence than the control group after seminar participation. 334 j. handy et al. / financial services review 29 (2021) 315–341 the coefficient estimates for gender in columns (1, 2) indicate that women report significantly lower levels of confidence (pre and post) in their future, while minorities show no statistical difference from caucasians. given the starting point with respect to financial knowledge for women, the finding is expected. in columns (3) and (4), we further examine the relationship between gender (minority) and confidence. in column (3), we add the interaction term gender*post. the results suggest that women have no difference in post-program confidence levels relative to the preprogram survey. however, the coefficient estimate for gender is still negative and significant, indicating women’s confidence is still lower than males. however, as was the case in table 3, gender*post groups all women together. as before, in column (4) we add the interaction term minority*post. the results suggest that there is no significant change in confidence levels in the post-program surveys. similar to the analysis in table 4, we introduce fem*treat*post, male*treat*post, and minority*treat*post to isolate the gender, minority, and treatment effects. results are presented in table 6. in column (1), we see that both women and men in the treatment (male*treat*post and female*treat*post) group have significantly higher levels of confidence after participating in the program, further support for hypothesis 4. the results in column (1) are important because the coefficient estimate for treatment is negative and significant, consistent with the treatment group being less confident than the control group on the presurvey, but the interaction terms, male*treat*post and female*treat*post, indicate that while both males and females in the program improve, they do so at levels that are now higher than the control group. we find complementary results in column (2) of table 6 for minorities in the treatment group; their confidence is significantly higher than the control group, additional support for hypothesis 4. the results in column (3) show that the results are generally robust when including all of the interaction terms together, although minority*treat*post is no longer significant at conventional levels. 6.3. section 5.3: robustness tests in total, our multivariate analysis supports all four of our hypotheses. however, one might argue that the measurement of confidence as a discrete variable ranging from 0 to 5 makes using ordinary least squares inappropriate because the least squares technique assumes a linear change from one value of confidence to another, which may not be true. as such, we also estimate eq. (1) for confidence using an ordered probit model. the results in terms of statistical significance and directional effect of variables on confidence are unchanged from those presented in tables 5-6. in total, our analysis is robust to a multivariate framework using alternative estimation techniques. 6.4. section 5.4: reconciling results for knowledge and confidence7 the results above demonstrate that pfp students see their knowledge scores and their confidence about their ability to manage their financial future rise significantly after seminar participation. the result for confidence is particularly striking; pfp students go from having significantly lower confidence than the control group to having significantly higher j. handy et al. / financial services review 29 (2021) 315–341 335 confidence than the control group after seminar participation. of course, increasing confidence may not be well placed if it is not calibrated to one’s financial knowledge. prior research shows that people who overestimate their skills or knowledge often make poor decisions and suffer negative consequences (e.g., kruger & dunning, 1999). indeed, if students simply believe that by participating in the seminar that they are now better able to handle their financial future, their “improved” confidence may actually result in poor financial choices. in contrast, if students believe they can better handle their financial future because they have more knowledge to apply, this increased confidence may result in more assertive and effective financial behaviors. indeed, researchers have identified that in addition to knowledge about finance, people must have a sense of financial self-efficacy to act on that knowledge (e.g., farrell, fry, & risse, 2016; lapp, 2010) to experience long-term financial success and stability. we speculate that our knowledge and confidence findings may represent a kind of financial self-efficacy (i.e., students gain knowledge and believe in their ability to apply that knowledge to positively affect their financial future). we now make an effort to gauge the “calibration” between student financial knowledge and confidence in their financial future. to do so, we use a simplified linear model of the form: confidence ¼ a þ bknowscore þ « (2) where we assume that a student’s confidence in their financial future is fully explained by their financial knowledge. we recognize that this is a strong assumption, but we believe that this allows us to cleanly evaluate whether participation in the pfp not only increases confidence but does so in a way that tracks increases in knowledge. we should note this calibration could come in multiple forms. for example, we may have students who, before the seminar, have significant confidence in their financial future but do not possess significant financial knowledge, classic overconfidence. in contrast, students that have high levels of financial knowledge may have low levels of confidence, classic under confidence. ideally, after participating in the pfp, a student’s confidence, while having improved, will also be better calibrated to their financial knowledge level, which has also improved. thus, the question we examine here is whether seminar participation leads to a better calibration between a student’s financial knowledge and confidence. to do so, we examine the residuals from eq. (2) for the control and treatment groups. if the seminar is useful in better aligning a participant’s knowledge and confidence, we expect the residuals for participants to decline from pre to post and the residuals for the control group to remain unchanged. before examining residuals from estimating eq. (2), we use the ratio of knowscore and confidence to gauge whether there is a relative change in values of knowscore and confidence from preto post-seminar. the results are presented in panel a of table 7. for the control group, the ratio of knowscore to confidence before the seminar is 5.657 and remains statistically similar at 5.313 after the seminar. therefore, any changes in confidence were roughly offset by changes in knowledge scores in a relative sense. in contrast, for the treatment group of seminar participants, the ratio drops from 6.987 to 4.715, statistically significant at better than one percent. from table 2, we know that both knowscore and 336 j. handy et al. / financial services review 29 (2021) 315–341 confidence increase for the treatment group. the result in panel a of table 7 indicates that seminar participant confidence changes much more than knowledge in relative terms. while improved confidence is a desired outcome, it is not if it does not coincide with a similar increase in knowledge level. seeing oneself as able to manage finances in the future would seem to suggest a willingness and confidence to apply one’s financial knowledge. in panel b of table 7, we examine the residuals of estimating eq. (2) for the control and treatment groups preand post-seminar. if the seminar simply leads to overconfidence, we would expect the residuals for the treatment group to be higher in the post-seminar survey than in the pre-seminar survey. first, students exhibit overconfidence relative to their financial knowledge. in all four cells of panel b, the residuals are greater than zero. however, in the treatment group row, residuals are statistically lower for the seminar participants (treatment group) but not the control group. this finding suggests that while pfp participants exhibit overconfidence, even after the seminar, there is better alignment between their financial knowledge and confidence in their financial future after attending the seminar, a finding not reflected in our control sample. 7. section 6: concluding remarks this paper summarizes the pfp, a six-week financial literacy improvement program for senior college students at a private liberal arts college, that focuses on major areas of table 7 reconciliation of financial knowledge and confidence panel a: variables pre-seminar ratio post-seminar ratio difference control group 5.657 5.313 �0.344 treatment group 6.987 4.715 �2.272*** ***p < 0.01, **p < 0.05. panel b: variables pre-seminar residual post-seminar residual difference control group 0.735 0.717 �0.018 treatment group 0.677 0.528 �0.149*** this table presents results examining the relationship between students’ knowledge scores as predictors of their confidence in their financial future. in panel a, we examine the ratio knowscore/confidence to provide intuitive motivation for this discussion. in panel b, we assume that confidence can be completely predicted by a student’s knowscore. as such, we estimate the single factor model identified as eq. (2) in the text: confidence = a + bknowscore + « for both the control and treatment groups both preand post-seminar. we then use model parameters to estimate a predicted confidence score, then take the residual between the predicted and actual scores. we then test the differences in residuals for the control and treatment groups between the preand post-seminar periods. ***p < 0.01, **p < 0.05. j. handy et al. / financial services review 29 (2021) 315–341 337 personal finance. an empirical analysis of the program shows that the pfp improves financial knowledge, and has a particularly positive impact on female participants. the results for minority participants, while encouraging, are not as robust as for women. both groups have been consistently shown to possess lower financial knowledge scores. additionally, participant confidence regarding their future financial outlook significantly improves for all participants: male, female, and minority. finally, we demonstrate that the improvement in confidence is well placed in that seminar participants demonstrate an improvement in their ability to calibrate their improved confidence to their improvement in financial knowledge. this study extends the previous literature in that the pfp is six-weeks long (that is longer than many other interventions), and the change in knowledge and confidence from before and after the program was compared with a control group. furthermore, the study measures several individual difference factors, allowing the researchers to utilize multiple regression analysis techniques. the study is not without its limitations. first, we do not measure changes in student behavior. our speculation about the connection between increases in knowledge and confidence may lead to increased financial self-efficacy can only be tested if we examine how these increases are reflected in behavior. in other words, we can examine if the combination of more knowledge and confidence predicts acting on that knowledge in financially sound ways. the authors are in the process of collecting follow-up behavioral data on the study participants reviewed here. second, our research is limited by the sample we use. due to the nature of the pfp, participants have been limited to only senior level students at a private liberal arts college, that is, the conclusions drawn from this study are limited to a very specific portion of the population. however, the authors are currently expanding the pfp to other universities and are currently summarizing results obtained from a faculty or staff version of the pfp. finally, given the demographics of the student population represented, we are hesitant to claim all college seniors will respond in similar ways to the program. more research is needed to validate the curriculum in a more diverse setting. ultimately, this paper expands upon financial literacy research and outlines a robust tool for addressing financial knowledge and confidence. notes 1 one could suggest that the use of a financial advisor could mitigate the need for improved financial literacy. however, handy and smythe (2020) and handy, smythe, and ricketts (2020) document that retail mutual fund investors face the potential of being misled into sub-optimal funds by well-intentioned advisors. 2 the program was first developed approximately 10 years ago but was not the focus of research until 2017. no program was offered in spring 2020 or fall 2020 due to the onset of the pandemic. a program was provided in spring 2021 but with an online delivery, which will allow us to examine if delivery method impacts our findings. the program is also being expanded to a new student population, as one of the authors has changed institutions. 338 j. handy et al. / financial services review 29 (2021) 315–341 3 this is taken from the white paper “closing the gap between knowledge and behavior: turning education into action” resulting from the symposium sponsored by the national endowment for financial education. 4 see knoll and houts (2012) for the list of questions, or they can be provided by the authors upon request. the knoll and houts questions include the core three questions introduced by lusardi (2008) and seven of the nine additional questions used by lusardi and mitchell (2007). 5 we did include a respondent’s ‘confidence in answers to the knowledge questions’ initially. unlike prior work, in a multi-variate framework, we found either no effect or a positive effect on knowledge. 6 the authors thank taylor vahle for her keen observation with respect to this and other points. 7 we would like to thank a reviewer for suggesting this analysis. appendix: summary of personal finance modules budgeting session the budgeting module begins with an explanation of what it means to maximize net worth, and defines what assets and liabilities are. in an effort to capture the students’ attention, there is a slide demonstrating approximately how much they might need to live comfortably in retirement. students are then provided a set of key strategies, that if used, should help maximize wealth (e.g., reduce spending and increase saving). concepts covered include: the purpose of a budget, steps to creating a budget, the difference between fixed and variable expenses and examples of each, how to estimate expenses, surprises likely to interrupt one’s budget that can be planned for (e.g., annual car maintenance), financial record keeping, reviewing/updating one’s budget, cash management, and the need to establish an emergency fund. throughout the session, faculty stress the need to regularly review budgets and that budgets reflect one’s stage in life. credit management session this session emphasizes the importance of maximizing wealth by effectively minimizing and managing liabilities. the focus is not that all debt is bad but that it must be used wisely and fit within the budget. concepts covered include: defining credit, providing examples of consumer credit, discussing credit costs, defining open-end and closed-end credit, providing and discussing an example of amortized loans, a specific look at credit cards and a discussion on how to choose a credit card, a discussion on credit card fraud and protective strategies, credit card uses and misuses, advantages and disadvantages of card use, reasons individuals find themselves carrying high debt levels, and credit-worthiness and credit scores. the discussion of credit scores goes beyond credit usage by explaining how credit scores are being used to set insurance premiums and make employment decisions. j. handy et al. / financial services review 29 (2021) 315–341 339 risk and return session this module is treated separately for two reasons. first, the authors concluded that the risk/return relationship is one most often “forgotten” and misunderstood by investors, especially those that are not financially inclined (a significant population in these sessions). second, the principles are essential to getting students to understand the need to invest in capital markets to achieve goals like retirement. this session reminds students of the importance of maximizing wealth and introduces how uncertainty or risk can impact decisions. concepts covered include: a historical look at risk and return, a formal definition of risk and return, fundamental rules of investing, implications of risk, and diversification (what it is, graphical representation, statistical information, and how to achieve in general). mutual funds session this module is included separately because mutual funds are the primary vehicle that students will likely use to participate in capital markets, and given the growing complexity of fee structures, especially in the advisor-sold channel (see handy et al., 2020), the authors decided to devote one module to funds. additionally, the module helps develop and reinforce the principle of diversification. the session emphasizes the importance of asset class diversification. students learn that mutual funds are one of the most popular asset growth tools. concepts covered include: defining a mutual fund, discussing the types of mutual funds (load vs. no-load), a significant focus on fund fees, load structure (e.g., a-class, b-class, or c-class), general fund investment types and information on each type including risk profiles, fund families, and how to evaluate mutual fund performance. retirement planning session this session begins by discussing the importance of wealth maximization, especially for this goal. the session covers popular retirement beliefs and myths, the importance of starting to save early, the state of social security, defined contribution plans and important terminology, traditional and roth iras, annuities, how to start a retirement account, and 529 college savings plans. risk management session this module is intentionally titled risk management instead of insurance to focus attention on why we need insurance—to protect net worth. the session reminds students of the importance of maximizing wealth but introduces them to the need to protect their assets from large loss via risk management. students learn general insurance terminology (e.g., coinsurance, deductible, or copay), and are exposed to health insurance, auto insurance, disability insurance, life insurance, and renter’s insurance. emphasis is placed on how insurance needs vary by life stage and insurance coverage should be revisited periodically. references allgood, s., & walstad, w. (2013). financial literacy and credit card behaviors: a cross-sectional analysis by age. numeracy, 6(2), 1–26. association for financial counseling and planning education. (2006). closing the gap between knowledge and behavior: turning education into action. journal of financial counseling and planning, 17, 73–90. bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological [database] review, 84, 191–215. bosshardt, w., & walstad, w. (2014). national standards for financial literacy: rationale and content. the journal of economic education, 45, 63–70. 340 j. handy et al. / financial services review 29 (2021) 315–341 cacioppo, j. t., & petty, r. e. (1982). the need for cognition. journal of personality and social psychology, 42, 116– 131. diener, e., emmons, r. a., larsen, r. j., & griffin, s. (1985). the satisfaction with life scale. journal of personality assessment, 46, 71–75. farrell, l., fry, t. r. l., & rissie, l. (2016). the significance of financial self-efficacy in explaining women’s personal finance behaviour. journal of economic psychology, 54, 85–99. fox, j., bartholomae, s., & lee, j. (2005). building the case for financial education. journal of consumer affairs, 39, 195–214. gerardi, k., goette, l., & meier, s. (2013). numerical ability predicts mortgage default. proceedings of the national academy of sciences, 110(28), 11267–11271. handy, j., & smythe, t. (2020). are multiple-share class funds poorly governed? financial services review, 28, 49–65. handy, j., smythe, t., & ricketts, j. (2020). analyzing the evidence: are multi-class mutual funds the right choice?” journal of index investing, 10(4), 49–58. henager, r., & cude, b. (2016). financial literacy and longand short-term financial behavior in different age groups. journal of financial counseling and planning, 27, 3–19. huston, s. (2010). measuring financial literacy. journal of consumer affairs, 44, 296–316. knoll, m., & houts, c. (2012). the financial knowledge scale: an application of item response theory to the assessment of financial literacy. journal of consumer affairs, 46, 381–410. kruger, j., & dunning, d. (1999). unskilled and unaware of it: how difficulties in recognizing one’s own incompetence lead to inflated self-assessments. journal of personality and social psychology, 77, 1121–1134. lapp, w. m. (2010). the missing link: financial self-efficacy’s critical role in financial capability. san francisco, ca: earn research institute. lusardi, a. (2008). financial literacy: an essential tool for informed consumer choice. nber working paper series, 14084. cambridge, ma: national bureau of economic research. lusardi, a., & mitchell, o. (2007). baby boomer retirement security: the roles of planning, financial literacy, and housing wealth. journal of monetary economics, 54, 205–224. lusardi, a., & mitchell, o. (2011). financial literacy and retirement planning in the united states. journal of pension economics and finance, 10, 509–525. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. [database]. journal of economic literature, 52, 5–44. lusardi, a., & mitchell, o. (2017). how ordinary consumers make complex economic decisions: financial literacy and retirement readiness. quarterly journal of finance, 07, 1750008–1750031. lusardi, a., mitchell, o., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences, and over-indebtedness. nber working paper, 14808. cambridge, ma: national bureau of economic research. mandell, l. (1998). our vulnerable youth: the financial literacy of american 12th graders. washington, dc: the jump$tart coalition for personal financial literacy. roberts, m. r., & whited, t. m. (2013). endogeneity in empirical corporate finance. in g. m. constantinides, m. harris, & r. m. stulz (eds), handbook of the economics of finance (pp. 493–572). new york, ny: elsevier. toker asad, c. (2015). financial literacy and financial behavior: assessing knowledge and confidence. financial services review, 24, 101–117. j. handy et al. / financial services review 29 (2021) 315–341 341 pii: 1057-0810(93)90005-b ftnancial services review, 3(l): 45-58 copyri&t 0 1993 by jai f&s inc. issn: 1057-0810 all rights of reproduction in any form reserved. real income growth and optimal credit use jessie x. fan y. regina cbang sherman hanna borrowing may be optimal ifreal income is expected to kcrease. rfincome growth is uncertain, optimal credit use is not obvious. a two period model of consumption for determining optimal credit use is presented, the impact of real income growth is malyzed with numerical analysis. the results may be use&l formal cozmselors and educators, as well as for insight into empirical patterns of et-e&t use. the income growth rate expected by the ~~e~~ plays a cracial role in determining opt credit ase for cm& co3zsamption. i. introduction economic i~ves~ent theory models developed by fisher (1930) and ~~hleifer (1970) suggest consumers may increase market ~p~~~ities and their utility through judicious selection of debts and assets (herendeen 1975). if a consumer is uncertain about future income, a small sustained growth (decrease} in real income or a substantial “expected” one-time increase (decline) might lead to borrowing (or saving) to smooth cons~on over life cycle, young cons~rs, especially students, and other families with temporarily low income might find borrowing rational. clearly, the use of consumer credit makes it possible for families and individuals to have the immediate consumption of goods and services and thus raise their level of living and satisfaction, however, the dramatic growth of consumer installmeut debt and the holding and use of credit cards from the past two decades (eastwood 1985; canner 1988), has led financial planners and educators to express alarm regarding whether consumers are becoming debt-ridden and overextended. the purpose of this paper is to describe a model for determining optimal credit use decisions for consumers with uncertain future income. the vehicle of analysis is the &s&x.fan . assistant professor, family and consumer studies department, university of utah. y. regina ciwng * post-doctoras researcher, family resource management department, the ohio state university. sherman hanna l ohio state university, family resource management department, columbus, oh 43210-1295. 46 financial services avow, 3(l) 1993 familiar two-period model of consumption. analysis is confined to credit for current consumption. ii. theliteratu~re there has been extensive discussion in the literature of optimal saving (borrowing) and consumption behavior under uncertainty either in the context of infinite time horizon or in two-period or rnulti~~~ ~~~ernpor~ models (e.g., leland (1968); levhari & srinivasan (1969); sandmo (1970); mirman (1971); dreze & modigliani (1972); sibley (1975); hey (1979); salyer (1988). in general, the authors analyze one or two variables at a time, assuming a value for each of the other parameters. for example, in two-period models the effects of income and interest rate uncer tainty on borrowing (or saving) decisions are analyzed, given an assumption of a certain lifetime. infinite horizon or finite horizon models explore effects of the discount factor (lifetime uncertainty) on borrowing (saving) behavior while assum ing absence of income and interest rate uncertainty. in the discussion of income uncertainty and saving behavior, it is assumed that the consumer’s beliefs about the value of future income can be summarized in a subjective probability density unction. on the basis of the probabi~ty density function, the consumer maximizes expected utility of consumption. leland (1968) uses a two-period model of consumption to demonstrate the effect of uncertainty on saving and concludes that with an additive utility function and the assumption of decreasing absolute risk aversion, the precautionary demand for saving is a positive function of uncertainty. sandmo (1970) discusses the effects of increased riskiness of future income on present consumption in a two-period model and proves that increased uncertainty about future income decreases consumption (increases saving). sibley (1975) extends the two-period result of the effects on optimal savings of increased risk in the future income to the multiperiod analysis discussed by leland (1968). sibley (1975) suggests that increased wage unce~~~ increases saving. for the case of a constant elasticity utility function, levhari and snivel (1969) show that optimal savings can increase with increasing uncertainty. however, those authors mainly emphasize the effects of the subjective probability density function as a projection of uncertain future income on saving (or borrowing) behavior. no study has been done relating levels of risk aversion, interest rates, and income growth rates to optimal borrowing in a model incorporating uncertainty. kinsey and lane (1978) point out that when consumption is accompanied by the use of consumer credit, utility maximization may be viewed in a lifetime sense, thus a life cycle approach to the allocation of income, consumption, and saving (crowing) is appropriate. addition~ly, by appropriate inte~retation, two-period models can describe completely the individu~‘s resource allocation problem during their lifetime, if the focus is on consumption in that period and total consumption in all future periods (hey, 1979). with additional assumptions on certain risk properties of utility functions, a two-period model with uncertainty for determining real income growth and optimal credit use 47 optimal credit use facing consumers is presented and illustrated with numerical analysis. implications for a life cycle model arc then discussed. factors affecting optimal credit use include the expected growth rate of real income, the variance of.future income, the consumer’s utility function (e.g., the parameter of risk aversion), the real interest rate, and the consumer’s personal discount rate. iii. a two-period model of consumption to begin, consider the following model: assume that consumers attempt to maxi mize the expected value of utility (7) for the two periods. they will make their borrowing (or saving) decision in conjunction with their known first period income. the second period consumption will, of course, be a random variable, dependent on the actual value of second period income which is assumed to be affected by income growth rate (or decrease rate) and the probability that income growth occurs, and also dependent on the interest rate of borrowing (or saving). it is assumed that there are two states of the world in the second period-real income either increases or stays constant. (the analysis could allow for other scenarios, but the discussion is limited to this scenario because it is the most plausible scenario for borrowing to be rational). c, and c, represent consumption in these states. finally, consumers are assumed to repay the loan in full in the second period. mathematically, the problem can be formulated as: t= u(c,) + p* u(c*)+(l-p)* u(c,) (1 +p) (1) the constraints are: cr = i s (2) c*=(l+g)*l+(l+r)*s (3) c,=z+(l+r)*s (4) variables: t= total two period utility i= year 1 income year 2 income = (1 + g) * i (if income increases in that year), otherwise, year 2 income = year 1 income c1 = consumption in year 1 s = the amount of savings in year 1 (negative value means borrowing.) j?inancial services review, 3(l) 1993 c, = consumption in year 2 if real income in year 2 increases c,= cons~ption in year 2 if real income in year 2 does not increase g = growth rate in real income r = real interest rate (note that r may be higher for s < 0, i.e., borrowing, than for s > 0) p= probability that real income increases p = personal discount factor a consumer may discount utility from future consumption because of the possibility of not being alive then, or because of other possible changes in capacity to derive utility from consumption. discounting because of the risk of death should be small for a young adult, although some younger consumers may discount future consumption because of “impatience” or limited ability to imagine utility as a middle aged or elderly consumer. it is possible to find the level of c1 and c, (and therefore savings or dissavings) which produce the highest feasible level of utility by using calculus. only general results can be derived from equation (l), unless restrictions are placed on the utility function. most studies of intertemporal consumption have used a constant elasticity utility function (hurd, 1989) which is additively time separable: the elasticity of marginal utility with respect to consumption is -x. the elasticity of intertemporal substitution in consumption is equal to l/x. when this type of utility function is used for analysis of risk, the parameter x is relative risk aversion. c is consumption per time period. estimates of relative risk aversion grossman and shiller (1981) have given x an interaction as ‘i... a measure of the concavity of the utility function or the disutility of consumption fluctuations” (grossman & shiller, 1981, p. 224). the higher the value of x, the more risk averse is the consumer, and the more rapidly marginal utility decreases as consumption or wealth increases. the analysis of economic behavior under un~~~nty uses relative risk aversion extensively. for intertemporal consumption, empirical estimates of x range from just under two (skinner, 1985) to 15 (hall, 1988). other estimates were between these two values. the wide range of estimates of the utility function parameter, x, relative risk aversion, may be due to several causes, including: real income growth and optimal credit use 49 1. many households face liquidity constraints (they have zero or low levels of assets, and therefore cannot dissave without borrowing at a much higher interest rate than can be obtained from safe investments). 2. the data sets used for empirical analyses did not contain appropriate variables. for instance, the theoretical measure of savings used in the literature is the amount not consumed. the measure used in many studies is the change in net worth. however, although it is often assumed that a consumer cannot identify a utility function explicitly, it may be possible to construct hypothetical examples that allow one to intuitively identify a unique utility function parameter. it is possible to create a scenario to obtain insight into the similar parameter for the intertemporal utility function. to obtain some insight into plausible values of relative risk aversion, x, consider the following hyporhetical situation: you are 20 years old, and know with certainty that you will live to be 100 in good health. everything about your personal situation will remain the same for the next 80 years. you want to spend all of your wealth by the day of your death. your non-asset income will be $20,000 per year in real (constant dollar) terms. you can obtain exactly 6% per year after inflation and taxes on investments. table 1 shows optimal consumption paths for different values of x, assuming p = 0. based on the hypothetical example, a value of x = 1 (which corresponds to a natural logarithm utility function) would seem extremely miserly, as you would spend only $4,323 of your $20,000 income at age 20 in order to enjoy $457,382 of consumption the last year of your life. a value of x = 6 might be representative of the typical american consumer, as the consumer would spend $16,929 out of $20,000 income at age 20, and could spend $36,817 at age 100. it seems likely that most americans would have a value of x between four and eight. kimball’s (1988) hypothetical example for relative risk aversion, may imply a value between four and eight (hanna 1988). the utility function u(w), and the expected utility eu(w) are specified as follows, table 1. optimal consumption by x, hypothetical example age x=1 x=2 x=3 20 4,323 11,104 13,942 30 7,742 14,859 16,931 40 13,865 19,885 20,560 50 24,83 1 26,611 24,968 60 44,468 35,611 30,320 70 78,635 47,656 36,820 80 141,614 63,774 44,713 90 255,400 85,344 54,299 100 457,382 114,210 65,939 x=4 15,421 17,840 20,637 23,874 27,618 31,948 36,959 42,754 49,459 x=5 x=6 16.323 16.929 18;341 18;656 20,608 20,558 23,155 22,655 26,017 24,966 29,233 27,512 32,846 30,318 36,905 33,410 41,467 36,817 x=20 19,073 19,637 20,217 20,815 21,431 22,064 22,716 23,388 24,079 50 financial services review, 3(l) 1993 table 2. hypothetical example of relative risk aversion relative risk aversion lowest value of1 0 0 1 25,ooo 2 33,333 3 37,796 4 40,548 6 43,665 10 46,299 20 48,209 u(w)=fy eu(w) = c pju(wj) (6) (7) where x = relative risk aversion level w = total wealth a modified version of kimball’s (1988) example developed by hanna (1988), could explain the concept of relative risk aversion in the context used. assume that you have one year to live, and may choose an investment to provide you with your consumption for the next year. once you choose, it will be impossible for you to obtain income from any other source. you have no assets of any kind. you may choose one of two plans: a or 3. plan a pays you a tax free real income of $50,000 per year, while plan b involves a gamble. if you choose plan b, the government in effect flips a coin, and there is a 50% chance of having a real income of $100,000 tax-free, and a 50% chance of some lower income i. at what level of i would you be indifferent between plan b and plan a. table 2 shows how your answer corresponds to your level of relative risk aversion. in the context of the expected utility model, relative risk aversion relates to the extra utility of increased consumption if the gamble pays off compared to the lost utility because of decreased utility if you lose the gamble. for instance, if you have a relative risk aversion level of four, you value the gain of utility from increasing your consumption from $50,~ to $ l~,~ the same as the loss of utility from decreasing your income from $50,000 to $40,548 (hamra, 1988). intuitively, a level of at least four would seem reasonable for most people. by combining inter-temporal consumption analysis with risk aversion, we can obtain the optimal amount of saving in terms of year 1 income, interest rate, income growth rate, and probability of that income increases. to give some intuitive insight into optimal credit a model with perfect certainty will be examined first. real income growth and optimal credit use 51 iv. optimal creditwhi-iperfzctcertainty if a consumer is certain that real income will increase with a growth rate g, and the consumer faces a real interest rate r, we can derive by calculus the optimal pattern of consumption (expressed as the growth rate of real consumption), as shown in equation (8). for plausible values of the real interest rate r, the personal discount rate, p, and relative risk aversion x, the optimal growth rate in consumption approximately equals the interest rate minus the personal discount rate, divided by relative risk aversion. note that the optimal growth rate in real consumption does not depend upon income patterns, although income patterns over time may influence the interest rate faced by a household, and also the feasibility of particular consump tion patterns. based on a regression of u.s. household expenditures on age and age squared, it can be estimated that real expenditures of households tend to increase with age until age 49 (bae, 1993, p. 137). consumption levels of individual households tend to be relatively stable over a lifetime, although there are several plausible theoretical explanations for this stable pattern (yunker, 1992). households with substantial amounts of financial assets face real, after tax interest rates of 0% to 6% on safe, liquid assets. households who would have to borrow might face real interest rates between 3% (home equity loan) and 20% (finance company.) the fact that the elderly are most likely of any age group to have declining levels of real consumption implies (at least within the life cycle framework) that the personal discount rate depends on the risk of death. there does not seem to be strong evidence of extra “impatience” by young consumers. if the personal discount rate is related to the risk of death, it is unlikely to be an important factor in credit decisions. it will be shown below that credit use can be explained by an observable factor, growth rates of real income, so there are advantages to dispensing with the unobservable factor of the personal discount rate. in any case, as the approximation in equation (8) shows, the effect of a higher personal discount rate can be approximated by using a lower real interest rate. (g cl) % =l+r -1 t 1 ‘-p cl l+p =x (8) equation (9) gives optimal savings as a proportion of year 1 income. the consumer’s relative risk aversion is x. for particular values of r and g, the greater the relative risk aversion, the more the consumer should borrow. this seemingly paradoxical result is due to the fact that the two period model with certainty involves no risk, but only intertemporal allocation. if the consumer faces a higher interest rate for borrowing than for saving, there may be some growth rates for which neither borrowing nor saving is optimal. if the ratio is negative, borrowing is optimal. 52 financialservicesrrview,3(1) 1993 l+r ’ t 1 -cl+& l+p [$+[+=] (9) the natural log utility function (u= ln[q) has been used frequently, although the example from table 1 implies extremely miserly behavior. it is simple to analyze. substituting the value of x = 1 in equation (9), then for reasonable values of r, equation (10) is a good approximation. s_(r--pi-g i2(1 + r) (10) note that if r-p is less than g, saving is negative, so that some borrowing is optimal. for a given value of r, as g increases, the optimal amount to borrow will increase. for instance, if g = 20%, p = 0% and r = 10% then s/z = -4.5%. if this year’s income is $10,000, the optimal amount to borrow is $450. for other values of x, equation (9) is somewhat complicated. however, an intuitive sense of the patterns can be obtained by using approximations, resulting in equation (11). if the real interest rate (r) minus the personal discount rate (p), divided by the relative risk aversion, is less than the expected growth rate in real income (g), borrowing is optimal. for instance, if r is 12%, p is 0%, and x is 6, then borrowing is optimal if g is greater than 2%. clearly, one does not need to rely on the assumption of a high personal discount rate to explain borrowing for current consumption. an analysis of households in the survey of consumer finances interviewed in 1983 and re-interviewed in 1986 (chang and lindamood, 1993) shows that, based on two year household income in 198211983 and two year household income in 1984/1985, the median growth rate in real income for households under the age of 35 was 20%. the annualized median growth rate for young households was almost 10%. for the households under 35, 25% had an annualized growth rate of 22% or higher. consumers who were certain of a high growth rate might borrow even if they faced rather high interest rates. ‘-p borrow if 7 < g v. optimalcreditw~iuncerta~y (11) there is no simple, closed analytical solution for optimal savings or credit with uncertainty. therefore, simulations were used to find optimal savings/credit. real income growth and optimal credit use 53 simulation results: factors affecting optimal credit use equations (1) through (4) were used with simulations to find the value of s that maximized expected lifetime utility for particular values of the parameters. in this section, we shall discuss and illustrate effects of real income growth on optimal savings/credit use, for two levels of relative risk aversion and three probability levels. in order to focus on scenarios with borrowing, it was assumed that the consumer faced either constant real income or a real income growth rate g with a probability p. in the cases of certainty (i.e., probability of income increase equals one), the greater the relative risk aversion, the less the consumer will save, or the more the consumer will borrow. the relative risk aversion is related to how much more utility the consumer will lose due to low consumption in year 1 than they will gain from higher consumption in year 2. for any given real income increase, the consumer will borrow more in order to smooth out consumption as much as is justified by the utility function and the real interest rate on loans. when uncertainty is added to the total period utility function (le., probability of income increases between zero and one), the borrowing-relative risk aversion relationship observed for certainty does not always hold. the simulations were based on the following assumptions: 1. the real interest rate on loans = 14.095% (e.g., nominal rate of 19.8% with 5% inflation.) 2. the real, after tax interest rate on savings = 1% (e.g., nominal interest rate of 8.4%, subject to 28% tax rate and 5% inflation.) 3. expected utility from all possible borrowing levels (at 14.095%) is com pared to expected utility from all possible saving levels (at 1%) and optimal saving/borrowing is that which produces highest expected utility. 4. the personal discount rate, p, is zero. figure 1 shows the relationship between the optimal ratio of amount saved in year 1 to year 1 income and the rate of increase of real income, for three probabilities {50%, 95% and 100%) that income will increase, each for two levels of relative risk aversion (one and six). table 3 shows similar info~ation, except expressed as the amount borrowed as a percent of year 1 income. there are threshold levels of real income growth for any borrowing to be optimal (table 4). for real income growth levels below these threshold levels, there is a range of growth levels for which neither borrowing nor saving is optimal. at very low levels of growth, a small amount of saving is optimal. for instance, for relative risk aversion of 1 .o, if the consumer is certain that real income will remain constant, the optimal amount to save out of year 1 income is 0.50% of income. if the consumer thinks there is a 50% chance that real income will increase by l%, optimal saving will be 0.25% of income for relative risk aversion of 1 .o but 0.0% for relative risk aversion level of 6.0. 54 ftnancial services revxeiw, 3(l) 1993 p=60%, x=6 --‘--.~r~~**~....*“.....~~.~*,....*..~.*...*....**......~~**~~.~~~~~~~*~*~* ------____ ---_----._--____._________ p=50%, x=1 -0% . -5% e 8 -10% + 5 -15% f “0 -20% 8 \ s -25% p ‘g -30% 2 -35% h 0 -40% ‘\ 45% (relative risk aversion denoted by x) \ _50% (probability that real income increases denoted by p) 0% 20% 40% 60% 60% 100% real growth (if real income increases) figure 1. optimal savings (borrowing) as % of income, by growth rate, probably and risk aversion in the cases of certainty, there is a virtually linear increase in the optimal amount to borrow as g increases. with x = 6.0, some borrowing is optimal even if the growth rate is only 5%. with x = 1.0, the growth rate must equal 20% for borrowing to be optimal. if the consumer is 95% sure that real income will increase, the pattern is almost the same as certainty if x = 1 .o, but the pattern is very different ifx = 6.0, with optimal borrowing at g = 100% forp = 95%, less than half the amount table 3. optimal amount to borrow, as percent of year 1 income, by growth rate(g), relative risk aversion (x) and probability that income increases growth rate 10% 20% 30% 40% 50% 60% 70% 80% 90% 1@3% x=1 x=6 p=50% 0.00% 0.00% 0.02% 1.60% 3.03% 4.34% 5.52% 6.61% 7.59% 8.49% p=ps% 0.00% 2.57% 6.66% 10.69% 14.65% 18.55% 22.36% 26.01% 29.68% 33.17% p=loo% 0.00% 3.10% 7.52% 11.95% 16.37% 20.80% 25.22% 29.65% 34.07% 38.50% p=50% 0.98% 2.25% 3.05% 3.54% 3.85% 4.05% 4.17% 4.25% 4.31% 4.35% p=95% 3.37% 7.42% 11.05% 14.08% 16.42% la.1096 19.26% 20.04% 20.55% 20.90% p=ioo% 3.69% 8.34% 12.99% 17.64% 22.29% 26.94% 31.59% 36.24% 40.89% 45.54% real income growth and optimal credit use table 4. minimum growth rate needed for borrowing to be optimal, by relative risk aversion(x) and probability that income increases probability x=1 x=6 50% 30% 5% 95% 14% 3% 100% 14% 3% for p = 100%. if there is a 50% chance that real income will increase, borrowing is not optimal for x = 1.0 unless growth of 35% or more is expected; while with x = 6.0, some borrowing is optimal even for growth of 5%, but the amount is limited. a consumer expecting a high probability of a substantial increase in real income may rationally borrow a large amount of money for current consumption. the importance of a correct assessment of the probability of an income increase may be seen in figure 1. for any particular level of relative risk aversion, optimal borrowing is substantially greater for higher probabilities. note, however, that regret is not assumed to enter the utility function. if a consumer with an income of $40,000 has a relative risk aversion of six and a probability of 95% that real income will increase by 50%, then there is a 5% probability of having the discomfort of repaying the $5,860 borrowed out of an unchanged income of $40,000. this will result in a drop in real consumption from $45,860 in year 1 to $33,314 in year 2, or a drop of 27.4%. vi. extensions of the model durable goods the analysis presented assumes that all spending is for current consumption, which may be realistic for a consumer who rents a home and leases automobiles. use of credit for some types of durable goods, such as automobiles and kitchen/laun dry appliances, may be rational even if real income is not expected to increase. however, for decisions about how expensive the durable good should be beyond minimum standards (e.g., reliable transportation), the analysis presented in this article may give some insights into how such choices should be made. extensions to more than two periods if the analysis is extended to three periods, but the assumption is made that the real income level during the third year is whatever the real income level is during the second year, then optimal credit for the first year is higher than the corresponding level shown in this article for the two period model. allowing for changes between 56 financial services avows 3(l) 1993 year 2 and year 3 introduces a much higher degree of complexity, and has not been addressed by the authors in this article. allowing for decreases in real income if there is a possibility that real income will decrease, optimal saving may be positive. for instance, if real income will either remain constant or decrease, and both states of the world are equally likely, the consumer should save some money from year 1 income in order to prevent too much of a decrease in year 2 consumption. if, however, the probabili~ that real income decreases is small, optimal saving may be low. extensions to more than two states of the world if there are more than two states of the world, analysis of optimal behavior is complex. for consumers with a very small chance of a large decrease in real income, and approximately equal chances of constant real income or a substantial increase in real income, the optimal saving (credit) pattern will be approximately the same as the patterns presented in this article for the two states of the world model. for some consumers, a small ~ssibili~ of a subst~ti~ decrease in income could be dealt with through help from relatives and the social safety net. consumers could also implicitly assume that decreases in real income could be dealt with by default. extensions to a life cycle model the two period model can be extended to a life cycle model if certainty is assumed. for probabilities greater than 98% that real income will increase, there may not be substantial differences in optimal credit use, if it can be assumed that real income will either increase or remain constant after the first year. for many households, the simplifying ~sumption that income will either increase or remain constant is very unrealistic. however, if there is a small probability that there will be a substantial drop in real income, a consumer who has taken on credit for current consumption has the option of default or some form of bankruptcy. these options, as well as the opportunity to repay credit early, can be viewed as put options which have implicit value for the borrower. to model the costs of bankruptcy is beyond the scope of this paper. taking default into account the model used in this article ignores the possibility of default and/or bank ruptcy. the results described are based on the ~sumption that the consumer must repay the loan in full. there are obvious costs of default and various forms of bankruptcy. if these costs were low, even more consumers would become overex tended. it is difficult to specify the monetary value of the costs of default, etc., but given the increasing number of bankruptcies, it is plausible that a priori, borrowing real income growth and optimal credit use 57 even if bankruptcy is possible is rational for some consumers. it is also plausible that many consumers may underestimate the true costs of bankruptcy, and therefore take too much risk with credit use. most consumers who use credit face at least a small risk of default. the analysis presented in this article provides a starting point to development of a realistic evaluation of rational credit use. vii. s~jmmary and conclusion a two-period model of consumption is developed to analyze optimal credit use decisions, based on the probability that future real income will increase, for different levels of relative risk aversion. with the assumptions that the utility function be additive between periods, and reflect constant relative risk aversion, effects of parameters on optimal borrowing and saving decisions and the interacting relation ships are discussed and demonstrated using numerical simulation technique and graphs. we have shown that the optimal amount of credit use increases with increasing income growth rate and with increasing probability of real income growth. there are threshold levels of real growth rates for borrowing to be optimal, at relatively low levels for certainty, but at high levels for lower probability levels that growth will take place. the optimism of the middle of the 1980s and the longest peacetime expansion (see discussion in chang and lindamood, 1993) may have led to the increase in consumer credit and may have contributed to the increase in personal bankruptcy rates. if young consumers in the future expect stagnant real income levels and have high levels of uncertainty about their prospects, the amount of consumer credit will continue to decrease. only optimistic consumers will borrow for current consumption. the simplest possible model of optimal credit use dealing with uncertainty has been developed. clearly it would be desirable to extend the model to the borrowing and saving decisions over more than two time periods. more complicated models such as multiperiod or a life span analysis of optimal saving and borrowing decisions, however, may require complex computer programming techniques. fur ther analyses using empirical data may be needed for comparisons between theo retical and empirical results. references bae, mikyeong 1992. analysis of household spending patterns. ph.d. dissertation, the ohio state university. canner, glennb. 1988. “changes in consumer holding anduseof credit cards, 1970-1986,“journal of retail banking, lo( l), spring, 13-24. chang, y. regina, and lindamood, suzanne 1993. “factors related to the risk of household income variability,” financial counseling and planning, 4,47-66. dreze, jacques h., and modigliani, france 1972. “consumption decision under uncertainty,” joumal of economic theory, 5(3), december, 308-335. 5s financial services review, 3(l) 1993 eastwood, david b. 1985. “a re-examination of consumer credit growth,” in karen p. schnittgrund (ed.) proceedings of the 31st annual conference of american council on consumer interests, missouri, columbia: american council on consumer interests, 169-174. fisher, irving 1930. the theory of interest. new york: macmillan co. grossman, sanford j., and shiller, robert j. 1981. ‘the determinants of the variability of stock market prices,” american economic review: papers and proceedings, 7 l(2). may, 222-227. hall, robert 1988. “intertemporal substitution in consumption,” journal ofpolitical economy, 96(2), 339-357. hamra, sherman 1988. “risk aversion and optimal insurauce deductibles.” american council on consumer interests proceedings, 141-147. herendeen, james 3. 1975. “the role of credit in the theory of the household,” the joamal of consumeraffairs, 8,157-181. hey, john d. 1979. uncertainty in microeconomics, new york: new york university press. hirshleifer, jack 1970. investment, interestandcapital, englewood cliffs: new jersey, prentice hall. hurd, michael d. 1989. “mortality risk and bequests,” econometrica, 57(4), 779-813. kimball, miles s. 1988. “farmers’ cooperatives as behavior toward risk,” american economic review, 78(l), 224-232. kinsey, jean, and lane, sylvia 1978. ‘the effect of debt on perceived household welfare,” the journal of consumer a$nirs, 12( 1), (summer), 48-62. leland, hayne e. 1968. “saving and uncertainty: the precautionary demand for saving,” quarterly journal of ecotzumics, lxxxil(3), 465473. levhari, david, and srinivasan, t. n. 1969. “gptimal savings under unce~n~~ review of ~ono~c~~~s, 36(106), 153-163. mirman, leonard 1971. “uncertainty and optimal consumption decisions,” econometrica, 39(l), 179-185. sandmo, agnar 1970. ‘the effect of uncertainty on saving decisions,” review of economic studies, 37(11 l), 353-360. salyer, kevin d. 1988. “the characterization of savings under uncertainty: the case of serially correlated returns,“’ economic letters, 26( 1), 21-27. skinner, jonathan 1985. “variable lifespan and the intertemporal elasticity of consumption,” review of economics and statistics, 67,616-623. sibley, david s. 1975. “permanent and transitory income effects in a model of optimal consumption with wage income uncertainty,” journal of economic theou, 1 i( i), august, 68-82. yunker, james a. 1992. “‘reiatively stable lifetime consumption as evidence of positive time preference,” jotuna of post kiss economics, 14(3), 347-366. pii: 1057-0810(95)90010-1 benson, ear1 a., dayid s. rysrrom, and greg t. smersh, ‘~omm~ssjon mot~v~ied trad ing patterns of brokers across the produc tion month,‘” 4(2): 8 l-95 boldin, robert j., and keith leggett, “bank dividend policy as a signal of bank quai ityp 4(l): 1-8 chang, y. regina. see fan, jessie x, cox, larry a., and sandra g. gustavson, “the market priming of disability income insur ance for individuals,” 4(2): ! os-1 22 crabb, ronald r., “efficient frontiers in estate p~arrrring,” 3( i ): r-27 daniels. kenueth n., neil b. murphy, andden nis m. g’toole, “an empirical analysis of the use of money orders, the payment sys tem of the poor,” 3(l): 75-81 devaney, sharon a., and ruth h. lynon, ~~hou~~~o~d ~nso~ven~y~ a review of household debt repayment, delinquency, and ba~r~~~c~~ 4(2): 137-156 fan, jessie x,, ‘i*, regina char& and sherman hanna, “‘real income grow&h and opti& credit use,” 311 f : 45-58 fields, joseph a,, david s. kidwel, and linda s. klein, ‘%oupon resets versus poison puts: an ~~d~vidua1.s valuation of event risk provisions in corporate deb&” 3(2): f43-156 f&j, john l. see smaby, timothy r. gustavson, sandra g. see cox, larry a. hamill, james r, and joe1 s. stemberg. “tax savings ~pportnn~t~es in estate freeze tm~a~t~o~: an a~~~~at~~ of the black &holes modes,” 4(i): 9-22 hanna, sherman. .re~ fan, jessie x. heck, jean louis, terry l. zivney, and naval k. modani, “a simplified approach to measuring bond duration,” 4(i): 31-40 ho, kw& moshe arye mifevsky, and chris munson, “asset a~l~at~o~~ life expec tancy and shortfall,” 3(2): 109-i% kidwell, david s. see fields, joseph a. klein, linda s. see fields, joseph a. knight, john r., and lewis mandeft, ‘“gptimaf holding period of assets that must he liq uidated,” 4(z): 97-1og leggett, keith. see boldin, robert j, littte, philip l. see ragsdale, cliff t. lytton, ruth h, see devaney, sharon a, mandell, lewis, se@ knight, john r, ~an~iero, george, sect mangier, susan mangiero, susan, and george mangiero, “quantifying present valuation errors,’ 4(l): 23-w miievsky, moshe arye. see ho, kwok modani, naval k see heck, jean louis, murphy, neil b. see daniels, kenneth n. g’toote, dennis m. see daniels, kenneth n. poofe, barbara s., “‘a pm~t~t~o~r?s perspee tive: comments on ‘analysis of u,s. sav ings bonds,“’ 4( 1): 57-60 poole, barbara s., “a practitioner’s hrspec tive: comments on *asset allocation, life ~pe~~n&y and shortfar,“’ 3(2): lsw58 166 financial services review. 3-4 1993-1995 potts, tom l., and william reichenstein, stock returns to enhance portfolio per “analysis of u.s. savings bonds,” 4(l): formance,” 3(l): 29-43 41-56 robinson, chris. see ho, kwok puelz, amy v., “the individual’s tax-exempt rystrom, david s. see benson, earl d. bond portfolio decision under income un certainty,” 3( 1): 59-73 seila, andrew f., see ragsdale, cliff t. smaby, timothy r., and john l. fizel, “fund closings as a signal to investors: investment ragsdale, cliff t. ragsdale, andrew f. seila, and philip l. little, “an optimization model for scheduling withdrawals from tax deferred retirement accounts,” 3(2): 93-108 randolph, william lewis, “the impact of mu tual fund distributions on after tax re turns,” 3(2): 127-141 reichenstein, william. see potts, tom l. reichenstein, william. see rich, steven p. rich, steven p., and william reichenstein, “using the predictability of long-horizon performance of open-end mutual funds that close to new shareholders,” 4(2): 71 80 smersh, greg t. see benson, earl d. stemberg, joel s. see hamill, james r. sullivan, a. charlene, and debradrecnik wor den, “credit cards and the option to de fault,” 4(2): 123-136 worden, debra drecnik. see sullivan, a. char lene zivney, terry l. see heck, jean louis, articletitles “analysis of u.s. savings bonds,” tom l. potts and william reichenstein,4( 1): 41-56 “asset allocation, life expectancy and short fall,” kwok ho, moshe arye milevsky, and chris robinson, 3(2): 109-126 “bank dividend policy as a signal of bank quality,” robert j. boldin and keith leg gett, 4( 1): l-8 “commission motivated trading patterns of brokers across the production month,“earl d. benson, david s. rystrom, and greg t. smersh, 4(2): 8 1-95 “coupon resets versus poison puts: an indi vidual’s valuation of event risk provisions in corporate debt,” joseph a. fields, david s. kidwell, and linda s. klein, 3(2): 143 156 “credit cards and the option to default,” a. charlene sullivan and debra drecnik wor den, 4(2): 123-l 36 “efficient frontiers in estate planning,” ronald r. crabb, 3(l): l-27 “an empirical analysis of the use of money orders, the payment system of the poor,” kenneth n. daniels, neil b. murphy, and dennis m. o’toole, 3( 1): 75-81 “fund closings as a signal to investors: invest ment performance of open-end mutual funds that close to new shareholders,” timothy r. smaby and john l. fizel, 4(2): 7 l-80 “household insolvency: a review of house hold debt repayment, delinquency and bankruptcy,” sharon a. devaney and ruth h. lytton, 4(2): 137-l 56 “the impact of mutual fund distributions on after tax returns,” william lewis ran dolph, 3(2): 127-141 “the individual’s tax-exempt bond portfolio decision under income uncertainty,” amy v. puelz, 3( 1): 59-73 “the market pricing of disability income in surance for individuals,” larry a. cox and sandra g. gustavson, 4(2): 109-122 167 “optimal holding period of assets that must be liquidated,” john r. knight and lewis mandell, 4(2): 97-108 “an optimization model for scheduling with drawals from tax deferred retirement ac counts,” cliff t. ragsdale, andrew f. seila, and philip l. little, 3(2): 93-108 “a practitioner’s perspective: comments on ‘analysis of u.s. savings bonds,“’ barbara s. poole, 4( 1): 57-60 “a practitioner’s perspective: comments on ‘asset allocation, life expectancy and shortfall,“‘barbaras. poole,3(2): 157-158 “quantifying present valuation errors,” susan mangier0 and george mangiero, 4( 1): 23 30 “real income growth and optimal credit use,” jessie x. fan, y. regina chang, and sher man hanna, 3(l): 45-58 “a simplified approach to measuring bond duration,” jean louis heck, terry l. zivney, and naval k. modani, 4( 1): 3 l-40 “tax savings opportunities in estate freeze transactions: an application of the black scholes model,” james r. hamill and joel s. stemberg, 4( 1): 9-22 “using the predictability of long-horizon stock returns to enhance portfolio per formance,” steven p. rich and william reichenstein, 3( 1): 29-43 pii: s1057-0810(00)00053-6 the impact of the pension fund on the decision to work one more year< walt woerheide* rochester institute of technology, 108 lomb memorial drive, rochester, ny 14623-5608, usa received 14 march 2000 abstract the decision to retire or work one more year is quite complex. one factor that plays a role in this decision is the net salary derived from working one more year. the type of pension a person has, defined benefit or defined contribution, will influence whether the net salary is larger than, equal to, or less than the stated salary. the larger the net salary relative to the stated salary, the more likely an employee is to continue working. this paper defines the net salary percentage for each type of pension plan and looks at empirical estimates for these values. © 2000 elsevier science inc. all rights reserved. jel classification:g23; j33 keywords:pension; defined contribution; defined benefit; retirement; net salary 1. introduction a critical decision we all must make is the age at which we will retire. the retirement age is often treated in economic models as one that is known with certainty (modigliani & brumberg, 1954). many financial planning models have the worker designate a retirement date, and then project what a worker’s retirement income would have to be to sustain the lifestyle during retirement to which he or she is accustomed. the most popularly quoted rules < the work on this paper was funded in part by a summer research grant from the rit college of business. * corresponding author. tel.:11-716-475-5268; fax:11-716-475-6920. e-mail address:wjwbbu@rit.edu (w. woerheide). financial services review 9 (2000) 17–31 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00053-6 of thumb are that a worker would need a retirement income equal to 60 to 70% of preretirement income (see, e.g., wiatrowski, 1993). this is known as the replacement ratio. empirical research suggests that replacement ratios are a function of the level of income and do not appear to exceed the 80% threshold (dexter, 1984 and palmer, 1989). the retirement decision can also be viewed as an option. that is, at the start of each year the worker decides whether to retire or to work one more year. it is this last approach that is taken in this paper. the sources of retirement income consist of three broad categories. these are: 1) governmental retirement benefits such as social security, 2) income from pensions and other tax-favored retirement programs such as iras, and 3) income from personal investments. people retiring would be able to receive income from all three sources. people who decide to work one more year would receive their salary and their projected retirement income in subsequent years would likely be higher, but they would lose one year’s worth of retirement income. thus, anyone deciding to work one more year would be earning a “net” salary equal to his or her wage income plus the present value of the increase in future retirement income, less the retirement income he or she would have received in that year. the decision to retire or work one more year is not, of course, based solely on this net salary figure. the decision to retire requires the employee to compare the utility of the “net” salary to the utility of an additional year’s worth of leisure. if the utility of the net salary exceeds the utility of leisure for the next year, the individual will work. if not, the individual will retire. the individual’s utility function will be shaped by his or her wealth and income from other sources, his or her health, and many other factors. the basic assumption of this paper is that the larger the net salary number, the more likely the individual is to work one more year. this is just another way of saying that people have a positive marginal utility of wealth, a common assumption in economic research. nonetheless, individuals may have a net salary less than their stated salary and still opt to work, or they may have a net salary greater than the stated salary and yet opt to retire. the net salary number only influences the retirement decision, it does not define it. the purpose of this paper is to examine how the type of pension a person has influences the retirement decision. a defined benefit (db) pension and a defined contribution (dc) pension will have different effects on the nominal value of the “net” salary. we cannot directly observe anyone’s utility function. but to the extent that utility is positively correlated with “net” salary, we can reach some conclusions about how db and dc pensions affect a person’s decision to retire now or to work one more year. one major conclusion is that the longer a person has been on the job with a db pension, the more likely that person is to retire. another conclusion is that with a dc pension, the decision to retire will be extremely sensitive to the value of the pension account relative to the salary at the time of retirement, and the individual’s life expectancy. several studies (banker’s trust, 1980; woerheide & fortner, 1991; kahl & williamson, 1994) have shown that there may be more involved in a pension than just the monthly retirement check. a pension may include such things as survivor protection, disability protection, retiree medical benefits, early retirement incentive programs, inflation adjustments, social security offset, and death benefits. also, pensions include elements of risk (mcleod, moody & phillips, 1992/93). this paper focuses only on the cash pay outs to a 18 w. woerheide / financial services review 9 (2000) 17–31 retiree and omits the nonpecuniary and nonretirement income benefits, as well as the elements of risk in the different types of pensions. 2. the defined benefit case let us consider the case of an employee covered in a db plan. if this employee retires today, the present value of his or her future cash inflows related to employment is the present value of the pension payments. this can be described as i0 5 o t50 d s0 * n* c ~1 1 k!t (1) where i0 5 present value of pension benefits if a person retires today, s0 5 “final” salary, d 5 number of years a person expects to live, n 5 number of years a person has worked in the db plan, c 5 pension benefit as a percentage of “final” salary per year worked, and k 5 appropriate discount rate. the payment of the salary and the payment of pension benefits are assumed to be on an annual basis for simplicity of discussion. the use of monthly payments would have no impact on the results presented herein. pension funds vary in how the “final” salary is computed. common variations are that it is defined as the salary during the last year of employment, the average salary during the last three years of employment, and the average salary during the last five years of employment. in this paper, “final” salary and the salary during the prior year are treated as synonymous. note that the summation in eq. (1) starts at zero. this means, of course, that the first pension payment would be received immediately (i.e., the pension is an annuity due). by working one more year at a salary that has grown by the rate “g” and then retiring, today’s value of the worker’s income stream becomes the current year’s salary plus the present value of pension benefits received during his or her retirement. this is expressed as i1 5 s0 * ~1 1 g! 1 o t51 d s0 * ~1 1 g!* ~n 1 1!* c ~1 1 k!t (2) where i1 5 present value of the combination of salary and benefits if the employee retires one year from today, and g 5 percentage increase in “final” salary for next year. in eq. (2), the first s0 represents the salary in the most recent year, and the second s0 represents the salary number upon which the pension benefit is computed. as pointed out earlier, these two could be different, and thus the associated growth rates would also be 19w. woerheide / financial services review 9 (2000) 17–31 different. if the salary growth rate for the coming year is greater than the growth rate of the “final” salary, the net salary coefficient to be derived would be lower, and vice versa. the “net salary” or marginal economic benefit of working one more year is the difference between i1 and i0: i1 2 i0 5 s0 * ~1 1 g! 1 o t51 d s0 * ~1 1 g!* ~n 1 1!* c ~1 1 k!t 2 o t50 d s0 * n* c ~1 1 k!t (3) this difference can be “simplified” to the following form i0 2 i0 5 s0 * ~1 1 g!* f1 2 n* c 1 1 g 1 o t51 d c 1 c~c* n* g/~1 1 g!! ~1 1 k!t g 5 s0 * ~1 1 g!* d (4) whered 5 1 2 n* c 1 1 g 1 o t51 d c 1 ~c* n* g/~1 1 g!! ~1 1 k!t . the decision to retire depends on whether the value of eq. (4) is enough to offset the loss of utility achieved from substituting one more year of work for an extra year of leisure. as indicated earlier, there is no one single value ofd that will guarantee a person would retire or would opt to work another year. although the current salary (s0) and the salary increase (g) are important variables, the salary coefficient termd is the most important term in eq. (4). the termd could be negative, but the more likely scenario is that it would be positive. some employers place a cap on the number of years that can be counted in computing the db pension benefit. thus, although the salary used in computing the pension benefit increases, the employee may not receive the benefit of getting another year’s credit in computing the pension payment. this procedure produces a simpler and smaller value for the salary coefficient term computed above. if the coefficient termd were negative, then the economically rational person would presumably retire as he or she is actually paying to work. the coefficient would be negative in cases where the product of years covered in the pension plan (n) and the pension percentage per year worked (c) exceeds the value of 1.0 by an amount large enough to offset the present value of the increase in future pension benefits. the latter would be low whenever life expectancy is low. for example, ifn 5 30, c5 0.04, g5 0.03, k5 0.10, and d5 2, then the financial benefit to working one more year (i.e., i1 2 i0) is equal to -3.50% of next year’s salary. it should be noted that even in a situation as extreme as this, where c is incredibly large, if the person expects to live at least three more years, then the coefficient becomes positive and stays positive for increasing life expectancies. 2.1. estimates ofd let us now consider estimates of the salary coefficient term. table 1 showsd as a function of the number of years the employee expects to live (i.e., d) and the number of years on the job (i.e., n). note that because the salary and pension payments are assumed to be received 20 w. woerheide / financial services review 9 (2000) 17–31 table 1 estimates of the salary coefficient term (d) under a db plan variable value d c 0.01 g 0.04 k 0.08 n 5 10 20 30 0 0.952 0.904 0.808 0.712 1 0.963 0.917 0.824 0.731 2 0.973 0.929 0.839 0.750 3 0.983 0.940 0.853 0.767 4 0.991 0.950 0.866 0.783 5 1.000 0.959 0.878 0.798 6 1.007 0.968 0.889 0.811 7 1.014 0.976 0.900 0.824 8 1.020 0.983 0.909 0.835 9 1.026 0.990 0.918 0.846 10 1.032 0.997 0.926 0.856 11 1.037 1.003 0.934 0.865 12 1.042 1.008 0.941 0.874 13 1.046 1.013 0.948 0.882 14 1.050 1.018 0.954 0.889 15 1.054 1.022 0.959 0.896 16 1.057 1.026 0.964 0.902 17 1.061 1.030 0.969 0.908 18 1.064 1.034 0.974 0.913 19 1.066 1.037 0.978 0.918 20 1.069 1.040 0.981 0.923 21 1.071 1.043 0.985 0.927 22 1.074 1.045 0.988 0.931 23 1.076 1.047 0.991 0.935 24 1.077 1.050 0.994 0.938 25 1.079 1.052 0.997 0.941 26 1.081 1.054 0.999 0.944 27 1.082 1.055 1.001 0.947 28 1.084 1.057 1.003 0.950 29 1.085 1.058 1.005 0.952 30 1.086 1.060 1.007 0.954 the salary coefficient term under a db plan is defined as d 5 1 2 n* c 1 1 g 1 o t51 d c 1 ~c* n* g/~1 1 g!! ~1 1 k!t . where d 5 number of years a person expects to live, n 5 number of years a person has worked in the db plan, c 5 pension benefit as a percentage of ‘‘final’’ salary per year worked, k 5 appropriate discount rate, and g 5 percentage increase in ‘‘final’’ salary for next year. 21w. woerheide / financial services review 9 (2000) 17–31 on the first day of each year, the vertical axis, which represents the life expectancy of the employee, represents a person dying after receiving the payment for that year. thus, the value of d5 1 represents a person collecting his or her salary now, and the pension benefit one year from now, and then dying before receiving the second pension payment. table 1 is constructed using the assumptions that the employee expects a salary increase of four percent (g5 0.04), has a pension of one percent of final salary for each year worked (c 5 0.01), and believes that the appropriate discount rate is 8% (k5 0.08). for example, if an employee has participated in the pension plan for 20 years (n 5 20), and expects to live 10, 20, or 30 more years, then the salary coefficients for the coming year are 92.6%, 98.1%, and 100.7% of the projected salary. the projected salary equals s0 *(1 1 g). the most subjective of the variables used in the analysis is probably the discount rate. payments from a db pension are relatively safe, particularly if the pension fund is fully funded and is insured by the pbgc. therefore, payments from an insured db pension should have a discount rate of no more than the pretax cost of debt. the selection of 8% as the discount rate represents an approximation of this value. 2.2. sensitivity of the coefficient termd to test the sensitivity of the numbers presented in table 1 to the parameter values chosen, partial derivatives or first differences are calculated as appropriate. also, each parameter is adjusted up and down within reasonable ranges and the impact reviewed. the resulting estimates of the coefficient terms are then compared to the base example in table 1. the partial derivative of the coefficient term with respect to the salary growth rate (g) is positive. however, changing the growth rate to 3% and 5% has negligible impact on the coefficient numbers produced. the partial derivative with respect to the discount rate (k) is negative. using 6% and 10% as the discount rate had negligible impact when life expectancy was at the lower end of its range. at the upper end (e.g., d5 30) the results were moderately sensitive to the discount rate used. a much more interesting calculation is the partial derivative of the coefficient term with respect to the pension benefit per year worked (c). the sign of the derivative is ambiguous as it depends on how long one expects to live after retirement. if one expects a “short” life, then the derivative is negative. thus, the larger the pension benefit per year worked, the lower the salary coefficient term becomes and the more likely the employee is to retire. if one expects a “long” life, then the larger the pension benefit per year worked the larger the coefficient term and the more likely the employee is to work another year. the point at which life expectancy (d) changes from “short” to “long” is a function of k, g, and n. although the derivative is relatively insensitive to differences in k and g over relevant ranges, it is quite sensitive to n. as n becomes shorter, the changeover value of n becomes shorter. for example, if k5 0.10, g5 0.05, andn 5 10, then the change from a positive to a negative derivative occurs at d5 11. if k 5 0.10, g5 0.05, andn 5 20, then the derivative changes from positive to negative at d5 39. thus, the longer one has worked at a job, the more likely the derivative is to be positive and hence higher values of c would make a person more likely to continue working. the briefer one has worked at a job, the more likely the derivative is 22 w. woerheide / financial services review 9 (2000) 17–31 to be negative and hence higher values of c would actually make a person more likely to retire. the size of the coefficient term for values of c other than 0.01 are easily computed as d 5 1 2 p* ~1 2 tv! (5) where p 5 a multiple of 0.01, and tv5 the number in table 1 that represents the combination of the number of years one has worked and the number of years one expects to live. thus, if one wanted to know the salary coefficient term for c5 0.02 whenn 5 30 and d5 20, then it equals 1 2 x (1 -0.923), or 0.846. it is clear from eq. (5) that when the table 1 value is near 1.0, the changes in the value of c will have minimal effect. but, where the table 1 value is substantially different than one, differences in c will have a substantial effect on the salary coefficient. the partial derivative of the coefficient term with respect to n is negative provided one assumes g, k. virtually all actuarial work in the area of pensions assumes that prospective salary growth rates are less than the expected return on the invested assets of the pension fund (wyatt company, 1981). thus, it is also likely that the salary growth rate would be less than the discount rate. the negative effect of n can also be verified by looking across any row in table 1 and noting that the coefficient term gets smaller as one moves to the right (i.e., as n gets larger). we cannot take the partial derivative of the coefficient term with respect to life expectancy as life expectancy defines the number of terms in the summation. however, we can take the first difference between the coefficient terms when the life expectancy is d11 years and when it is d years. this difference is always positive. this is verified by looking down any column in table 1. thus, the longer one expects to live, the larger the salary coefficient term and the less likely one is to retire because of the greater number of increased pension payments that will be received. note that life expectancy has a greater impact when n is large (e.g., 30 years), than when it is small (e.g., 5 years). in other words, people who physically feel like they could continue working would have the most fiscal incentive to do so. 3. the defined contribution case let us now consider the case of an employee covered by a dc plan. when this employee retires, he or she has two choices with regard to the pension account. one is to roll it over into an ira account, and the other is to annuitize it. if the roll over option is elected, then the employee may have a number of choices depending on his or her age. we will not consider these options as they take us well beyond the scope of this paper. if the employee opts to annuitize, then the assets in the account are used to purchase a life annuity with an insurance company. the portfolio manager of the assets in which the pension account money is invested and the insurance company may be part of the same company. when the employee is ready to retire, the dc pension account may be annuitized with the same company, or the employee may roll the account to another insurance company that 23w. woerheide / financial services review 9 (2000) 17–31 promises a higher monthly payment. a valid comparison of the effects of the pension plan type upon the retirement decision requires that we assume that the employee annuitizes his or her account upon retirement. if the employee retires today, the present value of pension payments is i0 5 o t50 d v/afq, j ~1 1 k!t 5 o t50 d m* s0/afq, j ~1 1 k!t (6) where v 5 the accrued value of the dc pension account, afq,j 5 annuity due factor used by the pension provider, based on the provider’s expectations the person will live q more years and that the provider uses a discount rate of j, and m5 the ratio of v to s0. the actual calculation used for the payment is more complicated than simply the present value interest factor of an annuity due based on the life expectancy of the individual. (see, e.g., chapter 9, cissell, cissell & flaspohler, 1990.) however, the actual calculation is immaterial to our results here. as in the db case, note that the pension payments start at time zero. if this employee decides to work one more year, then the present value of his or her income stream is the current year’s salary plus the present value of subsequent pension benefits received. this is expressed mathematically as i1 5 s0 * ~1 1 g! 1 o t51 d @m* s0 * ~1 1 i ! 1 s0 * ~1 1 g!* ~1 1 i !pp#/afq21,j ~1 1 k!t (7) where p 5 percentage of salary contributed to the pension account, and i 5 rate of return the employee expects to earn on the pension account. it is assumed that the pension contribution is made at the beginning of each year when the employee is paid. note that eqs. (6) and (7) incorporate two potentially different dates of death. the first, d, represents the number of additional years the employee expects to live. the second, q, represents the employee’s life expectancy based on mortality tables used by the provider of the annuity. the size of the annual pension payment is determined in part by q. the number of payments the employee expects to receive is defined exclusively by d. most people believe they will outlive the life expectancy defined by the mortality at the time of retirement. eqs. (6) and (7) similarly allow the rate of return expected on the pension account assets (i) to be different than the rate of return used in computing the pension payments (j). both of these rates may be different than the discount rate used by the employee in determining the present value of the future pension payments (k). the most crucial relationship among these variables is that j will be less than k. if the value of j is less than k, then the annuitization process represents a negative net present value decision. 24 w. woerheide / financial services review 9 (2000) 17–31 there are two reasons to believe that j is less than k. first, the pension provider incurs expenses. it costs money to run a portfolio, to send out statements, to send out monthly checks, and so forth. if j were equal to k, then there is no profit margin for the pension provider. thus, if j were equal to k, no company would ever sell an annuity. the second reason is that once a person annuitizes his or her dc pension account, the pension provider takes on a risk exposure that a retiree will live longer than expected. it is true that the pension provider benefits if a retiree dies prematurely. however, the pension provider must place a premium on the risk that the majority of retirees drawing pensions live longer than expected. the premium would be that the discount rate j used to determine the monthly benefit is less than the market rate k that the pension provider expects to earn on the pension assets. thus, because of its operating expenses and mortality risk, the discount rate j must be less than k. as noted earlier, if j is less than k, then the decision to annuitize represents a negative net present value. the value of this negative net present value is the price the individual pays the pension provider for the management of the account and for taking the risk the employee may live beyond his or her life expectancy at the time of retirement. it was also noted above that i (the rate of return earned on the dc account assets) could be different than j and k. however, it is likely the case that the expected value of i would be relatively close to k (the difference equaling the cost of managing the pension account), and in the following empirical estimates it is assumed these are equal. the “net salary” or marginal economic benefit to working one more year is, as in the db case, the difference between i1 and i0: i 2 i0 5 s0 * ~1 1 g! 1 o t51 d @m* s0 * ~1 1 i ! 1 s0~1 1 g!* ~1 1 i !* p#/afq21,j ~1 1 k!t 2 o t50 d m* s0/afq, j ~1 1 k!t (8) this difference can be “simplified” to: i1 2 i0 5 s0 * ~1 1 g!* f1 1 o t51 d @m* ~1 1 i !/~1 1 g! 1 p* ~1 1 i !#/afq21,j ~1 1 k!t 2 o t50 d ~m/~1 1 g!!/afq, j ~1 1 k!t 5 s0 * ~1 1 g!* d9 (9) where d9 5 1 1 o t51 d @m* ~1 1 i !/~1 1 g! 1 p* ~1 1 i !#/afq21,j ~1 1 k!t 2 o t50 d ~m/~1 1 g!!/afq, j ~1 1 k!t 25w. woerheide / financial services review 9 (2000) 17–31 note that eqs. (4) and (9) are identical except for the definition of the salary coefficient term. 3.1. the “neutrality” case eq. (9) looks complex because we have allowed the values of i, j, and k, and the values of d and q, to differ. if the pension manager and the employee share the same life expectancy (i.e., d5 q), and all three interest rates are the same (i.e., j5 k 5 i), then eq. (8) reduces to i1 2 i0 5 s0 * ~1 1 g! 1 m* s0* i 1 s0 * ~1 1 g!* p or, more simply 5 s0 * ~1 1 g!* f1 1 m* i 1 1 g 1 pg 5 s0 * ~1 1 g!* d0 (10) whered0511 m*i 11g 1p. this derivation requires the assumption that working one more year does not alter the expected date of death. hence, the annuity factor used by the pension provider for a person retiring today (afq,j) differs from the annuity factor that would be used one year later (afq-1,j) by a factor of exactly one year. in eq. (10),d0 is the salary coefficient term. it is easy to see that this term will always be greater than one if either m and i are positive and p is nonnegative, or p is positive and m and i have the same sign or at least one of them is zero. thus, when these two sets of parameters are equal, one’s effective salary for the coming year (as stated in eq. (10) will always be greater than the stated salary of s0 * (1 1 g). however, because it is unrealistic to assume that j and k are equal, we will not analyze this case any further. 3.2. empirical estimates ofd9 in a dc plan table 2 provides examples of values of the net salary coefficient term for a dc pension account. in all cases, we are considering an employee with a discount rate of 8% (k5 0.08), whose contribution to a pension fund equals 10% of current salary (p 5 .10), and whose salary will grow by 4% (g5 0.04) if the employee opts to work one more year. the employee believes the pension account will earn 8% (i5 0.08). the discount rate and the life expectancy used in determining the pension annuity are 7% (j5 0.07) and 17 years (q5 17). the unisex life expectancy for a 65-year old person is approximately 17 years. the net salary coefficients are shown in table 2 for values of m equal to 0.5, 4, 8, and 15. the rationale for these values is provided in the next section. as an example, if the employee expects to die in ten years (contrary to the pension fund’s expectation), and his or her accrued pension assets equal eight times his or her salary (m5 8), then the salary coefficient (d9) in 26 w. woerheide / financial services review 9 (2000) 17–31 table 2 estimates of the salary coefficient term under a dc plan variable value d k 0.8 j .06 i 0.08 p 0.1 g .04 q 17 m .5 4 8 15 1 0.954 0.632 0.264 20.381 2 0.969 0.681 0.353 20.222 3 0.983 0.727 0.435 20.076 4 0.995 0.770 0.512 0.060 5 1.007 0.809 0.582 0.186 6 1.018 0.845 0.648 0.303 7 1.028 0.879 0.709 0.410 8 1.037 0.910 0.765 0.510 9 1.046 0.939 0.817 0.603 10 1.054 0.966 0.865 0.688 11 1.062 0.991 0.909 0.767 12 1.068 1.014 0.951 0.841 13 1.075 1.035 0.989 0.909 14 1.081 1.054 1.024 0.972 15 1.086 1.073 1.057 1.030 16 1.091 1.089 1.087 1.084 17 1.096 1.105 1.116 1.134 18 1.100 1.120 1.142 1.180 19 1.104 1.133 1.166 1.223 20 1.108 1.145 1.188 1.263 21 1.111 1.157 1.209 1.299 22 1.115 1.167 1.228 1.333 23 1.118 1.177 1.245 1.365 24 1.210 1.186 1.262 1.394 25 1.123 1.195 1.277 1.421 26 1.125 1.203 1.291 1.446 27 1.127 1.210 1.304 1.469 28 1.129 1.216 1.316 1.490 29 1.131 1.223 1.327 1.510 30 1.133 1.228 1.338 1.529 the salary coefficient term under a dc plan is defined as d9 5 1 1 o t51 d @m* ~1 1 i !/~1 1 g! 1 p* ~1 1 i !#/afq–1,j ~1 1 k!t 2 o t50 d ~m/~1 1 g!!/afq, j ~1 1 k!t where g 5 salary growth rate, p 5 percentage of salary contributed to the pension account, i 5 rate of return the employee expects to earn on the pension account, v 5 the accrued value of the dc pension account, afq,j 5 annuity due factor used by the provider, based on the expectation the person will live q more years and that the fund uses a discount rate of j, and m 5 the ratio of the value of the dc account to the ‘‘final’’ salary. 27w. woerheide / financial services review 9 (2000) 17–31 eq. (9) is 0.865. that is, the employee is working for only 86.5% of his or her salary during that last year. there are several noteworthy points about table 2. one is how sensitive the salary coefficients are to the value of m. when m5 0.5, the coefficents range from 0.954 to 1.133. when m5 15, they range from20.381 to 1.529. this says that when m is “small,” the net salary for working one more year is about equal to the stated salary. but when m is “large,” there can be substantial differences between the net salary and the stated salary. another point is that when m5 15 and the employee expects to live less than four years (d5 1, 2, or 3), the salary coefficient term is actually negative. this means the person is paying to work. hence, when m is large and an employee is in ill health, it not only makes physical sense for an employee to retire, it also makes fiscal sense to retire. 3.3. estimates of plausible values of m as indicated above, a key variable in estimating the salary coefficient term is the ratio of the pension account at the time of retirement to the salary at the time of retirement (i.e., m). it turns out that the value of m can be defined by an iterative formula. consider the case of a person who works one year and then retires. if we continue our use of treating all payments as being made at the start of each year, then the ratio of the pension account at the end of the first year to the salary at the start of the first year (m1) is mi 5 v1 s0 5 p* s0 * ~1 1 i ! s0 5 p* ~1 1 i !. (11) the value of m at the end of the second year (m2) is m2 5 v2 s1 5 p* s0 * ~1 1 g!* ~1 1 i ! 1 s0 * m1 * ~1 1 i ! s0 * ~1 1 g! (12) 5 p* ~1 1 i ! 1 1 1 i 1 1 g * m1 for the n-th period, the value of mn is then defined as mn 5 vn sn21 5 p* sn21 * ~1 1 g!* ~1 1 i ! 1 sn22 * mn21 * ~1 1 i ! sn212 * ~1 1 g! (13) 5 p* ~1 1 i ! 1 1 1 i 1 1 g * mn21. note that if i 5 g, then mn 5 n*p*(11i). to ascertain some plausible values for mn, let us start with what might be considered a maximum case scenario. let us consider an employee who starts contributing to a dc pension at age 20. this employee works for 45 years with the same company (n 5 45), has a salary contribution rate of 10% (p 5 .10), a salary growth rate of 5% (g5 0.05), and his or her investments earn a rate of return of 8% (i5 0.08). in this situation, m45 5 14.47. at 28 w. woerheide / financial services review 9 (2000) 17–31 the other extreme, let us consider the case of an employee who does not start contributing to the pension on his or her current job until age 55. this employee works for 10 years (n 5 10), has a salary growth rate of 3% (g5 0.03), and has a salary contribution rate of 5% (p 5 .05). in this case, m10 5 0.674. in light of the extent to which some people have multiple job changes over their working lives, do not always participate in dc pensions, and do not always even participate in pension programs, this scenario may apply in some cases. the implication from these two extreme examples is that realistic values of m may range from as low as 0.5 to as high as 15. 3.4. sensitivity of the coefficient termd9 as in the db case, we can analyze the sensitivity of the coefficient term by taking partial derivatives of the coefficient term with respect to the various inputs, as well as considering actual variations in the values. the derivative of the coefficient term with respect to the salary growth rate is ambiguous. however, the impact of this parameter on the size of the term is trivial as alternative values of 3% and 5% produced negligible changes in the coefficients. the partial derivatives of the coefficient term with respect to the percentage of salary contributed by the employer (p) and the investment yield (i) are both positive. when values of 8% and 12% are used for p, there are again negligible changes in the coefficients. however, when values of 6 and 10% are substituted for the investment yield, there are substantial changes in the coefficient terms as m becomes larger. for example, if the employee expects to live 15 years, then when m5 0.5 the coefficient changes from 1.086 to 1.096 when i is increased from 8 to 10%. however, when m5 15, the coefficient term increases from 1.030 to 1.267. the partial derivative with respect to the discount rate is negative. but just as with the investment yield on the pension account, it is more critical when m is large, particularly when an employee expects to live a long time. for example, if the employee expects to live 30 years, then when the discount rate is increased from 8 to 10%, the coefficient falls from 1.133 to 1.104 when m5 0.5. alternatively, this same increase in the discount rate causes the coefficient to fall from 1.529 to 1.223 when m5 15. a more interesting case is the derivative of the coefficient term with respect to the pension fund account relative to current salary (m). the sign of the derivative is ambiguous. if one expects a “short” life, then the larger the accumulated pension account the lower the coefficient term. conversely, if one expects a “long” life, then the larger the accumulated pension account the larger the coefficient term. this is noted by looking across the rows on table 2. in rows 1 through 16, the coefficient term becomessmalleras m increases. in rows 17 and above, the coefficient term becomeslarger as m increases. variations in the discount rate (j) and in the life expectancy (q) used to determine the annuity have relatively little impact, except, again, when the value of the pension account is large relative to the final salary. 29w. woerheide / financial services review 9 (2000) 17–31 4. summary, conclusions, and extensions 4.1. summary and conclusions this paper focuses on a question many people will face in their lifetime, namely, should they retire now or work one more year and redecide. the decision requires comparing the utility of one year’s worth of net salary to the utility of one year’s worth of leisure. the utility function incorporates the wealth and substitution effects created by all the other assets and income the employee has, as well as many other variables such as health, hobbies, and so forth. we cannot examine utility functions, so we analyzed the percentage of next year’s salary one would actually earn by working under both a db and a dc pension. it is assumed that the higher this percentage is, the more likely a person is to work another year. several conclusions apply regardless of the type of pension plan in which one participates. first, the longer one expects to live after retirement, the higher the net salary from working one more year. second, the use of higher discount rates by an employee always reduces the net salary. third, the growth rate in salary for the coming year has a minimal effect on the salary coefficient term. thus, if you are contemplating retirement and your boss offers to change your salary increase for the next year from 3% to 4%, this is unlikely to alter your decision. some components in the decision to retire are unique to the type of pension plan one has. in the case of db pension plans, the longer one is covered by the plan, the lower the net salary for working another year. in the case of dc pension plans, increases in the value of the pension account relative to a worker’s current salary reduce the salary coefficient term if the worker expects to have a “short” life, and increase it if the worker expects to have a “long” life. as shown in table 1, the coefficient term under a db pension plan is relatively stable. for people who have been in a job for at least 20 years, the coefficient terms are less than one, but not dramatically so. the lower the coefficient term the more likely a worker is to retire because the opportunity cost of not working becomes lower. thus, we may infer that db pension plans encourage people who have been on the job for a long time to retire. the numbers in table 2 cover a larger range and are more likely to be larger than one. because a higher salary coefficient term encourages a person to continue working, we may infer that dc pension plans are more likely to provide this encouragement. 4.2. extensions in order to keep the analysis manageable and to keep the focus strictly on the impact of the pension program, many simplifications were made. first, all salary numbers were on a pretax basis. if a person works one more year, the increase in his or her wealth is actually the after-tax value of the salary plus the present value of the after-tax increase in pension benefits, less the after-tax value of what the first year’s pension payment would have been. the role of taxes needs to be examined. second, it was assumed that when the dc pension was annuitized that the annuity was a pure annuity for the employee with no survivor’s benefits and no guaranteed number of 30 w. woerheide / financial services review 9 (2000) 17–31 payments. the introduction of either survivor’s benefits or a guaranteed number of payments would likely have a significant effect on the numbers provided in table 2. third, social security was omitted. working one more year (or at least not starting social security withdrawals) has the benefit of increasing future social security payments. depending on how the internal rate of return on this deferral compares with the discount rate, the salary coefficient term could be enhanced or reduced. one effort has been made at incorporating social security into the decision to retire (newmark & walden, 1995), but that presentation is not consistent with the approach offered here. finally, the impact of working one more year on all other financial assets, such as iras, 401(k)s, and 403(b)s, was ignored. a more general treatment of the retirement decision would require incorporating these assets. acknowledgments the author gratefully acknowledges the substantive input of the two anonymous referees as well as the various discussants who provided input on presentations of earlier versions at various meetings. references banker’s trust company. (1980).corporate pension plan study: a guide for the 1980’s. cissell, r., cissell, h., & flaspohler, d. (1990).mathematics of finance(8th ed.). boston: houghton mifflin company. dexter, m. (1984).replacement ratios: a major issue in employee pension systems.washington, d.c.: public employees pension systems. kahl, d., & williamson, j. (1994).an analysis of trends in state retirement plans: the move from defined benefit to defined contribution plans.presented at the annual meeting of the academy of financial services, st. louis, mo. mcleod, r., moody, s., & phillips, a. (1992/93). the risks of pension plans: a review.financial services review,2, 131–156. modigliani, f.& brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section date. in k. kurihara (ed.),post keynesian economics(pp. 388 36). new brunswick: rutgers university press. newmark, c., & walden, m. (1995). should you retire at age 62 or at 65?financial counseling and planning, 6, 35–44. palmer, b. (1989). tax reform and retirement income replacement ratios.the journal of risk and insurance, 56, 702–725. wiatrowski, w. j. (1993). factors affecting retirement income.monthly labor review,march 1993, 25–35. woerheide, w., & fortner, r. (1991). comparing defined benefit and defined contribution pensions.journal of financial planning, 4,42–49. wyatt company. (1981).the 1981 survey of actuarial assumptions and funding.washington, d.c.: the wyatt company information center. 31w. woerheide / financial services review 9 (2000) 17–31 pii: s1057-0810(99)00042-6 an analysis of affinity programs: the case of real estate brokerage participation danielle lewisa, randy i. andersonb,* ,1, leonard v. zumpanoc adepartment of economics, college of business administration, southeastern louisiana university, hammond, la 70402, usa bschool of business, samford university, 800 lakeshore drive, birmingham, al 35229, usa calabama real estate research and education center, the university of alabama, 149 bidgood hall, tuscaloosa, al 35486, usa received 5 january 1999; received in revised form 7 july 1999; accepted 18 october 1999 abstract in this study, we examine the impact of affinity programs on the residential real estate brokerage market. the results indicate that affinity-participating firms employ more salespeople, operate more offices, are more likely to be franchised, and have more multiple listings service affiliations than their nonparticipating counterparts. we directly test for firm and industry efficiency using a bayesian stochastic frontier technique, and find strong evidence that non-affinity firms are much more efficient at allocating and utilizing their resources. these findings cast concerns on the industry in light of the growth of affinity programs. © 2000 elsevier science inc. all rights reserved. jel classification:l85; c11 keywords:real estate brokerage; affinity programs; efficiency * corresponding author. tel.:11-205-726-2128; fax:11-205-726-2464. e-mail address:rianders@samford.edu (r.i. anderson) 1 the work on this paper was started while the author was at northeast louisiana university. financial services review 8 (1999) 183–197 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00042-6 1. introduction when buying or selling real estate it is common to rely on the services of a residential real estate brokerage firm. in simplest terms, brokerage firms bring together buyers and sellers to conduct real estate transactions in return for a commission. for the individual real estate consumer, brokerage firm efficiency is crucial. if the brokerage industry is efficient, buyers and sellers benefit from reduced search time and low transactions costs. alternatively, economic theory predicts higher prices and lower quality of services in inefficient markets. regulators, policy makers, practitioners, and academics continue to spend considerable amounts of time monitoring and analyzing this industry because it impacts so many people. however, evaluating the efficiency of the residential real estate market is difficult because it is changing and evolving so rapidly. we are seeing mass consolidation, leaving fewer yet larger firms in the industry (zumpano, elder, and anderson, 2000). the structure of brokerage relationships is also changing. traditionally, the agent represented the seller in real estate transactions. today, other choices such as a buyer’s agent (regular and exclusive) and a disclosed dual agent exist and are growing in popularity. an exclusive buyer’s agent works for an office that does not take listings of any kind and represents only buyers and non-agency facilitator arrangements. a regular buyer’s agent works in a traditional real estate office that takes listings, but will work with a buyer under contract. disclosed dual agents represent both the buyer and the seller in real estate transactions and can arise in numerous settings. in addition to traditional multiple listing service (mls) use, the growth in technology, especially the internet, is changing the way people shop for homes and the way real estate firms conduct business. these changes are most certainly impacting the efficiency in which the market and industry operates (anderson, fok, zumpano, and elder, 1998). in this paper, we examine another change in the industry, the use of affinity programs, and attempt to ascertain how these programs will alter the competitive structure of the market. affinity relationships provide commission rebates, discounts, and other goods and services to individuals who are members of professional organizations, trade associations, unions, or organizations who have an agreement with a real estate company. the increase in affinity programs over the past few years and the related increase in referral fees are a major concern of the residential real estate brokerage industry (nar, 1996; dezube, 1996). given the increased involvement of the residential brokerage industry with affinity groups and the billions of dollars potentiality at stake, it is not an exaggeration to say that the affinity program issue may become one of the most important developments facing this industry. in this study, we present the first rigorous empirical investigation of affinity programs on firm and market efficiency using detailed financial data from a national sample of residential real estate brokerage firms. our analysis into the affinity question proceeds with further discussion on affinity programs within the real estate brokerage industry and a summary of the prior affinity research. section 3 provides the data that we use along with an analysis of the summary statistics for affinity and non-affinity firms. section 4 presents the bayesian stochastic frontier methodology that we use to estimate the efficiency levels by brokerage type, whereas section 5 presents the efficiency and scale results. section 6 concludes. 184 d. lewis et al. / financial services review 8 (1999) 183–197 2. affinity programs: background and prior research real estate firms have been in the business of relocation and referral-related residential sales for the past thirty years; however, it is the recent increase in referral fees that is of concern to the residential real estate brokerage industry (nar, 1996). thirty years ago the use of referral fees was initiated by brokers who desired to both send and receive prospective customers. for those brokers who received the prospective customer, the referral fee was considered a marketing cost. in fact, ball (1990) discusses the need to make the most of your referral network and business. at the time referral fees came into practice, it was rare for transferees to receive assistance from their employer in purchasing a new home or selling their existing residence. during the recession of 1981 through 1983, corporations began offering better relocation benefits to their employees. also during this time, the use of referral networks extended beyond independently owned, non-franchised firms as referral networks were also established among franchise firms (nar, 1996). the use of referral fees has continued to evolve. now, many corporations and relocation management companies demand such fees. in the past few years, corporations have begun charging referral fees as a means to reduce the skyrocketing expenses associated with relocation. also aware of rising relocation costs, relocation management companies have begun to charge referral fees in an effort to maintain or improve their profits. affinity relationships grew out of these relocation and referral networks. affinity groups are relatively new players in the real estate industry. in 1995, at the business issues committee of the national association of realtors© (nar), a working group was established to examine industry practices with respect to affinity relationships. (nar, 1996). companies that have established affinity relationships for their employees include the following: united air lines, sears, roebuck and company, united parcel services, and the prudential insurance co. of america. among the real estate firms involved in such partnerships are century 21, prudential, and coldwell banker. the structure of affinity benefits has varied according to the structure of each affinity program and according to state law. 2.1. potential benefits of affinity participation advocates of affinity programs argue that both consumers and brokers benefit. consumers benefit from reduced brokerage commission fees, and the cost savings on ancillary services associated with loan and property closings whereas brokers reap more business from affinity referrals. economic theory suggests that bundling real estate services, as could be the case with one-stop shopping arrangements, can result in significant cost savings to consumers from reduced search costs and economies of scale in the provision of bundled services. reich–hale (1999) show how this type of bundling of goods and products can be beneficial in the insurance industry. supporters of such programs also contend that they will tend to reduce real estate commissions for all consumers, whether or not they are members of affinity groups, by increasing overall price competition within the industry. advocates claim that small firms should benefit more than larger firms should from the affinity groups because larger firms can often 185d. lewis et al. / financial services review 8 (1999) 183–197 provide ancillary services to consumers cost effectively in-house. small real estate firms that could not otherwise provide competitive services to their customers can now do so by joining with an affinity organization. cendant mobility, the largest of the affinity organizations estimates that their company has paid out more than $20 million in rebates to more than 40,000 customers since 1990 (agency law quarterly, 1998). 2.2. possible limitations of affinity participation opponents of such programs counter that these programs may violate the real estate settlement procedures act’s (respa’s) prohibition against rebates, result in double referral fees, and could raise the cost of doing business for real estate professionals and consumers. the consumer law center and the consumer federation of americans, among others, claim that the money rebated to consumers is very small relative to the cost of affinity programs. participating real estate agents often have to pay affinity companies as much as 30% of their gross commissions in referral fees, very little of which ultimately passes to the consumer (agency law quarterly, 1998). imperfections in the market may occur due to imperfect information. cost reductions resulting from bundled services, rebates, and other goods and services provided by affinity partners may not always be realized. such savings presuppose that affinity service providers are less expensive than non-affinity competitors. for example, if a consumer receives a $300 rebate from an affinity broker but ends up paying $800 more for a mortgage through an affinity lender, the consumer is obviously not better off. in such cases sales agents, especially if working as a buyer’s agents, could violate their fiduciary responsibility to their client if they knowingly send clients to higher priced service providers. however, brokers could be threatened with the loss of future referral by informing customers that their affinity affiliates are more expensive. such coercion could create serious conflict of interest problems for brokers. for affinity programs to truly benefit the consumer, there must be an incentive structure in place that aligns the best interest of the consumer with the self-interest of all participating service providers, including the broker. another complaint against affinity groups is that much of their activities are outside the scope of state regulators and real estate commissions. in fact, the financial regulatory relief and economics efficiency act now before congress would, if passed, prevent states from prohibiting affinity rebates. at the other end of the spectrum, a number of states, including maryland, new jersey, and mississippi, either prohibit affinity rebates or consider them to be illegal inducements to consumers. limiting consumers to certain brokers, or alternatively, barring affinity rebates can present some freedom of choice problems, whereas over-riding state regulatory authorities can create some states rights issues. as the previous discussion indicates, affinity programs presently exist with a certain degree of controversy. however, as noted above, there has been little empirical evidence available to assess the impact of affinity programs on the real estate brokerage industry. in the next section we provide a summary the previous research on the affinity issue. 186 d. lewis et al. / financial services review 8 (1999) 183–197 2.3. prior empirical research surveys in 1996 and 1998 by the nar questioned brokers as to their perceptions about the impact of affinity programs on their business. the 1996 survey contains 722 useable responses (;7% response rate) from designated realtors© across the nation, using a random sampling procedure. the nar deemed the survey representative of the membership in all four-census regions, and hence conclusions are drawn for the entire population. the survey contains information from responses based on the previous year, 1995. a follow-up 1998 survey contains 635 useable responses (;6% response rate). this follow-up survey is not completely random. in addition to a random sample of the membership, the survey sampled 500 of the leading real estate firms as defined and identified by realtrends. again the responses are based on the previous business year, 1997. it is important to emphasize that these survey responses do not contain any financial data and are based solely on the opinions of respondents. moreover, even though the surveys were deemed reliable by the nar, the response rate is very small, and any conclusions should be interpreted with caution. the results indicate that affinity participation is increasing, but overall participation remains modest. in particular, the percentage of respondent firms that affiliate with affinity programs has risen from 16% in 1995 to 24% in 1997. mcmillan (1997) provides additional support for these results as he suggests that affinity programs are rapidly growing on the east coast and in california. however, he notes that in some locations the programs are just not catching on, which brings up issues on what settings or circumstances promote affinity affiliation. the results also indicate that once brokerage firms begin an affinity relationship, they usually participate in more than one program. in 1995 approximately 63% of affinity participating firms had more than one affinity partner. that number jumps to 79% of firms having multiple relationships in 1997. the surveys also reveal that brokerages primarily partner with corporations, professional associations, employers, and unions. the nar reports also note that affinity participation is greatest among the larger real estate companies and franchise firms. when examining the impact on firm performance, the survey results reveal that most respondents either did not know the impact of affinity programs on the various measures of performance, or that they thought that affinity programs did not impact performance. only 9% report an increase in agent productivity, and only 10% reported an increase in their firm’s listings. we must point out, however, that these responses were not based upon financial or accounting data, but rather, represented the opinions of respondents. hence, these findings must be interpreted with some caution. respondents may have had a difficult time isolating the effects of affinity participation on firm performance. the potential positive and negative impact of affinity on the industry and on profitability is further highlighted in berger (1997). the surveys also examine potential consumer benefits of affinity programs. the benefits most often take the form of commission rebates, special mortgage financing packages, free goods, and product discounts. sellers seem to be the principle beneficiaries of commission reductions, whereas buyers most often received financing inducements. in 1997, 76% of the affinity respondents indicated that they offered sellers commission reductions in the form of either a percentage discount (56%) or a fixed dollar discount (20%). some of these firms 187d. lewis et al. / financial services review 8 (1999) 183–197 offered commission rebates, although this was not as common as discounts. buyers were also offered commission reductions, however, offers of special services or goods were far more common. in general, the percentage of respondents offering benefits has increased between 1995 and 1997. 3. financial data and analysis as stated previously, to determine the actual impact on firm performance, we need to link affinity affiliation with actual brokerage financial figures and analyze those numbers to determine the implications of affinity relationships on performance. the financial data comes from another survey by the nar (nar, 1998). this survey contains 1996 income and expenses data. professionals who are certified real estate brokerage manager designees make up half of the sample with the remainder of the responses coming from a random sample of real estate brokerage firms that are members of the nar. the information includes the number of real estate listings and sales by each firm, net income, and the firm’s cost of listing and selling residential real estate. additionally, for the first time, the survey contains a variable indicating whether the firm participates in affinity programs and the corresponding costs associated with participation. following other studies (zumpano, elder, and crellin, 1993; zumpano and elder, 1994; anderson, fok, zumpano, and elder, 1998; anderson, lewis, and zumpano, 2000a; b; and lewis and anderson, 1999) that use a similar data source we include only firms who obtain at least 75% of their revenues from residential transactions. the final data set is made up of 176 firms, 92 of which are affinity participants and 85 nonaffiliated firms. the nar again denotes the sample as reliable and representative of the true population of real estate brokerage firms even though the overall survey has a low response rate at just over 3%. to the best of our knowledge, this sample does represent the only national data set that provides income and expense figures for real estate brokerage firms, thus providing us with the best opportunity to glean insights into the affinity issue on a nationwide basis. nevertheless, with the relatively small sample size and response rate caution must be exercised when examining and interpreting the results. we deem the results that follow as reliable for the sample set, but defer any generalizations for the population of all brokerage firms for future work with a larger data set. in table 1 (panel a) we examine the firm characteristics of both sample group types. affinity participants are larger as measured by the number of employees. affinity firms employ on average 79 salespersons and almost 13 non-sales persons. in contrast, the non-affinity firms in the 1996 sample employ only 22 salespersons and had just fewer than three and a half non-sales employees. affinity affiliates also tend to have a larger number of offices than their non-affinity counterparts. affinity firms operated, on average, four more offices than did non-affinity brokers. as with the prior surveys, we find that affinity firms are older (only significantly older at the 0.10 level), more likely to be associated with a franchise and have slightly more mls memberships. interestingly, the salespersons in non-affinity firms produce more revenue transactions on a full-time equivalent salesperson basis, indicating more productive employees. however, the difference is not statistically robust. in table 1(panel b), we examine several key financial variables. noting the size differ188 d. lewis et al. / financial services review 8 (1999) 183–197 ence in firms with respect to offices and employees, we expect and do find higher levels of listings, sales, total revenue transactions, gross revenues, and total adjusted profits for affinity firms relative to non-affinity firms. total adjusted profits are simply the firm’s net profits plus the distributions taken by the owners. however, when controlling for size effects by examining revenues and profits on a per transaction basis, differences in performance between the two groups disappear. in fact, profits per revenue transaction are actually greater for non-affinity firms than their affinity counterparts, although the difference is not statistically significant. finally, no statistical difference exists between the percentage of firms in 1996 that experience an increase in profitability. although not a test of firm and/or market operating efficiency, it seems as if the larger sample affinity firms are able to generate large output numbers, but unable to translate those earnings into higher profits. hence, for this sample set, affinity programs may help firms grow in size and perhaps gain market share, but may actually be reducing the efficiency in which firms conduct business. we highlight this point by noting that the agents for non-affinity firms actually produce more revenue transactions than their affinity salesperson counterparts. to formally test for x-efficiency difference in the two groups, we need to understand and define the firm as a production unit. in general, real estate brokerage firms produce revenue transactions (listings and sales). to be x-efficient, the firms must choose the optimal table 1 firm characteristics and financial performance panel a: the summary of firm characteristics: 1996 nar income and expense survey characteristic affinities non-affinities t-statistic number of sales people 79.13 22.24 3.37* number on nonsales people 12.70 3.44 3.16* number of offices 5.32 1.45 2.52* age of the firm 21.24 17.36 1.67 percent of firms franchised 39% 24% 2.28* number of mls subscriptions 2.52 2.17 3.38* revenue transaction per salesperson per year 17.88 22.39 21.85 panel b: average financial performance of affinity affiliates compared to non-affinity participants affinities nonaffinities t-statistic number of listings 544 212 2.75* number of sales 563 214 2.61* total revenue transactions 1107 426 2.60* gross revenues $4,107,672 $1,270,619 3.12* adjusted net income $245,617 $81,320 2.60* gross revenues/revenue transaction $3,808 $3,573 0.52 adjusted net income/revenue transaction $260 $360 20.36 percentage of firms experiencing an increase in profitability 49% 41% 1.41 * denotes that a significant difference between the two sample means at the 0.05 level of significance. 189d. lewis et al. / financial services review 8 (1999) 183–197 amounts of inputs and the optimal allocations of inputs such that total costs are minimized for a given number of revenue transactions. total costs consist of commissions paid to selling agents, the value of non-selling services provided by broker-owners, advertisement and promotional costs, the cost of buildings and occupancy, and all other production related expenditures. the selling expenses include multiple listing service (mls) fees that vary directly with sales, bonuses of sales managers based on sales-staff performance, commissions paid to owners, and commissions paid directly to the sales staff. within these expenses, the costs of affinity relationships are implicitly included. the cost function that we define and estimate in the next section expresses total costs as a function of output and input prices. hence, we convert the inputs of total cost into four input prices. we define the prices of labor (plab), physical capital (pcapital), advertising and promotions (pad) and other inputs (pother). wages of employees are total sales-related expenses plus salaries of all clerical, secretarial, and sales managers’ divided by the number of full-time equivalent employees. the rents on physical capital are total occupancy expense divided by the number of real estate offices. advertising and promotion expenses are expressed as a percentage of revenue transactions. and, “other” inputs are also expressed as a percentage of revenue transactions. all of these input prices are expressed in natural log form in the estimated model. as such, table 2 lists the mean and standard deviation for each of these four input prices and for total cost in natural log form. 4. the efficiency estimation methodology 4.1. the stochastic frontier methodology we use a stochastic frontier model to estimate firm efficiency and economies of scale. aigner, lovell, and schmidt (1977) and meeusen and van den broeck (1977) introduce stochastic frontier model to the finance and economics literature. in essence, the technique table 2 summary statistics (full sample) for the input prices and firm total costs in logarithmic forma mean standard deviation ln (plab) 10.29 1.42 ln (pocc) 9.99 1.05 ln (pad) 5.16 0.78 ln (pother) 6.16 0.88 ln (total costs) 13.37 1.61 a plab 5 prices of labor; pcapital5 price of physical capital; pad 5 price of advertising and promotions; pother 5 price other inputs. wages of employees are total sales-related expenses plus salaries of all clerical, secretarial, and sales managers’ divided by the number of full-time equivalent employees. the rents on physical capital are total occupancy expense divided by the number of real estate offices. advertising and promotion expenses are expressed as a percentage of revenue transactions. and, “other” inputs are also expressed as a percentage of revenue transactions. 190 d. lewis et al. / financial services review 8 (1999) 183–197 constructs an efficient frontier that represents the minimum total costs that a firm can incur given its outputs and input prices. after constructing the frontier, we then determine how far an individual firm is deviating from efficiency. with a stochastic frontier technique, we decompose deviations into two-components: random error and firm inefficiency. in other words, a firm can deviate from the efficient frontier by either incurring additional costs (inefficiency) or because of measurement error or bad luck, which we consider out of the firms control. the stochastic frontier approach allows us to determine what percentage of the deviation from the frontier is a function of inefficient operations and what portion of the deviation is simply measurement error. we make the common assumption that the random error term is two-sided, and is normally distributed with mean zero and variance,s2. in other words, the random error component can either increase or decrease total costs. we assume that the inefficiency component is distributed one-sided (exponentially), such that it can only increase the firm’s total costs. 4.2. the stochastic cost frontier the stochastic frontier approach is parametric, and as such it is necessary to specify a functional form. following previous efficiency studies for real estate brokerage firms, we utilize use the translog cost function to estimate cost efficiency: lntc~pi,y! 5 bo 1 o i51 4 bilnpi 1 o i51 4 o j51 4 bij lnpilnpj 1 b6lny 1 b7lny2 1 vi 1 zi (1) wheretc represents the firm’s total cost.tc(.) is the actual cost frontier that depends on four input prices,pi, and a single output,y. zi andvi represent a composed or two-part error term. zi is the non-negative stochastic error term reflecting firm inefficiency, and (vi), is the symmetric, two-sided error term that captures other deviations from the frontier such as measurement error. we make the usual assumption about the two-sided error term,vi ; iid n(0, s2). for the non-negative, one-sided error term, we assumezi follows an exponential distribution with shape parameterl that defines both the mean and variance of the exponential distribution. in our analysis, we allowl to take on two different values,l1 for affinity firms and l2 for non-affinity firms. this allows for the possibility of different mean inefficiency across the two groups of real estate brokerage firms. traditional approaches estimate eq. (1) by using either a corrected least squares approach or by maximum likelihood. in this paper, we use a bayesian estimation procedure. with a bayesian approach, we can use prior information about parameters from economic theory and/or previous studies. unlike the traditional statistics, in a simulation type procedure, we can calculate the precision of all parameter values, including individual firm efficiency and the returns to scale, by reporting a 90% confidence interval for each parameter. lastly, the bayesian technique allows us to estimate group type efficiency under a single efficient frontier. that is, we provide for inefficiency estimates for two groups in our sample under a single cost frontier. this methodology is superior to estimating two separate frontiers—one for affinity firms and another for non-affinity firms because frontiers tend to cross, making it impossible to interpret a single firm’s technical efficiency (refer to lewis and anderson, 1999 for a detailed methodological discussion on the advantages of estimating the condi191d. lewis et al. / financial services review 8 (1999) 183–197 tional efficient frontier). lastly, from the estimated parameters, we compute the odds that affinity affiliated brokerages are more efficient than non-affinity firms. 4.3. incorporating bayes rule and the stochastic frontier methodology in this paper we use bayes rule to combine our prior knowledge, which we call our prior density function with the knowledge we gather from the data described by the likelihood function to form a posterior density function. the expected value of the posterior density distribution is the weighted-average of the prior density function and the likelihood function. the more variance or uncertainty there is in the likelihood function or the smaller is the observed sample data set, the more emphasis the prior distribution has on the posterior. the more uncertainty associated with the prior density function and the more observed data, more emphasis is given to the data described by likelihood function. in other words, the data becomes more important than priors as the certainty and quantity of the data increase. 4.4. the prior density function we choose a uniform prior for the frontier’s coefficients and the standard error of the model implying that we do not possess any prior knowledge about those parameters. and, as noted above, the inefficiency component of the two-part error term is defined by an exponential distribution withlj being the shape parameter that defines the mean of the exponential density function conditional on whether the firm is an affinity member or not. j takes on a value of one if the firm partners in an affinity relationship and a two if the firm is not affiliated with an affinity. as in previous studies we chose a gamma prior forl21., and set the prior industry efficiency measure to 0.875. monte carlo integration allows us to derive the posterior marginal density functions that would otherwise be impossible to derive analytically. using the gibbs sampler we draw 25,000 observations from the conditional joint probability distribution functions, dropping the first 5,000 iterations to avoid sensitivity to starting values that may occur. we use the observations that we sample to form marginal posterior density functions for each of the model’s parameters. from the marginal posterior density functions, we calculate the expected value and 90% confidence intervals of each of the model’s parameters, 194 in all. economic theory suggests that the cost frontier is monotonically increasing and concave in input prices. another requirement imposed by economic theory is that the average cost function must be u-shaped. terrell (1996) demonstrates how to impose these restrictions. within the prior, we restrict the cost function to be monotonically increasing in input prices. in the translog cost frontier, this implies that the share equations must all be positive. we also impose the cost frontier to be concave in input prices. to assure that the average cost function is u-shaped within the translog model specified above we restrict the partial derivative of average firm total cost to firm output (revenue transactions) to be greater than or equal to zero at the minimum. 192 d. lewis et al. / financial services review 8 (1999) 183–197 5. the bayesian results 5.1. the efficiency results initially, we estimate the overall sample’s inefficiency (l)using a model that does not distinguish differences in the brokerage type. in other words the stochastic frontier’s parameter results do not take into account that there may be differences in the performance of affinity brokerages and non-affinity brokerages. the restricted base results show that sample real estate brokerage firms are approximately 81.5% efficient. these results are consistent with lewis and anderson (1999) and anderson, lewis, and zumpano’s (2000b) conclusion that brokerages, overall, are relatively efficient. it is also consistent with the findings of anderson, lewis and zumpano (2000a) that firms become more inefficient as the residential real estate markets strengthens. by imposing concavity, monotonicity, and the u-shaped average cost restrictions, the confidence intervals stated are much narrower than if no restrictions had been applied. subsequently, we compute efficiency measures contingent upon whether the firm has at least one affinity relationship or not. the results of this estimation are reported in table 3. we find that the sample firms with at least one affinity relationship are nearly nine times more cost inefficient than firms that have no affinity relationships. with the restricted bayesian stochastic cost frontier, affinity firms are approximately 35.2% inefficient and non-affinity firms are approximately 4.2% inefficient. this means that the average affinity participating firm in the sample could reduce its input costs by 35.2%, without decreasing output. on the other hand, non-affinity sample firms could only reduce input costs by 4.2% given their level of output. table 3 also establishes individual brokerage efficiencies for five arbitrary affinity affiliated firms and five arbitrary non-affinity firms. as noted in the table, the first example of the affinity firm shows that the firm is approximately 84.6% efficient, whereas the first example of the non-affinity firm shows that it is approximately 94.7% efficient. using the posterior marginal density functions constructed using the gibbs sampler, the probability that affinity firms are more efficient than non-affinity firms is less than one in 20,000. this converts to an odds ratio of 0.00005 to one. 5.2. economies of scale results to investigate optimal firm size, we calculate economies of scale from the efficient cost frontiers estimated above. a firm’s scale economies are 1 ­tc/­y evaluated at y, where y represents output as measured by revenue transactions. a result greater than one suggests the firm is operating at increasing returns to scale, a result less than one states that the firm is operating with decreasing returns to scale, and a result of one indicates the firm is operating with constant returns to scale. we calculate returns to scale two ways. first we use the base parameter estimates where the stochastic frontier does not 193d. lewis et al. / financial services review 8 (1999) 183–197 distinguish between types of brokerage firms–ones that participate in affinities and ones that do not participate and the parameter estimates. the next set of estimations the parameter estimates are from the stochastic frontier that is constructed conditional on the two separate firm types; affinity and non-affinity participants. the results from the former are in table 4 (panel a) and the results from the latter are in table 4 (panel b). in table 4 an overwhelming amount of evidence suggests that a majority of the sample firms are operating at increasing returns to scale. rather than reporting the returns to scale for all firms in the sample we report returns to scale for each quartile of brokerages based on the number of revenue transactions made. both sets of results show that brokerages in all table 3 conditional efficiency results for affinity and non-affinity firma arbitrary brokerage # posterior means [90% confidence interval] affinity brokerage 1 84.6% [64.3%, 98.7%] affinity brokerage 2 45.5% [29.6%, 98.7%] affinity brokerage 3 63.3% [42.1%, 88.5%] affinity brokerage 4 71.9% [49.3%, 95.1%] affinity brokerage 5 71.9% [49.7, 94.8%] nonaffinity brokerage 1 94.7% [82.4%, 99.8%] nonaffinity brokerage 2 96.3% [88.6%, 99.8] nonaffinity brokerage 3 94.3% [81.6%, 99.8%] nonaffinity brokerage 4 96.5% [89.4%, 99.8%] nonaffinity brokerage 5 96.9% [90.6%, 99.9%] affinity group inefficiency (l1) 35.2% [27.0%, 44.8%] affinity group inefficiency (exp(2l1) 70.3% [63.9%, 76.3%] non-affinity group inefficiency (l2) 4.2% [2.0%, 7.7%] non-affinity group efficiency (exp(2l2)) 95.9% [92.6%, 98.0%] relative group type inefficiency l1 l2 9.74 [4.50, 17.69] prob (l2,l2) 5 prob sl1 l2 , 1d probability that affinity brokerages are more efficient ,1 out of 20,000 odds that affinity brokerages are more efficient ,0.00005 to 1 a industry inefficiency, arbitrary examples of individual brokerage efficiency and the variance of the frontier’s measurement error are listed with 90% confidence intervals. 194 d. lewis et al. / financial services review 8 (1999) 183–197 quartiles are facing increasing returns to scale (irs). in fact, none of the sampled firms seem to be facing returns to scale that are less than one (decreasing returns to scale (drs)). zumpano, elder, and crellin (1993), zumpano and elder (1994), anderson, fok, zumpano, and elder (1998), and lewis and anderson (1999) also find that firms in the real estate brokerage industry are failing to take advantage of scale economies. the consolidation that is currently transpiring in this industry is evidence of a movement toward scale efficiency. the results obtained from the sample, contain information that is somewhat problematic from a normative and regulatory point of view. to the extent that affinity programs allow smaller firms to take advantage of economies of scale, affinity participation may be efficiency enhancing. on the other hand, if firms grow too large they may begin to operate less efficiently because of diseconomies of scale. we also know that affinity participation tends to be concentrated among the larger firms and affinity affiliations are increasing. although such diseconomies normally work to limit firm size, affinity programs may allow member firms to continue to grow despite inefficient operations because of a comparative advantage in generating listings and sales relative to non-affinity firms. if affinity groups provide a mechanism for less efficient firms to survive, or worse still, drive out more efficiently run firms, one result could be a less competitive market structure. this, in turn, could mean consumers could be left paying higher commissions for lower quality services. it must be pointed out that what has been said above is still highly conjectural, especially because we cannot quantify the consumer benefits of affinity programs. it is possible that the benefits of affinity participation outweigh the costs of these programs. moreover, it may not be the affinity programs, themselves, which are the culprits, but rather that affinity programs table 4 returns to scale results panel a. scale findings using the restricted base parameter results quartile 1y­tcy­y [90% confidence interval] returns to scale 25th (lny 5 4.63) 1.160 [1.077, 1.251] irs 50th (lny 5 5.38) 1.154 [1.072, 1.245] irs 75th (lny 5 6.63) 1.147 [1.067, 1.239] irs panel b: scale findings conditional on firm type quartile 1y­tcy­y [90% confidence interval] returns to scale 25th (lny 5 4.63) 1.22 [1.14, 1.30] irs 50th (lny 5 5.38) 1.21 [1.14, 1.29] irs 75th (lny 5 6.63) 1.20 [1.13. 1.29] irs 195d. lewis et al. / financial services review 8 (1999) 183–197 may encourage inefficient growth in firm size. equally important, the efficiency estimations undertaken in this study are based upon only one year of observations and affinity programs are still evolving. what the collective effect of affinity programs may be over time cannot yet be determined. 6. concluding remarks in this study, we provide the first rigorous empirical examination of the impact of affinity programs on real estate brokerage firms and the brokerage industry as a whole. most of the evidence obtained from the sample supports the notion that affinity participation is on the rise, but still remains somewhat modest. most of the sample firms that participate in affinity programs have multiple affinity partners, which usually take the form of corporations, unions, professional associations, and employers. further evidence suggests that these affinity participating firms employ more workers, operate more offices, are more likely to franchise, and have more mls affiliations. with respect to firm performance, an examination of the actual financial numbers shows that affinity programs can potentially help firms grow in output and revenue volume. however, the results also suggest that these additional revenues are not being translated into increased profitability. in fact, on a per revenue transaction basis, non-affinity firms in the sample produce higher levels of adjusted net income, although the difference is not statically significant. it seems that the additional revenues generated from affinity program participation are offset by commission discounts and the increased costs associated with increased production. to analyze the efficiency implications of affinity programs, we estimate x-efficiency levels for affinity-participating firms relative to non-affinity participating firms. for this sample, we find that affinity firms are much more inefficient than non-affinity firms. in fact, the probability that affinity firms are more efficient than non-affinity firms is less than one in 20,000, which translates into an odds ratio of 0.00005 to one. finally, we find further evidence that firms are operating at increasing returns to scale. as we note above, the results here are preliminarily and somewhat conjectural due to data limitations. however, the findings of this study serve as important benchmarks that can be used to monitor the growth and efficiency of affinity programs over time. the sample results obtained do provide strong and statistically significant evidence that affinity programs may hinder efficiency in the residential real estate brokerage market. given the growth of affinity programs, these empirical findings pose concerns for the individual real estate investor. in less efficient and competitive markets firms may not have to work as hard or as efficiently to compete and survive. in a brokerage context, consumers could potentially see higher fees or at least fees that are above a market competitive rate. moreover, the quality of service provided by firms may decrease, possibly resulting in higher search costs for the potential buyer or longer time on the market for the seller. the results suggest that regulators and policy makers may be justified in worrying about the growth of affinity programs within residential real estate brokerages, as it may ultimately be the individual investor in real estate that suffers. 196 d. lewis et al. / financial services review 8 (1999) 183–197 references aigner, d. j., lovell, c. a. & schmidt, p. (1977). formulation and estimation of stochastic frontier function models.journal of econometrics, 6,21–37. agency law quarterly. (1998).real estate intelligence report, vol. 9, no. 5, april anderson, r., fok, r., zumpano, l. v. & elder, h. w. (1998). measuring the efficiency of residential real estate brokerage firms.journal of real estate research, 16,139–158. anderson, r., lewis, d., & zumpano, l. v. (2000a). residential real estate brokerage efficiency from a cost and profit perspective.journal of real estate finance and economics, 20,3. anderson, r., lewis, d., & zumpano, l. v. (2000b). x-inefficiencies in the residential real estate market: a stochastic frontier approach.journal of real estate research. ball, s. (1990). make the most of referral business.today’s realtor, 23, 27. berger, w. (1997). just whom do affinity relationships benefit.today’s realtor, 29,46–49. dezube, d. (1996). hold on to your wallet: referral fees are growing.today’s realtor, 29,42–45. lewis, d. & anderson, r. (1999). is franchising more cost efficient? the case of the residential real estate brokerage industry.real estate economics, 27,543–560. meeusen, w. & van den brock, j. (1977). efficiency estimation from cobb-douglas production functions with composed error.international economic review,18435–444. mcmillan, d. (1997). affinity relationships hot elsewhere, but not here: real estate discount for business groups rare in portland.business journal-portland, 14,24–25. national association of realtors©.real. estate brokerage industry practices: affinity and referral/relocation relationships,1996. national association of realtors©.profile of real estate firms,washington, d.c. 1998. national association of realtors©.1997 real estate brokerage industry practices: affinity and referral/ relocation relationships,1998. reich–hale, d. (1999). affinity groups offer sales opportunities.national underwriter property & casualty-risk & benefits management, 103,12–13. terrell, d. (1996). incorporating monotonicity and concavity restrictions in flexible functional forms.journal of applied econometrics, 11,179–194. zumpano, l. v., elder, h. w., & anderson, r. i. (2000). the residential real estate market: an analysis of revenue and cost conditions over the last decade.journal of real estate research, 18,1/2. zumpano, l. v. & elder, h. w. (1994). economies of scope and density in the market for real estate brokerage services.american real estate and urban economics association journal, 22,497–513. zumpano, l. v., elder, h. w., & crellin, g. e. (1993). the market for residential real estate brokerage services: costs of production and economies of scale.journal of real estate finance and economics, 6,237–250. 197d. lewis et al. / financial services review 8 (1999) 183–197 pii: 1057-0810(95)90016-0 financial services review, 4(l): 31-40 copyright 0 1995 by jai press inc. issn: 1057-0810 ai1 rights of reproduction in any form reserved. a simplified approach to measuring bond duration jean l. heck terry l. zivney naval k. modani because interest rates vary over time, the realized return on a fixed-income investment will depend on the price at which the instrument is ultimately liquidated and the rate at which interim cash flows are reinvested. this variation in realized return, known as interest-rate risk, should be addressed by both individual and institutional investors. tools for measuring the impact and adjustingfor the efsects ofinterest nate changes on fixed-income inst~me~t pe~o~ance have long been available with duration and its com~an~n adjust ment factor, convexity. in this article, a simpltjied a~te~ati~ie to the tr~itional complex duration calculation is developed and demonstrated. thus, anyone who can calculate a bond price can quickly estimate the interest rate risk associated with a bond as well as calculate the expected bond price change for a given change in market yield-to+naturuy. i. introw~t~on because interest rates vary over time, the realized return on a fixed-income investment will depend on the price at which the instrument is ultimately liquidated and the rate at which interim cash flows are reinvested. this variation in realized return, known as interest-rate risk, should be addressed by both individual and institutional investors. tools for measuring &he impact and adjusting for the effects of interest rate changes on fixed-income instrument performance have long been available with duration and its companion adjustment factor, convexity.’ although the rigorous calculation of these measures is intellectually appealing to academics, and easy to employ in a high-tech institutional setting, such calculations can be daunting to individual investors. in this paper a different, yet simple, approach for m~suring duration and convexity is developed and demonstrated. ii. prior approaches to computing duration duration was originally derived as a superior way of capturing in a singlemeasure the impact of the pattern of a bond’s cash flows on its sensitivity to interest rate changes. two bonds jean l. heck l coilege of commerce and finance, villanova university, villanova, pa 19085. terry l. zivney l college of business, ball state university, muncie, in 47306. naval k. modani l college of business administration, university of central florida, orlando, pl 32816. 32 financial services review 4(l) 1995 with the same maturity and yield-to-maturity (yield) but different periodic coupon cash flows can result in very different levels of terminal wealth for their owners. as market interest rates fluctuate following the acquisition of a bond, the rate at which interim cash flows are reinvested will vary. also, if the bond is liquidated prior to maturity, the price received will depend on market interest rates at that time. the decrease (increase) in liquidation value seldom exactly offsets the increase (decrease) in reinvestment income that results from changing market interest rates. one way to avoid such terminal wealth risk is to hold only zero-coupon bonds that have a maturity equal to the planned holding period. however, because of a shortage of a range of “zeros” with desired risk/return relationships as well as certain tax implications associated with holding zeros, other strategies for addressing interest rate risk are necessary. in a book written for the national bureau of economic research in 1938, frederick macaulay first developed the concept of bond duration, duration basically measures the weighted average amount of time it takes to receive the present value of cash flows from a bond. duration = [x tc, / (1 + i)‘] / [x c, / (1 + i)‘] where c, is the cash payment received at time t and i is the yield-to-maturity. reilly (1989) shows that the duration of a 4% annual coupon bond that matures in 10 years with a yield-to-maturity of 8% is computed as follows: (1) (2) year cash flow 1 40 2 40 $ 40 10 1040 (3) (4) (5) (6) pv @8% pv of cf pv as % of price (1) x (5) .9259 37.04 .0506 .0506 .8573 34.29 .0469 ,093s .5oq2 . . . . . . 20.02 . . . . .0274 . . . . . .2466 . . . . . 4632 481.73 .6586 6.5850 731.58 l.woo 8.1193 duration = (6) + (5) = 8.12 years as noted by reilly (1989), generally: 1. bond duration is less than term to maturity; 2. there is an inverse relationship between duration and coupon rate; 3. there is an inverse relationship between duration and yield-to-maturity; and 4. bond price movements vary proportionally (linearly) with duration the estimated bond price change resulting from a change in yield is: where: ap = the change in bond price caused by ai p = the beginning price of the bond -&od = modified duration in years = duration / (l+iln) where n is the number of coupons per year ai = the change in yield-to-maturity thus, for the above bond, if the market yield-tc+maturity declined by 7.5 basis points, the change in price would be: a simplified approach to measuring bond duration 33 price $ p, _ . . . . . i i+ai figure 1. price/yield curve yield to maturity ap g (-8.1211 .os) x (-.0075) x (73 1.58) = 41.25 this indicates that if the market yield-t*maturity declines from 8% to 7.25%, the bond price would increase to: $731.58 + 41.25 = $772.83 duration can be viewed as the slope of a straight line tangent to the price/yield curve. figure 1 presents the price/yield curve for the 4 percent coupon bond in the preceding example. points along the curve represent prices for differing yields-tc+maturity. if the yield changes from i to i + a, the price of the bond would change from point po to point u. the slope of the tangent line at point po estimates the change in the bond price that would occur given a change in the yield. because the curve is convex, the accuracy of the estimate of price change depends on that degree of convexity. a convexity correction factor is often used to adjust the price change estimated by using the bond duration. that factor is given by: convexity = cvx = (62p/zii2)/ p = [1/(1+i)*]x[~c,(t2+t)/(1+i)‘] the total change in price resulting from a change in yield is therefore:2 ap z duration change + convexity change = -d,,,od x ai x p + 0.5 x p x cvxx (ai) from the above calculations, we observe that estimating the expected change in bond price for a given change in yield-to-maturity is quite complex and time consuming. however, duration has several valuable uses for investors, including individual investors. 34 financial services review 4(l) 1995 first, duration serves as a measure of interest rate risk. the prices of bonds with greater durations are more sensitive to interest rate changes. thus, knowledge of a bond’s duration provides a useful benchmark for comparing the riskiness of alternative bonds. second, investors with a specific investment planning horizon can select those bonds having durations most closely matching their planning horizon. interest rate risk is minimized for bonds whose duration matches the planning horizon. third, because the duration of a portfolio is the weighted average of the individual bond durations, investors can readily compute their total interest rate risk exposure from durations of the component bonds. numerous attempts have been made to simplify the estimation of macaulay’s duration measure. jess chua (1984) derived a closed-form formula that enables faster computation of duration. chua’s formula is: d = c[i(l +yy+l -cl +y>-ymvy2(1 +y>"l +(~(~/(l +y)m b where: d = duration in periods m = maturity in periods c = coupon in dollars per period f = face value y = yield to maturity per period b = value of the bond at y caks, lane, greenleaf, and joules (1985) (hereafter, clgj) utilized the linearity of duration to calculate duration for a coupon bond as a combination of the interest and matu~ty payments: d=n-(ufy) [n-(1 +y)a,,j where: d = duration in years n = years to maturity c = yearly coupon payment p = market price of bond y = yield to maturity in annual terms (all periods in clgj are annual periods) an = present value of an annuity yielding y for n years moser and lindley (1989) adapted the clgj formula to the case of multiple coupons per year with the following: d-n-c/py[n-(1 +y/k)]a,/k] where y/k is the rate per period and kn is the total number of periods. benesh and celec (1984) offered the following simplified formula with the alternative assumption that annual yields are arrived at by com~unding periodic yields rather than the usual assumption that annual yields are computed by multiplying the periodic yield by the number of periods per year: d = l/m + c/k[(acf,, n) + (n l)] / [cacf, + m] where: acf,,= (m/k)[(l +k/m)“i] c = annual coupon rate k = annualized yield to maturity from compounding periodic ~~ a simplified approach to measuring bond duration 35 m = number of payments per year n = total number of payments remaining until maturity the choices for calculating duration boil down to the original procedure involving numerous weighted present value calculations or the above “simplified” formulas. all of the choices are daunting for the typical individual investor. to address this problem, the following section develops a simplified procedure for calculating duration that is “truly” simple. iii. derivation of simplified duration formula most individual investors have access to a financial calculator and are capable of calculating a bond price. if not, learning to do so is quite simple. the derivation of the simplified formula is based on figure 2 and the definition of duration as the negative interest rate elasticity of the bond price. at the market yield-to-maturity, i, a bond would be priced at pa. the modified duration for a bond is related to the slope of the tangent line at m. however, for small changes in i, the slope of that tangent line at point m is equal to the slope of the line drawn from point l to point m. thus, it follows: and slope of tangent line = (p+-p_) / (2ai) = (po-p.) / ai = (p+-po) / ai slope of line from l to m = (m l) / ai therefore: lim (l-m)/ai lim (m-l/)/ai’ lim (p_-p+)/a2i) ai-sl ai+0 = = ai-0 po po po = modified duration therefore: d (l-m)/ai mod = (1) po also, convexity is proportional to the change in slope at point m: convexity = (slope” slopel)lai (2) po and, from figure 2, we see that: convexity = -(m-lj)/ai [-(l-m)]/ai = l+u-2m po p&i duration is reported in terms of a number of periods while the units of convexity are periods squared. with semi-annual coupons, duration would be in terms of half-years, while convexity is in terms of half-years squared. thus, with semi-annual coupons, semi-annual duration is divided by 2 to arrive at duration in years while convexity is divided by 4. 36 financial services review 4(l) 1995 price $ p_ -’ p,=m -. u -’ p+ -’ tangent slope = -d (p+-pjai x po/(1 +i) i-ai i i+ai figure 2. price/yield curve yield to maturity it should be noted that the above calculations for duration and convexity hold for general conditions about the shape of the yield curve. in particular, macaulay’s duration assumes that the yield curve is flat and that changes in the level of interest rates result in parallel shifts in the yield curve. the analytic “simplifications” described by others also rely upon these restrictive assumptions. our methodology is more general because we compute duration directly from the prices implied by any yield curve. thus, our method is not only simpler but in general more accurate than previous methodologies. a. example of duration and convexity calculations the simplicity of these newly derived duration and convexity calculations is best demonstrated through an example. consider an l&year to maturity, 12% coupon bond, which is selling to yield 9% (reilly, 1989, p. 432). using the traditional procedures, we find the following: price3 = 1265 modified duration = 8.38 years convexity = 107.7 ai = 100 basis points using the macaulay steps:4 z [60 x (1 + .045)-’ x 1 + 60 x (1 + ,045)~’ x 5 duration calculation: + 60 x (1 + .045)-*x 2 + 60x(1 +.0415~‘x3 + 60x(1 +.o45)ax4 + 60x(1 +.o45)6x6 + 60 x (1 + .045)-’ x 7 + 60~(1+.045)~x8 + . + . + . + . + 60x(1+.045)-‘3x33 +6o~(l+.o45)-~~x34 +60~(1+.045)-~~x35 +60x(1+.045)-‘6x36 + 1000 x (1 + .o45)-36 x 36]+ 1265 + 2 a simplified approach to measuring bond duration 37 duration = 8.76 modified duration = 8.76/l .045 = 8.38 convexity calculation: z [60x(1+.045)-‘x2 + 60x(1+.045)-‘x6 + 60x(1+.045)-‘x12 + 60x(1 +.045)ax20 + 6ox(l+.o45)-5x3o + 6ox(l+.o45)“x42 + 60x(1+.045)-‘x60 + 60a(l +~m5)-~x72 + . . . . . . . . . + . . . . . . . . . + . . + . . . . . . . . . . + 60x(1 +.o45)-33x 1092 + 60 x ( 1+.045)-34 x 1190 + 60 x (1 + ~i45)-‘~ x 1260 + 60 x (1 + .o45)-36 x 1332 + 1000x(1 +.045)-36x1332]x[1/(1.045)2]+ 1265-4 convexity = 107.70 for a 100 basis point decrease in yield, the predicted price change would be: predicted price change z duration change + convexity change = -dmod x ai x price + s x price x convexi x ai = -8.38 x (-.ol) x 1265 + 5 x 1265 x 107.70 x (-.ol)* = 112.82 predicted new price after ai = 1265 + 112.82 = 1377.82 the actual bond price with a yield to maturity of 8% would be $1,378.17. thus, using the usual measures of duration and convexity to predict the new bond price results in an error of 35 cents. b. using the heck/zivney/modani (hzm) steps to compute duration and convexity using our simplified procedures, calculate bond prices for a small change in i above and below the current semi-annual yield-t*maturity for the bond. because of rounding in the reilly example, our calculations will differ slightly.5 for example: at i = 4.5%, price (m) = 1264.99 at i = 4.49%, price (l) = 1267.11 at i = 4.51%, price (u) = 1262.87 thus: ai = .0450 .0449 = .oool d mod = [(l-m)/mj + ai + 2 = r(1264.99 1267.11)/1264.99] + .oool + 2 = 8.39 the percentage change in price is divided by 2 in the equation for modified duration to reflect semi-annual compounding. the formula for convexity involves dividing by 22, or4, to reflect the semi-annual compounding: convexity g (l+u-2m) + ai* + m + 4 38 financial services review 4(l) 1995 = (1267.11 -t1262.87 2 x 1264.99) + .oool* + 1264.99 + 4 = 107.70 the ai used to compute duration and convexity (generally a very small number) need not be the same as the ai used to estimate price changes. therefore, the predicted change in the price for a 100 basis point change in i (50 basis points each six-month compounding period) results in the following predicted price: price change 3 duration change + convexity change = -dmmod x ai x price + 0.5 x price x convexity x ai* = -8.39 x (-.ol) x 1264.99 + 0.5 x 1264.99 x 107.70 x (-.ol)* = 112.94 predicted price after ai = 1264.99 + 112.94 = 1377.93 the predicted price of $1,377.93 is only 24 cents from the actual price of $1,378.17 and, hence, more accurate than the traditional computations, which resulted in an error of 34 cents. the hzm estimated bond price change calculations involve seven separate mathemati cal steps. to summarize those seven steps, calculate: 1. bond price at current yield-to-maturity 2. bond price for a small increase from current yield-to-maturity 3. bond price for a small decrease from current yield-to-maturity 4. change in yield-to-maturity (step 1 yield minus step 2 yield) 5. modified duration 6. convexity 7. expected price change as sum of duration change and convexity change c. evaluating the performance of hzm the previous section demonstrated the ease with which the expected bond price change can be estimated using the hzm method. in the above example, the hzm and traditional macaulay results were quite similar. the obvious question the example raises is how consistent is the accuracy of the hzm method. to address this question, a simulation was performed. table 1 presents some of the results of a simulation that generated predicted bond price changes for both the hzm and traditional methods under varying scenarios. using a 10 percent coupon bond with maturities ranging from five to 30 years, predicted bond price changes were calculated for yield-to-maturity changes ranging from one to 100 basis points and for hzm ai ranging from .ol to 100 basis points. results presented include yield-to-maturities of 3% to 18% in three-percent intervals, for the hzm method using ai of .ol and one basis point as well as the traditional macaulay method. it can be seen in each cell of table 1 that the prediction error associated with the hzm method decreases with smaller ai; also, for small ai, hzm and macaulay yield nearly identical results. the last row and column of cells compares the average error under each a simplified approach to measuring bond duration 39 table 1 actual price change minus predicted price change using 100 basis point change in ytm yield-to-motrtritg on 10% coupon bond maturity method 3% 6% 9% 12% 15% 18% average error 5 years hzma=l.o 0.26* 0.2 i 0.18 0.15 0.12 0.10 0.174 hzm a = .ol macaulay 10 years hzma= 1.0 hzma=o.l macaulay 15 years hzma= i.0 hzm a = .ol macaulay 20 years hzma= 1.0 hzma=.ol macaulay 25 years hzma= 1.0 hzm a = .ol macaulay 30 years hzma= 1.0 hzm a = .ol macaulay average error hzm a = 1 .o hzm a = .ol macaulay 0.20 0.20 i.56 1.35 1.35 0.17 0.17 1.13 0.98 0.98 4.54 2.90 4.09 2.60 4.08 2.60 0.14 0.11 0.14 0.11 0.83 0.6 1 0.71 0.52 0.71 0.52 1.87 1.22 1.67 1.08 1.67 1.08 9.64 5.43 3.11 1.82 8.85 4.97 2.84 1.65 8.84 4.96 2.84 1.65 17.08 8.54 4.39 2.33 15.88 7.92 4.05 2.14 15.87 7.91 4.05 2.14 0.09 0.09 0.45 0.38 0.38 0.80 0.71 0.7 1 1.09 0.98 0.98 1.28 1.16 1.16 26.95 12.02 5.59 2.72 1.39 25.29 11.24 5.20 2.51 1.28 25.27 11.23 5.20 2.51 1.28 10.007 5.043 2.665 1.478 0.859 9.28 1 4.649 2.439 1.341 0.773 9.273 4.645 2.437 1.340 0.772 0.08 0.136 0.08 0.136 0.33 0.821 0.28 0.708 0.28 0.707 0.53 1.981 0.47 1.775 0.47 1.773 0.66 3.629 0.59 3.318 0.59 3.315 0.73 5.728 0.66 5.307 0.66 5.302 0.75 8.242 0.68 7.704 0.68 7.699 0.523 3.429 0.465 3.158 0.465 3.155 note: *for a 10% coupon bond maturing in 5 years, with a yield-tomaturity of 3%. and experiencing a 100 basis point change in yield, the difference between the actual change in price and the price change predicted using the hzm model (with a a = 1 .o) is 26 cents. a i = 1 .o means that a one basis point change was used to compute duration and convexity with the hzm method while a i = .ol means that a one-hundredth basis point change was used to compute duration and convexity. scenario. the results suggest that not only is the hzm method easier to employ, but it yields equally accurate results when computed using small ai. although the derivation of the hzm simplified formulas for duration and convexity calculations is somewhat complex, the resulting duration and convexity formulas are much easier for an individual investor to use. the hzm simplified method provides estimates of price changes which differ from those produced by traditional duration measures by less than one penny on average. anyone who can calculate a bond price (actually, three different bond prices) can now quickly estimate the interest rate risk associated with a bond as well as calculate the expected bond price change for a given change in market yield-to-maturity. also, unlike previously derived closed-form formulas, our formulas can be used in the general cases of sloping yield curves and nonparallel shifts in yield curves. 40 financial services review 4(l) 1995 1. for a detailed discussion of convexity, see riley (1989, pp. 427-431), francis (1991, pp.407-411), kolb (1992, pp. 245-251), and maginn and tuttle (1990, pp. 68-69). 2. the expression is the first two terms of a taylor expansion series. an explanation of the origins of the 5 and (ad2 terms can be found in reilly (p. 430) and dunetz and mahony (1988, p. 57). 3. reilly uses $1,265, although the actual price would be $1,264.99. the $1,265 figure is used here to maintain comparibility with the reilly text example. 4. for semi-annual coupon bonds, the duration calculation must be divided by 2. because convexity is the second derivative, the ai/ term is squared, yielding ai’/4. thus, the convexity calculation must be divided by 4 for semi-annual coupon bonds. 5. answers are shown to the nearest cent. all calculations are performed using the memory functions of the calculator to maximize the accuracy for small changes in i. references benesh, g.a., & celec, s.e. (1984). a simplified approach for calculating bond duration. financial review, 19(4), 394-396. caks, j., lane, w.r., greenleaf, r.w., &joules, r.g. (1985). a simple formula for duration. journal of financial research, 8(3), 245-249. chua, j. (1984). a closed-form formula for calculating bond duration. financial analysfs journal, 40(3), 76-78. dunetz, m.l., &mahoney, j.m. (1988). using duration and convexity in the analysis of callable bonds. financial analyst journal, n(3), 53-73. francis, j.c. (1991). investments: analysis and management. 5th edition. new york: mcgraw-hill. kolb, r.w. (1992). investments. 3rd edition. miami, fl: kolb publishing. macaulay, f.r. (1938). some theoretical problems suggested by the movement of interest rates, bond yields, and stock prices in the united states since 1856. new york: national bureau of economic research. maginn, j.l., & tuttle, d.l. (1990). managing investment portfolios. 2ndedition. new york: warren, gorham and lamont. moser, j.t., & lindley, j.t. (1989). a simple formula for duration: an extension. financial review, 24(4), 61 l-615. reilly, f.k. (1989). investment analysis and portfolio management. 3rd edition. chicago: dryden press. pii: s1057-0810(97)90016-0 financial services review, 6(3): 221-224 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. book, software, and web site reviews douglas kahl, editor university of akron the millionaire next door. thomas j. stanley and william d. danko marietta, ga: longstreet press, inc.; 1996 reviewed by: douglas r. kahl, professor of finance, the university of akron. thomas j. stanley and william d. danko surveyed 3000 heads of households expected to have high net worth. of the 1,115 respondents, 385 had household net worth of at least $1 ,ooo,ooo. they present the results of the eight page surveys in a 258-page book, the mil lionaire next door. based on the survey, in depth interviews and prior research stanley and danko paint a detailed portrait of the american millionaire. they will tell you what the millionaire is likely to drive; how much he will spend for a suit, watch and shoes; how long and how often he has been married; and, perhaps what is most important, how he became a millionaire. stanley and danko’s portrait may not surprise most experienced financial planners, but will come as a shock for many of those who want to sell goods and services to the wealthy, who want to become wealthy, or who simply want to appear to be wealthy. the results presented in the book are thoroughly discussed and are enhanced by many examples and illustrations. while the covers of the book may be a little farther apart than is abso lutely necessary, the book is well worth reading for the insights it provides into the modem american millionaire and his lifestyle. the financial services revolution. edited by clifford e. kirsch irwin publishing; 1997. reviewed by: j. tim query, dept. of risk management & insurance, university of georgia. the financial services revolution examines the important changes that have taken place in the last two decades in the financial services industry. its intention is to bridge the inter ests of policy makers, practitioners, and academics. to that end, submissions are included from practitioners, academics, and officials from regulatory agencies. editor clifford e. 222 financial services review 6(3) 1998 kirsch is chief counsel, variable products, at the prudential insurance company of amer ica. this book is divided into eight parts. part i examines the evolution of bank and insur ance company activities. an interesting survey of bank securities activities, including a review of the history of commercial bank securities activity since the civil war, provides a chronological narrative. another chapter examines the regulatory structure that applies to equity-based insurance products. the regulatory developments that have allowed banks to become sellers of securities and insurance products are reviewed, and issues relating to bank underwriting of these products are explored as well. part i concludes with a discus sion of the tension between bank regulators and insurance regulators over issues concem ing the authority of banks. part ii focuses on derivatives. an excellent overview of derivatives is presented that include an explanation of how these products work and why they are used. the coverage of derivatives begins with the basics, and is highly recommended for those who may not be familiar with this complex area of securities. the debate over what type of derivative activ ity should be covered by the commodity exchange act is examined extensively. the final chapter in part ii focuses on the issue of whether off-exchange derivative transactions should be subject to antifraud provisions of federal law. part iii demystifies securitization, explaining the process by which illiquid loans are converted into securities. a thorough discussion of future prospects for this financing vehi cle is accommodated here as well. part iv examines the policy issues related to the growing popularity of the defined contribution pension plan and in particular, the “401(k) plan,” a market in which banks, insurance companies, and mutual funds vigorously compete. reasons for the popularity of the 401 (k) plan, the regulatory framework governing communications to participants con cerning their investment options, and a review of recent initiatives of the department of labor related to dc plans are included in part iv. in part v the role of mutual funds in corporate governance contrasted with the role of banks and insurance companies is discussed. the issues of trading versus monitoring, “the prisoner’s dilemma”-type incentive problems driving shareholder actions, and various paradoxical conflicts of interest are identified and provide interesting reading. part vi focuses on issues relating to the mutual fund industry. situations originating from the growing use by mutual funds of computer networks and other types of electronic media as a means to reach investors are examined. the motivation behind recent sec reg ulations restricting money market funds to investments in top-quality securities with low volatility and a short maturity is explained. the concluding chapter in part vi deals with issues related to effectively disclosing risk to mutual fund investors. part vii looks at pooled investment management vehicles designed for the wealthy and non-u.s. investors. the focus is on issues related to hedge funds, which are sold to wealthy individuals and institutions and avoid much of the regulatory scheme imposed on mutual funds, although they are structurally similar to them. part vii also looks at offshore investment funds, which are pooled vehicles that are established by u.s. advisers for non u.s. resident investors. the final section, part viii, examines issues related to the regulatory framework applying to mutual funds, banks, and insurance companies. the question of whether a self regulatory organization similar to that which exists for brokerage firms, be established for mutual funds and advisers is addressed. the overlapping jurisdictions among federal and book, software and web site reviews 223 state regulatory agencies that apply to financial services regulation are also discussed. a perceptive look at the changes in the u.s. financial markets caused by the growth and expanded role of mutual funds provides stimulating insight. an explanation of functional regulation, a structure where financial activity is regulated by the same regulator regardless of the type of financial institution conducting the activity may serve as a preview of future regulatory strategy. part viii’s discussion is concluded by contrasting the various regula tory methods used to regulate risk in the financial services industry. wealth management. harold evensky chicago, il: irwin professional publishing; 1997 (isbn o-7863-0478-2) reviewed by: dale l. domian, associate professor of finance, memorial university of newfoundland. harold evensky’s wealth management provides a comprehensive and informative guide to financial planning. as noted by pahl (1996), there are several key steps to the financial planning process, including gathering data to identify the client’s objectives, developing the financial plan, and implementing and monitoring the results. while some planners may focus on specific areas such as tax or estate planning, others take a broader perspective to include all aspects of their clients’ financial lives. wealth management is directed toward the latter group, presenting a broad holistic approach for practitioners who truly are “wealth managers.” an underlying theme is that investment recommendations should be consistent with results from academic and professional research. evensky champions the work of brinson, hood, and beebower (1986), who show that 94 percent of the variation in portfolio returns is attributable to asset allocation, with just 4 percent due to security selection and only 2 percent from market timing. the book also recommends that projections of real returns and risk premiums be based on historical data, such as the well-known ibbotson series. a wealth manager can then determine an asset allocation which minimizes risk, while at the same time providing a sufficiently high expected return to meet the client’s needs. the book is organized in 16 chapters which progress through the key steps of the wealth management process. the first five chapters focus on the importance of a two-way exchange of information between wealth managers and clients. evensky provides his per sonal insights on gathering information from clients, and on explaining important finance principles so that clients will have realistic goals on what is obtainable. the section on taxes includes a discussion of portfolio tax management, refuting claims that substantial gains are available from active tax management strategies. chapters 6 and 7 provide a concise overview of statistics and portfolio theory. unfor tunately, several formulas contain errors, and some of the explanations are imprecise. readers without previous training in investment mathematics could be better served by the more detailed treatment in most investment textbooks (e.g., sharpe, alexander, and bailey [ 19951). asset allocation models and the development of investment policies are considered in chapters 8 through 10. the use of mathematical optimizers is recommended, with the pii: s1057-0810(99)00027-x family friendly firms: does it pay to care? dianna c. preecea,*, greg filbeckb acollege of business and public administration, university of louisville, louisville, ky 40292, usa bcollege of business administration, university of toledo, toledo, oh 43606-3390, usa abstract in this paper we examine the returns to a portfolio of 29 firms that are perceived as family-oriented. the sample is based on firms awarded the best 100 companies for working mothers in working mother magazine’sannual survey. there is much anecdotal evidence supporting the benefits of these programs, but little evidence relating family-oriented policies to shareholder wealth. we find, based on raw returns, that family-friendly firms do not earn statistically significant excess returns relative to a matched sample or to the s & p 500. based on risk-adjusted returns, the family-friendly portfolio outperforms the market, but underperforms a matched sample portfolio. © 1999 elsevier science inc. all rights reserved. jel classification:g10; g11 keywords:family friendly firms; portfolio returns; risk-adjusted returns 1. introduction investors continuously look for profitable investment opportunities. numerous studies have considered the factors that make a company worthy of investment. in this paper, the issue of “family-friendly” policies and the returns to investors who invest in family-friendly firms are examined. family-friendly issues have become increasingly important to corporate america. in 1996, the white house sponsored the first-ever conference on corporate citizenship. companies were showcased that, according to president clinton, “do the right thing” by their * corresponding author. tel.:11-852-4831; fax:11-502-7557. e-mail address:dcpree01@gwise.louisville.edu (d.c. preece) financial services review 8 (1999) 47–60 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00027-x employees. these companies offer on-site child care, flexible hours, and paternity leave. companies like corning and johnson & johnson were represented. vice-president al gore convened a second conference in 1996 to discuss how work affects family and vice versa. again, the leading firms were represented (moskowitz, 1996). firms instituting family-oriented policies have been on the rise since the middle to late 1980s. johnson & johnson, one of the most widely recognized leaders offering these benefits, launched most of its programs in 1989 (shalowitz, 1992). the issue has garnered such attention that many companies are now calling the programs work/life benefits to avoid alienating single and childless workers. family-oriented benefits are expected to expand well into the twenty-first century. for example, in a survey of 463 companies by the international foundation of employee benefits, 34% of firms expected to offer flextime by 2000 and 68% expected to offer either on-site child-care or child care subsidies (luciano, 1992). some firms are serious about family-oriented issues, whereas others are accused of having policies on the books but a corporate culture that discourages their use. according to a congressional study entitled “the changing workforce,” work/family issues are approached strategically as a means to improve recruitment, retention, and productivity by leading-edge companies (washington report, 1992). there is a core group of companies that are consistently cited as examples of firms committed to a family-friendly work environment. these firms do live up to their press. it is a sample of these firms that are examined in this paper. the stock market returns of a sample of 29 family-friendly firms relative to a matched sample and to a market portfolio are examined. annual and multi-year holding period returns are calculated. these returns are compared to the matched firms and to the s & p500. both raw and risk-adjusted returns are considered. 2. literature review there are two streams of literature that are relevant to this study: the literature on returns in the stock market and the literature on family-friendly firms. 2.1. stock market returns in a stream of literature spawned by the 1982 book by thomas j. peters and robert h. waterman,in search of excellence: lessons from america’s best run corporations,the returns to excellent firms were investigated. peters and waterman (1982) examined firms considered to be innovative and excellent by financial analysts, academics, executives, and consultants. companies were screened based on several factors including continuous innovation, size and financial performance. financial performance was measured by growth rates in equity and assets, profit margin, return on equity, return on assets, and on market-tobook-value ratios. based on the peters and waterman firms, clayman (1987) and kolodny, laurence, and ghosh (1989) considered returns from investing in these excellent firms. kolodny, laurence, and ghosh presented evidence that the peters and waterman firms did not outperform the market or a control sample. clayman found that the excellent firms did not outperform a 48 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 portfolio of “bad” firms, firms not considered excellent based on the peters and waterman criteria. in a 1994 follow-up study, clayman (1994) found conflicting evidence. over the period 1988 through 1992, the “good” companies outperformed the market whereas the “bad” firms under-performed the market. over the same period, however, the financial ratios of the “good” firms deteriorated whereas the ratios of the “bad” firms improved. hamilton, jo, and statman (1993) investigated the investment performance of sociallyresponsible mutual funds. both positive and negative factors were used in building the sample of funds. an example of a positive factor would be a firm with environmentally sound policies whereas a negative factor would include a company that produces weapons. the socially-responsible funds were measured against a sample of traditional mutual funds. hamilton, jo, and statman provided evidence that socially-responsible mutual funds do not earn statistically significant excess returns and that their performance was not statistically different from the performance of conventional funds. gorman, filbeck, and preece (1997) studied the returns from investing in a portfolio of the companies onfortune’s most-admired list. they compared the returns of the “mostadmired” firms (the firms at the top of thefortunelist) to the “least-admired” firms (the firms on the bottom of thefortunelist) and to the s & p500. they considered both annual returns and multi-year holding periods and found that generally the “most-admired” portfolio outperforms both the market and the “least-admired” firms on thefortune list. in 21 out of 22 years between 1973 and 1994, the “most-admired” firms outperformed the market and the “least admired” portfolio, based on raw returns. based on risk-adjusted returns, the “mostadmired” firms outperformed the market and the “least-admired” firms in the majority of years as well. finally, allen and kask (1997) examined the financial and market performance of socially-responsible firms. they defined social responsibility as a concern for the environment, community, women and minorities, and nuclear power. the sample included firms that have strengths and no weaknesses in the eight criteria defined by the domini social index. the index screens fortune 500 firms on a variety of positive and negative criteria and rates them based on strengths and weaknesses. allen and kask explored two regression models: one to assess the effect of social responsibility on profitability and the second to assess the effect of social responsibility on stock price performance. their results were surprising in that they found a positive effect on profits and a negative effect on stock price performance. allen and kask offered possible explanations for the conflicting results between the two models. they posit that either the result was sample specific or that it reflected investor confusion or a lack of information. 2.2. family-friendly firms there is a growing literature on firms that have family-oriented policies. since 1990, articles have appeared in publications such as thewall street journal (wsj), fortune, money magazine,andworking mother,in trade publications ranging frombusiness insuranceto modern office technology, and in academic publications such asemployee benefits journal andcompensation and benefits review(conference board, 1994). however, research in this area is largely anecdotal. little rigorous research has been done on the costs and benefits of 49d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 providing these benefits to workers. survey research on worker feelings about the policies has been the primary source of data regarding the benefits. why are so many firms adopting family-oriented policies? there are many benefits that accrue to the firm such as reduced absenteeism and enhanced recruitment. also, improved technology makes many of these policies easier to implement. certain types of work may now be performed at home with the advantage of computer technology. finally, firms are being forced by the marketplace to acknowledge and deal with family issues. according to thewsj, many “generation x” workers expect to lead more balanced lives than their parents and thus are judging firms’ attitudes towards work/family balance as part of their job decision-making process (shellenbarger, 1991). according to the new york based families and work institute, retention is the reason most often given for implementing work-life assistance policies (families and work institute web site, 1998). overall, 68% of the companies surveyed by the families and work institute find it difficult to fill vacancies for skilled workers and 40% find it difficult to fill hourly and entry-level positions. firms are responding in an effort to attract and retain high quality employees, in an age of low unemployment and significant competition for the most qualified workers. the issue has become so important that the council of institutional investors, a group of 94 pension funds that control $1 trillion in stocks between them, discussed how to encourage good workplace practices at their 1997 annual meeting. in addition, according to an ernst & young center for business innovation study, investor decisions are driven 35 percent by non-financial factors and the “ability to attract and retain” talented employees ranks fifth among 39 factors that investors use in picking stocks (shellenbarger, 1997). work-family research shows that child care or elder care problems lead to increased absenteeism, turnover, stress in the work place, and work disruptions. as a result, there is reduced productivity and morale problems in the workplace. according to the families and work institute, several unpublished corporate surveys indicated that child care responsibilities interfered with work for about half of the women and one-third of the men surveyed. corporate surveys also revealed that between 33–45% of employees with adult dependents work less effectively because of concerns about their relatives (shellenbarger, 1993). there is a debate about the effectiveness, given the cost, of child care centers provided by companies. on-site child care facilities are generally the most expensive of the familyfriendly benefits. the cost of these can be exorbitant. for example, according to woolsey (1992), a child care center built by johnson & johnson cost $5 million and a joint venture center built by all-state and other firms cost $1.6 million. anecdotal surveys of workers regarding on-site or near-site centers indicated less absenteeism and tardiness. rigorous research in this area has generally failed to substantiate this claim. for example, using companies’ administrative data, as opposed to worker opinions, a 1989 study by berkely planning associates found that only two out of five studies indicated absenteeism is actually reduced. in fact, studies attested that the benefits of child care facilities accrue more in the recruitment and retention areas than in reduced absenteeism and tardiness (shellenbarger, 1993). on-site child care facilities have positive benefits in addition to reduced absenteeism and enhanced recruitment and retention. according to thewsj, corporate child care centers yielded favorable publicity. in a study of a child care center that opened in monterey park, california, media coverage of the event led to 27 newspaper and magazine articles, two 50 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 evening news spots and a radio program (shellenbarger, 1993). according to moskowitz (1996), 75% of the 1996working motheraward winners had at least one on-site or near-site child care center. other family-friendly benefits are less expensive, and also yield positive results. flexible work schedules, work-place seminars, job sharing programs, and child and elder care referral services are inexpensive and may yield significant benefits in improved morale, reduced absenteeism and tardiness, and lower turnover. a study by the families and work institute showed that an average parental leave of four and one-half months costs the firm approximately 32% of an employee’s annual salary, whereas replacing the employee costs between 75–150% of the employee’s salary. the study also found that offering a generous familyleave package for maternity leave reduced the percentage of mothers who fail to return to work from 24% to 12% (scott, 1993). flextime, according to several studies, is inexpensive and provides substantial benefits to firms. companies indicate that flextime reduces tardiness and absenteeism and costs nothing in administration and training expenses. one corporate example of the success of these policies is first tennessee national. their workers make heavy use of flextime, flexplace, and employee-involvement programs. the corporation shifted its focus in 1993, with the arrival of new ceo ralph horn, to fostering an environment that motivated and supported employees. mr. horn argued that satisfied employees and satisfied customers are inextricably tied. he talks at length about these policies to securities analysts. prior to mr. horn’s taking over first tennessee, the company’s shares traded below the industry p/e; following his plan’s implementation, they have traded above the industry average (shellenbarger, 1997). one of the most important corporate awards to arise in recent years is the annualworking mother’s “100 best companies for working mothers.” theworking mothercontest began in 1986 when the magazine awarded 35 companies with “best company” status. since 1986 the number of companies has grown along with the number of entrants vying for the award. the number of participants reached and exceeded 1,000 in 1993 and continues to grow (fierman, 1994). working motherbases its award on five factors. they rate firms on pay, opportunities for women to advance, child care assistance and other family-friendly benefits. in 1996, workplace flexibility became a separate category. specific policies such as on-site or near-site child care facilities, flexible work schedules, job sharing, reduced work options, compressed work weeks, paid paternity leave, adoption benefits, leave to care for the elderly, and others, are examples of family-oriented policies according toworking mother. the sample in this study is based on firms that made theworking motherlist in either the majority of the years between 1986 to 1996 (at least six out of 11 years) or in all five years between 1992 and 1996. the firms that make theworking mother top 100 receive a tremendous amount of publicity, and it is cited as one of the key reasons many firms submit an application (shellenbarger, 1993). this has led some applicants to fear making the list and then subsequently being dropped. sprint faced negative publicity after making the list. many sprint workers expressed disbelief that they were working for the same firm that had won the award. this was reported in several publications including thewsj.sprint was later dropped from the working mothersurvey. in a subsequentfortune article on less family-friendly 51d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 firms, a sprint executive, discussing being dropped from the list, said “we have to run a business” (fierman, 1994). in sum, there is significant interest in these firms in both the political environment and in the financial marketplace. we intend to contribute to the existing research by providing evidence relevant to the stockholders, or perspective stockholders, of these firms. the research question addressed in this study is whether investors in family-oriented firms earn returns less than, equal to or greater than the market and a matched sample of companies. 3. hypothesis the general hypothesis regarding the performance of family-friendly firms relative to traditional firms is that the stock market returns of the two portfolios will be different. there are two alternative outcomes. first, the family-friendly portfolio could outperform the non-family friendly portfolio and the s & p500. this scenario would support the theory that workers in these firms are more productive, more efficient, and more focused on their jobs. the benefits of these policies, such as reduced turnover and attracting highly qualified workers, would outweigh the costs of the programs. also, investors may support these firms by increasing the demand for their goods and/or by buying the firm’s stock. both would lead to an improved bottom line and potentially higher returns in the stock market. second, the returns of the family-friendly portfolio could underperform the market or a matched sample portfolio. this implies that the many costs of implementing these programs would overwhelm the benefits. a time lag most likely occurs between implementation of the programs and the necessary front-end costs and the ultimate benefits that accrue to firms from increased productivity and enhanced recruiting. thus, firms committed to creating a familyoriented environment for their employees may see profits, and possibly stock price, suffer, at least in the short run. finally, the null hypothesis is that the raw and risk-adjusted returns of the family-friendly portfolio are equal to those of the market or a matched sample of non-family oriented companies. in this case, the costs and benefits of family friendliness would offset each other or investors are not specifically seeking or avoiding family-oriented companies for their portfolios. this possibility is most consistent with finance theory, which states that only risk factors affect return. non-risk factors such as family-friendliness should not have an impact on raw or risk-adjusted returns. 4. data and methodology the sample firms are collected from theworking mothersurvey.working motherhas published the survey each year since 1986. we include all firms that are publicly traded and have appeared in the survey in at least six of the 11 years of between 1986 to 1996, or each of the five years between 1992 to 1996. several firms that were ranked byworking mother are privately held, and thus return data does not exist. based on these criteria, we have a sample of 29 firms. 52 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 monthly closing prices and dividends were collected for both samples from compustat pc plus for the years 1987 through 1996. firms were matched on three criteria. first, the matched sample firms must never have appeared onworking mother’slist. firms were then matched based on industry classification and on market capitalization. table 1 includes a list of the family-friendly firms, the matched sample firms and the detailed matching criteria for each matched pair. total returns were calculated (both dividends and capital gains). annual returns were calculated using the geometric mean of the monthly returns of the entire portfolio. the monthly returns are equally weighted. ten annual holding-period returns and seven multi-year holding period returns were calculated. six, five-year averages were calculated, and one 10-year average over the entire sample period 1987 through 1996. we calculated raw returns and three risk-adjusted measures. compound annual family-friendly portfolio returns, matched sample returns and market raw returns are presented in table 2 for the single year and multi-year holding periods. a paired difference test was used to calculate a student’s t-test statistic test for differences in mean sub-sample returns. sharpe (1946, 1994) developed, and later clarified the use of, a risk-adjusted measure that measures return per unit of total risk. it is an appropriate measure of risk-adjusted return when the investor is not well diversified and is exposed to company specific risk. it is known as the reward-to-variability ratio and is calculated: sharpe index5 d1/sd1 3 =12 where d1 5 mean monthly difference between the portfolio or market return and the t-bill return, calculated over the appropriate holding period (12, 60, or 120 months), and sd1 5 the sample standard deviation of the monthly return differences. table 3 presents the sharpe index results. in addition, the treynor index, developed by treynor in 1965, measures return per unit of systematic risk. it is an appropriate measure of risk-adjusted return if the investor is well diversified and is not exposed to company-specific risk. it is calculated: treynor index5 d1/b where: d1 5 the mean monthly difference between the portfolio or market return and the t-bill return, calculated over the appropriate holding period (12, 60, or 120 months), and b 5 portfolio beta, or market beta (bm 5 1) betas are calculated based on monthly returns using a simple regression of the market returns and the portfolio returns. they are included in table 4 along with the treynor index results. in a 1968 paper, jensen developed a third risk-adjusted measure called jensen’s alpa. alpha indicates whether a portfolio exhibits above-average risk-adjusted returns. a positive (negative) alpha indicates that the portfolio consists of undervalued (overvalued) securities and is calculated by regressing the portfolio’s monthly risk premium on the market’s monthly risk premium. the regression equation appears below: rp 5 a 1 b(rb) 1 ei where: rp 5 excess return to the family-friendly portfolio ai 5 jensen’s alpha b 5 beta coefficient rb 5 excess return on the benchmark portfolio (family-friendly or market portfolio), and ei 5 error term table 5 presents the jensen’s alpha results. 53d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 table 1 family-friendly firms matched sample family friendly firm market value (millions) sic code matched sample firm market value (millions) sic code justification aetna, inc. 17085 6321 general re corp 16599 6331 sic two digit match, cap match american express 39270 6199 banc one 32637 6021 cap match, fm sic match (financials) apple computer 2215 3571 tandem 3596 3571 sic match, cap match at&t 59798 4813 sbc communications 53992 4813 sic match, cap match avon 9585 2844 clorox 7225 2842 sic three digit match, cap match barnett banks inc 10096 6022 fifth third bank 10037 6022 sic match, cap match ben & jerry’s homemade ice cream 93 2024 tcby 164 2024 sic match, cap match cigna 14307 6331 loews corp 12434 6331 sic match, cap match citicorp 62071 6021 bankamerica 52730 6021 sic match, cap closest (nationsbank better, but already on list) corning, inc. 14268 3220 armstrong world industries 3008 3089 fortune, fm sic match, closest cap match available dow chemical 22164 2821 monsanto 29165 2800 sic two-digit match, cap match du pont (e.i.) de nemours 75762 2820 american home products 53030 2834 sic two-digit match, cap (closest) eastman kodak 22043 3861 fuji photo film 21679 3861 sic match, cap match gannett 14091 2711 tribune inc. 6498 2711 sic match, closest cap genentech 7227 2834 rhone-poulenc rorer 12938 2834 sic match, cap match glaxo wellcome 74752 2834 smithkline 52211 2834 sic match, cap match plc beechham, plc both based out of england general motors 46618 3711 chrysler 25083 3711 sic match, closest domestic cap hewlett–packard co. 71073 3570 compaq 42941 3571 sic match, cap closest ibm 105026 3570 intel 150388 3674 cap match, fm sic match (business equipment) johnson and johnson 82766 2834 pfizer 75002 2834 sic match, cap match lincoln national company 7328 6311 sunamerica 7227 6311 sic match, cap match merck & co. 125640 2834 bristol myers squibb 78302 2834 sic match, cap match (closest available) mmm 39968 2670 kimberly clark 28359 2621 sic two-digit match, closest cap (continued on next page) 54 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 5. empirical analysis table 2 presents the annual raw rates of return for the family-friendly portfolio, the matched sample portfolio and the market index. multi-year holding period returns are also included in table 2. although the family-friendly firms outpaced the s & p 500index in seven out of 10 years, there are no statistically significant differences between the annual returns of the two portfolios. when comparing the family-friendly portfolio to the more appropriate matched sample portfolio, the differences in returns are significantly (p , .05) different from zero in only one year. in 1989, the matched sample returned 44.87% whereas the family-friendly portfolio returned 30.73% to shareholders. the family-friendly portfolio had higher returns than the matched sample portfolio in only two out of 10 years. for all six, five-year holding periods and for the overall 10-year holding period, the family-friendly portfolio had higher returns than the s & p 500index but had lower returns than the matched sample portfolio. the family-friendly portfolio outperformed the s & p 500, statistically at the 5% level, in one multi-year holding period between 1991 and 1995. the returns were 22.63% and 16.57%, respectively. in one five-year period, 1987 through 1991, and in the overall 10 year period, 1987 through 1996, the family-friendly portfolio statistically underperformed the matched sample portfolio. the matched sample’s return of 23.60% was significantly (p , .01) different than the 17.14% return for the family-friendly stocks over the total 10-year holding period. the matched sample also significantly (p , .05) outperformed the family-friendly portfolio in the period 1987 through 1991, 23.29% compared to 15.92%, respectively. overall, based on seventeen return comparisons between the family-friendly portfolio and the matched sample, there is statistical significance in only three periods. in the seventeen return comparisons between the family-friendly portfolio and the market portfolio, the difference in returns is significant in only one period. table 3 presents the sharpe index results. there was an even split between the familytable 1 (continued) family friendly firm market value (millions) sic code matched sample firm market value (millions) sic code justification motorola 47849 3663 sony 39115 3651 sic two-digit match, cap match nationsbank 51794 6021 morgan stanley, dean witter, and co. 50731 6211 cap match, fm sic match (financials) pitney bowes 10950 3579 seagate technology 10100 3572 sic three-digit match, cap match procter and gamble 103308 2840 philip morris 109273 2111 cap match, fm article (drugs) unum 6328 6321 conseco, inc. 7479 6321 sic match, cap match xerox corp. 26621 3861 emerson electric 26537 3823 sic two digit match, fm sic match (business equipment) cap match 55d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 friendly portfolio and the s & p500. each portfolio had the higher index in five out of 10 individual years. however, the family-friendly portfolio had the higher index in five out of seven multi-year holding periods. conversely, the sharpe index of the family-friendly firms underperforms the matched sample firms in eight out of 10 years and in six out of seven multi-year holding periods. table 4 presents the treynor index results. based on return per unit of systematic risk, the family-friendly returns are higher than thes & p 500 in seven of the 10 individual years and in all seven multi-year holding periods. in contrast, the family-friendly sample has a higher index in only two out of 10 years and in zero out of seven multi-year holding periods. differences between the sharpe and treynor risk-adjusted measures for comparisons of the family-friendly portfolio and the s & p 500portfolio are attributable to the level of diversification within each portfolio. specifically, the sharpe index uses total risk in the calculation of risk-adjusted returns. since the family-friendly portfolio is obviously less table 2 family-friendly investment strategy comparison of compound returns 1987–1996 year family-friendly (percent) matched sample (percent) t-test comparison of means (friendly versus matched sample) market index (percent) t-test comparison of means (friendly versus market) 1987 10.65 19.27 21.48 5.23 0.97 1988 9.62 18.90 21.68 16.81 21.73 1989 30.73 44.87 22.16a 1.49 20.13 1990 29.05 23.76 20.91 23.17 20.57 1991 45.12 44.07 0.03 30.55 1.70 1992 17.74 20.48 20.30 7.67 1.50 1993 10.89 15.29 20.60 9.99 0.17 1994 4.62 3.94 0.11 1.31 1.10 1995 39.93 46.19 20.75 37.43 0.48 1996 21.65 38.37 21.99 24.49 20.40 multiple year holding periods 1987–91 15.92 23.29 22.61a 15.36 0.38 1988–92 17.37 23.54 21.96 15.89 0.60 1989–93 17.64 22.78 21.58 14.50 1.10 1990–94 12.51 14.89 20.79 8.68 1.38 1991–95 22.63 24.91 20.71 16.57 2.34a 1992–96 18.38 23.90 21.60 15.47 1.17 1987–96 17.14 23.60 23.21b 15.41 1.03 summary table of raw return results type of comparison friendly versus matched sample—superior returns friendly versus market index—superior returns single year friendly—2 matched sample—8 friendly—7 market index—3 multiple year friendly—0 matched sample—7 friendly—7 market index—0 a statistically different (p , .05). b statistically different (p , .01). 56 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 diversified than the s & p500, it is exposed to a relatively higher level of total risk. the treynor index only considers market risk, eliminating the effects of company-specific risk that would be present in the family-friendly portfolio. thus, other things held constant, a well diversified portfolio (such as the s & p500) has an advantage over a less diversified portfolio (such as the family-friendly firms) when using the sharpe index. if an investor held only the family-friendly portfolio, the sharpe measure of risk-adjusted returns would be most appropriate. if the family-friendly portfolio is one of many held by the investor, the treynor index is the best measure of risk-adjusted return. table 5 presents the jensen’s alpha results. two different measures of alpha are used. first, jensen’s alphas are calculated for the family-friendly portfolio against the matched sample. consistent with the raw return data, negative alphas are observed from 1991 forward, except for 1996. in multi-year holding periods, jensen’s alpha values are negative for each five-year holding period until the 1991 through 1995 period. only for the five-year holding period from 1987 through 1991 do we observe a statistically significant alpha value, and it is negative. the negative coefficient implies that the family-friendly portfolio was overpriced compared to the matched sample benchmark during 1987 through 1991 time period. the family-friendly portfolio recorded positive jensen’s alpha values against the s & p 500 in table 3 family friendly investment strategy sharpe index measures year family friendly matched sample market index 1987 0.311 0.535 0.143 1988 0.322 0.921 0.978 1989 1.441 1.949 1.643 1990 20.586 20.333 20.503 1991 1.981 1.638 1.432 1992 1.422 1.465 0.568 1993 1.204 1.416 1.120 1994 0.113 0.057 20.193 1995 4.423 7.032 5.147 1996 1.320 2.447 1.447 multiple year holding periods 1987–91 0.500 0.764 0.518 1988–92 0.695 0.939 0.715 1989–93 0.770 0.965 0.693 1990–94 0.546 0.642 0.358 1991–95 1.533 1.518 1.172 1992–96 1.400 1.782 1.203 1987–96 0.728 1.013 0.699 summary table results for sharpe index measure type of comparison friendly versus matched sample—superior returns friendly versus market index—superior returns single year friendly—2 matched sample—8 friendly—5 market index—5 multiple year friendly—1 matched sample—6 friendly—5 market index—2 57d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 seven of the 10 one-year holding periods and in all but one multi-year holding period (a zero alpha was observed for 1985 through 1989). however, only in the five-year period from 1991 through 1995 was a statistically significant alpha value recorded against the s & p 500. jensen’s alpha values are consistent with the raw return data, indicating that the familyfriendly firms do not significantly outperform their benchmarks during the time studied. our results indicate that the raw returns of the family-friendly firms are not significantly different from the returns of a matched sample or of a market portfolio. these findings suggest that the market does not price family-friendly characteristics. given that finance theory suggests returns reflect risk characteristics, this finding is not surprising. 6. summary and conclusions our results are consistent with both the results presented in allen and kask (1997) and in hamilton, jo, and statman (1993). we find that investors do not necessarily “do well by doing good.” investors supporting family-oriented firms will not earn lower returns accordtable 4 family friendly investment strategy treynor index measures year family friendly matched sample market index beta value (family friendly) beta value (matched sample) 1987 9.680 17.809 4.380 1.075 0.992 1988 3.429 12.015 9.920 1.053 0.799 1989 18.304 30.650 20.300 1.086 0.833 1990 211.382 28.445 29.240 1.255 0.930 1991 33.820 33.489 22.610 0.995 0.796 1992 13.798 15.851 4.210 0.971 0.591 1993 10.509 11.761 6.870 0.734 0.466 1994 1.231 0.670 22.030 1.033 0.842 1995 29.599 33.223 26.880 0.973 0.656 1996 18.118 27.936 17.177 0.814 1.809 multiple year holding periods 1987–91 9.652 17.104 9.594 1.103 0.921 1988–92 10.042 16.712 9.560 1.120 0.851 1989–93 10.843 16.661 8.050 1.113 0.840 1990–94 7.475 10.665 4.484 1.117 0.845 1991–95 17.168 18.999 11.708 0.988 0.762 1992–96 14.694 17.888 10.621 0.897 0.714 1987–96 11.198 13.456 10.108 1.064 0.886 summary table of treynor index measures type of comparison friendly versus matched sample—superior returns friendly versus market index—superior returns single year friendly—2 matched sample—8 friendly—7 market index—3 multiple year friendly—0 matched sample—7 friendly—7 market index—0 58 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 ing to these results but will not outperform the market or a similar, non-family-friendly portfolio. the overall finding that the returns are not significantly different could also suggest that the costs of family-friendliness offset the benefits. many firms have only recently implemented these benefits. there can be significant up-front costs, depending on the type of benefits offered to workers. the payoff from many of these benefits, such as reduced turnover and improved recruiting, may slowly overcome the costs of implementation. additional research should focus on both the types of programs a firm has in place along with the amount of time the firm has been implementing family-oriented programs. the families and work institute has classified 188 firms into four stages beginning with firms who have very limit work/family programs to firms who have a holistic approach to the issues and a commitment to change the company culture (solomon, 1994). these stages could be used to assess the timing of costs and benefits accruing to firms instituting family-friendly programs. if firms are in the fourth stage of development (the most advanced stage), one could expect that the payoff from implementing these programs would have table 5 family friendly investment strategy jensen’s alpha measures year alpha (family friendly against matched sample) alpha (family friendly against market index) 1987 2.61 .47 1988 2.50 2.56 1989 2.47 2.18 1990 2.54 2.22 1991 .59 .93 1992 .34 .78 1993 .18 .22 1994 .06 .28 1995 .54 .21 1996 2.66 .06 multiple year holding periods 1987–91 2.43* .00 1988–92 2.25 .04 1989–93 2.16 .17 1990–94 2.06 .28 1991–95 .21 .45* 1992–96 .04 .30 1987–96 2.30 .10 summary table results for jensen’s alpha type of comparison friendly versus matched sample—superior returns friendly versus market index—superior returns single year friendly—5 matched sample—5 friendly—7 market index—3 multiple year friendly—2 matched sample—6 friendly—7 market index—0 * statistically different (p , .05). ** statistically different (p , .01). 59d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 caught up with or overcome the costs. in the earliest stages of development, the costs of implementation might still be quite significant, depending on the types of benefits put in place. this could help explain return differences between portfolios, if the firms were differentiated by stage of development. this differentiation could lead to an increased understanding of the returns to stockholders. references allen, g. c., & kask, s. b. (1997). socially responsible firms: financial and market performance. j bus econ perspect, 23(2), 86–96. clayman, m. (1994). excellence revisited.finan anal j, 51, 61–65. clayman, m. (1987). in search of excellence-the investor’s viewpoint.finan anal j, 43,54–63. fierman, j. (1994). are companies less family friendly?fortune, 129:64–67. gorman, r., filbeck, g. & preece, d. (1997). fortune’s most admired firms: an investor’s perspective.stud econ finan, 18(1), 74–93. families work institute web site. (1998). www.workandfamilies.org. conference board. (1994). firms share duties in work/family assistance. survey by the conference board. employee benefit plan rev, 49,38–40. kolodny, r., laurence, m. & ghosh, a. (1989). in search of excellence for whom?j portf manag, 15(3), 56–60. luciano, l. (1992). the good news about employee benefits.money, 21(june), 90–112. moskowitz, m. (1996). 100 best companies for working mothers. working mother magazine, 19, 10–70. peters, t., & waterman, r. (1982).in search of excellence: lessons from america’s best run companies.new york: harper & row. scott, m. (1993). downsizing firms place extra value on work and family programs.employee benefit plan rev, 48, 28–31. shalowitz, d. (1992). work/family benefits need not be costly to succeed.business insurance, 26, 10–11. shellenbarger, s. (1993). lessons from the workplace: how corporate policies and attitudes lag behind workers’ changing needs.hum res manag, 31(3), 157–169. shellenbarger, s. (1991). more job seekers put family needs first.wall street j, b1. shellenbarger, s. (1993). data gap: do family-support programs help the bottom line? the research is inconclusive.wall street j,june 21, r6. shellenbarger, s. (1993). concerns fight to be called best for moms.wall street j, september 14, b1. shellenbarger, s. (1993). best list for working mothers shows new faces: midwest firms, oil giants.wall street j, september16, a8. shellenbarger, s. (1997). investors seem attracted to firms with happy employees.wall street j,march 19, b1. solomon, c. (1994). work/family is a delicate balance.personnel j, 82, 72–87. washington report. (1992). traditional workplace practices are changing.office 116, 22. woolsey, c. (1992). continued growth seen in work-family benefits.business insurance 26, 13–14. 60 d.c. preece, g. filbeck / financial services review 8 (1999) 47–60 pii: s1057-0810(97)90017-2 book, software and web site reviews 223 state regulatory agencies that apply to financial services regulation are also discussed. a perceptive look at the changes in the u.s. financial markets caused by the growth and expanded role of mutual funds provides stimulating insight. an explanation of functional regulation, a structure where financial activity is regulated by the same regulator regardless of the type of financial institution conducting the activity may serve as a preview of future regulatory strategy. part viii’s discussion is concluded by contrasting the various regula tory methods used to regulate risk in the financial services industry. wealth management. harold evensky chicago, il: irwin professional publishing; 1997 (isbn o-7863-0478-2) reviewed by: dale l. domian, associate professor of finance, memorial university of newfoundland. harold evensky’s wealth management provides a comprehensive and informative guide to financial planning. as noted by pahl (1996), there are several key steps to the financial planning process, including gathering data to identify the client’s objectives, developing the financial plan, and implementing and monitoring the results. while some planners may focus on specific areas such as tax or estate planning, others take a broader perspective to include all aspects of their clients’ financial lives. wealth management is directed toward the latter group, presenting a broad holistic approach for practitioners who truly are “wealth managers.” an underlying theme is that investment recommendations should be consistent with results from academic and professional research. evensky champions the work of brinson, hood, and beebower (1986), who show that 94 percent of the variation in portfolio returns is attributable to asset allocation, with just 4 percent due to security selection and only 2 percent from market timing. the book also recommends that projections of real returns and risk premiums be based on historical data, such as the well-known ibbotson series. a wealth manager can then determine an asset allocation which minimizes risk, while at the same time providing a sufficiently high expected return to meet the client’s needs. the book is organized in 16 chapters which progress through the key steps of the wealth management process. the first five chapters focus on the importance of a two-way exchange of information between wealth managers and clients. evensky provides his per sonal insights on gathering information from clients, and on explaining important finance principles so that clients will have realistic goals on what is obtainable. the section on taxes includes a discussion of portfolio tax management, refuting claims that substantial gains are available from active tax management strategies. chapters 6 and 7 provide a concise overview of statistics and portfolio theory. unfor tunately, several formulas contain errors, and some of the explanations are imprecise. readers without previous training in investment mathematics could be better served by the more detailed treatment in most investment textbooks (e.g., sharpe, alexander, and bailey [ 19951). asset allocation models and the development of investment policies are considered in chapters 8 through 10. the use of mathematical optimizers is recommended, with the 224 financial services review 6(3) 1998 wealth manager providing constraints on minimum and maximum proportions for each asset class. the end result can be presented to the client as part of a detailed written finan cial plan. wealth managers typically implement their recommendations through mutual fund investments. chapters 11 through 13 discuss selection and evaluation of funds and their managers. some compelling evidence is presented for passive management techniques such as indexing. however, evensky observes that there are pragmatic reasons for includ ing some actively managed funds. many clients receive psychological benefits from attempting to beat the market, and may be unwilling to pay fees year after year to wealth managers who construct all-passive portfolios. acquiescing to this could be a bit unsettling to academics, but is probably necessary for practitioners working in the real world. the final three chapters of the book consider special topics. chapter 14 uses examples from evensky’s own firm to illustrate some of the business aspects of running a wealth management practice. this provides many useful ideas for planners at all levels of experi ence. chapter 15 discusses fiduciary investing in accordance with the employee retire ment security act (erisa) of 1974, the restatement of trusts third in 1992, and the uniform prudent investor act (upia) of 1994. readers unfamiliar with these acts will be pleasantly surprised to find that modem fiduciary law is firmly grounded in investment the ory. concluding remarks are made in chapter 16, again drawing on the author’s experience to present the philosophy of his own wealth management practice. wealth management is an excellent book which can be enthusiastically recommended to a wide range of readers. it provides first-rate investment advice for its intended audience of financial planners and other practitioners involved in managing multiple asset class investments. in addition, the book is also very suitable for the academic readers of finan cial services review; most already know the finance theories but may have little experi ence at addressing the concerns of financial planning clients. the practical knowledge from this book can give new insights on directions for research in personal financial man agement. references brinson, g.p., hood, r., & beebower, g.l. (1986). determinants of portfolio performance. finan cial analysts journal, 42, 39-44. pahl, d. (1996). an emerging partnership: afs and the cfp board. financial services review, 5, 71-21. sharpe, w.f., alexander, g.j., & bailey, j.v. (1995). investments, 5th edition. englewood cliffs, nj: prentice-hall. pii: 1057-0810(92)90005-w financial services review, 2(2): 97-l 10. copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. performance and risk exposure of international mutual funds larry r. lang robert m. niendorf this study examined whether internationally diversified mutualfinds increase a u.s. investor’s risk-adjusted return above that on a domestic benchmarkmutualfund. average returns on about one-half of the international funds exceeded the domestic benchmark fund’s return. the risk-adjusted returns on the international mutual funds were not significantly dtrerentji-om that on the domestic benchmark fund. these results dtrerfiom earlier studies which generally found superior returns on international mutual funds. the benefits for the u.s. investor of holding an internationally diversijied mutual fund appear to be limited for the period studied. i. intr~duc~~n over the past several years both business periodicals and the popular press have carried articles extolling the virtues of investing outside the united states. not only has the number of articles increased, but the target audience also has expanded. the typical article often concludes that there are likely to be potential benefits from investing in firms that have the majority of their operations outside the united states. many articles urge investors to consider the possible gains that might come from investing outside the domestic market. while both fixed return investments and common stocks are covered in the literature, the latter receive much more coverage and emphasis. with the proposed increased integration of the major economies of europe at the end of 1992, integration of financial markets also will likely increase. freer capital flows among the markets of the world is likely to be one outcome. even before europe 1992 was proposed, european investors considered external invest ing the norm rather than the exception. more limited domestic financial markets, coupled with a willingness to consider external financial markets, may explain part larry r. lang and robert m. niendorf l college of business administration, university of wisconsin, oshkosh, wisconsin 54901 98 financial services review, 2(2) 1993 of this more global investment horizon. until recently, however, such a trend has been much less evident in the united states. only in the past several years have small investors expanded their investment horizon beyond the domestic area to channel funds into external markets. domestic markets that not only are large, but also offer a wide array of options, have likely discouraged some small investors in the united states from seeking international investment options. at the same time, until recently the range of investment options that would allow small investors to participate in foreign markets has been somewhat limited. but the past several years have brought a rapid expansion in available international investment vehicles. current indications are that this growth will continue. the combination of the broadened array of international choices, and the increased awareness of the potential benefits from investing on an international scale, has encouraged small investors to expand beyond the domestic area into the broader international financial markets. at the same time the willingness of small investors to do this has encouraged financial intermediaries and service firms to develop still more intema tional investment vehicles. benefits from investing in financial markets other than the united states some earlier work (grubel, 1968; levy, 1970; swanson, 1979; eun, 1985, 1987; grauer, 1987) found that the rates of return on selected major equity markets outside the united states often were greater than the return on the united states’ market. during periods in the late 1970s and early 1980s the returns in major foreign markets often were higher than those in the domestic market. risk exposure results from those same studies tended to be more mixed. while risk in some markets was lower than in the united states’ domestic market, in a considerable number it was higher. the findings suggest that investors might be able both to enhance their returns and to lower risk by investing outside the united states. international mutual funds: an ideal vehicle for small investors investors with modest resources are likely to find that a mutual fund is the ideal vehicle for accessing the world’s major equity markets. first, the mutual fund should have better access to essential financial and economic data on individual firms, on the industries where those firms operate, and on those economies that have a pivotal role in the firm’s operations. further, the fund’s professional management group is more likely to have the expertise to interpret that data. at the same time, a moderate to large fund will be able to spread the costs of acquiring this data over its sizable investor base. combined, these factors should help to overcome one challenge to investing in international markets-obtaining and interpreting financial data. clearly, most small investors would be hard pressed to duplicate the fund’s access and interpretation expertise. performance & risk exposure of international mutual funds 99 second, because the mutual fund’s typical transaction will entail a much larger amount, economies of scale may allow it to reduce transaction costs. at the same time, volume purchases and sales may allow it to access some wholesale markets to further lower costs. one or both should drop a fund’s transaction costs below what small investors would pay to invest a modest amount directly in the foreign equity market. these cost savings can help to overcome a second challenge to investing in international markets-high transaction costs. finally, the mutual fund may be better able to offset the reduced liquidity that characterizes some foreign equity markets. with its diversified list of holdings, it will have a wide range of financial assets that might be liquidated to raise needed cash. also, assuming that investors continue an ongoing series of deposits, that cash inflow could meet ordinary redemption requests. overall, mutual funds appear to be an attractive way for small investors to access foreign equity markets. growth in the number of internationally diversified mutual funds over the past five years the number of internationally diversified mutual funds offered by united states based fund groups has expanded by nearly 50 percent. clearly, small investors now have many more options to expand their investment horizon beyond the united states. recent developments suggest that while the introduction rate for new funds has slowed, an investor’s choice is continuing to expand. since a broad international perspective is our concern, we exclude the rapid growth in highly specialized international funds such as those that concentrate either on a limited geographic area or on a single industry. performance of international mutual funds: earlier studies in an earlier study, mcdonald (1973) examined the performance of french mutual funds to determine if investing on an international scale was beneficial. for the period studied, he concluded that investing in financial markets other than the domestic one did offer definite benefits to a fund’s shareholders. not only was performance lowest for funds concentrating exclusively on french firms, but performance also rose with the level of investment in non-domestic markets. a later study by proffitt and seitz (1983) concentrated on internationally diversified mutual funds based in the united states. their results suggest that from 1974 to 1982, each international mutual fund in the sample outperformed the standard & poor’s 500 index during the period 1974 to 1982. furthermore, when those returns were adjusted for risk, the differences were statistically significant at a reasonable level of confidence. both studies suggest that investors would do well to consider internationally diversified mutual funds because their past superior risk-adjusted return may continue in the future. 100 financial services review, z(2) 1993 ii. purpose of this study this study examines whether u.s. investors could have improved their rate of return by purchasing an internationally diversified mutual fund to expand outside the domestic market. the study expands and extends on the earlier works by mcdonald (1973) and proffitt and seitz (1983). first, it covers an expanded and refined sample of international mutual funds. second, it extends the earlier works by examining the investment performance during the period of 1986 to 1990. if, indeed, financial markets are becoming more integrated, this recent period may reflect that. third, it overcomes a limitation of the earlier studies that forced the inclusion of some international funds that concentrated on a limited geographical area or on a single industry or product. fourth, the study uses a benchmark that is far more repre sentative of domestic equity mutual funds to judge the performance of the intema tional funds. the results will help answer the question: have internationally diversified mutual funds improved the returns of small u.s. investors by allowing them to expand outside the domestic market? research design the empirical portion of the study selected a sample of united states based internationally diversified mutual funds and compared their performance to that of a domestic benchmark mutual fund. the sections that follow discuss the sample criteria, return measures, mutual fund benchmark, overall market benchmarks, time period, and performance measures. sample of internationally diversified mutual funds used in study to be included in the sample, a mutual fund’s investment objective had to include a commitment to invest a significant fraction of its portfolio in the market able securities of foreign-based firms. since the emphasis was on funds that small investors could readily purchase, only funds offered by u.s. based mutual fund groups were candidates. as a further restriction, only mutual funds with more than 50 percent equities in their portfolio of securities were included. a qualifying fund had to have been readily available to potential investors during the entire study period from january 1986 through december 1990. it also had to have published data on net asset value as well as dividend and capital gains distributions. several groups of funds were specifically excluded because they lacked broad international diversification: 1) single country funds (e.g., the japan fund, the united kingdom fund); 2) funds that concentrated on a narrow geographic area such as europe, or the pacific basin; 3) funds that limit themselves to a single industry, product, or service (e.g., gold funds); 4) funds that held primarily u.s. equities with only a token portfolio of foreign-based firms. the resulting sample included mutual funds with an investment objective that permitted a significant fraction of the fund’s portfolio to be invested in foreign-based performance & risk exposure of international mutual funds 101 exhibit 1. internationally diversified mutual funds global mutual funds in sample dean witter worldwide putnam international equities first investors global sogen international keystone international templeton global new perspective templeton growth paine weber classic atlas templeton world international mutual funds in sample alliance international scudder international europacific growth templeton foreign kemper international trustees’ commingled kleinwort benson international united international growth t. rowe price international stock marketable securities. the sample funds did not confine themselves to a narrow spectrum of foreign markets or to a specific industry. qualifying mutual funds were stratified into two major groups: global mutual funds and international mutual funds. global mutualfunds. global mutual funds included those where the invest ment objective permitted them to purchase securities in any of the major world markets, including the united states. funds were classified based on the summary of each fund’s investment objective as published in wiesenberger’s investment companies service and in morningstar’s mutual fund values. to qualify, it was not necessary for the fund’s actual holdings of united states-issued securities to have been any prescribed fraction of the portfolio during the study period. all that was needed was that the fund’s investment objective permitted a significant fraction of u.s.-based equities in the portfolio. the top section of exhibit 1 lists the funds in the global sample. international mutual funds. international mutual funds included those where the investment objective specifically restricted their holdings to marketable securities of non-united states issuers. again, the summary of each fund’s invest ment objective published in wiesenberger and morningstar was used to classify it as part of the international group. while a typical objective restricted the holding of u.s. equities, most allowed the fund to hold its cash reserve in domestic money market instruments. the lower section of exhibit 1 lists the funds in the international sample. none of the mutual funds changed its investment objective during the period, so none had to be dropped or reclassified to a different group. holding period returns paralleling earlier work, the study computed each fund’s rate of return using monthly holding periods. holding period returns (hpr) were computed using: hpr = wavend navtwg) + dwdis + capgmdis navtxg 102 financial services review, 2(2) 1993 where: navb, = the net asset value of the fund’s shares at the beginning of the period; nav,, = the net asset value of the fund’s shares at the end of the period; div, = dividend distribution on each fund share for the period; and capgan,, = capital gain distribution per period for each fund share. a similar measure has been widely used in other mutual fund performance studies. the principal source of nav’s was wiesenberger’s investment companies service; the wall street journal was a secondary source for a limited number of situations. moody’s dividend record and standard and poor’s quarterly dividend record provided data on each fund’s dividend and capital gains distributions, with moody’s being the primary source. total investment return all holding period returns computed in the study are expressed in dollars, since u.s. based mutual funds convert all results to dollars. those returns reflect not only the fund’s investment performance but also its currency exchange gains and losses. when a fund marks its portfolio to market each day its nav will include potential capital gains and losses on the underlying securities plus currency gains and losses when the value of those securities is converted to dollars. since this study focused on small u.s. investors who would not likely be hedging in currency markets, total investment return was the appropriate measure. benchmark for domestic mutual funds to judge the performance of internationally diversified mutual funds, the study required a benchmark that was indicative of a mutual fund operating in the domestic market. rather than using a domestic market index such as the s&p 500 as others have, we wanted to use a benchmark that was representative of a domestic mutual fund. this would show an investor’s return after deducting management fees and other administrative costs. while an index like the s&p 500 may capture the performance of the overall market, it is likely to overstate the return because the fees and costs for operating a mutual fund are not considered. rather than adjust an index with some “average” or “typical” costs, the study selected the vanguard index 500 mutual fund as the representative benchmark for domestic mutual funds. that fund’s investment objective specifically states that it seeks performance that paral lels the return on the s&p 500 index. a review of the fund’s performance during the study period suggests that after allowing for the fees and operating costs, the fund was a reasonable proxy for an “adjusted” s&p 500. since many domestic mutual funds failed to match the return of the vanguard index 500, its use as the performance & risk exposure of internaiional mutual funds 103 domestic benchmark certainly does not understate the performance achieved by domestic funds. benchmarks for the overall market two different indices were used as performance benchmarks for the major international equity markets. for the global mutual fund group, the study used the morgan stanley capital international world index. this index is denominated in u.s. dollars and includes equities traded on stock exchanges in 21 different coun tries. this market weighted index includes companies which encompass approxi mately 60 percent of the market value of the common stocks traded in those 21 countries. the u.s. market is included along with 20 others. it was considered the appropriate benchmark because an internationally diversified mutual fund can invest in both the u.s. and foreign markets. a recent study by cumby and glen (1990) suggests that this morgan stanley index is a reasonably efficient benchmark for the world portfolio of equities. because funds in the international category cannot invest in marketable securities, other than money market instruments, of united states issuers, that group needed a different benchmark. the widely quoted morgan stanley capital intema tional europe, australia, and the far east (eafe) index was chosen. this dollar denominated index includes common stocks that are traded on the exchanges of 18 different countries, excluding the united states. again, this market value weighted index includes companies which encompass approximately 60 percent of the total market value of shares for those 18 countries. based in part on the work by adler and dumas (1983), the rate on go-day united states treasury bills was used as the riskless rate for purchasers of the internationally diversified mutual funds. the excess return for each of the funds was computed using this riskless rate. time period of the study the study covered the period from january 1986 through december 1990. selecting the period required balancing the need for acquiring a reasonably large sample of internationally diversified funds and the goal of covering a representative period. because many international funds have a relatively short operating history, launching the study prior to the mid-1980’s sharply reduced the sample size. one distinct advantage of the five year period chosen is that it includes considerable volatility in the exchange rate of the u.s. dollar. unfortunately, it was also a period of limited declines in the united states equity market. it does include the sharp reversals of october 1987, the lesser correction of october 1989, and the general drop of late 1990. overall, however, it was a period of generally rising prices in the united states’ equity market. 104 financial services review, 2(2) 1993 returns and performance measures the main question examined in this paper is whether recent performance results suggest to a u.s. investor that investing in u.s. based international mutual funds can increase risk-adjusted returns. the global and international mutual funds included in this study were tested using both the sharpe and treynor measures to rank fund performance. performance of each fund’s management was also exam ined using the jensen measure. sharpe measure. as its first performance measure, the study used the sharpe index (1966) to rank the funds. that measure uses each fund’s total risk, as measured by the standard deviation of its returns, to adjust the fund’s hpr. values for the sharpe measure (s) were determined for each fund using s = (hpri rf)/sdi where: hpr, = average monthly return for fund i; rf = average risk-free return; and sd, = standard deviation of hpr,. this is an appropriate consideration for an investor whose holdings are limited to a single fund or a small number of funds. total risk is the appropriate emphasis. treynor measure. the treynor (1965) measure also was used to rank the performance of each of the global and international funds. treynor’s measure assumes an investor holds a well-diversified portfolio which eliminates unsystem atic risk and leaves only systematic risk. as such, the appropriate risk measure becomes the fund’s beta. values for the treynor measure (t) were determined using t = (hpri rf)/bi where: hpr, = average monthly return for fund i; rf = average risk-free return; and bi = the beta for fund i. jensen measure. to examine the performance of each fund’s management, the study used jensen’s (1968) alpha measure. that measure is used to indicate whether management has earned any excess risk-adjusted return for the fund. jensen’s measure of excess risk-adjusted return assumes a well-diversified portfolio, and thus considers only systematic risk. a significant positive excess risk-adjusted return could be due to management’s ability to take advantage of favorable timing decisions to improve the fund’s return or to management’s ability to select undervalued assets. therefore, either positive or negative excess risk performance & risk exposure of international mutual funds 105 adjusted returns can be used to indicate the superior or inferior ability of fund management. jensen’s alpha measure is computed using a regression equation which allows for a nonzero intercept. the equation used in this study is hpri, rf, = ai + b,(hpr,, rf,) + ui, where: hpr, = monthly return for fund i in period t; rf, = risk-free return in period t; a, = intercept measuring abnormal return for fund i; bi = beta for fund i; hpr,, = monthly return for the market index in period t; and ui, = error term for fund i in period f. iii. research rmums exhibit 2 presents the returns, risk, and performance measures for the global funds included in this study. average monthly returns for three of the 10 global funds were higher than for the domestic benchmark mutual fund, vanguard index 500. monthly returns for the vanguard index 500 fund and four of the 10 global funds exceeded that of the market index, morgan stanley world index. on strictly monthly returns, the global funds present a mixed picture relative to the domestic benchmark fund. the two risk measures, standard deviation and beta, for the global funds cover a wide range of values. standard deviations of average monthly returns for six of the 10 funds exceeded that for the domestic benchmark fund. global fund standard deviations ranged from 4.796% to 10.870%; the domestic benchmark fund standard exhibit 2. monthly performance measures for ten global mutual funds 1986-1990 monthly standard treynor sharpe jensen mutual fund return deviation beta r2 measure measure measure dean witter worldwide 0.689 4.796 0.827 0.821 0.158 0.027 -0.0002 first investors global 1.564 6.428 1.009 0.679 0.997 0.156 0.0088 keystone international 0.761 5.737 0.985 0.814 0.206 0.035 0.0019 new perspective 1.126 10.870 0.944 0.208 0.602 0.052 0.0082 pain web classic atlas 1.649 6.154 0.919 0.616 1.186 0.177 0.0098 putnam intl. equities 0.966 6.885 0.894 0.465 0.456 0.059 0.0027 sogen international 1.052 6.337 0.547 0.205 0.902 0.078 0.0048 templeton global 0.493 5.544 0.725 0.47 1 -0.091 -0.012 -0.0031 templeton growth 0.834 5.025 0.705 0.543 0.391 0.055 0.0010 templeton world 0.691 5.104 0.724 0.554 0.183 0.026 -0.0001 vanguard index 500 1.111 5.618 0.758 0.502 0.729 0.098 morg.stan.world index 0.984 5.252 1.000 1.000 0.425 0.081 treasury bills-90 day 0.559 0.102 -0.006 106 financial services review, 2(2) 1993 exhibit 3. monthly performance measures for nine international mutual funds 1986-1990 monthly standard treynor sharpe jensen mutual fund return deviation beta r2 measure measure measure alliance international 1.041 6.048 0.676 0.506 0.714 0.080 0.0008 europacific growth 1.353 5.073 0.581 0.53 1 1.368 0.157 0.0022 kemper international 1.140 5.342 0.669 0.636 0.869 0.109 0.0008 kleinwort benson intl 1.153 12.784 0.791 0.155 0.752 0.046 0.0035 t. rowe price intl stk 1.265 10.447 0.752 0.210 0.939 0.068 0.0029 scudder international 1.237 6.000 0.705 0.559 0.962 0.113 0.0002 templeton foreign 1.661 4.965 0.522 0.448 2.111 0.222 0.0059 trustees’ commingled 1.564 10.952 0.666 0.150 1.509 0.092 0.0081 united intl growth 0.795 4.613 0.586 0.654 0.403 0.05 1 -0.0025 vanguard index 500 1.111 5.618 0.417 0.223 1.324 0.098 morgan stanley eafe 1.296 6.363 1.000 1 .ooo 0.738 0.116 treasury bills-90 day 0.559 0.102 -0.002 deviation was 5.618%; and the morgan stanley world index standard deviation was 5.252%, which was lower than the domestic benchmark fund. global fund betas ranged from .547 to 1.009, with six of the 10 funds having betas greater than the domestic benchmark fund. only one global fund had a beta greater than or equal to 1.0. betas were determined for global funds by regressing the fund return on the morgan stanley world index. exhibit 3 presents the returns, risk, and performance measures for the intema tional mutual funds included in this study. average monthly returns for seven of the nine international funds were higher than for the domestic benchmark fund. the average monthly return on the market index used for international funds, morgan stanley eafe, also was greater than the return on the domestic benchmark fund. three of the international funds had average returns greater than that of the eafe index. on a straight return basis, the sample of international mutual funds appears to have outperformed the domestic benchmark. standard deviations of returns for the nine international funds ranged from 4.613% to 12.784%. five of those nine funds had standard deviations greater than the benchmark fund’s 5.618%. the average return on the morgan stanley eafe index had a standard deviation of 6.363%, which is greater than that of the domestic benchmark fund. only three of the nine funds had standard deviations greater than that of eafe. betas for the international funds were determined using the morgan stanley eafe market index. international fund betas ranged from .522 to .791. all nine of the intema tional funds had a beta coefficient greater than the domestic benchmark fund beta. risk-adjusted returns on internationally diversified mutual funds results shown in exhibits 2 and 3 indicate that unadjusted returns on intema tionally diversified funds compare favorably with the return on a well-diversified pedormance & risk exposure of international mutual funds 107 domestic fund, the vanguard index 500. the risk level of each fund varies widely, with some below the domestic benchmark and some above. the study then com pared the risk-adjusted returns for internationally diversified mutual funds with the risk-adjusted return for a well-diversified domestic mutual fund. the following hypothesis was formed: h,: the risk-adjusted performance of internationally diversified mu tual funds is not significantly different from the performance of well-diversified domestic mutual funds. h,: the risk-adjusted performance of internationally diversified mutual funds varies significantly from the performance of well diversified domestic mutual funds. sharpe and treynor performance measures are presented in columns 6 and 7 of exhibits 2 and 3 for each of the global and international mutual funds in the study. globulfunds. values of the sharpe measure for the global funds ranged from -.012 to .177 (exhibit 2). only two of the 10 global funds had a risk-adjusted measure greater than that of the domestic benchmark fund. the domestic benchmark fund had a sharpe measure greater than that of the morgan stanley world index. this indicates that the domestic benchmark fund provides a greater premium per unit of risk than does the market index. treynor measures for the global funds ranged from -.09 1 to 1.186. only three of the 10 global funds had treynor measures greater than that of the domestic benchmark fund. as with the sharpe measure, the domestic benchmark fund had a treynor measure greater than that of the market index. five of the 10 global funds had treynor measures which exceeded that of the market index. international funds. values of the sharpe measure for the international funds ranged from .046 to .222 (exhibit 3). four of the nine international funds had a risk-adjusted measure greater than that of the domestic benchmark fund. the market index, eafe, had a sharpe measure greater than that of the domestic benchmark fund. only two funds had sharpe measures which exceeded that of eafe. three of the nine international funds had treynor measures greater than that of the benchmark mutual fund. the treynor measure of the domestic benchmark fund was greater than that of the morgan stanley eafe index. wilcoxon matched pairs sign test assumptions about the distribution of the sharpe and treynor measures cannot be made, thereby making it necessary to test the null hypothesis using a nonpara metric test. the wilcoxon matched pairs sign test, a nonparametric test appropriate for testing differences in matched sample pairs, was applied using a .05 significance level. results of this test for the sharpe and treynor measures for both global and international mutual funds are presented in exhibit 4. 10s financial services review, 2(2) 1993 exhibit 4. comparison of internationally diversified fund with domestic benchmark mutual fund: wilcoxon matched pair sign test treynor sharpe wilcoxon statistic p-value wilcoxon statistic p-value global funds -1.6004 0.1095 -1.6004 0.1095 international funds -1.3624 0.1731 -0.0592 0.9528 the wiicoxon matched pairs sign test does not indicate a significant difference at a reasonable level between the risk-adjusted sharpe and tteynor performance measures of either the global or the international mutual funds and the domestic benchmark mutual fund measure. thus, the null hypothesis that the risk-adjusted performance of intemation ally diversified mutual funds does not differ significantly from that of domestic mutual funds cannot be rejected at any reasonable level of significance. abnormal fund returns by management the existence of excess risk-adjusted returns on each of the global and international funds in this study was examined using the following hypothesis: h,: the excess risk-adjusted return on the internationally diversified mutual funds is not significantly different from zero. h,: the excess risk-adjusted return on the internationally diversified mutual funds varies significantly from zero. jensen’s measure was used to test for the existence of excess risk-adjusted returns on internationally diversified mutual funds. to determine whether managers of these internationally diversified funds are superior or inferior in their perform ance, jensen’s alpha was computed for each fund. the excess risk-adjusted fund returns were regressed on the excess risk-adjusted benchmark return. it was ex pected that funds with superior managers would consistently have significantly positive alphas. similarly, significantly negative alphas would indicate inferior fund management performance. examination of the jensen measures shows that seven of the 10 global funds had a positive intercept (exhibit 2) while eight of the nine international funds had a positive intercept (exhibit 3). a t-ratio was used to test whether the computed jensen alphas differed significantly from zero. the t-ratios ranged from -.46 to 1.30 for the global funds, and from -.70 to 1.23 for the international funds. however, none of the alphas are statistically significant at a reasonable level. the null hypothesis is not rejected, indicating that managers of internationally diversified mutual funds do not significantly outperform those of the benchmark index domestic fund. performance & risk exposure of international mutual funds 109 diversification of funds the efficiency with which a portfolio is diversified can be measured by the coefficient of determination, r*, with the market fund. the closer this measure is to 1 .oo, the closer the diversification of the portfolio is to the market index. the only relevant risk for a well-diversified portfolio is the systematic risk which remains. some funds may decide not to diversify completely but instead to retain a portion of nonsystematic risk as part of the fund’s investment objective. because of these differing objectives, it could be expected that the funds might display a range of r* values. low r* values would identify funds which have elected to retain nonsystematic risk. r* values for the global mutual funds (exhibit 2) were based upon the morgan stanley world index as the market measure, and ranged from i321 to .205. low r* values suggests that some funds elected to retain a sizable unsystematic risk component. r* values for the international funds (exhibit 3) were computed using the morgan stanley eape index, and ranged from. 150 to .654. again, the low r* values and their wide range suggests that some funds elected to retain a sizable unsystem atic risk component. iv. conclusions this study addressed whether internationally diversified mutual funds can be expected to increase a u.s. investor’s risk-adjusted return above the risk-adjusted return available on a domestic benchmark mutual fund. an inspection of the monthly returns presented in exhibits 2 and 3 shows that the average return on about one-half of the internationally diversified funds does exceed the average return on the domestic benchmark fund. when the returns are risk-adjusted using either the sharpe or the treynor measure, the internationally diversified mutual funds do not provide a risk-adjusted return that is significantly different at a reasonable level from the domestic benchmark mutual fund. these results differ from mcdonald’s (1973) conclusions that investing in an external market provides superior return performance. the results of this study also differ from those of proffitt and seitz (1983) who found universally that international funds significantly outperformed the s&p 500, a domestic market index. at least for the time period of this study, the benefits for the u.s. citizen of investing through an international or global mutual fund appear to be limited. the coefficients of determination for the global and international mutual funds in this study are generally quite low, with r* values for the global funds generally greater than those for the international funds. the range of values among the various funds may be due to each fund’s pursuing a different investment objective. given the low r* values for the internationally diversified mutual funds, we believe that funds do not diversify away the unsystematic risk in their portfolios. investors in these funds therefore must be concerned with the total risk of the fund, not just the 110 financial services review, 2(2) 1993 systematic or market risk, when selecting a performance measure. consequently, using either the treynor or the jensen performance measure, which assume that the fund is well-diversified and that the only relevant risk is the market risk as measured by the fund’s beta, may not be appropriate for many investors. we expect that u.s. investors, many of whom have only recently diversified on an international scale, are unlikely to hold a large number of internationally diversified funds. as such they are not likely to diversify away the unsystematic risk of their holdings. given this retention of unsystematic risk by the fund and by the investors, a fund’s total risk is likely to be a more appropriate risk adjustment measure. sharpe’s performance measure, which uses the fund’s standard deviation, would consider total risk rather than just the market risk component. in our opinion, the sharpe performance measure is appropriate for use by the typical small investor. references 1. adler, michael and bernard dumas. 1983. “international portfolio choice and corporation finance: a synthesis,” journal of finance, (june): 925-984. 2. cumby, robert e. and jack d. glen. 1990. “evaluating the performance of international mutual funds,” journal of finance, (june): 497-521. 3. eun, cheol s. and bruce g. resnick. 1985. “currency factor in international portfolio diversification,” columbia journal of world business, (summer): 45-53. 4. eun, cheol s. and bruceg. resnick. 1987. international diversification under estimation risk: actual vs. potential gains. in s. khoury and a. gosh, eds., recent developments in interna tional banking and finance. lexington, mass.: heath, 135-147. 5. grauer, robert r. and nils h. hakansson. 1987. “gains from international diversification: 1968-85 returns on portfolios of stocks and bonds,” journal of finance, (july): 721-739. 6. grubel, herbert g. 1968. “internationally diversified portfolios: welfare gains and capital flows,” american economic review, (december): 1299-1314. 7. jensen, michael c. 1968. “the performance of mutual funds in the period 1945-1963,” journal of finance, (may): 389416. 8. joy, 0. maurice, don b. panton, frank k. reilly, and stanley a. martin. 1976. “comovements of international equity markets,” the financial review: l-20. 9. levy, hiam and marshall samat. 1970. “international diversification of investment portfo lios,” american economic review, (sepember): 668-675. 10. mcdonald, j. 1973. “french mutual fund performance: evaluation of internationally diversified portfolios,” journal of finance, (december): 1161-l 180. 11. proffitt, dennis and neil seitz. 1983. “the performance of internationally diversified mutual funds,” journal of the midwest finance association, (december): 39-53. 12. sharpe, william f. 1966. “mutual fund performance,” journal of business, (january): 119 138. 13. solnik, bruno h. 1974. “why not diversify internationally rather than domestically?’ financial analysts journal, (july/august): 48-54. 14. swanson, joel. 1979. investing internationally to reduce riskand enhance return, new york: morgan guaranty trust company. 15. treynor, jack. 1965. “how to rate the performance of investment funds,” harvard business review, (january/february): 63-75. pii: s1057-0810(00)00042-1 student learning style and educational outcomes: evidence from a family financial management course jonathan foxa, suzanne bartholomaeb aconsumer and textile sciences, the ohio state university, 1787 neil avenue, columbus, ohio 43210, usa bhuman development and family science, the ohio state university, 1787 neil avenue, columbus, ohio 43210, usa abstract the academic performance of 419 undergraduate students in an individual financial management class was evaluated in light of their learning style, demographic background, academic history and time allocation. academic history and time use variables proved to be the only significant predictors of grades in the course. student learning style, as measured by kolb’s learning style inventory, was not a strong predictor of success in this financial management class, and it appears that no single type of learner best grasps financial management concepts. the implications of these finding lead to a discussion of instructional methods. © 1999 elsevier science inc. all rights reserved. jel classification:a220 keywords:learning style; personal finance instruction; instructional strategies 1. introduction instructors of individual financial management hope to contribute to the intellectual development of every student in their class. in working toward this goal, most instructors provide a variety of opportunities to learn. however, class size, time constraints, and the broad range of complex topics inherent to most financial management classes present * corresponding author. tel.:11-614-292-4561; fax:11-614-688-8133. e-mail address:fox.99@osu.edu (j. fox). financial services review 8 (1999) 235–251 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00042-1 significant challenges to the instructor. given these restrictions, a major challenge for instructors is making informed choices on which activities to use to best demonstrate and transfer the intended information and skills. a promising approach for assessing the appropriateness of learning opportunities to provide for students is the use of kolb’s learning style inventory (lsi). the use of this tool may contribute to improved quality and effectiveness of teaching, student learning, and academic success. the first purpose of this paper is to introduce this tool, and the learning style theory on which it is based, to personal finance educators. a second purpose is to explore the relationship of student learning style and academic outcome in a personal finance course. while the link between learning style and teaching effectiveness has been established in other academic disciplines, research has not been conducted in the area of individual financial management. though recent studies such as vihtelic (1996) identify the effectiveness of teaching financial concepts in a framework of personal finance, no study has yet linked learning styles to student outcomes in personal finance. additionally, courses in financial management involve a wide range of both mathematical and verbal application, thus providing a good environment for evaluation of the complete teacher-student learning process. further, stitt-gohdes (1999) discussed the implications of learning styles in the context of business teacher education and felt student focused and individualized instruction to be increasingly important given the changing demographics of the student population, changes in technology-based instruction, and the increase in the population of academically at-risk students. by diagnosing student learning style, instruction can be individualized (dunn, 1984), therefore addressing these new challenges and providing an opportunity for increased effectiveness in instruction and learning in financial management classes. 2. review of literature and theories of learning while learning can be defined as an internal process that occurs when an observable, permanent change takes place (kaplan & kies, 1993), a learning style can be described as the way people retain or absorb information (de bello, 1990). more formally, learning style can be defined as a biological and developmental set of personal characteristics defined by the way individuals process information (dunn, beaudry, & klavas, 1989; dunn, 1984). research has shown students to be characterized by different learning styles (kolb, 1981). depending on the learning style, students focus on different types of information, perceive information differently, and understand at different paces (barbe & milone, 1981; claxton & murrell, 1987; felder, 1993; felder & silverman, 1988; kolb, 1984; schmeck, 1988). financial management instructors may be able to benefit from the notion that students possess individual differences in learning style. greater insight into the learning process can yield different approaches and outcomes for both instructor and student. research has found that presenting information through a variety of approaches leads to more effective instruction (doyle & rutherford, 1984; kavale & forness, 1987; mccarthy, 1990; o’neil, 1990; snider, 1990). learning style has proven to have an impact on the effectiveness of student learning, resulting from student response to different teaching methods. for example, heitmeyer and thomas (1990) found students to be more comfortable with certain instruc236 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 tional strategies than others as a result of their preferred learning style. additionally, research has shown improved attitudes, behavior, and grades when the instructional environment complements the student’s learning style preference (dunn et al., 1989; marshall, 1991). conversely, ignoring learning styles, thus treating students as a homogenous group of learners, may have a negative outcome. for example, marshall (1991) suggested that students who respond to nontraditional learning were at risk because their learning style was not accommodated in the traditional school environment. 2.1. kolb’s model of learning styles according to kolb (1981), learning is a circular process moving through four stages. the learning process proceeds with the learner taking a concrete experience (concrete experience), observing and reflecting upon it (reflective observation), forming an abstract concept and/or generalization (abstract conceptualization), and testing the concept in a new situation (active experimentation). the process is circular because the learner approaches another concrete experience thus restarting the learning process, this time with the newly acquired concept(s) from the previous learning cycle(s). kolb operationalized his learning theory by formulating two dimensions, perceiving and processing. concrete experience (ce) and abstract conceptualization (ac) at the opposite ends of the continua of the perception dimension represent feeling and thinking, respectively. individuals who prefer learning through concrete experience are generally adaptable to new environments, fully engage in the moment and task at hand, and excel at hands-on learning. learners who tend toward abstract conceptualization engage in problem solving, deductive reasoning and enjoy the practical application of concepts or ideas (kolb, 1981). active experimentation (ae) and reflective observation (ro) are the opposing extremes of the processing dimension. learners along this dimension process information either actively, by doing or reflectively, by watching. it is important to note that students in varying degrees utilize the four modes of learning. kolb’s learning style inventory (lsi) classifies students into one of four learning styles based on how they rank order nine sets of four words. scores from the lsi are plotted along the dimensions placing the learner into a quadrant based on two learning modes. the quadrants represent the following four learning styles: diverger (ce-ro), assimilator (roac), converger (ae-ac), and accommodator (ae-ce). 1. diverger. a student classified as a diverger perceives information through concrete experience and processes it through reflective observation. a diverger may best process information by their feeling and by observation. divergers do well with viewing a concept or idea from many perspectives. divergers are characterized as being emotional, people–oriented, and imaginative. they are good at working in a group and at blending many different experiences or pieces of information into a whole (kolb, 1981). an important financial management concept is the time value of money. a diverger may best grasp the future value of an annuity by working through a problem set with varying interest rates and duration rates (concrete experience) followed by the comparing and contrasting of the resulting future values (reflective observation). 237j. fox, s. bartholomae / financial services review 8 (1999) 235–251 2. assimilator. an assimilator perceives information through abstract conceptualization and processes information through reflective observation. the assimilator may gain more from an assignment requiring the construction of a model. assimilators are more systematic in their approach to ideas and theories. an assimilator prefers to digest and think about the information (kolb, 1981). an effective appeal to an assimilator would be to require them to devise a comprehensive personal financial plan for a fictitious client (abstract conceptualization) and to keep a log or journal (reflective observation) of each component or step of the process, reinforcing the grasping/learning process. 3. converger. a converger perceives information through abstract conceptualization and processes it by active experimentation. a converger approaches ideas and theories systematically, and ideally transforms information by applying the ideas and information to practical situations, such as laboratory experiments. convergers are less peopleoriented and more technically minded (kolb, 1981). a lecture in tax planning (abstract conceptualization) followed by preparation of a family’s tax return (active experimentation) is a good example of a learning structure tailored to convergers. 4. accommodator. students classified into the accommodator learning style perceive information through concrete experience and process it through active experimentation. accommodators may learn most effectively through a hands-on experience, may prefer to engage in an activity related to the topic, or use the information in trial and error exercises. accommodators learn from interactions with others, and can be characterized as being risk takers, and enjoying new challenges and experiences (kolb, 1981). learning about credit usage through analyzing the terms of their credit cards (concrete experience) followed by ordering their own credit report, or examining credit reports exemplifying bad credit (active experimentation) would appeal to an accommodator. 2.2. learning styles and educational outcomes this section reviews the major empirical findings in the learning style literature. first, the effectiveness and difficulties of identifying learning styles is discussed. second, the literature identifying learning styles that characterize particular disciplines is reviewed. finally, previous attempts to link learning style to academic outcome are discussed. studies have found effective instruction to be accomplished through multiple approaches (lacina, 1991), suggesting the importance of recognizing multiple types of learners. claxton and murrell (1987) recommend that instructional methods include all four learning style modes to give each student with a unique learning style the opportunity to do well a quarter of the time. they found students retained 20% of information if instruction appealed only to abstract conceptualization and 90% of the information if teaching strategies related to all four learning styles. dunn (1984) found that students whose learning style matched with the teaching method and environment earned better grades. further, research has shown students retain information longer, apply the information more effectively, and maintain positive attitudes toward the course content when teaching strategies and methods are compatible with student learning style (felder, 1993). based on kolb’s learning style theory, studies have found that learning style can be 238 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 matched with particular disciplines. kolb (1981) found the learning styles of over 800 managers and business graduate students to vary with their undergraduate major. accommodators tended to major in business; convergers in engineering; divergers in history, political science, psychology, and english; and assimilators in sociology, mathematics, chemistry, and economics. while studies have been able to successfully classify students in the traditional academic disciplines, research has yet to examine an interdisciplinary area such as family and individual financial management. business related studies are a close approximation. in a sample of business graduate students, bergevin (1993) found a convergent learning style to be dominant among finance and accounting majors, whereas marketing and management students were predominantly classified as experience-oriented (accommodators and divergers). studies examining kolb’s learning style as a predictor of academic performance emerge from a variety of disciplines. some, for example, garvey and bootman (1984), found significant associations between predominate learning style and overall grade point average. however, others found no significant differences in mean overall gpa, mean class gpa and learning styles (heitmeyer & thomas, 1990) and found learning style to be an inadequate predictor of academic success (leiden, crosby, & follmer, 1990). no empirical studies have found, or even tested, the correlation between kolb’s lsi and academic success in an individual financial management course. paulsen & gentry (1995) concur that the empirical work in finance education relating learning (in their study, learning strategies) and academic performance is sparse. their study found learning strategies (e.g., time, study, effort) and motivational factors (e.g., goal orientations, text anxiety, selfefficacy) in combination with aptitude variables explained over half the variance in academic performance in a finance class. however, a few studies have found a link between learning styles and success in business classes. controlling for gpa, aptitude, and motivation, togo and baldwin (1990) found that students with a convergent learning style, compared to nonconvergent, performed better in a financial accounting course. similarly, carthey (1993) examined a group of students enrolled in intermediate accounting, principles of economics, business law, and principles of management courses. he found that students who had a predominantly convergent learning style performed better (measured as the average final grade in the courses) compared to the other learning styles. compared to students who received higher average grades, divergers had the weakest performance. these findings suggest that students who effectively acquire the concepts and skills to be successful in financial management may also demonstrate a specific learning style as described by kolb’s lsi. in this study we examined whether learning style is predictive of academic success in an individual financial management course, while accounting for several demographic, academic, and time use factors. 3. method this study is based on data from 419 students enrolled in four introductory undergraduate family financial management courses at a large midwestern university. the sample was composed of predominantly white students (88%) between the ages of 18 and 57. charac239j. fox, s. bartholomae / financial services review 8 (1999) 235–251 teristics of the sample are outlined by learning style in table 1. additionally, p-values for tests of significance between learning styles and participant characteristics are included in table 1. a one-page questionnaire that included kolb’s lsi along with demographic, academic and time-use questions was administered along with a scheduled quiz. instructions regarding kolb’s lsi were explained to the class once the quiz was distributed. students completed the questionnaire as part of the class quiz but no credit was earned for completion and no penalty given to students choosing not to participate. 3.1. measures 3.1.1. educational outcome the dependent variable of interest, student educational outcome, was measured by performance in the course, indicated by the total points earned in the class. total points as a continuous variable was preferred to categorizing the dependent variable by grade received. total points attainable were 1000. the point range among the sample of 419 students was from 90 to 1000. the mean of total points for the class was 763, with students on average earning a c for the class. a significance test revealed no differences in grade by type of learning style. table 1 characteristics of students by learning styles variable all diverger assimilator converger accommodatorp-valuea number of students 419 186 55 37 141 0.000 average class grade 763 755 772 770 768 0.757 demographic average age 22.6 22.2 22.0 23.0 23.4 0.077 percentage of male students 47% 42% 41% 57% 52% 0.185 race: percent non-white 11.6% 14.5% 10.9% 8.1% 9.2% 0.429 mother attended college 60% 62% 58% 54% 60% 0.792 father attended college 68% 68% 73% 75% 63% 0.384 academic total credit hours 137 135 131 141 141 0.087 percentage with low gpa 22% 23% 24% 16% 23% 0.833 percentage with high gpa 24% 23% 24% 27% 26% 0.862 percentage in major 41% 38% 42% 49% 41% 0.682 required course 79% 76% 72% 73% 86% 0.097 term 1 25% 24% 22% 27% 27% 0.840 term 2 24% 27% 22% 19% 21% 0.487 term 3 18% 19% 29% 23% 14% 0.083 term 4 32% 29% 27% 33% 38% 0.341 time-use average hours employed 16.6 16.1 15.3 21.9 16.7 0.108 average credit hours this term 15.1 15.3 15.2 13.5 15.1 0.505 a for discrete variables, values were derived from chi-square tests. for continuous variables,f-values were derived from one-way analysis of variance. 240 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 3.1.2. learning style used in over 150 studies, kolb’s learning style inventory was the instrument used to measure student learning style. besides its strong theoretical foundation, the practical usage of the inventory was preferred, as it is short and easy to administer and score. kolb’s lsi describes four styles of learning based on a ranking by students of nine sets of four words. each word describes the student’s preferences for a learning style, with a four representing the most preferred and a one the least preferred. for example, a set of words from the lsi is “intuitive,” “productive,” “logical,” and “questioning.” scores from summing the ranks of six words are assigned for each of the subscales (ce, ro, ac, ae). the scores are plotted along two dimensions identifying a person as an accommodator, diverger, assimilator, or converger. the instrument has established validity, yet has come under some criticism for its reliability and stability (atkinson, 1991). specifically, kolb (1976) established convergent validity among the four subscales scores, represented by the correlations between the words making up the four subscales and the total score (correlations ranged from 0.46 to 0.67). additionally, studies have estimated the split-half reliability with a range of 0.37 to 0.81 (carrier & melvin, 1982; kolb, 1976). the split-half reliability coefficients for our sample were 0.73 (ae-ro) and 0.61 (ac-ce), well within the range of previous studies using this instrument. the breakdown of learning style categories did not follow the usual prediction by kolb. his research classified 25% of students as accommodators, 25% as divergers, 17% as convergers, and 33% as assimilators. interestingly, assimilators, predicted to be almost a third of the class, consisted of only 13% of the sample whereas a majority of participants were accommodators or divergers (34% and 44%, respectively). only a small percentage of the financial management students were convergers (9%). this distribution of learning styles is not a complete surprise given the similarity between the disciplines of family financial management and business, and kolb’s (1981) previous finding that accommodators majored in business more often than any other major. 3.1.3. demographic variables five variables measuring individual and background characteristics of participants were introduced as controls. previous studies have found age, gender, race, and parents’ education to influence academic achievement (bellico, 1972; borde, byrd, & modani, 1998; mutchler, turner, & williams, 1987; sewell & shaw, 1968; simpson & sumrall, 1979). most instructors will agree that older students are usually better personal financial management students, given their past experiences with many of the topics and tools. studies support this notion with older students performing better in a business finance course (simpson & sumrall, 1979). findings regarding gender were conflicting, with females performing better than males in an accounting class (mutchler et al., 1987) and males outperforming females in a finance class (borde et al., 1998). several studies have examined how parents’ education level affects academic achievement of college students. sewell and shaw (1968), in a landmark study, found that the higher the parents’ educational level the greater the success and graduation rate of college students. thus, it was hypothesized that greater levels of parent educational achievement would be positively related to performance in the course. overall, it was hypothesized that older students with parents who had more formal education 241j. fox, s. bartholomae / financial services review 8 (1999) 235–251 would perform better in the financial management class. the direction of race and gender were not predicted. the questionnaire asked participants their age, gender, race, and their mother and father’s educational level. for data analysis, student age was input directly and a dummy variable was used for gender (15 male, 05 female), race (15 nonwhite, 05 white), and mother and father’s education level (15 attended college, 05 did not attend college). the average age was 22.6 years, 47% of the sample was male, and 12% were nonwhite. over 60% of the participants had a mother with some college education, whereas 68% had a father with some college education. significance tests revealed no differences in the proportion of students falling into a particular learning style by age, gender, race, and parent education. 3.1.4. academic variables several academic factors were included in the model as possible predictors of student performance in the individual financial management classes. 3.1.5. credit hours it was hypothesized that students who had a greater number of overall credit hours would perform better in the class. students indicated the total number of credit hours they had earned to date. approximately 191 credits are required for graduation, the sample average was 137, indicating that the average student held junior or early senior status. a difference test approached significance (0.09 level of significance) for number of credit hours by learning style, with assimilators reporting somewhat fewer total credit hours. 3.1.6. grade point average previous academic performance, measured as grade point average, should predict current performance in the course. studies have established that previous student academic achievement predicts future performance (astin, 1971; eskew & faley; 1988). for example, past academic performance was a predictor of current performance in an accounting course (eskew & faley, 1988), an economics course (bellico, 1972) and a finance course (sachdeva & sterk, 1982). students in the current sample were given a range to select from for their gpa. the ranges were collapsed into three categories representing high, low, and middle grade point averages. low gpas, 22% of the sample, included students with below a 2.3 gpa on the four-point university scale. high gpas consisted of 24% of the sample and included students with at least a 3.3 gpa. there were no significant differences in the distribution of low and high gpa students among learning styles. 3.1.7. major the decided major of the participant was thought to impact course performance. those students who are majoring in personal financial management have self-selected, possibly identifying a greater interest and motivation in the course content. previous studies have found that students who declared their major performed better than students who had not established their academic goal or occupational choice (lavin, 1965). further, students 242 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 majoring in finance outperformed nonmajors in a business finance class (simpson & sumrall, 1979). students wrote in their major and were grouped into family financial management majors (1) and nonmajors (0). nearly 41% of the sample was comprised of family financial management majors. there were no significant differences between decided major and learning style type. 3.1.8. course requirement students who were in the course as an elective may have a different motivational level than those in the course because it is required in a major or program of study. students enrolled as an elective presumably hope to attain useful skills to be applied to their own personal financial situation. reasoning for the inclusion of whether the course was required is somewhat different than whether the student was simply a declared family financial management major. several majors require this course in individual financial management. students selected whether they were taking the course as a fulfillment of a general education curriculum requirement (coded as 1), as a major requirement (coded as 1), or as an elective (coded as 0). nearly 80% of the class took the course to fulfill some sort of requirement in a degree program. a difference test approached significance (0.10 level of significance) for the proportion of students taking the class as a requirement across learning styles with the highest proportion of students required to take the class being accommodators. 3.1.9. term students were exposed to a variety of instructional methods and activities depending on which term they were enrolled in the course (see appendix for details on instructional methods by course term). the sample is comprised of students enrolled in four separate offerings of the same introductory family financial management course with roughly 25% of the sample coming from each of four terms when the course was offered. the same instructor taught all four terms of the financial management course, increasing the reliability of the study, yet reducing the generalizability of the findings. studies have found teaching style to be associated with academic outcome (felder, 1993), in this case, teaching style was controlled. another factor impacting educational outcome is the availability of instructional technology. courses taught in earlier terms did not draw material from a course website. students enrolled in later courses (terms 3 and 4) could access lecture notes, problem sets, and quiz answers from the course website. a previous study found that educational outcome was negatively impacted by the availability of lecture notes, with students who missed lecture and substituted accessible lecture notes not performing as well (kelley, 1975). in term 2, the homework assignments changed from optional, not graded, to required and graded exercises. other significant differences between terms included the addition of a personal financial portfolio in term 2, which may have led to greater personal involvement with the course material. by term 3 it was thought that learning form personal experience through the portfolio outweighed the effectiveness of wall street journal reading assignments, and this assignment was dropped. also in term 3 the number of applied math problems given as required homework was increased. closed book quizzes were added in term 2 as a means of 243j. fox, s. bartholomae / financial services review 8 (1999) 235–251 enforcing reading assignments. given the significant progression and development in the course, the term factor should make a difference in educational outcome in the course. 3.1.10. time use variables the amount of time available to commit to the personal financial management course should impact educational outcome. students who devote greater amounts of time to outside employment, other university courses, and organizational activities should have less time to spend on the financial management course. part-time employment has been negatively correlated with gpa (barone, 1993) and students with fewer employment commitments have been shown to perform better in business finance classes (borde et al., 1998; simpson & sumrall, 1979). students indicated their total hours of weekly paid employment; the number was input into the learning model directly. the average weekly employment was 16.67 hours. the number of credit hours taken during the term they were enrolled in the family financial management class was also directly input, with students taking an average 15 credit hours for the quarter. there were no significant differences in the number of hours employed or the number of credit hours taken for the term between the four learning styles. 4. analysis referring to table 1 and the descriptive breakdown by learning style, the most striking result is the number of students classified as divergers and accommodators. this clustering within these learning styles implies that most students in the family financial management course grasp information best through concrete experience. a full 78% of the students were classified as either a diverger or accommodator. further differences between learning styles were apparent with respect to gender, as male students in the individual financial management class tended to identify more frequently as convergers or accommodators. thus males preferred active experimentation in the information transforming process. convergers were also more likely to be family financial management majors. equally surprising was the fact that average grades did not vary across learning style. ordinary least squares regression was used to analyze the independent relationship between educational outcome and the independent variables measuring demographics, academic, learning style, and time-use characteristics. the regression model was: education outcome5 a 1 b1 (age)1 b2 (gender)1 b3 (race)1 b4 (mother’s education) 1 b5 (father’s education)1 b6 (total credit hours)1 b7 (low gpa) 1 b8 (high gpa)1 b9 (major) 1 b10 (required class)1 b11 (term 1) 1 b12 (term 2) 1 b13 (term 3) 1 b14 (diverger)1 b15 (converger)1 b16 (assimilator)1 b17 (hours employed)1 b18 (credit hours this term)1 e (1) table 2 presents the ols regression results. the adjusted r-square of 0.42 implies that demographic, academic, learning style, and time use variables explained a significant amount of variance in grades. pair-wise correlations between explanatory variables were generally low, with only a significant correlation found between mother and father’s education. when only one measure of parental education was included in the model the results were not 244 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 significantly different from those reported in table 2. as a further test for the presence of multicollinearity the variance inflation factor was calculated for each independent variable. again, the presence of multicollinearity was not detected. with respect to independent factors impacting learning outcome, academic and time-use variables appeared to contribute to the variance in grades, while demographic and learning style variables did not appear to explain differences in course performance. specifically, students with lower gpas tended to perform almost 32 points below the average gpa student, whereas students with higher gpas earned nearly 98 more points (almost one full letter-grade) than students with average grades. family financial management majors earned nearly 43 more points out of 1000 than their nonmajor counterparts. the strongest predictor of grades in the model was the term during which the course was taken. students who took the course during the first term of observation scored much lower than students taking the course during the fourth term. this is likely attributable to two factors, i) the instructor was teaching this course for the first time at this university, and ii) somewhat different teaching methods were employed each term. while the basic content, text, and examination methods were nearly identical across terms, several assignments and activities were adjusted between terms. for example, group problem solving sessions were only used in term 2 and homework problems were optional and not graded in term 1. hours of employment directly related to course outcome with each hour of additional table 2 demographic, academic, learning style, and time-use variables regressed on course performance variable coefficient standard error t-ratio p-value demographic age 0.51 1.30 0.39 0.70 gender 13.43 11.19 1.20 0.23 race 27.98 16.55 20.48 0.63 mother’s education 5.88 12.08 0.49 0.63 father’s education 4.39 12.68 0.35 0.73 academic total credit hours 0.14 0.13 1.09 0.28 low gpa 231.86 13.28 22.4 0.01 high gpa 97.82 13.59 7.20 0.01 major 42.66 11.67 3.66 0.01 required class 25.08 13.54 20.38 0.71 term 1 2160.43 14.30 211.22 0.01 term 2 5.82 14.45 0.40 0.69 term 3 22.24 15.38 20.15 0.88 learning style diverger 27.07 11.99 20.59 0.56 converger 214.03 19.66 20.71 0.81 assimilator 4.96 17.00 0.29 0.77 time-use hours employed 1.l5 0.41 2.80 0.01 credit hours this term 23.29 1.49 22.21 0.03 constant 763.59 48.44 15.76 0.01 adj r-square .42 245j. fox, s. bartholomae / financial services review 8 (1999) 235–251 employment adding 1.15 points to the average course grade. higher credit loads appeared to detract from academic performance as each additional credit hour correlated to a 3.3 point drop in the average student’s final grade. 5. discussion and implications for instruction unlike previous studies that have found age, gender, race, and parents’ education impacting academic achievement in college courses (bellico, 1972; borde et al., 1998; mutchler, turner, & williams, 1987; sewell & shaw, 1968; simpson & sumrall, 1979), our study did not find student characteristics to be a strong predictor of grades in the individual financial management class. we anticipated older students, and students with parents who had higher levels of education to perform better; this hypothesis was not supported. though learning style was not a significant predictor of performance in this class, it was observed that males were more prevalent in the converger and accommodator learning style, implying that they may benefit from activities that stress experimentation and “hands-on” problem solving to grasp individual financial management information. interesting findings were brought to light with regard to academic predictors, having implications for instructors of similar financial management classes. regression results clearly showed that students who reported a low gpa did not perform as well in the class, whereas students with a high gpa outperformed average and low gpa students. while this comes as no surprise, instructors could take proactive measures toward the “at-risk” group. for example, some students may lack the prerequisite knowledge necessary to succeed in an individual financial management course, therefore, additional study sessions or tutor hours could be extended to these individuals who have under performed in previous courses. other students may lack the effective and efficient study strategies to succeed in academic settings, thus time use and study recommendations specific to the financial management course could improve student performance. there also was a strong tendency for family financial management majors to outperform the rest of the class. motivation and inherent interest in the topics are the likely explanations for this result. an instructor could use motivated majors effectively in class if group activities and assignments facilitated communication between majors and nonmajors. perhaps identifying majors as group leaders in team building exercises, allowing the leaders to exhibit their connection and passion for the subject matter. the most significant predictor of outcome in the individual financial management course was term during which the course was taken. given the significant changes which took place at the end of the first term (described above and outlined in detail in the appendix) it was not surprising to find that grades improved after this first offering of the course. apart from the fact that it was the first time the instructor had taught the course, the addition of required homework problems, a personal financial portfolio, and closed book quizzes, all had a significant positive impact on learning outcomes. perhaps the most striking result from the analysis was that learning style had no impact on educational outcome in the individual financial management course. there are two plausible explanations for this finding. first, learning style may not impact student perfor246 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 mance in individual financial management classes. this being the case, then instructors should not evaluate student learning style and tailor assignments to the distribution of learners in a given class. however, the second plausible explanation for finding no relationship between learning style and success in this course could result from the fact that the instructional tools used in this course were equally appealing to all learning styles. students were given the option to complete suggested assignments, both for credit and noncredit, to enhance their learning. most assignments were to be completed outside of class, involving methods of self-discovery and independent study. however, the course also involved methods catering to a learner’s preference for concrete experience and active experimentation. in fact, the two-hour class over the four terms largely consisted of information lectures, best suited for assimilators and their tendency toward reflective observation and abstract conceptualization. during the lectures, significant time was spent going over applied problems in family finance, a technique best suited for a converger’s preference for active experimentation and abstract conceptualization. it could be this tendency toward appealing to multiple learning styles that explains the resulting even distribution of academic performance across learning styles. even if this is not the case, the desirable result from both the instructor and student perspective is no noticeable significant differences across learning styles. despite learning style not being a determinant of educational outcome in the financial management classes, other important implications for the financial educator emerge. with nearly 80% of the sample demonstrating a preference for grasping information through concrete experience, teaching techniques such as class discussions, brainstorming and group activities, simulations, debates, interviewing, practicums or internships, and independent study may generate the most positive classroom environment. equally important in this case is the fact that only 13% of the sample were classified as learning best through assimilative methods. assimilative learners learn best through a process of abstract conceptualization and reflective observation which is inherent in information lectures, abstract independent problem solving, theoretical analysis, and conceptual papers, techniques commonly used to teach individual financial management. this can be interpreted as weak evidence that more “traditional” teaching techniques may not be meeting the preferences of the majority of family financial management students. continued movement toward techniques of self discovery and group interaction seems to be in the students’ best interest. if the distribution of students over learning styles is similar in other universities, then moving instruction away from abstract conceptualization and toward active experimentation is warranted. however, instructors need to determine that their students are similar to those in this sample. the kolb learning style inventory used in this analysis is easy to administer and could be used to make this determination. studies on effective teaching support the notion that the most important element effecting learning outcome are the class activities/assignments required of the student. a range of instructional methods is available to the financial management instructor. immediate involvement in the learning experience works effectively for students responding to concrete experience; active experimentation learners prefer the hands on approach to learning. for these types of learners, felder (1993) suggested active student participation through encour247j. fox, s. bartholomae / financial services review 8 (1999) 235–251 aged or mandated cooperation (team-based) on assignments both in and outside class. students responding to abstract conceptualization prefer to take a rational and logical approach; students responding to reflective observation prefer to think about the information/ situation from many perspectives. for these students, felder (1993) suggested allowing students class time to think about the material being presented. the impact of time-use on educational outcome also has important implications for instructors of financial management. employment and academics appear to complement each other. assuming that employment hours were spent in work related to family financial management, then internships may provide the fertile ground for active experimentation with the material learned in class. this implies that instructors need to be linked to the employment and internship market to provide these complementary experiences. in this study we don’t know specifically what employment environment students are engaged in, thus strong conclusions cannot be drawn. on the other hand, students who are working may only be gaining tangible financial experience by being involved in the “personal finance of employment” (e.g., collecting a paycheck, observing withholdings, managing income, making retirement plan decisions). therefore work experience alone could be providing a means of active experimentation with personal financial management concepts and skills. additionally, there may be some self-selection bias present in this finding, as it is more likely that the “highly qualified” seek the challenge of employment demands in conjunction with taking college courses. heavy course loads also identify students at risk of under performing. this too is an easy factor for instructors to identify as students enter the class. students with heavy loads could be advised to reduce other outside commitments or reduce other course commitments when enrolled in rigorous family financial management classes. 6. conclusions the academic performance of 419 undergraduate students in an individual financial management class was evaluated in light of their learning style, demographic background, academic history and time allocation. academic history and time use variables proved to be the only significant predictors of grades in the course. students performing well in other courses also performed well in the financial management course. employed students performed better in the class whereas students taking heavy course loads received lower grades in the class. student learning style, as measured by kolb’s learning style inventory, was not a strong predictor of success in this financial management class, and it appears that no single type of learner best grasps financial management concepts. thus, within the financial management class instructors should consider using a range of instructional methods to ensure equal appeal to all types of learners. 248 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 appendix references astin, a.w. (1971).predicting academic performance in college. new york: free press. atkinson, g. (1991). kolb’s learning style inventory: a practitioner’s perspective.measurement and evaluation in counseling and development, 23,149–160. barbe, w. b. & milone, m. n. (1981). what we know about modality strengths.educational leadership, 45, 378–380. barone, f. j. (1993). the effects of part-time employment on academic performance.nassp bulletin, 77, 67–73. bellico, r. (1972). prediction of undergraduate achievement in economics. journal of economic education, fall, 54–55. bergevin, p. m. (1993). the relationship between language and learning style.the journal of language for international business, 4, 1–6. borde, s. f., byrd, a. k. & modani, n. k. (1998).determinants of student performance in introductory corporate finance courses. paper presented at the annual meeting of the southern finance association (sfa) annual meeting, key west, fl. instructional activities by terma instructional activities term 1 2 3 4 lecture: x x x x lecture notes posted on course website 2 2 x x personal financial portfolio: students develop and compile 2 x x x important documents related to their personal finances. includes stating financial goals, preparing income and expense statement and balance sheet, calculating financial ratios, preparing tax form, obtaining credit report, summarizing terms of insurance policies, obtaining prospectus for 3 mutual funds, and preparing a retirement plan worksheet. wall street journal reports: students maintain daily, reflective, journal writing about personal finance topics and movement in interest rates or stock market. x o 2 2 quizzes: students take fact based quizzes, closed note, closed book. 2 x x x exams: students answer problem solving or case study evaluation questions during open note, open book exams. x x x x homework problems: financial math questions, small case study analysis. o x x x in-class group discussion and problem solving 2 x 2 2 guest speakers x x x x course web-page: students could access on-line syllabi, lecture notes, homework answers, exam and quiz answers. webpage address: http://www.hec.ohio-state. edu/cts/sp260/index.htm 2 2 x x a note. x 5 method was employed; o5 optional for student. 249j. fox, s. bartholomae / financial services review 8 (1999) 235–251 carrier, c. a. & melvin, k. (1982).linking teacher theories to teacher practices. paper presented at the annual meeting of association for educational communication and technology, dallas, tx. carthey, j. h. (1993). relationships between learning styles and academic achievement and brain hemispheric dominance and academic performance in business and accounting courses. unpublished master’s thesis, winona state university, winona, mn. claxton, c. s. & murrell, p. h. (1987).learning styles: implications for improving educational practices (ashe-edic higher educational reports, no. 4).washington, dc: association for the study of higher education. de bello, t. c. (1990). comparison of eleven major learning styles models: variables, appropriate populations, validity of instrumentation, and the research behind them.journal of reading, writing, and learning disabilities international, 6, 203–233. doyle, w. & rutherford, b. (1984). classroom research on matching learning and teaching styles.theory into practice, 23, 20–24. dunn, r. (1984). learning style: state of the science.theory into practice, 23, 11–20. dunn, r., beaudry, j. s. & klavas, a. (1989). survey of research on learning styles.educational leadership, 46, 50–58. eskew, r. k. & faley, r. h. (1988). some determinants of student performance in the first college-level financial accounting course.the accounting review, 63, 137–147. felder, r. m. (1993). reaching the second tier-learning and teaching styles in college science education.journal of college student, 34, 286–290. felder, r. & silverman, l. (1988). learning and teaching styles in engineering education.engineering education, 78, 674–681. garvey, m. & bootman, j. l. (1984). an assessment of learning styles among pharmacy students.american journal of pharmaceutical education, 48, 134–140. heitmeyer, j. r. & thomas, h. b. (1990). cognitive learning style dimensions in post-secondary home economics students.journal of studies in technical career, 12, 139–152. kaplan, e. j. & kies, d. a. (1993). together: teaching styles and learning styles improving college instruction. college student journal, 27, 509–513. kavale, k. a. & forness, s. r. (1987). substances over style: assessing the efficacy of modality testing and teaching.exceptional children, 3, 228–239. kelley, a. c. (1975). the student as a utility maximizer.journal of economic education, spring, 82–92. kolb, d. a. (1976).learning style inventory.boston: mcber publishing. kolb, d. a. (1981). learning styles and disciplinary differences. in a. chickering (ed.),the modern american college(pp. 232–255). san francisco, ca: jossey-bass publishers. kolb, d. a. (1984).experiential learning: experience as the source of learning and development. englewood cliffs, nj: prentice-hall. lacina, l. j. (1991). empowering students in the college classroom through learning style research. college student journal, 25, 354–355. lavin, d. e. (1965).the prediction of academic performance. new york: russell sage foundation. leiden, l. i., crosby, r. d., & follmer, h. (1990). assessing learning-style inventories and how they predict academic performance.academic medicine, 65, 395–402. marshall, c. (1991). teachers’ learning styles: how they affect student learning.the clearing house, 64, 225–227. mccarthy, b. (1990). using the 4mat system to bring learning styles to schools.educational leadership, 4, 31–37. mutchler, j. f., turner, j. h., & williams, d. d. (1987). the performance of female versus male accounting students.issues in accounting education, spring,103–111. o’neil, j. (1990). findings of styles research murky at best.educational leadership, 48, 7. paulsen, m. b. & gentry, j. a. (1995). motivation, learning strategies, and academic performance: a study of the college finance classroom.financial practice and education, spring/summer, 78–89. 250 j. fox, s. bartholomae / financial services review 8 (1999) 235–251 sachdeva, k. s. & sterk, w. e. (1982). projecting finance student final course scores based on initial exam scores. journal of financial education, fall, 55–60. schmeck, r. (1988).learning strategies and learning styles. new york: plenum press. sewell, w., & shaw, v. (1968). parents’ education and children’s educational aspirations and achievements. american sociological review, 33, 191–209. simpson, w. g. & sumrall, b. p. (1979). the determinants of objective test scores by finance students.journal of financial education, fall, 58–62. snider, v. e. (1990). what we know about learning styles’ from research in special education.educational leadership, 48, 53. stitt-gohdes, w. l. (1999). teaching and learning styles: implications for business teacher education. in p. a. gallo villee and m. g. curran (eds.),the 21st century: meeting the challenges to business education(pp. 7–15). reston, va: national business education association. togo, d. f. & baldwin, b. a. (1990). learning style: a determinant of student performance for the introductory financial accounting course. in b. n. schwartz (ed.),advances in accounting(pp.189–199). greenwich, ct: jai press. vihtelic, j. l. (1996). personal finance: an alternative approach to teaching undergraduate finance.financial services review, 5, 119–131. 251j. fox, s. bartholomae / financial services review 8 (1999) 235–251 pii: s1057-0810(99)00028-1 a nineties perspective on international diversification michael e. hanna, joseph p. mccormack, grady perdue* university of houston2clear lake, 2700 bay area blvd., houston, tx 77058, usa abstract investors often look to international diversification as a means to reduce the risk of a stock portfolio while maintaining a given level of return. in this study we look at ten years of historical data from the stock markets in the g-7 countries. we see how diversification from an s & p 500 portfolio into a two-market (two-country) portfolio would have impacted the risk and return. across this ten-year period, we find that a portfolio consisting solely of the s & p 500dominates any portfolio that can be constructed from the s & p 500 and themajor market index of the g-7 countries. © 1999 elsevier science inc. all rights reserved. 1. introduction for over a decade academic researchers and investment advisors have been strongly recommending that investors diversify their portfolios by investing in international equities. supported by extensive academic research that has proclaimed the risk reduction advantages of international diversification, increasing numbers of american investors have established international equity components within the asset allocation of their portfolios. today investors are encouraged by both university textbooks and by investment advisors to follow the path of international diversification. investors are told that the primary reason they should look outside the united states and invest internationally is that they will enjoy increased diversification, meaning decreased volatility (risk) levels, as a result of including an international component within their portfolios. the ability to reduce risk without sacrificing return (modern portfolio theory) has developed as one of the major goals in portfolio management. as a side note to investors, it is added that returns from selected * corresponding author. tel.:11-281-283-3213; fax:11-281-283-3951. e-mail address:perdue@cl.uh.edu (g. perdue) financial services review 8 (1999) 37–45 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00028-1 international equity markets can sometimes even boost total portfolio return above that of a purely united states equity portfolio. however, risk reduction remains the primary goal of international diversification. today many private investors are told (usually in non-quantitative terms) that they can take advantage of the fact that not all markets move up or down at the same time. due to this less than perfect positive correlation between the american financial markets and financial markets in other countries, losses in the domestic market can often be offset by gains in those foreign markets that have a low correlation to our markets. by diversifying internationally the investor hopes to have international investments that are doing well when the united states portion of the portfolio is not. but are investors really getting what they are bargaining for when they follow the advice to invest internationally? when investors add an international asset class to their portfolio’s asset allocation, does the reduction in risk really include offsetting foreign gains to balance out domestic losses? is the current rate of return maintained while the risk is lowered? despite the arguments put forth in modern portfolio theory, most investors probably have no desire to diversify out of a rising domestic market and sacrifice that return. unfortunately, it is impossible to predict when the market will be rising or falling. thus, although diversification may help avoid large losses when the domestic market is falling, it also may prevent large gains when the domestic market is rising. the true reason that investors diversify, as the model shows us, is the same reason many investors hold bonds that offer lower yields than stocks—risk reduction. in the ideal case we move towards the efficient frontier of investments as we seek to reduce risk and hold expected return constant in the portfolio. 2. review of the literature the virtues of international investing have been sung for at least the past 25 years in our literature. in a significant study solnik (1974) states that the “primary motivation in holding a portfolio of stocks is to reduce risk,” and he went on to demonstrate exactly how the level of systematic risk in a portfolio is lowered when we pursue international diversification. solnik estimates that the variability of returns for “an internationally well-diversified portfolio would be half as risky as a portfolio of us stocks. . . ” more recently other studies have arrived at essentially the same conclusion—that international investing is useful in reducing risk. black and litterman (1991) find that the efficient frontier is pushed out further when international investment opportunities are included in the opportunity set, increasing the opportunity for risk reduction beyond a united states-only portfolio. michaud, bergstrom, frashure, and wolahan (1996) review data for the preceding twenty years and come to the conclusion that “international diversification increases return per unit of risk relative to a comparable u.s.-only portfolio.” although many studies concentrate on using market indexes to look for potential gains from international investing, potential gains may be found from other international investment vehicles. for example, wahab and khandwala (1993) argue that american drawing rights (adrs) can be used to gain the anticipated benefits of international diversification. 38 m.e. hanna et al. / financial services review 8 (1999) 37–45 coming to a different conclusion, however, russell (1998) finds that adrs seem to mimic the market where they are traded, rather than the market where the underlying security is found. russell also is dissatisfied with the result of closed-end country-specific funds, noting that they behave like the host market. his results on closed-end funds concur with the earlier findings of bailey and lim (1992). many studies focus on the use of market indexes for diversification. speidell and sappenfield (1992) raise concerns that the market indexes for the major world economies are becoming more highly correlated, and that historical diversification benefits may be fading away as economies and global events tie us together in a shrinking world. most (1996) finds that it is becoming increasingly difficult to find the diversification benefits in international markets. sinquefield (1996) continues this challenge to the use of europe australia/far east index (eafe) and other major indexes to diversify an s & p 500portfolio. like speidell and sappenfield he comes to the conclusion that actively managed emerging market portfolios may give greater diversification potential. coming to a different conclusion about the correlation of markets, michaud, bergstrom, frashure, and wolahan (1996) feel that the major market indexes are not really becoming more highly correlated. they feel that benefits from diversification still exist between markets. solnik, boucrelle, and le fur (1996) assert that the various financial markets show “correlation increases in periods of high market volatility.” this is not good for investors because it is precisely during volatile moments in the market that low correlation is most desired. aiello and chieffe (1999) examine international index funds to see if they can be used to gain the desired benefits of international diversification. however, they conclude that these funds do not accomplish that end. with similar results ho, milevsky, and robinson (1999) conclude that market performance in recent years would have allowed a canadian investor to benefit from international diversification, but that an american investor would not be so fortunate. the conclusions of these two studies are highly consistent with the results of our research. 3. methodology and data this study examines the risk and return effects that would have been realized by a hypothetical united states investor who elected to pursue international equity diversification by investing in the financial market indexes of the other six g-7 group of industrialized nations. those other six nations are canada, the united kingdom, france, germany, italy, and japan. data are analyzed on the reported value of a major market index for each of the seven markets. the market indexes under study in this research project are the s & p 500, the toronto stock exchange (tse) 300 composite index, the financial times index of london, the paris cac 40, the frankfurt dax, the milan mlbtel, and the tokyo nikkei 225. this study examines the ten years of equity market data from january 1988 to december 1997. all data that are analyzed for this study are monthly observations. values for each of the seven market indexes are obtained from the first joint trading day of each month, as 39m.e. hanna et al. / financial services review 8 (1999) 37–45 reported inthe wall street journal. exchange rate data are also collected for the same trading days as the stock index observations. this exchange rate information is used to convert market return data to united states dollar equivalent values. means and standard deviation (sd) are computed on the monthly return data (in u.s. dollars) for each of the seven indexes. these provide a relative comparison of the different markets both on a return basis and on a risk basis. these help to identify the foreign markets that would be best to consider for diversification. correlation coefficients are then calculated to describe the relationship between each foreign market index and the s & p 500. the markets with the lowest correlation would usually be good candidates for diversification of a portfolio. regression analysis is then used to test for a linear relationship between the s & p 500index and each of the market indexes individually. these regressions show where there is a statistically significant relationship between the s & p 500 andeach of the other markets. the coefficient of determination for each model is computed to measure the strength of the relationship. to provide further information about the impact of international diversification over this time period, sample portfolios are developed where the s & p 500 ispaired with each of the other markets. these range from a 100% s & p 500portfolio to a 60% s & p 500 and 40% other market portfolio. the return and standard deviation is computed for each of these to measure the performance. a tabulation is made with the data to indicate how often each of the other markets had monthly movements in the same direction as the s & p 500. clearly there are times when diversification provides a benefit—whenever the foreign market goes up during a month that the s & p 500goes down. however, there are other times when diversification causes the gains in the s & p 500 to beoffset by losses in the other market. the number of each occurrence over this ten-year period provides an indication of how often this occurs. 4. results the dollar-adjusted monthly average (geometric mean) return and sd of returns for each of the seven markets, are presented in table 1. as observed in this table, the united states table 1 rates of return and standard deviations market monthly geometric mean return annual geometric mean return standard deviation of monthly returns s & p 500 0.011206 0.143081 0.031528 toronto 0.005259 0.064964 0.037028 london 0.007746 0.097020 0.069647 paris 0.008069 0.101243 0.056440 frankfurt 0.011019 0.140548 0.051593 milan 0.003416 0.041770 0.085725 tokyo 20.002268 20.026882 0.078634 40 m.e. hanna et al. / financial services review 8 (1999) 37–45 market is the most stable (i.e., the smallest standard deviation of returns) across this period. the united states market also has the greatest monthly and annual geometric mean rate of return. of the seven markets reviewed in this study, the united states market is clearly the dominant market in this time period. its coefficient of variation (not presented in a table) is also the best for this period of study. the frankfurt dax has a comparable rate of return, but has a standard deviation of returns that is over sixty percent larger than that of the s & p 500. the sd of returns for the toronto 300 is the most comparable to the s & p 500, but the rate of return in the canadian market index is less than half of that experienced by the s & p 500. in table 2 the correlation coefficients between each of the seven markets are reported. as with all the analysis in this paper, the values reported are based upon returns that have been adjusted to united states dollar rates of return. all correlation coefficients are positive, indicating a clearly positive relationship between the markets. correlation with the united states market is highest with the toronto exchange, and lowest with the milan exchange. to statistically test for a linear relationship between the s & p 500 andeach of the other market indexes, six regression models are developed with the s & p 500 return as the dependent variable and the return of the other market as the independent variable. table 3 provides a summary of these results. from this we see statistically significant results for each of the markets except milan. the r2 for each of these shows the strength of the relationship. to gain an historical perspective on the impact of diversification, we develop portfolios of the s & p 500 witheach of the other six indexes individually. for each of these two-country portfolios, several weighting schemes are considered. the weights for the s & table 2 correlation coefficients between indexes s & p 500 toronto london paris frankfurt milan tokyo s & p 500 1 toronto 0.6023 1 london 0.3743 0.3542 1 paris 0.5006 0.3294 0.3930 1 frankfurt 0.4074 0.2650 0.3425 0.6707 1 milan 0.1607 0.1872 0.2178 0.1211 0.2003 1 tokyo 0.2371 0.2082 0.2603 0.3551 0.2300 0.3047 1 table 3 individual regression coefficients of foreign markets on s & p 500 market coefficient t-statistic r2 toronto 0.707 8.16a 36.3 london 0.827 4.37a 14.0 paris 0.896 6.26a 25.1 frankfurt 0.667 4.83a 16.6 milan 0.437 1.76 2.6 tokyo 0.592 2.64a 5.6 a statistically significant at the 0.01 level of significance. 41m.e. hanna et al. / financial services review 8 (1999) 37–45 p 500 portion of the portfolio vary from 100 percent to 60%. the mean return and standard deviation is computed for each of these portfolios. fig. 1 presents a graphical representation of the return and standard deviation of several possible combinations of the united states s & p 500 and thetoronto 300 index. the weights of the united states component vary in this figure from a 100 percent united states component to only 60% united states. the 100% united states component is at the upper left end-point of the set of portfolios, and the portfolio that is only 60% u.s. is at the lower right. this clearly shows that as the portfolio becomes more diversified (more weight in the foreign index), the mean return declines whereas the sd increases. this is precisely the opposite of what one wishes to obtain through international diversification. however, this was not too surprising since the united states market has both the highest rate of return and the lowest sd of returns. an analysis of the various other portfolios considered—comprised of united states and non-united states components (i.e., toronto, london, paris, frankfurt, milan, and tokyo)— each has essentially the same result. the 100% s & p 500portfolio is always the dominant portfolio. the greater the weighting of the s & p 500 component of the portfolio, the higher is the return and the lower is the level of risk. thus, in every case the undiversified (i.e., 100% s & p 500) portfolio dominates the diversified (i.e., two-country) portfolio. clearly this result is inconsistent with the arguments investors hear in favor of international diversification. for the results to be consistent with the theoretical expectations, certain conditions must exist in the data, and these are implicit (if not explicit) in the models proposed in textbooks. if the investor wishes to maintain a two-country portfolio with the same rate of return as an entirely united states portfolio, the return in the foreign market should be equal to (or greater than) the return in the united states market over this time period. this by itself, however, does not guarantee a lower level of risk. for risk to be less in the multi-country portfolio, its sd (which is a function of the weights of the two markets in the portfolio, the sd of the two markets, and the correlation coefficient) must be smaller than that of the united states market alone. 42 m.e. hanna et al. / financial services review 8 (1999) 37–45 4.1. an alternative perspective traditionally most textbooks stop with correlation between markets in explaining the advantages of international diversification. however, although correlation measurements are useful and an important component of modern portfolio theory, perhaps too few investors appreciate the implications of the movements between two markets that lead to the correlation value that is used in modern portfolio theory. this study breaks down the imperfect correlation between the united states and non-united states markets into a unique analysis that provides a simple illustration of a component of the level of correlation between markets. it has been said that when an american football team attempts to make a forward pass, there are three possible results—and two of them are bad. following that analogy this study reports on the four possible outcomes that can occur when an investor invests in both the united states market and another market. the four possible outcomes are that both markets can go up, both markets can go down, the domestic market can go up and the foreign market go down, and the foreign market can go up whereas the domestic market goes down. as is discussed below only one of these four possible outcomes is the scenario that is often explained to the investor as the justification for international diversification. this justification is the argument that domestic market losses can be offset by foreign markets gains due to less than perfect correlation between the markets. table 4 presents an analysis of the relationship between the direction of movement in the united states market and each of the six other individual markets in the g-7. by using a four-quadrant table for each market, it is possible to observe both the direction of movement in two markets and to what extent the investor really has those occasions where losses in the united states market are offset by gains in the non-united states market. the first foreign market that is reported in table 4 is the toronto market. in the upper left table 4 monthly market movements market market up market down toronto s & p up 59 18 s & p down 13 29 london s & p up 52 25 s & p down 18 24 paris s & p up 49 28 s & p down 11 31 frankfurt s & p up 58 19 s & p down 19 23 milan s & p up 38 39 s & p down 23 19 tokyo s & p up 42 35 s & p down 20 22 43m.e. hanna et al. / financial services review 8 (1999) 37–45 quadrant of the toronto data, one may observe that for 59 of the 119 monthly returns used in this study both thes & p 500 and the toronto 300 have positive returns. investing internationally may have done little or nothing for u.s. investors in these 59 periods, depending on the relative magnitude of the gains in toronto and new york. in the lower right quadrant one may observe that in 29 of the observations, both markets decline. again the magnitude of the loss in one market relative to another is important in arriving at the total portfolio loss, but there is no real offset between markets as international investors hope to enjoy. between these two quadrants we observe that 88 of all 119 observations (or 74% of the total) indicate the two markets moved in the same direction. in the upper right quadrant we observe that in 18 cases, the united states market is up and there is at least a partial offsetting loss in toronto. in these cases (15% of observations) international diversification clearly hurts portfolio returns. adding these observations to the previously discussed observations, one observes that in 106 of the 119 observations we can find little or no obvious benefit to returns from international diversification. finally, we observe the lower left quadrant. this quadrant represents the occasions in the study period when the united states market declines and the toronto market rises. there are only 13 observations in this quadrant, and these 13 observations are only 11 percent of all observations across this 10-year period. yet this is the quadrant that represents the primary reason we invest internationally. this is where we see losses in our home market being offset at least partially by the foreign market moving in the opposite direction. examining table 4 further to observe the four quadrants for each of the other five markets, we see that the pattern is basically the same in each case. the lower left quadrant ranges from a high of 23 observations (19%) for milan to a low of only 11 observations (9%) for the paris cac. this further challenges the argument that domestic losses are being offset by foreign gains. comparing table 2 correlation coefficients to table 4 with the data on the number of observations in the lower left quadrant for each market, no distinct pattern is discernible. high correlation between the united states market and another market does not clearly indicate what may be expected in the lower left quadrant in table 4. although higher correlation generally tends to coincide with fewer lower left quadrant observations, this is not always the case. for example toronto has the highest correlation, but the second fewest (not the fewest) number of observations in the lower left quadrant. thus, data in the four quadrants may provide information that is not clearly ascertained from the correlation coefficient. 5. conclusions modern portfolio theory is correct in how it seeks to minimize volatility around a given expected return. an investor may benefit significantly from the united states market and another market moving in opposite directions. however, this movement does not happen with enough frequency across the decade studied here to justify the assertion that foreign gains will compensate for domestic losses. data across this time period for the united states 44 m.e. hanna et al. / financial services review 8 (1999) 37–45 and its g-7 partners fail to support the expectation of offsetting movements that benefit the united states investor. if the goal of the investor is offsetting domestic market losses, then the investor facing an asset allocation decision must consider historical market patterns. the cases that are examined by this study do not show international diversification (based on market indexes) over the last ten years would have been as potent a tool as is been suggested in much of the academic research and by investment advisors. given this recent historical experience, the investor must ask if there is sufficient reason (in terms of risk or return) to pursue international diversification. we should acknowledge that the results of this study might be sample specific. although we used the complete data set of the decade from 1988 through 1997 in describing the relationship between the united states and other markets, it may be difficult to extrapolate the findings of this research to the future or to any other markets. nevertheless, investors may have more reasonable expectations of the potential benefits of international diversification by studying historical relationships. references aiello, s. & chieffe, n. (1999). international index funds and the investment portfolio.finan serv rev, 8(1), 29–37. bailey, w., & lim, j. (1992). evaluating the diversification benefits of the new country funds.j portf manag, 8(3), 74–80. black, f., & litterman, r. (1991). global portfolio optimization.finan anal j, 48(5), 28–43. ho, k., milevsky, m.a., & robinson, c. (1999). international equity diversification and shortfall risk.finan serv rev, 8(1), 13–27. michaud, r. o., bergstrom, g. l., frashure, r. d., & wolahan, b. k. (1996). twenty years of international equity investing: still a route to higher returns and lower risks?j portf manag, 23(1), 9–22. most, b. w. (1999). the challenges of international investing are getting tougher.j finan plan, 38–40(february), 42–46. russell, j. w. (1998). the international diversification fallacy of exchange-listed securities.finan serv rev, 7(2), 95–106. sinquefield, r. a. (1996). where are the gains from international diversification?finan anal j, 52(1), 8–14. solnik, b. h. (1974). why not diversify internationally rather than domestically?finan anal j, 30(4), 48–54. solnik, b., boucrelle, c., & le fur, y. (1996). international market correlation and volatility.finan anal j, 52(5), 17–34. speidell, l. s., & sappenfield, r. (1992). global diversification in a shrinking world.j portf manag, 19(1), 57–67. wahab, m., & khandwala, a. (1993). why not diversify internationally with adrs?j portf manag, 19(2), 75–82. 45m.e. hanna et al. / financial services review 8 (1999) 37–45 pii: s1057-0810(99)00037-2 hedging individual mortgage risk terry l. zivneya,*, carl f. luftb aball state university, muncie, in 47306, usa bdepaul university, 1 east jackson, chicago, il 60604, usa abstract this paper investigates the feasibility of an individual hedging the interest rate risk involved in planning to take out a mortgage at a future point in time. simulation using market data indicates that a simple futures hedge reduces the variation in mortgage capacity by about one half. expected mortgage capacity is very close to 100% of the original capacity at a very low cost. hedging the individual mortgage with a put futures option is less effective in reducing downside risk and has a higher expected cost. © 1999 elsevier science inc. all rights reserved. jel classification:g21; r21 keywords:mortgage; hedging; futures; housing economics 1. introduction one of an individual’s more frustrating financial events is watching an increase in mortgage interest rates make one’s dream house unaffordable. the housing search often begins with determining the maximum loan for which a borrower can qualify at the current level of interest rates. a 1% increase in rates can easily lead to a shortfall of 10% toward the purchase price of the house. although mortgage lenders offer a lock-in, or guarantee, in rates, these are generally good for a limited time and often require that a specific property has been identified. the housing search may easily take longer than the typical 30–60 day lock-in period. the delay is often compounded when the present home must be sold before purchasing a new home. * corresponding author. tel.:11-765-285-2198. e-mail address:00tlzivney@bsuvc.bsu.edu (t.l. zivney) financial services review 8 (1999) 101–115 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00037-2 homeowners often begin a serious search for a new home after selling their present home; yet, the decision to sell may well have been predicated upon expecting to be able to move up to a particular price that may be vulnerable to changes in interest rates. this paper examines the feasibility of individuals protecting themselves against unforeseen changes in interest rates that may lead to the disappointment of being unable to make the planned purchase of a home. 1.1. an illustration of the problem the smiths have saved $25,000 for a down payment on their dream home. this will serve as the 20% down payment on a $125,000 home, thus avoiding the extra costs of private mortgage insurance (pmi). at an annual percentage rate (apr) of 8%, the smith’s monthly mortgage payment on a 30-year mortgage for principal and interest will be $733.76. based on the smith’s income, their lender will qualify them for this monthly payment, but no more. alas, before the smiths locate a suitable home, interest rates rise to 9%. now, the monthly payments of $733.76 will only support a $91,193.63 mortgage, leaving the smiths $8,806.37 short. because of the smiths’ current incomes, they will not qualify for the $804.62 payments now needed for the $100,000 mortgage. for many home buyers the problem is even worse. for example, the normal practice for financing newly constructed homes is to obtain a short-term construction loan with the permanent mortgage rate not being determined until the house receives a certificate of occupancy, that can be many months after the start of construction. unlike the smiths, who were deeply disappointed in being unable to buy their dream home, the jones, who had a custom-built home constructed, are committed to paying the new rate. 1.2. a solution both the smiths and the jones could reduce a great deal of the uncertainty about their financing byhedgingtheir interest rate risk. the organized futures markets have developed over the years precisely to enable risk to be transferred from those unwilling to shoulder it to those willing to bear it. common textbook examples of those benefiting from hedging are the farmer who is unsure of the price of wheat he will receive at harvest and the baker who is unsure of the price she will pay for wheat in the future. the organized futures exchanges enable the farmer and baker to match these offsetting risks without having to be concerned with locating a person to take the other side of the deal. furthermore, the exchange’s clearinghouse effectively eliminates the credit risk one normally faces when dealing with individuals or corporations. both the farmer and baker need only be concerned with determining how much to buy or sell and the price they will receive or pay. 2. literature review hedging of mortgage-backed securities has become so commonplace that fernald, keane and mosser (1994) argue that hedging activity related to mortgage securities has affected the 102 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 level of treasury interest rates. as would be expected, a voluminous literature has developed to analyze this hedging by financial institutions. previous studies on hedging interest rate risk in mortgages have focused on the investor in mortgages rather than the individual home owner. for example, mortgages are typically pooled for resale in a secondary market. investors purchase securities that share in the cash flow from these mortgage pools. these mortgage-backed securities are subject to not only the pure interest-rate risk of treasury bonds, but also the risk of greatly-accelerated prepayment when rates decline. the original gnma futures contract failed in 1985 and the cash settled gnma futures contract failed in 1986 due to low trading volume. johnston and mcconnell (1989) present evidence that the demise of the gnma futures contracts was due to the poor design of the futures contracts. the variety of delivery options resulted in poor hedging quality relative to using treasury bonds in the difficult interest rate environment of the early 1980s. patel (1994) reports that a futures contract on mortgage interest rates also failed in the london market, apparently because of the inability to actually buy the index underlying the contract. follain and park (1989) show that the optimal hedge ratio for a lender using futures contracts varies with the level of interest rates. they use a regression approach to estimating the best hedge ratio. goodman and ho (1997) also find that the regression approach gives better hedge ratios than does using option-adjusted duration. breeden (1994) finds that hedging using either short-term eurodollar futures or long-term treasury bond futures reduced the volatility of investing in fnma mortgage pools by about 40% during the 1992–1994 period. breeden (1994) notes that the 7-year treasury note has the highest correlation with mortgage-backed securities. breeden (1991) shows that from a lender’s perspective, a mortgage is equivalent to issuing a bond and buying a call option, or equivalently, buying a put option. the exercise price of these options is the par value of the bond. breeden (1991) also finds that increased volatility in interest rates leads to a greater spread between treasury bond rates and mortgage rates because the imbedded prepayment option becomes more valuable. murphy and gordon (1990) suggest that put options on treasury futures may result in better hedges than hedging with the futures contracts themselves if downside risk is considered most important to the owner of a portfolio of mortgages. in contrast to the literature discussed above concerning hedging of existing mortgages is a smaller literature on hedging commitments to future loans. berkovitch and greenbaum (1991) note that banks use loan commitments to act as hedging instruments. the commitment fee paid by the potential borrower may be viewed as the cost of option to obtain a loan at a fixed rate; however, although option pricing models assume all-or-nothing behavior, bank’s loan commitments are usually exercised in part. maris and white (1989) use the black-scholes option pricing model to value this loan commitment for residential mortgages. they find that the embedded put option for a 45 day commitment is worth about 1% of the mortgage value, that is also a typical commitment fee. kutner and seifert (1991) use a similar methodology to estimate the value of a commitment for residential mortgages over a range of interest rate levels and volatilities. they find that 1–3% of the mortgage value is the fair price for a 30–45 day rate commitment using parameters from the 1985–1987 time period. hochstein (1998) quotes practitioners who note that lenders view hedging the mortgage pipeline (the time during which the rate commitment is made) as prohibitively expensive 103t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 compared with the expected benefit due to the short hedge times and generally low volatility. cross (1998) suggests that lenders can effectively manage pipeline risk by cutting the time between rate lock and closing the loan. in contrast to the extensive literature on institutional hedging of mortgage risks, almost nothing has been written about the interest rate risk facing an individual wanting to take out a mortgage at a future date. the most relevant paper is by sharp (1989) who uses an option pricing model to value the insurance premium (put option) charged in canada to partially protect homeowners against mortgage rate increases on their variable rate mortgages. he notes that the price of the option is so high that the insurance is rarely purchased. one other way an individual can deal with interest rate risk is to use an adjustable rate mortgage (arm). templeton, main and orris (1996) and chiang, gosnell and heuson (1997) both use simulations to assess the risks faced by individuals taking out arms. using an arm substitutes uncertainty about interest rates in the distant future for uncertainty about interest rates in the near future for the borrower expecting to get a fixed rate mortgage. this paper does not explore the use of arms. 3. methodology and data 3.1. hedging with futures contracts a futurescontract is often viewed as a standardizedforward contract. a forward contract is a contract to deliver something at a later date where the delivery price is fixed as of the date of the contract. thus, a loan commitment can be viewed as a forward contract. in practice, the borrower under the loan commitment is not obligated to take the funds, whereas the lender also often has some freedom to not deliver funds. viewing the commitment, however, as a forward contract aids in visualizing how futures contracts and options contracts may be used in the mortgage origination process. the good to be delivered is the loan and the price is the guaranteed interest rate. both the lender and borrower face the uncertainty of changing interest rates. the lender’s risk is that rates will fall and less income will be generated by the loan than expected, whereas the borrower must deal with the risk of rising rates and the prospect of paying more interest than anticipated. both parties can minimize the effects of changing interest rates via the futures market. the lender can establish a long position that will rise in value as rates decline by purchasing interest rate futures contracts. conversely, the borrower can create a short position by selling interest rate futures contracts that will increase in value as rates rise. futures contracts have several advantages over forward contracts because of the existence of the clearinghouse. the two parties to the contract do not have to know each other because the clearinghouse will match up their respective futures positions. neither party needs to be concerned with the others’ credit worthiness because of themargin requirement imposed by the clearinghouse. each party posts a good faith deposit of funds with the clearinghouse in a margin account. the clearinghouse requires that funds be deposited in a margin account when a futures position is established. this amount is knows as the initial margin. if losses are incurred by the futures position, then the clearinghouse permits the margin account to fall 104 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 to a minimum level known as the maintenance level. if the account falls below the maintenance level then the clearinghouse requires that enough funds be deposited to restore the account to the initial level. if gains accrue to the account then any funds that exceed the initial balance can be withdrawn. the clearinghouse adjusts the margin account each day to reflect gains or losses that accrue to the open futures position. this daily adjustment is known asmarking to marketand ensures that the necessary funds will be available to satisfy the contract. margin differs from the earnest money usually posted by the home buyer in that both parties to the futures contract make the same initial deposit and then are required to maintain a minimum balance in the margin account. furthermore, adjusting the margin account balance daily to reflect changes in interest rates ensures that both parties still have a strong interest in fulfilling each end of the bargain. either party can pass on their obligation to another at any time by entering into an offsetting contract at the then current price via the futures exchange and clearinghouse. thus, there is an effective exit possibility. of course, the one-size-fits-all nature of a standardized futures contract has some disadvantages. two of the drawbacks that are especially relevant to the case of hedging individual home mortgages are the limited number of discrete sizes of contracts (generally multiples of $100,000) and the less than perfect substitute contract for a mortgage. this relatively poor match between mortgages and futures contracts arises because of the mortgage prepayment option. because mortgages can be prepaid and retired before maturity, it makes the effective durationof a mortgage less than that of a treasury bond of the same maturity. if one wants to hedge a mortgage based on maturity, then a 30-year treasury bond contract is available; however, according to breeden (1994), the closest available substitute for a 30-year mortgage in terms of interest rate sensitivity is the 10-year treasury note contract. a third potential drawback to hedging individual mortgage rate risk with futures contracts is that hedging means that the individual is committed (at least financially) to taking out a mortgage at a future point in time. if the individual decides to not take out a mortgage, he must accept whatever losses or gains the futures contract has made without the benefit of the offsetting gain or loss in the size of the mortgage. therefore, hedging with futures contracts is a serious commitment. this commitment is quite appropriate when a contract for constructing a house has been signed, but may not be appropriate when a family is merely touring sunday open-houses. if one is serious about purchasing a house but is uneasy about the obligation imposed by a treasury bond futures position, then options on treasury bond futures are a viable alternative. the simplest and most effective strategy is to execute the purchase of a put option on a treasury bond futures contract. this strategy will protect against rising interest rates, yet will provide the flexibility to benefit from falling rates. when compared to a short futures position, the long put option position usually will require more of a cash outlay when the position is established, however. 3.2. institutional details treasury bond futures and treasury note futures contracts are traded on the chicago board of trade. both have a $100,000 face value and an assumed coupon of 8% (beginning 105t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 in january 2000, the assumed coupon will be 6% for the 10-year note and the 30-year bond). the initial margin for the 30-year t-bond contract is $2,700, whereas the maintenance margin is $2,000. initial and maintenance margins for the 10-year t-note contract are $1,620 and $1,200 respectively. a smaller contract would make hedging an individual mortgage potentially more attractive by reducing the error in the hedge ratio. such contracts, known as mini t-bond contracts, are traded on the mid america commodities exchange. both a 10-year treasury note and a 30-year treasury bond contract are available in $50,000 face value denominations. several authorities have noted that a 10-year note provides better tracking of mortgage rates than the 20or 30-year bond. the initial margin required on the 10-year note contract is $810 with a maintenance margin of $600. the corresponding margins on the 30-year bond contract are $1,350 and $1,000. the margins on the mini or half-sized contracts are half the margin on the full-sized contracts traded on the cbot. the smiths would need two of the mid american contracts to approximate the hedging effectiveness of one of the cbot contracts. in both cases, the margin they would be required to post is only a small fraction of the down payment they have already saved. thus, the hedge seems financially feasible. the treasury bond and note futures contracts are traded on the march-june-septemberdecember cycle. in general, the near-term contracts are considered most liquid; however, because the individual mortgage hedger is not concerned with large positions, adequate liquidity would exist with any of the four maturities. because the futures contract needs to be closed out before the contract month to avoid the possibility of having to deliver the underlying securities, the smiths are well advised to select a contract month beyond the time they expect to close on their mortgage. most individuals do not have a futures trading account; however, the financial capacity that is necessary to support a futures trading account is well within the reach of many small investors. the lind-waldock brokerage firm, one of the nation’s largest discount brokerage firms, has as normal requirements to open an account: $5,000 in liquid assets, an annual income of at least $25,000, and $50,000 in equity excluding one’s primary residence. these requirements are meant to insure that if the initial margin is depleted, then subsequent margin calls can be met. note that sharp decreases in interest rates would lead to large losses on the short futures position, thus prudence dictates that at least $10,000 in liquid assets be dedicated to a futures account. of course, the decrease in interest rates means that a larger mortgage can be qualified for at the same monthly payment, offsetting the loss in the futures position. moreover, the costs associated with both opening and closing a futures position, i.e., the round trip commissions, are only $29 for a futures contract and $35 for an options contract and are paid when the position is liquidated. a full service broker, however, may charge as much as $200 per contract and impose much stricter liquidity and equity requirements. although the smiths seem to have sufficient liquidity by virtue of the $25,000 available for their down payment, they may not have enough equity to establish a futures trading account. if, however, the smiths restrict their hedging activities to purchasing put options then it may be possible to establish a trading account with the $25,000 in cash. creation of such an account would be at the discretion of the brokerage house, and is by no means a certainty. on the other hand, the jones, who are constructing a $250,000 house, have 106 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 accumulated a $50,000 down payment, enough to satisfy normal equity requirements for establishing a futures trading account. 3.3. alternative goals the smiths and jones potentially have two alternative goals for their mortgage hedging strategies. the jones, who are committed to taking out a mortgage to fund their newlyconstructed home, may have a goal of minimizing the uncertainty of their mortgage capacity. this goal may be met by hedging with futures contracts at a low effective cost. the smiths, who are planning on buying a house at some indefinite future time, may be uncomfortable with the knowledge that hedging with futures means foregoing the opportunity to obtain a mortgage at a lower interest rates if rates decline before they find their home. yet the smiths are concerned about an unforeseen increase in mortgage rates ruining their chances of obtaining their dream home. furthermore, the smiths may decide to not buy a house at all. the smiths’ goal is to ensure that the effective mortgage rate they pay is no greater than the current rate. this goal may be met by purchasing a put option and effectively paying an insurance premium against an increase in rates. if the smiths decide against buying a home, they forfeit the price of the put option, but have no further financial obligation. 3.4. data the data underlying the analysis are from breeden and giarla (1992) who present mortgage interest rates for 30-year fixed-rate mortgages and futures contracts on 20-year treasury bonds for each month from january 1984 through december 1990. this period of time is one that was especially difficult for a hedger and therefore provides a powerful test of the potential effectiveness of hedging. first, interest rates were quite volatile, sometimes changing more than 3% over a 12-month period. second, the rates were on a long-term downward trend, meaning that on average, waiting to take out a mortgage would reward the individual with lower rates, albeit with the risk of sharp, sudden increases. this better-thanaverage performance of the unhedged mortgage makes a tough standard of comparison for the use of a hedge. finally, this period of time includes the time span when the gnma futures contract was failing, presumably because of ineffective hedging performance. fig. 1 displays the data for the fixed rate (gnma) mortgage interest rates and the short term t-bill rates for this period. these data are derived from market prices, rather than survey data as shown in templeton, main and orris (1996) that seems to be somewhat smoothed. there were a number of sizable reversals of interest rate trends during this period. 4. results 4.1. hedging individual mortgages with futures several different scenarios are examined to test the feasibility of hedging an individual’s mortgage origination using futures contracts. these scenarios differ in the time from the initiation of the hedge until the mortgage is taken down. 107t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 each month, the smiths are assumed to form a twelve month hedge in anticipation of taking out a 30-year fixed-rate mortgage sometime during the forthcoming year. the hedge is formed by selling one futures contract on a treasury bond. this contract has a face, or notional, value of $100,000, that nicely dovetails with the smiths’ plans for taking out a $100,000 mortgage. together with their savings of $25,000 for the down payment, the smiths will be able to afford a $125,000 home. at the beginning of this simulation, january 1984, the price of the one year futures contract is 69.16 (t-bond futures prices are expressed in decimal form with par being equal to 100) and the smiths established a short position in t-bond futures by selling a futures contract for $69,160. this steep discount from the t-bond future’s $100,000 face value is caused by market interest rates being far above the 8% rate underlying the futures contract. if the smiths finally take out a mortgage in december, and buy back the futures contract at its then current price of 71.06, or $71,060, they will lose $1900 on their futures transactions. meanwhile, 30-year mortgage rates have declined slightly, from 12.82% in january to 12.74% in december. the monthly payments they would have faced on the january mortgage would pay for a slightly larger mortgage in december, $100,570. thus, the additional $570 in leverage gained from the declining interest rates mitigated the $1900 loss in the futures market so that on balance the smiths are $1330 worse off from hedging in this example. fig. 1. historical end of month interest rates for mortgages and treasury bills. 108 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 suppose, however, the smiths had found their dream house somewhat earlier in the year, and had closed on their mortgage in june. in june, 1984, fixed-rate mortgages had increased to 14.67%! the original monthly payments projected in january could now support only an $88,210 mortgage, exactly what the smiths feared. fortunately, increasing interest rates had caused the price of the treasury bond futures contract to fall to 58.56, or $58,560. this means that when the smiths closed out their short futures position, they netted $10,600 in their margin account. as a result of their hedging activity, the smith’s total financing available for their house is $88,210 plus $10,600 plus the $25,000 savings, or $123,810. the smiths only had to raise an additional $1190 to move into their home. the analysis here ignores the interest earned on the $25,000 down payment while awaiting the closing of the transaction. the interest earned would increase the purchase capacity in both the hedged and unhedged cases, but would raise the apparent ending purchase capacity. repeating this example for the hedging periods available in our data shows that the average wealth at the end of the 11-month hedging period is 99.20% of the beginning home buying capacity. furthermore, as expected with hedging, the extreme outcomes declined dramatically. unhedged, the smiths faced a worst case shortfall of 15.21% in that era, and a standard deviation (sd) of capacity of 10.42%. standard statistical interpretation of the sd means that approximately 16% of the time the home buyer would likely be more than 10.42% short of the purchase price of the house. hedged, the worst case shortfall is 10.90%, whereas the sd is 5.02%. thus, hedging greatly improves the likelihood of the smiths being able to afford their dream home. repeating the analysis for the shorter hedge periods gives similar results. for the 79 five-month hedge periods, the unhedged worst case shortfall is 11.79% with a sd of 7.23%. hedged, the worst case shortfall is 12.36% (100 – 87.64) with a sd of 4.08%. the average purchase capacity at the end of the hedging period is 99.96% of the beginning capacity. when comparing the worst case shortfalls, the reader should note that the hedged and unhedged worst cases do not coincide. in general, the worst case hedged shortfall occurs when rates drop dramatically over the hedged period. the increased loan capacity is more than offset by the losses on the futures transaction. the worst case unhedged shortfall occurs when interest rates increase so that although the loan capacity is diminished no offsetting gain is available from the hedge not taken. the last column in table 1 shows the benefit that the borrower obtained via the hedge. this column portrays the change in the dispersion of the hedger’s borrowing capacity and the numbers are calculated as the percentage difference between the sd of the hedged position versus the unhedged position. for example, the 5.02% sd of the 11-month hedge is roughly half the 10.42% sd associated with the unhedged 11-month position. this lower sd represents a 51.8% improvement over the unhedged position. the data in this column reveal that the hedges reduced the sd of mortgage capacity for every hedge period. this reduces the smith’s exposure to changing interest rates, reducing the uncertainty about their borrowing capacity by between 43.1% to 51.8%. this finding is in accord with that of breeden (1994) for the effectiveness of hedging mortgage-backed securities in the more recent 1992–1994 period. 109t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 4.2. put options as insurance against interest rate increases the smiths may wish to consider buying a put option as an alternative to forming a hedge with futures contracts. the cbot trades put and call options on the treasury bond and treasury note futures contracts. there are no options available on the smaller mid american futures contracts. buying a treasury futures contract put option would give the smiths the opportunity to offset an increase in interest rates. if interest rates increase, then the right to sell the futures contract, that underlies the put option, would increase in value. the put option could then be resold to realize this increase in value. the value rises because the put option conveys the right to sell the underlying futures contract at specific price known as the exercise price. an increase in rates would cause the underlying futures contract price to fall below the previously established exercise price. the put option would then expire worthless, and the smiths would lose the amount paid for the put option. in general, the value of an option contract does not increase dollar for dollar with the value of the underlying futures contract. the most liquid contracts are for those options with the exercise price nearest the current futures price. for these puts thedelta is typically about one-half. that is, a decrease in the futures contract of one dollar would result in an increase in the option contract of fifty cents. therefore, to obtain a similar amount of hedging as obtained with one futures contract, the smiths would need to purchase two put options. if interest rates decline, the smiths would not exercise their put contract, but would instead take out a mortgage at the new, lower rates. the net cost of the hedging position would be the cost of the put contracts and the commissions. if the smiths decide against buying a house, they can simply sell the unnecessary put options for whatever the then-current price table 1 hedging mortgage using futuresa months delay unhedged hedged with treasury bond futures percent reduction average sd maximum minimum average sd maximum minimum sd with hedge 0 100.00 0 100.00 100.00 100.00 0 100.00 100.00 0 1 100.39 3.23 108.43 90.84 100.09 1.80 104.75 94.13 44.3 2 100.80 4.81 110.96 88.63 100.17 2.56 105.46 90.47 46.8 3 101.22 5.88 112.94 86.70 99.69 3.08 107.72 88.73 47.6 4 101.66 6.63 116.68 86.92 99.81 3.59 109.58 90.78 45.9 5 102.21 7.23 118.84 88.21 99.96 4.08 112.61 87.64 43.6 6 102.78 7.69 120.97 84.16 99.42 4.35 110.13 85.75 43.4 7 103.29 8.09 122.26 82.32 99.50 4.60 111.29 87.50 43.1 8 103.80 8.59 122.36 82.70 99.61 4.82 110.16 89.26 43.9 9 104.24 9.11 123.69 84.77 99.04 4.87 109.66 87.97 46.5 10 104.69 9.79 128.59 82.47 99.14 4.94 108.47 89.47 49.5 11 105.12 10.42 130.97 84.79 99.20 5.02 108.36 89.10 51.8 a percent of original mortgage capacity based on monthly mortgage payments remaining constant when there is a delay in closing mortgage. hedging is accomplished by selling one treasury bond futures contract at time 0 and closing the position when the mortgage is obtained. the contract sold is the fourth nearest contract. percent reduction refers to the percentage decrease in the standard deviations of mortgage capacity when hedged. 110 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 is. this price is guaranteed to be non-negative, and may even result in a large profit if interest rates haveincreasedbut the smith’s decide against buying a house. note that if the smiths had used futures contracts to hedge and decided against buying a house, they would likely reap a similar profit on unwinding their position when ratesincrease. if rates decline, however, then the smiths would face substantial losses in unwinding their futures contracts without being able to use the offsetting gain in mortgage capacity caused by the declining rates. as before, the breeden and giarla data from january 1984 through december 1990 for 30-year fixed-rate mortgages and 20-year treasury bond futures are employed in different scenarios to test the feasibility of hedging an individual’s mortgage origination by purchasing put options. although liquidity for one year put options on treasury futures contracts is virtually nonexistent, a complete analysis regarding the hedging alternatives potentially available to individual borrowers requires that we simulate the put option strategies. we used the black commodity option model examined by luft and fielitz (1986) to approximate the price of the put options on the 20-year t-bond futures contracts. the black model for a put on a futures contract is: p 5 e2rt@xn~2d2! 2 fn~2d1!# (1) d1 5 $ln~f/x! 1 0.5s2t%/$s√t% d2 5 d1 2 s√t where f is the price of the underlying futures contract, x is the exercise price, t is the time until maturity, r is the rate of return on a risk free bond that matches the option maturity, and s is the volatility of the price changes for the underling futures contract. the breeden and giarla data provided us with the means to observe or compute all the required inputs except the risk free rates. breeden and giarla also provide rates on 1-year, 6-month, and 3-month treasury bills that match their other observation dates. these rates are used during the simulations as required. volatility estimation always is a critical issue in option pricing. given the amount of monthly data that we had, we chose to estimate the return volatility based on the prior year’s worth of monthly returns. there are two consequences to this decision. first: it means that we calculate our annualized sd based on only 12 return observations. second: it means that the first put option hedge could not be constructed until january of 1985. if we lengthen the estimation interval, then we obtain more degrees of freedom, but run the risk of a poor estimate of current volatility because of extremely old return data. furthermore, we would have fewer hedging opportunities. if we shorten the estimation interval then we gain more opportunities to hedge, but our estimated sds become unreliable because of too few degrees of freedom. we feel comfortable that one year’s worth of return data is stable enough to captures the true volatility of the underlying futures contract, while providing us with an adequate number of hedging opportunities. using the breeden giarlia data, we construct one year at the money put option hedges for the smiths. we find that the average annualized put premiums are 4.36%, and that the put deltas are approximately 0.5 in absolute value as expected. these figures are generally 111t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 consistent with the average option values given by maris and white (1989), sharp (1989) and kutner and seifert (1991). the inverse of the put option delta’s absolute value establishes how many put options must be purchased when attempting to hedge a drop in the futures contract price that is the result of rising interest rates. given these deltas, the individual would need to buy two put contracts to give about the same protection as one futures contract. thus, the individual would need to pay about 8.72% of the anticipated mortgage to provide protection against an increase in interest rates for one year. individual borrowers may recover some of this premium, especially if they decide against taking out a mortgage. if interest rates increase, the individual may be able to sell the put option for more than the purchase price offsetting the lower mortgage they would qualify for. if rates, however, decline substantially, the put will become worthless. table 2 presents the results of the put option hedging simulations. the “unhedged” results are identical to those presented in table 1. the “hedged” results reflect the outcomes of the put option strategies. observe that on average, the put options preserve the borrowing capacity quite well. if the average hedging performance of the put options is compared with the average hedging performance of the futures contracts, the options dominate; however, when the percentage reduction columns are compared, the put options perform horribly. recall that the futures hedge reduces the borrower’s exposure to changing interest rates in every hedge period. the data in the last column of table 2 show that the put options reduce the dispersion associated with borrowing capacity for only two hedge periods: the 10-month and 11-month periods. moreover, the improvement is minor when compared to the improvetable 2 hedging mortgage using put optionsa months delay unhedged hedged with two treasury bond futures puts percent reduction average sd maximum minimum average sd maximum minimum sd with hedge 0 100.00 0 100.00 100.00 100.00 0 100.00 100.00 0 1 100.39 3.23 108.43 90.84 103.14 10.52 130.04 82.77 2226.7 2 100.80 4.81 110.96 88.63 101.79 10.40 127.91 78.52 2116.2 3 101.22 5.88 112.94 86.70 101.46 9.84 125.33 79.58 267.3 4 101.66 6.63 116.68 86.92 101.18 9.56 124.92 75.67 244.2 5 102.21 7.23 118.84 88.21 100.98 9.56 124.35 76.07 232.2 6 102.78 7.69 120.97 84.16 100.26 9.83 124.24 78.21 227.8 7 103.29 8.09 122.26 82.32 99.96 9.60 123.74 80.30 218.7 8 103.80 8.59 122.36 82.70 99.87 9.69 123.61 79.96 212.8 9 104.24 9.11 123.69 84.77 99.38 9.87 123.60 80.75 28.3 10 104.69 9.79 128.59 82.47 99.17 9.64 123.60 83.43 1.2 11 105.12 10.42 130.97 84.79 98.60 9.35 123.60 83.73 10.3 a percent of original mortgage capacity based on monthly mortgage payments remaining constant when there is a delay in closing mortgage. hedging is accomplished by buying two treasury bond futures put options at time 0 and closing the position when the mortgage is obtained. the put option bought is the fourth nearest contract. the striking price is the nearest to the money. percent reduction refers to the percentage decrease in the standard deviations of mortgage capacity when hedged. note: the negative percent reduction in standard deviation means that the short-term hedges increased the risk. the average premium paid for one put is 4.36 (a total of $8,720 for the two puts used in this example). 112 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 ment achieved via the futures contracts: 10.3% versus 51.8% for the 11-month hedges, and 1.2% versus 49.5% for the 10-month hedges. finally, the negative values indicate that the put option hedges actually increase the risk to the borrower. why do the options perform so poorly? we believe that the poor hedging performance is caused by the fundamental differences between the two types of instruments. first, when a futures hedge is constructed, only the initial margin deposit is required. this $2,700 requirement is only a fraction of the $8,720 required to establish a hedge via the purchase of two put options. second, the at the money options provide the borrower with full insurance against falling interest rates for the life of the option. thus, the longer the option’s term, the more expensive the insurance coverage. third, the put options also provide the flexibility to take advantage of more favorable rates. if rates fall sufficiently far, then the put option will expire worthless, and the borrower can capture the much lower rate, thus lowering the effective borrowing rate. conversely, the futures contract locks in a borrowing rate when the hedge is established. even if rates drop dramatically, the borrower’s effective rate is determined by the futures hedge at the time it is created. the ability to capture a lower effective rate, coupled with the long term insurance makes the put option very expensive. finally, if interest rates remain stable and prices do not change, then no gains or losses accrue to the futures margin account and the futures hedger neither suffers a loss nor enjoys a gain. the options hedger will lose the entire amount paid for the options, however. the reason is that if the market remains stable, then the put options never gain any intrinsic value and thus expire worthless. the potential gains and losses due to the put option’s flexibility are reflected in the range of outcomes reported in table 2. note the relatively large differences, when compared to the futures values in table 1, between the maximum and minimum values for all the put option hedges. given these results, we believe that the best alternative available to an individual borrower for hedging unwanted mortgage interest rate risk is provided by interest rate futures contracts. mortgage interest rate hedges constructed via put options written on interest rate futures contracts do little to reduce a borrower’s interest rate exposure, and are too expensive to justify the lower equity requirements. 4.3. potential difficulties all hedging programs face potential difficulties that may prevent the hedgers from completely meeting their objectives. one of these difficulties involves what is known as tracking error. tracking error springs from the commodity underlying the futures contract being a different commodity than that traded in the spot market. in the case of hedging mortgages, no futures contract exists on newly issued mortgages (the spot commodity). therefore, the hedger will need to use treasury bond or treasury note futures. ten-year treasury note futures have a very high, but not perfect, correlation with current mortgage rates. if t-note rates temporarily zig while mortgage rates zag, tracking error will occur. basis risk is another difficulty faced by a hedger. basis risk occurs because the hedge is rarely held to maturity. when the futures contract expires, the spot and futures prices are certain to be equal; however, before maturity, the futures and spot prices can diverge. 113t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 although the basis, the difference between the spot price and futures prices, tends to decline over time, this decline is not perfectly predictable except at the moment of delivery. a third potential difficulty with a mortgage hedging program for individuals is that the individual may decide not to buy a house. this means that the individual will need to unwind the hedging position. in the case of the futures contract, this is easily done by entering an offsetting contract. this offsetting contract will lock in any gains or losses in the futures market. because, however, the individual will not be taking out a mortgage, there is no offsetting loss or gain in the spot market. thus, if interest rates have fallen when the individual decides to close out the futures position, there will be a net loss or decrease in wealth. this take-down risk is diversifiable in the case of the lender, but is not for the individual. a hedge using options is also subject to volatility risk. volatility risk causes the hedge option to less perfectly track the underlying asset’s price. if interest rate volatility increases, as is often the case when rates increase, the value of the option increases explosively. conversely, if rates stabilize, the volatility declines, dragging down the value of the option, so that less of the original premium will be recovered when the hedging position is unwound. 5. conclusions simulation shows that hedging individual mortgage rate risk with futures contracts reduces the sd in wealth available for housing by about 50%. although total variation is reduced, substantial downside risk remains. the expected cost of this hedge is near zero. hedging with put options on futures contracts is much less effective in reducing downside risk, and has a substantially greater cost. the typical annualized put premium (the cost of buying the put options) is about 5–10% of the mortgage value in our simulation time period. the actual cost could be greater or less, depending upon interest rate dynamics. furthermore, there is minimal reduction in variation of wealth available for housing when hedging with put options. this paper has established that hedging interest rate risk before taking out a mortgage is feasible for individuals, and potentially effective by normal financial measures. yet, the practice is virtually unknown. a possible explanation is that individuals are not aware of the potential advantages to entering a mortgage hedge and need education on the process. this paper provides a basis for a financial planner to aid an individual in making this choice; however, a financial planner may want to consider how this process would seem to an individual client. a number of papers in behavioral finance have suggested that people perform mental accounting that segregates gains and losses. this means that, rather than integrating the gains and losses (the essence of a hedging policy), individuals focus on the losing side of the hedge. because all effective hedges will have a losing side, the individual needs to be educated in financial thinking. this is a time-consuming project, that needs to be accomplished each time a client faces this prospect. furthermore, about half of the time, an individual will observeex postthat they would have been better off without the hedge and perhaps blame the advisor. 114 t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 acknowledgment the authors thank barrett lewald of the lind-waldock brokerage firm for insightful discussions on the institutional details of trading futures and futures options. references berkovitch, e. & greenbaum, s. i. (1991). the loan commitment as an optimal financing contract.journal of financial and quantitative analysis, 26(1), 83–95. breeden, d. t. (1991). risk, return, and hedging of fixed rate mortgages.journal of fixed income 1(2), 85–94. breeden, d. t. (1994). complexities of hedging mortgages.journal of fixed income 4(3), 6–41. breeden, d. t. & giarla, m. j. (1992). hedging interest rate risks with futures, swaps, and options. in f. j. fabozzi (ed.),handbook of mortgage-backed securities(3rd edition, chap. 35). chicago: probus publishing. chiang, r., gosnell, t. f. & heuson, a. j. (1997). evaluating the interest-rate risk of adjustable-rate mortgage loans.journal of real estate research, 13(1), 77–94. cross, c. (1998). cutting the hedge gap.mortgage marketplace 21(25), 2. fernald, j. d., keane, f. & mosser, p. c. (1994). mortgage security hedging and the yield curve.frb new york-economic policy review, 19(2), 92–100. follain, j. r. & park, h. y. (1989). hedging the interest rate risk of mortgages with prepayment options.review of futures markets 8(1), 62–78. goodman, l. s. & ho, j. (1997). mortgage hedge ratios: which one works best?journal of fixed income, 7(3), 23–34. hochstein, m. (1998). rate jump halts refis, puts locked-in lenders in a bind.american banker, 163(200), 11. johnston, e. t. & mcconnell, j. j. (1989). requiem for a market: an analysis of the rise and fall of a financial futures contract.review of financial studies, 2(1), 1–24. luft, c. f. & fielitz, b. d. (1986). an empirical test of the commodity option pricing model using ginnie mae call options.journal of financial research, 9(2), 137–151. kutner, g. w. & seifert, j. a. (1991). pricing the interest rate commitment cost in residential real estate lending. real estate finance journal, 7(2), 74–76. maris, b. a. & white, h. l. (1989). valuing and hedging fixed-rate mortgage commitments.journal of real estate finance and economics, 2(3), 223–232. murphy, a. & gordon, d. (1990). an empirical note on hedging mortgages with puts.journal of futures markets, 10(1), 75–78. patel, k. (1994). lessons from the fox residential property futures and mortgage interest rate futures market. housing policy debate, 5(3), 343–360. sharp, k. p. (1989). mortgage rate insurance pricing under an interest rate diffusion with drift.journal of risk and insurance, 56(1), 34–49. templeton, w. k., main, r. s. & orris, j. b. (1996). a simulation approach to the choice between fixed and adjustable rate mortgages.financial services review, 5(2), 101–117. 115t.l. zivney, c.f. luft / financial services review 8 (1999) 101–115 income more important than financial literacy for improving wellbeing tracey westa, michelle cullb,*, dianne johnsonc adepartment of accounting, finance and economics, griffith university, gold coast, queensland, australia bschool of business, western sydney university, nsw, australia cdepartment of accounting, finance and economics, griffith university, gold coast, queensland, australia abstract as advocates of financial literacy education, it is a hard pill to swallow when data show little impact on financial behaviors. unfortunately, expectations that university students with higher levels of financial literacy have reduced money management stress and positive financial behavior, leading to higher levels of financial wellbeing, were expunged in this study. we did find, however, that being older and having higher levels of income contributed most significantly and consistently to explaining better financial wellbeing. proponents of financial literacy education should not despair but instead recognize the limits to transferring financial knowledge and set financial literacy and wellbeing goals based on evidence of what works. © 2021 academy of financial services. all rights reserved. jel classification: d1 household behaviour and family economics; i22 educational finance; financial aid; i240 education and inequality keywords: education; financial wellbeing; financial literacy; income; university students 1. introduction what is more important to an individual; financial literacy or financial wellbeing? sure, financial mistakes are costly, but to what extent do they impact on an individual’s level of satisfaction with their financial situation? further, how can financial literacy interventions effectively improve wellbeing outcomes? financial wellbeing therefore, is a topic of *corresponding author. tel.: +61 2 425 000 038; fax: +61 2 4620 3788. e-mail address: m.cull@westernsydney.edu.au (m. cull) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 187–207 increasing importance to academics, public policy officials, educators, financial managers, and employers (cfpb, 2015). while there is a plethora of studies available that advocate for consumer protection through financial literacy interventions (fernandes, lynch, & netemeyer, 2014) and policy support for financial literacy education in an attempt to increase economic participation, improve social inclusion and enhance economic health (asic, 2017; oecd, 2012), there has been less work done on what constitutes financial wellbeing or its role in overall wellbeing (netemeyer, warmath, fernandes, & lynch, 2018). this study utilizes the financial wellbeing framework of netemeyer et al. (2018) to measure the determinants of financial wellbeing of university students. university students are an important cohort of interest, and many institutions are discovering that first-generation, students of color, adults and military veterans are the “new majority” (lyon & matson, 2019). thus, universities represent a diverse array of people with differing socio-economic backgrounds and are at a pivotal point in their life, making decisions that affect their financial futures. using a survey of 420 students from an australian university in 2019, we quantify the impact of current money management circumstances, attitudes towards future finances and financial literacy on levels of financial wellbeing. ordered logit results find that the financial wellbeing framework applied to our sample does not explain financial wellbeing particularly well. financial literacy does not play an important role, but income does. these findings may be context specific, as students experience low and irregular incomes, which can lead to increased vulnerability to external shocks and uncertainty. however, many other workers experience irregular incomes due to the rise of the gig economy (farrell & greig, 2016; kaine, oliver, & josserand 2017; stewart & stanford, 2017). accordingly, this paper is set out as follows. section 2 provides an overview of the theoretical framework and a review of the literature. the data and methodology are discussed in section 3. results are presented in section 4 and the paper concludes with a discussion in section 5. 2. background there are a number of definitions of financial wellbeing that are being used in academic literature, industry reports and government policies (anz, 2018). however, due to the absence of a widely accepted definition and measure of financial wellbeing, efforts to examine the financial domain have been hampered (netemeyer et al., 2018). internationally, the consumer financial protection bureau (cfpb) report in 2015 provided a consumer driven definition of financial wellbeing as “a state of being wherein a person can fully meet current and ongoing financial obligations, can feel secure in their financial future, and is able to make choices that allow enjoyment of life” (p. 18). this definition has also been adopted by the oecd (oecd, 2020). the cfpb report highlights multiple factors that affect the level of financial wellbeing of an individual with a wide variation in how people in the united states feel about their financial wellbeing. key findings were that having a savings safety 188 t. west et al. / financial services review 29 (2021) 187–207 net had the strongest relationship to financial wellbeing, given that feeling financially secure is fundamental to the definition of financial wellbeing. certain experiences with debt and credit, however, seem to have the strongest negative relationship with financial wellbeing. individual characteristics were also factors associated with financial wellbeing; those with higher levels of education, older individuals and adults in better physical health tended to have higher levels of financial wellbeing (cfpb, 2015). average financial wellbeing, however, appeared to be the same for men and women. although financial circumstances were highly correlated with financial wellbeing scores, the report found that individuals with different experiences can arrive at the same score, suggesting that no single factor is responsible for, or indicative of, an individual’s level of financial wellbeing. in australia, research undertaken by muir et al. (2017) used an ecological systems approach to explore financial wellbeing. financial wellbeing was said to consist of three interrelated dimensions. the first was having adequate income to pay off debt, meet basic needs and cover unexpected expenses with some money left over. the second was feeling and acting in control of finances, and the third included feeling financially secure. financial wellbeing has objective (savings) and subjective (how the person is feeling) components. muir et al. (2017) use this lens to look at individual, household, family, peer-level, community, and social influences on financial wellbeing, and find that financial capability, financial inclusion, social capital, and economic resources (especially income) are among the strongest influencers of financial wellbeing. thus, improvement in each of these four areas are likely to enhance a person’s financial wellbeing both in times of financial adversity and in the context of everyday money management. similar to the (cfpb, 2015) research, having savings and building resilience for unexpected expenses were both important. in addition, cfpb research finds social capital to be significantly associated with financial wellbeing, that is, having support from others as well as access to resources if needed. further, a study by collins & urban (2020) in the united states found financial well-being to generally follow the life cycle, increasing with income and savings levels as well as with age. they also found that levels of financial wellbeing were not strongly associated with financial literacy. recent contributions to developing a better understanding of financial wellbeing have applicability to university students. netemeyer et al. (2018) suggests that the definition of financial wellbeing presented by the cfpb (2015) overweights current money management concerns. they argue that people under current money management stress can still expect to be financially better off in the future and explains prior studies of self-reported higher levels of financial wellbeing than would be expected giving current circumstances (berman, tran, lynch, & zauberman, 2016; finke, howe, & huston, 2017; johnson & krueger, 2006). as such, individuals who perceive their circumstances to be modifiable will be more likely to engage in self-improvement actions (summerville and roese, 2008). consequently, netemeyer et al. (2018) disentangled financial wellbeing into two related but separate constructs; current money management stress and expected future financial security, as shown in fig. 1. the research of netemeyer et al. (2018) was particularly interesting regarding the impact of income on financial wellbeing, as it was not found to be a direct positive predictor. instead, income moderates the effect of current money management stress on wellbeing, and as income levels rise the negative effect of money management stress on an individual’s wellbeing dissipates. holding constant other factors and the perceived t. west et al. / financial services review 29 (2021) 187–207 189 financial wellbeing constructs, income only increases overall wellbeing when current money management stress is high (netemeyer et al., 2018). that is, current money management stress has a serious detrimental effect on wellbeing among low-income individuals. for low income earners, the focus should be on reducing debilitating current money management stress that is not something that increasing financial literacy could likely achieve. for students, this framework is especially salient. students can be optimistic about their future income prospects due to their human capital investment in the program of study. thus, they may heavily weight the future and discount the present financial discomfort if seen as a short-term circumstance. current research by timmerman & volkov (2019) investigates the impact of career choice and education level on an individual’s overall wealth by finding the present values of future earnings for various occupations and makes some interesting comparisons by trading off against the human capital investment required. for example, one would expect future incomes to be higher for doctors and dentists than human resource advisers, but the human capital investment is often also higher. as all students are more likely to be relatively low-income earners due to giving up income generating opportunities to study, it is important to quantify the extent to which financial literacy moderates current money management stress. university campuses host a wide variety of students—from school leavers to career changers to employed professionals and international students. accordingly, it can be fig. 1. potential antecedents and consequences of perceived financial well-being. source: netemeyer et al. (2018, p. 72). 190 t. west et al. / financial services review 29 (2021) 187–207 difficult to make assumptions about the typical student. however, for many, study accompanies a period of reduced income and independent living for the first time, which provides a financial challenge. surveys often report that students struggle to afford basic study support tools such as textbooks, for example (dean & forray, 2018; senack, 2015). there is a wide literature on the financial stress of students, usually originating out of the united states, where student debts are similar to bank loans and total over $1.5 trillion (williams & oumlil, 2015). australian students are less likely to experience hardships caused by this style of student loan due to the income contingent nature of the australian government higher education loan program (help; west, 2020). however, the help student debt still has the ability to hinder students’ future borrowing capacity and a deficit in financial literacy can mean that these students are more likely to underestimate future student loan payments and hence be more vulnerable to unexpected financial shocks postgraduation (artavanis & karra, 2020). evidence also suggests that student debt anxiety is a factor that may affect the wellbeing of students. harrison and agnew (2016) found that when student confidence in their education as an investment was higher, debt anxiety was lower. having a tertiary degree is linked to financial wellness by a higher magnitude than student debt is linked to financial stress (henager & wilmarth, 2018). this distinction is important for students to understand when making study decisions. these perceptions were also found to be connected to subject choices. business students had higher confidence in a return on their education than that of social science students, a likely reflection of graduate salary expectations (luthans, luthans, & chaffin, 2019; peach & yuan, 2017). on the matter of study choice, cull & whitton (2011) survey 472 students in sydney and find that financial knowledge is dependent on field of study, income, and age. for example, science students scored better on questions regarding interest, but regarding fees, tax and student debt knowledge, income was a better predictor. u.s. studies also point to poor financial behaviors demonstrated by college students. for example, mae (2009) found that over half of college students had four or more credit cards, and 90% indicated using credit to pay for education expenses including text books, school supplies, and commuter costs. in addition, many survey respondents appeared to use credit cards to live beyond their means. studies also found poorer behaviors among females. female students in the united states were likely to carry a higher number of credit cards, exhibit more problematic credit card behaviors (e.g., not paying bills on time), and asking parents for help to pay bills (hancock, jorgensen, & swanson, 2013; norvilitis, szablicki, & wilson, 2003; worthy, jonkman, & blinn-pike, 2010). australian females are similar. ha (2013) surveyed 257 students at universities in melbourne and found that senior female students had irresponsible patterns of credit card use while junior students had responsible patterns. many female students sought financial help from friends and family or approached external sources of help such as financial counsellors, government, and non-government agencies. a further challenge to managing finances is the pressure to conform to social norms through consumer spending that is particularly challenging for young adults (georgarakos, haliassos, & pasini, 2014; spencer, nieboer, & elliott, 2015; vaitilingam, 2016). young people are more likely to hold potentially destructive beliefs about money, with materialism being a personality trait likely internalized early in life (mentzer, klontz, klontz, & britt, t. west et al. / financial services review 29 (2021) 187–207 191 2011; richins, 2004) and personality found to be an important predictor of financial satisfaction (tharp, seay, carswell, & macdonald, 2020). lifestyle aspirations spurred on by influence of various forms of media or peers are likely to increase young people’s reliance on debt (fear and o’brien, 2009). the introduction of new financial products such as afterpay and zip pay have further exacerbated the spending on non-essential items by young people, with almost a quarter of zip pay customers under the age of 24 (dutta, singh, & sultana, 2019). students, therefore, may undertake risky financial behaviors, especially low-income students due to the limited availability of financial resources (bester et al., 2008). finally, financial literacy researchers concur that a lack of knowledge of financial concepts before entering tertiary study contributes to an experience that can be financially stressful. studies find that in general, young people, older people, women, and minority groups have lower levels of financial literacy (west & worthington, 2018; wilkins, 2018). lusardi, mitchell, & curto (2010) found that financial literacy of young people was poor in the united states, leading to a long list of negative consequences. these consequences include problems with debt (lusardi & tufano, 2009), reduced stock market participation and risk taking (van rooij, lusardi, & alessie, 2007; west & worthington, 2014), lower likelihoods of choosing investments with lower fees (hastings & tejeda-ashton, 2008), lower likelihoods of accumulating wealth and managing wealth effectively (hilgert, hogarth, & beverly, 2003; stango & zinman, 2007) and lower likelihood of planning for retirement (bongini & cucinelli, 2019; lusardi & mitchell, 2006, 2007, 2009). more recent studies in the united states continue to show alarmingly low levels of financial literacy among undergraduate students with artavanisa and karra (2020) finding a literacy rate of 39.5% in addition to a large gender gap with female students exhibiting considerably lower literacy rates (26%) than their male peers (56%). compared with earlier studies in the united states, such as chen and volpe (1998), it seems that financial literacy has not improved and continues to limit the ability of students to manage their finances and make informed decisions. overall, the combined lack of financial knowledge and limited availability of financial resources for tertiary students may contribute to lower levels of self-reported financial wellbeing. students with lower levels of financial literacy are more likely to mismanage their finances now and, in the future, contributing to poor financial behaviors that are prevalent in today’s society (jorgensen, 2007). this is further supported by philippas and avdoulas (2020) who found that financially literate students have a 1.8 times higher possibility of having higher levels of financial well-being than financially illiterate students. they also found that the financial fragility of students had a significant impact on financial wellbeing with no financially fragile students showing higher levels of financial well-being. this study contributes to the literature by applying the financial wellbeing framework of netemeyer et al. (2018) to the australian university student context. as the literature highlights, this cohort has particular financial challenges and are actively making decisions that affect financial futures. accordingly, we hypothesize that expectations of the future play a more significant role in financial wellbeing outcomes than current financial stress, as students are likely to see their current situation as temporary. thus, the financial wellbeing model may have different outcomes when applied to university students than the general population. we also consider gender differences in financial literacy and how this impacts 192 t. west et al. / financial services review 29 (2021) 187–207 financial wellbeing. this study addresses these gaps in the literature and provides practitioners and educators with an understanding of where interventions are best targeted. 3. data and methodology this study applies the novel financial wellbeing framework as adopted from netemeyer et al. (2018) to investigate the role of financial literacy in improving financial wellbeing outcomes. data for this study was obtained from a survey of students from an australian university in 2019, ethics approval (2019/160). a monetary incentive by way of a prize draw was provided to improve response rates (yu et al., 2017); 420 students responded to the survey, providing a good sample size. however, it was only 0.9% of the total number of students enrolled at the university, even though all students were invited to participate via a broadcast email. when interpreting results, the relatively small sample size and distribution of the characteristics may not be representative of the larger tertiary student cohort. for example, 63% of respondents were female, while around 58% of university students at this institution were female. further, the average respondent is aged 23 years or younger. this presents implications for interpretation and generalization of results, as it may be that financial knowledge is not well developed in young adults and difficult to detect and measure accurately. table 1 provides further descriptive statistics of the sample. the average respondent is female, aged 23 or younger, earns under $20,000 a year, and is studying a subdegree qualification like a diploma, advanced diploma, or associate degree. a fundamental aspect of financial wellbeing is the overall financial profile of a person or household. however, two people or two households with the same financial situations might perceive their circumstances differently. perceived financial wellbeing, which consists of stress related to money management as well as feelings of security in one’s financial future, in fact, maps to financial wellbeing, as well as to overall subjective well-being (netemeyer et al., 2018). we apply a the netemeyer et al. (2018) financial wellbeing framework (as described in fig. 1), given by finsati ¼ ai þ b 1current money management stress þ b 2future expectationsþ b 3control þ b 4financial literacyþmi where, current money management stress is a set of variables used to describe students’ current money management circumstances. the framework by netemeyer et al. (2018) suggests that being late or making minimum payments on bills and credit cards, lack of self-control, materialism, and perceived financial self-efficacy are antecedents of current money management stress. we explore the relevance of a variety of indicators of current money management stress offered in the survey, and narrow down variables for inclusion in the regression through factor analysis. we include the variables in the analysis that are presented in table 2. t. west et al. / financial services review 29 (2021) 187–207 193 an investigation of the descriptive statistics shows that paying bills is a problem for students. over 50% of students regularly make only the minimum monthly payment on their credit cards or pay nothing (ccpay) and find it difficult to cover expenses and pay bills (bills). a slightly lower proportion (43%), indicate that it is difficult to come up with $500 to cover emergency expenses (emg500). these variables are coded into binary variables where 1 is equal to higher levels of financial stress, and an inverse relationship with financial wellbeing is predicted. lack of self-control is proxied by two variables, usedebt and spendmore. concerningly, 12.88% of respondents say they use debt so they do not miss out on student experiences, and 12.07% say they regularly spend more than they have by using credit or borrowing. the responses are coded so that a higher level equates to lacking self-control that is predicted to have an inverse relationship with financial wellbeing. the factor analysis for a set of questions relating to materialism showed that matimp and mathap had the highest loadings on the first factor. the descriptive statistics are interesting. just under 14% of respondents indicate that acquiring material possessions is an important achievement, while a much larger cohort (49.26%) indicate that they would be happier if they could afford to buy more things. no doubt the latter is representative of the constrained budgets of university students. a higher response level is expected to be associated with an inverse relationship with financial wellbeing. table 1 descriptive statistics set of personal factors proportion (%) mean sd what is your gender? gender 0.63 0.48 0– male 37.00 1– female 63.00 what age category are you in? agec 1.58 1.16 1– 23 or younger 52.05 2– 24 to 29 24.38 3– 30 to 39 16.44 4– 40 to 49 4.38 5– 50 to 59 2.19 6– 60 or over 0.55 what is your current annual income, including paid work, government benefits and other financial support? income 4.05 2.40 1– above $100,000 1.67 2– $80,000–$99,999 3.57 3– $60,000–$79,999 4.28 4– $40,000–$59,999 8.80 5– $20,000–$39,999 20.23 6– $1–$19,999 35.71 7– $0 4.76 what type of degree are you currently pursuing? edu 3.31 0.66 1– preparation program 1.57 2– diploma/advanced diploma/ associate degree 70.08 3– bachelor degree 23.62 4– postgraduate degree 3.94 5– phd 0.79 194 t. west et al. / financial services review 29 (2021) 187–207 perceived financial self-efficacy is related to control over one’s financial situation. two variables are included as proxies: finstress and conf. alarmingly, only 7.33% of respondents indicate that they do not feel stressed about their personal finances. responses to the question regarding how confident they feel about managing their finances is contradictory. only 3.37% of respondents say that they do not feel confident, meaning that most of the population has some level of confidence with managing their finances. we expect a negative coefficient for finstress and a positive coefficient for conf. table 2 current money management stress descriptive statistics current money management stress proportion (%) mean sd late minimum payments: when you get a credit card or other bill, do you usually: ccpay 0.48 0.50 1– make the minimum monthly payment/pay more than the minimum, sometimes pay nothing or miss the payment date 52.05 0– pay the full balance/someone else pays my bill 47.95 in a typical month, how difficult is it for you to do the following: to cover your expenses and pay all your bills? bills 0.50 0.50 1– always/often/sometimes 50.26 0– rarely/never 49.74 to come up with $500 to cover emergency expenses? emg500 0.57 0.50 1– always/often/sometimes 43.43 0– rarely/never 56.57 lack of self-control: i use debt so i do not miss out on “normal” student experiences usedebt 1.60 0.71 1– does not describe me 52.53 2– describes me very little/somewhat describes me 34.60 3– describes me very well/describes me completely 12.88 regularly spend more than i have by using credit or borrowing spendmore 1.58 0.70 1– does not describe me 54.43 2– describes me very little/somewhat describes me 33.50 3– describes me very well/describes me completely 12.07 materialism: some of the most important achievements in life include acquiring material possessions. matimp 1.81 0.66 1– does not describe me 32.68 2– describes me very little/somewhat describes me 53.41 3– describes me very well/describes me completely 13.90 i’d be happier if i could afford to buy more things. mathap 2.41 0.64 1– does not describe me 8.37 2– describes me very little/somewhat describes me 42.36 3– describes me very well/describes me completely 49.26 perceived financial self-efficacy. i feel stressed about my personal finances in general. finstress 2.44 0.68 1– does not describe me 7.33 2– describes me very little/somewhat describes me 41.08 3– describes me very well/describes me completely 51.59 i am confident i can manage my finances conf 2.55 0.56 1– does not describe me 3.37 2– describes me very little/somewhat describes me 38.46 3– describes me very well/describes me completely 58.17 t. west et al. / financial services review 29 (2021) 187–207 195 future expectations is a set of variables used to describe the student’s behaviors that are likely to lead to a positive financial outcome. the framework by netemeyer et al. (2018) suggests that perceived financial self-efficacy, positive financial behaviours, willingness to take investment risks and planning for the long term are antecedents for expected future financial security. we consider responses to several questions about the future that are pertinent to students as proxy for future expectations, as described in table 3. two variables that serve as proxy for positive financial behaviours were selected from a set of questions based on factor analysis: psav and pplan. regularly adding to savings is identified by 46.32% of respondents and 58.64% plan ahead for major purchases. higher scores are predicted to positively relate to financial wellbeing. the highest response to a single category for willingness to take financial risks is “i am not willing to take any financial risks” (36.93%). however, 42.93% of respondents did choose a category of willingness to take financial risks to various degrees. a high response is expected to relate positively to financial wellbeing. finally, planning for the long term is represented by plan. the longer the time period selected, the more positive an impact on financial wellbeing. while 29.33% of students are only planning for the next few months, a large portion (70.67%) of them are looking years ahead. table 3 future expectations descriptive statistics future expectations proportion (%) mean sd positive financial behaviors i add to my savings on a regular basis. psav 2.31 0.72 1– does not describe me 15.20 2– describes me very little/somewhat describes me 38.48 3– describes me very well/describes me completely 46.32 i plan ahead for major purchases. pplan 2.54 0.60 1– does not describe me 5.31 2– describes me very little/somewhat describes me 35.75 3– describes me very well/describes me completely 58.94 willingness to take investment risks which of the following statements comes closest to describing the amount of financial risk that you are willing to take with your spare cash? that is, cash used for savings or investment. frisk 2.40 1.03 1– i never have any spare cash 20.14 2– i am not willing to take any financial risks 36.93 3– i take average financial risks expecting average returns 29.26 4– i take above-average financial risks expecting to earn above-average returns 10.31 42.93 5i take substantial financial risks expecting to earn substantial returns 3.36 plan for money long-term in planning for saving and spending, which of the time periods are most important? plan 3.53 1.20 1– longer than 10 years 5.05 2– the next 5 to 10 years 15.14 3– the next few years 30.77 4– the next year 19.71 5– the next few months 29.33 196 t. west et al. / financial services review 29 (2021) 187–207 control is a set of constant personal characteristics, including demographics, socioeconomic status, and financial literacy. studies show that gender, income, education, age, and financial literacy are related to financial outcomes (lusardi & mitchell, 2011). the descriptive statistics of gender, agec, income, and edu are presented previously in table 1. table 4 provides detailed information about the financial literacy variable(s). as financial literacy is an assessment of objective knowledge of financial concepts, represented by responses to the “big three” questions on compound interest, inflation and diversification, it is included as a control and not a predictive variable so as not to confuse the constructs with a correlation with financial literacy. the responses to the three financial literacy questions in table 4 are of interest. “no” is the correct answer for all three questions. students generally do well on the first question regarding compound interest (finlc), with 70.95% of respondents correct. however, only 45.71% of students responded correctly for the second question on inflation (finli) and 34.29% were correct for the third question on diversification (finld). respondents to this survey underperform the general population, measured by responses to the same questions in the household, income, and labor dynamics in australia (hilda) survey (wilkins, 2018). for comparison purposes, 85.5% of the australian population select the correct answer for the compound interest question, 69.8% for the inflation question, and 74.9% for the diversification question (wilkins, 2018). we sum the responses to create a single financial literacy score (finlitscore). those with a score of 3 responded correctly to all three questions, which is just under a quarter of respondents (24.05%). table 4 financial literacy descriptive statistics financial literacy proportion (%) mean sd if you invested $100 today and the interest rate was 2% per year your bank account balance after five years would be exactly $102 finlc 2.38 1.05 3 no 70.95 2 yes 6.19 1 unsure 12.62 0 don’t care 10.24 after 1 year you would be able to buy more than today if you invested $100 in your bank account today at an interest rate of 1% per year when inflation is 2% per year finli 1.94 1.10 3 no 45.71 2 yes 14.29 1 unsure 27.86 0 don’t care 12.14 buying shares in a single company usually provides a safer return than buying units in a managed share fund finld 1.63 1.07 3 no 34.29 2 yes 5.48 1 unsure 49.29 0 don’t care 10.95 financial literacy score finlitscore 1.51 1.10 0– 0 correct 23.57 1– 1 correct 25.95 2– 2 correct 26.43 3– all correct 24.05 t. west et al. / financial services review 29 (2021) 187–207 197 finally, the dependent variable of interest is finsat. finsat is the response to “on a scale of 1 to 10, with 10 being totally satisfied, all things considered, how satisfied are you with your financial situation?” fig. 2 provides the distribution of responses. the mean is 5.35 (sd = 2.48), indicating that overall respondents are more satisfied than not with their financial situation. due to the ordered nature of this variable, we employ an ordered logit model for analysis. this analytical technique is appropriate as the dependent variable is discrete (that is, can only take the values of 1 through 10) and the values in each category have a meaningful sequential order (west & worthington, 2014). the ordered logit model estimates an underlying score as a linear function of the independent variables and a set of cut-points (cameron & trivedi, 2009), and the probability of observing outcome i corresponds to the probability that the estimated linear function plus random error is within the range of the cut-points estimated for the outcome: pr outcomej ¼ ið þ ¼ pr ki�1 < b ixij þ b 2x2j þ . . .þ b kxkj þ uj ≤ kið þ where uj is logistically distributed in the ordered logit, xkj is a vector of control variables with estimated coefficients b1, b2, . . . bk and cut-points k1, k2, . . . kk-1, where k is the number of possible outcomes, k0 is taken as –1, and kk is taken as +1. the estimated coefficients b and the cut-point parameters are obtained using maximum likelihood methods. the sign of the estimated coefficients can be immediately interpreted as determining whether the dependent variable increases with the independent variables (cameron & trivedi, 2009). 4. results table 5 provides the odds ratios and standard errors of eight ordered logit regressions. the f-tests for all models rejected the null hypothesis that all slope coefficients are zero at the 0.001 level, implying that they are appropriate for predicting financial wellbeing. the eight models are variants on the financial wellbeing framework and omit sets of variables to fig. 2. distribution of satisfaction with financial situation, where 10 is very satisfied. 198 t. west et al. / financial services review 29 (2021) 187–207 t ab le 5 o rd er ed lo g it re su lt s fo r fi n an ci al w el lb ei n g 1 2 3 4 5 6 7 8 p ar am et er e x p ec te d si g n o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o c c p a y � 0 .6 6 6 1 .2 2 0 1 .3 4 0 0 .7 0 9 0 .2 7 7 0 .5 6 2 0 .6 2 5 0 .2 9 8 b il l s � 0 .5 9 5 0 .6 7 0 0 .6 2 0 0 .5 7 2 0 .2 4 5 0 .2 9 4 0 .2 7 5 0 .2 3 7 e m g 5 0 0 � 0 .2 4 0 * * * 0 .2 8 9 * * 0 .3 0 3 * * 0 .2 4 8 * * * 0 .1 0 5 0 .1 3 3 0 .1 3 9 0 .1 0 9 u s e d e b t � 1 .5 6 3 * 1 .6 0 6 * 1 .6 9 8 * 1 .6 0 4 * 0 .4 1 8 0 .4 4 2 0 .4 7 2 0 .4 3 1 s p e n d m o r e � 0 .5 0 0 * 0 .4 8 2 * 0 .4 7 9 * * 0 .5 0 2 * * 0 .1 4 1 0 .1 4 3 0 .1 4 1 0 .1 4 1 m a t im p � 2 .5 1 9 * * 2 .2 9 6 * 2 .2 9 0 * * 2 .5 1 9 * 0 .8 0 1 0 .7 7 7 0 .7 7 8 0 .8 0 0 m a t h a p � 0 .7 1 7 * 0 .7 0 0 0 .6 9 7 0 .7 1 7 * 0 .2 0 8 0 .2 1 1 0 .2 1 1 0 .2 0 8 f in s t r e s s � 0 .6 2 0 * * 0 .7 4 0 0 .7 3 7 0 .6 2 1 * * 0 .2 0 8 0 .2 6 5 0 .2 6 1 0 .2 0 7 c o n f + 1 .3 1 8 1 .5 6 1 1 .5 6 3 1 .3 2 1 0 .4 1 9 0 .5 7 2 0 .5 7 4 0 .4 2 1 p s a v + 2 .0 4 6 * * * 1 .4 4 4 1 .5 0 2 2 .0 5 4 * * * 0 .3 0 6 0 .4 3 6 0 .4 5 3 0 .3 0 7 p p l a n + 1 .3 5 5 * 0 .9 9 1 0 .9 6 3 1 .3 5 2 * 0 .2 2 9 0 .3 2 4 0 .3 1 6 0 .2 2 9 f r is k + 1 .8 1 6 * * * 1 .5 7 6 * * 1 .5 2 6 * 1 .7 9 5 * * * 0 .1 9 0 0 .3 1 1 0 .3 0 2 0 .1 9 1 p l a n + 0 .7 5 5 * * 0 .8 2 4 0 .8 2 9 0 .7 5 8 * * 0 .0 6 1 0 .1 2 6 0 .1 2 7 0 .0 6 2 g e n d e r 1 .3 8 4 1 .2 2 0 1 .8 4 0 2 .1 8 8 1 .8 3 2 1 .2 5 9 0 .9 5 7 (c o n ti n u ed o n n ex t p a g e) t. west et al. / financial services review 29 (2021) 187–207 199 t ab le 5 (c o n ti n u ed ) 1 2 3 4 5 6 7 8 p ar am et er e x p ec te d si g n o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o o d d s ra ti o 0 .5 1 9 0 .2 6 3 0 .7 2 0 0 .9 0 5 0 .7 0 6 0 .2 7 9 0 .2 0 1 a g e c 0 .6 6 8 0 .8 7 1 0 .8 0 4 0 .7 8 1 * 0 .8 0 7 * * 0 .8 6 5 0 .8 4 2 * 0 .1 1 0 0 .0 7 7 0 .1 2 3 0 .1 2 2 0 .1 1 9 0 .0 7 7 0 .0 7 6 in c o m e 0 .7 4 3 * 0 .9 2 2 * 0 .8 6 6 0 .8 6 4 0 .8 5 5 * * 0 .9 2 0 * 0 .9 2 7 * 0 .0 8 9 0 .0 4 1 0 .0 8 4 0 .0 8 1 0 .0 8 0 0 .0 4 1 0 .0 4 1 e d u 1 .0 2 3 1 .1 9 8 0 .9 8 3 0 .9 2 6 0 .9 2 2 1 .1 8 6 1 .4 5 8 * * 0 .2 3 6 0 .1 7 5 0 .2 1 8 0 .2 1 0 0 .2 0 2 0 .1 7 5 0 .2 2 8 f in l it s c o r e + 1 .2 9 9 1 .1 7 1 * * 1 .1 7 7 1 .0 5 9 1 .1 9 3 * 0 .2 5 6 0 .0 9 4 0 .2 1 8 0 .1 0 2 0 .1 1 0 p se u d o r 2 0 .1 3 7 0 .0 8 3 0 .1 7 6 0 .1 7 2 0 .0 0 2 0 .1 3 8 0 .0 8 3 0 .0 1 0 l r x 2 7 9 .0 7 * * * 1 3 5 .7 8 * * * 9 0 .5 8 9 5 .6 3 * * * 0 .0 5 * 7 9 .8 5 * * * 1 3 6 .1 * * * 1 6 .2 7 * n o te s: m o d el s sp ec ifi ed as fo ll o w s: 1 c u r r e n t m o n e y m a n a g e m e n t s t r e s s + c o n t r o l 2 + f u t u r e e x p e c t a t io n s + c o n t r o l 3 c u r r e n t m o n e y m a n a g e m e n t s t r e s s + f u t u r e e x p e c t a t io n s + c o n t r o l 4 c u r r e n t m o n e y m a n a g e m e n t s t r e s s + f u t u r e e x p e c t a t io n s + c o n t r o l + f in a n c ia l l it e r a c y 5 + f in a n c ia l l it e r a c y 6 c u r r e n t m o n e y m a n a g e m e n t s t r e s s + c o n t r o l + f in a n c ia l l it e r a c y 7 + f u t u r e e x p e c t a t io n s + c o n t r o l + f in a n c ia l l it e r a c y 8 + c o n t r o l + f in a n c ia l l it e r a c y 200 t. west et al. / financial services review 29 (2021) 187–207 test for predictive power. the signs of the odds ratio indicate the effect on financial wellbeing. if the estimate is positive, then an increase in the dependent variable necessarily decreases the probability of being in the lowest financial wellbeing category and increases the probability of being in the highest financial wellbeing category. a summary of the models tested is provided in table 5. comparison of the pseudo r2 shows the model with the most explanatory power is model 3 (0.176), that includes all variables except finlitscore. all coefficients are positive in this model, with the significant factors including emg500, usedebt, spendmore in the set of current money management stress factors, matimp and frisk in the future expectations set, and no significant factors in the set of control variables. within the set of current money management stress variables, emg500 is highly significant and positive across all four models it is included in. positive responses to this binary variable indicate issues with accessing emergency funds, so the positive relationship with financial wellbeing is puzzling as we were expecting a negative association. other variables like usedebt, spendmore, matimp, mathap, and finstress vary in significance across models but were all positive, when negative signs were expected. we draw the conclusion that current financial stress, while a significant determinant of financial wellbeing, is not inversely related. therefore, students under weigh financial stress when asked to quantify their level of financial wellbeing, suggesting that other factors are more important. we sense what other factors are more important when examining the set of future expectations indicators. for the set of future expectations, psav, pplan, frisk, and plan had very significant positive coefficients when the current money management stress variables were excluded from the model, as well as strong coefficients. when the current money management stress variables were included (models 3 and 4), all but frisk lost significance. we infer from this that people who are willing to take financial risk have an innate understanding of the time value of money that translates into confidence in their financial futures. finally, of the set of personal attributes, income is significant in five out of eight models, and agec is significant in three out of eight models. this makes logical sense, as higher incomes, which are associated with being older, overcome barriers to perceived wellbeing by facilitating choice and lifestyle purchases. for example, higher incomes afford people to both purchase medicine and groceries, while people living hand-to-mouth have to trade-off between necessities. this finding contributes to discussions that a good income affords many benefits: financial stability, enables future planning, facilitates acting on money beliefs, and practice making financial decisions. interestingly, finlitscore is only significant in model 5 (as the only factor in the model) and model 8 (only includes control variables). both of these models have low explanatory power. given the focus on financial literacy in the hypothesis, we conducted further tests of the finlitscore as the dependent variable. results of the ordered logit model, marginal effects of the highest level (outcome 3), and an ordered logit retaining a sample of those that scored the highest level are provided in table 6. these models answer questions as to what factors are likely to contribute to having a high financial literacy score. the marginal effects of the highest outcome provide more consistency with our original expectations. ccpay and usedebt were significant with negative coefficients. people with higher levels of financial literacy, therefore, pay their bills on time and do not use debt. frisk is also t. west et al. / financial services review 29 (2021) 187–207 201 positive and significant, providing further evidence that people willing to take financial risk have a good level of financial knowledge. gender has a negative association that is well supported by the literature that females score less well on questions of financial literacy than men. our analysis concludes that financial literacy and financial wellbeing are not related constructs. however, we do find that good money habits like saving and planning contribute to higher levels of financial wellbeing, but mostly when current money management circumstances are excluded. when both data sets are included together, current financial stress dominates as a determinant of financial wellbeing. importantly though, current financial stress is not the single determinant of financial wellbeing, and respondents seem to include table 6 ordered logit results for financial literacy score ologit marginal effects parameter odds ratio odds ratio ccpay 0.488 �0.118 * 0.229 0.077 bills 1.791 0.096 0.779 0.072 emg500 0.756 �0.046 0.260 0.079 usedebt 0.545 * �0.100 ** 0.159 0.047 spendmore 1.073 0.012 0.313 0.048 matimp 0.952 �0.008 0.351 0.061 mathap 0.974 �0.004 0.293 0.049 finstress 0.917 �0.014 0.347 0.062 conf 1.362 0.051 0.501 0.060 psav 0.911 �0.015 0.281 0.051 pplan 0.824 �0.032 0.286 0.057 frisk 1.452 * 0.061 * 0.287 0.032 plan 0.919 �0.014 0.138 0.025 gender 0.404 ** �0.149 ** 0.162 0.063 agec 1.262 0.038 0.203 0.026 income 1.007 0.001 0.101 0.017 edu 1.472 0.064 0.352 0.039 pseudo r2 0.128 lr x2 43.47 ** 202 t. west et al. / financial services review 29 (2021) 187–207 their preparedness for the future into their level of financial wellbeing. this is especially salient for university students who expect their level of income to increase in the future as a result of graduating with a university qualification and ability to commit to working longer hours once their studies are completed. these findings contribute to the literature on financial wellbeing and the application of the netemeyer et al. (2018) financial wellbeing framework, as specified and with the limits of the data. for educators and practitioners, we highlight the importance of good financial savings habits, financial risk-taking and income in achieving higher levels of financial wellbeing in clients and students. finally we note that these findings may be context specific, as students experience low and irregular incomes, which can lead to increased vulnerability to external shocks and uncertainty. however, many other workers experience irregular incomes due to the rise of the gig economy. proponents of financial literacy education should persevere, recognizing the limits to transferring knowledge and set evidence-based goals for financial literacy education. 5. discussion and conclusion the financial wellbeing of students is of concern to universities both to facilitate learning and to prepare them for future financial decision-making as participants in the economy. as the increase in outstanding student debt and student loan defaults has raised concerns regarding the value of higher education outcomes and the consequences of over-indebtedness for young borrowers (artavanisa & karra, 2020; mueller & yannelis, 2019), it could further be argued that universities have a moral obligation to support students in managing their finances. financial literacy education seems an appropriate solution. however, if financial literacy is well proxied by the big three questions, then knowledge alone will not achieve the intended outcome. after all, understanding the concept of compound interest is of little use if living hand-to-mouth or if economic choice is confined due to household arrangement. therefore, public policy makers and educators that advocate for financial literacy interventions should add strategies for improving incomes to their arsenal. the playing field is not level, government safety nets are often inadequate, and non-participation in education and work can be intergenerational. however, to see more informed financial decision-making and to improve wealth and wellbeing outcomes would benefit many, and these factors are arguably more important. the findings of this study, while currently limited to one australian university, provide valuable insights into the financial wellbeing of university students and can be used to inform future actions to specifically improve the wellbeing of university students. the impact of income on student wellbeing and student welfare is significant. not only does an increase in income contribute to student wellbeing financially, but it may also assist in improving student grades by minimizing stress for those at the lower end of the income scale. while financial literacy education has traditionally been the response to improving university students’ financial wellbeing, this study has shown that for students on lower incomes, this this alone will not improve the financial wellbeing of students. a call to action t. west et al. / financial services review 29 (2021) 187–207 203 for more novel approaches to financial wellbeing from both universities and the government that address the needs of students on significantly lower levels of income is needed. this might include universities providing cheaper accommodation and meals for lower income students through subsidies or vouchers or providing more jobs to students on campus. other initiatives such as collaborating with industry to provide more cadetships, scholarships, and paid internships to students may also be beneficial. government welfare policies also need to reconsider the way that income is distributed so that university students are not penalized for furthering their education. for example, in australia, as it currently stands, a young university student’s access to welfare is directly linked to parental income, regardless of whether they remain in the family home or need to relocate to attend university, or in fact whether the parent or parents are even providing financial support to the student. in comparison, a young individual who has completed school but does not attend university is entitled to higher levels of welfare regardless of their parental income or whether they are living at home or not. innovative reform might include the provision of additional financial assistance to low income university students that is offered in conjunction with financial literacy education, or maybe tax incentives could be introduced for parents and spouses who provide financial support to dependents as they attend university. further research is needed to investigate the feasibility of such suggestions and extensions of this study at other institutions, along with qualitative studies would assist in realizing the true extent of the impact of low incomes on financial wellbeing of university students. acknowledgements this project was supported by a griffith business school new researcher grant. references anz. (2018). financial wellbeing a survey of adults in australia. available at https://www.anz.com/resources/ 2/f/2f348500-38a2-4cfe-8411-060cb753573d/financial-wellbeing-aus18.pdf?mod=ajperes artavanis, n., & karra, s. (2020). financial literacy and student debt. the european journal of finance, 26, 382–401. asic. (2007, february 16). asic’s national financial literacy strategy: a speech by greg medcraft, chairman, financial capability breakfast briefing. parliament house, canberra, australia. available at https:// download.asic.gov.au/media/4153965/greg-medcraft-speech-financial-capability-breakfast-16-february-2017. pdf berman, j. z., tran, a. t. k., lynch, j. g., & zauberman, g. (2016). expense neglect in forecasting personal finances. journal of marketing research, 53, 535–550. bester, h., chamberlain, d., de koker, l., hougaard, c., short, r., smith, a., & walker, r. (2008). implementing fatf standards in developing countries and financial inclusion: findings and guidelines. first initiative (world bank). avaialble at http://www.finmark.org.za/wp-content/uploads/2016/01/rep_ implementingfatfstandards_2008.pdf bongini, p., & cucinelli, d. (2019). university students and retirement planning: never too early. international journal of bank marketing, 37, 775–797. cameron, a. c., & trivedi, p. k. (2009). microeconometrics: methods and applications. melbourne: cambridge university press. 204 t. west et al. / financial services review 29 (2021) 187–207 cfpb. (2015). financial well-being in america, consumer financial protection bureau. available at https:// www.consumerfinance.gov/data-research/research-reports/financial-well-being-america/ chen, h., & volpe, r. p. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. collins, j. m., & urban, c. (2020). measuring financial wellbeing over the lifecourse. the european journal of finance, 26, 341–359. cull, m., & whitton, d. (2011). university students’ financial literacy levels: obstacles and aids. the economic and labour relations review, 22, 99–114. dean, k. l., & forray, j. m. (2018). buying or renting? how students are ’acquiring’ knowledge about management. journal of management education, 42, 163–167. dutta, s., singh, h., & sultana, n. (2019, february 14). how zip pay works, and why the extra cost of ‘buy now, pay later’ is still enticing. the conversation. available at https://theconversation.com/how-zip-payworks-and-why-the-extra-cost-of-buy-now-pay-later-is-still-enticing-110429 farrell, d., & greig, f. (2016). paychecks, paydays, and the online platform economy: big data on income volatility. in proceedings annual conference on taxation and minutes of the annual meeting of the national tax association (pp. 1–40). national tax association. available at https://www.jstor.org/stable/e26816497 fear, j., & o’brien, j. (2009). where does the buck stop? community attitudes to over-lending and over-spending. australasian accounting, business and finance journal, 3, 5–12. available at https://ro.uow.edu.au/aabfj/vol3/iss1/2/ fernandes, d., lynch, j. g. , jr., , & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors.management science, 60, 1861–1883. finke, m. s., howe, j. s., & huston, s. j. (2017). old age and the decline in financial literacy. management science, 63, 213–230. georgarakos, d., haliassos, m., & pasini, c. (2014). household debt and social interactions. review of financial studies, 27, 1404–1433. https://doi.org/10.1093/rfs/hhu014 ha, h. (2013). credit card use and debt by female students: a case study in melbourne. australia, youth studies australia, 32, 57–71. available at https://search.informit.com.au/fulltext;dn=769059811260161;res=ielhss hancock, a. m., jorgensen, b. l., & swanson, m. s. (2013). college students and credit card use: the role of parents, work experience, financial knowledge and credit card attitudes. journal of family and economic issues, 34, 369–381. harrison, n., & agnew, s. (2016). individual and social influences on students’ attitudes to debt: a crossnational path analysis using data from england and new zealand. higher education quarterly, 70, 332–332. hastings, j., & tejeda-ashton, l. (2008). financial literacy, information and demand elasticity: survey and experimental evidence from mexico. nber working paper 14538. cambridge, ma: national bureau of economic research. henager, r., & wilmarth, m. j. (2018). the relationship between student loan debt and financial wellness. family and consumer sciences research journal, 46, 381–395. hilgert, m., hogarth, j., & beverly, s. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, july, 309–322. johnson, w., & krueger, r. f. (2006). how money buys happiness: genetic and environmental processes linking finances and life satisfaction. journal of personality and social psychology, 90, 680–691. jorgensen, b. l. (2007). financial literacy of college students: parental and peer influences. thesis for master of science in human development, virginia polytechnic institute and state university. available at http:// citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.536.9199&rep=rep1&type=pdf kaine, s., oliver, d., & josserand, e. (2017). precarity in the gig-economy: the experience of rideshare drivers. in proceedings. association of industrial relations academics in australia and new zealand conference. canberra, australia, february 8–10. lusardi, a., & mitchell, o. (2006). financial literacy and planning: implications for retirement wellbeing. working paper, pension research council. wharton school: university of pennsylvania. lusardi, a., & mitchell, o. (2007). baby boomer retirement security: the role of planning, financial literacy and housing wealth. journal of monetary economics, 54, 205–224. t. west et al. / financial services review 29 (2021) 187–207 205 lusardi, a., & mitchell, o. (2009). how ordinary people make complex economics decisions: financial literacy and retirement readiness. nber working paper 15350. national bureau of economic research. available at http://www.nber.org/papers/w15350 lusardi, a., & mitchell, o. s. (2011). financial literacy around the world: an overview. journal of pension economics & finance, 10, 497–508. lusardi, a., mitchell, o., & curto, v. (2010). financial literacy among the young. journal of consumer affairs, 44, 358–380. lusardi, a., & tufano, p. (2009). debt literacy, financial experiences and overindebtedness. nber working paper 14808. national bureau of economic research. available at https://www.nber.org/papers/w14808 luthans, k. w., luthans, b. c., & chaffin, d. (2019). refining grit in academic performance: the mediational role of psychological capital. journal of management education, 43, 35–61. lyon, k., & matson, t. (2019). how university leaders can empower the ’new majority’ student. gallup education, august 9. available at https://www.gallup.com/education/264299/university-leaders-empowernew-majority-student.aspx mae, s. (2009). how undergraduate students use credit cards: sallie mae’s national study of usage rates and trends. available at https://news.salliemae.com/sites/salliemae.newshq.businesswire.com/files/doc_ library/file/salliemae_majoringinmoney_2016.pdf mentzer, j., klontz, t., klontz, b., & britt, s. l. (2011). money beliefs and financial behaviors: development of the klontz money script inventory. journal of financial therapy, 2(1),1. mueller, h. m., & yannelis, c. (2019). the rise in student loan defaults. journal of financial economics, 131, 1–19. muir, k., hamilton, m., noone, j., marjolin, a., salignac, f., & saunders, p. (2017). exploring financial wellbeing in the australian context. available at http://www.financialliteracy.gov.au/media/560752/ research-unsw-fla-exploringfinancialwellbeingintheaustraliancontextreport-201709.pdf netemeyer, r. g., warmath, d., fernandes, d., & lynch, j. g. (2018). how am i doing? perceived financial well-being, its potential antecedents, and its relation to overall well-being. journal of consumer research, 45, 68–89. norvilitis, j. m., szablicki, p. b., & wilson, s. d. (2003). factors influencing levels of credit-card debt in college students. journal of applied social psychology, 33, 935–947. oecd. (2012, august). oecd international network on financial education high-level principles on national strategies for financial education. available at https://www.oecd.org/daf/fin/financial-education/oecdinfe-principles-national-strategies-financial-education.pdf oecd. (2020). oecd/infe 2020 international survey of adult financial literacy. international network on financial education (infe). available at www.oecd.org/financial/education/launchoftheoecdinfeglobal financialliteracysurveyreport.htm peach, n. d., & yuan, h. (2017). assessing the financial knowledge, behaviours, and attitudes of undergraduates. e-journal of business education and scholarship teaching, 11, 27–38. available at https://eric.ed.gov/?id= ej1167326 philippas, n. d., & avdoulas, c. (2020). financial literacy and financial well-being among generation-z university students: evidence from greece. the european journal of finance, 26, 360–381. richins, m. l. (2004). the material values scale: measurement properties and development of a short form. journal of consumer research, 31, 209–219. senack, e. (2015). open textbooks: the billion dollar solution, report from the student public interest research groups. available at https://studentpirgs.org/sites/student/files/reports/the%20billion%20dollar% 20solution.pdf spencer, n., nieboer, j., & elliott, a. (2015).wired for imprudence. london: rsa. stango, v., & zinman, j. (2007). fuzzy math and red ink: when the opportunity cost of consumption is not what it seems.mimeo: dartmouth college. stewart, a., & stanford, j. (2017). regulating work in the gig economy: what are the options? the economic and labour relations review, 28, 420–437. 206 t. west et al. / financial services review 29 (2021) 187–207 summerville, a., & roese, n. j. (2008). dare to compare: fact-based versus simulation-based comparison in daily life. journal of experimental social psychology, 44, 664–671. tharp, d. t., seay, m. c., carswell, a. t., & macdonald, m. (2020). big five personality traits, dispositional affect, and financial satisfaction among older adults. personality and individual differences, 166, 110211– 110213. timmerman, i., & volkov, n. (2019). career and education choice as central elements of long-term financial planning. paper presented at academy of financial services 2019 annual meeting, minneapolis, mn, united states, october 2019. van rooij, m., lusardi, a., & alessie, r. (2007). financial literacy and stock market participation. nber working paper 13565. national bureau of economic research. available at http://www.nber.org/papers/ w13565 vaitilingam, r. (2016). improving people’s financial decision making. voxeu centre for economic policy research. available at https://voxeu.org/article/improving-people-s-financial-decision-making west, t. (2020). women pay their way on income contingent student debt. policy futures in education, 18, 937–956. west, t., & worthington, a. c. (2014). macroeconomic conditions and australian financial risk attitudes, 2001– 2010. journal of family and economic issues, 35, 263–277. west, t., & worthington, a. c. (2018). synthesizing an index of financial literacy using not-for-purpose survey data. economic notes, 47, 331–352. wilkins, r. (2018). financial literacy and attitudes to finances. the household, income and labour dynamics in australia survey: selected findings from waves 1 to 16 (pp. 117–126). victoria, australia: melbourne institute: applied economic & social research, university of melbourne. available at https:// melbourneinstitute.unimelb.edu.au/__data/assets/pdf_file/0005/2839919/2018-hilda-sr-for-web.pdf williams, a., & oumlil, b. (2015). college student financial capability. international journal of bank marketing, 33, 637–653. worthy, s. l., jonkman, j., & blinn-pike, l. (2010). sensation-seeking, risk-taking, and problematic financial behaviours of college students. journal of family and economic issues, 31, 161–170. yu, s., alper, h. e., nguyen, a., brackbill, r. m., turner, l., walker, d. j., maslow, c. b., & zweig, k. c. (2017). the effectiveness of a monetary incentive on survey offer response rates and response completeness in a longitudinal study. bmc medical research methodology, 17, 77–86. t. west et al. / financial services review 29 (2021) 187–207 207 pii: s1057-0810(96)90008-6 150 financial services review 5(2) 1996 finally, the site provides links to the ~i~n~~~~ analysts j~u~qz, the cfa digest, and the lsfa digest, all of which are published by the aimr. the new york stock exchange (nyse), the chicago board options exchange (cboe), and the financial management association websites reviewed by: thomas eyssell, associate professor, university of missouri-st. louis one of the biggest potential benefits of the worldwide web to financial users is the availability of up-to-date information on financial markets and instruments. two related websites of potential interest to fsr readers are those devoted to the new york stock exchange (nyse) and the chicago board options exchange (cboe). the cboe website is extremely user-~endly, and seems geared more toward the individual investor. upon reaching the site (http://www.cboe.com), one finds links to pages describing “what’s new,” cboe “products,” the “options institute” (the cboe’s education arm), as well as a “virtual visit” to the cboe. the latter consists of extensively captioned graphics of the trading floor, as well as a “virtual stroll” around the su~ounding chicago area. those contemplating the inclusion of options in a diversified portfolio, and those teaching basic options mechanics will find the “understanding options” link useful. included at this link are online versions of two cboe publications: “characteristics and risks of standardized options” and the options clearing corporation’s options prospec tus. these publications provide a great deal of nuts-and-bolts information sometimes glossed over in standard texts. finally, for those seeking to perform empirical research and analysis, the cboe page provides links to several sources of security and market data. the nyse website (http://www.nyse.com) is highlighted by an electronic version of the organization’s 1995 annual report. included at this site are several items of interest to the financial planner, teacher, or researcher. for example, one can go to an alphabetized list of nyse firms and, upon clicking on a ticker symbol, be immediately transferred to that firm’s homepage. (be aware, however, that not all nyse-listed firms appear.) the nyse website is a bit less user-friendly than that of the cboe. for example, clicking on “visit” takes one to a single image of the trading floor. in addition the educa tion/informational aspects are not nearly as extensive or compelling. on the other hand, it is possible to download, at no cost, daily closing price and volume data spanning several decades. in addition, one can obtain reasonably up-to-date lists of firm listings and dele tions, trading habits, discipline actions, and so forth. in any case, these two market sites are well worth the trip. the financial management association has put together an extensive website (http:// www.fma.org) which will be of interest to finance academics as well as to finance practi tioners. among other things, the interested websurfer can find out about upcoming confer ences, events, and publication dates through 1997, services for students, finance links, and, of course, placement services for employers and job seekers. one can search the “positions available” section by finance category (business finance, financial institutions and markets, general finance, insurance, investments, real estate, and other areas of finance), or alphabetically by school name. apparently, the list of available positions is updated periodically, since many of the listings sport a bright red “new!” tag. rook, software and web site reviews 151 the electronic “resume book,” on the other hand, is divided (at the time of this writing) into four categories--corporate finance, financial institutions, investments, and other areas of finance. as in the “positions available,” several listings are tagged as “new!,” indicat ing some updating of the list. those seeking positions are provided space for a relatively abbreviated vita in a standard format, which includes published papers, awards, presenta tions, teaching specialties, phone numbers, e-mail addresses, and so on. approximately 200 resumes appear, and a (somewhat ad hoc) comparison of the vitae of junior to more experienced job-seekers suggests that all receive about the same amount of space. finally don’t forget to look over the fma’s “guide to finance on the web,” one of the most comprehensive lists of links to professional associations, publications, resources, and other finance-related information anywhere. pii: s1057-0810(97)90035-4 72 financial services review 6(1) 1997 makes a pitch for its services throughout this link, but users can scan recommended per sonal budgets, balance sheets, and financial strategies easily and anonymously. the "per sonal finance center" link helps users understand financial issues facing most american individuals and families and how certain financial planning techniques can be used to meet financial objectives. users interested in finding objective information on financial plan ning topics like savings, college expenses, credit usage, insurance, mortgages, taxes, retire ment, mutual funds, and making a financial plan will love this link. family and consumer economists will also find a useful discussion and analysis of the individual financial life cycle presented here. a "business planning" link is available for those interested in learning more about maximizing business cash flows, providing competitive employee benefits, obtaining financing, and institutional investing. another feature that is unique to this web site is a key work search that allows users to query the merrill lynch database for interesting facts, fig ures, and research. researchers, educators, and students will fred this feature an invaluable source for data. compared to other brokerage sponsored web sites, merrill lynch's web page lacks graphical excitement (you can, however, download audio transcripts of each linked page), but for financial professionals, investors, researchers, educators, and students in need of solid information in a timely manner, this site fills the bill perfectly. another useful feature to note is that merrill lynch provides this service without obligation, cost, or a lot of self promotion, and in today's web world that's saying a lot. charter media's briefing.com web site reviewed by: robert l. albert jr., assistant professor of finance, morehead state university individual investors can now find a multitude of web sites which provide a wide array of relevant and timely financial market information. many sites are devoted to security prices and many others are devoted to fundamental characteristics of the firms. charter media's briefing.corn (http://www.brief'mg.com) is one site which provides security price information and fundamentals along with a real time commentary on the financial markets. briefing.corn is not a site burdened with many slow-loading graphics, so users can get to different levels of information quickly. briefing.corn provides comprehensive reports of economic, industry, and company news giving investors access to the relevant information necessary for a top-down approach to investing. charter media's research staff is comprised of former senior man agers and analysts from standard and poor's mms international who provide insightful interpretations of braking economic and industry developments. at the macroeconomic level, briefing.corn provides updates of the major economic variables along with forecasts of future economic activity. in their political brief, they provide a daily commentary on those political issues which are likely to have an impact on the financial markets. in addition, briefing.corn provides updated fed briefs which given an overview of fed policy and possible interest rate moves. briefing.corn's industry reviews provide a timely analysis of the major factors affect ing specific industries and highlight timely stocks within those industries. in alliance with book, software, and web site reviews 73 quote.com (http://www.quote.com), they also provide sector ratings and current price quotes within sectors. the site offers intraday updates on newsworthy stocks and analyses of both the debt and equity markets. also through quote.com, briefing.com provides quotes on individual securities (and on several market indices) which include pricing information along with fundamentals such as p/e ratios dividend yields and the 52-week price ranges. these 15 minutes delayed quotes are updated constantly throughout the trading day. members can also create a portfolio of up to 25 stocks in which they can monitor daffy updates of prices and volume. in addition, the portfolio feature will track gains and losses on all individual securities and the entire portfolio based on inputted purchase price. both intraday antl his torical price charts are also available. stock briefs on over 8,300 companies are also available. these briefs include a short business summary, key financial ratios, share-related information, short interest informa tion, institutional and insider ownership, historical growth rates in sales, eps and divi dends, and historical quarterly figures for revenue and eps. these briefs are updated with current data each monday. the monthly subscription fee for briefing.corn is $6.95 and includes access to quote.com. individual investors who trade through e-trade (the electronic brokerage f'mn) have free access to briefing.com. the money book of personal finance richard eisenberg and the editors of money magazine new york, ny: warner books inc.; 1996 (isbn 0-446-51981-2) reviewed by: douglas r. kahl, professor of finance, university of akron the money book of personal finance is a popular personal finance guide available in almost any bookstore not a personal finance textbook. however, i found that it works very well as the primary textbook in a personal finance course for nonbusiness majors. the book was a collaborative effort by at least thirteen writers and editors at money magazine. the clear, concise and authoritative writing style does not assume any prior background in finance or business. since it was not designed to be a course textbook, it lacks the usual classroom sup port package. there is no study guide, no instructor's manual, no chapter end problems and no package of overheads. for this reason, i would not recommend the money book of personal finance as a primary text when the instructor is teaching the course for the first time. for the instructor who teaches a personal finance service course for nonbusiness majors on a regular basis, this text could be an interesting and refreshing change. the book was very popular with my students who were about equally split between traditional and nontraditional backgrounds. both the traditional and nontradi tional students gave considerable credence to the book because of the perceived practi cal knowledge and applied expertise of the authors. the students also noted, with considerable approval, that the price was about one-fourth that of a standard text. they found the lower cost especially appropriate for a personal finance course. the book covers the usual personal finance topics in four major sections: getting started, reaching your financial goals, investing your money and your family pii: s1057-0810(98)90003-8 volume 7 number 1 1998 financial services review the journal of individual financial management editor karen eilers lahey 1998 james s. ang florida state university lawrence j. gitman san diego state university douglas r. kahl university of akron james e. larsen wright state university dixie l. mills illinois state university tony plath university of north carolina at charlotte a. charlene sullivan purdue university jill lynn vihtelic saint mary's college associate editors 1998-1999 vickie bajtelsmit colorado state university waldo l. born eastern illinois university larry a. cox university of mississippi jean louis heck villanova universi~ travis pritchett general reinsurance william reichenstein baylor university walter woerheide rochester institute of technology 1998-2000 raj aggarwal john carroll university james r. barth auburn university barry diskin florida state university thomas h. eyssell university of missouri, st. louis sandra g. gustavson university of georgia andrea hueson university of miami jeff madura florida atlantic university terry l. zivney ball state university stamford, connecticut @ jai press inc. london, england pii: s1057-0810(97)90014-7 r financial services review, 6(3): 201-219 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. asset allocation and investment horizon keith v. smith quarterly recommendations by national brokeragefirms since the third quarter of 1989 provide an opportunity to compare different approaches to asset allocation. to follow a brokerage firm’s recommendation every quarter is to practice tactical asset alloca tion. both the length of the investor’s decision horizon and brokerage commissions that are incurred when portfolio changes are made impact investment performance, and both contribute to the risk experienced by investors. buy-and-hold and strategic asset allocation would have served investors better than tactical asset allocation during the first half of the 1990s. i. introduction in a recent article in this review, walker and hatfield (1996) investigated the question of whether the published advice of professional analysts on individual securities documented in the wall street journal can be used advantageously by individual investors. the authors concluded that it may be possible for professionals to identify attractive securities, but that ability may be offset by the transaction costs necessary to implement their recommenda tions. the authors also reminded readers that individual security selection decisions logi cally follow asset allocation decisions. every quarter on an ongoing basis, the wall street journal also reports the recom mended asset allocations (i.e., stocks, bonds, and cash) of the large national brokerage firms.’ the recommendations tend to change over time, and so investment strategists at those firms must believe that they can successfully predict when stocks, bonds, and cash are likely to do well as asset categories. if investors follow the brokerage firms and change their asset allocations each quarter, they are engaging in market timing, an activity whose value has been seriously questioned by some researchers. in this paper, “investment horizon” is defined as the time span between investment decisions, and with the idea that investors may decide not to change their portfolio blend every quarter. if an investor changes the portfolio blend just two times per year, then he/ she behaves as if the investment horizon is six months. a change of blend every calendar quarter would reflect an investment horizon of three months. so even though the brokerage keith v. smith l professor of management, krannert graduate school of management, purdue university, west lafayette, in 47907; e-mail: kvsmith@mgmt.purdue.edu. 202 financial services review 6(3) 1998 firms are making asset allocation recommendations at least every quarter, investors may choose to adopt longer investment horizons and change their portfolio blends less often. the quarterly journal recommendations provide a window of opportunity for examin ing the impact of investment horizon on asset allocation decisions. after reviewing the rel evant literature on asset allocation, we carefully investigate how investors would have fared had they followed the brokerage firm recommendations over time. we examine the impact of brokerage commissions on ensuing portfolio performance. we consider the risk that investors face when the follow the asset allocation recommendations of the national brokerage firms. we find that strategies of buy-and-hold and/or strategic asset allocation would have served investors better than a strategy of tactical asset allocation during the first half of the 1990s. ii. nature and importance of asset allocation asset allocation is a decision-making process in which the investment funds of an individ ual or a group of individuals are allocated to investment categories rather than to individual assets. studies by brimson, et.al. (1986,199o) have shown convincingly that allocation of investment funds to asset categories is far more important than the selection of individual securities within each asset category. the simplest breakdown for asset allocation is into just two categories: fixed income (bonds) and equity (common stocks) securities. broader schemes of asset allocation include categories such as largeand small-capitalization common stocks, government and corporate bonds, real estate, gold, commodities, international securities, and venture capi tal opportunities. but the most common scheme for asset allocation is into three catego ries--stocks, bonds, and cash. that is what the brokerage firms tend to do, and that scheme is used in this study. a number of prior studies of asset allocation are germane to what we do here. a decade ago, sharpe (1986) suggested a useful taxonomy for asset allocation that included strategic, tactical, and insured approaches. the significant difference between strategic and tactical approaches to asset allocation is that a strategic allocution calls for an investor to hold constant a recommended blend over time, while a tucticd allocation periodically reassesses the portfolio blend and makes appropriate adjustments. in that sense, strategic asset allocation employs a longer investment horizon than does tactical asset allocation, which really is more of a market timing approach. an even less active approach is buy-and-hold, where no portfolio adjustments are made once an initial portfolio of asset categories is purchased. asset allocation decisions can be made in a variety of ways. four specific strategies were compared by perold & sharpe (1988). smith (1974) suggested a weighted-average measure of suitability for making asset allocations, while tarrazo (1997) provided an alter native approach using fuzzy-set theory. waring (1994) explained how mean-variance opti mization can be used to structure portfolios for 401(k) retirement plans. black & litterman showed how an optimizing approach can be applied to international bond portfolios, but with outlooks for interest rates and currencies being compared to expected returns from an asset pricing model. asset alhxztion and investment horizon 203 another approach in asset allocation studies is to examine investor portfolios to try and infer investment attitudes about risk. blume & friend (1975) examined federal reserve board data and concluded that individuals seem to maintain their percentage mix of riskier and safer investments and thus exhibit constant relative risk aversion. in contrast, cohn, et.al. (1975) used cross-sectional brokerage firm data to infer decreasing relative risk aversion, which is a tendency to put a larger percentage of wealth into riskier invest ments as wealth increases. there also are different findings from empirical investigations of just how well vari ous asset allocation strategies would have worked in practice. earlier studies such as sharpe (1975), henriksson (1984), and jeffrey (1994) provide expost evidence that market timing and tactical asset allocation do not add value. the mood of the popular press [e.g. clements (1995)] also is not very encouraging toward strategies of market timing and tac tical asset allocation. some hope was provided by phillips, et.al. (1996). the authors exam ined the performance of eleven managers who use tactical asset allocation to manage institutional portfolios. using performance data net of management fees, they found that the managers outperformed appropriate benchmarks during the 1977-87 period, but they did not continue to do that during the 1988-94 period. finally, it is appropriate to mention two recent studies that examined asset allocation with an ex anfe perspective. smith (1997) looked at the brokerage firm asset allocations in the sense of how close the recommended portfolios were to efficient portfolios in a mean-variance space. he found that no single firm is dominant at all, and the difficulty is that there is no agreement as to the necessary portfolio inputs that are necessary for the optimization. bierman (1997) also used an ex ante analysis to revisit the question of what happens to investment risk when the investment horizon gets longer. he concluded that lengthening the investment horizon does increase risk at least by some popular measures. iii. recommended blends by national brokerage firms since the middle of 1989, the wall street journal has published on a quarterly basis the recommended asset allocations of leading u.s. brokerage firms. although sixteen firms have participated, cs first boston and edward d. jones joined the group in 1994, and evelen first appeared in the second half of 1995. among the other thirteen firms, all have participated for at least two years, and eight firms have been involved since inception, the recommended blends of those thirteen firms, as reported on a quarterly basis in the journal, constitute the data for this investigation. it should be noted that while some of the broker age firms may change their recommendations more frequently, this study utilizes just the quarter-end blends of the brokerage firms. when on occasion a brokerage firm recom mended other categories such as real estate, gold, commodities, and foreign stocks, those assets were combined with common stocks into a single equity category. table 1 includes the recommended percentage allocations for common stocks by the thirteen national brokerage firms at the end of each quarter, from the third quarter of 1989 (89-3) through the end of calendar year 1995 (95-4). for each brokerage firm, for each quarter, and for the entire horizon, the maximum, minimum, range (maximum minus min imum), average, and standard deviation of the common stock recommendations are shown. the maximum common stock recommendation was 85%, the minimum value was 24%, t a b l e 1 c om m on st oc k r ec om m en da tio ns by b ro ke ra ge fi rm s q ua rt er e nd , 19 89 -1 99 5 (in p er ce nt ag e) b ro ke ra ge fi rm 89 -3 89 -4 90 -l 90 -z 90 -s 90 -4 91 -i 91 -z 91 -3 91 -4 92 -1 92 -2 92 -3 92 -4 93 -l 93 -2 a . g . e dw ar ds d ea n w itt er g ol dm an sa ch s k em pe r k id de r pe ab od y l eh m an b ro th er s m er ri ll l yn ch pa in e w eb be r pr ud en tia l r ay m on d ja m es sa lo m on b ro th er s sh ea rs on sm it‘ h b ar ne y 45 45 40 40 40 45 55 50 55 55 55 50 45 85 85 85 85 50 45 55 55 45 65 55 55 60 65 50 35 3. 5 35 50 60 60 75 75 75 75 75 50 45 45 50 60 60 70 70 70 70 70 70 70 50 45 60 60 75 60 50 50 70 70 50 50 40 50 45 55 55 55 60 65 60 55 60 24 25 56 56 56 59 57 55 64 67 57 64 13 70 35 40 40 40 65 80 60 75 65 65 45 45 60 60 75 65 65 65 70 70 70 70 60 55 45 60 60 65 65 65 65 65 60 60 60 55 55 55 50 50 45 50 45 50 50 55 50 60 55 50 50 60 55 71 69 55 70 55 55 50 50 50 50 50 50 55 60 60 50 70 75 55 73 70 60 45 50 m ax im um 85 85 85 85 65 65 80 io 75 15 75 75 75 75 70 75 m in im um 24 25 35 35 35 45 50 50 45 55 55 45 45 50 45 45 r an ge 61 60 50 50 30 20 30 20 30 20 20 30 30 25 25 30 a ve ra ge 55 .9 50 .5 52 .6 53 .6 50 .1 55 .9 61 .7 59 .0 62 .4 65 .2 60 .7 57 .4 57 .8 58 .0 57 .0 60 .3 st d d ev 15 .5 15 .2 16 .2 14 .2 10 .3 7. 6 8. 8 6. 2 9. 8 5. 5 6. 7 8. 9 11 .2 8. 9 8. 4 9. 4 % (t ab le 1 co nt in ue d) % b ro ke ra ge f ir m 93 -3 93 -4 94 -1 94 -2 94 -3 94 -4 95 -l 95 -2 95 -3 95 -4 m ax im um m in im um r an ge a ve ra ge sr d d ev s 5 a . g . e dw ar ds 65 65 60 50 45 40 40 50 45 55 65 d ea n w itt er 60 60 50 50 50 50 60 60 60 60 85 g ol dm an sa ch s 70 70 80 85 85 65 70 65 65 65 85 k em pe r 50 50 55 55 40 65 55 60 65 k id de r pe ab od y 70 70 60 50 66 70 l eh m an b ro th er s 75 50 50 50 45 45 40 55 65 70 75 m er ri ll l yn ch 60 60 60 50 50 50 55 55 50 50 65 pa in e w eb be r 77 68 58 52 49 48 52 58 58 64 77 pr ud en tia l 80 85 65 60 65 55 60 55 50 50 85 r ay m on d ja m es 60 70 65 65 60 55 65 65 65 70 75 sa lo m on b ro th er s 45 45 50 50 45 45 45 45 45 50 50 sh ea rs on 65 sm ith b ar ne y 55 55 50 50 50 50 50 55 60 60 60 40 25 49 .4 7. 2 45 40 60 .2 11 .7 35 50 64 .6 13 .8 6 40 25 52 .7 6. 2 3 45 25 63 .1 9. 2 8’ 40 35 56 .4 12 .1 3 40 25 54 .2 5. 5 24 53 58 .1 12 .3 35 50 59 .4 13 .3 45 30 63 .1 6. 5 45 5 46 .9 2. 4 50 15 59 .3 5. 1 45 15 51 .7 3. 9 m ax im um 80 85 80 85 85 65 70 65 65 70 85 m in im um 45 45 50 50 40 40 40 45 45 50 24 r an ge 35 40 30 35 45 25 30 20 20 20 61 a ve ra ge 63 .9 62 .3 58 .6 55 .6 54 .2 51 .6 53 .8 56 .6 56 .3 59 .4 57 .4 st d d ev 10 .4 10 .8 8. 4 10 .0 12 .2 7. 5 9. 3 5. 7 7. 7 7. 5 10 .8 n ot e: “8 93” in di ca te s th e en d of t he th ir d qu ar te r of 1 98 9. .% u u ?: w a ll st re et jo ur na l, q u a rt e rly is su e s. 206 financial services review 6(3) 1998 and the range was 61%. for the sample of thirteen brokerage firms, the average common stock recommendation was 57.4%, while the standard deviation was 10.8%. overall, there was considerable diversity in common stock recommendations, and that is one component of risk experienced by investors who follow the common stock suggestions of the national brokerage firms. brokerage firms also differed considerably in how their common stock recommenda tions changed from quarter to quarter. most brokerage firms provided recommendations that were multiples of 5%. an exception was paine webber whose recommendations changed by increments and fractions, suggesting that the firm was using an analytical model that generated more precise percentages. the range of common stock allocations was highest for paine webber (53%) and lowest for salomon brothers (5%). while salomon brothers’ asset recommendations did not begin until the end of 1992, they have been either 45% or 50% for common stocks in every subsequent quarter. over time, the average common stock recommendation was highest (65.2%) at the end of december 1991; it was lowest (50.1%) at the end of september 1990. table 2 provides further detail by including recommendations for bonds and cash, along with that for common stocks. the maximum, minimum, average, and standard devi ations of recommendations by brokerage firms for stocks, bonds, and cash during the 1989-95 period are reported. in addition to the 57.4% average for common stocks already reported, we see that the average recommendations were 30.6% for bonds and 12.0% for cash. the only firm recommending a zero percentage for bonds was prudential (on three occasions), while a total of eight firms recommended a zero percentage for cash at least once. ranges are not included in table 2, but further inspection of the quarterly data reveals that the range both for common stocks and cash was 61%, followed closely by a 55% range for bonds. average recommendations for each quarter also are not included in table 2, but it can be reported that the highest value for bonds was 37.1% at the end of 1990, while the lowest was 23.9% at the end of third quarter 1993. for cash, the highest average was 20.7% at the end of 1989, while the lowest was 2.6% just two years later. overall, there was con siderable diversity among the brokerage firms in their suggested allocations among the three asset categories. while some brokerage firms occasionally did not change their recommended blends from one quarter to the next, the historical pattern does suggest a strategy of tactical asset allocation by the national brokerage firms. part of the risk experienced by investors who follow the advice of the brokerage firms may be that considerable changes in their portfo lios must be made in order to remain in step with the tactical asset allocation recommenda tions. it remains to be seen here if such changes in asset allocations add value for investors. iv. perf’ormance benchmarks in order to investigate the impact of asset allocations on investor performance, the standard & poor’s 500 composite index (s & p 500) was used a measure of the level of common stocks. the level of bonds was portrayed by an index of long-term u.s. government issues (u.s. long), while the level of cash was portrayed by an index of 90-day u.s. treasury bills (u.s. bill).* t a b l e 2 s a a ss et a llo ca tio n r ec om m en da tio ns by b ro ke ra ge fi rm s, 19 89 -1 99 5 3 (in p er ce nt ag e) 1 c om m on st oc ks b on ds c as h b ro ke ra ge fi rm m ax im um m in im um a ve ra ge st d d ev m ax im um m in im um a ve ra ge st d d ev m ax im um m in im um a ve ra ge st d d ev g a . g . e dw ar ds 65 40 49 .4 1. 2 45 25 38 .7 4. 9 20 5 11 .9 4. 8 “0 ’ 5 d ea n w itt er 85 45 60 .2 1. 6 50 15 32 .1 9. 6 25 0 7. 7 1. 5 a g ol dm an sa ch s 85 35 65 .6 13 .8 55 15 28 .3 10 .6 30 0 7. 1 5. 9 k em pe r 65 40 52 .7 6. 2 30 20 24 .5 4. 5 30 10 22 .1 7. 2 k id de r pe ab od y 70 45 63 .1 9. 2 40 20 21 .6 5. 3 35 0 9. 2 13 .0 l eh m an b ro th er s 75 45 56 .4 12 .1 35 25 31 .4 3. 7 30 0 12 .3 9. 4 m er ri ll l yn ch 65 40 54 .2 5. 5 50 25 36 .5 7. 8 20 0 9. 2 5. 3 pa in e w eb be r 77 24 58 .1 12 .3 47 15 32 .8 8. 3 61 0 9. 1 15 .2 pr ud en tia l 85 35 59 .4 13 .3 55 0 27 .5 16 .0 40 0 13 .1 12 .8 r ay m on d ja m es 15 45 63 .1 6. 5 30 10 18 .8 5. 1 40 5 18 .1 8. 8 sa lo m on b ro th er s 50 45 46 .9 2. 4 35 30 31 .2 2. 1 25 15 21 .9 4. 2 sh ea rs on 65 50 59 .3 5. 1 40 20 32 .0 5. 1 20 5 8. 7 5. 3 sm ith b ar ne y 60 45 51 .7 3. 9 40 25 34 .0 4. 8 25 0 14 .2 5. 8 a gg re ga te 85 24 57 .4 10 .8 55 0 30 .6 9. 5 61 0 12 .0 9. 9 208 financial services review 6(3) 1998 a first performance benchmark in what follows is simply how each of those popular indexes did during a given period of time. in particular, we shall report the wealth relatives (non-annualized) for price appreciation for the given period. for example, during an ear lier time period, the s & p 500 stood at 349.15 at the end of the third quarter of 1989 (89-3), and it advanced to 451.67 by the end of the first quarter of 1993 (93-l). the resulting wealth relative benchmark for common stocks was 45 1.67 / 349.15 = 1.294. the corre sponding wealth relative for common stocks for a later time period (93-2 to 94-4) was 459.27 / 45 1.67 = 1.017, and for the entire time period, the benchmark wealth relative was 459.27 / 349.15 = 1.315. a second performance benchmark for each brokerage fnm is how an investor would have done if the recommended portfolio blends were not changed over time. it is essen tially a buy-and-hold strategy once an initial blend is implemented. buy-and-hold also minimizes transaction costs since no adjustments are made each quarter. the investment horizon of the investor is simply the total time period. two versions of buy-and-hold are calculated for each brokerage firm. one version uses the initial blend recommended by the firm at the beginning of the time period. because the achieved result depends only on the recommended blend at the beginning, we also examine a second version for buy-and-hold that uses the average blend for the brokerage firm over the total time period. investors would not know the average blend at the beginning of the time period, but it is included as a benchmark that better represents buy-and-hold for a given brokerage firm over time. a third performance benchmark is strategic asset allocation. a single blend is recom mended by each brokerage firm, but adjustments are made to rebalance/restore that partic ular blend at the end of each investment horizon, be it one quarter, six months, or a full year. as such, strategic asset allocation lies between the extremes of tactical asset alloca tion on the one hand, and buy-and-hold on the other. two versions of strategic asset allo cation are examined, and they parallel the two versions for buy-and-hold. one benchmark is strategic asset allocation based on the initial blend at the end of third quarter 1989, while the other benchmark is strategic asset allocation using the average blend recommended by each brokerage firm during the time period. v. empirical results the research questions are straightforward. how would an investor have fared had he/she followed the tactical asset allocation recommendations of each national brokerage firm on a quarterly basis? how does their achieved results compare with the performance bench marks discussed in the previous section? and what about risk? the research design was to examine for each brokerage firm its recommended asset allocations (for stocks, bonds, and cash) at successive quarter-ends, and to see how the resulting three-asset portfolio would have done over the next quarter, the next six months, and the next full year. because the brokerage firms tended to change their asset allocations each quarter, it did not seem fruitful to measure the achieved performance of a given port folio blend over longer time spans. such measurements were done for an earlier 14-quarter time period (89-3 through 93-l) during which ten brokerage firms made recommendations, a later 7-quarter period (93-l through 94-4) during which eleven firms were suggesting asset alhxtion and investment horizon 209 asset allocations, and the entire 21-quarter period (89-3 through 94-4) for which eight firms made recommendations each quarter. the results are presented in table 3. the values shown are wealth relatives for price appreciation based on the market indicators for common stocks, bonds, and cash, respec tively. the results are comparable across the sample of brokerage firms involved in each of the three time periods, as well as with the performance benchmarks that are included for each time period. again, the s & p 500 indicator is the benchmark for common stocks, the us long indicator is the benchmark for bonds, and the us bill indicator is the benchmark for cash. the “1 qtr” columns include results when the investor changed asset allocations every quarter; that is, when her/his investment horizon was three months, and thus coincided with that of the brokerage firm. the “2 qtr” columns show results that assume a six-month investment horizon. in other words, the investor acts on every other (i.e., second) quarterly asset allocation suggestion by the brokerage firm. similarly, the “4 qtr” columns in table 3 assume that the investor changed their portfolio blends only once per year, which is every fourth asset allocation suggestion by the brokerage firm. all three choices of investment horizon necessitate portfolio changes during a longer time period, and thus all three choices really are examples of tactical asset allocation. the largest wealth relative in each column is noted. for recommendations during the earlier time period (89-3 through 93-l), investors would have done better by following the recommended blends of kidder peabody and/or paine webber if their investment horizon was a single quarter. for investment horizons of two quarters, kidder peabody produced the highest wealth relative. and for investment horizons of four quarters, paine webber outperformed the other brokerage firms substantially--essentially a result of their 70% common stock recommendation at 89-3. for asset recommendations during the entire time period (89-3 through 94-4) inves tors should have followed goldman sachs if their investment horizon was either one or two quarters, but those investors would have done better following paine webber if their investment horizon was four quarters. for the later time period (93-l through 94-4), inves tors should have followed goldman sachs for investment horizons of one quarter or a full year. in contrast, they should have followed prudential if the investment horizon was six months in length. in summary, achieved performance by investors who followed the tactical asset allo cation recommendations of national brokerage firms depended on what time period they were in the security markets, which firm’s advice did they follow, but also the frequency of portfolio changes as determined by their choice of an investment horizon. no single bro kerage firm had the best results for all investment horizons and for all three time periods. table 3 also contains the comparable average wealth relatives for stocks, bonds, and cash during each time period. those performance benchmarks are for individual asset types, and thus no portfolio changes are necessary each quarter. because the brokerage firms recommended different blends of common stocks, bonds, and cash, the achieved blended results ended up being greater than the results just for bonds or cash, but less than the results if investors had decided to be in stock market completely. the choice of investment horizon thus had an impact on portfolio performance, but it was not the same in each time period. in the earlier time period, performance was on aver age slightly better for shorter investment horizons. for the later time period, performance was significantly lower for a six-month investment horizon. but for the entire time period, t a b l e 3 pr ic e a pp re ci at io n fo r r ec om m en de d a ss et a llo ca tio ns by n at io na l b ro ke ra ge fi rm s q ua rt er ly , 19 89 -1 99 5 e ar li er t im e p er io d e nt ir e t im e p er io d l at er p er io d b ro ke ra ge fi rm i q tr 2 q tr 4 q tr i q tr 2 q tr 4 q tr 1 q tr 2 q tr 4 q tr a . g . e dw ar ds d ea n w itt er g ol dm an sa ch s k em pe r k id de r pe ab od y l eh m an b ro th er s m er ri ll l yn ch pa in e w eb be r pr ud en tia l r ay m on d ja m es sa lo m on b ro th er s sh ea rs on sm ith b ar ne y 1. 23 6 1. 22 7 1. 21 8 1. 25 6 1. 27 8 1. 25 7 1. 20 1 1. 20 6 1. 20 9 1. 22 0 1. 23 5 1. 25 1 1. 26 7 1. 25 1 1. 21 0 1. 34 9* 1. 32 7* 1. 27 5 1. 01 3 1. 00 8 1. 00 0 1. 01 3 0. 86 3 1. 01 5 1. 06 7* 1. 00 3 1. 04 7* 1. 01 6 1. 02 2 1. 02 2 1. 28 3* 1. 29 9* 1. 29 0 1. 24 1 1. 23 9 1. 22 3 1. 26 5 1. 27 9 1. 25 7 1. 28 4* 1. 28 1 1. 33 9* 1. 30 8 1. 29 4 1. 41 0* 1. 27 4 1. 23 3 1. 19 3 1. 31 7 1. 29 9 1. 29 4 1. 24 8 1. 22 5 1. 23 5 1. 27 8 1. 29 6 1. 26 6 1. 25 9 1. 25 8 1. 27 0 1. 23 7 1. 23 8 1. 23 4 1. 25 4 1. 25 5 1. 25 1 1. 01 2 0. 85 7 1. 00 5 1. 00 4 0. 85 3 1. 01 7 1. 01 7 0. 86 1 1. 01 0 1. 01 3 0. 99 6 1. 02 0 1. 03 2 1. 03 0* 1. 03 0 1. 03 0 0. 87 8 1. 02 7 1. 00 0 0. 74 8 1. 00 0 a ve ra ge 1. 25 3 1. 24 6 1. 24 2 1. 28 2 s & p 5 00 1. 29 4 1. 29 4 1. 29 4 1. 31 5 u s l on g 1. 18 2 1. 18 2 1. 18 2 1. 03 8 u s b ill 1. 13 8 1. 13 8 1. 13 8 1. 02 1 n ot e: * h ig he st w ea lth r el at iv e (s ) fo r th at t im e pe ri od a n d in ve st m en t h o riz o n . 1. 28 3 1. 28 3 1. 02 0 0. 92 0 1. 01 8 1. 31 5 1. 31 5 1. 01 7 1. 01 7 1. 01 7 1. 03 8 1. 03 8 0. 87 9 0. 87 9 0. 87 9 1. 02 1 1. 02 1 0. 89 7 0. 89 7 0. 89 7 asset alhxtion and investment horizon 211 the average wealth relative were essentially the same for all three choices of investment horizon. that means that on average, it didn’t really matter how often investors followed the advice of the national brokerage firms. that is an interesting result, but as we shall now see, it is an incomplete picture of the impact of asset allocation recommendations on inves tor portfolio performance. vi. brokerage commissions the impact of investment horizon on asset allocation recommendations is not complete until transaction costs are considered. because all investors (both individual and institu tional) have different tax situations, we ignore taxes in this investigation, and thus transac tion costs are just brokerage commissions. if an investor’s portfolio blend is revised each quarter (1 qtr) to follow a brokerage firm’s recommendations, then brokerage commis sions will be incurred each quarter. in contrast, if the portfolio blend is revised only once per year (4 qtr), then brokerage commissions will be incurred only at the beginning and end of each twelve-month period. in addition, brokerage costs are not the same for all asset classes. commission costs for stocks are generally higher for stocks than they are for either bonds or cash. let s rep resent the one-way percentage commission for common stocks; let b represent the one-way commission for bonds; and let c represent the one-way commission for cash. usually, one would expect s > b > c. we assume that s, b, and c are the same for all brokerage firms. however, the dollar commissions for an investor following different brokerage firms will not be the same-because the suggested asset allocations by brokerage firms are different each quarter, and also because their suggestions vary from one quarter to the next. in what follows, we assume that s = l.o%, b = os%, and c = 0.0%.3 table 4 uses the asset allocation recommendations of a.g. edwards to illustrate how brokerage commissions are incorporated into the analysis. the left hand panel includes the three market indicators, while the middle panel includes the quarterly recommendations of a.g. edwards for the entire time period of the investigation. as of 89-3, a.g. edwards rec ommended 45% common stocks, 45% bonds, and 10% cash. the right hand panel of table 4 determines the impact of brokerage commissions for each of the three choices of investment horizon. the “sum” column indicates the available percentage of initial portfolio wealth before commissions at that point in time. the right-adjacent column in each instance (1 qtr, 2 qtr, or 4 qtr) indicates for that choice of investment horizon the wealth relative after com missions are paid to implement a.g. edwards’ suggested asset allocations. the available percentage before commissions at the start of the time period (i.e., 89-3) would be 100%. the after-commissions wealth relative at 89-3 would be (.ol) [loo% (45%)(1.0%) (45%)(0.5%) (10%) (o.o%)] = 0.99325, and thus reflects the initial brokerage costs to construct the recommended portfolio. suppose the investor decides on the shortest (i.e., quarterly) investment horizon. here is a breakdown of the available percentage three months later at 89-4: common stocks bonds cash available percentage (0.99325) (45%) (353.40/349.15) = 45.240% (0.99325) (45%) (104.74/101.37) = 46.182 (0.99325) (10%) (100.48/99.23) = 10.058 10 1.480% t a b l e 4 a na ly si s of t ac tic al a ss et a llo ca tio n r ec om m en da tio ns by a . g . e dw ar ds q ua rt er ly , 19 89 -1 99 4 m ar ke t in di ca to rs y ea r & q ua rt er s& p 50 0 us lu ng u s b ill % r uc om m m da ti on s st oc ks b on ds c as h su m r es ul ts fo r a lt er na ti ve in ve st m en t h or iz on s 1 q tr su m 2q tr su m 4 q tr 89 -3 34 9. 15 10 1. 37 99 .2 3 45 45 10 10 0. 00 0 0. 99 32 5 89 -4 35 3. 40 10 4. 74 10 0. 48 45 45 10 10 1. 48 0 1. 01 47 3 90 1 33 9. 94 97 .2 7 98 .2 0 40 45 15 96 .2 47 0. 96 18 8 90 -2 35 8. 02 98 .1 2 99 .0 2 40 45 15 98 .7 33 0. 98 71 9 90 -3 30 6. 05 93 .5 4 99 .6 2 40 40 20 91 .0 03 0. 90 94 7 90 -4 33 0. 22 10 1. 16 10 2. 48 45 40 15 97 .3 06 0. 97 25 9 91 -l 37 5. 22 10 0. 19 10 3. 10 55 35 10 10 2. 93 8 1. 02 85 7 91 -2 37 1. 16 98 .6 3 10 3. 60 50 40 10 10 1. 73 4 1. 01 65 7 91 -3 38 7. 86 10 4. 36 10 5. 60 55 40 5 10 6. 50 2 1. 06 44 5 91 -4 41 7. 09 10 8. 94 10 8. 65 55 40 5 11 2. 88 0 1. 12 86 8 92 -1 40 3. 69 10 5. 03 10 6. 75 55 40 5 10 9. 15 4 1. 09 15 3 92 -2 40 8. 14 10 6. 77 10 9. 28 50 40 10 11 0. 66 7 1. 10 61 3 92 -3 41 7. 80 11 3. 72 11 3. 47 45 40 15 11 5. 22 6 1. 15 17 4 92 -4 43 5. 71 11 .9 4 11 .9 7 50 35 15 11 6. 44 6 1. 16 38 1 93 -l 45 1. 67 11 9. 78 11 2. 97 45 40 15 12 1. 52 1 1. 21 44 0 93 -2 45 0. 53 12 1. 51 11 2. 64 55 40 5 12 1. 95 0 1. 21 82 2 93 -3 45 8. 93 13 0. 20 11 4. 01 65 30 5 12 6. 63 1 1. 26 41 9 93 -4 46 6. 45 12 5. 79 11 3. 41 65 30 5 12 6. 44 7 1. 26 42 8 94 1 44 5. 77 11 3. 80 10 7. 70 60 25 15 11 8. 85 1 1. 18 75 6 94 -2 44 4. 27 10 9. 30 10 5. 40 50 40 10 11 6. 96 1 1. 16 74 5 94 -3 46 2. 69 10 5. 60 10 4. 50 45 40 15 11 7. 48 4 1. 17 39 6 94 -4 45 9. 27 10 5. 60 10 1. 10 40 45 15 11 6. 43 2 1. 16 34 7 95 -l 50 0. 7 1 10 9. 80 10 3. 60 40 40 20 12 3. 06 0 1. 22 28 0 a ve ra ge 49 .3 5 39 .1 3 11 .5 2 b ro ke ra ge c om m is si on s 1 . o % 0. 5% 0. 0% 5 00 99 .3 25 0. 99 32 5 99 .3 25 0. 99 32 5 95 .9 36 0. 95 88 4 89 .8 54 0. 89 81 2 90 .3 94 0. 90 33 4 10 1. 76 2 1. 01 62 9 10 5. 41 5 1. 05 38 8 10 5. 25 8 1. 05 17 3 10 8. 05 8 1. 08 04 5 11 4. 14 5 1. 14 03 8 11 3. 80 3 1. 13 71 3 12 1. 54 0 1. 21 52 9 12 8. 22 3 1. 27 87 3 12 5. 42 4 1. 25 06 2 11 9. 85 3 1. 19 74 6 11 9. 02 2 1. 18 71 3 12 5. 69 9 1. 24 87 4 11 8. 11 7 1. 17 87 9 12 3. 96 1 1. 23 17 0 asset allocation and investment horizon 213 table 5 comparison between actual and recommended asset proportions at 89-4, and the necessary adjustments. acturrl weight recommended adjustment common stocks 45.240/101.480 = 44.580% 45% buy bonds 46.182/101.480 = 45.508% 45% sell cash 10.058/101.480 = 9.912% 10% buy these component and total percentages facilitate a comparison between actual and recom mended asset proportions at 89-4, and the necessary adjustments that must be made (see table 5). in this instance, the necessary adjustments are not large, because a.g. edwards did not change their recommended proportions from 89-3 to 89-4. the after-commission wealth relative at 89-4 would thus be given as (.01)(101.48)[1 (.ol) 1 .44580 .45 1 (.005) 1 .45508 .45 1 -01 = 1.01473 absolute values are used in the expression because some adjustments are to buy more of an asset category, while other adjustments are to sell some of an asset category. similar calculations are done for each quarter, except that at the end of the time period (95-l), the portfolio holdings are assumed to be sold. the final after-commission wealth relative (1 qtr) is 1.22280, which means that the investor’s wealth increased by 22.28%. that is a “cash-to-cash’ result in that brokerage commissions are incurred at the beginning, at each quarter adjustment, and at the end of the time period. alternatively, if the investor selected a six-month investment horizon, then portfolio adjustments would be made every other quarter. the result is an after-commission wealth relative (2 qtr) of 1.24874, which is about 2.6% higher than the former. and for a full year investment horizon with adjustments just once per year, the result is 1.23170 and thus between the results for shorter investment horizons. the results are comparable for each of the three investment horizons because portfolios are constructed at 89-3, and they are liq uidated at 951. differences are in how often the portfolios are adjusted. by changing all the commissions to zero, it is possible to obtain a before-commission wealth relative for each investment horizon, as well as a wealth relative that reflects aggre gate brokerage commissions for the time period. for example, the 1 qtr before-commis sion wealth relative was 1.25559 for a.g. edwards. because wealth relatives are multiplicative, the aggregate brokerage commission wealth relative also can be calculated. for a.g. edwards, it was 1.22280 i 1.25559 = .97388, or about 2.6% for the entire time period. table 6 presents in wealth relative terms the before-commission performance, the aggregate brokerage cost for the time period, and hence the after-commission performance. results are presented for three different investment horizons, as well as for each of eight brokerage firms that made recommendations during the entire time period (89-3 through 94-4).4 first, we note as expected that the wealth relatives for brokerage commissions increased with the length of the investment horizon. the effective commission averaged 3.0% (1 0.970) for a quarterly horizon, 2.6% for a six-month horizon, and 2.3% for a full-year investment horizon. second, the highest wealth relative is noted for each invest t a b l e 6 im pa ct of b ro ke ra ge c om m is si on s on p ri ce a pp re ci at io n fo llo w in g r ec om m en de d a ss et a llo ca tio ns by n at io na l b ro ke ra ge fi rm s e nt ir e t im e pe ri od 19 89 -1 99 4 i q tr h or iz on 2 q tr h or iz on 4 q tr h or iz on b ro ke ra ge f ir m b ef or e c om m is si on a ft er b ef or e c om m is si on a ft er b ef or e c om m is si on a ft er a . g . e dw ar ds 1. 25 6 0. 97 4 1. 22 3 1. 27 8 0. 97 7 1. 24 9 1. 25 7 0. 98 0 1. 23 2 d ea n w itt er 1. 22 0 0. 96 8 1. 18 2 1. 23 5 0. 97 3 1. 20 2 1. 25 1 0. 97 3 1. 21 7 g ol dm an sa ch s 1. 34 9* 0. 96 5 1. 30 2* 1. 32 7” 0. 96 8 1. 28 4* 1. 27 5 0. 97 5 1. 24 3 m er ri ll l yn ch i. 26 5 0. 97 4 1. 23 1 1. 21 9 0. 97 9 1. 25 2 1. 25 1 0. 98 1 1. 23 2 pa in e w eb be r 1. 30 8 0. 97 1 1. 27 1 1. 29 5 0. 97 3 1. 26 0 1. 41 0* 0. 97 9 1. 38 1* pr ud en tia l 1. 31 7 0. 95 6 1. 25 9 1. 29 9 0. 96 3 1. 25 1 1. 29 4 0. 96 6 1. 25 1 r ay m on d ja m es 1. 28 7 0. 97 4 1. 25 4 1. 29 6 0. 97 6 1. 26 5 1. 26 6 0. 97 7 1. 23 7 sm ith b ar ne y 1. 25 4 0. 97 8 1. 22 6 1. 25 5 0. 98 2 1. 23 2 1. 25 1 0. 98 3 1. 22 9 a ve ra ge 1. 28 2 0. 97 0 1. 24 4 1. 28 3 0. 97 4 1. 24 9 1. 28 3 0. 97 7 1. 25 3 s l& p5 00 1. 31 5 0. 98 0 1. 28 9 1. 31 5 0. 98 0 i. 28 9 1. 31 5 0. 98 0 1. 28 9 us l on g 1. 03 8 0. 99 0 1. 02 8 1. 03 8 0. 99 0 1. 02 8 1. 03 8 0. 99 0 1. 02 8 u s b ill 1. 02 1 1. 00 0 1. 02 1 1. 02 1 1. 00 0 1. 02 1 1. 02 1 i. 00 0 1. 02 1 n ot e: * h ig he st w e a lth re la tiv e fo r th a t in ve st m e n t h o riz o n asset allocation and investment horizon 215 ment horizon, both before and after brokerage commissions. investors who followed gold man sachs did the best for both 1 qtr and 2 qtr investment horizons, while investors who listened to paine webber did the best for the 4 qtr horizon. brokerage commissions overall did not change those particular results, but rankings of performance for all eight firms did change when commissions were brought into the picture. in other words, there were instances when commissions from more frequent port folio revision did offset the advantages of changing portfolio blends every quarter, or more frequently. naturally, it is the after-commission results that ultimately matter to investors. table 6 also reports the corresponding wealth relatives for individual invest ments (in common stocks, bonds, and cash, respectively) both before and after brokerage commissions. a key extension of the results in table 6 is to examine how well tactical asset alloca tion measures up to alternatives such as buy-and-hold and strategic asset allocation, when brokerage commissions are included. that was mentioned at the outset as one of the moti vating questions for this investigation. by holding the asset allocations of a given broker age firm constant over time, and repeating the calculations, the effect of strategic asset allocation can be analyzed. in turn, the effect of a buy-and-hold strategy can be determined if there are no interim adjustments between purchasing the three-asset portfolio at the beginning of the time period and selling that portfolio at the end. in table 7, the after-commission wealth relatives for tactical asset allocation by eight national brokerage firms during the entire time period (89-3 through 95-l) are included in the left-hand panel. the results of strategic asset allocation for the same period appear in the next two panels, followed by the buy-and-hold results in the right-hand panel. for both strategic asset allocation and buy-and-hold, there are two versions. one version uses the initiue brokerage firm suggestions at 89-3, while the other version uses the average recom mendations of each brokerage firm during the entire time period. for tactical asset allocation, investors following goldman sachs experienced the high est wealth relative for 1 qtr and 2 qtr investment horizons, while paine webber’s investors did the best for a 4 qtr investment horizon. for strategic asset allocation, dean witter pro vided the best performance if their investors maintained their initial recommendations, but goldman sachs was best if their investors utilized the average recommendations through out the time period. for buy-and-hold, the results are the same. investors following dean witter did the best if they maintained the initial (89-3) blend, while investors following goldman sachs did the best if they utilized their average suggestions in a buy-and-hold strategy. further perspective is available if the results in table 7 are examined by investment horizon rather than by type of asset allocation. for an investment horizon of one quarter, the best result was strategic asset allocation using dean witter’s initial blend of 85% com mon stocks and 15% bonds. for an investment horizon of six months, the best result is the same. but for an investment horizon of twelve months, the best result would have been for investors who followed paine webber’s asset allocation recommendations every four quarters. finally, if one examines just the average results for the sample of eight national bro kerage firms, the best result was buy-and-hold over the entire time period using the initial recommendations at the end of the third quarter of 1989. that result, 1.270, is second (in table 7) only to the 1.38 1 result that would have been achieved in investors followed paine t a b l e 7 c om pa ri so n of a ss et a llo ca tio n st ra te gi es b as ed on b ro ke ra ge fi rm r ec om m en da tio ns e nt ir e t im e pe ri od 19 89 -l 99 4 b ro ke ra ge f ir m t ac ti ca l a ll oc at io n s tr u te gi c a ll oc at io n (i n it ia l) s tr at eg ic a ll oc at io n (a ve ra ge ) b u yan d -h ol d i q tr 2 q tr 4 q tr 1 q tr 2 q tr 4 q tr 1 q tr 2 q tr 4 q tr in it ia l a ve ra ge a . g . e dw ar ds 1. 22 3 1. 24 9 d ea n w itt er 1. 18 2 1. 20 2 g ol dm an sa ch s 1. 30 2* 1. 28 4* m er ri ll l yn ch 1. 23 1 1. 25 2 pa in e w eb be r 1. 27 1 1. 26 0 pr ud en tia l 1. 25 9 1. 25 1 r ay m on d ja m es 1. 25 4 1. 26 4 sm ith b ar ne y 1. 22 6 1. 23 2 1. 23 2 1. 21 4 1. 21 8 1. 21 7 1. 35 3* 1. 35 3* 1. 24 3 1. 28 2 1. 28 5 1. 23 2 1. 23 3 1. 23 4 1. 38 1* 1. 13 1 1. 15 6 1. 25 1 1. 29 9 1. 30 2 1. 23 7 1. 26 1 1. 27 0 1. 22 9 1. 22 8 1. 23 6 1. 21 6 1. 22 8 1. 23 3 1. 29 9 1. 35 2* 1. 26 5 1. 26 9 1. 22 9 1. 28 4 1. 28 3* 1. 28 5* 1. 28 2* 1. 23 6 1. 24 6 1. 25 0 1. 24 7 1. 12 5 1. 25 7 1. 26 1 1. 25 9 1. 30 0 1. 26 4 1. 27 1 1. 26 5 1. 26 1 1. 27 1 1. 27 9 1. 27 0 1. 22 8 1. 23 2 1. 23 8 1. 23 3 1. 22 0 1. 23 4 1. 35 6* 1. 27 1 1. 28 8 1. 28 6* 1. 23 9 1. 25 2 1. 25 2 1. 26 3 1. 30 6 1. 26 3 1. 26 8 1. 27 7 1. 23 5 1. 24 0 a ve ra ge 1. 24 4 1. 24 9 1. 25 3 1. 25 0 1. 25 7 1. 25 0 1. 25 6 1. 26 1 1. 25 2 1. 27 0 1. 26 1 s& p5 00 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 1. 28 9 u s l on g 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 1. 02 8 u s b ill 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 1. 02 1 n o te : * h ig he st w e a lth re la tiv e fo r th a t in ve st m e n t h o riz o n . asset allocation and investment horizon 217 webber’s suggestions, but with only a single revision each year. that is strong evidence against the advisability of tactical asset allocation, especially if it is done quarterly. vii. risk the final question concerns the risk that investors face if they follow the asset allocation recommendations of the national brokerage firms. risk thus far has been reflected in the variability of common stock, bond, and cash recommendations from the third quarter of 1989 through the end of 1995 for the sample of thirteen brokerage firms. variability was seen to be greater for some firms and less for others. variability of recommendations also tended to change from one quarter to another during the entire time period. risk perspec tive also was reflected in the variability of appreciation for the market indicators chosen for common stocks, bonds, and cash, respectively. so investors’ first experience with risk is a result of their choices of asset categories to consider, and a particular brokerage firm’s rec ommendations to follow. we saw that price appreciation depends on the particular time period in which asset allocation recommendations are followed. if it happens to be a time span (like 89-3 through 95-l) when stocks do better than bonds or cash, then common stock recommendations of the most “bullish’ brokerage firms will lead to the best performance. it is quite the opposite in periods of time when bonds, or even cash, perform better than common stocks. so inves tors’ second experience with risk is when they select a particular time span in which to try to benefit by following the advice of a brokerage firm. investors also must select an investment horizon, defined in this paper as the time between investment decisions. if shorter horizons lead to greater turnover among the asset classes, and thus to greater brokerage commissions, it may actually detract from achieved performance. the title of this paper indeed suggests that investors’ third experience with risk is in their choice of a particular investment horizon. investors’fourth experience with risk is in their overall decision about asset allocation itself. our investigation was motivated by the tactical asset recommendations of the national brokerage firms. as one performance benchmark, we looked at strategic alloca tion, wherein investors re-balanced or restored the recommended asset allocations of a given firm at each investment horizon. as another performance benchmark, we examined buy-and-hold, wherein investors simply began with a recommended asset allocation and no further adjustments were made during the time period. the results presented in tables 6 and 7 readily attest to the wide variety of outcomes that occurred as a result of investor choices among buy-and-hold, strategic asset allocation, and tactical asset allocation fol lowing the recommendations of the national brokerage firms. to summarize the presence of risk, consider again the results for dean witter. if inves tors had followed that firm’s tactical recommendations for the entire period, their achieved performance would have been considerably below the sample average, regardless of invest ment horizon. if investors instead had adopted dean witter’s 89-3 recommendations, and made periodic adjustments to restore that blend (i.e., strategic asset allocation), they would have done better than average, and in fact better than by following any of the other seven firms. investors’ achieved performance would have been better than that if they had followed dean witter’s 89-3 recommendation, but in a buy-and-hold strategy. the best result, of 218 financial services review 6(3) 1998 course, would have been achieved by investors who simply bought and held common stocks for the entire time period. viii. conclusion an important message of this empirical study is that investment horizon is a critical factor in how individual investors make and implement their asset allocation decisions. brokerage commissions also play an important role as investors try to determine just how often they should follow the asset allocation recommendations of a national brokerage firm. and we have taken a qualitative look at some of the dimensions of risk that investors face in their asset allocation choices. in sum, the quarterly wall street journal practice of reporting the recommended asset allocations of national brokerage firms is an interesting look at how market prospects are viewed by those firms. however, to implement their tactical asset allocation recommenda tions every quarter may not be in the best interest of investors. alternatively, strategic asset allocation, and even the simpler buy-and-hold strategy, together with their savings of bro kerage commissions, may indeed be investment strategies that should be thoughtfully con sidered by many investors. notes 1. the series of articles under the general title “your money matters” are written by j. r. dorfman, staff reporter of the wall street journal. helpful comments on this paper were provided by mr. dorfman, as well as by professors s. badrinath, s. chakravarty, and m. cooper. 2. in their performance calculations, the wall street journal uses the s & p 500 and u.s. treasury bills as surrogates for common stocks and cash, respectively, but they use merrill lynch’s corporates/govemments domestic master bond index for bonds. the source of data for the three mar ket indicators used in this study was the security price index record statistical service, which is pub lished annually by the standard & poor’s corporation. 3. as long as s > b > c, modest changes in the commission levels for common stocks, bonds, and cash do not change the empirical results of this investigation. 4. to conserve space, and because the results are very similar, we do not extend the exhibits to include either the earlier or later time periods. references bierman, h. (1997). portfolio allocation and the investment horizon. journal of portfolio manage ment, 23,51-55. black, f., & litterman, r. (1991). asset allocation: combining investor views with market equilib rium. journal of fixed income, i, 7-l 8. blume, m., & friend, i. (1975). the asset structure of individual portfolios and some implications for utility functions. journal of finance, l&585-603. brimson, g. p., hood, l. r., & beelower, g. l. (1986). determinants of portfolio performance. financial analysts journal, 42, 39-44. asset allocation and znvestment horizon 219 brimson, g. p., et.al., (1990). determinants of portfolio performance ii: an update. financial ana lysts journal, 47,40-48. clements, j. (1995, january 17). market timing is a poor substitute for a long-term investment plan. wall street journal, c 1. cohn, r. a., lewellen, w. g., lease, r. c., & schlorbaum, g. g. (1975). individual investor risk aversion and investment portfolio composition. journal of finance, 30.605-620. hemiksson, r. d., (1984). market timing and mutual fund performance: an empirical investigation. journal of business, .57,3&45. jeffrey, r. h., (1984). the folly of stock market timing. harvard business review, 62, 102-l 10. perold, a. f., & sharpe, w. f., (1988). dynamic strategies for asset allocation. financial analysts journal, 44, 16-27. philips, t. k., rogers, g. t., & capaldi, r. e. (1996). tactical asset allocation: 1977-94. journal of port$olio management, 23,574. sharpe, w. f. (1975). likely gains from market timing. financial analysts journal, 31,60-69. sharpe, w. f. (1986). integrated asset allocation. financial analysts journal, 42, 15-24. smith, k. v. (1974). the major asset mix problem of the individual investor. journal of contempo rary business, 3,49-67. smith, k. v. (1997). portfolio analysis of brokerage firm recommendations. znstitute paper no. 1109, krannert graduate school of management, purdue university, west lafayette, indiana. tarrazo, m. j. (in press). an application of fuzzy set theory to the individual investor problem. finan cial services review. walker, m. m., & hatfield, g. b. (1996). professional stock analysts’ recommendations: implica tions for individual investors. financial services review, 5, 13-29. waring, m. b., (1994). 401(k) investment strategy and asset allocation: deja vu all over again. jour nal of investing, 3, 17-22. pii: s1057-0810(96)90006-2 financial services review, 5(z): 133-148 copyright q 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the effects of mutual fund managers’ ~haracte~stics on their por~olio ~e~ormance, risk and fees joseph h. golec the purpose of this study is to test whether a mutual fund managers’ characteristics help to explain fund per$ormance, risk and fees. the statistical tests consider per&or mance, risk andfees simultaneously to avoid biased results produced by earlier studies that ignore simultaneity. results show that a fund’s performance, risk andfees are sig nificantly impacted by its manager’s characteristics. all else equal, investors can expect better risk-adjusted performance from younger managers with mba degrees who have longer tenure at their funds. also, funds with low fees and more diversified po~olios~e~o~ better. the most sign~~cant predictor ofpe~o~ance is the length of time a manager has managed his or her find (tenure). funds that keep administrative expenses low also perform relatively well, but large management fees do not necessar ily imply poorer performance. apparently, a large management fee signals superior inves~ent skill which leads to better perfomtance. i. introduction managers make investment decisions based upon their personal abilities and risk prefer ences. this paper models a simultaneous system for a large sample of mutual fund manag ers in order to determine the effects that human capital characteristics have on fund return performance, risk and fees. that fund managers’ characteristics simultaneously determine their portfolio return performance and risk as well as their own compensation is not surprising; yet, earlier stud ies have not accounted for this simultaneity. for example, it follows from human capital theory that managers with greater human capital (intelligence, etc.) should produce better performance and receive better compensation. similarly, agency models, such as those of barry and starks (1984), starks (1987), cohen and starks (1988), and golec (1988, 1992) show that a manger’s portfolio risk choices will partly depend upon his or her risk-taking preferences because the volatility of a manager’s pay is affected by the portfolio’s perfor joseph ii. golec l associate professor of finance, clark university, graduate school of management, 950 main street, worcester, ma 01610. 134 financial services review s(2) 1996 mance. this study’s statistical approach accounts for the fact that performance, risk, and fees are interdependent. mutual fund performance alone is an important and popular finance topic because funds positive risk-adjusted returns has implications for market efficiency. most early studies, such as jensen (1968) and sharpe (1966), report that funds provide inferior perfor mance partly because of management fees and other expenses. recently, however, ippolito (1989), lee and rahman (1990), grinblatt and titman (1989, 1992), and hendricks, patel, and zeckhauser (1993) show that mutual funds can generate systematic positive risk adjusted returns. although ippolito’s sample of funds earned sufficient ask-adjusted returns to cover fees, elton, gruber, das, and hlavka (1993) question ippolito’s methods and suggest that funds do not exhibit positive risk-adjusted returns. whether mutual fund managers produce superior returns is controversial because most studies’ funds, sample periods, or performance measures are not comparable. unlike ear lier studies that try to determine if the average risk-adjusted fund pe~o~ance is positive, this study only requires that a performance measure rank funds appropriately. for example, if longer tenure implies greater human capital which, in turn, generates better performance, then job tenure should be positively related to performance. this positive relationship can be present even if all funds have negative risk-adjusted performance; long-tenured manag ers will simply have less negative ~~o~~ce. earlier studies consider relatively long time periods during which some funds change managers, risk, fees or objective, or liquidate. here, the cross-sectional data and shorter sample period reduce the degree of fund changes and survivorship bias (brown, goetz mann, ibbotson, & ross, 1992). the paper is organized as follows. section i discusses the statistical procedure used to account for simultaneity and defines the study’s endogenous and exogenous variables. sec tion ii describes the data. section iii presents each structural equation along with the results for each equation. section iv considers the issues of survivorship bias and perfor mance measurement. section v summarizes the results that have the most significant implications for investors’ choice among mutual funds and their managers. ii. three-stage least squares many earlier studies, such as sharpe (1966), jensen (1968), friend and blume (1970), ippolito (1989), grinblatt and titman (1989, 1992), hendricks, patel, and zeckhauser (1993) and elton et al. (1993), compare mutual funds’ risk-adjusted performance, as well as other endogenous variables (risk or fees), but ignore the fact that changes in perfor mance, risk, and fees tend to impact each other ~ontem~~eously. for example, a fund that increases fees will tend to have poorer performance, all else equal. in this case, fees enter as an independent variable in an equation explaining performance. clearly, errors in explaining fees will feed into errors in explaining performance. that is, a fund with unex plained large fees will have a large fee error, producing a relatively large performance error. this means that an independent variable (fees) will be correlated with the error term in the performance equation. ordinary least squares (ols) assumes independent variables and errors are uncorrelated; otherwise, ols coefficient estimates will be biased and inconsistent. mutual fund managers 135 three-stage least squares (3sls) offers consistent estimates and the large sample used in this study takes full advantage of this consistency. for example, 3sls eliminates the correlation between fees and errors in the performance equation by replacing actual fees by their estimated values obtained by regressing fees on fixed exogenous variables only. in other words, errors in fees do not feed into the performance regression because the fee errors are eliminated before the performance regression is estimated. in this study, the endogenous or simultaneously determined variables include portfolio yield and alpha (performance); portfolio beta and the standard deviation of residual portfo lio returns (risk); and expenses exclusive of management fees, management fees, and port folio turnover (fees). exogenous variables include manager age, tenure with the fund, years of education, whether or not the manager has an mba degree, management team size (usu ally one), fund age, fund assets, load charge, and fund objective. yield measures a manager’s propensity to choose high-dividend stocks. because man agement fees are paid as a proportion of fund assets, the more a fund pays out, the smaller its asset base and management fees, all else equal. managers may choose stocks with large dividends as a consequence of a “value” investing style or because they believe they can attract investors who prefer large dividends. conversely, a “growth” style or investors who prefer small dividends (tax avoidance) may imply small dividends. alpha is jensen’s measure of return performance adjusted for systematic risk. alpha measures the portfolio return attributable to the manager’s skill (or luck). systematic risk is measured by beta. unsystematic risk is the residual variation in portfolio return after accounting for variation due to beta risk. it measures the degree of portfolio diversification and beta stability. managers must deviate from a perfectly diversified fixed-beta portfolio if they wish to obtain a nonzero alpha. expenses exclusive of management fees measure administrative, operating, and cus tomer service expenses including 12b-1 fees. management fees are charged by the fund’s management company to cover the portfolio manager’s compensation, as well as their operating expenses, research support, and profit. although manager pay is not separately available, it is assumed that larger pay leads to larger fees. portfolio turnover measures the manager’s trading propensity. trueman (1988) sug gests that trading is a signal that a manager is gathering and trading on information. like management fees and expenses, increased turnover increases costs which are paid out of returns. on the other hand, management fees and turnover costs are presumably paid to facilitate return-producing input by fund managers. manager age measures experience but also gauges stamina for a demanding job. many in the mutual fund industry believe that investment management is so demanding that the negative impact of age on stamina leads to poorer performance. in addition, age indirectly measures time until retirement and, hence, the importance of future job income to the man ager. if tenure is a better measure of experience than age, age may largely capture the neg ative stamina effect. tenure measures the manager’s survivorship at the job. long tenure implies that the management company finds the manager’s ability and performance satis factory but may also indicate that the manager has few better opportunities because of spe cialized skills or an unspectacular performance record. years of education measures accumulated general knowledge while mba measures business-specific knowledge. an mba should know some basic tenets of investing as well as how to recognize firms with good management. team size will measure whether more heads are better than one or if investment decisions made by committee are ineffective. 136 financial services review 5(2) 1996 fund age measures a fund’s survivorship, its prestige, and the loyalty of its investors. fund assets measure a fund’s market acceptance, past growth, and economies of scale. load is an additional expense paid by investors of some funds and, thus, load funds may have to reduce other fees in order to compete. finally, the funds used in this study are pri marily stock funds with objectives including growth, aggressive growth, growth and income, small stocks, specialized, balanced, asset allocation, option equity, and intema tional. these fund variables will pick up average effects, such as the higher yields expected for growth and income funds. hi. the data the data sample spans 1988-1990 and is composed of 530 of the 979 mutual funds listed in the 1991 issue of mutual fund sourcebook volume i, published by momingstar inc. the distribution of funds by objective is 181 growth, 105 growth and income, 67 special, 50 small stock, 43 international, 41 balanced, 30 aggressive growth, 7 option equity, and 6 asset allocation. funds were excluded from the sample if momingstar did not report infor mation for them on the variables mentioned in the previous section; however, most of the funds excluded were eliminated because they had fewer than three years of performance history. the variables were calculated in the following ways: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. alpha, beta, and residual return standard deviation are calculated for each fund with monthly return data over 1988-1990 using the capital asset pricing model with treasury bills as the risk-free asset and the standard and poor’s 500 stock index (s&p 500) return as the market portfolio. yield is defined as annual fund income excluding capital gains divided by year end assets. expense ratio is the percentage of fund assets spent on operating expenses (excluding management fees and brokerage costs). management fee is the percentage of assets paid as management fees. turnover is the percentage of total assets sold during a year. fund assets are net year-end assets measured in millions. fund load is the percentage of new investments that must be paid to the fund as a sales charge. team size is the number of managers who make investment decisions for the fund (usually one). manager age, tenure, education, and fund age are measured in years, with 1990 as the end-year. when more than one manager is involved in the fund, the lead or more senior manager’s characteristics are used. dummy variables represent mba (mba=l, other=o) and fund objectives, where growth is the comparison type. table 1 lists the sample statistics for the variables. noteworthy is the fact that the aver age beta is less than one (0.84) and the average alpha is -2.83 percent per year. average mutual fund managers 137 turnover is nearly 100 percent (91.72%) with a maximum turnover of almost 800 percent. the average fund manager is 46 years old, holds an mba (64 percent) and has seven years tenure. the typical fund is 16 years old, has $280 million of assets, charges a 3.14 percent load, and has a growth objective (34%). the relatively large proportion of growth funds included in the sample reflects investor preference for such funds. iv. results the specification of each structural equation and its statistical results are presented together in order to focus the presentation. both the 3sls structural and reduced form coefficients are reported in the tables and statistically significant coefficients are starred (t statistics are available upon request). the reduced form coefficients will be discussed when they differ significantly from the structural coefficients. the structural coefficients represent the direct effects of the included right-hand-side exogenous and endogenous variables on the left-hand-side depen dent endogenous variable. by comparison, the reduced form coefficients combine the direct effect of an exogenous variable on the dependent variable, with the indirect effects implied by the endogenous variables that are included in the structural form but excluded from the reduced form. table 1 sample statistics for the variables used in simultaneous regression analysis endogenous variables yield (%) alpha (cub) beta residual s. dev. (%) expense ratio (%) management fee (%) turnover (%) mean standard deviation minimum maximum 2.59 2.21 0.00 13.10 -2.83 4.49 -23.80 16.05 0.84 0.29 -0.12 1.67 7.25 3.84 0.00 20.56 0.79 0.76 0.00 10.10 0.73 0.22 0.05 2.00 91.72 89.37 0.00 789.00 e.xogenous variables manager age (years) tenure (years) years of education mba degree team size fund age (years) fund assets (millions) load (%) 45.96 10.33 26.00 82.00 6.95 6.14 1.00 51.00 17.54 0.89 16.00 19.00 0.64 0.48 0.00 1.00 1.23 0.63 1 .oo 5.00 16.27 15.94 3.00 87.00 280.37 769.60 1.00 11980.00 3.19 2.76 0.00 8.50 fund objectives growth aggressive growth growth and income small stock special balanced asset allocation option equity 0.34 0.47 0.00 1.00 0.06 0.23 0.00 1.00 0.20 0.40 0.00 1.00 0.09 0.30 0.00 1.00 0.13 0.33 0.00 1.00 0.08 0.26 0.00 1.00 0.01 0.10 0.00 1.00 0.01 0.11 0.00 1.00 internati&ai 0.08 0.27 0.00 1.00 138 financial services review 5(2) 1996 a. the performance equations yield is included as a performance measure because many investors consider yield in their selection of a fund and because managers through stock selection have significant control of a fund’s yield. the structural equation for yield is: yield = (expense, fee, turnover, manager age, tenure, fund age, assets, objectives) (1) expenses, management fee and turnover should be negatively related to fund yield, all else equal. they represent costs that may be paid out of a fund’s cash flow which would otherwise go to shareholders. results for the yield regression reported in table 2 show that fund yield is significantly negatively related to management fees, as expected, while the expense ratio and turnover coefficients are insignificant. yield may be negatively related to manager age because older managers nearing retirement can boost fees somewhat by reducing payouts and growing assets. such short run behavior by older managers is documented in gibbons and murphy (1992) and dechow and sloan (1991). results, however, show no significant relationship between age and yield. yield may be positively related to tenure for precisely the opposite reason that age was predicted to be negatively related to yield. that is, long tenure may imply greater job secu rity, and hence, less short-run behavior by the manager. indeed, while selecting stocks pay ing high dividends reduces management fees now, high dividend yield may attract more investors to the fund in the future, increasing assets and fees. as predicted, results show that yield and tenure are significantly positively related. the relationship between fund age and yield may be positive. to the extent that inves tors prefer funds with larger dividends, funds providing larger dividends survive longer. table 2 reports a positive fund age coefficient, but the coefficient is statistically insignifi cant. as funds grow, managers typically invest in larger companies that usually pay rela tively large dividends. on the other hand, this effect could be offset because a larger divi dend payout means less assets, all else equal. the positive asset coefficient indicates that the effect of investing in larger companies dominates. furthermore, the 3sls structural coefficient (0.217) is smaller than the reduced form coefficient (0.283), indicating that large funds probably have proportionately smaller expenses which, in turn, lead to larger yields as well. this point illustrates the value of the simultaneous model. because fees enter the structural model, there will be an indirect effect of assets on yield through fees. as shown below, more assets lead to lower fees and, as noted above, lower fees lead to larger yields. the reduced form assets coefficient picks up both effects; hence it is larger than the structural coefficient. fund objective will impact yield since yield requirements may be written into a fund’s charter. the coefficients on the fund objective dummies are all as one might expect; for example, growth and income funds provide a 1.874 percentage point greater yield than growth funds (the comparison group) on average. note that the larger reduced form coeffi cient implies a 2.002 percentage point greater yield for growth and income funds because of the indirect effect of growth and income funds’ smaller expenses on yield. mutual fund managers 139 overall, the most notable result from the yield equation is that long-tenured managers tend to boost fund yield. investors who prefer larger yields, all else equal, should find them at funds with managers with relatively long tenure (greater than seven years). the alpha equation is: alpha = (beta, residual, expense, fee, turnover, manager age, tenure, education, mba, team size, fund age, assets) (2) friend and blume (1970) show that alpha and beta are weakly negatively related. residual standard deviation coefficient may be negatively related to alpha because noise trading by fund managers has negative performance consequences (see black, 1986). mis specification of the asset-pricing model can also lead to cross-sectional correlation between alpha, beta, and residual standard deviation. as expected, table 2 shows that alpha is negatively related to beta and residual standard deviation although the beta rela tionship is not significant. at a basic level, expense ratio, management fee and turnover should all be negatively related to alpha because the costs are deducted from shareholder returns. but the manage table 2 yield and alpha 3sls structural and reduced form regressions yield regressions structural reduced alpha regressions structurul reduced endrogenous variables intercept yield (%) alpha (%) beta residual st. dev. (%) expense ratio (%) management fee (%) turnover (%) exogenous variables 2.618* 2.559 1.210 -2.269 -0.469 -0.410* xl.014 -1.178* -1.590* 5.092* 0.003 0.006 manager age (years) tenure (years) years of education mba degree team size fund age (years) fund asset? (millions) load (%) fund objectives growth aggressive growth growth and income small stock special balanced asset allocation option equity international 0.001 a.003 0.053* 0.052* -0.047 0.065 a.070 0.001 0.005 0.217* 0.283* 0.012 -1.043* 1.874* -1.034* -0.449** 3.279* 2.880* 2.129* -0.237 -0.900* 2.002* 4.986* -0.321 3.469* 2.872* 2.057* -0.332 -0.076* -0.083* 0.165* 0.185* xi.084 0.165 0.943** 0.599 -0.337 -0.141 -0.012 -0.011 -0.010 0.013 -0.068 -2.152* 0.258 -1.429* -1.913* 0.585 1.542 0.863 -2.48 1* r-squared 0.42 0.43 0.12 0.10 notes: adivide coefficients by looo.*(**)significant at least at the 5 (io) percent level using a two-tailed test. 140 financial services review 5(2) 1996 ment fees and turnover costs are presumably paid to facilitate productive input from the manager. trueman (1988) suggests that turnover is a positive information signal. holm strom and ricart i costa (1986) and lambert (1986) suggest corporate managers try to sig nal their skill through the volume of capital investments. table 2 shows a strong positive relationship between alpha and management fee. the management fee coefficient is 5.092, indicating that a one basis point increase in manage ment fee increases fund alpha by about five basis points. the negative relationship between alpha and expense ratio indicates administration expenditure reduces alpha. each basis point increase in expenses leads to about a basis point (1.178) decrease in alpha. turnover is positively, but insignificantly, related to alpha. apparently, the positive information sig naling effect suggested by trueman (1988) is not strong enough to fully overwhelm the negative effect of trading costs. the standard human capital investment model, established by becker (1964), mincer (1973), and topel(1991), implies a positive relationship between measures of human cap ital such as tenure and education, and alpha. like years of education, mba should be pos itively related to alpha because specialized business education should lead to better performance. finally, if age largely measures stamina, then manager age and alpha should be negatively related. results show that education does not have the positive direct effect expected although the reduced form coefficient is positive. the positive mba coefficient is significant at the 10 percent level; the mba increases alpha by nearly one percentage point annually. when indirect effects are considered, the reduced form coefficient is smaller and statistically insignificant, indicating that the mba effect is somewhat weak. manager age is negatively related to alpha, supporting industry claims that younger managers cope more easily with the job’s demands. tenure and alpha are strongly posi tively related, indicating that experience pays and perhaps that poor performers are quickly eliminated. tenure is the strongest human capital measure; an additional year of tenure leads to a direct 0.165 increase in annual alpha. the full impact (measured by the structural coefficient) is 0.185, which means that tenured managers also keep costs or noise trading low, indirectly increasing alpha. team size has an indeterminate effect on alpha. perhaps funds with more than one manager may find two heads are better than one. alternatively, conflicts among managers may negatively impact alpha. while the team structural coefficient is negative, it is statis tically insignificant. do older funds produce better alphas? more established funds should be more experi enced at selecting better managers or keeping costs low. table 2 does not support this con tention. in fact, the fund age coefficient is negative, although statistically insignificant. many believe that as funds grow assets, performance suffers because larger assets reduce managers’ trading flexibility. nevertheless, grinblatt and titman (1989) after con trolling for expense and fee differences, find that assets and performance are unrelated. similarly, table 2 shows no significant relationship between assets and alpha. perhaps because some managers close funds to new investors when assets increase to a target (see mcgough, 1993b), few funds reach the point at which asset growth reduces performance. investors are often counseled that they will get better performance from a fund with low fees. the most important results from the alpha equation are that this statement is true for operating expenses but not management fees. indeed, management fees and perfor mance are strongly positively related. in addition, results show that investors should get mutual fund managers 141 better performance from well-diversified funds managed by younger, longer-tenured man agers with mba degrees. b. the risk equations the beta and residual standard deviation equations are: beta = (residual, turnover, manager age, tenure, mba, fund age, objectives) (3) residual = (beta, turnover, manager age, tenure, mba, team, fund age, assets, objectives) (4) many of the same independent variables are included in these two equations and are chosen for similar reasons. beta and residual risk may be positively related to one another because aggressive managers may try to reap high returns both by increasing beta and by concentrating their investments in fewer, well-researched stocks. therefore, residual risk appears in the beta regression and beta appears in the residual risk equation. table 3 shows that beta and residual return standard deviation are positively, although not statistically sig nificantly related. trueman (1988) suggests risk and turnover are likely to be positively related because high-risk stocks offer greater opportunity for gain (i.e., accurate information about their prospects is more valuable). as expected, table 3 shows turnover and both risk measures are positively related, but the turnover-residual standard deviation relationship is statisti cally insignificant. according to gibbons and murphy (1992) and fama (1980), manager age and risk should be positively related. poor performance hurts one’s reputation and reduces future job prospects and fees. younger managers with more time left in the labor market will want to avoid large negative outcomes more than managers approaching retirement. results show that manager age is positively but insignificantly related to risk except for the beta reduced form coefficient which is negative and insignificant. tenure should be negatively related to both risk measures if managers protect against losing a stable position by reducing risk. amihud and lev (198 1) use such agency argu ments to explain conglomerate mergers and amihud, kamin, and ronen (1983) show that manager-controlled (as opposed to owner-controlled) firms choose investment projects with less systematic and unsystematic risk. brown, harlow, and starks (1996) show that fund managers have compensation incentives to manipulate their risk levels. results show a negative, but statistically insignificant relationship. mba may be positively related to beta but not residual standard deviation because mbas are taught that only beta risk receives compensation in the market. hence, mbas are more likely than other managers to try to outperform the market index by increasing beta rather than residual risk. as predicted, mba and beta are significantly positively related although mba and residual standard deviation are not significantly related. team size has no clear impact on beta but one might expect that as the number of man agers grows, residual risk would fall because each individual in a team may wish to include his or her favorite stocks in the fund. results show a positive but insignificant relationship between residual risk and team size, however. 142 financial services review 5(2) 1996 table 3 beta and residual standard deviation 3sls structural and reduced form regressions beta regressions structural reduced stundard deviation regressions structural reduced endogenous variables intercept yield (%) alpha (%) beta residual st. dev. (8)” expense ratio (%) management fee (%) turnoverb (%) 0.780* 0.952* 4.832* 0.695 0.761 5.941* 0.759* 2.680 exogenous vuriubles managerb age (years) tenureb (years) years of education mba degree team size fzi ~~~~t~x~l)lions) load (8) 0.017 a.800 20.30 16.44 -2.878 -3.430** -19.80 -23.55 0.000 0.000 0.05 1* 0.044* 0.07 1 0.072 0.002 0.137 0.140 2.314* 1.684* -24.80* -25.13* -0.004 -0.299* 4.306* a.002 4koo6 fund objectives growth aggressive growth growth and income small stock special balanced asset allocation option equity international 0.087 -0.148* 0.099* -0.288* -0.377* -0.377% -0.402* -0.318* 0.164* 4.167* 0.130* x).229* x).392* 4).399* 4i410* -0.282* 3.634* -1.607* 3.259* 4.461* -2.255* -2.624* -0.713 6.200* 3.916* -1.740* 3.368* 4.393* -2.51 l* -2.901* -1.002 5.980* r-squared 0.35 0.34 0.58 0.57 nom: “@‘divide coefficients by 100 (1000). *(**) significant at least at the 5 (10) percent level using a two-tailed test funds that provide more systematic risk and less residual risk should earn larger aver age returns with relatively less noise trading, thereby improving their survival chances. consequently, fund age and beta (residual standard deviation) should be positively (nega tively) related. structural coefficients in table 3 show that fund age is positively related to beta and negatively related to residual standard deviation, as expected. this interpretation gains further support from the reduced form coefficients. the indirect effect of less noise trading is less turnover and less residual risk, both of which imply a smaller reduced form coefficient (1.684) than the structural coefficient (2.3 14) in the beta equation. in the resid ual standard deviation equation, the indirect positive effect of fund age on beta partly off sets the negative indirect effect of less turnover so that the reduced form coefficient is only a bit smaller (-25.13 vs. -24.80). fund asset size should have a negative effect on residual standard deviation because more assets require managers to invest in more companies. most funds have limitations on how much they can invest in any one stock. as predicted, results support a strong negative relationship between assets and residual risk. mutual fund managers 143 the coefficients on the objectives are all as one would expect. for example, aggres sive growth funds have relatively large betas and residual standard deviations while spe cialized funds which hold securities in one industry have relatively large residual standard deviations. overall, the most notable result from the beta equation is that larger portfolio turnover, fund age and a manager with an mba are associated with a larger fund beta. older, larger funds can be expected to deliver smaller residual risk. c. the fee equations fees are broken down into three main components and examined separately because earlier studies such as ippolito (1989) have done so. although each represents a cost to fund shareholders and are affected by some of the same variables, there are important dif ferences as well. the equations for expense ratio, management fee, and turnover are: expenses = (turnover, manager age, tenure, education, mba, team, fund age, assets, load, objective) management fees = (beta, residual, turnover, manager age, tenure, education, mba, team, fund age, assets, load, objective) (5) (6) turnover = (manager age, tenure, mba, team, fund age, assets, load, objective) (7) manager age, tenure, mba, team size, fund age, assets, load, and fund objective enter each equation. one would expect managers to improve their tenure chances and fund sur vival rates by reducing all types of expenses. mcgough (1993a) reports that older, tenured fund managers have reputations for keeping expenses low. similarly, mbas are often taught in investments classes that costs should be kept low to boost performance. on the other hand, human capital theory suggests that management fees should be positively related to age and tenure because age and tenure are measures of human capital. table 4 reports the regression results for expense ratio, management fee, and turnover. as expected, manager age is significantly negatively related to turnover but is not signifi cantly related to either expense ratio or management fee. also as expected, tenure is signif icantly negatively related to expense ratio but not significantly related to management fee or turnover. mba is negatively related to all three fee variables as predicted, although only the management fee coefficient is significant. this means that fund managers with mbas may accept lower compensation, perhaps reflecting strong competition among mbas for fund manager positions. the corresponding reduced form mba coefficient is larger than the 3sls structural coefficient due to the indirect impact on management fees through beta; that is, mbas choose larger portfolio betas on average and are compensated for bear ing the additional risk. a larger team size means more salaries, expenses and turnover; therefore, team should be positively related to all the fee variables. results show team size is significantly positively related to management fee as expected, but the other two coefficients are insignificant. older funds may purposely chose low fees as a means to survive. for example, the established vanguard funds have been successful by touting their low costs. indeed, table 144 financial services review s(2) 1996 table 4 expense ratio, management fee and portfolio turnover 3sls structural and reduced form regressions expense ratio management fee portfolio turnover regressions regressions regressions structural reduced strucrural reduced srructural reduced endogenous variables intercept yield (%) alpha (%) beta residual st. dev. (%)a expense ratio (%) management fee (%) turnoverb (%) 2.607* 2.600* -0.046 exogenous variables manager age (years) tenure (years) years of education mba degree team size fund age (years) fund asset? (millions) load (%) xwo4 -0.011** 4.081** a.054 a.018 -0.010* -0.108* 0.005 a.004 -0.011** -0.081** xl.054 a.018 -0.010* -0.108* 0.005 fund objectives growth aggressive growth growth and income small stock special balanced asset allocation option equity international 0.748* 0.027 -0.056 0.243* 0.09 1 -0.109 a.037 0.235* 0.746* 0.027 a.057 0.24 1 * 0.091 -0.110 0.043 0.236* r-squared 0.19 0.19 4.044 0.122** 2.765* 0.265 0.001 0.001 0.030* -0.062* 0.041* xnlo3* 4.034* -0.010* x).141* a.020 x).130* x).131* 0.005 0.135 0.122 -0.090 0.19 0.281 167.2* 167.2** 0.001 xwol 0.030* 4.057* 0.045* xnlo3* x).043* -0.010* -1.2428 a.493 -10.66 0.675 4).578* -1.760 -1.940 -1.242* -0.493 0.000 -10.67 0.675 x).578* -1.756 -1.940 0.004 4x091* x).019* a.028 a.1 lo* 0.007 0.044 0.038 62.79* ’ 62.80* -6.464 -6.464 6.925 6.925 33.84* 33.84* 6.311 6.311 0.178 0.178 -1.270 -1.271 -12.24 -12.24 0.21 0.11 0.11 nom: a cb)divide coefficients by 100 (iwj). *(**) significant at least at the 5 (io) percent level using a twwtailed test 4 shows that fund age is significantly negatively related to expense ratio, management fee, and turnover. scale economies should produce a negative relationship between the fees and assets. as expected, results show that assets are negatively related to all three fee types, although the turnover coefficient is statistically insignificant. this means that funds tend to charge smaller fees as they grow in size, spreading costs over more assets. because fund load is an extra marketing expense, load funds may have to keep other fees relatively low in order to compete. in addition, some no-load funds include 12b-1 mar keting fees in their expense ratio. thus, load funds may have smaller expense ratios by def inition since their largest marketing expense is broken out separately as a load. table 4 shows that load is negatively related to management fees and turnover, although only the management fee coefficient is significant. apparently, load funds trade off lower manage ment fees for up-front load fees. ma&ml fund managers 145 some of the fund objective coefficients are significant and have intuitively appealing signs. for example, aggressive growth funds’ annual turnover rate is 62.8 percentage points greater than that of ordinary growth funds. most fees for aggressive growth, special ized, and international funds tend to exceed those of growth funds. this is expected because these funds may require specialized management skills and more expensive administration. assuming that managers are risk averse, golec (1992) shows that management fees and risk should he positively related. because m~agement fees are a percent of assets, high-risk funds will have more volatile management fees. managers require greater aver age compensation in exchange for riskier fees. the management fee regression in table 4 supports this prediction; both beta and residual standard deviation coefficients are signifi cant and positive. turnover and expense ratio may be negatively related because competition between funds based on cost implies relatively high turnover costs must be offset by relatively low expenses. indeed, some funds avoid paying research costs by receiving their research from brokerage companies. they compensate brokers by directing more trades (“soft dollars”) to the brokers who supply research. results show a negative relationship, but the coeffi cient is insignificant. as noted above, turnover may signal management effort. assuming managers require compensation for this effort, turnover should be positively related to management fee. although the management fee regression shows turnover and management fees are posi tively related, the ~lationship is statistic~ly insigni~c~t. years of education should be positively related to management fee according to human capital theory, assuming that funds use education as a measure of human capital. indeed, table 4 shows that years of education and management fees are significantly pos itively related. in addition, assuming better educated managers can produce their own research or that they economize on other expenses, years of education and expense ratio may be negatively related. the significant negative coefficient for years of education in the expense ratio regression supports this claim. the most notable result from the fee regressions is that older and larger funds can be expected to deliver lower fees. in addition, older managers tend to trade less while long tenured managers tend to keep expenses low. v. conside~~on of survivo~~p bias and performance measurement performance evaluation of mutual fund managers is subject to a survivorship effect since very poor performers are likely to be fired and very good performers may leave volunt~ly for better opportunities. the survivorship effect may be relatively small in this study because of the short sample period. the relative numbers of good and poor performers who exit the industry along with the level of their performance will determine the net effect on the sample’s average alpha. either way, managers exiting the tails of the distribution will reduce the sample’s alpha variation and make it more dif~cult to find signi~cant structural relationships. indeed, the r-squared for the alpha equation is relatively low. this reduction of variance is of greater concern than the potential effect on average alpha because this 146 financial services review 32) 1996 study oniy requires alpha to rank performance of managers in a cross-section. by contrast, most other mutual fund performance studies are interested in using their performance mea sures as absolute measures of whether fund managers “beat the market” (i.e., whether the measure is positive). one drawback of this study is that the data source only provides alpha measured using the s&p 500. some recent studies have used alphas measured with multiple indexes, although ippolito (1989) and goetzmann and ibbotson (1994) use the s&p 500. many studies find that average fund performance changes with the index. average performance differences are less important to the cross-sectional analysis in this study because it relies on relative performance between managers. results may be affected if performance ranks are not stable over indexes and the wrong index is used. hendricks, patel, and zeckhauser (1993), who tried numerous single-index and multi index models, found little effect on rankings. on the other hand, grinblat and titman (1994) show that multiple-index characteristics-based models produce substantially differ ent performance rankings than single-index models, hence, the evidence on ranking stabil ity is mixed. this study’s results partly control for potentially m&specified alpha because the other components of the capm (beta and residual standard deviation) appear in the alpha equation. if the ranking is still improper, the results could be spurious, although it is also possible that improper ranking will produce noise and less significant results. vi. summary and conclusion this study analyzes mutual fund portfolio performance (yield and alpha), risk (beta and residual return standard deviation) and fees (expense ratio, management fees, and tum over) as endogenous variables in a system of simult~eous equations. earlier studies typi cally focus on only one or two of these variables using single equation methods. results of this study are summarized in light of their implications for investors choos ing among funds and fund managers. most investors are primarily concerned with the return they receive for bearing risk. one can expect better risk-adjusted performance (alpha) from a fund manager who is relatively young (less than 46 years old) yet has man aged a fund for a relatively long time (more than 7 years). results also show managers with mbas outperform those without. funds that keep administrative expenses low (less than 0.80 percent) produce better performance. but larger management fees (above 0.73 per cent) are associated with better performance, perhaps because larger fees are paid to better skilled managers. this means that investors should avoid funds with large operating expenses but not necessarily those with large management fees. results also show that investors should avoid funds whose portfolios contain much residual risk (more than 0.075 residual return standard deviation) because they tend to under-perform. a fund’s beta, turnover, team size, age and asset size as we11 as a manager’s years of education have no signi~c~t impact on risk-adjusted performance. investors seeking high yield, all else equal, should avoid funds with large fees, espe cially management fees, and choose larger (more than $280 million) funds managed by long-tenured managers. of course, such investors should also select funds with high-yield objectives, such as balanced funds. with regard to risk, investors should realize that by selecting high beta funds (greater than 0.84>, they often receive more residual risk and portfolio turnover as well. one way to mutual fund managers 147 limit this problem is to select managers with mbas because mbas provide relatively large betas without increasing residual risk significantly. another way to reduce the problem is to select older (older that 16.27 years) and larger funds because they tend to provide larger betas together with smaller residual risk, all else equal. managers apparently charge more to manage higher risk funds as compensation for more volatile fees. this result may have important implications for asset pricing since it may imply that investment managers require stock market compensation for holding portfolios with unsystematic risk. this could explain why levy (1978) and tinic and west (1986) find unsystematic risk priced in securities markets. managers of load funds apparently charge smaller management fees to partially compensate shareholders for load charges. load funds do not perform significantly better or worse than noload funds, however. strong competition among mbas for fund management jobs could explain why mbas charge smaller management fees even though they deliver larger alphas and betas and smaller expenses and turnover than other managers. apparently, successful funds rec ognize the bargain since even though fund managers with mbas are on average younger (45 vs. 48 years old) and less tenured (5.5 vs. 6.8 years), they manage larger ($302 vs. $228 million) and older (16.7 vs. 15.5 years old) funds. indeed, the competitive strength of mbas probably explains why they manage 64 percent of all funds. older and larger funds economize on expenses, management fees, and trading costs while keeping beta up and residual return variance down, all of which enhance fund sur vival and growth. finally, fund objective has a significant impact on many of the endoge nous variables, hence, controlling for fund type is important. acknowledgment: i would like to thank nick bramante for help gathering data and diane adams, maurry tamarkin, three referees and the editor for helpful comments. references amihud, yakov, & lev, baruch. (1981). risk reduction as a managerial motive for conglomerate mergers. the bell journal of economics, autumn, 60.5617. amihud, yakov, kamin, jacob y., & ronen, joshua. (1983). ‘managerialism,’ ‘ownerism’ and risk. journal of banking and finance, 7, 189-196. barry, christopher b., & starks, laura t. (1984). investment management and risk sharing with multiple managers. journal offinance, 39(june), 477-491. becker, gary s. (1964). human capital: a theoretical and empirical analysis, with special reference to education. new york: columbia university press (for nber). black, fisher. (1986). noise. journal of finance, 4i(july), 529-543. brown, keith c., harlow, w.v., & starks, laura t. (1996). of tournaments and temptations: an analysis of managerial incentives in the mutual fund industry. journal of finance, si(march), 85-l 10. brown, stephen j., goetzmann, william, ibbotson, roger g., & ross, stephen a. (1992). survivor ship bias in performance studies. review of financial studies, 5,553-580. cohen, susan i, & starks, laura t. (1988). estimation risk and contracts for portfolio managers. management science, 34(september), 1067-108 1. dechow, patricia m., & sloan, richard g. (1991). executive incentives and the horizon problem. journal of accounting and economics, 14,51-89. pii: s1057-0810(97)90018-4 financial services review, 6(3): 225-225 copyright 0 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. addendum grinder, b. (1997). an overview of financial services resources on the intemet. financial services review, 6, 125-140. the certified financial planner board of standards, inc. is the organization that awards the cfp and certified financial planner designation, not the institute of certified financial planners (icep). the cfp board is located on the internet at www.cfp board.org, for information on how to become certified, the list of registered educational institutions, disciplinary actions, practice standards developments and much more. this site also provides links to other financial services organizations, including educational institutions, membership organizations and regulators. pii: 1057-0810(95)90015-2 financial services review, 4( 1): 23-30 copyright 0 1995 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. quantifying time value errors george a. mangier0 susan m. mangier0 time valuation of cash flows is an essential part of personal financial planning and management. many financial arrangements are priced according to a cash-jlow valu ation model. expected cash flows associated with a stock or bond are discounted at an appropriate risk-adjusted rate in order to determine the fair value of the financial asset. home mortgage loans arepriced according to the discounted value of thefitureprincipal and interest cash flows. yet, despite the importance of the discounted cashflow method ology in pricing assets, computational errors are often made when discount factors are not calculatedprecisely. this article attempts to quantify the magnitude of the error when the mathematicalfirnction for present value is ignored and interpolation is used instead to determine the discount factor, i. introixjcti~n even though cash-flow valuation is an essential ingredient for financial planning and decision making, many borrowers are unfamiliar with time value mathematics. individual borrowers tend to rely on their bankers for accurate calculation of amortization schedules. however, bank errors do occur, especially with adjustable rate mortgages (arms), which require frequent recalculation. according to a study published in 1993, the federal reserve uncovered 881 faulty loans at six of 44 banks it scrutinized. an earlier government accounting office report quoted authorities as estimating that as many as 30% of arm adjustments are incorrect. adjustment errors result from three mistakes-using the wrong index to determine the reset rate, adjusting loans on the wrong day, and rounding off numbers incorrectly. the problem is so large that several banks have incurred penalties for their goofs and have had to settle with borrowers. moreover, some banks are being investigated by the federal trade commission for misleading practices in lending or sued for their malfeasance. for example, the use of an incorrect computation method on the part of a large california george a. mangiero l hagan school of business, iona college, 715 north avenue, new rochelle, new york 10801. susan m. mangier0 l sacred heart university, 515 1 park avenue, fairfield, ct 06432. 24 financial services review 4( 1) 1995 bank consistently led customers to think that borrowing rates were lower than they really were. class-action lawsuits against mortgage lenders are becoming more commonplace. when banks discover errors, they try to correct the problem with the borrower. in the case where the error leads to an underpayment of that period’s mortgage amount, the bank will attempt to extend the loan payback period or obtain a retroactive adjustment. an overpayment can be corrected by rebating the excess amount to the borrower. still, the customer has incurred opportunity costs by having paid more than was required. the arm market is large at over $14 billion. arm payment miscalculation is a serious issue. because of the prevalence of bank errors, consumer advocates are urging borrowers to verify the accuracy of their mortgage payments. borrowers have several choices. they can hire a financial advisor to audit the payment schedule. alternatively, they can render their own calculations to corroborate the payment schedule. aids such as the “arm check kit” from hsh associates provide instructions and tables for computing a mortgage amortization schedule. however, the verification process itself is fraught with difficulty. reliance on time value tables can lead to error for several reasons. first, table factors are often rounded to four decimal places, which in and of itself leads to inaccuracy. second, time value functions are not linear functions. this means that the approximation of a time value factor based on linearly interpolating between two time value factors when there is no table value available for the rate in question is incorrect. this article develops a formula to measure the error created when factors based on a linear interpolation are used in lieu of computing the actual factor value. the conditions under which the error is most acute are identified. specifically, the authors examine the sensitivity of error to levels of interest rates, the length of the time valuation period, and the difference between the two interest rates that are being used in the linear interpolation process. for ease of exposition, only the present value of a single cash-flow function assuming discrete compounding is discussed. however, the general idea of quantifying time valuation errors can easily be extended to problems involving either multiple cash flows, continuous compounding, and/or the future value function. to clearly see that the present value function is nonlinear, one need only plot the function as its inputs-time and periodic interest rate-change. figure 1 shows that for both 5% and lo%, the graph of the present value function for a single cash flow of $1.00 is curvilinear. a numerical example can also be used to emphasize the nonlinearity of the present value function. suppose that a financial planner wants to compute the present value of $500,000 for a client who is trying to assess how much must be deposited today in order to reach this objective in 20 years assuming an annual rate of return of 7.5%. the planner uses a discounting table with rate increments of 5%. using the linear interpolation method to compute the present value interest factor for 7.5%, the investor should make a deposit today in the amount of $131,375.’ the exact present value is $500,000 times (1 .075)-20, or $117.706.57. the magnitude of the error is $13,668.43, or 11.61% of the correct deposit amount. admittedly, this is an extreme example to illustrate the point inasmuch as table values are usually given in increments of 1% up to 10% and in increments of 2% for higher rates. 25 1o i / i , ) d / , i l : / i / 1 i i i i i_..l. i 2 3 4 5 6 7 8 9 1011121314151617181920 length of discounting period _.5% rate per period +_ 10% rate per period figure 1. present value of $1 .oo at 5% and lo%, respectively ii. a general model consider the general model in which the present value error is expressed as the linearly interpolated present value factor minus the actual present value factor. e={{[(y~-~*)(~~-rj1]x[(l +rt)-“-(1 +r2jn]) i-(1 +rjn]-[(i +r*)-n] (1) where: r* = the rate in question (not found in the table) pt = tength of the period rl = lower bound of the table rate interval r2 = upper bound of the table rate interval. redefining r2 as rl + .01 reflects the assumption of table factors being provided for 1% interest intervals. later, this int~va~ will be generalized and referred to as d. the distance between the rate in question and the lower boundary of the rate interval, r* rl, is redefined as a. equation (1) can thus he simplified as: e= ({[(.ol-@l.ol]~l(l +rjn-(1.01 +rt)-“]}+(l.ol +t~)p)-[(l +rt +oq-“].(2) note that when r* equals ri or r2, the only error present is that which results from using a rounded factor. no linear interpolation error exists when r* equals either the upper or lower boundary of a rate interval. to maximize the error e with respect to the distance between r* and rl, it is necessary to find the first derivative d(e)/& and set this to zero: 0= {-(l/.ol)x[(l -j-ii)-‘-(1.01 +r$“]) +[nx(l +rl +c$‘“+“] (3) letting k equal [ 100 x (1 + rl)-=] [i 00 x (1 .ol + z-t>-“], equation (4) gives a value for the distance between r* and r, which corresponds to the maximum error made when reiying on linear interpolation between table present value factors for a periodic interest rate, rl, and that same interest rate plus l%.* 26 financialservicesreview 4(l) 1995 a = (n/k)(l’“+‘) (1 + r,) (4) recall that the minimum interpolation error is zero, which occurs when r* equals either rl or r2. therefore, it is not necessary to take the second derivative of the error function to verify that cx is associated with the maximum error. iii. numericalverificationof a let rl equal .lo. thus, r2 = .l 1. when the time interval consists of two periods (n = 2), d(e)/da = -1.482 + 2(1 .lo + u)-~. setting the first derivative to zero, the result is that 1.482 = 2/( 1.10 + u)~. solving for a results in a value of .00498. this means that the maximum interpolation error occurs when r* is almost half way between rl and r2. consider the case where rt = 20 ceteris paribus. the variable k takes on a value of 2.460972090.3 using equation (4), a equals .004917 or (2012.4609 . . . )1’21 1.1. when ii is changed to 1 and no other variables change values, k equals .8 19.4 thus, a equals .0049886, or ( 1/.819)1’2 1 .l. when n is changed to 100, a still takes on a value of .00499. changing the numeric values of rl and r2 still results in an unchanged value of a. consider the case when r, equals so and r2 equals .5 1. for a one-period time interval, a still takes a value of .00499.5 importantly, however, it turns out that a does change value as the model inputs-r, and n-change. moreover, a is affected by relaxing the assumption that table values always represent periodic interest rates, which differ by 1%. the effects of these three model parameters on the magnitude of a are discussed in section iv. iv. sensitivityof a a is a function of three parameters-the percentage increment underlying table factors (d), the interval in which the target rate falls (rl), and the number of periods which define the investment horizon (n): c1 = rzll’(n+l)l (d-l x [( 1 + rjn (1 + rl + d)-“])[“(“+l)l (1 + r,) (5) note that when d equals 0.01, equation (5) becomes equation (4). for example, the a associated with finding the present value of funds for a 20-year period (n = 20) discounted at a rate of 8.25% (i.e., r* = 0.0825) will depend not only on ii equal to 20 and rl equal to 0.08 but also on d-that is, whether one is linearly interpolating between an 8% factor and a 9% factor versus interpolating between an 8% factor and a 10% factor. likewise, a will change if one lengthens or shortens the discounting period, n. using matlab, a program was written to compute a which is associated with the largest mathematical error as a function of d, ri, and n. the rl used to generate figures 2 and 3 equals 0%. in figure 2, however, d equals l%, while in figure 3, d equals 5%. both figures demonstrate that the error grows without bound as the n increases. the obvious ramification is that financial planners or lenders who erroneously use interpolated discount factors falling between zero percent and low interest rates are not only wrong in doing so but are especially wrong when working with long-term transactions. for rl greater than zero percent, the maximum interpolation error occurs sooner in the time period for larger rl values. figure 4 highlights this trend. when rl equals 5% and r2 quantifying time value errors 21 el.7 0.6 0.5 i 0.4 2 k 0.3 0.2 0.1 0 a 200 400 600 i300 1000 1 0.9 0.0 0.7 0.6 2 z 0.5 0.4 figure 2. error due to interpolation (rt=o%;d=l%) ............. ........ ...... ......... . ......... , ... ...... 200 400 600 000 1000 n figure 3. error due to interpolation (r-1 = 0%; d = 5%) 28 financial services review 4( 1) 1995 i 0.5 0 0 10 --_-;_i-i:::-il 20 3% 40 50 n figure 4. error due to interpolation (rj = 5%, lo%, 15%, 20%; d = 1%) equals 6%, the maximum interpolation error occurs when n equals 37. compare this to the timing of the maximum error which occurs when n equals 10 for an ri of 20% and an r2 of 2 1%. table 1 provides n values for various combinations of the lower and upper interest rate bound~ies. this same phenomenon is observed when the rate interval, d, is widened to 2%. recall that time value tables typically include factors in increments of one percent for periodic interest rates of 0% to 10% and use increments of two percentage points for rates greater than 10%. for example, to find an interpolated discount factor for a 22.25% rate, one would have to use the factors provided for 22% and 24%. figure 5 illustrates how the error function is maximized at smaller and smaller n values as the two boundary rates are increased. notice table 1 sensitivity of the timing of the maximum error to interest rates when d = 1% rl r2 maximum error 5% 6% 0.002 1864 10% 11% 0.0005842 15% 16% 0.000262 1 20% 21% 0.0001468 timing o~~~irnurn error cn) 37 20 13 10 quantifying time value errors 29 4 3 1 0 5 .5 -“’ .: ,. 4_. 3_ :. a. -el 5 10 15 25 n figure 5. error due to interpolation (rl = 22%, 24%, 26%, 28%; d = 2%) how the maximum point on each curve shifts to the left as rl increases. table 2 provides a summary of the rate combinations examined and the associated timing of the largest interpolation error. clearly, the interpolation methodology is fraught with problems. for a given lower rate boundary ri and a specified number of periods n, the interpolation error increases as the factor increment d widens. this conclusion is intuitive and implies that any individual using a factor table is best served by using as detailed a table as possible. for a given lower rate boundary and a specified distance between rl and r2, the interpolation error increases steadily, reaches a peak, and then decreases as n increases. for a given number of periods and a specified interest rate interval, the interpolation error decreases as the lower rate table 2 sensitivity of the timing of the maximum error to interest rates when d = 2% r1 22% 24% 26% 28% r2 24% 26% 28% 30% timing of maximum error maximum error in) o.ocw620 9 o.ocml3869 9 0.0003301 8 0.0002832 i 30 financial services review 4( 1) 1995 boundary increases. hence, in a low interest environment, there is more at stake when relying on interpolating from a time value table rather than computing the correct factors. clearly, verifying one’s arm amortization schedule is most accurate when precise time value mathematics is employed. aeknowi~gmen~: the authors wish to thank anonymous reviewers and the managing editor, barbara poole, for helpful comments. discussions with barbara, along with her research findings, enabled us to recognize the timeliness of our work and to identify the implications of our findings for the individual. notes 1. using a table, the 5% pv factor for 20 years rounded to four decimal places is .3769. the 10% factor for 20 years is .1486. the linearly interpolated factor for 7.5% is .3769 .5 (.3769 .1486), or .26275. multiplying .26275 by $500,000 results in a present value of $131,375. 2. equation (3) can be rewritten as equation (3’): [n i (1 + r, + a)““] = [-io0 i(1 .ol + r,)“] + i100 i(1 + r$] or [a i (1 + r, = #+tl = k. recipr~at~g each side and then multipiy~g both sides of the resulting equation by n gives (3”) (1 + r, + cqn+’ -n/k. taking the (n+l) root of each side results in(y) 1 +r,+a=(na)(“n+l). now it is possible to solve for cl. 3. the variable k equals 100 / (1.10)20 100 / (1.1 l)*‘, or 2.460972090. 4. k=loo/(l.lo)-loo/(l.ll),or.819. 5. k = 100/(1.50) 100/(1.51), or .4415011 io. references cissell, r., cissell, h., & flaspohler, d.c. (1990). muthemarics of jinance. 8th edition. boston: houghton mifflin. hamey, k. (1990). arm calculations should be checked. st. logs ~u~~-~~~~urc~, (october 2 i), 16. james, e.l. (1991). it pays to check changing arms. orla;ndo s’enrinel tribums, (february 2), 6 18. kobliner, b. (1993). avoid the pickpocket banks’ wrong arms and round hels. iwoney, (march), 434. misra, p. (1992). those hot-selling arm funds aren’t as safe as you think. money, (november), 39-40. pii: s1057-0810(99)80006-7 financial services review, 7(2): 107-128 copyright © 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. an analysis of personal financial literacy among college students haiyang chen and ronald p. volpe this study surveys 924 college students to examine their personal financial literacy; the relationship between the literacy and students' characteristics; and impact of the liter acy on students' opinions and decisions. results show that participants answer about 53% of questions correctly. non-business majors, women, students in the lower class ranks, under age 30, and with little work experience have lower levels of knowledge. less knowledgeable students tend to hom wrong opinions and make incorrect deci sions. it is concluded that college students are not knowledgeable about personal finance. the low level of knowledge will limit their ability to make informed decisions. i. ~ t r o d u c t i o n the ability to manage personal finances has become increasingly important in today's world. people must plan for long-term investments for their retirement and children's edu cation. they must also decide on short-term savings and borrowing for a vacation, a down payment for a house, a car loan, and other big-ticket items. additionally, they must manage their own medical and life insurance needs. unfortunately, studies have shown that americans have inadequate knowledge of per sonal finances (ebri, 1995; kpmg, 1995; psra, 1996, 1997; oppenheimer funds/girls inc., 1997; vanguard group/money magazine, 1997). they fail to make correct decisions because they have not received a sound personal finance education (hsr, 1993; hira, 1993; o'neill, 1993). this study has three purposes. first, it provides evidence of personal finance literacy among college students. second, it examines why some college students are relatively more knowledgeable than others. the analysis may help us identify factors that determine the level of competency possessed by college students. the third purpose is to examine how a student's knowledge influences his/her opinions and decisions on personal financial issues. haiyang chen• professor of finance, the williamson college of business administration, youngstown state university, youngstown, ohio 44555; phone: (330) 742-1883; fax: (330) 742 1459; e-mail: hychen@cc.ysu.edu. ronald p. volpe • professor of finance, the williamson college of business administration, youngstown state university, youngstown, ohio 44555. 108 financial services review 7(2) 1998 the paper is organized as follows. section ii reviews previous studies on financial lit eracy. section iii discusses methodology. section iv presents results. section v concludes the paper. ii. l i t e r a t u r e review most of the previous studies are conducted by practitioners in the financial service indus try. they focus on money management and investment-related issues. this emphasis is consistent with findings of the certified financial planners, indicating these issues are important areas of personal financial planning (nefe, 1993-1996). the results of these studies show that the participants generally answered fewer than 60% of survey questions correctly. prior studies of high school students consistently find that they are not receiving a good education in personal financial fundamentals and have poor knowledge (bakken, 1967; cfajamex, 1991; hsr, 1993; langrehr, 1979; naep, 1979). in a recent study of 1,509 high school seniors from 63 schools, mandell (1997) reports an average correct score of 57% in the areas of income, money management, savings and investment, and spending. his conclusion is that students are leaving schools without the ability to make critical deci sions affecting their lives. do adults have a good command of personal finance and investments? results of sev eral studies suggest that they do not. princeton survey research associates (1997) surveys 1,770 households nationwide on their financial knowledge and find an average correct score of 42%. this result shows that household financial decision makers do not have a good grasp of basic finance concepts. in another study of 522 adult women, 56% are found not very knowledgeable about investing (oppenheimer funds/girls inc., 1997). workers do not save adequately for retirement and make investment decisions that are too conservative. a kpmg (1995) survey of 1,183 employers finds employees contribute only about 5% of their income to 401k plans, although the typical plan allows a 14% con tribution. the evidence indicates that employees are not maximizing their benefits. addi tionally, the low savings rate and the low return from conservative investments may not provide enough income for a financially secure retirement. employee benefit research institute (1995) provides further evidence that most americans do not save sufficient retirement funds and may have a false sense of financial confidence and security. the study surveys 1,000 current workers and retirees on financial knowledge issues. about 71% of all workers and 81% of retirees score 60% or less. the institute of certified financial plan ners (1993) surveys 123 certified financial planner licensees and finds that financial illit eracy is a major problem when it comes to making individual financial decisions. poor knowledge of investment fundamentals is the most common problem encountered by their clients. the results of two national surveys suggest that investors do not have a solid knowl edge of investment issues. princeton survey research associates (1996) interviews 1,001 investors and finds that only 18% of them are financially literate. vanguard group/money magazine (1997) survey 1,467 mutual fund investors at 59 shopping malls across the coun try. the average correct score on a 20-question quiz is approximately 45%. financial literacy 109 most published studies focus on financial literacy among high school students and adults. few of them have examined college students except for danes and hira (1987) and volpe, chen, and pavlicko (1996). danes and hira (1987) survey 323 college students from iowa state university using a questionnaire covering knowledge of credit card, insur ance, personal loans, record keeping, and overall financial management. they find that the participants have a low level of knowledge regarding overall money management, credit cards, and insurance. they also find that males know more about insurance and personal loans, but females know more about issues covered in the section of overall financial man agement knowledge. married students generally are more knowledgeable about personal finance. volpe, chen, and pavlicko (1996) focus on knowledge of investment. they survey 454 students from a state university in the midwest and find that the average correct score of the participants is 44%, suggesting that they have inadequate knowledge. they also find that male students are more knowledgeable than female students, and business majors are more knowledgeable than non-business majors. while the prior research has provided evidence of people's personal finance knowl edge and improved our understanding of the issue, it suffers from several weaknesses. for example, both studies on college students use samples from a single university. many stud ies cover selected areas in personal finances, neglecting others. furthermore, the validity of the survey instruments is questionable because of the limited number of items included in the questionnaires. these limitations are compounded by the fact that many prior studies only report the levels of financial literacy without analyzing the factors that influence peo ple's knowledge. none of the previous studies have examined how an individual's knowl edge impacts their opinions regarding personal finance issues and financial decision making. iii. m e t h o d o l o g y this study uses a comprehensive questionnaire designed to cover major aspects of personal finance. it includes financial literacy on general knowledge, savings and borrowing, insur ance, and investments. the survey participants are asked to answer 52 questions including 36 multiple-choice questions of their knowledge on personal finance, eight questions of their opinions and decisions, and eight questions on demographic data. the survey is used in a pilot study to refine the instrument. the validity and clarity of the survey are further evaluated by two individuals who are knowledgeable in personal finance. the quality and consistency of the survey are further assessed using cronbach's alpha. a copy of the ques tionnaire can be found in the appendix. the responses from each participant are used to calculate the mean percentage of cor rect scores for each question, section, and the entire survey. consistent with the existing lit erature (danes & hira, 1987; volpe, chen, & pavlicko, 1996), the mean percentage of correct scores is grouped into (1) more than 80%, (2) 60% to 79%, and (3) below 60%. the first category represents a relatively high level of knowledge. the second category repre sents a medium level of knowledge. the third category represents a relatively low level of knowledge. previous research suggests that levels of financial literacy vary among subgroups of students (volpe, chen, & pavlicko, 1996). this study provides further evidence of the dif 110 financial services review 7(2) 1998 ferences using analysis of variance (anova). the differences are further analyzed using logistic regression models. the participants are classified into two subgroups using the median percentage of correct answers of the sample. students with scores higher than the sample median are classified as those with relatively more knowledge. students with scores equal to or below the median are classified as students with relatively less knowl edge. this dichotomous variable is then used in the logistic regression as the dependent variable, which is explained simultaneously by all of the independent variables. the independent variables used in the logistic regression are variables such as aca demic discipline, class rank, gender, race, nationality, work experience, age, and income. the coefficients represent the effect of each subgroup compared with a reference group, which is arbitrarily selected. for example, major is coded as 1 if a participant is a non business major, 0 otherwise. the reference category is a business major. if the logistic coefficient of the variable is negative, then it means that compared with business majors, the non-business majors are associated with decreased log odds ratio of being more knowl edgeable about personal finance. the logistic model takes on the following form: log [ p / ( 1 -p)] = b 0 + bi(major ) + b2(classrank1 ) + b3(classrank2 ) + b4(classrank3) + bs(classrank4) + b6(gender) + bt(race1) + bs(race2) + b9(race3) + blo(race4) + b 11(nationality) + bi2(experience1) + bia(experience2) + bi4(experience3) + b15(experience4) + bi6(age1) + bi7(age2) + b18(age3) + b19(income1) + b2o(income2) + b21 (income3) + e i (1) where p major = classrank1 = classrank2 = classrank3 = classrank4 = gender = race1 = race2 = race3 = race4 = nationality = experience 1 = experience2 = experience3 = experience4 = age1 = age2 = the probability of a student who is more knowledgeable about personal finance. 1 if a participant is a non-business major, 0 otherwise. 1 if a participant is a freshman, 0 otherwise. 1 if a participant is a sophomore, 0 otherwise. 1 if a participant is a junior, 0 otherwise. 1 if a participant is a senior, 0 otherwise. 1 if the participant is a male, 0 otherwise. 1 if a participant is white, 0 otherwise. 1 if a participant is african american, 0 otherwise. 1 if a participant is hispanic, 0 otherwise. 1 if a participant is american indian, 0 otherwise. 1 if the participant is a foreign student, 0 otherwise. 1 if a participant has no experience, 0 otherwise. 1 if a participant has more than 0 to less than 2 years of experi ence, 0 otherwise. 1 if a participant has 2 to less than 4 years of experience, 0 other wise. 1 if a participant has 4 to less than 6 years of experience, 0 other wise. 1 if a participant is in the age group of 18-22, 0 otherwise. 1 if a participant is in the age group of 23-29, 0 otherwise. financial literacy 111 a g e 3 i n c o m e 1 i n c o m e 2 i n c o m e 3 = l i f a pa r t i c ipan t is in the age g r o u p o f 30-39 , 0 o the rwise . = 1 i f the pa r t i c ipan t is in the i n c o m e g roup o f less t han $10 ,000 , 0 o the rwise . = 1 i f the pa r t i c ipan t is in the i n c o m e g roup o f $10 ,000 -$29 ,999 , 0 o the rwise . = 1 i f the pa r t i c ipan t is in the i n c o m e g roup o f $30 ,000 -$49 ,000 , 0 o the rwise . t a b l e 1 cha rac t e r i s t i c s o f the s a m p l e number of participants percentage a. education 1. academic disciplines a) business majors 431 52.6 b) non-business majors 389 47.4 2. class rank a) freshman 156 17.2 b) sophomore 157 17.3 c) junior 160 17.7 d) senior 326 36.0 e) graduate 106 l 1.7 b. demographic characteristics i. gender a) male 395 44.4 b) female 495 55.6 race a) white 763 85.0 b) african-american 59 6.6 c) asian 47 5.2 d) hispanic 14 1.6 e) native american 15 1.7 3. nationality a) usa 740 93.4 b) foreign (other than usa) 52 6.6 experience 2. years of work experience a) none 32 3.9 b) less than two years 78 9.5 c) two to less than four years 134 16.3 d) four to less than six years 194 23.6 e) six years or more 384 46.7 2. years of age a) 18 to 22 395 43.7 b) 23 to 29 289 32.0 c) 30 to 39 151 16.7 d) 40 and over 69 7.6 income i. last year's income a) under $10,000 184 21.7 b) $10,000 to $29,999 224 26.4 c) $30,000 to $49,999 192 22.6 d) $50,000 or more 248 29.2 c. d. 112 financial services review 7(2) 1998 to determine the impact of financial literacy possessed by the participants on their opinions, students are asked to rank personal finance issues using five categories: very important, somewhat important, not sure, somewhat unimportant, and very unimportant. they are also asked to make decisions on the related financial issues. as in the logistic regression analysis, the sample is partitioned into two groups of students with relatively more knowledge and those with relatively less knowledge. since the issues are related to each section in the survey, the section median percentage of correct answers is used to clas sify the sample. cross-tabulations and chi-square tests are used to determine if the differ ence of the two groups' opinions and decisions are statistically significant. iv. results and analysis the questionnaires are sent to 1,800 students from 14 college campuses. they include both public and private schools, main and branch campuses of large universities, and small com munity colleges in california, florida, kentucky, massachusetts, ohio, and pennsylvania. nine hundred twenty-four students from 13 campuses participated in the survey, represent ing a response rate of 51.33 %. detailed characteristics of the sample are presented in table 1. in terms of education, about 52.6% of the participants are business majors. thirty-six percent of the participants are seniors with the rest evenly distributed among freshman, sophomore, junior, and graduate students. in terms of demographic background, most of the participants are white and u.s. citizens. female participants represent about 55.6% of the sample. most participants have more than two years of work experience. about 75.7% of the students are from 18 to 29 years of age. missing responses cause the sample size to vary from 792 to 905; therefore, various sample sizes have been used to calculate valid per centages in table 1. a. overall results of the survey the overall results are presented in table 2. the mean percentage of correct scores is grouped into three categories: over 80, 60-79, and below 60. the highest score is presented first, which is followed by lower scores within each section. the overall mean percentage of correct scores is 52.87%, indicating on average the participants answered only about half of the survey questions correctly. the median percentage of correct scores is 55.56%. the reliability of the 36-question survey is 0.85. the large cronbach alpha indicates that the questionnaire is reliable, which further increases its validity. the findings suggest that college students' knowledge on personal finance is inadequate. one reason for the low level of knowledge is the systematic lack of a sound personal finance education in college curricula. most of the higher education institutions put little emphasis on students' personal finance education (danes & hira, 1987). even business schools do not require students to take a personal finance management course (bialasze wski, pencek, & zietlow, 1993). according to a survey by gitman and bacon (1985), only 5% of business school offers an undergraduate major in finance services. given the lack of personal finance education, it is not surprising the results show that college students have inadequate knowledge on personal finance. financial literacy 113 table 2 mean percentage of correct responses to each survey question, each section, and the entire survey level of personal finance knowledge low medium high below 60% 60-79% over 80% i. general knowledge personal finance literacy legal requirements for apartment lease apartment leasing costs asset liquidity spending vs. saving pattern checking account reconciliation net worth calculation 56.49 personal financial planning 52.38 tax credit vs. tax deduction 27.38 mean correct responses for the section median correct responses for the section ii. savings and borrowing creditworthiness consumer credit report sources deposit insurance checking account overdrafts compound interest 56.39 certificate of deposit terms 50.32 loan co-sign consequences 44.70 annual percentage rate 33.23 credit card use 23.81 mean correct responses for the section 54.47 median correct responses for the section 55.56 iii. insurance auto insurance rate determination reason to buy insurance health insurance characteristics insurance conflict resolution 48.70 homeowners' insurance characterb ~ics 48.8 ! term insurance characteristics 32.14 mean correct responses for the section 59.24 median correct responses for the section iv. investments mutual fund selection common stock investing for selected investment goals retirement benefits of early investment 53.68 mutual fund investment return 47.08 high risk return investment suitability 45.35 interest rate changes and treasury bond price 36.90 municipal bond investment 34.31 dollar-cost-averaging 33.23 investment diversification 30.1)9 mutual fund charges 29.00 foreign exchange rates 28.57 mutual fund ownership characteristics 12.45 mean correct responses for the section 40.37 median correct responses for the section 41.67 mean correct responses for the entire survey 52.87 median correct responses for the entire survey 55.56 75.11 74.03 73.48 70.89 62.55 63.70 66.67 76.95 72.08 69.16 63.64 74.35 64.94 66.67 64.94 64.50 80.95 86.47 114 financial services review 7(2) 1998 another reason for the low level of knowledge can be attributed to the young ages of the participants. as shown in table 1, about 44% of the participants are 18 to 22 years of age, and about 76% are under 30. the majority of them are in a very early stage of their financial life cycle. at this stage of the cycle, they are exposed to a limited number of financial issues related to general knowledge, savings and borrowing, and insurance. dur ing this period, most of their incomes are spent on consumption rather than investment. these factors may explain the differences in the mean percentages of correct answers for the sections of general knowledge (63.70%), savings and borrowing (54.47%), insurance (59.24%), and investment (40.37%). a further look into the scores on individual questions shows that students score higher on issues with which they are familiar. for example, the highest score is related to auto insurance. students are familiar with the issue because many of them own cars and have to pay a higher auto insurance premium. students also score rel atively high on apartment leases. they know more about these issues because they need to rent apartments during their college years. in contrast, students have little experience with tax, term life insurance, and most of investment topics. subsequently, they earn low scores in these areas. b. analysis of results by subgroups of the sample in this section, the relationship between personal financial literacy and participants' education, work experience, income and other demographic background are examined. table 3 shows the mean percentage of correct responses for section i (general knowl edge), section ii (savings and borrowing), section iii (insurance), section iv (invest ment), and the entire survey by different subgroups. anova has been used to detect if participants from various subgroups have different levels of knowledge. participants' educational background has a significant impact on their knowledge. the results for the entire survey clearly show that business majors are more knowledgeable than non-business majors. on average, the business majors answered 60.72% of the survey questions correctly; the non-business majors, 49.94%. this pattern of business majors answering about 8% to 12% more questions correctly than non-business majors is persis tent throughout the individual sections. the testing results of anova indicate that the dif ferences are statistically significant at the 0.01 level. the findings also suggest that participants from different class ranks have different levels of financial knowledge. generally, graduate students know more than the under graduate students, and junior and senior students are more knowledgeable than those from the lower ranks. again, the differences in the level of literacy among different ranks are statistically significant at the 0.01 level. table 3 shows participants' knowledge varies with their demographic characteristics. the percentages of correct answers from the female participants (50.77%) are lower than those from the male participants (57.40%). this pattern persists among all sections includ ing the overall results. the values of f-statistic suggest that these differences are highly significant. participants from dissimilar ethnic backgrounds have different levels of finan cial knowledge. although the different scores are statistically significant, no single sub group can claim the highest scores throughout the four sections. african-american participants earn the lowest scores throughout the various sections. foreign students also earn lower scores than their american counterparts. financial literacy 115 table 3 mean percentage of correct responses to each section by characteristics of sample and results of anova general savings & for the knowledge borrowing insurance investments sample a. education 1. academic disciplines a) business majors 72.85 61.23 66.47 48.36 60.72 b) non-business majors 59.98 53.47 57.97 35.73 49.94 fstatistic (87.34)** (30.11)** (29.95)** (94.49)** (103.66)** 2. class rank a) freshman 56.48 50.50 54.06 31.25 46.17 b) sophomore 63.69 51.66 59.02 36.78 50.93 c) junior 67.22 58.68 63.75 41.98 56.09 d) senior 63.02 53.00 58.64 42.20 52.85 e) graduate 75.37 66.25 67.71 55.35 65.12 f statistic (12.48)** (10.58)** (6.34)** (27.84)** (20.60)** b. demographic characteristics 1. gender a) male 67.68 59.13 63.16 45.51 57.40 b) female 62.04 52.01 57.88 37.85 50.77 f statistic (13.96)** (21.48)** (10.68)** (33.93)** (31.01)** 2. race a) white 64.89 55.60 61.25 41.73 54.24 b) african-american 56.69 44.82 46.61 31.78 43.74 c) asian 61.94 56.73 56.38 42.38 53.19 d) hispanic 73.81 63.49 57.14 40.48 57.34 e) native american 60.74 46.67 60.00 36.67 49.07 f statistic (2.64)** (4.05)** (5.40)** (3.64)** (5.05)** 3. nationality a) usa 67.54 58.71 63.67 43.05 56.52 b) foreign (not usa) 59.62 48.72 50.32 38.62 48.34 f statistic (7.26)** (i 1.92)** (17.77)** (2.53) (13.22)** c. experience 1. years of work experience a) none 57.97 42.01 45.83 29.69 42.53 b) less than two years 57.55 51.14 52.99 36.65 48.22 c) two to less than four years 58.21 51.82 55.97 34.36 48.42 d) four to less than six years 65.12 54.30 61.25 39.18 53.12 e) six yearsor more 73.09 63.72 68.66 48.78 61.91 f statistic (22.32)** (20.79)** (19.42)** (24.09)** (37.78)** 2. years of age a) 18 to 22 60.28 52.32 56.79 36.37 49.74 b) 23 to 29 68.78 59.63 63.68 44.90 57.63 c) 30 to 39 61.07 49.08 55.74 39.96 50.15 d) 40 and over 74.07 62.32 71.74 5 i.45 63.20 f statistic (13.70)** (11.57)** (11.43)** (17.98)** (20.16)** d. income 1. last year's income a) under $10,000 62.44 51.39 60.78 39.13 51.63 b) $10,000to $29,999 67.91 58.58 62.20 41.48 55.82 c) $30,000 to $49,999 65.45 58.62 61.11 42.97 55.53 d) $50,000 or more 68.46 59.50 63.71 44.40 57.41 f statistic (3.56)* (6.58)** (0.74) (2.77)* (4.69)** notes: *significant at the 0.05 level; **significant at the 0.01 level or greater. 116 financial services review 7(2) 1998 in terms of participants' work experience and ages, participants with more years of work experience are more knowledgeable than those with less experience. participants in the age subgroups of 23 to 29 and 40 or older exhibit greater knowledge than the other age groups. finally, it seems that participants with higher personal income answered more questions correctly than those with lower income. the survey has an age category of "60 or older" and an income category of "no income." however, the number of participants in these categories are very small. they are regrouped into the adjacent groups of "40 and over" and "under $10,000." results of the logistic regression are shown in table 4. as suggested by the high chi square values, the models have high explanatory power. another widely used measure of the overall fit of the model is to examine its ability to correctly classify observations. for the entire sample, 71.47% of the observations are correctly classified as compared with 50.03% chance classification. similar patterns can be found in the individual sections. in addition to the overall fit of the model, the coefficient of major for the entire sam ple is negative and significant at the 0.01 level. consistent with findings of anova, the result suggests that non-business majors are more likely to be less knowledgeable about personal finance than business majors. the significant negative coefficients for class rank variables indicate that participants from lower class ranks are more likely to be less knowledgeable than those from graduate classes. women participants are more likely to be less knowledgeable than men. participants with less work experience have high probability of being less knowledgeable than those with more experience. the differences between those with six or more years of experience and those with no experience and those with less than two years of experience are significant at the 0.07 and 0.06 levels respectively. partic ipants under age 30 are more likely to be less knowledgeable as compared with those 40 or older. although the coefficient of age3 still exhibits a negative sign, the difference between those who are in their thirties and forties or older is statistically insignificant. while race, nationality, and income variables affect level of knowledge in one way anova, they no longer have any significant impact in the logistic regression where all the variables are used simultaneously to explain the level of knowledge. with few exceptions, the results from logistic regressions for the individual sections are consistent with that of the entire sample. for example, the business majors perform consistently better than the non-business majors in every section of the survey. similarly, many coefficients for classrank, gender, and age variables in the individual sec tions carry the same signs and are significant as shown for the entire sample. few coeffi cients of race, nationality, and income are significant in the individual sections. the result that business majors are more knowledgeable is consistent with findings of previous research. the finding is not surprising because curriculum requirements of busi ness majors give them more opportunity to take finance and related courses. participants who are more senior in class rank have earned higher scores in the survey. one explanation is that by staying in universities longer, students will naturally pick up more about personal finance. our argument is that students do not gain more knowledge of personal finance by just spending more time in college learning other unrelated subjects. they learn the subject through a business course, seminars, or their own mistakes. our view is consistent with the finding of this study that business majors are more knowledgeable than non-business majors. a similar line of reasoning would apply to why the participants who are older or have more work experience earned high scores in the survey. they must have prior expo sure to personal finance. they are not more literate just because they are older. t a b l e 4 l og is tic r eg re ss io n a na ly si s of th e im pa ct o f p ar tic ip an ts ' e du ca tio n, e xp er ie nc e, d em og ra ph ic c ha ra ct er is tic s, an d in co m e on t he ir f in an ci al l ite ra cy m a jo r c l a s s r a n k 1 c l a s s r a n k 2 c l a s s r a n k 3 c l a s s r a n k 4 g e n d e r r a c e i r a c e 2 r a c e 3 r a c e 4 n a t io n a l it y e x p e r ie n c e 1 e x p e r ie n c e 2 e x p e r ie n c e 3 e x p e r ie n c e 4 a g e i a g e 2 a g e 3 in c o m e i in c o m e 2 in c o m e 3 e st im at ed c oe ffi ci en ts a nd th e le ve l o f s ig ni fic an ce f or v ar io us s ec tio ns a nd th e e nt ir e sa m pl e g en er al sa vi ng s & f or th e k no w le dg e b or ro w in g in su ra nc e in ve st m en ts sa m pl e -0 .9 91 1" * -0 .3 49 5* -0 .5 37 9* * -1 .1 7 5 3 "* -0 .8 96 5* * 1 .0 57 3 * * 1 .1 49 1 * * -0 .7 04 4 -2 .3 63 0* * 1 .8 61 3 * * -0 .9 17 3* * 1 .3 25 1 * * -0 .5 97 2 -2 .0 38 1 * * 1 .6 07 1 ** -0 .6 98 5* 1 .2 23 4* * -0 .5 33 7 1 .5 01 2* * 1 .1 10 7* * -0 .4 14 5 -0 .7 24 8* -0 .4 05 0 -0 .9 29 8* * -0 .6 69 7* 0. 16 13 0. 41 33 ** -0 .0 14 9 0. 68 62 ** 0. 63 31 * * 0. 85 46 -0 .2 74 0 0. 94 28 0. 44 97 0. 46 67 0. 52 09 1 .1 77 2* 0 .1 6 7 0 -0 .2 98 5 -0 .1 84 6 1. 93 26 * 1. 16 28 0. 99 68 0 .1 4 8 0 0. 71 66 0. 94 02 -0 .5 61 5 1. 67 78 0. 79 25 1. 34 56 -0 .0 99 2 1 .2 71 7 * * -0 .7 85 4 -0 .6 12 3 -0 .7 15 6 -0 .3 77 1 -0 .8 23 4 1 .2 64 4 -0 .9 24 8 1 .0 62 0 -0 .5 66 1 -0 .2 87 3 -0 .2 13 6 0. 15 15 -0 .6 61 1 -0 .8 08 3* * -0 .6 37 1 ** -0 .3 03 7 -0 .5 26 3 -0 .9 34 3* * -0 .1 62 3 -0 .5 57 4* * -0 .2 01 3 -0 .4 70 9* -0 .6 30 6* * -0 .9 84 4* * -1 .0 8 9 9 '* -1 .5 5 1 6 "* -1 .3 6 3 1 '* -1 .4 86 7" * -0 .4 77 0 -0 .7 99 5* * -1 .0 78 4" * -1 .2 4 1 2 "* -1 .1 91 4" * -0 .2 51 7 -0 .4 05 0* -0 .6 33 9 -0 .7 82 3* -0 .6 40 7 -0 .3 59 4 -0 .2 28 0 0 .3 8 8 6 0. 33 82 -0 .0 45 9 -0 .2 29 1 0. 36 23 -0 .1 87 7 0 .0 3 2 0 -0 .1 90 5 -0 .2 64 4 0. 18 93 -0 .0 66 8 -0 .0 13 5 -0 .1 66 4 c on st an t 0. 97 73 2 .2 1 3 8 "* 0. 30 96 2 .1 1 4 8 "* 2. 48 88 ** -2 l og l ik el ih oo d 89 9. 82 6 92 4. 75 3 85 8. 45 1 83 3. 36 5 84 2. 95 5 o ve ra ll c hi -s qu ar e (1 50 .4 85 )* * (1 34 .0 40 )* * (9 5. 63 0) ** (2 10 .4 46 )* * (2 15 .6 50 )* * a dj us te d r 2 0. 23 9 0. 21 5 0. 16 5 0. 32 3 0. 32 8 c or re ct c la ss if ic at io n 70 .2 9% 6 6 .8 8 % 72 .5 1% 7 2 .6 4 % 71 .4 7% c h an ce c la ss if ic at io n 5 0 .7 4 % 5 0 .0 2 % 56 .7 2% 5 1 .0 0 % 50 .0 3% n ot es : *s ig ni fi ca nt a t t he 0 .0 5 le ve l; ** si gn if ic an t a t t he 0 .0 1 le ve l o r g re at er . -" t a b l e 5 im pa ct o f t he p ar tic ip an ts ' fi na nc ia l k no w le dg e on t he ir o pi ni on a nd b eh av io r o o a . g en er al k no w le dg e 1. n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho v ie w m ai nt ai ni ng a de qu at e fi na nc ia l r ec or ds a s: v er y so m ew ha t so m ew ha t v er y im po rt an t im po rt an t n ot s ur e u ni m po rt an t u ni m po rt an t to ta l b . 2. st ud en ts w it h m or e pe rs on al 26 0 93 8 9 4 37 4 fi na nc e k no w le dg e 69 .5 % 24 .9 % 2. 1% 2. 4% 1. 1% 10 0% st ud en ts w it h l es s pe rs on al 36 3 12 5 30 8 4 53 0 fi na nc e k no w le dg e 68 .5 % 23 .6 % 5. 7% 1. 5% 0. 8% 10 0% c hi -s qu ar e = 7. 83 5, s ig ni fi ca nt a t t he 0 .0 98 l ev el . n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho a ct ua ll y m ai nt ai n fi na nc ia l r ec or ds : d et ai le d m in im um r ec or ds r ec or ds n o r ec or ds to ta l st ud en ts w it h m or e pe rs on al 16 9 18 2 20 37 1 fi na nc e k no w le dg e 45 .6 % 49 .1 % 5. 4% 10 0% st ud en ts w it h l es s pe rs on al 14 9 23 7 12 8 54 i fi na nc e k no w le dg e 29 .0 % 46 .1 % 24 .9 % 10 0% c hi -s qu ar e = 6 5. 90 3, s ig ni fic an t at th e 0. 00 1 le ve l. sa vi ng s an d b or ro w in g 1. n um be r an d pe rc en ta ge o f p ar ti ci pa nt s w ho v ie w th at s pe nd in g le ss t ha n th ei r in co m e is : v er y so m ew ha t so m ew ha t v er y im po rt an t im po rt an t n ot s ur e u ni m po rt an t u ni m po rt an t to ta l st ud en ts w it h m or e pe rs on al 33 7 53 11 2 3 40 6 fi na nc e k no w le dg e 83 .0 % 13 .1 % 2. 7% 0. 5% 0. 7% 10 0% st ud en ts w it h l es s pe rs on al 33 6 68 39 43 11 49 7 fi na nc e k no w le dg e 67 .6 % 13 .7 % 7. 8% 8. 7% 2. 2% 10 0% c hi -s qu ar e = 50 .8 13 , si gn if ic an t a t t he 0 .0 01 l ev el . z > z < n t~ < c . 2. n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho c ho os e to a ct o n th ei r sp en di ng : c or re ct ly in co rr ec tl y to ta l s tu de nt s w it h m or e pe rs on al 36 5 45 41 0 f in an ce k no w le dg e 89 .0 % 11 .0 % 10 0% s tu de nt s w it h l es s pe rs on al 35 1 16 3 51 4 f in an ce k no w le dg e 68 .3 % 31 .7 % 10 0% c hi -s qu ar e = 56 .2 23 , si gn if ic an t a t th e 0. 00 1 le ve l. in su ra nc e 1. n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho v ie w th at m ai nt ai ni ng a de qu at e in su ra nc e co ve ra ge i s: v er y so m ew ha t so m ew ha t v er y im po rt an t im po rt an t n ot s ur e u ni m po rt an t u ni m po rt an t s tu de nt s w it h m or e pe rs on al 14 5 83 29 4 2 f in an ce k no w le dg e 55 . 1 % 31 .6 % i 1 .0 % 1. 5% 0. 8% s tu de nt s w it h l es s pe rs on al 35 4 19 3 65 22 7 f in an ce k no w le dg e 55 .2 % 30 .1 % i 0 . 1 % 3. 4% 1. 1% c hi -s qu ar e - 2 .8 44 , s ig ni fi ca nt a t t he 0 .5 84 l ev el . 2. n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho c ho os e to a ct o n th ei r in su ra nc e: c or re ct ly in co rr ec tl y to ta l s tu de nt s w it h m or e pe rs on al 68 19 7 26 5 f in an ce k no w le dg e 25 .7 % 74 .3 % 10 0% s tu de nt s w it h l es s pe rs on al 11 5 54 4 65 9 f in an ce k no w le dg e 17 .5 % 82 .5 % 10 0% c hi -s qu ar e = 8. 02 0, s ig ni fi ca nt a t th e 0. 00 5 le ve l. to ta l 26 3 10 0% 64 1 10 0% (c on tin ue d) t a b l e 5 ( co nt .) d . in ve st m en ts 1. n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho v ie w p la nn in g an d im pl em en ti ng a r eg ul ar i nv es tm en t pr og ra m i s: v e ry so m ew ha t so m ew ha t v er y bn po rt an t im p o rt a n t n ot s u re u n im p o rt a n t u ni m po rt an t st ud en ts w it h m or e pe rs on al 22 0 10 0 24 10 4 fi na nc e k no w le dg e 61 .5 % 27 .9 % 6. 7% 2. 8% 1. 1 % st ud en ts w it h l es s pe rs on al 22 2 17 1 11 5 22 14 fi na nc e k no w le dg e 40 .8 % 31 .4 % 21 .1 % 4. 0% 2. 6% c hi -s qu ar e = 52 .1 02 , si gn if ic an t a t t he 0 .0 01 l ev el . 2. n um be r an d pe rc en ta ge o f pa rt ic ip an ts w ho c ho os e to a ct o n th ei r in ve st m en ts : c o rr ec tl y in co rr ec tl y t ot al st ud en ts w it h m or e pe rs on al 28 8 72 36 0 fi na nc e k no w le dg e 80 % 20 % 10 0% st ud en ts w it h l es s pe rs on al 29 1 27 3 56 4 fi na nc e k no w le dg e 51 .6 % 48 .4 % 10 0% c hi -s qu ar e = 74 .7 74 , si gn if ic an t a t th e 0. 00 1 le ve l. t ot al 35 8 10 0% 54 4 10 0% z z t" t~ < < ,,o 7~ o¢ financial literacy 121 the finding that women score lower than men is consistent with the existing literature (genasci, 1995; goldsmith & goldsmith, 1997; hsr, 1993; lewin, 1995; martinez, 1994; volpe, chen, & pavlicko, 1996; ). space limitations do not allow a comprehensive analysis of why women are less knowledgeable than men. yet given the fact that more and more women are joining the work force and they are expected to live longer than men, deficiency in their knowledge about personal finance needs to be addressed. c. consequences of having inadequate knowledge this section examines how a student's knowledge affects his/her opinions and deci sions about some personal finance issues. the sample is partitioned into two groups by each section's median score. students with section scores higher than the median are clas sified as those with relatively more knowledge. students with scores equal to or below the median are classified as those with relatively less knowledge. since many students' scores are equal to the median scores, the classification scheme changes the number of observa tions in the two groups from section to section. in addition, missing observations cause the total sample size to vary from 903 to 924. participants' responses to the importance of keeping financial records are reported in part 1 of section a of table 5 and what they actually decide to do in everyday life in part 2. about 95% of the participants from the more knowledgeable group rank keeping records as very important or somewhat important, and the rest of them believe otherwise. for the less knowledgeable group, about 92% view keeping records as important. the difference in opinions is significant at the 0.098 level. when asked what they would actually do, about 45.6% of students with more knowledge keep detailed financial records and only 29% of the less knowledgeable group are willing to keep such records. about 25% of the less knowl edgeable group keep no records at all, while the number is 5.4% for the more knowledgeable group. there is a statistically significant difference in behavior between the two groups. analysis has also been conducted using an alternative classification scheme. students with section scores equal to or higher than the median are classified as those with more knowledge, and those with scores below the median as those with less knowledge. this clas sification scheme makes the sample size in the more knowledgeable group larger than that of the less knowledgeable group. the results are similar to what have been reported above. the difference in opinions regarding record keeping is statistically significant at the 0.039 level. the difference in their decisions is significant at the 0.001 level. the results for savings and borrowing, insurance, and investment are similar to findings to be reported below. the more knowledgeable participants (96.1%) rank spending less than their income more important than the less knowledgeable group (81.3%). when provided with a hypo thetical situation of a spending decision, 89% of the more knowledgeable participants select the correct choice, compared to 68.3% of the less knowledgeable group. the chi square tests suggest that the differences in opinions and decisions are highly significant. in terms of insurance, more than 80% of the participants from both groups rank main taining adequate insurance coverages as important. the difference in opinions is not signif icant. however, more participants from the more knowledgeable group act correctly than the participants from the less knowledgeable group. the difference in their decisions can be seen from the significant test results and the fact that a higher proportion of the more knowledgeable participants act according to their opinions and a higher proportion of the less knowledgeable participants do not. 122 financial services review 7(2) 1998 section d shows that about 89.4% of the more knowledgeable group view planning and implementing a regular investment program as important, but the number is about 72.2% for the less knowledgeable group. not surprisingly, when offered an investment sit uation, 80% of the knowledgeable participants choose the correct action, and only 51% of the other group. both the differences in the opinions and decisions are highly significant. the above analysis suggests that the level of finance knowledge tends to influence people's opinions and affect their decisions. in the case of insurance, a statistical difference cannot be detected in their opinions. however, more participants from the less knowledge able group make the wrong choices. v. summary and conclusion this study surveys 924 students from multiple universities across the country to examine college students' knowledge of personal finance; the relationship between the financial lit eracy and participants' characteristics such as academic discipline, gender, and experience; and the consequences of having inadequate knowledge. results suggest that college students need to improve their knowledge of personal finance. although the questions included in the survey are fairly basic, the overall mean of correct answers for the survey is about 53%. none of the mean scores for each area of gen eral knowledge, savings and borrowing, insurance, and investments are above 65%. by far the weakest area is investment, where on average the participants answered about 40% of the questions correctly. lower levels of financial literacy are found among subgroups. they include those who are non-business majors, in the lower class ranks, women, under age 30, and have little work experience. it is also found that participants with less knowl edge tend to hold wrong opinions and make incorrect decisions in the areas of general knowledge, savings and borrowing and investments. while there is little difference in their opinions regarding insurance, the less knowledgeable participants are more likely to act incorrectly. the predictive ability of personal finance knowledge shows that improving college students' knowledge is important. without adequate knowledge, they are more likely to make mistakes in the real world. our conclusion is that college students are not knowledgeable about personal finance. the incompetency will limit their ability to make informed financial decisions. together with evidence provided by the research conducted in the past three decades, the findings of this study suggest that there is a systematic lack of personal finance education in our edu cation system. the lack of education has resulted in serious financial illiteracy found in the american public. the illiteracy and its costly consequences have made individuals worry about their finances to the extent that their productivity in workplaces is affected (chrgi, 1995). when individuals cannot manage their finances, it becomes a problem for the soci ety. this challenging issue needs to be addressed. acknowledgments: the authors are grateful to karen eilers lahey (editor) for her sug gestions. the authors also appreciate comments from three anonymous reviewers, and discussant and participants at the 1997 academy of financial services meeting where an early version of the paper was presented. the authors are responsible for all remaining errors. financial supports were provided by grants from the university research council at youngstown state university. financial literacy 123 appendix survey of personal financial literacy thank you for participating in our survey. this survey is intended to measure college stu dents ' knowledge of personal finance. the results will be used to help students improve their knowledge and colleges improve curriculums. directions: please use a #'2 lead pencil to mark your responses on the enclosed answer sheet. please select only one most appropriate answer for each question. please make marks that fill the circle. after completing the survey, please make sure that question num bers and answers correspond directly with those on the answer sheet. i. general personal finance know ledge 1. personal finance literacy can help you a. avoid being victimized by financial s c a l n s . b. buy the fight kind of insurance to protect you from .catastrophic risk. c. learn the fight approach to invest for your future needs. d. lead a financially secure life through forming healthy spending habits. e. do all of the above. 2. personal financial planning involves a. establishing an adequate financial record keeping system. b. developing a sound yearly budget of expenses and income. c. minimizing taxes and insurance expenses. d. preparing plans for future financial needs and goals. e. examining your investment portfolios to maximize returns. 3. the most liquid asset is a. money in a certificate of deposit account. b. money in a checking account. c. a car. d. a computer. e. a house. 4. your net worth is a. the difference between your expendi tures and income. b. the difference between your liabilities and assets. c. the difference between your cash inflow and outflow. d. the difference between your bank bor rowings and savings. e. none of the above. 5. assume you have dependent children, is a $500 tax credit per child or a $500 tax deduction per child more valuable to you? a. a $500 tax credit. b. a $500 tax deduction. c. they are the same. d. depends on your tax bracket. e. depends on the number of children you have. 6. you are not overspending if a. you write checks for more than what you have in your checking account. b. your monthly wages are $500 and credit charges $1,000. c. you frequently receive calls from col lection agencies. d. your monthly debt payment is 30% of your take-home pay. e. you meet your minimum monthly credit card payments. 7. i s not a cost of leasing an apartment. a. security deposit b. monthly rental payment c. expenses incurred for non-compliance of lease terms d. medical expenses of your friend who fell and broke his arm on the icy pave ment e. security deposit retained by the land lord for damages to property beyond normal wear and tear 8. if you signed a twelve month lease for $300/ month but never occupied the apartment, you legally owe the landlord a. your security deposit. b. your fwst month's rent of $300. c. your twelve month's rent of $3,600. 124 financial services review 7(2) 1998 d. nothing. e. whatever the landlord requires. 9. checking account reconciliation involves a. balancing bank statement with your checkbook records to determine if there are errors. b. reconciling current bank statement with the previous month ' s statement to determine if there are errors. c. subtracting outstanding checks to your checkbook balance to determine if your checks have been properly pro cessed. d. adding outstanding checks to your checkbook balance to improve your credit standing. e. none of the above. ii. your savings and borrowing 10. your savings accounts in a federally insured commercial bank are insured by a. sipc to the maximum amount of $10,000 per account. b. fdic to the maximum amount of $100,000. c. fdic to the maximum amount of $50,000 per account. d. slic to the maximum amount of $100,000. e. fnma to the maximum amount of $100,000 per account. 11. i f you invest $1,000 today at 4% for a year, your balance in a year will be a. higher if the interest is compounded daily rather than monthly. b. higher if the interest is compounded quarterly rather than weekly. c. higher if the interest is compounded yearly rather than quarterly. d. $1,040 no matter how the interest is computed. e. $1,000 no matter how the interest is computed. 12. which of the following investments requires that you keep your money invested for a specified period or face an early withdrawal penalty? a. certificate of deposit. b. checking account that pays interest. c. government savings bond. d. money market mutual fund. e. passbook savings account. 13. which of the following statements is true about the annual percentage rate (apr)? a. apr is the actual rate of interest paid over the life of the loan. b. apr is expressed as a percentage on an annual basis. c. apr is a good measure of comparing loan costs. d. apr takes into account all loan fees, e. all of the above. 14. you can receive your credit report from a. a credit union. b. a commercial bank. c. the better business bureau. d. a credit bureau, e. a professor. 15. which is false concerning credit cards? a. you can use your credit card to receive a cash advance. b. if your credit card balance is $1,000 and you pay $300, interest is charged on the unpaid balalnce of $700. c. the rate of interest on your credit card is normally higher than you can earn on a certificate of deposit. d. a credit card company will not charge you interest if you pay off the entire balance by the due date. e. you cannot spend more than your line of credit. 16. an overdraft a. occurs when you write a $1,000 dollar check when you have $500 in your account. b. is a stop-payment order written by the payee. c. will result in fines. d. all of the above. e. b o t h a andc. 17. you will improve your creditworthiness by a. visiting your local commercial bank. b. showing no record of personal bank ruptcies in recent years. c. paying cash for all goods and services. d. borrowing large amounts of money from your friends. e. donating money to charity. 18. if you co-sign a loan for a friend, then a. you become responsible for the loan payments if your friend defaults. b. it means that your friend cannot receive the loan by himself. financial literacy 125 c. you are entitled to receive part of the loan. d. both a and b. e. b o t h a a n d c . iii. your insurance 19. auto insurance companies determine your premium based on a. age of insured. b. record of accidents. c. type and age of vehicle. d. completion of a driver education course. e. all of the above. 20. the main reason to purchase insurance is to a. protect you from a loss recently incurred. b. provide you with excellent investment returns. c. protect you from sustaining a cata strophic loss. d. protect you from small incidental losses. e. improve your standard of living by fil ing fraudulent claims. 21. the main reason to purchase insurance is to a. after buying health insurance, you are normally covered for pre-existing con ditions. b. you have a better chance to choose doctors with a health maintenance organization rather than with a tradi tional health care insurance company. c. most policies contain deductible and coinsurance clauses. d. a policy purchased by the individual is cheaper than one purchased through a group. e. none of the above. 22. would not ordinarily be covered under a homeowners policy. a. war b. earthquake c. flood d. your being sued by someone for slander e. all of the above 23. which of the following statements is false? a. term insurance is an excellent invest ment vehicle. b. you receive no benefits when your term insurance policy expires. 24. c. a term insurance policy is the least expensive form of life insurance. d. a decreasing-term policy reduces cover age over time. e. a level-term policy guarantees a fixed premium over the life of the contract. you have a better chance of resolving a complaint against an insurance company by bringing the issue to a government agency at the a. federal level. b. state level. c. county level. d. township level. e. none of the above. iv: your investments 25. if interest rates rise, the price of a treasury bond will a. increase. b. decrease. c. remain the same. d. trade at a premium. e. be impossible to predict. 26. a dollar-cost-averaging approach to investing involves a. buying low and selling high. b. complex calculations of risk and return. c. selling securities to minimize capital 29. a high-risk and high-return investment strategy would be most suitable for a. an elderly retired couple living on a fixed income. b. a middle-aged couple needing funds for their children's education in two years. c. a young married couple without chil dren. d. all of the above because they all need high return. e. none of the above because they are equally risk averse. 30. which of the following is false? a. as shareholders of a mutual fund, you have a right to tell fund managers what securities to buy. b. a mutual fund is a diversified collection of securities used as an in vestment vehicle. c. a mutual fund is an investment corpora tion that raises funds from investors and purchases securities. d. your ownership in a mutual fund is pro portional to the number of shares you own in the fund. e. none of the above. 126 financial services review 7(2) 1998 31. the returns from a balanced mutual fund include a. interest earned on cash in the fund. b. dividends from common stock in the fund. c. interest earned on bonds in the fund. d. capital gains from stocks and bonds in the fund. e. all of the above. 32. no-load mutual funds are recommended over load funds because investors a. do not pay for 12b-i fees. b. can reduce their tax liability. c. are not charged with sales commissions. d. can avoid the funds' administrative expenses. e. believe that the funds have no manage ment charges. 36. if other factors remain the same, u.s. dol lar value of a japan fund will be a. higher if the dollar's value rises against that of the japanese yen. b. lower if the dollar's value rises against that of the japanese yen. c. unchanged if the japanese yen's value rises against that of dollar. d. lower if the japanese yen's value rises against that of dollar. e. impossible to determine if exchange rate changes between yen and dollar. v: your personal finance opinions. decisions. and education 37. assume you're in your early twenties and you would like to build up your nest egg for a secure retirement in 30 years. which of the following approaches would best meet your needs? a. start to build up your savings account at a federally insured bank. b. save money in certificate of deposit a~2ounts. c. put monthly savings in a diversified growth mutual fund. d. invest in long-term treasury bonds. e. accumulate money in a safe-box rented from a local bank. 38. assuming you are in your early twenties without any dependents, which of the fol. lowing would you do regarding your life insurance? a. you would buy a life insurance policy from an insurance agent. b. you would buy a term insurance policy. c. you probably do not need to buy any life insurance policy. d. you would buy flight insurance each time you travel by air. e. you would buy a cash value insurance policy. 39. you have just graduated from college and found a job earning $28,000 per year. you will pay $600 per month for five years for student loans. you have a monthly balance on each of your three credit cards. what should you do to improve your financial health? a. cut expenses and use your savings to pay down debt. b. keep the same spending pattern as in the past. c. apply for a consumer loan for a new car. d. eliminate debt by filing personal bank ruptcy. e. use your credit card to pay for a vacation in the bahamas. 40. do you maintain financial records? a. maintain very detailed records. b. maintain minimal records. c. maintain no records. using the scale given below please rank the importance of items numbered from 41 to 44. a b c d e very somewhat not somewhat very important important sure unimportant 41. maintaining adequate financial records. 42. spending less than your income. 43. maintaining adequate insurance coverage. 44. planning and implementing a regular investment program. vi. about yourself 45. what is your class rank? a. freshman d. senior b. sophomore e. graduate c. junior 46. what is your age? a. 18-22 d. 40-59 b. 23-29 e. 60 or older c. 30-39 47. what is your sex? a. male b. female 48. what is your race or ethnic background? a. white c. hispanic b. africand. american indian american e. asian financial literacy 127 49. 50. which best describes your or your fam ily's personal income last year? a. no income d. $30,000-$49,999. b. under $10,000 e. $50,000 or more. c. $10,000 $29,999. how many years of working experience do you have? include fullor part-time expe rience, internship, co-op, summer jobs, etc. a. none b. less than 2 years c. two to less than 4 years d. four to less than 6 years e. six years or more 51. what is your major field of study? a. business b. education c. liberal arts d. sciences or engineering e. others 52. are you a foreign sutdent? a. yes b. no c. two to less than 4 years d. four to less than 6 years e. six years or more vii. informed consent form as we explained in the cover letter, the purpose of this survey is to help you improve personal finance lit eracy. anonymity will be kept throughout the survey. however, to comply with youngstown state univer sity's policy of human subject research, we would like to ask you to sign or initial below to verify that your participation is voluntary. my participation is voluntary. if you would like to receive a copy of your personal finance intelligence report and summary of this research project, please write down your address. street address city, state, and zip thank you for your participation. r e f e r e n c e s bakken, r. (1967). money management understandings of tenth grade students. national business education quarterly, 36, 6. bialaszewski, d., pencek, t., & zietlow, j. (1993). finance requirements and computer utilization at aacsb accredited schools. financial practice and education, fall, 133-139. cambridge human resource group inc. (chrgi). (1995) latest workplace issues survey. chicago, il. consumer federation of america and american express company (cfajamex). (1991). student consumer knowledge: results of a national test. washington, d.c. danes, s. m., & hira, t. k. (1987, winter). money management knowledge of college students. the journal of student financial aid, 17(1), 4-16. employee benefit research institute (ebri). (1995). are workers kidding themselves? results of the 1995 retirement confidence survey. december, washington, dc. genasci, l. (1995, september 18). women and retirement: an unpleasant surprise: you're retired and broke. the vindicator, p. b5. gitman, l. j., & bacon, p. w. (1985). comprehensive personal financial planning: an emerging opportunity. journal of financial education, fall, 36-46. 128 financial services review 7(2) 1998 goldsmith, e., & goldsmith, r. e. (1997). gender differences in perceived and real knowledge of financial investments. psychological report, 80, 236-238. harris/scholastic research (hsr). (1993). liberty financial young investor survey. new york, ny. hira, t. (1993). financial management knowledge and practices: implications for financial health. paper presented at the personal economic summit '93, washington, d.c. institute of certified financial planners (icfp). (1993). planners say financial illiteracy plagues americans: inability to plan and poor understanding of investments cited. denver, co. kpmg peat marwick llp (kpmg). (1995). retirement benefits in the 1990s: 1995 survey data. june. new york, ny. langrehr, f. w. (1979). consumer education: does it change students' competencies and attitudes? the journal of consumer affairs, 13, 41-53. lewin, t. (1995, april, 20). income gaps for sexes is seen wider in retirement. the new york times, p. a19. mandell, l. (1997). personal financial survey of high school seniors. jump start coalition for per sonal financial literacy, march/april. washington, d.c. martinez, m. n. (1994). why women should save and plan more. hrmagazine, november, 104 105. national assessment of educational progress (naep). (1979). teenage consumers: a profile. den ver, co. national endowment for financial education (nefe). (1993-1996). cfp survey of trends in finan cial planning. denver, co. o'neill, b. (1993). assessing america's financial iq: reality, consequences, and potential for change. paper presented at the personal economic summit '93, washington, d.c. oppenheimer funds/girls inc. (1997, march). girls, money and independence. new york, ny. princeton survey research associates (psra). (1996). investor knowledge survey. investor protec tion trust, may. arlington, va. princeton survey research associates (psra). (1997, january/february). planning for the future: are americans prepared to meet their financial goals? nations bank~consumer federation of america. princeton, nj. vanguard group/money magazine. (1997). financial literacy of mutual fund investors, january, new york, ny. volpe, r. p., chen, h., & pavlicko j. j. (1996, fall/winter). personal investment literacy among col lege students: a survey. financial practice and education, 6(2), 86-94. pii: s1057-0810(96)90029-3 financial services review, 5(l): 83-86 copyright 0 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. book, software, and web site reviews douglas kahl, editor university of akron this is a new feature in financial services review. to make it a success, we need vol unteers willing to do the reviews. if you would be interested in doing an occasional review, please contact douglas kahl, department of finance, university of akron, akron. oh 44325403 or e-mail him at dkahl@uakron.edu. employee benefits (4th edition) by burton t. beam, jr., and john j. mcfadden chicago, il: dearborn financial publishing, inc.; 1996 (isbn o-7931-1504-3) reviewed by: douglas r. kahl, professor of finance, the university of akron burton t. beam, jr., and john j. mcfadden’s fourth edition of employee benefits has just been published by real estate education company, a division of dearborn financial publishing, inc. this is a serious book for serious students. the book has no cartoons, rel atively few exhibits, and is presented in two colors, black and white, with an occasional bit of grey to highlight an exhibit. the subject is covered from a benefits administrator’s point of view and is structured to be applicable to employers of all sizes and to multiple employer groups* employee benefits, an insignificant proportion of employee compensation 50 years ago, often constitutes 30% of employee compensation today. this thorough presentation of modern employee benefits makes clear the reasons for the rapid growth in the number, variety, and cost of employee benefits over the past five decades. the range of topics cov ered is extensive, from such standard topics as group insurance and retirement plans to much less common benefits such as adoption assistance and eldercare benefits. the text is arranged into five parts with 28 chapters. two of the parts, group insurance and retirement plans, make up the bulk of the book. as the baby boomers turn 50 and the work force ages, these two largest and most costly benefits can only increase in interest and importance. whether the rapid increase in cost and proportion of employee compensation will continue is less certain. the group insurance segment introduces the group insurance environment and pro vi&s a discussion of the most important tax and regulatory issues; group life insurance benefits, disability income benefits, medical and dental expense plans, and long-term care insurance. this segment also provides an excellent discussion of the history and develop ment in the area and a very up-to-date discussion of many of the current issues-such as 84 financial services review 5( 1) 1996 the development of and prospects for managed care plans, alternative funding methods for group insurance, and group insurance rate making. retirement plans is the second major section of the book, covering all of the major retirement plans. the discussions in this section are as up-to-date and detailed as can rea sonably be expected. the presentation in this section is extensive and well organized. in addition to a discussion of all of the major types of plans, the section provides a discussion of pension plan design, benefit formulas, plan funding, highly compensated employees, nonretirement qualified plans, and nonqualified deferred compensation plans. the presen tation covers plan installation, administration, investments, and termination. the issues in social insurance are given good coverage. there is a fairly detailed dis cussion of social security and medicare benefits with an excellent discussion of funding adequacy and alternatives. a brief discussion of unemployment benefits and workers com pensation programs is included. finally, the book contains a brief discussion of a very wide range of unrelated benefits frequently provided to employees. these benefits include personal time, service awards, adoption assistance, and parking. the discussion is concise, interesting and includes some rapidly developing topics. each chapter in the text ends with about a dozen discussion questions on the topics covered in the chapters. answers to the chapter-end questions are provided on an instruc tor’s manual disk, which also contains chapter outlines and a multiple choice test bank. no other ancillary materials are provided. this is a well-written, well organized, detailed and up-to-date text for a course in the rapidly developing area of employee benefits. in addition, this text would make an excellent addition to any business or business school library and an often-used desk reference book for anyone teaching in the area of per sonal finance. while many of the topics in this book are covered in other personal finance courses, the presentation of many important personal finance topics from the employer’s point of view provides some valuable insights. personal finance (4th edition) e. thomas garman and raymond e. forgue boston, ma: houghton mifflin company; 1994 (isbn o-395-66852-2) reviewed by: john clinebell, professor of finance, university of northern colorado e. thomas garman and raymond e. forgue’s personal finance, 4th edition, is an excellent textbook designed for a lower division (i.e., freshman/sophomore level) personal finance class. the informal writing style and level of the text imply that the authors are assuming that students have little to no background in finance. the text is clearly designed for a service course that addresses the needs of nonbusiness students. the book is organized in the following six broad topical areas: financial planning, money management, managing expenditures, income and asset protection, investment planning, and retirement and estate planning. by far the heaviest emphasis is placed on money management and managing expenditures with approximately one third of the book devoted to these topics. the topics introduction to financial planning and retirement and estate planning have the most limited coverage. topics related to insurance and invest ments have relatively equal emphasis. book, software, and web site reviews 85 the organization of the book and its emphasis on controlling debt and expenditures (the management of money) indicate that the book is best suited for a course designed to take a practical approach to specific aspects of personal finance rather than courses designed to take a more holistic approach to the financial planning process. further, the emphasis placed on career selection (chapter 2) suggests that this book may be better suited to traditional students. issues generally associated with nontraditional students such as changing careers and dual career families are given very little attention, the placement of the investment and retirement planning topics at the end of the book may cause difftcul ties in course planning for some instructors. the length of the book makes it unlikely that the entire book can be covered in one semester course. therefore, the critically important topics of investments and retirement planning will probably be taken out of sequence. although the text provides relatively broad coverage of personal finance topics, the grow ing importance of technology and on-line finance resources receives little attention. this lack of coverage will require instructors to develop supplemental readings and exercises to ensure students are exposed to this increasingly important topical area. the authors make excellent use of many pedagogical features. i especially enjoyed their use of boxed inserts such as ‘the top ten money mistakes that waylay young peo ple,” “how to set family financial goals,” and other “how to . . .” topics. these features make the text much more enjoyable and add significantly to the potential learning of the students. the objectives for each chapter are clearly defined at the beginning of the chapter and marginal notes are used to identify and highlight the chapter discussions. the authors provide comprehensive review questions at the end of each chapter and provide several end-of-chapter “action involvement activities. ” the “action involvement activities” are designed to encourage students to apply the concepts in the chapter to their own lives by visiting and dealing with businesses, creating personal budgets, etc. one potential weak ness i found with the text is the “decision-making cases” at the end of each chapter. these %ses” are basically problems rather than cases and although they are generally well writ ten, very few are provided. instructors who like to take a “hands on” or applied approach by assigning many problems to aid in covering the material may find the lack of problems at the end of each chapter frustrating. however, software is provided to aid in the analysis of some of the “cases” at the end of each chapter. overall, this is an excellent text for a personal finance class. instructors who prefer a textbook that is both comprehensive and descriptive in nature and who are teaching a course designed for traditional freshman and sophomore nonbusiness students should find this to be an excellent book. liffx on the internet reviewed by: douglas r. kahl, professor of finance, the university of akron located at &ttp://www.liffe.com/liffe/home.htm>, the home page for the london international financial futures and options exchange (liffe) offers finance students an opportunity to take an international look at derivatives. this a well done web site which will provide your students with an interesting and informative look across the atlantic. liffe bills itself as europe’s premier financial derivatives exchange, on the cutting-edge of technology, and intending to be the leading futures and options exchange in the world. among the many interesting offerings on the site is a discussion of the uk regulatory structure. the concise and informative discussion will provide students outside the uk 86 financial services review 5( 1) 1996 with a good look at a different regulatory philosophy, structure, and system. the potential for discussions, comparisons and contrasts could be pedagogically exciting in courses from derivatives to business law. u.s. students visiting the glossary will find many familiar terms as well as a few new terms to add to their vocabulary. on the lighter side, a visit to the photo library or the video clips library will put a face on the story. both are interesting, informative and well done. the discussion of lilfe’s automated pit trading system makes a strong case for liffe being on the cutting edge of technology. the automated pit trading system emu lates an open outcry trading pit in that all traders receive all bid/ask information simulta neously and any trader has an equal chance of accepting any bid or offer. the integral automated trading and order matching system provides a central limit order book, auto matic execution of stored orders touching the opposite side of the market, and a display of aggregate market bid/offer/volume information. a web site worth looking at. give it a try. pii: s1057-0810(00)00040-8 an integrative approach to using student investment clubs and student investment funds in the finance curriculum brian grindera,*, dan w. cooperb, michael brittc adepartment of management, eastern washington university, cheney, wa 99004, usa bschool of management, marist college, poughkeepsie, ny 12601-1387, usa cschool of social and behavioral sciences, marist college, poughkeepsie, ny 12601-1387, usa abstract the educational advantages of student investment clubs and student investment funds have been well documented. this paper suggests ways to integrate them into the finance curriculum and examines the benefits of integrating both funds and clubs into an instructional framework that provides important out-of-class experience to potential finance students, allows for the practical application of finance theory, and gives alumni an opportunity to remain involved with their alma mater. a discussion of how research in student learning styles can help instructors maximize the benefits of club and fund activity is also included. © 1999 elsevier science inc. all rights reserved. jel classification:i22 keywords:investments; clubs; funds; curriculum; application 1. introduction student investment clubs and funds are increasingly becoming an important part of the finance education process. two articles infinancial practice and education (fpe),“starting an investment club” (1996) by don r. cox and delbert c. goff and “financial innovation: the case of student investment funds at united states universities” (1994) by * corresponding author. tel.:11-509-359-4235; fax:11-509-358-2267. e-mail address:brian.grinder@ewu.edu (b. grinder). financial services review 8 (1999) 211–221 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00040-8 edward c. lawrence, have dealt with the benefits of these activities for students of finance. however, no one has yet examined the benefits of integrating these two activities into a comprehensive and practical finance program nor suggested how to best include these activities within a typical finance curriculum. investment clubs can help provide a practical base for students who are interested in finance and investments while investment funds offer the opportunity for more advanced hands-on training. before the benefits of integrating these activities into a finance program can be discussed, a separate examination of the nature of student investment clubs and student investment funds is warranted. 2. student investment clubs investment clubs can, as an extracurricular activity, provide a practical base for a finance program. the only requirement for joining an investment club is a desire to learn about investing and perhaps the willingness to commit one’s own funds to the process. finance departments are often at a disadvantage when it comes to recruiting their own university’s freshmen and sophomores. since most finance courses are offered at the junior and senior level, students at the freshman or sophomore level are often not aware of the advantages of pursuing a degree in finance. investment clubs can help offset this problem by actively recruiting freshmen and sophomores as members. good recruiting times include freshman orientation or during registration at the beginning of the term. it is important for the club to make its presence known on campus in order to successfully attract new members to the club. dedicated faculty sponsors must take a leadership role in the recruitment process in order to bring in new members on a consistent basis. once students are made aware of the club and begin to participate in club activities, the very nature of the club is bound to generate an interest in finance among club members. most investment clubs collect dues on a monthly or quarterly basis. these funds are then used to purchase securities for the club’s portfolio. special guests such as local stockbrokers or financial planners are often invited to club meetings in order to explain different investment strategies to the club. clubs can also make of use educational materials from the national association of investors corporation (naic) as well as other resources. as the club begins to build a portfolio and as the portfolio changes in value with club decisions and market fluctuations, members often become quite enthusiastic. they become deeply involved in the pursuit of new investment possibilities. the lively discussions of past investment successes and failures that often occur spontaneously help students to develop a good knowledge base. such interaction and enthusiasm creates a wonderful informal environment where students can learn a great deal about finance and have fun in the process. even if a student chooses not to major in finance, his or her involvement in an investment club may be the factor that leads that student into taking additional business/finance courses. furthermore, since topics such as budgeting, and the use of credit are often brought up for discussion at club meetings, those students who participate in the club, regardless of their major, will have a better understanding of how to manage their own finances. 212 b. grinder et al. / financial services review 8 (1999) 211–221 participation in an investment club will also give the members who choose to study finance valuable practical experience before they begin formal course work. club members will have already been exposed to a number of topics including stocks and bonds, financial markets, exchanges, efficient market theory, ratio analysis, asset allocation, and security selection. students who discuss their investment club experiences in the classroom, often help produce markedly higher levels of class participation. investment clubs also give finance faculty the opportunity to interact with students outside the realm of the classroom. faculty can get involved by sponsoring the club, providing investment funds for the club, or presenting investment information to the club. such interaction can have positive effects in terms of recruitment and retention of students in the finance program. however, faculty should always keep in mind that these benefits are secondary to helping club members learn sound investment principles that can be used in their own investment activities. 3. student investment funds cox and goff (1997) define a student investment fund as “one where university students have full discretion over the management of a real dollar portfolio typically a portion of the university’s endowment or money donated for the purpose of creating a student managed investment fund.” lawrence (1994) defines a student investment fund as a fund “where university students have full discretion over the management of areal dollar portfolio.” recent articles on student investment funds include block and french (1991), bhattacharya and mcclung (1994), and johnson, alexander and allen (1996). funds are generally obtained for these portfolios through individual giving, university endowments, or corporate donors. most student investment fund participants are required to enroll in some type of investments/portfolio management course (see lawrence 1994). the benefit for students in this arrangement is that they are able to apply the theories learned in the classroom to the management of the fund portfolio. student managers of investment funds who have also been members of an investment club will also be able to apply their club investing experience to the fund portfolio. there are significant differences between the investment activities of a club and of a fund. first, the size of the fund portfolio is likely to be much larger than the club’s portfolio. the median value of the funds surveyed by lawrence (1994) was about $200,000. the typical student investment club portfolio is usually much smaller than this. second, while clubs are free to set their own investment objectives, donors often have specific objectives in mind when they set up a student-managed fund. for instance, a retiring faculty member at a university in the pacific northwest and his wife recently donated a portfolio of stocks and mutual funds that is currently worth about $95,000 to the university. the objectives for this gift are 1) to provide a vehicle whereby students can learn the fundamentals of investing, and 2) to provide scholarships for students. the second objective will serve as a general guide for the students managing the fund. the goal is to grow the portfolio in order to provide 213b. grinder et al. / financial services review 8 (1999) 211–221 scholarships. investment decisions for this portfolio must always be made with this goal in mind. if fund donors are still living, it is essential to invite them periodically to the class to discuss their goals and objectives for the portfolio with the student managers. this interaction will allow students to develop a better understanding of the risk tolerances of the donors, and the objectives of the portfolio. fund donors are often very interested in students and are more than happy to meet with them from time to time. it is also useful to have the students make a formal portfolio performance review to the donors on a regularly scheduled basis. a second set of restrictions may be placed on a student fund if it is part of an endowment (see lawrence, 1994 p. 50). while such restrictions are sometimes looked upon negatively, they are valuable learning tools for those students who plan to make a career of managing the money of individuals or institutions. the goals, objectives, and risk tolerances of those individuals and institutions will not necessarily be the same as the goals and objectives of the fund manager. successful fund managers are able to grow a portfolio within the restrictions set down by the owners of the portfolio. it is clear that both clubs and funds are useful experiential learning activities. it is also clear that they differ substantively in the type of students that are involved as well as the complexity of the decision-making processes that takes place within each activity. club members often make decisions in a relatively unconstrained environment risking only their own pooled funds. while this is not a trivial process it is also not significantly bound by a formal set of investment policies or objectives common to investment funds. 4. pedagogic considerations as we consider developing an integrative instructional framework using clubs and funds, it is useful first to consider how students learn and then to consider how to best impact the learning process. students learn on different cognitive levels. these levels were first and perhaps most convincingly outlined by bloom in his seminal work on cognitive domains and educational goals (see bloom et al., 1956). bloom defined six levels of learning that extend across a continuum of student awareness. table 1 lists and briefly defines bloom’s domains. introductory courses typically focus on achieving student outcomes that demonstrate the first three of bloom’s levels (knowledge, comprehension, and simple application). more advanced courses, integrative cases, and complex simulations typically impact the student at table 1 bloom’s taxonomy of cognitive domains cognitive domain students demonstrate this domain when they can. . . 1. knowledge define, identify, or list information. 2. comprehension explain, distinguish, or summarize information. 3. application use, solve, or manipulate information. 4. analysis break down, differentiate, or discriminate information. 5. synthesis combine, compose, or rearrange information. 6. evaluation compare, contrast, or criticize information. 214 b. grinder et al. / financial services review 8 (1999) 211–221 the higher four comprehension levels (application, analysis, synthesis, and evaluation). students that struggle in senior level courses often do so because they were unable to master the lower levels of understanding. graduate studies are often built entirely around the upper three domains. in addition to this focus on intellectual cognitive abilities, most teachers feel that an important purpose for a student’s education is to influence student behavior in positive ways. just as students need direction and guidance in a purely intellectual context, they also need opportunities to develop and refine their own ethical decision-making framework. in recent years, additional work on the principles of student cognition has produced another useful approach to internalizing student learning levels and values known as the “affective domain.” learning levels within the affective domain are based on the principle that learning “affects” behavior (see krathwohl, 1964 or linn and gronlund, 1995). setting specific learning objectives and then creating educational situations that meet these objectives demonstrate student comprehension at the specific affective domain level. table 2 lists and briefly defines the affective domain. understanding the affective domain and designing educational opportunities to maximize student exposure to values-oriented decisions produces students who are logically more inclined to act responsibly outside the confines of the academic institution. good learning environments do not simply make use of a series of cognitive and affective levels. they must also consider the different ways that individual students learn. research in this area has produced many different types and definitions of learning styles. most college level instructors are aware that some students tend to be “concrete” learners while others are more “abstract.” probably the most widely known and usable framework in learning styles has come from dr. howard gardner (1983) of harvard university. by examining the different areas of life where people excel, gardner was able to identify seven kinds of intelligence. while people have abilities in nearly all of these areas, each of us tends to excel in only one or two of them. because of this, we seem to learn well when these specific abilities are tapped. table 3 lists gardner’s seven intelligences, describes the kinds of skills that students with that type of intelligence typically possess and suggests teaching techniques that may be best used to tap into those skills. traditionally the classroom tends to emphasize only the first two of these seven intelligences: verbal/linguistic and logical/mathematical. in the typical college course the student is required to: table 2 levels of affective domain affective domain students demonstrate this when they. . . 1. receiving listen, attend carefully, or show sensitivity toward an issue. 2. responding enjoy or show interest in an activity. 3. valuing appreciate or feel that something has value. 4. organization create (or organize) a value system or personal plan. 5. characterization by a value act on a value system, make a commitment to following or value complex certain values or plans. 215b. grinder et al. / financial services review 8 (1999) 211–221 ● attend lectures ● read the text ● write notes ● respond to questions or participate in class discussion ● do well on written tests or papers. all of these activities tap into verbal/linguistic and logical/mathematical skills. students whose strengths lie in these areas tend to benefit most from this approach. however, many students excel in other learning styles. for example, some concepts are not fully understood by some students until they can see the idea in a graph or chart (visual/spatial learners). others learn best by interacting with other students and discussing the issues (interpersonal learners), while others need to see the connection between the material and their own lives (intrapersonal learners). there is clearly a need for other types of educational experiences to help tap into these different aptitudes and predilections. as every instructor of finance knows, teaching the concepts, theory, mathematics, and applications of the field is a challenging undertaking. understanding bloom’s work on cognition helps teachers of finance set reasonable objectives and outcomes for each type of course taught. incorporating the work on affective domains helps our students develop a sense of enjoyment and an appropriate set of values for their future work in the profession. being aware of and responsive to the different ways our students learn increases the amount of learning that can occur. while a more in-depth discussion of each of these pedagogic topics is well warranted, this section was meant to simply provide a reasonable foundation for the following discussion on how to best include club and fund activities within the structure of our courses. 5. integrating learning theory and cognition into the club and funds armed with this basic understanding of the ways students learn and the different dimensions to the learning process, it is useful to consider some of the ways that club and fund table 3 gardner’s seven intelligences type of intelligence skills teaching techniques verbal/linguistic speaking, reading, writing textbooks, papers, class discussion, oral presentations logical/mathematical calculating, questioning, experimenting experiments, puzzles, word problems, calculations spatial drawing, visualizing videos, pictures, graphic organizers interpersonal relating to others cooperative learning projects intrapersonal reflecting on one’s actions, knowing oneself journals, self-paced learning projects bodily/kinesthetic building, moving, using one’s hands using manipulatives, role playing musical listening/creating sounds and rhythms creating or performing musical compositions 216 b. grinder et al. / financial services review 8 (1999) 211–221 activities fit within the work on cognition and learning. the investment club offers an early educational opportunity for novice financiers to learn by working with other students to uncover appropriate investment selections. as they strive to meet this primary objective they must quickly achieve the first three of bloom’s domains. faculty advisors, experienced club members, and outside experts all help immerse the novices in the basic principles of finance. more senior club members act as mentors and discussion leaders improving their own understanding as they interact with the less experienced club members. interpersonal learners make the best mentors and although they may struggle more with the mathematics and textual material, the payoff is the satisfaction they gain by interacting with (teaching) the novices in the club. faculty who understand the importance of different learning styles must take care to step back and allow this mentoring to take place. while each club member must pull their own weight and work hard individually, they must also learn to turn to one another for direction. verbal/linguistic and interpersonal learners find great benefit in this experience. all club members regardless of learning disposition must work together to first analyze and then convince the other club members that their recommendations are sound. the incentive structure for an investment club is entirely different from the classroom. within the club, members focus on meeting a shared set of investment goals. there is simply no opportunity for cheating and all club members share in the success or failures. students are not competing against one another, are not jealous or fearful about grades, and are equally motivated for the team to do well. when decisions produce negative consequences, club members must learn to deal with criticism, disappointment, and the very real loss of their own scarce investment capital. this is precisely the type of learning activity that allows students to come to understand the need for values and goals, and further helps them get a feel for whether they will enjoy working in the field of finance. this is also a good example of how a student activity can utilize the concepts of the affective domain. student managed funds also produce a team atmosphere that places students in a decisionrich environment. but this environment differs from the simpler club framework. within the fund, students must make decisions in a more constrained setting. further, these decisions must be analyzed, synthesized, and evaluated. each decision must be adequately justified before the group, the course instructor/faculty advisor, and often to at least one other student fund manager who has been specifically designated to study the decision independently. students must learn to work in an environment where each decision must be validated, where decisions have real consequences, and where the group’s decision may often differ from any individual student participant. bloom’s upper three domains are exercised regularly in this process. as students realize the consequences of their actions and as responsible faculty or investment professionals work with the students, the need for professional values and an ethical framework are also very likely to develop. this is a clear application of the principles of the affective domain. since clubs and funds differ in how they impact students, incorporating both of these activities into the curriculum provides important learning synergies. the goal is to produce a set of experiential activities that enhance student learning beyond what the typical curriculum would produce. 217b. grinder et al. / financial services review 8 (1999) 211–221 6. integrating club and fund activities within the finance curriculum student investment clubs and student-managed funds can be used together to provide a powerful learning experience for finance majors. fig. 1 provides a basic framework for this process. finance students often complain that the courses they take are so specialized and compartmentalized that they have difficulty tying everything together into a comprehensive whole. part of the problem is that most traditional students have no pertinent experience from which to draw. this can be remedied to a large extent by offering students the opportunity to join an investment club prior to their formal finance training. the students who have been introduced to the problems common to the investing process in the clubs are better prepared to understand the importance of key financial topics. few new finance students can properly connect financial practices and theories to the specific problems that the theory or practice was developed to help address. experience gleaned in club decisions translates directly within the first finance courses to a fuller understanding of the foundations of financial thought. another way to help remedy this problem is to offer a course in personal finance. vihtelic (1996) argues that a personal finance course should serve as the introductory finance course since “it better matches students’ interests, personal experiences, and cognitive structures.” this is true whether personal finance is offered as the introductory finance course or as a lower level elective. it is also true of investment clubs. the personal finance course is also a great place to recruit for investment clubs. students who are enrolled in a personal finance course probably already have a great deal of interest in finance. the combination of a structured course in personal finance along with the practical experience of the investment club will provide those students who go on to study finance as a major with a solid foundation. further most personal finance courses spend very fig. 1. flow chart of student participation in clubs and funds 218 b. grinder et al. / financial services review 8 (1999) 211–221 little time on investing. the course often ends just as the students are beginning to realize how interesting, fun, and profitable wise investing can be for them. when club members begin their upper level course work in finance, they should be well prepared to participate in a student-managed fund. student-managed funds are usually paired with a specific course such as portfolio management. practical application of the concepts learned in this more traditional classroom setting can be made, with the help of the professor, as the students are engaged in managing the fund. the fund acts as an ideal springboard for the topics of portfolio management. additionally, the course and fund give students the opportunity to gain experience in dealing with individuals who might not share their investment philosophies, time horizons or investment goals. of course, students may also continue to participate in an investment club while they are taking upper division courses. this gives them yet another opportunity to apply the concepts they are learning in the classroom and to share this growing knowledge with the younger club members. students who have participated in a club or fund or both often enjoy the experience so much that they want to continue it in some way. one way to accommodate this desire is to form a college sponsored investment club where both students and alumni can participate. this gives students a great chance to network with alumni who have established their careers, and it gives local alumni a way to remain active with their alma mater. students are often able to parlay their club and fund experience into jobs in the investment industry upon graduation. first job turnover should also be lower as students arrive at their first job with a clearer understanding of what an investment professional does for a living. 7. reflections it has taken a number of years to set up the overall program described in this paper. while it is fairly easy to start an investment club, student-managed investment funds can take years to establish. it took two years of negotiating with the donor, the university endowment association, the dean of the college of business, and a department chair to set up a fund at one of the author’s institutions. however, the time and effort is beginning to pay off, as students are now able to make practical application of the ideas and theories they are learning in the classroom. the enthusiasm and participation levels in the portfolio management class, which manages the fund, have been extraordinary. in fact some former students of the class missed the interaction and learning environment of the class so much that they were eager to form an investment club for alumni, upperclassmen, and retired faculty. this club, with only two or three meetings under its belt at the time of this writing, completes the integrative framework described above. some fine-tuning and assessment procedures need to be implemented, but the basic program is now in place and ready to be utilized to the fullest. a valuable side effect of faculty, students, and alumni working together are the long lasting ties created between all the participants. the act of investing together forges a bond that can last far beyond any quarter or semester. the continuing popularity of investment clubs outside of higher education is evidence of the strong ties that can form. since most academic institutions devote significant resources to the cause of maintaining relationships 219b. grinder et al. / financial services review 8 (1999) 211–221 with graduates, the fact that investment clubs and funds create such persistent connections between faculty and alumni should not be overlooked. investing activities often contribute to personal networks that can serve the student for their entire career. a student who was struggling in an introductory finance class recently acknowledged that part of her problem was that she had very little exposure to the world of finance prior to the course. thus nearly everything that was discussed in the class was very new to her, and she had difficulty relating to the subject matter. had she been involved in an investment club before taking the class, this student could have learned important information in an informal setting that would have made the course less threatening and less of a struggle for her. the great challenge is to direct students during their freshman and sophomore years into the investment club environment. for many students this is a time of great uncertainty about their future. the new experience of college is often overwhelming and frightening, and while they may have an interest in investing, other priorities often take precedence. furthermore many cash strapped students just can’t see how they can afford to invest in their present circumstances. an important goal of any student investment club is to educate the students on their campus and help them to realize that they really can’t afford not to invest! naturally, the real measure of success for any innovation in teaching or learning can only be gauged after the student has graduated. “real world” experiences such as club and fund activities are consistently identified by past students as the part of their education that they have found most valuable in their professional lives. alumni and business advisory groups constantly urge our institutions to immerse our students in more integrative learning experiences. many past students have contacted the authors explicitly to express their gratitude for the extra time and effort spent in the investing activities. some even call to let us know about hot stock tips! 8. conclusion a holistic approach to the finance curriculum should make full use of both investment clubs and investment funds. clubs offer an informal setting where freshmen and sophomores can begin to build a knowledge base. those who do not go on to major in finance will have learned valuable skills for investing and managing their own money. those who go on to major in finance will have built a solid foundation for upper division finance courses especially if their investment club activities are coupled with a course in personal finance. student-managed investment funds provide students with a more formal “financial laboratory” where the lessons of finance theory can be applied. it also gives students the opportunity to interact with fund donors and the school’s endowment association. fund management requires students to learn to make informed and responsible decisions in a way that simply cannot be taught in a traditional classroom setting. investment clubs are also an excellent vehicle for keeping alumni in touch with their alma mater. alumni benefit because they can continue to engage in an activity that interests them, students benefit because they are able to network with alumni, and schools benefit because of the increased interaction with alumni. both club and fund activities offer alternative ways to increase student comprehension by 220 b. grinder et al. / financial services review 8 (1999) 211–221 placing students in situations that impact a broader spectrum of cognitive domains. additionally, individual student learning is enhanced as students work together to meet group goals and objectives. students are able to work together in an enhanced learning environment that allows each member to take full advantage of their own skills and abilities. faculty participation, of course, is key to the success of both types of activities. clubs and funds offer an ideal nontraditional teaching and learning experience that can span a student’s entire college experience. references bhattacharya, t. k. & mcclung, j. j. (1994). cameron university’s unique student-managed investment portfolio. financial practice and education, 4,55–60. block, s. b. & french, d. w. (1991). the student-managed investment fund: a special opportunity in learning. financial practice and education, 1,35–40. bloom, b. s., englehart, m. d., furst, e. j., hill, w. h. & krathwohl, d. r. (1956).taxonomy of educational objectives, handbook i: cognitive domain. new york: david mckay. cox, d. r., and goff, d. c. (1996). starting and operating a student investment club.financial practice and education, 6,78–85. gardner, h. (1983).frames of mind: the theory of multiple intelligences.new york: basic books. johnson, d. w., alexander, j. f. & allen, g. h. (1996). student-managed investment funds: a comparison of alternative decision-making environments.financial practice and education 6,97–101. krathwohl, d. r., et al. (eds.). (1964).taxonomy of educational objectives: handbook ii, affective domain. new york: d. mckay. lawrence, e. c. (1994). financial innovation: the case of student investment funds at united states universities. financial practice and education, 4,47–53. linn, r. l. & gronlund, n. e. (1995).measurement and assessment in teaching. englewood cliffs, n.j.: merrill. vihtelic, j. l. (1996). personal finance: an alternative approach to teaching undergraduate finance.financial services review 5,119–131. 221b. grinder et al. / financial services review 8 (1999) 211–221 pii: s1057-0810(01)00081-6 authors allen, douglas e., “hedonic investment,” 9(4): 389– 403 bost, john c., “liquidating a remainder interest: simplifying personal finance,” 9(2): 183–195 brown, stewart l., “design considerations for large public sector defined contribution plans,” 9(3): 197–219 chapman, kenneth s., “risk tolerance and asset allocation for investors nearing retirement,” 9(2): 159–171 cherin, tony, “ liquidating a remainder interest: simplifying personal finance,” 9(2): 183–195 chung, richard, “market timing using strategists’ and analysts’ forecasts of s&p 500 earnings per share,” 9(2): 125–145 domian, dale l.,“risk tolerance and asset allocation for investors nearing retirement,” 9(2): 159–171 downe, edward a., “the effect of country-specific index trading on closed-end country funds: an empirical analysis,” 9(3): 259–277 elder, erick, “social security reform: the effect of investing in equities,” 9(1): 93–107 elder, harold w., “beliefs and actions: expectations and savings decisions by older americans,” 9(1): 33–47 englis, basil, “asset allocation decisions in retirement accounts: an all-or-nothing proposition?,” 9(1): 79–93 fraser, steve p., “strategic asset allocation for individual investors: the impact of the present value of social security benefits,” 9(4): 295–326 greninger, sue alexander, “retirement planning guidelines: a delphi study of financial planners and educators,” 9(3): 231–247 hampton, vickie l., “retirement planning guidelines: a delphi study of financial planners and educators,” 9(3): 231–247 hanna, sherman, “determinates of planned retirement age,” 9(1): 1–17 hariharan, govind, “risk tolerance and asset allocation for investors nearing retirement,” 9(2): 159–171 hearth, douglas, “performance persistence and management skill in non-conventional bond mutual funds,” 9(3): 247–259 hogarth, jeanne m., “consumer information search for home mortgages: who, what, how much, and what else?,” 9(3): 277–293 holland, larry, “social security reform: the effect of investing in equities,” 9(1): 93–107 jacquet, susan, “retirement planning guidelines: a delphi study of financial planners and educators,” 9(3): 231–247 jennings, william w., “strategic asset allocation for individual investors: the impact of the present value of social security benefits,” 9(4): 295–326 jordan, w. john, “stock selection based on morningstar’s ten-year, five-star general equity mutual funds,” 9(2): 145–159 king, david r., “strategic asset allocation for individual investors: the impact of the present value of social security benefits,” 9(4): 295–326 405 financial services review the journal of individual financial management index volume 9, 2000 kitt, karrol a., “retirement planning guidelines: a delphi study of financial planners and educators,” 9(3): 231–247 kryzanowski, lawrence, “market timing using strategists’ and analysts’ forecasts of s&p 500 earnings per share,” 9(2): 125–145 kuhlemeyer, gregory a., “the equity index annuity: an examination of performance and regulatory concerns,” 9(4): 327–342 lee, jinkook, “consumer information search for home mortgages: who, what, how much, and what else?,” 9(3): 277–293 loviscek, anthony l., “stock selection based on morningstar’s ten-year, five-star general equity mutual funds,” 9(2): 145–159 mcginnis, john d., “the asset allocation decision in retirement: lessons from dollar cost averaging,” 9(1): 47–65 mcgoun, elton g., “hedonic investment,” 9(4): 389– 403 miller, edward m., “exploitable patterns in retirement annuity returns: evidence from tiaa/ cref,” 9(3): 219–231 montalto, catherine phillips, “determinates of planned retirement age,” 9(1): 1–17 o’connor, matthew, “the effect of country-specific index trading on closed-end country funds: an empirical analysis,” 9(3): 259–277 olsen, kelly, “social security investment accounts: lessons from participant-directed 401(k) data,” 9(1): 65–79 philpot, james, “performance persistence and management skill in non-conventional bond mutual funds,” 9(3): 247–259 plath, d. anthony, “financial services and the african-american market: what every financial planner should know,” 9(4): 343–359 poole, barbara, “on time: contributions from the social sciences,” 9(4): 375–389 prather, larry j., “exploitable patterns in retirement annuity returns: evidence from tiaa/cref,” 9(3): 219–231 query, j. tim, “an analysis of the medical savings account as an alternative retirement savings vehicle,” 9(1): 107–123 richard, john e., “the information content of closed-end country fund discounts,” 9(2): 171– 183 rimbey, james, “performance persistence and management skill in non-conventional bond mutual funds,” 9(3): 247–259 rudolph, patricia m., “beliefs and actions: expectations and savings decisions by older americans,” 9(1): 33–47 sigrist, kevin w., “design considerations for large public sector defined contribution plans,” 9(3): 197–219. stevenson, thomas h., “financial services and the african-american market: what every financial planner should know,” 9(4): 343–359 trecartin, ralph r. jr., “the reliability of the bookto-market ratio as a risk proxy,” 9(4): 361–373 vanderhei, jack l., “social security investment accounts: lessons from participant-directed 401(k) data,” 9(1): 65–79 waggle, doug, “asset allocation decisions in retirement accounts: an all-or-nothing proposition?,” 9(1): 79–93 wiggins, james b., “the information content of closed-end country fund discounts,” 9(2): 171– 183 woerheide, walt, “the impact of the pension fund on the decision to work one more year,” 9(1): 17–33 yuh, yoonkyung, “determinates of planned retirement age,” 9(1): 1–17 titles “an analysis of the medical savings account as an alternative retirement savings vehicle,” j. tim query, 9(1): 107–123 “asset allocation decisions in retirement accounts: an all-or-nothing proposition?,” doug waggle and basil englis, 9(1): 79–93 406 index / financial services review 9 (2000) 405–407 “beliefs and actions: expectations and savings decisions by older americans,” harold w. elder and patricia m. rudolph, 9(1): 33–47 “consumer information search for home mortgages: who, what, how much, and what else?,” jeanne m. hogarth and jinkook lee, 9(3): 277–293 “design considerations for large public sector defined contribution plans,” stewart l. brown and kevin w. sigrist, 9(3): 197–219 “determinates of planned retirement age,” sherman hanna, catherine phillips montalto and yoonkyung yuh, 9(1): 1–17 “exploitable patterns in retirement annuity returns: evidence from tiaa/cref,” edward m. miller and larry j. prather, 9(3): 219–231 “financial services and the african-american market: what every financial planner should know,” d. anthony plath and thomas h. stevenson, 9(4): 343–359 “hedonic investment,” elton g. mcgoun and douglas e. allen, 9(4): 389–403 “liquidating a remainder interest: simplifying personal finance,” john c. bost and tony cherin, 9(2): 183–195 “market timing using strategists’ and analysts’ forecasts of s&p 500 earnings per share,” lawrence kryzanowski and richard chung, 9(2): 125– 145 “on time: contributions from the social sciences,” barbara poole, 9(4): 375–387 “performance persistence and management skill in non-conventional bond mutual funds,” douglas hearth, james rimbey, and james philpot, 9(3): 247–259 “retirement planning guidelines: a delphi study of financial planners and educators,” karrol a. kitt, sue alexander greninger, susan jacquet, and vickie l. hampton, 9(3): 231–247 “risk tolerance and asset allocation for investors nearing retirement,” dale domian, govind l. hariharan, and kenneth s. chapman, 9(2): 159– 171 “social security investment accounts: lessons from participant-directed 401 (k) data,” kelly olsen and jack l. vanderhei, 9(1): 65–79 “social security reform: the effect of investing in equities,” erick elder and larry holland, 9(1): 93–107 “stock selection based on morningstar’s ten-year, five-star general equity mutual funds,” w. john jordan and anthony l. loviscek, 9(2): 145–159 “strategic asset allocation for individual investors: the impact of the present value of social security benefits,” david r. king, steven p. fraser, william w. jennings, 9(4): 295–326 “the asset allocation decision in retirement: lessons from dollar cost averaging,” john d. mcginnis and premal p. vora, 9(1): 47–65 “the effect of country-specific index trading on closed-end country funds: an empirical analysis,” matthew o’connor and edward a. downe, 9(3): 259–277 “the equity index annuity: an examination of performance and regulatory concerns,” gregory a. kuhlemeyer, 9(4): 327–342 “the impact of the pension fund on the decision to work one more year,” walt woerheide, 9(1): 17–33 “the information content of closed-end country fund discounts,” james b. wiggins and john e. richard, 9(2): 171–183 “the reliability of the book-to-market ratio as a risk proxy,” ralph r. trecartin, jr., 9(4): 361– 373 407index / financial services review 9 (2000) 405–407 pii: 1057-0810(92)90006-x financial services review, 2(2): 11 l-1 30 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. what strategies are experienced estate planners recommending? evidence from survey data chris j. prestopino responses from 124 experienced estate planning attorneys in 33 states reveal the frequency with which their 1991 clients undertook over forty estate planning techniques commonly promoted in practitioner literature. nearly 83 percent of their clients had a net worth of at least $600,000, implying the need for careful consideration of these strategies. among the most popular techniques implemented include the durable power of attorney for property, durable power of attorney for health care, living will, credit shelter bypass plan, standard qtip trust, revocable living trust, and nonsimple will. among the least popular techniques include flower bonds, private annuity, grantor retained unitrust, large custodial gif, and grantor retained annuity trust. i. overview many financial services academics are likely to view estate planning as the most esoteric subject in the financial services curriculum. its practice is dominated by experienced attorneys with whom they share little in common. its subject matter is quite technical, and almost entirely legally related. few financial services academics have had formal education in estate planning; not many have taken an actual estate planning course, either in law school, in a ph.d. program, or in preparing for the cff. finally, financial academics have little convenient access to estate planning literature. while there is much written in the legal and trade journals, little of it appears in the traditional finance publications. although not very well understood, estate planning constitutes a major, integral subject within the area of individual financial management; financial services academics cannot ignore it. estate planning deals with too many financial chris j. restopino l califoinia state university, chico, ca 95929-005 1 112 financial services review, 2(2) 1993 planning-related problems encountered by individuals, particularly those actively planning in anticipation of their eventual death. this paper seeks to help finance academics become more acquainted with the principles of estate planning. by summarizing data obtained from the first survey of its kind, it will shed light on the frequency with which experienced estate planners recommend or use many of the estate planning techniques commonly promoted in practitioner literature. and by describing each of these techniques a bit, it should provide the finance academic with a general overview of estate planning principles, as well as an update of recent developments in the estate planning field. information about the attorneys and their clients any major attempt at tracking estate planning practices must focus on attor neys. while financial planners, accountants, life underwriters and other financial services professionals may offer clients general estate planning recommendations, only attorneys are permitted to render legal advice and, more importantly, only they can draft the documents that implement most estate planning techniques. attorney skills in estate planning vary widely, from the modest ability of the general practitioner who drafts a few simple wills a year, to the careful work of the expert, skilled in tax law, who limits his or her practice solely to estate planning matters and counsels many wealthy clients each year. this study focuses on attorneys more likely to fall in the latter, expert category. all are authors of estate planning articles in trade journals. in fact, that’s where their names were obtained: from the previous ten year’s issues of the industry’s two leading national trade publications: estate planning and trusts and estates. many of these attorneys have also written other professional articles and have spoken at leading estate planning institutes such as nyu, usc, miami, ucla, practicing law institute, southern california tax and estate planning forum, and notre dame. four-hundred-two attorney-authors were mailed a 59 question survey during a month long period in december, 199 1 and january, 1992. one hundred thirty nine responded (35 percent), and 124 responses were found usable. all questions requested information about their practices during the preceding year, which for most, represented all of 1991. the respondents come from a total of at least 33 states and the district of columbia: california: 24 attorneys; new york: 18; illinois: 11; missouri: 7; texas: 6; massachusetts and pennsylvania: 5; colorado, georgia and michigan: 4; maryland: 3; two attorneys from: district of columbia, north caro lina, nevada, ohio, oregon, rhode island, and virginia; and one each from: arkansas, connecticut, delaware, florida, hawaii, iowa, mississippi, montana, new jersey, south carolina, south dakota, tennessee, vermont, wisconsin, and west virginia. four attorneys did not reveal the state in which they conduct their main practice. table 1 summarizes other, miscellaneous data about these attorneys. the actual survey is reproduced in the appendix. what strategies are esme planners recommending? 113 table 1 miscellaneous attorney and client data interquartile range range mean std. dev. attorney data: years experience as ep attorney age % of entire practice is planning and drafting % of entire practice is planning, drafting, and administration % of planning, drafting, and administration practice is planning and drafting no. partners and associates in firm, including themselves client data: total number of clients in last 12 months number of clients with net worth greater than $600,000 % of clients with net worth less than $600,000 % of clients with net worth between $600,000 and $1,200,000 % of clients with net worth greater than $1,200,000 total of clients 2z l-100% lo-100% .5-100% l-600 3-300 3-275 o-85% o-80% o-100% 13-25 19.0 8 40-52 46.6 10 37-72% 54.0% 23% 50-100% 27% 60-90% 73.9% 73.5% 19% s-130 40-108 28-100 s-25% 14-40% 30-80% 90.7 90.0 73.2 17.2% 26.3% 56.5% 100.0% 130 70 58 19% 18% 28% years in practice measured by number of years in practice, this is truly a group of experienced estate planners. the average attorney has spent nineteen years in the field. the range extends from four years to 40 years, and the interquartile (iq) range is 13 to 25 years. ages of the respondents range from 29 to 68 (iq range: 40-52), with a mean of 46.6. ten percent are female. degree of specialization measured by degree of specialization, these planners also show great experi ence. before discussing the data, it may help to describe what skilled estate planning entails. speaking narrowly, the practice of estate planning involves planning and drafting wills, trusts, powers of attorney and other legal documents. however, in part to meet the additional needs of their clients, experienced estate planners must do more than just planning and drafting. first, as a direct consequence of and adjunct 114 financial services review, 2(2) 1993 to their planning, they will agree to administer trusts and estates before and after the client’s death. unlike estate planning, this administrative work largely involves more routine paperwork and simple financial management in the handling of a client’s legal affairs. second, estate planning attorneys also practice in the area of business law, since many of their clients have acquired the bulk of their wealth through business pursuits. finally, because most estate planning strategies are at least partly tax driven, they practice tax law, preparing tax returns, giving tax advice and representing clients before the i.r.s. and in court. thus, while estate planning attorneys can be expected to specialize in planning and drafting, they are also likely to engage in other related work. the group in this sample is no exception. sixty-six percent devote at least half of their entire practice time to the specific work of planning and drafting. the average respondent devotes 54.0 percent of total work time to these two challenging tasks. ninety-six percent also practice probate and trust administration. combined, the typical respondent works 73.9 percent of the time in estate planning, drafting, and administration, and devotes nearly three quarters (73.5%) of this combined time in planning and drafting, while about one quarter (26.5%) in administration. the remaining 26.1 percent of the typical attorney’s time is spent in other, mostly related areas. for example, seventy six percent devote some of their practice to at least one of the following areas: business law (28% of attorneys), taxation (21%), general practice (7%), and family law (4%). the clients they advised this survey asked several questions about number of clients advised and estimated client net worth. table 1 also summarizes client data. number of clients advised. most respondents gave estate planning advice to a sizable number of clients during 1991. the group’s average number of estate planning clients was 90.0, and three fourths of these attorneys advised at least 40 clients. unfortunately, the survey neglected to ask respondents to treat couples as one client, thus making the figures somewhat ambiguous. while some attorneys probably treated couples as two clients, others likely assumed them to be one. the upshot is somewhat of a bias in the estimate of number of clients advised. client net worth. as one would expect, most of this group’s clients are quite affluent. for any one attorney, the number of clients whose net worth exceeds $600,000, putting them in a position to have a potential federal transfer tax liability, ranged from three to 275 (i.q. range: 28-loo), with the average attorney having counseled 73 such clients. in terms of percentages, a mean of 17.2 percent of all clients were reported to have a “smaller” estate size (net worth of less than $600,000), 26.3 percent had a “medium” size estate ($600,000 to !$1,200,000), and 56.5 percent had a “larger” estate (greater than $1,200,000). thus, nearly 83 percent of all clients advised by the average surveyed attorney had a net worth of more than what strategies are estate planners recommending? 115 $600,000, large enough to need careful death tax planning help. three quarters of all attorneys surveyed had at least 30 percent of their clients having a net worth of more than $1.2 million, an amount large enough to consider lifetime transfer strategies to further reduce the death tax. summarizing, this group of attorneys has a clientele generally well off enough to require consideration of all of the estate planning techniques popularly discussed in the literature. ii. basic document preparation overview individuals commonly employ four distinct general methods of transferring property to their survivors: joint tenancies, simple will, non-simple will, and revocable living trust. each will be described briefly. joint tenancies title by joint tenancy with right of survivorship is a relatively automatic and inexpensive method of property disposition. however, it has the disadvantages of lack of flexibility, uncontrolled disposition, likely estate administration at the death of the surviving cotenant, premature loss of asset control, and possible greater income, gift and estate taxes. for these and other reasons, most estate planning attorneys are well known not to recommend joint tenancies except for a few of their least wealthy clients. simple will a simple will is a written will prepared for a family having a relatively small estate where death tax planning is not a significant concern. it usually leaves all property outright to the surviving spouse, if alive, otherwise outright to the children. it provides for no trusts, and contains no provisions designed to save taxes. it is more flexible and usually less tax costly than joint tenancies, and is frequently prepared for less wealthy clients, particularly by more inexperienced attorneys, who may (ill-advisedly) use it for other, wealthier clients, as well. nonsimple will a non-simple will is a longer, more carefully constructed will prepared by more skilled attorneys and designed especially for wealthier clients facing a potential death tax liability. it includes many provisions designed to both minimize taxes and arrange for more involved, non-outright distributions of property. in most situations, it provides for the creation of one or more trusts at the client’s death. 116 financial services review, 2(2) 1993 revocable living trust finally, a revocable living trust is often established during the client’s lifetime, primarily to minimize the cost of estate administration during incapacity and at death. its dispositive provisions are usually similar to those of the nonsimple will, and, after the client’s death, it can be just as effective as the nonsimple will. current use of these four transfer devices the attorneys in this survey were asked to describe what percentage of their clients used each of these four planning arrangements as the principal method of asset disposition. their responses are summarized in table 2. joint tenancies. for joint tenancies, no attorney indicated that greater than 50 percent of his or her clients used joint tenancies as the principal method of asset disposition. slightly less than one half (47 percent) had no clients using them, and only 25 percent (highest quartile) reported that greater than five percent of clients used them. the typical attorney reported 4.5 percent of clients using joint tenancies. this evidence supports the general belief that most experienced estate planning attorneys hardly ever recommend joint tenancies as the principal method of asset disposition. simple will. given the experience level of these attorneys and the sizable wealth of most of their clients, it is no surprise that most don’t recommend the simple will much, either. although 88 percent did draft the simple will for at least one client, the average attorney used it for only 13.8 percent of his or her clients. combining the figures for both joint tenancy and simple will, 18.3 percent of the average respondent’s clients used joint tenancies or simple wills. this closely relates to the finding mentioned above that 17.2 percent of the average attorney’s clients have an table 2. response to question: “whatpercentage of (all of your clients) used each of the following planning documents as the principal method of asset disposition?” interquartile % not using planning document range range mean at all std. dev. joint tenancies o-50% o-5% 4.5% 47% 11% simple will o-80% 5-15% 13.8% 12% 16% non-simple will o-100% lo-&o% 36.6% 9% 30% revocable living trust o-100% 15-75% 45.1% 8% 32% total of clients 100.0% what strategies are estate phaners recommending? 117 estate size small enough not to need tax planning. joint tenancies and the simple will are used more frequently for these clients. nonsimple will and living trust. the data confirms that these two complex instruments represent the experienced estate planning attorney’s primary asset disposition vehicles. the typical attorney in this survey used both documents combined for 81.7 percent of clients, with nonsimple wills prepared for an average of 36.6 percent of clients and living trusts an average of 45.1 percent. interestingly, while attorneys are generally perceived to be strongly preferential to one of these documents over the other, nine percent or less of the respondents to the survey indicated they used only one of the two of these exclusively for all wealthier clients. only nine percent did not draft any nonsimple wills and only eight percent did not draft any living trusts. thus, while most attorneys probably prefer to recommend one of these documents over the other, most will draft the less preferred one at least occasionally. apparently, most of these attorneys either unbiasedly select the instrument that more closely meets the client’s needs, or are in fact biased towards one or the other instrument, but nonetheless at least occasionally yield to the wishes of their more opinionated clients. iii. death tax planning with marital deduction and bypass provisions several questions were asked to reveal the popularity of alternative deathtime transfer techniques commonly used to reduce death taxes. the discussion that follows will assume planning for a married couple. in determining the preferences of attorneys and clients in this general area, the survey focused on two major aspects: bypass planning, which addresses the amount of property left to parties other than the surviving spouse (called the nonmarital share), and marital deduction distribu tion planning, which deals with the manner in which the marital share is left to the surviving spouse. in the discussion below, each is first described in general, and then analyzed in terms of the survey results. bypass planning bypass planning seeks to reduce death taxes by arranging for the first spouse to die to leave an amount of property to a nonspouse so as to reduce taxation at the surviving spouse’s later death. this property is said to ‘bypass’ the surviving spouse’s estate. overview of bypass planning in general, bypass estate planning offers three alternative deathtime transfer choices for married couples. they differ in the amount left to the nonmarital, bypass share: 1) no bypass (all to surviving spouse); 2) $600,000 to the bypass share (and 118 financial services review, 2(2) 1993 the rest to the surviving spouse); and 3) an amount greater than $600,000 to the bypass share (and the rest to the surviving spouse). the first, no-bypass strategy takes total (and unnecessary) advantage of the unlimited estate tax marital deduction, under which a spouse can transfer an unlimited amount of property, free of federal death tax, to the surviving spouse. for spouses selecting this alternative, a relatively large tax hit will occur at death of the second spouse, when the entire family estate, which is now owned by the surviving spouse, will be taxed at marginal rates effectively starting at 37 percent and rising to as high as 55 percent, for estates exceeding $2.5 million. the second strategy, the perceived most popular alternative to leaving every thing to the surviving spouse, has the first spouse to die leaving all but approximately $600,000 to the surviving spouse. as implied above, property passing to the bypass share is said to bypass the surviving spouse’s estate because it will escape estate taxation at his or her later death. a bypass of $600,000 is called a credit shelter bypass because $600,000 is the amount of the exemption equivalent of the federal unified credit; the first spouse to die can leave the first $600,000 tax free to anyone other than the spouse because every estate is entitled to a tax credit that effectively exempts this amount from tax. the third strategy makes the bypass share an amount greater than $600,000, and can often further reduce total estate taxes, despite the tax on the estate of the first spouse to die. the lower tax at the second death may more than offset the higher tax at the first death, particularly for estates that appreciate rapidly. regardless of its size, the bypass share is typically left in trust rather than outright, for the benefit of both the spouse and the children. the spouse is given a lifetime interest in all of the bypass trust income and power to invade the principal in certain events, while the children are given the remaining principal at the surviving spouse’s death. use of these bypass strategies the data on the use of these strategies is summarized in table 3. attorneys were asked to answer these questions only with reference to those clients sufficiently wealthy to have a potential estate tax liability. the typical respondent indicated that only 5.8 percent of wealthier clients preferred strategy 1, all to spouse with no bypass. seventy-five percent of these attorneys reported that no more than five percent of their clients chose it. in fact, more than one half of the attorneys had no clients at all choosing it. on the other hand, twenty one attorneys had at least 10 percent of their clients selecting this strategy, with the median attorney in this subset using it for 20 percent of clients, and two attorneys using it for 75 and 90 percent, respectively. the upshot is a skewed, somewhat bipolar distribution, which explains why the mean of all respon dents exceeds the third quartile level. two different groups are believed to represent this distribution: a large number of discriminating experts who use strategy 1 only what strategies are estate planners recommending? 119 table 3. use of bypass planning strategies (percentage of those clients having a potential estate tax liability) interquartile % not using planning strategy range range mean at all std. dev. no bypass &90% o-5% 5.8% 51% 14% $600,000 bypass lo-100% 85-l 00% 88.1% 0% 18% larger bypass &50% o-io% 6.0% 50% 10% total of clients 100.0% for smaller estates, and several other, less experienced attorneys who use it for far more clients. of the remaining 94 percent of clients advised by the typical attorney, 88.1 percent (out of 100 percent) selected strategy 2, credit shelter bypass of approxi mately $600,000 (i.q. range: 85100%). the other 6.0 percent chose strategy 3, a bypass exceeding $600,000 (i.q. range: o-10%). summarizing, planners are recommending strategy 2, the credit shelter bypass plan, to the overwhelming majority of their clients. this confirms the perceived general attitude widely held by the estate planning community that strategy 1 is too tax costly to elect, and strategy 3 is not popular because most clients do not want to incur a death tax during either their or their spouse’s lifetimes. marital deduction and distribution planning as implied above, selection of a bypass amount more or less automatically determines the amount passing to the surviving spouse. how the marital share can be distributed to that spouse is the subject of the next discussion. overview of planning for the marital share clients can dispose of the marital share in several different ways and still qualify for the estate tax marital deduction. they can leave it outright. or they can leave it in one of several trusts, of which the most common are the qtip trust and the power of appointment trust. both of these trusts require the trustee to distribute currently all trust income to the surviving spouse for life. the unique characteristic of the qtzp trust is its ability to totally deprive that spouse from exercising any control over use or disposition of the trust principal. the power of appointment trust is unique for just the opposite reason: it can extend to the surviving spouse an unlimited power to dispose of trust principal to anyone during his or her lifetime or at death. other less used marital dispositions include the estate trust, qualified domestic trust and a generation-skipping type of qtrp trust. 120 financial services review, 2(2) 1993 table 4. disposition of marital share to surviving spouse (percentage of those clients having a potential estate tax liability) interquartile % not using planning strategy range range mean at all std. dev. outright o-80% 5--50% 28.3% 19% 24% standard qtip trust o-loo% 20-60% 43.5% 7% 26% power of appointment trust o-loo% o-14% 14.0% 40% 24% estate trust o-90% o-o% 2.9% 85% 12% qualified domestic trust o-20% o-5% 3.3% 34% 4% reverse qtip election trust o-loo% 5-35% 24.3% 27% 27% any two marital trusts o-loo% o-30% 19.9% 26% 25% any three marital trusts o-80% o-o% 3.2% 22% 11% any four marital trusts o-loo% o-o% 1.5% 92% 11% attorney practices in disposing of the marital share attorneys were asked about the frequency with which their wealthier clients used these dispositions of marital deduction property. the results are shown in table 4. the most popular disposition was to a standard quip trust, with the typical attorney incorporating it in the plans of 43.5 percent of clients (i.q. range: 20-60%). twenty-eight point three percent of the typical attorney’s clients provided for an outright disposition to the surviving spouse (i.q. range: 5-50%), while 14.0 percent incorporated a power of appointment trust (i.q. range: &14%), 2.9 percent an estate trust and 3.3 percent a qualified domestic trust. in addition, 24.3 percent of clients used a generation-skipping qtip trust incorporating a “reverse election” (i.q. range: 5-35%). finally, since it is possible to include more than one marital disposition in a single plan, the survey asked about that and discerned that the typical attorney included two of these marital trusts in 19.9 percent of client estate plans (i.q. range: c&30%), three of these trusts in 3.2 percent of client plans, and four of these trusts in 1.5 percent of plans. in view of the fact that efficient planning for the generation-skipping transfer tax often requires the use of multiple marital trusts, it is somewhat puzzling to learn that these experienced planners did not use at least two marital trusts in larger numbers. iv. postmortem tax and liquidity planning the survey asked several questions relating to other deathtime planning techniques. attorneys were asked to answer these questions only with reference to those decedent-clients for whom a federal estate tax return was filed during the preceding twelve months. the results are shown in table 5. what strategies are estate pkmners recommending? 121 table 5. postmortem tax and liquidity planning (number/percentage of clients for whom a federal estate tax return wasjilid) interquartile % not using range range mean at all std. dev. o-so 4-10 8.8 n/a 8 number of estate tax returns filed by anyone estate tax returns actually prepared by respondent or respondent’s firm flower bonds section 6 166 deferral section 2032a valuation alternate valuation date o-100% 22.5-100% 68.5% 11% 41% o-20% o-o% .6% 94% 3% o-100% o-10% 10.0% 50% 20% o-100% o-o% 2.4% 85% 10% o-67% o-20% 10.7% 50% 15% federal estate tax return preparation a federal estate tax return must be filed for any decedent whose taxable base exceeds $600,000. one hundred seven out of 120 attorneys (89 percent) reported having at least one client who died during the previous twelve months for which an estate tax return had to be filed. the number of returns filed by either the attorney, the firm, or others, ranged from one to fifty (i.q. range: 4 to 10 returns). the typical attorney had 8.8 decedents for which returns were filed, and that attorney or his or her firm actually filed 68.5 percent of these returns. flower bonds only seven out of 122 attorneys (6%) reported purchasing flower bonds for their tax return decedents. flower bonds are certain u.s. treasury bonds purchased on behalf of the client shortly before death at discount and redeemed at face value in payment of the estate tax. during current periods of low interest rates, flower bonds are not attractive because their purchase discount is minimal. during 1991, that discount averaged only about five percent. section 6166 deferral section 6166 offers qualifying estates of closely held business-owning dece dents a method of paying the estate tax in 10 annual installments beginning four years after death. forty-six percent of attorney respondents indicated they elected section 6166 deferral for at least one of their estate tax return decedents, and eighteen percent so elected for twenty or more percent (iq. range: election made for o-10 percent of clients). the typical attorney made this election for 10.0 percent of this subset of clients. 122 financialservicesreview,2(2) 1993 section 2032a special use valuation section 2032 special use valuation enables the estate to value, for death tax purposes, the real estate of a farm or other qualifying business at its present use value (e.g., as a family farm business) rather than at its highest and best value (e.g., as property suitable for hi-rise development). only fifteen percent of responding attorneys reported electing section 2032a special use valuation for any decedents. and it appears that most of these fifteen percent filed for no more than one or two decedents. alternate valuation date election the federal estate tax return preparer may value assets as of date of death or as of alternate valuation date, which is exactly six months after date of death. one-half of the attorneys reported not ever electing the alternate valuation date for any of their decedent-clients during the most recent twelve months. for the other 50 percent, the range so electing was from one percent of their decedents to two-thirds, with a median of twenty percent of their clients. in view of the rapidly appreciating 1991 stock market, it is understandable that for many estates the later (alternate valuation) date would probably be rejected because it probably generated higher taxable values. v. plann~gforcloselyheldbusinesses the survey asked several questions relating to planning for closely held business clients. table 6 summarizes the results. the average attorney surveyed counseled 20.5 closely held business estate planning clients during the year. s corporation election on average, of these business clients, the typical attorney made a subchapter s corporation election for an average of 23.3 percent of them, or approximately five table6. closely held business planning (number/percent of closely held business clients) interquartile % not using planning strategy range range mean at all std. dev. number of c.h.b. clients o-100 7.5-25 20.5 n/a 20 subchapter s election o-100% o-40% 23.3% 35% 30% estate freezing recapitalization o-100% o-o% 4.0% 78% 12% partnership capital freeze o-50% o-o% 2.1% 85% 7% business buyout agreement o-100% 10-40% 29.3% 14% 29% what strategies are estate planners recommending? 123 clients. an s election is often undertaken so that the corporate business owner is able to enjoy lower individual tax rates and to be able to deduct business losses on his or her individual tax return. estate freezing recapitalization an estate freezing recapitalization is a recapitalization of corporate stock under which the client receives shares (such as preferred stock) that are not expected to appreciate substantially, and the client’s younger generation beneficiaries, typi cally the children, receive common stock, which is expected to appreciate rapidly. in this way, the client may be able to “freeze” the death tax value of the business at its current value. twenty-five out of those 113 respondents answering the question (22 percent) indicated that they implemented at least one corporate estate freezing recapitaliza tion. of these, the median percent of these clients doing this was 10 percent, meaning that on average, only about two business estate planning clients (10 percent of 20.5 business clients) undertook a recap for each of these 25 attorneys. the results are about the same for partnership capital freezes, with only 17 out of 113 attorneys (15 percent) undertaking at least one and the typical attorney in this small group doing about two of them. the passage of the greatly inhibiting internal revenue code section 2701, effective october 8, 1990, explains in part why these percentages are so low. actually, it is surprising to learn that some attorneys continued to do even that many freezes after that date. business buyout agreements a business buyout agreement is a contract obligating another party to purchase the decedent-client’s interest in a business at his or her death, or at some other event which could precipitate the client’s withdrawal from the firm. advantages include liquidity and a guaranteed market. ninety-five out of 111 attorneys (86 percent) had at least one client undertaking a business buyout agreement, and the average attorney arranged a buyout for about six clients (29.3 percent of 20.5 business clients). vi. lifetime transfers respondents were asked to indicate the percentage of all of their estate planning clients that undertook the following lifetime transfers in the past year. table 7 summarizes the results. noncharitable outright annual exclusion gift of $10,000 this is a gift tax-free transfer often undertaken by wealthier clients to reduce future death taxes, and by many clients simply to assist a child or other family member. no gift tax return need be filed if total gifts in a calendar year to any one 124 financial services review, 2(2) 1993 table 7. lifetime transfer planning cpercentuge of all clients) interquartile % not planning strategy range range mean using at all std. dev. nonchar. outr. $lok gift o-100% loao% 29.2% 2% 22% nonchar. custod. $lok gift o-95% 2-15% 12.1% 24% 17% nonchar. outr. >$ 10k gift o-25% o-10% 5.5% 24% 6% nonchar. custod. >$lok gift o-50% o-o% 1.3% 83% 5% nonchar. gift in trust o-90% 5-30% 19.1% 17% 18% irrevocable life insurance trust o-80% 10-30% 22.4% 7% 17% installment sale o-40% o-3% 2.1% 65% 5% private annuity o-10% o-o% .6% 85% 2% grantor retained annuity trusts o-20% o-2% 1.6% 68% 3% grantor retained unitrust o-15% so% .8% 80% 2% joint purchase o-25% o-o% .4% 92% 2% char. outr. gift at least $5k o-80% l-15% 11.5% 25% 15% char. split interest gift o-35% 2-5% 4.9% 39% 7% donee do not exceed $10,000, the amount of the annual gift tax exclusion. all but two attorneys had clients undertaking these gifts, with the typical attorney having 29.2 percent of clients making them (i.q. range: 10-40%). noncharitable custodial annual exclusion gift of $10,000 custodial gifts are irrevocable gifts to a custodian for the benefit of another person, usually a child, and are designed to fall under the provisions of the client’s state’s uniform gifts to minors act or uniform transfers to minors act. such gifts are much less complex than trusts to create and administer, but are far less flexible. ninety-three out of 123 attorneys who answered the question (76 percent) indicated they had at least one client making such a custodial gift, with the average attorney having 12.1 percent of clients making them (i.q. range: 2-b%). noncharitable outright gift exceeding the $10,000 annual exclusion large outright gifts are made by wealthier clients primarily to freeze their death taxable estate. seventy-six percent (94 attorneys) had at least one client making these gifts, with the typical attorney having 5.5 percent of clients making them (i.q. range: l-10%). noncharitable custodial gift exceeding the annual exclusion only twenty one attorneys, or 17 percent, had clients undertaking larger custodial gifting, and of these few attorneys, the percentage of total clients involved was about five percent. the mean for all attorneys was 1.3%. what strategies are estate phznners recommending? 125 noncharitable gifts in trust one hundred two of the attorneys (83 percent) had clients making noncharitable gifts in trust, with the typical attorney having about 19 percent of clients making them (i.q. range: 5-30 %). for these gifts, an average of sixty percent of the trusts included a crummey demand provision, while an average of six percent included a section 2503(b) provision, and 13 percent a section 2503(c) provision. these provisions am designed to qualify a future interest gift to a minor for the $10,000 annual gift tax exclusion. irrevocable life insurance trust an irrevocable life insurance trust is a trust which is owner and beneficiary of a policy on the life of the client. after the client’s death, the proceeds are retained by the trust for the benefit of the client’s spouse and children. its major advantage is exclusion of the policy proceeds from the death tax estate of both the client and the client’s spouse. one hundred fifteen attorneys (93 percent) helped clients draft these popular trusts, with the typical attorney arranging them for 22.4 percent of clients (i.q. range: lo-30%). installment sale by transferring appreciating property, an installment sale can offer the client a constant income stream for a fixed period and reduce death taxes. forty-three attorneys (35 percent) reported having clients who undertook an installment sale. of these, the average attorney reported an installment sale completed by five percent of their clients (for all attorneys: i.q. range: o-3%; mean: 2.1%). private annuity a private annuity involves the sale of an asset, usually to a family member, in exchange for the right to an annuity for life. it is a specialized technique, designed for a specific set of facts which usually apply to only a few estate planning clients. only 19 attorneys (15 percent) reported any clients engaging in a private annuity, and of these, the average was 2 percent of clients (mean for all attorneys: .6%). grantor retained trusts the next two lifetime transfers are examples of grantor retained trusts, in which the client creates and transfers property into an irrevocable trust, retaining a specific amount of income for a period of years. if the client dies within the period, the trust assets revert to the client’s estate and the trust will have not helped reduce the death tax. on the other hand, if the client survives the period, the assets vest in the “remaindermen,” typically the client’s children, and the clients death tax gross estate will not include the date of death value of the assets. thus, these transfers in trust are designed to freeze a portion of the client’s estate, and they must conform 126 ftnancialservicesreview,2(2) 1993 to the detailed rules under internal revenue code section 2702, which became effective in october, 1990. grantor retained annuity trust (grat) a grantor retained annuity trust is a grantor retained trust paying the client a fixed dollar annuity amount. forty attorneys (32 percent) had clients undertaking a grat, and of these, the typical number of clients was about five percent (for all attorneys: i.q. range: o-2%; mean: 1.6%). grantor retained unitrust (grut) a grantor retained unitrust is a grantor retained trust paying the client an annuity of a faed percentage of the current fair market value of the trust property. only twenty-six attorneys (20 percent) had clients undertaking a grut, and of these the typical number of clients was about two percent (mean for all attorneys: 0.8%). joint purchase of a life estate/remainder interest surprisingly, in spite of recent devastating tax legislation, 10 attorneys (8%) reported clients undertaking a joint purchase, and of these, the median number of clients was three percent. outright charitable gift of at least $5,000 most attorneys (75 percent) had at least one client making large outright charitable gifts, and of these, the average attorney had ten percent of his or her clients making this type of transfer (for all attorneys: i.q. range: l-15%; mean: 11.5%). charitable split interest gift common types of charitable split interest gifts include the charitable remain der annuity trust, charitable remainder unitrust, pooled income fund, and charitable lead trust. seventy-five attorneys (61 percent) reported clients making split interest gifts to charity, and of these the typical number of clients was five percent (for all attorneys: i.q. range: lo-30%; mean: 4.9%). vii. planningforincapacity estate planning offers two specialized non-trust techniques for helping clients plan for their own incapacity. both involve a variation of the general legal document called durable power of attorney. one variety is designed for management of the client’s property, while the other promotes efficient medical care decision making. table 8 summarizes the results. what strategies are estate planners recommending? 127 table 8. planning for incapacity @ercentuge of all clients) interquartile % not planning strategy range range mean using at all std. dev. dur. pwr. att’y. for property s-100% 7s100% 79.8% 0% 26% dur. pwr. att’y. for health care o-100% 75-100% 81.0% 1% 24% living will o-100% o-90% 51.3% 29% 42% durable power of attorney for property the durable power of attorney for property authorizes another person as agent to make financial decisions on behalf of an incapacitated client. the typical attorney surveyed prepared durable powers of attorney for property for 79.8 percent of clients. every single one of the attorneys drafted them, and a full three quarters of respondents prepared them for at least seventy five percent of clients. durable power of attorney for health care and living will a durable power of attorney for health care authorizes another person to make medical decisions on behalf of an incapacitated client. it often also includes a statement describing certain “heroic” life sustaining procedures the client does not wish to be undertaken. this statement can also be included in a separate document, called a living will. the figures on frequency of planning with the durable power of attorney for health care were nearly identical to those for the durable power of attorney for property, suggesting that health care powers have become nearly universally ac cepted. the typical attorney prepared these durable powers for 81 .o percent of clients and three quarters of the respondents drafted them for 75 percent of clients. along with this durable power, most attorneys surveyed were found to also draft a separate living will for clients. only 29 percent of attorneys indicated they did not draft separate living wills for any clients. those who did typically prepared them for nearly 88 percent of their clients 51.3%). (for all attorneys: mean = viii. conclusions this article has described the major techniques used by experienced estate planning attorneys, in the context of a summary of the results from a recent survey revealing the frequency with which each technique has been used. financial services academics may now have a better understanding of them, and a greater aware ness of the relative importance of each of these techniques to individual financial management. 128 financialservicesreview,2(2) 1993 appendix survey questionnaire surveyofestateplanningattorneys conducted by chris j. prestopino, ph.d. professor, california state university, chico for each question, please place in the appropriate space a number from 0 to 100, representing the approximate percentage of your clients whom you advised in the last year use each of the following strategies. estimates are acceptable. your response is anonymous. all identifying information will be held in the strictest confidence. results will only be reported for groups. thanks for sharing your valuable time. a: basicdocumentpreparation how many estate planning clients have you advised in the past year? clients what percentage of these clients use each of the following planning documents as the principal method of asset disposition? (note: all four answers should total 100%) 1. joint tenancies (whether recommended or not)? -% 2. simple will (no trusts; if both spouses alive, then conditional outright disposition between them, otherwise to children, etc.)? % 3. non-simple will (any other type of non-pourover will incorporating trusts or other more complex arrangements)? -% 4. revocable living trust with pourover will? -% should total-------100% b: lifetime transfers what percentage of all your estate planning cli ents advised in the past year undertook the following lifetime transfers? 1. noncharitable outright annual exclusion gift of $lo,ooo? 2. noncharitable custodial annual exclusion gift of $lo,ooo? 3. noncharitable outright gift exceeding the annual exclusion? 4. noncharitable custodial gift exceeding the annual exclusion? 5. noncharitable gift in trust? for these trusts, how often did they contain: a. ac rummey demand provision? b. a 2503(b) provision (mandatory annual income distribution)? c. a 2503(c) provision (mandatory corpus distribution at age 21)? 6. installment sale? 7. private annuity? 8. grantor retained annuity trust (grat)? 9. grantor retained unitrust (grut)? 10. joint purchase of a life estate/remainder interest? 11. outright charitable gift of at least $s,ooo? 12. charitable split interest gift 13. irrevocable life insurance trust what strategies are estate planners recommending? 129 c: maritaldeduci’ionandbypassplannin g how many of your cli ents advised in the past year were sufficiently wealthy to have apotentia estate lax liability? -clients 1. what percentage of these wealthier clients will use each of the following bypass planning strategies, in planning for the death of the first spouse. (note: all three answers should total 100%) a. no bypass (i.e., unlimited marital deduction)? b. a bypass of approximately $600,000? c. a bypass larger than $600,000? should total-- 2. what percentage of these wealthier client’s documents had a provision for a disclaimer to potentially increase the bypass? 3. with regard to the property qualifying for the marital deduction, what percentage of these wealthier client’s documents incorporate the follow ing marital dispositions? a. outright disposition of bulk of marital deduction property to surviving spouse? b. power of appointment marital trust? c. marital estate trust? d. standard q-tip marital trust? e. generation-skipping q-tip trust with “reverse election”? f. qualified domestic trust (qdtfor non-citizen spouses)? g. any two of the above marital trusts in the same plan? h. any three of the above marital trusts in the same plan? i. any four of the above marital trusts in the same plan? -% -70 d: other tax and liquidity plannin g in the past year, how many federal estate tax returns were filed (by you or others) for your deceased clients? _returns for what percentage of these tax return decedents did you or your office undertake the following? 1. purchase flower bonds? 2. elect section 6166 deferral? 3. elect section 2032a special use valuation? 4. actually prepare the federal estate tax return (you or your firm)? 6. of the federal estate tax returns you prepared or reviewed, approxi mately what percentage elected the alternate valuation date? -% -% -% -% -% e: planning for closely held businesses how many closely held business estate planning clients have you have had in the last year? clients for what percentage of these closely held estate planning clients did you imple ment the following arrangements? 1. s corporation election? -% 2. corporate estate freezing recapitalization? -% 3. partnership capital freeze? -70 4. business buyout (buy-sell) agreement? -% a. for what percentage of these buyout plan clients did you use the following? (note: should total 100%): 1) a cross purchase agreement? -70 2) an entity agreement? -% should total-------loo% 130 financial services review, 2(2) 1993 f: planning for incapacity in the past year, for what percentage of all your estate planning clients did you prepare the following documents? 1. durable power of attorney for proper@? -% 2. durable power of attorney for health care? -% 3. living will (as a separate document-not as part of a durable power of attorney for health care)? -% g: client profile during the past year, what percentage of all your es tate planning clients fell into each of the following three ltet worth categories? (note: combine husband and wife’s assets. all three questions should total 100%): 1. smaller estate (under $600,000)? -% 2. medium size estate ($600,000 to $1,200,000)? -% 3. larger estate (greater than $1,200,000)? -% should total-------100% h: personal information (will be held in strict confidence) 1. state in which you conduct your main practice _ 2. your age: _ 3. your gender (circle one): f m 4. number of years experience as an estate planning attorney: years 5. percent of your entire practice devoted to estate planning and probate: -% 6. your other most frequently practiced area (check only one): -none -family law -business law -personal injury general practice -other (please specify:) 7. of the time you devote to estate planning and probate, what percentage is spent doing estate planning, rather than probate: -% 8. total number of partners and associates in your tirm (include yourself): i: comments this survey may have failed to capture some of your important thoughts about the practice of estate planing. if so, please feel free to comment be low: pii: 1057-0810(95)90017-9 financial services review, 4(l): 41-56 copyright 0 199.5 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. analysis of u.s. savings bonds tom l. potts william reichenstein u.s. savings bonds are cum~iex con#r~c~. ~in~nc~l piu~ners have tr~itio~ally paid tittie attention to savings bonds, in part because they often o$er below-market interest rates. however, they sometimes offer above-market interest rates, especially when one learns how to view and value theiroptionfentures. all savings bonds contuinputoptions that protect the investor against a rise in interest rates. thus, they can be viewed as short-term, intermediate-term, or long-term bonds. the ee bonds also offer several tax options. we show how savings bonds can be used to beat the kiddie tax. to finance postsecond~~ education, and in retirement piann~ng. i. insi~~~dwti~n u.s. savings bonds are complex contracts. today, the treasury issues ee-series and hh-series savings bonds. both series offer put options that protect the investor against the risk of rising interest rates. the ee bond also offers a number of tax options. there has been little prior analysis of savings bonds. this lack of attention has not been due to the size of the market-at year-end 1992, individuals held $169.4 billion of u.s. savings bonds. rather, the lack of attention probably reflects their complexity and the fact that they often offer below-market interest rates. however, interest rates on savings bonds are administered, and administered rates sometimes result in above-market yields. the administered interest rates on savings bonds have occasionally exceeded market rates available on comparable treasury securities. for example, from october 1992 through february 1993, the ee bond offered a minimum yield of 6% if held for five years, while five-year, coupon-bearing treasury notes offered yields below 6%. in early 1993, several financial articles made this comparison. the rush to buy the above-market 6% savings bonds was so strong that at the end of february, the treasury stopped issuing the 6% ee contracts. funds invested in the 6% ee bonds continue to receive 6%. beginning in march 1993, new money invested in savings bonds had to settle for a 4% promised yield. the financial press and most financial planners failed to realize that the 4% yield still represented an above-market yield for some investors. savings bonds contain a tom l. potts and william reichenstein l hankamer school of business, baylor university, waco, tx 767988000. 42 financial services review 4(l) 1995 put option that allows investors to redeem the bonds at any time at a present price. thus, a savings bond can be thought of as a short-term bond, an intermediate-term bond, or a long-term bond. for example, since the savings bond could be redeemed in six months at par plus 2% interest, the ee bond is also a six-month security offering a 4% annual yield. until april 1994, yields on six-month marketable securities were consistently below 4%, often below 3%. yet, few financial planners encouraged their clients to buy 4% ee bonds instead of lower-yielding near-cash assets. we believe that most financial planners view savings bonds as a long-term security. the above example shows that learning to view a savings bond as a short-term or intermediate term security can uncover important investment opportunities. in addition to the put option, ee savings bonds offer several tax options. the tax options add yet another layer of complexity to these securities while offering other investment opportunities. to repeat, u.s. savings bonds do not always present above-market yields, but they sometimes do. financial planners must learn how to analyze savings bonds before they can recognize these unique opportunities. the purpose of this article is to demonstrate how these complex contracts can be viewed and used by investors. we use examples from 1993, but the approach can be applied to other times. we show how savings bonds can be viewed as short-term, intermediate-term, and long-term bonds. we also show how they can be used to beat the kiddie tax, to finance postsecondary education, and in retirement planning. the article is organized as follows: section ii looks at historic aspects of savings bonds and describes the terms of the ee-series and hh-series u.s. savings bonds; section iii presents generic models of the bonds noting the option features; section iv values hh bonds and examines their competitiveness as short-term to long-term assets; section v discusses three tax-based strategies using ee bonds; and the last section presents conclusions. ii. the ee and hh savings bonds series ee and hh savings bonds are the only ones currently being issued in the united states. for bonds issued after march 1, 1993, table 1 shows the minimum redemption values by investment horizon on a $1,000 par value ee savings bond. it also shows the coupon interest payments on a $1,000 par-value hh bond. the ee bonds are issued at 50% of face amount. they do not pay interest. instead, the minimum redemption value grows at a 4% annual rate (compounded monthly). they must be held for six months before the redemption value rises, but it rises monthly thereafter at a 4% annual rate. if redeemed at maturity in 18 years, they would be worth at least $1,026.00. if held at least five years, ee bonds pay the higher of 4% or a market-based variable yield. the variable yield equals 85% of the average yield during the holding period on live-year treasury notes. for example, suppose an investor buys an ee bond on march 1, 1993 and redeems it after five years. if the average five-year yield (as of the beginning of each six-month period) averages 5%, then the investor will receive 4.25% (5% x .85) instead of the 4% rate that is stated on the bond. the hh bond offers a 4% annual (2% semiannual) coupon interest payment. thus, a $1,000 bond pays $20 interest every six months. it must be held for the full six months to receive the $20 coupon interest. it can be redeemed at par at any time. analysis of u.s. savings bonds 43 table 1 the ee and hh savings bonds: beginning march 1993 ee minibus redemption values* hh cuupon interest period .4jier issue date (in do&z-s) (in dollars) at issue 500.00 na 6 months 7 months 8 months 9 months 10 months 11 months 1 .o year 1.5 years 2.0 years 2.5 years 3.0 years 3.5 years 4.0 years 4.5 years 5.0 years 18.0 years 510.00 512.00 513.60 515.20 516.80 518.80 520.40 530.80 541.60 552.50 563.60 575.00 586.60 598.40 610.50 1,026.oo 20 0 0 0 0 0 20 20 20 20 20 20 20 20 20 20 note: *since march 1, 1993, funds invested in ee savings bonds have offered a minimum return of 4% (compounded monthly) for up to 18 years. the bonds come in minimumdeno~mtions of $50 face value ($25 issue value). ee bonds must be held six months before the re~mption value rises, but it rises monthly after six months at a 4% annual rate (compounded monthly). the table shows all redemption values through one year but only selective redemption values thereafter. an investor must hold the hh bond for the full six months toreceive the $20 interest. each r~emption value for ee and hh bonds represents a put option on the bond’s price. for example, the investor can choose to redeem the ee bond for $5 io in six months (that is, exercise the put option at $5 lo), or redeem it at $512 in seven months, and so on. the hh bond contains an american put option with an exercise price of par. the next two sections analyze these put options in depth. table 2 gives other details of the ee and hh bonds. the ~nimum face value of ee and hh bonds are $50 and $500, respectively. neither bond can be transferred and, therefore, they cannot be used as collateral. thus, there exists no secondary market for the bonds. each year, each investor can purchase up to $15,000 issue price ($30,000 par value) of ee bonds. hh bonds cannot be directly purchased. rather, beginning six months after issue, ee bonds can be exchanged for hh bonds. savings bonds also offer implant tax advantages. interest on u.s. savings bonds is exempt from state and local taxes. federal taxes must be paid on the coupon interest on hh bonds in the year in which it is paid. in contrast, the ee bonds offer several tax options. table 3 summarizes these choices. the investor may choose to defer taxes until the bond is redeemed or matures (method 1) or pay taxes on the increase in redemption value as interest each year (method 2). furthe~ore, the investor may change from method 1 to method 2 and pay taxes on all interest accrued to date. after filling out irs form 3115, the investor can also switch from method 2 to method 1. most investors choose to defer taxes. the tax treatment of the realized interest can take one of three forms. 44 financial services review table 2 terms and conditions of ee and hh savings bonds issued after march 1 4(l) 1995 ,1993 ee bonds hh bonds minimum denomination $50 par value $500 par value issue price 50% of par value par value transferability not eligible for transfer same no secondary market not eligible for collateral interest rate tax status the larger of 4% or a market-based variable rate. the latter 4% denotes 85% of the average yield on five-year treasury notes. interest can be: (1) reported as it accrues or (2) tax deferred interest paid every six until the bond is redeemed, exchanged, or matures, at months. taxable which time the accumulated interest may be (a) taxed, in year paid. (b) tax deferred yet again, or (c) partially or fully tax exempt. original maturity 18 years 10 years final maturity 30 years 20 years it may be taxable in the year of redemption or at maturity. the redemption value can be rolled into hh bonds, in which case the interest continues to be tax deferred. for example, someone invests $5,000 ($10,000 face value) in ee bonds. in four years, they are worth $5,866. the investor could exchange the ee bonds for $5,500 of hh bonds and pay taxes on $366, or exchange the ee bonds plus $134 for $6,000 of hh bonds. in either case, the interest rolled into hh bonds is tax deferred until they are redeemed or mature. the interest may be partially or fully tax exempt if used to pay for qualified educational expenses. this tax option is popularly known as the education savings bond program. it is nothing more than an ee savings bond with interest deferred until redemption and used in the year of redemption to finance tuition and required table 3 tax options on ee savings bonds method 1: defer taxes on accrued interest until a. b. c. the bond is redeemed or matures, at which time taxes must be paid on accumulated interest; the bond is exchanged for hh bonds, in which case interest rolled into hh bonds continues to be tax deferred; or the bond is redeemed or matures and the proceeds are used in that year to pay for qualified postsecondary educational expenses. the accumulated interest may bc partially or fully tax exempt, depending on one’s level of income. method 2: pay taxes on interest as it accrues. that is, pay taxes on the increase in redemption values as interest each year. analysis of u.s. savings bonds 45 fees at most postsecondary schools. the bonds must be issued in one or both parents’ names; they cannot be registered in the student’s name. moreover, the exemption is subject to an income limitation, which will be discussed later. for further details, the investor should read “u.s. savings bonds for education,” published by the office of public affairs (u.s. savings bonds division, washington, dc. 20226). we examine tax-based investment strategies in depth in section v. iii. valuation models of savings bonds today in the united states, an individual can buy an ee savings bond or exchange an ee bond for an hh savings bond. in this article, we try to value the ee and hh savings bonds based on interest rates as of august 1993, the time of this writing. the hh savings bond is the easier bond to value. table 1 shows the coupon interest on today’s 4% coupon, $1,000 par value, hh savings bond. the bond also offers an american put option that allows an investor to sell it at par at any time. interest on the bond accrues every six months.’ if interest rates rise sharply, it may be worthwhile to redeem the bond between accrual dates. however, the put option would most likely be exercised immediately after the $20 interest has been earned. because the put can be exercised at any time, it is simultaneously a six-month bond, a one-year bond, and so on, out to a ten-year bond. we can value this bond as if it has any of these maturities. as we will see, the savings bond offers a yield that is above current (august 1993) short-term yields and below long-term yields on marketable treasury securities. the basic model for hh savings bonds is: vhh,n = vstr.,~ + vpur,n where v stands for value and hh, n denotes an hh savings bond of n-year maturity, ~tr., n denotes straight treasury debt of n-year maturity; and put, n denotes an american put option with an exercise price of par on a n-year bond. again, n can be any maturity from six months to ten years. there does not exist a secondary market for savings bonds. despite this, it is posstble for the value of the savings bond to exceed par. for example, prior to march 1993 the hh bond offered a 6% coupon rate. based on a yield of 5.31% on four-year, coupon-bearing treasury notes, the straight debt value of the hh contract (when viewed as a four-year security) was $1.0246 (par = $l).’ although the put option was out of the money, its value was still positive. thus, the true value of the savings bond was something in excess of $1.0246. the lack of marketability presents no problem for the efficient investor, who could buy the currently underpriced bond and hold it until it either matures or is no longer undervalued (at which time, he or she will exercise the put option on the bond).3 we see that the value of the hh bond, v,,, can exceed par. however, the put option guarantees that the value can never be less than par. it follows that the value of the savings bond must be at least par: vhh 2 par. figure 1 shows the values of a 6% coupon, ten-year straight treasury note and a 6% coupon, ten-year hh bond. when the market yield is well less than 6%, the value of the put option is negligible and the value of both bonds coincide. as the market yield approaches 6%, the put option becomes valuable and the value of the savings bond exceeds the straight 46 financialservicesreview 4(l) 1995 i i t 1'0 860'0 860'0 i p60'0 i 260'0 i i i i i i 9lo'o i plo’0 i zlo’o i : f lo’0 i 890'0 990'0 i p90‘0 ~~ 290’0 -90‘0 s ' -8so'o -9so'o -tso'o -zso'o -so'0 ~8po'o ~~ 9po'o -ppo'o -~ zpo'o -~to.0 ~8eo'o -9eo'o -~ peo'o -zeo'o ~~ fo‘o -820'0 i 9zo'o -pzo'o -zzo'o i i r zo'o analysis of u.s. savings bonds 47 debt value. for market yields above 6%, the put option dominates the valuation and sets the bond’s value at par. it is more difficult to value an ee savings bond than an hh bond. in addition to the put option, the ee bond offers several tax options. the value of the package of put and tax options is not equal to the sum of the separate options. as we shall see, investment strategies that are designed to capture the value of one option often limit the value of another option. for example, in order to get the tax exemption of interest, the investor may have to hold the bonds until his/her child enters college; that is, the investor must forego the put option until the child enters college. the value of the put option in ee and hh bonds depends on the volatility of interest rates and other factors that do not vary with the investor. in contrast, the value of the tax options on ee bonds depends on individual tax rates and, thus, its value varies among investors. a general model for the value of an ee bond is: vee,n = vstr,, ,, + v~~~ionspac~se, ,, where vdenotes value and ee,n refers to the ee bond; srr.,n refers to the straight bond; and options packnge,n refers to the package of put and tax options on an n-year bond. if we ignore the value of the tax options, then the structure of the ee bond is very similar to the structure of the hh bond. the structural difference is that the accrued interest on the ee bond is “reinvested” at 4%, thus raising the redemption price or exercise price on the put option. in contrast to the hh bond, the ee bond guarantees a 4% horizon return; that is, there is no reinvestment rate risk on the ee bond. we cannot calculate the precise value of an ee bond because the value of the tax options depends on each investor’s tax rate. however, the value of the hh-style bond represents a floor for the value of the ee bond. iv. valuing savings bonds with a put option as we have seen, the hh bond is simultaneously a six-month bond, a one-year bond, and so on out to a ten-year bond. and the value of the hh savings bond equals or exceeds its straight debt value plus the put value. in this section, we use august 16,1993 security prices to value the 4% coupon, hh savings bond when it is viewed as a six-month bond through a ten-year bond. we will see that the straight debt values and put values of the savings bond vary with maturity. for each maturity, we compare investments in two securities: an investment in straight treasury debt of the given maturity and an investment in the 4% coupon savings bond. since both strategies use treasury securities, we are comparing return prospects on investments of equal risk. in this section, we ignore the value of the tax options on ee savings bonds. again, the value of the hh bond may be viewed as a floor value for the ee bond. figure 2 shows the yield curves on august 16, 1993 on the hh savings bond and coupon-bearing treasury securities. there are two treasury yield curves at any time: the yield curve on marketable debt and the yield curve on savings bonds. again, the hh bond can be sold at par at any time. so, it is a bond with any maturity from six months through ten years. the 4% coupon on the savings bonds exceeds yields on straight debt for maturities less than 27 months. therefore, for investment horizons between six months (the accrual period on hh bonds) and 24 months, the hh bond guarantees an above-market yield with no additional risk. we next try to value the hh bond for investment horizons from six months through ten years. 48 financial services review 4( 1) 1995 -s’l -l -s’o -0 analysis of u.s. savings bonds 49 a. six-month horizon based on a six-month treasury bill discount interest rate of 3.11%, the straight debt value of a $1 savings bond is $l.0040.4 this savings bond also contains a put option. however, the investor must hold the hh bond for the full six months to receive the 2% interest. thus, unless treasury yields rise sharply, the put option will not be exercised before the interest accrual date in six months. if we ignore the ability to exercise the put option before six months, the value of the hh bond is $1.0040. some investors place funds in money market funds, cds, and other near-cash securities to reduce their exposure to interest rate risk. it is instructive to compare returns on savings bonds to returns from rolling over six-month cds. on august 16, 1993, six-month cd rates were about 3%.5 the ee savings bond dominates the cd strategy. it guarantees the larger 4% annual interest rate for the first six months6 it also offers a series of contingent, european call options on interest rates. at the end of six months, the investor has the right to reinvest the funds at 4% for another six months. at the end of the second six months, the investor has the right to reinvest the funds again at 4%. in total, the savings bond offers a series of 35 contingent call option on the interest rate at 4%. if an investor ever decides not to “buy” the 4% interest rate for six months (i.e., he or she exercises the put option on bond price and redeems the bond), then he or she forfeits all subsequent call options. table 4 compares returns on the old ee bond bought before march 1993 to a strategy of rolling over six-month cds. suppose the ee bond, which promised at least 6% if held for live years, was bought in february 1993. at that time, the six-month redemption value on ee bonds translated into a 4.16% interest rate, or about 1.4% above the then-current six-month cd rates. the immediate yield advantage was just the beginning of the story. instead of redeeming the bond in six months (i.e., august 1993). the investor had the option to hold it for a second six months and earn a 4.39% annual interest rate, [( ($521.60 $510.40) / $5 10.40) x 21. since cd rates in august 1993 were about 3%, the investor should have opted to hold the ee bond and earn the 4.39% yield. table 4 shows the redemption value after each six-month period as well as the interest rate on the next six-month interest rate option. for example, in february 1994 the investor had the option of cashing in the bond at $521.60 or accepting the 4.75% interest rate option-that is, holding the bond for another six months and earning 4.75%. interest rates on the series of options rise until they peak 4.5 years after issue at 8.18%. between years 5 and 12, the redemption values grows at a guaranteed minimum rate of 6%. the ee contract in february 1993 offered investors a real bargain compared to money market investments. first, it guaranteed a higher initial six-month yield. second, their redemption values implied a series of interest rate call options, which protect the investor against falling interest rates. these investors may enjoy above-market yields for 12 years. third, the put option provided protection against a rise in interest rates. finally, the ee-bond offers the tax options that are not available on cds. the 4% ee-series savings bond is not as attractive as the 6% ee bond. however, it continued to offer investors a better return than money market investments through march 1994. 50 financial services review 4( 1) 1995 table 4 comparing interest rates on six-month cd and ee bond: february 1993 period afrer issue date cd rate ee redemption value ee rate (in dollars) (beginning of period)(%) at issue 0.5 year 1 .o year 1.5 years 2.0 years 2.5 years 3.0 years 3.5 years 4.0 years 4.5 years 5.0 years 5.5 to 11.5 years 12 years 2.75% ? ? ? ? ? ? ? 5cq.00 510.40 521.60 534.00 548.00 563.00 580.00 599.60 621.20 645.60 672.00 1.016.40 4.16% 4.39% 4.75% 5.24% 5.55 5.97 6.76 7.20 7.86 8.18 6.00 6.00 no guaranty* note: *no guaranty: the guaranteed interest rate ends 12 years after issue. b. one-year horizon let us compare the returns on three investment strategies as of august 1993: l buy and hold a one-year, coupon-bearing note yielding 3.39%; l buy a 4% coupon savings bond, redeem it in one year; or l buy a 4% coupon savings bond, redeem it at par in six months, and reinvest at that time in a six-month treasury bill. the second strategy dominates the first; the straight debt value of the savings bond is $1.0059.7 the value of the put option reflects the ability to choose between the second and third strategies-that is, the choice to be made in february 1994 between accepting the 4% (annual) coupon on the savings bond or the then-current six-month treasury yield. to value the put option, notice that a put option on bond price is identical to a call option on interest rate. it gives the bondholder the right to buy a $1 par value, pure discount, six-month bond in february 1994 at $.9804 [$1/l .02]-that is, to buy a 4% annual interest rate for six months. we will use current treasury strip prices to value a call option on a forward contract that allows an investor to buy in february 1994 a $1 pure discount bond maturing in august 1994. based on august 16, 1993 prices, the implied six-month forward rate between february and august 1994 is 1.8258% and the corresponding price on the forward contract is $.9821.8 thus, the option in the bond has an intrinsic value of $.0017, or $.982 1 $.9804. its actual option value would exceed this amount due to its time value. we thus conservatively estimate the value of the savings bond for an investor with a one-year investment horizon at vhh = 1.0059 + .0017 = $1.0076. for the one-year investment horizon, it will be clear in six months whether the investor should exercise the put option. if in six months the six-month treasury yield exceeds 4%, the put option should be exercised. if not, it should not be exercised. the next example shows that it is not always clear whether the put option should be exercised. analysis of u.s. savings bonds 51 c. two-and-a-half-year investment horizon an investor could buy a 2.5year, coupon-bearing treasury note yielding 4.41% or a 4% coupon savings bond. the latter contains the put option that would most likely be exercised at the end of an interest accrual period in 0.5 years, 1 .o year, 1.5 years, or 2.0 years. the straight debt value of the savings bond is $0.9904 for a $1 par-value bond.’ the question is whether the put option, if sold separately in the financial market, would be worth $0.0096 since the bond can always be redeemed at par. as part of the savings bond, the put option must be worth at least $0.0096. if the put option sold separately is not worth $0.0096, the savings bond is not a good investment for the person with a 2.5year investment horizon. the put option gives the investor the right to cash in the savings bond at par and reinvest the proceeds at the then-prevailing treasury rate for the remainder of the 2.5-year horizon. it gives the investor the right to exchange the savings bond in six months for a two-year note, in one year for a 1.5-year note, in 1.5 years for a one-year note, or in two years for a 0.5-year note. is the put option worth the 0.4 1% initial yield disadvantage (4.4 1% -4%)? we do not know. if the yield on the initial 2.5-year note remains at or below 4.41%, then the straight debt will outperform the savings bond. if yields rise sharply, however, buying the savings bond, redeeming it at par after rates rise, and reinvesting at the then-current interest rate will produce the higher 2.5-year return. the 2.5-year returns on straight debt and savings bonds will be approximately equal if the put option is exercised in six months when the two-year yield is 4.51%. the 10 basis point (4.5 1 4.41%) yield advantage for four six-month periods roughly offsets the 41 basis point disadvantage for one period. the savings bond would also produce an equal 2.5-year return if: (1) the put is exercised in one-year and the 1.5-year treasury yield is 4.68%, (2) the put is exercised in 1.5 years and the one-year treasury yield is 5.03%, or (3) the put is exercised in two years and the 0.5-year yield is 6.05%. suppose that in six months, the two-year treasury yield is 4.6%. should the investor exercise the put option? exercising it and buying a two-year note would guarantee that the savings bond would produce a larger 2.5-yearreturn than the original straight debt. however, once the put is exercised, the investor loses the protection against rising interest rates. if rates continue to rise sharply, the investor would be better off delaying the exercise; in essence, the put option offers the investor a one-time exchange of the savings bond for the then-current yield. what is the optimal strategy for exercising the put option? despite tremendous efforts, we cannot answer the question with precision. at least two issues make valuing an option on bond price more difficult than valuing an option on stock price. first, we cannot agree on the nature of the process driving interest rates. are interest rate means reverting? do they follow a random walk? in contrast, we agree (at least for option models) that stock returns follow a log-normal distribution. to understand the second problem, note that the put option’s value depends on bond price volatility and the probability distribution of future bond prices. consider a stock currently priced at 40. the stock’s volatility per unit of time is assumed to be constant, so the probability distribution of the stock price widens as the investment horizon lengthens. the same is not true of an option on a bond price. through time, the maturity of the underlying bond shortens, which causes the volatility of the bond price to continually fall. for example, for the 2.5-year investment horizon, we had an option on a two-year bond in 52 financial services review 4(t) 1995 six months, an option on a 1 s-year bond in one year, and so on. moreover, the distribution of future bond prices is constrained by the guarantee that at maturity it will sell at par. in short, it is not clear whether or not the savings bond offers a good deal for the investor with a 2.5-year investment horizon. the savings bond’s straight debt value is $0.9904. if a separate put option would be worth more than $0.0096, the savings bond would be more than competitive. if less than $0.0096, it would be less than competitive. a key factor is the volatility of interest rates. the more volatile, the more valuable is the put option. d. ten-year horizon ten-year, coupon-bearing treasuries yield 5.89%. the straight debt value of the savings bond is $0.8587. the put option in the savings bond is thus worth $0.1413, because it allows it to be sold at $1.00 (par). it is clear that the value of the same option sold in the financial markets would be worth well less than this amount; the put option is so deep out of the money that it cannot be priced with current (august 16, 1993) market quotes. this implies that the hh savings bond offers a noncompetitive package for the investor with a ten-year investment horizon. if such an investor currently holds an hh bond, then he or she should exercise its overvalued put. if the bond’s put is not exercised, it will likely lose value as the maturity of the underlying bond shortens. v. tax strategies in section iv, we examined the value of the put option. in this section, we examine several investment strategies that are based on tax options. the tax options on ee bonds were described in section ii and summarized in table 3. a. beat the <‘kiddie tax” the “kiddie tax” section of the tax reform act of 1986 was written to limit parents’ ability to lower taxes by shifting income to their children. in 1993, for a child under age 14, unearned income in excess of $1,200 was taxed at the higher of the child’s or the parents’ marginal tax rate. this greatly reduces parents’ ability to reduce taxes by shifting income to a child under 14. once a child reaches 14, all of his or her income is taxed at the child’s tax rate. the ee bond offers parents the opportunity to shift earning assets to children under 14 without exceeding the income limit. parents can register (i.e., buy) ee bonds in the child’s name. suppose a couple registers $10,000 (redemption value) of ee bonds in johnny’s name on his fourth birthday. at this point, the interest can be treated in one of two ways. first, johnny can hold the bonds until he turns 14, when they will be worth at least $14,908. if redeemed, the $4,908 of deferred interest will be taxed at johnny’s tax rate. in the meantime, the parents can shift enough other assets into johnny’s name to fully use the income limit, currently $1,200, before he turns 14. second, if johnny is not using the income limit, he can elect to pay taxes on the interest as it accrues. as of 1993, the first $600 of unearned income is tax free. thus, if the $lo,~o ee bond is the only asset in johnny’s name, he should claim the interest annually. in the first year, he will claim $408 of interest and pay no taxes. in the second year, he will claim $424 of interest and avoid taxes [$10,832 $10,424]. if the accrued interest ever exceeds analysis of u.s. savings bonds 53 the income limit, the parents can change the tax treatment from the accrual method (method 2) to the deferral method (method 1). by registering savings bonds in a child’s name, parents can help finance any future expense-summer camp, room and board at college, or college tuition. the next strategy can only be used to finance tuition and fees at most higher educational institutions. b. educational savings bond program the accumulated interest on ee bonds may be partially or fully tax exempt if used to pay for qualifying educational expenses. to qualify, the bonds must be issued after 1989. they can be purchased by anyone, but they must be issued in one or both parents’ names. the bond cannot be issued in the child’s name, but he/she may be registered as the beneficiary. the bonds must be redeemed in the year the bond owner pays qualified educational expenses-tuition and fees-to an eligible educational institution. room, board, and books do not qualify as educational expenses. eligible institutions include colleges, universities, technical institutions, and vocational schools. as of 1994, the full interest exclusion applies only to couples filing jointly with modified adjusted gross income of $60,000 or less. (modified ac1 is adjusted gross income before this interest exclusion.) the exemption is phased out as the couple’s income rises until it is completely eliminated for couples with modified ac1 above $90,000. the exemption also phases out for single tilers. the corresponding 1994 amounts are $40,000 and $55,000. since the revenue reconciliation act of 1993, these limits are not indexed for inflation. some middle-income families can be reasonably certain that the ee interest will be fully tax exempt. these families might use ee bonds issued in the parents’ names to finance tuition and fees and ee bonds issued in the child’s name to finance room, board, and books.” figure 3 shows two term structures of interest rates for middle-income families. the first is the term structure for pure-discount treasury strip securities. the second shows a flat term structure at 5.56%, which corresponds to the before-tax equivalent yield of a fully tax-exempt ee bond [4%/( l .-.28)]. since the ee bonds have no reinvestment rate risk, we compare their yield to treasury strip yields. as of august 16, 1993, the term structures cross at seven years. consider a middle-in come family (i.e., who are in the 28% federal tax bracket and for whom the ee interest will be fully exempt) with a child entering college in five years. the figure implies that this family should prefer ee bonds yielding 4% after taxes to a treasury strip yielding 5.12% before taxes and 3.69% after taxes [5.12%/(1-.28)]. c. retirement planning investments in ee savings bonds can be used to defer interest income until retirement. an individual could purchase ee bonds during his or her working years and defer the recognition of the interest until his/her retirement years. this retirement planning strategy may prove even more beneficial for an individual who works in a high income-tax state and plans to retire in a low income-tax state. as with other tax-deferred retirement strategies, major advantages include not only the possibility of a lower tax bracket after retirement but also the greater accumulation of wealth since there is no loss of investment return due to income tax payments. 54 financial services review 4(l) 1995 t i 31 % , z,ll 11 = / 5‘91 91 s‘s1 sl 8’91 pi s’ec el s’zl zl -2‘11 11 s’ol 01 s’6 r 6 p ‘2’8 8 s‘l l s‘9 9 s’s s s’t t s’e e s’z 1 2 s’l [ 1 -i i s’o analysis of u.s. savings bonds 55 during retirement, the investor can choose either to keep the ee savings bond and allow the interest to accumulate or to exchange ee bonds for interest-paying hh bonds. the latter strategy will probably be preferred if the individual depends on the interest payments to meet his/her income needs. as mentioned before, the tax-deferred interest on ee bonds that is rolled into hh bonds is not taxable until the hh bond is redeemed or matures. vi. summary us. savings bonds are complex contracts. financial planners have traditionally paid little attention to savings bonds, in part because they often offer below-market interest rates. however, savings bonds sometimes offer above-market interest rates. by early 1993, many financial planners recognized the above-market yields offered on 6% ee savings bonds. the rush to capture these attractive yields forced the treasury to lower the promised interest rate on new funds invested in ee bonds to 4%. we suspect that few planners recognized that the 4% ee savings bonds offered above-market yields through march 1994 when viewed as a substitute for money market investments. in this article, we have tried to demonstrate how savings bonds can be analyzed for investors with short-term to long-term investment horizons. our approach should help planners identify unique investment opportunities. hh-series and ee-series savings bonds contain put options that protect the investor against a rise in interest rates. the ee bonds also offer several tax options. we have shown how savings bonds can be used to beat the kiddie tax, to finance postsecondary education, and in retirement planning. with a little imagination, an astute planner could think of other uses. acknowledgments: the authors thank a.j. senchack and an anonymous referee for helpful comments on an earlier draft. notes 1. interest on marketable treasury securities accrues daily. on a 4% coupon bond, interest accrues at $0.1096 per day ($40/365). an investor who sells a marketable bond 10 days after the last coupon payment date would receive $1 .lo accrued interest. in contrast, the investor in an hh savings bond gets no interest unless it is held for the full six months. on ee bonds, interest accrues (i.e., redemption value rises) monthly after six months. 2. in a financial calculator, insert fv= $1 par value, pmt = $.03 per six-month period, n = 8 six-month periods, and i = 5.31%/2 = 2.655%. 3. brennan and schwartz (1977) make the same point. “[the lack of marketability] presents no obstacle to the investor in an efficient capital market since the bond can be priced and the optimal exercise strategy determined even if trading of the bond is not allowed” (p. 68). even though the hh bond cannot be sold at a premium above par, the present value of its cash flows exceeds par. its value when held as an investment exceeds par. instead of holding the bond, suppose the individual wants to reduce his/her portfolio’s exposure to government securities and invest in common stocks. in this case, the hh bonds could be held (you do not liquidate an asset worth 102.46 for 100) and some other similar-risk asset liquidated to finance the increase in common stock. the lack of a secondary market would impact the utility of a savings bond only when (1) the individual wanted to change the composition of the portfolio and (2) there were no other marketable securities 56 financial services review 4(l) 1995 of comparable risk to the savings bonds that can be liquidated. in practice, we believe this would be a minor problem. 4. the $1 bond pays $1.02 in six months. the six-month treasury bill discount yield of 3.11% implies a present value of $1.02 (.98428), where: [.98428 = 1 .0311 (182/360)]. 5. the wall sweet journal reports an average three-month cd rate of 2.68%. since the six-month yield on treasury bills is only 13 basis points above the three-month yield, the 3% estimate for the six-month cd rate appears generous. 6. with a savings bond, an investor receives credit for the full month even if the funds are registered on the last day of the month. thus, the 2% return can be earned in five months and two days. an ee bond registered on january 31 can be redeemed in 152 days on july 1. the annualized return is 4.74%. [2%*(360/152)]. 7. the one-year yield on coupon-bearing treasuries is 3.39. on a financial calculator, insert n = 2, i = 3.39%/2 = 1.695%, pmt= $.02, fv= $1. 8. the average of bid and ask prices for february 1994 and august 1994 treasury strips are 98:15 and 96:22.5, where the latter denotes 96 plus 22.5/32% of par. the implied yield is [(98: 15/96:22.5) 11. the implied forward price is 96:22.5/98: 15, or $.9821. 9. on a financial calculator, insert n = 5, i = 2.205%, pmt= $0.02, fv= $1. 10. the interest on qualifying bonds will be excluded from federal income tax only if the qualifying tuition and fees paid during the year are equal to or more than the redemption proceeds (interest and principal) of qualified bonds, regardless of how the bond proceeds are actually used. if tuition and fees are less than the value of the bonds cashed, the exclusion is proportional to the percentage of the value that was used for tuition and fees. for example, if a bondholder redeemed $10,000 worth of bonds during the year but tuition and fees totaled only $8,000,80% of the interest income could be excluded from federal income tax. to avoid this problem, parents should buy several small-denomination ee bonds instead of one large-denomination bond. for example, the parents could buy 20 bonds worth $500 each instead of one bond worth $10,000. they could then liquidate an amount to ensure that the interest was fully tax exempt. reference brennan, m.j., &schwartz, e.s. (1977). savings bonds, retractable bonds, and callable bonds. journal of financial economics, 67-88. pii: s1057-0810(99)80008-0 from the editor karen eilers lahey this first issue of volume 7 for 1998 focuses on empirical tests of strategies that may be appropriate for individual investors. the lead article by clarie e. crutchley, carl d. hud son, and marlin r. h. jensen examines the impact of changes in the aggressive manage ment of the california public employees' retirement system (calpers). they discuss the history of activism by this leading state pension fund that annually publishes a list of firms that are being specifically targeted and monitored for changes in performance. the authors conduct tests of the changes, if any, in the portfolio of target firm's return. robert a. kunkel and william s. compton address an investment strategy that many professors may have a personal financial stakeholder interest in, the teachers insurance and annuity association-college retirement equities fund (tiaa-cref). they provide an excellent review of the literature on seasonal anomalies that include the january effect, the weekend effect, and the turn-of-the-month effect. a switching strategy based on this anomaly concept is then applied to the money and indexed stock funds offered by tiaa-cref funds and contrasted with a buy-and-hold strategy. the third article by dale domian and william reichenstein examine the predictive content of intermediate and long-term bond spreads for bond market prices utilizing the period 1942 to 1994. their unique results indicate that the bond market prices an interme diate-short spread, which tracks the reward for bearing duration risk. based on their empir ical findings, they offer advice to individual investors who wish to practice asset allocation strategies. f. larry detzel and robert a. weigand author one of two articles in this issue that examines the persistence performance of mutual funds or their managers. they focus on a model that relates mutual funds returns to the specific characteristics of the stocks held, such as market capitalization, book-to-value equity, earnings yield, and cash flow yield. their model helps to explain the previous empirical findings on persistence. the final article by gary e. porter and jack w. trifts approaches the persistence of mutual fund performance by looking at 93 fund managers who have run the same fund for at least ten years. they investigate the performance of long-term managers versus those with a shorter term. both studies on persistence utilize approaches that are different from previous research and produce interesting results. the issue closes with a book review of kwok ho and chris robinson's personal financial planning for canadians. they are bringing out an american version this year. pii: s1057-0810(01)00076-2 on time: contributions from the social sciences barbara s. poole* the american college, 270 bryn mawr ave., bryn mawr, pa 19010, usa received 23 february 2001; received in revised form 4 june 2001; accepted 21 july 2001 abstract this paper provides a brief review of the anthropology and psychology literature as it relates to time, an important variable in finance. first, the paper discusses ways that individuals represent time, and introduces cultural variations in the perception of time. then the experience of time passing, and behavioral pace, is discussed. the succession of time and the orientation toward past, present, and future, are described. the paper may provide implications for academics whose finance research is related to behavior over time. © 2001 elsevier science inc. all rights reserved. 1. introduction financial advisors and individuals rely on time value calculations for planning how to spread limited financial resources across a lifetime of needs. an individual’s perception of the future can influence his or her attitude toward planning, and perceptions about the passage of time can influence the urgency that the individual feels toward planning. these attitudes and perceptions can also influence current savings and spending behavior, which can support or hamper the individual’s attempts to achieve financial goals. although time is an important variable in time value calculations, financial researchers have paid little attention to the way that individuals perceive and experience time. the purpose of this paper is to provide a brief review of the temporal anthropology and psychology literature streams and to discuss possible implications for finance researchers. the paper is organized as follows. part ii discusses ways that individuals represent time, and introduces cultural variations in the perception of time. the experience of time passing, along with behavioral pace, is discussed in part iii. in part iv, the succession of time, time orientation, and attitudes toward the past and the future, are described. the paper concludes in part v. * tel.: �1-610-526-1341. e-mail address: barbarap@amercoll.edu (b.s. poole). financial services review 9 (2000) 375–387 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00076-2 2. representing and measuring time time is an important variable in time value calculations, a variable so significant that financial advisors and educators often illustrate the time horizon under discussion. however, cultural differences exist in the ways that individuals envision time. further, the precision of time measurement varies among cultures. this section will describe those variations and discuss possible implications for researchers. 2.1. representing time educators and financial planners usually represent a specific holding period or period of annuitization by depicting a time line. such a depiction generally is consistent with the individual’s concept of time. many individuals believe that the time continuum has a beginning before which, time did not exist. scientists associate this beginning with the big bang, while some religions associate this beginning with creation (hawking, 1988). individuals tend to think of time flowing in a forward-moving, linear fashion (melges, 1982). fig. 1 presents a representation of such a time line, starting with the big bang and extending forward. the representation of time described in the prior paragraph, and presented in fig. 1, is associated with western reasoning. however, the entire world does not adhere to this western representation of time. for example, speakers of bantu, a widespread language in central, eastern, and southern africa, depict time quite differently. bantu-speaking tribes in south africa, such as the hehe in southern tanzania, believe that time began with a supreme being as the source of the universe. these groups represent time as a revolving sphere that rolls forward along a spiral path into the endless future (msumange, 1998). the revolving sphere may be similar to the rotating earth, the source of day and night, while the spiral can represent the agricultural seasons or cycles of the moon. an example of how an african may represent time is presented in fig. 2. so although westerners and africans depict time differently, both cultures share two fundamental ideas concerning time. first, time had a beginning, before which, time did not exist. second, time is forward moving. however, neither of these ideas is universal. fig. 1. a western representation of time. fig. 2. an african representation of time. 376 b.s. poole / financial services review 9 (2000) 375–387 not all cultures agree that the continuum of time has a beginning. hindus believe that time has existed forever. they envision time as a revolving circle of birth, death, and rebirth. they feel a strong connection to prior and future revolutions of the circle, as they believe that good and bad conduct in one life is rewarded or punished in the next life (levinson, 1998). an example of a hindu’s representation of time appears in fig. 3. the cultures discussed thus far have agreed that time is a succession of events in the past, present, and future. however, the notion of the forward movement of time is not universal across cultures. for example, neither the hopi indians nor the people of the trobriand islands have words that distinguish between past, present, and future. rather than a linear representation of time, these cultures see time in a unified holistic pattern (melges, 1982). both the past and the future are blended with, and indistinguishable from, the present. these cultures see time as a fabric with an interwoven pattern of past, present, and future. fig. 4 is an example of such a depiction of time. so cultures vary in the ways that they represent time, reflecting such factors as religious fig. 3. a hindu representation of time. fig. 4. a hopi representation of time. 377b.s. poole / financial services review 9 (2000) 375–387 beliefs and the physical environment. while the culture can provide a general framework for envisioning time, each individual may envision a unique variation on the general cultural representation. of course, culture provides a framework, rather than a dictate, for envisioning time, and individuals may adopt representations or ideas usually associated with other cultures in constructing their own representations. an individual may also use different representations in different contexts, for example, the school year may be envisioned differently from the agricultural year. 2.2. measuring time the measurement of time is also a product of the individual’s culture. the accuracy of the measurement is associated with the characteristics of a geographical area. additionally, researchers have found other factors that are associated with clock accuracy. these time measurement issues are discussed further in this section. in the western representation of linear time, demarcations on the line represent uniform measurements, signifying specific time periods such as days, months, years, or decades. the time line will usually end at some specific time associated with the planning period, such as an assumed retirement or mortality age. time periods between critical points are clearly and uniformly represented. in the african culture, demarcation of time is less precise than in the western world. time is organized in terms of natural events that occur cyclically, such as the dry or rainy seasons, the agricultural cycles, and the cycles of the moon. shorter time periods are distinguished by the position of the sun, and the range of time in the night that a particular bird sings, crows, or clicks. important events in the past, such as wars, also provide temporal information. levine and norenzayan (1999) found significant differences in 31 countries’ temporal precision, as measured by the accuracy of public clocks. switzerland, italy, austria, and the u.s. were the top four countries in clock accuracy. greece, indonesia, and el salvador were at the bottom of the list. levine and norenzayan found that clock precision is related to the industrialization of the country as well as its population and economic health. some philosophers and economists have posited that a society’s temporal imprecision deters its industrialization. for example, msumange (1998) suggested that the inexactness of time measurement in third world countries, such as africa, has delayed the economic development of the area. he suggested that adoption of a precise “mathematical concept of time” is necessary for future development (p. 11). however, in the u.s., industrialization and the workings of the factory came first, and drove the need for a temporal precision that had been previously unimportant (hunt & hait, 1990). 2.3. summary cultural perspectives influence the ways that individuals represent time. the way that a culture represents time is a product of the religious and scientific beliefs of its population. in the next section, we will see that culture, as well as other variables, influences other aspects of the ways that individuals experience the passage of time. 378 b.s. poole / financial services review 9 (2000) 375–387 3. the duration of time individuals vary in the ways that they experience time passing, and in the speed of their behavioral pace. duration refers to the subjective experience of time, as uniquely perceived and interpreted by each individual. individuals use an inner tempo to judge how much time has elapsed. this inner tempo influences observable behavior, such as walking pace. in this section, inner tempo and observable pace are discussed, along with implications for advisors, educators, and researchers. 3.1. inner tempo inner tempo, also referred to as inner or temporal pace, is the individual’s experience of the passage of time. the literature describes the experience of time passage as “the appreciation of duration” (frederickson, 1988, p. 63). researchers assess participants’ inner tempo by asking participants to estimate the length of time that elapsed during a prior task. if the individual’s internal pace is fast, then external clock time seems slow (melges, 1982). psychological typing has ascribed internal pace to personalities, with type a personalities described as fast paced, time pressured workers who think and act quickly, and type b personalities described as steady workers unpressured by time (rao, reddy & samiuliah, 1997). time is valuable to highly motivated and high achievement-oriented individuals, who tend to be time possessive and concerned with the appropriate use of time. individuals judge time to have elapsed more quickly when they are busy than when they are idle, although this is only true for individuals who are achievement oriented (meade & singh, 1970). likewise, individuals judge that time has elapsed quickly when they have made progress toward a goal. however, individuals who are not achievement oriented make similar estimates of time regardless of whether they have made task progress (meade & singh, 1970). individuals judge that a shorter period of time has elapsed when they are interested in a task, as compared to when they find the task uninteresting (rotter, 1969). deliberate practice and naturalistic decision-making literature indicates that individuals with high levels of expertise can experience a quickened inner pace at critical points. ericsson and smith (1991) reported tennis players’ descriptions of external time slowing during crucial moments in an important game. klein (1998) studied experts, such as fire-fighters and aviators, who make life or death decisions under extreme time pressure. experts reported that ‘time stood still’ at critical times as they considered large amounts of information and selected among alternatives. some of the earlier work in the temporal psychology area evolved from medical practitioners, who observed temporal distortions in the thinking of patients with psychiatric disorders. time moves more slowly in depression (wyrick & wyrick, 1977), while manic patients experience time moving more quickly (melges, 1982). acutely psychotic patients experience a sense of timelessness, or disconnection from the experience of time passing (melges). environmental factors such as temperature changes, as well as the use of stimulants and psychedelic drugs, can speed up the internal clock. sensory deprivation and hypnosis can also induce changes to the sense of psychological time (melges). researchers have studied the accuracy of the inner clock as an indication of other personal 379b.s. poole / financial services review 9 (2000) 375–387 attributes. tsukanov (1991) studied the “quality of the inner clock,” or accuracy of participants’ time estimates. he associated accuracy with innate intelligence, using academic success in the soviet system as a proxy for intelligence. tsukanov’s participants consisted of (a) scientists, (b) university students, (c) poorly performing secondary education students, and (d) children diagnosed with very low intelligence. when judging the duration of a very short elapsed time, the scientists performed most accurately, and the low intelligence children least accurately. however, because the researcher did not consider other variables such as motivation, goal orientation, and maturity in the analysis, the conclusion that the ability to estimate the duration of time is related to intelligence is suspect. individuals, despite their personal experiences of time, rely on clock speed as the accurate determinant of elapsed time. rotter (1969) found that, by secretly adjusting clock time, test administrators could manipulate participants’ estimates of temporal duration. participants mistrusted their own estimate of a task’s duration, adjusting their estimates after obtaining access to a clock. these results suggest that individuals may be aware that their experiences of duration can be inconsistent or situational. 3.2. observable pace historians attribute the importance of time in the u.s. to the industrial revolution, when manufacturing was coordinated among machines and workers, according to rigid factory schedules. the workings of the factory required a new precision, and punctuality necessitated the invention of seth thomas clock company’s wind-up alarm clock in 1876. the impact of the industrial revolution permeated the american culture, and new sports created after 1860, such as basketball and football, were ruled by the clock, in contrast to un-timed earlier sports such as baseball (hunt & hait, 1990). the newly increased value of time spawned a new science of time and motion studies (e.g., taylor, 1947). as efficiency became more important, cultural pace increased. hall (1983) observed that cultures vary in overall pace, and compared the characteristics of fast and slow paced cultures. he described a fast paced, monochronic time culture, such as the u.s., as one run by the clock, focused on tasks and schedules rather than on people. the slower paced polychronic culture, characteristic of the third world, is run by relationships among people and does not consider time as a commodity that can be wasted. these descriptions parallel the psychological types a and b of individual personalities mentioned earlier. levine (1997) has studied the pace of life in various geographical locations for the last twenty years, and his observations have extended those of hall (1983). levine defined the pace of a location as the overall speed of experiences there, and measured walking speed, clock accuracy, and the average time taken by a postal clerk to fill a stamp order. findings indicated that faster pace is a function of industrialization, population density, a healthy economy, and a cooler climate. readers may be interested in the relative pace of u. s. life. of the 31 countries in the levine (1997) study, the u.s. ranked 16th in overall pace. the fastest paced locations were in switzerland, ireland, germany, and japan. the slowest ranked cities were located in economically underdeveloped countries, el salvador, brazil, indonesia, and mexico. western european countries and japan had increased in speed compared to similar 1980 data. meanwhile, the u.s. had increased in walking speed, but decreased in the other two measures since 1980. 380 b.s. poole / financial services review 9 (2000) 375–387 the pace of the surrounding environment can influence the temporal pace of the individual. an american visitor to switzerland, ireland, germany, or japan will notice themselves speeding up to keep up with the faster environment. most visitors to tropical islands must make an effort to slow down to the pace of the environment (levine & norenzayan, 1999). emphasis on efficiency and productivity is at odds with quiet reflection necessary for thoughtful learning and planning. experts have indicated that for learning to take place, individuals must reflect on prior experiences, build on positive feelings, deal with negative feelings, and re-evaluate the experience based on those reflections (boud, 1993). in fact, fidelity group of funds’ peter lynch recently suggested that advisors need to “slow down” and educate as well as advise clients so that they can make rational decisions (koco, 2000, p. 3). clearly, finance practitioners recognize that pace can have an impact on behavior. 3.3. summary an individual’s sense of time, and behavioral pace, can vary depending on the situation, and patients with psychiatric disorders experience temporal distortions. overall pace varies among cultures, and is related to industrialization as well as to population density, the economy of the area, and the climate. in the next section, we will discuss a second aspect of temporal experience, succession. 4. the succession of time cultures that distinguish between past, present, and future tend to agree that the present moment separates the past from the future. succession is the individual’s perception of past, present, and future along a forward flowing time sequence. researchers have found that individuals tend to have a predominant temporal orientation of past, present (infrequently), or future. individuals provide clues regarding their temporal orientation in their speaking and writing. psychologists agree that the frequency of a particular verb tense is an indication of a participant’s or patient’s temporal orientation. in research, analyzing the frequency of verb tense during a taped interview is one common technique for determining temporal orientation (e.g., frederickson, 1988). several researchers have designed questionnaires that enable respondents to select expressions that most accurately describe their feelings about the past, present, and future (e.g., braley & freed, 1971). meade (1971, p. 177) asked participants to tell stories based on present tense sentences for example, “d. s. receives his degree today,” then categorized the stories as to the temporal orientation of the major theme. cultural studies provide some clues for understanding differences in the temporal orientation of individuals. meade (1971) studied male college students in the u.s. and india, and found significant differences in temporal orientation between the two groups, with the students in india past oriented and the americans future oriented. meade (1972) continued his cross cultural comparison by studying seven indian subcultures, and found that three of the indian subcultures exhibited a future orientation. the past oriented subcultures were more likely to ascribe to the hindu belief that an individual’s own acts or personal efforts are 381b.s. poole / financial services review 9 (2000) 375–387 un-related to lifetime achievement. these beliefs may explain why a young man might be unconcerned with planning for his present life’s future. meade also found that the temporal orientation of a culture is related to its members’ motivation and attitudes toward work. members of the future oriented subcultures scored more highly on a motivational scale than members of the past oriented subcultures (mead, 1972). the future oriented indian subcultures and americans were also more likely to use references to personal motivation, work, or effort. 4.1. focus on the past or past-present calabresi and cohen (1968) categorized attitudes toward time, including the temporal orientation characteristics of each category. they found that past orientation is typically characteristic of time anxious individuals. associated characteristics include a need to control time, discomfort thinking about the future, and anxiety about the passage of time. these individuals tend to adhere to schedules and value routine. researchers have found that a past orientation frequently is inconsistent with psychological well being. past orientation has been associating with coping problems, inability to delay gratification, and poor self control (melges, 1963). medical researchers have observed significantly more preoccupation with the past among depressed patients as compared to a control group. depressed patients tend to feel hopeless about the future. their thoughts of time past also extended further back into the past than the temporal thoughts of nondepressed individuals (wyrick & wyrick, 1977). when braley and freed (1971) compared psychiatric outpatients to a control group, the researchers found that the outpatient group was significantly past-present focused, while the control group was significantly future focused. the ways that all individuals view their pasts are influenced by their interpretations of the present. schkade and kilbourne (1991) studied hindsight bias, a reinterpretation of the past based on a later outcome. the researchers found that more hindsight bias occurred when participants were surprised or disappointed by the outcomes, and when outcomes were negative rather than positive. these findings imply that later outcomes can color individuals’ views of their pasts. regardless of the individual’s primary temporal orientation, negative past experiences may serve as a reminder of risk, and prompt risk averse behavior. in a simulated study of investment decisions, thaler and johnson (1990) found that participants with previous wins made different decisions from those participants with previous losses. the findings suggest that individuals increase their tolerance for later risks after a successful risk taking past, and reduce their risk tolerance after an unsuccessful risk taking past. as tversky and kahneman (1986) pointed out, if an individual does not have a history or recollection of a negative event, that individual will expect that the event’s likelihood is low. if a loss or negative event is recalled easily, the individual is more likely to consider and alleviate downside risk. an individual with both negative experience investing, as well as a past orientation, would be expected to be very risk averse, although there is currently no empirical research in this area to substantiate this hypothesis. thaler and johnson’s (1990) findings that investors reduce their risk tolerance after an unsuccessful risk taking past seems to contradict the behavior of rogue and day traders. rogue traders increase the stakes with each prior loss, a phenomenon that shefrin and 382 b.s. poole / financial services review 9 (2000) 375–387 statman (1985) coined “get-even-itis.” the rogue traders had significantly higher stakes at risk than the student participants in the thaler and johnson gambling simulations, in fact these traders’ entire livelihoods were at risk. the risk involved may increase the traders’ adrenaline, which could increase their trading pace. the rogue traders also may be more likely to be type a, and have a short time perspective, as well as exhibit optimism bias and overconfidence, concepts discussed further in the following section. further study in the temporal psychology area may provide clues to identify potential rogue traders and the circumstances under which rogue trading may begin. the study may also provide further insight into the behavior of day traders. 4.2. focus on the future psychologists traditionally have recognized the importance of an individual’s vision of the future. kelly’s (1955) personal construct theory suggested that anticipation of future events is the primary focus of individual behavior and decision-making. maslow’s (1954) hierarchy of needs implies a present orientation in satisfaction of lower-level needs such as a food and safety, and a future orientation for the higher-level needs of achievement and self-actualization. even as he discussed finding concrete meaning in the present, frankl (1963, p.166) spoke of striving toward the future, “the call of a potential meaning waiting to be fulfilled. . . ” . kübler-ross (1999), who spent her scholarly life working with the dying, maintained a future focus, denying the existence of death, asserting that death is simply a passage to an after life. individuals with a future orientation are not only more likely to be emotionally healthy, but also more self satisfied and in control, compared to their past oriented counterparts. braley and freed (1971) found that future focused outpatients, compared to past-present focused outpatients, expressed significantly more satisfaction with themselves. when asked their ideal temporal orientations, both the past and future oriented groups did not differ significantly, agreeing on the preference for a future orientation. future orientation has been associated with coping skills, ability to delay gratification, and self-control (melges, 1963). a future orientation can be unhealthy if the individual does not perceive a sense of flow from present to future. braley and freed (1971) found that of their control group, those with extreme future orientations were more dissatisfied with themselves. braley and freed cautioned that “if the temporal orientation is focused too far into the future, the person fails to make sufficient contact with the present to establish a comfortable, and meaningful, sense of continuity” (p. 38). this observation suggests the importance of setting shorter-term as well as longer-term goals, and in showing that a client can, over time, meet long term goals. orientation on the present or future is typically characteristic of individuals with time possessiveness, a calabresi and cohen (1968) category of temporal attitudes. associated characteristics include greed toward time, upset over the passage of time, intolerance of wasting time, and disinterest in thinking of the past. a natural bias toward optimism, or wishful thinking, can affect individuals’ views of their futures (weinstein, 1980). behavioral economists have identified investors’ overconfidence in their predictions of the future, as well as over-reliance in the past as a predictor of the future (kahneman & riepe, 1998). individuals rely on their pasts when forming their expectations of the 383b.s. poole / financial services review 9 (2000) 375–387 future, but only to the extent that the past supports what they would like to believe. the further into the future that the individual considers, the greater the optimism bias (björkman, 1984). ito (1990) found that optimism bias varies based on individual characteristics and goals, despite similar past experiences. because of optimism bias, adults can develop persistent positive expectations of the future despite past negative experience (anderson & goldsmith, 1994). as a result, they are willing to assume higher risks for longer term investment horizons. planning activities are a means of attempting to control the future. in fact, “often the planning activities can give an exaggerated feeling of control of future events” (björkman, 1984, p. 35). thus planning can contribute to an individual’s feelings of overconfidence about future events. individuals weigh the strength of negative past experience against the desire to be optimistic. individuals with both a past temporal orientation and past negative experience may be more likely to develop pessimistic expectations of the future. however, no study has combined temporal orientation with past experience and future expectations. researchers have confirmed that time orientation and pace are inter-related. type a personalities are not only fast paced, but also preoccupied with future deadlines (rao, reddy & samiuliah, 1997). researchers have found that faster temporal pace is more likely to be associated with future oriented individuals (siegman, 1961). wyrick and wyrick (1977) found that the greater the pastpredominant orientation, the slower the temporal pace, and frederickson (1988) found a significant inverse relationship between past temporal orientation and fast temporal pace. 4.3. summary succession refers to the sense of past, present, and future, and individuals tend to have a dominant focus, or temporal orientation, toward the past or future. an individual’s temporal orientation can be associated with culture, personality type, and psychological well being. researchers have confirmed that time orientation and pace are inter-related. 5. conclusion much about market behavior remains unexplained by the traditional finance literature. this paper has looked to the anthropology and psychology literature to understand more about what is known about time, an important factor in finance. in this concluding section, possible implications of this literature for financial researchers are discussed. this paper began with a discussion of cultural representations of time. finance researchers model financial changes across a time horizon, creating ‘market time’ representations. finance researchers in the west have gone beyond the linear western concept of time in modeling financial market behavior. nonlinearity, business cycles, and structural changes that shift market movements across time are all commonly accepted. in all fields, researchers need to be aware of their own cultural biases. the extent that a researcher’s own representation of time can constrain attempts to model market time is an interesting challenge for researchers. the paper continued with a discussion of pace, which could have an impact on financial behavior. the research tells us that a fast paced individual is more likely to make quick 384 b.s. poole / financial services review 9 (2000) 375–387 decisions and to take immediate action. we also know that individuals adjust their temporal pace to match the environment. in the context of securities trading, a fast paced environment may foster higher trading levels and more frequent trading activity. a day trader may trade more frequently in a fast paced group environment than in solitude. research on the impact of individual and environmental pace on trading behavior might provide some insights into differences in trading volumes. environmental factors that quicken pace could also influence market behavior by accelerating trading. for example, psychologists know that a cold temperature encourages a faster pace, and the anthropologists have found an association between a cooler climate and faster cultural pace. in a trading environment, simply lowering the temperature might encourage a faster pace. wafting a tempting aroma of coffee to encourage caffeine intake might also step up the pace. although the marketing literature has looked at the relationship between sales environments and sales performance, this idea has not been raised in the finance arena. whether simple changes in the environment could lead to changes in trading volume and speed is not known. finally, this paper discussed temporal orientation. finance researchers have long recognized the importance of expectations of the future on financial behavior. much work in finance has been done to assess the expectations of the aggregate market. knowing the relative representation of past and future oriented individuals participating in the marketplace may prove helpful in explaining aggregate formation of expectations. study in the area of temporal orientation also may provide insight into the degree and direction, indicating optimism or pessimism, of market reactions to events. further research in the temporal psychology area may also provide additional insight into market behavior. one academic has suggested that increasingly widespread use of antidepressants for psychological treatment may be associated with recent market behavior. as indicated earlier, depressed patients experience a slow inner tempo, and exhibit a slow behavioral pace. treatment of depression tends to improve the patient’s expectations of the future, as well as to increase the patient’s pace. we already know that optimism is associated with increased risk tolerance. again, this might result in increased willingness to participate in the market, as well as willingness to pay higher prices in anticipation of high returns. nesse, a psychiatry professor, has speculated that the recent unusually long bull market may be related to an increase in the use of antidepressants, especially by well-to-do, highly stressed individuals. he suggested that “if investor caution is being inhibited by psychotropic drugs, bubbles could grow larger than usual before they pop, with potentially catastrophic economic consequences” (johnson, 2000). this article raised some possible applications of temporal psychology to problems typically studied in finance. researchers are encouraged to consider other roles that representation, pace, and temporal orientation may play in their areas of interest. references anderson, m. a., & goldsmith, a. h. (1994, fall). economic and psychological theories of forecast bias and learning: evidence from u. s. business managers’ forecasts. eastern economic journal, 20, 413–427. björkman, m. (1984). decision making, risk taking and psychological time: review of empirical findings and psychological theory. scandinavian journal of psychology, 25, 31–49. 385b.s. poole / financial services review 9 (2000) 375–387 boud, d. (1993). experience as the basis for learning. higher education research and development, 12(1), 33–44. braley, l. s., & freed, n. h. (1971). modes of temporal orientation and psychopathy. journal of consulting and clinical psychology, 36(1), 33–39. calabresi, r., & cohen, j. (1968). personality and time attitudes. journal of abnormal psychology, 73(5), 431–439. ericsson, k. a., & smith, j. (eds.). (1991). toward a general theory of expertise. ny: cambridge university press, frankl, v. e. (1963). man’s search for meaning. ny: washington square press, simon and schuster. frederickson, c. g. (1988). temporal experience: a two component model. perceptual & motor skills, 66(1), 63–68. hall, e. t. (1983). the dance of life: the other dimension of time. ny: anchor press/doubleday. hawking, s. w. (1988). a brief history of time. ny: bantam. hunt, d., & hait, p. (1990). the tao of time. ny: simon & shuster. ito, t. (1990, june). foreign exchange rate expectations: micro survey data. american economic review, 434–449. johnson, l. (2000, oct. 1). a market boom born of fear. sunday telegraph (london), p. 4. kahneman, d., & riepe, m. w. (1998). the psychology of the non-professional investor. journal of portfolio management, 24(4), 52–65. kelly, g. a. (1955). the psychology of personal constructs. 2 vols. ny: norton. klein, g. (1998). sources of power. cambridge, ma: mit press. koco, l. (2000, oct. 23). slow down, educate clients, lynch tells nava. national underwriter, 104 (43), pp. 3, 46. kübler-ross, e. (1999). the wheel of life: a memoir of living and dying. ny: simon & schuster. levine, r. (1997). a geography of time. ny: basic books. levine, r. v., & norenzayan, a. (1999). the pace of life in 31 countries. journal of cross-cultural psychology, 30(2), 178–205. levinson, d. (1998). religion: a cross-cultural dictionary. london: oxford university press. maslow, a. h. (1954). motivation and personality. ny: harper & row. meade, r. d. (1971). future time perspectives of college students in america and in india. the journal of social psychology, 83, 175–182. meade, r. d. (1972). future time perspectives of americans and subcultures in india. journal of cross-cultural psychology, 3(1), 93–99. meade, r. d., & singh, l. (1970). motivation and progress effects on psychological time in subcultures of india. the journal of social psychology, 80, 3–10. melges, f. t. (1982). time and the inner future. ny: wiley. msumange, g. (1998). timethe concept of future in the thought of the ‘hehe’ and other african tribes. unpublished manuscript, salvatorian major seminary, morogoro, tanzania. rao, g. b., reddy, k. s., & samiuliah, s. (1997). behaviour activity profiles and work values of employees. social science international, 13, 19–24. rotter, g. s. (1969). clock speed as an independent variable in psychological research. the journal of general psychology, 81, 45–52. schkade, d. a., & kilbourne, l. m. (1991). expectation-outcome consistency and hindsight bias. organizational behavior and human decision processes, 49, 105–123. shefrin, h., & statman, m. (1985). the disposition to sell winners too early and rider loses too long: theory and evidence. journal of finance, 40, 777–790. siegman, a. w. (1961). intercorrelation of some measures of time estimation. perceptual and motor skills, 14, 381–382. taylor, f. (1947). the principles of scientific management. ny: norton. thaler, r. h., & johnson, e. j. (1990). gambling with the house money and trying to break even: the effects of prior outcomes on risky choice. management science, 36, 643–660. 386 b.s. poole / financial services review 9 (2000) 375–387 tsukanov, b. i. (1991). the quality of an “inner clock” and the problem of intelligence. soviet journal of psychology, 12(3), 34–42. tversky, a., & kahneman, d. (1986). rational choice and the framing of decisions. journal of business, 59(4), s251–s278. weinstein, n. (1980). unrealistic optimism about future life events. journal of personality and social psychology, 39(5), 806–820. wyrick, r. a. & wyrick, l. c. (1977). time experience during depression. archives of general psychiatry, 34, 1441–1443. 387b.s. poole / financial services review 9 (2000) 375–387 pii: 1057-0810(93)90004-a financial services review, 3(l): 29-43 copyright (9 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any fom mserved. market timing for the individual investor: using the predictability of long-horizon stock returns to enhance portfolio performance steven p. rich william reichenstein recent research indicates that dividend yield and earnings-price ratio can partially predict long-horizon stock returns. we examine whether individual investors can successfidly construct timing portfolios based on either of these variables or a measure of the expected market risk premium. the out-of-sample tests in this study require that investors rely only on information that was available at the time of the ~r~t-t~~g decision. timing portfblios based on the market risk premix show the strongest ab~i~ to rime the market. we present an eco~mic ra~o~e for the results that is consistent with @scent ~~~. i. introduction several studies by leading financial economists conclude that long-horizon stock returns are partially predictable (e.g., campbell & shiller, 1988; fama & french, 1988a,b, 1989; and fama, 1991). these studies suggest that up to 25% of the variation in long-horizon returns on large (e.g., s&p) stocks is predictable. a larger portion of long-horizon small stock returns appears predictable. this study investigates the impli~tions of the market ~~ic~bili~ literature for market timing by individu~ investors. can the individual investor construct timing portfolios that outperform nontiming benchmark portfolios where the timing portfolios are based on variables that show a consistent relation with long-horizon stock returns? we address this question by constructing timing portfolios that combine the s&p index and treasury bills and, separately, portfolios that combine steven p. rieb l hankamer school of business, baylor university, waco, tx 76798-8004. william reiekenstein l baylor university, waco, tx 76798-8~. 30 financial slmv1ces revijzw, 3(l) 1993 small stocks and treasury bills. we then calculate several measures of portfolio performance, including raw return and risk-adjusted return. by every criterion, timing portfolios based on a measure of the expected market risk premium outperform s&p benchmark portfolios. of the small-stock timing portfolios, those based on the expected market risk premium perform best, but portfolios based on dividend yield and earning-price ratios also show some ability to time small stock returns. finally, we discuss the limitations of this study and their implica~ons for individ~ investors that plan to follow a market timing strategy. ii. review of berature several financial studies indicate that long-horizon stock returns are partially predictable. many economists argue that this predictability reflects the action of a rationally priced stock market with a time-varying market risk premium. simply put, average stock returns tend to be generous when the market is priced to offer a generous risk premium. nobel-la~~tc william sharpe, who sells market timing advice based on a relative market risk premium, expresses this view, “this is not a story that the market is screwed up. this is a story of how investors behave when their wealth changes. say i tell you the relative risk premium is 120-you’ll get 120% of the normal expected return for bearing risk. if that was the only thing i told you, you’d say, ‘let’s buy stocks.’ but what if i also told you the reason for that is we’re in a recession and someone in your family is out of work. if you’re average, you’ll say, ‘i guess these are offsetting. i’ll stay with what i’ve got.’ in fact, that’s what has to happen in an efficient market. the risk premium has to adjust to where the average investor doesn’t want to do anything” (berss, 1990, p. 78). fama and french also support this view. they conclude that the market risk p~~urn is highly autocorrelated, but slowly mean reverting’; that is, when the risk premium is large it tends to remain large for several quarters while slowly reverting to its historic mean, and vice-versa. this produces a predictable component of stock returns that increases roughly proportionately with the return horizon out to one or two years. we offer the following illustration of their story. let p represent the price of an index of stocks. for simplicity, we assume the zero growth model with projected year-ahead dividends d, equal to projected earnings e,. that is, d, = e, . in this model, where the required rate of return k = r + mrp, r denotes the risk-free interest rate, and mrp denotes the unobservable market risk premium. it follows that market timing for the zndividual znvestor 31 d,/p = e,ip = r + mrp. ub) as equation 1b shows, both dividend-price and earnings-price ratios perfectly reflect shocks to the market risk premium that are independent of dividend and earningsforecasts. these ratios are not important per se, but only to the degree that they reflect movement in the unobservable market risk premium. thus theory suggests that the predictive content of historic dividend-price and earnings-price ratios reflects their tendency to proxy for the unobservable market risk premium. suppose the risk-free interest rate is 5% and the market risk premium is 8%, its historic average. a shock raises the premium to 9.3%. the rise in the market premium drives the current stock price from $7.7d to $7d. if the premium followed a random walk then stocks would now offer a permanently higher expected return. however, the market premium slowly reverts to its mean. during the reversion, the expected risk premium exceeds its historic mean. after the reversion, the market risk premium is again 8% and the stock price $7.7d. this economic story can explain the evidence supporting the predictability of long-horizon stock returns. 1. model 1b explains why dividend yield and earnings-price ratio can predict stock returns. more generally, any variables that move with the time-vary ing market risk premium should be able to predict long-horizon stock returns. 2. mcqueen and thorley (199 1) found that a bad year in the stock market tends to follow two good calendar years and a good year tends to follow two bad years. this pattern is expected if movements in the market risk premium strongly influence individual year stock returns. 3. there appears to be a transitory, and thus predictable, component of stock price. in the example above, the shock to the market risk premium produced a transitory drop in stock price that is eventually offset by subsequent gains. equivalently, the mean-reverting market risk premium gives rise to negative autocorrelations in distant stock returns; the rise in the market risk premium produces a current period loss that is offset by higher subsequent returns. iii. construction of timing portfolios prediction variables in this study, we examine whether individual investors can successfully construct timing portfolios using three variables shown to predict stock returns. the first of the variables is the end-of-quarter earnings-price ratio on the standard and poor’s composite index, ep. the second is the median year-ahead dividend yield as published by value line, yield. 32 fecal services avow, 3(l) 1993 rp s&p 0.4 0.8 0.35 0.6 0.3 0.4 025 0.2 02 0 0.15 9.2 0.1 0.05 9.4 0 -0.6 66 6b 70 71 72 73 74 75 76 77 78 73 80 61 62 63 84 85 86 87 8b 0.4 0.35 0.3 0.25 0.2 0.15 0.t 0.05 0 rp i j 0.e 0.6 ” 0.4 0.2 0 -0.2 -0.4 -0.6 “.” 66 68 70 71 72 73 74 75 76 77 76 79 80 61 62 63 64 65 86 87 69 figure i. relationship between rp and cumulative six-quarters-ahead excess returns on s&p and small stock indexes the third predictor is a direct measure of the expected market risk premium: rp = (yzelz3 + capgains) r. cap~i~s denotes the median ammal capital gain return and is calculated (1 + appr~~“.*5 -1 where apprec is the median three-to-five year price appreciation potential as forecast by vazue line. the variable r denotes the bond-equivalent yield on three-month treasury bills. value market timing for the individual investor 33 line’s “summary and index” highlights yield and apprec weekly on the cover page. figure 1 shows the close relationships between rp and cumulative six-quarters-ahead excess returns on both the s&p and small stock indexes. the rp frequently hit a peak immediately prior to a strong six-quarter return. for example, rp reached a peak at the end-of-september 1974. this correctly predicted the peak excess s&p return from october 1, 1974, through march 3 1, 1976. the rp offers a number of important advantages compared to the ratio of past-four-quarters dividends to current stock price, the ratio most prevalent in prior studies of stock predictability. 1. it is a direct measure of the unobservable market risk premium, the alleged source of predictive content. 2. it uses forecasts of dividends and returns instead of historic information, and, not surprisingly, forecasts provide a better measure of expectations. 3. value line forecasts are generally accessible to the individual investor. in a recent study, reichenstein and rich (1992) examined the ability of ep, yield, and rp to explain s&p and small stock returns for 1968-1988 and for two equal subperiods. the market risk premium rp is the only candidate to show a consistent in-sample relation with s&p returns. the risk premium rp, earnings price ratio ep, and dividend-price ratio yield all show consistent in-sample relations with long-horizon small stock returns. the next section describes the timing strategies considered in this study. market timing strategies each quarter from the third quarter of 1978 (1978.3) through the fourth quarter of 1988 (1988.4) we form out-of-sample timing portfolios based on the three prediction variables. since rp is the only candidate to show a consistent in-sample relation to s&p stock returns, we determine whether it can enhance large (i.e., s&p) stock returns in out-of-sample tests. since all three show a consistent relation to in-sample small stock returns, we determine whether any of them can enhance small stock returns in out-of-sample tests. we examine two basic timing strategies, the first, the all-or-nothing strategy, calls for a portfolio that each quarter is either 100% in stocks or 100% in treasury bills. each quarter a regression of 1968.2-to-date stock returns on the prediction variable determines whether the timing portfolio invests in stocks or bills.2 consider the general regression: r(t,t+n-l)=a+b*x(t-l)+e(t), (2) where r(t, t + it 1) is the n-quarters-ahead market risk premium (market return less treasury bill return), x(t 1) is the time t-l value of the prediction variable, and 34 financial services review, 3(l) 1993 e(t) is the regression residual. if the fitted value 2 + &x(t 1) is positive, the all-or-nothing portfolio consists of stocks; if negative, it consists of treasury bills. for example, the 1968.2-1977.1 regression of six-quarter small stock excess returns on the value line market risk premium is:3 r(b, t+5) =-o-2717 + 1.83rp(b1). if the end-of-june 1978 value of rp exceeds 0.148 (0.2717/1.83), then the fitted value is positive and the timing portfolio invests in small stocks for the 1978.3 quarter. if less than 0.134, then it invests in bills. we update the regressions every four quarters.4 the forecast horizon in (2) is set at six quarters for rp and eight quarters for the other candidates.5 we compare all-or-nothing portfolio returns to that on a buy-and-hold benchmark portfolio that always allocates 100% to stocks. the second timing strategy, the variable-weights strategy, better reflects timing strategies in practice in that it allows portfolio weights of stocks and bills to vary according to market prospects, but it does not require either the 100% stocks or 100% bills extremes. this strategy calls for the timing portfolio to contain 25% stock when the market predictor signals “poor” stock prospects, 50% stock when it signals “average” prospects, and 75% when it signals “good” prospects. the balance of the portfolio is invested in t-bills. the benchmark portfolio begins each quarter with bill-stock weights of 50% each. stock prospects depend upon whether the prediction variable is high, average, or low by historic standards. poor and good stock prospects occur when the variable falls in, respectively, the bottom and top fourth of its 1968.1”to-date distribution. average stock prospects occur when the variable falls in the middle 50% of its distribution. distributions are updated every four quarters. table 1 suites the 1%8.1-1988.3 ~st~butions of the predictive vari ables and the 1968.2-1988.4 (continuously compounded) quarterly risk premiums on the s&p and small stocks. it provides historic perspective for individual investors who wish to use the variables. table 1. summary statistics: 19684988 variable mean rp 13.74% ep 8.77% yield 4.07% s&p 2.40% small stock 3.07% st. dev 6.34% 2.80% 1.04% 8.82% 13.62% 1st q 10.10% 6.08% 3.2% -2.61% -5.02% median 12.72% 8.37% 4.1% 3.05% 1.82% 3dq 16.91% 11.09% 4.9% 8.31% 12.19% note: rf’, ep, and yield refer to the value line estimates of the market risk premium, earnings-price ratio on s&p 500, and value line estimate of median year-ahead dividend yield. s&p and small stock show the distribution of quarterly risk pnxniuzn-total returns less treasury bill returns. market timing for the individual investor 35 s&p index rp benchmark equity weights of timing portjolios all-or-nothing variable weights 0 or 100% stock 25,50 or 75% stock 100% stock 50% stock small stocks all-or-nothing variable weights rp 0 or 100% stock 25,50, or 75% stock ep 0 or 100% stock 25,50, or 75% stock yield 0 or 100% stock 25,50, or 75% stock benchmark 100% stock 50% stock each portfolio contains stocks and/or treasury bills. the figure shows the equity weights. figure 2. outline of timing strategies the average annual risk premium (rp) of 13.74% equals a continuously compounded quarterly return of 3.22%. this exceeds both the average small stock premium 3.07% and average large stock premium 2.40%. we attribute this upward bias to the optimistic nature of investment advisory services. nevertheless, the results of this study suggest that movements in rp mirror movements in the unobservable expected market risk premium. figure 2 summarizes the market timing andbenchmark portfolios. for the s&p index, we compare returns on two rp-based portfolios with returns on their benchmark portfolios. the all-or-nothing portfolio contains either 0% or 100% stock. its returns are compared to the returns on a buy-and-hold 100% stock portfolio. the variable-weights timing portfolio contains 25%, so%, or 75% stock. its returns are compared to the returns on a constant-weights 50% debt-50% stock portfolio. for small stocks, we examine separately the market timing ability of rp, ep, and yield. iv. overview of tests in this section we examine whether rp can be used to enhance portfolio returns using large (i.e., s&p) stocks and whether rp, ep, or yield can be used to enhance portfolio returns using small stocks.6 the 1978.3-1988.4 out-of-sample perform ance of each timing portfolio is compared to the performance of a benchmark portfolio. we use five performance criteria to determine whether rp, ep, or yield can be used to time the market. the criteria include raw return, risk-adjusted return, and three capm-based regressions of timing ability. 36 financial services review, 3(l) 1993 s&p index rp benchmark small stock index rp ep yield benchmark an or nothing 1.49% 38.1% 1.38% 100.0% an or nothing 2.22% 54.8% 1.63% 76.2% 1.62% 45.2% 1.74% 100.0% variable weights 1.04% 47.0% 0.69% 50.0% variable weights 1.43% 47.0% 1.12% 55.4% 1.35% 42.3% 0.87% 50.0% note: all returns are continuously compounded. thus mean returns are geometric means. mean quarterly risk premium equals the average of portfolio returns minus the return on t-bills. all-or-nothing timing portfolio has either 100% stock or 100% treasury bills. its benchmark is a buy-and-hold stock portfolio. variable weights timing portfolio has weights of 75%. 50%. or 25% on stocks and the rest in treasury bills. its benchmark portfolio always maintains a 5c50 stock-bills mix. raw return table 2 presents the raw (unadjusted for risk) returns on timing and benchmark portfolios. the mean quarterly risk premium denotes the average of the portfolio return less the return on treasury bills. s&p index, both the rp-based all-or-nothing and variable-weights timing port folios earned higher raw returns than their benchmark portfolios. the mean risk premium for the all-or-nothing timing portfolio was 1.49%. its benchmark portfolio earned 1.38%. the mean risk premium on the variable-weights timing portfolio (1.04%) also exceeds its benchmark portfolio (0.69%). the all-or-nothing portfolio managed to earn a higher raw return than the buy-and-hold benchmark, despite being out of the market 62% of the time. small stocks. among the all-or-nothing timing portfolios, only the rp-based portfolio earned more (2.22%) than the benchmark (1.74%). it earned higher returns despite being out of the market about half of the time. the ep-based (1.63%) and yield-based (1.62%) all-or-nothing portfolios failed to match the benchmark. all of the variable-weights timing portfolios earned more than the benchmark. with the exception of the ep-based portfolio, the timing portfolios earned higher return despite a lower average exposure to stocks. the rp-based portfolio earned the highest return (1.43%). figure 3 shows that the higher raw returns on the timing portfolios produce substantial long-run wealth implications. for example, $10,000 invested in the variable-weights s&p benchmark portfolio (i.e., 50% s&p and 50% t-bills) at the beginning of the third quarter of 1978 would have grown to $33,303 by the end of 38 financial services review, 3(l) 1993 1988. however, had the $10,000 been invested in the rp-based variable-weights timing portfolio instead, it would have grown to $38,711. the results from timing with small stocks are even more impressive. the benchmark portfolio would have grown to $34,454 while the rp-based variable-weights portfolio would have grown to $43,722. as figure 3 shows, the gains from timing accrued slowly but steadily over the entire period. risk-adjusted return valid performance comparisons across portfolios should consider differences in portfolio risk. the sharpe (1966) ratio incorporates these differences by dividing the mean quarterly risk premium by its standard deviation, thus providing a measure of return per unit of risk. table 3 presents the sharpe ratios for the timing and benchmark portfolios. s&p index. both rp-based timing portfolios outperform their benchmark. the all-or-nothing portfolio earned twice as much return per unit of risk as the bench mark, 0.328 versus 0.163. the variable-weights portfolio earned 60% more return per unit of risk, 0.266 versus 0.163. small stock. all of the timing portfolios earned higher risk-adjusted returns than their benchmark portfolios. however, the rp portfolio (0.3 19) earned much larger risk-adjusted returns than the yield portfolio (0.180), its nearest competitor. among the variable-weights portfolios, rp again performed best (0.244) followed table 3. risk-adjusted returns: sharpe ratios portfolio an or nothing variable weights s&p index rp 0.328 0.266 benchmark 0.163 0.163 small stock index rp 0.319 0.244 ep 0.165 0.160 yield 0.180 0.234 benchmark 0.143 0.143 note: all returns are continuously compounded. thus mean returns are geomet ric means. the sharpe ratio reflects return per unit of risk and is defined as the ratio of mean-to-standard deviation of quarterly excess return. all-or-nothing timing portfolio has either 100% stock or 100% treasury bills. its benchmark is a buy-and-hold stock portfolio. variable-weights timing portfolio has weights of 75%,50%, or 25% on stocks and the rest in treasury bills. its benchmark portfolio always maintains a s&50 stock-bills mix. market isming for the individual investor 39 closely by yield (0.234). both rp portfolios earned at least 70% more return per unit of risk than the benchmark portfolios. regression tests of timing ability one weakness of the sharpe ratio is that it cannot test the statistical signifi cance of the performance differences between timing and benchmark portfolios. the three regressions presented here provide such tests. estimates of jensen (1968) regressions take the following form: rpt r--t = a + b(r,, r-ff) + e, where rpt denotes the portfolio return, and rpt rrt is the quarterly risk premium earned by the timing portfolio, r,,,, rfr is the market risk premium, e, is the error term, and a and b are regression coefficients. a positive intercept, a, indicates that the timing portfolio beats the benchmark portfolio on a risk-adjusted basis. a negative intercept indicates inferior performance. in general, a positive intercept can result from consistent selection of undervalued stocks, successful market timing, or both. for the portfolios considered here, a positive intercept indicates successful market timing since no individual stocks are selected. estimation of quadratic and dummy regressions provide direct tests of timing ability. they take the forms: rpt rrt = a + b(r,, rrt> + c(rti r-,t)’ + e, (quadratic) and rp, r-f, = a + b(r,, rfr) + c[w, rrt)l + e, (dummy) where d, = 0 if r, > rr, and -1 if r,,,, < rrt. in both regressions a positive “c” coefficient indicates successful timing ability.’ table 4 presents the results. s&p index. the regressions indicate superior timing for the bp-based all-or-noth ing portfolio. the jensen intercept of 1.09% (based on a quarterly-returns regres sion) indicates a compound annual return advantage of 4.43%. the positive “c” in the quadratic regression indicates a tendency for the timing portfolio’s beta to rise with the market return. the positive “c” in the dummy regression indicates that the average beta in an up market (i.e., r, > rfr) was 0.46 more than the average beta in a down market. this indicates that the timing portfolio tended to invest in stocks during an up market and in bills during a down market. both “c” coefficients imply timing ability at significance levels better than 10%. the regressions lend even stronger support to the claim of timing ability in the variable-weights portfolio. the “c” coefficients in the quadratic and dummy regressions show significance beyond the 1% level. the jensen intercept of 0.45% indicates an annual premium of 1.8%. flnancialservicesreview,3(1) 1993 table4. regression tests of timing strategies all or nothing variable weights portjxo aorc t-statistic aorc t-statistic s&p index rp jensen 1.09% 1.80* 0.45% 1.93* quad. 1.01 2.11* 2.93 4.65*** dummy 0.46 2.01* 0.62 3.94*** small stock in&x rp jensen 1.65% 1.85* 0.65% 1.97* quad. 1.44 3.99*** 2.73 5.93*** dummy 0.77 3.72*** 0.68 4.80*** ep jensen 0.48% 0.53 0.19% 0.47 quad. 0.36 3.29*** 1.73 2.41** dummy 0.50 2.12** 0.27 1.29 yield jensen 0.68% 0.71 0.59% 1.73* quad. 0.50 1.71; 1.69 2x5*** dummy 0.22 0.83 0.37 2.15** notes: * significant at 10% level. **significant at 5% level. *** significant at 1% level jensen’s regression: rpl rfr = a + b(rm rfr) + et quadratic regression: rpr rfr = a + b(r,,,t r@ + c(rmr r@2 + et dummy variable regression: rp rfr = a + b(rm ry) + c[dl(rm r@] + e, all returns am continuously compounded. thus mean returns am geometric means. all-or-nothing timing portfolio has either 100% stock or 100% treasury bills. its benchmark is a buy-and-hold stock portfolio. variable-weights timing portfolio has weights of 7546, 5046, or 25% on stocks and the rest in treasury bills. its benchmark portfolio always maintains a 50-50 stock-bills mix. small stock. each of the all-or-nothing timing portfolios demonstrates statisti cally significant timing ability at the 10% level or better at least once. however, the rp-based portfolio produced the strongest evidence of timing ability for each regression type. it produced a jensen intercept of 1.65% (6.76% annually) and “c” coefficients of 1.44 and 0.77 in the quadratic and dummy regressions. these are at least 42% larger than the nearest competitor. only the rp portfolio shows significant timing ability at the 10% level or better across all regressions. a similar story prevails for the variable-weights portfolios. all of these portfolios show significant timing ability at the 10% level at least once. both rp and yield show significant timing ability across all regression types. however, the rp portfolio consistently produces the strongest results. v. summaryand~mplications long-horizon stock returns are partially predictable. the $100,000 per year charged by william sharpe for his asset allocation advice suggests that pension funds can market timing for the individual investor 41 exploit this predictability (berss, 1990). but can the individual investor use market predictability to effectively time the market? the evidence presented here suggests that they can. the results imply that individual investors who specialize in large (i.e., s&p) stocks can successfully time the market with rp-an estimate of the expected market risk premium based on value line forecasts. the all-or-nothing timing portfolio beat the buy-and-hold s&p benchmark by every criterion. it earned higher raw returns, twice the risk-adjusted return, and it generated evidence of significant timing ability at the 10% level in all three regression tests. similarly, the rp-based variables-weights portfolio beats its s&p benchmark by every criterion. all of the regression tests support the claim of timing ability at the 10% level and two of the three tests were significant at the 1% level. the results also imply that investors can use rp to successfully time small stock returns. among the all-or-nothing strategies, the rp portfolio outperformed by every criterion the small stock benchmark portfolio. it also outperformed timing portfolios based on dividend yield (yield) and earnings-price ratio (ep). among variable-weights strategies, the rp portfolio performed best by every criterion. compared to the nontiming benchmark portfolio, it earned a 2.3% larger annual raw return and 70% larger risk-adjusted return. the rp portfolio also showed significant timing ability at the 1% level in two of the three tests and at the 10% level on the remaining test. yield beats the benchmark by every criterion, and ep beats it by most criteria. they all demonstrate significant timing ability at the 5% level in at least one regression test. why did rp, and to a lesser extent yield and ep, succeed in enhancing portfolio returns? more important, are they likely to do so in the future? we believe that the literature review on the predictability of long-horizon stock returns provides the answer. if the market risk premium varies through time, long-horizon stock returns should be predictable. these variables should move with the unobservable market risk premium and continue to predict long-horizon stock returns. will rp continue to predict as well in the future? possibly. in 1990, we noticed that value line’s prediction of median three to five year appreciation potential hit a low in august, 1987, shortly before the crash. this drew our attention, encouraged the development of the rp model, and later we saw how the model fit into the stock predictability literature. as we are all aware, a model usually performs better in the period in which its predictive content is “discovered” than in later periods. however, we believe, that rp will continue to predict returns better than dividend yield and earnings-price ratio because it is a direct measure (although possibly biased) of the unobservable market risk premium. finally, how could an individual implement an rp-based strategy? one approach would be to calculate rp, perhaps every three months, and to compare its current value to its historic distribution as presented in table 1. if rp is well below average, perhaps below the first-quartile value of 10.1 %, then the portfolio’s equity 42 financial services review, 3(l) 1993 weight can be reduced below the investor’s long-run target equity weight. if rp is well above average, perhaps above the third-quartile value of 16.9%, then equity weight can be increased. this type of strategy would allow an individual to increase equity exposure when the rewards to bearing stock market risk appear above average and to lower exposure when the rewards appear below average. in so doing, we believe that an investor will likely increase the portfolio’s long-run return without increasing the average risk exposure. we caution, however, against an all-or-nothing timing strategy and other timing strategies that require sharp swings in the portfolio’s equity position. to illustrate why, suppose an individual investor who maintains constant portfolio weights of 50% stock-50% debt. the stock predictability literature encourages varying the equity weight around 50%. but how far should the investor allow the weights to vary? the answer requires balancing the benefits of diversification across assets against the benefits of market timing. nobel-laureate paul samuelson (1990) recommends a modest range for the equity weights of perhaps 40% to 60% around the 50% long-run target weight. in essence, he argues that the benefits of diversifi cation across assets are more certain than the benefits of market timing. other economists would encourage a wider range. in summary, the stock predictability literature suggests that the individual investor might benefit from varying the equity weight around the long-run target weight based on market conditions, but the investor should probably exercise moderation in establishing the acceptable range of equity weights. acknowledgments: we acknowledge the helpful comments of an anonymous referee, frank l. page, and other participants at the academy of financial services 1992 annual meeting. notes 1. they conclude that nominal, real, and excess stock returns are highly autocorrelated, but slowly mean reverting. excess returns (returns less treasury bill returns) prove most useful for market timing purposes. see fama and french (1988a). 2. value line forecasts began at the end-of-quarter 1968.1. the 1968.1 value of x forecasts 1968.2 and later returns. value line did not publish median appreciation potential for a few weeks surrounding the 1972.3 quarter. thus rp has one missing observation. 3. the last observation for the dependent variable covers 1977.1-1978.2. the last observation for rp reflects the end-of-december 1976 value. 4. the regression results for 1968.2-1987.1 yield a critical value of 0.1037 (0.1819/1.7545). 5. these forecast horizons are set to maximize 1968.2-1978.2 predictive content. see reichenstein and rich (1992). 6. the small stock portfolio data came from ibbotson associates (1992). 7. for a discussion of the quadratic regression methodology see treynor and mazuy (1966) and admati, bhattacharya, pffeider, and ross (1986). for a discussion of the dummy variable regression methodology see henriksson and merton ( 198 1). market timing for the individual investor 43 itjwjzrences admati, anat r., sudipto bhattacharya, paul pfleider, and stephen a. ross. 1986. “on timing and selectivity,” journal offinance, 41:715-730. berss, m. r. 1990. “modem portfolio timing,” forbes, december 2476-78. campbell, john y., and robert j. shiller. 1988. “stock prices, earnings and expected dividends,” journal of finance, 43~661676. fama, eugene f. 1991. “efficient capital markets: ii,” journal of finance, 46:1575-1617. fama, eugene f., and kenneth r. french. 1989. “business conditions and expected returns on stocks and bonds,” journal of financial economics, 25:3-49. fama, eugene f., and kenneth r. french. 1988a. “dividend yields and expected stock returns,” journal of financial economics, 22:3-25. fama, eugene f., and kenneth r. french. 1988b. “permanent and temporary components of stock prices,” journal of political economy, 961246273. henriksson, roy d., and robert c. merton. 1981. “on market timing and investment performance ii. statistical procedures for evaluating forecasting skills,” journal of business, 54513-533. ibbotson associates. srocks, bonds, bills, and in&?&on 1992 yearbook, chicago, il. jensen, michael c. 1968. “the performance of mutual funds in the period 1945-1965,” journal of finance, 23:389416. mcqueen, grant, and steven thorley. 1991. “are stock returns predictable? a test using markov chains,” journal of finance, 461239-263. reichenstein, william, and steven p. rich. 1992. “predicting long-horizon stock returns: the ex-ante advantage,” baylor university. samuelson, p. a. 1990. “asset allocation could be dangerous to your health,” journal of portfolio management, 16(3):5-8. sharpe, william f. 1966. “mutual fund performance,” journal of business, 39: 119-138. treynor, jack l., and kay k. mazuy. 1966. “can mutual funds outguess the market?’ harvard business review, 44(4):131-136. pii: s1057-0810(96)90027-x financial services review, 5(l): 57-70 copyright 0 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. ratios and benchmarks for measuring the financial well-being of families and individuals sue a. greninger vickie l. hampton kamol a. kitt joseph a. achacoso financial planners and educators comprised a panel of 156 experts in this delphi study designed to identify and refine ratios and benchmarks for measuring financial well being. consensus between the two groups existed for benchmarks on 20 of 22 ratios in the areas of liquidity, savings, asset allocation, inflation prorection, tax burden, hous ing expenses, and insolvency/credit. consensus regarding the usefulness of specific ratios was observed for liquidity and tax burden but not for inflation protection and insolvency/credit. the preferred ratios were generally less complex and more easily measured than many of the ratios used in previous work. from the findings, a projile of jkmcial well-being for the typical family/individual was proposed. financial planning, as a profession, has evolved and matured over the past two decades. while many of the original certified financial planner (cfp) practitioners migrated from related disciplines such as accounting, insurance, and finance, a growing proportion of young people are seeking to pursue financial planning as their initial career choice (wechsler & longstaff, 1996). colleges and universities have responded by developing programs to meet the growing demand for planners in this emerging field. there are now more than 70 colleges and universities in the united states offering programs in financial planning (wechsler & longstaff, 1996). due to this rapid growth, standardization of fman cial measures and terminology in the field has been limited. in a recent discussion of the need for practice standards in financial planning, kochis (1996) contrasted the extensive development of such standards in medicine to their dearth in financial planning. he con cluded by saying, “in direct contrast, there are no defined standards of practice that are gen erally accepted by today’s financial planning profession” (p. 20). sue a. greninger, vickie l. hampton, kamd a. kit& and joseph a. achacoso l human ecology department, the university of texas, austin, tx 78712; e-mail: sgreninger@mail.utexas.edu, v.hampton@mail.utexas.edu, kkitt@mail.utexas.edu, and jocoso@utxvms.cc.utexas.edu. 58 financial services review 5( 1) 1996 the fundamental goal of this project was to identify and refine important, useful ratios and benchmarks for assessing the financial well-being of families and individuals. input via a delphi format was sought from both financial planning practitioners and academi cians to determine if a consensus regarding a comprehensive set of ratios and benchmarks has developed. these measures were viewed as potential diagnostic tools which could be used in monitoring financial progress and identifying problem areas. ii. review of the literature financial ratios have a long history in business as instruments for assessing the financial health of firms (horrigan, 1978). an early application of ratios centered on analyzing busi ness insolvency and bankruptcy (altman, 1968) while later work focused on the develop ment and refinement of a more comprehensive set of financial ratios and on evaluation of their effectiveness (chen & shimerada, 1981; brandt, danos, & brasseaux, 1989; lawder, 1989; kimmell, 1994; poston, harmon, & gramlich, 1994; gardiner, 1995; kane, 1995). one of the business applications of financial ratios that first interfaced with consumers involved credit approval standards including mortgage loans. in addition, debt and expen diture ratios have been included in credit scoring models employed by lenders to manage risk exposure a&to optimize profitability (devaney & lytton, 1995). the use of ratios as tools to assess the financial well-being of families and individuals has developed only over the past decade, evolving from early descriptive efforts to empir ical studies involving a variety of data sources. this analysis has its roots in griffith’s (1985) descriptive work where he reviewed personal finance books and found little speci ficity regarding ratios, norms, or other recommended measures for performing financial analysis. mason and griffith (1988) presented 20 ratios and then applied them to a hypo thetical case study. they noted that problems occur when trying to analyze personal finan cial statements because of lack of standardization in compiling such statements. lytton et al. (1991) also used ‘a hypothetical case study to illustrate the use of nine financial ratios. they concluded that financial ratios were broadly applicable and interpretable by financial professionals as well as by individuals and families. langrehr and langrehr (1989) studied 14 consumer debt ratios in search of a preferred measure for assessing ability to repay debt. they concluded that debt service to income was the best measure of ability to handle debt; however, these researchers called for continued development of reliable numeric guide lines. through empirical studies, researchers have attempted to address the usefulness of these and other financial ratios. prather (1990) attempted to establish norms for grifftth’s 16 ratios using data from the 1983 survey of consumer finances. she noted problems in structuring and interpreting several ratios and called for more work directed toward devel oping recommendations and standardization of ratios. a number of research studies have employed ratios to assess financial status or well being in one or more specific areas. these areas include the adequacy of emergency funds (iwuagwu, 1989; chang & huston, 1995; devaney, 1995; hanna & wang, 1995; hong & swanson, 1995), overall savings or overspending rates (burns & widdows, 1990; bosworth, burtless, & sabelhaus, 1991; bae, hanna, & lindamood, 1993), changes in net worth over time (hefferan, 1982; ,chang, 1994; fitzsimmons & leach, 1994), housing expense and affordability (fronczek & savage, 1991; bogdon, silver, & turner, 1993; oh, ratios and benchmarka 59 1995; paulin, 1995). household asset ownership and portfolio allocation (weagley & gannon, 1991; kennickell c shack-marquez, 1992; lee & hanna, 1995; kiao, 1995), and debt levels and insolvency risk (luckett, 1988; sullivan & fisher, 1988; sullivan, warren, & westbrook, 1989; scannell, 1990, livingstone & lunt, 1992; devaney, 1993, 1994; devaney & lytton, 1995; godwin, 1995; hong & swanson, 1995; yieh & w&lows, 1995). the literature indicates that a comprehensive set of ratios and benchmarks could be beneficial to both families and professionals when making current financial decisions and planning for future needs. such measures should be standardized, broadly accepted, and easily implemented. mason and griffith (1988) concluded that “input is needed from aca demicians and practitioners before competent ratio analysis can become standardized and widely used” (p. 83). lytton et al. (1991) noted that “establishment of guidelines for all ratios would be dependent upon (a) a consensus among professionals as to the most useful ratios; (b) broad application of the ratios by professionals; and (c) empirical research to determine appropriate numerical ranges” (p. 21). the present delphi study was initiated in an effort to address some of the concerns raised in the literature regarding standardization, consensus, simphcity, and broad-based application of financial ratios. m. objectives and methodology the three primary objectives of this research were to: l provide a forum for financial planners and educators to identify and refine useful ratios and benchmarks for measuring financial well-being; l determine the extent to which a consensus existed regarding these measures among the two professional groups; and l develop general guidelines for a financially “healthy” family/individual. while it is imperative to note that benchmarks must always be considered in a broad con text including factors such as life cycle, family type, financial status, economic environ ment, and personal objectives and goals, this research is a necessary first step toward the development of norms that allow for diversity among family types and economic condi tions. this study employed the delphi research method which was developed by the rand corporation in the 1950s as a spinoff of defense research. linstone and turoff (1975) described the delphi technique as “a method for structuring a group communication pro cess so that the process is effective in allowing a group of individuals, as a whole, to deal with a complex problem” (p. 3). these authors described a conventional delphi method as follows: a small monitor team designs a questionnaire which is sent to a larger respondent group. after the questionnaire is returned, the monitor team summan ‘zes the results and, based upon the results, develops a new questionnaire for the responding group. the respondent group is usually given at least one opportunity to reevaluate its original answers based upon examination of the group response @. 5). 60 financial services review 5(l) 1996 the participants in this research project included the research team, an advisory com mittee, and a panel of experts. initially the three-person research team compiled a list of existing financial ratios and benchmarks based on previous research findings and current personal finance textbooks. the advisory committee, composed of five financial planners and one additional educator, reviewed the list and suggested research design and sample selection methods. in spring 1994, 400 financial planners, randomly selected from cpp licensees, and 340 financial educators, selected from membership in the association for financial counseling and planning education, were invited to participate in this project. professionals agreeing to participate in the study included 122 financial planners and 159 educators-a combined acceptance rate of 38%. in round 1 of the study, respondents received an open-ended questionnaire that addressed various ideas identified in the literature as important aspects of financial well being. for each of these concepts, respondents were asked to list specific factors that should be considered when assessing the financial well-being of families and individuals. from round 1 responses, the research team designed two different round 2 question naires in an effort to keep questionnaires a reasonable length. the first round 2 question naire included (a) goal establishment, (b) financial reviews, (c) savings programs, (d) emergency funds, (e) liquidity ratios, (f) investment and diversification, (g) inflation protection, and (h) retirement. the response rate for this questionnaire was 63% of the original 281 experts. the second questionnaire in round 2 which yielded a slightly lower response rate (56% of the 281 experts) covered (a) housing, (b) credit, (c) insurance, (d) taxes, and (e) estate planning. results from both round 2 questionnaires were reported in the round 3 questionnaire, and the experts were asked to supply benchmark values for the most useful ratios that had emerged. this provided respondents an opportunity to reevaluate their original answers based upon examination of the group response as is rec ommended for a delphi project. the completion rate for this final round was 60% of the original 28 1 experts. round 3 respondents were asked to identify their primary type of work as practitioner, edu cator, or counselor. for the purposes of this paper, 85 practitioners and 7 1 educators were included for a total sample size of 156 financial planning experts. fourteen counselors were excluded from this analysis because the size of the group was too small for statistical testing. of the 156 who participated in round 3, the majority (60%) were male. there was a significant difference (p i .ool) in the gender of the planner and educator subsamples with the planners being primarily male (77%) and the educators being primarily female (60%). the average ages for the planner and educator subsamples were 47 and 49 years, respectively. as might be expected, there was a statistically significant @ i .ool) differ ence between planners and educators in level of education. the majority of planners (55%) had a bachelor’s degree or less while the majority of educators (58%) had a doctorate or juris doctorate degree. of the 85 planners, 80 (94%) were cfp licensees whereas only 3 1 of the 71 educators (44%) were. most planners (67%) received at least part of their income from commissions while very few educators (3%) received commissions. ratios and benchmarks 61 using the delphi approach, ratios and benchmarks resulted for seven general areas of financial planning including liquidity, savings, asset allocation, inflation protection, tax burden, housing expenses, and insolvency/credit. the following definitions also emerged: l liquid assets = cash and cash equivalents, checking accounts, savings accounts, money market accounts, money market mutual funds, and cds with maturities of i 6 months. l investment assets = all other assets held for investment purposes, not including use assets or equity in a home. l monthly expenses = average fixed and variable living expenses including debt/ credit repayment, taxes, and monthly allocations being set aside for irregular expenses such as auto insurance, vacations, gifts, etc. l current debt = all debt/credit obligations, charges, bills and payments due within 1 year. l payroll taxes = federal, state, and local income taxes and social security taxes. l property taxes = real estate and personal property taxes. l renter’s expenses = rent, renter’s insurance, and utilities. l homeowner’s expenses = principal, interest, taxes, insurance, homeowner’s association fees, utilities, maintenance, and repairs. the numerical ratios and values recommended by the panel of financial planners and educators for a typical family/individual are presented in table 1. in addition to means for the total sample, a comparison of means between the planners and educators is included. medians are also included since they are less influenced than means by outlying values. in general, results from the t-tests comparing the responses of planners and educators indi cated the existence of a strong consensus between the two professional groups. significant differences existed between planners and educators on only 2 of the 22 ratios, foreign investments + total investments and renter’s expenses + before-tax income. a. liquidity the two liquidity ratios which emerged as useful from this delphi study were liquid assets + monthly expenses and liquid assets + current debt. the median value for the ratio involving expenses was 300% suggesting a 3: 1 ratio of liquid assets to monthly expenses. the mean values, which ranged from 246% for the educators to 270% for the planners, were somewhat lower than the overall median. however, these findings support a general view that liquid assets should provide a minimum of 255 to 3 months of living expenses. on the second liquidity measure, a median value of 50% was recommended, with the mean being 88%. although the planners specified a higher mean than the educators, 94% vs. 79%, respectively, the difference was not significant indicating a consensus between the two groups. it is interesting to note that the panel of experts overwhelmingly preferred the liquidity ratio using monthly expenditures to the one using current debt. although respon dents were not asked to give reasons for their choices, at least two ideas come to mind. first, individuals tend to think of debt repayment as a part of monthly expenses and realize 62 financial services review 5( 1) 1996 table 1 recommended ratios for a typical family/individual total sample planners educators (n = 156) (n = 85) (n = 71) ratio median mean 2 sd mean mean liquidity liquid assets monthly expenses 300% 261 f 202% 270% 246% liquid assets torrent debt 50% 88 i 124% 94% 79% sovillgs savings gross income 10% 12i4% 13% 12% savings nete 10% 12*5% 13% 12% asset allocation liquid assets net worth net investment assets net worth 14% 17*11% 15% 20% 50% 56 zt 22% 57% 53% foreign investments total investments 10% 13 *7% 15% lo%** inflation protection % a in net worth rate of inflation % a in investment assets rate of inflation equity investments total investments 2 2 60% 4* 11 5 4* 10 4 59 i 18% 62% 3 4 54% tax burden $31,oim gross income payroll taxes gross 20% 19 *7% 20% 18% payroll + property taxes gross income $loo,ooo gross income payroll taxes gross income 25% 24*9% 24% 23% 30% 29i91 29% 29% payroll + property taxes gross income 35% 34 * 10% 33% 34% (continued) ratios and benchmarka 63 table 1 (continued) ratio total sample (n = 156) median mean f sd planners (n = 85) mean educators (n = 71) mean housing expenses renter’s expenses gross income homeowner’s expenses gross income 30% 29~8% 31% 28%* insolvencykkedit reasonable nomnortgage debt payments after-tax income 10% 14*9% 14% 15% total debt payments after-tax income total expenses after-tax income danger-point nonmortgagc debt payments after-tax income 35% 33 f 12% 34% 32% 80% 71 f 21% 72% 69% 20% 26 * 13% 27% 24% total debt payments after-tax income total expenses after-tax income 45% 46 * 14% 90% 85 f 20% 88% 82% notes: *p 5 .os. +*p s .ool it is necessary to pay more than just debt payments in the case of a financial setback. sec ond, monthly expenses is a simpler concept to conceptualize than current debt. b. savings when respondents were asked to specify the percentages of income that should be saved, identical values were recommended for both gross and net income by planners and educators alike. these results are problematic since it would be impossible for the same level of savings to comprise similar percentages of both beforeand after-tax income lev els. since the median recommended savings ratio for both incomes was 10% the research team felt that these responses might have been influenced by widely-held savings norms. consensus was again noted between the two subgroups with the planners averaging only one percentage point higher than the educators for both savings ratios, 13% and 12% respectively. this similarity was somewhat surprising since significantly more planners than educators (63% vs. 47%, p 5 .05) said that nonvoluntary or forced contributions should be included in the definition of savings. regarding the definition of income, gross income was preferred by a majority (58%) over net income, which was the choice of 64 financial services review 5( 1) 1996 approximately one-third of the respondents. thus, although there was a general consensus regarding recommended savings benchmarks, the planners appeared to embrace a broader definition of savings than the educators. a weakness of this study was that a definition of savings was not carefully specified in the final round 3 questionnaire due to a lack of con sensus in the earlier round. c. asset allocation in this study, the two ratios deemed most useful in analyzing assets were liquid assefs + net worth and investment assets + net worth. a difference between planners and educa tors that approached the .05 significance level was found regarding the recommended level of liquid assets, with educators’ mean value being 20% of net worth compared to the plan ners’ value of 15%. the median response for the total sample was 14% of net worth. finan cial experts agreed that net investment assets, not including equity in a home, should comprise slightly over one-half of net worth. recommended means of the planners and educators were quite similar on this ratio, being 57% and 53%, respectively. a third asset allocation ratio included in this study produced one of the few significant differences between planners and educators, with planners recommending a significantly higher per centage of net worth in foreign investments than was true of the educators. this finding may be reflective of differences in orientation and experience between the two groups. d. inilation protection two of the inflation protection ratios compared the annual rate of inflation to the annual percentage change in financial well-being as measured by (a) net worth and (b) investment assets. since these ratios compare annual percentage changes in both the numerator and the denominator, they are expressed in numbers rather than in percentages. for both ratios, the median responses were 2, indicating that net worth and investment assets should each increase twice as fast as the rate of inflation. for example, if annual inflation increased 3%, then net worth as well as investment assets would need to grow at 6% for that year. mean responses for these ratios were generally higher than the medians, ranging from 3% to 5%. a third ratio commonly utilized to gauge inflation protection, equity investments + total investments, could have been included among the asset allocation measures; however, it was incorporated here because investors often view equity investments as inflationary hedges. for the total sample, the median and mean were quite similar, 60% and 59%, respectively. as with the other inflation protection measures, planners and educators did not differ significantly, although planners recommended higher levels than educators, 62% and 54%, respectively. when respondents were asked which of the inflation protection measures they considered most useful, there was a significant difference @ i .ool) in the responses of the two subgroups. most of the educators (69%) preferred the measure involv ing net worth, whereas only 31% of the planners preferred this ratio. planners’ responses were split, with 43% preferring the investment assets measure and 26% preferring the equity investments measure. only 21% of the educators preferred the investment assets ratio, while 10% preferred the equity investment ratio. ratios and benchmarks 65 e. tax burden respondents were asked to specify reasonable tax burden values relative to gross income for (a) payroll taxes and (b) payroll plus property taxes. because of the progressive nature of the income tax structure in the united states, values were requested for two dif ferent gross income levels, $31,000 and $100,000. for the $31,000 income level, the median and mean for the total sample on the payroll tax measure were 20% and 19%, respectively. little difference existed in the responses of the planners and educators. con cerning the more complex tax measure that included real estate and personal property taxes, median and mean responses were again quite similar for the total sample, 25% and 24%, respectively, as well as for planners and educators who averaged within one percent age point of one another. at the higher gross income level of $100,000, the median and means were quite simi lar for both tax measures. means for the total sample and two subgroups were all 29% on the payroll tax ratio. when property taxes were included with payroll taxes at the higher income level, the median was 35%. the planners’ mean was 33%, and the educators’ mean was 34%. regarding the relative usefulness of the two tax burden ratios, the majority (54%) of planners chose the simpler payroll tax measure while the majority (57%) of edu cators preferred the payroll plus property tax measure. this difference, however, was not statistically significant. f. housing expenses the panel of experts were asked to specify percentages of gross monthly income that were reasonable for housing costs for both renters and homeowners. in the case of rental costs, the median response for the total sample was 301, and the mean was 29%. as pre viously discussed, planners averaged significantly @ i .05) higher than educators on this measure, 31% vs. 28%. respectively. for homeowners, the median and mean for the total panel were similar, 35% and 34%. respectively. there was no significant difference between planners and educators with regard to reasonable home ownership expenses. g. insolvency and credit respondents were asked to specify both reasonable and danger-point values relative to after-tax income for the following three measures: (a) nonmortgage debt payments, (b) total debt payments, and (c) total expenses. for the first ratio, involving only nonmortgage debt, the median value regarded as reasonable by the total sample was 10%. the means ranged from 14% to 15%, a little higher than the median. concerning the danger-point for this ratio, the experts’ median value was 20% while the averages recommended for this credit benchmark was 24% for educators and 27% for planners. reasonable values for the second ratio were higher than those for the first due to the inclusion of mortgage debt. the median and mean for this ratio were similar, 35% and 33%, respectively, as was true for the means of the planner and educator subgroups, 34% and 32%, respectively. danger-point values were also similar with the median and mean for the total sample being 45% and 46%, respectively. planners’ responses regarding the danger level, however, averaged six percentage points higher than the educators, 49% vs. 4346, respectively. this difference just missed the .05 significance level. 66 financial services review 5( 1) 1996 on the third ratio where total expenses relative to after-tax income were considered, the median response for a reasonable level was 80%. means for the total sample and two subgroups were all quite similar, ranging from 69% to 72%. the median response for the danger-point on this measure was 90%, with the mean being somewhat lower at 85%. plan ners again had a higher mean (88%) than educators (82%), but this was not statistically sig nificant. a significant difference @ i .05) did result when the respondents were asked which of the three ratios they felt were more useful in assessing exposure to insolvency and credit problems. one-half of the planners chose total expenses + ufier tan income as the most useful, whereas only 28% of the educators chose this ratio. the second ratio involving total debt payments relative to after-tax income was preferred by a slightly larger percent age of educators (43%) than planners (37%). twice as many educators as planners, 28% vs. 13%. respectively, preferred the ratio involving nonmortgage debt payments, or what is typically called the debt safety ratio. overall, 40% of the total sample preferred the more general ratio of expenditures to income, 40% preferred the broader debt to income ratio, and only 20% preferred the narrower debt safety ratio that is so widely used by lenders. v.summaryandconclusions there was considerable agreement among experts about the specific ratios and bench marks derived in this study. when the responses of planners and educators were compared on the major benchmarks presented in this study, the only statistically significant differ ences were for a relatively specific asset allocation ratio,foreign inveshnenrs i roral invesr menrs, and for renter’s expenses + before-rax income. however, in all other areas, liquidity, savings, inflation protection, tax burden, homeowner’s expenses, and insolvency and credit, the results of this study demonstrated that there was a consensus regarding the benchmark values. concerning the usefulness of specific ratios, there was general agreement between planners and educators on two of the four types of ratios where this information was requested. in the area of liquidity, experts agreed that liquid assets + monthly expenses was a more useful measure than liquid assets + current debt. in addition, the two tax burden ratios were judged similarly with regard to usefulness by both planners and educators. however, there were significant differences between planners and educators in the per ceived usefulness of the different ratios for inflation protection and insolvency and credit. in both, the planners preferred ratios that were conceptually simpler and operationally eas ier to discern. overall, the ratios chosen by the total panel of experts in this study were less complex and more easily measured than many of the ratios discussed in the previous liter ature. one exception was in the insolvency/credit ratios where after-tax income was uti lized. since gross income is more easily ascertained for planning purposes, the research team recommends the use of gross income for all ratios where income is a component. based on the results of this delphi study, a profile of financial well-being or “health” for the typical family or individual is proposed in table 2. there will no doubt be contro versy over whether this presentation oversimplifies the findings of this research or simply summarizes the findings more clearly, which is the intent of the investigators. as noted by devaney and lytton (1995) in their recent review of literature regarding insolvency, “the primary function of ratios should be to act as indicators or redflags . ..” (p. 148). the sug ratios and benchmarks 67 gested benchmarks may be viewed as “red flags” or indicators by which professionals, as well as families, might monitor financial health and well-being. finding values that lie out side the recommended parameters does not necessarily indicate a lack of financial health but instead points to an area worthy of further introspective thought and analysis. some of the benchmarks suggest minimum or maximum thresholds which may signal a greater like lihood for potential financial difficulties, such as the ratios for liquidity, savings, and insol vency and credit. others provide general reasonability guidelines in the areas of asset allocation, inflation protection, and major expenses for housing and taxes. comparisons with earlier work are difficult because of inconsistencies in the ratios’ components and definitions. however, similarities with previously-recommended bench marks were found in this study for liquid assets + monthly expenses and nonmortgage debt payments + after-tax income. conversely, in the case of the broader debt ratio, total debt area table 2 financial well-being profile ratio recommendation liquidity savings asset allocation inflation protection tax burden housing expenses insolvency/ctedit liquid assets monthly expenses savings gross income liquid assets net worth net investment assets net worth foreign investments total investments % a in net worth rate of inflation % a in investment assets rate of inflation payroll taxes gross income s 20% ($3 1,000 income) 5 30% ($100,000 income) payroll + property taxes gross income 5 25% ($31,000 income) .s 35% ($100,000 income) renter’s expenses gross income s 30% homeowner’s expenses gross income 5 35% nonmottgage debt payments 5 15% reasonable after-tax income 2 20% danger-point total debt payments after-tax income 2 250% (2 2% times monthly expenses) 2 10% 2 15% 2 50% 2 10% 22 22 s 35% reasonable 2 45% danger-point 68 financial services review 5( 1) 1996 payments + after-tax income, the benchmark reported in this study was higher than the thresholds formerly recommended and utilized. looking at the two general asset allocation ratios, the experts in this study prescribed a lower level of liquid assets and a higher level of investment assets relative to net worth than previously-published guidelines. in addition, the recommendations for inflation protection and reasonable housing expenses were also higher than those suggested in prior work. it is important to note that this research dealt with benchmarks that are generally appropriate for families and individuals, not considering differences in life-cycle stages, income levels (except the tax burden measure), risk tolerances, and economic conditions. however, financial experts are all too aware that a “typical” family/individual does not really exist. future research needs to move toward the development of norms that allow for diversity in family types, age or life-cycle stages, value orientations, and economic condi tions. the framework of ratios and benchmarks provided by this study can be used as a foundation for developing tolerances or ranges appropriate for more specific situations. acknowledgment: preparation of this article was supported in part by a grant from the certified financial planner board of standards. references altman, e.i. (1968). financial ratios, discriminant analysis, and the prediction of corporate bank ruptcy. the journal of finance, 23(4), 589-609. bae, m., hamra, s., & lindamood, s. (1993). patterns of overspending in u.s. households. finan cial counseling and planning, 4, 1 l-30. bogdon, a., silver, j., & turner, m.a. (1993). national analysis of housing affordability, adequacy, and availability: a framework for local housing strategies. washington, dc: urban institute. bosworth, b., burtless, g., & sabelhaus, j. (1991). the decline in saving: evidence from household surveys. in william c. brainard & george l. perry (eds.), brookings papers on economic activity (pp. 183-241). washington, dc: brookings institution. brand& l., danos, j.r., & brasseaux, j.h. (1989). financial statement analysis: benefits and pit falls. the practical accountant, 22(5), 68-70. burns, s.a., & widdows, r. (1990). sensitivity of a retirement analysis framework to changes in retirement analysis parameters. financial counseling and planning, i, 7 l-9 1. chang, y.r. (1994). saving behavior of u.s. households in the 1980s: results from the 1983 and 1986 survey of consumer finances. financial counseling and planning, 5.4564. chang, y.r., & huston, s.j. (1995). patterns of adequate household emergency fund holdings: a comparison of households in 1983 and 1986. financial counseling and planning, 6, 119-128. chen, k.h., & shimerada, t.a. (1981). an empirical analysis of useful financial ratios. financial management, lo(spring), 51-60. devaney, s.a. (1993). change in household financial ratios between 1983 and 1986: were american households improving their financial status? financial counseling and planning, 4,31-46. devaney, s.a. (1994). the usefulness of financial ratios as predictors of household insolvency: two perspectives. financial counseling and planning, 5, s-24. devaney, s.a. (1995). emergency fund adequacy among u.s. households in 1977 and 1989. con sumer interests annual, 41,222-223. devaney, s.a., & lytton, r.h. (1995). household insolvency: a review of household debt repay ment, delinquency, and bankruptcy. financial services review, 4(2), 137-156. ratios and benchwuuks 69 fitzsimmons, v.s., & leach, l.j. (1994). net worth change: beginning and expanding life cycle stages. financial counseling and planning, $6582. fronczek, p.j., & savage, h.a. (1991). who can afford to buy a house? current housing reports, h121/9101. washington, dc: u.s. department of commerce, economics and statistics administration, and u.s. bureau of the census. gardiner, m.a. (1995). financial ratio definitions reviewed. management accounting-london, 73(8), 32. godwin, d.d. (1995). are all debts created equally? consumer interests annual, 41, 141-147. griffith, r. (1985). personal financial statement analysis: a modest beginning. in v.l. langrehr (ed.). proceedings of the 3rd annual conference of the association of financial counseling and planning education, 3,123-131. hamra, s., & wang, h. (1995). the adequacy of emergency funds to cover household expenditures. consumer interests annual, 41,224-225. hefferan, c. (1982). determinants and patterns of family savings. home economics research jour nal, 11(l), 47-55. hong, g., & swanson, p.m. (1995). comparison of financial well-being of older women: 1977 and 1989. financial counseling and planning, 6, 129-138. horrigan, j.o. (ed.). (1978). financial ratio analysis: an historical perspective. new york: amo press. iwuagwu, u.s. (1989). financial ratios and perceived household financial security. unpublished m.a. thesis, university of wisconsin, madison. kane, g. (1995). using ratio analysis to hedge investment risks. trusts & estates, 134(6), 4245. kennickell, a., & shack-marquez, j. (1992). changes in family finances from 1983 to 1989: evi dence from the survey of consumer finances. federal reserve bulletin, 78, 1-18. kimmell, d.l. (1994). using ratios and graphics in financial reporting. the internal auditor, 51(5), 17. kochis, s.t. (1996). practice standards: building consumer confidence. in robert n. veres (ed.), the next agenda: a blueprint for the emerging financial planning profession {pp. 20-22). london: millennium press ltd. langrehr, v., & langrehr, f.w. (1989). measuring the ability to repay: the residual income ratio. the journal of consumer affairs, 23(2), 393406. lawder, k.e. (1989). ratios 101: back to the basics of financial analysis. business credit, 9i(6), 28-30. lee, h.k., & hamta, s. (1995). investment portfolios and human wealth. financial counseling and planning, 6, 147-152. l&tone, h.a., & turoff, m. (eds.). (1975). the delphi method: techniques and applications. bos ton: addison-wesley publishing co. livingstone, s.m., & lunt, p.k. (1992). predicting personal debt and debt repayment: psychologi cal, social and economic determinants. journal of economic psychology, 13,ll l134. luckett, c.a. (1988). personal bankruptcies. federal reserve bulletin, 74(g), 591603. lytton, r.h., garman, e.t., & porter, n.m. (1991). how to use financial ratios when advising cli ents. financial counseling and planning, 2,3-23. mason, j.w., & griffith, r. (1988, spring). new ratios for analyzing and interpreting personal financial statements. journal of the institute of certified financial planners, 9(l), 71-87. oh, d. (1995). households with rent burdens: impact on other spending and factors related to the probability of having a rent burden. financial counseling and planning, 6, 139-146. paulin, g. (1995). a comparison of consumer expenditures by housing tenure. the journal of con stutteraffairs, 29(l), 164198. poston, k.m., harmon, w.k., & gramlich, j.d. (1994). a test of financial ratios as predictors of turnaround versus failure among financially distressed firms. journal of applied business research, 10( 1). 41-56. 70 financial services review 5( 1) 1996 prather, c.g. (1990). the ratio analysis technique applied to personal financial statements: develop ment of household norms. financial counseling and planning, 1,53-69. scamrell, e. (1990). dairy farm families’ financial management. financial counseling and plan ning, 1, 133-146. sullivan, a.c., & fisher, m.f. (1988). consumer credit delinquency risk: characteristics of con sumers who fall behind. journal of retail banking, io(3), 53-64. sullivan, t.a., warren, e., & westbrook, j.l. (1989). as we forgive our debtors: bankruptcy and consumer credit in america. new york: oxford university press. weagley, r.o., & gannon, c.f. (1991). investor portfolio allocation. financial counseling and planning, 2, 131-154. wechsler, m.j., & longstaff, j.s. (1996). the agents of change. in robert n. veres (ed.), the next agenda: a blueprint for the emerging financial planning profession (pp. 3-6). london: mil lennium press, ltd. xiao, j.j. (1995). patterns of household financial asset ownership. financial counseling and plan ning, 6,99-106. yieh, k., & widdows, r. (1995). households showing financial characteristics of potential bank rupts. consumer interests annual, 41, 155-160. pii: s1057-0810(02)00102-6 ���� ��� � ��� ��� ��� ����� � ������ �� � ��� ��������� �� ���� �� �� ����� ��� ���� �� � �� � ���� ��� �� ���� �� � �� � �� �� ��� ���� ����� ����� �� �� ����� ���� � � ���!� ��� ���� ������ ���� ���� ��� ���� �� �" �� ���� # �� ��# ������ ��� ������� � � �� � ��� ��� ��� � � ���� �� ���� ���� ��� ��� �� � ������� � ��� $������� � ����� � %% ��� ���� ��� ���� � � # �� ��� ������� ���� ��� ���� ��� � ���� �� ���� � �� �� &�� � '�� ���� � (��! )� ���� �� �� *���� '��� � +����� )� ���� ��� *������ ��� ����� ��$�� ����"� � ��� ��� $������ � ���� ��� � ����� � � ��� $������� ����� �" # �� ������ �� ������ �� ��� ���� ������� ��� � � &���� & �������� �� ,���" � ����� )� ���� ��� &���� ��� ���� � ���� �-���� �� �� �������� ����� ��� � ��� � ��� $������ �� ��� ���� . ����� �� #��� �� ���� �" �� � ������ �� ��� /��� � 0 ������� � ��� ��� ��� � � ���� �� ���� ���� , ��� ��� � ��� �� ����"�� � ��� � ���� ������� ��� �� ���� ��� ��� ��� ��� ��� 11�� ����� "�� �� � ��� ���� ������ � 2���� �� � � �� � ��� �� #��3���� ��������44 �� &���� �� & �������� �� '������ �� 5�� � ��� �6����� ��� ������� � � � �� �� � � ��� � "�� ���� �� ���� �" �� #��� #��3���� ������ ��� � � ��� ��� � �� �� 2���� �� � � ��� ��� 2� ���� ���� �� ��� ��������� �� �� �� ����� �" ��� �� � ��!� � ������ � ��� ��� ��� �� � � ��� �� 2���� �� � � �� ��� #��3���� �������� �� ��6� � 6 ��� ���� �6�� �� ��� ��� ����� ������� �" ��� ������ ����� �� ������� 7�� ���!� �� � � ���� ���� ���� � ����� �� ��� ������ ���#��� �2�� ����2�� �� �2�� ����� ��� �� ������ � �� � �� ����� � �" �� � ��� �� ����� ���� �� �2�� ����� ��� �� ������ �� � ����� �� ����"�� � ��� #���� � ����� ���� �" �� � ��� ������� �� 2 �� �� � 6�� ��� ���� �6�� �� ��6 ������� ��� ������ ��� ���� �� ��� ��� �� � ������ � ��� � ������ ��� �� ��� 2��� � ����� ��� ����� 11 �� ����� � ��� ������ ����� �������8 ��� ����� � � � ���� ��� �������44 �� 5 �� �� 5� 9��� �"� �� 5 �� �� '� ������� � �6�� �� ��� ����� � � � ���� ��� ������ ������ ��� ��� ���� ����� �" ��� � ���3��6 ������� ����� � � � ���� ��� ������ ����� �� � ��� ��� ����� ����� � ��� � �� ��4� ����� ������� ��� 11� ������ ��� � ����� �� ���� �� �2�� ����2� �� �2�� ����� ��� �� ���� �� �����8 � +���� &���� � ����� ���44 �� :�� 9������� ���#� ������� �� 9��� ;���� ���!� �� ��� ������ � ��������� � �! ���#��� ��� �#� ����� � ����� �� �" � +���� &���� � ����� ��� �� ������� 2� ���� � "��� ����! ������� ��� ���� � �� � "��� ���� �" ������ �� #� �� ����#� � �2�� ����� ��� �� ���� �� � �� ���������� ����"� #����� �� ����� ��� ���� �� ����2�� � � �2�� ����2� ����� ��� ���� ���� � �� ����� ������� �� ������� � ����� � �" ��� � �� �� � 6 ��������� �� ����� ��� �#� ������ +� :�� � ;����� �� ���� *�#�����4� ��� ��� ��� ��� 11 �� �� � �� �� � ��8 ������ � ��� �" �������� � ����� ����� ����44 �6�� ��� �2�� ����� ��� �� ����� �� �������� � ���� �� ���� ��� '�� �# �� �%���� �<= ��>?3�@��a��ab < ��� ���� ������ � %��� c��� ��� �� ����� �� �� ���� ���� c��8 � � � > ? 3 � @ � � � � % � � � � � % 3 . ���� � ��� �� �� ����� ��� �� ����� ��� ����" ����� ������ �� ������� 2� ���� ����� � � ���� ��� � �� ����� ���� � ���#��� ��� ��������"� � �������� ����� ��� �� �� ��� ���� �� ��� ������ �� �������� ����� ����� �� ��� ����� 11*��" �� ��� ����"�� � ��� #���� �� �������� ����� �� ���"���� � ���������3 �� �� �� ��� 2��� ����� #���� � ��� ;'��44 �� :���� � ���� *���� �� 0�����" : � �����d�� ��� ���3���� �" ��� #���� � � ������ � � � ���� #�� ��� >� ����� �� �� �� ���� � .3����3��� � � � ��� 0���"���� � ��������� �� �� ���� ��� ��� �� �� � "� 2���� ���� � �� � �6��� � �" ����"�� ���� � "�� ��� � ���� ��� ����� ��� " ��� ������ �� �"�� �� ��6� �#� ��� ���� �6�� �� ��� ������ ��� ����� �� � ������ :� ����� �� ���� *�#�����4� ��� ��� ��� ��� 11������ � � ��������� ��� ��� �� ��� ������8 � ��# �������� �� �������� ���44 �������� � ��� ��� �'��� '��� �'��� ���3 � ��� ��� �'��� �� ���� ��6���� ���������� �� �� �" �e��� � � �� � ���3��6 ���������� �� ����� �� ������� ������� ���� #��� ������ �" � � ��� ��� �'� �� � '��� �'�� � '��� �'� ���� �� ���� ��� ��� �� -�6 � � �� �� ������� # �� ���" ��������� ��� d���� 11� ���6�� ��� �� � ��63 � ��� ��� �'��� '��� �'��� �� =���!� �����������44 �� ������� +� ;���� �� 9� ��� ;� c������� ���!� �� ����� ��6 ��� ���� ��������� ��� ���� �� ��� ������ ����� �" �� ��# � � ���� �� 2���� �� �������� ��!� ������� ��� ����� ��# ����� ��6 ��� �"� ��� �� �� ������ � ��� 2� ���� ��������3�������� =���!� ����� ���� ����� ���� �� ��� � �� ��������4� ����� ��� ��� �� ���� '��� �'��� 9��� 9� �� �d�� �� ��� ��� � �"�4� ��� ��� � ��� 11 �� ������ � ���! � ������ ����� ������� ��8 ���� �� ���� ��� 1f�# (��!4� &����"� ��� �"� c��"���4 44 �6�� ��� #�� ���� 11��� ��� ���� 44 ����� ������� �� ���"���� ��� � ������ � ���� �� ��� �"� ���"���� ��� �� ���3��� ���� *��! �" �� f�# (��!4� &����"� ��� �"� c��"���� ��� ������� ������ ���� ��� ���"�� ��� �� � ���� � ����� � ��� ����� ��� "������ ��� ����2� �� ���� ��� � "��� ��� ��6 ����!�� ��� ���� � ���� � �� ��� "������ ��� ����2�� 110��� ���� ����� �� �6��� � �����3���� �����" �"g44 �� :���� /� *�"" � �� 0���� * �� ���! �� �������4� ��hh>� ������ ���� ����� � � ����� �� ��� ����� �� ��� ���� ������ � �����3���� �����" �"� �� ������� ���� ������ �� �������� �� �� � �� ���� �� #� �� ��� ��� ��e� �� �������� �� �� ��� ���� � �! � ����� �� ���� �������4� ������� �� ������� 2� ���� ���� ����� �� �� �� ��� ��� �6��� � ��� �6 ������ � ��� �����3���� �����" �" ��������� ������"�� ��� �� �! c� ���� ��� �� ������� ,� ��! �"�4� ��� ���� 11� ����! ������ �� �� �� �� �" +��� �"����4� ����� /�6�44 ������ 2��� � ��� +��� �"����4� ����� /�6 ����� �� ����� #������ ������ �" 2��� ��� ����� ����� ����#� �������� �� ���� �� � ���� �� � ���� ����� # �� ��� � � �! ���������� �� ������� 2� ���� �� ����� �" � ��#�� � �! ����� ������ � "��� ������� ��� ���� ���� +���� ���� + ����!� �� :������ c����"�����4� ��� ��� 11��� ���� ���� � �� ������ ������ �� �8 � +���� &���� ������ � � ��� ��� ����44 �6�� ��� ��� ����� � ������ ������� ��� #��� ������ �" ��� ����� �� ����2�� � ������ �� � # �� ��� ���� ���� � ��� ��� ���� �� ��� ���� � ��63��� ��� �� �" � ����� ��� �� ������ � ���� ��� ���� ���� � �� ��� ����� �� �� ������ �� � ���� ���" � �� ��� d��� � �� ����� �� ������ 11�� ���� 7��� +��!�� &��� ���� ���� �"� �� ����! ���!�� ��� ��������44 �� &���� � '���� �� ���!� �� ��� ����� ���� � ���#��� ��������������� ����! ���!�� ������� �� �� ���� 7��� +��!�� &��� ���� ��7+&� ���� �" ���� ���#��� �h.� �� %���� �� % � �� ��� ����� � ��������� �� ���� �� �� �� ������ ��� ������ ������� ���� ��� �7+& ���� �" ���� ���� � � "� 2������ ��� � �� � ��� �� ������� ���!�� �������� 11������� �" � ����! ���!�� � � �" ������"�8 ��� ���� � ' � ����� +���"�����44 �� ����� �d�� �� , ����� +�+���� ���� �� �����" ������� �" �� ���� � ��� � � �" �� � �� � ' � ����� +���"����� �� ����� "�� �" ��� ��������� ��� �� ���� �������� �� ��� �� �� ��� �� ����� ����� �������� d� �� �� ��� �h?h �� �hhh� � ��� ��� � ������ �� ����� � ��� ��� �� �����d� ��� ��� ������� ���� ��� %�3����3��� � � �� ��6� ����� ��� ���� ������� � ����� �� �� � ���� ���� ������ �� 2��� ���!� �� ��� &�c �6�� ��� ��� ��� ����� �� ������� ����"�� � ��� 2����� ���� ����� �� ��� �� � �� ��� ������ ������ � � �� ���� �� ��� �" ����� 11 �� &�c + &��� 2��� �� �6�� ��� �� �������8 � ����� �� � ��� �� 2� ��"� ���� �" ����� �44 �� 9� 0�� ����� ���!� �� ��� �� 2� ��"� �������� ��� ���� �" ����� ��� �� ��� &�c + &��� 2��� �� �6�� ��� ��� �� ��# � ������� �� ��� &��� 2� � ���� �� c������ /��� � ���� �� � &��� 2��� �� �6�� ��� ��� ;� �6��� �� ��� ������ ��� ���� �" ����� ���"� �� �� �� ���� �" ��� ������ �� !� 5� , ����� �� 5 �� �� *� ������ � ��� � ��� ���� 11 �� � ���� �� +� ��� d�� �� ���8 ��# �������� ��� �� ��� 2����� ���� ������44 �6�� �� ��� � ���� �� +� ��� d�� �� ��� � �hhh �� ��# � ����"�� 2����� ���� ������ �� ������� ��""��� ���� ��� ��� ����� 2����� ���� ��� �� 11 ��� ��� ���� �������� ��44 ���� ������ ���� ��� � ���� ��� ���� �� d�� ��� � ��� ��� ��� �� � 11 ���� ���� �������� ��44 #� �� ������ �� ��� ����� � � �� 2���� �� ���� ���� 9��� &� ���6�� �� 9��� &������ &� ;��!� �� '����� /� +�������� � 11� "� � �� �� � �" � �� ���� �� ��� �" �����44 ������ d� ��� ����2��� ������ �� �������� ��� ����� ��� # �� �� � �" � 2���� �� � ����� �� �� ��� �" ����� ��� ���� �� ��� ���� 2�� � ���� �" �� ��� ���� �� �� ��������� 9��� �� +���$!�� � �� :���� � ���� *���� �6�� �� ��� "��# �" ������� � ��� �� ��� � � ��� � ��� ���� 11� ��� �� ��� �8 �� ���� �� ����������44 �� ������� ���! �� ��� ������� ��# �� ��# � ��� � ���� �� ���� � �� �� � ��� �� ��� � �� ������� ��" ���� �� ���� ��� ���� &� �� ������ /� 9�� &������� �� 9� ��� �� /���� � ��� � ��� ���� 11� ������ ���! �� ��� �" ������" �� �� )��� �e� �� ����� 3�� ��������� ������ ���44 � ����� ���� �� ��� �� �� ����� "�� �" ���� ��� �" ������" �� �� ��� � ��� ��� �������� �� ������ �� ������� ����� �� ���� ��� �� ���� ������� �� ������" �� �� ���� �������� �� ���� ����� �6� � � ��#�� ��� 2� ���� � ��� �� �� ���� ��� ��� ���� �� i c���4� >�� � �6� 0�� &�� � '�� ���� ��� (��! )� ���� �� � 7���� �� &��� � �� 0�� *�#����� &� '��� ��� +����� )� ���� �� � ���!��� � f�# j����� �"��� �� � � ���� �� ������� ���� ���� �� �� � $������� �� ��� � ��� #��! � ���� ���� ������ ��� � �� ����# �" � ��� � ���� ��� �� �� ����� ������� ���� ��� ����� &�� � '�� ���� #� ��� ���� �� � , ��� � ;�� '����� �� ���! j���� ����� "��� ��� ����� ���� � ���#��� � �! �� ������ � � �� ������ �� � �����3���� �� � �� ��� )���� �� �" ��� ��� � ������� �" ��� �� �� �������� ��������� ��� 2� �� ���� � ����3����� �� �� ���! �" �� ��� � ���� ��� ���� �� �������� ������������������ �� ��� � �� � #������ ��� ���!�� � ��� �" �� �� �#�� �� ���� 2� ���� ��� ��������� �� �� ��� ���� � ��� � "� � � �#� ���!�� /� ���� �� ��"�� �� � �! ���� � ��� � !��� �� ����#� �� ��"�� �� � �! ���� �� � � ����� ���� � � �����"�� � ����" �" ���!���� ��������� � ����� ���!��� ���� � �� ��� ����� � ��������� �� ���� �� �� �� ������ ��� k ����2� �� � ������� �� ������"� � ��� �" �"� ��� ��� ��� � � �#� ���!��� ����"� ��� ��� �� 2� ��� �� � ����� ���!��� ���� ���� � � "��� �"� ��� ��� ���2��� � �� � ���� � ������"�� �� ����# �" ����� ������ ���� �� ����� ���� ����� �� ��� �� ��� ������ � � � ��� ���������� ����� �" �� *���� '���� ��� ����� ������ ����� ���� �� � ������ �� ������ �" � ��� � � ��� �� � �� ��! �" ������� � � ��� ��� � ���!���� ������ � ��� � ��� �� ��"������� ������� ���� �� 2��� ������ �� &���!3+����� �� ,������ � ��� 11&�������� �� �� �������� ���"� � ��� ������ ��� �"� �� � ���8 ������� �� �� ���� � �������� � "�� �� ��������44 ���� ��� ������ ����"�� �� ��" ���� �� ������� �" ��� ������ ��� �"� ����� � ������� �� ��� ������� ��# ���� ������� �� � �� �� ������ ��� ����� � �� � �� ��� ����� � 6 � ����� �� �� �� ������� ������ �� � ������ � ������� � � ���� ����� 2� #� �� � ��� �� "��� ��� ��������� �� � ��� �� "�� �� �������� �� ��6� ������ �� c � �� � 5 ���� �� * � 11/��! �� ���!�� �� � d� �� ����� ����� � ������� � ����� �� e� ���!���8 ��� f�# j����� �����44 ���� ��� #������ ��������� ��� �" �� �� ����!� ��������� ������" �� �� �6��� � ���!3��3���!�� �� � d� � ���� ��� � �� � ������ ����� ���� �� e� ���!���� �� ����� ��� ���� �� � ������"� � 2���� �� �������� � f�# j����� #���� � � �� ���� 2��� �� �������� � �� �6 ��� �� �� � ������ ;����� : ���d� �� c�� �� � 11c��� �� � ���� 2��� �� � � � "��� �-�� ����� ����" �" ���!���44 �6�� �� �������� ����� ������� �� �� � ��� �� �������� � ��! �� 2���� �� ���!���� 5� �� ������ ��� ��"�� ���� �������� #��� �� ��� � �! �� ����� �" � � #���3 ���� 2� ���� �� �� ���� �" ������� ���� ������� �� � ����� ��� ��� �� �� ��� ���� �� ��� ��! �� ������� �� �" �� �������� ������ � ��� �������� = � �� ��� ����� � ��������� �� ���� �� �� �� ������ ��� pii: s1057-0810(96)90003-7 financial services review, s(2): 87-99 copyright 0 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. risk aversion measures: comparing attitudes and asset allocation diane k. schooley debra drecnik worden households’ reported willingness to take financial risk is compured to the riskiness of their po~olios, measured as risky assets to wealth. overall, their ~or~olio alio~ations are reliable indicators of attitudes toward risk, demonstrating an understanding of their relative level of risk taking. ~u~t~variate regression analysis using ~lu~t~pl~~ imputed data from the 1989 survey of consumer finances indicates that households generally exhibit decreasing relative risk aversion. further, investment in risky assets is significantly related to socioeconomic factors, attitude toward risk taking, desire to leave an estate, and expectations about the adequacy of social security and pension income. i. introduction individuals’ risk preferences reflected in asset allocation choices have been explored extensively both theoretically (arrow, 1965, 1971; pratt, 1964) and empirically (friend & blume, 1975; cohn, lewellen, lease, & schlarbaum, 1975; siegel & hoban, 1982, 1991; riley & chow, 1992). evidence on how various factors, especially wealth, impact risk aversion is mixed. this study affords an increased understanding of individuals’ behavior toward risk-one of finance theory’s most fundamental concepts. the first hypothesis examined in this study is that relative risk aversion (rra) calcu lated from the composition of a household’s portfolio and rra reported by the househoid in terms of willingness to take ftnancial risk are directiy related and can be used inter changeably to proxy risk aversion. this study is the first to compare these alternative measures of relative risk aversion. differences between the two relative risk aversion mea sures may indicate that some individu~s do not understand risk and therefore, may be taking more or less risk than they actually desire. the movement to defined contribution pension plans in recent years has put many individuals in the position of becoming portfo diane k. schooley l associate professor of finance, boise state university, boise, id 83725; e-mail: rmkschoo@cobfac.idbsu.edu. debra drecnik worden l assistant professor of business and economics, george fox university, newberg, or 97 132-2697; e-mail: dwordeu~georgefox.edu, 88 financial services review 5(2) 1996 lio managers by requiring them to make asset allocation decisions. if investors do not underst~d risk, the studies that use investors’ asset allocations to measure risk aversion may not actually measure attitude toward risk. the comparison of rra calculated from asset allocation to reported rra will increase the understanding of how individuals deter mine their portfolio risk, thereby improving financial and retirement planning decisions and investor education. in addition to comparing rra measures, this study examines the factors that may explain variations across households’ rras calculated from asset allocation. the second hypothesis is that rra calculated from a household’s portfolio is related to its wealth, income, full-time employment, race, gender, stage of life cycle, attitude toward risk taking, desire to leave an estate, and its expectations about the economy and the adequacy of social security and pension income for maintaining a standard of living after retirement. theory and a review of the literature are discussed next. section iii explains the data set and variables. the examination of the relationship between calculated and reported rel ative risk aversion is presented in section iv and the determination of household relative risk aversion is discussed in section v. section vi contains the summary and conclusions. ii. theory and literature review the model used here follows friend and blume (1973, who estimate rra by maximizing an investor’s expected utility function using a taylor series expansion. they define the proportion of an investor’s portfolio invested in risky assets (a) as: w, r-1 * 1 h a= - 02(r,x) (1-l)(l-h)c l-h * ph,m (1) where r, is the return on the market portfolio of all risky assets; r-is the return on the risk-free asset; t is the investor’s tax rate; h is the ratio of investor’s human capital to his total wealth; 8, n is the ratio of the covariance of rm and rh (the return on human capital) to 0,‘; and c is pratt’s measure of relative risk aversion: because beta is estimated from time-series data to be close to zero (fama & schwert, 1977; liberman, 1980), equation 1 becomes: e(r,-r-1 * 1 ix= o’(r,) (1 -t)(l -h)c risk aversion measures 89 equation 3 can be rewritten as: (l-t)( 1-h)a = mpr * ; (4) where mpr is the market price of risk, assumed constant across all households. therefore, because (l-t)( 1-h)a is proportional to c (i.e., rra) and can be observed, inferences about rra can be made from (l-t)( 1-h)ol. if (l-t)( 1-h)a increases (decreases) as wealth increases, the individual is said to exhibit decreasing (increasing) rra. some economists argue for utility functions whose properties reflect increasing rra (arrow, 1971) while others favor the log utility function, which reflects constant rra. because rra depends upon the form of utility function being con sidered, the question of individuals’ actual rra is, for the most part, an empirical issue. empirical analyses of household rra vary in results depending, in part, upon how wealth is measured. because individuals hold residential housing for consumption as well as investment purposes, wealth has been measured as net worth excluding home equity. using this measure, friend and blume (1975) find decreasing rra and siegel and hoban (1982), who limit their sample to households between the ages of 50 and 64, find constant rra. morin and suarez (1983) also find decreasing rra, but include home equity in the wealth measure and treat it as a riskless asset because of the low uncertainty of the real stream of benefits it provides. they also include personal property as a riskless asset. in this study, home equity is excluded from wealth, as are other consumption goods such as personal property, vehicles and recreational craft. individuals exhibit decreasing rra when wealth is measured as total assets rather than net worth (cohn, lewellen, lease & schlarbaum, 1975; riley & chow, 1992). this study employs a measure of wealth that is net of the debt incurred to accumulate it. some studies include human capital as a component of wealth. when human capital (as well as home equity) is incorporated into the model as a risky asset, friend and blume (1975) find constant rra while siegel and hoban (1982) find increasing rra. bellante and saba (1986) find the inclusion of human capital in wealth dramatically changes how rra varies across age categories. when human capital is not included, they find that rra increases significantly with age for heads of households over 45 years of age. however, when human capital is considered a part of wealth, rra tends to decrease with age for all age categories. these results indicate that unless human capital is recognized as a risky asset, the human capital effects mask the life cycle effects. thus, measures of human cap ital and life cycles are included in this study. id. methodology all of the variables used in this study are computed from the 1989 survey of consumer finances (scf). this survey, sponsored by the federal reserve board, was conducted by the survey research center at the university of michigan between august 1989 and march 1990. the purpose of the scf was to provide a comprehensive view of the financial behavior of households. detailed information was gathered on all assets, both real and financial, and liabilities of the household, as well as demographic characteristics such as 90 financial services review s(2) 1996 age, race, education, family composition, and employment status. attitudes toward credit use, savings, and risk taking also were measured. the survey is ~stinguished from other household surveys, not only because of the vast amount of information gathered, but also because of its sample design and its treat ment of nonresponses. research has shown that distribution of wealth in the united states is skewed, with a relatively small proportion of households holding a large share of the wealth. in order to obtain more detail on the financial behavior of those households hold ing a ~spropo~ionate share of the wealth, the scf employed a dual-~~e sampling design (see herringa & woodburn, 1991). the final sample of 3,143 respondents con sisted of 2,277 randomly selected households from across the u.s. and 866 high-income households selected from a list developed by the internal revenue service. this dual frame sampling design prohibits the use of this sample as representative of the u.s. popu lation. while this sample cannot be used to make statistical inferences about population means and distributions, inferences can be drawn about the relationships between vari ables within households. the 1989 scf also differs from similar surveys in its treatment of nonresponses. the method of multiple imputation, advanced by donald rubin, replaces each missing value with a set of values that represent a distribution of possibilities. thus, this method attempts to simulate the distribution of missing data and provide a more realistic measure of the variability around the unknown data than simpler methods of estimating missing values. models are used to impute five alternative values for each missing item; for nonmissing variables, the values are the same in each of five observations. the final database consists of five complete observations for each respondent, which are combined for the analysis (see rubin, 1987; kennickell, 1991). the measure of actual risk taking by households is the ratio of risky assets to wealth, that is, the dollar value of risky assets per dollar of wealth. following the typical de~nition of risk, a risky asset in this study is one that provides an uncertain nominal cash flow. thus, the measure of human capital is included as a risky asset. it is recognized that the riskiness of human capital, measured by the uncertainty in income streams, varies across occupation and industry. however, these data are coded in a way that such differences cannot be accounted for. those respondents who were currently employed full-time (64% of the sam ple) reported that they expected to continue working full-time for “n” years. household human capital is calculated as the present value of an n-year annuity of the current annual salary or earnings, discounted at 7 percent. essentially, this assumes a discount factor of 10 percent, but allows earnings to grow at a 3 percent rate for inflation. alternative discount rates have no significant impact on the resuits of this study. complete definitions of this study’s asset and wealth measures are as follows: risky assets: the market value of all real estate held for investment purposes, the mar ket value of mutual funds, corporate stock, and precious metals, the face value of all cor porate and government bonds, mounts accumulated in all other pension accounts, loans to individuals, and an estimate of human capital. risk-free assets: checking and savings balances, money market accounts, u.s. sav ings bonds, cash value of life insurance, call account balances, certificates of deposit, other cash balances, and iralkeogh balances in cds or money market accounts. wealth: risky plus risk-free assets minus the value of mortgage and consumer debt outstanding. the market values of those assets that could be held for consumption as well as investment purposes (vehicles, recreational craft, and residential and personal property) risk aversion measures 91 are excluded, as is the value of outstanding debt incurred to accumulate these assets. only personal assets and liabilities are included in these measures; those owned or owed by busi nesses are excluded. the sampling design employed with this survey yielded a sample of households with an average of over $1 million in wealth. in order to make the results of the study more gen eralizable and comparable to other studies, those households with wealth greater than $1 million are excluded from further analysis. the study will focus on the 2,239 households with positive wealth of a million dollars or less. table 1 presents the socioeconomic and attitudinal characteristics of this truncated sample. even with the sample truncation, the sample is relatively wealthy. mean wealth is nearly $295,000, median wealth is almost $248,000, and average household income is about $43,000. but the median household income of $30,000 is comparable to the national 1988 median of $27,225 (u.s. bureau of the census, 1992). respondents generally feel that there is only an average risk of a major depression in the u.s. economy over the next 10 years. the risk of double-digit inflation during the same time period is believed to be slightly higher. on average, respondents do not believe that their expected or current retire ment income from social security and pensions is adequate to maintain their living stan dard. at the same time, 50 percent of the respondents believe that leaving an inheritance or estate is important. table 1 socioeconomic and attitudinal characteristics of the sample (n=2239) financial characteristics risk-free assets risky assets human capital wealth risky assets/wealth household income net worth mean value $29,586 $277,616 $213,511 $294,825 0.807 $42,835 $162,935 (all assets all debt, excluding human capital) non-employed (no full-time employment) 29.9% attitude the economy over the next 10 years risk of major depression risk of double digit inflation 0 = no risk, lo=very great risk mean rating 5.24 5.11 attitude -retirement income rating of adequacy o=totally inadequate 5=enough to maintain living standards lo=very satisfactory mean rating 3.81 attitude -leaving an estate distribution (%) very important 19.2 important 30.8 respondent and partner differ 1.1 somewhat important 27.9 not important 21.0 92 financial services review 5(2) 1996 table 1 (continued) characteristic of head of household distribution (a) mean ratio risky assets/wealth life cycle: single, < 45 yr, no children 9.8 single parent, any age 5.8 married or with partner, < 45 yr 28.5 older, in labor force, 2 45 yr 31.6 older, retired, not in labor force, 2 45 yr 24.3 marital status: married or living with partner single gender: male female race: white black hispanic asian/american indian/other education: no high school diploma high school diploma some college college degrees 65.2 34.8 76.8 23.2 81.0 9.3 5.6 4.1 21.5 32.0 20.1 26.4 0.939 0.849 0.998 0.904 0.397 f = 269.9* 0.868 0.695 f = 92.0* 0.859 0.639 f = 116.7* 0.800 0.824 0.904 0.885 f = 3.98* 0.608 0.824 0.870 0.904 f = 55.9* notes: *mean variables are significantly different acrc~ss groups, at the i percent level. f statistic is derived from the analysis of the combined multiple impulations and can be interpreted here similarly to the chi-squared statistic. over 50 percent of the household heads in the sample are 45 years of age or older, with over one-half of those still in the labor force. about two-thirds of the respondents are mar ried or living as partners, and three-fourths of the households are headed by males. only 19 percent of the respondents are nonwhite, and nearly 50 percent of the heads of household have at least some post secondary education. a. univariate analysis the univariate analysis presented in the second part of table 1 indicates that the mean level of risk taking, as defined by the ratio of risky assets to wealth, does vary significantly across demographic groups in the sample. older households whose head is no longer in the labor force have, on average, less than half the value of risky assets per dollar of wealth than other households. these households have less human capital than those in other cycles of life. households consisting of couples in their family formation years exhibit the highest value of risky assets per dollar of wealth. an examination of marital status reveals that single respondents have significantly fewer risky assets per dollar of wealth than other households. one explanation may be that households of couples are more likely to have two incomes and thus a larger amount of risk aversion measures 93 human capital. in addition, 38 percent of single respondents are in the older stage of the life cycle and not in the labor force. the portfolios of households headed by females have sig nificantly fewer risky assets per dollar of wealth than those headed by males. one reason for this result may be the coding procedure used in the creation of the data set, rather than inher ent gender differences in risk aversion. the responses of opposite-sex couples were coded such that the male is the “head of household.” therefore, the marital status and gender of the household head are highly correlated. across the race categories, white households have the lowest value of risky assets per dollar of wealth, while hispanic households have the highest. as the education level of the household head increases, so does the value of risky assets per dollar of wealth. while this is related to human capital (higher education is asso ciated with higher earning streams), it may also be the case that a more highly educated household would make more financially sophisticated, and thus riskier, investments. the multiv~ate analysis presented in section v will examine the relationship of each of the household’s socioeconomic characteristics to the value of risky assets per dollar of wealth, holding all else constant, to provide a clearer understanding of the effect of these factors. one other determining factor in portfolio composition is the household’s reported attitude toward risk taking, which is examined in the following section. iv. calculated vs. reported rjzlative risk aversion one-way analysis of variance is used to test whether the means of calculated rra are sig ni~cantly different across the household’s reported attitude toward risk taking. the results indicate whether calculated and reported rra are measuring the same construct (relative risk aversion). calculated rra ((l-h)@ is the ratio of risky assets to wealth, where the numerator and denominator include human capital. while equation 4 illustrates that observed rra should be (1-t)( 1-h)a, the tax rate (t) is difficult to calculate from this data set. bellante and saba (1986) find that adjusting for taxes does not affect their results and friend and blume (1975) show that not taking tax differentials into account may only slightly bias the rra estimate downward. reported rra is a categorical response vari able derived from the response to the question: “which of the statements on this page comes closest to the amount of financial risk that you (and your husband/wife) are willing to take when you save or make investments?’ 1. take substantial financial risks expecting to earn substantial returns. 2. take above average financial risks expecting to earn above average returns. 3. take average financial risks expecting to earn average returns. 4. not willing to take any financial risks. results from the one-way analysis of variance for testing whether asset allocation is different across responses to the above question are found in table 2. the mean values of risky assets to wealth are significantly different across the four response categories and are in the expected order of size. those respondents willing to take no risk have the lowest mean ratio of risky assets to wealth, with the value increasing with the willingness to take risk. a r-test for differences between categories indicates that there is no significant differ ence in the mean values for the “substantial” vs. the “above average” responses. however, 94 financial services review 5(2) 1996 table 2 mean values of risky assets to wealth across reported risk aversion risk measure risky assets wealth (% of sample) reported amount cjf risk willing to take test substantial above average average none statistic .982 ,941 ,858 ,722 33.04* (3.9%) (9.1%) (41.1%) (45.9%) nofrs: *fstatistic indicates significant differences in mean values across groups, at the i percent level. n = 2239. risky assets are those measured assets whose cash flows are uncertain (including human capital). there is a significant difference between the mean values for all other categories. these results indicate that the households surveyed understand their relative level of risk taking. while the mean risk measure for those who reported that they are not willing to take any financial risk is very high (.722), it is significantly less than the mean value for the “average” risk category. one explanation for the high value is that there was no category indicating a willingness to take “less-than-average” financial risk. it is possible that many of the respondents would have chosen this category rather than the “no risk” response. fur ther, the definition of risky assets is quite broad; in particular, it includes all accumulated pension funds that are not ira/keogh balances invested in cds or money market accounts. these funds are either invested in stock or interest-bearing accounts. pension funds com prise a considerable proportion of the assets owned by these households; on average, accu mulated pension funds equal 21.6 percent of a household’s total financial assets. v. determination of household relative risk aversion multivariate regression analysis is used to test the second hypothesis that calculated rra (i.e., portfolio composition) is a linear function of the household’s socioeconomic charac teristics and attitudinal factors such as its attitude toward risk taking, desire to leave an estate, expectations about the economy, and the adequacy of social security and pension income for maintaining a standard of living after retirement. the linear model estimated is y=xr+p (5) where y is the household’s dollar value of risky assets per dollar of wealth (with higher val ues indicating lower rra), xl3 is the matrix of variables and parameters determining y, and u denotes the random component-attributes of the household that are not observed or can not be measured, but impact the portfolio composition. definitions of the socioeconomic and attitudinal explanatory variables included in the model are found in table 3. note that the univariate analysis presented in table 1 indicates a relationship between the level of education achieved by the household head and the value of risky assets per dollar of house hold wealth. however, high positive correlation between education and household income prohibits the inclusion of both variables in the final analysis. similar estimation results are risk aversion measures 95 table 3 definitions of explanatory variables z.n wealth: natural logarithm of the dollar value of household wealth. household income ($cvo)c 1988 before-tax household income from all sources. non-employed: 1 for those households where neither the head of household where neither head of household or partner (for couples) is a full-time wage earner; 0 otherwise. race -nonwhite: 1 if head of household is hispanic, african-american, or other nonwhite race: 0 otherwise. female: 1 if head of household is female; 0 if male. life cycle of household head family formation: 0 if head of household is < 45 years old, married, with or without children (in constant). mean age = 35; mean number of dependents = 2.8. young single: 1 if head of household is < 45 years old, single, without children; 0 otherwise. mean age = 32; mean number of dependents = 0.1. single parent: 1 if head of household is any age, single, with children; 0 otherwise. mean age = 39; mean number of dependents = 2.0. older working: 1 if head of household is 2 45 years old, in labor force; 0 otherwise. mean age = 56; mean number of dependents = 1.4. older refired: 1 if head of household is 2 45 years old, retired, or otherwise not in labor force; 0 otherwise. mean age = 7 1; mean number of dependents = 0.7. estate: 1 if respondent believes it is very important or important to leave an estate or inheritance to surviving heirs; 0 otherwise. depression; values of 0 to 10 indicating the respondent’s expectation of the u.s. economy experiencing a major depression within the next 10 years; 0 = almost no risk, 10 = very great risk. inflation: values of 0 to 10 indicating the respondent’s expectation of the u.s. economy experiencing double digit inflation during the next 10 years; 0 = almost no risk, 10 = very great risk. refirement income; values of 0 to 10 indicating the respondent’s rating of the retirement income expected (or cur rently receiving) from social security and job pensions; 0 = totally inadequate, 5 = enough to maintain living standards, 10 = very satisfactory. risk taking substantial: 1 if the respondent is willing to take substantial financial risks expecting to earn substantial returns; 0 otherwise. above average: 1 if the respondent is willing to take above average financial risks expecting to earn above average returns; 0 otherwise. average: 1 if the respondent is willing to take average financial risks expecting to earn average returns; 0 otherwise. none: 0 if the resnondent is not willinn to take any financial risk. obtained when variables measuring the education of the household head are substituted for household income in the model. the analysis is performed on each of the five imputations in the data set. the estimated parameters from each are combined, taking into consideration the variation across the imputations. using the multiple imputations in this manner increases the efficiency of the estimated parameters and test statistics. the use of a single imputation for the estimation of the nonresponses leads to biased results (see rubin, 1987). because of this methodology, an f statistic is reported to test the significance of each estimated coefficient, rather than the traditional t-statistic. the observed level of significance, the p-value, is reported with each statistic to facilitate the evaluation of the results, which are reported in table 4. the model’s overall explanatory power is significant, with adjusted r2 for the separate imputations ranging from 48 to 52 percent. the estimated coefficient on the log of wealth is significantly positive, indicating decreasing rra. that is, when other socioeconomic factors and the measured expectations and attitudes are held constant, this study finds that 96 financial services review 5(2) 1996 expianafury variable table 4 regression analysis of risky assets to wealth estimated cmffici6tlt f statistic p-value ln wealth household income ($000) non-employed household head demographics race -nonwhite female stage of life cycle young single single parent older working older retired attitude~xpectations estate depression inflation retirement income risk taking substantial above average average .06t* 162.40 .000003 .oo -.288* 112.12 .oss* ii.11 ,000 -.025 1.83 .178 ,030 1.49 ,223 ,040 1.52 .218 -.015 .71 ,399 -.168* 32.28 boo .030** 5.34 ,021 -.003 .97 ,325 ,003 .97 .325 -.oo7* 8.02 ,005 .129* 13.34 ,000 ,029 i.51 ,219 ,010 .48 .490 .ooo ,969 .ooo constant .213* 10.43 .ool notes: dependent variable mean value = ,807. mean wealth = $294,825. n = 2239. overalt fstatistic = 123.26*. (r2 for separate imputation regressions are reported in appendix.) ‘significant at the 1 percent level. an f statistic, rather than the traditional 1-statistic, is calculated from the estimated parameters and parameter variances across the five imputations. the p-value is the observed level of significance associated with each f statistic. **significant at the 5 percent level. increases in househoids’ holdings of risky assets per dollar of wealth are positively related to increases in their wealth. while household income is not significant, whether a household head, and/or partner, are full-time wage earners is significantly related to the holdings of risky assets per dollar of wealth. the negative sign on the coefficient may indicate that those households with no full-time earnings are less willing to hold risky assets. in addition, this categorical variable reflects whether a household has estimated human capital; holding all else constant, house holds with a zero value for human capital are expected to have fewer risky assets. because this variable holds constant the inclusion of human capital in risky assets, the impact of other demographic variables can be more clearly estimated. gender of the household head is not significant, once such factors as life cycle and employment are held constant. but, nonwhites have significantly higher risky assets to wealth than do whites, holding other factors constant. siegel and hoban (1991) find that race does not have a significant effect on a similar risk measure. this study’s results may differ because siegel and hoban’s anal ysis adjusts for several socioeconomic factors not included here, such as home ownership, self-employment, health limitations, and family size. the nonwhite households in this risk aversion measures 91 study’s sample have significantly more dependents and less education than whites and are less likely to be homeowners. the coefficients for the life cycle of the household head reveal that older households whose head is retired, or otherwise not in the labor force, have significantly lower risky assets relative to wealth than do households in their family formation years. in response to the question on financial risk taking, 64 percent of the older retired households reported that they would take no risk at all. further, the mean value of risky assets to wealth for those households is .324, a value much lower than .722, the mean ratio for all households who reported that they would not take financial risk (table 2). the fact that fewer house holds in this group have estimated human capital is being held constant with the inclusion of the “non-employed” variable. even when the smaller percentage of pension assets for the older retired households is considered (9 percent of total financial assets vs. the overall average of 22 percent), this difference still indicates a tendency to choose less risky invest ments relative to wealth. note that other households do not differ significantly from those in their family forma tion years in the holdings of risky assets relative to wealth. this result may suggest that when other socioeconomic factors are held constant, family responsibilities do not impact relative risk aversion. the desire to leave an inheritance (estate) is significantly positively related to the level of risky assets relative to wealth. this result, which siegel and hoban (1991) also find, pro vides evidence that individuals recognize the positive relationship between leaving an inheritance and investing in relatively riskier assets. an interesting relationship that has not been explored in other studies is that between asset allocation and the rating of the adequacy of social security and pension income for retirement. this study finds that the less adequate those sources of retirement income are expected to be, the more risk households take in their portfolios. again, these results pro vide evidence that investors recognize the need to take more risk in order to earn a higher portfolio return. households’ expectations about a future depression or about inflation are not significantly related to the ratio of risky assets to wealth. dummy variables capturing the household’s attitude toward risk taking indicate that holding all socioeconomic factors constant, those respondents who claimed that they were willing to take substantial financial risk to earn substantial returns actually have a signifi cantly higher ratio of risky asset to wealth, compared to those who were not willing to take any risk. yi. summary and conclusions as more households are taking responsibility for the asset allocation of their portfolios, an understanding about attitudes toward financial risk and its relationship to expected return is of growing interest. this study finds that households in this sample do allocate portfolio holdings consistent with their professed attitudes toward taking risk to increase returns. risky assets are defined to include the value of financial assets that provide an uncertain cash flow, the market value of real estate held for investment purposes, and an estimate of human capital. these findings suggest that a household’s relative risk aversion (rra) can be assessed by responses to questions about risk aversion, as well as by measuring asset 98 financial services review 5(z) 1996 allocation. the implication is that the households sampled do understand the basic risk/ return relationship; an investor must be willing to accept more uncertainty (higher risk) to earn higher expected returns. some have suggested that households are taking more risk by choosing “safe” investments such as cds and savings accounts because these investments may not provide returns suffl~ient to rn~nt~ purch~ing power, al~ough the cash flows from the investments are relatively certain. nevertheless, this study’s results indicate that households still recognize the traditional meaning of financial risk as variability (or uncer tainty) of returns. regression analysis of the ratio of risky assets to wealth indicates that this sample of households exhibits decreasing rra. that is, as wealth increases, households allocate a greater portion of their portfolios to risky assets, holding constant attitudes about risk and the economy, as well as socioeconomic factors. those households where neither the head or partner is a full-time wage earner have significantly fewer risky assets relative to wealth, a factor that may simply capture the effect of no estimated human capital. nonwhites have higher risky assets to wealth than do whites, a topic for future research. older retired households allocate less of their portfolios to risky assets than households in their family formation years. of particular interest is household attitude toward social security and pension income. those households who have less confidence in these sources of income for maintaining living standards have larger portions of their portfolios invested in risky assets, implying the reco~ition that higher expected returns are associated with higher risk. finally, the results reveal that individu~s in this sample understand the relative level of riskiness in their portfolios; those who say they are willing to take substantial risk to earn higher return do have riskier portfolios, as compared to those who are not willing to take any financial risk at all. these results provide further understanding of the factors that influence individuals’ asset allocation. as cited throughout the paper, previous evidence on individuals’ rr4 is mixed, which is likely due to the different samples and the measures of wealth used in each study. the determination of individuals’ risk-taking attitudes and behavior may be so complex that it is not possible to characterize households as exhibiting a particular rr4. acknowledgment: the authors thank two anonymous reviewers for their helpful com ments and express appreciation to professor george mccabe, marcey abate, and jian zhao of the statistics department at purdue university for developing the methodology used in the analysis. reterences arrow, k.j. (1971). essays in the theory of risk-bearing. chicago: markham publishing company. arrow, k.j. (1965). aspects of the theory of risk bearing. yfjo johnsson lectures, helsinki. bellante, d., & saba, r.p. (1986). human capital and life-cycle effects on risk aversion. joumae of financial research, 9,41-5 1. cohn, r.a., lewellen, w.g., lease, rx!., & schlarbaum, g.g. (1975). individual investor risk aversion and investment portfolio composition. journal of finance, 30,605~620. fama, e., & schwert, g. (1977). human capital and market equilibrium. journal of financial ecq nomics, 4,95-125. risk aversion measures 99 friend, i., & blume, m.e. (1975). the demand for risky assets. american economic review, 65, 900-922. herringa, s.g., 8~ woodbum, r.l. (1991). the 1989 survey of consumer finances: sample design documentation. mimeo, survey research center, university of michigan, ann arbor. kennickell, a.b. (1991). imputation of the 1989 survey of consumer finances: stochastic relaxation and multiple imputation. proceedings of the section on survey research methods, american statistical association, atlanta, ga. liberman, j. (1980). human capital and the financial capital market. journae of business, 53, 165 191. morin, r.a., & suarez, f. (1983). risk aversion revisited. journal of finance, 38, 1201-1216. pratt, j.w. (1964). risk aversion in the small and in the large. econometrica, 32, 122-136. riley, w.b., jr., & chow, k.v. (1992). asset allocation and individual risk aversion. financial analysts journal, 48, 32-37. rubin, d.b. (1987). multiple imputation for nonresponse in surveys. new york: john wiley and sons. siegel, f.w., & hoban, j.p., jr. (1991). measuring risk aversion: allocation, leverage, and accumu lation. journal of financial research, 14, 27-35. siegel, f.w., & hoban, j.p., jr. (1982). relative risk aversion revisited. review of economics and statistics, 64, 481-487. u.s. bureau of the census. (1992). statistical abstract of the united states: 1992, 112th ed. wash ington, dc. pii: s1057-0810(00)00052-4 determinants of planned retirement age catherine phillips montaltoa,*, yoonkyung yuhb, sherman hannac aassistant professor consumer and textile sciences department, the ohio state university, 1787 neil avenue, columbus, oh 43210-1295, usa blecturer in consumer sciences, 203-dong 1403-ho, hyundai apartment, dowha-dong, mapo-ku, seoul korea cprofessor, consumer and textile sciences department, the ohio state university, 1787 neil avenue, columbus, oh 43210-1295, usa abstract determinants of planned retirement age are analyzed. the prediction equation indicates that planned retirement age increases substantially as people get older, and increases somewhat with higher noninvestment income. social security reform should recognize that the capacity to continue working and the ability to afford to retire both influence the age at which people plan to retire. the range of planned retirement ages suggests that research on the adequacy of retirement preparation should focus on planned retirement age. financial planners should consider the finding that planned retirement age increases with age. © 2000 elsevier science inc. all rights reserved. jel classification:d12; j26 keywords:planned retirement age; retirement adequacy 1. introduction the financial viability of the social security program in the united states is being debated in response to the aging of the population. the percentage of the u.s. population made up of persons 65 years of age and over is 15% today and is projected to increase to 20% over the next 30 years (u.s. bureau of the census, 1998). the demographic pressures of * corresponding author. tel.:11-614-292-4571; fax:11-614-688-8133. e-mail address:montalto.2@osu.edu (c.p. montalto). financial services review 9 (2000) 1–15 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00052-4 population aging will require forward-looking action from policy makers to preserve the financial viability of the social security program. one proposed change is to raise the age of eligibility for full retirement benefits further or more rapidly than the currently planned gradual increase from 65 to 67 over the next 25 years. proposals have also been made to increase the early retirement age from age 62 to age 65 (mitchell & quinn, 1995). the rationale for increasing retirement age is to reduce the long-term deficit in the social security trust fund. increasing retirement age would increase the number of years a worker spends in the workforce thereby increasing the amount the worker contributes to the trust fund. additionally, increasing the retirement age would decrease the number of years a retiree spends in retirement thereby reducing the benefits drawn out. the primary justification given for increasing the retirement age is the longer life expectancy and improved health of the nation’s elderly. raising the retirement age may improve the financial solvency of the social security system, but it will also affect the economic well-being of individuals. implications for individual well-being depend on the importance of social security income, the impact of delayed receipt of social security income, the ability to continue working to the age of eligibility, and individual preferences related to retirement age. additionally, raising the age of eligibility for social security benefits has possible spill-over effects to other government programs, such as supplemental security income and disability insurance (bovbjerg, 1998). previous research documents that a worker’s decision to retire is influenced by rules governing pensions and social security benefits, wealth, characteristics of jobs held by elderly workers, health insurance coverage, and social norms (fields & mitchell, 1984; hurd, 1997). during this century, social norms and enacted legislation have resulted in retirement at earlier ages. labor force participation rates of older males declined throughout most of this century, and then stabilized in the mid-1980s. today, older men (55 to 64 years of age) frequently leave full-time career jobs, but continue working part-time or part-year rather than completely withdrawing from the labor force (employee benefit research institute, 1999). although age 65 is currently the age of eligibility for full retirement benefits under the social security program, everyone does not plan to retire at age 65. the range of planned retirement ages is quite large. five percent of today’s workers plan to retire before age 55, while 22 percent plan to stay in the labor force until at least age 66 (retirement confidence survey, 1999). despite evidence of variation in the age at which individuals plan to retire, some recent studies (mitchell & moore, 1997; bernheim, 1996) evaluating retirement wealth adequacy of preretirees assume age 65 as the retirement age. the standardized assumption of retirement at age 65 without allowing for individual differences can result in significant overestimation or underestimation of the adequacy of retirement wealth. clearly, assumptions made about planned retirement age are critical in determining whether people have saved ‘enough’ for retirement. yuh, montalto, and hanna (1998) and yuh, hanna, and montalto (1998) find that planned retirement age has a substantial impact on the estimated adequacy of preparation for retirement. even though about 75% of workers elect to retire before age 65, there are proposals to increase the minimum age to receive any social security retirement pension (apfel, 1998). therefore, it is worthwhile to study factors related to planned retirement age to see which types of workers would be impacted most by increases in the minimum age for 2 c.p. montalto et al. / financial services review 9 (2000) 1–15 receiving social security benefits. additionally, planned retirement age is an important variable in developing rational savings plans for retirement, so improving understanding of planned retirement age has implications for financial planning. this research investigates the determinants of planned retirement age. few studies have addressed this issue directly. although previous studies on retirement behavior have analyzed the observed age of retirement among retireesex-post,little research has focused on the planned retirement age of preretired workersex-ante.understanding the determinants of the age that current workers plan to retire is important because the planned retirement age of preretirees is a crucial factor affecting saving and investment decisions during the working years. additionally, proposed increases in the minimum age for receiving social security benefits will have the most impact on workers that plan to retire before the minimum age. the paper is organized as follows. section 2 provides a review of relevant literature and presents the conceptual framework underlying our estimation of determinants of planned retirement age. the methodology is presented in section 3, and the results and discussion are provided in section 4. the summary and policy implications are presented in section 5. 2. literature review there have been many studies on retirement issues since the 1970s. typically, actual retirement has been treated as a choice variable in the literature, and various economic factors have been shown to play an important role in the retirement decision. 2.1. related empirical research boskin (1977) tries to explain the long-term decline in the labor-force participation of all male age-groups. using data from the panel study of income dynamics for 1968 through 1972, he finds that the value of current annual social security retirement benefits has a pronounced effect on the decision to retire. the level of net earnings has a strong negative effect on the probability of retirement. quinn (1977) examines the microeconomic determinants of early retirement among white married men aged 58–63 using the 1969 retirement history study. the relative impact of three sets of factors in explaining older men’s labor-force participation decisions are investigated: personal and financial characteristics, local labor market conditions, and certain attributes of the individual’s job. quinn finds that health status and current eligibility for social security and other pensions are the most important determinants of retirement, and that there is a definite interaction between the two—persons in poor health are more likely to retire in response to financial incentives from social security and other private pensions. kotlikoff (1979) estimates a model for expected age of retirement using data from the national longitudinal survey (nls) of older men. private pension coverage is an important predictor of expected retirement age. coverage under a private pension plan is associated with expected retirement 1.2 years earlier; for government pension coverage the impact is 1.8 3c.p. montalto et al. / financial services review 9 (2000) 1–15 years. age has a positive and significant effect, and the health and employment attitudinal variables all have the anticipated negative effects. diamond and hausman (1984) examine factors that affect the actual retirement decision using the national longitudinal survey of older men. the presence of pensions and social security benefits, the level of permanent income, and poor health have strong, positive effects on the probability of retirement. they argue that planned retirement dates change over time. in fact, while planned retirement age has some predictive power for actual retirement age, much unexplained variance remains. honig (1996) uses data from the first wave of the health and retirement survey and finds evidence that expected and observed retirement functions are similar. honig suggests that retirement expectations may accurately forecast retirement behavior. burtless and moffitt (1985) develop and estimate a model of the joint choice of retirement age and post retirement hours of work by the aged population using data from the longitudinal retirement history survey (lrhs). they find that social security influences both retirement age and choice of post retirement hours of work, but the magnitude of the effect on the age of retirement is small. they also find that an earlier retirement age is related to poor health, lower levels of education, and higher pre retirement wage rates. burtless (1986) develops a retirement age model and estimates the model using the longitudinal retirement history survey. poor health, being married, household size, and wealth in excess of $25,000 all reduced the age of retirement. samwick (1998) investigates the incentive effects of social security and pension benefits on retirement using data from the 1983 survey of consumer finances and the corresponding pension provider survey. the results suggest that the retirement decision is much more sensitive to changes in retirement wealth than to the level of retirement wealth. further, changes in retirement wealth are primarily determined by pensions, and not social security. samwick finds small effects of social security on retirement, and much more substantial effects of pensions. uccello (1998) examines the relative importance of health status, income, employment characteristics, and demographic characteristics in the decision to retire using data from the 1990 survey of income and program participation and the 1994 wave of the health and retirement survey. simulations reveal that health insurance coverage solely through one’s employer and presence of a working spouse have the largest negative effects on the expected level of retirement. pension coverage, employment in a physically demanding occupation, and being nonwhite have the largest positive impact on the expected level of retirement. the previous research focuses primarily onex-postanalyses of the observed retirement age of retirees using a work-leisure model or a life cycle labor supply model. typically, data on actual retirement behavior is used to estimate the probability of being retired as a function of social security and pension benefits, and other demographic characteristics. the results consistently confirm that higher earning power and good health reduce the probability of retirement, while eligibility for and higher levels of retirement benefits, and higher financial wealth increase the probability of retirement. previous research has not analyzed factors affecting the age that currently employed workers plan to retire. 4 c.p. montalto et al. / financial services review 9 (2000) 1–15 2.2. conceptual model a currently employed individual choosing a planned retirement age must consider whether resources will be adequate, whether working will be possible, and also his or her individual preferences for leisure. the ability to afford to retire is influenced by the individual’s accumulated financial resources as well as the earned retirement benefits. the ability to continue to work is influenced by individual productivity and health, as well as characteristics of jobs. preferences for leisure may be influenced by social norms, but also vary across individuals at a given point in time. the data set used in the empirical analysis allows examination of several of these factors. a definition of retirement is required before the determinants of planned retirement age can be analyzed. however, there is no consensus in the literature on the definition of retirement (gustman, mitchell & steinmeier, 1995). various definitions of retirement have been used by economists and other social scientists, including: self-reported retirement; termination of work or looking for work; termination of full-time work; working less than a given number of hours; leaving the main employer (a long-term job); and receipt of an employer-provided pension or social security benefits. this study defines retirement as occurring when an individual stops working full-time which is the definition most commonly used in empirical studies (sickles & taubman, 1986; diamond & hausman, 1984). 3. methodology 3.1. data data for this study are drawn from the public use tape of the 1995 survey of consumer finances (kennickell, starr-mccluer & sunde´n, 1997). the survey of consumer finances (scf) is a triennial survey sponsored by the federal reserve with the cooperation of the department of the treasury. the purpose of the scf is to provide comprehensive and detailed information on the financial characteristics of u.s. households. a total of 4,299 families were interviewed in the 1995 scf survey. the 1995 scf has five complete data sets called “implicates” as a result of multiple imputation to handle missing data. this study uses repeated-imputation inference (rii) techniques to combine the five different data sets to make valid inferences (rubin, 1987; montalto & sung, 1996). the survey of consumer finances was chosen for this study because it provides information on a broad age-range of the u.s. population, and it specifically asks currently employed respondents to provide their planned retirement age. to analyze determinants of planned retirement age, heads of household age 35 to 70 years who were currently working full-time were selected, resulting in a sample of 1,607 individuals. fig. 1 shows the cumulative distribution of the planned retirement age. about 17% of the sample planned to retire by age 55, 35% planned to retire before age 62, and 51% planned to retire by age 62. almost all respondents (89%) planned to retire by age 65, and 91% planned to retire by age 67. 5c.p. montalto et al. / financial services review 9 (2000) 1–15 3.2. sample selection planned retirement age in this study is defined as the age at which the individual plans to stop working full-time. thus, the variable of interest is only observed for those individuals currently working 35 hours per week or more. if current hours of work and planned retirement age are correlated, then analysis of planned retirement age using only the sample of individuals currently working full-time will produce inconsistent estimates of the parameters of the planned retirement age equation. a positive correlation between current hours of work and planned retirement age is plausible, since a “taste” for work would likely result in more hours of work and a later age of planned retirement. if this “taste” for work is not controlled in the planned retirement age equation, a specification error is committed by omitting a relevant variable. this type of specification error is commonly referred to as selection bias. in other words, an estimation of the effect of age on planned retirement age may be biased because older workers with a preference for earlier retirement are selected out of the sample of individuals currently working full-time. heckman’s (1979) two-step estimation procedure is used to estimate a planned retirement age equation that includes a variable to correct for potential selection bias. in the first step, probit analysis is used to estimate the probability of working full-time for the full sample of heads of household age 35 to 70 years. the probit results are then used to calculate the selection bias correction variable (also referred to as the inverse mills ratio) for each observation. in the second step, the determinants of planned retirement age are estimated by ordinary least squares on the sub sample of heads of household age 35 to 70 years who were currently working full-time. the selection bias correction variable is used as an independent fig. 1. cumulative distribution of planned retirement age. 6 c.p. montalto et al. / financial services review 9 (2000) 1–15 variable in this equation, thereby producing consistent estimates of the parameters of the planned retirement age equation. 3.3. probability of working full-time equation the probability of currently working full-time is estimated with a probit regression on the sample of heads of household age 35 to 70 years (n 5 2,731). the dependent variable is a dichotomous variable equal to one if the respondent is currently working full-time, zero otherwise. independent variables include variables capturing potential barriers to full-time employment as well as the standard human capital variables. the probit estimating equation contains eighteen independent variables and can be represented as full-time 5 b0 1 b1 black non-hispanic1 b2 hispanic1 b3 other 1 b4 unmarried male, living alone1 b5 unmarried female, living alone 1 b6 unmarried male, living with others1 b7 unmarried female, living with others 1 b8 poor health1 b9 children under 6 years1 b10 children 6 to 17 years 1 b11 high school1 b12 some college1 b13 college grad1 b14 age 1 b15 age over 451 b16 age over 551 b17 age over 651 b18 experience (1) potential barriers to full-time employment are measured with categorical dichotomous variables for race/ethnicity and marital status/living arrangement, and dichotomous variables for self-reported poor health of the respondent, presence of children under age 6 in the household, and presence of children 6 to 17 years of age in the household. human capital is measured with categorical dichotomous variables for education, a spline variable for respondent’s age (suits, mason & chan, 1978), and a continuous variable measuring the respondent’s previous years of full-time work experience. means for the continuous variables and percentages for the dichotomous variables are presented in table 1 . the probit equation is estimated on the combined data from the five implicates of the survey of consumer finances, resulting in unbiased coefficient estimates. in order to correct the standard errors for imputation error, the estimated covariance matrix of the estimated coefficients is needed. the probit procedure in sas does not generate this matrix when the model includes dichotomous variables as dependent or independent variables (sas institute inc., 1990, p. 1338). as a result, the standard errors cannot be corrected for imputation error, and the statistical significance of the coefficient estimates may be overestimated. however, since the purpose of the probit equation is to generate the selection bias correction variable, the criteria of unbiased coefficient estimates is relatively more important, and the significance of relationships of lesser importance in this application. 3.4. planned retirement age equation the determinants of planned retirement age are estimated by ordinary least squares on the sub-sample of heads of household age 35 to 70 years who were currently working full-time. repeated-imputation inference (rii) techniques are used to combine data from all five implicates of the survey of consumer finances to generate the coefficient estimates of the 7c.p. montalto et al. / financial services review 9 (2000) 1–15 planned retirement age equation. standard errors are corrected for imputation error to enable valid tests of significance of coefficients. the dependent variable is the planned retirement age of the respondent. the independent variables are selected in accordance with the conceptual model where a currently employed individual considers the adequacy of retirement resources, the feasibility of continued employment, and individual preferences for leisure when selecting the age at which retirement will occur. the independent variables include financial variables and variables capturing access to resources, characteristics of employment, and respondent demographic characteristics and perceptions. the selection bias correction variable is included as an independent variable to correct for potential selection bias. means for the continuous variables and percentages for dichotomous variables are presented in table 2 . the ordinary least squares regression equation contains twenty seven independent variables and can be represented by table 1 descriptive statistics and probit analysis of the probability of working full-time variable mean percent1 probit regression2 estimate std. error p-value3 working full-time (dependent variable) 67.4% intercept 0.5026 0.6099 0.4099 respondent’s race/ethnicity (reference category: white non-hispanic) black non-hispanic 12.6% 20.2844 0.1116 0.0109* hispanic 5.7% 20.1621 0.1540 0.2926 other races 4.2% 0.1842 0.1566 0.2396 marital status/living arrangement (reference category: married or living with partner) unmarried male, living alone 7.3% 20.2421 0.1189 0.0417* unmarried female, living alone 11.9% 20.2662 0.1207 0.0274* unmarried male, living with others 3.9% 20.3032 0.1782 0.0889 unmarried female, living with others 13.1% 20.2832 0.1088 0.0092** respondent self reports poor health 6.3% 21.4016 0.1607 0.0001*** presence of children, age 6 in the household 13.4% 20.0097 0.1081 0.9288 presence of children 6 to 17 in the household 35.9% 0.2058 0.0823 0.0124* respondent’s education (reference category: less than high school graduate) high school graduate 30.6% 0.2418 0.1042 0.0203* some college education 23.2% 0.2924 0.1091 0.0073** college graduate or more 27.6% 0.5225 0.0992 0.0001*** respondent age (spline variable) years of age 49.77 20.0108 0.0148 0.4646 years of age over 45 6.87 20.0515 0.0241 0.0328* years of age over 55 2.49 20.1280 0.0236 0.0001*** years of age over 65 0.31 0.1050 0.0452 0.0201* respondents previous years of full-time work experience 25.6 0.0452 0.0041 0.0001*** 1 descriptive statistics are weighted and estimated using rii techniques. 2probit analysis is unweighted and estimated on the pooled sample; standard errors are not corrected for imputation error and the statistical significance of the coefficient estimates may be overestimated. 3* p , .05, ** p , .01, *** p , .001. source: 1995 survey of consumer finances, combined data set, n5 13,655 (2,731 in each implicate). 8 c.p. montalto et al. / financial services review 9 (2000) 1–15 planned retirement age5 a0 1 a1 ln noninvestment income1 a2 ln financial assets 1 a3 ln nonfinancial assets1 a4 ln debt 1 a5 ln ira/keogh 1 a6 ln defined contribution1 a7 defined benefit1 a8 employed spouse/partner 1 a9 household size1 a10 retirement saving goal1 a11 poor health 1 a12 self-employed1 a13 technical1 a14 service1 a15 precision/repair 1 a16 operators1 a17 farming1 a18 life expectancy1 a19 age 1 a20 age-squared1 a21 black non-hispanic1 a22 hispanic1 a23 other 1 a24 high school1 a25 some college1 a26 college graduate 1 a27 selection bias correction variable (2) the financial variables include amounts of noninvestment income, financial assets (excluding ira/keogh and defined contribution values), nonfinancial assets, defined contribution benefits, ira/keogh, and debt. these amounts are measured as the natural logarithm (ln) to reduce heteroskedasticity (unequal variance of the disturbances). an indicator variable is included for ownership of a defined benefit plan. access to resources is measured with an indicator variable for an employed spouse or partner, a continuous variable for household size, and an indicator variable equal to one if retirement is one of the top three household saving goals. higher levels of financial variables, lower levels of debt, and increased access to resources through family members or saving behavior increase the ability to “afford” to retire and are expected to decrease the planned age of retirement. alternatively, employment of a spouse or partner may suggest interdependent decision making regarding the timing of retirement and may increase the planned retirement age. larger household size may also increase the level of resources needed in retirement, thus increasing the planned age of retirement. characteristics of the respondent’s employment are measured with indicator variables for self-reported poor health, and for self-employment of the respondent, and categorical dichotomous variables for respondent’s occupation. the occupation controls are rather crude since the information in the data set only identifies six broad categories of occupation. this information is used in an attempt to control for differences across these occupation categories in the characteristics of jobs held by elderly workers. poor health may reduce the ability to continue working thus lowering the planned retirement age. the effect of health problems on the ability to work may also depend on the type of job one has, the opportunities for accommodating health problems, and the opportunities to switch to less demanding jobs. some of this effect may be picked up by the occupation variables. self-employment may enable one to extend the working life at their own discretion, and is expected to be positively associated with planned retirement age. respondent’s demographic characteristics are measured with linear and quadratic terms for respondent’s current age, and categorical dichotomous variables for race/ethnicity, and for the highest level of educational attainment. the respondent’s perception of life expectancy is measured with a continuous variable. the availability of reduced social security benefits at age 62, and the increase in the benefit level per year that receipt is deferred (up to age 65) is actuarially fair for a person with average life expectancy, and better than fair for someone with longer than average life expectancy. however, for persons whose life expectancy is lower than the average, social security wealth decreases the longer they 9c.p. montalto et al. / financial services review 9 (2000) 1–15 postpone benefits beyond age 62, creating an incentive to begin taking benefits at age 62 rather than later (economic report of the president, 1999). lower life expectancy is thus expected to decrease the planned retirement age, and therefore we expect a positive relationship between life expectancy and planned retirement age. table 2 descriptive statistics and ordinary least squares regression of planned retirement age variable mean percent1 ols regression2 estimate std. error p-value3 planned retirement age (dependent variable) 61.97 intercept 69.3167 6.7182 0.0001*** financial variables/access to resources log(non-investment income) 10.72 0.3885 0.1781 0.0309* log(financial assets excluding ira/keogh and defined contribution) 8.70 20.1797 0.0720 0.0127* log(nonfinancial assets) 10.99 20.1976 0.0833 0.0179* log(debt) 9.24 0.0448 0.0415 0.2812 log(ira/keogh) 3.44 20.0759 0.0357 0.0335* log(defined contribution) 4.28 20.0520 0.0311 0.0954 defined benefit ownership 34.2% 20.7557 0.3411 0.0268* employed spouse/partner 47.3% 0.3024 0.3162 0.3388 household size 3.0 0.0769 0.1168 0.5102 retirement is a saving goal 34.1% 20.4173 0.3179 0.1893 characteristics of employment respondent self-reports poor health 1.0% 0.5543 1.9092 0.7716 respondent is self employed 10.8% 0.3040 0.3985 0.4456 respondent’s occupation (reference category: managerial and professional specialty) technical, sales, administrative support 24.6% 0.4040 0.4192 0.3355 service 9.0% 21.4772 0.7020 0.0354* precision production, craft and repair 12.9% 21.2691 0.6032 0.0355* operators, fabricators, and laborers 20.0% 20.0331 0.5526 0.9522 farming, forestry, and fishing 1.8% 22.0350 1.1511 0.0772 respondent demographic characteristics and perceptions respondent’s life expectancy (years) 80.22 0.0511 0.0163 0.0024** age of respondent 46.09 20.8195 0.2483 0.0010** age of respondent squared 2184.14 0.0116 0.0026 0.0001*** respondent’s race/ethnicity (reference category: white non-hispanic) black non-hispanic 10.7% 22.4825 0.6583 0.0002*** hispanic 5.1% 22.3700 0.8655 0.0066** other races 4.7% 20.9494 0.7092 0.1813 respondent’s education (reference category: less than high school graduate) high school graduate 30.2% 0.7226 0.6793 0.2877 some college education 25.6% 1.3220 0.7007 0.0593 college graduate or more 33.6% 1.7421 0.7366 0.0181* selection bias correction variable 0.28 0.6124 1.1247 0.5861 model f-statistic5 16.0438 (p-value5 0.0001). adjusted r-square ranges from 0.2093 to 0.2194. 1 descriptive statistics are weighted and estimated using rii techniques. 2 regression analysis is unweighted and estimated using rii techniques. 3* p , .05, ** p , .01, *** p , .001. source: 1995 survey of consumer finances (1,607) households in each implicate). 10 c.p. montalto et al. / financial services review 9 (2000) 1–15 4. results and discussion 4.1. probability of working full-time the probit results for the probability of working full-time are consistent with a priori expectations and previous research (table 1). for a 40 year old, white, married respondent, in good health, with no dependent children, a high school diploma, and 22 years of previous full-time work experience, the probability of working full-time is 90%. the probability of currently working full-time increases with education and work experience. for the reference case, the probability of working full-time is only 86% for someone who has not completed high school, and increases to 94% for a college graduate. each additional year of previous full-time work experience increases the probability of currently working full-time by 0.76 percentage points. for the reference case, 25 years of previous full-time work experience increases the probability of currently working full-time to 92.5%. age is inversely related to the probability of currently working full-time within the sample of people 35 to 70 years old, with the most noticeable declines occurring after age 50. for the reference case, the probability of currently working full-time is 92% at age 35, 90% at age 40, 89% at age 45, 82% at age 50, then declines to 73% at age 55, 37% at age 60, 10% at age 65, and 4% at age 70. the probability of currently working full-time is lower for black, non-hispanic respondents than for otherwise similar white, non-hispanic respondents (85% vs. 90% given the characteristics of the reference case). compared to respondents who are married or living with a partner, unmarried respondents living alone, and unmarried female respondents living with at least one other person, are less likely to currently work full-time. self reported poor health of the respondent also reduces the probability of currently working full-time. for the reference case, poor health reduces the probability of currently working full-time to only 46%. 4.2. planned retirement age the retirement age prediction equation explains approximately 21% of the variance, and 13 of the 27 variables are significant at the 0.05 level or better (table 2). the levels of financial assets (excluding ira/keogh and defined contribution values), nonfinancial assets, and other private pension funds significantly lower the planned retirement age. levels of financial assets and nonfinancial assets lower the planned retirement age relatively more than levels of ira/keogh accounts or defined-contribution pension plans. ownership of a defined-benefit pension significantly decreases the planned retirement age. because the dependent variable is planned retirement age, the coefficients of dummy variables can be interpreted as the effect on planned retirement age, holding all other variables constant. for instance, all other things equal, those who have a defined benefit pension have a predicted retirement age 0.76 years lower than otherwise similar households without a defined benefit pension. planned retirement age does not vary much across the six broad occupation categories. more detailed information on occupation may be necessary to accurately measure this effect. 11c.p. montalto et al. / financial services review 9 (2000) 1–15 being employed in less-skilled occupations (service; precision production, craft and repair) relative to managerial and professional specialty occupations decreases the planned retirement age by less than 1.5 years. being black non-hispanic, or hispanic relative to white non-hispanic decreases the planned retirement age significantly. the planned retirement ages for black non-hispanic respondents and hispanic respondents are 2.4 and 2.2 years lower, respectively, than that of otherwise similar white non-hispanic respondents. planned retirement age increases with noninvestment income, anticipated life expectancy of the householder, the combined effect of linear and quadratic age variables, and education. the selection bias correction variable is not statistically significant. the effect of noninvestment income, although significant, is generally small. for instance, at the mean values of other variables, the predicted effect of income increasing from $10,000 per year to $50,000 per year is a 0.63 year increase in planned retirement age; but an increase from $50,000 per year to $100,000 per year is only a 0.27 year increase in planned retirement age. planned retirement age of respondents who have graduated from college is 1.7 years higher than that of otherwise similar respondents who have not finished high school. fig. 2 shows the effect of current age on planned retirement age. at the mean value of other variables, the effect of increasing current age from 35 to 45 is a 1.08 year increase in planned retirement age, while an increase from 45 to 55 is a 3.41 year increase, and an increase from 55 to 65 is a 5.72 year increase. 5. summary and implications 5.1. summary the regression results suggest that financial preparation for retirement, as well as demographic characteristics and perceptions, including current age and anticipated life expectfig. 2. predicted planned retirement age by current age. 12 c.p. montalto et al. / financial services review 9 (2000) 1–15 ancy, strongly affect planned retirement age. the results also suggest that adjustments to planned retirement age take place over time. these adjustments may be in response to the realization that accumulated resources are not adequate to meet needs in retirement, thus causing workers to postpone retirement. an alternate explanation is that a generational change may result in lower planned retirement ages for younger cohorts of workers since norms regarding retirement age and financial instruments used to save for retirement have changed over time. these differences may produce systematic differences between younger cohorts and older cohorts of workers in the age at which they plan to retire. either explanation has implications for proposed changes in government policy. 5.2. implications for financial planning as fig. 1 demonstrates, there is a wide range of planned retirement ages. the effect of current age on planned retirement age (fig. 2) suggests that some workers’ plans are not achieved, or that there is a generational change in planned retirement ages. financial planners should try to assess the likelihood that a client’s planned retirement age can be achieved, both in terms of the risks of loss of a high income job before the planned retirement age and the chance that the client will have to work longer than planned. if there is a generational reduction in planned retirement ages, financial planning for retirement will become more challenging and perhaps further increase the need for financial professionals to assist workers. 5.3. implications for public policy and future research the ability of workers to adapt to further increases in social security retirement age depends on their capacity to extend their working lives and to accumulate enough savings to offset a delay or reduction in social security income. alternatively, workers face retirement with reduced income. adequacy of retirement resources is influenced by family income and wealth. policies to increase pension coverage as well as private savings would help counter the negative effects of a decrease in social security income. the ability to extend the working life is influenced by health status as well as characteristics of the job. some workers will have difficulty extending their working lives. important questions to address with additional research include: will employers be willing to retain and/or hire older workers? what will happen to older men and women who are not healthy enough to work full-time or who are unable to find jobs? what will happen to older workers in physically demanding jobs? the relationship between age and planned retirement age should be further explored with different data sets, including the 1998 survey of consumer finances, in order to test the generational explanation of the positive relationship between age and planned retirement age. there are implications for proposed changes in government policy whether workers adjust their planned retirement age over time, or more recent cohorts of workers plan to retire at younger ages than earlier cohorts. this study focuses on financial, employment and demographic characteristics as determinants of planned retirement age. however, attitudinal and psychological factors might also 13c.p. montalto et al. / financial services review 9 (2000) 1–15 affect planned retirement age. in reality, individual responses to work and retirement incentives often vary substantially even among persons who appear to have much in common in terms of background characteristics and financial circumstances. thus, unobserved, unmeasured individual differences might play an important role in retirement decisions. a comprehensive theory of work and retirement should be able to explain the substantial variations in retirement decisions that are observed among apparently similar individuals (leonesio, 1996). research that improves our understanding of factors related to planned retirement age will improve our ability to analyze policy issues, including proposed changes to the social security program, as well as provide better insight into how to influence individual behavior related to planning and saving for retirement. references apfel, k. s. (1998, february 26). testimony before house ways and means subcommittee on social security by social security commissioner kenneth s. apfel. bernheim, b. d. (1996).the merrill lynch baby boom retirement index: update ‘96.stanford university, merrill lynch & co., inc. boskin, m. j. (1977). social security and retirement decisions.economic inquiry, 15(1), 1–25. bovbjerg, b. d. (1998, july 15). social security reform. raising the retirement age improves program solvency but may cause hardship for some. government accounting office, t-hehs-98–207. burtless, g. (1986). social security, unanticipated benefit increases, and the timing of retirement.review of economic studies, 53(5), 781–805. burtless, g., & moffitt, r. a. (1985). the joint choice of retirement age and postretirement hours of work. journal of labor economics, 3(2), 209–236. diamond, p. a., & hausman, j. a. (1984). individual retirement and savings behavior.journal of public economics, 23,81–114. economic report of the president.(1999). washington, d.c.: u.s. government printing office. employee benefit research institute. (1999). retirement patterns and bridge jobs in the 1990s.ebri issue brief no. 206.author. fields, g. s., & mitchell, o. s. (1984). economic determinants of the optimal retirement age: an empirical investigation.journal of human resources, 19(2), 245–262. gustman, a. l., mitchell, o. s., & steinmeier, t. l. (1995). retirement measures in the health and retirement study.journal of human resources, 30(1995 supplement), s57–s83. heckman, j. (1979). sample selection bias as a specification error.econometrica, 47,153–161. honig, m. (1996). retirement expectations: differences by race, ethnicity, and gender.the gerontologist, 36(3), 373–382. hurd, m. d. (1997). the economics of individual aging. in m. r. rosenzweig & o. stark (eds.),handbook of population and family economics,volume 1b (pp. 891–966). new york: new holland. kennickell, a. b., starr-mccluer, m., & sunde´n, a. e. (1997). family finances in the u.s.: recent evidence from the survey of consumer finances.federal reserve bulletin, 83(1), 1–24. kotlikoff, l. j. (1979). testing the theory of social security and life cycle accumulation.the american economic review, 69(3), 396–410. leonesio, m. v. (1996). the economics of retirement: a nontechnical guide.social security bulletin, 59(4), 29–50. mitchell, o. s., & moore, j. f. (1997).retirement wealth accumulation and decumulation: new developments and outstanding opportunities.the wharton financial institutions center, university of pennsylvania, working paper 97–12. 14 c.p. montalto et al. / financial services review 9 (2000) 1–15 mitchell, o. s., & quinn, j. f. (1995). final report to the technical panel on trends and issues in retirement savings. washington, d.c.: 1994–95 advisory council on social security. montalto, c. p., & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances.financial counseling and planning, 7,133–146. quinn, j. f. (1977). microeconomic determinants of early retirement: a cross-sectional view of white married men.the journal of human resources, 12(3), 329–346. retirement confidence survey. (1999). changing nature of retirement: fact sheet. [www document]. url http://www.asec.org/rcshm.htm rubin, d. b. (1987).multiple imputation for nonresponse in surveys.new york: john wiley & sons. samwick, a. a. (1998). new evidence on pensions, social security, and the timing of retirement.journal of public economics, 70(2), 207–236. sas institute inc. (1990). sas/stat© user’s guide, version 6, fourth edition, volume 2. cary, nc: sas institute inc. sickles, r. c., & taubman, p. (1986). an analysis of the health and retirement status of the elderly.econometrica, 54, 1339–1356. suits, d. b., mason, a., & chan, l. (1978). spline functions fitted by standard regression methods.the review of economics and statistics, 60(1), 132–139. uccello, c. e. (1998).factors influencing retirement: their implications for raising retirement age.(aarp public policy institute, report no. 9810). washington, d.c.: american association of retired persons. u. s. bureau of the census. (1998). population profile of the united states 1997.current population reports, p23–194. washington, d.c.: u.s. government printing office. yuh, y., hanna, s., & montalto, c. p. (1998). mean and pessimistic projections of retirement adequacy. financial services review, 7(3), 175–193. yuh, y., montalto, c. p., & hanna, s. (1998). are americans prepared for retirement?financial counseling and planning, 9(1), 1–12. 15c.p. montalto et al. / financial services review 9 (2000) 1–15 pii: s1057-0810(00)00044-5 an integrated model for financial planning natalie chieffe*, ganas k. rakes finance department, ohio university, college of business, athens, oh 45701, usa abstract financial planning is a broad subject that requires an integrating overview. the model for financial planning incorporates the time and the expected nature of financial events. the categories of the model include 1) money management issues that the individual faces as short-term expected events, 2) issues of meeting unexpected financial events through an emergency fund and insurance, 3) investing to reach the individual’s intermediate and long-term goals, 4) transference planning and other long-term issues whose time frame is unknown. the model has applications for “do it yourself” investors, financial planners, and students. the framework successfully integrates the broad range of topics typically covered in financial planning and personal finance courses. © 1999 elsevier science inc. all rights reserved. jel classification:a22, g29 keywords:financial planning; financial education 1. introduction financial planning and personal finance are topics of growing student interest and of concern for many adults. the field is well supplied with books and articles in publications ranging from the popular press to academic journals. for example, eaton (1993) discusses saving for retirement, fevurly (1991) describes how to save for children’s college education, gray (1993) illustrates the effect of inflation on savings, clements (1993a,b) explains the importance of diversification. except for textbooks and general financial guides, most of these sources deal with a single aspect, product or problem in the area of financial planning. * corresponding author. tel.:11-740-593-9320; fax:11-740-593-9539. e-mail address:chieffe@ohio.edu (n. chieffe). financial services review 8 (1999) 261–268 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00044-5 the general topic of financial planning and personal finance is so broad and has so many subtopics that it is very interesting to teach. but it is difficult to keep the material organized so that the relative importance of each area is clear. this paper presents a process or model that provides an integrating overview of financial planning that could be useful to furnish a focus that cuts across topics and products. the approach outlined below was used in financial planning classes at two universities and in seminars for working adults. the typical reaction from the adults is, “this makes sense. it is so clear. why didn’t i think of it myself?” the approach has been useful in organizing the college course syllabus. the students in these classes have stated that the model helped them see how the topics relate to each other. the model also has applications to help “do it yourself” investors and financial planners. the framework integrates the broad range of topics and products typically covered in financial planning and personal finance courses. it is introduced to students at the beginning of the course as a part of the overview of financial planning. the model is also used in principles of insurance and principals of investments classes to help students perceive the role of the individual course material in the broader context of financial planning. thus, this integrated model for financial planning has proven to be a useful contribution to the teaching and learning process in these related areas. 2. the model the model, as shown in table 1, is a 23 2 matrix that symbolizes the area positioning for the various topics in financial planning and personal finance. each topic that is discussed table 1 the financial planning model current period future period planned financial events money management investing for goals budgetinga investment planninga income (education planninga living expenses and other lt or it goals): savings stocks and bonds credit mutual funds real estate income tax planninga: retirement planninga: gifts pension funds taxes iras & annuities 401ks and 403bs unplanned financial events emergency planning transference planning risk managementa: estate planninga: emergency fund wills line of credit tax planning insurance: trusts property life insurance health business agreements liability charitable bequests a major areas as defined by the cfp board. 262 n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 in a financial planning class or presented in a financial plan can be positioned in one of the four categories of the model. the first element to be considered is time. some financial events occur in the current period, others are expected in a future period. the length of the current period can be equivalent to the length of time between regularly scheduled payments from wages or other sources. alternately, the current period can be defined as the foreseeable period of a year or two. the future period includes financial events occurring beyond the current period’s budget cycle. these events are defined by the individual’s goals for the intermediate term (one to five years from the present) or for the long term (five years or more from the present). the second element to be considered is whether or not the occurrence of or the date of the financial event can be foreseen. some events, such as the due date on the next tax bill or the date a child will begin college, are known with some accuracy. others, such as medical bills or the date of one’s death, cannot be predicted. these unexpected expenses can be met through appropriate planning. 2.1. money management the first category of the matrix (table 1) concernsmoney managementissues that the individual needs to handle as short-term expected events. in the current period the individual generally knows what income and expenses are expected. individuals who need to concentrate on this area of financial planning are in the early part of their careers or are those whose living situation has recently changed. cash management, including budgeting and planned saving, may be of immediate importance to them. college students also hope to acquire good cash management skills prior to full-time employment. they are interested in learning how to obtain the greatest benefit from the money they earn. the money management category feeds each of the other categories and is referenced at many points of the planning process. financial planners recommend the use of a budget so that available funds are directed to the most important goals. however, davis and carr (1992) show that while most households have a budget, only a small percentage have a written budget. most of the financial planning textbooks provide a budget method and software which the students can use to prepare a personal budget. while the students perceive this to be painful at first, the budget is the primary vehicle for money management. it is also a way to ensure that the available funds are directed to the most important goals. it is usually enlightening for a student to track expenses for a month and then compare actual expenditures to planned expenditures. at that point the class is usually ready to discuss methods to ensure a surplus rather than a deficit of cash. most americans find it difficult to save money. according to the “flow of funds accounts” in thestatistical abstract of the u.s. 1998, personal savings as a percentage of disposable personal income fell from 12.1% in 1980 to 7.4% in 1996. it should become clear to the students that at any income level one needs a plan in order to manage expenses and save some money. unless the instructor is successful in convincing students of the importance of budgeting, a careful study of financial planning is unlikely to have lasting benefits. the students should also discuss how their financial circumstances will change through time and how the selection of alternatives at each stage of their life will affect their standard of living. 263n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 at this point it is appropriate to discuss the use of credit. most college students are well aware of the availability of credit since they receive frequent offers from credit card companies. they also are concerned with credit management. many of them have friends who are deeply in debt because they use credit cards too freely. others are concerned about how long it will take to repay their student loans. one appropriate exercise is to use time-value-of-money techniques from the introductory finance class and calculate the true cost of using credit for various items. this leads to a discussion of reasons for using credit and how to balance short-term consumption needs with the long-term cost. we also calculate a hypothetical budget for their first year of employment. the conflict they will face becomes obvious—they want to purchase new household goods, a car, and clothing while trying to pay down the debt they incurred as students. 2.2. emergency planning however, regardless of how well a person has planned, in the short term the individual may also meet some unexpected events. the second category,emergency planning, involves these issues. financial planners generally recommend that individuals accumulate emergency funds of about three to six months of expenses. these funds are kept in a savings account, money market fund, or short-term certificates of deposit. however, very few households actually meet the 3-month guideline. chang and huston (1995) and huston and chang (1997) show that most households do not have the recommended levels of liquid assets. chang, hanna, and fan (1997) suggest that this behavior may be rational for those households who do not expect a decline in real income. many individuals do not feel that accumulating funds for emergencies is as important as accumulating funds for other goals. an alternative strategy that we have recommended to clients and to students is to keep an open line of credit either with a major credit card or a home equity line of credit. this technique bypasses the opportunity cost of keeping funds in low-interest bearing accounts or of foregoing current consumption to accumulate liquid assets. of course, the individual needs a certain amount of self-discipline to reserve the open line of credit for a true emergency. other types of emergencies are more readily addressed with health and property insurance. individual clients who are in the early stages of their careers are often concerned with property and health insurance coverage. these concerns arise for other financial planning clients only when they’ve experienced a major lifestyle change. most college students are familiar with health and automobile insurance. many of them are covered under their parents’ policies and are aware that this coverage may end when they graduate. they also know that this type of coverage can be critical and are interested in learning about the different types of health policies that may be available through their future employers. students in our personal financial planning classes usually know that automobile insurance is necessary to conform to the law in most states and can save a person from financial distress. many of them express dissatisfaction with the premiums they now pay. but they are unaware of the differences in types of policies and that they should shop for the best policy. many of them, however, have never considered the need for homeowners or tenants insurance. they know their parents have coverage and feel that they themselves don’t own 264 n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 enough property to insure. after we discuss the facets of homeowners policies some of them reconsider. the funds for the emergency savings account and for insurance premiums are appropriated through the budgeting process in the money management category. sometimes it may be necessary to reduce expenditures for other discretionary items in order to obtain adequate funds for these items or to maintain the open line of credit. many students in our classes graduate with significant financial obligations incurred during their college years. they reveal that they have accrued large amounts of credit card and student loan debt. they will need to plan to meet this existing debt and to prepare for unexpected financial problems. it may take efforts for some years after graduation to reach a financial position where they may begin effective wealth accumulation. 2.3. investing for goals the third category,investing for goals, includes the intermediate-term and long-term goals which the individual expects to reach. individuals with ill-defined or undefined goals may never have the incentive to invest for the long-term or may be more easily tempted by short-term investment techniques. to help students and clients define their goals we ask them to consider the question, “if you could have anything you wanted, what would that be?” this question guarantees a smiling audience! then they answer more questions about this dream, “when would you like to plan to have this? how much would it cost to have it today? is there any other goal which is more important to you than this one?” after each person has prioritized the goals listed, we begin to analyze each one for the present value of the cost, the estimated time to reach the goal and the degree of risk acceptable in reaching the goal. the goals are as numerous and various as the students and clients. some choose retirement as the most important goal; others choose to save for the down-payment on their first home or for their children’s college education. time-value-of-money concepts are used to determine the amount of monthly investment that is needed. we show the students how to choose appropriate investment vehicles based on their risk aversion tendencies, investment horizon and position in the life cycle (see modigliani and ando, 1963). we discuss why certain categories of investments are more appropriate for certain goals (see butler and domian, 1993, for simulations which illustrate the probable returns for various investment horizons). some attention is devoted to employer pension fund analysis and tax-sheltered investment opportunities. the changing rules on vesting, portability, ira’s and concerns about the social security system require attention. taking these factors into account, appropriate investments can be matched to each goal. hensel, ezra and ilkiw (1991) show that the type of investment is much more important than the choice of specific investment within the category. the effects of inflation and investment returns are included in the analysis. funds for investing for goalsare allocated through the budget in themoney managementcategory and are directed to the appropriate investments. 265n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 2.4. transference planning the fourth category is fortransference planningand other long-term issues whose time frame is unknown. this encompasses estate planning, business continuation plans, life insurance and various types of trusts for the care of dependents, the disposal of assets and techniques to minimize estate taxes. this category evolves over the individual’s lifetime as certain continuance and transference needs increase. most young adults do not need to direct substantial funds to this area until they have accumulated assets and/or have dependents. many of our students tell us that they have life insurance policies which were purchased by their parents. others are quite averse to owning life insurance. individuals with no dependents need only enough life insurance to cover their final expenses and debts in excess of their current assets. individuals with dependents are often interested in providing income for survivors. 3. applications table 2 indicates the textual coverage of the content included within the financial planning model. the referenced texts were selected because of their wide acceptance in university level personal finance and financial planning courses. other books devoted to the area would likely work equally well. table 2 chapters from selected textbooks which fit into the financial planning model textbooka money management investing for goals gf 2, 3, 4, 5, 6, 7, 8 9, 13, 14, 15, 16, 17, 18 gj 2, 3, 4, 6, 7 5, 11, 12, 13, 14 hpr 4, 5, 6, 7, 11, 12 13, 14, 15, 16, 17 k 4, 5, 6, 7, 8 12, 13, 14, 15, 16, 17 rmc 2, 3, 4, 5, 6, 7 12, 13, 14, 15, 16 rmac 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 15, 16, 17, 18 wf 2, 3, 4, 5, 16 9, 10, 11, 12, 13, 14, 15, 17 emergency planning transference planning gf 10, 11 12, 19 gj 9, 10 8, 15 hpr 8, 9, 10 k 10, 11 18 rmc 8, 10, 11 9, 17 rmac 12, 13, 14 19 wf 7, 8 6, 18 a gf, garman, e.t. and forgue, r.e.personal finance, 6th ed.; gj, gitman, l.d. and joehnk, m.d.personal financial planning, 8th ed.; hpr, ho, k., perdue, g. and robinson, c.personal financial planning, the u.s. edition; k, keown, a.j.personal finance: turning money into wealth; rmac, ramaglia, j.a. and macdonald, d.b. personal financial management; rmc, rejda, g.e. and mcnamara, m.j.personal financial planning, 1st ed.; wf, winger, b.j. and frasca, r.r.personal finance: an integrated approach, 5th ed. 266 n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 in addition to using the model as an organizational device to integrate material and provide a focus for students, the framework works well for specific student assignments in a problem-based approach. some instructors choose to include case studies in the financial planning course. table 3 shows how the financial planning model relates to these project assignments. the specific assignments require relatively little information from the instructor. typically, data concerning current assets, liabilities, income, ages and size of family are all that is required. additional details can enrich the experience. for example, providing a simulated 1040 for the most recent year is a vehicle to add exploration of income tax aspects of financial planning. 4. conclusions the complete financial planning model outlined in this paper divides all of the content material into four general categories to provide a simple but comprehensive overview of the entire area. essentially, personal financial activities are either current period or future period events and involve circumstances that can be expected and therefore planned or circumstances that are unexpected and thus create emergencies. using this overview of the material, students can continually place seemingly unrelated material in sufficient context to retain a useful perspective of the complete course. it is possible for an instructor to cover the material in any order or format that seems most successful for them and still have an integrating device in the background for occasional reference. likewise, any text book can be used with this financial planning model as a tool to relate different parts to the total course. as previously mentioned, this model has been successfully used by the authors in several contexts. table 3 case studies for the financial planning model stage in life possible goals single, newly employed money management and emergency planning: emergency fund debt reduction married, no children investing for goals: tax planning house down payment vacation retirement married, children money management and investing for goals and transference planning: education retirement estate planning near retirement investing for goals and transference planning: retirement income estate planning 267n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 the model also works well for financial planning clients who have already moved through some stages in their financial life cycles. the planning goals which are defined early in the client-planner relationship can be inserted into the appropriate categories. these individuals appreciate the organization of the financial planning model. they can see how their current and projected financial situations compare. acknowledgment the authors appreciate the suggestions of professor jill lynn vihtelic of st. mary’s college, notre dame, in. references butler, k. c. & domian, d. l. (1993). long-run returns on stock and bond portfolios: implications for retirement planning.financial services review, 2(1), 41–49. chang, y. r., hanna, s. & fan, j. x. (1997). emergency fund levels: is household behavior rational?financial counseling and planning, 8(1), 47–55. chang, y. r. & huston, s. (1995). patterns of adequate household emergency fund holdings: a comparison of households in 1983 and 1986.financial counseling and planning, 6,119–128. clements, j. (1993a). recipe for successful investing: first mix assets well.the wall street journal, (oct. 6), c1. clements, j. (1993b). three steps to becoming a successful investor.the wall street journal, (june 11), c1. davis, e. p. & carr, r. a. (1992). budgeting practices over the life cycle.financial counseling and planning, 3, 3–16. eaton, l. (1993). cloudy sunset: a grim surprise awaits future retirees.barron’s, (july 12):8–9. fevurly, k. r. (1991). personal financial planning: how to provide for the cost of a college education.journal of accountancy, february,81–87. garman, e. t. & forgue, r. e. (2000).personal finance(6th ed.). boston: houghton mifflin company. gitman, l. j. & joehnk, m. d. (1999).personal financial planning(8th ed.). fortworth, tx: dryden press. gray, w. s. (1993). historical returns, inflation and future return expectations.financial analysts journal, 49(4), 35–45. hensel, c. r., ezra, d. d., & ilkiw, j. h. (1991). the importance of the asset allocation decision.financial analysts journal, 47(4), 65–72. ho, k., perdue, g., & robinson, c. (1999).personal financial planning, the u.s. edition. north york, on: captus press inc. huston, s. j. & chang, y. r. (1997). adequate emergency fund holdings and household type.financial counseling and planning, 8(1), 37–46. kapoor, j. r., dlabay, l. r. & hughes, r. j. (1999).personal finance and personal financial planner package, 5th ed. boston: irwin-mcgraw-hill. keown, a. j. (1998).personal finance: turning money into wealth.upper saddle river, nj: prentice-hall, inc. modigliani, f. & ando, a. (1963). the life cycle hypothesis of saving: aggregate implications and tests.the american economic review, 53, 55–84. ramaglia, j. a. & macdonald, d. b. (1999).personal financial management. cincinnati: south-western college publishing. rejda, g. e. & mcnamara, m. j. (1998).personal financial planning, 1st ed. reading, ma: addison wesley longman. rosefsky, r. s. (1996).personal finance,6th ed. new york: john wiley. u.s. department of commerce. (1998).statistical abstract of the u.s. winger, b. j. & frasca, r. r. (1996).personal finance: an integrated approach, 5th ed. upper saddle river, nj: prentice hall. 268 n. chieffe, g.k. rakes / financial services review 8 (1999) 261–268 pii: s1057-0810(96)90024-4 financial services review, 5(l): 13-29 copyright q 1996 by jai press inc. issn: 1057-08 10 all rights of reproduction in any form reserved. professional stock analysts’ recommendations: implications for individual investors m. mark walker gay b. hatfield conclusions regarding analyst performance often depend on the evaluation technique employed. using a wide variety of techniques, wefind that although there is some evi dence that analysts do have the ability to identify undervalued and overvalued securities, individual investors generally experience inferior portfolio performance by following analyst recommendations published in the ‘market highlights” section of usa today (even before transaction costs are inch&d). as a result, individual investors should view studies that purport to show superior performance with skepticism. this statement ispar titularly true when the assertions are based on stock index comparisons. an individual investor’s search for an optimal portfolio typically consists of two key deci sions: (a) how to allocate funds across asset classes, and (b) which securities to purchase within each class. to aid investors with the security selection decision, many brokerage firms offer investment advice. professional stock analysts, for example, regularly issue buy, sell, and hold recommendations. the interesting question that arises is as follows: if an investor had acted on a particular recommendation, how would his investment have per formed over the following calendar year? would a recommended buy have provided excess returns? would a recommended sell have resulted in a loss if the investor had not acted on the negative recommendation? despite the considerable amount of time and effort that brokerage fums devote to fun damental and technical analysis, many academicians and investment practitioners question the notion that “wall street” research can be used to enhance portfolio performance. in one up on wall street, for example, peter lynch (1989) argues that professional analysts generally miss the best investment opportunities because they tend to issue buy recommen dations after a firm’s stock price has risen dramatically (lynch refers to this situation as “street lag.“) lynch believes that street lag occurs because many analysts follow only those stocks which have attracted the attention d large institutional investors. by that time, how m. mark walker and gay b. hatfield l department of economics and finance, 301 conner hall, university of mississippi, university, ms 38677. 14 financial services review 5( 1) 1996 ever, lynch generally finds that a firm’s stock price already reflects most of the good news about the company. in fact, lynch (1989) believes that a large increase in institutional ownership, coupled with favorable comments by professional analysts, often represents an opportune time to sell. dorfman (1993, april) reaches a similar conclusion. he cites evi dence that, of the 12 stocks that were listed in a wall street joumaz article (on march 6, 1992) as being the most popular among 200 money managers, 9 declined in price over the next 12 months (the s&p 500 index rose approximately 10% during this period). finally, most academicians seem to believe that markets are at least semi-strong form efficient, and that it is difficult for investors (even professional stock analysts) to continually identify undervalued securities (e.g., see roll, 1994). to test the academicians’ hypothesis, the wall street journal has been conducting a series of six-month contests that compare the investment performance of professional stock analysts to: (a) the dow jones industrial average (djia) and (b) a group of randomly selected stocks (the dartboard portfolio). the results for all contests show the professional analysts winning 3 1 times while the djia has won 26 times.’ the average six-month gain is 8.1% for the professional analysts and 3.8% for the djia, but the test does not examine risk-adjusted returns. moreover, each analyst is limited to one stock, and returns include capital gains and losses but not dividends. the purpose of this study is two-fold. first, this study reviews the techniques that are commonly used to evaluate investment performance: stock index comparisons, event study methodologies, and the sharpe, treynor, and jensen measures. second, this study exam ines the question raised earlier; that is, if an investor followed the professional analysts’ recommendations, what would the result have been over a period of time? when a profes sional analyst makes a buy or sell recommendation, his standard time horizon for perfor mance is six months to a year; therefore, this time frame has been adopted in this study. the results of this study indicate that the conclusion one draws regarding analyst per formance depends critically on the evaluation technique employed. stock index compari sons, for example, often produce biased results because the recommended securities exhibit different risk than the benchmark selected. similarly, event study methodologies often produce biased results because the recommended securities exhibit abnormal perfor mance during the estimation period selected. based on an analysis of each technique, we believe that the sharpe, treynor, and jensen methodologies provide the most defensible results. while these methodologies generally have not been used to test analyst perfor mance, they have been used to evaluate portfolio managers. the results of this study also indicate that while analysts do identify mispriced securities, it is difficult for individual investors to capitalize on investment recommendations. when transaction costs are included, investors generally earn normal returns even when trades are assumed to occur at pre-recommendation prices. investors who purchase securities following recommendation announcements generally earn lower returns than investors who trade at pre-recommendation prices. this result is consistent not only with the announcement effect documented in previous studies, but also with market efficiency in the semi-strong form. ii. background a market is said to be efficient if security prices reflect all available information, and infor mation is freely and quickly disseminated in an unbiased manner (fama, 1970). financial hfessional stock anazysts ’ recommendations 15 theorists have identified three conditions that make a market efficient: (a) there are a large number of profit-maximizing investors; (b) transaction costs are insignificant; and (c) investors have free and equal access to all relevant information. while most observers believe that the u.s. stock markets are semi-strong form efficient (i.e., security prices adjust rapidly to reflect publicly-available information), financial research indicates that security prices do not reflect private information.* as a result, fundamental and technical analysis may be a worthwhile endeavor. investors who are more adept at analyzing and interpreting publicly-available information, and investors who can uncover nonpublic information, should earn positive risk-adjusted returns. grossman and stiglitz (1980) point out, however, that investors will search for new information only if the cost and effort pro duces higher investment retums.3 previous studies that examine analyst recommendations generally fall into one of two categories: (a) studies that test accuracy, and (b) studies that test information content. in the early 1980s several researchers tested the accuracy of value line investment survey time liness rankings (value line analysts forecast stock price performance over a twelve-month period and rank stocks from 1 (outperform) to 5 (underperform)). the results of these stud ies were mixed. copeland and mayers (1982), for example, found no evidence that inves tors who followed an active trading strategy earned positive abnormal returns by investing in stocks with a particular value line ranking. while a portfolio of rank 5 stocks did expe rience statistically significant abnormal returns of approximately -3% (based on a 26week holding period), copeland and mayers (1982) argued that the cost of implementing a short sale trading rule would offset the gain. holloway (1981), on the other hand, concluded that investors who bought and held value line rank 1 stocks did outperform the market. holloway (1981) also tested an active trading strategy that involved rebalancing the portfolio weekly to reflect ranking changes, but the results were significant only when transaction costs were ignored. in a later study holloway (1983) found that an active trading strategy did result in significant abnormal returns even when transaction costs were included. while holloway (1983) used friday’s closing prices to calculate returns (value line recommendations were published on fri day), he indicated that timing was critical. when returns were calculated using prices for the following monday, the returns for the rank 1 portfolio were significantly lower. more recently, financial researchers have used an event study methodology to test the information content of analysts’ recommendations. numerous studies (givoly & lakon ishok, 1979; groth, lewellen, schlarbaum, & lease, 1979; bjerring, lakonishok, & ver maelen, 1983; liu, smith, & syed, 1990; barber & loeffler, 1993) have documented positive abnormal returns around the announcement of analysts’ recommendations.4 liu et al. (1990). for example, analyzed the reaction of stock prices to security recommendations listed in the “heard on the street” column in the wall street journul. based on the finding that investors earned positive cumulative abnormal returns (cars) of approximately 3.4% over a 21-trading day period centered around the announcement date, they concluded that analyst recommendations convey new information to the market (the information hypoth esis). barber and loeffler (1993) analyzed recommendations listed in the monthly “dart board” column of the wall street journuz. while the professional analyst stock picks earned cars of approximately 4% on the publication date, the cars were partially reversed over the next 25 days, which caused barber and loeffler to conclude that invest ment recommendations have both an information and a price pressure effect. the price 16 financial services review 5( 1) 1996 pressure hypothesis suggests that the abnormal returns associated with investment recom mendations are caused primarily by the actions of naive investors. ijl data this study examines professional analysts’ recommendations announced in the “market highlights” section of usa today’ that pertain to firms listed on either the new york stock exchange or the american stock exchange. the study period is january 1988 to december 1990. return data for individual securities and the market were taken from the crsp (cen ter for research in security prices) tapes. the initial sample contained 374 investment rec ommendations, but 24 recommendations were excluded due to insufficient price data during the estimation periods, and 21 observations were excluded because the time span between conflicting recommendations was less than 125 trading days (approximately 6 months). as a result, the final sample contains 329 recommendations.6 the following example illustrates the issues related to conflicting recommendations. on october 30, 1990, prudential bathe issued a sell recommendation for armstrong world industries. ten trading days later (on november 13, 1990) smith barney issued a buy recommendation. if the focus of this study were to analyze the recommendations of a particular brokerage firm, then armstrong world industries would remain in the “sell” subsample until prudential bathe upgraded its opinion of the stock.7 instead, this study assumes that the smith barney “buy” recommendation negates the prudential bathe “sell” recommendation. moreover, because the time period between the two recommendations is short (less than 125 trading days), the prudential bathe observation is excluded from the sample. the focus of this study is to test the accuracy of “wall street” research over a rel atively long time horizon rather than to isolate the short-term impact of recommendation announcements on stock prices. during a telephone interview, a usa today employee indicated that only stocks whose price had been affected by the recommendation would actually be included in the newspaper article.* because the objective of this study is to evaluate the performance of brokerage house investment recommendations, we want a sample that contains only those recommendations that changed investors’ expectations. we acknowledge, how ever, that our results and conclusions may apply only to this subset of analysts’ recom mendations. it it important to note that investment recommendations usually are disclosed to the brokerage firm’s institutional and retail clients prior to publication in usa today. a usa today employee indicated, however, that the time span between disclosure to clients and publication in usa today is relatively short (i.e., 1 to 3 days). peter lynch (1989) argues that investors are least likely to earn abnormal returns by following recommendations on stocks that have attracted the attention of large institutional investors. because 238 recommendations (70% of the total sample) involve firms that are included in the s&p 500 index, one would expect the stock prices of the sample firms to be particularly efficient. table 1 shows the number of analyst recommendations by year. two hundred and forty-five recommendations (75%) are positive (strong buy, buy, or reiterate buy), and 84 recommendations (25%) are negative (sell or a change from buy to hold). the data provide rofessional stock analysts’ recowmentiations 17 table 1 brokerage firm investment recommendations announced in usa to&y (1988-1990) positive recommendations negative recommendations strong reiterate buy to market buy’ buy+ buy hold selp total retwd 1988 3 20 8 7 38 12.4% 1989 8 36 37 16 9 106 21.2 1990 i 66 70 35 20 198 -6.6 total 18 116 111 57 27 329 % of total 5.5% 35.3% 33.7% 17.3% 8.2% 100% notes: l includes changes from buy to strong buy, moderately attractive to very attractive, and recommendations that reiterate a strong or aggressive buy. ’ includes changes from neutral to above average, and stocks placed on the firm’s &ommended list. * includes changes from above average to average, and stocks dropped from the firm’s recommended list. o includes recommendations that reiterate a sell. 1 the 1%month return on the s&p 500 stock index. support for the belief that analysts have a predilection for making positive recommenda tions to avoid offending current or potential investment banking clients. the percentage of negative recommendations, which increases between 1988 and 1990, coincides with the decrease in the stock market (as measured by the s&p 500 index). in 1988,18% of the recommendations were negative, but in 1990,28% of the recommen dations were negative. given the industry’s aversion to issuing negative recommendations, one might expect negative recommendations to be more accurate than positive recommen dations. this study tests this hypothesis and examines the sensitivity of the results to dif ferent holding periods. academicians who analyze analyst performance generally use an event study method ology. studies published in the popular press, on the other hand, often utilize stock index comparisons. the following section critiques the most commonly used methodologies for evaluating analyst performance. iv. evaluating portfolio performance a. stock index comparisons while financial theory indicates that investment performance should be evaluated using risk-adjusted returns, numerous benchmarks have been applied in practice. dorfman (1994), for example, reports the results of a study by zacks investment research, inc. that examines the performance of stocks recommended by 16 major brokerage firms during 1993. based on the finding that 12 of the 16 firms’ stock picks outperformed the s&p 500 index, and 9 of the 16 firms stock picks outperformed the wilshire 5000 index (which contains non-s&p 500 stocks), dorfman concludes that the analysts performed extremely well. a more detailed analysis, however, indicates that the choice of an appropriate bench mark is crucial. if one assumes that there is a 50-50 chance that a brokerage firm’s stock recommenda tions will outperform the market in a given year (define this occurrence as a success), then 18 financial services review 5( 1) 1996 the binomial probability model indicates that there is a 3.8% chance of observing 12 or more successes in 16 trials. the low probability strongly suggests that the analysts were able to identify undervalued securities. if the wilshire index is used, on the other hand, the binomial model fails to reject the null hypothesis (i.e., the probability of observing 9 or more successes in 16 trials is 40.4%). the key issue is whether the s&p 500 index or the wilshire index is the appropriate benchmark. dorfman cites evidence that the analysts tend to recommend small (i.e., non-s&p 500) stocks, but dorfman does not conduct the bino mial tests. moreover, because risk-adjusted returns are not examined, the magnitude of the analysts’ relative performance cannot be evaluated. b. event study methodologies the current study employs the following market model to calculate the excess return, or prediction error (fe+ for each firm j at event day t: pe’, = rjt c% + pp,,)* (1) rjl is the rate of return on security j for day t, and r,t is the return on the crsp value weighted index on day t. the coefficients 3 and pj are ordinary least squares estimates of the intercept and slope, respectively, from a pre-event market model regression for days -500 to -25 1.’ day zero (t = 0) is defined as the last trading day before a recommendation is reported in the “market highlights” section of usa today. prediction errors are esti mated over the interval t = -5 days prior to the announcement of the brokerage house invest ment recommendation to t = +250 days after the announcement. the cumulative prediction error (cpe) from day tl to day t, for each recommendationj is: t2 cpej = cpejt. tl (2) cumulative prediction errors are estimated over various intervals. for a sample of n securities, the mean cumulative prediction error (mcpe) is defined as: mcpe = (l/n) 2 cpej. j= 1 (3) the expected value of mcpe is zero in the absence of abnormal performance (i.e., if mcpe equals zero, then one cannot reject the null hypothesis that investors earned normal returns). the test statistic is based on an aggregation of mean standardized cumulative predic tion errors (mscpe) (see appendix). the test statistic for a sample of n securities is: z = i (mscpe,)/fi j= 1 (4) each mscpej, is assumed to be distributed unit normal in the absence of abnormal performance. under this assumption, z is also unit normal. hfessional stock analysts ’ recommcndalions 19 event study methodologies evaluate investment performance by subtracting a secu rity’s expected rate of return from its actual rate of return (see equation 1). when a secu rity’s expected rate of return is estimated using parameters calculated from a preor post event estimation period, the methodology tests only whether a security’s performance dur ing the event period differs from its performance during the estimation period. as a result, any abnormal returns measure relative performance rather than absolute performance. if no abnormal returns are observed during the event period, the researcher can conclude only that the security’s performance did not change relative to the estimation period. it is not possible to rule out, however, the possibility that the security was a good investment during both periods. lo the null hypothesis of no abnormal performance during the estimation period (i.e., $, = 0) is tested for each category p. c. sharpe, treynor, and jensen measures while the sharpe, treynor, and jensen measures are commonly used to evaluate the performance of portfolio managers, these techniques generally have not been used to ana lyze the performance of professional stock analysts. shatpe’s measure examines average excess return per unit of total risk: sharpe = ( rp rf) / cq, where rp rf = the average monthly excess return on a portfolio of stocks with a particu lar analyst recommendation (where rf equals the one-month return on a three-month t bill), and crp = the standard deviation of returns for portfolio p. treynor’s measure, on the other hand, examines average excess return per unit of sys tematic risk: treynor = ( rp rf ) i &, where &, = the beta coefficient for portfolio p. ‘lie sharpe and treynor measures for the market portfolio are calculated in a similar fashion using the crsp value-weighted index (the results using the crsp equal-weighted index are essentially the same). jensen’s measure examines excess return as a function of systematic risk. r pr rjt = ap + pp cr,, rft), where rmt = the monthly return on the crsp value-weighted or equal-weighted index. the coefficients ap and fip are estimated using ols regression. if ap is statistically different from zero, then the null hypothesis of no abnormal performance is rejected. this study calculates sharpe, treynor, and jensen measures for two trading strategies. one strategy assumes that an investor buys a recommended stock at the beginning of the month that contains the publication date (t = 0) and holds the stock for 13 months (i.e., the holding period is t = 0 to t = +12). the second strategy assumes that an investor buys a rec ommended stock at the beginning of the monthfollowing the publication date and holds the stock for 12 months (i.e., the holding period is t = +l to t = +12). as a result, each perfor mance measure is calculated using pre-recommendation prices (prices on which the ana 20 financial services review 5( 1) 1996 table 2 portfolio construction for strong buy category no. company announcement month holding periods strategy i strategy 2 (t=otot=+12) (t = +i tot = +12) 1 walt disney mar 1988 mar 1988-mar 1989 apr 1988-mar 1989 2 clark equip apr 1988 apr 1988-apr 1989 may 1988-apr 1989 3 ford nov 1988 nov 1988-nov 1989 dee 1988-nov 1989 18 georgia gulf dee 1990 dee 1990-dee 1991 jan 1991-dee 1991 lysts’ recommendations are based) and post-recommendation prices (prices at which investors are likely to trade). strong buy, buy, reiterate buy, downgrade to hold, and sell recommendations are examined separately. the sharpe, treynor, and jensen measures for the buy, reiterate buy, and downgrade to hold categories are calculated using monthly returns from january 1988 to december 199 1 (48 months). the performance measures for the strong buy and sell categories are cal culated using 46 and 35 monthly returns, respectively. the first strong buy recommenda tion is published in march 1988; the first sell recommendation is published in february 1989. portfolio construction for the strong buy category is illustrated in table 2 (n = 18 rec ommendations). on march 10, 1988, usa today reported a strong buy recommendation on walt disney. strategy 1 assumes that investors buy walt disney on march 1, 1988 and sell the stock on march 31, 1989. similarly, investors buy clark equipment on the first trading day of april 1988 and sell on the last trading day of april 1989. the april 1988 return for the strong buy portfolio, rp,, is an equal-weighted average of the monthly returns on walt disney and clark equipment. there are 46 monthly returns for the strong buy cat egory (march 1988-december 1991). these 46 monthly returns are used to calculate the sharpe, treynor, and jensen measures. the portfolio construction for the post-recommen dation strategy is similar except investors buy walt disney stock on april 1, 1988 and sell on march 31,1989. v. results a. event study results to evaluate analyst performance using an event study methodology, one should recall that event studies measure relative performance: the techniques test only whether stock performance during the event period differs from that observed during the estimation period. as a result, investors should consider the cumulative prediction errors earned dur ing the event period, and the abnormal performance (if any) observed during the estimation period itself.’ ’ an initial examination of the strong buy category, for example, suggests inferior per formance (see table 3). based on a pre-event estimation period, investors earned cumula t a b l e 3 e ve nt s tu dy m et ho do lo gy r es ul ts : 6 a nd 1 z m on th h ol di ng p er io ds in te rv al st ro ng b uy b u y r ep ea t b uy b uy t o h ol d se ll (t ra di n g d ay s) m c p e 2 m c p e 2 m c p e z m c p e z m c p e z p m -e ve nt e st im at io n p er io d [5, 12 51 -. 04 95 [5. 25 01 -. 14 10 11 , 1 25 1 -0 91 9 [l , 2. 50 1 -. 18 34 e st im at io n p er io d p ar am et er % .m o 66 zst at 4. 40 ** * p os te ve n t e st im at io n p er io d [5, 12 51 .0 01 6 [5, 25 0] -. -b 42 3 r1 , w -. 04 36 11 .2 50 1 -. 08 75 e st im at io n p er io d p ar am et er % j d o 01 8 tst at .6 9 -1 .3 2 .0 30 4 2. 22 )‘ . -2 .4 0* * .0 28 8 1. 57 -2 .1 9’ . -. 01 33 -. 36 -3 .0 1* ** -. 01 49 -. 26 -. 02 .0 41 6 3. 25 ** * .0 65 5 3. 36 ”’ -. 06 65 .0 9 -. 12 09 -1 .9 7” -. 54 .0 49 1 3. 27 ** ’ -s w 35 2. 80 ** -. 09 79 .0 3 -. 17 80 -1 .3 4 -9 0 -i i0 33 .9 3 -. 02 50 1. 21 -. 02 3 1 1. 61 -. 02 69 -. 18 -1 .1 7 .0 04 1 1. 61 -. 05 29 1. 25 -. 05 45 1. 10 -. 08 39 -. 05 .0 00 12 .0 00 25 1. 46 3. 09 ** * .o oo o4 .3 0 .0 10 2 1. 33 -. 08 71 -4 .2 1* ** -. 01 83 -. 65 -. 12 90 -4 .0 3* ** -. 02 82 -1 .0 2 -. 04 32 -2 .1 7* ’ -. 05 67 -2 .3 4. ’ -. 08 50 -2 .5 7* * -. 00 02 0 -. 00 00 4 -1 .8 0* -. 23 -. 06 72 -2 .0 2* * -. 09 22 -2 .0 6” .0 28 3 1. 00 .0 03 2 .0 8 .8 0 -1 .6 6 .o oo o4 .1 5 n ot es : l * l ** ** * in di ca te si gn if ic an ce at th e 1 0% . 5% . an d 1% le ve l, re sp ec ti ve ly . t h e s am pl e si z es fo r ea ch ca te go ry ar e a s f ol lo w s: st ro ng b uy ( 18 r ec om m en da ti on s) , bu y (1 16 ). r ei te ra te b uy ( il l) , bu y to h ol d (5 7) . an d a l( 27 ). 22 financial services review 5( 1) 1996 tive abnormal returns of -18% over the 12-month period following recommendation announcements. stocks that received a strong buy recommendation, however, performed extremely well during the estimation period. the null hypothesis of no abnormal perfor mance is rejected at the 1% level fip = boo66 and f = 4.40). the bias over a 250-day inter val is 16.5% (.00066 x 250 = .1650). together, these results indicate that (a) strong buy recommendations typically follow a period of superlative performance, and (b) the supe rior performance does nor continue past the recommendation announcement date. a simi lar argument applies to the repeat buy category. negative recommendations, on the other hand, tend to be the most accurate. investors lose approximately 8.5% following downgrades from buy to hold (t = -2.57, which is sig nificant at the 5% level). stock price performance during the estimation period is not sta tistically significant (7, = .oooo8 and t = .80). in addition, there is some evidence that investors earn negative abnormal returns following sell recommendations. the average abnormal return during the estimation period is -xl0021 per day (t = -1.66, which is signif icant at the 12% level), and no significant change in this performance is observed following sell recommendations. the results for the post-event estimation period, which are reported in the lower panel of table 3, indicate that investors generally earn normal returns by following analyst rec ommendations. only the estimation period parameter for the repeat buy category is statis tically significant @, = -.0002 and t = -1.80, which is significant at the 10% level). in sum, the results indicate that both preand post-event estimation period parameters can be biased (particularly for the repeat buy category). regardless of the estimation period selected, as one would expect investors generally earn higher returns by purchasing stocks at pre-recommendation prices (see, e.g., the r = -5 to t = +250 interval). announcement day effects are discussed in the next section. b. announcement day effects while the objective of this study is to examine a long-term holding period following analysts’ recommendations, it is interesting to note the returns surrounding the announce ment date. the announcement effect for positive recommendations (strong buy, buy and repeat buy) is approximately 4% (see table 4). negative recommendations (buy to hold and sell) are associated with mcpes of approximately -4% to -6%. in general, the abnormal returns are focused on the announcement date (t = 0). for the strong buy and buy categories, the returns over the five-day period ending one day before the announcement date (r = -5 to r = -1) are not statistically significant. the cumulative abnormal return for the repeat buy category is .85% over this interval (t = 2.28, which is significant at the .05 level), but an additional 3% is earned on the announcement date. for the sell category, 67% of the cumulative abnormal return observed over the t = -5 to t = 0 interval is earned on the announcement date. as a result, relatively small price changes are observed between the date recommendations are disclosed to clients and the date recom mendations are disclosed to the general public (t = 0). recall that the usa today publica tion date is t = +l. relatively small price changes are also observed on or immediately following the usa today publication date. stocks that receive a strong buy recommendation earn positive abnormal returns of approximately 1.77% over the t = + 1 to t = +5 interval (t = 1.93, which t a b l e 4 e ve nt s tu dy m et ho do lo gy r es uj ts : a nn ou nc em en t d ay e ff ec ts in te n s tr on g b uy b u y r ep ea t b uy b uy t o h ol d (t ra di ng d ay s) m c p e z m c p e z m c p e z m c p e z l -5 , 11 a0 34 .2 0 .o o lo .1 9 .0 0 8 5 2 .2 8 ** -. 0 0 7 8 -1 .7 4 : r1 , -1 1 .o m 2 .8 0 a0 6 2 4 .8 8 ’* ’ .0 0 7 0 4 .8 2 ”’ -. cm 6 7 -3 .7 1 ** * 10 . 0 1 .0 3 9 0 8 .8 8 ** * .0 4 1 6 2 8 .5 8 *‘ * .0 3 0 4 2 2 .0 4 ** * -. 0 3 8 5 -2 1 .3 7 ”’ [l . 11 .0 1 6 7 3 .4 6 ** * a0 4 7 3 .9 4 ** * a0 1 2 9 9 -. 0 0 2 8 -1 .8 3 ’ l 51 .0 17 7 1. 93 ’ .0 10 9 3. 63 ** * a0 54 2. 38 ” -. 01 19 -3 .2 3* ” n or cs : 8. = *, d = ** m d ic at e s ig n it ic an ce at th e 10 % . 5% an d 1% l ev el , re sp ec ti ve ly . ‘t k sa m pl e s iz c s fo r ea ch c at eg or y ar e as f ol lo w s: s tr on g bu y (1 8 re co m m en da ti on s) . bu y (1 16 ). r ei te ra te b uy ( 11 i) . b uy t o ho ld ( 57 ). a nd s .a i (2 7) . s el l m c p e z -. 0 3 1 0 -5 .1 6 ”’ -. o lo o -3 .9 6 ** * -. 0 6 2 3 -2 2 .9 4 ”’ .0 0 2 6 .5 5 .0 10 2 1. 38 24 financial services review 5( 1) 1996 is significant at the .10 level). buy recommendations gain 1.09%, repeat buy recommenda tions gain .54%, downgrades from buy to hold lose 1.19%, and sell recommendations experience normal returns. despite the price change requirement for usa today to publish recommendations, the results of this study are similar to the findings of previous research. liu et al. (1990) reported mcpes of 3.0% for buy recommendations and -3.6% for sell recommendations over the six-day period ending with t = 0. they concluded that the publication of recom mendations in the wall street journal’s “heard-on-the-street” column (which was the source of their recommendations) impacts security prices. as their forecast period con tained only 2 1 trading days centered around the publication date, they did not consider the long-term impact. our findings also support the results in barber and loeffler (1993), who examined the recommendations in the monthly “dartboard” column in the wall street joumaz. they found that analyst buy recommendations resulted in a mcpe of 3.53% on the publi cation date (t = 12.19). while they concluded that the initial price response was partially reversed over the following 25day period, they did not test for bias in the estimation period parameters. because the results in tables 3 and 4 do not include brokerage commissions, the actual returns experienced by most investors would have been less. bodie, kane, and marcus (1993) indicate that the commissions charged by full-service brokerage firms often represent about 2% of the transaction value. based on a hypothetical purchase of 200 shares at $26 per share, for example, bodie, kane, and marcus indicate that full-ser vice brokers would charge about $135 while discount brokers would charge approxi mately $6 1. c. sharpe, treynor, and jensen measures the sharpe, treynor, and jensen measures also provide evidence that professional ana lysts do identify mispriced securities (see table 5). the results in the upper panel of table 5 are based on pre-recommendation prices (i.e., the holding period is 13 months: t = 0 to t = + 12). the sharpe measures for the strong buy, buy, and reiterate buy categories do not exhibit a consistent pattern, but the treynor measures exceed the market portfolio ratios. the jensen alpha measures are positive, but the results are not statistically significant. l2 the results for negative recommendations (sell and downgrades from buy to hold) are more pronounced. in each case the sharpe and treynor measures are less than the market portfolio benchmarks. jensen’s alpha measures are negative, and the results for the down grade to hold category are statistically significant at the 5% level (c+, = 1 .ol% per month and t = -2.28). none of the results in table 5 reflect transaction costs. whether investors who act after the publication in usa today can benefit from analyst recommendations is more problematic (see the lower panel in table 5). not only do the sharpe and treynor measures indicate inferior portfolio performance, but the jensen alpha measures are negative (though not statistically significant). in general, the beta coefficients for each category are greater than 1, and the null hypothesis that bp = 1 can be rejected at the 10% level. as a result, professional analysts tend to issue recommendations on stocks that exhibit above-average risk. t a b l e 5 a n al ys t in ve st m en t r ec om m en d at io n s: s h ar p , t re yn or , an d j en se n m ea su re s (1 98 819 90 ) s ha ip e’ s m ea su re t re yn or ’s m ea su re a na ly st m ar ke t a na ly st m ar ke t je m en ’s m ea su re r ec om m en da ti on p or rf ol io p o* ol io p or tf ho 13 -m on th p er io d be gi n n in g w it h a n n ou n ce m en t m on th [ t = 0 ; t = + 1 21 1. s tr on g b u y .1 68 .1 68 .7 6l .2 . b u y .2 42 .2 07 1. 07 6 3. r ei te ra te b u y .1 78 .2 07 .8 29 4. d ow n gr ad e t o h ol d .0 16 20 7 .0 72 5. s el l -. 05 0 .1 52 -. 28 7 12 -m on th p er io d fo ll ow in g an n ou n ce m en t m on th [t = + 1 ; t = + 1 21 1. s tr on g b u y .0 98 .1 87 44 7 2. b u y .1 30 .1 89 .5 72 3. r ei te ra te b u y .1 38 .1 89 .5 95 4. l h vn gr ad e t o h ol d a9 6 .1 89 .4 18 5. s el l .l o o .1 73 .6 00 p or tf ol io .6 68 .1 63 .2 8 1. 65 4. 59 ** * .8 18 .3 18 .8 3 1. 23 2. 51 ** * .8 18 .0 13 .0 3 1. 18 1. 68 * .8 18 -1 .0 13 -2 .2 8* * 1. 36 3. 26 ”’ 64 8 -1 .2 67 -1 .3 8 1. 35 1. 67 ’ .7 43 .7 49 .7 49 .7 49 .7 40 -. 48 9 -. 87 1. 65 4. 70 ”’ -. 22 1 7. 60 1. 25 2. 79 ** * -. 2c m -x i0 1. 29 3. 66 ”’ -. 45 8 -1 .2 5 1. 38 4. 22 ’* ’ -. 17 5 -. 19 1. 25 1. 15 t t 26 financial services review 5( 1) 1996 vi. conclusions the results of this study have both theoretical and practical implications. from a theoretical standpoint this study tests whether professional stock analysts can identify undervalued securities (a test of strong form market efficiency). from a practical standpoint this study tests whether investors can profit from analyst recommendations. two factors should be considered. first, potential gains must be evaluated net of transaction costs. second, while an analyst may identify an undervalued security and issue a buy recommendation, an inves tor who purchases the security typically must pay a price that reflects any information embedded in the recommendation announcement. as a result, an investor may not be able to capitalize on a recommendation even if it is “correct” ex post. this study tests the accuracy of analyst recommendations published in the “market highlights” section of usa today. investment performance is evaluated for two groups of investors: (a) institutional and retail clients who trade before recommendations are dis closed to the general public, and (b) individual investors who trade after recommendations are published in usa today. the use of recommendations published in usa today has the advantage that these are recommendations available to a wide range of investors. still, because usa today publishes only those recommendations that have affected stock prices, it is important to note that the conclusions of this study may apply only to this subset of analysts’ recommendations. the results of this study support the conclusion that professional stock analysts do iden tify mispriced securities. similar to previous research, we find an announcement effect of approximately 4% for positive recommendations. when a twelve-month holding period is examined and returns are based on pre-recommendation prices, the sharpe, treynor, and jensen measures consistently rank the analyst buy portfolios above the sell and hold port folios. even though the returns are not statistically significant, the returns may exceed the costs of implementing an active strategy for large institutional investors (bodie et al., 1993). individual investors, on the other hand, tend to experience subpar returns by following analyst recommendations. when returns are based on post-recommendation prices, the sharpe and treynor measures indicate inferior performance relative to the benchmark port folio. the jensen alpha measures are negative, but the results are not statistically signifi cant. because the results do not reflect transaction costs, the returns for most individual investors would be even lower. these results support peter lynch’s concept of “street lag” (i.e., by the time an individual investor can act on an analyst’s recommendation, the “good news” is already impounded in the stock price). finally, the results indicate that the conclusion one draws regarding analyst perfor mance often depends on the methodology employed. stock index comparisons are the least reliable because recommended stocks often exhibit greater risk than the benchmark selected. if an event study methodology is used, on the other hand, researchers should test whether the market model intercept for the estimation period is equal to zero. ignoring this test can lead to bias in the abnormal returns reported for the event period (see also cope land & mayers, 1982; edmister, graham, & scott, 1994). acknowledgment: the authors wish to thank keith womer, the participants at the 1994 financial management association meeting, and two anonymous reviewers for helpful comments and suggestions. the usual disclaimer applies. 28 financial services review 5( 1) 1996 show that the dodd-warner test statistic is biased and that the bias increases with the length of the interval examined. as a result, we use the test statistics suggested by karafiath and spencer (1988) and mikkelson and partch (1988). these test statistics are smaller than would be obtained if the serial correlation in the prediction errors were ignored, the formula for the test statistic is: mscpej = (l&-t1 + 1)) ; pejt/var ; pejt t= t, t= t, where ti is the fust day of the interval, t2 is the last day of the interval, and the denomina tor is the square root of the variance of the cumulative prediction errors of firmj. the vari ance is defined to be: a4sej is the standard deviation of the regression, t is the number of days in the interval (t2 t1 + l), ed is the number of days in the estimation period for the market model, r,,,t is the market return on day t, and r, is the mean market return during the estimation period. because the weights used in calculating the mscpe-statistic ire a modified inverse of the standard deviation of the cumulative prediction errors, the z-statistic can differ in sign from the average prediction error (since returns of securities with lower variance are given greater weight). iteferences barber, b.m., & loeffler, d. (1993). the dartboard column: second-hand information and price pressure. journal of financial and quantitative analysis, june, 273-284. bjerring, j., lakonishok, j., & vermaelen. t. (1983). stock prices and financial analysts recommen dations. journal of finance, 38, 187-204. bodie. z.. kane, a., & marcus, a.j. (1993). investments. boston, ma: irwin. copeland, t.e., & mayers, d. (1982). the value line enigma (19651978): a case study of perfor mance evaluation issues. journal of finuncial economics, june, 289-322. dodd, p., & warner, j.b. (1983). on corporate governance: a study of proxy contests. journal of financial economics, april, 401-438. dorfman, j.r. ‘(1993, april 19). heard on the street. wall street journal, p. c2. dorfman. j.r. (1993. september 15). all star analysts 1993 survey. wall street journal, p. r9. dorfman, j.r. (1994, february 1). paine webber picks topped a good year for wall street. wall street journal, p. c 1. edmister, r.o., graham, a.s., & scott, t. (1994). selection bias in event studies: the case of dart board stocks. presented at the 1994 financial management association meeting, st. louis, mo. fama, e.f. (1970). efficient capital markets: a review of theory and empirical work. journal of finance, may, 383-417. prafessional stock analysts’ recommendations 27 notes 1. wall street journal, march 7, 1995, section c. 2. empirical studies indicate that corporate insiders and stock specialists generally do profit by having access to nonpublic information (see most investment texts, e.g., reilly & norton, 1995, for a review of this literature). 3. ippolito (1993) reviews studies that test mutual fund performance and finds support for the grossman and stiglitz hypothesis: mutual funds appear to earn positive risk-adjusted returns, but the returns are offset by higher operating expenses and trading costs. as a result, active and passive investors earn the same rate of return net of expenses. 4. research has found that announcements by investment advisory agencies (moody’s, s&p, value line) also provide information to the market (see griffin & sanvicente, 1982; holth ausen & leftwich, 1986). 5. this data source was selected because it reaches a broad spectrum of investors. 6. the 329 sample recommendations involve 204 firms: 138 firms received 1 recommenda tion, 39 firms received 2 recommendations, 11 firms received 3 recommendations, 8 firms received 4 recommendations, 5 fiis received 5 recommendations, 1 firm received 6 recommendations, 1 firm received 8 recommendations, and 1 firm received 9 recommendations. 7. zacks investment research inc. of chicago used this technique to evaluate 32,000 recom mendations made by 1,275 analysts during 1992 (dorfman, 1993, september). the results of that study indicated that following the analysts’ stock recommendations would have produced a return of 8.9% (compared to a total return of 7.6% on the s&p500 index). in 26 of the 30 industries, however, an inves tor would have outperformed the analysts’ picks by buying an equally-weighted portfolio consisting of nonrecommended firms in the same industry. mr. ryan, a research manager at zacks, attributed this result to the superlative investment performance of small-capitalization stocks during 1992. 8. the sample recommendations are typical of the following excerpt from the “market high lights” section of usa today (march 10,1988, page 3b): “walt disney gained 1 to 63-l/4 on a ‘strong buy’ from cyrus j. lawrence, and analog devices jumped l-1/8 to 14-3/4 after it was recommended by goldman, sachs.” the announcement date (t = 0) for this example is wednesday, march 9, 1988. 9. analyst performance is also evaluated using (a) a post-event estimation period (t = +25 1 to t = +500 trading days), and (b) the crsp equal-weighted index as the market return. because the findings of this study are not affected by the choice of the market index, the results for the equal weighted index are not reported separately. 10. if 3 in equation 1 is positive and statistically significant, then the test would be biased against finding a positive abnormal return. however, a stock with a positive x during the estimation period and no abnormal return during the event period would still be regarded as a good investment. 11. copeland and mayers (1982) also note the problems that arise when the estimation period parameters are biased. edmister et al. (1994) test for bias in the pre-event estimation period parame ters. we test for bias in both the preand post-event estimation periods. 12. bodie et al. (1993) argue that large institutional investors may be able to justify fundamen tal and technical analysis even though the returns are not “significant” using conventional statistical tests. the results for the buy category, for example, suggest an abnormal dollar return of approxi mately $3.82 million per year on a $100 million portfolio (.00318 x 12 months/yr x $100 mil. = $3.82 mil.). even though the return is not statistically significant, it may exceed the cost of implementing an active strategy. appendix standard event study methodology is used to estimate the excess returns (see dodd & warner, 1983). for intervals longer than one day, however, karafiath and spencer (1988) professional stock analysts ’ recommendations 29 givoly, d., & lakonishok, j. (1979). the information content of financial analysts forecast of eam ings. journal ofaccounting and economics, december, 165-185. griffin, p.a., & savicente, a.z. (1982). common stock returns and rating changes: a methodologi cal comparison. journul of finance, march, 103-l 19. grossman, s.j., & stiglitz. j.e. (1980). on the impossibility of infotmationally efficient markets. american economic review, june, 393408. groth, j.c., lewellen, w.g., schlarbaum, g.g., & lease, r.c. (1979). an analysis of brokerage house securities recommendations. financial analysts journal, january-february, 32-39. holloway, c. (1981). a note on testing an aggressive investment strategy using value line ranks. journal of finance, june, 711-719. holloway, c. (1983). testing an aggressive investment strategy using value line ranks: a reply. journal of finance, march, 263-270. holthausen, r.w., & leftwich, r.w. (1986). the effect of bond rating changes on common stock prices. journal of financial economics, september, 57-89. ippolito, r.a. (1993). on studies of mutual fund performance, 1962-1991. financial analysts jour nal, january/february, 42-50. karatiath, i., 8t spencer, d.e. (1988). assessing the validity of the standardized cumulative predic tion error and alternative test statistics. working paper, massachusetts institute of technol ogy, boston, ma. liu, p., smith, s.d., & syed, a.a. (1990). stock price reactions to the wall street journal’s secuti ties recommendations. journal of financial and quantitative analysis, september, 399-410. lynch, p. (1989). one up on wall street. new york penguin books. mikkelson, w.h., & partch. m.m. (1988). withdrawn security offerings. journal of financial and quantitative analysis, june, 119134. reilly, f.k., & norton, e.a. (1995). investments. new york dryden press. roll, r. (1994). what every cfo should know about scientific progress in financial economics: what is known and what remains to be resolved. financial management, summer, 69-75. pii: s1057-0810(00)00058-5 social security reform: the effect of investing in equities erick elder, larry holland* college of business administration, university of arkansas at little rock, little rock, ar 72207-1099, usa abstract several proposals have been developed to reform the social security system to ensure that it is fully funded. the investment of a portion of social security funds in equities has often been proposed as a means to avoid increasing payroll taxes. this paper develops a general equilibrium model to demonstrate that investing social security funds in equities will decrease the return on equities and increase interest rates on bonds, which also leads to an increase in general income taxes. thus, investing social security funds in equities simply shifts a potential increase in payroll taxes to an increase in income taxes. © 2000 elsevier science inc. all rights reserved. jel classification:e44; e62; g12 keywords:social security reform; general equilibrium; portfolio choice; asset pricing 1. introduction a recent report of the advisory council on social security (advisory council, 1997) shows that the u.s. social security system will not be fully funded over the next 50 years. forecasts show that with current and projected contributions and outflows, the social security trust fund will be completely depleted by the year 2034 and benefits would have to be reduced beyond that date by 29%. therefore, there are several proposals to reform the social security system, many of which suggest investing of a portion of the social security trust fund into the equities market in order to obtain a higher return on accumulated funds. in this way, the social security system could supposedly become fully funded without raising payroll taxes. * corresponding author. tel.:11-501-569-3042; fax:11-501-569-8871. e-mail address:lcholland@ualr.edu (l. holland). financial services review 9 (2000) 93–106 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00058-5 there is some debate concerning the impact of this type of social security reform on asset returns. clearly, social security reform would have no effect on asset returns if the following two assumptions are true: (1) individuals perceive social security as part of their savings (or at least individuals must take into consideration the investment of social security when making their own allocation decision) and (2) no individuals are constrained in terms of their desired allocation, that is, every individual should be able to invest their entire retirement account (including social security) in bonds and/or equities according to their preference. serious questions as to the validity of these assumptions exist. first of all, under the current defined benefit system, individuals should not care about the investment of social security funds because the benefits are fixed regardless of the allocation decision. secondly, geanokoplos, mitchell, and zeldes (1998) point out that if households are constrained in their investment choices, there would be macroeconomic consequences in terms of changes in asset returns due to the increase in demand for equities. this could occur if some individuals want all their retirement funds in equities (extreme risk takers) or in bonds (highly risk averse investors) or if some individuals do not have access to credit markets. if either of these two assumptions are violated, then asset returns will be affected to some degree by social security reform, and specifically by shifting funds to the equities market. the intuition behind this effect on interest rates and asset returns is that, if either of the two above assumptions is violated, then a shift in funds to equities is equivalent to a change in preferences. in our view, these two assumptions are likely to be violated, and this forms the basis for the model in this paper. in this paper, we formally model the impact of social security reform on asset returns in the presence of constrained individuals. the model shows unambiguously that interest rates on bonds will increase and the return on equities will decrease as social security funds are invested in the equities market. we find that the magnitude of the change in asset returns could be greater than one percentage following a forty percentage shift of social security funds to the equities market. the topic of social security reform has attracted a large literature and has been the topic of numerous conferences (federal reserve bank of boston, 1997, american economic association 1996, and the national academy of social insurance, 1998). several researchers have commented, without the use of a formal model, that such a reform might affect the overall economy in terms of changing interest rates and equity returns. diamond (1996), stein (1997), stiglitz, munnell, and frankel (the council of economic advisers) in their economic report of the president (1997), and geanokoplos et al. (1998) mention the potential impact of reform on asset returns. in addition, it is interesting that the advisory council’s analysis of the effect of three different reform programs did not take into account the effect on interest rates from the implementation of the reform itself. other researchers have developed formal models to study the impact of social security reform, but their focus is on topics other than the impact on asset returns. these studies generally use the life-cycle framework developed by auerbach and kotlikoff (1987) (e.g., kotlikoff, smetters, and walliser, 1998; bohn, 1997; smetters, 1998; and diamond, 1997). the familiar life-cycle approach might seem to be the natural framework for analyzing social security since the very essence of a pay-as-you-go social security system is the transfer of assets from generation to generation. this is especially true if one is interested in a welfare analysis of the social security system, for which a life-cycle framework is ideal. in 94 e. elder, l. holland / financial services review 9 (2000) 93–106 contrast, the primary purpose of this paper is to examine the instantaneous effects of social security reform on asset returns. to this end, the benefits of using a life-cycle model (e.g., bohn, 1997 and diamond, 1997) are not obvious and can actually cloud the issue. this is because a major component of the life-cycle models is the saving-consumption and/or labor-leisure decision, which may be altered by (or may alter) returns. the empirical evidence is ambiguous concerning the relationship between returns and these decisions (e.g., hall, 1988 and campbell and mankiw, 1989). the strength of our results is that they do not depend on any changes in these decisions. we construct a representative-agent model to analyze the effects of social security reform, addressing the constraints of geanokoplos et al. (1998). our model uses the optimal-portfolio selection rule derived by merton (1969) in continuous time and samuelson (1969) within a discrete time framework in which the optimal-portfolio selection rule is a function of a given set of expected asset returns. we extend this analysis to a general equilibrium framework to endogenously determine these expected returns. it is assumed that social security rules determine the allocation of a portion of each individual’s wealth. furthermore, individuals with different levels of risk aversion attempt to achieve an optimal allocation of their wealth between a risky asset (equities) and a risk-free asset (bonds). some individuals may be unable to achieve their optimal allocation, and therefore are constrained (by social security) to hold funds in excess of their optimal allocation in either the bond or equity market. this leads to a change in asset returns. an interesting implication is that with an increase in the interest rates on government debt, income taxes must be raised to pay the higher interest on the national debt in order to maintain the same budget deficit position as prereform. thus, investing social security funds in equities to some extent simply shifts a potential increase in social security taxes to an increase in general income taxes. this impact is of immense importance when discussing potential reform because it significantly alters the merits of the potential reform plans. in section 2, we first construct a model in which there are constrained individuals. we then analyze in section 3 the qualitative impact on asset returns from shifting social security funds to the equities market and briefly report examples of the potential magnitude of the effects this reform may have on bond and equity returns. in section 4, we explore comparative statics to determine the impact of various other parameter values. we present extensions of the model in section 5 to include income taxes and show why income taxes would increase if social security funds are invested in equities. we also briefly explore the potential implications of increasing the payroll taxes in section 6. finally, section 7 concludes the paper, providing a summary and possible extensions of our research. 2. the model the model in this section assumes that individuals consider social security as part of their savings and therefore take into account the allocation decisions of the social security trust fund when making their own allocation decisions. following merton (1969), there is one risk-free asset (bonds) and one risky asset (equities). bonds return a real risk-free rate, rb . 0, while the return on equities is normally distributed with mean re and variances2. suppose 95e. elder, l. holland / financial services review 9 (2000) 93–106 that there arem different types of individuals where each individuali maximizes his utility using a power utility function of the type u(ci)5ci gi/gi, wheregi , 1 for all i (this precludes the possibility of a risk-loving individual). the widely used power utility function is not necessary for the qualitative results of this paper, but it is convenient because it offers an explicit solution to the optimal portfolio rule. furthermore, the power utility function is consistent with the empirical results of friend and blume (1975). this type of utility function exhibits constant relative risk aversion with a coefficient of relative risk aversion of 1-gi. individuals are ordered on the interval [1,m] by their degree of risk aversion where individual 1 is the most risk-averse and individualm is the least risk-averse (gm . gi). in addition, each individual is assumed to have a wealth level ofwi. maximizing utility entails choosing optimal consumption levels as well as an optimal allocation of wealth between the risk-free and risky asset. it can be shown that each individuali would optimally like to allocate a fraction of his wealth,vi to equities and 1-vi of his wealth,wi, to holding bonds where vi 5 re 2 rb ~l 2 gi!s2 (1) (see merton, 1969, page 250 or ingersoll, 1987, page 275 for a derivation of the optimal portfolio allocation.) individuals are forced to contribute to social security, and the accumulated funds in the social security trust fund are allocated without any consideration for individuals’ wishes. letn $ 0 be the prereform fraction of social security funds allocated to equities (note that this is the same for everyone and under the current regime isn 5 0) and n9 is the post reform fraction (where future values are denoted by a9). finally, let hi be the fraction of an individual’s wealth controlled by social security. therefore, social security allocates (1-n) hi fraction of an individual’s wealth to the bond market andnhi fraction of an individual’s wealth to the equities market. furthermore, it is also assumed that individuals cannot sell stock short. therefore, some individuals may be unable to achieve their optimal allocation. for any given return pair, (re,rb), individuals can be grouped into three categories when 0# n # 1: the most risk-averse individuals will be constrained to hold funds in excess of their optimal allocation in the equities market, the least risk-averse individuals will be constrained to hold funds in excess of their optimal allocation in the bond market, and the remaining individuals will not be constrained in either market. we will treat each of the three different categories of individuals in turn. given this ordering of individuals, there exists a marginal investor, represented byie, such thatvie 5 nhie (individual ie’s optimal allocation to the equities market is exactly equal to the amount social security allocates to the equities and bond markets). individuals of typex , ie (or equivalently, all investors more risk averse thanie) are constrained to hold more funds in the equities market than they desire, that is,nhx . vx but (1-n)hx , 1-vx. these individuals will allocate a total ofnhxwx to the equities market and their remaining funds, (1-nhx)wx, to the bond market. furthermore, there is also a marginal investor, ib, such that 12vib 5 (1 2n)hib (individual ib’s optimal allocation to the bond market is exactly equal to the amount social security allocates to the equities and bond markets). individuals of typez. ib (or equivalently, all investors less risk averse than ib) are constrained to hold more funds in the bond market than they desire, that is, (1-n)hz . 1-vz but nhz , vz. individuals in this group will allocate (1-n)hzwz to bonds and (1-(1-n)hz)wz 96 e. elder, l. holland / financial services review 9 (2000) 93–106 to equities. finally, there is also a group of individuals,y @ [i e, ib] which are not constrained to be in either market, that is, (1-n)hy # 1-vy andnhy # vy. individuals in this group will allocatevywy to the equities market and (1-vy)wy to the bond market. it is assumed that there is a fixed quantity of bonds and stocks, qb and qe respectively, and changes in the return of a given asset are determined by the current prices, pb and pe respectively. also by assumption, there is general agreement (i.e., homogeneous expectations) on the expected future prices of bonds and stocks, e[pb9] and e[pe9], based on the expected growth in the real assets of the economy, which means that the expected-future prices are constants. conceptually, one can think of bonds being paid off with certainty at face value at maturity and the real assets of each firm being liquidated to pay off the stockholders at a known expected value. therefore, forn . 0, the market clearing conditions for the bond and equity market are respectively given by o x51 ie21 nx~1 2 nhx!wx 1 o y5ie ib ny~1 2 vy!wy 1 o z5ib11 m nz~1 2 n!hzwz 5 qbpb (2) o x51 ie21 nxnhxwx 1 o y5ie ib nyvywy 1 o z5ib11 m nz~1 2 ~1 2 n!hz!wz 5 qepe (3) note that pe 5 e[pe9] exp(-re) and pb 5 e[pb9] exp(-re), where exp(-r) represents the exponential function raised to the -r power, which is the present value factor using continuous discounting. substituting in (2) and (3) yields o x51 ie21 nx~1 2 nhx!wx 1 o y5ie ib ny~1 2 vy!wy 1 o z5ib11 m nz~1 2 n!hzwz 5 qbe@p9b# exp~rb! (4) o x51 ie21 nxnhxwx 1 o y5ie ib nyvywy 1 o z5ib11 m nz~1 2 ~1 2 n!hz!wz 5 qee@p9e# exp~re! (5) this is the most general form of the model. however, the model can be simplified by assuming there are only 3 categories of individuals rather thanm. with this simplification, it is easier to demonstrate the effects on bond and equity returns. suppose, for example, that each category of individuals discussed above is represented by a single type of individual, that is, nx individuals with the samegx are constrained to hold excess funds in the equities market, ny individuals with the samegy are not constrained to hold excess funds in either market, and nz individuals with the samegz are constrained to hold excess funds in the bond market. thus, the total number of individuals is nx 1 ny 1 nz 5 n. for simplicity assume that wx 5 wy 5 wz 5 w andhx 5 hy 5 hz 5 h. then the market clearing conditions for the bond and equity markets, (4) and (5), can be rewritten as nx~1 2 nh! 1 ny~1 2 vy! 1 nz~1 2 n!h 5 b exp~rb! (6) 97e. elder, l. holland / financial services review 9 (2000) 93–106 nxnh 1 nyvy 1 nz~1 2 ~1 2 n!h! 5 e exp~re! (7) where b5 qb e[pb9]/w and e5 qe e[pe9]/w are standardized values for bonds and equities (relative to total wealth), respectively. eqs. (6) and (7) can be solved for rb and re, the bond and equity returns, by substituting eq. (1) for the optimal portfolio allocation,vy. 3. qualitative effects of shifting funds on asset returns using eqs. (6) and (7), it is possible to analyze the qualitative impact of shifting social security funds on asset returns. eqs. (6) and (7) can be rearranged to show that re is a function of rb andn and that rb is a function of re andn. re 5 f~rb,n! (8) rb 5 g~re,n! (9) taking the total derivative of (8) and (9) with respect ton results in dre dn 5 frb drb dn 1 fn (10) drb dv 5 gre dre dv gv (11) from which it is possible to show that dre dv 5 gv~ frb 2 l ! l 2 frb gre , 0 (12) dre dv 5 2 gv~ g 2 l ! l 2 frb gre . 0 (13) where frb 5 1 1 (b(1 2 gy)s 2)/(ny exp(rb)) . 1, gre 5 1 1 (e(1 2 gy)s 2)/(ny exp(re)) . 1, and gn 5 (nxh 1 nzh)(1 2 gy)s 2)/ny . 0. (see appendix a for complete details of the proof.) thus, as the social security trust fund shifts its investment from government securities to equities, re decreases and rb increases. it is possible to think of the chain of events following reform as (1) social security reallocates some of the trust fund to the equities market, (2) unconstrained, typey individuals reallocate their portfolio in order to maintain their prereform allocation, (3) constrained individuals now have more funds in the equities market (and fewer in the bond market) and therefore the equity return falls and the bond return rises, (4) typey individuals then reallocate their portfolio (moving some funds back to the bond market because re-rb has decreased) which offsets some of the changes in re and rb in step 3, and finally (5) steps 3 and 4 are repeated until a new equilibrium is achieved. 98 e. elder, l. holland / financial services review 9 (2000) 93–106 it is interesting that the change in re and rb following social security reform is not dependent upon the level of risk aversion of the constrained individuals. the intuition is straightforward. it is assumed that individuals that are constrained in a given market prior to reform remain constrained following reform. whether an individual is constrained is dependent upon his risk aversion (as shown in appendix b). the amount of funds that a constrained individual invests in a given market is by definition fixed by the constraint, regardless of the degree of constraint. therefore all individuals constrained to hold funds in the bond (equities) market will invest the same amount of funds regardless of the degree of risk aversion. it is also straightforward to understand why there are no effects on re or rb if no one is constrained. this can be shown by noting the bond and equity market-clearing conditions under the situation in which no one is constrained. prereform, these conditions are respectively o i51 m vi 5 e exp~re! (14) o i51 m ~l 2 vi! 5 b exp~rb! (15) which is exactly the same as the post reform conditions, indicating that no change in either re or rb occurs following reform. if no one is constrained, then all individuals can simply reallocate the fraction of their wealth which they have control over in order to fully achieve their prereform allocation (hence leaving returns unchanged). therefore, effects on re and rb occur only if there are individuals who are constrained. if individuals become unconstrained then that individual’s action no longer has an effect on re and rb (but re and rb could still be changing if others continue to be constrained or if new individuals become constrained). this is consistent with the results given in (12) and (13) by noting that the “no constrained individuals” case is equivalent to nx 5 nz 5 0 (in which case gn 5 0 and dre/dn 5 0 and drb/dn 5 0). the actual magnitude of these effects is an empirical question. however, we can use eqs. (6), (7), (12), and (13) to calculate the potential magnitude of changes in interest rates and equity returns for a wide range of parameter values forn, ny, nz. the changes in rb and re, following a shift of forty percentage of the trust fund into the equities market, range from dre/dn 5 21.21%, drb/dn 51.30% (ny 5 1, nz 5 1, n 5 0.4) to dre/dn 5 20.03%, drb/dn 5 0.03% (ny 5 10, nz 5 1, n 5 0.1). this range of estimates is consistent with the estimated increase in interest rates of 1.44% in elder and holland (1999), which uses a different approach to empirically estimate the effects on interest rates. the estimates from these calculations are based on several assumptions. the relative size of ny and nz is analogous to the size of the population which is unconstrained relative to the amount of the population which is constrained. these results set nx 5 0 because currently there are no social security funds invested in the equities market so there cannot be anyone constrained to hold excess funds in the equities market. the assumption that nx remains equal to zero as funds are shifted into the equities market is rather conservative. if we were to allow 99e. elder, l. holland / financial services review 9 (2000) 93–106 initially unconstrained individuals to become nx type individuals the results below would be magnified. the calculations set the relative size of the bond and equities market to match the actual relationship; the flow of funds accounts of the united states (1999) shows that in 1998 the market capitalization of the equities market was $15,438 billion and the market capitalization of the bond market (including treasury and government agency securities, corporate and foreign bonds, and municipal bonds) was $12,387 billion. 4. other comparative statics it is also interesting to examine how the changes in rb and re are affected by changes in nx, nz, andgy. qualitatively, these effects are found by taking the cross derivatives of (12) and (13). first of all, it can be shown that with respect to nx or nz that dre/(dndnx)5 dre/(dndnz) , 0 and drb/(dndnx) 5 drb/(dndnz) . 0, meaning that as more individuals are constrained the effects of reform on re and rb are magnified. to understand this, note that as more individuals are constrained (regardless of where they are constrained), a larger amount of funds is not adjusted in each market (following the reallocation of they individuals in step 2). this leads to a larger initial change in returns (step 3 above) and finally, a larger final effect of reform on returns. the other interesting part of these results is that the effect of reform on returns is not dependent upon where the constrained individuals are constrained, but just dependent on thetotal number of constrained individuals. for every 1% of the trust fund that social security shifts to the equities market, the allocation ofz individuals (most risk-averse) moves closer to their optimal allocation byh% while the allocation ofx individuals (least risk-averse) diverges from their optimal allocation byh%. regardless of whether the constrained individuals are moving closer to their optimal allocation or not, the same amount of funds are being moved to the equities market in step 3 above; hence the effect of reform on returns is only dependent on the number of constrained individuals and not where they are constrained. another factor determining the magnitude of the change in returns is the level of risk aversion of the unconstrained individuals, since these individuals can partially (or completely) undo the changes made by social security. this result is obtained by taking the cross derivative with respect togy. it can be shown that dre/dndgy . 0 and drb/dndgy , 0. this means that when unconstrained individuals are more risk-averse, the effects on rb and re will be larger following reform. to understand this, first note that the allocation decision of less risk-averse individuals is more sensitive to changes in re-rb. if the risk premium rises, then less risk-averse individuals increase their allocation to equities much more than do more risk-averse individuals. it follows that the less risk-averse they individuals are in step 4 above, the larger will be the amount of funds shifted back to the bond market by those individuals, hence offsetting more of the initial change in returns. analogous results obtain if individuals do not perceive social security funds as a part of their wealth (see appendix c for details) which may be relevant if individuals do not expect to receive any benefits from social security in the future. 100 e. elder, l. holland / financial services review 9 (2000) 93–106 5. implications of changing asset returns some of the social security reform proposals suggest that it is possible to avoid an increase in the social security tax rate by simply shifting trust fund money into the equities market. social security investments in the trust fund will benefit from a movement of funds to the equities market in two ways: (1) higher interest rates will increase the return on funds invested in the bond market, and (2) funds moved to the equities market will have a higher expected return than the bond market. however, this is not the complete story. the increase in interest rates also has substantial implications for the federal government and fiscal policy. with a rise in interest rates, the cost of servicing the total federal debt will increase. the total government debt in 1999 ($5,606 billion) is currently substantially larger than the trust fund ($855 billion) and is forecasted by the cbo (2000) to be about twice as large over the next decade (government debt of $6,300 billion and trust fund of $3,325 billion in 2010). therefore, the cost to the federal government will outweigh the benefits that accrue to social security because higher interest payments are made to all bondholders, not just social security. this increase in interest costs will necessitate an increase in other tax revenues relative to taxes without reform. one implication of a change in interest rates and equity returns is a redistribution of income that would result depending on the particular tax system. first of all, there would be a transfer of income from equity holders to bond holders as interest rates increase. secondly, general income taxes would increase for bond holders, equity holders, and wage earners in order to pay for the increasing cost of debt service. overall, there would therefore be a redistribution of income from wage earners and equity holders to bondholders. this is further complicated by the fact that a substantial portion of u.s. government debt (approximately one third of publicly held debt) is held by foreigners. this creates a political dilemma in which there is distribution of income from american taxpayers (u.s. wage earners and holders of capital) to foreign bondholders. regardless of the tax scheme, moving social security funds to the equities market would necessitate higher income taxes due to the effect this type of reform has on interest rates. 6. increase the payroll tax one of the major alternatives to shifting funds is to increase the payroll tax and continue investing the full trust fund in treasury securities. note that this would substantially increase the size of the trust fund (and reduce the amount of publicly-held debt). intuitively, we would predict that an increase in the social security trust fund would cause interest rates to fall because of the decreased amount of debt the government must market to the public. along a similar, but qualitatively opposite, line of thought as above, the borrowing costs of the federal government would fall which would allow a decrease in overall taxes. this decrease in interest costs would allow for a tax reduction offsetting some of the increase in social security payroll taxes. furthermore, in this case there would be a transfer of income from foreign bondholders to american taxpayers. even though either reform regime (shifting funds to equities or increasing payroll taxes) is expected to save the social security system, 101e. elder, l. holland / financial services review 9 (2000) 93–106 the secondary effects of the two different reforms from a change in interest rates are significantly different and should be considered before any specific reform is decided upon. 7. summary and conclusion several proposals have been developed to reform the social security system in order to ensure that it is fully funded. the investment of social security funds in equities has often been proposed as a means to avoid increasing payroll taxes. the general equilibrium model developed in this paper demonstrates some of the effects of these reform proposals. we show that investing social security funds in equities will increase interest rates on bonds and decrease the return on equities. furthermore, such a shift in social security investments will lead to a necessary increase in general income taxes (assuming the level of total federal debt remains larger than the social security trust fund) and possible income redistribution. thus, to some extent, investing social security funds in equities simply shifts a potential increase in payroll taxes to an increase in income taxes. dynamic extensions of the model may be interesting to pursue by giving more insight into the long-run effects of any reform. it would be desirable to take into account the long-run changes which may occur due to lower weighted average cost of capital and subsequent increases in real assets of the firm or distortions which may occur in the labor market due to higher social security payroll taxes. appendix a. a.1. proof it is possible to rearrange (6) and (7) to get re as a function of rb andn and similarly for rb re 5 fnx 2 nxvh 1 ny 1 ny ~1 2 gy!s2 rb 1 nzh 2 nzhv 2 b exp~rb! g ny ~1 2 gy!s2 5 f~rb, v! (16) rb 5 fnxhv 1 ny ~1 2 gy!s2 re 1 nz 2 nzh 1 nzhve exp~re! g ny ~1 2 gy!s2 5 g~re, v! (17) taking the total derivatives of (16) and (17) with respect ton results in dre dv 5 frb drb dv 1fv (18) 102 e. elder, l. holland / financial services review 9 (2000) 93–106 drb dv 5 gre dre dv 1 gv (19) where gre 5 1 1 e~1 2 gy!s2 ny~exp~re!! (20) gv 5 ~nxh 1 nzh!~1 2 gy!s2 ny (21) frb 5 1 1 b~1 2 gy!s2 ny~exp~rb!! (22) fv 5 2 ~nxh 1 nzh!~1 2 gy!s2 ny 5 2gv (23) using (18) and (19) we now have the following dre dv 5 frb drb dv 2 gv (24) drb dv 5 gre dre dv 1 gv (25) which is a system of two equations in two unknowns, dre/dn and drb/dn. this system has the solution dre dv 5 gv~ frb 2 1! 1 2 gre frb (26) drb dv 5 2 gv~ gre 2 1! 1 2 gre frb (27) it can be shown that frb . 1, gre . 1, and gn . 0 and therefore dre dv , 0 (28) drb dv . 0 (29) appendix b. in our analysis, we assumed that the most risk-averse individual remained constrained in the equities market, the least risk averse individual remained constrained in the bond market, 103e. elder, l. holland / financial services review 9 (2000) 93–106 and the intermediate individual remained unconstrained following a change inn. in order to ensure these results remain valid, it is necessary to find an upper bound ongx and a lower bound ongz. in order to ensure that the most risk-averse individual is originally constrained to be in the equities market, it is necessary that, initiallynh . vx and post reformn9 h . vx9 (wherevx 5 (re-rb)/(1-gx)s 2) andvx9 5 (re9-rb9)/((1-gx)s 2) which implies gx , mins1 2 re 2 rb hn s2 , 1 2 re9 2 rb9 hn9s2 d (30) where re9 and rb9 are the post reform returns andn9 is the post reform fraction the social security trust fund allocates to equities. the least risk-averse individual is assumed to originally be constrained to hold funds in the bond market, and following reform is assumed to remain constrained. in order for this to be true, it is necessary that prereform (1-n)h . 1-vz and post reform (1-n9)h . 1-vz9 wherevz andvz9 are similar to the above. therefore, a lower bound forgz can be solved for as gz . maxs re 2 rb @~1 2 v!h 2 1#s2 1 1, re9 2 rb9 @~1 2 v9!h 2 1#s2 1 1d (31) it can be shown thatgx , gz. finally, typey individuals must be identified by agy such that maxs1 2 re 2 rb hvs2 , 1 2 re9 2 rb9 hv9s2 d , gy , mins re 2 rb @~1 2 v!h 2 1#s2 1 1, re9 2 rb9 @~1 2 v9!h 2 1#s2 1 1d (32) in order to remain unconstrained prior to and after reform. the above restrictions concerning gx, gy, gz (together with the assumption that all three individuals are risk-averse to some degree) are sufficient to show that re . 0 and re-rb . 0. appendix c. c.1. trust-fund actions do not affect individual’s decisions if individuals do not perceive funds in the social security trust fund as a portion of their savings then the trust fund can be treated as just another very large investor. therefore, an individual of typei has control overŵi þ wi of which they would optimally allocate (12 vi)ŵi to bonds andviŵi to equities. also, definessfas the amount of funds that social security controls. if there are three groups of individuals identified by concerninggx, gy, and gz with respective population sizes of nx, ny, nz, then we can defineg* which solves the equation nxvx 1 nyvy 1 nzvz 5 nv* (33) 104 e. elder, l. holland / financial services review 9 (2000) 93–106 where n5 nx1ny1nz. the solution to this equation,g*, satisfies the following equation: 1 2 g* 5 ~nx 1 ny 1 nz!~1 2 gx!~1 2 gy!~1 2 gz! nx~1 2 gy!~1 2 gz! 1 ny~1 2 gx!~1 2 gz! 1 nz~1 2 gx!~1 2 gy! (34) without loss of generality, we can assume that there is only one type of individual, identified by g* , which leads to the bond and equity market-clearing conditions ~1 2 v* ! 1 ~1 2 v!ssf5 b exp~rb! (35) v* 1 v~ssf! 5 e exp~re! (36) wherev* 5 (re 2 rb)/((1 2 g*)s2), ssf 5 ssf/ŵ, b 5 qb e[p9b]/ŵ, e 5 qee[p9e]/ŵ. by a similar methodology as above, (35) and (36) can be rearranged so that re 5 f~rb, v! (37) rb 5 g~re, v! (38) taking the total derivatives of (37) and (38) with respect ton, the resulting two equations can be solved for dre/dn and drb/dn as drb dv 5 2 gv~gre 2 1! 1 2 frb gre . 0 (39) dre dv 5 gv~frb 2 1! 1 2 frb gre , 0 (40) where frb 5 1 1 b(1 2 g*)s2/exp(rb) . 1, gre 5 1 1 e(1 2 g*)s2/exp(re) . 1, and gn 5 ssf (1-g*)s2 . 0. these results are similar to the results given in section 2 with re decreasing and rb increasing following a shift of social security trust fund money into the equities market. references advisory council on social security. (1997).report of the 1994–1996 advisory council on social security, vol. 1: findings and recommendations.washington: government printing office. american economic association. (1996).the american economic review, 86,358–377. auerbach, a., & kotlikoff, l. (1987).dynamic fiscal policy.cambridge: cambridge university press. bohn, h. (1997). social security reform and financial markets. in s.a. sass and r.k. triest (eds.),social security reform conference proceedings,federal reserve bank of boston. campbell, j. y., & mankiw, n. g. (1989). consumption, income, and interest rates: reinterpreting the time series evidence.nber macroeconomics annual, 185–216. congressional budget office. (2000).the budget and economic outlook: fiscal years 2001-2010.washington, government printing office. 105e. elder, l. holland / financial services review 9 (2000) 93–106 diamond, p. a. (1996). proposals to restructure social security.journal of economic perspectives, 10,67–88. diamond, p. a. (1997). macroeconomic aspects of social security reform.brookings papers on economic activity (forthcoming). elder, e., & holland, l. (1999). implications of social security reform on interest rates: theory and evidence. working paper; university of arkansas at little rock. federal reserve bank of boston. (1997). social security reform conference proceedings: links to saving, investment, and growth. in s. a. sass and r. k. triest (eds.),social security reform conference proceedings, federal reserve bank of boston. federal reserve board of governors. (1999). flow of funds accounts for the united states. washington, dc. friend, i., & blume, m. (1975). the demand for risky assets.the american economic review, 65(december), 900–922. geanokoplos, j., mitchell, o., & zeldes, s. (1998). would a privatized social security system really pay a higher rate of return?. in r. d. arnold, m. j. graetz, & a. h. munnell (eds.),framing the social security debate (pp.137–157). washington, dc: brookings institution press. hall, r. e. (1988). intertemporal substitution and consumption.journal of political economy, 96,339–357. ingersoll, j. e. (1987).theory of financial decision making.savage, md: rowman & littlefield publishers. kotlikoff, l., smetters, k., & walliser, j. (1998). privatizing social security in the us: comparing the options. technical paper series: macroeconomic analysis and tax analysis divisions.congressional budget office. merton, r. c. (1969). lifetime portfolio selection under uncertainty: the continuous-time case.review of economics and statistics, 51,247–257. national academy of social insurance. (1998). conference proceedings. in r. d. arnold, m. j. graetz, & a. h. munnell (eds.),framing the social security debate. washington, d.c.: brookings institution press. samuelson, p. a. (1969). lifetime portfolio selection by dynamic stochastic programming.review of economics and statistics, 51, 239–246. smetters, k. (1998). privatizing versus prefunding social security in a stochastic economy.technical paper series: macroeconomic analysis and tax divisions.congressional budget office. stein, h. (1997). social security and the single investor.the wall street journal,wednesday, february, 5, a18. stiglitz, j. e., munnell, a. h., & frankel, j. a. (council of economic advisors). (1997).economic report of the president.washington, dc: u.s. government printing office. 106 e. elder, l. holland / financial services review 9 (2000) 93–106 pii: s1057-0810(97)90027-5 from the editor karen eilers lahey volume 6, number 1, is the first quarterly issue of financial services review (fsr). it is an important move because the journal will now adhere to a standard academic publication schedule. this issue represents the culmination of many hours of hard work over an extended period of time by many members of the academy of financial services (afs). it is very fitting that the lead article is written by frank k. reilly, a past-president of afs (1990-1991) and a long-time active participant in the organization. his article entitled, "the impact of inflation on roe, growth and stock prices" examines the det rimental effect of inflation on stock returns and changes in the roe during the last 40 years. he provides readers with a clear visual picture that shows that total asset turn over and net profit margins have declined while f'mancial leverage has increased. "the congressional calendar and stock market performance" by reinhold p. lamb, k.c. ma, r. daniel pace, and william f. kennedy reveals an unusual calendar effect. their sample period is 97 years from 1897 to 1993 and focuses on the dow jones industrial average to measure returns. they find a statistically significant posi tive return for investors when congress is not in session, and this finding is not explained by the january effect. the article by robert l. albert jr., timothy r. smaby, and h. david robison is enti tled "short selling and trading abuses on nasdaq.'" it tests the effect of short selling on the nasdaq market compared to short selling on the nyse/amex markets. they find that short sellers earn significant abnormal returns by selling into rising markets and do not exacerbate price drops. further, they suggest that recent regulatory changes in short selling on the nasdaq market may not be appropriate. financial planners will be able to share the article "financial planning and college saving recommendations: let's set things straight" with their clients. thomas h. eyssell examines standard tables that show the cost of college and the savings necessary to fund a child's education. these tables appear to overstate the amount that parents must save for college. additionally, a series of tables provide for different asset allocations and assump tions about the growth rates for tuition rates in the future. o. felix ayadi in his article, "adverse selection, search costs and sticky credit card rates" examines the sensitivity of credit card rates to the costs of funds in the united states. he estimates that credit card rates adjust at a rate of 15 percent per quarter. the review feature examines two personal financial books---one a textbook by git man and joehnk called personal financial planning that is reviewed by sue greinger, and the other is a book intended for the general public by eisenberg called the money v vi financial services review 6(1) 1997 book of personal finance that is reviewed by doug kahl. in addition, john grable reviews the merrill lynch web site, and robert l. albert jr. reviews the chartei media's web site. pii: s1057-0810(96)90021-9 volume 5 number 1 1996 financial services review the journal, of individual financial management editqr karen eilers laheiy 1995-1996 waldo l. born eastern illinois university lany a. cox university of mississippi sharon a. devaney purdue university jean louis heck villanova university walter j. wcerheide rochester institute of technology associateeditors 1995-1997 raj aggrawal john carroll university james r. barth auburn university barry diskin florida state university thomas h. eyssell university of missouri, st. louis sat&a g. gustavson university of georgia andrea hueson vniversiry of miami jeff madura florida atlantic university terry l. zivney ball state university 1!395-1998 james s. ang florida state university lawrence j. gitman san diego state university douglas r. kahl university of akron james e. larsen wright state university dixie l. mills illinois state university tony plath university of north carolina at charlotte a. charlene sullivan purdue vniversiry jill lynn vihtelic saint mary’s college 0 ai jai press inc. greenwich, connecticut london, england pii: s1057-0810(00)00043-3 does education affect how well students forecast the market? john c. alexander jr.*, robert b. mcelreath department of finance, clemson university, 314 sirrine hall, clemson, sc 29634-1323, usa abstract this study examines the results of a student stock market forecasting project used in our basic and advanced investments classes. students fail to outperform a random walk model over the entire period, but do perform well in some subperiods. students receiving an above average grade in the basic investments class provide more accurate forecasts than all other groups of students. further, poorer performing students tend to be more pessimistic in their expectations of the market. the results suggest that education improves the forecasting ability of students. © 1999 elsevier science inc. all rights reserved. jel classification:d84; e37; g10; i21 keywords:market forecasts; expectations; education 1. introduction this study presents the results of a stock market prediction project that has been used in our investments classes. the project is an excellent way to get students to examine and think about what is happening in the stock market at an early point in a class. besides being a helpful tool in motivating class discussion, the results from the study are also interesting in their own right since they provide evidence on the relationship between forecasting ability and education. studies of stock market predictability have been of interest to both academics and * corresponding author. tel.:11-864-656-0547; fax:11-864-656-3748. e-mail address:alexanj@clemson.edu (j. alexander). financial services review 8 (1999) 253–260 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00043-3 practitioners alike. several studies examine the forecast performance of economists relative to the stock market (e.g., lakonishok (1980), brown and maital (1981), pearce (1984), dokko and edelstein (1989), and de bondt (1991, 1993). these studies use the economists’ forecasts as a proxy for the expectations of the sophisticated investors or smart money. this study is unique in that it examines the forecasting ability of students. our sample of students includes those who have completed a basic investments class and those who have not. thus, the results provide insight into the benefits of investment knowledge in one aspect of understanding market behavior. in addition to examining the forecast performance and bias of the students, we control for attributes such as gender and academic performance. 2. data the data used for the current study consist of three-month-ahead forecasts of the dow jones industrial average (djia) provided by college students extending from september 1991 through september 1993. these students are full-time undergraduate finance students in juniorand senior-level investments classes. the nature of the college curriculum and records allows us to control for prerequisites, gender, and academic performance. more specifically, the prerequisite for junior-level investments is basic corporate finance, and the prerequisite for senior-level investments is investments at the junior level. the prerequisite class is most often completed in the prior semester. at the start of the semester, junior-level investments students may be familiar with the djia, but typically cannot tell you its composition or level. in contrast, at the start of the senior-level investments class, the students not only know the content and level of the djia and how it moves relative to other indexes and economic variables, but also have an understanding of forecast models and random walks. we proxy the expectations of the less educated investor using the junior-level forecasts and the expectations of more educated investors using the senior-level forecasts. the students’ expectations are surveyed using the handout shown in fig. 1. this handout asks the students to provide an estimate of the djia approximately three months ahead. both the juniorand senior-level courses in the sample are taught by the same professor. this ensures that each class receives the same data on the same day, and that no biases are introduced. to provide the students with incentive to participate and to give some thought to the survey, bonus points are awarded to the closest two estimates, or anyone predicting the closing value within 0.05%. the bonus consists of 5 points added to the second exam score, which comprises 25% of the final grade. this incentive appears to work, with participation averaging 94%. the high participation is no surprise since this is basically a game with zero costs (no justification for the forecast is required) and a potential gain. one week prior to finals, the winners are announced so that students know their standing prior to sitting for the last exam. in addition to controlling for the student’s level of financial knowledge relative to investments, we control for the grade point average (gpa) at the beginning of the semester, the grade in the prerequisite class (pg), and the gender of the student. this allows us to observe the affect these variables may have on the student expectations. the two samples, consisting of 140 junior-level and 122 senior-level investments students, are very similar. 254 j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 both the juniorand senior-level samples have an average gpa of 2.75. the percentage of juniors with a prerequisite grade of greater than or equal to b is 52%, whereas the percentage for seniors is 61%. we fail to reject the null of no difference in the juniorand senior-level classes for both gpa and prerequisite grade. fig. 1. a sample of the questionaire distributed to junior and senior level investment students at the start of the semester. 255j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 3. methodology random walk forecasts are proxied by the level of the djia published in thewall street journal on the due date of the student forecasts. the relative predictive ability of the forecasts is determined using two error metrics. the mean prediction error (mpe) is defined as follows: mpe5 1/n o n51 n ~at 2 fn,t!/at (1) where: n is the number of observations or forecasts; at is the actual level of the djia for period t; and fn,t is the forecasted level of the djia by student n for period t. deflating the forecast error by the actual helps control for differences in the market level over time. the second error metric used in the analyses is the mean of the absolute values of the prediction errors (mape). while the mpe metric provides an indication of any bias in expectations, forecast errors of differing sign may cancel out each other cross-sectionally; thus, the mape metric is a better measure of accuracy. the normality of the forecast errors is tested using a kolmogorov d-statistic. we reject the null that the forecast errors are distributed normally. differences in forecast accuracy across selected groups of students is determined using a nonparametric wilcoxon rank sum analysis. in addition to the accuracy of the forecasts, we also check for any bias that the students may have in their market expectations. if students are optimistic in their forecasts, we would expect the mpe to be negative. alternatively, if they are pessimistic, we would expect the mpe to be positive. the proportion of students with a forecast less than the actual is also examined. more rigorous tests for unbiasedness use the following regression equation. at 5 ßft 1 ut (2) the intercept of the equation is constrained to equal zero. if the forecast (ft) is an unbiased prediction of the actual (at), then the estimate of ß should not differ significantly from one. if students are optimistic (i.e., expectations are biased upward), ß is less than one. in contrast, if students are pessimistic, ß is greater than one. an f-test determines whether ß is significantly different than one. 4. results table 1 summarizes the accuracy of the student forecasts compared with a random walk for the entire sample period and by semester. for the entire sample period, the random walk forecasts are more accurate than the student forecasts (mape equals 2.16% and 3.29%, respectively). the students outperform the random walk model in only one out of four semesters (spring 93). to get a better understanding of forecast accuracy it is helpful to look 256 j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 at what the djia was doing during the forecast period and before. analyzing the change in the djia during the forecast period, the largest move up in the djia was 4.60%, and the largest move down was22.58%. examining the change in the djia for the period three months prior to the forecast due date, the largest move up was 2.64%, and the largest move down was23.46%. with respect to bias over the entire sample period, both the student and random walk forecasts appear to be slightly pessimistic. when examining the data by period, we find that students tend to be pessimistic when the prior change in the djia is up, and optimistic when the prior change is relatively flat or down. this may appear in conflict with the findings of de bondt (1993) where nonexperts expect a continuation of past trends. table 2 summarizes the nonparametric wilcoxon rank sum analysis of the difference in forecasting accuracy among selected groups of students. the alternative hypothesis is that the mape is smaller for the samples listed in the second row of each designated group. examining group 1, which controls for gender, we find that the forecasts of male students are more accurate than the forecasts of female students. we find no difference in forecast accuracy for group 2, which controls for the different class levels of the students. further, although the mape appears smaller for the high gpa students relative to the low gpa students (group 3), and for students receiving a high grade in the prerequisite course relative to students receiving a low grade in the prerequisite course (group 4), the differences are not significant. we also fail to find significant differences when separating the different class levels by high and low gpa (groups 5 and 6). further, we find no difference in the forecast accuracy of juniors receiving a high or a low grade in the prerequisite class (group 7). in contrast, seniors with a high grade in the prerequisite class have better forecast accuracy than seniors with a low grade in the prerequisite class (group 8). since the lowest mape belongs to seniors with a high grade in the prerequisite class, we compare their accuracy to all other students (group 9). we find that their forecast accuracy is significantly better, at the 12% level, than all other students. the prerequisite course for junior investments students is “fundamentals of corporate finance” and the prerequisite course for senior investments students is “fundamentals of investments.” thus, it appears that the knowledge gained from an above average undertable 1 forecast accuracy of students and a random walk by period and over all periodsa period sample size student mape student mpe random walk mape random walk mpe change in djiab change in prior djiac fall 91 67 3.83% 23.10% 2.65% 22.65% 22.58% 1.02% fall 92 62 2.79 21.02 .38 .38 .38 23.46 spring 93 66 3.56 3.18 4.39 4.39 4.60 2.64 fall 93 67 3.03 2.11 1.36 1.36 1.38 2.51 all 262 3.29 .23 2.16 .80 na na a mpe5 1/n sn51 n (actual level2 forecasted level)/actual level; mape is the mean of the absolute values of the prediction errors and na is not applicable. b change in djia5 (djia on final date2 djia one day prior to due date)/djia one day prior to due date. c change in prior djia5 (djia one day prior to due date2 djia three months prior to due date)/djia three months prior to due date. 257j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 standing of investments improves the forecast accuracy of the seniors. in contrast, the knowledge gained from the corporate course had no effect on the forecast accuracy of the juniors. this differential forecast accuracy suggests that investors may benefit from an above average understanding of the financial markets prior to their entry into the market. table 3 summarizes the tests of potential bias in expectations across the different groups. the table includes the mpe, the percentage of forecasts below the actual, and the more rigorous regression results. recall that a ß of oneindicates that expectations are unbiased. evidence of pessimistic expectations would include an mpe greater than zero, more than 50% of the forecasts below the actual, and a ß greater than one. we find some evidence that students in general are pessimistic in their expectations. examining differences in expectations based on gender (group 2), we find that the expectations of females are pessimistic, whereas the males are unbiased. when controlling for the class level, we find that juniors are pessimistic and the seniors are unbiased in their forecasts of the market. this may suggest that as students learn more about the market their expectations increase. in contrast, we find that students with a low gpa (group 4) or a low grade in the prerequisite class (group 5) exhibit some pessimism, whereas the forecasts of students with a high gpa or high prerequisite grade in the prerequisite class are unbiased. thus, the observed pessimism may be more a function of an individual’s performance. additional analysis of the junior class provides evidence that this pessimism is caused primarily by those juniors having a low gpa table 2 nonparametric wilcoxon rank sum analysis of the difference in forecast accuracy among students across selected groups: the alternative hypothesis is that the distribution of absolute mean percentage errors (ampe) is smaller for samples listed in the second row of each designated groupa group samplesb sample size mape mpe z-test p-value 1) females 80 3.90% 1.08% 1.51* .07 males 182 3.03 2.14 2) juniors 140 3.25 .50 20.02 .49 seniors 122 3.37 2.08 3) gpa , 3.0 165 3.49 .47 2.55 .29 gpa $ 3.0 97 2.98 2.17 4) pg, b 115 3.60 .58 .56 .29 pg $ b 147 3.06 2.04 5) juniors, gpa, 3.0 90 3.36 .73 2.19 .42 juniors, gpa$ 3.0 50 2.99 .08 6) seniors, gpa, 3.0 75 3.63 .16 2.80 .21 seniors, gpa$ 3.0 47 2.96 2.45 7) juniors, pg, b 67 3.22 .81 2.54 .29 juniors, pg$ b 73 3.24 .21 8) seniors, pg, b 48 4.14 .26 1.48* .06 seniors, pg$ b 74 2.88 2.29 9) all others 188 3.46 .44 21.10 .12 seniors, pg$ b 74 2.88 2.29 a mpe5 1/n sn51 n (actual level2 forecasted level)/actual level; mape is the mean of the absolute values of the prediction errors. b gpa is the grade point average taken at the beginning of the semester, and pg is the letter grade received in the prerequisite course. * significant at the 10% level. 258 j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 (group 6). comparing the results from tables 2 and 3, we find that those student groups exhibiting the best forecast accuracy also tend to be unbiased in their expectations. 5. conclusions using student forecasts of the djia, we find that students fail to outperform a random walk forecast over the entire sample period. more specifically, the students only outperform a random walk model in one out of four semesters. comparing the forecast accuracy among different groups of students, we find that senior-level students who received an above average grade in the “fundamentals of investments” class provide more accurate forecasts than all other students in the sample. the forecasting ability of these seniors may result from their investment knowledge, which may allow them to apply more sophisticated information and techniques in estimating the djia. this finding suggests that an above average understanding of investments prior to entering the market may benefit investors. with respect to any bias in expectations, the junior-level and female students in our sample tend to be pessimistic in their expectations, whereas the senior-level and male students tend to be unbiased. the pessimism of the juniors can be attributed to poorer performing students (measured by their gpa and grade in prerequisite class). this suggests table 3 tests for the biasedness of student forecasts: the null hypothesis of unbiasedness is that the coefficients associated with ß will not differ significantly from 1 (regression equation: actual value5 ß [forecasted value]) group samplesa mpeb % underc ß f-test p-value 1) all students .23% 47.3% 1.003 1.76 .19 2) females 1.08 58.8 1.012 3.68* .06 males 2.14 42.3 1.000 .00 .99 3) juniors .50 51.4 1.007 4.58* .03 seniors 2.08 42.6 .999 .01 .94 4) gpa , 3.0 .47 47.3 1.006 2.17 .14 gpa $ 3.0 2.17 47.4 1.000 .01 .93 5) pg, b .58 46.1 1.007 2.08 .15 pg $ b 2.04 48.3 1.001 .11 .75 6) juniors, gpa, 3.0 .73 51.1 1.009 4.14* .04 juniors, gpa$ 3.0 .08 52.0 1.004 .60 .44 7) seniors, gpa, 3.0 .26 42.7 1.002 .06 .81 seniors, gpa$ 3.0 2.29 42.6 .997 .37 .54 8) juniors, pg, b .81 53.7 1.011 4.38* .04 juniors, pg$ b .21 49.3 1.004 .85 .35 9) seniors, pg, b .26 35.4 1.002 .06 .80 seniors, pg$ b 2.29 47.3 .997 .25 .61 a gpa is the grade point average taken at the beginning of the semester, and pg is the letter grade received in the prerequisite course. b mpe 5 1/n sn51 n (actual level2 forecasted level)/actual level. c % under is the percent of forecasts that are below the actual. * significant at the 10% level. 259j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 that the expectations of investors may be influenced by their demographics, exposure to information, and/or their job performance. acknowledgments the authors thank donald nast, the originator of the student questionaire, and the reviewers for their helpful comments. references brown, b. w. & maital s. (1981). what do economists know? an empirical study of experts’ expectations. econometrica, 49,491–504. de bondt, w. f. (1991). what do economists know about the stock market?journal of portfolio management, 17, 84–91. de bondt, w. f. (1993). betting on trends: intuitive forecasts of financial risk and return.international journal of forecasting, 9,355–371. dokko, y. & edelstein r. (1989). how well do economists forecast stock market prices? a study of the livingston surveys.american economic review, 79,865–871. lakonishok, j. (1980). stock market return expectations: some general properties.journal of finance, 35, 921–930. pearce, d. k. (1984). an empirical analysis of expected stock price movements.journal of money, credit, and banking, 16,317–327. 260 j.c. alexander, r.b. mcelreath / financial services review 8 (1999) 253–260 pii: s1057-0810(99)00032-3 computerized stock screening rules for portfolio selection steven c. gold*, paul lebowitz department of finance, accounting and mis, rochester insitiute of technology, college of business, max lowenthal bldg., 1 lomb memorial drive, rochester, ny 14623, usa abstract recent studies have uncovered several systematic patterns that increase the probability that individual investors can select stock portfolios with excess returns. this study tests the feasibility of using a commercially available computerized stock screening program for investors to take advantage of these patterns. the screening program searches the three major exchanges and selects stocks on both fundamental and technical indicators: low price-to-sales ratio, small firm size, accelerating stock prices above their 50 day moving average, high trading volume, and high earnings growth. of the 18 models tested between 1994 and 1998, those that allow for selection between exchanges yield portfolio returns that significantly exceed the average market indices. © 1999 elsevier science inc. all rights reserved. 1. introduction stock screening programs, similar to the ones used by professional portfolio managers, are becoming available on the internet at little or no cost to the individual investor. anderson (1998), in an article “screening for investment gold,” overviews four such sites with relatively sophisticated programs. stock screening programs make it easier and quicker to tailor a portfolio to fit the desired style and preferences of the investor. styles may range from investing for value to growth. but whatever the style, the goal is to develop a screening program to find systematic patterns that will increase the performance of a portfolio. this paper investigates the use of one highly ranked investor-screening platform from telescan, an easy-to-use software package that provides nearly 300 screening variables that cover both fundamental and technical indicators. the user can customize a model to include * corresponding author. tel.:11-716-475-2318; fax:11-716-475-6920. e-mail address:scgbbu@rit.edu (s.c. gold) financial services review 8 (1999) 61–70 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00032-3 up to 40 different indicators simultaneously and can scan on relative values, e.g., the stocks with the highest relative strength, or on fundamental variables, like earnings per share momentum. according to the efficient market hypothesis, past price movements in a competitively traded financial market do not help in predicting future prices. however, many recent articles question the efficient market hypothesis and support the notion that stock market excess returns can be predicted by publicly available information (e.g., see gencay, 1996; fama and french, 1995; pesaran and timmerman, 1995; ferson and harvey, 1993). although it is still commonly believed the u.s. stock market is semi-strong efficient, walker and hatfield (1996) argue that financial research indicates that security prices do not reflect all publicly available information. for this reason, investors may find it worthwhile to use computerized screening programs that take advantage of large databases and advances in information technology to efficiently select stocks. a number of studies on the profitability of filter rules on exchange rates lend credibility to this hypothesis (e.g., levich and thomas, 1993; taylor, 1994). financial gains from the use of computerized screening programs are consistent with the weak form of the efficient market hypothesis offered by jensen (1978), i.e., prices reflect information up to the point where the marginal benefits equal the marginal costs of the information. the marginal costs of making informed investment decisions have been declining significantly due to high-speed and low-cost computer technology. this lower marginal cost may account for the increase in short-term speculative trading by individual investors. the recent literature clearly indicates that benefits, in the form of excess returns, can be realized by taking advantage of systematic patterns that seem to exist in the stock market, e.g., reaction and drift effects, earnings and forecast surprise effects, and performance persistence. a well-known study by jegadeesh and titman (1993), for example, documents that investment strategies which buy stocks that have performed well in the past and sell stocks that have performed poorly in the past generate significant positive returns over 3 to 12 month holding periods. this outcome is consistent with delayed price reactions to firm-specific information. the following sections involve: (1) reviewing investment strategies for portfolio selection that have been shown to yield superior returns; (2) combining these superior strategies in a screening model to explore the possibility of any synergy gains; and (3) testing the model to assess its performance relative to the market indices. 2. promising investment strategies a review of the recent finance literature identifies several causal variables predicting excess returns in the market. this section will discuss these factors and their implications for use in a stock-screening model. 2.1. price/sales ratios researchers have studied several indicators to identify undervalued stock and predict excess returns. these include relatively low values for price-to-earnings, price-to-book, and 62 s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 price-to-sales ratios. these researchers argue that stocks that have low values for these measures are not currently popular with investors and create a potential for greater price appreciation. recent evidence suggests the sales-to-price ratio is a more reliable indicator than other measures of undervalued stock, so this ratio is utilized in the screening model in this study. barbee et al., (1996) offers three reasons why the sales-to-price ratio may be a more reliable predictor for stock returns than p/e ratios. first, annual sales historically are a better indicator of long-run expected profits than current reported profits. second, earnings are more likely to be affected by short-term policies than sales revenues. sales figures tend to be less subject to manipulation. third, sales-to-price cannot have negative values as can p/e. vandell (1986) argues that focusing primarily on p/e could result in two types of investment errors: (1) avoiding firms with low earnings that have a temporarily high p/e ratio but are expected to grow profitably in the future, and (2) selecting cyclical stocks when their p/e ratios are low but their profits are at their peak. the price-to-book ratio (the inverse of book-to-market value) has received some attention in the literature, but the reviews are mixed. a low price-to-book ratio may select a company with low earnings prospects and greater risk (kothari, shanken, and sloan, 1995). 2.2. earnings momentum earnings prospects are considered to be an important indicator of stock returns. a study by vandell (1986) develops a screening model for stock selection based on earnings momentum. this study finds that screening on low price-to-earnings values would be successful in predicting returns only if earnings-per-share expectations are high. more recently, tam, kiang, and chi (1991) employ an induction methodology to analyze a database for commonalties. they find changes in quarterly earnings as one of the driving variables predicting the potential for excess returns. similarly, harris and marston (1994) find it necessary to control for earnings prospects to predict excess returns. earnings momentum in this study is measured as the weighted average of quarterly growth rates in earning per share (eps) over the past year. the recent quarters are given the most weight. 2.3. market capitalization several studies support an inverse relationship between small firm size or market capitalization and stock returns. market capitalization is the firm’s stock price multiplied by the number of common shares outstanding. lo and mackinlay (1988) find positive autocorrelation between weekly returns and stock portfolios grouped by size. dennis et al., (1995) find a significant relationship between firm size, book-to-market equity and excess returns. the optimal portfolios are those with the smallest firms and highest book-to-market equity. the portfolios with the largest firms and smallest book-to-market equity underperformed the market. 63s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 2.4. price and volume reactions reinganum (1988) identifies relative price as one of the primary characteristics of the 222 stock market winners between 1970 through 1983. relative price performance is a weighted average measure of past price changes of each stock in comparison to all stocks. advocates of this measure contend that a stock will generally lose relative strength before a significant drop in price occurs. relative price strength is a significant factor in selecting successful stock portfolios in studies by tam, kiang, and chi (1991) and jegadeesh and titman (1993). recent studies suggest that an effective investment strategy needs to examine movements in both price and volume. kim and verrecchia (1991) argue that price changes are associated with the market’saveragebeliefs, while trading volume is thesumof all individual trades. hiemstra and jones (1994) suggest that volume serves as a proxy for information flow and that a positive relationship exists between trading volume and absolute stock returns. stickel and verrecchia (1994) present evidence that price changes are more likely to be reversed following low trade volume than high volume. they argue high volume reflects a greater probability that the trading stems from informed investors. trading volume in this study is measured by an accumulation distribution over the past 50 days. for each of these days, the stock’s volume is multiplied by the change in price and summed. this indicates the amount of money moving into the stock. 2.5. moving average rules technical analysts have developed numerous moving average rules. the basic rule involves calculating a moving average of past prices. the length of the period used in the moving average is commonly between 20 and 200 days. the length selected by the investor reflects the time horizon being predicted. this study uses a 50-day moving average. the moving average model assumes that there are systematic patterns in market prices that can be used in forecasting. if the current price moves above the moving average (or some band about the average), a buy signal is generated. in this situation technical analysts believe that the mood has changed from a declining to a rising stock pattern. if the current price falls below the moving average (band), a sell signal occurs. until recently, technical trading rules involving moving averages were not considered to have much forecasting value. however, advances in computer technology have increased the sophistication of such models and new evidence is now being uncovered. gencay (1996) shows substantial gains in forecast accuracy through the use of such models with a forecast horizon of 20 days. these findings are also consistent with studies showing evidence of systematic patterns in daily, weekly, and monthly returns (see haugen and jorian, 1996; cutler, poterba, and summers, 1991). 3. screening model and methodology the screening model in this study attempts to find a portfolio of stocks that are not only under-valued but have relatively high growth potential. this model is developed by combining the superior investment strategies previously reviewed. possible synergy gains are 64 s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 derived from the interaction of these strategies using fundamental, industry sector, and technical information. the screening model usesa three-step approach. first, fundamentalvariables are used to screen for smaller companies with low price-to-sales ratios, positive returns on equity, and high earnings per share growth. second,strong industrysectors are filtered, searching for positive relative price performance, because those industries do better under certain macroeconomic conditions and are positioned for future growth. stocks within these industries are then screened for high relative price performance. third,technicalfilters are used to select stocks that are over their 50-day moving average price and have a high accumulation distribution (trading volume activity measured in dollars). a commercially available computerized search and screening program, prosearch from telescan, inc., is used to select and rank order a portfolio of stocks that most closely meet a set of criteria based on the most promising screening indicators identified in the finance literature. a methodological advantage of a screening program is that it does not require the specification of a functional form to predict returns, as is the case with linear or non-linear regression analysis. the telescan database universe consists of 9, 000 listed stocks on the nyse, amex, and nasdaq exchanges and is restricted by telescan to the most recent three-year moving window. several investment models are tested between the time frame november 1, 1994 to august 31, 1998. the rolling results cover almost four years. this period encompasses a long bull market and the sell off from theasian contagion, thereby including a mix of market conditions. the investment strategy is designed to pick a portfolio of stocks, between 20 and 50, and rebalance on a quarterly basis. the number of different stocks is kept relatively small to make the portfolio financially accessible to the individual investor. the specific search criteria is defined below. at the beginning of each quarter, the search program is executed and selects for purchase the top stocks from the universe that best fit the screening criteria. dollars invested are distributed equally between the selected stocks. the total return (net of transactions costs) for the selected portfolio is compared to the market return at the end of the quarter. the portfolio is liquidated and a new search is made on the same criteria. dollars are once again invested equally among selected stocks in the new portfolio. transaction costs assume electronic trading with average fees between $7.50 and $15.00 per trade. 4. screening criteria the desired portfolio is selected using a two-step screening criteria. first, the stocks are filtered based on the criteria values in table 1. those firms that do not meet all the criteria in table 1 are eliminated from the potential portfolio. second, after filtering on the above criteria, the remaining stocks arerank orderedbased on a weighted average index of the percentile rank of the indicators in table 2 and their respective weights. because all criteria are weighted equally (at 100%), the index used to rank each stock is simply the average of the percentile ranking of each indicator. the ranking is done by the prosearch program once the indicators, criteria and weights are specified. 65s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 5. results the stock screening program is used to select a portfolio of stocks that best fit these criteria on a quarterly basis. although only 3 to 4 years of data are available, 18 different models are tested within that timeframe and are summarized in table 3. models 1 to 9 search for the best stocks each quarter in all three exchanges, but have different time intervals and portfolio size, ranging between 20 and 50 different stocks. models 10 through 18 restrict the searches to a specific exchange. the start date depends on the available three year moving window in telescan’s database, with the end date restricted to the most recent full quarter of data given the start date. data are available from november 1994 for model 1, but are no longer accessible when the remaining models are tested. returns are calculated net of round trip transactions costs, assuming that trades are transacted electronically, without the direct use of a broker. fees for trading electronically vary between $7.50 and $15.00 per transaction and are assumed to average $10.00. assuming a 100% turnover per quarter, transaction costs for a portfolio with 20 to 50 different stocks range between $400 and $1, 000 quarterly. overall transaction costs are calculated to average just under 1% of the portfolio, with an average portfolio size between $40, 000 and $100, 000. the quarterly average portfolio net returns for models 1 through 9 are compared to the average return in all three exchanges, i.e., the nyse, amex, and nasdaq. in all cases, the portfolio net returns exceed the average index returns (without costs). the results are statistically significant in 7 out of the nine models. models 10 through 18, that arbitrarily restrict the program to select stocks within one table 1 screening criteria indicator screening criteria price/sales ratio between 0 and 1 return on equity value greater than zero earnings per share momentum highest 50% of stocks in market price performance highest 50% of stocks in market industry price performance highest 50% of stocks in market accumulation distribution highest 50% of stocks in market table 2 ranking criteria indicator rank criteria weight price/sales ratio lowest percentile in the market 100% earnings per share momentum highest percentile in the market 100% market capitalization lowest percentile in the market 100% price performance highest percentile in the market 100% industry price performance highest percentile in the market 100% moving avg. ratio50 days highest percentile in the market 100% accumulation distribution highest percentile in the market 100% 66 s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 exchange, do not perform as well. although excess returns are still positive in 8 out of the 9 models, only two are statistically significant at the 0.05 level of confidence. this result is not unexpected because the flexibility of the model to choose between markets and industries is limited. this result also confirms the power of the unrestricted screening model to switch between exchanges to select the most promising companies. although most excess returns in models 10 through 18 are not significant, the binomial proportionality test statistic associated with all nine models together shows statistical significance at the 0.01 level. the null hypothesis tested assumes the simple fraction of profitable models (those with positive excess returns after the payment of transaction costs) should equal 0.5. the null hypothesis is rejected given 8 out of 9 models with positive excess returns. further evidence of support for the screening model may be found by comparing the performance of fund managers to the market indices. for the past five consecutive years, bary (1999) finds fund managers underperformed the s & p 500. in 1998, 86% of fund managers trailed the s & p 500, whereas 90% lagged the index in 1997. treanor (1999) reports the combined actuarial performance service (caps) median performance of pension fund managers for the past five years is only 13.1%, well below client requirements. the superior portfolio returns of the screening model could not be explained by risk. the mean of the portfolio betas (obtained from the telescan database) for models 1 to 9 is shown table 3 parametric tests of the screening model model exchange start date end date no. of stocks quarterly average portfolio net return quarterly average exchange return quarterly excess return test statistic 1 all 11/1/94 4/30/98 35 8.9% 5.3% 3.6% 1.78b 2 all 12/1/95 8/31/98 35 7.7% 2.7% 5.0% 2.75b 3 all 1/1/95 6/30/98 35 6.2% 3.5% 2.7% 2.09b 4 all 11/1/95 7/31/98 20 6.8% 4.8% 2.0% 1.36 5 all 12/1/95 8/31/98 20 7.7% 2.7% 5.0% 3.05b 6 all 1/1/96 6/30/98 20 6.8% 3.5% 3.3% 2.49b 7 all 11/1/95 7/31/98 50 6.4% 4.8% 1.6% 1.27 8 all 12/1/95 8/31/98 50 9.1% 2.7% 6.4% 3.32b 9 all 1/1/96 6/30/98 50 6.2% 3.5% 2.7% 1.95b 10 nyse 11/1/95 7/31/98 35 5.7% 5.4% 0.3% 0.93 11 nyse 12/1/95 8/31/98 35 4.4% 3.9% 0.5% 0.99 12 nyse 1/1/96 6/30/98 35 5.1% 4.4% 0.7% 1.08 13 amex 11/1/95 7/31/98 35 6.9% 2.7% 4.2% 1.78 14 amex 12/1/95 8/31/98 35 6.3% 0.7% 5.6% 2.12b 15 amex 1/1/96 6/30/98 35 4.0% 1.3% 2.7% 1.36 16 nasdaq 11/1/95 7/31/98 35 6.2% 6.3% 20.10% 0.87 17 nasdaq 12/1/95 8/31/98 35 9.7% 3.6% 6.1% 2.26b 18 nasdaq 1/1/96 6/30/98 35 6.7% 5.0% 1.7% 1.11 mean 6.7% 3.7% 3.0% a in models 1–9 ‘‘all’’ refers to exchanges including the nyse, amex, and nasdaq. models 10–18 are restricted to specific exchanges. portfolio returns are net of transactions costs. b indicates statistical significance at the 0.05 level. 67s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 in table 4. the average beta of the portfolios is only 0.72. although not shown in the table, the highest portfolio beta is 0.99 and the lowest portfolio beta is 0.48. the betas may be biased downward owing to the selection of relatively smaller firms in the screening criteria. scholes and williams (1977) suggest beta adjustments to account for non-synchronous trading. however, examination of the data indicates that this is not a problem. firm size in the portfolios, although below the market mean, have market capitalization averaging $531 million, ranking firms in the 32nd percentile. high trading volume is also a selection criteria. the average firm in the portfolio is ranked in the 85th percentile in terms of accumulation distribution. the other characteristics of the portfolio, displayed in table 4, are consistent with the a priori screening criteria. the price-to-sales ratio averages 0.44, which is in the lower 24% of the market in ranking, considering all exchanges (nyse, nasdaq, and american). the price-to-earnings (p/e) and price-to-book ratios are not considered in the selection criteria, but average in the middle range of the market, i.e., 57–60% ranking. the size of the firms in the market, measured by market capitalization, is in the lower end. the average firm’s total market capitalization in the portfolio is $531.42 million, which places it in the lower 32% in market rank. the earnings per share are $0.61, which is in the middle of the market in percentile ranking at 59%. however, the portfolio roe is relatively high, placing it at 91% in market rank. price performance, that measures the relative price increases over a year, is in the top 90% of all stocks in the market. the moving average price ratio ranks in the top 89%, confirming that the price at selection is relatively high in relation to its 50-day moving average. trading volume measured as accumulation distribution is relatively high, ranking in the top 85%. the selected exchanges are shown at the bottom of table 4. most stocks in the portfolios are selected consistently from the nasdaq exchange. the nasdaq averaged 68.34% of the portfolio dollars, but it is as low as 49.4% in some quarters and as high as 86.1% in others. on average, only 12.89% of the portfolio stocks are from the nyse and 18.77% from the ame. this tracks well with the ranking criteria that gives priority to industries with high table 4 average portfolio characteristics: models 1–9 portfolio mean percentile (%) rank in market beta 0.72 — price/sales 0.44 24 p/e ratio 20.73 60 price/book 2.08 57 capitalization (mil$) 531.42 32 roe 0.46 91 earnings/share ($) 0.61 59 price/rank (%) 118.50 90 ma price ratio (%) 115.35 89 accum. distr. (mil$) 79.87 85 nasdaq (%) 68.34 — ame (%) 18.77 — nyse (%) 12.89 — 68 s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 relative price growth. because the nasdaq had a concentration of high growth stocks during the study period, it confirms the power of the model to switch to high growth industries. 6. conclusions advances in communication and technology now place powerful screening tools in the hands of serious investors. they can rapidly search large databases in real time with complex screening rules to select the best stocks in the universe that fit a set of criteria and they can find results in seconds. the costs of using professional screening programs have been declining, making these programs accessible to the individual investor. the screening model in this study is robust enough to work well with as little as 20 different stocks in the portfolio, requiring an investment of only $40, 000. without considering any tax consequences, the portfolio returns of the screening models significantly exceed the average exchange indices and are almost twice the median returns achieved by professional fund managers. what accounts for the success of the screening model in this study? four considerations are important. first, the evidence is consistent with delayed price reactions to firm-specific information. perhaps the market is slow to react because large institutional investors’ costs (both direct and indirect) are too high, leaving an opportunity for the efficient individual investor to move more rapidly by taking advantage of recent developments in information technology. second, the indicators are selected based on a review of the empirical literature. only included are those indicators that proved successful in prior studies. third, both fundamental and technical filters are linked to take advantage of synergistic effects. finally, strong industry sectors are given priority to account for current and changing macroeconomic conditions. although the results are favorable, the data is limited, covering only a three-to-four year time horizon. the time interval includes the asian crisis and a market decline, but most quarters are in a growth market. more work is needed to confirm the model, especially in declining markets. references anderson, j. a. (1998). screening for investment gold.black enterprise, 29(3), 93–97. bamber, l. s.(1995). differential price and volume reactions to accounting earnings announcements.accounting review, 70(3), 417–441. barbee, w. c., jr., mukherji, s., & raines, g. a. (1996). do sales-price and debt-equity explain stock returns better than book-market and firm size?financial analysts journal(march/april), 56–60. bary, a. (1999). the trader: indexers trounce active managers and roil market, too.barrons, 79(1), 3–5. cutler,d. m., poterba, j. m., & summers, l. h. (1991). speculative dynamics.review of economic studies, 58, 529–546. dennis, p., perfect, s. b., snow, k.n., & wiles, k. w. (1995). the effects of rebalancing on size and book-to-market ratio portfolio returns.financial analysis journal(may-june), 47–56. fama, e. f., & french, k. r. (1995). size and book-to-market factors in earnings and returns.the journal of finance, 50(1),131–155. 69s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 ferson, w. e., & harvey, c. r. (1993). the risk and predictability of international equity returns.review of financial studies, 6, 527–566. gencay, r. (1996). non-linear prediction of security returns with moving average rules.journal of forecasting, 15, 165–174. grundy, b., & mcnichols, m. (1989). trade and the revelation of information through prices and direct disclosure.review of financial studies, 2, 495–526. harris, r., & marston, f. (1994). value versus growth stocks: book-to-market, growth, and beta.financial analysts journal, 50(5). haugen, r., & jorion, p. (1996). the january effect: still there after all these years.financial analysts journal, (jan.-feb.), 27–31. hiemstra, c., & jones, j. d. (1994). testing for linear and non-linear granger causality in the stock price-volume relation.the journal of finance, xlix(5), 1639–1664. jegadeesh, n., & titman, s. (1993). returns to buying winners and selling losers: implications for stock market efficiency.the journal of finance, xlviii(1), 65–91. jensen, m. c. (1978). some anomalous evidence regarding market efficiency.the journal of financial economics, 6, 95–101. kim, o., & verrecchia, r. (1991). trading volume and price reactions to public announcements.journal of accounting research, 29,302–321. kothari, s. p, shanken, j., & sloan, r. g. (1995). another look at the cross section of expected stock returns. the journal of finance, 50(1), 185–224. levich, r. m., & thomas, l. r. (1993). the significance of technical trading-rule profits in the foreign exchange market: a bootstrap approach.journal of international money and finance, 12, 451–474. lo, a. w., & mackinlay, a. c., (1988). stock market prices do not follow random walks: evidence from a simple specificationtest.review of financial studies 1, 41–66. pesaran, h. .m., & timmermann, a. (1995). predictability of stock returns: robustness and economic significance.the journal of finance, l(4), 1201–1228. reinganum, m. r. (1988). the anatomy of a stock market winner.financial analysts journal,(march/april), 16–28. scholes, m., & williams, j. (1977). estimating betas from nonsynchronous data.journal of financial economics, 5, 309–327. stickel, s. e., & verrecchia, r. e. (1994). evidence that trading volume sustains stock price changes.financial analysts journal, 50(6), 57–67. tam, k. y., kiang, m. y., & chi, r. t. h. (1991). inducing stock screening rules for portfolio construction. journal of the operational research society, 42(9), 747–756. taylor, s. j. (1994). trading futures using a channel rule: a study of the predictive power of technical analysis with currency examples.journal of futures market, 14, 215–235. treanor, j. (1999, march 21–22). most fund managers fail the test.the guardian. vandell, r. p. (1986). a purposeful stride down wall street.journal of portfolio management, 12(2), 31–39. walker, m. m., & hatfield, g. b. (1996). professional stock analysts’ recommendations: implications for individual investors.financial services review, 5(1), 13–29. 70 s.c. gold, p. lebowitz / financial services review 8 (1999) 61–70 pii: s1057-0810(99)00002-5 from the editor karen eilers lahey i am delighted to tell you that three of the articles that are being published in this issue are recipients of the paper awards given for manuscripts presented at the 1998academy of financial servicesannual meeting. the first award is provided bythe association of individual investors and is given to dale domian, david a. louton, and charles e. mossman for their presentation entitled, “the rise and fall of the ‘dogs of the dow.’” they provide an explanation of why this simple decision rule for investing in the dow stocks originally worked and no longer is successful. the second award is provided by theamerican college and the recipients are chris robinson and elton g. mcgoun for their presentation entitled, “the sociology of personal finance.” this theoretical piece discusses how individuals view the subject of money and the rational behavior assumption that underlies the finance discipline’s view of personal and corporate finance. hopefully, it will provide you with a fresh look at this important question. the third award is provided bytexas instrumentsand the winners are yoonkyung yuh, sherman hanna, and catherine phillips montalto. their presentation is entitled, “mean and pessimistic projections of retirement adequacy.” they use available data to analyze household financial preparation for retirement using a ratio of projected wealth and a pessimistic projection. the fourth article is written by william reichenstein and is entitled, “calculating a family’s asset mix.” he examines the question of the appropriateness of including the ownership of a mortgaged home as an asset. this paper compliments the presentation by yuh, hanna, and montalto which looks at retirement adequacy that includes home ownership with and without a mortgage the last article is entitled, “credit union structure: an examination of potential risks.” in it, robert j. boldin, keith leggett and robert strand, examine the risk faced by individuals who place their savings in credit unions. financial services review 7 (1998) v 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00002-5 pii: s1057-0810(00)00054-8 beliefs and actions: expectations and savings decisions by older americans harold w. elder*, patricia m. rudolph department of economics, finance & legal studies, university of alabama, box 870224, tuscaloosa, al 35487, usa abstract to understand the interaction of savings behavior, pension fund participation and expectations of retirement well being, we ask two questions. are expected pension benefits a substitute for accumulated savings in replacing preretirement income? are individuals’ expectations concerning their retirement standard of living realistic based on their accumulated savings and pension plan participation? first-wave data from the health and retirement study (hrs) are analyzed using a probit regression. the results are consistent with the idea that pension benefits are substitutes for saving and that accumulated savings have a significant impact on the expected standard of living but pension plan participation does not. © 2000 elsevier science inc. all rights reserved. jel classification:j14; d1; e21 keywords:savings; consumption; pensions 1. introduction survey data as well as anecdotal evidence indicate the existence of a significant discrepancy between the expectations of economic well being in retirement and the savings behavior of americans. the ability of a retiree to maintain his preretirement standard of living depends on his ability to replace preretirement income from a combination of social security, private retirement benefits and income from investments. social security benefits serve as the * corresponding author. tel.:11-205-348-8976; fax:11-205-348-0590. e-mail address:helder@cba.ua.edu (h.w. elder). financial services review 9 (2000) 33–45 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00054-8 primary source of retirement income for many, but for most—especially higher income individuals—those benefits do not provide sufficient income to maintain the preretirement standard of living. income from accumulated savings, liquidation of capital and private pension benefits must fill the gap. the divergence of individual expectations and the level of accumulated savings has at least two potential sources. the first is that people do not perceive the need to accumulate savings, because they expect pension benefits to “fill the gap.” the second is that people have unrealistic expectations concerning their likely standard of living in retirement. in this paper, data from the first wave of the health and retirement study (hrs) are used to analyze savings behavior and the formulation of expectations concerning retirement well being. the hrs is an excellent resource for the examination of these issues, since it is a longitudinal database of households in which at least one member was age 51 to 61 at the time of the start of the survey (1992). the data contained in the survey include information on economic and demographic characteristics, as well as detailed information on individual health and disabilities, employment history, and retirement plans, including the individual’s expectations about his standard of living in retirement. the first section of our analysis focuses on past savings behavior as reflected in accumulated savings. if individuals use savings and pension participation as a means of smoothing consumption over their lifetimes, then savings and pension participation may be substitutes. we assess the impact on accumulated savings of pension participation, as well as such economic and demographic characteristics as income, occupation, education, race, marital status and number of children. the second section of the paper focuses on whether individuals have realistic expectations of their standard of living after retirement. if individuals have realistic expectations of the effect of retirement on their standard of living, then factors that affect their ability to replace their preretirement income should be systematically related to their expectations. the impact of accumulated savings and pension participation on the individual’s expectations is tested in an ordered probit model that includes demographic characteristics. our results are consistent with the notion that individuals view pension plan participation as an alternative to savings even when differences in income, education, gender, age, number of children and health are taken into account. in forming their expectations of retirement well being, the results indicate that individuals do take into account their own levels of accumulated savings but do not consider pension plan participation. these results can potentially provide useful information about the extent of the divergence between beliefs and actions in savings decisions. it can also provide insights as to the impact of this behavior on the economic status of the upcoming cohort of older americans. such information could play an important role in public policy making, particularly as it relates to the proposals to reform the social security system. 2. the analysis of savings, consumption and retirement the literature that addresses the issues surrounding savings and retirement is vast. hurd (1990) and poterba (1996) provide good overviews of the literature, along with insights into 34 h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 the direction of future research. some researchers like lazear (1994) focus on the level of aggregate savings and the potential impact on economic growth. others, like cantor and yuengert (1994), are more interested in the adequacy of individuals’ saving for retirement. regardless of the perspective, most of the work in this area has the life-cycle theory of consumption and savings as its theoretical foundation. according to the life-cycle theory, individuals make decisions to maximize lifetime utility. thus, during years when income is earned, savings decisions are made so that part of the income can be consumed after retirement. the main goal of saving is to smooth consumption over the lifetime. this theory takes the view that individuals have foresight and the planning capability to project future income streams, future price changes, future expected returns on investments, and to make utility maximizing decisions. this leads directly to the first question addressed in this paper: are pension benefits a substitute for accumulated savings? because the goal of saving is to maintain the preretirement standard of living, many researchers focus on the replacement ratio, the ratio of postretirement income to preretirement income. retirement income is provided by social security benefits and pension benefits as well as income generated by accumulated saving. if the individual approaches retirement with a desired replacement ratio, the amount of saving she will need to have accumulated will be less if pension benefits will replace part of the preretirement income. to attempt to answer this question, we look at the impact of participating in a pension fund on the level of accumulated savings. the amount of savings necessary to replace an acceptable portion of income will be related to the level of income, how much income will be replaced by pension benefits and the time remaining to save. demographic and life situation (marital status, number of children and health) may also be systematically related to accumulated savings. in the second part of our analysis, we address the question: do individuals have realistic expectations of their economic well being in retirement? two recent papers by banks, blundell and tanner (1998) and bernheim, skinner and weinberg (1997) raise questions about how realistic individuals are in the formation of their expectations. they consider the consumption drop at retirement as evidence that the individuals may not be the far-sighted, utility maximizers represented in the life-cycle theory. banks, blundell and turner explain the observed drop in consumption in terms of the “systematic arrival of adverse information,” while bernheim, skinner and weinberg write of a “surprise” at retirement. either explanation would imply that individuals’ expectations are not realistic. whether an individual expects to be better off, the same, or worse off after retirement should be based on his ability to replace preretirement income and continue preretirement levels of consumption. to form realistic expectations, the individual should consider her claim on retirement benefits and accumulated saving. these variables along with demographic and life situation variables should have an impact on the expected living standard in retirement. 35h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 3. data and empirical model the hrs is a survey specifically designed by an interdisciplinary panel to gather a broad range of information pertinent to savings and retirement decisions. the national institute on aging (nia) is the sponsoring organization and the data are collected by the institute for social research at the university of michigan. juster and suzman (1995) provide an introduction to the data in a supplemental issue of thejournal of human resourcesthat is devoted to the hrs. current information about the data collection process and the data itself can be downloaded from the institute for social research web site, http://www.umich.edu/ ;hrswww/. the survey includes questions on retirement planning, pension participation, net worth, income and employment history, as well as health status and familial relationships. alternative data sets, such as the panel study of income dynamics and the surveys of consumer finances, were designed to focus on households’ economic decisions and, accordingly, do not include questions about many important noneconomic factors. in this paper, the first wave of the hrs, collected in march 1992, is used. to be age-eligible at that time a person had to be between 51 and 61 years of age. partners were also interviewed, so some of the respondents are not “age eligible.” from the 12,652 total respondents, 5,600 were age-eligible and answered that they were “not retired at all.” of those, 4,978 had complete data and are included in our analysis. the subset of the hrs data used in our statistical analysis includes income and net worth information, expected standard of living in retirement and demographic information. definitions, means and standard deviations for the variables used in this study are contained in table 1. in the hrs, income and wealth information is collected on a household level, so both the respondent’s and the partner’s (if there is one) income and net worth are included. the individual’s expected standard of living reflects that one person’s perception. demographic information is also on an individual level. 3.1. wealth and income measures the household income figure includes income from all sources—earnings, pensions and social security, investment income, and welfare payments. accumulated savings or net worth measures reported in the hrs include both financial assets and housing equity. the question of which assets should be regarded as accumulated savings for retirement is the subject of some debate. because of this debate, we will use several measures of accumulated saving. along with many researchers, we assume that individuals are unwilling or unable to consume housing equity and so exclude housing equity from accumulated saving. we also exclude the value of any vehicles, since they are generally not used to provide resources for consumption during retirement. in the results presented below, housing equity is excluded; however, the inclusion of housing equity does not significantly affect the results. the household’s current level of accumulated financial assets (savefin) is measured by the sum of liquid assets, such as bank accounts, cds, treasury bills, stocks, mutual funds and 36 h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 bonds. a second, more inclusive measure of accumulated savings is also examined. this measure, savetot, includes financial assets as well as such nonfinancial assets as business equity and real estate (other than equity in the household’s first or second residence). finally, table 1 variable definitions, means and standard deviations variable name variable definition mean (std. dev.) n 5 4978 male binary variable that takes on a value of one if the respondent is male and zero otherwise. .504 (.500) white binary variable that takes on a value of one if the respondent is white and zero otherwise. .741 (.438) married binary variable that takes on a value of one if the respondent has a partner and zero otherwise. ‘‘partner’’ refers to a spouse or live-in companion of the same or opposite sex. .756 (.429) age number of years of age. 55.5 (3.10) edyrs number of years of education completed. 12.50 (2.93) hlth binary variable that takes on a value of one if the respondent states that their current health is excellent, very good or good, and zero otherwise. .925 (.263) hhinc the total household income from all sources including earned income, investment income, pensions and transfers payments. (the log of hhinc is actually used in the estimates. $56,537 (50349) totkids number children, either living at home or outside household 1.74 (2.14) kidsah number of children living at home .312 (0.680) yrsemp number of years employed with the current employer. 15.41 (10.97) selfemp binary variable that takes on a value of one if the respondent is selfemployed and zero otherwise. .128 (.334) retplan binary variable that takes on a value of one if the respondent participates in a retirement plan offered through the place of employment and zero otherwise. .626 (.484) sevefin the value of the household’s liquid assets plus stocks plus bonds. $41,073 (157,783) savetot the value of the households assets, less housing equity, ira value and the value of household vehicles. $130,797 (448,760) retsave the respondent’s expectation of the household’s liquid assets and reserves at retirement without ira or pension funds. $108.669 (346,774) nwira the value of all of household’s ira assets. $19,041 (55,419) wlthinc the ratio of savefin to hhinc. .722 (3.50) increp the ratio of the future value of savefin to retsave. 1.75 (12.51) standrd coded response to the question, ‘‘do you expect your living standards to increase a lot (4), increase somewhat (3), stay about the same (2), decline somewhat (1) or decline a lot (0)?’’ 1.60 (0.750) 37h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 the part of savings that is specifically set aside in an ira (nwira) is used as another measure of accumulated retirement saving. in planning to replace some percentage of preretirement income, the dollar amount of savings may not be as important as the accumulated savings relative to the income that needs to be replaced. the same dollar amount of savings might be sufficient to replace all of a low-income individual’s income but might replace only a very small percentage of a high-income individual’s income. to capture the relative need for saving, the ratio of nonpension financial assets to household income (finsave/hhinc) is used. because the sample consists of those who are “not at all retired,” additional savings may be accumulated in the period prior to retirement. as part of the retirement planning section of the survey, individuals were asked what they expected to have in “savings and reserves” exclusive of ira and pension fund assets. this variable (retsave) reflects the individual’s expectations concerning future savings. individuals may have unrealistic expectations of their ability to save in the future. to try to capture how reasonable those savings goals are we compute the ratio of the future value of the actual accumulated savings to the expected savings at retirement (increp). the numerator is what they will have at retirement if they simply invest what they have already accumulated. the denominator is what they expect to have at retirement. the larger increp, the larger the percentage of their expected savings and reserves have already been accumulated. 3.2. expectation standard of living the second part of our research focuses on the expectation of the standard of living after retirement. the hrs provides two methods to assess the individual’s expectations. the first is simply to ask, “do you expect your living standard to increase a lot (4), increase somewhat (3), stay about the same (2), decline somewhat (1), or decline a lot (0)?” the second method is to proxy the individual’s expected ability to replace preretirement income with pension benefits. we use a binary variable indicating pension plan participation and the number of years employed by the current employer to proxy expected pension benefits. a binary variable for self-employed is also included to indicate the individual’s work situation. the hrs contains the information necessary to estimate a dollar amount of expected pension benefits; however, the computation is complicated. mitchell and moore (1997) discuss the complexities of estimating pension wealth as well as the problems of estimating “adequate” retirement saving. mcgarry and davenport (1997) deal specifically with some of the assumptions involved in the estimation process. a question of interest in future research is how well individuals understand their pension benefits and whether their expectations correspond with the benefits projected by their employers. 3.3. demographic variables in addition to the variables that directly reflect the ability to replace income, variables reflecting demographic and life situation differences are included in both the accumulated saving and standard of living models. the respondent’s race is coded as one if white and zero otherwise. the education level is measured in years of education. since marital and health 38 h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 status may both affect the need for accumulated savings, “married” is used to designate those who indicate they have partners of the same or opposite sex, and health status is an indicator variable that is based upon a self-reported scale of physical health. the total number of children and the number of children at home may affect the ability to save for retirement. 4. statistical methodology in the first part of our analysis, the level of accumulated savings is modeled using a sample selection technique rather than a simple regression. the level of accumulated savings is a function of the binary variable for pension plan participation and other explanatory variables. the use of this binary variable may introduce a sample selection bias if pension plan participation is systematically related to the other explanatory variables. to take account of this potential problem, a two-stage heckman procedure is used. in the first stage, a probit model of pension plan participation is estimated using the other explanatory variables as right hand side variables. from this estimate, the inverse mills ratio (generally referred to as “lambda”) is calculated and included as an explanatory variable in the second-stage, leastsquares regression. if the lambda variable is significant, the use of the pension plan participation variable in a simple regression model would introduce a sample selection bias. the second component of the analysis investigates the “expected standard of living in retirement” equation, by estimating an ordered probit model. as can be seen in the form that the question about retirement living standard is posed (see above), respondents provide information about their expected standard of living based upon an ordinal ranking of the choices, choosing the selection that most closely corresponds to their true level of expectations. an ordered probit model can provide useful insights about this type of question (see zavoina & mcelvey, 1975). this model is a latent regression procedure that assumes that an underlying measure of expectations exists but that its value cannot be observed. thus, the true model is y* 5 b9x 1 e, (1) but y* is unobserved. instead, we observe values ofy, which correspond to this person’s expected standard of living, with values starting at 0 and increasing by units of one as the expected standard of living increases. these responses correspond to a set of parameters, usually labeledm’s, that partitions the distribution ofy*. the estimation procedure thus determines the probability that the valuey* falls into a range of themi’s as established by the observed values of y (in this case, the responses to the questions of expected standard of living). this model assumes thate is normally distributed and the mean and variance ofe are normalized to zero and one, respectively. 5. empirical results 5.1. actual and expected savings accumulations before examining the findings concerning the factors affecting accumulated savings by the respondents, it may be helpful to consider the determinants of pension plan participation. 39h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 while these estimates are produced primarily to evaluate and control for selection bias, they also provide a profile of the type of individuals who are more (or less) likely to receive pension benefits upon retirement. these results, as are the accumulated and expected savings estimates, are displayed in table 2. in general, the results shown in these estimates, found in the first column of the table, are quite consistent with one’s intuition about a typical individual who is a participant in a pension plan. the estimates show that participants are more likely to be male, to have more education and higher income, and to have worked longer for their employers. as one would expect, individuals who are self employed are significantly less likely to be involved in a pension plan. interestingly, younger individuals are more likely to participate in a pension plan. along the same lines, married respondents are significantly less likely to participate. the total number of children (totkids) is not significant but the number of children in the home (kidsah) has a significant negative impact on pension participation. it may be that table 2 sample selection model of accumulated savings dependent variable first stage probit estimates retplan second stage estimates savefin savetot nwira retsave constant 24.02 2695746 2301408 2265833 1297383 (7.485) (12.696) (14.768) (14.047) (10.724) male .1376 21253.0 233406.7 28610.4 33820.6 (2.281) (0.205) (1.921) (4.084) (2.507) white .0051 1596.5 22759.8 6257.1 24592.4 (0.099) (3.030) (1.516) (3.439) (2.112) married 21953 222525.8 279910.2 20.400 287628.5 (3.423) (3.815) (4.755) (0.000) (6.719) age 20313 2490.5 3023.6 1697.1 23455.5 (4.400) (3.340) (1.426) (6.590) (2.099) edyrs .0829 3222.8 7753.8 1187.5 6619.3 (9.545) (3.156) (3.684) (3.366) (2.945) hlth — 21144.7 212920.4 1331.1 9505.3 (0.139) (0.577) (0.467) (0.537) loghhinc .4446 55160.5 224918 15913.3 154362 (22.766) (11.702) (16.915) (9.770) (14.904) yrsemp .0514 697.9 4778.6 232.15 23221.3 (22.233) (1.776) (4.343) (0.237) (2.700) selfemp 22.58 20607.1 74967.4 17944.1 10932.1 (28.238) (0.529) (1.481) (2.859) (0.276) totkids .0002 2830.7 577.2 2406.5 471.0 (0.014) (0.568) (0.139) (0.805) (0.146) kidsah 2.0985 1516.6 239746.6 23349.0 228208.5 (2.766) (0.396) (3.647) (2.532) (3.335) retplan — 252172.8 2374513 6700.2 2224576 (1.980) (5.110) (0.735) (3.911) lambda — 288842.1 207961 27552.3 129951.8 (1.855) (4.824) (1.404) (3.842) chi-squared statistic/ adjusted r2 2174.0 0.068 0.144 0.094 0.104 note: absolute value of t-statistics in parentheses. 40 h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 children at home represent a drain on resources so that pension fund participation is delayed or foregone entirely. once children have left the home they no longer have an impact on participation. table 2 also contains the estimates of the relationship between accumulated savings and participation in a pension plan for four measures of accumulated savings: financial assets (savefin), all assets except equity in one’s residence and ira accumulations (savetot), estimates of accumulated ira assets (nwira), and anticipated financial assets at retirement (retsave). these estimates are generally consistent across most of the variables, though there are some notable differences between “ordinary” savings and iras. accordingly, these will be discussed separately. accumulated savings portray a consistent and intriguing picture of savings behavior. individuals who have accumulated more total dollars of savings are significantly more likely to be white, older, have more formal education, have higher income, and have more work experience. notably, males have significantly higher levels of savings than females, and married respondents have lower levels of savings than their unmarried counterparts. the more children in the home, the lower the level of accumulated saving as measured by savetot. health status (current) does not play a significant role in the level of savings, and interestingly, the results show no statistically measurable differences in the level of savings for self-employed individuals as compared to those who are employed by others. this is noteworthy, because self-employed individuals are far less likely to have pension assets to draw on in their retirement. their savings may, however, take a different form, as the ira estimates indicate. for the first two measures of savings accumulations (savefin and savetot), there are indications that individuals treat their savings and pension benefits as substitutes. the coefficient on the pension participation variable is negative and significant in both estimates of accumulated savings. one should note that many of the studies using techniques to detect and control for the presence of selection bias provide separate estimates for the two groups—in the case here, this would be pension plan participants and nonparticipants. the approach we take is to determine what is usually called the “treatment” effect and estimate only one equation with a dummy variable. see barnow, cain and goldberger (1981) for a discussion of this approach. the results show convincing evidence of selection bias at work in this process, as shown by the statistical significance of lambda. taken together, retplan and lambda show that the impact of pension plan participation is larger than estimates that do not take into account the selection effects on savings. in general, the estimates show that those who participate in a pension are more likely to save than those who do not, so the impact of pensions on savings accumulation would be understated if this bias were not controlled in the estimates. a simple way to check this is to examine an ordinary least squares estimate of the accumulated savings equations. the ols estimates show that little or no substitution behavior is present. (these estimates are available from the authors upon request.) now consider the fourth column of the table 2, in which estimates of accumulated ira assets are presented. there are several differences in the factors that are important in determining the size of ira accumulations, as compared to ordinary savings. for instance, while there are no gender differences in the accumulation of ordinary savings, males are 41h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 significantly more likely to have higher ira assets, and, in contrast to the findings for ordinary savings, married and unmarried respondents display no significant differences in the amount of ira savings. the years of work experience are not statistically significant in the ira equation, unlike the results for both of the regular savings equations. most notably, those who are self-employed are significantly more likely to have larger accumulations of ira assets, though it is reasonable to expect these individuals to use this type of financial vehicle, since pensions are often unavailable. the total number of children is insignificant; however, the number of children at home has a significant negative impact on nwira. finally, participation in a retirement plan is unrelated to ira accumulations, again different from accumulations of savings outside of retirement plans. additionally, there is scant evidence of selection bias in the case of ira savings. the final column in table 2 shows estimates based upon the (self-reported) expectations of respondents about the level of savings they will have in retirement. there is remarkable consistency here between the results for this equation and the estimates for both financial assets and overall savings amounts. note that the question about expected savings explicitly asked the respondent to answer based upon overall savings but not to include either iras or pension benefits. thus, the comparison between columns two, three and five is natural. the only differences in the estimates are found in the gender variable: males are significantly more optimistic about their level of retirement savings, in that males have a significantly higher level of anticipated savings, while there are no significant gender differences in actual accumulations. there are also some slight differences on the basis of age. older individuals expect to have lower retirement savings than younger respondents do. these individuals are likely to be closer to retirement than are younger individuals; therefore, this may not constitute a real difference in the estimates but, rather, better information about actual accumulations. indeed, if the similarities across the two types of estimates are any indication, there is consistency between actual accumulations and the expected levels of savings upon retirement. of course, these results do not address the issue of savingsadequacy.additional work on this needs to be done to determine the underlying relationship between the size of accumulated savings and expected retirement levels. 5.2. expectations about changes in the living standard the results for the estimates of expected living standard can be found in table 3. all of the results examine the relationship between the anticipated living standard and objective measures of the individual’s ability to replace income. the differences displayed across the columns depend upon the savings or income measures included in the specification. the first column includes income and accumulated financial assets plus the value of ira assets. the subsequent columns move from absolute measures of income or wealth to relative measures that attempt to measure the extent to which income replacement is to be accomplished. the dependent variable is defined as arelative change in the living standard, that is, whether they expect their standard of living to improve, stay the same or worsen. if one were measuring the absolute standard of living, the results presented in the table may appear counterintuitive, but not in relative terms. an example of this is found in the results for the first column. the income variable is negative and not statistically significant. the anticipated 42 h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 change in living standard after retirement is unrelated to income level, even though people with higher incomes are likely to have higher living standards than do people with lower incomes. their expectations about their future standard of living—relative to their current standard of living—are unrelated to an absolute measure of income. many of the other variables work in a manner comparable to these findings. respondents who are white may expect (because their socioeconomic status may be higher) that this level may not be maintained in retirement, while minority individuals—at a lower level—may view the potential change in their living standard more positively. similarly, the results for the respondent’s education level are negative and (marginally) statistically significant. education may increase income in absolute levels, but may not positively affect the anticipated retirement living standards. table 3 ordered probit model of expected standard of living retirement variable (1) (2) (3) (4) constant 2.420 2.048 2.204 2.202 (6.410) (6.698) (7.158) (7.149) male .2372 .2306 .2383 .2381 (5.593) (5.431) (5.620) (5.615) white 2.1833 2.1724 2.1835 2.1839 (5.121) (4.863) (5.151) (5.163) married .0289 .0331 .0193 .0188 (0.737) (0.925) (0.539) (0.525) age 2.0131 2.0110 2.0130 2.0129 (2.5580 (2.157) (2.534) (2.523) edyrs 2.0098 2.0070 2.0106 2.0107 (1.696) (1.278) (1.935) (1.950) hlth .1356 .1374 .1330 .1327 (2.655) (2.697) (2.612) (2.605) hhinc (log) 2.0219 — — —(0.872) totkids .0117 .0113 .0123 .0123 (1.163) (1.124) (1.223) (1.228) kidsah .0133 .0127 .0129 .0129 (0.522) (0.512) (0.520) (0.520) selfemp .0924 .1085 .0920 .0927 (1.804) (2.138) (1.804) (1.818) retplan .0420 .0408 .0396 .0394 (1.140) (1.144) (1.109) (1.103) savefin .33 e206 — — —(2.991) nwira .13 e205 — .13 e205 .13 e205 (2.586) (2.885) (2.862) wlthinc — .0105 .0092 .0089 (2.444) (1.901) (1.768) increp — — — .0009 (0.669) chi-squared statistic 112.5 87.4 107.1 107.7 note: asymptotic t-statistic in parentheses. 43h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 it does appear that (echoing the results from the expected retirement savings level) male respondents are more optimistic about their living standard than are females. in contrast to the results for the savings estimates in table 2, marital status is unrelated to expected changes in living standard. individuals who are self-employed have expectations of an improved living standard. both the total number of children and the number of children at home are insignificant. finally, those individuals who consider themselves to be currently in good health view their future living standard positively. it is possible that individuals whose current health is good (which is correlated with future health) do not anticipate significant outlays for health care costs that could impair their ability to consume otherwise. turning to the variables related to financial status as it relates to retirement, these variables produce an interesting set of results. the accumulation of assets is, in general, positively related to the expected standard of living. those who have larger accumulations of ira assets or other financial assets (savefin) anticipate an improved standard of living. the relative measure of wealth to income (wlthinc) is also positively associated with the expected living standard. the ratio of actual accumulated savings to the expected accumulation of savings at retirement, increp, is not statistically significant. the more the household has saved, the more likely it is to believe that it will maintain its living standard. on the other hand, what they have saved relative to what they expect to save may not convey as much information as to the direction of their standard of living. the pension plan participation variable is insignificant in all of the equations. participation in a pension plan is not statistically related to anticipated changes in standard of living. the lack of significance of this variable is intriguing. apparently, people believe that their accumulated savings will affect their retirement standard of living but not pension fund participation. this discrepancy may be related to the tendency seen in the first section of our analysis for individuals to substitute saving for retirement participation. one interpretation of this result is that individuals who do not participate in pensions may have made sufficient other arrangements so that their expectations about their future living standard do not differ from those who participate in pension plans. other potential explanations of the insignificance of the pension plan participation variable could be that they are not certain what benefits they will actually receive. this could be a lack of understanding of their pension plan or a lack of confidence that they will receive the promised benefits. savings may be given greater weight because saving is controlled directly by the respondent. alternatively, it may be that the measure of pension benefits is inadequate. future research should involve a dollar estimate of the expected pension benefits rather than a dummy variable for participation. 6. summary and conclusion to understand the discrepancy between the expectations of economic well being in retirement and the savings behavior of americans, we examined at two questions. first, do individuals view expected pension benefits as a substitute for accumulated savings in the process of replacing preretirement income? our results are consistent with the idea that pension benefits are seen as substitutes for saving. second, are individuals’ expectations concerning the standard of living they will enjoy in retirement realistic based on their 44 h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 accumulated savings and pension plan participation? our results indicate that accumulated savings have a significant impact on the expected standard of living but pension plan participation does not. the insignificance of the pension plan participation variable is intriguing. as is often the situation, a variety of explanations for the lack of significance are possible, but none is satisfying. future research that focuses on how well individuals understand their retirement plans and how realistically they anticipate the benefits they will receive may explain why pension plan participation does not affect the expected standard of living. references banks, j., blundell, r., & tanner, s. (1998). is there a retirement-savings puzzle?american economic review, 88, 769–788. barnow, b., cain, g., & goldberger, a. (1981). issues in the analysis of selectivity bias. in: e. stromsdorfer & g. farkas (eds.),evaluation studies review annual(vol. 5 pp.43–59). beverly hills, california: sage. bernheim, b. d., skinner, j., & weinberg, s. (1997). what accounts for the variation in retirement wealth among u.s. households?nber working paper no, 6227. cantor, r., & yuengert, a. (1994). the baby boom generation and aggregate savings.federal reserve bank of new york, quarterly review, 19,76–91. hurd, m. (1990). research on the elderly, economic status, retirement and consumption and saving.journal of economic literature, 28,565–637. juster, f. t., & suzman, e. (1995). an overview of the health and retirement study.the journal of human resources, 30(suppl.), s7–s58. lazear, e. p. (1994). some thoughts on savings. in d. a. wise (ed.),studies in the economics of aging(pp. 143–167). chicago: national bureau of economic research. mcgarry k. & davenport, a. (1997). pensions and the distribution of wealth.nber working paper no, 6171. mitchell, o., & moore, j. (1997). retirement wealth accumulation and decumulation: new developments and outstanding opportunities.nber working paper no, 6178. poterba, j. a. (1996). personal saving behavior and retirement income modeling: a research assessment. in nation research council,assessing knowledge of retirement behavior(pp. 123–148). washington, dc: national academy press. zavonia, r., & mcelvey, w. (1975). a statistical model for the analysis of ordinal level dependent variables. journal of mathematical sociology, 4,103–120. 45h.w. elder, p.m. rudolph / financial services review 9 (2000) 33–45 pii: s1057-0810(96)90009-8 financial services review the journal of individual financial management index volume 5,1996 authors achacoso, joseph a. see greninger, sue a. brooks, robert, “computing yields on enhanced cds,” 5( 1): 3 l-42 brown, stewart l., “churning: excessive trad ing in retail securities accounts,” 5( 1): 43 56 ciccotello, conrad s., “equity fund size and growth: implications for performance and selection,” 5(l): 1-12 clinebell, john, review of “personal finance,” 5( 1): 84-85 eyssell, thomas, review of “new york stock exchange (nyse), the chicago board options exchange (cboe), and the finan cial management association websites,” 5(2): 150-151 golec, joseph h., “the effects of mutual fund managers’ characteristics on their portfo lio performance, risk and fees,” 5(2): 133 148 grant, c. terry. see ciccotello, conrad s. greninger, sue a., “ratios and benchmarks for measuring the financial well-being of families and individuals,” 5( 1): 57-70 hampton, vickie l. see greninger, sue a. hatfield, gay b. see walker, m. mark kahl, douglas r., review of “employee bene fits,” 5( 1): 83-84 kahl, douglas r., review of “liffe on the internet,” 5(l): 85-86 kitt, karrol a. see greninger, sue a. main, robert s. see templeton, william k. orris, j.b. see templeton, william k. pahl, dede, “an emerging partnership: afs and the cfp board,” 5(l): 71-81 schooley, diane k., “risk aversion measures: comparing attitudes and asset allocation,” 5(2): 87-99 smaby, timothy r., review of “associations for investment management and research (aimr) web site,” 5(2): 149-150 templeton, william k., “a simulation approach to the choice between fixed and adjustable rate mortgages,” 5(2): 101-l 17 vihtelic, jill, “personal finance: an alternative approach to teaching undergraduate finance,” 5(2): 119-131 walker, m. mark, “professional stock ana lysts’ recommendations: implications for individual investors,” 5(l): 13-29 worden, debra drecnik. see schooley, diane k. 155 156 financial services review 5(2) 1996 titles “a simulation approach to the choice be tween fixed and adjustable rate mort gages,” william k. templeton, robert s. main, and j. b. orris, 5(2): 101-117 “an emerging partnership: afs and the cfp board,” dede pahl, 5(l): 71-81 “churning: excessive trading in retail securi ties accounts,” stewart l. brown, 5( 1): 43 56 “computing yields on enhanced cds,” robert brooks, 5(l): 31-42 “the effects of mutual fund managers’ char acteristics on their portfolio performance, risk and fees,” joseph h. golec, 5(2): 133 148 “equity fund size and growth: implications for performance and selection,” conrad s. ciccotello and c. terry grant, 5(l): 1-12 “personal finance: an alternative approach to teaching undergraduate finance,” jill lynn vihtelic, 5(2): 119-131 “professional stock analysts’ recommenda tions: implications for the individual inves tors,” m. mark walker and gay b. hatfield, 5( 1): 13-29 “ratios and benchmarks for measuring the financial well-being of families and indi viduals,” sue a. greninger, vickie l. hampton, karrol a. kitt, and joseph a. achacoso, 5( 1): 57-70 “risk aversion measures: comparing attitudes and asset allocation,” diane k. schooley and debra drecnik worden, 5(2): 87-99 review of “associations for investment man agement and research (aimr) web site,” timothy r. smaby, 5(2): 149-150 review of “employee benefits,” douglas r. kahl, 5( 1): 83-84 review of “liffe on the internet,” douglas r. kahl, 5( 1): 85-86 review of “new york stock exchange (nyse), the chicago board options exchange (cboe), and the financial man agement association websites,” thomas eyssell, 5(2): 150-151 review of “personal finance,” john clinebell, 5( 1): 84-85 pii: s1057-0810(00)00055-x the asset allocation decision in retirement: lessons from dollar-cost averaging premal p. voraa,*, john d. mcginnisb apenn state great valley, 30 e. swedesford rd., malvern, pa 19355, usa bpenn state altoona, 3000 ivyside park, altoona, pa 16601, usa abstract how should a retiree allocate his wealth between stocks and bonds? we address this question by studying whether it would have been better to have consumed periodically from stocks than from bonds over the seven decades of u.s. financial markets beginning in 1926 and ending in 1995. we find that retirees would have consistently done better by investing in stocks as opposed to bonds. when we analyze dispersion in consumption around its mean we find that there are greater chances for low consumption from the bond portfolio and greater chances for high consumption from the stock portfolio. thus, we challenge the conventional wisdom that one should move away from stocks and towards bonds as one ages. © 2000 elsevier science inc. all rights reserved. jel classification:g1; g11; n2 keywords:retirement planning; asset allocation; dollar-cost averaging 1. introduction how should a retiree allocate his wealth between stocks and bonds? this asset allocation decision is of interest to all investors (brinson, hood & beebower, 1986), but it is of particular significance to retirees, whose human capital is close to zero making it most difficult to handle adverse investment results (posner, 1995). we address this question by * corresponding author. tel.:11-610-648-3374; fax:11-610-725-5224. e-mail address:fpv@psu.edu (p.p. vora). financial services review 9 (2000) 47–63 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00055-x studying whether it would have been better to have consumed periodically from stocks than from bonds over the seven decades of u.s. financial markets beginning in 1926 and ending in 1995. our study is different from other studies of stock and bond performance. the existing literature assumes either that money is periodically invested in financial assets (e.g., butler & domian, 1993) or that a set amount is invested in financial assets at the beginning of a time span with no additional infusions or withdrawals (ibbotson & sinquefield, 1989; siegel, 1994). we look at performance from the perspective of an individual who is constantly “disinvesting” from financial assets. our hypothesis is that it is better to consume from stocks than from bonds. we are led to this hypothesis by the literature on the fallacy of dollar-cost averaging (e.g., constantinides, 1979; knight & mandell, 1993; rozeff, 1994). specifically, we argue that if dollar-cost averaging into stocks is suboptimal, then dollar-cost disinvesting from stocks ought to be sound; that is, because it is imprudent to gradually transfer one’s wealth from cash to securities, it makes sense to gradually transfer one’s wealth from securities to cash (for consumption). in order to test this hypothesis, we begin with a representative retiree who has a life expectancy ofn years after retirement. without loss of generality, we assume that this retiree has wealth of $1 at the beginning of his retired life which is invested in a stock portfolio. using monthly return data on the center for research in security prices’ (crsp) value weighted (vw) portfolio, we calculate the amount that this retiree could have spent every month from the dividends and capital gains for then years, that is, if the retiree followed a dollar-cost disinvesting strategy, we calculate the equal monthly amount that the retiree could have consumed forn years. in order to be able to make comparisons between asset classes of different risks, we then repeat the procedure for a portfolio consisting solely of long-term treasury bonds (safe class), an annuity based on the yield to maturity on a portfolio of long-term aaaand baa-rated corporate bonds and treasury bonds (moderate risk class), and an annuity based on the yield to maturity on a portfolio consisting of only baa-rated corporate bonds (moderately high risk class) and compare the magnitude of these consumption amounts to the stock portfolio based (risky class) amount. we find that retirees would have consistently done better by investing in the stock portfolio as opposed to investing solely in t-bonds or in the broadly diversified bond portfolio or even the moderately risky bond portfolio. this result is particularly true asn gets larger—over longer periods of time stocks look better relative to t-bonds, the broadly diversified bond portfolio, and the moderately risky bond portfolio. we also examine the impact of inflation on consumption. once we allow consumption to grow at the rate of inflation we find even stronger support for our result that disinvesting from stocks is better than from bonds. we also analyze the likelihood, across different asset classes, of states in which consumption is extremely low or high due to extreme returns. we find that there are greater chances for low consumption from t-bonds and greater chances for high consumption from the stock portfolio, that is, there is more “downside risk” from t-bonds than from stocks and there is greater “upside potential” from stocks than from t-bonds. we also find that as the horizon increases, retirees are guaranteed a low consumption from t-bonds and, unless circumstances are exceptional, they are guaranteed better consumption from stocks. 48 p.p. vora / financial services review 9 (2000) 47–63 in order to pinpoint these exceptional circumstances, we construct and analyze the best portfolio of stocks and t-bonds by calculating the optimal weights that should have been attached to stocks and bonds for every period in our study. we find that the frequency with which the optimal portfolio should have been 100% invested in stock is much greater than the frequency with which the optimal portfolio should have been 100% invested in bonds or in a mix of stocks and bonds. we also find a strong correlation between the occurrence of contractions in the economy and the optimality of the t-bond portfolio or the portfolio that has a mix of t-bonds and stocks. this evidence goes against the conventional wisdom that more wealth should be allocated to bonds as one ages. one such rule of thumb is that the percentage of wealth invested in stocks should be equal to 100 minus the investor’s age. for a 65-year old, this would mean no more than 35% of wealth should be invested in stocks. our results indicate that such an allocation would be very costly to older investors in terms of their standard of living—yet, we are aware that many invest very little in stocks relative to bonds. we speculate that this may be related to the fact that the current cohort of elderly investors grew up during the great depression and world war ii and may be extremely sensitive, perhaps overly so, to the risk associated with loss of wealth from stocks. in the next section we summarize some of the studies that look at the asset allocation decision and present our hypotheses; in section 3 we describe our data and methods; in section 4 we present our results, and we conclude in section 5. 2. literature review and hypotheses in this section, we review some prior work on optimal asset allocation between stocks and bonds for retirees and the findings on dollar-cost averaging (dca). it appears that little work has been done to address the former issue while substantial work has been done on the latter topic. we were able to find only two studies that looked squarely at the question of optimal allocation for retirees. bengen (1994) finds, for a retiree who has a horizon of at least 30 years, that a 50–50 to a 75–25 stock-bond combination is optimal in retirement if the primary goal of the retiree is to not outlive his assets. in related literature, jagannathan and kocherlakota (1996) pose the question “why should older people invest less in stocks than younger people?” they consider a number of alternative explanations and they conclude that the only plausible explanation is that the value of their human capital (which is considered riskless) decreases as they age so that they need to shift more financial wealth into risk less assets in order to reach their optimal portfolio. although siegel (1994) does not look at optimal asset allocation in retirement, he presents an impressive array of statistics, particularly in the first two chapters, to make the case that stocks are far superior to bonds in terms of risk and return, especially over long holding periods. on the topic of dca, constantinides, (1979) shows that it is a sub optimal investment policy because it depends upon the composition (and not just size) of the investor’s wealth while the optimal nonsequential policy is independent of the composition of the investor’s wealth. knight and mandell (1993) demonstrate numerically that a risk averse investor would get lower utility under dca versus a lump sum or an optimal rebalancing strategy. 49p.p. vora / financial services review 9 (2000) 47–63 additionally, they provide empirical evidence that the utility of risk averse investors would have been lower in the 1962–92 period under dca as opposed to a lump sum or optimal rebalancing strategy. in a similar vein, rozeff (1994) analytically and empirically demonstrates the superiority of a lump sum strategy over dca. thus, there appears to be a consensus in the literature that dca is inferior to a lump sum strategy of investing into stocks. in the face of all this research why does dca persist? statman (1995) offers four behavioral explanations for the resilience of dca, although he provides no evidence on which is the most plausible explanation. our study is somewhat related to bengen’s because we want to study whether stocks are better in retirement than bonds. our study is related to the dca literature in the following respect: a retiree’s objective is fundamentally different from an individual who is building assets during his earning life. a retiree is concerned with the amount of consumption that his assets will allow him. in order to consume he will disinvest periodically from his financial assets. we call this activity dollar-cost disinvesting and it is the reverse of dca. our main hypothesis is that dollar-cost disinvesting from stocks will lead to a higher magnitude of consumption than from bonds. this hypothesis follows from the dca literature that consistently shows that dca into stocks is sub optimal. our second hypothesis is that dollar-cost disinvesting from stocks gets continuously better relative to bonds as the retirement horizon gets longer. although we are unaware of studies that examine the suboptimality of dca for longer versus shorter periods of time, our second hypothesis follows from the various studies such as siegel’s that show that stocks look better in comparison to bonds as the horizon gets longer. although our study is related to bengen’s, there are significant differences between the two. for instance, it is unclear whether optimal asset allocation in retirement should be considered only from the perspective of individuals who expect to spend at least 30 years in retirement. with no mandatory retirement age in the u.s. in most professions, it is entirely possible that people who choose to keep working longer may be looking at a shorter retirement horizon. thus, we consider various retirement horizons ranging from 5 to 30 years. additionally, bengen focuses on the longevity of the portfolio, and particularly upon the impact of significant downturns in the market on longevity; we believe that the main objective of retirees is to enjoy the utility they derive from consumption so we focus on the magnitude of consumption that is possible under different asset allocations. 3. data and methods we present below a simple model that is used to arrive at the equal monthly consumption amount for different portfolios together with a description of our data sources. dca assumes that a fixed dollar amount is moved from one type of financial asset (like cash) into another type (like stock) periodically. conceptually, dollar-cost disinvesting is similar to dca except that we reverse the process by moving a fixed dollar amount from a stock or bond portfolio into cash which is spent immediately on consumption of goods or services. as in the dca papers, we use a month as a convenient unit of analysis for moving dollars between assets. 50 p.p. vora / financial services review 9 (2000) 47–63 we also believe that a month is a convenient period for analyzing the dollar-cost disinvesting decision because most households use a month as a convenient period for budgeting. we calculate the monthly equal consumption that could be possible for an individual who hasw0 of wealth invested in a portfolio of investments at the beginning of his retirement and starts to spend the money right away. we assume that all the money is spent by the end of the retirement period. for a two-month long retirement period we would solve forc the consumption amount in the expression ((w0 –c)(11r1) –c)(11r2) –c 5 0, wherert is the rate of return on a portfolio in month t. when we extend this formulation ton years we solve for c in ~. . . ~~~~w0 2 c!~1 1 r1! 2 c!~1 1 r2! 2 c!~1 1 r3! . . .2 c!~1 1 r12n! 2 c! 5 0 (1) eq. (1) is linear in one unknown and it has a unique solution for a givenw0. this approach abstracts away the uncertainty in the rate of return because the consumption amount is based on ex-postreturns. our goal is to make some statements regarding optimal asset allocation for retirees based upon events that have already occurred, and thus our approach is no different from the approach adopted in the studies discussed in the previous section. eq. (1) is used to estimate one consumption amount if a retiree’s wealth is 100% invested in the crsp vw portfolio and another consumption amount if a retiree’s wealth is 100% invested in t-bonds. we obtain the monthly returns on the crsp vw portfolio and the monthly holding period returns on an equally weighted t-bond portfolio from the crsp tapes for january 1926 through december 1995. clearly, this period was chosen for analysis due to the easy availability of data. however, because this period of 70 years (or 840 months) has seen extraordinary fluctuations in stock prices and movements in interest rates, results based on this period will be more robust than in periods of relative calm. in addition to being able to invest their wealth in stocks and t-bonds, retirees have had the opportunity, for a number of decades, to invest in annuities (poterba, 1997). if a retiree invests in acertain annuity,the invested wealth gets amortized at a fixed interest rate over a fixed number of years, while in alife annuitythe invested wealth gets amortized at a fixed interest rate over the remaining life of the retiree. the interest rate that is used to amortize the investment is based on interest rates prevailing in the market at the time the annuity is bought, and even at a given time it can vary with the risk of the asset on which the annuity is based. we want to study whether such annuities outperform stock and t-bond based investments in retirement. thus, we compare the performance of retirement assets based on monthly returns on the crsp vw portfolio to a fixed annuity based on the holding period returns on a t-bond portfolio, an annuity based on the yield to maturity on a broadly diversified bond portfolio of medium risk, and an annuity based on the yield to maturity on a baa-rated corporate bond portfolio of moderately high risk. the broadly diversified bond portfolio is created by assigning equal weights to a portfolio of aaa-rated corporate bonds, baa-rated corporate bonds, and t-bonds. we amortizew0 on a monthly basis overn years at the monthly yield to maturity prevailing on the first day of retirement. one drawback of the above model is it does not account for inflation because the annuity remains fixed over the retirement period. we also separately allow for growth in the consumption amount at the rate of inflation in order to keep the retiree’s standard of living 51p.p. vora / financial services review 9 (2000) 47–63 constant. thus, for the stock portfolio and the t-bond portfolio, we solve forc in the following expression which captures the essence of eq. (1) and at the same time allows for inflation adjustments: ~. . . ~~~~w0 2 c!~1 1 r1! 2 c~1 1 i1!!~1 1 r2! 2 c~1 1 i1!~1 1 i2!!~1 1 r3!. . . 2 c p t51 12n21 ~1 1 i t!)~1 1 r12n! 2 cp t51 12n ~1 1 i t! 5 0 (2) where it is the rate of inflation at time t. we obtain the monthly cpi for all urban wage earners from the bureau of labor statistics (1997) and we calculate the monthly rate of inflation based on this cpi. for the annuities we account for inflation by settingrt5r1 for all t52,3,. . . . . ,12n in eq. (2) because the annuities based on the yield to maturity on the broadly diversified bond portfolio and the annuity based on baa-rated corporate bond portfolio depend solely on the interest rate at the commencement of the retirement period. with the growing life expectancies in the u.s. during 1926 through 1995 and the changing dynamics of employment it is difficult to make statements regarding what the appropriate retirement horizonn should be. we address this issue by looking at six values forn ranging from 5 years, which is considered a short retirement, to 30 years, which is considered a long retirement, with 5-year increments. with 840 months of data at our disposal, there are 840–12n different periods of length 12n that could be analyzed. even whenn is set to its maximum value of 30 there are 480 possible unique periods of analysis. it is redundant to study periods beginning in montht and beginning in the vicinity of montht because the results are likely to be very similar. thus, we impose the following constraints on the number of periods that are studied: (i) for eachn at least 30 periods are studied, and (ii) for the sake of consistency the number of months between the beginning of one draw and the beginning of the next remains constant. for an example that will clarify the above procedure, consider the retirement horizon of 5 years, whenn 5 5. if the first retirement period begins in january 1926, it will end in december 1930. if we allow the second retirement period to begin in march 1928 and end in february 1933 then the gap between the beginning of the first and the beginning of the second is 26 months. with this gap, we have a total of 31 retirement periods with the last beginning in january 1991 and ending in december 1995. if the gap were to be increased to 27 months or more then it is not possible to have at least 30 retirement periods within the 840 months of data that we have. if the gap is decreased, then we have too many retirement periods. we find that for every 5-year increment inn if the gap is decreased by two months then we are able to meet both of the above constraints. an additional benefit is that we uniformly end up having 31 retirement periods of analysis for alln with the first period always beginning in january 1926 and the last period always ending with december 1995. summary statistics for the returns in the data are presented in table 1 . the average monthly return on the crsp vw portfolio is 0.9606% while the standard deviation of its rate of return is 5.5%. compared to the rates of return on the other portfolios, it appears that the crsp vw has higher risk but also offers a higher rate of return. in the next section we turn our attention to the results. 52 p.p. vora / financial services review 9 (2000) 47–63 4. results 4.1. main results in table 2 we report the results on dollar-cost disinvesting for various values of the retirement horizon,n, and for the portfolios that we have selected. for eachn and for each portfolio we present the average monthly consumption amount, which we termedc in eq. (1), for the 31 retirement periods that are included in the analysis. thus, an individual with a retirement horizon of 5 years who invested $100 into the crsp vw portfolio could have spent $2.19 every month immediately upon retiring till the end of 5 years. instead, if he had opted to invest in a portfolio of t-bonds, he could have spent $1.89. we provide thep-value of the wilcoxon sign rank statistic for the null hypothesis that the mean difference between the stock based consumption amount and the t-bond based consumption amount is zero in parentheses. our analysis of whether the consumption numbers as well as the differences in consumption between different classes of assets are normally distributed or not suggests that normality cannot always be assumed. thus, we resort to nonparametric test statistics which impose few distributional assumptions on the data for all our hypothesis tests (conover, 1980). we also provide the number of periods out of 31 in which the stock based consumption was greater than t-bond based consumption below thep-value of the wilcoxon sign rank statistic. finally, the number that appears in brackets provides an indication of how much bigger (in percentage terms) the consumption based on the stock portfolio is compared to consumption based on the t-bond portfolio. thus, the consumption based on the stock portfolio over 5 year horizons is significantly greater than the consumption based on t-bonds at the 1% level. additionally, in 23 out of 31 five-year periods the stock based consumption was greater than t-bond based consumption. based on the sign test with a probability of success equal to 50%, in a random sample of 31 draws, we would find 23 or more successes to occur less than 1% of the time. finally, the stock portfolio allows 15.87% greater consumption per 100 dollars invested than the t-bond portfolio, that is, the difference of 30 cents between $2.19 and $1.89 is 15.87% of $1.89. when stock based consumption is compared to t-bond based consumption over different horizons, a clear pattern emerges. stock based consumption is always greater than t-bond based consumption regardless of the retirement horizonn. this difference is always statistically significant at the 1% level. the percentage economic difference increases monotonically with the length of the retirement horizon. to see this economic difference more clearly consider an individual who has accumulated $100,000 by retirement. if this money is table 1 summary statistics on monthly returns for portfolios portfolio mean standard deviation s&p 500 index 0.9606% 5.50% t-bond holding period 0.435% 1.57% aaa, baa, t-bond annuity 0.502% 3.05% baa bond annuity 0.588% 3.27% 53p.p. vora / financial services review 9 (2000) 47–63 invested in the stock portfolio, with a five year horizon it translates into an additional $300 per month of consumption as opposed to the t-bond portfolio. for a 30 year horizon, the stock portfolio generates $420 more each month than the t-bond portfolio. thus, our results unequivocally suggest that a stock based portfolio is better than a t-bond based portfolio. when stocks are compared to an annuity based on the broadly diversified investment grade bond portfolio (annuity a) we find similar results. consumption based on the stock portfolio is statistically significantly greater for all horizons from 5 to 30 years. the statistical and economic differences also grow monotonically as the retirement horizon lengthens. we note that consumption based on annuity a is greater than consumption based on the t-bond portfolio for all horizons, but only marginally so. we also compare consumption based on the stock portfolio to consumption based on the yield on baa-rated corporate bonds (annuity table 2 monthly consumption based on u.s. financial market historya horizon (years) stock t-bonds annuity a annuity b 5 2.19% 1.89% 1.94% 1.98% (,0.01) (,0.01) (0.02) 23*** 23*** 22*** [15.87%] [12.89%] [10.61%] 10 1.31% 1.06% 1.10% 1.16% (,0.001) (,0.01) (0.02) 24*** 22*** 21** [23.58%] [19.09%] [12.93%] 15 1.05% .75% .79% .86% (,0.001) (,0.001) (0.01) 21** 20** 20** [40.00%] [32.91%] [22.09%] 20 .91% .59% .63% .69% (,0.001) (,0.001) (,0.01) 24*** 24*** 21*** [54.24%] [44.44%] [31.88%] 25 .86% .49% .54% .59% (,0.001) (,0.001) (,0.001) 25*** 25*** 24*** [75.51%] [59.26%] [45.76%] 30 .85% .43% .47% .53% (,0.001) (,0.001) (,0.001) 29*** 26*** 26*** [97.67%] [80.85%] [60.38%] a monthly consumption is the consumption at the beginning of every month for every dollar of wealth, with nothing left over at the end of retirement. the percentages are an average over 31 retirement periods of equal length depending on the retirement horizon, with the first period beginning in january 1926 and the last period ending in december 1995. thep-values of the wilcoxon sign rank statistic for the null hypothesis that the t-bond and annuity portfolios provide consumption equal to the stock portfolio are presented in parentheses. number of periods out of 31 in which consumption from the stock portfolio is greater than that from the other portfolios is presented in italics. in brackets is how much more consumption the stock portfolio yields as a percentage of the other portfolios. * significant at the 10% level. ** significant at the 5% level. *** significant at the 1% level. 54 p.p. vora / financial services review 9 (2000) 47–63 b). although consumption based on this annuity is superior to consumption based on annuity a, it is statistically and economically dominated by the stock portfolio and the dominance is stronger for longer horizons. thus, we find that stocks are better than t-bonds and are also better than annuities based on treasury and corporate bonds. one very important concern for retirees is the effect of inflation on consumption. in table 3 we present our results on the real consumption that the retiree can afford during retirement based on the solution to eq. (2). compared to the percentages appearing in table 2, the consumption allowances appearing in table 3 are smaller. however, we find that the dominance of the stock portfolio over the others is generally larger in economic terms while tests for differences in mean consumption based on stocks versus consumption based on table 3 monthly real consumption based on u.s. financial market historya horizon (years) stock t-bonds annuity a annuity b 5 1.99% 1.73% 1.76% 1.81% (,0.01) (,0.001) (0.02) 23*** 23*** 22*** [15.03%] [13.07%] [9.94%] 10 1.12% .90% .94% .99% (,0.001) (,0.001) (0.02) 24*** 22** 21** [24.44%] [19.15%] [13.13%] 15 .84% .59% .63% .67% (,0.001) (,0.001) (,0.01) 25*** 22*** 20** [42.37%] [33.33%] [25.37%] 20 .71% .43% .47% .51% (,0.001) (,0.001) (,0.001) 26*** 26*** 23*** [65.12%] [51.06%] [39.22%] 25 .65% .35% .38% .42% (,0.001) (,0.001) (,0.001) 28*** 27*** 25*** [85.71%] [71.05%] [54.76%] 30 .64% .29% .31% .36% (,0.001) (,0.001) (,0.001) 29*** 27*** 26*** [120.69%] [106.45%] [77.78%] a monthly consumption is the consumption at the beginning of every month for every dollar of wealth, with nothing left over at the end of retirement. the percentages are an average over 31 retirement periods of equal length depending on the retirement horizon, with the first period beginning in january 1926 and the last period ending in december 1995. thep-values of the wilcoxon sign rank statistic for the null hypothesis that the t-bond and annuity portfolios provide consumption equal to the stock portfolio are presented in parentheses. number of periods out of 31 in which consumption from the stock portfolio is greater than that from the other portfolios is presented in italics. in brackets is how much more consumption the stock portfolio yields as a percentage of the other portfolios. * significant at the 10% level. ** significant at the 5% level. *** significant at the 1% level. 55p.p. vora / financial services review 9 (2000) 47–63 t-bonds, annuity a, and annuity b maintain their level of statistical significance. thus, our main results are robust to inflation. 4.2. transactions costs and taxes the results presented in tables 2 and 3 and our discussion in the preceding section, present a strong case for investing in a broadly diversified stock portfolio in lieu of t-bonds or annuities based on bonds, particularly for retirees with long horizons. in this section we provide a brief discussion of the impact of taxes and transactions costs on our conclusions. formal comparisons of transactions costs between stocks, bonds, and annuities have yet to appear in the literature. however, it has been recognized that transactions costs are driven by dealer inventory considerations, by liquidity of the traded asset, and by adverse selection in the market for the traded asset (huang & stoll, 1997). mitchell, poterba, and warshawsky (1997, p. 2), find that a 65-year old purchaser of a life annuity can expect to pay 15% to 20% of their amortized wealth as a transaction cost. they also note that adverse selection is a significant problem in the market for annuities aseveryannuity buyer knows more about their own health than any annuity seller. in the bond and the stock markets, however, adverse selection is a problem only when dealers are trading with buyers and sellers who are better informed than they are. because theaverageinvestor in the bond and stock markets is unlikely to possess inside information, and therefore, unlikely to be better informed than dealers, we believe that adverse selection is a smaller problem in the bond and stock markets relative to the annuity market. additionally, the bond and stock markets are significantly more liquid than the market for annuities. therefore, if transactions costs were to be factored into our analysis, our result that retirees should consider investing in stocks will be even stronger. the analysis of how taxes will affect our results will depend on whether disinvesting is being done from a “qualified” pension account like a 401(k), a 403(b), or a deductible ira or from outside such an account. because contributions to a qualified account are taxdeductible, the withdrawals from such an account will be wholly taxable regardless of whether the funds are invested in stocks, bonds, or annuities. in this instance, taxes have no impact on our conclusions. however, if disinvesting is being implemented from outside such a qualified account, then stocks will generally impose less of a tax burden than bonds or annuities. in a broadly diversified stock portfolio like the crsp vw portfolio or the s&p 500 index, many companies pay no dividends. therefore, a greater portion of the monthly cash flow received from disinvesting such a stock portfolio will come from capital gains relative to the cashflow from disinvesting a bond portfolio or an annuity based on corporate bonds. because the statutory tax rate on capital gains is less than on ordinary income (for most people), disinvesting from a stock portfolio will impose a slightly lower tax burden as compared to disinvesting from bonds or from an annuity. thus, from a tax perspective our results remain unchanged if disinvesting is being carried out from a qualified account, and our case for a retiree investing in a stock portfolio becomes stronger if disinvesting is being implemented from outside a qualified account. 56 p.p. vora / financial services review 9 (2000) 47–63 4.3. consumption risk a common objection that is raised to the type of analysis that is presented in the preceding section, which focuses on average outcomes, is that it fails to account for the higher dispersion in consumption around the mean that accompanies investing in stock portfolios. for a retiree, this dispersion manifests itself in fluctuations in purchasing power, or in “consumption risk”. therefore, although on average the retiree might have high consumption from stocks, there will be some states where the stock-based consumption is extremely low. such low consumption will be reflected in a low level of utility during that state. the standard expected utility maximization models suggest that on anex-antebasis the retiree will try to balance his portfolio between risky and risk less assets so that expected marginal utility across the two classes is equal at the optimum (for a mathematical derivation of this idea see chapter 2 in merton 1992, especially eq. 2.4. the comparative statics of the optimum will then suggest how the optimum will change with respect to changes in the distribution of the risky portfolio—the standard conclusion is that the optimum portion of wealth invested in the risky asset decreases with increases in the risk of that asset. in order to address this objection, we present in fig. 1 the frequency plots for the 31 inflation-adjusted consumption amounts that were calculated on the basis of eq. (2). the frequency plots capture succinctly the magnitude of consumption, but more vitally, they capture the dispersion of the magnitude around the mean, and most importantly, the likelihood of extreme consumption outcomes. thus, our response to the above objection is based on empirical evidence on the likelihood of low consumption states from stocks versus other portfolios. in the frequency plots, the magnitudes and frequencies of consumption for the crsp vw portfolio, the t-bond based investment, and the best or optimal portfolio allocated between stocks and t-bonds are shown in different shades. the optimal portfolio is derived by calculating a, the portion of wealth invested in the crsp vw portfolio and (1-a) invested in t-bonds that would have maximized a retiree’s consumption during a particular sub period of our analysis. we use a computer program to find this a for every sub period. our program begins with a51.0 and calculates the consumption possible then decreases a by 0.05, then repeats the procedure till a50.0. the optimal a is the one that maximizes consumption. we will discuss below the composition of the optimal portfolio once we make some inferences based on the frequency plots. we present in fig. 1 three frequency plots, one corresponding to each of the 10-, 20-, and 30-year horizons. the 5-, 15-, and 25-year plots add little to the conclusions that we come to below, and therefore we choose to leave them out of the analysis. in order to make the interpretation of the frequency plots easier, it may be helpful to examine one such plot in some detail. by looking at the left hand side of the frequency plot for the 10-year horizon the following statements can be made: a retiree with a 10-year horizon would have consumed 0.72% or less of his wealth every month only once out of the 31 periods of data that are analyzed, if the retiree had invested in the optimal stock-bond portfolio. had this retiree invested in the crsp vw portfolio, he would have consumed 0.72% or less of his wealth three out of 31 times, and with t-bonds this number would have been five out of 31 times. the next three bars that appear in the frequency plot suggest the following: this retiree would have consumed 0.93% of his wealth every month in seven out 57p.p. vora / financial services review 9 (2000) 47–63 fig. 1. frequency plot of consumption amounts. 58 p.p. vora / financial services review 9 (2000) 47–63 of the 31 periods, with the optimal stock-bond portfolio or the crsp vw portfolio. however, with the t-bond portfolio this number rises to 16 out of 31 times. if 0.72% and 0.93% are considered to be low consumption, then it appears that the frequency of low consumption for the optimal portfolio and the stock portfolio is smallerthan for the t-bond portfolio. thus, although there appears to be more dispersion of consumption around its mean for the optimal and stock portfolios, there is more downside risk in the t-bond portfolio than in the other two portfolios. this is a new and very important finding that, to our knowledge, has not appeared anywhere in the literature. similarly, by analyzing the right hand side of the frequency plot, we find that frequency of high consumption (1.35%, 1.55%, and greater than 1.55%) is always greater for the optimal and the stock portfolios relative to the t-bond portfolio. in the extreme, the frequency of consuming more than 1.55% of wealth every month is zero for t-bonds while it is two for both the optimal and the stock portfolios. we conclude on the basis of the frequency plot that there is more upside potential in the stock and the optimal portfolios than the t-bond portfolio. this pattern gets stronger as the investment horizon gets longer. as the horizon increases from 10 to 20 to 30 years the frequency of low consumption from t-bonds increases while the frequency of low consumption from the crsp vw portfolio and the optimal portfolio decreases. at the 30-year horizon, the frequency plot shows that a retiree who invested in a t-bond portfolio would have been able to spend 0.48%or lessof his starting wealth every month with no uncertainty while a retiree who invested in the crsp vw portfolio or the optimal portfolio would have been able to consume 0.48%or more of his wealth with virtually no uncertainty. in fact, in only one period he is able to consume less than 0.48% of his wealth. thus, the little dispersion in consumption around its mean for the t-bond portfolio means that a retiree is practically guaranteed a low standard of living (relative to the stock based or optimal portfolios) while for the crsp vw or optimal portfolios low consumption occurs infrequently and high consumption always occurs more frequently than for the t-bond portfolio. we conclude from this analysis that as the horizon gets longer retirees would be well advised to invest in stocks over t-bonds. the “good” thing about the little dispersion in consumption based on the t-bond portfolio is that it is easy for the retiree to determine what portion of his wealth he should consume while with the large dispersion in consumption based on the crsp vw or optimal portfolios it is difficult to come to any general conclusions what consumption should be. however, this is really not a problem because the retiree cannot possibly go wrong by consuming 0.48% of his wealth from the crsp vw or optimal portfolios knowing that he can spend more if the market turns out to be favorable while recognizing that there is virtually no risk that he will outlive his wealth over the 30-year horizon. finally, we would like to point out that most retirees in the u.s. are eligible for social security payments. these payments are like interest payments from a relatively safe bond (clements, 1999). for retirees who are eligible for social security payments, the monthly consumption that we calculate is in addition to the floor consumption that the social security payments will allow them. this fact should mitigate, to a great extent, any remaining doubts that stocks offer the best consumption possibilities for retirees. 59p.p. vora / financial services review 9 (2000) 47–63 4.4. composition and timing of optimal portfolio we now turn our attention to the composition of the optimal portfolio. in panel a of table 4 we present summary statistics on the frequency with which the optimal portfolio is 100% invested in the crsp vw portfolio versus 100% invested in the t-bond portfolio versus a combination of stocks and t-bonds (an interior optimum). one obvious conclusion that we can make is that there are significantly more periods in which the optimal portfolio is 100% invested in stocks versus being 100% invested in t-bonds or a combination of stocks and t-bonds. another conclusion that can be made is that the number of periods in which the optimal portfolio is invested in stocks increases with the retirement horizon. this increase is generally at the expense of the number of periods in which the optimal portfolio is 100% invested in t-bonds. we can gain some additional insights into the retirement asset allocation decision by studying the correlation between the stage of the business cycle and the optimality of stocks versus t-bonds versus a combination of stocks and t-bonds. we obtain u.s. business cycle data for 1926 to 1995 from the national bureau of economic research (nber). for every month in which the economy expanded, we set the dummy variable contract to zero and when the economy contracted we set it to 1. additionally, for every month in which a 100% table 4 composition and timing of the optimal portfolio panel a: summary statistics on composition of optimal portfolio. horizon (years) number of periods out of 31 in which optimal portfolio has 100% stock optimum 100% t-bond optimum an interior optimum 5 22 7 2 10 21 4 6 15 21 1 9 20 21 1 9 25 25 0 6 30 25 0 6 panel b: correlation between business contractions and composition of optimal portfolio. horizon (years) 100% stock optimum 100% t-bond optimum an interior optimum 5 20.07** 0.15*** 20.02 10 20.19*** 0.14*** 0.19*** 15 20.05 0.12*** 0.05 20 20.09*** 0.12*** 0.09*** 25 20.12*** — 0.13*** 30 20.14*** — 0.14*** * significant at the 10% level. ** significant at the 5% level. *** significant at the 1% level. 60 p.p. vora / financial services review 9 (2000) 47–63 stock portfolio was optimal, we set the dummy variable stocks to 1, and to zero otherwise. likewise, for every month in which the 100% t-bond portfolio was optimal, we set the dummy variable t-bonds to 1, and to zero otherwise, and for every month in which a mix of stocks and t-bonds and stocks was optimal, we set the dummy variable interior to 1, and to zero otherwise. in panel b of table 4, we present the correlation coefficients of contract with stocks, t-bonds, and interior. we find that the correlation coefficient between contract and stocks is negative and statistically significant for all horizons except 15 years. for the horizons for which a correlation coefficient between contract and t-bonds can be calculated, we find it to be positive and statistically significant. the correlation coefficient between contract and interior is, with one exception, positive and statistically significant. when it is negative, it is statistically insignificant. based on these correlation coefficients, we conclude that a 100% stock portfolio is optimal during economic expansions. when the economy contracts, either t-bonds or a mix of t-bonds and stocks are optimal. mcnees (1987), however, demonstrates that it is difficult to predictex-antewhen the economy will next contract. in order to determine whether our main results are robust to the unpredictability of business cycles, we run one final test. we use each of the 840 monthly returns for stocks, t-bonds, and the two annuities to calculate monthly consumption, but instead of using them in the chronological sequence that they appear in our tests reported in table 2, we mix them up by drawing a random sequence of 840 returns, with replacement, from the set of returns that are at our disposal. the monthly consumption percentages based on the random sequence of returns and calculated from eq. (1) appear in table 5 together with the p-value of the wilcoxon sign rank test statistic. before we turn our attention to a discussion of the results we would like to place a note of caution here: there probably are table 5 monthly consumption based on u.s. financial market history with randomized returnsa horizon (years) stock t-bonds annuity a annuity b 5 2.08% 1.87% 1.93% 2.00% (0.016) (0.065) (0.225) 10 1.19% 1.02% 1.14% 1.18% (0.013) (0.200) (0.327) 15 1.20% 0.77% 0.78% 0.86% (,0.001) (,0.001) (,0.001) 20 0.98% 0.67% 0.72% 0.81% (,0.001) (,).001) (0.008) 25 0.93% 0.60% 0.66% 0.67% (0.039) (0.027) (0.048) 30 0.81% 0.54% 0.57% 0.71% (,0.001) (,0.001) (0.016) a monthly consumption is the consumption at the beginning of every month for every dollar of wealth, with nothing left over at the end of retirement. the percentages are an average over 31 retirement periods of equal length depending on the retirement horizon. the returns are drawn randomly (with replacement) from a set that contains actual monthly returns based on the history of u.s. financial markets. thep-values of the wilcoxon sign rank statistic for the null hypothesis that the t-bond and annuity portfolios provide consumption equal to the stock portfolio are presented in parentheses. 61p.p. vora / financial services review 9 (2000) 47–63 linkages between the stock market, the t-bond market, and the market for annuities that are not fully understood currently. by randomizing the returns data, we ignore the linkages that might have existed between the different returns data that we have at our disposal. we find that the stock portfolio based consumption is consistently greater than consumption based on other portfolios. however, our results are not as strong statistically as the results reported in table 2. in particular, at horizons of 5 and 10 years, the consumption based on annuity b is statistically virtually indistinguishable from stock based consumption. at the 10-year horizon, consumption based on annuity a is also statistically virtually indistinguishable from stock based consumption. apart from these two statistically insignificant results, the rest of the results that appear in the table are all statistically significant but in some cases not as significant as their counterparts in table 2. however, based on the results that appear in table 5 we conclude that our main result that at long horizons retirees are better off by being in stocks is robust to the unpredictability of recessions. 5. conclusions we study the history of u.s. financial markets from the perspective of a retiree because the voluminous literature on u.s. financial market history and on optimal asset allocation largely ignores the retiree’s asset allocation decision. we assume that retirees are primarily interested in maximizing their consumption stream. therefore, we calculate and compare the magnitude of consumption that would have been possible under various asset classes for various retirement horizons. we find that consumption based on a stock portfolio has been much higher compared to consumption based on a portfolio composed of t-bonds, or annuities based on the yield on a broadly diversified portfolio of investment-grade bonds or even moderately risky bonds. this statistical and economic dominance of stocks over bonds increases as we increase the time horizon for investment. when we compare the real consumption that is possible across different asset classes we find that stocks perform even better. we also find that the likelihood of low consumption is higher with t-bonds and this likelihood increases with the time horizon, while the likelihood of high consumption is higher with stocks and this likelihood also increases with the time horizon. interestingly, the few cases in which other than a 100% investment in stock was superior, seem to cluster in periods when the economy contracted. thus, we conclude that unless the economy is beset by an economic contraction, a retiree will be better off with a 100% allocation of wealth to stocks. however, we recognize thatex-ante the ability of any individual to predict the onset of an economic contraction is low. therefore, we provide additional evidence on monthly consumption when returns are drawn randomly from the set of monthly returns that we have. our results based on randomized returns are qualitatively consistent with our other results, but statistically they are not quite as significant, particularly at short horizons. thus, our study provides strong evidence that individuals should seriously consider remaining in stocks even after retirement, particularly at long retirement horizons, due to the larger size of the stock based consumption amounts as opposed to fixed-income security based consumption. 62 p.p. vora / financial services review 9 (2000) 47–63 acknowledgments the authors wish to thank the editor (vickie bajtelsmit) and two anonymous reviewers for their helpful comments. references bengen, w. p. (1994). determining withdrawal rates using historical data.journal of financial planning, 7(4), 171–180. brinson, g. p., hood, r. l., & beebower, g. (1986). determinants of portfolio performance.financial analysts journal, 42(4), 39–44. butler, k. c. and domian, d. l. (1993). long-run returns on stock and bond portfolios: implications for retirement planning.financial services review, 2(1), 41–49. bureau of labor statistics. visited august 7, 1997.bureau of labor statistics data: consumer price portfolioall urban consumers[www document]. urlhttp://stats.bls.gov/cgi-bin/surveymost?cu clements, j. (november 23, 1999). figuring asset mix? avoid mix-ups.the wall street journal,c1. conover, w. j. (1980).practical nonparametric statistics.new york, ny: john wiley & sons. constantinides, g. m. (1979). a note on the suboptimality of dollar-cost averaging as an investment policy. journal of financial and quantitative analysis, 14(2), 443–450. federal reserve board. visited january 30, 1997.federal reserve statistical release: selected interest rates [www document]. urlhttp://www.bog.frb.fed.us/releases/h15/data.htm huang, r. d., and stoll, h. a. (1997). the components of the bid-ask spread: a general approach.review of financial studies, 10(4), 995–1034. ibbotson, r. g. and sinquefield, r. a. (1989).stocks, bonds, bills, and inflation, historical returns (1926-1987). homewood, il: dow jones-irwin. jagannathan, r., & kocherlakota, n. r. (1996). why should older people invest less in stocks than younger people?federal reserve bank of minneapolis quarterly review, 20(3), 11–23. knight, j. r., & mandell, l. (1993). nobody gains from dollar cost averaging: analytical, numerical and empirical results.financial services review, 2(1), 51–61. mcnees, s. k. (1987). forecasting cyclical turning points: the record in the past three recessions.new england economic review,(march/april), 31–40. merton, r. c. (1992).continuous time finance.cambridge, ma: blackwell publishers. mitchell, o., poterba, j. m., & warshawsky, m. j. (1997). new evidence on the money’s worth of individual annuities.nber working paper 6002. national bureau of economic research. visited march 3, 2000.u.s. business cycle expansions and contractions [www document]. url http://www.nber.org/cycles.html posner, r. a. (1995).aging and old age.chicago, il: university of chicago press. poterba, j. m. (1997). the history of annuities in the united states.nber working paper 6001. rozeff, m. s. (1994). lump-sum investing versus dollar-averagingjournal of portfolio management, 20(2), 45–50. siegel, j. j. (1994).stocks for the long run: a guide for selecting markets for long-term growth.new york, ny: mcgraw-hill. statman, m. (1995). a behaviorial framework for dollar-cost averaging.journal of portfolio management, 22(1), 70–78. 63p.p. vora / financial services review 9 (2000) 47–63 pii: s1057-0810(99)00008-6 the sociology of personal finance chris robinsona, elton g. mcgounb,* aassociate professor of finance, schulich school of business, york university, north york, ontario m3j 1p3, canada bassociate professor of finance, department of management, bucknell university, lewisburg, pa 17837, usa abstract finance in general and personal finance, in particular, assume that there is a pure market money. the financial resources of a business or a household are taken to be a single mass made up of indistinguishable dollars, marks, yen, pounds, francs, or whatever. consequently, we are free to devise “rational rules” for managing this mass, for prescribing how a business or household should choose the appropriate forms of money and the appropriate accounts for money without having to look more closely at the money itself. in this paper, we argue that rational behaviour is a more complex and richer process than simply valuing market money, since there are qualitative characteristics attached to any money, however defined. © 1998 elsevier science inc. all rights reserved. 1. different views of money the economic literature makes numerous references to the symbolic value of money in its role as a medium of exchange. a man does not hold money for its own sake, but for its purchasing power—that is to say, for what it will buy. therefore his demand is not for units of money as such, but for units of purchasing power. (keynes, 1965, p. 53) money is, as the classical economists said, both a medium of exchange and a “measure of value.” it is symbolic in that, though measuring and thus “standing for” economic value * corresponding author. tel.:11-570-577-3732; fax:11-570-577-1338. e-mail address:mcgoun@bucknell.edu (e.g. mcgoun) financial services review 7 (1998) 161–173 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00008-6 or utility, it does not itself possess utility in the primary consumption sense—it has no “value in use” but only “in exchange,” that is for possession of things having utility. (parsons, 1967, p. 306) as money is a generalized medium, it does not symbolize a specific thing; rather, it symbolizes the utility obtainable when the money is exchanged for things (mcgoun, 1997). since as far as economics is concerned money can be exchanged by any one for any thing, money is free to symbolize any thing any one happens to desire. this makes it an especially potent symbol; it is “pure” wealth. exchange value forms the substance of money, and exchange value is wealth. money is therefore, on another side, also the embodied form of wealth, in contrast to all the substances of which wealth consists. . . . money is therefore the god among commodities. (marx, 1973, p. 221) gold [money] is a wonderful thing! its owner is master of all he desires. gold can even enable souls to enter paradise. (columbus, in his letter from jamaica, 1503, quoted in marx, 1976, p. 227) finance in general and personal finance in particular make use of this purity of money. the financial resources of a business or a household are taken to be a single mass made up of indistinguishable dollars, marks, yen, pounds, francs, or whatever. it certainly does not matter where money has come from; once it becomes part of the mass any previous identity which may have been associated with its origin is lost. parts of the mass may be held in different forms of money (cash, transactions accounts, securities, etc.), but there are no constraints on conversion of one form into another. and the mass may be broken down into a number of budgetary accounts (operating expenditures, capital expenditures, household funds, retirement funds, etc.), but transfers from one account to another can occur freely at the discretion of whoever it was who created the system. consequently, we are free to devise “rational rules” for managing this mass, for prescribing how a business or household should choose the appropriate forms of money and the appropriate accounts for money without having to look more closely at the money itself. what we mean by “rational rules” is that no non-economic values enter into the calculations; that is, that the purpose of the rules for managing money is to make more (or lose less) of it. and whatever form or whatever accounting we may choose for the money within our business or household, it is all the same money, market money, when it leaves the business or household and exercises itself as “purchasing power.” because of this seductive purity of money, there is a tendency for it to penetrate more and more aspects of life. zelizer (1994, p. 11) summarizes in the following way this traditional view of money, which is implicitly employed in finance: 1. the functions and characteristics of money are defined strictly in economic terms. as an entirely homogeneous, infinitely divisible, liquid object, lacking in quality, money is a matchless tool for market exchange. even when the symbolic meaning of money is recognized, it either remains restricted to the economic sphere or is treated as a largely inconsequential feature. 162 c. robinson et al. / financial services review 7 (1998) 161–173 2. all monies are the same in modern society. what simmel (1978) called money’s “qualitatively communistic character” denies any distinction between types of money. only differences in quantity are possible. thus, there is only one kind of money— market money. 3. a sharp dichotomy is established between money and non-pecuniary values. money in modern society is defined as essentially profane and utilitarian in contrast to noninstrumental values. money is qualitatively neutral; personal, social, and sacred values are qualitatively distinct, unexchangeable, and indivisible. 4. monetary concerns are seen as constantly enlarging, quantifying, and often corrupting all areas of life. as an abstract medium of exchange, money has not only the freedom, but also the power to draw an increasing number of goods and services into the web of the market. money is, thus, the vehicle for an inevitable commodification of society. 5. there is no question about the power of money to transform nonpecuniary values, whereas the reciprocal transformation of money by values or social relations is seldom conceptualized or else explicitly rejected. while this traditional view of money may be very useful for a number of purposes, especially business purposes, another view of money might also be useful for the purpose of understanding personal finance. zelizer (1994, p. 18) also summarizes in the following way a non-traditional view of money, which contrasts with the list in the preceding section: 1. while money does serve as a key rational tool of the modern economic market, it also exists outside the sphere of the market and is profoundly influenced by cultural and social structures. 2. there is no single, uniform, generalized money, but multiple monies: people earmark different currencies for many or perhaps all types of social interactions, much as they create distinctive languages for different social contexts. and people will in fact respond with anger, shock, or ridicule to the “misuse” of monies for the wrong circumstances or social relations, such as offering a thousand-dollar bill to pay for a newspaper or tipping a restaurant’s owner. money used for rational instrumental exchanges is not “free” from social constraints but is another type of socially created currency, subject to particular networks of social relations and its own set of values and norms. 3. the classic economic inventory of money’s functions and attributes, based on the assumption of a single general-purpose type of money, is unsuitably narrow. by focusing exclusively on money as a market phenomenon, it fails to capture the very complex range of characteristics of money as a social medium. a different, more inclusive coding is necessary, for certain monies can be indivisible (or divisible but not in mathematically predictable portions), non-fungible, non-portable, deeply subjective, and therefore qualitatively heterogeneous. 4. the assumed dichotomy between utilitarian money and non-pecuniary values is false, for money under certain circumstances may be as singular and unexchangeable as the most personal or unique object. 163c. robinson et al. / financial services review 7 (1998) 161–173 5. given these assumptions, the alleged freedom and unchecked power of money become improbable. cultural and social structures set inevitable limits to the monetization process by introducing profound controls and restrictions on the flow and liquidity of monies. 2. “rationality” and “irrationality” in financial economics, we accept market money as the basic foundation of all of our thinking. we can reduce any set of cash flows, or any asset held for its cash flows, to a single market value by appropriate discounting, allowing for differences in risk. conversely, we define asset values as the discounted value of future cash flows. we make different assets or forms of market money comparable, and can then rank their values. behaviour towards different forms of market money that are equal by this calculus is rational behaviour if it treats them the same, and irrational otherwise. in this paper, we argue that rational behaviour is a more complex and richer process than simply valuing market money, since there are qualitative characteristics attached to any money, however defined. a different treatment or attitude or valuation of sums of market money that are equal in market terms may be perfectly rational when the qualitative characteristics differ materially. the current academic literature in finance suffers from selective amnesia when it comes to utility maximization. the only argument in the formal models is wealth, and thus we have explicitly reduced utility maximization to wealth maximization, subject to whatever budget constraints are relevant in a given research question. as finance researchers we know that utility maximization should also include non-wealth considerations, but the tractability and elegance of the current models is so attractive that we ignore these other arguments in the objective function. any field of inquiry that purports to describe and understand observable phenomena must embrace all of the reality, sooner or later. we do not claim that apparently irrational behaviour with respect to money never exists. individuals or social groups can suffer from misperceptions or lack of knowledge of the quantitative characteristics of money in some form that will cause them to behave in ways which appear irrational, where we would agree that their understanding is wrong in a purely objective fashion. the dividing line between an objective reality and subjective perception is not a defined boundary, and what one person thinks is totally irrational may be the manifestation of some subjective characteristic of money valued by another person. let us consider some examples of asset allocations to illustrate the issue: family a . . . obeys the muslim injunction against receiving or paying interest (sharia). this family’s investment portfolio is entirely in common equity and real estate. 164 c. robinson et al. / financial services review 7 (1998) 161–173 family b . . . believes inflation will increase rapidly around the world in the near future. accordingly, this family’s portfolio consists of real estate (heavily levered with fixed rate debt, since the lenders have not allowed for the higher inflation rates, in the family’s opinion) and gold. family c . . . often finds themselves at social events at which investments are discussed. this family’s investment portfolio consists of stocks and other investments about which there has been considerable recent publicity, and they trade very frequently. family d . . . subscribes to a number of economic databases on the internet and follows them assiduously. at any given time, they hold a very small number of investments in their portfolio; however, its contents change rapidly as they trade daily based upon their assessments of the impact of macroeconomic variables on market performance. family e . . . has a very successful business that grew out of a hobby of the spouse who formerly did not work outside the home. all of its profits are either reinvested in the business or in short-term money-market instruments. family f . . . received a large inheritance of government securities from a wealthy relative who had grown up during the depression. they have not touched the money since receiving it quite some years ago and have no special plans for it. these families are not atypical, and we think that most of us have personally known several of them, or families very much like them. and we all know that from the standpoint of traditional finance, these families are acting irrationally. they are not maximizing their wealth, as a result of: (1) concerning themselves with non-economic factors (families a and c); (2) forming irrational expectations (families b and d); (3) making insufficient investment in portfolio management (families e and f); (4) and, of course, not adequately diversifying their portfolios (all families). but we would be hard-pressed to argue that from a sociological or psychological standpoint these families are acting irrationally. there are certainly no reasons why personal financial management should be subordinate to religious interests, why personal anxieties should not be important considerations in personal financial management not only in risk tolerance but also in the formation of expectations, why personal financial management should not also be a form of entertainment, or why families not very interested in it should spend much time at personal financial management. 165c. robinson et al. / financial services review 7 (1998) 161–173 at present, there is rising interest in something calledbehavioral finance.but what behavioral financecurrently is is work which refers to the body of knowledge of cognitive psychology as support for what might be calledquasi-rationalassumptions, to use thaler’s (1992) phrase, but employs the same methods and the same methodology as traditional finance. of course, one might argue that these new assumptions are morerealistic than those of implicitly self-proclaimedrational finance and reflect a concern withrealismat odds with friedman’s (1953) instrumentalism. but they themselves are certainly unrealistic, and arguably even more unrealistic than the current assumptions since there is no reason to believe that individuals with biases might aggregate in such ways as this work assumes. it is necessary to distinguish betweenlogical andrational (devlin, 1997). what traditional finance callsrationality is logic and what it pejoratively callsquasi-rationalitylegitimately deserves the more reputable termrationality, given the current theories of evolutionary psychology (pinker, 1997). we actually do observe thelogical behavior assumed in traditional finance—in subjects who have suffered certain forms of brain damage (damasio, 1994). we intend to show in this paper just how there are rational actions which do not currently fall under finance’s definition of the term. 3. the colour of money what all of this means is that when we concern ourselves with real households inhabited by real people rather than economic units made up of genderless economic persons, the colour of money matters. therefore, personal finance research should consider the qualitative characteristics, and how they affect both our observations and empirical work, and the normative advice that we give. let us propose some headings in order to organize our ideas of how the colour of money matters: where money comes from: the payment form, source or recipient; in what form money is held: coins, bills, bank accounts, credit cards, securities, real assets; how money is labeled: special accounts, who controls it, who spends it, credit and debit cards versus cash; what sums of money are involved. there are numerous familiar examples to illustrate these points. we list and explain some of them in the spirit of demonstrating that personal finance research and teaching should expand their horizons. we do not claim to provide a new theory of personal finance. 3.1. where money comes from ethical investment funds cater to the increasing acceptance that corporations should act to maximize something more than shareholder wealth. ethical funds assess the suitability of companies based on ‘screens’ that either disqualify them completely or give them scores that are compared. for example, most ethical funds will not hold securities of tobacco companies. they will assess the commitment of chemical companies to minimize pollution and hold only 166 c. robinson et al. / financial services review 7 (1998) 161–173 those that meet a higher standard, however determined. the excluded companies may be very profitable. we know from basic finance theory that constraining the set of allowable investments forces the investor to hold an inefficient portfolio, and thus any ethical investor is getting a poorer risk-return combination than he or she could get. in fact, if enough investors refuse to hold certain companies, we would expect abnormal profits to be available for the buyers of the unwanted securities. in budgeting, we treat all sources of income identically and add them up. however, a formal model based on some theory of smoothing lifetime income and consumption would allocate a large part of any windfall into savings rather than expenditures. families may not behave as the models suggest they should. one family may view all receipts as income to be spent at once with no saving. thus, a windfall source of cash may do nothing to alleviate a low standard of living, since it is consumed very quickly. another family may view a windfall as an amount that must be set aside and placed in savings. a particular social view of welfare income is the preference of many governments to provide it in a tied form. food stamps and housing subsidies are obvious examples. we know that utility would be maximized by giving the same economic value without constraints, as tied programmes usually cost more to administer, and more money would be available for distribution if it were given as straight cash. nonetheless, tied programmes persist because society (or at least the part of society that pays more tax than it receives in welfare) has a different image of welfare income. it must be spent on only the necessities of life, as defined by those who know better. ‘excessive’ amounts of booze, cigarettes, entertainment and children’s toys are undesirable. we are unaware of any personal finance research that tries to estimate the lost utility due to tied programmes. gifts may be another form of money, when given as cash or near-cash items. for example, in some cultures, wedding gifts always take the form of cash so that the newly-weds can buy whatever they need to start life together. in other cultures a gift of cash from anyone except perhaps the parents is considered vulgar, and the friends and relatives must buy and wrap real assets—dishes, small appliances, wine, etc. if the culture that gives the money imposes no implicit restrictions on how the couple spends it, then the cash gifts have more utility because there is no constraint. however, if the society expects that only certain sorts of things should be bought with money gifts, then binding constraints exist again, and the money is not ‘worth’ as much as purely market money would be. how executives are paid may matter, too, when bonus and salary are considered. in a recent example, gibbon (1998) reported that royal oak mines, a gold mining company, lowered the exercise price of previously-issued executive stock options because the share price had dropped so far the options were almost worthless. many shareholders were very upset, since of course they had also lost money on the share price decline, and lowering the exercise price diluted their value even further. the motion passed by a close vote at the annual meeting, although most motions at annual meetings are rubber-stamped. the ceo said the options were the only means of retaining senior staff, since their salaries and bonuses had been frozen for two years, while royal oak wrestled with bringing a new mine into production at a time when gold prices were falling. it seems that the executives considered the options to be salary, perhaps with a bit of 167c. robinson et al. / financial services review 7 (1998) 161–173 variability in total amount. nonetheless, they had been consuming or planning to consume both salary and options. many shareholders believe the options are a form of reward or incentive, and if the company does badly the value should be low or zero. when the company originally granted the options, the shareholders and executives would have agreed on the market money value of the options, valued by an appropriate-options pricing model. what they did not recognize was that the two sides had a very different view as to the qualitative nature of the payment being given. an old aphorism in personal finance is to spend only your income but never encroach on your capital. this aphorism and the strong beliefs underlying it create a complex problem in personal finance, especially at the retirement stage of the life cycle. in purely rational or market money terms, only one interpretation is valid: a very high utility of bequest function. persons have limited lives and therefore they can afford to consume their capital, too. for all but the quite wealthy few, consumption of capital in retirement is a necessity for a comfortable lifestyle or even bare subsistence. uncertainty about the rate of return will cause you to rationally curtail your consumption to what is estimated to be sustainable, given an uncertain life expectancy (see milevsky et al., 1997), but the consumption amount includes both income and capital. pensions and annuities pay out both capital and income in a level or indexed stream of cash flow. thus, the form in which the money is received from two identical endowments may affect the amount consumed. one endowment might be invested entirely in treasury bills, paying 4% interest. the other endowment might be in an equity mutual fund with 2% dividend yield and 8% unrealized capital gains. if the families follow the traditional belief that you should never consume capital, the first family will have twice the consumption of the second family, even though the second family actually earned more. miller and modigliani (1961) created a world in which dividends are irrelevant. in the real world we know they are relevant, since we observe their continued existence in spite of double taxation. part of the dividend puzzle may relate to this belief that you shouldn’t consume capital. the dividend payout may be consumed, and higher dividend payout companies permit more consumption. which member of the household earns an income amount also affects how it is valued, and the valuation varies widely according to the circumstances of the family and the society in which it is situated. children’s earnings from a paper route in a middle class canadian or us family are essentially a luxury, or totally discretionary income for the child. even if the family imposes certain discipline (“save 10%,” “buy your own treats with the money”), it is not a part of the income that the family uses for its household budget. on the other hand, a family on welfare may need the children to earn some money to buy the latest new clothing, and hence the children’s earnings become part of the budgeting process. a family in a third world country may require the children’s earnings to avoid hunger. a different aspect of who earns the income is how it is allocated to spending and saving. until quite recently in history, women did not work outside the home once they were married. now two income families are common, and the allocation of the two incomes does not necessarily correspond to formal budgeting models. regardless of which spouse earns them, take-home dollars are identical in market money, but they are not equal in their value 168 c. robinson et al. / financial services review 7 (1998) 161–173 to some families. a common scheme is to save the entire second income, yet it would not be unusual for the “second” income to have become substantially larger than the “first” one. the budgeting process we teach in a personal finance course advises you to add up all the income and determine how much to spend on different categories. the residual is saving. the process is iterative, because one of the family’s goals will be to achieve savings targets, and hence the family should budget to reduce expenditures if the residual is insufficient. if the family uses the model of saving the entire second income, it is skipping the budgeting process, and may save less or more than the goals it would set if it used the other process. a similar effect occurs if different household expenditures are assigned to each income earner. thus, who earns the money affects the budgeting process, even though the dollars are the same. a more subtle effect is the way in which second incomes are characterized. a term that has become almost pejorative is “pin money,” the descriptor for small amounts of incidental income earned by a wife who does not work full-time. the implication is that it is frivolous, or unimportant, and thus worth less than the same number of dollars earned by the main breadwinner. a wife’s personal income may be valued differently by each spouse as well. a wife who has been dependent upon her husband for all her money for many years may place greater value on income that she earns when she starts working, than she did on the same number of dollars received from her husband. objectively, it may cost quite a lot to earn such income —child care, a second car, new clothing, more expensive meals—but the psychological feeling of independence makes the second income more valuable. in addition, the social benefits of working outside the home after years of housework may provide even greater psychic benefits. consider the following situation as an experiment that could be conducted, or as a behaviour that could be observed if sufficient data could be collected. a family unit can increase its income by a specific take home amount (that is, after taxes and deductions) in one of two ways. the principal breadwinner can work some extra hours, or the homemaker spouse can work at a part-time job. the two are mutually exclusive because one spouse must be home to care for the children. the risk of the two income amounts is identical. we have already suggested that other factors may affect the choice of which method they use to increase income. now, consider how they spend or save the extra income. would the extra amount be allocated exactly the same, regardless of who earned it? furthermore, would the allocation of the income depend on whether the homemaker was male or female? we don’t know the answers, but our belief is that for some families the money would be allocated the same no matter who earns it, and for others the allocation would be dependent on who earned it. 3.2. in what form money is held one way to categorize the holding of money is as currency, coinage, bank account accessed by debit card, credit card paid in full each month, or bank account accessed by check. the rational way to treat money in these different forms is simply a comparison of transaction costs and opportunity costs of money held in a bank account paying low interest. for example, bills and currency are most convenient for small transactions or for paying 169c. robinson et al. / financial services review 7 (1998) 161–173 merchants or machines that do not accept anything else, but they also carry a theft risk and the nuisance of having to go to a financial institution or an atm to get cash. a credit card may carry a fee, and also a risk of theft and forgery, but it is very convenient, and the free credit period allows money to be held temporarily in a form earning more than a bank account. it may also convey various other financial benefits such as affinity points or free insurance. there are other aspects to the form of holding money, however. the best known is the nature of spending different forms of money. if a family goes out for a night on the town and takes a credit card, will it spend more or less than if it takes a wallet with $300 in it? does it create a different feeling to pay for an expensive audio system with a roll of $1,000 bills, or does tendering the most exclusive platinum credit card create a better image? does it matter what denominations of bills or coins are used? for example, many people save pennies. every time they have an odd amount to pay in cash in a store, they pay with an even amount, and all the pennies in pocket or purse go into a bowl at the end of the day. a family could save more easily by putting a dime or quarter into the bowl each day and avoid having to roll all the pennies some day. however, the idea of the huge bowl of pennies is somehow appealing, and makes savings seem more concrete, as well as less painful to achieve. another example is the canadian $2 bill on the prairies. this bill no longer exists, having been replaced by the $2 coin. during its long history, it was said that it was uncommonly tendered in the prairies provinces because it had been the price of a prostitute for a long time, and to be seen using one had an aura of immorality. this story was firmly believed by many people, which would make it self-fulfilling to some extent. one of the authors believes that he saw far fewer $2 bills than he would expect during his various trips to the prairie provinces, compared with what he saw in the central provinces, but that may be an illusion. 3.3. how money is labeled the labels we put on different sums of money are closely related to the forms in which they are held. even with identical forms, however, the labels have different meanings, and create different behaviour. the point in labeling is generally in the budgeting system, and is a means of control or enforced saving. creating different accounts, or holding money in cash form for long periods, are inefficient in a rational sense. transaction costs rise and income on the money may be lower. nonetheless, enforced saving is necessary for many families. we have already mentioned that money may be permanently labeled by source, and from then on its fate is may be quite precisely determined. but we can be equally adamant adhering to our voluntary labels. even without tax implications, money set aside for retirement or education will only in extreme circumstances be used for something else, even though it may not be all that necessary for its originally intended purpose. if a certain sum is put away in a christmas club account or a vacation club account, it will be used for that purpose, even though that might mean a more lavish christmas or vacation than would really be prudent given other circumstances. in fact, the very existence of special retirement, 170 c. robinson et al. / financial services review 7 (1998) 161–173 education, christmas, and vacation accounts points out the importance of labels in disciplining personal financial management. 3.4. what sums of money are involved we have long known that money does not obey the laws of mathematics. although utility might, wealth does not, since utility and wealth are not directly proportional. the very foundations of finance in the work of daniel bernoulli in the eighteenth century assumed that monetary losses and gains were less important to a richer person, who would be willing to undertake gambles that a poorer person would rationally avoid. more recently, work in behavioral finance has shown that losses and gains are asymmetrical; that is, that more utility is lost with an investment loss than is gained with an equivalent gain. this is not so surprising, nor so “irrational” from a utility standpoint, since it certainly makes sense that we would value more what we have had for a while than what we have not had time to accustom ourselves to yet. other examples of the phenomenon are less familiar. there are some things that are obtainable for large sums of money that are unobtainable with sums of smaller sums. one example suggested by simmel (1978) is political influence. the principle of one person one vote notwithstanding, someone whose net worth is one million dollars is likely to have more influence over government decision-making than ten persons who have one hundred thousand dollars each. at the same time, a person who has ten million dollars probably will not have more influence than ten millionaires. similarly, having sufficient money to purchase the least expensive house in a prestigious neighborhood can give someone vastly greater social standing than someone who can not quite afford it; but not that much less social standing, at least as far as residence is concerned, than someone who is able to live in one of the most expensive there. numbers and sums themselves have psychological value. one of the most common examples of the phenomenon is stock indices, round numbers of which may represent “floors” or “ceilings.” once the dow has passed 9,000 on the way up, a certain optimism may be triggered that accelerates the rise. or when the nikkei finally falls below 15,000, a despair falls over the market which may lead to a continued decline. more relevant to personal finance is another common example which appeared in the preceding paragraph. there is something about being a “millionaire” that one does not have being a “nine hundred ninety thousand-aire.” or if one has set an investment target of $100,000, once having reached that amount it is possible or perhaps permissible to imagine possibilities that were not thought of before that time. amounts of money have sociological as well as psychological value. as noted by zelizer (1994), there is a protocol associated with using certain forms of money for certain purposes that is not necessarily related to transactions costs. small purchases can generally not be made with small bills or with credit cards, but some can. more significant along these lines are the values of gifts. there are fairly precise and often quite elaborate, but unwritten, rules concerning how much you can spend on what occasion. and there can be quite severe penalties for falling short or exceeding that amount by not so large margins. 171c. robinson et al. / financial services review 7 (1998) 161–173 4. why does this matter to academic personal finance? the colour of money affects human behaviour individually, in families and in larger groups or societies. the reader can think of many more examples than we have presented in this paper, and can enlarge on every aspect of what we have outlined. the reader may also disagree with specific examples, in the belief that they reflect purely irrational behaviour rather than social norms. we think the general argument is undeniably valid, however. the question then is, so what? first, teachers of personal finance need to reflect both the rational economic models for students to learn to improve personal financial management (their own or that of their future clients) and the reality of what people do and why. financial advisers of every form must deal with their clients in a world that does not obey the limiting assumptions of our formal financial models. if they insist on trying to force their clients into uniform plans that do not take account of the broader social situations these families are situated within, they will either lose the clients or possibly convince them to act in ways that may reduce their utility. second, researchers in personal finance need to recognize that their apparent empirical findings may be coloured, changed or even invalidated by the sociological phenomena and situational nature of the family, because they have assumed these away. perhaps the small firm effect or the january effect has rational causes grounded in our society, causes that are not directly related to maximizing wealth and hence not easily found by purely rationalist, objectivist financial research or even the purely quasi-rationalist but nonetheless still objectivist financial research of behavioral finance. third, related to our second point, is the question of the appropriate methods in personal finance research. the dominant paradigm comes from corporate finance and investments, and is rigidly quantitative. there is a huge literature on method and methodology in the social sciences in general that finance has ignored. in particular, we rarely use qualitative methods, though they are much commoner everywhere else in the social sciences. to give one example, we might learn a great deal about budgeting if we participated directly in the budgeting processes of families and recorded our observations in a systematic way. this paper is not the place to go into extensive detail, but several different ways of doing it— action research, case study method, grounded field theory—exist and have been developed thoroughly. bettner et al. (1994) provide an introduction to qualitative research in the context of corporate finance and investments. fourth, we think that investigation of these sociological or psychological aspects of personal finance offers a great opportunity to expand the scope of the field. many of the examples we suggest lead to clear research questions. we observe that personal finance research as represented in the programmes of theacademy of financial servicesand the association for financial counseling and planning educationand their respective journals, financial services reviewand financial counseling and planningare narrower than is necessary or appropriate. investments, government income/pension/welfare/taxation laws and institutional arrangements seem to dominant the discourse. we think there is a lot more to personal finance, though we agree that these are critical topics, and do not wish to minimize the usefulness of these lines of inquiry. 172 c. robinson et al. / financial services review 7 (1998) 161–173 references bettner, m., robinson, c., & mcgoun, e. (1994). the case for qualitative research in finance.international review of financial analysis 3(1), 1–18. damasio, a. r. (1994).decartes’ error: emotion, reason, and the human brain.new york: g. p. putnam’s sons. devlin, k. (1997).goodbye descartes: the end of logic and the search for a new cosmology of the mind.new york: wiley. friedman, m. (1953). the methodology of positive economics. inessays in positive economics, (pp. 3–43.) chicago: the university of chicago press. gibbon, a. (1998). royal oak to reduce exercise price of options.the toronto globe and mail, june 27, b1. keynes, j. m. (1965).a treatise on money.london: macmillan & company. marx, k. (1976).capital. london: penguin books. marx, k. (1973).grundrisse.new york: random house. mcgoun, e. g. (1997). hyperreal finance.critical perspectives on accounting 8, 97–122. milevsky, m. a., ho, k., & robinson, c. (1997). asset allocation via the conditional first exit time or how to avoid outliving your money.review of quantitative finance and accounting 9, 53–70. miller, m., & modigliani, f. (9161). dividend policy, growth and the valuation of shares.journal of business 34, 411–433. parsons, t. (1967).sociological theory and modern society.new york: free press. pinker, s. (1997).how the mind works.new york: w.w. norton & company. simmel, g. (1978).the philosophy of money.london: routledge & kegan paul. thaler, r. h. (1991).quasi-rational economics.new york: russell sage foundation. zelizer, v. a. (1994).the social meaning of money.new york: basicbooks. 173c. robinson et al. / financial services review 7 (1998) 161–173 pii: s1057-0810(99)00031-1 international equity diversification and shortfall risk kwok hoa, moshe arye milevskyb, chris robinsonb,* aatkinson college, york university, north york, ontario m3j 1p3, canada bschulich school of business, york university, north york, ontario m3j 1p3, canada abstract international equity diversification benefits canadian investors very substantially by reducing shortfall risk, as shown by results of a model that minimizes the risk of shortfall from a desired consumption level for a retired investor with an unknown date of death and stochastic investment returns. it does not benefit american investors materially. the united states equity market is a large proportion of the international equity market that is available to individual investors, and united states returns are highly correlated with other markets. © 1999 elsevier science inc. all rights reserved. jel classification:g11; g15; g23 keywords:international diversification; shortfall risk; asset allocation; asian options 1. introduction the benefits of international portfolio diversification have received considerable attention in the investment literature. using data on major international markets, numerous studies have documented the benefits from the viewpoint of reducing total risk without sacrificing expected returns. a smaller number of papers have questioned the benefits of international diversification. in this paper we use a different technique to try to resolve this conflict. specifically, we use minimization of shortfall risk as the choice criterion for an individual investor who is retired, and hence has no further labour income to support consumption, only an endowment from which he or she can consume both principal and income. our model * corresponding author. tel.:11-416-736-5072; fax:11-416-736-5687. e-mail address:findoctr@interlog.com (c. robinson) financial services review 8 (1999) 11–25 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00031-1 allows the individual to allocate this endowment among various assets classes to determine which combination will provide the smallest risk of falling short of a pre-determined desired level of consumption. we compare the minimum shortfall risk portfolios for different possible choice sets. we find that a retired canadian investor can do much better, where “better” means lower shortfall risk, by investing in a portfolio that includes international equity, compared with a portfolio that allows only domestic assets. by contrast, an american investor in the same situation (i.e. same age, gender, wealth and desired consumption) cannot reduce shortfall risk materially by investing in international equity. our approach allows both the rate of return and the date of death of the individual to be stochastic. 2. literature review in one of the earliest studies, grubel (1968) finds that international diversification pays off from the viewpoint of an american investor—a diversified portfolio of international stock indices dominates the united states index in terms of risk and return. in two related studies, levy and sarnat (1970) and grubel and fadner (1971) confirm this result and show that the japanese and south african indexes play an important role in the efficient portfolios. errunza (1977) extends the work to diversification in less-developed countries. more recent studies confirm and extend these results: adler and dumas (1983), bailey and stulz (1990), cosset and suret (1995), grauer and hakansson (1987), hunter and coggin (1990), jorion (1985, 1989), levy and lerman (1988) and obstfeld (1994). some studies have considered the situation of investors outside the united states. mcdonald (1973) and solnik (1979) find that international diversification is attractive for investors whose home countries are not the united states. on the other hand, however, there are studies which question the benefit of international diversification. sinquefield (1996) concludes that the empirical evidence for the 1970–94 period does not support the belief that the international equity market has higher expected returns than the american equity market, nor that it can substantially diversify american portfolios. odier and solnik (1993) show that the benefits depend in part on whether one expects cross-country correlations, market volatilities and currency risk to change. while currency risk remains a small, although significant, component of total equity risk, there is no easy formula for determining whether to hedge currency risk, or how much of it to hedge. goldberg and heflin (1995) suggest that increasing the degree of international involvement decreases systematic risk but increases total risk. haavisto and hansson (1992), in a study of nordic stock markets, conclude that the optimal portfolios are extremely concentrated and include in general only finnish and swedish assets, and that a long-term investor would have done very well by keeping an unhedged and diversified nordic portfolio. these studies use the techniques of modern portfolio theory. they calculate the correlation coefficients of returns in various markets, and the returns of combined portfolios. the criterion for benefits of international diversification is related to the degree to which standard deviation of a combined portfolio is lower than a single country portfolio. some studies apply a utility function explicitly. 12 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 3. minimizing shortfall risk as a criterion we use a different method to investigate the benefit of international diversification. we postulate that individual investors are likely to wish to minimize the probability of falling short of predetermined goals for wealth and consumption. the shortfall approach thus combines both an investor’s objective and a measure of risk. other papers have also recommended this method, including liebowitz and kogelman (1991), milevsky et al. (1997) and tse et al. (1993). in this paper we focus on the asset allocation decision faced by a person who is retired and must live solely on the capital saved during the working life, and the income from it. the person has an unknown time remaining until death. the date of death is distributed according to standard mortality tables. the person has wealth, or capital, ofw, and wishes to consume a fixed real amount,c, every year in retirement. the investment returns are uncertain and he or she will take thisc first from income, and then from capital if income is insufficient. this person wishes to avoid shortfall in consumption every year, and therefore the objective is minimize the probability of being able to consume less thanc in any year prior to death. you could think of the problem in two equivalent forms. the retiree has a sum of money that fluctuates each year as it earns income and he consumes part of it. what is the probability that the money will run out before the date of death, given a distribution of investment returns? alternatively, you could think of it as a stochastic present value problem.w is the initial investment. the future consumption is the uncertain future cash flow, with an uncertain number of years. what is the probability that the present value of the future consumption stream will be greater thanw? in either case, the probability specified is the shortfall probability. a solution for the probability for any given distribution would also permit identification of the minimum shortfall risk portfolio for any number of assets. this is a problem in stochastic optimization with two random variables—date of death and investment return. milevsky et al. (1997) provide a solution in a form which requires simulation. milevsky and robinson (1999) derive an analytic solution using insights from the pricing of asian options. the solution provides the probability of shortfall for any allocation amongn investment assets, given: ● initial wealth,w ● the periodic amount to be withdrawn (i.e. principal 1 income) for consumption,c ● a mortality table or mortality function for the person ● the distribution of continuously-compounded investment returns for each class of assets. the details appear in milevsky and robinson (1999). we include a summary of the derivation in appendix 1. one useful feature of the solution is that the result is identical for all ratios ofw/c, regardless of the size ofw andc. accordingly, we will refer tow/c in the rest of the paper, rather than separate values. the reader should remember that this refers to initial wealth, and fixed real consumption every year until death. it is not a constant ratio every year, becausew is variable. to visualize what a wealth to consumption ratio might mean, and what values might be reasonable, let us consider how much a person might need to consume per year in retirement. 13k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 we are abstracting from the form of the income—taxes are ignored, pensions are implicitly capitalized, etc. suppose you want your financial capital to provide $30,000 per annum in real dollars, until you die. if you have $360,000 now,w/c5 12, which is the lowest value we will provide in our empirical section. if you have $720,000,w/c 5 24, which is the highest value we will use. the absolute level of wealth is not relevant in how investments should be allocated, but rather the wealth relative to what you require from it. to look at it another way, suppose you are exactly a millionaire at age 65. the popular view is that a millionaire can enjoy a luxurious retirement lifestyle. a modest consumption of $50,000 per annum is equivalent tow/c 5 20, and as we shall see, this supposedly modest income for a millionaire is not all that secure. we emphasize that the values are in real dollars, and that both income and principal are consumed—no bequest is assumed. the value ofw at any future time is random, with the randomness coming from the random return function. if inflation is positive, the actual amount drawn out in nominal dollars increases every year. if you do consume approximately c, then on average the future values ofw/c will decline for most situations. that is, most people will be obliged to consume part of the principal every year, and therefore as their wealth declines, so doesw/c. as a practical matter, we will work with only three assets at a time in this paper, one risk-free and two risky. presentation of allocations among more than three assets on a two-dimensional piece of paper is very confusing and consumes far more space without adding further understanding of the questions we consider in this paper. an earlier version of this paper used the simulation method and two assets, but we reached the same conclusions. 4. international equity diversification and shortfall risk in different countries we want to answer three related questions: 1. does international equity diversification reduce shortfall risk for canadians and is it material? 2. does international equity diversification reduce shortfall risk for americans and is the reduction material? 3. is international diversification relatively less effective in reducing shortfall risk for americans than it is for canadians? the answer to the first two questions is very likely positive to some extent, since diversification reduces total risk as long as the assets are imperfectly correlated. if the mean returns are too low on one asset, shortfall risk may not change even when it is introduced into the opportunity set. 4.1. the data for canada, in canadian $: ● t-bills: annualized rate of return of 91-day canadian treasury bills 14 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 ● bonds: scotia mcleod long bond index, total rate of return ● tse: toronto stock exchange 300 index, total rate of return ● s&p: standard and poor’s 500 index, total rate of return ● world: morgan stanley capital international (msci) global portfolio total rate of return. ● eafe: msci’s index portfolio for europe, australia, far east, total rate of return. for the united states, in united states $: ● t-bills: annualized rate of return of three month united states treasury bills ● bonds: 30 year federal bonds, total rate of return ● s&p: standard and poor’s 500 index, total rate of return ● world: morgan stanley capital international (msci) global portfolio total rate of return. ● eafe: msci’s index portfolio for europe, australia, far east, total rate of return. we use annual rates of return and convert them to real rates using the consumer price index of the respective countries. we show the continuously-compounded mean real rates of return, standard deviations and correlations for the period 1957–1997 in table 1. although the standard deviation is non-zero for t-bills, we will still use it as the risk-free rate in our table 1 continuously compounded annual real rates of returna canadian american mean standard deviation mean standard deviation t-bills 0.026 0.030 0.013 0.022 bonds 0.039 0.105 0.019 0.112 tse 0.051 0.156 s & p 0.076 0.163 0.062 0.161 world 0.070 0.170 0.059 0.169 eafe 0.071 0.214 0.065 0.213 correlation matrices canada bonds tse s & p world tse 0.27 s & p 0.55 0.79 world 0.50 0.80 0.84 eafe 0.32 0.69 0.60 0.92 united states bonds s & p world s & p 0.37 world 0.47 0.87 eafe 0.34 0.64 0.93 a the returns and correlation coefficients are 1970–1997 for world and eafe indexes; 1957–1997 for the others. scotia mcleod, inc. and morgan stanley capital international, inc. provided the raw returns from which the distributions are calculated. 15k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 analysis, since it is the closest proxy we have. the international indexes offer much higher returns than canadian equity, and somewhat higher volatility. the distribution of the date of death is drawn from the 1996 individual annuity mortality —basic (iam) tables, which insurance companies in both the united states and canada use when pricing individual life annuities. we have smoothed it using a gompertz distribution, as developed in milevsky and robinson (1999). we use separate tables for males and females. mortality tables exist for many finer distinctions like smoker vs. non-smoker and different provinces. if we used different tables we would get different estimates of the shortfall probabilities, but the conclusions of the paper would be unchanged, unless we were using a table for some group with very short life expectancies, relative to the general population. the iam tables assume adverse selection and hence the mortality function assigns higher probabilities to longer life than do standard tables for the entire population in either canada or the united states. 4.2. empirical results using the shortfall model and the data already described, we estimated the shortfall probabilities for a variety of the parameters:w/c, age and gender. the asset allocations were in five percentage point units. thus, for one person we would calculate shortfall for all the possible asset allocations of three different assets adding to 100%, with each asset having in turn allocations by five percentage increments from 0 to 100%. this procedure generates an enormous amount of data very quickly. the results are quite consistent in the evidence they provide on the three questions posed earlier. we present the results in 20% increments for each asset and for each country for a single situation: a female aged 65 withw/c5 16. then, to summarize a wider range of situations, we present the minimum shortfall risk probability (msp, hence) and the allocation that yields that probability, for a variety of parameters. leaving aside the question of which equity portfolio to choose and which country the investor lives in, we first summarize some results concerning the choice among t-bills, bonds and equity (whether domestic or foreign). these results are similar to those in milevsky et al. (1997), which looked at canadian equity, t-bills and bonds. the risk premium of equity over t-bills was somewhat higher in the 1997 paper, and yet the general results are comparable, showing that the conclusions are fairly robust to changes in the estimates of the return distribution. ● women have higher shortfall probabilities than men, all else constant. ● the allocation to equity that minimizes shortfall decreases as age andw/c rise. men and women still require at least some equity in their portfolios unless they are very old and/or have a highw/c ratio. ● a 100% allocation to t-bills or bonds has a higher shortfall risk than a 100% allocation to equity for all but the oldest people or higher values ofw/c. ● changes that move a portfolio towards the msp reduce shortfall risk at a declining rate. if a portfolio has proportions that are close to the msp, then the difference in shortfall probabilities between it and the msp portfolio is quite small. for example, if your minimum risk occurs at 50% equity, and your current portfolio is 100% t-bills, every 16 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 5% you switch into equity up to 50% will decrease your risk, but at a decreasing rate. the change from 100% t-bills to 95% t-bills and 5% equity will reduce shortfall risk by a greater amount than will a change from 55% t-bills, 45% equity, to a 50%–50% portfolio. table 2 summarizes some results for a female aged 65, withw/c 5 16. in panel a she is a canadian choosing among canadian t-bills, bonds and equity (tse 300). the column headings are the percentage of the portfolio allocated to equity; the row headings are the allocation to bonds. the allocation to t-bills is (100%–equity%–bonds%). each entry in the table is the probability of shortfall for that particular asset allocation. the msp and the allocation that produces it, to the nearest 5% increment, are in the lower right hand. for example, a portfolio invested 40% in the tse 300, 40% in bonds and 20% in t-bills has a shortfall probability of 36.5%. panel b replaces bonds with the world equity index. all the shortfall probabilities for allocations that include world equity are lower. the msp is reduced from 36.1% (50% tse, 45% bonds, 5% t-bills) to 27.8% (75% world, 25% t-bills). this is a material change in table 2 asset allocation and shortfall risk for a canadian woman, aged 65, withw/c 5 16a tse 0% 20% 40% 60% 80% 100% a: bondsb 0% 0.488 0.428 0.393 0.383 0.388 0.405 20 0.455 0.401 0.374 0.369 0.378 40 0.433 0.386 0.365 0.363 60 0.421 0.381 0.364 80 0.419 0.384 100 0.423 b: world equityc 0% 0.488 0.428 0.393 0.383 0.388 0.405 20 0.378 0.346 0.339 0.348 0.368 40 0.311 0.306 0.316 0.337 60 0.283 0.293 0.314 80 0.278 0.297 100 0.288 c: s & p 500d 0% 0.488 0.428 0.393 0.383 0.388 0.405 20 0.361 0.330 0.324 0.334 0.355 40 0.280 0.276 0.289 0.312 60 0.241 0.254 0.277 80 0.229 0.250 100 0.232 a the column headings at the top of each panel show the allocation to the tse 300. the row headings are the allocations to bonds, world equity index and s & p 500, respectively, in panels a, b, and c. the allocation to t-bills is 100%2 column heading%2 row heading%. the entries are the shortfall probabilities for the given allocations. b minimum shortfall probability5 36.1% tse 50%; bonds 45%; t-bills 5%. c minimum shortfall probability5 27.8% tse 0%; world 75%; t-bills 25%. d minimum shortfall probability5 22.9% tse 0%; s & p 500 85%; t-bills 15%. 17k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 both asset allocation and shortfall probability. panel c provides an even greater shortfall risk reduction when the s & p 500 issubstituted for canadian bonds and the msp falls to 22.9%. in both international substitutions, the tse is dominated for a 65 year old woman with w/c 5 16. we have not shown the result if the choice were among eafe, tse and t-bills, but it lies between the others. the msp is 32.9% for an allocation of tse 15%; eafe 45%; t-bills 40%. the reader may suspect that the results are partly because we dropped the long bonds out of the picture. in fact, the msp allocation is identical when the s & p 500 is theinternational index and either long bonds or the tse is included. that is, the msp is 22.9% with an allocation of: 0% bondsor tse; 85% s & p 500; 15% t-bills. we are not aware of any metric for comparing the relative benefits of different portfolios in reducing shortfall risk; so we rely upon common sense in determining what is a material improvement. our judgement is that international equity diversification provides a material benefit to canadian investors. the shortfall probabilities would be reduced somewhat more for at least some situations (i.e. for different ages, gender and values ofw/c) if we included more than three assets. the expansion would be quite difficult to present and would not change the answer to the questions we are asking. table 3 displays the same three panels as table 2, but now calculated for an american woman, aged 65, withw/c 5 16. panel a is the domestic portfolio choice among t-bills, long-term bonds and the s & p500. panel b replaces the bonds with the world portfolio and panel c replaces the bonds with the eafe portfolio. the allocations in panel a that include a substantial amount of bonds and/or t-bills have higher shortfall probabilities than the comparable canadian entries. the real return to bonds and t-bills in the united states has been significantly lower than it has in canada, and the standard deviation on the bonds is higher in the united states than in canada. the s & p 500 has done much better than the tse 300, however, whether you measure it in canadian $ or united states $, and so the risk premium in the united states is much higher than in canada. the msp for the domestic allocation is slightly lower than in canada at 32.7%, and it requires 100% invested in the s & p 500. when we turn to panels b and c, we find that international diversification has an effect, but it is material only when the person holds a portfolio very distant from the minimum shortfall risk portfolio. for example, in panel b, a woman investor who holds 40% s & p, 40% bonds and 20% t-bills faces a shortfall probability of 44.5%. by replacing the bonds with the world equity index, she reduces the shortfall probability to 34.3%. however, the msp is 32.6% with an allocation of: s & p 90, world 10%; t-bills 0%. this portfolio is scarcely distinguishable from the domestic case where the msp is 32.7% with 100% invested in the s & p 500. eafe offers higher return and standard deviation than the s & p 500 or the world portfolio, and is less-correlated with the s & p 500 than is the world portfolio, because it does not include american equity, as does the world index. it provides only a small improvement in the msp to 31.9% for an allocation of s & p 500 80%, eafe 20%. if we take minimizing shortfall risk as the objective, then international diversification appears to have minimal benefits for an american woman at age 65, with moderate wealth 18 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 relative to desired consumption, but substantial benefits for an identical canadian woman. in table 4 we focus on msp portfolios for a range of age andw/cand both males and females to show that this observation holds more generally. the entries in table 4 display the msp allocations with the msp in parentheses underneath each allocation. we provide results for males and females aged 65 and 75, withw/c ratios of 12, 16, 20 and 24. the investors can choose between two sets of assets in each country. in canada, the choices are either (1) domestic equity, bonds, bills, or, (2) domestic equity, s & p 500, bills. in the us, the choices are either (1) domestic equity, bonds, bills, or, (2) domestic equity, eafe index, bills. in each case we chose the non-domestic portfolio that gives the best diversification results, but other choices would not materially affect our conclusions. table 4 shows the same pattern as tables 2 and 3 showed. the canadian investor changes his or her portfolio allocation a great deal and reduces shortfall risk significantly when the s & p 500 replaces bonds in the available assets. the american investor makes almost no change to the allocation and the shortfall risk falls very little, when the eafe index replaces bonds. this is despite the fact that the canadian bond index has outperformed the american table 3 asset allocation and shortfall risk for an american woman, aged 65, withw/c 5 16a s & p 0% 20% 40% 60% 80% 100% a: bondsb 0% 0.661 0.540 0.438 0.373 0.339 0.327 20 0.648 0.531 0.437 0.379 0.349 40 0.638 0.529 0.445 0.392 60 0.633 0.534 0.458 80 0.632 0.544 100 0.637 b: worldc 0% 0.661 0.540 0.438 0.373 0.339 0.327 20 0.548 0.445 0.377 0.341 0.327 40 0.456 0.386 0.348 0.331 60 0.400 0.359 0.340 80 0.374 0.351 100 0.367 c: eafed 0% 0.661 0.540 0.438 0.373 0.339 0.327 20 0.537 0.435 0.368 0.333 0.319 40 0.450 0.382 0.343 0.325 60 0.411 0.368 0.346 80 0.404 0.378 100 0.419 a the column headings at the top of each panel show the allocation to the s & p500. the row headings are the allocations to bonds, world equity index and eafe index, respectively, in panels a, b, and c. the allocation to t-bills is 100%2 column heading%2 row heading%. the entries are the shortfall probabilities for the given allocations. b minimum shortfall probability5 32.7% s & p 100%; bonds 0%; t-bills 0%. c minimum shortfall probability5 32.6% s & p 90%; world 10%; t-bills 0%. d minimum shortfall probability5 31.9% s & p 80%; eafe 20%; t-bills 0%. 19k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 table 4 asset allocations yielding minimum shortfall probability (msp)a canada united states tse-bonds-bills in % (msp) tse-s & p-bills (msp) s & p-bonds-bills (msp) s & p-eafe-bills (msp) age 65 w/c 5 12 male 60-40-0 0-100-0 100-0-0 75-25-0 (0.48) (0.34) (0.43) (0.43) female 85-15-0 0-100-0 100-0-0 75-25-0 (0.64) (0.45) (0.57) (0.56) w/c 5 16 male 40-40-20 0-75-25 90-0-10 75-20-5 (0.24) (0.16) (0.23) (0.22) female 50-50-0 0-85-15 100-0-0 80-20-0 (0.36) (0.23) (0.33) (0.32) w/c 5 20 male 30-25-45 0-60-40 70-0-30 55-20-25 (0.11) (0.07) (0.12) (0.12) female 30-30-40 0-60-40 80-0-20 65-20-15 (0.17) (0.10) (0.18) (0.18) w/c 5 24 male 25-20-55 0-55-45 60-0-40 50-10-40 (0.05) (0.03) (0.06) (0.06) famale 25-25-50 0-55-45 60-0-40 50-15-35 (0.07) (0.04) (0.09) (0.09) age 75 w/c 5 12 male 40-35-25 0-80-20 90-0-10 75-15-10 (0.22) (0.16) (0.21) (0.21) female 50-50-0 0-90-10 100-0-0 80-20-0 (0.35) (0.25) (0.33) (0.32) w/c 5 16 male 30-30-40 0-65-35 75-0-25 60-15-25 (0.09) (0.06) (0.09) (0.09) female 30-30-40 0-70-30 80-0-20 65-15-20 (0.15) (0.10) (0.15) (0.15) w/c 5 20 male 25-25-50 0-55-45 60-0-40 50-15-35 (0.04) (0.03) (0.04) (0.04) female 25-25-50 0-55-45 60-0-40 50-15-35 (0.06) (0.04) (0.07) (0.07) w/c 5 24 male 25-25-50 0-50-50 60-0-40 50-10-40 (0.02) (0.01) (0.02) (0.02) female 25-25-50 0-50-50 55-0-45 45-15-40 (0.03) (0.02) (0.04) (0.03) a the four columns display the asset allocations that yield msps for four different choice sets for an individual investor. the individual entries show the allocations to the asset classes in the column heading, in percent, with the shortfall probability in parentheses underneath it. the first two columns are canada: domestic equity, bonds, and bills; domestic equity, s & p 500, and bills. the second two columns are the united states with the same domestic choice and with the eafe index replacing bonds for the last column. the rows are different investors. the first set are aged 65, male and female, with differentw/c ratios. the second set are aged 75. 20 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 bond index (in terms of domestic currency) very substantially, and therefore the international substitution for an american investor is replacing a very low return asset in the choice set. 5. discussion these results should not surprise us. the united states equity market is a substantial part of the international equity market, while canada is only about 3%. furthermore, the performance of american companies in the index is correlated with economic performance globally, because many of the largest american companies that make up the index have subsidiaries in other countries. therefore, the diversification that the international equity can provide to americans is quite limited. canada has more of its economy based on international trade than the united states, because the canadian economy is based on commodities. however, the canadian companies in the tse 300 have fewer operations outside canada, and hence their returns are less-correlated with the equity returns elsewhere. furthermore, the tse 300 has a heavier weighting in natural resources (forest products, mining, oil and gas) than the s & p 500 or any other significant national or international index. one odd, and unexpected, part of the results is the role of the debt securities. united states debt securities have much lower realized rates of return than canadian. international debt rating agencies give american government slightly higher debt ratings than canadian government debt, which seems reasonable. however, if debt ratings in the very long run are to accurately predict returns, there must be some difference in the realized default rate, and this has not been the case. as a result, the canadian domestic investor has received much higher yields on domestic debt, but can also invest in united states or international equity to benefit from diversification. it seems that perhaps the research should focus on international debt diversification, but that is also problematic. the united states dollar has been generally rising against the canadian dollar for a long time, and hence an american investor would have received much lower returns from canadian debt. we seem to have created a paradox in these results, therefore, since it seems that canadians can do better than americans, even though all could access the same choice set. the paradox is resolved if purchasing power parity holds. in that case, a canadian retiree would require a higherw/c ratio than an american retiree in order to enjoy the same real standard of living. whether this is true we leave to the economists to debate. our use of shortfall risk as a criterion produces results that are similar to utility maximization with increasing relative and absolute risk aversion. to see this, consider holding constant the age, gender and desired consumption,c. the only thing that varies isw, which will thus causew/c to vary. increasing wealth in this situation leads to lower allocations to equity, as the reader can see by reading down the columns in table 4. thus, a very wealthy person, relative to his or her desired consumption, should invest primarily in t-bills, if the criterion is to minimize shortfall risk. in finance theory, it is generally assumed that individuals have decreasing or constant relative risk aversion. in practice, our result seems counter-intuitive to what we generally believe wealthy individuals do. if the wealthy person also has a bequest motive, then increasing equity investment is preferred, as milevsky et al. (1997) show. 21k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 thus, our shortfall model may not be a good normative model for persons who are very rich relative to their level of consumption. the very rich do not have to worry about survival. for the great majority of people, for whom shortfall below the desired level of consumption during retirement is a realistic possibility, the shortfall criterion appears to provide reasonable results for asset allocation. another way to look at the results is to focus on what income might be sustainable from a given wealth, with reasonable assurance. as long as the individual’s wealth is strictly less than the risk-free discounted value of perpetual consumption, there is always a non-zero probability of ruin. returning to table 4, the entries forw/c 5 20 or 24 show shortfall probabilities of 2–7%, which we might consider as fairly secure.the asset allocations in this situation are balanced, on the order of 50% debt securities, 50% equity; although the specific form of debt or equity differs. 6. limitations of the study and directions for future research the biggest question mark is the validity of the return distributions we use. these are historic, real rates, and they are reasonably accurate representations of what investors would have realized over long time periods in the past. jorion and goetzmann (1999) present results of an investigation into long run rates of return and stability of those rates for many countries, including canada and the united states. while they characterize the united states as practically the only country with a very long-run stable series of equity returns, canada certainly has a substantial history as well, longer than most other countries. nonetheless, these returns histories are taken over time periods in which the underlying economies were changing. we cannot have much confidence that the past observations are drawings from the returns distribution that applies now and in the future, even in canada and the united states. we have no alternative, unfortunately. a forecast of long-run rates of return and the covariance matrix of them is no more valid, since it is ultimately conditioned on what evidence we have from the past. this problem affects a great deal of the research in investments and personal finance, and we simply have to be aware that the shortfall probabilities we calculate are dependent on the validity of the returns distributions that we assume. gibson (1996) suggests an heuristic method, which is to use the historical data to estimate the variance-covariance matrix, and then use a forecast of expected returns. we know of no forecast that is necessarily more accurate or reliable than using historical data for all the parameters. our model uses very long run returns, and we have never seen forecasts for such long periods. we are dealing with periods that could be as much as 40 years (a retiree aged 65 living to the age of 105). historical data has the advantage that it is at least objective. in addition, any return forecast must be mathematically consistent with the variance-covariance matrix, or the results from any analysis will be invalid. there are other limitations that could perhaps be overcome: ● all the returns are before transactions costs. transactions costs would reduce the gross returns of every portfolio, but perhaps not by equal percentages. 22 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 ● all returns are before income tax. taxes have no effect inside sheltered retirement portfolios (ira or 401k in the united states; rrsp and rrif in canada). for investment outside a shelter in canada, interest income is taxed at a higher rate than international equity, which in turn is taxed at a higher rate than canadian equity. this would favour domestic equity even more strongly, but might reduce or even eliminate the benefit of international diversification. while we do not believe that the omission of transactions costs and taxes would change the conclusions of the paper, a formal modelling of these two factors is the obvious direction for future research on this question. 7. conclusion the evidence is very strong that international equity diversification reduces shortfall risk for canadians significantly. it does not appear to benefit americans materially, because their equity portfolio is already closely-related to the international equity portfolio. finally, we note that our results apply to investors atomistically in canada. if every canadian individual investor read this paper and immediately shifted his or her portfolios into international equity, we have no idea how much effect that might have on canadian markets, if any. acknowledgments the authors thank narine kaltakjian for research assistance, william reichenstein, steve krull and david stangeland and participants at the 1997 asac and 1998 afs conferences for helpful comments, and scotia mcleod inc. and morgan stanley capital international inc. for providing the raw returns from which the return distributions are calculated. appendix: derivation of the probability of shortfall the probability of shortfall (ruin) is derived in milevsky and robinson (1999). however, for the sake of completeness, we reproduce here the formula and refer the interested reader to the above-referenced paper. (also available for download at the website: www.yorku.ca/ academics/milevsky) the probability is computed in three stages. first, we must compute the present value of a life annuity, denoted bya(x), where the discounting is done at the rate ofx. this calculation must be conducted using the appropriate insurance mortality table. we then define the two new variables: m1 5 a~m 2 s2!, m2 5 a~m 2 s2! 2 a~2m 2 3s2! m 2 2 s2 (1) 23k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 wherem, s are the portfolio mean and standard deviation respectively. the technical term for these quantities is the first and second moment of the stochastic present value. the next step is to compute: a 5 2m2 2 m1 2 m2 2 m1 2 , b 5 m2 2 m1 2 m2m1 (2) finally, the probability of shortfall is: p~ruin! < gs c w u a,bd (3) where c/w is the consumption to wealth ratio, and theg(yua,b), denotes the cumulative distribution function (cdf) of the gamma random variable. this function is readily available in all commercial spreadsheets such as excel or lotus and can be easily implemented in practice. in fact, the g(.) term is quite similar to the n(.) function, which should be well known to the users of the black-scholes equation. where n(.) represents the area-underthe-curve for the standard normal distribution, g(.) represents the same concept for the gamma distribution. in our case, the gamma density is parameterized by two variables (similar to a mean and variance), which we denote by the greek letters alpha and beta. the gamma distribution can take on only positive values and can be implemented also in terms of the chi-square distribution. once again, from a slightly more technical perspective, the inverse of the stochastic present value (spv) of lifetime consumption is taken to be gamma-distributed. therefore, the probability of ruin is the probability that the spv is greater than the initial wealth that is available to support the consumption. it is important to note that the above formula is only anapproximation to the true probability of shortfall. the exact expression can not be obtained analytically—to the best of the authors’ knowledge. however, the approximation is quite accurate when benchmarked against monte carlo simulations. furthermore, one can actually demonstrate that as the time horizon of the investor increases, the expression actuallyconvergesto the true probability of ruin. we refer the interested reader to milevsky and robinson (1999). references adler, m. & dumas, b. (1983). international portfolio choice and corporation finance: a synthesis.j finance, 43, 925–984. bailey, w. & stulz, r. (1990). benefits of international diversification: the case of the pacific basin stock markets.j portfolio management 16, 57–61. cosset, j. & suret, j. (1995). political risk and the benefits of international portfolio diversification.j int business stud, 26(2), 301–318. errunza, v. (1977). gains from portfolio diversification into less developed countries’ securities.j int business stud,fall/winter, 83–99. gibson, r. (1996).asset allocation: balancing financial risk(2nd ed). new york: mcgraw-hill. goldberg, s., & heflin, f. (1995). the association between the level of international diversification and risk.j int financial management accounting 6, 1–25. 24 k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 grauer, r., & hakansson, n. (1987). gains from international diversification: 1968–85 returns on portfolios of stocks and bonds.j finance 47,721–741. grubel, h. (1968). internationally-diversified portfolios: welfare gains and capital flows.am econ rev 58, 1299–1314. grubel, h., & fadner, k. (1971). the interdependence of international equity markets.j finance 31, 89–94. haavisto, t., & hansson, b. (1992). risk reduction by diversification in the nordic stock markets.scand j econ 94(4), 581–588. hunter, j., & coggin, t. (1990). an analysis of the diversification benefit from international equity investment. j portfolio management 16, 33–37. jorion, p. (1985). international portfolio diversification with estimation risk.j business 58, 259–278. jorion, p. (1989). asset allocation with hedged and unhedged foreign stocks and bonds.j portfolio management 15, 49–54. jorion, p., & william n. goetzmann. (1999). global stock markets in the twentieth century.j finance, 54(3), 953–979. levy, h., & lerman, z. (1988). the benefits of international diversification in bonds.finan analysts j 44, 56–64. levy, h., & sarnat, m. (1970). international diversification of investment portfolios.am econ rev 60, 668–675. liebowitz, m., & kogelman, s. (1991). asset allocation under shortfall constraints.j portfolio management 17, 18–23. mcdonald, j. (1973). french mutual fund performance: evaluation of internationally-diversified portfolios.j finance 33, 1161–1180. milevsky, m., ho, k., & robinson, c. (1997). asset allocation via the conditional first exit time.or how to avoid outliving your moneyrev quantitative finance accounting 9, 53–70. milevsky, m., & robinson, c. (1999). self annuitization and ruin in retirement.working paper, schulich school of business, york university. (also available for download at the website: www.yorku.ca/academics/ milevsky) obstfeld, m. (1994). risk taking, global diversification, and growth.am econ rev 84, 1310–1329. odier, p., & solnik, b. (1993). lessons for international asset allocation.finan analysts j 49, 63–77. sinquefield, r. (1996). where are the gains from international diversification?finan analysts j 52,8–14. solnik, b. (1979). why not diversify internationally rather than domestically?finan analysts jl 35,48–54. tse, m., uppal, j., & white, m. a. (1993). downside risk and investment choice.finan rev 28, 585–606. 25k. ho, m.a. milevsky / financial services review 8 (1999) 11–25 pii: 1057-0810(94)90018-3 financial services review, 3(2): 127-141 copyright 0 1994 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the impact of mutual fund distributions on after-tax returns william lewis randolph this paper analyzes the impact of mutual fund distributions on after-tax returns. mutual fund objective and management style are the two most importantfactors which determine the proportion of thefund’s total return that ispaid out in distn’butions. the larger thefund’s annual distributions, the greater the amount of return lost to taxes. the correlation between portfolio turnover and after-tax return is examined and found to be low. a measure of effective portfolio turnover is developed to show the real effect of turnover on distributions. within some mutualfund categories, the investor can increase after-tax returns by one or two percent by selecting funds with low distributions. i. introduction investors who save for retirement or estate purposes have a love-hate relationship with mutual fund distributions. large distributions usually mean the fund is doing well, but if it is a regular mutual fund account, the investor has to pay taxes on these distributions. taxes reduce the rate at which savings grow. while most investors are aware of the adverse effect that fees and expenses have on the performance of their investments, the impact of taxes on investment return is just as critical. an investor must achieve a certain after-tax return in order to meet investment goals. unfortunately, available information on mutual fund performance does not consider the impact of taxes on return. the aim of this paper is to analyze the impact of mutual fund distributions on after-tax returns. total return from a mutual fund investment is composed of dividend or interest payments and capital gains. the dividend or interest component of return will be referred to as yield. the capital gain component of total return is further divided into realized and unrealized capital gains. capital gains are realized when a fund sells a security or investment that has increased in value. in order to maintain a tax exempt status, mutual funds must distribute income from dividends and interest and realized capital gains at least annually. in a regular mutual fund account, the investor pays taxes on these distributions at the applicable tax rate. the larger the taxable distributions, the greater the loss of potential return on the savings plan. this loss of potential return because of taxes is referred to as “tax drag.” william lewis randolph l department of finance, school of business, norfolk state university, norfolk, va 23504. 128 financial services review, 3(2) 1994 unrealized gains made by a mutual fund show up as increases in the net asset value (nav) of the fund and are not taxed until the shares are sold. all other things being equal, the taxable investor who saves for retirement or estate purposes is better off if the return received is in the form of unrealized gains. unrealized gains compound tax-deferred and narrow the difference between before-tax and after-tax returns. a mutual fund’s objective and management style are the most important factors which determine the proportion of the fund’s total return that is paid out in distributions. in analyzing distributions and after-tax returns by mutual fund objective, fund categories of aggressive growth, growth, growth and income, balanced, and bond are examined. funds that invest in stocks will have distributions composed of dividends and realized capital gains. since stocks realize most of their total return from capital gains, capital gains distributions will be very important in determining after-tax returns in stock funds. these capital gains distributions are usually much less significant in bond funds, where total return and distributions are both directly related to bond yield. the nav of the bond fund will vary with changes in interest rates, but over the long run the return is based primarily on interest income. mutual fund management style is examined based on the size of the capital gains distribution a fund management generates annually relative to other funds with the same objective. the distribution of capital gains is related somewhat to reported portfolio turnover, but more directly to the number of winners the fund management sells during the year.* jeffrey and arnott (1993) address this issue by focusing on reported portfolio turnover and its potential for creating a tax liability which will drastically reduce the performance of a fund. although reported portfolio turnover information is readily available, reported turn over does not correlate well with actual mutual fund after-tax performance. turnover mutual fund turnover can be the result of management strategy or may be caused by flows of investor money into and out of the fund. the taxable investor is concerned with the sale of fund assets which cause capital gains to be realized and distributed to the shareholders. this type of turnover creates a tax liability and diminishes the potential growth of the investment. most turnover is the result of management strategy. for example, a sector rotation strategy will result in high turnover as the fund management moves from one sector to another in anticipation of economic events. realized capital gains, and the associated tax liability which reduces after-tax return are not necessarily bad for investors. however, investors should expect that active portfolio managers will do at least as well as passive managers on an after-tax basis. turnover is not uniform across a fund’s holdings and does not necessarily cause capital gains to be realized. for example, a fund can experience a very high portfolio turnover by frequently trading poorly performing stocks. this type of turnover does not necessitate capital gains distributions, because the stocks that are traded have not appreciated in value. on the other hand, low portfolio turnover that involves selling winners can create very large capital gains distributions. turnover can also be caused by net shareholder redemptions. when savings are flowing out of a mutual fund, the fully invested fund must sell assets to redeem shares. if the fund manager chooses to sell assets that have experienced capital gains, net shareholder redemptions can cause capital gains to be realized. it is possible that recent investors in a after-tax mutual fund returns 129 fund may have to pay taxes on capital gains distributions for realized gains in which they did not participate because the gains occurred before the investor invested in the fund. if a fund has a strategy that calls for very little turnover, like an index fund, then a significant increase in turnover can be caused by withdrawals from the fund. when management must sell a uniform proportion of all stocks in the fund to maintain the index weighting, this can cause an index fund to realize capital gains that the manager would not ordinarily realize. other than these forced selloffs, index funds can be expected to have low portfolio turnover and, therefore, low capital gains distributions. in this paper the concept of effective portfolio turnover is developed and compared to reported portfolio turnover. effective portfolio turnover is a derived figure based on the percentage of unrealized capital gains which are realized and distributed on an annual basis. effective turnover is a surrogate for uniform turnover. for example, if a new growth fund starts the year with a nav of $10.00, one year later has a capital gain of $1.00 per share, and pays out a capital gains distribution of $0.25 per share then the effective turnover would be 25 percent; the realized capital gain divided by the total capital gain. the fund’s nav would end the year at $10.75. if this fund was fully invested at the beginning of the year, took in no additional funds, and sold 25 percent of each holding at the end of the year and reinvested (uniform turnover), the reported turnover and effective turnover would both be 25 percent. uniform turnover does not occur in actual portfolio management. suppose the fund management reports turnover of 150 percent in the above example. then the high reported turnover is primarily the result of trading losers. or if the fund reported a turnover of 15 percent, then the management is selling more winners than losers in its turnover. in either case the reported turnover is not uniform across the portfolio and is different from the effective turnover. effective turnover is superior to reported turnover as an indicator of fund management behavior that will result in the distribution of capital gains. below, effective turnover will be derived for a sample of mutual funds and compared with the reported turnover of the funds. investment tax strategy to maximize long-term wealth accumulation, the investor should maximize contribu tions to qualified plans which allow income to be invested on a before-tax basis. examples include tax-deductible iras, 401(k), 403(b), sep, and keogh plans. after these programs have been fully used, it is assumed that the investor will invest after-tax dollars into taxable mutual fund accounts or into non-qualified plans that allow distributions to accumulate tax-deferred. the investor must pay taxes on all distributions made by regular mutual funds. the variable annuity plan (vap) and non-deductible ira are two non-qualified tax-deferred accounts available to investors. in these plans distributions accumulate tax-deferred. no taxes are paid until savings are withdrawn. the non-deductible ira is generally superior to the vap because of lower expenses, and it should be used first. but since the non-deductible ira is limited to $2000 per year, many investors will need to invest funds beyond this amount and will use either a vap, regular mutual fund, or a combination of these vehicles. the vap is offered by insurance companies and has an insurance feature that increases the cost of the plan. the insurance feature guarantees that the investor’s beneficiary receives at least an amount equal to the net investment in the plan. when participating in a vap, the 130 financial services review, 3(z) 1994 investor can select from several investment alternatives, usually ranging from bond funds to growth funds. the non-deductible ira has an even greater selection of investment objectives to choose from and minimal or no extra expenses. both non-qualified plans penalize the investor 10 percent if funds are withdrawn prior to age 59 ih. this could limit their attractiveness for some investors. while the age withdrawal restrictions and the insurance feature of the vap makes the non-qualified plans not directly comparable to a regular mutual fund account, it is instructive to compare the after-tax return of these alternative savings plans. the extra expense of variable annuity plans varies widely. vanguard has a plan that adds costs of 0.55 percent per year to fund operating expenses. according to morningstar, total expenses for the average vap is 2.05 percent.2 this is twice the total expense of the average vanguard annuity plan. aside from the extra expense, vaps can have some unpleasant features. some states tax initial and subsequent investments in vaps. some states tax payments from the plan if the owner chooses a lump sum distribution. the taxable portion of distributions from vaps is taxed at ordinary income tax rates which could be a disadvan tage for investors in the top tax bracket. in the following analysis, non-qualified plans and regular mutual funds are compared as long-term savings vehicles. the effect of turnover on after-tax return is analyzed for the regular mutual fund account. in order to make comparisons, an assumed marginal tax rate of 35 percent is used. to simplify the analysis, no distinction is made between income and capital gains tax rates.3 for the regular taxable mutual fund account, taxes on fund distribu tions are assumed to be paid in the year of the distribution. for the non-qualified tax-deferred plans, distributions accumulate tax-deferred. effective turnover and after-tax return in the following examples, dividends and interest income and realized capital gains are assumed to be distributed at year end by the mutual fund. the investor then pays taxes on these distributions at the assumed 35 percent rate, and reinvests the remainder. two different after-tax measures are used to assess investment performance: the after-tax return and the rate of wealth accumulation. to compute the after-tax return (atr), taxes are paid on each annual distribution, the after-tax proceeds of the distribution are reinvested, and then taxes are paid on the unrealized capital gain at the end of a selected holding period. the atr can be used in planning a retirement savings program when the investor is interested in a return net of all taxes. to compute the rate of wealth accumulation (rwa), taxes are paid on each annual distribution, the proceeds are reinvested, but the investment is not sold. an investor accumulating wealth for estate purposes would use this computation since there is no capital gains tax liability for the beneficiaries. the rwa is also useful for retirement planning since most investors plan in terms of accumulating a target amount of wealth which will generate income in their retirement years. if a fund’s annual effective turnover rate is 100 percent, there are no unrealized capital gains and these two rates will be equal and can be computed by multiplying total return by one minus the tax rate. since reinvestment of the after-tax proceeds of the distributions changes the basis for the investment, all other solutions for atr or rwa must be computed taking each year of the investment process into consideration. based on the high-return/low-yield and low-return/high-yield relationship that is normally found in the market, the following returns were constructed for the major mutual fund investment objectives: ajler-tax mutual fund returns 131 fund objective total return yield component capital gains component aggressive growth growth growth sk income balanced taxable bond municipal bond 12% 1% 11% 11% 2% 9% 10% 4% 6% 9% 5% 4% 7% 7% 0% 5% 5% 0% these hypothetical returns and yields are meant to approximate long-term market conditions. table 1 presents the computations of atr and rwa for the above investment objectives in a taxable mutual fund and for the two tax-deferred plans. the two tax-deferred plans will be discussed in the next section. for the taxable account, the after-tax returns are computed at various rates of effective turnover. under the aggressive growth objective at 100 percent turnover the atr and rwa are both 7.80 percent. this is equal to the total return of 12 percent times one minus the tax rate [12%*(1 .35) = 7.80%]. as the rate of turnover decreases, the atr and rwa both increase, with rwa growing at a faster rate since the final tax bite is not taken. the four fund categories which invest in stocks behave in a similar manner, with the riskier funds having the higher returns at any turnover rate. the two bond objectives are not affected by turnover since capital gains are not considered to be part of their total return. the taxable and tax exempt bond funds have an assumed two percent difference in yield which causes the tax exempt bond fund to have the higher after-tax return at the assumed 35 percent tax rate. now analyze table 1 from the point of view of an investor selecting from almost identical growth funds. both funds have a total return of 11 percent and a dividend yield of ‘soox i 0.00% 20.00% 40.00% 60.00% bo.oox 100.00% turnover 0 tax-exempt + taxable bond 0 balanced a growth & income x growth v aggressive growth figure 1. rate of wealth accumulation 132 financial services review, 3(2) 1994 table 1. after-tax returns of selected investment objectives objective aggressive growth growth growth & income type of plan atr taxable with turnover of: 100% 7.80% 50% 7.95% 40% 8.03% 30% 8.16% 25% 8.27% 20% 8.41% 15% 8.61% 10% 8.88% 5% 9.25% 0% 9.72% tax-deferred: vap 8.99% ira 9.91% rwa 7.80% 8.01% 8.12% 8.32% 8.50% 8.76% 9.16% 9.76% 10.59% 11.65% 11.00% 12.00% atr rwa atr rwa 7.15% 7.15% 7.27% 7.32% 7.33% 7.41% 7.43% 7.57% 7.50% 7.70% 7.61% 7.90% 7.76% 8.21% 7.97% 8.69% 8.25% 9.38% 8.63% 10.30% 8.07% 10.00% 8.99% 11.00% objective 6.50% 6.50% 6.57% 6.61% 6.61% 6.67% 6.67% 6.77% 6.71% 6.85% 6.78% 6.97% 6.86% 7.16% 6.99% 7.46% 7.17% 7.92% 7.41% 8.60% 7.17% 9.00% 8.07% 10.00% balanced taxable bond tar-exempt bond type of plan atr taxable with turnover of: 100% 5.85% 50% 5.90% 40% 5.92% 30% 5.95% 25% 5.98% 20% 6.02% 15% 6.07% 10% 6.14% 5% 6.25% 0% 6.41% tax-deferred: vap 6.28% ira 7.17% rwa atr rwa atr rwa 5.85% 4.55% 4.55% 5.00% 5.00% 5.92% 4.55% 4.55% 5.00% 5.00% 5.96% 4.55% 4.55% 5.00% 5.00% 6.02% 4.55% 4.55% 5.00% 5.00% 6.07% 4.55% 4.55% 5.00% 5.00% 6.15% 4.55% 4.55% 5.00% 5.00% 6.27% 4.55% 4.55% 5.00% 5.00% 6.46% 4.55% 4.55% 5.00% 5.00% 6.77% 4.55% 4.55% 5.00% 5.00% 7.25% 4.55% 4.55% 5.00% 5.00% 8.00% 4.55% 6.00% 5.00% 5.00% 9.00% 5.40% 7.00% 5.00% 5.00% note: the 20 year after-tax return (atr) and rate of wealth accumulation (rwa) are presented for six investment objectives based on hypothetical total returns and yields. included are results using a regular taxable mutual fund account with various rates of effective turnover and two tax-deferred plans; the variable annuity plan (vap) and a non-deductible ira. two percent. the difference between the funds is in the management style as it affects turnover and realized capital gains. using rwa as the decision rule, a fund with an effective turnover of 10 percent will have a return of 8.69 percent, while a fund with a turnover of 50 percent will have a rwa of 7.32 percent. the point is, an investor’s after tax return can be increased by 1.37 percent by selecting a fund with low effective turnover. the relationship between after-tax return and effective turnover might surprise many investors. a small amount of turnover has a large effect on after-tax return. the non-linear relationship between turnover and after-tax return can readily be seen in figure 1 where the rwa data is presented. as turnover increases from zero, the returns rapidly decrease towards a.fker--tm mutual fund returns 133 the minimum for each investment objective. half of the extra after-tax return that can be obtained by compounding the unrealized capital gains is lost at turnover rates above 10 percent. at an effective turnover of 50 percent, almost all of this compounding effect is lost. for fund objectives where turnover affects return, we can see the dramatic effect the first 10 or 15 percent rate of effective turnover has on the realized return. non-qualified tax-deferred accounts included in table 1 are returns based on an investor pursuing one of the six investment objectives using a non-deductible ira or a vap. in order to compare the non-deductible ira and the vap to the regular mutual fund account, it is assumed that the investor can open an ira account at no extra expense and that the extra expense of the vap will cause a decrease of one percent in total return. since we are assuming no extra expense using the ira, the ira dominates the vap and regular mutual fund. the vap, due to the one percent loss in total return because of extra expenses, must be compared to the regular mutual fund on a case by case basis. the vap is clearly superior to the taxable account in accumulating wealth for estate purposes if the investor prefers either growth and income, balanced, or a bond fund objective. in contrast, if the investor uses an aggressive growth fund or a growth fund, then the vap is superior for rwa if the comparable taxable fund has an effective turnover greater than one or two percent. since any inves~ent program will have a few percent effective turnover, the vap will generally be superior to taxable funds for wealth accumulation over a 20 year period. the after-tax return (atr) is the appropriate yardstick to use when comparing the vap to a taxable account for the investor who will use the accumulated wealth for retirement. for the aggressive growth and growth objectives, the taxable account must have an effective turnover of less than eight percent in order to be superior to the vap. for the balanced fund and growth and income fund objectives the taxable account must have an effective turnover of less than five percent. as will be shown in the analysis of actual fund performance, very few funds have an effective turnover of less than five percent. the investment horizon and the amount of the added expense for the vap are the two important assumptions used in comparing the vap and taxable account. the vap returns are below those shown in table 1 when the extra expense is greater than one percent, and are above table i results when the extra expense is less than one percent. for example, under the bond fund inves~ent objective, the vap has the same atr as the taxable mutual fund at a one percent extra expense level. if the extra expenses were 0.6 percent, then the atr would increase to 4.89 percent, making the vap superior. these examples assume a twenty year investment horizon. for shorter horizons, the vap will be less competitive when compared to a taxable account with the same objective. growth stock example before examining actual after-tax mutual fund performance results, an example is presented to demonstrate the dynamics of tax drag and turnover. the example is an investment in a new growth stock mutual fund that will have a total return of 11 percent composed of a dividend yield of two percent and acapital gain of nine percent. this fund will have an initial nav of $10.00 and an effective turnover of 20 percent. the effective turnover of 20 percent means that the fund will turnover a uniform 20 percent of its holdings which will result in 20 percent of its cumulative unrealized capital gains being realized each year. 134 financial services review, 3(2) 1994 table 2. growth stock fund example investment results year 1 year 2 year 5 net asset value $10.72 $11.35 $12.79 (after distributions) dividends earned and distributed 0.20 0.21 0.25 capital gain earned 0.90 0.96 1.11 cumulative umealized capital gain 0.90 1.68 3.49 (before distribution) capital gain realized & distributed 0.18 0.34 0.70 (20% of above amount) taxes paid 0.13 0.20 0.38 (35% of combined distributions) shares owned 1.000 1.023 1.144 rate of wealth 9.67% 9.43% 8.91% accumulation (rwa) tax drag 1.33% 1.57% 2.09% (total return rwa) after-tax return (atr) 7.15% 7.23% 7.41% (if sold at the end of vear) year 20 $15.26 0.30 1.37 6.58 1.32 1.59 2.805 7.90% 3.10% 7.61% note: this table presents investment results for a hypothetical new growth stock mutual fund which has an 11 percent total return composed of a 2 percent dividend yield and a 9 percent capital gain rate. the fund has an initial net asset value (nav) of $10.00 and an effective turnover rate of 20 percent. the investor buys one sham, pays taxes on the distributions, and reinvests the balance of the distribution in new shares. the results from holding this investment for one, two, five, and 20 years are presented in table 2. distributions are reinvested after paying a 35 percent tax. the rwa decreases over time due to the relative increase in realized capital gains. this decrease in rwa results in the tax drag increasing as the holding period increases. after one year, the atr is equal to the total return times one minus the tax rate [ 1 1%*( 1 .35)]. over time the atr increases due to the compounding gain from deferring the tax liability. with moderate to high turnover and long holding periods, the significance of the unrealized capital gain and its tax liability is small compared to the overall value of the investment. this causes rwa and atr to converge over time. historical data to study the impact of turnover on actual investment results, five years of data on no-load and low-load mutual funds were examined. the data for this analysis was taken from the 1993 aa11 mutual funds publication.4 samples of funds in the following categories were scrutinized; aggressive growth, growth, growth and income, and balanced funds. funds that were included in the sample had asset size greater than $100 million, a fiscal year ending december 31, and five years of data. the data include before-tax performance results, distributions, net asset value (nav), and reported portfolio turnover during the five year period ending december 31, 1992. using this data, after-tax performance results were constructed. also, the amount of unrealized gain that had accrued at the end of the five year period was computed for each fund. after-tax mutual fund returns 135 table 3. mutual fund objective and after-tax returns (1987-1992) attribute aanressive growth investment objective growth growth & income balanced number of funds total return rate of wealth accumulation (rwa) tax drag potential tax drag percent tax drag used unrealized capital gain as percent of wealth yield size (million$) reported turnover effective turnover 12 18.68% (4.46%) 15.87% (4.44%) 2.80% (0.98%) 6.54% (1.56%) 44.57% (19.03%) 37.16% (14.43%) 0.36% (0.77%) 388 (366) 131.07 (65.93) 20.0 25 15.84% (3.41%) 13.33% (3.67%) 2.51% (0.96%) 5.54% (1.19%) 48.61% (23.27%) 31.85% (15.11%) 1.88% (0.99%) 576 (446) 102.92 (131.82) 17.2 9 14.78% (1.18%) 12.15% (1.42%) 2.62% (0.51%) 5.12% (0.40%) 51.82% (11.83%) 25.29% ( 6.77%) 3.46% (0.75%) 1,601 (1,663) 62.25 (47.40) 15.8 12 12.61% (1.34%) 9.96% (1.36%) 2.66% (0.35%) 4.37% (0.47%) 61.64% (11.59%) 18.12% (5.83%) 5.49% (0.85%) 1,425 (1,746) 81.73 (104.26) 12.0 note: this table presents information on four different mutual fund categories. the original data is from the 1993 aa11 mutual funds publication. averages, with standard deviations in parentheses, are presented for selected fund characteristics. the derived effective turnover is also presented. the statistics based on this analysis are presented in table 3. the average for various attributes of the samples are presented with standard deviations in parentheses. the average total return for this period was well above historical norms for the various objective categories. the after-tax return used is the rate of wealth accumulation (rwa); the amount by which an investment in the fund would have changed on an annual basis after paying taxes on the distributions and reinvesting the remainder. the tax drag is the difference between total return and the rwa. the maximum return that could be lost to taxes (100 percent effective turnover) is called the potential tax drag, and is computed by multiplying the total return by the tax rate (35 percent in our analysis). the percent tax drag used is computed by dividing the tax drag by the potential tax drag. the percent tax drag used is a function of the yield and the effective turnover of the fund. the unrealized capital gain as a percent of end-of-period wealth gives an indication of the potential tax liability imbedded in the share price. the reported turnover shown in table 3 is the average of the annual reported turnover for the five year period. if a fund with 100 percent turnover literally sold every security and reinvested the proceeds, then all gains accrued in the past would be realized. this is obviously not what happens with most high turnover funds. the correlation between the reported turnover and return lost to taxes for the aggressive growth and growth funds was only 0.250. a high turnover management style does not necessarily lead to large capital gains distribu m ut ua l f un d g ab el li g ro w th t a b l e 4 . 5i p er fo rm an ce c ha ra ct er is ti cs o f m ut ua l f un d sa m pl e w it h g ro w th o bj ec ti ve p ot en ti al t ax t ax d ra g r ep or te d e ff ec ti ve t ot al r et ur n r a w t ax d ra g d ra g u se d c ap it al g ai n t ur no ve r t ur no ve r d iv y ie ld 21 .4 0% 20 .0 4% 1. 36 % 4. 88 % 1. 41 % t . r ow e p ri ce n ew a m g r 20 .6 0% n ic h ol as l ii te d e d 19 .6 0% m on et ta 21 .0 0% s tr on g d is co ve ry 20 .6 0% a co rn 18 .6 0% f id el it y t re n d 17 .8 0% g in te l 16 .5 0% d re yf u s a pp re ci at io n 16 .0 0% v al u e l in e f u n d 17 .3 0% g ab el ti a ss et 16 .3 0% v an gu ar d in de x e xt m kt 15 .2 0% c li pp er 15 .6 0% v an gu ar d p ri rn ec ap 14 .4 0% f ou n de rs g ro w th 15 .3 0% v an gu ar d m or ga n g ro w th 15 .8 0% c ol u m bi a g ro w th 15 .7 0% p en n sy lv an ia m u tu al 14 .4 0% s tr on g o pp or tu n it y 13 .5 0% w il li am b la ir g ro w th 16 .2 0% t . r ow e p ri ce c ap a pp 14 .0 0% t . r ow e p ri ce g r s th 12 .5 0% b os to n c o. c ap a pp 10 .5 0% m at h er s 9. 30 % t . r ow e p ri ce n ew e ra zk !z ? 20 .0 2% 18 .1 3% 17 .1 0% 16 .9 4% 16 .1 0% 15 .6 8% 14 .5 8% 14 .2 2% 14 .0 7% 14 .0 3% 13 .9 5% 13 .2 9% 12 .8 8% 12 .5 6% 12 .4 5% 12 .1 0% 12 .0 9% 11 .9 9% 11 .6 4% 10 .7 2% 10 .1 0% 6. 78 % 6. 33 % & il ?& 0. 58 % 1. 47 % 3. 90 % 3. 66 % 2. 50 % 2. 12 % 1. 92 % 1. 78 % 3. 23 % 2. 27 % 1. 25 % 2. 31 % 1. 52 % 2. 74 % 3. 35 % 3. 60 % 2. 31 % 1. 51 % 4. 56 % 3. 28 % 2. 40 % 3. 72 % 2. 97 % z& z? 7. 49 % 7. 21 % 6. 86 % 7. 35 % 7. 21 % 6. 51 % 6. 23 % 5. 78 % 5. 60 % 6. 06 % 5. 71 % 5. 32 % 5. 46 % 5. 04 % 5. 36 % 5. 53 % 5. 50 % 5. 04 % 4. 73 % 5. 67 % 4. 90 % 4. 38 % 3. 68 % 3. 26 % m 18 .1 7% 8. 08 % 21 .3 7% 53 .0 1% 50 .7 0% 38 .3 4% 33 .9 8% 33 .3 1% 31 .7 4% 53 .2 8% 39 .7 7% 23 .4 9% 42 .3 6% 30 .0 7% 51 .1 3% 60 .4 9% 65 .4 9% 45 .8 3% 31 .9 7% 80 .4 2% 67 .0 1% 54 .8 2% 10 1. 25 % 91 .2 4% i? & q e ? 55 .9 5% 57 .% % 51 .2 5% 39 .6 5% 37 .5 4% 41 .3 8% 42 .0 7% 40 .3 0% 40 .4 0% 31 .1 1% 36 .5 7% 42 .4 2% 34 .7 5% 37 .8 6% 29 .7 0% 25 .7 7% 22 .8 8% 31 .6 3% 35 .7 5% 12 .5 7% 19 .6 7% 23 .5 3% -2 .4 0% 3. 73 % 4. 24 % 60 .2 0% 39 .0 0% 22 .8 0% 18 3. 20 % 66 2. 00 % 29 .6 0% 53 .6 0% 69 .4 0% 92 .4 0% 11 1. 00 % 52 .4 0% 13 .8 0% 34 .0 0% 16 .6 0% 18 0. 20 % 49 .6 0% 15 9. 40 % 22 .6 0% 26 8. 60 % 31 .4 0% 79 .2 0% 33 .8 0% 10 7. 60 % 18 6. 60 % 14 .0 0% 2. 25 % 8. 22 % 36 .9 9% 14 .6 5% 33 .2 2% 12 .6 1% 4. 81 % 11 .6 6% 24 .4 6% 11 .8 4% 5. 99 % 14 .7 2% 11 .8 5% 17 .1 2% 30 .8 6% 35 .1 0% 19 .3 7% 5. 98 % 59 .7 0% 25 .3 6% 34 .5 7% 56 .8 6% 41 .5 7% x u z 22 .7 1% 16 .0 5% -0 .0 3% 1 . o o % 1. 06 % 1. 06 % 1. 35 % 1. 57 % 3. 16 % 1. 63 % 1. 75 % 0. 80 % 2. 47 % 2. 48 % 1. 03 % 0. 95 % 2. 17 % 1. 76 % 2. 22 % 3. 28 % 1. 34 % 3. 65 % 1. 77 % 1. 96 % 4. 53 % u q !& a ve ra ge s ta n da rd d ev ia ti on 15 .8 4% 3. 41 % 13 .3 3% 3. 67 % 2. 51 % 0. 96 % 5. 54 % 1. 19 % 48 .6 1% 23 .2 6% 31 .8 5% 15 .1 1% 10 3. 92 % 13 2. 82 % 1. 88 % 0. 99 % after-tax mutual fund returns 137 tions. this is because mutual fund managers do not sell a uniform proportion of their assets. on average, fund managers sell more losers than winners. in order to relate capital gains distributions to turnover, an effective turnover estimate was derived based on the amount of total return lost to taxes because of realized capital gains. the effective turnover figure is the amount of uniform turnover that would cause the fund category to distribute the amount of capital gains that were distributed over the five year period? although the derived effective turnover rate is only a small proportion of the average reported turnover rate, it still has a substantial effect since only a small amount of effective turnover creates a large amount of tax drag. the effective turnover for the growth fund category will be examined in detail below. the tax drag is fairly uniform across investment objective categories, ranging from 2.51-2.80 percent. since total return decreases from aggressive growth to balanced funds, the potential tax drag also decreases. this results in the aggressive growth funds having the lowest percent tax drag used and the balanced funds the highest. the primary cause of tax drag varies by fund objective. the aggressive growth funds have very low yields, and the tax drag is caused almost entirely by realized capital gains. the balanced funds have high yields which contribute over 70 percent of the calculated tax drag with the remainder due to realized capital gains. it is important to note the variance in the tax drag is relatively low for balanced funds and high for growth funds. this indicates that funds with low tax drag are more easily found in the growth fund categories. growth funds due to their high total returns and low yields, growth funds are frequently used by investors desiring to maximize wealth accumulation over a long-term investment horizon. the five year performance results for the growth fund sample are presented in table 4. the funds are ranked by rwa in descending order. the effective turnover estimate presented in this table is derived by computing the capital gain distributed using the nav at the beginning of the five year period as the basis. an effective turnover is computed for each year. the number reported in table 4 is the average for the five year period. this reported number is biased upward somewhat since the unrealized capital gain is not being considered.6 the top three funds had low turnover and low amounts of return lost to tax drag. the next two funds, monetta and strong discovery, had high turnover and a relatively large tax drag. two funds, t. rowe price new america growth and strong discovery, were tied for third with a total return of 20.60 percent. but the after-tax returns of these two funds differs by more than three percent. this large difference in after-tax results is due to the difference in effective turnover and dividend yield. strong discovery realized large capital gains in 1991 and 1992 which caused it to have a high tax drag. this large difference in after-tax return is important information to the investor selecting between these two funds for use in a taxable savings plan. william blair growth shares had the largest tax drag, 4.56 percent, but a reported turnover that was well below average for this group. the effective turnover, however, is the highest in this group of funds. the fund’s management turned an above average total return into a below average after-tax performance by consistently realizing their capital gains. over the five year period this fund had a total of $5.92 in capital gains, of which $4.74 were realized and distributed. the nav increased from $8.21 to $9.39, making their unrealized capital gains the fourth lowest in the growth fund sample. this fund is a good example of 138 financial services review, 3(2) 1994 why the prospective investor cannot be guided by low reported turnover in seeking to avoid tax drag. for the growth fund sample, the correlation between return lost to taxes and reported portfolio turnover is 0.277 while the correlation between return lost to taxes and effective turnover is 0.680. the after-tax returns and tax drag data provided in table 4 are useful when an investor is selecting a fund from a particular objective category for a taxable savings plan. in this case, tax drag, along with the fees and expenses of the fund needs to be considered, and in most cases minimized. if savings are going into a tax-deferred plan, then the tax drag can be ignored. when an investor splits savings between tax-deferred and taxable accounts, funds with high tax drag should be placed in tax-deferred vehicles, and funds with low tax drag in taxable funds. undistributed income and capital gains usually a tax liability is associated with earned income or realized capital gains. but this is not the case for an investor who has to pay taxes on income or capital gains that provide no increase in wealth. at the time of purchase, the share price of a mutual fund may reflect undistributed income, capital gains, and unrealized appreciation of securities. any income or capital gains from these amounts which are later distributed are fully taxable in a taxable plan. investors in stock mutual funds are advised to make large purchases after a fund makes its major annual distribution in order to avoid incurring the tax liability associated with this distribution. even if an investor makes purchases after the annual distribution, net share holder redemptions or management decisions can result in the fund liquidating positions, and realizing capital gains in which the new investor did not participate. the following example is used to demonstrate nonparticipant capital gain tax risk. assume an investor makes an investment in an index fund with a nav of $20,50 percent of which represents unrealized capital gains. the investment is made just before a market drop of 10 percent, which in turn leads to other investors redeeming 20 percent of the shares of the fund. these redemptions cause turnover which results in 20 percent of the remaining capital gain of $9 to be realized. later, the investor will receive a capital gain distribution of $1.80 which creates a tax liability of $0.63 (assumed tax rate of 35%) that causes a loss of 3.15 percent ($0.63/$20.00) on the original investment. thus, this investor will lose 10 percent because of the market decline plus 3.15 percent due to the tax liability. the investor can avoid the tax liability by redeeming the shares prior to the fund’s distribution date, but the 10 percent market loss is realized when this is done. while this example might be considered extreme, an investor in a mutual fund is at some risk of paying taxes on gains in which he or she did not participate. this can occur in bull markets when managers realize gains for market timing purposes or other tactical reasons. it is difficult for the long range investor in stock mutual funds to avoid this risk. one way is to invest in funds which have little or no unrealized gains. funds of this type include new funds, funds that have done poorly in the past, or funds that have a history of taking their gains quickly. this is not a very attractive group of funds from which to pick. a method of minimizing the nonparticipant capital gain tax risk when using funds with large unrealized capital gains is to only invest in funds that have a relatively long record of asset growth. this type of fund will have accumulated shares over a long period of time, and will be able to initially sell shares with the highest cost basis, typically the last purchased. when this type of fund is forced to sell assets due to net redemptions, a lifo approach will ajter-tax mutual fund returns 139 minimize realized capital gains. fund managers are aware of the tax consequences of their turnover and, hopefully, are motivated to keep their taxable shareholders happy by minimiz ing tax drag. the index fund one of the funds included in the growth and income sample is the vanguard index trust-500. this fund is designed to mimic the returns of the s&p 500 index by investing, using index weighting factors, in the index companies. the vanguard index trust-500 had the lowest amount of return lost to taxes in the growth and income sample (1.65%), the lowest reported turnover (lo%), and an effective turnover of only 4.2 percent. the average effective turnover rate for this fund group was 15.8 percent. these results mean that over a 20 year period vanguard zndex trust-500 will beat the average growth and income fund in a vap if the vap plan has total expenses greater than 1.19 percent. recall that the average total expense for a vap fund is 2.05 percent. the vanguard index trust-500, with 1992 annual expenses of 0.19 percent, and less than five percent effective turnover, certainly sets the standard in the growth and income objective category. in keeping with its low tax drag, the vanguard index trust-500 fund accumulated unrealized capital gains of 37 percent in the five year period. since inception, the nav of the fund has gone from $10 to $40.97. the new investor might be concerned about the risk of paying tax on prior gains because of the large unrealized capital gain imbedded in the nav, but this risk is minimized due to the constant growth in fund assets over the last 12 years. during this period the fund has been continuously buying shares of the companies in the index. if redemptions force a sale of a percentage of these shares, the lifo approach will result in minimal capital gains being realized. ii. summary and conclusions paying taxes on distributions from mutual funds reduces the rate at which savings grow. when taxes are deferred, compounding increases the ultimate after-tax return. in order to achieve savings goals, investors need to realize certain after-tax results. unfortunately, after-tax performance results are not reported by most of the services that report on mutual fund performance. reported mutual fund turnover has been suggested as an indicator of potential loss of return due to taxes on distributions. the results described in this paper show that reported turnover has a low correlation with computed after-tax returns. this lack of correlation is the result of mutual fund managers selling more losers than winners. high reported turnover does not necessarily mean that a fund manager is realizing past capital gains. the five year historical results reported on above show that a mutual fund can have above average before-tax returns, but below average after-tax returns. these results also show that up to three percent difference in after-tax performance can be attributed to fund management style. it is important that investors focus on the impact that distributions and taxes have on their savings plans in order to maximize after-tax return. it is common for an investor to have some savings in qualified tax-deferred plans and the remainder in taxable savings plans. it is also common for investors to diversify among different risk categories. the strategy to use in this situation is to have investments with high 140 financial services review, 3(2) 1994 tax drag in the tax-deferred plan and investments with low tax drag in the taxable plan. since average tax drag is relatively consistent across fund categories, more detailed information on the alternative funds is needed to optimize this decision. the data presented in table 4 show that low tax drag funds are in the growth fund category. it is unlikely that a low tax drag bond fund exists due to the high level of distributions required of this type of fund. without a detailed analysis, the investor is probably better off having the bond or fixed income portion of savings in the tax-deferred plan. taxable plans would usually be the better place for stock funds. since investors do not have adequate information to minimize taxes and maximize wealth accumulation, it is recommended that mutual fund reporting services provide an after-tax return based on a typical, but standardized tax rate. the rate of wealth accumulation (rwa) is recommended as the standard after-tax return. this return can easily be provided, along with the total return, for five and 10 year historical periods. notes 1. reported portfolio turnover is defined as the lower of purchases or sales divided by average net assets. 2. morningstar is an independent mutual fund rating service. they cover over 3000 mutual funds and provide both print and electronic subscription service to investors. 3. in 1994, married-filing-jointly taxpayers with income between $36,000 and $89,150 have a marginal tax rate on income of 28 percent and a tax rate on long term capital gains of 28 percent. investors in the highest tax brackets face a marginal tax rate of 39.6 percent on income, but the rate on capital gains remains at 28 percent. since most states also have income taxes, the rate of 35 percent used in the following examples typifies an investor with a marginal federal tax rate of 28 percent living in a state with moderate to high state income tax rates. 4. since only funds with a size larger than $100 million were used, there is a survivorship bias in these results. for example, the aa11 average growth fund realized a 14.8 percent return over the five year period compared to this study’s growth fund sample average of 15.8 percent. this paper’s historical analysis is not meant as a general examination of mutual fund performance. its purpose is to analyze reported turnover and after-tax results. for a recent review of studies on mutual fund performance and market efficiency see ippolito (1993). 5. effective portfolio turnover estimates for the various categories of funds were derived by modeling the mutual fund category attributes. the inputs to the model were average total return, unrealized capital gains, and yield. 6. the average effective annual turnover reported in table 4 (22.71%), is greater than the derived figure reported in table 3 (17.2%). the database used for this study was limited to five years. computing annual effective turnover for a fund over part of its existence will result in over estimating the effective turnover due to unrealized capital gains imbedded in the nav at the beginning of the study period. after-tax mutual fund returns 141 references american association of individual investors. (1993). the individual investor’s guide to no-load mutualfinds. (12th ed.). chicago, il: author. ippolito, r.a. (1993). on studies of mutual fund performance, 1962-1991. financial analysts journal (jan./feb.), 42-50. jeffrey, r.h., & arnott, r.d. (1993). is your alpha big enough to cover its taxes?. the journal of portfolio management (spring), 15-25. optimism, overconfidence, and insurance decisions jennifer coatsa, vickie bajtelsmita,* adepartment of finance and real estate, colorado state university, 1272 campus delivery, fort collins, co 80523, usa abstract we report experimental evidence regarding overconfidence, optimism, and insurance decisions. our design distinguishes between an individual’s optimism bias and overconfidence bias, a contribution particularly important for understanding insurance decisions related to risks beyond the purchaser’s control. results show that optimistic participants incur a higher total cost of risk and are more likely to underinsure than non-optimistic participants, even when purchasing insurance maximizes expected payoffs. in contrast, we find that overconfidence does not significantly affect the decision to insure. however, participants with higher overall overconfidence show larger differences in insurance behavior when the risk of loss arises from their own mistakes. © 2021 academy of financial services. all rights reserved. keywords: c9 design of experiments; c91 laboratory experiments; d81 decision-making under uncertainty; overconfidence; insurance demand 1. introduction optimism bias, or the tendency to assign higher subjective probabilities to favorable outcomes, is well documented in the psychology and economics literature. in fact, after decades of controlled studies in psychology, the only individuals identified as consistently free from this bias are the clinically depressed (pyszczynski, holt, & greenberg, 1987). overconfidence, which can be viewed as a special case of optimism, relates to having a biased perception of one’s own skills, prospects, or knowledge. both optimism and overconfidence theoretically affect decision-making under conditions of risk and uncertainty (de bondt & thaler, 1985). although previous studies differ in how these factors are defined and operationalized, the *corresponding author. tel.: +1-970-491-0610. e-mail address: vickie.bajtelsmit@colostate.edu (v. bajtelsmit) 1057-0810/21/$ – see front matter © 2021 academy of financial services. all rights reserved. financial services review 29 (2021) 1–28 conclusions overwhelmingly suggest that underestimation of the risk of negative outcomes has an economically significant effect on individual and societal well-being.1 general optimism about events outside of one’s own control may cause individuals to underestimate their actual risk, which may lead them to make suboptimal financial decisions. for example, underestimation of expected losses may result in reduced demand for insurance (kunreuther & pauly, 2004) and optimism about market performance may affect portfolio allocations (jacobsen, lee, marquering, & zhang, 2014). similarly, overestimation of one’s financial skills may result in excessive and costly financial market trading (barber & odean, 2001) or investment in suboptimal business projects (malmendier & tate, 2005).2 in this article, we focus on the degree to which heterogeneous risk perceptions, which may be affected by both optimism and overconfidence, influence insurance and risk management decisions. to the extent that these biases result in underestimation of personal risks, individuals are hypothesized to have reduced demand for insurance and greater total cost of risk. although overconfidence and optimism are widely discussed in the psychology and economics literatures, the influence of these behavioral biases on insurance decision-making has received less attention. information asymmetry in insurance markets, in which applicants for insurance know more about their own risk characteristics than do insurers, is shown theoretically to create the potential for market failure. with heterogeneous risk types, equilibrium solutions result in separating contracts that encourage individuals to select price and coverage policies that are appropriate to their risk type.3 for example, a high risk individual might prefer a contract that provides relatively full coverage for a higher premium rate, whereas an individual with a lower risk of loss might select a partial-coverage policy for a lower premium rate. critical to the success of self-selection equilibria models is that individuals can correctly self-identify their risk type. bajtelsmit and thistle (2015) develop a model in which noisy or imperfect information about risk types could result in suboptimal insurance and risk management decisions. in this article, we consider information imperfection caused by individual psychological biases that influence an individual’s risk perceptions. our study uses a novel experimental design which connects, through large monetary incentives, an earnings task, a frequency estimation task, and insurance decisions. participants develop subjective estimates of their own and others’ performance on an earnings task and decide, in light of their subjective probabilities of loss, whether to fully insure against loss of earnings. in this article, we first focus on the origins of subjective probability estimation, as it depends on task performance, and then develop measures to distinguish between optimism and overconfidence biases in the estimation task. finally, we investigate how the biases affect the risk management decision to fully insure against loss. our analysis of insurance decisions in this article continues a stream of research from a large experimental design which, as a whole, considers an extensive set of variables shown theoretically to affect risk management and insurance decisions.4 the experiment design allows us to distinguish between an individual’s optimism, or the general tendency to overestimate favorable outcomes (“wishful thinking”), and their overconfidence, the tendency to overestimate their own performance or skills (“i think i’m better than i am”) and the relationship between these outlooks.5 we assess optimism with a treatment manipulation in which a participant’s expected payoff is independent of their own performance on a task, but positively related to the performance of others on the same task. 2 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 participants’ outlooks range from underconfident to highly overconfident in their own performance and from pessimistic to optimistic about others’. participants who are optimistic about others’ performance are generally confident in their own as well. we find that overconfidence bias does not reduce the likelihood of insurance purchase, but that the highly optimistic are less likely to purchase insurance and, controlling for risk preferences, tend to underinsure. the remainder of the article is organized as follows. section 2 describes related literature on insurance and overconfidence. section 3 explains the experimental design and hypotheses. we provide results in section 4, and discuss conclusions in the final section. 2. related literature the empirical link between overconfidence and the purchase of insurance is an important contribution in support of recent theoretical findings (huang, liu, & tzeng, 2010; spinnewijn, 2013) and leads to policy implications applicable to insurance contract design and government policy interventions. huang et al. (2010) show theoretically that hidden overconfidence can lead to insurance market equilibria that are consistent with observed anomalies in insurance markets. they suggest that insurers could use an overconfidence proxy to screen prospective policyholders so as to achieve an advantageous selection equilibrium and that regulators might need to intervene in markets that are operating imperfectly as a result of this type of asymmetric information.6 in recent years, researchers have distinguished several different categories of overconfidence bias. however, overconfidence, as a subset of optimism bias, has often remained confounded with a general optimistic outlook. the optimism bias leads individuals to underweight the probability of negative outcomes that are beyond their control, such as an airline crash or a wildfire, and overweight the probability of positive outcomes that are beyond their control, such as winning a lottery. spinnewijn (2013) refers to this phenomenon as “baseline optimism.” for example, landry and jahan-parvar (2011) suggest that failure to purchase flood insurance may be related to residents’ reliance on community protection policies (seawalls, beach replenishment, etc.). if the households are overly optimistic regarding the success of the communities’ risk management efforts, then they may underinsure against loss. in addition to a baseline level of optimism, individuals may be overconfident with respect to their influence and/or abilities, which then leads to unrealistic optimism about outcomes, a quality spinnewijn (2013) terms “control optimism.” royal and tasoff (2017) show theoretically that overconfident agents are likely to reduce their payoffs by investing too much in capital that complements their ability and too little in capital that substitutes for their own ability. their experimental results support the theoretical predictions. in the insurance context, this could lead an individual to underinsure if they unrealistically believe they can reduce the frequency or severity of loss through their own actions. a large literature in behavioral economics distinguishes further between different types of control optimism. “absolute overconfidence” is a term applied to individuals who overestimate their own knowledge, ability, or performance against a given benchmark, such as the prediction that j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 3 they will perform better on a test or run faster than they actually do (moore & healy, 2008). perhaps the most well-known type of overconfidence is the “better-than-average effect,” also termed “over-placement,” which refers to overestimation of one’s relative performance in a group. an important quality of over-placement is “reference group neglect” in which individuals rank their placement equally high among groups of self-selected members and groups of exogenously assigned members (camerer & lovallo, 1999; moore & healy, 2008). other types of control optimism include the “illusion of control,” which refers to the case in which individuals erroneously perceive their actions to influence independent events, and “calibration-based overconfidence,” or overconfidence in the precision of knowledge.7 fellner and krugel (2012) find evidence that overconfidence assessed with different methods actually reflects separate and distinct biases. several recent studies suggest that heterogeneity of risk perceptions (spinnewijn, 2013) or risk preferences (de meza & webb, 2001) can explain the negative correlation between risk and insurance coverage found in some markets.8 for example, as modeled in sandroni and squintani (2007), differences in risk perceptions may lead some high-risk types to believe that they are low-risk types, which can decrease investment in precaution and/or reduce demand for insurance at offered prices. huang et al. (2010) model optimistic individuals who have subjective loss probabilities that are lower than their objective loss probabilities and “rational” individuals who assess their loss probability correctly. in their model, higher optimism leads to lower likelihood of insurance purchase. arad (2014) presents a study in which participants are aware of objective probabilities but may assign a higher or lower likelihood due to their own personal motivations that are unrelated to the random event. arad labels this phenomenon “magical thinking” and notes that beliefs about one’s own good or bad luck, regardless of probability distribution can also lead to suboptimal insurance decisions. honl, meissner, and wulf (2017) develop a model in which individual risk-taking behavior depends on cognitive processing of outcomes and probabilities, affect in judgment and decision-making, and upon contextual factors. to the extent that insurance purchasers do not know the objective likelihood of a loss event, it is important to consider biases in estimating risk, and insurance decisions in light of subjective probability estimation. while our experiment is not motivated as a study of gender effects or risk preferences per se, extant studies suggest that it is important to control for both factors in the analysis. recent studies designed to investigate gender differences in decision-making find that optimism about conditions or others’ performance varies across men and women. in an analysis of survey data that includes several different indicators of optimism, jacobsen et al. (2014) find men to be more optimistic than women in their expectations about the general economic outlook. in a study spanning three years, foster and frijters (2014) find that male university students are more consistently overconfident than females about their future grades. in an experiment task where payoffs depend on team performance, kuhn and villeval (2015) find that both men and women expect to outperform others on their team, although women are more optimistic than men about their team members’ performance. laboratory experiments designed to model insurance decisions offer the opportunity to measure and control for beliefs, risk attitudes, and the set of risk management alternatives. for example, harrison and ng (2016) find that, after measuring and controlling for risk 4 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 preferences, experimental participants tend to make welfare-reducing insurance decisions. in another controlled laboratory experiment, jaspersen and aseervatham (2017) find that insurance demand decisions are often driven by biases and the use of heuristics. laury and mcinnes (2003) find that information provided by actuarially fair insurance prices can reduce experimental participants’ reliance on heuristics and improve decisions. when experiment participants are allowed to insure against losses that depend on relative performance, hales and kachelmeier (2008) show that insurance decisions are affected by biases in performance estimation. for a comprehensive survey on the experimental literature on insurance demand, see jaspersen (2016) who concludes that the decision context, decision task, and the use of salient incentives all heavily influence experimental results. in this article, we present the results of an experimental study in which we first elicit participants’ beliefs about risk perceptions, and then investigate their subsequent decisions in laboratory insurance decision tasks. 3. experiment procedures and design and hypotheses in this section, we first describe the design and procedures used in this laboratory experiment, and then explain how this design can be used to test various hypotheses related to the effects of overconfidence and optimism on risk management decisions. the experiment is designed with incentive-compatible earnings and risk management tasks. the design also includes an indirectly incentive-compatible frequency estimation task to elicit subjective probabilities for losses that depend on the participants’ own ability and losses that are outside of their influence.9 the reported subjective probabilities allow us to measure participants’ overconfidence in their own performance as well as baseline optimism regarding favorable outcomes. we use these to estimate the impact of overconfidence and optimism biases on incentivized insurance purchase decisions under different risk conditions. 3.1. procedures overview students were recruited from business classes at a large university to participate in a paid experiment. all sessions were conducted with z-tree (fischbacher, 2007) in a networked computer lab with partitioned stations. we conducted six sessions, each with ten participants, between june and october 2013. the experiment proceeded through two stages with steps as summarized in table 1, including participation payment, earnings task, instructions, estimation task, and risk management task. after participants were paid $15 up-front in cash (never at risk of loss) and learned the experiment procedures, they earned $60 for correctly answering at least eight questions on a quiz comprised of twenty questions drawn from previous driver licensing exams for the state in which their university was located.10,11 for each question, participants were asked to indicate whether they were sure they had answered it correctly. to incentivize participants to carefully make these assessments, the instructions clearly explained the relationship between their quiz performance, others’ quiz performance, the probability of loss, and their expected earnings from the experiment. after participants demonstrated their understanding of these j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 5 t ab le 1 e x p er im en ta l p ro ce d u re s f ir st st ag e s te p 1 • $ 1 5 in ca sh im m ed ia te ly u p o n en te ri n g la b . p ar ti ci p at io n p ay m en t an d p ro ce d u re s o v er v ie w • a g en d a fo r ea rn in g s an d ri sk m an ag em en t ta sk s. s te p 2 • t w en ty m u lt ip le ch o ic e d ri v in g q u iz q u es ti o n s d is tr ib u te d o n p ap er . e ar n in g s ta sk • m o n et ar y in ce n ti v es to an sw er co rr ec tl y . • in d ic at e w h et h er su re o f an sw er . s te p 3 • r is k m an ag em en t ta sk in st ru ct io n s d is tr ib u te d an d re ad al o u d . in st ru ct io n s • r el at io n b et w ee n ea rn in g s, es ti m at io n , an d ri sk m an ag em en t ta sk s ex p la in ed w it h ex am p le s. • in st ru ct io n s as se ss m en t an d re v ie w . s te p 4 • e n te r d ri v in g q u iz fi n al an sw er s an d w h et h er su re o f ea ch . e st im at io n ta sk • e n te r es ti m at e o f o w n sc o re . • e n te r es ti m at e o f av er ag e sc o re ea rn ed b y o th er p ar ti ci p an ts . s ec o n d st ag e s te p 1 • n o m is ta k es , o w n m is ta k es , an d o th er s’ m is ta k es tr ea tm en ts . r is k m an ag em en t ta sk s • p re ca u ti o n , in su ra n ce , an d in it ia l p ro b ab il it y o f lo ss tr ea tm en t m an ip u la ti o n s p re se n te d in ra n d o m o rd er . s te p 2 • c o m p le te al l d ec is io n s, th en re v ie w ea ch , o n e at a ti m e. r ev ie w al l d ec is io n s • m u st ac ti v el y co n fi rm o r re v is e ea ch in d iv id u al d ec is io n . s te p 3 • r an d o m d ra w b y a p ar ti ci p an t d et er m in es tr ea tm en t ap p li ed fo r se ss io n p ay m en t. s el ec ti o n o f a tr ea tm en t fo r p ay o ff • p ar ti ci p an ts ar e ea ch p ai d ac co rd in g to th ei r d ec is io n in th e d ra w n tr ea tm en t. 6 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 relationships, we proceeded with the estimation task in which they recorded an estimate of the number of questions they had answered correctly and an estimate of the average number correct for the other participants in the session. in the second stage, participants made risk management and insurance decisions in several treatments. participants finalized their choices in the program only after experiencing all decisions for the experiment, and no losses or outcomes were realized until after confirmation of all decisions. a random draw at the end of the experiment was used to select the treatment used to determine payoff. 3.2. treatment design we use a within-subjects design with 60 participants each completing 8 treatments.12 this results in 480 participant-treatment observations, some of which are used in the primary analyses, and some of which are used only to check participant rationality or as controls. to minimize order effects, treatment manipulations are randomized across participants in each session. in each treatment, participants are exposed to a risk of losing $45 from their $60 earnings. treatment manipulations include the initial loss probability (10% or 32%), and the determinants of overall loss probability. for each initial loss probability, three treatment manipulations that differ in the way that quiz performance determines the overall probability of loss as follows: • no mistakes: the risk of loss is implemented as a computer-generated random number—explained with the analogy of a random draw from 100 white and orange ping-pong balls. the risk of loss is expressed as a percentage (10% or 32% orange balls) and also described in terms of number of orange and white balls, respectively. participants are told that, if an orange ball is drawn, they lose $45. • own mistakes: as in the no mistakes treatments, there is a draw from a known distribution of orange and white ping-pong balls (10% or 32% orange). participants are told that they lose $45 if an orange ball is drawn but, if a white ball is drawn, there is a random draw from their own driving quiz questions. if they answered the drawn question correctly, they do not lose any money, but if they answered it incorrectly, they lose $45. • others’ mistakes: as in the own mistakes treatments, there is a draw from a known distribution of orange and white ping-pong balls (10% or 32% orange). participants are told that they lose $45 if an orange ball is drawn but, if a white ball is drawn, there is a random draw from a different participant’s driving quiz questions. if the other participant answered it incorrectly, they lose $45. before learning about whether they experienced a loss, participants made risk management (insurance or precaution) decisions described as the option to pay a dollar cost from their earnings to replace orange balls with white balls. in each case, they were presented with a menu of incremental options as in the example in the appendix. consistent with jasperson and aseervatham (2015), who emphasize the importance of a choice frame in laboratory experiments of insurance decisions, the insurance and precaution decisions are framed as choice tasks rather than elicitations of willingness to pay for insurance. the cost j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 7 of insurance is $14.50 and, if purchased in any treatment, it reduces their probability of loss to zero. each $1.50 spent on precaution reduces the probability of loss by one percentage point in the low probability of loss treatments and by four percentage points in the high probability of loss treatments.13 in the no mistakes treatments, participants could reduce their initial probability of loss to zero through buying the maximum level of precaution, making this choice equivalent to full insurance. buying full precaution is more expensive than insuring in the low probability treatments, but less expensive in the high probability treatments. our within-participants design allows us to observe that participants who wish to reduce risk to zero make rational choices between precaution and insurance. in the mistakes treatments where the probability of loss depends on quiz performance, insurance is the only option for reducing risk to zero, because even with the purchase of full precaution (replacing all orange balls with white balls), the risk of loss from mistakes remains. in the no mistakes treatments, participants know their risk of loss with certainty whereas, in the mistakes treatments, they must make subjective assessments over the risk of quiz mistakes to determine their probability of loss. table 2 summarizes the way in which participants’ estimates of quiz scores impact estimates of the probability of loss prior to any investments in risk mitigation. table 3 summarizes the optimal insurance decisions for a risk-neutral participant in each of the treatments used in this study.14 the treatments that do not provide an insurance option are used to categorize and control for participants’ risk preferences but are not otherwise included in the primary analysis. the no mistakes low probability treatment provides a check of participant rationality but is not included in the primary analysis. several treatments allow for the possibility of underinsurance by risk-neutral (or riskaverse) participants. in all of the high (32%) initial probability of loss treatments, purchasing insurance is optimal for risk-neutral (or risk-averse) participants in that it results in the highest expected payoff. in the low (10%) initial probability treatments with loss probability independent of performance, the highest expected payoff results from not purchasing any risk mitigation. however, in the treatments where loss probability depends on performance, purchasing insurance is optimal for risk-neutral or risk-averse participants who performed poorly on the quiz. participants maximize expected payoffs by purchasing insurance if less table 2 probability of losing $45 before risk management decisions in different treatments treatments initial probability of loss low (10%) high (32%) own mistakes: dependent on own performance 0:10 þ 0:90 � 20�ownquiz scoreð þ 20 0:32 þ 0:68 � 20�ownquiz scoreð þ 20 others’ mistakes: dependent on others’ performance 0:10 þ 0:90 � 20�others’ avg: scoreð þ 20 0:32 þ 0:68 � 20�others’ avg: scoreð þ 20 no mistakes: independent of performance 0.10 0.32 8 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 than 76% of quiz questions are answered correctly. therefore, overestimation of performance can lead to underinsurance. 3.3. the incentive for accurate estimation in this experiment design in addition to the show-up fee paid at the beginning of stage 1, participants received a payment at the end of stage 2 (in private and in cash) based upon their performance, risk management decisions, and chance. monetary incentives connect the earnings, estimation, and risk management tasks across the two stages. a higher score on the driving quiz decreases the probability of loss in the own mistakes treatments in the same way that loss event probability estimation and insurance decisions correspond to expected wealth effects in practice. that is, the estimation task facilitates comparison by the participants between their expected payoff without insurance versus their payoff with insurance. more accurate score estimates improve a participant’s ability to make an optimal insurance decision. in summary, higher scores on the quiz reduce the risk of loss in the mistakes treatments, but overestimation of scores could lead to suboptimal insurance decisions and lower expected payoffs, and this correspondence rewards participants for accuracy in their estimation.15 for example, a risk-neutral participant with an initial 10% probability of loss who estimates scoring 90% correct on the driving quiz in the own mistakes treatment (expected loss of $8.55) is better off not purchasing insurance for $14.50. however, if the participant’s actual performance on the driving quiz is 60% (expected loss of $20.70), the expected payoff is higher with insurance. participants recorded their estimated scores only after they had received all the instructions and passed the instructions assessment, demonstrating they understood how both the probability of mistakes on the quiz and the cost of insurance affected their expected payoffs. at that point, participants were aware of how their quiz table 3 optimal risk management decisions to minimize expected loss in each treatment treatments initial probability of loss low (10%) high (32%) own mistakes: risk depends on own performance insure if quiz score < 76% no risk mitigation if > 76% insure others’ mistakes: risk depends on others’ performance insure if quiz score < 76% no risk mitigation if > 76% insure no mistakes: risk is independent risk mitigation-precaution and insurance no risk mitigation full precautiona risk mitigation-precaution onlyb no risk mitigation full precaution a participants pay to reduce the initial probability of loss before a mistake is drawn in increments of 10 percentage points. in the no mistakes treatments, the purchase of full precaution reduces the probability of loss to zero and is, therefore, the risk mitigation equivalent to buying insurance in this design. full precaution is the more efficient means to reduce risk to zero in the high probability treatments, while insurance is the more efficient means in low probability treatments. b precaution is the only risk management tool. these treatments are used only to categorize and control for risk attitudes. j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 9 scores and score estimation accuracy combined with their risk management and insurance decisions to affect their earnings in the experiment. as in actual insurance purchase decisions, a participant in the experiment who chooses not to insure based on an optimistic or overconfident estimate of loss probability faces a larger total cost of risk compared with insurance decisions based on more accurate assessments. all responses were completely anonymous. there were no financial or risk management incentives to report higher or lower scores than estimated, and optimal decisions depended on using best estimates. therefore, participants had incentives to accurately estimate their risk of loss and faced no incentive to inaccurately report their estimates.16 consequently, the earnings task is directly incentive-compatible, and the corresponding estimation task is indirectly incentive-compatible, with respect to maximizing final payoffs in the experiment. participants completed all decisions before receiving their earnings from the estimation and risk management task. after all decisions were completed, reviewed, and confirmed, a public random draw of a numbered ping-pong ball by a participant determined the treatment used to pay out earnings. each participant’s individual earnings for the chosen treatment depended on a computer-generated random number representing either an orange ball or white ball, and if applicable, random selection of quiz question. 3.4. hypotheses in this section, we draw on the existing literature to categorize participants as optimistic or overconfident based on decisions made in the experiment and develop hypotheses about the expected impact of optimism and overconfidence on participants’ insurance decisions. participants who overestimate their own quiz scores are classified as overconfident in their own knowledge or abilities. we measure a participant’s overconfidence as the percentage by which they overestimate (or underestimate) their own score and we call this measure the gross overconfidence bias.17 gross overconfidence bias ranges from negative (under confident) to positive (overconfident) so that results closer to zero reflect smaller biases. we measure a participant’s general level of optimism (or pessimism) about an unknown probability of loss outside of his or her own control, as the percentage by which they overestimate (or underestimate) the average score of other participants. our measure of optimism bias also ranges from negative (pessimistic) to positive (optimistic), with measures closer to zero reflecting smaller biases. because an individual may be generally optimistic or pessimistic about outcomes that beyond their control, the accuracy of the participant’s estimate of others’ average score is used as a proxy for a general tendency to underestimate or overestimate the risk of loss in the experiment, independent of their own knowledge or ability. participants who overestimate others’ scores are classified as optimistic because overestimation of scores corresponds to underestimating the probability of loss due to errors that are beyond their own control. an optimistic outlook that causes participants to underestimate the chance of loss in general may also influence participants’ estimation of their own scores. in other words, a participant’s overestimation of their own score may be attributable to overconfidence in their own abilities, a general optimistic outlook, or a combination of the two biases. therefore, we also measure net overconfidence bias as the difference between the errors in estimates of own and others’ scores. 10 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 again, there is no compensation associated with relative performance, or reward for above average performance in this experiment. the larger the positive bias in a participant’s own estimate compared with the bias in estimating others’ average score, the higher that participant’s net overconfidence bias. participants who overestimate their own performance by the same or smaller percentage than they overestimate others’ do not exhibit the net overconfidence bias. based on the theoretical literature on insurance demand, and the literature on optimism and overconfidence, we develop the following hypotheses to be tested using the experiment design described in the previous section: hypothesis 1: individual biases a. gross overconfidence bias: participants will exhibit gross overconfidence bias and will, therefore, overestimate their own performance. this measure of bias includes the possibly confounded effects of general optimism and overconfidence in their own ability to minimize the risk of loss. b. optimism bias: participants will exhibit optimism bias and will, therefore, overestimate others’ average performance. c. net overconfidence bias: participants will exhibit net overconfidence bias and will overestimate their own performance to a larger extent than they overestimate others’ performance. after adjusting for general optimism about overall performance, participants’ estimates of their own performance will reflect a positive bias. hypothesis 2: effect of biases on decision-making by causing individuals to underestimate the risk of loss, optimism and overconfidence biases will lead to underinsurance against losses. we define underinsurance as declining to insure when the expected payoff is higher with insurance than without it.18 a. gross overconfidence bias will increase the likelihood of underinsurance. our measure of gross overconfidence bias directly reflects an underestimate of the probability of loss in the own mistakes treatments. therefore, the gross overconfidence bias will be more likely to increase underinsurance against a loss depending on a participant’s own performance, than a loss depending on others’ mistakes. b. optimism bias will increase the likelihood of underinsurance. errors in assessing risk due to general optimism do not depend on the source of loss. however, our measure of optimism bias directly reflects an underestimate of the probability of loss in the others’ mistakes treatments. therefore, the optimism bias will be more likely to increase underinsurance against a loss resulting from others’ mistakes than a participant’s own mistakes. c. net overconfidence bias should not affect the probability of underinsurance in this design because the estimated probability of loss does not depend on relative performance in any way. j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 11 hypothesis 3: effect of biases on the total cost of risk in this experiment, participants can spend money on partial risk mitigation and insurance. participants who do not either insure or purchase precaution face expected losses that depend on their quiz performance, while those who insure incur the known cost of the insurance premium. therefore, our measure of the total cost of risk is the sum of the expected loss resulting from the risk event and the known amount spent on precaution or insurance. if the overconfident or optimistic choose to pay for precaution rather than fully insuring, they may actually increase their total cost of risk. we hypothesize that: a. higher gross overconfidence bias will be associated with greater total cost of risk. the increase in the total cost of risk will be higher when the probability of a loss depends on one’s own performance (because the overconfidence bias directly measures errors in assessing risk in the own mistakes treatment). b. higher optimism bias will be associated with greater total cost of risk. the increase in the total cost of risk will be higher when the probability of loss depends on others’ performance (because the optimism bias directly measures errors in assessing risk in the others’ mistakes treatment). c. net overconfidence, the difference between a participant’s gross overconfidence and optimism, does not inform the estimation of risk in either mistakes treatment. therefore, it will not affect participants’ total cost of risk. the next section presents and discusses the outcomes of participant decisions. then it introduces controls used in the analysis and presents tests of these hypotheses. 4. results and analysis 4.1. summary statistics table 4 presents summary statistics for the participants’ performance on the earnings and estimation tasks, and summarizes the definitions used for participants’ optimism, gross and net overconfidence biases. the summary statistics show that participants overestimate their own performance to a greater extent (9.26%) than they overestimate others’ performance (2.10%) and that there is a great deal of within-sample variation in these bias measures. given our experiment parameters ($45 loss, $14.50 insurance premium, and precaution alternatives), risk-neutral and risk-averse participants are predicted to purchase insurance (or the full precaution equivalent) for all high initial probability treatments. if they do not purchase insurance, their expected payoff decreases with lower quiz scores in the mistakes treatments. however, risk-seeking participants in the 32% initial probability no mistakes treatments may prefer not to purchase insurance in the mistakes treatments as well. under the low initial probability of loss, the expected loss is $4.50 in the no mistakes treatments, and a risk-neutral participant should not purchase insurance. however, in the low initial probability mistakes treatments, any participant who answers 76% or fewer quiz questions 12 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 correctly has higher expected payoffs from purchasing insurance, making insurance the optimal choice for risk averse and risk neutral participants. as discussed above, overconfidence and/or optimism regarding quiz scores could lead participants to underestimate their risk of loss, which could lead to underinsurance in light of their risk preferences. while table 4 shows the average levels of the biases, table 5 categorizes participants in terms of which biases they exhibit. the table illustrates that participants are relatively unlikely to exhibit one of these biases and not the other. they are more likely to have gross overconfidence and optimism bias or show neither bias. results in tables 4 and 5 reveal that while 50% of participants exhibit some degree of optimism, the average level of optimism does not differ significantly from zero. however, 60% of participants exhibit gross overconfidence, and the average level of gross overconfidence is significantly above zero. it follows that their average net overconfidence bias is also significant. in fact, 60% of participants exhibit the net overconfidence bias. in summary, we find descriptive evidence in support of parts a (gross overconfidence bias) and c (net overconfidence bias) of hypothesis 1, but not part b (optimism bias). we summarize the insurance and precaution purchase decisions in table 6. in the no mistakes treatments, most participants insure against loss, all purchase some form of risk mitigation in the initial 32% probability of loss treatments, and most purchase some risk mitigation in the initial 10% probability of loss treatments. participants insure against loss more frequently under the others’ mistakes treatment than the own mistakes treatment. some participants purchase precaution to reduce their risk of loss. for example, in the no mistakes 32% initial probability treatment, those participants who do not insure, purchase table 4 summary statistics: driving quiz, estimation task, and biases, n = 60 participants mean minimum maximum standard deviation estimated own quiz score (out of 20) percent correct 16.4 82% 11 55% 19 95% 2.06 estimated others’ quiz score (out of 20) percent correct 15.4 77% 10 50% 18 90% 1.94 actual quiz score (out of 20) percent correct 15.1 75% 12 60% 18 90% 1.50 gocbias ¼ own estimate� own score own score 9.26%*** �18.75% 41.67% 15.80% optimism bias ¼ estimated others’score� others’ score others’score 2.10% �33.86% 19.46% 12.89% noc bias = overconfidence bias – optimism bias 7.16%*** �23.83% 39.65% 14.72% ***significantly different from zero at the 99% confidence level, according to the t test and nonparametric signed rank test. goc and noc denote gross and net overconfidence, respectively. table 5 classification of participants by optimism and gross overconfidence not overconfident overconfident (gross overconfidence) not optimistic 33% 17% optimistic 7% 43% differences in proportions are significant at the 99% level in a x2 test. j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 13 t ab le 6 p er ce n ta g es ch o o si n g ea ch ri sk m it ig at io n al te rn at iv e, b y tr ea tm en t m is ta k es tr ea tm en t n o m is ta k es o w n m is ta k es o th er s’ m is ta k es in it ia l p ro b ab il it y o f lo ss 3 2 % 1 0 % 3 2 % 1 0 % 3 2 % 1 0 % b u y in su ra n ce o r fu ll p re ca u ti o n a 6 5 % 4 3 % 5 7 % 5 5 % 6 8 % 6 7 % b u y p ar ti al p re ca u ti o n (a v er ag e re d u ct io n in in it ia l p ro b ab il it y )b 3 5 % (� 2 0 % ) 3 8 % (� 5 % ) 4 3 % (� 2 2 % ) 2 3 % (� 5 % ) 3 2 % (� 2 1 % ) 1 7 % (� 3 % ) n o ri sk m it ig at io n 0 % 1 9 % 0 % 2 2 % 0 % 1 6 % a t h e p u rc h as e o f fu ll p re ca u ti o n re d u ce s th e p ro b ab il it y o f lo ss to ze ro an d is , th er ef o re , th e eq u iv al en t to b u y in g in su ra n ce in th is d es ig n (a lt h o u g h n o t th e sa m e co st ). b p ar ti ci p an ts p ay to re d u ce th e in it ia l p ro b ab il it y o f lo ss b ef o re a m is ta k e is d ra w n in in cr em en ts o f 1 0 p er ce n ta g e p o in ts . t h e n u m b er s in p ar en th es es sh o w th e av er ag e p er ce n ta g e p o in t re d u ct io n in in it ia l p ro b ab il it y . f o r ex am p le , in th e n o m is ta k es 3 2 % in it ia l p ro b ab il it y tr ea tm en t, 3 5 % o f p ar ti ci p an ts re d u ce th ei r p ro b ab il it y o f lo ss b y an av er ag e o f 2 0 % . 14 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 sufficient precaution to reduce the initial risk from 32% to 12% on average. in the no mistakes 10% initial probability treatments, participants who buy precaution instead of insurance reduce the risk of loss from 10% to 5% on average. 4.2. risk attitudes and gender controls the primary purpose of this research is to analyze the relationship between overconfidence, optimism, and insurance purchase. however, participants make insurance and precaution decisions in light of their estimation of the risk of loss and their attitudes about accepting different levels of risk. previous literature suggests that overconfidence bias (that affects estimation of risk) may vary systematically with gender. at the same time, optimal insurance purchase decisions will differ based on risk attitudes. this subsection discusses summary statistics particular to the risk attitude and gender control variables included in the analysis. risk-neutral or risk-averse participants should purchase insurance whenever the expected loss exceeds the insurance premium. all else equal, risk-seeking participants would be less likely to purchase insurance. we use the precaution only no mistakes treatments to provide information about participants’ risk attitudes. the within-subject design allows us to observe individual participants’ choices across each different treatment. in the precaution only no mistakes treatments, full precaution provides equivalent risk mitigation to insurance. under an initial probability of loss of 10%, expected payoff is decreasing in precaution, and full precaution (the equivalent of insurance) costs roughly three times the expected loss. under an initial probability of loss of 32%, expected payoff is increasing in precaution, and the cost of full precaution is only 83% of the expected loss. therefore, we can identify participants whose behavior is consistent with risk-seeking preferences in the 32% probability of loss treatments, and those who make choices consistent with risk aversion in the 10% probability of loss treatments. we analyze each participant’s choices across no mistakes precaution only treatments to broadly classify risk attitudes. we classify participants as risk averse if they purchase any precaution in the 10% initial probability treatment and also purchase full precaution in the 32% initial probability treatment. participants who exhibit risk-averse behavior under the 10% initial probability treatment, but risk-seeking behavior under the 32% initial probability treatment are considered to be “reflexive.”19 we classify participants as risk-seeking if they purchase less than full precaution in the 32% initial probability treatment and also do not purchase any precaution in the 10% initial probability treatment. the risk-seeking classification also controls for an interpretation of optimism in which some participants are optimistic about their “luck” in the outcome of a random draw with a known distribution, even if they are realistic about the distribution itself. because risk-seekers could optimally choose to remain uninsured in cases where purchasing insurance is optimal for risk-neutral or risk-averse individuals, we control for evidence of risk-seeking behavior in our analysis of the relationship between optimism, overconfidence, and insurance purchase. to avoid multicollinearity problems between this control variable and others, we use the participants’ decisions in the precaution only no mistakes treatments j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 15 (that are not included in the main regressions) only to estimate risk attitudes of the participants. table 7 presents the classification of participant risk attitudes, optimism, and overconfidence. risk-seeking behavior is evident in 17% of the sample. as discussed above, participants who overestimate others’ average score are classified as optimistic and table 4 showed no statistically significant optimism bias in the sample (though this is conservative given the small sample size). however, table 7 reveals that half of the participants do not overestimate others’ scores. (the participants who overestimate others’ scores have an average error of 11.74%, while those who underestimate others’ scores have an average error of only 7.5%.) we classify participants who overestimate their own score as exhibiting gross overconfidence in their own knowledge or abilities; but some or all of that overconfidence may be due to general optimism. therefore, we also report the participants classified as exhibiting net overconfidence bias. comparing these metrics by gender, we find that a larger percentage of men are overconfident, both before and after adjusting for general optimism. results in tables 5, 6, and 7 confirm the presence of optimism and overconfidence. table 8 summarizes the respective sample correlations of these measures, along with the categorical variables risk-seeking and male. correlations with the categorical variable male in table 8 show that there are no significant gender differences on average in gross overconfidence or optimism biases. however, women’s average optimism is slightly higher and gross overconfidence slightly lower than men’s. when these effects are combined, the differencein-differences between men and women is significant for net overconfidence bias. men exhibit a larger difference between gross overconfidence in their own quiz performance and optimism about others’ quiz performance, compared with women. these results are consistent with those reported in kuhn and villeval (2015), although the earnings tasks in the two experiments are very different. table 7 classification of risk attitudes, optimism, and overconfidence percent of participants percent of females percent of males risk seeking 17% 17% 17% optimism 50% 50% 50% gross overconfidence 60% 54% 64% net overconfidence 67% 63% 69% table 8 biases, risk attitude, and gender correlations gross overconfidence optimism net overconfidence risk seeking male gross overconfidence 1 optimism 0.49*** 1 net overconfidence 0.65*** �0.35*** 1 risk seeking 0.08 0.02 0.06 1 male 0.19 �0.14 0.33*** 0 1 ***significant at the 99% level. 16 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 4.3. the effect of overconfidence and optimism on the decision to insure we now turn our attention to the second set of hypotheses, which considers the effects of the biases on the decision to insure and on the total cost of risk. because full precaution is equivalent to insurance in the no mistakes treatments, participants who purchase either of those options in the no mistakes treatments are counted as buying insurance. although some participants purchase partial precaution rather than insurance in the mistakes treatments, this decision results in their being underinsured when the probability of loss is 32% or higher (unless they are risk-seeking). participants who are overconfident in their own quiz performance or precaution decisions, or optimistic with respect to others’ quiz performance, will underestimate this probability and therefore could make suboptimal insurance decisions in the mistakes treatments. hypothesis 2 predicts that the likelihood of underinsurance in the others’ mistakes treatment will increase with higher levels of optimism bias. it also predicts that insurance decisions in the own mistakes treatments will depend on the gross overconfidence bias, which as discussed above, may also include general optimism. we expect both the optimism bias and the gross overconfidence bias to increase the likelihood of underinsurance in the own mistakes treatment. fig. 1 presents the incidence of underinsurance (from a risk-neutral perspective) for participants in each treatment, according to whether they exhibit optimism (panel a), gross overconfidence (panel b), or net overconfidence (panel c). the optimism bias is associated with underinsurance in the others’ mistakes low probability treatment, but also in the no mistakes and own mistakes high probability treatments. we find that, as expected, the gross overconfidence bias is associated with underinsurance when payoffs depend on one’s own performance, but not others’ performance, and the net overconfidence bias has an insignificant impact on underinsurance in this experiment. we further analyze the effect of the biases on underinsurance through logit regressions presented in table 9, in which the dependent variable is a dummy variable where underinsurance =1 indicates underinsurance from a risk-neutral perspective. categorical independent variables include gender (female is the omitted category), risk treatment type (10% initial probability is the omitted category), and risk-seeking attitude (no evidence of risk seeking is the omitted category).20 regression coefficients are presented in terms of log odds. due to the strong relationship between the biases, we run separate regression models, including different bias measures as independent variables in each. the top panel of table 9 presents analysis of the mistakes treatments (n = 240 decisions, with 60 standard error clusters). conceptually, we would expect optimism and gross overconfidence (that encompasses optimism) to influence insurance decisions in both mistakes treatments. at the same time, our measure of optimism directly informs subjective probability of loss in the others’ mistakes treatment and our measure of gross overconfidence directly informs subjective probability of loss in the own mistakes treatment. therefore, we are especially interested in the significance of interaction terms for the biases and mistakes treatment. in particular, we would expect the gross overconfidence measure to have a more significant impact on underinsurance in the own mistakes treatments and general optimism to be more influential in the others’ mistakes treatments. j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 17 fig. 1. panel a: percent underinsuring by treatment and optimism. panel b: percent underinsuring by treatment and gross overconfidence (goc). panel c: percent underinsuring by treatment and net overconfidence (noc) notes: differences in proportions are significant at the 90% level (*) based on a x2 test. p(loss) represents the initial probability 32% and 10% treatments; no mistakes, own mistakes, and others’ mistakes denote no mistakes, own mistakes and others’ mistakes treatments, respectively. 18 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 t ab le 9 d et er m in an ts o f u n d er in su ra n ce , lo g it re g re ss io n co ef fi ci en t es ti m at es an d (s ta n d ar d er ro rs ) m is ta k es tr ea tm en ts , n = 2 4 0 , st an d ar d er ro rs cl u st er ed b y p ar ti ci p an t m o d el 1 : g ro ss o v er co n fi d en ce m o d el 2 : o p ti m is m m o d el 3 : n et o v er co n fi d en ce m ai n ef fe ct s in te ra ct io n ef fe ct s m ai n ef fe ct s in te ra ct io n ef fe ct s m ai n ef fe ct s in te ra ct io n ef fe ct s in te rc ep t �1 .5 0 5 0 * * * (0 .3 9 7 2 ) �1 .6 0 8 9 * * * (0 .4 1 0 4 ) �1 .7 8 9 4 * * * (0 .4 0 9 5 ) �1 .8 0 6 5 * * * (0 .4 1 2 1 ) �1 .4 9 8 6 * * * (0 .4 2 4 8 ) �1 .5 7 0 6 * * * (0 .4 3 5 7 ) o th er s’ m is ta k es �0 .7 1 2 2 * * * (0 .2 3 9 3 ) �0 .4 9 6 5 * (0 .2 5 0 7 ) �0 .7 6 0 7 * * * (0 .2 5 3 2 ) �0 .7 2 0 2 * * * (0 .2 3 6 9 ) �0 .7 3 3 2 * * * (0 .2 4 1 9 ) �0 .6 0 5 5 * * (0 .2 6 0 0 ) g o c b ia s 0 .7 3 8 8 (1 .5 1 4 0 ) 1 .7 0 1 3 (1 .6 0 5 4 ) g o c � o th er s’ m is ta k es �2 .1 8 0 4 * (1 .2 4 8 1 ) o p ti m is m b ia s 5 .0 2 8 1 * * * (1 .5 4 2 7 ) 5 .4 4 0 7 * * * (1 .6 5 0 8 ) o p ti m is m � o th er s’ m is ta k es �0 .9 3 3 2 (1 .7 8 6 0 ) n o c b ia s �3 .1 2 8 8 * (1 .7 9 8 8 ) �2 .1 2 9 8 (1 .7 5 3 2 ) n o c � o th er s’ m is ta k es �2 .4 3 0 6 (1 .5 3 1 1 ) m al e 0 .5 6 0 0 (0 .4 6 4 4 ) 0 .5 6 4 7 (0 .4 6 7 2 ) 0 .8 2 7 1 * (0 .4 4 3 9 ) 0 .8 2 8 7 * (0 .4 4 6 0 ) 0 .9 2 8 7 * (0 .4 9 1 5 ) 0 .9 3 6 1 * (0 .4 9 2 4 ) in it ia l p ro b ab il it y = 3 2 % 0 .6 2 4 3 * (0 .2 7 1 6 ) 0 .6 2 8 5 * * (0 .2 7 4 3 ) 0 .6 6 6 9 * * (0 .2 9 0 8 ) 0 .6 6 7 7 * * (0 .2 9 1 9 ) 0 .6 4 2 8 * * (0 .4 9 1 5 ) 0 .6 4 4 9 * * (0 .2 8 4 9 ) r is k -s ee k in g 1 .6 7 4 6 * * * (0 .3 8 7 6 ) 1 .6 8 9 0 * * * (0 .3 9 7 7 ) 1 .8 0 3 9 * * * (0 .4 4 3 9 ) 1 .8 0 5 7 * * * (0 .3 6 8 2 ) 1 .8 3 0 5 * * * (1 .7 9 8 8 ) 1 .8 5 9 9 * * * (0 .3 9 5 9 ) n o m is ta k es p (l o ss ) = 3 2 % tr ea tm en t, n = 6 0 m o d el 1 : g ro ss o v er co n fi d en ce m o d el 2 : o p ti m is m m o d el 3 : n et o v er co n fi d en ce in te rc ep t �1 .1 6 1 5 (0 .5 2 0 5 ) �1 .2 1 9 7 * * (0 .4 9 5 3 ) �1 .0 9 1 2 * * (0 .4 8 1 8 ) g o c b ia s 1 .0 1 5 2 (1 .8 6 1 8 ) o p ti m is m b ia s 2 .4 2 9 8 (2 .3 5 7 4 ) n o c b ia s �0 .7 2 5 2 (0 .6 0 3 7 ) m al e 0 .7 0 4 6 (0 .5 9 4 6 ) 0 .8 5 6 3 (0 .6 0 1 5 ) 0 .8 3 3 8 (0 .6 0 3 7 ) n o te s: * , * * , * * * d es ig n at e si g n ifi ca n ce at th e 9 0 % , 9 5 % , an d 9 9 % co n fi d en ce in te rv al s, re sp ec ti v el y . t h e d ep en d en t v ar ia b le is ze ro if th e p ar ti ci p an t d id n o t u n d er in su re , an d 1 if th ey d id . c o ef fi ci en ts ar e p re se n te d in lo g o d d s te rm s. g o c an d n o c d en o te g ro ss an d n et o v er co n fi d en ce , re sp ec ti v el y . j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 19 controlling for the biases, initial probability of loss, and risk-seeking preferences, we find that participants are significantly less likely to underinsure in the others’ mistakes treatments than in the own mistakes treatments. model 1 shows that gross overconfidence itself is not a significant predictor of underinsurance. however, the interaction effect in the second column reveals that the difference in the probability of underinsuring between the own and others’ mistakes treatment does depend on gross overconfidence. as gross overconfidence increases, there is a greater increase in the likelihood of underinsuring in the own mistakes compared with the others’ mistakes treatments. in other words, participants with higher gross overconfidence are even more likely (compared with those who are less overconfident) to make suboptimal insurance decisions when the risk of loss depends on their own performance than when it depends on others’ performance. this is somewhat consistent with the hypothesis 2 predictions about biases. in model 2, we find that participants who exhibit higher levels of optimism bias are significantly more likely to underinsure. however, the interaction between optimism and the others’ mistakes treatment is not significant, which suggests that the effect of optimism is no more influential in the others’ mistakes treatment than in the own mistakes treatment. finally, in model 3, we find that net overconfidence bias has a negative relationship with underinsurance that is weakly significant and independent of the mistakes treatment. we attribute this result to the fact that net overconfidence is decreasing in optimism, which as discussed above, has a strong positive and significant relationship with the probability of underinsuring across both mistakes treatments. the second panel of table 9 displays results for the no mistakes condition. in this treatment, participants can only underinsure in the 32% initial probability of loss treatment, because purchasing no risk mitigation is the payoff-maximizing choice in the 10% initial probability treatment. therefore, we examine underinsurance decisions separately for the no mistakes, 32% probability of loss treatment (n = 60), and find that, as expected, none of the biases have a significant effect.21 this suggests that participants’ bias stems from their beliefs about the probability of loss rather than optimism or pessimism about luck when confronting a known distribution. however, the insignificant effect on the insurance decision may also be attributable to the smaller sample size using only one treatment for this statistical test. 4.4. the effect of overconfidence and optimism on the total cost of risk as described in hypothesis 3, the relationship between the total cost of risk (the sum of the expected loss resulting from the risk event and the known amount spent on precaution or insurance) and the biases provides a way to examine the cost of underinsurance, in particular when participants have access to alternative risk mitigation measures instead of insuring. for example, given the experiment parameters in the 10% initial loss probability own mistakes treatments, a participant who answers 70% of the quiz questions correctly and insures faces a total cost of $14.50 (out of their $60 earnings). if they underinsure by neither purchasing insurance nor precaution, they face an expected loss of $16.70 in the initial 10% probability treatment. however, if that participant pays for precaution to reduce the initial 20 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 probability of loss to 5%, then the total cost of risk, including the cost of precaution, is actually even higher at $22.60. participants who are overconfident about the effects of precaution increase their total cost of risk when they choose to purchase partial precaution instead of either insuring or doing nothing. table 10 presents results of a generalized least squares regression estimating the impact of the biases and control variables on the total cost of risk. as in the previous regressions, we separately analyze the effects of the three types of biases. the results for model 1 show that gross overconfidence bias is a statistically and economically significant factor increasing the cost of loss in the own mistakes treatment. furthermore, the effect is significantly lower for participants in the others’ mistakes treatments as compared with the own mistakes treatments. overconfidence in one’s own ability is more costly when the expected loss depends on own performance. in model 2, we find that optimism bias also increases the total cost of risk, but there is no significant difference in this relationship across mistakes treatments. this too is expected, since errors in assessing risk due to general optimism do not depend on the source of loss. finally, model 3 confirms that net overconfidence bias does not significantly increase the total cost of risk. however, as the combined results for gross overconfidence and optimism imply, the impact of net overconfidence on the total cost of risk is lower in the others’ mistakes treatment than in the own mistakes treatment. as in the previous table, the smaller sample size when using only the data from the no mistakes treatment (n = 60) reduces the power of the test but, in this case, we still find the effect of net overconfidence to be marginally significant. 5. conclusions this article contributes to the literature by providing experimental evidence regarding the effects of overconfidence and optimism on insurance decisions. in the existing literature, optimism is sometimes confounded with overconfidence, and we contribute an innovative design that distinguishes between overconfidence regarding the likelihood of a favorable outcome resulting from one’s own performance versus optimism regarding the likelihood of a favorable outcome outside of one’s own influence. we find that these psychological biases have important implications for insurance in that they can cause individuals to underestimate their risk and, therefore, underinsure, resulting in a higher total cost of risk. this effect is stronger for insurance over risks that depend on one’s own performance as compared with exogenous risks. our results contribute to the growing body of literature on the effect of psychological biases on financial decisions and reinforce the importance of careful measurement of overconfidence bias in laboratory experiments. this distinction is particularly important for understanding insurance decisions related to risks outside of the purchaser’s control. the relationship between overconfidence and optimism may also help explain the perceived tradeoffs between risk mitigation and insurance decisions. our results show that, after controlling for risk-seeking behavior, participants are more likely to make suboptimal insurance decisions when the risk of loss depends on their own j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 21 t ab le 1 0 d et er m in an ts o f to ta l co st o f ri sk , g en er al iz ed le as t sq u ar es re g re ss io n co ef fi ci en t es ti m at es an d (s ta n d ar d er ro rs ) d et er m in an ts o f u n d er in su ra n ce in m is ta k es tr ea tm en ts , n = 2 4 0 , st an d ar d er ro rs cl u st er ed b y p ar ti ci p an t m o d el 1 : g ro ss o v er co n fi d en ce m o d el 2 : o p ti m is m m o d el 3 : n et o v er co n fi d en ce m ai n ef fe ct s in te ra ct io n ef fe ct s m ai n ef fe ct s in te ra ct io n ef fe ct s m ai n ef fe ct s in te ra ct io n ef fe ct s in te rc ep t 1 4 .6 3 1 1 (0 .3 9 8 8 ) 1 4 .8 3 7 3 * * * (0 .4 0 3 4 ) 1 4 .5 5 8 3 * * * (0 .3 9 3 1 ) 1 4 .5 2 9 7 * * * (0 .3 9 2 2 ) 1 4 .8 0 5 6 * * * (0 .4 3 7 3 ) 1 4 .6 5 0 4 * * * (0 .4 3 4 0 ) o th er s’ m is ta k es �0 .4 8 4 7 (0 .3 3 0 2 ) 0 .0 3 2 0 (0 .2 9 0 0 ) �0 .4 8 4 7 (0 .3 3 0 2 ) �0 .4 2 7 5 (0 .3 1 0 2 ) �0 .4 8 4 7 (0 .3 3 0 2 ) �0 .1 7 4 3 (2 .6 1 5 4 ) g o c b ia s 3 .2 6 3 7 (2 .0 0 6 4 ) 6 .0 5 4 1 * * * (2 .7 7 0 3 ) g o c � o th er s’ m is ta k es �5 .5 8 0 8 * * * (0 .4 3 5 2 ) o p ti m is m b ia s 5 .6 8 8 5 * * * (1 .6 2 5 4 ) 7 .0 4 9 4 * * * (1 .9 7 4 4 ) o p ti m is m � o th er s’ m is ta k es �2 .7 2 1 8 (1 .9 3 7 3 ) n o c b ia s �0 .4 8 4 7 (0 .3 3 0 2 ) 1 .4 1 7 9 (0 .5 8 9 8 ) n o c � o th er s’ m is ta k es �4 .3 3 6 9 * (2 .3 5 5 5 ) m al e 0 .4 0 9 9 (0 .5 3 6 9 ) 0 .4 0 9 9 (0 .5 3 8 1 ) 0 .8 1 7 0 (0 .5 1 5 9 ) 0 .8 1 7 0 (0 .5 1 7 0 ) 0 .6 7 8 0 (0 .5 4 0 1 ) 0 .6 7 8 0 (0 .5 4 1 3 ) in it ia l p ro b ab il it y = 3 2 % 2 .1 9 5 9 * * * (0 .4 3 0 6 ) 2 .1 9 5 9 * * * (0 .4 3 1 5 ) 1 .8 2 7 2 * * * (0 .4 9 3 7 ) 2 .1 9 5 9 * * * (0 .4 3 1 5 ) 2 .1 9 5 9 (0 .4 3 0 6 ) 2 .1 9 5 9 * * * (0 .4 3 1 5 ) r is k -s ee k in g 1 .7 5 9 6 * * * (0 .5 1 8 6 ) 1 .7 5 9 6 * * * (0 .5 3 8 1 ) 1 .8 2 7 2 * * * (0 .4 9 4 7 ) 1 .8 8 3 1 * * * (0 .5 4 0 8 ) 1 .8 8 3 1 * * * (0 .5 4 1 9 ) d et er m in an ts o f to ta l co st o f ri sk in n o m is ta k es p (l o ss ) = 3 2 % tr ea tm en t, n = 6 0 m o d el 1 : g ro ss o v er co n fi d en ce m o d el 2 : o p ti m is m m o d el 3 : n et o v er co n fi d en ce in te rc ep t 1 2 .5 6 5 0 * * * (0 .2 3 6 7 ) 1 2 .6 4 4 6 * * * (0 .2 2 2 2 ) 1 2 .5 8 * * * (0 .2 1 9 4 ) g o c b ia s 0 .6 9 1 0 (0 .8 1 3 3 ) o p ti m is m b ia s �1 .0 4 3 9 (1 .0 3 9 6 ) n o c b ia s 1 .7 3 8 5 * (0 .9 7 8 0 ) m al e 0 .0 2 1 2 (0 .9 2 9 9 ) 0 .0 2 3 6 (0 .2 5 6 8 ) �0 .1 0 6 2 (0 .2 5 8 8 ) n o te s: * , * * , * * * d es ig n at e si g n ifi ca n ce at th e 9 0 % , 9 5 % , an d 9 9 % co n fi d en ce in te rv al s, re sp ec ti v el y . g o c an d n o c d en o te g ro ss an d n et o v er co n fi d en ce , re sp ec ti v el y . 22 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 performance than others’ performance. optimistic participants are more likely to underinsure than non-optimistic participants. overconfidence does not have a significant effect on the decision to purchase insurance, although participants with higher overall overconfidence show larger differences in behavior when they are responsible for the risk of loss than when it is beyond their control. analysis of the total cost of risk, including both the expected loss and the cost of precaution or insurance, under different treatments provides similar evidence about the influence of these psychological biases. overconfidence and optimism both significantly increase the total cost of risk. however, overconfidence has a significantly lower effect on total cost when the loss event is triggered by someone else’s error. optimism bias increases the cost by about the same amount regardless of whether the risk depends on one’s own actions. these results suggest that general optimism extends to outcomes that depend on one’s own ability, but overconfidence in one’s own performance does not affect the decision to mitigate risks due to factors outside of one’s own control, such as those resulting from nature or from others’ errors. the laboratory evidence reported in this article offers a potential explanation for the underinsurance against catastrophe that has been observed in the market. beliefs, together with risk tolerance, preferences, or general probability misperceptions may provide alternative explanations for some of the observed insurance decision puzzles. for example, jacobsen et al. (2015), who study asset allocation decisions under uncertainty, find that optimism about outcomes and optimism about the level of risk are as important as risk aversion in explaining asset allocation. spinnewijn (2013) notes that, while heterogeneity in beliefs informs insurance policy design, it is very difficult to obtain direct evidence about beliefs. laboratory results such as ours provide a first step in connecting beliefs to insurance decisions, with more control than surveys or behavioral proxies for perceptions. while convenience samples of students provide for a strong degree of laboratory control, it would interesting to further explore these issues with a more diverse participant population. future research should also consider in greater detail the influence of these biases on individual perceptions about the effectiveness of different risk mitigation alternatives. notes 1 theoretical explanations in economics and evolutionary biology have also illustrated conditions under which optimism or overconfidence bias can be individually welfare-improving (compared to rational expectations), a second best solution in the presence of other biases, and even a necessary adaptation for species survival. see, for instance brunnermeier and parker (2005), besharov (2004), and johnson and fowler (2011). 2 this definition of overconfidence assumes misperceptions rather than a well-calibrated assessment of one’s own relative abilities. well-calibrated confidence does not produce lower results. for example, fielder (2011) finds that virtual traders who self-report as better than average, in fact earn above average virtual profits. j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 23 3 the adverse selection literature in insurance originated with the seminal work of rothschild and stiglitz (1976). see dionne, fombaron, and doherty (2013) for a more complete summary of this extensive literature. 4 in bajtelsmit, coats, and thistle (2015), the authors focus on the other considerations addressed by the experiment design, including the effect of ambiguity on risk management decisions and the tradeoff between taking precaution and purchasing insurance, as well as replicability of other researchers’ results. 5 we acknowledge that there is a great deal of inconsistency in terminology and trait measurement across the overconfidence literature, both theoretical and empirical. see clark and friesen (2009) and spinnewijn (2013) for more complete reviews of this literature. 6 de meza and webb (2001) show that insurers can design contracts that will result in an equilibrium which they term “advantageous selection” in which the risk-averse agent buys insurance and also invests in some precaution. the risk-neutral agent does not take precaution or buy insurance. 7 however, recent research suggests that calibration-based overconfidence observed in confidence interval reporting may be overstated because of the measurement instrument (blavatskky, 2009; cesarini, sandewall, & johannesson, 2006; glaser, langer, & weber, 2013; soll & klayman, 2004). 8 see chiappori and salanie (2013) for a review of this literature. 9 experiments by cesarini et al (2006), blavatskyy (2009), and clark and friesen (2009) suggest that frequency estimation tasks provide better measures of overconfidence relative to confidence interval estimation tasks because of the improvement in incentive-compatibility, better alignment of accuracy and information, and because framing a forecast as a frequency is a much more natural cognitive task. 10 the earnings, probability estimation, and risk management tasks were explained in a power point presentation at the front of the room, with the instructions read aloud. participants took an instructions assessment to confirm they understood how their earnings would be determined and were able to ask questions. 11 participants were required to have a valid state driver’s license as a condition of participation in the experiment. the driving quiz was designed to include a sufficient number of easy questions such that all participants were expected to be able to achieve the minimum score, but also some more difficult questions to minimize the number who could achieve a perfect 20 out of 20 correct. 12 the within-participants design provides for much greater statistical power than a between-participants design of the same size (see bellamare, bissonnette, and kroger, 2014 and charness, gneezy, and kuhn, 2012). we obtain more participant observations and cluster standard errors at the participant level in our analysis. 13 the return to taking precaution is higher for the high-risk treatments than the low-risk treatments due to the assumption of greater productivity of precaution 24 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 under high initial risk. for a theoretical justification, see bajtelsmit and thistle (2015). 14 the full design also included four additional treatments, in which the probability of loss depended on own and others’ performance under high and low initial probabilities of loss. however, in those treatments, participants could neither reduce loss probability to zero, nor insure against loss and, therefore, we do not include or analyze them in this article. see bajtelsmit et al. (2015). 15 although there are other methods of incentivizing participants, such as scoring rules that reward estimates but penalize errors, participant risk preferences have been shown to affect their choices. see, for example, andersen, fountain, harrison, and rutström (2014) and harrison, martinex-correa and swarthout, (2014). other probability elicitation mechanisms depend on independence between agents’ actions and the probability of the risky event (armentier and treich, 2013) and on the agent having no stake in the risky event (karni, 2009). 16 estimating one score to use in risk management decision making, but recording a different score, while technically possible, would be inconsistent with payoff maximization efforts because it would increase the cognitive difficulty of the risk management task and potentially increase the chance of making a costly risk management error and, therefore, would not be incentive compatible with maximizing experiment payoffs. 17 we thank an anonymous referee for suggesting the “gross overconfidence” and “net overconfidence” labels. 18 because this definition only applies to risk-averse and risk-neutral participants, we control for risk-seeking behavior in our analysis. 19 this classification is discussed in detail in an earlier article by the authors where the no mistakes treatments is compared to the alternative to buy insurance against the no mistakes precaution only treatments and show that (1) participants purchase the more efficient means of risk mitigation, and (2) that participants are consistent in their risk mitigation decisions across treatments. comparison of the precaution only mistakes treatments to the precaution only no mistakes treatments shows that participants respond predictably to the lower effectiveness of precaution by purchasing less precaution in the mistakes treatments. 20 identified through behavior in the precaution only no mistakes treatments. 21 the risk-seeking dummy variable is not included because, in the treatments used to determine risk attitudes (precaution only no mistakes), a risk-seeking identification is indistinguishable from underinsurance in the high probability of loss insurance treatments. appendix: examples of choices available in high and low initial probability treatments j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 25 references andersen, s., fountain, j., harrison, g. w., & rutström, e. e. (2014). estimating subjective probabilities. journal of risk and uncertainty, 48, 207–229. arad, a. (2014). avoiding greedy behavior in situations of uncertainty: the role of magical thinking. journal of behavioral and experimental economics, 53, 17–23. armentier, o., & treich, n. (2013). eliciting beliefs: proper scoring rules, incentives, stakes, and hedging. european economic review, 62, 17–40. bajtelsmit, v., coats, j., & thistle, p. (2015). the effect of risk ambiguity on risk management choices: an experimental study. journal of risk and uncertainty, 50, 249–280. bajtelsmit, v., & thistle, p. (2015). liability, insurance and the incentive to obtain information about risk. the geneva risk and insurance review, 40, 171–193. initial high probability of loss example: choose one of the following options below. decision up-front cost to replace orange balls new number of orange balls new number of white balls probability orange ball is drawn a $0 32 68 32% b $1.50 28 72 28% c $3.00 24 76 24% d $4.50 20 80 20% e $6.00 16 84 16% f $7.50 12 88 12% g $9.00 8 92 8% h $10.50 4 96 4% i $12.00 0 100 0 j (insurance) $14.50 32 68 n/a your decision in scenario 2: ____________________ initial low probability of loss example: choose one of the following options below. decision up-front cost to replace orange balls new number of orange balls new number of white balls probability an orange ball is drawn a $0 10 90 10% b $1.50 9 91 9% c $3.00 8 92 8% d $4.50 7 93 7% e $6.00 6 94 6% f $7.50 5 95 5% g $9.00 4 96 5% h $10.50 3 97 3% i $12.00 2 98 2% j $13.50 1 99 1% k $15.00 0 100 0 l (insurance) $14.50 10 90 n/a your decision in scenario 8: ___________________ 26 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 barber, b. m., & odean, t. (2001). boys will be boys: gender, overconfidence, and common stock investment. the quarterly journal of economics, 116, 261–292. bellemare, c., bissonnette, l., & kroger, s. (2014). statistical power of within and between-subjects designs in economic experiments. discussion paper no. 8583. germany: institute for the study of labor. besharov, g. (2004). second-best considerations in correcting cognitive biases. southern economic journal, 71, 12–20. blavatskyy, p. r. (2009). betting on own knowledge: experimental test of overconfidence. journal of risk and uncertainty, 38, 39–49. brunnermeier, m., & parker, j. (2005). optimal expectations. american economic review, 95, 1092–1118. camerer, c. f., & lovallo, d. (1999). overconfidence and excess entry: an experimental approach. american economic review, 89, 306–318. cesarini, d., sandewall, o., & johannesson, m. (2006). confidence interval estimation tasks and the economics of overconfidence. journal of economic behavior & organization, 61, 453–470 . charness, g., gneezy, u., & kuhn, m. (2012). experimental methods: between-subject and within-subject design. journal of economic behavior and organization, 91, 1–8. chiappori, p., & salanie, b. (2013).asymmetric information in insurance markets: predictions and tests. in g. dionne (ed.), handbook of insurance (pp. 397–422). new york, ny: springer. clark, j., & friesen, l. (2009). overconfidence in forecasts of own performance: an experimental study. the economic journal, 119, 229–251. de bondt, f., & thaler, r. (1985). does the stock market overreact? the journal of finance, 40, 793–805. de meza, d., & webb, d. c. (2001). advantageous selection in insurance markets. the rand journal of economics, 32, 249–262. dionne, g., fombaron, n., & doherty, n. (2013).adverse selection in insurance contracting. in g. dionne (ed.), handbook of insurance (pp. 231–280). new york, ny: springer fellner, g., & krugel, s. (2012). judgmental overconfidence: three measures, one bias? journal of economic psychology, 33, 142–154. fiedler, m. (2011). experience and confidence in an internet based asset market experiment. southern economic journal, 78, 30–52. fischbacher, u. (2007). z-tree: zurich toolbox for readymade economic experiments. experimental economics, 10, 171–178. foster, g., & frijters, p. (2014). the formation of expectations: competing theories and new evidence. journal of behavioral and experimental economics, 53, 66–81. glaser, m., langer, t., & weber, m. (2013). true overconfidence in interval estimates: evidence based on a new measure of miscalibration. journal of behavioral decision making, 2, 405–417. hales, j., & kachelmeier, s. (2008). predicting relative performance in economic competition. journal of behavioral finance, 9, 187–192. harrison, g., martinez-correa, j., & swarthout, t. (2014). eliciting subjective probabilities with binary lotteries. journal of economic behavior & organization, 101, 128–140. harrison, g., & ng, j. (2016). evaluating the expected welfare gain from insurance. journal of risk and insurance, 83, 91–120. honl, h., meissner, p., & wulf, t. (2017). risk attribution theory: an exploratory conceptualization of individual choice under uncertainty. the journal of behavioral and experimental economics, 67, 20–27. huang, r., liu, y., & tzeng, l. (2010). hidden overconfidence and advantageous selection. the geneva risk and insurance review, 35, 93–107. jacobsen, b., lee, j., marquering, w., & zhang, c. (2014). gender differences in optimism and asset allocation. journal of economic behavior & organization, 107, 630–651 . jaspersen, j. (2016). hypothetical surveys and experimental studies of insurance demand: a review. journal of risk and insurance, 83, 217–255. jaspersen, j., & aseervatham, v. (2017). the influence of affect on heuristic thinking in insurance demand. journal of risk and insurance, 84, 239–266. johnson, d., & fowler, j. (2011). the evolution of overconfidence. nature, 477, 317–320. j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 27 karni, e. (2009). a mechanism for eliciting probabilities. econometrica, 77, 603–606. kuhn, p., & villeval, m. (2015). are women more attracted to co-operation than men? the economic journal, 125, 115–140. kunreuther, h., & pauly, m. (2004). neglecting disaster: why don’t people insure against large losses? journal of risk and uncertainty, 28, 5–21. landry, c., & jahan-parvar, m. (2011). flood insurance coverage in the coastal zone. journal of risk and insurance, 78, 361–388. laury, s., & mcinnes, m. (2003). the impact of insurance prices on decision making biases: an experimental analysis. journal of risk & insurance, 70, 219–233. malmendier, u., & tate, g. (2005). does overconfidence affect corporate investment?: ceo overconfidence measures revisited. the journal of finance, 11, 649–659. moore, d., & healy, p. (2008). the trouble with overconfidence. psychological review, 115, 502–517. pyszczynski, t., holt, k., & greenberg, j. (1987). depression, self-focused attention, and expectancies for positive and negative future life events for self and others. journal of personality and social psychology, 52, 994–1001. rothschild, m., & stiglitz, j. (1976). equilibrium in competitive insurance markets: an essay on the economics of imperfect information. the quarterly journal of economics, 90, 629–649. royal, a., & tasoff, j. (2017). when higher productivity hurts: the interaction between overconfidence and capital. journal of behavioral and experimental economics, 67, 131–142. sandroni, a., & squintani, f. (2007). overconfidence, insurance, and paternalism. american economic review, 97, 1994–2004. soll, j., & klayman, j. (2004). overconfidence in interval estimates. journal of experimental psychology: learning, memory, and cognition, 30, 299–314. spinnewijn, j. (2013). insurance and perceptions: how to screen optimists and pessimists. the economic journal, 123, 606–633. 28 j. coats, v. bajtelsmit / financial services review 29 (2021) 1–28 pii: s1057-0810(00)00059-7 an analysis of the medical savings account as an alternative retirement savings vehicle j. tim query*,1 university of georgia, terry college of business, 206 brooks hall, athens, ga 30602, usa abstract personal savings as a percentage of disposable income have dropped steadily since the early 1980s. savings have continued to decline in 1999, as the savings rate—savings as a percentage of after-tax income—dropped to a record low of minus 0.7% in april 1999, according to the department of commerce. the study finds that msa-type accounts are a viable supplement to retirement savings, but should not be used as a replacement for existing retirement alternatives given their current structure. results show that future health care expenditures are an important factor in the success or failure of msas as supplemental retirement accounts. medical savings accounts are currently eligible for long-term care expenses, and to the extent that such expenses occur during retirement, msa balances could be used to pay for retirement expenses. in that respect the accounts already capture the characteristics of a retirement savings account. a comparison of the roth ira with the msa as defined by the 1996 hipaa legislation is also conducted. © 2000 elsevier science inc. all rights reserved. 1. introduction personal savings as a percentage of disposable income have dropped steadily since the early 1980s. in 1997, the savings rate by this measure stood at less than four percent—the lowest rate in 59 years. research for merrill lynch’s annual baby boom retirement index has consistently shown that the baby boomer generation has fallen as much as two-thirds * tel.: 11-706-542-4290; fax:11-706-542-4295. e-mail addressess:jtq@arches.uga.edu (j.t. query), tquery@titan.iwu.edu (j.t. query). 1 illinois wesleyan university, department of business administration, 301 e. beecher street, 324cla, bloomington, il 61702. financial services review 9 (2000) 107–123 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00059-7 behind the rate of savings that they need to maintain their current standard of living in retirement. a 1995 kpmg survey of 1,183 employers found that employees contribute only about one-third of the amount they are eligible to put into their 401(k) plans. savings have continued to decline in 1999, as the savings rate—savings as a percentage of after-tax income—dropped to a record low of minus 0.7% in april, according to the department of commerce. savings were in negative territory for five of the previous six months. even among those who do participate in retirement savings there is a tendency to be exceedingly cautious in their investment choices. as demonstrated by bajtelsmit (1996), overly conservative pension investments by plan participants could have extreme consequences for retirement income security. many individuals and financial advisors are re-examining the medical savings account for its potential role as an alternative savings vehicle. federal guidelines state that an msa account-holder can use msa funds for certain medical expenses, but it is not required. medical bills can be paid with money from some other account (if available) without touching the funds in the msa. early indications suggest that many msa owners may be exercising that option. medical savings accounts are currently eligible for long-term care expenses, and to the extent that such expenses occur during retirement, msa balances could be used to pay for retirement expenses. in that respect the accounts already capture the characteristics of a retirement savings account. this paper examines the feasibility of medical savings accounts as a practical savings opportunity. the numerical analysis is a refinement of previous simulations in a number of areas (see jensen and morlock, 1994, bond et al., 1996). rates used for inflation, stock returns, and yields on bills are based on forecasts by ibbotson and associates for the years 1999–2025. taxable and tax-deferred growth are computed and contrasted. the composition of assumed investments in the msa more closely emulate actual practice than do prior related studies, and claims scenarios are based on an extensive national bureau of economic research study. also, the maximum allowable contribution is indexed in order to align the amounts more closely with the parameters found in the 1996 health insurance portability and accountability act (hipaa) pilot program. a comparison of the roth individual retirement account (ira) with the msa as defined by the 1996 hipaa legislation is also conducted. the public policy objectives behind the creation of the msa and roth ira—saving for health care expenditures and retirement, respectively—are obviously different. however, consideration should be given to an integration of these objectives into a more flexible and comprehensive savings vehicle. if the composition of legislation currently being drafted is any indication, the structure of these two accounts may contain even more similarities in the future. at the present time the option of saving through an msa-type account is limited to the self-employed, employers with 50 or fewer employees, employees of firms with msa plans (not necessarily eligible for federal tax incentives), and individuals not covered elsewhere. the results of this study should be of interest to individuals already using such plans, and to employers who have not previously considered the use of msa plans as a benefit option. 108 j.t. query / financial services review 9 (2000) 107–123 2. an overview of medical savings accounts the introduction of medical savings accounts was motivated by two primary factors: rising health costs and a steady decline in private health insurance. with an msa, people pay premiums to an insurer, but the premiums are lower because the policy has a high deductible that covers only catastrophic expenses. a major medical policy with a $2,500 deductible can often be purchased for about one-half the cost of a policy with a $250 deductible. part of the money saved by the lower insurance premium is deposited to the medical savings account. individuals can withdraw funds from their msa to pay smaller health care expenses below the deductible of their health insurance policy. underlying the concept of an msa is the notion that the traditional health care financing system is inefficient for a number of reasons. first of all, there is limited price comparison in the delivery of medical services, as little incentive exists for the fully insured to reduce costs that are primarily borne by the insurance company. currently, about 95% of all hospital bills and 83% of physicians’ fees are paid by private and public third-party payers. on the average, every time a patient spends a dollar in the medical marketplace, 79 cents is paid by someone else (robbins et al. 1994). in addition, studies have confirmed that the level of deductibles has an influence on the utilization of health care, with the system providing services that patients would not authorize if they had to pay for them out-of-pocket. the health insurance experiment, conducted between 1974 and 1981, looked at the effect of health insurance on spending for medical purposes by enrolling people in a variety of health insurance plans (for a description of the experiment see newhouse, 1993). by monitoring medical spending in these plans, analysts at rand were able to estimate dissimilar medical spending behavior among the various samples based on different out-of-pocket costs. they found that when the out-ofpocket price is zero, individuals use approximately 50% more medical care than when insured pay 95% of the cost out-of-pocket up to a maximum of $1,000 per year. msa-type health plans have been offered by a limited number of companies since at least the early 1980s (doerpinghaus, 1996). missouri became the first state to pass a law recognizing use of msas for state income tax purposes in 1993. around 15 states now have similar legislation. the first federal medical savings account legislation was introduced on may 21, 1992. various pieces of federal legislation have been offered throughout the first half of the 1990s to allow federal tax deductibility of medical savings accounts, but the first successful, although limited attempt to allow such accounts did not occur until 1996. in 1996, congress created a demonstration project permitting small employers and the self-employed to establish up to 750,000 tax-free medical savings accounts (msas). however, lawmakers imposed a number of restrictions that limit who can purchase msas and that constrain the ability of msas to work properly. for example, the program is limited to employers having 50 or fewer employees and the self-employed. as a result, the project is primarily aimed at the small business sector, where traditional health insurance benefits are less likely to be offered and impose the greatest financial burden on employers. in response to the criticisms of the limitations imposed, legislation is being drafted to enlarge the project. in 1999 expanded medicare health plan choices included a msa demonstration project. as many as 390,000 beneficiaries can join msas under the project. 109j.t. query / financial services review 9 (2000) 107–123 under the 1996 health insurance portability and accountability act of 1996, msa is defined as a personal savings account from which unreimbursed medical expenses, including deductibles and co-payments, can be paid. an msa must be in the form of a tax-exempt trust or custodial account established in conjunction with a high-deductible health plan. preferred provider organizations are the most common type of plan, but traditional indemnity plans are also widely available. other plan types including health maintenance organizations, exclusive provider organizations and point-of-service plans, are also available to a lesser degree. a benefit of the msa is that participants under most plans have complete choice as to the doctor or health care provider they wish to use. another advantage is the scope of services that qualify for msa payments. eligible medical expenses are mostly the same ones that are deductible for federal income tax purposes—ignoring the 7.5% of adjusted gross income limitation for the taxpayer that itemizes. in contrast to section 125 plans, or flexible spending accounts (fsas), msas do not penalize account holders for saving more than is used in a given year. under fsas, employees must use their account balance by year-end or lose the funds. this encourages employees to increase utilization near year-end to “use up” any unspent funds. msas allows unused funds to roll over to the next year, discouraging inefficient usage of medical services. critics of the program claim that msas benefit the healthy and wealthy at the expense of other individuals, as only high-risk people will be left in other plans to distribute the risk in the risk pool. many critics of medical savings accounts also point to the financial hardships imposed upon a family with high deductibles. however, this argument is weakened when viewed in the context of the employee’s cost-sharing arrangement in traditional health plans. average employee contributions to health insurance plans have risen consistently since 1983. in that year, single coverage contributions averaged $10 a month and family coverage $33 a month. by 1995, required contributions were three and four times higher for single and family coverage, respectively, than in 1983. during this time period, medical prices, as measured by the medical care component of the consumer price index, doubled. employee contributions increased at about the same rate as the medical care component of the cpi until the mid-1980s, and since then have outpaced the cpi medical care component. critics of medical savings accounts also claim that they provide a disincentive towards preventive care. however, experience suggests that the reverse is true. a survey of golden rule employees who have msas found that 20% actually used their msa for a medical service they would not have purchased under the traditional insurance plan. the experience of companies with a 5 or 6year history of using msas, such as dominion resources, found that their annual premium cost have risen at less than 1% a year. the problems of neglecting preventative care presumably would have shown up in the form of increased insurance premiums. concrete evidence that would support this criticism is not found in the rand study. this governmental study also found that with large differences in total use of medical services, the health outcomes were not significantly different. in other words, increased cost sharing did not have a significant negative effect on the participants’ health status (newhouse, 1993). medical savings accounts are not limited to those falling under the 1996 hipaa legislation. many organizations have already implemented a medical savings account type of health insurance plan, even when such plans do not qualify for federal tax deductibility under 110 j.t. query / financial services review 9 (2000) 107–123 hipaa. they include dominion resources, forbes, golden rule insurance co., quaker oats, and the united mine workers union. a few major msa plan providers have categorized the demographics of their msa owners and applicants. according tomsanews,an online newsletter sponsored by the golden rule insurance company, the market leader in providing medical savings accounts, the demographics of msa owners are as follows: y 72% are families y 55% have children y 10% are single parents y 27.8% are single y 16% of msa applicants were previously uninsured y 77% of msa applicants are self-employed y 23% work for small businesses y the average age of the primary insured is 44 as of may 1999, golden rule insurance co. had a total of 37,323 individual msa plans in force. at fortis health, another leading provider of msa plans, approximately 80% of sales are to individuals and 20% of sales are to small businesses. the medical savings concept has also been embraced internationally. since 1984, singapore has provided its citizens with health care through a form of medical savings accounts. singapore spends only 3.1% of its gross domestic product (gdp) on health care, while the u.s. spends about 14%, yet singapore’s hospitalization rate is about equal to that of hmos in the united states (massaro & wong, 1996). in 1993, china began experimenting with msas. today, about 5 million chinese workers have them. 3. literature review previous research in the area of medical savings accounts has been somewhat inhibited due to limited participation at this point in time. for example, the government accounting office (gao) had been directed by congress to survey enrollees, employers and financial institutions in the hipaa qualified plans and submit a report by january 1, 1999, but was unable to do so at a reasonable cost due to relatively low enrollment. as a result, research has been focused on three primary areas: (1) customer interest in and satisfaction with the plans, (2) theoretical economic arguments and simulations supporting or refuting benefits and/or criticisms of the plans, and (3) simulations evaluating the savings component of the plans. an overview of prior research in each of these areas follows. most surveys of individuals regarding medical savings accounts indicate a high level of interest in the concept. those who are already enrolled in such programs appear to be satisfied with their experience so far. a survey jointly conducted by the kaiser family foundation and harvard university found that medical savings accounts are indeed popular. in a random survey of 1,011 adults to determine whether or not people would choose an msa given the opportunity, they found that if the deductible were $2,000, 43% said they would be “very likely” or “somewhat likely” to choose an msa. if the deductible were 111j.t. query / financial services review 9 (2000) 107–123 $5,000, 37% would be likely to choose an msa. these results were consistent with market studies by the national blue cross blue shield association that found 43% of employees would “definitely or probably” switch to an msa if it were offered to them (kaiser-harvard, 1996). in a survey of employees at golden rule insurance company, 65% of msa enrollees rated the msa plan as “excellent” and 32% rated it as “good.” the survey was conducted independently by the luntz research group. despite the current paucity of statistically meaningful empirical data, many researchers have utilized economic models to simulate the effects of various public policies related to medical savings-type accounts. in a 1998 paper, heffley and miceli examined the economics of incentive-based health care plans. using a model that allowed the consumer to invest in healthy activities, they examined the efficiency properties of incentive plans and compared them to traditional plans. they found that properly constructed incentive plans have the capacity to induce socially efficient levels of healthy activities and preventative care, raising the expected wealth of consumers without reducing insurers’ profits. bond, heshizer and hrivnak (1997) analyzed health insurance cost data from ohio public employers and private firms that had adopted medical savings accounts. their study showed that ohio public employers could reduce their health insurance costs an average of 12% for single coverage and 34% for family coverage with msas, compared to traditional plans. under the msa plans, employee out of pocket costs (opc) would also be lower compared to traditional plans. researchers at the rand institute contend that msas would be attractive to those who expect high health-care costs, because potential out-of-pocket expenses under traditional insurance are higher than under msas. the interventions the researchers evaluated differ in the deductibles of the catastrophic plan and in whether the employee or employer funds the msa. if all insured nonelderly americans switched to msas, their health care expenditures would decline by between 0% and 13%, depending on how the msas are designed. however, not all nonelderly americans would choose msas; taking into account selection patterns, health spending would change by11% to 22%. in a comparison of employee-funded and employer-funded msas, traditional health insurance, and hmos, rand found that 57% of the population would choose an msa. employees choosing the employer-funded msa would have an average income of $29,000, while those remaining in fee-for-service plans and those choosing an hmo would have an average income of $28,000 and $43,000, respectively (see keeler et al., 1996). a simulation done by blue cross/blue shield of ohio provides the basis for that organization’s strong opposition to msas. using a sample of over 38,000 families, blue cross/blue shield documented claims of $159.3 million. assuming family units were given an msa of $3,000 and a $3,000 deductible plan was purchased for them at a cost of $1,200, bc/bs projected a deficit of $50 million from writing an msa plan. however, as pointed out by bond et al. (1996) the study ignores the fact that the 68% of families with claims under the $3,000 deductible would have a total of $53.7 million remaining in their msas. zabinski et al. (1999) used micro simulation methods to examine the equilibrium effect of medical savings accounts combined with catastrophic health plans on health care and nonhealth care expenditures, tax revenues, insurance premiums, and exposure 112 j.t. query / financial services review 9 (2000) 107–123 to risk. they contend that if msa-chps are offered alongside comprehensive plans, biased msa-chp enrollment can lead to premium spirals that drive out comprehensive coverage. the third area of research focuses on the savings component of the medical savings account plans. one question that is critical to the viability of the plans involves the level of health care expenses paid out of the account versus the amount of funds placed in the account. a 1989 survey of about 1 million individuals in large self-insured health plans (adjusted to 1994 dollars), found that about one-third filed no claims, 73% filed claims for less than $300, and 89% filed claims for less than $2,000. it should be noted that self-insured health plans usually have very generous benefits (jensen & morlock, 1994). the national bureau of economic research studied 300,000 employees of fortune 500 companies from 1989–1991 and found that workers who are sick in one year tend to have higher-than average medical expenses in the next few years. however, when medical costs exceed the deductible they would pay nothing or a low co-payment, up to the maximum out-of-pocket. this appears to be in agreement with a study by berk and monheit (1992) concluding that about ten percentage of the population is responsible for three-quarters of all u.s. health care spending in a given year. the average expenditure for people in the top 1% of spenders in 1987 was $63,497 (in 1995 dollars). among people in the top 5% of spenders, the average expenditure was $24,735 (in 1995 dollars). an msa model was also created using the nber study. the economists assumed maximum annual deductibles of $2,250 for individuals and $4,000 for families—and deposits of up to three fourths of the deductible in an msa each year. they determined that by retirement, 90% of the workers would have saved more than $25,000 each in their msas and more than half would have saved $50,000 or more. only 5% of the workers would have saved less than 20% of their employers’ contributions over their lifetimes. in 1997 the medisave america council calculated the projected growth of msas assuming an 8% annual return on the maximum allowable annual family contribution of $3,375. in five years an msa would grow to $21,384, $166,802 in 20 years and more than $1.4 million in 45 years. with family medical expenses of $500 annually, the amounts would be $18,216, $142,091 and more than $1.2 million. at $1,000 of annual expenses, the numbers are $15, 408, $117, 379 and $991, 387. for individuals the maximum contribution is $2,250 a year. the account would grow without medical expenses to $9,266, $72,281 and $610,486. at $500 of expenses per year, the account would grow to $7,682, $59,925, and $506,129 and those at the $1,000 expense level would be $2,930, $22,858 and $193,060. jensen and morlock (1994) calculated the growth of msas under different scenarios to estimate the impact of compounding msa deposits would have on savings. assuming a 20-year-old employee, and an eight percentage return on an $1,800 annual investment, the employee would have the following amounts based on these different medical expense scenarios: sufficient time has accrued for some initial data collection on the actual experience of 113j.t. query / financial services review 9 (2000) 107–123 participating employees in selected msa health plans. preliminary results generally support forecasts for a significant accumulation of funds in the accounts. at the golden rule company, the average refund returned to employees was $603 in 1993, $1,002 in 1994, $997 in 1995, $976 in 1996, and $925 in 1997. total cumulative money and earnings left in funds from all golden rule msa policyholders through 1997 was over $22 million dollars. also in 1996, the health and hospital corporation of marion county, indiana returned an average refund of $557 to its 314 employees, with 12% of employees receiving refunds of around $1,000. a 1995 survey of 17 firms using msas found average remaining balances of around $600 for single msa coverage. the amount was approximately $900 for family coverage. up to 80% of the employees in these 17 firms had funds remaining in their msas at the end of the coverage year (barchet et al., 1995). 4. using the msa as a primary savings vehicle while federal guidelines state that an msa account-holder can use msa funds for certain medical expenses, it is not a requirement. medical bills can be paid with money from some other account without touching the funds in the msa. early indications suggest that many msa owners may be exercising that option. a survey of msa policyholders at time insurance co. in milwaukee found that 65% said they did not want time to automatically send them a check from the account when they file a claim. if they are not saving specifically for a big medical purchase then most may be using their accounts as an additional savings vehicle (panko, 1997). a merrill lynch financial consultant specializing in msas in california found that about 80% of employees view the plans as a way to supplement their long-term savings. which month the medical savings account is opened does make a difference in that year. the savings element is affected by the timing of the purchase but the deductible is not. for example, if you open an msa in october, you still have the full deductible, but you can only fund three twelfths (3/12s) of the savings account, so purchases are more appealing early in the year. much of the funding for the msa can come from the reduction in insurance premiums. the savings in premiums when changing from a low-deductible to a high-deductible plan can vary anywhere from 20%–60%. average annual medical expenses years amount accumulated zero 5 $ 11,404 zero 25 $142,118 zero 45 $751,367 $250 5 $ 9,821 $250 45 $647,010 $1,000 5 $ 5,068 $1,000 45 $333,941 114 j.t. query / financial services review 9 (2000) 107–123 from a public policy perspective, the argument for encouraging savings via msas is appealing. according to the u.s. census bureau, the number of americans most likely to need long term care (85 years and older) will double in the next 25 years, and the number of americans over 90 will triple. allowing individuals in their late thirties and early forties to have an msa in which they could build up two or three decades of savings would give these individuals the funds to pay for drug therapies, nursing home care, and in-home care. they will not be forced to turn to medicaid or medicare programs when they need long term care, thus saving the u.s. government hundreds of millions of dollars in the future. after considering the cost of administration, penalties, and contribution limits, it may be more prudent to maximize funding of a qualified retirement plan, such as a 401(k) plan, simple ira, or roth ira before using an msa for retirement purposes. however, there are feasible scenarios where an msa would be attractive as an alternative or supplemental savings vehicle. there are those who have already contributed the maximum allowable to existing retirement plans, yet have sufficient compensation to fund an additional taxadvantaged opportunity. these individuals, if they qualify, would find the medical savings account to be an attractive tool for supplementing their savings. also, some may view msas as a prime way to build up a tax-deferred nest egg without bumping up against section 415 (pension) limits. a second scenario involves the individual contemplating the risk of medical expenditures without other health care coverage and with no retirement plan. for that person, it may be advisable to create and fund the msa first. a comparison of the msa eligible for federal tax preferences and the roth ira in their current formats, respectively, is presented in the next section of the paper. 5. msas vs. roth iras there are a number of similarities in the legislative structure of msa accounts and roth iras—and some notable distinctions. an insurance company or bank, as well as selected other financial institutions, can be a msa trustee or custodian. any entity already approved by the irs as a trustee or custodian for iras is qualified as well. 5.1. contributions for roth iras and msas, individuals’ contributions must generally be made by april 15 of the year following the year for which the contributions are made. all contributions to the msa and roth ira must be made in cash, and neither can be invested in life insurance contracts. the maximum contribution allowed in the medical savings account for 1998 was $3,375.00 for a family and $1,462 for an individual. the msa contribution may not exceed compensation. beginning after 1998, the high deductible dollar amounts are indexed for inflation in $50 dollar increments based on the consumer price index. the actual msa contribution that can be deducted is limited to 1/12 of the annual amount times the number of months an individual is eligible for msa participation. 115j.t. query / financial services review 9 (2000) 107–123 the maximum contribution allowed in a roth ira is $2,000 for an individual. a working couple with earned income can contribute $2,000 each into the roth ira. a nonworking spouse can also contribute up to $2,000 each year. the total amount of contributions to all iras (both traditional and roth iras) for any taxable year may not exceed the lesser of $2,000 or 100% of compensation for the taxable year. in addition, the maximum contribution permitted under a roth ira is phased-out from $2,000 to $0 for individuals earning above a certain level of adjusted gross income. in-90 msa contributions can be made by the employer or the employee, but not both in the same year. contributions made by the employer, within limits, are excluded from the employee’s compensation. contributions made by individuals, with limitations, are deductions from adjusted gross income. roth ira contributions are funded entirely by the employee and are not tax deductible. unlike regular iras, which do not allow contributions after age 701⁄2, the msa and roth ira do not put a maximum age on allowable contributions. an excess contribution to an msa occurs when the contributions exceed the deductible limits or are made for an ineligible person. any excess contribution made by the employer is included in the employee’s gross income, and the account holder is subject to a 6% excise tax on excess contributions for each year they are in an account. if the excess contributions are removed prior to the last day prescribed by law for filing an income tax return, the excise tax can be avoided. the treatment of excess contributions to a roth ira is essentially the same as the msa. 5.2. withdrawals regular iras require withdrawals to begin on the account no later than age 701⁄2. there is no such requirement with msas or roth-iras. qualified medical withdrawals are permitted tax free in msas for a wide range of medical expenses. qualified expenses are anything the internal revenue service allows for a medical tax deduction. in addition, premiums for cobra continuation coverage, long-term care insurance or services, and premiums for health insurance coverage while receiving unemployment compensation are also permitted tax free. nonqualified withdrawals before age 65 are subject to a 15% penalty and are included in gross income for taxation purposes. when the insured msa account holder reaches 65 years of age, withdrawals can be made with no penalty. any withdrawal from the roth ira that is not a qualified distribution is first considered a taxand penalty-free distribution of contributions. once an amount equaling the cumulative contributions to the roth has been recovered tax-free, all further distributions that are not qualified will be subject to ordinary income tax. a 10% penalty tax is also assessed if the owner is not at least 591⁄2 years old. a qualified distribution from a roth ira is both (1) made after a five-year holding periodand (2) described by any one of the following: (1) made on or after the date the account holder reaches age 591⁄2; (2) made to the designated beneficiary after the account holder’s death; (3) made because of the account holder’s permanent disability; (4) a “qualified first-time homebuyer” distribution. 116 j.t. query / financial services review 9 (2000) 107–123 5.3. rollovers/transfers a rollover from one msa to another will not trigger income tax or penalties if made within 60 days of distribution. also, one year must pass between tax-free rollovers. while there are a number of restrictions on the rollover of a traditional ira into a roth ira, the rollover of the same type of ira plans operates under the same basic rules as msas. transfers of msa accounts pursuant to a divorce or separation order are not taxable. if the account holder dies or becomes disabled before age 65 the penalty tax does not apply. if the msa account holder dies, the medical savings account is included in the gross estate for estate tax purposes. if the beneficiary is the surviving spouse, the msa belongs to the spouse and he or she can deduct the account balance in determining the account holder’s gross estate. the surviving spouse is then allowed to use the msa for his or her own medical expenses. if the beneficiary is anyone other than the spouse, or there is no beneficiary, the msa ceases to exist, and the beneficiary is required to include the fair market value of msa assets in gross income for the taxable year that includes the date of death. if eligible to receive an eligible rollover distribution from an employer’s qualified retirement plan resulting from divorce or similar proceedings, all or part of the distribution may be rolled over to an ira on a tax-free basis. a surviving spouse of a deceased employee may also be permitted to make a tax-free rollover contribution to a traditional ira. 5.4. taxation earnings in the msa fund grow tax-deferred, and are tax-free if used for qualified medical expenses. contributions up to the allowable amount are federally tax deductible if made by the employee. employer contributions are not included in the employee’s taxable income. earnings on roth ira contributions accumulate on a tax-deferred basis and may ultimately be tax-free if the earnings are part of a qualified distribution. a qualified distribution is generally a distribution made after age 591⁄2 and after the roth ira account is at least five years old. contributions to a roth ira are not deductible for federal income tax purposes. 5.5. summary while many similarities exist between the two types of accounts, there are some significant differences. the msa can be funded by either an employer or an employee, while the roth ira can only be funded by an employee. if used strictly for qualified medical expenses, the possibility exists for a tax-deductible msa contribution combined with tax-free earnings. the minimum age for penalty-free nonmedical withdrawals from the msa is 65, compared to a minimum age of 591⁄2 for roth iras. in addition, employee contributions to the roth ira can be withdrawn at any time subject to restrictions. under current law, the maximum amount that can be deposited into an msa is greater than the maximum roth contribution for an individual. msa contributions are made with before-tax dollars while roth ira contributions must be made with after-tax dollars. 117j.t. query / financial services review 9 (2000) 107–123 6. numerical analysis this paper uses a series of single point forecasts to illustrate the viability of medical savings-type accounts as a retirement savings alternative. to support the theoretical exposition of the medical savings account as a viable savings alternative, a numerical analysis is performed simulating real-world conditions. the model uses an assumed general inflation rate of 3.1% annually. the cost of health care claims increases at a 4.7% rate. the general inflation rate is based on ibbotson’s forecast for the period 1999–2025. the health care inflation rate is based on a historical comparison of the cpi (all items) with the cpi (medical care) and cpi (medical care services). account set-up or maintenance costs to the msa owner are assumed to be zero. in actual practice set-up fees range from around $0 to $19 for the leading msa custodians, and $0 to $5 for monthly maintenance fees. to ensure an acceptable level of liquidity, the first $1,000 deposited in the msa is left in a marketable securities fund invested in treasury bills earning 4.5%. the remainder of msa funds are invested in large cap stock fund with an average return of 11.6%. many existing msa plans offered by financial services companies require a minimum amount be maintained in a liquid money-market type of account or an interest-bearing demand deposit account. early versions of medical savings accounts limited investment of the savings portion in accounts paying a low fixed rate. due to the increase in account balances and development of the account over time, many financial services firms offer an array of investment choices. for example, the msa offered by mellon financial corporation allows balances in excess of $3,500 to be transferred to a dreyfus brokerage account. those funds can then be invested in a wide variety of mutual funds and equity investments. the rates of return on assets are based on ibbotson’s forecasted returns for stocks and bills for the period 1999–2025. the maximum allowable insurance deductible for the individual/family is assumed, and the maximum msa contribution allowed, 65/75% of the maximum allowable deductible, is deposited into the account. the maximum deductible allowed is assumed to increase at 3.1% per year. beginning after 1998, the high deductible dollar amounts are indexed for inflation in $50 increments based on the consumer price index. the individual/family income tax rate is assumed to be 28%. the individual/family is eligible for msa participation for the entire 12 months of each year modeled. the “low” “average” and “high” claims levels are initially set at $500, $1,000, and $2,000 respectively. these amounts are based on the survey referred to in the jensen and morlock paper. the 1989 survey of about 1 million individuals in large self-insured health plans (adjusted to 1994 dollars), found that about one-third filed no claims, 73% filed claims for less than $300, and 89% filed claims for less than $2,000. the cost of health care claims increases at an assumed rate of 4.7% annually. 7. results the results of the numerical analysis generally support the conclusions of previous simulations regarding the growth of funds in msa accounts. as shown in table 1, column 118 j.t. query / financial services review 9 (2000) 107–123 [a] the msa for the family would grow to a tax-deferred balance of $25,419 after five years and $156,099 after twenty years assuming no claims or that claims are paid from a non-msa source. these totals compare favorably with after-tax balances of $23,134 and $95,155 for the same time frames (column [b]). column [c] examines the effect of a relatively low level of claims ($500, indexed for inflation) withdrawn from the account annually. tax-deferred growth would be reduced to $21,992 and $109,182 for the five-year and twenty-year time horizons, respectively. at an “average” claim level of $1,000 annually (indexed for inflation), the msa account holder would end up with $18,565 and $62,266 remaining in the account after paying these claims. the balances remain positive even at a “high” claim level of $2,000 initially, then indexed for inflation. after subtracting the “high” claim payments from the msa, thousands of dollars would remain in the msa family account. as seen in table 2, the results are similar for the individual msa with respect to growth and no claims paid out of the account. tax-deferred growth would result in a five-year balance of $12,084 and a twenty-year balance of $156,099. the balances would remain positive with a “low” claims level and an “average” claims level as well. under the individual plan scenario, however, the fund balance in the plan could change drastically if claims begin to rise at a level approaching the $2,000 indexed threshold. table 1 ending balances in medical savings account for a family years in plan [a] tax-deferred growth (tdg) no claims [b] taxable growth no claims [c] tdg with low claims* [d] tdg with average claims* [e] tdg with high claims* 5 $ 25,419 $ 23,134 $ 21,992 $ 18,565 $ 11,710 10 $ 73,687 $ 61,528 $ 63,442 $ 53,198 $ 32,709 15 $ 162,213 $ 123,398 $ 139,053 $ 115,894 $ 69,576 20 $ 321,248 $ 221,073 $ 274,332 $ 227,416 $133,583 40 $3,480,546 $1,509,879 $2,941,673 $2,402,799 $325,052 * assumes no limit on out-of-pocket costs; in 1998 the opc was $5,500 for families. table 2 ending balances in medical savings account for an individual years in plan [a] tax-deferred growth (tdg) no claims [b] taxable growth no claims [c] tdg with low claims* [d] tdg with average claims* [e] tdg with high claims* 5 $ 12,084 $ 10,715 $ 8,657 $ 5,230 2$ 1,624 10 $ 35,537 $ 28,044 $ 25,293 $ 15,048 2$ 5,441 15 $ 78,621 $ 54,768 $ 55,462 $ 32,303 2$ 14,015 20 $ 156,099 $ 95,155 $ 109,182 $ 62,266 2$ 31,566 40 $1,697,192 $561,449 $1,158,319 $619,445 2$458,302 * assumes no limit on out-of-pocket costs; in 1998 the opc was $3,000 for individuals. 119j.t. query / financial services review 9 (2000) 107–123 column [e] of table 1 illustrates that at an initial “high” claim level of $2,000 annually and rising, the balance of the account would have a deficit after the first year. the msa limit for first year contributions would be around $1,508. this figure is based on a scenario of the maximum individual deductible of $2,250 increasing by a projected 3.1% c.p.i. to $2,320. based on this amount, the maximum msa funding allowable for that year would be 65% of $2,320, or $1,508. at the “high” level of claim activity the deficit would continue to grow and the account balance would stand at negative $14,015 after 20 years. the cumulative deficit would grow to negative $458,302 after 40 years. the effect of the different contribution maximums for individuals and families on plan balances is shown graphically. it should be noted that in most cases the level of claims for a family would, on average, be higher than the level of claims for an individual. to test for the robustness of the data presented, another forecast was made using an annual growth rate of six percentage for health care expenditures. this growth rate is closer to actual price increases for various components of health care during the 1990s. as expected, the increase in the rate of medical inflation decreases the balances accumulated. the results of the analysis using a higher growth rate for claims are shown in tables 3 and 4 below. this model admittedly contains limitations that obviously have a significant effect on the results. it should be noted that this model assumes no cap on out-of-pocket expenditures. under the 1996 health insurance portability and accountability act, there is an out-ofpocket cap of $3,000 and $5,500 on individual and family plans, respectively. such a cap would alter the results shown here dramatically. however, even with such a cap the difference would have to be covered by some entity, such as private insurance or government table 3 ending balances in medical savings account for a family (assuming annual growth rate of health care expenditures equals six percent) years in plan tdg with low claims* tdg with average claims* tdg with high claims* 5 $ 21,971 $ 18,406 $ 11,391 10 $ 63,366 $ 52,159 $ 30,626 15 $ 138,867 $ 112,383 $ 62,543 20 $ 273,937 $ 218,174 $115,085 40 $2,936,804 $2,224,307 $967,923 * assumes no limit on out-of-pocket costs; in 1998 the opc was $5,500 for families. table 4 ending balances in medical savings account for an individual (assuming annual growth rate of health care expenditures equals six percent) years in plan tdg with low claims* tdg with average claims* tdg with high claims* 5 $ 8,576 $ 5,069 2$ 1,947 10 $ 24,771 $ 14,004 2$ 7,529 15 $ 53,702 $ 28,782 2$ 21,057 20 $ 104,554 $ 53,009 2$ 50,081 40 $1,069,000 $440,808 2$815,575 * assumes no limit on out-of-pocket costs; in 1998 the opc was $3,000 for individuals. 120 j.t. query / financial services review 9 (2000) 107–123 subsidized health care coverage. in either case the costs would be significant, resulting in either rising insurance premiums or higher government expenditures. in addition, the model assumes that claims are level (adjusted for inflation) throughout one’s lifetime. as evidenced by a watson wyatt worldwide study (1996) this is not indicative of a person’s medical expenditures during a typical life cycle. the average per capita health expenditure by age group in 1995 indicates an increasing level of spending dependent upon age group. from ages 5–44, the average per capita health expenditure was below $2,000; from ages 45–64 the average was between $3,000 and $6,000; and from ages 65–89 the average was between $8,000 and $13,000. under this scenario, it might be possible for the individual msa account to accumulate enough capital in the early years to offset the effect of large withdrawals to pay claims in the later years. 8. conclusion this paper examines the feasibility of medical savings accounts as a practical savings opportunity. the numerical analysis is a refinement of previous simulations, as described in the literature review, in a number of areas. actual forecasted rates of inflation, stocks, and bills for the years 1999–2025 are obtained from ibbotson and associates. taxable and tax-deferred growth are computed and contrasted. the composition of assumed investments in the msa more closely emulate actual practice. claims scenarios are based on an extensive national bureau of economic research study. the maximum allowable contribution is indexed in order to align the amounts more closely with the parameters found in the 1996 hipaa pilot program. a comparison of the msa and the roth ira finds advantages and disadvantages in each plan. strategies using a combination of a medical savings-type account and a retirement account may optimize the benefits available to qualified taxpayers. to the vast majority of individuals, the medical savings account represents a viable personal savings vehicle. the 90% of the population that incur only one-fourth of the total u.s. health care expenditures would end up with a significant account balance upon retirement. for the ten percentage of the u.s. population that incur three-fourths of the total u.s. health care expenditures, the medical savings account is not a practicable retirement funding alternative. under it’s present structure, the msa should not be considered as a complete replacement for other retirement savings vehicles such as the 401(k) and the various iras available. future health status is the primary determinant of the success of the msa as an effective savings tool. the uncertainty of an individual’s future health expenditures makes the medical savings account an “all-or-nothing” proposal in some respects. most objective analyses acknowledge that medical savings accounts have both pros and cons. however, the debate continues on whether msas would provide a net benefit or burden to the existing health care system. the argument for a medical savings-type account or a “super ira” that combines attributes of the msa and the roth ira is compelling. this is especially critical to those within twenty-five years of retiring, as a means for “catching up” to acceptable savings levels for a comfortable retirement. nevertheless, the challenge will be 121j.t. query / financial services review 9 (2000) 107–123 to insure that the unhealthy ten percentage are not left behind or remain in a position that burdens the entire health care system. references american academy of actuaries. (1995). medical savings accounts: cost implications and design issues. washington, d.c. public policy department of the american academy of actuaries. anon. (1996). real health-care reform.investor’s business daily,october, 16. barchet, s., anderson, j., & chapman, l. (1995). medical savings accounts: an option to reduce health care costs and increase health care satisfaction.aca journal, 4,34–47. bajtelsmit, v. l. (1996). conservative pension investment: how much difference does it make?benefits quarterly, 12,35–39. beam, b. t. jr., & tacchino, k. b. (1997). medical savings accounts.journal of the american society of clu & chfu, 51, 8–12. berk, m. l., & monheit, a. c. (1992). the concentration of health expenditures: an update.health affairs, 11, 145–149. bond, m. t., heshizer, b. p., & hrivnak, m. w. (1996). medical savings accounts: why do they work?benefits quarterly, 12,78–83. bond, m. t., heshizer, b. p., & hrivnak, m. w. (1997). medical savings accounts: a health insurance option for the public sector?public personnel management, 28,535–544. bradley, c. f., & dalton, m. a. (1997). medical savings accounts: a critical analysis.journal of financial planning, 10,72–77. burry, j. (1994). medical savings accounts: bad medicine for the u.s. health care system. blue cross/blue shield of ohio. doerpinghaus, h. (1996). managing medical savings accounts.benefits quarterly, 12,69–77. heffley, d. r., & miceli, t. j. (1998). the economics of incentive-based health care plans.journal of risk and insurance, 65,445–465. jensen, g. a., & morlock, r. j. (1994). why medical savings accounts deserve a closer look.journal of american health policy,may/june. kaiser-harvard program on the public and health/social policy. (1996). survey of americans on health policy. questionnaire and national toplines –july 30, 1996. harvard school of public health, boston, ma and kaiser family foundation, menlo park, ca. keeler, e. b., malkin, j. d., goldman, d. p., & buchanan, j. l. (1996). can medical savings accounts for the nonelderly reduce health care costs?journal of the american medical association, 275,1666–1671. kpmg peat marwick llp. (1995).retirement benefits in the 1990s: 1995 survey data.new york, ny. kpmg. luntz research companies. january. (1995). massaro, t. a., & wong, y. n. (1996). medical savings accounts: the singapore experience. ncpa policy report no. (203). april,1996. mcdevitt, r. d., & schieber, s. j. (1996).from baby boom to elder boom, providing health care for an aging population.washington, d.c.: watson wyatt worldwide. merchant, t., & rusk, m. (1997). medical savings accounts: counting on (and counting up!) the benefits.credit world, 86,13–15. national center for policy analysis. (1994). medical savings accounts: the private sector already has them. ncpa brief analysis no. 105 (20 april). newhouse, j. and the insurance experiment group. (1993).free for all: lessons from the rand health insurance experiment.cambridge, ma: harvard university press. ozanne, l. (1996). how do medical savings accounts affect medical spending?inquiry, 33,225–236. panko, r. (1997). five msa sellers, five strategies.best’s review (life/health), 98,60–61. pauly, m. v. (1994). an analysis of medical savings accounts: do two wrongs make a right? a paper delivered at the american enterprise institute, april 18. 122 j.t. query / financial services review 9 (2000) 107–123 robbins, g., robbins, a., & goodman, j. c. (1994). inefficiency in the u.s. health care system: what can we do? national center for policy analysis, ncpa policy report no. 182, april. saunders, l. (1997). psst! super-ira.forbes, 159,170–172. scandlen, g. (1998). medical savings accounts: obstacles to their growth and ways to improve them. national center for policy analysis, dallas, tx, policy study no. 216. swartz, k. (1996). medical savings accounts and research.inquiry, 33,216–219. williamson, c. (1999). ibbotson forecasts dow at 120,000 within 25 years.pensions and investments,april 5, (1999), 1,46. zabinski, d., selden, t. m., moeller, j. f., & banthin, j. s. (1999). medical savings accounts: micro simulation results from a model with adverse selection.journal of health economics, 18,195–218. 123j.t. query / financial services review 9 (2000) 107–123 pii: 1057-0810(93)90006-c financial services review, 3(l): 5%73 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the individual’s tax-exempt bond portfolio decision under income uncertainty amy v. puelz in this article, an individual’s tax-exempt bond porrfolio decision is investigated. a model capturing the relationship between income uncertainty and optimal portfolio choice is defined when an individual decision-maker has the opportunity to hold higher yielding private-activity bonds. the findings in this article show that in most cases risk-averse individuals will maximize the expected utility of after-tax income by holding a large proportion of private-activity bonds in their portfolio even under income uncertainty and the risk of a minimum tax liability. those individuals who would benefitfrom holding private-activity bonds in a tax-exempt portfolio are identified and the magnitude of the benefit is quantified, i. intr~duc~~n the tax reform act of 1986 created two general classes of municipal investments, essential-purpose and nonessential-purpose or private-activity bonds. essential purpose bonds yield interest income that is exempt from any form of federal income tax. however, interest income on private-activity bonds issued after august 7, 1986 is classified by the internal revenue service as preference income and is exempt from federal income tax only if an individual is not subject to the alternative minimum tax.’ the alternative minimum tax (or minimum tax) is a flat tax rate applied to minimum taxable income which is the total of regular taxable income plus preference income and adjustments. an individual is subject to the minimum tax if the minimum tax liability exceeds the regular tax liability.* the probability of an individual being subject to minimum tax increased with 1990 tax reform when the minimum tax rate was increased from 21% to 24%. the separation of municipal bonds into two distinct classes alters individual decision-making regarding the tax-exempt portfolio because of the relatively high yields and the uncertain tax treatment of private-activity bonds. in this article, the amy v. helz l department of management information sciences, edwin l. cox school of business, southern methodist university, dallas, texas, 75275. 60 financialservicesreview,3(1) 1993 tax-exempt portfolio decision under income uncertainty is examined given the expanded choice set that includes private-activity bonds. the tax-exempt bond literature has focused on both sides of the market. supply-side research has dealt with the creation and issuance of municipal debt and includes such topics as structuring bond issues (cohen & hammer, 1966; puelz & lee, 1992); insurance and signalling (kidwell, sorensen, & wachowicz, 1987; hsueh & liu, 1990; puelz, 1991); and underwriter bids (bierwag, 1976; braswell, stunners, 1982; nauss, 1987). demand-side research has dealt with topics such as market segmentation (fischer, 1980; kidwell & koch, 1983); market inefficiencies (speer, 1987; kochin & parks, 1988); risk premiums (mcinish, 1980; gehrlein & mcinish, 1985); and the portfolio decision relative to tax-exempt bonds. it is this last topic that is addressed in this article. several authors have written about the descriptive charac teristics of private-activity bonds and the minimum tax (petersen, 1987, 1988; aalberts & utley, 1988; brown, 1988; porterba, 1989; bettner, 1990; day, 1991). however, there has been no rigorous analysis of the portfolio decision given the new environment of post-1986 tax reform, although the portfolio decision has been addressed in the context of income certainty (puelz & puelz, 1991). this article provides a more general approach by examining individual decision-making under uncertain income and uncertain after-tax retums.3 the institutional literature addressing the allocation of private-activity bonds to a portfolio suggests that an individual who is uncertain as to their tax status should reduce or eliminate their holdings of private-activity bonds to avoid triggering the minimum tax (brown, 1988; bettner, 1990; hoffman, smith, willis, & raabe, 1991). the results in this article show, contrary to the conventional wisdom, that in many cases a risk-averse individual will maximize the expected utility of after-tax income by holding private-activity bonds in their portfolio even under income uncertainty and the risk of a minimum tax liability. through a simulation procedure, those individuals who would benefit from holding private-activity bonds in a tax-exempt portfolio are identified and the magnitude of the benefit quantified. in the next section of the article the model of individual portfolio choice is developed. this is followed by a comparison of the utility maximizing portfolio allocation derived from the model presented in this article and the naive portfolio allocation of 100% to essential-purpose bonds for different individual and market charac teristics. finally, the relationship between income uncertainty and optimal portfolio allocation is explored for different income levels, bond yield differentials, and portfolio sizes. themodel as a starting point consider the simple case of an individual with certain income who wants to select the bond portfolio that maximizes after-tax income. the decision considered is the proportion of investable wealth to allocate to private activity bonds, with the remainder of investable wealth allocated to essential-pur individd tax-exempt bond p@iiilia 61 pose bonds. to facilitate the comp~son, the bonds mature in one time period, are not sold short, and in all other aspects are identical except for their tax treatment and yields.4 the allocation in the certain income case is straight forward. an individual who is not subject to the minimum tax will allocate 100% to private-ac tivity bonds because of their associated higher yields. an individual who is subject to the minimum tax because of preference income (other than private-activity bond income) or adjustments will allocate 100% to essential-purpose bonds.5 however, an individual whose preference income earned on private-activity bonds could trigger a minimum tax liability will allocate a proportion to private-activity bonds such that the minimum tax liability equals the regular tax liability.6 now consider a more general model that specifies an individu~s ~~e~~n after-tax income, 2, a fiction of ocean pre-tax income and uncertain tax liability, as where, iv= cy.= l-a= b= r, = rp = t, = z{lv) = %m = %m = p’ uncertain regular taxable adjusted gross income (dollars), propo~ion of the tax-exempt portfolio allocated to p~vate-activity bonds, proportion of the tax-exempt portfolio allocated to essential-purpose bonds, wealth allocated to tax-exempt bond portfolio (dollars)7, yield on essential-purpose bonds, yield on private-activity bonds, alternative minimum tax rate, regular tax rate (a function of es), regular taxable income exemption (a function of $j), alternative ~nimum taxable income exemption (a function of i$, proportion of income from preference items other that p~vate-activity bonds and minimum taxable income adjustments. the first term, fl+ (1 a)bz?, + abrp -t pn, represents pre-tax income. the maxi mand function is the tax liability. within this maximand an individual pays the maximum of the regular tax liability, the left hand side, or the minimum tax liability, the right hand side. the exemption and the tax functions are described in appendix a. the optimal allocation to private-activity bonds under income uncertainty is a function of the ~~ationship of the after-tax returns, and the individu~‘s anacin characteristics and risk preferences. an expected utility m~imizing individu~ will 62 financial services review, 3(l) 1993 choose the proportion of the portfolio allocated to private-activity bonds, a, that satisfies the first-order condition z[ flx]l/aa = 0. since (1) is a non-differentiable, non-continuous function, simulation is employed to derive the optimal expected utility maximizing allocation to private-activity bonds, a’. comparative analysis in this section, the model derived in the previous section is simulated to derive optimal utility maximizing portfolios containing private-activity bonds and essen tial-purpose bonds. model parameters are varied over reasonable ranges and the optimal portfolio strategies are compared to naive strategies of portfolios containing only essential-purpose bonds. after-tax income for this comparative analysis is calculated under the assumption of joint-filing status by a married couple. house hold income, for which one individual acts as decision-maker, is assumed to follow a pareto distribution (quandt, 1966). income uncertainty is measured by the dispersion factor (df). income certainty corresponds to a df of one and higher income uncertainty corresponds to higher values of df.8 in addition, risk prefer ences of the individual decision-maker are characterized by a function displaying decreasing absolute risk aversion, u{ x} = log{ x}. reasonable ranges for p and b were determined to be from 0.15 to 0.25 and from 0.5 to 1.5 respectively. these ranges are based on alternative minimum tax computations from sample income tax returns (day, 1991, p. 22). the simulation steps are presented in detail in appendix a. in table 1 the portfolio decisions are listed when income uncertainty is relatively low (df is 1.05). the first two columns of numbers under each income category indicate the expected utility maximizing percentage of the portfolio allocated to private-activity bonds. the next two columns of numbers in each income category are the estimated mean difference between the after-tax income when the combined utility maximizing portfolio is selected (i,) and when the pure essential-purpose portfolio is selected (i,). the numbers not in parenthesis are those derived when the spread between private-activity and essential-purpose bonds is 20 basis points and those numbers in parenthesis are for when the spread is 70 basis points.’ in all cases where the estimated mean is reported, the paired t-test of the alternative hypothesis ha: z, z, > 0 was significant at the 0.001 level. in almost all cases, except when p and b are both relatively high, individuals with household median income of $150,000 allocate close to 100% to private activity bonds. this is because the risk of minimum tax is low and therefore there is a high likelihood of realizing the additional return on private-activity bonds. as median income increases, the allocation to private-activity bonds falls rapidly with increasing levels of p and b because of the higher probability of a minimum tax liability. however, when p is 20% or less, the utility maximizing portfolio contains a portion of private-activity bonds. in addition, private-activity bonds are only eliminated from the utility maximizing portfolio for median incomes and p levels t a b l e 1 . !? a c om pa ri so n of t he e xp ec te d a ft er -t ax i nc om es o f th e c om bi ne d e xp ec te d u ti lit y m ax im iz in g p or tf ol io ( z e) an d th e p or tf ol io c on ta in in g o ni y e ss en ti al -p ur po se b on ds ( z e) w he n th e d is pe rs io n f ac to r of i nc om e is s et a t 1. 05 . 3 g p, p ro po rt io n of i nc om e m ed ia n in co m e = is o ,@ m ed ia n in co m e = 20 0, oo o m ed ia n in co m e = 30 0, oo o fr om p re fe re nc e he m s i (o th er r ha n p ri va te -a & v b , t ax -e xe rn ft b on d p or tf o y a* , o ph al a* , o pr im al a* , o pr im al it y b on ds ) an d a dj w t a & xa do n to a & nx zd on ~ fq a ll oc ar io n to si ze $ e& y; dr m p nw zv yc ty es ti m at ed m ea n m en ts p nv ;t ac cy ty es ti m at ed m ea n p ri va te -a ct iv it y ic ic i, i, b on ds ( % ) 15 % 50 % lo o (1 00 ) 15 0 (5 25 ) 10 0 (1 00 ) 20 0 (7 00 ) 10 0 (1 00 ) 10 0% lo o (1 00 ) 30 0 (1 05 0) 10 0 (1 00 ) 40 0 (1 40 0) 98 ( 92 ) 58 8 (1 92 3) 15 0% 10 0 (1 00 ) 45 0 (1 57 5) 82 ( 79 ) 47 9 (1 61 4) 65 ( 61 ) 58 5 (1 92 2) 20 % 50 % 10 0 (1 00 ) 15 0 (5 25 ) 92 ( 99 ) 16 9 (6 19 ) 53 ( 50 ) 15 9 (5 18 ) 10 0% 10 0 (1 00 ) 30 0 (1 05 0) 46 ( 50 ) 16 5 (6 21 ) 26 ( 25 ) 15 6 (5 18 ) 15 0% 83 ( 82 ) 36 1 (1 22 7) 31 ( 33 ) 16 3 (6 19 ) 17 (1 7) 15 3 (5 16 ) 25 % 50 % 95 ( 10 0) 12 5 (4 83 ) 0 (0 ) () 0 (0 ) () 10 0% 47 ( 54 ) 11 9 (4 79 ) 0 (0 ) () 0 (0 ) q ;i ; 15 0% 32 ( 36 ) 12 0 (4 86 ) 0 (0 ) () 0 (0 ) n ot e: t he n um be rs n ot in p ar en th es es a re w he n th e yi el d di ff er en tia l is 2 0 ba si s po in ts a nd th os e nu m be rs in p ar en th es es a re w he n th e yi el d di ff er en tia l is 7 0 ba si s po in ts . t he e ss en tia lpu rp os e bo nd y ie ld i s 6. 73 % . in a ll ca se s w he re t he e st im at ed m ea n is r ep or te d th e pva lu e fr om t he p ai re d tte st o f th e al te m at iv e hy po th es is h a: i, i r > 0 w as le ss t ha n 0. 00 1. t he b on ds a re id en tic al w ith th e ex ce pt io n of y ie ld s an d ta x tr ea tm en t. a ft er -t ax i nc om e is c al cu la te d as su m in g a jo in t fi lin g st at us fo r a m ar ri ed c ou pl e. t he ri sk p re fe re nc es o f th e in di vi du al d ec is io nm ak er a re c ha ra ct er iz ed b y de cr ea si ng a bs ol ut e ri sk a ve rs io n, u {x ) = l q g {x ). t a b l e 2 . a c om pa ri so n of t he e xp ec te d a ft er -t ax i nc om es o f th e c om bi ne d e xp ec te d u ti lit y m ax im iz in g p or tf ol io ( i, ) an d th e p or tf ol io c on ta in in g o nl y e ss en ti al -p ur po se b on ds ( i, ) w he n th e d is pe rs io n f ac to r of i nc om e is s et a t 1. 35 . p, p ro po rt io n of in co m e m ed ia n in co m e = 15 0, oo o m ed ia n in co m e = 20 0, oo o m ed ia n in co m e = 30 0, 00 0 fr om p re fe re nc e it em s b , ta v. ex et n t b on d p or tf o f c l, op tim al (o th er t & z p ri va te -a ct iv 19 c & e $ ~~ m $~ dl a, , a ll oc at io n, f a: o pt im a1 a, op t& al it y b on ak ) a nd a dj us tm en ts p ri i;~ ~~ tt y 0 es ti m pt ed lm em a & xa ti on . to ls o ’& p ng tg lc cu y es ti ye ed lm em a hc at io n. to p ri v ~i v ~t l~ ty es ti yt 5m er m 15 % 50 % . 10 0 (1 00 ) 10 0 (1 00 ) 20 6 $0 ) lo o ( 10 0) 30 & 1~ 50 ) 10 0% lo o (1 00 ) 30 0 (1 05 0) 10 0 (1 00 ) 40 0 (1 40 0) 7l e6 ) 41 5 (1 37 5) 15 0% 10 0 (1 00 ) 45 0 (1 57 5) 74 ( 80 ) 41 6 (1 48 5) 47 ( 44 ) 41 3 (1 37 1) 20 % 50 % lo o (l o 0) 15 0 (5 25 ) 46 ( 95 ) 58 ( 34 6) 2l w ) 50 ( 21 9) 10 0% 72 ( 10 0) 10 5 (1 05 0) 23 ( 49 ) 49 ( 34 7) 11 (1 2) 48 ( 21 1) 15 0% 47 ( 75 ) 83 ( 68 2) 15 (3 3) 56 ( 36 0) 7 (8 ) 50 ( 21 8) 25 % 50 % 0 (7 9) (1 7) 0 (0 ) cd 0 (0 ) 10 0% 0 (4 0) (6 8) 0 (0 ) 0 (0 ) i; -; 15 0% 0 (2 6) (8 ) 0 (0 ) r; -; 0 (0 ) -t -1 n ot e: t he n um be rs n ot in p ar en th es es a re w he n th e yi el d di ff er en tia l is 2 0 ba si s po in ts a nd th os e nu m be rs in p ar en th es es a re w he n th e yi el d di ff er en tia l is 7 0 ba si s po in ts . t he e ss en tia lpu rp os e bo nd y ie ld i s 6. 73 % . i n al l c as es w he re t he e st im at ed m ea n is r ep or te d tb e pva lu e fr om t be p ai re d tte st o f th e al te rn at iv e hy po th es is h a: ic ie > 0 w as le ss th an 0 .0 01 . t he b on ds a re id en tic al w ith th e ex ce pt io n of y ie ld s an d ta x tr ea tm en t. a ft er -t ax i nc om e is c al cu la te d as su m in g a j oi nt f ili ng s ta tu s fo r a m ar ri ed c ou pl e. t he r is k pr ef er en ce s of th e in di vi du al d ec is io nm ak er a re c ha ra ct er iz ed b y de cr ea si ng a bs ol ut e ri sk a ve rs io n, u (x ) = l o g (x ). individual tax-exempt bond pot$folio 65 above $2~,~ and 25% respectively. comp~ng the decision when the spread between p~vate-activity and essential-pu~ose bonds is 20 basis points to the decision when the spread is 70 basis points there is relatively little change in d across all income levels. individuals with household median incomes of $150,000 and 25% of income from preference items slightly increase their holding of private activity bonds as the yield spread increases because they can capture the additional return without significantly increasing their risk of a minimum tax liability. in contrast, individuals with a greater risk of a minimum tax liability do not shift to private-activity bonds as the yield spread increases. the next set of comparisons is identical to those presented in table 1, except the unce~ainty associated with income, df, is increased to 1.35 from 1.05. the results are presented in table 2. in all cases where estimated means are reported, the paired t-test of the alternative hypothesis el,: 1, 1, > 0 is significant at the 0.001 level. the effect of greater uncertainty is consistent among all individuals in that private-activity bond holdings are more rapidly eliminated from the portfolio with increasing probability of minimum tax (i.e., increasing p and b). however, the only cases where individuals completely eliminate private-activity bonds from their portfolio are when p is 25% or more. individuals with household median income of $150,000 continue to hold private-activity bonds even at high levels of p and b. in most instances, the optimal proportion of private-activity bonds held is greater when the yield spread between private-activi~ and essential-pu~ose bonds is 70 basis points as opposed to 20 basis points. the fact that in some cases the proportion of private-activity bonds held in the portfolio drops as the yield on these bonds increases is due to the fact that the higher yield results in a greater risk of minimum tax liability. hence, the risk-averse investor may actually reduce their holding of private-activity as the yield increases if the additional risk is too high. the yield spread is a much more significant factor in the allocation decision when income uncertainty is high. in summary, this comparative analysis illustrates the relationship between the individual’s vulnerability to the minimum tax, household income uncertainty and the market yields on private-activity bonds relative to essenti~-pu~ose bonds. those individuals with household median incomes below $150,~ and those with low levels of preference income (other than private-activity bond income) and/or a small tax-exempt portfolios, should hold a large proportion if not all of their tax-exempt portfolio in private-activity bonds. in addition, the greater the uncer tainty of income the greater the impact of market yield spreads on the allocation decision. prxvate-activwy bond allocation in this section the effect of ~ce~ainty on the optimal allocation to private ~tivity bonds is presented. as in the previous section, after-tax income is calculated under 66 financial services review, 3(l) 1993 the assumptions of joint filing status by a married couple and income following a pareto distribution. risk preferences of the individual decision-maker are again characterized by decreasing absolute risk aversion. the allocation decision is presented for different median income levels, different bond yield spreads, and different tax-exempt portfolio sizes. optimal private-activity bond allocation relative to income the first set of simulation results compares the optimal expected utility maximizing portfolio for different median income levels. tax-exempt portfolio size (b) is assumed to be 100% of median income and proportion of income from preference items other than private-activity bond income plus adjustments (p) is 20%. the private-activity bond yield for the comparison is set at 6.93% or 20 basis points greater than the essential-purpose yield of 6.73.” median income is varied from $150,000 to $400,000. the optimal allocations (a*) are presented in figure 1. first consider the optimal portfolio for each individual under household income certainty (of = 1). an individual with household income of $150,000 will not be subject to the minimum tax regardless of the allocation to private-activity bonds and will therefore allocate 100% of the portfolio to private-activity bonds. individuals with household incomes of $200,000 and $300,000 will allocate a portion of the portfolio to private-activity bonds such that the regular tax liability equals the minimum tax liability. the individual with household income of $300,000 as compared to the individual with household income of $200,000 has a higher effective regular tax rate but also has a significantly higher effective mini mum tax rate and therefore allocates a smaller portion (27% as opposed to 57% for incomes of $200,000) of the portfolio to private-activity bonds. the higher effective minimum tax rate is due to the minimum taxable income exemption phaseout that occurs for high income households. for example, a married couple filing a joint return will have a $40,000 minimum taxable income exemption that is phased out at a rate of 25% for every dollar minimum taxable income exceeds $150,000. the $40,000 exemption is phased out completely at an income of $310,000. the individual with household income of $400,000 as compared to the individual household income of $300,000 has a higher effective regular tax rate and virtually the same minimum tax rate and therefore allocates a larger portion of the portfolio (53% as opposed to 27% for the income of $300,000) to private-ac tivity bonds. now consider the change in ct* relative to income uncertainty. individuals with household median income of $150,000 have a very low probability of being subject to the minimum tax and therefore hold 100% private-activity bonds until df is greater than 1.16. as df increases above 1.16 the risk of a minimum tax becomes significant enough to induce the individual to reduce their holding of private-activity bonds. iniiividunl tar-exempt bond por@oi?o 67 d op allocation to prlvate aetivity 602 i i 0% i i 1 , 1 1.05 1.1 1.15 12 125 df dispersion factor n $1se,ooo + $2oomo 0 t=omo a @oqooo figure 1. the optimal allocation to ovate-activi~ bonds (a*) relative to the dispersion factor of income (d8’) for different median incomes. note: income follows a pareto distribution and after-tax income is calculated assuming a joint filing status for a married couple. the risk preferences of the individual decision-maker are characterized by decreasing absolute risk aversion. investable wealth (b) is 100% of median income, the proportion of income from preference items (other that private-activity bond income) and adjustments (p) totals 20%. the bonds mature in one time period, are not sold short, and are in all other aspects identical except for their tax treatment and yields. the yield of the private-activity bond and the essential-pur pose bond are 6.93% and 6.73% respectively. individu~s with hou~hold medii incomes of $2~,~ will experience a higher effective ~irn~ tax rate under higher levels of ~~~~nty because of the minimum taxable income exemption phaseout that occurs in this example be tween $150,000 and $310,000. therefore the optimal holding of private-activity bonds (a*) decreases with increasing uncertainty (of’). the optimal allocation to private-activity bonds relative to uncertainty levels off at high levels of uncertainty because the probability of income falling above the upper limit of the phaseout range increases and the effective minimum tax rate is constant in income above this upper limit. individuals with household median income levels of both $3~,~ and $4~,~ will slightly reduce a* as df increases because at these income levels the effective minimum tax rate is virtually constant in income. 68 financial services review, 3(l) 1993 100% a’ 90% t 0pti-l 60% allocation 70!4 to private1 607: 507. 40x 307. 207; 10x o%l ’ i i i i i i 1.05 1.1 1.15 12 125 df dispersion factor 0 6.637: + 7.037: 0 7232 a 7.437. figure 2. the optimal allocation to private-activity bonds (a*) relative to the dispersion factor of income (of) for different private-activity bond yields. note: income follows a pareto distribution and after-tax income is calculated assuming a joint filing status for a married couple. median income is $200,000. the risk preferences of the individual decision-maker are characterized by decreasing absolute risk aversion. investable wealth (b) is 100% of median income. the proportion of income from preference items (other that private-activity bond income) and adjustments @) totals 20%. the bonds mature in one time period, are not sold short, and are in all other aspects identical except for their tax treatment and yields. the yield of the essential purpose bond is 6.73%. to summarize the results presented in figure 1, individuals with household median income levels below $150,000 will typically maximize expected utility by holding 100% private-activity bonds. however, if income is highly variable (in this example a df of 1.16 for a median income of $150,000) optimal private-activity bond holdings will fall below 100%. individuals with median incomes near the lower limit of the minimum taxable income exemption phaseout range will reduce private-activity bond holding at a faster rate with income uncertainty than individu als with incomes near or above the upper limit minimum taxable income exemption phaseout range. individual tax-exempt bond porlfolio 69 this example illustrates, as one might expect, that the portfolio decision is greatly influenced by the median income level. it is interesting to note that only individuals with relatively high income levels combined with high levels of income uncertainty will hold less that one-half of their tax-exempt portfolio in private activity bonds. hence, private-activity bonds should not be arbitrarily eliminated from the tax-exempt portfolio if the individual is at risk of being subject to the minimum tax. optimal private-activity allocation relative to bond yields the relationship between p~vate-activi~ bond yields and essential-purpose bond yields is de~ndent on the characteristics of the bonds being compared. for purposes of illustrating the effect of the yield differential on the portfolio allocation decision the essential-purpose bond portfolio yield is held constant at 6.73% and the private-activity bond portfolio yield is varied from a low level to a high level. the assumptions from the previous example hold except private-activity bond yield is varied from 6.83% to 7.43%, and median household income is $200,000. the simulation results are presented in figure 2. when the private-activity yield is relatively low at 6.83%, a* is 58% under income certainty (df = 1). as the yield on private-activity bonds increases the individual under household income certainty holds a smaller proportion of private activity bonds in the portfolio. al~ough this may seem~ounter-intuitive, d is lower for higher yields under certainty in order for the portfolio to satisfy the after-tax maximizing condition that the regular tax liability equals the minimum tax liability. however, &x*ldf is lower for lower private-activity yields. this means a* falls at a more rapid rate with increasing uncertainty the lower the yield on private-activity bonds. when the yield on private-activity bonds is high relative to essential-purpose bonds, in this example 7.43%, the proportion of private-activity bonds held in the portfolio changes very little with uncertainty. optimal private-activity allocation relative to the percentage allocated to tax-exempt bond portfolio the final set of simulations examines the allocation decision relative to the size of the tax-exempt bond portfolio. the amount invested in the tax-exempt bond portfolio will influence the allocation decision in that the larger the portfolio, all else equal, the greater the probability the individual’s household will be subject to the minimum tax.” the assumptions made are those described in the first two examples except bond portfolio size, b, is varied from 50% to 150% of median income which is set at $200,000. the results are presented in figure 3. as one would expect, the larger the size of the tax-exempt portfolio the smaller the allocation to private-activity bonds. however, &x*&w is lower for low levels of b. when b is 150% of median income, a* is almost constant in uncertainty. 70 financial services review, 3(l) 1993 a* opm itlknxtioii to private activity 100x 90x box 70x 6oz 507. 407. 307; 207. 105: i i i i i 1 1.05 1.1 1.15 12 125 df dispersion factor 0 507. of med. income + 100x of med. income 0 1507. of med. income figure 3. the optimal allocation to private-activity bonds (a*) relative to the dispersion factor of income (of) for different allocations to the tax-exempt bond portfolio size. note: income follows a pareto distribution and after-tax income is calculated assuming a joint tiling status for a married couple. median income is $200,000. the risk preferences of the individual decision-maker are characterized by decreasing absolute risk aversion, the proportion of income from preference items (other that private-activity bond income) and adjustments (p) totals 20%. the bonds mature in one time period, are not sold short, and are in all other aspects identical except for their tax treatment and yields. the yield of the private-activity bond and the essential-purpose bond are 6.93% and 6.73% respectively. conclusion in this article, the tax-exempt bond portfolio decision is explored when an individual has the option of purchasing private-activity bonds. it is shown that the introduction of private-activity bonds affects bond portfolio decision-making when an individ ual’s household is subject to uncertain taxable income that may result in a minimum tax liability. the results in this article show that many individuals will maximize the expected utility of after-tax income by holding a large portion of private-activity bonds in their tax-exempt portfolio even when faced with the risk of a minimum tax liability. when applied in a portfolio planning framework, significantly greater after tax income may be realized if private-activity bonds are included in the tax-exempt portfolio. through simulation, the magnitude of the benefit derived from holding lndh&hal tar-exempt bond portfolio 71 p~vat~-activity bonds is quoting relative to indi~du~ household and market characteristics. median income level, yield d~fe~nti~, and tax-exempt bond port folio size are all shown to have a significant impact on the portfolio decision relative to income uncertainty. appendix a simulation steps the simulation steps 1 through 3 are repeated 10,000 times. the 01 that yields the greatest average utility is selected as the optimal cc*. steps: 1. income (n) is generated by approximating a pareto distribution. this is accomplished by generating a lognormal random variable with a median (m) and dispersion factor (of’) and setting all values less than the mode of the lognormal variable equal to the mode (sachs, 1982. p. 111). 2. after-tax income (x) for all ci = 0.0 to 1.0 (in steps of 0.01) is derived by x=n+(l -a)br,+abr,+pn where, e,(n) is $40,000 if n is less than $150,000 or the maximum of $0 or $40,000 minus 25% of the difference between the minimum taxable income and $150,000, e,(n) is $5300 reduced by 2% for each $2500 (or fraction of) that adjusted gross income exceeds $150,000, z{ .} follows the 1991 federal income tax schedule yi for a married couple filing a joint return. 3. the utility of each after-tax income is set equal to u{x} = log(x). (a-2) notes 1. some private-activity bonds are qualified as tax-exempt under the tax code section 501(c)(3). 2. for a detailed discussion of private-activity bonds the reader is referred to brown (1988), petersen (1988), bettner (1990), or day (1991). 3. piros (1987) provides a model of individual choice under uncertainty relative to taxable and tax-exempt bonds. although this article focuses on essential-purpose and private-activity bonds, the analysis is similar in that uncertain income results in uncertain tax treatment. however, unlike the analysis by piros, the mi~mum tax function is noncontinuous necessitating the use 72 financial services review, 3(l) 1993 of simulation. the security portfolio choice decision given various sources of uncertainty has been extensively addressed in the literature (i.e., kwan & yip (1987), kwan (1988), chamber lain & cheung (1990)). 4. this simplifying comparison allows the focus of this article to be on how the allocation to private-activity bonds affects individual utility. it is not the purpose of this article to address allocation within private-activity bonds or, for that matter, within essential-purpose bonds. however, a simulation approach similar to this could be employed as a decision-making tool to address the time dependent portfolio allocation problem. 5. the relationship of after-tax returns in an efficient market where individual income is certain is rt, > re > (1 t,)r, where r, is the yield on essential-purpose bonds, rp is the yield on private-activity bonds, and tq is the minimum tax rate. refer to puelz and puelz (1991) for a detailed discussion of the relationship of after-tax returns given income certainty. 6. this optimality condition under certainty is illustrated in puelz and puelz (1991). 7. it is assumed that the investor has made all other portfolio decisions, e.g., the stock portfolio, so the only investment under consideration is the net investable wealth in the tax-exempt portfolio. 8. in the pareto distribution the income range from median income divided by df to median income multiplied by df contains 68 percent of the income distribution. for example, if an individual’s median income and dispersion factor are $100,000 and 1.5 respectively then the probability of income falling between $66,667 and $150,000 is 0.68. 9. the average private-activity and essential-purpose bond yields on insured bonds offered in the blue list september 1, 1989 were 6.93% and 6.73% respectively. private-activity bond yields have in the past been as much as 70 basis points higher than comparable essential-purpose bonds (brown, 1988). 10. see note 9. 11. these same basic results are found when p is varied. references aalberts, r. j., and j. c. utley. 1988. ‘the tax-free municipal bond: a security in decline,” journal of taxation and investments, 6: 267-285. bettner, j. 1990. “more people face the minimum-tax formula,” wall street journal, october 29. bierwag, g. 0. 1976. “optimal tic bids on serial bond issues,” management science, 22: 1175 1185. braswell, r. c., and d.l. sumners. 1982. “an analysis of the tic criterion for accepting tax-exempt bond bids under certainty,” decision sciences, 13: 88-100. brown, d. 1988. “impact of the tax reform act of 1986 on municipal bond yields,” municipal finance journal, 9: 111-146. chamberlain, t. w., and c. s. cheung. 1990. “optimal portfolio selection using the general multi-index model: a stable paretian framework,” decision sciences, 21: 563-571. cohen, k. j., and f. s. hammer. 1966. “optimal coupon schedules for municipal bonds.” manage ment science, 3: 161-166. day, c. 1991. “individual income tax rates, 1987.” statistics of income, 11: 13-25. fischer, p. j. 1980. “on the extent of segmentation in the municipal securities market,” journal of money credit and banking, 12: 7 l-83. gehrlein, w. v., and t. h. mcinish. 1985. “cyclical variability of bond risk premia,” journal of banking and finance, 9: 157-165. individual tax-exempt bond portfolio 73 hoffman, w. h., j. e. smith, e. w. willis, and w. a. raabe. 1991. ina’ividual income taxes, 1992 edition, st. paul: west publishing company. hsueh, l. p., and y. a. liu. 1990. “the effectiveness of debt insurance as a valid signal of bond quality,” journal of risk and insurance, 57: 691-700. kidwell, d. s., andt. w. koch. 1983. “market segmentation and theterm structureof interest rates,” journal of money credit and banking, 15: 40-55. kidwell, d. s., e. h. sorensen, and j. m. wachowicz. 1987. “estimating the benefits of debt insurance: the case of municipal bonds,” journal of financial and quantitative analysis, 22: 299-3 13. kochin, l. a., and r. w. parks. 1988. “was the tax-exempt bond market inefficient or were future expected tax rates negative?’ journal of finance, 43: 913-931. kwan, c. c. y., and p. c. yip. 1987. “optimal portfolio selection with upper bounds for individual securities,” decision sciences, 18: 505-523. kwan, c. c. y. 1988. “optimal portfolio selection and cutoff rates of security performance: a multi-index case,” decision sciences, 19: 682-699. mcinish, t. h. 1980. ‘the determinants of municipal bond risk premiums by maturity,” journal of financial research, 3: 129-138. nauss, r. m. 1987. “true interest cost in municipal bond bidding: an integer programming approach,” management science, 32: 870-877. petersen, j. e. 1987. “examining the impacts of the 1986 tax reform act on the municipal securities market” national tax journal, 40: 393-402. petersen, j. e. 1988. ‘the municipal bond market in a changing economy,” public budgeting and finance, 8: 22-34. piros, c.d. 1987. “taxable vs. tax-exempt bonds: a note on the effect of uncertain taxable income,” journal of finance, 42: 447-45 1. porterba, j. m. 1989. “tax reform and the market for tax-exempt debt.” regional science and urban economics, 19: 537-562. puelz, a. v., and r. puelz. 1991. ‘the effect of alternative minimum tax bonds on optimal bond portfolio choice,” proceedings of the 1991 annual meeting of the decision sciences institute, 1: 138-140. puelz, a. v. 1991. “call provisions and the cost effectiveness of debt insurance,” municipal finance journal, 12: 23-34. puelz, a. v., and s. m. lee. 1992. “a multiple-objective programming model for structuring tax-exempt serial revenue bonds,” management science, 38: 1186-1200. quandt, r. e. 1966. “old and new methods of estimation and the pareto distribution,” metrika, 10: 55-82. sachs, l. 1982. applied statistics: a handbook of techniques. new york. springer-verlag. speer, p. d. 1987. “toward an efficient market in new municipal bonds,” municipal finance journal, 8: 329-331. pii: s1057-0810(99)00007-4 the rise and fall of the “dogs of the dow” dale l. domiana, david a. loutonb,*, charles e. mossmanc acollege of commerce, university of saskatchewan, saskatoon, saskatchewan, s7n 0w0, canada bdepartment of finance, bryant college, 1150 douglas pike, smithfield, ri 02917-1284, usa cfaculty of management, university of manitoba, winnipeg, manitoba, r3t 2n2, canada abstract the dow dividend strategy recommends the highest-yielding stocks from the 30 dow industrials. these stocks have come to be known as the “dogs of the dow” since they often include some of the previous year’s worst performers. while the strategy’s successes—and more recently, its failures—have been well documented in the popular press, there have not been any convincing explanations of why the strategy worked. this paper demonstrates that the behavior of these stocks is consistent with the market overreaction hypothesis. in years before the stock market crash of 1987, the dogs were indeed “losers” which went on to become “winners.” but in the post-crash period, the high-yield stocks actually outperformed the market during the previous year. the dow dividend strategy is no longer selecting the true dogs. © 1998 elsevier science inc. all rights reserved. 1. introduction investors have always yearned for ways to beat the market. in recent years, one popular strategy involves the 30 stocks in the dow jones industrial average (djia). according to the dow dividend strategy (dds), a portfolio comprised of the ten highest-yielding djia stocks usually outperforms the dow. initial explanations of the strategy’s success concentrated on the dividends themselves. some explanations involved nothing more than the observation that for a given percentage change in a stock’s price, a higher dividend produces a higher total return. but by the * corresponding author. tel.:11-401-232-6343; fax:11-401-232-6319. e-mail address:dlouton@bryant.edu (d.a. louton) financial services review 7 (1998) 145–159 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00007-4 mid-1990s, market observers realized that the dds had often selected the previous year’s worst performing djia stocks. dds portfolios came to be known as the “dogs of the dow” if a company maintains a constant quarterly cash dividend even though its stock price is falling, the dividend yield must necessarily rise. thus, a high dividend yield may be a proxy for a low past return, and the dow strategy’s success may be a winner-loser phenomenon rather than a yield effect. academic research in market overreaction can therefore provide a methodological framework for examining the dds. this paper examines connections among past returns, dividend yields, and future returns during 1964 through 1997. our dds portfolios consist of the ten highest-yielding djia stocks at the start of each year, and we also analyze portfolios of the ten lowest-yielding stocks. to exclude the effects of the 1987 stock market crash, we examine results over 1964–1986 and 1989–1997 in addition to the full sample. the post-crash subsample also represents a period when the dds became widely known in the popular press. the paper is organized as follows. section 1 presents claims which have been made about the dow dividend strategy. section 2 reviews the market overreaction literature. the methodology employed is discussed in section 3. our tests for overreaction among the djia stocks are presented in section 4. concluding remarks are made in section 5. 2. the dow dividend strategy one of the first reports of the superior performance of high-yielding djia stocks appeared in the wall street journalon august 11, 1988. john slatter, an analyst with prescott, ball & turben, inc., examined the total returns of the ten highest dividend yielding dow stocks for the years 1973 through 1988 and found that they outperformed the djia overall. expanded studies subsequently appeared in books by o’higgins and downes (1991) and knowles and petty (1992). these studies continued to show superior returns from the dds since 1973. knowles and petty also showed that the ten highest-yielding stocks outperformed the dow over a longer period of time from 1957 through 1991. several major brokerage firms, including merrill lynch, prudential securities, and dean witter, followed up with their own studies which provided further empirical evidence to support the earlier results. table 1 summarizes the average annual returns of the ten highest-yielding stocks compared to the dow average, as reported by various studies. table 1 reported returns from the dow dividend strategy study period return on 10 highest yielding stocks return on dow jones industrial average slatter 1973–1988 18.39% 10.86% knowles and petty 1973–1990 17.81% 11.41% o’higgins and downes 1973–1991 16.61% 10.43% prudential securities 1973–1992 16.06% 10.91% average annual returns from the ten highest yielding dow stocks are compared to annual returns on the entire dow jones industrial average. the returns include the reinvestment of dividends. 146 d.l. domian et al. / financial services review 7 (1998) 145–159 prompted by this evidence, merrill lynch, prudential securities, and painewebber cosponsored a unit investment trust (uit) called the defined assets fund select ten portfolio, based on the dow dividend strategy. this type of fund is attractive to the sponsors because of the low cost of implementing and administering such a simple investment strategy. no large staff of highly paid research analysts is required and because these funds are set up as unit investment trusts, they are, by definition, unmanaged. once the portfolio of the ten highest yielding djia stocks is constructed, it is not changed during the one-year life of the fund. at the end of one year, the fund is liquidated at a price determined by the market values of the stocks as of the termination date. investors can choose to receive the proceeds or roll them over into a new uit at a reduced commission charge. the basic dow dividend strategy is straightforward and is executed as follows: step 1: select any starting day (the first trading day of the year is most common) and construct an equally weighted portfolio consisting of the ten stocks in the djia 30 with the highest current dividend yield. step 2: hold the portfolio for one year. on the anniversary date, determine the total value of the portfolio including all dividends and other cash distributions along with the closing values of the stocks. rebalance the portfolio by investing 10% of the total value in each of the ten highest yielding djia stocks. stocks which have dropped off the top-ten yield list should be sold and replaced with the new additions to the list. step 3: repeat the process on each anniversary date. while actual results from various studies differ depending on starting dates used and how dividend yields are defined, all have arrived at similar conclusions about the success of the dow dividend strategy. table 2 shows an annual comparison of the actual performance of the djia and an equally weighted portfolio of the ten highest dividend yielding stocks as reported in a uit prospectus (prudential securities, 1993). results are for the 20 years from january 1973 through december 1992, assuming that total return proceeds are reinvested at the beginning of each calendar year in the ten highest yielding djia stocks in equal dollar amounts (calculated on the previous year’s closing stock prices). results do not include transaction costs or taxes. the dds portfolios had an average annual total return of 16.06% versus 10.91% for the djia. 3. market overreaction the literature relating dividend yields and stock returns is extensive and well established; see, i.e., elton and gruber (1970); black and scholes (1974); black (1976); miller and scholes (1978); litzenberger and ramaswamy (1979); blume (1980); christie (1990). in contrast, studies of market overreaction are more recent. de bondt and thaler (1985) examine the question of stock price predictability in terms of earlier work in experimental psychology. the overreaction hypothesis states that the behavioral tendency of people to “overreact” to surprises extends to the way stock prices are determined. in particular, it suggests that stock prices systematically overshoot because individuals focus excessively on 147d.l. domian et al. / financial services review 7 (1998) 145–159 short-term events such as changes in earnings patterns. evidence of such behavior would be a violation of weak-form market efficiency. two hypotheses are tested: (1) extreme movements in stock prices will be followed by subsequent price movements in the opposite direction; and (2) the more extreme the initial price movement, the greater will be the subsequent adjustment. in their 1985 study, de bondt and thaler (1985) examine the cumulative average residuals of winner and loser portfolios formed in each of 16 non-overlapping three-year periods from january 1933 to december 1980. they find that loser portfolios outperform the market, on average, by 19.6% for the three-year postformation period. winner portfolios underperform the market by about 5.0%. these results are consistent with the overreaction hypothesis. de bondt and thaler (1987) followed up their original study in response to suggestions by some critics that the overreaction effect was, in fact, a rational response to changes in risk (see brown et al., 1988), or that it was primarily caused by mean-reverting factor risk premia. the extended study also addresses unresolved issues relating the overreaction effect to size effects (see zarowin, 1990) and seasonality, as well as the asymmetric nature of the corrections of the winners as compared to those of the losers. to retest the overreaction hypothesis, de bondt and thaler construct rank portfolios of stocks with extreme capital gains (winners) and extreme capital losses (losers) based on past market-adjusted excess returns taken over formation periods of up to five years. using table 2 capital gains, dividends, and total returns year 10 highest yielding stocks dow jones industrial average capital gain dividend yield total return capital gain dividend yield total return 1973 26.22% 5.20% 21.02% 216.58% 3.46% 213.12% 1974 216.32 7.37 28.95 227.57 4.43 223.14 1975 48.78 7.95 56.73 38.32 6.08 44.40 1976 27.70 7.10 34.80 17.86 4.86 22.72 1977 26.75 5.92 20.83 217.27 4.56 212.71 1978 26.92 7.11 0.19 23.15 5.84 2.69 1979 3.97 8.41 12.38 4.19 6.33 10.52 1980 17.83 8.54 26.37 14.93 6.48 21.41 1981 20.94 8.29 7.35 29.23 5.83 23.40 1982 17.24 8.22 25.46 19.60 6.19 25.79 1983 30.20 8.25 38.45 20.30 5.38 25.68 1984 0.24 6.65 6.89 23.76 4.82 1.06 1985 21.45 6.97 28.42 27.66 5.12 32.78 1986 23.74 6.13 29.87 22.58 4.33 26.91 1987 1.87 5.10 6.97 2.26 3.76 6.02 1988 15.80 5.80 21.60 11.85 4.10 15.95 1989 20.28 6.94 27.22 26.96 4.75 31.71 1990 213.00 5.06 27.94 24.34 3.77 20.57 1991 28.32 5.22 33.54 20.32 3.61 23.93 1992 3.44 4.82 8.26 4.17 3.17 7.34 the data reported in this table are obtained from prudential securities (1993). 148 d.l. domian et al. / financial services review 7 (1998) 145–159 varying post-formation test periods, they show that sharp price reversals occur for both the winner and loser stock portfolios. in other words, the losers become winners and vice versa. overall, the losers outperform the winners by an average of 31.9%, and as in their previous study, the overreaction effect is asymmetric. test period returns also show a strong seasonality effect, with a large part of the losers’ excess returns occurring in the month of january for up to five years following portfolio formation. they show that the winner-loser effect cannot be attributed to changes in capm betas and that it is not primarily a size effect. chopra et al. (1992) also find an economically important overreaction effect even after adjusting for size and beta. their evidence suggests that the overreaction effect is distinct from tax-loss selling effects. furthermore, they find that the effect is stronger for smaller firms. jegadeesh and titman (1995) find that contrarian investment strategies are profitable, primarily due to the overreaction of stock prices to firm-specific information. renewed doubts about the existence of market overreaction are raised by conrad and kaul (1993) and ball et al. (1995). conrad and kaul (1993) focus on biases in computed returns due to the cumulation of monthly returns containing measurement errors. they show a large upward bias in the cumulative returns of the lowest priced stocks. ball et al. (1995) also document problems in measuring returns. however, loughran and ritter (1996) dispute conrad and kaul’s methodology, and rozeff and zaman (1998) find overreaction in portfolios which are not affected by the problems raised by ball et al. (1995). 4. methodology we follow the empirical testing procedures employed by de bondt and thaler (1985) in their original study. whereas de bondt and thaler formed winner and loser portfolios conditional on past excess returns, we form portfolios of high-yield and low-yield stocks based on the dividend yields at the beginning of each year. as in the de bondt and thaler study, the tests in this study assess the extent to which systematic nonzero residual return behavior in the twelve-month period after portfolio formation is associated with systematic residual returns in the twelve-month preformation period. stock return data from the center for research in security prices (crsp) are used for the period between january 1963 and december 1997. we use crsp daily data for computing dividend yields; the first full calendar year on these tapes is 1963. the s&p 500 is the benchmark portfolio for market returns. consistent with the de bondt and thaler study, we use market-adjusted excess return residuals estimated asûjt 5 rjt 2 rmt. de bondt and thaler show that the results of their empirical analysis are not affected by the method for determining return residuals. we follow a five-step testing procedure similar to de bondt and thaler (1985, pp. 797–798): 1. for each stockj and each montht, return residuals are determined as described above. at the beginning of each year, the dividend yield for each stock in the djia is determined. the stocks are then ranked according to dividend yield. the ten stocks with the highest dividend yields comprise the high-yield portfolio, while the ten lowest 149d.l. domian et al. / financial services review 7 (1998) 145–159 form the low-yield portfolio. all portfolios are equally weighted. this procedure is repeated each year from 1964 to 1997. although the djia consists of exactly 30 firms at any point in time, 43 different firms appeared in the dow for at least a portion of our sample period. 2. for each high-yield portfolio, 24 average portfolio residualsarh,n,t are calculated for each of the twelve months before and twelve months after the formation date (i.e., the first trading day of the year). the twelve preformation months are denoted byt 5 211 to t 5 0, with t 5 0 representing the prior month of december. similarly,t 5 11 to t 5 112 denote the twelve postformation months witht 5 11 representing the month of january. twenty-four average portfolio residualsarl,n,t are also determined for each of the low-yield portfolios. 3. for each month fromt 5 211 to t 5 112, we compute an average of the average portfolio residuals over the sample period for both high-yield (aarh,t) and low-yield (aarl,t) portfolios. cumulative average average residuals are computed for the highyield portfolios over the twelve preformation months according to the formula: caarh,t 5 o s5211 t aarh,s (1) for t 5 211 to 0. the low-yield cumulative residualscaarl,t are calculated similarly. postformation cumulative residuals are computed separately, restarting the cumulation at the formation point. the high-yield residuals are: caarh,t 5 o s51 t aarh,s (2) for t 5 1 to 12. postformation low-yield cumulative residuals are similar. 4. if dividend yields are related to stock price changes during the preformation months, then we would expect thatcaarh,t , 0 andcaarl,t . 0 for t # 0. the overreaction hypothesis then predicts that fort . 0, caarh,t . 0 andcaarl,t , 0. this implies that, for time periodst # 0, [caarh,t 2 caarl,t] , 0 and, fort . 0, [caarh,t 2 caarl,t] . 0. to determine whether, at any timet, the difference in returns between the high-yield and low-yield portfolios is statistically significant, we find a pooled estimate of the population variance: st 2 5 o n51 n ~carh,n,t 2 caarh,t! 2 1 o n51 n ~carl,n,t 2 caarl,t! 2 2~n 2 1! (3) with two samples of equal sizen (the number of portfolio formation years in the sample period), the variance of the difference of sample means equals2st 2/n and the t-statistic is: 150 d.l. domian et al. / financial services review 7 (1998) 145–159 tt 5 ~caarh,t 2 caarl,t! î2st 2/n (4) for each of the twelve preformation and twelve postformation months, relevant t-statistics can be found but, as noted by de bondt and thaler, they do not represent independent evidence. 5. to determine whether a high-yield average residual for some montht is significantly different from zero, we first compute the sample standard deviation: st 5 îo n51 n ~ arh,n,t 2 aarh,t! 2 ~n 2 1! (5) the t-statistic is tt 5 aarh,t st /în (6) similar procedures apply for the low-yield portfolio. 5. testing the dow dividend strategy we test the dds to consider whether the superior performance of high yielding stocks is actually an overreaction effect. some high yields may result from recent stock price declines rather than explicit dividend policy decisions. our goal is to determine whether high-yield stocks are losers in the preformation months, and whether the subsequent outperformance is in fact de bondt and thaler’s “winner-loser” overreaction effect. a second objective is to compare the performance of the dds over different subperiods. we want to determine whether the underlying dynamics of the dow dividend effect remained stable during the entire 1964–1997 sample period. as documented in table 1, the dds has been extensively publicized since 1988. also, inclusion of 1987 and 1988 may bring about potential confounding effects due to the stock market crash of 1987. we therefore choose 1964–1986 and 1989–1997 as the two subsamples. we apply the tests described in section 3 to the djia stocks. the results for the full 1964–1997 sample period are presented in table 3. for these 34 years, the portfolios of ten high-yield stocks outperform the s&p 500 during the twelve months after portfolio formation by 4.76% (see panel a). our findings are consistent with the claims made by proponents of the dds who use the djia rather than the s&p 500 as a benchmark. in contrast, the portfolios of ten low-yield stocks approximately match the market, underperforming by just 0.52% during the twelve months after portfolio formation. the difference between the cumulative average average residuals of the two portfolios [caarh,12 2 caarl,12] is 5.28%. 151d.l. domian et al. / financial services review 7 (1998) 145–159 table 3 residuals from dividend yield portfolios, 1964–1997 month high-yield low-yield differences aarh,t caarh,t aarl,t caarl,t caarh,t 2 caarl,t panel a 11 0.0225 0.0225 0.0025 0.0025 0.0200 (4.08)** (4.08)** (0.49) (0.49) (2.68)** 12 0.0027 0.0253 0.0076 0.0100 0.0152 (0.70) (3.73)** (2.42)* (1.58) (1.64) 13 0.0077 0.0330 20.0000 0.0100 0.0230 (1.94) (4.12)** (20.01) (1.33) (2.10)* 14 0.0053 0.0382 0.0016 0.0116 0.0267 (1.11) (4.50)** (0.50) (1.39) (2.24)* 15 0.0051 0.0434 20.0050 0.0065 0.0368 (1.47) (4.66)** (21.62) (0.70) (2.78)** 16 0.0012 0.0446 20.0030 0.0035 0.0410 (0.31) (4.28)** (20.68) (0.34) (2.77)** 17 0.0010 0.0456 0.0011 0.0047 0.0409 (0.35) (3.89)** (0.28) (0.43) (2.56)* 18 0.0038 0.0494 0.0015 0.0062 0.0432 (1.06) (4.02)** (0.39) (0.53) (2.53)* 19 0.0057 0.0551 20.0068 20.0006 0.0558 (1.15) (4.24)** (22.24)* (20.05) (3.21)** 110 20.0053 0.0498 20.0110 20.0116 0.0614 (21.10) (4.00)** (22.46)* (21.02) (3.64)** 111 20.0009 0.0489 0.0048 20.0068 0.0557 (20.19) (3.40)** (1.01) (20.54) (2.92)** 112 20.0013 0.0476 0.0016 20.0052 0.0528 (20.30) (3.16)** (0.37) (20.36) (2.54)* average average residuals (aars) and cumulative average average residuals (caars) are presented for the full sample period, 1964 to 1997. the high-yield portfolios include the ten highest-yielding stocks from the dow jones industrial average. the low-yield portfolios include the dow’s ten lowest yielding stocks. this panel shows post-formation monthst 5 11 to t 5 112. numbers in parentheses aret-statistics. panel b 211 0.0074 0.0074 0.0117 0.0117 20.0043 (1.55) (1.55) (2.63)* (2.63)* (20.66) 210 0.0001 0.0075 0.0105 0.0222 20.0147 (0.02) (1.18) (3.23)** (3.70)** (21.69) 29 0.0020 0.0094 0.0059 0.0281 20.0187 (0.55) (1.47) (1.79) (4.00)** (21.96) 28 0.0000 0.0094 0.0066 0.0347 20.0253 (0.00) (1.32) (1.75)* (4.29)** (22.34)* 27 20.0026 0.0069 0.0040 0.0388 20.0319 (20.78) (0.87) (1.22) (4.76)** (22.81)* 26 20.0090 20.0021 0.0022 0.0410 20.0430 (22.58)* (20.23) (0.53) (4.50)** (23.31)** 25 20.0036 20.0057 0.0068 0.0478 20.0534 (21.13) (20.55) (1.92) (4.73)** (23.69)** 24 0.0018 20.0039 0.0051 0.0529 20.0568 (0.61) (20.33) (1.28) (4.86)** (23.58)** 23 20.0017 20.0056 0.0026 0.0555 20.0611 (20.34) (20.47) (0.90) (4.80)** (23.69)** 22 20.0170 20.0226 0.0027 0.0582 20.0809 (23.08)** (21.97) (0.64) (4.61)** (24.74)** 21 20.0090 20.0316 0.0103 0.0685 20.1001 (21.61) (22.13)* (2.40)* (5.28)** (25.09)** 0 20.0051 20.0367 0.0096 0.0781 20.1148 (21.05) (22.26)* (2.52)* (5.38)** (25.26)** average average residuals (aars) and cumulative average average residuals (caars) are presented for the full sample period, 1964 to 1997. the high-yield portfolios include the ten highest-yielding stocks from the dow jones industrial average. the low-yield portfolios include the dow’s ten lowest yielding stocks. this panel shows preformation monthst 5 211 to t 5 0. numbers in parentheses aret-statistics. ** significant at the 1 percent level. *significant at the 5 percent level. 152 d.l. domian et al. / financial services review 7 (1998) 145–159 our results also reveal that the high-yield stocks underperform the market by 3.67% in the twelve months before portfolio formation (see panel b). these high-yield stocks are indeed, in de bondt and thaler parlance, losers. the low-yield stocks outperform the market in the twelve preformation months by 7.81%, establishing them as winners. the difference between the two cumulative returns is 11.48%. furthermore, our findings are consistent with de bondt and thaler in that the overreaction effect is asymmetric; i.e., it is larger for the high-yield stocks than the low-yield stocks. also consistent with de bondt and thaler, we find evidence of a seasonality effect, particularly among the high-yield stocks. the overreaction effect is much more pronounced in january than in subsequent months. in montht 5 11, the high-yield portfolio earns an excess return of 2.25%. figure 1 combines the preformation and postformation periods to show the excess returns over 24 months, cumulating average average residuals fromt 5 211 to t 5 12. for the full two years, the high-yield stocks outperform the s&p 500 by a relatively modest 1.09%. the choice of january as the starting month is arbitrary. while previous dds studies typically follow this convention, it should be noted that the uits co-sponsored by merrill fig. 1. cumulative residuals for 1964–1997. 153d.l. domian et al. / financial services review 7 (1998) 145–159 lynch et al. use different starting months throughout the year. however, the studies and the uits all use annual rebalancing of the portfolios, so we use twelve month postformation periods throughout this paper. interestingly, de bondt and thaler (1985) and chopra et al. (1992) find that portfolios formed on the basis of one-year returns display return momentum instead of overreaction. that is, the previous year’s losers continue to underperform in the next year, while the winners continue to outperform. in an effort to examine robustness of the dds throughout the full sample period, we test the 1964–1986 and 1989–1997 subperiods separately. results for the 1964–1986 pre-crash subperiod are presented in table 4. during the twelve month period following portfolio formation (see panel a) the high-yield portfolio outperforms the s&p 500 by 5.11%. the low-yield portfolio underperforms the benchmark by 3.21% during the same period. the difference between the cumulative residuals of the two portfolios [caarh,12 2 caarl,12] is 8.32%. thus, the dow dividend effect is somewhat stronger during this subperiod than during the full sample period. we find that the high-yield stocks underperform the market by 4.67% in the twelve months before portfolio formation (see panel b). the low-yield stocks outperform the market by 6.16% during this period. figure 2 follows the format used in fig. 1, cumulating residuals over 24 months. for 1964-1986, the preformation and postformation plots of the high-yield stocks are almost mirror images aboutt 5 0, and after two years these stocks gain only 0.44% on the market benchmark. results for the 1989–1997 post-crash subperiod are presented in table 5 and fig. 3. during the postformation period (see panel a) the high-yield portfoliosunderperformthe market portfolio by 1.13% while the low-yield portfolios underperform by a slightly larger amount, 2.78%. the difference between the cumulative average residuals of the two portfolios [caarh,12 2 caarl,12] is 1.65% with at-statistic of only 0.37. thus, a dow dividend effect does not seem to exist during this subperiod, although it should be noted that the lack of statistical significance is partly due to the small sample size. during the preformation subperiod (see panel b), both portfolios outperform the s&p 500, the high-yield by 2.75% and the low-yield by 1.06%. these results contrast sharply with the performance of these portfolios in the pre-crash period, when the high-yield stocks were “losers.” the underlying dynamics of the dow dividend effect have not remained stable during the entire 1964–1997 sample period. two observations can be made to conclude this section. first, any capital market anomaly may disappear after it becomes widely known by investors. mcqueen et al. (1997) suggest that this may have happened to the dow dividend strategy, and furthermore that the dds did not beat the djia economically after adjusting for taxes, transactions costs, and the higher risk from holding an undiversified portfolio of only 10 stocks. other examples of “investor learning” are documented by mittoo and thompson (1990) for the firm size anomaly, and mcqueen and thorley (1997) for gold stocks and gold prices. a second possible explanation for the strategy’s recent failures is that it is no longer selecting the true dogs. during the post-crash period, dividend yield has not been an inverse proxy for past performance of the dow stocks. as noted above, the highest yielding stocks outperformed both the s&p 500 and the low-yield portfolio during the preformation period. 154 d.l. domian et al. / financial services review 7 (1998) 145–159 table 4 residuals from dividend yield portfolios, 1964–1986 (pre-crash period) month high-yield low-yield differences aarh,t caarh,t aarl,t caarl,t caarh,t 2 caarl,t panel a 11 0.0236 0.0236 0.0001 0.0001 0.0234 (3.64)** (3.64)** (0.02) (0.02) (2.69)* 12 0.0018 0.0254 0.0066 0.0067 0.0187 (0.39) (3.15)** (1.47) (0.81) (1.62) 13 0.0066 0.0319 20.0044 0.0023 0.0296 (1.26) (3.14)** (21.03) (0.24) (2.09)* 14 0.0069 0.0388 0.0028 0.0052 0.0337 (1.26) (3.74)** (0.65) (0.48) (2.25)* 15 0.0051 0.0439 20.0120 20.0068 0.0507 (1.20) (3.54)** (23.92)** (20.57) (2.94)** 16 20.0017 0.0422 20.0055 20.0123 0.0545 (20.40) (3.18)** (20.90) (20.85) (2.77)* 17 0.0026 0.0449 0.0012 20.0111 0.0559 (0.73) (3.09)** (0.22) (20.73) (2.66)* 18 0.0034 0.0483 0.0023 20.0088 0.0571 (0.77) (3.03)** (0.57) (20.56) (2.56)* 19 0.0092 0.0575 20.0046 20.0134 0.0709 (1.49) (3.31)** (21.30) (20.87) (3.05)** 110 20.0056 0.0519 20.0140 20.0274 0.0793 (20.89) (3.25)** (22.49)* (21.92) (3.70)** 111 20.0056 0.0463 20.0019 20.0293 0.0756 (20.89) (2.37)* (20.32) (21.85) (3.01)** 112 0.0049 0.0511 20.0028 20.0321 0.0832 (0.89) (2.55)* (20.64) (21.83) (3.12)** average average residuals (aars) and cumulative average average residuals (caars) are presented for 1964–1986, the period preceding the 1987 stock market crash. the high-yield portfolios include the ten highest-yielding stocks from the dow jones industrial average. the low-yield portfolios include the dow’s ten lowest yielding stocks. this panel shows post-formation monthst 5 11 to t 5 112. numbers in parentheses are t-statistics. panel b 211 0.0057 0.0057 0.0065 0.0065 20.0008 (0.96) (0.96) (1.25) (0.25) (20.10) 210 20.0029 0.0028 0.0095 0.0161 20.0132 (20.74) (0.35) (2.42)* (2.16) (21.21) 29 0.0017 0.0045 0.0028 0.0189 20.0144 (0.36) (0.61) (0.67) (2.18) (21.27) 28 0.0013 0.0058 0.0090 0.0279 20.0222 (0.25) (0.67) (1.94) (2.70)* (21.64) 27 20.0014 0.0044 0.0034 0.0313 20.0269 (20.35) (0.44) (0.89) (2.96)* (21.85) 26 20.0124 20.0080 0.0002 0.0315 20.0395 (22.92)** (20.71) (0.05) (2.44)* (22.32)* 25 20.0025 20.0105 0.0065 0.0380 20.0485 (20.67) (20.85) (1.47) (2.51)* (22.48)* 24 0.0011 20.0095 0.0093 0.0473 20.0567 (0.27) (20.63) (2.27)* (2.96)* (22.58)* 23 20.0002 20.0096 0.0018 0.0490 20.0586 (20.02) (20.62) (0.49) (2.96)* (22.59)* 22 20.0191 20.0287 0.0005 0.0495 20.0782 (22.81)** (22.13)* (0.09) (2.72)* (23.46)** 21 20.0161 20.0448 0.0047 0.0542 20.0991 (22.08)* (22.46)* (0.91) (2.82)* (23.74)** 0 20.0018 20.0467 0.0074 0.0616 20.1083 (20.36) (22.40)* (1.92) (2.87)* (23.74)** average average residuals (aars) and cumulative average average residuals (caars) are presented for 1964–1986, the period preceding the 1987 stock market crash. the high-yield portfolios include the ten highest-yielding stocks from the dow jones industrial average. the low-yield portfolios include the dow’s ten lowest yielding stocks. this panel shows preformation monthst 5 21 to t 5 212. numbers in parentheses are t-statistics. **significant at the 1 percent level. *significant at the 5 percent level. 155d.l. domian et al. / financial services review 7 (1998) 145–159 even if the market overreaction phenomenon still exists, investors have not been given the opportunity to exploit it using dow stocks. 5. conclusion our analysis of the dow dividend strategy is generally consistent with the overreaction hypothesis. during 1964–1997, portfolios of the ten highest yielding dow stocks underperform the market in the twelve preformation months, and outperform the market in the twelve months following formation. portfolios of low-yield stocks outperform the market in the preformation period, and slightly underperform in the following twelve months. furthermore, the overreaction effect is asymmetric and more pronounced in january, as previously found by de bondt and thaler. results from the pre-crash 1964–1986 period are similar to the 1964–1997 findings. the most notable difference is the more pronounced underperformance of low-yield stocks in the fig. 2. cumulative residuals for 1964–1986, pre-crash period. 156 d.l. domian et al. / financial services review 7 (1998) 145–159 table 5 residuals from dividend yield portfolios, 1989–1997 (post-crash period) month high-yield low-yield differences aarh,t caarh,t aarl,t caarl,t caarh,t 2 caarl,t panel a 11 0.0165 0.0165 0.0176 0.0176 20.0011 (1.18) (1.18) (1.45) (1.45) (20.06) 12 0.0048 0.0213 20.0014 0.0162 0.0051 (0.76) (1.41) (20.21) (1.00) (0.23) 13 0.0065 0.0278 0.0015 0.0177 0.0101 (0.72) (1.69) (0.29) (1.02) (0.43) 14 20.0027 0.0251 20.0036 0.0141 0.0110 (20.24) (1.48) (20.46) (0.67) (0.40) 15 20.0058 0.0193 0.0119 0.0261 20.0067 (21.09) (0.98) (0.90) (1.39) (20.25) 16 20.0057 0.0137 20.0038 0.0223 20.0086 (20.83) (0.70) (20.58) (1.21) (20.32) 17 0.0017 0.0153 20.0080 0.0142 0.0011 (0.26) (0.81) (21.32) (0.65) (0.04) 18 20.0021 0.0132 0.0000 0.0142 20.0010 (20.25) (0.86) (0.00) (0.85) (20.04) 19 20.0099 0.0033 20.0253 20.0111 0.0144 (21.16) (0.17) (24.70)** (20.55) (0.52) 110 20.0038 20.0004 20.0178 20.0290 0.0285 (20.41) (20.02) (21.30) (21.11) (0.80) 111 20.0162 20.0166 0.0020 20.0270 0.0104 (21.43) (20.57) (0.27) (21.02) (0.26) 112 0.0053 20.0113 20.0008 20.0278 0.0165 (0.55) (20.37) (20.08) (20.88) (0.37) average average residuals (aars) and cumulative average average residuals (caars) are presented for 1989–1997, the period following the 1987 stock market crash. the high-yield portfolios include the ten highest-yielding stocks from the dow jones industrial average. the low-yield portfolios include the dow’s ten lowest yielding stocks. this panel shows post-formation monthst 5 11 to t 5 112. numbers in parentheses are t-statistics. panel b 211 0.0345 0.0345 0.0207 0.0207 0.0137 (2.37)* (2.37)* (1.26) (1.26) (0.62) 210 0.0148 0.0493 20.0006 0.0202 0.0291 (1.78) (3.19)* (20.10) (1.09) (1.21) 29 0.0169 0.0661 0.0006 0.0207 0.0454 (1.98) (3.15)* (0.06) (1.00) (1.54) 28 20.0002 0.0660 0.0021 0.0228 0.0432 (20.01) (2.54)* (0.26) (0.91) (1.20) 27 20.0131 0.0529 0.0016 0.0244 0.0284 (22.28) (1.80) (0.22) (1.00) (0.74) 26 20.0087 0.0442 0.0111 0.0356 0.0086 (20.88) (1.20) (1.28) (1.42) (0.19) 25 20.0045 0.0397 20.0046 0.0310 0.0087 (20.63) (1.13) (20.44) (1.18) (0.20) 24 0.0033 0.0430 0.0014 0.0324 0.0106 (0.48) (1.31) (0.14) (1.36) (0.26) 23 20.0035 0.0395 20.0195 0.0129 0.0266 (20.64) (1.15) (23.48)** (0.50) (0.62) 22 20.0025 0.0370 20.0029 0.0100 0.0270 (20.31) (0.92) (20.27) (0.30) (0.52) 21 20.0094 0.0276 0.0054 0.0154 0.0122 (20.97) (0.69) (1.05) (0.53) (0.25) 0 20.0001 0.0275 20.0048 0.0106 0.0169 (20.15) (0.65) (20.36) (0.35) (0.33) average average residuals (aars) and cumulative average average residuals (caars) are presented for 1989–1997, the period following the 1987 stock market crash. the high-yield portfolios include the ten highest-yielding stocks from the dow jones industrial average. the low-yield portfolios include the dow’s ten lowest yielding stocks. this panel shows preformation monthst 5 21 to t 5 212. numbers in parentheses are t-statistics. **significant at the 1 percent level. *significant at the 5 percent level. 157d.l. domian et al. / financial services review 7 (1998) 145–159 postformation period. in contrast, the post-crash results exhibit greater variability, perhaps due in part to the shorter sample period. over 1989–1997, high-yield stocks have small excess returns during the first three months, but then drop back. thus, during the years in which the dds was becoming popular, the strategy itself was no longer successful. this study did not consider the risk characteristics of the high and low yield portfolios. however, as noted in section 2 above, de bondt and thaler (1987) find that the winner-loser effect cannot be attributed to changes in risk as measured by capm betas. they also find that the winner-loser effect is not primarily a size effect; there is no small firm effect in our study since dow stocks are typically among the largest firms. there remains the possibility that the dow dividend strategy, and even the entire market overreaction literature, reflects nothing more than data mining. according to fischer black, “most of the so-called anomalies that have plagued the literature on investments seem likely to be the result of data mining” (black, 1993, p. 9). fama (1998) observes that some anomalies are overreactions while others are underreactions, and the approximately equal split between the two is consistent with market efficiency. furthermore, fama claims that both types of anomalies usually vanish when different methodologies are adopted. in light of fig. 3. cumulative residuals for 1989–1997, post-crash period. 158 d.l. domian et al. / financial services review 7 (1998) 145–159 these concerns, the best investment strategy may be the simplest—buy and hold a welldiversified portfolio. references ball, r., kothari, s., & shanken, j. (1995). problems in measuring portfolio performance: an application to contrarian investment strategies.journal of financial economics 38, 79–107. black, f. (1976). the dividend puzzle.journal of portfolio management 2, 5–8. black, f. (1993). beta and return.journal of portfolio management 20, 8–18. black, f., & scholes, m. (1974). the effects of dividend yield and dividend policy on common stocks and returns. journal of financial economics 1, 1–22. blume, m. (1980). stock returns and dividend yields: some more evidence.review of economics and statistics 62, 567–577. brown, k., harlow, w. v., & tinic, s. (1988). risk aversion, uncertain information, and market efficiency. journal of financial economics 22, 355–386. chopra, n., lakonishok, j., & ritter, j. (1992). measuring abnormal performance: do stocks overreact?journal of financial economics 31, 235–268. christie, w. (1990). dividend yield and expected returns: the zero-dividend puzzle.journal of financial economics 28, 95–126. conrad, j., & kaul, g. (1993). long-term market overreaction or biases in computed returns?journal of finance 48, 39–63. de bondt, w., & thaler, r. (1985). does the stock market overreact?journal of finance 40, 793–808. de bondt, w., & thaler, r. (1987). further evidence on investor overreaction and stock market seasonality. journal of finance 42, 557–581. elton, e., & gruber, m. (1970). marginal stockholder tax rates and the clientele effect.review of economics and statistics 52, 68–74. fama, e. (1998). market efficiency, long term returns, and behavioral finance.journal of financial economics 49, 283–306. jegadeesh, n., & titman, s. (1995). overreaction, delayed reaction, and contrarian profits.review of financial studies 8, 973–993. knowles, h., & petty, d. (1992).the dividend investor. chicago: probus publishing co. litzenberger, r., & ramaswamy, k. (1979). the effects of personal taxes and dividends on capital prices. journal of financial economics 7, 163–195. loughran, t., & ritter, j. (1996). long-term market overreaction: the effect of low-priced stocks.journal of finance, 51, 1959–1970. mcqueen, g., shields, k., & thorley, s. (1997). does the “dow-10 investment strategy” beat the dow statistically and economically?financial analysts journal 53, 66–72. mcqueen, g., & thorley, s. (1997). do investors learn? evidence from a gold market anomaly.financial review 32, 501–525. miller, m., & scholes, m. (1978). dividends and taxes.journal of financial economics 6, 333–364. mittoo, u., & thompson, r. (1990). do the capital markets learn from financial economists? university of manitoba working paper. o’higgins, m., & downes, j. (1991).beating the dow.new york: harpercollins. prudential securities (1993). prospectus, defined assets fund, select ten portfolio, 1993 winter series. new york: prudential securities. rozeff, m., & zaman, m. (1998). overreaction and insider trading: evidence from growth and value portfolios. journal of finance 53, 701–716. zarowin, p. (1990). size, seasonality, and stock market overreaction.journal of financial and quantitative analysis 25, 113–125. 159d.l. domian et al. / financial services review 7 (1998) 145–159 pii: s1057-0810(99)00021-9 comment on kenneth s. bigel’s paper robert p. gossa,* acertified financial planner board of standards, inc., 1700 broadway, suite 2100, denver, co 80290-2101, usa consumer research has shown that two core values are at the heart of personal financial planning relationship between financial advisors and their clients-competency and trustworthiness. in 1993 a national consumer survey of 500 households with income of $50,000 or more, commissioned by the certified financial planner board of standards, inc. (cfp board), illustrated that consumers believed the following components, of the certification the board offered, are of particularly high importance: ● the existence of a professional code of ethics; ● successful completion of a certification examination; ● individuals being subject to disciplinary action; ● the requirement of practical work experience; ● a financial planning-specific curriculum; and ● the need for continuing education. the availability of these key components as a part of our certification is, perhaps, the single most important reason why the cfp designation has become popular with the public over the years. dr. kenneth s. bigel’s study is the first known survey dealing specifically with the ethics of those holding the cfp credential, and among the first in testing for manifestations of higher ethical development in investment advisers and financial planners. for these reasons alone the work is significant. but his hypotheses also attempt to determine whether there a difference in ethical development because of several elements the value of which have been debated extensively within the financial planner and adviser community for more than decade. is ethics associated with the manner in which an adviser is paid for his or her services? is moral judgment affected by education, age, gender, or the length of work experience? our experience at the cfp board may shed some light on one finding running counter to bigel’s hypothesis that ethical development should increase with career tenure. he found that * corresponding author. tel.:11-303-830-7500; fax:11-303-860-7388. e-mail address:rgoss@cfp-board.org (r.p. goss) financial services review 7 (1998) 237–256 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00021-9 those cfp licensees with less than ten years of career tenure manifested statistically significant higher moral judgment than those with more than ten years of experience. it may be instructive for readers to know that in 1987 the cfp board undertook its first financial planner job analysis and developed its educational standard—a model financial planning curriculum. this standard has now been adopted by more than 90 colleges and universities in the u.s. for their formal education of financial planners. both the curriculum and the comprehensive certification examination, which we first put into effect in 1991, cover to some extent ethical and professional considerations in financial planning. integrity, objectivity, competence, fairness, confidentiality, professionalism and diligence in performing financial planning activities are core principles. further, we updated and emphasized the code of ethics and professional responsibility beginning in 1992, and in 1994 we began requiring a specific ethics course as part of the continuing education requirements that we first inaugurated in 1989. lastly, as cfp licensees know, we have become more diligent in highlighting our increased disciplinary activity handled by the peer review group known as the board of professional review. for example, in the six years between 1987 and 1993 we handled 273 disciplinary cases, while in the five years from 1994 to 1998 we completed 721. in 1999 we anticipate closing another 225 ethics cases. the cumulative result of these developments has been an expectation that those persons recently certified by us, many of which have less work experience or tenure in the field, should in fact have a greater awareness and sensitivity to the moral dimensions of their work as financial planning practitioners. bigel’s other conclusion, that no statistically significant differences were found in the ethics of investment planners based upon how they were compensated for their work, might indeed need more future study. the consumer press notion that advisers who are compensated solely by fees from clients are without conflicts of interest is too simplistic to be accurate. potential or real conflicts of interest may exist for advisers no matter how they are compensated, and this is a basis for the kind of mandatory disclosure of such conflicts and sources of compensation required in the cfp board’s code of ethics. the current trends toward both more fees and less commission and toward greater numbers of advisers and planners receiving some combination of both, again is likely to increase the sensitivity and awareness of cfp licensees to the necessary moral judgements called for in this rapidly growing financial planning field. cfp licensee standards there are two standards pertinent to this research—ethical or conduct standards and practice standards. below is the text of the currentcode of ethics and professional responsibilityas well as practice standards to which cfp licensees adhere. code of ethics and professional responsibility preamble and applicability the code of ethics and professional responsibility (code) has been adopted by the certified financial planner board of standards, inc. (cfp board) to provide principles and 238 r.p. goss / financial services review 7 (1998) 237–256 rules to all persons whom it has recognized and certified to use the cfp certification mark and the marks cfp and certified financial planner (collectively “the marks”). the cfp board determines who is recognized and certified to use the marks. implicit in the acceptance of this authorization is an obligation not only to comply with the mandates and requirements of all applicable laws and regulations but also to take responsibility to act in an ethical and professionally responsible manner in all professional services and activities. for purposes of this code, a person recognized and certified by the cfp board to use the marks is called a cfp designee or certified financial planner designee. this code applies to cfp designees actively involved in the practice of personal financial planning, in other areas of financial services, in industry, in related professions, in government, in education or in any other professional activity in which the marks are used in the performance of their professional responsibilities. this code also applies to candidates for the cfp designation who are registered as such with the cfp board. for purposes of this code, the term cfp designee shall be deemed to include candidates. composition and scope the code consists of two parts: part i—principles and part ii—rules. the principles are statements expressing in general terms the ethical and professional ideals expected of cfp designees and which they should strive to display in their professional activities. as such the principles are aspirational in character but are intended to provide a source of guidance for a cfp designee. the comments following each principle further explain the meaning of the principle. the rules provide practical guidelines derived from the tenets embodied in the principles. as such, the rules set forth the standards of ethical and professionally responsible conduct expected to be followed in particular situations. this code does not undertake to define standards of professional conduct of cfp designees for purposes of civil liability. due to the nature of a cfp designee’s particular field of endeavor, certain rules may not be applicable to that cfp designee’s activities. for example, a cfp designee who is engaged solely in the sale of securities as a registered representative is not subject to the written disclosure requirements of rule 402 (applicable to cfp designees engaged in personal financial planning) although he or she may have disclosure responsibilities under rule 401. a cfp designee is obligated to determine what responsibilities the cfp designee has in each professional relationship including, for example, duties that arise in particular circumstances from a position of trust or confidence that a cfp designee may have. the cfp designee is obligated to meet those responsibilities. the code is structured so that the presentation of the rules parallels the presentation of the principles. for example, the rules which relate to principle 1—integrity, are numbered in the 100 to 199 series while those rules relating to principle 2—objectivity, are numbered in the 200 to 299 series. compliance the cfp board of governors requires adherence to this code by all those it recognizes and certifies to use the marks. compliance with the code, individually and by the profession as a whole, depends on each cfp designee’s knowledge of and voluntary compliance with 239r.p. goss / financial services review 7 (1998) 237–256 the principles and applicable rules, on the influence of fellow professionals and public opinion, and on disciplinary proceedings, when necessary, involving cfp designees who fail to comply with the applicable provisions of the code. terminology in this code “client” denotes a person, persons, or entity who engages a practitioner and for whom professional services are rendered. for purposes of this definition, a practitioner is engaged when an individual, based upon the relevant facts and circumstances, reasonably relies upon information or service provided by that practitioner. where the services of the practitioner are provided to an entity (corporation, trust, partnership, estate, etc.), the client is the entity acting through its legally authorized representative. “cfp designee” denotes current licensees, candidates for certification, and individuals that have any entitlement, direct or indirect, to the federally registered service marks cfp® and certified financial planner®. “commission” denotes the compensation received by an agent or broker when the same is calculated as a percentage on the amount of his or her sales or purchase transactions. “conflict(s) of interest(s)” denotes circumstances, relationships or other facts about the cfp designee’s own financial, business, property and/or personal interests which will or reasonably may impair the cfp designee’s rendering of disinterested advice, recommendations or services. “fee-only” denotes a method of compensation in which compensation is received solely from a client with neither the personal financial planning practitioner nor any related party receiving compensation which is contingent upon the purchase or sale of any financial product. a “related party” for this purpose shall mean an individual or entity from whom any direct or indirect economic benefit is derived by the personal financial planning practitioner as a result of implementing a recommendation made by the personal financial planning practitioner. “personal financial planning” or “financial planning” denotes the process of determining whether and how an individual can meet life goals through the proper management of financial resources. “personal financial planning process” or “financial planning process” denotes the process which typically includes, but is not limited to, the six elements of establishing and defining the client-planner relationship, gathering client data including goals, analyzing and evaluating the client’s financial status, developing and presenting financial planning recommendations and/or alternatives, implementing the financial planning recommendations and monitoring the financial planning recommendations. “personal financial planning subject areas” or “financial planning subject areas” denotes the basic subject fields covered in the financial planning process which typically include, but are not limited to, financial statement preparation and analysis (including cash flow analysis/ planning and budgeting), investment planning (including portfolio design, i.e., asset allocation, and portfolio management), income tax planning, education planning, risk management, retirement planning, and estate planning. “personal financial planning professional” or “financial planning professional” denotes a 240 r.p. goss / financial services review 7 (1998) 237–256 person who is capable and qualified to offer objective, integrated, and comprehensive financial advice to or for the benefit of individuals to help them achieve their financial objectives. a financial planning professional must have the ability to provide financial planning services to clients, using the financial planning process covering the basic financial planning subjects. “personal financial planning practitioner” or “financial planning practitioner” denotes a person who is capable and qualified to offer objective, integrated, and comprehensive financial advice to or for the benefit of clients to help them achieve their financial objectives and who engages in financial planning using the financial planning process in working with clients. part i—principles introduction these principles of the code express the profession’s recognition of its responsibilities to the public, to clients, to colleagues, and to employers. they apply to all cfp designees and provide guidance to them in the performance of their professional services. principle 1—integrity a cfp designee shall offer and provide professional services with integrity. as discussed in composition and scope, cfp designees may be placed by clients in positions of trust and confidence. the ultimate source of such public trust is the cfp designee’s personal integrity. in deciding what is right and just, a cfp designee should rely on his or her integrity as the appropriate touchstone. integrity demands honesty and candor which must not be subordinated to personal gain and advantage. within the characteristic of integrity, allowance can be made for innocent error and legitimate difference of opinion; but integrity cannot co-exist with deceit or subordination of one’s principles. integrity requires a cfp designee to observe not only the letter but also the spirit of this code. principle 2—objectivity a cfp designee shall be objective in providing professional services to clients. objectivity requires intellectual honesty and impartiality. it is an essential quality for any professional. regardless of the particular service rendered or the capacity in which a cfp designee functions, a cfp designee should protect the integrity of his or her work, maintain objectivity, and avoid subordination of his or her judgment that would be in violation of this code. principle 3—competence a cfp designee shall provide services to clients competently and maintain the necessary knowledge and skill to continue to do so in those areas in which the designee is engaged. one 241r.p. goss / financial services review 7 (1998) 237–256 is competent only when he or she has attained and maintained an adequate level of knowledge and skill, and applies that knowledge effectively in providing services to clients. competence also includes the wisdom to recognize the limitations of that knowledge and when consultation or client referral is appropriate. a cfp designee, by virtue of having earned the cfp designation, is deemed to be qualified to practice financial planning. however, in addition to assimilating the common body of knowledge required and acquiring the necessary experience for designation, a cfp designee shall make a continuing commitment to learning and professional improvement. principle 4—fairness a cfp designee shall perform professional services in a manner that is fair and reasonable to clients, principals, partners, and employers and shall disclose conflict(s) of interest(s) in providing such services. fairness requires impartiality, intellectual honesty, and disclosure of conflict(s) of interest(s). it involves a subordination of one’s own feelings, prejudices, and desires so as to achieve a proper balance of conflicting interests. fairness is treating others in the same fashion that you would want to be treated and is an essential trait of any professional. principle 5—confidentiality a cfp designee shall not disclose any confidential client information without the specific consent of the client unless in response to proper legal process, to defend against charges of wrongdoing by the cfp designee or in connection with a civil dispute between the cfp designee and client. a client, by seeking the services of a cfp designee, may be interested in creating a relationship of personal trust and confidence with the cfp designee. this type of relationship can only be built upon the understanding that information supplied to the cfp designee or other information will be confidential. in order to provide the contemplated services effectively and to protect the client’s privacy, the cfp designee shall safeguard the confidentiality of such information. principle 6—professionalism a cfp designee’s conduct in all matters shall reflect credit upon the profession. because of the importance of the professional services rendered by cfp designees, there are attendant responsibilities to behave with dignity and courtesy to all those who use those services, fellow professionals, and those in related professions. a cfp designee also has an obligation to cooperate with fellow cfp designees to enhance and maintain the profession’s public image and to work jointly with other cfp designees to improve the quality of services. it is only through the combined efforts of all cfp designees in cooperation with other professionals, that this vision can be realized. 242 r.p. goss / financial services review 7 (1998) 237–256 principle 7—diligence a cfp designee shall act diligently in providing professional services. diligence is the provision of services in a reasonably prompt and thorough manner. diligence also includes proper planning for and supervision of the rendering of professional services. part ii—rules introduction as stated in part i—principles, the principles apply to all cfp designees. however, due to the nature of a cfp designee’s particular field of endeavor, certain rules may not be applicable to that cfp designee’s activities. the universe of activities by cfp designees is indeed diverse and a particular cfp designee may be performing all, some or none of the typical services provided by financial planning professionals. as a result, in considering the rules in part ii, a cfp designee must first recognize what specific services he or she is rendering and then determine whether or not a specific rule is applicable to those services. to assist the cfp designee in making these determinations, this code includes a series of definitions of terminology used throughout the code. based upon these definitions, a cfp designee should be able to determine which services he or she provides and, therefore, which rules are applicable to those services. rules that relate to the principle of integrity rule 101 a cfp designee shall not solicit clients through false or misleading communications or advertisements: (a) misleading advertising: a cfp designee shall not make a false or misleading communication about the size, scope or areas of competence of the cfp designee’s practice or of any organization with which the cfp designee is associated; and (b) promotional activities: in promotional activities, a cfp designee shall not make materially false or misleading communications to the public or create unjustified expectations regarding matters relating to financial planning or the professional activities and competence of the cfp designee. the term “promotional activities” includes, but is not limited to, speeches, interviews, books and/or printed publications, seminars, radio and television shows, and video cassettes; and (c) representation of authority: a cfp designee shall not give the impression that a cfp designee is representing the views of the cfp board or any other group unless the cfp designee has been authorized to do so. personal opinions shall be clearly identified as such. rule 102 in the course of professional activities, a cfp designee shall not engage in conduct involving dishonesty, fraud, deceit or misrepresentation, or knowingly make a false or 243r.p. goss / financial services review 7 (1998) 237–256 misleading statement to a client, employer, employee, professional colleague, governmental or other regulatory body or official, or any other person or entity. rule 103 a cfp designee has the following responsibilities regarding funds and/or other property of clients: (a) in exercising custody of or discretionary authority over client funds or other property, a cfp designee shall act only in accordance with the authority set forth in the governing legal instrument (e.g., special power of attorney, trust, letters testamentary, etc.); and (b) a cfp designee shall identify and keep complete records of all funds or other property of a client in the custody of or under the discretionary authority of the cfp designee; and (c) upon receiving funds or other property of a client, a cfp designee shall promptly or as otherwise permitted by law or provided by agreement with the client, deliver to the client or third party any funds or other property which the client or third party is entitled to receive and, upon request by the client, render a full accounting regarding such funds or other property; and (d) a cfp designee shall not commingle client funds or other property with a cfp designee’s personal funds and/or other property or the funds and/or other property of a cfp designee’s firm. commingling one or more client’s funds or other property together is permitted, subject to compliance with applicable legal requirements and provided accurate records are maintained for each client’s funds or other property; and (e) a cfp designee who takes custody of all or any part of a client’s assets for investment purposes, shall do so with the care required of a fiduciary. rules that relate to the principle of objectivity rule 201 a cfp designee shall exercise reasonable and prudent professional judgment in providing professional services. rule 202 a financial planning practitioner shall act in the interest of the client. rules that relate to the principle of competence rule 301 a cfp designee shall keep informed of developments in the field of financial planning and participate in continuing education throughout the cfp designee’s professional career in order to improve professional competence in all areas in which the cfp designee is engaged. 244 r.p. goss / financial services review 7 (1998) 237–256 as a distinct part of this requirement, a cfp designee shall satisfy all minimum continuing education requirements established for cfp designees by the cfp board. rule 302 a cfp designee shall offer advice only in those areas in which the cfp designee has competence. in areas where the cfp designee is not professionally competent, the cfp designee shall seek the counsel of qualified individuals and/or refer clients to such parties. rules that relate to the principle of fairness rule 401 in rendering professional services, a cfp designee shall disclose to the client: (a) material information relevant to the professional relationship, including but not limited to conflict(s) of interest(s), changes in the cfp designee’s business affiliation, address, telephone number, credentials, qualifications, licenses, compensation structure and any agency relationships, and the scope of the cfp designee’s authority in that capacity. (b) the information required by all laws applicable to the relationship in a manner complying with such laws. rule 402 a financial planning practitioner shall make timely written disclosure of all material information relative to the professional relationship. in all circumstances such disclosure shall include conflict(s) of interest(s) and sources of compensation. written disclosures that include the following information are considered to be in compliance with this rule: (a) a statement of the basic philosophy of the cfp designee (or firm) in working with clients. the disclosure shall include the philosophy, theory and/or principles of financial planning which will be utilized by the cfp designee; and (b) resumes of principals and employees of a firm who are expected to provide financial planning services to the client and a description of those services. such disclosures shall include educational background, professional/employment history, professional designations and licenses held, and areas of competence and specialization; and (c) a statement of compensation, which in reasonable detail discloses the source(s) and any contingencies or other aspects material to the fee and/or commission arrangement. any estimates made shall be clearly identified as such and shall be based on reasonable assumptions. referral fees, if any, shall be fully disclosed; and (d) a statement indicating whether the cfp designee’s compensation arrangements involve fee-only, commission-only, or fee and commission. a cfp designee shall not hold out as a fee-only financial planning practitioner if the cfp designee receives commissions or other forms of economic benefit from related parties; and (e) a statement describing material agency or employment relationships a cfp designee (or firm) has with third parties and the fees or commissions resulting from such relationships; and (f) a statement identifying conflict(s) of interest(s). 245r.p. goss / financial services review 7 (1998) 237–256 rule 403 a cfp designee providing financial planning shall disclose in writing, prior to establishing a client relationship, relationships which reasonably may compromise the cfp designee’s objectivity or independence. rule 404 should conflict(s) of interest(s) develop after a professional relationship has been commenced, but before the services contemplated by that relationship have been completed, a cfp designee shall promptly disclose the conflict(s) of interest(s) to the client or other necessary persons. rule 405 in addition to the disclosure by financial planning practitioners regarding sources of compensation required under rule 402, such disclosure shall be made annually thereafter for ongoing clients. the annual disclosure requirement may be satisfied by offering to provide clients with the current copy of sec form adv, part ii or the disclosure called for by rule 402. rule 406 a cfp designee’s compensation shall be fair and reasonable. rule 407 prior to establishing a client relationship, and consistent with the confidentiality requirements of rule 501, a cfp designee may provide references which may include recommendations from present and/or former clients. rule 408 when acting as an agent for a principal, a cfp designee shall assure that the scope of his or her authority is clearly defined and properly documented. rule 409 whether a cfp designee is employed by a financial planning firm, an investment institution, or serves as an agent for such an organization, or is self-employed, all cfp designees shall adhere to the same standards of disclosure and service. 246 r.p. goss / financial services review 7 (1998) 237–256 rule 410 a cfp designee who is an employee shall perform professional services with dedication to the lawful objectives of the employer and in accordance with this code. rule 411 a cfp designee shall: (a) advise the cfp designee’s employer of outside affiliations which reasonably may compromise service to an employer; and (b) provide timely notice to the employer and clients, unless precluded by contractual obligation, in the event of change of employment or cfp board licensing status. rule 412 a cfp designee doing business as a partner or principal of a financial services firm owes to the cfp designee’s partners or co-owners a responsibility to act in good faith. this includes, but is not limited to, disclosure of relevant and material financial information while in business together. rule 413 a cfp designee shall join a financial planning firm as a partner or principal only on the basis of mutual disclosure of relevant and material information regarding credentials, competence, experience, licensing and/or legal status, and financial stability of the parties involved. rule 414 a cfp designee who is a partner or co-owner of a financial services firm who elects to withdraw from the firm shall do so in compliance with any applicable agreement, and shall deal with his or her business interest in a fair and equitable manner. rule 415 a cfp designee shall inform his or her employer, partners or co-owners of compensation or other benefit arrangements in connection with his or her services to clients which are in addition to compensation from the employer, partners or co-owners for such services. rule 416 if a cfp designee enters into a business transaction with a client, the transaction shall be on terms which are fair and reasonable to the client and the cfp designee shall disclose the risks of the transaction, conflict(s) of interest(s) of the cfp designee, and other relevant information, if any, necessary to make the transaction fair to the client. 247r.p. goss / financial services review 7 (1998) 237–256 rules that relate to the principle of confidentiality rule 501 a cfp designee shall not reveal-or use for his or her own benefit-without the client’s consent, any personally identifiable information relating to the client relationship or the affairs of the client, except and to the extent disclosure or use is reasonably necessary: (a) to establish an advisory or brokerage account, to effect a transaction for the client, or as otherwise impliedly authorized in order to carry out the client engagement; or (b) to comply with legal requirements or legal process; or (c) to defend the cfp designee against charges of wrongdoing; or (d) in connection with a civil dispute between the cfp designee and the client. for purposes of this rule, the proscribed use of client information is improper whether or not it actually causes harm to the client. rule 502 a cfp designee shall maintain the same standards of confidentiality to employers as to clients. rule 503 a cfp designee doing business as a partner or principal of a financial services firm owes to the cfp designee’s partners or co-owners a responsibility to act in good faith. this includes, but is not limited to, adherence to reasonable expectations of confidentiality both while in business together and thereafter. rules that relate to the principle of professionalism rule 601 a cfp designee shall use the marks in compliance with the rules and regulations of the cfp board, as established and amended from time to time. rule 602 a cfp designee shall show respect for other financial planning professionals, and related occupational groups, by engaging in fair and honorable competitive practices. collegiality among cfp designees shall not, however, impede enforcement of this code. rule 603 a cfp designee who has knowledge, which is not required to be kept confidential under this code, that another cfp designee has committed a violation of this code which raises 248 r.p. goss / financial services review 7 (1998) 237–256 substantial questions as to the designee’s honesty, trustworthiness or fitness as a cfp designee in other respects, shall promptly inform the cfp board. this rule does not require disclosure of information or reporting based on knowledge gained as a consultant or expert witness in anticipation of or related to litigation or other dispute resolution mechanisms. for purposes of this rule, knowledge means no substantial doubt. rule 604 a cfp designee who has knowledge, which is not required under this code to be kept confidential, and which raises a substantial question of unprofessional, fraudulent or illegal conduct by a cfp designee or other financial professional, shall promptly inform the appropriate regulatory and/or professional disciplinary body. this rule does not require disclosure or reporting of information gained as a consultant or expert witness in anticipation of or related to litigation or other dispute resolution mechanisms. for purposes of this rule, knowledge means no substantial doubt. rule 605 a cfp designee who has reason to suspect illegal conduct within the cfp designee’s organization shall make timely disclosure of the available evidence to the cfp designee’s immediate supervisor and/or partners or co-owners. if the cfp designee is convinced that illegal conduct exists within the cfp designee’s organization, and that appropriate measures are not taken to remedy the situation, the cfp designee shall, where appropriate, alert the appropriate regulatory authorities including the cfp board in a timely manner. rule 606 in all professional activities a cfp designee shall perform services in accordance with: (a) applicable laws, rules, and regulations of governmental agencies and other applicable authorities; and (b) applicable rules, regulations, and other established policies of the cfp board. rule 607 a cfp designee shall not engage in any conduct which reflects adversely on his or her integrity or fitness as a cfp designee, upon the marks, or upon the profession. rule 608 the investment advisers act of 1940 requires registration of investment advisers with the u.s. securities and exchange commission and similar state statutes may require registration with state securities agencies. cfp designees shall disclose to clients their firm’s status as registered investment advisers. under present standards of acceptable business conduct, it is proper to use registered investment adviser if the cfp designee is registered individually. if 249r.p. goss / financial services review 7 (1998) 237–256 the cfp designee is registered through his or her firm, then the cfp designee is not a registered investment adviser but a person associated with an investment adviser. the firm is the registered investment adviser. moreover, ria or r.i.a. following a cfp designee’s name in advertising, letterhead stationery, and business cards may be misleading and is not permitted either by this code or by sec regulations. rule 609 a cfp designee shall not practice any other profession or offer to provide such services unless the cfp designee is qualified to practice in those fields and is licensed as required by state law. rule 610 a cfp designee shall return the client’s original records in a timely manner after their return has been requested by a client. rule 611 a cfp designee shall not bring or threaten to bring a disciplinary proceeding under this code, or report or threaten to report information to the cfp board pursuant to rules 603 and/or 604, or make or threaten to make use of this code for no substantial purpose other than to harass, maliciously injure, embarrass and/or unfairly burden another cfp designee. rule 612 a cfp designee shall comply with all applicable post-certification requirements established by the cfp board including, but not limited to, payment of the annual cfp designee fee as well as signing and returning the licensee’s statement annually in connection with the license renewal process. rules that relate to the principle of diligence rule 701 a cfp designee shall provide services diligently. rule 702 a financial planning practitioner shall enter into an engagement only after securing sufficient information to satisfy the cfp designee that: (a) the relationship is warranted by the individual’s needs and objectives; and (b) the cfp designee has the ability to either 250 r.p. goss / financial services review 7 (1998) 237–256 provide requisite competent services or to involve other professionals who can provide such services. rule 703 a financial planning practitioner shall make and/or implement only recommendations which are suitable for the client. rule 704 consistent with the nature and scope of the engagement, a cfp designee shall make a reasonable investigation regarding the financial products recommended to clients. such an investigation may be made by the cfp designee or by others provided the cfp designee acts reasonably in relying upon such investigation. rule 705 a cfp designee shall properly supervise subordinates with regard to their delivery of financial planning services, and shall not accept or condone conduct in violation of this code. financial planning practice standards statement of purpose for practice standards practice standards are being developed and promulgated by the certified financial planner board of standards, inc. (cfp board) for the practice of personal financial planning. these standards: (1) assure that the practice of financial planning by certified financial planner designees (cfp designees) is based on agreed upon norms of practice; (2) advance professionalism in financial planning; and (3) enhance the value of the personal financial planning process. origin of practice standards the cfp board is a professional regulatory organization founded in 1985 to benefit and protect the public by establishing and enforcing education, examination, experience, and ethics requirements for cfp designees. through its certification process, the cfp board has established fundamental criteria necessary for competency in the personal financial planning profession. through its code of ethics and professional responsibility, the board has identified the ethics standards to which personal financial planning professionals should adhere. now, consistent with its objective to promote and maintain professional standards and continuing competency among cfp designees, the cfp board addresses standards of practice for personal financial planning. the cfp board has established the board of practice 251r.p. goss / financial services review 7 (1998) 237–256 standards, a subsidiary board comprised exclusively of cfp practitioners, to draft these practice standards. practice standards defined a practice standard establishes the level of professional practice that is expected of cfp designees engaged in personal financial planning. the services provided depend on the facts and circumstances of a particular situation. practice standards apply to cfp designees in performing the tasks of personal financial planning regardless of the person’s title, job position, type of employment, or method of compensation. practice standards should be considered by all personal financial planning professionals when performing the financial planning task or activity addressed by the standard but are enforceable by the cfp board only against cfp designees. conduct inconsistent with a standard in and of itself is not intended to give rise to a cause of action nor to create any presumption that a legal duty has been breached. the standards are designed to provide cfp designees a structure for identifying and implementing expectations regarding the professional practice of personal financial planning. they are not designed to be a basis for legal liability. practice standards are not intended to prescribe step-by-step procedures for providing any particular service. such procedures may be provided in practice aids developed by various financial planning organizations and other sources. practice standards are being developed for selected financial planning activities identified in a financial planner job analysis first conducted by the cfp board in 1987 and updated in 1994 by ctb/mcgraw-hill, an independent consulting firm. financial planning process related practice standard series establishing and defining the relationship with the client 100 gathering client data including goals 200 analyzing and evaluating the client’s financial status 300 developing and presenting financial planning recommendations and/or alternatives 400 implementing the financial planning recommendations 500 monitoring the financial planning recommendations 600 compliance the practice of financial planning consistent with these standards is required for cfp designees and will be enforced by the cfp board. 252 r.p. goss / financial services review 7 (1998) 237–256 practice standard 100-1 establishing and defining the relationship with the client defining the scope of the engagement the scope of the engagement shall be mutually defined by the financial planning practitioner and the client prior to providing any financial planning service. explanation of this practice standard prior to providing any financial planning service, a financial planning practitioner and the client shall mutually define the scope of the engagement. the process of “mutually-defining” is essential in determining what activities may be necessary to proceed with the client engagement. this is accomplished by: ● identifying the service(s) to be provided; ● disclosing financial planning practitioner’s compensation arrangement(s); ● determining the client’s and the financial planning practitioner’s responsibilities; ● establishing the duration of the engagement; and ● providing any additional information necessary to define or limit the scope. the scope of the engagement may include one or more financial planning subject areas. it is acceptable to mutually define engagements in which the scope is limited to specific activities. this serves to establish realistic expectations both for the client and the practitioner. this practice standard does not require the scope of the engagement to be in writing. however, as noted in section 3, there may be certain disclosures that might be required to be in writing. as the relationship proceeds, the scope may change by mutual understanding. this practice standard shall not be considered alone, but in conjunction with all other practice standards. effective date this practice standard shall be effective january 1, 1999. practice standard 200-1 gathering client data determining a client’s personal and financial goals, needs and priorities a client’s personal and financial goals, needs and priorities that are relevant to the scope of the engagement and the service(s) being provided shall be mutually defined by the 253r.p. goss / financial services review 7 (1998) 237–256 financial planning practitioner and the client prior to making and/or implementing any recommendations. explanation of this practice standard prior to making recommendations to a client, a financial planning practitioner (practitioner) and the client shall mutually define the client’s personal and financial goals, needs and priorities. in order to arrive at such a definition, the practitioner will need to explore the client’s values, attitudes, expectations, and time horizons as they affect the client’s goals, needs, and priorities. the process of “mutually-defining” is essential in determining what activities may be necessary to proceed with the client engagement. personal values and attitudes shape a client’s goals and objectives and the priority placed on them. accordingly, these goals and objectives must be consistent with the client’s values and attitudes in order for the client to make the commitment necessary to accomplish them. goals and objectives provide focus, purpose, vision, and direction for the financial planning process. it is essential that objectives relative to the scope of the engagement are determined and that they are clear, precise, consistent, and measurable. the role of the practitioner is to facilitate the goal-setting process in order to clarify, with the client, goals and objectives, and, when appropriate, the practitioner must try to assist clients in recognizing the implications of unrealistic goals and objectives. this practice standard addresses only the tasks of determining a client’s personal and financial goals, needs and priorities; assessing a client’s values, attitudes and expectations; and determining a client’s time horizons. these areas are subjective and the practitioner’s interpretation is limited by what the client reveals. a practitioner performing the activity of “gathering client data” should consider together the various practice standards applicable to such activity. this practice standard shall not be considered alone, but in conjunction with all other practice standards. effective date this practice standard shall be effective january 1, 1999. relationship of this practice standard to the cfp board’s code of ethics and professional responsibility this practice standard relates to the cfp board’scode of ethics and professional responsibilitythrough the code’s principle 4-fairness rule 402 and principle 7 diligence rule 702. principle 4 states that “a cfp designee shall perform professional services in a manner that is fair and reasonable to clients . . .”. although, as stated earlier, there is no requirement that the scope of the engagement be in writing, rule 402 in thecode of ethics and professional responsibilityrequires a financial planning practitioner to make “timely written 254 r.p. goss / financial services review 7 (1998) 237–256 disclosure of all material information relative to the professional relationship. in all circumstances such disclosure shall include . . . sources of compensation”. principle 7 states that “a cfp designee shall act diligently in providing professional services. rule 702 requires that financial planning practitioners enter into an engagement only after obtaining sufficient information to satisfy that “the relationship is warranted by the individual’s needs and objectives; and the cfp designee has the ability to either provide requisite competent services or to involve other professionals who can provide such services.” anticipated impact of this practice standard upon the public.the public is served when the relationship is based upon a mutual understanding of the engagement. clarity of the scope of the engagement enhances the likelihood of achieving client expectations. upon the financial planning profession.the profession benefits when clients are satisfied. this is more likely to take place when clients have expectations of the process, which are both realistic and clear, before services are provided. upon the financial planning practitioner.a mutually defined scope of the engagement provides a framework for the financial planning process by focusing both the client and the practitioner on the agreed upon tasks. this enhances the potential for positive results. practice standards 200-2 gathering client data determining a client’s personal and financial goals, needs and priorities a financial planning practitioner shall obtain sufficient and relevant quantitative information and documents about a client applicable to the scope of the engagement and the services being provided prior to making and/or implementing any recommendations. explanation of this practice standard prior to making recommendations to a client and depending upon the type of client engagement and its scope, a financial planning practitioner (practitioner) shall determine what quantitative information and documents are sufficient and relevant. a practitioner shall obtain sufficient and relevant quantitative information and documents pertaining to the client’s financial resources, obligations, and personal situation. this information may be obtained directly from the client or other sources through interview, questionnaire, client records and documents. a practitioner shall communicate to the client a reliance on the completeness and accuracy of the information provided and that incomplete or inaccurate information will impact conclusions and recommendations. 255r.p. goss / financial services review 7 (1998) 237–256 if a practitioner is unable to obtain sufficient and relevant quantitative information and documents to form a basis for recommendations, the practitioner shall either: (a) restrict the scope of the engagement to those matters for which sufficient and relevant information is available; or (b) terminate the engagement. a practitioner shall communicate to the client any limitations on the scope of the engagement, as well as the fact that this limitation could affect the conclusions and recommendations. this practice standard shall not be considered alone, but in conjunction with all other practice standards. effective date this practice standard shall be effective january 1, 1999. relationship of this practice standard to the cfp board’s code of ethics and professional responsibility this practice standard relates to the cfp board’scode of ethics and professional responsibilitythrough the code’s principle 7—diligence, and rules 701 through 703. rule 701 states that “a cfp designee shall provide services diligently.” rule 702 requires a financial planning practitioner to “enter into an engagement only after securing sufficient information to satisfy the cfp designee that . . . the relationship is warranted by the individual’s needs and objectives . . .”. in addition, rule 703 requires a financial planning practitioner to “make and/or implement only recommendations which are suitable for the client.” anticipated impact of this practice standard upon the public.the public is served when the relationship is based upon mutually defined goals, needs, and priorities. compliance with this practice standard reinforces the practice of putting the client’s interest first which is intended to increase the likelihood of achieving the client’s goals and objectives. upon the financial planning profession.compliance with this practice standard emphasizes to the public that the client’s goals, needs, and priorities are the focus of the financial planning process. this encourages the public to seek out the services of a financial planning practitioner who uses such an approach. upon the financial planning practitioner.the client’s goals, needs and priorities help determine the direction of the financial planning process. this focuses the practitioner on the specific tasks that need to be accomplished. ultimately this will facilitate the development of appropriate recommendations. 256 r.p. goss / financial services review 7 (1998) 237–256 pii: s1057-0810(99)80009-2 financial services review, 7(1): 1-10 issn: 1057-0810 copyright © 1998 by jai press inc. all rights of reproduction in any form reserved. shareholder wealth effects of calpers' activism claire e. crutchley, carl d. hudson, and marlin r.h. jensen in the past decade, institutional investors have become more active in monitoring man agement and voting the shares they control. the california public employees' retirement system ( caipers) was a leader in this wave of activism. this study investi gates the long-term returns an investor with public information could earn by buying a portfolio of firms targeted by caipers and whether the success of caipers' activism depends on the aggressiveness of the targeting. the evidence supports the idea that vis ible and aggressive activism leads to substantial increases in shareholder wealth while a quieter activism does not. i. introduction the past several decades have witnessed a trend towards the institutionalization of the sav ings process; more than half of all stocks outstanding are held through institutions rather than direct holdings of securities by individuals. the increased popularity of pension funds since world war ii and the more recent development and expansion of keogh plans and ira accounts have resulted in the delegation of a large degree of corporate control to the managers of financial institutions. prior to the 1980s such institutional investors were rel atively passive and tended to vote shares in accordance with management's wishes. how ever, as the holdings of institutional investors increased, so did the pressure to produce attractive returns. in an attempt to improve performance, institutional investors have become much more active in monitoring management and voting the shares they control. the california public employees' retirement system (calpers) under the direction of ceo dale hanson was an early leader of this wave of activism. each year since 1987, calpers has targeted a small group of firms in its portfolio that it perceives as having management problems. articles in the wall street journal have reported that investors could earn very high long-term returns by simply buying the firms that calpers targeted. however, in 1994 hanson resigned, and was replaced by james burton who continues to target finns, but whose activism is not as vocal as hanson's. this study investigates the long-term returns that an investor with public information could earn by buying a portfolio claire e. crutehley, carl d. hudson, and marlin r.h. jensen ° 303 lowder business building, auburn university, auburn, al 36849, fax: 3341844-4960; e-mail: claire@business.auburn.edu. 2 financial services review 7(1 ) 1998 of firms targeted by calpers and whether the success of calpers' activism in the hanson era differs from the post-hanson era. ii. background the history of calpers' activism can be separated into three stages; dale hanson, cal pers' ceo from 1987-1994, spearheaded the first two. calpers began its aggressive shareholder activism campaign in 1987 when many boards of directors were enacting poi son pills and staggered boards to avoid being bought out in the very active takeover market of the 1980s. the initial activism in the 1987-1990 time period was geared towards elimi nating those poison pills and changing corporate governance structures (nesbitt, 1994). for example, in 1987 calpers introduced anti-poison pill resolutions at amr corpora tion and influenced management at aluminum company of america to withdraw their poi son pill (wall street journal [wsj], kilman, 1988). in 1990, calpers' focus changed and it began to target firms based on poor stock market performance. it examined firms in the bottom half of the standard & poors (s&p) 500 index and hired consultants to prepare detailed reports that were used to select the finns to target. calpers first sent a letter to the targeted firms' board of directors that asked management to meet with calpers' staff and discuss problems with the company. in 1993 bill crist, the president of calpers' board, said, "our objective is not to instill fear, but to encourage good performance" (the washington post, vise, 1993). however, if management was not responsive to calpers' concerns, calpers took action. calpers introduced shareholder proposals in many firms, voted against the management slate of directors, and also retargeted firms the following year again asking for changes. for exam ple, in 1994 calpers publicly announced its support for shareholder activist robert a.g. monks against the wishes of sears' board of directors (los angeles times, silverstein, 1991). calpers also announced its intention to vote against the board of phillip morris when they would not meet with calpers' staff (wsj, hwang, 1995). in the early years, the fact that calpers targeted a firm would only become public when a shareholder resolution was proposed. however, in 1992 calpers began publicly announcing the list of the target firms. hanson said, "a number of companies won't move unless they have to deal with [the problem] because it's in the public eye" (business week, dobrzynski, 1992). presently, calpers continues to focus on poor stock price performance and publicly announces the target firms. however, this stage is different due to the change in manage ment styles; "to be less visible is not to be less effective," says current ceo, james bur ton. however, robert monks, a shareholder activist through lens corporation says, "what gave calpers the power was the personality of dale hanson" (rehfeld, 1997). in 1995, the first year after hanson resigned, richard h. koppes, general counsel of calpers, announced the targeting of nine firms and promised to file shareholder proposals for com panies who didn't make significant changes in response to calpers (business week, schine, 1995). in 1996, koppes left calpers and he now feels calpers is not continuing the public activism as it should; koppes said, "five years of going to the press, now they don't go" (rehfeld, 1997). the size of the firms that are targeted also changed following hanson's shareholder wealth effects 3 departure. unlike early years when very large firms were targeted, now mainly midsize companies are targeted. calpers' strategy changed from targeting firms such as ibm to firms such as edison brothers stores and venture stores incorporated. rehfeld (1997) documents charming shoppes' experience with being targeted in 1996 during the post-hanson era. the treasurer of charming shoppes, bernard brodsky, received calpers' letter and agreed to meet with calpers' representatives. in the meeting, he answered questions and documented changes being made. however, there was no follow-up to this meeting from calpers. after several months brodsky called to find out whether calpers had more concerns and was told channing shoppes was fine even though over that time period they recorded losses and their stock price fluctuated greatly. there have been several empirical studies documenting the stock return effects of shareholder activism. nesbitt (1994) documents huge gains to shareholders following the calpers' letter targeting a firm for poor performance. he does not find gains for firms tar geted for corporate governance issues. these results are for companies targeted from 1990-1992, during hanson's tenure as ceo of calpers. smith (1996) documents significant positive two-day and long-term stock returns fol lowing the announcement that calpers is targeting a firm for performance in 1989-1993. wahal (1996) studies activism by several pension funds. on average, he finds neither long-term nor short-term abnormal stock returns following pension fund activism. how ever, he finds evidence that when calpers pursues the activism, there are short-term (six-day) abnormal gains in the market following the day a letter is sent to management. neither smith nor wahal find any evidence of improvement in accounting measures fol lowing the activism. strickland, wiles, and zenner (1996) find significant positive two-day stock returns to the announcement that the shareholder activist group united shareholder association (usa) negotiated an agreement with management. however, unlike the evidence on cal pers' activism by smith and wahal, they do not find the significant returns at the announcement of the shareholder proposal. therefore, the usa activism itself is not good news, only the successful resolution causes increased shareholder value. opler and sokobin (1995) study the effects of activism by the council of institutional investors, and find that in the long-term there are significant gains to the shareholders of the targeted firms. the council of individual investors is a group that was formed by calpers, but it has a low profile targeting strategy. akhigbe, madura, and tucker (1997) study both insti tutional and individual activism. they find positive long-term stock returns from activism, and find that activism instituted by individual investors leads to higher stock returns than proposals by institutions such as calpers. overall, the evidence supports the idea that outsiders pursuing active monitoring can cause increases in stockholder returns. the evidence on calpers' activism is the stron gest, showing positive short and long-term returns to shareholders after calpers' targets firms for poor performance. these studies all examine targets prior to the resignation of dale hanson. this paper will extend the previous papers by adding calpers' targets from 1994-1997. we also calculate returns following the public announcement date of the tar gets to test whether investors with only public information could earn the high returns doc umented in the previous studies. 4 f inancial services r eview 7(1) 1998 iii. data and methodology the sample in this study consists of all public announcements of calpers' targets from 1992-1997. we collect data on stock returns for these companies, stock returns for the standard & poors (s&p) 500 index, evidence of significant changes in the firms following the targeting, the ownership structure and other attributes of the firm. the announcement date is the date calpers' target list was reported in the wall street journal (wsj) over a six-year period (wsj, march 20, 1992; january 22, 1993; january 19, 1994; february 3, 1995; february 6, 1996; and february 11, 1997). the first date, march 20, 1992, is the first time calpers publicly announced its list of target companies. the initial sample contains the 63 firms that appear in the six wsj articles (table 1). several firms have been targeted more than one year; this indicates that calpers was not satisfied with the firm's response to their initial letter. a study by anand (1994) indicates that of the ten original targets in 1993, calpers only followed through with shareholder proposals on three of the firms by 1994. this indicates that seven of the ten targeted firms addressed calpers' concerns in some way. however, others were retargeted, such as ibm that was targeted in 1992, 1993 and 1994. this study includes only the first year in which a firm was targeted publicly, since investors could have different expectations following a second calpers' target. the final sample contains 47 targeted firms. table 1 firms targeted by calpers as reported in the wall street journal 3/23/92 american express control data corporation chrysler corporation dial corporation hercules incl itt corporation l ibm polaroid ryder system incorporated salomon incorporated time-warner usair group 2/03/95 boise cascade 2 first mississippil jostens incorporated kmart corporation melville corporation navistar international 1 oryx energy corporation u.s. shoe company 1 zurn industries 1/22/93 advanced micro devices boise cascade champion international chrysler corporation 1 ibm l macfrugals bargains pennzoil polaroid 1 sears sizzler international time-warner 1 westinghouse electric 2/06/96 applied bioscience international bassett furniture industries charming shoppes inc edison brothers stores inc melville corporation1 oryx energy corporation l rollins environmental services stride rite corporation u.s. surgical corporation venture stores inc 1/19/94 boise cascade l cpi corporation eastman kodak first mississippi ibm 2 navistar international usx corporation u.s. shoe company westinghouse electric i zenith electronics corporation 2/11/97 apple computer bassett furniture industries l fleming cos novell inc reebok international ltd rollins environmental services j sensormatic electronics corp stride rite corporation l summit technology inc sybase inc notes: i. calpers targeted this firm in the prior year. 2. calpers targeted this firm in the prior two years. shareholder wealth effects 5 table 2 wall street journal reports of activities associated with targeted firms in the target year total sample 1992-1994 1995-1997 number of firms period period management activities ceo change 7 5 2 top manager change 6 3 3 restructuring 13 7 6 divestiture 14 8 6 debt issue 4 4 0 stock issue 7 6 ! cost cutting 14 i 0 4 executive compensation decrease 5 4 1 executive compensation increase 3 3 0 other debt downgrade 14 10 4 debt upgrade 8 8 0 the purpose of calpers' activism is to cause management of the targeted firms to change policies so that shareholders' (and calpers') wealth can be increased. the wall street journal index was examined for reports of significant structural changes in the tar get firms in the year following the target announcement. many firms reported major changes in operations (table 2). major changes include restructuring, divestitures, general cost cutting, and top management changes. also, several firms issued either debt or equity during the target year. managerial incentives to make significant changes and the resulting shareholder returns may be impacted by the structure of the firm' s equity ownership. jensen and meck ling (1976) assert that firms with low insider ownership need more active monitoring. firms with high insider ownership may already have the incentives to operate at a high level of efficiency. consistent with this argument, smith (1996) finds that calpers does not target firms with high insider ownership and low institutional ownership. data on own ership structure was collected from the compact disclosure data files; the ownership vari ables collected are the percentage of stock held by insiders, institutional owners, and large (greater than five percent) owners. the success of shareholder activism may also be a function of the size of the firm; larger firms are less likely to experience a large gain, all else equal. the size of each firm is measured by the market value of its equity the year prior to targeting, which is calculated as closing price times number of shares outstanding as given by the standard & poors compustat database. the measure of shareholder's excess return is the targeted company's holding period return less the stock return of the s&p 500 index. the cumulative excess performance is calculated as: n n cerin = l-i (1 + rit ) l"l (1 +st ) t = l t = l 6 financial services review 7(1) 1998 table 3 descriptive statistics of calpers targets [the mean and standard deviation (in parentheses) are listed for each variable] total 1992-1994 1995-1997 targeted sample time period time period variable n = 47 n = 27 n = 20 market size of the firms (in millions of dollars) 3666.73 5474.13 a 1226.74 a (7987.91 ) ( 10,167.45) ( 1451.54) percentage of the stock held by institutional investors 56.68 60.35 a 51.71 a (15.34) (11.33) (18.68) percentage of the stock held by insiders 5.73 6.18 5.13 (10.08) (12.38) (5.96) percentage of the stock held by 5% owners 24.69 23.18 26.73 (18.60) (20.46) (16.02) note: athe firms in the 1992-1994 time period are significantly different from the firms targeted in the 1995-1997 time period at the 0.05 level. where cerin is the n-day cumulative excess return for firm i, tit is the daily return for firm i, and s t is the daily return on the s&p 500 index. the cumulative return is calculated each month from the day following the announcement through the first year thereafter. consis tent with prior studies, we expect to see long term gains to stockholders' wealth following the public announcement of a firm being targeted by calpers. cumulative monthly excess returns are examined the year following the announce ment for the entire sample and for two sub-periods, 1992-1994 and 1995-1997. the later period corresponds to targets announced in the post-hanson era. using standard t-tests, we test whether the cumulative excess returns are greater than zero for each month, indicating whether a portfolio of calpers' targets would earn higher returns than the s&p 500 index. we also test whether the returns in the two sub-periods are statistically different. other factors such as ownership structure, finn size, and the type of significant mana gerial activities that were undertaken during the year after targeting are examined for effects on stockholder returns. both firm size and the percentage of the stock held by insti tutional investors are significantly different for the two periods (table 3). finally, the excess stock returns are regressed against a dummy variable for time period and the other variables. the regression results will indicate whether the relationship between the excess returns and the other variables is conditional on the time period. iv. results the mean cumulative excess return (cer) is reported monthly for the whole sample and for the two sub-periods (table 4, figure 1). for the total sample, the cer is positive for each month, which indicates that an investor holding a portfolio of the firms targeted by calpers would earn higher returns than holding the s&p 500 index. however, the excess returns are statistically different from zero only up to eight months following the announcement. for the entire year, the cumulative excess return of the portfolio of targeted firms is not different than the return of the s&p 500 index. thus, the act of targeting has shareholder wealth effects 7 t a b l e 4 c u m u l a t i v e e x c e s s r e t u r n s ( c e r ) o f c a l p e r s t a r g e t s a f te r the wall street journal a n n o u n c e m e n t d a t e a time pehod (monks after ~e wsj announcement difference between calpers targets time periods 1992-1994 1995-1997 total sample time period time period n = 47 n = 27 n = 20 test cer cer cer statistic 1 2.47% 0.95% 4.54% -0.77 2 6.98%* 3.57% 11.59%* -1.57 3 8.63%* 6.03%* 12.14% -0.79 4 11.05%* 8.03%* 15.18% -0.75 5 5.47% 4.76% 6.42% -0.18 6 5.80% 7.13%* 4.00% 0.38 7 8.27%* 10.41%* 5.37% 0.61 8 8.35%* 11.30%* 4.36% 0.90 9 4.09% 11.93%* -6.50% 2.24 b 10 1.82% 12.54%* -12.60% 2.91 b 11 0.65% 13.07%* -19.18%* 3.35 b 12 0.17% 14.79%* -19.56%* 3.4@ notes: athe cumulative excess return is calculated as the mean difference between the return for the calpers targets minus the return for the standard & poors 500 index, bthe cumulative excess returns for the firms in the 1992-1994 targeted group are significantly different from the 1995-1997 targeted group at the 0.05 level. *significantly different from zero at the 0.05 level. 0.2 / 0.15 [ . . . . . . . . . . . . . . . . . . . . . . . . . . ;-~: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.t ................................................. / ; ................ ~ . ~ . ~ ' ~ , ~ ' , ~ , . . . . . . . . . . . . . . . . . . . . . . . : --- * / n~ _ 4 %% f ~ . . . . ",,, " " " ~ ~ .~ .~. 0 1 2 3 4 5 6 7 8 '. 9 10 11 12 -0.05 : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -0,15 ......................................................................... :.~ ............. -0.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . "~ l:."""" -0.25 figure 1. calpers targeting: excess returns 8 financial services review 7(1) 1998 some short run impact, but there is no lasting effect. there may, however, be differing effects depending on the year of the targeting. when the whole sample is split into sub-periods, two distinct patterns appear. for the 1992-1994 period, the cumulative return for the targeted firms is positive and significant. after twelve months the average return is 14.79% greater than that of the s&p which is sta tistically different from zero with 95% confidence. for the 1995-1997 period, the perfor mance of the targeted portfolio is only statistically above that of the s&p 500 index through month two, and by the end of 12 months, the performance on these later caipers' targets is significantly below that of the s&p 500 index by 19.56%. investors appear to find the targeting during this period to be a positive signal only in the short run. to determine whether the difference in the returns between the two sub-periods is due to differences in the firms' responses or differences in calpers's strategy, we regress the excess returns on ownership structure, firm size, actions taken following caipers target ing, and a dummy variable representing the time period in which the firm was targeted (table 5). the only variable found to be significant is the time period dummy variable; firms which were targeted in the later period had significantly lower excess returns than the firms targeted in early periods. the specific actions taken by companies in the year after the target do not appear to have any effect on returns. a possible explanation is that different actions are needed in different firms, so restructuring is not superior to cost cutting; firms act in ways to enhance value. the difference between the results in the two different time periods was not caused by size differentials in the firms targeted, as size does not have a significant effect on stock returns. another possible explanation for the significantly lower excess returns in the later time period is the performance of the overall stock market. if target firms have high returns, but the s&p 500 index had extraordinary returns, this would be shown as a negative excess return. however this is not the case. we examine the overall raw returns for the targeted table 5 cross-sectional regression estimates of the twelve-month cumulative excess returns on various independent variables a (test-statistics are in parentheses) intercept insider inst five cost restruct size time parameter 0.31 0 . 7 6 0.03 0.58 0.19 0 . 1 9 0 . 3 6 -0.38*** estimate (0.85) ( -1 .13) (0.10) (1.52) (0.93) ( -1 .43) ( -0 .78 ) ( -3 .45) adjusted r 2 21 .16% f value 2.76** notes: atbe cumulative excess return is calculated as the mean difference between the return for the calpers targets minus the return for the standard & poors 500 index. bth¢ variable insider represents the percentage of stock held by insiders in the firm as reported by disclosure; inst represents the percentage of stock held by institutional investors as reported by disclosure; five represents the per centage of stock held by five percent owners as reported by disclosure; cost is a dummy variable representing 1 if the fu'm has announced a cost-cutting measure in the wall street journal during the year following the target announcement and 0 if not; restruct is a dummy variable representing i if the firm has announced a restructuring measure in the wall street journal during the year following the target announcement and 0 if not; size in the natural log of the stock price of the firm multiplied by the number of shares outstanding, and time is a dummy variable rep resenting 1 if the firm was targeted in the 1992-1994 time period and 0 if not. ***the coefficient is significantly different from zero at the 0.01 level. ** the coefficient is significantly different from zero at the 0.05 level. shareholder wealth effects 9 0.25 e | .! z= =e o 0.2 0.15 0.1 ¸ 0.05 -o.os figure 2. months after targeting r loosl~o7 talced firms s & p 5oo [ calpers targeting 1995-1997: raw returns vs. s&p firms and the s&p 500 index in the later time period (figure 2). while the s&p 500 index was doing very well, the average target earned a higher return through eight months follow ing the target. in the last four months of the year, all gains to the target f'maas were erased. v. summary and conclusion this paper examines whether investors can earn returns higher than a market index by pur chasing a portfolio of t-n-ms after the announcement that calpers has targeted these firms. the evidence from this study indicates that the returns earned depend upon the era of cal pers' activism. an investor following the strategy of buying the portfolio of first time tar gets and holding these stocks for a year would have earned high returns in the 1992-1994 time period. however, an investor continuing this strategy in 1995-1997 would have earned much less than the s&p 500 after controlling for firm size, ownership structure, and the types of managerial actions taken. this later time period coincides with the era of less visible activism. a related question is whether pension fund investors benefit when managers are active monitors of firms in their portfolio. the evidence in this paper supports the idea that very visible and aggressive activism does cause substantial increases in shareholder wealth. however, a quieter activism does not yield the same results. and while calpers has become a quieter activist in the last few years, other pension funds and mutual funds con tinue to aggressively monitor the firms in which they invest. in conclusion, our results indi cate that unless management is pressured into making substantial changes investors will not benefit from shareholder activism. 10 financial services review 7(1) 1998 references akhigbe, a., madura, j., & tucker, a. l. (1997). long-term valuation effects of shareholder activ ism. applied financial economics, 7, 567-573. anand, v. (1994, january 24). calpers gunning for poor performers, 3 companies being targeted by fund. pensions & investments, 22, 4, 80. dobrzynski, j. h. (1992, march 30). calpers is ready to roar, but will ceos listen. business week, 44--45. hwang, s. l. (1995, april 14). calpers to vote against board of philip morris. the wall street jour nal, b4. jensen, m. c., & meckling, w. h. (1976). theory of the firm: managerial behavior, agency costs and capital structure. journal of financial economics, 3, 305-360. kilman, s. (1988, may 13). holder resolutions against poison pills win more support at annual meet ings. the wall street journal, 4. nesbitt, s. (1994). long-term rewards from shareholder activism: a study of the calpers effect. journal of applied corporate finance, 6, 75-80. opler, t. c., & sokobin, j, (1995). does coordinated institutional activism work? an analysis of the activities of the council of institutional investors. working paper, ohio state university. rehfeld, b. (1997). low-cal calpers. institutional investor, 31, 41-49. schine, e. (1995, february 13). this gadfly is really buzzing calpers is issuing fresh reprimands to laggard boards. business week, 48--49. silverstein, s. (1991, april 18). calpers backs activist for seat on sears board. los angeles times, d2. smith, m. p. (1996). shareholder activism by institutional investors: evidence from calpers. jour nal of finance, 51, 227-252. strickland, d., wiles, k. w., & zenner, m. (1996). a requiem for the usa is small shareholder mon itoring effective. journal of financial economics, 40, 319-338. vise, d. a. (1993, january 28). new force in boardroom. the washington post, a1. wahal, s. (1996). public pension fund activism and firm performance. journal of financial and quantitative analysis, 31, 1-23. pii: s1057-0810(97)90026-3 volume 6 number 1 1997 financial services review the journal of individual financial management editor karen eilers lahey 1997 raj aggawal john carroll university james r. barth auburn university barry diskin florida state university thomas h. eyssell university of missouri, st. louis sandra g. gustavson university of georgia andrea hueson university of miami jeff madura florida atlantic university terry l. zivney ball state university associate editors 1997-1998 james s. ang florida state university lawrence j. gitman san diego state university douglas r. kahl university of akron james e. i..arsen wright state university dixie l. mills illinois state university tony plath university of north carolina at charlotte a. charlene sullivan purdue university jill lynn vihtelic saint mary's college 1997-1999 vickie bajtdsmit colorado state university waldo l. born eastern illinois university lawrence a. cox university of mississippi jean louis heck villanova university joan lamm-tennant villanova university travis pritcheu university of south carolina william reichcnsmin baylor university walter woerheide rochester institute of technology greenwich, connecticut @ j m press inc. london, england pii: s1057-0810(00)00051-2 from the special issue editor vickie l. bajtelsmit it is my great pleasure to have served as special issue editor for volume 9, number 1 of thefinancial services reviewand to offer you eight outstanding articles covering a variety of issues related to retirement. the topic of this issue “ensuring retirement income adequacy in the next century” is quite timely and has implications for all of us, whether or not we are doing research in retirement investment, pensions, or social security. the retirement of the baby boom is fast approaching and, despite the strength of the economy in recent years, there is much concern among policy-makers that this generation is financially ill-prepared for retirement. this is not an insignificant problem in light of the demographic changes facing this country and others worldwide. the aging of the population is predicted to place great strain on financial markets as the flow of investment funds reverses direction, and retirees gradually liquidate asset portfolios to fund cash flow in retirement. the social security administration predicts that, although we can expect many more years in which payroll tax inflows exceed benefit outflows while the baby boom enjoys their peak earning years, their retirement which begins in 2008 will eventually bankrupt the social security trust fund in the absence of significant reform, and this will occur sooner if politicians continue to look for ways to spend the surplus. in recognition of the “three-legged stool” concept of retirement planning, this issue of thefinancial services reviewaddresses problems and solutions in all three areas: private pensions, social security, and individual savings. how will individuals respond to a shortfall of financial resources as retirement approaches? a large percentage of u.s. workers do not have access to employer-provided pensions, and some cross-sections of individuals, most notably women, have significantly lower pension sponsorship and participation rates. furthermore, u.s. savings rates are low relative to other developed countries and studies have reported that the average household in the u.s. has too little retirement wealth accumulation. individuals approaching retirement under these circumstances have few choices. two of the articles in this issue address the timing of retirement. in “determinants of planned retirement age,” catherine montalto, yoonkyung yuh, and sherman hanna use the health and retirement study dataset covering older households to consider the factors that influence planned retirement age. although individuals may have plans to retire earlier while they are young, planned retirement age increases with age, perhaps an indication that inadequacy of retirement wealth is forcing financial services review 9 (2000) v–vii 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00051-2 longer working periods. walt woerheide takes a different approach to this question in his paper, “the impact of the pension on the decision to work one more year,” in which he estimates the net benefit to be received by working an additional year (salary plus additional pension accruals less foregone retirement benefits). in “beliefs and actions: expectations and savings decisions by older americans”, harold elder and patricia rudolph explore the interesting question of the inter-relationship between savings behavior, pension plan participation, and expectations with respect to retirement standard of living. their finding that individuals consider personal savings and pension plan savings to be substitutes is an important explanation for low savings rates in light of the increasing numbers of defined contribution plans and their financial performance over the last decade. however, the fact that individual saving, but not pensions, is found to have an affect on expected standard of living in retirement may imply that individuals do not fully understand the value of private pensions. in light of social security reform proposals that suggest creation of individual investment accounts and/or investment of social security trust fund in equities, several papers in this issue are particularly interesting. although previous studies have suggested that stocks generally have outperformed bonds in the long run, common wisdom has argued in favor of reducing equity exposure as retirement nears. in “the asset allocation decision in retirement: lessons from dollar-cost averaging,” premal vora and john mcginnis consider the retirement consumption effect of stock versus bond investing during the retirement period. employing a “dollar-cost disinvesting” methodology, they show that retirees over that last several decades would have consistently done better by investing in stocks. however, it is clear from the other papers in this issue that actual investment decisions by individuals may not produce optimal retirement outcomes. jack vanderhei and kelly olsen use a unique database which includes information on nearly thirty thousand 401(k) plans and more than six million participants in “social security investment accounts: lessons from participant directed 401(k) data” and apply their findings to an analysis of the advisability of giving investment control over social security accounts to individual participants. based on recent investment decisions by 401(k) participants, they conclude that a large percentage of individuals are very conservative in their investment choices, a finding that is consistent with that of doug waggle and basil englis who examine individual retirement account investments in their article “asset allocation decisions in retirement accounts: an all-ornothing proposition.” their results, based on a large survey sample of consumers, suggest that many individuals are totally undiversified and that a large percentage have their iras invested entirely in cash. whether or not individuals can be expected to optimally invest their individual social security accounts, there are potentially large macroeconomic spillover effects from shifting payroll taxes to equities. in “social security reform: the effect of investing in equities,” erick elder and larry holland provide a convincing argument that shifting social security funds from government securities to equities implies a decrease in the return on equities, an increase in bond yield and ultimately an increase in general income taxes. the final paper in this issue provides an interesting analysis of an alternative retirement savings vehicle. in “an analysis of the medical savings account as an alternative retirement savings vehicle,” j. tim query suggest that, while the medical savings account is not vi v.l. bajtelsmit / financial services review 9 (2000) v–vii a substitute for traditional retirement savings vehicles and pensions, individuals with low expected medical expenditures could accumulate significant account balances by the time of retirement. vickie l. bajtelsmit department of finance and real estate colorado state university fort collins, co 80523, usa viiv.l. bajtelsmit / financial services review 9 (2000) v–vii pii: 1057-0810(92)90007-y financial services review, 2(2): 131-156 copyright 0 1993 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the risks of pension plans robert w. mcleod sharon moody aaron phillips this paper identifies and describes the risks to which the prospective pensioner is exposed. an understanding of the types of plans and the risks associated with each will assist the individual pensioner with making a proper analysis of the safety of his/her-plan and acquaint the pensioner with the role of the employee retirement income security act (erisa) and the pension benefit guaranty corporation (pbcg) in safeguarding pension assets. i. intr~duc~~n when it comes to managing their own investment portfolios most individual investors are concerned with the amount of risk they will incur relative to the expected return. they are aware of the various risks inherent in most investment securities such as purchasing power, liquidity, marketability, portfolio, reinvest ment, and default. however, when it comes to their pension plans, many individuals do not allocate an adequate amount of time to understanding the risks inherent in their plans nor do they fully utilize or understand the options that may be offered to them. as a result many employees take their pensions for granted and assume that their pension benefits will be sufficient to meet their expenditures in retirement. many of the risks individuals face regarding their pensions arise from financial management decisions made by plan sponsors which directly impact the individual, but over which the individual has little or no control. the major risks to the plan participant include (either directly or indirectly) all of the risks normally associated with investments along with additional risks involving the selection of the type of plan that is offered to the employee, the level of funding for the plan, the investment robert w. mcleod l department of economics and finance, university of alabama, tuscaloosa, alabama 35487; sharon moody l department of accountancy, george washington university, washington, dc 20052; aaron phillips l department of finance and real estate, kogod college of business, washington, dc 20016. 132 financial services review, 2(2) 1993 selection and asset allocation, as well as the risks of plan termination plan modifi cation, and regulatory non-compliance. the focus of this article is to review the risks to the individual who participates in the various types of pension plans. we will begin with a review of private defined benefit pension plans and their associated risks and then move to public defined benefit plans. we will then shift our attention to defined contribution plans and conclude with a review of the government’s role in guaranteeing pension obligations. the last section will offer our conclusions and suggestions for further research. ii. private defined benefit pension plans all pension plans present the participant or beneficiary with some types of risk taking. the extent and types of risk, however, vary significantly between different plans because of the ability to shift various pension risks between employer and employee. the pension choice normally involves a decision to select either a defined benefit pension plan or a defined contribution pension plan. the defined benefit pension plan covers most individuals covered by pension plans (ippolito, 1985b, p. 1031) and is referred to as a formula plan because the benefit is customarily based upon some multiple of the employee’s earnings and years of employment with the sponsor. the participant in a defined benefit plan is generally more certain about the amount of the promised retirement benefit than in the case of a defined contribution plan because the participant can use the pension formula to estimate his/her promised benefit. even though the defined benefit plan appears to offer the participant more certainty as to the expected benefits, there are a number of risks inherent in such plans. the most important risks are the firm specific risks associated with the financial health of the sponsor and the risks associated with the funding status of the plan. the funding status is affected by the investment performance of the plan assets and the actuarial assumptions used in computing plan liabilities. in addition there are risks to the participant arising from plan terminations and regulatory compliance. firm specific risks the plan sponsor is liable for the promised pension, but only if it remains a viable concern. a sponsor which experiences financial distress is no longer in a position to make good on its pension promises. this means that the participant in a defined benefit plan is exposed to a great deal of unsystematic risk. each firm has its own business and financial risks associated with its operations. the ability of the firm to honor its pension obligations is based on its financial condition. the financial strength of the sponsor not only affects its ability to fund its pension obligations, but also puts the employee at risk for his/her job. if the plan sponsor were to experience financial distress resulting in bankruptcy, the loss of the employee’s job the risks of pension plans 133 coupled with the possible loss of pension benefits presents an extremely high level of risk th2t is not usually considered by the employee/plan participant. funding risk one of the biggest risks to participants in defined benefit pension plans is that of the bankruptcy of the plan sponsor coupled with 2 severely underfunded plan. in order to determine the amount of exposure to funding risk, the plan participant or his/her advisor would begin by reviewing the funding requirements for the plan 2nd then look at the disclosure requirements. determining funding status participants in defined benefit pension plans should be aware of the funding status of their plans. the employee retirement income security act of 1974 requires firms to report any underfunded accrued vested’ liability in excess of plan assets, thus providing plan participants with an indication of the relative safety of their plan. this reporting can be found in the corporation’s annual report 2nd 10-k filings 2s well 2s on the employer’s annual reports of employee benefit plans (form 5500). if the form 5500 is not available, the information can be requested from the summary annual report available from the plan administrator. in order to provide more uniform 2nd economically relevant information about the funding status of pensions, the financial accounting standards board (fasb) issued statement #36 in 1980, which required disclosure of the accrued pension liability 2nd the market value of plan assets 2s 2 footnote to the balance sheet. a major limitation to fasb 36 was that it did not consider future salary 2nd benefit increases in determining the present value of pension liabilities. also it allowed the use of 2 wide range of interest rates for this calculation. further changes in pension reporting were the result of the issuance of fasb statement #87 in 1985, which required that the underfunded liability 2ppe2.r on the balance sheet. in addition it required 2 footnote disclosing the pension liability with 2nd without consideration for projected salary increases. fasb 87 also reduced the flexibility of the plan sponsor in choosing discount rates for valuation of plan liabilities. funding policies up until the enactment of erisa, defined benefit plan participants were at risk for virtually the entire amount of the promised retirement income from the plan because there were no regulations mandating corporate funding, nor were there regulations for requiring disclosure of the funding level of the plan. however, 2s warshawsky (1989) points out, despite the funding requirements of erisa, the plan sponsors still retain considerable flexibility in determining the amount of their annual contributions to their pension plans. whether the pension is overfunded, fully 134 financial services review, 2(2) 1993 funded, or underfunded depends not only on the financial health of the sponsor, but also on the sponsor’s policy toward funding pension benefits within the maximum/minimum corridor established by the plan’s actuary. at this point a review of the theory of funding defined benefit pension plans will illustrate the sources of funding risk to the plan participant. sharpe (1976) established that corporate policy on the funding level of a pension may not matter, if the corporation is required to insure the plan’s promises. treynor (1977) observed that lacking insurance company willingness to provide the requisite coverage, the pension benefit guaranty corporation (pbgc) will become the insurer of benefit promises. the result is that “. . . it still pays the employer to make overgenerous pension promises”(p. 636). this is done without regard for the ability to deliver on these promises. the theory established by sharpe (1976) derives from a world with no taxation. the tax motivations for fully funding include (1) the contribution to the plan, up to the maximum allowable contribution, is a tax deductible business expense, thus requiring governmental taxing authorities to bear part of the cost of the pension contribution, (2) the amount in the pension plan grows at a non-taxed rate to the corporation, thus denying taxing authorities revenue on what otherwise would be considered corporate assets, and (3) the current earnings on pension plan assets represent saved future contributions by the corporation. black (1980) and tepper (198 1) independently establish that corporations should take advantage of the tax laws by fully funding their defined benefit pension plans and investing the pension fund’s assets in bonds or other fixed dollar investments. francis and reiter (1987) conclude that overfunding of pensions is motivated by tax benefits and the desire to store financial slack, while underfunding is driven by the desire to reduce debt costs through internal borrowing. malley and jayson (1986) state that the funding decision is influenced by the sponsors current financial position and investment opportunities. bulow (1992) notes that large underfunded liabilities are much more common in union defined benefit pension plans. ippolito (1985a) finds that this practice turns workers into bondholders of the firm, a behavior which further binds the worker to the sponsor. friedman (1983) believes that firms time their pension contributions to smooth earnings. however, some firms take advantage of the pension funding laws to minimize their investment in their pension plans so as to maximize their reported earnings and/or use of cash generated from operations that would have been channeled into the pension plan. this is due to the fact that the smaller amount that goes into the pension plan, the lower the charge against earnings for that period and the lower the cash drain. therefore, this provides an incentive to minimize the pension contribution when earnings are low or cash is short. for example, lockhart (1992) points out that twa never asked for a waiver of funding, never missed a payment and always made at least the minimum necessary contribution, yet at year end 1991 it was $900 million underfunded. the risks of pension plans 135 this method of underfunding essentially makes the firm a low-cost borrower from the other plan sponsors covered by the pbgc, since the pbgc requests premium increases on all sponsors when it faces a cash flow crisis. according to the pbgc’s executive director, james lockhart, approximately half the current premium “. . . represents a subsidy from the well-funded to the underfunded plans, creating a direct incentive to underfunding.“* the variation in the amount of a firm’s contribution to its pension plan can be explained by a number of reasons. if the contribution made in any period differs from that which is required on an actuarial basis, the plan will become something other than fully funded. however, even if the firm has consistently made the actuarially correct contribution, the plan could be underfunded due to the investment performance of assets in the plan. variation in the value of plan assets gives rise to funding risk and the types of assets in the plan give rise to this variation. we will now look at how investment risk affects funding status. investment risks the investment performance of plan asset managers is dependent upon asset allocation, security selection, and market performance. prior to the passage of erisa, one could assume that the pension liability was analogous to risky debt. using the approach of copeland and weston (1988) and assuming that the pension is uninsured and that this is an all equity firm, the end of period payoff to the pension beneficiary is shown in figure 1. the dollars of end of year payoff are on the vertical axis while the market value of the firm (v) and the market value of pension assets (a) are on the horizontal axis. v+a<b v+a>b figure i. end-of-period pension fund payoffs. 136 financial services review, 2(2) 1993 $ figure 2. the pension beneficiaries is equivalent to risky debt (long in a riskless bond and short in a put option). the plan participant will receive the promised benefit (b) only as long as the market value of the total assets, v + a, is greater than or equal to the promised benefit, b. the line oxb in figure 1 represents the payoff to the pension beneficiary. concentrating on the pension beneficiaries’ position, this can be modeled as the equivalent of owning a risk-free bond having an end-of-period value which is equal to b and selling a put option (p) on the assets of the firm. as can be seen from figure 2, the pension beneficiaries’ position is the sum of the payoffs on the risk-free bond and selling the put option which is the equivalent of holding risky debt. this is the same payoff as shown in figure 1. if pension insurance were considered, the payoff to the beneficiary would be as shown in figure 3 where gb is the pbgc guaranteed benefit and pb is the promised benefit. the horizontal axis is the funding ratio (fr) and the vertical axis is the dollars of payment. as can be seen the pension beneficiary would receive the pb as long as the plan is fully funded or better (fr 2 1 .o). when the fr < 1 .o the actual benefit would be less than the pb to the point where the guaranteed benefit begins. the payoff to the pension beneficiary resembles a collar with the cap as the pb and the floor the gb. this is equivalent to the purchase of a risk free bond, sale of a put, and purchase of another put with a lower exercise price. assuming that shareholders wealth is a call option on the assets of the firm, the level of investment risk or types of assets selected by pension managers can be explained. in the event that the firm is under financial distress and its pension plan is underfunded, the sponsor could change the asset mix of the pension in order to maximize the value of the call option. since the pbgc establishes a floor for the participants and the pbgc claim on the equity of the corporation in bankruptcy is the risks of pension plans 137 fr figure 3. payoff with pbgc guarantee. generally worthless, the optimal strategy from the point of view of the shareholders of a financially distressed company, according to copeland and weston (1988), would be to put all of the pension assets into very risky stocks. if the stocks performed well, the company could have an overfunded pension, if not the pbgc and the plan participants absorb the losses. however, if the firm were not in financial distress, they summarize that the all bond pension portfolio is preferable to share holders. treynor (1977) also addresses the issue of asset management of defined benefit plans relative to the role of the pbgc as guarantor. he concludes that due to the existence of the pbgc . . . “it still pays the employer to put heavy pressure on the manager of his pension funds to manage them aggressively” (p. 636). bodie (1992) states that the pbgc insurance creates an incentive for an underfunded pension plan to invest in risky assets due to the existence of a put option which increases in value with increases in the risk of the underlying portfolio. black (1980) and tepper (1981) as mentioned previously suggest fully funding and investing in bonds and other fixed income securities. contrasting the views of treynor and bodie (1992) to that of black and tepper suggests a difference in pension asset management. treynor’s and bodie’s conclu sions support aggressive management, implying equities, while black and tepper endorse the use of bonds. to integrate these views, bodie, et al. (1987) examined the proposition that pension assets and liabilities are parts of an extended balance sheet and that pension funding policy is integral to corporate financial policies. this approach would integrate the tax induced behavior of black and tepper with the insurance role of the pbgc, as perceived by treynor and bodie, and test the pension asset strategy. they find that the proportion of plan assets in fixed income securities is inversely related to funded status of the plan (plan assets to liabilities ratio) and sponsor’s bond rating which are two measures of corporate risk taking. this would 138 financial services review, 2(2) 1993 support the contention of bodie, et al. that the pension is an extension of the corporation. alderson (1990) examined the influence of the changes in pension regulation introduced by the omnibus budget reconciliation act of 1987 (obra) on the pension policies of defined benefit plan sponsors. he concluded that the tighter restrictions on the discretionary capabilities to terminate underfunded plans has transferred a greater share of the financial risks from the government to shareholders and ultimately to the plan participant. he concludes that this has increased the need for a lower risk policy in regard to investment selection and management of pension assets. as one would expect, the asset mix and, therefore, the investment performance of pension funds is of particular importance to the funding status of pension plans. over the 1987 to 1990 period, according to miller (1991), distributions from pension plans exceeded contributions with the net outflow in 1990 amounting to $13 billion. in order to compensate for this negative flow, the sponsor must make either greater contributions in the future (increase funding) or achieve greater investment performance (perhaps by incurring greater risks or by employing superior investment managers). prudence in investment management the pension trustees, or plan administrators, while employed by the plan sponsor, have a fiduciary responsibility for the plan. as fiduciaries, they are required to perform their duties for the sole benefit of plan participants and their beneficiaries; with the skill, care, prudence, and diligence under the circumstances prevailing that a prudent man3 acting in like capacity and familiar with such matters would use, by diversifying the investments of the plan to minimize losses in accordance with the documents and instruments governing the plan4 in many cases this would involve the retention of the services of a professional investment manager. however, hiring professional investment managers does not assure the plan sponsor or the participant of either good investment performance or a source of recovery for poor investment performance. for example, weyerhaeuser failed to win a judgment against one of its investment managers for poor performance in spite of the investment manager earning only $34.9 million on $2.52 billion under management over a three year period approximately, 0.5% per year, clearly indicating that the sponsor (and ultimately the employee) is the one at risk.5 whether a firm takes an aggressive or conservative investment approach with its pension assets would appear to be influenced by both the current funding status of the plan and the financial condition of the sponsor. the current funding status of the plan is determined by the actuarial assumptions used to calculate the pension liability. actuarial (valuation) risks the employer hires the enrolled actuary (ea) to perform an annual valuation report and to complete a schedule b for corporate tax reporting. the ea is an the risks of pension plans 139 independent professional bound by rules of the profession and by the irs. as a result of this annual valuation, the ea recommends an amount for the employer to contribute to the plan. this recommended contribution is within a band of actuarially acceptable minimum and maximum contributions. the requirements for the mini mum contribution are established by erisa and determined as follows: 1. all normal costs attributable to benefit claims deriving from employee services in a given year must be paid that year; 2. any experience losses (caused by a decline in the value of the securities in the fund, by unexpected changes in employee turnover, or by changes in actuarial assumptions about the discount rate) must be amortized over a period not to exceed 15 years; and 3. supplemental liabilities resulting from increased benefits or underfunded past service costs must be amortized over a period not to exceed 30 years (40 years for companies with pre-erisa supplemental liabilities).6 the maximum allowable contribution is determined by the internal revenue service regulations. the limit is the actuarially determined normal cost of the plan plus any amount which is necessary to amortize over ten years any experience and supplemental losses. despite the fact that erisa and the irs have established the permissible range of annual contributions to defined benefit plans, sponsors still retain considerable flexibility in determining the amount of the contribution to make to their pension plans within the parameters set by regulation. however, the assumptions which the ea has to make in order to complete the annual valuation are the basis of this flexibility. among the assumptions that affect the valuation and, therefore, the funding status of a plan are the interest rate used for discounting, the rate of wage inflation, and the rate of price inflation, participant mortality and morbidity. any changes in the actuarial assumptions and methods are included in the valuation report and schedule b. the ea must be prepared to justify any changes based on either plan experience or long term market expectations. actuarial changes would not only affect the present value of the pension fund liability, but also the shareholder’s expectations about the riskiness and level of future cash flows of the firm. following the approach of weston and copeland (1988), the value of shareholders wealth (s) is shown in (1) as the market value of the firm (v) minus the market value of its pension fund liabilities (pfl) and other debt (b). s=v-pfl-b (1) the market value of the pension fund liabilities is determined by the market value of pension plan assets (pa), the present value of expected contributions adjusted for tax benefits (pc), and the present value of expected pension fund benefits from past and future service (pb). this is shown in (2). 140 financial services review, 2(2) 1993 pfl=pb-pa-pc (2) of concern to us here is the difference between the book value of the pension fund deficit and the market value. the main cause of any difference is the discount rate used. the appropriate discount rate used for pension valuation has changed as a result of erda. prior to erisa the expected benefits (pb) would have been discounted at the firm’s cost of junior, or subordinated, debt because the pension benefit obligation was not considered to be a senior claim on the assets of the sponsoring corporation. with the passage of erisa the pension obligation became a senior obligation falling right behind tax liabilities in the order of distribution of assets. the result is that the present value of expected benefits is discounted at a lower discount rate, thereby increasing the value of the pension benefits. this change resulted in a wealth transfer to pension beneficiaries from shareholders as the relative increase in pb from (2) would increase pfl. this would in turn, everything else being equal, result in a decrease in s from (1). whether a plan is currently underfunded or overfunded depends on the present value of the future pension obligation relative to the present value of the pension plan assets. if a firm chooses a high assumed discount rate, the present value of the future obligation becomes less and requires lower contributions to the plan. (it should be noted that, if the cash available to the corporation from lower pension contributions is expensed, the effect on shareholder wealth is the same and changing the actuarial assumptions to change current contributions is futile at best.) current accounting practice, however, indicates that the discount rate applied must reflect the market determined rates available for investment. consequently, when interest rates are declining, present values of future pension obligations increase. this could turn an overfunded pension plan into an underfunded one, or’exacerbate an under funded situation. the ability to alter the discount rate used to calculate pension liabilities and, therefore, to affect the funding status of the pension plan is a risk to plan participants which is tempered somewhat by regulation and the professional standards of the actuary. lower rates make the present values of future pension payments larger (thus justifying larger current period contributions). as such, bodie et al. (1987) report that firms that are more profitable avail themselves of the tax advantages of funding pension plans by encouraging their plan actuaries to assume lower discount rates in determining the pension liability. this behavior is attributable partly to corporations building financial slack (bodie et al. (1987); francis and reiter (1987); and stone, 1987). correspondingly, firms with low profitability would encourage the use of greater discount rates in determining pension obligations and thereby reduce the firm’s cash outflow by reducing the size of the required pension payment. a low profit position would, accordingly, place the firm in a low tax position and, therefore, in a position to benefit only negligibly from the favorable tax treatment of the periodic pension expense. feldstein and merck (1983) have established that tirrns with the greatest pension obligations relative to smaller levels of pension assets assume the largest rates of return to be earned on those assets; therefore, those plans the risks of pension plans 141 which present the greatest risk to plan participants from being underfunded are also subject to the sponsor’s risk taking from assuming too large a rate of return. a study by schwimmer illustrates the effects of changing the discount rate on the funding status of various plans. by lowering the discount rate from 8% in 1990 to 7.75% in 1991 bell atlantic’s underfunded projected benefit’ liability increased by $385 million (32 percent), to a new total of $1.598 billion. vosti (1991a) found that a reduction of the discount rate from 9.65% in 1990 to 8.5% in 1991 increased chrysler corporation’s underfunded projected benefit obligation by $770 million to $4.39 billion (an increase of 21 percent). vosti (1991b) further states that by decreasing the rate from 10% in 1990 to 9.3% at the end of 1991, general motors’ (gm) underfunded liability increased to $8.6 billion. however, according to the pbgc’s calculations the underfunded liability for gm was $11.8 billion at the end of 1991. in interest rate related research, barrett and pfenenger (1989) suggest that pension liabilities should be discounted at the risk free rate of return. d’arcy and chen (1988) find that the stock market returns of firms’ reducing assumed discount rates outperform the market and firms raising discount rates under perform the market. they attribute this performance difference to the fact that since lower rates represent a more conservative (higher) value of the pension liability, the investment managers must become more aggressive in order to earn higher returns which would increase the value of the funds assets. warshawsky (1989) further examines the issue of how corporations remain within erisa guidelines and yet maintain different funding levels for their plans. he concludes that although investment performance can account for some of the difference, for the most part the choice of actuarial cost methods* and assumptions explains the variation in funding ratios. regardless of the actuarial method used, willinger (1992) proposes the use of a contingent claims model to reduce the possible variation in the reporting of the pension liability associated with the actuarial model. financial accounting standards board (fasb) statement no. 87 states that the assumed discount rate shall reflect the rates at which the pension benefits could be effectively settled. the discount rate fitting this description include a variety of choices which would provide for the possibility of manipulating current and historical pension obligations. he suggests that using the contingent claims model would limit the ability of the sponsor (in conjunction with the plan’s actuary) to manipulate the pension liability as previously discussed in this section. assumptions are important because they are the basis of all funding and liability calculations. each year the plan endures demographic changes such as turnover, early retirement, aging of the covered population, and hiring of new employees, as well as changes tied to assumptions such as asset appreciation. large fluctuations in demographics or assets can create volatility in the normal cost and the unfunded liability projections from one year to the next. so the actual experience of the plan can impact the normal cost of funding and the stability of the plan. 142 financial services review, 2(2) 1993 as has been discussed, the selection of actuarial assumptions and methods affects the funding status of the defined benefit plan. since these assumptions are made by plan actuaries, the participant is at risk for any errors in these assumptions that would result in an underfunded position for firms experiencing financial distress. while the effects of actuarial assumptions on funding status are important there are other factors which affect the risk of defined benefit plans. among these are the risk of plan termination and non-compliance. plan termination risk overfunding a pension plan does not necessarily reduce all of the risks faced by participants in defined benefit plans. many corporations in an attempt to reduce future funding costs or to recapture excess pension assets have terminated their defined benefit plans. the issue of plan termination is of interest to plan participants for two reasons, both of which connote pension risk. first, as mittelstaedt and reiger (1993) observe, when terminating defined benefit plans fums must satisfy vested and previously non-vested liabilities “. . . using salaries in effect at the legal date of termination.“(p. 3) the ramification is that salary progression ceases and the future retiree will receive benefits calculated upon the salary level in place at the termina tion date and not at retirement. the second plan termination risk is really a type of investment risk and results from the firm purchasing annuities to satisfy the legal obligation to the participant. once the annuity is purchased this form of benefit promise falls outside of the realm of the pbgc and is not guaranteed by them. in addition the plan sponsor does not have any liability for the purchase of annuities from an insurance company that subsequently fails according to pbgc opinion letter no. 91-4. table 1. defined benefit pension plan terminations number of terminations by type of termination (1) number of dollar amount of a b c d total participants reversions (2) year 1986 43 12 84 50 249 261,769 $4,284.3 1987 28 70 106 72 276 235,826 1,954.9 1988 29 33 127 73 262 272,107 2,206.2 1989 13 23 97 39 172 161,220 843.5 1990 6 17 38 12 73 55,537 304.2 total 119 215 452 246 1032 986,459 $9,593.1 notes: (1) pbgc classifies successor plan to a termination as a = spinoff b = defined benefit plan (complete termination) c = defined contribution plan (complete termination) d = no new plan (complete termination) (2) in millions source: “seppaa completed reversion cases,” pbgc, february 7, 1991. the risks of pension plans 143 if the plan is terminated and the participant receives a paid up annuity, the safety of the pension is solely a function of the ability of the insurance company to fulfill its obligations. zall (1992) explains the magnitude of this risk by showing that 170 life insurance companies failed from 1975 to 1990 with forty percent failing in 1989 and 1990. (it should be noted that only four major insurers have failed.) zall goes on to mention that more than 300 defined benefit plans have been terminated which purchased annuities from insurers with questionable financial condition. as a result of these terminations, there is a growing exposure of plan participants to life insurance company failure. the sources of protection to plan participants associated with a failed insurance company are a recently enacted rule by the pbgc in june, 1992 requiring firms to give participants 45 days notice prior to any distribution of assets of the names of the insurers which are being considered for providing annuities to replace the pension benefits and state guarantee associations. the pbgc rule arose from the failure of executive life, which had provided annuities to many corporations that had terminated defined benefit plans.’ corporations also have been terminating defined benefit pension plans to recapture the value of excess contributions to the plan. this coincides with stone’s (1987) observation that firms build financial slack in their pension plans and then terminate them to access this slack. one of the problems in this termination process is that rarely do firms replace the terminated defined benefit pension plan with another defined benefit plan. as table 1 indicates, over the period 1986 to 1990, 1032 defined benefit pension plans were terminated and only 215 (20.8%) were replaced with another defined benefit plan. of the remaining terminations, 246 (23.8%) were not replaced at all and 452 (43.8%) were replaced with defined contribution plans. the recapture by corporations of over $9.5 billion is sufficient motivation for this action, but over 986,000 employees were affected in the process. mittelstaedt and reiger (1993) find that firms terminating defined benefit pension plans and replacing them with defined contribution pension plans show excess positive stock returns, thereby documenting a wealth transfer from pension plan participants to stockholders. the trend away from offering defined benefit plans as reported in table 1 was also noted by lockhart (1990), that in the 1980s the percentage of employees covered by defined benefit plans declined from 80 percent to under 70 percent. compliance risk another risk that participants face is that the plan is found to be in violation of anti-discrimination laws. such non-compliance occurs when the internal reve nue service (irs) deems that a particular pension plan does not equitably treat all participants or potential participants. if an existing plan fails discrimination testing, the contributions to the pension plan or the earnings from the plan are immediately taxable income to the participants. the amount of the contribution or the earnings that are taxable depends upon a complex formula of vesting. 144 financial services review, 2(2) 1993 compliance with changes in pension regulations involving integration of pension benefits with social security can have an effect on the funding status of a defined benefit plan. currently almost 60% of participants in defined benefit pension plans have their yearly retirement benefit integrated with social security (see maher and ketz, 1991). however, in response to the tax reform act of 1986 which addressed the discriminatory nature of integration with social security in regards to lower paid employees, the irs implemented new rules in 1991 to limit the adverse effects. for those plans that continue to use the social security offset there will continue to be the problem of estimating the social security portion of the participants benefit so as to arrive at the sponsors contribution to the plan which could result in either overfunding or underfunding. in addition this adds a new dimension to the risks to the participant and the sponsor based on possible changes in eligibility requirements for social security. there are additional risks for those employees participating in a defined benefit plan sponsored by a relatively small employer because of the nature of the employer and the regulation imposed on it. the small employer has more firm specific risk, and is subject to more restrictions (called top heavy rules) on benefit variations among employees. these extra restrictions increase the risks of non compliance of the plan. additional plan risks the defined benefit pension plan participant faces risk on three additional fronts: changes in taxation of pensions, use of the sponsor’s own securities to fund the plan, and revisions to the plan document. in regards to the first risk, there is concern that with a new president and a congress intent on reducing the deficit, there may be additional limitations placed on the level of tax deductible contribu tions to defined benefit pension plans. any change in deductible funding rate would probably result in a reduction in contributions for at least the short term as companies reassessed their costs of contributing to such plans. in the long run, such changes could have the effect of reducing the attractiveness of such plans and accelerating the conversion of defined benefit plans into defined contribution plans or eliminat ing retirement plans altogether. a second area involves the use of the corporation’s own securities to fund the plan contributions which is a permissible practice under erisa. however, it exposes the participant to an additional level of risk. in the event the sponsor bankrupts, not only would the employee lose his/her job, the pension plan would contain the now worthless securities of the sponsor, jeopardizing the employee’s pension benefits. even if the sponsor does not bankrupt, there is a lack of personal diversification on the part of the employee whose current and future fortunes are tied to the sponsor. the contribution of a firm’s securities to its pension plan is typical when cash flow problems arise. according to vosti (1992b) general motors (gm), responding to cash flow needs, in 1992 planned to contribute $500 million of its own common the risks of pension plans 145 stock to reduce its pension liability. gm had not made a pension plan contribution since 1987 due to credits it had built up. however, its 1987 contribution was $1.04 billion of its own new preferred stock. while there is no risk to participants if the firm remains financially healthy, chernoff (1992) notes that the pbgc did inherit $26 million of worthless preference stock from wheeling-pittsburgh steel when the pbgc took over that plan. another aspect of risks of defined benefit plans not previously covered is that of an amendment to the plan. although it is through this mechanism that retiree benefits are frequently raised, the plan participant should be aware of the risk that the adjustment might result in a reduction of future benefits. albert and schelberg (1992) illustrate this risk by referring to kreutzer versus the a.o. smith corporation. in this case the company had amended the severance benefit formula printed in the supervisor’s manual, but failed to notify the supervisors. seven supervisors re quested benefits according to the original formula and were denied. in subsequent litigation the court ruled that the firm had not acted in bad faith and had not tried to conceal the changed benefit formula. consequently, the supervisors were not entitled to benefits other than those provided under the newer formula. the plan participant is clearly required to stay abreast of any changes the sponsor makes. iii. public defined benefit pension plans while erisa provides some security for corporate sponsored defined benefit plans, it does not cover public employee plans. this means that in addition to having all of the risks associated with a private defined benefit plan, public plans are not subject to federally mandated minimum funding requirements or vesting schedules. the following sections will cover the differences in risk between public and private defined benefit plans. funding risk as a result of the lack of regulation concerning the level of funding of public defined benefit plans, their funding ratios can vary more widely than those of private plans. according to clark (1991b) many public defined benefit plans are operated on a pay-as-you-go basis and are severely underfunded. for example, the massa chusetts system is less than 20 percent funded. other states are required by state law to be fully funded. this necessity to fully fund creates political problems for municipalities during economically challenging periods. actuarial risks clark (199 1) also points out that public funds have been creative in an attempt to alter their funding status during the 1990-1992 economic slowdown. many plans 146 financial services review, 2(2) 1993 have adjusted the assumed rate of return to be earned on plan assets for purposes of reducing the present value of future obligations. while corporations were lowering their discount rates to more nearly reflect current market rates of return, many public employee funds have been increasing their rates to diminish obligations. as an example of the effects of this type of activity clark (1991) mentions that the metropolitan transit authority of new york city (mta) was faced with an operating shortfall and did not want to increase fares. the mta changed its rate of return assumption on plan assets from 8.25% to 9%, thereby reducing its pension contribution by $40 million. new york is not alone, however. among the twelve public funds increasing discount rates, louisiana increased its 1991 rate to 8.3% from 7.5% to save $24 million and california proposed to increase its 1991 rate to 9.5% from 8.5% to save $300 million. plan alterations in addition to changing actuarial assumptions, public defined contribution plans have shifted the burden of funding to employees. in a study by vosti (1992c), arizona, for example, in 1989 reduced its cash outflow by dropping its contribution rate to 2 percent of salary from the previous 4.7% level. minnesota reduced its contribution rate from 4.5% to 4.3%. these reductions in contributions, if not offset by increases in employee contributions, are attained at the potential expense of the state taxpayers, who will have to make up shortfalls through increased taxes in future years. to limit this future confrontation with taxpayers, states are looking at alternatives. one such alternative is adoption of defined contribution plans for new employees. colorado and oklahoma each approved defined contribution plans for municipalities and universities in their states. trends in public defined benefit plans wentz, et al. (1991) point out that there were three trends which became evident in the 1980s which would have an adverse impact on public defined benefit pension plans. these trends were the ending of the economic expansion of the 1980s; the advent of lower interest rates; and the expansion in the number of state and local government employees. due to these events the funding for pension plans in many states has been adversely affected. however, this underfunding is not necessarily apparent to the public (even though disclosure is required under government accounting standards board (gasb) statement no. 5). this is due to the fact that many of the states increased their assumed rate of return on pension assets thereby minimizing the underfunding of the plans. public pension plans may soon be following the lead of private corporations in shifting more of the burden of providing retirement benefits to the employee by terminating their defined benefit plans and adopting defined contribution plans. the risks of pension plans 147 iv. defined contribution plans defined contribution plans would generally include 401(k) and profit-sharing plans for private employees, 457 plans for employees of state and local government, and 403(b) plans for non-profit groups such as teachers and hospital employees. in this type of plan the sponsor and/or the employee make deposits into the plan on a regular basis. at retirement, the plan participants will have a given pool of funds to draw upon for their retirement needs. funding risk defined contribution plans promise no fixed benefit as the ultimate retirement benefit is determined by the rate of return earned on funds contributed to his/her account and the amount contributed. by definition defined contribution plans are always fully funded and, therefore, there is no funding risk as in the case of defined benefit plans. in essence, the sponsor has shifted all of the risks to the employee. however, there is the possibility that the sponsor would not put the required contribution into the pension, but that would be obvious to the participant upon receipt of his/her pension statement which records all contributions and earnings performance. investment risks the participants in defined contribution plans are captive to the rates of return earned by the investment options in the plan. in addition the participant in many cases is responsible for making asset allocation/investment decisions for his/her contributions through the selection from among a number of investment options. once the initial decisions are made as to the location of contributions, the participant is still directly involved in the investment decision making process through the ability to transfer from one investment to another within the plan. this responsibility for asset allocation or rebalancing the portfolio is generally far beyond the capabili ties of most employees and usually results in the continuation of the initial invest ment selection. with the increase in defined contribution plans and the greater responsibility placed on the employee to plan for his/her retirement, obviously comes greater risks along with the opportunity for greater returns. unfortunately, in cases where this involves the selection of investment options, it would appear that employees are not investing their money wisely. according to schultz (1992a) three-fourths of em ployees with self-directed defined contribution plans have no investment in stocks. there appears to be a tendency toward being too conservative in the investment of these funds. this investment strategy is not one that offers much hope of protecting the purchasing power of the pension. in order to protect the employees from themselves, the plan trustees have a fiduciary responsibility to uphold. if they do not, they may be liable for damages incurred by the participants. but what of the case where the employee makes the 148 financial services review, 2(2) 1993 selection of the investment vehicle(s) for his/her contribution? are the trustees still held responsible for bad investment decisions on the part of participants? the section on fiduciary responsibilities addresses this issue. liquidity risks another risk that occurs in some defined contribution plans comes about due to provisions in the plan document allowing for loans to plan participants. if the participant has this capability and exercises this option the plan is obligated to provide the funds. if this were to occur at a time of rising interest rates or falling stock prices, the plan may have to liquidate securities at unfavorable prices and could incur losses on the pension portfolio. as a result of this risk, the plan would have to hold larger amounts in liquid assets which would normally result in lower invest ment returns. as an option, the plan trustees may have the ability to borrow to meet this liquidity need in order to avoid selling securities at an unfavorable price. however, the plan would still incur the cost of borrowing. fiduciary responsibilities in an attempt to define the responsibilities of the plan fiduciaries in the case where the participant has the ability to choose among a number of investment options, the department of labor has issued regulations under erisa section 404(c) which limits the liability of plan fiduciaries who satisfy the requirements of this section. according to buck consultants (1992) the key provisions for compliance with section 404(c) are that the participant has the ability to exercise meaningful control over the assets in the account. in order to meet the requirements of having mean ingful control the participant must have the opportunity to: 1. choose from a broad range of investment alternatives. 2. give investment direction with respect to each investment alternative available under the plan with a frequency which is appropriate in light of its market volatility. 3. diversify investments within and among investment alternatives. 4. receive sufficient information to make informed decisions. a plan is considered to offer a broad range of investment alternatives only if it allows the participant to diversify investments in order to minimize the potential risk of large losses; materially affect the potential return on assets under the participants control; and choose from at least three investment alternatives each significantly different in its risk and return characteristics. if section 404(c) regulations are followed, then the plan fiduciary is exempted from certain liability for investment losses incurred by the plan participant. for plans that qualify under section 404(c) the risk of loss in the defined contribution plan rests solely on the participant. the risks of pension plans 149 risks of guaranteed investment contracts (gics) with the growth of defined contribution pension plans came the problem of selection of investment options within the plans. many trustees choose to offer gics offered by insurance companies. according to the general accounting office 28 percent of defined contribution plan assets are currently invested in gics.” due to the failure of executive life and other insurers there has been increas ing scrutiny of the safety of investments in pension plans. while the term guaranteed investment contract sounds good, the guarantee is only as good as the insurance company which is backing the gic. although it is true that to date no retiree has lost benefits due to an insurance company failure, one has to be concerned about the safety of gics. according to a study by todd and wallace (1992) currently only 17 states explicitly guarantee gics with two other states under court order to do so. in 14 states neither the guaranty fund nor the courts say whether gics are guaran teed. puerto rico and 17 of the states have laws which explicitly deny coverage to gics. however, even if there are explicit state guarantees, the pension beneficiary should be aware of another set of risks involving the funding status of the guaranty fund and the maximum guarantees provided. v. the roleofthe government inreducingpensionrisk erda in its desire to provide a safety net for at least a portion of the employees retirement benefits congress passed the employee retirement income security act (erisa) in 1974 which set mandatory funding requirements for defined benefit pension plans, established minimum vesting requirements, created the enrolled actuary (ea) designation, and required that the ea provide an independent valu ation and tax report of the pension. among the reasons for the passage of erisa were actions by employers, who offered defined benefit plans, such as firing employees right before they would have become vested in their pension benefits. this practice placed the employee at risk for his/her pension right up to the vesting date which was usually the normal retirement date. some of this uncertainty was eliminated with passage of the erisa and the tax reform act of 1986 which requires all pension plans subject to erisa to vest their employees according to one of the following schedules: 1. cliff vesting: employees are fully vested after five years of service. 2. graded vesting: employees are vested 20% after 3 years of service and then increasing at the rate of 20% per year to become fully vested after seven years of service. 150 financial services review, 2(2) 1993 even though there are mandated vesting schedules, this does not mean that all employees are covered by the pension. the employee must meet certain criteria before he/she is eligible to participate in the plan. irs rules allow the plan to exclude new employees under the age of 25, with less than one year of service, or with less than 5 years until normal retirement age defined in the plan. plans may require that each participant work a certain number of hours during the plan year, such as 1000 or 2000 hours, in order to be eligible. this may exclude many part time employees from the plan. provisions for handling breaks in service are also included in plan specifications. therefore, an employee is at risk of losing covered status for the year, due for example, to a change in the number of hours worked or to an uncovered break in service. pension benefit guaranty corporation erisa also created the pension benefit guaranty corporation (pbgc). the pbgc is a government sponsored insurance plan designed along the lines of the federal deposit insurance corporation and operated under the u.s. department of labor. it obtains its funds by assessing plan sponsors an annual premium per covered employee. the current premium assessed by the pbgc for single employer plans is the sum of a flat premium of $19 per employee plus a variable premium of $9 per $1,000 of underfunded vested liability (subject to a cap of $53 per participant). funding status is determined as the ratio of pension plan assets to accrued benefits, both those vested and those earned but not vested. the pbgc acts as an insurer as long as the firm remains financially viable and maintains minimum contributions to its pension plan. the actual behavior of the pbgc is more like deposit insurance. as more firms’ underfunded plans are taken over by the pbgc, it must raise premiums, which started at only $1 per employee when erisa was first passed. according to lockhart (1992) the pbgc’s premium receipts are currently sufficient to pay the benefits received by retirees of the 1650 plans the pbgc currently administers. however, if there were an increase in the number of plans taken over by the pbgc combined with an increasing number of retirees, its cash flow would become negative in a very short period of time. the existence of the pbgc does not eliminate all of the default risk of the plan participants. in order to protect itself, the pbgc has authority to terminate a defined benefit plan in the event the sponsor is deemed to be financially unsound, thereby limiting the accumulation of further benefit liability. in such a termination the pensioner should be aware that the pbgc does not insure or guarantee all benefits in a covered plan. it covers basic benefits only. therefore upon termination of a plan, the pbgc takes over administration of the plan and makes pension payments in accordance with the terminated plan’s agreement. however, these payments are limited to a maximum guaranteed monthly benefit which is adjusted annually with inflation.” for plans which provide employees with cost of living adjustment (cola) clauses and with insurance benefits, there is even more risk of loss due to the lack of pbgc coverage of these benefits. the risks of pension plans 151 current risks to the pbgc the pension funding problems associated with the bankruptcy of ltv and other plan sponsors are illustrative of potential liability of the pbgc for coverage of underfunded pension plans which it insures. according to lockhart (1991) of the total underfunding of $40 billion in defined benefit plans, $13 billion is associated with financially troubled firms which present a serious risk to the pbgc. this represents an increase of 75% over the previous period. estimates for 1992 are that the level of underfunded liabilities will be approximately $43 billion according to estimates by the office of management and budget found in abken (1992). the 50 top companies with the largest underfunded pension liabilities as of the end of 1991 are shown in exhibit 1. as you can see from exhibit 1 the underfunded liability of these fifty compa nies exceeded $29.4 billion and the average funding ratio was 71%. if just the guaranteed benefits were considered, the amount of underfunding was “only” $24.2 billion.‘* the future of the pbgc according to abken (1992) the problems with pension guarantees may seem like a repeat of the recently experienced deposit insurance difficulties. however, the problems with the pbgc are not new as the net-worth deficits have persisted since its creation. in fact many of the top 50 underfunded companies in exhibit 1 have been on the list for decades. the major issue is whether or not we have learned from the deposit insurance problems and can apply what we have learned to the pension insurance problem before it becomes a taxpayer’s nightmare. there are a few particular issues currently facing the pbgc which will affect the riskiness of the guarantee provided to covered pensions. one of these issues is the contribution of non-cash assets (except for employer stock and diversified real estate holdings) unless a specific exemption is approved. chemoff(l992b) points out that under erisa the reason that a firm cannot contribute non-cash assets is the concern of the pbgc that it would eventually be saddled over-valued assets which have limited marketability. however, as one might expect plan sponsors want more flexibility in the selection of assets for contribution to the pension plan. as a result litigation has been filed to challenge the pbgc’s restrictions and the u.s. supreme court has agreed to consider the case. if the supreme court relaxes these restrictions, then the plan participants may incur greater risk due to the potential for the plan to become overloaded with assets with limited marketability or inflated values. a broader question concerning the future of the pbgc is its own funding status. the pbgc currently operates on a cash basis, essentially a pay-as-you-go process. this method fails to consider the long term impact of pbgc obligations over time. as an example, lockhart (1992) noted that when pan am’s plans were terminated, the loss to the pbgc was over $600 million while the loss reported as part of the federal budget was only $10 million. chemoff (1992a) estimated that the 152 financial services review, 2(2) 1993 exhibit 1. top 50 companies with the largest unfunded pension liability unfunded unfunded guaranteed benejit benefit total liability guaranteed liability benefit (funding benefit (funding company name assets liability ratio) liabiltiy ratio) ravenswood aluminum corn. $ 10 $ 90 $ 80(11%) $ 85 s 75 (12%) ltv corp. morrell(john) & co. uniroyal goodrich tire co. keystone consolidated ind. new valley corp. loews corp. sharon steel corp. la&de steel carter hawley hale chrysler corp. american national can borg-warner bridgestone-firestone rockwell international anchor glass co. national intergroup tram world airlines occidental petroleum corp. acf industries budd co. cyclops industries inc. tenneco inc. bethlehem steel corp. white consolidated industries foxboro co. crown cork & seal co. inc. varity alleghenyludium goodrich (b.f.) navistar international james river corp. maxxam inc. general motors corn clark eouinment ’ northwest -airlines ast holding reynolds m&b occre & co. rohr inc. honeywell inc. sx corp. rjr nabisco holdings corp. goodyear tire & rubber co. burlington northern pacificorp national steel corp. westinghouse electric armco steel l.p. kimberly-clark corp. 425 3;; 85 331 129 122 i? 4,855 516 121 305 437 159 434 592 115 124 333 275 191 3,492 230 129 528 243 313 490 2,050 214 617 38,903 234 357 199 542 1,185 326 367 815 727 1,146 43 4 478 482 4,275 672 780 3,415 124 945 205 712 277 243 132 151 8,874 929 217 546 797 281 758 986 187 199 520 424 294 5,347 345 189 723 332 426 661 2,759 281 810 50,730 305 460 257 691 1,613 406 456 1,008 897 1,402 523 578 572 4,999 867 853 2,990(12) ’ 73 (41) 554 (41) 120 (42) 381 (46) 147 (47) 121 (50) 63 (53) 70 (54) 4,019 (55) 4&! [zzj 241 (56) 360 (55) 122 (57) zz: i;:; ;z i:;{ ;:z $z; 103 (65) 1,854 (65) 116 (67) 60 (68) ‘:; gz; 113 (73) 171 (74) 709 (74) 68 (76) 193 (76) 11,827 (77) 71 (77) 102 (78) 58 (77) 149 (78) 4;y {ii; 90 (80) 193 (81) 169 (81) 2;; t:;; 100 (83) 90 (84) 724 (86) 195 (78) 73 (91) 3,244 118 898 195 677 263 231 126 143 8,430 883 206 519 741 267 721 936 177 189 494 403 279 5,079 328 180 687 315 405 628 2,621 267 769 48,194 290 437 244 657 1,417 386 433 958 852 1,332 497 549 543 4,749 737 838 2,819 (13) ’ 67 (43) :g ::; 346 (49) :; :z; 2; iz; 3s;; ;;;; 85 (59) 213 (59) 304 (59) 108 (60) 287 (60) 344 (63) 62 (65) 65 (66) 162 (67) 128 (68) 88 (68) 1,587 (69) 98 (70) 51 (72) ‘z ;;;; 92 (77) 138 (78) 5z: {z{ 152 (80) 9,291 (81) 55 (81) :z ifi{ 115 (83) 2:: r::; 67 (85) 143 (85) 125 (85) 186 (86) 66 (87) 71 (87) 62 (89) 474 (90) totals (in millions) $70,375 $99,795 $29,420 (7 1%) $94,615 $24,239 (74%) the risks of pension plans 153 pbgc’s present value of future liabilities is $43 billion although its most recently reported annual loss was only $2.5 billion for 1991. finally, there is the question of the pbgc’s status in bankruptcy court. clearly, the ltv case indicates corporations cannot arbitrarily claim the inability to fund existing plans, put them to the pbgc, and then initiate new plans. the bankruptcy status of the pbgc, is as a dual class citizen. the pbgc has a priority claim for the guaranteed benefits. for those plan participants due more benefits than the pbgc guaranteed maximum, legislation in 1987 allows the pbgc to share recovered assets with workers for non-guaranteed benefits. however, zall (1992) notes that historically the pbgc only recovers through bankruptcy courts 10 to 15% of assets claimed. lockhart (1992) points out that approximately 86% of the pbgc’s claims are in the general unsecured creditor status. vi. s~jmmaryandconclusions employees who are participants in retirement plans offered by their employers are subject to a number risks. the type and degree of risk varies depending on the type of retirement plan-defined benefit or defined contribution. in the private defined benefit plan the participant is subject to the risk that the sponsor of an underfunded plan bankrupts or the plan is terminated by the pbgc. the plan could be under funded because of inadequate contributions due to inaccurate actuarial assumptions or because of poor investment performance. in any event the participant may not receive all the pension to which he/she was entitled. a fully funded plan can also be terminated by the sponsor and benefits to the participant distributed in the form of a paid up annuity. in this case the pbgc is no longer involved and the participant is at risk in the event of the failure of the insurance company which wrote the annuity. private defined benefit plans basically have all of the risks of private plans perhaps with the exception that the risk of bankruptcy of the sponsor may be less. the major risk that the participant faces is that of the shifting of the burden of the pension obligation to the employee or the termination of the defined benefit plan in favor of a defined contribution plan. defined contribution plans present the participant with almost all of the same risks as those faced by an individual investor. the exception is in the selection of assets to include in the plan, which are limited by the plan trustees or in some cases by erisa. erisa and the pbgc provide limited safeguards to plan participants by guaranteeing a basic benefit up to a maximum adjusted for inflation and setting fiduciary standards for plan sponsors and trustees and charging penalty rates on premiums for underfunded plans. in addition for defined contribution plans erisa encourages plan sponsors to provide flexibility in self-directed plans through section 404(c) by limiting liability to the plan sponsors and trustees. a number of issues concerning pension funds deserve further attention of researchers. first of all, since one of the greatest risks to participants of corporate 154 financial services review, 2(2) 1993 defined benefit plans comes from the inability of the sponsor to make contributions due to financial distress, research should be focused on how the individual can hedge this risk using options and futures contracts. a second area of research involves the determination of the optimal number and type of asset classes to be offered in self-directed defined benefit pension plans. the requirements of erisa section 404(c) only state that the number of options must be at least three and that each alternative must also be diversified and has materially different risk and return characteristics. the question remains as to what is materially different. finally, there would appear to be a need to develop guidelines for the selection of insurance companies which provide annuities to plans which are terminated. under the current pbgc regulations, there is no further involvement on its part once the plan is terminated and the annuities are purchased. in addition under current law the plan sponsor is not liable in the event the insurance suffers financial difficulties in the future which would jeopardize the participants pension benefits. therefore, establishing criteria for the selection of the insurance company would be a desirable research project. notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. accrued vested benefits are those benefits which the participant has earned by satisfying a minimum employment period criterion, and which have become an obligation of the firm, regardless of whether or not the plan participant remains with the firm beyond that time period. these comments by mr. lockhart are found in “grey peril,” the economist, may 11,1991, p. 78. the prudent man rule is being replaced by the prudent investor rule as a result of work by the american law institute which resulted in a restatement of the common law of trust investments. erisa section 404(a)( l)(a-d). for a discussion of the weyerhauser suit see “verdict inconclusive in manager lawsuit,” pensions & investments, march 30, 1992, p. 4. copeland and weston (1988, p. 642). projected benefit obligation exceeds the accumulated benefit obligation because it factors in periodic assumed increases in salary, thereby raising the pension liability above the accumulated benefit obligation, which does not include a provision for salary increases. for a discussion of actuarial costs methods see winklevoss (1977). for a further discussion of this requirement see “annuity notice to be required,” employment benefit plan review, december 199 1, pp. 52-53. “retirement plans,” employee benejit review, june 1992, p. 68. the 1992 guaranteed maximum benefit is $2,353.27. the source of this information is “news,” pension benefit guaranty corporation, november 19, 1992. abken, peter a. 1992. “corporate pension and government insurance: deja vu all over again?,” economic review, federal reserve bank of atlanta, march/april: 1-16. the risks of pension plans 155 albert, rory j. and neal s. schelberg. 1992. “courts bar recovery for erisa procedural violations,” pension world, 28 (may): 27-28. alderson, michael j. 1990. “corporate pension policy under obra 1987,” finunciul management, 19: 86-97. barrett, w. brian and john w. phenenger, ii. 1989. “proper cash-flow discounting for pension fund liabilities,” financial analysts journal, (march/april): 68-70. black, fischer. 1980. “the tax consequences of long-run pension policy,” financial analysts journal, (july/august): 21-28. bodie, zvi, 1990, “pension funds and financial innovation,” financial management, 19: 1 l-22. bodie, zvi, jay 0. light, randall merck, and robert a. taggart, jr. 1987. “funding and asset allocation in corporate pension plans: an empirical investigation,” issues in pension eco nomics, bodie, shoven, and wise, (eds.) chicago, il: university of chicago press. bulow, jeremy i. 1982. “what are corporate pension liabilities?,” quarterly journal of economics, (august): 435452. buck consultants, inc. 1992. for your information, november 3,1992, pp. 1-6. chemoff, joel. 1992a. “pbgc in accounting fray,” pensions & investments, (march 2): 6. chemoff, joel. 1992b. “supreme court to rule on non-cash assets,” pensions & investments, (june 22): 39. d’arcy, stephen p. and k.c. chen. 1988. “the effect of changes in pension plan interest rate assumptions on security prices,” journal of economics and business, 40: 243-252. clark, stephen e. 1991. “are public funds mortgaging their future. 7,” institutional investor, (july 25): 61. copeland, thomas e. and j. fred weston. 1988. financial theory and policy, third edition, reading, ma, addison-wesley publishing, pp. 638-656. feldstein, martin and randall merck. 1983, “pension funds and the value of equities,” financial analysts journal, (september/october): 29-39. francis, j. and s. reiter, 1987, “determinants of corporate pension funding strategy,” journul of accounting and economics, (january): 35-59. friedman, b. 1983. “pension funding, pension asset allocation and corporate finance,” in financial aspects of the united states pension system, z. bodie and j. shovens @is.), chicago, university of chicago press, 1983, pp. 107-152. ippolito, richard a., 1985a. “the economic function of under-funded pension plans,” journal of law & economics, (october): 61165 1. ippolito, richard a., 1985b. “the labor contract and true economic pension liabilities,“americun economic review, (december): 1031-1043. lockhart, james b. 1992. ‘securing the pension promise,” vitalspeeches ofthe day, (may 15): 472-3. maher, john j. 1987. “pension obligations and the bond credit market: an empirical analysis of accounting numbers,” accounting review, (october): 785-798. maher, john j. and j. edward ketz. 1991. “defined-benefit versus defined-contribution pension plans: how to compare,” compensation & benefits review, (may/june): 49-56. malley, s. and s. jayson. 1986. “why do financial executives manage pension funds the way they do?,” financial analysts journal, (november/december): 56-62. miller, lori. 1991. “cash flow problem worsens for corporates,” pensions & investments, (may 20): 91. mittelstaedt, h. fred and philip r. regier. 1993. “the market response to pension plan termina tions,” accounting review, (january): l-27. oldfield, george s. 1977. “financial aspects of the private pension system,” journal ofmoney credit & bunking, (february): 48-54. schultz, ellen e. 1992a. “passing the buck, in new pension plans, companies are putting the onus on workers,” wall street journal, (july 7): al, a5. 156 financial services review, 2(2) 1993 schultz, ellen e. 1992b. “ignorant of retirement nest egg? watch out!,” wall street journal, (september 11): cl, c9. schwimmer, anne. 1992. “discount rates dropping,” pensions & investments, (may 11): 18. sharpe, william f. 1976. “corporate pension funding policy,” journal of financial economics, (june): 183-193. stone, mary. 1987. “a financing explanation for overfunded pension plan terminations,” journal of accounting research, (autumn): 3 17-326. tepper, irwin. 1981. “taxation and corporate pension policy,” journal of finunce, (march): 1-13. todd, richard m. and neil wallace. 1992. “spdas and gics: like money in the bank?’ quarterly review, federal reserve bank of minneapolis, (summer): 2-17. treynor, jack l. 1977. ‘the principles of corporate pension finance,” journal of finance, (may): 627-638. vosti, curtis. 1992a. “chrysler pension fund gets hit,” pensions l investments, (march 30): 1 vosti, curtis. 1992b. “gm stock contribution eases underfunding,” pensions & investments, (may 11): 27. vosti, curtis. 1992c. “governments look at defined contribution,” pensions & investment, (may 11): 3. warshawsky, mark j. 1989. the adequacy of funding of private defined benefit pension plans, finance and economics discussion series, board of governors of the federal reserve system. wentx, arthur g., allen s. anderson, and a. frederic banda. 1991. quality ratings of state pension funds: impact of underfunded pension liabilities, working paper, the university of akron. willinger, g. lee. 1992. “a simulation comparison of actuarial and contingent claims models for underfunded pension liabilities,” quarterly journal of business and economics, 3 l(1): 12-97. winlclevoss, h. 1977. pension mathematics: with numerical illustrations, homewood, il, richard d. irwin. winklevoss, h. 1982. “plasm: pension liability and asset simulation model,” journal of finance, (may): 585-594. zall, milton. 1992. “understanding the risks to pension benefits,” personnel journal, (january): 64. pii: s1057-0810(99)00024-4 planning to move to retirement housing karen m. giblera,*, george p. moschisb, euehun leec avisiting associate professor, department of real estate, j. mack robinson college of business, georgia state university, p.o. box 4020, atlanta, ga, 30302-4020, usa bcenter for mature consumer studies, georgia state university, atlanta, ga, 30303, usa csejong university, koonja-dong kwangjin-gu, 1433-747 seoul, korea abstract as the size and diversity of the older segment of our population grows, they will need increased assistance in planning for and choosing appropriate housing from the array of options offered in the marketplace. this national survey of people age 55 and older indicates interest in retirement community housing among all socioeconomic groups, but especially among women and better-educated seniors. facilities that offer access to medical services, transportation, and shopping while providing housekeeping and personal care services along with social activities will appeal to these consumers. © 1999 elsevier science inc. all rights reserved. keywords:economics of the elderly; housing demand 1. introduction as the american population ages more attention is being focused on the housing needs and preferences of older americans. currently, there are more than 33 million people in the u.s. age 65 and older. as health and medical treatments improve, more people are surviving to reach and remain in the elderly population longer. in the last 40 years, the number of americans age 85 and older has increased by more than 250%. by 2010, the 85 and older population is expected to number 6 million, a 100% increase over 1990. the census bureau estimates that by 2050, 20% of the u.s. population will be 65 or older (hobbs, 1996). the elderly population is made up of a diverse group of americans with differing needs * corresponding author. tel.:11-404-651-4612; fax:11-404-651-3396. e-mail address:kgibler@gsu.edu (k.m. gibler) financial services review 7 (1998) 291–300 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00024-4 and resources. each must choose where and how to live as they age. changes in health, family composition, and leisure interests lead many older americans to consider an alternative to aging in place in a traditional single-family home. the housing industry, recognizing these opportunities and constraints, is responding by providing a variety of retirement housing options that merge real estate and support services. the purpose of this paper is to review the seniors housing options currently available to older americans and use the results of a national survey to examine the related preferences of those in or near retirement age. 2. background 2.1. housing preferences surveys have show that at least 80% of older americans want to age in place (aarp, 1992; 1996). however, about 5% of elderly americans move each year. most of these movers change residences within the same metropolitan area or within the same region, but one-quarter move to a different state (hobbs, 1996). elderly movers appear to be both pushed and pulled into moves. younger, more affluent, healthy seniors tend to be attracted to amenity-rich retirement communities and make long-distance moves to reach them. failing health, lack of social support, and inability to maintain the home trigger moves by older, slightly impaired seniors into a relative’s home or a supportive housing situation (de jong et al., 1995; speare et al., 1991). these seniors are seeking help with activities of daily living (adls) such as bathing, dressing, and walking or instrumental activities of daily living (iadl) such as preparing meals and doing housework. 2.2. housing options the seniors housing industry consists of a diverse collection of housing and service offerings, including everything from traditional single-family houses to nursing homes. the levels of services and medical care provided usually defines seniors housing types, however, no standard definitions exist. some of the terms used to describe the offerings in the seniors housing industry include: independent living, retirement community, active adult community, rest home, congregate housing, board and care home, assisted living facility, intermediate care facility, skilled nursing facility, nursing home, continuing care retirement community, and lifecare facility. the physical design ranges from detached single-family homes to institutional facilities with semi-private bedrooms. the level of services provided ranges from none to 24-hour medical and personal services. independent retirement housing refers to seniors-only single-family houses, condominiums, or rental apartments that provide no personal services. often called active adult or retirement communities, these developments may provide amenities and recreational facilities, such as golf courses. these are available in all price ranges. congregate care provides independent rental apartments with access to a common dining 292 k.m. gibler et al. / financial services review 7 (1998) 291–300 facility. housekeeping, recreation, and transportation services are often provided. residents must be healthy enough to care for themselves and use outside medical services. these may cost up to $2,000 per month, depending on the range of services offered and whether fees for service are charged separately or included in the monthly rent. assisted living facilities (sometimes called residential care, board and care, or sheltered care) are designed for a frailer resident. residents occupy private or semiprivate rooms with access to central dining and activity rooms. services include meals, assistance with adls and iadls, security, emergency call system, transportation, medication management, laundry, housekeeping, and social programs. assisted living facilities association surveys indicate the average basic per-diem rate for a private room is $72, with cost varying from $20 to $200 per day. this averages $2,200 per month (alfa, 1996). some subsidized facilities, however, keep their rents as low as $600 a month. medicaid reimbursement for assisted living service costs is just getting underway through waiver programs (brecht, 1996). thus, most residents are private pay. custodial care provides unskilled nursing care 24-hours-a-day in a private unit. skilled nursing facilities provide 24-hour-a-day nursing and medical care to residents. the average daily charge for care in a nursing home is $127 a day, about $3,800 a month (levit et al., 1996). one-third of residents are private pay and medicaid supports two-thirds. a continuing care retirement community (ccrc) or lifecare community offers a range of living units from independent to skilled nursing facilities, often at one location. ccrcs are usually rental facilities, with or without a refundable entrance fee. residents enter the independent units when in good health, but remain and move to more supportive housing within the facility when additional services are needed. the original concept was that the resident would pay the entrance fee to guarantee a place in whatever level of care was needed in the future, but this is evolving into a fee for service arrangement (scribner & dalkowski, 1998). the role of hmos in providing health care for ccrc residents is creating a growing complication for this segment of the industry. 2.3. financial resources older americans obtain the cash flow needed to pay for housing from various income sources. few are in the workforce drawing a salary. over 90% of americans age 65 and older receive income from social security, with its benefits the primary source of income for more than 60% of the elderly and the only source of income for 14% of beneficiaries. over two-thirds (67%) receive asset-generated income. just under half (42%) receive pension, ira, keogh, 401(k) or other retirement benefits (u.s. department of health & human services, 1996). one-fourth of households headed by a senior citizen have an annual income of less than $10,000, 41% receive between $10,000 and $24,999, and 23% receive between $25,000 and $49,999 (hobbs, 1996). more of the seniors of the next century, especially the women, are expected to receive pension payments and those payments are expected to average higher than those current retirees receive, raising the average income of the elderly. the median net worth of elderly households in 1991 ranged from $104,354 for those age 65 to 69 to $76,541 for those 75 or older, with home equity accounting for 42% of net worth (hobbs, 1996). interest-earning assets are also an important component of elderly net worth. 293k.m. gibler et al. / financial services review 7 (1998) 291–300 2.4. housing choices of americans age 65 and older, 30% live alone, 55% live with a spouse, 13% live with other relatives and 2% live with nonrelatives only. a majority of elderly householders live in single-family detached houses that they own in mixed-age neighborhoods, especially younger couples with higher incomes. three-fourths of the elderly live in metropolitan areas, with a majority in the suburbs. although their homes may be more than 30 years old, most are physically sound (hobbs, 1996). precise figures are not available for the number of elderly living in each type of seniors housing because of the variety of labels given them as well as the lack of central data collection. asha (1998) estimates there are approximately 15,500 professionally owned and managed seniors-housing residences in the u.s. with a total capacity of 1.6 million residents. the 2bema network inc. estimates there are 10,100 seniors apartment, congregate housing, and retirement communities in the u.s. (kelman, 1998). the estimated number of residents of congregate facilities ranges from 200,000 to 500,000 (feinberg, 1993). typical residents are middleto upper-income widows age 75 to 85 who are in good health (asha, 1994). the 2bema network inc. estimates the number of assisted living facilities in the u.s. at 27,500 (kelman, 1998). the estimated number of assisted living facility residents ranges from 500,000 to 1 million, however, confusion in the definition of housing types may overstate this number (alfaa, 1993; feinberg, 1993). the typical resident of assisted living facility is a widowed female in her 80s with an annual income of around $20,000. she is suffering three adl disabilities associated with advanced age. many residents stay in an assisted living facility for only two to three years, then move to obtain greater care in a hospital or nursing home (alfa, 1996). the number of ccrcs is estimated at 1,200 (kelman, 1998), with an estimated population of between 200,000 and 500,000 (alfaa, 1993: feinberg, 1993). the typical entrant to an independent unit in a ccrc is a white, better educated, higher income, widowed female in her 70s without children nearby and with few disabilities, however ccrcs tend to attract higher numbers of married couples than other types of seniors housing. the average length of stay in a ccrc is 7.5 years (asha, 1994; kichen & roche, 1990; newcomer et al., 1995). there are more than 18,000 nursing homes in the u.s. (kelman, 1998) housing approximately 5% of americans age 65 and older and nearly one-fourth of the oldest old (national center for health statistics, 1995). still other older americans live with family members in their homes. but increased mobility, more working women, a declining birth rate, and increasing single parenthood and divorce have led to sociological changes that contribute to younger americans not caring for elderly family members in their homes (scribner & dalkowski, 1998). 2.5. housing planning most middle-aged and older americans do not have any plans for how and where they will live during later life (aarp, 1996). studies indicate that most older people are unfamiliar with their housing and care options, so they are not prepared to make an informed choice 294 k.m. gibler et al. / financial services review 7 (1998) 291–300 (gibler et al., 1998; harvard & louis harris & associates, 1996). the many terms, definitions, and variety of housing and service combinations available make it difficult for the elderly to compare their housing and service options. few have any personal experience with seniors housing beyond visiting relatives or friends in nursing homes. seniors often believe the only option available when they can no longer care for themselves in an independent single-family house is to move to a nursing home, a place they believe people with no other choices go to die. among those older consumers who have heard of seniors housing options, many have misconceptions, including the belief that congregate housing and ccrcs are places only for sick people and that lifecare communities only offer planned recreation and meals (gibler et al., 1998; harvard & louis harris & associates, 1996). consumers who are familiar with the options cite the importance of privacy and independence in choosing their housing unit. equally as important are supportive services within the community, especially health care, transportation, emergency call system, meal service, planned activities, and housekeeping (gibler et al., 1998). older consumers are limited in their understanding of the financial aspects of seniors housing as well. some developments offer fee-simple ownership. others are leaseholds. the rental fee may or may not include the cost of supportive services. the ccrc and lifecare communities have often been structured with large entrance fees. financial questions about refundability of entrance fees and health care payments by hmos will continue to complicate this option. the lack of knowledge and understanding of seniors housing options extends to another important group of consumers—the elderly’s children. many older residents consult family members about potential moves (aarp 1992; alfaa, 1993). the children, especially daughters and daughters-in-law, can be very influential in seniors housing decisions. thus, the growing elderly american population is faced with a variety of housing options. to help these consumers make better housing choices, a better understanding of their needs and preferences is essential. the purpose of this paper is to further explore the potential consumers of retirement housing, their purposes in moving and their preferences in their housing. this is done through a survey of older americans to identify who plans to move into retirement housing in the future and what they want in that housing. 3. methodology as part of a larger consumer study of americans of all ages, a national mail survey is used to learn more about the retirement housing interests of those approaching and already in retirement age. questionnaires are mailed to 10,000 adults in a stratified random sample in which the number of people selected in each state is proportionate to the state’s share of the u.s. population. a total of 1,463 usable responses are received from people age 55 and older. this age group is selected for this analysis because these people are either considering retirement housing now or will be in the near future. the survey gathers socioeconomic information about the respondents. they are questioned whether they currently live in a retirement community and if they plan to live in one 295k.m. gibler et al. / financial services review 7 (1998) 291–300 in the future. those planning to live in retirement housing are asked what services and location characteristics are important in choosing a specific retirement community. to further understand their decision process and expectations, these respondents are asked why they think people move out of single-family homes in mixed age neighborhoods and into retirement communities. 4. results and analysis the demographic characteristics of the survey respondents age 55 and older are presented in table 1. this group is comprised mostly of married couples living by themselves (empty nesters). almost half (43%) are age 65 to 74. most are white (97%) retired (81%) males (64%) with an annual income of less than $35,000 (64%). the majority of the respondents (69%) do not have any college education. the survey reaches people in all regions of the country, with two-thirds of respondents from north central and south. (housing characteristics are presented in table 2.) almost all table 1 demographic characteristics of survey respondents characteristic % of respondents (n 5 1,463) gender female 36 male 64 race black 2 white 97 other 1 education high school or less 69 at least some college 31 age 55–64 28 65–74 43 75 and older 29 income less than $20,000 32 $20,000–$34,999 32 $35,000–$49,999 17 $50,000 or more 19 marital status married 71 not married 29 employment status of respondent retired or never employed 81 employed 19 employment status of spouse retired or never employed 79 employed 21 296 k.m. gibler et al. / financial services review 7 (1998) 291–300 of the respondents (96%) currently live in non-retirement housing. when asked about their future housing plans, 26% of the respondents indicate plans to move into either a retirement community without health-care services, a retirement community with health-care services, or a nursing home. this may appear to be a relatively high proportion of respondents considering previous studies (aarp, 1992; aarp, 1996) have found that at least 80% of older americans wanted to never move. however, those studies ask how strongly people agree with the statement, “what i’d really like to do is stay in my own home and never move.” this study asks in what type of housing people plan to live in the future. thus, some seniors may be happy with their current homes and would prefer to avoid the disruption of moving, but they may realize they cannot obtain all the services there that they may want and need in the future. thus, they may plan to move to a retirement community in the future to satisfy certain needs, not because it is their optimal choice. a comparison of those who plan to live in retirement housing with those who do not indicates a higher interest in retirement housing among females and better-educated respondents, as is shown in table 3. people in all age and income groups, employed and retired, married and widowed indicate a similar level of interest in retirement community living. the higher interest among the better educated may be related to them having greater awareness and understanding of the options available in retirement housing. thus, they would more likely recognize the benefits that retirement community housing offers. other less-educated consumers may be under false impressions that retirement communities are simply nursing homes or that all retirement facilities are very expensive, as gibler et al. (1998) find. most women are aware that they face potential widowhood in later life. thus, they would only be practical to plan for some type of appropriate housing for an elderly person living alone. the respondents are asked to indicate what characteristics are important in choosing a table 2 current housing situation of survey respondents characteristic % of respondents (n 5 1,463) household size 1 23 2 65 3 or more 12 region east 18 north central 30 south 33 west 19 type of current housing retirement 4 non-retirement 96 plan for future housing retirement 26 non-retirement 74 297k.m. gibler et al. / financial services review 7 (1998) 291–300 home. the important considerations to those planning to live in a retirement community, based on the number who said that item is important, are: access to medical services, access to planned social activities, access to public transportation, location near hospitals and shopping centers, and access to personal and home care services, as is shown in table 4. these findings are similar to earlier studies and reinforce the importance placed on both the social aspects of retirement community living as well as the medical and personal support services. the respondents are also asked to indicate what they think are the reasons people move into retirement communities. the most often cited reasons are to increase social contacts and activities. most also believe people move to retirement communities to obtain access to personal care services, to avoid household chores, or after the loss of a spouse (table 5). again, these respondents are indicating that the retirement community is seen as fulfilling two needs: a social need, especially for widows, and a physical/medical need for those suffering physical limitations. table 3 differences in those planning and not planning to live in retirement community characteristic % people planning to live in retirement community (n 5 381) % people not planning to live in retirement community (n 5 1,082) x2 p gender 5.839 .016 female 41 34 male 59 66 race 0.987 .611 black 2 2 white 98 98 other 1 1 education 4.353 .037 high school or less 65 71 some college 35 29 age 55–64 28 27 1.277 .735 65–74 43 45 75 and older 28 27 income less than $25,000 31 32 1.298 .730 $25,000–$34,999 32 32 $35,000–$49,999 18 16 $50,000 or more 19 20 marital status 0.018 .895 married 71 72 not married 29 28 employment status 0.333 .564 retired 82 80 employed 18 20 household size 1.324 .857 1 25 23 2 64 65 3 or more 11 12 298 k.m. gibler et al. / financial services review 7 (1998) 291–300 5. conclusions and recommendations the size and composition of the u.s. elderly population is changing. as the number and diversity of this segment of our population grows, they will need increased assistance in understanding and evaluating their housing options. based on this survey, the interest in retirement community housing appears to cut across socioeconomic lines, indicating possible demand for a wide variety of retirement housing facilities. better-educated consumers are more likely to plan to move to a retirement community sometime in the future, perhaps reflecting a better awareness and understanding of the housing options available. interest is also relatively higher among females, perhaps indicating their awareness of potential widowhood and the social life that other residents of a retirement housing community might provide at that time. this is also supported by the number of people who think the reason people move out of their homes and into a retirement community is because of loss of a spouse. in choosing a specific retirement community, these retirees look for access to medical services (including hospitals and home care), public transportation, and shopping. they table 4 preferences of those planning to live in a retirement community characteristic % rating characteristic as important in choosing retirement community (n 5 374) access to medical services 65 access to planned social activities 59 access to public transportation 58 location near hospitals 57 location near shopping centers 56 access to personal & home-care services 53 home or personal security 47 distance from friends and relatives 47 table 5 reasons why people move out of single family homes and into retirement communities reason for moving % of respondents planning to move who think reason applies (n 5 375) more social contacts and activities 77 access to personal care services 61 unwilling/unable to do house chores 59 loss of spouse 56 freedom and independence 52 reduce housing costs 50 need continuous health care assistance 38 closer to relatives 30 299k.m. gibler et al. / financial services review 7 (1998) 291–300 expect the retirement community to offer them greater social contacts and activities while reducing their household chores and providing personal care services. thus, the location relative to support services and their access are critical in the selection of any retirement community. yet, the decision will remain personal and distinctive as seniors choose housing that best suits their individual needs within their social, economic, and physical constraints. references american association for retired persons. (1992).understanding senior housing for the 1990’s. washington, dc: aarp. american association for retired persons. (1996).understanding senior housing into the next century. washington, dc: aarp. assisted living facilities association of america. (1993).an overview of the assisted living industry. fairfax, va: alfaa. assisted living federation of america. (1996).an overview of the assisted living industry. washington, dc: alfa. american seniors housing association. (1994).seniors housing: the market-driven solution to long-term care. washington, dc: asha. american seniors housing association. (1998).seniors housing construction report. washington, dc: asha. brecht, s. b. (1996). trends in the retirement housing industry.urban landnovember, 33–39. de jong, g. f., wilmoth, j. m., angel, j. l., & cornwell, g. t. (1995). motives and geographic mobility of very old americans.journal of gerontology: social sciences 50, s395–s404. feinberg, p. (1993). senior housing: investments coming of age.real estate forumseptember, 84–95. gibler, k. m., lumpkin, j. r., & moschis, g. p. (1998). retirement housing and long-term health care: attitudes of the elderly. in m. a. anikeeff & g. r. mueller (eds.),research issues in real estate: vol. 4. seniors housing(pp. 109–130). boston: kluwer. harvard school of public health & louis harris & associates. (1996).long term care awareness survey. cambridge, ma: harvard. hobbs, f. b. (1996).651 in the united states(current population reports special studies p23–190). washington, dc: u.s. bureau of the census. kelman, h. (1998). prime time for elder care.real estate forumaugust, 38–40, 42, 44, 46, 48, 50, 52–57. kichen, j. m., & roche, j. l. (1990). life-care resident preferences. in r. d. chellis & p. j. grayson (eds.),life care: a long-term solution?(pp. 49–60). lexington, ma: lexington. levit, k. r., lazenby, h. c., braden, b. r., cowan, c. a., mcdonnell, p. a., sivarajan, l., stiller, j. m., won, d. k., donham, c. s., long, a. m., & stewart, m. w. (1996). national health expenditures, 1995.health care financing review 18(1), 175–214. national center for health statistics. (1995).health, united states, 1994.hyattsville, md: public health service. newcomer, r., preston, s., & roderick, s. s. (1995). assisted living and nursing unit use among continuing care retirement community residents.research on aging 17(2), 149–167. scribner, d., jr., & dalkowski, j. a. (1998). the evolution and status of seniors housing terminology. in m. a. anikeeff & g. r. mueller (eds.),research issues in real estate: vol. 4. seniors housing(pp. 73–88). boston: kluwer. speare, a., jr., avery, r., & lawton, l. (1991). disability, residential mobility, and changes in living arrangements.journal of gerontology: social sciences 46, s133–s142. 300 k.m. gibler et al. / financial services review 7 (1998) 291–300 pii: 1057-0810(94)90016-7 financial services review, 3(2): 93-108 copyright 0 1994 by jai press inc. issn: 10.57-0810 all rights of reproduction in any form reserved. an optimization model for scheduling withdrawals from tax-deferred retirement accounts cliff t. ragsdale andrew f. seila philip l. little as a growing number of americans reach refiremen? age, more and more people are facing important decisions about how to withdraw savingsfrom tax-deferred retirement accounts (tdras). these decisions are complicated by the federal tax code which imposes a number of rules and regulations on these withdrawals. since these decisions collectively involve billions of dollars, the potential lossfrom even slightly suboptimal decision making is very large. in this paper, we present a mathematical programming model that can be used to assist retirees ami/or their advisors in determining the optimal schedule of withdrawals from tdras. i. introduction as the general population of the united states ages, more and more americans are having to make important decisions about how to withdraw money from tax-deferred retirement accounts (tdras). tdras include individual retirement accounts (iras), qualified corporate retirement plans, and tax-deferred salary reduction plans covered by section 403(b) of the internal revenue service (irs) code. generally speaking, tdras are an attractive investment alternative for two primary reasons: 1) they offer a way for some taxpayers to reduce their current tax liability; and 2) interest earnings on these investments are sheltered from taxes until they are withdrawn. typically the argument is made that people should invest in tdras so their investment will grow faster (i.e., money is being com pounded rather than taxed) and likely will be subjected to a lower tax rate at the taxpayer’s retirement. cliff t. ragsdale l department of management science, virginia polytechnic institute and state university, blacksburg, va 24061-0235; andrew f. seila l department of insurance, real estate, legal studies and management science, terry college of business administration, university of georgia, athens, ga 30602-6255; philip l. little l department of accounting, western carolina university, cullowhee, nc 28723. 94 financial services review, 3(2) 1994 american demographics (1989) reports that investors have recently put as much as $100 billion annually into iras alone. thus, the amounts and timing of the withdrawals from tdras can have a significant impact on the amount of taxes one pays and the accumulation of personal wealth however, this decision is complicated by the federal tax code which imposes a myriad of rules and regulations on these withdrawals. in addition to the normal income tax that applies to withdrawals from tdras, there are also penalty taxes associated with withdrawing too much or too little and making withdrawals too soon or too zate. coupled with these rules are a number of more subtle investment issues which further complicate the process of determining the optimum amount to withdraw each year. in this paper we first briefly review the rules regarding distributions from tdras in order to demonstrate and motivate the need for an optimization model that can assist investors (or their advisors) in determining how to withdraw money from these accounts. next, a mathematical programming model for this problem is proposed and described in detail. finally, an example is provided to demonstrate the potential merit of the proposed model versus a number of heuristics that have been suggested in the literature. ii. the problem a complete description of the regulations regarding tdra distributions is available in irs publication 590 (1992) or less technical summary articles (e.g., abramson, 1989; katz, 1990; mccommbe, 1989). for the reader’s convenience and reference, we have also summarized these rules in table 1. it goes without saying that there are some minor exceptions to these rules. premature distributions tax congress has established various rules concerning when withdrawals can be made from tdras. generally speaking, the tax law is written with the intention that withdrawals from tdras begin no earlier than age 59.5. with few exceptions, any money withdrawn from tdras before age 59.5 is subject to a 10 percent premature distribution tax penalty in addition to regular income taxes. table 1. summary of tdra tax rules (na = not applicable) age when withdrawal is made rule premature distribution tax’ minimum distribution tax2 excess distribution tax3 before 59.5 from 59.5 to 70.5 10% na 5z 1% after 70.5 go 15% notes: ‘applies to the total amount withdrawn in any year. 2applies to the amount by which the minimum legally required withdrawal in any year exceeds the actual amount withdrawn. 3applies to amounts in excess of $150,000 withdrawn in any year. this tax also has an impact on the determination of estate taxes at the taxpayer’s death. withdrawal from tax-deferred retirement accounts 95 minimum distribution tax as the name implies, tdras were originally intended to provide income to retirees not estates for their heirs. thus, additional rules require that minimum withdrawals be made from tdras on an annual basis beginning no later than the year in which the taxpayer reaches age 70.5. the actual withdrawal for this first “required” year (the year in which age 70.5 is reached) may be deferred as late as april 1 of the following year if desired. this deferral provision can be advantageous since it also defers the taxes on the first “required” year’s withdrawal for a year. however, since the taxpayer must also make a withdrawal for the second “required” year (the year in which age 7 1.5 is reached), effectively making two withdrawals in this second year could place some income into a higher tax bracket which might more than offset the benefit of deferring the taxes on the first “required” year’s withdrawal (e.g., katz, 1990, p. 56). at any rate, annual withdrawals must also be made in each of the following years (where ages 72.5,73.5, . . . are reached). the actual minimum amount that must be withdrawn in each year beginning at age 70.5 is determined as follows. for each separate tdra the taxpayer owns, a theoretical minimum figure is calculated by dividing the balance in the account at the beginning of the year by the joint life expectancy of the taxpayer and the designated beneficiary for the account (e.g., johnson, 1990). a schedule of life expectancy factors is supplied in irs publication 590 (1992) to assist in determining this minimum amount. the minimum figure for each account is theoretical in that the irs does not actually require the money to be withdrawn from this particular account, provided the sum of the actual withdrawals is at least as much as the sum of the theoretical figures (e.g., geller, 1988; solbee, 1988). for instance, table 2 shows the calculations for a taxpayer with two tdras where the minimum required withdrawal is $72,172. this taxpayer may satisfy the irs requirements by with drawing uf least $72,172 from either account or in any combination between accounts. of course, the manner in which the taxpayer elects to split the withdrawal between these accounts will determine the amounts left in the accounts and, therefore, also affects the minimum required withdrawals in subsequent years. for retirees who are age 70.5 or older and fail to make the minimum required withdrawal, a minimum distribution tax penalty of 50 percent applies to the difference between the actual amount withdrawn and the minimum required withdrawal. for instance, table 2. minimum distribution tax example account i account 2 age balance 70 $7oo,ooo 71 ? 72 ? 73 ? 74 ? joint life expectancy’ 26.2 25.3 24.4 23.5 22.7 balance $900,000 ? ? ? ? join? life expectancyz 19.8 19.0 18.2 17.3 16.5 minimum required withdrawa (at age 70) = w+y=$72,172 notes: lassuming 40 year old beneficiary *assuming 72 year old beneficiary 96 financial services review, 3(2) 1994 if the minimum required withdrawal is $100,000 and the taxpayer only withdraws $50,000, a nondeductible penalty tax of $25,000 (i.e., 0.50(100,000 50,000)) must be paid (e.g., mccommbe, 1989). excess distributions tax in keeping with the philosophy that tdras should be used to provide retirement income, the excess distributions tax is intended to discourage people from using tdras to amass or receive excessive retirement benefits. this tax imposes a penalty on the amount by which the total withdrawal from all tdras exceed $150,000 in any year. for any such “excess distribution” made prior to age 59.5 the tax is effectively five percent of the amount exceeding $150,000 (and the 10 percent premature distribution tax applies to the entire amount withdrawn). for “excess distributions” made at or after age 59.5 the tax is 15 percent of the amount exceeding $150,000. note that this tax applies even to those who are 70.5 or older and are required to make a “minimum withdrawal” (as discussed above) in excess of $150,000. for instance, if the minimum required withdrawal is $175,000 and the taxpayer withdraws this amount, he or she must pay an excess distribution tax penalty of $3,750 (i.e., 0.15(175,000 150,000)) in addition to the normal income taxes which apply. of course, if this taxpayer tries to avoid the excess distribution tax by withdrawing only $150,000 the 50 percent minimum distribution tax would levy a penalty of $12,500 (i.e., 0.50(175,000 150,000)). estate taxes an estate tax return must be filed if a taxpayer’s gross estate exceeds $600,000 at the time of death. generally, the balances in tdras are included in the valuation of one’s estate at the time of death and are subject to normal estate taxes. however, the law allows special exclusions which effectively eliminate normal estate taxes on accounts where one’s spouse is the beneficiary or where the remaining balances are left to qualified charitable organiza tions (e.g., irs publication 448, 1992). the excess distributions tax described above also plays a role in determining the taxes due on one’s estate. again, since tdras are only intended to provide retirement income for a given taxpayer, the tax law maintains that there should not be an “excessive accumulation” of funds left in these accounts at the taxpayer’s death. congress has decided that a “reasonable” amount to have in tdras at one’s death should total no more than the present value of a $150,000 annuity for the remaining actuarial life expectancy of the taxpayer at the time of their death. thus, one’s estate must pay a 15 percent penalty tax on the amount by which the total value of the deceased’s tdras exceed this “reasonable” amount. this “excess accumulation” tax applies to all tdras regardless of beneficiary designations. iii. a difficult decision from the previous discussion it is clear that the question of how one should go about making withdrawals from tdras can be difficult. for many, it is probably challenging enough to make withdrawals that are within the irs guidelines. however, if one attempts to make withdrawals that are not only “legal” but also maximize the value of one’s benefits, the withdrawals from ta-deferred retirement accounts 97 problem enters a new realm of difficulty. in either case, a number of practical questions must be addressed such as: 1) when should withdrawals begin? 2) how long should they continue? 3) how much should be withdrawn each year? 4) from which accounts should withdrawals be made? clearly, the premature distributions tax can be avoided by not making withdrawals before age 59.5 and the minimum distribution tax can be avoided by making the required minimum withdrawal beginning at age 70.5. thus, with respect to the first question above we can generally say that withdrawals should begin no sooner than age 59.5 and no later than 70.5. a more specific answer would require consideration of an individual taxpayer’s income needs. similarly, with respect to the third question above we can generally say that the taxpayer should make at least the minimum required withdrawal beginning at age 70.5 in order to avoid the onerous minimum distribution tax penalty. however, in some instances it may be necessary and/or wise to withdraw more than the minimum required amount. the “best” or optimal answer to all of the questions listed above requires one to consider the simultaneous impacts of a number of subtle factors over a period of years. for instance, when deciding from which accounts to actually withdraw money one must consider the rates of return on the various accounts and the schedule of life expectancy factors which will apply to future balances in the accounts. one might intuitively sense that the optimal withdrawal policy would first involve making withdrawals from the accounts with the lowest rate of return. however, it is possible that the beneficiary designations on higher yielding accounts will, in subsequent years, impose higher minimum required withdrawals which might force more taxable income into higher tax brackets. for instance, let us suppose that the rate of return on account 1 in table 2 is less than that the return on account 2. if the required withdrawal for year 1 is made from account 1 this will obviously cause more money to accumulate in account 2 which, in turn, will cause the minimum required withdrawal in subsequent years to be higher (due to the smaller joint life expectancy value on this account). part of this higher minimum required withdrawal may fall into a higher tax bracket that may more than offset the higher earnings on this account. thus, in some situations it might actually be best to withdraw money from accounts earning the highest rates of return. similarly, it is easy to believe that one should avoid making withdrawals for as long as possible (if not needed as current income) as this allows the investment to continue to earn interest and defer taxes. however, it is possible that in some cases withdrawals should begin before age 70.5 to help avoid exposure to the excess distribution tax. simultaneously evaluating all these factors affecting the decision can quickly over whelm us and prompt many to adopt heuristic withdrawal policies (e.g., gould, 1988; mcintosh & hollinrake, 1988; quinn, 1988; saftner & fink, 1990; sage, 1988; tiaa cref, 1991; tritch, 1988). two such policies are summarized below: minimal withdrawal policy: withdraw the maximum of the minimum required by the irs or the minimum needed to reach the desired level of retirement income. the actual withdrawal is made from the account(s) paying the lowest rate(s) of interest. 98 financial services review, 3(2) 1994 proportional withdrawal policy: same as above except the actual withdrawal is made from all the accounts in proportion to their balances at the beginning of the year. (note: this is the default withdrawal policy used by tiaa-cref.) notice that these policies ensure that the taxpayer withdraws at least as much as required by the irs to avoid the 50 percent minimum distribution tax penalty. while such policies may provide good “rules of thumb” for the average investor to follow, specific individuals can lose thousands of dollars by making suboptimal withdrawals using these heuristics (e.g., katz, 1990; ragsdale, seila, & little, 1993). if one considers the number of individuals with tdra investments and the amount of money deposited in these accounts, the total potential loss to individual taxpayers as a result of poor or even slightly suboptimal decision making is considerable. thus, even heuristics that generally work well still may leave specific individuals with very suboptimal withdrawal schedules. iv. an optimization model a taxpayer facing the decisions described above might be interested in determining the schedule of withdrawals that maximizes the net (after tax) present value (npv) of the withdrawals made over their life expectancy plus the npv of the remaining tdra balances passing to their beneficiaries (all within irs regulations). in this section, we present a mathematical programming model that can be solved to determine the schedule of withdraw als that achieves this objective. assumptions since our model considers a series of withdrawals over a number of years, it is clearly unrealistic to assume that the tax law will not change during this time. on the other hand, it is also clearly impossible to foresee what these changes will be-particularly those of a political or economic nature. however, other changes do seem reasonably certain. for instance, it is reasonable to expect the tax rate schedules to change every year to account for inflation. today’s tax rate schedules will almost certainly not apply 10 years from now. but we might expect the current schedules adjusted for inflation may reasonably estimate what may exist 10 years hence. similarly, the $150,000 limit involved in today’s excess distri bution tax and estate tax will almost certainly be adjusted for inflation in the future. thus, our model accounts for inflationary changes in the tax law wherever appropriate. however, it does not account for unforeseen structural changes in the tax law. such changes would have to be incorporated into the model as they occur. a number of additional assumptions are also reflected in this model. 1. we assume that withdrawals are made at the end of each year to allow the taxpayer to accumulate as much tax-deferred income as possible. the formulation can easily be modified so that withdrawals are assumed to occur at the beginning of each year or on some other periodic basis, if so desired. 2. we assume that no contributions have been made to any accounts on a non-taxable basis. additional variables and constraints, currently omitted for simplicity, can be added to our model to accommodate non-taxable contributions. 100 financial services review, 3(z) 1994 withdrawals are determined by dividing the balance in each account at the beginning of the year (b$ by the joint life expectancy factor (a$ supplied by the irs. c cwij + wn:j by/uo) 2 0, i = n, (4) j=l i(wo-bijia,)>o, i=n,+i,...,n (5) j=l as mentioned in section 2.3, amounts withdrawn from tdras in excess of the “reasonable withdrawal limit” (presently $150,000) in any year are subject to a 15 percent excess distributions tax. constraint equations (6) and (7) force the variable ei to equal the amount by which withdrawals exceed the inflation-adjusted “reasonable withdrawal limit” (_‘$$?j in year i. any such “excessive” withdrawals are then subjected to the 15 percent excess distribution tax as shown in equation (1). ~wij-eis~y i=l,..., n,,n,+2 ,..., n (6) j=l “i c(wu+w,;j)-eiisi, i=n,+, (7) j=l in equations (8) and (9) the total taxable income from withdrawals (wij) and other sources (oi) is allocated to the variables that represent the total income falling into each of the three different personal income tax brackets each year (i.e., xii, xi2, and xi3). given the current personal tax rate structure, the objective function in equation (1) will ensure that taxable income will be first allocated to xi, and then to xi2 and xi3 since the tax rates for each of these income brackets (represented by rfh ) are monotonically increasing (i.e., t’; < t$ c t$ ). inflation-adjusted upper bounds (rnyk ) for the income brackets are given in equation (10). ~xi~-~wy-oi=o, i= 1,. . ,n,,n,+p,. . . ,n (8) k=l j=l i xiki (wg + dnj) oi = 0, i=n,+, (9) k=l j=l xikim$ i=l,...,n;k=l,2 (10) the constraints in equations (11) through (13) indicate that each account’s balance at the beginning of each year (b$ should equal the prior year’s beginning balance plus the interest earned, less any amount withdrawn from the account. constraint equations (12) and withdrawals from tar-deferred retirement accounts 101 (13) apply, respectively, to the years in which the taxpayer reaches age 70.5 (n,) and 71.5 (n,,). these constraints make the necessary adjustments to the beginning balances for years ++i and 4+2 if the deferral provision described in section 2.2 is utilized. notice that if the deferral provision is not used (i.e., if all w,:j = 0) constraint equations (12) and (13) assume the same form as equation (11). bi+i,j(l+rii)bii+wii=o,i=l ,..., n,._l,n,+2, . . . . n;j=l,...,nl (11) bi+l,j(1 +ru)by+wy+w$ =o,i=n,;j=l,. . . ,nl (12) bi+l,j(l+r~)b~+w~-~~~wn~j =o,i=n,+i;j=l,...,ni (13) the remaining constraints in the model have to do with estate taxes. as mentioned earlier, an estate tax return does not have to be filed if the value of a taxpayer’s estate does not exceed a certain cutoff point (presently $600,000). we use y to represent this cutoff point (adjusted for inflation) and b to represent the estimated future value of the taxpayer’s estate excluding tdras. since more than one beneficiary can be designated for a given tdra, it is also necessary to consider the percentage bj) of each tdra which passes to the taxpayer’s spouse or to a qualified charitable organization (which are both exempt from estate taxes) in determining the taxable value of an estate. in equation (14) the surplus variable s will equal the amount by which the taxpayer’s gross taxable estate exceeds y. similarly, if estate taxes must be paid the optimal value for the binary variable h will be one in equation (15) (where m represents a very large number). “1 c(l-~j)bn+lj+8-~i~ j=l (14) slm?l (15) from equations (14) and (15) we know that if estate taxes must be paid, h = 1 and the taxable value of the estate is y+s. in this case, it is necessary to allocate the taxable value of the estate to variables in equation (1) which represented the different estate income tax brackets (i.e., the yk). this is accomplished in equation (16). upper bounds for the different estate income tax brackets are given in equation (17). also notice that since the yk are penalized in equation (1) an optimizer will attempt to sets = 0 and h = 0 whenever possible. thus, in equations (14) and (15), s and h will assume strictly positive values if and only if estate taxes must be paid. 19 cy,-s-m=0 k=l (16) yk<m;,k=l,...,18 (17) given the current estate tax rate structure, the objective function in equation (1) will ensure that the taxable value of the estate will be first allocated to yi and then to y2 and so forth up to yt7 since the tax rates for each of these income brackets are monotonically 102 ftnah’cial services review, 3(2) 1994 increasing (i.e., 0.18 = t: c ti < . . . < tt7 = 0.55). however, t& = 0.60 and t&, = 0.55. thus, when necessary, the objective in equation (1) would favor skipping y18 and allocating the remaining estate income to yi9 since this has a lower tax rate. constraint equations (18) and (19) ensure that this does not happen. it is clear from equation (19) that yi9 can assume a positive value if and only if the binary variable x is zero. however, if x = 0 then by equations (17) and (18) it must be that y18 = rn&. thus, the inclusion of equations (18) and (19) in the model ensure that estate income is allocated to y19 if and only if all other yk (including yis) have reached their upper bounds. (18) y19 m( 1 x) i 0 (19) the last constraint in our model deals with the “excess accumulation” tax. the constraint in equation (20) forces the variable e to equal the amount by which the ending tdra account balances exceed the inflation-adjusted “reasonable” amount (b) that a taxpayer should have in their tdras at the time of their death (see appendix for a detailed definition of 9 ). note that the “excess accumulations” represented by e are subjected to a 15 percent tax in equation (1). finally, bounds and integral conditions for the variables in this model are given in equations (21) and (22). xik, yk, ei, e, wv. w& , by 2 0, ‘d i, j, k (21) lee {o,l) (22) in general, this model consists of 2n(n, + 2) + 23 variables and 46 + n,) n, + 25 constraints. in most real-world situations the formulation of this model would require several hundred variables and several hundred constraints. since the model has only two integer (binary) variables and all the constraints are linear, a problem of this size would be fairly easy to solve using a personal computer and almost any commercial optimization package. v. an example we shall illustrate the potential benefits of our model by analyzing the decision problem faced by the hypothetical taxpayer described below. example: suppose a 68 year old taxpayer has two tdras. the first account has a current value of $2400,000 and is invested in a money market fund yielding a 7.25% return. the taxpayer’s 68 year old spouse is the beneficiary for this account. the second account has a current value of $1,500,000 and is invested in a mutual fund yielding a 7.5% return. the withdrawals from tar-deferred retirement accounts 103 table 3. data items required for model formulation personal information economic information information required for each tdra present age of taxpayer and discount rate current account balance spouse expected rate of inflation expected annual investment return life expectancy of minimum desired level of annual age of oldest beneficiary on each account taxpayer and spouse retirement income (plus a minimum desired annual growth factor for the above amount) other non-tdra annual retirement percentage of account balance bequeathed income (plus an estimated annual to spousal beneficiary growth factor for the above amount) expected value of tax payer’s estate percentage of account balance bequeathed (excluding tdras) to charitable organizations taxpayer’s 35 year old daughter is the beneficiary for this account. the taxpayer has other taxable retirement income of $50,000 per year which is expected to increase with inflation at 4% each year and desires a minimum retirement income of $75,000 per year increasing by 5% per year. the taxable value of the taxpayer’s estate is expected to be $2,000,000 (excluding tdras) at the time of his or her death. finally, assume the taxpayer uses the “married filing jointly” tax rate schedule and that “risk free” investments are available yielding 5.25% annually. while the model described above might appear somewhat daunting, it is important to realize that it can be made totally transparent to its user when implemented as part of a computerized decision support system. such a system would provide a user-friendly interface to guide the user through entering the various data items shown in table 3. the system would then automatically formulate and solve the model without any additional user intervention. once the “optimal” schedule of withdrawals over the taxpayer’s life expectancy has been determined the user would be provided with a summary analysis of the different withdrawal policies as described below. analyzing the alternatives there are two major issues for the taxpayer in our example to consider: the amount of money withdrawn from the accounts and the amounts that will be left to the beneficiaries. thus, we first compare the results of the optimal withdrawal policy identified by our model to the two heuristic policies in terms of the npv of the withdrawals and npv of ending tdra balances left at the taxpayer’s death. these results for our example are given in table 4. as expected, the optimal withdrawal policy determined using the model results in the highest total npv; in this case, over $900,000 more than the minimal withdrawal policy and over $340,000 more than the proportional policy. interestingly, this solution provides not only the greatest amount of retirement income for the taxpayer (via withdrawals), but also results in the largest benefits (net of estate taxes) at the taxpayer’s death. this, of course, is due to the relatively large amounts of estate taxes required to be paid under the heuristic withdrawal policies. to better understand the reason for this, consider the schedules of withdrawals and ending account balances presented in tables 5 and 6. 104 financial services review, 3(2) 1994 table ,4. comparison of present values of withdrawals & ending balances (original example) wihdrawal policy minimal proportional optimal pv of withdrawals $2544,496 $2,692,088 $2.740.709 pv of income taxes on withdrawals (842,932) (910,824) (933,190 npv of withdrawals $1.701.564 $1,781,264 $1.807.519 pv of ending balance 2.050.674 1,866,016 l817.441 pv of estate taxes on ending balance (l204.765) (532,270) (168,560) npv of ending balance $845,909 $1,333,746 $1.648.881 total npv $2,547,473 $3,115,010 $3,456,400 in table 5 we see that under the minimal withdrawal policy, withdrawals are first made from account l-the account with the lowest investment yield. thus, as shown in table 6 this policy causes the balance in account 1 to decline steadily beginning at age 70 until it reaches zero. this leaves a substantial balance in account 2. in contrast, table 5 shows that the optimal policy follows the same withdrawal pattern as the minimal policy up through age 73 and afterward withdrawals begin to be made from account 2. thus, the optimal policy leaves a substantial balance in account 1 and zero in account 2 as shown in table 6. of course, the proportional policy withdraws money from both accounts over the years and leaves balances in each account at the taxpayer’s death. the reason for the relatively large amount of estate taxes for the heuristic withdrawal policies in table 4 should now be clear. in this example, the taxpayer’s spouse is the beneficiary on account 1 and, as discussed above, a surviving spouse is effectively exempt from normal estate taxes on tdras. thus, the estate taxes listed for the optimal withdrawal table 5. comparison of withdrawal amounts age 68 69 70 71 72 73 14 75 76 77 78 79 80 81 82 83 84 85 minimal proportional optimal account i account 2 account 1 account 2 account 1 account 2 $25,000 $0 $14,286 $10,714 $25,000 $0 26,750 0 15,270 11,480 26,750 0 175,237 0 100,070 75,405 175,237 0 186,513 0 106,871 80,726 186,513 0 199,036 0 114,497 86,696 199,036 0 211,598 0 122,239 92,785 211,598 0 224,358 0 130,271 99,123 156,705 67,652 238,171 0 139,032 106,048 0 239,169 252,615 0 148,352 113,435 0 257,483 266,123 0 157,304 120,576 0 275,690 281,678 0 167,777 128,921 0 297,121 295,873 0 177,702 136,886 0 318,314 3 10,246 0 188,102 145,257 0 341,165 324,652 0 198,979 154,039 0 365,870 338,192 0 209,152 162,319 0 390,030 350,305 0 220,401 171,478 0 418,322 363,087 0 231,126 180,276 0 446,094 9,963 362,674 241,423 188,785 0 475,444 withdrawals from tax-deferred retirement accounts 105 age 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 table 6. comparison of ending balances minimal proportional optimal accounr i account 2 account 1 accounr 2 account 1 account 2 $2,120,000 $1,612,500 $2,130,714 $1.601.786 $2,120,000 $1,612,500 2,246,950 1,733,438 2,269,921 1,710,440 2,246,950 1,733,438 2,234,617 1,863,445 2,334,420 1,763,319 2,234,616 1.863.445 2,210,113 2,003,204 2,396,795 1,814,842 2,210,112 2,003,204 2,171,311 2,153,444 2.456.066 1.864.259 2,171,309 2.153.444 2.117.133 2,314,953 2,511,891 1,991,293 2j.17.130 2.314.953 2,046,267 2,488,574 2,563,732 1,955,517 2,113,917 2,420,922 1,956,450 2,675,218 2,610,571 1,996,133 2,267,176 2,363,323 1,845,677 2,875,859 2,651,485 2,032,408 2,431,546 2,283,089 1,713,366 3,091,549 2,686,413 2ji64.262 2,607,833 2.178.631 1,555,907 3,323,415 2,713,401 2,090,161 2,796,901 2,044,907 1,372,838 3,572,671 2,732,420 2,110,037 2,999,676 1,879,%1 1,162,128 3,840,622 2,742,419 2,123,033 3,217,153 1,679,793 921,725 4,128,668 2,742,265 2,128,222 3,450,3% 1,439,908 650,358 4,438,319 2.73 1,927 2,125,520 3,700,550 l157.870 347,205 4,771,193 2,709,591 2,113,455 3,968,840 826,388 9,290 5,129,032 2,674,911 2,091,689 4,256,581 442,274 0 5,151,036 2,627,418 2,059,781 4,565,183 0 policy in table 4 represent only the 15 percent excess distribution tax on balances exceeding the “reasonable” value one should have in tdras at the time of death. the estate taxes under the minimal and proportional policies include this excess distribution tax plus the normal estate taxes due on the balances in account 2 (with the taxpayer’s child as beneficiary). vi. discussion the preceding example highlights a number of interesting issues. first, it is clear that the use of heuristics can lead to decisions that can result in very large financial losses. of course, the example presented here is carefully chosen to illustrate this fact and one might argue that it is obvious that the minimal withdrawal policy will perform poorly in this situation. indeed, other examples exist which show the minimal withdrawal policy to outperform the propor tional policy and still other examples show that these heuristics can also be optimal. however, determining all the conditions under which these heuristics can be expected to perform well or poorly is, in itself, a difficult or impossible task considering the infinite combinations of beneficiary designations, life expectancies, and interest rate assumptions that one can encounter. our model avoids this exercise by directly determining the optimal schedule of withdrawals under any set of conditions. additionally, even in cases where the model only serves to confirm one’s intuition about the optimal withdrawal schedule, there can be a significant psychological benefit in knowing with certainty that the best possible decision has been made for the given assumptions. 106 financial services review, 3(2) 1994 table 7. comparison of present values of withdrawals & ending balances (modz>ed example) pv of withdrawals pv of income taxes on withdrawals npv of withdrawals pv of ending balance pv of estate taxes on ending balance npv of ending balance total npv minimal $2,892,660 (1,003,087) $1,889,573 1,653,690 (143,998) $1,509,692 $3,399,265 withdrawal policy proporrional optimal $2,884,784 $2,892,660 (999,464) (1,003,087) $1,885,320 $1,889,573 1,636,203 1,653,690 (141,375) (143,998) $1,494,828 $1.509.692 $3,380,148 $3,399,265 second, this example also highlights a number of withdrawal policy matters where one’s intuition can fail. for instance, our example dramatically demonstrates how undesir able it can be to leave money in a tdra with a non-spouse beneficiary. this might lead one to believe that, given the option, it is best to first make withdrawals from accounts with non-spouse beneficiaries. however, the schedule of withdrawals in table 5 clearly indicates that this is not always the case since the optimal policy here involves first making withdraw als from account 1 where the spouse is the beneficiary. similarly, our example might lead one to believe that it is best to make one’s spouse the beneficiary on all tdras so as to avoid the payment of normal estate taxes at one’s death. (note that this alternative may not always be possible since not all taxpayers are married and, even if they are, they may be unable to change the beneficiary designations on certain accounts.) the results for this scenario are presented in table 7. in table 7 we see that if the spouse is the beneficiary on both accounts, the minimal and optimal withdrawal policies are identical while the proportional policy is only slightly suboptimal. notice, however, that the npv of the withdrawals under this scenario are larger than for the example in table 4. this is due to the fact that the joint life expectancy of the taxpayer and his or her spouse is smaller than the joint life expectancy of the taxpayer and his or her child. this, in turn, forces the required minimum withdrawals that begin at age 70.5 to be larger. thus, under this scenario the taxpayer receives $82,053 more in npv during their lifetime via withdrawals, but the npv being left to his or her spouse is reduced by $139,188. thus, the overall effect of making the spouse the beneficiary on both accounts is to decrease the npv of the taxpayer’s total benefits by about $57,000 compared to the optimal solution in the original example. of course, there might be some who would prefer the solution offered by this second scenario even though the total npv of the benefits are smaller. this illustrates yet another important aspect of the model. by allowing a person to easily play out such “what if?” scenarios, our model can determine not only the optimal withdrawal policy for a given set of conditions, but also allow the user to explore how changes in these conditions impact the solution. this can lead to the discovery of “better” (higher npv) solutions. it can also lead to the discovery of solutions which have a smaller npv but greater utility to individual decision makers. in either case, the model can provide the user with a greater understanding and sharper intuition about the problem they face and an objective means for assessing the trade-offs among the various possible decisions. withdrawals from tax-deferred retirement accounts 1w vii. conclusions in this paper we presented a mathematical programming model for assisting retirees (or their advisors) in determining how to make withdrawals from tdras. this model can be used to determine the withdrawal schedule that maximizes the npv of the taxpayer’s retirement benefits. the potential benefits of this model were illustrated relative to a number of heuristic withdrawal policies likely to be used in practice. it is important to note that the use of this model is not intended to be a one-time occurrence. while the model determines the optimal schedule of withdrawals over a number of years, the taxpayer really is only immediately interested in what action they should take in the current year. thus, the model can and should be updated on an annual basis to reflect changes in investment returns, life expectancies, beneficiary designations and, of course, structural changes in the tax law. when used in this manner our model offers the taxpayer the assurance of knowing what the best possible withdrawal decision is made each year based on the information at hand. finally, as mentioned at the outset, there are some exceptions to the tax rules embodied in our model that could have significant impacts on individual taxpayers. thus, our model should not be used as a replacement for tax advisors but in tandem with qualified financial planning professionals. appendix glossary of terms in the model aij bij 4 6 e: f 3 ei 2 n 4 nl oi pi pvi life expectancy factor at year i for investmentj (from irs [ 1991b] tables). balance at beginning of year i in tdra investmentj. desired minimum total taxable income from all sources in year i. the “risk free” discount rate. estimated total taxable future value of the taxpayer’s estate (excluding tdras). the estimated rate of inflation. the “reasonable” withdrawal limit adjusted for inflation (i.e., ._!zi = $150,000 x (1 +ni>. amount in excess of the “reasonable” withdrawal limit zi withdrawn from tdras in year i. an arbitrarily large positive number. a constant representing the maximum amount of personal income allowed in personal tax bracket k in year i (adjusted of inflation). a constant representing the maximum amount of estate income allowed in estate tax bracket k (adjusted for inflation). the life expectancy of the taxpayer at distribution year 1 (or the year in which the first withdrawal is made). the distribution year in which the tax payer turns age 70.5. the number of tdra investments. other (non-tdra) taxable income in year i. the percentage of the balance in tdra j which at the taxpayer’s death passes to his or her spouse or to a qualified charitable organization. present value interest factor at year i (i.e., pvi = (1 + 6)-q. 10s financial services review, 3(2) 1994 rij 3% yj w' "rj xik yk the expected rate of return in year i on tdra investmentj. the “reasonable” limit on the value of the taxpayer’s tdras at the time of their death (i.e. 9 = %n x (1 (1 + s)-“d)/s where nd is the remaining actuarial life expectancy at the taxpayer’s death.) amount by which the final value of tdras exceed .% . the maximum amount of gross estate value allowed without having to file an estate tax return adjusted for inflation (i.e., y= 600,000 x (1 +a”). the tih marginal personal income tax rate. the ich marginal estate tax rate. withdrawal in year i from tdra investmentj. withdrawal for year n, from tdra investment i made in year n, + 1. amount of total taxable income in year i subject to tax rate tfl amount of total estate subject to tax rate ti references abramson, e.m. (1989). iras: calculating your withdrawals. modern maturity (august september), 79-80. american demographics. (1989). 11(3), 19. geller, sm. (1988). minimum payments from iras. cpa journal, 58, 114. gould, c. (1988, march 10). the incredible lra withdrawal game. the new york times, p. 13f. internal revenue service. (1992). publication 448: federal estate and gift taxes. washington, dc: u.s. department of the treasury. internal revenue service. (1992). publication 590: individual retirement arrangements. washington, dc: u.s. department of the treasury. johnson, d.g. (1990). recalculating the life expectancy election. trusts & estates, 129( 1 l), 8-16. katz, c.i. (1990). planning to avoid and minimize penalty taxes upon withdrawing money from iras. the practical accountant (september), 54-66. mccommbe, l.r. (1989). when (and how much) to withdraw from an ira. personnel (february), 58-62. mcintosh, j.a., & hollinrake, j.d. (1988). minimizing the 15 percent tax on excess plan distributions. taxes, 66(4), 263-272. quinn, j.b. (1988, august 22). the next burning ira question: when to take the money out. the washington post, p. wb55. ragsdale, c.t., seila, a.f., &little, p.l. (1993). optimizing distributions from tax-deferred retirement accounts. personal financial planning, 5(3), 20-28. saftner, d., & fink, p. (1990). timing withdrawals from retirement accounts can increase tax savings. taxation for accountants, 44, 172-178. sage, j.a. (1988). selection of qualified retirement plan distribution options reflecting tra ‘86. taxes (april), 301-310. solbee, w.l. (1988). ira minimum distribution rules eased. the journal of taxation, 68,388. tiaa-cref. (1991). a practical guide to minimum distribution. new york: teachers insurance and annuity association college retirement equities fund. tritch, t. (1988). the fine art of drawing down tax-deferred accounts. money (july), 135-136. pii: s1057-0810(99)80013-4 financial services review, 7(1): 57-68 issn: 1057-0810 copyright © 1998 by jai press inc. all rights of reproduction in any form reserved. performance persistence of experienced mutual fund managers gary e. porter and jack w. trifts this study examines the performance of 93 fund managers over the 10 year period 1986 through 1995 using relative percentile ranks based on quarterly compounded, annual total returns measured against funds with the same investment objective. on average, managers with lo-year track records at the same fund do not perform better than man agers with shorter track records. also, for these experienced managers, superior performance in one five-year period is not predictive of superior performance over the next five years. however, inferior performance persists, particularly for funds with above average expense ratios. i. introduction tough guys finish first, notes the ad for a popular mutual fund. another ad touts a fund's number one ranking, along with its management's experience. still another reminds the investor that, while past returns are not indicative of future returns, their fund has 'beaten' other similar funds several years in a row. a casual survey of barron's or other investors' publications suggests that a mutual fund's ranking versus its peers is of great importance. while numerous studies have examined the issues of relative fund performance and perfor mance persistence by comparing mutual fund returns, few have directly examined the rel ative performance rankings of individual mutual fund managers. this study measures the performance of experienced fund managers and their ability to demonstrate consistent performance with the same fund over a ten-year period. by studying managers with this length of tenure at a single fund, the impact of changes in per formance related to different styles of individual fund managers is reduced. the intuitively appealing percentile ranking measure used by morningstar and other mutual fund rating services is the methodology used in this study. while total return is important, investors may use performance rank as a proxy for explicit comparisons of total return. the perfor mance measure is the rank of each manager's annual returns relative to all other managers within the same investment objective. measuring performance in this way controls for the gary e. porter • assistant professor of finance, college of business administration, university of central florida, orlando, florida 32816. jack w. trills ° professor of finance, crummer graduate school of business, rollins college, winter park, florida 32789. 58 financial services review 7(1) 1998 general level of risk assumed by the investor in the fund. managers have the ability to define distinct risk levels within objectives, so the analysis controls for exposure to system atic risk and to the risk associated with a style of investment strategy. finally, by placing more emphasis on annual performance than on total return over a specific period, this methodology provides evidence regarding the consistency of manager performance over time. this study provides evidence on the following research questions. one, as a group, do managers with 10-year track records at the same fund consistently outperform their peers? two, are mutual fund performance rankings in one five-year period predictive of the rank ings over the next five-year period? ii. literature review a number of recent studies have examined the issue of performance persistence in mutual funds. grinblatt and titman (1992) analyze performance of 279 funds over the period of 1975 to 1984 using a benchmark technique and find evidence that performance differences between funds persists over time. hendricks, patel, and zeckhauser (1993) study 165 no-load growth-oriented funds over the period 1974 to 1988 and obtain similar results. in a study of 728 mutual fund returns over the period 1976 to 1988, goetzman and ibbotson (1994) find that two-year performance is predictive of performance over the successive two years. volkman and wohar (1995) extend this analysis to examine factors that impact performance persistence. their data consists of 322 funds over the period 1980 to 1989, and shows performance persistence is negatively related to size and negatively related to levels of management fees. studies of performance persistence in mutual funds are not without contrary evidence. carhart (1997) shows that expenses and common factors in stock returns such as beta, mar ket capitalization, one-year return momentum, and whether the portfolio is value or growth oriented "almost completely" explain short term persistence in risk-adjusted returns. he concludes that his evidence does not "support the existence of skilled or informed mutual fund portfolio managers" (carhart, 1997, p. 57). in the kahn and rudd 1995 study of 300 equity funds and 195 bond funds between 1983 and 1993, only the bond funds show evi dence of persistence. in an article in this issue, detzel and weigand (1998) use a regression residual technique to control for the effects of investment style, size and expense ratios. they find, after controlling for these variables, no evidence of performance persistence. two other studies have used performance ranks. dunn and theisen (1983) rank the annual performance of 201 institutional portfolios for the period 1973 through 1982 without controlling for fund risk. they found no evidence that funds performed within the same quar tile over the ten-year period. they also found that ranks of individual managers based on 5-year compound returns revealed no consistency. bauman and miller (1995) studied the persistence of pension and investment fund performance by type of investment organization and investment style. they employed a quartile ranking technique because they noted that "investors pay particular attention to consultants' and financial periodicals' investment per formance rankings of mutual funds and pension funds" (bauman & miller, 1995, p. 79). they found that portfolios managed by investment advisors showed more consistent per formance (measured by quartile rankings) over market cycles and that funds managed by banks and insurance companies showed the least consistency. they suggest that this result may be caused by a higher turnover in the decision-making structure in these less consistent performance persistence 59 funds. this study controls for the effects of turnover of key decision makers by restricting the sample to those funds with the same manager for the entire period of study. iii. data and methodology our sample consists of mutual funds selected from morningstar's mutual funds ondisc ® database for the years 1986 through 1995. this source allowed the identification of fund managers for each fund, each year of the test period. a. sample selection the database was screened for funds with the following characteristics: 1. funds with at least one year (12 months) of total returns covering the calendar years 1986 through 1995. 2. the sample includes only "identifiable" managers. it does not include funds which list "management team" or "multiple managers" as manager because our objective is to draw inferences about the performance of managers who are known to inves tors. 3. to insure that there are enough funds to produce meaningful performance ranks within investment categories in a given year, only categories that contain at least 10 funds over the entire 10-year period are included in the study. 4. funds that had a change in objective or investment style over the period are not included in the sample. the final sample consists of 791 mutual funds. the funds represent the following seven investment objective categories defined by morningstar: 1. growth, which seeks capital appreciation by investing in equity with aboveaver age earnings potential. 2. aggressive growth, which seeks rapid growth of capital by making investments with greater than average risk, leveraged positions, and high turnover. 3. balanced, which invests in fixed proportions of stocks and bonds. 4. equity-income, which invests in equity securities with above-average yields. 5. growth and income, which invests in equity securities with above-average yields with some potential for growth appreciation. 6. small company, which invests in stocks of small companies, as determined by market capitalization or the value of assets. 7. specialty precious metals, which invests in equities of companies in the explora tion, distribution, or processing of precious metals. from this sample, all funds listing the same manager for the 1986-1995 period are identified. this group of 93 funds and their "experienced" managers is the subject of this study, and their performance is measured relative to the entire sample of 791 funds. 60 financial services review 7(1 ) 1998 b. performance analysis our analysis consists of four parts. first, the relative performance is measured for all 791 funds. second, the performance of the group of experienced managers is compared to all others. third, the performance persistence of the group of experienced managers over two successive five-year periods is compared and finally, the impact of expense ratios on our results is examined. to measure relative performance, compound annual returns are computed from monthly returns for each of the 791 funds in the sample. percentile rankings are then assigned to each manager, relative to other funds within the same investment category, for each year. with this measure, higher percentile ranks indicate superior performance. for example, the manager of a balanced fund in the 75th percentile has bettered 74 percent of all fund managers with the same investment objective that year, regardless of the age of the competing funds or the length of tenure of their managers. it is important to note that the percentile ranks are computed using all managers in the investment category for which there are data. while the focus of the study is the performance of managers with a 10-year track record, performance is measured against all peers within their fund's investment cat egory. next, the fiveand ten-year average performance percentiles based on annual rankings for each fund are computed. the measure, referred to throughout the paper as the "mean performance rank," or mpr, is the average rank achieved by a manager over a specific number of years. "average mpr" refers to the arithmetic mean mpr for a group of funds over a specific number of years. for example, we estimate the mpr of each of the 93 man agers with 10 years experience at the same fund. each of these 93 mprs represents a man ager's average ranking over the 10-year period. so, the average mpr of 48.7 shown in table 1 represents the overall level of performance for these 93 managers, over the 10-year period, relative to the other managers in the sample. in the first test, the sample of 10-year managers is divided into two groups by their mpr over the first five years of the sample period (1986-1990). managers with mprs above the 50th percentile are considered superior. those at, or below, the 50th percentile are considered inferior performers. next, the mprs of these two groups are calculated over the subsequent five years. if performance relative to their peers is non-random, the average mprs of the superior (inferior) performers should continue to be greater (less) than 50 in the 1991-1995 period. dividing the sample into two groups may distort the data because funds with margin ally poor performance are in the same category as funds with extremely poor performance, and vice versa for superior performers. therefore, our tests are repeated using extreme measures of superior and inferior performance. the managers with the 10 highest mprs are designated superior, while those producing the i0 worst mprs are deemed inferior. fund advertisements would certainly tout any manager among the top ten performers as likely to repeat their superior performance. finally, to test the importance of expense ratios in predicting performance persistence, the performance of subgroups of funds with above, and below, average expense ratios is examined. mutual fund performance studies have shown that expense ratios are inversely related to performance. golec (1996) concludes that investors should avoid funds with large operating and management fees. elton, gruber, das, and hlavka (1993) and ippolito (1989) provide evidence that funds with lower fees outperform funds with higher fees. gruber performance persistence 61 (1996) finds that the expenses of top performing funds are not different than for the funds with average performance, while fees of the poorer performers tend to be above average. also, fees for better performers tend to increase more slowly than fees of poorer performers. iv. results test results and implications are divided into three categories. first, the performance of the group of experienced managers is compared to the performance of all others in the sample for the entire 10-year period. next, performance persistence of the experienced managers is analyzed for two 5-year periods within the test period. finally, expense ratios are exam ined to identify their relationship with performance and performance persistence. a. performance of experienced managers relative to others table 1 shows the range of annum performance rankings, mean performance rankings, and expense ratios for the sample. the sample of managers with at least 10 years tenure at one fund (10-year managers) consists of 93 funds with a mean tenure of 17.4 years. the mean tenure for the control sample of 698 funds is 6.4 years. the range of 10-year mprs is narrower for the 93 managers with at least 10 years of tenure with the same fund compared to all other funds. if managers with longer tenure at the same fund are able to outperform those with shorter track records, their 10-year average mpr should be significantly above 50 and sig nificantly different from the average mpr of the sample group. however, the average mpr of experienced managers, 48.7, is not different from 50 (t = -1.07) and the average mprs of the two groups are not significantly different (t = -0.56). similarly, the mean expense ratios of the two groups of funds are not significantly different (t = 0.02). table 1 performance for 791 mutual funds including performance of 93 managers with a minimum of 10 years tenure with the same fund (1986--1995) managers o f the same f u n d for at least difference 10 years all other funds ten year vs. others percentile ranks n = 93 n = 698 (p-value) 1 range of 10-year mean performance ranks 7.2 to 73.2 0.33 to 902 average 10-year mean performance ranks 48.7 49.5 (0.1407) 3 (0.1144) average 10-year mean fund expense ratio 1.342% 1.339% average manager tenure 6.4 years 17.4 years -0 .8 (0.5608) 0.003% (0.9828) notes: annual performance ranks are determined by ranking annual momingstar total returns within each investment objec tive. ! the p-value indicates die likelihood that the two variables have equal population means. thus, significant differences are indicated by low p-values (i.e., less than .05 for significance at the 5 percent level). 2 these extremes are not unlikely because the minimum fund performance period is one year. in an investment objective with 300 funds, the fund in next to last place would have a performance rank of 0.33. 3 value in parentheses is significance level of t-test, ho: average mpr = 50. 62 financial services review 7(1) 1998 similarly, if experienced managers have any advantage over their peers, they should obtain a significantly higher number of superior annual rankings when their annual returns are ranked among the returns of 698 funds whose managers had less experience. however, there is no evidence of consistently superior performance. of the 930 possible annual per centile rankings of experienced managers (93 funds times l0 annual performance rank ings), only 430 were greater than the 50th percentile and 500 were at, or below, that level. if performances were random, the probability that there would be at most 430 annual supe rior performances out of 930 is only 1.18%. these results suggest that managers with lengthy experience at a fund have no particular ability to outperform other mutual funds within the same objective. the results suggesting that experienced managers, on average, perform no better than their less experienced peers for the sample period are particularly meaningful because our data set of 10-year managers imposes a survivorship bias on the test. malkiel (1995) pro vides evidence of positive return survivorship bias in equity mutual funds. he reports the mean annual returns of all general equity funds for the period 1982-1992 were signifi cantly greater for funds surviving the 1983-1992 period than for those which failed during the period (no failures were counted in 1982). however, there is no evidence regarding the direction of the performance bias that managers surviving for l0 years with the same fund may impart on our tests. if managers are retained (dismissed) for their ability to demon strate consistent superior (inferior) performance, the sample of experienced managers should contain more superior performers than inferior ones over the 10-year period. one interpretation of our result is that managers are not retained (dismissed) for demonstrating superior (inferior) performance. b. performance persistence of experienced managers the results from our analysis of lo-year managers provides evidence that, overall, experienced managers do not outperform their less experienced peers. the following tests are designed to provide evidence about the ability of superior experienced managers in one period to repeat their superior performance in the next. the performance persistence of experienced inferior managers is also examined. superior performance the results of tests of rankings persistence over two successive five-year periods are reported in table 2. as before, performance is determined relative to the 50th percentile. managers whose mpr over the period 1986-1990 was greater than the 50th percentile are deemed superior, while managers with mprs equal to, or less than, the 50th percentile are considered inferior. as shown in table 2, panel a, 43 of the 93 managers averaged above the 50th percentile, relative to all managers in the same investment category in our sample, over the period 1986-1990. the average mpr for these superior performers is 60.3, signif icantly greater than 50 at the 1% level. in the subsequent period, however, the average mpr of these superior managers fell to 51.1, an average mpr which is not significantly different from 50, and which represents a statistically significant decline of 9.2 points (t = 3.78). the number of managers demonstrating superior performance for the period 1991 1995 also dropped. only 23 of the 43 funds which were rated superior in the prior period performance persistence 63 table 2 test of persistence in performance rankings of managers with a minimum of 10 years tenure at the same fund (1986-1995) period i period 2 difference 1986-1990 1991-1995 (p-value) panel a: superior managers: mpr > 50 superior managers 1986--1990 60.3 51.1 -9 .2 5-year average mpr (0.0001)l (0.6036) (0.0003) number of superior managers 43 23 2 0 of 93 (0.7965) 2 (0.3804) 3 number of annual ranks > 50 144 109 -35 of 215 4 (0.0001)5 (0.4458) panel b: inferior managers mpr _ 50 inferior managers 1986-1990 39.4 46.0 6.6 5-year average mpr (0.0001) (0.0360) (0.023 l ) number of inferior managers 50 27 -23 of 93 (0.7330) (0.3359) 3 number of annual ranks <_ 50 166 139 -27 of 250 (0.0001) (0.0438) notes: mpr is the mean annual performance rank for each manager for the time period specified. 1 for mprs, figure in parentheses represents the significance level associated with the two-tail t-test, ho: mpr = 50. for differences in mprs between periods, value in parentheses represents the significance level associated with two-tail t-test, ho: mpr 1 mpr 2 = 0. 2 probability of observing at least n successes in 93 trials. 323 of the original 43 superior managers rated superior performance over the period 1991-1995. 27 of the original 50 inferior managers rated inferior performance over the subsequent period. p-value in parentheses represents the proba bility, p, associated with binomial test, p = probability of observing at least n successes in 43 trials for superior and 50 trials for inferior managers. 4 numbers of annual performance ranks satisfying condition out of 215 possible ranks for 43 superior managers in each 5-year period. 5 figure in parentheses represents probability, p, associated with binomial test, p = probability of observing at least 144 successes in 215 trials. same test is conducted for inferior managers. had mprs higher than 50 in the subsequent period. based on the normal approximation of the binomial distribution, there is a 38.15% probability that of the 43, at least 23 would have repeated as superior managers if the results were random. the implication is that there is at least a four in ten chance that 23 managers would demonstrate superior performance in the 1991-1995 period. the ability of 23 managers out of 93 to finish above the 50th per centile for two consecutive periods is also consistent with even odds (50% likelihood in the first period, times 50% in the second, times 93 managers equals 23.25 expected occur rences). there are 215 annual performance ranks for the 43 experienced managers in the period 1991-1995 (43 managers times 5 years). only 109 annual performance ranks by these superior managers were greater than 50 during this period, compared to 144 in the period 1986-1990. based on the binomial distribution, 109 successes out of 215 trials is not unlikely if the process is random, suggesting that the evidence on the drop in mprs reported above was not the result of a couple of particularly poor managers. 64 financial services review 7(1) 1998 inferior performance the evidence suggests that superior management performance over a five-year period is not predictive of continued superior performance over the subsequent five-year period. is inferior performance predictive of continued inferior performance? table 2, panel b presents the analysis of the 50 managers with inferior performance over the 1986-1990 period. the average mpr, which was 39.4 over the initial period, improved to 46.0 in the latter period. the increase of 6.6 points is statistically significant at the 5% level, (t = 2.32, two-tail test). however, the average mpr of the inferior performers remains a statistically significant 4.0 points below the 50th percentile in the subsequent period (t = -2.11, two-tail test, significant at 5%). so, while the performance of the inferior managers reverted toward the 50th percentile in the subsequent period, it remained below average relative to all other managers. a comparison of the mean performance of superior managers and inferior managers during the 1991-1995 test period reveals a convergence to average performance. the aver age mprs for inferior managers, 46.0, is different than that of the superior managers, 51.1, in the latter period at the 10% significance level (t = 1.82, two-tail test). twenty seven of the 50 managers with inferior performance in the first period remained inferior performers in the second period. this is consistent with random performance as, under the binomial distribution, the likelihood that a minimum of 27 inferior managers would repeat is 33.59%. however, there is a low probability, 4.4%, that as many as 139 of the annual ranks of this group could be inferior in the subsequent period. that is, as a group, these managers had more individual years of inferior performance that would have been predicted by chance. the results of table 2 show that managers outperforming a majority of their peers from 1986 through 1990 failed to repeat their superior performance in the subsequent 5 year period. the performance of managers who under-performed their peers in the earlier period improved in the subsequent period, but remained inferior to the subsequent perfor mance of superior managers. the results suggest that, while the performance of superior managers reverts toward the 50th percentile, the performance of inferior managers does not fully revert. performance extremes when experienced managers with the 10 highest mprs are designated superior, and those producing the 10 worst mprs are deemed inferior, the results of these tests, shown in table 3, are qualitatively identical to those reported when dividing the sample relative to the 50th percentile. the average mpr for the superior group fell from 71.7 to 51.2, a sta tistically significant drop of 20.5 points. of the 10 superior managers in the period 1986 1990, only 1 fund repeated their top ten performance in the subsequent period. the ability of 1 fund of the 93 to rank in the top ten for two consecutive periods is consistent with a random performance process. only six of the top ten funds for 1986-1990 subsequently had mprs above 50, and only 25 of the 50 annual observations were above the 50th per centile over the 1991-1995 period. these results strengthen the evidence against perfor mance persistence by showing that even top management performance over a five-year period, on average, is not predictive of continuing above average performance in subse quent periods. performance persistence 65 table 3 test of persistence in performance rankings of managers with a minimum of 10 years tenure at the same fund (1986-1995) period i period 2 difference 1986--i 990 1991-1995 (p-value) panel a: superior managers top ten mprs superior managers 1986-1990 71.7 51.2 -20.5 5-year average mpr (0.0001)l (0.2389) (0.0005) number of managers 10 1 9 in top ten number of annual ranks > 50 44 25 -19 of 50 fund years (0.0001)2 (0.5561) panel b: inferior managers bottom ten mprs inferior managers 1986-1990 22.3 40.6 18.3 5-year average mpr (0.0001) (0.0564) (0.0029) number of managers 10 3 7 in bottom ten number of annual ranks < 50 43 29 -14 of 50 (0.0001) 2 (0.1611) notes: mpr is the mean annual performance rank for each manager for the time period specified. 1 for mprs, figure in parentheses represents the significance level associauxl with the two-tail t-test, ho: mpr -50. for differences in mprs between periods, value in parentheses represents the significance level associated with two-tail t-test, ho: mpr 1 mpr 2 ~= 0. 2 the probability of observing at least n successes in 50 trials. the results for the 10 worst performers are also consistent with the results of our prior tests. the average mpr of the ten worst performers improved by 18.3 points, a statistically significant improvement, but performance in the subsequent period is significantly less than the 50th percentile at the 10% level (t = 1.95, two-tail test). performance of superior managers was greater than performance by inferior managers in the subsequent period by 10.6 points, which, like our earlier results, is significant at the 10% level (t -1.67, two-tail test). these results provide strong evidence that the performance of even top managers reverts to average, and marginal evidence that the performance of the worst managers does not fully mean revert. c. performance persistence and expense ratios in this section, the analysis is designed to determine to what extent expense ratios are associated with the performance rankings of the experienced managers. the results in table 4 are grouped by time period. as shown in table 4, panel a, expense ratios across funds are significantly lower for superior funds, averaging 0.47% less than expense ratios of inferior funds over the period 1986-1990 (t = 3.96, two-tall test) and 0.727% lower for the period 1991-1995, significant at the 5% level (t = 4.37, two-tail test). this result is con sistent with evidence provided by volkman and wohar (1995), and others, who find evi dence that funds with low management fees show persistent positive performance while those with high fees show persistent negative performance. panel a in table 4 also pro 66 financial services review 7(1) 1998 table 4 relationship between fund expense ratio and performance persistence for managers with a minimum of 10 years tenure with the same fund (1986-1995) period i period 2 difference 1986-1990 1991-1995 (p-value) panel a: mean expense ratios of 93 experienced managers superior managers, n = 43 1.011% 0.978% 0.033% 5-year expense ratio (0.4168) inferior managers, n = 50 1.481% 1.705% -0.224% 5-year mean expense ratio (0.2291) difference -0.470% -0.727% (p-value) (0.0001 ) (0.0001 ) 5-year performance/fund average mpr average mpr expenses: 1986-1990 period i period 2 difference by fund 1986-1990 1991-1995 (p-value) panel b: impact of fund expense ratios on manager performance persistence superior performance/ 62.4 42.7 -19.7 high expense, n = 8 t (0.0054) i (0.1341) 2 (0.0027) 4 superior performance/ 59.8 53.0 -6.80 low expense, n -35 (0.0001) (0.1919) (0.0243) inferior performance/ 34.7 37.0 2.3 high expense, n -19 (0.0001) (0.0002) (0.5954) inferior performance/ 42.3 51.6 9.3 low expense, n --31 (0.0006) (0.4604) (0.0028) notes: mpr is the mean annual performance rank for each manager for the time period specified. i number of firms of 93 satisfying condition. t-test, degrees of freedom = n*5l years. 2 p-value from t-test under the null hypothesis that the average mpr is different from 50. 3 p-value from two-tail t-test under the null hypothesis that the averages of mpr 1 mpr 2 = 0. vides evidence that for both superior and inferior performers, there is no significant change in expenses as a proportion of net asset value over the ten-year period. results in panel b of table 4 are divided such that, as before, superior managers had mprs greater than the 50th percentile for the period 1986-1990, while all others are deemed inferior managers. expense ratios are divided into high and low categories. expense ratios were listed as high if the fund's ratio was greater than the full sample mean (1.37 percent) during the period 1986-1990, and listed as low, otherwise. panel b shows that, regardless of expense level, performance of superior managers reverts to levels near 50 in the subsequent period. however, the average mpr of the low expense funds in the 1991-1995 period is 10.3 points greater than the average mpr of high expense funds (t = 1.93, two-tail test, significant at 10%). the performance of inferior managers with low expense ratios improves to average for 1991-1995, while the average mpr of those with high expense ratios remains signifi cantly below average at the 5% level. the average mpr of low expense funds during 1991-1995 is significantly greater than the performance of the high expense funds by 14.6 points at the 5% level (t = 3.60, two-tail test). the evidence presented in table 3 suggests performance persistence 67 that, not only do successful managers, on average, fail to reproduce their prior success rel ative to their peers, but managers of funds with relatively high expense ratios tend to per form, on average, more poorly than their peers in subsequent years. v. conclusion using relative annual performance ranks, this study finds no evidence that experienced mutual fund managers outperform their less experienced peers. also, the superior relative performance of experienced managers measured over the five-year period 1986--1990 was not predictive of superior performance over the subsequent five years. while superior per formance is not persistent, there is evidence that inferior performance does persist. poorly performing managers tend to improve their rankings in the next period, but their perfor mance remains below that of superior managers. one reason for this appears to be differ ences in expense ratios. managers with inferior performance and greater than average expense ratios during 1986-1990 perform more poorly during 1991-1995 than funds with below average expense ratios. the results hold after controlling the systematic risk of the fund and risk exposure related to individual management styles defined by morningstar. the results of this study are consistent with those of detzel and weigand (also in this issue). using a regression residual approach, they also find no persistence in fund perfor mance after controlling for investment objective and other common portfolio factors. explicit control for investment objective and style category, as well as longer performance periods, are key facets of the methodologies of our two studies. though each employs unique methodologies relative to recent work in this area, these studies provide further evi dence that neither expertise nor past performance is indicative of future, long-run, superior mutual fund performance. acknowledgment: the authors thank editor karen eilers lahey and the anonymous ref erees for their valuable contributions. thanks also to qingsheng mou and neil monaghan for their research assistance and jim gilkeson and stan atkinson for their helpful com ments. we also appreciate the comments generated by a presentation to ucf faculty, particularly, stan smith, shawn phelps, john cheney, and ronnie clayton. references bauman, w. s., & miller, r. e. (1995). portfolio performance rankings in stock market cycles. financial analysts journal, 51, 79-87. carhart, m. m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57-82. detzel, f. l., & weigand, r. a. (1998). explaining persistence in mutual fund performance. finan cial services review, 7(1 ), 45-55. dunn, p. c., & theisen, r. d. (1983). how consistently do active managers win? journal of portfolio management, 9, 47-51. elton, e. j., gruber, m., das, s., & hlavka, m. (1993). efficiency with costly information: reinter pretation of evidence from managed portfolios. the review of financial studies, 6, 1-22. 68 financial services review 7(1) 1998 goetzman, w. n., & ibbotson, r. g. (1994). do winners repeat? journal of portfolio management, 20, 9-18. golec, j. h. (1996). the effects of mutual fund managers' characteristics on their portfolio perfor mance, risk, and fees. financial services review, 5, 133-147. grinblatt, m., & titman, s. (1992). the persistence of mutual fund performance. journal of finance, 47, 1977-1984. gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51,783-810. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974-1988. journal of finance, 48, 93-130. ippolito, r. a. (1989). efficiency with costly information: a study of mutual fund performance. quarterly journal of economics, 104, 1-23. kahn, r. n., & rudd, a. (1995), does historical performance predict future performance? financial analysts journal, 51, 43-52. malkiel, b. g. (1995). returns from investing in equity mutual funds 1971 to 1991. journal of finance, 50, 549-572. volkman, d. a., & wohar, m. e. (1995). determinants of persistence in relative performance of mutual funds. journal of financial research, 18, 415-430. pii: s1057-0810(96)90022-0 from the editor karen eilers lahey financial services review (fsr) is the official publication of the academy of fbndal services. this refereed academic journal encourages rigorous empirical research that explores all aspects of individualjhancial management. the current issue consists of six articles that provide boundary spanning efforts across a variety of financial topics and should be of interest to both academics and financial ser vice providers. the first three examine the selection of mutual funds, individual stocks, and enhanced cds, while the fourth tests a new financial criteria for determining the churning of individual accounts. the last two articles focus on cooperative efforts between academ ics and practitioners interested in furthering research in financial services. the first article, “bquity fund size and growth: implications for performance and selection,” is written by conrad ciccotello and terry grant. their empirical research tests the impact of fund size on the performance and subsequent selection of mutual funds. they conclude that individuals who are have an aggressive growth objective should select smaller funds which have the potential to be the next “magellan.” for investors whose objective is growth and income, they find no systematic relationship between performance and size. in addition, they provide an excellent review of the recent surge of academic stud ies that reflect the extraordinary growth of mutual funds during the 1!39os. mark walker and gay hatfield choose to test a widely available source (usa today) for the selection of individual stocks in their article, “professional stock analysts’ recom mendations: implications for individual investors.” they first explore a variety of tech niques that include stock index benchmarks, event study methodology, and the sharpe, treynor, and jensen indexes to measure analysts’ ability to identify mispriced stocks. results indicate that the most defensible technique is the sharpe, treynor, and jensen mea sures. walker and hatfield then test individual investor’s ability to profit from analysts* recommendations and find that it is difficult for individual investors to benefit. the third article shifts the individual investment selection choice to bank certificates of deposit and is entitled, “computing yields on enhanced cds,” robert brooks provides a framework and a model for comparing the effective annual rate for standard cd products and those cds with an embedded derivative. this research brings together two strands of academic research: (a) cds from the perspective of the role of government guarantees and the factors that influence changes in cd rates, and (b) interest rate contingent claims valu ation for the enhanced cd product. brooks’ clear message is that enhanced cd valuation is much more complex than the simple cds of an earlier time period, and both individual v vi financial services review 5( 1) 1996 investors and bank executives must carefully analyze their objectives in selecting these new products. stewart brown’s article, “churning: excessive trading in retail securities accounts” provides another boundary spanning effort at the intersection of legal and financial analysis in security brokers handling of individual accounts. drawing on his experience as an aca demic and an expert witness, he argues for an alternative financial model (commission to equity ratio) to the current criteria (turnover ratio) in determining if churning has occurred. his arguments for a different criteria are illustrated with summary data from 23 actual churning cases. additionally, he provides statistics on actual commissions paid for both equities and options. in an effort to bring about consensus between academics and financial planners, sue greninger, vickie hampton, karrol kitt, and joseph achacoso employ a delpi approach in their research. their article entitled, “ratios and benchmarks for measuring the finan cial well-being of families and individuals” reviews academic literature on ratio analysis starting with the traditional business applications and then moves to its use for individuals. they report on 22 ratios that are recommended in their study and then propose a smaller list of ratios to be included in a financial well-being profile. their suggested series of ratios include measures of liquidity, savings, asset allocation, inflation protection, tax burden, housing expenses, and insolvency. the last article, “an emerging partnership: afs and the cfp board” by dede pahl discusses the needs and benefits of a partnership between academic members of the acad emy of financial services and the cfp board. de& is an active personal boundary span ner whose official title is director of certification for the cfp board and a board of directors member of the academy of f’inancial services. she explains the history and mission of the cfp board and the partnership that is developing between the two organi zations. the last item listed in the contents is a new section that is being initiated with this issue of fsr. douglas kahl has volunteered to be responsible for this section, which is entitled “book, software and web site reviews.” he has started the section by reviewing a book written by beam and mcfadden that is titled, employee benefits. john clinebell has reviewed garman and forgue’s personal finance textbook. the section concludes with a review of the web site for the london international futures and options exchange. doug is very interested in hearing from you if you are willing to volunteer to do a review or offer materials to be reviewed. i would like to thank the editorial board for all of their advise and the reviews they have promptly returned as well as the 54 ad hoc reviewers who have responded to my request to help shape the future of the financial services review. i look forward to hearing your reactions to the articles in this issue. i will publish interesting responses in the new member section, afs notes section. in addition, your ideas for future articles can be transmitted to me in the form of letters, e-mail, or hits on our fsr web site. the address to the web site is [http: //www.vakron.edu/cba/fsr/frs.htm/l. pii: s1057-0810(99)00009-8 mean and pessimistic projections of retirement adequacy yoonkyung yuha, sherman hannab,*, catherine phillips montaltoc awon-building 6th floor, yumri-dong, mapo-ku, seoul korea (rok) bprofessor, consumer and textile sciences department, the ohio state university, 1787 neil ave., columbus, oh 43210-1295, usa cassistant professor, consumer and textile sciences department, the ohio state university, 1787 neil ave., columbus, oh 43210-1295, usa abstract retirement adequacy is estimated using a 1995 united states sample of households. based on mean lognormal portfolio projections and current contribution rates, 52% of households are adequately prepared for retirement. based on pessimistic projections, only 42% of households are adequately prepared. a regression of the ratio of projected wealth to needs at retirement shows that adequacy increases with stock share (mean projection) and the impact increases with time until retirement. with pessimistic projections, there is no significant relationship between stock share and the adequacy ratio. planned retirement age and household spending behavior are each significantly related to the adequacy ratio. © 1998 elsevier science inc. all rights reserved. 1. introduction the elderly population in the united states is growing at a much faster rate than the population as a whole. the number of persons 65 years old and over in 1996 (34 million) was 11 times larger than in 1900 (3 million). over this same period, the number of persons under 65 years old only tripled. the elderly population is projected to more than double by the middle of the 21st century to 79 million, at which time elderly persons will represent 20% of the united states population (u.s. bureau of the census, 1998). while the number of older persons is rapidly growing, the financial situation for future retirees remains uncertain. retirement income is commonly assumed to come from the triad of social security, private * corresponding author. tel.:11-614-292-4584; fax:11-614-292-7536. e-mail address:hanna.1@osu.edu (s. hanna) financial services review 7 (1998) 175–193 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00009-8 pensions, and personal saving. planned reductions in social security benefits for retirement before age 67 and the shift away from defined benefit pension plans (u.s. general accounting office, 1996) increase the importance of personal saving as a source of retirement income. personal savings rates in the united states have decreased recently to very low levels, reaching20.2% in september, 1998 (u.s. department of commerce, 1998), so individuals must carefully determine how much to save for retirement and how to invest savings in order to be prepared financially for retirement. retirement savings can be invested in a variety of ways, ranging from traditional savings accounts with relatively low rates of return, to publicly traded stocks and mutual funds offering relatively high rates of return. previous research has documented variation across individuals in the choice of investment vehicles, or portfolio allocation (bajtelsmit and van derhei, 1996). however, an individual’s retirement funds are invested for a particular period of time during which market rates of return can vary. an individual investor may receive an unusually high or unusually low rate of return. while it is routine to use average rates of return to project retirement resources, this may overor under-estimate the ultimate accumulation depending on the actual market rates of return. a better picture of the range of possibilities can be provided by making more than one projection of retirement resources. the purpose of this study is to investigate the adequacy of retirement wealth using both mean and pessimistic projections of retirement wealth. unique contributions of this research include use of household specific information on planned retirement age and portfolio allocation, projection of retirement wealth using asset specific growth rates, estimation of retirement needs based on household expenditure functions, and comparison of adequacy based on mean and pessimistic projections. 2. review of the literature 2.1. related empirical research analysis of retirement wealth adequacy requires information on the resources that will be available in retirement, as well as the amount needed to finance consumption during those years. retirement adequacy can be defined as having resources exceed the amount needed to finance desired retirement consumption. a “retirement gap” exists when resources are less than the amount needed. duncan et al. (1984) use this framework to determine savings goals for retirement. in order to implement such a framework, it is necessary to estimate the resources an individual will have accumulated at the date of retirement, as well as the amount needed to finance consumption during the retirement years. a variety of techniques have been used in previous research to estimate the amount needed to finance retirement consumption and to project the resources available for retirement. 2.1.1. retirement needs estimation of retirement needs is often based on the life cycle hypothesis, and the assumption that individuals desire to smooth the level of consumption over their lifetime (modigliani and brumberg, 1954). bernheim et al. (1997) challenge the validity of standard 176 y. yuh et al. / financial services review 7 (1998) 175–193 life cycle models to explain actual variations in saving and wealth. the most commonly used method to estimate the level of retirement need is to specify the percentage of preretirement income that represents the desired consumption level in retirement. this percentage is commonly referred to as the “replacement rate.” duncan et al. (1984) set the replacement rate equal to 100% in their standard model, but use rates in the range of 70% to 90% in their calculation of hypothetical cases. other researchers adopt similar replacement rates in empirical research (burns and widdows, 1988, 1990; mitchell and moore, 1997). palmer (1989, 1994) calculates replacement rates based on data from the consumer expenditure survey. these replacement rates (based on gross income) range from 65% to 85%, vary with marital and employment status, and generally decline with income. palmer’s replacement rates are used by other researchers to estimate retirement needs (li et al., 1996). approaches other than the replacement rate are used to estimate retirement needs. moore and mitchell (1997) jointly estimate replacement rates and savings rates given current earnings and projected assets. yuh et al. (1998) use the household’s level of preretirement consumption as a proxy for the household’s desired level of retirement consumption. 2.1.2. retirement wealth in order to estimate the level of retirement wealth, it is necessary to determine which resources will be available for retirement, as well as to determine the value of the accumulated resources at the point of retirement. empirical measures of retirement wealth commonly include the value of financial assets, but the treatment of nonfinancial assets, particularly the value of home equity, varies. duncan et al. (1984) define retirement income to include social security, private pensions, house equity at retirement, and other assets not earmarked for other purposes (such as children’s education). burns and widdows (1988), li et al. (1996), moore and mitchell (1997), and yuh et al. (1998) use similar definitions of retirement income that include the value of home equity. burns and widdows (1990) examine the sensitivity of retirement savings rates to the treatment of home equity. bernheim (1996) excludes home equity from the calculation of assets available to finance consumption during retirement. home equity accounts for the largest share of total household wealth in the united states. it can be converted to a more liquid form by selling the house or using debt instruments such as second mortgages, home equity loans, and reverse mortgages. while most people do not sell their homes in retirement or use reverse mortgages to finance retirement consumption, home equity represents an important potential resource. furthermore, a homeowner will be better off in retirement than an otherwise similar renter, so inclusion of home equity results in more valid comparisons between owners and renters. using a comprehensive measure of asset availability is particularly important when evaluating the resources that could be used to finance expenditures during retirement, including costs of long-term care (mitchell and moore, 1997; andrews, 1993). once the components of retirement wealth are defined, it is necessary to determine the value of the accumulated resources at the point of retirement. li et al. (1996) use panel data containing information on households at the point of planned retirement. retirement wealth is calculated as the household’s net worth at the point of planned retirement plus the present value of income streams from social security and other pension plans. much research 177y. yuh et al. / financial services review 7 (1998) 175–193 focuses on determining the “future” retirement wealth adequacy of currently pre-retired households, particularly baby boomers. for pre-retired households, wealth at the point of retirement in the future must be projected, and therefore information is needed on the future rates of return, or growth rates, for assets. various approaches have been used to project the value of retirement wealth, including use of a common growth rate for all financial assets (burns and widdows, 1988), and use of asset specific growth rates (moore and mitchell, 1997; yuh et al., 1998). for example, burns and widdows (1988) use growth rates of 0% and 3%. moore and mitchell (1997) project individual components of net financial wealth assuming the growth rates are geometric averages of historical real returns. while this approach is an improvement over the use of a common growth rate for all financial assets, it ignores risks associated with investments due to changes in interest rates over time. yuh et al. (1998) also project individual components of financial and nonfinancial wealth, but use both average and pessimistic growth rates generated from historical rates of return and a lognormal forecasting model. 2.1.3. savings rate in general, the previous research suggests that pre-retired people are not adequately prepared financially for their retirement and thus need additional savings in order to have adequate retirement wealth. moore and mitchell (1997) examine the adequacy of asset holdings of persons on the verge of retirement using data from the health and retirement study (hrs). they compare the projected value of assets at retirement with estimated retirement needs, and determine the level of saving needed to maintain the retirement consumption. retirement is assumed to occur at ages 62 and 65. they conclude that the majority of older households will not be able to maintain current levels of consumption into retirement without increasing savings. in particular, the median hrs household would have to save an additional 16% of earnings to maintain the pre-retirement consumption level for age 62 retirement, or an additional seven percent of earnings for retirement at age 65. in a related study, mitchell and moore (1997) use the hrs data to examine the adequacy of retirement wealth for a household with characteristics similar to hrs median characteristics (a married couple household, husband and wife both age 56 in 1992, with an annual household income of $46,000). wealth accumulation is projected for retirement at age 65 assuming a portfolio of 60% bonds and 40% stocks. the wealth accumulation is compared to a retirement needs calculation using replacement rates of 70% and 80%. substantial shortfalls in retirement wealth accumulations are found, and the authors conclude that the median american on the verge of retirement has accumulated too little wealth to support a comfortable retirement. bernheim (1996) calculates the ratio of actual savings to savings needed to maintain the preretirement level of living during the retirement years for respondents to a merrill lynch survey. a computer simulation model is used to determine the prescribed savings levels and these levels are compared with actual savings behavior. an adequacy index is developed based on actual savings as a percentage of prescribed savings for three cases: pessimistic, optimistic, and midpoint. the index indicates a significant shortfall in the retirement savings of the baby boom generation. the overall index at the midpoint (36%) indicates that the 178 y. yuh et al. / financial services review 7 (1998) 175–193 typical baby boom household needs to nearly triple its rate of saving to maintain the preretirement consumption level in the retirement years (bernheim, 1996, p. 22). burns and widdows (1988) apply the framework developed by duncan et al. (1984) to data from the 1983 survey of consumer finances to estimate savings rates needed to adequately fund baby boomers’ retirement. sizeable retirement gaps are generally found across all age and income groups. the authors conclude that the average family needs to increase the current level of saving in order to meet retirement needs. 2.1.4. correlates of retirement wealth adequacy previous research has analyzed the correlates of retirement wealth adequacy. li et al. (1996) use data from the national longitudinal survey of older men to compare needed resources and actual resources at the expected date of retirement for each household. the results suggest that men born between 1907 and 1921 are not well prepared financially for retirement. only 46% of the sample has accumulated retirement wealth at the expected retirement age that is adequate to maintain the pre-retirement consumption level during retirement. retirement age is found to be an important factor affecting retirement wealth adequacy. being white, having a longer planning horizon, planning to retire at age 65 or later, and asset ownership all increase the probability of having adequate retirement wealth. yuh et al. (1998) use data from the 1995 survey of consumer finances to analyze retirement adequacy of pre-retired households and estimate that slightly more than half of these households will be able to maintain the preretirement consumption level during the years following retirement. the probability of having adequate retirement wealth is found to increase with income, to be higher for households that have defined benefit or defined contribution pension plans, and for households that own their home mortgage free. the two most important factors related to retirement wealth adequacy are planned retirement age of the householder and household spending behavior. planned retirement age is positively related to the probability of adequate retirement wealth. spending at least as much as household income decreases the probability of adequate retirement wealth. a common limitation that cuts across previous research on retirement adequacy is uniform assumptions that do not allow for variation across households. these assumptions often relate to planned retirement age, portfolio allocation, growth rates for assets, and retirement needs. by not allowing for variation across households, the corresponding estimates of retirement adequacy are prone to over or under represent actual adequacy. for example, planned retirement age affects both the amount of time prior to retirement during which assets accumulate as well as the amount of time that will be spent after retirement. retirement at later ages, ceteris paribus, increases the time over which assets accumulate (thus increasing retirement resources), as well as decreases the amount of time spent after retirement (thus decreasing retirement needs). both of these factors would influence the measure of retirement adequacy. therefore, planned retirement age is an important variable in estimation of retirement adequacy, and information on the actual planned retirement age should be used instead of assuming retirement at given ages. similarly, household specific information on portfolio allocation and retirement needs, as well as asset specific growth rates, will improve the accuracy of estimates of retirement wealth adequacy. this study addresses several of these limitations. in contrast to previous studies, house179y. yuh et al. / financial services review 7 (1998) 175–193 hold specific information on planned retirement age, portfolio allocation, and asset specific growth rates are used to project retirement wealth. rather than using set replacement rates, household expenditure functions are used to estimate retirement needs. the ratio of projected wealth to needs at retirement, a continuous measure of wealth adequacy, is analyzed in contrast to dichotomous indicators analyzed in previous work (li et al., 1996; yuh et al., 1998).the adequacy of retirement wealth is analyzed based on both mean and pessimistic projections of asset growth in order to consider the range of possibilities. 2.2. conceptual framework under a life cycle model, assets are accumulated during an individual’s work life mainly to finance consumption after retirement when earned income is reduced. a generally accepted goal of retirement planning is to provide enough income in retirement to prevent the level of living from dropping much below the preretirement level (schulz, 1992). thus, retirement wealth can be defined asadequateif total retirement income is equal to or greater than the total desired retirement consumption level (cf. hatcher, 1997). the desired retirement level of living can be estimated from information on the preretirement level of living, assuming that individuals would like the same consumption level after retirement as before retirement. retirement wealth adequacy at the point of retirement (age r) can be defined as follows: ar 1 o t51 t2r bt /(1 1 r)t $ o t51 t2r ct /(1 1 r)t (1) where ar 5 total asset accumulation upon retirement (age r), bt 5 pension income at age t, ct 5 consumption level at age t, r 5 retirement age, and t 5 age at death. according to this equation, retirement wealth at the point of retirement is adequate if the sum of the accumulated assets plus the present value of pension income (including social security and annuities) is at least as large as the present value of retirement consumption. 3. methodology to operationalize the conceptual model retirement wealth must be clearly defined and methods for projecting the levels of retirement wealth and retirement needs must be selected. (for more details on the methodology refer to yuh, 1998, and yuh et al., 1998). 180 y. yuh et al. / financial services review 7 (1998) 175–193 3.1. empirical definition of retirement wealth a comprehensive measure of retirement wealth is used in this study. retirement wealth is defined to include financial assets, nonfinancial assets including housing wealth, and retirement income from defined contribution plans, defined benefit plans, and social security. in order to determine the level of retirement wealth at the planned retirement age, the value of current assets must be projected forward. this requires information on future rates of return for these assets. 3.2. projection of future rates of return total wealth available for retirement from financial assets, nonfinancial assets, and defined contribution plans is projected using future real rates of return for each asset category. future real rates of return are projected separately for stocks, bonds, money market instruments, business assets, and real estate assets using data on historical rates of return and a lognormal forecasting model (ibbotson associates, 1995). the lognormal forecasting model is used because, unlike the normal model, the lognormal model does not project negative values and therefore may produce more plausible predictions (crow and shimizu, 1988). using the lognormal model, it is straightforward to form probabilistic forecasts of both compound rates of return and ending period wealth values. wealth at time n (assuming reinvestment of all income and no taxes) is: ln(wn) 5 ln(w0) 1 ln(1 1 r1) 1 ln(1 1 r2) 1 . . . 1 ln(1 1 rn) (2) where wn 5 the wealth value at time n w0 5 the initial investment at time 0 r1, r2.....rn 5 the total returns on the portfolio for the rebalancing period ending at times 1, 2, and n. the geometric mean return over the same period, rg, is: rg 5 (wn/w0) 1/n 2 1 (3) where rg 5 the geometric mean return n 5 the inclusive number of periods. in the lognormal forecasting model, the expected value (m) and standard deviation (s) of the natural logarithm of the return relative of the portfolio can be calculated from the expected return (m) and standard deviation (s) of the portfolio as follows: m 5 ln(1 1 m) 2 ~s2/2) (4) s5 {ln[1 1 ~s/1 1 m!2]} 1/2 (5) where 181y. yuh et al. / financial services review 7 (1998) 175–193 ln 5 the natural logarithm function. given the logarithmic parameters of a portfolio (m and s), a time horizon (n), and the z-score of a percentile (z), the percentile of the geometric mean return for an asset i is calculated as: ri 5 exp {mi 1 z(si /n 1/2)} 2 1 (6) where ri 5 percentile of the geometric mean return of asset i mi 5 expected value of natural logarithm of the return relative of asset i si 5 standard deviation of natural logarithm of the return relative of asset i z 5 the z-score of the percentile n 5 investment horizon. using this equation, it is possible to calculate the various percentiles of the geometric mean return over various time horizons. in order to compare adequacy under mean and pessimistic conditions, rates of return at the 50th percentile and the 5th percentile are selected. the rate of return for the 50th percentile of each asset is used as the projected return for the mean portfolio performance, and the rate of return for the 5th percentile is used as the projected return for the pessimistic portfolio performance. data for historical rates of return from thestocks, bonds, bills and inflation yearbook published by ibbotson associates (1995) are used to provide information on the mean and variance of the real rate of return for specific asset categories. the 1995 yearbook provides historical return data from january 1, 1926 through december 31, 1994 for six categories of financial assets: small capitalization stocks, large stocks (s&p 500), corporate bonds, intermediate government bonds, long term government bonds, and treasury bills. real estate returns from 1947 to 1982 estimated by ibbotson and siegel (1984) are used to produce lognormal projections of future real rates of return for real estate assets. this real estate dataset is comparable to the historical return data in the ibbotson yearbook, and is the longest period of annual return data for real estate available. information is available for residential real estate, farm real estate, business real estate, and composite real estate (average of the three categories). 3.3. estimation of retirement needs following the assumption of the life cycle savings model (modigliani and brumberg, 1954) it is assumed that households desire to maintain the preretirement level of living during retirement. retirement needs are defined as the total wealth needed to provide the level of preretirement consumption during all years of retirement. wn 5 c*{[1 2 ~1 1 rr) 2d]/rr% (7) where wn 5 retirement need (present value of total consumption needed in retirement), c 5 annual consumption during retirement, rr 5 (expected) real interest rate from retirement to death, and 182 y. yuh et al. / financial services review 7 (1998) 175–193 d 5 retirement period (the number of years from retirement age to death). 3.3.1. annual consumption during retirement a household expenditure function is used to predict annual consumption during retirement. the household expenditure function is estimated using data from the interview component of the 1993–1994 consumer expenditure survey. the consumer expenditure survey is conducted by the united states bureau of the census for the bureau of labor statistics (u.s. bureau of labor statistics, 1996) and is the most comprehensive source of detailed information on expenditures for goods and services by households in the united states. for this study households that are interviewed in four consecutive quarters (excluding the initial bounding interview) between the second quarter of 1993 and the fourth quarter of 1994 are retained. for each household, data on the four consecutive quarters of expenditure are summed to obtain actual annual household expenditures. all dollar values are adjusted to 1994 dollars. a box-cox test is used to determine the best functional form for the expenditure equation, and the double-log model is selected: ln (ci) 5 f [ln (incomei), zi] where zi is a vector of household characteristics excluding the income variable. a chow test is used to compare separate regressions for households that do and do not spend less than income to a regression on the pooled sample of households. the chow-test rejects the pooled model at the 1% level of significance, indicating that the two separate regressions provide a better fit than the regression on the pooled sample. these regression tables are available from the authors. for each household in the sample, the appropriate household expenditure function (separate functions for households that spend less than income and households that do not) is used to predict annual consumption in the year preceding retirement. the predicted preretirement consumption level is used as a proxy for the desired level of retirement consumption. 3.3.2. real interest rate the appropriate real interest rate (rr) for discounting total retirement needs should be based on a household’s investment behavior. it is typically assumed that retired people invest very conservatively because of their low level of risk tolerance during retirement. in this study a real discount rate of 2.3% is used to calculate total retirement needs. 3.3.3. retirement period the retirement period is determined as the difference between an individual’s expected age at death and age at planned retirement. expected age at death is estimated by gender and marital status using actuarial annuity tables published by the internal revenue service. ordinary single life annuities are used for single people and ordinary joint life and survivor annuities are used for married couples (internal revenue service, 1998, tables i and ii). 183y. yuh et al. / financial services review 7 (1998) 175–193 table 1 sample characteristics and wealth-needs ratio by characteristics (mean and pessimistic portfolio projections) variables % wealth-needs ratio (mean) wealth-needs ratio (pessimistic) total 100.0 131.7 103.8 education less than high school grad. 9.8 93.7*** 88.6*** high school graduate 29.8 120.9 100.3 some college 26.5 126.6 100.2 college or more 33.8 156.5 114.3 race/ethnicity white, non-hispanic 81.0 136.7*** 106.1*** black, non-hispanic 10.3 100.2 87.4 hispanic 4.2 100.8 84.8 other, non-hispanic 4.5 143.0 119.0 excellent health yes 36.7 143.9*** 110.2*** no 63.3 124.6 100.1 marital status couple 69.8 137.2*** 107.3*** unmarried male 9.7 134.6 98.4 unmarried female 20.5 111.8 94.5 occupation professional, managerial, specialty 32.1 158.1*** 117.2*** technical, sales, admin. support 25.2 132.6 103.1 service 8.3 104.7 93.9 precision production, craft, repair 12.8 122.8 97.8 operators, fabricators, laborers 20.1 107.2 93.4 farming, forestry, fishing 1.4 104.0 73.9 self-employed yes 7.0 193.3*** 110.2* no 93.0 127.1 103.4 household income $0 , income# $32,000 24.7 108.5*** 91.0*** 32,000, income# 45,000 25.8 114.4 95.6 45,000, income# 71,000 24.7 131.5 106.6 income. 71,000 24.7 173.2 122.5 ownership of db plan yes 36.1 146.7*** 124.7*** no 63.9 123.3 92.0 housing tenure own without mortgage 16.4 147.3*** 123.1*** own with mortgage 62.9 137.1 106.6 rent 20.7 102.9 80.0 planned retirement age 61 or earlier 34.6 114.1*** 90.1*** 62–65 55.4 132.3 105.7 66 or later 10.0 189.8 141.2 have retirement as a saving goal yes 35.1 157.5*** 120.2*** no 64.9 117.8 95.0 use of financial planner yes 24.7 145.0*** 111.0*** no 75.3 127.4 101.5 (continued on next page) 184 y. yuh et al. / financial services review 7 (1998) 175–193 3.4. data and sample the data analyzed in this study are from the public use tape of the 1995 survey of consumer finances (scf; kennickell et al., 1997). the scf is a triennial survey sponsored by the federal reserve with the cooperation of the department of the treasury. the 1995 scf was conducted by the national opinion research center (norc) at the university of chicago between july and december 1995. the purpose of the scf is to provide comprehensive, detailed information on the financial characteristics of u.s. households. a total of 4,299 families are interviewed in the 1995 scf. the 1995 scf has five complete data sets called “implicates” as a result of multiple imputation to handle missing data. this study uses repeated-imputation inference (rii) techniques to combine the five different data sets to make valid inferences ( montalto and sung, 1996; rubin, 1987). households are included in the sample if the householder is age 35 to 70, works full-time, and indicates the age at which s/he plans to stop full-time work. the age cutoffs are necessary since income and portfolio projections are used to examine retirement wealth adequacy. portfolio projections are simulated based on the household’s current portfolio and financial situation (yuh, 1998). households are excluded from the study if information on the age at which the householder plans to stop working full-time is not available. additionally, households are included only if they have positive non-investment income and total annual household income above the poverty threshold. a total of 1,387 households meet all of the criteria for inclusion. table 1(continued) variables % wealth-needs ratio (mean) wealth-needs ratio (pessimistic) stock share 0% 41.7 98.4*** 90.7*** 0% , stock, 13.5% 18.2 156.8 117.2 13.5 # stock, 36.5 20.3 142.9 109.1 stock$ 36.5% 19.9 167.2 113.8 spending$ income yes 51.1 91.8*** 74.7*** no 48.9 173.4 134.3 take high financial risk yes 20.7 153.6*** 114.0*** no 79.3 126.0 101.2 expect income growth yes 17.0 143.0*** 102.4 no 83.0 129.4 104.1 subjective life expectancy live # 24 years 24.4 128.2*** 111.1*** 24 , live # 32 24.3 129.7 100.3 32 , live # 42 25.8 141.7 106.3 live . 42 25.5 126.8 97.9 analysis of variance f-test for difference of means is statistically significant, * p# 0.05, *** p # 0.001 source: 1995 survey of consumer finances, combined data set, n5 6,310 (1,262 in each implicate) 185y. yuh et al. / financial services review 7 (1998) 175–193 3.5. household retirement wealth adequacy each household provides detailed information on assets that is used to estimate retirement wealth. future rates of return projected with the lognormal forecasting model are used to project future real accumulations separately for business assets (using the returns on small table 2 regression of wealth-needs ratio (%) on household characteristics, mean portfolio projections variable estimate std error p-value intercept 2208.1532 72.7955 0.0050** less than high school: reference high school graduate 8.0712 16.5438 0.6257 some college 211.5437 17.3004 0.5047 college or more 216.9703 18.9179 0.3702 white, non-hispanic: reference black, non-hispanic 29.1628 14.6630 0.5321 hispanic 24.5669 22.9794 0.8427 other, non-hispanic 35.5614 17.8932 0.0486* excellent health 6.5972 8.1427 0.4180 married couple: reference unmarried male 41.9206 13.3632 0.0018** unmarried female 13.7288 13.2616 0.3007 household size 20.8594 5.0021 0.8639 proportion of members, 18 25.6501 26.6756 0.3382 professional, managerial, specialty: reference technical, sales, admin. support 26.5285 11.0332 0.5547 service 211.2904 17.5343 0.5197 precision production, craft, repair 229.9816 15.6144 0.0562 operators, fabricators, laborers 230.9372 13.9746 0.0269* farming, forestry, fishing 231.5121 35.4799 0.3745 self employed 92.9234 11.2567 0.0000*** log of household income 35.8350 6.2803 0.0000*** db pension ownership 21.4864 8.1988 0.0088** rent 252.6381 14.1551 0.0002*** own with mortgage 243.2536 10.2051 0.0000*** own without mortgage: reference retire at 61 or earlier: reference retire at 62–65 27.8241 9.1217 0.0024** retire at 66 or up 85.3864 14.7443 0.0000*** retirement saving goal 2.7374 8.3980 0.7446 use of financial planner 0.8698 9.4058 0.9265 stock share 25.0523 49.0215 0.9181 investment horizon 20.4789 0.7873 0.5436 investment horizonp stock share 9.4966 2.5664 0.0004*** spending$ income 282.4638 8.7886 0.0000*** high risk taking 17.0423 8.7914 0.0526 expect income growth 11.1496 10.0444 0.2680 expected life expectancy 20.2688 0.3782 0.4778 f 5 21.9227, p-value5 0.0000 r-square5 0.3757 to 0.4017 combined data set, number of observations in each implicate5 1,262 * : p-value# 0.05, ** : p-value# 0.01, *** : p-value# 0.001 186 y. yuh et al. / financial services review 7 (1998) 175–193 capitalization stocks,) stocks and the stock components of mutual funds (using the returns on large stocks,) bonds (using the returns on corporate bonds), money market instruments (using the returns on treasury bills), and real estate assets (using the returns on composite real estate). total defined benefit pension wealth is estimated from the household’s self-reported information on expected benefits from defined benefit pension plans. the geometric mean of the nominal rate of return for long-term corporate bonds (ibbotson associates, 1995, pp. 38–39), 5.4%, is used as the discount rate for calculating defined benefit pension wealth. the 1995 scf does not provide direct identification of social security coverage. about 95% of jobs in the u.s. are covered by social security. the sample in this study consists of pre-retired households with at least one full-time worker, so all households are assumed to be covered by social security. the annual social security benefit is estimated using current social security replacement ratios based on current age, planned retirement age, current earnings, and marital status (social security administration, 1995). the replacement ratio represents the portion of preretirement salary that social security income will replace. the estimated annual social security benefit is adjusted for early retirement or delayed retirement as indicated by the planned retirement age. the present value of social security at the point of planned retirement is estimated using the real discount rate used by the social security administration (2.3%) in their long range projections (moore and mitchell, 1997). one limitation of this study is that income taxes on retirement income are not taken into account. no previous study has explicitly taken income taxes on retirement income into account, probably because of the complexity of the task. the treatment of pension and annuity income is complex, including the uncertain effect of roth iras over the next 20 or 30 years. given the income distribution of elderly households, it is likely that a majority face an average federal income tax rate under 10%. the potential bias from ignoring income taxes on retirement income may be high for higher income households and zero for low income and most moderate income households. 4. findings and discussion 4.1. descriptive statistics the dependent variable analyzed is the wealth-needs ratio expressed as a percentage: [projected retirement wealth / total retirement needs]*100 to reduce the amount of variance in the dependent variable, households with a wealthneeds ratio greater than 1,000 based on the mean case projection are dropped, resulting in a sample of 1,262 households. two different ratios are computed, based on the mean projection and the pessimistic projection. the median ratio is 102% for the mean projection and 87% for the pessimistic projection. for the mean projection, 25% of the households have a wealth-needs ratio of 68% or less, and for the pessimistic projection, 25% of the households have a ratio of 61% or less. about 52% of the households in the sample have adequate wealth for retirement at the 187y. yuh et al. / financial services review 7 (1998) 175–193 planned retirement age under the mean case projection. only 42% of the households in the sample have adequate wealth for retirement at the planned retirement age under the pessimistic case projection. sample characteristics are provided in the second column of table 1. only 36% of the households own defined benefit pension plans. about 40% of the households hold 13.5% or more of their non-housing assets in stock. over half (55%) of the householders plan to retire between age 62 and 65, and 35% of the households have retirement as a major saving goal. about half of the households (51%) indicate their spending is at least as high as income last year, and the majority do not expect future real income growth (83%). about one fourth of the householders (25%) expect to live an additional 42 years or more, and another fourth (24%) expect to live an additional 24 years or less. about 21% of the households are willing to take high financial risk to earn high returns. 4.2. analysis of variance results from analysis of the wealth-needs ratio by each variable are provided in the third column of table 1 for the mean case projection, and in the last column of table 1 for the pessimistic case projection. analysis of variance f-tests are used to identify the categories of independent variables with significant differences in mean wealth-needs ratios, not controlling for other factors. for the mean case projection, all of the independent variables are significantly related to the wealth-needs ratio at the 0.1% level or better. for the pessimistic case projection, all independent variables, with the exception of the expectation of real income growth, are significantly related to the mean wealth-needs ratio at the 5% level or better. the wealth-needs ratio is positively related to education, household income, planned retirement age, and the share of non-housing assets held in stocks, and varies with household spending behavior. the wealth-needs ratio (mean case projection) ranges from 94% for households with a householder who has not graduated from high school to 157% for households with a householder who is a college graduate. the ratio ranges from 109% for households with annual income of $32,000 or less, to 173% for households with annual income over $71,000. households with a householder who plans to retire at age 66 or later have a much higher mean wealth-needs ratio (190%) than those with a householder who plans to retire before age 62 (114%) or between age 62 and 65 (132%). households with a zero stock share have a mean wealth-needs ratio of 98%, compared to a ratio of 167% of households with a stock share or 36.5% or more. mean wealth-needs ratios are higher for households that spend less than income (173%) compared to those that do not (92%). 4.3. determinants of retirement wealth adequacy multivariate ordinary least squares (ols) regression analyses are performed to estimate the effect of each independent variable while simultaneously controlling for the effects of all other independent variables. the measure of retirement wealth adequacy used in the analyses is the wealth-needs ratio expressed as a percentage. separate regressions are performed for adequacy ratios based onmean portfolio performanceprojections of total retirement wealth (table 2) andpessimistic portfolio performanceprojections of total retirement wealth (table 188 y. yuh et al. / financial services review 7 (1998) 175–193 3). total retirement needs are estimated from a household expenditure function for both cases. since this dependent variable captures the amount of total retirement wealth in the households relative to their needs, it measures the extent of adequacy of retirement wealth in each household. table 3 regression of wealth-needs ratio (%) on household characteristics, pessimistic portfolio projections variable estimate std error p-value intercept 67.7287 44.0674 0.1304 less than high school: reference high school graduate 6.8667 9.3918 0.4648 some college 23.4516 9.7785 0.7241 college or more 25.2969 10.4124 0.6110 white, non-hispanic: reference black, non-hispanic 26.1461 8.3860 0.4637 hispanic 0.3468 13.3589 0.9793 other, non-hispanic 34.6908 10.7793 0.0021** excellent health 1.7044 4.6951 0.7168 married couple: reference unmarried male 19.3435 7.3878 0.0089** unmarried female 3.4093 7.4554 0.6475 household size 21.0024 2.6199 0.7021 proportion of members, 18 16.1367 14.5988 0.2701 professional, managerial, specialty: reference technical, sales, admin. support 0.6829 6.3033 0.9139 service 28.7984 10.1076 0.3843 precision production, craft, repair 213.6439 8.9079 0.1275 operators, fabricators, laborers 215.4189 7.8768 0.0504 farming, forestry, fishing 236.0423 20.2499 0.0754 self employed 20.3252 6.3783 0.0017** log of household income 6.6942 3.7562 0.0818 db pension ownership 27.2414 4.6763 0.0000*** rent 240.6923 7.9816 0.0000*** own with mortgage 228.1107 5.8227 0.0000*** own without mortgage: reference category retire at 61 or earlier: reference retire at 62–65 25.0765 5.1477 0.0000*** retire at 66 or up 77.3792 8.5035 0.0000*** retirement saving goal 2.5834 4.6160 0.5758 use of financial planner 1.8390 4.9622 0.7112 stock share 59.8913 46.4871 0.2088 investment horizon 20.9314 0.4250 0.0290* investment horizonp stock share 1.5370 2.6086 0.5605 spending$ income 256.7591 5.1399 0.0000*** high risk taking 12.4335 4.9803 0.0126* expect income growth 2.8483 5.7522 0.6211 expected life expectancy 20.0577 0.2134 0.7871 f 5 17.6192, p-value5 0.0000 r-square5 0.3373 to 0.3527 combined data set, number of observations in each implicate5 1,262 * : p-value# 0.05, ** : p-value# 0.01, *** : p-value# 0.001 189y. yuh et al. / financial services review 7 (1998) 175–193 4.4. discussion of regression results for the mean portfolio performance regression, race/ethnicity, marital status, occupation, self-employment, income, ownership of a defined benefit pension, housing tenure, planned retirement age, the interaction of stock share and investment horizon, and spending behavior are significantly related to the mean wealth-needs ratio at the 5% level or better (table 2). results for the pessimistic portfolio performance regression are similar with the exception that the investment horizon and high risk tolerance are significant, while occupation, income, and the interaction of stock share and investment horizon are not significant (table 3). although significant in the bivariate analysis, education, health status, having retirement as a saving goal, stock share, expectation of real income growth, and subjective life expectancy do not have statistically significant effects on the wealth-needs ratio when the other independent variables are controlled. the predicted wealth-needs ratio increases with the log of household income for the mean portfolio projection but not for the pessimistic projection. the increase in the predicted wealth-needs ratio is large for an increase from very low income to middle income (e.g., $10,000 to $40,000) but small for increases above $40,000. for the mean portfolio projection, the wealth-needs ratio is related to an interaction term for investment horizon and stock share, but not to stock share or investment horizon variables individually. the net effect of all three variables is that the predicted wealth-needs ratio increases with horizon for values of stock share over 5%. the predicted wealth-needs ratio increases with stock share for horizons of at least 1 year. for the pessimistic portfolio projection, the wealth-needs ratio is not significantly related to stock share by itself or to the interaction term for investment horizon and stock share, but it is related to the investment horizon variable. the net effect of all three variables is that the predicted wealth-needs ratio decreases with horizon for values of stock share under 60%. planned retirement age also has a large effect on the wealth-needs ratio. for the mean portfolio projection, those who plan to retire at age 66 or later have a predicted wealth-needs ratio 85 percentage points higher, and those who plan to retire between age 62 and 65 have a predicted wealth-needs ratio 28 percentage points higher than otherwise similar households who plan to retire before age 62. for the pessimistic portfolio projection those who plan to retire at age 66 or later have a predicted wealth-needs ratio 77 percentage points higher, and those who plan to retire between age 62 and 65 have predicted wealth-needs ratio 25 percentage points higher than otherwise similar households who plan to retire before age 62. spending as much as or more than income has a large effect on the wealth-needs ratio. for the mean portfolio projection, those who report spending at least as much as income have a predicted wealth-needs ratio 82 percentage points lower than otherwise similar households who spend less than income. for the pessimistic portfolio projection, those who report spending at least as much as income have a predicted wealth-needs ratio 57 percentage points lower than otherwise similar households who spend less than income. to provide some idea of the magnitude of the effects of planned retirement age and overspending behavior, predicted wealth-needs ratios at retirement are calculated based on the regression results in tables 2 and 3. a hypothetical household is defined to have mean values for continuous variables (median value for household income) and the most common 190 y. yuh et al. / financial services review 7 (1998) 175–193 values for dummy variables in the model. the predicted probabilities of having adequate wealth at retirement for this scenario are presented in table 4. the importance of spending less than income can be clearly seen—even those planning to retire before age 62 have a projected wealth-needs ratio at retirement of 135% if they spend less than their income, compared to 52% for comparable households that spend at least as much as income (mean portfolio projection). retiring at a later age has a large impact on the wealth-needs ratio. for households currently spending at least as much as income, those planning to retire before age 62 have a projected wealth-needs ratio at retirement of 52%, compared to a ratio of 80% for those retiring between ages 62 and 65, and a ratio of 138% for those planning to retire after age 65. 5. conclusions this study projects that almost half of u.s. households headed by a worker age 35 to 70 will not be able to maintain the current level of spending in retirement, even if investments achieve an average rate of return in the future. the proportion unable to maintain the level of spending increases to 58% with pessimistic investment projections. these estimates are based on current projections of social security pensions and ignore the effect of income taxes, so the situation could be worse than reported. planned retirement age and household spending behavior are important factors affecting the adequacy of retirement wealth. later retirement increases the number of years to accumulate retirement resources and decreases the number of years in retirement. in addition, retirement age is directly related to pension availability and the level of pension benefits. spending less than income implies saving, and thus increases the opportunity to save for table 4 predicted retirement wealth-needs ratio for a hypothetical scenario, for mean and pessimistic portfolio projections, by spending and retirement age retire at predicted retirement wealth-needs ratio mean projection pessimistic projection spend$ income spend, income spend$ income spend, income , 62 52.2 134.6 50.9 107.7 62–65 80.0 162.5 76.0 132.8 $ 66 137.6 220.0 128.3 185.0 predicted wealth-needs ratios at retirement were calculated based on the mean portfolio projection (table 2) and pessimistic portfolio projection (table 3). a hypothetical household is defined to have mean values for continuous variables (median value for normal income) and the most common categories for dummy variables in the model, except for health and having retirement as a savings goal. thus, the example household is assumed to have the following characteristics: white non-hispanic married couple, with one child, a college educated householder in excellent health employed in a professional occupation, annual household income of $45,000, no defined benefit pension plan, retirement portfolio with a 14% stock share, owns a house with a mortgage, 16 years away from retirement, not expecting future income growth, does not use a financial planner, does not expect income growth, not a risk taker, has retirement as a saving goal, and expects to live 34 years. 191y. yuh et al. / financial services review 7 (1998) 175–193 retirement. overspending decreases the wealth accumulations for retirement and increases estimated retirement consumption needs. based on mean projections, the interaction between the stock share and the investment horizon (number of years until retirement) is an important factor affecting retirement wealth adequacy. a higher stock share with the same investment horizon or a longer investment horizon with the same stock share significantly increase the adequacy of retirement wealth. the lack of a significant negative effect of stock share on the wealth-needs ratio implies that increasing the stock share will not impose a risk for households in general, even if projected future real returns for investments are at the levels of the lowest 5% for time periods in the past. (households with little diversification may be at risk, but low diversification could not be measured accurately in the dataset.) aggressive investment or saving strategies should be encouraged especially for individuals who have longer investment horizons. moreover, asset allocation decisions within retirement saving programs are important for individual investors given the increase in 401(k) and related retirement saving programs and the decrease in defined benefit plans since the 1980s. evidence of higher rates of return for stocks in the long run should be used to encourage stock investment within retirement savings programs. clearly though, a simple first step to an adequate retirement is getting spending under control. the fact that education was not significant in the regressions suggests that there is not an inherent barrier to teaching workers about saving for retirement. many of the variables with substantial significant effects in the regressions are factors that households can control, as illustrated in table 4 for spending and retirement age. references andrews, e. s. (1993). gaps in retirement income adequacy. in r. schmitt & pension research council (eds.), future of pensions in the united states. (pp.1–52). philadelphia: university of pennsylvania press. bajtelsmit, v. j., & van derhei, j. a. (1996). risk aversion and retirement income adequacy. in o. s. mitchell (ed.), positioning pensions for the twenty-first century. philadelphia: university of pennsylvania press. bernheim, b. d. (1996).the merrill lynch baby boom retirement index: update ’96. stanford university: merrill lynch. bernheim, b. d., skinner, j. & weinberg (1997, september). what accounts for variation in retirement wealth among u.s. households? mimeo, stanford university. burns, s. a., & widdows, r. (1988). an estimation of savings needs to adequately fund baby boomers’ retirement. in v. hampton (ed.),proceedings of the 34th annual conference of the american council on consumer interests(pp. 15–8). columbia, mo: american council on consumer interests. burns, s. a. & widdows, r. (1990). sensitivity of a retirement analysis framework to changes in retirement analysis parameters.financial counseling and planning 1, 71–91. crow, e. l., & shimizu, k. (1988).lognormal distribution. new york: marcel dekker. duncan, g. j., mitchell, o. s., & morgan, j. n. (1984). a framework for setting retirement savings goals.journal of consumer affairs 18(1), 22–46. hatcher, c. b. (1997). a model of desired wealth at retirement.financial counseling and planning 8(1), 57–64. ibbotson associates. (1995).stocks, bonds, bills, and inflation yearbook. chicago, il: ibbotson associates. ibbotson, r. g., & siegel, l. b. (1984). real estate returns: a comparison with other investments.areuea journal 12(3), 219–242. internal revenue service. (1998).how to use actuarial tables[online]. available: http://www.irs.ustreas.gov/ prod/forms2pubs/graphics. 192 y. yuh et al. / financial services review 7 (1998) 175–193 kennickell, a. b., starr-mccluer, m., & sunde´n, a. e. (1997). family finances in the u.s.: recent evidence from the survey of consumer finances.federal reserve bulletin 83(1), 1–24. li, j., montalto, c. p., & geistfeld, l. v. (1996). determinants of financial adequacy for retirement.financial counseling and planning 7, 39–48. mitchell, o. s., & moore, j. f. (1997).retirement wealth accumulation and decumulation: new developments and outstanding opportunities. wharton financial institutions center working paper 97-12, the wharton school of the university of pennsylvania, philadelphia. modigliani, f., & brumberg, r. (1954). utility analysis and the consumption function: an interpretation of cross-section data. in k. kurihara (ed.),post keynesian economics(pp. 388–446). new brunswick, nj: rutgers university press. montalto, c. p., & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances.financial counseling and planning 7, 133–146. moore, j. f., & mitchell, o. s. (1997).projected retirement wealth and savings adequacy in the health and retirement study. pension research council working paper 98-1. the wharton school of the university of pennsylvania, philadelphia. palmer, b. a. (1989). tax reform and retirement income replacement ratios.the journal of risk and insurance 56, 702–725. palmer, b. a. (1994). retirement income replacement ratios: an update.benefits quarterly 10(2), 59–75. rubin, d. b. (1987).multiple imputation for nonresponse in surveys. new york: wiley. schulz, j. h. (1992).the economics of aging. westport, ct: auburn house. social security administration. (1995).annual statistical supplement, 1995 to the social security bulletin. washington, dc: us department of health and human services, social security administration. u.s. bureau of labor statistics. (1996).consumer expenditure survey: 1994 interview survey cd rom/public use tape documentation. washington, dc: u.s. bureau of labor statistics. u.s. bureau of the census. (1998). population profile of the united states 1997.current population reports, p23-194. washington, dc: u.s. government printing office. u.s. department of commerce. (1998, november 3). personal income and outlays: september 1998, bea 98-34. washington, dc: author. u.s. general accounting office. (1996, october 3). private pensions: most employers offer pensions that use defined contribution plans, ggd-97-1. washington, dc: author. yuh, y. (1998). adequacy of preparation for retirement: mean and pessimistic case projections. unpublished doctoral dissertation. the ohio state university, columbus, oh. yuh, y., montalto, c. p., & hanna, s. (1998). are americans prepared for retirement?financial counseling and planning 9(1), 1–12. 193y. yuh et al. / financial services review 7 (1998) 175–193 pii: s1057-0810(00)00065-2 liquidating a remainder interest: simplifying personal finance john c. bost, tony cherin* san diego state university, san diego 92182-8236, usa received 11 may 1999; received in revised form 10 march 2000; accepted 26 june 2000 abstract there are many ways in which decedents leave property in trust for their heirs. one technique is to grant a life estate to surviving children. the purpose of this paper is to describe verbally, and through example, an approach to liquidating a life estate. this simplification in personal finance involves a “buyout” of the interests of the remaindermen. the result is dissolution of the trust, leaving the income beneficiaries to manage, as owner in fee, the remaining assets as they wish, without the expense and complexity associated with maintaining a trust. © 2000 elsevier science inc. all rights reserved. jel classifications:g290; g120; d46 keywords:personal finance; remainder interests; liquidation; valuation; remaindermen 1. introduction there are many ways in which decedents leave property in trust for their heirs. one technique is to grant a life estate to surviving children that allows them to reap the income thrown-off from trust assets while, for all intents and purposes, leaving the corpus intact for later distribution to those with a remainder interest. the latter are usually the settler’s (grantor’s) grandchildren, but, in some instances, the settlor may choose to name charitable/ tax-exempt organizations (for example, educational institutions) as the remaindermen. while * corresponding author. tel.:11-619-594-5657; fax:11-619-594-1573. e-mail address: tcherin@mail.sdsu.edu (t. cherin). financial services review 9 (2000) 183–195 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00065-2 such perfectly acceptable estate planning methods benefit younger generations and, eventually, a favored charity, they do involve complexity and expense. among such complications and costs one would include appointing and compensating trustees. moreover, in their fiduciary capacity, these trustees must hew to the investing wishes of the grantor of the trust, which may not be in the best interest of the income beneficiaries. furthermore, fiduciary tax returns must be prepared and submitted annually. additionally, trust income tax brackets are much more compressed than the income tax brackets for individuals and, as a result, the highest rates are quickly reached unless most of the income is distributed to the income beneficiaries. finally, in some instances, there may be a directive to hold the trust’s financial assets with yet another party, such as an investment firm, incurring even greater expense. the purpose of this paper is to describe verbally, and through example, an approach to liquidating a life estate in situations where the income beneficiaries are capable of managing the corpus of the trust and to demonstrate that this can be done to the mutual benefit of both the income beneficiaries and the remaindermen. this simplification in personal finance involves a “buyout” of the interests of the remaindermen using but a portion of the assets of the trust as payment. to put it formally, the income beneficiaries purchase all of the remainder interests in the trust. the result is dissolution of the trust, leaving the income beneficiaries to manage, as owner in fee, the remaining assets as they wish, without the expense and complexity associated with a trust. 2. a brief review of some relevant literature while the body of literature relating to the specific issue discussed in this article is sparse, there are tangential discussions on liquidating a remainder interest and charitable remainder trusts. in his article on charitable remainder trusts (crt), fooden (1996) notes that a crt “. . . provides a mechanism for the transfer of property from one generation to the next without incurring any gift or estate tax liability.” moreover, the author observes that it is the only method for transferring the full value of “highly appreciated assets” to heirs. in the same vein, moyers, spiegel and baum (1997) state that in order to receive a higher tax benefit, the tangible appreciated person property (tapp) crt procedure should be taken into account “. . . when a client has valuable, low-basis tangible personal property that cannot be liquidated without incurring substantial capital gains.” at the same time, each situation is unique regarding a remainder interest or trust and should be individually evaluated to find the option providing the greatest financial benefit and smallest tax consequence. siegel and swerdlin (1996) consider the situation where the interests of the income and charitable beneficiaries of a trust do not overlap. the distinctive tax structure of such “split-interest trusts” allows a fiduciary to ally the wishes of the parties by investing for capital appreciation instead of income. this technique results in higher after-tax income to the donor (income beneficiary) and an anticipated gain to the charity. in his article on selling a remainder interest, croman (1994) states that a sale of a remainder interest could also reduce the estate taxes to be paid while still carrying out the 184 j.c. bost, t. cherin / financial services review 9 (2000) 183–195 wishes of the decedent. the sale, which involves the payment by the buyer of the full value of the remainder interest, either in cash, exchange of property, a promissory note or an annuity, is recommended as a possible estate planning process. of course, the irs has created regulations and recommendations to be followed regarding the liquidation of remainder interests and trusts. herman-giddens (1998) discusses the proposed regulations issued in 1997 by the irs. they include matters related to “. . . flip unitrusts, the time for paying the annuity or unitrust amount, appraising unmarketable assets, . . . special valuation rules for crts, and prohibiting the allocation of precontribution capital gain to trust income.” many of the regulations are aimed at reducing abuses to the system; however, they will not seriously limit or impede planning by a donor who has a “genuine charitable intent.” (since implemented, these regulations are cited undertreasury regulations andunited states codesin the references.) while it would be impossible to affix an exact figure to the number of instances in which the “buyout” option should be considered, it can be said that anytime there are remaindermen other than family, typically charities and/or educational institutions, named in the trust this alternative may very well be appropriate. finally, although other contributions in this arena have addressed the possibility of valuing remainder interests, there has been no consideration of actually carrying out the process in practice and then convincing the remaindermen to accept a negotiated amount. this work meets both goals, dealing with many of the potential idiosyncrasies and difficulties attendant to the procedure. 3. estate planning goal the goal is to greatly simplify a complicated financial arrangement by changing it from an expensive, complex set of split interests (a subset of present interests and another subset of future interests in a long-term trust) into outright ownership. to accomplish the goal, it is necessary to convince the remaindermen that it is in everyone’s best interest to liquidate the trust. 4. general method 4.1. entering negotiations with the remaindermen there are several steps that must be carried out in implementing this strategy of liquidating a trust. without a doubt, the most crucial of these is the preparation and delivery of a thorough and easily understood explanation of the proposal to the remaindermen. it goes without saying that the efficiency of this process is inversely related to the number of the remaindermen and directly correlated to their level of financial sophistication. institutional remaindermen can generally be counted on to have a current understanding of the technique and to readily assimilate the nuances of a particular situation. in contrast, individual remaindermen may find the approach difficult to comprehend initially, although most will 185j.c. bost, t. cherin / financial services review 9 (2000) 183–195 quickly appreciate that receiving something of value now could be worth as much or more than receiving a greater, but perhaps more uncertain, amount in the future. individual remaindermen must also factor in their own mortality, something that charitable institutions can ignore. how does one provide and convey an understandable interpretation of the “buyout” option? most likely it will be initiated through the mail, followed by numerous telephone conversations with the remaindermen and/or their representatives. the content of the initial communications with the remaindermen is likely to be most productive if it includes: y how many other remaindermen there are and a promise to eventually reveal (assuming negotiations progress) whom the others are. y the nature and the present value of the trust’s assets. y what proportional share of the trust’s assets each remainderman is to receive. for example the letter might state: you are one of four remaindermen, each with an equal share. the value of the trust is approximately $1,400,000. thus each remainderman’s interest would be $350,000 (25%), if the income beneficiaries died today. y specific information about the income beneficiaries’ interests and general information about the income beneficiaries themselves. y a disclosure of any circumstances under which the income beneficiaries can spend trust principal (i.e., corpus). y the grounds for opening negotiations and the arguments for accepting a reasonable “buyout” offer, rather than waiting for the trust to run its course. for example: the trust allows distribution of principal to the income beneficiaries to provide for entry into a profession or business. “a” is in his early fifties and “b” is in her late forties. the temptation to draw out the corpus of the trust to start a business is quite strong. there is nothing in the trust document itself that would prevent them from using trust funds to purchase a business, manage it for a reasonable period of time, and then sell it. there is no requirement that proceeds from the sale of the business be put back into the trust. early in the negotiations, the income beneficiaries’ representative should include a “sanitized” copy of the trust document, that is one that blacks out the identity of the settlors, the trustee, and the individual beneficiaries. this will allow the remaindermen to assess the terms of the trust, especially those that convey special privileges with regard to withdrawing corpus or those that allow the trustee to make or hold investments that tend to favor the income beneficiaries over the remaindermen. 4.2. factors that determine present value the entire “buyout” strategy is based upon what is a relatively straightforward concept for personal finance professionals the time value of money or, more precisely, the present value of a dollar. the case is presented to the remaindermen in the form of two mutually exclusive options; one that can be exercised immediately and another that can be exercised only upon 186 j.c. bost, t. cherin / financial services review 9 (2000) 183–195 the death of the trust’s income beneficiary (or a successor income beneficiary, generally a spouse). in other words, each remainderman chooses between taking the present value of the future worth of its proportional share of the trust assets now or waiting for the demise of the income beneficiary and, in the case of a trust term defined by the joint lives of a married couple, a surviving spouse. both options, admittedly, come with “strings attached.” in the case of the former, at least two major variables are open to negotiations between the income beneficiaries and the remaindermen. one bone of contention concerns the future value of the remainder interests. a realistic valuation of the remainder interests rests upon predictions of asset growth rates, asset weights within the portfolio, and the projected life expectancy of the income beneficiaries. this, of course, results in a discussion of how to determine theratesof growth of the various trust portfolio assets and, concomitantly, theterm of growth of these assets. 4.3. life expectancy thetermof growth means that the life expectancy of the income beneficiary or, in the case of a surviving spouse, the joint life expectancy of both must be determined. treasury tables are generally used for this purpose, but, because they are based on averages, they do not necessarily represent the negotiators’ points of view. in brief, in any given instance, the table-generated life expectancy may be, from a negotiator’s standpoint, either too long or too short. a remainderman would always argue for a shorter life expectancy while an income beneficiary would vie for a longer one. obviously, a 50-year-old income beneficiary who is jogging or bicycling to work everyday, doesn’t smoke or hang-glide, and has parents who lived into their nineties, has a strong argument that the tables greatly understate life expectancy and, thus, the life estate period. as a consequence, the present value of the remainder interest is greatly overstated (to the ultimate benefit of the remaindermen). if, on the other hand, one is representing a chain smoking “couch potato” whose parents died in their mid-fifties and who considers operating the tv remote to be regular exercise, one would not raise the issue of individual variation in longevity, but would rather keep this part of the negotiations focused on the tables. assuming one chooses to use the federal table rate for valuing the remainder interest to determine the present value of each remainderman’s proportional share, the following wording might be an appropriate pattern for structuring the letter: the july 1997 federal rate for valuing remainder interests is 8%. jack is 55 years old and his wife, sally, is 50 years old. the remainder factor using the federal rate tables found in alpha volume, table r (2)-part 4, for two lives (55, 50) is 0.10516. this factor times 50% (your client’s remainder interest) of the july 1997 fair market value of the trust ($2,000,000) is $105,160 (that is, 50% * $2,000,000 * 0.10516). this is the amount that my client is willing to pay in exchange for your client’s release (or assignment) of its remainder interest. this offer is contingent upon the other 50% remainderman also accepting a similar offer. again, reiterating that the tables are remiss in accounting for a particular individual’s health status and, thus, over or understate the value of the remainder interest, it is also important to note that they do not take into consideration special provisions in a particular trust that change the life estate from a straight life estate (or an annuity for life) to something else. 187j.c. bost, t. cherin / financial services review 9 (2000) 183–195 4.4. distribution of principal it is these specific provisions in any given trust that may persuade the remaindermen to accept a “buyout” proposal even though, initially, they are inclined to reject it. such provisions include, but are not limited to, the distribution of principal to the income beneficiary pursuant to an “ascertainable standard” power of appointment or some other, even less restrictive, provision for example, the trust may make corpus available if the income beneficiary wants to start a business. under an “ascertainable standard” power, the terms of the trust give the holder of the power the right to invade the principal for reasons of “health, education, support, or maintenance.” the power to invade principal under an “ascertainable standard” is found in many irrevocable trusts because the internal revenue code (irc) endorses these prerogatives as “limited powers” rather than “general powers” (irc §2041[b][1][a]). this means that, even though the holder of such a power (for example, the income beneficiary) can benefit from its exercise, when the holder dies the trust property is not included in the estate. it can be readily seen that the remaindermen, in choosing between the two mutually exclusive options, are faced with the time-honored conundrum of whether “a bird in the hand is worth two in the bush.” indeed, the more opportunities that exist for the income beneficiary to invade corpus, the greater the likelihood that, when the time comes, all birds will have flown and just the bush will remain. under such circumstances, a convincing case can be made to the remaindermen for taking “a bird in the hand.” they will be swayed in favor of the “buyout” option if they believe that the income beneficiaries will exercise their power to provisionally invade the corpus of the trust and/or that the costs associated with running the trust will, over time, greatly diminish its value. 4.5. trust powers naturally, where there is a discretionary power to distribute trust corpus to the income beneficiaries, the extent to which the remaindermen’s interest is decreased depends upon who holds this power, the degree to which the amount that can be distributed is limited either in dollar amount or by circumstance, and the probability that the discretionary power to distribute will be exercised. the authority to make distributions is usually in the form of a power of appointment. the term “power of appointment” refers to the ability of a person (the “holder” of the power) to transfer (“appoint”) title to property, even though this “holder” does not necessarily own it, from the owner (generally a trustee) to someone else. if a power is a general power, it will be included in the power holder’s estate when he/she dies. by irc definition, a general power is “a power which is exercisable in favor of the decedent [power holder], his estate, his creditors, or the creditors of his estate. . . ” (irc §2041[b][1]). all other powers are deemed to be limited powers that are not included in the holder’s estate. because of special irc provisions, a power limited by an “ascertainable standard” and another power, called the “5 & 5” power, are two of the most common “powers” found in trusts. a power limited by an “ascertainable standard” is, in the language of the irc, a power that can be exercised by the holder-beneficiary only for the purpose of benefiting his/her “health, 188 j.c. bost, t. cherin / financial services review 9 (2000) 183–195 education, support, or maintenance” (irc §2041[b][1][a]). this limited power can have a definite impact on the remainder interest and, thus, can be used as a bargaining chip in selling the “buyout” option to the remaindermen under certain circumstances. for instance, if the trust beneficiary has only modest income and/or the trust assets do not throw off significant cash flow for use by the beneficiary, it is much more likely that a power limited by an “ascertainable standard” will be exercised and corpus will be invaded, diminishing the amount ultimately to be received by the remaindermen. perhaps of greater consequence to the remaindermen (and the value they will eventually receive) is the “5 & 5” power which can be classified as either a general or limited power depending on who holds it. this power allows the power holder to distribute up to 5% or $5,000 (whichever is greater) of the trust corpus annually, on a noncumulative basis with no justification necessary. the irc (§2041[b][2]) allows the lapse (during the holder’s lifetime) of a general power to be treated as if no gift occurred provided the power that lapsed was limited to no more than the greater of $5,000 or 5% of the value of the trust. obviously, the influence of such a power, if it exists, in promoting the “buyout” option is huge. table 1 above is an attempt to assign numerical “negotiating weights” to the existence of each of these two powers in a trust. the higher the number in the last column (on a scale of 1–10), the greater the expected decrease in the remainder value and, consequently, the greater its negotiating weight. the table indicates that the lowest negotiating weight goes to an “ascertainable standard” with an independent trustee and a wealthy beneficiary because it is highly unlikely that the power will be exercised. why so? if the trustee exercises it in favor of a beneficiary who is not really in need, the trustee opens itself to being sued by the remaindermen. compare this situation to the “5 & 5” power held by the beneficiary. it receives a 10 because it can be safely assumed that the beneficiary will definitely exercise such a power at every opportunity even if he or she is very wealthy. 5. trust provisions that can be quantified some provisions in a trust, such as a 5 & 5 powerheld by an income beneficiary, can be valued because it is correct to assume that the beneficiary will exercise this right of table 1 negotiating weights provision power holder beneficiary is: relative decrease in remainder value negotiating weight ascertainable standard trustee wealthy very slight 2 ascertainable standard trustee poor large 8 ascertainable standard beneficiary wealthy slight 4 ascertainable standard beneficiary poor large 8 5 & 5 power trustee wealthy slight 4 5 & 5 power trustee poor large 7 5 & 5 power beneficiary wealthy very large 10 5 & 5 power beneficiary poor very large 10 189j.c. bost, t. cherin / financial services review 9 (2000) 183–195 withdrawal each and every year, and at the earliest opportunity. likewise, one can assume that withdrawal will occur if a provision permitting a specific sum to be withdrawn for a particular purpose (for example, up to $100,000 for a down payment on a home) exists. even if the withdrawal must be delayed, its present value can be determined. other provisions, such as trustee fees and the cost of accountants to prepare trust income tax returns, can generally be quantified with some degree of accuracy. the trustee fees are usually some small percentage of the trust’s value (for example, 1% of year-end corpus value) and the cost of tax return preparation is an annually occurring expense that can easily be valued. 6. trust provisions that cannot be quantified and their effect on valuation provisions that cannot be precisely valued are those that allow considerable discretion as to the amount and timing of distributions of corpus. included in this category are powers held by an independent trustee where the beneficiary’s needs must be taken into account. such powers might be those limited by an ascertainable standard. for example, if the income beneficiary is in need of extra money immediately, will this be true in the future? or, if wealthy now, will he or she be able to maintain that level of wealth? obviously, the beneficiary’s immediate circumstances will be given some weight, but how much is open to argument. needless to say, the poorer the income beneficiary, the stronger the case for distributions of corpus and the lower the value of the remainder interest. naturally, the reverse is true if the income beneficiary is wealthy or if the trust itself is so large that the income throw-off by itself will make the beneficiary affluent. other provisions whose impact on the remainder interest cannot be precisely measured include: circumstances relating to the income beneficiary such as y his or her health y whether he or she has medical insurance y his or her lifestyle (for example, is he or she a smoker or nonsmoker, a nondrinker, a social drinker, or a problem drinker?) matters relating to the trust such as y the relationship between the trustee and the income beneficiary y whether there is a provision specifying that the income beneficiary’s interest takes preference over that of the remaindermen y the restrictions in the trust regarding the types of investments that can be made by the trustee (state laws generally favor diversified portfolios for trusts. the uniform probate investor act adopted by many states makes trust asset diversification the general rule, e.g., ca probate code §16048: “in making and implementing investment decisions, the trustee has a duty to diversify the investments of the trust unless, under the circumstances, it is prudent not to do so.” however a restrictive clause in a trust 190 j.c. bost, t. cherin / financial services review 9 (2000) 183–195 would have to be followed unless the trustee obtained court approval to invest differently.) y the extent to which the trust incurs transaction costs, including commissions, capital gains taxes, and investment advisory fees the circumstances of the remaindermen such as y the possibility of adverse publicity if a “deal”is struck or being deemed heartless if one is not struck y one or more of the remaindermen being a minor (here, court approval of a “buyout” may be required. and a court may require the appointment of an attorney to represent the “best interest” of the minor. his or her understanding of present value concepts and the ensuing benefits of such an arrangement may be lacking.) 7. valuation: a complex example based on an actual case the following hypothetical example is based on an actual case. the successfully concluded negotiations took place several years ago. this example is used to present the numbers and to suggest language for a written approach to the remaindermen with likely responses and appropriate counter responses included. the settlor (grantor) established a living trust, retaining all interests until his death. upon his death, the trust became irrevocable. the settlor’s 54-year-old son was given a life estate in the trust, with his 48-year-old spouse (the settlor’s daughter-in-law) receiving a following life estate in the event she outlived him. the trust did not condition her contingent income interest on her being still married to the son at his death (that is, she would have the income interest simply by outliving him). two highly regarded universities, a charity involved in the law, and an adult grandson were named as the remaindermen, each to receive a 25% share. when the settlor died the trust corpus was worth approximately $1,390,000 after all death taxes had been paid. the trust terms allowed the trustee to invade corpus for the benefit of the income beneficiary based upon an ascertainable standard related to “health, education, support, or maintenance.” another provision stated that the interests of the income beneficiaries were to be considered before those of the remaindermen. and, yet another, allowed the trustee to distribute corpus so as to allow the income beneficiaries to start a business. these provisions that favored the income beneficiaries, and the fact that the beneficiaries were given all income rather than a fixed dollar or fixed percentage amount, kept the trust from qualifying for a charitable estate tax deduction (see irc §2055). in addition to this favorable treatment, both income beneficiaries were in very good health. in implementing the remaindermen “buyout” offer, several steps were crucial. first, a determination of the appropriate growth rates was made for the different components of the initial trust portfolio. historical rates of capital appreciation were gathered fromstocks, bonds, bills, and inflation 1995 yearbook(ibbotson associates, 1995, chicago, il). second, a weighted average was used to determine the growth of the portfolio as a whole. table 2 portrays the expected portfolio growth. third, this expected growth rate was projected out for a 38-year joint life expectancy 191j.c. bost, t. cherin / financial services review 9 (2000) 183–195 starting with the portfolio value of the trust as of march 31, 1995. this projection took into account the capital gains tax (combined federal and state of 29%, which one might want to reduce to 22%, or even less, in light of the 1997 tax act’s capital gains rate reduction) on a 20% annual turnover of the portfolio, transaction costs of 0.5%, and annual trustee’s fees charged to corpus of 0.5% of the value of the corpus. fourth, the present value of the trust at the end of the 38 years was calculated using 9.2% discount rate, which was the average irc section 7520 rate for the months just preceding the negotiations. based on the example data just presented, the present value for the trust remainder interest was $83,177. therefore, each 25% share was worth $20,794, which was rounded up to $21,000 as an initial offer. it is worth mentioning that copies of schedules showing the growth of trust assets (taking into account taxes, transaction costs, and trustee fees) were attached to the letters to the remaindermen. in addition, the letters set forth the calculation of the valuation of the income and the remainder interest, as well as the assumptions underlying those calculations. moreover, it should be recognized that the determination of an appropriate discount rate might be a sticking point. while irc section 7520 directs the treasury to provide rates for valuing income and remainder interests on a monthly basis, income beneficiaries would naturally argue for a higher rate while the remaindermen would advocate a lower one. in this particular case both parties used the 9.2% discount rate for their calculations. fifth, an argument was then made that, given the power to withdraw corpus to start a business and the excellent health of the income beneficiaries, the use of the treasury table factors greatly overstated the remainder values. (at the same time, the income beneficiaries fully realized that making an offer that was ridiculously low would create an environment in which it would not be worth the remaindermen’s trouble to negotiate a settlement). furthermore, it was pointed out that some of the assumptions regarding costs were conservative, resulting in a higher value of the remainder interest than actual practice might justify. the assumption of a 20% turnover per year in the portfolio is very conservative as studies of mutual funds show that turnover rates generally exceed 50% per year, and increased turnover rates accelerate the recognition of capital gains and increase transaction costs, both of which act as a break on capital growth. finally, since this was a private inter vivos trust, the income beneficiaries chose not to disclose to the charitable remaindermen each other’s identity until all four remaindermen table 2 expected portfolio growth investment % of corpus expected cap. appr. per year weighted cap. appr. bonds 50% 0.0% 0.00% stocks, large cap 30% 5.4% 1.62% stocks, small cap 10% 12.5% 1.25% foreign stocks (22 year history) 10% 9.7% 0.97% expected ann. portfolio appr. 3.84% 192 j.c. bost, t. cherin / financial services review 9 (2000) 183–195 were in agreement, at least in principle, to negotiate a transfer of (sell) their remainder interests. the wording in the letter to the remaindermen was along these lines: this letter is to open discussion; it is not an offer that can be immediately accepted since my clients wish to retain the option to go forward only if settlement can be reached with all four remaindermen. from our initial conversations, it looks like this can be accomplished. once all four of you have indicated an interest in selling your remainder interest, you will be given the names of the other remaindermen. it is my clients’ intent that each remainderman receive the same payment, one that represents a realistic payment of the value of the interest that you each give up. we believe the offer is exceedingly fair for a number of reasons: the tables greatly overstate the remainder value, given the ease with which the income beneficiaries can remove the corpus from the trust; since the trust terms express a preference for the income beneficiaries, the trustee has the option of increasing income at the expense of growth; the joint life tables understate my clients’ joint life expectancy given that neither of them smoke, both exercise, both enjoy good health and both have excellent medical benefit packages through their employment; and the assumption of a 20% turnover per year in the portfolio is very conservative as studies of mutual funds show that turnover rates generally exceed 50% per year, and increased turnover rates accelerate the recognition of capital gains and increase transaction costs, both of which act as a break on capital growth. once all the remaindermen indicated their willingness to sell and the asking prices were within a range acceptable to the income beneficiaries, the charitable remaindermen were put in touch with one another. the income beneficiaries’ son (the settlor’s grandson who was appropriately represented by independent counsel) agreed early in the negotiations to take whatever share the three charitable remaindermen took. the charities’ initial counter offers (in response to the $21,000 “buyout” offer) were respectively: $29,000, $30,000, and $41,000. just the latter seemed out of line on the high side. interestingly, the charitable remaindermen actually ignored the income beneficiaries’ power to withdraw corpus to start a business and its depressing effect on the value of their remainder interests. apparently, this would have dropped the remainder value so low that further negotiating would have been pointless and the “wait and see” option would have been adopted instead. consequently, the income beneficiaries decided not to push this issue and its associated lowering effect on the remainder value. after conferring with one another, the charities offered $30,000 per interest and accepted a counter offer of $25,500 per interest. the trust was liquidated in august of 1995. at the time of its liquidation the trust’s value had risen to approximately $1,500,000. hence, after the payment of $102,000 to the remaindermen (and compensating the trustee and the attorney), the income beneficiaries received, free of trust, an amount just slightly in excess of $1,390,000. not only were the income beneficiaries happy with the result, the director of planned giving at one of the charities wrote a thank you letter stating: on behalf of all of us at [the university], i want to thank you, [the income beneficiaries], and [the trustee] for your help in making the [settlors] trust gift to [the university] possible. in the many years that i have been involved at [the university], this gift was unique. needless to say, we are pleased and grateful to [the settlors] for their thoughtful vision and interest. 193j.c. bost, t. cherin / financial services review 9 (2000) 183–195 please will you convey our gratitude to your clients. . . . again, i thank you for your assistance in this matter and send best regards from [the university]. clearly, he understood the present value concept. 8. situations where liquidation is inappropriate quite obviously, there are instances in which liquidating a trust would be an unfortunate strategy. for example, where the income beneficiaries are incapable of wisely handling an outright inheritance, tying up the property in a trust may be in the best interests of all persons concerned. even if the settlor’s children are quite capable of handling the property, the settlor may still prefer to restrict their interest to that of a life estate. (this may be particularly true if there are no grandchildren and none are likely.) in short, the settlor may like the idea that heor she,and not the children, has purposefully chosen the ultimate owners of the property (the remaindermen). this desire might be particularly strong if the remaindermen are the settlor’s favorite charities. of course, liquidation of the trust would be a moot point if the settlor placed a clause in it that prevented the remaindermen from selling their interests. (for example, “any attempt by a remainderman to sell its interest shall cause that remainderman’s interest to terminate, and said interest shall instead go to charity ‘x.’ ”) the income beneficiaries might be well advised not to pursue a “buyout” strategy if they are having personal financial problems and/or are faced with possible lawsuits. a life estate, especially if the trust contains a “spendthrift provision,” is less attractive to creditors than property owned outright by debtor. a remainderman may be quite loathe to sell its remainder interest if “bad press” might result from the disclosure that a one million dollar gift in the future was “traded away” for a mere fraction of said value today. of course, this gross misinterpretation is a realistic concern due only to the general public’s lack of understanding of the time value of money and/or public relations difficulties that some charities have had in the recent past. unfortunately, charities may be concerned that potential donors may hear about such settlements and, fearing that their “generous” gifts may eventually be whittled down to something that seems quite paltry, decide to give it to some other charity. after all, with a future $2,000,000 gift a library might well bear the donor’s name, whereas today’s $100,000 gift might result in the name appearing on a long list of donors on some recognition plaque in the reference section. 9. conclusion this paper has attempted to describe verbally, and through example, an approach to liquidating a life estate in situations where the income beneficiaries are competent to handle the corpus of the trust and to establish that this can be done to the shared benefit of both the income beneficiaries and the remaindermen. this simplification in personal finance involves employing a portion of the trust assets to “buyout” the interests of the remaindermen. the 194 j.c. bost, t. cherin / financial services review 9 (2000) 183–195 consequence is a termination of the trust, leaving the income beneficiaries to manage the remaining assets, without the costs and impediments associated with a trust, and the remaindermen with immediate cash to spend or invest as they see fit. references california probate code, division 9, part 4, chapter 1, article 2.5, section 16048. croman, earl l. “remainder interest sales may again be viable,”taxation for accountants, 1994,v52. (6), 332–338. fooden, bart l. “charitable remainder trusts,”cpa journal, 1996,v64. (9, sep), 44–49. herman-giddens, gregory. “revise tax planning for charitable remainder trusts,”taxation for accountants, 1998,v60. (5, may), 260–265. ibbotson associates. (1995).stocks, bonds, bills, and inflation 1995 yearbook.chicago, illinois. internal revenue code. (1999). irc §2041[b][1][a] internal revenue code. (1999). irc (§2041[b][2]) moyers, michael k., alan d. spiegel and e. richard baum. “charitable remainder trusts offer noncharitable benefits,”taxation for accountants, 1997,v58. (5, may), 285–291. siegel, laurence b. and eric i swerdlin. “using a growth strategy for charitable remainder trust portfolios,” journal of taxation, 1996,v84. (3, mar), 150–156. treasury regulations, section 1.664–1. charitable remainder trusts. treasury regulations, section 1.664–2. charitable remainder annuity trust. treasury regulations, section 1.664–3. charitable remainder unitrust. united states codes, title 26, internal revenue code, section 2041. powers of appointment. united states codes, title 26, internal revenue code, section 2055. transfers for public, charitable, and religious uses. 195j.c. bost, t. cherin / financial services review 9 (2000) 183–195 pii: s1057-0810(00)00057-3 asset allocation decisions in retirement accounts: an all-or-nothing proposition? doug waggle*, basil englis campbell school of business, berry college, mount berry, ga 30149-5024, usa received 28 june 1999; received in revised form 6 november 1999; received in second revised form 10 march 2000 abstract an examination of survey responses about individual retirement account (ira) holdings reveals that individuals often take all-or-nothing approaches in their decisions to diversify across the asset categories of cash, bonds, and equity. two thirds of survey respondents put their entire ira holdings into a single asset category. a surprisingly large proportion of funds is held in cash, while only a minimal amount is invested in bonds. these findings also contrast with those of bodie and crane’s (1997) examination of tiaa-cref participants, which is heavily weighted with individuals holding fixed income annuities. our results suggest that there is a compelling need for risk education for investors. © 2000 elsevier science inc. all rights reserved. jel classification:d12; g11 keywords:individual investors; retirement accounts; asset allocation 1. introduction the portfolio decision-making of individual investors is an issue of increasing importance as more and more individuals take personal control of their investment and retirement accounts. defined contribution retirement plans, as contrasted with the more traditional defined benefit retirement plans, are becoming the norm rather than the exception. for * corresponding author. tel.:11-706-290-2681; fax:11-706-238-7854. e-mail addresses:dwaggle@berry.edu (d. waggle), benglis@berry.edu (b. englis). financial services review 9 (2000) 79–92 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00057-3 example, jacobius (1999) presents survey results showing that 51% of employees with retirement plans are in defined contribution plans. 401(k) and 403(b) plans generally give employees of all levels and backgrounds discretionary control over their retirement futures by allowing them to choose how to allocate their retirement contributions among a limited number of selections in various asset categories. participants in individual retirement accounts (ira) and keogh plans have even more freedom in allocating their investments among a nearly unlimited number of financial asset choices. portfolio decisions of individual investors are also relevant to the current debate regarding proposed reform of the u.s. social security system. one strategy under consideration would place a portion of social security taxes in self-directed individual retirement accounts (see olsen & baylyff, 1998 and georges, 1998a, 1998b for discussions of this proposal). this would mean that virtually all wage earners would be involved, to some degree, in asset allocation decisions. many policy makers are concerned that the majority of u.s. citizens are ill-prepared to handle this responsibility. poor investment decision-making by individuals in their retirement accounts could have a detrimental impact on their future income (see kim & wong, 1997). suboptimal investment of these funds would result in reduced wealth for older americans, a group that peterson (1999) says will grow considerably over the next several years. therefore, it is essential to develop a clear understanding of individual investor decision-making as a potential guide for financial advisors and policy makers. in this paper, we study the asset allocation decisions of individuals in their retirement accounts and examine some of the factors affecting those decisions. our data are drawn from the macromonitor survey conducted by the stanford research institute (sri). the survey provides self-reported data from a representative sample of over 3,900 american consumers. the protocol includes detailed information on respondents’ portfolio positions and a broad range of demographic and psychographic information. specifically, we look at how individuals allocate their personal ira and simplified employee pension plan (sep-ira) portfolios between cash, bonds, and equity. ira holdings may come from contributions by the individual or from rollovers of employer-sponsored 401(k) or 403(b) retirement plans. sep-iras are geared toward small business owners with no employees. both of these plans allow individuals discretionary control over their investments. portfolio decisions within ira accounts are unique in providing investors with nearly complete discretionary control over their funds with an investment time horizon that is generally known in advance. the investment time horizon is an important factor in portfolio asset allocation decisions. the investments in these plans are also unencumbered by employer-imposed choices and are made without an employer’s guidance. investors are saving for a common goal (retirement) with a definite time horizon, and there are severe penalties for taking funds out of iras before reaching retirement age. prior to the taxpayer relief act of 1997, funds withdrawn from ira accounts before retirement, other than for the purchase of a first home, were subject to ordinary income tax and a 10% penalty. the number of conditions under which the 10% penalty could be avoided was increased considerably in 1997. the remainder of the paper is organized as follows: first we discuss considerations associated with the asset allocation decisions of individual investors. next, we describe the 80 d. waggle, b. englis / financial services review 9 (2000) 79–92 stanford research institute survey from which our data are drawn. in the third section we discuss our empirical results, and in the final section we present our conclusions. 2. asset allocation decisions of individual investors there are several key factors that generally influence investor asset allocation decisions. a primary decision-making factor is the expected investment time horizon which, in the case of iras, is the expected length of time until retirement and the initial withdrawal of funds. while individuals may begin withdrawing funds at retirement, they will likely plan on receiving income from their accounts over the remainder of their lives. the proxy for the beginning of retirement is typically the individual’s age. the younger the individual is, the longer the expected time until retirement. there are numerous articles addressing the issue of how investor asset allocation decisions are affected by changes in investment horizon timelines. early work by samuelson (1969) argues that portfolio allocation is independent of the time horizon. later works by samuelson (1989, 1990, 1994) and kritzman (1994) use the assumption of constant relative risk aversion to support the initial argument that investors should prefer the same mix of assets regardless of the time horizon. olsen and khaki (1998) and bierman (1998) argue that utility theory can support either increasing or decreasing equity allocations with changes in the investment horizon. butler and domian (1991), thaler and williamson (1994), thorley (1995), and bierman (1997) analyze historical data on returns of asset classes and find that as the investment horizon lengthens, investors should allocate higher levels of equity to their portfolios. these findings are consistent with practitioner advice (such as that offered by bogle, 1994) and the oft-cited rule of allocating 100 minus the individual’s age to equity. thus, a 30-year-old individual would have a 70% equity investment, and a 65-year-old individual would have only 35% equity. bodie and crane (1997) examine a survey of tiaa-cref participants and find actual investment behavior consistent with this advice. levy and gunthorpe (1993) and hodges, taylor, and yoder (1997), on the other hand, use multiperiod models employing mean-variance analysis and sharpe’s (1994) ratios, respectively, and conclude that investors should allocate less to equity as the investment horizon lengthens. bodie, merton, and samuelson (1992) add an individual’s labor supply to the analysis and end up supporting the practitioner view. young individuals, with most of their working careers in front of them, should be more willing to take on risky investments and, thus, should allocate higher levels to equity. an individual’s labor supply is relatively safe and the younger he/she is the larger the future working capacity. if risky equity investments take an unfavorable turn, younger individuals have sufficient working time to make up the deficit, while older individuals nearing retirement have a much lower labor supply with which to recoup losses. to compensate for this, older individuals should allocate less to risky equity. overall wealth should influence this age-related risk sensitivity as well. economic theory suggests that higher wealth levels are generally consistent with higher percentage equity investments in retirement accounts. for example, the assumption of decreasing relative risk aversion implies that individuals are less sensitive to proportional 81d. waggle, b. englis / financial services review 9 (2000) 79–92 changes in wealth as the level of wealth increases. in other words, someone with a high level of wealth is better able to absorb proportional losses associated with higher risk investments and still be in a position to provide for an acceptable level of retirement income. the impact of wealth on retirement savings is constrained at very low levels of wealth since individuals typically first need to obtain sufficient levels of cash to meet the demands of everyday transactions and to prepare for unforeseen emergencies before turning to retirement investing. practitioners advise that individuals should set aside three to six months of living expenses in safe, nonretirement accounts before putting nonretirement funds in bonds and equity. thus, individuals with low net worth might be holding 100% of their assets in cash in nonretirement accounts just to meet contingencies. perhaps the most pervasive advice offered by investment professionals (see, e.g., bogle, 1994) is that individuals diversify their portfolios both across and within asset categories regardless of other individual investor characteristics. although holding a small proportion of cash might be wise in certain situations, retirement accounts should be predominantly invested in equity and long-term bonds. investment advisors have also suggested concentrating long-term equity in nonretirement accounts and taxable bonds in tax deferred retirement accounts to maximize tax efficiency. the argument is that long-term capital gains on equity are taxed at a lower rate than interest payments on bonds, so a greater tax advantage is gained by holding equity in nonretirement accounts. even without the differential tax rates, financial advisors have suggested that individuals can simply hold on to their stocks and avoid paying the taxes on the unrealized gains. this advice ignores the fact that the average equity mutual fund turns over its entire portfolio in about a year and is not managed with tax efficiency in mind. the tax inefficiency of mutual funds coupled with the consistent double-digit returns of equity in the 1990s might be a reasonable argument in favor of holding more equity in tax-deferred accounts. bodie and crane (1997) did not find evidence that people were concentrating equity in nonretirement accounts. other factors that may affect portfolio decision-making include the marital status and education level of the individual and whether or not a home is owned. if one’s home is considered to be a relatively safe asset, then home ownership may allow individuals to diversify their portfolios into more risky equity investments. a couple with two incomes may feel more secure than an individual relying on a single salary, and, thus, might be willing to take on more risk in their investment portfolio. a more educated individual may have a better understanding of the stock market and be more receptive to investments in equity. bodie and crane (1997) found home ownership and college education (when omitting job category variables), but not marital status, to be significant factors affecting equity allocation. demographic variables may also help explain the investment decisions of groups, but they are imperfect predictors of individual investment behavior. individuals may have different attitudes toward risk that overshadow other characteristics when it comes to portfolio decisions. while different economists may have opposing views regarding general risk tolerance, all agree that no two individuals are alike. the sri survey provides responses that give some useful insight into the risk tolerance of the individuals. 82 d. waggle, b. englis / financial services review 9 (2000) 79–92 3. description the macromonitor is sri consulting’s consumer financial decisions database and marketing program. in the present analysis, we use data from the 1996–97 survey, which includes interviews with 3,931 financial decision-makers drawn from a nationally representative sample of american households. the survey sample is constructed using a randomdigit-dialing sampling frame based on both listed and unlisted numbers. prospective households are first contacted by telephone to solicit cooperation for a mail survey concerning financial decision-making. only individuals who agree to participate are sent a questionnaire along with a small incentive and a return stamped envelope. postcard and telephone follow-ups are used to encourage participation. this approach results in a 40% response rate among households who originally agreed to participate. a stratified sample is used that over-represents high-income households. of the households originally contacted, over 1,700 have annual incomes that exceed $75,000 or total assets over $300,000 (excluding primary residence). the macromonitor questionnaire is an 85-page protocol that represents a broad array of attitudinal, behavioral, and motivational information. demographic variables include age, income, education, gender, race, marital status, family structure, occupation, employment status, and business ownership. a consumer’s reported current financial status reflects the incidence and dollar amount of holdings in financial products and a behavioral inventory assesses household use of comprehensive array of financial services. the sample used in this paper includes 942 families with iras or sep-iras and with no missing ira/sep-ira financial data. these are the only retirement categories for which detailed financial information was collected. iras and seps are voluntary plans that are more likely to attract individuals in upper income categories with sufficient funds to set aside for their retirements. not surprisingly, our sample is more educated and affluent than the population as a whole. the majority of the household heads have completed college (71%), and most are married (81%). the average annual income of the group was $102,000 and most own their own homes (92%). average net worth, including the value of homes, is $725,000. the average age of the heads of households in our sample is 51. our use of ira data to gain insights into the asset allocation decisions of individual investors in retirement accounts has some limitations. while we have complete retirement data on iras, individuals may have other retirement accounts that enter into their asset allocation strategies. for example, individuals may hold equity in their ira accounts and bonds in their 401(k) accounts; and we can only observe the former. despite this fact, for the reasons previously stated, observation of activity within ira accounts does provide much useful information. table 1 shows a breakdown of the sample size by age of the head of household and family net worth, including home value. for presentation purposes, we have broken both age and net worth into four categories, giving us a total of 16 different groups. as expected, there is a direct relationship between age and net worth such that the youngest people are predominantly in the lowest net worth quartile, while the oldest people are primarily in the upper net worth quartiles. 83d. waggle, b. englis / financial services review 9 (2000) 79–92 4. empirical results 4.1. descriptive statistics in table 2 we show the retirement portfolio allocation decisions between cash, bonds, and equity as a function of the age of the head of household and net worth. cash includes savings accounts, savings certificates, money market accounts, and money market mutual funds. the bond category includes bonds of all types and bond mutual funds, as well as any annuities. equity includes holdings of both stocks and stock mutual funds. the portfolio allocation percentages are simply the dollar amount in each category divided by the total assets in the retirement account. the table shows that investment decisions are clearly related to both age and net worth. higher age groups invest less in equity, and higher net worth groups invest more. table 1 sample size by age and net worth net worth quartile age ,45 45–54 55–64 651 total lowest 144 58 19 14 235 second 84 80 43 29 236 third 49 100 42 44 235 highest 27 66 73 70 236 total 304 304 177 157 942 table 2 asset allocation by age and net worth net worth quartile asset age ,45 45–54 55–64 651 lowest cash 37.1% 46.4% 70.9% 78.6% bonds 9.7% 11.0% 4.2% 21.4% equity 53.1% 42.6% 24.9% 0.0% second cash 31.3% 32.6% 46.1% 71.3% bonds 6.1% 10.6% 12.8% 13.0% equity 62.6% 56.9% 41.1% 15.7% third cash 18.9% 23.9% 29.1% 55.1% bonds 12.5% 14.0% 17.3% 13.8% equity 68.6% 62.1% 53.6% 31.1% highest cash 25.3% 18.0% 19.5% 34.4% bonds 3.5% 12.5% 16.8% 13.9% equity 71.2% 69.5% 63.7% 51.7% the table presents the average decision of respondents regarding the percent of total retirement dollars invested in each asset category based on the family net worth and the age of the head of household. for example, the percentage cash is total cash in the retirement account divided by the total assets in the retirement account. for each age/net worth category, the sum of cash, bonds, and equity equals 100%. (there may be small differences due to rounding.) 84 d. waggle, b. englis / financial services review 9 (2000) 79–92 a surprising observation is the predominance of cash in the retirement portfolios. across all 16 age by net worth groups, the average allocation to cash is 34.4%, with bond and equity holdings at 11.9% and 53.7%, respectively. of particular note is the finding that cash is weighted much more heavily than bonds. in fact, 76.3% of the total sample holds no bonds whatsoever in their retirement accounts. based on expectations derived from finance theory, bonds are underrepresented in the portfolios and cash is over-represented. as previously noted, the classic recommendation for long-term retirement accounts is to place assets primarily in long-term bonds and equity. cash should be held in nonretirement accounts. while we expect a distribution about a perceived optimal percentage allocation to equity (or bonds or cash) with gradual declines in allocation to either side of the ideal, tables 3 and 4 reveal that the distributions are somewhat bimodal in nature. for example, for respondents under age 45, the average equity allocation is 59.8%, so we might expect that most individuals would have between 50 and 75% of their portfolios allocated to equity. table 3, however, reveals that only 6.9% of respondents made this decision. instead, over 70% of investors in this age group held either 100% equity (44.4% of the sample) or no equity (30.3% of the sample). a similar bimodal pattern is observed for all age groups with the sharpest decline by age shown for the 100% equity allocation category. similar bimodal patterns are seen in table 4, which shows equity allocations as a function of net worth. for table 3 equity allocation decisions by age age %equity 0 .0–,25 25–,50 50–,75 75–,100 100 ,45 30.3% 3.0% 5.3% 6.9% 10.2% 44.4% 45–54 29.6% 1.6% 7.9% 10.2% 9.9% 40.8% 55–64 30.5% 6.8% 6.8% 15.3% 11.9% 28.8% 651 55.4% 1.9% 5.7% 8.3% 10.8% 17.8% the table presents the percentage of respondents in each age category choosing a particular equity allocation. for example, 30.3% of those under 45 years of age put no money at all in equity, while 44.4% put all of their funds in equity. each of the rows sums to 100%, with minor exceptions for rounding. table 4 equity allocation decisions by net worth net worth quartile %equity 0 .0–,25 25–,50 50–,75 75–,100 100 lowest 45.5% 3.4% 6.0% 6.0% 3.4% 35.7% second 40.7% 0.8% 6.8% 6.8% 5.1% 39.8% third 29.4% 3.4% 6.8% 12.3% 16.6% 31.5% highest 21.6% 4.7% 6.4% 14.0% 16.9% 36.4% the table presents the percentage of respondents in each net worth category choosing a particular equity allocation. for example, 45.5% of those in the lowest net worth category put no money at all in equity, while only 6.0% of those in that category put between 50 and 75% of their funds in equity. each of the rows sums to 100%, with minor exceptions for rounding. 85d. waggle, b. englis / financial services review 9 (2000) 79–92 the highest net worth quartile, the average equity investment is 62.6%. table 4 shows that 21.6% of the highest net worth group put no money in equities while 36.4% put all of their money into them. all-or-nothing investments in equity appear to be the norm rather than the exception. table 5 looks at the distribution of equity allocation decisions based on whether or not respondents own a home, completed college, or are married. the bimodal nature of the data are again revealed. while marital status does not appear to be an important factor, lack of home ownership is reflected with a more pronounced bimodal distribution. over 80% of those without a home were invested all or nothing in equity compared to 69% of homeowners. the most revealing observation is that fully 50.2% of those without a college degree chose to put no money at all into equity while only 27.8% of those with college degrees made the same decision. households led by individuals without college degrees appear far more likely than their more educated counterparts to avoid equity investments altogether. tables 6 and 7 consider the lack of diversification across asset categories. table 6 shows that for those 65 and older, the average allocations to cash, bonds, and equity were 50.9%, 14.4%, and 34.7%, respectively. however, more careful examination reveals that 42.7% of this group was invested in all cash, 6.4% was invested in all bonds, and 17.8% was invested in all equity. thus, for those 65 and over, 66.9% of the retirement accounts are invested in a single asset category. for the entire sample, 67.3% of the accounts were not diversified across asset categories. for the lowest income quartile, shown in table 7, 37.4% chose all cash, 6.4% chose all bonds, and 37.4% chose all equity in their retirement investments. fully 79.6% of those in the lowest net worth quartile did not diversify across asset categories. in as much as ira accounts are representative of total retirement holdings, these findings strongly suggest that most individual investors are not diversifying their retirement portfolios at all. when individuals do diversify, they tend to hold too much cash and to underinvest in bonds. table 5 equity allocation decisions based on home ownership, completion of college, and marital status %equity 0 .0–,25 25–,50 50–,75 75–,100 100 own home no 40.8% 2.8% 4.2% 8.5% 4.2% 39.4% yes 33.5% 3.1% 6.6% 10.0% 11.1% 35.7% college degree no 50.2% 1.8% 6.2% 6.2% 4.8% 30.8% yes 27.8% 3.6% 6.6% 11.1% 12.9% 37.9% married no 38.9% 3.4% 4.6% 6.3% 13.1% 33.7% yes 33.2% 3.0% 6.9% 10.6% 10.0% 36.3% the table presents the percentage of respondents choosing a particular equity allocation based on home ownership, completion of a college degree by the head of the household, and marital status. for example, 50.2% of respondents where the head of household did not complete college put no money at all in equity while only 27.8% of those who completed college did the same. each of the rows sums to 100%, with minor exceptions for rounding. 86 d. waggle, b. englis / financial services review 9 (2000) 79–92 in addition to the demographic factors of age, net worth, education level, marital status, and home ownership, our analysis examines data relevant to the risk tolerance of individual investors. different individuals who are the same age with identical financial situations and educational backgrounds might have completely different tolerance for risk. while finance theory might suggest that younger individuals should invest more heavily in equity, a young individual with very limited risk tolerance might be reluctant to do so. the sri survey protocol includes several measures designed to gauge the investor’s attitudes toward risk and investing. we focus on two items that provide self-reported measures of the risk tolerance of individuals. table 6 asset allocation decisions and failure to diversify across asset categories by age asset allocation investment decision age cash% bonds% equity% no cash all cash no bonds all bonds no equity all equity all one category ,45 31.5% 8.6% 59.8% 56.9% 25.0% 83.6% 4.9% 30.3% 44.4% 74.3% 45–54 29.2% 12.2% 58.6% 53.6% 21.1% 76.0% 5.3% 29.6% 40.8% 67.1% 55–64 33.8% 14.6% 51.6% 45.2% 24.3% 67.2% 2.8% 30.5% 28.8% 55.9% 651 50.9% 14.4% 34.7% 28.0% 42.7% 73.2% 6.4% 55.4% 17.8% 66.9% full sample 34.4% 11.9% 53.7% 48.8% 26.5% 76.3% 4.9% 34.3% 35.9% 67.3% the cash%, bonds%, equity% columns present the average decision of respondents regarding the percent of total retirement dollars invested in each of the asset categories based on the age of the head of household. for example, the percentage cash is total cash in the retirement account divided by the total assets in the retirement account. the sum of cash, bonds, and equity is 100%, with small differences possible due to rounding. the various investment decisions show the percentage of each age category making the particular decision. for example, 56.9% of those with heads of households under 45 chose to hold no cash, while 25% chose to hold all cash in their retirement accounts. a full 74.3% of those under 45 chose either all cash, or all equity. table 7 asset allocation decisions and failure to diversify across asset categories by net worth asset allocation investment decision net worth quartile cash% bonds% equity% no cash all cash no bonds all bonds no equity all equity all one category lowest 44.6% 10.3% 45.1% 47.7% 37.4% 83.4% 6.4% 45.5% 35.7% 79.6% second 39.3% 9.7% 51.0% 49.2% 33.5% 83.1% 3.4% 40.7% 39.8% 76.7% third 29.6% 14.3% 56.1% 46.8% 19.6% 71.5% 6.4% 29.4% 31.5% 57.4% highest 24.2% 13.2% 62.6% 51.7% 15.7% 67.4% 3.4% 21.6% 36.4% 55.5% the cash%, bonds%, and equity% columns present the average dicision of respondents regarding the percent of total retirement dollars invested in each of the asset categories based on the family net worth. for example, the percentage cash is total cash in the retirement account divided by the total assets in the retirement account. the sum of cash, bonds, and equity is 100%, with small differences possible due to rounding. the various investment decisions show the percentage of each net worth category making the particular decision. for example, 47.7% of those in the lowest net worth category chose to hold no cash, while 37.4% chose to hold all cash in their retirement accounts. a full 79.6% of those with the lowest net worths chose either all cash, all bonds, or all equity. 87d. waggle, b. englis / financial services review 9 (2000) 79–92 the first questionnaire item we analyze relates to the importance of earning high yields. survey participants were asked whether theyagree mostly, agree somewhat, disagree somewhat,or disagree mostlywith the statement“it is wise to put some portion of savings in uninsured investments to get a high yield.”table 8 shows the breakdown of responses by age. nearly 35% of the sample disagreed (somewhat or mostly) with this statement, which may help to explain why so many investors hold 100% of their retirement assets in cash. a second item deals with the respondent’s willingness to incur high risk in order to reap a large potential benefit. respondents were asked whether theyagree mostly, agree somewhat, disagree somewhat,or disagree mostlywith the statement“i am willing to take substantial risks to realize substantial financial gains from investments.”the two items are not fully independent, as agreement with this statement means the respondent probably agreed with the first statement, but the relationship does not hold in the opposite direction. while older respondents might generally be more risk averse, there are certainly young investors who are risk-averse and older individuals who are willing to accept more risk. as shown in table 8, about 55% of the sample population somewhat or mostly disagreed regarding their willingness to take substantial risks to obtain substantial returns. not surprisingly, for individuals 65 or older, this percentage increases to 75%. although analysis of such measures of risk aversion may prove of interest, the investment behavior of individuals is not always consistent with their stated risk tolerance (as demonstrated in the work of jianakoplos & bernasek, 1998). table 8 survey questions assessing attitudes toward risk by age “it is wise to put some portion of savings in uninsured investments to get a high return.” age age mostly agree somewhat disagree somewhat disagree mostly ,45 26.6% 43.2% 22.9% 7.3% 45–54 29.1% 38.1% 21.1% 11.7% 55–64 29.7% 38.3% 19.4% 12.6% 651 17.2% 30.5% 26.5% 25.8% full sample 26.5% 38.6% 22.2% 12.7% ‘‘i am willing to take substantial risks to realize substantial financial gains from investments.’’ age age mostly agree somewhat disagree somewhat disagree mostly ,45 15.8% 36.8% 32.6% 14.8% 45–54 11.3% 36.7% 31.7% 20.3% 55–64 12.6% 30.9% 25.1% 31.4% 651 2.6% 21.2% 31.8% 44.4% full sample 11.6% 33.1% 30.8% 24.5% the table presents the percentage of respondents responding in a particular fashion to the above statements based on the age of the heads of households. for example, 26.6% of respondents with heads of household under 45 mostly agreed with the first statement about high yields. each of the rows sums to 100%, with minor exceptions for rounding. 88 d. waggle, b. englis / financial services review 9 (2000) 79–92 4.2. regression analysis the final part of our analysis is designed to examine the relative contributions of several demographic factors on individual investor retirement account asset allocations. leastsquares multiple regression is employed to explain some of the dispersion in equity allocation decisions. we use the logistic transformation of the percentage invested in equity (equity/total assets) as the dependent variable equity%. since the percentage invested in equity is constrained to a minimum of 0% and a maximum of 100%, the logistic transformation provides the most suitable fit for our data. the logistic transformation is the natural log of the percentage equity divided by one minus the percentage equity. substitutions of 0.001 for 0% equity and 0.999 for 100% equity were made. the regression equation takes the following form: equity%i 5 b1 1 b2 age i 1 b3 net worthi 1 b4 own home i 1 b5 collegei 1 b6 married i (1) age is the age of the survey respondent and net worth is the family unit net worth in thousands of dollars. own home, college, and married are all dummy variables that are one if the statement is true and zero otherwise. the results of the regression equation are shown in table 9 . age, net worth, and college are all significant at the 0.05 level. the coefficient on age has the expected negative sign showing that the percentage of assets invested in equity decreases as the investor ages. the positive coefficient on net worth shows that the equity allocation in retirement accounts increases with higher net worth levels. completion of college is also consistent with a higher allocation to equity. 5. conclusions we have examined the asset allocation decisions that a large sample of individuals made in their retirement accounts. our sample group is wealthier and more educated than the population as a whole, so they should be expected to make investment decisions that are table 9 regression of equity allocation decision independent variables coefficient t-statistic intercept 2.8668 2.74** age 20.1007 26.52** net worth 0.0003 2.29* own home 1.0210 1.35 college 1.7269 4.14** married 0.2882 0.57 f 5 14.0410, p-value# .01. * significant at the 5 percent level. ** significant at the 1 percent level. 89d. waggle, b. englis / financial services review 9 (2000) 79–92 generally better informed than the typical investor. despite this, we find that the sample studied tended overall to hold less diversified retirement portfolios than would be expected, to underinvest in bonds, and to hold high levels of cash in their portfolios. the results shed light on the role of several demographic and psychological characteristics on asset allocation patterns. our findings also provide clues about the likely behavior of investors as they gain control over retirement funds through either employer-sponsored, self-directed retirement accounts or changes in the current social security system. as a cautionary note, we once again point out that our sample data only includes ira accounts, and individuals may be diversifying across other retirement accounts that we cannot observe. a particularly striking result of our analysis is the finding that a majority of investors made all-or-nothing asset allocation decisions. over two-thirds of investors put their ira funds entirely in cash, bonds, or equity. diversification across asset categories was the exception, rather than the rule, for all age and net worth groupings. in addition, cash held a much higher position in the accounts than advocated by investment professionals. over 34% of retirement funds were held in cash with 26% of respondents investing their entire accounts in cash. investors with high cash levels may be very cautious or may hold ira accounts at banks where noncash investment opportunities are not as salient to customers. bonds were not widely held by individuals, with three fourths of the sample holding no bonds at all. the dispersion of equity investments was somewhat bimodal in nature. no equity was held in 34% of accounts, while 36% of accounts were all equity. the level of bond ownership and diversification patterns in our sample differ considerably from those observed by bodie and crane (1997) in their analysis of asset allocation decisions by tiaa-cref participants. however, it is important to note that bodie and crane report only average allocations in retirement accounts and, thus, it is not possible to determine whether or not their sample exhibits bimodal tendencies. bodie and crane found average bond percentage allocations of over 40% taken across all net worth levels. this is in sharp contrast with the present findings based on an unrestricted asset allocation situation (ira and sep-ira accounts). as noted by bodie and crane (1997), their results may be unique to tiaa-cref accounts. much of this discrepancy relates to the fact that tiaa, the fixed income annuity that is the oldest tiaa-cref option, controls a significant percentage of participant money and is included in bodie and crane’s bond category. once an investment is made in the annuities, investors who want to change their selection can only do so over a ten-year period. furthermore, some institutions even require that participants annuitize at retirement. for the general population with an ira at a mutual fund family, such as fidelity or vanguard, investors choose between an array of mutual fund options and can generally move their money at will. the findings of the sri survey suggest that the holdings of tiaa-cref participants may not be indicative of the population as a whole. some key demographic factors impact asset allocation decisions. we find that older individuals hold a smaller proportion of assets in equity. a college education is consistent with higher investments in equity. of those in our sample without college degrees, half had their entire retirement accounts invested completely in cash. higher levels of net worth are also related to higher equity allocations. marriage and home ownership do not affect asset allocation decisions. all of these demographic findings except for the lack of significance of home ownership are comparable to bodie and crane’s (1997) findings. 90 d. waggle, b. englis / financial services review 9 (2000) 79–92 the wisdom of diversifying across asset categories is widely touted by investment professionals. based on behavior in retirement accounts, individuals have taken little notice. the present findings suggest that more effective educational efforts in the area of asset allocation and diversification in general and specifically in connection with retirement planning need to be directed at individual investors. while not directly related to asset allocation, murray (1999) shows that employer educational programs increased enrollment and the level of contributions in 401(k) plans. educational efforts aimed at allocation decisions could have a similar result. unknowledgeable investors focused on the short-term volatility of the stock market may miss the long-run potential that equity can provide. on the other hand, portfolios comprised completely of equity might expose investors to unnecessary risk. putting retirement funds in cash may seem to be a safe alternative to stocks, but this strategy ignores purchasing power risk, which can be quite significant in the long run. if safety is a primary concern, bonds may be a better alternative than cash. to make informed investment decisions in their retirement accounts, investors need to fully understand the different features of the asset classes; and it seems clear that they do not. the investment behavior of individuals also has ramifications for financial services firms. financial services professionals may need to examine more carefully why consumers appear averse to investing in bonds and to seek ways of overcoming this aversion. although beyond the scope of the current paper, there are several financial services marketing issues that are relevant as well. for example, it may turn out that financial services professionals need to develop new investment products that will be easier for consumers to understand and that more clearly connect with consumers’ investment goals. bodie and crane (1999), for example, propose and discuss the benefits of a new retirement product. the industry might need to place stronger emphasis on “pre-packaged” retirement investments (e.g., balanced mutual funds) that provide “appropriate” levels of diversification suited to the investment goals and risk tolerance of different groups of consumers. acknowledgments we thank two anonymous reviewers and frank stephenson for their helpful comments. references bierman, h. jr. (1997). portfolio allocation and the investment horizon.journal of portfolio management, 23, 51–55. bierman, h. jr. (1998). a utility approach to the portfolio allocation decision and the investment horizon. journal of portfolio management, 25,81–87. bodie, z., & crane, d. b. (1999). the design and production of new retirement savings products.journal of portfolio management, 25,77–82. bodie, z., & crane, d. b. (1997). personal investing: advice, theory, and evidence.financial analysts journal, 53, 13–23. bodie, z., merton, r. c., & samuelson, w. f. (1992). labor supply flexibility and portfolio choice in a life cycle model. journal of economic dynamics and control, 16,427–449. 91d. waggle, b. englis / financial services review 9 (2000) 79–92 bogle, j. c. (1994).bogle on mutual funds.homewood, il: richard d. irwin. butler, k. c., & domian, d. l. (1991). risk, diversification and the investment horizon.journal of portfolio management, 17,41–47. georges, c. (1998a). overhaul of social security is endorsed by panel.wall street journal,may, 18, a3. georges, c. (1998b). social-security privatization effort makes headway.wall street journal,june, 24, a24. hodges, c. w., taylor, w. r. l., & yoder, j. a. (1997). stocks, bonds, the sharpe ratio, and the investment horizon.financial analysts journal, 53,74–80. jacobius, a. (1999). retirement programs cover 70% of employees.pensions and investments, 27,25. jianakoplos, n. a., & bernasek, a. (1998). are women more risk averse?economic inquiry, 36,620–630. kim, c., & wong, k. m. (1997). asset allocation strategies for personal pension contributions.financial practice and education, 7,35–46. kritzman, m. (1994). what practitioners need to know . . . about time diversification.financial analysts journal, 50, 14–18. levy, h., & gunthorpe, d. (1993). optimal investment proportions in senior securities and equities under alternative holding periods.journal of portfolio management, 19,30–36. murray, m. c. (1999). employee education boosts participation, 401(k) sponsors find.national underwriter, 103,25–28. olsen, k. a., & baylyff c. (1998). keeping track of social security reform proposals: an update.ebri notes, april, 6–11. olsen, r. a., & khaki m. (1998). risk, rationality, and time diversification.financial analysts journal, 54, 58–63. peterson, p. g. (1999). gray dawn: the global aging crisis.foreign affairs, 78,42–55. samuelson, p. a. (1969). lifetime portfolio selection by dynamic stochastic programming.review of economics and statistics, 51,239–46. samuelson, p. a. (1989). the judgement of economic science on rational portfolio management: indexing, timing, and long-horizon effects.journal of portfolio management, 16,4–12. samuelson, p. a. (1990). asset allocation could be dangerous to your health.journal of portfolio management, 16, 5–8. samuelson, p. a. (1994). the long-term case for equities.journal of portfolio management, 21,15–24. sharpe, w. f. (1994). the sharpe ratio.journal of portfolio management, 21,49–58. thaler, r. h., & williamson, j. p. (1994). college and university endowment funds: why not 100% equities? journal of portfolio management, 21,27–37. thorley, s. r. (1995.). the time diversification controversy.financial analysts journal,51, 68–76. 92 d. waggle, b. englis / financial services review 9 (2000) 79–92 financial services review, 33(1) 86 an investigation of the relationship between gender and investor behavior during a market correction matthew sommer,1 megan mccoy,2 and hanna lim3 abstract this study used primary data collected during october 2022 from 2,119 u.s. retail investors to investigate how individuals were coping with the declining stock market and rising inflation. using a path analysis, this study sought to explain the relationships between gender, financial stress, investment overconfidence, and trading behavior. first, a positive relationship was found between males and moving from stocks and bonds to cash. next, the results indicated that females were more likely to have experienced financial stress and males were more likely to have displayed investment overconfidence. both financial stress and investment overconfidence were positively related to moving from stocks and bonds to cash. the indirect effects of financial stress and investment overconfidence, however, were small and only partially mediated the relationship between gender and trading behavior. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation sommer, m., mccoy, m., & lim, h. (2025). investor behavior during a market correction. financial services review, 33(1), 86-101. introduction the purpose of this study was to investigate the relationship between gender and moving stock and bond holdings to cash during a sudden and significant capital market decline. these findings are important because a challenge faced by many financial advisors is helping clients avoid making rash decisions during periods of extreme market volatility (gennaioli et al., 2015). investors who allow their emotions to dictate their actions during these periods are likely to fall into the trap of selling stocks at the worst possible time, which is immediately after they drop in value. it is perhaps not surprising therefore, that individual 1 corresponding author (matthew.sommer@janus.com). janus henderson investors, denver, co, usa. 2 kansas state university, manhattan, ks, usa. 3 california state university, fullerton, fullerton, ca, usa. investors underperform the stock market by an average of 3.0% per year (dalbar, 2023). a more complete understanding of the psychological determinants of male and female investment tendencies will help financial advisors provide the intervention needed to encourage positive, long-term behaviors that close the performance gap (kinniry et al., 2014). in our conceptual model, financial stress and investment overconfidence were explored as mediators to the relationship between gender and investment behavior. during periods of economic hardships such as the covid-19 pandemic and the great recession, females exhibited higher https://creativecommons.org/licenses/by-nc/4.0/ mailto:matthew.sommer@janus.com https://creativecommons.org/licenses/by-nc/4.0/ sommer et al. 87 levels of financial stress (fox & bartholomae, 2020; haslet et al., 2021; peck, 2020). lazarus and folkman’s (1984) transactional theory of stress and coping would suggest that respondents feeling high levels of discomfort would seek to eliminate the root cause and thus, would be more likely to seek the safety of cash. other researchers have found that males were more likely to display investment overconfidence by examining behaviors such as excessive trading (barber & odean, 2001) and shorter time horizons (ferriera-schenk et al., 2021; paisarn et al., 2021). kaheman’s (2011) system 1 and system 2 framework would suggest that overconfident investors are more likely to falsely believe they can successfully time the market. the market correction that occurred during 2022 offered an ideal setting to explore this study’s research question. during the first nine months of the year, the s&p 500 declined by 23.9% while the bloomberg u.s. aggregate bond index fell by 14.6% (bloomberg, 2024). to understand how individual investors were coping with the markets, an online survey was sent to clients of a large u.s. asset manager in early october, resulting in a sample of 2,119 respondents. among this sample, 13% reported having moved stocks and bonds to cash. considering that from october 2022 through december 2023 the s&p 500 rose by 35.8% and the bloomberg u.s. aggregate bond index increased by 7.5% (bloomberg, 2024), investors who sold a portion of their stocks and bonds likely did not fully participate in the eventual market rebound and therefore, were prone to underperforming the broader markets as predicted by dalbar (2023). literature review financial stress financial stress arises when individuals are unable to meet current and ongoing financial obligations (friedline et al., 2020). triggers of financial stress may include worrying about paying bills, losing jobs, providing for children, and saving for retirement (malhotra & witt, 2010). prolonged feelings of financial stress can negatively affect financial satisfaction (lee & dustin, 2021), life satisfaction (stein et al., 2013), financial well-being (heo et al., 2018), psychological well-being (afifi et al., 2017), physical well-being (skinner et al., 2004), and lead to depression (guan et al., 2022). as aforementioned, during periods of economic hardship such as the covid-19 pandemic and the great recession, females were found to have higher levels of financial stress than males. for example, during the covid-19 pandemic, women were disproportionately impacted by financial shocks caused by layoffs, pay cuts, or both (fox & bartholomae, 2020). hasler et al. (2021) found that females had higher levels of both financial stress and financial anxiety compared to men during the pandemic, after controlling for socio-economic status and other demographic characteristics. similarly, simha et al. (2020) found higher levels of stress among u.k. female respondents after controlling for financial vulnerability. peck (2020) suggested that because females had fewer economic resources prior to the covid-19 pandemic, the losses incurred during this period created even higher levels of uncertainty and anxiety. similar conclusions regarding gender differences and financial stress were drawn during the great recession. for example, older adult women, and women of color were more likely to experience mortgage trouble and asset depletion during and after the great recession compared to their male counterparts (castro-baker et al., 2017). afifi et al. (2018) found that during the great recession, women had higher levels of financial stress than their partners when discussing household finances and other money matters. heretick (2013) concluded that both males and females were equally financially stressed, however, the reasons differed. women were more likely to report feelings of anxiety and worry, whereas men were more likely to report shame and guilt. the connection between financial stress and poor investment decision making is abundant within the literature. for example, bernaola et al. (2020) found that respondents who reported higher levels of anxiety displayed less patience with investments that had declined in value. rahman and gan (2020) also found a positive relationship between feelings of anxiety and poor investment decision-making such as demonstrating a misalignment between time horizon and security financial services review, 33(1) 88 selection. among investors in pakistan, moueed and hunja (2020) concluded that under stressful conditions, investors had less control over their thinking and were unable to make optimum use of their cognitive skills. as a result, stressed investors were likely to make rash decisions based on sudden fluctuations in the stock market or lack of diversification within their portfolios. stress has also been found to have an impact on other financial planning behaviors, although the conclusions were mixed. for example, fan and henager (2021) found that feelings of financial stress were negatively related to short-term behaviors such as having emergency funds and paying off credit cards in full. interestingly, financial stress was positively related to longterm behaviors such as calculating retirement needs and saving for retirement. both shortand long-term behaviors were, in turn, related to overall financial well-being. fiksenbaum et al. (2017) found that stress was a motivating factor that increased individuals’ willingness to change certain behaviors that would reduce economic hardship. in this case, stressed respondents were more likely to reduce their spending or find new avenues to increase their income. heo et al. (2024) found that the negative effect of financial stress on financial behavior was weakened during the covid-19 pandemic, suggesting that an appropriate level of stress may serve as a coping mechanism during challenging periods. investment overconfidence overconfidence occurs when individuals think that they know more than they actually do (charupat et al., 2005). a common approach used by researchers to detect and measure overconfidence was to compare an individual’s self-assessed subjective knowledge to how well individuals scored on a short financial literacy quiz. for example, mokhtari and chawla (2023) computed the difference between subjective knowledge and the number of questions answered correctly as a proxy for the degree of overconfidence. another approach identified overconfident individuals as those with high subjective scores but low objective scores using quartiles (robb et al., 2015; zahirovic-herbert et al., 2016) or by comparing means (aristei & gallo, 2021; pearson & korankye, 2022; yeh & ling, 2022). a third approach regressed subjective knowledge on objective knowledge and used the residual term to capture overconfidence (kim et al., 2022; piehlmaier, 2022). researchers have theorized that males were more likely to display investment overconfidence and therefore, trade more frequently than females. excessive trading is considered detrimental to maximizing long-term investment returns due to market friction and mistimed trades (willows & west, 2014). overconfidence has also been linked to holding shorter-term investment horizons (ferriera-schenk et al., 2021; paisarn et al., 2021). one of the first studies regarding this topic examined the trading behaviors of 35,000 households from 1991 through 1997 (barber & odean, 2001). the researchers found that males traded 45% more than females, and trading costs reduced male’s returns by 2.65% compared to a 1.72% reduction for females. similarly, a study regarding the trading behavior of 19,021 south african investors found that over a five-year period from 2007 to 2011, males traded more than females and experienced a greater variance of returns (willows & west, 2014). on a riskadjusted basis therefore, it was concluded that females were better investors than males. controlling for the ‘big five’ personality traits, zhang et al. (2014) found that males traded more than women in both price rising and price falling scenarios in a simulated stock market experiment. researchers that explicitly measured overconfidence painted a much more nuanced relationship between gender and trading activity. for example, cueva et al. (2019) reported that while males did trade more than females, differences in the measured levels of overconfidence did not explain the gender gap in trading activity. competitiveness, risk aversion, and financial literacy were also ruled out as possible explanations. instead, the researchers suggested that perhaps sensation-seeking and gambling attitudes might explain the differences. glaser and weber (2007) found that investors who believed that their investment skills were above average were found to trade more, although no differences were found between males and females. similarly, deaves et al. sommer et al. 89 (2009) found that overconfidence was associated with greater trading volume, although gender did not play a role in the study’s regression models. on the other hand, while fellner-röhling and krügel (2014) found no relationship between overconfidence and trading volume, men traded more than women at higher levels of risk aversion. the gender trading gap vanished as risk aversion decreased. conceptual model and hypotheses development the purpose of this study was to investigate the relationships between gender and trading behavior during a market correction while exploring the mediating role of financial stress and investment overconfidence. regarding these mediators, there are two competing points of view. first, researchers have consistently found that females were more likely to experience financial stress than males, particularly during times of economic uncertainty (fox & bartholomae, 2020; hasler et al., 2021; peck, 2020). lazarus and folkman’s (1984) transactional theory of stress and coping suggested that individuals assess stimuli as having a positive effect, no effect, or negative effect on their well-being. in the case of the latter, stressful stimuli that is perceived as harmful or threatening generates negative emotions, and a secondary appraisal is conducted to determine what can be done to manage and potentially remove the stressor. there are two coping strategies through which stress can be managed: problem-focused and emotional-focused (lazarus & folkman, 1984). problem-focused coping strategies attempt to directly manage the stressful event, while emotional-focused coping strategies seek to regulate the negative feelings caused by the event. the process is iterative as individuals continually reappraise their environment and results of adopting coping efforts. unsuccessful adaptation may lead to the use of additional coping strategies, and continued failure may result in psychological distress. in this case, our working hypothesis is that females are likely to feel higher levels of financial stress during a market correction and to mitigate or eliminate this perceived threat, are more likely to exit the capital markets. the conflicting argument, however, is that males are more likely to feel overconfident in their investment abilities compared to females (willows & west, 2014; zhang et al., 2014). kahneman (2011) suggested that there are two complementary modes of thinking that help individuals assess information and make decisions. system 1 operates automatically and quickly, with little or no effort. system 2 allocates attention to effortful mental activities as needed, including complex calculations. the operations of system 2 often involve choice and concentration. system 1 saves time and energy while system 2 allows for deliberate and careful decision-making. while these systems often work in harmony, misjudgments are likely to occur when difficult decisions are guided by system 1. investment overconfidence occurs because system 1 thinking seeks information that easily comes to mind and constructs a coherent story that makes sense (kahneman, 2011). as a result, important information not readily recalled or known is excluded from consideration. additionally, system 1 thinking is prone to judgment errors including the false belief that knowing the past is knowing the future, inaccurately assessing abilities and knowledge relative to others, unable to discern the differences between luck and skill, and overly relying on intuition. kahneman (2011) stated, “subjective confidence in a judgment is not a reasoned evaluation of the probability that this judgment is correct. confidence is a feeling that reflects the coherence of the information and cognitive ease of processing it” (p. 212). we predict a positive relationship between males and investment overconfidence, and as a result, males are more likely to attempt to successfully time a highly volatile stock market. at this point, we posit that a relationship exists between gender and trading behavior but are uncertain about the direction given conflicting mediating factors. additionally, we believe that these factors, financial stress and investment overconfidence, may help explain the relationship between gender and trading behavior but are uncertain which factor is more dominant. financial services review, 33(1) 90 formally stated, therefore, this study’s hypotheses are: h1: gender is related to moving from stocks and bonds to cash. h2: males are negatively related to financial stress. h3: financial stress is positively related to moving from stocks and bonds to cash. h4: males are positively related to investment overconfidence. h5: investment overconfidence is positively related to moving from stocks and bonds to cash. methodology data and sample this study was conducted in partnership with a leading global asset manager. one of the manager’s lines of business is a direct channel that caters to u.s. retail investors who have established accounts without the assistance of a financial professional (although some investors may use a financial professional for other aspects of their wealth). within this channel, only the asset manager’s proprietary mutual funds are available for purchase. the direct channel was closed to new investors in 2009 but reopened in july 2020. at the end of 2021, the mean and median account balances were $104,614 and $35,782 respectively, and the mean age was approximately 56. in october 2022, an online survey was electronically mailed in batches based on the alphabetical order of the account owner’s last name. the purpose of the survey was to gain insights into how individual investors were coping with recent market volatility and high inflation. the criteria for selection were a balance greater than $0 and an email address on file. respondents were not provided an option to skip questions but could have terminated the survey at any time. after a period of one week and collection of 2,119 responses, the survey ended. dependent variable respondents were asked “as a result of financial market performance and current inflationary environment in 2022, have you moved out of stocks and/or bonds and into cash?” a binary variable was coded as ‘1’ for yes, ‘0’ otherwise. independent variables male respondents were coded as ‘1’ and female respondents were coded as ‘0.’ financial stress was operationalized using the financial anxiety scale (archuleta et al., 2013). respondents were asked on a scale of 1 to 7, where 1 means “never” and 7 means “always,” how often each of the following statements apply to them. the seven statements were ‘i feel anxious about my financial situation,’ ‘i have difficulty sleeping because of my financial situation,’ ‘i have difficulty concentrating on my school/or work because of my financial situation,’ ‘i am irritable because of my financial situation,’ ‘i have difficulty controlling worrying about my financial situation,’ ‘my muscles feel tense because of worrying about my financial situation,’ and ‘i feel fatigued because i worry about my financial situation.’ following the approach used by archuleta et al (2013) and grable et al. (2015), scores were estimated by summing each item. factor loadings achieved 0.68 and above (table 1), and cronbach’s alpha was 0.95. table 1. factor loadings for financial anxiety scale item factor loading i feel anxious about my financial situation 0.6801 i have difficulty sleeping because of my financial situation 0.9086 i have difficulty concentrating on my school/work because of my financial situation 0.9269 i am irritable because of my financial situation 0.8859 i have difficulty controlling worrying about my financial situation 0.9158 my muscles feel tense because of worrying about my financial situation 0.8963 i feel fatigued because i worry about my financial situation 0.9127 sommer et al. 91 overconfidence was operationalized by utilizing the residuals from an ols regression of respondent subjective knowledge on objective knowledge (kim et al., 2022). a single subjective knowledge item asked respondents “how would use assess your overall financial knowledge?” on a scale of 1 to 7, where 1 means “very low” and 7 means “very high” (finra investor education foundation, 2022). objective knowledge was assessed as the number of correct answers to lusardi and mitchell’s (2011) ‘big three’ financial literacy multiple-choice items regarding compounding, inflation, and diversification. ‘don’t know’ responses were coded as incorrect. ‘prefer not to say’ was not provided as an option. socio-demographic characteristics were included in the analysis as categorical variables. these categorical variables included age (29 or younger, between 30 and 39, between 40 and 49, between 50 and 59, between 60 and 69, between 70 and 79, and 80 and older), ethnicity (white and nonwhite), education attainment (high school, some college, bachelor’s degree, and post-graduate degree), household income (less than $50,000, between $50,000 and $99,999, between $100,000 and $199,999, and $200,000 and greater), investments including retirement accounts (less than $500,000, between $500,000 and $999,999, between $1,000,000 and $1,999,999, and $2,000,000 and greater), self-assessed health status (excellent, very good, good, fair, and poor), employment status (employed, partially retired, fully retired, and out of the workforce but not retired), and marital status (married/partnered and single). given the small number of responses, some categories were combined including age, health status, and employment status. empirical strategy to explore the relationships between gender, financial stress, investment overconfidence, and trading behavior a path analysis was specified. a path analysis is used to study complex models where variable a is related to variable b, which in turn is related to variable c (streiner, 2005). it is important to note that a path analysis cannot be used to establish causation or whether a specified model is correct, but it can help identify the direct, indirect, and total effects of the variables under consideration. in our model, we explored the direct effects between gender and moving out of stocks and bonds to cash and the indirect effects through financial stress and investment overconfidence, while controlling for several socio-demographic characteristics. results descriptive statistics this study’s descriptive statistics can be found in table 2. among the sample of 2,119 respondents, approximately 13% reported having moved from stocks and bonds into cash due to the market performance and inflationary environment in 2022. approximately three-quarters of respondents were male (74%). the mean financial stress score was 14.43 (on a scale of 7 to 49) with a standard deviation of 8.41. regarding the financial literacy items, the mean subjective knowledge score was 4.99 (on a scale of 1 to 7) and the standard deviation was 1.17. on average, respondents answered 2.74 of the three financial literacy questions correctly, and the standard deviation was 0.56. a plurality of respondents (35%) was between ages 60 and 69. the majority were white (89%) and college educated (35% had a bachelor’s degree and 42% had a post-graduate degree). more than half of respondents reported household income above $100,000 (57%) and investment assets above $1,000,000 (51%). most respondents were in excellent or very good health (67%), fully retired (52%), and married or partnered (67%). a higher percentage of males (15%) moved out of stocks and bonds compared to females (10%). on average, males reported lower levels of financial stress (13.77 versus 16.28) and higher levels of subjective confidence (5.18 versus 4.54) and objective knowledge (2.80 versus 2.56). all three mean differences were statistically significant (p<0.001). among the other demographic variables, a greater percentage of males reported household income of $100,000 and greater (60% versus 48%) and investments of at least $1,000,000 (54% versus 50%). lastly, 73% of males were married or partnered compared to 51% of females. financial services review, 33(1) 92 table 2. descriptive statistics variable full sample % (n = 2,119) male % (n = 1,562) female % (n = 557) moved out of stocks and bonds to cash 13.26 14.53 9.69 male 73.71 financial stress (mean; 7-49) 14.43 (8.41) 13.77 (7.83) 16.28 (9.62) subjective confidence (mean; 1-7) 4.99 (1.17) 5.18 (1.07) 4.54 (1.28) objective knowledge (mean; 0-3) 2.74 (0.56) 2.80 (0.48) 2.56 (0.73) age less than 50 6.61 5.95 8.44 between 50 and 59 19.58 19.01 21.18 between 60 and 69 35.06 35.08 35.01 between 70 and 79 28.05 29.45 25.85 age 80 and older 10.24 10.50 9.52 white 88.77 88.48 89.59 education attainment: high school 4.62 4.55 4.85 some college 18.74 18.25 20.11 bachelor’s degree 34.87 34.19 36.80 post-graduate degree 41.76 43.02 38.24 household income: less than $50,000 11.18 8.71 18.13 between $50,000 and $99,999 31.57 30.86 33.57 between $100,000 and $199,999 39.69 42.00 33.21 $200,000 and greater 17.56 18.44 15.08 investments: less than $500,000 23.64 20.55 32.32 between $500,000 and $999,999 25.06 24.46 26.75 between $1,000,000 and $1,999,999 25.15 25.22 24.96 $2,000,000 and greater 26.14 29.77 15.98 self-assessed health: excellent 23.69 23.69 23.70 very good 44.03 44.11 43.81 good 24.96 25.03 24.78 fair/poor 7.31 7.17 7.72 fully retired 51.72 52.05 50.81 married/partnered 67.11 72.79 51.17 note: standard deviation in parentheses. sommer et al. 93 path analysis the specified path model and standardized coefficients can be found in figure 1. assessing the model fit, the standardized root mean squared residual (srmr) was 0.031, the root mean square of approximation (rmsea) was 0.073, and the comparative fit index (cpi) was 0.957. regarding the first hypothesis, a positive relationship was found between males and moving from stocks to bonds to cash in 2022 (β = 0.065, p = 0.003). regarding the next two hypotheses, a negative relationship was found between males and financial stress (β = -0.131, p<0.001), and financial stress was in turn, positively related to moving from stocks and bonds to cash in 2022 (β = 0.104, p<0.001). regarding the final two hypotheses, a positive relationship was found between males and investment overconfidence (β = 0.231, p<0.001), and investment overconfidence was in turn, positively related to moving from stocks and bonds to cash in 2022 (β = 0.049, p = 0.025). although support was found for hypotheses 2 through 5, the indirect effects were small. multiplying the estimates for each path indicated an indirect effect of -0.013 for financial stress and 0.011 for investment confidence. when summed, the total indirect effect was -0.002. combining the indirect effect of -0.002 with the direct effect between males and moving from stocks and bonds to cash (0.067) yielded a total effect of 0.065. a summary of the direct effects, indirect effects, and total effects can be found in table 3. figure 1. standardized path coefficients for prediction of moving out of stocks and bonds to cash financial services review, 33(1) 94 table 3. direct, indirect, and total effects for the hypothesized model (n = 2,119) standardized coefficient direct effect indirect effect total effect path to financial stress: male -0.131*** na -0.131*** path to investment overconfidence: male 0.231*** na 0.231*** path to moved out of stocks and bonds to cash: male 0.067** -0.002 0.065** financial stress 0.104*** na 0.104*** investment overconfidence 0.049* na 0.049* *p < 0.05. **p < .01. ***p < 0.001 discussion this study sought to investigate the relationship between gender and moving from stocks and bonds to cash during a market correction. the steep downturn in the stock and bonds markets during the first nine months of 2022 offered an ideal time to explore this topic. first, a positive relationship was found between males and exiting the capital markets in favor of cash. despite finding some evidence about the mediating role played by financial stress and investment overconfidence, the results of a path model indicated there is much more to the story. while strong support was found for the study’s hypotheses, the indirect effects were small meaning one or more variables not identified in our model accounted for gender differences in trading behavior. also, the direction of the indirect effects was opposite, effectively canceling each other out. the following discussion will review how our study builds upon the profession’s existing understanding of the relationship between gender, financial stress, investment overconfidence, and trading behavior while also offering alternative explanations for our results that may be ripe for further investigation. while the market downturn of 2022 was shortlived and did not compare in magnitude to the great recession or the covid-19 pandemic, this event offers insights into how different investors coped with market volatility. as predicted, females in october 2022 were experiencing higher levels of stress compared to their male counterparts. this finding confirms the earlier conclusion of hasler et al. (2021), peak (2020), and castro-baker et al. (2017) regarding gender differences during periods of economic difficulties. further, this study’s conclusions regarding the connection between higher stress levels and suboptimal investment behaviors agrees with earlier findings by bernaola et al. (2020), rahman and gan (2020), and maueed and hunja (2020). this study also found that by comparing what investors think they know to what investors actually know (kim et al., 2022), males were more likely to have an unfounded confidence in their financial abilities. this bias has been used to explain excessive trading as males are more likely to falsely believe they can time the market (barber and odean, 2001). one of the unique aspects of this study was that the administration of the survey instrument occurred during a market trough in october 2022. unlike the barber and odean (2001) study which tracked investors over several years, this study investigated investor behaviors during a particularly challenging period. the connection between overconfidence and moving out of stocks and bonds in the face of a rapidly declining market offers new insights about this bias. while financial stress and investment overconfidence partially explained gender differences in trading behavior during the 2022 market correction, our analysis indicates that there may be other factors behind the relationship between males and exiting the capital markets. for example, because males are likely to have a higher risk tolerance, and thus are likely to hold a greater allocation to stocks (heo et al., 2016) males simply have more to lose than females sommer et al. 95 during a market correction. as stock prices decline, investors with greater exposure may more be tempted to “cut their losses” and seek to preserve principle in the safety of cash and cash equivalents. additionally, market declines may be less salient to investors with smaller stock allocations, and therefore these investors may be more likely to embrace the status quo and refrain from making changes to portfolios. a second possible explanation may be that males are less likely to use and trust a financial advisor compared to females (collins, 2012). the guidance provided by advisors during periods of market volatility about the benefits of maintaining a long-term perspective are invaluable. kinniry et al. (2014) estimated the economic benefits of a financial advisor’s advice was an incremental 3 percent per year, half of which was attributable to ‘behavioral coaching.’ according to kinniry et al. (2014) volatile markets influence investors’ confidence and financial advisors can act as ‘emotional circuit breakers’ by helping clients overcome the natural tendency to sell high and buy low. a final explanation may be the sensation-seeking and gambling attitudes of males suggested by cuervo et al. (2019) who also found that overconfidence did not fully explain the gender trading gap. perhaps males are more prone to excessive trading not only in the hopes of maximizing risk-adjusted returns but also to experience the thrill of stock investing. limitations to address some of the limitations of this study, future research that incorporates asset allocation, financial advisor use, and non-financial motivations behind stock investing would potentially add to the existing body of knowledge. in addition to omitted variable bias, there were three other limitations regarding the sample and survey instrument. first, the asset manager’s clientele skewed towards an older, highly educated, and wealthier cohort which may not be representative of the general population or u.s. retail investors. a more diverse sample may have yielded different results. second, expanding the ‘big three’ objective knowledge items to the ‘big five’ that includes additional items regarding the relationship between interest rates and bond prices and mortgage amortization (lusardi & mitchell, 2011) may have yielded a more complete picture of respondents’ financial literacy. lastly, capturing the percentage of respondent portfolios that had shifted from stocks and bonds to cash may have improved the richness of the analysis. implications our study of gender-related trading differences offers valuable takeaways for financial practitioners. during this period, females were more likely to feel financial stress and males were more likely to feel overconfident in their financial abilities. both the feelings of financial stress and investment overconfidence were linked to moving out of stocks and bonds, which in hindsight may have been a mistake. consider that from october 2022 to december 2023, the s&p 500 gained 35.8% and the bloomberg u.s. aggregate bond index rose 7.5% (bloomberg, 2024). investors who panicked in 2022 likely fell into the trap of buying high and selling low, putting their long-term financial goals such as retirement, funding a college education for a child, or buying a new home in jeopardy. practitioners able to identify signs of stress and overconfidence and intervene with the appropriate tools and techniques may be able to provide a differentiated client experience. at the same time, properly advised investors are positioned to reap the benefits of the incremental returns quantified by kinniry et al. (2014). one of the first steps a practitioner might consider with clients experiencing financial stress is a risk tolerance reassessment. typically, an initial assessment is done by using a questionnaire and associated scoring methodology, with higher scores indicative of a higher risk tolerance leading the practitioner to recommend a greater equity allocation. only a limited number of these commercially used risk-tolerance assessments have been peer-reviewed within the academic community, and as a result may lack reliability and validity (kuzniak et al., 2015). faulty assessments may lead to practitioner recommendations that are not aligned with an investor’s true willingness and ability to assume risk. panicked selling in 2022 may have been partly the result of equity allocations that far financial services review, 33(1) 96 exceeded actual financial risk tolerances. grable and lytton (1999) have developed one of the only peer-reviewed, publicly available risk assessment tools at no cost. the tool consists of 13 multiplechoice style questions with a very straightforward scoring methodology. incorporating this tool into a client service model, particularly during times of market stress as a robustness check to the initial risk assessment, may be very beneficial to both the practitioner and investor. another technique that practitioners may use to assist financially stressed investors is to improve their financial literacy. financial literacy consists of not only objective knowledge but also having the confidence to apply that knowledge (huston, 2012). higher levels of financial literacy have been linked to lower levels of financial stress (xiao & kim, 2021; zhang & chatterjee, 2023). a related concept to confidence is self-efficacy, which refers to people’s beliefs in their capabilities to meet a certain goal or objective (bandura, 1997). letkiewicz et al. (2016) offers practitioners suggestions for improving investor financial self-efficacy through performance accomplishments, vicarious experience, verbal encouragement, and physiological states. accomplishments help build an investor’s confidence and provide motivation to engage in a new task. letkiewicz’s et al. (2016) suggested structuring financial decisions that allow for small accomplishments while learning new skills. one relatively simple example is to establish an emergency fund. moon et al. (2023) found that emergency funds were an effective way to mitigate financial stress, especially during challenging economic periods such as the covid-19 pandemic. cash reserves can also play an important role for retirees. practitioners might suggest that retirees have access to enough cash to meet one to two years of living expenses (benz, 2022). this cash buffer may give investors the peace of mind necessary to stay the course during periods of market volatility knowing no immediate lifestyle changes will be necessary. vicarious experiences occur when an individual sees a peer successfully reaching a goal or objective (letkiewicz et al., 2016). practitioners should be prepared to offer anonymous case studies or vignettes about how similarly situated investors reacted during challenging times or successfully reached certain financial milestones. verbal encouragement also includes constructive feedback. many investors have experienced success in other aspects of their lives and practitioners should remind these individuals the same long-term perspective that was a key element to their personal accomplishments can be applied to investing. lastly, investors feeling nervous or anxious are likely to have low levels of self-efficacy. one way to alleviate investor stress is to establish basic ground rules for engaging during difficult economic periods. an example may be to ask for an in-person meeting to occur, that includes the spouse or partner, before any major changes are made to the portfolio’s asset allocation. this pause reassures the investor that a plan is in place should the current crises worsen, while avoiding rash decisions that can easily be executed over the phone or electronically. overconfident investors present a slightly different challenge. while some investors may share their concerns regarding household finances, workplace uncertainty, or the broader economy thus providing clues of stressful feelings, it is less likely that an investor would admit or even recognize overconfidence in their financial abilities. as part of the new client onboarding process, practitioners might consider administering the three objective questions and one subjective question used in this study (lewis, 2019). the larger the disparity between a client’s self-assessed subjective knowledge score and the number of objective knowledge items answered correctly, the more overconfidence the investor is displaying. regarding existing clients, practitioners might explain that as their service model evolved, a need has been recognized to build the financial literacy of not only existing investors who may be interested but also that of family members. part of the exercise is to establish a baseline through the administration of four questions that will be revisited and tracked over time. the important point for practitioners to stress is not how the questions are answered presently, but rather, the investor’s financial literacy improvements over time. once a financial practitioner identifies an overconfident investor, the educational process sommer et al. 97 might start by addressing objective knowledge. adil et al. (2021) found that objective knowledge had a negative moderating effect on the relationship between overconfidence and suspect investment decision making. since overconfident investors are more likely to engage in market timing, practitioners must be prepared to explain the futility of these actions. a tenet commonly repeated within the financial services industry is “time in the market, not market timing.” as an example, according to an analysis conducted by janus henderson investors (2024), $10,000 invested in the s&p 500 from 1988 through 2022 would have grown to $33,098. if the 10 best trading days, however, were missed during this period the investment would have only grown to $15,163, and if the 20 best trading days were missed the investment would have declined to $8,899. providing investors with these simple messages through easy-to-read illustrations will help reinforce key learnings. managing an investor’s subjective knowledge is likely to prove more challenging than simply providing facts and supporting data. in these cases, two techniques that may be helpful are subjective probability interval estimation (spies) (lurtz, 2020) and premortem planning (klein, 2007). spies is a graphical representation of all possible outcomes. premortem planning starts by posing the question, “what is the worst outcome and why would that occur?” next ask, “what is the best outcome and why would that occur?” presenting both good and bad outcomes reminds investors of suboptimal outcomes not previously considered. in the case of the 2022 market correction, the rebound in 2023 was sudden and dramatic. this period in history can be used to remind market timers that they have to be right twice: once when they sell and again when they buy. evidence was found that supported this study’s hypotheses regarding the relationships between gender, financial stress, investment overconfidence, and moving from stocks and bonds to cash during a market correction. financial stress and investment overconfidence, however, only partially mediated gender differences, inviting opportunities to further explore this important topic. this study adds to the existing body of literature by providing new insights regarding trading behavior during a very challenging investment climate. financial practitioners can use these findings to enhance their client relationships by taking proactive steps to mitigate financial stress and temper investment overconfidence. references adil, m., singh, y., & ansari, m. d. (2021). how financial literacy moderate the association between behaviour biases and investment decision? asian journal of accounting research, 7(1), 17-30. afifi, t. d., davis, s., merrill, a. f., coveleski, s., denes, a., & shahnazi, a. f. (2018). couples’ communication about financial uncertainty following the great recession and its association with stress, mental health and divorce proneness. journal of family and economic issues, 39(2), 205-219. archuleta, k. l., dale, a., & spann, s. m. (2013). college students and financial distress: exploring debt, financial satisfaction, and financial anxiety. journal of financial counseling and planning, 24(2), 50-62. aristei, d., & gallo, m. (2021). financial knowledge, confidence, and sustainable financial behavior. sustainability, 13, 10926. bandura, a. (1997). self-efficacy: the exercise of control. freeman. barber, b. m., & odean, t. (2001). boys will be boys: overconfidence, and common stock investments. the quarterly journal of economics, 116(1), 261-292. benz, c. (2022, november 23). the bucket approach to building a retirement portfolio. morningstar. https://www.morningstar.com/portfolios/ bucket-approach-building-retirementportfolio bernaola, d. m. v., willows, g. d., & west, d. (2020). the relevance of anger, anxiety, gender and race in investment decisions. mind & society, 20(1), 1-21. https://www.morningstar.com/portfolios/bucket-approach-building-retirement-portfolio https://www.morningstar.com/portfolios/bucket-approach-building-retirement-portfolio https://www.morningstar.com/portfolios/bucket-approach-building-retirement-portfolio financial services review, 33(1) 98 bloomberg l.p. (2024). s&p 500 index and bloomberg u.s. aggregate bond index, 1/1/2022 to 9/30/2022, 10/1/2022 to 12/31/2023. retrieved from bloomberg database july 8, 2024. casto-baker, a., west, s., & wood, a. (2017). asset depletion, chronic financial stress, and mortgage trouble among older female homeowners. the gerontologist, 59(2), 230-241. charupat, n., deaves, r., & lüders, e. (2005). knowledge vs. knowledge perception: implications for financial professionals. journal of personal finance, 4(2), 5061. collins, j. m. (2012). financial advice: a substitute for financial literacy? financial services review, 21, 307-322. cueva, c., iturbe-ormaetxe, i., ponti, g., & tomás, j. (2019). boys will still be boys: gender differences in trading activity are not due to differences in (over)confidence. journal of economic behavior & organization, 160, 100-120. dalbar. (march 30, 2023). investors panic in 2022 and lose more than indices suggest. https://www.dalbar.com/portals/dalbar/c ache/news/pressreleases/qaib23pr.p df deaves, r., lüders, e., & luo, g. y. (2009). an experimental test of the impact of overconfidence and gender on trading activity. review of finance, 13(3), 555575. fan, l., & henager, r. (2022). a structural determinants framework for financial well-being. journal of family and economic issues, 43(2), 415-428. fellner-röhling, g., & krügel, s. (2014). judgmental overconfidence and trading activity. journal of economic behavior & organization, 107(partb), 827-842. ferreria-schenk, s., dickason-koekemoer, z., & shah, n. h. (2021). factors influencing individuals’ short-term investment intentions. international journal of economics and financial issues, 11(4), 73-81. fiksenbaum, l., marjanovic, z., & greenglass, e. (2017). financial threat and individuals’ willingness to change financial behavior. review of behavioral finance, 9(2), 128-147. finra investor education foundation (july 2022). financial capability in the united states: highlights from the finra foundation national capability study. https://www.finrafoundation.org/sites/fin rafoundation/files/nfcs-report-fifthedition-july-2022.pdf fox, j., & bartholomae, s. (2020). household finances, financial planning, and covid-19. financial planning review, 3(4), 1-9. friedline, t., chen, z., & morrow, s. (2021). families’ financial stress and well-being: the importance of the economy and economic environments. journal of family and economic issues, 42(s1) s34-s51. gennaioli, n., shleifer, a., & vishny, r. (2015). money doctors. journal of finance, 70(1), 91-114. glaser, m., & weber, m. (2007). overconfidence and trading volume. geneva risk insurance review, 32, 1-36. grable, j. e., heo, w., & rabbani, a. (2015). financial anxiety, physiological arousal, and planning intention. journal of financial therapy, 5(2), 1-18. grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8(3), 163181. https://www.dalbar.com/portals/dalbar/cache/news/pressreleases/qaib23pr.pdf https://www.dalbar.com/portals/dalbar/cache/news/pressreleases/qaib23pr.pdf https://www.dalbar.com/portals/dalbar/cache/news/pressreleases/qaib23pr.pdf https://www.finrafoundation.org/sites/finrafoundation/files/nfcs-report-fifth-edition-july-2022.pdf https://www.finrafoundation.org/sites/finrafoundation/files/nfcs-report-fifth-edition-july-2022.pdf https://www.finrafoundation.org/sites/finrafoundation/files/nfcs-report-fifth-edition-july-2022.pdf sommer et al. 99 guan, n., guariglia, a., moore, p., xu, f., & aljanabi, h. (2022). financial stress and depression in adults: a systematic review. plos one 17(2), e0264041. hasler, a., lusardi, a., & valdes, o. (2021, april). financial anxiety and stress among u.s. households: new evidence from the national financial capability study and focus groups. finra investor education foundation. https://gflec.org/wpcontent/uploads/2021/04/anxiety-andstress-report-gflec-finrafinal.pdf?x85507 heo, w., grable, j. e., & o’neill, b. (2017). wealth accumulation inequity: does investment risk tolerance and equity ownership drive wealth accumulation? social indicators research, 133(1), 209225. heo, w., moon, k., & ahn, s. y. (2024). unmasking the phantoms of generations: a comparative study of financial responsibility and stress across generational divides. journal of financial therapy, 15(1), 1-25. heo., w., saboe-wounded head, l., cho, s.h., & lee, p. s. (2018). association between personality and financial well-being: strategies for more effective financial education, counseling and advising. consumer interests annual, 64, 1-5. heretick, d. m. l. (2013). clinicians’ reports of the impact of the 2008 financial crisis on mental health clients. journal of social, behavioral, and health sciences, 7(1), 122. huston, s. j. (2012). financial literacy and the cost of borrowing. international journal of consumer studies, 36(5), 566-572. janus henderson investors (2014). taking the long view. https://cdn.janushenderson.com/webdoc s/flyer_takingthelongview_final.pd f kahneman, d. (2011). thinking, fast and slow. farrar, straus, and giroux. kim, h., kim, k. t., & hanna, s. h. (2022). the effect of investment literacy on the likelihood of retail investor margin trading and having a margin call. finance research letters, 45, 102146. kinniry, f., jaconetti, c., dijoseph, m., & zilbering, y. (2014). putting a value on your value: qualifying vanguard advisor’s alpha. https://dta0yqvfnusiq.cloudfront.net/cost e56324383/2019/03/advisor-value5c9513e6b07df.pdf klein, g. (2007, september). performing a project premortem. harvard business review. https://hbr.org/2007/09/performing-aproject-premortem kuzniak, s., rabbani, a., heo, w., ruizmenjivar, j., & grable, j. e. (2015). the grable and lytton risk-tolerance scale: a 15-year retrospective. financial services review, 24(2), 177-192. lazarus, r. s., & folkman, s (1984). stress, appraisal, and coping. springer. lee, y. g., & dustin, l. (2021). explaining financial satisfaction in marriage: the role of financial stress, financial knowledge, and financial behavior. marriage & family review, 57(5), 397421. letkiewicz, j., robinson, c., & domian, d. (2016). behavioral and wealth considerations for seeking professional financial help. financial services review, 25, 105-126. lewis, m. b. (2019). an exploration of overconfidence in the utilization of financial advisors. journal of personal finance, 18(2), 39-49. https://gflec.org/wp-content/uploads/2021/04/anxiety-and-stress-report-gflec-finra-final.pdf?x85507 https://gflec.org/wp-content/uploads/2021/04/anxiety-and-stress-report-gflec-finra-final.pdf?x85507 https://gflec.org/wp-content/uploads/2021/04/anxiety-and-stress-report-gflec-finra-final.pdf?x85507 https://gflec.org/wp-content/uploads/2021/04/anxiety-and-stress-report-gflec-finra-final.pdf?x85507 https://cdn.janushenderson.com/webdocs/flyer_takingthelongview_final.pdf https://cdn.janushenderson.com/webdocs/flyer_takingthelongview_final.pdf https://cdn.janushenderson.com/webdocs/flyer_takingthelongview_final.pdf https://dta0yqvfnusiq.cloudfront.net/coste56324383/2019/03/advisor-value-5c9513e6b07df.pdf https://dta0yqvfnusiq.cloudfront.net/coste56324383/2019/03/advisor-value-5c9513e6b07df.pdf https://dta0yqvfnusiq.cloudfront.net/coste56324383/2019/03/advisor-value-5c9513e6b07df.pdf https://hbr.org/2007/09/performing-a-project-premortem https://hbr.org/2007/09/performing-a-project-premortem financial services review, 33(1) 100 lurtz, m. (2020). how overconfidence can aid entrepreneurs and the challenges of advising those who can’t see the risks they’re taking. kitces. https://www.kitces.com/blog/entreprene urs-risk-taking-confidenceoverestimation-overplacementoverprecision-subjective-probabilityinterval-estimates/ lusardi, a., & mitchell, o. s. (2011). financial literacy and retirement planning in the united states. national bureau of economic research (working paper 17108). https://www.nber.org/papers/w17108 malhotra, n., & witt, m (2010). determinants of financial and retirement planning: a female perspective. journal of business & economic research, 8(6), 41-46. mokhtari, m., & chawla, i. (2023). financial literacy overconfidence, mobile financial service use, and high-cost borrowing. journal of financial counseling and planning, 34(2), 184-203. moon, k., heo, w., lee, j. m., & grable, j. e. (2023). financial stress and covid-19: a comprehensive analysis of the factors associated with the pandemic. risks, 11, 218. moueed, a., & hunjra, a. i. (2020). use anger to guide your stock-market decision making: results from pakistan. cogent economics & finance, 8, 1-29. paisarn, w., chancharat, n., & chancharat, s. (2021). factors influencing retail investors’ trading behavior in the thai stock market. australasian business, accounting & finance journal, 15(2), 26-37. pearson, b., & korankye, t. (2022). the association between financial literacy confidence and financial satisfaction. review of behavioral finance, 15(6), 935-946. peck, j. a. (2020). the disproportionate impact of covid-19 on women relative to men: a conservation of resources perspective. gender, work, & organization, 28(s2), 484-497. piehlmaier, d. m. (2022). overconfidence and the adoption of robo-advice: why overconfidence investors drive the expansion of automated financial advice. financial innovation, 8(14), 1-24. rahman, m., & gan, s. s. (2020). generation y investment decisions: an analysis using behavioural factors. managerial finance, 46(8), 1023-1041. robb, c. a., babiarz, p., woodyard, a., & seay, m. c. (2015). bounded rationality and use of alternative financial services. journal of consumer affairs, 49(2), 407435. simha, a., prasad, r., ahmed, s., & rao, n. p. (2020). effect of gender and clinicalfinancial vulnerability on mental distress due to covid-19. archives of women’s mental health, 23(6), 775-777. skinner, m. a., zautra, a. j., & reich, j. w. (2004). financial stress predictors and the emotional and physical health of chronic pain patients. cognitive therapy and research, 28(5), 695-713. stein, c. h., hoffman, e., bonar, e. e., leith, j. e., abraham, k. m., hamill, a. c., kraus, s. w., gumber, s., & fogo, w. r. (2013). the united states economic crises: young adults’ reports of economic pressures, financial and religious coping and psychological well-being. journal of family and economic issues, 34(2), 200210. streiner, david l. (2005). finding our way: an introduction to path analysis. the canadian journal of psychiatry, 50(2), 115-122. willows, g., & west, d. (2014). differential investment performance in south africa based on gender. international business & economics research journal, 14(1), 221-236. https://www.kitces.com/blog/entrepreneurs-risk-taking-confidence-overestimation-overplacement-overprecision-subjective-probability-interval-estimates/ https://www.kitces.com/blog/entrepreneurs-risk-taking-confidence-overestimation-overplacement-overprecision-subjective-probability-interval-estimates/ https://www.kitces.com/blog/entrepreneurs-risk-taking-confidence-overestimation-overplacement-overprecision-subjective-probability-interval-estimates/ https://www.kitces.com/blog/entrepreneurs-risk-taking-confidence-overestimation-overplacement-overprecision-subjective-probability-interval-estimates/ https://www.kitces.com/blog/entrepreneurs-risk-taking-confidence-overestimation-overplacement-overprecision-subjective-probability-interval-estimates/ https://www.nber.org/papers/w17108 sommer et al. 101 xiao, j. j., & kim, k. t. (2022). the able worry more? debt delinquency, financial capability, and financial stress. journal of family and economic issues, 43(1), 138-152. yeh, t., & ling, y. (2022). confidence in financial literacy, stock market participation, and retirement planning. journal of family and economic issues, 43(1), 169-186. zahirovic-herbert, v., gibler, k. m., & chatterjee, s. (2015). financial literacy, risky mortgages, and delinquency in the u.s. during the financial crises. international journal of housing markets and analysis, 9(2), 164-189. zhang, j., wang, h., wang, l., & liu, s. (2014). is there any overtrading in the stock markets? the moderating role of the big five personality traits and gender in a unilateral trend stock market. plos one, 9(1), e87111. zhang, y., & chatterjee, s. (2023). financial well-being in the united states: the roles of financial literacy and financial stress. sustainability, 15(5), 4505. pii: 1057-0810(94)90017-5 financial services review, 3(2): l-126 copyright q 1994 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. asset allocation, life expectancy and shortfall kwok ho moshe arye milevsky chris robinson an analytical model provides a solution to the retirement problem of how to allocate investment between risky and risk-free assets. the objective is to minimize the probability that the retiree will be unable to consume at the desired level over his/her expected lifetime. the procedure incorporates mortality tables, real or nominal rates of return, initial wealth, and desired consumption levels. numerical examples using standard mortality tables, historic rates of return on canaa’ian equity and treasury bills, and a range of realistic values for wealth and consumption show that equity should play a much bigger role in retirement portfolios than other wn’ters advise. i. introduction how does a person who is retired invest his or her wealth to maximize the probability of a secure and sufficient income? this asset allocation decision is very critical, because the person no longer has the opportunity or time to recover from mistakes with increased earnings from work. as an increasing proportion of the population of the developed nations enters retirement years, this personal finance problem is becoming of particular interest. the retiree faces several issues in making the investment decision: 1. how much annual income does one need to provide the desired standard of living? 2. how long does the money have to last? another way to put this is to ask how the retiree balances lower consumption against running out of money before death. 3. how does the decision incorporate inflation? 4. how should the investment be allocated among the various classes available shares, bonds, etc.’ kwok ho, moshe arye milevsky, and chris robinson, faculty of administrative studies, atkinson college, york university, north york, canada m3j lp3. 110 financial services review, 3(z) 1994 basic financial planning answers the first question and we assume the income figure required is known. the third question involves using either real returns and constant dollars or nominal returns and nominal dollars throughout the analysis. the approach we use in this paper works equally well with either one, although for convenience we use real returns and constant dollars in the numerical examples.* this paper provides an analytic solution to questions two and four, under reasonable assumptions. we incorporate standard mortality tables into the decision to arrive at an expected rate of return needed to finance future consumption, with each year’s consumption weighted by the probability of survival, and given the initial wealth available to generate the income. the approach is perfectly generalizable to any mortality schedule or to any individual’s preferred risk schedule. for example, an individual may decide that he or she wants to be sure of consuming until age 90, and weight each year at 100 percent.3 malkiel(l990) in his chapter on the life cycle guide to investing provides an explicit answer to the allocation question without the same analytic process: as investors age they should start cutting back on the riskier investments and start increasing the proportion of the portfolio committed to bonds. by the age of fifty-five, investors should start thinking about the transition to retirement and moving the portfolio toward income production. . . in retirement, portfolio mainly in a variety of intermediate-term bonds (five to ten years to maturity) and long-term bonds (over ten years to maturity) is recommended. the small proportion of stocks is included to give some income growth to cope with inflation. (pp. 356-7) in the graphs that follow the chapter, he recommends investors in the late sixties and beyond hold 60 percent bonds, 30 percent equity and 10 percent in a money market fund. investors in their mid-fifties are recommended to have 50 percent in stocks, 45 percent in bonds. we compare a numerical example generated in our model with malkiel’s advice. the investment allocation in this paper incorporates the required rate of return to minimize the probability of failing to meet that rate of return on average over the weighted lifespan remaining to the person. this implied utility function of minimizing shortfall is somewhat similar to the approach taken by leibowitz and kogelman (1991). they “measure risk by the “shortfall probability” relative to a minimum return threshold.” a fund manager can choose any combination of minimum return and probability and allocate the assets between a risk and risk-free asset to attain a desirable position. their procedure does not endogenize the time horizon of the investor, since fund managers do not necessarily have a specific time constraint. they do observe that for longer time horizons, the proportion invested in equity rises. many researchers have considered the general question of which investment horizon to use and what effect different horizons have on how we view risk and return. in general, they find that risk declines if assets are held without trading for long periods. different assets perform better in shorter periods of time so the benefits of changing portfolio composition are considerable if the investor times successfully.4 the conclusion for asset allocation is that you should use more equity for longer horizons. lloyd and modani (1983) conclude: in general, the usefulness of time diversification is more evident for portfolios containing common stock. further, the riskiness of any portfolio position is unclear unless the number of time periods the portfolio will be held is also considered. (p. 11) asset allocation, life expectancy and shortfall 111 butler and domian (1993) use a simulation to find that equity is almost certain to be superior to bonds for holding periods exceeding 10 years, and is likely to be better for shorter holding periods. since we are solving the problem for an individual retiree, we incorporate this time dimension explicitly. in addition, we require annual consumption from the portfolio, which does not appear in other researchers’ treatments of this problem. substitu tion of standard canadian mortality tables and reasonable estimates of return and variance for canadian t-bills and equity provides surprising results. only at quite high wealth levels or well into retirement do the portfolios contain less than 100 percent equity. not surpris ingly, 100 percent equity is optimal for women at an older age than men, since women have a longer expected lifespan to finance. this result highlights the contradiction in the obser vation that women are generally seen to invest in less risky portfolios than men do. the unrecognized risk for retirees is the risk of living too long. in the rest of the paper we proceed as follows. the next section formulates and solves the retiree’s asset allocation problem. most of the mathematical details are left to an appendix. the following section provides the numerical results. we then examine the problem when 100 percent equity is insufficient, and provide an heuristic solution to the question of optimal leverage on personal (margin) account. we discuss the implications of our results for retirement planning. finally, we conclude with a brief discussion of possible improvements and extensions. ii. developing a solution formulation of the problem we wish to solve the problem of how a retiree should allocate his/her wealth between a risky and a default risk-free asset. we consider how age, mortality rates (or equivalently, life expectancy of a person at any given age), initial wealth, and the desired level of consumption affect the allocation decision. assume that, at the point of retirement, the individual of n years of age has wealth of w dollars. assume that he has no other source of income so that his current and future consumption is entirely financed from this sum and earnings on it. he will invest w in a portfolio of risky and risk-free assets in order to support the level of desired consumption until death. let c, be this desired annual consumption in nominal dollars5 we do the analysis in before-tax dollars, because the details of tax rules are too difficult to incorporate. let ip, be the probability that the individual aged n will survive one year to age n + 1. for the first year after retirement, the expected consumption is then ,p, . c,. for the second year after retirement, the expected consumption is *p,, . c,. if the mortality table ends at age t, the expected consumption time path after retirement will be 1,p; c,,,p; c,, . . , t_,,pn c,,) . the probabilities and the life expectancy for any given age can be found in standard mortality tables. letting d be one plus the minimum rate of return necessary to support the expected consumption, we have: t-npn ’ ct-n dt-” ’ given the wealth, consumption and life expectancies, there is an unique solution for d, which is the level of return required to avoid disaster. that is, d is one plus the minimum financial services review, 3(2) 1994 rate of return that an individual with initial wealth w must earn to have enough to consume c, per annum, given the average mortality rate. although we will examine this more formally later, we note that the larger the value of n, the lower the value d for a given wand c,. in other words, older individuals may earn less in order to maintain their consumption because they have fewer years to live. this is consistent with the observation that older investors usually invest more in ‘safer’ assets, which provide lower rates of return. our analysis provides an explicit way to determine when they should switch to ‘safer’ assets. if we perform the analysis in real dollars, which is equivalent to assuming that the level of inflation is certain, then c, is a constant, c.6 we can simplify equation (1) for computation purposes to: p p w=ln+2+*. .+ t-npn c d d2 dr_n (2) the solution d is now in real terms. we use constant dollars and real rates of return in our numerical illustrations in a later section for ease of exposition, but the theoretical development is the same. without loss of generality, we assume that there are two assets: treasury bills (t-bills) and a diversified equity portfolio. the individual allocates w between the two. t-bills are free from default risk, but not from interest-rate risk in the long-run. a security is completely risk-free only if it pays off a known and certain amount of consumption at exactly the date required by the investor. an important point to note is that t-bills are risky, in the sense that they have a standard deviation in either real or nominal returns. a person who holds a t-bill until maturity will get exactly the promised rate of return, but if inflation changes during the period, the return is risky in terms of the consumption it permits. empirically, we observe that the time series of real t-bill rates has significant variability. we treat the t-bill rate of return as a random variable, and hence even a portfolio invested 100 percent in t-bills has some risk. the investor must redo the calculations and rebalance the portfolio periodically because the required d changes as one ages. in practical terms, annual rebalancing seems reasonable, since mortality tables report one year age differences. an analytic solution the individual’s problem is to allocate wbetween t-bills and shares so as to minimize the probability of failing to earn the minimum gross rate of return d on average over the remaining years of one’s life. we assume: 1. rates of return on equity and treasury bills, are normally (as opposed to lognor mally) distributed. this assumption is not crucial for optimal results, however it enables us to secure an analytic solution to our problem. 2. rates of return on each asset are serially uncorrelated. thus, we consider a series of decisions in a static framework, without the dynamic consideration of what they will do each year when they come to rebalance their portfolios. 3. returns on each asset are uncorrelated with the other. this assumption can be relaxed, and a solution is given in the appendix. asset allocation, life expectancy and shortfall let us use the following notation and terminology: 113 a is the proportion of w invested in t-bills pln is the average annual one plus rate of return on treasury bills, (or any other relatively safe investment .) of, is the variance of the annual rate of return on treasury bills. pe4 is the average annual one plus rate of return on equity, (or any other relatively risky investment.) cy& is the variance of the annual rate of return on equity. denote by: clp(a)=cltr.a+cl,q.(l--o1) (3) ~~(a)=~~~.az+a$.(l-cl)* (4) which represents the mean and variance of the rate of return (which is normally distributed), of the investor’s portfolio, assuming that he has placed a proportion 01, of his wealth, in treasury bills, and a proportion 1 a in equity. for a given i&, i&, oy,, o& we are looking for an asset allocation proportion a’ that will minimize the probability of earning an annual rate of return that is less than the required rate d. thus, we are trying to solve the following stochastic optimization problem: s.t. o<all (5) note that the one plus rate of return cannot fall under zero. nevertheless, we integrate from -00 for completeness. the solution of this equation for the optimal a* is detailed in the appendix. the result is: (6) this fairly simple result yields quite intuitive properties. we can see under what conditions the retiree picks a comer solution. if d 2 ptr then they must ‘gamble’ by placing 100 percent in equities. if o,, = 0 and d < pl,, then the optimal solution is entirely t-bills, because they will guarantee the minimum required consumption. in between these extremes, higher returns on either asset or lower variances on either asset drives the solution towards investment in that asset, just as we would expect. iii. numerical results q: a 65 year old canadian male has retired with an initial wealth of $850,000. he plans to consume $65,000 per year (in real terms). how much should he allocate, at this time, to treasury bills, and how much should he allocate to equity, to minimize the probability of outliving his money? 114 financial services review, 3(2) 1994 assume that the average real return from treasury bills is two percent per year with a volatility (standard deviation) of 3.5%, and that the average real return on equity is eight percent per year with a volatility (standard deviation) of 17.5 percent. these values are derived from hatch and white (1988) using exponential smoothing to get a continuous function that places somewhat more weight on recent observations. a: using a canadian actuarial mortality table we can find the required rate of return d that the individual must earn per year so as to equate the present value of his consumption requirements to his current wealth, that is, we solve equation (2). thus, the individual’s wealth should be the equivalent of a whole life annuity immediate at the required rate of return d. in this problem d = 1 .ol, thus, the individual must earn at least one percent per year effective. computationally, utr = 1.02 and peq = 1.08 likewise ts,, = 0.035 and rse4 = 0.175. plugging the above values into equation (6) we obtain that a* = 0.78 13 and thus the individual should place 78.13 percent of his current wealth (or $664,100) in treasury bills, and 21.87 percent of his current wealth (or $185,900) in equity. note that even though d < pltr, 0, > 0; so the optimal allocation includes equity. the risk-free rate is only free of default risk. each year the above computation must be done anew, (i.e., the portfolio must be re-balanced once a year) because the individual’s d, one plus the required rate of return, will change as time progresses. to generalize the picture, we calculate a range of results for variations in initial wealth, desired consumption, age, and sex. we combine the mortality rates for females and males at various ages with wealth and consumption in constant dollars to obtain d in real terms. using the same returns and variances as in the example, we obtain table 1. the value of d, one plus the required rate of return, are shown in table 2. table 1 has a block of es in the upper left denoting all equity portfolios, which are preferred whenever the required rate of return equals or exceeds the t-bill rate (we will explain shortly). below them are a few bold-face numbers ranging from 0.171 to 0.849. these are interior optima where the required rate falls between zero and the tbill rate. finally, the lower part of the table has values of a ranging from 0.865 to 0.955. these are portfolios where d < 1 (see table 2). that is, the portfolio need not earn positive returns, but must not lose more than a very small percentage of its value. regardless of how secure the consumption seems to be, the optimal portfolio includes some equity. the extent to which all equity portfolios dominate is quite surprising at first glance. equity is always characterized as the riskiest security, even in a portfolio. in fact, the greatest risk for a retiree is outliving the available wealth, and given a relatively long lifespan, high risk/high return investments are necessary to minimize this risk. thus we see that for a reasonable range of wealth/consumption ratios, an all-equity allocation is preferred into normal retirement years, and is essential for early retirees, even if they have very substantial wealth. numerically, the upper limit of equation (6) is a = 0.96 for the returns and standard deviations in the example. as a practical matter, an a > 0.9 is essentially all t-bills. we can draw more specific observations from table 1: 1. the equity requirement is greater for women than for men. we show only five year intervals, and women should invest in all equity until they are about five years older than men with the same wealth-to-consumption ratios. 2. women with quite low wealth to consumption ratios-seven or less-should invest in all equity as late as 80 years of age. t a b l e 1. o pt im al a ll oc at io n b et w ee n t -b il ls a nd e qu it ie s p a : w om en w ea lth to c on su m pt io n r at io [ 6 7 8 9 10 10 .5 11 11 .5 12 12 .5 13 13 .5 i4 14 .5 1. 5 16 e e e e e e e e e e e e e e e e f q e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e p e e e e e e e e e e e e e e 0. 17 1 0. 71 0 e e e e e e e e e 0. 38 6 0. 69 0 0. 78 0 0. 82 3 0. 84 9 0. 86 6 0. 88 6 e e e e 0. 51 2 0. 76 6 0. 83 3 0. 86 4 0. 88 2 0. 89 4 0. 90 2 0. 90 8 0. 91 3 0. 91 7 0. 92 0 0. 92 4 i k e e 0. 79 5 0. 89 1 0. 91 4 0. 92 0 0. 92 4 0. 92 8 0. 93 0 0. 93 2 0. 93 4 0. 93 5 0. 93 7 0. 93 8 0. 93 9 0. 94 0 0 0. 86 2 0. 91 8 0. 93 2 0. 93 8 0. 94 1 0. 94 2 0. 94 4 0. 94 4 0. 94 5 0. 94 6 0. 94 6 0. 94 7 0. 94 7 0. 94 8 0. 94 8 0. 94 9 0. 94 3 0. 94 7 0. 94 9 0. 95 0 0. 95 1 0. 95 1 0. 95 2 0. 95 2 0. 95 2 0. 95 2 0. 95 3 0. 95 3 0. 95 3 0. 95 3 0. 95 3 0. 95 4 ? b : m en w ea lth fo c on su m pt io n r at io g 6 7 8 9 10 10 .5 11 11 .5 12 12 .5 i3 13 .5 14 14 .5 15 16 e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e e 0. 29 8 0. 62 3 0. 78 4 e e e e e e e e 0. 25 4 0. 67 0 0. 77 4 0. 82 1 0. 84 8 0. 86 5 0. 87 7 0. 89 3 e e e e 0. 65 1 0. 79 5 0. 84 5 0. 87 1 0. 88 6 0. 89 6 0. 90 4 0. 90 9 0. 91 4 0. 91 7 0. 92 0 0. 92 4 e e 0. 76 0 0. 88 3 0. 91 0 0. 91 6 0. 92 1 0. 92 5 0. 92 8 0. 93 0 0. 93 2 0. 93 4 0. 93 5 0. 93 6 0. 93 7 0. 93 9 0. 74 3 0. 90 3 0. 92 4 0. 93 3 0. 93 7 0. 93 9 0. 94 0 0. 94 1 0. 94 2 0. 94 3 0. 94 4 0. 94 4 0. 94 5 0. 94 5 0. 94 6 0. 94 7 0. 93 3 0. 94 0 0. 94 4 0. 94 6 0. 94 8 0. 94 8 0. 94 9 0. 94 9 0. 94 9 0. 95 0 0. 95 0 0. 95 0 0. 95 1 0. 95 1 0. 95 1 0. 95 1 0. 94 9 0. 95 0 0. 95 2 0. 95 2 0. 95 3 0. 95 3 0. 95 3 0. 95 4 0. 95 4 0. 95 4 0. 95 4 0. 95 4 0. 95 4 0. 95 4 0. 95 4 0. 95 5 a ge 50 55 60 65 70 75 80 85 90 a ge 50 55 60 65 70 75 80 85 90 n ot es : t hi s ta bl e pr es en ts t he f ra ct io n of in ve st m en t ca pi ta l at re tir em en t to b e al lo ca te d to t -b ill s (a lp ha f ro m e qu at io n (6 )) . t he se e st im at es u se th e hi st or ic al r et ur ns in h at ch a nd w hi te ( 19 88 ) a nd th e re qu ir ed r et ur ns ( 6) f ro m t ab le 2 . t he v al ue s of a lp ha i n th e ta bl e m us t b e in te rp re te d ca re fu lly . t he ta bl e sh ow s al ph as f or d if fe re nt r et ir em en t ag es a nd le ve ls o f w ea lth r el at iv e to th e co ns um pt io n in c on st an t do lla rs to b e fu nd ed b y th e w ea lth . f or e xa m pl e, a ra tio o f 10 co ul d be $ 40 ,0 00 d es ir ed c on su m pt io n in r ea l t er m s to b e fu nd ed b y $4 00 ,0 00 in s av in gs . t he ‘ e ’ e nt ri es a re a ll eq ui ty p or tf ol io s. t he n on -e v al ue s ar e op tim al p or tf ol io s co nt ai ni ng b ot h t -b ill s an d eq ui ty . t he v al ue s in b ol dfa ce a re p or tf ol io s w he re t he v al ue o f d lie s be tw ee n ze ro a nd th e t -b ill r at e, w he re th e al lo ca tio n de ci si on is p ar tic ul ar ly s ig ni fi ca nt . i f t he re qu ir ed r ea l r at e of re tu rn is n on -p os iti ve , th e pr oc ed ur e pr od uc es a llo ca tio ns ra ng in g fr om a bo ut 8 6 pe rc en t (i f th e ra te i s ze ro ) to 9 6 pe rc en t. h er e th e al lo ca tio n de ci si on i s no t to o si gn if ic an t, al th ou gh s om e eq ui ty i s al w ay s de si re d. t he ri sk o f fa ili ng t o ea rn e no ug h is e qu ite l ow , a nd c ha ng in g to a 1 00 p er ce nt t -b ill p or tf ol io w ou ld n ot r ai se t he r is k si gn if ic an tly . v i t a b l e 2 . r eq ui re d r at es o f r et ur n w ei gh te d by s ur vi va l p ro ba bi lit ie s (f or d if fe re nt w ea lth /c on su m pt io n r at io s) a : w om en w ea lth /c on su m pt io n r ad io 9 10 10 .5 i1 il .5 12 12 .5 13 13 .5 a ze ri 7 8 14 14 .5 15 16 50 1. 15 9 1. 13 4 1. 11 6 1. 10 1 1. 08 8 1. 08 3 1. 07 8 1. 07 4 1. 07 0 1. 06 6 1. 06 2 1. 05 9 1. 05 5 1. 05 2 1. 05 0 1. 04 4 55 1. 15 5 1. 13 0 1. 11 1 1. 09 6 1. 08 3 1. 07 8 1. 07 3 1. 06 8 1. 06 4 1. 06 0 1. 05 6 1. 05 2 1. 04 9 1. 04 6 1. 04 3 1. 03 7 60 1. 14 9 1. 12 3 1. 10 4 1. 08 8 1. 07 5 1. 06 9 1. 06 4 1. 05 9 1. 05 5 1. 05 1 1. 04 7 1. 04 3 1. 04 0 1. 03 6 1. 03 3 1. 02 8 6j 1. 13 9 1. 11 3 1. 09 3 1. 07 7 1 . c6 3 1. 05 7 1. 05 2 1. 04 7 1 . cm 2 1. 03 8 1. 03 4 1. 03 0 1. 02 6 1. 02 3 1. 01 9 1. 01 3 70 1. 12 3 1. 09 7 1. 07 6 1. 05 9 1. 04 5 1. 03 9 1. 03 3 1. 02 8 1. 02 3 1. 01 8 1. 01 4 1. 01 0 1. 00 6 1. 00 3 0. 99 9 0. 99 3 75 1. 09 8 1. 07 1 1. 04 9 1. 03 2 1. 01 7 1. 01 1 1. 00 5 0. 99 9 0. 99 4 0. 99 0 0. 98 5 0. 98 1 0. 97 7 0. 97 3 0. 96 9 0. 96 2 80 1. 06 0 i. 03 1 1. 00 9 0. 99 1 0. 97 6 0. % 9 0. 96 3 0. 95 7 0. 95 2 0. 94 6 0. 94 2 0. 93 7 0. 93 3 0. 92 9 0. 92 5 0. 91 8 85 1. 00 0 0. 97 1 0. 94 8 0. 92 9 0. 91 3 0. 90 6 0. 89 9 0. 89 3 0. 88 7 0. 88 2 0. 87 7 0. 87 2 0. 86 8 0. 86 3 0. 85 9 0. 85 1 90 0. 90 4 0. 87 3 0. 84 9 0. 82 9 0. 81 2 0. 80 4 0. 79 7 0. 79 0 0. 78 4 0. 77 8 0. 77 2 0. 76 7 0. 76 2 0. 75 7 0. 75 3 0. 74 5 b : m en w ea lth /c on su m pt io n r at io a ge 6 7 8 9 10 10 .5 11 11 .5 12 12 .5 13 13 .5 14 14 .5 15 16 5 50 1. 15 3 1. 12 8 1. 10 9 1. 09 3 1. 08 1 1. 07 5 1. 07 0 1. 06 6 1. 06 1 1. 05 7 1. 05 3 1. 05 0 1. 04 7 1. 04 3 1. 04 0 1. 03 5 e 55 1. 14 6 1. 12 0 1. 10 0 1. 08 5 1. 07 2 1. 06 6 i. 06 1 1. 05 6 1. 05 2 l.c .4 7 1. 04 4 1. 04 0 1. 03 6 1. 03 3 1. 03 0 1. 02 4 p 60 1. 13 4 1. 10 8 1. 08 8 1. 07 2 1. 05 9 1. 05 3 1. 04 8 1. 04 3 1. 03 8 1. 03 4 1. 03 0 1. 02 6 1. 02 2 1. 01 9 1. 01 6 1. 01 0 65 1. 11 8 1. 09 1 1. 07 1 1. 05 4 1. 04 1 1. 03 5 1. 02 9 1. 02 4 1. 01 9 1. 01 5 1. 01 0 1. 00 7 1. 00 3 0. 99 9 0. 99 6 0. 99 0 b 70 1. 09 4 1. 06 7 1. 04 6 1. 02 9 1. 01 5 1. 00 9 1. 00 3 0. 99 8 0. 99 3 0. 98 8 0. 98 4 0. 98 0 0. 97 6 0. 97 2 0. 96 9 0. 96 2 5 75 1. 06 0 1. 03 3 1. 01 1 0. 99 4 0. 98 0 0. 97 3 0. 96 7 0. 96 2 0. 95 7 0. 95 2 0. 94 7 0. 94 3 0. 93 9 0. 93 5 0. 93 2 0. 92 5 ij 80 1. 01 2 0. 98 5 0. % 3 0. 94 5 0. 93 1 0. 92 4 0. 91 8 0. 91 2 0. 90 7 0. 90 2 0. 89 7 0. 89 3 0. 88 9 0. 88 5 0. 88 1 0. 87 4 b 85 0. 94 6 0. 91 8 0. 89 6 0. 87 8 0. 86 3 0. 85 6 0. 85 0 0. 84 4 0. 83 9 0. 83 4 0. 82 9 0. 82 4 0. 82 0 0. 81 6 0. 81 2 0. 80 5 90 0. 85 1 0. 82 3 0. 79 9 0. 78 0 0. 76 4 0. 75 7 0. 75 0 0. 74 4 0. 73 8 0. 73 3 0. 72 7 0. 72 2 0. 71 8 0. 71 3 0. 70 9 0. 70 1 e n ot es : t h es e va lu es a re 1 + r at e of r et u rn = d b as ed u po n s ta n da rd c an ad ia n m or ta li ty ta bl es . e qu at io n (2 ) is s ol ve d fo r th e ra te o f in te re st th at eq u at es a c on st an t d ol la r $ va lu e fo r co n su m pt io n , w ei gh te d by p ro ba bi li ty o f li vi n g to t h e en d of e ac h y ea r, w it h t h e cu rr en t i n ve st ab le w ea lt h o f th e in di vi du al . s in ce t h e co n su m pt io n is a $ co n st an t ( se e e qu at io n 2 ). w ea lt h a n d co n su m pt io n ca n b e su m m ar iz ed in a s in gl e ra ti o. f or e xa m pl e, a n in di vi du al w it h s 40 0, o o o to in ve st w h o w is h es to c on su m e “w $4 0$ 00 p a in c on st an t d ol la rs h as a w ea lt h /c on su m pt io n ra ti o of 1 0. t h is y ie ld s th e sa m e d as i f on e h ad $ 60 0, 00 0 to i n ve st a n d w is h ed t o co n su m e $6 0, 00 0 p. a. 8 asset allocatim, life expectancy and shortfall 117 3. virtually all women should invest in all equity at age 65 or earlier. 4. men with quite low wealth should invest in all equity as late as 75 years of age. 5. virtually all men should invest in all equity at age 60 or earlier. the specific values of alphaderived from this procedure must be interpreted with some caution, which is why we have shown ‘e’ instead of the specific values. the definition of the problem requires that 0 s a i 1. this is the same as saying that the required d cannot exceed one plus the treasury bill rate. as soon as it does, we would want no treasury bills in the portfolio. the intuition is that you cannot minimize the probability of falling below a rate of return by including in the portfolio any asset which is expected to earn less than that rate of return. including a high risk, high return asset like equity may yield a greater loss on some occasions, but the probability of earning more than the required minimum is still higher. given enough years of returns, the long-run return will converge to the expected return. since so many people are in a position where they need more return to minimize shortfall risk than 100 percent equity will provide, we model borrowing in the next section. iv. optimal margin position as long as the borrowing rate is less than the return on equity, borrowing to buy more equity provides a higher rate of return than 100 percent equity, but it is also more risky. persons normally borrow on margin or demand loans, which charge floating rate interest. therefore, although the equity returns will fluctuate in real terms, the interest expense is essentially fixed in real terms. the investor is faced with the annual (one plus) rate of interest charged on margin loans denoted by r, together with the previously-mentioned l_~,,~, <t,~. the choice variable is q which represents the proportion of wealth that the investor should borrow on margin. that is, q = 0.5 means that the investor should borrow 50 percent of his current liquid wealth wc and invest the proceeds in the equity market. note that we are assuming: rb < d the interest rate on margin is less than the minimal required rate of return and is less than the expected rate of return on equity. note that no assumptions are being made about the relationship between the variables peq and d; hence d can indeed be larger than ue4. we provide the mathematical solution in appendix 2. it turns out that there is no global minimum probability. intuitively, if the equity rate exceeds the borrowing rate, you can always get a little lower probability of failing to consume a given level of income by borrowing more and investing it. the objective function does have a well-defined asymptotic value, which can be obtained by taking the limit as q + 00. in other words, the asymptotic minimum is the probability of failing to earn the required rate of return as the individual’s margin approaches infinity. we derive a reasonable and practical heuristic solution to calculate the margin, q“, that is required for the individual to get ‘very close’ to the asymptotic minimum. we denote this distance to the asymptotic minimum by e, which can be any arbitrarily-small fraction. 118 financial services review, 3(2) 1994 the probability approaches the asymptotic minimum fairly quickly, and so values of the order of one to five percent give reasonable answers. what follows is a table of values of q” for various values of d with an e = 0.02 under the assumption that one plus the margin interest rate is rb = 1.03 while one plus the expected rate of return on equity is peq = 1.08 and the standard deviation is cre4 = 0.175. these values are again all in real terms. d 1.040 1.045 1.050 1.055 1.060 1.070 1.080 1.100 4” 0.1019 0.6527 1.2037 1.7547 2.3056 3.4071 4.5094 6.7132 this illustrates the fact that once the required rate of return on investors’ wealth d exceeds the margin interest rate rb, by more than a little, the actual ‘optimal’ margin rate q” becomes unreasonably high, and on the practical level a lottery ticket. v. retirement planning implications the most important implication is that retirees should be holding a lot of high risk, high return assets in their portfolios at much later ages than most financial planners and authors have recommended. equity and treasury bills are merely representative of classes of assets and the implications extend to all situations. there are more portfolios with 100 percent equity for women than men, because they have longer life expectancy. we have not seen any personal finance source advising women that equity is even more important for them than for menthe implicit assumption is that the allocation should be the same for both sexes. if the unsystematic comments of investment advisers are valid, it appears that women are less likely than men to buy shares. the evidence in this paper shows that this tendency is riskier, not safer, since women thus have an even higher probability of outliving their capital. suppose we consider malkiel’s advice, as quoted earlier, to move into long-term bonds as retirement nears. let us assume that long-term bonds are expected to yield a five percent real rate.7 as long as obbonds < oeq, the standard deviation is irrelevant in determining whether or not equity dominates. for every case in which l&,,& _ < d, the probability of not having enough to consume is minimized (as much as is feasible) by an all-equity portfolio. inspection of table 2 reveals that a five percent required rate would turn 15 of the all-equity cases into internal optima with a mixture of bonds and equity. a greater number of all-equity cases remain, and they can occur as late as age 75 for men and age 80 for women. thus, malkiel’s advice is not appropriate for all retirees. finally, borrowing does not seem to be a good plan to cover shortfall in most cases. a person who is short a lot must borrow an unreasonably large amount to reach desired consumption. someone close to the line should be able to reduce current consumption, rethink planned future consumption, and/or defer retirement date. asset allocation, life expectancy and shortfall 119 vi. conclusion we have developed a procedure for a retiree to estimate how to allocate his/her investable wealth between any two classes of investment assets. this procedure combines mortality tables, desired consumption, initial investable wealth, and the age of the retiree in a single analytic result. numerical examples using realistic values show that if the risky asset is equity and the safe asset treasury bills, the allocation favors all equity in a large number of cases, even well into retirement years. a similar result will hold for long-term bonds. the procedure is easily adaptable to any mortality schedule, initial wealth, desired consumption, and set of returns and standard deviations. the procedure is generalizable to situations where the individual’s initial wealth includes one or more pension plans, sponsored by government, employer or self. defined contribution plans simply provide a part of the initial wealth to be allocated. the pension provided by the defined benefit plan can be deducted from the desired consumption to arrive at a shortfall between retirement income and consumption. this shortfall is treated as the consumption value to be financed by the non-pension investments.’ the procedure described in the previous paragraph is technically correct. our model uses minimization of shortfall instead of a particular form of utility function, as is commonly done elsewhere. this increases the generalizability of the results, since a person can choose various levels of desired consumption retirement and see what the investment implications are. however, the following problem arises. two persons faced with the same shortfall may assess it very differently if one depends on investment for all their income, and the other depends on a large part from, say, an employer pension plan. the person with the separate employer plan is much more secure, and would presumably have a different utility for the extra consumption than would someone for whom missing it is a serious risk. we make that tradeoff in order to maintain a paper that provides practical solutions, since derivation of specific utility functions is quite problematic. the procedure applies to a wide variety of problems, but there are some improvements which would make it more valid. in principle, it should handle the problem of a retirement couple as well, but we have not analyzed the practical difficulties. all the changes occur in the determination of d. since there is now a joint probability of either one or two persons consuming an amount which is dependent on who survives how long, the mechanics would be much more complex. once a d is estimated the procedure continues as for a single person. since the probability is good that one (unspecified) member of the couple will live to a substantial age, the requirement for a high return portfolio increases. the derivation assumes no serial correlation of returns. if year-to-year returns are serially correlated, the minimization of the probability of not earning enough over the entire lifetime is a much more complicated procedure which may or may not prove tractable mathematically. serially-uncorrelated returns allow minimization of the product of the probabilities, and a closed-form solution. we do not consider the dynamic behavior of the retiree faced with this problem each year. every year the retiree can rebalance the portfolio or change consumption. they face a new table 1, and can choose to adjust accordingly. for example, a bad year may induce lower consumption, and this affects all future probabilities of failing to earn enough. 120 financial services review, 3(2) 1994 although we do not give any details, in the appendix we do present the analytical solution to the problem if the returns on the two assets are correlated (see equation (25)). we do not provide numerical examples. appendix 1 solving for the optimal asset allocation the solution to equation (5) is the optimal proportion that is to be invested in treasury bills, and should be expressed as the following function: (7) to simplify the notation: l denote the probability density function (p.d.f) of the standard normal distribution, (with mean zero and variance one.) by: l denote the cumulative density function (c.d.f) of the standard normal distribution’ by: u o(u) = j cp(y) dy (9) let us make the following substitution in the integral of equation (5) so as to simplify the integrand. x l&(o) (10) y= o,(a) the above implies that: dy=---ldx q(a) (11) thus we can substitute and transform the stochastic optimization problem in equation (5) to the elementary representation: d-cl,(a) o,(a) (12) s.t. 0 i a 5 1 asset allocation, lge expectancy and shortfall 121 which can be expressed, using equation (8), as: min b(a) = j q(y) dy (13) s.t. 0 i a s 1 however, using the notation introduced in equation (9), we can rewrite the above as: min b(a)sq xt. 0 i a i 1 (14) in order to find an a* that will minimize the above, we must take the derivative of the objective function, b(a), with respect to a and set it equal to zero. that is, the first order condition (f.o.c.) is: ?!!m=, (1% aa or we can write explicitly: (16) taking derivatives with respect to a by using the chain rule, we derive that the first order condition for optimality can be expressed as: -i.$ (a) . q,(a) 0,’ (a) . (d (17) &a) the objective now is to find an a* that will solve the above equation. the next step is to realize that the first term in equation (17), namely a’, is the derivative of the cumulative distribution function. thus, using the fundamental theorem of calculus and the definition in equation (9), we can rewrite the first order condition as: (18) let us analyze equation (18). it should be obvious that cp, which is a probability density function, will never take on the value zero. thus, the only way that the left hand side of equation (18) can evaluate to zero, is by the following being satisfied: p; (a) . q(a) 0; (a) . (d y(a)) = 0 (19) in order to solve the above equation for a (or isolate a), we must compute &(a) and $(a) by taking the derivatives of equations (3) and (4) to obtain: 122 financial services review, 3(2) 1994 (21) substituting equations (3) and (4) together with equations (20) and (21) into the first order condition of equation (19), we can rewrite equation (19) as: x(~-ptr.a+lq.u-an (22) solving for a (with some tedious algebra and the help of the symbolic computational language maple v), we find that there is only one value of c1 that will satisfy equation (22) and it is equation (6): checking second order conditions (s.o.c.) one can verify that the second derivative of b(a) with respect to a is positive at a*. (based on intuition, one could assert without resorting to s.o.c. that a* must be a minimum because of diversification arguments.) hence, a* is a unique (global) minimum, and we have obtained a closed form solution for the proportion a* that must be allocated to treasury bills so as to minimize the probability of earning a rate of return that is less than the required amount. let us perform some comparative statics by inspecting equation (6), we can recognize that ceteris paribus as a; decreases a* will increase. intuitively this means that as treasury bills become less risky one should allocate a higher proportion of his wealth to treasury bills, and vice versa if the riskiness of treasury bills were to increase. mathematically this can be represented as: likewise, ceterisparibus as g& increases, a* will decrease. as equity becomes riskier one should allocate more to treasury bills. mathematically this can be expressed as: the same kind of argument will show that as ulfr increases, a* increases; and as l&q increases, .a* will decrease. mathematically: and: asset allocation, life expectancy and shortfall 123 let us rewrite equation (6) as: (23) we see by inspection that as d increases, a* decreases. hence, as the required rate of return goes up, the individual must ‘gamble’ by investing more in equity. mathematically: aa* -so it is important to note that by the definition of the problem a is restricted to be between zero and one. on the other hand, there exist realistic situations where plugging the numbers into equation (6) will produce an a* that is negative or larger than one. in the event that should happen, most probably the required rate of return is too high; that is, there is no way to minimize the probability of default because b(a) is a monotonic function that has no interior minimum. thus, one should choose the boundary value closest to the a*. the above criteria will probably not be appropriate for individuals with either very high or very low d’s. finally, the problem can be solved analytically for the case in which the returns on the two assets are correlated. letting p be the correlation between the two assets, equation (4) becomes: o~(a)=o:,.a2+o~~~(1-a)2+2a(l-a).p.a,,.o,. proceeding to solve it in a similar fashion yields the following: (24) appendix 2 determining the optimal margin the (levered) investor’s wealth at the end of one year will be a random variable, if one starts with wa to invest: w,x(l+q)x~-wwoxqxrb, (26) where x is the random variable, which represents the (one plus) rate of return from equity with mean ple4 and standard deviation creq. as previously-defined, q is the margin, expressed as a fraction of the investor’s total assets and r, is the borrowing rate. we create a new random variable “r such that: “y=(l+q)xx-qxrb, (27) 124 financial services review, 3(2) 1994 which represents the random levered (one plus) rate of return from the margined position. e[“yi = (1 + 9) x pe4 4 x rb; (28) likewise s.d.[yj = (1 + q) x oe4. (29) this represents the higher expected return, as well as higher risk, in the margined position. in addition “y is normally distributed by simple additivity. as before, we seek the optimal q* that will minimize the probability of earning a (one plus) rate of return that is under d. mathematically: (30) proceeding with the derivation and conditions as in appendix 1, we are seeking to minimize the function: min q, d (i+ 4) x peq + 4 x rb drb (lb4 .)i (31) 4 (1+4x%, u+q)xgq %q . setting up the first order conditions for optimality implies finding a q* such that: q’ drb _ (kb rb) (l+q’)xo, %q 1 = o ’ (32) which is equivalent to finding a q* that satisfies: however, there is no positive value of q* that will satisfy the above first order condition. in fact, by inspection, one can ascertain thatfir all values of q the derivative is negative, this in turn implies that the objective function is decreasing in q. but all is not lost. the objective function has a well defined asymptotic value which can be obtained by taking the limit as q + 00. the asymptotic value of the objective function will be a _ (& rb) [ 1 0 eq (34) so in fact, the probability of earning a rate of return that is under d, by buying on margin, will approach a limit as the amount of margin goes out to infinity. it is important to realize, that the above mentioned probability is not a probability of starvation (or failing to meet the consumption requirement). it is the probability of earning, in one year, a rate of return that is under the required level. if indeed one earned an above average return in the following year then one could recuperate the losses and meet all future consumption requirements. the overall probability of starvation is the probability of repeat asset allocation, life expectancy and shortfall 12s edly earning a rate of return under the required rate, over long periods of time, so as to make the required consumption infeasible. this probability is not addressed in the paper. let us continue with our margin discussion. in theory the asymptotic value means that there is no optimal margin per se. yet, we can find a q” that will bring us within e of the minimal probability. this q” will satisfy q, drb (1 + 47 x %q acknowledgments the authors thank gilles bernier, david brophy, dale domian, myron gordon, sherman hanna, barbara poole and tom salisbury for helpful comments. address all communica tions to chris robinson, faculty of administrative studies, york university, north york, canada m3j lp3 (email: as0oo037@orion.yorku.ca) notes 1. we assume throughout that investment in any one asset class is approximately diversified, perhaps through a mutual fund. 2. gitman and joehnk (1990, exhibit 15.1, p. 541) and turner, le rossignol, rinfret, and daw (1991, form 27, p. 115) make the mistake of capitalizing a constant dollar value using a nominal discount rate,which yields too low an estimate of required savings. this error seems common among financial planners. 3. financial planners generally seem to assume that a perpetual dollar return equal to consumption is required, explicitly ignoring the possibility of consuming capital over time. see previous note for two examples. this error overstates the savings required, and thus the two errors offset to some extent. 4. see, for example, benari (1990), butler and domian (1991), grauer and hakans son (1982) and lloyd and modani (1983). 5. in retirement planning, the usual approach is to have the retiree define the lifestyle they want and then price each expenditure of that lifestyle. the total expenditure required to achieve the desired lifestyle in each year is c,. the value of c, can range from bare subsistence to great luxury, provided it is feasible. 6. if inflation is certain, then we can convert real to nominal returns using the fisher equation without affecting the analysis. we do not consider uncertain inflation. 7. the long-run historical experience in canada is actually a rate lower than the t-bill rate, but this is surely not a good basis for an expected rate. 8. if the pension is indexed, then everything can be treated in real terms, as in equation (2). if the pension is unindexed, then it pays in nominal dollars, and equation (1) is appropriate. 9. that is, the probability that a standard normal random variable will take on a value less than or equal to u. graphically, it is the area under the p.d.f. from u to negative infinity. 126 financial services review, 3(2) 1994 references benari, y. (1990). optimal asset mix and its link to changing fundamental factors. the journal of portfolio management (winter), 1 l-l 8. butler, k., & domian, d. (1993). long-run returns on stock and bond portfolios: implica tions for retirement planning. financial services review, z(l), 41-49. butler, k., & domian, d. (1991). risk, diversification, and the investment horizon. the journal of portfolio management (spring), 41-47. gitman, l., & joehnk, m. (1990). personal financial planning (5th ed.). new york: dryden press. grauer, r., & hakansson, n. (1982). higher return, lower risk: historical returns on long-run, actively-managed portfolios of stocks, bonds and bills, 1936-1978. finan cial analysis journal (march-april), 39-53. hatch, j., & white, r. (1988). canadian stocks, bonds, and inflation, 1950-87. research foundation of the institute of chartered financial analysts. leibowitz, m., & kogelman, 5. (1991). asset allocation under shortfall constraints. the journal of portfolio management (winter), 18-23. life tables, canada, andprovinces 198.5-871canl cs8.5 82-0038, no. 131. health reports supplement, vol. 2, no. 4. ottawa, 1990: statistics canada. lloyd, w., & modani, n. stocks, bonds, bills and time diversification. the journal of portfolio management (spring), 7-l 1. malkiel, b. (1990). a random walk down wall street. new york: w. w. norton and company. turner, m., le rossignol, d., rinfret, c., & daw, r. (1991). canadian guide to personal financial management (6th ed.). toronto: prentice hall canada. financial services review, 32(4) i volume 32 issue 4 from the editor the first paper in this issue of financial services review begins with an invited manuscript that discusses the multiple ways artificial intelligence (ai) is impacting the financial planning profession. written by drs. manuela faulhaber and charles chaffin, artificial intelligence in accounting, medicine, and law with potential implications for financial planning shows that financial planners and financial planning regulators have much to learn from the profession's affiliated fields. one reason that i asked drs. faulhaber and chaffin for their paper is that i see parallels between the profession of financial planning and the world of academic publishing. consider how chatgpt considers the role of ai in academic publishing (prompt asked september 27, 2024): ai is transforming academic publishing by streamlining various stages of the research and publication process. ai-powered tools help researchers with literature reviews, data analysis, and manuscript drafting, significantly reducing the time and effort required to produce high-quality papers. in publishing, ai is used for automated peer review, plagiarism detection, and manuscript evaluation, improving the review process's speed and accuracy. moreover, ai enables better discoverability of papers by generating precise recommendations for readers based on their interests. while generally positive, ai's use prompts ethical use concerns and worries about bias, integrity, and fairness. the faulhaber and chaffin paper provides a foundation on which academy of financial services members can begin to consider how to utilize ai in the future. the second paper in this issue (should investors defer long-term gains in taxable portfolios?) was authored by dr. jeff whitworth. dr. whitworth points out that investors sometimes postpone the sale of appreciated stock to defer capital gains taxes. he then shows that while useful in reducing taxes, this strategy can cause a portfolio to become overly concentrated, which can increase volatility while reducing long-term returns. this paper adds to the literature by examining the tradeoff between tax efficiency and diversification. in the end, via monte carlo simulations, dr. whitworth documents that diversification is more important for increasing terminal wealth. dr. whitworth goes on to suggest that investors are generally better off rebalancing almost completely each year, even though this may require selling winners and paying capital gains taxes. the next paper in this issue of fsr (racial/ethnic disparities in financial advice seeking: a decomposition analysis) was co-authored by di qing and miranda reiter. their work adds to the growing body of articles published in fsr and other journals which show that seeking financial advice and working with a financial planner differs by race and ethnicity. drs. qing and reiter document that black and white consumers are more likely than asian and hispanic consumers to use financial planners for saving and investment decisions, with significant differences existing between white and hispanic consumers and narrower differences between black and hispanic consumers. their work is important in showing that risk tolerance, financial knowledge, and income are important determinants that help explain racial/ethnic differences in financial planner use. i hope that the leaders of certification boards, policymakers, and regulators take to heart what drs. qing and reiter discuss in this paper. financial services review, 32(4) ii in the fourth paper, drs. jia qi, yu zhang, and sheri worthy (determinants of health and desirable financial behaviors: the meditation effect of financial knowledge) provide evidence that the social determinants of health (sdh) (e.g., a person's environment, including economic stability, educational access and quality, health care access and quality, neighborhood and built environment, and social and community context) are related to desirable financial behaviors. in their models, they used financial knowledge as a mediation variable. this research team is among the first to show that financial knowledge plays a significant role in mediating the relationship between sdh and desirable financial behaviors. for those who are not familiar with the concept of sdh, this paper can serve as an easily accessible introduction to the concept. this issue of fsr concludes with a paper that explores an important topic that is embedded in the practice of financial planning: phone call phobia. in their paper (improving communication with financial consumers: insights from a study of phone call phobia), drs. wookjae heo, yi liu, and jae min lee describe what many financial planners encounter on a daily basis—client phone call phobia, a symptom of avoiding real-time communication that ranges from mild nervousness to a debilitating fear of making or receiving phone calls. drs. heo, liu, and lee examine factors related to three types of phone call phobia (i.e., employer, family, and friends) and six types of communication preferences (i.e., face-to-face, phone calls, letters, email, text messaging, and online messaging apps). what they found is profound. financial characteristics (i.e., financial status, job insecurity, and financial-psychological factors) were observed to be the primary factors contributing to higher levels of phone call phobia. financial planners and others who believe their clients may be exhibiting signs of phone call phobia would be well advised to read this paper. let's shift to two other topics. the first is the academy of financial services (afs) conference. the 2024 conference was held in columbus, ohio. the event was amazing. the quality of the research presentations and the overall level of camaraderie were unbelievably good. if you have not attended an afs conference, or if it has been a while since you've gone, i strongly encourage you to consider joining afs in 2025. there will be opportunities for dual submissions to fsr and the conference. go to the afs website to learn more (https://www.academyfinancial.org/). the second topic is one of some sadness (on my part). the next issue of fsr will be my last as editor. given work, family obligations, and health issues, i have decided to step aside. fortunately, dr. stephen horan has agreed to take the role of editor. this will allow opportunities for financial services review and financial planning review to synergistically align to enhance fsr's publication and outreach efforts. more information about the future of fsr will be provided in the next couple of issues. in the meantime, i encourage you to submit your work to fsr. until next time, all the best, john e. grable, ph.d., cfp® editor financial services review, 33(1) i volume 33 issue 1 from the editor i have some exciting updates to share about financial services review (fsr). first, as you can see from the number of papers in this issue, the journal is in a strong position and growing. we've seen a notable increase in both the quantity and quality of submissions, and our acceptance rate continues to decline, reflecting rising reviewer expectations that align with the goal of enhancing the journal's impact. another development is that the editorial board and i are moving forward with an application for possible scopus indexing. while there are no guarantees, i do believe that fsr is in a strong position for consideration. as an example, papers from fsr have been cited nearly 1,200 times in scopus journals. the journal’s impact is expansive. the final update is that this issue includes some outstanding papers. let’s quickly review these papers. the first paper (the impact of the covid-19 income shock on debt management: a mediation analysis) was co-authored by drs. ouyang congrong, thomas crandall, and swarn chatterjee. this study examines the relationship between large, unexpected income drops during the covid-19 pandemic and individuals' use of stimulus checks to pay off debt. this research team found that those with significant income reductions were more likely to use their checks for debt management. the paper also highlights the role of perceived financial control in the decision-making process, showing that individuals who felt a lack of control over their finances were more likely to use stimulus funds for debt relief. the next paper (retirement expectations vs. reality: if covid-19 did not impact retirement expectations significantly, what did?) was written by drs. zhikun liu, david blanchett, qi sun, and naomi fink. this study analyzes the retirement expectations and actual retirement ages of older americans, finding that while many expect to delay retirement, the covid-19 pandemic had no significant impact on this trend. they did find that key factors influencing retirement decisions include health, wealth, marital status, education, and disability status. their paper underscores the importance of understanding how these factors can help explain the gap between expected and actual retirement expectations. drs. emily koochel, megan mccoy, and sonya lutter wrote the third paper in this issue (protecting well-being through financial shocks). this study explores how resources, including personal qualities like self-control and perceived health, can buffer the effects of financial stress and improve financial well-being. they found that beyond income, factors such as access to resources and a strong sense of control play a critical role in enhancing resilience to financial challenges. dr. megan mccoy and her colleagues, drs. ives machiz, portia johnson, and kenneth white, contributed the next paper (resilient personality or financial resilience framework for coping with physical and mental health during the covid-19 pandemic). this study examines how two resilience frameworks—the financial resilience framework and the resilient personality model—were related to physical and mental health outcomes during the covid-19 pandemic. their analysis indicates that while both frameworks contribute to resilience, financial resilience is a stronger predictor of positive health outcomes than a resilient personality. the fifth paper (an investigation of the relationship between gender and investor behavior during a market correction) was written by drs. matthew sommer, megan mccoy, and hanna lim. this study investigates how u.s. retail investors coped during a recent declining stock market and a period of rising inflation. they found that men were more likely to shift investments from stocks and bonds to cash, while women experienced more financial stress. their work shows that financial stress and financial services review, 33(1) ii investment overconfidence are linked to shifts in trading behavior, although the indirect effects are minor and only partially mediated by gender. the next paper (retail investors and investment fraud victims: is there a connection?) was contributed by christopher rand, melisande mccrae, and jason martin. this study examines the characteristics of investment fraud victims, finding that both overconfident and financially literate investors are more likely to be victimized (although they differ in their investment behaviors). overconfident investors were observed to be more comfortable with risk and use background checks on investment professionals, while financially literate investors engage in more frequent trading and stock allocation, with males and younger investors being more likely to experience fraud. drs. yu zhang, khurram naveed, and jia qi wrote the next paper in this issue (crypto investment: the role of investment motivations, investment confidence, and risk perceptions). this study examines the psychological and behavioral factors influencing cryptocurrency investment, finding that motivations and investment confidence positively impact current and future investment decisions, while risk perceptions act as a barrier. their findings suggest that understanding these psychological factors can improve investor profiling and enable financial advisors to offer more personalized guidance in the volatile cryptocurrency market. in the next paper (assessing the impact of rebalancing on equal-weighted and value-weighted portfolios over five decades), drs. rama malladi and alexander stanoyevitch discuss how transaction costs affect the performance gap between equal-weighted portfolios (ewps) and valueweighted portfolios (vwps). they found that ewps generally outperform vwps unless transaction costs become excessive. the results highlight that while ewps offer significant diversification benefits, transaction costs play a critical role in diminishing their outperformance. dr. eun jin kwak and i wrote the second-to-last paper in this issue (a domain specific measure of investment risk preference). this paper proposes a new measure of risk preference that combines aspects of revealed-preference measures and propensity assessment techniques. the resulting test adds to the literature by advancing the precision of investment risk assessment tools. this issue of fsr concludes with a paper (does overspending harm retirement preparation?) written by drs. christina lynn, stuart heckman, michael kothakota, and derek lawson. this study explores the relationship between overspending and retirement preparation, testing the behavioral life cycle hypothesis using three data sets. findings from their study reveal that overspending is only weakly linked to poor retirement preparation, suggesting that overspending behavior may only sometimes reflect one's retirement readiness. dr. lynn et al. provide detailed policy solutions to address overspending and improve financial awareness. i would like to end by giving you an update about my role as editor. i announced in the last issue that volume 33 issue 1 would be my last as editor. i made that announcement prematurely. i’ve been asked to stay on through 2025 or until the academy of financial services (afs) can find a new permanent editor. as i told the afs board, i am totally committed to ensuring that fsr continues to remain impactful. in this regard, i am officially announcing that i will serve as interim editor for the foreseeable future. i am so grateful for the guidance provided by those serving on the advisory and editorial boards. the individuals listed on the masthead (and online) are amazing professionals. finally, thank you for being an active participant in fsr. each one of you, as a contributor or reader, has helped shape the content and quality of the financial services review. until next time, all the best, john e. grable, ph.d., cfp® pii: s1057-0810(96)90025-6 financial services review, 5( 1): 3 142 copyright 0 1996 by jai ress inc. issn: 1057-0810 all rights of reproduction in any fom reserved. computing yields on enhanced cds robert brooks in this paper, we seek to provide a framework for comparing certificates of deposit (cd) products that vary in their features. there are now fixed-rate cds with no early withdrawal penalties as well as floating-rate cds with guaranteed floors. with the model developed here, we examine the required change in the effective annual rate required in basis points to make cd products with embedded derivatives (called enhanced cds) comparable with the standard cdproducts (ones with large early with drawal penalties). thisfiamework is beneficial for both retail customers seeking to make rational comparisons and bank executives seeking to provide optimal liability products and seeking to manage the resulting interest rate risk i. introduction the purpose of this paper is to provide a general framework for comparing cd products with emhedded options. the objective here is to move toward a methodology that would comply with the general intent of the truth in savings act of 1991. section 262 (h) states that the purpose of this act is: . . . to require the clear and unifom~ disclosure of (1) the rates of interest which are payable on deposit accounts by depository institu tions; and (2) the fees that are assessable against deposit accounts, so that consumers can make a meaningful comparison between competing claims (emphasis added) of depository institutions with regard to deposit accounts. thus, we seek to provide a mechanism for comparing enhanced cd products. specif ically, we build upon the “annual percentage yield” (apy) legal concept contained in the truth in savings act. our results are presented, however, in “effective ammal rate” (ear) form which is more appropriate. that is, we demonstrate how to compute the apy and ear for enhanced cds. this framework will he useful for both depositors and bankers. depositors will find this useful for comparing cd products, and bankers will be able to use the framework to make more informed decisions on which products to offer. robert brooks l the university of alabama, 200 alston hall, tuscaloosa, al 35487; e-mail: rbrcoks@slston.cba.ua.edu 32 financial services review 5( 1) 1996 retail banking is undergoing dramatic change. there is the mentality that cds are a rather mundane source of bank funding. this mentality is best captured by the following quote from chet currier (1987, p. 39): no matter how it is structured, a cd is fundamentally a simple proposition-a loan of money by the buyer to a bank or savings institution for a specified period of time, in return for interest on that money. this view of cds has cost the banks market share. however, recent changes have somewhat revived the cd market. in an article entitled, “yields on cds lag behind rate increase,” which appeared in the wall street journal (1996, pp. cl, cls), lynn asinof reported: indeed, people today have some $925 billion in small-denomination cds, according to federal reserve board figures. although that is far less than in 1991, when cd assets topped out at over $1.1 trillion, it is up significantly from 1994, when assets dropped well below $800 billion. this wall sfreet journal article goes on to illustrate how banks are attempting to revive this market: at the same time, some banks are promoting cd deals aimed at encouraging the cre ation of additional banking relationships or attracting larger deposits.... hibemia national bank, new orleans, offers a two-year cd with the one-time option of bump ing the rate up if interest rates rise. (p. c15) in this paper, we provide a methodology for comparing hibemia national bank’s offer to other two-year cd products. with deregulation of the banking industry coupled with historically low interest rates in the early 199os, retail banks began offering a variety of incentives to keep funds as well as attract new funds. one of the more esoteric offerings is the world soccer cd that would double the interest rate if the united states won the world soccer cup (peterson bank, 1994). the retail bank, peterson bank in chicago, actually purchased an insurance policy against the risk. admittedly, the main purpose was to gain publicity-which it did, receiv ing notice on the front pages of the wall street journal and usa today. there are health care cds that allow money to be withdrawn with no early withdrawal penalty if funds are used for health-care expenses (o’connell, 1994) and even a “hunter’s special” that pays a weatherby mark v deluxe rifle instead of cash interest (far&am, 1994). the primary motivation for these products is marketing. however, many retail banks are now routinely offering cds with features such as no early withdrawal penalty, the abil ity to add to the original deposit, and a variable rate with a guaranteed floor. for example, the bank leumi trust’s triple choice time deposit cd is a two-year cd that allows depositors to add funds in the first year, make one withdrawal (up to 50%), and obtain an upward rate reset once during the term of the cd at the depositor’s discretion (o’connell, 1994). southtrust bank offers a two-year cd that has a guaranteed floor rate, which floats with the 13-week treasury bill, allows one early withdrawal at no penalty, and permits depositors to add to the cd at any time. this product was advertised as the “maximum advantage two-year cd account.” certificate of deposits are an important source of bank funding. a survey of small banks in the southeast united states revealed that cds represented an average of 45% of computing yields on enhanced cds 33 the total assets and ranged from a low of 32% to a high of 61%. clearly, it is essential that liability managers at banks understand the ramifications of embedding derivatives in their cd products. also, bank customers need to be aware of the opportunities these new prod ucts provide. ii. literatijre review academic research related to cds focus primarily on the role of government guarantees. for example, cook and spellman (1994) develop a theoretical model for cd rates related to the risk of default by the bank as well as third party government guarantees. they find empirical evidence around the collapse of the federal savings and loan insurance corpo ration that these risks were adequately priced. cooperman, lee and wolfe (1992) and cook and spellman (1991) provide further research in this area. a second strand of academic research attempts to identify the factors driving changes in cd rates. for example, fraser (1995) uses principal components analysis to ascertain that treasury bill rate changes effect cd yields in the united kingdom. in the united states, knez, litterman and scheinkman (1994) use a factor approach to identify that 86% of the variation in money market returns are explained by three factors, and a four factor model explains 90% of this variation. using cointegration analysis based on an error cor rection model, fung and isberg (1992) document a causal relationship where u.s. markets influence international markets in the 1981-1983 period. however, from 1984 to 1988, this causality seemed to reverse itself. the academic research related to valuing interest rate contingent claims is vast. the seminal paper by black and scholes (1973) on option pricing also addressed bond valua tion. two surveys of this vast literature include courtadon (1993) and ho (1995). tuckman (1995) offers more details on analytical techniques. in this paper, we seek to merge the recent advances in interest rate contingent claims valuation with the enhanced cd prod ucts. iii. cd valuation in developing a cd valuation model that incorporates embedded derivatives, several issues must be addressed. first, the value of the exotic option will increase as the maturity of the cd is lengthened. even the most enhanced product will not be associated with much risk if the maturity is only three months. however, some cds have very long maturities. the hunter’s special mentioned above had a nine-year maturity. second, the ability to early withdrawal is actually an option held by the depositor to put the cd back to the bank. the depositor has the option of selling the cd to the bank at a price less than its “market value.” what sort of early withdrawal penalty is appropriated for bankers and depositors alike? as some banks are dropping the early withdrawal penalty altogether, how much reduction in yield should they demand in return? third, the ability to reset the interest rate and the timing of this ability need to be addressed. for example, some two-year cd products aliow depositors to reset the interest 34 financial services review 5( 1) 1996 rate higher during the first 10 days of the second year. other two-year cd products allow depositors to reset the interest rate at any time during the two-year period. still other two year cd products automatically reset the interest rate during the life of the cd. some framework is needed to convert these various alternatives into tangible apys or ears for comparative purposes. fourth, most floating-rate cds provide a floor on interest rates while some have caps. the market value of a floating-rate cd with an embedded option (cap or floor) is equal to the value of a floating-rate cd with no option minus the value of the cap or plus the value of the floor. fifth, some cds allow depositors to add additional funds to the cd during its life. finally, the standard adjustments for compounding method and payment frequency must be made. the comparison of enhanced cds with standard instruments follows a two-step proce dure. first, the current value will be estimated for both the enhanced cd and the standard cd using the same reference interest rate (see the appendix for a sketch of how this would be accomplished). second, the price differential between these two cds will be used to establish the ear or apy for the enhanced cd. once the value of the enhanced cd is established, we can estimate either the effective annual rate (ear) or the annual percentage yield (apy). the ear is the appropriate inter est rate if it were compounded only once a year. the apy is the per period interest rate multiplied by the number of periods in a year. it is important to emphasize at this point that our objective is one of assessing relative value. that is, what is the value of the enhanced cd relative to a standard cd whose value we are assumed to know? because there is no secondary market in enhanced cds, it is impossible to use a direct market comparable valuation approach. it is standard practice in the derivatives markets to value off-market instruments using some valuation methodology calibrated to at-market instruments. here we assume the current standard cd trades at par, even though there is obviously profit to the bank, because our focus is on relative value. currently, bank cd rates are highly correlated with libor rates. the only unobserved parameter is volatility and it can be approximated from the appropriate interest rate deriv ative market, such as caps, floors, and swaption markets. it is not uncommon to determine the value of a financial security that has no liquidity. for example, fasb is presently assessing what to do with the fair value of executive stock options even though they are nonmarketable, require vesting, and so forth. there is a whole industry related to business valuation, where many times there are no direct market com parables. in derivative markets, standard practice is to rely on valuation models that have been appropriately calibrated. in the end, bankers and depositors must assess the validity of the valuation model and the inputs used. the depositor must carefully examine the enhanced features of these cds and deter mine whether the enhanced features are sufficiently beneficial to offset the loss in yield. one approach would be for the depositors to assign a price to the enhanced features based on their preferences. this depositor specific price would be used to establish an adjusted ear or apy. suppose the standard two-year, fixed-rate cd is quoted at 5% (ear,, where s denotes the “standard cd”). a banker is considering offering another comparable cd product with a low early withdrawal penalty (say 14&y penalty rather than 182&y), and this lower early withdrawal penalty resulted in an enhanced cd value of 100.20% of par based on an acceptable valuation model that has been appropriately calibrated (denote this value as computing yields on enhanced cds 35 encdvulue; hence, depositors place a 20 basis point value on the lower early withdrawal penalty), then the eab for this enhanced cd (ear,) would be computed as follows: encdvalue = par(1 + annualpremrate)“365, or annualpremrate = > 365/r _ 1 thus, ear,, = loo(ear, + annualpremrate) ear, = 1w[o.05+{(~~'730i}] = 100[0.05+ {o.ool]] = 5.11, (1) (2) (3) where t is the number of days in the deposit. in this case, the lower early withdrawal pen alty benefits depositors at an ear of 10 basis points per year. a similar analysis could be conducted using apy. based on the data above, the peri odic apy (apyj of the regular cd would be expressed as: apyp = (1 + ear)““, (4) where n is the number of compounding periods per year. if we assume quarterly com pounding, we have a periodic apy of: apyp = (1 + 0.05)‘” 1 = 0.012272 = 1.2272%. (5) the annualized apy for the standard cd (apy,) is: apy, = 0.012272 x n = 0.012272 x 4 = 4.9089%. (6) for the enhanced cd, we have: apy,, = apy, = 100~.049089+{~-l}x($] = 5.0089%. (7) the ear” is equal to the standard ear plus an adjustment reflecting the added value of the embedded options. hence, there are two factors impacting ear, the current cd rate reflected in ears, plus the value of the embedded option. this additional factor causes some standard pricing relationships to no longer hold. for example, we could observe a higher em,, and a higher market value of the enhanced product (encdvulue), due to increased volatility in the market. 36 financial services review 5( 1) 1996 we turn now to examine some of the general implications of this model. iv. cd valuation model implications in the examples and illustrations that follow, we assume a 5% flat term structure. the recombining binomial lattice used here is easily adapted to nonflat term structures. thus, fixed-rate cds with large early withdrawal penalties and floating-rate cds with no floors or caps will both be valued at par. suppose yield volatility is known to be 15%. how could we compare the following two-year cd products? a) standard fixed-rate cd (large early withdrawal penalty) b) standard floating-rate cd (no floor or cap) c) enhanced fixed-rate cd (only 7-day early withdrawal penalty) d) enhanced floating-rate cd (a guaranteed 5% ear floor) clearly, product c is preferable to a because of the limited early withdrawal penalty and product d is preferable to b because of the guaranteed floor rate. thus, we know the ear will be higher for products c and d. what is less clear is the comparability of prod ucts c and d. some depositors will have a strong preference for c and others for d depend ing in part on depositors’ perceived liquidity needs and expectations regarding future interest rates. using the binomial approach to valuation, we find the ear of both c and d to be approximately equal at 5.24%. hence, the ability to put back a fixed-rate cd with only a seven-day penalty is roughly equivalent to holding a floating-rate cd with a guar anteed floor of 5%. these enhanced features are worth roughly 24 basis points each. it is important to emphasize that the individual depositor will select from these prod ucts based on their own preferences. thus, the depositors will select a pure floating-rate cd if they have a strong view that interest rates are going to rise. uncertainty regarding this view will perhaps drive depositors to prefer a floating-rate cd with a floor. similarly, a depositor with no liquidity needs and a strong preference for a fixed rate would select a fixed-rate cd with heavy early withdrawal penalties. however, a fixed-rate depositor with some liquidity needs may select a fixed-rate cd with limited early withdrawal penalties. figure la illustrates in basis points the additional value of reducing the early with drawal penalty on fixed-rate, one-year cds under various volatility assumptions. the lower the early withdrawal penalty, the greater the value in basis points. also, the higher the volatility, the greater the value in basis points. figure lb illustrates the same pattern with greater magnitude for two-year cds. thus, a depositor seeking to reduce the early withdrawal penalty must be willing to pay the increased embedded option value. specifi cally, they must be willing to accept a lower stated rate. figure 2a illustrates in basis points the additional value of providing guaranteed floors on floating rate, one-year cds. we see tbe higher the guaranteed floor, the greater the value in basis points. higher volatility makes the impact of a floor even greater. figure 2b illus trates that for two-year cd products the value of the floor is even greater (but clearly the value is not double). y 38 financial services review 5( 1) 1996 floor rote 1.5% 1 .o% figure 243. impact of floor on ear one-year cds _.i._ floor rate 1 5% 1 .o% figure 2b. impact of floor on ear two-year cds computing yields on enhanced cds 39 depositor preferences will determine which enhanced cd product they prefer. a val uation model such as the one presented here would be beneficial in assisting depositors in sorting out their choices. as the cd product offerings increase, depositors will need to clearly understand the relative value trade-offs. v.applying thismodelinpractice thus far we have observed that embedded options in cd products are valuable and should be appropriately addressed when offering or buying new cd products. specifically, a lower early withdrawal penalty is valuable to the fixed rate cd depositor and hence will translate into a higher option-adjusted ear or apy than the standard cd paying the same fixed rate. a floating rate cd with a guaranteed floor will have a higher option-adjusted ear or apy when compared to a floating rate cd with no floor. does the higher ear or apy mean bankers should avoid offering enhanced cds? of course not. each depositor is unique and has different preferences and beliefs regarding the future course of interest rates. a depositor who believes interest rates will rise shortly will not be interested in a fixed rate cd with large early withdrawal penalties. a depositor that is uncertain about their liquidity needs will avoid cds with heavy early withdrawal penalties. the banker’s objective should be to offer appropriately priced cds that allow for product differentiation which results in a competitive advantage over banks offering stan dard products. brooks and white (1996) focus on the marketing aspects of enhanced cd products. the depositors objective should be to select the appropriate cd products best suited for them. for illustrative purposes, suppose there are two banks in a given market, stoic bank and modem bank. at two-year maturities, the stoic bank offers a fixed rate cd at 5% with 182&y early withdrawal penalty. the modem bank also has a similar cd but would like to expand its offering to attract other depositors. for simplicity, assume the yield curve is flat and cd rate volatility is 10% (about normal). the modem bank decides to offer a menu of pmd ucts at the two-year maturities as follows (see table 1). the problem is establishing the appropriate rate for products 3 and 4. using the valu ation model, we find the annual premium rate of the embedded options are 0.167% for prod uct 3 and 0.168% for product 4. thus, the appropriate ear for product 3 is roughly 4.833% (5% 0.167%) and for product 4 is 4.832% (5% 0.168%). however, some depositors will have strong preferences for products 3 or 4 and will be willing to accept even lower rates. table 1 modem bank products at 2-year maturities annual producr type earli j-r, premium rate 1 fixed rate cd 5% 5% 0% 2 floating-rate cd 5% 5% 0% 3 fixed rate cd, no ewp ? 5% 0.167% 4 floating-rate cd, 5% floor ? 5% 0.168% note: * ewp denotes early withdrawal penalty. 40 financial services review 5( 1) 1996 the benefits to the bank of attracting new depositors and gaining market share may well offset the additional administrative costs. the administrative costs should be minimal because the marketing channels have already been established. offering enhanced cds requires bankers to have a good grasp of their interest rate risk. alternatively, one benefit of good asset/liability management is the ability to comfortably offer new deposit products. the benefits to the depositor of expanded cd offerings may well offset the additional education costs involved in understanding the relative trade-offs. depositors must take the time and effort to clearly understand and assess the different features of these new cd products. one cost of offering enhanced cds is increasing a bank’s interest rate risk. the embed ded options will typically be used against the bank at a time when it is disadvantageous to them. for example, depositors will early withdraw when rates go higher causing the bank to lose lower cost of funds. the resulting interest rate risk can be effectively managed, if necessary, with over-the-counter derivative products such as interest rate caps and floors. one benefit of conducting this type of analysis is the ability to compute an option adjusted net interest spread (or margin). in the illustration above, it would appear as if the modem bank had a higher net interest rate spread if no adjustments were made to the quoted ear or apy. that is, the modem bank would be issuing deposits at 4.833% and 4.832% as opposed to 5%. appropriately managed, enhanced cds are an effective way to expand market share and profitability. in the same way depositors must understand that even though they receive a lower stated yield on enhanced cds, they are receiving valuable options. these options increase flexibility which, at times, would be extremely valuable to depositors. one interesting issue is the failure of depositors to behave rationally. anecdotal evi dence from a few bankers suggests that depositors will fail to exercise valuable options embedded in cd products. for example, liquidity considerations will cause some deposi tors to abandon valuable embedded options. hence, one interesting benefit to banks is offering appropriately priced embedded options and receiving the benefit of depositors lack of optimal exercise. for example, a bank offered a two-year cd product with the option to set the rate higher for a few weeks after the first year if one-year cd rates had risen after a year. this product was offered in late 1993. in late 1994 a significant percent age of cd customers failed to exercise their right to reset the interest rate higher even though rates had risen significantly. interestingly, this suggests that two cd products--one that resets upward automatically and one that requires the depositor to request the higher rate-will have different values to the bank. vi. concluding comments in summary, we examined the relative pricing of enhanced cd products. the objective was to provide a mechanism to facilitate comparison of vastly different cd products. further insights could be obtained if empirical evidence is gathered from banks that offer these products. this research is vital for bankers in their management of the resulting interest rate risk of enhanced cd offerings. for example, a five-year cd with no early withdrawal pen alty will have risk characteristics that are dramatically different from those of a five-year cd with a high early withdrawal penalty. without a viable method to evaluate their enhanced cd products, bankers cannot assess the interest rate risk of their institutions. computing yieijs on enhanced cds 41 depositors must realize that there is more than two dimensions (yield and maturity) when analyzing various cd products. depositors are accustomed to choosing a maturity but now must assess the type of rate, fixed or floating, and the value they place on liquidity. if nothing else, cds are no longer a mundane security. appendix a sketch of valuing the enhanced cd we use the following notations: current value of cd; maturity of cd in years; reference interest rate at time s; ceiling interest rate at time s; floor interest rate at time s; periodic interest rate (or coupon) paid for the cd; appropriate deposit amount upon which to compute interest payment; indicator function based on whether the cd can be reset (denoted by r); price of cd at t based on the ceiling rate, c; price of cd at t based on the floor rate,f; price of the cd at t based on the current market interest rate, r(t); dollar interest penalty at t for early withdrawal; and accrued interest since last interest payment (accrued at the appropriate interest rate). the model developed here is based on the following major assumptions: 1. the reference interest rate follows a multiplicative binomial process that con verges to a bivariate lognormal density as the time step tends to zero; 2. the appropriate discount rate is the reference interest rate; and 3. markets are complete and riskless arbitrage opportunities do not exist. this analysis could easily be adjusted to handle a discount rate at a fixed spread to the reference interest rate. for example, the reference interest rate may be 13-week treasury bills and the appropriate discount rate is approximately 50 basis point under 13-week bills. using standard present value arguments and rational behavior by depositors, we can value the cd as follows: t f r(s)ds po = b o [dpwc,(c(t)), mdc,cf(t)), c,(dt)))) 0 + ~,([rl){max(d, -dp, + az,, min(p, t, max(p’ 1, p,, ,)))}]dt. 42 financial services review 5( 1) 1996 this cd valuation model is implemented using a standard recombining binomial lat tice. this binomial method applied to interest rate securities is known to be slightly biased (e.g., see windas, 1993). we adjust for this bias to make the standard cd product value at par whether it is a fixed-rate cd or a floating-rate cd. acknowledgments: the author gratefully acknowledges the helpful comments of the journal editor karen eilers lahey, journal referees, jon carter, lawrence co&ran, cindy russo, benton gup, billy helms, hazel johnson, robert mcleod, pat rudolph, and darin white. the author also appreciates the assistance of pradeep kumar. references asinof, l. (1996, may 2). yields on cds lag behind rate increase. wall street journal, pp. cl & c15. black, f., & scholes, m. (1973). the pricing of options and corporate liabilities. journal of political economy, 81,637-654. brooks, r., & white, d. (1996). don’t copy competition: lead with new products. bank marketing, 28(5), 13-17 cook, d.o.. & spellman, l. (1991). federal financial guarantees and the occasional market pricing of default risk: evidence from insured deposits. joumul of banking and finance, 15(6), 1113 1130. cook, d.o., & spellman, l. (1994). repudiation risk and restitution costs: toward understanding premiums on insured deposits. journal of money, credit and banking, 26(3), 439-459. coopertnan, e.s., lee, w.b., & wolfe, g.a. (1992). the 1985 ohio thrift crisis, the fslic’s sol vency, and rate contagion for retail cds. journal of finance, 47(3), 919-1082. courtadon, 9. (1993). a survey of bond option pricing models. in r.j. schwartz & c.w. smith (eds.), advanced strategies in financial risk management (pp. 3-67). new york: new york institute of finance. currier, c. (1987). the investor’s encyclopedia. new york: business news. famham, a. (1994). a bang for your buck in boulder. fortune, 130(2), 17-20. fraser, p. (1995). an empirical analysis of the relationship between uk treasury bills and the term structure of certificates of deposit. bulletin of economic research, 47(2), 143-160. fung, h.-g., & isberg, s.c. (1992). the international transmission of eurodollar and u.s. interest rates: a cointegration analysis. journal of banking and finance, 16(4), 757-769. ho, t.s.y. (1995). evolution of interest rate models: a comparison. journal of derivatives, summer, 9-20. knez, p.j., litterman, r., & scheinkman, j. (1994). explorations into factors explaining money market returns. journal of finance, 49(.5), 1861-1882. o’connell, v. (1994). turn a deaf ear to these jazzy new cds. money, 23(8), 30. peterson bank successful in world cup marketing foray. (1994). bank marketing, 26(8), 6. tuckman, b. (1995). fixed income securities tools for today’s markets. new york: john wiley 8~ sons. windas, t. (1993). an introduction to option-adjusted spread analysis. new york: bloomberg l.p. pii: s1057-0810(00)00045-7 cash flow: a quick and easy way to learn personal finance ronald r. crabb* finance and business law departments, university of wisconsin-whitewater, 5002 carlson hall, whitewater, wi 53190-1797, usa abstract a cash flow spreadsheet methodology simplifies solving a number of common personal finance problems. at schools where a basic finance course is not a prerequisite for the personal financial planning course, this methodology makes learning personal finance skills and concepts quick and easy. it reinforces time value of money mathematics when students have had a prerequisite finance course. creating the cash flow spreadsheets, editing them for errors, and testing the finished spreadsheets for logical consistency imprints in minds of most students the skills and concepts you are trying to teach. © 1999 elsevier science inc. all rights reserved. jel classification:c6; c63 keywords:goal seek; solver; cash flow spreadsheet 1. introduction students understand cash flow. they work hard in the summer, storing up cash for the school year. many of them live at home during the summer to avoid paying rent and to avoid buying food. when the school year starts, they begin to hemorrhage cash. many of them go to the financial aid office to borrow $2,500 at a zero rate of interest until six months after they graduate. a few of those borrowers do not really need the cash, but they do understand that it makes sense to borrow money if the u.s. government is paying the interest while they attend school and also for the first six months after they graduate. students understand cash flow. * tel.: 11-262-472-1326; fax:11-262-472-4863. e-mail address:crabbr@uwwvax.uww.edu (r.r. crabb). financial services review 8 (1999) 269–282 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00045-7 hence, the easiest way to teach students how personal finance math works is to show them the cash flows that underlie the financial transactions that they currently are (or will be) entering into. students borrow money to go to college. they begin to experiment with credit cards. they will be buying houses. they will be saving for retirement. how much should they save? how big will the mortgage payment be? how expensive are credit cards? how big will the stafford loan payment be six months after graduation? 2. the student loan consider a student who takes out a stafford loan. the student borrows money by signing a note for $2500, but the student only receives $2400. a $100 origination fee, 4% of the loan amount, covers up-front expenses and provides insurance to repay the loan if the student should die before the loan has been repaid. assume that the interest rate on the loan is 7.5%, the current stafford loan rate, during the first two years of the repayment period, and then jumps to 8.25%, the maximum stanford loan rate, during the remaining eight years of the repayment period. to minimize the size of the tables in this article, assume that payments are made on a quarterly basis; in reality, payments would be made monthly. if the student borrows $10,000 over a four-year period, how large will her or his payment be when she or he begins to repay the loan? the cash flow is easy to set up on a spreadsheet: there are only a few simple equations. see table 1 for the results of those equations. the next few paragraphs detail the classroom procedure that i use to create a stafford loan excel 97 cash flow spreadsheet. this is a mental exercise, not a mechanical one. the learning and understanding does not come from simply typing a few equations and looking at the finished cash flow picture. i require my table 1 an excel spreadsheet for a student loana a b c d e 1 time period loan amount (boq) interest expense loan payment loan amount (eoq) 2 1 $10,000.00 $187.50 $300.00 $9,887.50 3 2 $ 9,887.50 $185.39 $300.00 $9,772.89 4 3 $ 9,772.89 $183.24 $300.00 $9,656.13 5 4 $ 9,656.13 $181.05 $300.00 $9,537.18 6 4 $ 9,537.18 $178.82 $300.00 $9,416.01 7 6 $ 9,416.01 $176.55 $300.00 $9,292.56 8 7 $ 9,292.56 $174.24 $300.00 $9,166.79 9 8 $ 9,166.79 $171.88 $300.00 $9,038.67 10 9 $ 9,038.67 $186.42 $300.00 $8,925.09 11 10 $ 8,925.09 $184.08 $300.00 $8,809.17 40 39 $ 4,385.68 $ 90.45 $300.00 $4,176.13 41 40 $ 4,176.13 $ 86.13 $300.00 $3,962.27 a boq, beginning of quarter; eoq, end of quarter. 270 r.r. crabb / financial services review 8 (1999) 269–282 students to think about what formula or number belongs in each cell. beginning with a blank excel spreadsheet, i type the labels shown in table 1 and the 1 shown in cell a2. i then ask my students “how much did the student borrow?” usually about half of the class says simultaneously “$10,000,” and i type a 10000 into cell b2. i then ask for a formula to compute the interest expense. almost always, some student says “50.075*b2,” and i type that formula into cell c2. rather than saying “it’s wrong.” i say: “it’s almost right. how does the equation need to be modified to compute thequarterly interest expense?” given the hint, the same student, or a different student, says “50.075/4*b2,” and i edit the formula in cell c2. after explaining that cell d2 needs a number which is bigger than the interest expense shown in cell c2, else the ending loan balance will be bigger than the beginning loan balance, i enter a $300.00 into cell d2. the $300.00 in cell d2 is an arbitrary entry. it is not the loan payment amount; it will need to be adjusted (by goal seek in an excel 97 environment or back solver in a lotus 1–2-3 environment) when the spreadsheet equations underlying the cash flow are finished. i now ask for a volunteer to give me an equation to compute the loan balance at the end of the first quarter. i am indifferent as to whether the equation volunteered is correct or incorrect. if correct, i go on to row two. if incorrect, i say: “it’s almost right. how does the equation need to be modified to compute the ending loan balance?” after the formula 5b21c2-d2 is entered into cell e2, i move to the next row. i ask my students for a formula to compute the time period, and, after getting a correct answer or after editing an incorrect answer, enter5a211 into cell a3. i ask my students for a formula for the beginning loan balance at the beginning of the second quarter, and, after getting a correct answer or editing an incorrect answer, enter5e2 into cell b3. i click on cell c2, and i ask my students if the formula in cell c2 is appropriate for cell c3. after getting the answer “yes.” i copy the c2 formula into cell c3. in cell d3 i type5d2. i click on cell e2, and i ask my students if the formula in cell e2 is appropriate for cell e3. after getting the answer “yes.” i copy the cell e2 formula into cell e3. i then click on cells a3 through e3, and copy the formulae down through row 10. i click on cell c10, and ask my students what change needs to made to this formula. after a student replies that the 0.075 needs to be changed to 0.0825 because the interest rate changes from 7.5% to 8.25% at the beginning of the third year of the loan, i edit cell c10 to50.0825/ 4*b10. i click on cells a10 through e10, and copy the formulae down through row 41. finally, i click on cells b2 through e41, and format to dollars by clicking the $ icon on the toolbar. the process of requiring your students to think about what equation or number is going into each cell forces them to become active learners. since the balance of the loan after 40 payments is not zero, i ask someone in the class to volunteer a new number for the loan payment amount. when a student says “$350.00,” i type 350 in cell d2, and the loan balance is now $906.37. after entering 375 into cell d2, the loan balance is ($621.59). a 365 entry results in a loan balance of ($10.41). it doesn’t take too long to get the loan balance close to zero; however, the closest a trial-and-error solution can come is minus one cent. the inability to get an exact zero answer allows for a discussion of how loans are actually amortized in the real world. to amortize his/her loan, the stafford loan borrower makes 39 equal payments of $364.83 and one final payment of $364.82. while a trial-and-error solution is certainly possible,1 it is easier and quicker to set the cursor on cell e41, go up to the toolbar and click on tools, and then click on goal seek. see fig. 1. 271r.r. crabb / financial services review 8 (1999) 269–282 the goal seek tool requires three input values to solve a problem: a target cell (cell e41, the ending loan balance), a value for the target cell, and the cell address that you want to change to cause the target cell to take on the desired value. in theto value: box, enter a zero. in the by changing cell: box, enter cell d2. click on the ok button, and the goal seek algorithm will iterate towards a solution. the goal seek algorithm will try different values for the loan payment amount to find the loan payment amount that causes the loan balance, after 40 loan payments, to compute to a zero value. see table 2. the cash flow goal seek methodology enhances learning because the student can literally see (on the monitor or on a printed hard copy) exactly how the cash flows from him/herself to the lender (or how a retirement fund is accumulated and liquidated). the student can see the interest expense (or earnings); the student can see the loan balance declining to zero over the life of the loan (or the retirement fund balance maximizing at retirement before declining to zero at the end of the retirement period). the process of writing the equations, targeting the ending loan balance to zero, and using first trial-and-error and then goal seek allows most students to quickly conceptualize what, for many, have been difficult concepts to comprehend. this is actually much easier to do in real time in a classroom than it is to explain in words. working live in a classroom with a computer overhead display and a blank excel spreadsheet, this problem can be set up and solved in about three minutes. if you have access to a computer lab classroom, then your students can work with you. early in the semester the time will increase to about ten minutes because your students with weak spreadsheet skills will fig. 1. setting up goal seek to solve for the loan payment. 272 r.r. crabb / financial services review 8 (1999) 269–282 make typos, formulae errors, and copy-down errors. over the course of the semester most of your students will become proficient spreadsheet users as they master personal finance problem solving skills and learn personal finance concepts. whether you work with your fig. 2. the constant dollar spreadsheet and solver parameters. table 2 the goal seek solution to the student loan problem a b c d e f g 1 time period loan amount (boq) interest expense loan payment loan amount (eoq) pvifs 2 1 $10,000.00 $187.50 $364.83 $9,822.67 0.9816 35 34 $ 2,355.51 $ 48.58 $364.83 $2,039.27 0.5069 36 35 $ 2,039.27 $ 42.06 $364.83 $1,716.50 0.4967 37 36 $ 1,716.50 $ 35.40 $364.83 $1,387.07 0.4866 38 37 $ 1,387.07 $ 28.61 $364.83 $1,050.85 0.4768 39 38 $ 1,050.85 $ 21.67 $364.83 $ 707.69 0.4672 40 39 $ 707.69 $ 14.60 $364.83 $ 357.46 0.4577 41 40 $ 357.46 $ 7.37 $364.83 $ 2 0.4485 42 41 pvifa is 5 27.4100 43 42 payment is 5 $364.83 273r.r. crabb / financial services review 8 (1999) 269–282 students and they get instant feedback in a lab or delayed feedback when working on their own after class, most students find the cash flow goal seek methodology easy to understand. a number of my brighter students have told me that “this is like cheating. we’re not using the interest factors we learned in basic finance.” while the methodology is so simple that it may seem like cheating, it is just a different way of solving the loan amortization problem. the simplicity of the cash flow methodology coupled with the process of writing the cash flow equations, targeting the ending loan balance to zero, solving the problem with goal seek, and examining the spreadsheet solution for logical consistency (no equations errors et al.) imprints the amortization process in the minds of most students. the real issue is understanding, not methodology, not machinery. from an instructional standpoint, the important questions are: 1. do students understand the loan amortization process; 2. can students determine the size of a loan payment; and 3. can these concepts quickly and easily become integral parts of the student’s knowledge base? for most students, the cash flow goal seek methodology enables students to answer “yes.” to the above three questions. the cash flow goal seek methodology has four major advantages over the two traditional approaches (using a financial calculator or using pvifa tables and a simple calculator) to finding loan payments amounts. first, the entire process of creating the cash flow spreadsheet allows students to visualize and understand exactly how the loan amortization process works. second, the process of checking the spreadsheet for errors forces students to think about the relationships that must exist in a properly created amortization table. is the loan balance continuously decreasing? is the loan balance at the beginning of one time period equal to the ending loan balance of the prior period? is the loan payment amount constant (except for the last payment if using answers computed to the nearest penny vis-a`-vis using 15 digit goal seek answers)? does the interest expense continuously decrease as the loan payments are made? is the final loan balance equal to zero? does the sum of the loan payments, minus the sum of the interest expense, compute to the initial loan amount? third, financial calculators or simple calculators and pvifa tables do little to foster a conceptual understanding of the loan amortization process. input some numbers; an answer appears on the display screen. the purely mechanical process of computing an answer using a hand-held calculator does not foster an understanding of the loan amortization process. fourth, not only does the financial calculator not provide any insight as to how the loan amortization process works, the “black box” answer is, for most real world loan problems, not checkable. it simply is not feasible to create amortization tables using hand calculators and paper and pencil for most real world loans that have from 60 to 360 monthly payments. the cash flow methodology allows students to simultaneously find the solution to the loan problem and to see exactly how the loan balance decreases to zero over the 40 quarter time period. students need to look at the cash flow columns to make sure that an equation error has not gotten into the cash flow algorithm. when students are examining their spreadsheets for errors, they are thinking; when they find errors, they are learning; after they have fixed 274 r.r. crabb / financial services review 8 (1999) 269–282 their errors, they have learned exactly how the amortization process works. in a learning context, this methodology creates an active learning situation. students become comfortable with loan payment computation and loan amortization. depending upon your needs and your students’ backgrounds and their needs (e.g., have they already had a basic finance course? is this their first course in finance? will they be taking more courses in finance using this course as a prerequisite?), it is easy to integrate the basic finance course into your financial planning course or vice versa. with the addition of a single column to the spreadsheet shown in table 2, you can show your students the traditional pvifa approach to solving this problem while also showing them a new way to compute pvifa factors. 3. the pvifa approach for a student loan in cell g2 of the spreadsheet shown in table 2 enter51/(110.075/4). in cell g3 enter 5g2/(110.075/4). click on cell g3, and copy down through row 10. edit cell g10 to be 5g9/(110.0825/4) to reflect the change in the interest rate. click on cell g10 and copy down through row 41. set the cursor in cell g42. go up to the toolbar and double click the summation sign to sum up cells g2 through g41. the resulting sum, 27.41, is the present value annuity factor (pvifa) for this loan problem. in cell g43 divide $10,000 by that annuity factor. the quotient, $364.83, computes the exact same payment as determined via the cash flow methodology. students who have had a basic finance class learn a new way to compute annuity factors, and their memories of how to solve loan amortization problems are refreshed; students who have not had a finance class can see how annuity factors are built and used.2 i reinforce student learning by using the cash flow goal seek methodology to solve credit card and mortgage loan problems. following that, students have a reasonably good understanding of the loan amortization process and are ready to move to the next topic: retirement funding. 4. cash flow retirement planning perhaps some of the nastiest mathematics that students can encounter in a financial planning course is the math that underlies retirement planning. rich fortin (1997), the author of “retirement planning mathematics” (hereinafter referred to as rpm), used 17 equations in the body of his article and another 11 equations in the appendix to show the present value/future value equations which need to be solved to determine how much to save for retirement. he demonstrated that the retirement savings approach which people might be most likely to use did not have a closed form solution. like loan payment computation, a cash flow methodology greatly simplifies retirement planning mathematics. to save time and space, and to allow for a comparison between a standard present value/future value approach to retirement mathematics and a cash flow approach, this article will assume the same case facts as the rpm example. those facts follow. assume an age 25 student has just graduated from college. she expects to earn $25,000 in her first year of employment. she expects to work for 30 years, and she expects to be retired for 25 years. she 275r.r. crabb / financial services review 8 (1999) 269–282 expects an inflation rate of 3.2% in both her salary and in prices, and she wants a constant dollar annuity equal to one half of her final salary. she assumes that she will earn an 8% nominal interest rate on her investments, and she makes all of her computations on an end-of-the-year basis. she considers three approaches to funding her retirement. she could save a constant dollar amount every year, she could save a constant percentage of her salary each year, or she could save some initial percentage of her salary, and then increase that percentage each year.3 because the nature of retirement planning is highly “what if?” based (what if she actually earns 12%? what if she only earns 4%?), an alternate excel tool, solver, is used in place of goal seek. solver uses a more complex trial and error algorithm; as a result, solver takes more computer time to get an answer the first time it is used. however, this one-time cost is outweighed when testing multiple interest rates because solver remembers the target cell, remembers the zero value for that target cell, and remembers the cell that needs to be changed to cause the target cell to compute to a zero value. once the solver algorithm has been defined for a particular problem, it allows the student to quickly do “what if?” analysis. create a cash flow spreadsheet that shows the accumulation and liquidation of a retirement fund under the constant dollar approach. once the formulae are written (see the appendix for the equation details), bring up solver to compute the amount of money that needs to be saved on an annual basis. like goal seek, solver requires that you specify three parameters: the target cell, the value for the target cell, and the cell whose value you want to change to make the target cell compute to the value you want that cell to have. in this case the target cell is the retirement fund balance at the end of the retirement period. hence, have the cursor in cell e56 when you bring up solver. in fig. 2, the second line of the solver box shows three possible options for the target cell:max, min, value of . set the pointer in the little circle to the left of the wordsvalue of, and click the mouse button. we want the target cell, e56, to compute to zero, and, since the solver default value is zero, this parameter is now correctly specified. enter an i2 into theby changing cell box. click the solve button in the solver box, then click the ok button when the “solver has found a solution” box appears on the monitor. the solution to this problem is shown in table 3. the resulting spreadsheet allows your students to see what happens to the retirement fund of the hypothetical age 25 student as she accumulates capital for her retirement and then liquidates that capital during her retirement. as a byproduct of solving the retirement funding problem, the size of the retirement gap4 shows on the spreadsheet in cell e31, the year in which our hypothetical student retires. again, this is much easier to do in real time in a classroom than it is to explain in words. from start to finish on a blank excel spreadsheet, solving the constant dollar problem takes about ten minutes. as before, the student needs to examine his/her spreadsheet for logical consistency. is the retirement fund at the beginning of one time period equal to the to retirement fund at the end of the previous period? do the interest earnings increase continuously through the 30 year accumulation period, and continue to increase during the liquidation period, although at a slower rate, until the annuity payments become larger than the interest earned? is the salary of the hypothetical student increasing continuously from when she begins work until when she retires? does the sum of the interest earned over the entire accumulation and liquidation 276 r.r. crabb / financial services review 8 (1999) 269–282 period plus the sum of the dollars invested during the accumulation period equal the sum of the retirement annuity payments paid out during the liquidation period? is the ending retirement fund at the end of the liquidation period equal to zero? the process of examining the spreadsheet for logical or mathematical errors, and fixing those errors, again creates an active learning environment that helps students conceptualize the retirement funding process. modifying the table 3 spreadsheet to solve the constant percentage problem and to solve the constant growth of an initial percentage problem is straightforward. copy the entire spreadsheet shown in table 3 and paste it into two blank spreadsheets. on each sheet, edit a few cells, copy down the new formulae, and bring up solver to solve the problem. in about five minutes you can solve both problems. see the appendix for the equation modifications needed to create these two spreadsheets. before going to the appendix, however, try to modify the equations yourself to see if you can create these spreadsheets. irrespective of how our hypothetical student chooses to fund her retirement, she needs $469,526 in her retirement fund when she retires, assuming an 8% nominal interest rate. all types of sensitivity analysis are now possible for all three approaches. the simplest, changing the nominal interest rate, takes about 20 seconds per each interest rate. type in the new rate, bring up solver, and push the solve button. table 4, in the three columns that begin with year one, shows the required fund at retirement (which is also the retirement gap) for interest rates from 4% through 12%. what if the student choosesnot to begin saving for her retirement in her first year of employment? what if she puts off saving for retirement until her sixth year of employment? on each spreadsheet, zero out (that is, type in a zero) cell e6, the retirement fund at the end table 3 the constant dollar solution a b c d e f g h i 1 year retirement fund (beginning of the year) interest earned dollars invested (or) dollars paid out retirement fund (end of the year) nominal investment interest rate salary (& price) inflation rate end-ofyear salary constant dollar payment amount 2 1 $ 2 $ 2 $ 4,145 $ 4,145 8.0% 3.2% $25,800 $4,145 3 2 $ 4,145 $ 332 $ 4,145 $ 8,621 $26,626 6 5 $ 18,677 $ 1,494 $ 4,145 $ 24,315 $29,264 11 10 $ 51,757 $ 4,141 $ 4,145 $ 60,043 $34,256 16 15 $100,364 $ 8,029 $ 4,145 $112,538 $40,099 21 20 $171,783 $13,743 $ 4,145 $189,670 $46,939 26 25 $276,720 $22,138 $ 4,145 $303,003 $54,946 31 30 $430,908 $34,473 $ 4,145 $469,526 retirement gap $64,318 32 1 $469,526 $37,562 $(33,188) $473,900 33 2 $473,900 $37,912 $(34,250) $477,562 36 5 $482,378 $38,590 $(37,644) $483,323 41 10 $474,470 $37,958 $(44,065) $468,362 46 15 $422,886 $33,831 $(51,582) $405,135 51 20 $300,310 $24,025 $(60,380) $263,955 56 25 $ 65,444 $ 5,236 $(70,679) $ (0) 277r.r. crabb / financial services review 8 (1999) 269–282 of year five. on the constant growth of an initial percentage spreadsheet, one additional edit needs to be made. since the initial percentage saved by our hypothetical student was assumed to be 5%, cell j7 needs to redefined as 5%. bring up solver, and push the ok button. while numbers will appear on the monitor in accumulation years 1 through 5, solver disregards these numbers because of the $0 entry into cell e6. table 4, in the three columns beginning with the words year six, shows the effect of waiting until the end of six years before beginning to save for retirement for all three approaches. a more interesting “what if?” analysis, feasibility analysis, can be done on the constant growth spreadsheet. column j shows the percentage of salary saved each year. this column lists what percentage of her salary the hypothetical student must save each year, given the nominal rate of return. if the student chooses to invest conservatively at 4%, by the time she gets to age 65, she needs to save 113% of her salary. this is not feasible. even if she chooses to invest a bit more aggressively, at 6%, she still needs to save 63% of her salary in her last year of employment. again, this is not feasible. sensitivity and feasibility analyses allow students to quickly grasp important retirement concepts. again, the real issue is understanding, not methodology, not machinery. from an instructional standpoint, the important questions are 1. do students know how to compute the savings amount required to fund a person’s retirement; 2. do they understand the accumulation and liquidation process inherent in retirement funding; 3. can they compute the size of the retirement gap; 4. do they understand the need to begin saving for their retirements early in their careers; 5. do they understand the need to invest aggressively when they are young (or, alternately stated, understand the long-run cost of investing conservatively when they get old); and, table 4 sensitivity analysis and funding amounts (in dollars or percentages) nominal rate real rate retirement fund(gap) year one cash flow constant dollar year six cash flow constant dollar year one cash flow constant percentage year six cash flow constant percentage year one cash flow constant growtha year six cash flow constant growthb 4% 0.78% $728,311 $12,986 $17,488 33.7% 41.2% 11.4% 15.1% 5% 1.74% $647,123 $ 9,740 $13,559 25.8% 32.5% 10.3% 13.9% 6% 2.71% $578,313 $ 7,315 $10,541 19.8% 25.6% 9.1% 12.7% 7% 3.68% $519,700 $ 5,502 $ 8,217 15.2% 20.2% 7.9% 11.4% 8% 4.65% $469,526 $ 4,145 $ 6,423 11.7% 16.0% 6.5% 10.1% 9% 5.62% $426,364 $ 3,128 $ 5,034 9.0% 12.7% 4.9% 8.6% 10% 6.59% $389,056 $ 2,365 $ 3,956 6.9% 10.1% 3.0% 7.0% 11% 7.56% $356,653 $ 1,792 $ 3,117 5.3% 8.1% 0.6% 5.1% 12% 8.53% $328,379 $ 1,361 $ 2,463 4.1% 6.5% 22.4% 3.0% a assumes savings 5% of her end-of-year one salary. b assumes savings 5% of her end-of-year six salary. 278 r.r. crabb / financial services review 8 (1999) 269–282 6. can these concepts quickly and easily become integral parts of the student’s knowledge base? for most students, the cash flow solver methodology produces “yes” answers to most of the above questions. 5. evaluating alternative investments consider an individual investor who needs to choose between two possible investments. one costs $10,000, and the investor expects that s/he should double her/his investment in nine years. the other investment also costs $10,000, and the investor expects to triple her/his investment in 14 years. which investment offers the investor the highest rate of return? the cash flow methodology spreadsheet solution to this problem is shown in table 5. the $10,000 investment is set up to grow an arbitrary interest rate. use goal seek to target the eoy (end of year) table 5 cash flow investment spreadsheet, irr computationa a b c d e f g h i j k 1 8.0% 8.2% 2 year investment (boy) implicit interest investment (eoy) year investment (boy) implicit interest investment (eoy) ($10,000) 3 1 $10,000 $ 801 $10,801 1 $10,000 $ 816 $10,816 $0 4 2 $10,801 $ 865 $11,665 2 $10,816 $ 883 $11,699 $0 5 3 $11,665 $ 934 $12,599 3 $11,699 $ 955 $12,654 $0 6 4 $12,599 $1,009 $13,608 4 $12,654 $1,033 $13,687 $0 7 5 $13,608 $1,089 $14,697 5 $13,687 $1,117 $14,805 $0 8 6 $14,697 $1,177 $15,874 6 $14,805 $1,209 $16,013 $0 9 7 $15,874 $1,271 $17,145 7 $16,013 $1,307 $17,321 $0 10 8 $17,145 $1,373 $18,517 8 $17,321 $1,414 $18,734 $0 11 9 $18,517 $1,483 $20,000 9 $18,734 $1,529 $20,264 $0 12 10 $20,264 $1,654 $21,918 $0 13 11 $21,918 $1,789 $23,707 $0 14 12 $23,707 $1,935 $25,643 $0 15 13 $25,643 $2,093 $27,736 $0 16 8.5% 14 $27,736 $2,264 $30,000 $30,000 17 year investment (boy) implicit interest investment (eoy) irr 5 8.2% 18 1 $10,000 $ 849 $10,849 19 2 $10,849 $ 921 $11,769 20 3 $11,769 $ 999 $12,768 21 4 $12,768 $1,083 $13,851 22 5 $13,851 $1,175 $15,027 23 6 $15,027 $1,275 $16,302 24 7 $16,302 $1,383 $17,685 25 8 $17,685 $1,501 $19,186 26 8.5 $19,186 $ 814 $20,000 a target cells and variable cells are bold. boy, beginning of year; eoy, end of year. 279r.r. crabb / financial services review 8 (1999) 269–282 investment value to $20,000 at the end of the 9th year for the first investment and $30,000 at the end of the 14th year for the second investment. that is, in theto valuebox, type 20000 or 30000. the variable cell is the interest rate. the 14 year investment offers the highest expected rate of return, 8.2%, versus 8.0% for the nine year investment. like many investment choices facing individual investors, this solution is quite sensitive to small changes in time. if the nine year investment horizon were just one half a year less, that is, the investor expected to double her/his money in eight and one-half years rather than nine years, then that investment would yield 8.5%, and it would be preferred over the 14 year investment. students owning financial calculators can input $10,000 as the present value, $20,000 (or $30,000) as the future value, nine (or 14) as the time period, and ask the calculator to compute the interest rate that solves each problem. if your students have had a prerequisite finance course, then you can solve this investment problem using the irr function, again integrating the finance course into your personal finance course. set up the string of numbers shown in column k, cells 2 through 16, and type 5irr(k2:k16) in cell k17. after formatting cell k17 to one decimal point, the irr for the 14 year investment is 8.2%. using an fvif table, your students would drop down the period column to 14, go horizontally across the table looking for a 3 in the 14th period row. your students would find a 2.937 in the 8% column and a 3.342 in the 9% column, and could interpolate 8.2% as the answer. 6. conclusions there is no need for a prerequisite basic finance course for personal financial planning when using this methodology. by appealing directly to life experiences which students have had or will have, this methodology engages students’ interest in learning. the information is structured such that it can be readily grasped. avoiding present and future value math (and later showing how that math works), providing visual numerical pictures of the entire process, and requiring students to write, edit, and fix the equations underlying the amortization process allows those with little experience to simultaneously learn financial planning math skills and financial planning concepts. many personal finance problems involve cash flow, and any problem that involves cash flow is solvable by this methodology. while simplifying the math makes a cash flow methodology both easy-to-use and useful, especially for the retirement problem that does not have a closed form solution, perhaps its greatest benefit lies in its requiring students to understand exactly what is happening when they take out loans, plan their retirements, and choose between alternative investments. notes 1. to do this by trial and error, set cell f25e41, making it possible to see the final loan balance and the loan payment simultaneously. since a $300 loan payment results in a positive loan balance, try a number larger than $300. whether you type $350 or 280 r.r. crabb / financial services review 8 (1999) 269–282 $400 or $500 or even $1000 does not matter. a visual trial-and-error process is actually rather efficient. it usually takes less than a minute to find a loan payment amount that will get the ending loan balance to1/a few cents. it has instructional value to demonstrate that the goal seek algorithm is nothing more than a trial-anderror algorithm. it is not a “black box”; there is nothing mysterious going on inside of the computer as goal seek iterates to a value of 364.829755293873 for the loan payment amount. 2. because many students do not really understand the annuity concept when they enter a financial planning class, either because they have not had a basic finance class or the concept failed to gel when they took a basic finance class, another useful exercise at this point in the classroom is to show them what the annuity factor 27.41 means. in cell j2, enter5g42. enter50.075/4*j2 into cell k2. enter 1 into cell l2. enter 5j21k2-l2 into cell m2. enter5m2 into cell j3. copy-down the formula from k2 into cell k3. set cell l35l2. copy-down the formula from cell m2 into cell m3. click on cells j3 through m3. copy down through row 10. edit the interest formula in cell k10 to 50.0825/4*j10. click on cells j10 through m10, and copy their formulae down through row 41. in cell m41, the student will see a $0.00. you invest $27.41 in the bank. the bank pays you interest of 7.5% annually (compounded quarterly) for two years, then 8.25% annually (compounded quarterly) for eight years. at the end of each quarter, you withdraw one dollar. at the end of 40 quarters, your bank account has a zero balance. the present value of a $1.00, payable at the end of each quarter for 40 quarters, given the above interest rates, is $27.41. by looking at how the cash flows through a bank over a ten-year period, your students can see exactly how annuities work. note that you could put any number into cell j2, target cell m41 to a zero value, and then use goal seek to find the initial sum of money, $27.41, needed to provide a dollar per quarter for 40 quarters. 3. these three approaches (constant dollar, constant percentage, and constant growth of an initial percentage) are the three approaches that the major mutual fund providers put on their web sites for use by their customers to answer the question “how much should i save to provide for my retirement?” the rpm article explored the mathematics underlying those three approaches. 4. when our hypothetical student retires at age 55, to provide a constant dollar annuity equal to one-half of her final salary for 25 years, she needs $469,526, assuming an 8% nominal interest rate. in the same sense that $27.41 was the amount of money needed to provide a $1.00 per quarter for 40 quarters, $469,526 is the amount of money needed to provide an initial annuity of $33,188 in her first year of retirement and $70,679 in her last year of retirement. by assumption, she has not yet accumulated any funds for retirement when she begins working. since $469,526 (her future retirement fund) minus $0 (her current retirement fund) is equal to $469,526, the cash flow methodology computes the retirement gap as a byproduct of solving the retirement funding problem. alternately, $469,526 is the present value of the 25 retirement annuity payments, computed on an ordinary annuity basis. 281r.r. crabb / financial services review 8 (1999) 269–282 acknowledgment the author wishes to thank two anonymous reviewers. without their input, this paper would not exist. his thanks are also extended to his secretary and his wife for reading all drafts of this paper, checking cell addresses for accuracy and grammar for consistency. appendix a.1. equation instructions for the constant dollar spreadsheet shown in fig. 2. after entering the labels in cells a1 through i1, a 1 is entered in cell a2, a 0 in cell b2, 5$f$2*b2 in cell c2,5$i$2 in cell d2,5sum(b2:d2) in cell e2, 8% in cell f2, 3.2% in cell g2, 525000*(11$g$2) in cell h2, and $2400 in cell i2. as before (the loan and mortgage cash flow spreadsheets), some cells in row 3 are different from those in row 2. cell a3 is 5a211; cell b3 5 e2. click on cells c2, d2, and e2, and copy their formulae into cells c3, d3, and e3. cell h3 is5h2*(11$g$2). click on cells a3 through h3, and copy down their formulae through row 56. because she stops working at age 55, clear the contents of cells h32:h56. enter a 1 incell a32 to define the first year of retirement. in d32, enter5 20.5*h31*(11$g$2). in cell d33, enter5 d32*(11$g$2); then copy this formula down through cell d56. format to dollars without any cents. a.2. equation editing instructions for the constant percentage spreadsheet to edit the constant percentage spreadsheet, change the label in cell i1 to “constant percentage.” reformat cell i2 to a percentage format by typing an arbitrary 10.00% in cell i2. change the formula in cell d2 to5h2*$i$2. click on cell d2, and copy this formula down through row 31 (or working year 30). bring up solver with e56 as the target cell, target cell e56 to zero, type in i2 as the variable cell, and then click on the solve button. a.3. equation editing instructions for the constant growth in an initial percent spreadsheet to edit the constant growth of an initial percentage spreadsheet, change the label in cell i1 to “constant growth rate.” reformat cell i2 to a percentage format by typing an arbitrary 10.00% in cell i2. type the label “percent of salary saved” in j1 and type 5.00% in j2. cell d2 5h2*j2. click on cell d2, and copy down through cell d31. cell j35j2*(11$i$2). click on cell j3, format to percent with two decimals, and then copy down through cell j31. bring up solver with e56 as the target cell, target cell e56 to zero, type in i2 as the variable cell, and then click on the solve button. reference fortin, r. (1997). retirement planning mathematics.journal of financial education, 23,73–80. 282 r.r. crabb / financial services review 8 (1999) 269–282 pii: s1057-0810(01)00085-3 � ������� � �� ��� � �� � ��� � � � � �� � � � � � �� �������� � �� ��� � � �� � ����� �������� � ��� �� �� ���� ���� � ��� �� � ���������� ����� � ��� � � �� �� �� ���� � ������ ����������� ���� � � � ! � ������" � � �# $%&'%� ��# ! � �� � " ����� "##$% � � �� � � � �� � ���� $ &�� �� � "##$% ��� �� � "# &�� �� � "##$ �������� '�� ��� � (��� � � �� � �� ) � ������� * � � � � �� +,-. � � � � � �� �������� +,�. ��� �� ������ * � �� � ����� �������� / 0 � * � �������� ����� �� �������� � ��� * ��� ��� �� �� ) �� � ����� � ���� � ,�1��� / '� �� � � ���� �� ��*� � � � * �� � �� � ��������� ��� �� ��� � ,-1��� ���2�� / � �� ��� �� ��������� ��� (� �� � �� �� �� ���� �� 3� � �� ���� �� ����� * ��� � � �� � � �� � ��� �� �� � �� ���/ '� * ��� � ��� � ���� ��*� � ���) ��������� ����� �� ��*� � �� � * ��� ���� �� ���) � � �� �� �� �� ���� �� ���� (� �� � �� �� ��������� ��*� � ���� �� � ,� ��������� ����� �� ����� � � �� �� � ,-1��� / � "##$ 4���� � � �� 5� �� � ��� � 6 �/ (��) ���* , � � � ��% , � � �� �������� % �� � ����� �������� � ���� ������ ! � ���� � � ���� ��� � � � ������ ���) � ��� ���� � ��� ��� �� �� � � �� � �� ) � ���� � ���� ��1��� �� � � ��� � � ��������� � ���� � � �� �� ) �� �� ����� � * �� � � �� � � ��� � �� � �� � ��� � ��� � �/ ! � � ���) � ���� 7��� � �� � � �� ����� +$889. � ���: ��� * � ��� � � �� ;����� ������� � ������ � � � ,���� +"###. � �� ) ��� �� � � � � � ��������� ����� �� � ��� � �� � � �� ������ 7��� � � � � +"###. � �� ������ ���� � � ��� �� � � �������� ��� �� � ��� / �� � �� �� ���� � � �� ��:����� �� ���� � ��� �� � �� �� � �������� �������� �� � � �� � � ��� � � ��� ���� � ��� � � � ��� � �� �� ���)/ ��������� ������ ��������� ��� � ��� � �� � ��� � � �� ������� �� � � � � �� � ����� * ���� ����������� �� �������� ���� � ��� � � * �� ������� �� ��* (��� �� ��/ �� � � +"##$. �� ���� ��� ��� � �� � � * � � �� �� � ���� 5 *� � ��� � * ���3 �� ������� � � '/ !�� �� � ���� � ���� ! �� � $# +"##$. <=>?? $#@=1#9$#a#$ab > ��� � ���� � � "##$ 4���� � � �� 5� �� � ��� � 6 �/ 466� � $ # @ = 1 # 9 $ # + # $ . # # # 9 @ 1 < � ���� �� �� * ������/ ' �/� �$18$"1c9$1#<=8% ��(� �$18$"19=$1$9<@/ �+���, �������* :�� �� d*� ��/ �� +�/ ��� �� ./ 4��� � ����� � �� � ������ "#��� "##$ ������� � ��� ����� ���� � ee! ��� � � ��� � '��� f � ! ����� -�� �33 � �� -�,, ������ . ����,/ � ��� � ���� � � � � � � ���� � � �� ������� �� ��������� � �� ���� ����� ��� �� � ��� � � ��� �� � �� � �� � � � � �� +,-. �� � � � �� �������� +,�./ '�� �� � �� � � � ������ ��� �� � � � ��� �� 0 �� � ��� �� � � / �����1 � ��� ������ � �� � ��� ������� � � ���� � ������ ���� � ����� �� ��� / �� ����� ����� ,�)���� � � g� �� � ���� ��� �� �� ������ � ����� �� ��� / ����� �� ��� � �� � *� ����� ������� * � ����� � ��� � � � � �� �������� ��� �� � ���� � � � �� ��� ��� / g� ���� � * � �� � ,-1 � � ,�1��� � � ��������� � � ���� � * ��� ���� ������� �� � �� �� � � �� � �� )� � � � � �� � � ����� � ����� � ���� �� �� � � � ��� � � � ��� / 0 � � ,�1��� � � �� � �� ) ���� � ��������� � � ���� � ��� � � � ��� � � �� * � �� ��������� �3 � � �� * ���� � � �� ���� ���� �� �� ���) �/ 0 � � ,-1��� �� ����� �a �� �� � �� �� � � �� � �� )� � � ��������� � � �� �� ee����33 �����*� ���� ��� � � �� ���� ������� � ������ � � � ��� � � � ����/ ������*�� ���� � � (� � � �� �� ��� � �� � � � � � � ����� * � �� *� ������������ � � � ���� � ���� � ��� ��� � ���� � �� ��� ���� � ��� � � � ��� � � � � �� � �� ) �� �� ��� / 6 ��� ��� �� �� � ����� � ���� �� ��� � ������ � � �� ������� * �� �� �� �� �� � � � ������ ����� ���) ��� � � ������ �� � ��� � � � ��� ���� � ,-1��� � ����� �/ � �� �� � �� � 1����� �������� �� � � �� � � ��� ���� � ,�1��� � �� ���� �/ �� ����� �� ��� � �� ������ * �� � ��� � � �� � 2�� �� �� ,-1��� ���� �� ,�1��� �� � � ��� �/ � ���� ���� � ������� � �� �� ��� � ;��� ����� �� �� �� �������� �� � � � � � �� � ��������� � � �� � � � � ,�1��� � � � �� � �� � ��� �� � � � �� � �� �� ��������� � � �� � � � � ,-1��� / '� � *� � � ������ �� � � ��� ����� � �� �� ���) ��� � � �� � �� ��* �� �� ������ �� � �� �� � �� ������ �� � � �� � ����� � �� � � �� � � ��� �� ���� � � � ���� * �� �� ,�1��� �� � 2�� / � ������ �� � �� � ��� ����� � �� �� � ��� � � +�� * ����.� �� � �� �������� � � � � �� � �� ���� * ��:� �� � +�hf� . ��� � 2���� � �� � � ��� �� ���� � � � ���� * �� �� ,-1��� �� � 2�� / '� ���� ���� ��� � � �� � ��������� � �� � ����� * ����� �� b?#�###� ���� � ��� � @i ����� � ������ ��� ��� � ��� +����� � ����� � ���� @i ���� �� �� �� � � � ,�.� ���� ��� � ��� � �� � �� <i � ������ � � � (� �� � ���) ��� ��� � � �� ��� � ��� �� <# �� <@ � �� / &�� ���� ��� ��� �� �� � ;��� � �� �������� ���� �� ��������� �� ������*� �� � ��/ 6 ��� ���� ��� �� ������ * ���� � �� �� ����� � ���� / '� ���� ���) ����� � ��� � �������� � 1�� ��������� ��� � � * � �� ���� �� �� �� ������ / '� � ������ �� � � �� * ��� �� ,�1��� �� � � 2�� �� �� � ��� � � ��������� � � * �� �������� ��$� ����$���� � ,�1��� �� � � ���� +$. �� � �� ���� � � �� ���) �� �� � *� � * ������ �� *����� ��� � ����� � � ��� ����� � ������ ��� �� ��3 ������ * � �� ��* �� �������� / <9 (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& '� � � ��� �� ,-1��� � �� � � �� � ��� (� ��� �� �� � ��� � � �� ��� � <# �� "@ � �� � ����� ���� �� � ���� �� � �� �� � ��� � � � ��� � ���� �� � ����� ���� � �� � � � �� � �������� � � "i � �� � � � �� �� ���� �������� � �� �� �� � � ��� �����/ �� �� �� ��* ���� ����� � �� �� � ��� ��������� * �� ,� ��� ���� ��� �� ���� ����� � / 6� ��� � �� �� ��* �� �� �� � � � � �� � �� �� �� ��*� � �� � * � �� / '� �� � � �� � � � � � ��� ���������/ '� � ��� � �� � � �� � �� � �� � �� <i � � �� ��� �hf� ���� � �� � �� �����(���� �� ������� ��� �� � 2���� ��� � ��/ '� ) � �� (���� � ���� � ��� �1��1� �� �� �������� � � ������ �� � �/ � � �� � � � � �� ��� �� � 2�� �� � �� � � ��� ���� �� ���$� �� � ,-1��� �� � ���� +". �� ��� � �� � �� ���)� ��� � �� �� ��* ������ �� � �� ��* ��� ��� � � � � �� ������� � � �� � 2���� ��� / '��� $ �� � �� � ��� �� � 2�� ��� �� ��� � ���� / � ���������� ����� ����� h � �� � � �� � �� � 2�� � � � ��� ��� �� � ��� �� �� (� � � � �� ��� � �� ��������� �3 ��� ���� � � �� � � ��� � � � � ,�1��� / ��� ���������3 �) �� � � �� � ���� �� � �� � �� ��� � � ��� �������� % ���) � � ��� � �� � ��������� �� ���* ��������j���� ���) � � �� � � � ��������� �� �� *1� �� � � �� � *��� �������� �� � / '��� $ � ��� �� � 2�� <# � �� ���)� <# � �� � ��� � <@ � �� ���)� "@ � �� � ��� � 7 �� �� � 2�� 7 �� �� � 2�� $ ?### $ ?### " ?$"# " ?$"# < ?"?? < ?"?? "8 8$@" <? $#c#8/<? <# 8?"c <@ $#8"=/c" <$ �@c@@= <c �=c?8</< <" �@9"@? <= �=9=99/$ @8 �$"8<89 @8 �$@#8cc c# �$<<"9# c# �$@@?8@ '�� ���� �� � �� � ��� �� � 2�� ��� �� ��� � ���� / '� �� ���� �� � 2�� � �� � ��� � ��3 � 2�� �� �� ,�1��� ��������� � � � � ��� ��$� ����$���� � ,�1��� �� � � ���� �� � �� ���� � � �� ���) �% �� � *� � * ����� +b?#�###.% �� *����� ��� � ����� +<i.% ��� ����� � ������ ��� �� ��3 ������ * � �� ��* �� �������� +@� @ � $#i./ '� *���� �� � 2�� � �� � �� ����� � � �� ���� � ,-1��� � � � � ��� ���� �� ���$� �� � ,-1��� �� � ���� �� � �� ��� � �� � �� ���)� ��� � �� �� ��* ������ �� � �� ��* ��� ��� � � � � �� ������� +"i.� �� � 2���� ��� / ��� ��� � � ���� � <i �hf� ��� � ��/ (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& <8 � ��� � ��� � ���* ��� ���) � � �������� �� � � � � ���� � �� 6���� � +"###. �� � �� � ���� $8@# >$888� �� � � �� �������� ����� �� �� ���� ����� ���� ��� +� � � ����� � � � ������� � . �� �� � � / '�� � ����� �� �� � gg66 ��� � � � � �� � �� �� �� �� * ��*�� � ���� �� ���� ����� ���� ��� �� �� � ��� � � � �� � �� � � ���� * �� ���� ���� ���� �� �� �� �� �� ���� � � ��� � gg66 ������ � �� ����� �� �� ���/ ����� ���� � �� � � �� � ���� �� � � 3 � ��� �� ������� �� ������ � � � �� ���� ���� ��� �/ 0 � * �� ���� ����� � � � ������ ����� �� �������� �� ���� ���� ��������� �� � � �� �� �� � � � � ����� � � � ������� � ������� ������ * �� �� ������ �� �� �� ������� ���� �� ��/ k�� �� �� �� � � � � ��� � � �� � ������ �� �� :�� � �� �������� � �������� �����/ 6 �� �������� �� ���) � � �� � � ��� ��� � � �� $?/9?i � � c/?"i� � � ����� � �� $c/c$i � � $#/"=i� � � ���� ��� � � � ������� � � �� � ��� � ��� �� ?=/"=i/ '� ����� � � � ������� � �� ����� ��) ���� �� �� $8@#>$888 � ���� � �� � � / �� � ����� �������� �� � � �� * ��� c# ���� �� � ��� � ��� ��� ���) � � �� � ���� ���� � �������� ����� �� �������� � �� ���� ���� � �� �� * � �� ��� ��� �� ����� ��*�� � ���� �� � �� (� c# � �� / 6 �� (� � � � � ����� ��������� �� "#i ���) l9#i �� � � ����� � �� �� ������ � � ��1��1� �� � ��� ���� / 6 �� �� ��� � � �� ������ ���� �� �� ��������� � � ���� � � ��� � �� �� ��� "#i �� �� � ���� � � � � � ���) � � 9#i � � � � � �� � / '� � ���� * ��������� � ��� � ����� � ���� �� ���� �� �� * �� � 2�� �*�� �� � � � '��� $ +�� � ��� � ,� �� � � 2�� ��) �� ��������� ���� *��� � � ,�� � ���2�� ��) �� � ��� . �� �������� ��� ���� �� ��������� � ����� ��� � ��/ '�� ���� ��� � � � �� � $�### ��� � �/ /� �� � �� $�### �� ����� �� �� ��������� �������� � ��� ��� �� ��� ���)1� ��� � � � ���� � +�. <# � �� �� ���)� <# � �� � ��� � � � +��. <@ � �� �� ���)� "@ � �� � ��� �/ � �� ����� � �� �� � � � ��� �� �� ��������� �� 3� �� ��� �� �� � �� � �� � ��� � � ��� � � ����j� / '� ����� ��� � �� ��� �� � ��� $�### �� ����� � � � ��� � ��� �� "#ia9#i ��������� / 6 �� �� �� � � �� ����� � � ��� �� �1 ��1� ��� � � � ��� � * ��������� � ���� � �� �/ 6 �� �� �� � � �� ����� �� �� ��� �� �� � �� � ����� �� ��������� � (��� � � � �� �/ & (�� �� ��������� � ��� * � �� <#i ���) � =#i �� � � � �� ��� ���� � � � �� � ���� �� � ��� ��� � �� * ��������� ���� � � � � �� (��� � � � � * ��� �/ '� � � ���� * ������ � ���� �� �� ��� * �� :� � �� ��������� � �� � ���� �� �� �� �� ���� � � ��������� � �� ��������� � ��� / '� �� ��������� � ��� * � �� � ��� � $#i� ��/� �� �� �� * �� �� � ��������� �� 9#i ���) � "#i �� � / '� ��������� �� ��� * � � � -�� � � ���� � � ��� � +$89c. ��� � ���� ��� ��� 8#i �� ������ � � �� � � ��� �� � � � � � �� ��������� ��������� ��(/ ��� ;� � � ���� � �� � � � ���� �� � �� * / ��� (���� � �� ��� � )�� � � !��� � +$88?. � � ��� � )�� �� � � !��� � +$88=. � � ���� �� � � ��� � �� � � � � �* ���� � ���� �� �� �� � ��� � � �� � � ;���� / � ������� '� � �� * ��� �� ���� � � ���)a�� � ��������� ��( ��� �� <# � �� �� ���)� <# � �� �� � ��� � � (���� �� � ���� � � '��� " � � </ ?# (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& '��� " �� � �� � �� * ��� �� �� ����� ���� �� ��� �� � <# � �� �� ���) � � <# � �� �� � ��� � �/ 6� ��� ���� � �� ��������� �� ���) � �� ��������� � �� � � �� *� �� � � �� ����������� ���� �� � � � �� �������� ��� ���� � ��� �� ����� �� ���2�� �� �� � � � � �� ��� +� � �� � � � � �� ��* �� ��� ./ ��� (���� � �� � ��� �� �� @#a@# ��������� � ����� �� � � � c8i ��� ��� ���� �� � � � �� �������� ��� ���� ����� �� � ��� � � ���2�� ��� �� � � � � �� ��� / '� �� * � ���� �*�� ������ �� �� * �� �� ��� � ��� �� ��������� ���� �� �� � �� �� � ��� � � � ����� � ��� �� <# � �� � ��� ��� ����� ���� ��� ��� ����� ����� � �� � � � � �� ��� ���2�� / &�� ���� �� ��*� � ��� � ���� �*�� �� � ������� �� �� �� �� ���� ����������� �� � � �� � * ��� �� �� � �� �� ����� � ��� �� ��) �� �����*� �� �� � <# � �� �� � ��� � � ��� �����/ 6 �� @#a@# ��������� (���� � �� �� ���� �� * � ���� � b"@�#"=� ���� � ��(���� �� b$$c�c@#�"c< � � � � ��� ���� �� b=�#@=�<@=/ ��� �� ����� *� �� ��� ��� ���� � � �� � �� ���� �� � �� �����/ ����� +"###. �� ���� �� ��� � �� ��� ���� � � �� � ����� / k�� ��� ��� �* � �� '��� " ���� � <# � �� ���) m <# � �� � ��� � ,� �� � 2�� � � �� (� � ,�� � 2�� �� � � ��� � � � ���� ��������� ���) a�� � ���� ��� �� ���� �� * � ���� ��(���� �� * � ���� � ��� �� * � ���� �� �� � � � ���� "#a9# $@i b""$@? b$"?#"999 b$c"@@=8 $"i� $9i <#a=# <$i b88?#< b"9c8""<? b"c<#$?< "=i� <@i ?#ac# @"i b<=?8 b@89=<#9c b<=9#8$$ ?9i� @ci @#a@# c8i b"@#"= b$$cc@#"c< b=#@=<@= c@i� =<i c#a?# 9#i b8?8?9 b"$=?##@<9 b$#==@=?8 ==i� 9<i =#a<# 9=i b$@$?c8 b<8cc88$8? b$c=$@c"9 9?i� 98i 9#a"# 8#i b?#=$?= b=?8c<@8=" b"@"88$== 99i� 8"i '� ��� ��� � ����� �� � �� ��* �� �� �� � �� ,� � ��� � � � �� � � �� (� � �� ,�� � 2�� ���� * � ��� � �/ '� �� ����� ��(���� � � � ��� �� * � ���� �*�� � �� � �� � � � �� � ���� ����� � �� �� ��� ��� ������ �� �� �������� / '� �� �� � � � ���� �� �� 88i 61���� ��� �� ��������� ��������� / '��� < ������ � <# � �� ���) m <# � �� � ��� � ,� �� � 2�� �� �� � � ,�� � 2�� �� � � ��� � � � ���� ��������� ���) a�� � ������ ��� g�� � �� ��������� ��� � ��� ��������� ��� "#a9# 9@i ? �� $< �� <#a=# c8i @ �� $@ �� ?#ac# ?9i c �� $c �� @#a@# <$i = �� $= �� c#a?# "#i c �� $= �� =#a<# $<i c �� $= �� 9#a"# $#i @ �� $9 �� '� ������ ��� � ����� �� � �� ��* �� �� �� � �� ,� � ��� � � � �� �� �� � � �� ,�� � 2�� ���� * � ��� � �/ '� ��� � � � � ��� ��������� ��� � �� � �� � *�� �� ��� �� ����� �� ,�1��� �� ��� �� �� �/ (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& ?$ ����� � � �������� �� � �������� � �� �� � � � ����/ '� ��� � �� ��� +� � ������ . �� � ��� � � � � �� ����� �� �������� + ��� �� ������ ./ g��� � � �� ��� � �� ����� � � � � �� ��������� ��������� � �� ���� �� � � ��$� �� �� *� �� � ��� �� ;��� �� @ +g �� � $88$� ��/ "@$.� �� �� ����� �� �������� � � *��� �����(������ �� �� �� �����/ g��� � �������� �j �� $�### � � �� � � ��������� �� � � � �� �� �������� � � �� ���� ���/ g��� �� � ������ �� ��������� �� ������� ��� �� �� �� � � � ���� ����� �� �� ��������� ��������� � � �� 7 ��$� ��$�" �$�" � �� ��� � � � ���� +<. �� � �� �� ���� ��������� + ��� � �� ��� a ��� � �� ����� .% 7� 71���� ���� ���� �� � �� �� � �� � � � � �� �� �� � % �� ��� � �� ����� / k�� �� �������� �j +�. �� $�### � � � ��� * � 88i �� �� � � � ���� ����� ���� �� � �� � 71���� �� "/@9� 5;/ +<. � ��� ��� �� "�@9���$� ���$�" �$###�$�" � 88� �� ��� � � � ���� +?. f��)� * �*�� �� �� @#a@# ��������� � �� ��� ��� +�. � c8i ���� � 88i �� �� � � � ���� �� c@/"i>="/9i/ '��� < �� � ���� ��� �� �� ����� ���� �� ������ / 6 �� @#a@# ��������� �� �� ������ ��� � <$i� � � � * ���� ��������� �� �� �� ����� �� � ��� �� ����� �� � � � � �� ���2�� ��� �� ��� <# � �� / '� ��� � �� � ���� ��� �� @#a@# ��������� �� � = � �� � �� ���� � � � ��� ����� �� �� �� �� � � � * � � � � �� �������� ��� / '� � ��� ��������� ��� ��� ���� ���� �� ������ �� ��� �� �� � ����� �� �� $=�� � ��/ 6 � � �� *��� ��� � �� � ��� ��� �� �� ��� � �� ��� �� ������ ���� ������ � �� � ���� �� ��������� �� ���) � �� ���������/ '��� �� � ��� � � ��������� �� ��� �� �� � � �� �� ����� �� ��� �� ��������� ��� ���� ���� �� � � �� � �� ) � ���� �/ '��� ? ���� � <@ � �� ���) m "@ � �� � ��� � ,� �� � 2�� � � �� (� � ,�� � 2�� �� � � ��� � � � ���� ��������� ���) a�� � ���� ��� �� ���� �� * � ���� ��(���� �� * � ���� � ��� �� * � ���� �� �� � � � ���� "#a9# "=i b$@?@? b"<c=#?<9 b""?c$<9 "<i� <$i <#a=# @"i b$=$8c b<?@==@"9 b"8?"=c$ ?9i� @ci ?#ac# =@i b$=$== b@c$9c$@@ b?8##?9@ =$i� =8i @#a@# 99i b"@"@8 b$#"9$#8<" b==9?"8# 9@i� 8$i c#a?# 8?i b$@@@9c b$==8=9?cc b$<#<$?#$ 8"i� 8ci =#a<# 8=i b?#@#<@ b"8=#=?c<= b"#88=9#< 8ci� 89i 9#a"# 89i b$<""#9 b?9"8@@==? b<"??@<8? 8=i� 88i '� ��� ��� � ����� �� � �� ��* �� �� �� � �� ,� � ��� � � � �� � � �� (� � �� ,�� � 2�� ���� * � ��� � �/ '� �� ����� ��(����� � � � ��� �� * � ���� �*�� � �� � �� � � � �� � ���� ����� � �� �� ��� ��� ������ �� �� �������� / '� �� �� � � � ���� �� �� 88i 61���� ��� �� ��������� ��������� / ?" (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& '��� ? � � @ �� � �� �������� � ��� ��� <@ � �� �� ���) ������ � �� "@ � �� �� � ��� � �/ '� � ��� ��� ���� � (� �� �% � ���� � ���) ��� ���� � �� * � � ��� � � � ���� ���� ��� ��� �� �������� �� ,-1��� � � �� �� � ��������� � � ����/ � �� * � ���) ��� � � ���� � � ��� � � � ���� * ��� ��� ��� � � �� ���* � �� * � ����/ �� � ��� � � �� ����� �� � ����� �� �� �� � �� ���) �� �� �� � � � �� �������� ��������� � ���� � � � � � ����� � �� � � ���: �� � � ��� � � � / � ������� � �� '��� " � � ? ��� ���� �� ���)� * �� ��� � �� �� ��� ��� � �� � ��� ��� � � ��������� / ��� (���� � �� @#a@# ��������� �� ��� ��� *� ���� c8i �� 99i/ '� � � � � �� ��� ���� �� ��� ����� �� ����� * �� ��� � �� � �� ���) � � � � � � ��� � * �� � ��� � � � �� / ��� (���� � �� �� ��� ���� � �� ��� � �� � �� �� <#� � ������� �� � ��� � � � �� ���� � �� � �� ��� ���) ��� ��� <# � �� / f� ���� �� � ��� � ���� �*�� � ����� ���� �� ��� � ����� � �� �� � � � ��*� ����������� �� � � �� � * ��� �� �� �/ � ���������� '� ��� � ��*���*�� � � �� � �� ) � ������� * ,-1 � � ,�1��� �� ������ * � �� � ����� �������� / '� � � �� � �� ) �� � ,�1��� � �� ��� * ��� ��� � +$. ���� �� ��*� � � � * �� � �� � ��������� ��� �� ��� � �� ,-1��� ���2�� � ������������ ���� �� ��� � ���) ��������� �� "#a9# � � <#a=# � � ,�1��� � � � +". ���� � *���� �� ��������� � �� ��� �� ��������� (� �� � �� �� �� ���� �� 3� � �� ���� �� ����� * ��� � � �� � � �� � ��� �� �� � �� ���/ '� � � ��� �� (� �� * � ,-1 ��� 3 ������ �� ��� � ��� � �� ��*� ���) ��������� ��� � � =#a<# � � 9#a"#% ��������� ���� ����� �� ��*� � �� � * ��� ���� �� ���) � � �� �� �� ���� �� ���� (� �� � �� �� ��������� ��*� � ���� �� �� ��������� �����*�/ �� ����� �� ��*� � ��� � ���� �*�� �� � ������� �� �� �� �� ���� ����������� �� � � �� � * ��� �� �� � �� �� ����� � ��� �� ��� ����� ����� �� ,-1��� � �� �� � �� � ��� � � ����j� ��� ��� � ����/ '��� @ ������ � <@ � �� ���) m "@ � �� � ��� � ,� �� � 2�� �� �� � � ,�� � 2�� �� � � ��� � � � ���� ��������� ���) a�� � ������ ��� g�� � �� ��������� ��� � ��� ��������� ��� +�� . "#a9# =<i ? �� $< �� <#a=# ?9i @ �� $c �� ?#ac# "@i @ �� $= �� @#a@# $"i c �� $= �� c#a?# ci c �� $c �� =#a<# <i = �� $< �� 9#a"# "i c �� $$ �� '� ������ ��� � ����� �� � �� ��* �� �� �� � �� ,� � ��� � � � �� �� �� � � �� ,�� � 2�� ���� * � ��� � �/ '� ��� � � � � ��� ��������� ��� � �� � �� � *�� �� ��� �� ����� �� ,�1��� �� ��� �� �� �/ (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& ?< ���������� -�� � � k/ 4/� ���� f/ !/� m ��� �� k/ f/ +$89c./ , � ��� � � �� ��������� � ������ � � ����� ��, #��,���� . ����,� &2+?.� <8>??/ �� � � � �/ +"##$./ ! ��� � � ��� � ���� � � � ����� ��� � k ��� * k�� * ����� � -�,, ������ . ����,/ ������� � k/� ������ � �/� m ,���� � ,/ +"###./ !� ) ��� �� � � � � � ��������� ��� � � ��� ��� * � ��� � �� ����� ��, ����� �� 0����)� 8� $@8>$=#/ �� �/� ��� � )�� �/� m !��� � � �/ +$88?./ � � ��������� � ��� (� ��� ��� � � ��������� ����� ��, ����� �� 0����)� $� $#8>$"c/ 6���� � � ����� +"###./ �� 9�� " ���� "�,,�� ��� ��:��� �� �����*�� 6��/� !/k/ 6���� � � ����� / ��� � )�� �/� �� �/� m !��� � � �/ +$88=./ � � ��������� ��� �� �� ����� �� �� � (�� ��� �� ��� �� ����� ������� * ���� �� �� 0����) ;����������� ����� � ��� # ������� 8� @<>=#/ �� ������ �/� 7��� 7/� m � �� �/ +"###./ , � ��� � � �� ��� � � ��� � � �* � ����� ��, ����� �� 0����)� 8� $>$@/ ������ �/ +"###./ h����� 1�� � � � �� * ���� ���� � �� � ����� ��������� � ���)� * ��� �/ �� � ���� ! ����� �� � ���� 5 *� 6 �/ �������� � �����aa���/������ /�� / ��a���a���/����/ g �� � !/ +$88$./ <��� �� �� � � "������� �������� �/ ,��� 4� � �����*�/ 7��� 7/� �� ������ �/� m � �� �/ +$889./ � � � � � ��� ��� ���: ���� �� � ��� � � �� ;����� ����� ��, ����� �� 0����)� 4� $=@>$8</ ?? (� . ���� � �� �,� / ����� ��, ����� �� 0����) % 12%% 3 $45&& http://prc.wharton.upenn.edu/prc/prc.html a comparison of state university defined benefit and defined contribution pension plans: a monte carlo simulation introduction cash flows simulation methodology results conclusion references pii: s1057-0810(99)00020-7 do dividend reinvestment plans contribute to industrial firm value and efficiency? andrew saporoschenkoa,* acollege of business administration, university of akron, akron, oh 44325-4803, usa abstract this paper examines four hypotheses to test whether industrial firms offering drips are structured to operate efficiently, given that drips are a cheap source of outside financing for a firm. the hypotheses are derived from the corporate finance literature and partially based on a review of the previous drip-related finance literature. evidence is found that larger firms will be more likely to offer drips, which supports a bookkeeping hypothesis. no difference in valuation is found between drip firms and the other industry-matched firms using a tobin’s q proxy. several cash flow measures are found not to be higher for drip firms. © 1999 elsevier science inc. all rights reserved. 1. introduction dividend reinvestment plans (drips) allow investors to purchase additional shares of stock with reinvested dividends. the popular media encourages and popularizes individual investor interest in drips. for instance, vita nelson’s moneypaper newsletter, tv commercials, and web site prominently tout drips as advantageous to individual investors. the moneypaper web site (http://www.moneypaper.com) has the following statement, mainly in regard to drips: “by utilizing the information we make available, the small investor finally has a fighting chance. our commonsense approach helps you save on fees and commissions, invest in quality, and take advantage of strategies to reduce risk.” the motley fool investor web site (http://www.fool.com/foolfaq/foolfaq0015.htm) offers the following comments concerning drips, “basically, drips are an inexpensive and simple way to increase your * corresponding author. tel.:11-330-971-6331; fax:11-330-972-5970. e-mail address:asaporo@uakron.edu (a. saporoschenko) financial services review 7 (1998) 273–289 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00020-7 equity position in some really fine companies.” the american association of individual investors distributes a yearly drip guide to its members. this study examines whether drips offer additional value to investors. matched industry samples of drip and non-drip offering firms are compared based on several variables including a firm valuation measure—tobin’s q. this study also investigates other factors that might affect the value of drips to investors by examining how corporate capital structure and free cash flow relate to the probability of a firm offering a drip and how the agency conflicts between management and shareholders are influenced by drips. thus, various firm operating characteristics are examined to determine whether manager incentives for firms offering drips are aligned with those of investors and whether managers make efficient capital structure decisions. 2. drip background 2.1. description of drips an investor who joins a drip will have her dividends automatically reinvested in the purchase of additional shares of the firm offering the drip upon payment of dividends. an investor can usually choose to have only part of her dividends reinvested. discount drips allow the shareholder to purchase the additional shares at a discount to the market price. the selection of the purchase price varies, but is based on the market price of a share. often the market price on the dividend payment day or a five day average around the payment date is used. usually, the drip corporate sponsor retains the shares purchased through the drip on account. drips also usually require ownership in street name. drips are currently taxed as if dividends are received by the shareholder instead of being reinvested and thus no tax effects are proposed in this paper. two distinct types and one combined type of drip plan exist. through a new issue (or treasury) plan, a firm sells shares from its treasury stock or issues new shares. the new issue plan acts as a financing source for the firm. new issue drips are a cheap source of cash but the existence of these drips can increase free cash flow abuse by managers, thus lowering a firm’s value, and affecting the monitoring of firm management. also, new issue drips provide a continuous source of cash to firms, thus drips can be viewed as a source of financial slack. market drips (also known as repurchase drips), contrarily, provide no cash to a firm. a firm offers a market drip as a service to shareholders. shares are usually purchased by a bank trustee on a stock market or through a negotiated purchase. the sponsoring firm usually maintains an arms-length relationship in relation to these purchases. a possible benefit of these drips may be based on an extension of the dividend signaling literature. the act of paying dividends serves as a costly signal of firm value. shareholders are not obligated to reinvest the dividends through a drip and indeed many choose not to reinvest, thus indirectly disciplining management and maintaining firm value. indeed, scholes and wolfson (1989) indicate many shareholders choose not to reinvest their dividends. however, once 274 a. saporoschenko / financial services review 7 (1998) 273–289 the offer of pay out has been made, reinvestment of dividends in the purchase of new shares should aid in maintaining the price of the stock through buying pressure. possible benefits of drips to individual investors include low or no brokerage fees to buy additional shares with the reinvested dividends, the chance to buy additional shares at a discount to a market price, increased stock price due to buying pressure, and increased firm value (stock price) due to drips serving as a cheap source of firm financing. possible disadvantages of drips for individual investors are related to indirect frictions in the financing of a firm which lower firm value. 2.2. industry distribution of drips utilities and financial firms including real estate reinvestment trusts (reits) are more likely to have a drip and are more likely to offer discount drips, using the drips as a source of financing (scholes & wolfson, 1989). the industry distribution of drips is important in reviewing the monitoring structures which protect individual shareholders who lack the analysis capabilities of institutional investors. since utilities and financial firms are highly regulated, the degree of regulation appears to play a role in firms offering drips. regulators offer a monitoring function, thus lessening the need for investment and commercial bankers in their monitoring role. the regulated firms are then able to take advantage of the lower cost of drip financing yet maintain efficient monitoring of management. 3. drip literature review most of the drip literature looks at shareholder wealth (value-enhancing or valuedecreasing) effects upon drip-related announcements. peterson et al. (1987) find insignificant average abnormal returns in a day zero,11 window surrounding sec filings for new shares to be issued by non-utility firms using drips. chang and nichols (1992) and peterson et al. (1987) find positive announcement effects for announcements regarding the 1981 tax legislation for qualifying utility drip firms. roden and stripling (1996) find significant wealth effects for announcements of drips by utilities in a 15 day period before the announcements, but not during the announcement period. dubofsky and bierman (1988) find significantly positive abnormal returns upon announcements of discount drips for a 46 firm sample from 1975 to 1983 in a three-day window surrounding the announcements. perumpral et al. (1991) investigate the stock market reaction upon announcements of drip plans with the time of announcement indicated in letters from the firms offering the drips. for their entire sample, they find an average positive significant abnormal return in the month of announcement and significant positive announcement effects for market plan and original issue plan subsamples. no significant effect is found for discount drips for both first and subsequent plan announcements. allen (1991) uses the sec filing date for initial industrial firm drips between 1974 and 1987 and a control sample of initial general cash offerings to measure the wealth effects of initial offerings of drips. day zero,11 abnormal returns are significantly negative for both drips and general cash offerings though the negative reaction is significantly larger for the 275a. saporoschenko / financial services review 7 (1998) 273–289 general cash offerings, supporting an asymmetric information argument since the general cash offering are to new, less-informed investors. cross-sectional tests for industrial firms using announcement day zero,11 abnormal returns as the dependent variable indicate no significant relationship to: (1) the discount from share market price at which dividends are reinvested; (2) the number of shares indicated in an offering; (3) if additional shares can be purchased at a discount through the plan; (4) a dummy variable for a period between 1982–1985 when utility drips received preferential tax treatment; and (5) an option which allowed preferred equityholders and bondholders to participate in the drips. dhillon et al. (1992) find that announcements of industrial new issue drip initiations result in significantly less negative wealth effects than for other new equity issue announcements. the negative wealth effects are also significantly smaller for discount industrial new issue drip initiations than for the equivalent no discount initiations. these results provide evidence that drips alleviate the informational asymmetries between managers and investors. overall, the drip event study findings are ambiguous. these results are surprising given the cost savings that a firm should obtain in using a new issue or combined drip instead of investment bankers to raise capital. the findings may be due to agency inefficiencies heightened by drips or by a relatively small proportion of financing being provided by drips. another stream of the drip literature examines the new issue drip directly as a source of outside firm financing. finnerty (1989) formally models new issue drips as a source of financing for firms. of interest to individual investors is his finding that when new issue drips issue shares at a discount, participating investors gain at the expense of nonparticipating investors. scholes and wolfson (1989) view discount drips as providing an investment banking function in raising new capital for a firm. both the firm and current investors participating in the discount drips can split the underwriting costs which would otherwise accrue to investment bankers. by enrolling in drips for firms which provided discounts on share purchases with reinvested dividends and may have also allowed purchase of shares at below market prices for drip participants, scholes and wolfson are able to obtain actual net profits of about $421,000 from an initial $200,000 investment after brokerage fees, losses from a hedging program, and other transactions costs. tamule et al. (1993) investigate whether firms will offer new issue drips to raise cash. they create samples of new issue drips that are industry matched with market drips. they find the new issue drip firms have significantly higher debt-to-asset ratios and insider ownership but are significantly less likely to pay out dividends. the finding for the debt-toasset ratios supports their hypothesis that firms with a greater need for outside financing will use new issue drips. 4. hypotheses in this section, four hypotheses are examined to explain how firms offering drips are structured to operate efficiently, given that drips are a cheap source of outside financing for 276 a. saporoschenko / financial services review 7 (1998) 273–289 a firm. the hypotheses are: (1) the free cash flow-debt hypothesis; (2) the monitoring hypothesis; (3) the informational asymmetries hypothesis; and (4) the bookkeeping hypothesis. each of these hypotheses is based on previous work in the corporate finance literature. cash can be raised quickly through an existing drip. finnerty (1989) provides a salomon brothers estimate of approximately $4.6 billion raised through new issue drips in 1983. after amending its drip in april 1990, to allow a 5% discount from market price for all drip-purchased shares, south jersey industries raised $8 million dollars by june 1990. scholes and wolfson (1989) indicate that the bank of america raised $350 million in a year-and-half period between 1987 and 1989 by introducing a discount drip plan while j.p. morgan raised $92 million (44% of its common and preferred dividends for the entire year) when it offered a discount drip plan for nine months in 1985. new issue drips and, potentially, combined drips, put cash into the hands of management on a regular schedule which exacerbates the possibility of free cash flow abuse as outlined by jensen (1986, 1989). he maintains that managers will often waste free cash flow — funds a firm should potentially return to shareholders rather then waste on value decreasing (negative npv) projects. several motivations have been forwarded for this behavior including empire building and hubris. the sources of free cash flow can be either internally generated funds or externally raised funds. jensen also posits that capital structure does affect free cash flow abuse (the financing decisions), most prominently through debt, because the promised interest payments of debt force managers to pay out cash rather then spend it. the free cash flow hypothesis is well tested in the finance literature (for example, lehn & poulsen, 1989; lang & litzenberger, 1989; lang et al., 1991). management has an increased tendency to abuse cash flow from new issue drips, and possibly, combined drips, since it has this cash inflow on a regular basis rather then going to the market only when positive npv projects credibly exist. this problem should be especially prominent for discount drips. free cash flow abuse may lower the market valuation of equity, canceling most gains from receiving shares without paying brokerage fees or receiving the shares at a discount. debt bonds the firm to pay out cash and thus lowers the possibility of free cash flow abuse. if firms have high cash flow then a high debt level can overcome the tendency of managers to waste this cash flow. conversely, a firm with low cash flow will not need a high debt level to alleviate agency costs and will want to keep its debt level relatively low to avoid financial costs of distress. if firms have high levels of cash flow after dividend payments but before interest payments, than these firms should have higher debt levels if managers have a tendency to abuse free cash flow. thus, the target capital structure of new issue and combined drip firms should be maintained so as to increase efficient operation by decreasing the potential of free cash flow abuse through relatively high debt levels while still using drips as a cheap source of financing. this is termed thefree cash flow-debt hypothesis. firms with high debt levels, ceteris paribus, are expected to be more likely to have new issue or combined drips. firms with more growth opportunities are expected to have less free cash flow and thus less free cash flow abuse (and more need for outside funding). these firms should also be more likely to offer new issue and combined drips to obtain outside financing. because new issue and combined drips result in firms obtaining equity financing directly 277a. saporoschenko / financial services review 7 (1998) 273–289 from investors, they forego the outside monitoring of bankers and intense scrutiny of new offerings of equity through securities markets. if firms are to maintain an efficient capital structure, this lack of monitoring would result in a negative probability for firms to have drips. thismonitoring hypothesiswould depend on the percentage of equity and total capital raised through the drips. on the other hand, equity issuance through drips will tend to prevent a dispersion of ownership that would occur using underwritten offerings of capital, with concentrated ownership assumed to provide a more efficient monitoring function. the anecdotal evidence of higher individual investor proportion of ownership for drip firms would tend to mitigate this concentrated ownership effect. firms with high r&d expenses and sales expenses should be more difficult to monitor, since these two variables may proxy for growth opportunities which are inherently more difficult to monitor, and thus a negative relation could be found between the probability of firms having new issue drips or combined drips and high r&d or sales expenses. this effect contradicts the free cash flow effect of growth opportunities, thus the overall tendency of drip firms to have high growth opportunities is ambiguous. the management monitoring and the free cash flow-debt hypotheses are conditional on the findings of eckbo and masulis (1992) and smith (1977) that the lower direct flotation costs of rights offerings make rights offerings a substantially cheaper method of equity issuance for firms than underwritten offerings. both rights offerings and drips provide a means for current investors to invest further cash in a firm. there are tax differences between cash flows to investors through drips and through rights offerings, since drip cash flows are taxed as dividends, even though the dividends are reinvested. however, eckbo and masulis (1992) provide a strong argument that drips have partially replaced rights offerings as a means of raising cash: “insofar as drips are similar to rights offer, one would expect rights issuers to adopt drips more often than firm commitment issuers.” because drips can be viewed as a partial substitute for rights offerings, firms should realize a distinct cost saving by using drips. firms will select drips as a significant form of financing as long as other indirect costs can be alleviated to a large enough extent that drips are an efficient form of financing. in other words, only those firms which maintain an efficient or non-wealth-reducing capital structure for the purpose of alleviating the indirect costs of drips are expected to offer (and retain) drips. if combined, market, and new issue drips lower the costs of informational asymmetries, firms with a tendency toward higher informational asymmetries should be more likely to have drips. this is termed theinformational asymmetries hypothesis. scholes and wolfson (1989) propose a lowering of informational asymmetries through new issue drips. if relatively undervalued firms are willing to sell shares at a price lower than the market price over an extended period of time, adverse selection costs may be lowered since good news about the firms is substantiated over time. also, new issue and combined drips allow for periodic, steady cash inflows and are thus a source of firm slack, which should be of value to a firm (myers & majluf, 1984). tied to the informational asymmetries argument is asmall firm effect. smaller firms should benefit more from the alleviation of informational asymmetries by the use of drips. blackwell and kidwell (1988) find that smaller utility firms tend to issue private placements 278 a. saporoschenko / financial services review 7 (1998) 273–289 since they lower informational asymmetries and slovin et al. (1992) find significant positive abnormal returns for smaller firms upon announcement of both new loans and loan renewals, since bank loan monitoring also alleviates informational asymmetries. bookkeeping costs for drips may be substantial with economies of scale and scope prevalent, thus drips should be proportionately more expensive for smaller firms. for instance, legal departments in large firms may also review a prospectus for a drip. bank trustees hired for other stockholder-related activities can also provide drip processing support. thebookkeeping hypothesiswill cause small firms to be less likely to have drips, thus contradicting the small firm informational asymmetry effect. thus, the overall tendency for small firms to offer drips is ambiguous. 5. data and method 5.1. sample drip brochures (firms offering market drips do not have to offer actual prospectuses) and prospectuses, together termed as drip descriptions, were collected from drip trustees and individual firms between december, 1995 and february, 1996. the collection of actual drip descriptions allows for an accurate identification of drip types. the initial sample consists of 238 drip descriptions. the descriptions are indexed by date, if indicated, and by statements which indicate the type of plan. a typical market plan description directly states that drip shares are purchased on the open market or the plan may state the drip involves purchase of shares only on the open market (or through negotiated purchase by the bank trustee) such as the following standard boiler plate statement for the brown group, inc. drip brochure, dated march, 1990: the purchases ...... may be made on any securities exchange where such shares are traded, in the over-the-counter market, or by negotiated transactions and may be subject to such terms of price as boatmen’s [bancshares] may agree. neither brown group nor any shareholders shall have any authority or power to direct the time or price at which shares may be purchased, or the selection of the broker or dealer through or from whom purchases are to be made. plans which offered a firm the option of either issuing treasury stock, issuing new shares, or directing market purchases of its shares are termed combined plans. to abstract from industry effects, two matched samples are created for comparison with the drip firms. dhillon et al. (1992) use industry-matched samples of (1) newly-initiated discount, new issue drips; (2) newly-initiated, non-discount, new issue drips; and (3) non-drip firms, but they do not differentiate their non-drip sample by whether the firms paid out dividends. they measure differences in earnings before taxes, the price-to-book ratio, the debt-to-equity ratio, and current ratios for the year after drip initiation. industrial drip firms exhibit significantly lower price-to-book ratios and current ratios, as well as a higher growth rate in total assets. for this study, the drip sample consists of 92 non-utility industrial firms, all which were 279a. saporoschenko / financial services review 7 (1998) 273–289 in existence prior to 1994. of these, eight are new issue, 20 are combined, and 64 are market plans. utility and financial firms are eliminated due to the idiosyncratic nature of their industries’ capital structures and regulatory environments. additional firms are eliminated due to lack of compustat data. for various tests, the drip sample is divided into market, new issue, and combined plan subsamples. two other samples were also formed, composed of firms which are matched by industry with the drip firms. the div sample consists of firms who paid dividends in 1994 but did not offer a drip (there are 92 matched sets of firms in the drip-div comparison sample). the nodiv sample consists of firms who did not pay dividends in 1994 (there are 89 matched sets of firms in the drip-nodiv sample). the inclusion of both the div and nodiv samples allows for partial differentiation of dividend effects from drip effects. if a significant difference is discovered between the drip sample and the div sample but not the drip sample and nodiv sample, this is stronger evidence of a drip effect instead of a dividend effect. matched t-tests are conducted to allow for statistical tests based on comparisons between drip and non-drip firms in the same or similar industries. there are some limitations to using a matched sample. some tests such as differentiation of growth opportunities between drip and non-drip firms may not be very powerful since growth opportunities tend to be influenced by industry. on the other hand, a matched sample allows for the clearer differentiation of drip effects from industry effects, especially as it has been shown that drips tend to cluster in certain industries. to overcome some of the limitations of a matched sample, logistic regressions and valuation tests are estimated, where all drip firms are pooled together and compared to the other two samples, which are also pooled. 5.2. test variables and methodology firm valuation (q) is proxied by the sum of market value of equity, book value of long-term debt, and the book value of preferred stock, all of which is divided by the book value of total assets at the end of 1994. amit et al. (1989) provide evidence that this measure is highly correlated with tobin’s q. firm size is proxied by the natural logarithm of book value of total assets (lsize) at the end of 1993. growth opportunities are proxied by r&d-expenses-to-total assets (rdasst) and by sales expense-to-total assets (salexp). two cash flow variables are calculated. the first cash flow measure (cshflw1) used is the lehn and poulsen (1989) measure of free cash flow, normalized by the total book value of assets. a summary of specific empirical variables and estimated signs for the logistic regressions follows: cshflw1 — because interest is deducted from net income, and if debt plays a bonding role, then there should be no significantly greater probability of drip firms having a higher value of the lehn and poulsen cash flow measure. for example, a firm with high free (undistributed) cash flow before interest payments should have this free (undistrib280 a. saporoschenko / financial services review 7 (1998) 273–289 uted) cash flow significantly decreased by interest payments if the high free cash flow, high debt-to-asset level is efficient. cshflw2 — cshflw2 is cshflw1 without interest deducted. if new issue or combined drip firms tend to generate more undistributed cash flow then these drip firms should be more likely to have a higher value of cshflw2. dtoa — the ratio of book value of total liabilities to book value of total assets. high levels of debt should be more likely for new issue and combined drip firms because of the larger potential for free cash flow abuse. rdasst, salexp — firms with higher growth opportunities should have less free cash flow abuse and should thus have a higher probability of having new issue and combined drips. also, firms with high growth opportunities may lower their inherent informational asymmetries through drips. however, firms with growth opportunities will be more difficult to monitor. the realized sign on the rdasst and salexp variables is thus indeterminate. lsize — the natural logarithm of book value of total assets. informational asymmetries are higher for smaller firms since they lack extensive security analyst following and receive less scrutiny from regulators and industry governing bodies. thus, smaller firms should benefit from the alleviation of informational asymmetries offered by the use of drips. proportional bookkeeping costs of drips should, however, be larger for smaller firms. the expected sign on lsize is thus indeterminate. icn — operating income before depreciation income normalized by total assets is included as a measure of profitability/financial distress. equation one was estimated to measure the effect of drips on firm valuation. an example of a similar method to measure firm valuation would be lang and stulz (1994), who use tobin’s q to measure the contribution of diversification to firm value. q 5 a0 1 a1drip 1 a2dtoa 1 a3cshflw1 1 a4lsize (1) separate regressions were estimated where the drip dummy equals 1 for a pooled sample of drip, div, and nodiv firms. the div and nodiv samples were also pooled separately with the following: (1) market drip sample; (2) combined drip sample; (3) combined and new issue pooled drip sample; and (4) new issue drip sample. the other independent variables are chosen as control variables, which should affect firm value. although the valuation test is admittedly noisy, it provides some discriminatory power to differentiate drip vs. non-drip firm valuation. dichotomous logistic regressions were also estimated using maximum likelihood estimation to obtain the probability of a firm having a drip. the dependent variable is equal to 0 if a firm has a drip in 1994. the regressions were separated into four sections; one set of regressions is estimated with the cshflw1 variable and one set with the cshflw2 variable; also both sets of regressions were estimated separately with one of the growth opportunity proxies, salexp or rdasst. 281a. saporoschenko / financial services review 7 (1998) 273–289 6. results 6.1. sample statistics sample statistics are presented in table 1 for the industrial drip and matched industry firm samples. the average size of the drip firms ($7,674 million in total assets) is larger than those of the div ($1,733 million in total assets) and nodiv samples ($1,095 million in total assets). general housewares, the smallest firm in the drip sample ($98 million in table 1 sample summary statistics variable n mean std. dev. minimum maximum drip sample total assets 92 7674.91 14,638.92 98.35799 91,896.00 lsize 92 8.00925 1.4032916 4.5886137 11.4284128 tot. dividends 92 185.698 357.96 1.0890 2623.00 q 92 1.25756 0.69455 0.29875 4.83937 q (lagged) 92 1.37651 0.73319 0.33453 4.93527 cshflw1 92 0.0819230 0.0372664 20.0249086 0.1827844 cshflw2 92 0.0997680 0.0368283 20.0103269 0.1966496 incn 92 0.0561200 0.0693949 0.000316058 0.3650174 dtoa 92 0.6067851 0.1487477 0.2377968 0.9061820 salexp 92 0.2453796 0.1810896 nil 1.0609109 rdasst 92 0.0186885 0.0275613 0 0.1130393 dividend (div) sample—matched firms that paid dividends but did not have drips in 1994 total assets 92 1733.06 3525.21 9.97800 21,463.00 lsize 92 6.39093 1.46960 2.30004 9.97409 tot. dividends 92 31.8512 68.7107 0.07900 340.000 q 92 1.36663 1.11323 0.26004 8.76430 q (lagged) 92 1.38931 0.90379 0.00820 4.98289 cshflw1 92 0.0860348 0.0533645 20.1197285 0.2101816 cshflw2 92 0.1012306 0.0527817 20.1077531 0.2149301 incn 92 0.0791372 0.0936032 0.001687058 0.4721314 dtoa 92 0.5178858 0.1824835 0.0966009 1.0099085* salexp 92 0.2538047 0.2117171 nil 1.0122536 rdasst 92 0.2538047 0.0274004 0 0.1459930 non-dividend-paying sample (nodiv)—matched firms that did not pay a dividend in 1994 total assets 89 1094.91 2697.18 16.3770 19,486.00 lsize 89 5.91251 1.35048 2.79588 9.87745 tot. dividends 89 n/a n/a n/a n/a q 89 1.24907 0.93463 0.20176 4.97667 q (lagged) 89 1.47672 1.26612 0.08518 6.61897 cshflw1 89 0.0637312 0.0862548 20.2873609 0.3181978 cshflw2 89 0.0924216 0.1171183 20.2611841 0.3395661 incn 89 0.0561200 0.0693949 0 0.593591 dtoa 89 0.6146045 0.2628324 0.0764271 1.6058428* salexp 89 0.2395089 0.2231384 nil 1.1183610 rdasst 89 0.0274172 0.2231384 0 0.3445305 * one firm in the div sample and five in the nodiv sample have debt-to-asset ratios greater than one due to accumulated losses resulting in negative equity. 282 a. saporoschenko / financial services review 7 (1998) 273–289 total assets), is substantially larger than the smallest firms in the div and nodiv samples. the average amount of dividends paid out by drip firms ($186.698 million) is higher than that of the div sample firms ($31.851 million). the normalized undistributed average cash flow (cshflw1) for the nodiv sample is substantially smaller than for the other two samples. drip firms (.606) and nodiv firms (.615) have a higher average debt-to-asset ratio (dtoa) than div sample firms (.518). minimum values for the debt-to-asset ratio are also substantially smaller for the div and nodiv samples compared to the drip sample. 6.2. matched t-test results the most notable result is for firm size (lsize), as indicated in table 2. drip firms are significantly larger at a .01 level than non-drip firms across all plan types, except the new issue drip vs. nodiv sample, providing support for the bookkeeping hypothesis. otherwise, there is no pattern in the significance of variables to support the other hypotheses. also, there are no significant differences indicated in table 2 between the samples for q. thus, no evidence is provided that drip firms have higher market valuations based on the matched t-tests. the lack of significance for cshflw2 indicates that drip firms do not generate higher undistributed cash flow. there is also no evidence (cshflw1 is not significant) to indicate that new issue and combined drips generate substantially more free cash flow that will be abused by managers. however, an interesting question is, “do new issue and combined drip firms distribute a higher portion of dividends (normalized by total assets) than other, same industry firms, knowing that some of the cash from these dividends will return shortly to the firm?” the normalized difference of total common stock dividends paid out is significantly greater for the new issue drip sample in 1992 and for the full sample and market drip samples in 1994, but otherwise is found to be insignificant. additionally, normalized operating income (incn) is significantly higher for market drip firms when compared to matched non-dividend firms (nodiv). the debt-to-asset ratio is significantly higher for market drip firms when compared to other same-industry dividend-paying firms (div sample). because drip firms are larger in size, any found effects may be due to firm size. pearson correlation coefficients for the pooled sample of drip, div, and nodiv firms (273 observations) are calculated. firm size (lsize) is significantly correlated only with the debt-to-asset ratio variable with a correlation coefficient of 0.2645 and the normalized sales expense ratio with a correlation coefficient of20.16600. 6.3. valuation results table 3 provides the results of the valuation regressions. the drip dummy is not significant in any of the regressions. regressions estimated with only the drip dummy and firm size variables did not alter the results. thus, no evidence is provided that drip firms provide incremental increases in firm valuation. justifying their inclusion in the regressions, both cash flow (cshflw1) and the debt-to-asset ratio (dtoa) are usually highly significant 283a. saporoschenko / financial services review 7 (1998) 273–289 t ab le 2 m at ch ed -s am pl e t -t es ts fo r 19 94 da ta d r ip sa m pl e vs . d iv sa m pl e d iff er en ce s f ul ls am pl e (9 2 ob s. ) m ea n pva lu e p re -1 99 4 sa m pl e (5 7 ob s. ) m ea n pva lu e m ar ke t p la n sa m pl e (6 4 ob s. ) m ea n pva lu e c om bi ne d p la n sa m pl e (2 0 ob s. ) m ea n pva lu e n ew is su e p la n sa m pl e (8 ob s. ) m ea n pva lu e ls iz e 1. 61 83 2 0. 00 01 * 1. 52 80 7 0. 00 01 * 1. 70 40 0 0. 00 01 * 1. 36 00 5 0. 00 01 * 1. 57 85 6 0. 03 33 * in c n 2 0. 02 30 2 0. 03 56 2 0. 03 02 3 0. 04 43 2 0. 02 57 2 0. 05 94 2 0. 02 85 6 0. 22 66 0. 01 24 8 0. 65 90 q 2 0. 09 07 3 0. 52 09 2 0. 31 27 5 0. 09 57 2 0. 12 76 9 0. 51 80 0. 06 84 7 0. 62 79 2 0. 19 30 7 0. 38 08 c s h f lw 1 2 0. 00 41 1 0. 52 07 2 0. 00 20 7 0. 81 80 2 0. 00 56 2 0. 49 73 0. 00 54 1 0. 64 90 2 0. 01 58 7 0. 32 51 c s h f lw 2 2 0. 00 14 6 0. 81 53 2 0. 00 18 2 0. 83 80 2 0. 00 29 4 0. 70 92 0. 00 90 5 0. 48 35 2 0. 01 58 9 0. 32 26 d t o a 0. 08 89 0 0. 00 04 * 0. 07 30 7 0. 04 14 0. 08 12 8 0. 00 71 * 0. 13 77 7 0. 01 37 0. 02 77 1 0. 76 99 s a le x p 2 0. 00 84 3 0. 65 19 2 0. 00 46 9 0. 82 47 0. 01 26 8 0. 50 80 2 0. 08 44 8 0. 14 72 0. 01 28 6 0. 77 18 r d a s s t 0. 00 13 9 0. 62 37 2 0. 00 15 4 0. 64 63 0. 00 09 7 0. 76 08 0. 00 10 6 0. 89 61 0. 00 56 2 0. 38 95 d r ip sa m pl e vs . n o d iv sa m pl e d iff er en ce s f ul ls am pl e (8 9 ob s. ) m ea n pva lu e p re -1 99 4 sa m pl e (5 5 ob s. ) m ea n pva lu e m ar ke t p la n sa m pl e (6 1 ob s. ) m ea n pva lu e c om bi ne d p la n sa m pl e (2 0 ob s. ) m ea n pva lu e n ew is su e p la n sa m pl e (8 ob s. ) m ea n pva lu e ls iz e 2. 02 70 2 0. 00 01 * 1. 75 03 1 0. 00 01 * 2. 05 37 6 0. 00 01 * 2. 13 03 0. 00 01 * 1. 56 48 5 0. 01 32 in c n 2 0. 03 51 6 0. 00 88 * 2 0. 03 23 2 0. 04 89 2 0. 03 91 0 0. 02 40 2 0. 02 44 3 0. 31 24 2 0. 03 19 6 0. 45 64 q 0. 03 49 1 0. 76 82 2 0. 12 07 8 0. 39 71 2 0. 02 63 3 0. 87 53 0. 22 36 3 0. 05 68 0. 03 00 7 0. 85 81 c s h f lw 1 0. 01 85 9 0. 07 82 0. 01 73 9 0. 19 36 0. 01 08 5 0. 32 17 0. 04 34 0 0. 13 30 0. 01 55 7 0. 74 04 c s h f lw 2 0. 00 76 7 0. 45 33 0. 00 50 9 0. 69 59 2 0. 00 27 9 0. 78 94 0. 03 96 8 0. 15 88 0. 07 45 2 0. 87 20 d t o a 2 0. 01 11 1 0. 71 39 2 0. 05 43 5 0. 17 93 2 0. 02 51 6 0. 53 84 0. 03 55 6 0. 41 42 2 0. 02 06 9 0. 80 08 s a le x p 0. 00 73 6 0. 73 28 0. 01 43 7 0. 59 78 0. 01 13 2 0. 68 41 0. 02 86 1 0. 44 702 0. 07 59 5 0. 27 59 r d a s s t 2 0. 00 82 8 0. 12 53 2 0. 00 68 9 0. 17 93 2 0. 01 19 7 0. 10 26 0. 00 00 9 0. 99 07 2 0. 00 10 5 0. 92 92 n o te s: d iff er en ce s fo r al lv ar ia bl es ar e ob ta in ed by su bt ra ct in g th e va lu e of th e no nd r ip sa m pl e va ria bl e fr om th e va lu e of th e d r ip sa m pl e va ria bl e. t he pva lu es ar e fo r tw osi de d tte st s. t he pr e19 94 sa m pl e el im in at es d r ip fir m s (a nd th ei r m at ch ed fir m s) fo r w hi ch th e d r ip de sc rip tio ns w er e da te d 19 94 . * in di ca te s si gn ifi ca nc e at a .0 1 le ve lo r be tte r. 284 a. saporoschenko / financial services review 7 (1998) 273–289 t ab le 3 v al ua tio n re gr es si on s— (d ep en de nt va ria bl e q ) e qu at io ns (1 ) p ar . e st .pva lu e (2 ) p ar . e st .pva lu e (3 ) p ar . e st .pva lu e (4 ) p ar . e st .pva lu e (5 ) p ar . e st .pva lu e (6 ) p ar . e st .pva lu e (7 ) p ar . e st .pva lu e in te rc ep t 1. 96 66 .0 00 1* 1. 74 80 .0 00 1* 1. 53 25 .0 14 6 1. 28 98 .0 01 5* 2. 02 71 .0 00 1* 2. 18 04 .0 00 1* 1. 25 06 .0 02 1* d r ip d um m y 0. 02 85 .8 19 92 0. 05 70 .6 80 1 2 0. 25 43 .1 06 9 0. 19 50 .2 09 8 0. 25 07 .0 96 7 0. 16 48 .3 74 2 2 0. 27 30 .1 64 5 c s h f lw 1 3. 59 93 .0 00 1* 0. 97 89 .2 34 3 1. 50 10 .4 74 9 3. 35 91 .0 53 7 0. 50 31 .5 01 0 7. 64 69 .0 00 1* 4. 77 86 .0 00 5* d t o a 2 1. 49 91 .0 00 1* 2 1. 66 79 .0 02 2* 2 1. 86 79 .0 01 7* 2 1. 32 27 .0 12 5* 2 1. 76 41 .0 01 8* 2 1. 98 99 .0 00 3* 2 1. 01 87 .0 03 9* ls iz e 2 0. 01 31 .7 24 5 0. 02 38 .5 60 1 0. 08 66 .1 08 4 0. 02 76 .6 32 5 2 0. 01 46 .7 52 5 2 0. 05 15 .3 96 2 0. 07 47 .2 02 9 ad j-r 2 .2 03 6 .5 75 4 .5 59 1 .2 38 0 .3 07 4 .2 80 0 .1 88 5 f -v al ue .0 00 1 .0 07 8 0. 00 95 .0 08 5 .0 02 6 .0 00 1 .0 00 1 h et er os ce da st ic ity te st .0 20 4 .5 11 4 .3 05 2 .8 02 8 .8 95 7 .1 21 4 .0 00 1 o bs er va tio ns 27 3 16 16 40 38 12 8 12 5 n o te s: * in di ca te s si gn ifi ca nc e at a .0 1 le ve lo r be tte r. h et er os ce da st ic ity te st is w hi te ’s te st us in g fir st an d se co nd m om en ts . (1 ) is es tim at ed fo r th e fu ll sa m pl e of 27 3 fir m s (a ll d r ip , d iv an d n o d iv fir m s) . (2 ) is es tim at ed fo r ne w is su e d r ip s an d th ei r m at ch ed d iv fir m s. (3 ) is es tim at ed fo r ne w is su e d r ip s an d th ei r m at ch ed n o d iv fir m s. (4 ) is es tim at ed fo r co m bi ne d d r ip s an d th ei r m at ch ed d iv fir m s. (5 ) is es tim at ed fo r co m bi ne d d r ip s an d th ei r m at ch ed n o d iv fir m s. (6 ) is es tim at ed fo r m ar ke t d r ip s an d th ei r m at ch ed d iv fir m s. (7 ) is es tim at ed fo r m ar ke t d r ip s an d th ei r m at ch ed n o d iv fir m s. 285a. saporoschenko / financial services review 7 (1998) 273–289 in the regressions. the cshflw1 parameters are all positively signed, while the dtoa parameters are all negatively signed. 6.4. logistic regressions results the results of the logistic regressions are presented in table 4. the firm size (lsize) parameter is negatively significant and has a low odds ratio for the full sample and certain other regressions indicating that drip firms are more likely to be large. after firm size and table 4a logistic regressions—(dependent variable—drip dummy with independent variable cshflw1) equations (1) par. est. p-value odds (2) par. est. p-value odds (3) par. est. p-value odds intercept 7.3809 .0001 999.00 2.1906 .6353 8.940 19.549 .1435 999.00 cshflw1 20.2961 .9263 0.744 2.3418 .7831 10.400 11.202 .6658 999.00 dtoa 0.9058 .3298 2.474 5.0039 .2785 148.99 6.9614 .3338 999.00 salexp 21.2287 .0875 0.293 0.3365 .9425 1.400222.471 .1102 0.000 lsize 20.9749 .0001 0.377 20.7684 .1699 0.464 23.0747 .1119 0.046 drip5 0 181 8 8 drip5 1 92 8 8 equations (4) par. est. p-value odds (5) par. est. p-value odds intercept 7.0299 .0323 999.00 6.7892 .0537 888.19 cshflw1 29.4897 .4362 0.000 28.1343 .5111 0.000 dtoa 24.8494 .1868 0.008 4.7376 .3236 114.15 salexp 0.7663 .6735 2.152 21.1770 .6863 0.308 lsize 20.4932 .1659 0.611 21.3022 .0054 0.272 drip5 0 20 19 drip5 1 20 19 equations (6) par. est. p-value odds (7) par. est. p-value odds intercept 6.1603 .0001 473.56 8.0831 .0001 999.00 cshflw1 4.5668 .3607 46.231 20.6858 .8825 0.504 dtoa 0.0582 .9684 1.060 2.3616 .0700 10.608 salexp 21.6577 .1253 0.191 21.4865 .1510 0.226 lsize 20.8507 .0001 0.427 21.3023 .0001 0.272 drip5 0 64 61 drip5 1 64 64 note: using the sas convention, a negative parameter and low odds ratio indicates that large values of the independent variables contribute positively to the probability of a firm having a drip. (1) is estimated for the full sample of 273 firms (all drip, div and nodiv firms). (2) is estimated for new issue drips and their matched div firms. (3) is estimated for new issue drips and their matched nodiv firms. (4) is estimated for combined drips and their matched div firms. (5) is estimated for combined drips and their matched nodiv firms. (6) is estimated for market drips and their matched div firms. (7) is estimated for market drips and their matched nodiv firms. 286 a. saporoschenko / financial services review 7 (1998) 273–289 other variables are taken into account, the debt-to-asset ratio (dtoa) is only significant in the market drip vs. nodiv regressions, where the debt-to-asset ratio tends to be smaller for the market drip firms. the normalized sales expense (salexp) parameter is negatively-signed with a low odds ratio in all the regressions except the table 4a regression (2) and (4) and the table 4b regressions (2) and (4). only the full, pooled sample parameter is, however, significant. this result provides limited evidence for the alleviation of informational asymmetries by drips, since drip firms are more likely to have a higher sales expense ratio and thus more growth opportunities. table 4b logistic regressions—(dependent variable—drip dummy with independent variable cshflw2) equations (1) par. est. p-value odds (2) par. est. p-value odds (3) par. est. p-value odds intercept 7.2815 .0001 999.00 1.7015 .7195 5.482 20.421 .1404 999.00 cshflw2 0.5908 .8485 1.805 3.7680 .6585 43.295 3.8327 .8782 46.186 dtoa 0.9563 .2814 2.602 5.1509 .2593 172.58 5.9254 .3722 374.43 salexp 21.2180 .0900 0.296 0.8514 .8598 2.343222.091 .1214 0.000 lsize 20.9774 .0001 0.376 20.7533 .1759 0.471 23.0240 .1209 0.049 drip51 92 8 8 drip50 181 8 8 equations (4) par. est. p-value odds (5) par. est. p-value odds intercept 7.1350 .0369 999.00 6.9394 .0650 999.00 cshflw2 29.3053 .4491 0.000 28.0790 .4977 0.000 dtoa 24.4814 .1862 0.011 4.9895 .2870 146.86 salexp 0.6822 .7096 1.978 21.2529 .6667 0.286 lsize 20.5132 .1400 0.599 21.3226 .0046 0.266 drip51 20 19 drip50 20 19 equations (6) par. est. p-value odds (7) par. est. p-value odds intercept 6.1903 .0001 488.02 7.8165 .0001 999.00 cshflw2 4.5301 .3621 92.768 1.7741 .6875 5.895 dtoa 20.1702 .9042 0.843 2.4722 .0466 11.848 salexp 21.6266 .1342 0.197 21.4893 .1479 0.226 lsize 20.8487 .0001 0.428 21.3074 .0001 0.271 drip51 64 64 drip50 64 61 note: using the sas convention, a negative parameter and low odds ratio indicates that large values of the independent variables contribute positively to the probability of a firm having a drip. (1) is estimated for the full sample of 273 firms (all drip, div and nodiv firms). (2) is estimated for new issue drips and their matched div firms. (3) is estimated for new issue drips and their matched nodiv firms. (4) is estimated for combined drips and their matched div firms. (5) is estimated for combined drips and their matched nodiv firms. (6) is estimated for market drips and their matched div firms. (7) is estimated for market drips and their matched nodiv firms. 287a. saporoschenko / financial services review 7 (1998) 273–289 7. summary evidence is provided that drips do not increase the value of a firm for investors when firms offering drips are compared to firms not offering drips in the same industry. these results hold separately when drip firms are separated into new issue, combined, and market plan samples. evidence is provided that industrial firms which offer drips are larger than other firms in the same industry, including both those who pay dividends and those who do not pay dividends. this supports the bookkeeping rationale that larger firms will tend to have drips because of economies of scale and scope. market drip firms are found to have significantly higher operating income in 1994 than dividend and non-dividend-paying firms in the same industry. firm managers appear to offer drips to attract individual investors. drips can serve as a cheap source of outside financing for firms, thus providing possible additional value to investors. however, the motives of managers for offering drips must be identified and investigated to determine if drips, overall, do contribute to increased firm value. no evidence is found to indicate that industrial drip firms maintain an efficient debt structure, based on the type of drip the firms offers, nor does the evidence in this study indicate that drips are offered to decrease informational asymmetries. however, no evidence is found to indicate that industrial drip firms generate more free cash flow than their industry-matched counterparts. since the results in this study fail to provide evidence that drips increase firm value, the dividend reinvestment puzzle of why firms offer drips remains to be further investigated. acknowledgment an earlier version of this paper was presented at the 1996 midwest finance association meeting. references allen, g. c. (1991). dividend reinvestment plans: the effects on shareholder wealth and equilibrium security returns. unpublished dissertation, university of south carolina-columbia. amit, r., livnat, j., & zarowin, p. (1989). a classification of mergers and acquisitions by motives; analysis of market response.contemporary accounting research 6, 143–158. blackwell, d. w., & kidwell, d. s. (1988). an investigation of cost differences between public sales and private placements of debt.journal of financial economics 22, 253–276. chang, o. h., & nichols, d. r. (1992). tax incentives and capital structures: the case of dividend reinvestment plans.journal of accounting research 30, 109–125. dhillon, u. s., dennis, j. l., & ramirez, g. g. (1992). dividend reinvestment plans: an empirical analysis. review of quantitative finance and accounting 2, 205–213. dubofsky, d. a., & bierman, l. (1988). the effect of dividend reinvestment plan announcements on equity value. akron business and economic review 19, 58–68. eckbo, b. e., & masulis, r. w. (1992). adverse selection and the rights offer paradox.journal of financial economics 32, 293–322. 288 a. saporoschenko / financial services review 7 (1998) 273–289 finnerty, j. d. (1989). new issue dividend reinvestment plans and the cost of equity capital.journal of business research 18, 127–139. jensen, m. c. (1986). agency costs of free cash flow, corporate finance, and takeovers.american economic review 76, 323–329. jensen, m. c. (1989). eclipse of the public corporation.harvard business review 67, 61–75. lang, l. h. p., & litzenberger, r. h. (1989). dividend announcements: cash flow signaling vs. free cash flow hypothesis?journal of financial economics 24, 181–192. lang, l. h. p., & stulz, r. m. (1994). tobin’s q, corporate diversification, and firm performance.journal of political economy 102, 1248–1280. lang, l. h. p., stulz, r. m., & walkling, r. (1991). a test of the free cash flow hypothesis: the case of bidder returns.journal of financial economics 29, 315–336. lehn, k., & poulsen, a. (1989). free cash flow and stockholder gains in going private transactions.journal of finance 44, 771–787. myers, s., & majluf, n. s. (1984). corporate financing and investment decisions when firms have information that investors do not have.journal of financial economics 13, 187–221. perumpral, s., keown, a. j., & pinkerton, j. (1991). market reaction to the formulation of automatic dividend reinvestment plans.review of business and economic research 26,48–58. peterson, p. p., peterson, d. r., & moore, n. h. (1987). the adoption of new-issue dividend reinvestment plans and shareholder wealth.the financial review 22, 221–232. roden, f., & stripling, t. (1996). dividend reinvestment plans as efficient methods of raising equity financing. review of financial economics 5, 91–100. scholes, m. s., & wolfson, m. a. (1989) decentralized investment banking: the case of discount dividendreinvestment and stock-purchase plans.journal of financial economics 23, 7–35. slovin, m. b., johnson, s. a., & glascock, j. l. (1992). firm size and the information content of bank loan announcements.journal of banking and finance 16, 1057–1071. smith, c. (1977). alternative methods for raising capital: rights versus underwritten offerings.journal of financial economics 5, 173–207. tamule, h. b., bubnys, e. l., & sugrue, t. f. (1993). dividend reinvestment plans and pecking order capital structure behavior: an empirical investigation.journal of economics and finance 17, 91–102. the moneypaper web site. available online: http://www.money.paper. the motley fool web site foolish faqs drips. available online: http://www.fool.com/foolfaq/ foolfaq0015.htm. 289a. saporoschenko / financial services review 7 (1998) 273–289 pii: s1057-0810(00)00047-0 learning by doing: offering a university practicum in personal financial planning thomas h. eyssell* umsl financial planning and counseling program, university of missouri-st. louis, 8001 natural bridge road, st. louis, mo 63121, usa abstract given the importance of interpersonal relationships in the financial planning process, it is surprising that relatively few registered financial planning programs explicitly provide students the opportunity to meet with real clients on a one-to-one basis. this paper describes the structure of a financial planning practicum developed for the purpose of providing such experience to future financial planners. it is hoped that this information will encourage others to consider offering experiential learning opportunities to those seeking positions in the financial services industry. © 1999 elsevier science inc. all rights reserved. jel classification:a22, g20 keywords:financial planning; practicum 1. introduction over the last two decades, the practice of personal financial planning has evolved from a relatively narrow specialty undertaken largely by accountants, attorneys, and insurance agents, into a profession unto itself, characterized by certification testing, professional designations, and self-regulating bodies. one thing that hasn’t changed, however, is the importance of the client-planner relationship. successful financial planners must be able to inspire trust, dig deeply into their clients’ financial histories, and convey complicated * tel.: 11-314-516-6273; fax:11-314-516-6420. e-mail address:c1803@umslvma.umsl.edu (t.h. eyssell). financial services review 8 (1999) 293–303 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00047-0 information in an understandable fashion. in short, they must be adept at building and maintaining long-term relationships with clients whose backgrounds place them at various levels of the socioeconomic stratum. teaching relationship skills is, unfortunately, more challenging than teaching a particular theory or financial model, because it requires personto-person interaction. this paper describes one institution’s experience in designing and offering a “practicum” in financial planning, the purpose of which is to provide prospective financial planners experience in dealing with actual clients on a one-to-one basis. 2. the education of a personal financial planner at the end of june, 1999, ninety-nine american colleges and universities offered certified financial planner programs registered with the certified financial planner board of standards (hereafter “cfp board”), owners and licensors of the certified financial planner (“cfp”) designation. of the programs registered, approximately half are for-credit programs leading to undergraduate or graduate degrees, and the remaining programs are not-for-credit, continuing education, or university extension-based programs.1 registration of a program with the cfp board primarily entails satisfying requirements as to topical coverage and instructional quality (as indicated by the credentials of the instructors). from the viewpoint of the student, completion of the courses in a registered program provides a “gateway” through which one is allowed to sit for the certified financial planner certification examination. (it should be noted that, in order to be registered as a certified financial planner licensee and to use the cfp mark, one must not only pass the certification exam, but must also fulfill a work experience requirement and agree to abide by the cfp board “code of ethics and professional responsibility.” special regulations apply to those with other related certifications and educational attainments.) interestingly, examination of several university-based programs reveals substantive differences in program configurations. while virtually all of the registered programs considered offer courses in the fundamental topical areas covered by the cfp examination—investment planning, income tax planning, insurance planning, retirement planning, and estate planning, a relatively small number of programs offer the opportunity for “hands-on” experience for students.2,3 3. the practice of financial planning and the need for interpersonal skills financial planners work in a “relationship business.” as such, the ability to interact positively with clients and to develop long-term relationships is crucial to the development and long-term growth of a successful financial planning practice. in describing the qualities possessed by successful financial planners, rowland (1997) states that “good interpersonal skills and the ability to build relationships” are of key importance. (p. 43) further, she suggests that the ability to educate clients (which, in turn, suggests communicative ability) is of crucial importance to the financial planner. in the words of a planner interviewed by 294 t.h. eyssell / financial services review 8 (1999) 293–303 rowland: “there’s something crucial about the educational process, aboutcommunicating concepts and ideas and helping people to learn.” (rowland, p. 44) the importance of the client-planner relationship is also reflected in the following quote from evensky’swealth management(1997): as is the case with any professional financial planner, the wealth manager’s focus is his client. his efforts are devoted to helping clients achieve life goals through the proper management of their financial resources. whereas the money manager may not necessarily know if his client is male or female, single or married, a doctor, lawyer or candlestick maker,the wealth manager will know all this, as well as the client’s hopes, dreams, goals, and fears.(p. xiv, italics added) further, the importance of relationship-building is evident in the explication of the six-stage “financial planning process” promoted by the cfp board: (1) establishing client-planner relationships (2) gathering client data and determining goals and expectations (3) assessing client’s financial status (4) developing and presenting the financial plan (5) implementing the financial plan (6) monitoring the financial plan clearly, the essence of the financial planning process is meeting with clients, gathering detailed (and sometimes sensitive) information about their personal lives, and making recommendations for changes in behaviors and beliefs that may go back decades. as such, interpersonal skills such as listening, questioning, and empathizing are as important as the analytical skills necessary to prepare, implement, and monitor the financial plan. but how (and how often) are interpersonal skills taught? a popular set of cfp examination self-study materials frequently mentions the interpersonal aspects of the job but provides little in the way of direct training in this area.4 and, as noted above, relatively few programs explicitly offer students the opportunity to interact with “real” clients in a carefully designed and monitored educational environment.5 in short, while interpersonal adeptness is an important part of the skill set required of practicing financial planners, it appears to be given less prominence than the technical skills in many financial planning educational programs. this paper describes the trials and tribulations associated with the implementation of a program developed to address this need. 4. the practicum in personal financial planning in 1995, the members of my finance department decided to offer more focused training in finance by dividing the curriculum into several “tracks”—corporate finance, investments, financial institutions, treasury management, and personal financial planning. a welcome byproduct of this curricular change has been the facilitation of study toward the attainment of professional designations by our students. we have developed links with organizations offering certification in the areas of financial analysis, treasury management, and financial planning. 295t.h. eyssell / financial services review 8 (1999) 293–303 the financial planning track has proven to be quite popular with business students–the introductory course (which serves as a “gateway” to the program) regularly attracts 50–60 students per semester during the regular academic year and another 20–30 students in the summer session.6 (this total constitutes approximately 25% of our finance majors.) the course provides students with an overview of the main topical areas of financial planning, as well as introductions to relevant regulations, planner-client issues, and the practice of personal financial planning. the centerpiece of the course is the requirement to prepare a comprehensive financial plan as a semester project. in addition to the introductory course, the financial planning curriculum includes required courses in investments, income taxation, insurance, retirement planning, and estate planning. courses are taught by a combination of regular faculty and adjunct faculty. the latter are experts in a particular field (e.g., estate planning) and are drawn from the pool of local financial planning practitioners. it is important to note the importance of creating and maintaining strong links between the educational institution and local financial planning practitioners. our institution reached out to the local chapters of the institute of certified financial planners (icfp) and the international association for financial planning (iafp).7 members of these groups have been very supportive of every aspect of our university program and have been willing to act as speakers in various courses, to hire students in internship positions, to provide scholarship and award money for outstanding financial planning students, and to serve as members of the advisory board created to help administer the practicum. in short, i would urge anyone considering the creation of a practicum course to actively involve members of the local financial planning community. 4.1. structure of the practicum in january, 1999 we offered a practicum in personal financial planning for the first time. as noted above, the practicum has two main thrusts. the first is to provide students with exposure to clients and to provide practice in information-gathering, data analysis, plan preparation, and plan presentation. secondarily, students gain experience in using the financial planning software they will find in the offices of future employers and in working in teams to prepare a polished finished product. enrollment in the practicum is limited to students who have completed the introductory financial planning course and at least one other course in the financial planning track. we also require that students perform volunteer (i.e., unpaid) service 3–5 hr per week for ten weeks during the semester at the local consumer credit counseling service (cccs) under the direct supervision of the director of counseling and a consumer credit counselor.8 more specifically, the structure of the practicum is as follows. 4.1.1. course preparation the practicum course requires that a significant amount of groundwork be laid prior to the opening of the semester. first, because of the emphases placed on community service and counseling in our program, we chose to work jointly with the local office of the consumer credit counseling service (cccs).9 early on, several arrangements must be made with the 296 t.h. eyssell / financial services review 8 (1999) 293–303 cccs. details as to the number of students that can be expected each semester (we limit enrollment to ten per course) and exactly what the student volunteers will be doing while at the cccs office must be spelled out.10 with respect to the latter, our students perform such tasks as observing counseling sessions, contacting clients, and contacting creditor organizations. at the same time, the instructor should be making arrangements for the acquisition of those people who will serve as financial planning “clients” for the student planners. in our case, the cccs director of counseling provided us with a list of former clients who had successfully completed the cccs “debt management program” (which effectively clears the client’s financial slate), met certain parameters (e.g., home ownership, current employment), and indicated a willingness to participate in the volunteer financial planning program. (in describing the program to cccs officials, we emphasized that financial planning represents a logical extension of the debt management program once a client’s debts have been brought under control, s/he should seek to plan for order and safety in future financial dealings.) it is at this time that one or more local financial planners should be approached for participation in the program. the financial planner is asked to attend the first meeting of each student team with its client, review and critique the financial plan the student team prepares, and then observe the final meeting of the teams and the clients. given that the practitioner has input throughout the program, it is very important that s/he be an enthusiastic participant. at the same time, the instructor should be willing to accept suggestions from the planner, particularly those related to the contents, recommendations, and presentation of the final plan document. finally, if it is intended that students use financial planning software in the preparation of their financial plans, it is necessary to obtain and familiarize oneself with it. 4.1.2. weeks 1–4 the early weeks of the course are largely preparatory in nature. during this period, the course meets on a traditional schedule, and students work through a case using the financial planning software. there are three benefits of this approach. first, students become familiar with the software, and learn what information will be needed to generate satisfactory output. second, students quickly become accustomed to working in teams. finally, preparation and discussion of a case analysis refreshes and enhances the material covered in the other financial planning courses, and provides the instructor with the opportunity to point out gaps in the information provided, inferences to be drawn, and additional questions to be asked. during this period, we find it useful to take students to the offices of a local financial planner, who gives a presentation on the financial planning process, with particular emphasis on the importance of trust and relationship-building. because of our belief in the importance of interpersonal skills, we also provide for guest lecturers during this period to address the students on the topics of listening skills and the development and use of effective interviewing techniques. some instructors may wish to use this period to involve the students in role-playing exercises and/or short case analyses to further hone their listening and investigative skills. 297t.h. eyssell / financial services review 8 (1999) 293–303 4.1.3. weeks 4–13 in the fourth week of class, the course meeting schedule changes dramatically. because students begin their volunteer work at this point, traditional class meetings are reduced to include only periodic updates and meetings for special purposes. in our case, students are taken to the cccs offices and given an orientation by the director of counseling and the director of education. over the next few days, student work schedules are finalized, and students are placed with counselors, with whom they will meet with cccs clients. each student observes multiple cccs sessions. this is important, because cccs clients range from the relatively well-off who simply wish to better manage their debt load, to those who are in desperate financial straits. as a result, the interviews range from being fairly pleasant, low-key discussions, to being tearful, high-stress encounters. this portion of the course greatly enhances student interviewing and interpersonal skills because they observe counselor interactions with clients of varying socioeconomic backgrounds and different personality types. simultaneously, practicum clients are selected and final arrangements for the first meeting with the students are made. the client participant is mailed a simple fact-finder and a financial goal sheet and requested to complete them to bring to the first meeting. (it is particularly important to stress to the client that his/her financial goals should be as specific as possible with respect to time horizon, level of importance to the client, and estimated cost, since the financial plan is essentially a “road map” to the attainment of the client’s financial goals.) students are asked to develop a set of questions and are coached further in information-gathering techniques. next, each student team meets with a client participant under the supervision of the instructor and the financial planner. students perform virtually all of the questioning, with the instructor and/or the financial planner stepping in only if the situation warrants it. it has been our experience that, having observed a number of credit counseling sessions by this point, students begin to have a good feel for interviewing and questioning techniques and, in many cases, to develop some self-confidence. over the remaining weeks in this period, the students take the information obtained and, with the assistance of the instructor and using the software, generate an initial financial plan for the client participant. while traditional class meetings have ceased at this point, the instructor should strive to keep close tabs via phone or e-mail on the progress of each student team in preparing its financial plan. unless one restricts course enrollment to a relatively small number, this aspect of the course can become time-consuming for the instructor. additionally, close monitoring ensures that any unforeseen problems or questions that may arise in the work environment can be quickly addressed. 4.1.4. weeks 13–14 as the end of the semester approaches, the initial financial plan is presented to the instructor and the financial planner for comment and revision. this is essentially a dry run of the eventual formal presentation to the client and is extremely useful in identifying gaps and/or weak points in the plan. students must list completely all assumptions made in the generation of the plan, and provide a set of “action steps” to be recommended to the client. in general, the plan should include recommendations which will lead to accomplishment of 298 t.h. eyssell / financial services review 8 (1999) 293–303 the client’s goals, and include sections covering the following topics: evaluation of the client’s current financial position, with particular emphasis on both shortand long-term liquidity needs; analysis of the client’s current life insurance coverage and recommendations for changes in the amount and/or types of insurance needed, if appropriate; assessment of the client’s retirement program, including plans made through his/her employer, as well as personal savings and investment; and analysis of the client’s investment portfolio relative to the return required to meet his/her goals, conditioned by the client’s stated risk tolerance.11 the instructor and the practitioner are encouraged to probe every detail of the plan at this point. this is likely the first financial plan the students have prepared (apart from that done for the introductory course, in our case), and should be considered a “first draft” of the final document. necessary revisions to the plan are made, and students continue to practice their presentation technique. 4.1.5. week 15 in the final week of the course, the client participant returns and the students present the completed financial plan under the supervision of the instructor and the participating financial planner. the client participant is encouraged to ask questions of the students, and, once again, the instructor and the planner serve primarily as observers. the presentation ends with the aforementioned list of financial recommendations or “action steps” that the client can implement should s/he so choose. assigning grades in the practicum is a bit different than in a more traditional course. first, given the nature of this course, i choose not to give examinations. rather, students are graded on the quality of their work at the cccs (on which they are required to write a brief commentary at the end of the semester), on the quality of the plan document and on the presentation of it to the client participant. also, since students are working in teams, some instructors may wish to allow them to grade other team members on the quantity and quality of their efforts in preparing the plan. 4.2. benefits, costs, and caveats i believe the most important benefits of the practicum are the client interaction and the real-world lessons learned by the student participants. first and foremost, the students learn the importance of the ability to deal with people on a one-to-one basis in a situation that is complicated by the requirement that the client lay bare his/her financial situation. on the more concrete side, student participants are likely to discover the following facts, which will serve them well should they eventually enter the financial services industry: (1) many people have only a vague understanding of common financial terms. the most basic notions of risk, reasonable rates of return in various asset classes, long-run implications of different asset allocations (particularly in the context of defined contribution retirement plans), and differences in the costs and types of life insurance policies are not well understood by the public at large. (2) few people have well-thought out financial goals. what goals they do have are often not highly specific as to priority, nor are they precise (or realistic) as to the financial resources required for attainment. 299t.h. eyssell / financial services review 8 (1999) 293–303 (3) client-provided financial information is often incomplete and/or contradictory. this may be due to several factors, including the client’s lack of understanding, as well as the client’s reluctance to totally expose himself/herself financially. students should, following their practicum experience, have a much better idea of what questions to ask and of what client-provided data to be suspicious. of course, one can lecture on these aspects of the financial planning process in the traditional manner, but it is my experience that a much stronger impression is made when students deal with these complications firsthand. in short, a well-designed practicum serves to reinforce what students have learned previously, as well as to add to their knowledge base. an additional benefit of such a program is that it is consistent with the outreach and community service components implicit in most universities’ missions–a program such as this is viewed by many external constituents as evidence of the university “giving something back” to the community and providing a useful service, while increasing the quality of the educational process. finally, the existence of a practicum enhances the university-community link by requiring joint effort between educators and practitioners. in our case, it was extremely gratifying to find that members of the local icfp society were very willing to assist in this course, viewing it as an opportunity to performpro bonowork. the benefits of such a course do not come without cost, however. we offer some caveats for those considering the implementation of a similar course. first, great care must be taken to ensure that students do not “overreach their expertise” in the financial plan document, intentionally or otherwise. we repeatedly stress to the students that any of the recommendations they make are to be general and generic in nature–no specific firm’s products or services are to be discussed. second, it can be difficult to convince people unfamiliar with the financial planning process to disclose the details of their financial lives to strangers. it is incumbent upon the practicum instructor to make clear to potential client participants that, in return for their participation in the program, they will receive, at no cost or obligation, an extensive financial plan that should greatly aid them in attaining their financial goals. i stress this point in the introductory letter sent to potential client participants. the letter also describes the financial planning process and the considerable benefits of personal financial planning for the individual. (the client pool we draw from is largely unfamiliar with the financial planning process.) early discussions should provide the client with a clear understanding of the nature and the scope of the financial planning process. of course, the client should be assured that any and all information provided will be held in the strictest confidence and will not be divulged to any parties outside of the program for any reason whatsoever. two important questions the instructor must address prior to offering the practicum are (1) how to obtain clients and (2) how many clients to obtain. as noted above, we obtained clients with the assistance of the local cccs. others may wish to obtain clients via word-of-mouth or on-campus advertising. the number of clients to serve is directly related to the number of students in the practicum. we severely limit enrollment in the course in order to ensure that the students are of the highest caliber and are likely to undertake the task with the appropriate seriousness. 300 t.h. eyssell / financial services review 8 (1999) 293–303 perhaps the principal limitation on the number of clients, though, is the level of faculty involvement required. as suggested by the course schedule above, the number of hours of classroom contact time in this course is limited largely to the first and last months of the semester. nonetheless, the time commitment required of the faculty member is fairly heavy. a substantial amount of groundwork must be laid before the course can begin, and scheduling issues are ongoing throughout the semester: student work arrangements must be made, potential clients must be contacted, and meeting dates with students, clients, financial planners, and the instructor must be coordinated. unlike the traditional lecture course, in which the incremental student imposes minimal marginal effort on the part of the instructor, marginal effort in the practicum seems to increase substantially with the number of course participants. 5. concluding remarks based on their comments, i believe the students who participate in our practicum in financial planning find it an immensely rewarding and eye-opening experience. for some, it is the first time in their lives that they have performed community service; for others, it is the first time in their lives (or at least in their college careers) that they have been made responsible for the completion of a task which extends far beyond the typical “homework” assignment. the completion of this task puts them into meaningful contact with individuals in the community who work in the students’ desired profession. in short, students gain valuable firsthand knowledge attainable in very few other ways. notes 1. the list of registered programs is available on the web page of the cfp board of standards at http://www.cfp-board.org. 2. program information was solicited by e-mail from the directors of registered programs that appeared on the cfp board of standards website in april, 1999. sixty-four program directors (or their representatives) had responded by the time of writing. it is from this sample that the conclusion is drawn. 3. we distinguish here between programs that offer coursework whichrequiresclientbased experiences and those that simply “encourage” students to seek out such experiences. 4. the certified financial planner board of standards also registers self-study programs leading to the certified financial planner examination. a popular program requires students to study and pass, on a sequential basis, materials in the following topical areas: financial planning process, insurance planning, investment planning, income tax planning, retirement planning and employee benefits, and estate planning. upon successful completion of all of the topical exams, the student is eligible to sit for the comprehensive certification exam. 5. this lack of explicit emphasis may be due in part to the nature of the program 301t.h. eyssell / financial services review 8 (1999) 293–303 registration criteria of the cfp board: the institutional registration materials emphasize topical coverage requirements and instructor qualifications, rather than experiential learning opportunities. and, in fairness, it should be noted that the criticism that the teaching of qualitative skills is underemphasized has been leveled at business education for decades. as kenneth hastie pointed out in his presidential address to the financial management association in 1982, “business sees far too many individuals who have all the intellectual and analytical skills necessary to be successful, but who are unable to communicate those ideas and conclusions. and that is regrettable.” hastie (1982), p. 60. 6. the prerequisite for the introductory financial planning course is the first financial management course, which is part of the core business curriculum. similar to most institutions, the financial management course has as prerequisites courses in accounting, economics, and statistics. 7. in mid-1999, the national icfp and iafp organizations announced their intent to merge into a single 25,000-member organization, the financial planning association (fpa). 8. in addition to providing students with the opportunity to observe financial counselors “in action,” this requirement promotes community service. it may also provide a favorable self-selection bias for students considering enrollment in the course as only those who are sincerely interested in financial planning and counseling will devote several hours per week for no compensation. 9. other consumer-oriented organizations might also be contacted. for example, the web page of the national foundation for consumer credit (nfcc) indicates that organization has 1,450 offices nationwide, many of which are cccs offices. the web page can be found at http://www.nfcc.org. i would like to thank a reviewer for pointing this out. 10. convincing one’s department chair or dean to offer a course with an enrollment of ten students might be difficult; in our case, the instructor is simply given credit for one-half course for each practicum section taught. 11. those familiar with the contents of a comprehensive financial plan may question the omission of a segment on estate planning. it is my belief that, without detailed assistance from a financial planner who is highly qualified in the field of estate planning, any but the most general and innocuous recommendations in the area are potentially inappropriate and may open the door to unwanted liability issues for the instructor, the students, and the university. acknowledgment special thanks are due to cheryl fannin, cccs director of counseling, victoria jacobson, cccs director of education, jack white, rfp, cfp, mike kickham cfp, and carl goodin, cfp, for their assistance and willingness to participate in this project. i would also like to thank two anonymous reviewers for their comments on an earlier draft of this paper. 302 t.h. eyssell / financial services review 8 (1999) 293–303 references college for financial planning. (1997).the financial planning process and insurance. denver: the college for financial planning. evensky, h. (1997).wealth management. chicago: irwin/mcgraw-hill. hastie, k. l. (1982). a perspective on management education.financial management, 55–62. rowland, m. (1997).best practices for financial advisors. princeton, nj: bloomberg press. 303t.h. eyssell / financial services review 8 (1999) 293–303 pii: s1057-0810(01)00071-3 consumer information search for home mortgages: who, what, how much, and what else?< jinkook leea, jeanne m. hogarthb,* adepartment of housing & consumer economics, university of georgia, dawson hall, athens, ga 30606-3622, usa bconsumer policies, consumer & community affairs, federal reserve board, washington, dc 20551, usa received april 2000; received in revised form january 2001; accepted february 2001 1. introduction a prerequisite for consumers to make good decisions is to have all the relevant information about the decision at hand, and therefore mandatory information disclosure is often adopted as a public policy remedy to assist consumer decision making (durkin & elliehausen, 1990; oced, 1992). in the consumer mortgage loan market, the truth in lending act (p.l. 90–321; 15 usc §1601) and the real estate settlement procedures act (p.l. 93–533; 12 usc §2601) require such mandatory disclosure of information (federal reserve board, 1997; retsinas, 1997). unlike the well-acknowledged role of informationper se,our understanding of consumer information search behavior when shopping for credit, such as a mortgage loan, is very limited (lee & hogarth, 1999a). addressing this issue, the purpose of this study is to investigate how consumers search for information, particularly when looking for a home mortgage loan. the insights gained from this analysis will guide us to more effective ways to help consumers choose a mortgage loan, and, as a result, improve market efficiency and effectiveness. the remainder of the paper is structured as follows. first, we present background information related to mortgage markets, followed by a review of previous studies on < the analysis and conclusions set forth in this paper represent the work of the authors and do not indicate concurrence of the federal reserve board, the federal reserve banks, or their staff. * corresponding author. tel.:11-202-785-6024; fax:11-202-728-5850. e-mail address:jeanne.m.hogarth@frb.gov (j.m. hogarth). financial services review 9 (2000) 277–293 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(01)00071-3 mortgage shopping behavior. then, using data from the 1997 university of michigan’s survey of consumers, we identify consumers’ information search patterns when shopping for a mortgage loan. specifically, we examine the interdependence of information search activities, providing further insights on consumer search behavior. finally, we provide a discussion for the research and consumer outreach communities. 2. background both the truth in lending act (tila) and the real estate settlement procedures act (respa) mandate information disclosure related to home mortgage loans. the main objectives of the statutes are to aid consumers in making informed credit decisions and to promote price competition by facilitating consumers’ comparison-shopping. although the idea of price disclosure is simple, tila and its implementing regulation, federal reserve regulation z, have grown to require an extensive list of information. the “fed box” disclosures include the annual percentage rate (apr), finance charge, amount financed, and total of payments. disclosures about credit report fees, flood insurance, origination fees, credit life insurance, and numerous other charges are also required if they are necessary to obtain the loan. respa is administered under regulation x from the department of housing and urban development (hud) (meier & garman, 1995; retsina, 1997). the act contains both disclosure and substantive provisions. it requires that certain disclosures be given at various points in mortgage transactions to ensure that consumers receive timely and useful information about the costs associated with the transaction. under respa, within three days of application, lenders are required to provide the mortgage loan borrowers with a good-faith estimate of closing costs and an information booklet explaining the settlement statement. while information related to mortgage loans is fully available as required under tilarespa, several researchers point out that consumers often lack understanding of disclosed information in the mortgage loan market (lee & hogarth, 1999b; chang & hanna, 1992; thakor, beltz & barefoot, 1993). this lack of understanding stems from the complexity of financial information disclosed as well as the proliferation of product choices available (kimball, frisch & gregor, 1997). for example, many mortgage lenders offer a wide variety of mortgage products including fixed or variable rates or some hybrid of the two, varying terms to maturity, and different combinations of rate and points. because the details of transactions may differ substantially, the disclosed information may not always permit easy comparisons. the language of the creditor is also difficult to understand for many consumers (chang & hanna, 1992), and very few people in fact understand the apr (lee & hogarth, 1999b). the timing of the information disclosures may also compound difficulties in comparisons. when shopping, consumers are often quoted a contract interest rate and points plus fees; the apr—the metric for loan-to-loan comparisons—is not required to be disclosed until three days after application. while this delay enables financial institutions to provide an accurate apr, it means consumers must actually apply for a loan—a step usually considered as the culmination of the shopping process—before they know the “true” interest rate. 278 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 3. literature review 3.1. consumer mortgage shopping behavior in the real estate finance literature, consumers’ choice of a mortgage loan is conceptualized as a function of the loan-to-value ratio, the likelihood of prepayment, the points-interestrate tradeoff and the type mortgage sought (fixed or adjustable). sources of uncertainty, including future interest rates, housing prices, and general economic conditions, are other factors in the mortgage decision (follain, 1990). previous researchers have tended to focus on only one dimension of the decision. yang (1992), for example, focused on estimating the points-interest-rate trade-off for both purchase money and refinancing mortgages. lino (1992), on the other hand, focused on consumers’ choice of a fixed or adjustable rate mortgage based on the consumers’ assessment of interest rate risks. it is, however, critical to investigate all the elements of a loan to obtain a full understanding of consumer shopping behavior for a mortgage loan. stigler’s (stigler, 1961) seminal work on the economics of information provides a theoretical basis for consumer’s information search behavior when faced with choosing a home mortgage loan. according to stigler, market price dispersion is the major reason for consumers to engage in information search. consumers can close price gaps by searching for price information, but at some cost in terms of time and money. a consumer will search for lower prices as long as the marginal benefits from additional search exceed the marginal costs of search. chang and hanna (1992) discussed the benefits and costs of search for credit (general credit, not specific to home mortgage loans). immediate benefits include lower interest rates and finance charges; indirect benefits include better money management, greater savings and convenience from using appropriate credit, and gains in financial knowledge and experience gathered from the search process. the costs of search include the opportunity cost of time and physical and mental effort spent in search process. they noted that in today’s complex financial market, the cost of information search may be much greater for a money borrower than a goods buyer because the cost of this service is often not fully or clearly disclosed until the time of application. lee and hogarth (1999b) showed that many consumers are confused about the price of a home mortgage loan as measured by the difference between the contract interest rate and the apr. also, the terms used with closed-end credit products may be more complicated than those used in open-end credit products, such as credit cards. psychological dislike or enjoyment of search influences consumers’ perception of the benefits and costs of search (babin, darden & griffin, 1994). also, “high-risk” consumers may be less willing to engage in search in order to avoid being turned down. on the other hand, if they are turned down or obtain a smaller loan than needed, they may be forced to engage in further search activities. it might be important to distinguish between information search behavior by refinancers and other mortgage borrowers, since their behaviors are governed by different situations. unlike purchase-money mortgage borrowers whose home buying activities compete with their time devoted to mortgage shopping activities, refinancers usually have a longer lead 279j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 time and more flexibility to engage in information search activities. refinancing also suggests that consumers have experience with the mortgage loan process at least once with their current mortgage, unlike first time homebuyers. on the other hand, compared with refinancers, homebuyers have more recent exposure to real estate agents who may serve as an information and referral source. refinancing has been cast as an information asymmetry problem, with consumers knowing when (i.e., under which conditions) they will refinance; lenders are assumed not to have this information (yang & maris, 1993). refinancing is posited to be a function of mortgage value, ability to repay, amount of equity, and credit worthiness, all within an exogenous interest rate environment (bennnett, peach & peristiani, 2000). dickinson and heuson (1993) also include personal decisions as reasons for refinancing in their model of refinancing decisions. thus, borrowers with similar loans react differently to the same market stimulus. in a 1999 survey sponsored by the mortgage bankers association of america (mba), respondents overwhelmingly cited using the phone as their favorite method to obtain information on mortgages; personal contact was the second-most cited method (duncan, 1999). friends and family members were the most-often cited source of information, followed by ads and real estate professionals; for refinancers, experience was also cited as a source of information. the mba study also asked consumers what terms and conditions they compared. four-fifths shopped for “interest rates” (duncan notes especially that they did not say “apr”), 23% asked about fees, 8% asked abut the type of loan and payment schedule, and 3% asked about documentation requirements. half of respondents shopped three or more lenders, while “less than 30% shopped only one lender” (duncan, p. 47). while half of the respondents contacted mortgage brokers and one third contacted mortgage lenders, duncan indicates that there was “massive confusion” on the part of the respondents as to which was which. 3.2. consumers’ information search pattern most previous researchers defined information search either explicitly or implicitly as “the degree of attention, perception, and effort directed toward obtaining environmental data or information related to the specific purchase under consideration” (beatty & smith, 1987, p. 85). information search behavior can be also distinguished as either internal or external (beales et al., 1981). internal information search refers to consumers’ retrieval of memory – knowledge from previous search, experience with products, or passively acquired information during normal daily activities. external information search behavior includes consulting with friends, family, expert consumers, sellers, and third-party experts, reading books, magazine articles, consumer ratings, advertising, and direct inspection. consumers gather information from a variety of different sources. although a variety of different taxonomies of external information sources has been introduced, the sources of information can be classified into direct inspection, seller-provided, personal (family and friends), and third party. seller-provided information has been further categorized into direct-from-seller and advertisements. the major sources of information when shopping for a home mortgage loan include financial institutions such as banks, savings and loans, credit 280 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 unions, mortgage companies, and other financial institutions, advertisements, third-party generated information such as ratings, information from real estate professionals (especially in the case of first-time home buyers), and family and friends. information search is difficult to quantify. over the last three decades hundreds of articles have been published, and numerous measures of search have been used, including both single aspects of search behavior and aggregate measures of search (see lee & hogarth, 2000 for a review of measures of search). while some researchers used multiple measures in their study, most researchers employed only one or two measures, failing to capture all the aspects of information search. given that interdependencies among different search activities are not known, some aspects of search may have been ignored. consequently, conflicting results are obtained, and we do not know if the discrepancies are due to different measures or to different behaviors. information search activities might be interdependent (e.g., using information from thirdparty sources may discourage consumers from obtaining information from advertisements). in fact, several researchers have already suggested that there is a need to consider the interrelationships among various information sources in designing information programs. for example, smith (1993) suggested that advertising influenced how consumers interpret other product-related information. furthermore, consumers’ confidence in their existing information or knowledge may influence how much they search for additional information and the information sources they seek out. however, overconfidence in their knowledge, and the calibration of that knowledge, may lead some consumers to do less-than-optimal searching (alba & hutchinson, 2000). more broadly, previously acquired information in external search may influence additional search activities as well as how consumers interpret the information from that search. addressing this issue, we examine potential interdependencies among information search activities using single and composite measures of information search. specifically, we investigate the following research questions: how do consumers search for information when choosing a home mortgage? who do they get information from? what terms and features do they search for? how much information do they seek out? are there any interdependencies among these information search behaviors? when using a particular information source, what other sources do consumers use? when searching for a specific term, what else do they search for? 4. methods 4.1. data the surveys of consumers were initiated in the late 1940s by the survey research center at the university of michigan. the purpose of these surveys is to measure changes in consumer attitudes and expectations with regard to consumer finance decisions. each monthly telephone survey of 500 households includes a set of core questions covering 281j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 consumer attitudes and expectations along with socioeconomic and demographic characteristics (see curtin, 2001 for more information). in february and march 1997, the federal reserve board commissioned additional questions on the surveys of consumers, including specific questions on consumers’ information search behavior when shopping for a home mortgage. in particular, we asked the number and type of lenders contacted; the number and types of information sources used in mortgage shopping; and the number and types of loan terms and features used in comparison-shopping. for these surveys 1,001 households were interviewed by telephone. among those, only the respondents who applied for or refinanced a home mortgage loan (that is, shopped for a primary mortgage) during the past five years are included in this analysis (n 5 219; a demographic profile of the sample is provided in table 1). for some analyses, we will be able to separate refinancers from “other mortgage borrowers”; however the data do not allow us to determine if these other mortgage borrowers are first time home buyers, other purchase money mortgage borrowers, or some other category of mortgage borrower. households who indicated they obtained home equity loans are not included, although some refinancers may have taken some cash out of their homes as part of the refinancing. for example, if a borrower had $100,000 remaining on the mortgage and had a home worth $150,000, the refinancing could have been for $120,000 – $100,000 to pay off the former mortgage and $20,000 in cash for some other purpose. again, the data do not allow us to determine the pure refinancers from the cash-out refinancers. 4.2. measures the following questions were asked to identify consumer’s search behavior when shopping for a home mortgage. number of lenders contactedrefers to the number of different lenders or brokers a respondent contacted before applying for the mortgage (srinivasan & ratchford, 1991). it is a continuous variable. source of informationreflects where a respondent obtained information. respondents were asked what sources of information they used when shopping for a mortgage loan. each source of information was prompted. y banks, savings and loans, or credit unions y mortgage company y other financial institutions y real estate agents y advertisements – tv and newspapers y comparative ratings published in newspapers, magazines, or internet y friends or relatives a set of binary variables was created to indicate whether each of the above particular sources of information was used (15 used; 05 did not use). number of information sources consultedwas then constructed by summing the number of the above information sources used. 282 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 table 1 profile of sample demographic characteristics respondents who applied for a home mortgage past 5 years (not refinancing) respondents who refinanced a home mortgage past 5 years all respondents age1 18–24 6.1% 3.4% 9.2% 25–34 35.9% 13.6% 22.1% 35–44 32.8% 46.6% 24.9% 45–54 18.3% 26.1% 18.0% 55–64 3.8% 6.8% 8.7% 65 or older 3.1% 3.4% 17.0% education1 less than high school 3.1% 0.0% 10.6% high school graduate 25.2% 25.0% 31.5% some college 19.9% 34.1% 21.2% college degree or more 50.3% 49.8% 34.6% income1 under $10,000 1.5% 0.0% 5.4% $10,000–19,999 5.3% 4.6% 13.7% $20,000–29,999 9.9% 6.8% 13.7% $30,000–39,999 12.2% 9.1% 12.9% $40,000–49,999 8.4% 13.6% 9.7% $50,000–59,999 16.8% 13.6% 9.0% $60,000–74,999 12.2% 21.6% 10.5% $75,000–99,999 9.9% 11.4% 6.1% $100,000 or more 19.9% 18.1% 10.3% below $35,000 (n.e.c.) 0.8% 0.0% 0.9% above $35,000 (n.e.c.) 0.8% 0.0% 0.7% race/ethnicity1 white except hispanic 84.0% 89.8% 78.2% black except hispanic 3.1% 4.6% 8.1% hispanic 4.6% 3.4% 6.8% others 5.4% 1.1% 3.6% marital status1 married 72.5% 75.0% 57.4% separated/divorced 11.5% 15.9% 13.5% widowed 2.3% 1.1% 8.7% never married 13.0% 6.8% 19.6% region west 22.9% 25.0% 21.3% midwest 31.3% 28.4% 26.9% northeast 15.3% 14.8% 18.7% south 30.5% 31.8% 33.2% number 131 88 1001 1 percentage doesn’t add up to 100% due to missing information. n.e.c., not elsewhere classified. 283j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 terms compared. respondents were asked what loan terms and features they compared when shopping for a home mortgage loan. each term was prompted. y interest rates y whether the interest was fixed or variable y points or origination fees y mortgage broker fees y application fees y annual percentage rate y monthly payment amounts y optional credit life insurance y other fees, such as late payment fees y other terms a set of binary variables was created to indicate whether each of the above particular terms and features was compared (15 compared; 05 did not compare). number of terms comparedwas then constructed by summing the number of the above terms compared (ozanne, brucks & grewal, 1992). extent of search. respondents were asked how much comparison shopping they did when they last applied for a home mortgage loan or refinanced their mortgage (srinivasan & ratchford, 1991). the response to this question was on 5-level likert-type scale, ranging from almost no shopping (1) to a great deal (5). 4.3. analysis after performing an initial descriptive analysis, we began exploring interdependencies among of information search activities. first, we examined the interdependencies among the information sources consulted and among the loan terms and features consumers compared, using log-linear models. log-linear models treat all variables as response variables and are most appropriate when the focus is on statistical independence/ dependence among variables. we performed log-linear model analysis using the sas catmod (categorical data modeling) procedure. the catmod procedure allows the examination of interactions, and the likelihood ratio indicates the goodness of fit of the tested model. log-linear modeling procedures begin by estimating the most complex interaction model that incorporates all the possible interaction terms and then iterating to find the model with the best fit by eliminating insignificant interaction terms. for the information sources model, we began by examining 10-way interactions, then 9-way interactions, 8-way interactions, and so forth. for the loan terms and features model, we began by examining 9-way interactions, and iteratively worked through 8-way, 7-way, and so forth. finally, interdependencies among the composite measures (number of lenders contacted, number of sources consulted, number of terms compared, and perceived extent of search) were examined using correlation analysis. 284 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 5. findings 5.1. information search 5.1.1. who do consumers get information from? consumers’ information search behaviors varied a great deal (table 2). about one-fourth (23%) of other mortgage borrowers contacted only one lender compared with only one in seven (14%) of refinancers who contacted only one lender. the mean and median number of lenders compared were 4 and 3, respectively for all mortgage borrowers; there was no statistical difference between refinancers and other borrowers with regard to number of lenders contacted (p 5 0.65). an average consumer obtained information from two sources. lenders, specifically, banks, savings and loans, and credit unions, were the most popular sources for information. while there was no difference between refinancers and other mortgage borrowers in terms of the total number of information sources consulted (p 5 0.90), some notable differences were found with regard to the use of specific sources of information. understandably, larger proportions of other mortgage borrowers (vs. refinancers) obtained information from real estate agents, (29% and 8%, respectively). first-time homebuyers and purchase-money borrowers (included in the other mortgage borrowers category) often have a limited time frame (usually two weeks, included in the purchase offer contract) in which to apply for mortgage. also, homebuyers are more likely to be working with a real estate professional than a homeowner who is refinancing. refinancers were more likely to look for information from advertisements compared with other mortgage borrowers (26% vs. 13%, respectively). 5.1.2. what terms and features do consumers search for? a greater proportion of refinancers reported examining a wider range of terms compared with other mortgage borrowers, although these differences were not statistically significant. most consumers (78%) compared some aspect of interest rates (i.e., interest rate or apr), type of loan (i.e., fixed or variable rate; 76%), and fees (i.e., points or origination fees, application fees, mortgage broker fees; ranging from 54% to 65%). only a quarter of mortgage borrowers compared optional credit life insurance and other fees, such as late payment fees. the mean and median numbers of terms considered were 5.9 and 6, respectively. 5.1.3. how much information do consumers seek out? two-fifths (40%) claimed to do almost no searching or only a little searching; about one-third (32%) reported doing a moderate amount of search and over one-fourth (28%) claimed to do a lot or a great deal of searching. recall that the median number of loans compared was 3 and the median number of terms and features considered was 6. it is evident that consumers consider this a “little” or a “moderate” amount of search, but at the median this requires consumers to process 18 pieces of information (6 features on 3 loans). 285j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 5.2. interdependency of information search—what else do consumers search for? 5.2.1. information sources used using a log-linear model, we examined whether obtaining information from one particular source was associated with obtaining information from another source; results are presented table 2 information searched other mortgage borrowers refinancers all mortgage borrowers number of lenders contacted 1 22.8% 13.8% 19.2% 2–3 32.3% 44.8% 37.4% 4–5 27.5% 20.7% 24.8% 6–9 12.6% 13.8% 13.1% 10–20 4.7% 6.9% 5.6% mean 3.95 4.15 4.03 median 3 3 3 source of information obtained info from (multiple responses allowed) banks, s&ls, credit unions 56.1% 63.6% 59.2% mortgage company 34.6% 42.1% 37.6% other financial institution 6.9% 10.2% 8.3% real estate agent*** 29.3% 8.0% 20.6% advertisements––television & newspapers** 13.1% 26.1% 18.4% comparative ratings—newspapers, magazines, or internet 22.3% 18.2% 20.6% friends or relatives 30.8% 23.9% 28.0% total number of sources considered mean 1.95 1.92 1.94 median 2 2 2 specific terms compared (multiple responses allowed) interest rate 78.9% 85.2% 78.4% fixed or variable interest rate 72.9% 81.6% 76.4% points or origination fees 62.3% 70.1% 65.4% mortgage broker fees 52.8% 56.8% 54.4% application fees 60.5% 60.2% 60.4% apr 68.2% 78.2% 72.2% monthly payment amounts 64.3% 76.1% 69.1% optional credit life insurance 24.6% 25.0% 24.8% other fees, such as late payment fees 24.6% 28.4% 26.2% total number of terms considered mean 5.39 5.59 5.94 median 7 6 6 overall extent of search almost no 23.4% 14.8% 19.9% a little 19.5% 21.6% 20.4% moderate 29.7% 35.2% 31.9% a lot 14.1% 15.9% 14.8% great deal 13.3% 12.5% 13.0% total 131 88 219 note. chi-square statistics are significant at: **p-value, 0.01, *** p-value, 0.001. 286 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 in table 3. one out of seven (14.7%) refinancers and one-fourth (25.5%) of other mortgage borrowers reported that they did not use any information sources when obtaining a mortgage. on the other hand, less than two percentage of all mortgage borrowers used all the sources of information that were prompted. the results of the log-linear model suggest 3-way interactions among: 1) banks/s&ls/ table 3 interdependency among use of information sources: maximum likelihood analysis—log-linear model source df chi-square prob refinancing 1 2.15 0.1428 bank (bank, s&l, credit union) 1 9.46 0.0021 refinancing * bank 1 1.14 0.2857 mortgage companies 1 3.04 0.0812 refinancing * mortgage companies 1 3.67 0.0555 bank * mortgage companies 1 5.12 0.0236 other financial institution (ofi) 1 4.64 0.0312 refinancing * ofi 1 0.34 0.5627 bank * ofi 1 0.90 0.3416 mortgage companies * ofi 1 0.03 0.8624 real estate agent 1 3.84 0.0501 refinancing * real estate agent 1 1.71 0.1915 bank * real estate agent 1 1.77 0.1828 mortgage companies * real estate agent 1 5.27 0.0217 ofi * real estate agent 1 0.30 0.5830 advertisement 1 6.13 0.0133 refinancing * advertisement 1 1.61 0.2042 bank * advertisement 1 5.72 0.0168 mortgage companies * advertisement 1 3.23 0.0721 ofi * advertisement 1 0.59 0.4407 real estate agent * advertisement 1 3.45 0.0633 ratings 1 0.03 0.8561 refinancing * ratings 1 0.11 0.7392 bank * ratings 1 1.22 0.2701 mortgage companies * ratings 1 4.10 0.0430 ofi * ratings 1 0.12 0.7333 real estate agent * ratings 1 7.24 0.0071 advertisement * ratings 1 0.24 0.6238 family 1 0.06 0.8007 refinancing * family 1 0.00 0.9938 bank * family 1 0.54 0.4634 mortgage companies * family 1 0.40 0.5286 ofi * family 1 0.05 0.8181 real estate agent * family 1 2.18 0.1400 advertisement * family 1 2.54 0.1112 ratings * family 1 1.22 0.2698 banks * mortgage companies * real estate agent 1 5.48 0.0192 banks * real estate agent * advertisement 1 4.32 0.0378 banks * real estate agent * ratings 1 7.23 0.0072 mortgage companies * real estate agent * ratings 1 2.15 0.1422 banks * mortgage companies * family 1 6.38 0.0115 mortgage companies * real estate agent * family 1 0.41 0.5228 likelihood ratio 27 22.49 0.7123 287j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 credit unions (hereafter referred to as banks), mortgage companies, and real estate agent; 2) banks, real estate agents, and advertisement; 3) banks, real estate agents, and comparative ratings; 4) banks, mortgage companies, and family. there was also a significant two-way interaction between using information from mortgage companies and comparative ratings.1 if consumers didnotobtain information from real estate agents, but did obtain information from mortgage companies, they were also likely to obtain information from banks. on the other hand, if consumers obtained information from real estate agents, but didnot obtain information from mortgage companies, they were more likely to obtain information from banks. with respect to the three-way interaction among banks, advertisements, and real estate agents, if consumers used information from advertisements, but not from real estate agents, they were more likely to get information from banks. on the other hand, consumers who did not get information from advertisements tended to use information from real estate agents regardless of whether or not they obtained information from banks. for the interaction among banks, real estate agents, and comparative ratings, we find that those who obtained information from comparative ratings were more likely to get information from real estate agents, regardless of whether they obtained information from banks. for those who used third party ratings information, the use of real estate agent-provided information influenced their use of bank-provided information: if consumers did not get information from real estate agents, they were more likely to get information from banks. with respect to the three-way interaction of banks, mortgage companies, and family, when information is not obtained from family or friends, a positive relationship between consulting banks and mortgage companies exists: if consumers obtained information from mortgage companies, they were also more likely to obtain information from banks. on the other hand, when information was provided by family or friends, the relationship between obtaining information from mortgage companies and banks became negative: those who did not get information from mortgage companies were more likely to get information from banks. the two-way interaction between obtaining information from mortgage companies and comparative ratings shows that consumers who obtained information from comparative ratings were also likely to get information from mortgage companies. 5.2.3. loan terms compared using a second log-linear model, we examined whether comparing a given loan term was associated with comparing other loan terms or conditions; results are presented in table 4. one out of seven refinancers (15%) reported they did not compare any terms and one-fourth (26%) of other mortgage borrowers reported not comparing any terms. on the other hand, 14% of all borrowers (refinancers and other mortgage borrowers) reported they compared all the terms that were prompted, and another 17% of all borrowers examined all the terms except credit life insurance and other fees. no complex 3-way interactions were found. significant positive two-way interactions were found for 1) mortgage broker fees and application fees; 2) apr and monthly payment; and 3) credit life insurance and other fees. the two-way interaction of mortgage broker fees and application fees suggest a significant positive relationship between comparing mortgage broker fees and application fees: those who compared application fees tended to compare 288 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 table 4 interdependency among terms compared: maximum likelihood analysis—log-linear model source df chi-square prob refinancing 1 7.87 0.0050 interest rate 1 0.47 0.4919 refinancing * interest rate 1 1.13 0.2884 fixed or variable rate 1 0.10 0.7558 refinancing * fixed or variable rate 1 0.11 0.7456 interest rate * fixed or variable rate 1 1.29 0.2555 points or origination fees 1 0.00 0.9940 refinancing * points 1 0.05 0.8266 fixed or variable rate * points 1 0.43 0.5129 mortgage broker fees 1 1.55 0.2139 refinancing * mortgage broker fees 1 0.03 0.8519 fixed or variable rate * mortgage broker fees 1 1.45 0.2290 points * mortgage broker fees 1 2.79 0.0948 application fees 1 1.29 0.2553 refinancing * application fees 1 2.88 0.0899 fixed or variable rate * application fees 1 0.78 0.3771 points * application fees 1 0.03 0.8532 mortgage broker fees * application fees 1 7.94 0.0048 apr 1 1.49 0.2225 refinancing * apr 1 0.13 0.7157 fixed or variable rate * apr 1 0.79 0.3728 points * apr 1 0.00 0.9742 mortgage broker fees * apr 1 0.00 0.9484 application fees * apr 1 0.87 0.3497 monthly payment 1 1.95 0.1623 refinancing * monthly payment 1 0.01 0.9228 fixed or variable rate * monthly payment 1 1.58 0.2085 points * monthly payment 1 0.40 0.5250 mortgage broker fees * monthly payment 1 0.14 0.7083 application fees * monthly payment 1 0.09 0.7697 apr * monthly payment 1 4.17 0.0412 credit life insurance 1 0.19 0.6650 refinancing * credit life insurance 1 0.09 0.7631 points * credit life insurance 1 0.11 0.7370 mortgage broker fees * credit life insurance 1 0.03 0.8704 application fees * credit life insurance 1 0.02 0.9020 apr * credit life insurance 1 0.09 0.7602 monthly payment * credit life insurance 1 0.65 0.4219 other fees 1 0.03 0.8653 refinancing * other fees 1 0.00 0.9859 points * other fees 1 0.67 0.4148 mortgage broker fees * other fees 1 1.06 0.3031 application fees * other fees 1 0.02 0.8821 apr * other fees 1 0.22 0.6384 monthly payment * other fees 1 0.84 0.3601 credit life insurance * other fees 1 24.67 0.0000 likelihood ratio 17 9.23 0.9143 289j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 mortgage broker fees, and vice versa. similarly, those who compared apr were more likely to compare monthly payment, and those who compared other fees were also likely to compare credit life insurance. 5.2.4. interdependency among composite measures table 5 presents correlation coefficients among number of lenders contacted, number of sources consulted, number of terms compared, and perceived extent of search. although all of the search measures were significantly and positively correlated, the magnitudes of correlation coefficients generally are not great. for example, the correlation coefficient between number of terms compared and number of lenders considered is 0.44, suggesting that these are two different dimensions of search. similarly, the correlation coefficients among number of lenders considered, number of information sources consulted, and number of terms compared all suggest significant relationships, but each seems to include unique characteristics that were not captured by the other measures. the correlation coefficients between the perceived extent of search and these other measures are around 0.45–0.59, indicating the limitation of employing a single measure and the importance of measuring different dimensions of search behavior. 6. discussion, conclusions, and implications this paper has demonstrated that consumers use a variety of sources of information, they search for a variety of terms, and on average, they seek out a moderate amount of information. furthermore, there are significant interdependencies in the features compared in the search process. for mortgage loans in particular, consumers used information from banks and mortgage companies as substitutes when they did not obtain information from real estate agents but used banks and mortgage companies as information-complements when they also obtained information from real estate agents. in the same vein, consumers used information from banks and mortgage companies as complements when they did not obtain information table 5 interdependency among information search measures: number of lenders contacted, number of sources consulted, and number of terms compared; correlation coefficients (p-value) number of lenders considered number of sources consulted number of terms compared extent of search number of lenders contacted 1.0000 0.5438 0.4377 0.5891 (0.0000) (0.0001) (0.0055) (0.0001) number of sources consulted 1.0000 0.6268 0.4674 (0.0000) (0.0001) (0.0001) number of terms compared 1.0000 0.4552 (0.0000) (0.0001) extent of search 1.0000 (0.0001) 290 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 from family and friends but used banks and mortgage companies as information-substitutes when they also obtained information from family and friends. the interaction among the features that consumers compared is of potential interest to policy makers. when deciding what terms and conditions to include in mandatory disclosures, it may be useful to know, for example, that consumers who compare interest rates or aprs are likely to also compare monthly payments. thus, it may be better to require disclosure of a feature consumers are not going to actively seek out on their own rather than one that, although important, they will “always” look for. the current truth in lending act disclosures provide information on the apr, the finance charge (which includes a wide variety of fees and charges), the amount financed, the total of payments, a payment schedule (the number and amount of the payments and when these are due), “opt-in” provisions for credit insurance, whether there is a security interest, filing fees, late charges, and prepayment penalties. in addition, the real estate settlement procedures act requires a preliminary good faith estimate of settlement costs including an itemized list of a wide range of fees. in all, there are dozens of features and costs disclosed per loan, far in excess of the combination of terms, lenders, and information sources consumers report using when shopping. however, while the results of this study suggest that consumers do not always use all of this information in comparison shopping, it is clear that different consumers shop for different sets of features; thus information on a range of features needs to continue to be available. in july, 1998, the federal reserve board and the department of housing and urban development issued a joint report to congress, recommending changes to the disclosures provided to consumers during the mortgage application process (federal reserve board, 1998). these recommendations included various ways of streamlining and simplifying mortgage loan disclosures, thus providing consumers with more meaningful cost information. in addition to simplifying disclosures, the joint report also addressed the issue of timing of information and disclosures. the regulators recognized the tension between providing information early in the search process and providing firm, accurate information about the terms, conditions, and closing costs of the specific loan the consumer is considering. the hud recommendation with respect to timing was for disclosures to “be provided before the consumer pays any significant fees” (federal reserve board, 1998, p. 42). the board’s recommendation was that “initial cost disclosures (includingreliable closing costs and an estimatedapr and interest rate) be provided no later than three days after application” (federal reserve board, 1998, p. 43, italics are in the original document). both regulators indicated that statutory changes would be needed in order to implement these recommendations. congress did not act on this report at the time, but the issue continues to be on their agenda. 6.1. limitations of this study and future research one of today’s more important sources of information, the internet, was not included in this study. the internet provides both a source of information as well as a set of sellers for comparisons. numerous sites provide general consumer education about the home buying 291j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 and mortgage shopping processes. however, recent studies on consumers’ use of the internet show mixed results. about one third (37%) of consumers reported using the internet as an information source when shopping for a home (littlefield et al., 2000). in 1998, “the internet was not yet a factor in the thinking of many borrowers or lenders” (duncan, 1999, p. 42). and while industry reports that on-line traffic of new customers has increased 150% between 1999 and 2000, consumers seem loath to buy insurance or apply for loans on line (business wire, 2000). consumers seem to be using the internet for mortgage shopping, but not for initiating the mortgage transaction (myer, 2000). in a survey by fanniemae, 56% of consumers indicated they used the internet to search for a mortgage, but only 21% used the internet to price a loan and even fewer—4%—applied for a mortgage on line (raines, 2000). future studies on information search activities may need to separate the general information search from the price shopping aspects of the internet. notes 1. figures on these interaction effects are available from the authors. references alba, j. w., & hutchinson, j. w. (2000). knowledge calibration: what consumers know and what they think they know. journal of consumer research, 27(september), 123–156. babin, b. j., darden, w. r., & griffin, m. (1994). work and/or fun: measuring hedonic and utilitarian shopping value.journal of consumer research, 20(march), 644–656. beales, h., mazis, m. b., salop, s. c., & staelin, r. (1981). information remedies for consumer protection. american economic review, 71(may), 410–413. beatty, s. e., & smith, s. m. (1987). external search effort: an investigation across several product categories. journal of consumer research, 14(june), 83–95. bennett, p., peach, r., & peristiani, s. (2000). implied mortgage refinancing thresholds.real estate economics, 28(3), 405–435. business wire (2000). most web-based financial services have yet to win consumer acceptance, according to new study. retrieved from the world wide web at www.mba.org/briefs/2000/1221e.html on december 27. capon, n., & burke, m. (1980). individual, product class, and task-related factors in consumer information processing.journal of consumer research, 7(3), 314–326. chang, y. r., & hanna, s. (1992). consumer credit search behavior.journal of consumer studies and home economics, 16,207–227. curtin, r. t. (2001). surveys of consumers. retrieved from the world wide web at http://athena.sca.isr.umich.edu/scripts/info/info.asp on february 14. dickinson, a., & heuson, a. j. (1993). explaining refinancing decisions using microdata.journal of the american real estate and urban economics association, 21(3), 293–311. duncan, d. (1999). how do borrowers shop?mortgage banking,december, 38–48. durkin, t. a., & elliehausen, g. e. (1990). the issue of market transparency: truth-in-lending disclosure requirements as consumer protections in the united states, in: robert n. mayer (ed),enhancing consumer choice,(pp. 255–265). missouri, co: american council on consumer interests. federal reserve board. (1997). truth in lending,federal reserve press release,march 31, 12cfr part 226, regulation z; docket no. r-0954. federal reserve board and the department of housing and urban development. (1998). joint report to the congress concerning reform to the truth in lending act and the real estate settlement procedures act. 292 j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 retrieved from the world wide web at www.federalreserve.gov/boarddocs/rptcongress/tila.pdf on march 14. follain, j. r. (1990). mortgage choices. journal of the american real estate and urban economics association, 18(2), 125–144. kimball, r. c., frisch, r., & gregor, w. t. (1997). alternative visions of consumer financial services.journal of retail banking services, 19(1), 1–10. lee, j., & hogarth, j. m. (1999a). returns to information search: consumer mortgage shopping decisions. financial counseling and planning journal, 10(1), 49–66. lee, j., & hogarth, j. m. (1999b). the price of money: consumers’ understanding of aprs and contract interest rates.journal of public policy and marketing,66–76. lee, j., & hogarth, j. m. (2000). relationships among information search activities when shopping for a credit card.journal of consumer affairs, 34(2), 330–360. lino, m. (1992). factors affecting borrower choice between fixed and adjustable rate mortgages. journal of consumer affairs, 26(2), 262–273. littlefield, j. e, bao, y., & cook, d. l. (2000). internet real estate information: are home purchasers paying attention to it?journal of consumer marketing, 17(7), 575–590. meier, k. j., & garman, e. t. (1995).regulation and consumer protection.houston, tx: dames. myer, w. h. (2000). the 2000 internet survey.mortgage banking, 61(1), 42–52. organisation for economic co-operation and development (oecd). (1992).consumer information about financial services,paris, france: oced. ozanne, j. l., brucks, m., & grewal, d. (1992). a study of information search behavior during the categorization of new products.journal of consumer research, 18(march), 452–463. raines, f. d. (2000). here come the home loan web-surfers.american banker, 165(192), 10–12. retsinas, n. p. (1997). testimony of assistant secretary for housing federal housing commissioner before the u.s. senate committee on banking, housing and urban affairs, subcommittees on financial institutions and regulatory relief and housing opportunity and community development, july 15. smith, r. e. (1993). integrating information from advertising and trial: processes and effects on consumer response to product information.journal of marketing research, 30(may), 204–219. srinivasan, n., & ratchford, b. t. (1991). an empirical test of a model of external search for automobiles. journal of consumer research, 18(september), 233–242. stigler, g. j. (1961). the economics of information.journal of political economy, 69(june), 213–255. thakor, a. v., beltz, j. c., & barefoot, j. a. s. (1993).common ground: increasing consumer benefits and reducing regulatory costs in banking,general banking study, prepared for the herbert v. prochnow educational foundation of the graduate school of banking at the university of wisconsin-madison, barefoot, marrinan and associates, inc. yang, t. l. t. (1992). self-selection in the fixed-rate mortgage market.journal of the american real estate and urban economics association, 20(3), 359–391. yang, t. l. t., & maris, b. a. (1993). mortgage refinancing with asymmetric information.journal of the american real estate and urban economics association, 21(4), 481–510. 293j.m. hogarth, j. lee / financial services review 9 (2000) 277–293 pii: s1057-0810(99)80007-9 financial services review, 7(2): 129-143 copyright © 1998 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. 'putting your money where your mouth is' a profile of ethical investors diana beal and miche l l e g o y e n this research sought to determine why people chose to invest in earth sanctuaries lim ited (esl), which conserves ecosystems and breeds endangered species as its corporate mission, an unequivocally ethical objective. investors were surveyed to assess the rela tive importance of financial versus ecological considerations. demographic and investor behavior attributes o f esl shareholders were compared with those of austra lian shareholders as reported by an australian stock exchange survey. the results showed that the environmental mission of esl took pre-eminence over financial con siderations for these investors. comparison of the two groups revealed significant differences in most variables. reflecting the shift in societal values, ethical or social investment has emerged in finance and accounting in the last decade as a topic of research concern. ethical investment has been known by alternative nomenclature, including environmental, 'green' and socially responsible investment (sri). the term 'ethical' tends to be used in the u.k. and australia whilst the latter terminology tends to be used principally in the u.s. both terms appear to suggest that underlying values are fixed, non-controversial and unequivocal. this, how ever, is not the case. the additional dimension which is necessary for a commercial venture to be seen as an ethical investment is contentious and varies between individuals. to avoid the implication of absolutism, the u.k.-based ethical investment research information service (eires), an independent research organization, suggests that ethical investment means making investments which reflect the investor's values. ethical invest ments thus have characteristics which allow investors to integrate their personal values and social concerns with their investment objectives. thus, businesses actively supported by such investments ( ' investees ') operate in industries or carry out activities of which the investors approve and want to support. entine (1997) has taken a more hard-line approach and has suggested somewhat scathingly that investees are organizations with a social agenda, as if that were a shameful motivation. knowles (1997), on the other hand, regards ethical businesses as those which try to improve society. diana beal ° faculty of commerce, university of southern queensland, toowoomba 4350, australia. michelle goyen • university of southern queensland. 130 financial services review 7(2) 1998 i. investments available to e t h i c a l investors ethical investors interested in putting their funds to work in companies of which they approve are rather limited in the scope of activities and number of investees. additionally, the lack of ready information about all of the activities of corporations is a constraint to investment. investors may invest directly in the primary or secondary market when they find a company of which they approve, they may use the growing number of investment brokers who offer a service in screening the corporate world for likely ethical investments, or they may invest in trust funds especially set up to attract the ethical investor. warren, stevens, and mcconkey (1990, p. 74) state that "little empirical research exists concerning individual investment behaviour." the motivation of ethical investors investing directly or through the activities of investment brokers has received even less attention in the literature. after investigating the factors that led to the exclusion of stocks from individual's ethical shareholdings, anand and cowton (1993) identified the need for research into the demographic characteristics of ethical investors. ethical investment as an identifiable and discrete force in financial markets has tended to be referred to as investment in specialist trust funds, which are managed by professional funds managers to reflect supposed ethical concerns and to provide a service to ethically concerned investors. 'supposed' is used advisedly, because trust managers can only respond to what is thought to be important in the market, devise a new trust, open it to the public, advertise it and see what happens. the trust funds which reflect investors' needs will grow and be successful; the less well focussed will wither and close. the situation is analogous to the futures market where well traded and popular contracts survive and the less well traded are eventually withdrawn. to gain some guidance for their ventures, funds managers undertake market research so that they can devise portfolios which have a reasonable chance of success in attracting funds. australian ethical investment ltd (aet), the only australian funds manager oper ating exclusively in the sri markets, for example, undertook such a survey for the 1996 1997 period. aet found that environmental concerns were high in priority for australian ethical investors. the issues with highest priority were proactive rather than reactive in nature, being environmental protection, sustainable land use, forest logging/woodchipping, reafforestation, and efficiency of energy and resource use. of lesser importance, but soil of concern, was the screening out of armaments, repressive regimes, unfair work practices, uranium and nuclear industries, and racism and discrimination (aet, 1997a). one of the problems encountered by funds managers is finding enterprises which meet their portfolio design criteria. apart from ethical issues, funds managers generally want optimal return with acceptable risk, so that they can attract investment funds. because it is easier to analyze the activities of small companies, funds tend to flow to smaller companies where ethical qualities are more evident (aet, 1997b; cullis, lewis, & winnett, 1992). additionally, larger corporations are often involved in a multitude of activities where an ethical tradeoff is required (aet, 1997b; cullis, lewis, & winnett, 1992). do the 'good' activities outweigh the ethically unacceptable? are the corporation's activities on balance ethically acceptable? moreover, trust managers are often forced to invest in suspect larger firms in order to increase the return and reduce the risk on invested funds. to a large degree, ethical investors rely on the judgement of funds managers and their supposedly superior information as to the ethical acceptability of investee firms. however, the decisions of managers may not in reality reflect the value framework of individual ethical investors 131 investors. lists of investments held by australian ethical trusts (cummings & burritt, 1997, for example) include a surprisingly high number of mining companies and petroleum explorers. investors take portfolio decisions made by their funds managers on trust until evidence, one way or the other, becomes available. trust fund managers, hoping they are correctly interpreting the motivation of ethical investors, advertize their products with calls to boycott enterprises 'which harm our world' (htr ethical fund) and with talk of 'the moral obligation of our generation to guard the heritage of the earth' (scottish equitable ethical trust). thus, ethical funds managers hope to succeed by providing appropriate services which meet the needs of investors to support businesses which reflect their individual values. additionally, whether ethical investors are a subset of the same group as 'ordinary' investors has not been investigated. how do the characteristics of ethical investors compare with the attributes of 'ordinary' investors, for example, those holding shares listed on the australian stock exchange? this exploratory research aims to shed some light on these questions. ii. t h e selected c o r p o r a t i o n because trust funds support a wide range of business activities, not all of which may be totally acceptable to all individual investors, the design for this research called for a focused, unequivocally ethically acceptable corporation to provide a population of undeni ably ethical investors. the selected corporation also had to be publicly owned, so that a valid comparison of shareholders could be made with the australian shareholder popula tion. one of the few corporations in australia to qualify is earth sanctuaries ltd (esl). because this corporation is involved with nature conservation, environmental protection and sustainable land use, the issues of apparently highest concern to ethical investors, it was the obvious candidate for selection for the study. esl is australia's only public nature conservation company. its core business is nature conservation and, as nature conservation is generally a non-market good, it employs the demand for tourism goods and services as a tool to generate revenue and a market return. the corporation aims to conserve australian indigenous animals, plants and ecosys tems within a commercial environment with minimal financial risk to stakeholders, and to develop sanctuaries in the most important ecosystems. it aims within 10 years to have 20 of australia's most endangered animals including woylies, bilbies, boodies, potoroos and pademelons (all species of small arid-zone kangaroos) living and breeding successfully within its securely fenced sanctuaries. also within its mission are education, interpretation to visitors and consultancy services. directors of esl believe that the various government agencies charged with care of the australian natural environment lack vision, and that bet ter leadership is required. the corporation attempts to provide that leadership (esl, 1997). the corporation currently controls and is developing five sanctuaries, some of which are large, even by national park standards. the key to conservation within the sanctuaries is seen to be the elimination and exclusion of feral cats, pigs, goats and foxes, which have devastated the populations of many small mammals, marsupials and birds, and perhaps driven some to extinction. whilst the erection of many kilometres of vermin-proof fencing is capital-intensive and a major financial undertaking, it is this commitment which differ entiates the activities of this corporation from other nature conservation efforts in austra 132 financial services review 7(2) 1998 lia, especially government initiatives. governments generally appear willing to commit land resources, but unwilling to provide the necessary funds, nor establish the market mechanisms, which would allow properly-resourced management of parks (beal, 1996). esl was started by dr. john wamsley (phd in mathematics) on a small degraded former dairy farm by replanting local floral species. the land area controlled by esl has been successively increased to more than 200 000 acres, principally in arid zones. the entity became a public company in 1993 and, in order to facilitate capital formation and trading, intends to list in the near future. when this research was being planned in october 1997, there were about 1600 shareholders. this number has increased daily, so that there were 1650 in late november, and nearly 2300 by the end of june 1998. the founding fam ily and directors held about 60% of the company, superannuation and public trustee com panies held about 8%, and individuals and private companies held about 32% at june 30, 1998 (www.esl.com.au). expansion of the company has been achieved through the issue of shares under condi tions laid out in prospectuses approved by the australian securities commission. the pro spectuses have been made available through stockbrokers and at the sanctuaries. additionally, esl has maintained a comprehensive web site, www.esl.com.au, which con tains a copy of the current prospectus, and has attracted both australian and overseas inves tors. shares were issued in 1993 at $0.75; the issue price has progressively increased so that the price in 1998 has been $1.50. a small proportion of the issued shares find their way to the secondary market and have generally sold at about $1.20. esl is in a strong position financially with about $10 million in assets and $250,000 in liabilities at june 30, 1997. operating profit after tax for 1996-1997 was about $100,000 of which 67% was paid out. a dividend of 0.4 cents per share, fully franked for paid com pany tax, was paid in that year (australia has a dividend imputation system where share holders are credited with the tax paid by the company on their share of paid-out profits; this tax credit then reduces liability for personal income tax.) dividend yield was 0.33%. some may suggest that esl shareholders are in fact not 'real' investors, but donors. it could be argued that issuing shares is an innovative way to attract donations. contraindica tions to this proposition include: the existence of both a separate charitable foundation which receives donations to further the conservation work of the finn and a secondary mar ket in the shares (donations are non-tradeable); quite substantial minimum purchase requirements for primary share issues; investor interest in financial stability and the corpo ration's plan to float shortly on the asx. in addition, several public trusts and institutional investors hold shares when legal requirements for trusts in australia prevent charitable donations on the behalf of beneficiaries. the first objective in this research was to identify the primary reasons private indi vidual investors are willing to place money in a company that has openly constrained the traditional objective of wealth maximization by the objective of nature conservation. drawing the sample from existing shareholders of esl was expected to capture a subset of investors who are most highly motivated by ethical considerations. as an unlisted company, esl's shares are not exchange traded. even though a secondary market exists for the shares, the relatively low liquidity indicates that esl investors are more likely to view their shares as longer term investments and less likely to be motivated by short term gains. ethical investors 133 i ll e t h i c a l investment research ethical investment research over the last decade or so has included the examination of both the positive and negative portfolio design criteria given in trust investment policies, com parison of the rate of return from ethical trusts with that of 'ordinary' investment portfolios, ethical investment as an agent of social change and corporate social disclosure (csr). rockness and williams (1988) found a broadly similar set of issues concerning ethical investors in the u.s. as was identified for australian investors by the aet. the 1997 report on responsible investing trends in the united states (social investment forum, 1997), however, found that tobacco screens were used by over 97% of ethical funds, while environmental screens were used by only 37%. the increased relative importance of tobacco as a concern for responsible investors in the u.s. should be viewed in the light of the recent success of legal actions against tobacco companies. the u.s. funds may have been motivated more by the threat of declining tobacco companies' share prices than a change in ethical focus. after investigating trusts' portfolio selection criteria, perks, rawlinson, and ingram (1992) reported that u.k. ethical investment trusts consistently maintained that they avoided supporting tobacco and alcohol companies, armaments manufacturers and dealers, gambling, repressive political regimes, the nuclear industry and various aspects of animal exploitation including animal experimentation, chemical testing and wild animal products. on the other hand, trusts encouraged firms which had enlightened employee policies and conditions, supported their communities and protected the natural environment. dunham (1990) identified acid rain, pollution of beaches and rivers and nuclear accidents as the main concerns for those british investors who have an environmental focus. british ethical investors have a somewhat different focus from their australian counterparts. it is to be expected that different geographic localities, societies and business environments will pro duce divergent ethical concerns. perks, rawlinson, and ingram (1992) were interested in how investors can force change upon companies so that their behavior becomes less exploitive and more sustain able. given that funds managers such as pension funds, insurance companies and trusts manage the majority of listed capital and that this is a highly competitive market where short-term rates of return determine investment managers' perceived success, they theo rized that university investors would perhaps be the most likely group of investors where social responsibility would be apparent in making investment decisions. accordingly, they investigated universities' investment policies and found half the respondents said they con sidered ethical issues in making investments. however, further investigation revealed that in fact the university investment managers did not have a firm ethical foundation for their decisions. the only investment they sought to exclude consistently was anything con nected with south africa (governed by a high profile repressive regime at the time) and they used their investment powers to support nothing in particular. if universities, which are generally thought to be the residence of liberal thinkers, cannot lead investment towards desirable social goals and institutional investors are unlikely to take a strong lead ership role, then questions are raised as to the identity of social investors, their characteris tics and motives. dividends, expected returns and the firm's financial stability have been identified in the extant literature (baker & haslem, 1974) as critical considerations for the individual investor's investment decision. perceived risk and diversification needs are identified by 134 financial services review 7(2) 1998 nagy and obenberger (1994) as 'traditional' considerations. the notion that 'you can make a difference' in addition to making financial gains can be heady stuff to the ordi nary citizen. marketing material often says that ethical trust funds can help to change business practice. yet, the reality is that funds managers typically do not invest in corpo rations which cannot meet their criteria, or they sell holdings in companies when they become aware of unacceptable practice. aet notes 'when...ethical parameters stray from preconceived courses, the decision to divest or sell the assets may be made' (aet, 1997a, p. 2). arnold and hammond's (1994) analysis of the south african divestment movement found the social investment funds were slow to adjust portfolio composition in the face of increasing criticism of the sullivan principles, which set minimum ethical standards for u.s. companies operating in south africa. in relation to 13 u.k. funds, eires found funds managers mostly do not attempt actively to influence company policy or practice (eires, 1989). avoiding or selling a small holding in a corporation is unlikely to change the share price significantly nor to affect it for more than a few days. thus, divestment policies are unlikely to force change upon affected corporations. the marketing material of ethical trusts typically emphasizes that the returns from eth ical investment are as good as those from 'ordinary' portfolios (dei, 1998). in other words, 'you can gain comparable financial rewards and influence the business world and society as well, with no cost to yourself.' funds managers are in business to attract funds. the financial markets cynic would have to ask whether it 's true that investors can increase the number of their investment goals and still achieve them all without some decrease in the quality of the achievement of some or all of them. moore (1988) doubted that ethical trust managers, seeking to maximize returns within a reasonable risk framework, can exercise responsible share ownership and fall within the definition of ethical investors. luther and matatko (1994) reported returns from ethical trust funds broadly matched the returns of the whole market. constructing portfolios from 159 listed u.s. firms on var ious ethical screens, diltz (1995, p. 77) concluded that, "taken as a whole, the evidence regarding the effects of ethical screening reveals no overwhelming impact." looking at specific screens, diltz found some evidence to support a favorable market attitude toward good environmental performers. using a sample of unit trusts in the u.k., mallin, saa douni, and briston (1995) found that ethical funds tended to outperform non-ethical funds on a risk-adjusted basis. however, they concluded that the overperformance was weak and may have been specific to the time of their study. the evidence does not, therefore, provide conclusive support for excess returns on ethical investments. guerard (1996) found no sta tistical difference between socially screened and unscreened stocks for the 1987-1994 period. he concluded that returns are more dependent on manager performance than on social screening. corporate social responsibility is an emerging area of concern in the accounting liter ature, with a major theme being whether corporations conducting their operations with a social conscience are rewarded with economic advantage. the link between csr and eth ical investment is of course that funds managers and individual investors must have rele vant and accurate information before an investment can be effected (domini & kinder, 1984). perforce, investors place heavy reliance on annual reports for information, and gen erally they prove inadequate (perks, rawlinson, & ingram, 1992; rubenstein, 1992). the current adequacy of csr is clearly a constraint to effective ethical investment. ethical investors 135 keim (1978) argued that wealth maximization motives of the firm and social respon sibility activities can be consistent. bruyn (1987) set himself the mammoth task of trying to develop a conceptual foundation and body of theory for social investment. essentially, bruyn's theory rests on the fundamental proposition that social investors are committed to economic development and optimal financial returns, given that social and economic val ues are maximized jointly. owen (1990) was severely critical of this approach. owen has major reservations that social investment can be at all congruent with the tenets of the resurgent free market philosophy currently enjoying supremacy globally. similarly, after highlighting socially damaging examples of the consequences of the "obsession with shareholder wealth," beaver (1995, p. 53) states that the "process of developing a more ethically based shareholder mentality will take time." in essence, the prior literature has revealed that ethical concerns vary across national borders, and this can limit the ability of some studies to be generalised. the potential tradeoff of risk, return and socially responsible behavior means that ethical funds managers cannot guarantee that issues of importance to each individual investor are anticipated in their investment decisions. this research contributes to the existing literature by giving attention to the motivations of individual shareholders acting to increase their own utility. iv. data c o l l e c t i o n in addition to the objective of conducting exploratory research in an area little investigated, a second objective was to compare some characteristics of esl shareholders to those of the australian shareholder population. the asx conducted a comprehensive survey of aus tralian shareholders in 1997 to assess some aspects of their investment behavior and demo graphic characteristics. the sample size of 2400 and the contemporaneity of the asx survey suggested that it could be used as a suitable base for the assessment of the austra lian shareholder population. a. selection of the sample at the time the sample was being selected, there were about 1650 shareholders of esl. on the basis that about 300 responses would be necessary for statistical reliability, it was decided to send out about 800 questionnaires, in the expectation of a slightly better than usual response rate than is normally achieved by postal surveys. as the focus of this research is on individual investors, corporations holding shares in esl were not included in the sample. names of shareholders resident overseas were also omitted from the shareholder list for comparability to the asx survey. additionally, dr wamsley and identifiable members of his family were not included in the survey. the sam ple was chosen by selecting every second name on the shareholder list, after the initial selection of the shareholder corresponding to a randomly selected number between 1 and 10 (ray & ravizza, 1988). this method can be expected to be more efficient than simple random sampling, because it allows the selection of a spread of shareholders from the ear liest through to the latest. the sample selection process produced a sample of 825 austra lian shareholders. 136 financial services review 7(2) 1998 table 1 investors ' reasons for holding shares in esl considered considered number of not considered unimportant important reason for purchasing esl shares respondents (%) (%) (%) conservation of endangered animals 692 help to save endangered ecosystems 689 conservation of endangered plants 678 provision of sanctuaries 647 help protect ecosystems 633 provision of educational services 637 financial stability of esl 589 provision of recreational services 622 share price (capital) growth 617 portfolio diversification 608 dividends 580 0.1 .02 97.3 0.6 0.7 95.1 2.7 1.7 88.3 6.5 2.9 77.9 7.1 5.4 70.7 10.8 12.1 52.6 16.3 25.0 35.3 16.9 29.9 25.6 16.6 39.4 20.3 38.8 37.4 9.9 28.6 56.6 3.8 note." the row total percentages are less than 100, because the score 3 responses are not included here. b. development and administration of the questionnaire the questionnaire comprised 19 questions printed on three panels of a single folded a3 sheet of paper. an introductory letter written on university letterhead with an esl logo pasted to the top right comer was included as the first panel on the sheet. the letter invited shareholders to participate in the survey. the instructions placed at the top of the survey indicated that any or all parties involved in shareholdings in joint names could complete the survey. respondents were guaranteed anonymity. fourteen of the questions were derived from the asx survey. these were used to determine demographic details and investor behavior characteristics. the remaining ques tions were exploratory in nature and sought to determine why conservation was important to investors. one question investigated the reasons respondents invested in esl. this question offered eleven possible reasons (see table 1) and asked respondents to rank importance on a five-point likert scale. the possible reasons proffered included financial attributes drawn from the investor behavior literature and some key stated objectives of esl as identified from prospectuses and annual reports. one further question asked respondents to use a likert scale to rate their perceptions of the security of their investment in esl. during the design phase, the questionnaire was pilot tested with colleagues and share holders of other corporations. valuable suggestions were made about the phrasing of ques tions and layout, and amendments were made in line with these suggestions. the data collection design included a double mailing in order to maximize the response rate. the use of a fol low-up survey also provided the opportunity to assess any potential sources of non response bias. survey forms in the first mailing were posted on 2 december 1997, and the second mailing was made on 17 december. respondents were requested to return the com pleted surveys within one week. a free-post return envelope was enclosed with each mail ing. of the 825 survey instruments posted out, 715 were returned, 615 or 74.5% from the first mailing and 100 or a further 12.1% from the second mailing, giving an 86.6% response rate. some joint shareholders availed themselves of the opportunity to complete ethical investors 137 their survey as individuals, bringing the total number of responses to 739. the response rate was greater than 80% from shareholders living in all states, except south australia where it was 75.7%. mail surveys rarely achieve response rates of 80 to 90 percent (zikmund, 1997). we attribute the abnormal response rate to the inherent characteristics of the individuals who are esl shareholders. by "putting their money where their mouths are," these people dis play their commitment to the environment. this research allowed esl shareholders to have their say on issues important to them. a contributing factor to the high response rate might also be the unusually high representation of postgraduates in the esl shareholder population. these people are likely to be sympathetic towards research efforts. as the sur vey's response rate far exceeded the accepted minimum of a 50% response rate necessary for reliability (cryer & miller, 1991; zikmund, 1997), possible sources of non-response bias were considered a priori to be negligible. v. research findings: reasons for investing in esl the survey sought to determine why esl shareholders had invested in the company. five point likert scales were used to assess the importance of eleven factors in the investment decision. table 1 identifies these factors and shows the number of responses to each ques tion. the second column shows the percentage of respondents who indicated that they had not considered that particular factor. the third column provides the percentage of respon dents who rated the factor as unimportant (scores of 1 and 2 on the scale). the final column indicates the percentage of respondents who considered the factor to be important (scores of 4 and 5 on the scale). respondents were invited to report any other reasons for their pur chase of the company's shares. financial returns were not the primary motivating factor behind the investment deci sion of the vast majority of respondents. conservation of animals, plants and ecosystems were considered to be highly important. of those who answered question on the impor tance of the financial stability of esl, 35.3% rated this factor as 'important' or 'very important,' while 16.3% had not considered financial stability at all when making the investment decision. capital growth was rated highly by 20.6%, while only 3.7% rated div idends highly. approximately three-quarters of respondents either did not consider portfo lio diversification or considered this factor to be unimportant reason for making the decision to invest in esl. these results are in stark contrast to the importance placed on financial indicators by members of the australian shareholders' association and invest ment analysts (deegan & rankin, 1997). as previously mentioned, respondents were encouraged to identify any other reasons for their investment decision. the most frequent response in this section was an admiration for the work of dr john wamsley, the founder of the company. the fact that investors con sidered esl to be an "ethical" investment was also a recurring reason. other reasons included "putting your money where your mouth is," "for future generations" and "eradi cation of ferals." many investors were encouraged by the success esl was achieving in the area of conservation, especially when compared to government achievements. others liked the idea of private enterprise conservation activities. 138 financial services review 7(2 ) 1998 vi. comparison of esl investors with the australian asx investor population visual inspection of the data suggests that typical esl shareholders are more likely to be female than typical asx shareholders, educated to a tertiary level and more likely to post graduate degree level. they are professionally employed or retired, likely to have a lower household income, high household assets and a share portfolio containing more than 11 stocks. two tables which summarize the responses received are provided. demographic char acteristics are presented in table 2 and investor behavior characteristics in table 3. some respondents in our sample were under eighteen years old and one adult (while giving an australian address) was not an australian resident. these people were excluded from com parison with asx data as it relates only to australian adults. esl shareholders were gen erally more forthcoming in their response rates to individual questions than were the shareholders in the asx survey. to facilitate meaningful comparisons, the relative per centages of those responding to these questions rather than the percentages of those sur veyed have been used. the data collected were either categorical or ordinal by nature. such data do not sup port the assumptions of parametric testing. seigel (1956) states that nonparametric tests are more powerful in such circumstances, with the power of the tests increasing as sample size increases. thus, chi-squared testing was chosen as the means of testing for independence of the two samples. the relatively high levels of non-response to the household assets and household income questions in the asx survey indicate that our results for these variables should be interpreted with caution. we are unable to draw any inferences on the distribu tion of non-respondents for the asx survey as we only have access to the results of their study and not to their raw data. using the standard significance level of five percent, most of the demographic charac teristics and all of the investor behavior characteristics were significantly different between the two samples. only marital status (which was included to complete the replication of the asx data) could be considered to be undifferentiated between the samples. one startling difference lies in the usage of broker services. the australian share ownership survey (asx, 1997) revealed that 86% of shareholders use the services of a broker to some extent. as esl is an unlisted public company, it was not expected that bro kers would be an important source of information, and this expectation was supported by the survey findings that 2.5% of the sample received a recommendation on the purchase of esl shares from a broker. only 21% of the sample received an investment recommenda tion from any source. of those, the most frequently cited sources of recommendations was from relatives (33.8%), friends (30.5%), brokers (11.9%), media (6.6%) and accountants (2.6%). a five-point likert scale question used to assess the importance of recommenda tions received by the investors gave an average score of 3.6, indicating that recommenda tions were considered as neither important nor unimportant. although the chi-squared tests do not indicate which of the categories caused the dif ferences between the two samples, these can be inferred by inspection of the numbers of responses in each of the cells. of the surveys returned, 490 respondents completed every question. to protect the analysis from any possible systematic bias associated with those respondents who did not answer all questions, the fully completed questionnaires were used to test for differences among the shareholders of esl. eth i ca l inves tors 139 table 2 comparison of demographic characteristics between esl and asx shareholders esl asx p value sex male 53.0 58.0 female 47.0 42.0 0.00462* marital status single 15.8 12.8 partnered/married 71.3 71.0 0.95958 separated/divorced/widowed 12.9 16.2 age 18-24 1.7 6.7 25-34 9.7 19.7 35-44 27.7 24.8 45-54 25.7 22.9 55+ 35.2 25.9 0.0000 education status year 10 and under 10.4 23.9 year 12 9.7 22.3 trade diploma 28.9 20.1 undergraduate 26.1 25.2 postgraduate 24.9 8.5 0.0000 socio-economic status lower 6.8 24.0 lower-middle 8.2 13.3 middle (administrative, middle manager) 12.8 27.7 upper-middle (managerial) 14.2 19.3 upper (professionals) 58.0 15.7 not defined (including retired or unemployed) 29.1 17.0 0.0000 household income less than 20 000 10.1 4.9 21 000 to 30 000 12.0 10.5 31 000 to 40 000 14.3 14.3 41 000 to 50 000 12.3 16.2 51 000 to 70 000 19.2 23.3 71 000to 100 000 16.2 16.2 more than 100 000 15.9 14.6 refusal/not stated 2.6 1 5 . 3 0.00000 household assets less than 25 000 18.8 21.9 26 000 to 50 000 14.2 21.9 51 000to 100 000 12.1 19.6 101 000 to 300 000 28.7 19.2 more than 300 000 26.2 17.5 refusal/not stated 4.5 21.8 0.0000 geographic location metro 61.0 53.0 regional 39.0 47.0 not stated 2.4 0 0.00005 new south wales/act 30.3 33.9 victoria 23.6 26.5 queensland 7.1 20.2 south australia 31.4 8.0 western australia 4.5 9.6 tasmania 2.0 1.4 northern territory i 0.9 0.4 not stated 0.2 0 0.0000 note: *using a two-tailed fisher exact test. a. testing for non-response bias even though the proportion of non-respondents was small (13.4%) by comparison to most other surveys, tests were conducted to determine if it was likely that the non-respon dents differed from respondents in a systematic manner. the 60 responses from the second 140 f i n a n c i a l services r e v i e w 7(2) 1998 table 3 comparisons of investor behavior between esl and asx shareholders esl asx p value number of stocks in portfolio one 15.8 25.6 two 14.1 13.9 three 9.4 14.4 four 6.5 7.2 five 6.9 9.5 six to ten 20.6 18.8 eleven and over 26.7 10.6 not stated 1.7 3.0 0.0000 one to two years 20.6 17.8 three to five years 30.8 18.8 six to ten years 17.0 26.5 eleven to twenty years 15.9 16.9 over twenty years 15.7 20.0 0.0000 under 5 000 23.6 27.2 5 000 to 10 000 11.8 19.9 10001 to 25 000 12.9 22.2 25001 to 50 000 12.9 12.8 50001 to 100 000 11.1 6.1 over 100 000 27.7 11.8 not stated 3.8 16.2 0.0000 length of share ownership value of share portfolio mailing that had all questions answered were compared to the first 60 responses that were complete. the one significant difference that emerged was the geographic location of respondents. this difference is attributable to the relative time it takes for mail to reach those members of the sample located close to the researchers' home state compared to those residing in the more distant states. as such, the difference is trivial. combined with the significance levels found in assessing the differences between esl shareholders and those of the australian shareholder population, there is no reason to suggest non-respon dent bias would alter the results of this study. vii. goal of investors the goal of utility maximization is a fundamental assumption underlying much economic theory. finance theory operationalizes utility maximization via the assumption that wealth equates with utility. in finance, the maximization of shareholder wealth is taken as the pri mary goal for corporations. upon investigation of the motivation for investment in esl, the results have shown that these investors assessed the financial characteristics of the investment as less important than environmental considerations when they decided to "put their money where their mouths are." these investors would not, apparently, be satisfied with wealth maximization and demand that a psychic utility component be included in their returns. presumably, this psychic utility comes from providing capital that is used for the protection and conservation of native flora and fauna. demographic comparisons were made of the esl shareholders and the more tradi tional (and presumably wealth maximizing) shareholders as represented by the asx sur vey. the results show that esl shareholders are indeed different in many fundamental ethical investors 141 aspects. esl shareholders were generally older and more likely to be female than the total shareholder population. they were more likely to be metropolitan residents than regional with significantly higher levels of education, socio-economic status and household assets. table 2 also shows that reported household earnings were significantly different from the asx survey. an alternate explanation stems from the esl sample including many retired or semi retired people. according to life cycle theory, these retirees would have lower cash incomes, but would have sizeable unrealized gains from their relatively large levels of household assets. to get an idea of how well the life cycle theory might fit the esl respon dents, the sample was split by income level. the lowest two income groups were used to represent 'low income earners.' these were then compared to the highest two income groups. (statistical results are not reported here). the low income group had significantly lower levels of household assets, smaller portfolios of much lower value and had held shares for relatively short periods of time. thus, the differences from the asx survey do not appear to be explained by the life cycle theory. viii . conclusion with the exception of 'state of residence,' all of the variables tested were significantly dif ferent. as one would expect, the low income group included more respondents who were female, single, non-urban residents, more than 55 years old retirees who had, on average, a much lower level of education. many esl shareholders had much lower levels of income, thus casting doubt on the ' income affordability' factor suggested in the previous section. the esl shareholders had, on average, owned shares for a shorter length of time than the participants in the asx study. portfolio size was also significantly different between the two surveys. many esl shareholders had much smaller portfolios, while many had much larger than would be suggested by conformity to the asx data. with the sample split on income, the low income group had significantly lower levels of household assets, smaller portfolios of much lower value and had held shares for relatively short periods of time. dif ferences from the asx information do not appear to be driven by retirees with relatively low incomes and high household assets. although the ability to generalize the results of the study may be limited, this research has identified key differences in the characteristics of a subset of environmental investors when compared to 'average' investors. further research is required to determine the extent to which these differences are observed across other groups of environmental investors. esl shareholders appear willing to subordinate short term financial gains in order to pro vide an enhanced natural heritage for future generations. more importantly, the investment motivations of esl shareholders emphasize the divergence of the needs and preferences of this group from mainstream australian inves tors. our results show that wealth maximization was not the primary motivation of esl shareholders. this finding highlights the importance of advisers' investigating and giving consideration to total utility profiles when advising clients. similarly, ethical funds could benefit from offering a greater number of differentiated products to meet specific needs of groups of individuals. ethical funds have struggled for recognition of legitimacy among 142 financial services review 7(2) 1998 mainstream commentators. this research may strengthen the claims for recognition by the ethical investment ' communi ty . ' acknowledgments: the authors would like to thank scott mccarthy, two anonymous referees, and seminar participants at the usq and the university of tasmania seminar series for their useful comments on earlier drafts of this papers. the research was funded by the faculty of commerce, university of southern queensland. r e f e r e n c e s australian ethical trusts (aet). (1997a, june). aim high. australian ethical trusts (aet). (1997b, december). aim high. anand, p., & cowton, c. j. (1993). the ethical investor: exploring dimensions of investment behav iour. journal of economic psychology, 22(4), 483-496. arnold, p., & hammond, t. (1994). the role of accounting in ideological conflict: lessons from the south african divestment movement. accounting, organizations and society, 19(2), 111 126. australian stock exchange (asx) (1997). 1997 australian shareownership survey. australian stock exchange limited. baker, h. k., & haslem, j. a. (1974). toward the development of client-specified valuation models. journal o f finance, 29, 1255-1263. beal, d. j. (1996). the behaviour of agency costs relevant to visitation at national parks of south western queensland. economic analysis and policy, 26(1 ), 1-24. beaver, w. (1995). corporations' misguided obsession with shareholder wealth. business and soci ety review, 95(fall), 49-53. bruyn, s.t. (1987). thefield of social investment. cambridge, ma: cambridge university press. cryer, j. d., & miller, r. b. (1991). statistics for business: data analysis and modelling. boston: pws-kent. cullis, j. g., lewis, a., & winnett, a. (1992). paying to be good? u.k. ethical investments. kyklos, 45, 3-24. cummings, l. s., & burritt, r. l. (1997). is there anything special about the social disclosures of ethical trust investees in australia ? unpublished paper presented at the university of southern queensland seminar series. deegan, c., & rankin, m. (1997). the materiality of environmental information to users of annual reports. accounting, auditing and accountability journal, 10(4), 562-583. directed ethical investment (dei). (1998). what is ethical investment? http://www.peg.apc.org/ -dei/what/html accessed 5/1/1998. diltz, j. d. (1995). the private cost of socially responsible investing. applied financial economics, 5, 69-77. domini, a. l., & kinder, p. d. (1984). ethical investing. reading: addison wesley. dunham, r. (1990). ethical funds--no bar to profit. accountancy, june, 111. eires. (1989). the financial performance of ethical investments. london: eires. entine, j. (1997). vivisecting the anti-vivisectionist movement. drug and cosmetic industry, 169(1), 38-41. earth sanctuaries limited (esl). (1997). annual report 1997. stifling: esl. guerard, j. b. (1996). is there a cost in being socially responsible in investing? http://www.socialin vest.org/mosessy6.htm accessed 13/4/1998. ethical investors 143 keim, g. (1978). corporate social responsibility: an assessment of the enlightened self-interest model. academy of management review, january, 57-68. knowles, r. (ed.). (1997). ethical investment. marfickville, nsw: choice books. luther, r. g., & matatko, j. n. (1994). the performance of ethical unit trusts: choosing an appropri ate benchmark. british accounting review, 26(1), 77-89. mallin, c. a., saadouni, b., & briston, r. j. (1995). the financial performance of ethical investment funds. journal of business finance and accounting, 22(4), 483-497. moore, g. (1988). towards ethical investment. gateshead: traidcraft exchange. nagy, r. a., & obenberger, r. w. (1994). factors influencing individual investor behaviour. finan cial analysts journal, 50, 63-68. owen, d.l. (1990). towards a theory of social investment: a review essay. accounting, organisa tions and society, 15(3), 249-265. perks, r. w., rawlinson, d. h., & ingram, l. (1992). an exploration of ethical investment in the uk. british accounting review, 24(1), 43-65. ray, w. j., & ravizza, r. (1988). methods towards a science of behaviour and experience. belmont, ca: wadsworth inc. rockness, j., & williams, p. f. (1988). a descriptive study of social responsibility mutual funds. accounting, organisation and society, •3(4), 397-4 11. rubenstein, d. (1992). bridging the gap between green accounting and black ink. accounting, organisation and society, 17(5), 501-508. seigel, s. (1956). nonparametric statistics for the behavioural sciences. tokyo: mcgraw-hill kogakusha. social investment forum. (1997). 1997 report on responsible investing trends in the united states. http.//www.socialinvest.org/lnvsritrends.htm accessed 2/2/1998. warren, w. e., stevens, r. e., & mcconkey, c. w. (1990). using demographic and lifestyle analysis to segment individual investors. financial analysts journal, march-april, 74-77. zikmund, w. g. (1997). business research methods (sth ed.). fort worth: the dryden press. har court brace. pii: s1057-0810(00)00056-1 social security investment accounts: lessons from participant-directed 401(k) data jack l. vanderheia,* ,1, kelly a. olsenb,2 a1000 great springs road, rosemont, pa 19010, usa b500 e street, 9th floor, sw, washington, dc 20036, usa revised 16 march 2000 abstract newly available 401(k) participant investment data may have implications for individual social security account (ia) proposals. we found that women with wages between $25,000 and $50,000 have a significantly greater probability of investing a small percentage of their 401(k) in equities than their male counterparts, but those with salaries over $75,000 have a smaller probability. hence, women’s less aggressive investment behavior may be primarily due to younger cohorts and may not apply above a threshold wage. however, overall, 28.4% of men and 33.8% of women are conservative investors, suggesting the possible risk low ia accumulations under some proposals. © 2000 elsevier science inc. all rights reserved. 1. introduction and background obtaining private market returns though social security individual accounts (ias) is a relatively novel and popular idea (bowman, 1999; upston, 1998). this type of reform * corresponding author. tel.:11-610-525-6139. e-mail address:jack@vanderhei.com (j.l. vanderhei). 1 jack l. vanderhei is a faculty member at temple university’s fox school of business and management (department of risk, insurance, and healthcare management) and the research director of the ebri fellow’s program. 2 kelly a. olsen is a policy analyst with the social security administration’s office of retirement policy. all statements made in this paper reflect only those of the authors and do not represent or reflect whatsoever any positions held by the administration or the office of retirement policy. financial services review 9 (2000) 65–78 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00056-1 proposal is the alternative to the traditional, unpopular program fixes of tax increases and benefit cuts that will inevitably need to be part of a comprehensive approach to long-term social security policy. however, the idea raises concerns over the extent to which benefits would vary across participants. while most research thus far has focused on the effect of market volatility on social security benefits across workers participating in ia plans (e.g., ball 1998), the objective of this analysis is to ascertain the extent to which individual investment behavior would be a source of benefit variation. the extent of such variation, relative to current law, is likely to be a consideration used by policymakers and the public in assessing predicted outcomes under ia plans and for the possible design and/or redesign of the plans based on those expectations. to demonstrate the varying degrees to which final social security benefits would depend on individual investment decisions under proposed policies, this article first describes five individual account (ia) proposals that were introduced in 1999. then, a review is provided of previous studies that estimate 401(k) participant investment behavior. in the third section of this article, we describe how newly available data on 401(k) participant investment behavior can be used to identify groups that tend to be overly conservative investors in their 401(k) plans, and we present our findings. finally, the implications of these results are discussed in terms of the ia plans profiled. 2. specific proposals for individual social security investment accounts 2.1. kolbe-stenholm plan (1999 version) the kolbe-stenholm proposal (table 1) would maintain oasdi taxes at their current-law rate (12.4% in 2000 and beyond) and divert 2 percentage points of this 12.4% into ias. voluntary contributions of up to $2,000 per year could be added, and low-income workers would be provided a matching contribution on a sliding scale based on income. investment choices would be limited and designed to resemble those available to federal workers who participate in the thrift savings plan (tsp). tsp participants are currently limited to investment in equity index funds, treasury bonds, and debt index funds, and a small-cap and an international fund are scheduled to be added. in order to “make room” for the diversion of 2% of taxable payroll to individual accounts, kolbe and stenholm would significantly reduce the current-law defined benefit. in order to protect workers to some extent from down-side investment risk, this plan proposes a sliding scale minimum defined benefit social security payment equal to 60 to 100% of the poverty level for retired workers with between 20 and 40 years of covered employment. for workers who want to retire with social security benefits above the poverty level, individual account investment behavior would be extremely important to total social security benefits under the kolbe-stenholm plan. 66 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 2.2. nick smith’s plan (1999) like the kolbe-stenholm plan, rep. nick smith’s most recent plan (table 1) is the quintessential “carve-out” plan, whereby a percentage of social security payroll taxes would be redirected to individual accounts, and social security’s defined benefit would be reduced so that it could be financed with the resulting (lower) payroll tax rate. under nick smith’s 1999 plan, the payroll tax “carve-out” would begin at 2.6 percentage points from 2001–2036. investment selection would be modeled on the tsp for smaller account balances. after account assets had grown to a sufficient (to-be-determined) size, they could be transferred into an ira type of vehicle that would allow for broader investment selections. at time of benefit claim, the participant’s social security defined benefit would be reduced by the amount of his or her hypothetical account balance—not actual account balance. the hypothetical account balance would be calculated using a rate based on long-term u.s. bonds held by the public, plus 0.7 percentage points. under the smith plan, participants’ social security benefits—which would be reduced from current-law levels—would be further reduced by the amount of benefits that could be provided by a life annuity from the hypothetical account balance. through the combination of the ia and the defined benefit, participants could receive more than the scaled-down social security defined benefit (i.e., the benefit before the hypothetical annuity is subtracted) if their actual ia returns are greater than the return used to compute their hypothetical annuity. on the other hand, they would receive less than the scaled-back defined benefit if their ia investments earned less than the returns used to compute the hypothetical annuity. as a result, the number of individuals who might only invest in long-term bonds and not earn on their assets the additional 0.7% used to compute their hypothetical annuity for purposes of reducing their defined benefit from social security is of interest. in addition, the number of people who would earn returns substantially higher than the long-term bond rate plus 0.7% is a behavioral assumption of particular interest in assessing the distributional outcome under this plan. 2.3. moynihan-kerrey plan (1999) this plan permits a worker to elect individual account contributions equal to 1% of taxable payroll, which his or her employer would be required to match (for a total of 2% of taxable payroll beyond the payroll tax specified to finance the plan’s reduced social security defined benefit) (table 1). a wide range of investments would be available to workers who directed their contributions to an ira or any other type of voluntary investment account, as determined by the secretary of treasury. other workers could elect to contribute to a voluntary investment fund that would provide investment options similar to those offered to federal workers under their thrift savings plan. the individual accounts would be exclusively supplemental to the social security defined benefit in the sense that participation is voluntary. on the other hand, these accounts could be considered to be a partial replacement for social security benefits, since current-law social security benefits would be reduced under the moynihan-kerrey plan based on an individual’s cohort and other characteristics. either 67j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 t ab le 1 s el ec te d in di vi du al ac co un t po lic y op tio ns pr op os ed in 19 99 p la n s up pl em en ts or re pl ac es a po rt io n of s oc ia l s ec ur ity ? f un di ng so ur ce in ve st m en t ch oi ce s oc ia ls ec ur ity be ne fit of fs et ba se d on ac co un t ba la nc e? m in im um be ne fit gu ar an te e be yo nd s oc ia l s ec ur ity de fin ed be ne fit ? r ep s. ji m k ol be (r ., a z ) an d c ha rle s s te nh ol m (d ., t x ) (m ay 19 99 p ro po sa l)r ep la ce s m an da to ry 2 pe rc en ta ge po in ts of s oc ia ls ec ur ity pa yr ol lt ax w ith sc al ed m at ch in g co nt rib ut io ns fo r lo w -in co m e w or ke rs on vo lu nt ar ya co nt rib ut io ns li m ite d ch oi ce s, su ch as th os e av ai la bl e to fe de ra l w or ke rsb n o y es , be tw ee n 60 an d 10 0 pe rc en t of po ve rt y le ve l, de pe nd in g on ye ar s of w or k r ep . n ic k s m ith (r ., m i) c r ep la ce s m an da to ry in di vi du al ac co un t co nt rib ut io ns eq ua lt o 2. 6 pe rc en t of ta xa bl e pa yr ol lf ro m 20 01 -3 6d li m ite d ch oi ce s, su ch as th os e av ai la bl e to fe de ra l w or ke rs un til ac co un ts be co m e la rg e en ou gh to tr an sf er in to an ir a ve hi cl e w ith br oa de r in ve st m en t ch oi ce s y es , de fin ed be ne fit is re du ce d by th e am ou nt of ac co un t co nt rib ut io ns , ac cu m ul at ed hy po th et ic al ly at a ra te ba se d on lo ng -t er m u .s . bo nd s he ld by th e pu bl ic pl us 0. 7 n ot ap pl ic ab le , si nc e de fin ed be ne fit pa id to in di vi du al de pe nd s on hy po th et ic al in di vi du al ac co un t ba la nc e pe rc en ta ge po in ts . p ar tic ip an ts ca n re ce iv e ad di tio na lb en efi ts if in di vi du al ac co un t re tu rn s ex ce ed th os e us ed to co m pu te th e hy po th et ic al ba la nc e. s en s. d an ie lp at ric k m oy ni ha n (d -n y ) an d b ob k er re y (d -n eb )e a rg ua bl y ei th er v ol un ta ry in di vi du al ac co un t co nt rib ut io ns eq ua lt o 2 pe rc en t of ta xa bl e pa yr ol l( 1 pe rc en t ea ch fr om em pl oy er s an d em pl oy ee ) e st ab lis he s a vo lu nt ar y in ve st m en t fu nd w ith in ve st m en ts lik e th os e of fe re d to fe de ra l w or ke rs bu t al so ir a ty pe ac co un ts n o n o (c o n tin u e d o n n e xt p a g e ) 68 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 t ab le 1 (c o n tin u e d ) p la n s up pl em en ts or re pl ac es a po rt io n of so ci al se cu rit y? f un di ng so ur ce in ve st m en t ch oi ce s oc ia ls ec ur ity be ne fit of fs et ba se d on ac co un t ba la nc e? m in im um be ne fit gu ar an te e be yo nd s oc ia l s ec ur ity de fin ed be ne fit ? r ep . jo hn k as ic h (r -o h ) r ep la ce m en t ir re vo ca bl e op tio n to pa rt ic ip at e in in di vi du al pl an ; go ve rn m en t w ou ld co nt rib ut e be tw ee n 1 an d 3. 5 pe rc en t of ta xa bl e pa yr ol l (d ep en di ng on ea rn in gs ) to in di vi du al ac co un t in ve st m en ts ch oi ce s lik e th os e of fe re d to of fe de ra lw or ke rs y es , in ex ch an ge fo r pa rt ic ip at in g in in di vi du al ac co un t pl an , pa rt ic ip an ts m us t ag re e to a sc al ed ba ck de fin ed be ne fit n o r ep s. b ill a rc he r (r ., t x ) an d e . c la y s ha w (r ., f l) f r ep la ce s m an da to ry 2 pe rc en ta ge po in ts of s oc ia l s ec ur ity pa yr ol lt ax a ll as se ts w ou ld be in ve st ed in a qu al ifi ed m ut ua lf un d m ai nt ai ne d w ith a po rt fo lio al lo ca tio n of 60 pe rc en t st oc k in de x fu nd s an d 40 pe rc en t co rp or at e bo nd s y es , pa rt ic ip an ts ar e gu ar an te ed at le as t th e cu rr en tla w s oc ia l s ec ur ity be ne fit an d ca n re ce iv e ad di tio na l be ne fit s if in di vi du al ac co un t ba la nc es ex ce ed th e am ou nt ne ed ed to fu nd cu rr en tla w s oc ia l s ec ur ity be ne fit n ot ap pl ic ab le , si nc e de fin ed be ne fit pa id to in di vi du al de pe nd s on in di vi du al ac co un t ba la nc e; to ta lb en efi t ca nn ot be lo w er th an cu rr en tla w s oc ia l s ec ur ity be ne fit a a dd iti on al $2 ,0 00 pe r ye ar co ul d be co nt rib ut ed by th e in di vi du al on a vo lu nt ar y ba si s. b f ed er al w or ke rs w ho pa rt ic ip at e in th e f ed er al t hr ift s av in gs p la n ar e cu rr en tly lim ite d to in ve st m en t in eq ui ty in de x, t re as ur y bo nd , an d de bt in de x fu nd s. c u pd at e of th e s oc ia ls ec ur ity s ol ve nc y a ct of 19 95 . d a m ou nt of ac co un t co nt rib ut io ns be yo nd 20 36 de pe nd s on th e am ou nt of o a s d i in co m e th at w ou ld no t be ne ed ed to in or de r to m ai nt ai n so lv en cy of th e de fin ed be ne fit pl an on a pa yas -y ou go ba si s w ith a tr us t fu nd ra tio of ab ou t 50 pe rc en t of an nu al co st . e s oc ia ls ec ur ity s ol ve nc y a ct of 19 99 (s . 21 ). f t he s oc ia ls ec ur ity g ua ra nt ee p la n. 69j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 way, individual investment behavior could significantly affect retirees’ well being under this reform option. 2.4. kasich plan (1999) the kasich plan gives workers an irrevocable option to participate in an ia system, whereby the government would contribute to their personal accounts (table 1). the amount that the government would contribute would be between 1 and 3.5% of the individual’s taxable payroll, with higher percentages going to lower-earning workers. workers would be provided the same type of investment options that are available to federal workers under the tsp. in exchange for accepting government contributions to their account, the individual would agree to accept a social security defined benefit that would be reduced by the number of years for which account contributions were made. obviously, the extent to which individual investment behavior can produce returns that offset or exceed the worker’s reduced defined benefit is crucial to predicting outcomes under the kasich plan. 2.5. the social security guarantee plan (1999) this plan takes a novel approach to individual accounts (table 1). in a sense, the archer-shaw plan is a new type of hybrid between defined benefit and defined contribution plans. while archer-shaw includes individual accounts and allows participants to benefit from up-side investment risk, the social security program eventually recaptures the entire account balance and protects workers from downside investment risk, inflation risk, and longevity risk. although individuals would manage their own accounts, such management would be limited to the choice of one’s qualified mutual fund provider. for purposes of social security accounts, these providers would be required to offer only investments with a portfolio allocation of 60% indexed equities and 40% corporate bonds. at retirement or disability, the social security administration would use a formula to determine whether a participant’s account balance was adequate to provide a benefit equal to that available under the currentlaw social security program. if the result of this formula shows that the account to be inadequate, then social security’s defined benefit program would cover the difference, and social security would pay benefits in the form of a life annuity, just as under current law. alternatively, if the result of the formula was a higher benefit than the beneficiary would be entitled to under current-law, then social security pays the beneficiary a life annuity greater than the current-law benefit. either way, the entire balance of the individual account is gradually transferred to the social security trust funds, and for about half of all participants who will live beyond their life expectancy, social security’s defined benefit program would pay benefits entirely after the exhaustion of their individual account assets. because individual account investment behavior is so constrained under archer-shaw, this plan’s outcome depends least on individual asset allocation when compared to other plans described above. 70 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 3. 401(k) participant behavior 3.1. relevance and applicability to ias due to their preeminence in the defined contribution plan world, 401(k) plans are the closest existing model to ias in terms of how large numbers of workers would invest their account balances. however, it should be noted that the populations covered by 401(k) plans differ substantially in terms of income and other demographic variables from the 96% of the u.s. workforce that is covered by social security (vanderhei, 1999a). in addition, to the extent that 401(k) behavior is influenced by the existence of a totally defined benefit social security promise, current 401(k) investment behavior may differ from ia investment behavior. in 1981, the irs issued preliminary regulations that set the stage for the proliferation of 401(k) plans, which accounted for over three-quarters of the net increase in all private pension plans from 1980–1995. by 1995 (the most recent data available), 401(k) plan contributions accounted for more than half of all new defined benefit and defined contribution retirement plan contributions (u.s. department of labor, 1999). as a result, more and more individuals have a direct stake in the equities market, as evidenced by the number of households that own mutual fund shares. mutual funds are one gauge of how 401(k) plans have brought individuals into the market, since 35% of mutual fund assets are held in retirement plans (investment company institute, 1999). 3.2. previous research on 401(k) investment behavior despite the importance of 401(k) asset allocation behavior as the closet available model by which to estimate possible ia behavior, previous 401(k) studies unfortunately have provided limited information. as a result, analysts generally have estimated the outcomes of various ia plans by assuming: that all individuals invest identically; that individuals will invest in a range of hypothetical portfolios (e.g., all bonds, all equities, or mixed); or that the experience of a limited sample applies to all workers. the database used for this analysis bypasses several limitations inherent in earlier studies of 401(k) investment behavior and thereby precludes the need to use such simplifying assumptions to estimate the outcomes of various ia plans. much of the previous research in this field examined aggregate 401(k) plan data (buck consultants, 1997; hewitt associates, 1997; profit sharing/401(k) council of america, 1997; kpmg, 1998; william m. mercer, 1997; and cerulli associates, inc., 1998). although this allowed analysts to aggregate plan data to determine overall asset allocation behavior, it did not allow any examination of the impact of demographic characteristics. still other researchers have used surveys of 401(k) participants to analyze participant activity and decision-making in 401(k) plans. one of the more frequently used is the survey of consumer finances (scf), a stratified random sample of u.s. households, administered by the federal reserve board. although the survey has the advantage of providing information on asset holdings outside the participant’s 401(k) plan, it only asks the respondents to indicate plan asset allocations as “mostly in stock,” “mostly in bonds,” or “split between.” any analysis 71j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 of this data therefore must either restrict itself to these three categories or utilize ad-hoc assumptions with respect to the actual distributions. similar problems exist with the national longitudinal survey of mature women (papke 1998). at least three research projects have used 401(k) administrative records; unfortunately they have all utilized relatively small samples. goodfellow and schieber (1997) investigated the investment elections of 36,000 participants in twenty-four 401(k) plans. the total number of participants in the plans analyzed in their study ranged from around 150 to 6,000. in addition, yakoboski and vanderhei (1996) analyzed the asset allocation decisions of 401(k) plan participants working for three large employers (at&t, ibm corporation, and new york life insurance company) with a total of 180,000 employees. finally, hewitt associates has developed an index to track the investment activity of 401(k) participants. this index is based upon 1.4 million 401(k) participants with approximately $62 billion in collective assets. currently, this index reflects experience of large corporations and does not provide any analysis of employee demographics. two other micro level defined contribution databases have been analyzed but constitute different types of plans. hinz, mccarthy, and turner (1997) investigate asset allocations among federal thrift savings plan participants, and ameriks, king and warshawsky (1997) perform a similar analysis on the tiaa-cref population. 3.3. the ebri/ici database in comparison to previous studies, the database created by the employee benefit research institute (ebri) and the investment company institute (ici) for 1996 contains 6.6 million active participants, who hold nearly $250 billion worth of assets in 27,762 401(k) plans of all sizes. measured against the universe of 401(k) plans, the 1996 database accounts for 9% of all plans, 18% of all participants, and 31% of all assets. the distribution of participants, plans, and assets in the ebri/ici database for 1996 is similar to that reported for the universe of plans by cerulli associates, inc. (1998). for each of five plan size classifications, the share of the database’s assets falling within those categories is very close to the share found in the universe for that size category. similarly, the share of the database’s participants and plans within these size categories is approximately the same as that in the universe (vanderhei et al., 1999b). 4. analysis of participant-directed asset allocation much research has focused on the effects of market volatility risk in distributing social security benefits across workers participating in ia plans (e. g., ball 1998). the objective of this analysis is to ascertain the extent to which individual investment behavior is a potential source of variation in social security benefits between ia participants. the extent of such variation in benefit distribution across beneficiaries, relative to current law, is likely to have implications for assessing outcomes under ia plans and for possibly (re)designing ia plans based on those expected outcomes. however, it is important to note that just because beneficiaries under one ia proposal may experience more disparate benefits than under 72 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 current-law does not mean that one proposal is necessarily inferior to another. for example, a plan could theoretically produce more disparate benefits across beneficiaries than currentlaw but also provide higher benefits to all, or be more desirable on other grounds. it is indisputable, though, that the extent to which benefit distribution will differ from current law social security benefits isone important concern when comparing the desirability of different ia plans. for purposes of this analysis, investment options have been grouped into three broad asset classes: 1. diversified equity fundsconsist of pooled accounts primarily investing in stocks. these funds include mutual funds, bank collective trusts, life insurance separate accounts, and other pooled investments. in addition, 60% of any monies invested in balanced funds were added to this category. 2. fixed income fundsincluded bond funds and 40% of any balanced funds in addition to money funds, guaranteed investment contracts (gics) and other stable value funds (synthetic gics providing benefit payments at book value or similar instruments). 3. company stockwhich is equity in the participant’s employer. because vanderhei et al. (1999b) emphasize that overall 401(k) asset allocation is largely a function of the investment menus provided to workers (e.g., the existence of company stock as an optional and/or required investment for employee and employer contributions), we limit our study to a subset of participants completely free of any potential company stock contamination. we chose a sample of approximately 122,000 participants with age, salary, gender and complete information on year-end 1996 asset allocations. none of these participants were exposed to a company stock investment as either an optional or a required contribution. similar to earlier studies by bajtelsmit and vanderhei (1997) and hinz, mccarthy and turner (1997), we begin by estimating the effect of gender, age and salary on asset allocations for the total account balance. given the nonlinearity of the impact of age and salary found in yakoboski and vanderhei (1996), bajtelsmit and vanderhei (1997), and goodfellow and schieber (1997), we chose to segment the age category into ten year increments (20–30, 30–40, 40–50, 50–60 and over 60) and salary into $25,000 increments (less than $25,000, $25,000-$50,000, $50,000-$75,000, $75,000-$100,000, and over $100,000). we also chose to analyze the impact of gender by including both a dummy variable for females as well as an interaction term with each of the age and salary categories. table 2 shows double-censored tobit regression results when the dependent variable is defined as the percentage of equity in the total account balance. the model consists of a gender variable (female), four salary dummy variables, four age dummy variables and separate interaction terms between gender and the salary and age dummy variables. the salary variables (sal2550, sal5075, sal75100, and sal100p) bracket the participant’s salary in $25,000 increments with the last variable including all participants earning in excess of $100,000. the age variables (age3040, age4050, age5060, age60p) divide the participants into ten year age brackets with the last variable including all participants over age 60. the various interaction terms (e.g., isal2550) represent dummy 73j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 variables that are equal to one if the participant is a female and the respective salary or age variable is equal to one. each of the coefficients for salary are positive and significant, indicating that persons earning more than $25,000 tend to invest in equities to a greater extent than those earning less than $25,000. however, the coefficients are not monotonically increasing but follow a “u” shape with the minimum (but still positive) value for those earning between $50,000 and $75,000. each of the age coefficients are also significant: however, as expected, the coefficients for those over 40 are all negative (indicating a smaller propensity to invest in equities than those in their twenties). after a small positive coefficient for participants in their thirties, the coefficients become increasingly negative for older age cohorts. the coefficient for the female dummy variable is negative (indicating that women have a smaller propensity to invest in equities) and statistically significant. in addition many of the interaction terms between gender and salary or age are also significant, suggesting that the relationship between these variables and equity allocation differs by gender. the interaction terms between gender and $25,000 and $75,000 salaries are negative and statistically significant while the interaction terms for women at salaries above $75,000 are insignificant. this suggests that some of the gender differences in asset allocation reported earlier (sunden, 1998; papke, 1998) may be due primarily to lower-earning females and that, above a threshold wage, the distinctions may no longer be material. two of the interaction terms for age and gender categories are statistically significant. the coefficient for those between age 30 and 40 is negative while the coefficient for females ages 50 to 60 is positive. this suggests that women may be more risk averse in their asset allocation decisions while young and that this discrepancy lessens and even reverses later in life. table 2 estimates of the effect of gender, age and salary on equity allocation for 401 (k) total account balances variable estimate std err pr. chi intercpt 0.44567686 0.006469 0.0001 sal2550 0.14827351 0.0066276 0.0001 sal5075 0.09016767 0.007006 0.0001 sal75100 0.1228902 0.008916 0.0001 sal100p 0.2406459 0.010615 0.0001 age3040 0.04015406 0.007449 0.0001 age4050 20.0456495 0.007371 0.0001 age5060 20.1040694 0.008217 0.0001 age60p 20.2480606 0.014716 0.0001 isal2550 20.1509172 0.0086698 0.0001 isal5075 20.0574428 0.010559 0.0001 isl75100 0.01385501 0.016792 0.4093 isal100p 0.03445817 0.024106 0.1529 iage3040 20.0294692 0.01073 0.0060 iage4050 20.0013867 0.010614 0.8961 iage5060 0.03041728 0.0123 0.0134 iage60p 20.017036 0.022798 0.4549 female 20.0305001 0.009192 0.0009 74 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 however, some prediction of groups that will be expected to generate less-than-average investment income based on their under aggressive investment behavior is perhaps of more relevance to the current public policy debate. bajtelsmit (1996) demonstrates the potential for very diverse outcomes for persons of similar demographics given varying investment behavior. table 3 provides estimates from a probit model for the sample described above that predicts whether a participant will hold a very small percentage of their total account balance in equities. admittedly, there is no clear-cut threshold as to what should be considered a small percentage, but we chose 20% as a cutoff value. overall, we find that 28.4% of all males, and 33.8% of all females fall into this category of conservative investors. the variable definitions are the same as in table 2. the results in table 3 generally follow those in table 2 as expected. all the noninteractive age terms are statistically significant as are all but one of the salary terms. the salary categories follow the same “u” shaped curve as before and after an initial negative coefficient for participants in their thirties (which suggests they are less likely to invest a “small” percentage of their assets in equities than participants in their twenties), the coefficients are all positive and monotonically increasing with age. the female dummy variable is statistically significant at the ten percentage level, as are as all but one of the gender-salary interaction terms. the women with wages between $25,000 and $50,000 experience a significantly greater probability of investing a small percentage of their 401(k) assets in equities, while those with salaries over $75,000 have a significantly smaller probability of investing a small percentage of their 401(k) assets in equities than their male counterparts. only one of the age-gender interaction terms is significant. females between 30 and 40 table 3 probit estimates for the probability that a 401 (k) participant will hold less than 20 percent of the total account balance in equities variable parameter estimate standard error pr . chisquare intercpt 20.5168 0.0136 0.0001 sal2550 20.2003 0.0134 0.0001 sal5075 20.0199 0.0145 0.1706 sal75100 20.1316 0.0188 0.0001 sal100p 20.4162 0.0239 0.0001 age3040 20.0881 0.0160 0.0001 age4050 0.0513 0.0156 0.0010 age5060 0.1803 0.0172 0.0001 age60p 0.4484 0.0294 0.0001 isal2550 0.2691 0.0182 0.0001 isal5075 0.0202 0.0218 0.3531 isl75100 20.0797 0.0359 0.0265 isal100p 20.1982 0.0581 0.0006 iage3040 0.0620 0.0226 0.0061 iage4050 0.0246 0.0222 0.2671 iage5060 20.0336 0.0254 0.1867 iage60p 0.0666 0.0450 0.1389 female 0.0363 0.0192 0.0581 75j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 have a significantly higher probability of small equity investments than do males; however there is no significant difference after age 40. this suggests that the less aggressive 401(k) investment behavior observed for women may be primarily due to younger female cohorts. the overall impact of the estimated model in table 3 is summarized in table 4. the estimated probabilities for each age and salary category are displayed by gender. 5. social security implications/discussion our results suggest that differences which other researchers have found between men’s and women’s 401(k) investments may not apply above a threshold wage. assuming this applies to social security accounts, the bad news for advocates is that despite recent progress in closing the gap between men’s and women’s earnings, women’s labor force participation rates continue to be lower than men’s at every age. also, despite the fact that a larger percentage of women are entering the workforce than ever, women’s earnings are still substantially lower than those of men. women who worked full-time, year-round in 1955 earned about 64% of what their male counterparts earned. not much has changed since then; women who work full-time, year-round today earn about 70% of what their male counterparts earn—an increase of just 6 percentage points (9%) in over 40 years. this means that the gap between men’s and women’s social security benefits that exists under today’s program due to work history differences could widen under an ia system due to differences in investment behavior. of the five proposals profiled above, only the archer-shaw plan would protect certain groups of workers (such as lower-earning women) that were shown above to be most likely to invest overly conservatively, and this plan would be most apt to distribute social security benefits across workers in the same pattern as current-law. the kolbe-stenholm proposal would mitigate the extent to which some persons within a cohort would receive different social security benefits based on different investment choices by including a minimum table 4 estimated probabilities that a 401 (k) participant will hold less than 20 percent of the total account balance in equities, by age, salary and gender male 20–29 30–39 40–49 50–59 601 ,$25,000 30.3% 27.3% 32.1% 36.8% 47.3% $25,000–$50,000 23.7% 21.0% 25.3% 29.6% 39.4% $50,000–$75,000 29.6% 26.6% 31.4% 36.1% 46.5% $75,000–$100,000 25.8% 23.1% 27.5% 32.0% 42.1% .$100,000 17.5% 15.4% 18.9% 22.6% 31.4% female 20–29 30–39 40–49 50–59 601 ,$25,000 31.5% 30.6% 34.3% 36.9% 51.4% $25,000–$50,000 34.0% 33.1% 36.9% 39.6% 54.1% $50,000–$75,000 31.6% 30.6% 34.3% 36.9% 51.4% $75,000–$100,000 24.5% 23.6% 26.9% 29.3% 43.0% .$100,000 13.7% 13.1% 15.4% 17.2% 28.1% 76 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 benefit guarantee of up to the poverty level. in contrast, by providing investment choices like those available to federal workers and by not providing a minimum benefit guarantee, the kasich, moynihan-kerrey, and smith plans would likely result in greater differences in individual social security benefits between men and women than exist under current law. our results also suggest some good news for those who advocate recent ia proposals. first, our results suggest that average investment disparities will lessen—or even disappear—as the gap between women’s and men’s incomes continues to narrow. second, the identification of those most at risk for very conservative investments, such as lower-earning, younger women, may help policymakers target them for special social security account investment education or special social security benefit protections (such as the minimum benefit guarantees) suited to their needs. such targets may be desirable, because: (1) those most at risk from overly conservative investing in 401(k)s are those who will probably need retirement income supplementation from social security the most, and (2) this analysis has shown that young females are even more conservative in their investment of 401(k) assets than overall gender differences would suggest, and it is precisely younger age groups, of course, whose contributions would have the longest period of time to accumulate investment income (barring any loan or withdrawal behavior). references ameriks, j., king, f. p., & warshawsky, m. (1997). premium allocations and accumulations in tiaa-cref— trends in participant choices among asset classes and investment accounts.tiaa-cref research dialogues, 51. bajtelsmit, v. l. (1996). conservative pension investment: how much difference does it make?benefits quarterly, 12(2), 35–39. bajtelsmit, v. l., & vanderhei, j. l. (1997). risk aversion and pension investment choices. in mitchell, gordon et al. (eds.),positioning pensions for the twenty-first century.philadelphia, pa: university of pennsylvania press. ball, r. m. (1998).straight talk about social security: an analysis of the issues in the current debate.new york: the century foundation press. bowman, k. (1999).social security: a report on current polls.social security and pension studies. the american enterprise institute, washington, dc: april. buck consultants. (1997).401(k) plans: survey report on plan design—1997(9th ed.) new york: buck consultants. cerulli associates, inc. (1998).the cerulli report: the state of the pension and retirement markets.financial press copy. clark, r. l., & schieber, s. j. factors affecting participation rates and contribution levels in 401(k) plans. in o. s. mitchell and s. j. schieber (eds.),living with defined contribution pensions. remaking responsibility for retirement.philadelphia, pa: university of pennsylvania press. clark, r. l., goodfellow, g. p., schieber, s., & warwick, d. (1998). making the most of 401(k) plans: who’s choosing what and why? pension research council working paper. goodfellow, g. p., & schieber, s. j. (1997). investment of assets in self-directed retirement plans. in m. gordon, o.s. mitchell, & m. twinney (eds.),positioning pensions for the twenty first century.philadelphia, pa: university of pennsylvania press. hewitt associates (1997).survey findings: 401(k) trends and experience, 1997.lincolnshire, il: hewitt associates. hinz, r., mccarthy, d., & turner, j. (1997). are women conservative investors: gender differences in 77j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 participant-directed pension investments. in mitchell, g. et al. (eds.),positioning pensions for the twentyfirst century.philadelphia, pa: university of pennsylvania press. investment company institute. (1999).u.s. household ownership of mutual funds in 1998,washington, dc: march. kpmg peat marwick. (1998).retirement benefits in the 1990s: 1998 survey data.washington, dc: kpmg peat marwick. mercer, w. m. (1997).survey on employee savings plans, 1997.new york: william m. mercer. office of management and budget. (2000).budget of the united states government. fiscal year, 2000. papke, l. e. (1998). how are participants investing their accounts in participant-directed individual account pension plans?the american economic review, 88(2), 212–216. profit sharing/401(k) council of america. (1988).41st annual survey of profit sharing and 401(k) plans, reflecting 1997 plan year experience.chicago: profit sharing/401(k) council of america. sunden. (1998). gender differences in the allocation of assets in retirement savings plans.the american economic review,may. u. s. department of labor, pension and welfare benefit administration. (1998).abstract of 1994 form 5500 annual reports.washington, dc: u.s. government printing office. upston, a. (1998). public opinion on social security.ebri notes, 19(10). vanderhei, j. (1999a). demographic differences between the population covered by employment-based defined contributions plans and the total work force covered by social security. in d. salisbury. (ed.).beyond ideology: are individual social security accounts feasible?.washington, dc: employee benefit research institute. 117–122. vanderhei, j., galer, r., quick, c., & rea, j. (1999b). 401(k) plan asset allocation, account balances, and loan activity. ebri issue brief, 205. yakoboski, p. j., & vanderhei, j. l. (1996). worker investment decisions: an analysis of large 401(k) data.ebri issue,brief no. 176 (employee benefit research institute). 78 j.l. vanderhei, k.a. olsen / financial services review 9 (2000) 65–78 pii: s1057-0810(02)001142 (������ @���� � � ��4��� (��� �������� +�/���� 4�����-8��� ���������c��� �� �!"#$ ��1!�%1 (�� d����� �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� �� �!"#$ %&�!0�% (���� 4������ ��4��� (��� �������� +�/���� 4�����8��� ���������c��� �� �!"#$ ��1!�%1 7���f)���8� � � + � ��.������� � �*�$ >� ?��/ *�� ������������ �� �!"#$ %%�!%09 78+����� � ��3��� � � ��� ����� �� �������� � ���������� ������� ������� �� �!"#$ %�&!%%� 787����� �������� ��+��������� � '��8) 7��)�� ������ '�������$ � � ���� �* ��+ ����� 7������������ �� �!"#$ �10!�2b 7�����)�� 7�� � ����� ��a����3�� ��������� a����� 7����� 5����$ � 7���� ����� �������� �* � � :/�������� �� �!"#$ �"9!�b� 6��������� � @��/ ��� ��a����3�� ��������� a����� 7����� 5����$ � 7����-����� �������� �* � � :/�������� �� �!"#$ �"9!�b� 6������ ������ ��6���*���� 4������*�8����� �� � ?�� �� .�*��������� +������� 7��)����� �� �!"#$ 0�0! 0�" 6�������� �**��� ? � ��� ������������� �* �������8��3�� .���� ��� .���� ��� "�� )# .������������� �� �!"#$ 21!��� 6��*���� � � 5 � �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� �� �!"#$ %&�!0�% �� ���� ?����� ��.�������/���� ���)!������ ��������� �/ �� � � ����� 7��)��� ������ � � ����� ��������� �� �!"#$ %"&!%1% ���8 �������� ;������� ��� � a���� �* ���������� .�8��� '������$ � � a���� �* 7������� ������������� �� �!"#$ �&!09 '8 ������ 5������8) 6 � ��� '��8) '���8���� 7���� �����7������������'������������ �!"#$�%&!�"" '8���� ;������ ( � ��� � 5����8��� 7������=����� �8�$ >�< 6���/�8����� *�� � � 5����8� �����8������� �� �!"#$ �&1!%�2 '��� � '�����/� ��� � 5����8� �* ���)�� ������ ����� ����8�����$ '��� +�����8� *��� ��>�< g��)�� ������� '������ 6����������� �� �!"#$ ���!��b '/��=��� � � � ��� � 5����8� �* ���)�� ������ ����� ����8�����$ '��� +�����8� *��� ��>�< g��)�� ������� '������ 6����������� �� �!"#$ ���!��b ��=��� � ���� ��+��������� � '��8) 7��)�� ������ '�������$ � � ���� �* ��+ ����� 7������������ �� �!"#$ �10!�2b ����� ���� �� ������ ��5������ :/�� 7��)�� ��������� 7������� ��� '��8) 7��)�� 6��*�����8���� �� �!"#$ �b0!�1� ;������ ;������ � � �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� �� �!"#$ %&�!0�% ;���� ������� + � ���� .������������ �* � � ���������8� �* 5����8��� ����8� :**���� 3� ;�3-3���� '���8����� �� �!"#$ 9!�2 e� ��� ����)� ��.�������/���� ���)!������ ��������� �/ �� � � ����� 7��)��� ������ � � ����� ��������� �� �!"#$ %"&!%1% ������ ��� ������ (��) �� ������� '��������� *�� � ' +,���� ������-+�� .��������� ���/�������� '�� � ��������� � �� �������� ��� �**��� � ����� �� �!"#$ %01!%"2 ��� ���/������ �* '���� ���������� 4�*���� ����*�� ��� 4�*���� ������3����� 6������ 6����$ � 7���� ����� '������������ @�� � ������ ' �<� 5��3��� ��� � � ?����� �� �!"#$ 01!"" ��� ����� �� �������� � ���������� ������� ������� � � � ��������� � � � ����� � ?�8)� ��� ��3��� � 78+����� � �� �!"#$ %�&! %%� ��� ������������� �* ���-4���8��3�� .���� ��� .���� ��� "�� )# .������������� '��/ �� 7 ?���� ��� �**��� ? 6�������� �� �!"#$ 21!��� ��� '��8) '���8���� 7���� ����� 7������������ '���� ������ 5������8) 6 '8 ����� ��� '������ � +�)���� �� �!"#$ �%&!�"" ���� .������������ �* � � ���������8� �* 5����8��� ����8� :**���� 3� ;�3-����� '���8����� ������ ' ��88������ ��� ������� + ;���� �� �!"#$ 9!�2 ���������� �* .��������� � ��8�� *�� ����������$ � >�< �//���8 ��� 6���/�8������� a '������� @��� ��� ��� ' ��� (�<���8�� �� �!"#$ 19!2b �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� � � 5 6��*���� ;������ � ;������ ��� d���� (�� �� �!"#$ %&�!0�% �������������� ��� ������8� ����� �� ���������� '������ 4�8������$ ���������� +�����8� �* '��0�b ����� � ��� � � � �� �� �� ������ �� ������ ������� ������8 ������ 4�**����8����� 7������ ����)-7��/ � ��� 6��� �������� �� �!"#$ %10!%&� ��4��� (��� �������� +�/���� 4�����-���� ���������c��� @���� � (����� ��� 4����� (���� �� �!"#$ ��1!�%1 ��+��������� � '��8) 7��)�� ������ '�������$ � � ���� �* ��+ ����� 7������������ � ��� ��=�� ��� ������� 787����� �� �!"#$ �10!�2b ��5������ :/�� 7��)�� ��������� 7������� ��� '��8) 7��)�� 6��*�����8���� ���� �� ����� ����� �� �!"#$ �b0!�1� ��.������� � �*�$ >� ?��/ *�� ������������ � � + 7���f)���8 ��� @���� +����� (� ��� �� �!"#$ %%�!%09 ��.�������/���� ���)!������ ��������� �/ �� � � ����� 7��)��� ������ � � ����� ��������� +��8 ������� ?���� �� ���� ��� ����) e� ��� �� �!"#$ %"&!%1% ��(����������� � ����� �� >�� ;��� 3� ?���� ��� .�8��� ��� 4������/ �8 � ���8�������8� *�� � � 5���� � ��� ;���� �* � � ?�'��� @���� +����� (� �� ��� 4������ @��� �� �!"#$ 99!10 ��6���*���� 4������*�8����� �� � ?�� �� .�*��������� +������� 7��)����� 7�8 ��� + ?����� ?���� @����=� ��� ����� 6������ �� �!"#$ 0�0!0�" ��� � �56�7 �����*�8����� +���������� 6��8���$ � 4��8������ �* � � 7���*��� ����** '8����� 7�� ����� 4���� �� 3�� �� �!"#$ �21!�&9 ��� � 4�8����� �� 4�*��$ 5�8���� �**�8���� +�/����� 4�*����� .�8���������� 7 @�3�� ?����� ��� ' ��� (�<���8�� �� �!"#$ "9!9" ��� � 5����8� �* ���)�� ������ ����� ����8�����$ '��� +�����8� *��� ��>�< g��)�� ������� '������ 6����������� � � '/��=�� ��� '�����/ '��� � �� �!"#$ ���!��b ��� � 5����8��� 7������=����� �8�$ >�< 6���/�8����� 5�� � � 5����8� �����8������� +��)� ; ���3��� ��� ;������ ( '8���� �� �!"#$ �&1!%�2 ��� � a���� �* ���������� .�8��� '������$ � � a���� �* 7������� ������������� ;������ ; ������� ��� ;������ ���8 �������� �� �!"#$ �&!09 ��a����3�� ��������� a����� 7����� 5����$ � 7��������� �������� �* � � :/�������� 7�� � ���� 7�����)� ��� @��/ �� 6��������� � �� �!"#$ �"9!�b� ����� � ��� � � � �� �� �� ������ �� ������ ������� 0�1 pii: s1057-0810(01)00066-x design considerations for large public sector defined contribution plans< kevin w. sigrista,1, stewart l. brownb,* aflorida state board of administration bflorida state university, tallahassee, fl 32306, usa received 14 may 2000; received in revised form 20 october 2000; accepted 1 december 2000 abstract the paper identifies differences between private (401 (k)) plans, which have evolved under erisa and existing public plans, which have not. examination of model legislation reveals that public plans should largely conform to erisa going forward and reflect best practices in the private sector. empirical analysis of equity mutual funds with $1.8 trillion in assets and institutional equity accounts with $98 billion in assets demonstrates efficiencies in separately procured institutional investment, administrative and educational services relative to retail investment products. the analysis points to tension between the duties of trustees and the demands of participants requesting large numbers of retail investment options. © 2001 elsevier science inc. all rights reserved. jel classification:g23; g28; h55; h72; j26 1. introduction the public sector increasingly uses defined contribution (dc) pension plans, either as supplement or in place of the more traditional defined benefit (db) pension plans. optional dc programs have been in place for a number of years for public university faculty, but in recent years a number of large public employee retirement systems have adopted or are in < the views expressed in this paper are solely those of the authors and do not represent the positions of their employers. the authors would like to express their appreciation for helpful comments from jim francis, russ ivinjack and john freeman. * corresponding author. tel.:11-850-644-9657; fax:11-850-644-4225. e-mail addresses:sigrist_kevin@fsba.state.fl.us (k.w. sigrist), sbrown@cob.fsu.edu (s.l. brown). financial services review 9 (2000) 197–218 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00066-x the process of implementing dc arrangements (e.g., michigan, colorado, washington, louisiana, florida, montana, ohio, and south carolina). moreover, proposed social security reforms that would introduce self-directed accounts are essentially federal dc accounts. thus, investment policy issues unique to public sector dc plans are important and timely. there is little literature exploring the special circumstances of public sector dc plans and thus no clear enunciation of best practices. however, there are apparent differences in the organization and administration of public and private sector dc plans. the purpose of the paper is to address the policy issues encountered in the design of large public sector dc programs. many corporate dc plans, which have evolved under the employee retirement income security act (u.s. department of labor, 1974), utilize an institutional approach where services are unbundled; that is, trustees contract separately for components of investment advisory, educational and administrative services. morgan stanley (1999) reports that of 375 401(k) plans surveyed in 1999, about half of the programs purchased unbundled investment products and administrative services and more than half of the programs offered some institutional (nonretail) investment products to their participants. another survey reported that 80% of 401(k) plans with more than 10,000 participants were at least partially unbundled in 1999 (barra rogerscasey/ioma, 1999). this has the impact of minimizing plan expenses and increasing participant benefits. generally such plans are consistent with modern portfolio theory (mpt) and offer limited investment options. it is not unusual to find index funds in such plans. in contrast, public dc plans have evolved without the benefit of erisa. enabling laws and statutes are unique to each state. such plans often lack a trustee governance structure and commonly utilize a bundled investment architecture where vendors provide investment advisory, educational and administrative services under one umbrella fee. vendors in this sector include variable annuity and mutual fund companies and fees are often higher than the fees of unbundled dc plans. it is common to find multiple competing vendors in the same plan, negating some of the advantages of economies of scale and sometimes engendering conflicts of interest among vendors (jacobius, 2000). public plans typically have dozens and sometime hundreds of different investment options (kpmg, 1999; national association of deferred compensation administrators (nadca), 1998). trustees of new public dc pension programs, including potential social security reforms, must ultimately choose between the two approaches or implement a hybrid program. the salient issues are the number of investment options available to participants and the manner in which those products and other services are procured: 1) name-branded retail investment options which bundle together investment, administrative and educational services versus 2) an unbundled procurement of private label institutional/wholesale investment options which may be independent of administrative and educational services. other issues remain important, such as active versus passive management. however, actively and passively managed portfolios coexist comfortably in most corporate dc plans and we view this issue as outside the scope of the paper. the issue of the number and diversity of investment options is a thorny thicket constantly faced by public sector and corporate dc trustees. choice is commonly equated with freedom; the more the better. thus, for plan participants, there is an intuitive appeal and 198 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 comfort that arises from the ability to choose among a broad array of branded, hence familiar, investment products. vendors reinforce the importance of product diversity and trustees can be put in the uncomfortable position of defending design decisions with seemingly arcane academic arguments. however, even in a theoretically pure liberal democracy, freedom is constrained to protect property rights. similarly, trustees must critically evaluate the expected marginal costs and benefits to the overall group of participants associated with providing a diverse choice among duplicative investment products. in the main, corporate dc trustees have largely opted for a cost-effective approach that limits the investment options to a handful (i.e., 10 or less options) of relatively distinct products. moreover, ennisknupp (1999) surveyed corporate plan administrators and over two-thirds of 110 respondents replied that institutional accounts are more effective than mutual funds. we explore these issues in more depth in the body of the paper. it is organized as follows: we identify the major types of public plans, identify unique pressures on the organizers/ trustees of such plans, discuss the evolving legal infrastructure underpinning public pension plans and discuss model legislation: the 1997 uniform management of public employee retirement systems act (national conference of commissioners on uniform state laws, “mpers”). we then explore the implications of mpers for the investment/administrative design of a public dc plan and analyze a unique data set comparing the administrative and investment advisory costs associated with institutional accounts and mutual funds. finally, we explore the implications of a hybrid approach where institutional accounts and mutual funds coexist. mpers and mpt strongly suggest that trustees/fiduciaries implement an unbundled institutional approach. analysis reveals substantially greater economies of scale associated with institutional than mutual fund accounts. further analysis reveals that a hybrid structure has serious shortcomings because it results in higher costs (lessened economies of scale) for a subset of the participants: those who choose the institutional investment options. the higher costs associated with retail accounts significantly reduces the future benefits of participants. with an initial account value of $500 million, a 30-year horizon and annual market appreciation of 10%, large-cap mutual fund fees would lead to an account value that was $870 million lower than that produced by a large-cap institutional fund account. further, in hybrid structures, we show that for a 10-year horizon, participants that prefer the low-cost institutional option over the higher cost name-brand options are forced to pay about 40% more in cumulative fees because assets move into the branded options. 2. first principles: trusts, trustees and fiduciary duties garla (1998) provides a taxonomy of public-sector tax-qualified and nontax-qualified dc plans. section 403(b) and section 457 plans (designated by the internal revenue code section) are by far the most common public-sector dc-type plans. section 403(b) plans allow employees of state educational organizations and governmental entities to elect to have their employer make tax-deferred contributions for them to purchase an annuity contract or make contributions to a custodial account for investment in mutual fund shares. section 457 199k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 plans are deferred compensation arrangements, for employee contributions, that can be established by a state or other governmental entity. in 1996, section 457 was amended so that as of january 1, 1999 all assets and income of 457 plans must be held in trust for the exclusive benefit of participants and beneficiaries. section 401(k) plans are available to a small number of public employers who had them adopted prior to may 6, 1986. the more recent wave of new and large public sector dc plans have been established as section 401(a) money purchase plans. in 1986, the federal thrift savings plan (a 401(a) plan), currently with $85 billion in assets, was authorized for federal employees. as a rule, public sector retirement systems are exempt from the employment retirement income security act (erisa) of 1974 (employment retirement income security act, 1974), which governs corporate 401(k) programs. without the guiding principles of erisa, the laws regulating public sector dc plans have evolved separately and vary significantly across states. moreover, both db and dc public sector retirement plans have often lagged developments in modern financial theory and practice. garla (1998) surveys governmental retirement system law and reports the widespread legacy of statutory lists and other legal guidelines that govern permissible investments. legal considerations are typically ignored in the investment policy literature because erisa has provided a legal framework. absent erisa, specific laws and statutes are necessary to initiate a public sector program and govern responsibilities for the design and administration of the program. therefore, this section discusses the key legal, fiduciary and governance foundations that are necessary to establish a sound investment policy in the public sector context. 2.1. interested parties and the political system besides erisa, another difference between public sector plans and corporate plans is that interest groups routinely attempt to use the political process to enhance their well being and profitability—a phenomenon known generically as rent-seeking. rent-seeking is not the same as lobbying activity. the latter principally communicates information on political positions to decision-makers and their staff. such communications may include proprietary information and analysis relevant to assessing the economic impact of potential decisions. rent-seeking is an exercise in redistributing economic resources by obtaining preferential tax, regulatory or procurement policies. in the context of public retirement systems, rent-seeking behavior can take several forms. first, potential vendors may try to manipulate decision-makers through the statutory, budgetary or administrative law processes. second, potential vendors have apparently established economic and financial relationships with traditional lobbying organizations, such as local government associations (e.g., associations of counties and leagues of cities), unions and professional associations (pinkston, 2000; jacobius, 2000). third, some lobbying organizations, such as local government associations, may have captive money management operations. fourth, other interest groups can be expected to lobby for economically targeted investment policies. finally, trustees of a retirement system may face rent-seeking activity within the ranks of their professional staff, whose interests may be mis-aligned with those of participants, particularly if hiring and firing of staff is restricted by civil service laws. 200 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 the existence of interest group pressures dictate that implementing statutes must incorporate good financial and governance policy to ensure that participants and taxpayers are sheltered from the rent-seeking behavior of interest groups. such behavior generally increases the costs of dc plans and thus reduces the eventual benefits that accrue to participants. it is speculative to predict whether, and under what circumstances, these types of political efforts might cause public sector decision makers to take actions potentially adverse to the interests of participants. for instance, trustees are generally accomplished professionals with integrity and substance and possessing sufficient resources to help them recognize and understand their duties to participants. at the same time, the history of public sector db plans is instructive. political pressures have led to legal restrictions on investments (constitutional and statutory), economically targeted investments, social investments and occasional budget actions affecting contributions or plan assets (garla, 1998; garthwaite, 1999; useem & hess, 1999). 2.2. the uniform management of public employee retirement systems act importantly, there is movement toward standardizing the legal infrastructure governing public sector plans and incorporating some of the key components of erisa. the uniform management of public employee retirement systems act (“mpers”) (national conference of commissioners on uniform state laws, 1997) develops a uniform legal framework for the administration and operation of state and local public sector db and dc plans. mpers sets out much of the legal infrastructure that is necessary to facilitate an effective investment policy and the balance of this section explores certain aspects of mpers. one of the guiding principles of mpers (and erisa) is the requirement that assets of retirement systems be held in trust and that trustees have theexclusiveauthority to invest and manage those assets. further, mpers establishes that trustees should be sufficiently independent to effectively and efficiently perform their duties. among the exclusive powers that mpers confers are: the power to establish an administrative budget sufficient to perform the trustee’s duties and as appropriate and reasonable draw upon assets of the retirement system to fund the budget; the power to obtain by contract the services necessary to exercise the trustees’ duties; the power to procure and dispose of goods and property necessary to exercise the trustees’ powers and perform the trustee’s duties. trustee independence is essential because it permits them to perform their duties in the face of political pressure from “interested parties.” in the absence of independence, trustees may be forced to decide between fulfilling their fiduciary obligation to participants or following the suggestions of groups whose interests may not be aligned with those of the participants. the surest protection for trustees’ independence is through a constitutional amendment. a supermajority requirement for future changes to trustees’ powers, creates a close approximation to constitutional protection. of course, an independent board of trustees raises the specter of unethical behavior that 201k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 harms participants and taxpayers (self-dealing, bloated expenses etc.). the antidote to such behavior is to subject trustees to: stringent fiduciary duties and standards of care; a requirement to publish an investment policy statement consistent with those duties; a requirement to regularly report on investment performance net of fees and relative to financial market benchmarks; annual independent audits; and personal liability for any losses resulting from the breach of duties, with limited ability to shift fiduciary liability onto others. 2.3. mpers and the prudent expert rule the one major shortcoming of mpers is its incorporation of a prudent person standard of care rather than the more stringent prudent expert standard embodied in erisa and common law. the prudent expert standard requires that the retirement program’s assets are invested, on behalf of the participants, with the care, skill, and diligence that a prudent investor acting in a like manner would undertake. under this standard, if a fiduciary is not an expert, they have an obligation to obtain expert advice. the official mpers commentary explains that the prudent expert standard is too exacting given the diversity among existing public retirement systems. in other words, a lower standard is a legacy of existing public sector plans that evolved without erisa standards. invoking the lesser prudent man standard favors trustees at the expense of participants because trustee’s actions are judged on the basis of how trustees in other public plans have behaved. such behavior may not be exemplary. the prudent expert standard applies to corporate retirement plans under erisa and private trusts under the common law. there does not appear to be a compelling argument for a weaker standard for public sector plan participants. 2.4. requirements for trustees under mpers mpers requires that trustees must discharge their duties for the sole interest and exclusive purpose of providing benefits to plan participants and beneficiaries and defraying reasonable expenses of administering the plan. trustees must also discharge their duties by impartially taking into account any differing interests of participants and beneficiaries, by incurring only costs that are appropriate and reasonable and in accordance with a good faith interpretation of the laws governing the program. fiduciaries responsible for the investment or management of retirement assets are further required to: “consider a broad range of economic and financial circumstances in establishing investment strategies, to diversify investments, to make a reasonable effort to verify facts relevant to the investment program, and may consider the benefits created by an investment in addition to investment return only if the trustee determines the investment was prudent even without the collateral benefits.” collateral benefits refer to social investing, economically targeted investing and so forth mpers confers an affirmative duty for fiduciaries to adopt a detailed investment policy statement that incorporates the investment objectives, the desired rates of return and acceptable levels of risk for each asset class, guidelines for the delegation of authority and information on the types of reports to be used to evaluate investment performance. mpers 202 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 requirements comport reasonably well with current corporate practice. table 1 indicates the subject areas contained in the investment policy statements of a sample of 250 corporate 401(k) dc programs (barra rogerscasey/ioma, 1999). this source indicates that about one-half of corporate 401(k) programs have formally adopted investment policy statements, although chambers (1999) notes that there is no requirement under erisa to have an investment policy statement. dc programs are intended to be participant-directed where participants bear the risks of their investment decisions. therefore, mpers effectively incorporates section 404(c) of erisa: if a participant exercises control over the assets in their account, no program fiduciary is liable for any loss to a participant’s account which results from such participant’s exercise of control. to obtain this protection for program fiduciaries, participants must be given meaningful, independent control over the assets in their account with the opportunity to: choose from a broad range of investment alternatives that allow diversification within and among such alternatives; give investment instructions with a frequency that is appropriate in light of the market volatility of the investment alternatives; and obtain sufficient information to make informed investment decisions. the incorporation of erisa 404(c) is an important legal foundation in the design of a public sector dc plan. in order to obtain 404(c) protection, program fiduciaries must establish a program with investment, administrative and educational features that meet minimum standards. however, the fact that participants are in control of certain investment decisions does not relieve trustees of the duty to provide an appropriate design and conduct on-going monitoring of the program to ensure it operates as planned and remains competitive table 1 areas covered in 401(k) investment policy statements, 1999 all plans large plans monitoring investment options 80% 88% determining type of investment options 78% 76% setting 401(k) plan objectives 70% 68% selecting investment options 67% 74% determining number of investment options 61% 65% benchmarking investment options 61% 76% terminating investment options 60% 65% amending investment policy 48% 38% designating roles and responsibilities 46% 56% plan expenses 36% 41% enforcing investment policy 35% 26% administration guidelines/requirements 31% 18% communication guidelines/requirements 27% 21% other 4% 0% memo: number with statement/total sample 250/446 34/56 source: barra rogerscasey/ioma annual defined contribution survey. large plans have at least 10,000 participants. 203k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 with other alternatives. public sector trustees have a duty to establish and maintain “best practices” in the design of a dc program. as we shall see, cost considerations are central to those decisions. 3. public dc program objectives having reviewed the legal and fiduciary framework within which trustees should operate, it is relatively straightforward to identify a dc program’s objectives: 1. offer a diversified mix of low-cost investment options that span the risk-return spectrum and give participants the opportunity to accumulate portable retirement benefits. 2. offer investment options that avoid excessive risk, have a prudent degree of diversification relative to broad market indices and provide a long-term rate of return, net of all expenses and fees, that achieves or exceeds the returns on comparable market benchmark indices. 3. offer participants meaningful, independent control over the assets in their account with the opportunity to: a) obtain sufficient information about the plan and investment alternatives to make informed investment decisions; b) direct contributions and account balances between approved investment options with a frequency that is appropriate in light of the market volatility of the investment options; c) direct contributions and account balances between approved investment options without the limitation of fees or charges; and d) remove accrued benefits from the plan without undue delay or penalties. 4. offer participants cost-effective administrative and educational services that will help the plan maintain compliance with u.s. department of labor section 404(c) regulations and provide participants with impartial and balanced information about investment choices that will help facilitate their portfolio decisions. most of the objectives are direct offshoots of mpers, the prudent expert standard of care and the 404(c) requirements. others such as portability are not controversial. however, objective 4 deserves detailed discussion. 3.1. impartial and balanced educational services corporate dc educational services have significantly evolved and expanded under recent federal regulations. corporations were naturally averse to providing education to 401(k) participants that might later be interpreted by courts as investment advice. therefore, in 1996, the u.s. department of labor promulgated interpretive bulletin 96–1 (ib 96–1) to encourage the provision of educational services under 404(c) of erisa. ib 96–1 identified broad safe harbors for education, including: general financial and investment information, asset allocation models and interactive investment materials (e.g., worksheets and software). 204 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 the department of labor also officially stated their interest in the provision of impartial and balanced educational services. for example, to comply with the ib 96–1 safe harbors, asset allocation models and interactive investment materials must incorporate generallyaccepted investment theory and disclose all material assumptions. similarly, if interactive materials and allocation models identify specific investment products, the participant must be alerted to the existence of other substitutes and be given sources of information on those alternative products. in ib 96–1, the department stated that their intent behind these requirements was, “to address the concern that a service provider could effectively steer participants to a specific investment alternative by only specifying one particular fund in connection with an asset allocation model.” the department’s concerns are of central importance for public sector dc plans because they are not subject to the prohibited transactions regulations of erisa. under such regulations, investment product providers and other fiduciaries are generally prohibited from actions that would benefit them at the expense of the participants; for example, advising a participant to choose one of their investment offerings. therefore, unless public sector trustees take an affirmative action, in law or the program objectives, participants will not be afforded one of the basic protections enjoyed by members of corporate 401(k) programs. unfortunately, mpers does not invoke the prohibited transactions regulations of erisa. public sector trustees have a strong incentive to arrange cost-effective educational services to help facilitate investment decisions and improve retirement benefits. first, providing impartial and balanced educational information appears to be broadly consistent with the 404(c) objective of ensuring that participants remain in control of their assets. second, provision of even low levels of impartial educational services along the lines of ib 96–1 should entail significant increases in retirement benefits. studies of consumer investment behavior indicate a sizeable group of consumers do not understand or follow basic investment principals related to asset allocation and cost management.1 moreover, the rule of thumb is that more than 90% of dc retirement benefits will be determined by an investor’s asset allocation (ibbotson & kaplan, 2000) and every 100 basis points of excess cost lowers a dc account balance by approximately 20% over 30 years. trustees are also in a position to utilize the program’s group purchasing power to acquire a standardized package of educational services for substantially lower cost than can be acquired by individuals at retail prices. effectively, the long-term cost to taxpayers of a dc program should be lower per dollar of ultimate retirement benefit with relatively low educational expenditures. of course, trustees have a duty to monitor the cost effectiveness of their chosen package of educational services, since they have a duty to incur only costs that are appropriate and reasonable. 4. implications for general program design and procurement the objectives imply that the selection of all program vendors must be guided by “best-in-class” principles, unless there are significant perceived advantages to purchasing products and services bundled. however, there are several compelling arguments for pursuing an unbundled architecture with separately negotiated and procured investment, admin205k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 istrative and education services. first, it enhances the trustees’ negotiating position, as the purchasing agent for the participants. second, it provides the opportunity for a full disclosure and review of performance and all costs borne by participants and taxpayers. importantly, investment and administrative services are subject to different cost economics and there appears to be significantly different price structures in the retail and wholesale/institutional marketplaces (a point we empirically test below). for instance, a u.s. department of labor (1998) report on 401(k) costs stated: “larger plans enjoy potentially significant economies of scale. in the case of investment expenses, they have access to more providers offering a wide range of investment vehicles at lower cost. very large plans may be able to reduce investment expense even more through fee-reduction negotiations with the providers or use lower-cost institutional accounts. in other expense categories, the combination of flat (or nearly flat) fees regardless of plan size, plus declining per-capita charges for basic administration fee, reduce per-participant administrative costs among larger plans.” finally, an independent education vendor is needed to provide participants with impartial and balanced information about investment choices. an independent educator can sidestep the inherent conflicts of interest that arise when an investment product vendor educates and counsels on asset allocation, financial planning issues and investment fees/costs in the context of their products. an independent education vendor is also the natural agent to empower participants with regular, understandable and standardized disclosure of product fees and investment option performance versus market index benchmarks. 5. implications for investment program design and procurement the number and types of investment options offered in a dc plan are critical. the overriding objective should be to provide a range of options that allows participants to choose a point as close as possible to the efficient frontier consistent with individual risk preferences. within the context of the program objectives, investment options should be consistent with mpt, finance theory and best practices. there are three interrelated issues influencing the number and types of investment options: diversification, costs and active management. assume for a moment that markets are perfectly efficient and consider the issue of diversification. in a perfectly efficient market, index funds are the investment vehicle of choice and investment management costs are minimal. how many and what types of investment vehicles are necessary to “span the risk/return spectrum?” at least three: us stocks, us bonds and cash equivalents. for many years, those were essentially the options available to federal employees in the federal thrift savings plan. these three choices minimally meet the diversification requirement of 404(c). however, we would argue this concise list should be expanded to include two categories: foreign stocks and inflation-indexed bonds. an asset type should be defined at the highest pragmatic level of aggregation possible, where the next highest level of aggregation would combine securities with materially different legal, financial and economic characteristics. foreign 206 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 stocks and inflation-indexed bonds have materially different characteristics than the three options. providing five index fund options representing the broadest representations of these asset categories would minimally satisfy the program’s investment objectives. among the set of liquid, highly diversified and unleveraged portfolios composed of public market securities, these five index options would qualify as low-cost and effectively span the lion’s share of the risk and return spectrum. importantly, these options span the short-term market risk dimension and the long-term income replacement risk dimension. the latter is defined by the possibility that a participant’s account balance will prove to be insufficient to maintain a reasonable post-retirement standard of living. over time, an employee’s income tends to grow with inflation, improving productivity and prevailing market wages for similar occupations and levels of managerial responsibility. u.s. stocks have historically provided high average annual real returns and this long-term real return has been largely unaffected by the rate of inflation (boudoukh & richardson, 1993). the 1997 introduction of inflation-indexed bonds (tips) in the u.s. created a new asset type with extremely attractive characteristics for tax-deferred retirement accounts. ibbotson and kaplan (2000) examined performance for balanced mutual funds and large db pension funds and found that 99% to 112% of the total return level was explained by the fund’s long-term asset allocation (the authors subdivided u.s. stocks into largeand smallcap sectors). thus, five indexed portfolios representing broad asset types would allow plan participants to articulate their risk preferences. the average contribution of active management in the ibbotson and kaplan study was less than invigorating. however, the appropriate use of active management strategies is outside the scope of this paper. for guidance, it is useful to turn to best practices in the 401(k) arena where plan sponsors are subject to erisa. table 2 provides summary data on 401(k) programs’ investment options (barra rogerscasey/ioma, 1999). despite diversity in the aggregate, this same survey indicated that the median number of options was between 8 and 9 in 1999 and the majority of program administrators believed that the number of investment options became excessive around 11. a representative list of 8 options that generally comports with these corporate practices and recent academic research would be: money market option; total market nominal bond option; inflation-indexed bond option; total market u.s. stock index option; u.s. small stock option; u.s. value stock option; u.s. growth stock option; and foreign stock option. in addition, two or three balanced funds, optimized to reside on the efficient frontier, would provide one-stop shopping for participants reluctant to perform their own asset allocation analysis. waring et al. (2000) make a persuasive case that dc investment options should predominantly be optimized balanced funds. in their model design, asset type options should function as “specialty options.” in part, they argue that participants should not be put in an environment that facilitates focusing on the short run game of trying to pick hot funds, to the exclusion of focusing on more important long run considerations, such as asset allocation. this model is not commonly used in corporate or public sector dc plans. as indicated in the introduction, the relatively low numbers of options in corporate dc plans stand in stark contrast to the high numbers of investment options that are often put before participants in public sector plans. moreover, public sector 403(b) and 457 dc plans often utilize multiple full-service bundled providers to generate diversity of investment 207k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 options within the program; a practice that is virtually unheard of in 401(k) programs. clearly, duplicative administrative and educational services bloat costs and reduce overall benefits. moreover, bundled providers have incentives to favor their branded alternatives, which creates an inherent conflict of interest in any educational services they might offer. in the next section, we empirically identify performance factors that support an unbundled procurement of private label institutional/wholesale investment options. 6. comparative cost analysis table 3 provides cost data from 1996 that illustrates the basic attraction of institutional separate accounts relative to mutual funds: costs are far lower. institutional mutual funds are special share classes with high initial minimum balances (e.g., $100,000) and limited provision of administrative services to the client, as well as lower charges for marketing and distribution. institutional separate accounts typically operate wherein a custody bank safekeeps securities in a group account and a manager(s) is authorized to trade securities in that account; individual and collective recordkeeping is performed by the custody agent or a third party recordkeeper. institutional commingled accounts operate in a similar fashion as an institutional mutual fund, but pricing is generally at the level of separate accounts. table 2 investment options offered by corporate 401(k) programs in 1999 percentage of plans all plans offering option large plans balanced fund 77% 77% international equity 71% 70% u.s. large cap equity 70% 72% u.s. equity—indexed 66% 75% u.s. bonds 63% 65% u.s. small cap equity 61% 51% stable value (gics/synthetics) 60% 74% money market 56% 42% u.s. mid cap equity 48% 40% company stock 28% 56% global equity 27% 21% lifestyle options 19% 25% emerging markets 15% 9% international bonds 8% 2% real estate investment trusts 6% 5% self-directed brokerage 9% 4% mutual fund window 7% 4% other 26% 30% memo: respondents 448 57 source:barra rogerscasey/ioma annual defined contribution survey. large plans have at least 10,000 participants. 208 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 it is important to note that the mutual fund costs in table 3 include investment and administrative fees, but the institutional separate accounts include only investment fees. however, institutional accounts with more than $25 million in assets should enjoy even lower investment advisory fees because their fees for institutional separate and commingled accounts (institutional accounts) are generally negotiated on a sliding fee basis according to the level of assets under management. in contrast, retail mutual funds charge all owners of each share class the same expenses, regardless of whether the owner has 100 shares or 2,000,000 shares. the empirical analysis in this section compares investment advisory fee levels for institutional accounts and mutual funds and tests whether either vehicle lowers their investment advisory fees as assets under management increase. we find that institutional accounts have much lower investment advisory fees than retail mutual funds and provide significantly greater fee concessions as assets under management grow. in contrast, mutual funds provide fee concessions for administrative services. however, these services can be independently procured at costs competitive with the mutual funds’ charges. our empirical analysis uses two data sets compiled by freeman and brown (2001) to compare the costs of actively managed domestic equity mutual funds and externally managed equity institutional investment products. readers are referred to that paper for methodological details. freeman and brown sent inquiries to the 100 largest db public pension funds listed in the january 25, 1999 edition of pensions and investments. data for 1999 was collected on 220 individual external actively managed domestic equity portfolios with a total of $97.5 billion in assets. the average portfolio size was $443 million, with the range extending from $15 million to $4.8 billion. morningstar’s principia pro compilation for october 1999 was the chief source of mutual fund data. after eliminating funds with zero assets and missing data, the sample consisted of 4,943 equity funds. multiclass funds were aggregated into single funds where weighted averages of various expense ratios were obtained, using subfund assets as weights. investment advisory fee and administrative fee ratios were separately compiled, where administable 3 average mutual fund expense ratios and separate account management fees data from 1996 and expressed as percent of assets fund categories most common retail mutual funds in dc plans institutional mutual funds $25 million separate account indexed u.s. stocks 0.27% 0.35% 0.13% active large stocks 0.83% 0.91% 0.63% active small stocks 1.06% 1.01% 0.95% foreign stocks 1.33% 1.15% 0.75% active u.s. bonds na 0.69% 0.37% source: cerulli associates data cited in u.s. department of labor publication: study of 401(k) plan fees and expenses, april 13, 1998, contract no. j-p-7-0046, task order 209k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 trative fees were defined to exclude marketing and distribution fees (e.g., 12b-1). administrative fees include transfer agent fees, custodial services, accounting fees and director’s fees. screens were applied to generate a sample of mutual funds closely corresponding to characteristics of portfolios of public pension funds. the final mutual fund sample consisted of 1,343 funds representing a total market value of about $1.77 trillion. the average portfolio size was $1.3 billion, with the range extending from $15 million to $92.2 billion. within the sample are 1205 retail mutual funds with a market value of $1.71 trillion and 138 institutional mutual funds with a market value of $56 billion. institutional mutual funds were defined as single-class funds having a minimum initial balance of $100,000 or more. importantly, there are no significant differences between the cost economics of providing investment advisory services to mutual funds and db pension fund separate accounts. investment management firms commonly offer nearly identical investment products in both the institutional and retail markets. investment management firms also access the same markets for inputs: human capital, information and technology. thus, investment advisory services are essentially a commodity and fungible between mutual funds and institutional accounts. however, there are significant administrative cost differences between mutual funds and db pension fund separate accounts. mutual funds are valued daily, rather than monthly. also, retail mutual fund administrative fees incorporate accounting/record-keeping costs at the individual account level. these accounts are much smaller than the typical db institutional account. smaller account size entails higher average administrative fees. therefore, we separate administrative fees and investment advisory fees in the following analysis. this paper expands on freeman and brown’s analysis in several ways. first, a more efficient simultaneous equation model is used to test for different economies of scale for mutual fund administrative fees and investment advisory fees. second, cost equations utilize dummy variables to capture different cost structures for large-cap mutual funds, midcap mutual funds and small-cap mutual funds. regression equations shown in 1 and 2 below were run on the mutual fund data and the results are presented in table 4. expense ratios are scaled in basis points and size is scaled in millions of dollars under management. because of the exhaustive classification of the mutual funds, no intercept is included in the equations. 1. administrative expense ratio5 a(1)*(natural log of size)1a(2)*(large-cap dummy)1a(3)*(mid-cap dummy)1a(4)*(small-cap dummy) 2. investment advisory expense ratio5 b(1)*(natural log of size)1b(2)*(large-cap dummy)1b(3)*(mid-cap dummy)1b(4)*(small-cap dummy) the regression equations are highly statistically significant, as are all of the individual coefficients, except the size coefficient for institutional mutual funds. significant economies of scale are apparent for both administrative expenses and investment advisory expenses. however, scale economies are roughly 2.5 times greater for administrative fees than investment advisory fees. this difference is statistically significant for the retail mutual funds and all mutual funds. for example, for a $500 million investment in a large-cap retail mutual fund portfolio, total administrative costs are about one-half the investment advisory fee.2 the gap in expense ratios between retail and institutional mutual funds is generally consistent with the fees shown in table 3; noting that this latter report aggregated admin210 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 istrative and investment advisory fees. the smaller, although still significant, economies of scale present in the administrative services of institutional mutual funds is likely a result of the fact that they are bundled with materially lower administrative services and fees (see fig. 1). since institutional mutual fund share classes often perform limited individual services, they have to prepare fewer prospectuses, allocate fewer phone representatives, compile and maintain limited account data and so forth regressions of the form described in 3 below were run on the pension fund data. expense ratios are scaled in basis points and size is scaled in millions of dollars under management. because of the lack of an exhaustive classification of the pension fund data, an intercept is included in the equation. 3. investment advisory expense ratio5 c(1)1c(2)*(natural log of size)1c(3)*(largecap dummy)1c(4)*(mid-cap dummy)1c(5)*(small-cap dummy) table 5 contains the pension fund regression results and the comparable mutual fund results from table 4. the analysis indicates that at low asset levels, investment advisory fees for the pension funds and institutional mutual funds are comparable, but are significantly table 4 simultaneous equations estimates of active domestic equity mutual fund administrative and investment advisory fees per $10,000 of assets under management coefficient estimates (t-statistics) all mutual funds retail mutual funds institutional mutual funds administrative expense ratio natural log of size 26.4 26.9 23.6 (214.1) (214.3) (23.0) large-cap dummy 71.5 75.3 45.9 (28.4) (28.0) (7.1) mid-cap dummy 72.0 74.8 53.7 (29.5) (28.8) (8.4) small-cap dummy 71.8 75.2 48.8 (28.8) (28.8) (7.1) r-squared 18.7% 20.8% 9.6% investment advisory expense ratio natural log of size 22.5 22.8 22.8 (26.8) (27.4) (21.7) large-cap dummy 79.1 82.3 64.9 (35.3) (35.3) (7.0) mid-cap dummy 88.5 89.7 91.2 (37.8) (37.1) (9.8) small-cap dummy 95.6 97.4 93.1 (46.1) (46.8) (10.0) r-squared 14.8% 15.0% 24.9% sample size 1343 1205 138 assets in billions $1,769 $1,713 $56 chi-square test on difference of natural log size coefficients chi-square/probability 39.6/0.0% 39.3/0.0% 0.13/71.5% 211k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 below that of retail mutual funds for the large-cap category. significant economies of scale are apparent in the pension fund account investment advisory fees. the slope coefficients for the pension fund regression is 2 times greater than the mutual fund regression reflecting that pension fund advisory fees are twice as sensitive to assets under management than mutual fund fees. at larger asset levels, institutional mutual funds have lower fee structures than table 5 estimates of active domestic equity pension fund and mutual fund investment advisory fees per $10,000 of assets under management coefficient estimates (t-statistics) pension fund retail mutual funds institutional mutual funds investment advisory expense ratio natural log of size 26.1 22.8 22.8 (25.0) (27.4) (21.7) intercept 66.8 na na (8.6) large-cap dummy 25.1 82.3 64.9 (22.1) (35.3) (7.0) mid-cap dummy 13.5 89.7 91.2 (2.6) (37.1) (9.8) small-cap dummy 24.8 97.4 93.1 (5.8) (46.8) (10.0) r-squared 50.8% 15.0% 24.9% sample size 220 1205 138 assets in billions $98 $1,713 $56 fig. 1. domestic equity administrative expense ratios for active retail mutual funds and active institutional mutual funds 212 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 retail mutual funds, but do not appear to be competitive with the pension fund institutional accounts (see fig. 2). the economic impact of the differential fee structures is significant. for example, for a $500 million investment, the coefficients in table 5 predict an investment advisory fee of 65 basis points for the large-cap mutual fund versus 24 basis points for a large-cap pension fund account. for an initial account value of $500 million, a 30-year horizon and annual market appreciation at 10%, the large-cap mutual fund fees would lead to an account value that was $870 million lower than that produced by the large-cap pension fund account. it is clear why institutional accounts are considered a best practice arrangement in the corporate 401(k) arena: they have investment advisory costs that are one-half to two-thirds lower than retail mutual funds and are able to deliver additional cost savings as assets under management grow. similarly, mutual fund’s practice of passing through the benefits of economies of scale in administrative services, but not in investment advisory services is an important motivator for trustees to negotiate and procure unbundled investment products and administrative services. freeman and brown (2001) conclude that the chief reason for substantial investment advisory fee level differences between equity db pension fund portfolio managers and equity mutual fund portfolio managers is that advisory fees in the pension field are determined in a marketplace where arm’s length bargaining occurs. they argue that mutual fund shareholders as a rule do not benefit from arm’s length bargaining due to a dysfunctional regulatory system, and case law that, to date, has unduly favored advisory firms at the expense of their captive funds’ shareholders. mutual fund boards are typically dominated by employees of the investment advisory firm, even though the mutual fund is to be operated in the interests of the mutual fund shareholders. similar to investment advisory services, administrative services are essentially a commodity and there is limited product differentiation. unlike advisory services, however, fig. 2. domestic equity investment advisory fees for active retail mutual funds, active institutional mutual funds and active db pension accounts 213k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 competition appears to be more vibrant in the administrative services arena. it is common for smalland midsized mutual funds to purchase administrative services from third parties; rather than from their affiliated investment management companies. the natural economies of scale in administrative lines of business are demanded by the mutual fund companies and then, at least in part, passed on to mutual fund shareholders. under mpers trustees have a duty to, “make a reasonable effort to verify facts relevant to the investment program.” the basic industry cost information presented above strongly argues that trustees procure separate bids for administrative services and investment advisory services. trustees and consultants should be able to obtain sufficient information to identify a normal operating profit rate for the stand-alone administrative services, given the dc plan’s size and other circumstances. moreover, all else equal, a large dc plan with an average dc account balance that exceeds average mutual fund account balances should be able to use an independent administrative vendor to meet or beat the cost structure of retail mutual funds. as noted above, flat or declining per-capita charges for basic administration fees, reduce per-participant administrative costs among larger plans when expressed as a percentage of assets (u.s. department of labor, 1998). separate procurement of administrative services eliminates the mutual fund as a middle-man and any associated profit margin built into its procurement of services for shareholders. 7. the trade-off between participant choice and effectiveness the existence of economies of scale in institutional investment products sheds a negative light on the practice of offering multiple versions of the same type of investment option; for example, two active large-cap options. the common rationale for offering multiple versions of the same option is to afford participants choice; with the expectation that participants might be able to exit a poorly performing fund and choose a substitute that performs better going forward. however, trustees must critically evaluate the expected marginal costs and benefits associated with providing choice on this dimension. even with a sophisticated participant base, it seems unlikely that the expected benefits of expanding the choice of duplicative investment options would materialize. there is a major gap between the information retail investors can obtain on retail products and the information professional staff can obtain on institutional products. in the search and monitoring phases for institutional pension fund managers, professional staff typically gathers and analyzes detailed information on aggregate performance and dispersion across subaccounts, investment strategies, real-time holdings data and trading activity, portfolio management, trading and research personnel, risk management processes, trading arrangements, trading effectiveness, and so forth. in contrast, retail mutual fund investors are generally limited to historical aggregate return information, semiannual holdings data and historical turnover statistics. unfortunately, recent studies (carhart, 1997; kahn & rudd, 1995; malkiel, 1995) show that past performance has virtually no ability to predict future performance by domestic equity managers and has mixed results for fixed income managers. on the other hand, the expected marginal costs resulting from expanded choice are certain and material. first, for a given asset base, expanding the set of options compromises a 214 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 program’s ability to maintain low costs. for instance, splitting a $1 billion large-cap active institutional mandate into three institutional mandates causes costs to rise by 34%. additionally, under an institutional product structure, there are cost externalities created by facilitating choice among duplicative products. these externalities are a nontrivial matter, given the duty of impartiality: trustees’ obligation to consider the impact of design decisions by taking into account any differing interests of participants. table 6 provides an estimate of the externality created when two mutual funds are introduced to compete with an identical institutional fund. we assume that after the expansion of choice, each option gets one-third of the assets that had originally been invested in the institutional option. this assumption is conservatively based on data from fidelity institutional retirement services company (1999) and the recent allocation of participant balances within the state of michigan dc plan. for a ten year horizon, participants that prefer the low-cost generic institutional option over the higher cost name-brand options are forced to pay about 40% more in fees because assets move into the branded options. this analysis also points to the unavoidable tension between participant desires, especially the demands of vocal subsets of participants requesting more investment options, and the trustees’ duties to the overall group of participants. using the cost analysis from above, it is possible to demonstrate that fiduciary duty precludes trustees from maximizing participants’ choice of investment options; to do so would be in direct conflict with the trustees’ obligation to discharge their duties for the sole interest and exclusive purpose of providing benefits. imagine that a poll of participants were conducted regarding their nominees for investment options, prior to drafting an investment policy statement or conducting an education program. it seems reasonable to expect that their preferences would resemble the retail mutual fund market, with a smattering of bank deposits and insurance products. adopting an investment policy statement that incorporated each of these nominated products would maximize product choice from each individual’s perspective. however, the high cost of such a “maximal product choice design” would substantially erode terminal retirement benefits for participants in the aggregate, relative to benefits provided through a set of low-cost investment options that represent the basic asset types. consolidating all assets in comparable retail investment options (e.g., large-cap domestic equities) into a single comparable retail product would create significant economies of scale in administration and advisory fees for participants. in turn, that retail product would be higher cost than an equivalent institutional mutual fund share class; which would be higher cost than an institutional separate account. finally, in terms of investment performance, there is no basis to expect that higher cost table 6 allocation of 10-year cumulative costs for $1 billion active large cap domestic equity structures. all institutional funds and institutional/mutual fund mix (cost in millions) all institutional 1 institutional/2 mutual funds direct cost to mutual fund investors $17.56 $71.41 direct cost to institutional users $8.78 $8.78 indirect cost to institutional users $3.60 total costs $26.34 $83.79 215k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 investment products (i.e., retail vs. institutional) are more effective. in fact, large plan db and dc practice suggests quite the opposite. domestic equity mutual funds are rarely used within large db plans and ennisknupp (1999) reveals the vast majority of corporate dc plan administrators believe institutional accounts are more effective than mutual funds. 8. conclusion new public sector dc plans face important challenges in program design. however, there is a clear legal, fiduciary and governance infrastructure that should be adopted. with such an appropriate infrastructure, public sector plans should largely conform to erisa going forward and reflect best practices in the private sector. the major challenge facing public sector trustees is to reconcile fiduciary duties with the desires of individual plan participants. erisa, mpt and best practices in the corporate arena strongly suggest that the number of investment options should be limited and an institutional structure should be utilized where all costs are identified separately and tightly controlled. in contrast, participants and other interested parties can be expected to argue vociferously for an expansive list of recognizable, branded investment options. in the context of proposed social security reforms, the massive stakes for retail investment firms can be expected to generate proportionally large rent-seeking activities. based on empirical analysis presented above, the two views appear difficult to reconcile. the retail approach, although intuitively appealing and familiar will result in higher costs and lower aggregate participant benefits in the long run. ultimately, trustees must choose between what is right and what is easy and popular. giving trustees sufficient independence and subjecting them to stringent fiduciary duties should ensure that they make the appropriate design choices. notes 1. a 1999 national survey of defined contribution plans by john hancock financial services indicates that less than one-quarter of respondents consider themselves knowledgeable investors. forty-one percentage believe that money market funds include equities and 49% believe they contain bonds. most have no strategy or goal for allocating their retirement investments. most say they would pull money out of equities or reduce future allocations to equities if the stock market experienced a significant downturn (john hancock financial services, 1999). recent academic studies indicate that individual 401(k) investors naively diversify among the available investment options and individual investors tend to chase the hottest performing sectors, trading excessively and generating poor results (benartzi & thaler, 2001; barber & odean, 1999; benartzi, 2001). a 1996 survey conducted by the securities and exchange commission and the comptroller of the currency found that fewer than one in five mutual fund investors could give any estimate of expenses for their largest mutual fund, and that fewer than one in six fund investors understood that higher expenses can lead to lower investment returns. 216 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 2. a full information likelihood technique was used to estimate the simultaneous equations. in addition, standard errors and covariances were estimated in a single equation context using newey-west hac and white heteroskedasticity-consistent technique. neither technique materially affected the overall results. the general results were also insensitive to excluding the smallest and largest 10% of the retail or institutional mutual fund sample. results were generally insensitive to approaches using morningstar’s equity style categories to segregate the sample or develop exogenous variables. regressions were also performed on subsets of the mutual fund universe using simple one variable equations in order to determine whether scale economies might differ for large-cap mutual funds, midcap mutual funds and small-cap mutual funds. results for administrative expenses for retail mutual funds are generally consistent with those from the systems approach presented in table 4. however, investment advisory expenses appear to differ materially for the midcap mutual funds (no economies of scale) and small-cap mutual funds (economies of scale comparable to administrative expense) relative to results in table 4. nonetheless, at $500 million of assets under management the small-cap mutual fund coefficients in table 4 predict an investment advisory fee of 80 basis points versus the 76 basis points predicted by the single equation coefficients. the predicted results for midcap funds are similarly close. references barber, b. m., & odean, t. (1999). the courage of misguided convictions.financial analysts journal, 55(6), 41–55. barra rogers casey/ioma. (1999).1999 annual defined contribution survey.darien, ct. benartzi, s. (2001). excessive extrapolation and the allocation of 401(k) accounts to company stock, forthcoming;the journal of finance. benartzi, s., & thaler, r. h. (2001). naı¨ve diversification strategies in defined contribution savings plans, forthcoming;american economic review. boudoukh, j., & richardson, m. (1993). stock returns and inflation: a long-horizon perspective.american economic review, 83(5), 1346–1355. carhart, m. m. (1997). on persistence in mutual fund performance.the journal of finance, 52(1), 57–82. chambers, j. c. (1999). investment policy statements for defined contribution plans.journal of pension benefits, 7(1), 29–33. ennisknupp. (1999).survey of corporate plan sponsors,chicago, il. employee retirement income security act (“erisa”) of. (1974). sec. 2(a), pub. l. no. 93–406,88 stat. 829 (1974) (codified at 29 u.s.c. ch. 18), 29 u.s.c. sec. 1001(a) (1994). fidelity institutional retirement services company. (1999).building futures: how american companies are helping their employees retire, a report on corporate defined contribution plans.boston, ma. freeman, j., & brown, s. l. (2001). mutual fund advisory fees: the cost of conflicts of interest.journal of corporation law,forthcoming, 25 (3). garla, b. e. (1998)., investments of governmental plans of states and their political subdivisions. 415 pli/tax 391,practicing law institute. garthwaite, c. (1999). playing politics with retirement accounts.american legislative exchange council issue analysis. ibbotson, r. g., & kaplan, p. d. (2000). does asset allocation policy explain 40, 90, or 100 percent of performance?financial analysts journal,56(1), 26–33. jacobius, a. (2000). nea 403(b) endorsement shifting to security benefit as nationwide exits market.pensions and investments, 28(21), 44–45. 217k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 john hancock financial services. (1999). insight into participant investment knowledge and behavior, sixth defined contribution plan survey. boston, ma. kahn, r. n., & rudd, a. (1995). does historical performance predict future performance.financial analysts journal, 51(6), 43–52. kpmg peat marwick llp. (1998).retirement benefits in the 1990s: 1998 survey data.new york, ny. kpmg. malkiel, b. g. (1995). returns from investing in equity mutual funds 1971 to 1991.the journal of finance, 50(2), 549–572. morgan stanley dean witter investment management defined contribution services. (1999).1999 survey of the 410(k) market,new york, ny. national association of deferred compensation administrators (nadca). (1998).1997 survey of 457 plans. national association of deferred compensation administrators. lexington, ky. national conference of commissioners on uniform state laws. (1997).uniform management of public employee retirement systems act.chicago, il. pinkston, w. (2000). new retirement program is created for georgia counties.wall street journal, southeastern edition. september, 20. useem, m., & hess, d. (1999). governance and investments of public pension funds.pension research council working paper, 99–11.wharton school. philadelphia, pa. u. s. department of labor, pension and welfare benefits administration. (1996). 29 cfr part 2509 interpretive bulletin 96–1;participant investment education,federal register61 (113), 29586–29590. washington, d.c. u. s. department of labor, pension welfare benefits administration. (1998).study of 401(k) plan fees and expenses, contract no. j-p-7–0046, task order 1.washington, d.c. waring, m. b., harbert, l. d., & seigel, l. b. (2000). mind the gap! why dc plans underperform db plans, and how to fix them,investment insights, barclays global investors, 3 (1), san francisco, ca. 218 k.w. sigrist, s.l. brown / financial services review 9 (2000) 197–218 pii: s1057-0810(97)90021-4 financial services review, 6(2): 97-107 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. an application of fuzzy set theory to the individual investor problem manuel tarrazo this study reviews the problem of the individual investor and applies to it a methodol ogy based on fuzzy sets and the theory of possibility. the investment decision is characterized by uncertainty, imprecision and complexity, which lessen the eflective ness of conventional calculus andprobability tools. in contrast, fuzzy set theory and its modeling language provide objects of analysis and algebra that are well suited to this problem. new concepts such as ‘fuzzy portfolio weights” are introduced. the result of our research is a qualitative, general, and practical model for individual investors’ decision making, which is based on smith’s (1974) asset-mix model. i. introduction financial planning and investing have not become any easier. the stakes seem higher today because “normal” standards of living require more funds than ever. there are uncer tainties regarding social security and the health system, and insecurity regarding job per manency. therefore, individuals must assume greater responsibility for their future. with respect to previous research, the first model developed to assist investors in secu rity selection is owed to markowitz (1959, 1991) and is part of what is known as portfolio management. considerations regarding transaction costs and minimum budgets needed to achieve diversification made portfolio management most useful to large or institutional investors. the literature on limited diversification tries to adapt portfolio theory to small investors as in, for example, jacob (1974) and brennan (1975). in spite of their practicality, interest devoted to limited diversification models waned with the arrival of mutual funds. shortcomings of the methodology employed also exacted a high price in terms of analytical and computational cost, as in the model by pogue (1970) and precluded further advancement. state-preference models were developed to cope with uncertainty regarding the environment where investment decisions take place, hirshleifer (1965), myers (1968), and kraus and litzenberger (1975). state-preference models cannot be used in actual investment decisions because they are too precise, use utility functions, manuel tarrazo, ph.d. l assistant professor of finance, mclaren school of business, university of san francisco, 2130 fulton st., san francisco, ca 94117-1080; e-mail: tarrazom@usfca.edu. 98 financial services review 6(2) 1997 and require perfect knowledge (including returns on contingent securities that are not risky themselves). lastly, they require the user to input the very key item the model should pro vide: the optimal weights for the portfolio. zadeh (1965) is the seminal study on fuzzy sets. the inability of the conventional mathematical apparatus to assist in explaining systems that are characterized by uncer tainty, complexity, and vague or imprecise concepts and variables motivated this researcher. financial systems have these qualities, perhaps because they are ever-changing. these systems call for a different methodology and different research goals, as stated in zadeh’s principle of incompatibility: “as the complexity of the system increases, our abil ity to make precise and yet significant statements about its behavior diminishes until a threshold is reached beyond which precision and significance (or relevance) become almost mutually exclusive characteristics” (zimmermann, 1991, p. 3). ever since zadeh’s (1965) publication, there has been an explosion of papers, mono graphs and applications, mostly in the areas of information theory, computer sciences, and engineering. fuzzy set-based research has provided concepts, objects and frameworks of analysis for including imprecision, complexity and uncertainty, and it is the only formal methodology which permits linguistic variables and approximate reasoning to be used (zadeh, 1979, 1975a, 1975b, 1975c, 1973). why aren’t there more contributions of fuzzy set theory in economics and finance? kickert notes that applications in fuzzy sets often are designed to highlight the abilities of fuzzy methodology rather than first selecting the critical problem and then applying the methodology. kicker? also indicates the preference of researchers for what is known as “fuzzyfying” a problem, which is extending standard models into fuzzy versions (kickert, 1978, especially pp. 2, 22). what is gained in the application of fuzzy sets to the individual investor problem? in brief, fuzzy sets provide the algebraic tools and operations to permit the asset allocation problem to be solved in a logically proper manner. one cannot apply arithmetic and calcu lus to variables that are only known approximately. moreover, the methodology presented is also closer to the type of reasoning which is actually used by investors themselves. ii. smith’s qualitative model smith (1974) indicates that his main motivation in developing his model is to study the asset allocation problem, which was probably the most important problem faced by indi vidual investors, and yet had received little attention in the literature. to develop his model, smith breaks away from convention and develops a qualitative, but practical, inf. about assets -> investment attributes <inf. individual 1 multiasset portfolio decisions figure 1. the elements of the investment decision. an application of fuzzy set theory 99 model. above all, he concentrates on the essential elements of the problem: the types of assets, the most salient properties of these assets, and the characteristics of investors as fig ure 1 shows. smith uses four asset categories, (a) savings accounts, (b) corporate bonds, (c) com mon stock, and (d) real estate, with four asset properties, (a) liquidity (l), (b) income (i), (c) appreciation (a), and (d) safety (s). he also classifies investors by age into five groups. these variables are divided into two matrices: one relating different assets to their proper ties, and another relating asset properties to different age groups. the coefficients in each matrix are derived from experience and empirical observation, as well as from normative prescriptions regarding investors’ holdings at different stages in their life cycle. using matrix operations, smith obtains a weighted suitability index, which can be used to allocate wealth. reading the product matrix by columns, suitable allocations for each age group can be seen in table 1. for example, the second column in the product matrix shows recommended allocations for the 35-year-old: 19%, 16.5%, 32%, 32.5% in a savings account, corporate bonds, common stocks, and real estate, respectively (these numbers appear boldfaced). reading the matrix by rows, it is observed how holdings of each asset may evolve over time in each investor’s portfolio. the distribution of assets over time is consistent with investment advice based on the life-cycle hypothesis. smith indicated his model had the following limitations: (a) it needed clarification concerning how to obtain probability assignments for matrices; (b) it had no sense of opti mality; (c) it did not include issues bearing on the management of security holdings over time, for example, transaction costs; (d) it held suitability measures for asset types constant for all age levels. smith’s departure from conventional modeling must be appreciated. it represents a “back to basics” approach and, to a certain extent, rescues qualitative analysis from obliv ion. the following sections will show the fertility of smith’s departure when the adequate tools and framework are employed. table 1 smith’s model matrix 1. assets-properties matrix 2. properties-ages l i a s 25 35 45 55 65 sav. 0.6 0.3 0.0 0.4 l 0.3 0.15 0. i 0.1 0.0 cb 0.2 0.5 0.1 0.3 i 0.1 0.00 0. i 0.2 0.6 cs 0.2 0.1 0.4 0.2 a 0.0 0.60 0.7 0.5 0.0 re 0.0 0.1 0.5 0.1 s 0.6 0.25 0.1 0.2 0.4 product matrix: assets\aee: 25 35 45 55 65 savings 0.45 0.190 0.13 0.20 0.34 cbonds 0.29 0.165 0.17 0.23 0.42 cstocks 0.19 0.320 0.33 0.28 0.14 restate 0.07 0.325 0.37 0.29 0.10 totals 1.00 1.00 1.00 1 .oo 1 .oo loo financial, services review 6(z) 1997 iii. relational equations-possibilistic modeling “relational equations-possibi~stic modeling” (rep) is the modeling language of fuzzy sets and uses algebra instead of calculus. when problems and variables cannot precisely be reduced to numbers, algebra tools and the concept of a partial ordering still choices to be made, for example, by comparing and ranking the alternatives. in this context, rep repre sents a problem-determined bridge between exactness and inexactness, which also inte grates qualitative and q~titative info~ation. rep is an example of logical analysis. it begins by defining a problem and then iden tifies relevant elements (categories), relationships, and variables (langer, 1967). in the algebra of sets, the problems and the elements of analysis are known as the universe of dis course, and the categories are referred to as classes, the group of objects that share some thing in common, and sets. sets are composed of elements, which roughly correspond to the variables of calculus. a. fuzzy sets concepts the area of fuzzy sets can be regarded both as a formal theory and as the modeling lan guage aforementioned (zimmermann, 199 1, p. 6). the formal theory is based on algebra of fuzzy sets, which is, in turn, built upon the concept of a fuzzy set. a fuzzy set is a set in the universe of discourse defined by a graded membership function. for example, the sets “young” (y) and “old” (0) in the universe “active investors” (i) are fuzzy sets. these sets can be described as y = {x, uy (x)} = {(10,0.6), (20,o.q (30, l), (40,0.8), (50,0.6)}, and 0 = {x, uo (x)} = {(40,0.6), (50,0.8), (60, 1), (70,0.8), (80,0.6)}, and y, 0 c i. each set is described by an ordered pair: the first element indicates the age, the sec ond element the grade of membership to the set, uy(x) is the membership function. for example, the pair (40, 0.8) in the subset “young” can be read as, a 40-year-old is only a 80% member of this group. the same 40-year-old would be a 60% member of the set “old.” the best representatives of the (fuzzy) adjectives “young” and “old” are 30 and 60 years, respectively. note that there is some grading, as well as an overlap, between a fuzzy set and its complement. this is different from the sets employed in boolean algebra, in which an investor would belong to either the “young” or the “not young” set, but never to both (law of the excluded middle). the selection of elements above a given level of mem bership (say u(x) > 0.6) is called an alpha-cut. memberships can be normalized to 1 and can be constructed to be symmetric, but these are nonessential properties. the values for the membership function used in the set young (y), above, have a possibility meaning (zadeh, 1978). these values have a meaning of compatibility, which can be of a linguistic, or economic nature (a savings account is 60% liquid, common stock is 20% safe, at age 65 income and safety replace appreciation as the most important attribute of securities, etc.), as in smith’s model. y, 0, and i are examples of fuzzy sets, with fuzzy mem~rships and a fuzzy universe definition. the same types of fuzziness and sets appear when assets are grouped into classes (fixed income versus variable income) or into subclasses (growth versus income stocks) and in the labeling of economic situations (recessions, transitions, and periods of prosperity). note that no amount of additional data can eliminate the uncertainty regarding these sets. these constructs are used in financial counseling because they are useful. for exam ple, a brokerage service may want to learn how many of their clients are really “active an application of fuzzy set theory 101 investors.” the adverb “really” is another fuzzy qualifier that is useful to communicate. fuzzy set theory has been developed to provide support to everyday reasoning. rather than purifying concepts so that they conform to the boolean framework, fuzzy set theory con forms itself to individuals. major operations are defined in fuzzy sets such as the cartesian product, the algebraic product, etc. and in set-theoretic operations such as union and intersection, the max-opera tor (logical “or”) and the min-operator (logical “and’). for example, if there is the percep tion that something is missing between “young” and “old,” the intersection of these two fuzzy sets would provide us with a third set that could be called “mature” and which fills the logical void, that is, “mature” = y n 0 = m = { (40,0.6), (50,0.6)}. b. fuzzy relationships and their composition relationships are generalizations of functions. functions assign a unique object (value of the function) for each object in their domain (argument). functions pair values of variables, while relationships pair variables or sets themselves. much of the scientific effort is aimed at uncovering and discovering such relationships because they lend a degree of perma nency to our knowledge. a fuzzy relationship is a relationship defined over fuzzy sets. with discrete supports, that is, numbers, matrices express relationships. strength is the most important character istic of a relationship, and it is represented by the elements in the relational matrix. in order to study the composition of fuzzy relationships, let us first establish two fuzzy relationships, for example rl = “security xi is a close substitute for xj” and r2 = “perfor mance in economic cycles.” let r2 have the following membership function defined over the elements of the class economy = {prosperity-zl, transition-z2, recession-z3}, which is represented in table 2. the composition offuzzy relationships amounts to compiling or merging the informa tion from each of the relationships being composed. the max-min composition is fre quently used to compose fuzzy relationships and is also a fuzzy set defined by a membership function: rl l r2 = {l(x,z), max {min (pr1 (x,z), ur2 (x,z>}}] ix e x,z e z}. table 2 first example of composition ri xl x2 x3 xi 1 .o 0.8 0.4 x2 0.8 1 .o 0.2 x3 0.4 0.2 i .o r2 zi 22 23 xl 1.0 0.6 0.4 x2 0.6 0.8 0.4 x3 0.2 0.6 1.0 1.0 0.8 0.4 r3 = (ri l r2) = 0.8 0.8 0.4 0.4 0.6 1.0 102 financial services review 6(2) 1997 the composition of the two relationships into r3, which is shown in table 2, could be interpreted as securities that are close substitutes and perform well in economic cycles. the membership functions for the composition are calculated in a manner that is sim ilar to matrix multiplication. merge the elements in each row in the first matrix with those in the first column of the second matrix, which results in three ordered pairs, {( 1, i), (0.8, 0.6), (0.4,0.2)}, and select the minimum values in each of these pairs. this results in a set of three elements: { 1,0.6,0.2}. select the maximum element from these values, 1, which will be the first element in the first column and row in the r3 = {ri l r2) set, and can be written as t = r3(x, z) = {rl(x, y) l r2(y, z)}. dubois and prade (1980, p. 74) provide a very clear interpretation of the max-min rule. the membership functions in each relational matrix can be regarded as links in the chain of reasoning. this chain is only as strong as its weakest element (min evaluation), while the strength of the relationship between each of the sets (say, x and y) is that of their strongest chain (max evaluation). the max-min composition in this study is based on partial order ings called brow&an lattices, which are a further refinement of boolean lattices, ruther ford (1966) and abbot (1969). lattice algebra is a special type of algebra which emphasizes ordering. relationships can be established among several symbolic objects, x, y, and z, which provides a wider scope of analysis and brings us to the more familiar territory of simulta neous equations systems. fuzzy relational equations express causality implied by the rela tionship among x, y, and z. assume, for example, that yi’s represent the fuzzy class of “age” = { young-yl, mature-y2, senior-y3). the left-hand-side matrix, fuzzy relation rl, could express “adequacy of security xi for group age yi.” and the composition of these two relationships, r3 = ri l r2, could be interpreted as “adequacy of securities xi’s for age groups yi’s through economic cycles zi’s,” which is shown in the upper section (example 2) of table 3. the information in the composition matrix can be read as follows: “security one is the most adequate if we expect prosperity. securities one and two are comparable, but better than security three, in transition periods. security three is the most appropriate if a reces sion is expected.” these propositions would be adequate for any of the three age groups. table 3 further examples of compositions example 2. yl y2 y3 %i 22 23 xi 1.0 0.8 0.4 yi i.0 0.6 0.4 x2 0.8 1 .o 0.2 y3 0.6 0.x 0.4 x3 0.4 0.2 1 ,o y3 0.2 0.6 1 .o example 3. x1 x2 x3 %i z2 23 yl 1.0 0.8 0.4 xi 1 .o 0.6 0.4 y2 0.8 1.0 0.2 x2 0.6 0.8 0.4 y3 0.4 0.2 i .o x3 0.2 0.6 i .o an application of fuzzy set theory 103 assume now that the left-hand-side matrix is transposed. the left-hand-side matrix (rl) could now take the meaning of “holding of securities xi’s by ages yi’s,” the right hand-side matrix (r2), could mean “performance of securities xi’s in economic cycles zi’s,” and the composition r3(y, z) = {rl(y, x) l r2(x, z)} could be interpreted as “per formance by age groups across economic cycles,” as in example 3 in table 3. a hierarchy of classes can be established and they can be composed successively into two main overarching classes, which resemble the use of intermediate variables. for exam ple, suppose that learning about classes (m, q) requires that a knowledge tree be built, in which information is processed at different levels of analysis or layers. one may start with intermediate relations rl’s in level one, compose then into r2’s at level two, and finally obtain the integrated, all-encompassing r3 target relationship. table 4 shows a diagram of this process. the following three considerations are critical in modeling by relational equations: (a) to determine a relatively sufficient universe of discourse, (b) to determine what is known, and what is to be known, and (c) to establish proper causality. the first element is similar to having conformable matrices. each column in the rela tional matrix rl(x,y) represents a variable (set). a corresponding variable (yi), or link, in r2(y,z) is needed if it is to be composed with relation r2(y,z). with respect to the second element, examples 2 and 3 show that relational composi tions have different effects with respect to which variable becomes implicit in the informa tional structure formed by x, y, and z. in example two, the y set, age groups becomes implicit, and the result is a relationship in terms of x and z, as in the case of simultaneous equations systems. in example 3, it is x which becomes implicit, as the following compact notation shows: r3(x,z) = {rl(x,y) l r2(y,z)} and r3(y,z) = {ri(y,x) l r2(x,z)} for examples 1 and 2, respectively, shown in table 4. the third element is significant because much of the validity of knowledge rests on the notion of causality. two variables may have a very large correlation coefficient but that does not mean they are related in any way. a regression of the set of variables x = {xl = xl, . . . . xt, x2, x3} on the set y = {y = yl, .., yt} may have a large r-squared value and yet the “x” variables-set may have nothing to do with the variable “y”. what is critical in scientific work (econometric, biomedical, etc.) is the reasoning that justifies using set x as a basis for y. recall that a relationship is a set of ordered pairs, by definition, and that this “ordering” is meant to carry causality content. in other words, fuzzy relationships may or may not imply causality, but causality is required in fuzzy relational equations, as in any other type of equation. relational equations modeling is not more difficult, and is equally (or more) rigorous from a formal viewpoint, than conventional modeling. the relational equations-possibilis table 4 relationships and multilayered compositions level i: level 2: level 3: ri i(m,n) r i2(n,o) r13(0, p) r14(p, q) r21(m, 0) r22(0, q) r31(m, q) 104 financial services review 6(2) 1997 tic approach is simply different from the calculus-plus-probability “gold exchange,” and it stresses different aspects of the decision-making process. readers who feel reluctant to leave the “calculus + probability” safe harbor of numerical exactness may ponder shackle’s observation: “(p)robability is amodel of thought, not achar acter of the natural world” (1972, p. 385). as usual, logical reasoning must weave variables together if the resulting information tapestry is to make any sense and have any meaning. iii. application to the individual investor problem reformulating smith’s model into the relational equations-possibilistic setting is easy. it amounts to composing information on two relationships: (a) between securities and their associated properties, and (b) between properties of the securities and age groups. the latter one is the main “class-ification” to identify investors in smith’s model. note that security properties are the commonality, or link, that permits the composition of the two relationships. smith’s numbers are used in the relational compositions to facilitate comparisons. there are several critical differences: 1. the modeling framework now is different. the rep framework is of a wider scope than the purely numerical one, given the use of fuzzy sets (classes and variables) and membership functions endowed with linguistic and possibilistic content and truth values. 2. the receptacles of information (relational matrices and their coefficients) are also different. the matrices now are relational matrices, which provide informa tion about approximate relationships among variables. the coefficients of these matrices inform us about approximate adequacy, possibility, or compatibility between the sets in each of the classes. recall that the analysis must be endowed with flexibility to enable it to handle uncertainty, imprecision, and complexity. this flexibility is appropriate and necessary in each of the areas of the problem. for example, real estate may not appear as such a good investment when evaluated strictly in terms of returns, but all individual investors need a place to live. 3. causation and inference rules are different than in smith’s model. rep modeling requires only conjectures or approximate causation and approximate inference, while smith’s methodology presumes exact causation and exact inference. table 5 presents the composition of the two previous fuzzy relationships and a com parison of results with smith’s, respectively. several points regarding the results need comment. first, the results differ from smith’s, even though the ranking of securities in each portfolio is similar. the results obtained by matrix multiplication correspond to the arith metic product rule, while the (fuzzy) numbers in the first composition have been obtained by the max-min composition rule. there are some cases in which they may coincide; in general they will not. in some cases, the max-min rule may resemble a rounding or averag ing of the arithmetic rules, but the differences are more profound than that. an application of fuzzy set theory 105 table 5 individual investor’s problem sav. cb cs re r3 = ri ’ r2 = r1 (assets, a-properties) l r2(a-properties, age) = iu(assets, age) l i a s 25 35 45 55 0.6 0.3 0.0 0.4 l 0.3 0.15 0.1 0. i 0.2 0.5 0.1 0.3 i 0. i 0.0 0.1 0.2 0.2 0.1 0.4 0.2 a 0.0 0.6 0.7 0.5 0.0 0.1 0.5 0. i s 0.6 0.25 0. i 0.2 65 0.0 0.6 0.0 0.4 max-min rule: assets\age: 25 35 45 55 65 savings 0.40 0.25 0.10 0.20 0.40 cbonds 0.30 0.25 0.10 0.20 0.50 cstocks 0.20 0.40 0.40 0.40 0.20 restate 0.10 0.50 0.50 0.50 0.10 totals 1 .oo 1.40 1.10 i .30 1.20 max-min rule (normalized): assets\age: 25 35 45 55 65 savings 0.40 0.18 0.09 0.15 0.33 cbonds 0.30 0.18 0.09 0.15 0.42 cstocks 0.20 0.29 0.36 0.3 1 0.17 restate 0.10 0.36 0.45 0.38 0.08 totals i .oo 1.00 1 .oo 1 .oo i .oo arithmetic product rule, smith (1974): assets\age: 25 35 45 55 65 savings 0.45 0.19 0.13 0.20 0.34 cbonds 0.29 0.165 0.17 0.23 0.42 cstocks 0.19 0.32 0.33 0.28 0.14 restate 0.07 0.325 0.37 0.29 0.10 totals 1 .oo 1.00 1 .oo 1 .oo 1.00 the critical fact is that the arithmetic product rule is only valid with non-fuzzy, or pre cise (boolean) “crisp” sets. when the classes, sets, indicators and inferences are not pre cise, the logical composition rules for fuzzy sets must be used. the individual investor’s problem simply does not have the precision required by the arithmetic product rule. the rankings are similar because the “ordering” induced by the supporting lattice (a partially ordered set) is approximately the same. that is, the most heavily weighted securities in the arithmetic product rule (smith’s case) are also the most heavily weighted securities in the relational composition. second, the optimal weights in the relational composition do not add up to one. this is a by-product of the approximate composition rule and the use of fuzzy sets, which results in fuzzy portfolio weights. it is both easy and instructive to normalize fuzzy weights divide each weight by its column cumulant-so that they can be compared to those obtained from the arithmetic product rule. 106 financial services review 6(2) 1997 the arithmetic product weights look impressively informative. for individuals of ages 3.5, 45, and 55, the max-min composition does not seem to be able to determine security holdings as precisely as the product rule. however, there is a problem with the “precise” weights of the arithmetic rule: there is not enough knowledge to guarantee their accuracy. “fuzzy weights” do not give such a false security. note that there is no limit to how many fuzzy relationships can be composed as long as they are logically integrated. that is, the lists of assets, investors’ characteristics, etc. can be expanded, but the models introduced suffice to present the new methodology. more over, each user of this methodology is expected to develop custom-made or situation-spe cific models, to which the rep methodology lends itself well. the rep model addresses some of the limitations of smith’s model. it provides alter native ways to obtain coefficients and assignments. a sense of optimizing is implicit in the model via possibility theory, since the model provides the best course of action given the information available. suitability measures for asset types can be changed for different ages, and additional fuzzy relationships can be included. in sum, the rep model achieves what smith set out to do in a more satisfactory in a logically proper manner. rep modeling is inductive in that the integration of the premises permits insights to be gained that are not evident in the premises themselves. it opens a door to approximate rea soning, which is “a type of reasoning which is neither very exact, nor very inexact, . . and provides a way of dealing with problems which are too complex for precise solution” (cf. zadeh, 1975d, p. 2). iv. concluding comments this study applied to the case of the individual investor’s problem a new methodology based on fuzzy sets. in this problem, the classes, variables, indicators, causality, and infer ences are at best approximate, which calls for solutions and reasoning which will also be approximate at best. it also requires a type of flexibility and imprecision that can not be obtained in the conventional (calculus plus probability) framework. the relational equations-possibilistic modeling is only part of the ongoing research on “approximate reasoning,” which shows considerable potential for providing solutions to investors. research in approximate reasoning points toward the possibility of building an evolving model or automaton soon (klir & yuan, 1995, p. 349) which could manage investments. before this “automatic pilot” arrives, relationships between investment pro fessionals and their clients are likely to continue to be very important. the methodology presented can contribute to enhancing that relationship, since it properly incorporates the formality requirements of investment professionals with clients’ needs and reasoning in a flexible framework that one can both relate to and understand. references abbot, j. (1969). sets, lattices, and boolean algebras. allyn & bacon. brennan, m. (1975). the optimal number of securities in a risky asset portfolio when there are fixed costs of transacting: theory and some empirical results. journal of financial and quantita tive analysis, 10(3), 483-496. an application of fuzzy set theory 107 dubois, d., & prade, h. (1980). fuzzy sets and systems. new york: academic press. hirshleifer, j. (1965). investment decision under uncertainty. choice theoretic approaches, 49(4). jacob, n. (1974). a limited diversification portfolio selection model for the ismall investor. the journal of finance, 27(2), 847-856. kickert, w. (1978). fuzzy theories on decision-making. leiden, netherlands: martinus nijhoff. klir, g., & yuan, b. (1995). fuzzy sets andfuzzy logic: theory and applications. englewood cliffs, nj: prentice hall. kraus, a., & liztenberger, r. (1975). market equilibrium in a multiperiod state preference model with logarithmic utility. the journal of finance, 30(5), 1213-1227. langer, s. (1967). an introduction to symbolic logic. new york: dover. markowitz, h. (1991). foundations of portfolio theory. the journal of finance, xlvi(2), 469-477. markowitz, h. (1959). portfolio selection [monograph]. new york: cowless foundation. myers, s. (1968). a time state preference model of security valuation. journal of financial and quantitative analysis, 3(l), l-33. pogue, g. (1970). an extension of the markowitz portfolio selection model to include variable transaction costs, short sales, leverage policies, and taxes. the journal of finance, 35(5), 1005-1027. rutherford, d. (1966). introduction to lattice theory [university mathematical monographs]. lon don: oliver & boyd. shackle, g. (1972). epistemics and economics. cambridge, ma: cambridge university press. smith, k. (1974). the major asset mix problem of the individual investor. journal of contemporary business, winter. zadeh, l. (1979). a theory of approximate reasoning. in j. hayes, d. michie, & i. mikulich (eds.), machine intelligence (vol. 9, pp. 149-194). new york: halstead press. zadeh, l. (1978). fuzzy sets as a basis for a theory of possibility. fuzzy sets and systems, 1, 3-28. zadeh, l. (1975a, b, and c). the concept of a linguistic variable and its application to approximate reasoning: parts i, ii, and iii. information sciences, 8, 199-25 1, 301-357; 9,43-80. zadeh, l. (1975d). calculus of fuzzy restrictions. in l. zadeh, k. fu, k. tanaka, & m. shimura (eds.), fuzzy sets and their applications to cognitive and decision processes (pp. l-39). new york: academic press. zadeh, l. (1973). outline of a new approach to the analysis of complex systems and decision pro cesses. ieee transactions on systems, man, and cybernetics, 3, 28-44. zadeh, l. (1965). fuzzy sets. information and control, 8(3), 338-353. zimmermann, h. (1991). fuzzy set theory and its applications, 2nd ed. new york: kluwer aca demic publishers. pii: s1057-0810(97)90008-1 financial services review the journal of individual financial management index volume 6,1997 authors albert, robert l., jr., review of “charter media’s briefing.com web site,” 6( 1): 72 73 albert, robert l., jr., “short selling and trad ing abuses on nasdaq,” 6(l): 27-39 ang, james s., “personal bankruptcy costs: their relevance and some estimates,” 6(2): 77-96 austin, jeffrey r. see cram, terry l. ayadi, 0. felix, “adverse selection, search costs and sticky credit card rates,” 6(l): 53-67 bertin, william see prather, laurie cagle, julie a. b., “conversions of mutual savings institutions: do initial returns from these ipos provide investors with windfall profits?’ 6(2): 141-150 crain, terry l., “an analysis of the tradeoff between tax deferred earnings in iras and preferential capital gains,” 6(4): 227 242 domian, dale l., “performance and persistence in money market fund returns,” 6(3): 169 183 domian, dale l., review of “wealth manage ment,” 6(3): 223-224 dresnack, william h. see singh, sandeep eyssell, thomas h, “financial planning and college saving recommendations: let’s set things straight,” 6( 1): 41-52 fatemi, ali m. see ang, james s. gable, john, review of “the merrill lynch web site (www.ml.com),” 6(l): 71-72 greninger, sue, review of “personal financial planning, 7th edition,” 6( 1): 69-7 1 grinder, brian, “an overview of financial ser vices resources on the internet,” 6(2): 125 140 horan, stephen, “an analysis of nondeduct ible ira contributions and roth ira con versions,” 6(4): 243-256 kahl, douglas r., “the challenges and oppor tunities of student-managed investment funds at metropolitan universities,” 6(3): 197-200 kahl, douglas r., review of “the millionaire next door,” 6(3): 221 kahl, douglas r., review of “the money book of personal finance,” 6(l): 73-74 kahl, douglas r., review of “the mutual fund investor’s center,” 6(2): 153-154 kennedy, william f. see lamb, reinhold p. kim, sharon see milevsky, moshe arye lamb, reinhold p., “the congressional calen dar and stock price performance,” 6( 1): 19 25 lee, ryan b., review of “risk management and insurance,” 6(4): 295-296 liano, kartono see manakyan, herman ma, k. c. see lamb, reinhold p. manakyan, herman, “performance of mutual funds before and after closing to new investors,” 6(4): 257-269 marchand, james, review of “the financial services and financial institutions,” 6(2): 151-152 mcleod, robert see horan, stephen 299 milevsky, moshe arye, ‘the optimal choice query, j. tim, review of ‘the financial ser of index-linked gics: some canadian evivices revolution,” 6(3): 221-223 dence,” 6(4): 27 l-284 query, j. tim see lee, ryan b. nelson, susan logan, “the use of professional designations in the real estate industry,” 6(2): 109-124 nelson, theron r. see nelson, susan logan reichenstein, william see domian, dale l. reilly, frank k, ‘the impact of inflation on roe, growth and stock prices,” 6(l): 1-17 robison, h. david see albert, robert l. jr., pace, r. daniel see lamb, reinhold p. peterson, jeffrey h. see horan, stephen prather, laurie, “a simple and effective trad ing rule for individual investors,” 6(4): 285-294 pritchett, s. travis, “incorporating historical investment performance in projecting life insurance cash values,” 6(3): 155-167 porter, gary e. see cagle, julie a. b. singh, !&deep, “market knowledge in man aged municipal bond portfolios,” 6(3): 185 196 smaby, timothy r. see albert, robert l. jr., smith, keith v., “asset allocation and invest ment horizon,” 6(3): 201-219 tarrazo, manuel, ‘an application of fuzzy set theory to the individual investor problem,” 6(2): 97-107 query, j. tim, review of “american risk and insurance association (aria) web site,” 6(2): 152-153 300 financial services review 6(4) 1997 titles “adverse selection, search costs and sticky credit card rates,” 0. felix ayadi, 6(l): 53-67 “an analysis of nondeductible ira contribu tions and roth ira conversions,” stephen horan, jeffrey h. peterson, and robert mcleod, 6(4): 243-256 “an analysis of the tradeoff between tax deferred earnings in iras and preferential capital gains,” terry l. cram and jeffrey r. austin, 6(4): 227-242 “an application of fuzzy set theory to the individual investor problem,” manuel tar razo, 6(2): 97107 “asset allocation and investment horizon,” keith v. smith, 6(3): 201-219 ‘the challenges and opportunities of student managed investment funds at metropolitan universities,” douglas r. kahl, 6(3): 197 200 “the congressional calendar and stock price performance,” reinhold p. lamb, k. c. ma, r. daniel pace, and william f. kennedy, 6( 1): 19-25 “conversions of mutual savings institutions: do initial returns from these ipgs provide investors with windfall profits?” julie a. b. cagle and gary e. porter, 6(2): 141150 “financial planning and college saving rec ommendations: let’s set things straight,” thomas h. eyssell, 6( 1): 41-52 ‘the impact of inflation on roe, growth and stock prices,” frank k. reilly, 6( 1): 1 17 “incorporating historical investment perfor mance in projecting life insurance cash values,” s. travis pritchett, 6(3): 155-167 “market knowledge in managed municipal bond portfolios,” sandeep singh and will iam h. dresnack, 6(3): 185-196 financial services review 6(4) 1997 301 “the optimal choice of index-linked gics: *some canadian evidence,” moshe arye milevsky and sharon kim, 6(4): 27 l-284 “an overview of financial services resources on the internet,” brian grinder, 6(2): 125 140 “performance and persistence in money market fund returns,” dale l. domian and will iam reichenstein, 6(3): 169183 “performance of mutual funds before and after closing to new investors,” herman manakyan and kartono liano, 6(4): 257 269 “personal bankruptcy costs: their relevance and some estimates,” james s. ang and ali m. fatemi, 6(2): 77-96 review of “american risk and insurance asso ciation (aria) web site,” j. tim query, 6(2): 151-154 review of “charter media’s brieting.com web site,” robert l. albert jr., 6(l): 72-73 review of ‘the financial services and finan cial institutions,” james marchand, 6(2): 151-154 review of ‘the financial services revolu tion,” j. tim query, 6(3): 221-223 review of “the merrill lynch web site (wwwmlcom),” john gable, 6(l): 71-72 review of ‘the millionaire next door,” dou glas r., kahl, 6(3): 221 review of ‘the money book of personal finance,” douglas r. kahl, 6( 1): 73-74 review of ‘the mutual fund investor’s cen ter,” douglas r. kahl, 6(2): 151-154 review of “personal financial planning, 7th edition,” sue greninger, 6( 1): 69-7 1 review of “risk management and insurance,” lee, ryan b. and j. tim query, 6(4): 295 296 review of “wealth management,” dale l. domian, 6(3): 223-224 “short selling and trading abuses on nasdaq,” robert l. albert jr., timothy r. smaby, and h. david robison, 6( 1): 27-39 “a simple and effective trading rule for indi vidual investors,” laurie prather and will iam bertin, 6(4): 285-294 “the use of professional designations in the real estate industry” susan logan nelson and theron r. nelson, 6(2): 109-124 pii: s1057-0810(01)00082-8 ��������� � ��� ��� �� ����� ����� �� ��� �� � �� ��� � �� ���� ����� ����� ������� ������� ������ ���������� ����� �� ���� � � � ��������� ����� � ������ ����������� ���� ���� ������ !����"������� ! #$#%#� �� �������� � ����� !!"# �������� �� ���� �� ���� "$ %��� !!"# �������� " &�����'�� !!" �������� ��� �� ����� � � ������� ������ ���� ����� ����� � ���� ����� � �� '��� ��� ��� ��( )�*����� ��� �������� �� ��� �'����� �� ����� �� ���������� ��� ��� �� � ��+��( ��� ����� ������� ���� �������� �������� �� ��� ����� �'����� �� ��� � �� ���� ����� '� ���� �� ���� ��� ����������� '�� ��� ��� �������������� ������� ���� �� ����������� ������������ ����� ���� ��� ",-,.",,, ������ ��� ������ �'������ ( /� � � ������ �����������0���������� ��������� �� ���� ���� ��� ��� �� ��� ���������( ��� �������� �'������ �� ��� �� ����� ��� � � �������� ��������� �� ��� ���������� ������ � �� �� ��� ����� � � ��������� ���� � ��� ���� ������( � !!" �� ����� &������ 1��( ��� �� �� �� �����( �&' ����� � ��� �( �"2 )��* �"�( ��� �� ����� # �� 0��3� ��� ����������� ��� ��� # 4������������ ��� ���������� �����������0 ���������� ����� � ���� ������ 1��� ��� ��� ��*�� ������� ��� ���� ����� ����� �� ���� ����� � �� ���������� ��� ��� ��( 5�� ������� ����� �� �������� �� � � �� ��� ������������ �� ��� �� ����� � ������� ������ ���� ����� ������6�� ���� �� �� ������� ����������� '� ������ ���� �� � �� ��� � ���� ��� �+������ �� ���������� ����� � �� ��� � �� ������� �'���0����� � �+������ ������ ( ��� �� ����� ������ 6����������� ����� �� � � �� ��� �������� ������������ �� � �� ��� � � ��� �� ���� �6���� ��������� ������6�� �� �������� ��������� ( 5���� ��� �� ����� ����� ������������ ����� '� �� �� ������ ��������� 7�������� &������ �����* "! 8 !!"9 "-�.":; "!$-0!:"!<!"<= . �� ����� ������ � !!" �� ����� &������ 1��( ��� �� �� �� �����( >11� & " ! $ 0 ! : " ! 8 ! " 9 ! ! ! : 0 : � ?���� ������ ������( ���(� �"0;"!0;;!0";-!# ��+� �"0;"!0""!0",:;( &+���� �""����( ������@ 3�(��� 8�( �����9( ������� �� �������� ��6����� ������ � �� ������� ��������� ��� ����������� ���� � �� ��� � � �� �������� �� �+������ ������ ( ��� ��� �� ����� ����� �� ��� '� �� �� ��� ������� � ���� ��� ��� �� ������ ��� ��� ������ '�� ���� �� ����� ������ ��������'�� 8a��� � b� ��� �� � c b�a����� ",, # b��� ",,-# &�������� ����������� c /����� ",,,9( d� ���� ��� ��������'����� �� ��� ������ � ��� �������� �� ��� �'����� �� ��� �� ����� �� ���������� ��� ��� �� 8�� ����� '������� 9 � ��+��( 1� ��� ����� *� ��� �� ��� ��'��� '� �+��� ����� �� ��� ��� ����������� ��� ����� �'����� �� � ����� ����� ��� �� ����� ��*� � ��� � �� ���� ������ � ���� ��� �� �� ����� ����������� � 1��( /� � � ��� ������ ����������� �6���� ����� ���� ��� *�� �� � �� ���������� �� ��������� ��� ����� ������ ��� ��� ������ �� b������ 1��� ����� >������ � 1��(� � ������� �������� ���� ����� ���� ��� ������ ���� ������� ��� �� ����� 7��� ��� ��� b������ ����� >�� ���( ���� ��� ������� ���� 8",,�9� b�� �� ��� /����� 8",,$9� >������ ��� a����� 8",,:9� ��� ?������� ��� ?������� 8",,,9 ��� ����� �� ��� �� ��� � ���� ��� ��� �� ����� ����� � ���� ����� ���� '��� ����������� '� ���� ��� ( 1� ������ �� *� ���� �� ��� � ��� �� ����� ����� � ���� �� �������� '��� ����������� '� ���� ��� �� ��� ����0*����( � ���� ��� ����� ������� ���� ������ �������� �� ��� �� ����� �'������ �� ��� ���( ��� ��������� �� ��� ����� � �� ������ � �����* ( 1� &������ � *� ������� � �����* �� ��� �������� ������� ����������( &������ � �� ���'� ��� ��� �� ����� ����� � �� ��� ��� ��� ���� � �� �� ��� ����� � �� �� ����������� ��� ��� �� ����� �'�����( &������ 2 ������� ������ �����������0���������� ��������� �� ��� �e�� �� ���� ���� ���� � 8"9 ������0 �����0���������� � �� ����� ������ # 8 9 �����������0���������� � �� �� 0��3� ��� ������ # 8�9 �����������0���������� � �� � ������������� �� � �� ��� �� ����� �'�����# ��� 829 �����������0���������� � �� � ���������� �� � �� ��� �� ����� �'�����( /� �� �� ��� �� ��� �� ��� ������( � ������ ��� ������ ��� � ��� ����� �� &������ $( � ���������� ������ ����� ���� '��� ���� ����� �� ��� �'����� �� ����� ����� ���� ����� ���� �� �� ���� ��� ��� �� ���� ��� ����� ",-! ( ������ ����� �� ��� ��� �� ����� ��� �� ������ ���� '� ?���� b��� ������� ��� ?��� 8",, 9� ����'���� ��� ������ 8",,29� ��� ��� 8",,$9� d������ ����'����� ������� ��� /����� 8",,-9� a�� ��� 7�� ��� ���� � ��� &����� 8",,,9� ��� ���������� 1� �� ���� ��� 1� ���� 8 !!!9 e�� ������ �������� �� ��� �� ����� �'�����( � ������ ������ ��� � ������� �� ?��'� ��� ���� 8",,!9� ���� f������� ��� �� ��� 8",,"9� d��� ��� /�� �� 8",,29� ��� f��� ?��� � ��� ?��� 8",,:9 *�� ������ ��� ����� ��� ���������� ��� �� ���� �� �� g(&(0'� �� ������������� ������ ���� ( 1� ������ �� b�� ��� ������ 8",,!9� /�� �� 8",,"9� 7�� �� ��� &����� 8",,;9� � *��� � a���� ��� %��3� ��� 8",,-9 e�� ��� �������� �� ��� �� ����� ��� ( ��� ��� �� ���� �� ��� �� ������ �� ��� �� ����� �'����� ��� '� �+������� '� ��� ���� ���� ��* ������ ���� ��������� �� ��� �� ����� ( ��� �� ����� ����������� � �� ��� ��������� �� ���� ��� ��� ���� ����� ��* ������ 8f��� c ��� � ",,!9( ������ ��� )����� 8",,;� ",,-9 �������� ��� ����������� �� ��* ������ ����� �� ���� �� '� ������ ����� ��� ��,��� ��� e�� ��� �������� �� ��� �� ����� �'����� ��� ��* ������ ���� ��� �� � ��� ���� ( "-2 .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ���� ����� ��� �� ��� �� ������ �� ��� �� ����� � � ���� ��� ����� � � �'������ '� /� ���� &������ � ��� >��� 8",, 9 ��� ?����� ��� )����� 8",,:9( a��� ����� � � ���� ���� ��� ������ ����� ?����������� � ��* ������ ��'�� ��� ����� � ��� ��� ��������� ��� ����������� �� ����� ����� ��� �� ����� ( ������ � ��� ��� �� ����� ��������� �� ��� � ����� ��� ����� ����� *�� �������� �+����� ����� '�� ��� ��� �������������� ��� ������ � ��� ����� � ��� �������� '� ��� ������ ����� ��* ������ �� ��� ���� ����� ����� ���� �� ��� ������ ���� �� *���� ����� *��� ����������� ��� ��� ������ �� ��� ����� ( &������ �� ?����� ��� )����� 8",,:9� '�� ���� � /� ��� �� ��( 8",, 9� *� � � ���� ����������� ����������� ���������� �� � ���� � ��� �� ���� ������ 8a������ c ?������ ",,29( 1� � ��� �� ����� ����� � � �� *�� � �� *��� �� ��� ����� ��������� ���� � ��� ��� ������ �� ���� 8a������� c h��� ����� ",,"9( 1� ��������� ��� �� � ���� ���� ��*�� �� ������ ��� �� ����� �'����� ����� ���� � ����� �� ������ *� � � *�� �� �� ���� �� ������� ���� �� ��� ��� ������ 8��������� �� ��(� !!!9( ?����� ��� )����� 8",,:9 ��� ��� e� � �� � � ����� 8�� ���� ���6����9 �' �������� ��� �� ����� ����� ���� ��� ���6����� *��� *���� ����� i �������������� ��� �' ����� � ������� �� �������� � ��e���� ��� �� ����� �'�����( � ���� %-#3.& ��" ��� ���:�� �����, ������,� ��� � �� ���� ����� � �� ����������� �� � ����� ���� ����� ���� �� �� �� �� ����� 7����� �� 7��� ���� ",-, ��� �� � ��� �� ��� ����� �������������� �� �� ������������ ����������� ������ ��� ������ ��� ��� ��� �� ����� 8�����9� ����� � ����������� '�� �� <������ ��� e������� ������� *�� ��������� ���� � �� ������ ���� ( ��� �����* � �� 0���� � �������� �� ���� �� ���� ����� ���� ��� ������� ( ���� ����e������ ��� '������ ���� ��� � ���� ����������� ��� ��� ��� �������� �� ������� 8������� �� � �� ��*�0��� �� 9 ��� '� �������� *��� ������� �� '� � ����������� ����������� ������ ���� �� ��� � �� ��� *��� ��� �� � ������'�� ��� ���� ��� ������� ������ ������ ( �������� �� ���� �'���0����� � ������ ��� '� �������� *��� �� �� � �� � ��� ��� �� ����� ����� �( ��� � � ������������ �� ��� �� ����� �� �� ���� � ����� '��� ��*�� ����� ���� ��� �� ��� �� �*� '���� � �� ��� � � ��� ��� �� �( ��� ����� � � *�� ����0����� � �� ���� ;! ���������� ��������� � *���� ��� '� '������ ������ ���� ���� ���� ���� � ������������ ��������� ��������� ��� ��������( 7���������� ��������� ��� �������� ������� �� ��j����� �+��������� ��� ������� ��� ?�� ���� >���� 1���+� ��� ?�a 7����� 1���+� ��� d�* %��� 7����� 1���+� ���������� ��������� ����� � ��� ��� g(&( ������( �������� ��������� ��� �������� ������� �� ��� �������� ������ �� ��� 7������ �� ���� a�� � ������ � ���� �������� � ��� ����� ����� ��� ����� ����� � *��� � '��� ��������� ��������� ��� � ������� ������� � ������� � ��* �� � ��� ��* ��* ( 1�������� �� ���� ��� �������� �� '��� ��� '��� ��� ��� ��� �� ��� �������� ���������� �� ������ ��� ������� ������ ���� ������ � ����� � ����<��� ������ ��� ������� ����� � ��� '� '��� ��������( �������� ��������� ��� ������� �� ������ ��� �� �������� ��� �� ��� ������� ��� ����� ������ ( ��� ������� ��� ����� � *�� �� �� ����������� ��� �������� ��������� ��� �� �� *�� �� �� �������� ��� �������� .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� "-$ ��������� ( ���������� �� ������ ��� ����� � ������������ '� �� ��� 6������������� ������� *������ ��� �'3������ ��3� �����( ��� ����� �� '��� ����� ��������� ����������� ���� ",;, ��� �� � �� ���������� ���� ��� �������� � �� �� ���� � ���� ����� ���� ��� ������� ���� ",-,( ����� ��� �������� �+������ ����� ��� ����� �� ����� '� ��� ��� �����*�� ��� ��������� �� '�� ��� ��� � ��� ��� � �� ���������� ���� ��� ��� �����*�� *�� ��( ������ ��� ���� � �����+������� = '������ �� ������ � �� ( %-53����� �� ����� �� ��" �� ���� � 5��� ��� ",-,.",,, ���� ������� ����� �� ���� 2: ����� ���� ��� � 2 '�� ��� 2 ��� � ����� �� �'��� �*� � ��� � ����( g �� ����� '�� ��� ��� �������������� ��� ��� ����e� ���� ��� � �������������� *��� ����������� ��� ��� ������ �� b������ 1��� ����� >������ � 1��(� *� ��������� ����� *�� �� ����� ������ � 3�� ��� � !0���� ������ ���� %��� ":� ",-, �� 7�'����� " � ",,, 8"�! $ *�� 9( /� � � *�� �� ������ '���� � ��� ����� �������������� ��� ������� *�� ��( ����� ����� ������ ��� ��� ���� � �� ��� �� ��� !!! 8� ����� ������ ����+9 ��� ��� &c> $!! 8��3� ��� ��� �������� 9( ��� ����� ������ e ��� �� ��� ������� ���� ��� �� ���� ����� ��� �� �'��� (�k �������� ��� ������� �� ="!!�!!! �� �� ( 4��� ���� ����� �� � ��� ���� 8���'�'�� �� ���� !(�k9 ���� �� � �� ��� ������� �� ��� ��� b������ ������ � ���� ��� �� ������ ���� ����*�� *������ 8�� ��� � � ��* *����� ��� ����9 �� � ����� ��� �� ������� �� �� �� �( ��� �0����� �0'��� � � �� �� ���+� ��� �� 0���� ����� 3�� ��� ��� �� ��� !!! � *��� � ��� &c> $!! ������ ��� � �� ��� ��� ���� ����� *�� �� ��� �� ������ � 3�( /���� ��� &c> $!! 1���+ � *����� ���� ����� � � '������� ��� ��������� ������������ �� ���+�� ��� � �������������� ��� ( &���� ��� � � ��� b����� ���*�� 7��� 1���+ � � '������� � *� ��� ���� ��� �� ��� !!! 1���+ � ���� ����� �������� '������� ��� ��� ���� �� ���� 8 ��*�� ��� ���� �������������� ��� 9 ���� ��� ����� '� ��� ���� ���( ���������� �� ��� � �� ��� &c> $!! ���� ��� 6������������� ������ �� ��� � � �� ��� �� ��� !!! ��� ��� �������� �� ��� ���� ����( �� ���� �� ��� *������ ��� ����������� ��� ��� �� ����� �'����� �� ��� ��� � � ���� ����� ��� ������ !0���� ������ � ������� ���� ���� ���������� �6��� ���� ������ ( ��� e� � �'������ ��� � ���� %��� ":� ",-, �� �� � � ;� ",:2 ��� ��� ����� ��� �+���� ���� �� � � ;� ",:2 �� ��� � ",:,( ��� ����� �'������ � ���� ��� � ",:, �� 4����'�� �!� ",,� *���� ��� ������ �'������ � ���� 4����'�� �!� ",,� �� 7�'����� " � ",,,( ��� ��� �� �� ��� �'������ � ��� �+����� $ ���� ���� ���� *��� ������� � �� ��� ������ � ��� 8������ 9 ���� ( ��� �� ������� �� ��� �� ��� �� ��� ��� �� �� ��� �'������ *� ���� �� ���� '� �+������ ��� ����������� ��� ����� �'����� �� ��� � �� �+ �6��� �'������ �� � ���� ������ �� ",:"( &���� *� ��� ��� �' ���� ��� � ��e���� ���� � � �� ��� ��� �������� ��� ��� ������ !0���� ���� ������ � *��� � ��� ���� �'������ ����( � ����������������������� ����� ���� �� ������� /� ������ ������ ��������� �� ���� ����� � � ��� �� ��� �� � �� � �������� �������� �����������0���������� �����( /� e� � ������� ��� ����������� �� ��� � �� "-; .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ����� ������ ( )�*����� ��� ����������� ��� ��� ��� '� �� ��� �� ������ ���� �� � ���� �� ( /� ���� �������� ���i ����������� *��� ������ ��� ��� �� �� 0��3� ��� ������ ( 7������� '���� � �� 0��3� ��� ��� ��� �� ��� ������� ��� ���� ����� �� ��� � *� ������� � ������������� ��� � ���������� �� � �� ���i ��� �� ����� �'�����( ;-#!�� ���� �+�������� � ����, � ��� ������� 5�� ������������ �����������0���������� ������ ������ ����� ������ ( ��'�� " ������ ����� � ����� *�� �� �+�� ������ �� ���� 3� � 3� � ��� �� '������� � 3� � 3� � � �� ��� �� ��� !!! 1���+ � 3�� ��� ��� ������ !0���� ���� ������ � *��� � ��� ���� �'������ ( ���� ������������ ��'�� " ��* ��� ����� ����� �� ��� � ����� ������ ����� ��� �����*�� ���� ���� ����� ������� � ��� � ��0���0��� �� ��� ��� ��� �� � �� 8��*�9 ��� ��� � ���0 ��0���0��� �� ��� ��� ��� �� � �� 8��*�9( �� ��0��� �� � ��e��� � ��� *���� ��� ��� �� ������ �6��� �� �+���� ��� �� 0���� ����� 3� � 3� � !� ��� � ��*�0��� �� � ������������� '� � ��� �� ������ �� ���� ��� �� 0���� ����� 3� � 3� � !( ��� �� ��� �� ��'�� " �������� ���� ���i ����� *�� �� ������ �+���� ��� � �� �� '������� �� ��� �� ��� ���� ������ �� ���( 5��� ��� ",-,.",,, ������� ��� *� ��0���0 ��� �� ;"(2k �� ��� ���� ��� �������� � ����� *�� �� �+�� ������ �� !( k� *���� �+���� ��� ��� �� *�� �� �+�� ������ �� !("!k( ���� ����e������ ��� ����� *� ��0���0��� �� �� ��0��� �� ��������� �;(;k �� ��� ���� ������ �� ����� � *�� �� �+�� ������ �� "($;k ��� ��0���0��� �� 2(:k �� ��� ���� �� ��*�0��� �� ��������� ������ �� ����� � *�� �� �+�� ������ �� �"(2 k( /���� ��� ������ �� '� ���� ��� �� �������� � � ��e���� ������� �� ��0��� �� ��������� � �� ����� ����� � � ��� �� ����� � �� � ��j����� �� ��� �� � ������� *��� ���0 ��0���0��� �� 8&������ � ",, # /� ���� ",,-9( 5��� ��� ������ ������ �� ���� ��� ������� � ��� � ����� � *�� �� �+�� ������ �� � ( �k �� ���� ���� ��� � �+�� ������ �� "(2:k( ��� ���������� '��*��� ��� ����� � *�� �� �+�� ������ �� � ( �k *��� ��� � ���0��0 ���0��� �� �� ��*�0��� �� ��������� 8��� �� �������9 ��� ��� ����� � *�� �� �+�� ������ �� "(2:k *��� ��� � ���0��0���0��� �� �� ��0��� �� ��������� 8��� ��� ��� �����������9 � ���� ������� � ��e���� �� ��� !($k ����� �� '����� 8� � ��;"9( ��� �� � �� ��� ������ ���� �� � ���e��'�� �� '� ���0��0���0��� �� �� ��*�0��� �� ��������� ���� �� ���� �� ��� ��0��� �� *�� ��� �� �� ����������( ��� ���� �0 ���� ��� �� ��'�� " ��� � ��e���� �� ��� ������������ $k ����� �� '����� � ���� �� ��� ����� �'������( ��� ���������� �� ����� ;� ",:�� ��� �+�� ��� �� ���� � ��� ���� ����� ������� ��� ��� � �� 5���'�� ",� ",:-( ����� ��� 7������ �� ���� 6��� �� ��*���� ������ � ���� �� �� ��� � �� ��� ��� ��� �� ��� �� ��� �� ���� �� �6���� �+�� ��� �� 4����'�� ,� ",:-( ��� � ���������� ���� ��� ������'���� ������ �� �� ������� ������ � ��� � ��� '������� ( )�*����� ��� �������� ������ '�� ����� ���� ��� �� ",::.",:,� ��������� �� �� ��� ����� ����� '����� ��� ����� ��� ��� �( ��� �������� ����� ����� �� ���i ��� �� ����� ����� � � �� � ���� �� ���� '� ���������� *�� �� �+�� ������ ��� �� ���� ����� ��� �� !(!2k ��� *�� 8 (�k ��������<$ *�� 9( ��� *�� �� �+�� ������ ��� �� ���� ����� ��� ��� ���� ���������� �� ��* ���� ��� ���� ���i �������� *������ ��� ="!! �� ������� ���� ������ �� ��� ������ ��� �� ����� ����� � �� ��� � = �"�2 *���� �+���� ��� �������� *����� �� .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� "-=:"$ �� ��� �� ��� ����� � ��� ��� ������ !0���� ������( ��� �� ��� ��� ��� ����� ���������� �� 7� ( "( ��� ��� �� ����� ����� � �� ��� ��� � � �� ������ �0'��� ��� ��� �� ��� !!! 1���+ �� ��� ����e� ���� ���� ��� '�� ��� ��� �������������� �� ����� ��� �����������( ��� �� ��� ����� � ��� � � �� '���� 0���0������ ��� �� ��� !!! 1���+( 7� ( " �� � ������ ���� ��� *�� �� ���������� ������ ������ �� ��� ���� ��� ������ ��'�� " ��� ����������� �� ��� '� � �� ����� ������ ��� ��� �� ��� !!! � ��� '������� 1��� ����� ������� >������� � �� ���� ����� � �+�� ������ /�� ����0����� � '������� � 3� � 3� �+�� ������ ���� 3� � 3� �0 ���� ���� >���� �� ������ ��� ��� ������ ������ 8%��� ":� ",-, �� 7�'����� " � ",,,9 ��� ��# ��0��� �� �;(;k "($;k !($-k !($-k ��� ��# ��*�0��� �� 2(:k �"(2 k �!(�$k �!(�$k ��� ���# ��0��� �� ",(,k "(2:k !( ,k ! ��� ���# ��*�0��� �� ":(-k � ( �k �!(2"k ! �(;"�� ����� *�� �� �+�� ������ !("!k !( k >���� a� ������ ��� ��� e� � �'������ 8%��� ":� ",-, �� �� � � ;� ",:29 ��� ��# ��0��� �� :(:k "(, k !($$k !($$k ��� ��# ��*�0��� �� ";($k �"(; k �!( -k �!( -k ��� ���# ��0��� �� ;( k "(;2k !(2�k ! ��� ���# ��*�0��� �� :($k � ( -k �!(;2k ! (� � ����� *�� �� �+�� ������ !(!-k !( :k >���� ?� ������ ��� ��� ����� �'������ 8�� � � ;� ",:2 �� ��� � ",:,9 ��� ��# ��0��� �� 2!(!k "(2;k !($:k !($:k ��� ��# ��*�0��� �� �!(2k �"( -k �!(�:k �!(�:k ��� ���# ��0��� �� ";(:k "(�2k !( k ! ��� ���# ��*�0��� �� " (:k � (��k �!(�!k ! "( � ����� *�� �� �+�� ������ !(" k !( !k >���� d� ������ ��� ��� ����� �'������ 8��� � ",:, �� 4����'�� �!� ",,�9 ��� ��# ��0��� �� 2;(;k "($!k !(-!k !(-!k ��� ��# ��*�0��� �� ��($k �"(�:k �!(2;k �!(2;k ��� ���# ��0��� �� :(,k "(";k !("!k ! ��� ���# ��*�0��� �� ""(!k � (�2k �!( ;k ! (-2�� ����� *�� �� �+�� ������ !(!:k !( 2k >���� �� ������ ��� ��� ������ �'������ 84����'�� �!� ",,� �� 7�'����� " � ",,,9 ��� ��# ��0��� �� � (2k "(2 k !(2;k !(2;k ��� ��# ��*�0��� �� !( k �"($"k �!(�"k �!(�"k ��� ���# ��0��� �� ;("k "(2,k !(�,k ! ��� ���# ��*�0��� �� "(�k � (!$k �!(22k ! "(-!� ����� *�� �� �+�� ������ !("!k !("$k � �0 ���� ��� �� ��� ���������� '��*��� ��� ����� � ������ �� ��0 ��� ��*�0��� �� *��� ��� � ���0��0���0 ��� ��( ���� 1������� � ��e����� �� ��� !($ ��� $(!k ����� �� '����� 8���0���� �� �9� �� ���������( "-: .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� !0���� ������ � ";(:!k� *���� � �� ��� ���� ��� ���������� ������ ������ �� ��� '��0���0���� ����� � �� ""( �k( 5������� ��� �� ��� ���� ��� ���� ������ �������� ���� �� ��� ���� ���i ������� � ��e������� ��� ��� ���� ����� ��� 8� *��� � ����� �� � 9 ��� ��e������� ��*� ���i ��� �� ����� ����� � ������� �� ��� ����� ������ ���� � �� ��� 8'��0���0����9 �����������( ;-5!�� ���� �+�������� � ����, ���:+�"<����" ������ �������� /� � � ���� ��� ��� �� � ������ ��� �� 0��3� ��� ������ �� ���� ��� &����� ��� ���� ��� ���� ����� ��� ���� ����� ��� ���� ��� ��� ������ ��� )����� ��� ��� ( ��� �� � ������ ��� ��� �� �� 0��3� ��� ������ � ��� &����� 8�9 �����( ��� &����� 1���+� �� ��� ��� �� ����� �+�� ������ ��� ���� �� ����� �� *���� �� � ��e��� � ��� ������� ���������� � � 3� � 3� ��3� � 3�� 8"9 *���� 3� � 3� ��� �83� � 3�9 ��� ��� ���� ��� ������� ��������� �� ��� �+�� ������� �� ���������( ��� ���������� 8������������ 9 �� '������� �� � �� 0��3� ��� '� � �� �� &����� ����� � �� ��� 8��*��9 ���� ���� �� ��� '������� ( ���� ����� ��� ���� ����� 8��� ",,-9 ��� ������ ��� )����� 8�)� ",,-9 ����� � �� 0��3� ��� ����������� ��� ��� ���� '� ��* ����� �� '� � ����� � ���i ������ ���������� 8������������ 9 ��� ������ �� �� '������� ( ��� �� ��� ��� ��� ��� � �'������ '� ���������� ��� ���� *�� �� ������ �� � ��������� *��� � ��� ���� ���������� 8��<��9 ������ 3� ��� ��� �������� ���������� 8"� ��<��9 ������ 3� ��� ������� �� ��� ���� � �����* � �� � 3� � �� �� � � �3� � 3�� 8 9 7� ( "( �������� *����� ��� ���������� ������ ������ ����� ������ ������ ����� �� ( .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� "-, *���� �� ��� �� ��� ��� ������� ��������� �� ��� ���� *�� �� ������ �� ���� 3�� ��� ��� ��� �� ����+� 3�� �� ���������( ��� �� ��� ��� ����� �� ��� ������ �� ���� 3�� '� '����*�� �� ��� �� 0���� ���� � ���� �� �� ��� ��� ����� �� �� �� ������� ��������� � ��� '������� ������� ��3�� � ��3��( ��� ��3� ��� ������ �� ���� ��� � ���� �������� �������� *��� ��� ������ �� ��� '������� ���� ��� ��� ���� ������( ��� ���������� 8������������ 9 �� '������� �� � �� 0��3� ��� '� � �� ��� ���������� '��*��� �� ��3� ��� ������ ��� ���� �� ��� '������� � ��� 3�� � �� ����� 8�� �����9( ������ ��� )����� e� � ��� ���� �)"� � �'������ '� �'������� ���� ���i ���� *�� �� ������� 3�� ��� ������� ������ �� � ��������� *��� � ��� ���� ���������� 8��<��9 ������ 3� ��� ��� �������� ���������� 8"� ��<��9 ������ 3�� ��� ������� �� ��� ���� � �����* � �)" � 3� � 3� � �� �� � � �3� � 3�� � � 8�9 �)" ��� ��� ��� ���������� '��*��� ��� ���� *�� �� ������ �� ��� ��� ���� �� ��� '������� � 3� � 3�� *���� ��� ���� *�� �� ������ �� ��� '������� � ������� �� �� ��*� � �� ��� �0'��� � ���� �� �� ��� ��� �� ����� �� ������� ��������� � ���i ������( ��� ���������� 8������������ 9 �� ������� '������� �� � �� 0��3� ��� '� � �� �)" � �� ����� 8�� �����9( 4��� ���� ��� �)" ��� ��� ������� '� ��� �� �� ������� ��������� �� ���i *�� �� ������ � �6��� �� ��� ���������� '��*��� ��� &����� ����� �� ��� ��� ���� �� ��� ��� ��( 1� � ������ ll�����ii �� ?����� ��� )����� 8",,:9 *�� ����� ����� ���� �� � �� �+������� ��� ��� �� ��� ��� �� 0��3� ��� ����������� �� ��� �� ����� ( ��� ����� ��� ��� ���������� '� ������ ��� )������ �) � � ��� ���������� '��*��� ��� ���� *�� �� ������ �� ��� ��� ���� �� ��� '������� � 3� � 3�� *���� ��� ������ �� ��� � ������� �� �� ��*� � �� ��� �0'��� � ���� �� �� ��� ��� ������� ��������� � ��� ������ �� ��� '������� � �) � 3� � �� �� � � �3� � 3�� � � � 3� 829 � *��� ��� e� � ��� ���� ��� ���������� 8������������ 9 �� '������� �� � �� 0 ��3� ��� '� � �� �) � �� ����� 8�� �����9( 1� ��������� ���� ���� �) � � �������� �6��� �� ��� 3� 8������ c )������ ",,-9( ��� �� 0��3� ��� ��� ��� �� ���i ����������� �������� �� ��� �*� '������� ��� ��� ����� �� ��'�� ( �� ��� �� ��'�� ���e�� ���� ��� ������������ �� '������� � � ��� ���� '� ��� �� ��� !!! 1���+� �� � �� 0��3� ��� '� � ��� ��� ������ !0���� ������ ��� ��� �'������ ( 1� >���� �� ��� ��� ������ ",-,.",,,� ��� *�� �� &����� ����� � !("22 ��� ��� *���� ���� �� ��� ��� �� � !(!2 ( &��������� �� >���� a� ��� �� 0��3� ��� *�� �� ������ �� ��� ��� ��� ������ ������� �� � !�2-�� � *��� �'��� ���� �� ��� ��� �� '������� � 3� � !� ��( 7������� '��� *�� �� ��3� ��� ������ �)" 8!(";k9 ��� �) 8!( 2k9 ��� �� ����� ��� ��� ������ ������ 8��� ��� ��� �'������ 9( ��� �� 0��3� ��� ����������� �� ��� � �� ����� ��� *��� ��� &c> $!! � � �� � ��� ��� �� '������� ( 1� ����������� �� &����� ����� �� !(";" � ��*�� ���� ���� �� ��� ��� �� ��� ��� ������ �'������( &��������� ��� �� ��� ��� �� ��� � �� ���� ���� �� ��� '������� ��� ��� �) ��� ��� ��� �� ����� �� ���� �'������( 7��� ��� �6���� ����� � *� ������ ���� ":! .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� �� ����� �� �� ��� �+�� ��� �� ���� "($ ���� �� 4����'�� �!� ",,� ��� �� ��� ��� 3��� '�� ���� ��� '����� 7�'����� -� ",,$( ����*���� ��� &c> $!! 1���+ *� ������� ��* ������ �� � m &������ '�� ����� ���� ��� *��� ���� �� ",,-� ��� *��� ���0��0���0��� �� ��� ��� �� � �� ��� ���� �� ����� "$� ",,*���� ��� &c> $!! ����� ���� �!k ����� ",,-( ;-%!�� ���� �+�������� � ����, � � ���������� ���� ���:�� �����, ������� /� � � � ������������� ��������� ��������� '� >� ���� ��� ��������� 8",, 9 ��� �������� '� >���n�� 8",,:9 ��� ��������� ��� ��� �� ����� �'����� �� ���( >� ���� ��� ��������� ����� � � �� � �� ��� ���������� '��*��� ��� ������ ��� �+������ �������� � �� ������� ���������� �� ��� �� ��������� ( ��� �� � � �������� �� � ��� �0 ���� ���� ���������� � �����* � � � !� !� �� ! � � !� "� 8$9 *���� ! � �=����� � =������*� �=��� � ��� �������� � �� ��� ���� ���� '��� ��0 ��� ��*�0 ��� �� ��������� ��� ��������� ������ ��� '� ��� �����( =��#�� 8=���#��*�9 � ��� ���'�� �� *�� ��� � ��0���0��� �� 8���0��0���0��� ��9 �� �� 8��*�90��� �� ��������� � ��� =��� � ��� ����� ���'�� �� *�� �� ��� �� � ������( 5� ��� ����� ����� !� � �!����!����� ��'�� �� 0��3� ��� ��� ��� �� ��� ����������� ���� ������ &����� 1���+.�� ��� !!! &����� 1���+.&c> $!! ��� a������� ��� a������� >���� �� &����� 8�9 ��� ��� ������ ������� !("22 !(!2 !(" ; !(!:, 7�� � �'������' !(":!(! : !("!: !(! : &����� �'������� !(" : !(!$" !("�� !(!,: ����� �'������� !("2; !(!2� !("!; !(!-$ 7����� �'������� !(""$ !(!2: !(";" !("-� ��.�� ��� !!! ��.&c> $!! �).�� ��� !!! �).&c> $!! ��� a������� ��� a������� �)" �) �)" �) >���� a� ���� ����� ��� ���� ����� 8��9 ��� ��� � *��� � ������ ��� )����� 8�)9 ��� ��� ������ ������� !(2-k !( �k !(2 k !(�2k !(";k !( 2k !(!;k !(!,k 7�� � �'������' !(;-k !( :k !(2-k !( :k !( 2k !(�,k !("�k !( !k &����� �'������� !(2$k !( ;k !($"k !(2"k !(" k !(",k !(!;k !("!k ����� �'������� !(�,k !(",k !(�!k !( 2k !("-k !( "k !(!$k !(!;k 7����� �'������� !(�$k !( !k !(2!k !(2 k !(!,k !("$k �!(! k �!(! k � ������ ������� %��� ":� ",-, �� 7�'����� " � ",,,( ' 7�� � �'������� %��� ":� ",-, �� �� � � ;� ",:2( � &����� �'������� �� � � ;� ",:2 �� ��� � ",:,( � ����� �'������� ��� � ",:, �� 4����'�� �!� ",,�( � 7����� �'������� 4����'�� �!� ",,� �� 7�'����� " � ",,,( .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ":" �"� !����"� !���� � ��� ��������� �������� � �� ������� ��� �� ������ � ����� ��� ���� ������� � �� �� ��� �� ����� �'�����( !�� � �=��#�� � =��#��*� �= ��� � ��� �������� � �� ��� ���� ��� � ��0���0��� �� ��� !��� � = ���#���=��� � ��� �������� � �� ��� ���� ��� ��� ��� �� � ��( � ! � �"�=����!��"� !�� � ��� �������� �� !( � !� � �"�=����� !��� � "� 0 �!����"� !��� � �"�=����� !�� � "� �!�����"� !���� � ��� �������� �� !�( 5��� �� ����� ��� ���� ������� � ��e���� ����� �� � ������� �������� �� ��� �� ����� ��� � ���� ����� �'����� ��6���� ��� �������� � �� ������ ������� ���������� � !� �� �+���� ��� ��������� ���������� ����� ��� ���� ������� � �� �� ��� �� ����� �'������ !�( � ��*� �� >���� � �� ��'�� �� ��� ��� ��� � ��� ���� '� ��� �� ��� !!! 1���+� *� �� $;($k o!��� � $-,<"�! $p �� ��� ���� ��� ��*� 2�($k �� ��� ���� ���� ��� ������ ",-,. ",,,( � ��������� '������ ��� *� ��0���0��� �� ;"(2k o!�� � ; ,<"�! $p �� ��� ���� ��� ���0��0���0��� �� �:(;k �� ��� ����( ���� ������������ ��� ���������� �� ��� ����� *��� ������� $$(�k o! � ��-$� ", �<"�! $p �� ��� ����( ��� �+������ �������� � �� ��'�� � 4������������ �� � �� ��� ��� �� ����� �'����� �� ��� !!! &c> $!! =�� =��� ! !� � =�� =��� ! !� � >���� �� �0 ���� ��� ��� ��� ������ ������ 8%��� ":� ",-, �� 7�'����� " � ",,,9 g�0��� �� �-$ $-, �;: $-2 d�*�0��� �� $2 22; ;" 2$" ����� ; , "�! $ $$(�k $"($k ($$�� ; , "�! $ $2(2k $"(2k (!2� >���� a� �0 ���� ��� ��� ��� e� � �'������ 8%��� ":� ",-, �� �� � � ;� ",:29 g�0��� �� -"2;$ "�! d�*�0��� �� 22 " ! $; "������ " " ;$-(�k 2,($k ($-�� " " ;$2(-k $!("k "($! >���� ?� �0 ���� ��� ��� ��� ����� �'������ 8�� � � ;� ",:2 �� ��� � ",:,9 g�0��� �� "!! "2 ,"22 d�*�0��� �� -; "!: -, "!; ����� "-; $! $ (:k $ (:k !(!" "-; $! 2,(;k $�("k �"( � >���� d� �0 ���� ��� ��� ��� ����� �'������ 8��� � ",:, �� 4����'�� �!� ",,�9 g�0��� �� ""! "�" "!" ; d�*�0��� �� -, "!$ : ""! ����� ":, �; $-(;k $�(�k "(;-� ":, �; $-( k $!(-k "(,,� >���� �� �0 ���� ��� ��� ��� ������ �'������ 84����'�� �!� ",,� �� 7�'����� " � ",,,9 g�0��� �� :: "$, ,, "-2 d�*�0��� �� $$ ""� 22 ,: ����� "2� $�(-k $!(2k "(!, "2� $;( k $!(-k "(,!� =��� ���'�� �� *�� ���� ��� � ��0���0��� ��# =���� ���'�� �� *�� �� ��� �� � ������# !� �������� � �� ���� ���� ��� ��������� ������ ��� ��� �� ��������� # !�� �+������ ���������� �� ������� ������ � ����� ��� ���� ������� � �� �� ������ ��� �'�����( � � �!� !����� ! � � !� "� � *���� � ! ��� � !� ��� ��� �������� �� ! ��� !�� �� ���������( ���� 1������� � ��e����� �� ��� !($ ��� $(!k ����� �� '����� 8���0���� �� �9� �� ���������( ": .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ������� ���������� � !�� � $"($k( 5������� ��� *� ���� ���� ��� *��� �� ��� �� ������ � � ���� ��� ���� ������� � �� �� ������ ��� �'����� ��� '� ��3����� �� ��� !($k ����� �� � ��e����� �� � ���0���� �� � 8� � �$$9( ��� *� �� � ���� ���� ��� *��� �� ��� �� ������ � ����� ��� e� � �'������( ��� �0 ���� ��� �� ($��� ��� ",-,.",:2 ������ � ���� ������� � ��e���� �� ��� !($k �����( b� � ��e���� ������ ��� �'����� � ������� ��� ��� ����� �'������ 8� � "�;-� � ��e���� �� ��� $k ����� �� '�����9( )�*����� ��� ������� � �� �� ������ ��� �'����� ������ '� ��3����� ��� ��� ����� �'������ �� ��� "!k ����� �� � ��e����� �� '�����( ��'�� � �� � ������ ���i �'����� �� ������ � ��� ��������� �� ��� ��� �� � �� ��� &c> $!! 1���+( /���� ��� �0 ���� ��� �� (!2 ��� ��� ������ ������ � � ��e���� �� ��� $k ����� �� '������ *� e�� ���� �� �������� �� ���i ��� �� ����� �'����� *��� � �� ��� �� ��� !!! �� ���+� ��� ��� ��( ��� �� ��� � ��� � ���� *��� ��� � �� ?����� ��� )����� 8",,:9 ��� f� ��� 8",,!9 *�� ������ ���� ��� �� ����� �'����� ������ ��� ���� � � �������� �� 4�&d�q 8 ����0e�� ��� 9 ��� ��� *�� � � �������� �� ��� &c> $!! 8��� �0e�� ��� 9( ��� �� ��� � �� � ��� � ���� ���� ��� ������ ���� �'������ � �+���� ��� ��� �� � ��� *���� ��� �0 ���� ��� � �� ��� &c> $!! � � ��e���� �� ��� $k ����� �� '����� 8� � "�,!9 *����� ��� �0 ���� ��� � �� ��� �� ��� !!! � ��� � ��e����( ;-;!�� ���� �+�������� � ����, � ��������� ���� ���:�� �����, ������� /� � � � �� �� ���0'� �� �����������0���������� ����� ���� ������� ������� ��� ��� �� ����� �'����� ���� ������ �� ����� ��� ��� � ��� ��������� ��� �� � ��� ����+ ��'���� �( ��� �*�0'��� �� �� ��� ����� � �������� '� f�� �� ��( 8",,:9 �� ���� �� ��� ��� ���������� ��� ��� �� ����� �'����� �� ���� �� �� ������������� ������ ���� ��� � �+��� �� � �����* � 3�� � 3�� � �� �"�3�� � 3��� � � >� � �� 8;9 *���� 3�� � 3�� � *�� �� �+�� ������ �� ��� �� ���� �# 3�� � 3�� � *�� �� �+�� ������ �� ��� ��� �� �� ���� �# >� � ��+�!���3�� � 3��� # � � ��������� ����# ��� �� � ������ ����� ���� �� ���� �( /��� 3�� � 3��� �6( 8;9 '����� 3�� � 3�� � �� �"�3�� � 3��� � ��� *���� �" � ��� ��0��� �� '���� ��( /��� 3�� � 3��� �6( 8;9 '����� 3�� � 3�� � �� ��" � � ��3�� � 3��� � �� 8 ���� 3�� � 3�� � �>�9� *���� ��" � � � � ��� ��*�0��� �� '���� ��( ���������� �� � �� � � ( � ����� ��� ��� ���� � �� ��� '��� �� � ������� �� ����� �� ��� ��� �� ����� *��� ��� �� ��������� ���� �( &���� ��� � ������� �� �� � ��� �� ����� ��� ��� ��� ��� ���� ���� ����� ��� �� ����� �'����� � ��� ���� � ��� ���� � �� '��� ���� �� ��0�� � ��*�0 ��� �� ���������( 7�� � ����� ��� �� ������ ��� ��0��� �� '���� ��� ����� '� ������ ���� ��� ��*�0��� �� '���� ��� ���� � � �� � '� �� ����� ��� ���� ������� � ��e����( � �� ����� ��� ���� ������� � ��e���� �0 ���� ��� ��� � ������� �������� �� ���� ��� ��� �� ����� � ���� ��� �+������ ����� �� � � ���� ����� ��� ���� ������� � �� �� ��� �� ����� �'�����( ��� ��������� ���� � � � ��� ��� �� ��� ������ �� ����� ��� �� ��� ��� �� ������ ����� e������ ��� �� ��� �� ����� �'�����( &���� ��� ��� ��� ��� ������ ���� '�� ���� ������� � ������ � �� ��� ��� �� ���������� � � �� ��� �� ��� ��� ����� ���� �� ��������� ���� ����������� ������ ��� �� ��� ����+ ��'���� � 8/�� ��� ",,"9( .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ":� /� ������ �������� ��� � 6���� ������ 85b&9 ��� ��� �� ��� !!! � ��� ��� �� ����+ �� � ������ ��� ��������� �� ��� �*�0'��� ����� �����( 5b& ��������� � ������ ��� �������� �� ��'�� 2( 7�� ��� ������ ������� �� ��� ������ ���� ��� �+��'�� ��� �����'�� ��� �� ��� �� ����� ( 1� >���� � �� ��'�� 2� � � �� ����� ��� ���� ������� � ��e���� �� ��� !($k ����� �� '����� 8� � ""� 9( � � �� � �� ����� ��� ���� ������� � ��e���� �� ��� ������������ $k ����� �� '����� ��� ��� �'������ ( 7������� ��� ���� ������� � ��e���� �� ����� � ������ ���� ��� ��� ��� ��� �� ����� ��� '� �������� �� ����� ������ �� ������ �� ����� ��������� ( 4� ����� ��������� ���� ��� �� � �' ����� �� ������ ����� ����� '�� ����� � �� �������� �� �� �+��������� ��� �� ��� � 8a���� c %��3� ���� ",,-# /�� ��� ",,"9( ! "����� �� ���������� /� � � �+������ '�� ��� ��� �������������� ������� '� ��� ������ � ������������ ����� ��������� '� � ������������ �� ��� �� ����� � ��� ��� ����e� ���� ��� � �������������� *��� ����������� ��� ��� ������ �� b������ 1��� ����� >������ � 1��(� �� ��* ��* ���� ��� ��� � � ������ �����������0���������� ��������� �� �� ���� �� ��� *������ ��� �� ����� ��� �����( /� ������� �������� �� ��� ������� ����������� ��� ��� �� ����� ��� �� ��� � �� ���� ������ � �������� ���� � !0���� ���� ������ 8���� ",-, �� ",,,9 ��� � �� ��� �� ��� !!! 1���+ � � '������� ( 5��� ��� ������ ������� ��� ������������ �� '������� *��� ��� �� 0��3� ��� ��� ��� ( ��� �� � �+��'���� � ���� �'����� �� ������ � ��0 ��� ��*�0��� �� ��������� � ��� �0 ���� ��� '��� � ��e���� �� ��� !($k ����� �� '�����( 1� ��������� ��� '��� �� � ������� �� ����� �� ��� �� ��0��� �� ��'�� 2 >��������� �� � �� ��� ��� �� ����� �'����� d�������� h����'�� � �" � � 3 = >���� �� �� �� ��� ����e����� � �� ��� ������ ������ 8%��� ":� ",-, �� 7�'����� " � ",,,9 3�� � 3�� �!(!! 8�2(2"9�� !(-"; 8 $(",9 !(2-� 8""( 9 $!("�! $ >���� a� �� �� ��� ����e����� � �� ��� e� � �'������ 8%��� ":� ",-, �� �� � � ;� ",:29 3�� � 3�� �!(!! 8� (",9� !(;:! 8" (2,9�� !($", 8;(!!9�� 2-(; ;>���� ?� �� �� ��� ����e����� � �� ��� ����� �'������ 8�� � � ;� ",:2 �� ��� � ",:,9 3�� � 3�� �!(!! 8��(":9�� !(-$$ 8"�(:!9�� !($;: 8-(;29�� $;(: $! >���� d� �� �� ��� ����e����� � �� ��� ����� �'������ 8��� � ",:, �� 4����'�� �!� ",,�9 3�� � 3�� �!(!!� 8��(" 9�� "(!" 8":(:"9�� !($-: 8;(:-9�� -$( �; >���� �� �� �� ��� ����e����� � �� ��� ������ �'������ 84����'�� �!� ",,� �� 7�'����� " � ",,,9 3�� � 3�� ! 8�!( �9 !(2-" 8:(!;9�� !(";: 8"(,29� �-(, �� ��� ��� ���� � �*�0'��� �� �� ��� � �� ��� *�� �� �+�� ������ � 3�� � 3��� � ��� ��������� �����'��( ��� ����������� �����'�� ��� ��� �� �+�� ������ � 3�� � 3��� � �� ��� �� ��� !!! � 3�� ��� � �����'�� �������� ��� ��*�0��� �� �� �������� >� � ��+�!���3�� � 3��� ( ��� ���'�� �� �������� � ��� ��� �0 ���� ��� ��� ����� � ��e����� �����( ���� 1������� � ��e����� �� ��� !($ ��� $(!k ����� �� '����� 8���0���� �� �9� �� ���������( � � � �� � ��� *���� �� � ��� ��0��� �� '��� ��� �� � ��� ��*�0��� �� '���( ":2 .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ��������� � ������ ���� �� '��� �� ��*�0��� �� ��������� ���� ��� ����� ����e����� � � � �� ��� �*�0'��� �� �� ��� ��� �� ����� ��� � ��e���� �� ��� !($k ����� �� '�����( � �+������� ��� �� ��� ��� ��+�� *��� � � �� ���i ��� �� ����� �'����� ���� ������ �'������ �� �����+������� $ ���� ( ��� �� ��� ��� ��� ",-,.",:2 ��� ��� ",:,. ",,� �'������ ��� � ��e���� �� ��� !($k ����� �� '����� ����� ��� ��� ��������� �� ( )�*����� ��� �� ��� ��� ��� ",:2.",:, ��� ��� ",,�.",,, �'������ ��� ��� � ��e���� �� ��� ������������ ����� �� $k �� '����� � �� ��� ������������� ��� ���������� �� � �� ���i ��� �� ����� �'�����( 5������� ��� e���� ������� ��� �������� �� ���i ��� �� ����� ��� �� ���� ���� ���� � ������� ������ �� ���� � � ��� �� ������� ������( #��������� a�� ��� ?(� 7�� ��� /(� ���� � d( )(� c &������ �( 8",,,9( ?���������� ��� �� ����� *��� '������� ���� ��� ( � ����� ����� ��� & � �� �� �5� "",."2:( a�������� �( b(� c h��� ����� �( >( 8",,"9( ��� ���� ��� �� ����� ����� �� ( ����� ��� ������� � ������ ;7� -:.:2( a����� r( s(� c %��3� ���� h( 8",,-9( � ���+��������� �� ��� ��� ��0����� ��� �������0 �������� ����������� �� ������ ���� ( ����� ��� ������� � ������ �%� 2.�!( a������� %(� c ?������ >( �( 8",,29( d�� ��� �� ����� ������ *�� �� ��� ���� *����t � ����� ! �� �� 1���,������ 5�� �,.22( a��� � /(� b� ��� �� � %(� c b�a����� a( 8",, 9( &����� ��������� ������ ���� ��� ��� ����� ��� ��������� �� ��� ������ ( � ����� ����� �� ;7� "-�"."-;2( ?���� ?( �(� b��� ?( 7(� ������� &(� c ?���� �( 8",, 9( � ��� 0 �������� ����� � �� ������ ���� i ��� �� ����� ��� ������� �������� ���( � ����� �������� ����� � ��" �����,� #$� ;$,.;-;( ?������ d( �(� c )������ �( b( 8",,:9( ��� ����������� �� ����� ����� ��� �� ����� � d���� �������� ���� �+������ ����� �� ( >���� ��� ����� �� ��� ",,: 7�������� ���� ����� � �������� ������ ������ ( ?��'�� �( �(� c ����� %( d( 8",,!9( ��������� ��� ����������� �� ��� ������������� ������ ���� ( � ����� ����� �� ;�� 2,-.$ "( ?�������� �( �(� c ?�������� �( �( 8",,,9( ��� �� ����� � &���� ��� ��� �������� � �� ��� h1u( ����� ��� ������� � ������ ��� -�.:"( d������ f(� ����'����� �(� ������� &(� c /����� � �( 8",,-9( ��� ���� ������ ���� ����������� *��� ���������� ���0'� �� '������� ( � ����� ����� �� �5� "!�$."!$:( d��� � /( �(� c /�� ��� d( �( 8",,29( 1��� ����� ����������� �� ������������� ������ ���� ( � ����� ����� ��� 3����� �� #7� "."2( ���� ?( &(� f������� �(� c �� ��� � a( �( 8",,"9( g(&(0'� �� ������������� ������ ���� � � ����������� ����������( � ����� ! �� �� 1���,������ #7� ::.,2( 7�� ��� /(� c &������ �( 8",,;9( ��� ���� ���� ����� � ��� ����������� �� ���� �� �������� ��������� ( � ����� ����� �� �#� 2 $.2; ( ���������� /( 4(� 1� �� ���� %(� %�(� c 1� ����� r( 8 !!!9( ������� ��� ������� �� ����� ����� ( � ����� ����� ��� ��" ?����������� �������� %�� $-. ,!( ������� %( �(� c )������ ?( �( 8",,;9( ��� �� ����� �'����� ��� ���������� ������� �� ���� ����� ��* ������ i � �� ���������� �������������� ( � ����� ����� ��� & � �� �� ;5� �,-.2 "( ������� %( �(� c )������ ?( �( 8",,-9( ������ ��� ����������� �� ��� �� ����� ��* ������ ( ����� ��� ������� � ������ �%� $2.;;( ����'����� �(� c ������� &( 8",,29( � ���� �� ������ ���� ������ ��� ����������� ���������� ������6�� ( � ����� ����� ��� ��" ?����������� �������� 5$� 2",.222( f���� �(� c ��� � &( �( 8",,!9( ��� �������� �� ��� �� ����� ������ � 4�* ������ ��� � ��� �� ����� ���� ( � ����� ����� ��� @������"���� �� #� "$!.";;( .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� ":$ f��� �(� ?��� � b( �( /(� c ?���� f( ?( 8",,:9( 1������������ ������ ���� ���������� ��� ��� �� ����� ����� �� ��� ��*�0��� �� ��������� ( ����� ��� 3����*� %%� " -."22( f� ���� �( /( 8",,!9( ��� �� ����� *��� ���� ��� � ��� �0e�� ��� � >�������� ��� ��� ��6����� ���������� �'�����( ����� ��� ������� � ������ ;�� ;�.;,( b�� ��� �( �(� %�(� c /����� � �( d( 8",,$9( ��� �� ����� ��� *�� �� ��� ���� *����( � ����� ! �� �� 1���,������ 5#� -2.:"( b��� /( 8",,-9( ��� �� ����� ��� ����0���� ������ � ���� ( � ����� ! �� �� 1���,������ 5%� �$.2;( b��� ?(� c ������� &( 8",,!9( ��� �� ����� � ����������� ��� ������ ���� ������������ �� ��������� �+���������( � ����� ��������� �%� ;". -:( ��� ���� a( �( 8",,$9( ������ ���� ���� ��� �� �6���� ������ ���� � ",-".",,"( � ����� ����� �� ��� $2,. $( ���� ������ 7(� c ���� ������ b( 8",,-9( �� 0��3� ��� �����������( � ����� ! �� �� 1���,������ 5%� 2$. $2( >���n��� �( 7( 8",,:9( ������������ � ��e���� ��� ��� �� ������ � ( .�� ����� ��� 3����*� %%� ;$.-;( >� ����� �( )(� c ���������� �( 8",, 9( � ����� ������������� �� � �� ���������� �����������( � ����� �������� ��" & � �� �������� �� #�� 2;".2;$( >������� b(� c a������ /( %( 8",,:9( ��� ����������� �� �� ����� ���� ���� � ��� ��� �� ����� ( 3����* ����� ��� & � �� �� 7� ".��( ����� &( >(� c ������� ����� /( 8",,�9( ��� �� ����� ��� ��� ���������� ���� ���� g �� ��� ��������'����� �� ��� 0������� ��� ������ �� ������� ��������� �����������( ����� ��� ����� �� 3����*� %� ,.2�( &������ � �( �( 8",, 9( ��� �� ����� � a����� ���� � '��0���0���� ����� �( ����� ��� ������� � ������ ;9� 2;. $!( &�������� �(� ����������� �(� c /����� )( 8",,,9( d���0 ������ � ��������� ������ ���� ������������ ��� ��� '��� ����( � ����� ����� �� �;� ";2-.";,"( /� ���� %( ?( 8",,-9( /�� ��� �� ����� *�� ( � ����� @�������,� �� -:.:"( /� ���� %(� &������ � &(� c >���� �( 8",, 9( ��� �� ����� *�� *���� �� ������ �� � �� ��� ���� *����( � ����� ! �� �� 1���,������ #9� :;.,!( /�� ��� �( %( 8",,"9( ��� ����������� �� �������� � �� ����������( ����� ��� ������� � ������ ;7� ;�.-"( ":; .�/��� 01 1���� 2 ����� ��� ����� �� 3����* #� 45��#6 #7%8#9� evaluating a stock market timing strategy: the case of rte asset management introduction literature review data rte and its market timing strategy return calculations and composition performance-evaluation: methodology and results performance-evaluation using total returns performance-evaluation using risk-adjusted return measures performance-evaluation using a nonparametric test of market timing ability performance-evaluation using a parametric test of market timing ability summary and conclusion references pii: s1057-0810(01)00067-1 the effect of country-specific index trading on closed-end country funds: an empirical analysis matthew o’connor, ph.d.a,*, edward a. downe, ph.d.b aquinnipiac university, 275 mount carmel avenue, hamden, ct 06518 buniversity of new haven, 300 orange avenue, west haven, ct 16516 received 13 march 2000; received in revised form 17 august 2000; accepted 24 january 2001 abstract the american stock exchange initiated trading in 17 world equity benchmark shares (acronym “webs”™) in april 1996. webs are index funds designed to track the morgan stanley capital international (msci) indexes. we examine the effect of this event on closed-end country funds (cecfs) and find that percentage discounts increase. cecfs with a corresponding webs index experience the largest increase and also show a decline in trading volume. we attribute these results to (1) the effects of increased competition and (2) a reduction in the market segmentation premium. since additional webs have begun trading, and other approvals may follow, similar effects could be experienced in the future. © 2001 elsevier science inc. all rights reserved. jel classification:g140, g150 keywords:investments; market efficiency; international financial markets 1. introduction closed-end funds are mutual funds that issue a fixed number of shares that trade on a stock exchange. owners of closed-end funds liquidate their shares by selling them to other investors. whereas open-end funds stand ready to redeem shares at net asset value (nav), closed-end funds usually sell at a discount and sometimes at a premium to nav. however, when the funds are initiated, they always sell at a premium. the start-up premium results * corresponding author. tel.:11-203-582-8297; fax:11-203-582-8664. e-mail address:matthew.oconnor@quinnipiac.edu (m. o’connor). financial services review 9 (2000) 259–275 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00067-1 from underwriting fees and up-front costs associated with the flotation. attempts to fully explain discounts or premiums on a rational basis have generally failed, and this has become known as the “closed-end fund puzzle.” nevertheless, it is not surprising that closed-ending is often more attractive than openending for funds specializing in a foreign country, i.e. funds in which the underlying assets trade in a foreign market. such closed-end country funds (cecfs) may be illiquid, and closed-ending ensures that assets will not have to be sold off at fire sale prices to meet redemptions during market declines. hence, cecfs represent rational international investment vehicles for many domestic investors. beginning april 5, 1996, the american stock exchange initiated trading in world equity benchmark shares (acronym webs™) designed to track the morgan stanley capital international (msci) indexes for 17 countries. msci indexes have been the traditional benchmarks for international portfolios. each webs index series holds a representative sample of the underlying securities in a corresponding msci index. since webs can be sold to the fund or redeemed by the fund in aggregations of securities called “creation units,” webs can be easily arbitraged and therefore track closely the values of their respective msci index. however, irregularities in portfolio sampling techniques and other regulatory constraints (mainly tax-related) allow webs values to vary slightly from msci indexes. few open-end mutual funds specialize in a single foreign market. hence, prior to the arrival of international index funds, cecfs were often the only economical avenue for individual investors to diversify into single market foreign investments. therefore, webs are an important integrating event and should impact cecf discounts. it should be noted that the week prior to the introduction of webs,deutsche morgan grenfellintroduced eight foreign country basket index funds (cbs). although quite similar to webs, investors never fully accepted cbs, and the series was discontinued on february 10, 1997, less than a year after trading began. however, as an integrating product, cbs may have had a similar impact on cecf discounts. nevertheless, since cbs were introduced at almost exactly the same time as webs, were never accepted by investors, and have similar impact, we couch our discussion in terms of webs. this paper uses a modified event study methodology to examine the impact of webs trading on mean cecf percentage discounts. we divide cecfs into two groups: those with and those without a corresponding (same country) webs index. we separately estimate the impact of webs trading on each group. we construct tests to examine the short-, intermediate-, and long-run impact of webs trading. we also use a two-factor difference in means test to examine the differential impact on percentage discounts for cecfs with and without a corresponding webs index. finally, we use the same methodologies to examine the impact of webs on trading volumes. 2. theory and literature review considerable effort has been spent to explain deviations between closed-end fund prices and navs. explanations in the literature fall into two categories: (1) behavioral theories, 260 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 which focus on the irrational expectations of individual investors, and (2) economic theories that attempt to explain discounts within the framework of the efficient market hypothesis. among the behavioral proponents, delong et al. (1990) and lee, shleifer, and thaler (1991) attribute fluctuations of the discount or premium to changes in “investor sentiment.” these authors suggest that the closed-end fund market in the u.s. is characterized by negligible institutional participation and is dominated by individual investors, who are less likely to trade on fundamentals. the behavioral model suggests two possible reactions to the introduction of webs trading. investors may fail to appreciate the importance of webs trading, in which case no impact on discounts will be found. alternatively, investors may irrationally bid down cecf values. in contrast, economic explanations suggest rational premiums and discounts. the four classic economic explanations include: (1) biases in navs, (2) tax timing issues, (3) agency costs, and (4) international market segmentation. nav biases reflect differences between personal and fund accounting of capital gains liabilities. they also reflect rational discounting of some portfolio assets due to liquidity concerns. however, neal and wheatley (1998) find no evidence of a liquidity effect in closed-end fund discounts. the tax timing explanation states that closed-end fund investors give up valuable tax-timing advantages. seyhun and skinner (1994) and brickley, manaster, and schalheim (1991) find empirical support for a tax-timing effect. the agency cost explanation suggests that discounts are, in part, a rational response to excessive management fees, poor management performance, and managerial entrenchment. pontiff (1995) and barclay, holderness, and pontiff (1993) find evidence that agency costs are reflected in closed-end fund prices. richard and wiggins (2000) find that cecf premiums are able to predict nav returns after controlling for the return on foreign markets, indicating that at least part of the discount reflects rational assessments of managerial abilities. finally, the market segmentation explanation suggests that if a closed-end fund provides access to a restricted market, fund prices may reflect a rational premium. the restrictions need not be governmentally imposed. they may simply reflect the cost to domestic investors of creating well-diversified, single-country foreign investments. bosnerneal et al. (1990), errunza (1991), and levy-yeyati and ubide (1998) find evidence of a market segmentation premium in cecf prices. economic explanations suggest webs trading should tend to reduce market segmentation premiums, and the effect should be more evident in funds with a corresponding webs index. further, some investors may abandon excessive fee cecfs for lower fee webs, putting downward pressure on prices and increasing discounts. again, the effect would tend to be more evident in cecfs with a corresponding webs index. unlike an overreaction, these changes would tend to be permanent. the primary motivation for our paper is to empirically examine the effect of webs trading on cecf discounts. our tests also shed light on theoretical explanations for closedend fund discounts. in particular, the behavioral model will be supported by either: (1) no change in cecf discounts, or (2) a reversal pattern in which discounts initially increase but subsequently fall. on the other hand, a permanent increase in discounts favors the economic model. since our tests on percentage discounts do not specifically differentiate between a market segmentation premium effect and the agency cost effect, we also examine the impact 261m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 on cecf trading volume. if the agency explanation is correct, an impact on cecf volume should be evident. 3. data and methodology the data for this study are from thewiesenberger investment company. our sample comprises weekly closing market prices and navs. the data also include average daily trading volumes (in shares transacted) for each week in the study. the discount is determined as the nav minus the price of the fund, and the percentage discount is defined as the discount divided by the nav. the data are from all cecfs listed on the new york stock exchange and american stock exchange. we also include a country-specific canadian fund that trades on the toronto stock exchange, because a canadian webs is listed. all of the funds are categorized by thewall street journal(wsj) as “specialized equity funds.” we limit this study to country specific funds. we exclude blended funds (funds that invest in more than one country) because the impact of webs trading may be ambiguous. to be included in the study we require funds to trade over the 160-week time period from march 18, 1994 through april 4, 1997. since the introduction of a webs index may have a stronger impact on a cecf from the same country, we divide the sample into funds with and funds without a corresponding webs index. for cecfs trading over the 160-week period, we find 18 funds with and 25 funds without a corresponding webs index. appendix a lists the cecfs by group and shows the corresponding webs index for the first group. table 1 presents mean percentage discounts and grand average daily volumes for the 160-week period. note that volume data were unavailable for 4 funds. 3.1. modified event study the traditional approach to measuring event impact is cumulative abnormal residual (car) analysis. however, campbell, lo, and mackinlay, (1997 henceforth clm) caution against using car analysis when events are clustered. event clustering results in crosscorrelations among residuals, violating the normal car assumption of independence. since we examine the impact of a simultaneous event on all funds, we clearly have clustering. clm suggest a system of equations model used by shipper and thompson (1983 henceforth st), which we use below. the pre-event estimation window includes the 104-week period from march 18, 1994 through march 8, 1996. the four weeks culminating with the april 5, 1996 introduction of webs trading are the event window. observations during the event window are excluded to avoid unintentional biases in the estimation of post-event mean discounts. the remaining 52 weeks from april 12, 1996 through april 4, 1997 comprise the post-event evaluation period. we subdivide the post-event evaluation period into overlapping 13-, 26-, and 52-week periods. the model is constructed as a system of equations with each individual equation corresponding to a particular fund. we define t as the number of observations per equation. in 262 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 particular, t will equal the sum of preand post-event weeks. for example, if the post-event window is 13 weeks, then t will equal 521 13 5 65 observations. let yj be a t3 1 vector of observed percentage discounts for firm j. let d be a t3 1 vector of dummy variables with zeroes in the pre-event estimation period and ones in the post-event test period. finally, let «j be a t3 1 vector of serially independent error terms for firm j. then for each individual firm, the model can be written as: yj 5 mj 1 bjd 1 «j with «j ; n(0, s2) (1) let j equal the number of firms in a particular group. then, following st, the j individual equations are stacked into a system of equations as: y 5 xg 1 e (2) where: y 5 fy1···yj g, x 5 f x# 0 0 x# g, x# 5 [1 d], e 5 f«1··· «j g, andg 5 3 a1 b1··· aj bj 4 the model is robust to correlation of the«j across the j equations. hence, it is appropriate for clustered events. with a known residual covariance matrixs, gls provides the best unbiased maximum likelihood estimate ofg. using an ols estimate for s, the joint gls estimate ofg is consistent and asymptotically efficient. since the explanatory variables are identical across equations, ols provides identical parameter estimates as gls. for all tests, the null hypothesis is that the introduction of webs has no effect on percentage discounts. a positive and significantbj coefficient indicates a fund’s mean percentage discount increases after the introduction of webs. however, following st, we also construct two sets of hypotheses regarding the joint impact of webs trading. the first joint hypothesis is that the sum of thebj equal zero: h0: o j51 j bj 5 0 (3) paraphrasing st, this test is analogous to testing the sample-wide abnormal returns in a traditional car methodology. we also test the hypothesis that the individualbj parameters are jointly zero: h0: bj 5 0@j (4) 263m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 this test may be useful if some funds react positively while others react negatively, hence canceling each other out in the sum. st give the following test statistic to evaluate the two cross-equation hypotheses: (a2 aĝ)9[a(x 9(ŝ21vi)x)21a9]21(a2 aĝ) (5) in the first joint test a5 0, and a is a row vector of ones and zeros. in the second test, a is a column vector of zeros, and a is a matrix of ones and zeros. as t goes to infinity, (5) is distributed asx2(q) in the limit, where q represents the number of restrictions. note that the hypothesis test from equation (3) has one restriction. the hypothesis test in equation (4) will have q5 j restrictions, because the system has j separate equations, one for each cecf. we separately estimate the system of equations model for each group of funds (i.e., those with and without corresponding webs). further, we estimate separate models for the three post-event time periods. in the first run, we use a 13-week post-event window. this indicates the short run impact of webs trading. to examine the persistence of fund discount responses over the intermediate and long run, we also estimate the models using 26and 52-week post-event windows. 3.2. the two-factor difference in means test extending our analysis, we also conduct a two-factor difference in means test. this allows us to examine whether webs had a significantly larger impact on corresponding cecfs. the model’s first factor is time and has two levels: the preand post-event observations. the second factor is fund group: cecfs with and without a corresponding webs index. to construct the model, again let y be a fund’s observed percentage discount. the dummy regression is: y 5 m 1 ad1 1 gd2 1 qd1d2 1 « « ; n(0, s2) (6) where: d1 5 h 1 if the observation is in the pre-event window 21 if the observation is in the post-event window d2 5 h 1 if the observation is from a cecf with a corresponding webs index 21 if the observation is from a cecf with no corresponding webs index we estimate the model as a single regression on a pooled sample of percentage discounts from all 43 funds. hence,m is the average discount over the entire sample period for all funds. thea parameter measures the effect due to the time factor,g measures the effect due to fund type, andq measures the interaction effect between the two factors. in order to investigate the differential impact on cecfs, let the subscript i equal 1 for a pre-event observation and 2 otherwise. let the subscript k equal 1 for an observation from a cecf with a corresponding webs and 2 otherwise. then the mean percentage discount (mik) for each level of each factor can be recovered from the model via substitution. for example: 264 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 table 1 descriptive statistics fund name mean percentage discount mean daily volume funds with a corresponding webs austria fund 16.9% 22,097 canadian general investments ltd 23.7% na emerging mexico fund 2.2% na first australia fund 14.1% 45,420 france growth fund 17.1% 28,544 germany fund 16.4% 44,572 growth fund of spain 18.5% na italy fund 12.2% 32,964 japan equity fund 28.1% 54,402 japan otc equity fund 24.8% 47,528 malaysia fund 5.7% 31,929 mexico equity & income fund 2.2% 50,991 mexico fund 6.4% 317,097 new germany fund 21.0% 87,823 singapore fund 21.7% 23,921 spain fund 15.0% 27,183 swiss helvetia fund 10.5% 27,619 united kingdom fund 15.5% na funds without a corresponding webs argentina fund 0.4% 49,137 brazil fund 4.4% 86,095 brazilian equity fund 2.2% 45,249 chile fund 9.7% 39,204 china fund 21.8% 40,359 first israel fund 2.9% 40,211 fidelity philippine fund 18.8% 130,587 greater china fund 6.9% 34,686 india fund 6.5% 14,098 india growth fund 27.0% 11,471 irish investment fund 13.3% 34,384 jardine fleming china region fund 5.2% 124,198 jardine fleming india fund 1.1% 31,099 korea equity fund 3.4% 37,033 korea fund 29.1% 136,821 korean investment fund 3.0% 39,131 morgan stanley indian fund 1.9% 55,896 new south africa fund 18.8% 76,657 pakistan investment fund 12.8% 39,120 roc taiwan fund 0.0% 25,720 taiwan fund 22.9% 19,338 templeton china world fund 7.3% 78,475 thai capital fund 6.7% 48,554 thai fund 6.4% 27,581 turkish investment fund 215.6% 28,825 calculations based on 160 total observations. na 5 not available. 265m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 5 m11 5 m 2 a 2 g 1 q m21 5 m 1 a 2 g 2 q m12 5 m 2 a 1 g 2 q m22 5 m 1 a 1 g 1 q (7) the relevant null hypotheses are: 1. h0: m21 2 m11 5 0 the change in the mean percentage discount for funds with a corresponding webs index is zero. 2. h0: m22 2 m12 5 0 the change in the mean percentage discount for funds with no corresponding webs index is zero. 3. h0: m21 2 m11 2 (m22 2 m12) 5 0 the change in the mean percentage discount for funds with a corresponding webs index is no different than for funds without a corresponding webs index. coefficient restriction tests of these hypotheses can be constructed from the definitions of the mik. 4. results 4.1. results of the modified event study table 2—panel a presents results for the j5 18 funds with corresponding webs indexes. for each individual equation, the intercept shows a fund’s mean percentage discount over the pre-event estimation window. each fund’s intercept remains constant, regardless of the length of the post-event window. thebj coefficients estimate the change in the mean percentage discount from the preto the post-event period. for the 13-week window, 15 of the 18 funds (83%) have positivebj coefficients, with 14 of the positive estimates significant at the 1% level. the average change in percentage discount is 6.8%. more importantly, the hypothesis that the sum of thebj coefficients equals zero is strongly rejected. this result is analogous to finding significant cars in traditional event study methods. not surprisingly, the hypothesis that thebj coefficients are jointly zero is also strongly rejected. the 26-week and 52-week post-event runs confirm that the increase in percentage discount persists. in the 26-week run, 17 of the 18 funds (94%) have positivebj estimates. of these, 15 are significant at the 1% level. the average change in percentage discount is 7.7%, and not surprisingly, both joint hypotheses of no change in percentage discount are strongly rejected. similar results occur in the 52-week run. the significant and persistent increase in percentage discounts is consistent with an economic explanation of closed-end fund discounts, and inconsistent with a behavioral 266 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 table 2 modified event study results (dependent variable5 percentage discount) fund number fund name intercept b̂j for post-event window of: 13-weeks 26-weeks 52-weeks panel a: funds with corresponding webs 1 austria fund 14.5%** 6.7%** 7.2%** 6.8%** 2 canadian general investments ltd 25.4%** 20.3% 21.5%* 25.1%** 3 emerging mexico fund 26.0%** 22.1%** 22.7%** 23.6%** 4 first australia fund 12.4%** 4.2%** 4.4%** 4.8%** 5 france growth fund 15.4%** 4.2%** 4.7%** 5.1%** 6 germany fund 15.0%** 3.7%** 4.6%** 4.0%** 7 growth fund of spain 18.3%** 20.7% 0.2% 0.7% 8 italy fund 9.9%** 5.3%** 6.3%** 6.7%** 9 japan equity fund 28.1%** 20.9% 2.8% 0.6% 10 japan otc equity fund 27.7%** 6.2%** 10.7%** 9.3%** 11 malaysia fund 3.2%** 8.1%** 8.1%** 7.4%** 12 mexico equity & income fund 26.1%** 22.7%** 23.6%** 24.2%** 13 mexico fund 1.0% 13.7%** 14.3%** 15.7%** 14 new germany fund 19.5%** 5.5%** 5.5%** 4.3%** 15 singapore fund 24.7%** 2.3% 5.5%** 9.2%** 16 spain fund 12.0%** 7.0%** 7.6%** 8.6%** 17 swiss helvetia fund 7.4%** 7.1%** 8.2%** 9.2%** 18 united kingdom fund 14.6%** 4.6%** 3.4%** 2.5%** average 7.6% 6.8% 7.7% 7.6% % positive 83% 94% 94% x2 statistics h0: sum of parameters5 0 72.275** 180.734** 348.169** h0: each parameter5 0 213.281** 432.698** 860.429** panel b: funds with no corresponding webs 1 argentina fund 23.0%** 7.5%** 8.5%** 10.6%** 2 brazil fund 20.7% 9.1%** 12.3%** 15.0%** 3 brazilian equity fund 24.4%** 13.0%** 18.1%** 19.8%** 4 chile fund 9.8%** 21.4% 0.3% 0.5% 5 china fund 27.9%** 10.5%** 14.1%** 19.1%** 6 first israel fund 22.7%* 16.3%** 16.1%** 16.7%** 7 fidelity philippine fund 18.2%** 2.4%** 1.9%** 1.6%** 8 greater china fund 2.3%** 9.4%** 12.0%** 14.3%** 9 india fund 9.5%** 210.6%** 26.5%** 27.6%** 10 india growth fund 26.2%** 24.7% 20.3% 21.5% 11 irish investment fund 12.6%** 1.8% 3.4%** 3.1%** 12 jardine fleming china region fund 0.1% 8.4%** 12.6%** 15.6%** 13 jardine fleming india fund 1.0% 1.1% 2.1% 1.9% 14 korea equity fund 5.0%** 21.6% 22.1% 24.4%** 15 korea fund 28.7%** 0.3% 20.2% 21.1% 16 korean investment fund 3.1%** 3.1% 2.4% 0.0% 17 morgan stanley indian fund 5.4%** 28.9%** 25.8%** 29.5%** 18 new south africa fund 18.8%** 0.9% 0.9% 0.9% 19 pakistan investment fund 18.0%** 211.3%** 212.2%** 215.3%** 20 roc taiwan fund 20.9% 25.1%* 24.4%* 4.0%** 21 taiwan fund 24.5%** 27.6%* 23.6% 6.4%** (continued on next page) 267m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 model, which suggests either no reaction or an overreaction to an integrating event. from the economic perspective, the increase could result from either a reduction in the market segmentation premium or a migration to index funds with lower fees. however, this test does not differentiate between the two explanations. one interesting observation from panel a is the relative magnitudes of thebj estimates for the three mexican funds. for example, in the 13-week window, thebj estimates for the emerging mexico fund, the mexico equity & income fund, and the mexico fund are 22.1%, 22.7%, and 13.7%, respectively. these estimates are much larger in magnitude than those of any other funds. this raises the possibility that our results are overly sensitive to changes in mexican fund discounts. to test this possibility, we re-run the model dropping the three mexican funds. although not shown, the hypothesis that the individualbj estimates are jointly zero and the hypothesis that the sum of thebj estimates equals zero again reject at the 1% level. these results hold across all three post-event windows. hence, we feel our results are not unduly influenced by the mexican funds. table 2—panel b examines the effect of webs trading on funds with no corresponding webs index. for the 13-week post-event window, only 9 of the 25 funds (36%) show a significantly positive (1% or 5% level) increase in mean percentage discount. hence, over the short-term, a lower proportion of these funds experience a webs induced impact. also, the average change in percentage discount is 2%, well below the 6.8% change experienced by funds with a corresponding webs index. finally, the hypothesis that the sum of thebj parameters equals zero cannot be rejected at the 5% level. for the 26-week period, 11 out of 25 funds (44%) experience a significantly positive increase in mean percentage discount evaluated at the 5% level. the average change in percentage discount is 3.6%, and the test table 2(continued) fund number fund name intercept b̂j for post-event window of: 13-weeks 26-weeks 52-weeks panel b: funds with no corresponding webs 22 templeton china world fund 2.9%** 11.4%** 13.0%** 13.4%** 23 thai capital fund 10.2%** 23.4%** 24.3%** 210.5%** 24 thai fund 10.7%** 23.5%** 25.8%** 213.0%** 25 turkish investment fund 222.7%** 13.6%* 17.7%** 22.6%** average 2.6% 2.0% 3.6% 4.1% % positive 60% 60% 64% x2 statistics h0: sum of parameters5 0 2.72 16.546** 41.663** h0: each parameter5 0 453.925** 939.345** 1590.159** *significant at 5%; **significant at 1%. results are from a systems of equations regression of weekly percentage discound against a dummy variable, d, that is constructed with ones in the post-event window. for each individual firm j, the equation is: yj 5 mj 1 bjd 1 «j, with «j ; n(0, s2). 268 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 that the sum of the coefficients equals zero rejects at the 1% level. the 52-week run has similar results. table 2 suggests that webs trading impacts both groups of funds. however, the effect appears to be larger for funds with a corresponding webs index. hence, we specifically test for a differential using the two-factor difference in means test. 4.2. results of the two-factor test we run the two-factor model for a pooled sample of all 43 cecfs. analogous to the systems of equations models, we run the model separately for each of the three post-event windows. table 3—panel a reports the parameter estimates, and panel b presents the results of coefficient restriction tests on each of the three null hypotheses. focusing on the 13-week post-event period, we reject the first two null hypotheses. these findings are consistent with the results from the modified event study. in particular, mean discounts for both categories of funds increase after the introduction of webs trading. moreover, we reject the third null hypothesis, confirming a statistically significant differential impact. these results hold for both the intermediate and long-term windows as well. overall, the stronger impact on corresponding closed-end funds is consistent with the economic explanation. these results are easily observed in chart 1 below. the chart shows the increase in average percentage discounts above pre-event means, for cecfs with and without corresponding webs indexes. clearly cecfs with a corresponding webs index have larger and more persistent increases in percentage discounts relative to cecfs with no corresponding webs. 4.3. volume analysis since the previous tests do not directly differentiate between a market segmentation effect and an agency effect, we also examine the impact on trading volume. if cecf agency costs are excessive, then investors should migrate to lower fee webs, reducing cecf volumes. as before, the impact should be more pronounced in funds with a corresponding webs index. to test this hypothesis, we re-run the systems of equations models with volume data replacing percentage discounts. in particular we now let yj be a vector of average daily volumes (observed weekly) for fund j. as before, we run the models separately by fund group and by post-event measurement window. table 4—panel a presents the results of the volume analysis for funds with a corresponding webs index. for this model, a negativebj coefficient indicates a decrease in trading volume after the introduction of webs trading. for the 13-week window, 11 of the 14 (79%) individual funds experience a decrease in mean trading volume. however, only one decrease is significant (5% level). furthermore, we fail to reject either the null hypothesis that all funds have zerobj parameters or that the sum of the parameters is zero. thus in the short-term, the direction of the volume change is consistent with reduced demand, but the level is not significant. for the 26-week window, only 9 of the decreases are significant at the 1% or 5% levels. further, the hypotheses that the parameters sum to zero and are jointly zero reject. results of the 52-week window are similar. compared to percentage 269m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 discounts, the change in trading volume appears to materialize more slowly and impact fewer funds. table 4—panel b presents the results of the volume analysis for the 21 firms with no corresponding webs index. for the 13-week post-event window 13 of the 21 (62%) funds experience decreased volumes, but only 1 volume decrease is significant at even the 5% level. also, the sum of thebj parameters is not significantly different from zero, indicating that the introduction of webs trading does not have a significant effect on the trading volumes of this group. over the 26-week window, the number of funds with lower volume remains at 13, with two of the changes significant at the 5% level. also, the null hypothesis that the sum of the volume changes equals zero cannot be rejected. similar results occur as we extend the window to 52 weeks. interestingly, the null hypothesis that thebj parameters are jointly zero easily rejects for all three event windows. this occurs because some of the volume changes are positive, some are negative, but the average change is not significant. overall, these results suggest a negligible impact on trading volume for funds with no corresponding webs. as with discounts, we next examine the differential impact on each group’s trading volume. table 5 presents the results of the two factor test on volume data. panel b contains the results of the three hypothesis tests. for the 13-week post event window the average daily trading volume for funds with a corresponding webs index decreases by 14,340 shares, which is significant at the 5% level. in contrast, funds with no corresponding webs index show on average an insignificant increase in volume of 252 shares. however, the differential 270 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 short-term volume impact is not significant. by the end of the 26-week window, the relative decrease in volume for funds with a corresponding webs index is significant, and this result persists into the 52-week window. these results corroborate a slowly emerging decrease in demand for cecfs with a corresponding webs index. this provides indirect evidence of investor migration to the lower fee webs indexes, which is consistent with an agency cost explanation of closed-end fund discounts. however, since the tests on percentage discounts show a more immediate and wide-spread impact, we cannot rule out a market segmentation effect. 5. summary this paper examines the impact of country specific index trading on closed-end country fund discounts. cecfs provide individual investors with diversified, country specific, foreign investments. however, like all closed-end funds, they often trade at a discount table 3 two factor difference in means test (dependent variable5 percentage discount) post-event window of: 13 weeks 26 weeks 52 weeks panel a: parameter estimates m 7.2%** 7.9%** 8.0%** a 2.2%** 2.9%** 3.0%** g 23.7%** 23.5%** 23.4%** u 21.2%** 21.0%** 20.9%** panel b: hypotheses test results ho (m21 2 m11) 5 0 6.8%** 7.7%** 7.6%** (m22 2 m12) 5 0 2.1%** 3.7%** 4.2%** (m21 2 m11) 2 (m22 2 m12) 5 0 4.7%** 4.0%** 3.4%** *significant at 5%; **significant at 1%. results are from a pooled regression of percentage discounts against three dummy variables. letting y be a percentage discount, the regression model is y5 m 1 ad1 1 gd2 1 qd1d2 1 «, where: d1 5 h1 if the observation is in the pre-event estimation window 21 if the observation is in the post-event window and d2 5 h1 if the observation is from a fund with a corresonding webs index 21 if the observation is from a fund with no corresponding webs index the parameterm is the average percentage discount for all funds,a measures the effect due to the time factor, g measures the effect due to fund type, andq measures the interaction effect between the two factors. let the subscript i equal one for a pre-event observation and two otherwise. let the subscript k equal one for an observation from a fund with a corresponding webs index and two otherwise. then the mean percentage discount (mik) for each level of each factor is recoverable from the regression model. 271m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 to nav, which may be a concern to individual investors. country specific indexes such as webs provide similar international diversification benefits without the concern over discounts. explanations of discounts fall into two categories: behavioral and rational economic explanations. behavioral modes suggest either no reaction or an overreaction from investors in response to the introduction of webs. rational economic models suggest table 4 modified event study results (dependent variable5 average daily volume). fund number fund name intercept b̂j for post-event window of: 13-weeks 26-weeks 52-weeks panel a: funds with corresponding webs 1 austria fund 23,821** 211,336 28,356 23,925 2 first australia fund 47,950** 212,281 215,152 27,090 3 france growth fund 29,818** 27,688 29,914 22,313 4 germany fund 47,193** 15,701 219,359* 26,464 5 italy fund 29,930** 547 7,493 9,625 6 japan equity fund 62,249** 237,641* 237,418** 223,327** 7 japan otc equity fund 52,992** 219,615 223,954 216,834 8 malaysia fund 56,566** 213,250 215,285 215,106 9 mexico equity & income fund 366,385** 278,608 2126,024 2141,585** 10 mexico fund 87,098** 548 1,321 4,207 11 new germany fund 24,407** 26,668 25,168 2622 12 singapore fund 27,986** 28,010 211,679* 21,998 13 spain fund 28,332** 26,340 210,517 22,800 14 swiss helvetia fund 34,115** 25,815 29,996* 28,015* average 65,632 213,604 220,286 215,446 % negative 79% 86% 86% x2 statistics h0: sum of parameters5 0 2.721 8.823 9.765** h0: each parameter5 0 12.500 36.912** 31.007** panel b: funds with no corresponding webs 1 argentina fund 55,172** 21,711 213,807 215,536 2 brazil fund 89,375** 220,4241 220,659 212,123 3 brazilian equity fund 44,358** 15,141 13,741 4,686 4 chile fund 37,132** 9,802 210,713 4,987 5 china fund 44,346** 4,977 211,419 210,438 6 first israel fund 37,761** 15,592 2,685 6,267 7 fidelity philippine fund 128,634** 3,412 8,647 7,825 8 greater china fund 34,954** 24,600 22,351 21,083 9 india fund 16,248** 27,740 27,356* 26,256* 10 india growth fund 9,983** 4,871 3,390 4,521* 11 irish investment fund 36,663** 22,833 28,106 24,994 12 jardine fleming china region fund 114,103** 247,334 233,811 30,006 13 korea fund 29,208** 27,203 27,038 4,125 14 korean investment fund 38,387** 26,156 28,002 24,654 15 pakistan investment fund 119,164** 71,651* 30,593 54,364* 16 roc taiwan fund 43,872** 212,432 215,834* 215,532** 17 taiwan fund 57,038** 28,838 29,188 22,644 (continued on next page) 272 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 table 5 two factor difference in means test (dependent variable5 average daily volume). post-event window of: 13 weeks 26 weeks 52 weeks panel a: parameter estimates m 55,186** 52,726** 55,760** a 23,522 25,947** 22,949* g 23,274 22,734 22,148 u 3,648 4,187 4,774** panel b: hypotheses test results ho (m21 2 m11) 5 0 214,340** 220,268** 215,446** (m22 2 m12) 5 0 252 23,520 3,650 (m21 2 m11) 2 (m22 2 m12) 5 0 214,592 216,748* 219,096** *significant at 5%; **significant at 1%. results are from a pooled regression of each week’s average daily volume against three dummy variables. letting y be a weekly observation of average daily volume, the regression model is y5 m 1 ad1 1 gd2 1 qd1d2 1 «, where: d1 5 h1 if the observation is in the pre-event estimation window 21 if the observation is in the post-event window and d2 5 h1 if the observation is from a fund with a corresonding webs index 21 if the observation is from a fund with no corresponding webs index the parameterm is the grand mean daily volume for all funds,a measures the effect due to the time factor,g measures the effect due to fund type, andq measures the interaction effect between the two factors. let the subscript i equal one for a pre-event observation and two otherwise. let the subscript k equal one for an observation from a fund with a corresponding webs index and two otherwise. then the grand mean daily volume (mik) for each level of each factor is recoverable from the regression model. table 4(continued) fund number fund name intercept b̂j for post-event window of: 13-weeks 26-weeks 52-weeks panel b: funds with no corresponding webs 18 templeton china world fund 73,343** 16,103 7,295 10,695 19 thai capital fund 36,819** 214,504 8,919 8,319 20 thai fund 24,863** 21,487 22,517 3,340 21 turkish investment fund 16,096** 2165 1,827 10,771 average 51,787 292 23,510 3,650 % negative 62% 62% 43% x2 statistics h0: sum of parameters5 0 0.055 1.017 1.722 h0: each parameter5 0 55.359** 66.137** 83.790** *significant at 5%; **significant at 1%. results are from a systems of equations regression of each week’s average daily volume against a dummy variable, d, that is constructed with ones in the post-event window. for each individual firm j, the equation is: yj 5 mj 1 bjd 1 «j, with «j ; n(0, s2). 273m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 that discounts will increase for cecfs with a corresponding webs index due to increased competition and reduced market segmentation premiums. we use a modified event study to test the impact of country specific index trading on cecf discounts. our results are consistent with a rational economic explanation of the discount. in particular, the percentage discount increases after the introduction of webs trading. the increase is stronger for cecfs with a corresponding webs index and remains permanent over a 52-week interval. funds with a corresponding webs index also experience a decrease in trading volume, further suggesting a migration away from cecfs. our results are particularly important to individual investors. country specific indexes, such as webs help investors optimize risk-return tradeoffs by lowering the costs of international diversification. competitive pressures force cecf managers to improve performance and lower agency costs, or face additional migratory pressures. nevertheless, a caution to individual investors is in order. our results suggest that cecf investors should be wary of widening discounts for funds facing the introduction of a competing country specific index. appendix a closed-end country funds by group cecfs with a corresponding webs index cecfs without a corresponding webs index fund name and symbol webs index and symbol fund name and symbol austria fund (ost) austria (ewo) argentina fund (af) first australia fund (iaf) australia (ewa) brazil fund (bzf) canadian general investors (cgi) canada (ewc) brazilian equity fund (bzl) france growth fund (frf) france (ewq) chile fund (ch) germany fund (ger) germany (ewg) china fund (chn) new germany fund (gf) germany (ewg) first israel fund (isl) italy fund (ita) italy (ewi) first philippine fund (fpf) japan equity fund (jeq) japan (ewj) greater china fund (gch) japan otc fund (jof) japan (ewj) india fund (ifn) mexico equity & income fund (mxe) mexico (eww) india growth fund (igf) emerging mexico fund (mef) mexico (eww) irish investment fund (ifl) mexico fund (mxf) mexico (eww) jardine fleming china region fund (jfc) singapore fund (sgf) singapore (ews) jardine fleming india fund (jfi) spain fund (snf) spain (ewp) korea equity fund (kef) growth fund of spain (gsp) spain (ewp) korea fund (kf) swiss helvetia fund (swz) switzerland (ewl) korean investment fund (kif) united kingdom fund (ukm) united kingdom (ewu) morgan stanley india fund (iif) new south africa fund (nsa) pakistan investment fund (pkf) roc taiwan fund (roc) taiwan fund (twn) templeton china world fund (tch) thai capital fund (tc) thai fund (ttf) turkish investment fund (tkf) 274 m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 references barclay, m., holderness, c., & pontiff, j. (1993). private benefits from block ownership and discounts on closed-end funds.journal of financial economics, 33, 263–291. bonser-neal, c., brauer, g., neal, r., & wheatley, s. (1990). international investment restrictions and closed-end country fund prices.journal of finance, 28, 523–547. brickley, j., manaster, s., & schallheim, j. (1991). the tax-timing option and the discounts on closed-end investment companies.journal of business, 64, 287–312. campbell, j. y., lo, a., & mackinlay, c. (1997). the econometrics of financial markets. princeton: princeton university press. delong, j., shleifer, a., summers, l., & waldmann, r. (1990). noise trader risk in financial markets.journal of political economy, 98, 703–738. errunza, v. (1991). pricing of national index funds.review of quantitative finance and accounting, 1, 91–100. lee, c., shleifer a., & thaler, r. (1991). investor sentiment and the closed-end fund puzzle.journal of finance, 46, 76–110. levy-yeyati, e., & ubide, a. (1998). crises, contagion and the closed-end country fund puzzle. working paper, international monetary fund. neal, r., & wheatley, s. (1998). do measures of investor sentiment predict returns?journal of financial and quantitative analysis, 33 (4), 523–547. pontiff, j. (1995). closed-end fund premiums and returns: implications for financial market equilibrium.journal of financial economics, 37, 341–367. richard, j.e., & wiggins, j. b. (2000). the information content of closed-end country fund discounts.financial services review, 9 (2), 171–182. seyhun, h., & skinner, d. (1994). how do taxes affect investors’ stock market realizations? evidence from tax-return panel data.journal of business, 67, 231–262. shipper, k., & thompson, r. (1983). the impact of merger-related regulations on shareholders of acquiring firms. journal of accounting research, 21 (1), 184–221. 275m. o’connor, e.a. downe / financial services review 9 (2000) 259–275 pii: s1057-0810(01)00075-0 the reliability of the book-to-market ratio as a risk proxy ralph r. trecartin jr.* department of business & economics, suny–brockport, 350 new campus drive, brockport, ny 14420, usa received 13 march 2000; received in revised form 9 march 2001; accepted 25 june 2001 abstract this study examines whether the book-to-market ratio consistently explains the cross-section of stock returns through time. the results reveal that the book-to-market ratio is positively and significantly related to return in only 43% of the monthly regressions. other value/growth variables such as cash flow,” “sales growth,” and “size”; perform even more erratically than the book-to-market ratio, and are thus less likely to be viewed as legitimate risk proxies. © 2001 elsevier science inc. all rights reserved. jel classification:g11; g12 keywords:risk factors; market efficiency; contrarian strategy; portfolio performance 1. introduction a particular investment strategy or style is often followed by investors forming actively managed portfolios. in recent years, the “value style” investment strategy has generated considerable interest and debate. many investment advisors and a number of academics advocate investing in “value firms” that are considered to be relatively un-popular. these firms have high book equity-to-market equity ratios (be/me), high cash flows, and low sales growth rates. value firms appear to earn much higher long-term returns than those with low be/me, low cash flows, and high sales growth rates. of potential interest to investors is whether they can use a value strategy over a short-term horizon and still expect to earn superior returns on a consistent basis, as the literature seems to indicate. * tel.: �1-716-395-5678. e-mail address:rtrecart@brockport.edu (r.r. trecartin). financial services review 9 (2000) 361–373 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00075-0 the value effect may not be consistently dependable thus causing professional portfolio managers to avoid investing in these securities. under this scenario these managers will avoid buying extreme value firms due to the risk of short-term underperformance. even for individual investors who do not have to worry about capital flight from their portfolios, a series of monthly returns with low performance may be enough to cause them to jettison this strategy. if short-term returns are not reliable, only individual investors with a long-term focus may be able to withstand short-term performance setbacks or reverses and invest consistently in the most unpopular securities. this study examines whether the book equity-to-market equity ratio and other value/ growth variables predict returns consistently from 1963 to 1997 using monthly intervals. can the individual investor using a value investment strategy expect at any point in time to outperform a growth strategy over subsequent months? average returns are reported over long intervals as in other studies. subperiods are then examined over ten-year, five-year, and one-year periods. the study documents the dependability of returns, or the lack thereof, for value firms, and also indicates whether be/me or some competing variable captures the most variation in return. the paper is organized as follows: in section 2 a review of the background literature is undertaken. data specifications and issues are presented in section 3. empirical test results for competing univariate variables are displayed in section 4. subperiod results for five and ten-year periods are analyzed in section 5. monthly regression coefficients averaged on a yearly basis are displayed in section 6. a summary of the findings may be found in section 7. 2. literature foundations fama and french (1992, 1993, 1995) have contributed to the popularity of the be/me variable for predicting stock returns. the strength of this variable allows it to act as the central factor in their asset-pricing model. other authors have also found a significant positive correlation between be/me ratios and the cross-section of stock returns including stattman (1980), rosenberg, reid, and lanstein (1985), chan, hamao, and lakonishok (1991, 1993), capaul, rowley, and sharpe (1993), lakonishok, shleifer, and vishny (1994), davis (1994), and chan, karceski, and lakonishok (1998). fama and french (1992) interpret be/me as a proxy for risk. fama and french (1995) provide evidence, which indicates that high be/me firms have a degree of financial distress via depressed earnings. their perspective is that high be/me firms have both high risk and high return levels. thus, the market appears to behave efficiently. lakonishok et al. (1994), and haugen and baker (1996), provide a competing explanation and suggest that the relationship is caused by inefficient markets and investor overreaction. their position is that investors bid “growth” fi rm prices up too far, causing high market values and extremely low be/me ratios but they bid the price of “value” fi rms down too low causing high be/me ratios. be/me, cash flow, and sales growth variables are considered alternative variables for categorizing value/growth firms by lakonishok et al. (1994). high be/me, high cash flow, and low sales growth firms are labeled as “value” fi rms. they find that cash flow yield and sales growth 362 r.r. trecartin / financial services review 9 (2000) 361–373 variables capture larger average premiums than be/me and size in one-to-five year holding periods. davis (1994) on the other hand does not find a significant sales growth variable. some recent research tends not to support the risk proxy theory, although it is unresolved as to whether the premium captured by be/me is due to an unknown source of risk or due to market inefficiencies. daniel and titman (1997) find no evidence for a distress factor. la porta, lakonishok, shleifer, and vishny (1997) find that much of the difference in returns between value and growth firms can be attributed to relatively larger positive earning surprises for value stocks. this finding tends to support the investor overreaction theory rather than a risk based explanation. dechow and sloan (1997) counter the overreaction literature by showing that prices are affected by misplaced investor belief in biased analyst forecasts. whether these biased forecasts contribute to a risk based explanation is yet to be determined. loughran (1997) and loughran and ritter (2000) call into question the usefulness of the be/me effect altogether. loughran (1997) examines the book-to-market effect in detail by firm size, exchange listing, and calendar seasonality and concludes that the fama and french findings are driven by the january effect and low returns on small, young, growth stocks. he finds that book-to-market has no explanatory effect in the largest size quintile, which has 73% of the market capitalization. he assumes that the effect is not useful for professional investors who must take into consideration other factors such as liquidity when forming portfolios and investment strategies. he uses this argument to explain why value fund and growth fund manager performance is so comparable as in malkiel (1995). lakonishok, shleifer, and vishney (1994), and haugen (1995) give a different rationale as to why fund managers may avoid investing in value firms on the extreme end of the continuum. they reason that professional money managers cannot risk having a portfolio that substantially underperforms the market even on a short-term basis because performance is measured and rewarded monthly or quarterly. haugen (1995) claims that the value effect is a “golden opportunity” for individual investors who can continue to earn above normal long-term returns. he suggests that the effect should continue to persist because of the potential short-term uncertainty in returns which drives away the professional managers. dhatt, kim, and mukherji (1999) present findings that partly disagree with loughran’s (1997) premise. they show that an exploitable value premium exists for stocks in the russell 2000 index in general, and for liquid stocks other than the smallest size quintile in particular. furthermore they document a value premium outside of january. this finding is consistent with the findings of loughran (1997) for nasdaq and amex stocks but not for nyse firms. chan et al. (1998) document strong size, be/me, and dividend yield comovements in stock returns. in addition to the most widely used fundamental factors, they also examine a number of competing macroeconomic, statistical, and technical factors as well as an overall market factor. they do not examine the sales growth variable among their other factors. chan et al. (1998) also point out that comovement in returns is not the same as a priced factor in returns. for example, their strongest fundamental factor, the size variable, tends to load on large firms part of the time and then switches over to small firms. this causes the size factor to provide the highest variation among fundamental factors over time. their findings do not directly contribute to a further understanding of whether these factors are risk based or 363r.r. trecartin / financial services review 9 (2000) 361–373 caused by behavioral aberrations, but they do provide information on which variables generate the largest variation in returns. in the following report, an examination of the short-term nature of the value effect provides evidence concerning the usefulness of the strategy, and also contributes new information to the risk-proxy versus overreaction debate. in constructing the study three issues are taken into account. first, can the long term returns documented in the above studies, be consistently captured on a short-term basis? if so, then money managers would be expected to participate if loughran’s (1997) liquidity theory is not applicable. if not, then this may explain why managers avoid these securities. second, does the be/me variable do the best job of predicting return, or are there better alternative value/growth variables? and third, when the results are known do they support the risk proxy theory or the investor overreaction explanation? in an efficient market one would expect that risk and return would be highly related on a reliable basis through time. any useful risk proxy will be expected to explain variation in return on a consistent basis. if instead the relationship between the risk proxy and return is not reliable through time, it can be argued that the superior returns are generated during some time periods because of investor overreaction or by chance rather than because of risk. or, one may argue that markets are efficient only part of the time. 3. data and empirical design all firms on nyse, amex, and nasdaq are included in the study if they meet other data requirements described below. in this study the emphasis is on short-term performance of the value effect, and its reliability through time. stock return is used as the dependent variable in monthly regressions as in fama and french (1992). the independent variables are drawn from the pool of value/growth variables discussed in the literature. the composition of these variables is discussed below. seven additional years of data are included beyond the years examined by fama and french (1992) providing return data from july 1963 to december 1997. accounting variables for use in this study are taken from fiscal year-end in calendar year t-1. for example, the accounting data for book equity is gathered from the compustat active and research files starting with december 1962 and ending with december 1996. market value of equity is taken at the end of june in year t and comes from the crsp files. the accounting data are taken in year t-1 to explain returns for each month starting in july of year t to june of year t�1. daily return data are compounded to form monthly returns as in loughran (1997). a company must have a crsp stock price for december of year t-1 and june of year t to be included in the study. the company must have monthly returns for at least 24 of the 60 months prior to july of year t. each company must have book equity for its fiscal year (ending in any month) of calendar year t-1. these restrictions, imposed by fama and french, are followed in this study. an additional restriction is added for convenience in comparing studies. a firm must have sales in at least two adjacent years during the five years preceding year t in order to calculate sales growth rates. the firm must also record earnings. 364 r.r. trecartin / financial services review 9 (2000) 361–373 listed below in table 1 are variable descriptions and brief explanations as needed. one major dilemma deals with firms that exhibit negative cash flows. for most of this study, negative cash flow firms are assumed to be most similar to growth firms and assigned a rank of 1 in the cash flow decile variable described in table 1. negative cash flow firms are included so as to maintain as large a data set as possible. including negative cash flow firms in the data set creates the strongest and most significant results for the be/me variable. alternate regressions were also run using only positive cash flow firms and a negative cash flow dummy. another test was conducted assuming negative cash flow firms to be most like “value firms.” negative book equity firms are not included in the regressions. so, negative cash flow firms included in the data set are more apt to be temporarily depressed, or very rapid growth firms using more resources than are generated internally. another data issue involves the form of the sales growth variable that is most effective. anthony and ramesh (1992) use a median sales growth (mgs) variable while lakonishok et al. (1994) use both equal weighting sales growth and weighted sales growth procedures. a weighted sales growth variable is primarily used in this study (see table 1) but other forms of the variable are also tested. 4. empirical results for competing univariate variables the focus of this section is on statistically significant “value/growth” variables in monthly regressions, and whether these variables are reliable or dependable through time. table 2 summarizes monthly regression results for single variable models including book-to-market, table 1 variable descriptions variable name variable description ln(be/me) book equity (be) is compustat data item # 60. this includes common stock outstanding, capital surplus, and retained earnings. market equity (me) is reported on the last trade date in december and consists of compustat item #24* #25 (shares outstanding * price) cash flow earnings before extraordinary items (data item # 18) plus depreciation (item # 14) all divided by market value. cash flow deciles or the raw cash flows above are sorted into deciles and assigned a 1.0 for (cfdecile) the lowest cash flow firms, a 2.0 for the next lowest decile, and so on up to a 10 for the highest cash flow firms. wgs weighted growth in sales over the last five years if available. (30% weighting to the most recent year, 25% for the next most recent year down to 10% for the fifth year). if there are less than five years the weights are adjusted to add up to 100%. wgs deciles wgs sorted into decile portfolios and assigned a rank from 1 to 10 from lowest to highest growth. ln(wgs) the natural log of wgs deciles. ln(me) the natural log of june ending market value denominated in millions of dollars. market equity (me) for the “size” variable is taken from crsp. the most recent reported shares outstanding are multiplied with the price reported on the last trade date of june in year t. 365r.r. trecartin / financial services review 9 (2000) 361–373 size, sales growth, and cash flow. the coefficient mean (or average slope) for each variable is the time series average of the regression coefficients taken from 414 monthly regressions that start in july 1963 and proceed to december 1997. these coefficients vary over the 34.5-year period but on an average explain a statistically significant portion of the variation in return. all the variables are significant and have signs that correspond with the results from earlier studies. ln(be/me) and cash flow variables are positively related to return while size and sales growth variables are negatively correlated with return. the strongest variables are ln(be/me) and cash flow deciles. the next strongest variables are the natural log of weighted sales growth, and size. of all the variables analyzed in a univariate fashion, ln(be/me) has the highest significance level, but only fractionally higher than cash flow deciles. examination of each individual monthly regression, reveals that ln(be/me) has more significant monthly regressions of the appropriate sign than any other variable in the study. out of a universe of 414 monthly regressions, 43.0% are significantly positive while only 18.4% work in the opposite direction. thirty-nine percentage (38.6%) of the monthly regression coefficients cannot be established as different from zero. the cash flow decile variable has fewer statistically positive monthly regression coefficients than ln(be/me), and about the same number of months where the results move in the wrong direction (40.6% positive, 18.6% negative, and 40.8% inconclusive). for ln(wgs) 30.4% have significant negative monthly coefficients, while 15.5% of the months perform contrary to expectations. over half the monthly regressions coefficients (54.1%) are insignificant. for ln(me) similar percentages are 40.3% significantly negative, 31.6% significantly positive, and 28% inconclusive. on the univariate level the results indicate that ln(be/me) would probably be the most effective predictor of return over time due to its more consistent relationship with return. in this context there is no one unique definition of reliability. what is meant by reliability for a variable, is that it is statistically significant with the proper sign for the period of table 2 average slopes from month-by-month regressions of stock return on variables of interest july 1963 to december 1997 variable coefficient mean t statistic p value number sign pos months numbers sign neg months ln(be/me) 0.48% 6.07 0.000* 178 76 cashflow 0.11% 5.59 0.000* 168 77 deciles ln(wgs) �0.21% �3.80 0.000* 64 126 ln(me) �0.14% �3.11 0.002* 131 167 each variable is formed as described in table 1. the coefficient mean (or average slope) is the average regression coefficient taken from 414 monthly regressions that start in july 1963 and proceed to december 1997. the t-statistics and p values are taken from single sample t-tests in which the time series of the regression coefficient is tested for the hypothesis that the mean is not different than zero. asterisks found next to individual p values highlight significance levels at 5% or lower. in addition, the number of significant positive and negative (5% level) monthly regression coefficients are recorded. 366 r.r. trecartin / financial services review 9 (2000) 361–373 examination, and that it correctly predicts returns in more than 50% of the time periods. the position taken in this paper holds that a strategy is not considered reliable if it captures variation in return less than 50% of the time. clearly, if one accepts this definition of reliability, the value effect as captured by be/me is unreliable. when examining the value effect using alternative risk proxies the level of reliability will have some impact on the explanation as to whether the effect is caused by investor overreaction or is based on risk. risk can change through time, but in the long tradition of the capital asset pricing model, it may be assumed that a risk proxy should consistently differentiate differences in risk to the exclusion of other nonrisk based factors. if on the other hand, investor overreaction or fad investing is driving returns, it may well be that a particular strategy’s performance is cyclical through time. since a risk proxy benchmark for reliability has not been established in the literature, the best that can be done is to present statistics that are useful for comparison. each individual investor, depending upon his or her level of risk aversion, can determine whether the strategy warrants usage. each individual will determine a minimum level of reliability and choose strategies accordingly. 5. subperiod results time consistency an analysis of subperiod results for each model is useful in answering the following questions. do the large average returns garnered from contrarian or risk proxy strategies derive from consistent common stock return behavior? or, are the results due to a few exceptional months or years? table 3 casts light on the issue by dividing the data into ten-year subperiods. in table 3 the first 360 months of the study period is divided into three 120-month intervals. the last two columns present the most recent 20–1/2 years providing a different time period view. this view avoids the possibility that inclusion of nasdaq firms from the early 70’s is somehow altering the results. some overlap in time periods occurs due to this added view. with a ten-year horizon each of the models presented exhibit statistically significant results as expected for the majority of the study period except for the size variable. ln(me) is significant in only two time periods examined, and these two periods have some overlap between 1977 and 1983. the sales growth variable is not significant in the earliest time period between july 1963 and june 1973. during one of the ten-year periods, the be/me variable exceeds the 50% reliability rule that is introduced in the last section. between july 1977 and june 1987 be/me is significantly related to return in 53% of the monthly regressions. other time periods and other variables do not accomplish this distinction except for the size variable in the period of july 1973 to june 1983. it appears that the value effect is an unreliable predictor of return even though it is significantly related to return on average over 10 year periods. ln(be/me) is a significant variable in each of the ten-year periods, though over half the monthly regression coefficients are not positive for three of the time periods examined. 367r.r. trecartin / financial services review 9 (2000) 361–373 likewise the cash flow decile variable is significant in most of the ten-year periods, though the percentage of positive coefficients never exceeds 48%. the results found in subperiod analysis reveals that patience is a key attribute when any of these strategies are used as an investment rule. over longer-time horizons such as the tenyear periods shown in table 3, average premiums on ln(be/me) are statistically significant and positive. this is not the case for certain shorter horizons as in the five-year subperiods shown in table 4. wide variation in the strength of each model is evident. the coefficient mean in the first five years is ten times larger than the following five years. the coefficient mean in the third five-year period is over 13 times as large as the second period and over three times as large as the fourth. it is evident that an investor following a value strategy, will have to weather many months of negative or insignificant return performance to capture a large positive average value premium. table 3 average slopes from month-by-month regressions of stock return on variables of interest: ten year sub-period results july 1963 to june 1973 july 1973 to june 1983 july 1983 to june 1993 july 1977 to june 1987 july 1987 to dec 1997 ln(be/me) coef. mean 0.34% 0.52% 0.54% 0.49% 0.48% t statistic 2.29 2.69 5.44 3.86 5.28 p value 0.024* 0.008* 0.000* 0.000* 0.000* % sign pos 31% 46% 50% 53% 45% % sign neg 19% 25% 13% 22% 13% cfdecile coef. mean 0.10% 0.09% 0.12% 0.13% 0.07% t statistic 3.26 2.48 3.59 4.34 1.97 p value 0.001* 0.015* 0.000* 0.000* 0.052 % sign pos 31% 41% 48% 48% 43% % sign neg 15% 17% 18% 18% 26% ln(wgs) coef. mean �0.01% �0.36% �0.24% �0.28% �0.21% t statistic �0.10 �3.28 �2.36 �2.98 �2.19 p value 0.918 0.001* 0.020* 0.003* 0.030* % sign pos 12% 20% 18% 19% 17% % sign neg 18% 38% 35% 38% 32% ln(me) coef. mean �0.12% �0.32% �0.01% �0.14% �0.10% t statistic �1.43 �3.39 �0.10 �2.04 �1.31 p value 0.157 0.001* 0.923 0.043* 0.192 % sign pos 27% 28% 37% 33% 35% % sign neg 29% 53% 38% 46% 45% each variable is formed as described in table 1. the coefficient mean (or average slope) is the average regression coefficient taken from the 120 monthly regressions for each ten-year period. the t-statistics and p values are taken from single sample t-tests in which the time series of the regression coefficient is tested for the hypothesis that the mean is not different than zero. in addition, the percentage of significant positive and negative (5% level) monthly regression coefficients are recorded. 368 r.r. trecartin / financial services review 9 (2000) 361–373 in five-year periods the findings are ambivalent. results indicate that even the powerful be/me variable is reliably related to return for more than 50% of the months, during only one five-year period. other value/growth variables are also not reliably related to return. firm market value for example, is not a good univariate predictor of return for many time periods. it may be that size is randomly unrelated to the other investment strategies and thus provides a degree of diversification benefit when included in multivariate models. or it may be that the results are simply time period specific. if this is the case, then there is no assurance that size will continue to perform well in the future. the ideal risk proxy would consistently be related to return in each five-year (or more frequent) period. cash flow deciles and be/me appear to follow the same pattern in returns to some extant. both variables are significant during the same five-year time spans, except table 4 average slopes from month-by-month regressions of stock returns on variables of interest—five year subperiods 7/63 to 6/68 7/68 to 6/73 7/73 to 6/78 7/78 to 6/83 7/83 to 6/88 7/88 to 6/93 7/93 to 12/97 ln(be/me) coef. mean 0.62% 0.06% 0.79% 0.26% 0.67% 0.40% 0.54% t statistic 2.71 0.32 2.42 1.23 4.71 2.97 4.02 p value 0.009* 0.749 0.019* 0.225 0.000* 0.004* 0.000* % sign pos 33% 28% 47% 45% 60% 40% 48% % sign neg 12% 27% 18% 32% 12% 15% 13% cfdecile coef. mean 0.15% 0.06% 0.13% 0.06% 0.23% 0.02% 0.10% t statistic 3.39 1.28 2.14 1.28 5.42 0.37 1.68 p value 0.001* 0.206 0.036* 0.207 0.000* 0.713 0.099 % sign pos 35% 27% 42% 40% 58% 38% 44% % sign neg 7% 23% 10% 23% 12% 25% 31% ln(wgs) coef. mean 0.11% �0.13% �0.53% �0.19% �0.33% �0.14% �0.27% t statistic 0.68 �0.91 �3.49 �1.21 �2.59 �0.92 �2.46 p value 0.499 0.367 0.001* 0.232 0.012* 0.364 0.017* % sign pos 8% 12% 13% 27% 15% 22% 11% % sign neg 12% 23% 40% 37% 40% 30% 31% ln(me) coef. mean �0.39% 0.14% �0.38% �0.26% 0.13% �0.15% �0.07% t statistic �3.29 1.25 �2.38 �2.57 1.60 �1.27 �0.71 p value 0.002* 0.216 0.021* 0.013* 0.115 0.210 0.480 % sign pos 12% 42% 25% 30% 45% 28% 41% % sign neg 40% 18% 53% 52% 28% 48% 43% each variable is formed as described in table 1. the coefficient mean (or average slope) is the average regression coefficient taken from the 60 monthly regressions for each five-year period. the t-statistics and p values are taken from single sample t-tests in which the time series of the regression coefficient is tested for the hypothesis that the mean is not different than zero. in addition, the percentage of significant positive and negative (5% level) monthly regression coefficients are recorded. 369r.r. trecartin / financial services review 9 (2000) 361–373 for the last two time periods from july 1988 to december 1997. during these periods cash flow deciles are not significantly related to return, while the be/me variable continues to differentiate between high and low return investments. the sales growth variable is quite unreliable in short time spans but, like size, this variable may add diversification benefits in multivariate strategies. 6. monthly regression coefficients averaged on a yearly basis individual investors and portfolio managers can never be sure that past patterns in returns will continue in the future. if the value firm effect is based on underlying risk factors, one should expect it to continue. on the other hand if the effect is based upon investor overreaction, than one may expect the effect to phase in and out with investing fads or disappear altogether. from the information presented in table 5 it is easy to identify when the variable in question has had periods of failure. failure entails the average coefficient falling into negative territory for ln(be/me) and cash flow deciles. this indicates that growth firms outperformed value firms during this period, contrary to what is expected. violation of the value firm effect is evident for ln(wgs) and ln(me) when the average coefficients rise into positive territory. ln(be/me) and cash flow deciles appear to be fairly strong variables over some periods of time on an average basis. ln(be/me) is more reliable than any of the other variables with only four years in which the coefficient drops into negative territory. cash flow deciles become negative in eight years. ln(wgs) is positive in eight out of thirty-five years, with several more years very close to zero. clearly ln(wgs) performs opposite to contrarian strategy predictions during the period before that in which lakonishok et al. (1994) analyze their results. the variable has performed as expected most of the time in recent years. ln(me) has positive coefficients in over one-third of the years. ln(me) does not perform well during the mid-eighties in particular. table 5 indicates that there are not many years in which the contrarian be/me strategy dramatically fails (growth firms earn more on average than value firms). for ln(be/me) this occurs between july 1970 and june 1972, july 1979 to june 1980, and july 1989 to june 1990. there are many years where the strategy does not appear to differentiate well between the two styles of investing. during the entire 34.5-year period there are only 16 years (46%) in which 50% or more of the monthly be/me coefficients are positive and significant. 7. summary and interpretation of findings evaluation of a value investment strategy reveals that the high returns found over long time horizons are not uniform or dependable over short time intervals. in this study the book equity-to-market equity ratio (be/me) is regressed monthly against returns for the years of 1963–1997. the book-to-market effect (be/me) is statistically related to return as predicted in less than 50% of the monthly time periods examined. also, the variable is not always significant 370 r.r. trecartin / financial services review 9 (2000) 361–373 in five-year subperiods. however in ten-year periods be/me is significantly related to return. thus the data supports the view that the be/me variable is not a reliable predictor of return over short time horizons. an investor can capture superior returns only if the holding period is extended to cover fairly long intervals. it is the author’s opinion that short term be/me unreliability does not negate the usefulness of the value effect for a patient investor as evidenced by the long term positive and statistically significant coefficients presented in this and other studies. but, there is no table 5 monthly regression coefficients averaged on a yearly basis year ln(be/me) pos/neg cfdecile pos/neg ln(wgs) pos/neg ln(me) pos/neg 1962 0.52% 4,0 0.08% 2,1 �0.42% 0,5 0.11% 3,2 1963 0.17 3,2 0.06 5,1 �0.30 0,1 �0.25 1,5 1964 0.58 6,3 0.27 8,1 0.06 0,0 �0.46 1,7 1965 0.97 4,1 0.17 2,0 0.34 1,0 �0.57 2,4 1966 0.87 3,1 0.17 4,1 0.88 4,1 �0.78 0,6 1967 0.40 5,1 0.12 4,0 �0.34 0,1 0.02 2,2 1968 0.08 4,4 0.09 3,4 �0.41 2,5 0.32 6,2 1969 �0.49 2,4 �0.12 1,5 0.75 4,0 �0.20 4,5 1970 �0.38 1,6 �0.11 0,4 �0.01 1,1 �0.04 5,2 1971 0.68 5,1 0.32 8,1 �0.66 0,7 0.63 8,0 1972 0.46 4,3 0.03 4,1 �0.77 2,6 �0.15 4,4 1973 0.65 6,4 0.03 3,4 �0.55 1,5 �0.35 4,6 1974 1.13 3,3 0.26 6,0 �0.60 1,5 �0.34 5,3 1975 0.84 8,0 0.20 6,0 �0.25 3,4 �0.35 2,7 1976 0.85 7,1 0.11 6,1 �0.49 1,5 �0.70 0,12 1977 0.13 6,4 0.04 3,1 �0.15 4,4 �0.24 4,6 1978 �0.45 1,4 �0.06 2,4 0.30 4,1 �0.05 5,3 1979 0.22 6,5 0.04 6,3 0.11 4,3 �0.39 3,8 1980 1.07 9,2 0.34 10,2 �0.55 2,7 �0.05 4,4 1981 0.31 5,4 �0.07 3,4 �0.66 2,7 �0.57 2,10 1982 1.31 10,1 0.36 9,0 �0.71 1,7 0.12 6,3 1983 0.59 8,2 0.36 9,2 �0.01 2,3 0.34 7,2 1984 0.27 2,0 0.10 4,1 �0.12 1,2 0.15 5,2 1985 0.62 9,3 0.12 6,3 �0.56 2,7 0.00 4,5 1986 0.56 7,1 0.22 7,1 �0.25 3,5 0.05 5,0 1987 0.55 6,1 0.19 7,1 �0.20 1,4 0.08 5,3 1988 �0.21 1,2 0.03 4,4 0.02 2,2 0.11 5,4 1989 0.27 5,3 �0.06 5,4 0.05 5,4 �0.12 3,8 1990 0.64 6,3 �0.14 3,3 �0.32 2,5 �0.47 3,7 1991 0.75 6,0 0.07 4,3 �0.26 3,3 �0.34 1,7 1992 0.86 8,0 0.13 7,2 �0.40 1,3 �0.24 3,7 1993 0.13 5,2 �0.02 3,4 �0.10 1,1 0.00 5,3 1994 0.11 2,2 �0.14 3,7 �0.18 1,5 �0.37 3,8 1995 0.89 8,3 0.35 8,2 �0.24 3,5 0.27 8,2 1996 0.87* 3,0* 0.24* 3,2* �0.62* 0,3* 0.05* 3,3* the years represent the “variable” formation period from which accounting data is drawn. the coefficient means represent the average monthly regression coefficient of the variable regressed against stock return, during the period july of year t � 1 to june of year t � 2. for example, the average negative regression coefficient on ln(be/me) for 1978 is for the 12-month return period july 1979 to june 1980. * 1996 results are based on returns from july 1997 to december 1997 only. pos/neg records the number of statistically significant positive and negative coefficients respectively. 371r.r. trecartin / financial services review 9 (2000) 361–373 certainty that the historical data will predict trends that will persist into the future. the professional investment community may be aware of the “value” effect, but be unwilling to risk the possible short-term underperformance resident in such a strategy. they are more likely to invest in securities more closely aligned with common performance measurements such as the s&p 500 index. in this study three questions were examined. first, can the long term returns documented in the literature, be consistently captured on a short-term basis? the answer is no, not on a reliable basis through time. second, does the be/me variable do the best job of predicting return, or are there better alternative value/growth variables? although be/me is weak at times, and is positive and statistically significant in only 43% of the monthly regressions, the be/me ratio is a more consistent predictor of return than other competing value/growth variables such as cash flow, size, and sales growth. and third, when the results are known do they support the risk proxy theory or the investor overreaction explanation? because the be/me effect is not reliable over short horizons an argument can be made that either the market is not efficient, or that the be/me variable is not an adequate proxy for risk. one would expect a useful risk proxy to be related to return on a reliable basis if markets are efficient. perhaps there is a consistent relationship between true underlying risk and return through time. as a proxy for risk the be/me variable does not adequately predict return on a consistent basis, and so the results of this study do not provide support for the risk proxy theory. rather, it is plausible that some investor overreaction is behind the positive but variable returns derived from the value effect. investing fads (value style or growth style) can be expected to come in and out of favor with investors. an investor overreaction story would help explain why the effect is stronger during some time periods than others. references anthony, j. h., & ramesh, k. (1992). association between accounting performance measures and stock prices. journal of accounting and economics, 15, 203–227. capaul, c., rowley, i., & sharpe, w. f. (1993). international value and growth stock returns. financial analysts journal, (january/february), 27–36. chan, l. k., hamao, y., & lakonishok, j. (1991). fundamentals and stock returns in japan. the journal of finance, 46, 1739–1789. chan, l. k., hamao, y., & lakonishok, j. (1993). can fundamentals predict japanese stock returns? financial analysts journal, (july-august), 63–69. chan, l. k., karceski, j., & lakonishok, j. (1998). the risk and return from factors. journal of financial and quantitative analysis, 33(2), 159–188. daniel, k., & titman, s. (1997). evidence on the characteristics of cross sectional variation in stock returns. the journal of finance, 52(1), 1–33. davis, j. l. (1994). the cross-section of realized stock returns: the pre-compustat evidence. the journal of finance, 49, 1579–1593. dechow, p. m., & sloan, r. g. (1997). returns to contrarian investment strategies: tests of naı̈ve expectations hypotheses. journal of financial economics, 43, 3–27. dhatt, m. s., kim, y. h., & mukherji, s. (1999). the value premium for small-capitalization stocks. financial analysts journal, 55(5), 60–68. fama, e. f., & french, k. r. (1992). the cross-section of expected stock returns. the journal of finance, 47, 427–465. 372 r.r. trecartin / financial services review 9 (2000) 361–373 fama, e. f., & french, k. r. (1993). common risk factors in the returns on stocks and bonds. journal of financial economics, 33, 3–56. fama, e. f., & french, k. r. (1995). size and book-to-market factors in earnings and returns. the journal of finance, 50, 131–155. haugen, r. a., & baker, n. l. (1996). commonality in the determinants of expected stock returns. journal of financial economics, 41, 401–439. haugen, r. a. (1995). the new finance: the case against efficient markets. englewood cliffs, new jersey: prentice-hall. lakonishok, j., shleifer, a., & vishny, r. w. (1994). contrarian investment extrapolation, and risk. the journal of finance, 49, 1541–1578. la porta, r., lakonishok, j., shleifer, a., & vishny, r. (1997). good news for value stocks: further evidence on market efficiency. the journal of finance, 52(2), 859–874. loughran, t. (1997). book-to-market across firm size, exchange, and seasonality: is there and effect? journal of financial and quantitative analysis, 32(3), 249–268. loughran, t., & ritter, j. (2000). uniformly least powerful tests of market efficiency. journal of financial economics, 55, 361–389. malkiel, b. g. (1995). returns from investing in equity mutual funds 1971 to 1991. journal of finance, 50, 549–572. rosenberg, b., reid, k., & lanstein, r. (1985). persuasive evidence of market inefficiency. journal of portfolio management, 11, 9–17. stattman, d. (1980). book values and stock returns. the chicago mba: a journal of selected papers, 4, 25–45. 373r.r. trecartin / financial services review 9 (2000) 361–373 financial services review, 33(1) 142 assessing the impact of rebalancing on equal-weighted and value-weighted portfolios over five decades rama malladi1 and alexander stanoyevitch2 abstract this study investigates the impact of transaction costs on the performance differential between equal-weighted portfolios (ewps) and value-weighted portfolios (vwps). employing a comprehensive dataset of 181 stocks from 1970 to 2023, we utilize paired two-sample tests to identify statistically significant differences in turnover and riskadjusted returns. our findings reveal a substantial performance advantage for ewps, with annualized return surpluses ranging from 115 to 188 basis points over vwps, depending on the assumed transaction cost level. notably, this outperformance persists until transaction costs reach a critical threshold of 728 basis points of portfolio turnover. the analysis further demonstrates that ewps outperform vwps in 94.5% of scenarios devoid of transaction costs, declining to 84% when incorporating realistic cost assumptions. these results highlight the potential of ewps to exploit diversification benefits but also emphasize the crucial role of transaction costs in moderating their outperformance. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation malladi, r., & stanoyevitch, a. assessing the impact of rebalancing on equal-weighted and valued-weighted portfolios over five decades. financial services review, 33(1), 142-164. introduction since the inception of the s&p 500 in 1957, market capitalization weighting has stood as the dominant indexing approach, steadfastly maintaining its influence over indices. the 1970s witnessed the emergence of index mutual funds, followed by the introduction of index etfs in the 1990s. as of the close of 2022, the combined net assets in these categories had skyrocketed to $10.9 trillion. notably, index mutual funds and index etfs jointly constituted 46 percent of assets in long-term funds, a substantial leap from 1 corresponding author (rmalladi@csudh.edu). california state university, dominguez hills, carson, ca, usa. 2 california state university, dominguez hills, carson, ca, usa. the 22 percent recorded in 2012 (investment company institute, 2023). aligned with the capital asset pricing model (capm) proposed by sharpe (1964) and lintner (1969), the market-cap weighted approach—also known as the value-weight (vw) approach—has been a consistent force. drawing on the foundational principles of markowitz (1952), sharpe (1964), and others, the finance sector has translated these insights into trillions of dollars allocated to or measured against market indexes https://creativecommons.org/licenses/by-nc/4.0/ mailto:rmalladi@csudh.edu https://creativecommons.org/licenses/by-nc/4.0/ malladi & stanoyevitch 143 such as the s&p 500 or russell 1000 (r. d. arnott et al., 2005). the efficient market hypothesis (emh) by fama (1970) posits that a vw index signifies an optimal mean-variance efficient investment. however, debates on market efficiency, notably fueled by roll's critique (1977), persist. various alternatives to vw indices, including fundamental indexing (r. d. arnott et al., 2005), smart beta (amenc et al., 2011; amenc & goltz, 2013; amenc & sourd, 2005), and equalweighted (ew) indices (demiguel et al., 2009; malladi & fabozzi, 2017; plyakha et al., 2012), have been proposed. fundamental indexing incorporates factors like earnings to capture robust financial fundamentals, while smart beta utilizes alternative factor weighting methods like volatility to enhance returns (amenc et al., 2016). the primary objective of this paper is to conduct a performance comparison between equalweighted portfolios (ewp) and value-weighted portfolios (vwp), with a particular emphasis on the influence of transaction costs. the intentional concentration on an equal-weighted index in this study is designed to ensure clarity and specificity, facilitating a more thorough examination of the effects of equal weighting without introducing the complexities associated with smart beta or fundamental indices. theoretically, ewp can outperform a given vw index using the same index constituents. this allows us to create an ewp from a passive investable vw index and surpass its return. the findings of this study will be of interest to both investors and academics. investors will benefit from a better understanding of the risks and returns of ewps, as well as the factors that influence their performance. academics will benefit from a more rigorous and comprehensive empirical analysis of ewps. the rest of the paper is organized as follows: the next section reviews the current literature. the section on data explains how 181 stock data from 1970 to 2023 are used in this study. the methodology section describes the exhaustive random sampling method deployed in this paper, followed by the results and discussion. the final section presents conclusions and identifies the scope for further research. the literature of equal-weighted portfolios academic studies (demiguel et al., 2009; malladi & fabozzi, 2017; plyakha et al., 2012, 2021) suggest that portfolios employing equalweighting (or 1/n portfolios) exhibit superior performance compared to other portfolio strategies. the benefits of an equal-weighted portfolio (ewp) are well documented. none of the 14 models assessed in demiguel et al. (2009) across seven empirical datasets consistently outperforms the ewp in sharpe ratio, certainty-equivalent return, or turnover. malladi and fabozzi (2017) demonstrated, using theory, simulation, and realworld data from 1926 to 2014, that ewp outperforms the value-weighted portfolio (vwp). moreover, in a two-stock, two-period setting, they demonstrate that a significant portion of the excess return is attributable to portfolio rebalancing. importantly, they illustrate that, despite higher turnover costs associated with equal weighting, the excess returns surpass the increased expenses, making equal weighting economically justified. plyakha et al. (2021) find that, despite accounting for a fifty basis point transaction cost, the monthly-rebalanced ewp surpasses vwp in terms of total mean return and oneand four-factor alphas. three historical explanations in academic literature account for the distinctions between ewp and vwp: the noisy market hypothesis, illiquidity, and autocorrelation (pae & sabbaghi, 2015). the noisy market hypothesis of arnott (2006) posits that market errors or value tilting lead to inflated market capitalizations of overvalued stocks, resulting in lower expected returns for large-cap stocks and higher expected returns for undervalued small-cap stocks. as a result, the vwp is sub-optimal (hsu, 2006). the elevated illiquidity premium linked to small firms can lead to an ewp demonstrating superior returns and increased volatility compared to a vwp (amihud, 2002; malladi & fabozzi, 2017). the heightened autocorrelation of an ewp could result in a higher return compared to a vwp (atchison et al., 1987). two recent additions to the three prior explanations include size and the rebalancing financial services review, 33(1) 144 effect. plyakha et al. (2012, 2021) reported that 58% of ewp's total excess mean return over vwp is attributable to the systematic component, compensating for exposure to smaller stocks as anticipated. however, 42% is derived from the difference in alphas, primarily influenced by the rebalancing effect (i.e., monthly rebalancing to maintain constant weights in the ewp). additionally, malladi and fabozzi (2017) found that 85% of ewp's excess returns can be attributed to the rebalancing effect. divergences exist between these two studies regarding the analysis timeframe and transaction costs. the former investigated up to 300 stocks between 1967 and 2009, accounting for 50 basis points (bps) in transaction costs. conversely, the latter study examined 500 stocks from 1926 to 2014, factoring in a transaction cost of 169 bps. swade et al. (2023) show that ewp consistently surpasses vwp, regardless of the rebalancing frequency spanning from 1 month to 60 months. notably, the highest level of outperformance occurs with monthly rebalancing. while ewp generally outperforms vwp in the long term, there are instances, particularly in short-term periods, where ewp may underperform vwp (taljaard & maré, 2021). the existing literature on ewp broadly suggests that they generally outperform vwp in the long term. however, some important questions have not yet been fully answered, such as the statistical significance of excess returns in the presence of transaction costs, the long-term consistency of excess returns, the impact of turnover on excess returns, and the potential for additional returns by optimizing efficiency to minimize transaction costs. in this study, we aim to address these unanswered research questions: • statistical significance of excess returns: do excess returns demonstrate statistical significance in the presence of transaction costs across extended time frames? • long-term consistency of excess returns: how consistent are excess returns over decades for the financial instrument, and what factors contribute to this consistency? • impact of turnover on excess returns: what are the measurable effects of turnover on the excess return, providing insights into the financial instrument's dynamics? • efficiency optimization and additional returns: by optimizing efficiency to minimize transaction costs, what additional returns can a firm generate, and to what extent does this contribute to financial performance? data we create ewps and vwps of three stocks from a pool of 181 stocks, using monthly returns from the center for research in security prices (crsp) from january 1, 1970, to january 1, 2023, to enable comparisons and validations through mathematical proofs. however, in the subsection titled "robustness checks in a larger portfolio," we show that our findings with three stocks hold true with a portfolio of 10 stocks as well. the 181 stocks in this paper meet two criteria: 1) 636 monthly returns from 01/01/1970 to 01/01/2023, and 2) a minimum market capitalization of $1 billion on 01/01/2023 to ensure enough liquidity for monthly portfolio adjustments. there are 971,970 unique ways to select three stocks from a pool of 181, as per the combination formula 𝑛𝐶𝑟 = 𝑛! 𝑟!(𝑛−𝑟)! , considering unordered selection. extending malladi and fabozzi (2017) two-stock portfolio model, we applied it to a three-stock scenario. we randomly chose 1,000 combinations from the 971,970 possibilities through exhaustive random sampling, calculating portfolio performance metrics and summary statistics from 01/01/1970 to 01/01/2023. tables 1 and 2 present summary statistics and a list of the 181 stocks, respectively. malladi & stanoyevitch 145 table 1. summary statistics for variables in the study monthly data stock price stock return shares outstanding (in thousands) market cap (in millions, $) mean 47.22 0.0124 315,957 16,527 median 36.92 0.0103 80,944 3,034 standard error 0.12 0.0003 2,345 122 standard deviation 42.39 0.0909 795,794 41,248 kurtosis 29.65 11.4846 63.84 35,665 skewness 4.27 0.8622 6.86 5,265 minimum 0.38 (0.8323) 451.00 0.69 maximum 747.63 2.1352 11,144,681 581,099 n, (181 x 636) 115,116 115,116 115,116 115,116 • the monthly data to build ewp and vwp from 01/01/1970 to 01/01/2023 is shown below. • a total of 115,116 monthly data points (181 stocks x 636 months) are used for analysis. • a median company in this study has a stock price of $36.92, yielded a monthly total return of 1.03% (or annualized return of 13.08%), 80.944 million shares outstanding, and a $3.034 billion market cap. • one can observe that monthly stock returns exhibit positive skewness and excess kurtosis. table 1 summarizes key statistics related to the study's variables. it covers monthly data collected from 01/01/1970 to 01/01/2023, totaling 115,116 data points from 181 stocks observed over 636 months. the median company featured in the study has a stock price of $36.92, delivering a monthly total return of 1.03% (equivalent to an annualized return of 13.08%), with 80.944 million shares outstanding and a market capitalization of $3.034 billion. an important observation from this data is that monthly stock returns exhibit a statistical distribution characterized by positive skewness and excess kurtosis, indicating certain asymmetry and heavy-tailedness in the return data. table 2 displays a roster of the 181 stocks employed in this study. the criteria for selecting these stocks are uncomplicated and revolve around two conditions: firstly, these stocks needed to exhibit uninterrupted monthly returns from 01/01/1970 to 01/01/2023. secondly, as of 01/01/2023, they were required to possess a market capitalization of at least $1 billion. this latter condition was essential to guarantee ample liquidity for the monthly portfolio rebalancing process, a critical component of the study's analytical framework. previous academic studies (demiguel et al., 2009; malladi & fabozzi, 2017; plyakha et al., 2012, 2021) also established a minimum market capitalization threshold to prevent the inclusion of companies with very small market capitalizations that pose challenges for the rebalancing process. financial services review, 33(1) 146 table 2. list of 181 stocks used in this study the selection criteria are straightforward with two conditions: 1) stocks have continuous monthly returns from 01/01/1970 to 01/01/2023; 2) stocks have $1 billion market capitalization on 01/01/2023 (to ensure enough liquidity for monthly portfolio rebalancing). methodology this section explains the steps for the ewp and vwp portfolio construction. for illustration, we construct two portfolios (ewp and vwp) comprising three stocks. we randomly picked these three stocks from our 181 stock pool. as shown previously, 971,970 unique ways of picking three stocks from this pool exist. the first subsection below demonstrates ewp, and the second one displays vwp. equal-weighted portfolios (ewp) consider v0 as the starting portfolio value (in this case, $100) at month t = 0. if m is the number of stocks in the portfolio, the initial value invested in a stock i, where i = 1,2,…,m, is vi,0 = v0 / m. the maximum value of m is 181 in our dataset. the initial quantity of stock i in the portfolio at the beginning of the month t = 0, qi,0 = vi,0 / pi,0, where pi,0 is the price of stock i at month t = 0. malladi & stanoyevitch 147 the value of t ranges from 0 to 636 months. the portfolio value before transaction cost (tc) at the end of month t, denoted as 𝑉𝑡[𝑁𝑜 𝑇𝐶], is computed with equation (1). 𝑉𝑡[𝑁𝑜 𝑇𝐶] = ∑ (𝑉𝑖,𝑡−1[𝑁𝑜 𝑇𝐶] × (1 + 𝑚 𝑖=1 𝑅𝑖,𝑡)) (1) where 𝑅𝑖,𝑡 is the monthly return of stock i, between months t-1 and t. 𝑅𝑖,𝑡, return between t and t-1, is obtained from the column ret in the msf table in the crsp. 𝑃𝑖,𝑡 is obtained from the column prc in the msf table in the crsp database. in the crsp database, returns (ret) are already adjusted for stock splits and dividends─ however, prices (prc) and shares outstanding (shrout) are not. the ewp monthly return without tc at the end of month t, denoted as 𝑅𝑡[𝑁𝑜 𝑇𝐶], is computed with equation (2). 𝑅𝑡[𝑁𝑜 𝑇𝐶] = 𝑉𝑡[𝑁𝑜 𝑇𝐶] 𝑉𝑡−1[𝑁𝑜 𝑇𝐶] − 1 (2) the number of stocks in the ewp at the end of month t, qi,t, is computed with equation (3). if there is a stock split or buy-back during the month, the qi,t is adjusted accordingly. the portfolio is rebalanced after the end of each month (before the next month's cycle begins) so that an equal amount of portfolio value is invested in each stock, i.e., 𝑉1,𝑡 = 𝑉2,𝑡 = ⋯ = 𝑉𝑚,𝑡. 𝑄𝑖,𝑡 = 𝑉𝑡[𝑁𝑜 𝑇𝐶] 𝑚 × 𝑃𝑖,𝑡 (3) the ewp monthly turnover (to) at the end of month t, denoted as 𝑇𝑂𝑡, is computed with equation (4). 𝑇𝑂𝑡 = ∑ (𝑎𝑏𝑠(𝑄𝑖,𝑡−𝑄𝑖,𝑡−1)) 𝑚 𝑖=1 ∑ 𝑄𝑖,𝑡−1 𝑚 𝑖=1 (4) portfolio turnover incurs transaction costs (tc) and diminishes portfolio returns. nevertheless, there remains no consensus on the exact tc. for instance, plyakha et al. (2021) apply a 50 basis point (0.50% bps) tc, citing french's (2008) claim that trading costs for u.s. equity decreased from 0.55% of total market cap in 1980 to just 0.21% in 2006. malladi and fabozzi (2017), however, opt for a more conservative tc of 169 bps, based on edelen et al.'s (2013) finding that the average tc is 1.69% for large funds (averaging $2.88 billion in assets) and 1.19% for small funds (averaging $164 million in assets), based on data from 3,799 open-end u.s. equity mutual funds, using quarterly portfolio holdings from morningstar spanning 1995 to 2006. our study employs the conservative 169 basis point transaction cost and demonstrates that if ewp outperforms vwp at 169 bps tc, it will thus also outperform at the lower 50 bps. the ewp value with tc at the end of month t, denoted as 𝑉𝑡[𝑇𝐶], is computed with equation (5). 𝑉𝑡[𝑇𝐶] = 𝑉𝑡[𝑁𝑜 𝑇𝐶] × (1 − 𝑇𝐶 × 𝑇𝑂𝑡) (5) the 𝑉𝑡[𝑇𝐶] is recursively incorporated into equation (1) starting from the first month (i.e., t = 1) to calculate the monthly portfolio value and return of ewp for all following months until 01/01/2023. the initial condition property that 𝑉𝑡=0[𝑇𝐶] = 𝑉𝑡=0[𝑁𝑜 𝑇𝐶] = 𝑉0 is useful in this recursive process. the ewp monthly return with tc at the end of month t, denoted as 𝑅𝑡[𝑇𝐶], is computed with equation (6). when tc = 0, 𝑅𝑡[𝑁𝑜 𝑇𝐶] = 𝑅𝑡[𝑇𝐶]. 𝑅𝑡[𝑇𝐶] = 𝑉𝑡[𝑇𝐶] 𝑉𝑡−1[𝑇𝐶] − 1 (6) value-weighted portfolio (vwp) we assume that the initial vwp value, denoted as �̂�0, is identical to the starting portfolio value, v0, in ewp, which stands at $100 at month t = 0. we employ hat-accented notation (ˆ) for vwp to differentiate it from ewp. in addition, vwp incorporates an additional variable, denoted as w, representing the value-weight of each stock. the weight of stock i in vwp (based on the market cap) at the beginning of month t, denoted as 𝑊𝑡, is computed with equation (7). 𝑊𝑖,𝑡 = 𝑃𝑖,𝑡 × 𝑄𝑖,𝑡 ∑ 𝑃𝑖,𝑡 × 𝑄𝑖,𝑡 𝑚 𝑖=1 (7) where 𝑄𝑖,𝑡 is the number of stocks outstanding for stock i, in month t. it is obtained from the column shrout in the msf table in the crsp and represents the unadjusted number of publicly held shares recorded in 1000s. it represents the actual, undiluted value, so fractional shares are possible. considering m is the number of stocks in the portfolio, the initial value invested in a stock i, (i = 1,2,…,m), is �̂�𝑖,0 = 𝑊𝑖,0 × �̂�0. the maximum financial services review, 33(1) 148 value of m is 181 in our dataset. the initial number of stocks in the portfolio at the beginning of the month t = 0, �̂�𝑖,0= �̂�𝑖,0 / �̂�𝑖,0, where �̂�𝑖,0 is the price of stock i at month t = 0. the value of t ranges from 0 to 636 months. the portfolio value before transaction cost (tc) at the end of month t, denoted as 𝑉�̂�[𝑁𝑜 𝑇𝐶], is computed with equation (8). �̂�𝑡[𝑁𝑜 𝑇𝐶] = ∑ (�̂�𝑡−1[𝑁𝑜 𝑇𝐶] × 𝑊𝑖,𝑡 × (1 + 𝑚 𝑖=1 𝑅𝑖,𝑡)) (8) the vwp monthly return without tc at the end of month t, denoted as 𝑅�̂�[𝑁𝑜 𝑇𝐶], is computed with equation (9). 𝑅�̂�[𝑁𝑜 𝑇𝐶] = �̂�𝑡[𝑁𝑜 𝑇𝐶] �̂�𝑡−1[𝑁𝑜 𝑇𝐶] − 1 (9) the stock quantity of stock i in the vwp at the end of month t, �̂�𝑖,𝑡, is computed with equation (10). the portfolio is rebalanced after the end of each month (before the next month's cycle begins) such that the weighted amount invested in each stock equals the market cap weights, i.e., �̂�𝑖,𝑡 �̂�𝑡 = 𝑊𝑖,𝑡. if there is a stock split or buy-back during the month, the �̂�𝑖,𝑡 is adjusted accordingly. �̂�𝑖,𝑡 = �̂�𝑡[𝑁𝑜 𝑇𝐶] × 𝑊𝑖,𝑡 𝑃𝑖,𝑡 (10) the vwp monthly turnover (to) at the end of month t, denoted as 𝑇�̂�𝑡, is computed with equation (11) and the vwp value with tc, denoted as �̂�𝑡[𝑇𝐶], is computed with equation (12). 𝑇�̂�𝑡 = ∑ (𝑎𝑏𝑠(�̂�𝑖,𝑡−�̂�𝑖,𝑡−1)) 𝑚 𝑖=1 ∑ �̂�𝑖,𝑡−1 𝑚 𝑖=1 (11) �̂�𝑡[𝑇𝐶] = �̂�𝑡[𝑁𝑜 𝑇𝐶] × (1 − 𝑇𝐶 × 𝑇�̂�𝑡) (12) the �̂�𝑡[𝑇𝐶] is recursively incorporated into equation (8) starting from the first month (i.e., t = 1) to calculate the monthly portfolio value and return of vwp for all following months until 01/01/2023. the initial condition property that �̂�𝑡=0[𝑇𝐶] = �̂�𝑡=0[𝑁𝑜 𝑇𝐶] = �̂�0 is useful in this recursive process. the vwp monthly return with tc at the end of month t, denoted as �̂�𝑡[𝑇𝐶], is computed with equation (13). when tc = 0, �̂�𝑡[𝑁𝑜 𝑇𝐶] = �̂�𝑡[𝑇𝐶]. �̂�𝑡[𝑇𝐶] = �̂�𝑡[𝑇𝐶] �̂�𝑡−1[𝑇𝐶] − 1 (13) results and discussion this section summarizes the results presented above. figure 1, titled "portfolio ending value (kernel density diagram)," illustrates the kernel density for four portfolio ending values described in section (4) (i.e., 𝑉𝑡[𝑁𝑜 𝑇𝐶], 𝑉𝑡[𝑇𝐶], �̂�𝑡[𝑁𝑜 𝑇𝐶], and �̂�𝑡[𝑇𝐶]). a kernel density diagram is a non-parametric way to estimate the probability density function of a random variable. it is a smooth curve that shows how likely it is to find a data point at any given value. kernel density diagrams are useful for visualizing data distribution and comparing the distributions of different data groups. malladi & stanoyevitch 149 figure 1. portfolio ending value (kernel density diagram) • this figure displays the kernel density for four portfolio ending values (vwp, ewp; with & without transaction costs, tc). • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. the average portfolio values are shown below. each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. tc = 1.69% of portfolio turnover. • a 'shorter and right-shifted ewp kernel density' suggests a lower and right-shifted data distribution with most points concentrated on the higher values, indicating a shift in central tendency towards higher values. • transaction costs significantly affect ewp negatively, whereas their impact on vwp is relatively minor. figure 1 presents the results of constructing 1,000 ewp and vwp portfolios, each comprised of three randomly selected stocks, and then showcases the average portfolio values. these portfolios are examined over the entire period, employing monthly returns from january 1, 1970, to january 1, 2023. the shorter and right-shifted ewp kernel density indicates a concentration of data points toward higher values, suggesting that ewp portfolios have a higher potential to generate higher returns than vwp. the kernel density diagram also shows that transaction costs significantly negatively impact ewp portfolios but only have a minor impact on vwp portfolios. this is because ewp portfolios rebalance more financial services review, 33(1) 150 frequently than vwp portfolios, resulting in higher turnover and transaction costs. it is important to note that the results shown in figure 1 are based on historical data and may not represent future performance. investors should carefully consider their risk tolerance and investment goals before deciding whether to invest in ewp or vwp portfolios. table 3 provides a detailed overview of the four portfolios' monthly and annualized returns. the table shows that the median and mean annualized returns are highest for ewp portfolios without transaction costs and lowest for vwp portfolios with transaction costs. this means that ewp portfolios have the potential to generate higher risk-adjusted returns than vwp portfolios, but they may have a higher standard deviation (riskier). the table also shows that transaction costs significantly negatively impact the performance of ewp portfolios. table 3. monthly and annualized returns of four portfolios monthly return and risk (m=3, n = 635,000) vwp ewp no tc tc no tc tc mean 0.0108 0.0107 0.0124 0.0115 median 0.0109 0.0108 0.0120 0.0113 stdev 0.0648 0.0647 0.0641 0.0640 annual return, risk, and risk-adjusted return vwp ewp no tc tc no tc tc mean 0.1373 0.1363 0.1592 0.1477 median 0.1390 0.1380 0.1533 0.1441 stdev 0.2243 0.2243 0.2219 0.2216 sharpe ratio 0.3891 0.3847 0.4918 0.4410 monthly portfolio turnover vwp ewp mean 0.0042 0.0488 median 0.0010 0.0384 stdev 0.0252 0.0441 correlations ewp (no tc) ewp (tc) vwp (no tc) vwp (tc) ewp (no tc) 1.0000 ewp (tc) 0.9944 1.0000 vwp (no tc) 0.6354 0.6568 1.0000 vwp (tc) 0.6356 0.6570 0.9998 1.0000 • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. • each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. • since each stock has 636 monthly prices (or 635 returns), one can have n = 635,000 (635 x 1,000). • the annualized returns are derived from the monthly returns using the formula: 𝑅𝐴𝑛𝑛𝑢𝑎𝑙𝑖𝑧𝑒𝑑 = (1 + 𝑅𝑚𝑜𝑛𝑡ℎ)12 − 1. • the median annualized returns in descending order for the portfolios are as follows: tc = 1.69% of portfolio turnover. malladi & stanoyevitch 151 • ewp (no tc): 0.1533, ewp (tc): 0.1441, vwp (no tc): 0.1390, and vwp (tc): 0.1380. • the mean values are 0.1592, 0.1477, 0.1373, and 0.1363, respectively. • the standard deviations of annual returns are 0.2219, 0.2216, 0.2243, and 0.2243, respectively. • assuming a 5% risk-free rate, the sharpe ratios, computed as (mean – rf)/stdev, are 0.49, 0.44, 0.39, and 0.38, respectively. • the correlation matrix in the lower panel shows that vwp and ewp have a monthly correlation of 0.6354 without tc and 0.6570 with tc. to calculate the annualized returns, the formula 𝑅𝐴𝑛𝑛𝑢𝑎𝑙𝑖𝑧𝑒𝑑 = (1 + 𝑅𝑚𝑜𝑛𝑡ℎ)12 − 1 is applied to the monthly returns. this transformation accounts for compounding over one year. the resulting median annualized returns for the portfolios, presented in descending order, are as follows: ewp (no tc) with a value of 0.1533, ewp (tc) with 0.1441, vwp (no tc) with 0.1390, and vwp (tc) with 0.1380. additionally, the mean values for these portfolios are calculated, and they are as follows: 0.1592 for ewp (no tc), 0.1477 for ewp (tc), 0.1373 for vwp (no tc), and 0.1363 for vwp (tc). furthermore, the standard deviations of the annual returns are calculated, yielding values of 0.2219 for ewp (no tc), 0.2216 for ewp (tc), 0.2243 for vwp (no tc), and 0.2243 for vwp (tc). the vwp and ewp have a monthly correlation of 0.6354 without tc and 0.6570 with tc. the 10-year us treasury bond yield at the end of each year is sourced from the federal reserve of st. louis (fred) and compiled by professor damodaran3. the average yield from 1970 to 2023 is rounded to 5% and used as the risk-free rate. this rate enables the computation of sharpe ratios to assess the risk-adjusted performance of each portfolio. the sharpe ratio, computed as (mean annualized return – annual risk-free rate) / standard deviation of annualized return, quantifies the return achieved per unit of risk taken. the observed sharpe ratios, denoting the risk-adjusted performance, were determined as 0.49 for ewp (no tc) and 0.44 for ewp (tc), followed by 0.39 for vwp (no tc) and 0.38 for vwp (tc), portraying a comprehensive perspective on the relative risk-adjusted performance of these portfolios. figure 2 depicts the four portfolios' historical median and mean values. figure 3 displays the four portfolios' mean annualized return kernel densities. the mean portfolio values are higher than the median. without transaction costs, ewp's ending value surpasses vwp 94.5% of the time (and 84.0% with tc). ewp's mean and standard deviation of monthly returns exceeds vwp 90.9% of the time without tc (80.5% with tc). 3 historical returns on stocks, bonds and bills: 19282023: https://pages.stern.nyu.edu/~adamodar/ https://pages.stern.nyu.edu/~adamodar/ financial services review, 33(1) 152 figure 2. historical portfolio median (left) and mean (right) value • the y-axis scale on the right chart is higher than that of the left (10 x 104 compared to 7 x 104). • the two figures display ending values for four portfolios (vwp, ewp, with and without transaction costs, tc). tc = 1.69% of portfolio turnover. • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. the median portfolio values are on the left, and the mean is on the right. each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. malladi & stanoyevitch 153 • without tc, ewp's ending value surpasses vwp 94.5% of the time (and 84.0% with tc). • the ewp’s mean/standard deviation of monthly returns surpasses vwp 90.9% of the time without tc (80.5% with tc). • for $100 invested on 01/01/1970, the median portfolio ending values on 01/01/2023 are as follows: ewp (no tc): $69,638, ewp (tc): $43,079, vwp (no tc): $25,543, and vwp (tc): $24,320. the mean values are $99,731, $57,795, $37,268, and $35,488, respectively. figure 3. portfolio mean annualized return (kernel density diagram) • this figure shows the mean annualized return kernel density for four portfolios (vwp, ewp; with & without transaction costs, tc). • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. the annualized returns are derived from the monthly returns using the formula: 𝑅𝐴𝑛𝑛𝑢𝑎𝑙𝑖𝑧𝑒𝑑 = (1 + 𝑅𝑚𝑜𝑛𝑡ℎ)12 − 1. • each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. tc = 1.69% of portfolio turnover. • transaction costs notably harm ewp but have a minor impact on vwp. • the peak of a kernel density diagram represents the mode or the most probable value in the dataset. the two ewps have higher modes. • the median annualized returns for the portfolios are as follows: ewp (no tc): 0.1533, ewp (tc): 0.1441, vwp (no tc): 0.1390, and vwp (tc): 0.1380. the mean values are 0.1592, 0.1477, 0.1373, and 0.1363, respectively. the standard deviations of annual returns are 0.2219, 0.2216, 0.2243, and 0.2243, respectively. financial services review, 33(1) 154 for an initial investment of $100 on january 1, 1970, the median portfolio ending values on january 1, 2023, are as follows: ewp (no tc) at $69,638, ewp (tc) at $43,079, vwp (no tc) at $25,543, and vwp (tc) at $24,320. the mean values for the portfolios are $99,731, $57,795, $37,268, and $35,488, respectively. when the mean portfolio values are higher than the median, the distribution of portfolio values is positively skewed. this means there are more portfolios with lower values than those with higher values, but the few portfolios with very high values are enough to pull the mean above the median. positively skewed distributions are often seen in financial markets like the stock market. this is because a few stocks outperform the market by a large margin. investors should be aware of the skewness of a distribution before investing in it. positively skewed distributions can offer the potential for high returns but also have a higher risk of loss. such a distribution is expected and can be approximated as log-normal if the price process follows a geometric brownian motion. for a robustness check, as shown in table 4, a paired two-sample test was conducted to assess the similarity of means between vwp and ewp. this test, evaluating whether the paired sample means are equivalent, consistently yielded a pvalue of 0 across all three instances in the data. this outcome strongly suggests substantial divergence between the mean values of vwp and ewp in each case, consequently leading to the rejection of the hypothesis proposing their equality. table 4. statistical significance tests paired test type, same mean mean vwp mean ewp p-value (two-sided) ci of difference t-stat df portfolio monthly turnovers 0.0042 0.0488 0.0000 *** 0.0447 to 0.0444 807.73 634,999 annualized returns (no tc) 0.1373 0.1592 0.0000 *** 0.0229 to 0.0209 41.47 999 annualized returns (with tc) 0.1363 0.1477 0.0000 *** 0.0124 to 0.0105 23.07 999 standard deviation of returns (no tc) 0.2243 0.2219 0.015 * 0.0042 to 0.0005 2.44 999 standard deviation of returns (tc) 0.2243 0.2216 0.0067 ** 0.0045 to 0.0007 2.72 999 • a paired samples test assesses the difference between two sets of paired observations or measurements. it is used when each observation in one set relates to a specific observation in the other. • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023, i.e., 636 monthly prices (or 635 returns). tc = 1.69% of portfolio turnover. • in a paired t-test, when comparing two groups or conditions, the degrees of freedom (df) are calculated as n 1, where n is the number of paired observations. this accounts for the constraint imposed by using the paired differences to estimate the mean difference. • the same mean test evaluates whether the averages of the paired samples (vwp and ewp) are identical or distinct. the confidence interval (ci) of the mean difference is provided for each test. • across all three tests in the provided data, a p-value of 0 was obtained, implying significant divergence between the means of vwp and ewp in each scenario. • this substantiates the rejection of the hypothesis, stating that the means are equal. • we conclude that the turnover and returns of vwp and ewp exhibit statistically significant differences in their means. malladi & stanoyevitch 155 • the asterisk (*) denotes the significance level, * indicates a p-value less than 0.05, ** for a p-value less than 0.01, and *** for a p-value less than 0.001. figure 4 analyses monthly portfolio turnover for four portfolios. the left part of the figure indicates that ewp exhibits significantly higher monthly portfolio turnover than vwp, with a median ratio of 34.7 and a mean ratio of 11.4 of ewp monthly turnover to vwp monthly turnover. this heightened turnover increases transaction costs, primarily impacting ewp (tc) returns, which are notably lower than ewp (no tc). figure 4. monthly portfolio turnover (kernel density diagram on the left and historical trend on the right) financial services review, 33(1) 156 • these figures show four portfolios' monthly portfolio turnover (vwp, ewp; with & without transaction costs, tc). tc = 1.69% of portfolio turnover. • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. • the figure on the left shows that ewp has an order of magnitude higher portfolio turnover than the vwp. • the ratio of ewp monthly turnover to vwp monthly turnover has a median of 34.7 and a mean of 11.4. • high turnover results in high transaction costs, primarily causing ewp (tc) returns to be significantly lower than ewp (no tc) returns. • the median annualized returns for the portfolios are as follows: ewp (no tc): 0.1585, ewp (tc): 0.1476, vwp (no tc): 0.1374, and vwp (tc): 0.1365. the mean values are 0.1600, 0.1485, 0.1376, and 0.1366, respectively. the annual returns exhibit standard deviations of 0.2198, 0.2195, 0.2232, and 0.2232, respectively. • the figure on the right shows the relative turnover of ewp compared to ewp since 1970. ewp turnovers increase when returns are large (either positive or negative). the next figure (5) shows the monthly stock return trend. figure 4 shows that ewp portfolios are more expensive to maintain than vwp portfolios. this is because ewp portfolios need to be rebalanced more frequently, which results in higher transaction costs. the figure on the right suggests that ewp portfolio managers are more active traders when market returns are high. on the right side of the figure, the relative turnover of ewp compared to vwp since 1970 is displayed. notably, ewp turnovers increase during periods of large positive or negative returns. this relationship between turnover and returns is further explored in figure 5, which examines the monthly stock return trend. figure 5 presents the average monthly stock returns for all 181 stocks included in the study. the chart highlights substantial stock return fluctuations, particularly in 1974, 2008, 2000, and 2018. interestingly, these same years coincide with a pronounced spike in ewp turnover, as observed in figure 4. this alignment malladi & stanoyevitch 157 suggests a relationship between significant fluctuations in stock returns and corresponding increases in ewp turnover, shedding light on the interplay between market dynamics and portfolio turnover during these particular years. figure 5. average monthly stock returns • this chart displays the monthly average return of all 181 stocks in this study from 01/01/1970 to 01/01/2023. • this chart reveals large stock return fluctuations in 1974, 2008, 2000, and 2018. • the same years exhibit a noticeable ewp turnover spike (figure 4, right). figure 6 presents the portfolio mean annualized returns by decade, showcasing the culmination value of a portfolio after a $100 investment made at the initiation of each decade. when considering instances lacking transaction costs, the ewp consistently outperforms, yielding a higher ending value in each of the five decades. however, when transaction costs are factored in, the ewp yields a higher ending value in four out of five decades, except from 1983 to 1992. financial services review, 33(1) 158 figure 6. portfolio mean annualized returns by decade • the provided chart illustrates the decade-end value of a portfolio following a $100 investment at the start of each decade. • ewp consistently yields a superior ending value across all five decades in scenarios without transaction costs. • ewp demonstrates a higher ending value in four out of five decades, excluding the 1983-92 period. tc = 1.69% of portfolio turnover. various comparisons are conducted with plyakha et al. (2021) at 50 basis points (bps) and malladi and fabozzi (2017) at 169 bps. plyakha et al. note that at 50 bps tc, the equal-weighted portfolio (ewp) surpasses the value-weighted portfolio (vwp) with mean annual returns of 13.19% and 10.48%, respectively, equating to an excess return of 271 basis points. similarly, malladi and fabozzi demonstrate at 169 bps tc that the ewp outperforms the vwp, securing mean annual returns of 13.88% and 12.87%, respectively, with an excess return of 101 basis points. figure 7 displays the impact of tc on the ewp and vwp returns. additionally, our findings reveal a statistically significant excess annualized return of 188 basis points for the ewp over the vwp when tc = 50 bps and 115 basis points when tc = 169 bps. notably, figure 7 shows that when tc reaches 728 basis points of portfolio turnover, the ewp and vwp portfolio returns converge, suggesting a pivotal threshold for these portfolios. malladi & stanoyevitch 159 figure 7. impact of transaction costs on portfolio return transaction cost (tc, bps) 0 50 100 169 annualized returns (vwp) 0.1374 0.1371 0.1368 0.1364 annualized returns (ewp) 0.1593 0.1559 0.1525 0.1479 excess return (ewp vwp) 0.0219 0.0188 0.0157 0.0115 standard deviation of returns (vwp) 0.2215 0.2215 0.2215 0.2215 standard deviation of returns (ewp) 0.2191 0.2190 0.2190 0.2189 • the chart demonstrates how tc (on the x-axis) affects ewp and vwp portfolio returns (on the y-axis). • 1,000 ewp and vwp portfolios are constructed using three randomly selected stocks. each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. thick lines depict exhaustive random sampling, while lighter dotted lines represent trendlines. • plyakha et al. (2021) show that at 50 bps, tc ewp outperforms vwp, with mean annual returns of 13.19% and 10.48%, respectively (or 271 bps excess returns). • malladi and fabozzi (2017) demonstrate that at 169 bps tc, ewp outperforms vwp, with mean annual returns of 13.88% and 12.87%, respectively (or 101 bps excess returns). • we report that the ewp achieves a statistically significant excess annualized return over the vwp of 188 bps when tc = 50 bps and 115 bps when tc = 169 bps. • when tc reaches 728 bps of portfolio turnover, ewp and vwp returns converge. robustness checks in a larger portfolio so far, we have used three stocks in our analysis to facilitate easier comparisons and validations through mathematical proofs. in this subsection, we increase the number of stocks to ten to demonstrate the robustness of our results. the findings with a larger number of securities y = -0.0038x + 0.1633 r² = 0.9939 y = -0.0003x + 0.1378 r² = 0.9945 0.1300 0.1350 0.1400 0.1450 0.1500 0.1550 0.1600 0 50 100 169 an nu al iz ed r et ur n tc = 728 bps ewp vwp financial services review, 33(1) 160 maintain the integrity of our initial results with three stocks. there are 8.074 quadrillion unique ways to select ten stocks from a pool of 181, as per the combination formula 𝑛𝐶𝑟 = 𝑛! 𝑟!(𝑛−𝑟)! , considering unordered selection. given the impracticality of exhaustively analyzing quadrillions of portfolios, we created 1,000 portfolios of ten randomly selected stocks each for our analysis. the results of both threeand ten-stock portfolios are presented in table 5. the results of the analysis are robust, as evidenced by consistent patterns across different portfolio types and varying numbers of stocks (m). for both vwp and ewp, as the number of stocks increases from three to ten, the standard deviation of returns decreases significantly, as expected (e.g., annual stdev decreased from 22% to 16% for ewp and to 14% for vwp). ewp mean and median returns exceed those of vwp in both cases (i.e., m = 3 and m = 10). ewp maintained higher returns than vwp in both cases. both strategies experienced significant risk reduction and improved sharpe ratios with an increased number of stocks. transaction cost impact decreased significantly for ewp as the number of stocks increased, while it remained minimal for vwp in both cases. these results align with portfolio theory: increased diversification leads to lower risk, equal-weighting outperforms value-weighting in terms of returns and risk-adjusted returns, and the impact of transaction costs on equal-weighted portfolios decreases with more stocks, likely due to reduced rebalancing needs. the consistent decrease in returns as transaction costs increase, the consistent outperformance of ewp over vwp, and the reduction in risk as the number of stocks in the portfolio increases reinforce the reliability and robustness of the results, demonstrating that the observed trends are not anomalies but reflective of underlying market dynamics and portfolio characteristics. conclusions in conclusion, the paper's results based on the methodology and data presented demonstrate several key findings. ewp portfolios exhibit the potential for higher risk-adjusted returns, particularly in the absence of transaction costs. transaction costs significantly impact the performance of ewp portfolios. moreover, ewp portfolios tend to have higher turnover, leading to increased transaction costs, especially during periods of significant market returns. these results highlight the trade-offs between higher return potential and increased costs associated with ewp portfolios. investors should carefully consider their risk tolerance and investment objectives when choosing between ewp and vwp portfolios, considering the complex relationship between turnover, transaction costs, and market dynamics. future research in this area may encompass the following directions: • transaction cost optimization: investigating strategies and methodologies to minimize transaction costs in ewp portfolios, potentially through advanced algorithms or trading techniques. • incorporating various transaction costs by periods: we kept transaction costs fixed to ensure consistency with previous studies and facilitate cross-comparison, as shown in figure 7. incorporating transaction costs by period is a valuable direction for future research. • we randomly chose 1,000 combinations (of three stocks) from the 971,970 possibilities through exhaustive random sampling (from the available pool of 181 stocks). this pool could typically originate from an investment firm's stock selection methodology, an index, or similar sources. researchers are encouraged to experiment with different pools and varying numbers of stocks in the portfolio. • asset class expansion: extending the research to include various asset classes beyond equities, such as fixed income, real estate, or alternative investments, to understand how ewp and vwp strategies perform in diverse investment landscapes. • machine learning applications: leveraging machine learning and predictive modeling techniques to forecast future returns, turnover, and transaction costs for both ewp malladi & stanoyevitch 161 and vwp portfolios, offering predictive insights into portfolio management. • global market comparisons: conducting comparative analyses of ewp and vwp strategies in different global markets and examining the influence of cross-border investment considerations. these research avenues can collectively contribute to a more comprehensive understanding of the dynamics and implications of ewp and vwp portfolio management in various contexts, offering valuable insights for investors and portfolio managers. financial services review, 33(1) 162 table 5. robustness check in a 3-stock (left panel) and 10-stock (right panel) portfolio • 1,000 ewp and vwp portfolios are constructed using three (left panel) or ten (right panel) randomly selected stocks. • each run examines the portfolio for the whole period using monthly returns from 01/01/1970 to 01/01/2023. • since each stock has 636 monthly prices (or 635 returns), one can have n = 635,000 (635 x 1,000). • the annualized returns are derived from the monthly returns using the formula: 𝑅𝐴𝑛𝑛𝑢𝑎𝑙𝑖𝑧𝑒𝑑 = (1 + 𝑅𝑚𝑜𝑛𝑡ℎ)12 − 1. • the results of the analysis are robust, as evidenced by consistent patterns across different portfolio types and varying numbers of stocks (m). for both vwp and ewp, as the number of stocks increases from three to ten, the standard deviation of returns decreases significantly, as expected (e.g., annual stdev decreased from 22% to 16% for ewp and to 14% for vwp). ewp mean and median returns exceed those of vwp in both cases (i.e., m = 3 and m = 10). • despite these reductions in returns, the standard deviations of the portfolios show minimal change, indicating that risk levels remain relatively stable regardless of transaction costs and the number of stocks in the portfolio. • the sharpe ratio, a measure of risk-adjusted return, consistently decreases with the inclusion of transaction costs, highlighting the impact on overall performance. • the consistent decrease in returns and sharpe ratios across different configurations (monthly and annual returns, with and without transaction costs) reinforces the reliability and robustness of the results, demonstrating that the observed trends are reflective of underlying market dynamics and portfolio characteristics. monthly return and risk (m=3, n = 635,000) vwp ewp no tc tc no tc tc mean 0.0108 0.0107 0.0124 0.0115 median 0.0109 0.0108 0.0120 0.0113 stdev 0.0648 0.0647 0.0641 0.0640 annual return, risk, and risk-adjusted return vwp ewp no tc tc no tc tc mean 0.1373 0.1363 0.1592 0.1477 median 0.1390 0.1380 0.1533 0.1441 stdev 0.2243 0.2243 0.2219 0.2216 sharpe ratio 0.3891 0.3847 0.4918 0.4410 monthly return and risk (m=10, n = 635,000) vwp ewp no tc tc no tc tc mean 0.0099 0.0099 0.0123 0.0123 median 0.0117 0.0116 0.0144 0.0143 stdev 0.0417 0.0417 0.0465 0.0465 annual return, risk, and risk-adjusted return vwp ewp no tc tc no tc tc mean 0.1259 0.1249 0.1582 0.1574 median 0.1498 0.1486 0.1873 0.1862 stdev 0.1445 0.1445 0.1610 0.1611 sharpe ratio 0.5254 0.5185 0.6721 0.6668 financial services review, 33(1) 163 references amenc, n., & goltz, f. (2013). smart beta 2.0. the journal of index investing, 4(3), 15–23. https://doi.org/10.3905/jii.2013.4.3.015 amenc, n., goltz, f., & lodh, a. (2016). smart beta is not monkey business. the journal of index investing, 6(4), 12–29. https://doi.org/10.3905/jii.2016.6.4.012 amenc, n., goltz, f., & martellini, l. (2011). “a survey of alternative equity index strategies”: a comment. financial analysts journal, 67(6), 14–16. https://doi.org/10.2469/faj.v67.n6.8 amenc, n., & sourd, v. l. (2005). portfolio theory and performance analysis. john wiley & sons. amihud, y. (2002). illiquidity and stock returns: cross-section and time-series effects. journal of financial markets, 5(1), 31–56. https://doi.org/10.1016/s13864181(01)00024-6 arnott, r. (2006). orthodoxy overwrought. institutional investor. https://www.researchaffiliates.com/content/d am/ra/publications/pdf/p-2006-dec-anoverwrought-orthodoxy.pdf arnott, r. d., hsu, j., & moore, p. (2005). fundamental indexation. financial analysts journal, 61(2), 83–99. https://doi.org/10.2469/faj.v61.n2.2718 atchison, m. d., butler, k. c., & simonds, r. r. (1987). nonsynchronous security trading and market index autocorrelation. the journal of finance, 42(1), 111–118. https://doi.org/10.2307/2328422 demiguel, v., garlappi, l., & uppal, r. (2009). optimal versus naive diversification: how inefficient is the 1/n portfolio strategy? the review of financial studies, 22(5), 1915– 1953. https://doi.org/10.1093/rfs/hhm075 edelen, r., evans, r., & kadlec, g. (2013). shedding light on “invisible” costs: trading costs and mutual fund performance. financial analysts journal, 69(1), 33–44. https://doi.org/10.2469/faj.v69.n1.6 fama, e. f. (1970). efficient capital markets: a review of theory and empirical work. the journal of finance, 25(2), 383–417. https://doi.org/10.2307/2325486 french, k. r. (2008). presidential address: the cost of active investing. the journal of finance, 63(4), 1537–1573. https://doi.org/10.1111/j.15406261.2008.01368.x hsu, j. c. (2006). cap-weighted portfolios are sub-optimal portfolios (ssrn scholarly paper 934425). https://papers.ssrn.com/abstract=934425 investment company institute. (2023). 2023 investment company fact book, p. 22. https://www.icifactbook.org/ lintner, j. (1969). the valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets: a reply. the review of economics and statistics, 51(2), 222–224. https://doi.org/10.2307/1926735 malladi, r., & fabozzi, f. j. (2017). equalweighted strategy: why it outperforms valueweighted strategies? theory and evidence. journal of asset management, 18(3), 188– 208. https://doi.org/10.1057/s41260-0160033-4 markowitz, h. (1952). portfolio selection*. the journal of finance, 7(1), 77–91. https://doi.org/10.1111/j.15406261.1952.tb01525.x pae, y., & sabbaghi, n. (2015). equally weighted portfolios vs value weighted portfolios: reasons for differing betas. journal of financial stability, 18, 203–207. https://doi.org/10.1016/j.jfs.2015.04.006 plyakha, y., uppal, r., & vilkov, g. (2012). why does an equal-weighted portfolio outperform valueand price-weighted portfolios? (ssrn scholarly paper id 2724535). social science research network. https://doi.org/10.2139/ssrn.2724535 financial services review, 33(1) 164 plyakha, y., uppal, r., & vilkov, g. (2021). equal or value weighting? implications for asset-pricing tests. in c. zopounidis, r. benkraiem, & i. kalaitzoglou (eds.), financial risk management and modeling (pp. 295–347). springer international publishing. https://doi.org/10.1007/978-3030-66691-0_9 roll, r. (1977). a critique of the asset pricing theory’s tests part i: on past and potential testability of the theory. journal of financial economics, 4(2), 129–176. https://doi.org/10.1016/0304405x(77)90009-5 sharpe, w. f. (1964). capital asset prices: a theory of market equilibrium under conditions of risk. the journal of finance, 19(3), 425–442. https://doi.org/10.1111/j.15406261.1964.tb02865.x swade, a., nolte, s., shackleton, m., & lohre, h. (2023). why do equally weighted portfolios beat value-weighted ones? the journal of portfolio management, 49(5), 167–187. https://doi.org/10.3905/jpm.2023.1.482 taljaard, b. h., & maré, e. (2021). why has the equal weight portfolio underperformed and what can we do about it? quantitative finance, 21(11), 1855–1868. https://doi.org/10.1080/14697688.2021.1889 020 disclosure statements funding: this study has not been funded by any entity. conflict of interest: authors do not have any conflicts of interest associated with this article. ethical approval: this article does not contain any studies with human participants or animals. pii: s1057-0810(96)90007-4 financial services review, 32): 149-151 copyright 0 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. book, software, and web site reviews douglas kahl, editor university of akron this is a new feature in financial services review. to make it a success, we need vol unteers willing to do the reviews. if you would be interested in doing an occasional review, please contact douglas kahl, department of finance, university of akron, akron, oh 443254803 or e-mail him at dkahl@ljakron.edu. association for investment management and research (aimr) web site reviewed by: timothy r. smaby, assistant professor of finance, penn state at erie, the behrend college. the mission of the association for investment management and research (aimr) is to serve its more than 27,000 members and, through them, to serve investors by educating and examining investment professionals. aimr, through the institute of charted financial analysts, grants the prestigious chartered financial analyst (cfa) designation. the (aimr) makes available at its home page (http://www.aimr.com) useful information of interest to finance professors interested in pursuing the cfa designation, seeking external funding, or keeping abreast of developments among investment practitioners. the web site is not fancy by currently world-wide-web standards, but it is well organized and easy to use. finance professors planning to pursue the cfa designation, or those encouraging their students to do so, will find extensive information available on the cfa exam most useful. here you will find exam deadlines, dates and fees, an outline of the topics covered in the exam, and recent passing rates. you can request registration materials online. my only crit icism concerns the lack of information on the special program and fees for college and uni versity faculty interested in the cfa designation. finance faculty feeling the pressure to pursue external funding for their research will find the information on the research foundation of the icfa helpful. the research foun dation funds and publishes a diverse assortment of research studies in the form of mono graphs and tutorials. lists of currently funded projects and published studies will give you an idea of the research topics of interest to the foundation. aimr’s advocacy program, by identifying and monitoring legislative, regulatory, and professional issues and developments that affect the domestic and global capital markets and investment practice, will enable you to keep abreast of “hot issues” and develop new research ideas. 150 financial services review 5(2) 1996 finally, the site provides links to the ~i~n~~~~ analysts j~u~qz, the cfa digest, and the lsfa digest, all of which are published by the aimr. the new york stock exchange (nyse), the chicago board options exchange (cboe), and the financial management association websites reviewed by: thomas eyssell, associate professor, university of missouri-st. louis one of the biggest potential benefits of the worldwide web to financial users is the availability of up-to-date information on financial markets and instruments. two related websites of potential interest to fsr readers are those devoted to the new york stock exchange (nyse) and the chicago board options exchange (cboe). the cboe website is extremely user-~endly, and seems geared more toward the individual investor. upon reaching the site (http://www.cboe.com), one finds links to pages describing “what’s new,” cboe “products,” the “options institute” (the cboe’s education arm), as well as a “virtual visit” to the cboe. the latter consists of extensively captioned graphics of the trading floor, as well as a “virtual stroll” around the su~ounding chicago area. those contemplating the inclusion of options in a diversified portfolio, and those teaching basic options mechanics will find the “understanding options” link useful. included at this link are online versions of two cboe publications: “characteristics and risks of standardized options” and the options clearing corporation’s options prospec tus. these publications provide a great deal of nuts-and-bolts information sometimes glossed over in standard texts. finally, for those seeking to perform empirical research and analysis, the cboe page provides links to several sources of security and market data. the nyse website (http://www.nyse.com) is highlighted by an electronic version of the organization’s 1995 annual report. included at this site are several items of interest to the financial planner, teacher, or researcher. for example, one can go to an alphabetized list of nyse firms and, upon clicking on a ticker symbol, be immediately transferred to that firm’s homepage. (be aware, however, that not all nyse-listed firms appear.) the nyse website is a bit less user-friendly than that of the cboe. for example, clicking on “visit” takes one to a single image of the trading floor. in addition the educa tion/informational aspects are not nearly as extensive or compelling. on the other hand, it is possible to download, at no cost, daily closing price and volume data spanning several decades. in addition, one can obtain reasonably up-to-date lists of firm listings and dele tions, trading habits, discipline actions, and so forth. in any case, these two market sites are well worth the trip. the financial management association has put together an extensive website (http:// www.fma.org) which will be of interest to finance academics as well as to finance practi tioners. among other things, the interested websurfer can find out about upcoming confer ences, events, and publication dates through 1997, services for students, finance links, and, of course, placement services for employers and job seekers. one can search the “positions available” section by finance category (business finance, financial institutions and markets, general finance, insurance, investments, real estate, and other areas of finance), or alphabetically by school name. apparently, the list of available positions is updated periodically, since many of the listings sport a bright red “new!” tag. pii: s1057-0810(99)00029-3 international index funds and the investment portfolio scott aielloa, natalie chieffeb,* aliberty mutual group, boston, ma, usa bohio university, college of business, finance department, athens, oh 45701, usa abstract financial advisors often recommend that investors diversify their investments internationally and also use mutual funds with the lowest expenses. recently it has been possible to use both of these strategies by purchasing an international index fund. this study considers international index funds as a means of portfolio diversification. performance is evaluated using monthly return data on nine international indexes from january 1989 through december 1997. returns are measured against the s & p 500index returns. the results of statistical tests suggest that international index investing does not offer superior returns compared to the s & p 500 index but diversification benefits do exist. © 1999 elsevier science inc. all rights reserved. jel classification:g110; g29 keywords:international indexes; individual investors; index funds 1. introduction individual investors find many pieces of advice directed toward them. some of this advice recommends that they should diversify internationally, they should include high-growth emerging market stocks in their portfolios, and they should buy low-expense index funds. in this paper we review the literature in these three areas. we then determine empirically if international index investing can better diversify a domestic portfolio. the results of this study may help investors determine if international investing combined with index investing can improve their portfolios. * corresponding author. tel.:11-740-593-9320; fax:11-740-593-9539. e-mail address:chieffe@ohio.edu (n. chieffe) financial services review 8 (1999) 27–35 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00029-3 investors have several options when choosing international equities. a domestic investor can invest directly in foreign securities that trade on foreign exchanges. but, the direct investor pays higher fees, may be handicapped by a lack of information and faces additional currency and political risks. thus, other options may be indicated. one such option is the purchase of euroequities, securities that are listed on any of the foreign stock exchanges and also on an american exchange. these dual-listings give americans the chance to take advantage of international investing while avoiding the disadvantages of direct investment. multinational corporations (mncs) that are based in the united states have significant exposure in foreign countries and offer a second option. a portfolio containing mncs gains the increased diversity of foreign investment, but at the lower costs of domestic market investment. a third option is american depository receipts (adrs). these receipts for shares of foreign companies held by u.s. banks offer the american investor international diversification. the bank converts the dividends from the foreign currency and pays them in dollars to the shareholder. adrs are popular because investors do not have to leave the domestic market to invest in foreign firms. a fourth, and often preferred, option for international investment is the purchase of shares in an internationally diversified mutual fund or closed-end fund. of these funds, international index mutual funds are relatively new, and have not gone through the same tests as domestic funds. although many pension funds and institutional investors already invest in international mutual funds, only a few studies have examined their performance. it seems reasonable that if there are benefits from both international investing and domestic index investing, there may be benefits from international index investing. the next section reviews the relevant literature in this area. a description of the research design and data follows in the third section. the fourth section contains the empirical analysis. the last section offers the applications and conclusions. 2. effects of international diversification and index investing 2.1. international investing hunter and coggin (1990) show that international diversification can reduce investment risk to about 56% of the level that can be achieved with national diversification. russell (1998) considers the options that investors have for international investing. he tests whether u.s. exchange-listed securities such as adrs, mncs, and closed-end country funds behave more like the new york composite index than the market they represent. his study suggests that these securities do not provide diversification benefits for the u.s. investor. cumby and glen (1990) evaluate the performance of international mutual funds to determine whether the managers are responsible for above average returns or if the returns are simply the result of international diversification. they find that these internationally diversified mutual funds have superior performance when compared to the u.s. stock market. the authors attribute this solely to the benefits of international diversification, rather 28 s. aiello, n. chieffe / financial services review 8 (1999) 27–35 than to the performance of the active managers. after testing these funds against an international index, cumby and glen conclude that there is “no evidence that the funds, either individually or as a whole, provide investors with performance that surpasses that of a broad, international equity index over this same period”. they also find that the funds systematically under-perform the indexes during the october 1987 stock market crash. though their proponents often argue that active managers are able to avoid such crashes by moving cash in and out of securities, these findings suggest those active managers only magnify the losses. eun, kolodny, and resnick (1991) test u.s.-based international fund performance in terms of mean-variance efficiency. they evaluate the funds using the jensen (1968), sharpe (1966), and treynor (1965) measures and find that the funds provide a valuable opportunity for international diversification. of the thirteen funds, ten outperform the s & p 500 index (s & p 500) when evaluated with the sharpe ratio. when these funds are compared to the morgan stanley capital international (msci) world index (w) only two funds outperform the benchmark. these researchers also study the complementary effects of adding international funds to a u.s. benchmark, s & p500. they conclude that a u.s. investor may benefit from adding any international fund, except the canadian fund. solnik (1995) shows that the variability of returns for an internationally diversified portfolio is one-tenth that of a typical domestic diversified portfolio. apap and collins (1994) show that the performance of international mutual funds exceeds that of u.s. domestic mutual funds. the fluctuations of the u.s. dollar exchange rate have a negligible effect in the long run when the mutual fund invests in more than one country. on the other hand, droms and walker (1994) find that international mutual funds do not offer excess risk-adjusted returns. however, the funds do offer an efficient means of investing in a broadly diversified portfolio of common stocks with returns that are appropriate for the risk exposure. ho, milevsky, and robinson (1999) focus on retired investors who are consuming accumulated capital and the income from it. they show that retired u.s. investors do not benefit from international equity diversification. 2.2. investing in emerging markets divecha, drach, and stefek (1992) show that modest investments in emerging markets are likely to reduce overall portfolio risk. masters (1998) suggests that the ideal diversified portfolio consists of at least 6% allocated to emerging markets. khanna (1996) recommends investing in emerging markets because their rapid economic growth provides for good returns. even though the returns are volatile, these markets are poorly correlated to developed markets and, thus, offer good portfolio diversification benefits. speidell and sappenfield (1992) discuss the fact that global diversification depends on the correlations among countries. they suggest that as the developed markets move less independently of each other, the correlations among them increase. this decreases the potential for diversification benefits. on the other hand, emerging markets are less correlated with the developed markets; this increases their relative diversification advantage. 29s. aiello, n. chieffe / financial services review 8 (1999) 27–35 2.3. index funds gruber (1996) questions why so many investors choose actively managed mutual funds when their performance, on average, is inferior to that of index funds. bogle (1998) examines the relationships among risk, return, and cost. he shows that low-cost passively managed index funds deliver the highest risk-adjusted return for each category of mutual funds. malhorta and mcleod (1997) show that mutual fund expenses are inversely related to the investor’s returns. hogan (1994) recommends that financial advisors use index mutual funds in an asset-class investing strategy. hogan stresses that this strategy is consistent with modern portfolio theory and presents important cost advantages for clients. 2.4. international indexes although it is certain that international index funds have a lower correlation to the u.s. stock market than do domestic index funds, it is not clear whether they outperform typical domestic index strategies or actively managed international strategies. one difficulty in determining the performance of international index funds is that there is no accepted standard index like the s & p 500 (masters 1998). more than 38% of the international equities are included in the msci european australasia and far east index (eafe), but there are a variety of other indexes to choose from. emerging market indexes were created in 1985 by the international finance corporation and in 1988 by msci. investors expect indexes to closely track the market they follow, in this case the emerging markets. one of the problems associated with international indexes is that each is different in the way it tracks its underlying market. another problem with emerging markets indexes is country weights. the global capitalization weight for a particular country can change dramatically over a short period of time. because the benchmark tracks only emerging markets, new developed markets must be removed and new emerging markets must be added. unstable and confusing weights lead to higher costs caused by higher turnover ratios. transaction costs are another factor that lead to inefficiency and these are about five times those encountered in u.s. index funds. 2.5. current study cumby and glen (1990), eun, kolodny, and resnick (1991), solnik (1995), apap and collins (1994), and droms and walker (1994) find advantages in international mutual funds. divecha, drach, and stefek (1992), masters (1998), khanna (1996), and speidell and sappenfield (1992) recommend investing in emerging markets. gruber (1996), bogle (1998), malhorta and mcleod (1997), and hogan (1994) find advantages in index funds. in this paper, we ask, if investors should choose international mutual funds, and if they should choose index funds, should they then also choose international index funds including emerging markets index funds? we find no conflict with ho, milevsky, and robinson (1999) who find no benefit in international diversification for retired u.s. investors. unlike their study, this study concerns the accumulation phase. 30 s. aiello, n. chieffe / financial services review 8 (1999) 27–35 3. research design to prove that international index funds are an excellent way for the domestic individual to invest, we perform tests to determine whether international indexes outperform domestic indexes and to determine whether there are the diversification benefits to investing in international index funds. we first test the monthly return data of nine international indexes to determine whether they outperform the s & p 500benchmark on a risk-adjusted basis. these tests include the sharpe, treynor, and jensen measures. the sharpe (1966) measure is the ratio of average risk premium to the total risk the portfolio faces during the evaluation period. this sharpe measure is relevant for the investor choosing a specific fund for a major portion of his/her portfolio: si 5 r# i 2 r# f si (1) where si is the sharpe measure for indexi ; r# i is the average return on indexi ; r# f is the average risk-free rate; si is the standard deviation of returns on the index. the treynor (1965) measure uses systematic risk and allows the investor to compare the individual index returns to the market return, disregarding diversification: ti 5 r# i 2 r# f bi (2) where ti is the treynor measure for indexi ; bi is the beta or systematic risk for indexi . for the jensen (1968) measure we regress excess portfolio returns against excess domestic market returns. the intercept represents the jensen measure. when the intercept is positive and statistically significant, superior performance is noted. this measure can also measure individual fund performance: ai 5 ~r# i 2 r# f! 2 @bi~r# m 2 r# f!# (3) where a is the jensen measure; r# m is the average return on the market. we then ask, are there diversification benefits to investing in international index funds? to answer this question we examine the correlation of the international index funds with the s & p 500. we attempt to answer the following questions: 1) does a linear relationship exist between the individual international index returns and the s & p 500returns?; 2) does a linear relationship exist between the returns on the entire portfolio of international indexes and the returns on the s & p 500? we use monthly return data from nine international indexes in this study. these indexes 31s. aiello, n. chieffe / financial services review 8 (1999) 27–35 are those which international index mutual funds would attempt to match. the sample period is january 1988 through december 1997. we chose the indexes on the basis of region and data availability. the indexes encompass every region of the world and are formed, designed and updated by msci. the s & p 500 is thebenchmark u.s. index. the three month u.s. treasury bill yield is the risk-free rate. we compute excess returns from this security. 4. analysis of data and results table 1 lists the risk-return data for the nine international indexes and the s & p500. over the 1988 to 1997 period the average monthly return on the international indexes is 0.0076 or 0.76% per month. the average monthly return on the s & p 500 is 0.0120 or 1.2%. of the nine international indexes only two return more than the s & p 500index. the emerging markets global index (emg) posts the largest monthly returns at 1.37% and the pacific index (p) returns the smallest at 0.04%. the lowest standard deviation (sd) of returns is that of the americas free index (af) at 0.0345. the sd of returns on the s & p 500 is 0.0348. the index with the greatest sd (0.0833) is the europe & middle east index (eme). the overall average sd of returns is 0.0560. the investor who seeks the lowest risk based on sd would consider thes & p 500 or the af. table 2 lists the portfolio performance measures for the indexes. based on the sharpe measure, or full risk of the fund for each index, only af outperformed the s & p 500. but af has an s & p 500 component and is, therefore, somewhat correlated. the worst performer is p with a sharpe measure of20.0638. this statistic is much lower than the s & p 500 sharpe measure of 0.2140. we fail to reject the null hypothesis that the average performance of international indexes is equal to or less than that of the s & p 500index in terms of the sharpe measure (excess return to variability). table 1 returns for international indexes and s & p 500 index average monthly return standard deviation developed markets standard and poor’s 500 (s & p 500) 1.20% 0.0348 europe australasia far east (eafe) 0.50% 0.0496 world ex-u.s. (w) 0.50% 0.0481 pacific (p) 0.04% 0.0662 emerging markets europe & middle east (eme) 0.80% 0.0833 emerging markets free (emf) 1.09% 0.0621 emerging markets global (emg) 1.37% 0.0612 combined markets americas free (af) 1.21% 0.0345 ac world ex-u.s. (acw) 0.50% 0.0466 asia pacific ex-japan (ap) 0.80% 0.0520 average international (without s & p 500) 0.76% 0.0042 32 s. aiello, n. chieffe / financial services review 8 (1999) 27–35 on the basis of the treynor measure four indexes, eme, af, emerging markets free (emf), and emg, outperform the s & p 500. however, the average index has a 0.0050 treynor measure, lower than the s & p 500treynor measure of 0.0074. we fail to reject the null hypothesis that the average performance of international indexes is equal to or less than that of the s & p 500index in terms of the treynor measure (excess returns to nondiversifiable risk). however, four individual indexes outperform the s & p 500. the results of the jensen measure of abnormal performance are very similar to those of the treynor measure. af and the emerging markets indexes, eme, emf, emg, outperform the s & p 500 interms of the jensen measure. the s & p 500outperforms the average of the international indexes which was only 1.2692. we fail to reject the null hypothesis that the average performance of international indexes is equal to or less than that of the s & p 500 index in terms of the jensen measure. however, when we test individual indexes against the s & p 500, three emerging market indexes outperform the domestic benchmark: the emg, emf, and eme indexes. these results agree with sinquefield (1996) who found that investing in eafe and similar indexes does not diversify a u.s. portfolio. table 3 lists the results of the regression of the international indexes on the s & p500: esp5005 a 1 eap 1 eeme 1 eacw1 eeafe1 ew 1 ep 1 eemg 1 eemf 1 « (4) wheree is the excess returns (monthly index return2 monthly t-bill return). we eliminated af as an independent variable because it is highly correlated to the dependent variable. the variation in eafe, w (world ex-us), p, and emf significantly explains some variation in the s & p 500index. eafe and p are negatively correlated to the s & p 500; w and emf are positively correlated to the s & p 500. table 4 lists the results of nine regressions of the individual international indexes on the s & p 500: esp5005 a 1 ei 1 « (5) wherei is one of the international indexes. table 2 performance measures for international indexes and s & p 500 index sharpe treynor jensen rank rank rank s & p 500 0.21 2 0.01 4 1.89 5 ap 0.07 5 0.00 6 1.13 6 eme 0.04 6 0.01 1 3.71 1 af 0.22 1 0.01 5 1.92 4 acw 0.01 7 0.00 7 0.17 7 eafe 0.01 9 0.00 9 0.15 8 w 0.01 8 0.00 8 0.14 9 p 20.06 10 20.01 10 21.85 10 emf 0.10 4 0.01 3 2.89 3 emg 0.15 3 0.01 2 3.19 2 average international 0.06 0.01 1.27 33s. aiello, n. chieffe / financial services review 8 (1999) 27–35 the regression of af indicates a high degree of correlation; the others explain only a small portion of the variation in the s & p 500. clearly, diversification potential exists when investing internationally. the variation in the above indexes explains between 0.31% and 94% of the variation in the s & p 500. this shows that one can diversify with many of the indexes. selecting a fund that is based on an index that has little correlation to the u.s. market will allow a portfolio to grow consistently and steadily. 5. conclusions and applications the results of this study suggest that investment into international mutual funds that are based on international indexes may offer significant diversification benefits. however, the performance of the average international index does not outperform the s & p 500benchmark. the sharpe, treynor, and jensen measures prove that the emerging market indexes do outperform the s & p 500. applying this study to the current investment climate is extremely important and useful table 3 regression of the international indexes on s & p 500 index coefficient t-statistic constant 0.003 1.098 ap 20.016 20.199 eme 20.050 21.501 acw 21.959 21.475 eafe 29.582 25.906* w 12.856 6.157* p 20.599 25.046* emg 0.153 1.440 emf 0.143 1.929* * statistically significant at the 95% level. table 4 individual regressions of international indexes on s & p 500 index constant coefficient t-statistic r2 ap 0.0086 0.3413 6.1680* 0.2438 eme 0.0112 0.0239 0.6017 0.0031 af 20.0008 1.0086 43.1363* 0.9404 acw 0.0096 0.3603 5.7459* 0.2186 eafe 0.0098 0.3178 5.2955* 0.1920 w 0.0097 0.3379 5.5116* 0.2047 p 0.0113 0.1664 3.5008* 0.0941 emf 0.0093 0.1932 3.8442* 0.1113 emg 0.0083 0.2275 4.6052* 0.1523 * statistically significant at the 99% level. 34 s. aiello, n. chieffe / financial services review 8 (1999) 27–35 because more people are investing in mutual funds. international index funds provide another alternative for investors. this study urges caution for those investors who seek to maximize returns. however, it does suggest that the diversification one can gain from international index funds is significant and important. references apap, a., & collins, h. (1994). international mutual fund performance: a comparison.manag finan 20(4), 47–54. bogle, j. c. (1998). the implications of style analysis for mutual fund performance.j portf manag 24(4), 34–42. cumby, r. e., & glen, j. d. (1990). evaluating the performance of international mutual funds.j finan 45, 497–521. divecha, a. b., drach, j., & stefek, d. (1992). emerging markets: a quantitative perspective.j portf manag 19(1), 41–50. droms, w. g., & walker, d. a. (1994). investment performance of international mutual funds.j finan res 17, 1–14. eun, c. s., kolodny, r., & resnick, b. g. (1991). u.s.-based international mutual funds: a performance evaluation.j portf manag 17, 88–94. gruber, m. j. (1996). another puzzle: the growth in actively managed mutual funds.j finan 51(3), 783–810. ho, k., milevsky, m.a., & robinson, c. (1999). international equity diversification and shortfall risk.finan serv rev 8(1), 13–27. hogan, p. h. (1994). portfolio theory creates new investment opportunities.j finan plan 7(1), 35–37. hunter, j. e. & coggin, t. d. (1990). an analysis of the diversification benefit from international equity investment.j portf manag 17(1), 33–36. jensen, m. c. (1968). the performance of mutual funds in the period 1945–1964.j finan 23, 389–416. khanna, a. (1996). equity investment prospects in emerging markets.j world bus 31(2), 32–39. malhorta, d. k., & mcleod, r. f. (1997). an empirical analysis of mutual fund expenses.j finan res 20, 175–190. masters, s. j. (1998). the problem with emerging market indexes.j. portf manag 24, 93–100. russell, j. (1998). the international diversification fallacy of exchange-listed securities.finan serv rev 7(2), 95–106. sharpe, w. (1966). mutual fund performance.j bus 39(1), 119–38. sinquefield, r. (1996). where are the gains from international diversification?finan anal j 52(1), 8–14. solnik, b. h. (1995). why not diversify internationally rather than domestically?finan anal j 51(1), 89–94. speidell, l. s., & sappenfield, r. (1992). global diversification in a shrinking world.j portf manag 19(1), 57–67. treynor, j. (1965). how to rate management of investment funds.harv bus rev 43(1), 63–75. 35s. aiello, n. chieffe / financial services review 8 (1999) 27–35 pii: s1057-0810(99)80010-9 financial services review, 7(1): 11-23 issn: 1057-0810 copyright © 1998 by jai press inc. all rights of reproduction in any form reserved. a tax-free exploitation of the turn-of-the-month effect: c.r.e.f. robert a. kunkel and wil l iam s. c o m p t o n by applying knowledge of the "turn-of-the-month" effect investors will improve the risk-adjusted performance of their retirement accounts by using a simple and easily implemented "switching" strategy. our exploitation of the turn-of-the-month anomaly achieves a 17. 7 percent average annual rate of return by switching between a money market account and a broad market indexed stock account. this is compared to a 15.6 percent average annual rate achieved by simply buying and holding the stock account, or a 5.8 percent rate on the money market account. additionally, volatility is cut in half and there are no tax consequences or transactions fees when the switching strategy is used within a retirement account. our results suggests that this strategy might be suc cessfully implemented, under current tax laws, in qualified retirement plans and in variable annuities. i. introduction the turn-of-the-month effect in stock returns has received much attention recently, espe cially by those who have attempted to document opportunities to exploit this apparent mar ket anomaly. in one recent study, henzel and ziemba (1996) demonstrate a trading strategy which achieves superior performance by switching between an interest beating cash account and the s&p 500 index around the turn-of-the-month. while ignoring transfer costs and the tax consequences, they claim the results would appeal to institutional inves tors concerned with the timing of purchases and sales. this paper examines whether indi vidual investors can exploit the turn-of-the-month effect in retirement accounts and variable annuities. by applying a similar switching strategy in a tax-deferred, no transfer cost retirement fund we find that individual investors can exploit the turn-of-the-month effect and earn superior risk-adjusted returns while avoiding the transfer costs and tax con sequences of account switching. william s. compton • department of marketing & finance, western illinois university, macomb, il 61455-1390. fax: (309) 298-2198; e-mail: ws-compton@wiu.edu. robert a. kunkel • school of business, eastern illinois university, charleston, il 61920-3099. fax: (217) 581-6247; e-mail: cfrak@eiu.edu. 12 financial services review 7(1) 1998 ii. literature review evidence of seasonal anomalies in stock returns has generated considerable public interest in recent years and a significant amount of research has been devoted toward documenting their existence and potential for generating abnormal returns. much of the empirical evi dence suggests that these abnormal returns are economically insignificant once transac tions costs and tax consequences are considered. the january effect, in which average stock returns are higher in january than in any other month, is perhaps the best known and most extensively documented seasonal anom aly. the persistence of this phenomenon over the years, despite the attention it has gener ated in the popular press, has been reaffirmed in innumerable academic articles since it was first observed more than 50 years ago (wachtel, 1942) and rediscovered more recently (rozeff & kinney, 1976). among the possible explanations proposed by wachtel are year end selling of stocks for tax loss purposes and a "general feeling of good fellowship and cheer" during christmas holidays (wachtel, p. 186). the tax-loss selling hypothesis is gen erally considered the most likely explanation for the january rebound and has received the strongest support in the academic literature. early work by keim (1983), roll (1983), and reinganum (1983) link the observed january seasonal to small firm return patterns in january and this connection has been rein forced by others. lakonishok and smidt (1984), ritter (1988), and johnston and cox (1996) find that tax motivated trading of small capitalization stocks by individual investors drives the january rebound. haugen and lakonishok (1987), and ritter and chopra (1989) suggest that portfolio rebalancing (or window dressing) by professional portfolio managers to clear smaller, lesser known companies off the books is another likely source of the apparent anomaly. several studies attribute the january effect to variation in risk premia or expected returns and suggest either that the assumed positive risk-return tradeoff is restricted to small stocks in january (tinic & west, 1984) or that smaller stocks are simply riskier in january than at other times of the year (chan, chen, & hsieh, 1985; rogalski & tinic, 1986). the information hypothesis proposed by ritter (1988) suggests that informed investors are perceived to have a larger relative advantage when trading after the turn-of the-year as management becomes aware of non-public information. since individual inves tors tend to buy a disproportionate number of smaller stocks this informational advantage results in an increase in the volatility and systematic risk of smaller stocks in january. in an extensive study of equally weighted deciles of stocks on the nyse from 1926 through 1993, haugen and jorion (1996) find that the january premium persists for all but the largest decile of stocks with no significant reduction in magnitude. however, more recent evidence indicates that the january effect may be running its course, except for the smallest capitalization stocks. riepe (1998) documents a diminishing of the effect by examining market-value weighted deciles of stocks between 1926 and 1997. the author concludes that the introduction of futures on the s&p 500 and value line in 1982, and especially the creation of futures contracts on the russell 2000 by the chicago merchantile exchange in 1993, has contributed to the market's ability to exploit the effect. this, and other evidence (star, 1996), suggests that opportunities for exploiting the effect with small cap stocks are limited by liquidity, transactions costs, and the availability of futures. the monday seasonal in equity returns, also referred to as the weekend effect, is another apparent anomaly which continues to generate interest and research. early evi dence that monday returns are negative and significantly different from other daily returns a tax-free exploitation 13 is presented in cross (1973), french (1980), and gibbons and hess (1981). keim and stambaugh (1984) document 55 years of the effect on the s&p 500 and flannery and pro topapadakis (1988) find the pattern is shared by 11 asset groups, including treasury secu rity of varying maturities and three stock market indices. possible explanations include the timing of corporate news announcements, trading patterns of institutional and individual investors, and settlement procedures. in studies of corporate news announcements patell and wolfson (1982) and penman (1987) find that bad news announcements are more likely to occur during closed market periods or on mondays. the relationship between negative markets and the monday effect has been substantiated by dyl and maberly (1988), fishe, gosnell, and lasser (1993), athanassakos and robinson (1994), and others. abraham and ikenberry (1994) find that positive (negative) monday returns follow positive (negative) fridays and suggest that the monday effect is a consequence of both corporate news announcements and a monday bias toward sell transactions by individual investors following weekend decision making. a number of studies identify an imbalance in buy and sell orders on mondays caused by the trading activity of individual and institutional investors as a possible source of the weekend seasonal (lakonishok & maberly, 1990; lakonishok, shleifer, & vishny, 1992; miller,1988). these studies suggest that the monday effect results from a combination of factors, including a selling bias by individual traders and reduced institutional trading activity on mondays when brokers are planning strategy for the week. recent evidence that the monday effect is weakening, at least for large capitalization stocks, indicates trading to exploit the anomaly. in an article examining daily returns from 1962-1993, kamara (1997) finds that the monday seasonal disappears after 1982 for stocks in the standard and poor's (s&p) 500 but continues to be present in the smallest decile of stocks on the nyse, reflecting the ability of institutional investors to trade against the effect. another recent study by chow, hsiao, and solt (1997) also reports a diminishing of the effect between 1970 and 1993 for those stocks with lower transactions costs. a lesser known and more recently discovered seasonal pattern is the monthly calendar anomaly known as the turn-of-the-month effect. it has been identified in a number of stud ies in both individual stocks and in various stock market indices. ariel (1987) first reported a monthly seasonal pattern in the returns of equally-weighted and value-weighted stock portfolios between 1963 and 1981, using data obtained from the center for research in security prices (crsp). in that study, stock returns in the first half of the month, identified as the first nine trading days of the month plus the last trading day of the previous month, are considerably higher than stock returns in the second half of the month, identified as the last eight trading days of the month, exclusive of the last trading day. this pattern exists in both large and small capitalization stocks and is independent of other known calendar anomalies, such as the january effect. ariel notes that the phenomenon is especially strong in the five day period between the last trading day of one month the fourth trading day of the next month (trading days -1 through +4). in a subsequent study of various seasonal patterns on the dow jones industrial aver age over a ninety year period between 1897 and 1987, lakonishok and smidt (1988) dis cover a persistent monthly seasonal limited to trading days -1 through +3. the authors suggest that the monthly jump in returns may be liquidity driven and a result of the buying and selling activity of pension fund managers around the turn-of-the-month. ogden (1990) provides evidence that the anomaly is driven by liquidity and suggests that a "standardiza 14 financial services review 7(1) i 998 tion of payments system" in the united states is responsible for the monthly seasonal. cash receipts such as wages, dividends, interest, and principal payments at the end and begin ning of the calendar month are quickly reinvested, resulting in a surge in stock returns. examining value (equally)-weighted stock indices for the eighteen year period from janu ary 1969 through december 1986 the author reports an average cumulative return during the turn-of-the-month of 0.5132 (0.8468) percent. however, with round trip transactions costs conservatively estimated at 0.46 percent, a speculator is unlikely to generate suffi cient profits from trading on this information. henzel and ziemba (1996) demonstrate how the monthly seasonal can be exploited, using data from the s&p 500 index between 1928 and 1993 and the trading days identified by ariel (trading day -1 to +4), by switching between the index and an interest bearing cash account at the turn-of-the-month. the strategy produces an average annual return of 10.13 percent over the sixty-five-year period of the study, compared to an average annual return of 9.50 percent for a simple buy-and-hold strategy on the index. they also demon strate that the effect is not the result of a few "significant days" and that large gains and losses are proportionally distributed between the turn-of-the-month period and the rest of the month. an even stronger turn-of-the-month effect is identified for the five days between trading days -2 and +3. this shifting of the critical turn-of-the-month period has been observed in other studies and may be related to futures trading in anticipation of the effect (henzel, sick, & ziemba, 1994). this study builds on henzel and ziemba by testing whether individual investors can exploit the turn-of-the-month effect, and avoid the transactions costs, by implementing the switching strategy in a tax-deferred, no cost retirement plan. if this pattern in returns is driven by liquidity, pension fund buying, and futures activity, then it should be observable in most broad stock market accounts available in retirement plans and variable annuities. by timing trades to the turn-of-the-month period, individual investors could enhance the performance of retirement plans and annuities which allow frequent transfer at no cost. the rest of the paper proceeds as follows. the next sections describe the data used in the study and our efforts to determine if the turn-of-the-month effect exists in the data. this is fol lowed by empirical tests of the switching strategy, the results, and our conclusions based on these results. iii. data and methodology the retirement fund data used in this study were obtained by the authors from the corporate office of the teachers insurance and annuity association-college retirement equities fund (tiaa-cref). tiaa-cref is a nationwide retirement system for employees at col leges, universities, and other nonprofit educational and research institutions in the united states. it is the largest private retirement system in the world, reporting assets in excess of $200 billion as of december 31, 1997. tiaa was established in 1918 as a nonprofit orga nization for the advancement of teaching as a profession and to provide life insurance, pen sion products, and long-term disability and care insurance. cref was established in 1952 as a companion, nonprofit open-end investment company offering new financial products, including the world's first variable annuity. as of june, 1998 cref offers one traditional a tax-free exploitation 15 annuity, four equity accounts, two fixed-income accounts, a money market account, a social choice account, and a real estate account. the study begins on april 1, 1988, the day that cref introduced its money market account, and ends on the last day of december, 1997. we examine four sample periods, one nine-year period and three sub-periods, in which to evaluate the trading strategy. returns for the switching strategy and the buy-and-hold strategy are computed using the daily closing unit values for the stock account and money market account. the raw data were transformed into daily holding period returns from april, 1988 through december, 1997. we also examine two market indices, the dow jones industrial average and standard & poor's 500 index, for evidence that the monthly seasonal identified in earlier studies continues to exist. we obtain daily closing values for the dow jones industrial average (djia) from the dow jones averages 1885-1995, (1996), the wall street journal index (1996), and the dow jones averages: the market's measure--dow data, [online](1997). closing values for the standard & poor's 500 index (s&p 500) are obtained from standard & poor's security price index record (1996) and standard & poor's current statistics (january, 1998). finally, the treasury bill rates used to calculate excess returns in the risk-adjusted performance measures are obtained from the federal reserve bank of st. louis, research division, (1997), federal reserve economic data (fred) [online] data files of daily historical three-month treasury bill rates. the switching strategy we investigate involves the transfer of funds between a cash account and a market portfolio. the money market a~count is selected as the cash account because it is the only money market fund available to eligible participants of cref. the stock account, cref's flagship fund, is selected as the market portfolio because it is the most broadly diversified equity account offered by cref, with almost 80 percent of the account indexed to the u.s. stock market. it is the best proxy for the market portfolio that is available through cref, as evidenced by the fact that the correlation of returns between the stock account and the s&p 500 index is 0.972 during the study. the switching strategy is similar to the one explored by henzel and ziemba (1996). funds are moved from the money market account to the stock account at the beginning of the turn-of-the-month period (hereafter, simply tom), and then switched back to the money market account on the last day of the tom period. all transfers between accounts are executed by cref at the accumulated unit values at the close of that business day. par ticipants in cref can place an order to transfer funds from any account, 24 hours a day, seven days a week, using cref's automated telephone service or the internet advanced communication and transaction system. there are no restrictions on transfers between accounts and no transaction fees for the transfer. finally, since these accounts are part of a qualified retirement plan, no tax liabilities are triggered. although the study is implemented using cref data, it has implications for invest ments in variable annuities and in other tax-deferred retirement plans, such as iras, 401 (k)s, 403(b)s, which allow frequent trading. a recent article in a leading financial news magazine states that "... with no tax consequences and little or no transaction costs, the 401(k) offers the best possible environment for rapid-trading, market-timing tactics" (wil cox, p. 53). since most retirement plans, such as 401(k)s and 403(b)s, offer a stock account and money market account, individual investors would have a practical, operational way to exploit the tom effect in a tax-deferred, no cost, unlimited switching setting. 16 f inancial services review 7(1) 1998 in the next section we examine the daily returns of the djia, s&p 500, and the stock account over the period from april, 1988 through december, 1997. we demonstrate that the tom calendar anomaly exists in the stock account and continues to be present in the two broad market indices. we then proceed to test the switching strategy in section v. iv. initial evidence initial examination of the data indicates the tom pattern exists in the stock account, which is our proxy for the market portfolio, and continues to exist in the djia and the s&p 500 index. table 1 shows average daily returns and t-statistics for the eighteen days around the turn-of-the-month for the stock account and for the two indices. in the table, trading day 1 is the first trading day of the month, trading day -1 is the trading day just prior to the first trading day of the month, etc. the results in table 1 show a strong tom effect in the stock account, significant at the 10 percent level at least, extending over a six trading day period between trading days --4 and +2. no other trading day in the month shows a return that is significantly greater than zero at even a 10 percent level. the average daily return achieved during the six critical days in the stock account represent an annualized table 1 average daily returns (%) and t-values during the turn-of-the-month period for the djia, s&p 500, and cref stock account (april 1988 to december 1997) djia s&p 500 cref-stock average average average trading day return (%) t-value return (%) t-value return (%) t-value -9 -0.0366 -0.43 -0.0568 -0.73 -0.0662 -0.95 -8 -0.0637 -0.84 -0.0244 -0.33 -0.0369 -0.57 -7 -0.1027 1.34 -0.0905 1.32 -0.0941 1.61 -6 0.0482 0.59 0.0341 0.46 0.0315 0.51 -5 0.0360 0.38 0.0101 0.11 0.0147 0.19 --4 0.1368 1.64 0.0963 1.28 0.1121" 1.92 -3 0.0750 1 . 1 2 0.1383"* 2.13 0.1268** 2.34 -2 0.1203 1.62 0,1290* 1.72 0.1479* * 2.33 i 0.0981 1 . 4 0 0.1733** 2 . 5 1 0.2414*** 4.44 1 0.2398** 3.14 0.2209** 2.96 0.1817** 2.93 2 0.2426** 3.18 0.1998"* 2.64 0.2315'** 3.64 3 0.0198 0.28 0.0280 0.41 0.0686 1.16 4 -0.0313 -0.44 -0.0533 -0.77 -0.0178 -0.29 5 -0.0035 -0.05 0.0095 0.14 0.0011 0.02 6 0.0796 1.27 0.0393 0.58 0.0373 0.65 7 0.0409 0.58 0.0171 0.25 0.0149 0.26 8 -0.0377 -0.50 -0.0197 -0.28 -0.0172 -0.28 9 0.0820 0.92 0.0741 0.86 0.0463 0.65 notes: t-values test the null hypothesis that the average daily return is not significantly different from zero. significance levels are for one-tailed tests. ***significant at the 1% level; **significant at the 5% level; *significant at the 10% level. trading day represents trading days around the turn-of-the-month. trading day 1 is the first trading day of the month; trading day 1 is the day prior to the first trading day of the month. the turn-of-the-month effect begins on trading day -4 and extends through trading day +2, a tax-free exploitation 17 table 2 average daily returns (%) and daily return standard deviations (%) for the dow jones industrial average (djia), the standard and poor's 500 index (s&p 500), the stock account (stock), and the money market account (mmkt). figures are for the six trading days around the turn-of-the-month (tom), for the rest of the month (rom), and for the entire month (month) tom rom month average standard average standard average standard return deviation return deviation return dev ia t ion f-value djia 0.1521 0.8086 0.0228 0.8535 0.0596 0.8428 11.87"** s&p 500 0.1596 0.7828 0.0159 0.8104 0.0569 0.8051 16.09"** stock 0.1736 0.6432 0.0145 0.6931 0.0598 0.6829 27.55*** mmkt 0.0233 0.0177 0.0222 0.0161 0.0225 0.0166 2.39 notes: the turn-of-the-month (tom) days cover the period between the first two days of the month and the last four days of the previous month. the f-value is a test of the null hypothesis that the average daily return during the turn-of-themonth is equal to the aver age daily return during the rest-of-the-month. the average daily returns are arithmetic mean returns for the period between april 1988 through december 1997. there are 2,465 total returns, with 702 occurring in the tom period and 1763 occurring in the rom period. ***significant at the 1% level. compound return of approximately 55 percent while the rest of the month (hereafter, sim ply rom) represent an annualized return of less than 4 percent. in the two indices, the first two trading days of the month show positive average returns that are significant at the 5 percent level. additionally, the s&p 500 shows positive and significant returns on the last three trading days of the month. no other day of the month shows a significant positive average return for either index. there is additional evidence of a tom effect in table 2, which includes average daily returns, return standard deviations, and an analysis of variance between the tom and the rom periods for the two cref accounts and the two indices. the results in the table are consistent with earlier studies. the average daily return for the s&p 500 index in our sam ple during the tom period is 0.1596 percent. henzel and ziemba (1996) report an average daily return for the s&p 500 index for the tom period in their study of 0.1236 percent. odgen (1990) reports an average daily return on the crsp value-weighted index during the tom period of 0.1283 percent. the average daily return for the s&p 500 during the rom period in this study is 0.0159 percent, compared to 0.0137 percent average daily return in the ogden sample, and minus 0.0235 percent average daily return reported by henzel and ziemba. the results also clearly show that the tom effect exists in the stock account. the average daily returns in the tom period, rom period, and during the entire month are 0.1736 percent, 0.0145 percent, and 0.0598 percent, respectively. with approximately 21 trading days in the average month, the six days of the tom period account for approxi mately 83 percent of the average month's total return for the stock account. finally, the f-value in the analysis of variance shows the tom period is significantly different from the rom period at the 0.01 level for the stock account and for both indices. the relevant issue now is what the effect will be on the investor's portfolio when attempts are made to exploit the monthly pattern that we have identified. the next section examines 18 f i n a n c i a l s e r v i c e s r e v i e w 7(1) 1998 whether individual investors can exploit the tom effect by implementing a switching strategy with the stock account and money market account. v. performance results using the switching strategy in this section we examine the performance results of a traditional buy-and-hold strategy versus a switching strategy based on the tom pattern that we have identified in the stock account. the buy-and-hold strategy is implemented by placing $1,000 in the stock account on april 4, 1988. the markets are closed on april first and weekends, making april 4 the first trading day of the month for that year. this money is left in the stock account for the duration of the study. the switching strategy is also implemented by plac ing $1,000 in the stock account on april 4. all accumulated funds are then moved to the money market account at the close of business on april 5, the second trading day of the month. finally, accumulated funds are switched back from the money market account into the stock account at the close of business on the fifth trading day prior to the end of the month. this switching of funds is repeated each month, moving funds into the stock account during the tom period and back into the money market account during the rom period, through the end of december, 1997. table 3 average daily compound return (%), return standard deviation (%), beta, sharpe, treynor, and appraisal ratios for the period between april 1988 and december 1997 and for three sub-periods period average return growth of daily standard sharpe treynor appraisal $1,000 return deviation beta ratio ratio ratio investment 4/198812/1997 mmkt 0.023 0.017 . . . . $1,740 stock 0.057 0.683 0.824 0.0625 0.00052 0.0638 $4,118 switch 0.065 0.350 0.219 0.1429 0.00228 0.1367 $4,921 4/198812/1991 mmkt 0.030 0.019 . . . . $1,331 stock 0.057 0.793 0.852 0.0484 0.00045 0.0809 $1,724 switch 0.085 0.389 0.210 0.1689 0.00313 0.1741 $2,233 1/199212/1994 mmkt 0.014 0.010 . . . . $1,112 stock 0.025 0.509 0,817 0.0299 0.00019 0.0791 $1,209 switch 0.043 0.278 0.247 0.1205 0.00136 0.1386 $1,388 1/199512/1997 mmkt 0.021 0.013 . . . . $1,176 stock 0.090 0.684 0.787 0.1105 0.00096 0.0481 $1,976 switch 0.061 0.362 0.220 0.1293 0.00213 0.0898 $1,588 notes: annualized returns for the money market account (mmkt) and the stock account (stock), are 5.8 percent and 15.6 percent, respectively. annualized return for the switchin 8 strategy (switch) is 17.7 percent. there are 2,465 daily returns between april 1988 through december 1997. betas for the stock account and the switchin 8 strategy are calculated by regressing the dally returns for each onto the daily return for the s&p 500 index for the entire period and each subperiod. a tax-free exploitation 19 $6,000 $5 ,~ t $4,000 ~ $3,000 j j y $1,00o month-year i ~switch i ~mmr'f figure 1. growth of $1,000 between april 1988 and december 1997 in the money market account (mmkt), the stock account (stock), and the switching strategy (switch) performance results for the switching and the buy-and-hold strategies indicate that the switching strategy clearly outperforms the buy-and-hold strategy. table 3 shows various summary statistics and risk-adjusted performance measures for the entire period and for the three sub-periods. between 1988 and 1997 the switching strategy achieves an average daily compound return of 0.065 percent over the 2,465 trading days, compared to an aver age daily compound return of 0.057 percent for the buy-and-hold strategy and a 0.023 per cent for the money market account. this means that, beginning with a $1,000 investment, the accumulated value grows to $4,921 over the 117 months of the study in the switching strategy, compared to $4,118 for a similar $1,000 investment in the buy-and-hold strategy. the growth of a $1,000 investment in both strategies and in the money market account is shown graphically in figure 1. not included in the table are the results from combining the money market account with the djia and the s&p 500 in a switching strategy. over the entire period of the study, a $1,000 invested in the djia grows to $4,198 with the strategy, versus $3,978 without the strategy. a $1,000 investment in the s&p 500 index grows to $4,432 with the strategy, and $3,748 without. funds in the switching strategy are held in the low risk money market account over seventy percent of each month, significantly reducing the investment's risk as measured by total volatility of return. return standard deviation for the switching strategy is reduced to almost half that of the buy-and-hold strategy, from 0.683 percent to 0.35 percent. betas for both strategies are calculated by regressing returns onto the s&p 500 index. we find a reduction in this measure of market risk from 0.82 for the stock account to 0.22 for the switching strategy. also reported in table 3 are the summary results for each sub-period that we examined. the risk-adjusted performance measures included in table 3 are the sharpe ratio, the treynor ratio, and the appraisal ratio. while each is consistent with conventional mean 20 financial services review 7(i) 1998 variance optimization, they differ in their assumptions about the investor's complementary portfolios. the sharpe ratio measures average excess return per unit of total risk as mea sured by standard deviation of return (sharpe, 1966). it is appropriate when a portfolio rep resents the investor's entire investment and is being compared to a benchmark or another portfolio. a higher ratio indicates superior risk-adjusted performance. over the nine years of the study this ratio is 0.14 for the switching strategy, compared to 0.062 when simply buying-and-holding the stock account. the treynor ratio, measuring average excess return per unit of systematic risk as mea sured by beta, is appropriate when a portfolio is held in combination with other portfolios in a larger investment fund (treynor, 1965). in this case mean excess return should be weighed against the portfolio's systematic risk and a higher ratio indicates superior risk adjusted performance. over the nine years of the study this measure is 0.0023 for the switching strategy and 0.00052 for the buy-and-hold strategy. the third investment assumption has the investor holding the strategy in combination with a passive market index. when the two portfolios are optimally mixed, the appropriate performance measure is the appraisal ratio. this measures the abnormal return of the active portfolio relative to the passive portfolio, per unit of diversifiable risk (bodie, kane, and marcus, 1996). abnormal return is computed as the intercept (jensen's alpha) in a regres sion of the portfolio's excess returns onto the excess returns of the market. the s&p 500 index is used as the proxy for the market in this regression. the standard deviation of the residuals in the regression is the measure of diversifiable risk. a higher ratio indicates superior risk-adjusted performance. over the nine years of the study the appraisal ratio is 0.1367 for the switching strategy and 0.0638 for the buy-and-hold strategy. finally, henzel and ziemba note that low correlations between the switching strat egy's returns and returns on alternative investments provide additional diversification ben efits when the strategy is used in combination with other assets. they report a correlation of 0.46 between the switching strategy and the s&p 500 index. we find that the correlation of the stock account with the s&p 500 index drops from 97 percent to 50 percent when the stock account is combined with the money market account in the switching strategy. based on the three different measures of risk-adjusted performance and the low correla tions, the switching strategy is the superior investment strategy for all investors. sub period results are reported in table 3 and are consistent with the overall period. vi. conclusion in this study we test whether individual investors can exploit the turn-of-the-month effect with a simple switching strategy in a tax-deferred, no cost retirement plan. using daily closing prices for a money market account and a broadly diversified stock account, we find that individual investors who use the switching strategy can improve the performance of their retirement accounts. the superior performance of the switching strategy over a simple buy-andhold strategy is demonstrated using three measures of risk-adjusted performance that are each consistent with a different investment plan or objective. whether the invest ment strategy is intended as the sole investment of a portfolio, whether it is intended to be held in combination with many other investments, or whether it is to be combined with a a tax.free exploitation 21 passive market portfolio, the switching strategy achieves superior risk-adjusted perfor mance under each scenario. a consequence of the switching strategy is the free-rider problem that may exist in the many types of retirement accounts, including qualified retirement plans and variable annu ities, where the strategy could be used. this is potentially a problem when no fees are lev ied on exchanges between accounts and no restrictions are placed on the number of exchanges allowed. although plan participants can earn superior risk-adjusted returns through frequent switching, this increases the fund's operating cost and places an unfair burden on the plan participants who do not employ the switching strategy. to address this problem investment companies may eventually limit the number of transfers between accounts and/or charge a fee for each transfer. there are also public policy issues to consider. efforts in washington to eliminate the tax-exempt status of certain events related to variable annuities, such as account switching, are not uncommon and widespread use of switching strategies may add fuel to the forces that are demanding a new tax on exchanges. however, we must recognize that when such opportunities exist, investors will seek to exploit them. a recent article in the wall street journal suggests: "if you are going to trade and still have some slim hope of beating the market, you really need to do your buying and selling in a retirement account" (clements, p. c 1). as long as current tax laws and investment company policies allow frequent switch ing, investors may consider the turn-ofthe-month switching strategy as an easy way to enhance the performance of their retirement portfolio while at the same time achieving a considerable reduction in risk. references ariel, r. (1987). a monthly effect in stock returns. journal of financial economics, 18, 161-174. abraham, a., & ikenberry, d. (1994). the individual investor and the weekend effect. journal of financial and quantitative analysis, 29, 263-277. athanassakos, g., & robinson, m. (1994). the day-of-the-week anomaly: the toronto stock exchange experience. journal of business finance & accounting, 21,833-856. bodie, z., kane, a., & marcus, a. (1996). investments (3rd ed.). chicago: irwin. chan, k., chen, n., & hsieh, d. (1985). an exploratory investigation of the firm size effect. journal of financial economics, 14, 451--471. chow, e., hsiao, p., & solt, m. (1997). trading returns for the weekend effect using intraday data. journal of business finance & accounting, 24, 425-444. clements, j. (1998, march 24). if you fancy yourself a stock trader, put your retirement account to work. the wall street journal, p. c 1. cross, f. (1973). the behavior of stock prices on fridays and mondays. financial analysts journal, november-december, 67--69 dow jones & company, inc. (1996). the wall street journal index 1996. ann arbor, mi: umi. dow jones & company, inc. (1997). the dow jones averages: the market's measure dow data. [online]. available: http://averages.dowjones.com/frameset.html. dyl, e., & maberly, e. (1988). a possible explanation of the weekend effect. financial analysts journal, 44, 83-84. federal reserve bank of st. louis, research division. (1997). federal reserve economic data (fred), [online]. available: http://www.stls.frb.org/fred/data/wkly/dtb3 22 financial services review 7(1) 1998 fishe, r., gosnell, t., & lasser, d. (1993). good news, bad news, volume, and the monday effect. journal of business finance & accounting, 20, 881-893. flannery, m., & protopapadakis, a. (1988). from t-bills to common stocks: investigating the gener ality of intra-week return seasonality. journal of finance, 43, 431-450. french, k. (1980). stock returns and the weekend effect. journal of financial economics, 8, 55--69. gibbons, m., & hess, p. (1981). day of the week effects and asset returns. journal of business, 54, 579-596. haugen, r., & jorion, p. (1996). the january effect: still there after all these years. financial ana lysts journal, 52, 27-31. haugen, r., & lakonishok, j. (1988). the incredible january effect: the stock market's unsolved mystery. homewood, il: irwin. henzel, r., & ziemba, w. (1996). investment results from exploiting turn-of-the-month effects. journal of portfolio management, spring, 17-23. henzel, r., sick, g., & ziemba, w. (1994). the turn-of-the-month effect in the u.s. stock index futures markets, 1982-1992. review of futures markets, 13, 827-856. johnston, k., & cox, d. (1996). the influence of tax-loss selling by individual investors in explain ing the january effect. quarterly journal of business & economics, 35, 14-20. kamara, a. (1997). new evidence on the monday seasonal in stock returns. journal of business, 70, 63-84. keim, d. (1983). size-related anomalies and stock return seasonality: further empirical evidence. journal of financial economics, 12, 13-32. keim, d., & stambaugh, r. (1984). a further investigation of the weekend effect in stock returns. journal of finance, 39, 819-835. lakonishok, j., & maberly, e. (1990). the weekend effect: trading patterns of individual and insti tutional investors. journal of finance, 45, 231-243. lakonishok, j., shleifer, a., & vishny, r. (1992). the impact of institutional trading on stock prices. journal of financial economics, 32, 23-43. lakonishok, j., & smidt, s. (1984). volume and turn-of-the-year behavior. journal of financial economics, 13, 435-455. lakonishok, j., & smidt, s. (1988). are seasonal anomalies real? a ninety-year perspective. review of financial studies, 1,403--425. miller, e. (1988). why a weekend effect? journal of portfolio management, 14, 43-48. ogden, j. (1990). turn-of-month evaluations of liquid profits and stock returns: a common explana tion for the monthly and january effects. journal of finance, 45, 1259-1272. patell, j., & wolfson, m. (! 982). good news, bad news, and the intraday timing of corporate disclo sures. accounting review, 57, 509-527. penman, s. (1987). the distribution of earnings news over time and seasonalities in aggregate stock returns. journal of financial economics, 18, 199-228. pierce, p. s. (ed.). (1996). the dow jones averages 1885-1995. chicago: irwin. reinganum, m. (1983). the anomalous stock market behavior of small finns in january: empirical tests for tax-loss selling effects. journal of financial economics, 12, 89-104. riepe, m. (1998). is publicity killing the january effect? journal of financial planning, 11, 64-70. ritter, j. (1988). the buying and selling behavior of individual investors at the turn of the year. jour nal of finance, 43, 701-717. ritter, j., & chopra, n. (1989). portfolio rebalancing and the turn-of-the-year effect. journal of finance, 44, 149-166. rogalski, r., & tinic, s. (1986). the january size effect: anomaly or risk mismeasurement? finan cial analysts journal, 42, 63-70. roll, r. (1983). vas ist das? the turn of the year effect and the return premium of small firms. jour nal of portfolio management, 9, 18-28. a tax-free exploitation 23 rozeff, m., & kinney, w. (1976). capital market seasonality: the case of stock returns. journal of financial economics, 3, 379-402. star, m. (1996). the january effect: dead or just missing in action? pensions & investments, 24, 3, 31. sharpe, w. r. (1966). mutual fund performance. journal of business, 39, 119-138. the tiaa-cref accounts (1997). profiles on performance: 1997. new york: teachers insurance and annuity association college retirement equities fund. tinic, s., & west, r. (1984). risk and return: january vs. the rest of the year. journal of financial economics, 13, 561-574. treynor, j. l. (1965). how to rate management of investment funds. harvard business review, 43, 63-75. wachtel, s. (1942). certain observations on seasonal movements in stock prices. journal of busi ness, 15, 184-193. wilcox, m. (1997). get the max from your 401 (k) plan. kipplinger's personal finance magazine, 51, 52-54. pii: s1057-0810(99)00023-2 mutual fund shareholders: characteristics, investor knowledge, and sources of information gordon j. alexandera, jonathan d. jonesb, peter j. nigroc,* acarlson school of management, university of minnesota, minneapolis, mn 55455, usa brisk management division, office of thrift supervision, washington, dc 20552, usa cpolicy analysis division, office of the comptroller of the currency, washington, dc 20219, usa abstract this paper examines responses from a survey of 2,000 randomly selected mutual fund investors who purchased shares from six different distribution channels. the survey provides data on the demographic, financial, and fund ownership characteristics of mutual fund investors. it also provides data on investors’ knowledge of the costs and investment risks of mutual funds and the information sources these investors use to learn about these costs and risks. our survey results strongly suggest there is room for improvement in the level of financial literacy of mutual fund investors. © 1999 elsevier science inc. all rights reserved. 1. introduction over the past twenty or so years, mutual funds have become an increasingly popular investment vehicle. ownership of stock, bond, and money market mutual funds rose from 6% of u.s. households in 1980 to 42% in 1998, while the total assets held by mutual funds soared by almost 4,000%, increasing from $135 billion to roughly $5.5 trillion at year-end 1998 (investment company institute, 1999). this dramatic growth has raised policymakers’ * corresponding author. tel.:11-202-874-4799; fax:11-202-874-5394. e-mail address:peter.nigro@occ.treas.gov (p.j. nigro) the office of thrift supervision and the office of the comptroller of the currency, as a matter of policy, disclaim any responsibility for any publication or statements by any of their employees. the views expressed herein are those of the authors and do not necessarily reflect the views of either of these agencies or of the authors’ colleagues on the staffs of the agencies. financial services review 7 (1998) 301–316 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00023-2 concern with the level of investor knowledge regarding the costs and risks associated with mutual funds and with the types of distribution channels permitted to sell mutual funds. to provide perspective on these concerns, the office of the comptroller of the currency (“occ”) and the securities and exchange commission (“sec”) contracted with a market research firm to conduct a nationwide telephone survey of a randomly selected sample of 2,000 mutual fund investors (see alexander et al., 1996). the mutual fund survey collected two kinds of data. first, the survey collected data on the demographic, financial, and fund ownership characteristics of mutual fund shareholders. these data permit analysis of how these characteristics differ across the major distribution channels used by mutual fund purchasers. second, the survey collected data on mutual fund investors’ familiarity with certain costs and investment risks associated with mutual funds as well as the information sources these investors used to learn about these costs and risks. in this paper, we provide background information on mutual fund investors, assess their degree of knowledge about the costs and investment risks associated with mutual funds, examine the determinants of their financial literacy, and examine whether or not certain distribution channels (e.g., pension plans or banks) pose unique regulatory concerns. the remainder of this paper is organized as follows. section 1 briefly reviews previous survey research on mutual fund investors. section 2 examines the demographic and financial characteristics of mutual fund shareholders. section 3 examines investor familiarity with the costs and certain investment risks associated with mutual funds, along with the information sources used by these investors in making mutual fund purchases. section 4 develops a measure of overall investor financial literacy and examines its determinants. finally, section 5 discusses the policy implications of the survey results and draws conclusions. 2. previous literature several recent mutual fund investor surveys provide conflicting evidence on investment risk disclosures and the level of investor knowledge. for example, the american association of retired persons et al. (1994) concluded from their survey that “the vast majority of american bank consumers are unaware of the risks and fees involved in the sale of uninsured investment products, such as mutual funds and annuities.” in sharp contrast, the consumer bankers association (1994), a banking trade group, found that few bank customers held the misconception that mutual funds purchased through a bank are federally insured. the conflicting evidence in these two initial surveys generated several other surveys. for example, a 1995 prophet market research mystery shopping study that employed unidentified testers to examine disclosure concluded that banks do a better job than brokerage houses and insurance companies in educating customers about the risk of investment products (kimmelman, 1995). a second round of bank mystery shopping by the same company in january 1996, however, yielded less favorable results about bank sales representatives’ disclosures of the risks, fees, and expenses associated with mutual funds. bank representatives countered that because such disclosures are not typically made until the sales are about to be closed, mystery shoppers would not receive them (plasencia & cope, 1996). the federal deposit insurance corporation (“fdic”) conducted a shopping survey of 302 g.j. alexander et al. / financial services review 7 (1998) 301–316 non-deposit investment sales at fdic-insured depository institutions (market trends, 1996) and found that bank sales representatives were more likely to make required disclosures in face-to-face discussions with investors than over the telephone. concerns about investor understanding of the costs and risks of mutual funds extend beyond investors who obtain their fund shares through a bank-related channel. for example, a recent survey of pension plan participants (mostly 401(k) plan participants) by john hancock mutual life insurance co. reported that more than one-third of the respondents believed it was impossible to lose money in a bond fund (an additional 12% were not sure), while 12% believed it was impossible to lose money in a stock fund or said they did not know (schultz, 1995). more generally, a survey commissioned by the investor protection trust (crenshaw, 1996) found that fewer than one-fifth of all individual investors (in stocks, bonds, funds, or other securities) could be considered “financially literate” based on their responses to a quiz. furthermore, chen and volpe (1998) found that a large percentage of college students are not knowledgeable about personal finances. finally,moneymagazine and the vanguard funds group jointly conducted a 20-question survey of 1,467 mutual fund investors and found that most investors have inadequate knowledge about their mutual fund investments (updegrave, 1996). the occ/sec survey focuses on investor knowledge rather than disclosure. one key distinction from previous surveys, however, is that detailed information on the type of distribution channel used in purchasing mutual funds was collected. this permits an examination of the differences in the demographic and financial characteristics of purchasers, as well as differences in the degree of financial literacy, by distribution channel. 3. demographic and financial characteristics of mutual fund investors the major demographic characteristics considered in the survey included age, income, education, and gender. purchasers from six distribution channels, including stockbrokers (both full-service and discount), commercial banks (both banks and savings associations, hereafter banks), mutual fund companies, insurance companies, employer-sponsored pension plans, and “other” (e.g., financial planners) were examined. these distribution channels are not mutually exclusive. that is, an investor who purchases a mutual fund directly from a fund company may purchase another one from a bank or a brokerage firm. as a result, the percentages reported in any given row for the following tables often sum to more than 100% and the chi-squared statistics in the tables test for significant differences between bank and non-bank purchasers, broker and non-broker purchasers, and so on. panel a of table 1 shows the number of respondents for each of the six channels. note that while there were 2,000 respondents to the survey, the sum of the respondents in the channels is 3,232 (summing across the row), indicating there are a large number of multiplechannel purchasers. it should also be noted that not all respondents provided answers to all questions, so the number of responses can vary by question. panels b and c of table 1 show both demographic and financial data on investors. as shown in panel b, 58.6% of survey respondents were males. investors who purchased mutual funds directly from a fund company were significantly more likely to be male (69.4%), while 303g.j. alexander et al. / financial services review 7 (1998) 301–316 bank purchasers (50%) were equally divided between male and female, indicating that banks reach a somewhat different segment of the population than that reached by other mutual fund providers. the median age of a mutual fund shareholder in the survey is 43 years. younger investors are significantly more likely to invest in mutual funds through their pension plans (e.g., 401(k) plans), reflecting the increased usage of defined contribution plans by employers in recent years. in terms of income, mutual fund investors have a median household income of $58,800, which is close to the median household income of fund owners reported elsewhere (investment company institute, 1999). mutual fund purchasers using brokers, those buying through pension plans, and those buying directly from the fund company report notably higher median incomes than those purchasing through other distribution channels. finally, in terms of education, mutual fund investors are well educated, with 54.6% having at least completed college. broker (62.8%) and direct fund company (68.5%) customers are more likely to have at least a college degree than customers in the other distribution channels, while bank (49.3%) customers are less likely. table 1 demographic characteristics and financial experience of respondents distribution channel used bank broker pension direct insurance other total a. number of respondents 294 638 1,118 569 521 92 2,000 b. demographic characteristics male 50.0% 62.5% 62.3% 69.4%* 54.9% 57.6% 58.6% median age 45* 47* 41* 44* 44 44 43 median income $55,200 $67,600* $62,100* $67,000* $59,200* $58,400 $58,800 college grad. 49.3%* 62.8%* 57.5% 68.5%* 55.3% 52.2% 54.6% c. financial characteristics seasoned investor1 85.2% 91.1%* 85.0% 89.7%* 90.5%* 83.5% 85.2% individual stocks 44.6* 72.6* 51.8 58.4 47.4 42.4 50.8 individual bonds 34.4 39.0* 30.4 33.4 34.4 29.4 31.1 cds 47.6* 41.7* 30.8* 34.3 36.3 28.3 34.9 money market 50.7* 46.2* 36.5* 37.3 36.3 38.0 38.3 deposit account annuities 31.0 31.0* 25.1 25.0 45.5* 25.0 26.7 primary residence 77.6 88.6* 81.0 82.1 84.6* 71.7* 80.9 1 purchased mutual fund prior to 1993. notes:because the distribution channels are not mutually exclusive, a chi-squared statistic is used to test for significant differences in the percentages between bank and non-bank purchasers, broker and non-broker purchasers, pension and non-pension purchasers, direct and non-direct, insurance and non-insurance and “other” and non-other. to save space, the cell values corresponding to non-bank purchasers, non-broker purchasers, non-pension purchasers and so on are not reported in the table. an “*” denotes a cell value that is statistically significantly different at the five percent level from the corresponding value for all other purchasers not using the particular distribution channel being examined. nonparametric tests for differences in the percentage values yield similar results and are not reported. a nonparametric test for median values is used to test for significant differences in the median age between bank and non-bank purchasers, broker and non-broker purchasers, direct and non-direct, pension and non-pension purchasers, insurance and non-insurance and “other” and non-other channel. 304 g.j. alexander et al. / financial services review 7 (1998) 301–316 panel c of table 1 reports the length of time that respondents have been fund investors, i.e., investor seasoning. as shown in the table, the average mutual fund shareholder was not a new investor in mutual funds since about 85% of the survey respondents purchased a mutual fund prior to 1993. purchasers of mutual funds from brokers, fund companies, and insurance companies were significantly more likely to be experienced investors. the panel also indicates that the typical mutual fund shareholder owned several other types of financial assets besides mutual funds. roughly 51% owned individual stocks, 31% owned individual bonds, 35% owned certificates of deposit (“cds”), 38% had money market deposit accounts (“mmdas”), and 27% owned annuities. furthermore, about 81% of the sample owned their primary residence. purchasers of mutual funds from brokers were significantly more likely than all other purchasers to own each type of financial asset listed and their primary residence. in contrast, bank purchasers were significantly less likely to own individual stocks but were significantly more likely to own cds and mmdas. pension plan investors were significantly less likely to own cds and mmdas, whereas insurance company investors were more likely to own annuities and their primary residence. panel a of table 2 presents data on the types of mutual funds owned by purchasers using the various distribution channels. in general, each type or category of fund represents a table 2 ownership attributes distribution channel used bank broker pension direct insurance other total a. type of fund owned stock 64.8%* 82.3%* 80.1* 85.3%* 58.5%* 75.9% 72.9% bond 40.3 45.6* 39.3* 39.7* 41.0* 34.5 36.1 money 44.6* 39.4 39.1 38.8 65.5* 32.8 39.2 other 15.5 19.1 12.4* 21.8* 28.6* 20.7 14.6 median number 2* 2* 2* 2* 2* 1 1 of channels used b. number of funds owned one 22.9% 12.5%* 18.3%* 13.4%* 18.9%* 32.1%* 23.3% two 20.6 15.7* 20.3 17.2* 22.3 10.7* 21.0 three 19.8 14.9 17.0 12.5* 15.2 16.7 16.1 four or more 36.8 56.9%* 44.4* 57.0* 43.6* 40.5 39.6 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% median number 3 41* 3* 41* 3 3 3 of funds owned c. type of largest fund owned stock 49.8%* 69.7%* 68.0%* 73.9%* 59.5%* 59.3% 63.8% bond 14.7* 11.6 8.1* 7.9* 9.0 20.4* 10.6 money 25.3* 11.6* 14.1* 9.7* 20.1* 13.0 16.3 other 10.2 7.3* 9.8 8.5 11.3 7.4 9.3 notes:1. a “*” denotes a cell value that is statistically significantly different at the five percent level from the corresponding value for all other purchasers not using the particular distribution channel being examined. 2. fund owners with four or more funds are represented by 41 since the exact number of funds, if over three, was not requested in the survey. 3. see notes to table 1. 305g.j. alexander et al. / financial services review 7 (1998) 301–316 different combination of possible risk and return. over 72% of respondents own stock mutual funds, nearly 40% own money market mutual funds, and about 36% own bond funds. broker, pension plan, and direct purchasers were significantly more likely to own stock funds, whereas bank and insurance company purchasers were significantly less likely to own them. in contrast, bank and insurance company customers were significantly more likely to own money market mutual funds than were the customers of other sales channels. lastly, broker, pension plan, direct, and insurance company customers were significantly more likely to own bond funds than are other sales channel customers. panel b of table 2 reveals that the median number of funds owned by the respondents is three. furthermore, 55.7% (5 16.1%1 39.6%) of the respondents reported owning three or more mutual funds. more than two-thirds of broker and direct mutual fund purchasers own three or more mutual funds, with more than half of both groups owning four or more mutual funds. the median number of funds owned by the respondents was significantly greater for broker, direct, and pension plan investors relative to non-broker, non-direct, and non-pension plan purchasers, respectively. panel c of table 2 presents the type of mutual fund in which the respondents hold their largest investment. the largest fund type may indicate some measure of the risk preferences of investors, or alternatively, the knowledge of investors. about 64% of the respondents report that their largest investment is in a stock fund. broker, pension plan, and direct purchasers were significantly more likely to have their largest investment in a stock fund. on the other hand, bank and insurance company purchasers were significantly less likely to have their largest investment in a stock fund. bank purchasers were the largest investors in bond and money market funds with a percentage (40.0%5 14.7%1 25.3%) that is significantly greater than that for non-bank purchasers. 4. sources of information and knowledge of mutual fund investors this section examines the sources of information that investors use to learn about mutual fund investments, as well as the level of financial literacy displayed by survey respondents. the analysis shows which investors, categorized by distribution channel, are aware of the returns and risks associated with mutual fund purchases, along with the role played by the mutual fund prospectus and other sources of information in their learning about mutual fund investments. 4.1. sources of information panel a of table 3 indicates that the mutual fund prospectus was the single most widely used source of information, with 57.7% of respondents having cited it as a source of information in making their most recent mutual fund purchase. survey respondents also reporteded that they relied heavily on, in decreasing order, employer-provided printed materials (44.5%), financial publications like newspapers and magazines (42.0%), family or friends (37.6%), and meetings or presentations at work (33.5%) in choosing their most recent 306 g.j. alexander et al. / financial services review 7 (1998) 301–316 mutual fund investments. furthermore, 31% of the survey respondents stated that brokers provided information used in making their most recent mutual fund investment decisions. the prospectus was used by over 50% of the respondents regardless of the distribution channel used to make the purchase (except for “other”). for those who purchased mutual funds directly from a fund company, the prospectus and financial publications were the two most widely cited sources of information. not surprisingly, bank and broker purchasers were much more likely to cite bankers and brokers, respectively, as sources of information than purchasers who used other distribution channels, while pension plan purchasers were more likely to cite employer-provided printed materials and meetings or presentations at work. panel b of table 3 presents respondents’ perceptions of the best source of information for their most recently purchased mutual fund. generally, respondents cited the information source most closely associated with the distribution channel that they used in making their purchase as the most important. for example, a significant percentage of bank purchasers (19.4%), broker purchasers (39.0%), and pension plan purchasers (39.3%) named banker, broker, and employer-provided printed materials, respectively, as the best source of information. this is consistent with earlier observations on the results presented in panel a. table 3 information sources used in purchasing most recent mutual fund distribution channel used bank broker pension direct insurance other total a. information sources prospectus 51.2%* 56.5% 60.8%* 74.0%* 59.1% 49.4% 57.7% broker 27.4 61.6* 24.8* 29.6 31.7 31.8 31.0 family or friends 40.4 34.3* 33.6* 30.5* 42.4* 36.5 37.6 financial publications 41.4 49.8* 41.3 67.9* 39.7 34.1 42.0 banker 41.1* 6.9* 7.0* 4.3* 10.5 4.7 10.3 insurance company 0.0* 0.6* 0.6* 0.5* 6.0* 0.0 1.6 fund company 0.0 0.2 0.3 0.7* 0.0 0.0 0.3 employer 34.4* 23.3* 65.0* 25.9* 35.6* 35.3 44.5 meeting/presentation 23.9* 18.3* 46.6* 17.1* 31.1 27.1 33.5 other 4.6 4.8* 3.5 5.9* 3.5 5.9 3.5 b. best source of information prospectus 13.9% 13.0% 16.8%* 20.5%* 17.4% 13.4% 15.2% broker 11.0* 39.0* 11.7* 14.9 16.0 22.0 16.9 family or friends 20.9* 13.3* 10.9* 12.6* 20.4* 24.4* 16.3 financial publications 13.6 21.6* 16.6 36.7* 12.6* 14.6 17.1 banker 19.4* 2.0* 1.9* 0.9* 4.4 1.2 4.2 insurance company 0.0 0.3 0.1* 0.0 1.6* 0.0 0.4 fund company 0.0 0.0 0.0 0.0 0.0 0.0 0.0 employer 18.7* 9.1* 39.3* 10.9* 21.4* 23.2 26.7 meeting/presentation 0.4 0.2 0.7 0.2 0.8 0.0 0.6 other 2.2 1.6 2.0 3.3 5.2 1.2 2.6 notes:1. a “*” denotes a cell value that is statistically significantly different at the five percent level from the corresponding value for all other purchasers not using the particular distribution channel being examined. 2. “employer” denotes “employer-provided printed materials” and “meeting/presentation” denotes “meetings or presentations at work.” 3. see notes to table 1. 307g.j. alexander et al. / financial services review 7 (1998) 301–316 surprisingly, direct plan purchasers cited financial publications (36.7%) more than the prospectus (20.5%) as the best source. overall, survey respondents most often cited employer-provided printed materials as the best source of information about their most recently acquired mutual funds. this result would seem to be best explained by the large number of respondents who had purchased funds through pension plans. after employer-provided materials (26.7%), the sources of information most frequently cited as the best were, in decreasing order, financial publications (17.1%), broker (16.9%), family or friends (16.3%), and the prospectus (15.2%). 4.2. knowledge of risk, expenses, and performance panel a of table 4 presents data on mutual fund investor awareness of certain investment risks involved with stock, bond, and money market mutual funds. most mutual fund purchasers know that it is possible to lose money in stock, bond, and money market mutual funds (94.0%, 71.8%, and 63.9% know this, respectively). as shown in panel b, the difference in the percentages of investors who believe that stock and bond mutual funds can table 4 investor knowledge of risk associated with mutual funds distribution channel used bank broker pension direct insurance other total a. is it possible to lose money in this type of fund? stock fund yes 93.9% 96.9%* 94.6% 97.9%* 92.3% 92.4% 94.0% no 2.7 0.9* 1.5 0.5* 2.5 2.2 2.0 dk/refused 3.4 2.2* 3.9 1.6* 5.2 5.4 4.1 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% bond fund yes 72.8% 79.5%* 73.6%* 85.6%* 68.7% 67.4% 71.8% no 13.3 8.2* 12.1 6.2* 13.2 18.5 12.3 dk/refused 14.0 12.4* 14.3* 8.3* 18.0 14.1 16.0 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% money market fund yes 64.0% 63.0% 64.9% 67.5%* 66.8% 64.1% 63.9% no 20.1 23.0 20.3 21.8* 20.0 19.6 20.5 dk/refused 16.0 14.0 14.9 10.7* 13.2 16.3 15.7 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% b. cross-fund differences stock vs. bond funds stock funds 93.9% 96.9% 94.6% 97.9% 92.3% 92.4% 94.0% bond funds 72.8 79.5 73.6 85.6 68.7 67.4 71.8 difference 21.1 17.4 21.0 12.3 23.6 25.0 22.2 (t-statistic) (8.0*) (10.7*) (15.8*) (8.6*) (11.7*) (4.7*) (22.1*) bond vs. money market funds bond funds 72.8% 79.5% 73.6% 85.6% 68.7% 67.4% 71.8% money mkt funds 64.0 63.0 64.9 67.5 66.8 64.1 63.9 difference 8.8 16.5 8.8 18.1 1.9 3.3 7.9 (t-statistic) (2.6*) (7.3*) (5.1*) (7.9*) (0.74) (0.55) (6.1*) notes:1. dk denotes “don’t know.” 2. a “*” signifies statistical significance at the five percent level; a paired t-test was used in testing the difference between stock and bond funds and between bond and money market funds. 3. see notes to table 1. 308 g.j. alexander et al. / financial services review 7 (1998) 301–316 lose money is a statistically significant 22.2%; the difference when bond and money market funds are compared is 7.9%, which is smaller but still statistically significant. the differences between the stock and bond fund percentages are generally similar across all distribution channels. a similar observation can be made when the bond and money market fund percentages are compared, except for the insurance and “other” distribution channels where the differences are small and insignificant. overall, broker and direct purchasers seem most knowledgeable about the possibility of losing money in all three types of mutual funds. although not reported in the tables, the respondents’ beliefs about being able to lose money in stock, bond, and money market funds were also examined by four demographic pieces of information: age, income, education, and gender. notable observations are (1) college graduates are significantly more likely to believe one can lose money in a stock fund; (2) knowledge that bond funds can lose money is related to age (older investors are more likely to know), income (wealthier investors are more likely to know), and gender (males are more likely to know); and (3) respondents younger than 35 are less likely to believe that one can lose money in a money market fund. no other significant differences were observed. table 5 reports data on investor familiarity with mutual fund operating expenses. the first two panels present information on the percentage of respondents who could provide some expense estimates for their largest mutual fund. as shown in panel a, the level of expenses did not seem to be an important factor in the purchasing decision of many respondents. only 18.9% of the respondents could give an estimate of expenses for their largest mutual fund, although broker and direct purchasers were significantly more likely to be able to do so. the percentages of respondents who could provide even an approximation of actual expenses table 5 knowledge and beliefs about annual expenses distribution channel used bank broker pension direct insurance other total a. knowledge of largest fund’s expenses yes 15.3% 23.0%* 19.8% 35.0%* 20.7% 17.4% 18.9% no 84.7 77.0 80.2 65.0 79.3 82.6 81.2 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% b. knowledge of expenses at time of purchase yes 46.1% 49.5%* 40.5%* 59.7%* 47.8%* 28.0* 43.0% no 53.9 50.5* 59.5* 40.3* 52.2* 71.9* 57.1 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% c. expected performance of fund with higher than average expenses above average 23.8% 19.3% 19.7% 16.6% 22.9% 20.3% 19.9% about average 66.5 63.3 64.4 62.9 63.6 56.3 64.4 below average 9.7* 17.4 15.9 20.6* 13.5 23.4 15.7 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% d. expected performance of fund with good performance in the previous year above average 19.5% 24.9% 25.3% 29.8%* 27.3 23.6% 24.1% about average 75.6 68.0 68.8 62.2* 69.1 69.4 70.6 below average 4.9 7.1* 5.9 8.0* 3.6 6.9 5.3 total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% notes:a “*” signifies statistical significance at the five percent level; a paired t-test was used in testing the difference between stock and bond funds and between bond and money market funds. see notes to table 1. 309g.j. alexander et al. / financial services review 7 (1998) 301–316 were even smaller. respondents earning less than $75,000 were significantly less likely to provide an expense estimate. males and college graduates were significantly more likely to provide an expense estimate. respondents who could not provide an expense estimate for their largest fund were asked if they knew of their largest funds’ expenses at the time of purchase. panel b of table 5 reports that only 43.0% of the respondents claimed to have known any of their largest fund’s expenses at the time they first invested in the fund. broker, direct, and insurance company purchasers were significantly more likely to have claimed to have known the annual expenses of their funds at the time of initial purchase. college graduates and males were significantly more likely to have responded that they knew the funds expenses. pension plan purchasers, in contrast, were significantly less likely to have known their funds’ annual expenses. annual expenses of funds may be of less significance to mutual fund shareholders who purchase their shares through employee pension plans, however, as a participant in a typical defined contribution plan is presented with a choice of different funds with different investment objectives. although the choice of funds is typically designed to allow the employee to allocate assets among broad categories of investments (e.g., stocks, bonds, or money market investments), the employee is usually not presented with a choice of different funds with the same investment objective. as a result, the cost of holding a particular fund would appear to be of lesser importance to an employee who purchases fund shares through an employee pension plan than to an investor who purchases funds through other distribution channels. panel c of table 5 reports investors’ beliefs about the relationship between expenses and mutual fund performance. in general, there is an inverse relationship between fund performance and expenses, particularly for bond and money funds (see blake et al., 1993; carhart, 1997; elton et al., 1996). about 20% of the survey respondents believed that mutual funds with higher expenses produced better results, while 64% believed that funds with higher expenses produced average results. only 16% of the survey respondents believed that higher expenses led to lower than average returns. bank customers were significantly less likely than non-bank customers to expect an inverse relationship, while direct fund purchasers were significantly more likely to expect an inverse relationship. the relationship between performance and expenses was also examined by the respondents’ largest fund type. the only statistically significant difference involved respondents who named money market mutual funds as their largest type. these respondents were significantly less likely to believe that higher expenses led to lower than average returns. panel d of table 5 reports investor perceptions about the year-to-year performance of mutual funds. mutual funds must present historical fund returns over the past ten years in the prospectus which, in turn, must be presented to investors before they make their purchase decision. although many investors tend to choose mutual funds largely on the basis of past performance, empirical evidence on the historical relationship between returns in successive years suggests there is either a slightly positive relationship or none at all, depending on the time period, sample, and methodology utilized (see blake et al. 1993; bogle, 1992, 1994; brown & goetzmann, 1995; brown et al., 1992; carhart, 1997; elton et al., 1996; goetzmann & ibbotson, 1994; hendricks et al., 1993; kahn & rudd, 1995; malkiel, 1995). approximately 24% of the respondents believe that a fund that has performed well last 310 g.j. alexander et al. / financial services review 7 (1998) 301–316 year will have an above average return this year; 71% believe the fund will have an average return; and 5% believe the fund will have a below average return. direct purchasers are significantly more likely than non-direct purchasers to believe that returns in successive years are positively related. interestingly, as reported in panel b of table 3, 36.7% of direct purchasers named financial publications as the best source of information. these publications are the most likely places for performance advertisements to appear, thereby implying— but not stating—that there is a direct relationship between past and future performance. several other interesting results are not reported in the tables. for example, respondents who owned either stock or bond funds were significantly more likely than, respectively, non-stock owners or non-bond owners to know that average stock market returns exceed the return on u.s. treasury bills. however, the difference between money market and non-money market owners was not significant. in terms of demographic characteristics, college graduates, males, and respondents with higher income were significantly more likely to believe that the average return on stocks is greater than that on treasury bills. although this analysis was extended by classifying respondents by the largest fund they owned, those whose largest holding was a stock fund were found to be significantly more likely to know that stock market returns on average exceed treasury bill returns. in contrast, investors whose largest holding was either a bond or money market fund were significantly less likely to know this. 5. investor financial literacy in this section, we examine the level of financial literacy of the mutual fund survey respondents. we conduct this analysis in several steps. first, we construct an aggregate measure of overall investing and mutual fund knowledge for each respondent. this measure is called the respondents’ quiz score. second, we test for statistically significant differences between high and low quiz score groups by demographic and financial characteristics and by sources of information. finally, we employ a logit model to assess the factors that are most important in explaining differences in overall investor literacy as measured by the quiz score. the quiz score should not necessarily be interpreted as indicating whether any particular financial intermediary has been more or less successful in educating investors. quiz scores may show that in general, more knowledgeable investors choose to purchase funds through particular channels. for example, more financially literate investors may be more comfortable with the idea of purchasing directly from a fund company and may be more likely to maintain an account with a stockbroker. as a result, it can not be inferred that salespeople in these distribution channels necessarily do a better job of disclosing risks and costs. indeed, alexander et al. (1997) found evidence that a mutual fund investor’s level of financial literacy and choice of distribution channel are jointly determined. 311g.j. alexander et al. / financial services review 7 (1998) 301–316 5.1. quiz score measure of investor financial literacy the measure of overall investor knowledge is based on the responses to a subset of questions in the mutual fund survey. the quiz consists of nine questions and the number of correct responses is called the quiz score. the quiz score measures investing knowledge in general and mutual fund investment knowledge in particular on the part of mutual fund shareholders. the questions involve the respondents reporting whether or not they know: (1) that it is possible to lose money in a stock mutual fund; (2) that it is possible to lose money in a bond mutual fund; (3) that money market mutual funds are not insured; (4) that there are thousands of mutual funds to choose from in making an investment decision; (5) that stock market returns are, on the average, greater than the return on u. s. treasury bills; (6) what the term net asset value (nav) means; (7) what the term redemption means; (8) what the term derivatives means; and (9) what the term present value means. quiz questions six through nine are relatively weak measures of investor knowledge, as respondents were given credit for those questions if they claimed to know what these various terms mean, even though no attempt was made to verify the accuracy of their responses. the results reported here, however, are essentially unchanged when quiz scores are based solely on the responses to the first five questions. 5.2. quiz score analysis table 6 presents the mean of the respondents’ quiz score by distribution channel as well as for those respondents who did not use the channel. also, we report the results of conventional t-tests of the equality of mean quiz scores for each type of distribution channel. table 6 mean of the quiz score by distribution channel distribution channel mean for channel users mean for non-channel users difference t-statistic bank 4.77 5.08 20.31 22.10* broker 5.48 4.81 0.67 6.21* direct 6.26 4.50 1.76 17.44* pension 5.14 4.89 0.25 2.39* insurance company 4.82 5.11 20.29 22.38 other 4.92 5.04 20.12 20.42 mean for all channels 5.03 notes:a “*” signifies statistical significance at the five percent level. a difference in means test is used to test for significant differences in quiz scores that adjusts for unequal variances when necessary. the absolute value of the t-statistic is reported. see notes to table 1. 312 g.j. alexander et al. / financial services review 7 (1998) 301–316 the typical mutual fund shareholder had a quiz score of five out of a possible nine. investors purchasing directly from fund companies scored much higher than any other fund group. broker and pension plan purchasers also scored significantly higher than those buying mutual fund shares through other distribution channels. however, bank and insurance company purchasers received significantly lower mean quiz scores than other survey respondents. although not presented, we examine financial literacy results by the number of channels used by the respondent to purchase mutual funds, and several demographic and financial characteristics, as well as best source of information. first, multiple-channel purchasers have significantly higher quiz scores than those who used only a single channel (5.70 vs. 4.44, respectively), with the largest difference, except other, being in the pension channel (5.78 vs. 4.21). this is of particular interest, given the recent rapid growth in 401(k) plans. second, average quiz score was higher for males and for those respondents who work in the financial services industry, and generally increased with age, education, and income. finally, in terms of best source of information, respondents who reported that financial publications and the prospectus were the best sources of information scored significantly higher on the quiz. in contrast, those respondents who relied on family or friends, bank representatives, employerprovided printed materials, and insurance company representatives scored significantly lower. 5.3. quiz score logit analysis in the previous section, we examined quiz scores by considering the individual factors one at a time. in comparison, a multivariate analysis assesses how the quiz score varies by particular factors, holding constant the effects of a wide set of other factors. we use a multivariate analysis based on a logit model. this analysis makes it is possible to assess the linkage between the quiz score to demographic and financial characteristics, and to other factors such as sources of information and distribution channel used. table 7 reports the results for this model, displaying maximum-likelihood coefficient estimates and their asymptotic t-statistics along with a chi-squared measure of overall goodness of fit and its p-value, the proportion of correctly predicted quiz scores, and the total number of observations. overall, the model can be viewed as a way of seeing if the various items of information provided by respondents can be used to predict whether they are above or below average in their financial literacy. thus, the dependent variable in the multivariate model is a discrete random variable that takes on a value of either one or zero for each respondent, depending on whether the respondents quiz score placed him or her in the top or bottom half of the quiz score distribution. while dividing the sample into halves is arbitrary, the results do not differ when the sample is divided into tertiles, quartiles, and quintiles. similarly, the results were also insensitive to different indices of knowledge measured by several alternative subsets of the quiz questions. we use a large number of explanatory variables in the estimation. included among these are several dummy (or indicator) variables. the demographic dummy variables male, college_grad, work_fin_inst, age, num_funds, income, and seasoned take on a value of 1 (0 otherwise) if the respondent is a male, is a college graduate, works at a financial institution, is older than 43 years of age, owns three or more funds, has 313g.j. alexander et al. / financial services review 7 (1998) 301–316 household income greater than $75,000, and purchased a mutual fund prior to 1993, respectively. also included are dummy variables for the best source of information used in purchasing the most recent mutual fund. to avoid collinearity problems, a separate dummy variable was not included for respondents naming “other” as the best source of information. the dummy variables publications, prospectus, broker, banker, employer, and family take on a value of 1 (0 otherwise) if the best source of information is, respectively financial publications, mutual fund prospectuses, brokers, bankers, employerprovided printed materials, and family or friends. it should be noted the number of respondents in the multivariate analysis is 1,554 since 446 respondents did not provide answers to one or more of the questions such as age or income. the results of this exercise indicate that there is a significant positive relationship between the quiz score and five demographic explanatory variables—being a male, a college graduate, working at a financial institution, owning three or more funds, and earning income greater than $75,000. furthermore, those respondents who indicated that either financial publications or mutual fund prospectuses were their best source of information earned significantly higher quiz scores. the overall fit of the multivariate model is good, as indicated by the significantly low p-value associated with the chi-squared statistic. in addition, the model was able to correctly identify high and low quiz score respondents for slightly more table 7 multivariate logit estimation of determinants of quiz scores variable coefficient estimate t-statistic male 0.8320 6.75* college_grad 0.6753 5.54* work_fin_inst 1.1758 5.00* age 0.1618 1.34 num_funds 0.2530 2.08* income 0.5370 4.20* seasoned 0.3609 1.54 publications 0.9376 3.82* prospectus 0.5981 2.45* broker 0.0707 0.29 banker 20.3925 21.03 employer 20.4336 21.85 family 20.3051 21.20 chi-squared statistic 302.5 (p-value) (0.000) proportion predicted correctly 0.701 number of observations 1554 note: quiz score is a dummy variable with a 1 indicating the respondent scored in the top half of the distribution and 0 otherwise. the dummy variables, male, college_grad, work_fin_inst, age, num_funds, income and seasoned take on a value of 1 (0 otherwise) if the respondent is a male, a college graduate, works at a financial institution, older than 43 years of age, owns three or more funds, has household income greater than $75,000, and purchased a mutual fund prior to 1993, respectively. also included are dummy variables for the best source of information used in the respondents’ most recent mutual fund purchases. the dummy variables publications, prospectus, broker, banker, employer, and family take on a value of 1 (0 otherwise) if the best source of information is financial publications, the mutual fund prospectus, broker, banker, employer-provided printed materials, and family or friends, respectively. 314 g.j. alexander et al. / financial services review 7 (1998) 301–316 than 70% of the respondents based on their demographics and best source of information, providing additional evidence that the model has a good fit. 6. conclusion our results show that the typical mutual fund investor surveyed is older, wealthier, and better educated than the average american. the results of the survey suggest, however, that investor knowledge of the expenses and risks associated with mutual funds can be improved. although the average fund shareholder has invested in funds for several years, most fund shareholders do not appear to appreciate the relationship between fund expenses and performance. in addition, a substantial number of fund investors still believe that they cannot lose money in a bond fund. the survey results also suggest that more can be done to make mutual fund prospectuses more useful to investors, especially since over 40% of those surveyed stated that they never used the prospectus. moreover, the survey respondents considered the prospectus only the fifth best source of information about the funds that they purchased. two rules that were recently adopted by the sec are significant and timely steps in this direction. the first rule requires the use of “plain english” that avoids legalese in disclosure documents such as prospectuses. the second rule allows shortened but more focused prospectuses, known as profile prospectuses, to be sent to potential mutual fund purchasers. readers interested in these rules will find them posted at the sec’s web site: www.sec.gov/rules/final/33-7497.txt and www.sec.gov/rules/final/33-7513.htm, respectively. although broker and direct fund company purchasers are relatively more knowledgeable about the costs and risks of mutual fund investments than non-broker and non-direct fund company purchasers, it is likely that investors self-select into the various distribution channels. for example, more knowledgeable investors may be more comfortable with the idea of purchasing from a fund company or a broker. as a result, salespeople at banks and insurance companies may face greater challenges in educating their typical mutual fund buyers. the survey should not be read as indicating that salespeople in broker and direct distribution channels necessarily do a better job of disclosing risks and costs than their counterparts in other distribution channels. the ongoing challenge of raising the level of investor comprehension of the costs and risks associated with mutual fund investments extends well beyond simply designing regulatory requirements. ultimately, the goal of better educated investors can only be achieved by a concerted joint effort involving numerous parties, including plan sponsors, brokers, fund companies, and governmental regulatory agencies. references alexander, g. j., jones, j. d., & nigro, p. j. (1996). report on the occ/sec survey of mutual fund investors. washington, dc: office of the comptroller of the currency and the u. s. securities and exchange commission. 315g.j. alexander et al. / financial services review 7 (1998) 301–316 alexander, g. j., jones, j. d., & nigro, p. j. (1997). investor self-selection: evidence from a mutual fund survey. managerial and decision economics 18(special issue, nov.–dec.), 719–729. american association of retired persons, consumer federation of america, and the north american securities administrators association, inc., (1994, january). outline for minimums for hill reform action...most consumers donot know risk of uninsured investments at banks, report. blake, c. r., elton, e. j., & gruber, m. j. (1993). the performance of bond mutual funds.journal of business 66(july), 371–403. bogle, j. c. (1992). selecting equity mutual funds.journal of portfolio management 18(winter), 94–100. bogle, j. c. (1994).bogle on mutual funds: new perspectives for the intelligent investor. burr ridge, il: irwin. brown, s. j., & goetzmann, w. n. (1995). performance persistence.journal of finance 50(june), 679–698. brown, s. j., goetzmann, w. n., ibbotson, r. g., & ross, s. (1992). survivorship bias in performance studies. review of financial studies 5, 553–580. carhart, m. m. (1997). on persistence in mutual fund performance.journal of finance 52(march), 57–82. chen, h., & volpe, r. (1998). an analysis of personal financial literacy among college students.financial services review 7(2), 107–28. consumer bankers association. (1994, november).consumer investment product survey. crenshaw, a. b. (1996, may 19). before risking the money, invest in financial literacy.the washington post, pp. h1, h4. elton, e. j., gruber, m. j., & blake, c. r. (1996). the persistence of risk-adjusted mutual fund performance. journal of business 69(april), 133–157. goetzmann, w., & ibbotson, r. g. (1994). do winners repeat?journal of portfolio management 20(winter), 9–18. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974–1988.journal of finance 48(march), 93–130. investment company institute. (1999). 1999 mutual fund fact book. available online: www.ici.org. kahn, r. n., & rudd, a. (1995). does historical performance predict future performance?financial analysts journal 51(nov.–dec.), 43–52. kimmelman, j. (1995, june 9). banks best at disclosing risks, study finds.american banker, p. 10. malkiel, b. (1995). returns from investing in equity mutual funds 1971 to 1991.journal of finance 50(june), 549–572. market trends. (1996, may 5).survey of nondeposit investment sales at fdic-insured institutions. plasencia, w., & cope, d. (1996, january 19). blasted on mutual fund disclosure, banks hit back.american banker. schultz, e. (1995, december 22). helpful or confusing? fund choices multiply for many retirement plans.the wall street journal, pp. c1, c25. updegrave, w. l. (1996). fund investors need to go back to school.money(february) 98–100. 316 g.j. alexander et al. / financial services review 7 (1998) 301–316 financial services review, 33(3) 1 is using a financial advisor related to cryptocurrency investment? alex brockbank,1 charlene kalenkoski,2 christopher browning, 3 and michael guillemette4 abstract do financial advisors recommend cryptocurrency investment within a household portfolio? cryptocurrencies have emerged in popularity, especially in the united states, as households seek to maximize returns. financial advisors are expected to provide beneficial advice for a household in managing financial decisions including investments. the existing literature has examined this relatively new form of investing and found determinants for cryptocurrency investment but has not sufficiently explored the association between the investor’s use of a financial advisor and the decision to invest in cryptocurrency. with data from the 2018 wave of the national financial capabilities study (nfcs), this paper examines the relationship between the use of a financial advisor and cryptocurrency investment for american investors. the results suggest that investors who use a financial advisor are more likely to invest in cryptocurrencies. additional determinants seen in previous works are also confirmed in the current study; showing that investors who are younger, married, and have higher subjective financial literacy are more likely to have cryptocurrency investments compared to individuals who are older, unmarried, and have lower levels of subjective financial literacy. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation brockbank, a, kalenkoski, c., browning, c., & guillemette, m. (2025). is using a financial advisor related to cryptocurrency investment? financial services review, 33(3), 1-19. introduction cryptocurrencies have emerged as a popular and exciting investment opportunity. in early 2017, the global cryptocurrency market capitalization was just under $20 billion and a year later that market cap was closer to $500 billion—extending beyond $3 trillion in 2025 (best, 2025). this 1 corresponding author (alex.brockbank@uvu.edu). utah valley university, orem, ut, usa. 2 james madison university, harrisonburg, va, usa. 3 texas tech university, lubbock, tx, usa. 4 texas tech university, lubbock, tx, usa. phenomenon is also taking hold in the united states. in 2022, the u.s. cryptocurrency market size alone was valued at just under $1.2 billion with researchers expecting an annual growth rate of about 12% until 2030 (grand view research, 2023). furthermore, about two out of three americans say they are familiar with https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(3) 2 cryptocurrency and about 28% of american adults own cryptocurrency in the u.s. (blackstone, 2025). the opportunity and ability to invest in cryptocurrency has continued to increase over time. from the inception of bitcoin, the first cryptocurrency investment option, to the first exchange-traded fund (etf) that included cryptocurrency becoming available in the u.s., to even more availability and ease of investment (zweig, 2021). there is plenty of reason for excitement, but also hesitation about cryptocurrencies and the volatility it introduces to investment decisions. when navigating complicated and risky investments, individuals may hire a financial advisor to help them select and optimize these investments within their portfolios, thereby mitigating risks and increasing returns. while current literature has thoroughly examined cryptocurrency as an investment option (andrianto & diputra, 2017; inci and lagasse, 2019), including the determinants of cryptocurrency investment (xi et al., 2020; kim et al., 2023) and its impact on an investment portfolio (wong et al., 2018), little extensive research has been conducted to determine the role and influence of a financial advisor on cryptocurrency investment. this problem—and gap in the body of literature—is addressed in the current study. does an individuals’ use of a financial advisor have an influence on their decision to invest in cryptocurrencies? the current study addresses this question, along with furthering the research on the associations between other measures of personal human capital and cryptocurrency investment. first, the relevant literature on cryptocurrency and financial advisors will be thoroughly examined. a theoretical framework is discussed and hypotheses developed. a statistical estimation is completed using a probit model on data from the 2018 wave of the national financial capabilities study (nfcs). the findings show that investors using a financial advisor have a higher probability of investing in cryptocurrency. this finding suggests that individuals who use a financial advisor are more likely to use beneficial complex investment options, specifically cryptocurrency, within their investment portfolio. there are valuable implications from the results of the current study. from the financial advisor’s perspective, as the popularity of cryptocurrencies continue to increase, clients will have more questions about the opportunity and a greater desire to participate and reap the benefits of investment. for clients, if they desire to better use complex investment tools, they can benefit from hiring a financial advisor. literature review and theoretical framework human capital and existing cryptocurrency investment research research has examined common traits among the individuals who choose to invest in cryptocurrencies and found certain demographic variables to be influential in the decision. among them, gender, education, income, and wealth are significant predictors of cryptocurrency investment—most investors are men, higher educated, have higher income, and have accumulated greater wealth (lammer et al., 2020; xi et al., 2020). another association exists between investor age and the cryptocurrency investment decision, wherein younger investors are more likely to invest compared to older individuals (xi et al., 2020). hackethal et al. (2022) show that cryptocurrency investors are more likely to be engaged in risk-seeking financial behaviors. social influence and public perception are associated with the decision to invest in cryptocurrency (gupta et al., 2020). ayedh et al. (2020) find that profitability, compatibility, awareness, and facilitating conditions impact cryptocurrency investment decisions. other studies have similarly found that potential future gain (profitability) is a significant factor in cryptocurrency investment (yilmaz & hazar, 2018; arias-oliva et al., 2019). the existing literature finds conflicting results with the relationship between financial literacy and cryptocurrency investment. arias-oliva et al. (2019) find evidence that financial literacy has no statistical significance; however, zhao & zhang (2021) show that financial literacy was a brockbank et al. 3 statistically significant determinant, but that investment experience predicts cryptocurrency investment more accurately than financial literacy. they further delve into the significance of financial literacy and show that only subjective financial knowledge, determined by asking the individual to assess their knowledge, is found to have a significant positive association. the relationship between objective financial knowledge, determined by a scored questionnaire on important financial topics and cryptocurrency investment, is not significant (zhao & zhang, 2021). kim et al. (2023) find cryptocurrency investment to have a negative association with objective investment literacy, but a positive association with subjective literacy measures. an earlier study also found that subjective financial knowledge was significantly related to investment behavior regardless of objective financial knowledge levels (allgood & walstad, 2016). previous literature has examined portfolio construction utilizing cryptocurrencies and found that cryptocurrencies add beneficial diversification results within an investment portfolio (andrianto & diputra, 2017; chuen et al., 2018; wong et al., 2018; inci and lagasse, 2019). wong et al. (2018) further examine cryptocurrency investments and find that different coins have different hedging or diversification benefits, thus adding to the potential of the investment, but also adding to the complexity and volatility. human capital theory posits that obtained education, training, and skills will increase an individual's abilities, including financial capabilities (ibrahim, 2018), and overall productive efficiency (becker, 1962). additional research has examined domain-specific human capital and emphasized the importance of specialization within different industries on outcomes (gibbons & waldman, 2004; sullivan, 2010; rauch & rijsdijk, 2013). studies have reviewed domain-specific expertise within the context of financial advisor services and found similar benefits of specialization (hershey et al., 1990; hanna & lindamood, 2010; pilote et al., 2024). the current study also uses the theoretical framework of information economics. consumers engage in the search for information to maximize their utility, but the cost of search— including monetary spending, time consumption, and opportunity cost—is negatively related to consumer engagement in search (stigler, 1961; lin & lee, 2004). cryptocurrency investments are complex in nature (wong et al., 2018; stosic et al., 2019) and can be extremely volatile (dasman, 2021), which increases the cost of the search for information and reduces participation. smith et al. (1999) find that higher knowledge reduces cost and increases the efficiency of the information search process. it can, therefore, be expected that greater domain-specific human capital increases an individual’s willingness to complete meaningful research of complex investments—such as cryptocurrency investments. while the ownership of human capital is generally nontransferable (lev & schwartz, 1971), hiring professionals with domain-specific human capital, such as a financial advisor, is an effective way to increase decision-making capacity by utilizing the hired professional’s expertise and training (hershey et al., 1990). one role of a financial advisor is to help clients navigate the risk and complexities of investing (kitces, 2021). studies find that households using a financial advisor tend to be wealthier, have higher incomes, and have greater financial literacy than those who do not use a financial advisor (smith et al., 2012). additional research also shows that financial advisors play a crucial role in helping individuals navigate significant life events, and that experiencing these major events significantly influence the decision to seek out a financial advisor (cummings & james, 2014; sommer & macdonald, 2022). educational achievement and risk tolerance are also positively associated with a consumer's decision to seek help from a financial advisor (hanna, 2011; barthel & lei, 2021). hypotheses within the scope of the current study, financial advisors are assumed to have high domainspecific (financial) human capital relevant to investments. given this assumption, and the financial services review, 33(3) 4 framework of information search and human capital theory, it is expected that financial advisors will be more likely to engage in the search for information on cryptocurrency investments to better maximize the utility of their clients. thus, leveraging the human capital of a financial advisor should lead to an increase in the utilization of beneficial complex investment options, such as cryptocurrency, within an investment portfolio. therefore, the main hypothesis of the current study is as follows: hypothesis 1: the use of a financial advisor is associated positively with cryptocurrency investment. human capital theory and the existing literature on cryptocurrency also motivate the following hypotheses: hypothesis 2: the key demographic variables (education, gender, income, and investment assets) are each associated positively with cryptocurrency investment. hypothesis 3: age is associated negatively with cryptocurrency investment. data the data utilized are from the 2018 national financial capability study (nfcs) state-bystate and follow-up investor surveys. the later surveys, from 2021, were considered for use alone, or in combination with 2018 data, but the key variable measuring the use of a financial advisor was significantly altered and, therefore, it was determined that the use of 2018 survey results are sufficient for the current study. the survey is a project of the finra investor education foundation (finra foundation). the state-by-state survey is conducted to provide an extensive analysis of the financial capabilities of u.s. adults. nfcs state-by-state data are nationally representative of american adults (age 18+). a portion of these survey respondents are then asked additional questions for the investor survey. the investor survey data provide a more in-depth analysis regarding the investing decisions of the state-by-state survey respondents. the respondent count for the investor survey is 2,003. knowing that approximately two-thirds of americans say they are familiar with cryptocurrency, as stated in the introduction, and that about 28% of american adults own cryptocurrency (blackstone, 2025), a dataset that examines u.s. investors within the scope of the current paper. the nfcs data fit these parameters. while the initial data contain 2,003 respondents, investors who have not heard of cryptocurrency and those that did not answer, or chose not to answer, the applicable questions from the survey for the variables of the present study are removed from the analysis sample. the final count after these exclusions is 1,556 respondents. weights are provided and applied to approximate the investor population in terms of age and education. two-sample t-tests are completed to compare the analysis sample to the original sample (shown in table 1). the results show insignificant differences between the two samples for gender, ethnicity, marital status, age, and income; indicating that the analysis sample is not statistically different from the original sample on these dimensions. there is, however, a statistically significant difference found between the two samples in education, though the magnitude of difference is negligible. the analysis sample is slightly higher educated than the original survey sample. brockbank et al. 5 table 1. demographic variables mean comparison via t-test variable full mean1 (standard error) analysis mean2 (standard error) p-value (n = 2,003) (n = 1,556) gender 0.5667 (0.0111) 0.5925 (0.0125) 0.1208 ethnicity 0.8178 (0.0086) 0.8130 (0.0099) 0.7147 marital status 0.6370 (0.0107) 0.6343 (0.0122) 0.8669 age 4.6505 (0.0328) 4.6407 (0.0372) 0.8437 education 2.6410 (0.0229) 2.7204 (0.0251) 0.0201 ** income 2.2341 (0.0217) 2.2661 (0.0246) 0.3298 1: full mean represents the mean of the nfcs investor survey complete sample. 2: analysis mean represents the mean of the current paper analysis sample. *significance at the 10% level, **significance at the 5% level, ***significance at the 1% level the dependent variable is the choice of respondents to invest in cryptocurrency. it is a dummy variable that takes a value of 1 if the respondent is currently invested in cryptocurrency (either directly or through a fund) and 0 otherwise. the main explanatory variable captures the investor's use of a financial advisor when making investment decisions. investors are asked on the survey whether they use a “financial advisor other than stockbrokers” when “making an investment decision.” this variable is dichotomous and takes a value of 1 if the respondent answered yes, and 0 otherwise. additional key independent variables for this study are measures of human capital. these include two measures of financial literacy (subjective investment literacy and objective investment knowledge) and two measures of education (education level and participation in a financial education program). for subjective investment literacy, respondents are asked to selfassess their overall knowledge about investing from 1 to 7. this variable is reduced to include three groupings (below average, average, and above average). this reduction is based on the variable statistics, with “average” representing respondents who self-assessed the mean rating of 5 (a score of 1-4 is labeled “below average”, the reference category, while “above average” is a score between 6-7). the nfcs survey asks ten questions to assess the respondent’s investment knowledge, four of them cover basic investment knowledge topics including stocks, bonds, the relationship between risk and return, and average return. the objective investment knowledge variable sums the correct responses for these four questions and takes a value from 0 (no correct responses, the reference category) to 4 (all correct answers). answers of “don’t know” and “prefer not to say” are coded as incorrect responses for each investment question. education level is a categorical variable including no college (reference group), some college, a bachelor’s degree, and a post-graduate degree. participation in a financial education is a dummy variable that takes a value of 1 if the respondent received and participated in a financial education program, and 0 otherwise. other demographic and economic factors are included as independent variables for control purposes. these include age bracket, gender, ethnicity, marital status, income range, employment status, investment assets, and risk aversion. gender, ethnicity, and marital status are dichotomous variables that take a value of 1 if the respondent is male, white, or married, and 0 otherwise. age bracket ranges from 18-24 (youngest bracket) to 65+ (oldest bracket) with 10-year increments for each bracket. the lowest age group (aged 18-24) is used as the reference category. income range is coded into four categories: less than $50k (reference group), $50k-$100k, $100k-$150k, and greater than $150k. employment status includes selfemployed, employed full-time, employed parttime, retired, and those not currently employed. financial services review, 33(2) 6 the investment assets variable measures the approximate value of the respondents’ nonretirement accounts and is coded into seven groups: $10k or less (reference group), $10k$50k, $50k-$100k, $100k-$250k, $250k-$500k, $500k-$1 million, and $1 million or more. risk aversion is measured on a scale of 1 to 10 (selfreported on the survey as risk willingness and reverse coded to represent risk aversion). due to a small number of respondents within the extremities of the ten categories, this variable is condensed by combining the lowest scores (1 and 2) and then the highest scores (9 and 10). the resulting variable of risk aversion is eight categories, with a value of 1 being those very willing to take risk (lowest risk aversion) and 8 for those not at all willing to take risk (highest level of risk aversion). model the current study estimates the following probit model to examine the cryptocurrency investment decision of investors: 𝑦𝑖 ∗ = β0 + x𝑖𝛼 + 𝑍𝑖𝜆 + 𝑒𝑖 𝑒𝑖 ~ 𝑁(0,1) 𝑦𝑖 = { 1 𝑖𝑓 𝑦𝑖 ∗ > 0 0 𝑖𝑓 𝑦𝑖 ∗ ≤ 0 where 𝑦𝑖 ∗ is a latent variable representing whether the investor, i, chooses to invest in cryptocurrency or not. the observed variable, 𝑦𝑖, takes a value of 1 if the respondent invests in cryptocurrency and 0 otherwise. the main explanatory variables (use of a financial advisor, objective investment knowledge, subjective investment literacy, financial education participation, and education) are represented by 𝑋𝑖 in the model. the matrix, 𝑍𝑖, includes all the demographic and economic variables in the study. these variables are age, gender, ethnicity, marital status, income level, employment status, investment assets, and risk aversion. the intercept for the model is β0. the coefficients for the main explanatory variables are represented by 𝛼, and λ is the vector of coefficients for the control variables. the error term, 𝑒𝑖, is assumed to follow the standard normal distribution. to test the robustness of the full model, eight additional probit models are estimated. each new model iteration removes a single dependent variable from the full model. the removed variables were determined based on existing literature and include objective investment literacy, subjective investment literacy, education, age, gender, marital status, income, and investment assets. results table 2 provides descriptive statistics for the dependent and explanatory variables. about 85% of the analysis sample is not currently invested in cryptocurrencies and approximately 60% of the sample is using a financial advisor . brockbank et al. 7 table 2. descriptive statistics variable percent variable percent cryptocurrency investment ethnicity no 84.86% non-white 23.34% yes 15.14% white 76.66% use financial advisor marital status no 39.71% not married 41.27% yes 60.29% married 58.73% objective investment knowledge income range 0 none correct 2.17% less than $50k 26.21% 1 11.77% $50k $100k 43.19% 2 21.14% $100k $150k 19.25% 3 31.52% greater than $150k 11.35% 4 all correct 33.39% employment status subjective investment knowledge self-employed 9.83% below average (1-4) 32.98% employed full-time 42.44% average (5) 34.88% employed part-time 7.32% above average (6-7) 32.14% retired 30.43% financial education participation not currently employed 9.97% no 72.68% non-retirement investment assets yes 27.32% $10k or less 19.25% education level $10k $50k 18.90% no college 16.92% $50k $100k 14.54% some college 35.52% $100k $250k 16.43% bachelor's degree 28.55% $250k $500k 14.17% post graduate degree 19.02% $500k $1 million 8.81% age group $1 million or more 7.91% 18 24 7.90% risk aversion 25 34 17.41% 1 – low (very willing) 16.21% 35 44 12.27% 2 14.73% 45 54 15.81% 3 21.26% 55 64 17.35% 4 14.14% 65+ 29.26% 5 13.91% gender 6 7.83% female 42.21% 7 6.12% male 57.79% 8 – high (not willing) 5.80% observation count = 1,556 analysis using the 2018 nfcs surveys. survey weights are applied. the analysis sample has high self-rated investment literacy and high objective investment knowledge. about half of the sample has a bachelor’s degree or higher, with about 19% financial services review, 33(2) 8 having a post-graduate degree, and only 27% have participated in a financial education program. most of the analysis sample is male, white, and married. almost half make over $50k per year and have between $50k and $500k in non-retirement investment assets. table 3 provides a breakdown of the analysis sample between respondents who have a cryptocurrency investment and those who do not. as expected, based on the findings of existing literature, the sample of cryptocurrency investors are a greater percentage male, young, and nonrisk averse (risk seeking). the sample of cryptocurrency investors also shows a higher percentage of respondents using a financial advisor and self-rating investment knowledge compared to the non-cryptocurrency invested sample. table 3. descriptive statistics by crypto investment no cryptocurrency have cryptocurrency variable percent use financial advisor no 41.00% 32.52% yes 59.00% 67.48% objective investment knowledge 0 none correct 1.64% 5.18% 1 9.82% 22.71% 2 20.98% 22.01% 3 31.58% 31.21% 4 all correct 35.98% 18.89% subjective investment knowledge below average (1-4) 35.17% 20.73% average (5) 37.07% 22.59% above average (6-7) 27.76% 56.68% financial education participation no 74.83% 60.64% yes 25.17% 39.36% education level no college 16.07% 21.66% some college 34.53% 41.02% bachelor's degree 29.24% 24.67% post graduate degree 20.15% 12.64% age group 18 24 4.66% 26.05% 25 34 13.61% 38.72% 35 44 11.56% 16.25% 45 54 16.88% 9.81% 55 64 19.26% 6.67% 65+ 34.04% 2.49% gender brockbank et al. 9 female 43.16% 36.91% male 56.84% 63.09% ethnicity non-white 19.84% 42.92% white 80.16% 57.08% marital status not married 40.23% 47.13% married 59.77% 52.87% income range less than $50k 24.49% 35.90% $50k $100k 43.21% 43.10% $100k $150k 19.90% 15.60% greater than $150k 12.40% 5.41% employment status self-employed 9.65% 10.87% employed full-time 38.56% 64.19% employed part-time 6.81% 10.21% retired 35.14% 4.04% not currently employed 9.84% 10.69% non-retirement investment assets $10k or less 18.66% 22.60% $10k $50k 17.42% 27.21% $50k $100k 13.85% 18.39% $100k $250k 16.98% 13.30% $250k $500k 14.72% 11.08% $500k $1 million 10.11% 1.52% $1 million or more 8.27% 5.90% risk aversion 1 – low (very willing) 13.49% 31.46% 2 13.83% 19.79% 3 21.37% 20.65% 4 15.42% 6.99% 5 14.09% 12.94% 6 8.27% 5.40% 7 6.88% 1.83% 8 – high (not willing) 6.67% 0.94% observation count 1,405 151 table 4 shows the marginal effects and standard errors for the probit model. the association between using a financial advisor and investing in cryptocurrency is positive and statistically significant, consistent with hypothesis 1. investors using a financial advisor have a higher financial services review, 33(2) 10 probability of investing in cryptocurrency of about 0.03. table 4. marginal effects of cryptocurrency investment variable (reference group) marginal effect std. err. p-value 95% conf. int. use financial advisor (no) yes 0.0318 * 0.0192 0.097 -0.0057 0.0693 obj. inv. knowledge (none correct) 1 -0.1303 * 0.0747 0.081 -0.2768 0.0161 2 -0.1937 *** 0.0720 0.007 -0.3349 -0.0525 3 -0.1790 ** 0.0706 0.011 -0.3174 -0.0407 4 all correct -0.2255 *** 0.0714 0.002 -0.3654 -0.0857 sub. inv. knowledge (below average) average (5) 0.0167 0.0232 0.472 -0.0288 0.0622 above average (6-7) 0.1315 *** 0.0271 0.000 0.0785 0.1845 fin. educ. participation (no) yes 0.0065 0.0207 0.752 -0.0341 0.0472 education level (no college) some college 0.0032 0.0319 0.921 -0.0593 0.0657 bachelor's degree -0.0198 0.0307 0.520 -0.0801 0.0405 post graduate degree 0.0183 0.0360 0.611 -0.0522 0.0888 age group (18-24) 25 34 -0.1545 ** 0.0775 0.046 -0.3063 -0.0026 35 44 -0.2631 *** 0.0761 0.001 -0.4122 -0.1140 45 54 -0.3294 *** 0.0748 0.000 -0.4760 -0.1828 55 64 -0.3508 *** 0.0745 0.000 -0.4969 -0.2048 65+ -0.4047 *** 0.0734 0.000 -0.5485 -0.2609 gender (female) male 0.0277 0.0213 0.194 -0.0141 0.0695 ethnicity (non-white) white -0.0190 0.0228 0.404 -0.0638 0.0257 marital status (not married) married 0.0427 ** 0.0201 0.033 0.0034 0.0820 income range (less than $50k) $50k $100k -0.0158 0.0290 0.586 -0.0726 0.0410 $100k $150k -0.0130 0.0356 0.716 -0.0827 0.0568 greater than $150k -0.0583 * 0.0353 0.099 -0.1274 0.0109 employment status (full-time) self-employed -0.0240 0.0318 0.450 -0.0863 0.0383 employed part-time 0.0472 0.0401 0.239 -0.0314 0.1257 retired -0.0351 0.0321 0.274 -0.0981 0.0278 not currently employed -0.0628 ** 0.0317 0.047 -0.1249 -0.0008 brockbank et al. 11 table 4 continued variable (reference group) marginal effect std. err. p-value 95% conf. inv. non-ret. inv. assets ($10k or less) $10k $50k 0.0271 0.0328 0.410 -0.0373 0.0914 $50k $100k -0.0209 0.0343 0.542 -0.0880 0.0463 $100k $250k -0.0169 0.0339 0.618 -0.0834 0.0495 $250k $500k 0.0150 0.0360 0.678 -0.0557 0.0856 $500k $1 million -0.1008 *** 0.0338 0.003 -0.1671 -0.0346 $1 million or more -0.0175 0.0458 0.702 -0.1073 0.0723 risk aversion (1 – lowest) 2 -0.0064 0.0327 0.844 -0.0705 0.0576 3 -0.0163 0.0326 0.618 -0.0801 0.0476 4 -0.0750 ** 0.0344 0.029 -0.1424 -0.0076 5 -0.0156 0.0364 0.668 -0.0869 0.0557 6 -0.0257 0.0440 0.559 -0.1119 0.0605 7 -0.0625 0.0427 0.144 -0.1462 0.0213 8 highest -0.0985 ** 0.0446 0.027 -0.1860 -0.0111 number of observations = 1,556 pseudo r-squared = 0.3219 *significance at the 10% level, **significance at the 5% level, ***significance at the 1% level analysis using the 2018 nfcs state-by-state and investor surveys. survey weights are applied. table 4 also shows a negative association between objective investment literacy and cryptocurrency investment, but a positive association between the highest level of subjective investment literacy and cryptocurrency investment. age is related significantly and negatively to cryptocurrency investment. older investors have a substantially lower probability of investing in cryptocurrency compared to younger investors (up to a marginal effect of about -0.40 for those ages 65 and above compared to those between the ages of 18-24). investor risk aversion is also related negatively to cryptocurrency investment in that those not willing to take risks (the highest level of risk aversion) have a lower probability of investing in cryptocurrency of approximately 0.099. results additionally show that married investors have a higher probability of owning cryptocurrencies compared to non-married investors by about 0.043. table 5 shows the results of the additional models estimated to test robustness and find the marginal effect of the key variable, the use of a financial advisor, to remain relatively stable in significance and magnitude on cryptocurrency investment. financial services review, 33(2) 12 table 5. model robustness testing (w/ marginal effects) variable (reference group) full model w/o obj inv know w/o sub inv know w/o educ w/o age w/o gender w/o mar status w/o income w/o inv. assets use fin adv (no) yes 0.0318 * 0.0368 * 0.0383 * 0.0316 * 0.0384 * 0.0306 0.0346 * 0.0314 0.0266 obj inv know (0 correct) 1 -0.1303 * -0.0796 -0.1293 * -0.1115 -0.1286 * -0.1354 * -0.1297 * -0.1045 2 -0.1937 *** -0.1527 ** -0.1941 *** -0.2075 ** -0.1881 *** -0.2011 *** -0.1953 *** -0.1714 ** 3 -0.1790 ** -0.1294 * -0.1814 ** -0.1959 ** -0.1705 ** -0.1868 ** -0.1816 *** -0.1565 ** 4 all correct -0.2255 *** -0.1724 ** -0.2271 *** -0.2493 *** -0.2185 *** -0.2364 *** -0.2284 *** -0.2080 *** sub inv know (below average) average (5) 0.0167 0.0134 0.0168 0.0039 0.0183 0.0169 0.0170 0.0166 above average (6-7) 0.1315 *** 0.1227 *** 0.1304 *** 0.1269 *** 0.1332 *** 0.1335 *** 0.1326 *** 0.1253 *** fin educ participation (no) yes 0.0065 0.0109 0.0101 0.0059 0.0313 0.0064 0.0077 0.0078 0.0047 education level (no college) some college 0.0032 0.0087 0.0023 -0.0095 0.0034 -0.0062 -0.0016 0.0027 bachelor's degree -0.0198 -0.0229 -0.0225 -0.0305 -0.0197 -0.0219 -0.0260 -0.0208 post graduate degree 0.0183 0.0145 0.0067 -0.0188 0.0195 0.0123 0.0028 0.0181 age group (18-24) 25 34 -0.1545 ** -0.1569 ** -0.1353 * -0.1528 ** -0.1607 ** -0.1323 * -0.1557 ** -0.1671 ** 35 44 -0.2631 *** -0.2755 *** -0.2447 *** -0.2593 *** -0.2649 *** -0.2360 *** -0.2698 *** -0.2718 *** 45 54 -0.3294 *** -0.3482 *** -0.3167 *** -0.3234 *** -0.3341 *** -0.3032 *** -0.3345 *** -0.3415 *** 55 64 -0.3508 *** -0.3688 *** -0.3436 *** -0.3455 *** -0.3527 *** -0.3237 *** -0.3554 *** -0.3632 *** 65+ -0.4047 *** -0.4220 *** -0.3898 *** -0.3991 *** -0.4069 *** -0.3791 *** -0.4090 *** -0.4151 *** gender (female) male 0.0277 0.0210 0.0297 0.0283 0.0251 0.0267 0.0291 0.0265 ethnicity (non-white) white -0.0190 -0.0239 -0.0304 -0.0168 -0.0537 ** -0.0181 -0.0156 -0.0192 -0.0208 marital status (not married) married 0.0427 ** 0.0501 ** 0.0469 ** 0.0400 ** 0.0161 0.0421 ** 0.0369 * 0.0443 ** brockbank et al. 13 variable (reference group) full model w/o obj inv know w/o sub inv know w/o educ w/o age w/o gender w/o mar status w/o income w/o inv. assets inc range (less than $50k) $50k $100k -0.0158 -0.0356 -0.0376 -0.0153 -0.0284 -0.0191 -0.0048 -0.0224 $100k $150k -0.0130 -0.0416 -0.0402 -0.0115 -0.0221 -0.0146 0.0041 -0.0283 greater than $150k -0.0583 * -0.0755 ** -0.0747 ** -0.0514 -0.0813 ** -0.0611 * -0.0374 -0.0824 ** emp status (full-time) self-employed -0.0240 -0.0327 -0.0177 -0.0211 -0.0592 -0.0235 -0.0251 -0.0220 -0.0244 employed part-time 0.0472 0.0302 0.0253 0.0507 0.0292 0.0409 0.0507 0.0517 0.0498 retired -0.0351 -0.0343 -0.0458 -0.0325 -0.1619 *** -0.0384 -0.0306 -0.0311 -0.0421 not currently employed -0.0628 ** -0.0694 ** -0.0685 ** -0.0611 * -0.0657 -0.0695 ** -0.0634 ** -0.0614 * -0.0631 ** inv assets ($10k or less) $10k $50k 0.0271 0.0332 0.0388 0.0268 0.0399 0.0284 0.0280 0.0258 $50k $100k -0.0209 -0.0086 0.0196 -0.0227 -0.0038 -0.0196 -0.0208 -0.0255 $100k $250k -0.0169 -0.0140 0.0142 -0.0187 -0.0305 -0.0163 -0.0177 -0.0231 $250k $500k 0.0150 0.0242 0.0511 0.0111 -0.0269 0.0162 0.0125 0.0060 $500k $1 million -0.1008 *** -0.0973 *** -0.0707 ** -0.1020 *** -0.1200 *** -0.0988 *** -0.1029 *** -0.1124 *** $1 million or more -0.0175 -0.0291 0.0268 -0.0220 -0.0225 -0.0173 -0.0206 -0.0316 risk aversion (lowest) 2 -0.0064 -0.0019 -0.0218 -0.0090 -0.0343 -0.0042 -0.0064 -0.0037 -0.0028 3 -0.0163 -0.0155 -0.0378 -0.0182 -0.0396 -0.0173 -0.0198 -0.0155 -0.0115 4 -0.0750 ** -0.0724 ** -0.1096 *** -0.0805 ** -0.1100 *** -0.0756 ** -0.0780 ** -0.0734 ** -0.0664 * 5 -0.0156 -0.0051 -0.0576 -0.0182 -0.0567 -0.0219 -0.0144 -0.0142 -0.0096 6 -0.0257 -0.0261 -0.0678 -0.0294 -0.0676 -0.0314 -0.0258 -0.0195 -0.0236 7 -0.0625 -0.0631 -0.1039 ** -0.0692 * -0.1122 ** -0.0679 -0.0647 -0.0637 -0.0629 8 highest -0.0985 ** -0.1010 ** -0.1428 *** -0.1045 ** -0.1457 *** -0.1056 ** -0.0933 ** -0.0944 ** -0.1040 ** pseudo r-squared 0.3220 0.3051 0.2889 0.3199 0.2476 0.3199 0.3172 0.3197 0.3109 number of observations = 1,556 *significance at the 10% level, **significance at the 5% level, ***significance at the 1% level analysis using the 2018 nfcs state-by-state and investor surveys. survey weights are applied. financial services review, 33(2) 14 the model fit, measured by pseudo r-squared, also remain stable in most estimated models— with one major exception. when age is removed from the model, pseudo r-squared drops from 0.322 (full model) to 0.2476—a difference of 0.0744. this change in explained variance suggests that age has substantial explanatory power in the original model. to gain better knowledge of the relationship between the use of a financial advisor and cryptocurrency investment within the context of the current study, additional literature on the topic is examined. research shows that older individuals are significantly less likely to invest in cryptocurrency (xi et al., 2020) while younger individuals have historically been less likely to utilize the services of a financial advisor, though newer generations are seeking out services earlier (nourallah et al., 2023). lin (2024) observes that millennials search for formal financial advice on average at age 29, while generation x and baby boomers do so on average at ages 38 and 49, respectively. although this trend suggests earlier use of a financial advisor, those in the youngest age group in the sample of the current study still, on average, do not use a financial advisor. these findings suggest that neither of the outermost age categories can be considered the pliable middle, those willing to consider both the use of a financial advisor and cryptocurrency investment. due to these findings, an additional robustness check on the full probit model is estimated in which the youngest and oldest age groups (ages 18-24 and 65+) are removed from the sample to reduce the potential bias that these age categories may present. table 6 describes the findings of this new model which show an increase in the magnitude and significance of the relationship between the use of a financial advisor and cryptocurrency investment. among these middleaged investors (aged 25-64), those who use a financial advisor have a higher probability of having cryptocurrency investments by about 0.05 (compared to 0.03 in the full model). table 6. marginal effects on cryptocurrency investment (w/o age groups 18-24 & 65+) variable (reference group) marginal effect std. err. p-value 95% conf. int. use financial advisor (no) yes 0.0524 ** 0.0257 0.041 0.0020 0.1028 obj. inv. knowledge (none correct) 1 -0.1828 0.1187 0.123 -0.4155 0.0498 2 -0.2655 ** 0.1154 0.021 -0.4918 -0.0393 3 -0.2562 ** 0.1135 0.024 -0.4787 -0.0337 4 all correct -0.3165 *** 0.1140 0.006 -0.5399 -0.0930 sub. inv. knowledge (below ave.) average (5) 0.0509 * 0.0296 0.086 -0.0071 0.1089 above average (6-7) 0.1813 *** 0.0363 0.000 0.1101 0.2524 fin. educ. participation (no) yes 0.0226 0.0282 0.423 -0.0327 0.0780 education level (no college) some college -0.0346 0.0478 0.469 -0.1284 0.0592 bachelor's degree -0.0378 0.0467 0.418 -0.1292 0.0536 post graduate degree -0.0028 0.0527 0.957 -0.1061 0.1005 age group (25-34) 35 44 -0.1098 ** 0.0462 0.017 -0.2003 -0.0193 45 54 -0.1756 *** 0.0441 0.000 -0.2620 -0.0891 55 64 -0.1997 *** 0.0447 0.000 -0.2873 -0.1121 brockbank et al. 15 gender (female) male 0.0442 0.0296 0.135 -0.0138 0.1022 ethnicity (non-white) white -0.0060 0.0296 0.839 -0.0640 0.0519 marital status (not married) married 0.0333 0.0278 0.231 -0.0212 0.0877 income range (less than $50k) $50k $100k -0.0134 0.0409 0.743 -0.0936 0.0668 $100k $150k -0.0141 0.0498 0.778 -0.1117 0.0836 greater than $150k -0.0774 * 0.0465 0.096 -0.1685 0.0137 employment status (full-time) self-employed -0.0503 0.0382 0.188 -0.1253 0.0246 employed part-time 0.0452 0.0583 0.438 -0.0690 0.1594 retired -0.0599 0.0450 0.183 -0.1482 0.0283 not currently employed -0.0463 0.0470 0.325 -0.1384 0.0458 non-ret. inv. assets ($10k or less) $10k $50k 0.0232 0.0450 0.607 -0.0651 0.1114 $50k $100k -0.0772 * 0.0429 0.072 -0.1613 0.0069 $100k $250k -0.0185 0.0462 0.689 -0.1089 0.0720 $250k $500k -0.0075 0.0485 0.877 -0.1025 0.0875 $500k $1 million -0.1265 *** 0.0454 0.005 -0.2154 -0.0376 $1 million or more 0.0048 0.0659 0.942 -0.1244 0.1339 risk aversion (1-lowest) 2 -0.0194 0.0445 0.663 -0.1067 0.0679 3 -0.0389 0.0442 0.378 -0.1255 0.0476 4 -0.0763 0.0481 0.113 -0.1707 0.0180 5 -0.0341 0.0518 0.510 -0.1356 0.0674 6 -0.0746 0.0606 0.218 -0.1933 0.0442 7 -0.0752 0.0580 0.195 -0.1889 0.0384 8 highest -0.1617 *** 0.0527 0.002 -0.2649 -0.0585 number of observations = 889 pseudo r-squared = 0.2255 *significance at the 10% level, **significance at the 5% level, ***significance at the 1% level analysis using the 2018 nfcs state-by-state and investor surveys. survey weights are applied. discussion, limitations, and suggestions for further research human capital theory proposes that increased knowledge should improve decision-making abilities (becker, 1962). given the diversification benefits that cryptocurrency provides (andrianto & diputra, 2017; wong et al., 2018), using the domain-specific human capital of a financial advisor should increase the likelihood of an individual to invest in cryptocurrency. the finding that using a financial advisor is associated positively with cryptocurrency investment financial services review, 33(2) 16 confirms the theoretical expectation and is consistent with hypothesis 1. inconsistent results are found regarding hypothesis 2, in that education and gender are not significantly associated with cryptocurrency investment. additionally, the few areas of significance for income and the level of investment assets are negatively associated with cryptocurrency investment, contradictory to hypothetical expectations and previous literature (lammer et al., 2020; xi et al., 2020). those with the highest level of income, greater than $150k, are significantly less likely to be invested in cryptocurrencies compared to those with income less than $50k. these discrepancies may be explained through the data (international vs u.s. data and/or sample size) or other means, though specifics are only speculation and fall outside the focus of the current study but could provide the foundation for future research. consistent with hypothesis 3, a strong negative relationship between age and cryptocurrency investment is shown in the findings (xi et al., 2020). many additional findings are consistent with theoretical expectations and strengthen the current body of literature on cryptocurrencies. those with higher levels of risk aversion are less likely to invest in cryptocurrencies, consistent with the findings of hackenthal et al. (2022). those with higher subjective investment knowledge have a greater probability of having cryptocurrency investments, consistent with human capital theoretical expectations and existing literature (zhao & zhang, 2021; kim et al., 2023). however, those with a higher objective investment knowledge score have a lower probability of having cryptocurrency investments. this finding is also consistent with kim et al. (2023), but contrary to the findings of arias-oliva et al. (2019) who found financial literacy to be a non-significant determinant of cryptocurrency investment. future research of these apparent discrepancies could be helpful. these findings also suggest that there may be a gap in the subjective and objective investment knowledge measures among cryptocurrency investors, which could provide the foundation for a future research project. it is important to note that the current paper is the first to our knowledge to specify “investment knowledge” instead of a general financial knowledge, and this distinction could be a relevant factor in future research endeavors. lastly, the economic theory of marriage suggests that individuals get married because they believe it will make them better off through pooled resources, combined human capital, risk sharing, and opportunities for specialization (becker, 1973). risk-sharing and combined human capital may explain the increased probability of cryptocurrency investment among married investors compared to non-married ones. a limitation of the current study concerns cryptocurrency risk for investors and advisors. while existing research finds evidence of the positive benefits of cryptocurrency investments, they also acknowledge the extreme volatility. future research could further examine the real risk associated with cryptocurrency investment and compare the volatility with more established investment options, especially post “crash” of late 2022 and the collapse of ftx—one of the largest digital exchange platforms for cryptocurrencies (dai et al., 2022; smith, 2023). additional research also could examine how investors define the use of a financial advisor and determine whether the relationship is formal or informal, and how the advisor provides investment advice. respondents in the current study are asked only if they use a financial advisor when making an investment decision, but not the depth of use (or advice given). other limitations of the study are specific to the data available for cryptocurrency investment. the current analysis sample evaluates the binary decision to invest in cryptocurrency (yes or no) but does not examine the allocation percentages of these investments. information about the volume of investment would provide better insight and transparency into the investors’ intent. there is a vast difference between investing 5% of a portfolio into cryptocurrencies compared to 50% or more, and studying individuals with various levels of cryptocurrency investment would be valuable. longitudinal data also would prove useful in determining a causal relationship between the dependent and main explanatory variable, where the current study (and data) could only determine a correlated brockbank et al. 17 relationship between the use of a financial advisor and cryptocurrency investment. although these limitations exist, it is maintained that this study provides evidence to address a gap in the literature on the relationship between the use of a financial advisor and cryptocurrency investment. conclusion this paper shows that investors who use a financial advisor have a higher probability of investing in cryptocurrency. this finding suggests that financial advisors may be recommending cryptocurrency investment within client portfolios to take advantage of the diversification benefits that are experienced when doing so (andrianto & diputra, 2017; chuen et al., 2018; wong et al., 2018; inci and lagasse, 2019). financial advisors have greater domainspecific human capital in financial matters (through education, experience, and training) and are more likely to engage in the search for information on complex topics. therefore, investors who leverage their financial advisor’s domain-specific knowledge will make informed financial decisions, including the utilization of investments that benefit the diversification of the investment portfolio—i.e., cryptocurrency, which is confirmed by the findings of this study. with the increasing use and interest in cryptocurrency investments, especially in the united states, individuals are challenged with making complex decisions within their investments. financial advisors have a fiduciary responsibility to understand and make recommendations that benefit their clients, including the use of cryptocurrency when necessary. the findings of the current study suggest that these recommendations may already be taking place and emphasize the role of the financial advisor in these decisions. investors expect financial advisors to help optimize their investment returns and financial goals. does an individuals’ use of a financial advisor have an influence on their decision to invest in cryptocurrencies? evidence from the current study answers that question and shows that investors who use a financial advisor are more likely to utilize the diversification benefits of cryptocurrency to optimize their investment portfolio. references allgood, s., & walstad, w. b. (2016). the effects of perceived and actual financial literacy on financial behaviors. economic inquiry, 54(1), 675–697. andrianto, y., & diputra, y. (2017). the effect of cryptocurrency on investment portfolio effectiveness. journal of finance and accounting, 5(6), 229–238. arias-oliva, m., pelegrín-borondo, j., & matíasclavero, g. (2019). variables influencing cryptocurrency use: a technology acceptance model in spain. frontiers in psychology, 10, 475. ayedh, a., echchabi, a., battour, m., & omar, m. (2020). malaysian muslim investors’ behaviour towards the blockchain-based bitcoin cryptocurrency market. journal of islamic marketing, 12(4), 690– 704.this barthel, a. c., & lei, s. (2021). investment in financial literacy and financial adviceseeking: substitutes or complements? the quarterly review of economics and finance, 81, 385–396. becker, g. s. (1962). investment in human capital: a theoretical analysis. journal of political economy, 70(5, part 2), 9–49. becker, g. s. (1973). a theory of marriage: part i. journal of political economy, 81(4), 813–846. best, r. (2025). crypto market cap 2010-2025. statista. https://www.statista.com/statistics/7308 76/cryptocurrency-market-value/ blackstone, t. (2025). 2025 cryptocurrency adoption and consumer sentiment report. security.org. https://www.security.org/digitalsecurity/cryptocurrency-annualconsumer-report/ cummings, b. f., & james iii, r. n. (2014). factors associated with getting and dropping financial advisors among older https://www.security.org/digital-security/cryptocurrency-annual-consumer-report/ https://www.security.org/digital-security/cryptocurrency-annual-consumer-report/ https://www.security.org/digital-security/cryptocurrency-annual-consumer-report/ financial services review, 33(2) 18 adults: evidence from longitudinal data. journal of financial counseling and planning, 25(2), 129–147. dai, p. f., goodell, j. w., huynh, l. d. t., liu, z., & corbet, s. (2022). understanding the transmission of crash risk between cryptocurrency and equity markets. financial review, 58(3), 539– 573. dasman, s. (2021). analysis of return and risk of cryptocurrency bitcoin asset as investment instrument. in n. m. alsharari (ed.), accounting and finance innovations (pp. 1–14). intechopen. gibbons, r., & waldman, m. (2004). taskspecific human capital. american economic review, 94(2), 203–207. grand view research. (2023). u.s. cryptocurrency market size and share report, 2030. grand view research. https://www.grandviewresearch.com/ind ustry-analysis/us-cryptocurrencymarket-report gupta, s., gupta, s., mathew, m., & sama, h. r. (2020). prioritizing intentions behind investment in cryptocurrency: a fuzzy analytical framework. journal of economic studies, 48(8), 1442–1459. hackethal, a., hanspal, t., lammer, d. m., & rink, k. (2022). the characteristics and portfolio behavior of bitcoin investors: evidence from indirect cryptocurrency investments. review of finance, 26(4), 855–898. hanna, s. d. (2011). the demand for financial planning services. journal of personal finance, 10(1), 36–62. hanna, s. d., & lindamood, s. (2010). quantifying the economic benefits of personal financial planning. financial services review, 19(2), 111–127. hershey, d. a., walsh, d. a., read, s. j., & chulef, a. s. (1990). the effects of expertise on financial problem solving: evidence for goal-directed, problemsolving scripts. organizational behavior and human decision processes, 46(1), 77–101. ibrahim, m. (2018). interactive effects of human capital in finance–economic growth nexus in sub-saharan africa. journal of economic studies, 45(6), 1192–1210. kim, k. t., hanna, s. d., & lee, s. t. (2023). investment literacy, overconfidence and cryptocurrency investment. financial services review, 31(2/3), 121–132. lammer, d. m., hanspal, t., & hackethal, a. (2020). who are the bitcoin investors? evidence from indirect cryptocurrency investments (safe working paper series no. 277). leibniz institute for financial research safe. lev, b., & schwartz, a. (1971). on the use of the economic concept of human capital in financial statements. the accounting review, 46(1), 103–112. lin, n. (2024). younger generations seek out advisers earlier. planadviser. https://www.planadviser.com/youngergenerations-seek-advisers-earlier/ lin, q., & lee, j. (2004). consumer information search when making investment decisions. financial services review, 13(4), 319–332. nourallah, m., öhman, p., & amin, m. (2023). no trust, no use: how young retail investors build initial trust in financial robo-advisors. journal of financial reporting and accounting, 21(1), 60–82. pilote, p. e., boulianne, e., & magnan, m. (2024). impact of the financial advisor on clients’ financial outcomes: an integrative model. financial services review, 32(3), 32–67. rauch, a., & rijsdijk, s. a. (2013). the effects of general and specific human capital on long–term growth and failure of newly founded businesses. entrepreneurship theory and practice, 37(4), 923–941. smith, h. (2023, may 22). bitcoin price crash: what caused the latest collapse?. times money mentor. https://www.grandviewresearch.com/industry-analysis/us-cryptocurrency-market-report https://www.grandviewresearch.com/industry-analysis/us-cryptocurrency-market-report https://www.grandviewresearch.com/industry-analysis/us-cryptocurrency-market-report https://www.planadviser.com/younger-generations-seek-advisers-earlier/ https://www.planadviser.com/younger-generations-seek-advisers-earlier/ brockbank et al. 19 https://www.thetimes.co.uk/moneymentor/article/is-bitcoin-crash-coming/ smith, h., finke, m., & huston, s. (2012). the influence of financial sophistication and financial planners on roth ira ownership. journal of financial service professionals, 66(6), 69. smith, g. e., venkatraman, m. p., & dholakia, r. r. (1999). diagnosing the search cost effect: waiting time and the moderating impact of prior category knowledge. journal of economic psychology, 20(3), 285–314. sommer, m., lim, h., & macdonald, m. (2022). financial advisor use, life events, and the relationship with beneficial intentions. financial services review, 30(1), 69–88. stigler, g. j. (1961). the economics of information. journal of political economy, 69(3), 213–225. stosic, d., stosic, d., ludermir, t. b., & stosic, t. (2019). exploring disorder and complexity in the cryptocurrency space. physica a: statistical mechanics and its applications, 525, 548–556. sullivan, p. (2010). empirical evidence on occupation and industry specific human capital. labour economics, 17(3), 567– 580. xi, d., o’brien, t. i., & irannezhad, e. (2020). investigating the investment behaviors in cryptocurrency. the journal of alternative investments, 23(2), 141–160. yilmaz, n. k., & hazar, h. b. (2018). predicting future cryptocurrency investment trends by conjoint analysis. journal of economics, finance and accounting, 5(4), 321–330. zhao, h., & zhang, l. (2021). financial literacy or investment experience: which is more influential in cryptocurrency investment?. international journal of bank marketing, 39(7), 1208–1226. zweig, j. (2021, october 22). why your adviser might start talking up bitcoin. the wall street journal. https://www.wsj.com/articles/why-youradviser-might-start-talking-up-bitcoin11634914812. https://www.thetimes.co.uk/money-mentor/article/is-bitcoin-crash-coming/ https://www.thetimes.co.uk/money-mentor/article/is-bitcoin-crash-coming/ pii: s1057-0810(99)00019-0 book, software and website reviews history of the eighties, lessons for the future,vol. 1 federal deposits insurance corporation, 1997, 572 pp. american finance for the twenty-first century robert e. litan, united states department of treasury, nov. 7, 1997, 162 pp. asset securitization—liquidity and funds management, controller’s handbook. washington, d.c. nov. 1997, 90 pp. the purpose of this review is to alert the reader to the recent publication of three separate reports by federal agencies. all of the reports are potential additions to a reading list in a financial institutions course. the thesis of the fdic report, histories for the eighties, lessons for the future, is that bank failures in the 1980s did not have a short list of causes. rather they resulted from the concurrence of a number of forces-economic, financial, legislative, supervisory, and managerial-which working together produced a decade of bank crisis. the first section of the report provided a review if the challenges to the banking industry created by the operating environment in the 1970s. the legislative mandates of the 1980s and 1990s are summarized with special attention paid to the impact these changes had on bank failures. the section concludes with a discussion of the interagency conflicts between fdic and occ. in the second section of the study, sectoral and regional crises are considered. the topics covered include: ldc debt, mutual savings banks, the “too big to fail” issue, and the cause of various regional bank crises. the third section of the study discusses the common elements in the economic environment that characterized the various bank crises and the set of bank behaviors that amplified their impact. the section concludes with an interesting section on fraud and financial misconduct followed by a brief statement on the difficulty the regulatory community has in separating managerial deficiencies and the slack internal control system from fraud. the final section deals with the behavior and performance of the supervisory agencies during this period and the legal and regulatory constraints they operate under. the thrust of this discussion is to present the thesis that the supervisory function was handicapped by a philosophy that stresses identifying unacceptable levels of risk as observed on the financial reports of the banks. the study suggests that the more appropriate policy would have been to take a proactive view if risk that stressed policies that would help identify and contain risk financial services review 7 (1998) 317–319 1057-0810/98/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00019-0 as it developed. in a review of the role of the fslic was limited in how it could respond to industry problems by severe budget constraints caused by the differentially poor funding it received relative to other federal regulatory agencies. a review if occ examiners reports for the period indicates that examiners repeatedly reported problems with the internal control systems over lending operations but they were ignored until the late eighties. it is interesting to note that academic literature has devoted considerable time reflecting on the adverse incentive system created by the flat-rate deposit insurance premiums; an element in the discussion that is not given much creditability by this study. in preparing this report the fdic has been successful in creating a readable commentary on forces that influenced bank failures in the decade of the eighties. the stated purpose of american finance for the twenty-first century is to discuss the changes in the financial service industry through 2010, and to review the adequacy of existing statutes and regulations. the general thrust of the report is to provide an all encompassing review of the forces influencing the financial system (without covering any topic in depth). this makes the report excellent reading for an undergraduate issues oriented course. one of the pluses to the report are the brief historical summaries of policy issues. the assertion, which i regard as essentially correct, is that today’s policy initiatives and regulations are oriented toward preventing a second great depression. what is needed is a paradigm shift that builds a policy that seeks to contain failure within boundaries and recognize that markets reward players that embrace change. the report also argues that the financial markets are becoming more complex, and that legislative and regulatory activities that seek to divide the financial world into discrete segments is doomed to failure. this policy should be oriented toward identifying, isolating and disposing of trouble spots that endanger the stability of the whole system. the report is built around five distinct themes. the study opens with a snapshot view of the financial services industry and its function. the discussion is then broadened to include globalization and a variety of changes in the information system. these forces of change, it is argued, have unleashed a new round of competition that requires a change in the orientation and thrust of our anti-trust efforts. the new policy must monitor the “gatekeepers and keyholders” that control networks and technology. the goal of such a policy should be to prevent a financial meltdown. the concluding section of the study deals with the broad social theme that we need to work to open the financial system to all americans. the occ asset securitization—liquidity and funds management, controller’s handbook is divided into three chapters, two of which provide an excellent overview of the structure of the securitization transaction, the mechanics of cash flow, and the risks associated with the securitization activities. although the discussion on the structure of securitization can be found in a variety of sources, the detail and simplicity with which the process is covered and the example provided—the credit card securitization process—make the discussion a winner for an undergraduate presentation. the discussion of the risk associated with the securitization is simple and much more complete than that found in alternative sources. the third chapter is not of general interest but rather provides a discussion of the examination objectives as they relate to the securitization 318 book review / financial services review 7 (1998) 317–319 process. other comptroller’s handbooks of possible interest include, capital and dividends and mortgage banking. austin h. spencer professor of economics and finance, western carolina university, school of business, economics & finance, cullowhee, nc 28723-9034, usa e-mail address:aspence@wcu.edu 319book review / financial services review 7 (1998) 317–319 pii: s1057-0810(96)90005-0 personal finance: an alternative approach to teaching undergraduate finance jill lynn vihtelic effective teaching invites students into the discipline and helps them to see and make connections between the discipline’s content and their lives. this paper identifies an alternative approach to effective teaching of undergraduatefinance. personalfinance, as opposed to managerialfinance, provides a more appropriatefoundation on which to center the undergraduate$nance curriculum. it better matches students’ interests, per sonal experiences, and cognitive structures. this paper takes the position that personal finance should precede managerial finance as the introduction and start to the finance major in the undergraduate business curriculum. i. introduction a problem shared by many business schools these days is that of declining enrollments. nationwide, the american assembly of collegiate schools of business (aacsb, 1996) reports bachelors enrollments at accredited business schools off 16.4 percent from 1990 levels. similarly, the number of finance bachelors degrees conferred dropped 13.5 percent between 1989 and 1993 (u.s. department of education, 1991, 1995). some schools have the compounded predicament of a simultaneous loss of the “middle” group of students; finance professors talk of experiencing a bi-modal distribution within their classes. while much debate surrounds these issues, questions about teaching effectiveness come to mind. could it be that the introductory organizational structure of the discipline is responsible for limiting students’ entry to finance? how can finance professors, through their teaching, pedagogy, and classroom climate, better invite students into the discipline? the purpose of this paper is to identify an alternative approach to teaching finance which will (a) invite a larger, diverse group of students into the discipline and (b) improve undergraduate finance education for all. it takes the position that personal finance provides an effective approach to teaching undergraduate finance. personal finance should be taught as the first course in the undergraduate finance curriculum. it should be the introductory jill lynn vihtelic l p.o. box 46, madeleva hall, saint mary’s college, notre dame, in 46556. 120 financial services review 5(2) 1996 and prerequisite course for the major. personal finance issues should be the central and uni fying thread throughout the undergraduate curriculum. basic financial concepts, vocabu lary, and principles should be taught first in terms of personal finance. it should serve as the foundation building block for further financial knowledge. corporate finance topics should be introduced to the curriculum after students have a firm base on which to build and accept new knowledge. part i of this paper reviews previous research on finance education. part ii discusses student-centered learning. an examination of student-centered ways of knowing applied to teaching finance follows in part iii. the paper concludes by describing personal-centered finance and outlining steps required toward a theory of instruction in finance. ii. literature review a. the undergraduate finance curriculum much has been published about financial education and the undergraduate curriculum. currently, and throughout its history, finance education has centered on business finance theory and practice. brigham (1973) provided the rationale that as part of the business school curriculum, the finance core should remain oriented toward business finance. tra ditionally, the finance curriculum has had three prongs: corporate or business financial management, investments, and financial markets and institutions (brigham, 1973). how ever, there is often a wide gap and general confusion between what executives, accrediting agencies, and finance faculty members view as the ideal curriculum. based on their survey of bankers, corporate financial officers, and finance academicians, demong, pettit, and campsey (1979) suggested that a finance curriculum for the future should accentuate com munication and people skills as well as analytical skills. they found that academicians believed there to be an increasing need for quantitative and computer skills, while practi tioners emphasized people and communication skills. conflicts are also apparent in comparing the porter-mckibbin (1988) report, commis sioned by the aacsb, to the mcwilliams and pantalone (1994) survey results on 800 finance executives. new aacsb standards, partly in response to the porter-mckibbin report, are designed to allow for greater diversity in mission and curriculum for business schools, and are thought to encourage a broader undergraduate education. yet when asked to define the optimal curriculum for an undergraduate finance student, the financial execu tives chose a curriculum which placed emphasis on business, and especially on finance and accounting courses (mcwilliams & pantalone, 1994). while the debate over the curriculum for finance students continues, the introductory finance course content remains focused on business. berry and farragher (1987) surveyed 549 academic members of the financial management association (76 percent of which were aacsb-accredited institutions) to determine characteristics of the introductory finance course. they found topic coverage to be fairly evenly distributed among financial management topics, with capital budgeting getting the most attention, followed by the time value of money, and the cost of capital/capital structure. bialaszewski, pencek, and ziet low (1993) reported that although much flexibility is allowed in the design of aacsb accredited programs, all (100%) of their survey respondent group required financial man personal finance 121 agement i as the introductory finance course. this was especially noteworthy since the bialaszewski, et al. (1993) respondent group offered many “nontraditional” finance courses; 38 percent regularly offered personal financial management; 25 percent offered life insurance; 13 percent offered estate planning; 25 percent offered property and liabil ity insurance; and 65 percent offered real estate. at a tutorial presentation, merton and bodie (1995) demonstrated how the first course in finance can be a general introduction to the whole field of finance. they suggested that various subfield topics be presented within a single unifying framework. b. the personal financial planning major the concept of personal financial planning (pfp) as a major or upper level finance course in the business program has also received academic scrutiny. many found the finan cial services industry in general, and specifically the field of pfp, in need of people well educated and prepared to enter the field (daigler, 1979; gitman & bacon, 1985; higgins, 1983; lavine, 1987; ulivi, 1982). in response to this need, daigler (1979) suggested a cur riculum for a pfp major. along the same line, higgins (1983) argued that colleges of busi ness and departments of finance should provide undergraduate students with the option to prepare for careers in financial planning. his model recognized that all of the primary areas of financial planning, namely insurance, taxation, and investments, with the possible exception of pension and estate planning, were already offered by well established depart ments of business administration. higgins advocated that implementing the pfp major could be accomplished easily by adding to the already existing courses an upper-level introduction to financial planning course. since most universities were already teaching a pfp course geared to nonmajors, the up-grade/switch to a comprehensive professional course was judged quite feasible (d’ambrosio 1980; higgins 1983). ulivi (1982) proposed three alternatives for universi ties to follow in meeting the challenge posed by the increased importance of financial services: (a) offer a one-semester course in comprehensive financial planning, (b) offer an interdisciplinary certificate of financial planning, or (c) offer a new major in financial services. gitman and bacon (1985) presented an explanation of the comprehensive financial planning process to argue for establishment of a pfp major. they described the basic ele ments of a comprehensive personal financial plan as closely related to traditional business finance education topics, yet sufficiently novel to warrant the establishment of a new major. for example, comprehensive personal financial plans include elements familiar to traditional business finance majors such as the statement of financial position, the state ment of cash flows, a cash budget, an investment portfolio analysis, and an analysis of insurance needs. other elements of a comprehensive plan are not traditionally part of the business finance curriculum. examples of these include employee benefits analysis, retire ment planning, estate planning, planning for educational funding, issues relating to closely held businesses, estate and gift tax planning, and goal setting. within the same article, gitman and bacon (1985) also reported the results of their 1984 survey of business school financial curriculums. as expected, respondent schools clearly emphasized traditional business finance topics. only 5 percent offered an under graduate financial services curriculum, but 21 percent of those respondents not then offer ing a financial services curriculum indicated that they intended to do so in the future. 122 financial services review 5(2) 1996 c. finance pedagogies many publications describe finance pedagogies. kalogeras (1976) provides a sum mary and critique of various pedagogies including modular learning, case method, the game approach, computer assisted instruction, and the traditional lecture. some articles explore ways to better explain financial concepts to students. for examples see horvath (1985) “a pedagogical note on inter-period compounding and discounting,” or kochman (1986) “intrayear discounting: uses and misuses.” recently, many authors describe ped agogies designed to transform the classroom into an active and realistic forum. for exam ple, dyl (1991), then chan, weber and johnson (1995) and graham and kocher (1995) describe using popular movies in the introductory finance class to better students’ under standing of the business. nofsinger (1995) describes a multi-media project for the capstone course in finance which integrates the real-world topics of ethics and social responsibility with the finance of mergers and acquisitions. lange (1993) and yoon (1995) present spreadsheet approaches to teaching realistic financial concepts. lawrence (1994) details experiences of establishing real as opposed to simulated student investment funds at us universities. while pedagogic articles abound, relatively little has been done to assess various ped agogic approaches. krishnan, bathala, bhattacharya, and ritchey (1996) have recently surveyed students from three different business schools to evaluate their perceptions and expectations about finance and the introductory finance course. preliminary results lead the authors to conclude that a significant number of business students find the introductory finance class neither useful nor interesting. differences between the number of preand post-course survey responses suggest that a large percentage (about 25%) of students drop the course. in forthcoming work, these authors plan to address why the introductory finance course fails to engage a significant number of students. they will also offer their suggestions to make the finance course work more appealing. d. finance student learning working with education and psychology experts, gentry has explored finance student learning (helgesen & gentry 1988; pratt & gentry 1994; paulsen & gentry 1995; helge sen & gentry 1995). pratt and gentry (1994) observed that finance professors are serious about their teaching effectiveness and are conscientious in trying to help students learn. helgesen and gentry (1995) noted, however, that professors’ research activities are closely aligned to the theoretical structure of the discipline as opposed to research focused on learning within the discipline. pratt and gentry (1994) explained that there are several dif ferent learning styles and found that finance students seldom learn the same way as their professors. differences also exist with respect to learning styles among students (helgesen & gentry 1995); female business students process information in significantly different ways than male business students; nonmajors process information differently than business majors. helgesen and gentry (1995) commented on the statistical under-representation of african americans as business students. they suggest that finance professors interested in diversity must explore the discourse of the discipline to make their classrooms equitable to all types of students. personal finance 123 iii. student-centered learning emphasis on student experiences rather than teacher authority can be viewed as the critical factor in achieving true understanding. with this approach, the teacher’s role is to help stu dents code information in ways that are appropriate to the individual. this requires teachers to evaluate students and their previous experiences in order to help them place new knowl edge in their repertoires. this is very different from the traditional approach to teaching finance that views the class as content driven, based on theory, and rewards products (cor rect answers) rather than process and inquiry. john dewey’s (1974) dictum to “start where the students are” is often ignored by col lege finance teachers who find it overwhelmingly complex to evaluate students’ diverse educational, cultural, work, and family life experiences. in graduate finance programs, rel evant work experience often provides the common base on which to ground advanced stud ies. a work experience requirement is not feasible, however, for acceptance to undergraduate programs. most undergraduate finance curriculums prescribe prerequisites such as principles of accounting and economics and often a math course such as statistics. but unless these courses are grounded in students’ experiences, they may also fail to make connections. for example, at many schools the principles of finance course prerequisites include a full year (two courses) in accounting. none-the-less, many principles of finance texts and professors start off the semester with a thorough review of financial statements. it is perplexing that often students who were very successful in accounting, even declared accounting majors, cannot distinguish between the income statement and the statement of financial position when in thefinance classroom. why this happens can be best understood in terms of missed connections. the accounting knowledge for many of the students is coded too functionally and not connected to other knowledge held. it only works in their accounting classes to answer accounting problems. similarly, connections can fail within the finance classroom. ratio analysis is a staple of finance texts and courses. yet it is not unheard of for teachers of the business policy cap stone course to complain that students do not know ratio analysis. could it be that students are blatantly lying when they deny having a thorough introduction to ratio analysis? per haps, but it is also possible that students are sincere in their denial; the knowledge connec tion fell short. in order for students to use specific finance know-how and techniques outside the finance classroom, the finance knowledge must be generically coded for future access. educational psychologist jerome bruner’s (1966, 1973, 1977, 1986) structuralist views on knowledge and education provide a practical corollary. he contends that a person actively constructs knowledge by relating new information to a previously acquired frame of reference, called the coding system. knowledge is an active process that requires leam ers to code information into storage systems for future retrieval and use. bruner asserts that the problem with memory is not storage space, but retrieval. the best coding systems for efficient retrieval are expressed in terms familiar to the learner. as the learner masters domains of knowledge through discovery, she re-codes the storage system to make it more generic and thus more useful. bruner (1977) makes four claims as to the importance of discipline structure: l understanding fundamentals makes a subject more comprehensible; l unless detail is placed into a structured pattern, it is rapidly forgotten; 124 financial services review 5(2) 1996 l understanding fundamental principles and concepts is the main way to transfer training; l emphasis on structure allows re-examination of fundamentals, thus narrowing the gap between elementary and advanced knowledge. bruner contends that all teaching and learning should be connected to the broader fun damental structure of the subject. in addition to being intellectually exciting, learning is more likely to be remembered if it is tied together. such emphasis on structure promotes transference of knowledge; good theory explains and is remembered. further, the role of structure in learning takes on increased importance as today’s students have limited expe riential exposure to the materials they learn. that finance requires an organizational structure is neither a new nor a controversial idea. from its inception, the principles of managerial finance course has served as the base structure of the discipline. traditional finance curriculums offer the principles of manage rial finance course first and as a prerequisite to further finance course work. upper level courses typically concentrate on one section from the principles framework, teaching more advanced concepts in areas such as working capital management, investments, or capital budgeting. many curriculums offer a senior capstone experience, which revisits theory and application at a higher and deeper level of understanding and instruction. although it is clear that the traditional structure has lead to successful coding systems and transference of finance knowledge for many, there are others whose access to the discipline and its powers may have been unwittingly barred by the traditional managerial/corporate structure. could it be that many students become turned off to finance because it seems disconnected from their lives? does newly learned managerial-focused finance knowledge simply get lost in students’ coding and retrieval systems because connections are not made to students’ lives? “an alternative way to “start where the students are” is to start with personal finance. teaching personal finance first promotes connected knowing by linking what the students already know from their own financial experiences to the new knowledge presented in the course. the framework for knowing finance can thus be constructed by the individual learner, encouraged through self-discovery, and specific to each knower. teaching per sonal finance as the first course in the discipline uses student centered experiences as the platform from which finance learning proceeds. as a coding system, personal finance is an effective cognitive structure since it is familiar to learners. personal finance provides a beneficial organizing structure for the discipline since it is couched in terms of students’ interests and life experiences. students have a positive atti tude toward learning personal finance since they recognize the materials as being highly relevant to future use. they have many questions they want answered and they are aroused with interest. for example, as they get closer to graduation students become increasingly aware of the values of job options they hold, the need to consider and protect against the risks of poor health and loss of income, and liability exposures due to car ownership. they can appreciate the benefits of learning topics such as cash budgeting before they are truly independent of their families and facing a real life cash crisis. all have some previous per sonal financial experience and a certain degree of finance mastery. they have checking accounts, most have credit cards and some savings and investing experience--even if only with a bank savings account or series ee bonds. new and diverse situations, from changes in financial markets and instruments to changes in the family life cycle and goals, force stu dents to apply old knowledge to new situations, making it more likely that the financial personal finance 125 knowledge will be coded generically as opposed to concretely. finance can be illustrated to students in terms of their own lives, which they can understand, appreciate and practice. personal finance can abet the invention and creation of student coding systems. the personal financial planning process involves stages of planning activities from gathering and processing and analyzing client information, to developing and presenting the plan to the client, and finally to implementing and monitoring the plan (hallman & rosenbloom, 1987; gitman & joehnk, 1996). by emphasizing and teaching the process, answers become less significant than inquiry and procedure. as noted by higgins (1983), comprehensive personal financial planning considers a broad array of topics including insurance, taxation, investments, and retirement planning. students are encouraged to look at financial posi tions and goals in a holistic manner. for example, cash flow shortages are not seen as merely a budgeting problem, but also as the consequences of life-style and career choices, and spending and saving patterns. students’ intuition and personal experiences can be val idated as appropriate sources for personal financial solutions. analytical procedures can be applied by students to test their hunches, encouraging them to try out ideas and invent new solutions. iv. student-centered knowing a quick fix and reorganization of the curriculum placing personal finance as the first course taught in the curriculum will have only limited success unless it is coupled with an expansion of the discipline. economist jean shackelford (1992, p. 57 1) asserts that “how one teaches is as important as what one teaches.” including students’ ways of knowing finance, matched to appropriate pedagogies, will more effectively invite them to join the discourse of the discipline. the personal financial management course offers a unique opportunity within the finance curriculum for learning to be student-centered. silence in the classroom is often misinterpreted by the professor (belencky, et al., 1986). rather than being an indication of lack of ability or preparedness, silent students are often turned off due to the clash between their own view and that presented by the disci pline, the professor, and the dominant group of classroom peers. instead of viewing class room silence as a student deficiency, finance professors should heed silence in the classroom as a call to incorporate alternative learning approaches and pedagogies to increase participation. to find successful examples of alternative pedagogies, one has only to look to gradu ate business education where these teaching methods dominate. instruction is often rooted in authentic, real-world situations making use of rich and complex interdisciplinary cases and/or consulting work. graduate students have the opportunity to work in groups and engage in meaningful functional tasks. the textured nature of real cases demonstrates to graduate learners the importance of multiple perspectives and considering all sides to every problem. the small group format is often used to give graduate students the opportunity to reflect on their projects and then to talk to the professor in shared learning environments. professors model the skills to be taught by sharing expert thinking with students and by coaching them which entails more collaboration than evaluation. most graduate programs emphasize a collaborative approach which prepares students for the teamwork environ ment of u.s. businesses. 126 financial services review 5(2) 1996 to avoid “turning off’ students, undergraduate finance should be taught incorporating the alternatives found in graduate business education. these pedagogies can be used effec tively in all finance classes, including corporate and managerial finance classes. personal finance, however, provides a unique opportunity within the undergraduate finance curricu lum to incorporate not only alternative pedagogies, but also alternative ways of knowing. the content of personal finance lends itself to shape learning episodes consistent with different modes of knowing. for example, personal finance centers on questions of human import. ethical themes of care and duty can be easily incorporated in discussions centering on family financial goal setting. personal finance is experiential and can include reflective practice. it offers students an oppo~unity to practice and reflect using their own life cir cumstances, their families, and friends, cases can be used to encourage dialogue, asking students what they would do in this situation, and focus on how they would implement pro posed changes. personal finance is process rather than product oriented; by teaching the personal financial planning process, answers are viewed as less significant than inquiry and procedure. personal finance is holistic; comprehensive personal financial planning consid ers a broad array of topics and encourages students to look at financial positions and goals in a holistic manner. personal finance is intuitional; students’ intuition and personal expe riences can be validated as appropriate sources for financial solutions. finally, personal finance is experimental. analytical procedures can be applied to test student’s hunches, encouraging learners to try out their own ideas and invent new solutions. v. personal-centered finance as previously noted, many authors have discussed the concept of and pedagogical issues su~ounding personal financial planning education as a major or upper level course in the business program (daigler, 1979; d’ambrosio, 1980; gitman & bacon, 1985; higgins, 1983; lavine, 1987; ulivi, 1982). merton and bodie (1995) presented the notion that the first course in finance should be an introduction to the whole field. further, the first course should be taught within a single unifying framework. this paper takes the position that per sonal financial m~agement should be used as the unifying structure and framework for presenting principles of finance. it has been suggested that most principles of finance courses, and the textbooks, try to cover too much material and present too much detail for an introductory course (merton & bodie, 1995; krishnan, et al., 1996). finance professors need to decide on the top ten “big ideas” necessary to acquaint, welcome, and encourage students to the field of finance. once decided, these big ideas can be effectively introduced from the perspective of per sonal finance. table 1 shows an example of the large overlap between managerial finance topics and personal finance topics in two popular texts. while some differences in topic coverage can be found, the similarities are striking. the only business finance topic without a corollary in personal finance is the weighted average cost of capital. note, however, that even the component costs of capital (kp, kd, k,) can be approached from the personal side of the transaction as the investor’s required return on investment. personal finance topics are sufficiently broad to provide a strong foundation on which to base further finance learning. a thorough presentation of personal finance as the preliminary course can provide students with an alternative yet rigorous introduction to the discipline. personal finance 127 table 1 comoarison between two texts: chauter titles. numbers. and tonics fundamentals of financial management (brigham and houston 1996) personal financial planning (gitman and joehnk 1996) overview of financial management, #l, goals of the corporation; importance of financial management; careers in finance financial statements, cash flow, and taxes, #2, balance sheet; income statement; statement of cash flows; the federal income tax system analysis of financial statements, #3, ratio analysis the financial environment: markets, institutions, and interest rates, #4, financial markets; term structure theories risk and rates of return, #5, stand alone and portfolio risks; capm time value of money, #6, present and future values; uneven cash flow streams bonds and their valuation, #7, finding expected interest rate k,+ default risk stocks and their valuation, w, common and preferred stock, rights and privileges the cost of capital, #9, wacc; mcc; k,; kd; k, the basics of capital budgeting, #lo, npv; irr; mirr risk and other topics in capital budgeting, #l 1, estimating cash flows; the optimal capital budget capital structure and leverage, #12, target capital structure; business and financial risk, capital structure theory dividend policy, #13, theories; policies; practices; factors influencing financial forecasting, #14, projected financial statement method: afn managing current assets, #15, cash management; cash budget; techniques; marketable securities; inventory control understanding the financial planning process, #l, types of financial goals; the rewards of sound financial planning; planning your career your financial statements and plans, #2, balance sheet; income and expenditures statement; cash budget managing your taxes, #3, principles of federal income taxes; calculating and filing taxes solvency, liquidity, savings, and debt service ratios presented in #2; book value, roe, eps, p/e., beta in #i i making security transactions, #12, securities markets and information; managing investments; risk-return relationship, investment risks, and yields covered in #l 1 time value of money concepts, present and future values presented in #2 investing in stocks and bonds, #l 1, stocks; bonds; preferreds; convertibles; yields and returns investing in mutual funds, #13, basics; types and services; asset allocation stocks covered in ,911 and #12; making housing and automobile decisions, #5; buy vs. lease; financing insuring your life, f%, life insurance insuring your health, #9, health and long-term care: disability income insurance protecting your property, #i 0, property and liability insurance; auto insurance borrowing on open account, #‘6, basic concepts of credit. using consumer loans, #7, managing your credit cash and stock dividends, dividend reinvestment plans, growth discussed in #l 1 meeting retirement goals, #l4, pension plans; annuities; 401(k); social security preserving your estate, #15, wills; trusts; gift and estate taxes managing your cash and savings, #4, role of cash management; the new financial marketplace this paper calls for finance professors to join the education movement to collabora tively develop a theory of finance instruction which specifies: 1. the experiences which most effectively implant in the student a predisposition toward learning; 128 financial services review 5(2) 1996 but what these experiences should be we cannot at present say. collaborative research with colleagues from education and psychology could provide answers. what financial, life, and learning experiences embed in students a thirst for learn ing finance? what causes students to yearn for and to appreciate the power of financial knowledge? as financial experts, finance professors’ input and research efforts are required in this endeavor. a favorable predisposition to learning is also dependent on cultural, motiva tional, and personal factors. for example, the relationship between the college finance teacher and the student might be a factor in whether the student likes finance or not. setting an inviting class climate in finance may be motivational to students as well as “conducive to more effective teaching, and therefore, more effective learning on the part of all students” (sandler & hoffman, 1992, p. 1). there has been much written recently about girls and math, and how girls are denied an equal education (sadker, 1994). but finance academe has not yet addressed the implications of these findings. for instance, if we find that women students do not relate to learning finance when they have come from a family where the man makes all the financial decisions, then what can we do as teachers to stimulate interest? the impact of culture, motivation and personal factors on finance pedagogy needs our scholarly attention and exploration. 2. ways to structure the knowledge so that it can be readily grasped; this paper argues that personal finance presents a powerful structure for sim plifying information, for generating new propositions, and for increasing the manipulability of financial knowledge for undergraduate students. 3. the most effective sequences in which to present materials; for example, should concrete examples always be presented before notation? what symbolic forms of representation do learners hold at various stages of understanding? how do individual factors, such as differences with respect to experiences or past learning, impact the optimum sequence? research is required that identifies the basic building blocks of financial learning and explores their sequencing. students enter finance classes with a wide variety of previous personal financial experiences. the sequencing of finance materials should be designed to provide the missing yet essential building blocks for those with little experience. 4. the nature and pacing of rewards. rewards are either provided from extrinsic sources to the learner such as the teacher, or from intrinsic sources within the learner. intrinsic rewards are more sustainable and thus more desirable. the ideal situation is for knowledge of results and the ability to make corrections to be the learner’s reward; this allows for the learner to achieve independent mastery. too often in finance, as with other college classes, the reward from the knowledge of results is provided only from the graded chapter test, quiz, or homework assignment. an exam becomes the end of the unit rather than a way to communicate the learner’s progress and take corrective actions. the goal is for students to recognize on their own when comprehension is incomplete and to seek help from the instructor. personal finance 129 a curriculum must be built and tested by close observational and experimental meth ods to assess it fully. there are many opportunities and challenges for finance faculty to make important contributions toward a theory of finance instruction. vi. summary application of educational and learning theory supports the position that personal financial management should be positioned as the cornerstone, first course taught in the finance cur riculum. students learn complex concepts best when structured in ways that have meaning and relevance to them. connected knowing is fostered by centering new learning on stu dents’ past experiences. of all the finance courses, personal financial management is most relevant to undergraduate college students; it provides the best scaffold for linking new finance learning to their current and near future lives. the implications of this paper’s position and call for change in business school curric ulums would be varied. some programs already offer an upper-level personal finance course. for these programs changing the course sequence to teach personal finance first and managerial finance second would be relatively simple, although it may be difficult to add a required course to the major. other business school programs offer personal finance only as a service course to nonmajors. for these schools, the transition would be more dif ficult since the course would have to be redesigned with more rigor and reallocation of fac ulty course assignments would be required. finally, there are programs where no personal finance course is taught what-so-ever. the restructuring of the curriculum for these schools would most likely face faculty resistance and would be extremely difficult since they have not previously taught or valued personal financial education. it is encouraging to note, however, that the aacsb’s new standards make diversity among business school pro grams a possibility. how best to introduce students to and engage students in finance is a matter of bold trials and dogged assessment. a theory of finance instruction needs to be developed to specify the requisite topics and experiences, discipline structure, topic sequence, and reward system that best achieves student growth. how we teach is as important as what we teach, and what we teach needs to be presented in ways consistent with students’ views. good teaching of personal finance as the first course in the undergraduate finance curriculum would provide a broad funda mental structure for the discipline to which all subsequent learning can be easily connected. it will also promote transfer of knowledge in later undergraduate finance and business course work. references aacsb. (1996, may 17). business school enrollments: summary report. saint louis, mo: aacsb. belenky, mary field, clinchy, blythe mcvicker, goldberger, nancy rule, & tarule, jill mattuck. (1986). women’s ways ofknowing. new york: basic books inc. berry, t.d., & farragher, e.j. (1987). survey of introductory financial management courses. jour nal of financial education, fall, 65-72. 130 financial services review 5(2) 1996 bialaszewski, d., pencek, t., & zietlow, j. (1993). finance requirements and computer utilization at aacsb accredited schools. financial practice and education, fall, 133-139. brigham, eugene f. (1973). future direction of the basic finance course. journal of financial edu cation, fall, 13-19. brigham, eugene f., & joel f. houston. (1996). fundamentals offinancial management: the con cise edition. fort worth: the dryden press. bruner, jerome. (1986). actual minds, possible worlds. cambridge: harvard university press. bruner, jerome. (1973). beyond the information given: studies in the psychology of knowing. edited by jeremy m. anglin. new york: w.w. norton & company. bruner, jerome. (1977). the process of education. cambridge: harvard university press. bruner, jerome. (1966). toward a theory of instruction. cambridge: harvard university press. burger, david l. (1992). the importance of living in the case. paper presented at the annual meet ing of the american association of colleges for teacher education, san antonio, tx, febru ary 25-28. available from the educational resources information center. chan, kam c., weber, marsha, & johnson, michelle. (1995). using ‘other peoples money’ in the classroom. financial practice and education, 5(1 spring/summer): 123-127. daigler, robert t. (1979). ‘personal financial planning’ as a new major within finance. journal of financial education, fall, 23-28. d’ambrosio, charles a. (1980). financial theory in personal finance courses. journal of financial education, fall, 17-18. demong, r.f., pettit, l.c., & campsey, b.j. (1979). finance curriculum for the future: perceptions of practitioners versus academicians. journal of financial education, fall, 45-48. dewey, john. (1974). john dewey on education: selected writings. chicago: university of chicago press. dyl, edward a. (1991). wall street: a case in ethics, financial practice and education, i(1 spring), 49-5 1. gitman, lawrence j., & bacon, peter w. (1985). comprehensive personal financial planning: an emerging opportunity. journal of financial education,’ fall, 36-46. gitman, lawrence j., & joehnk, michael d. (1996). personal financial planning, 7th ed. fort worth: the dryden press. graham, lise, & kocher, claudia. (1995). a note on using ‘other people’s money’ in the class room. financial practice and education, 5( 1 spring/summer), 128-129. hallman, g. victor, & rosenbloom, jerry s. (1987). personal financial planning, 4th ed. new york: mcgraw hill book co. helgesen, m., & gentry, j. (1988). cognitive processing preferences and their relationship to finance as a discipline, curriculum and profession. paper presented at the financial manage ment association annual meeting, new orleans, la, october. helgesen, m., & gentry, j. (1995). the relationship between learning styles of college students and sex, race and academic major. paper presented at the financial management association annual meeting, new york, october. higgins, david p. (1983). integrating a personal financial planning major into existing curricula. journal of financial education, fall, 3 1-36. horvath, p.a. (1985). a pedagogical note on inter-period compounding and discounting. financial review, february, 116118. kalogeras, gus. (1976). in defense of the lecture method of instruction. journal of financial educa tion, fall, 45-47. kochman, l.m. (1986). intrayear discounting: uses and misuses. journal of financial education, fall, 12-16. krishnan, v. sivarama, bathala, chenchu t., bhattacharya, t.k., & ritchey, robert. (1996). stu dent perceptions and expectations: a new approach to designing the introductory finance personal finance 131 course. unpublished and preliminary paper presented at the midwest finance association annual conference, chicago, march. lange, david. (1993). applications-oriented spreadsheet instruction: a solution to spreadsheet unfa miliarity, financial practice and education, 3( 1 spring/summer), 73-76. lavine, alan. (1987). brick walls and ivory towers. financiae planning, may, 150-152. lawrence, edward c. (1994). financial innovation: the case of student investment funds at united states universities. financial practice and education, 4( 1 spring/summer), 47-54. mcwilliams, v.b., & pantalone, cc. (1994). structuring the finance curriculum: a survey. finan cial practice and education, spring/summer, 37-46. merton, robert c., & bodie, zvi. (1995). a new approach to the introductory finance course. tutorial presented at the financial management association annual meeting, new york, october. nofsinger, john r. (1995). ‘barbarians at the gate’: a capstone corporate finance project. financial practice and education, 5( 1 spring/summer), 11% 122. paulsen, michael b., & gentry, james a. (1995). motivation, learning strategies, and academic per formance: a study of the college finance classroom. financial practice and education, spring/summer, 78-89. pratt, r.w., & gentry, j.a. (1994). attentional and interpersonal characteristics of finance profes sors: comparisons with college students and business executives. financial practice and education, faluwinter, 66-76. porter, l.w., & mckibbin, l.e. (1988). management education and development: drif or thrust into the 21st century. new york: mcgraw-hill. sadker, myra, & sadker, david. (1994). failing at fairness: how america’s schools cheat girls. new york: charles scribner’s sons. sandler, bemice resnick, & hoffman, ellen. (1992). teaching faculty members to be better teach ers: a guide to equitable and effective classroom techniques. washington, dc: association of american colleges. shackelford, jean. (1992). feminist pedagogy: a means for bringing critical thinking and creativity to the economics classroom. aea papers and proceedings 82(2 may): 570-576. ulivi, richard m. (1982). teaching financial planning to business students. journal of financial education, fall, 5-8. u.s. department of education. (1995). digest of education statistics, p. 256. u.s. department of education. (1991). digest of education statistics, p. 249. yoon, yeomin. (1995). applications-oriented spreadsheet instruction in finance model building: an extension. financiae practice and education, 5( 1 spring/summer), 143-147. pii: s1057-0810(99)00018-9 authors alexander, gordon j., “mutual fund shareholders: characteristics, investor knowledge, and sources of information,” 7(4): 301–316 beal, diana, “‘putting your money where your mouth is’ a profile of ethical investors,” 7(2): 129–143 bieker, richard f., review of “the wall street journal interactive edition internet site,” 7(3): 218– 221 bigel, kenneth s., “the correlations of professionalization and compensation sources with the ethical development of personal investment planners,” 7(4): 223–236 boldin, robert j., “credit union structure: an examination of potential risks,” 7(3): 207–215 brooks, robert, “managing college tuition inflation using surplus framework methodology,” 7(4): 257–271 chen, haiyang, “an analysis of personal financial literacy among college students,” 7(2): 107–128 compton, william s., “a tax-free exploitation of the turn-of-the-month effect: c.r.e.f.,” 7(1): 11–23 crutchley, claire e., “shareholder wealth effects of calpers’ activism,” 7(1): 1–10 detzel, f. larry, “explaining persistence in mutual fund performance,” 7(1): 45–55 domian, dale l., “term spreads and predictions of bond and stock excess returns,” 7(1): 25–44 domian, dale l., “the rise and fall of the ‘dogs of the dow,’” 7(3): 145–159 gibler, karen m., “planning to move to retirement housing,” 7(4): 291–300 goss, robert p., “comment on kenneth s. bigel’s paper,” 7(4): 237–256 goyen, michelle, “‘putting your money where your mouth is’ a profile of ethical investors,” 7(2): 129–143 hanna, sherman, “mean and pessimistic projections of retirement adequacy,” 7(3): 175–193 hudson, carl d., “shareholder wealth effects of calpers’ activism,” 7(1): 1–10 jensen, marlin r.h., “shareholder wealth effects of calpers’ activism,” 7(1): 1–10 jones, jonathan d., “mutual fund shareholders: characteristics, investor knowledge, and sources of information,” 7(4): 301–316 kahl, douglas r., review of “money logic: financial strategies for the smart investor,” 7(3): 217 kahl, douglas r., review of “personal financial planning,” 7(1): 69–70 kunkel, robert a., “a tax-free exploitation of the turn-of-the-month effect: c.r.e.f.,” 7(1): 11–23 lee, euehun, “planning to move to retirement housing,” 7(4): 291–300 leggett, keith, “credit union structure: an examination of potential risks,” 7(3): 207–215 louton, david a., “the rise and fall of the ‘dogs of the dow,’” 7(3): 145–159 mcgoun, elton g., “the sociology of personal finance,” 7(3): 161–173 323 financial services review the journal of individual financial management index volume 7, 1998 mochis, george p., “planning to move to retirement housing,” 7(4): 291–300 montalto, catherine philips, “mean and pessimistic projections of retirement adequacy,” 7(3): 175– 193 mossman, charles e., “the rise and fall of the ‘dogs of the dow,’” 7(3): 145–159 nigro, peter j., “mutual fund shareholders: characteristics, investor knowledge, and sources of information,” 7(4): 301–316 porter, gary e., “performance persistence of experienced mutual fund managers,” 7(1): 57–68 reichenstein, william, “calculating a family’s asset mix,” 7(3): 195–206 reichenstein, william, “term spreads and predictions of bond and stock excess returns,” 7(1): 25–44 reichert, carolyn, “closed end investment companies: historic returns and investment strategies,” 7(2): 83–93 ridge, jenny, “innovations in savings schemes: the bonus bonds trust in new zealand,” 7(2): 73–81 robinson, chris, “the sociology of personal finance,” 7(3): 161–173 russell, judson w., “managing college tuition inflation using surplus framework methodology,” 7(4): 257–271 russell, judson w., “the international diversification fallacy of exchange-listed securities,” 7(2): 95–106 saporoschenko, andy, “do dividend reinvestment plans contribute to industrial firm value and efficiency?” 7(4): 273–289 spencer, austin h., review of “histories of the eighties, lessons for the future, vol. 1, american finance for the twenty-first century, and asset securitization—liquidity and funds management, controller’s handbook,” 7(4): 317–319 strand, robert, “credit union structure: an examination of potential risks,” 7(3): 207–215 timmons, j. douglas, “closed end investment companies: historic returns and investment strategies,” 7(2): 83–93 trifts, jack w., “performance persistence of experienced mutual fund managers,” 7(1): 57–68 volpe, ronald p., “an analysis of personal financial literacy among college students,” 7(2): 107–128 weigand, robert a., “explaining persistence in mutual fund performance,” 7(1): 45–55 young, martin, “innovations in savings schemes: the bonus bonds trust in new zealand,” 7(2): 73–81 yuh, yoonkyung, “mean and pessimistic projections of retirement adequacy,” 7(3): 175–193 titles “a tax-free exploitation of the turn-of-the-month effect: c.r.e.f.,” robert a. kunkel, william s. compton, 7(1): 11–23 “an analysis of personal financial literacy among college students,” haiyang chen, ronald p. volpe, 7(2): 107–128 “calculating a family’s asset mix,” william reichenstein, 7(3): 195–206 “closed end investment companies: historic returns and investment strategies,” carolyn reichert, j. douglas timmons, 7(2): 83–93 “comment on kenneth s. bigel’s paper,” robert p. goss, 7(4): 237–256 “credit union structure: an examination of potential risks,” robert j. boldin, keith leggett, robert strand, 7(3): 207–215 “do dividend reinvestment plans contribute to industrial firm value and efficiency?,” andy saporoschenko, 7(4): 273–289 “explaining persistence in mutual fund performance,” f. larry detzel, robert a. weigand, 7(1): 45–55 “innovations in savings schemes: the bonus bonds trust in new zealand,” jenny ridge, martin young, 7(2): 73–81 324 index / financial services review 7 (1998) 323–325 “managing college tuition inflation using surplus framework methodology,” judson w. russell, robert brooks, 7(4): 257–271 “mean and pessimistic projections of retirement adequacy,” yoonkyung yuh, sherman hanna, catherine philips montalto, 7(3): 175–193 “mutual fund shareholders: characteristics, investor knowledge, and sources of information,” gordon j. alexander, jonathan d. jones, peter j. nigro, 7(4): 301–316 “performance persistence of experienced mutual fund managers,” gary e. porter, jack w. trifts, 7(1): 57–68 “planning to move to retirement housing,” karen m. gibler, george p. mochis, euehun lee, 7(4): 291– 300 “‘putting your money where your mouth is’ a profile of ethical investors,” diana beal, michelle goyen, 7(2): 129–143 review of “histories of the eighties, lessons for the future, vol. 1., american finance for the twentyfirst century, and asset securitization—liquidity and funds management, controller’s handbook,” austin h. spencer, 7(4): 317–319 review of “money logic: financial strategies for the smart investor,” douglas r. kahl, 7(3): 217 review of “personal financial planning,” douglas r. kahl, 7(1): 69–70 review of “the wall street journal interactive edition internet site,” richard f. bieker, 7(3): 218– 221 “shareholder wealth effects of calpers’ activism,” claire e. crutchley, carl d. hudson, marlin r. h. jensen, 7(1): 1–10 “term spreads and predictions of bond and stock excess returns,” dale l. domian, william reichenstein, 7(1): 25–44 “the correlations of professionalization and compensation sources with the ethical development of personal investment planners,” kenneth s. bigel, 7(4): 223–236 “the international diversification fallacy of exchange-listed securities,” judson w. russell, 7(2): 95–106 “the rise and fall of the ‘dogs of the dow’,” dale l. domian, david a. louton, charles e. mossman, 7(3): 145–159 “the sociology of personal finance,” chris robinson, elton g. mcgoun, 7(3): 161–173 325index / financial services review 7 (1998) 323–325 financial services review, 32(4) 1 artificial intelligence in accounting, medicine, and law with potential implications for financial planning: a review of literature manuela e. faulhaber1 and charles chaffin2 abstract generative artificial intelligence (ai) is rapidly reshaping multiple fields. generative ai is a type of ai that can create new content or information from scratch, rather than simply manipulating or organizing existing data. this has the potential to revolutionize the way that financial advisors interact with clients and manage their businesses. however, there are many unknowns as it relates to the level and degree of disruption that generative ai can bring to financial planning. this paper explores the interaction of ai with financial planning, drawing insights from the practices of accounting, medicine, and law. while the primary focus remains on financial planning, this interdisciplinary approach aims to enrich understanding while examining parallels and emerging trends across diverse professional domains as each of the four professions reviewed in this paper integrates client needs, preferences, and goals into their decision-making processes. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation faulhaber, m. e., & chaffin, c. (2024). artificial intelligence in accounting, medicine, and law with potential implications for financial planning: a review of literature. financial services review, 32(4), 1-11. introduction according to the world economic forum's 2023 "future of jobs report," jobs that necessitate "human skills such as judgment, creativity, physical dexterity, and emotional intelligence" are the least likely to be replaced by artificial intelligence (shine, 2023). within financial planning, ai can revolutionize client communication, streamlining backend operations, and redefining the role of advisors. advisors can generate client communication quickly and efficiently, developing tailored emails and other correspondence to clients during market volatility and client life changes. back 1 corresponding author (manufa@iastate.edu). iowa state university, ames, ia, usa 2 iowa state university, ames, ia, usa office work in developing financial plans will likely be impacted as generative ai can be used to develop plans quickly and more efficiently. one of the most promising applications of generative ai in financial planning is in the area of client communication. advisors can use generative ai to generate personalized and tailored emails, letters, and other forms of correspondence to clients. this can save advisors time and effort, while also ensuring that clients receive timely and relevant information. for example, advisors can use generative ai to generate personalized investment recommendations, based on each client's https://creativecommons.org/licenses/by-nc/4.0/ https://paperpile.com/c/vjdbn8/rpo2 mailto:manufa@iastate.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(4) 2 circumstances and risk tolerance. generative ai can also be used to streamline back-office operations in financial planning. for example, advisors can use generative ai to automate the process of creating financial plans. this can free up advisors' time, allowing them to focus on more strategic tasks, such as developing client relationships and providing personalized advice. in addition, generative ai can augment the knowledge and skills of the financial advisor, providing clients with more comprehensive and sophisticated advice. for example, advisors can use generative ai to generate hypothetical scenarios, allowing clients to see how different investment decisions might affect their financial future in much more specific, tangible ways. the use of generative ai in financial planning is still in its early stages, but it has the potential to revolutionize the way that advisors do business. advisors who are early adopters of generative ai will be well-positioned to succeed in the future. here are some specific examples of how generative ai can be used in financial planning (spiegel, 2023): • client communication: o generate personalized emails and letters to clients o create tailored investment recommendations o develop hypothetical scenarios • back-office operations: o automate the process of creating financial plans o generate reports and presentations o manage client data • advisor development: o provide training and education on new investment products and strategies o generate ideas for new marketing campaigns o develop new business plans despite its transformative potential, apprehension looms among financial advisors regarding the extent of ai's disruption to the profession. within larger firms, ai is being used more frequently, whereas, within smaller firms, advisors may only use chatgpt to draft client communication, not having the resources to use ai for the development of comprehensive financial plans. in many cases, ai is most effective at increasing assets via client communication, specifically what to deliver and when to deliver to a client relative market, life event, or other economic commentary. drawing insights from parallel professions like accounting, medicine, and law, this paper delves into the disruptive impact of ai, not only on day-to-day operations but also on crucial aspects such as talent acquisition and client engagement. generative ai is already having a significant impact on financial planning. however, there are still a great deal of unknowns as to the implications on the workforce, competencies, and business models that generative ai will have within the next several years. this is why we explore the impact of generative ai on accounting, medicine, and law as it relates to the business processes, workforce, and compensation structures—our hope is to provide context on how ai can impact financial planning. accounting the accounting profession has not been immune to the growing adoption of artificial intelligence. from data entry to financial analysis, ai has the potential to automate many of the routine and repetitive tasks that accountants currently perform. this has led to a significant increase in the implementation of ai in accounting, with 24% of top-performing client advisory services practices incorporating ai, as per the 2022 cas benchmark survey conducted by cpa.com (cpa.com, 2022, p. 19) while ai excels in repetitive tasks, it lacks the subject matter expertise required for certain specialized tasks. however, the use of ai in accounting allows professionals to transition away from daily repetitive tasks and focus on more complex and strategic responsibilities where ai may not be as effective. this enables accounting professionals to serve more clients, https://paperpile.com/c/vjdbn8/xjf0k faulhaber & chaffin 3 provide tailored services to clients with more complex needs, and increase the revenue of the accounting firm. additionally, implementing ai in accounting can reduce stress levels by allowing accountants to transition from repetitive tasks to more complex and strategic responsibilities. incorporating ai in accounting offers a solution to the pressing concerns of accountant burnout and staffing shortages. by automating routine tasks, ai empowers accountants to shift their focus from mundane duties to more complex, strategic responsibilities. this, in turn, has a positive impact on accountant well-being and overall job satisfaction. here is how ai is revolutionizing the accounting landscape: 1. streamlined workflow and efficiency: ai automates tasks such as data entry, invoice processing, and financial reporting, freeing up accountants' time for more value-added activities. this streamlining of workflows enhances productivity, enables faster turnaround times, and reduces the risk of errors. 2. improved accuracy and consistency: ai algorithms analyze vast amounts of data with precision and consistency, minimizing the potential for human error. this enhances the reliability of financial statements and improves compliance with accounting standards. 3. real-time insights and decision-making: aipowered analytics provide real-time insights into financial performance, enabling accountants to make informed decisions promptly. this agility gives businesses a competitive edge and fosters a proactive approach to financial management. 4. enhanced audit and risk management: ai assists in identifying inconsistencies, anomalies, and potential risks within financial data. this facilitates timely and effective audits, strengthens internal controls, and ensures compliance with regulatory requirements. 5. personalized client service: ai enables accountants to offer tailored and personalized services to clients. by leveraging data-driven insights, accountants can better understand client needs, anticipate challenges, and provide proactive solutions. 6. enhanced job satisfaction and reduced burnout: by eliminating repetitive and mundane tasks, ai empowers accountants to engage in more intellectually stimulating and challenging work. this shift enhances job satisfaction, reduces stress, and fosters a sense of accomplishment, contributing to a more positive work environment. 7. upskilling and continuous learning: the integration of ai demands continuous learning and upskilling. accountants are encouraged to embrace new technologies, develop programming skills, and deepen their understanding of data analytics. this fosters a culture of innovation and adaptability within the accounting profession. 8. future-proofing the accounting profession: by embracing ai, accounting firms position themselves for future success. as technology continues to advance, ai will play an increasingly pivotal role in the accounting landscape. firms that invest in ai today will be well-equipped to meet the demands of tomorrow. the integration of ai in accounting is not merely a technological advancement; it represents a transformative shift that empowers accountants to unlock their full potential. by leveraging the capabilities of ai, accounting firms can enhance efficiency, improve accuracy, gain real-time insights, and deliver exceptional client service. as a result, accountants can focus on strategic and advisory roles, leading to increased job satisfaction, reduced burnout, and a sustainable future for the accounting profession. however, while ai can respond to prompts and input, trained on big data, it cannot make original decisions or develop new ideas (badonia, 2023). human touch is of utmost importance when it comes to nuanced decision-making, empathetic client interactions, or addressing complex scenarios. while ai can enhance efficiency in accounting firms, human expertise remains invaluable. https://paperpile.com/c/vjdbn8/by3x2 financial services review, 32(4) 4 there is a steadily increasing shortage of accountants in the accounting profession. a 2022 deloitte poll found that 82% of hiring managers for accounting at public companies and 69% at private companies mentioned talent retention as a challenge (deloitte, 2022). simultaneously, ai in the accounting market is expected to exceed 39.57 billion usd by 2030, exhibiting a compound annual growth rate (cagr) of 45.31% during the forecast period of 2023-2030 (sns insider, 2023). accountants have to be able to think logically to make well-informed decisions, utilize their creativity to devise solutions for clients and communicate effectively. the u.s. bureau of labor statistics projects a 4% growth in employment for accountants and auditors from 2022 to 2032, in line with the average for all occupations. this growth is influenced by factors such as globalization, economic expansion, and a complex regulatory landscape. approximately 126,500 annual job openings are anticipated, driven by replacements for workers changing occupations or retiring (u.s. bureau of labor statistics, 2023a). while technological advancements, including cloud computing and ai, may automate routine tasks for accountants, they are expected to enhance efficiency rather than reduce overall demand. this automation will likely increase the importance of accountants' advisory and analytical responsibilities. hence, the impact of ai on accounting suggests that rather than leading to job losses, it has the potential to enhance efficiency and productivity in the profession. as the accounting landscape is constantly changing with the inclusion of ai, accountants are compelled to adapt to these evolving technological trends which also mark a significant shift in the required skills for accountants. key areas include data analysis, cyber security, the strategic utilization of ai as well as a pivotal shift from a traditional role to a more advisory role. according to a recent survey conducted by the american institute of cpas, it was found that 44% of surveyed firms are presently providing advisory services and 70% intend to expand their advisory offerings within the next five years (schraeder, 2023). accountants are expected to provide strategic financial guidance for clients, leveraging their analytical skills for complex tasks and industry knowledge to offer valuable insights to their customers. furthermore, the implementation and utilization of ai in accounting has been shown to assist organizations in minimizing their exposure to accounting risks, such as outstanding debts or tax penalties (farkas, 2023). the reliability of ai in handling routine data entry tasks significantly reduces the risk of human error. ai has the ability to process vast amounts of numerical data exponentially faster and with greater accuracy than accountants. its computational speed surpasses human capabilities. it can analyze and interpret quantitative information in a fraction of the time it would take a human accountant. furthermore, the accuracy of ai in handling quantitative tasks is unparalleled. the algorithms ai uses are designed to execute calculations precisely and do not leave room for human error (oecd, 2021). this precision ensures reliable results and can improve integrity for accounting firms. additionally, another strength of ai is managing and analyzing large quantities of data. it can process financial transactions, conduct market trend analyses, or perform complex calculations. these tasks would be impractical or timeconsuming for a human accountant to manage manually. ai has shown itself to be an efficient tool for tasks that require precision and mathematical rigor. while ai excels in quantitative tasks, as previously discussed, ai falls short in the realm of qualitative skills, compared to human accountants. human accountants have the ability to understand the broader context surrounding financial data and decisions. they can consider industry dynamics, market trends, and the unique circumstances of each client and they can offer strategic advice (gaetano, 2023). accountants can leverage their experience and industry knowledge to interpret complex situations and ask clients the right questions at the right time based on their specific needs and scenarios. this may also help them to foresee potential challenges, identify opportunities, and make https://paperpile.com/c/vjdbn8/jgcsk https://paperpile.com/c/vjdbn8/gwin9 https://paperpile.com/c/vjdbn8/oxmt2 https://paperpile.com/c/vjdbn8/ppsn4 https://paperpile.com/c/vjdbn8/pzmrl https://paperpile.com/c/vjdbn8/j0b5h https://paperpile.com/c/vjdbn8/bvpq2 faulhaber & chaffin 5 decisions beyond quantitative data alone. it can be concluded that the qualitative value provided by human accountants is unmatched. like financial planning, human impact in the field of accounting is irreplaceable. a human touch is needed to form trusting business relationships (marciano, 2023). this also includes accounting professionals who can emphasize with their clients about, for instance business anxiety, understand their business vision, and guide them in decision-making. the relational aspect of relationships accountants have with their clients cannot be replicated by ai. human accountants have a profound understanding of financial principles and the financial industry—attributes that are difficult for ai to replicate. furthermore, humans play an essential role in maintaining ethical integrity of financial practices, ensuring adherence to regulations, and making decisions with a personal touch (bontrager, 2023). understanding ethics is crucial to maintaining industry standards and fostering trust with clients. in the realm of accounting, human involvement remains irreplaceable, just like in financial planning. developing strong business relationships necessitates the human touch. accountants, for example, can relate to their clients' concerns, such as business anxiety, grasp their business objectives, and guide them in making wise decisions. artificial intelligence (ai) cannot duplicate the relational aspect of accountants' relationships with their clients. human accountants have a deep understanding of financial principles and the industry, which are qualities that ai finds difficult to duplicate. additionally, humans play a vital role in upholding the moral fiber of financial practices, adhering to rules, and making decisions with a personal touch. maintaining industry standards and gaining clients' trust rely heavily on ethical comprehension. accountants' ability to grasp the emotional and psychological elements of clients' financial decisions gives them a significant edge over ai. humans can empathize with clients who are experiencing financial anxiety or stress, providing emotional support and a personalized touch that ai lacks. this empathetic approach helps build stronger client relationships and fosters trust, leading to improved communication and decision-making. furthermore, human accountants' deep understanding of the specific nuances and complexities of their clients' industries allows them to provide more tailored and effective advice. by leveraging their knowledge and experience, accountants can offer customized solutions that address the unique challenges and opportunities faced by their clients. this level of personalization is difficult for ai to achieve, as it cannot comprehensively grasp the intricacies of various industries and their impact on financial decisions. as ai continues to advance, certain aspects of accounting work may likely become automated. however, the human element will always be essential in providing high-quality accounting services that meet the complex needs of clients. by combining the strengths of ai and human expertise, accounting professionals can deliver a comprehensive and personalized approach that maximizes value for their clients. in addition to technical skills, accountants are also being encouraged to develop a range of soft skills that are essential for success in the modern workplace. these skills include being a team player, having strong interpersonal skills, valuing professional and personal integrity, being creative, and being empathic. these traits are difficult for ai to replicate, and they are therefore essential for accountants who want to distinguish themselves in the job market. accountants who possess these skills will be well-positioned to succeed in the future of accounting. they will be able to work effectively with ai technology, and they will be able to provide valuable insights to their clients and organizations. here are some specific examples of how accountants can use their analytical and strategic thinking skills to complement ai technology: • accountants can use ai to automate repetitive tasks, such as data entry and analysis. this can free up accountants' time so that they can focus on more complex and value-added activities, https://paperpile.com/c/vjdbn8/q4ung https://paperpile.com/c/vjdbn8/obmcy financial services review, 32(4) 6 such as providing insights to clients and organizations. • accountants can use ai to identify trends and patterns in data. this information can be used to make better decisions and to develop more effective strategies. • accountants can use ai to create and evaluate financial models. this can help accountants to better understand the financial implications of different decisions. accountants who possess these skills will be in high demand in the years to come. they will be well-positioned to succeed in the future of accounting, and they will be able to make a significant contribution to their organizations. medicine rapid advancements in ai have uncovered new possibilities for the healthcare sector. one application is in the realm of telehealth and telemedicine. ai can be seamlessly integrated into clinical practice to enhance and improve patient care and diagnostic accuracy. according to alowais et al. (2023), ai-driven virtual healthcare can successfully simulate conversations, diagnose diseases, formulate personalized treatment plans, and support medical professionals in decision-making. these systems can provide personalized care to individual patients. with this evolution, the primary emphasis shifts beyond task automation, directing attention to technologies that can improve patient care in a healthcare setting. studies conducted by fitzpatrick et al. (2017) and williams and andrews (2013) emphasize the effectiveness and accessibility of ai in providing mental health support to patients who suffer from symptoms of anxiety and depression. internetbased cognitive behavioral therapy (cbt) has emerged as a promising psychotherapeutic intervention that can be facilitated with the help of ai. fitzpatrick et al.'s study focused on the feasibility and acceptability of a fully automated conversational ai agent that engaged with participants by delivering a self-help program. results indicated that participants in the ai agent group significantly reduced their symptoms of depression over the study period compared to a control group. this study serves as one example that demonstrates the potential of conversational ai to deliver therapy effectively. similarly, williams and andrews explored the effectiveness of internet-delivered cognitive-behavioral therapy (icbt) and found significant reductions in depressive symptoms. these studies underscore the potential of ai-driven mental health support. ai has had a remarkable impact on diagnostic accuracy in healthcare. ai can analyze extensive sets of medical data, identify intricate patterns, and provide predictions. this can facilitate early and precise diagnoses and revolutionize the landscape of healthcare. the significance of automating the interpretation of chest radiographs is underscored by a study that aimed to evaluate the performance of an ai tool (plesner et al., 2023). the tool's assessment was based on (a) the number of autonomously reported chest radiographs, (b) its sensitivity in detecting abnormalities, and (c) its performance compared to clinical radiology reports. the ai tool, in comparison to clinical radiology reports, exhibited a sensitivity of 99.1% for detecting abnormal radiographs and an even higher sensitivity of 99.8% for identifying critical abnormalities. this performance surpassed the sensitivity of radiologist reports, which stood at 72.3% and 93.5%. this study highlights the potential of ai to achieve high accuracy in identifying abnormalities in chest radiographs with further implications in detecting and treating a number of patient needs. furthermore, ai has the ability to identify subtle nuances and detect these abnormalities early in the diagnostic process which can contribute to a proactive approach to healthcare and therefore improve patient outcomes. plesner et al.’s (2023) findings underscore the potential of ai to significantly reduce the risk of misdiagnoses. it has emerged as a promising solution for an already existing global shortage of trained radiologists reported by the radiological society of north america (henderson, 2022). by outperforming human experts in accuracy and sensitivity, ai introduces a level of reliability and constituency that is very important in the field of medical diagnostics. the integration of ai in diagnostic assistance represents a shift in https://paperpile.com/c/vjdbn8/p8dlv/?noauthor=1 https://paperpile.com/c/vjdbn8/g8sv3 https://paperpile.com/c/vjdbn8/g8sv3 https://paperpile.com/c/vjdbn8/dm4fg faulhaber & chaffin 7 healthcare, ultimately having the potential to reduce errors and improve patient outcomes. as the demand for quality healthcare increases, healthcare systems are dealing with time constraints and excessive workloads. these can compromise the quality of patient care. ai can be utilized by trained medical health professionals for multiple tasks such as diagnostics, data analysis, health insurance tasks, or treatment planning. ai contributes to more efficient healthcare delivery, reduces costs, and facilitates better patient experiences. the integration of ai into medical workflows holds the potential to significantly enhance efficiency and accuracy (krishnan et al., 2023). ai assists healthcare professionals in managing their workload efficiently and can provide them with valuable insights. as a result, the collaboration between ai and humans can address the challenges of healthcare and improve overall healthcare delivery. however, while ai has proven itself to be very helpful and supportive, regulatory bodies such as the food and drug administration (fda) maintain a crucial requirement. the fda mandates that a human must serve as the "ultimate arbiter" of what the machine-learning algorithm finds (park, 2022). this regulatory oversight ensures that the ethical and decisionmaking aspects of healthcare remain within the control of trained medical professionals. like financial planning and accounting, the role of a human touch remains indispensable in healthcare. medical healthcare professionals possess a unique, holistic understanding of patients, taking into account not only their symptoms but also the emotional and social aspects of their well-being. this human-centric approach, focused on human touch is crucial for providing patient-centered healthcare (drouin & freeman, 2020). furthermore, participants in a study by longoni et al. (2019) exhibited strong reluctance across various procedures, ranging from skin cancer screenings to pacemaker implant surgeries, when ai was proposed as the service provider. participants expressed a preference for human care providers, even if it meant a greater risk of inaccurate diagnosis or surgical complications. it was identified that the resistance to ai was rooted in the belief that ai cannot account for individual characteristics and circumstances. people perceive themselves as unique and extend this belief to their health needs. the study found that participants were less likely to utilize ai services and preferred human providers. additionally, they reported wanting to pay less for ai-based healthcare, implying a lesser value. proper physician contact can model calmness and a sense of protection for patients (spivack, 2023). spivack argued that, despite technological advancements, no substitute exists for the essential human-to-human connection in healthcare. in conclusion, the integration of ai into healthcare has to consider the inherent value of human touch. while ai offers remarkable capabilities, patients' beliefs, reluctance, and a strong need for human connection remain. law recent advancements in technology— specifically ai—are playing an important role in shaping the future trajectory of the legal profession. as mentioned by segarra (2023), the integration of ai into legal practices can change the nature of lawyers' day-to-day work. attorneys can utilize ai in various capacities. this can range from document creation and processing to database classification. additionally, ai can help streamline other tasks, such as due diligence, document review, contract management and review, and data analysis as well as enhancing efficiency and productivity (altfee, 2023). however, despite these efficiency gains, the human element remains indispensable in legal practice. while ai can automate mundane tasks for lawyers, the complexity of legal strategy, decision-making, and ethical considerations still need human involvement (segarra, 2023). furthermore, while the usage of ai presents significant advantages, it also carries risks. relying excessively on ai for tasks that require human judgment and oversight can lead to serious repercussions. for instance, in the case of mata versus avianca, a lawyer utilized chatgpt, an openai program, to draft a legal brief. however, the program "hallucinated" and produced a fictitious case law (mcgregor & atherton, 2023). the lawyer who used ai expressed deep regret for relying on chatgpt and vowed never to use it again without absolute https://paperpile.com/c/vjdbn8/f2wev https://paperpile.com/c/vjdbn8/nldur https://paperpile.com/c/vjdbn8/hsbqs https://paperpile.com/c/vjdbn8/hsbqs https://paperpile.com/c/vjdbn8/0xpsf/?noauthor=1 https://paperpile.com/c/vjdbn8/wfrah https://paperpile.com/c/vjdbn8/jj1rc/?noauthor=1 https://paperpile.com/c/vjdbn8/nv0vr https://paperpile.com/c/vjdbn8/pzb5m https://paperpile.com/c/vjdbn8/pzb5m financial services review, 32(4) 8 verification of its authenticity. the judge described the legal submission as filled with fake judicial decisions, quotes, and citations, and called it an "unprecedented situation.” the case was dismissed and the judge imposed a $5,000 fine on the lawyer and their firm. additionally, they were mandated to send letters of apology to six actual judges who were wrongly attributed as the authors of the fabricated opinions cited in their legal documents (weiser, 2023). this example emphasizes the importance for legal professionals to understand and work within the limitations of ai tools. ai may lack the capacity for logical reasoning and to identify factual inconsistencies in responses when attempting to fulfill a request. generative ai presents opportunities for legal professionals to expand their firms by enabling them to take on more clients and handle additional work, due to the ability to increase productivity. furthermore, ai holds promise for bridging the justice gap and expanding access to legal counsel, particularly for individuals from low-economic communities. many individuals still view hiring a lawyer as financially prohibitive, with 80% of low-income individuals unable to afford legal representation (beckman, 2023). "donotpay" is hailed as the "world's first robot lawyer", and shows the potential of ai in addressing legal challenges. its ceo joshua browder reports over 2 million successfully resolved cases through ai. the company mainly focuses on supporting individuals with legal conflicts related to medical bills(donotpay, 2023). currently, donotpay maintains hundreds of thousands of active subscribers. the impact of ai in the legal industry can also pave access to solving justice issues. the impact of ai on the legal field is undeniable. while senior legal professionals often proceed with their duties relatively unaffected, junior lawyers and legal support staff must adapt to the evolving circumstances. they must embrace ai, understand its constraints, and uphold ethical norms (segarra, 2023). according to a recent survey conducted by the thomson reuters institute, 82% of legal professionals acknowledged the potential ai capabilities in legal work. however, only 51%, believe that ai should be implemented in legal work. this difference suggests that while legal professionals understand the potential benefits of ai, they show hesitation regarding the reliability of ai tools and work accuracy (warren, 2023). the employment outlook for lawyers appears promising with a projected growth rate of 8% from 2022 to 2032 (u.s. bureau of labor statistics, 2023b). this exceeds the average growth rate for all occupations reported by the labor department. furthermore, ai presents opportunities for legal professionals to expand their firms by enabling them to take on more clients and handle additional work, due to the ability to increase productivity. this can increase the time spent on mundane tasks which also may lead to a decrease in billable hours per client. this shift could prompt some firms to have to reevaluate their billing model. there may be the chance to move towards a more "value-based" billing process, where it is ensured that proper compensation for legal services is provided while also providing clients with additional transparency (lexisnexis, 2023). while the billable hour method may still be utilized, clients may receive greater value from an attorney's time as a result of this shift. moreover, attorneys may be able to offer flat-rate pricing, when taking on more clients or handling additional work, through the help of ai. this shift in billing practice is most likely appreciated by clients. additionally, lawyers may have to spend less time on mundane tasks and are therefore able to bill for activities that provide greater value for the clients, such as high-value activities that require legal expertise or critical thinking. overall, the efficiency gains brought by ai underscore the potential unsustainability of the billable hour model. law firms cannot ignore these technological advancements as lawyers can become more efficient with the help of ai (altfee, 2023). according to the generative ai & the legal profession: 2023 survey report, lawyers report recognizing the potential benefits of ai in the legal space, however, the continued importance of human attorneys to prevent overreliance on technology is also emphasized. legal ai is viewed more as a tool to complement rather than replace human lawyers (generative ai & the legal profession 2023 survey report, 2023). in addition, the ability to use ai to develop wills, https://paperpile.com/c/vjdbn8/jxulv https://paperpile.com/c/vjdbn8/xowfb https://paperpile.com/c/vjdbn8/xowfb https://paperpile.com/c/vjdbn8/np8sg https://paperpile.com/c/vjdbn8/np8sg https://paperpile.com/c/vjdbn8/jj1rc https://paperpile.com/c/vjdbn8/vh0zu https://paperpile.com/c/vjdbn8/uqf87 https://paperpile.com/c/vjdbn8/uqf87 https://paperpile.com/c/vjdbn8/ens8y https://paperpile.com/c/vjdbn8/nv0vr https://paperpile.com/c/vjdbn8/4b7bb https://paperpile.com/c/vjdbn8/4b7bb faulhaber & chaffin 9 trusts, and other legal documents can expand the market to consumers who otherwise would not be served, with consumers with more complex situations seeking additional legal advice. financial planning helps guide individuals and businesses in achieving their financial goals. the integration of ai has started to reshape financial planning and offers opportunities for continuous innovation. this last section explores the intersection of financial planning and ai, highlighting the significance of ai across diverse industries. financial planning financial planning, coupled with the integration of artificial intelligence (ai), presents a transformative landscape of opportunities for optimizing client outcomes and achieving operational efficiencies. financial firms have been leveraging ai for various tasks, including fraud detection and credit scoring, for some time now. however, the advent of generative ai has opened new doors for financial advisors to seamlessly incorporate this technology into their daily workflows. tasks such as conducting comprehensive research, analyzing stock market trends, and generating insightful reports can be significantly enhanced with the assistance of ai. furthermore, chatbots have emerged as valuable tools for financial advisors, simplifying tasks like drafting personalized emails to clients. however, it is crucial to recognize that the information provided by ai may not always be entirely accurate or comprehensive. without incorporating specific client details, the advice received may lack the necessary customization to address individual needs. therefore, financial advisors must engage in asking follow-up questions and be mindful of the limitations of ai in providing financial advice. the integration of ai offers financial advisors the ability to analyze large data sets rapidly and reliably. ai-driven financial tools can provide automated budgeting and expense tracking, personalized debt management strategies, enhanced financial literacy resources, and practical tips for smart investing. by utilizing historical data, machine learning algorithms, and statistical models, ai can identify trends and patterns to suggest portfolio optimization and ensure compliance by flagging potential issues. while ai presents numerous benefits, it is essential to acknowledge the potential legal, ethical, and regulatory considerations associated with its implementation. financial advisors must navigate these considerations responsibly and ethically to maintain the trust and confidence of their clients. a recent survey conducted by f2 strategy revealed that over half (51%) of wealth management firms are actively engaged in ai projects, while 49% have yet to initiate such projects (f2 strategy, 2023). among the ongoing projects, optical character recognition, predictive analysis, and a combination of both are the most commonly utilized ai applications. notably, 62% of surveyed wealth management firms rated their ai knowledge as 5 out of 10 or lower. this finding highlights the need and opportunity for enhancing understanding and awareness in this domain. firms can benefit immensely from additional education about ai's capabilities, value, practical applications, and limitations. the u.s. bureau of labor statistics projects a rapid growth rate of 13% for the employment of personal financial advisors from 2022 to 2032. this surpasses the average growth rate for all occupations. the integration of ai is reshaping traditional practices and redefining the role of financial advisors in the years to come. financial planners can extract valuable insights and best practices from related sectors such as accounting, medicine, and law to effectively navigate the influence of ai and provide holistic financial guidance to their clients. legal, ethical, and regulatory considerations this section aims to provide guidance on navigating legal, ethical, and regulatory considerations in ai-driven financial planning. financial planners must familiarize themselves with the legal and regulatory framework governing ai applications in financial services. this could include laws such as the gdpr or ccpa. like other aspects of financial planning practice, with an awareness of regulatory developments and compliance requirements, financial services review, 32(4) 10 financial planners can mitigate potential legal risks and ensure adherence to industry standards. additionally, central to ai-driven financial planning is the responsible handling of client data. financial planners must implement reliable data privacy and security measures to protect sensitive information from unauthorized access. this may include encryption protocols, access, and regular audits to ensure compliance with data regulations. additionally, transparency in data collection and usage can foster trust and demonstrate commitment to client confidentiality. financial planners should openly discuss the role of ai in their practice, including its benefits, limitations, and potential risks. transparency about data usage, algorithmic decision-making, and risk management processes can instill confidence in the process. by proactively addressing these potential issues and implementing best practices, financial planners can uphold client trust, mitigate risks, and utilize the potential of ai technologies in financial planning. conclusion in conclusion, this paper underscores the importance of leveraging insights from accounting, medicine, and law to enhance the understanding of ai in financial planning. by embracing ai and continuously enhancing their knowledge and skills, financial advisors can stay at the forefront of innovation, deliver exceptional client service, and contribute to the overall financial well-being of individuals and families. references alowais, s. a., alghamdi, s. s., alsuhebany, n., alqahtani, t., alshaya, a. i., almohareb, s. n., aldairem, a., alrashed, m., bin saleh, k., badreldin, h. a., al yami, m. s., al harbi, s., & albekairy, a. m. (2023). revolutionizing healthcare: the role of artificial intelligence in clinical practice. bmc medical education, 23(1), 689. altfee. (2023, october 24). how ai for lawyers is changing the legal industry. https://www.altfeeco.com/resources/ai-forlawyers badonia, s. (2023, january 21). chatgpt vs. human writers: the abilities ai can’t replicate. the startup. https://medium.com/swlh/chatgpt-vshuman-writers-the-abilities-ai-cantreplicate-8ec346844ebb beckman, m. (2023, april 10). artificial intelligence presents opportunities and challenges for the legal ecosystem. https://nysba.org/artificial-intelligencepresents-oppor-tunities-and-challenges-forthe-legal-ecosystem/ bontrager, m. (2023, june 6). how ai is transforming the accounting industry — and what the future will look like. entrepreneur. https://www.entrepreneur.com/sciencetechnology/how-ai-is-transforming-theaccounting-industry/452837 cpa.com (2022, december 1). cas benchmark survey. https://www.cpa.com/sites/cpa/files/202212/cas-benchmark-survey-22-cpacom.pdf deloitte. (2022, september 21). most public companies are challenged in attracting finance and accounting talent, deloitte finds. deloitte united states. https://www2.deloitte.com/us/en/pages/abo ut-deloitte/articles/press-releases/mostcompanies-challenged-to-attract-financeand-accounting-talent.html donotpay. (2023). https://donotpay.com/about/ drouin, o., & freeman, s. (2020, january 22). health care needs ai. it also needs the human touch. https://www.statnews.com/2020/01/22/heal th-care-needs-ai-it-also-needs-humantouch/ farkas, c. (2023, june). ai in accounting and tax compliance: the possibilities and pitfalls. https://accounting.nridigital.com/iab_jun23 /ai_in_accounting_and_tax_compliance_th e_possibilities_and_pitfalls f2 strategy. (2023, october 26). wealth management firms begin to implement ai with an eye on security and regulation. https://www.f2strategy.com/insight/wealthmanagement-firms-begin-to-implement-aiwith-an-eye-on-security-and-regulation http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/p8dlv http://paperpile.com/b/vjdbn8/nv0vr http://paperpile.com/b/vjdbn8/nv0vr http://paperpile.com/b/vjdbn8/nv0vr http://paperpile.com/b/vjdbn8/nv0vr http://paperpile.com/b/vjdbn8/nv0vr https://www.altfeeco.com/resources/ai-for-lawyers https://www.altfeeco.com/resources/ai-for-lawyers http://paperpile.com/b/vjdbn8/by3x2 http://paperpile.com/b/vjdbn8/by3x2 http://paperpile.com/b/vjdbn8/by3x2 http://paperpile.com/b/vjdbn8/by3x2 http://paperpile.com/b/vjdbn8/by3x2 http://paperpile.com/b/vjdbn8/by3x2 https://medium.com/swlh/chatgpt-vs-human-writers-the-abilities-ai-cant-replicate-8ec346844ebb https://medium.com/swlh/chatgpt-vs-human-writers-the-abilities-ai-cant-replicate-8ec346844ebb https://medium.com/swlh/chatgpt-vs-human-writers-the-abilities-ai-cant-replicate-8ec346844ebb http://paperpile.com/b/vjdbn8/xowfb http://paperpile.com/b/vjdbn8/xowfb http://paperpile.com/b/vjdbn8/xowfb http://paperpile.com/b/vjdbn8/xowfb http://paperpile.com/b/vjdbn8/xowfb http://paperpile.com/b/vjdbn8/xowfb https://nysba.org/artificial-intelligence-presents-oppor-tunities-and-challenges-for-the-legal-ecosystem/ https://nysba.org/artificial-intelligence-presents-oppor-tunities-and-challenges-for-the-legal-ecosystem/ https://nysba.org/artificial-intelligence-presents-oppor-tunities-and-challenges-for-the-legal-ecosystem/ http://paperpile.com/b/vjdbn8/obmcy http://paperpile.com/b/vjdbn8/obmcy http://paperpile.com/b/vjdbn8/obmcy http://paperpile.com/b/vjdbn8/obmcy http://paperpile.com/b/vjdbn8/obmcy http://paperpile.com/b/vjdbn8/obmcy https://www.entrepreneur.com/science-technology/how-ai-is-transforming-the-accounting-industry/452837 https://www.entrepreneur.com/science-technology/how-ai-is-transforming-the-accounting-industry/452837 https://www.entrepreneur.com/science-technology/how-ai-is-transforming-the-accounting-industry/452837 http://paperpile.com/b/vjdbn8/jgcsk http://paperpile.com/b/vjdbn8/jgcsk http://paperpile.com/b/vjdbn8/jgcsk http://paperpile.com/b/vjdbn8/jgcsk http://paperpile.com/b/vjdbn8/jgcsk http://paperpile.com/b/vjdbn8/jgcsk http://paperpile.com/b/vjdbn8/jgcsk https://www2.deloitte.com/us/en/pages/about-deloitte/articles/press-releases/most-companies-challenged-to-attract-finance-and-accounting-talent.html https://www2.deloitte.com/us/en/pages/about-deloitte/articles/press-releases/most-companies-challenged-to-attract-finance-and-accounting-talent.html https://www2.deloitte.com/us/en/pages/about-deloitte/articles/press-releases/most-companies-challenged-to-attract-finance-and-accounting-talent.html https://www2.deloitte.com/us/en/pages/about-deloitte/articles/press-releases/most-companies-challenged-to-attract-finance-and-accounting-talent.html http://paperpile.com/b/vjdbn8/np8sg http://paperpile.com/b/vjdbn8/np8sg https://donotpay.com/about/ http://paperpile.com/b/vjdbn8/hsbqs http://paperpile.com/b/vjdbn8/hsbqs http://paperpile.com/b/vjdbn8/hsbqs http://paperpile.com/b/vjdbn8/hsbqs http://paperpile.com/b/vjdbn8/hsbqs http://paperpile.com/b/vjdbn8/hsbqs https://www.statnews.com/2020/01/22/health-care-needs-ai-it-also-needs-human-touch/ https://www.statnews.com/2020/01/22/health-care-needs-ai-it-also-needs-human-touch/ https://www.statnews.com/2020/01/22/health-care-needs-ai-it-also-needs-human-touch/ http://paperpile.com/b/vjdbn8/pzmrl http://paperpile.com/b/vjdbn8/pzmrl http://paperpile.com/b/vjdbn8/pzmrl http://paperpile.com/b/vjdbn8/pzmrl http://paperpile.com/b/vjdbn8/pzmrl https://accounting.nridigital.com/iab_jun23/ai_in_accounting_and_tax_compliance_the_possibilities_and_pitfalls https://accounting.nridigital.com/iab_jun23/ai_in_accounting_and_tax_compliance_the_possibilities_and_pitfalls https://accounting.nridigital.com/iab_jun23/ai_in_accounting_and_tax_compliance_the_possibilities_and_pitfalls faulhaber & chaffin 11 gaetano, c. (2023, august 8). big four: ai will augment, not replace, accountants. accounting today. https://www.accountingtoday.com/news/bi g-four-squarely-on-team-human-when-itcomes-to-ai-says-tech-not-there-to-replaceaccountants generative ai & the legal profession 2023 survey report. (2023). https://www.lexisnexis.com/pdf/ln_generat ive_ai_report.pdf henderson, m. (may 10. 2022). radiology facing a global shortage. radiological society of north america. https://www.rsna.org/news/2022/may/glob al-radiologist-shortage krishnan, g., singh, s., pathania, m., gosavi, s., abhishek, s., parchani, a., & dhar, m. (2023). artificial intelligence in clinical medicine: catalyzing a sustainable global healthcare paradigm. frontiers in artificial intelligence, 6, 1227091. lexisnexis. (2023, november 6). how generative ai is disrupting law firm billing practices. community. https://www.lexisnexis.com/community/ins ights/legal/b/thought-leadership/posts/howgenerative-ai-is-disrupting-law-firmbilling-practices longoni, c., bonezzi, a., & morewedge, c. k. (2019). resistance to medical artificial intelligence. the journal of consumer research, 46(4), 629–650. marciano, j. (2023, april 17). how marketing and finance can click better. https://www.datarails.com/how-marketingand-finance-can-click-better/ mcgregor, s., & atherton, d. (2023, may 4). incident 541: chatgpt reportedly produced false court case law presented by legal counsel in court. https://incidentdatabase.ai/cite/541/ oecd. (2021). artificial intelligence, machine learning and big data in finance: opportunities, challenges, and implications for policy makers. https://www.oecd.org/finance/artificialintelligence-machine-learningbig-data-infinance.htm. park, a. (2022, november 4). how ai is changing medical imaging to improve patient care. https://time.com/6227623/aimedical-imaging-radiology/ plesner, l. l., müller, f. c., nybing, j. d., laustrup, l. c., rasmussen, f., nielsen, o. w., boesen, m., & andersen, m. b. (2023). autonomous chest radiograph reporting using ai: estimation of clinical impact. radiology, 307(3), e222268. schraeder, k. (2023, november 29). thriving in the ai-driven landscape: the shift to advisory services for cpas. https://www.dryrun.com/blog/thriving-inthe-ai-driven-landscape-the-shift-toadvisory-services-for-cpas segarra, p. (2023, november 2). linkedin. https://www.linkedin.com/pulse/futurelawyers-age-ai-adapting-changesidehustle-law-8g48e/ shine, i. (2023, may 17). these are the jobs that ai can’t replace. world economic forum. https://www.weforum.org/agenda/2023/05/ jobs-ai-cant-replace/ sns insider. (2023, may 5). artificial intelligence (ai) in accounting market to surpass usd 39.57 billion by 2030 owing to rising need for faster and more accurate accounting in private and public firms. https://www.globenewswire.com/en/newsrelease/2023/05/05/2662590/0/en/artificial -intelligence-ai-in-accounting-market-tosurpass-usd-39-57-billion-by-2030owing-to-rising-need-for-faster-andmore-accurate-accounting-in-privateand-public-firms-resear.html spiegel, j. (2023, august 30). how advisors are increasing efficiency and impact with ai. blackrock. https://www.blackrock.com/us/financialprofessionals/insights/how-advisors-use-ai http://paperpile.com/b/vjdbn8/bvpq2 http://paperpile.com/b/vjdbn8/bvpq2 http://paperpile.com/b/vjdbn8/bvpq2 http://paperpile.com/b/vjdbn8/bvpq2 http://paperpile.com/b/vjdbn8/bvpq2 https://www.accountingtoday.com/news/big-four-squarely-on-team-human-when-it-comes-to-ai-says-tech-not-there-to-replace-accountants https://www.accountingtoday.com/news/big-four-squarely-on-team-human-when-it-comes-to-ai-says-tech-not-there-to-replace-accountants https://www.accountingtoday.com/news/big-four-squarely-on-team-human-when-it-comes-to-ai-says-tech-not-there-to-replace-accountants https://www.accountingtoday.com/news/big-four-squarely-on-team-human-when-it-comes-to-ai-says-tech-not-there-to-replace-accountants http://paperpile.com/b/vjdbn8/4b7bb http://paperpile.com/b/vjdbn8/4b7bb http://paperpile.com/b/vjdbn8/4b7bb http://paperpile.com/b/vjdbn8/4b7bb https://www.lexisnexis.com/pdf/ln_generative_ai_report.pdf https://www.lexisnexis.com/pdf/ln_generative_ai_report.pdf http://paperpile.com/b/vjdbn8/dm4fg http://paperpile.com/b/vjdbn8/dm4fg http://paperpile.com/b/vjdbn8/dm4fg http://paperpile.com/b/vjdbn8/dm4fg http://paperpile.com/b/vjdbn8/dm4fg https://www.rsna.org/news/2022/may/global-radiologist-shortage https://www.rsna.org/news/2022/may/global-radiologist-shortage http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/f2wev http://paperpile.com/b/vjdbn8/ens8y http://paperpile.com/b/vjdbn8/ens8y http://paperpile.com/b/vjdbn8/ens8y http://paperpile.com/b/vjdbn8/ens8y http://paperpile.com/b/vjdbn8/ens8y http://paperpile.com/b/vjdbn8/ens8y https://www.lexisnexis.com/community/insights/legal/b/thought-leadership/posts/how-generative-ai-is-disrupting-law-firm-billing-practices https://www.lexisnexis.com/community/insights/legal/b/thought-leadership/posts/how-generative-ai-is-disrupting-law-firm-billing-practices https://www.lexisnexis.com/community/insights/legal/b/thought-leadership/posts/how-generative-ai-is-disrupting-law-firm-billing-practices https://www.lexisnexis.com/community/insights/legal/b/thought-leadership/posts/how-generative-ai-is-disrupting-law-firm-billing-practices http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/0xpsf http://paperpile.com/b/vjdbn8/q4ung http://paperpile.com/b/vjdbn8/q4ung http://paperpile.com/b/vjdbn8/q4ung http://paperpile.com/b/vjdbn8/q4ung http://paperpile.com/b/vjdbn8/q4ung https://www.datarails.com/how-marketing-and-finance-can-click-better/ https://www.datarails.com/how-marketing-and-finance-can-click-better/ http://paperpile.com/b/vjdbn8/pzb5m http://paperpile.com/b/vjdbn8/pzb5m http://paperpile.com/b/vjdbn8/pzb5m http://paperpile.com/b/vjdbn8/pzb5m http://paperpile.com/b/vjdbn8/pzb5m http://paperpile.com/b/vjdbn8/pzb5m http://paperpile.com/b/vjdbn8/pzb5m https://incidentdatabase.ai/cite/541/ http://paperpile.com/b/vjdbn8/j0b5h http://paperpile.com/b/vjdbn8/j0b5h http://paperpile.com/b/vjdbn8/j0b5h http://paperpile.com/b/vjdbn8/j0b5h http://paperpile.com/b/vjdbn8/j0b5h http://paperpile.com/b/vjdbn8/j0b5h http://paperpile.com/b/vjdbn8/j0b5h https://www.oecd.org/finance/artificial-intelligence-machine-learningbig-data-in-finance.htm. https://www.oecd.org/finance/artificial-intelligence-machine-learningbig-data-in-finance.htm. https://www.oecd.org/finance/artificial-intelligence-machine-learningbig-data-in-finance.htm. http://paperpile.com/b/vjdbn8/nldur http://paperpile.com/b/vjdbn8/nldur http://paperpile.com/b/vjdbn8/nldur http://paperpile.com/b/vjdbn8/nldur http://paperpile.com/b/vjdbn8/nldur https://time.com/6227623/ai-medical-imaging-radiology/ https://time.com/6227623/ai-medical-imaging-radiology/ http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/g8sv3 http://paperpile.com/b/vjdbn8/ppsn4 http://paperpile.com/b/vjdbn8/ppsn4 http://paperpile.com/b/vjdbn8/ppsn4 http://paperpile.com/b/vjdbn8/ppsn4 http://paperpile.com/b/vjdbn8/ppsn4 http://paperpile.com/b/vjdbn8/ppsn4 https://www.dryrun.com/blog/thriving-in-the-ai-driven-landscape-the-shift-to-advisory-services-for-cpas https://www.dryrun.com/blog/thriving-in-the-ai-driven-landscape-the-shift-to-advisory-services-for-cpas https://www.dryrun.com/blog/thriving-in-the-ai-driven-landscape-the-shift-to-advisory-services-for-cpas http://paperpile.com/b/vjdbn8/jj1rc http://paperpile.com/b/vjdbn8/jj1rc http://paperpile.com/b/vjdbn8/jj1rc http://paperpile.com/b/vjdbn8/jj1rc https://www.linkedin.com/pulse/future-lawyers-age-ai-adapting-change-sidehustle-law-8g48e/ https://www.linkedin.com/pulse/future-lawyers-age-ai-adapting-change-sidehustle-law-8g48e/ https://www.linkedin.com/pulse/future-lawyers-age-ai-adapting-change-sidehustle-law-8g48e/ http://paperpile.com/b/vjdbn8/rpo2 http://paperpile.com/b/vjdbn8/rpo2 http://paperpile.com/b/vjdbn8/rpo2 http://paperpile.com/b/vjdbn8/rpo2 http://paperpile.com/b/vjdbn8/rpo2 https://www.weforum.org/agenda/2023/05/jobs-ai-cant-replace/ https://www.weforum.org/agenda/2023/05/jobs-ai-cant-replace/ http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 http://paperpile.com/b/vjdbn8/gwin9 https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html https://www.globenewswire.com/en/news-release/2023/05/05/2662590/0/en/artificial-intelligence-ai-in-accounting-market-to-surpass-usd-39-57-billion-by-2030-owing-to-rising-need-for-faster-and-more-accurate-accounting-in-private-and-public-firms-resear.html http://paperpile.com/b/vjdbn8/xjf0k http://paperpile.com/b/vjdbn8/xjf0k http://paperpile.com/b/vjdbn8/xjf0k http://paperpile.com/b/vjdbn8/xjf0k http://paperpile.com/b/vjdbn8/xjf0k https://www.blackrock.com/us/financial-professionals/insights/how-advisors-use-ai https://www.blackrock.com/us/financial-professionals/insights/how-advisors-use-ai financial services review, 32(4) 12 spivack, s. (apr, 20 2023). human touch: an essential connection between patients and doctors. https://blogs.einsteinmed.edu/blog/2023/04 /20/human-touch-an-essential-connectionbetween-patients-and-doctors/ u.s. bureau of labor statistics. (2023a, september 6). accountants and auditors. https://www.bls.gov/ooh/business-andfinancial/accountants-and-auditors.htm u.s. bureau of labor statistics. (2023b, september 6). lawyers. https://www.bls.gov/ooh/legal/lawyers.htm warren, z. (2023, august 15). impact of ai on law firms of every size. reuters. https://legal.thomsonreuters.com/blog/aisimpact-on-law-firms-of-every-size/ weiser, b. (2023, june 22). chatgpt lawyers are ordered to consider seeking forgiveness. the new york times. https://www.nytimes.com/2023/06/22/nyre gion/lawyers-chatgpt-schwartz-loduca.html http://paperpile.com/b/vjdbn8/wfrah http://paperpile.com/b/vjdbn8/wfrah http://paperpile.com/b/vjdbn8/wfrah http://paperpile.com/b/vjdbn8/wfrah http://paperpile.com/b/vjdbn8/wfrah http://paperpile.com/b/vjdbn8/wfrah https://blogs.einsteinmed.edu/blog/2023/04/20/human-touch-an-essential-connection-between-patients-and-doctors/ https://blogs.einsteinmed.edu/blog/2023/04/20/human-touch-an-essential-connection-between-patients-and-doctors/ https://blogs.einsteinmed.edu/blog/2023/04/20/human-touch-an-essential-connection-between-patients-and-doctors/ http://paperpile.com/b/vjdbn8/oxmt2 http://paperpile.com/b/vjdbn8/oxmt2 http://paperpile.com/b/vjdbn8/oxmt2 http://paperpile.com/b/vjdbn8/oxmt2 http://paperpile.com/b/vjdbn8/oxmt2 https://www.bls.gov/ooh/business-and-financial/accountants-and-auditors.htm https://www.bls.gov/ooh/business-and-financial/accountants-and-auditors.htm http://paperpile.com/b/vjdbn8/uqf87 http://paperpile.com/b/vjdbn8/uqf87 http://paperpile.com/b/vjdbn8/uqf87 http://paperpile.com/b/vjdbn8/uqf87 http://paperpile.com/b/vjdbn8/uqf87 https://www.bls.gov/ooh/legal/lawyers.htm http://paperpile.com/b/vjdbn8/vh0zu http://paperpile.com/b/vjdbn8/vh0zu http://paperpile.com/b/vjdbn8/vh0zu http://paperpile.com/b/vjdbn8/vh0zu http://paperpile.com/b/vjdbn8/vh0zu https://legal.thomsonreuters.com/blog/ais-impact-on-law-firms-of-every-size/ https://legal.thomsonreuters.com/blog/ais-impact-on-law-firms-of-every-size/ http://paperpile.com/b/vjdbn8/jxulv http://paperpile.com/b/vjdbn8/jxulv http://paperpile.com/b/vjdbn8/jxulv http://paperpile.com/b/vjdbn8/jxulv http://paperpile.com/b/vjdbn8/jxulv http://paperpile.com/b/vjdbn8/jxulv https://www.nytimes.com/2023/06/22/nyregion/lawyers-chatgpt-schwartz-loduca.html https://www.nytimes.com/2023/06/22/nyregion/lawyers-chatgpt-schwartz-loduca.html financial services review volume 32 number 4 (2024) volume 32, no. 4 2024 editor: john e. grable, ph.d. cfp ® university of georgia advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp®, baylor university (emeritus) martin seay, ph.d., cfp ® , kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp ® , san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama jasmine fang, ph.d., massey university, new zealand mark fedenia, ph.d., university of wisconsin stu heckman, ph.d., cfp ® , texas tech university william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea thomas langdon, ph.d., roger william university, bristol, ri terrance martin, ph.d., winston-salem state university mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc lance palmer, ph.d., cfp ® , cpa®, university of georgia abed rabbani, ph.d., cfp ® , university of missouri chris robinson, ph.d., york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida issn online 1057-0810 print 1873-5673 contents grable, john e., from the editor, i-ii. faulhaber, manuela e. & chaffin, charles. artificial intelligence in accounting, medicine, and law with potential implications for financial planning: a review of literature. 1-12. whitworth, jeff. should investors defer long-term gains in taxable stock portfolios? 13-26. qing, di & reiter, miranda. racial/ethnic disparities in financial advice seeking: a decomposition analysis. 27-50. qi, jia, zhang, yu, & worthy, sheri. social determinants of health and desirable financial behaviors: the mediation effect of financial knowledge. 51-77. heo, wookjae, liu, yi, & lee, jae min. improving communication with financial consumers: insights from a study of phone call phobia. 78-102. academy of financial services officers president shawn brayman smb research consulting executive vice president program michelle cull western sydney university vice president finance thanh ngo east carolina university vice president communications kirsten macdonald griffith university vice president international relations jasmine fang massey university vice president marketing & pr cora pettipas hsbc global wealth vice president membership matt goren dalton education, cerifi immediate past president tom potts baylor university editor, financial services review john e. grable, ph.d., cfp® university of georgia directors jason anderson university of kansas norah feng massey university wookjae heo purdue university thomas korankye the university of arizona barry mulholland university of akron mustafa nourallah mid sweden university richard stebbins university of alabama yu (yulia) chang kansas state university past presidents inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 32, no. 4, 2024 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • shinae l. choi, ph.d. university of alabama • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • stu heckman, ph.d., cfp®, texas tech university • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • thomas langdon, ph.d., roger william university, bristol, ri • terrance martin, ph.d., winston-salem state university • mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., cfp®, cpa, ca, york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • brian boscaljon, penn state behrend • axton betz-hamilton, south dakota state university • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph.d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • lu fan, ph.d., cfp®, university of georgia • patti fisher, ph.d., virginia tech • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph.d. stetson university, deland, fl • philip gibson, ph.d., cfp®, winthrop university • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • john nofsinger, ph.d., university of alaska anchorage • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peerreviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ pii: s1057-0810(01)00068-3 exploitable patterns in retirement annuity returns: evidence from tiaa/cref edward m. millera, larry j. pratherb,* adepartment of economics and finance, university of new orleans, new orleans, la 70148, usa bdepartment of economics and finance, east tennessee state university, box 70686, johnson city, tn 37614, usa received 18 september 2000; received in revised form 19 december 2000; accepted 7 february 2001 abstract evidence suggests that predictabilities in asset class returns exist but transactions costs prevent exploiting them using individual securities. extant research also shows that these relationships may by exploitable through the trading of mutual funds but fails to examine whether this relationship exists within an individual fund family. this paper finds that tiaa/cref retirement annuities exhibit predictable elements that could be exploited by informed traders. the proposed trading strategy dominates a buy-and-hold strategy by producing higher raw and risk-adjusted returns. additionally, the risk is greatly reduced. © 2001 elsevier science inc. all rights reserved. jel classification:g12 keywords: return predictability; technical trading rules; lead and lag relationships among asset classes; retirement annuities; technical trading rules 1. introduction there is now substantial evidence that stock prices exhibit an appreciable amount of predictability (e.g., chalmers, edelen, & kadlec, 1999; goetzmann, ivkovic´, & rouwenhorst, 2000; hamao, masulis & ng, 1990; lo & mackinlay, 1999). this is in spite of the traditional efficient market theory argument that if an appreciable predictive element were * corresponding author. tel.:11-423-439-5668; fax:11-423-439-8583. e-mail addresses:emmef@uno.edu (e.m. miller); prather@etsu.edu (l.j. prather). financial services review 9 (2000) 219–230 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00068-3 found, investors would trade so as to exploit the predictability. miller and prather (1999) find similar return predictabilities using mutual fund investment objectives as asset class proxies. however, extant research on asset class return predictabilities does not address any specific fund family. rather, it explores the aggregate return predictabilities using asset class return indexes. since najand and prather (1999) find that common fund classification schemes fail to capture all elements of risk, it is not possible to determine whether return predictabilities exist or whether a predictable component is exploitable within any given fund family. this article addresses the issue of whether these predictabilities exist in tiaa/cref retirement annuities. this line of investigation is important for several reasons. first, tiaa/cref is one of the largest providers of retirement annuities to members of universities. therefore, it is important to determine whether the effects found in theoretical portfolios constructed by researchers are observable in portfolios that fund managers construct and investors purchase. secondly, if return predictabilities exist, can knowledgeable investors use these predictabilities to garner higher risk-adjusted returns through frequent trading? evidence suggests that some investors may be attempting to move funds on a systematic basis. a recent article inthe participant,the periodic newsletter for tiaa/cref participants, expressed concern that some investors are making frequent trades and that this is problematic for tiaa/cref. additionally, a new section was added to the college retirement equities fund prospectus, dated may 1, 2000, concerning the implementation of a market timing policy. this policy restricts the number of trades permitted in an effort to reduce the negative impact transactions costs incurred by tiaa/cref due to this frequent trading. since this policy would not prevent trading on the turn-of-the-month effect found by kunkel and compton (1998), it is possible another driver of frequent trading exists. empirical testing uses the granger causality method to examine daily returns of retirement annuities during a 1453-day period. the granger causality method enables us to test directly whether knowing past returns on one asset class will aid in the forecast of returns for another asset class. next, we develop a trading strategy that attempts to exploit observed statistical relationships. then we test the trading strategy on a holdout sample and find that it dominates a buy-and-hold strategy in terms of raw and risk-adjusted returns. finally, we show that the new market timing policy implemented by tiaa/cref does not eliminate the ability of informed traders to exploit predictable components of returns. as an example of the effectiveness of this trading rule, an initial investment of $100,000 at the beginning of the holdout period would have increased to $186,450 if invested entirely in the index fund whereas it would have grown to $253,180 if moved systematically. thus, frequent trading would lead to a profit of $66,730 while reducing risk. 2. background while individual stocks show only a low predictability, when these stocks are combined into portfolios, the predictability may be higher or even different in nature. lo and mackinlay (1999) found this for small stocks. small stocks had a low correlation with a lagged index of larger stocks and tended to have a negative autocorrelation over short periods. this 220 e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 is probably the result of idiosyncratic risk. however, when the stocks were combined into portfolios, they showed an appreciable positive correlation with the earlier performance of large stock indices. this results because the noise in the time series of individual stocks is diversified away when the stocks are combined into portfolios, leaving systematic effects to be revealed. miller and prather (1999) show that studying mutual funds is one way to quickly study portfolios, since each fund represents a portfolio. further, they show that, on average, exploitable regularities exist. this is important since mutual funds can be frequently traded with negligible costs. within a fund family, transfers can usually be made by phone at no cost. even between fund families, trades can often be made at zero or low cost. usually one can sell shares in a mutual fund back to the fund for net asset value and the proceeds can then be reinvested in a new no-load fund at no expense. retirement annuities function in much the same way as mutual funds; therefore, they are expected to exhibit similar patterns. a benefit of using retirement annuities to exploit predictable patterns in security returns, over individual stocks or even mutual funds, is the lack of taxes and transaction costs. trading mutual funds, or retirement annuities, lacks the same self-correcting forces that stock trading has. an example can illustrate the self-correcting nature of trading. suppose that a rule is found that predicts that growth stocks will rise by an appreciable amount from tuesday’s close to wednesday’s close. a single investor may be able to profit from this rule by buying at the close on tuesday and selling at wednesday close. however, if many investors discover the rule, and buy at tuesday’s close and sell at wednesday’s close, they will bid the tuesday’s closing price up and lower wednesday’s closing price, tending to eliminate the effect. the same effect that produces the predictability in the stock prices would be expected to produce predictability in the net asset values of a mutual fund or retirement annuity. thus, an investor might be able to make a profit by buying the mutual fund, or retirement annuity, at the close on tuesday (done by a phone call before the close) and then selling it at the close on wednesday. clearly, a single investor could profit from such predictability. however, large numbers of investors might also profit from it. imagine investors moving a sum of money that would eliminate the effect if the trading were done directly in stocks. the same trading would be unlikely to eliminate the effect if the trading was done in mutual funds or retirement annuities. the reason is that the fund manager does not have the influx of funds until after the close tuesday (when the trades are actually done). therefore, if the fund manager takes no action until the new funds are received, the effect could continue since no action was taken to affect stock prices. alternately, suppose the fund manager responds to the influx of new funds by buying more stocks immediately. two possibilities exist. first, if his buying is too little to affect the prices, the effect continues. however, if his buying is large enough to affect prices, his buying raises the prices of the stocks his fund buys, and the net asset value is even greater when it is calculated at wednesday’s close. the trading actually accentuated the price change (this would be especially likely for a specialized fund whose manager buys and sells within a small list of stocks that may be limited in liquidity, such as a country fund). thus, day-to-day predictabilities could persist even if they were known. 221e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 3. data and methodology 3.1. data in order to ascertain if predictabilities in retirement returns exist, we carry out empirical analysis. the sample consists of 1,453 daily returns on the following tiaa/cref retirement annuities: money market, bond, stock, index, growth, and global. daily return data for each of the selected retirement annuities during the period may 2, 1994 through december 31, 1999 was obtained from tiaa/cref. the objectives of these funds remained constant during the period of our study. 3.2. computation of returns daily returns are computed for each retirement annuity by taking the change in net asset value (nav) for an investment for each of the 1,453 days in our sample, as shown in eq. (1) ri,t 5 navi,t 2 navi,t21 navi,t21 (1) where ri,t is the return on retirement annuity i during day t, navi,t is the net asset value of an investment in retirement annuity i at time t. 3.3. methodologies employed several methodologies have been employed in the literature to ascertain lead and lag relationships. however, one of the most popular is the granger causality test. the popularity of the granger causality test stems from its ability to use an observable time series to test whether using information contained in that time series would be beneficial in forecasting another time series. granger’s (1969) is the only established method that permits testing whether one portfolio’s returns are predictable by another portfolio’s returns after controlling for autocorrelation. therefore, it is useful in inferring the relative predictability between stochastic variables to ascertain a lead-lag structure. the granger approach to the question of whether x causes y is to determine the amount of the current y that can be explained by past values of y and then to ascertain whether adding lagged values of x can improve the explanation. y is said to be granger-caused by x if x helps in the prediction of y, or equivalently if the coefficients on the lagged xs are statistically significant. it is important to note that the statement “x granger causes y” does not imply that y is the effect or the result of x. granger causality measures information content but does not indicate causality in the common use of the term. formally, a time series {xt} “granger causes” another time series {y t} if using past values of x can improve the forecast of y. since the objective of this research is to determine whether past returns from one retirement annuity are useful in predicting the future returns of another retirement annuity, this paper follows richardson and peterson (1999) and uses techniques based upon granger causality to ascertain lead and lag relationships in returns between various retirement annuities. 222 e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 our test for return predictability uses eqs. (2) and (3): ri,t 5 a 1 o k51 n bi,k ri,t2k 1 o k51 n bj,k rj,t2k 1 «t (2) rj,t 5 d 1 o k51 n gi,k ri,t2k 1 o k51 n gi,k rj,t2k 1 yt (3) where n is the number of lags estimated; ri,t is the return series for asset class i; rj,t is the return series for asset class j;a and d are the estimated intercepts;bi,k and gi,k are the coefficients for asset class i’s return series lagged t-k periods;bj,k andgj,k are the coefficients for asset class j’s return series lagged t-k periods; andet andyt are the normally distributed error terms. the granger causality tests consist of whether all the coefficients of the lagged xs in eq. (2) may be considered to be zero, and similarly whether the coefficients of the lagged ys in eq. (3) are zero. thus, the null hypotheses being tested are that x does not granger-cause y and that y does not granger-cause x. 4. empirical results of lead and lag relationships among asset classes in order to conduct empirical investigation, we first divided our sample into two subsamples with an approximately equal number of observations. this division allows testing lead and lag relationships and the development of trading rules with one subsample and then testing the dominance of those rules over a buy and hold strategy using the holdout sample. the first subsample contains 725 daily observations from may 2, 1994 through february 28, 1997. the holdout sample contains 728 daily observations from march 3, 1997 through december 31, 1999. the first step in assessing whether it is possible to utilize return patterns from one asset class to foretell the average future return of another asset class is to determine the instantaneous correlation. if instantaneous correlations are high, the impact of any trading strategy is mitigated. table 1 provides the instantaneous correlation and covariance matrices for the six asset classes in our sample. the instantaneous correlation table is of interest in its own right since it shows that diversification between different asset classes is possible. the relatively low correlations between global funds and other categories of funds suggest that global funds can indeed be used to reduce risk when combined with domestic u.s. funds. the highest correlation (0.974) is between index and the stock funds. the high correlation between the index and the stock fund is because approximately two thirds of the stock fund is indexed. the instantaneous correlation structure suggests that investors may be able to benefit from an asset reallocation strategy if observing the returns on one asset class could provide information about the future returns of another asset class. 223e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 4.1. tests of granger causality since investors may be able to benefit from an asset reallocation strategy if observing the returns on one retirement annuity could provide information about the future returns of another retirement annuity, we examine whether that situation exists (on average) using granger causality. table 2 provides the results of granger causality tests of lead-lag table 1 instantaneous correlation and covariance between asset classes panel a instantaneous correlation stk mm bond glob gro indx stk 1.000000 0.036410 0.530942 0.742750 0.942812 0.973581 mm 0.036410 1.000000 0.110867 0.007747 0.004376 0.030099 bond 0.530942 0.110867 1.000000 0.269797 0.474017 0.551139 glob 0.742750 0.007747 0.269797 1.000000 0.629086 0.607968 gro 0.942812 0.004376 0.474017 0.629086 1.000000 0.961849 indx 0.973581 0.030099 0.551139 0.607958 0.961849 1.000000 panel b instantaneous covariance matrix bond glob gro indx mm stk bond 7.42e-06 3.63e-06 8.67e-06 9.11e-06 3.81e-08 7.41e-06 glob 3.63e-06 2.44e-05 2.09e-05 1.82e-05 4.82e-09 1.88e-05 gro 8.67e-06 2.09e-05 4.51e-05 3.92e-05 3.70e-09 3.24e-05 indx 9.11e-06 1.82e-05 3.92e-05 3.68e-05 2.30e-08 3.03e-05 mm 3.81e-08 4.82e-09 3.70e-09 2.30e-08 1.59e-08 2.35e-08 stk 7.41e-06 1.88e-05 3.24e-05 3.03e-05 2.35e-08 2.62e-05 table 2 pairwise granger causality tests for a one-period lag stk mm bond glob gro indx stk 5.2915 0.4090 60.4311 0.6969 0.2362 (0.0217) (0.5227) (0.0000) (0.4041) (0.6272) mm 0.6290 3.6568 0.2450 0.9709 0.6189 (0.4280) (0.0563) (0.6208) (0.3248) (0.4317) bond 0.4412 4.4005 32.5721 0.0833 0.1137 (0.5068) (0.0363) (0.0000) (0.7730) (0.7361) glob 0.0474 2.5997 3.0471 0.7126 1.0155 (0.8277) (0.1073) (0.0813) (0.3989) (0.3139) gro 0.0390 2.9648 0.0051 59.4262 0.1944 (0.8435) (0.0855) (0.9431) (0.0000) (0.6594) indx 1.4763 4.7076 0.0929 69.0951 1.2504 (0.2247) (0.0304) (0.7607) (0.0000) (0.2639) a matrix of the f-statistic (p-value in parenthesis) for the granger causality test of lead and lag relationships is presented for the initial sample of 725 trading days between may 2, 1994 and february 28, 1997. the test statistics are for testing the null hypothesis that the asset class returns in the row do not granger cause the asset class returns in the column for a one-day lag. 224 e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 relationships of the thirty possible test pair combinations over the 725-day sample period. these tests of granger causality are stated in null hypothesis form and examine a one-period (trading-day) lag. interestingly, seven test pairs yielded significant lead-lag relationships at the five-percent level (the null hypothesis was rejected) and three additional test pairs were significant at the ten-percent level. the finding of seven significant causalities at the five-percent level is far above the one or two that would be expected from chance alone. this suggests that trading patterns may exist and that investors may develop a personal strategy to maximize their utility by using this lead-lag information to enhance the movement of funds among preselected retirement annuities that provide the desired level of risk. one caveat exists. granger causing a relationship does not mean one asset class causes another, it simply means that a statistical relationship exists such that knowing past returns of one class will help determine future returns in another. empirical analysis suggests that the most significant statistical relationships are for: (1) index leading global, (2) stock leading global, (3) growth leading global, (4) bond leading global, (5) stock leading money market, (6) index leading money market, and (7) bond leading money market. each of these relationships is significant at better than the five-percent level. 4.2. practical limitations as miller and prather (1999) point out, one limitation of trading on a one-day lag is that it presupposes that an investor could determine returns on the asset class during any given day, sell the fund at the close, and immediately reposition the assets. however, since this study is using retirement annuities as the securities in the asset class instead of individual stocks, this is not possible. the returns are not known with certainty until after the close of business. this permits the fund to compute net asset value (nav) after closing prices of the securities in the portfolio are determined. sell orders are executed at the closing nav on the day the order is processed and buy orders are executed at the closing nav on the day the buy order is processed. therefore, if after observing past changes in nav, an order were placed to sell/buy, the prices would be the prices on event day t1 1, not day t. miller and prather (1999) deal with this complication by examining relationships based on a two-day lag and form a conservative trading strategy using the two-day lagged returns. however, they state that it is reasonable to believe that investors could monitor index returns and make decisions to trade based on the index that proxies for their current asset class. we find a similar two-day lag relationship with tiaa/cref data (not reported in the paper). however, we base our trading rule on the ability to monitor the index and therefore, use a one-day lag to report results. this is explained in the following section. 4.3. exploitation of return predictability exploiting information provided by the granger causality tests requires moving assets based on the strength of the statistical relationship. therefore, it is necessary to examine the granger f-statistic (p-value) to find the strongest statistical relationship. results in table 2 suggest that the strongest relationship is that index causes global. this is consistent with hamao, masulis and ng (1990) who document spillover effects between u.s. and interna225e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 tional markets. one explanation for this is that it arises from an asynchronous pricing problem (e.g., chalmers, edelen, & kadlec, 1999; goetzmann, ivkovic´, and rouwenhorst, 2000; varela, 1997). when one buys a fund the net asset value is based on the latest available prices as of the time the new york stock exchange closes (4 p.m. eastern time). however, for asian and european stocks the latest available prices are many hours old. some pieces of news that have affected the u.s. stocks prices probably have also affected the values of these foreign stocks, but the prices will not actually change until the asian and european markets open the next day. this creates an opportunity to trade at stale prices. this relationship is fortunate for traders because the returns on an index fund closely approximate the returns on the index it tracks. the tiaa/cref prospectus (may 2000) indicates that the index fund tracks the russell 3000 index. the russell 3000 index is an unmanaged index of 3,000 of the largest u.s. companies (based on market capitalization). the index fund does not hold all 3000 stocks but uses a sampling technique to closely replicate the index. the goal of tiaa/cref management is to attempt to match the total return on the russell 3000 index but with any index fund, they may not always do so. as a practical matter, the tracking error (1-r2) of index funds is typically very small. while the values of the index fund are calculated only at the end of the trading day, the index itself is available during the day. thus, the result of a strategy of making decisions on the closing price of the index fund can be approximated by checking the index just before the closing, and then trading based on the performance of the index. this is important since tiaa/cref permits telephone or the inter/act internet service transactions. any transactions received before the market closes will be executed that day. the tiaa/cref prospectus indicates that the global fund invests at least 65% of its assets in equity securities of foreign and domestic companies. typically, at least 40% of assets are allocated to foreign securities and 25% of assets are allocated to domestic securities. the remaining 35% of assets are distributed between foreign and domestic securities based upon market conditions. the cref global fund usually has only a minority of its funds in foreign stocks. many other fund families include foreign stock funds that hold most of their equity positions in asian and european stocks. if there is an exploitable effect for the cref funds, a similar strategy would probably be even more profitable for many other fund families and retirement plans that permit frequent trading. since investors can monitor the russell 3000, our trading strategy is based on a one-day lag. this strategy calls for moving assets from index to global on a positive return in index. this move is made since the investor can use knowledge of return predictability to predict a positive global return. the investor stays in global until a negative return is observed for index. upon observing a negative return on index, the investor returns to index (to avoid a negative return in global during the upcoming period). since global does not lead index, there is no reason to believe that index will be negative in the next period. this pattern is consistent with the findings of hamao, masulis and ng (1990). it is worthwhile examining the magnitude of the predictable component to determine the potential gain from trading. table 3 provides results of regressing returns of the global annuity on one-day lagged returns of the index annuity. results suggest that a positive one-percent return by the index annuity on day t should lead to an average positive return on the global annuity of 0.3% return on day t1 1. 226 e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 a practical constraint faced by an informed trader is that tiaa/cref will not permit frequent movement of assets from one asset class to another. the cref prospectus dated may 1, 2000, states “participants who make more than three transfers out of any account (other than the money market account) in a calendar month will be advised that if this transfer frequency continues, we will suspend their ability to make telephone, fax, and internet transfers” (p. 36). this places constraints on knowledgeable investors attempting to capitalize on return predictabilities. an examination of the data shows that the index returns were positive on 387 days in the second sample period (53%). returns for global were similar with 414 positive daily returns (57%). additionally, global had a positive return on 254 of the 387 trading days following a positive return on the index (66%). therefore, an examination of the return distributions is required prior to attempting to form any implementable rule since the frequent trading suggested by this simple rule would be prohibited by tiaa/cref. in order to limit the number of trades, the sample was examined to ascertain the characteristics of the distribution of returns. the goal is to be able to trade only on the most positive index returns, not just any positive index return. as a starting point, we selected a return slightly higher than the top twenty percent (0.0093) as the trigger point for moving into global. prior to executing the strategy, we also examined the number of returns in this top twenty percentage category and how frequently positive index returns were followed by positive global returns. of the series of returns, 146 were above the selected trigger point. interestingly, 99 of the 146 large positive index returns were followed by positive global returns (68%). we also selected a trigger point to move back into index based on negative index return. again, the trigger point was a bottom twenty-percent return (20.0068). this type of constrained trading should be useful in reducing the number of trades and avoiding large negative returns, while capitalizing on expected large positive returns. 4.4. risk and return of trading strategies the returns, risks, and sharpe (1966) measures for both buy-and-hold strategies and the proposed trading rule are presented in table 4. the sharpe measure (s) is computed as s5 [rp-rf]/sp. the trading rule portfolio has the highest average daily return of any of the portfolios (0.00142) compared to 0.00086 for stock (lowest equity return) and 0.00113 for table 3 predictable return components of selected trading strategy index leading global variable coefficient std. error t-statistic prob. r2 index(-1) 0.302026 0.028150 10.72927 0.0000 0.1375 a 0.000286 0.000172 1.65577 0.0982 the regression of global returns on the one-day lagged predictor variable (index) is presented to determine the magnitude of the predictable element of returns. columns one through six present the variable, coefficient, standard error, t-statistic, probability (p-value), and coefficient of determination (r2) of the regressions, respectively. 227e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 growth (highest equity return). it also has a lower total risk, as measured by standard deviation (0.01062), than either the index (0.01132) or the growth portfolios (0.01298). this characteristic leads to the sharpe measure of the trading rule portfolio (0.11488) being higher than any buy-and-hold strategy. none of these outcomes are surprising given the strong granger relationship. trading on granger relationships has increased the probability of obtaining a positive return in the period following a positive return in the index and reduced the probability of receiving a large negative return in global. this tends to increase average returns and to reduce variability of returns (risk), which leads directly to the higher sharpe measure. as a test of robustness, the jensen (1968) measure of risk-adjusted returns was computed to determine whether the positive risk-adjusted returns were statistically significant. the jensen measure is computed as rp-rf 5 a 1 bp (rm 2 rf) 1 e. for our purposes, the ordinary least squares regression uses the cref money market returns as the risk-free proxy and the cref index returns as the market proxy. this approach was selected since tiaa/ cref investors can form a portfolio using these asset classes to achieve their desired level of systematic risk. this is a realistic approach since it represents actual returns investors could achieve after costs. table 5 shows that the model fit is good since more than 88% of the trading rule returns are explained by the model. this is comforting since brown and brown (1987) and lehmann and modest (1987) find that the market proxy is important when attempting to determine true performance. risk-adjusted returns of the trading rule are positive and statistically significant at better than the one-percent level. additionally, systematic risk of the trading rule (b 5 0.88) is less than that of the index. table 4 risk and return of sample portfolios asset class return std. dev. sharpe stk 0.00086 0.01015 0.06535 bond 0.00022 0.00230 0.00732 glob 0.00091 0.00921 0.07639 gro 0.00113 0.01298 0.07188 indx 0.00092 0.01132 0.06346 rule 0.00142 0.01062 0.11488 column one is the asset class, columns two and three are the average arithmetic return and standard deviation of returns. column four provides the sharpe measure. table 5 risk-adjusted return of trading rule a t-statistic p-value b r2 n 0.00059 4.40436 0.00001 0.88287 0.88637 728 (.00013) (0.01173) the results of the market model regression are presented. columns one through three present the risk-adjusted return (a), the t-statistic, and p-value for the two-tailed hypothesis test that risk-adjusted return equals zero. columns four through six present the systematic risk (b), coefficient of determination (r2), and number of observations (n). standard errors are in parenthesis below the coefficient estimates. 228 e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 given the superior sharpe and jensen performance measures for the trading strategy, it appears that informed investors were able to exploit asset class return predictabilities in tiaa/cref family retirement annuities. to achieve this result, 139 trades were conducted over 728 trading days for an average trading frequency of approximately 3.8 trades per month. given the current restrictions that limit trading out of a given annuity to three trades per month, it is important to determine if this constraint had been violated by our trading rule. examination of the trading pattern reveals that in seven of the 34 months the trading limit would have been reached in at least one of the annuities and that in three months, some trades may have been prohibited. to ascertain the impact of this, we took a worst-case approach and assumed that the trades would have been prohibited. therefore, the investor would be required to remain in the current asset class once the trading limit was reached. results show that the limits had little impact on previous results. the average return for the constrained trading strategy is 0.00133 and the standard deviation is 0.010574, which produces a sharpe measure of 0.106961. this compares favorably with the results of buy-and-hold strategies in table 4 and is slightly less than the unconstrained rule. results of the jensen measure are also similar to the unconstrained rule. risk-adjusted return is 0.0005 and statistically significant with a p-value of 0.00018. additionally, risk and model fit are essentially unchanged from the unconstrained model. 5. conclusion recent evidence suggests that asset class returns possess a predictable component. however, transactions costs such as commissions and bid-ask spreads prevent investors from exploiting this predictability. this study examines asset class returns of tiaa/cref retirement annuities for several reasons. first, investors can trade these retirement annuities without incurring trading costs. additionally, these annuities should exhibit return predictabilities similar to mutual funds and recent evidence suggests that mutual fund asset class indices exhibit a predictable component. our primary goal is to learn whether tiaa/cref retirement annuities exhibit asset class return predictability and to see if a profitable trading scheme exists. we performed granger causality tests and found that the retirement annuity asset class returns exhibited a predictable element. using the strength of the predictable element, we developed a trading strategy and tested that trading rule on a holdout sample. examination of risks and returns of the trading rule and buy-and-hold strategies show that the trading rule has superior sharpe and jensen performance measures compared with a buy-and-hold strategy. further, the recent trading constraints implemented by tiaa/cref do not eliminate the ability of informed investors to enhance their reward-to-risk ratio through trading based on statistical relationships. this finding has implications for management at tiaa/cref, portfolio managers, and investors seeking the best risk-return relationship. references brown, k., & brown, g. (1987). does the composition of the market portfolio really matter?journal of portfolio management,(winter), 26–32. 229e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 chalmers, j., edelen, j., & kadlec, g. (1999). the wildcard option in transacting mutual-fund shares. unpublished working paper, university of pennsylvania. college retirement equities fund prospectus, individual, group, and tax-deferred variable annuities. (may 1) 2000, pp. 1–45. goetzmann, w., ivkovic´, z., & rouwenhorst, g. (2000). day trading international mutual funds: evidence and policy solutions. unpublished working paper, yale university. granger, c. (1969). investigating causal relationships by econometric models and cross-spectral models.econometrica, 37,424–438. hamao, y., masulis, r., & ng, v. (1990). correlations in price changes and volatility across international stock markets.the review of financial studies, 3,281–307. jensen, m. (1968). the performance of mutual funds in the period 1945–1964.journal of finance, 23,389–419. kunkel, r., & compton, w. (1998). a tax-free exploitation of the turn-of-the-month effect: c.r.e.f.financial services review, 7,11–23. lehmann, b., & modest, d. (1987). mutual fund performance evaluation: a comparison of benchmarks and benchmark comparisons.journal of finance, 42,233–265. lo, a., & mackinlay, a. (1999).a non-random walk down wall street.princeton, new jersey: princeton. miller, e., & prather, l. (1999). lead and lag relationships among asset classes: evidence from the mutual fund industry. (working paper, east tennessee state university). najand, m., & prather, l. (1999). the risk level discriminatory power of mutual fund investment objectives: additional evidence.journal of financial markets, 2,307–328. sharpe, w. (1966). mutual fund performance.journal of business, 39,119–138. richardson, t., & peterson, d. (1999). the cross-autocorrelation of sized-based portfolio returns is not an artifact of portfolio autocorrelation.the journal of financial research, 22,1–13. varela, o. (1997). efficient market implications of institutional practices in obtaining net asset values for asian-market based mutual funds in the u.s. unpublished working paper no. 8–97, university of new orleans). 230 e.m. miller, l.j. prather / financial services review 9 (2000) 219–230 pii: s1057-0810(01)00072-5 strategic asset allocation for individual investors: the impact of the present value of social security benefits steve p. frasera, william w. jenningsb,*, david r. kingc adoctoral pre-candidate, department of finance, college of business administration, university of south florida, tampa, fl, usa bassistant professor of finance, hq usafa/dfm, u.s. air force academy, academy, co 80840, usa cdoctoral pre-candidate, department of strategy, college of business, indiana university, bloomington, ia, usa received 17 january 2001; received in revised form 1 march 2001; accepted 3 april 2001 abstract this paper demonstrates the dramatic effect of social security wealth on individuals’ asset allocation. we first discuss why social security wealth should be included in portfolio asset-mix decisions. we then draw parallels between social security benefits and inflation-indexed treasury bonds to help quantify the present value of social security benefits. finally, we show the portfolio impact of including social security wealth under several asset-mix decision rules. excluding social security wealth from the asset mix decision results in suboptimal portfolios. including social security wealth provides an incentive for including more stock in the asset mix. © 2001 elsevier science inc. all rights reserved. jel classification: g11; h55; j26; j17; k13 keywords: social security; portfolio choice; mean-variance optimization; valuation 1. introduction the question of retirement income adequacy is a matter of national concern given an aging population and the imminent baby-boomer retirement. although defined contribution plans * corresponding author. tel.: �1-719-333-4130; fax: �1-719-333-6880. e-mail addresses: sfraser@coba.usf.edu (s.p. fraser), wj@williamjennings.com or william.jennings@usafa. af.mil (w.w. jennings), drking@indiana.edu (d.r. king). financial services review 9 (2000) 295–326 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00072-5 have overtaken traditional pensions, one defined benefit plan—social security—affects virtually everyone. only a limited group of workers is excluded from social security. thus, comprehensive retirement planning should consider social security benefits. a typical financial planning approach deducts social security payments from a retirement incomeneeds amount to compute a retirement income gap that investments must fund. however, this income-adequacy approach ignores the effect of social security on portfolio asset-mix decisions. this is an inconsistency—if income-adequacy computations include social security, strategic asset allocation should also incorporate social security. a fundamental aspect of financial planning is ensuring there is adequate retirement income. one means of doing so is to maximize the utility of one’s retirement benefits— including social security. to that end, we consider three questions: first, should investors consider social security when making asset mix decisions? second, how should an investor determine the value of social security benefits? third, how does proper valuation of social security affect the asset mix of a financial portfolio? we address these questions in separate sections below. scott (1995, 1997), reichenstein (1998, 2000) and others advance a concept we label an individual’s “total portfolio.” this total portfolio includes the traditional financial portfolio (stocks, bonds, cash, etc.) but adds other relevant assets. scott and reichenstein both argue that total portfolios should include the present value of social security benefits (which we call “social security wealth”). few would argue that social security is unimportant, but even the financially sophisticated regularly exclude social security wealth from portfolio analysis. as an example, two popular financial planning textbooks (gitman & joehnk, 1999, and kapoor, dlabay & hughes, 1999) acknowledge that social security reduces, to a degree, the need for other retirement income but ignore social security wealth’s impact on portfolio composition. we take scott’s and reichenstein’s valuable insights and advance them with a detailed analysis of how to measure and include social security as a portfolio asset. despite the exclusion of social security from formal portfolio analysis, many people implicitly include social security as a portfolio asset as a matter of practice. feldstein (1974) shows “social security substantially depresses personal savings.” if social security empirically substitutes for savings, then the typical social security beneficiary is implicitly incorporating benefits in his or her portfolio. hubbard (1985) shows that social security wealth affects both an individual’s choice to include particular investments and their subsequent portfolio weighting. individuals are inherently including social security in the asset mix decision, and portfolio analysis literature should include social security as well. to address this issue, we first quantify the present value of social security benefits. we use the similarities between inflation-indexed treasury bonds and social security benefits to value those benefits. both social security and treasury inflation protection securities (tips) have cash flows linked to the consumer price index (cpi), and both are federal obligations. while there is slight ambiguity in the tips-social security analogy (which we discuss in section 3.2 below), tips are good approximations for social security wealth. with this wealth value computed, we investigate the effect of social security benefits on strategic asset allocation under several common asset-mix decision rules. there are numerous approaches to the asset allocation decision ranging from the simple to complex. we consider simple fixed mixes, life-cycle rules and consensus expert views. in 296 s.p. fraser et al. / financial services review 9 (2000) 295–326 each case, the impact on an individual’s asset allocation of including social security wealth is dramatic. beyond these simple diversified portfolio approaches, we assess the impact of social security under mean-variance optimization. social security benefits shift the meanvariance efficient frontier and offer substantial risk-reduction opportunities. since social security wealth behaves much like an inflation-indexed treasury bond, considering it in the total portfolio has a striking impact on the optimized financial portfolio. by answering the three questions (should we include social security? what is its value? and what is its impact?), we create arguments for adding stock to the asset mix. siegel (1998) and others prescribe higher stock allocations than typical asset allocation models because of their long-run risk-return benefits. for some investors, investing in bonds can be comforting. perhaps this is the case because their predictable cash flow allows them to plan their spending or because bond’s stable returns can be reassuring. the notion that social security payments are very similar to treasury bond coupon payments is simple, intuitive, and persuasive—once the similarity is considered. to the extent it is persuasive, this analysis encourages increased stockholdings to achieve target asset allocations. we organize the remainder of this paper as follows: section 2 discusses the inclusion of social security wealth in an investor’s portfolio. section 3 presents our arguments for using tips to value social security as well as our specific valuation method and the underlying assumptions. section 4 shows the asset allocation implications of valuing social security wealth. section 5 summarizes our findings, discusses extensions and includes a simple one-paragraph description of how to apply our findings. 2. social security wealth as a portfolio asset despite widespread analysis of optimal portfolio asset allocation policy, there is little consensus on which assets to include. given the dominance of strategic asset allocation decisions over security selection and market timing decisions (brinson, hood & beebower, 1986), there is a critical need for accuracy in determining the portfolio. reichenstein (1998, 2000) reviews and analyzes the inclusion and valuation of portfolio assets. he advances and enhances the scott (1995, 1997) view of an “investment portfolio”; that is, assets that generate spending money or that can be sold for spending money. the “investment portfolio” is larger than a pure financial portfolio. scott and reichenstein both consider social security wealth a portfolio asset. beyond the limited individual finance literature, there is legal economics research on the value of retirement benefits (e.g., stoller, 1992; rosenman & fort, 1992). valuation is relevant to divorce, personal injury and wrongful death cases. stoller notes, “the present value of a spouse’s pension can easily be the most valuable asset that a couple possesses.” as a valuable asset and as the most-portable pension, social security wealth should affect portfolio analysis as well. social security is a savings program in some respects. foregoing consumption now in order to consume later is a prime reason for saving. similarly, social security reallocates consumption across time (albeit via forced intergenerational transfers). this similarity supports thinking of social security wealth—like other savings—as a portfolio component. 297s.p. fraser et al. / financial services review 9 (2000) 295–326 inflation offers another argument for including social security in the portfolio. inflation is a big risk to retirees since they have limited opportunities to save more, work longer or work harder after an inflationary period. the inflation protection inherent in social security has historically been a wealth protector (dalio & bernstein, 1999). to the extent that portfolios exist for future consumption, social security’s inflation protection suggests it be considered part of the portfolio. counterarguments to including social security wealth also exist; one is simplicity. the value of social security benefits relies on many hard-to-estimate variables. however, simply ignoring social security wealth can hurt portfolio efficiency. another counterargument is pessimism. planning for the worst can lead to the pleasant condition of having greater retirement wealth than required. this approach appeals to the risk averse. however, the pessimistic approach ignores the economic reality that a suboptimal risk-return tradeoff is a disservice to even the most risk-averse. the most pessimistic view of social security is that the system will fail before an individual retires. in response, there are several proposals for social security reform including those that incorporate elements of privatization. if that were to occur, the system would include asset accounts that could hold stocks and bonds. even partially privatized social security has clear portfolio implications. alternative reforms that are more needs-based behave like portfolio insurance. if a reformed social security system has portfolio management implications, then the current social security should be included in portfolios as well. 3. valuation to be included in a portfolio, social security wealth must be valued. to value social security wealth, we first discuss the nature of social security benefits. we then consider our market-based proxy for social security cash flows—inflation-indexed treasury bonds—and draw parallels between the two. finally, we undertake the actual valuation. 3.1. social security benefits the social security act encompasses numerous social insurance programs. this paper focuses only on its retirement aspects. for simplicity, we ignore other aspects of the act including nonspouse survivor benefits, disability insurance, death benefits, medicare, unemployment insurance, black lung benefits and supplemental security income. within the retirement aspects, we consider only the old age and survivors insurance (oasi) program— and then only the wage earner retirement, spousal retirement and widow(er) retirement benefits. our approach is conservative; considering other social security benefits would increase the value of social security benefits and amplify any impact of social security wealth on asset allocation. social security retirement benefits are available to retired insured workers age 62 and over, the spouse (age 62 and over) of retired insured workers and the surviving spouse (age 60 and over) of a deceased insured worker. almost everyone in these age groups is included—social security excludes only pre-1984 federal workers, some state employees, election officials, 298 s.p. fraser et al. / financial services review 9 (2000) 295–326 railroad employees, student nurses, paperboys and girls, clergy who opt out, and members of certain religious sects like the old order amish. a worker becomes fully insured after 40 quarters of earning a minimal amount (indexed to $3,120 annually in 2000). full retirement age was 65, but began rising in 2000 and will reach age 67 in 2020. (full retirement age is based on year of birth; details are in thomas (1999, p. 76).) reduced retirement benefits are, and will continue to be, available at age 62. such reduced benefits were 80% of the primary insurance amount (pia) in 1999 and gradually decline to 70% when the full retirement age is 67. if an individual continues to work past full retirement age, social security benefits increase. a piecewise linear function of the average indexed monthly earnings (aime) credited to a worker’s account determines the pia benefit amount. in 2000, the pia consists of 90% of the first $531 of aime, 32% of the next $2,671 of aime and 15% of aime above that. aime adjusts a worker’s historical wages for average wage inflation. the maximum earnings credited to an account is the maximum amount subject to social security tax (indexed to $76,200 in 2000), so earning more than that maximum does not increase benefits. the maximum monthly pia a worker may receive is $1,433 in 2000. spouses of insured workers are entitled to the greater of their own pia or one-half of the worker’s pia. a widow(er) past full retirement age is entitled to 100% of the worker’s pia although reductions similar to those for early retirement are available after age 60. social security benefits may be subject to income tax. if social security is the only source of income, benefits are generally tax-exempt. if annual taxable income for 2000 (including tax-exempt interest) is greater than $34,000 if single or $44,000 if married, 85% of the benefits are taxable. all social security benefits can be lost if the recipient keeps working and has annual wage income above certain thresholds (e.g., $10,080 if under full retirement age). note that the senior citizens’ freedom to work act of 2000 only eliminated the retirement earnings test for individuals aged 65–69. additionally, social security benefits are effectively lost if the worker’s other pensions are coordinated with social security (as with the federal employee’s retirement system and some private pensions). thomas (1999) provides additional detail about social security. 3.2. inflation-indexed treasury securities we use treasury securities as our market-based proxy for valuing social security wealth. treasury securities, and specifically treasury inflation protection securities (tips), are appropriate for two reasons. first, both social security and tips are senior federal obligations effectively backed by the full faith and credit of the government. second, tips are appropriate because they distribute cash flows indexed to inflation (i.e., that are constant real payments). although the future of social security is the subject of much speculation, it can be valued with sufficient detail for our analysis. using tips allows us to quantify the rate of return on inflation-indexed federal obligations; today’s tips real yield is appropriate for discounting the projected constant real social security benefit payments. tips, like social security, have unknown nominal cash flows but predictable real cash flows. both instruments include embedded protection against changes in inflation—a feature that ordinary bonds omit. tips are bonds and notes with semiannual interest payments based 299s.p. fraser et al. / financial services review 9 (2000) 295–326 on a fixed coupon rate. however, the underlying principal amount adjusts for inflation, and the coupon payment is calculated based on the inflation-adjusted principal amount. using tips yields to quantify values of real federal obligations requires acknowledging some caveats. first, cash flows from tips and social security depend on different versions of cpi. tips measure inflation with the consumer price index for urban consumers; social security benefits increase with the consumer price index for urban wage earners and clerical workers (phoa, 1999; thomas, 1999). however, the two cpis are highly correlated (�0.99) and almost identical in magnitude, so this difference is inconsequential. second, tips adjust for inflation daily and use a nonseasonally adjusted cpi. within a year, tips inflation accrual can vary dramatically and affect yields (kan, 1999). since social security does not rely on seasonal adjustments, using tips yields—with their seasonalities—can lead to slight mispricing. in our valuation, the impact is negligible. seasonal adjustments are greatest november-march; however, kan (1999) shows the average seasonal/nonseasonal difference is only one basis point in june, our valuation month. another caveat involves taxes. all coupon payments on tips are federally taxable, but, at most, 85% of social security payments are taxable. this difference is economically insignificant for this analysis. it has less than 3% impact on our valuation of social security benefits, so we ignore this difference. however, tips incur an additional tax on the cpi-driven increase in principal amount. each year, as the principal value increases with inflation, the increase is taxable at the ordinary income tax rate. there is no similar tax on the increase in future social security benefits. this has three consequences for our analysis. first, tips yield may be higher if investors are compensated for this tax. if the yield is higher, then using tips undervalues social security wealth. adjusting for this difference would only heighten the portfolio impact of including social security as a portfolio asset. second, tips interest is exempt from state income tax. social security benefits are not necessarily exempt, but many states do have significant exemptions or deductions for pension income. third, this taxation substantially limits the advantages of inflation-indexed bonds for (taxable) individuals under mean-variance optimization (kinney, 1999) but does not affect social security similarly. a final caveat to the tips proxy for social security concerns duration. tips and social security with the same maturity (i.e., bond maturity equal to life expectancy) have different sensitivities to interest rate changes as measured by macaulay duration because tips have a large terminal cash flow and social security does not. because we focus on point estimates of social security wealth, we ignore the duration issue. we can safely do so because duration has minimal consequences on tips pricing. the tips yield curve is quite flat; bonds with different coupons but similar maturity have similar real yields (as of june 2000). 3.3. valuation detail assets are worth the present value of expected future cash flows. we have major valuation components described (i.e., social security benefits described and a good discount rate proxy) but must make a few more assumptions. valuing pensions requires estimates about coverage levels, work life, job tenure, age, mortality, discount rates, taxation, salary in300 s.p. fraser et al. / financial services review 9 (2000) 295–326 creases and inflation. this present value approach is the micro equivalent of macrovaluations of gross national social security wealth in the literature (feldstein, 1974). even the micro approach is not novel. reichenstein (1998, 2000) argues for including the present value of social security benefits in asset mix decisions, but does not quantify it because of social security’s complexities. scott (1995) argues for inclusion of social security wealth as a fixed income component of the total portfolio, but in discussing pensions is ambiguous in her prescription. first, she relates social security to defined benefit pension plans, but there suggests incorporating the value of current payments. second, she argues for excluding distant and uncertain cash flows. one interpretation of these two points is that social security is a portfolio asset only for retirees. third, she offers limited guidance in selecting a discount rate. her pension example uses nominal treasury yields, but this ignores social security’s inflation adjustment as well as risk adjustments (e.g., default risk in her corporate pension). we build on reichenstein’s (1998, 2000) and scott’s (1995) valuable insights. before going into details, we should clarify the appropriate valuation framework. social security wealth is not liquid, that is, it cannot be sold or explicitly be used as collateral. lack of marketability affects value. this is the essence of the valuation framework that applies a marketability discount to value an illiquid asset. since the marketability discount increases with the difficulty of selling an asset, the “fair market value” of social security wealth under this paradigm is zero. (note, however, that if social security wealth could be sold, its near-certain cash flows would minimize any marketability discount (see pratt, reilly & schweihs, 1996, p. 358).) rather than fair market value, the individual-specific valuation is relevant to an individual’s asset allocation decision. this is investment value, or “value to a particular investor based on individual investment requirements” (pratt et al., 1996). using investment value is valid because even nonsalable assets have individual-specific value. we describe our specific methodology and assumptions in the subsections below. in turn, we discuss the timing of social security cash flows, the size of social security cash flows, discounting the cash flows and our valuation results. 3.3.1. timing of cash flows several cash flow timing issues are relevant to our valuation. we discuss retirement timing, age and mortality in the following paragraphs. social security includes a retirement-timing option. reduced benefits are available 3–5 years before the statutory full retirement ages of 65–67. we focus our primary analysis on those who take or took social security at age 62 (but consider age 65 in the appendix). we do this for two reasons: first, detweiler (1999) evaluates the early retirement decision; he calculates present values and concludes that retired men should start social security at age 62 when real returns are more than 2.25% and that retired women should do so for real returns above 4.5%. second and more importantly, most people start benefits at 62 (congress, 1996). we calculate the present value of social security benefits for a wide range of ages as of 2000. this date is relevant since we incorporate the planned increases in statutory “full retirement” ages. that is, the present value reflects the early retirement penalty, which varies 301s.p. fraser et al. / financial services review 9 (2000) 295–326 according to birth year (thomas, 1999). we assume spouses are the same ages (but consider alternatives in the appendix). we use irs life expectancy tables (irs publication 939, 1997), which force two simplifying assumptions. first, the primary irs tables do not distinguish between men and women—death does. second, we ignore the correlation between health and wealth (smith & kington, 1997). the high earner we evaluate below is more likely to have the resources to pay for quality health care and nutrition (evensky, 1997). 3.3.2. size of cash flows in our valuation, we focus on high-income individuals. we do so without much loss of generality. social security benefits are increasing in earned income; further, savings are likely increasing in income. that is, both social security wealth and financial portfolio value are likely correlated with earnings. although the exact quantities will differ, qualitative conclusions drawn from high earners are relevant across most income classes. in support of this view, we consider a lower-income case in the appendix. we assume our worker earns $76,200 in 2000—the maximum amount credited to a social security account. we assume the worker earns the inflation-adjusted equivalent of $76,200 in all years before retirement. this has several consequences. first, if earned in the topearning thirty-five years of working lifetime, this qualifies for maximum retirement benefits. recall that maximum monthly benefits are $1,433 in 2000. we assume retirees qualify for this maximum pia. second, this means the worker never receives any real wage increases. third, we use this earnings stream to compute financial portfolio values when we discuss asset allocation implications. as we detail there, these last two consequences are conservative—they reduce social security’s impact on portfolios. recall that income earned (above certain thresholds) while receiving social security reduces benefits. we avoid this complication by assuming the worker truly retires at 62 and has no earned income. (how work affects your benefits, a social security administration publication (2000), says, “we do not count nonwork income such as investment earnings, interest, pensions, annuities, capital gains, and other government benefits” for this earnings test.) we assume the high-earner retiree receives sufficient unearned income from other sources to force the maximum 85% taxation of social security benefits. that is, the retiree’s annual income from all sources exceeds $34,000 if single or $44,000 if married. this assumption makes social security benefits very similar to other income with respect to taxes and forces a lower value of social security benefits. if a retiree had lower income from other sources, this would shield some social security benefits from taxes. this higher tax shield would necessitate decreasing the discount rate on social security benefits to keep it equivalent to taxable tips yields; such a lower discount rate would increase the value of social security wealth. lower social security values are conservative with respect to our conclusion that social security has dramatic portfolio implications. we compute the present value of these cash flows—on a pretax basis. this pretax basis allows direct comparability of social security wealth to other portfolio assets. while reichenstein (1998, 2000) argues (appropriately, in our view) for after-tax valuation of portfolio assets, pretax valuation is much more common. pretax valuation also avoids the complexities of social security taxation (thomas, 1999). 302 s.p. fraser et al. / financial services review 9 (2000) 295–326 in summary, we discount the maximum pia (adjusted for early retirement) over the timeline discussed. this maximum pia is constant in real, not nominal, dollars. for married workers, we assume the spouse never worked; the spouse receives benefits based on the worker’s income, not his/her own. we add the 50% spousal benefit over the same timeline used for single retirees. finally, we then replace the worker-plus-spousal 150% with the survivor’s [either worker’s or widow(er)’s] 100% for the years the joint life expectancy exceeds the single life expectancy. 3.3.3. discounting the cash flows we discount the constant real cash flows using a 3.95% real yield to maturity on inflation-indexed treasury securities from june 2000; we do so based on the similarities between tips and social security that we noted earlier. our approach is not the first to use real rates to value social security benefits. newmark and walden (1995) suggest a 1%–3% range is appropriate for most individuals. steuerle (1996) uses a 2% rate. neither, however, used market-based real rates. our market-based rates provide an unbiased valuation. in the remainder of this subsection, we consider alternate discount rate approaches. the discount rate for valuing social security benefits should be based only on expected future cash flows. it should not be a function of imputed returns on past contributions. many find the imputed returns on social security are very low or negative (see steuerle & bakija, 1994). leimer (1995) and munnell (1998) calculate them to be less than 2%. further, investment in social security is not a choice. if it were, these imputed returns would be more relevant to our valuation and asset mix analysis. our approach relies on the market yield-to-maturity curve rather than attempting to develop a theoretical spot rate curve to discount each social security payment at an exact rate. using market yields is not problematic—the tips yield curve is very flat, so the inaccuracy is slight. we do not make risk adjustments to the discount rate; we rely on the risk-free tips return. instead, we incorporate risk via the options approach discussed in section 3.4. 3.3.4. results of discounted cash flow table 1 presents our estimates of the net present value of social security benefits. it relies on the formula described above and detailed here. valuess � 1 �1 � tips yield 12 �months until retirement � �1 � �1 � tips yield 12 ��months in retirement� tips yield 12 � benefit (1) it reflects our assumptions about coverage levels, work life, job tenure, age, mortality, discount rates, taxation, salary levels, and inflation. (it does not reflect the risk-adjustment 303s.p. fraser et al. / financial services review 9 (2000) 295–326 described in section 3.4 below.) note that social security wealth for couples is about two-thirds greater than for individuals. roughly three-quarters of the increase is due to the spousal benefit with the remainder coming from the widow(er)’s benefit. for both couples and individuals, social security wealth increases until retirement and then declines. this pattern is similar to that expected of a financial portfolio. consider a simple example from table 1—social security wealth in 2000 for a single sixty-two-year old. the full retirement age for those born in 1938 is 65 years and 2 months; accordingly, she is retiring 38 months early and are only entitled to a monthly benefit 79.17% of the maximum $1,433 pia, or $1,134. the 79.17% reflects 5/9ths of 1% reduction for the first 36 months and 5/12ths of 1% reduction for the two extra months. according to the irs unisex life-expectancy tables this sixty-two-year old can expect to live 22.5 years and will spend 270 months in retirement. using eq. (1) and the 3.95% real tips yield, the present value of social security wealth is $202,731. while monthly benefits will increase in nominal terms from the $1,134, they stay constant in real terms and are correctly valued with the real tips yield. now add two changes—marriage and valuation before retirement; consider social security wealth in 2000 for married fifty-year olds. the full retirement age for those born in 1950 is 66 years; so for a fifty-year-old retiring at 62, they are retiring 48 months early and are only entitled to a monthly benefit 75% of the maximum $1,433 pia, or $1,075. note that we use the same $1,433 pia; while this amount will increase in nominal terms over the next twelve table 1 estimates of high-earner social security wealth age single married 25 $ 38,645 $ 65,636 30 $ 47,376 $ 80,096 35 $ 57,889 $ 97,790 40 $ 70,958 $119,330 45 $ 92,146 $154,504 50 $114,898 $191,272 55 $142,046 $234,679 60 $181,883 $298,157 62 $202,731 $330,863 65 $190,006 $318,356 70 $162,964 $275,189 75 $135,538 $234,351 80 $108,821 $192,451 85 $ 82,967 $151,247 estimated present value of social security benefits given assumptions listed below and explained in the text. coverage maximum pia ($1,433 in 2000), adjusted for early retirement penalties wages social security maximum ($76,200 in 2000), adjusted for inflation job tenure from age 21 to 61 retirement age 62 mortality based on gender-neutral irs actuarial tables discount rate tips yield to maturity (3.95% in june 1999) spouse non-working, entitled to 50% of worker pia (as spouse) or 100% of worker pia (as widow(er)) 304 s.p. fraser et al. / financial services review 9 (2000) 295–326 years before retirement, it is constant in real terms (and, therefore, appropriately discounted with the tips yield). according to the irs life-expectancy tables, a single fifty-year old life expectancy is 33.1 years and a joint life expectancy is 39.2 years. the couple can expect 21.1 years of 150% retiree-plus-spouse benefits and an additional 6.1 years of 100% retiree or widow(er) benefits. think of this as equating to two components—a $1,075 benefit starting in 144 months that lasts 326.4 months and an additional $537 benefit starting in 144 months that lasts 253.2 months. using eq. (1) and the 3.95% tips yield, the present value of the $1,075 benefit is $133,823; the present value of the $537 benefit is $57,449. together these values are the $191,272 shown in table 1. 3.4. risk adjustment: the binomial options overlay the present values of social security benefits are calculated from constant real cash flows (if alive and social security operates as planned); they do not specifically include risk adjustments to the discount rate. to incorporate risk, we prefer to overlay a binomial-option valuation approach because it allows cleaner specification of assumptions. for example, rather than estimating a discount rate adjustment for the vagaries of potential benefit changes, we use probabilities to capture the same effect. our recommended approach includes a probability of living long enough to start benefits at age 62. it also includes a probability of social security continuing to exist. conversely, social security not existing is equivalent to a 100% benefit reduction. this simple approach encompasses a range of other possibilities; for example, a 5% probability of zero social security benefits is equivalent to a 10% probability of benefits being reduced 50%. fig. 1 portrays our approach graphically and eq. fig. 1. binomial options evaluation tree. 305s.p. fraser et al. / financial services review 9 (2000) 295–326 (2) does so mathematically. pvrisky � p�qpvriskfree � �1�q� pvnoss� � �1�p� pvdead � pqpvriskfree (2) where pvrisky � risk-adjusted and survival-adjusted present value, p � probability of survival to age 62 of the beneficiary (and spouse), q � probability of survival for the social security system, pvriskfree � net present value computed above using eq. (1) with the risk-free rate, pvnoss � the present value without social security, and pvdead � the present value when dead. both pvnoss and pvdead are zero. the risk-adjusted present value simplifies to pqpvriskfree. for example, if there is a 95% chance of survival of both the beneficiary and the social security system, the risk-adjusted present value is 90.25% of the net present value computed in table 1. in our valuation, we assume—for simplicity—this value is 100% of the unadjusted present value (i.e., pq � 100%). given the demographic trends influencing social security, this assumption is debatable, but quantifying the probability of survival of the social security system is beyond the scope of this paper. those more pessimistic should use a lower percentage when applying this binomial options approach. the binomial options overlay approach to uncertainty about social security’s future is a simple and tractable alternative to adjusting the tips yield with a risk premium. for example, reconsider the fifty-year-old high-earning couple; under the assumptions used in this paper (and as we demonstrated in section 3.3), their present value of social security benefits is $191,272. a 2% risk premium lowers the present value of social security benefits by $68,135. this is the same as assuming the probability of social security not surviving is 35.6%. similarly, a 1% risk premium is equivalent to a 20.0% probability of no social security. we suspect that most consumers of financial planning advice (and many financial professionals) will find the probability approach in the options overlay more intuitive than the risk premium. 4. strategic asset allocation implications having computed the present value of social security benefits, we can investigate the effect of social security wealth on the asset mix decision. we apply four asset-mix decision rules—simple fixed mixes, life-cycle rules, consensus expert views and mean-variance optimization. in each case, the impact is dramatic. before continuing, we should clarify our terminology. the “financial portfolio” is the classic focus of the investment manager—stocks, bonds, and so forth. the “total portfolio” includes all relevant assets discussed in scott (1997) and reichenstein (1998, 2000). somewhere between these portfolio concepts is an “expanded portfolio” that is the financial portfolio plus social security wealth. our analysis focuses only on the expanded portfolio. while doing so is incomplete from a total portfolio viewpoint, this approach isolates social security’s impact. 306 s.p. fraser et al. / financial services review 9 (2000) 295–326 to demonstrate the impact of social security wealth on portfolios, we must first quantify the portfolios affected. note that social security wealth is measured in dollars, but asset-mix decision rules are expressed in percentages. to bridge this gap, we compute four realistic, hypothetical portfolios. the first three hypothetical portfolios are based on savings rates of 5, 10 and 15% of income per year over a working lifetime (age 21–61). ten-percent and fifteen-percent savings rules are common advice (clements, 1998) and apply to wages earned. recall that we assumed earnings of $76,200 in 2000, adjusted this amount for inflation, and assumed it is earned in all years before retirement. (this unrealistically assumes maximum earnings at age 21 and proportional lifelong savings. this is both a simplifying and a conservative assumption. high early savings and lifelong savings both increase the financial portfolio, which diminishes the impact of social security wealth on the expanded portfolio.) savings grow at a conservative rate of 6.5%. for retirees, we assume savings stopped at retirement and spending begins based on a rolling 20-year 6.5% annuity; we update the payout each year to include a new 20-year remaining life. this avoids relying on a specific insurer’s life annuity payout rate, but our approach’s average payout approximates several annuities we did examine. the fourth hypothetical portfolio uses the formula from the millionaire next door (mnd) bestseller. stanley and danko (1996, p. 13) suggest an investor is a “prodigious accumulator of wealth” if the following equation holds: net worth �� age � income 10 (3) we assume our fourth portfolio equal to this value through the beginning of retirement. for example, our fifty-year-old high-earner has a financial portfolio of $381,000 [50 $76,200 10]. 4.1. four asset allocation strategies the simplest strategic asset allocation that we consider is the classic fixed mix of 60% stocks and 40% bonds (maginn & tuttle, 1990). the standard interpretation of this and other asset allocation rules is that they apply to financial portfolios only, not more inclusive portfolios. the 60/40 fixed mix, while simple, does not recognize life-cycle changes. wealth, risk aversion and investment horizon all change over a life cycle. in response, bogle (1994) and others advocate a simple rule of thumb where a portfolio’s bond weight (in percentage) should be the investor’s age. a 40 year old should be 60% in stock, and a 60 year old should be 40% in stock. a consensus of many investment experts offers potential enhancements to both the classic 60/40 mix and the simple life-cycle mix. malkiel (1996) offers “life-cycle. . .savings allocations” which vary with an investor’s age and correspond with an investor’s tolerance for risk. his guide provides four sample portfolios ranging from a 70/30 split for the midtwenties age group to a 30/70 mix for those in the late-sixties-and-beyond group. malkiel’s portfolios are reasonable guides to this class of asset-mix decision rules. reichenstein (1996) 307s.p. fraser et al. / financial services review 9 (2000) 295–326 shows consensus among asset allocation strategies from several sources including malkiel; fisher and statman (1997) also illustrate commonalities across expert asset allocations. the fourth strategic asset allocation approach we consider is mean-variance optimization. under mean-variance optimization, we examine portfolio implications for a range of risk aversion levels. the four asset allocation strategies are summarized in table 2. 4.2. implications for the classic 60/40 mix table 3 treats social security wealth as a bond and shows expanded portfolio asset mixes when the financial portfolio is a 60/40 fixed mix. a fifty-year-old 5% saver who thinks he owns more stocks than bonds (the 60/40 mix) actually only owns half as much stocks as bonds in his expanded portfolio (35% stocks/65% bonds) when social security wealth is included. that is, holding 60% stocks and 40% bonds in the financial portfolio really equates to holding 35% stocks and 65% bonds (actual bonds plus social security wealth) in the expanded portfolio. we rely on the following formula: % stocksexpanded portfolio � % stocksfinancial portfolio � wealthfinancial portfolio wealthfinancial portfolio � wealthsocial security (4) fifty-year-old couples, with more social security wealth, are even more underweighted in stocks. better savers, with larger financial portfolios, obviously have a larger stock allocation in their expanded portfolio, but none of the four types of savers has even a 50% stock allocation. in no case are seemingly equity-heavy portfolios even majority-equity portfolios. because social security wealth grows at the same time as the financial portfolio grows, these patterns are relatively consistent over a lifetime. fig. 2 shows the resultant expanded portfolio asset mix for a fifty-year-old 10%-saver couple when the financial portfolio is a 60/40 fixed mix. the discussion above is descriptive—it describes the real consequences of applying the fixed mix to a financial portfolio. what follows is prescriptive—it describes how to implement a financial portfolio in order to achieve a desired expanded portfolio. table 4 shows the required asset mixes in the financial portfolio to obtain a 60/40 fixed mix in the expanded portfolio; that is, table 4 shows how to implement a 60/40 fixed mix in the expanded portfolio by adjusting the financial portfolio. we use the following formula: table 2 summary of the four asset allocation strategies asset-mix decision rule description citation fixed mix hold 60% stocks and 40% bonds. maginn and tuttle (1990) life cycle mix hold (1-age)% in stocks and (age)% in bonds. bogle (1994) consensus expert mix hold stocks and bonds in accordance with malkiel’s “life-cycle savings allocations.” malkiel (1996) mean-variance optimization hold stocks and bonds in proportions found on the mean-variance efficient frontier. markowitz (1959) 308 s.p. fraser et al. / financial services review 9 (2000) 295–326 % stocksfinancial portfolio � % stocksexpanded portfolio � �wealthfinancial portfolio � wealthsocial security) wealthfinancial portfolio (5) again, consider a single fifty-year-old 5% saver; he or she should be very slightly leveraged in her financial portfolio. this slight leverage (103% equities) is equivalent to margin borrowing. the fifty-year-old couple should be somewhat more heavily leveraged (131% in equity). margin borrowing is only required for the worst savers among our four types of fifty year olds; a fifty-year-old 10% saver should have 81% stock in her financial portfolio if single or 95% if married. fig. 3 shows the required financial portfolio asset mix for a fifty-year-old 10%-saver couple when the desired expanded portfolio is a 60/40 fixed mix. as one can see in table 4, social security wealth has most significant impact for those just beginning to save (youngest), those no longer saving (oldest), and those that save the least. all of the leveraged asset mixes described in this paper are feasible. purchasing $100 of stock on margin with $50 equity is equivalent to a portfolio with 200% stock/-100% bonds if one classifies the margin debt, a fixed obligation, as a (short) bond. any asset allocation with the bond weighting greater than or equal to �100% is possible with ordinary margin accounts. stock index futures offer even greater leverage potential. we considered margined table 3 implications of a 60/40 mix in the financial portfolio for the expanded portfolio demography perceived financial portfolio resultant expanded portfolio including social security wealth 5% saver 10% saver 15% saver mnd saver age status stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds 30 single 60% 40% 29% 71% 40% 60% 45% 55% 50% 50% married 60% 40% 22% 78% 32% 68% 38% 62% 44% 56% 40 single 60% 40% 36% 64% 45% 55% 49% 51% 49% 51% married 60% 40% 28% 72% 38% 62% 44% 56% 43% 57% 50 single 60% 40% 35% 65% 44% 56% 49% 51% 46% 54% married 60% 40% 28% 72% 38% 62% 43% 57% 40% 60% 60 single 60% 40% 33% 67% 43% 57% 47% 53% 43% 57% married 60% 40% 26% 74% 36% 64% 42% 58% 36% 64% 62 single 60% 40% 31% 69% 41% 59% 46% 54% 42% 58% married 60% 40% 24% 76% 34% 66% 40% 60% 35% 65% 65 single 60% 40% 29% 71% 40% 60% 45% 55% 41% 59% married 60% 40% 22% 78% 32% 68% 38% 62% 34% 66% 70 single 60% 40% 27% 73% 38% 62% 43% 57% 40% 60% married 60% 40% 20% 80% 30% 70% 36% 64% 33% 67% 80 single 60% 40% 24% 76% 35% 65% 40% 60% 38% 62% married 60% 40% 17% 83% 26% 74% 32% 68% 29% 71% expanded portfolio stock/bond mix resulting from adding social security wealth to a financial portfolio with a 60/40 fixed mix. social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). 309s.p. fraser et al. / financial services review 9 (2000) 295–326 positions rather than constrained allocations in order to highlight the dramatic impact of incorporating social security wealth in portfolios. to demonstrate the calculations behind tables 3 and 4, consider a hypothetical investor with a $200,000 financial portfolio invested in a 60/40 mix. that is, $120,000 is in stocks and $80,000 is in bonds. assume social security wealth is valued at $100,000. the $120,000 in stocks is 40% of the $300,000 expanded portfolio. to achieve a 60/40 mix in the expanded portfolio, 60% of $300,000, or $180,000, must be invested in stocks. this value would represent 90% of the $200,000 financial portfolio. the values in this example also work with eqs. (4) and (5). 4.3. implications for the simple life-cycle mix if our fifty-year-old investor follows our second asset allocation strategy, the simple life-cycle mix, it indicates a 50% position in bonds. for the 5% saver, the target 50/50 (stock/bond) mix in the financial portfolio results in an expanded portfolio mix of 29/71 if single and 23/77 if married. see table 5. even though the life-cycle mix dictates a majority-equity position for those younger than fifty years old (because the bond weight fig. 2. implications of social security wealth on asset allocation. 310 s.p. fraser et al. / financial services review 9 (2000) 295–326 equals the investor’s age), only the youngest, best savers have such a majority-equity position in their expanded portfolios. table 6 shows how to implement a life-cycle mix in the expanded portfolio by adjusting the financial portfolio. if she really desired a 50/50 mix in her expanded portfolio, the required mix in the fifty-year-old investor’s financial portfolio is 85/15 if single or a margined 109/-9 if married. to offset the bond-like value of social security benefits, all investors, except the oldest, need at least a majority-equity position in their financial portfolio. 4.4. implications for the consensus expert mix table 7 consolidates the expert’s views and describes the impact of social security wealth on the expanded portfolio asset mix. we make two simplifications—we combine cash with bonds and we ignore the within-class allocations. for our same fifty-year-old 5% saver, the consensus expert target of 53/47 (interpolated from malkiel, 1996) yields an actual expanded portfolio asset mix of 31/69. again, only the youngest and best savers following the asset-mix decision rule have more than a majority-equity position. table 8 shows how to implement the consensus expert mix in the expanded portfolio by adjusting the financial portfolio. to achieve the desired 53/47 split in the expanded portfolio would require a 91/9 table 4 implications of a 60/40 mix in the expanded portfolio for the financial portfolio demography desired expanded portfolio resultant financial portfolio 5% saver 10% saver 15% saver mnd saver age status stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds 30 single 60% 40% 122% �22% 91% 9% 81% 19% 72% 28% married 60% 40% 165% �65% 112% �12% 95% 5% 81% 19% 40 single 60% 40% 100% 0% 80% 20% 73% 27% 74% 26% married 60% 40% 128% �28% 94% 6% 83% 17% 83% 17% 50 single 60% 40% 103% �3% 81% 19% 74% 26% 78% 22% married 60% 40% 131% �31% 95% 5% 84% 16% 90% 10% 60 single 60% 40% 108% �8% 84% 16% 76% 24% 84% 16% married 60% 40% 139% �39% 99% 1% 86% 14% 99% 1% 62 single 60% 40% 115% �15% 88% 12% 78% 22% 86% 14% married 60% 40% 150% �50% 105% �5% 90% 10% 102% �2% 65 single 60% 40% 122% �22% 91% 9% 81% 19% 87% 13% married 60% 40% 164% �64% 112% �12% 95% 5% 105% �5% 70 single 60% 40% 131% �31% 96% 4% 84% 16% 90% 10% married 60% 40% 181% �81% 120% �20% 100% 0% 110% �10% 80 single 60% 40% 148% �48% 104% �4% 89% 11% 96% 4% married 60% 40% 215% �115% 137% �37% 112% �12% 123% �23% required financial portfolio asset allocation necessary to achieve a 60/40 fixed mix in the expanded portfolio, which includes social security wealth. social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). stock positions greater than 100% are purchased on margin; those greater than 200% require stock futures to implement. negative bond positions represent margin borrowing. 311s.p. fraser et al. / financial services review 9 (2000) 295–326 mix in the financial portfolio. again, to offset social security’s impact on the financial portfolio, most investors need more than 50% stock in their financial portfolio. while these results are similar to the life-cycle approach, we include them to demonstrate the veracity of our result that including social security wealth has a profound impact on strategic asset allocation—whatever the asset-mix decision rule. 4.5. implications for optimization we have shown social security wealth has a clear impact on the portfolio asset mix under three common decision rules. here we discuss its impact under the more sophisticated markowitz (1959) approach that calculates the efficient frontier of minimum risk portfolios for a given return level—mean-variance optimization (mvo). as before, we can state that excluding social security wealth will greatly affect an individual’s optimal portfolio. many researchers (siegel, 1998; bodie, 1990; brynjolfsson & fabozzi, 1999; chen & terrien, 1999; dalio & bernstein, 1999; lamm, 1998; phoa, 1999; rudolph-shabinsky, 2000) find that adding tips to stock/bond portfolios enhances portfolio efficiency and that tips displace ordinary bonds along much of the efficient frontier. tips shift the meanvariance efficient frontier because of their impact on the optimization inputs—expected means, variance and correlations. we consider these in turn. first, the expected return on tips is low when compared to stocks and bonds in nominal terms, but they compare favorably in real terms. second, the variance of tips is lower than other asset classes. while fig. 3. adjusting the financial portfolio to reflect social security wealth. 312 s.p. fraser et al. / financial services review 9 (2000) 295–326 ordinary bonds’ variance reflects changes in both real yields and inflation expectations, tips’s variance only reflects changes in real yields. dalio and bernstein (1999) estimate this is responsible for only 30% of bonds’ overall variance. third, tips have an unusually low correlation with other asset classes. tips have low correlation with nominal bonds because of their opposite reaction to inflation. tips have low correlation with stocks because stocks are empirically poor inflation hedges. for long horizons, dalio and bernstein (1999) estimate tips-nominal bond correlation at �0.01 and tips-stock correlation at �0.24. brynjolfsson and rennie (1999), chen and terrien (1999) and phoa (1999) also estimate negative or very low correlations. social security, like tips, affects the optimal portfolio through mvo inputs—expected means, variance and correlations. in section 3.3, we related social security returns to tips returns. variance on social security wealth, like tips, is low and reflects only changes in real yields (but social security variance could be even lower since it excludes tradinginduced noise). likewise, tips correlation is a reasonable guide to social security correlation. social security has low correlation with bonds because of its opposite relation to inflation. for most workers, social security wealth is essentially unrelated to stock returns; low or negative correlations for social security wealth are appropriate. since social security behaves much like tips with respect to optimization inputs, social security wealth will also table 5 implications of a life-cycle mix in the financial portfolio for the expanded portfolio demography perceived financial portfolio resultant expanded portfolio including social security wealth 5% saver 10% saver 15% saver mnd saver age status stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds 30 single 70% 30% 34% 66% 46% 54% 52% 48% 58% 42% married 70% 30% 25% 75% 37% 63% 44% 56% 52% 48% 40 single 60% 40% 36% 64% 45% 55% 49% 51% 49% 51% married 60% 40% 28% 72% 38% 62% 44% 56% 43% 57% 50 single 50% 50% 29% 71% 37% 63% 40% 60% 38% 62% married 50% 50% 23% 77% 31% 69% 36% 64% 33% 67% 60 single 40% 60% 22% 78% 29% 71% 32% 68% 29% 71% married 40% 60% 17% 83% 24% 76% 28% 72% 24% 76% 62 single 38% 62% 20% 80% 26% 74% 29% 71% 27% 73% married 38% 62% 15% 85% 22% 78% 25% 75% 22% 78% 65 single 35% 65% 17% 83% 23% 77% 26% 74% 24% 76% married 35% 65% 13% 87% 19% 81% 22% 78% 20% 80% 70 single 30% 70% 14% 86% 19% 81% 21% 79% 20% 80% married 30% 70% 10% 90% 15% 85% 18% 82% 16% 84% 80 single 20% 80% 8% 92% 12% 88% 13% 87% 13% 87% married 20% 80% 6% 94% 9% 91% 11% 89% 10% 90% expanded portfolio stock/bond mix resulting from adding social security wealth to a financial portfolio with a life-cycle asset allocation. the life-cycle mix invests a percentage of the portfolio in bonds equal to the investor’s age. social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). 313s.p. fraser et al. / financial services review 9 (2000) 295–326 enhance portfolio efficiency and be part of optimal portfolios. see bajtelsmit and turner (1998). in tables 9 and 10, we demonstrate a particular implementation of mean-variance optimization (mvo) with social security. tables 3 through 8 (which relate to the other three asset-mix decision rules) vary in three dimensions—age, savings rates and marital status. mvo adds many more dimensions of variability including risk aversion, means, variances and correlations. even a small-scale optimization with three assets (bonds, stocks and social security wealth) adds 10 degrees of freedom (3 expected returns, 3 standard deviations, 3 correlations and risk aversion). rather than attempt to build tables comparable to tables 3 through 8 across 13 dimensions, we illustrate how social security wealth affects optimal strategic asset allocation in a scenario and leave it to the reader to conduct mvo with inputs aligned with their beliefs. we first compute the mvo frontier with two assets—stocks and bonds. we use expected returns of 10.5% and 6.5%, respectively. the less-than-historical equity risk premium is conservative (cornell, 1999). we use a 0.6 correlation between stocks and bonds and standard deviations of 20% and 9%, respectively. from this two-asset frontier, we impute table 6 implications of a life-cycle mix in the expanded portfolio for the financial portfolio demography desired expanded portfolio required financial portfolio 5% saver 10% saver 15% saver mnd saver age status stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds 30 single 70% 30% 142% �42% 106% �6% 94% 6% 85% 15% married 70% 30% 192% �92% 131% �31% 111% �11% 95% 5% 40 single 60% 40% 100% 0% 80% 20% 73% 27% 74% 26% married 60% 40% 128% �28% 94% 6% 83% 17% 83% 17% 50 single 50% 50% 85% 15% 68% 32% 62% 38% 65% 35% married 50% 50% 109% �9% 79% 21% 70% 30% 75% 25% 60 single 40% 60% 72% 28% 56% 44% 51% 49% 56% 44% married 40% 60% 93% 7% 66% 34% 58% 42% 66% 34% 62 single 38% 62% 73% 27% 55% 45% 50% 50% 54% 46% married 38% 62% 95% 5% 66% 34% 57% 43% 65% 35% 65 single 35% 65% 71% 29% 53% 47% 47% 53% 51% 49% married 35% 65% 96% 4% 65% 35% 55% 45% 62% 38% 70 single 30% 70% 66% 34% 48% 52% 42% 58% 45% 55% married 30% 70% 90% 10% 60% 40% 50% 50% 55% 45% 80 single 20% 80% 49% 51% 35% 65% 30% 70% 32% 68% married 20% 80% 72% 28% 46% 54% 37% 63% 41% 59% required financial portfolio asset allocation necessary to achieve a life-cycle asset allocation in the expanded portfolio, which includes social security wealth. the life-cycle mix invests a percentage of the portfolio in bonds equal to the investor’s age. social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). stock positions greater than 100% are purchased on margin; negative bond positions represent margin borrowing. 314 s.p. fraser et al. / financial services review 9 (2000) 295–326 risk-aversion coefficients, �, for 100/0, 90/10, 80/20, and so forth, portfolios. these riskaversion coefficients maximize utility under the standard framework: max up � e�rp� � �e��p� (6) higher � implies higher risk aversion. while the levels of the risk-aversion coefficients in tables 9 and 10 appear arbitrary, they are designed to map to “round number” asset mixes in the financial portfolio. we then apply these risk-aversion coefficients to mvo in expanded portfolios that include social security wealth. we consider single and married fifty-year-olds with different savings. we use the table 1 present value of social security benefits and the hypothetical financial portfolios for 5%, 10%, 15% and mnd savers to calculate the percentage of the expanded portfolio that social security wealth represents. we constrain the mvo to hold social security wealth in this constant proportion. unlike the other three asset-mix decision rules, we treat social security wealth as a distinct asset class—not an ordinary bond. we use a 6.45% expected return on social security that reflects 2.5% inflation and the 3.95% real tips yield used throughout this paper. we use a 3% standard deviation, �0.2 correlation with stocks and 0.0 correlation with nominal bonds. the optimization inputs are illustrative, not authoritative, but represent a harmonization of many sources (bodie, 1990; brynjolfsson & fabozzi, 1999; chen & table 7 implications of malkiel mix in the financial portfolio for the expanded portfolio demography perceived financial portfolio resultant expanded portfolio including social security wealth 5% saver 10% saver 15% saver mnd saver age status stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds 30 single 65% 35% 32% 68% 43% 57% 48% 52% 54% 46% married 65% 35% 24% 76% 35% 65% 41% 59% 48% 52% 40 single 60% 40% 36% 64% 45% 55% 49% 51% 49% 51% married 60% 40% 28% 72% 38% 62% 44% 56% 43% 57% 50 single 53% 47% 31% 69% 39% 61% 43% 57% 41% 59% married 53% 47% 24% 76% 33% 67% 38% 62% 35% 65% 60 single 43% 57% 24% 76% 31% 69% 34% 66% 31% 69% married 43% 57% 19% 81% 26% 74% 30% 70% 26% 74% 62 single 39% 61% 20% 80% 27% 73% 30% 70% 27% 73% married 39% 61% 16% 84% 22% 78% 26% 74% 23% 77% 65 single 37% 63% 18% 82% 24% 76% 28% 72% 25% 75% married 37% 63% 14% 86% 20% 80% 23% 77% 21% 79% 70 single 30% 70% 14% 86% 19% 81% 21% 79% 20% 80% married 30% 70% 10% 90% 15% 85% 18% 82% 16% 84% 80 single 30% 70% 12% 88% 17% 83% 20% 80% 19% 81% married 30% 70% 8% 92% 13% 87% 16% 84% 15% 85% expanded portfolio stock/bond mix resulting from adding social security wealth to a financial portfolio with a consensus-expert asset allocation. the consensus-expert mix is per malkiel (1996). social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). 315s.p. fraser et al. / financial services review 9 (2000) 295–326 terrien, 1999; dalio & bernstein, 1999; ibbotson associates, 1999; lamm, 1998; phoa, 1999; rudolph-shabinsky, 2000). for simplicity, we require asset holdings to be in whole percents and constrain stock positions in the expanded portfolio to 0% to 100%. table 9 shows the mvo efficient portfolios for the different imputed risk aversion levels. for example, a single 10%-saver 50-year-old whose risk aversion recommends a 90%/10% stock/bond mix if optimizing only her financial portfolio (i.e., � � 0.279) would choose to hold a 69%/5%/26% stock/bond/social security mix in her expanded portfolio. table 10 shows this is requires a 93%/7% stock/bond mix in the financial portfolio to implement. as before, integrating social security wealth under this asset-mix decision rule requires more stockholdings; however, note that the required financial portfolios in table 10 require only slight adjustments (1–17% shifts) from the a priori financial portfolios with the same risk aversion. these small shifts are for constant risk aversions � (typically diagonal moves in risk-return space); integrating social security into the asset mix decision also allows i) risk reductions for constant expected returns (horizontal moves) or ii) return enhancements for constant risk levels (vertical moves). these horizontal and vertical moves in risk-return space require larger reallocations—akin to those under the first three asset-mix decision rules. fig. table 8 implications of malkiel mix in the expanded portfolio for the financial portfolio demography desired expanded portfolio required financial portfolio 5% saver 10% saver 15% saver mnd saver age status stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds 30 single 65% 35% 132% �32% 99% 1% 87% 13% 78% 22% married 65% 35% 179% �79% 122% �22% 103% �3% 88% 12% 40 single 60% 40% 100% 0% 80% 20% 73% 27% 74% 26% married 60% 40% 128% �28% 94% 6% 83% 17% 83% 17% 50 single 53% 47% 91% 9% 72% 28% 66% 34% 69% 31% married 53% 47% 115% �15% 84% 16% 74% 26% 80% 20% 60 single 43% 57% 77% 23% 60% 40% 54% 46% 60% 40% married 43% 57% 99% 1% 71% 29% 62% 38% 71% 29% 62 single 39% 61% 75% 25% 57% 43% 51% 49% 56% 44% married 39% 61% 97% 3% 68% 32% 58% 42% 66% 34% 65 single 37% 63% 75% 25% 56% 44% 50% 50% 54% 46% married 37% 63% 101% �1% 69% 31% 58% 42% 65% 35% 70 single 30% 70% 66% 34% 48% 52% 42% 58% 45% 55% married 30% 70% 90% 10% 60% 40% 50% 50% 55% 45% 80 single 30% 70% 74% 26% 52% 48% 45% 55% 48% 52% married 30% 70% 107% �7% 69% 31% 56% 44% 62% 38% required financial portfolio asset allocation necessary to achieve a consensus-expert asset allocation in the expanded portfolio, which includes social security wealth. the consensus-expert mix is per malkiel (1996). social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). stock positions greater than 100% are purchased on margin; negative bond positions represent margin borrowing. 316 s.p. fraser et al. / financial services review 9 (2000) 295–326 table 9. implications of mvo for the expanded portfolio including social security risk aversion actual financial portfolio resultant expanded portfolio including social security wealth 5% saver 10% saver 15% saver mnd saver status � stocks bonds stocks bonds soc sec stocks bonds soc sec stocks bonds soc sec stocks bonds soc sec single fifty-year-old 0.274 100% 0% 63% �4% 41% 76% �2% 26% 82% �1% 19% 78% �1% 23% 0.279 90% 10% 57% 2% 41% 69% 5% 26% 75% 6% 19% 71% 6% 23% 0.286 80% 20% 51% 8% 41% 62% 12% 26% 67% 14% 19% 64% 13% 23% 0.296 70% 30% 45% 14% 41% 54% 20% 26% 58% 23% 19% 56% 21% 23% 0.310 60% 40% 39% 20% 41% 47% 27% 26% 50% 31% 19% 48% 29% 23% 0.330 50% 50% 33% 26% 41% 39% 35% 26% 42% 39% 19% 40% 37% 23% 0.360 40% 60% 27% 32% 41% 32% 42% 26% 34% 47% 19% 33% 44% 23% 0.410 30% 70% 21% 38% 41% 24% 50% 26% 26% 55% 19% 25% 52% 23% 0.500 20% 80% 15% 44% 41% 17% 57% 26% 18% 63% 19% 17% 60% 23% 0.691 10% 90% 8% 51% 41% 9% 65% 26% 9% 72% 19% 9% 68% 23% 1.273 0% 100% 2% 57% 41% 2% 72% 26% 1% 80% 19% 2% 75% 23% actual financial 5% saver 10% saver 15% saver mnd saver stocks bonds stocks bonds soc sec stocks bonds soc sec stocks bonds soc sec stocks bonds soc sec married fifty-year-old 0.274 100% 0% 53% �7% 54% 66% �3% 37% 74% �2% 28% 70% �3% 33% 0.279 90% 10% 48% �2% 54% 60% 3% 37% 67% 5% 28% 64% 3% 33% 0.286 80% 20% 43% 3% 54% 54% 9% 37% 60% 12% 28% 57% 10% 33% 0.296 70% 30% 38% 8% 54% 47% 16% 37% 53% 19% 28% 50% 17% 33% 0.310 60% 40% 33% 13% 54% 41% 22% 37% 46% 26% 28% 43% 24% 33% 0.330 50% 50% 28% 18% 54% 35% 28% 37% 38% 34% 28% 36% 31% 33% 0.360 40% 60% 23% 23% 54% 28% 35% 37% 31% 41% 28% 29% 38% 33% 0.410 30% 70% 18% 28% 54% 22% 41% 37% 24% 48% 28% 22% 45% 33% 0.500 20% 80% 13% 33% 54% 15% 48% 37% 16% 56% 28% 16% 51% 33% 0.691 10% 90% 8% 38% 54% 9% 54% 37% 9% 63% 28% 9% 58% 33% 1.273 0% 100% 3% 43% 54% 2% 61% 37% 2% 70% 28% 2% 65% 33% expanded portfolio stock/bond/social security wealth mix resulting from adding social security wealth to a financial portfolio with a mean-variance optimal asset allocation for a given risk aversion level �. � is the risk aversion from the utility maximization problem e(r) � �e(�). social security wealth is treated as tips and uses values from table 1 computed using real tips yields for a fifty-year-old high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). 317 s.p . f raser et al./f inancial services r eview 9 (2000) 295–326 4, which shows the efficient frontiers with and without social security for single fifty-yearolds, illustrates these horizontaland vertical-shift possibilities. fig. 4 shows that for all four savings levels, social security wealth shifts the efficient frontier. this is consistent with the research showing tips increasing efficiency. the efficient frontiers for expanded portfolios are shorter because required holdings in social security wealth are displacing high-risk/high-return stock (which we limited to 100% of the expanded portfolio); the 5% saver’s efficient frontier is shortest because social security wealth is most significant in her expanded portfolio. leverage could extend the expanded portfolio frontiers. table 10 implications of mvo in the expanded portfolio for the financial portfolio risk aversion required financial portfolio 5% saver 10% saver 15% saver mnd saver status � stocks bonds stocks bonds stocks bonds stocks bonds single fifty-year-old 0.274 107% �7% 103% �3% 101% �1% 101% �1% 0.279 97% 3% 93% 7% 93% 7% 92% 8% 0.286 86% 14% 84% 16% 83% 17% 83% 17% 0.296 76% 24% 73% 27% 72% 28% 73% 27% 0.310 66% 34% 64% 36% 62% 38% 62% 38% 0.330 56% 44% 53% 47% 52% 48% 52% 48% 0.360 46% 54% 43% 57% 42% 58% 43% 57% 0.410 36% 64% 32% 68% 32% 68% 32% 68% 0.500 25% 75% 23% 77% 22% 78% 22% 78% 0.691 14% 86% 12% 88% 11% 89% 12% 88% 1.273 3% 97% 3% 97% 1% 99% 3% 97% married fifty-year-old 0.274 115% �15% 105% �5% 103% �3% 104% �4% 0.279 104% �4% 95% 5% 93% 7% 96% 4% 0.286 93% 7% 86% 14% 83% 17% 85% 15% 0.296 83% 17% 75% 25% 74% 26% 75% 25% 0.310 72% 28% 65% 35% 64% 36% 64% 36% 0.330 61% 39% 56% 44% 53% 47% 54% 46% 0.360 50% 50% 44% 56% 43% 57% 43% 57% 0.410 39% 61% 35% 65% 33% 67% 33% 67% 0.500 28% 72% 24% 76% 22% 78% 24% 76% 0.691 17% 83% 14% 86% 12% 88% 13% 87% 1.273 7% 93% 3% 97% 3% 97% 3% 97% required financial portfolio asset allocation necessary to achieve a mean-variance optimal asset allocation in the expanded portfolio, which includes social security wealth. risk aversion levels � are from the utility maximization problem e(r) � �e(�); specific values of � displayed in this table relate to asset allocations in table 9. expanded portfolio weights are computed from rounded values in table 9. social security wealth is treated as tips and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). stock positions greater than 100% are purchased on margin; negative bond positions represent margin borrowing. 318 s.p. fraser et al. / financial services review 9 (2000) 295–326 in short, integrating the present value of social security benefits shifts the mean-variance efficient frontier and offers substantial risk-reducing or return-enhancing opportunities. since social security wealth has risk-return-correlation characteristics much like an inflationindexed treasury bond, considering it in the total portfolio has a striking impact on the optimized financial portfolio. 5. conclusion 5.1. summary the typical income-adequacy retirement planning approach, which deducts social security benefits from required retirement cash flows, ignores asset mix implications of social security wealth. we assert that including social security cash flows but ignoring asset mix implications is inconsistent. this paper addresses three questions related to this inconsistency (should we include social security? what is its value? and what is its impact?). we first consider the qualitative and quantitative arguments that this is an inconsistency and that social security should be included in asset mix decisions. next, we highlight commonalities in the payment stream from social security benefits and tips and use those commonalities fig. 4. efficient frontiers with and without social security wealth for a single 50-year-old. 319s.p. fraser et al. / financial services review 9 (2000) 295–326 to value social security wealth. third, we show that under any of the asset-mix decision rules considered, social security wealth clearly affects strategic asset allocation. common interpretation of the four asset-mix decision rules we consider is that they apply to asset mixes in financial portfolios rather than expanded portfolios that incorporate social security wealth. we show that adding social security wealth to the financial portfolio results in a dramatically different asset mix in the expanded portfolio. for many, social security wealth inverts a seemingly stock-heavy portfolio into a bond-heavy one. conversely, we show that achieving a desired expanded portfolio mix requires significantly more stock holdings in the financial portfolio. for some, margined stock positions are required to offset the bond-like present value of social security benefits. while we consider a range of circumstances—age, marital status, savings-level and (in the appendix) gender, income-level and retirement age, we encourage applying our approach to individual circumstances. how, then, should an individual value their social security benefits for inclusion in a portfolio? first, we advise completing the form ssa-7004-sm, request for earnings and benefit estimate statement, which is available on www.ssa.gov. in completing the form, pay careful attention to the instructions about average future annual earnings—exclude any cost-of-living increases. (alternatively, www.ssa.gov/planners/calculators.htm offers three benefit calculators.) second, find the wall street journal yield quote for inflation-indexed treasury securities whose maturity approximates your years to retirement. third, estimate your life expectancy. irs publication 939 has actuarial tables, but you may want to deviate from its median values. use the yield and life expectancy to discount the estimated benefit. remember this is a two-stage process—compute the value of the expected benefit annuity as of your expected retirement date and then discount the value of that future lump sum to obtain today’s value. see eq. (1). 5.2. implications and extensions the key practical impact of this paper is to demonstrate that the present value of social security benefits is properly considered part of investors total portfolio and that excluding social security has a striking impact on their finances and investment decisions. a key contribution of this paper is providing a well-reasoned means of quantifying social security wealth. once a dollar value is assigned to social security, it can be properly integrated into a portfolio. rarely do asset-mix prescriptions incorporate marital status; none that we consider does so. some suggest that single investors can accept more risk (e.g., gutner, 2000). the present value of a couple’s social security annuity is more valuable than a single person’s annuity. if asset-mix decision rules apply to financial portfolios, couples are inherently implementing a less risky expanded portfolio; conversely, singles already have a more risky expanded portfolio—without adjusting the financial portfolio. if, instead, the asset-mix decision rules apply to expanded portfolios, couples require riskier, stock-heavy financial portfolios to offset their greater social security wealth. another implication of this paper is that question about the prospects for social security introduce a paradox. a lower probability of social security surviving implies social security wealth is lower. given a specific asset mix decision rule for the expanded portfolio, this 320 s.p. fraser et al. / financial services review 9 (2000) 295–326 implies a smaller equity allocation in the financial portfolio since the expanded portfolio has less fixed income. however, less social security wealth also reduces the size of the expanded portfolio—suggesting a need for more equity to grow the portfolio to meet future cash flow needs. we focus on first-order strategic asset allocation implications under asset mix decision rules. expanding the analysis to include a liability representing future cash flow needs is left to future work. putting aside questions about social security’s future, this analysis creates arguments for adding stock to the asset mix. for some investors, bonds are comforting—to the potential detriment of their portfolio. the similarity between inflation-indexed social security payments and inflation-indexed tips payments is simple and persuasive. the idea of an expanded portfolio (with comforting bond-like social security wealth) encourages and enables increased stockholdings to achieve target asset allocations. there are many other possible extensions to this investigation. we ignore a range of additional nonretirement social security benefits; including them would only increase the value of social security. one could extend the analysis to reflect the probability of outliving the actuarial life expectancy. we assume couples were the same age and used gender-neutral actuarial tables. we make several simplifying assumptions related to considering only a high earner who receives maximum social security benefits, but social security’s social insurance aspects favor low and middle earners. if private savings is proportional to income, the portfolio impact for low and middle earners will be even greater than the dramatic impact we show for high earners. all of these impacts could be quantified in detail; instead, we briefly consider gender, income-level and retirement-age in the appendix. however, the point of this paper is to demonstrate a method of quantifying social security wealth and demonstrate its portfolio impact; rather than attempt to accommodate infinite variations, we suggest that our method be applied to an individual’s specific circumstances. to be realistic, the dual prescriptions of this paper—accurately valuing social security wealth and then integrating that wealth into asset mix decisions—will most likely be implemented by financial educators or financial planners. we agree with kritzman (1992) that “the key challenge. . . is to present asset allocation (to individuals) in a way that appeals to an individual’s intuition without compromising the integrity of the analysis.” as stated above, we find the parallel between social security payments and inflation-indexed treasury bond coupon payments to be intuitive and persuasive. we hope that this intuition on valuing and integrating social security wealth will aid financial planners and educators to better implement rigorous asset allocation decisions with their clients or students. acknowledgments the opinions included are those of the authors and not necessarily those of the us air force academy, the us air force or any other federal agency. we anticipate maintaining related materials at www.williamjennings.com. we are grateful to vickie bajtelsmit, tim burch, conrad ciccotello, karen lahey (the editor), bill reichenstein, c. e. steuerle, participants at the 2000 academy of financial services conference in seattle, and four anonymous reviewers for their input to this project. 321s.p. fraser et al. / financial services review 9 (2000) 295–326 appendix a. generalized results we can generalize qualitative aspects of our results. our base case analysis made certain assumptions about gender, income levels, and the retirement age decision; here, we demonstrate the dramatic impact of including social security wealth in an expanded portfolio for men, women, middle-income workers, and those that retire at age 65. we consider each permutation separately and examine its impact on fifty-year-old workers. because of data availability, we use 1999 information in this appendix. without loss of generality, we focus on the 60/40 (stocks/bonds) fixed mix decision rule. all of our other assumptions remain unchanged. a.1. gender impact given equal earnings, men and women have different social security wealth because their life expectancies differ. since women have longer life expectancies, they have greater social security wealth. women, therefore, will have a greater bond allocation than men in their expanded portfolio for any asset-mix decision rule. panel a of table 11 shows this gender impact. we use irs gender-specific actuarial tables for life expectancies. (note that life expectancies under these tables for pre-1986 annuities are not comparable to the more modern irs gender-neutral actuarial tables; however, they do serve our purpose in this appendix by allowing comparison between men and women.) the gender impact is greater for single individuals but is quite small even then. for fifty-year-old workers pursuing a 60/40 fixed mix, single women’s expanded portfolio stock allocation is within 3% of men. for married couples, the difference is less than 1%. this result holds when we change our assumption of equal-age couples to a case when the nonworking spouse is ten years younger. accommodating gender differences in life expectancy has minimal impact on our conclusion that social security wealth dramatically affects the asset mix decision. a2. income level impact given the same percentage savings rates, people with different income levels will have different asset allocation consequences of including social security wealth in their expanded portfolios. income level affects both the amount of social security benefits and the size of financial portfolios. individuals with different incomes have different expanded portfolio consequences because social security benefits are a concave function of income. middleincome workers have proportionally more of their wages replaced by social security benefits than high-income workers do. accordingly, middle-income workers have greater social security wealth relative to their financial portfolios; thus, they will have a greater bond allocation than high-earners in their expanded portfolio for any asset-mix decision rule. note that using the high-income worker as our base case was a conservative assumption since it reduced the impact of social security wealth on the expanded portfolio. we compare a $72,600 high-income worker with a $36,456 middle-income worker. the $36,456 amount reflects someone earning 160% of the national average wage in 1999 and is approximately one-half the high-earner income. given equal percentage savings, financial 322 s.p. fraser et al. / financial services review 9 (2000) 295–326 portfolio values will also be approximately one-half those of the high earner. their social security benefits, however, will be substantially greater than one-half those of the high earner. according to the social security web site (1999), this middle-income worker can expect a pia of $1,231, or about 90% of the high-earner’s maximum pia. note that using the high-earner as our base case was a conservative assumption since it reduced the impact of social security wealth on the expanded portfolio. panel b of table 11 shows this income impact. for fifty-year-old workers pursuing a 60/40 fixed mix, the middle-income expanded portfolio stock allocation is within 9% of the high-income allocation. the impact would be greater still upon adding the value of the tax shield on social security benefits for lowand middle-income brackets. while this quantitative difference is economically significant, it does not affect our qualitative conclusion that social security wealth has a dramatic impact on asset allocation. it does encourage using individual-specific valuations as discussed in section 5. table 11 implications of a 60/40 mix in the financial portfolio for the expanded portfolio demography perceived financial portfolio resultant expanded portfolio including social security wealth 5% saver 10% saver 15% saver mnd saver stocks bonds stocks bonds stocks bonds stocks bonds stocks bonds panel a: gender impact age status gender 50 single male 60% 40% 39% 61% 47% 53% 51% 49% 49% 51% female 60% 40% 36% 64% 45% 55% 49% 51% 47% 53% m � f � 2.7% �2.7% 2.0% �2.0% 1.6% �1.6% 1.8% �1.8% spouse is 50 male 60% 40% 29% 71% 39% 61% 44% 56% 42% 58% female 60% 40% 29% 71% 39% 61% 44% 56% 41% 59% m � f � 0.8% �0.8% 0.7% �0.7% 0.6% �0.6% 0.7% �0.7% spouse is 40 male 60% 40% 28% 72% 38% 62% 43% 57% 40% 60% female 60% 40% 27% 73% 38% 62% 43% 57% 40% 60% m � f � 0.4% �0.4% 0.3% �0.3% 0.3% �0.3% 0.3% �0.3% panel b: income impact age status income 50 single $72,600 60% 40% 34% 66% 44% 56% 48% 52% 46% 54% $36,456 60% 40% 26% 74% 36% 64% 42% 58% 39% 61% married $72,600 60% 40% 27% 73% 37% 63% 43% 57% 40% 60% $36,456 60% 40% 19% 81% 29% 71% 35% 65% 31% 69% panel c: retirement age impact age status retire at 50 single 62 60% 40% 39% 61% 47% 53% 51% 49% 49% 51% 65 60% 40% 35% 65% 44% 56% 48% 52% 46% 54% married 62 60% 40% 28% 72% 38% 62% 43% 57% 40% 60% 65 60% 40% 27% 73% 37% 63% 42% 58% 39% 61% expanded portfolio stock/bond mix resulting from adding social security wealth to a financial portfolio with a 60/40 fixed mix. social security wealth is treated as a bond and uses values from table 1 computed using real tips yields for a high-earning worker. savings rates are applied over a working lifetime to create financial portfolio values; mnd saver’s financial portfolios are computed for “prodigious accumulator[s] of wealth” per stanley and danko (1996). 323s.p. fraser et al. / financial services review 9 (2000) 295–326 a3. retirement age impact all else equal, people planning on waiting until age 65 to draw social security will have different asset allocation consequences of including social security wealth in their expanded portfolios than those that start at age 62. retirement age affects both the amount of social security benefits and the size of financial portfolios. waiting to retire will reduce the early retirement penalty and increase the financial portfolio; however, waiting decreases the expected duration of receiving benefits. given all of our assumptions, retiring at age 65 implies greater social security wealth than retiring at age 62. those that retire at age 65 have greater social security wealth relative to their financial portfolios; thus, they will have a greater bond allocation than those retiring at age 62 in their expanded portfolio for any asset-mix decision rule. note that using the worker retiring at age 62 as our base case was a conservative assumption since it reduced the impact of social security wealth on the expanded portfolio. we also believe it a more likely scenario for the reasons discussed in section 3.3. panel c of table 11 shows the retirement age impact. for fifty-year-old workers pursuing a 60/40 fixed mix, the delayed-retirement expanded portfolio stock allocation is within 4% of the early-retirement allocation. accommodating retirement age differences has minimal impact on our conclusion that social security wealth has a dramatic impact on asset allocation. a4. summary of generalized results as with the base case discussed in the body of the paper, the generalized results demonstrate the dramatic impact of including social security wealth in an expanded portfolio. all else equal, gender differences and retirement-age differences had minimal impact on the strategic asset allocation. because social security replaces more income for lower income workers, income differences affect the asset mix decision more than gender or retirement-age differences. even with income differences, the stock allocation was within 9% of the high-earner base case in the body of the paper. note, crucially, that because social security replaces more income for lower income workers, excluding social security wealth is a bigger error for them. thus, focusing on high-earners in the body of the paper is conservative—it understates the case for including social security wealth. in short, none of the generalized results affects the conclusion that social security wealth has a dramatic impact on asset allocation and should be included in strategic asset allocation. references bajtelsmit, v. l., & turner, j. a. (1998). a portfolio model of retirement income including social security. working paper, colorado state university. bodie, z. (1990). inflation, index-linked bonds, and asset allocation. the journal of portfolio management, winter, 48–53. bogle, j. c. (1994). bogle on mutual funds. new york: irwin press. brinson, g. p., hood, l. r., & beebower, g. l. (1986). determinants of portfolio performance. financial analysts journal, july/august, 39–44. 324 s.p. fraser et al. / financial services review 9 (2000) 295–326 brynjolfsson, j., & fabozzi, f. j. (eds.) (1999). handbook of inflation indexed bonds. new hope, pa: frank j. fabozzi associates. brynjolfsson, j., & rennie, e. p. (1999). asset allocation with tips. in f. j. fabozzi (eds.), handbook of inflation indexed bonds. new hope, pa: frank j. fabozzi associates. chen, p., & terrien, m. (1999). tips as an asset class. working paper, chicago: ibbotson associates. clements, j. (1998). 25 myths you’ve got to avoid if you want to manage your money right. new york: simon & schuster. congress (1996). overview of entitlement programs: 1996 green book. house, committee on ways and means. washington, dc: u.s. government printing office. cornell, b. (1999). the equity risk premium: the long-run future of the stock market. new york: john wiley & sons. dalio, r. t., & bernstein, d. s. (1999). shifting the efficient frontier with inflation-indexed bonds. in j. brynjolfsson & f. j. fabozzi (eds.). detweiler, j. h. (1999). a note on the 62–65 social security decision. national estimator, spring, 39–42. evensky, h. (1997). wealth management: the financial advisor’s guide to investing and managing client assets. new york: mcgraw-hill. feldstein, m. (1974). social security, induced retirement, and aggregate capital accumulation. journal of political economy, 82, 1095–1117. fisher, k. l., & statman, m. (1997). investment advice from mutual fund companies: closer to the talmud than to markowitz. the journal of portfolio management, fall, 9–25. gitman, l. j., & joehnk, m. d. (1999). personal financial planning (8th ed.). fort worth: the dryden press/harcourt brace college publishers. gutner, t. (2000, march 24). investment strategies that alter as you age. hers.online column. available at www.businessweek.com/bwdaily. hubbard, r. g. (1985). personal taxation, pension wealth, and portfolio composition. review of economics and statistics, 67, 53–60. ibbotson associates. (1999). stocks, bonds, bills and inflation 1999 yearbook. chicago: ibbotson associates. internal revenue service (1997). general rule for pensions and annuities. (publication 939, rev. june 1997.) kan, w. (1999). understanding the cpi. in j. brynjolfsson & f. j. fabozzi (eds.). kapoor, j. r., dlabay, l. r., & hughes, r. j. (1999). personal finance (5th ed.). boston: irwin/mcgraw-hill. kinney, c. f. (1999). the individual investor and inflation-indexed bonds: tax issues and retirement planning. in j. brynjolfsson & f. j. fabozzi (eds.). kritzman, m. (1992). asset allocation for individual investors. financial analysts journal, january-february, 12–13. lamm, r. m. (1998). asset allocation implication of inflation protection securities. the journal of portfolio management, summer, 93–100. leimer, d. r. (1995). a guide to social security money’s worth issues. social security bulletin, 67, 24–34. maginn, j. l., & tuttle, d. l. (eds.) (1990). managing investment portfolios (2nd ed.). charlottesville, va: association for investment management and research. malkiel, b. g. (1996). a random walk down wall street: including a life-cycle guide to personal investing (6th ed.). new york: w. w. norton and company. markowitz, h. (1959). portfolio selection: diversification of investments. new york: john wiley and sons. munnell, a. h. (1998). rates of return on social security contributions. testimony before the budget committee task force on social security, u.s. senate. january 21, 1998. available at www.socsec.org. newmark, c. m., & walden, m. (1995). should you retire at age 62 or 65? financial counseling and planning, 6, 35–44. phoa, w. (1999). quantitative approaches to inflation-indexed bonds. in j. brynjolfsson & f. j. fabozzi (eds.). pratt, s. p., reilly, r. f., & schweihs, r. p. (1996). valuing a business. new york: mcgraw-hill. reichenstein, w. (1996). basic truths about asset allocation: a consensus view among the experts. aaii journal, october, 1996. reichenstein, w. (1998). calculating a family’s asset mix. financial services review, 7 (3), 195–206. 325s.p. fraser et al. / financial services review 9 (2000) 295–326 reichenstein, w. (2000). calculating the asset mix. journal of wealth management, 3 (2), 20–25. rosenman, r., & fort, r. (1992). the correct value of social security contributions in personal injury and wrongful death settlements. journal of forensic economics, 5, 149–158. rudolph-shabinsky, i. (2000). inflation-protection securities in an asset allocation framework. investment counseling for private clients ii: aimr conference proceedings, 3, 26–38. scott, m. c. (1995). defining your investment portfolio: what should you include? aaii journal, november, 1995. scott, m. c. (1997). re-examining your asset allocation in light of vested pension benefits. aaii journal, january, 1997. siegel, j. (1998). stocks for the long run. new york: mcgraw-hill. smith, j. p., & kington, r. (1997). demographic and economic correlates of health in old age. demography, 34, 159–170. social security administration website (1999, october 19). illustrative benefit table for workers with high earnings. available at: www.ssa.gov/ocat/cola/illushigh.html. accessed november, 14, 1999. social security administration (2000). how work affects your benefits (publication 10069). stanley, t. j., & danko, w. d. (1996). the millionaire next door: the surprising secrets of america’s wealthy. marietta, ga: longstreet press. steuerle, c. e., & bakija, j. (1994). retooling social security for the 21st century. washington, dc: the urban institute press. steuerle, c. e. (1996). confronting the challenges presented by an aging population. statement before the labor and human resources subcommittee on aging, u.s. senate, may 14, 1996. available at �www.urban.org . stoller, m. a. (1992). estimating the present value of pensions: why different estimators get varying results. journal of legal economics, 2, 49–61. thomas, w. w. (ed.) (1999). social security manual. ’99 edition. cincinnati: national underwriter company. 326 s.p. fraser et al. / financial services review 9 (2000) 295–326 pii: s1057-0810(00)00041-x undergraduate research: the senior thesis in finance ellie fogartya, herbert mayob athe college of new jersey business library, the college of new jersey, box 7718, ewing, nj 08628-0718, usa bdepartment of finance, the college of new jersey, school of business, box 7718, ewing, nj 08628-0718, usa abstract an undergraduate senior thesis offers the individual student an opportunity to pursue a topic of special interest or in greater depth than available in a traditional course. this paper describes the senior thesis in finance required of all finance majors enrolled in the college of new jersey, specifies the objectives of the thesis, and offers the results on an assessment survey of alumni, who completed the senior thesis. © 1999 elsevier science inc. all rights reserved. jel classification:a220 keywords:senior thesis; undergraduate research 1. introduction instituting a senior thesis for undergraduate majors can offer both students and faculty an opportunity for a challenging academic experience. by emphasizing a single topic, which is covered in depth, the senior thesis serves objectives that are distinct from traditional courses. it can tie together material learned in various business courses and allow the student to explore one topic in detail. the senior thesis can also be used to investigate an area in finance not offered by the major, which broadens both the appeal of the school’s program and marketability of the student. this paper discusses the senior thesis in finance at the college of new jersey (tcnj). * corresponding author. tel.:11-609771-3014; fax:11-609-6375129. e-mail address:mayoher@tcnj.edu (h. mayo). financial services review 8 (1999) 223–234 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00041-x the paper begins with a brief description of the tcnj finance program. next follows the mechanics of the course, objectives of the senior thesis, assessment, and special problems. the primary purpose of this paper is to report the tcnj experience and initial efforts to assess it. faculty at other institutions who are considering initiating a senior research project for majors should also find the discussion of the mechanics and implementation problems useful. 2. the tcnj finance program tcnj is medium-sized school of about 6,000 students with average sat scores of 1,200. the school of business has approximately 1,000 undergraduate students with 200 freshman admits and 50 transfers each year. (no graduate degrees are offered by the school of business.) there are 30 to 34 faculty in the school of business, depending on how administrators who also teach are counted. the finance program has about 160–180 majors (about 40 graduates a year) currently being taught by four faculty. when the senior thesis was started, the number of majors was about 120 with three full-time faculty. while new faculty do not supervise theses during their first year, they are integrated into the program during the spring semester of the second year. adjunct faculty are used to teach finance courses but not the senior thesis. the number of adjunct faculty varies each semester, as regular faculty are granted release time for research or administrative duties. the finance curriculum is part of the school of business, which offers traditional business programs in accounting, economics, general business, international, management, and marketing as well as finance. the finance program consists of 1. a general education core 2. a core of business courses consistent with aacsb requirements 3. six courses (18 credits) in finance above the intro course required of all majors in the school of business. four of the six finance courses are required (introduction to investments, working capital management, capital budgeting, and senior thesis), and two “options” are chosen from such courses as corporate valuation, international finance, commercial banking, derivatives securities/portfolio management, and estate and retirement planning. if students were not required to complete the senior thesis, they would substitute one of the options. thus, the alternative to the senior thesis (assuming the number of finance courses remains at six) would be to require the student to take three finance options. national interest in capstone experiences increased during the 1980s in response to critical studies of american higher education. (see, for instance, the national commission on excellence in education.a nation at risk: the imperative for education reform, 1983 or the association of american colleges.integrity in the college curriculum: a report to the academic community, 1985.) the senior thesis in finance is one possible capstone experience. in a wonderful analogy, heinemann (1997) classified senior experiences as either “domes,” which are designed to integrate and unify a body of knowledge, or “spires,” which are designed to focus on further exploration in a specific area. 224 e. fogarty, h. mayo / financial services review 8 (1999) 223–234 the tcnj senior thesis in finance is designed as a “spire.” the school of business has its own capstone course (strategic management) which integrates material form the various functional areas of business. the required courses in corporate finance give students a thorough grounding in financial theory and applications and the required course in investments repeats some of this material from a different perspective. hence, a unifying capstone course is unnecessary. the senior thesis in finance was started in 1992 and initially only a handful of students (transfers into the program) had to complete the thesis. students who enrolled prior to establishing the requirement were exempted. during the phase-in period, one faculty member handled all the theses except for a few specialized topics that were covered by another faculty member on an ad hoc basis. senior thesis sections are scheduled as a regular course with specified hours and specified rooms. while traditional courses have 25–30 students, enrollment in a section of senior thesis is capped at ten students. under optimal conditions, this distribution implies each of the four faculty would have one section of ten students during the spring semester. this distribution of the workload assumes the number of majors remains stable at approximately 40 students each year and all students complete the thesis during the spring semester. however, the number of majors does fluctuate, so a given section may have fewer or more than ten students. in addition, not all students can complete the thesis during the spring. to meet this need, one section is offered during the fall semester. since the senior thesis is treated as any other course, students sign up for individual sections. the faculty member responsible for the section is identified in the roster of courses. from prior course work, students know the areas of the various instructors (e.g., investments, international, or financial institutions) and enroll in the appropriate section based either on their desire to work in a specific area or to work with a specific instructor. as with any other course, the instructor cannot refuse to work with an enrolled student. how an individual section is conducted depends on the instructor. usually during the first week a formal class is held to identify topics and start the general process described in the next section. additional classes may be held at intervals during the semester to meet specified progress milestones. since the section has a specified meeting time and place, both student and faculty know there is a time available at which they can meet. this avoids the problem of not being able to find a mutually convenient time for meetings but does not preclude meetings at different times or at different places such as faculty offices, computer labs, or the library. 3. the mechanics of the tcnj senior thesis while there are differences among each instructor’s approach, the following gives the general process for the student’s completing the senior thesis. the first step is for the student to identify a topic. undergraduate courses in the functional areas of business generally start at the junior level, so students will have completed at best a modest amount of coursework in finance. students with topics need guidance to refine and make the thesis manageable for 225e. fogarty, h. mayo / financial services review 8 (1999) 223–234 the short period (one semester) during which they have to complete the project. students tend to be too ambitious with regard to scope, and the better defined or focused the topic, the easier the project will be for both the student and the faculty member. students who have not identified a topic require more initial assistance. a series of questions such as (1) what areas did you enjoy in your course work? (2) what do you plan to do upon graduation? or (3) do you have a special interest that may be tied to finance? may help the student identify a particular area of interest. flexibility on the part of the faculty member helps. for example, one finance student expressed a desire to work for the fbi upon graduation, which led to a senior thesis on white-collar crime: bank fraud. while bank fraud is not a traditional or typical finance topic, the thesis resulted in the student interviewing fbi agents as well as researching traditional finance literature. the topic also helped the student understand how the finance major could contribute to a career in law enforcement. individual students are permitted to pursue similar topics. for example, comparing mutual fund performance is an area that is potentially useful to a finance major considering a career in financial planning. students who are working on similar topics are encouraged to share information, but each thesis is individually written and goes through the process of review and revision. there are no joint projects. the knowledge and interests of the faculty also affect the selection of a topic. a student, however, may want to pursue a topic about which the faculty member has, at best, modest knowledge and, at worst, knows virtually nothing. an alternative approach discussed in amyotte (1989) is to determine a set of senior projects and let the student select a topic from the list. having a list improves the faculty member’s comfort level but relieves the student of having to identify a topic. the list also may reduce student interest and defeat one of the objectives of the senior thesis to pursue an area of special interest. since the discomfort problem can only occur if the student suggests the topic, instructors have to balance their discomfort level with the possible enthusiasm the student may have for the topic. unless the topic is far afield, it is unlikely that someone cannot be found to supervise the thesis. after selecting the topic, the student is asked to complete a literature search. since one on the objectives discussed in the next section is exposure to academic literature, each student is encouraged to find readings in thejournal of finance, financial management, thejournal of portfolio management, or thefinancial analysts journal. the business reference librarian then helps identify which publications are immediately available and which require interlibrary loan. (the school of business covers the costs of interlibrary loans.) if the topic requires data, the librarian can also help the student locate or access a database. obtaining data are often a major constraint on the undergraduate thesis. inaccessibility of data available in a timely fashion or the need to limit the amount of data forces the student to answer many questions. should the sample size be reduced? should annual, monthly, or weekly data be used? what time period should be covered? in addition to the bibliography, the student submits a statement of purpose. the statement is corrected for grammar, form, and content. students subsequently submit a section on methodology, a presentation of results, and other appropriate sections as completed. submissions are returned with comments and encouragement. this system contributes to original work and reduces the potential for cheating. 226 e. fogarty, h. mayo / financial services review 8 (1999) 223–234 faculty comments serve several purposes. first, the suggestions help refine the topic and make it more manageable. second, the student becomes more aware of the importance of grammar, organization, and structure. third, the comments help the student develop selfcriticism. many students will have to write more during their careers than they currently realize and now is the best time to start becoming more aware of the need to develop self-criticism and improve writing skills. after the paper is finished (or virtually finished except for the last rewrites), the student gives a 15 to 20 min oral presentation. four to six students give their papers at a session. after completion of the presentation each student attending the session is expected to ask at least one question which forces everyone to listen. the topics covered by the students have been varied, especially those selected by the students themselves. a brief list of titles is provided in table 1. students who want to work with a specific instructor may select a topic that both instructor and student find interesting. these topics may not be as varied, since the faculty member may direct the selection process. since many finance students prefer investments, a large percentage of theses (perhaps as many as 50%) are completed in that area. the growth in enrollment in international finance has generated topics in that area, which in some cases combine investments and international. while publication is not an objective of the senior thesis, an occasional thesis has been published or presented at professional meetings. (see table 2 for titles of articles and presentations that have evolved from the senior thesis.) all senior theses that are published or presented at professional meetings are the joint work of the student and the sponsoring faculty. refining the thesis for possible presentation or publication requires considerable work and effort by both the student and the faculty member. even if a student does exceptional work, the review process operates against publication. the need to revise and resubmit requires the student to continue working on the thesis after graduation. most students’ jobs and other constraints make subsequent revisions difficult, but the successes given in table 2 are exceptionally gratifying to both the student and the sponsoring faculty member. table 1 selected titles of senior theses a comparison of 401(k) plans art as an investment credit card arbitrage escrowed to maturity bonds: call provisions and yields financial leverage and risk in the retail department stores industry insuring mutual funds with stock index put options manager tenure as a predictor of mutual fund performance municipal bond defaults, 1988–1992 predicting airline bankruptcy price appreciation of vintage claret residual price difference between japanese adrs and their underlying securities returns to investments in cartoon art socially responsible investing 227e. fogarty, h. mayo / financial services review 8 (1999) 223–234 4. senior thesis objectives the objectives for the senior thesis include the following. 4.1. opportunity to pursue an individual interest one important advantage of the senior thesis is the opportunity for the student to pursue something of personal interest. for those students, who have a general area they wish to pursue or a well-defined topic, the senior thesis provides an opportunity and incentive to explore their special interest. the thesis may also be used to cover a topic that may be important for career goals but that is not taught in existing courses. for example, a student interested in pursuing a career in financial planning may write a thesis on charitable annuities, which are not currently covered in existing courses. 4.2. development and application of research skills the senior thesis presents an opportunity to apply analytical skills learned in specific finance courses and to use statistical, accounting or other analytical tools learned in various courses. while not every student pursues a topic that applies specific skills (e.g., regression analysis), an effort can be made by the faculty to encourage topics that involve analytical tools learned in finance and other courses. table 2 senior theses that have been published or presented at professional meetings kagan, g. & mayo, h. (1995). risk-adjusted returns and stock market games.journal of economic education, 26, 39–50. landis, m. & patrick, t. (1998) mutual fund risk and return.proceedings of the northeast business & economic association, annual meetings. paladino, m. & mayo, h. (1995). investments in reits do not help diversify stock portfolios.real estate review, 23–26. paladino, m. & mayo, h. (1997). investments in reits do not help diversify stock portfolios, an update. real estate review, 39–40. parisi, m. & patrick, t. (1997). the risk between moody’s bond ratings and the altman z score for firms in the transportation industry: a case study.proceedings of the northeast business & economic association, annual meetings. pikul, j. & mayo, h. (1999). performance and eligibility for arbitration or free agency and salaries of professional major league baseball players, the 1994–1995 experience.journal of sport & social issues, 23, 353–361. philbrook, l. & patrick, t. (1999). the relationship between bond ratings and yield curves.proceedings of the northeast business & economic association, annual meetings. wunder, g. & mayo h. (1995). study supports efficient-market hypothesis.journal of financial planning, 128–135. 228 e. fogarty, h. mayo / financial services review 8 (1999) 223–234 4.3. oral and written communication skills the development of oral and written communication skills is one of the most important goals of the senior thesis. each student completes a written paper and gives an oral presentation of the results. 4.4. interaction between student and faculty increased interaction between faculty and student is more than a fringe benefit of the senior thesis. it forces one-on-one interaction between the student and the faculty member. while the process is time consuming and requires willingness on the part of the faculty member to work individually with the undergraduates, both the student and the faculty member can benefit. for example, letters of recommendation with substantive comments concerning the student’s academic ability and communication skills are easier to write. relationships after graduation are easier to maintain, which helps current students who are looking for internships or a job after graduation. 4.5. literature exposure many undergraduate courses in business do not expose students to primary research. all courses have time constraints. the use of problems, cases, and other instructional tools in business courses often comes at the expense of academic literature. one goal of the senior thesis is to introduce this type of academic literature. while most undergraduates cannot be expected to read and understand the majority of material in academic journals, they can be made aware of its existence. this exposure to research materials gives the student some indication of what advanced study in the discipline will cover. 4.6. product differentiation tcnj operates in a competitive market for quality students. several colleges exist within a radius of fifty miles, and many offer programs in business with finance majors. the ability to study with an individual professor at the undergraduate level differentiates the tcnj program from the majors offered by other schools in our immediate geographic area. finance programs are offered at lehigh university, seton hall university, rider university, villanova university, and lafayette college. each of these programs has a course with a title such as directed research, independent research, independent study, or special topics. several (e.g., lafayette and villanova) have a thesis as a requirement in their honors program. none of the programs, however, requires a senior thesis of all undergraduate finance majors. the same concept applies to the students who have to compete against graduates from other colleges and with other majors from tcnj. just as the tcnj finance program uses the senior thesis to differentiate the school’s program from competing finance programs, individual students may use the thesis to differentiate themselves from their competitors. since finance majors have had similar classes, the senior thesis becomes a means to indicate the 229e. fogarty, h. mayo / financial services review 8 (1999) 223–234 uniqueness of the individual. some students have reported that interviewers do become fascinated with the topic and that a large proportion of the interview was devoted to the thesis. while it may be impossible to know if the thesis results in better employment opportunities, it provides an opening for the student to discuss specific skills and to impress the interviewer. 5. assessment while courses have objectives, the determination if the objectives are met requires assessment. identifying strengths and weaknesses may help provide evidence of the senior thesis’s effectiveness. in addition to determining if the specific objectives were being met, we wanted to learn if the graduates believed that the thesis should be continued or if it would be more beneficial to substitute another finance course. when the senior thesis was instituted, the finance faculty members were concerned with the writing abilities of their students. interest in improving writing skills was also a concern of the school of business, which was pursuing aacsb accreditation. since written communication skills are an important component of aacsb accreditation standards, the school was considering alternatives means to integrate writing into the curriculum. finance, however, was the only program to institute a senior thesis and questions arose concerning its appropriateness for a business program (the thesis stresses academic writing) and whether the thesis could serve as a model for the school or the college. (other programs in the school of business require a thesis for honors students but not for all students. economics and marketing require senior projects, which may be broader than the academic emphasis placed on the senior thesis in finance. the college as a whole is considering some type of senior experience for all students.) the initial consideration for assessing the senior thesis was to use the evaluation form that students complete for every course. it became apparent, however, that this form would not do. questions such as fairness of tests or the use of class time do not apply to the senior thesis. surveys of graduates made by the school of business were too general and at best produced only anecdotal evidence concerning the senior thesis. a separate questionnaire was sent to ninety-two finance majors who had graduated after the senior thesis had been instituted and whose addresses could be obtained from the college’s alumni office. the questionnaire was constructed by one finance faculty member and the business reference librarian. the questionnaire was purposely kept brief in an effort to encourage a better response rate. the questionnaires were sent and collected by the business reference librarian instead of the faculty member. a total of 33 (35.9%) completed questionnaires were returned. the questionnaire consisted of twenty-one statements and the individual was given five choices: strongly agree (5), agree (4), neither agree nor disagree (3), disagree (2), or strongly disagree (1). the statements were designed to cover the specific objectives, and most of the statements were written in the positive so that the larger number is a positive response. the statements (and the objectives they relate to), the average score, the standard deviation, and possible interpretation concerning the course’s objectives follows. 230 e. fogarty, h. mayo / financial services review 8 (1999) 223–234 5.1. objective: opportunity to devote time to a topic of special interest statement: the thesis gave me the opportunity to spend time working on a finance topic of interest to me. —4.5 (av), 0.62 (sd) i would have preferred an assigned topic instead of developing my own topic. —1.6(av), 0.77 (sd) i would have preferred choosing my topic from a list of possible subjects. —2.5 (av), 0.92 (sd) these results suggest that students thought the thesis gave them the opportunity to study a topic of special interest. while few supported being assigned a topic, they seemed somewhat ambivalent about been supplied a list from which to choose a topic. 5.2. objective: development and application of research skills statement: i made use of research skills learned in other classes. —3.8 (av), 0.86 (sd) the thesis did not build on any previous finance course. —1.7 (av), 0.78 (sd) these results suggest that the senior thesis did build on finance courses and that some students did use research skills learned in other courses. the results also suggest that some students did not use skills learned elsewhere. this could be caused by insufficient skills being required by the thesis or the necessary skills were learned as the thesis progressed and not in other courses. 5.3. objective: develop oral and written communication skills statement: discussing my topic with others (students, faculty, librarians) helped improve my oral communication skills. —3.7 (av), 0.83 (sd) my thesis presentation helped me develop my oral communication skills. —3.5 (av), 0.77 (sd) writing and editing my thesis helped me to develop my written communication skills. —4.5 (av), 0.51 (sd) these answers suggest that the students believe the senior thesis helped their written communication skills and suggest that having to explain the topic helped improve oral communication skills. an opportunity may exist to further develop oral communication skills if students can be placed in a setting in which they have to explain the thesis (e.g., class meetings in which they discuss their topics and progress). 5.4. objective: increased interaction between faculty and the individual student statement: the thesis allowed me to work closely with the professor. —4.8 (av), 0.91 (sd) there was enough faculty/student interaction while i was working on my thesis. —4.6 (av), 0.61 (sd) these results indicate that increased interaction is perceived as an objective and the objective is accomplished. 231e. fogarty, h. mayo / financial services review 8 (1999) 223–234 5.5. objective: exposure to academic and professional literature not normally covered in course work statement: researching my topic exposed me to research literature in academic finance journals. —4.2 (av), 0.71 (sd) while this result suggests that students did get the desired exposure, the students in the survey predate the widespread use of the internet by students. whether this objective would be as readily achieved today is doubtful unless the instructor explicitly requires a specified minimum number of academic citations. 5.6. objective: differentiation of our graduates from those who attend other schools statement: when interviewing for a job, the thesis helped separate me from other applicants. —3.6 (av), 0.77 (sd) my current position is related to my thesis topic. —2.7 (av), 1.41 (sd) while anecdotal evidence strongly suggests that students use the thesis to differentiate themselves from other students, these results confirm that the thesis does differentiate but that the tendency is not strong. while few students appear to use the topic or general area in their employment, the large standard deviation suggests that some students are in positions related to the topic while others are not (i.e., the distribution is bimodal). 6. additional information from the survey resources are obviously an important part of a senior thesis. a byproduct of the senior thesis is that a closer working relationship can develop between the library and computer staffs and the faculty. collection development of the library resources may be more specifically coordinated to support the thesis program. as bailey (1985) points out, the librarian can assess the research capabilities of the library after leading students through the library’s resources. two statements concerning resources were included in the survey: statement: i was able to obtain the materials i needed through the library’s services. —3.8 (av), 1.02 (sd) the computer facilities were adequate to complete my thesis. —3.5 (av), 1.50 (sd) in terms of college resources the library fared better than the computer facilities. since computer facilities are periodically upgraded, the survey, however, may not indicate if current facilities are sufficient. in addition to resources, the senior thesis is expensive in terms of alternatives. that is, students complete the thesis at the expense of an alternative course. faculty members have to supervise the students and since all majors must complete the thesis, it reduces faculty availability for other assignments. (there is also an implicit assumption that faculty are willing to supervise the senior thesis on a continuing basis.) perhaps the biggest problem associated with the senior thesis is faculty resources. if ten students constitutes one class and four sections are taught during the spring semester, that is the equivalent of one faculty 232 e. fogarty, h. mayo / financial services review 8 (1999) 223–234 member’s total load. thus, the senior thesis may increase the need for adjuncts or increase the size of other classes. one suggestion for overcoming these problems is to have the thesis as an honors project only. the survey included four statements concerning these issues. statement: an additional finance course would have been more valuable than the senior thesis. —1.8 (av), 0.69 (sd) the thesis should only be required as part of an honors program in finance. —1.9 (av), 0.95 (sd) i think the thesis should be retained as a requirement for graduation. —4.3 (av), 0.90 (sd) a thesis should be required of all majors in the school of business. —4.3 (av), 0.96 (sd) the results of all four statements indicate that the finance graduates favor the requirement for all students. the respondents viewed the senior thesis as being more valuable than another course in finance, that it should not be limited to the best (“honors”) students, that is should be retained as a requirement, and that the requirement should be extended to all students in the school of business. similar results were found in a survey of princeton university’s class of 1954, in which the undergraduate thesis stood out among the students’ memories twenty-fiveyears after graduation. the positive impact was remembered in statements such as “taking a mass of information, arranging it in a logical sequence and drawing conclusions . . . was a significant challenge” or “potential for personal growth . . . or in meeting the challenge is in my opinion enormous (heath, 1979).” 7. special problems while the senior thesis can be an excellent learning experience for many students, it is naive to conclude that the senior thesis is a positive experience for all participants. for motivated students, who develop their own topics and can work at their own initiative, the senior thesis is a valuable learning experience. some students become wrapped up in the process and develop a sense of accomplishment. this is particularly true for students who have a topic of personal interest. while it cannot be verified, anecdotal evidence suggests that the best work is often done by the above average student (gpa of 3.0 to 3.4) but not the best students (gpa exceeding 3.4). this belief is one of the arguments for requiring the senior thesis for all students and not limiting the thesis to a selected few students. in addition to generating interest in a specific topic, students often become interested in the learning process as they rewrite and revise their work. while such enthusiasm does not apply to all students, it does apply to many, and the level of this enthusiasm appears to be greater than for a formal course. even for weaker students with major gaps in their backgrounds, the senior thesis can offer an opportunity to improve important skills and develop the responsibility associated with working on your own. however, as prud’homme (1981) so aptly expresses it: “. . . the thought of dragging an unmotivated, unwilling, and ill-prepared student through a year of senior thesis research sends chills up my spine.” the unmotivated and unwilling student poses a real dilemma. while fear or lack of motivation has resulted in some students changing their major, some students start the process and then disappear so that the student becomes one course short of 233e. fogarty, h. mayo / financial services review 8 (1999) 223–234 graduation. while being abd is not unusual for graduate students, such a status generally does not apply to undergraduates. it raises an important question: should a student be denied graduation for failing to complete one requirement that is unique to the specific program? currently, the answer depends on the individual instructor. since each instructor determines the student’s grade, what constitutes acceptable work or what differentiates grades is left to the discretion of the instructor. 8. concluding comment the tcnj senior thesis in finance offers an opportunity to develop the research and communication skills of students who will soon become professionals. whether the students enter industry or attend graduate or professional school, research will be an important component of their lives. few students realize how important research and communication skills may be for their careers. the tcnj senior thesis in finance gives our faculty an opportunity to focus on these skills in a manner that is not possible in an individual finance course and thus differentiates our majors from those graduating from institutions not offering this academic experience. references amyotte, p. r. (1989). an alternative approach to the undergraduate thesis.chemical engineering education, 23, 28–30. association of american colleges. (1985).integrity in the college curriculum: a report to the academic community. washington, dc. bailey, w. (1985). thesis practicum and the librarian’s role.journal of academic librarianship, 11, 79–81. heath, r. (1979).princeton retrospectives. princeton, nj: the class of 1954. heinemann, r. (1997). the senior capstone, dome or spire?“ paper presented at the 83rd nca annual convention, chicago, il, november, 20–23. national commission on excellence in education. (1983).a nation at risk: the imperative for education reform. washington, dc: u.s. department of education. prud’homme, r. k. (1981). senior thesis research at princeton.chemical engineering education, 15, 130–132. 234 e. fogarty, h. mayo / financial services review 8 (1999) 223–234 pii: s1057-0810(99)00011-6 calculating a family’s asset mix william reichenstein* baylor university, hankamer school of business, p.o. box 98004, waco, tx 76798-8004, usa abstract two conclusions are reached about how a family should calculate its asset mix. first, if the assets will be used to finance retirement needs, the asset mix should be based on after-tax values, because goods and services are purchased with after-tax dollars. this novel conclusion rejects current practice. the second conclusion concerns which assets and liabilities should be included in the portfolio. if the purpose of the calculation is to consider a family’s retirement needs, the asset mix should include the promises of defined-benefit pension plans and social security, and the family’s mortgage should be treated as a short bond position. also, if the family is willing to downsize or borrow against the residence, part of its value should be included in the portfolio. © 1998 elsevier science inc. all rights reserved. 1. introduction there is wide agreement that the asset allocation decision is the most important investment decision an individual or family will make. how the family breaks up its portfolio between stocks, bonds, and, perhaps, other asset classes is more important than the choice of individual securities within each asset class (see brinson et al., 1986; reilly and brown, 1997; and sharpe, 1990, to name but a few). from this perspective, one would think that there is wide agreement about how the family should calculate its asset mix. in reality, there has been surprisingly little written on this issue. the goal of this paper is to begin to address this question in a rigorous fashion. in this paper, i present the financial position of a hypothetical but typical family. i then ask, what is the best view of its asset mix? the calculation of its asset mix depends upon answers to two questions. first, what assets and * corresponding author. tel.:11-254-710-6146; fax: 11-254-710-1092. e-mail address:bill_reichenstein@baylor.edu (w. reichenstein) financial services review 7 (1998) 195–206 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00011-6 liabilities should be included in the portfolio? second, should the assets and liabilities be expressed in market values or in after-tax values? i address the second question first. if the assets are intended for income needs during retirement then i believe the asset mix should be based on after-tax values, because goods and services are purchased with after-tax dollars. current practice advocates the use of market values. so, if my thinking is correct the profession has been miscalculating families’ true asset mixes, and the measurement errors can be substantial. the other question asks what should be included in the portfolio. i present the hypothetical family’s asset mix based on several views of its portfolio—that is, decisions about what to include. the narrowest view includes financial assets only. broader views consider the family’s personal residence and mortgage, promises of a defined-benefit pension plan, or both. i then ask, what is the best view of its portfolio? 2. pre-tax and after-tax dollars for simplicity suppose mr. and mrs. jones have $1 in a stock fund held in a deductible pension plan and $1 in bonds held in a taxable account. the deductible pension plan could be any plan where the contribution is tax deferred. these include the deductible ira, 403(b), 401(k), keogh, sep-ira and others. the cost base and market value of the bonds are $1. what is their asset mix? based on market values, it is 50% stocks and 50% bonds. the use of market values ignores the fact that the pension account contains pre-tax dollars and the taxable account contains after-tax dollars. the joneses must pay for goods and services with after-tax dollars. the $1 in the pension will buy fewer goods than the $1 in the taxable account.assuming the funds are to be used to buy goods and services during retirementthen the market value of pension assets must be converted to after-tax values. this is done by multiplying pension assets by the factor (12 tn), where tn is the expected tax rate in retirement. (the used-during-retirement assumption allows us to ignore estate-planning issues.) suppose the expected return on stocks is k% and the funds will be withdrawn in retirement in n years. the expected pre-tax value of the pension in retirement n years hence is $1(11 k)n. if the expected tax rate in retirement is 35% then the expected tax liability at time n is $0.35(1 1 k)n. to find the present value of this tax liability we must determine the appropriate discount rate. this tax liability increases and decreases with the value of the pension asset. so, the riskiness of the tax liability is precisely equal to the riskiness of the pension asset. discounting this expected tax liability at k% for n years reduces its present value to $0.35; that is, {$0.35(11 k)n}/(1 1 k)n 5 $0.35. so, we can convert the $1 of pre-tax pension assets to an after-tax basis by multiplying by (12 tn). similarly, if $1 of bonds are held in the pension account, the expected pre-tax value of the pension n years hence is $1(11 i)n , where i is the expected return on the bonds. the expected tax liability n years hence is $0.35 (11 i)n. discounting at i percent for n years reduces the present value of the expected tax liability to $0.35. the argument that the discount rate for the tax liability should be k% for stocks and i% for bonds is essentially the 196 w. reichenstein / financial services review 7 (1998) 195–206 same as modigliani and miller’s (1958) capital-structure argument that the discount rate for the tax savings from debt should be the debt’s interest rate. if the expected tax rate in retirement is 35%, the jones’ after-tax asset mix is $0.65 stocks and $1 bonds, or 39.4% stocks and 60.6% bonds. before calculating the current asset mix, one must first convert deductible pension assets to after-tax dollars by multiplying by the factor (1 2 tn), where tn is the expected tax rate in retirement. generalizing, it is not appropriate to calculate the asset mix until all investments have been converted to an after-tax basis. this perspective was first presented in reichenstein (1999). this article extends that work. the next section presents the financial position of a hypothetical but typical family. it then explains how market values of assets held in deductible pensions, tax-deferred annuities, and taxable accounts can be converted to after-tax values. 3. the smith family mary and bob smith, both age 65, have no children remaining at home. bob recently retired from eds after 30 years of service. the smiths moved frequently during bob’s career. they have lived in albuquerque for the past seven years. they are in the combined state-plus-federal 35% tax bracket and expect to remain in that bracket after retirement. table 1 presents the financial position of the smiths. they hold $15,000 of exxon stocks in a taxable account. it has a cost base of $5,000. they have a stock fund worth $165,000 that is held in a deductible pension. corporate bonds, with a cost base and current market value of $50,000, are held in a taxable account. they have $70,000 in a bond subaccount held in a nonqualified tax-deferred annuity; it was funded five years ago with a $50,000 investment and has accumulated $20,000 of tax-deferred interest. table 1 presents this asset as a $70,000 bond held in a tax-deferred annuity. table 1 the smith’s assets and liabilities in market and after-tax values asset market value after-tax value savings vehicle financial assets exxon stock $ 15,000 $ 12,300 taxable account stock fund $165,000 $107,250 deductible pension bonds $ 50,000 $ 50,000 taxable account bonds $ 70,000 $ 63,000 tax-deferred annuity total $300,000 $242,550 other assets and liabilities home $250,000 $250,000 mortgage $200,000 $200,000 insurance: cash value $ 60,000 $ 51,250 contingent assets insurance: death ben. $195,000 $195,000 defined-benefit plan $300,000 $195,000 197w. reichenstein / financial services review 7 (1998) 195–206 they bought a house seven years ago for $150,000. it is currently worth $250,000, and it was recently refinanced with a $200,000 fixed-rate mortgage. the eds defined-benefit (db) plan will pay bob $26,000 a year for the rest of his life. if bob dies first, it will pay mary $13,000 a year for the rest of her life. for simplicity, i accept the irs’ current assumption that they each have 20-year life expectancies and assume that the present value of $26,000 a year for 20 years is $300,000; the discount rate is 5.93%. if bob dies today, the present value of mary’s $13,000 a year for 20 years is $150,000. the present value of the db plan should reflect the joint probability distribution for bob and mary’s life expectancies. i leave this detail to future researchers. the benefits are fully taxable and, therefore, are worth $195,000 and $97,500 after taxes [e.g., $300,000(12 0.35)]. mary is the beneficiary of a $195,000 life insurance policy on bob. it has a cash surrender value of $60,000. table 1 presents the after-tax values of the smith’s assets assuming that during retirement they will have a combined federal-plus-state tax rate of 35% and a combined capital gains tax rate of 27%. to repeat, the smiths must buy goods and pay for services with after-tax dollars. calculating the asset mix based on market values implicitly assumes that their expected tax rate in retirement will be zero. the smiths may not be certain what their retirement tax rate will be, but the implicit expectation of zero is clearly less than optimal. the exxon stock is held in a taxable account and has a cost base of $5,000. based on the 27% tax rate, the $10,000 capital gain is worth $7,300 after taxes. so, the stock is worth $12,300 after taxes—the original $5,000 plus the $7,300 capital gain. the before-tax value of an asset with an unrealized capital loss should also be converted to an after-tax value. suppose a capital asset has a cost base of $10,000, a market value of $7,000, and the loss will be realized this year while the family is in the 35% marginal tax bracket. the after-tax value of the asset is $8,050, that is, $7,0001 $3,000 (0.35). as discussed earlier, the $165,000 of deductible pension assets converts to $107,250 of after-tax assets, [$165,000 (12 0.35)]. the $50,000 of bonds held in a taxable account is already after-tax dollars. in this case, no difference exists between the cost base and the current market value. if a difference exists, however, one should convert the unrealized gain or loss into an after-tax equivalent. the bond subaccount held in a nonqualified tax-deferred annuity is a hybrid asset consisting of pre-tax and after-tax funds. in essence, it is a bond that is subject to the tax structure facing tax-deferred annuities. the original investment of $50,000 is after-tax funds. the $20,000 of tax-deferred interest is before-tax funds. it converts to $13,000 of after-tax funds. so, the annuity is worth $63,000 after taxes. current tax law allows the exclusion of up to $500,000 in capital gain on a family’s personal residence. so, no taxes are expected on the $100,000 unrealized capital gain, and the pre-tax and after-tax values of the personal residence are $250,000. the mortgage is a short bond position. the family should calculate the after-tax value of the mortgage. the smiths recently refinanced their home. so, the market value of the $200,000 mortgage is approximately equal to its book value. and, the $200,000 book value is an after-tax liability. a family may have a fixed-rate mortgage with a below current-market interest rate. if so, the market value of the mortgage would be less than its book value. the $60,000 cash surrender value of the death benefit is a hybrid security. part of it—i assume $35,000—represents principal, which is after-tax dollars. the remaining $25,000 is 198 w. reichenstein / financial services review 7 (1998) 195–206 accumulated interest, which is pre-tax dollars. the after-tax value of the cash surrender value is $51,250. 4. calculations of the asset mix table 2 presents calculations of the smiths’ stocks-bonds-real estate asset mix based on several portfolio views, that is, views about what to include in the portfolio. the first two views use market values. they are included to represent traditional approaches of calculating the asset mix and to serve as benchmarks for the corresponding after-tax asset mixes. the first two portfolio views are the ones used by peavy and sherrerd (1990) throughout cases in portfolio management. they use market values to calculate a family’s asset mix, a practice followed throughout the profession.cases, with its guideline answer, serves as the capstone to the three-year chartered financial analysts program. in addition, the surveys of consumer finance and survey of income and program participants use market values when calculating a family’s asset mix (see kennickell et al., 1997, and poterba et al., 1994). they make no distinction between the before-tax dollars in family’s keogh accounts, 401(k) plans, and so on and the (usually) after-tax dollars in family’s taxable accounts. similarly, studies that examine a family’s retirement preparedness typically compare projected consumption with projected sources including before-tax payments from defined-benefit pension plans and social security. no adjustment is made for taxes on pension and social security benefits. table 2 views of the asset mix portfolio view stocks bonds real estate market values financial assets only 60.0% 40.0% 0.0% home equity 51.4% 34.3% 14.3% after-tax values financial assets only 51.4% 48.6% 0.0% home equity 42.3% 40.0% 17.7% balance sheeta 42.3% 230.8% 88.5% scottb 25.0% 75.0% 0.0% cash flowc 42.9% 57.1% 0.0% expandedd 27.9% 37.1% 35.0% estate planninge 20.8% 35.7% 43.5% a in thousands of after-tax dollars, stocks include $12.31 $107.25. bonds are $501 $63 2 $200 mortgage. real estate is $250. total net worth is $292.55. b in thousands, the stocks are worth $119.55 and the bonds are worth $501 $631 $51.251 $195 or $359.25 after taxes. c in thousands, the stocks are worth $119.55 and the bonds are worth $159.25 or $501 $631 $51.251 $195 2 $200. d in thousands, stocks are worth $119.55. the bonds are worth $159.25 or $501 $63 1 $51.251 $1952 $200. real estate is worth $150. e in thousands, stocks are worth $119.55. bonds are worth $501 $631 $1951 $97.52 $200 or $205.5. real estate is worth $250. 199w. reichenstein / financial services review 7 (1998) 195–206 yet, consumption expenditures use after-tax funds while defined-benefit plans and social security provide before-tax funds. yuh et al. (1998b) note this deficiency that exists in all prior studies, including their own, but as yet, no one has adjusted for it. these examples demonstrate that the profession currently advocates the use of non-tax-adjusted, market values when calculating the asset mix and the adequacy of retirement income. i call the first view the market-value financial assets only view. as its name implies, it only considers the $300,000 market value of financial assets. it says the asset mix is 60% stocks, 40% bonds, and 0% real estate. the second view, which i call the home equity view, considers the financial assets and the $50,000 of home equity. prior studies that consider a family’s preparedness for retirement usually include home equity as an asset. these include: duncan et al. (1984); burns and widdows (1988, 1990); li et al. (1996); moore and mitchell (1997); and yuh et al. (1998a). bernheim (1996) excludes house equity. this view ignores $200,000 of real estate value, the $200,000 mortgage, and the cash value of the insurance policy. according to this view, the smiths’ asset mix is 51.4% stocks ($180,000/$350,000), 34.3% bonds ($120,000/$350,000), and 14.3% real estate ($50,000/$350,000). table 2 also presents these same two portfolio views using after-tax dollars. the after-tax financial assets only view considers the $242,550 of financial assets. it says the smiths’ asset mix is 51.4% stocks and 48.6% bonds. the after-tax home equity view adds the home equity to the financial assets. it says the smiths’ asset mix is 42.3% stocks, 40.0% bonds, and 17.7% real estate. let us first discuss the effects of substituting after-tax values for market values when calculating these two narrow asset mixes. at least three observations are noteworthy. first, using after-tax values reduces the stock weights in both narrow views by about 9%. a frequent rule of thumb says that each asset-class weight should remain within 10% of its target weight. that is, a 10% or larger deviation is considered substantial. by this standard, the inappropriate use of market values caused a nearly substantial measurement error. for many families, it would cause a substantial measurement error. second, advocates of the financial assets only view claim to allocate the financial assets with an “appropriate recognition” of the outside real estate exposure. such subjective treatment lacks precision. there may be as many interpretations of “appropriate recognition” as there are financial analysts. in short, this view is not wrong but it has clear limitations. third, i believe the equity only view is dead wrong. it implicitly assumes that the $200,000 mortgage serves as a perfect hedge for $200,000 of home value. stated differently, it implicitly assumes that the values of the mortgage and home are perfectly correlated. in reality, their values do not move in unison. moreover, this netting-out treatment is inconsistent with the way we treat other financial transactions. suppose someone borrows $4,000 to margin a $10,000 purchase of intel stock. we would not say he has a $6,000 exposure to intel stock. we would recognize that he has a $10,000 exposure. if intel’s stock price rises 20%, he gains $2,000. if it falls 25%, he loses $2,500. similarly, if real estate values rise 20% the smiths gain $50,000. the smiths indeed have a $250,000 exposure to real estate, not a $50,000 exposure. this position may have been first presented in reichenstein and delaney (1995) and delaney and reichenstein (1996). the fixed-rate mortgage is equivalent to a short position in corporate bonds. the smith 200 w. reichenstein / financial services review 7 (1998) 195–206 family “issued” a $200,000 fixed-rate “bond.” suppose the $50,000 corporate bond in the smiths’ portfolio is an 8% coupon bond (that was bought at par and selling at par) and the mortgage has an 8% fixed rate. this bond offsets $50,000 of the mortgage. both “bonds” have the same before-tax and after-tax interest rates and both are callable. the smiths’ true financial position is essentially the same whether they keep the corporate bond or liquidate it and prepay part of the mortgage. suppose they liquidate the bond and prepay $50,000 of mortgage. according to the home equity view, this decreases their (net) bond position by $50,000 and increases their real estate by $50,000. in reality, (net) bonds and real estate are unchanged. the financial assets only view counts the corporate bond as part of the portfolio but excludes the mortgage. yet, they are offsetting. it makes no sense to include one but exclude the other. the mortgage is part of the family’s financial assets; it is a negative asset or a liability. in short, neither view’s treatment of the mortgage is adequate. to clarify why it is not appropriate to net out the home and mortgage, consider the deductible pension, where itis appropriate to net out the before-tax market value of the pension assets and the tax liability on the pension. the market value of the pension is an individual’s asset, its tax liability is his or her liability, and the values of the pension asset and pension liability areperfectly correlated.if the value of pension assets doubles, the liability also doubles. it follows that we can net out the pension asset and pension liability. that is, we can convert before-tax pension dollars to after-tax dollars by multiplying by (12 tn). in contrast, the values of the home and mortgage are not perfectly correlated. if the home value rises 10%, the value of the mortgage does not rise a corresponding amount. in fact, historical returns on residential real estate and bonds have not been closely correlated. so, it is inappropriate to net out the two. the balance sheet view removes this misrepresentation of the smiths’ portfolio. in particular, it includes the financial assets, the $250,000 home value, and treats the mortgage as a $200,000 short bond position. it excludes the defined-benefit pension plan and life insurance. according to the after-tax balance sheet view, the smiths’ asset mix is 42.3% stocks,230.8% bonds, and 88.5% real estate. the scott view reflects the thinking of maria crawford scott (1995), editor of the (american association of individual investors)aaii journal. she believes the “portfolio should consist of financial assets that you would be willing to sell for spending money or that generate some form of spending money, either now or sometime in the future” (p. 15). as such, she includes the present value of expected defined-benefit pension payments and the cash surrender value of insurance. (actually, she considers the market values of these assets, but i substitute the after-tax values.) she excludes the home and mortgage. the home value is excluded because it is considered primarily a consumption good; it is prepaid housing. she says, “most families do not purchase homes strictly for investment purposes. usually, if you sell a home, you must buy another one to live in. if you plan to ‘downsize’ at some point, wait until you receive the sales proceeds, and then reassess and rebalance the portfolio” (p. 17). this view says the asset mix is 25.0% stocks and 75.0% bonds. in reality, scott (1995) and other studies advocate including the present value ofbefore-tax social security benefits as part of the smith’s portfolio. i advocate using the present value of after-taxsocial security benefits. i choose to ignore social security in this theoretical paper due to its unique complexities at this time. for example, many individuals’ social security payments 201w. reichenstein / financial services review 7 (1998) 195–206 are reduced if their earned income exceeds a modest level. will these restrictions continue? what changes do you expect in the social security plan, and what is the timing of the expected changes? what is the appropriate discount rate to calculate the present value of these risky government promises? what is the proper treatment of these (currently) inflation-indexed payments? in short, i agree with prior studies that a family should include the present value of social security payments in its portfolio. i differ with prior studies in that i would use the present value of after-tax payments instead of before-tax payments. in addition, i suspect that it is a lot more difficult than usually assumed to generate “good” estimates of the present value of a family’s social security payments, either before or after taxes. although the scott view has merit, i have two concerns about it. first, scott’s (1995) criterion for including or excluding an asset or liability seems to be whether it affects cash flows. she includes pensions and social security and excludes the home and mortgage. however, the mortgage affects cash flows. so, it seems to me that it should logically be included in this view. second, many individuals and couples expect to receive spending money from their home. for example, i know a couple who owns a home with an estimated market value of $750,000. when the first spouse dies, the survivor plans to sell the home and move to a condominium. in reality, their home contributes to an overexposure to real estate in their asset mix. any view that ignores the home’s value (or only considers it after the sale) would fail to recognize this overexposure. i propose two modifications of the scott view. the first i call the cash flow view. it is like the scott view except it considers the mortgage. according to this view, the smiths’ asset mix is 42.9% stocks and 57.1% bonds. this mix differs substantially from the scott view’s asset mix of 25% stocks and 75% bonds. i believe the cash flow view provides a useful perspective for a family (or a surviving spouse) that would not be willing to sell the home or borrow against it. in particular, i agree with scott (1995) that the home should be excluded when considering the smiths’ ability to pay for non-housing consumption needs during retirement. however, the cash flow view would not be appropriate if the family would be willing to sell the home or borrow against it. the second modification of the scott view is the expanded view. assume the smiths expect to downsize and move into a $100,000 home. they should consider $150,000 of the $250,000 home as real estate and ignore the other $100,000. in essence, the $100,000 is prepaid housing. if the smiths are willing to borrow against their home, they should include the home equity in the family’s portfolio. the expanded view includes financial assets, $150,000 of home, mortgage, cash surrender value, and present value of defined-benefit pension plan. this asset mix is 27.9% stocks, 37.1% bonds, and 35.0% real estate. table 2 also presents the estate planning view. it reflects after-income-tax values at bob’s death. it includes financial assets, the home and mortgage, the death benefit of bob’s life insurance policy, and the present value of the after-tax value of mary’s $13,000 a year payment from the db plan. it excludes the cash surrender value. this view of the smiths’ asset mix is 20.8% stocks, 35.7% bonds, and 43.5% real estate. an estate planning view could also be developed for the less likely scenario that mary should die first. it would exclude the death benefit but include the cash value of bob’s life insurance. 202 w. reichenstein / financial services review 7 (1998) 195–206 5. which portfolio view is “best”? the answer to this question depends in part on the purpose of the calculation. naturally, if the calculation is for estate-planning purposes then the estate planning view is best. but let us look at the range of asset weights from the other after-tax views, which assume the smiths enjoy a happy retirement. stocks’ weight ranges from 25.0 to 51.4%. bonds’ weight ranges from230.8% to 75%, or more than 100%. real estate’s weight ranges from 0% to 88.5%. these ranges are huge. moreover, these dramatic differences are due entirely to differences in opinion about what belongs in the portfolio. objections about how i calculate an asset or liability’s value would not change the reality that the asset mix changes dramatically with the decision about what belongs in the portfolio. which view is best? one way to shed light on this question is to see if the asset mix changes substantially when the family’s true financial position changes substantially. table 3 presents asset mixes based on three substantially different financial positions. the base financial position is the one described thus far. the smiths have a defined-benefit pension plan worth $195,000 and a $200,000 mortgage. in the no-db-plan financial position, the smiths have the $200,000 mortgage but do not have the db pension plan. in the no-mortgage financial position, they have the db pension plan but do not have the $200,000 mortgage; the mortgage has been paid off. i believe these three financial positions are substantially different. as such, i believe one test of a realistic portfolio view is that its asset mix should be substantially different for each financial position. table 3 asset mixes across three financial positions financial positions stocks bonds real estate financial assets only view base $119.55 (51.4%) $113 (48.6%) $0 (0.0%) no db plan $119.55 (51.4%) $113 (48.6%) $0 (0.0%) no mortgage $119.55 (51.4%) $113 (48.6%) $0 (0.0%) balance sheet view base $119.55 (42.3%) 2$87 (230.8%) $250 (88.5%) no db plan $119.55 (42.3%) 2$87 (230.8%) $250 (88.5%) no mortgage $119.55 (24.8%) $113 (23.4%) $250 (51.8%) scott view base $119.55 (25.0%) $359.25 (75.0%) $0 (0.0%) no db plan $119.55 (42.1%) $164.25 (57.9%) $0 (0.0%) no mortgage $119.55 (25.0%) $359.25 (75.0%) $0 (0.0%) cash flow view base $119.55 (42.9%) $159.25 (57.1%) $0 (0.0%) no db plan $119.55 (142.7%) 2$35.75 (242.7%) $0 (0.0%) no mortgage $119.55 (25.0%) $359.25 (75.0%) $0 (0.0%) expanded view base $119.55 (27.9%) $159.25 (37.1%) $150 (35.0%) no db plan $119.55 (39.2%) 2$35.75 (211.7%) $150 (49.1%) no mortgage $119.55 (19.0%) $359.25 (57.1%) $150 (24.9%) in the base financial position, the smiths have a defined-benefit (db) pension plan worth $195,000 and a $200,000 mortgage. in the no-db scenario, they do not have the defined-benefit pension plan. in the no-mortgage scenario, they have paid off their mortgage. 203w. reichenstein / financial services review 7 (1998) 195–206 the financial assets only view paints the same picture of the smiths’ financial position in all three scenarios. the balances sheet view paints the same picture whether or not bob smith is eligible for the db pension. the scott view paints the same financial picture whether the mortgage balance is $200,000 or zero. although not shown, the home equity and estate planning views also fail this reality check. the cash flow and expanded views are the only ones to pass this reality check. therefore, when estimating cash flow needs i believe they provide the best measures of the family’s true financial position and are the best views of the family’s asset mix. the cash flow view appears best if bob and mary (or the surviving spouse) are not willing to sell the home or to borrow against it. the expanded view appears best if they (or the surviving spouse) are willing to sell the home or borrow against it. one might ask if it makes any difference how we calculate the smiths’ asset mix as long as we know their true financial position. and, their true financial position does not change with the portfolio view. i believe the answer is yes, because it affects investment decisions. suppose mary smith inherits $100,000 after taxes. in which asset class should she invest the funds? bogle (1994), among others, advocates a popular rule of thumb that says the portfolio’s stock weight should be about 100 minus the investor’s age. for simplicity, let us assume that this rule of thumb fits the smiths’ risk tolerance. at 65, it suggests that the smiths should have a 35% stock exposure. now consider the asset mixes as measured by portfolio views. the market-value financial assets only view, which has been the most popular view in practice, says the smiths have a 60% stock exposure. it says they are overexposed to stocks and should invest all of the $100,000 in bonds. the after-tax financial assets only view also says they are overexposed to stocks and should invest the $100,000 in bonds. the after-tax cash flow view says they have a 42.9% stock exposure or about 8% above the 35% target. to achieve the 35% target, they should invest about $14,000 in stocks. if the expanded view is appropriate, the smiths have a 27.9% stock exposure. in this case, they should invest about $66,000 in stocks. in sum, the appropriate investment decision for their risk tolerance depends upon their portfolio view. the portfolio view affects investment decisions and, therefore, it is important. 6. summary there is wide agreement that the asset allocation decision is the most important investment decision an individual or family will make. yet, there is little written about what assets and liabilities should “count” in the family’s portfolio. moreover, prior studies have not adjusted asset values for taxes, so that they reflect after-tax purchasing power. in this paper, i presented the financial position of the smiths, a family that has a simple portfolio. in addressing these issues, i reached the following conclusions. first, if the assets are intended for retirement then the asset mix should be stated in after-tax values because goods and services are purchased with after-tax dollars. second, the best view of the portfolio depends upon the purpose of the calculation. if the purpose is for estate planning, the estate planning view is best. but what if the purpose is 204 w. reichenstein / financial services review 7 (1998) 195–206 to calculate the family’s cash flow needs during retirement? based on a criterion established in this paper, i conclude that the best portfolio view (or even a good view) must include the promises of a defined-benefit pension plan. it also must include the family’s mortgage— something that the profession has seldom done when calculating the asset mix. to be specific, i recommend either the cash flow or the expanded view. these views consider the present value of after-tax pension payments to be a “bond” in the family’s portfolio. similarly, the present value of after-tax social security promises is a “bond.” the recommended views treat the mortgage as a short position in bonds. suppose a family has $100,000 in 8% coupon corporate bonds (held in a taxable account) and a $100,000 mortgage with an 8% interest rate. its financial position would be the same whether it retains both or liquidates the bonds and prepays the mortgage. it is, thus, inconsistent to consider the bonds part of the portfolio and to exclude the mortgage. similarly, the common practice of including “home equity” in the asset mix produces logical inconsistencies. the difference between the two recommended views is whether the family’s personal residence should be considered part of the portfolio. if the family is not willing to sell the residence and downsize or to borrow against it, i believe it should not be considered part of the portfolio. in this case, the family’s residence is considered pre-paid housing. if the family is willing to sell the residence and downsize, i believe it should include in its portfolio the after-tax value of the freed funds from downsizing. if it is willing to borrow against the residence, the portfolio should include the home equity. finally and most importantly, i offer my views to encourage response. i am sure that the issues addressed in this paper are important to my family and to others. i am much less sure that my answers are “best” or even “good.” acknowledgments i thank two anonymous reviewers and karen eilers lahey for helpful comments. references bernheim, b. d. (1996). the merrill lynch baby boom retirement index: update ’96. stanford university, merrill lynch. bogle, j. c. (1994).bogle on mutual funds. new york: irwin press. brinson, g. p., hood, r., & beebower, g. (1986). determinants of portfolio performance.financial analysts journal 43(4), 39–44. burns, s. a., & widdows, r. (1988). an estimation of savings needs to adequately fund baby boomers’ retirement. in v. hampton (ed.),proceedings of the 34the annual conference of the american council on consumer interests(pp. 15–18). columbia, mo: american council on consumer interests. burns, s. a., & widdows, r. (1990). sensitivity of a retirement analysis framework to changes in retirement analysis parameters.financial counseling and planning 1, 71–91. delaney, c., & reichenstein, w. (1996). an expanded portfolio view includes real estate and human capital.aaii journal 18(6), 7–11. duncan, g. j., mitchell, o. s., & morgan, j. n. (1984). a framework for setting retirement savings goals.journal of consumer affairs 18(1), 22–46. 205w. reichenstein / financial services review 7 (1998) 195–206 kennickell, a. b., starr-mccluer, m., & sunden, a. e. (1997). family finances in the u.s.: recent evidence from the survey of consumer finances.federal reserve bulletin(january), 1–24. li, j., montalto, c. p., & geistfeld, l. v. (1996). determinants of financial adequacy for retirement.financial counseling and planning 7, 39–48. modigliani, f., & miller, m. h. (1958). the cost of capital, corporate finance and the theory of investment. american economic review 48, 261–297. moore, j. f., & mitchell, o. s. (1997).projected retirement wealth and savings adequacy in the health and retirement study. pension research council working paper 98-1. the wharton school of the university of pennsylvania, philadelphia. peavy, iii, j. w., & sherrerd, k. f. (1990).cases in portfolio management.charlottesville, va: association for investment management and research. poterba, j. m., venti, s. f., & wise, d. a. (1994). targeted retirement saving and the net worth of elderly americans.american economic review 84(2), 180–185. reichenstein, w. (1999).savings vehicles and taxation of individual investors.charlottesville, va: research foundation of the institute of chartered financial analysts. reichenstein, w., & delaney, c. (1995). portfolio management: an alternative approach to asset allocation. journal of investing 4(3), 57–62. reilly, f. k., & brown, k. c. (1997).investment analysis and portfolio management(5th ed.). orlando, fl: dryden press. scott, m. c. (1995). defining your investment portfolio: what should you include?aaii journal 17(10), 15–17. sharpe, w. f. (1990). asset allocation. in j. l. maginn & d. l. tuttle (eds.).managing investment portfolios: a dynamic process(2nd ed., pp. 1–71). charlottesville, va: association for investment management and research. yuh, y., montalto, c. p., & hanna, s. (1998a). are americans prepared for retirement?financial counseling and planning 9(1), 1–12. yuh, y., hanna, s., & phillips montalto, c. (1998b). mean and pessimistic projections of retirement adequacy. financial services review7(3), 00–00. 206 w. reichenstein / financial services review 7 (1998) 195–206 financial services review, 33(3) i guest editorial navigating contemporary fintech solutions: revealing potential and challenges peter öhman and mustafa nourallah centre for research on economic relations (cer), mid sweden university izidin el kalak alfaisal university, ksa towards fintech advancements individuals, families, businesses, and communities have witnessed significant advancements in the field of financial technology (fintech), which shows promising prospects but also substantial challenges (fulk et al., 2018). according to the international monetary fund (2023), regions worldwide have seen an increase in the use of digital financial services. in 2014, 44% of people worldwide made or received a digital payment; this percentage increased to 52% in 2017, and by 2021, the statistics estimated that almost 64% of the world's population used digital payments (demirgüç-kunt et al., 2022). exploring the origins of fintech solutions takes us on a journey back to the 15th century, when financial transactions were carried out using coins. it also directs our attention to analog technologies that enable people to extract information from documents and transmit it across long distances (alt, 2018). the launch of smartphones in 2007 represents a significant breakthrough in fintech history and led to what is now known as the fintech revolution. advanced technologies, such as blockchain and machine learning, have given rise to new generations of fintech solutions that enable individuals, and businesses, to make informed financial decisions and manage their finances more efficiently (nguyen et al., 2023). the financial industry is currently experiencing the development of contemporary fintech solutions, including financial robo-advisors and the potential issue of central bank digital currency (cbdc). these solutions, among others, will likely reshape the financial landscape (nourallah et al., 2025). qin et al. (2024) suggested a strong and positive correlation between fintech and the green environmental index, while d'acunto & rossi. (2023) argued that financial robo-advisors may help improve financial literacy, and andolfatto et al. (2021) predicted a positive role of cbdc in enhancing financial inclusion. in parallel, these solutions come with their own set of challenges, including behavioral issues related to emerging e-banking technologies (gomber et al., 2017), and contemporary solutions can also be associated with behavioral biases (barber et al., 2022; welch, 2022). in their turn, bartlett et al. (2022) found that fintech lenders’ fees could be higher for minority borrowers compared to non-minority ones, while recent studies shed light on issues related to the ‘one size financial services review, 33(3) ii fits all’ approach used by financial robo-advisors and asked for considering cultural differences when developing these solutions (nourallah, 2023). moreover, das (2019, p. 1004) highlighted the dilemma of ”garbage-in, garbage-out”, stating that ‘more data does not mean better results’. berg et al. (2019) further argued that even basic digital footprint information, such as the type of mobile system, can disclose important information about individuals’ financial behavior and predict their default. this special issue aims to explore the advancements offered by contemporary fintech solutions while also examining the concerns associated with these solutions. it also encourages discussions beyond the traditional work of fraud and security, addressing various concerns surrounding these solutions. contributions to this special issue we are grateful for the contributions submitted to our special issue. following the peerreview process, four papers have been accepted for publication. the first study by brockbank et al. used human capital theory to examine whether the use of financial advisors is associated with household investment in cryptocurrencies, using data from the 2018 national financial capabilities study (nfcs) in the united states. brockbank et al. find that households consulting financial advisors are more likely to invest in cryptocurrencies. additionally, they discover that younger individuals, married people, and the ones with higher subjective financial literacy are more likely to hold cryptocurrency investments compared to other investors. the second study examines how the use of fintech influences consumers' emergency fund savings, applying the theory of planned behavior and technology adoption models. using data from 453 u.s. respondents collected in july 2021, chen et al. employed structural equation modeling to examine the links between intention to use fintech and actual saving outcomes. the results show that while positive attitudes toward emergency savings surprisingly reduced the intention to use fintech, subjective norms and perceived behavioral control increased it. perceived behavioral control also directly supported having adequate savings. the intention to use fintech was linked to saving apps and websites, but only website use significantly improved emergency savings. the study concludes that both the intention and the actual use of fintech tools are critical pathways connecting psychological factors to successful saving behavior, providing practical insights for financial institutions, advisors, and policymakers seeking to enhance household financial resilience. in the third study, diab et al. utilized self-control theory and balanced panel data from 2014, 2017, and 2021 to investigate the factors influencing the tendency to borrow from family and friends in the european union, highlighting the crucial role of debit card usage and saving habits in mitigating this behavior. this study questions the effectiveness of current public financial education initiatives and underscores the need for more effective policy development within the evolving fintech landscape. by shedding light on these findings, the paper increases our knowledge of the relationship between debit card usage and borrowing behavior within the european union. last but not least, shekhar & ramesh investigate how cbdc design impacts implementation in emerging economies with strong digital infrastructures. based on interviews financial services review, 33(3) iii with 22 experts, they find that 1) a two-tier, non-interest-bearing model with offline access supports stability and inclusion, 2) cbdcs should complement, not replace, existing payment systems, and 3) phased implementation with clear metrics and partnerships ensures adoption. a four-layer design framework highlights the interplay of technology, security, finance, and user experience. the findings guide policymakers to prioritize interoperability, stability, and efficient integration. conclusion and future research avenues technological advancements have fundamentally reshaped the structure of the broader financial sector (nourallah & öhman, 2021). these transformations have created promising opportunities for delivering innovative and high-quality financial services, while simultaneously introduced new challenges that warrant careful investigation. this special issue focuses on fintech, exploring its emerging opportunities and associated challenges. based on the findings from the four empirical studies, it is evident that fintech plays a dynamic role in the lives of individuals, families, businesses, and communities. one notable trend is the increasing interest in incorporating cryptocurrencies into household investment portfolios in the united states (brockbank et al., 2025). another important finding from the same part of the world highlights the role of fintech platforms in enhancing households' ability to build and maintain emergency funds (chen et al., 2025). meanwhile, in the european context, evidence suggests that the use of structured financial tools can mitigate harmful borrowing behaviors, such as informal lending from friends and family (diab et al., 2025). in the context of india, a conceptual framework for the potential implementation of cbdc is proposed (shekhar & ramesh, 2025), offering insights into the institutional and technological requirements for its adoption. looking ahead, future research in fintech could focus on three key areas that are particularly significant. while considerable research has examined the impact of fintech applications on improving access to financial services and advancing financial inclusion, there seems to be a need for studies on their role in enhancing individuals' capacity for sound financial decision-making. this gap underscores the importance of future research aimed at investigating how fintech applications can support users in making informed, responsible, and contextually appropriate financial choices. a second research area concerns the potential of fintech applications to improve financial well-being and mitigate financial stress. given the integration of these tools into everyday financial life, it seems important to assess whether and how they contribute to individuals’ sense of control, stability, and confidence in their financial decisions. the third and most forward-looking area relates to the role of artificial intelligence (ai) in the continued development of fintech applications. the integration of ai offers the potential to overcome many of the shortcomings and risks previously identified in the fintech literature. this could lead to more personalized, adaptive, and ethically grounded financial tools that extend beyond the boundaries of traditional financial applications. nevertheless, these tools still require regulation by human-established authorities. financial services review, 33(3) iv references andolfatto, d. (2021). assessing the impact of central bank digital currency on private banks. the economic journal, 131(634), 525–540. alt, r., beck, r., & smits, m. t. (2018). fintech and the transformation of the financial industry. electronic markets, 28, 235–243. barber, b. m., huang, x., odean, t., & schwarz, c. (2022). attention‐induced trading and returns: evidence from robinhood users. the journal of finance, 77(6), 3141–3190. bartlett, r., morse, a., stanton, r., & wallace, n. (2022). consumer-lending discrimination in the fintech era. journal of financial economics, 143(1), 30–56. berg, t., burg, v., gombović, a., & puri, m. (2020). on the rise of fintechs: credit scoring using digital footprints. the review of financial studies, 33(7), 2845–2897. brockbank, a, kalenkoski, c., browning, c., & guillemette, m. (2025). is using a financial advisor related to cryptocurrency investment? financial services review, 33(3), 1–19. chen, y., asebedo, s. d., little, t. d., & ning, w. (2025). from intention to adequate emergency fund savings through fintech use: evidence from a u.s. survey study. financial services review, 33(3), 20–47. d'acunto, f., & rossi, a. g. (2023). robo-advice: transforming households into rational economic agents. annual review of financial economics, 15(1), 543–563. das, s. r. (2019). the future of fintech. financial management, 48(4), 981–1007. diab, s., nourallah, m. & öhman, p. (2025). borrowing from family and friends: study of the european union. financial services review, 33(3), 48–60. demirgüç-kunt, a., klapper, l., singer, d., & ansar, s. (2022). the global findex database 2021. the world bank. https://doi.org/10.1596/978-1-4648-1897-4 fulk, m., grabel, j., watkins, k., & kruger, m. (2018). who uses robo-advisory services, and who does not? financial services review, 27(2), 173–188. gomber, p., koch, j. a., & siering, m. (2017). digital finance and fintech: current research and future research directions. journal of business economics, 87, 537–580. international monetary fund. (2023). financial access survey 2023 trends and developments. data.imf.org/fas nguyen, d. k., sermpinis, g., & stasinakis, c. (2023). big data, artificial intelligence and machine learning: a transformative symbiosis in favour of financial technology. european financial management, 29(2), 517–548. nourallah, m. (2023). one size does not fit all: young retail investors’ initial trust in financial robo-advisors. journal of business research, 156, 113470. nourallah, m., & öhman, p. (2021). impact of advanced technologies on consumer finance and retail investment: mobile bank applications and robo-financial advisors. in impact of globalization and advanced technologies on online business models (pp. 1–15). igi global. financial services review, 33(3) v nourallah, m., öhman, p., walther, t., & nguyen, d. k. (2025). financial robo-advisors: a comprehensive review and future directions. available at ssrn https://dx.doi.org/10.2139/ssrn.5215748 qin, l., aziz, g., hussan, m. w., qadeer, a., & sarwar, s. (2024). empirical evidence of fintech and green environment: using the green finance as a mediating variable. international review of economics & finance, 89, 33–49. shekhar, v. & ramesh, s. (2025). central bank digital currency: perspectives on design choices and implications, with a focus on e-rupee. financial services review, 33(3), 61–79. welch, i. (2022). the wisdom of the robinhood crowd. the journal of finance, 77(3), 1489– 1527. https://dx.doi.org/10.2139/ssrn.5215748 pii: s1057-0810(96)90026-8 financial services review, 5(l): 43-56 copyright 8 1996 by jai press inc. issn: 1057-0810 all rights of reproduction in any form rcscrved. churning: excessive trading in retail securities accounts stewart l. brown churning involves excessive trading by stockbrokers in order to generate commissions. current practice uses the turnover ratio to detect excessive trading. the turnover ratio is a flawed indicator of the actual hartn of excessive trading which is commissions. this paper examines the intersection of law and financial analysis in the retail securities arena. a unique set of data from 23 actual churning cases is used to argue that the tum over ratio should be replaced by a more direct measure of the trading costs: the commission to equity ratio. an appropriate benchmark related to the return on common stocks is suggested to gauge excessive trading in a commission context. i. introduction individual investors sometimes invest through retail brokerage firms. one advantage of this approach is that the investor gains the counsel of a stockbroker who has expertise in matters relating to investments. unfortunately, investor’s experiences with brokerage firms are sometimes unsatisfactory, and there has been an explosion of lawsuits against stockbro kers in the last 15 years. two types of cases constitute the majority of such claims: suitabil ity claims and churning claims. suitability cases involve allegations that the broker made recommendations of unsuitable securities. churning cases involve allegations that the bro ker over-traded the account in order to generate commissions and not to benefit the client. the genesis of claims against stockbrokers for wrongful conduct is the anti-fraud rule, rule (lob+, under the securities exchange act of 1934. this rule is supplemented by rules promulgated by self regulatory organizations (sros), such as the new york stock exchange and the national association of securities dealers. for a stockbroker to become registered with the sec, or become a member of a sro, he or she must agree to follow “just and equitable principles of trade.” churning claims arise out of the inherent conflict of interest involved in the normal procedure used to compensate stockbrokers. because stockbroker compensation is typi cally based upon the volume, size, and type of transactions, customers suffer higher costs and lower returns when unnecessary transactions take place in the account. churning stewart l. brown l florida state university, tallahassee, fi 32306. 44 financial services review 5( 1) 1996 claims involve allegations that a broker abused his or her position by trading too often in order to generate commissions. churning is essentially an agency/conflict of interest issue (anderson, heacock, & hill, 1987). the purpose of this paper is two fold. first, current legal practice uses a flawed finan cial ratio, the turnover ratio, to detect excessive trading. this paper will critique the tum over ratio and develop a better metric of excessive trading: the commission to equity ratio. second, the paper introduces academics and practitioners who deal with individual inves tor issues to the concept of churning and how to detect excessive trading. most academics are familiar with the general concept but may have little practical knowledge. the paper will present the important issues in sufficient depth that an interested reader will be better informed in the classroom and in other academic and practical arenas. although dealing peripherally with legal issues, this paper is not a legal paper. attor neys may find the paper of interest, however, there will be a minimum of legal references. interested readers are referred to goldberg (1991), jensen (1991), loss and seligman (1989), and winslow and anderson (1990) for voluminous case citations. ii. elementsofchurningcases one definition of churning is as follows: “churning occurs when a securities broker engages in excessive trading in disregard of his customer’s investment objectives for the purpose of generating commission business,” (loss & seligman, 1989, p. 3874). there are three elements of a churning case: control, excessive trading, and scienter (fraudulent intent). a. control if a stockbroker is to be found liable for churning an account, the finder of fact (judge, jury, or arbitration panel) must first determine that the broker was in control of the account. typically, control is the most hotly contested element in churning cases. there are two types of control: express and implied. the best example of express control over a brokerage account is when there is a writ ten discretionary trading authorization. in such an account, the stockbroker receives blan ket permission by the customer to execute trades at the stockbroker’s discretion. control is typically established with little contest when there is a written discretionary trading autho rization present. however, this is rarely the case. in the absence of express control, implied or “de facto” control must he demonstrated by credible evidence in the form of written and oral testimony. the finder of fact must make a determination about who is ultimately in control of trading or who is “calling the shots.” there are certain factors and customer characteristics which the finder of fact usually considers in determining whether the stockbroker had sufficient control to call the shots for an account (goldberg, 1991). these factors include the customer’s sophistication, prior securities experience, and the amount of trust and confidence placed in the stockbroker. it is also relevant to examine whether the client conducted independent research, read current financial publications, and whether the client was provided with totally truthful informa churning 45 tion. in examining the trading that took place, if a high percentage of transactions were rec ommended by the broker and the client usually acquiesced, then the broker may have been in control of the account. the fact that the client occasionally executed a trade or refused a broker’s recommendation does not establish that the client was in control. b. excessive trading it is well established in case law that excessive trading may only be gauged in light of the nature of the account, the dominant element of which is the investment objective of the client. in determining the investment objective, it is sometimes useful to examine new account forms for indicated investment objectives. unfortunately, there is little standard ization among firms, and categories vary widely. common categories include safety, con servation of principal, income, safety and income, growth, growth and income, aggressive income, aggressive growth, trading, speculation, and long term growth. the problem is compounded by the frequent occurrence that more than one category is checked and these are often conflicting. it is often more fruitful to look at the financial situation of the client. for instance, on one end of the spectrum, an elderly client with a retirement account investing irreplaceable funds should have conservative investment objectives. on the other end of the spectrum, a young executive with a good income, investing a small percentage of his or her wealth, could well have more speculative investment objectives. another consideration is the sophistication and knowledge of the client. more sophisticated and knowledgeable clients can generally tolerate higher levels of risk. ultimately, a determination of the appropriate level of risk tolerance in the account must be made in order to determine if the account has been excessively traded: the degree of tolerance of risk among clients is obviously a seamless spectrum; nonetheless, for legal purposes, case law seems to have coalesced around three general categories, as follows: 1. conservative-low risk. accounts of this type generally require safety and conservation of principal. common investment objectives include conservation of principal, income, and perhaps growth and income. such accounts can ill afford excessive trading costs and low risk securities are generally appropriate. 2. investment-medium risk. such accounts are ordinary investment accounts. although not classified as speculators, this type of investor is willing to assume some degree of investment risk. common investment objectives include growth and long term growth. a buy and hold strategy is appropriate although such accounts can afford some trading and medium risk securities are generally appropriate. 3. speculative+high risk. these accounts are the most aggressive. common investment objectives include trading, aggressive growth, and speculation. sub stantial trading is suitable in such accounts and high risk securities are generally not unsuitable. once a determination of the risk tolerance and general nature of the account is made, a quantitative analysis of the trading in the account is conducted. the analysis typically includes a calculation of the turnover ratio in the account and other, more direct measures financial services review 5( 1) 1996 of trading costs. it is also common to examine the holding periods of the securities in the account. case law has developed certain standards for gauging the excessiveness of trading in relation to the nature of the account. these measures and standards of excessive trading will be examined below. c. scienter the final element of a churning claim is scienter or “evil intent.” once it has been determined that the broker controlled the account and that the account was excessively traded, it must also be established that the stockbroker traded the account with the intent to defraud, or at least with reckless disregard of the best interests of the client. most courts simply infer this element when the first two have been established (goldberg, 1991). iii. measures and legal standards of excessive trading a. the turnover ratio until recently, most courts have dealt with the issue of excessive trading by examining the turnover ratio, which generally represents the number of times the equity in the account is liquidated and reinvested in a given period. typically, turnover ratios are calculated on an annual basis. for instance, a nonmargined account would experience an annual turnover rate of one if all the investments in the account were sold and the proceeds reinvested dur ing the same year. turnover may be calculated using periods other than exactly one year by using the following formula: turnover ratio = total purchases x 12 average rquity x n ’ the turnover ratio divides total purchases over the period in question by the average account equity. this number, which represents total turnover, is then annualized by divid ing by the total number of months (n) in the period in question and multiplying by 12. the period of analysis is normally in months for retail securities accounts since state ments are typically produced monthly. the ratio can be annualized using the number of days in the analysis and 365 in the numerator. the turnover ratio normally looks at purchases in the account and ignores sales. the ratio may be calculated with the average of purchases and sales in the numerator or with the smaller of purchases and sales. in addition, in an institutional context, adjustments are commonly made for increases or decreases in funds in the account (schreiner, 1980). under normal circumstances these adjustment result in only minor changes in the calcu lated turnover ratio. the formula presented is consistent with current case law. account equity is the amount the client would realize if all securities in the account were liquidated and margin loans repaid. conceptually, average account equity is the amount of money available to the stockbroker to turnover the account and thus to generate commissions. churning 47 b. legal standards for determung excesslve trading in gauging whether or not the trading in a particular account is excessive, certain stan dards have been established in case law. goldberg (1991) suggests that an annualized turn over ratio in excess of two times is indicative of active trading, a ratio of four or greater raises the presumption of excessive trading, and when the turnover ratio is greater than six the presumption becomes conclusive. this has become lmown as the 2-4-6 formulation. several early cases, including sec cases, found excessive trading in accounts with conservative investment objectives that had turnover ratios in the neighborhood of two times on an annual basis. other cases find excessive trading in normal investment accounts with turnover ratios in the neighborhood of four times, and turnover ratios in excess of about six times are typically sufficient to infer excessive trading even for accounts with speculative investment objectives. this position is buttressed by the seminal law review article entitled “churning by securities dealers,” in the hurvard law review (1967, p. 876) which contains the statement: the turnover rates found to be excessive vary widely . . . while few cases involved tum overs as frequent as one per month, turnovers averaging once every other month have occurred inore frequently. since the amount of activity permissible will vary with indi vidual circumstances, these figures cannot be a erm guide; smaller turnovers may be objectionable in some cases and larger ones permissible in others. nonetheless, it is pos sible to generalize from the sec cases that a complete turnover more than once every two months is likely to be labeled excessive, and this conclusion appears reasonable. this passage appears to the be genesis of much subsequent case law. it follows that a turnover of six is indicative of excessive trading even for accounts with very aggressive investment objectives. it is common for excessive trading to be found in accounts with turnover ratios of less than six but also with more conservative investment objectives. thus, a turnover ratio of two might be excessive in the account of a customer with conser vative investment objectives but not excessive in the account of a customer with specula tive objectives. the argument has been made that the turnover standards are out of date. winslow and anderson (1990) argue in favor of a different standard for judging whether an account has been excessively traded. they suggest that the proper standard of comparison is the aver age turnover ratio of a group of mutual funds with investment objectives similar to the cus tomer in question. the argument is that the turnover in accounts managed by professionals such as mutual fund managers is driven solely by professional judgment and not the desire for commissions. brokers hold themselves out as financial experts with skill in managing money, and brokers also earn commissions when securities are traded. thus, it is relevant to compare trading activity generated by brokers to the trading activity generated by pro fessionals who have no conflict of interest. table 1 summarizes the winslow and anderson mutual fund turnover results. average turnover rates range from 0.53 for the most conservative group of funds to 1.18 for aggressive growth funds. the average of the average turnover rates in their sample is 0.8 times on an annual basis. mutual funds with different investment objectives exhibit substantially different average turnover ratios. for instance, the three fund categories that could be characterized as conservative (income, balanced, and equity income) had an 48 financial services review 5( 1) 1996 table 1 fund category and turnover statistics fund category mean turnover rare aggressive growth 1.18 balanced 0.66 equity income 0.70 growth 0.98 growth-income 0.53 international 0.55 option-income 1.45 small company 0.54 income 0.58 average 0.80 source: winslow & anderson (1990). sd mean plus 2 sd 0.72 2.62 0.58 1.81 0.53 1.76 0.61 2.19 0.55 1.64 0.42 1.38 0.74 2.93 0.39 1.32 0.40 1.39 average turnover ratio of 0.65 times. given the standard deviations around these means, essentially all of the more conservative funds were turned over less than two times. funds with more aggressive investment objectives (growth-income, growth, and aggressive growth) had an average turnover of 0.9 times, and standard deviations indicate that essen tially all of these funds had turnover ratios less than three times. the standard deviation numbers should be interpreted with some caution since it is unlikely that the distribution of turnover ratios in normal. the turnover ratio is bounded on the downside by zero, and thus the distribution is likely to have a positive skew. given that professional money managers turn over their portfolios an average of less than one time per year, case law standards, that is, the 2-4-6 formulation appears overly gen erous to brokers. this is especially true when it is noted that retail securities customers pay much higher commissions than institutions such as mutual funds. winslow and anderson (1990, p. 357) conclude that a lower hurdle to demonstrate excessive trading is warranted: we believe that some use of a weaker form of presumption would be helpful to the courts. depending to some extent on the specific investment objectives as reflected in the mutual fund data . . . an appropriate annual turnover rate should be seen as lying in the neighborhood of one; rates increasing beyond that should be viewed with skepti cism. as the rate increases much beyond one, the broker should bear the burden of explaining the higher than normal rate. once rates rise to about three, there should be little room for argument about the excessive trading element of the claim, in the absence of an investment motive or strategy that is not accounted for in our data. a recent supreme court decision (mcmahon v. shearson lehman, 1987) forced most securities cases into arbitration and, with few exceptions, case law is frozen in essentially the same place it was in 1987. thus, the winslow and anderson argument has not been fac tored into case law. c. problems with the turnover measure the turnover ratio is an indirect measure of trading activity and trading costs. the harm of excessive trading is not transactions volume per se but the unnecessary trading costs imposed on an account. moreover, the motivation for excessive trading is commis sions which accrue to the stockbroker and the brokerage firm. if commission rates were churning 49 fixed, as they were during the period when case law standards evolved, then the turnover ratio would have a fixed relationship with trading costs, and there would be no problem in using turnover ratios to gauge excessive trading. however, commission rates are now com petitive, and there has been a proliferation of other types of securities traded in brokerage accounts, many of which carry higher commission rates than listed equities. for instance, it is not unusual to see mutual funds carried in retail brokerage accounts. such instruments carry front end or back end loads, typically in the range of 4-8%. thus, a complete turnover (purchase and sale) of mutual funds could cost perhaps 8% while a com plete turnover in listed equities might cost 3% if one-way commissions are 1.5%. thus, similar trading volume in listed equities and mutual funds would generate similar turnover ratios with much different trading costs. it has become common place to see exchange traded options in retail securities accounts. commission rates on options are typically in the range of 3-7% of the transaction amount (see appendix). for certain types of options strategies, the turnover ratio may dra matically understate the level of trading activity. for instance, if the strategy used is to sell calls and allow them to expire or be exercised, then the turnover rate associated with options trading would be zero since no purchases would occur. this problem has been rec ognized,(sec special options study, 1968). another potential problem with using turnover to measure excessive trading occurs when principal trades on otc stocks take place in the account. such trades carry no explicit commissions, but there are trading costs, and stockbrokers are compensated in the form of sharing a part of the markup on the security. the national association of securities dealers has a policy which suggests that profits on principal transactions should normally not exceed 5%. the policy was adopted after a survey of mark-ups charged by nasd members in retail transactions indicated that 47% of the transactions were marked up less than 3%, another 24% were marked up between 3 and 5% and the remainder were marked up more than 5% (nasd, 1944). commission on pink sheet stocks sometimes run as high as 15-202 per trade. in summary, many securities that are commonly traded today carry different and gen erally higher commission rates than the commission rates on listed equities. the turnover ratio and associated case law standards may have worked well in a simpler era when listed equities constituted the vast bulk of trading in retail securities accounts. this is no longer the case, and as a result, the turnover ratio is a flawed measure of the degree of excessive trading in retail securities accounts. what is needed is a direct measure of trading costs in retail securities accounts. the next section presents such a measure. d. a better measure of excessive trading-the commission to equity ratio the 1968 sec special options study suggests the commission to equity (c/e) ratio as an alternative to the turnover ratio. the suggestion was made in the context of options trad ing but is equally applicable to all securities trading. it has the advantage of measuring trading costs directly. the c/e ratio is calculated by dividing the total commissions in the account by the average equity in the account and then annualizing the number, that is: ueratio = total commissions x 12 average equity x n * (2) 50 financial services review 5( 1) 1996 the interpretation of the c/e ratio is both interesting and useful. the c/e ratio repre sents the minimum annual rate of return on the account equity that would have to be earned to break even and cover commissions paid. if the c/e ratio is 16% then the account would have to earn 16% annually just to cover commissions and leave the account equity intact. the c/e ratio is a better measure of trading activity than the turnover ratio because it looks at trading costs directly and includes commissions on both purchases and sales. the turn over ratio includes only purchases in the account. the total cost to equity (tc/e) ratio is a variant of the c/e ratio which provides further useful information. the tc/e ratio is calculated in the same manner as the c/e ratio but includes margin interest and fees in the numerator. the tc/e ratio is interpreted as the minimum annual return necessary to break even and cover total account costs. the tc/e ratio is always greater than or equal to the c/e ratio. it will be equal to the c/e ratio if there is no margin interest or fees. the tcye ratio must be interpreted with caution. the argument can be made that it is a biased measure of trading activity in the sense that it is possible to have a margin balance and pay margin interest when there is no trading activity at all. it is a valid measure of the return necessary to break even but includes some costs that are not actual trading costs. often, the tcye ratio is presented in conjunction with the c/e ratio and is useful as an indi cation of the costs imposed on the account by trading on margin. since margin trading and the resultant risks are suitability issues, the tc/e is useful in those cases where both churn ing and suitability are issues. cost ratios are finding their way into administrative case law. in a recent sec case against brokers accused of churning several retail accounts, the administrative law judge explicitly accepted the c/e ratio “as a valid indicator of excessive trading,” (johnson, 1992). the decision is useful in validating trading costs as an indicator of excessive trad ing. unfortunately, the decision offered no guidelines to gauge excessive trading, only that the return necessary to break even and cover trading costs is a valid indicator. what is needed is a reasonable benchmark against which to gauge the excessiveness of the c/e kltio. e. turnover and commission costs the relationship between turnover and trading costs is intuitive. consider two accounts; both have a turnover ratio of four, but in one account commissions are charged at institutional rates of about 40 basis points per turnover (20 basis points one way). a turn over of four would be associated with trading costs of 1.6%, probably too high but not excessive enough to charge a stockbroker with fraudulent behavior. the other account might have trading costs of 3% per turnover (see appendix) and would have trading costs of 12% which might well be excessive and fraudulent. total commission costs in an account over a period of analysis must be equal to the total transaction volume times the average commission rate per transaction during the same period. if c is total commissions paid over a period, tv is the total volume of transactions (both purchases and sales), and acr is the average commission rate on the securities traded, then: c= tvxacr. (3) churning 51 commission costs can vary because of transactions volume and commission rates. if we know the total volume of transactions and total commissions, the average commission rate may be calculated as: acr = c/t?‘. (4) recalling that total transaction volume is the sum of total purchases and total sales, if purchases are equal to sales, then total volume may be represented by two times purchases (p). substituting 2p for tv in equation 3 and rearranging yields: 2acr = cip. (5) total commissions c and purchases p are numbers that are normally calculated in an analysis of excessive trading. it is of interest to interpret these numbers in a c/e and turn over context. dividing the c/e ratio by the turnover ratio may be interpreted as the average cost per turnover, where one turnover is both a purchase and a sale of all of the securities in the account: cost per turnover = c/e ratio i turnover ratio. (6) since a turnover is the equivalent of a purchase and a sale of account equity, the cost per turnover will be twice the average commission rate, exactly the result of equation 5. this is not surprising since the only difference between the turnover ratio and the c/e ratio is that the former has purchases in the numerator, and the latter has commissions in the numerator. the remaining terms (average equity, n and 12) cancel, leaving up. iv. churning cases this section presents summary numbers for 23 actual churning cases. of the 23, 20 were civil cases, and 3 were administrative law cases (2 sec and 1 state). in 21 of the 23 cases, the stockbroker suffered some penalty in the form of an arbitration award, administrative sanction, or monetary settlement. seventeen of the cases were against a total of 6 major brokerage firms, 6 were against regional firms. in all the cases there were allegations of unsuitable trading as well as churning. twenty of the 23 were traded on margin, and 15 traded options in some form and amount. listed equities comprised slightly more than half of the trading in these accounts. option trading and unlisted equities accounted for about 20% each, and the remainder of the trading was a mixture of bonds, mutual funds, preferred stock, warrants, and small amounts of limited partnerships. on balance, these cases constitute a wide cross-section of garden variety churning and suitability cases. trading in these accounts took place from the early 1980s through early 1994. a. overview of “churned” accounts table 2 presents some summary numbers for the 23 churning cases listed chronologi cally. column averages are presented at the bottom of the table. c as e # p er io d m on th i n a na ly si s t a b l e 2 su m m ar y a na ly si s fo r 23 c hu rn in g c as es a ve ra ge t ot al t ot al m ar gi n t ot al t ur no ve r c /e r at io t c /e r at io c os t pe r e qu ity p ur ch as es in te re st c om m is si on s r at io (% b) (% ) t ur no ve r (% ) 1 2 3 4 5 6 i 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 9/ 82 -1 2j 84 9/ 82 11 87 9/ 84 -7 18 6 9/ 85 -1 2j 91 4/ 86 -3 & o 6/ 86 -1 01 90 7/ 86 -3 /8 8 91 86 -7 18 8 l/8 71/ 90 l/ 87 lo /8 7 41 87 ~ 6l a9 41 87 -7 18 9 a/ 87 -m l 3/ 89 -4 /9 4 6/ 89 -a /9 0 7/ 89 -9 l9 4 2& l91 92 4l 91 -9 l9 2 lo l9 16l 92 6l 92 -1 19 4 1 l /9 26/ 94 l/9 34/ 94 1 l /9 31/ 95 a ve ra ge s 28 53 23 76 48 53 21 22 37 10 27 28 53 62 14 51 20 17 9 20 20 16 15 14 ,7 37 26 1, 64 7 49 ,1 46 1, 30 1, 08 8 29 ,7 5 1 3 14 ,5 03 11 2, 79 7 3, 69 34 73 23 8, 64 8 2, 44 1, 88 8 77 ,8 02 76 0, 81 8 60 ,5 31 1, 03 7, 14 5 48 ,1 92 55 8, 58 9 33 ,1 66 68 1, 44 6 69 1, 29 3 3, 79 4, 51 4 25 3, 55 1 3, 43 3, 82 2 44 ,3 36 78 2, 03 7 89 ,2 05 87 9, 55 9 46 ,7 00 2, 93 3, 80 3 11 3, 91 3 2, 12 0, 00 3 62 ,2 16 l 47 7. 88 1 9, 53 3 21 9, 83 6 84 ,7 34 1. 53 8. 42 3 49 3, 97 8 5, 97 9, 23 3 11 3, 80 0 2, 22 6, 33 7 62 ,2 93 49 1, 54 9 73 ,5 24 93 4, 55 3 13 7, 98 0 95 2, 07 4 12 7, 90 5 1, 68 7, 57 5 10 ,5 91 11 ,1 17 69 ,4 33 25 ,0 99 39 ,5 83 9, 75 9 2, 63 6 4, 82 8 3, 55 4 15 ,1 48 2, 01 9 21 ,7 87 19 ,3 02 21 ,2 31 1, 25 4 9, 20 9 1, 27 8 7, 20 6 5, 41 3 14 ,8 53 15 ,0 65 19 ,2 84 7. 6 56 .1 86 .9 7. 4 54 ,5 80 6. 0 25 .1 30 .3 4. 2 21 ,9 73 5. 5 38 .5 28 .5 7. 0 76 ,8 36 5. 2 10 .8 20 .5 2. 1 10 5, 78 1 2. 6 11 .1 13 .7 4. 3 33 18 92 2. 2 9. 9 21 .4 4. 5 53 ,4 82 9. 8 50 .5 59 .7 5. 2 30 ,2 38 6. 3 34 .2 37 .2 5. 4 25 ,4 69 6. 7 24 .9 29 .6 3. 7 13 2, 28 2 6. 6 23 .0 23 .6 3. 5 99 ,0 84 6. 0 17 .4 20 .0 2. 9 25 ,5 47 7. 6 24 .7 26 .6 3. 3 44 ,9 92 2. 2 11 .4 11 .4 5. 1 82 ,5 92 12 .2 34 .2 43 .3 2. 8 41 ,9 20 16 .0 31 .5 46 .1 2. 0 62 ,5 4 1 5. 6 23 .7 31 .7 4. 2 10 ,8 31 13 .8 68 .2 76 .1 4. 9 63 ,9 7 1 12 .8 53 .3 61 .0 4. 2 15 6, 52 7 16 .1 42 .2 42 .2 2. 6 60 ,5 34 11 .7 31 .9 35 .8 2. 7 17 ,9 87 4. 7 17 .3 24 .3 3. 7 46 ,9 04 9. 5 47 .8 53 .4 5. 0 23 ,8 44 5. 5 13 .8 22 .4 2. 5 56 ,1 34 7. 9 30 .5 37 .2 4. 0 churning 53 on average, the owners of these accounts alleged excessive trading over about a 2.5 year period (3 1 months). the range was from 9 to 76 months. in general, the time period of analysis corresponds to the period the account was under the control of the stockbroker in question. the average annual turnover ratio for these cases is 7.9 times. the range of tum over ratios is from 2.2 to 16.1. only 3 of the 23 cases had turnover ratios less than four, 4 had turnover ratios less than five, and 8 (35%) of the cases had turnover ratios less than six. thus, the majority of cases had an indication of conclusive excessive trading (turnover greater than six) regardless of the investment objectives of the client. the average annual c/e ratio was 30.5% for these cases, an indication that these accounts had to earn that annual rate to cover trading costs. the range is from 9.9% for case 6 to 68.2% for case 17. the average annual tc/e ratio was 37.2% with a range from 11.4% for case 13 to 86.9% for case 1. b. determinants of trading costs in section iii-e the relationship between turnover and trading costs was developed; total commissions are the product of trading volume times the average commission rate, and the c/e ratio is the product of the turnover ratio times the cost per turnover. the cost per turnover will be approximately twice the average commission rate. since the c/e and turnover ratios are normally calculated in churning cases, the average cost per turnover may also be calculated. the last column of table 2 presents the associated cost per turnover for each case. there is a strong relationship between the two activity measures; the cases with the highest turnover ratios also have the highest c/e ratios. the 12 cases with the lowest turnover ratios (average 4.9) have an average c/e ratio of 19.7%. the 11 cases with the highest tum over ratios (average 11.3) have an average c/e ratio of 42.3%. the standard deviation of average cost per turnover is 1.4%, an indication of consid erable commission rate variation in these accounts. variation in account commission costs can be attributed both to transaction volume (turnover) and commission rate variation (cost per turnover). a simple linear regression of c/e’s on turnover indicates that turnover explains only about half of the variance of trading costs (c/e ratios) in this sample of retail securities accounts. the remainder of the variance is explained by differences in average commission rates (cost per turnover). the range of cost per turnover is from 2.0 to 7.4%. thus a tum over ratio of four could be associated with a c/e ratio as low as 8% or as high as 29.6%. this potential variability in trading costs for the same level of turnover reinforces the notion that the turnover ratio is a flawed and incomplete indicator of excessive trading. the variability of cost per turnover or, equivalently, average commission rates can be explained by different instruments being traded and the associated commissions rates charged. the appendix presents an analysis of commission rates on listed equities and options traded in theses accounts. the weighted average one-way commission cost on listed equities was about 1.5%. this translates to a cost per turnover of about 3%. however, 15 of the 23 accounts traded options, and the weighted average one-way option commis sions in these accounts was 3.7%. this translates into a 7.4% cost per turnover for an account which traded only options. indeed, in case 1, which had a cost per turnover of 7.4%, more than 80% of the commission costs ($15,671) were commissions on options transactions. 54 financial services review 5(l) 1996 v. c/e excessive trading guideline the commission/equity ratio is superior to turnover because it measures trading costs directly and has the useful interpretation that it is the minimum annual rate of return on the securities in the account necessary to break even and cover trading costs. unfortunately, there are no case law precedents to guide courts and arbitration panels in determining what level of commission costs should constitute excessive trading. a reasonable guideline or benchmark would thus be useful to the courts, as well as to academics and practitioners who deal with individual investor issues. one approach to the solution of this problem is to relate the maximum allowable c/e ratio to average market rate of return on securities. if there is little likelihood that the account can consistently earn a positive rate of return because of the costs imposed, then trading is excessive. it is well known that the long run average annual rate of return on common stocks is about 12% (ibbotson associates, 1996). common stocks are the most prevalent investment vehicle in brokerage accounts. if an account can be generally labeled an investment account, and if trading costs are imposed in excess of about 12%, then there is no reasonable expectation that the account will consistently earn a positive rate of return. in such instances, trading costs are excessive, and a broker in control of such an account should be viewed as having traded the account excessively. this approach is a simple and elegant solution that takes into account market realities. moreover, the levels of trading which appear excessive are roughly congruent with exist ing case law and the winslow and anderson results for mutual funds. winslow and anderson argue that turnover in excess of three times should be viewed as conclusively excessive. since the average cost per turnover for the 23 accounts reviewed here was 4% a turnover ratio of three with a cost per turnover of 4% corresponds to a c/e ratio of 12%, exactly the guideline suggested. if only listed equities were traded in an investment account, then a turnover ratio of three coupled with a cost per turnover of 3% (see appendix for average equity trading costs) translates to a c/e ratio of 9%. thus, suggesting that a 12% c/e ratio is excessive is generous to stockbrokers. it is useful to put this benchmark into context. consider the average mutual fund with normal investment objectives. in the winslow and anderson sample these correspond roughly to investment objectives of growth-income, growth, and aggressive growth. the mean annual turnover rate for these funds is 0.9 times. it is well known that trading costs in an institutional context are in the range of 3 to 7 cents a share with a mean of about 5 cents. in the appendix, it is shown that commissions averaged 43 cents per share and 150 basis points. a 5 cent institutional commission trans lates to about 40 basis points per turnover. thus, the typical mutual fund manager with nor mal investment objectives and no personal gain from trading has a turnover of about 0.9 with a cost of 0.4% per turnover. this imposes annual trading costs (c/e) of about 0.36%. the 12% guideline suggested as excessive is more than 30 times the average costs imposed in an institutional context. since the majority of mutual fund managers cannot con sistently earn a risk adjusted return in excess of a’ simple buy and hold strategy after costs (jensen, 1968; malkial, 1995), it is not unreasonable to suggest that a stockbroker who imposes costs in excess of 30 times institutional costs is trading the account excessively. the 12% guideline/benchmark should not be viewed as a bright line test. rather the unique circumstances of each case should be considered. for instance, accounts with more churning 55 conservative investment objectives should reasonably tolerate lower levels of trading than investment or speculative accounts. accounts with aggressive investment objectives might tolerate slightly higher rates depending on the nature of the securities traded. the detection of excessive trading in retail securities accounts is not a settled issue. the legal profession has relied on the turnover ratio as the principal metric of excessive trading, and case law standards have evolved around that measure. this paper explores the concept of churning and excessive trading in retail securities accounts. data from 23 actual churn ing cases are used to demonstrates that the commission to equity ratio is a superior measure of excessive trading. commission costs an analysis was conducted of the commission costs of the trading conducted in the 23 accounts analyzed in the paper. agency commission costs on listed equities and options are readily available from monthly statements and confirmation slips. mark-ups on unlisted equities and bonds are typically not available, although this information can occasionally be obtained through the legal discovery process. table a-l presents the results of the analysis of commissions for listed equities and listed options. these costs are one-way commission costs, and no distinction is made between purchases and sales. for listed equities, the weighted average one-way commis sion costs on 2,3 16 round lots trades was 1 .5%, and the weighted average commission cost per share was $0.43. odd lot trades cost 2.7% one way and $1.31 per share in these accounts. trades of just 100 shares cost 1.8% one way and $0.90 per share. in all cases the weights used are the trade size divided by the total trade value in a category. the weighted average one-way commission costs on 2,220 listed option trades was 3.7% and $12.83 per contract. the commission costs are weighted by trade size. table a-l commissions on listed equities vs. commissions on options trade size number of trades weighted averages cost/ percent share commissions trade size number of trades weighted averages cost/ percent share commissions odd lot trades 2,328 $1.31 2.7% 1 contract 2,102 $31.06 7.8% round lot trades 2,6 12 $0.90 1.8% 5 contracts 2,568 $14.81 4.7% 100 shares 1000 shares 2,403 $0.37 1.5% 10 contracts 2,590 $13.85 3.6% > 1000 shares 2,404 $0.21 1.5% 20 to 50 contracts 2,307 $10.82 3.3% overall round lot 2,316 $0.43 1.5% overall 2,220 $12.83 3.7% 56 financial services review 5( 1) 1996 anderson, s., heacock, m., & hill, k.p. (1987). churning: an ethical issue in finance. business and professional ethics journal, 6( 1). 317. goldberg, s.c. (1991). piaba’s 1991 public investor’s recover guide and arbitrator source book to stockbroker fraud and securities arbitration. austin, tx: public investors arbitration bar association. jensen, mc. (1968). the performance of mutual funds in the period 19451964. journal of finance, 23,389-416. jensen, m.c. (1991). abuse of discretion claims under rule lob-5 churning, unsuitability, and unau thorized transactions. securities regulation law journal, 18,374-399. johnson, se. (1992, june). admin. proc. file no. 3-7528, 1992 lexis 1598. ibbotson associates. (1996). stocks, bonds, bill, and inflation 1996yearbook. chicago, il: ibbotson associates. loss, l., 8~ seligman, j. (1989). securities regulation-volume viii, 3rd ed. boston, ma: little, brown & co. malkial, b.g. (1995). returns from investing in equity mutual funds 1971 to 1991. the journal of finance, 50(2), 549-572. mcmahon v. shearson lehman. (1987). 482 u.s. 220,226. national association of securities dealers. (1944). 17 sec 459, gl-g6. note. (1967). churning by securities dealers. harvard law review, 80, 869-876. schreiner, j. (1980). portfolio revision: a turnover constrained approach. financial management, 9(l), 67-75. u.s. securities and exchange commissions. (1968). 96th congress, 1st session, report of the spe cial study of the options market. winslow, d.a., & anderson, s.c. (1990). a model for determining the excessive trading element in churning claims. north carolina law review, 68(2), 327-361. financial services review, 33(2) 74 student willingness to borrow for higher education stuart heckman,1 jodi letkiewicz,2 and hanna lim3 abstract a human capital model is used to examine students’ willingness to borrow to pay for a college degree. we hypothesize direct costs of education and education goals to be positively associated and current income and alternative financial support to be negatively associated with willingness to borrow. using college student data from the 2020 study on collegiate financial wellness, we found that 13% of college students are not willing to borrow to pay for school, 31% are willing to borrow up to $20,000, and 27% are willing to borrow up to $50,000. overall, we find evidence that students’ willingness to borrow corresponds to the rational decisions predicted from human capital theory. higher tuition costs, educational goals, and fields of study with higher expected pay were all positively associated with willingness to borrow. income is positively correlated with willingness to borrow at the lower end of income, but as income increases, the amount students are willing to borrow is less. alternative financial support from either scholarships, grants, or family is negatively associated with willingness to borrow, which is all consistent with the twoperiod human capital model. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation heckman, s., letkiewicz, j., & lim, h. (2025). student willingness to borrow for higher education. financial services review, 33(2), 74-92. introduction among policymakers, researchers, and media outlets, concerns abound regarding the diminishing value of a college degree and the widespread borrowing behavior of college students. these are important concerns since current public policy in the u.s. is designed to ensure equal access to post-secondary educational opportunities. while equal access is the intended effect of financial aid, students have diverse economic backgrounds and circumstances which may result in high student debt that can significantly impact long-term 1 corresponding author (stuart.j.heckman@ttu.edu). texas tech university, lubbock, texas, usa. 2 california state university-northridge, los angeles, california, usa. 3 california state university-fullerton, fullerton, california, usa. financial planning goals for individuals and families. federal financial aid policy subsidizes higher education approximately $125 billion per year through grants, loans, and work-study (federal student aid, 2021). this figure does not include other public subsidies through tax credits (i.e., the american opportunity tax credit and the lifetime learning credit), a deduction for student loan interest, and tax advantaged investment accounts (e.g., 529 plans). the justification for this substantial public investment is typically grounded in human capital theory because society benefits from having a more https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ heckman et al. 75 productive workforce. however, the implications of student borrowing extend beyond individual students to the broader economy, affecting everything from risk assessment in lending practices to the development of financial products tailored to manage educational debt. understanding students' willingness to borrow is crucial for financial planners and advisors who must help clients navigate the complex interplay between education financing and long-term financial goals, ultimately influencing wealth accumulation and economic stability for both individuals, families, and the economy as a whole. while the rationale for the public subsidy of higher education is clear from the perspective of policymakers, the rationale for investing (both money and time) in their own college education is less understood from the student perspective. although there is a rich literature regarding the pecuniary returns and nonpecuniary benefits (ma & pender, 2023; oreopoulos & salvanes, 2011) of attending post-secondary education, there remains a deficient understanding of student valuations of higher education. borrowing for college is different from other consumer loans in that while individuals immediately consume the goods purchased with consumer loans such as credit card debt, borrowing for college is usually used for future production over the long-term (li, 2013). furthermore, some have argued that the consumption value of higher education could be quite substantial, i.e., students may be willing to pay a high price to simply have the college experience without regard to the financial payoff (alstadsæter, 2010; jacob et al., 2013), and it makes the decision to borrow for college education more complicated. the literature gives little guidance regarding the relative importance of these considerations as students make borrowing decisions. from an economic standpoint, some have suggested that declining to use loans to invest in human capital is an irrational decision (cadena & keys, 2013). on the other hand, it may be rational for students to decide not to borrow if they do the calculations and determine their choice of major and labor market prospects will make repaying their loans difficult (boatman et al., 2017). regardless of the question of whether borrowing is rational, there is evidence that student wellness may be suffering due to higher education financing decisions. heckman et al. (2014) found that 71% of college students report feeling financially stressed and that expecting to have debt at graduation was associated with higher likelihoods of financial stress. perhaps this stress is valid given that many students struggle with repayment (avery & turner, 2012) and more than a million students defaulted for the first time in 2016 (frotman & williams, 2017). the number of defaults continued to climb until 2020 (u.s. department of education, 2020) when the coronavirus aid, relief, and economic security act (cares act) halted collection on defaulted loans, credited payments towards income-driven repayment (idr), and temporarily blocked the accrual of interest for a set period of time (cares act, 2020). examining students' willingness to borrow for higher education provides crucial insights that can inform financial advisors, institutions, and policymakers about the borrowing behaviors of a key demographic. this research not only reveals the extent to which borrowing decisions align with human capital theory expectations, but also contributes to curriculum development in financial services education, ensuring future professionals are equipped to address the challenges faced by students in financing their education and repaying their debt. the findings offer important perspectives on students' perceptions of the value of a college degree, including the accuracy of information on costs and returns to education. such insights can inform policies designed to improve access to higher education among prospective students from diverse socioeconomic backgrounds and support the retention of current students in an uncertain economy. further, findings from this study can be useful in preventing student loan defaults, which is an important concern among policymakers and higher education administrators. in a study of ohio colleges, approximately 25% of students anticipated defaulting on their loans and students who believed that higher education was a good investment were less likely to anticipate default (fox et al., 2017). subsequently, helping students make good borrowing choices could improve the financial services review, 33(2) 76 likelihood of repayment. this study seeks to advance the literature by examining factors related to students’ willingness to borrow in order to pay for a college degree, utilizing responses from a national dataset that directly queries current college students about their personal borrowing limits. literature review value of college degree current consensus from the economic literature is that the monetary returns to college education outweigh the costs (oreopoulus & petronijevic, 2013; ma & pender, 2023). the literature also suggests that the wage premium has been growing (avery & turner, 2012; carnevale, cheah, & wenzinger, 2021). recent studies report that those with a bachelor’s degree earn 6575% more than those with no education past a high school diploma and approximately $1.2 million more over their lifetimes (carnevale, cheah, & wenzinger, 2021; ma & pender, 2023). in addition to the private returns to education, researchers argue that there are many public benefits to higher education (e.g., see damon & glewwe, 2011). for the purposes of our discussion, it is sufficient to keep the focus on the private returns to education, which include many benefits beyond the obvious monetary returns. oreopoulus and salvanes (2011) outline a host of other non-pecuniary private benefits, including improved work environments, sense of accomplishment, autonomy, job security, opportunity for social interaction, and prestige. they also point out that increased education can lead to better marriage opportunities, improved health choices, and an enjoyable college experience, concluding that the combined effect of pecuniary and non-pecuniary returns would be quite large (oreopoulos & salvanes, 2011). those with college degrees also face much lower unemployment rates, although there is substantial variation considering factors such as an institution’s reputation, a student’s major, and the skills they learn (sigelman & selingo, 2021). during difficult economic periods, such as a recession (hoynes et al., 2012) and the covid19 pandemic (daly et al., 2020), those with college degrees have been found to suffer less from unemployment than those without college degrees. in terms of other non-pecuniary benefits, ma and pender (2023) report that those with at least a bachelor’s degree are more likely to be engaged in the community, involved with their children’s activities, and more likely to live a healthy lifestyle. borrowing decisions research on student borrowing decisions has mostly focused on financial resources of students and found some disparity in borrowing decisions depending on their available resources (baum & payea, 2012; cha & weagley, 2002; cunningham & santiago, 2008; goldrick-rab & kelchen, 2015; perna, 2008). cha and weagley (2002) find that students with higher income, which includes parental income for dependent students, are significantly less likely to borrow to pay for college than students with low-income. similarly, students from low-income families are more likely to borrow (baum & payea, 2012; baum & schwartz, 2015). this discrepancy results in students from low-income households having more debt (households earning less than $30,000 average debt load =$16,500) compared to those from higher income households (households earning more than $120,000 average debt load = $14,000) (baum & payea, 2012). this suggests an unequal debt-to-income burden for low-income families. although students with more assets were less likely to borrow, they borrowed higher amounts than students with fewer assets when they did borrow (cha & weagley, 2002). in addition to having a direct impact on the available financial resources, research shows that family socioeconomic background has other effects on borrowing choices. for example, several studies (cunningham & santiago, 2008; goldrick-rab & kelchen, 2015; perna, 2008) find evidence of debt aversion among students from lower socioeconomic backgrounds. goldrick-rab and kelchen (2015) explore the topic of debt aversion using a sample of pell grant recipients from the wisconsin scholars longitudinal study. the researchers report that parental education was positively correlated with the decision to take student loans. students from families with lower socioeconomic status (lowses) are more loan averse, particularly those heckman et al. 77 growing up in poverty or lacking financial support from their families. based on their findings, the authors suggest that an aversion to debt might be related to cultural and community norms and these factors are important to consider when developing college financing interventions (goldrick-rab & kelchen, 2015). furthermore, a survey of undergraduate students reports that 45% of the neediest students decided not to take the student loans offered to them, which rendered them unable to cover the expenses needed to complete their education (cunningham & santiago, 2008). in an examination of high school students’ willingness to borrow for college, perna (2008) finds that loan perceptions are heavily influenced by the messages students receive from parents and teachers regarding loans. she also finds that lowincome students tend to be less informed and to view the use of student loans as riskier compared to students from high-income families. her analysis shows that students and parents generally thought about student loan borrowing in terms of costs and benefits, as predicted by human capital theory (perna, 2008). results from boatman and evans (2017) are consistent with perna’s (2008) finding that student loan borrowing is generally consistent with human capital theory. boatman and evans (2017) report that awareness of income-based repayment and higher financial literacy were both associated with a greater willingness to borrow. the authors suggest that knowledge about how federal loans work and repayment options after college alters the cost-benefit analysis and makes borrowing more acceptable (boatman & evans, 2017). several studies show clear relationships between cost and willingness to borrow. goldrick-rab and kelchen (2015) find that higher net tuition prices were correlated with greater willingness to borrow. in the analysis of ohio state university’s internal data, hart and mustafa (2008) find that the net cost of attendance, defined as total costs less aid that does not need to be repaid, is significantly positively related to student loan amounts. furthermore, they find that higher loan limits, while not affecting low-income borrowers, positively affect the amount that middleand upper-income students borrowed. research has also consistently shown that those who expect higher wages are more willing to borrow. cha and weagley (2002) find that higher expected post-graduation wages are related to higher amounts of student loan debt. goldrickrab and kelchen (2015) find that higher achieving students and those with higher expected earnings in the future are more likely to borrow. this is potentially an important concern for policymakers because previous research reports that earnings expectations are heavily influenced by ses background of individuals. low-ses students had substantially lower earnings expectations, which may indicate that these students anticipated labor market discrimination or were otherwise systematically underestimating the returns to education (delaney et al., 2011). while the research literature shows evidence that students generally make borrowing decisions based on the costs and benefits of higher education, there is concern that students are borrowing too much. avery and turner (2012) examine this concern and report that students were not borrowing excessively and suggest that people may be displaying a form of cognitive bias when they put more weight on extreme cases than paying attention to the average student. people tend to pay attention to vivid or extreme cases and extrapolate those to the general population. there is, of course, evidence that suggests some students experience financial distress due to their student debt (despard et al., 2016; heckman et al., 2014; martin & dwyer, 2021; mckinnery & burridge, 2015). individuals with student loans experience higher likelihoods of material hardships such as trouble meeting basic needs such as food, medical care, and shelter (despard et al., 2016). heckman et al. (2014) finds that any debt, including but not exclusive to student loans, is positively associated with financial stress among college students. mckinnery and burridge (2015) report that community college students with loans are more than twice as likely to drop out as non-borrowers. the effects of debt and stress have been found to be more significant among black and hispanic students (martin & financial services review, 33(2) 78 dwyer, 2021). although some literature suggests that students are not systematically overborrowing and that borrowing generally seems to correspond to costs and benefits of higher education, there is evidence to show that borrowing decisions are associated with financial distress. gap as discussed above, previous researchers were interested in factors related to borrowing decisions, focusing on explaining which individuals borrow or not or which individuals borrow too little or too much. while the current study is still interested in factors related to borrowing, we further the literature by utilizing a national dataset with unique data to examine responses to a hypothetical question about the amount students are willing to personally borrow to pay for their degree. this hypothetical question could give us insight into students’ thought processes and valuation of a college degree. to our knowledge, there are currently no studies that ask students directly for the dollar amount of student loan debt that they would be personally willing to accumulate to pay for their college degree. the current literature is primarily based on actual student loan borrowing behavior or hypothetical questions about general loans or debt. therefore, this study aims to contribute to the growing body of research by studying student willingness to borrow. theoretical framework the theoretical foundation of this study comes from the economics of education and human capital theory. human capital theory in the context of education was first introduced by becker (1964) and examines the relationship between investment in education and training (i.e., human capital) and lifetime earnings. following human capital theory, individuals are willing to invest in education in line with the expected return on that investment. investments in education include both direct expenses (i.e., cost of tuition) and the opportunity cost of foregone earnings while obtaining the education. the investment in education will then be made by individuals who expect the benefits will outweigh the costs. the foundation for our analysis is a basic twoperiod human capital model presented by daniele checchi (2006, pp. 20-23) as a simplification of the well-known ben-porath (1967) model. in his model, the optimal fraction of time that should be devoted to education in the first period (t) for individual i, denoted 𝑆𝑖𝑡 ∗ , is a function of ability, initial endowment of human capital, earnings premium, discount rate, effort, and the direct costs of schooling. because it is possible for the direct costs of the optimal amount of schooling to be greater than the income available in the first period, individuals may borrow in the first period to maximize lifetime utility or draw upon other financial resources (e.g., personal or family wealth) to finance the human capital investment. given that we analyze a sample of college students, presumably there should be no individuals for which 𝑆𝑖𝑡 ∗ ≤ 0. assuming that 𝑆𝑖𝑡 ∗ > 0, the optimal amount that individual i should be willing to invest in the first period, 𝑀𝑖𝑡 ∗ , is equal to the direct cost of education (𝛾) multiplied by the optimal amount of time for education (𝑆𝑖𝑡 ∗ ). the direct cost of education can be proxied by the tuition cost and the optimal amount of time for education can be proxied by the individual’s educational goal. the optimal amount to invest for education may be reduced by either current income or alternative financial resources. assume that current income is the product of wages (𝛽𝑡) and the amount of time spent in the labor market (1 − 𝑆𝑖𝑡 ∗ ). alternative financial resources, denoted (𝐹𝑖𝑡), include but are not limited to, personal or family wealth and nonloan aid. therefore, 𝑀𝑖𝑡 ∗ is reduced accordingly: 𝑀𝑖𝑡 ∗ = (𝛾𝑆𝑖𝑡 ∗ ) − 𝛽𝑡(1 − 𝑆𝑖𝑡 ∗ ) − 𝐹𝑖𝑡 [1] the axiom of a “rational agent” in economic theory proposes that an individual will account for all available information, costs, and benefits when determining a course of action that is in their best interest (simon, 1955). while there is some controversy over the term “rational” (desjardins & toutkoushian, 2005), we use “rational” in this study to simply describe adherence to the economic model outlined above and “irrational” to describe a deviation from the expected model. heckman et al. 79 hypotheses equation [1] implies the following relationships: 𝑀𝑖𝑡 ∗ = ( + 𝛾 , + 𝑆𝑖𝑡 ∗ , − 𝛽𝑡 , − 𝐹𝑖𝑡 ). [2] therefore, the following hypotheses are postulated: h1: direct costs of education will be positively associated with willingness to borrow. h2: education goals will be positively associated with willingness to borrow. h3: current income will be negatively associated with willingness to borrow. h4: alternative financial support will be negatively associated with willingness to borrow. methodology data and sample this study uses data from the 2020 study on collegiate financial wellness (scfw), previously titled the national student financial wellness survey (nsfws). the scfw is a survey of college students examining the financial attitudes, practices, and knowledge of students from higher education institutions across the us. the online survey was administered by the center for the study of student life and college of education and human ecology at the ohio state university and was launched in february 2020. the survey is comprised of a random sample of students from 60 two-year and four-year institutions. in total, 29,883 students responded to the survey and the final sample available for analyses in the current paper includes 24,121 respondents. the institutional data are matched with the institution postsecondary data system (ipeds) of the national center for education statistics providing information about each institution, such as tuition, faculty-to-student ratios, and rate of students receiving pell grants. two dependent variables are created from the categorical responses on the following question: “assuming you are paying, or had to pay, for college on your own, how much debt would you be willing to personally accumulate in order to complete your current degree?” first, a categorical variable was created to distinguish between students who are willing to borrow for a college degree, those who are not willing to borrow, and those who do not know if they are willing to borrow. while we recognized that students responding “don’t know” to the question may be classified as willing to borrow, we found factors that distinguish this group from the others and decided to keep them as a separate group for the first stage of the analysis. for example, students who responded “don’t know,” on average, report more support from their parents, lower gpas, and have higher tuition costs. second, among those who reported that they are willing to borrow for a college degree, a continuous variable was created to represent how much debt the students are willing to accumulate. we transformed categorical answers to continuous values by taking the midpoint from each range of categories. there were eight possible ranges, ranging from zero (not willing) to “over $60,000.” students responding “don’t know” were excluded from the second analysis reducing the sample size for that model. empirical model although we are interested in the underlying value or amount that the student is willing to invest, 𝑀𝑖 ∗ is not observed. for convenience, we drop the t subscript when discussing the empirical model. 𝑀𝑖, the amount of student loan debt that individual i would be willing to personally accumulate in order to pay for college, is observed in the data. we use two empirical models to test the hypotheses regarding the amount students are willing to borrow. the first is a multinomial regression and the second is a tobit regression. both models utilize maximum likelihood for parameter estimation. a multinomial logistic regression is a method that generalizes logistic regression to multiclass problems where there are more than two discrete outcomes. multinomial logistic regression uses a linear predictor function 𝑓(𝑘, 𝑖) to predict the probability that the observation 𝑖 has outcome 𝑘 as modeled below: 𝑓(𝑘, 𝑖) = 𝛽𝑘 . 𝑥𝑖 [3] financial services review, 33(2) 80 where 𝛽𝑘 is the set of regression coefficients associated with outcome k, and 𝑥𝑖 is the set of explanatory variables associated with observation i. we use this model to compare three distinct groups – those willing to borrow, those not willing to borrow, and those who do not know if they are willing to borrow. to obtain actual probabilities rather than relative probabilities, we report average marginal effects from the multinomial logistic regression model. the tobit model is a hybrid of a probit and an ordinary least squares (ols) regression and allows us to model both the choice to borrow and the extent to which one is willing to borrow. the tobit model is particularly suitable for this analysis because it addresses the unique characteristics of the sample, which includes both students who are unwilling to borrow and those who are willing to borrow. in this context, the dependent variable—how much students are willing to borrow—is censored at zero; that is, for those who choose not to borrow, the amount is not just unobserved but is actually zero. by employing a tobit model, we can simultaneously capture two critical aspects of borrowing behavior – the decision to borrow and the extent of borrowing. marginal effects are calculated to reflect the change in the average amount someone is willing to borrow, including both those who are willing to borrow (uncensored) and those who are not (censored at zero). we use the following tobit model to analyze the amount a student is willing to borrow for higher education. 𝑀𝑖 = 𝛽0 + 𝑋𝛽 + 𝑢 [4] where 𝑢|𝑋 ~𝑁 (0, 𝜎2) 𝑀 = { 𝑀𝑖 𝑖𝑓𝑀𝑖 > 0 0 𝑖𝑓 𝑀𝑖 ≤ 0 independent variables the independent variables include control variables, cost variables, educational goal variables, current income, and financial support variables. several variables control for student characteristics such as students’ class rank, gender, race/ethnicity, and whether or not they are first generation or nontraditional students. research has documented earnings gaps between men and women and between white and racial/ethnic minorities. therefore, women and non-white students may be less willing to borrow than men and white students. dummy variables of “male,” “female,” and “other or prefer not to say” were created for students’ gender, with “male” coded as the reference category. dummy variables of “white,” “black,” “hispanic,” “asian,” and “other” were created for students’ racial/ethnic identities, with “whites” coded as the reference category. dummy variables of “year 1,” “year 2,” “year 3,” “year 4,” and “more than 4” were created for students’ class rank and “year 1” was used as a reference category. a student is considered first generation if the student reported that the highest level of educational attainment of their parent/guardian was less than a bachelor’s degree. a nontraditional student is coded as such if they report supporting a child or family member(s) or if they are at least 24 years old. cost. the cost-related variables are the net price for an undergraduate degree and an indicator for tuition type. the net price variable comes from the matched ipeds data. net price means the amount that a student pays to attend an institution (tuition and room and board) after adjusting for scholarships and grants. while we do not directly observe the actual amount a student pays for tuition, we believe using the average net price of the institution where they are enrolled is a good proxy for cost of attendance since majority of full-time undergraduate students receive aid (college board, 2023). the tuition type variable was created using the two variables from scfw – the administrative variable on whether the institution is public or private and the respondents’ response to whether they qualify for in-state tuition or out-of-state tuition (including international student tuition). one variable for tuition type was created with three categories of 1) public in-state, 2) public out-of-state, and 3) private. educational goals. variables expected to influence the optimal amount of borrowing for education include gpa, major, planned educational attainment, and student perception of tuition as an investment. the variable used for gpa is the student’s self-reported gpa. after dropping the outliers in gpa values (gpa above heckman et al. 81 5.0), gpa variable ranges from 0 to 4.93. binary variables were created to indicate students’ majors based on responses to the broad category of majors. respondents were able to select multiple categories from a list that includes 1) arts or humanities, 2) business, 3) education, 4) health or medicine, 5) social sciences, 6) stem (science, engineering, technology, or math), 7) vocational, and 8) other. next, a binary variable was created to distinguish those who strongly agree or agree that the cost of tuition is a good investment for their financial future from those who disagree or strongly disagree. finally, students were asked “what type of degree are you currently pursuing?” and “what is the highest degree you plan to obtain?” the options for type of degree they are currently seeking were “2-year degree,” “4-year degree,” “certificate” and “other.” a categorical variable was created to indicate the four categories. if students responded to the question about the highest degree they planned to achieve with “master’s”, “professional’, or ‘doctoral,” the graduate school variable is coded a 1, zero if otherwise. current income. students were asked how many hours they work and the hourly rate they earn. from those questions, we create two variables. one is created for employment (yes/no) and the other is a calculated annual income based on hours worked and hourly rate. alternative financial support. receipt of scholarships and grants and family financial support were used to measure student access to alternative financial resources. the survey asks students “please indicate how much of your college/university expenses are paid for by the following: (1) parents or other family members from their current income or past savings, or (2) scholarships or grants that don't need to be repaid.” continuous variables were created based on the responses from none (1) to all (4). results sample description this paper explores which students are willing to borrow for their education and how much they are willing to borrow. just over 70% of the sample indicated a willingness to borrow, 16% said they did not know how much they were willing to borrow, and 13% said they were not willing to borrow. these figures are substantially lower than the percentage of debt averse students, which was 48%, found by goldrick-rab and kelchen (2015), though the sample in their study was exclusively low-income college students. we believe the different proportions between the two studies have to do with the difference in samples (low-income vs. general population) and how the questions were asked. the discrepancy also likely highlights stronger debt-aversion among lowincome students, which our broad sample may not adequately capture. student rank is fairly well balanced with 26% of the sample in their first year, 23% in their second year, 25% in their third year, 20% in their fourth year and the remaining 6% taking more than four years. women represent 67% of the sample, men make up 30% and those who prefer not to answer or indicate “other” comprise 3% of the sample. approximately two-thirds of the sample identify as white, 6% are black, 11% are hispanic, 8% are asian, and 10% responded as a race or ethnicity other than those listed. on average, tuition (net price) is approximately $19,000 per year and the mean amount students are willing to borrow is just over $27,000 for students who indicated they were willing to borrow. the sample have high educational goals overall with 64% percent indicating they plan to pursue some form of graduate education and the mean gpa is 3.38. approximately 71% of the sample believe that college is a good investment. this figure drops to 65% for those who are unwilling to borrow. thirty percent of the sample indicated stem as their major(s), followed by arts and humanities (21%), health (19%), business (18%), and social science (17%). just over 60% of the sample are employed and the mean earnings for those who work is just over $6,500, which is expected to reduce the amount of money borrowed to pay for college education. furthermore, students have alternative financial support; the sample mean for scholarship or grants covered is 2.3 (somewhere between “some” and “most”), and for parental support it is 2.01 (“some”). financial services review, 33(2) 82 table 1. sample description total sample not willing willing don't know variable n=24,121 n=3,095 n=17,050 n=3,976 sample mean 0.13 0.71 0.16 max debt ($) $22,942 $27,107 controls rank year 1 0.26 0.25 0.25 0.30 year 2 0.23 0.24 0.23 0.25 year 3 0.25 0.25 0.25 0.22 year 4 0.20 0.19 0.21 0.18 more than 4 years 0.06 0.07 0.06 0.05 first generation 0.41 0.44 0.41 0.37 non-traditional 0.15 0.20 0.14 0.12 gender male 0.30 0.33 0.31 0.22 female 0.67 0.63 0.67 0.75 prefer not to say 0.03 0.03 0.02 0.03 race/ethnicity white 0.65 0.56 0.67 0.64 black 0.06 0.07 0.05 0.07 hispanic 0.11 0.13 0.10 0.09 asian 0.08 0.12 0.08 0.09 other 0.10 0.13 0.09 0.11 cost net price ($1,000) 1.91 1.84 1.90 2.00 public in-state tuition 0.68 0.71 0.69 0.62 public out-of-state tuition 0.18 0.16 0.18 0.21 private 0.13 0.13 0.13 0.18 educational goals gpa 3.38 3.38 3.38 3.34 good investment 0.71 0.65 0.73 0.70 major: arts & humanities 0.21 0.24 0.19 0.24 major: business 0.18 0.19 0.19 0.13 major: education 0.07 0.07 0.07 0.08 major: health 0.19 0.15 0.19 0.21 major: social science 0.17 0.16 0.17 0.19 major: stem 0.30 0.29 0.31 0.26 major: vocation 0.01 0.01 0.01 0.01 major: other 0.06 0.07 0.05 0.08 two year goal 0.03 0.04 0.02 0.03 four year goal 0.96 0.94 0.96 0.95 certificate 0.01 0.01 0.01 0.01 other 0.01 0.01 0.01 0.02 grad school plans 0.64 0.58 0.64 0.64 heckman et al. 83 income employed 0.61 0.58 0.63 0.56 student income ($1k) 6.55 7.05 6.70 5.55 alternative financial support scholarships/grants (14) 2.31 2.40 2.32 2.20 parental support (1-4) 2.01 2.07 1.97 2.15 multivariate results the combined results of the multinomial logit and the tobit provide insight into both the decision to borrow and the amount students are willing to borrow. results from the multinomial logistic regression are presented in table 2 and the tobit results are presented in table 3. marginal effects were computed for the multinomial regression and are presented in table 2. the results discussed in this section will focus on these marginal effects and the standard coefficients from the tobit results. control variables. as students progress through school, the likelihood they are willing to borrow or to indicate that they knew, one way or another, increases. the willingness to borrow of students in later years may reflect a recognition of the benefits of education. compared to the first-year students, the later year students may value the college experience based on their time spent in college. those who were enrolled for more than four years are 5.4 percentage points more likely to say they are willing to borrow than students in their first year in college. first generation students are not statistically different than their counterparts in terms of willingness to borrow or how much they are willing to borrow. this differs from findings in a study by furquim et al. (2017) that reported first generation students were both more likely to borrow and willing to borrow more and from the findings by goldrick-rab and kelchen (2015) that more parental education leads to less debt aversion. non-traditional students display more reluctance to borrow. they were 3.1 percentage points less likely to say they are willing to borrow and 4.4 percentage points more likely to say they are unwilling to borrow. in addition, the amount they are willing to borrow is negative and significant (me = -1.424, p < .001). this may stem from the fact that they are older, more mature, and may have more financial resources at their disposal. women are less likely to be willing to borrow, but also less likely to be unwilling to borrow, and more likely to be unsure about their willingness to borrow (i.e., “don’t know”). students who identify as a race other than white were more likely to be unwilling to borrow than white students and less likely to say they are willing to borrow. this trend continues when looking at how much students are willing to borrow, broken down by race/ethnicity. students identifying as black, hispanic, asian, or other are all willing to borrow less than their white counterparts. this is in line with boatman et al. (2017) who found that hispanic students were more likely to be debt averse but differs from findings by goldrick-rab and kelchen (2015) who found that black students were more willing to borrow. this is an important finding that needs further research. cost. the findings provide evidence to support h1 that the direct cost of education is positively associated with the amount willing to borrow. net price was negatively correlated with unwillingness to borrow and positively correlated with a willingness to borrow. students attending universities with higher net price were more likely to say they did not know how much they would borrow. this may reflect an acknowledgement that the high cost of their college creates some uncertainty about the amount of loans that will be required to finish the degree. students from higher cost schools reported being willing to accumulate significantly more student loan debt (me = 5.485, p < .001). students paying out of state tuition at public schools are less likely to be unwilling to borrow, more likely to be willing to borrow, and more likely to be uncertain about borrowing that those financial services review, 33(2) 84 paying in-state tuition. the amount they are willing to borrow is also much higher (me = 5.631, p < .001). these findings are consistent with other studies (goldrick-rab & kelchen, 2015; hart & mustafa, 2008) which found that students attending universities with higher net costs were less likely to be debt averse and were willing to borrow more. it’s worthwhile to note that students from private schools have opposite results than those paying out-of-state tuition; they are more likely to be unwilling to borrow and less likely to be willing to borrow than students from in-state public schools. they are also willing to borrow less (me = -2.365, p < .001). we think the private school findings are picking up on family affluency rather than cost of attendance. educational attainment. we find partial support for the hypothesis that those with higher educational aspirations are more likely to borrow (h2), in line with the expectation of human capital theory. gpa is positively correlated with an unwillingness to borrow and those with higher gpa are less likely to be uncertain about borrowing. the second analysis indicates that gpa is negatively associated with the amount students are willing to borrow (me = -2.142, p < .001). this differs slightly from goldrick-rab and kelchen’s (2015) findings that students who expressed an unwillingness to borrow had lower gpas than students who were willing to borrow. the way the question in the scfw is asked may explain the more nuanced finding and there are other possible explanations for this finding. one possibility is that students with higher gpas may plan to finish on time or even early, whereas students with lower gpas may take longer to finish their degree. students with higher gpas may receive more financial support (grants/scholarships) or may simply be more pragmatic about their borrowing behaviors. to further this point, students with high gpas are less likely to respond “don’t know” when asked whether they are willing to borrow. students who believe tuition is a good investment were 4.3 percentage points more willing to borrow and were willing to borrow more (me = 3.138, p < .001). this supports the notion that they are making rational borrowing decisions premised by human capital theory. students with an intention to pursue advanced education were less likely to say they were unwilling to borrow and more likely to say they were willing to borrow. this is likely due to the costs associated with many advanced degrees. students who planned to attend graduate school were willing to borrow more (me = 2.821, p < .001). student choice of major and type of institution also show rational decision making by students. students in more traditionally low-paying majors such as arts and humanities are less likely to be willing to borrow and are willing to borrow less (me = -1.755, p < .001). student majoring in business, health or medicine, and stem fields are more willing to borrow for their education. there is some deviation, however, in how much they are willing to borrow with only students in health or medicine majors willing to borrow more (me = 3.086, p < .001). students pursuing a 4year degree were willing to borrow far more than students pursuing a 2-year degree (me = 2.006, p < .019). current income. results from the study provide mixed support for h3 (current income will be negatively associated with willingness to borrow). the results indicate that employment is positively correlated with willingness to borrow and negatively correlated with an unwillingness to borrow, and students who are employed are willing to borrow more (me = 1.430, p = <.001). however, student income is positively associated with an unwillingness to borrow and students with more income are willing to borrow less (me = -0.136, p < .001). these are somewhat contradictory findings – on the one hand, students who are employed are more willing to borrow and on the other hand, student income is negatively correlated with how much someone is willing to borrow. what we might be picking up on here is that students who are working may be doing so to pay some of their living expenses rather than to pay for their education, a reflection of the high cost of room, board, and tuition and the financial squeeze on families. students with higher incomes may be working full time and able to pay for their tuition and/or be debt averse, so are working more to pay for their education. alternative financial resources. the findings on access to alternative financial resources provide support for h4 (alternative financial heckman et al. 85 support will be negatively associated with willingness to borrow). those receiving more financial support from either scholarship/grant were more likely to say they were unwilling to borrow and were willing to borrow less (me = 3.041, p < .001). likewise, students receiving more financial support from their parents are more likely to say they are unwilling to borrow and are willing to borrow less (me = -2.192, p < .001). this runs counter to the findings by cha and weagley (2002) who found that while students with more wealth were less likely to borrow, they tended to borrow more when they did borrow. interestingly, those with parental support were also more likely to say they did not know how much they were willing to borrow, perhaps showing some uncertainty around continued support, higher educational aspirations, or ignorance of their financial obligations. similarly, previous study found that college students whose family was the primary source of funding for their expenses were more likely to be unaware about how much student debt they have (letkiewicz et al., 2019). table 2. 2020 multinomial results on willingness to borrow: average marginal effects not willing willing don't know variable dy/dx s.e. pvalue dy/dx s.e. pvalue dy/dx s.e. pvalue control rank (ref: year 1) year 2 .006 .006 .338 .006 .009 .463 -.012 .007 .080 year 3 .001 .006 .848 .034 .008 <.001 -.035 .007 <.001 year 4 -.007 .007 .287 .045 .009 <.001 -.038 .007 <.001 more than 4 -.011 .010 .239 .054 .013 <.001 -.043 .011 <.001 first generation .008 .005 .101 .005 .007 .475 -.012 .005 .020 non-traditional .044 .006 <.001 -.031 .009 .001 -.013 .008 .115 gender (ref: male) female -.020 .005 <.001 -.036 .007 <.001 .056 .005 <.001 prefer not to say 021 .015 .155 -.100 .020 <.001 .079 .017 <.001 race/ethnicity (ref: white) black .038 .010 <.001 -.067 .013 <.001 .029 .011 .008 hispanic .039 .008 <.001 -.027 .010 .009 -.012 .008 .126 asian .073 .009 <.001 -.066 .011 <.001 -.007 .008 .430 other .048 .008 <.001 -.058 .010 <.001 .010 .008 .213 cost tuition cost (net price) -.036 .005 <.001 .021 .007 .003 .014 .006 .011 tuition type (ref: public in-state) public out-of-state .025 .005 <.001 .006 .008 .451 .019 .006 .003 private .023 .012 .062 -.047 .015 .002 .025 .012 .041 educational goals gpa .016 .004 <.001 .011 .006 .058 -.027 .005 <.001 good investment -.041 .005 <.001 .043 .006 <.001 -.002 .005 .727 financial services review, 33(2) 86 major: arts & humanities .020 .006 .001 -.032 .009 <.001 .012 .007 .094 major: business .003 .007 .662 .056 .010 <.001 -.059 .008 <.001 major: education -.003 .009 .741 .001 .012 .967 .002 .010 .802 major: health -.027 .007 <.001 .024 .009 .010 .003 .007 .728 major: social science -.012 .007 .080 .010 .009 .281 .002 .007 .781 major: stem -.001 .006 .844 .025 .009 .004 -.023 .007 .001 major: vocational .023 .022 .294 .040 .033 .230 -.063 .030 .036 major: other .027 .009 .003 -.044 .013 .001 .016 .010 .115 degree goals (ref: 2 yr) 4-year -.012 .013 .373 .019 .019 .323 -.007 .016 .653 certificate -.011 .030 .698 .047 .040 .240 -.036 .032 .263 other -.039 .022 .081 -.026 .034 .447 .065 .030 .033 planning on grad school -.022 .005 <.001 .025 .006 <.001 -.003 .005 .612 income employed -.035 .006 <.001 .042 .008 <.001 -.007 .007 .301 student income .002 .000 <.001 -.001 .001 .126 -.001 .000 .003 alternative financial support scholarship or grants .020 .003 <.001 .001 .004 .787 -.021 .003 <.001 parental support .021 .002 <.001 -.032 .003 <.001 .011 .003 <.001 n=24,121 limitations this research is not without its limitations. there may be selection effects due to the sampling design. institutions voluntarily participated in the scfw which may induce a selection effect if these institutions were systematically different than the population of higher education institutions in the us. additionally, students were randomly sampled but voluntarily completed the surveys. although the sample looks comparable to national statistics on the college student population, we cannot rule out selection effects. the distribution of students across different institutions is not in line with national averages. the sample in this study is more heavily weighted towards 4-year public schools when compared to the national average (81% vs. 43%). the sample also excludes students from 4-year for-profit institutions and under-represents 2-year public schools (9% vs. 30%) and 4-year private schools (10% vs. 21%). our sample might be why we find a significantly smaller number of students who are debt averse as compared to other studies (e.g., boatman et al., 2017; goldrick-rab & kelchen, 2015). heckman et al. 87 table 3. 2020 tobit regression results on amount willing to borrow variable dy/dx (me) std. err. p-value control rank (ref: year 1) year 2 -0.109 .373 .770 year 3 0.338 .372 .363 year 4 1.049 .397 .008 more than 4 2.520 .621 <.001 first generation 0.356 .284 .211 non-traditional -1.424 .408 <.001 gender (ref: male) female -0.426 .292 .144 prefer not to say/other -2.464 .851 .005 race/ethnicity (ref: white) black -2.337 .575 <.001 hispanic -2.830 .428 <.001 asian -4.953 .458 <.001 other -2.150 .437 <.001 cost tuition cost (net price) 5.485 .311 <.001 tuition type (ref: public in-state) public out-of-state 5.631 .366 <.001 private -2.365 .603 <.001 educational goals gpa -2.142 .259 <.001 good investment 3.138 .290 <.001 major: arts & humanities -1.755 .386 <.001 major: business 0.530 .426 .213 major: education -0.640 .546 .241 major: health 3.086 .411 <.001 major: social science 0.317 .396 .423 major: stem 0.538 .380 .157 major: vocational -0.092 1.384 .947 major: other -1.662 .608 .006 degree goals (ref: 2 yr) 4-year 2.006 .837 .019 certificate 2.415 1.839 .184 other 4.695 1.565 .002 planning on grad school 2.821 .278 <.001 income employed 1.430 .358 <.001 student income -0.136 .023 <.001 alternative financial support scholarship or grants -3.041 .161 <.001 parental support -2.192 .148 <.001 n=20,145 financial services review, 33(2) 88 conclusions, discussion, and implications the decision to borrow for college is a complex process and self-assessment of the risk involved in borrowing is one aspect contributing to this complexity (dowd & coury, 2006; heckman & montalto, 2018). nevertheless, this research indicates that many students are rational about the amount of debt they are willing to accumulate. as expected from the two-period human capital model (checchi, 2006), the results show that higher costs of education (i.e., attending at institution with higher tuition, expecting to take longer to complete a degree) and educational goals (i.e., perceiving tuition as a good investment, pursuing more than 2-year degree, planning to attend graduate school, enrolled in a high-paying major) were positively associated with willingness to borrow. findings on income were mixed. there does appear to be a positive effect of income on willingness to borrow at the lower end of incomes, but as income increases, the amount willing to borrow is less. alternative financial support (i.e., more financial support from family or financial aid from scholarships and grants) is negatively associated with willingness to borrow, which is also consistent with the two-period human capital model (checchi 2006). while there have been concerns about whether students choose not to borrow enough for college education (cadena & keys, 2013) and whether students over-borrow (avery & turner, 2012), our findings showed that students’ willingness to borrow is aligned with the directions from the economic model of cost and benefit analysis. human capital theory is based on the economic assumption that individuals are rational with full information. since our findings confirmed that college students’ willingness to borrow reflects their rationale given the information, accurate information needs to be available for students to make optimal decisions on borrowing. evans and boatman (2019) found that providing information on income-based repayment options to high school students before they decide on college enrollment improved the likelihood of enrollment for populations who are averse to borrowing. professionals advising families on college planning can leverage tools like the net price calculator to provide comprehensive guidance (u.s. department of education, n.d.). these provide information on net price that the family has to pay after accounting for the scholarships and grants from the sticker price. one study found that there still exist some gaps between npc estimates and actual costs due to the variations in individual financial aid packages, so some caution is needed in accurately pricing an individuals’ educational costs (anthony et al., 2016). in addition to the cost of attendance, there are information gaps in estimating the benefits from higher education as well. while median earnings for those with a bachelor’s degree are approximately $29,000 more than those with a high school degree, considerable heterogeneity exists in incomes of those with college degrees (ma & pender, 2023). for example, in 2018 and 2019, mid-career median earnings were $43,700 for early childhood education majors but they were $100,000 for computer science majors (ma & pender, 2023). detailed information not only on the cost but also on the benefit of college degree will help students make optimal decisions regarding college enrollment, major selection, and borrowing decisions. for professionals advising students and their families, their expertise should include realistic income projections based on chosen majors or fields of study along with long-term strategies for managing student debt, incorporating federal tax benefits, income-based repayment options. by integrating these elements, we can equip students and their families with the comprehensive information needed to make informed, rational decisions about higher education investments. the results confirm many things as expected, however there are bigger questions that need to be investigated. the first question concerns college access. students who enroll in college seem to make rational borrowing decisions. but, what about the students who decide not to enroll? the finding that students with lower educational attainment goals are less willing to borrow might be rational given the information they are using to make that decision, but is that in their best interest? are these students systematically underweighting the value of a degree, or making a rational decision based on their career goals and/or knowledge of the declining wage premium heckman et al. 89 (valletta, 2017)? the findings that socioeconomic factors are more to blame for dropping out of college and aversion to borrowing point to a potential problem. for example, non-traditional students were found to have lower willingness to borrow in our analysis. if the goal of public financing of education is to allow for equal access, then this population should be examined more carefully and more efforts should be spent on which policy tools are effective to assist them to make rational decisions on attending and borrowing for colleges. the second question raised by our results concerns persistence. what if the previously ‘rational’ decision turns out to be sub-optimal when circumstances change? the rise in past due balances and delinquency rates of student loans (li, 2013; muller & yannelis, 2019) casts doubt on previously ‘rational’ borrowing decisions. for example, the sharp increase in the direct cost for education, depressed future job market, or losing current income or financial support can make college students feel that completing their undergraduate degrees is no longer a good investment. those who already borrowed heavily for their college education may be struggling with financial stress and may fail to persist. even if they persist, the debt burden may impact students’ post-graduation plans (valez et al., bentz, 2019). for those who decided to enroll based on their original rationale, more attention is needed when the circumstances related with borrowing decisions change. on top of all of this are the recent policy decisions affecting student loans since the start of the pandemic including the repayment pause, restarting payments in 2023, and the save plan introduced in 2023. these changes and the inherent complexity of the student loan system in the u.s. may be leading to a sense of overwhelm and confusion for borrowers. college enrollments have been decreasing since their peak in 2013 (national student clearinghouse, 2017). while this trend might be due to economic conditions such as job growth, it might also be due to growing education costs and reluctance to take on debt. this space should be watched closely, particularly by those in public policy. if post-secondary education becomes too expensive for potential students, then the mechanisms used to support higher education ought to be revisited. this might be through increasing public expenditures, restructuring loan programs, reigning in costs, or making some post-secondary education, like community college, free for all students. if the problem is one of risk aversion, then educating students on the specifics of student loans, such as repayment options, might help alleviate concerns. references alstadsæter, a. (2010). measuring the consumption value of higher education. cesifo economic studies, 57(3), 458479. anthony, a. m., page, l. c., & seldin, a. (2016). in the right ballpark? assessing the accuracy of net price calculators. journal of student financial aid, 46(2), 25-50. avery, c., & turner, s. (2012). student loans: do college students borrow too much-or not enough? journal of economic perspectives, 26(1), 165-192. doi:10.1257/jep.26.1.165 baum, s., & payea, k. (2012). trends in student aid 2012. the college board. baum, s., & schwartz, s. (2015). student aid, student behavior, and educational attainment. in a.a. author (ed.), decision making for student success (pp. 38-62). routledge. becker, g. (1964). human capital. 2nd edition. columbia university press. ben-porath, y. (1967). the production of human capital and the life cycle of earnings. the journal of political economy, 75(4), 352-365. boatman, a., & evans, b. j. (2017). how financial literacy, federal aid knowledge, and credit market experience predict loan aversion for education. the annals of the american academy of political and social science, 671(1), 49-68. boatman, a., evans, b. j., & soliz, a. (2017). understanding loan aversion in education: evidence from high school seniors, community college students, and financial services review, 33(2) 90 adults. aera open, 3(1), doi: 2332858416683649. cadena, b. c., & keys, b. j. (2013). can selfcontrol explain avoiding free money? evidence frominterest-free student loans. review of economics and statistics, 95(4), 1117-1129. cares act, pub. l. no. 116–136, 134 stat. 281 (2020). carnevale, a. p., cheah, b., & wenzinger, e. (2021). the college payoff: not all college degrees are created equal. georgetown university center on education and the workforce. retrieved from https://cew.georgetown.edu/wpcontent/uploads/cewcollege_payoff_2021-fr.pdf cha, k.-w., & weagley, r. o. (2002). higher education borrowing. financial counseling and planning, 13(1), 61–73. checchi, d. (2006). the economics of education: human capital, family background and inequality. cambridge university press. college board. (2023). trends in college pricing and student aid 2023. retrieved from https://research.collegeboard.org/media/ pdf/trends%20report%202023%20upd ated.pdf cunningham, a. f., & santiago, d. a. (2008). student aversion to borrowing: who borrows and who doesn't. institute for higher education policy. retrieved from https://files.eric.ed.gov/fulltext/ed5036 84.pdf daly, m. c., buckman, s. r., & seitelman, l. m. (2020). the unequal impact of covid19: why education matters. frbsf economic letter, 17(5), 1-5. damon, a., & glewwe, p. (2011). valuing the benefits of the education provided by public universities: a case study of minnesota. economics of education review, 30(6), 1242-1261. doi:10.1016/j.econedurev.2011.07.015 delaney, l., harmon, c., & redmond, c. (2011). parental education, grade attainment and earnings expectations among university students. economics of education review, 30(6), 1136-1152. doi:10.1016/j.econedurev.2011.04.004 desjardins, s. l., & toutkoushian, r. k. (2005). are students really rational? the development of rational thought and its application to student choice. in j. c. smart (ed.), higher education: handbook of theory and research (vol. 20). springer. https://doi.org/10.1007/14020-3279-x_4 despard, m. r., perantie, d., taylor, s., grinstein-weiss, m., friedline, t., & raghavan, r. (2016). student debt and hardship: evidence from a large sample of low-and moderate-income households. children and youth services review, 70, 8-18. dowd, a. c., & coury, t. (2006). the effect of loans on the persistence and attainment of community college students. research in higher education, 47(1), 33-62. fox, j. j., bartholomae, s., letkiewicz, j. c., & montalto, c. p. (2017). college student debt and anticipated repayment difficulty. journal of student financial aid, 47(2), n2. frotman, s., & williams, r. (2017). new data documents a disturbing cycle of defaults for struggling student loan borrowers. consumer financial protection bureau. retrieved from https://www.consumerfinance.gov/about -us/blog/new-data-documentsdisturbing-cycle-defaults-strugglingstudent-loan-borrowers/ furquim, f., glasener, k. m., oster, m., mccall, b. p., & desjardins, s. l. (2017). navigating the financial aid process: borrowing outcomes among firstgeneration and non-first-generation students. the annals of the american academy of political and social science, 671(1), 69–91. goldrick-rab, s., & kelchen, r. (2015). making sense of loan aversion: evidence from https://cew.georgetown.edu/wp-content/uploads/cew-college_payoff_2021-fr.pdf https://cew.georgetown.edu/wp-content/uploads/cew-college_payoff_2021-fr.pdf https://cew.georgetown.edu/wp-content/uploads/cew-college_payoff_2021-fr.pdf https://research.collegeboard.org/media/pdf/trends%20report%202023%20updated.pdf https://research.collegeboard.org/media/pdf/trends%20report%202023%20updated.pdf https://research.collegeboard.org/media/pdf/trends%20report%202023%20updated.pdf https://files.eric.ed.gov/fulltext/ed503684.pdf https://files.eric.ed.gov/fulltext/ed503684.pdf https://doi.org/10.1007/1-4020-3279-x_4 https://doi.org/10.1007/1-4020-3279-x_4 https://www.consumerfinance.gov/about-us/blog/new-data-documents-disturbing-cycle-defaults-struggling-student-loan-borrowers/ https://www.consumerfinance.gov/about-us/blog/new-data-documents-disturbing-cycle-defaults-struggling-student-loan-borrowers/ https://www.consumerfinance.gov/about-us/blog/new-data-documents-disturbing-cycle-defaults-struggling-student-loan-borrowers/ https://www.consumerfinance.gov/about-us/blog/new-data-documents-disturbing-cycle-defaults-struggling-student-loan-borrowers/ heckman et al. 91 wisconsin. in h. hershbein & k. hollenbeck (eds.), student loans and the dynamics of debt (pp. 317–378). w.e. upjohn institute for employment research. https://doi.org/10.17848/978088099487 3.ch10 hart, n. k., & mustafa, s. (2008). what determines the amount students borrow? revisiting the crisis–convenience debate. journal of student financial aid, 38(1), 17–39. heckman, s. j., lim, h., & montalto, c. p. (2014). factors related to financial stress among college students. journal of financial therapy, 5(1), 19–39. https://doi.org/10.4148/1944-9771.1063 heckman, s. j., & montalto, c. p. (2018). consumer risk preferences and higher education enrollment decisions. journal of consumer affairs, 52(1), 166–196. https://doi.org/10.1111/joca.12139 hoynes, h., miller, d. l., & schaller, j. (2012). who suffers during recessions? journal of economic perspectives, 26(3), 27–48. jacob, b., mccall, b., & strange, k. (2013). the consumption value of education: implications for the postsecondary market. american economic association papers and proceedings, 103(5), 18811915. letkiewicz, j. c., lim, h., heckman, s. j., & montalto, c. p. (2019). parental financial socialization: is too much help leading to debt ignorance among college students? family and consumer sciences research journal, 48(2), 149–164. li, w. (2013). the economics of student loan borrowing and repayment. business review, 3, 1-10. ma, j., & pender, m. (2023). education pays 2023: the benefits of higher education for individuals and society. college board. retrieved from https://research.collegeboard.org/media/ pdf/education-pays-2023.pdf martin, e. c., & dwyer, r. e. (2021). financial stress, race, and student debt during the great recession. social currents, 8(5), 424–445. mckinnery, l., & burridge, a. b. (2015). helping or hindering? the effects of loans on community college student persistence. research in higher education, 56(1), 299–324. national student clearinghouse research center. (2017). current term enrollment estimates spring 2017. retrieved from https://nscresearchcenter.org/wpcontent/uploads/currenttermenrollment -spring2017.pdf oreopoulos, p., & petronijevic, u. (2013). making college worth it: a review of the returns to higher education. future of children, 23(1), 41–65. oreopoulos, p., & salvanes, k. g. (2011). priceless: the nonpecuniary benefits of schooling. journal of economic perspectives, 25(1), 159–184. https://doi.org/10.1257/jep.25.1.159 perna, l. w. (2008). understanding high school students willingness to borrow to pay college prices. research in higher education, 49(7), 589–606. https://doi.org/10.1007/s11162-0089095-6 sigelman, m., & selingo, j. (2021). making the bachelor's degree more valuable. the burning glass institute. retrieved from https://www.burningglassinstitute.org/re search/making-the-bachelors-degreemore-valuable simon, h.a. (1955). a behavioral model of rational choice. quarterly journal of economics. 69(1): 99–118 u.s. department of education. (2020). direct loans entering default. retrieved from https://studentaid.gov/sites/default/files/ dlenteringdefaults.xls https://www.google.com/search?q=https://www.google.com/search%3fq%3dhttps://doi.org/10.17848/9780880994873.ch10 https://www.google.com/search?q=https://www.google.com/search%3fq%3dhttps://doi.org/10.17848/9780880994873.ch10 https://www.google.com/search?q=https://www.google.com/search%3fq%3dhttps://doi.org/10.4148/1944-9771.1063 https://doi.org/10.1111/joca.12139 https://research.collegeboard.org/media/pdf/education-pays-2023.pdf https://research.collegeboard.org/media/pdf/education-pays-2023.pdf https://nscresearchcenter.org/wp-content/uploads/currenttermenrollment-spring2017.pdf https://nscresearchcenter.org/wp-content/uploads/currenttermenrollment-spring2017.pdf https://nscresearchcenter.org/wp-content/uploads/currenttermenrollment-spring2017.pdf https://www.google.com/url?sa=e&source=gmail&q=https://doi.org/10.1257/jep.25.1.159 https://www.google.com/search?q=https://www.google.com/url%3fsa%3de%26source%3dgmail%26q%3dhttps://doi.org/10.1007/s11162-008-9095-6 https://www.google.com/search?q=https://www.google.com/url%3fsa%3de%26source%3dgmail%26q%3dhttps://doi.org/10.1007/s11162-008-9095-6 https://www.burningglassinstitute.org/research/making-the-bachelors-degree-more-valuable https://www.burningglassinstitute.org/research/making-the-bachelors-degree-more-valuable https://www.burningglassinstitute.org/research/making-the-bachelors-degree-more-valuable https://studentaid.gov/sites/default/files/dlenteringdefaults.xls https://studentaid.gov/sites/default/files/dlenteringdefaults.xls financial services review, 33(2) 92 u.s. department of education. (n.d.). net price calculator center. retrieved from https://collegecost.ed.gov/net-price valletta, r. g. (2017). recent flattening in the higher education wage premium: polarization, skill downgrading, or both? in education, skills, and technical change: implications for future us gdp growth. university of chicago press. https://collegecost.ed.gov/net-price pii: s1057-0810(02)00097-5 ��� ����� � ��� � ��� ���� � �� �� �� ��� ����� � ����� � ��� � ���� ������ �� ������������ ������ ��������������� ��� �������� ���� �� �� �� ������� ������� �� ����� ��� ���� ��� ������ � ��!� � ���� �� "�#������� ���$����� ��%� ����� �& ������� ���'��� �����( �& )*������ )��(� ���#� ���� +��� �, -.-/� ��� �������� �� ��� � �! ������� �� ������ � �" #�$�� �� � �! ����$��� �% &��� �� � � �������� �� �'� ��� ������ �� ������� ��� ����� � ����� � ��� � ��� ���� �� �� ���� $ ����� ���� ���� � ��� �'$����� $ ����� ����� � ���� ���� � �� �� � �� ��'�� � �������� � ��� ���$���� (���(� ��� (���(� � � ����� � ��� ���� )������� (� ������� ���� ����������� ������ � ��� ��� � �� * ��' $ ����� ����� � ����� � ��� � ��� ���� � �� � ���� �� ���� � ��� $� � ����� ���� �� ������� ��� �� ��� � � �������� � ��� ��� � ���+� ����� ����������� ,�������� �� -������ #������ .��� 012 �(����&�������3 /��! /��! ��"! ���! 0�1 4��!� ��3 ,�������! ,� � ���� ������! 2��������! ���� � ��� � ���� ������ ���� � ��� $������� �� � �� � �� $����� � 3������� $�������� ��� ��$���� �$$ ���� �� ��� � ��� $������� ������ (��� � �� ��� ����� � ��� � ��� �$������� -'$����� $������ $�� ���� ��� #����� #��� ��� ����3�� � � �������� � ���� �$������ ����� �� �� $��� �� ���� � ��$ ���� ��� ������ ��� $� � ���� ��� ���� ��$������� $������� � � ������� �� (��� $������� � ��� � ��� ���� � ��� ���� � � ��� ������� �� (��� ����������� ��� ����� )�������� #� ����� �����( � 4� �5 �%617 � 7�* 8� 9 �9: 6 ��� ��� ���� � ,�������� �� -������ #������ .��� ,..� # � 7 � * 8 � 4 � 5 % � * 7 ���� �$������ �'$ ����� � � ����� � ��� ����� � ��� � � ��� ������� ��� ����� � ��� ;#<)<� ��� ;�#� <� )� �� � ��� ���� ��� �� ������� -'��� �$ ��������� � ��� ������������� �� (��� �� �'$����� �� ����� � ��� ������� � � ��������� �$�� �=����� ���� (��� �� ��� ��������� �� �� ����� ��� ����� +� �� �$��� 4��� (((�(����� >���������� 5� � ?� ��$������ ����� � ����� ��*��%*111*@�1 ! �'� ��*��%*111*%��7� 15���( ��� �����3 (>a(����� >���������� � (����� �>�������a��� ��� � �� 4���� ��������5� ����b ������* �����a����� ���� 4�� ������������5� � ����� ��*�7@*�� *"�@"! �'� ��*�7@*�� *� %�� http://www.williamjennings.com ����������� �� � ��� ���� ���� �� ������������� #���� ���� $������� ��� ��� ������ ��� $� � ���� ���� ��� ��� � ��� ����� ���� ������ �� ������� �� $� � � ��� � ���+� $� � ����� .� ���� $�$� � (� �'� ��� ������ �� �������� ��� $ ����� ���� ���� � � ��� $ ����� ����� � ��� � ����� ��� � ��� �� �� �c ����� � ��� � ��� ���� �� ���� �������� �� � � ���� � ���� �����! ��� ������ �'� ���� ���� $� ���� �� ���� ���� � ��� ������ � �� $� ��� $������� ��� #����� #��� ��� ����3��� ���� �$$ �'� ����� ��%� � ����������� ��� ����� � ����� � ��� � ��� ���� � ��(� ��� � �������� �� ��� �(� ����� ��� ����� � ��� � ��� ����� $ ������ �� ���� ������ ������ ����� �� ���� �� �� �������� #� ���� � �� �� � �'� $��� � � �������� �� ������ ����3�� � �� � ��� � � �� ����� d�e� ��� 3 �3���� � ��� $����� � 3�� �� ����� � � �� � ��� � ��� ���� ��� � � ������ �� �� ��� �� � � ������ $������� ���� ��� ���� ff������* ��$�++ g������ ���� ������� ��� � ��� ��$������ ���� �(���� ���� � � �� ���� � ����� ��� $� ��� � ������ �� �� � ������ � ��� ��� �� ��� ���� ��� $������ ���������� (��� >�� �(������� �� ��3���*����3� $������ $���� 4��� ����� h )� ��� � �%%@5 ������� � ����� � ��� � ��� ���� � ������ � �������� � �� ��� 3 �� ��� � ���� ��� ��� ���� � �� ����'�� � ��i������ 2������ ����� � ��� � ��� ����3�� ����(� �� �� �� ���� ��� �(� � $� ���� � $��������� � ����� ����������� )� ��� ��e� � ��������� .�<� ��� @ �4e5�� ���� �� ���� � ����� � ��� � ��� � � ��'����� �� ��� ��� ���� �� � ���� ����� ������������+� 4�%%8� � 5 ������� ���� ����� ���������� ������ �� ����� �� � �� *��' ������� #������ (� ��� ��� �� ��� ����� ���(��� ��i�����*����'�� � ���� � ����� ��� ����� � ��� � ��� ����3�� �� ����� ����� ����3��� ?��� i�(� � ���� � ���� � .�i����� , �������� #��� ����� 4�.,#5 ��� ����� � ��� � ��� � � ���e�� �� ��� ?���� � , ��� .���' 4?,.5� ��� ���� ���� � ���� i�(� � � ���e�� �� ��� ��� �� ���� � ���� ����� �� ��� ����� ������� � ������ ��� $ ����� ����� � ����� � ��� � ��� ����3�� �� ����� �� ��� � ���+� $� � ����� ��� ff� ��++ � ��� $� � ���� �������� 3������� ������ ��� ��� $ ����� ����� � ��� � ��� ���� � �� �� � ���� �� ����� � ��� � ��� ���� � 4#����� �%%75� <� ������������ 4�%%8� � 5 ��� ) ��� � ��������� ��� 0��� 4� �5 �� ���� ���� ��������� ����� $�����*����� �� ��� $� � ���� ��� � � �������� � ��� ��� � ���+� ����� ����������� )� �'� $��� ��$$��� � � ���+� ��e ���� ���� ����� � � 7 97 ����e�9����� ����� ���������� ��� �� ��������� ��� 3������� ������c��� � ��������� ��3������ � ��� � ��� $� � ����c�� ��� ���� �� ��� ����� �'� .�� � �� $� � ����� (���� �������� ����� � ��� � ��� ���� �� �� �� �79�7 �� ����e�9 ������ . ��� � ��� ���� �� ������j� ���� ����� � ��� � ��� ���� � �� � ���� �� ��� $� � ���� ���� �� (��� ��� ��� �� ��� ����� ����������� #������ � ����(� ��� ���� ��� � ���� ��$$� �� �� ��� ����� ������� �� ��� � ��������� �$$ ���� � ����������� � � ���+� ����� �'� #������ 1 $ ������ ������� � ��� ;�#� ����� � ��� � ��� ����� � �� ���� ������� ������ �� ������� ���� ����� � ��� $ ����� ����� � ��� � ��� ���� � �� �� �� $ ����� �� ���� ����� ������ ��� � ���=�� ��� )������� (� $ ����� �� �'� $�� ���� �� ���� ���� ��� � $� ����� � ��������� ��� $ ����� ����� � ��� � ��� ���� � �� �� �c�� ���� ���� ����� � ��� � ���c�� � � ���+� $� � ����� ��� 3��� ������� ���������� � ������ �� ��� ���������� .� ���� �������� (� $ ����� �(� ���3������� � ��� � ��������� �$$ ���� � ����������� � � ���+� ����� �'� )� ��� (� ����( ��� ���� ��� � ���� ��� ����� ��� ����� � �� � �*��' � +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< �� ��� � �� *��' ��������� � $� � ���� �� $������� #������ (� ������� ��� =������� � (���� ������� ������ ��� 3������� ������ �������� �� � � ��������� $� � ����� ������ �� �������� �� �� �'$����� ���( � � � ���+� � �� $� � ����� 8�6� )�&� �5��= #� �*� �&�� 5��= &*��� ������������ 4�%%8� � 5 � ���� ���� � � ���+� ����� ���������� ������ �� ����� �� � �� * ��' ���� ������� ����� ��� �� ����� � � $� ������ (��� � �� *��' ����� ��� � ��������� �$$ ���� �� ����������� ��� ����� ���������� ���� ��� ����������� ���(��� �� � �*��' ��� � �� *��' ����� )� �'� $��� ��$$��� � � ��� ��� � :� � ���� ��� ���� �� � @ �4e5 ��� � :� � ���� ��� ���� �� � ��'���� �������� .� ��� ��'���� �������� ���� � e�� ����� ��� ���e ����� � � :� � � ��� � ��������� �$$ ���� (���� ������� ���� � ����� �������� (� �� :� � � k��� ��� @ �4e5 �������� �� � �*��' ����� (���� ��� ��'���� ������� �������� � �� *��' ����� < ��� �� ������ ��� �� ��� ����� ��� �� ����� � ��� :� � �� ��� @ �4e5 �� �� ��� � ��� :� � �� ��� ��'���� �������� < �� *��' ���� ������ �� �� $� �� �� � �� *��' ����� <���� ������ ������ 3 �� �� ����� ��� �� � �� *��' ������� ��� ��� � ���+� ����� ���������� ������ �i��� ����� � �� *��' ������� ������������ 4�%%8� � 5 ��������� ������ �� ������� ��� ����� ���� � � e�� ������ �� � �� *��' ������ � ������ ���� �� ���������� $������� 4����� @ �4e5� @ 14�5� 0����� ����5� ���*���������� .�<�� ���* =����3�� ��'*�� � �� ���������� ��� ��'���� ��������� 8�8� +��� >>��*���?? �� ��� &���(� %� �&�(��@ #���� 4�%%75� ������������ 4�%%8� � 5� ��� ) ��� �� ��� 4� �5 � ��� ���� � � ��� ���� ��� � ��� $� � ���� ������ �i��� ��� $ ����� ����� � ��� � ��� ���� � �� �� � ���� �� #����� #��� ���� �� $��� $�������� ��� ����� � ��� � ���� ���� � ��� ���� � �� �� � � � � ����������� �������� �� ������� � ��� � ��� $ �$� ������ 4��� ,��� ��� 2����� h ����� �%%@! 0�����e���� #�� *��?��� � h #������ �%%�! k��� l����� h ��������� �%%8� � ��� ���� �5� ��� �'������ (��� ����������� � � ���+� ����� ���������� 4��� ,���� h #�� � �� �%% ! #������� � � � ��� ��� ���� �5� #����+� � ��� ��� � �������� (����� �� ������� �� ����� �� ��� � ��� $� � ���� �� �� ������� ff3������� ������ ���� ��� (���� �� (������ �� ���� � �$������ ���� � ���� ���� ��� �� � � � �$������ ����� ����� ��( � �� ��� � �� ��� ��� � � � ��++ 4�%%7� $� �75 #�� ���� �������� ��� $ ����� ����� � ��� a�&� �5 ��= #����� #��� ��� $�� ���� ��� ���� ��� � ���*���� � ���� i�(�� .� ���� $�$� � (� ������� ��� $ ����� ����� � ����� �&�� 5��= � ������ ������ �� ���� �������� �� � �*��' � � �� *��' ������ � ��� � ��� ���� � �� �� �� ���� ��������� ��� � � �������� � ��� ��� ���� � ��� � ��� � ���+� ����� ����������� ����� ��� �������� �� ������� � �'����� ��� ��� � ��� ���� � �� �� �� �� � $� ���� ��� (����� (� ��� �� � �*��' � � �� *��' ������� � � � �������� ���������� ������� ��� � � � �� �� ��� ����� � ��� � ��� ����� �c)���� ,��� l��� �� ��� ��� �-&;m 4ff�-&;m++ �� ���������� � ��� �������� �� ����3�� �� ��� ��� � ����� � ��� � ��� ����� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< �� � ����� ���� ��� �%8" ������ � ���� � ��� �� � <�� 4��� ;�#� ?���� ����� � � n�@ %55� )���� ,�� �$$���� �� �� ���� � �� � (�� ���� �� ��� �� ���� �� � � #�$�� �� 8� �%8 � l��� �� �� �$$���� �� �� ���� � �� � (�� ���� �� � #�$�� �� 8� �%8 �� ���� 1�� �%8"� #� ���� � �� � (�� ���� ��� �� ���� � �� ���� 1�� �%8" ���� � �� �7 ��� � � �� ����� ����� ��� l��� �� �� ����� � ����� ��� �-&;m ����� ��� ������ � :1 � �� �� �������� ������ -'��$� �� ������� �� �� �������� �� ���� � �� � ��� �� �� � ��� � �� �� �������� � ��� � ��� $��� o��� ���� �� � �� #� ���� <���� ��� � �� �p�? ����(� ���%� ������ � ���� ����� ��� ��� � ����� ��� �� ��� ����� � �� $����� ��� � � �� ����� ����� � �� � �j�� e�� ��$���� � ���� ����� � )�( ����� � $� ������ ��� ������ ���(��� ��� �� �� ����� � ������ �� ���� ��� ���� � ��� � ���� ����� � �� ���� ���� ���� ������������ ��� ���� ������ �� � ����� (�� ���� �� � �� ���� 1�� �%8"� ����� ���� ������ (��� ��$��� �� ���������� �� �� �������� � =���e ��������� ����� �� ��� ��������� �����$��� (� $ ����� ����(� �������� ���� ���� ����� (�� �'$��� ���� ����� � �� �� � 4����� �8 ��� � � ����� 5� ����� ��e� 4����� �� ?������ � /��� ��5� ���9� ���� ��� ��� �� 4����� � ���� ���� 8q5 ������ � ��� ���� �-&;m ��� ��e� ��� :1 � �� �� �������� ������ ��� (������ $��� ������ �� 4� �5 $ ������ � � � �������� ���������� � ��� ������ � ����� (�� ���� �� � �� ���� 1�� �%8"� �� �� e�� ��� ���� � ���� ����� � � g��� ,�� < ����� ,� �������� ��� ����*� *������ ��>��� ���� �� ����� ��� ������ g��� ,�� < ���� ��� ,� ������� � � ��$�� (��� ��$���� ����� � (��� ���� � � �� � e�� ��� ���� ?� ��� <����� ,�� ��� 4?<,5 �� ��� $ ����� � g��� ,�� < ���� ��� ,� �������� ;��� )���� ,��� g��� ,�� < ���� �� ��� � ��� �� ��� 3��� ���� $��� �'������� ������� ���� �� � ��j� ���� ����� ;��� l��� �� �� ��� �-&;m� g��� ,�� < ���� �� ��� � ��� �� ��� ��� ��� ���� $�� � ��� ������� 1" ������ �'������� �������� ����� � ?� $� ���� � �� �� ;�#� ����� � ��� � ��� ����� � )���� ,�� l��� �� �� �-&;m g��� ,�� < ���� )���� ���� $��� �'������� ������� <�� ��� ���� $�� �� ������� 1" ������ �'������� ������� <�� ��� ���� $�� �� ������� 1" ������ �'������� ������� ,� ������� � g��� ,�� < ���� ������� �� ��� � ��� 7 q � ��7q 4��� � � �� ���� � � 5 7 q � ��7q 4��� � � �� ���� � � 5 g� � � ��� "�� @ q � 1�7q 4��� � � �� ���� � � 5! ��������� ��� "�� 7 q � ��7q 4��� � � �� ���� � � 5 ?���*� *������ ��>��� ��� g��� ,�� < ���� ��� ����� (��� ?,. ��i����� ��� g��� ,�� < ���� ��� ����� (��� ?,. ��i����� ��� g� � � ��� "�� g��� ,�� < ���� ��� ����� (��� ?,. ��i����� ��� ���� �q! �� ��� "�� g��� ,�� < ���� �� ��>����� �� � ���� ���� l��� �� ��! � �� ��� "�� ���� ��>����� g��� ,�� < ���� ��� ����� (��� ?,. ��i����� ��� ���� �q �� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< ,� ������� � )���� ,�� ��� l��� �� �� �� 7 �� ��7� 4��� � � �� ���� � � 5� -���������� ���� ���� � ��� �� ������� 7 q � ��� g��� ,�� < ���� � �� � ��� � � �� ����� ��� ��� ,� ������� ��� ����� �� ��7q � ��� � )� �'� $��� � �� ���� ��� ��� � �� ���1 ��� � � �� ����� ��� ,� ������� �� 71��7q� ��� ,� ������� � �-&;m �� @ �� 1�7� 4��� � � �� ���� � � 5� ���� ���� ���� � �� � ��� � � ��� �� ������� @ q � g��� ,�� < ����� ��� ��� ,� ������� ��� ����� �� 1�7q � ��� � < �� 1 ��� � � �� ����� � ��� �� ������� �7q � ��� g��� ,�� < ���� ���� ��� �� �� ����� �� p��� ��� ,� ������� ������ �7q� �� ���� ��� ��� ����� ;���e� ��� $ ����� ����� $�������� ����� � ��� � ��� $�� ���� ������ ���� � ������ ��>��� ����� ;��� )���� ,�� ��� l��� �� ��� g��� ,�� < ���� ��� ����� �������� (��� ��� ?,.*� ��i����� ���� (�� � ?,.*� ������� ��� ?���� � , ��� .���' � ; ��� ���� -� �� � ��� ?�� ���� �� e� �� ;��� �-&;m� g��� ,�� < ���� ��� ����� �������� �� �q ���� ���� ��� ?,.*� ��i����� ���� o� ?pd< �� � ����� � ?,.*� �� ���� ���� �q� ;��� �-&;m� �� ��� "� ��� � �� � ���*�� � ff�����*�$++ ���� ��>���� ��� � ��� $�� �� ��� ����� �� (���� �� ���� l��� �� ��� ���� �����*�$ �=�� �� �(� ��>��� ����� g��� ,�� < ���� �� ������� �� ��� ����� �� (���� ���� ���� (������ ��� �q ��i����� ��� ��� ,� ������� �� ������� �� ��� ����� ,� ������� ���� l��� �� ��� ;��� �-&;m� � �� ��� "� g��� ,�� < ���� ����� ��� ����� �������� �� �q ���� ���� ��� ?,.*� ��i����� ���� ��� &)<# (������ 4� �5 $ ������ � � �������� �� � ����� ����� ��� ��� � ��� $����� � ������ ������������� �� �������� �� ��� ��������� ��6� "��*�� �� �� ������ ��� �� ������� �� � � ����������� ��� $ ����� ����� � ����� � ��� � ��� ���� �� ����� ������� ��� ��� � �� � �*��' � � �� *��' ���� �� ��� �$$ �$ ���� �������� ���� ��� ��� $ �>����� ���� � �� �� � ��� 3 �� �(� ������ ��� �� ���� �� =���e��� )� ������ ��������� �� ��� � (� ���� ��� ��� ����� � ����� � ��� � ��� ���� � �� ��� $ ����� ����� � $ �>����� � �� *��' ���� �� )� $ ����� ����� ������������� ��� �������� ��� �� ��� ����� �� ��� �.,# (��� ��� ��� ������� �� ��� ������ � ��� �'$����� ���� � �� �� � �.,# $�� � 3'�� ��� ��� � ���� ��� ���� ��� �� $ ����$��� ��� ��� $ ����$�� �� ��� ����� ���� ��� ����� �� ?,.*;� ��� ?,. ��i����� ��� � ; ��� ?���� � �� ����� $ ����$�� ��� ���� ��� $�� ���� ��� ���� (��� ��� ����� � ��� ?,.� ���� � � ��i�����*$ ������� ���� ������ <� ����� ��� �.,# ��� ����� ��� �� ���� �� �������� ��� ��i�����*��>����� � ��� ����� � ��� � ��� ���� � �� �� � ���� ���� � � ���� �� �������� ��� ���� �� ��� �� ���� ���� � � ���� �� �������� �� ���� ���� �� )� �%�@6�%%%� ��� �� ������� ��� 3����� ���(��� ?,.*���� ?,.*; �� �%%%8� ��� ��� � �� � ���������� �� � ���� ���(��� ���� ����*�� �j�� �� ������� ��i����� ����� ?����=������� ��� � �.,# ������� (���� ��� �� ?,.*;� ���� ��� ���� �� ���� ���� � ����� � ��� � ��� ���� �� (���� ����� �� ?,.*�� l���� � ��� (� ���� � ��� ����������� ��� � � �� ��� �� ?,.� ) ��� �� ��� 4� �5 ��e��(����� ��� ?,. �� � ���� ��� ���� ������� �� ����� �.,# �� ����� ��� ���� � �� �� �� ��� �������� ��� ��� ������ � �.,# ���(����� ��� ��� ��� ����� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< �1 ��� 3��� ����� �� ��� $ �>����� ���� � �� �� � �� ���� ��� ��� $ �*��' ����� � ����� � ��� � ��� ���� � �� ��� $ ����� ����� � �'$����� $ �*��' ���� �� (���� �� ��� $ � ������+� � ��������� ����� � ���� ����� �� �����+� ������ <� (� ����� ���� ���� �� �� � ��� � ��� $ ����� ����� � ���� � �� ���� �� ����+� �'$����� �� �� ��8� "((*�� ����� �� ���� ��� ��� � �� *��' $ ����� ����� � ����� � ��� � ��� ���� � � ������ ���� ��� ������ � ���� �� ����(�� ?<, � g��� ,��< ����� ,� ������� 4�5 , ����� ����� �� � � ��'�� � ?<,������$�� 4�5 , ����� ����� � �� ��'�� � ?<,������$��� ��� �� 415 (�� � � �� ��� �'$����� � ����� ��' ���� �����$��� � � �'$������ �� ��� ��'� $� �� �$� ��� ��� ���� �� ������ � � ���� ���� �� #������ 1� g� � � ����� ���� ������� � ����������� � �����$��� ������� � �� $�� �� ���� ����� � ��� $ ������� �=�������� )� �'� $��� ���� � d�e� �� ������ ��� ��� �� �� ��� @@�� �� ����� �� ��� )���� ,�� � ��� l��� �� �� ����� � l� �� ��� � ��� �� �� ��� ,� ������� �� 7 q� . ��� g��� ,�� < ���� �� :@ � � ��� ?<, �� :� � � .� ��� 3 �� ��� � ��� � ���� �� (��� ������ :��""� � ����� :� � 9��� <��� ��� � @q �.,# ������ (� ���� ��� ��� $ �*��' ����� � d�e�+� ����� � ��� � ��� ���� � �� :� � 4�%��%5 � :187�8 � (�� � �%��%� ��� �����$�� � ����� 1� �� ��� $ ����� ����� � ����� � ���� � � ��� ?<, �� :�� #���� ����� � ��� ��� ��� ������ ��������� ���� ?<, �� �� �� �� ��� �� ��� ������ ����e � ����� �����$��� � � ��(��� �'$ ����� �� � ����$�� � ?<,� ����� � ������� ��� ����������� � � �����$�� ���� )���� ,�� � l��� �� �� ����� �� #���� d�e� >��� ��� ��� ��� � �� � � q $ ��������� ���� �� (��� ������ ��� 3 �� :� ?<,� �� ���� ��� ��� �����$�� �� � ������ $�� �� ������� �� ��� ��������� � ���� ��� � (��� �� �� �������� ������� ������� g���� �� �$����� � ������ ������ 4#������ � <���� ���� � 5� �� ��� � �%%8" 1 $ ��������� � ����� ����� �� ��� @7 ��� ����� �������� ��� ������ :� ��� $�� ���� #���� �� ���� $�� ���� ��� ���� (��� ��i������ ��� ��� $�� ���� � ��� ��������� ��� �'$����� ��� $�� ��� �� k�� � �� ����� : �%%8" 1� ��� � �� � �%%� %� $ ��������� ���� �� (��� ���� �� ��� @"� ��� ������ ��� ��� � :� ��� $�� ���� ���� $ ����� �� �$����� �� ���� ��� �� �� �� �����c��� ��� � ��� � ������ ������c(�� � ��� � �� � � @ $ ��������� ���� �� (��� �� ����� ��� ������ ��� ���� :� ��� $�� ���� ��� $ �*��' �����$�� �� ��� �� � ��� $ ����� ����� � ���� ��� +� �'$����� ��� ���� �� <��� ��� � @q �.,# ������ �� �� ��� �� � ��� � �%%8" 19��� @�� � � �%%� %�9 ��� @�� � � � � � � � @94�� @5�"� ��� +�(( �� ��� 0�* ��( $ ������ �.,# ������ �� ��� ��� � ��� � ���� � g����� o���� h g���� ����� ���� .�i�����*.���'�� � ���� � #��� ������ ��� $ �*��' ����� � d�e�+� ����� � ��� � ��� ���� � �� :� � 4�%��%5 � :187�8 � . �� �'$���� �� �� �� ��� �8q ��' � ��e��� ��� � �� *��' ����� �� ����� :����8 � �� � ��� ���� d�e� ������ ������� ����� � ��� � ��� ���� � �� ��� $� � ���� ��� ������� �� � ���� (� �� ����� :����8 � �� ��'��� �@ +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< ����� � ?����������� � �����$��� <�� , ��������� � ����� ����� ���� g��� ,�� < ���� 4:5 ) ������ ���� $�� ��� 4:5 -'$����� ��� $�� ��� 4:5 , ����� ����� � �'$����� ��� $�� ��� 4:5 <��� $����� @@*��� *��� ������ ��� ���� )���� ,�� � l��� �� �� ��� � ��� ����� � @@ �� � � � @7 �%%8" 1 � �%%8" 1 �%%8 "94�� @5� @" �%%� %� � �%%� %� �%%� %�94�� @5� �% �" 8�8� � �" 8�8� �" 8�8�94�� @517 8 �7�1 7 � �7�1 7 �7�1 7 94�� @51" �� � @ � � @ � @94�� @5�" �����$�� � �%��% <��� $����� 7 *��� *��� ������ ��� ���� �-&;m ��� � ��� ����� � 7 �� � � � � 7� �%%�8"� �94�� �5� � �%8�%8� �%8�%8�94�� @5 7� �%%7@�8 �94�� �5� � �%�78 @ �%�78 @94�� @5� "� �%7� @ �94�� �5�� � �87��87 �87��8794�� @5�� "� �%@%"78 �4�� �5" ���@ ���@%��� ���@%���94�� @5�� "1 �%@�1@� �4�� �57 ���@ �����8� �����8� 94�� @5�1 �� � @ �94�� �57� ���@ � 1 � 194�� @5� �����$�� � �8��1 � ���� �����$�� �� ��� $ ����� ����� � �'$����� ��� $�� ����� (�� � :� �� ��� ������ ��� � ��i�����*��>����� $�� ���� ��� �����$�� ��� �� �$$���� �� ��� ?� ��� <����� ,�� ��� �� ���� ��� ��� ����� � ��� � ��� ����3��� ��� :� �� �� ������� ���� ��� �������� ���� ��� ��� �� @q� � ���� �����$�� �� ��� �� � $ ����� ����� � �'$����� ��� $�� ����� (�� � :� �� ��� ?� ��� <����� ,�� ���� ��� �����$�� ��� �� �$$���� �� ��� ?� ��� <����� ,�� ��� �� ���� ��� ��� ����� � ��� � ��� ����3��� ��� �������� ���� ��� ��� �� @q� <� ��� "� ��� ������� ��� ) ������ �� �7�79 �7 7� (�� � �7�7 ��� �7 7 � � ��� ,� �������� � g��� ,�� < ���� �� � � ��� � �� ��� "�� ��$��������� <� ��� "�� ��� g��� ,�� < ���� �� ��>����� �� ��� ����� �� (���� ���� ���� � g��� ,�� < ���� ��� ���� ��>����� ����� ����� � ��� � ��� (��� ?,. ��i����� ��� ������� � ?,.*�q� )� �-&;m� ��� ����� � $� ������ �� ���� �� �� ��� � �� ��� @@� ��� �����$�� � ���� 7 *��� *��� ��� ������ ��� �� �� ��%��7� � ���@���88� � �8��1� + �+ � 0�� � �� $ � + � ���� �� ����� �� �� � � ��� ( � � #���� �#��! 6 � 78 � � 6 9 6 / : ; < � 7 ��;� � ���b*� �& �* ��������� ��� � � � �� ����� �� �� $�������� � ������ � � �� ����� � ���� ����� ��� ����� � ����� � ��� � ��� ���� �� )� ��� (� ���� ��� ��� ����� �� �� ������� ��� (��� ��� $�� ���� ���� ��� �� ��� ��������� � ��� ��� � ��� e�� �� ���� ����� � ��� � ��� �� � ��� �������� ��� � �� ���� �������� o� ���� ���� ��� ���� ��� �� �'$ ����� �� � � ��� .� ���� ���� (�� � � ������� � �� ���� ���� ��� ���� .� ������� ��i������ ��� �� ������� � ��� ���� #�$$��� ��� ���� � � ����� � ��� � ��� �� :� � � ��� � ���� �� ������� ���� ��� $ ����� ����� � �� � �*��' $�� ���� ������� � ��� � ��� � ����� � � :� � 4�� .� 59 4�� �5 ��� :� � 4�� .� 5�94�� �5�� ��$��������� #����������� 4�� .� 54�� ��5 � 4�� �5 $ ������ :� � 94�� ��5 ��� :� � 94�� ��5�� ��$��������� ���� ��� ���� ��� ��� ��� �� ������ �� ���� ��� ��� $ ����� ����� � ��� :� � ��� ���� � �� �� � . ��� :� � �� �� � ���� � �������� ��� $ ����� ����� � � �� *��' $�� ��� ������� � ��� � ����� (���� �� :� � 4�� .� 594�� �5� ����� ���� ��� ��i����� ��>��� ��� (���� ���� ���� ��� #����������� 4�� .� 54�� ��5 � 4�� �5� (� ��� :� � 9 4�� .� 54�� ��5�� ;���e� ��� ������� ���� ��� ����������� � ��� $ ����� ����� � � ����� ������� �=�� �� �(� ���� ����c�� ���� ��� � ��� �'$����� ��� ��� ��� �� ���� ��� � �'$����� ��i������ ;���� �� ������� ��� � $� �� � ������ �$(� � ���� ����� ����� #������ (� ��������� ��� ����� � ����� � ��� � ��� ���� � �� ��� $ ����� ����� � �=%����� ���� c�!�� #� � $ � ��������� (���� $ � � �� ���� ��� �� ����� �� ����� ���� c�!� �� �*$� (�&� �=%������� � ���� i�(� � � $� ��� ����� ���� �� � �'$�������� d�� �� ��� � �� d�e�� ��� @@*��� *��� ������ ��� ��� ��� l�� �� � �'$������� �� 1"�7 ��� �� ��� $ ����� ����� � � :� � � ��� ������� ��� � 1"�7 ��� � (��� ���������� �� @q �� ����� :1%7�8 � ���� $ ����� ����� �'����� ��� $ ����� ����� � �'$����� ���� i�(�� (���� �� :187�8 � -'$� � �������� ��������� ���� ��� $ ����� ����� � ���� i�(� �� ���� �� � �'$������� �� ������������ �������� �� �� ���� ��� $ ����� ����� � �'$����� ���� i�(�� )� ������ ��� ���� ���� ���� ����� �� ���� i�(� �� ���� �� � �'$������� � � �$$ �'� ����� �=��� �� ���� ���� ����� �� �'$����� ���� i�(�� (���� � $���� ���� ��� � � ���� ���� ���� � ��� l�(��� � (� ������ ���� ��� ��� ����� � ����� � ��� � ��� ����� �� ��� ���$��+� �� � �'$�������� )� � "7*��� *��� ���$��� ��� ���� �� ��� �� �� ��� >���� �� � �'$������� � �" ��� �� g�� ��� $�� ���� �� ��� ���$�� ��$��� �$�� (����� ��� ����� � ��� �� �� ��� 3 �� � ���� �� ���� ��� ��� ���� ��� ���� ��� �� ����� � ��(� ���� ���� � ���� ��� ���� ��� ��� ���� �������� <� ���� ���+� ���� ������� ��$���� �$�� (����� ��� ���� $����� 3� ��� ���������� )� �'� $��� ��� ���� � 3������� $������� �� �� ���� � ���� �� ���� (��� ��� ������� ��� � �� ���� ���� ����� �� � $����� � �� (��� �� ����� ����� � ��� � ��� ���� � ���� ��� d�e� (��� ���� ����� � ����� $����� � (���� $���� � ����� ����� �� ����� � ��� � ��� ���� � ���� �� ������ �� ������� ����� ���� ����� �� ��� ��� ��� ���� ������ ��� ��� $ ����� ����� � �'$����� ���� i�(c��� $ � ������+� � ��������� ����� � ������������ ��� ��� � $�������� � ������ ����� �� ��� �� � ���� ���(��� � ������$��� ����� ��� � $*� ������ ������ ��� �'$����� ����� � d�e�+� ����� � ��� � ��� ���� � �� :187�8 �� � � ��'�� � :����8 � �� ��'��� d�� �� �� $� � �(� $��e���� � $�� ����� 3 ��� �'$����� $�� ���� � d�e�+� ����� � ��� � ��� ���� ������� $�� ���� � � @ �4e5 �" +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< ���� �������� � ����� �� $� � ���� (��� :� � � 4j� �*���$��5 �.,# ����� ��� ��� ���� ��� � 1��"7 ��� � 4����� �.,# ����� (� �� :� � ��� � ���� ��� � 1� ��� � ��� :�1� ��� �� �� ��� 18�� ��� 5� ��� $�� ��� $��e���� � � ���� ���e�� �� ��� ;�#� ���� � ���� <��� ��� � �.,# ����� � @q� ���� ���� ���� � $ ����� ����� � ����� :187�8 �� � � ��'�� ��� :����8 � �� ��'��� ��� ����� �� $� � ���� ��� ��� � $� ���� ��������� �� $� �� �� ����� � ��� � ���� . d�e� ������ ��� $ � ��� ���� ��� ����� �� $� � ���� (��� �������� �� $�� :� � ��i����� ��>����� � ��� � ����� � ��� 1��"7 ��� �� .� ���� ���� d�e�+� ����� � ��� � ��� ���� � ������� #�$$��� d�e� ��� � ������� � � $ ������ � ���� (�� ����� �$����� �� �� �� ���� � ��� ������� +� 3������� ������ �� �� (��� �� ��� ����� ��� �&�� 5��= ����� � ����� � ��� � ��� ���� � �� ������ �� � ���� ����� #���� �� � ���� ���� ����3�� � � ��' �'� $�� �� ����� ��� � :����8 �� �� � $������ ���� � �������� ����� �� ��� ������$��� ����� � ����� � ��� � ���c���������� �$�� �� ������ �� ����� 1��"7 ��� �c���� � $*� ������ ������ �� ���� ��� ��� �� �� ��� ����� ��� ��j� � ��� ����� ����3�� ���� ��� �'��$���� � ���� � $� ���� ������� (� ������� ���� d�e� ������ ���( ����� � ��� � ��� �� � :����8 �&�� 5��= ����� !����"��� .��� � ����$��� � � ��(��� �'$ ����� �� � �����$�� � ?<,� ���� ��� ��� ������ ��������� ��� ����� �� �� �� �� ��� ��� ����� ������ $�� ���� �� ������� ���� � �� ���� �� ��� ��� ���� � <�6� *(��%(�� *��� ����( ��� ��� ��$� �� �� ������� ������ 1 ��� @ $ ����� �����$��� � ��� ��� ���� ��� )���� ,�� ��� l��� �� �� ����� �� ����� 1 $ ������ �����$��� � ������ ����� ������ � ����� (���(� � ��� (���(�� ��� �� ��� ������ ����� $ ������� ��� ���� ����� ����� � ������ ����� ��� �� � ����� �$$���� �� ������ � ����� &�� �� ���� ����� �� � �'$�������� ������ � ���� ���� �� �� �����$��� ���� ������ ���� �� ��� �� ���� �� ��� �(� �� � $����� � ����� 1� ��� �(� ���� $����� � ����� 1 $ ����� ��� ���� � �����$��� � � �� ������ ��� �$���� ��� � �� ������ � ��� �$����� (�� � ���� �������� $� ��� (�� ��� ����� � ��� ��� #�$$��� #�� ��� #�j����� ���� ��� @@� � � � ���� #�� ��� �� ���� )���� ,�� � l��� �� �� (��� g��� ,�� < ���� � :" � ��� � ��� � � �� ����� . ������� �� (���� ���� ������� :1 � � ��� � (���� (� ���� ?<,�� . ���� ����� ��� #� ���� � g���3� ,���� ���� ?<, �� "�7q ���� � :�8� 7 � r4 �7 5 4:" � 5 4�� � "75s� (���� (� ���� ?<,�� . #�� ���� �� � � #�j���� ������ ��� "�� ��� (��� ������ 77q � ?<,�� #�� (��� �������� �� ������ 77q � ?<,� ����� ��� ������ "�� ��� 17q � ?<,� ��� �� �� � �����$��� �� ��� �(� ���� $����� � ����� 1 � � �'$ ����� ������� �� ?<,�� (���� �� �� �� �� ��� �� ���� +� �� ��� ������ $�� ���� ����� @ $ ������ ���� � �����$��� ���� )���� ,�� ��� l��� �� �� � � ��� ��� ��� ��� ��� � ��� � ��� ��� ��� (�� ������ ��� #� ���� g���3� ,��� 4#g,5� ���� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< �� � ��� ��� ��� ������ ��� #g,� .� ���� ��� ��� �� �� ���� �������� �� ����� (��� ��� �$����� ���� ��� ������ ��� �$���� (����� ��� � �� ���� �� �� ��� ��� ��� � � � �(� ���� � �����$��� �� ����� @� p�� ��� ���� �� ��� ������� ��� (� � � � ��� �� � ���� ��� ���� ���� �� ��� (� � �� 1 ��� � ������ � #�$$��� d�e�� ��� @@� ��� �� (��� g��� ,�� < ���� � :@ � ��� ,� ������� � 7 q� l� ��� ��� (� �� (�� �� 1 ��� � ������ � �$� � ��� #g,� . ������ 4� � ��� ��� ��� �� ��� #g,5� ��� ?<,� (���� �� :� � � #���� �� ��� ��� (� � ����� ��� #g,� ��� ?<,� �� :�8�� � rg��� ,�� < ����� ,� �������� ��� � "7�s� (�� � "�7q �� ��� ���� � ��� ����� 1 �����$��� � ������ ���� ��� � ����� (���(� �� ��� (���(� ���� )���� ,�� � l��� �� �� ��� � ��� ����� <�� #����� ��� #����� � ��� ����(� ����( ��7q 1q @q ��7q 1q @q ��7q 1q @q ��7q 1q @q @@ �1�%1 ���� �%��% �7� @ �1��� �%�%7 � �@% �%��% �"�%% ����� �%�8 ���@� 7 ���@7 � � � ����� ���"8 ����1 �8�7 ���@� �"�@@ �@��� �8��� ����1 �7��8 7" �8�"8 ���"@ �7�81 � � @ �8�81 �"��" �1�%� �1��1 ��� 1 �@��% �@� ���"1 7% ����� �"�1@ �@��8 �8�"1 ���7% �7��8 ��� � ���@7 � �@7 ���%� ����7 ��� % "� �7��1 �@�%% �1�"8 ����% �"�1 �@��@ �7��1 �@�%% �1�"8 ����% �"�1 �@��@ "7 �@��@ �1�"1 ���7@ �7��@ �@�%% �1�"� �@��@ �1�"1 ���7@ �7��@ �@�%% �1�"� "8 ����� ����8 ���1% �@��% �1�"� ���7� ����� ����8 ���1% �@��% �1�"� ���7� �@ %�%� %�"� %� " ���@@ ��� @ � �1� %�%� %�"� %� " ���@@ ��� @ � �1� 8 ��1 ���@ "�81 8��� 8�71 8� % ��1 ���@ "�81 8��� 8�71 8� % ��� ��7q� 1q� ��� @q � � �.,# ������� ��� �����$��� ���� � ��� ����� � $� ������ ��� �� �� ��� @@� �����$��� � � �$$���� �� ��� �� ���������+� ?� ��� <����� ,�� ��� �� ���� ��� ��� $ ����� ����� � �'$����� ���� i�(�� ����� @ �����$��� � � ��� ���� ��� � ���� ����� � ��� ��� ���� )���� ,�� � l��� �� �� ��� � ��� ����� <�� �� ��� ��� �� ��� � ��� �� � �� � ��� �� � 1 ��� � ������ l������ �� � ��� l������ 1 ��� � ���� ��7q 1q @q ��7q 1q @q ��7q 1q @q ��7q 1q @q @@ �7��� �1��7 � � � �7�@� �1�@" � ��� �7�8@ �1�8� � �@� �7�"8 �1�"8 � �1� 7 ����8 ����1 �8�"� �1� � ���@� �8��8 �1�%% ����8 �%�1% �1�17 ����� �8�%� 7" � �� �8�%� �"�8@ � �@� �%��8 ��� @ � �%7 �%�"7 ���@� � ��@ �%�@� ����8 7% �8�"� ���"@ �7�87 �%� ���%1 �"� " �%�7" �8�@1 �"�@� �%�17 �8��@ �"�1� "� ����� �"�11 �@��% ���7@ �"�"� �7� � �8��� ����7 �7�@" �"� " �7�1 �1�%1 "7 �7��� �@�%% �1�"% �"� " �7��% �1�%1 �"�"" �7�8@ �@�@ �@�7% �1�%7 ���8� "8 �@��@ �1�"@ ���7" �@�78 �1�%7 ���8� �7��% �@�7� �1�1 �1��� ���" ���"� �� ����7 ����� ���@ �1��� ���7% ���"" �1��� �1��" ����" ���"" ����7 � �7� �@ ����% � �%� � ��� ���"7 ����@ � �@% ����7 ���8� ��� � ��7 %�%1 %�17 8 8�71 8�1� ��% 8�88 8�"@ 8��% %�@@ %��8 8�"% ��"7 ��@� ���1 ��� ��7q� 1q� ��� @q � � �.,# ������� ��� �����$��� ���� � ��� ����� � $� ������ ��� �� �� ��� @@� �����$��� � � �$$���� �� ��� �� ���������+� ?� ��� <����� ,�� ��� �� ���� ��� ��� $ ����� ����� � �'$����� ���� i�(�� �8 +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< #g,� . d�e� ���� 3 ��� ��� �� ������ (� � ������� 77q � ���6 ����� ��� ������ ��� "�� ��� 17q � ���6 ��� �� �� � ��� �����$��� �� ����� @ � � ����$��� � ?<,�� (���� �� �� �� �� ���� �� ��� ������ $�� ���� ��� �����$�� � � � ��� ��� ��� �� �� ����� @ �� �� �� ���� ��� �����$�� � � ������ ��� ��� �� �� ����� 1� ��� � ���* ��� �����$�� �i���� �'$����� ���� i�(� � ��� ��� �� ����� %(*� �'$����� ���� i�(� �� ��� (� � � �� ���� 3 ��� ��� ������ ��� �����$�� �i���� ���� �'$����� ���� i�( �� ��� ��� � ������ ��� �� � ����������$ �'���� ���(��� �����$��� � ������ � ���� ��� � ��� � ��� ��� ���� )��� �������� � �������� #g,� � ���� �� ������ ��� ���$� � ���� $�$� � ��� �� �����$��� � � �� � ������� ���� (����� �������� ��� #g, �e�� ����� �� ������ ������ ���� (���� ��� ��� ��� ��e� #g,� �� �� $������� �� ������� #g, (��� �$����� �������� #��� � �������� �� ��� �� ��� � ���� .� �� ���� $�������� (��� �$����� �������� �� ������ ���� ���� ��� ��� #g, ���� ��� (� �����j�� ��� �� �� ���� $������� �� $� ����� #�$$�� ����� #g,� . d�e� (����� #g,� ��� $������ (����� �� �%��%� �� � � � :187�8 � ����� �� �� @@*��� *��� ��� ��� �.,# ��� �� @q� . d�e� ������ #g,� d�e� ��� ��� (� � ���� $������ (����� � � � �� ��� � � ��� "�7�� � � ����� :1�7�1 � ��� $ ����� ����� � �'$����� ���� i�(� �� �������� ����� (������ #g,� ��� �� � ���������� ����� � � ��� ����� � � �� �� ��������� (��� ��� � ��� �� 1 ��� � ������ � ��� ��(� �.,# ������� <����� ��� �������� � ��� #g, �������� 4��������� ������ � ���� �������� ��e ���� ����� ���������� �� �� �������� ����5 �� ������ ��� ���$� � ���� $�$� � ��� ��� $ ����� ����� � �'$����� $������ ���� i�(� �� �������� ����� (������ #g,� <�8� *(��%(�� *��� 1��, ��� ����� ������ ������ � $� ������ (��� ����� �� ��� � ���� �-&;m ��� � ��� ����� �� ��� � � o��� �������� ��� ������ $ ����� �����$��� � � (��� ���� � ���� �� ���� ����� ��� � ��� ���� ?����������� � �����$��� ���� �-&;m � � ���� ����� � � �� $��'� ?������ �(� ��������� �(�� � ���� � (�� � � �� ����� "�*��� *���� ���� ���� �� ��� �� ���� ������� ��� ��� ��������� �� �� � �'��$� <��' ��� �� � �� � ��� � �� ��� @� ��� g� � ��� �� � �� �1 ��� � �� ��� @@� <��'+� ������ $�� ��� ��� ����� ���� ������ ������ ��i����� �� �q � ��� � � ��� �� (���� g� � ��� ������ ������ ��i����� � �� ��� �� <� ��� "�� <��' (��� ����� �$ � � � ��� � � �q ��i����� ��� � ����� .� ��������� <��'+� ,� ������� (��� ��� � @ q �� 7 q� (���� g� �+� ,� ������� (��� ��� � 7 �7q �� 7��7q� ����� 7 $ ������ �� �� � ���� ����( �� ���� �� ���� ��� ��� ���� � �����$�� ���� �-&;m ��� � ��� ����� � ������ ���� ��� � ����� ��� ����(��� �'� $�� ������ ���� ��( �� ��� ����� 7� #�$$��� g� � �� 7 *��� *��� ��� ��� �� �� ��� @@ (��� �1 ��� � � �� ����� l�� ?<, �� 7 �7q � g��� ,�� < ���� �� ��� 7 � #���� �� ��� �� " ��� � �� ��� � ��� ��� g��� ,�� < ���� ��� ������ ��� ?,. ��i����� ��� �� �q � ��� � ��� ���� " ��� �� <� ��� "�� ��� � � � �(� ��>��� ����� )� ��� ��� g��� ,�� < ���� ��� ����� �� ��� ����� �� (���� ���� ���� (������ ��� ?,. ?<,! ���� ��������� ��� ��� g��� ,�� < ���� �� ��� "� �=���� ��� g��� ,�� < ���� �� ��� 7 �� �� 4�� �5"� #������ ��� ,� ������� �� ��� ����� � 7 �7q �� 7��7q� ��� ,� ������� � �� ���� ��� �� � �� �1 ��� � � �� ���� ���� l��� �� ��� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< �% ?� $� �� �� ��� ��� ?<, �� ��� 7 � ��� ��� ?<, �� ��� "� (��� �� �� ��� ��q �� �� � r ��� ��"�7��7�7 �7�� � �s� ����� 7 �=�� �� �������� �'$��������� )� � @q �.,# ������ $ ����� ����� � g� �+� $�� ���� �� ���� ��� "� �� ��� $ ����� %��7� ?<,� ��� ?<, �� ��� 7 � ��� $ ����� ����� � $�� ���� � �� ��� "� �� ��� $ ����� ��88� ) ������� ?<,� ��� ) ������ �� 7��797 �7 � ���@� (�� � 7 �7 ��� 7��7 � � ��� ,� �������� � g��� ,�� < ���� ���� ��� ��� �� ������� �� � � ��� � �� ��� "�� ��� $ �*��' ����� � g� �+� ��� � ��� ���� � �� ����� �8��1� ?<, (�� � �8��1 �� %��7� ��88� ���@� ��� ��(� $���� � ����� � $ ������ ������� � ��� ������������ ���� � �� ��� �(�� � ���� �� <��' ��� g� �� <� ��� 7 � ��� $ �*��' $ ����� ����� � <��'+� ��� � ��� ���� � �� �%� ?<,<��'� ��� $ �*��' $ ����� ����� � g� � �� �8��1� ?<,g� �� <������� <��' ��� ��� �� �� �����$��� �� �� � ����$�� � ��� � ���� ����3�� #���� �� ��� �� ���� � ��� $ ����� ����� � g� �+� ����� � ��� � ��� (��� ��(��� �� � ���� ���� ��� $ ����� ����� � <��'+�� ����� " $ ������ �����$��� � � ��� ��� ��� � ��� � ��� ��� ��� (�� �$� � ��� #g,� ��$��������� #�$$��� d�e�� ��� @@� ��� �� � �� �@ ��� � � �� ���� (��� g��� ,�� < ���� � :@ � ���� �-&;m ��� ��� ,� ������� �� 7 q� l� ��� ��� (� �� (�� �� ����� 7 �����$��� � ������ ���� ��� � ���� ���� �-&;m ��� � ��� ����� <�� ����� )� ���� ��7q 1q @q ��7q 1q @q @@ g� � � "� �1�@ ���%� ��� 7 �1�@7 ���%� ��� % < �� "� 8�"� ��71 7�8� %�@" 8��@ "�1 7 g� � � "� %�81 %�7% %��7 %�8� %�"1 %��8 < �� "� � �� %�"7 ��88 ����1 � �71 8�71 7" g� � � "� 7�@7 7�18 7��" 7�@" 7�@ 7��8 < �� "� �1�@ ���@@ � ��" �@�"1 �1�7� ���"� 7% g� � � "� ��88 ��8� ��8@ ��8% ��8� ��87 < �� "� �7� 8 �@�� ���"7 �"�@ �7�18 �1�" "� �@��8 �1�"@ ���7� �7�@7 �@�� �1�18 "7 �1� @ ���7� ���7" �@��� �1�"1 ���7 "8 ���8 ���1" � �78 �1� � ���7@ ���78 �@ %�1� %� @ 8�77 � �"7 � ��% %�"7 8 "�%� "�8� "�7@ 8��% 8��� ��"� ��� ��7q� 1q� ��� @q � � �.,# ������� ��� �����$��� ���� � ��� ����� � $� ������ ��� �� �� ��� @@� �����$��� � � �$$���� �� ��� �� ���������+� ?� ��� <����� ,�� ��� �� ���� ��� ��� $ ����� ����� � �'$����� ���� i�(�� )� ���� ����( "�� ��� �����$�� �� �� $���� �� ��� � � "������$��� � ) ������� �� �� "������$��� (�� � ) ������ �� ��� ,� ������� � g��� ,�� < ���� ���� ��� ��� �� ������� � �� "� ������� �� ��� ,� ������� � g��� ,�� < ���� ���� ��� ��� �� ������� �� � � "�! ���� ���� ������ ��� �� ����� ���$*�$ ���� � � �-&;m� 1 +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< 1 ��� � ������ � �$� � ��� #g,� . ������� ��� ?<,� (���� �� :� � � #���� �� ��� ��� (� � ����� ��� #g,� ��� ?<,� �� :�8�� � "�7q ���� ��� �� ��� ���� � ��� #g,� . d�e� ���� �� � � �� ������ ��� "�� ��� �� ������ (� � ������� 77q � ���6� ���� ��� ��� ��(��� � ��� ��� (���� ���� ������ ��� "�� ��� ���*�� � ����� �$ ���� �� ����� ��� g��� ,�� < ���� ������� �$ � ��� ��� � � �q ��i����� ��� ��� ��� ,� ������� ��� ����� �� " q� ��� ����� ���� )���� ,�� ��� l��� �� ��� ���� ���� ��� �* #�#��$ �%�*��� ������ ��� "�� ��� ������ �� ������ �� 17q� .� ����� "� ��� �����$�� ����������� �� $� ����� � � ��� �� � ��� ��� �� ������� ���� ��� #g,� <�;� � �%� ���� �& *(��%(�� .��� � �����$��� �'����� ����*$ ����� $ �$� ����� ������� ������� � ��� � ��� ���� � �� �� �� ����������� ��e� ������� � ����� <������� (� ��� �����$��� �� ����� 1 �� �� ���� ��� ��� ����*$ ����� $ �$� ����� ���� � � $ ����� �� ��� �����$���� ��� �����$��� � � ��������� �� ��� ����� � �.,# ������ �� ���� � ���� ����������� �� �� �� � ������ ��� ��� ���� � ���� ��� ��� <� ��� @@� � ������ ���+� �����$�� �� �7��q �� �� (��� ��� �.,# ����� " �����$��� � � ��� ���� ��� � ���� ��� ��� ���� �-&;m ��� � ��� ����� <�� �� ��� ��� �� ��� � ��� �� � �� � ��� �� � 1 ��� � ������ l������ �� � ��� l������ 1 ��� � ���� ��7q 1q @q ��7q 1q @q ��7q 1q @q ��7q 1q @q @@ g� � � "� �1�71 �1� @ ����" �1�71 �1� 7 ����" �1�77 �1� � ����8 �1�7@ �1� 7 ����� < �� "� %�7� 8�1 "�1" %��� 8�@" "�@� � � 7 8��@ "�"� %�%1 8�"@ "�" 7 g� � � "� %�%� %�"8 %��1 %�%� %�"8 %��1 %�%@ %�� %��@ %�%� %�"8 %��1 < �� "� ���8 � �" 8�" ���1@ � ��� 8��@ ���@@ ����7 %� � ���78 � �18 8�@ 7" g� � � "� 7�@8 7�@� 7�1 7�@8 7�@� 7�1 7�@% 7�@1 7�1� 7�@8 7�@� 7�1 < �� "� �@�� �1�7% ���"% �1��% ����� � �71 �7�@" �@��� ����� �1�7� ���71 � ��@ 7% g� � � "� ��8% ��88 ��87 ��8% ��88 ��87 ��8% ��88 ��87 ��% ��8% ��8" < �� "� �"�@� �7�@7 �1�"8 �@�@@ �1�71 ���%7 ����% �"��% �@��� �@�1 �1�1% ���8� "� �7�@% �@��7 �1�@@ �7��@ �@�%� �1�"� �"��� �7�@� �1�%8 �@�77 �1�% ����1 "7 �@��% �1�"" ���7@ �@�77 �1�8% ����� �7� " �@�1" �1��� �1�1@ ����% ���8 "8 �1� 7 ���71 ���7% �1�11 ����8 ����% �1�87 �1��� ����� ���� ���"7 � �81 �@ � �7@ � �� %�78 � �8@ � �@� %�8� ���1� � �%8 � ��� %�"� %�11 8�8 8 8� % ��8% ��7� 8�@ 8��8 ���� 8�%� 8�"� 8��1 ���% ���� "�8� ��� ��7q� 1q� ��� @q � � �.,# k������ ��� �����$��� ���� � ��� ����� � $� ������ ��� �� �� ��� @@� �����$��� � � �$$���� �� ��� �� ���������+� ?� ��� <����� ,�� ��� �� ���� ��� ��� $ ����� ����� � �'$����� ���� i�(�� )� ���� ����( "�� ��� �����$�� �� �� $���� �� ��� � � "������$��� � ) ������� �� �� "������$��� (�� � ) ������ �� ��� ,� ������� � g��� ,�� < ���� ���� ��� ��� �� ������� � �� "� ������� �� ��� ,� ������� � g��� ,�� < ���� ���� ��� ��� �� ������� �� � � "�! ���� ���� ������ ��� �� ����� ���$*�$ ���� � � �-&;m� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< 1� ����� �� 1q ������� � @q� <� ��� �@� ��� �����$�� �� "��q �� �� � .� ��������� � � ����� ������ �� ����� ���� ����������� �� �� �� �� ��( ������ ���� �� ���� ������� )� �'� $��� ��� @@*��� *��� ������ ���+� �����$�� ��� ����� ��%� ��� � (��� ��� �.,# ����� ���� � @q �� 1q ��� � ��������� �� �� ���1 ��� � (��� ��� ����� ���� ��� �� ��� � 1q �� ��7q� )������� ��� ��j� � ��� �����$�� ��� ����� ��������� ���(��� ��$������� � ������ ��� ���� <� ��� @@ ��� � �.,# ����� � @q� � ������ ��� ��� � �����$�� � �%��% ��� � �� � �'$������� � ����� 1" ��� �� <� ��� "�� ��� ��� �� � �'$������� ���� � ��� �����$�� ��� ���� ��� ����� ����� 1 q �� �1�"8� .� �������� �� ��� ����*$ ����� $ �$� ���� ����� ������ �� � ��������� � ���� ��� �����$���� <� ������ ��� �� ���� ����� �� � �'$�������� ������ (� �� ���� ����� �����$��� ���� ������ ��� #� ��� ��� (���(� ���� ����� �����$��� ���� (���(� �� �� ��� ��� ��� ���� ����� �����$��� ���� �������� (���(�� ��� (���(� �� ���� ������ � >���� �� � �'$��������� �'������� ������ �� � �'$���������� ���� � ��� �� � $��3�� (��� ��� ��� ��+� �$���� �� ������ � # ����������� �� ��� �""����� �� ����������$ ��� ����� ��% /�� ��� ��� "@� �� � ����� � ��� ��� ,� � ��� "�� �� ��� (� �� /�� �� ��� �� � ��� $���* ����� � >�� 1 ��� � ��� ��� ,� ������� ��� ��� ���� ���� :1 � �� ����e� �� @ �4e5 $���� ��� :1 � �� ����� �� ��'���� ��������! ��� ���� ����� ��� � e�� ����� � ��� ����� � � :1 � � ���� �'$��� �� �� �� ��� �8q ��' � ��e�� �� ��� ��� � ���� ���� ?<, � ����� � ��� � ��� �� :1 � � 4)� �� $������� (� ���� � ���� /�� �� ��� ��� �� � � ������ ��3���*����3� $������ ���� (���� �=�� � ��������� ��� ����� ����� ���� ��� $� � �����5 ���� �� ��� ����� �' �� 7 q ����e� ��� 7 q ������ ���� �� ���� �� ��� ����� �'t ?������ �(� ���(� � ,� ��� /�� ��+� ����� �'� . ���� ��������� ��� ����� �' ����� �� ��� � ��������� �$$ ����� ���� ������� ���� ���� �������� ��� ���� �� ����� �'� ���� � ��������� $� � ���� �������� :1 � �� ����e� ��� :1 � �� ������ o�( ��� �� ��������� ���� ff� ��++ $� � ����� (���� �������� ���� ����� � ��� � ��� ��� ������������ �i���� � �� *��' ������� �� ��� 3 �� ����� � ��� ������ �� � �� *��' ������� ��� :1 � �� ��� @ �4e5 $���� �� $ �*��' ����� �� .� �$ ������ :��"� � �� ��'��� :1 � 4�� ��85� ��� :1 � � ����� �� ��� ��'���� ������� �� � �� *��' ����� �� ��� ���� ��� ���� �����$��� �����+� �.,# ����� �� @q� ) � ����� @� ��� �����$�� � � "7*��� * ��� ��� ��� �� (��� � (� � 1 ��� � ������ �� �1�%1� /�� �� �� �*��� ������ ���� "7 ��� ,� �� � 4��� 15 ��� � ������ ���� /�� ��� #� (� ���� �� ���� $����� � ��� ������� ���� �����$�� �� �@��� � �1�%1� �1"� � 8� ��� �1" �� ���*��� � ��� �� � ���� ���(��� ��� �����$��� � ���� ��� "7 ��� ��� "� (��� � (� � 1 ��� � ������ � �91 4�7� �� �1�%15� ��� � 8 �� ���*��� � ��� �� � ���� ���(��� ��� �����$��� � ��� ��� "7 (��� � (� � ��� �� � ��� ��� ��� ��� "7 (��� � (� � 1 ��� � ������ � �91 4�1�%1� �1�"%5� ��� $ ����� ����� � ���� � �� *��' ����� � ��� � ��� ���� � �� ����� :1 "�% � � :1 � 4�@���5 4�� ��85� g���� �� � �� *��' ���� ��� ��������� ��� ����� � ����� � ��� � ���� ���� $� � ���� �������� :8���% �� ������� ���� � �� ����� �' �������� � �"q ����e� 4:��"� 9 1� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< :8���% 5 ��� �@q ������ (���� �� ����� ��� ��� ����e $� ���� ���� ���� � 4��� ����e ���� ����5� �� ������� ��� ���� �� 7 q ����e �'$��� �� ���� ����� ������ ��� :��"� � �� ��'�� � @ �4e5 ���� ��� :�%7�@7 � ��� ��'���� ������� �� ����e�� <��� ��������� ���� ����� ������ ��� :1 � � ��� ��'���� ������� �� ����e� ��� :����@7 � � �� *��' @ �4e5 ���� 4� :�7@�8 � $ �*��' @ �4e5 ����5 �� ����e�� ��� ������ ���(��� ����� �(� ���� ������ $� � ����� ��$���� �$�� ��� ������� +� ����� �������� ������� ���� ��� �� ��� ��� �� $�������� ������ ����e� � ����� �� ������� �� $�������t #����� 4�%%%5 ��� #����� ��� #��� 4�%%85 � ��� ���� � � ��� ���� ������ ����e ������� � ������ ������ ����e� �� $�������� ��� ������������ 4� �5 � ��� ���� � ���� ��� �$$����� ����������� ���� ��$�� �� ������ ��� ���$� � ���� $�$� � �� ����� � ,� ��� /�� ��+� � ��������� ����� �' �� ���� � �� ����� �'� (� ���� �� ���� ��� � ���� � ����� ��' ��� �� ��� ��� � ���� .� ��������� ��� �����$�� �i���� �� � ��� ���� $������ ��������� ��� ���� $���� ���� ��� ����� ��� ������ �i��� ���� ���=�� �� � �'$�������� <������� ,� ��� /�� ��+� � �� ����� �' ��� ���� �� ���� ����� (� ���� ���� �� �� ����� �� ��� �� �� ���� ��� � ���� � �� $� � ���� ���� �� ��� �� ���� � ��������� $� � ����� g� ���� ��� ��'��� ��� � ��������� $� � ���� � $������� ���� �� ���� � ����� ��' ��� (��� �� j� �� .��� ��� ����� � ��� � ��� � $���� ��� ����� �� j� �! �� ���� ���� ��� �'� $�� �� ���� ���� �� �� � $� ���� �� $ �$� �� ���� � � � ���+� ����� �'� & '��������� �� � ��� ���� �� �� ������������ �� ������� $������ ���� � �� �� � �� ��� � ��� $�������� ��� ���� � $������ (����� �� ��� � ��� ����� ����������� #����� 4�%%�5 ������ ff��� $ ����� ����� � � �$����+� $������ ��� ������ �� ��� ��� �������� ����� ���� � ���$�� $���������++ ��� � � �� $������ (����� (��� � ��� $� � ���� ��������� �� � ��� ��� ����� $�� ��� $���� � � $������� �e�� ��� ����*��e�� .� $� ������ � (� � ��� ���� ��� ��i�����* ��>������ ����� $�� ���� � ����� � $������� �e�� ��� ��e� � ���� � .�i����� , �������� #��� ������ ��� ����� ����� � $������� ���� ������� <� ��� ������ ����� �� #������ " �� ���� ����� � $� � ���� ������ (��� � �� ��� ����� ���������� ��� �� � ����� ��� � ����� ��� � $������ (����� �� ���� ��� /���� ��� ������� � $� ����� � ����� ���������� ��� ���� ��� ��������� ��� � e�� �� ��� 4g ������ l���� h g����(� � �%8"5� ��� � �� � � ������ ���� � ���� ��� �� ���� ����� ��� $� � ����� �� ����� �� ��� (� e � #���� 4�%%75 ��� ���� � �� ������������ ��� � $� ����� � ��������� $������ (������ /���� ��� ����������� ����*��e� ����� ���������� (��� $������ (������ (� ��� ��� ������� ��� ����� � $������ (����� �� � ���� �� � � ���+� ff� ��++ $� � ����� ��� � � � ���� �� � $��������� �� ����� ��� g������ $������ (����� ��� ���� � � ��� ����� ����e*����� $� � ���� ���� � ����*����� ���� �'������� $������ (����� ��� ���� � �����3���� � $��� �� ������ ��� ���������� ����e��� � $������ (����� �� � ���� � ����� � �� ���� � ��������� � � ����e �� ��� 3������� $� � � ��� ff� ��++ $� � ����� �� 3�� �� ����������� �$$������ �� ����e � $������ (����� �� � ����! ���� ��������� ���� $ ��� $� ������� � �� � ������� � � ��� � ��� ����� �=���� �� �� ����� )������� �'������� ����*��e� $������ (����� � ����� �' ��������� ������ �� ���*�$�� �� $� � ������ .��� ��� $������ (����� ���� ���� ��� ���� �� 4��� $ ��� ���� �$�� ��5 ��e9 ��� � +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< 11 � ���� �� � $�� ����� ���� ��� � ��������� ���( � ��� $� � ���� �� ����e��� �� �� �������� �� ��� ff� ��++ � ��� $� � ����� ���� ���� ������� � (�� ���� � $������ (����� � � ������ �$ ��� �*������� ��� � ��e* �������� �$$� ��������� ���� $�$� �e�� � ���� ������� �� ���� �� ������� )� ��� �� �$ ������ ��� ����� ����� ��� ���� ������� � �� �� �� ���� � ���� ��� �c��� #���� 4�%%75 ������� ����� �'$����� ���(� � � ���+� ff� ��++ $� � ����� �� ������� $������ (������ ��� ������������ 4�%%8� � 5 ������� ����� � �� *��' ��������� � $� � ���� ������� ��� ��� ) ��� �� ��� 4� �5 ������� ���� �.,# ������ � � ��� �� �� ������� ��� ����*���e�� ��i�����*����'�� ���� i�(�� #������ �� �� ���� ���� ��� � ���=��� ��� ��� � $ ����� ����� � �'$����� ���� i�(� � $������ ��������� �� ���� � ��� � � ��$����� ��� ���� �������� �� � �� � $ ����� ����� � ���� i�(� �� ���� �� � �'$�������� ��� �� �� �$$���� ��� ��� �������� �� � ��j���� ��� � $� ���� $������ ����� c���� � ��� ;�#� ����� �� ��(����� $����� ��� ����� � ����e ?�� �� ?���������� /���� &����� #���� ) ��� � #���� / ���� 0� �� d���� 4��� ����� 5� &��� 0���� &����� #(������ $� ����$���� �� ��� � � <���� � � )�������� #� ����� ��� � ���� �� �� ����� ��� �� �� ����� ��� � � ��� � ���$ �� ���� ��������� ���������� g ������ /� ,�� l���� d� ��� h g����(� � /� d� 4�%8"5� &��� ������ � $� � ���� $� � ����� ��������( ���(���� 0�* ��(� ����9<������ 1%6@@� &)<# 4&� ���� )������ ��� <��������� #� ����5 (������ 4� �5� ��� �� ��� ���*����� %��� <��������� (((�� ��� ��9 ����9 ��� ��9� <������� &��� �� �1� � �� ) ��� � #� ,�� ��������� �� ��� h 0���� &� �� 4� �5� #� ������ ����� ���������� � ���������� ������� �� ��� � $��� � ��� $ ����� ����� � ������ ���� ��� ����3��� ��������( �� #���� �#��!� /4@5� �%761�"� 0�����e���� <� g�� #�� *��?�� �� ��� h #������ <� -� 4�%%�5� )� ��� 3������ �� ��� ;�#�� ������ �������� � ��� #� ��� � ?���� � )�������� ���� �( ��� #� )*((����� ����� �� �6�@� $��� ������ �� (������ 4� �5� ���8���� <��������� $��� ������ ��� <������� �� �� �� � �� ,����� �� ��� ...� h #�� � �� 0� )� 4�%% 5� ����� �� %� �&�(�� ����$������ <���������� � .����� ��� ������ ��� ��� ����� ��� ?�� ������������ 2<� ,��� ��� �� ��� 2����� #� )�� h ����� &� <� 4�%%@5� �� ����� ��� � ��� ������ ��� ��� ��� (� �� � ���� �� < � ������ ��� ���� 1������� �#��!� ��4�5� �8 6�87� ������������� �� 4�%%85� ?���������� � � ���+� ����� �'� ��������( �� #���� �#��!� -415� �%76� "� ������������� �� 4� 5� ?���������� ��� ����� �'� 0�* ��( �& +��(�� ���$����� 4$ �������� 0�* ��( �& � �#��� �� �&�(�� ���$�����5� ;4�5� � 6�7� ������������� �� 4� �5� <���� ���������� ��� ����� �������� ��������� ���������� 0�* ��( �& +��(�� ���$����� 4$ �������� 0�* ��( �& � �#��� �� �&�(�� ���$�����5� �4�5� �"6�"� #����� �� ?� 4�%%75� &�3���� ��� ������ ��� $� � ����� ���� ������ ��� �������t ��� ���� ����������� �& "���#��*�( "�#���� � 7��""9 0�* ��(� 6-4� 5� �76��� #������ �� g� 4�%%%5� ��� �������� ��� ���������� � ������ �� $������ ��� ������������ ������� ��������� d������( )* ��* �& 1������� ���� ��� �� e��� ,�$� ��� � � 4�� ��5� #������ �� g�� h #��� � ?� 4�%%85� d��� �� ����� ���������� � ��� � ��� �������� 0�* ��( �& � �#��� �� �&�(�� ���$������ 84�5� �16�"� 1@ +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< http://www.dfas.mil/money/retired/ http://pay2000.dtic.mil #������ � <���� ��� (������ 4� 5� ����� @67 ���� �,*� ����� � ?� ����� l������ ��� ����� @6" )� ��� �,*� ����� � ?� ����� l������� <��������� (((������ �9 ���� ��� <������� ��� �� � � #������� #� 4� 5� ��$ � ��� � ���*$������� $�� ����� ,� � �� � ���$��� � �� �� 1 � #����� � �� <� 4�%%�5� -��� ����� ��� $ ����� ����� � $�������� ��� �� � ��� ���� ��� � ��� �� ���� ������� 0�* ��( �& 2�$�( 1��������� 8� @%6"�� ��� ������ ��� h )� ��� � �� 4�%%@5� ��� $������ $������ ���������� (��� �������� � $���� �� ��������( ��*���(��$ ��� �(�����$� <� � �6��"� k��� k�� l����� #�� h ��������� ?� ,� 4�%%85� ���� ��� $���� ����� $ �>������� � ��� � ��� ���=����� ��������( �� #���� �#��!� -415� +�+� 0�����$� +� ����������� � ��������( �� #���� �#��! 6� 78��69 6/:;< 17 http://www.soa.org/research the value of retirement income streams: the value of military retirement introduction review of the literature before-tax versus after-tax funds what "counts" in the family portfolio? u.s. military retirement systems issues, illustration, and critique of our estimates issues illustration critique of our estimates multiples multiples under final pay and high three systems multiples under redux retirement system properties of multiples illustration of new approach to calculating the asset mix conclusion acknowledgements references pii: s1057-0810(01)00078-6 �� �������� �� � � ��� ������� � � �� � � � ��� � ������ �� ����� ��� � �� �� � �� � �� �� ���� ��� ������� �� � �� �������� �� � � ���� �� � � �������� ������� ����� � ��������� ���� ��� ��� ��� � � � ���� �� ��� ������� � � ����� � ��� ���� ��� ��� ��� �� ����� � ��� !""#$ �� ����� �� ������� � �% !& ������ !""#$ �'��� ( ��'��%��� !""# �������� )������� �� �� �� ������ ���� � ����� ��� � �� �� � � �� � � � ��� �� *�� � �� ��� ��� � � � ��� ��'��� � ��� ���� � � �����+ ������� � �� �� ���,�� ��'�� � � �� % ����� ��������� � %��� � �� �� �� �� ������ �� � ��� � �� �� � �� ��� � � �� ��������� �� �� ��� ���� ����� ��'�� %'��-���� ��� ��� � ���'���� � �� ������� �� % ������� � �� �� � � � '� ����� � % ��� �� � � ����� � �� ������%����� ������%���� �� %� � -� �� ��� �� � -� ./��� ���� �0 �� � � � � ���� ��� � ������ �� � � �� �� % �� �������� �� � �� �� � �� ��� � ��� ���� 1�% � ��� � � ���� ���� ���� ����� %'��-���� ��� �� ��� �� '� ����� � % ��� ��� ��� � � � � ��� ���'���� �� 2�� ������� ������� �� � ����� � � ������� � �� '� ����� � % ��� % ���� �� ��� ����� '%��� � ��� ���� ���������� � %� � ���%� � � �� � � � ��� �� � !""# �������� � ��� � )� � ��� ������ ��������� �������� 3�� � � � ��� �$ ����� ��� � �� ��$ ���� � ���'���� � � ���� ������ 2�� � ������������ � � '� ����� �������� �� ���� ��'�� ������� ��� � �� � � � ��� �� 1 ������ %� � ��% ��� � �� �� ���� � �� ��� ��� � �� � �������� �� �� ���� �� % ����� ��� � �� ��� � ��� ��%� ���� � �� �� � � 4������ 5#6678� ��� � ���� �� ����� '� ��� �� �� ��,��� ��� � �� �� ��� ��� �� �� �� �� � � � ��� �� �� �����9������� �� � � � % � ,��� ������� �� � � � �� �� : � � �� 5#6668 � �� �� �� ��� � � �;�� �� � � � ���� ������ 1 ������ ����� ��� �� � ��� % � ,��� ������� ���� � � ���� �9� ����� %� ����� �� � � ������� � � ���� ������������� � ��%� ���� � � � ��� � � %� � � ��� ���� ��;� *� �� �� 4 ����� 5!"""8� � � � ���� � ������ �� � � ��� ��� �� ����' ������� ��� ��� �� ��� '�� �� � � �� � � � '� �� ��<����� ��� =���� ��%� �� ��;� � >� ' �� ?��; � � � %� 3�� � � � ����� �� ������ #" 5!""#8 (@#a #"(7�"a#"<"#<b @ ��� �� �� % ���� � !""# �������� � ��� � )� � ��� ������ ��������� c))d � # " ( 7 � " a # " 5 " # 8 " " " 7 a � e � � ����' ����� ��� �� 2���d �#��"��e(#�#7##$ � -d �#��"��e(#��!#6� !"���� ������� � ���� f������� 5���� �� ���� 8� �� g �� 5!"""8 � �� �%���� ������� �� ; % �;�� '��� �% � � � ��� ��%� ��� ��� � ��� ..�-'����� �00 �� � ����� ��� %��� � �� � ��� ��%0� ����� ����� �� � � � % � ,��� ��� ��� �� � � �; �������� ������� �� ������������ ���� �� � � ���� � ������ � =� ��� ���� � ��� ��� ��%� ���� � � � ��� '�� ��� �%' �� �� ��� � ��� ��� �� � � �������� ������ ���� �� � �� �� ��� ��� � ������� � � ��� �� �� ���������� ���%'�� � �� �� � ��� ������� � � ��� ��� � � �� ��� � ��� ��� �� ��� �� �� � �� �� �� � �� ��� ���� � �� � ������ �� ���� ���� ��� c��� '� �� � ���%'�� � �� �� � � ���� ��� � �� ;� ������� ����� � � ���� � ���� � � ���� ��� � ' ���� 2��� ' '�� ���� � ' �� ��� ����� ���� �� �������� ���� ��� ������� � � �� � � � ��� � '� ����� �� ����� ��� � �� ��� : �� � �������� � �� ������ %����'�� ��������� ���� � ��� ��� � %� 5 � ��%�� �8 �� � � � ��� � �� % � �+ 1 � � �� ��� ��� � %' �� � .����0 ��� �+ h������ ����� ����� ������ �� �%' �� �� � � �� �������� � ���%��� �� � ��� �� � � � ����� �� ��������� � ��� � �� � ����%��� � ������ ' ��� � � �� ��� ������� '��� ��, �� � � �� � � � �� ��� �� �� ��� ���� � � %'��������� '��� � � �� � � � '� ������ i ���� � � ��� �������� � '� ����� ��� � � %����� % �;�� ���%��� ����� ��� ��� ��� � �� ��% � �� ��� %' ���� 5i ���� !"""$ / �� � #6668� ���� ��� � %� ���%��� % � ������ ����� ��� ��� � � �� � � � �� ������ i ���� 5!"""8 ���� �� � ���%��� ���� � '� �� � �� �� �� � �� ���� �� �� �� ����� �������� � � � � ��j� ���� ��� ����� � % �� �� � � � �������� ��������� * ���� 5!"""8 ����� �� � ����� ��� � �� ���� � � � �� � � ���%��� ���� ������ � ..'� ����� �� � '�����00 ��� � ��� ��� �� '� ���� � �� �� ���� � �� �� ��� �������� � �� � � � ��� �� % �� ����� � �� ������� � � �� ��� ��� � � �� ��� ��� ��� �� � %����%� 2 �� � ��� �� �%��� � � �� � � � ��� � � ���� �� ���%��� 3 � �%'��� � � ��� � ��� ���%��� %���� ��� % �� �������� ��� � ����� ��� ��� �� � ���� ��� � ��� ���%��� � � �� �� ���� ������� ��� � �� �� ���� % �� �������� ��� � � � %� ��'�� � 9����� ��� 2�� ��� %���� ���� ���� �������� ��� � � ��� % ��� � ���� ����� � � �� � � �������� � ������ � � �%'��� =� ��� ���� � ��� ��� ��� � ��� � �� ������ � �� ����� �� � �������� � � ���� � �� �� � '��% � ���� �� �� � � � ��� ������%��� ' ��� �� � 2�� ���� � ���� ' '�� �� �� ��,�� � � �� ��d �� �� � ! ���� �� �� ��� � � �� %��� ��� �� �� � & '� ����� ��� ������� �� ��� ���� �� �� �� � � ����� ��%% �� �� � ���� ��� � ���� #�$� %��� � � & � ���� ���'� ���� �� ����� ���� ��� ����� '%��� � ����� �� � � � �� '��� � � '� ����� 2�� � � �� � ���� � �� �� � �� � ��'������ ������ ���� ��������� ������ � �� ��� �� �� � � � %'��-���� i������ � ��� ' ����� � � � �� � ��� ���� ���� �� ���� �� �� � �� �� � � ����� �� ��� � �� � � � � ����'������ ��'�� � �� �� �� �� �� ���� �� ����� � : �� � c���� �� �� � � �� ���� �� � ��� ����� � ��� ��� �� ����� ��-�� ;� %���� � � �((������� �� !�� ����� � 5: �� � k : �� �� #6668� � � � ��� ��� � � ''� -�% ���� ��� ' ��� � ��� ���� � �� �� �� � � � �-������� ��%% �� ��� �% �� � � �� ��� ��� ''� �� ��� �� e %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 #�$�$� ����� ���� 2�� *� ��0� �� � �� &7 �� �� ��� i�;� �� � �; �� � '�������� �� � ��� b7"�""" ��� ���� � � � �� �������%� � %�% ;��� 2�� *� ��0� � �� �� �������� ��� �� �� ������ � � � �� �-'� ���� ����� ����� ����� 2�� *� ��0� �� � %� ����� 5���� % ��� ��8 �� � �� ������ ��� � ��� � ''� -�% ���� b!""�"""� #�$�#� ����� ���� 2 % �%��� �� �7 �� �� �� �-� ����� ���� � ��������� �' � �� �� 2 %0� � � �� �� b#""�""" '�� �� �� ��� �%��� �� (" �� �� ��$ ��� �������� � �;� � � �%������ ���� ����� �� � ����� b!��""" ��� ���� ��� � � �� ������� �� % ��� ���� % ��� ��� 2��� �� ��� #e �� #!� 2 % �� ��� � �� �� ��� �� �� � ������� 2�� �%���0� �� � %� ����� 5���� % ��� ��8 �� � �� ������ ��� � ��� � ''� -�% ���� b(("�"""� #�$�2� )��� ��� ���� � �� � �� g���� ��%���� � �� � �� �� e(� � �� � �� ' ����� �� �� ������� ��%� � �� � '� �� � ���� ��� �������� �� ��� ��% �� ��� �� � ������ �� � �; � ����� ��� g���� �� � � ��� ���� � �������� 2�� �'�� � � ��� �������� � ����� �� �� �� % #a � &6� �� #� �� �� �������� 2�� ��%���� �0� � �� '��% �� ������� � �� �� � �� � � %�� 5 �� ���� % ��� ���8� 2���� ��� � ��� �� ''� -�% ���� b#(�"""�"""� �� ���� ����� ' ���� � � � 9����� �� �� � ��� ���� � ��� � �� � � � ��� �� 2�� �� � � ����� ����' �� � ��� ;� ������ %'����� ����� %'��������� '��� � � �� � � � '� ������ 2�� 9����� �� �� � � �� ��d )������� � 1 � %� � ���������� � ����� 5��%' ��%8 �� � �� �� ���� � � �'�� ���� � ������%��� � ���� ����� ������%��� � ��+ 3��� � �'�� �� 1 � %� � ���������� ���� ����� � � '� �� �� � 5��%' ��%8 �� � �� �� � �� ��� ������ � �� � ���� �� � � � %��� ����+ ����4����� � �'�� �� 1 � %� � ���������� � ������% ��� ������ �� %� ����� � � '� �� �� � �� � �� �� 5'�� % ���8 � �� ��� ������ � �� � ���� ����+ � ������ �� 5����� ��������� 8 �� � �� ��� � � �'��0� ���� �� �� � �� � � ����� ������%��� ' ��� �� + ������ �� � � ���� � � ���� �� �� ��� �� ��%�� ��� '�� �� � ��� % > � ���� � ���� lh��� �� ;�� ������ �� � � �� ;�� � ���� �� ��m� � ���� 5�6 1 � ��� � �� ��� ��� ���� �� �� �� 5�-'���� ��� �� �� �� � ��� ���� ��%���8 �� � �� �� �� �� �� �� ��� ������ � �� � ���� ������ ��� 2� ���� � � )�� � � �� )�� � �+ !����� 5�6 1 � %� � � � �� ���� ��� � ��� �� ��� �� ��� ������ � �� � ���� ���� � � �+ #�#� *�4 & � ���� ������ ��'���� �� ��-� �� �� � %'�� � ��������� � ������ � � � ��� 9����� ��� =�� � � �� � ��%�� �� '� ��� �� � '��� � � ��� ��� � ��� ����� ��� �� � � � ��� �� �� ����� �� ���� ���� � ����'������ �� � � ������ �� �� � �� �� � �������� � �� %� �� � �� �� � �������� � �� % � ��� � ��� ��� ������� ��� ��� � �� � �� �� �� �� � � ������� �� %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 7 ��� �� ��� '���� � �� �� � � ������ ���� 2�� ���� ����� ��� � ' !� ��� �� � � ������� � ��� � ��,�� ��'�� �� �� �� ��'��� ������ � �������� �% �� ��� ������� ������ ����� �� � �� � ����� �� ���� � ����� �� ����� ��� ��� '� �� %� �� � �� ��� �� � �' � %���� �� � ����� � �� � � ����� ��� �% �� �� �� � � �� � � ������� �� �� �%'�� �� ����� ����� %� :��� � ���� �� ��% �� %'���� ����� � ����-�� � ��� ���� ����� �� ������ ����� � ��� ��� ��'�� � '� �� �� �� � �� �� �������� ��� 2��� ��� �� � %� �� � �;��� � ��� �� � � ������� � ����'������ ��������� � ��� ������� � % �� ��'��� �� �%'�� �� n � � %� ��� � ����� �� � % �� ������ �� �� ���� % �� �� � � �������� i�� � ����� � �� ���� �� ���� 2���� �� � �� ��� �� � � ������� � �� � � �������� 2��� �� � ���� � �� � � ��� ������� � ����� � �� � � ������� ��%��� �� ���� �� % �� ��������� h���� ��� ���� � ��� !� ���� �� � � ������� � ������ �� ������� #& ��������� i�� � ����� �� %%����� �� �� � �������� � �� % � �����%��� ��� � ���� ��� � ����� ��%��� � ��� � ���� �� � � �������� �� ���� ��� ������ �� � �� %� ��� � �����,�� !& �� � ��� � ' !� �� � � �������� �� ���� ��� � �� �� � � �� �� ���� '�� �� � � �� �� �' �� � ����� ��� ��� ����� ��� �� � � � '� �� ��� �� � � ������� ����� ��� �� � � � � �������� � � ��� � '�� ���� �� � � %'���� ������ ��� 3 � �%'��� ��� ������%��� 9����� � �� � � '�� �� � � ..������%��� c� ����� o���� � � �� � �00 � '' ��� � ..1 � %� � � ) ���� � %��� %� ������%��� � ��+00 ����� ��'������ ��� �� � � '�� ��� �� �� �� ���� � ����� � � � � 9����� �� �� ���� ����� ��� �� ������ �� ������ ��� ���� �� ��� ����� � �� � � ��� � � �-'���� ���;� �''����� �� �� ��� ����� ��������� � � � � 9����� �� � %� � ��� ����� �� ��� ����� �� ���� ���� ;� �� �� � %� ���� �� ���, ��� ����� �� � � �� ������ 2��� � � ��� � ��� ��� � � ����'������ �� � � %��� � � ��� �� ��� � � �� � % �� � ��� % �� ' '�� � ����� �� �� %��� � � �� � � � '� ����� ��� �% �� �� ��� ���� � � � �� �� � �� ��< � � �� � �� �� � ��� �� ���� �� �� ������ i� � � ��� ��� �% �� � '� ����� � ��� ��� ����� ������ �� � � '��� � ��� � �� � � '� �� � ��������� ������ h' � ������� �������� �� � � ��� ����� ���%������� �� �� �� �� ���� ���� �� ������ �� � � �� � � ������ ��� � �% � 9����� � �� �;��� 2�� � �%�� ��� ���� ��� � %� ���������� � �� � � ' ���� �� ������ )� ���� ��� �� ���%�� ��� ��� �� �� �� �� �����9���� ����� �� % ��� � ������ 2�� � ���� ���� �� �� � � �������� ��� 9����� � �� �;��� � � % �� %% � 5����� ���� � ��8 �� ��� ������%��� 9����� �� � %� ����� %'���� ��� ��� � ����� � � � ����� � � ������%���� � '' ��� � ��� ����� � � ��%' ��% �� ��� �� � ������%���� )� ��� ���� ��� � � ���� ��� ������ ��� 9����� � �� �;� �� �� ''�� �� �� % ��� � ������ #�2� 3��� ������� ���� 3 � ��� ..���� ���� �00 ' ��� � � ��� ������ � %'�� � �� �� �� �������� �� '��� � � �� � � � '� ����� � � ��� �� �� ���������� �� ��� � ���� ����� h����� �� ��� ���� �;�� � ���' �� � ��� � %� 9����� �� � ����� '�� � �� ���� � ���� ���� ����� ��� ������ � ����� � � � 2�� �������� �� �� %�%���� � ��� '�� �� 5�� �8 � �� �� �3c2i * �� ���������� i ����� � � ��� � c� �� %� �������� � ���� ����� � '��' � �� �� a %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 http://webcrawler.com http://yahoo.com http://searchenginewatch.com http://webinfosearch.com � ������� 2�$� *�4 ������ 2 ��� # '� ����� ��%% �� � ��� ��� ��� � �� � %��� � �� �� � �� ���� �� ��� ����� �� ����� � ����� �� ��� 2�� � ��� � ��� �� ������ � ��� ���� ����� %���� ��� � � ������ �� 9���� � ��� �� �� ����� � %��� � �� �� �� 3 � ��� *� �� ������%���� ��� %� � ����% �� � ����� � � ����� ������ 5 � ������%���8 �� � �� b#��i� ����� ��� � ��� �� �� ��� b!�ai� �� %%����� �-' ���� � h��� �9������ � � ��� *� �� � %��� ������%��� ' ��� �� ��� ��� � �� (7�ep� ��� ��� � ��� �� e"p� ���� � ���'���� � ����� � �� �� �� � �� % �� �� % ��� 2 ��� # ��%% �� � ��� ������� �� � %��� � �� �� ��� 578 i� � � ��� 5������ �8 �� �� �� ���� �� � �� *� �� ������%��� b 5!&8 #�""�e !767�� 77"�" /��� ����� � � b 5��8 #"aa�� !a6(�6 �a"�! :�� ������ ����� � � b 568 !�& &�& #�" )�����%����d ���� �� �� � p 5##8 h� �9���� (7�e e"�" #a�7 )����� �� � � �9���� #"�7 !&�# 6�& * ��� !&�" �"�" #"�� � �� a�7 !#�" 7�6 � �� )�� �� �� �� b 5#68 (!�6 e(�! #7�7 ��� �� � b 5!#8 e�� #&��( !6�� *� �%��� ������%��� b 5!&8 #a(!�6 �66e�! #"7&�" /��� ����� � � b 5��8 e�e�" ##a(�& &!"�� :�� ������ ����� � � b 568 "�7 (�a #�6 )�����%����d ���� �� �� � p 5##8 h� �9���� e"�� ("�" #e�( )����� �� � � �9���� #!�" !6�" #"�e * ��� #6�# !&�e 7�& � �� a�( &"�" #"�# � �� )�� �� �� �� b 5#68 ##�" #6�# (�" ��� �� � b 5!#8 #7�7 #e!�a �(�# �� ��%���� � )�����%����d ���� �� �� � p 5##8 h� �9���� &!�e e"�" #6�# )����� �� � � �9���� 7�� #a�" e�a * ��� &&�! ((�" #7�6 � �� !e�a #""�" &"�7 ��� �� � b 5!#8 &6a!�& (&a6�7 #a"7�! 2��� � ��� �� �� ��� ������� �� % ��� � �� �� � � ����������� '��� � � �� � � � '� ����� 9����� ��� 2�� ' ���� ��'������ ��� *� ��� �%���� �� ��%���� � � �� �� �� ���'� ������� : �� � ���' ���� �� �� �� �� ���� ������%��� �� �� �� �� �� '�� ��� ���� 7 �� ��� ��%��� � ��������� � � � � ' ��� �� � 9����� �� %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 6 *� �� � ��� ���� � %��� � �� �� �� 3 � �%'��� ��� � ��� � ��� �%��� ������%��� 9����� � �� �� ��� b(i� 2 ��� ! ���'� �� ��� ��� ���� � � �� �� � 5�?8 � � � � � ��� �� � � � '� ����� 9����� ��� 2�� �? ��j ��� ��� �� �� �� ���� �� � �� ��� %� � �� ���� � �� � %' ��� � � � �� �� � ���� ���' ���� � �� ��������� %� � � ����� *�� � � � �? � ��� � ��� �� ��� �� �� ���,�� �?� �� %'��� ��� �� �� ���,�� �? �� �������� �� ��� �% ����� �? � � �� � 9����� �� �� �� ���,�� �? ���� � � %���%�% � ��� � ��� 2�� ������� ����� ��� �� � �� � ��� �% ����� �� �� ���,�� �? �� �� ��� *� �� ��� *� �� �� ��� �� ����� ��� �'�� � � ��� � ���� ��� � ��� �� ��%'���� �� � � � � �� �� � 3 � ����� 9����� �� % ���� �� % *� �� � �%��� � ��%���� � ��� ��� ������� �� ������ �?�� i �� �� �� ���,�� �?� �� � ����� #"@!"p ������ � �� ���� )� ��� �� � ���� ����� � �� � � �%'��� ��� �%��� �? �� #�p ������ �� � ��� *� �� �?� 2��� ������ ��'' ��� ��� ������ � �� � ��� ��� � ����� ��%'��� ���� �� �� ������ �� � %'�� ���� ���� � �� % �� � %% ���� ���� ���� �� � �� ���� %'��-� *�� ��� ����������� �� ���� � 2 ��� ! �� �� � �� '��%��� � �� �� �� � � � � ��� ��� ���'���� � � ���� � �� ��'�� � �� �� �� )� % �� ���� ��� �� �� ���,�� �? � � �%��� �� ��� �������� ��������� �� % �� � � *� ��� 2�� �%���0� �� �'�� ���� ��� � ���%���� � � �� �� �� ��� � ���� =�� ������� � �� % ���� ������ �� �� � ��� ��� � ������� � � �� � � ������ � �� � '���� ��� ��� � ��� �� % b!""g � b(("g ��� � ���� 1 ���� % �� ���� ������� �� � ������ �� ' ���� �� � ���� � %���� �� �� ��� �? � %� ���� � ������� �� �� ���� %� ���� %���� � � '���� ��� ..���� ������00 � ��� � �� �� � ��� � � ��������� ��� %���� � ���� ? �� � � �� % �� �%' �� �� � � ��� ��� � ��� ��� %���� �? � �� ��j �� � �� � � �� %� �� � ������ �� �� � ��� %� ����� 2 ��� ! :��'���� � � ��� ������� �� � %��� ��'�� � �� �� ��� 578 � ��� ���� � � �� �� � 5�� �� ���,�� ��� ����8 *� �� �%��� ��%���� � ������%��� 5!&8 "�(( 5#�""8 "�(a 5#�"(8 /��� ����� � � 5��8 "��� 5#�""8 "�(" 5#�#�8 :�� ������ ����� � � 568 "��& 5#�""8 !�7# 5e�&"8 )�����%����d 5 ���� �� �� �8 5##8 h� �9���� "�&! 5#�#68 "�!7 5#�""8 "�(a 5!�#(8 )����� �� � � �9���� "�a7 5#�""8 "�aa 5#�"#8 "�6! 5#�"e8 * ��� "��( 5#�#a8 "�&a 5#�""8 "�(� 5#��!8 � �� "�6# 5#�""8 #�#a 5#�&"8 #�#� 5#�!(8 2 � � ' ��� �� "�e� 5#�""8 "�ea 5#�"e8 "�a" 5#�!(8 � �� )�� �� �� �� 5#68 "�&& 5#�""8 "��( 5#�&e8 ��� �� � 5!#8 ��(6 5#"�!"8 !�(� 5(�e�8 "��( 5#�""8 2��� � ��� �� �� ��� ��� ���� � � �� �� � 5�?8 � � ��� ��� ������� �� 9����� � � � � � � ��� ����� � %��� � �� �� �� 2�� �� �� ���,�� ��� ���� �� �� �� �� ' ��������� � ��� � ��� ������� �� ��� %���%�% � � � � 9����� �� 2�� ����� � %����� � �� �� �� ���� ��� � ��� �� %'��-��� % ���� �� % *� �� ��� ��� ��%���� �� )� ��� ������%��� 5 ���� �� �� �8 9����� �� ��� � � � ' ��� �� �? �� � �9� ��� �������� ��� �� � ��� ���� %' ���� �?�� 7 �� ��� ��%��� � ��������� '�� 9����� �� #" %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 �� ��� �� ���%��� �� � � ��� %��� �� � ��� � ���� � ��� � ���� ��� � �� %� ��� ���� ��� � �%'���� � '� -� � � ����� %'��-���� ��� �� �� � '���� ���� ����������� ���%��� �� � ' ����������� % �� ������� �� � � '�� � �� '�� � i���� � ���� �� �� ������ �� �� � ��� � � �%'��� %���� ��;� � ���� � ��� ��� �-��������� � � �� � � � ��� �� 2�� ��� �� � ������� ''� � � �� � � % �� �� 2 ��� !� )� ���� ��� ��� *� �� ������� � �� ��� ������� �?� 2���� �� ������ �� ��� ����%��� % �� ��� ��� %'�� �� � � � �� �� ��� ����� ������� �� ��� �� � � ��� �� ��� � ��� �� �� � ������ i�;� *� �� � 2 % �%���� ����� � ��� �� � � �� � �� '� ���� ,�� ����% �� � � � ���� ��� � %� � � � ������ ��� ����%���� % ��� � ���� �� �� ��� � % � �-'� �� ���� ������� h���;� ��� ��%���� � ��� ����� ��� %� � ��� �� � ����% �� �� � �� b&�6ai� ��� ��� �� ����% ��� �� *� �� �� �%��� �� be�� �� �� ��� �� b#7�7 �� �� ��� ���'� ������� 2�� ��������� �?� �� ����� ��� �� 9���� � ���� 2 ��� & ������� ��� ��� ��� �� �? �� �� � ������ � � � %��� � �� �� ������ �9� � ������� 2��� ������ '� ����� � �����%��� � ��� ���'���� � � ��� � �� �� � 2�� � ��� �� �� �� � ��� ��� ����� ��� % �� �������� ��� � � ��� � �� )�� ������ � �� ���� ����� � � ����% �� 9����� ��� =� ��� ���� � ��� ��� ������ ����� � � ����% ��� �� ��� �� � %'�� �� �� �-����� ��� ������� ��� �� �?�� 2�#� � ��(� �� � ����� 2�� '�� ��� �� � ������ �� ����� 5�� �� �� � �� ' ����� �� ��� � %� ���������� � �� � �8 � ���� � �� �� � � � '� ����� � % ��� )� � %� �� �� �� � � ��� ��� ������ 2 ��� & � �;��� � ��� � �� �� � � � '� ����� � % �� 2��� � ��� �� �� �� �� ��� ��� �� � ��� ��� ���� � � �� �� � � � � � � ��� 9����� �� �;��� 2�� ��� �� �� � ;�� ��� ��� � %��� � �� �� �� %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 ## ����� � � �� ��� � �� ������ � ����% ��� ������ � � �� 9���� ����� 2�� � %� � �� � �� � � �%'��� ' ���� � �� e"p � ��� ��� ������ ����� �� � ���� �� �� ��� ("p � ��� � �� ������ � ������ ������ � � �� ��� ���� �� �� � �� ������%��� ����� � � %� � � ���� ���� �� � #"p� �� � ��� � ������ � � � �� �� ��� ���� ����� � � � ��� �� � ������ ������ � � ��� � � ����� � �� � ��� =� ��� � �� ������ �� � � �%'��� %���4'����� ����� ��� ��� � �� � � ���� �� #! �� � ��� &# ������ �� ��� � �� % � � � � �� � �� %���4'����� '� �� � � ����� ����� ���� � ;� ����� � � �� ���� ��� � �� � �� ��� � ���� ����� �� � � �� � �� � ��� ������ ���� �� ��� �� ���� � �������� % ���� ��� ���; �� ���� ���� 3 � �%'��� �� '��� � ���� �� ��� ���; � ��� � �� � �� � � � ��� g�'������ ������ ��� ���� �� � ;�� � ��� %���4'����� ���� 5 � '' ��� � � ���� ' �� � ��� g�'������ ����8� %���4'����� ����� � �� � �� �� ���� � % ;� '� ��� ����� �� � '�� � �� ��� � �� � � ��� � '� � ����� � %� � ��� � ' � ��� %���4'������ �� ��� ���� �� ���� � '��� � ��,� ��� ��� ��� �� � � �� � �� '� ���� ������ � ������ � �� ������ ��� � ��� � �� �� ����� �� ' ���� � � ��� � �� � �� ���� ��� ��� � � � %���� � �� � ����� � �� � � �� �� ��� � � � ����� ����� 3 � ��� ������ ����� � �� #& �� � !( ����� ���� ������ �� �� �� � ���� ���� %����� ���;� � ��� ������ � ���� 5�������������� ��8� )� ��� �� � ��� %���4'����� ���� �� �� % �� � ��� ����� �� ��� �'��� �� �� � ���� ���� %����� '� ������ ���; � ��� ������ ����� � � � �� � � � � ���� ����� ���� �� ���� � � � � �� ���� � � ��� � �� � � ���; � � ���� ��� �� �������� � ��� � %��� � ��� ' �� � ���������� %� � ��� ����� ���� ������ �� �� �� � ���� ���� � ��� � ����%�%���� �� %��� �� %� 2��� �� ���� �� % %��� � �� � �� � � � �� !"" ����� ��� �� � ��� ��� ������ ��� � % �� ����� ��� �� �� �% �� �� �������� ��� �� � �� �� ' ������ ��� '����� �� ���� � �� � ���� � ������� 2�� ���� � ���� '� ����� ����� � � �� �� � � �� �� �' � ��� ��� ��� ��� ������ ��%'�� ��������� ���� �� ��� %�%����� 3 � �%'��� ���� � � �% �� ����� ��� � % � �� ��%'�� ������ ' ��� =� �� � ����� ��� ����� ��� � � � ���; �� � �� ��� � %������ ��;� .. �� ; ���� � �����%��� � �� ��� ������ ����� � � ������00 ���� ��� ���; �� �� ;��� ��� ���� �� � ;�� � ��� :�� ������ )���� � � � �� � � � ��� � ��� �� � �� �0� ��� ' ��� 2�2� *�4 ����'� ���� ������ 2 ��� � '� ����� ��%% �� � ��� ��� � '� ����� �� ��� ���� ���� ��� o �� �� �� � ���� ���� �� '� ���� �-�� ���� �� '�� ��� � � �� 9�� ;�� ���'������ 2�� � ��� � ��� � �� 9���� � ���� 2 �%��� � � ��� �� ���� ��� � %' ��� �� '��� �% �� � ������ 3����� �� %'��� � � ��� ������ ���� ��� � ���'���� � �� ��� ��� � � %��� � �� �� � )� ��� *� �� ��� ��� ��� �� � �� � �? 5���8 � �? 5���� ���� �8 �� ���� �� � �� 5 � �� a!p8� 1 ������ �� ��� �%��� ��� ���� � �� ��%�� � � �� ##!p� )� �� ���� � �� �� � �� ��� ��� � ��� ���� � ���� ��� � �� ���;�� � ��� ��� ����� %'��-��� 5*� �� � �%���8� �� ��� ���� �� � �� %�� % �� %'��-� ��� � �� %�� ���� � %% ���� �� ���� ������� � � �� ��� � �� �� �� �������� 1 ������ ��� � �� � �� #! %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 http://www.calcbuilder.com http://www.life-line.org http://life-line.com http://www.member.cunamutual.com � �� *� �� �� �%��� �� � �� � �� � ��� ���������� �� � ��� ��� � �� �� � � ��� '� -� � � ���� ��� � ��� ���� � ��� � ������ 2 ��� ( �� �� � � ��� �� ���� ��� � ���'���� � %' ��� � �� � % ���� ��� ��� � �� ��� ������ % �� �������� �� ��� � �� )�� ������ � �� ���� ����� � � ���� 9����� ��� ����� �� �� �� ��� ������ ����� � �� �� ��� �� � ����% �� � � � � ���� ���� ��� ������%��� ����� ����% �� � �� � �� �9� � �� � �� �� � ������� ���� �� ��� ��� �� ���� ��� %' ��� � ������� ��� �� ���� ��� � � � %'����� �� �� ����������� � ��� ��� ��� ����� �� � �� � � �� �� � � � ��� �� 3 � ��� ���� ���� ..��� �00 ��� �� 2 ��� � ��%% �� � ���� ���� � ������� �� � %��� � �� �� ��� i� � � ��� 5������ �8 �� �� �� ���� �� � �� *� �� ������%��� b #!e!�& !!�7�� 7"e�" /��� ����� � � b a&a�# ##6e�� &7(�e :�� ������ ����� � � b #�a !�7 #�# )�����%����d ���� �� �� � p h� �9���� e(�( &"�" 6�& )����� �� � � �9���� #&�" !"�" (�6 * ��� #!�( !(�" a�! � �� a�" #"�" &�( � �� )�� �� �� �� b &e�" #!6�# �(�# ��� �� � b "�" "�" "�" *� �%��� ������%��� b #(e��# !��(�# 6&!�& /��� ����� � � b 7ea�# #7""�" �a7�! :�� ������ ����� � � b #�! (�" #�6 )�����%����d ���� �� �� � p h� �9���� (7�( &"�" #"�# )����� �� � � �9���� #&�" !"�" (�6 * ��� #6�( &(�" #"�� � �� 6�" #(�" ��e � �� )�� �� �� �� b ##�! !a�" #!�# ��� �� � b !7�� !7&�6 ae�e �� ��%���� � )�����%����d ���� �� �� � p h� �9���� �"�" &(�" ##�a )����� �� � � �9���� 6�( #(�" ��� * ��� &(�" &"�" ##�# � �� #(�( &"�" a�& ��� �� � b !##(�7 7a(e�� !(e&�! 2��� � ��� �� �� ��� ������� �� % ��� ���� ���� � ���' ���� � � ����������� '��� � � �� � � � '� ����� 9����� ��� 2�� ' ���� ��'������ ��� *� ��� �%���� �� ��%���� � � �� �� �� ���'� ������� : �� � ���' ���� �� �� �� �� ��� �� ������%��� �� �� �� �� �� '�� ��� ���� %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 #& �� � �� � %'�� ��� � ��� �� � � ��% ��� � �� ��� � � ����� �����%� � �� � � � ���� ����� � �������% ������ 2��� ����� �� ��� ����� �� ��� �*� * �;��� ��� � 5�� ��% ��� !"""8� =�� ����� ��� �� � � ����� ������ 9���� �������� ���� ��� � �� �� � �� ��� ��� � ��% ���� =�� �� � � �� �� � ����� ��� �� ����%� ���� � ��������� ���%'�� �� �� %��� �� � ����� ������� � %� � ����� �� ����%� �������� �� � �� ����� � ��� % ���� � % �� ��� ��� ��������� 3 � �%'��� ������ � ��� ���� � �� � � �� � � �� � � �; �� ��� ���>� � �� �������� % ������ =���� �������� �� %���� �� ���� �>� ����q� %� ������%��� � �� � � ����� ������ ' ��������%��� ��j �� �� ����� ����� � � �� 3 � ���� ��� �� �� �� � � � � ' ����� � � �� � � �� �� �� 1�% � ������ % ;� %��� ;��� =� ��� ���� ���� � ���� � ���� ������ ������� ' ��� �' ��� � �� ���������'� ������%���� ������� ' ��� �' ��� ��� � �� ���� ������ � �� ��� 9����� �� �� �� ��� � '� �� ��� ..���� �����00 � � % �� ��� �% �� �� ����� � �������� ��% �;�� �� � �� � � ���� ��� � ��� ���� � � ��� ����� �� % ������� ��� �% �� � � ��� =�� �������� ��� � �� � ;�� % ��� �� � � �� � ����� '��' �� ��� ������� 1�% � �� � � � ���� �� �� � �� ����� .. �� ��� ��00 c �� � �� � ������������� ���� ���� �� �� % % ����� �� �� � ���� ���� ��� ����� � %������ ���9��� ����� �� �� � ��� � %�� � %��� � � �� ����� ���� �� %���� ��;��� ���� �� �� % ��� �� �� �� � �'�� 5� � �%'��� ..#"" %���� � �� ��00 �� �9������8� )������������� ��� ��� ������ ����� � �� �� � �� ��� ���� �� �� � ���' ���� � ��� ���� ���� �� �� ���� ����� %���� �� ��� � ����� �� ��� � �� �� ���� % ����� ��������� 2��� � �� �� �� ����� � � ����� � �� � � ����� �� � � �� ����� � �� ��� ..� �'��00 � ���� �� �� �� 2 ��� ( � %' ��� � � ��� ��� ���� ��� � ���'���� � �� �� � � � '� ����� �� 2��� � ��� �� �� ��� � �� �� �� � � � '� ����� �� � ��� ��� ���� � � �� �� � 5�?8� 2�� � �� �� �? 5��� ��� �8 ������� �� �? 5����� ��� �8� 2�� � �� � ���� �9� � ������ � �� � %��� � �� �� � �� ����� ����� �� � � � � � ��� �?� � � ��� �� � � � � % �� 9����� �� �� ����� � � 5���� �� ��� ������8� � �� )��� ������%���� ���� �� �� �� �� ��� �� � 5������� ��� *� �� ��8� #� %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 /��� ���� �� %���� �� �-����� � �� �� � �� ��� � �� ��� � ��� ��� � '� % �� ��� �� '� �� �� � ����� ��� )� �� ������ ������ '��� '� ���� �� � � ��� �� � � ��� ���� � �� � '� �� �� 2�� � �� �� � �� ���� ��� � �� ���� ����� ��� �� �� ��� � ��� �� � ��� �-���� �� � '� % �� � %��� ���� '��� ��� �� � � � ��� � ..�� ��� �� � � ����00 2�8� ��(������� � ��� �� �'���� 2�� ������� � �� ����� � ���%�� �%'�� �� ��� 2�� ���� �� �� � ��� ��� � � ��� ��� �� � � % � ����� � )� � � � �� � �� �� �� � � �� �� � ������� � �� ����� %'��-��� �� ��'� � 9����� �� 1 ������ ���� ��� � � � � �� � � � � 2�� ������� � � ��� ��� � �� � ���� ���� �� '� ���� �� ���� ��������� � ����� %'��-���� ��� % �� � �� �� �� � %� �� � � � '� ����� � % ���� =� ��� �� ��� ��� ��� � �� ��� �' 9���� ���� �� ���%� � ������� � � � %����� % �;�� ���%���� )� ��� *� �� ��� � � �%'��� ��� ��� �� �? �� ��� ��� � � �� � % ��� �� �� ��� � ��� 5a!p8 �� � �� � � ��� ���� ���� ��� 2�� � �� ��%�� �� ��� �%��� ��� ��� � ##!p �� �� ����� � � ���� ��������� �� % ������ 2�� ������� ������� �� � ��� ��� � �� ������ � � � �� �� ��� �� � ���� � ��� �� �� � � ���� � ��� � ���%���� � ���%��� %��� ��� � ������ %������ � ��� ''� � � � � ����'����� � � �� ���������� )� ��� �� � % ���� ������ ��� � �� � � � %�� ���� 2�/� ��(������� � ��� ��� & � ���� (�� � � � �'���� 2��� ' '��0� ������� � �������� � ��� �� ��� ��� ���� � � �� ��� ������ � �� � � � '� ����� �������� ' ������ �� � ������ ��� ��� �� ���� ��� �� � ���� ����� �� � � � ��� '� ����� � � ����� ���� ���� �� ��%�� �� ��� ��� �� �� � � � ����� � '� ������ � ����� ���� ��'� � ' ������� 2� �%'��� �� ���� ���� 5!"""8� �� ��� ����� ��� > ��� ��� �� �� i��� � �� i' ���$ �� 2� %��� 5!"""8� �� �%���� ��� ������� ������� ��� �� �� ����� k n ���� 2��� ' '�� �%���� �� ;�� ��%���� �� �� � � �� ������ �%' �� 2�� ���� �� ����� ��� � ���� 2�� ������� ������� �� � ��� ��� � � ���� �� � � ��� > � '� ������ ��� � � %����� % �;�� ������ �� � � ��� *� ��0�� �� � �� ��� � ��� � � ��� b!""g� ��� � � ���� ������ ���� � ��� ������ � ��� � ��� 5� ��� ������ � ������%��� �����8 ����� ��� %� �� �� � '� ���� %��� � �� % �� ������� ��� ������ � ��� ����� ��� � �� �� ���� � � '� �� ��� '�� �� �� =�� ������� ���� ��'' �� � ����� ��� ��� � ' �� � ��� �������� %� � ���% � � �� � � � ��� � � ��� %����� % �;��� � �������� ���� ��� 5#6668� ��� ��� ����� ���� '� ����� %'��%��� �� ��� � � �� � ���� � ��� ������ � ��� �� �� %%��� �� ���� ������� * ��� 5#6668 ' ��� �� ��� ��� � % �� � ���� � � ���� '� ������ 2 ;�� �� �� � ������ �� ������� �� � �� �� ���� � � ������ �� � � ��� �%���0� 5��� � ��� ��� b(""�"""8� ��� � �� � � �� �� � �� ��� � ���� ��� � �� ���� � ��� )� ���� ���� �� �� � � ����� 9� ���� ��� ���� � ��� ���� ����� �������� ��� �� ���� ������ 2�� ��%���� � �������� � ������ ����� � � �� � ��� ���� ��� � ��� ����� ���� ��'� ��� �� ������ ���� �� ��� %�,�� ��� �� �� c��� � �� 3��� 5!"""8 ����� ����� ..'��� �� �� ���00 ������ � �� ��� ��� � �� ������ �� � %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 #( ��% �� �������� � ������ �� )� ����� ���� ��� ��� �� ��� �� % �� � %%��� �� � � � �� � � �� � ��� � � �� �� ��� ����� �� � � � '� ����� � % ���� 2 ��� ( �� �� �� �� ��� ���� � ���� ��� �� ��� ��� � �� ��� ���� > � � � �� ������ � �� ���� ����� � � ���� 9����� ��� * ��� � ����� �������� ������ ��� � � ����� � � �� %���� �� � � ���� �� � � ��� ������ � ���� ������� )������������� g��� 5!"""8 ��� ����� � � 3������� � � �� ����� � �� ��� ��� �� ����� � � 9� ���� ����� ������%��� '� ����� � �� �� � �� �� ��� �9� � � �� � %��� � �� �� �� ���������� �� � ������%��� � �� � �� % � ���� �� � � ��� ��� ������ �� � � %����� % �;�� ������ �� ' ��� �� �� :�� ������ ����% ��� �� �� � �� �� � �� �� �� /��� ���� �� ���� � �� % �� �������� ���� ��� �� � ����% ���� 1 ������ �� ����� � �-�%'�� � �� �� ��� ��� % �� ��� �� � ������ � � �� % ��� %����� % �;��� )� ��� ..'��� �� �� ��� % � ��%���00 ���� �� ��� ��;� ��� ��%���� �0�� ����� ��� �� � -�� ���� ��% �� ������ �� ������ ���� ��� � �� ��� ������� � �� � �� ����� c ��� �� �� �� � �� '��� '� ��� % �� ����������� � ��� � % �� �� �������� � ���� �%'����� � � ��� ����� % � �� �� � ��� ��� � ��� � �������� > � ���� ���� � �;� /��� ���� �� % � ���� � � �� ' ��� �� � ��'�� � ������%��� �� � �� � ��� ��� � ��� � � � ��� 2�� �������� � ������ ������� ��� '' ����� /��� ��� � � ' ��� �� �� �� � �� %� � % �� �������� �� � �� � � �� ��� � ��� ���� 2�� �%'�� �� � �� �� � ���� ���� �� �� ��� ' � % �� ������ � � ���� � �� � ��� � � � ����� ������� ���� � ��� %����� % �;�� ������ � ���������� =�� � � �� ������ ;��� ���� ���� � � �� � '� ���� � �����%��� � ��� ������� � � �� � � � ��� � ����� ������ �� ����� �� � �� ��� ������� � �� �� ���,�� ��'��� �� % ����� � ����� '� ����� �� ������ �� � ��� ��� � �� �� ��� �� ���'���� � � ����� ��'��� �� %� % �� %'��-� /��� ��� � �� ���� ��������� � ����� %'��-���� � � �� �� � � � '� ����� � % ���� ����� � ����� � � ' ������ �%����� ��� ��� � �� % �� �������� � 9����� �� ��� ��� � ���� ����� � � ������ �� � ������� 5�� % � �� ������ �8 �� � ���� ��� �� /��� ��� � � � � ��� ���'���� � � ��� �� � �� ���� �� �� � 9����� ��� =�� ���� � � �� ��� ������ ��� �������� � �� �����%� ��� �%����� ��� ��� � � ����� � � �� � � � ��� �� *� �>� ������ �������� ��� ��������� �� �� ;������ � ��� ��� � ��� ���� � ..����00 ��� �� ���� ���� � � ����� ���%��� � �� � � � � �� �� ����� �� ������� ��� �� � � ��� ������� )� ��%� �� ������� ������� �� � � � %����� % �;�� ���%���� ��� ��� �� �� � � ��� ����� � � �� � � � ��� �� 2�� ���� � � �� � � �%'�� �� �� � � �� � � � ����� � '� ������ �� � �� ���������� �� ��������� ��� ���; ��� ������� ����� ��� �� ���� ���� � ��� �� ��� �� �� �� ��� ���� � � ���� �%' � ��� ������ � ��� '��� � � �� � � � '� ����� ��� �� � 5 �� ��� � ��� �� � � � ����� �� ��������8� =�� ������� ������� �� � ��� ��� �� �� � � � ��� ���� > � �� � %� �� �� �� � � ����� � � �� �� %� � -� =���� �� �� �� � � ��� �� � �� ���� �� �� �� �� ��� ������ � � �� ������ �� % ��� � ���� ���� � ��'��� #e %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 ����� �� � /��� � ��������� �� 9����� � ������%��� �� � � %< ���������% �� �� %��� � % � �%�������� � �� % � %�������' �� �;� % �������� % � �� �� % ;�'������� % �� �% ���% � ��%���� % ����� % �������� % � �� %�� % � �;����� % %��� ������� % � ��� % �% ���� % ������%������������ % % �� � � �� % ���� �;����� % ������� % � ������� % ��%�� ������ % ���� ��� % � ������ % ��� � % � � % '��� %���� % � %����� � �� % %��� ��-'����� % > ��� � ;� % �� ��'�� �� % � �� � % ''�����%�������� % ���; ������ % /��� ����� � � �� ��� % 9� �� �� �� ��� % � ����� % �� � ����� �� ���'��� % ��� %� % 9� ��;��� % %���%� % ����%� % ��������9� ��� % ������� % � � %���������� � �� % ��9� �������� � �� % ���9� ��� % 9� ��� ��� � �� % 9� ��� % ������ � �� % ���� �;����� % ���������� �� % ��� ���� ��%��� % %������� % � ������� ��� % �������9� ��� % �������� % ;� ����� ������� ''�� % ���������� � �����%������ � �� % >��� 9� ��� % ���������� � % ��� ���� � �������� �� % �������������� � ����� % ���% ���� % ���%����'� �� % ���� 9� ��� % '� �� � � �� % ��;���%��������� � �� % ����������� % ���� � % � %�� % �9� ��� % �'� ���%����� � ��� �'� % '�����9� ���� % '� ���%�� � ���� % ��������������� % :�� ������ ����� � � ���������� �� � ����������%��� �� % ����� % % � � ��� �� % �% ��% ���� % �����9� ���� % � ����� � ���� ��� �'� % % ��� ���� �� % ���� ��� �� % % ��� ��� % % ��� ���� % ����,� �������� % ���� �'� % ���� %%������ % '� �'���� % ������� �� �� ��� ��� �� ���� �� �� �' ���� �%����� ����� % �� ;, � % ���� ����������� � �9� ���� % %�%���� �� %��� �� % �� � � ������ % )�����%����d ���� �� �� � �����' �� �;� % > ��� � ;� % � �� � % ��' ��� �� � % �������� % � �� �� % ����� % �% ��% ���� % % ��� ���� ��%��� % ���� ��� % ;� � '�� �� % �� �� %��� � % )� %� � -d � �� ������ � � ��� ��� % ���������� % ���� %��� �� �;� % � �������� % �����&e"� % � ����'������ % �������� % ������� �'� % � � ��� �� % �,,;#"�7� % �� �'���������� % � �� % �����> ��� �� % �>��������� % �������� % � ����� �������� % �% ��% ���� % %� � ' � % ���� ��� % � ������ % �� ��� �%� % '������� �� % ������ �� � % ;�'������� % ����� % 9�� ;��� % % ��� ���� �� % � �� �� % %��� ��� % � �������� % % ��� % ��� �� � � �;����� % ���� ��� ��� % � %����� ����� ��'� � ��� % �������� % � �� �� %� % �� ��� �%� % ��� ������ % �������� % �� �� %��� � % ���� ���� % ���������� � % �% ��% ���� % ��� ��'� ������ % ������ ���� ,��� % '�� �'� ��� � % % � �����% �� % j ��� �'� � ��� % ���� �;����� % %�;� ������ % �� ��� ������ %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 #7 http://finncalc.com/afretin.htm http://transamerica.com http://farmersinsurance.com http://home.wingspanbank.com http://fidelity.com http://schwab.com http://kiplinger.com http://dionmoneymanagement.com http://nefin.com http://nbfunds.com http://dcihome.com http://banksite.com http://americanfunds.com http://nasd.com http://onmoney.com http://retirement.russell.com http://mostchoice.com http://newyorklife.com http://usnews.com http://founders.com http://columbiafunds.com http://estrong.com http://waddell.com http://cigna.com http://olac.com http://putnaminv.com http://womensfinance.com http://americanexpress.com http://johnhancock.com http://troweprice.com http://valic.com http://oppenheimerfunds.com http://erekafunds.com http://finaid.com http://quoteadvantage.com http://cnalife.com http://tiaa-creff.org http://bygpub.com http://genam.com http://quotekey.com http://amterm.com http://accuterm.com http://freelifequote.com http://insweb.com http://economylifeinsurance.com http://allquotesinsurance.com http://lstquote.com http://quotenavigator.com http://accquote.com http://cwinsurance.com http://newyorklife.com http://life-line.org http://moneycentral.msn.com http://metlife.com http://garden-state.com http://intelliquote.com http://lifenet.com http://kwot.net http://insurehoppe.com http://life-insurance-term-insurance.com http://justaquote.com http://sunlife-usa.com http://hutchinsoncreditunion.com http://ter-life-insurance-4u.com http://termonly.com http://termlifepros.com http://reliaquote.com http://pga�nancial.com http://ntktermlifeinsurance.com http://budgetlife.com http://byfsco.com http://nams.com http://iquote.com http://spectruminsurancegroup.com http://phillyquotes.com http://proterm.totalsw.com http://westernreserve.com http://life-line-org http://northwesternmutual.com http://nefn.com http://macatawaagency.com http://smartmoney.com http://4bestquotes.com http://tcu.edu http://gaudreaugroup.com http://moneyadvisor.com http://cyberconnect.com http://moneycafe.com http://moneyclub.com http://venezia-insure.com http://hfgroup.com http://�rstcommunity.com http://planplus.com http://andrewsfcu.org http://ss-cu.-org http://sunfcu.org http://fpa.net http://smith-fought.com http://stockzoo.com http://4disabilityinsurancequotes.com http://member.cunamutual.com http://finnancialserv.com http://wingspanbank.com http://johnhancock.com http://valic.com http://finportfolio.com http://findelity.com http://schwab.com http://gefn.com http://smartmoney.com http://moneycentral.msn.com http://estrong.com http://oakwoodcapital.com http://transamerica.com http://vanguard.com http://algerfunds.com http://firstamericanbank.com http://calcbuilder.com http://austin360.com http://roth.upi-net.com http://fidelity.com http://calvertgroup.com http://gocarolinas.com http://wzzk1047.com http://realpittsburgh.com http://ocnow.com http://news-journal.com http://gjsentinel.com http://nbfunds.com http://sandiegoinsider.com http://smartmoney.com http://msco-cpa.com http://estrong.com http://ccbonline.com http://statefarm.com http://prudential.com http://wellsfargo.com http://kiplinger.com http://cnnfn.com http://quicken.com http://moneycentral.com http://schwab.com http://mutuals.com http://4anything.com http://mony.com http://vanguard.comalgerfunds.com�rstamericanbank.comcalcbuilder.comaustin360.comroth.upi-net.com�delity.comcalvertgroup.comgocarolinas.comwzzk1047.comrealpittsburgh.comocnow.comnews-journal.comgjsentinel.comnbfunds.comsandiegoinsider.comsmartmoney.commsco-cpa.comestrong.comccbonline.comstatefarm.comprudential.comwellsfargo.comkiplinger.comcnnfn.comquicken.commoneycentral.comschwab.commutuals.com4anything.commony.com http://banksite.com http://thehartford.com http://tamu.edu http://willsandprobate.com http://lifenet.com http://rowbotham.com http://statefarm.com http://estateweb.com http://fidelity.com http://transamerica.com http://agedwards.com http://sunlife-usa.com http://smartmoney.com http://estateplanning.com http://riddleandbrazil.com http://capitalplaninc.com http://marcweissman.com http://florida-probate.com http://newyorklife.com http://mikecollins.com http://florida-trust �������� ������ 2�� ��� �� �� �� ����� � g ��� / ���� ��� ���� �� �� � ��% �� ��������� i ���� 4� � �� �� � c�����'� � � ���'��� %%����� �� � �� �� � i ���� � � �%������ ���� ��'' ��� 2�� ��� �� ;� ������ ��� ����� �� �� � � � �' �� ����' � ��� 3 ��� �� � � � 3�� � � � c� ������ 2��� � �; � � '��� �%�� ����� �� �� ������ � ��� 4� ��� �� �� h��������� ���� � � 3 ��� �� �� ���������� �� ��% �� 5!"""� ��'��%���8� ����� ��� �� � � � '� ����� ����� ��� �� � ��� ��� �� ��� �� 9� � :�'� ��� ;#� ##�� * ���� � �� 5!"""� �'��� #78� =����� ��� � � �� ���� '� ����� ��� � ��� � � � � ������ ��� #1� �7� * ���� � 5#666� o ��%��� #8� o�� ���� �d c� ���� �� � � ��� ��� �� � �� ������ ���; ����� ���� �������� ������� � � ���� �� � �� !"6@!#"� *� ��� �� k 4 ������ �� 5!"""8� : �� ��� �������� % ;�� % �;��� % �� %'�������+ ������ � �� % ��� ���� ����� � � ��������� � �;��� c '��� �� 3� g������ � � � � 4 ����%���� 1 �� �� h���������� : �� �� i�� k : �� �� � 5#6668� %���� � � �((������� �� 5!�� ���8� ��� � ��� /�d : �� � c���� �� ��� 4������� *� 5#6678� �� ������� � �� � � � ����� �� ��� �� �� � ��� ��������� � � ���� �������� )������ <� #!(@#�"� ����� �� 5!"""� i � !68� i��� � ������ � �� ��� � ��� �� � ��� � � � � ������ ��� #1� !a� ���� �� 5#666� :� �%��� #8� 2�� ����� ����d 2�� % �� ���� ���� ��������� � � '� ����� �� �� �� �� � �� '� ����� � �� ����� �� ����� ����� � � ���� �� � �� #�&@#�e� g���� �� 5!"""� i � � e8� 3������� � �� ��� ����� � � 9� ��� '� ����� ����� 7���� �� � ���������� $,8� &� / �� � 2� 5#666� o ��%��� #8� )�0� � ������ �� % ����d � ��� ���� � � ����� '���� � �� � ����'������ '� ����� ���� �� � �;��� � ���� � � ���� % ���� �� �;� � ��� 9� ����� �� 9� ���� � %'����� �� � � ���� �� � �� #&7@#�!� i ����� *� 5!"""8� ��� � �0� ���%��� ��� ��� �� �� � �;��� '� �� ��+ � � ��� ��� �� � ��� ���� �� �������� �� '' ���������� �%������ c �%���� = �� � � c '�� ������� ������ )������ �� 9 �� %������� i ���� �� 5!"""� �'��� #8� i �;����� ��� ������q�%�� ��� ��� %'����� �� ��� ��� ���0� >��� ��� �� �� � �� ��� �� � � � '� ����� �� ���� ���' % �;����� ����� '� �� ��� � � ���� �� � �� #!�@#!a� c��� �� ��� k 3���� g� 5!"""� 3���� �� #8� g��'��� ����� �� ����d � �����,������ �� �������� ��� ���� � � ��� '��� �� �� ��� ����� � �� � � � �;� � � ���� �� � �� #&&@#&e� � >� ' �� ��� ?��; � � � %� i�� k g �� � �� 5!"""8� : �� ��� 9� ���� � ����� ��� %�� �-'����� � �� �� ���� �� ��� %'������� �� �� �� � � �� %%�� � ��%�+ � �;��� c '��� h��������� � � ������ �� � ��� ��� k : � � ��� *� 5#6668� *��� � ��� ���d 2�� � ' (" �� � � � ���������� =� ��� #15#!8� #7a@#aa� 2� %���� i� 5!"""� ��� (8� �2� ��� ����� k n ��� ' �� �' � ���� �� � � � ��� �� %��('��������� 28� #e� #a %��� %���������� )�!� *��� + � � ���� �������� )����� $, -#,,$. /0$1 an investigation of the consistency of financial advice offered by web-based sources introduction data case and financial issue data brady case smith case remington case web financial advice sources live advisor data results web advice independent sites web versus live advice implications for consumers implications for the financial planning industry conclusion acknowledgements appendix a references pii: s1057-0810(02)001130 ����� ��� � �� ���� ������� ��������� � � � � � � ��� ����� �� �������� � ���������� ������� ������� �� �!"#$ %�&!%%� ��������� '�� � � ��� ������ (��) �� ������� '��������� *�� � ' +,���� ������-��� .��������� ���/�������� �� �!"#$ %01!%"2 �� 3�� 4����� ��� � �56�7 �����*�8����� +���������� 6��8���$ � 4��8������ �* � � 7���*��� ����** '8����� 7�� ����� �� �!"#$ �21!�&9 ����� �**��� � � ��� ������ (��) �� ������� '��������� *�� � ' +,���� ������-��� .��������� ���/�������� �� �!"#$ %01!%"2 ��88������� ������ ' � ���� .������������ �* � � ���������8� �* 5����8��� ����8� :**���� 3� ;�33���� '���8����� �� �!"#$ 9!�2 ����)-7��/ �� 7������� �������������� ��� ������8� ����� �� ���������� '������ 4�8������$ ���������� +�����8� �* '��������8 ������ 4�**����8����� �� �!"#$ %10!%&� �������� � ��� ��� ������ (��) �� ������� '��������� *�� � ' +,���� ������-��� .��������� ���/�������� �� �!"#$ %01!%"2 +�)���� '������ � � ��� '��8) '���8���� 7���� ����� 7������������ '���� ������ �� �!"#$ �%&!�"" 5��3��� ' �<�� ��� ���/������ �* '���� ���������� 4�*���� ����*�� ��� 4�*���� ������3����� 6������ 6����$ � 7���� ����� '������������ �� �!"#$ 01!"" �������� 6���� �������������� ��� ������8� ����� �� ���������� '������ 4�8������$ ���������� +�����8� �* '��������8 ������ 4�**����8����� �� �!"#$ %10!%&� ���3���� +��)� ; � ��� � 5����8��� 7������=����� �8�$ >�< 6���/�8����� *�� � � 5����8� �����8������� �� �!"#$ �&1!%�2 ������� +��8� ��.�������/���� ���)!������ ��������� �/ �� � � ����� 7��)��� ������ � � ����� ��������� �� �!"#$ %"&!%1% ?����� 7�8 ��� + � ��6���*���� 4������*�8����� �� � ?�� �� .�*��������� +������� 7��)����� �� �!"#$ 0�0!0�" ?������ 7 @�3��� ��� � 4�8����� �� 4�*��$ 5�8���� �**�8���� +�/����� 4�*����� .�8���������� �� �!"#$ "9!9" ?����� � �� ��� ���/������ �* '���� ���������� 4�*���� ����*�� ��� 4�*���� ������3����� 6������ 6����$ � 7���� ����� '������������ �� �!"#$ 01!"" ?�8)� � ����� � � ��� ����� �� �������� � ���������� ������� ������� �� �!"#$ %�&!%%� ?����� '��/ �� 7 � ��� ������������� �* ���-����8��3�� .���� ��� .���� ��� "�� )# .������������� �� �!"#$ 21!��� �������� ;������ ; � ��� � a���� �* ���������� .�8��� '������$ � � a���� �* 7������� ������������� �� �!"#$ �&!09 � ������ @��� ��� ���/������ �* '���� ���������� 4�*���� ����*�� ��� 4�*���� ������3����� 6������ 6����$ � 7���� ����� '������������ �� �!"#$ 01!"" @��� 4������� ��(����������� � ����� �� >�� ;��� 3� ?���� ��� .�8��� ��� 4������/ �8 � ���8�������8� *�� � � 5���� � ��� ;���� �* � � ?�'��� �� �!"#$ 99!10 @����=� ?����� ��6���*���� 4������*�8����� �� � ?�� �� .�*��������� +������� 7��)����� �� �!"#$ 0�0!0�" @��� ���� a '�������� ���������� �* .��������� � ��8�� *�� ����������$ � >�< �//���8 ��� 6���/�8������� �� �!"#$ 19!2b (� ��� @���� +������ ��.������� � �*�$ >� ?��/ *�� ������������ �� �!"#$ %%�!%09 (� ��� @���� +������ ��(����������� � ����� �� >�� ;��� 3� ?���� ��� .�8��� ��� 4������/ �8 � ���8�������8� *�� � � 5���� � ��� ;���� �* � � ?�'��� �� �!"#$ 99!10 (�<���8�� ' ���� ���������� �* .��������� � ��8�� *�� ����������$ � >�< �//���8 ��� 6���/�8������� �� �!"#$ 19!2b (�<���8�� ' ���� ��� � 4�8����� �� 4�*��$ 5�8���� �**�8���� +�/����� 4�*����� .�8���������� �� �!"#$ "9!9" ��������� �������� ������ � � ������ �* .��������� 5����8��� 7��������� 0�9 (������ @���� � � ��4��� (��� �������� +�/���� 4�����-8��� ���������c��� �� �!"#$ ��1!�%1 (�� d����� �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� �� �!"#$ %&�!0�% (���� 4������ ��4��� (��� �������� +�/���� 4�����8��� ���������c��� �� �!"#$ ��1!�%1 7���f)���8� � � + � ��.������� � �*�$ >� ?��/ *�� ������������ �� �!"#$ %%�!%09 78+����� � ��3��� � � ��� ����� �� �������� � ���������� ������� ������� �� �!"#$ %�&!%%� 787����� �������� ��+��������� � '��8) 7��)�� ������ '�������$ � � ���� �* ��+ ����� 7������������ �� �!"#$ �10!�2b 7�����)�� 7�� � ����� ��a����3�� ��������� a����� 7����� 5����$ � 7���� ����� �������� �* � � :/�������� �� �!"#$ �"9!�b� 6��������� � @��/ ��� ��a����3�� ��������� a����� 7����� 5����$ � 7����-����� �������� �* � � :/�������� �� �!"#$ �"9!�b� 6������ ������ ��6���*���� 4������*�8����� �� � ?�� �� .�*��������� +������� 7��)����� �� �!"#$ 0�0! 0�" 6�������� �**��� ? � ��� ������������� �* �������8��3�� .���� ��� .���� ��� "�� )# .������������� �� �!"#$ 21!��� 6��*���� � � 5 � �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� �� �!"#$ %&�!0�% �� ���� ?����� ��.�������/���� ���)!������ ��������� �/ �� � � ����� 7��)��� ������ � � ����� ��������� �� �!"#$ %"&!%1% ���8 �������� ;������� ��� � a���� �* ���������� .�8��� '������$ � � a���� �* 7������� ������������� �� �!"#$ �&!09 '8 ������ 5������8) 6 � ��� '��8) '���8���� 7���� �����7������������'������������ �!"#$�%&!�"" '8���� ;������ ( � ��� � 5����8��� 7������=����� �8�$ >�< 6���/�8����� *�� � � 5����8� �����8������� �� �!"#$ �&1!%�2 '��� � '�����/� ��� � 5����8� �* ���)�� ������ ����� ����8�����$ '��� +�����8� *��� ��>�< g��)�� ������� '������ 6����������� �� �!"#$ ���!��b '/��=��� � � � ��� � 5����8� �* ���)�� ������ ����� ����8�����$ '��� +�����8� *��� ��>�< g��)�� ������� '������ 6����������� �� �!"#$ ���!��b ��=��� � ���� ��+��������� � '��8) 7��)�� ������ '�������$ � � ���� �* ��+ ����� 7������������ �� �!"#$ �10!�2b ����� ���� �� ������ ��5������ :/�� 7��)�� ��������� 7������� ��� '��8) 7��)�� 6��*�����8���� �� �!"#$ �b0!�1� ;������ ;������ � � �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� �� �!"#$ %&�!0�% ;���� ������� + � ���� .������������ �* � � ���������8� �* 5����8��� ����8� :**���� 3� ;�3-3���� '���8����� �� �!"#$ 9!�2 e� ��� ����)� ��.�������/���� ���)!������ ��������� �/ �� � � ����� 7��)��� ������ � � ����� ��������� �� �!"#$ %"&!%1% ������ ��� ������ (��) �� ������� '��������� *�� � ' +,���� ������-+�� .��������� ���/�������� '�� � ��������� � �� �������� ��� �**��� � ����� �� �!"#$ %01!%"2 ��� ���/������ �* '���� ���������� 4�*���� ����*�� ��� 4�*���� ������3����� 6������ 6����$ � 7���� ����� '������������ @�� � ������ ' �<� 5��3��� ��� � � ?����� �� �!"#$ 01!"" ��� ����� �� �������� � ���������� ������� ������� � � � ��������� � � � ����� � ?�8)� ��� ��3��� � 78+����� � �� �!"#$ %�&! %%� ��� ������������� �* ���-4���8��3�� .���� ��� .���� ��� "�� )# .������������� '��/ �� 7 ?���� ��� �**��� ? 6�������� �� �!"#$ 21!��� ��� '��8) '���8���� 7���� ����� 7������������ '���� ������ 5������8) 6 '8 ����� ��� '������ � +�)���� �� �!"#$ �%&!�"" ���� .������������ �* � � ���������8� �* 5����8��� ����8� :**���� 3� ;�3-����� '���8����� ������ ' ��88������ ��� ������� + ;���� �� �!"#$ 9!�2 ���������� �* .��������� � ��8�� *�� ����������$ � >�< �//���8 ��� 6���/�8������� a '������� @��� ��� ��� ' ��� (�<���8�� �� �!"#$ 19!2b �����) �� 7��)�� ��� '�=� �� 4����������� �* ������� �� '���� .���,��� 7��)���$ � � >�< e������ ������� � � 5 6��*���� ;������ � ;������ ��� d���� (�� �� �!"#$ %&�!0�% �������������� ��� ������8� ����� �� ���������� '������ 4�8������$ ���������� +�����8� �* '��0�b ����� � ��� � � � �� �� �� ������ �� ������ ������� financial services review, 33(3) 48 borrowing from family and friends: study of the european union sara diab,1 mustafa nourallah,2 and peter öhman3 abstract informal borrowing from family and friends suffers from the lack of formal agreements and can lead to severe consequences. self-control theory suggests some strategies for improving saving tendencies, which can reduce such borrowing. to examine what factors can enhance these strategies in the european union, this study analyzes balanced panel data from the global findex and eurostat databases for the years 2014, 2017, and 2021, identifying a pivotal role for debit card use and saving behavior in addressing informal borrowing. the study also raises questions about the effectiveness of public financial education and emphasizes the importance of improving related policies in the fintech landscape. by elucidating these findings, this paper deepens our knowledge of the relationship between debit card use and borrowing practices in the european union. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation diab, s., nourallah, m. & öhman, p. (2025). borrowing from family and friends: study of the european union. financial services review, 33(3), 48-60. introduction borrowing is a central theme in human history. while individuals in ancient societies relied extensively on borrowing from their communities, modern history has been characterized by borrowing from financial institutions. a problem is that the lower the ability of households to access formal financial credit, the greater their likelihood of encountering financial issues. in such situations, households may be forced to seek financing from nearby communities, such as family and friends (lee & persson, 2016). seeking informal loans can thus be perceived as an indicator of less favorable socioeconomic conditions. 1 corresponding author (sara.diab711@gmail.com), lebanese american university, beirut, lebanon. 2 mustafa.nourallah@miun.se, mid sweden university, sundsvall, sweden. 3 peter.ohman@miun.se, mid sweden university, sundsvall, sweden. borrowing from family and friends can be beneficial but also lead to adverse outcomes, including potential personal conflicts. additionally, such arrangements often lack the flexibility to be rescheduled (karaivanov & kessler, 2018), and in the event of default, such borrowing may result in financial instability for the lending party (blanc et al., 2015). between 2017 and 2021, there was a notable rise in borrowing from family and friends in parts of the european union (eu). for example, there was a 73.33% increase in such borrowing in greece (demirgüç-kunt et al., 2022). this trend could be attributed to higher living costs aggravated by external factors, which triggered a surge in gas prices and led to high inflation rates. https://creativecommons.org/licenses/by-nc/4.0/ mailto:sara.diab711@gmail.com mailto:mustafa.nourallah@miun.se mailto:peter.ohman@miun.se https://creativecommons.org/licenses/by-nc/4.0/ diab et al. 49 borrowing behavior can be related to self-control theory and two related strategies (thaler & shefrin, 1981): the control-based strategy focuses on households’ financial management in order to prevent overconsumption, whereas the incentivebased strategy concerns the importance of saving money for the future. these strategies require the use of tools, and financial technology (fintech) offers various opportunities to help households manage their personal finances. debit and credit cards are commonly used fintech tools, and their use has grown remarkably in households. demirgüç-kunt et al. (2022) surveyed debit card use by european households, reporting an increase from 68% in 2011 to 88% in 2021. using debit cards can help households in two ways: first, to enforce a reasonable spending limit and, second, to establish and adhere to a saving plan. nourallah et al. (2024) reported that debit card use would likely enhance financial capabilities, and stango and zinman (2023) argued that education was a crucial factor enhancing cognitive skills, which in turn could mitigate the biases that usually affect financial decisions. when examining borrowing from family and friends, focusing on the eu rather than on a single country is essential due to the interconnected nature of the economies of eu member states. external shocks, economic conditions, and policy decisions in one country can have significant effects across the entire union (nourallah et al., 2024). by evaluating data from various countries in the eu, we can identify broader trends that might not be apparent when studying an individual country, enabling the implementation of more effective policies to help individuals and households overcome financial challenges. we further argue that it is vital to focus on factors that can mitigate the negative impact of borrowing from family and friends, in order to support policymakers with insights into household financial conditions in various eu member states. in this regard, it is important to consider the arguments of nourallah and öhman (2021) about the role of appropriate fintech solutions and of lusardi et al. (2021) about the effectiveness of savings plans in helping households to manage expenses and minimize undesirable financial behavior. the aim of this study is, therefore, to empirically investigate how the use of debit cards and saving behavior can affect borrowing from family and friends in the eu context. this study also tests intercalibrations of other factors that may affect households’ informal borrowing. the results suggest a significant effect of using debit cards and practicing saving behavior in terms of controlling borrowing from family and friends. however, educational background does not affect the targeted borrowing behavior. in addition, it is important to recall that inflation can provoke borrowing behavior due to its effect on prices. in line with hamid et al. (2023), our study of informal borrowing supplies policymakers with empirical knowledge that can promote financial resilience in the eu. in fact, the study offers two significant contributions to the fields of fintech and household finance. the first contribution lies in identifying the potential for fintech solutions to enable more efficient financial management practices. the study concludes that using appropriate fintech solutions such as debit cards can help promote financial independence and will most likely help households to navigate financial setbacks and improve financial stability. the second contribution concerns the limited effectiveness of traditional financial education and training in fostering sound financial decisionmaking. this makes it possible to question the effectiveness of financial literacy policies in the eu. while financial education remains important (kaiser et al., 2022), it may not necessarily address the complexities of modern financial behaviors. this study provides evidence supporting the need for enhanced financial literacy programs within the eu, particularly those tailored to addressing the various unique situations of households. the rest of the article is structured as follows: section 2 presents the literature review and section 3 the methods. the results are reported in section 4, while the conclusion, policy recommendations, and suggestions for future research are addressed in the final section. financial services review, 33(3) 50 literature review the self-control theory of thaler and shefrin (1981) explores the dilemma of setting consumption limits and the resulting issues that arise from the conflict between consumption and saving. the theory suggests two main strategies for resolving these issues, either setting strict rules to control consumption or altering the incentives to save money. the two strategies require tools in order to be properly implemented, and bank debit cards are such tools. through using debit cards, households can follow a strict rule that limits their consumption (bachas et al., 2021). households could also sort their expenses into predefined categories, enabling them to assess their overall consumption in a period. nevertheless, the development of online stores and the availability of various digital payment methods in the fintech landscape have introduced challenges, such as impulse buying. meyll and walter (2019) provided evidence of a relationship between innovative payment methods and surges in individuals’ overall spending, which might affect consumption and saving behavior. when households face financial setbacks, they often rely on readily accessible resources, such as emergency funds, to navigate these challenges (demirgüç-kunt et al., 2022). these emergency funds are typically built through consistent saving, highlighting the importance of applying a disciplined approach (asebedo et al., 2019; despard et al., 2020). the more a household saves, the better equipped it is to build a sufficient emergency fund. such financial protections not only provide immediate substitutes during unexpected situations but also promote long-term financial stability, reducing the stress associated with unexpected expenses or income disruptions. browning and lusardi (1996) stated that a primary motivation for households to save money is to enhance their ability to manage unforeseen contingencies, and tufano (2009) argued that saving is an irreplaceable element of households’ sound financial management that enables them to invest money and increase their wealth. notably, households that lack emergency fund savings to cover unexpected life events may be forced to take disadvantageous loans. besides debit cards and saving behavior, educational background is essential to making sound financial decisions (nourallah et al., 2024). in a meta-analysis, kaiser et al. (2022) concluded that financial education has a positive effect on financial behavior. similarly, lusardi et al. (2021) reported that well-educated households tend to manage their money properly, which likely helps them deal with financial shocks. relatedly, nokulunga and klara (2023) found that people with low education levels have a higher probability of using the informal rather than formal financial sector, which may negatively affect their financial well-being. it is worth highlighting that limited access to borrowing options from financial institutions often compels households to seek alternative sources of financing (xiao & tao, 2021), such as borrowing from individuals in their social networks, thereby hindering them from achieving long-term financial stability. higher financial well-being means a better quality of life and less stress related to financial concerns. the literature describes various consequences of financial well-being and emphasizes the negative societal impacts when a significant percentage of households face financial issues (brüggen et al., 2017). lower levels of financial well-being result in financial vulnerability (beckmann & kiesl-reiter, 2023), which, in turn, can hinder households from accessing financial credit, compelling them to borrow money from surrounding communities, such as family and friends. inflation and gross domestic product (gdp) per capita are other factors that can affect household financial management and all kinds of borrowing (nourallah et al., 2024). taken together, it can be hypothesized that the use of debit cards, saving behavior, educational background, the opportunity to borrow from a formal financial institution, financial well-being, inflation, and gdp all affect the behavior of borrowing from family and friends. method in this study, we extract annual data for a sample of 24 eu member states for 2014, 2017, and 2021 to examine factors that might affect diab et al. 51 borrowing from family or friends. due to the lack of data, luxembourg, slovenia, and the slovak republic are excluded. the selected time frame is determined by data availability. we use data from the world bank’s global findex database on borrowing from family or friends (bff), savings (saving), and borrowing from a formal financial institution (bfi) (demirgüç-kunt et al., 2022). inflation (inf) and the growth rate of gross domestic product per capita (gdp) are taken from the world development indicators database (the world bank, 2024). moreover, we use data on the total number of debit (debit) and credit (credit) cards from the financial access survey (international monetary fund, 2024). since debit and credit are expressed in billions, whereas the remaining variables are expressed as percentages, these two variables are standardized by subtracting their respective means and dividing by their standard deviations to ensure comparability. information about the adult participation rate in learning (apl) and financial well-being (fwb) is based on data from eurostat (n.d.). we further incorporate data on the participation rate of youth and adults in formal and non-formal education and training (educ) from the unesco institute for statistics (n.d.). in line with nourallah et al. (2024), the variable saving is computed as an average of the percentages of respondents who save money for any reason, those who save for old age, and those who save at any financial institution. calculating the average of these three distinct percentages provides a holistic measure of saving behavior. moreover, fwb is assessed through two variables: the average rating of satisfaction, and the distribution of the population aged 18 and over by health status (very good). the appendix reports all employed variables along with their definitions and the sources from which we extract them. at the top, we find the dependent variable, i.e., borrowing from family and friends, followed by the independent variables and finally the two control variables, i.e., inflation and gdp. to study the determinants of borrowing from friends or family, we employ the following panel regression model: bffi,t = β0 + β1*debiti,t + β2*savingi,t + β3*educi,t + β4*bfii,t + β5*fwbi,t + β6*infi,t + β7*gdpi,t + εi,t where bffi,t represents the percentage of borrowing from family and friends, debiti,t the standardized total number of debit cards, savingi,t the percentage of respondents saving money, educi,t the participation rate of youths and adults in education and training, bfii,t borrowing from a formal financial institution, fwbi,t financial well-being, infi,t the inflation rate, and gdpi,t the gross domestic product per capita, all for country i across time t. finally, εi,t is the stochastic error term. the hausman test does not reject the null hypothesis at the 5% significance level (prob = 0.5525 > 0.05), indicating that the random-effects model is appropriate for capturing unobserved heterogeneity. ignoring this heterogeneity could lead to omitted variable bias. it is also worth noting that the random-effects model is an appropriate specification because the data are drawn from a survey with a randomly selected sample and because the data for some variables remain constant over time. hence, employing a fixed-effects model may lead to collinearity issues. moreover, we conduct the woolridge test for serial correlation, failing to reject the null hypothesis of no autocorrelation at the 5% significance level (prob = 0.1895 > 0.05). we further conduct the ramsey reset test for omitted variables in the random-effects model. the results indicate a failure to reject the null hypothesis of no omitted variables at the 5% significance level (prob = 0.3267 > 0.05), showing that our model is well-specified. to ensure our results’ robustness and account for potential model misspecifications, we employ ordinary least squares (ols) and panel randomeffects approaches to estimate our model. ols provides a straightforward method for estimating relationships among variables, and panel randomeffects models offer additional advantages, such as controlling for unobservable characteristics that are individual-specific and time-invariant. by doing this, we aim to validate the consistency of our findings across different strategies and enhance the reliability of our conclusions. financial services review, 33(3) 52 results by analyzing the data on the percentage of people borrowing from family or friends in 2021 compared with 2014 across the 24 eu countries under study, we can observe the emergence of distinct trends in figure 1. the percentage of such borrowing remained stable between 2014 and 2021 in austria, belgium, cyprus, and the czech republic. an increase is observed in 2021 as opposed to 2014 in bulgaria, denmark, france, germany, greece, malta, and poland. the remaining countries experienced a decline in the borrowing percentage during the same period. figure 1. borrowed from family or friends (% age 15+) figure 1 compares the level of borrowing from family or friends (% age 15+) in 24 countries representing member states of the european union, i.e., all countries except luxembourg, slovenia, and the slovak republic, between 2014 and 2021. data sources: the global findex and eurostat databases. table 1 reports the mean, median, minimum, 25th percentile (first quartile), 75th percentile (third quartile), and maximum for all the variables. regarding borrowing from family and friends, the interval extends from 0.051 to 0.317, which indicates that all countries in the union have some informal borrowing, although to different extents. table 2 displays the results obtained from estimating our model using ols, i.e., model (1), and the panel random-effects approach, including one variable at the time, i.e., models (2)–(7). model (1) indicates that the total number of debit cards, saving behavior, and gdp per capita negatively influence the likelihood of borrowing from family or friends, while borrowing from a formal financial institution and the inflation rate positively affect the household’s informal borrowing. however, utilizing the ols approach might lead to erroneous causal inferences among the variables, as it does not adequately address individual-specific effects or time-invariant unobserved heterogeneity. hence, models (2)– (7) represent the estimated results of applying a panel random-effects methodology. when including the education variable alone, the results reveal a significant negative impact on borrowing from family or friends, as shown in model (2). yet, the significance is lost when other variables are included in the model. as demonstrated in models (2)–(7), the total number of debit cards, saving behavior, and the inflation rate play crucial roles in shaping households’ borrowing behavior within their social networks. the reliance on debit cards for financial control is in line with the conclusion of bachas et al. (2021). moreover, the fact that households with higher levels of savings tend to rely less on informal borrowing sources is in line with the argument of nourallah et al. (2024), who discussed the role of saving in improving households’ financial management. the findings regarding debit cards and saving emphasize the importance of financial prudence and preparedness in lowering the need for external financial assistance from social networks. at the same time, higher inflation rates may intensify financial strain on people, leading to higher levels diab et al. 53 of borrowing from family or friends as a coping mechanism. the result concerning the interconnectedness between macroeconomic conditions and personal financial habits echoes the argument of zinman (2015), who emphasized the importance of investigating such issues. table 1. descriptive statistics variables minimum first quartile median mean third quartile maximum bff 0.051 0.115 0.149 0.157 0.198 0.317 debit –0.801 –0.565 –0.422 2.230*10–9 0.253 3.738 credit –0.538 –0.444 –0.374 4.190*10–9 0.010 3.998 saving 0.140 0.351 0.497 0.480 0.609 0.784 educ 11.020 14.200 17.320 19.754 20.650 42.230 apl 0.011 0.057 0.088 0. 114 0.145 0.347 bfi 0.112 0.262 0.359 0.360 0.477 0.580 fwb 0.347 0.437 0.489 0.493 0.550 0.647 inf –0.014 0.005 0.014 0.0160 0.024 0.051 gdp –0.009 0.018 0.040 0.040 0.055 0.147 table 1 presents the descriptive statistics for the variables used in the models. borrowing from a formal financial institution (% age 15+), as denoted by bff, is the explained variable, debit, credit, saving, educ, apl, bfi, and fwb are the explanatory variables, while inf and gdp are the control variables. the debit and credit variables are standardized by subtracting their means and dividing by their standard deviations to ensure comparability to other variables measured in percentages. the data are drawn from three waves of panel data spanning the years 2014, 2017, and 2021, sourced from the global findex (demirgüç-kunt et al., 2022) and eurostat (n.d.) databases. the dataset encompasses all eu member states, excluding luxembourg, slovenia, and the slovak republic. to increase the robustness of our findings and ensure the reliability of our results, we conduct sensitivity tests. in the eu context, it is plausible to argue that, due to the proliferation of educational platforms and diverse continuing education programs, people may engage in various forms of informal education. therefore, we re-estimate our model by replacing the educ variable with apl, which is a broader measure of education that captures the multifaceted nature of learning behavior among adults. apl accounts for ongoing formal and informal learning activities and continuous skill development, reflecting the overall educational exposure of households. as demonstrated in table 3, the results of this alternative specification align with those presented in table 2. notably, the number of debit cards, savings, and inflation are statistically significant across the ols and panel random-effects regressions. borrowing from a formal institution and gdp per capita are statistically significant across both models, providing reassurance regarding the robustness of these relationships. our findings underscore the significance of considering broader educational measures, such as apl, in capturing the diverse nature of adults’ learning behavior in the eu. financial services review, 33(3) 54 table 2. ols results variab les model (1) model (2) model (3) model (4) model (5) model (6) model (7) debit –0.019*** – 0.021*** –0.014** –0.015** –0.017** –0.018** –0.020*** (0.0058) (0.0080) (0.0066) (0.0069) (0.0073) (0.0076) (0.0072) saving –0.284*** – 0.161*** –0.154*** –0.189*** –0.186*** –0.259*** (0.0592) (0.0412) (0.0527) (0.0646) (0.0653) (0.0785) educ 0.00007 –0.00018 –0.00028 –0.00037 –0.00017 (0.0006) (0.0010) (0.0010) (0.0011) (0.0010) bfi 0.152* 0.0716 0.0849 0.145 (0.0771) (0.0771) (0.0816) (0.0880) fwb –0.016 –0.0533 –0.0294 (0.0936) (0.0964) (0.0916) inf 1.122** 0.946* (0.5120) (0.5070) gdp –0.410** –0.363 (0.1940) (0.2280) constant 0.258*** 0.174*** 0.245*** 0.245*** 0.240*** 0.263*** 0.262*** (0.0387) (0.0110) (0.0204) (0.0217) (0.0225) (0.0473) (0.0446) observatio ns 71 71 71 71 71 71 71 r squared 0.392 0.110 0.322 0.321 0.331 0.333 0.390 number of countries 24 24 24 24 24 24 24 table 2 presents the estimation results. model (1) presents the results of regressing the model using ols with robust standard errors to eliminate heteroskedasticity. models (2)–(7) present the results of estimating the model using the panel random-effects model, including one variable at a time. the data are drawn from three waves of panel data spanning the years 2014, 2017, and 2021, sourced from the global findex (demirgüç-kunt et al., 2022) and eurostat (n.d.) databases. the dataset encompasses all eu member states, excluding luxembourg, slovenia, and the slovak republic. note: ***, **, and * denote the 1%, 5%, and 10% significance levels, respectively. standard errors are shown within parentheses. for model (1), robust standard errors are reported. the debit variable is standardized by subtracting its mean and dividing by its standard deviation to ensure comparability to other variables measured in percentages. diab et al. 55 table 3. ols results (when replacing educ with apl) variables model (1) model (2) debit –0.021*** –0.022*** (0.0059) (0.0073) saving –0.242*** –0.205** (0.0610) (0.0827) apl –0.103 –0.153 (0.0913) (0.1180) bfi 0.158** 0.149* (0.0759) (0.0885) fwb –0.0157 –0.0273 (0.0900) (0.0924) inf 1.083** 0.880* (0.5151) (0.4970) gdp –0.479** –0.424* (0.2080) (0.2321) constant 0.255*** 0.253*** (0.0349) (0.0442) observations 71 71 r squared 0.401 0.396 number of countries 24 24 table 3 presents the estimation results of replicating the models in table 2 while replacing educ with apl. model (1) presents the results of regressing the model using ols with robust standard errors to eliminate heteroskedasticity. model (2) presents the results of estimating the model using the panel random-effects model, including all variables. the data are drawn from three waves of panel data spanning the years 2014, 2017, and 2021, sourced from the global findex (demirgüç-kunt et al., 2022) and eurostat (n.d.) databases. the dataset encompasses all eu member states, excluding luxembourg, slovenia, and the slovak republic. note: ***, **, and * denote the 1%, 5%, and 10% significance levels, respectively. standard errors are shown within parentheses. for model (1), robust standard errors are reported. the debit variable is standardized by subtracting its mean and dividing by its standard deviation to ensure comparability with other variables measured in percentages. although debit and credit cards both involve transactions, they represent two distinct financial behaviors. to explore how borrowing from family or friends may be influenced by these different behaviors, we replicate the test presented in table 3 and include the total number of credit cards. the presence of the debit and credit variables allows us to examine the interplay among various financial instruments. including the total number of credit cards in the model adds the element of debit cards and acts as a robustness check regarding our findings on the significance of the debit variable. our findings provide evidence that the relationship between the total number of debit cards and borrowing from family or friends holds even when considering a broader range of financial behavior. the additional results reveal that debit, saving, and inf remain statistically significant across both models, as shown in table 4, aligning with those presented in table 3. additionally, gdp is significant across both models. however, unlike the results presented in table 3, bfi no longer exhibits statistical significance. financial services review, 33(3) 56 table 4. ols results (when replacing educ with apl and including credit in the models) variables model (1) model (2) debit –0.023** –0.025** (0.0092) (0.0118) credit 0.0078 0.0084 (0.0089) (0.0112) saving –0.250*** –0.216** (0.0658) (0.0870) apl –0.069 –0.117 (0.0929) (0.1240) bfi 0.136 0.131 (0.0833) (0.0934) fwb 0.0016 –0.0104 (0.0945) (0.0939) inf 1.074** 0.913* (0.5261) (0.5069) gdp –0.492** –0.457* (0.2179) (0.2390) constant 0.253*** 0.252*** (0.0355) (0.0446) observations 68 68 r squared 0.388 0.384 number of countries 23 23 table 4 presents the estimation results of replacing educ with apl and including credit in the models. model (1) presents the results of regressing the model using ols with robust standard errors to eliminate heteroskedasticity. model (2) presents the results of estimating the model using the panel random-effects model, including all variables. the data are drawn from three waves of panel data spanning the years 2014, 2017, and 2021, sourced from the global findex (demirgüç-kunt et al., 2022) and eurostat (n.d.) databases. the dataset encompasses all eu member states, excluding luxembourg, slovenia, and the slovak republic. by including the credit variable, we lost three observations due to the absence of data for france. note: ***, **, and * denote the 1%, 5%, and 10% significance levels, respectively. standard errors are shown within parentheses. for model (1), robust standard errors are reported. the debit and credit variables are standardized by subtracting their means and dividing by their standard deviations to ensure comparability with other variables measured in percentages. conclusion, policy recommendations, and suggestions for future research households may depend on informal borrowing due to a low level of financial resilience (lusardi et al., 2021), including inadequate safety nets. nevertheless, such behavior can lead to serious financial problems and personal conflicts (karaivanov & kessler, 2018). applying the selfcontrol theory and using balanced panel data from 2014, 2017, and 2021, this study focuses on various factors that can affect the behavior of borrowing from family and friends and underscores the pivotal role of debit card use and saving behavior in reducing the tendency for such borrowing. despite the development of many advanced fintech payment tools, debit cards are still popular. according to the european central bank (2024), card payments accounted for 54% of all non-cash transactions in the first half of 2023. this highlights the widespread use of debit and credit cards in the eu. moreover, many diab et al. 57 contemporary fintech solutions such as mobile wallets require that account holders charge their accounts by transferring money via debit cards. due to the functionality of debit cards (which do not allow holders to spend more than the money they actually possess), households can control their spending. the work of bachas et al. (2021), which emphasizes the positive effect of using debit cards in increasing overall saving and controlling consumption, can shed light on the present findings. therefore, we conjecture that the use of debit cards could improve households’ financial resilience and their capability to deal with unexpected financial shocks. policymakers should therefore promote the adoption and usage of debit cards over credit cards due to the role the former has in decreasing informal borrowing. policymakers should also raise awareness of the benefits of formal borrowing such as consumer protections, ensure the availability of affordable formal borrowing options for low-income households, promote access to bank loans and microfinance, and strengthen social safety nets to act as a buffer against financial shocks and hardships. in such a financial landscape, financial robo-advisors are promising technology tools because they can help households conduct sound financial management at a reasonable cost and without time or place constraints (d’acunto & rossi, 2023). the finding related to the role of savings in reducing informal borrowing aligns with arguments presented by, for example, browning and lusardi (1996), tufano (2009), and despard et al. (2020). they have argued that a reason for saving and establishing emergency funds is to mitigate unexpected events, emphasized the importance of the precautionary principle in dealing with contingencies, and suggested that saving and sound financial management can successfully stimulate the investing of money. this study identifies a significant effect of saving behavior in addressing undesirable informal borrowing. however, it is essential to go one step further and investigate what factors can prompt saving behavior, particularly among low-income households. also, when focusing on saving behavior, financial robo-advisors can be useful for households (nourallah et al., 2023). we also encourage future research to explore interventions to enhance financial decisionmaking and sustainable saving practices and consider how to improve financial well-being across diverse socioeconomic contexts. to address the potential impact of general knowledge on borrowing behavior, this study utilizes a measure based on the percentage of households participating in formal or non-formal education or training within the last 12 months. the results reveal an insignificant effect of education on informal borrowing behavior. this suggests that the education provided to the public may not effectively enhance households’ financial knowledge or address money-related issues. hence, and in line with lusardi et al. (2021), we recommend improving financial literacy education and embedding more financial knowledge into formal education curricula in order to focus on money management skills such as financial planning and debt management. this would seem to be increasingly important, as the search for effective retirement savings strategies has more than doubled over the past two decades (lusardi & mitchell, 2023). related to this, we suggest future financial literacy education studies around the world. references asebedo, s. d., wilmarth, m. j., seay, m. c., archuleta, k., brase, g. l., & macdonald, m. (2019). personality and saving behavior among older adults. journal of consumer affairs, 53(2), 488–519. bachas, p., gertler, p., higgins, s., & seira, e. (2021). how debit cards enable the poor to save more. the journal of finance, 76(4), 1913– 1957. https://doi.org/10.1111/jofi.13021 beckmann, e., & kiesl-reiter, s. (2023). financial literacy and financial wellbeing: evidence from eastern europe in a high inflation environment. journal of financial literacy and wellbeing, 1(2), 263–367. https://doi.org/10.1017/flw.2023.12 blanc, l. j., porpiglia, a., teppa, f., zhu, j., & ziegelmeyer, m. (2015). household saving behaviour and credit constraints in the euro area (ecb working paper no. 1790). european central bank. browning, m., & lusardi, a. (1996). household saving: micro theories and micro facts. https://doi.org/10.1111/jofi.13021 https://doi.org/10.1017/flw.2023.12 financial services review, 33(3) 58 journal of economic literature, 34(4), 1797–1855. brüggen, e. c., hogreve, j., holmlund, m., kabadayi, s., & löfgren, m. (2017). financial wellbeing: a conceptualization and research agenda. journal of business research, 79, 228–237. https://doi.org/10.1016/j.jbusres.2017.03.01 3 d’acunto, f., & rossi, a. g. (2023). robo-advice: transforming households into rational economic agents. annual review of financial economics, 15(1), 543–563. https://doi.org/10.1146/annurev-financial110921-013217 demirgüç-kunt, a., klapper, l., singer, d., & ansar, s. (2022). the global findex database 2021. the world bank. https://doi.org/10.1596/978-1-4648-1897-4 despard, m. r., friedline, t., & martin-west, s. (2020). why do households lack emergency savings? the role of financial capability. journal of family and economic issues, 41, 542–557. eurostat. (n.d.). eurostat database. retrieved from: https://ec.europa.eu/eurostat/web/main/data/ database hamid, f. s., loke, y. j., & chin, p. n. (2023). determinants of financial resilience: insights from an emerging economy. journal of social and economic development, 25(2), 479–499. international monetary fund. (2024). financial access survey database. https://data.imf.org/?sk=e5dcab7e-a5ca4892-a6ea-598b5463a34c kaiser, t., lusardi, a., menkhoff, l., & urban, c. (2022). financial education affects financial knowledge and downstream behaviors. journal of financial economics, 145(2), 255–272. https://doi.org/10.1016/j.jfineco.2021.09.022 karaivanov, a., & kessler, a. (2018). (dis)advantages of informal loans: theory and evidence. european economic review, 102(c), 100–128. lee, s., & persson, p. (2016). financing from family and friends. review of financial studies, 29(9), 2341–2386. https://doi.org/10.1093/rfs/hhw031 lusardi, a., hasler, a., & yakoboski, p. j. (2021). building up financial literacy and financial resilience. mind & society, 20(2), 181–187. https://doi.org/10.1007/s11299-020-00246-0 lusardi, a., & mitchell, o. s. (2023). the importance of financial literacy: opening a new field. journal of economic perspectives, 37(4), 137–154. meyll, t., & walter, a. (2019). tapping and waving to debt: mobile payments and credit card behavior. finance research letters, 28, 381– 387. https://doi.org/10.1016/j.frl.2018.06.009 nokulunga, m., & klara, m. (2023). determinants of using formal vs informal financial sector in brics group. finance research letters, 55(b), 103956. https://doi.org/10.1016/j.frl.2023.103956 nourallah, m., & öhman, p. (2021). impact of advanced technologies on consumer finance and retail investment. in r. ho, a. h. h. ng, & m. nourallah (eds.), impact of globalization and advanced technologies on online business models (pp. 1–15). igi global. https://doi.org/10.4018/978-1-79987603-8.ch001 nourallah, m., öhman, p., & amin, m. (2023). no trust, no use: how young retail investors build initial trust in financial roboadvisors. journal of financial reporting and accounting, 21(1), 60–82. nourallah, m., öhman, p., & hamati, s. (2024). financial technology and financial capability: study of the european union. global finance journal, 62, 101008. https://doi.org/10.1016/j.gfj.2024.101008 stango, v., & zinman, j. (2023). we are all behavioural, more, or less: a taxonomy of consumer decision-making. the review of economic studies, 90(3), 1470–1498. https://doi.org/10.1093/restud/rdac055 thaler, r. h., & shefrin, h. m. (1981). an economic theory of self-control. journal of political economy, 89(2), 392–406. https://doi.org/10.1086/260971 the european central bank. (2024). payments statistics: first half of 2023. european central bank. https://www.ecb.europa.eu/press/stats/payse c/html/ecb.pis2023~b28d791ed8.en.html https://doi.org/10.1016/j.jbusres.2017.03.013 https://doi.org/10.1016/j.jbusres.2017.03.013 https://doi.org/10.1146/annurev-financial-110921-013217 https://doi.org/10.1146/annurev-financial-110921-013217 https://doi.org/10.1596/978-1-4648-1897-4 https://ec.europa.eu/eurostat/web/main/data/database https://ec.europa.eu/eurostat/web/main/data/database https://data.imf.org/?sk=e5dcab7e-a5ca-4892-a6ea-598b5463a34c https://data.imf.org/?sk=e5dcab7e-a5ca-4892-a6ea-598b5463a34c https://doi.org/10.1016/j.jfineco.2021.09.022 https://doi.org/10.1093/rfs/hhw031 https://doi.org/10.1007/s11299-020-00246-0 https://doi.org/10.1016/j.frl.2018.06.009 https://doi.org/10.1016/j.frl.2023.103956 https://doi.org/10.1016/j.gfj.2024.101008 https://doi.org/10.1093/restud/rdac055 https://doi.org/10.1086/260971 https://www.ecb.europa.eu/press/stats/paysec/html/ecb.pis2023~b28d791ed8.en.html https://www.ecb.europa.eu/press/stats/paysec/html/ecb.pis2023~b28d791ed8.en.html diab et al. 59 the world bank. (2024). world development indicators database. the world bank. retrieved from https://databank.worldbank.org/source/world -development-indicators the world bank. (n.d.). metadata glossary. the world bank. retrieved from https://databank.worldbank.org/metadataglo ssary/world-developmentindicators/series/fp.cpi.totl.zg tufano, p. (2009). consumer finance. annual review of financial economics, 1(1), 227–247. https://doi.org/10.1146/annurev.financial.05 0808.114457 unesco institute for statistics (uis). (n.d.). uis database. retrieved from: http://data.uis.unesco.org/ xiao, j. j., & tao, c. (2021). consumer finance/household finance: the definition and scope. china finance review international, 11(1), 1–25. zinman, j. (2015). household debt: facts, puzzles, theories, and policies. annual review of economics, 7(1), 251–276. https://doi.org/10.1146/annurev-economics080614-115640 https://databank.worldbank.org/source/world-development-indicators https://databank.worldbank.org/source/world-development-indicators https://databank.worldbank.org/metadataglossary/world-development-indicators/series/fp.cpi.totl.zg https://databank.worldbank.org/metadataglossary/world-development-indicators/series/fp.cpi.totl.zg https://databank.worldbank.org/metadataglossary/world-development-indicators/series/fp.cpi.totl.zg https://doi.org/10.1146/annurev.financial.050808.114457 https://doi.org/10.1146/annurev.financial.050808.114457 http://data.uis.unesco.org/ https://doi.org/10.1146/annurev-economics-080614-115640 https://doi.org/10.1146/annurev-economics-080614-115640 financial services review, 33(3) 60 appendix: definition of variables borrowed from family or friends (% age 15+) “the percentage of respondents who report borrowing any money from family, relatives, or friends in the past year.” (demirgüç-kunt et al., 2022) the global findex database saved any money (% age 15+) “the percentage of respondents who report personally saving or setting aside any money for any reason and using any mode of saving in the past year.” (demirgüç-kunt et al., 2022) saved for old age (% age 15+) “the percentage of respondents who report saving or setting aside any money in the past year for old age.” (demirgüç-kunt et al., 2022) saved at a financial institution (% age 15+) “the percentage of respondents who report saving or setting aside any money at a bank or another type of financial institution in the past year.” (demirgüç-kunt et al., 2022) borrowed from a formal financial institution (% age 15+) the percentage of respondents who report borrowing any money from a bank or another type of financial institution or using a credit card in the past year. use of financial services, number of cards, debit cards the total number of debit cards in circulation (excluding expired and withdrawn cards) of all financial institutions in the reporting jurisdiction. (imf, n.d.) international monetary fund – financial access survey use of financial services, number of cards, credit cards the total number of credit cards in circulation (excluding expired and withdrawn cards) of all financial institutions in the reporting jurisdiction. (imf, n.d.) education “percentage of youth and adults in a given age range (e.g. 15–24 years, 25–64 years, et cetera) participating in formal or nonformal education or training in a given time period (e.g. last 12 months).” (uis, n.d.) unesco institute for statistics adult participation rate in learning the adult participation rate in learning covers participation in formal and non-formal education and training. it encompasses all learning activities undertaken with the aim of improving knowledge, skills, and competences within the personal, civic, social, or employment-related domains. eurostat financial well-being average rating of satisfaction overall life satisfaction. distribution of population aged 18 and over by health status very good distribution of population aged 18 to 64 years who responded that their health status is very good. (eurostat, n.d.) inflation, consumer prices (annual %) “inflation as measured by the consumer price index reflects the annual percentage change in the cost to the average consumer of acquiring a basket of goods and services that may be fixed or changed annually.” (the world bank, n.d.) world bank – world development indicators gross domestic product (gdp) per capita growth “annual percentage growth rate of gdp per capita based on constant local currency. gdp per capita is gross domestic product divided by midyear population.” (world bank, n.d.) pii: s1057-0810(01)00084-1 ��� ������� � ����� ��� ��� ����� � � �������� �������� � �� ���� �� ����� ����� ���� ����� ����� ����� ����������� �� � � �� ��� � ������ �� ���� �� ��� �������� ���� � ��� ���� ������ �� ���� ��� �������� �� �!"#� �$� ������ %���� �& $��� ��� '���� �(# ���� � �� ��� )����� ���� �� �""#� �$� ������� �! "#�� $%%%& �������� � ������� ���� $' "#�� $%%�& ����� �� (% )� ���� $%%� �������� ��� �#����� �� ��� � #�� �� � �*��� � �������� ��� ����� �� � � �� ���# �� � �� ���� ��� � ��+ �� �� ���# �� �� ����� ��� �� ,�� � �� � ���� �� �!!�#���� �� .� �#��� /� � ���� ����� � � ��� � ������ � ���� �� � ��� �+�� ���� �� ���� �� ��� �������� �� ���# �� � � � � ��������0� ������� � ��1� �� ���# �� � � � ��� �� ��� �� ����� ��� � �������� �� ���� ��������� ��� ��� �� ���# � � � ���� �� ��� ��� �� �� �������� �� ���# �� ����� ��� ��� � ��� �+�� ���� ��� ����� � �������� �� ���# �� �� ��� ���������� �� + �� �� �#���� �#� ���� ����� �� � ����� ���� �� ���� �� ��� �� ������� �������� �������� ����� ��� � � �� �*�� � �� �� � �������� �� ��� � �� ���# �� � � $%%� 2������� ���� �� 3 �� 4�� ���� � ��������� *��� ���+ �#���� �� .� �#��� /� � ���& 5���� & ���� � ���� ������ � ����� � ���� �� ������ ����� ���� ���� ��#���� ������ ���#�� �� ��������6 ��� ����� �� ����� ��� �� � � � �����#�� ���� ��� 7���� ������ ���#�� � ���� �������� �������� �� ��� �� � � �� ����� �� �+ �� �� ���# �� ����� �� �� �� � ���8#� � ��� ��� �� �#����� 2������� ��� ����� �� ����� ��� � ���� ������ �� ��� ��#��� �� �� ����� � ���� �#� � �� � ���8#��� �� �� ������ ���#�� � ��� �� � � ��� ������ �� ����� �� ���� ��� ����� ��� � ��� ��� �� �� �� �� �#�� ��� � � � � � ������� � ���� � �� ��������� #�� �� ����� �� ���� � �� �� �������� � �� �� �������� ��� ����� �� ����� ��� � ��� �#�� � ����� � ��#��� �� �� ����� � ����� �� ���� ������� � � ���9� ���� �������� �� ���# � � ��� ��� � :#� �� � �������� ��� ����� �� � � �� ���# �� ������ � �#�� ��� �� /� � ���� �������� ����� �% ;$%%�< =-6-= �%-'>%?�%@%�@a 6 ��� ��� �� �� � $%%� 2������� ���� �� 3 �� 4�� ���� � ��������� 533 � � % ' > % ? � % ; % � < % % % ? = > � � .������� �� � �# ���� ���� ��>-%=>$?%>b�b(& ��* ��>-%=>$?%>b(!'� �,�� � ���������+ ������ c# ����# ;���� ����� <� ���������� ��c�� �������� ;�� ����� ��<� � ���� ��>!?->b$b>?'?'� ��� �� ��� �� ��� �*��� � � �������� ��� ����� �� � � �� ���# �� � �� ���� ��� � �������� ��� ����� �� ����� ��� � �� ��>����� /��� � �� �!!�#���� �� .� �#��� /� � ��� ;�./< �� �� ��������� ��� #��� ���� �� � �� ��� �� ����� � ����� �� ������� � �� � �� ��� #��� �� �� ����� ��� � � � �������� ��� ����� �� ���� � �� �#�� ��� �� ���������� �� �!!�./ �������� �� ����� # � � � �*�� � �������� � �����+� ����� �� �� ���� ����� ��� ��� ����� �� � ��� �� ���# �� �� ��� �������� �� ��� � �� ���#> �� �� ����>����� �� ����# �� ��� �� � � �������� �� ��������� ���� �� ���� �� � �� � #�� ���� ���� ���� ���� �� �� ��+ �� �� �+ �� ��� ��� � � ����� �� �������� ��1�� ���� �� ��� �� �� �� ���# �� � ����� ��+ �� �� ���# �� ��� � ����� �� �������� ������� ��� �� ��� �� ���# �� ���# � � � �� � ������ ��� �� �# �� ���#�� �� � ��� ��� ��#�� � �� ������ � � ������ ��� ����� � � � #�� �� # ������ � �� ��� �� ��� �������� ��� ����� �� � �������� ��� ����� �� ����� ��� � � � �� ���# �� � � �#�� ��� �� ��� ����� �� �� � +��0� ������ �� �� ��� � �������� ���������� �� ���������� � �� ����>����� �� ����# �� � � �� �*�� � �� �� ��� � �� �������� �� ��������� ��� ���� � ����� � �������� ��� ����� �� � � �� ���# �� �� ��� 3 ���� �� � �� � �� ��� ����� � �� �������� �� ���# �� � ��� �� � ���� � +��� ������� � �������� �� ��� �� � � �� � �� � ��������� �������� �� ���# �� �� �� � ���� ������� �� � ������� � �� � �� �� ������� ���� �� �� � � � ���#� � � �����>��� ���� � � #� �� � �����#��� � ���� �� ����� ������ � ��������� �� ������� ���������� �� ����������� �� ���� ��� � 4 ���� � #�� �� d��� ;�!!'< #��� ��� ��������� � � � ���9� ���� � �� ���� � � =%�;1< �� ���# �� ��� � �� � �� � ���� �� �!!�#���� �� .� �#��� /� � ���� ��� ���#� � � ���� � �� �������� ���� � �� ���� ��� �� ����� � ��#�� �* ���� � � � =%�;1< ��� ��� #��� �#�� �� �������� �� ��� �� ��� �� � � �������� ��� ���# �� �� 3 ���� �� � �� + �� �� ���� � ���� ��� ���� �� �� ��� �� ���# �� � � � ���1 ����� ��� ����� � ����� � � ����� �� ���1� �� d���0� ;�!!'< ���#� � ������� � ���1 ����� �� �#���� � ������� � #�� �� e#� � � f�g� �� ;�!!b<� ,�� � �� � ���� �� �!!$ �#���� �� .� �#��� /� � ���� �� ���#� � ���� e#� � � f�g� ��0� ;�!!b< �#� �>����� � � ������ � ���� � �� ���1 ������ � �����#��� ���� ������� ���# � �� ��+ �� �� ���# �� �# ��� 3 ���� �� � ��#������� �� � ����� ��#�� �� ������ � � ���� �1����� ���#�� �� � �� ��� ���� ����1 �� ����� �� � ��� � ���� ����� ������ � ��+ �� �� ���# �� �# ��� ) ��� + �� �� �� e#� � � f�g� �� ;�!!b< � ���� � �� �# �� � ������ �� ����� �� �������� � ������� h� � �� =%�;1< ��� � ��� �� ��������� �� �� � �������� � � �������� �� ���# �� �� 4� ��#�� �� �������� �� � ��8#���� � �� ���# � � � =%�;1< ��� � �� � ������ � �� ��� #�#���� � �� ���� �� � �������� �� ��� ��������� �� �������� ���� �� �� � ����� ��� �� ���� ��������0� �� ���# �� #� � � �����+� ����� ��� �� ������� 4������ � � 2�� � � ���5����� ;�!!=<� ?(i �� ��� ����� � � ������� =%�;1< ��� � ������� �� ��� � �� ��� � �� ���# �� ����� � ���� �������� ���# �� ���� ��� �� �������� �� ���# �� �#�� �� � ���+ ����� � �� ���# �� � j� ������� �� ��� ������� �� ��� � =%�;1< ��� � �� �� ��� �� �� ���# �� ����� �� �� ��� ���� 4 =%�;1< �#���� �� �#� �� �� ������� h�*� � � � 7������� ;�!!!< �� ���� =%% �� ��� .������ �� +��� � ���� �� � � ������ � �� � � �� #���� �� ��������� ��1� � =b )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" ������ �� ��� �� ���# �� � � � ���� =%�;1< ��� �� �����+������ �� ����� ��� �� +��� ��1� � ������ �� ��� �� ���# �� � � ������� ���� (?i � �!!? � -=i � �!!!� /#� �������� �� ����� ��� �� +��� ����� ��� � � ��#�� �� ��� � �� ���# �� � ����� � # ��� ��� ���� �� ������#� ���� � '-i� 3 ���� �� � �� + �� �� � ���� � � ���� ���� � � �������� ��� ����� �� �� ��� ��� ����� ��� �� +��� ����� � � �� ?%i6�%%i �� ���� �������� ��������� ��� ����� � ��� � ������� ���� (-i � �!! � =$i � �!!!� ��� �# ���� � ���# � �� � ������ � �������� ��� ����� �� �� �� � ������ � ��� �� ������ � � ���� �� � ��� �� :�*����� �� /� ����� ��� ��� ����� �� ��� ���� ���1�� �� �� � �� ='i �� ��� ��� ���� �� 1 ��� ���� ���� ������ � ������ ���� �� �* $ ������ d��� ;�!!'< ��� � �����+����� �� �������� ��� ����� �� � � �� ���# �� ������ � �� ����� ��� � ����� ���� � �������� �� ��� � �� ���# �� ��� #��� 3 � ���� �� �# ��� ���#��� �� �� �*�� � �� �� � �������� �� �� �� � ��������� ��� #�� ��� ���1���� ������ �� ��� � � ��� � �� ���# � ���� �� � ���#� �� ��� ��� #��� ��� � � �� �#� � #�� �� � ���1 � ��� ���#� ���� ������� �� �� � � �� ���� �� ���� �� ��� �������� ��� ����� �� � � �� ���# �� ������ ����� � �� �������� �� ��� � �� ���# �� ��� #��� )#� � #�� ���� � ���9�� � ���� �� �� ��� #�� � � ��#��� � d���0� ;�!!'< � #�� �� ���������� � �� ����> ����� �� ����# � � � �� ����� �� ��� ��� #�� � �������� ��� ����� �� � � �� ���# �� ������� 4 ���� ����� �� k���� � /���� �� � � �����#�� ;�!!?< #��� �� � ���� �� 4���� �!!( .#��� 5��#�� �� �#���� � � � � �#���� �� 2������� k� �+ � �#������ ;.5�< � � ���9� ��� ���1��� #�� =%�;1< ��� �� ����� ������� ��� + �� �� � ���� � �� b-i �� �������� ���1��� ��� ����� � � =%�;1< ��� �� ���� �����+������ � �� (bi �� ���1��� �� � ������ � ����� �� a�-�%%% �� ���� ��� ����� �� ������� ?�i �� ���1��� �� � ������ � ����� �� � ���� a'-�%%% ��� ����� �� /#� �������� �� + �� �� ���� �� ��� ����� �� � � =%�;1< ��� � ������� �� �������� �� ��� � �*�� �� �������� �� �� � �� �� ��� � ���� � ������ � ����� �� ��� ����� �� ���������� 4�� ��� ���� �� �� ��� ��� � �� � �%% �� ���� ��� ����� � ��� ��8#���� � +�� �� /��� -�-%% � #���� �� � �� 3 �� ���� ���� �������� � ����� �� � ��� ��������� �� ��� ����� �� � �� ���# �� �� ��� ���# �� �� �*�� ���� � � �������� � ,�� � �� �������� � � �������� �� ���# �� �� � � /��� -�-%%� 5��1� ;�!!-< ����# �� � ������� �� ��� � �� ���# �� �� � ��� ���� ��� � #����� )� �� =�b?' ������� �� �� �=i �� ��� � ��� � �� ��� � �� ���# �� � � ���� ��� ��� ���8#� �� ��� � �� ���# �� �� � ��� -%i �� ���� ������ � ��� ����� �� ���# �� � ���������� �������� �� $-i � � '-i� ������ ������ ������ � �������� ��� ����� �� � ��� �� ��� �� ����� ��� � 5��1�0� ;�!!-< ���#� � � ���� � �� ��� ����� �� �� ��� ����� ������� �� �� � +�� ��9�� ���� ���#� �� �#�� �� � 4 �����0� ;�!!$< + �� � �� ���� �� ���� ��� ����� �� � ������� +���� �������� ����#�� �#�� +��� ��� �� �� ���� � ����� �� �� �+ � ������� � �� ���� �� ���� ���������� ��������� ������ +��� � � � ���� ������ �������� #� ����� ���������� �� �� �� ��������� � ��8#����� � ��� ��� ��� � ��� � ����� ������� � � � �!!? �#���� �� �#� �� �� �� 5��+ ����� �@=%�;1< .�# ��� �� 4������ ;�!!?<� � �� ������� �� ������ +��� ��#�� ���� � ����� ��� ����� �� �� �� 5��1�0� ;�!!-< � ������� �������� ����� �� �� ���� �� ����� � �������� �� ���# �� � ���#� � � ���� � ���� � �� �� � � �%i �� ��>��#�� � � � ������ � �������� �� ���# �� � �� (?�!bi� �������� �� + �� �� � ���� � �� ��� �������� )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" =' �� ��� � ���� � �� �%i ����� ���#� � � �������� �� ���# �� � �� ��� � � �� ������ � ��� �+�� �� � �� �*��� �� �� .� ���# �� � � ������ �� '�(bi �� ��� � (%i � � =%i �� �� �# �������� �� !�!'i �� ��� � =%i � � -%i �� ��� /#� �������� ���� �� ����� �� �� � �������� � �� ���# �� � � ���� �� �� �� ��� ���� ��#�� �� � ���#� �� � ������� � �� �� �� 4 �� �� � ��� �*��� � �� �� �� �� �� �������� �� ��� � �� � � �������� ��������� ���� �� � �� ��� � � ������ �� ���# �� ���� �� ���# �� � � ����� � �� �� �� ��*��#� �� ��� � �� �+ � 4� ��#�� ��������� ���� ��� #���� ����� � ��� ����� ��� ���� � �� ���# � � � ���� �� �� ������� �#� � ���� �� � ���� ������ �� �� �� ��� �� � ���� ����� ��� �# � ������ �� �#�� � # �� ����� � 5��1� ;�!!-< �� ��#��� �� ���� �� � ���� � ������� �� � �� ���# �� �������� � ��� � � ���� �� �� �� ��� 4 ����� ���� ��� �� �� �� � ��� �+�� �� � ������� �������� �� ���# �� �� .� �������� � ������ �� �� �� �� ��������� ������ � �� #���� �������� ���� �� ���# �� �� 5��1� ;�!!-< �#���� � �� ��� ��� � ���� � � ll ���� ���� �00 ��������� ) � ���1 ��� �� ��� � #�� ������ ���� #�� � �� /��� -�-%% � ����# � �� �������� �� ��� � ����� ���� ��� /��� -�-%% ����1� ��� �� � �� �� ���# �� � ���� �� �� �������� � � �� ���������� �������� �� � �� �������� �� ���# �� ��� �� ��������� �� ���� �� m#� �������� �� ��� � �� ���# �� �� /�� � � � ��� �� �������� ��� �� ��8#���� � ��1� � ���� �� �� �� ���# �� � ����� � ���� ��� �� � �������� � �� �� �� ���� ���#�� ��� �� ������� �������� ����� ��� ;4f5< �� �� ������ ����� �� �� ��������� �*����� �� 4f5 �� �� � ������ ����� �� �� ��������� �� � ���# ��#�� � �� ��� � ���� �� � �������� � �� �� �� h� � �� � �� ����� � ������ ����� �� �� �������� �� � �������� ��� �� � ���� � -i �� �� �� �� ����� � �� ��� � ���� �� � ��� �� � ��*�� � ���� ������ 3� � ��� ��� ��� ��� ���� � �� ���# �� � ����� � ������ �� � �� � �������� � ��� ��8#����� �� �� �������� �� ��� � ����� ��� ����#�� �� ���� �� /��� -�-%% ���� �� ����� � ��� 3 ���� �� � ���� =%�;1< ��� � ���� � ���+ ����� � ����� � � ����� ��������� ��� ���� � ���� � ������ �� ��� �� ���# �� #� � �� ��� �� =����� � �� �� 3 �� �� ��� #� .���� ����� �� ���# �� � ��� ������ ��� # ���� �� � � � ��������� �� ��������� ���� ���� � �� =%�;1< ��� �� �� �� ��� � 3 �#�� ���� ��� �� �������� �������� �� ���# �� ����� � � �� �� � :� � �� �������� �� ��� � ����� ��� ��� ����#�� �� #�� � �� /��� -�-%% �� ���# �� � ��8#���� � ���� � �������� � �� �� � � � ������ �� ��� ���+ ����� � �� ���# �� �� 3 � � ��� � ����� ���� ������ �� ���� #�� � �����#�� �� �� �� ��� �� ��� �� �� ���� �� �!!�#���� �� .� �#��� /� � ���� 2������ � #���� �� �� ����� � �� �������� �� ��� � ���� ������� ��*�� ���#� �� 4 ����� ;�!!$< ��# � � ���� ��� ������� �� �� ��� �������� �� ��� � � � ��� ����� �� �� � �� ��� �!?? .5�� ��1������ 5��1� ;�!!-< ��# � ������� ���#� � #�� � �� � �������� �� �� 3 �� 5��1� � � 5� ���� ;�!!-< #��� �� � ���� =( +��� � � ���� ��# � ����� �� �� �������� ��� ����� �� � ������� �� � �� �*�� � �� �� �������� �� ��� �� 3 ���� �� � ��� ��# � ���1 ����� �� �� ��� ����� �� � ������� �� � �� �� ��� � ������ �������� �#�1�� 5� ����� � � 7����* ;�!!=< ��# � � ���� �� ���� �� ��� �� ��� � � � ��� ����� �� � 3 ���#�� �� � �� ��#�� �� ���� � #�� �� ��� �� �� � �� � � ����#� ��9�� +��� 4 �!!? =%�;1< ��� �#���� �� k#�1 .� �#� � � 3 �� ;�!!?< ������� ���� � ���� � � ���#� �� ���� ������ �#����� � �����*��� ��� =�%%% ���� �9� �� �� )� �� 'b+��� �� ����� ���� b=b ����� �� ��� ���� � � =%�;1< ��� � ����� ��#�� �� ���#� � � ���� � �� �� ������� =? )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" �������� ��� ����� �� �� � � =%�;1< ��� � �� ''i� ��� ������� �������� �������� �� � �� ��� ���� (i � ?i� 3 ���� �� � ?!i �� �� ��� � �������� �������� �� ��� � �� ���# �� �� ��� ��� ����� �������� �� �� �� � �� � -%i �� �� �� � ����� ���� �� �� ����� �� � ����� � � ���� /#� �������� =bi �� ����� �� � ����� � ��*��#� �� �� �� bi �� ������� ������ � =%�;1< ��� � � � �� ��������� ?'i �� ��� � �#������ �� k#�1 .� �#� � � 3 �� ;�!!?< ���� � ��� �������� ��� ��� ����� �� 3 ���� �� � !$i �� ��� � ����� �������� �� ��������� ) ��� + �� �� �� k#�1 .� �#� � � 3 �� ;�!!?< � ���� � �� ������ +���� �� � � ������� �� !�$$��������� � � � ��� ��� �� � ��+ �� �� �+ � � ��+ �� �� ���# �� ��� �� �������� +��� �� � ����� ���������� ������� � =�b�(� � � � ��� ��� =%�;1< ��� � �*��#������� 3 ���� �� � ���� �#���� �� �#�� �� � � ��� �������� + �� �� � �� ��� ����� �� � � � 1 ������� ���# � �������� ��� ����� �� ����� ��� �� ���� ����� ������� �� �� � +�� ��9�� 4� ��#�� ��� �#���� �������� � ���� � �� � � �� ���� � ����� �� ���# �� ����� ��� �� � ���� � ������� #� � � � ���� � � �������� ������� �� � �� ��� �������� �� ��� � �� �� � � �������� �������� �� ��� ������ � �� � ��� �� �� �� ��� ���� �� ��� � ��� �� �� #� � ���� �� ���� � ����� �� �� �� ���� �� �� ��� �� ��� ��� �� ��� ����� ����� ��� 3 ���� �� ���� �� �� ���#� �� � ��� � � � �������� � ��1� ���� � � 5��� ;�!!?< �*��� �� �� ���1 � � �� #� ����� � �� ���� � � �� ����� ��� �� �������� �� ���# �� � � ��+ �� �� ���# �� ��� � � �� ���� �� �������� � ��1� 5�� ������ ���� �� � �#� �� #�� � �������� � ��1� ����#�� �� ��� � � � ������� ����� � ��1 ��� ������ �� ��# � �� �� ������� �� #� ��� �� �������� � ��1 ��� ������� � �� ������� �� #� � �� ������� ����� � ��1 � ��*� �������� ���� � �� � ��1 � ��* � �� ��� ����� ���#��� �� ���1� � ���� ��� �� � ��� ��� � #�� ���� �� �� �� ���� �� �!!�#���� �� .� �#��� /� � ��� ��������� ���� �� /������ ������ ��� �� � ������� �� �� � �� � � �� ��� �� 3 ���� f������ � �� f���� �� �� �����#��� ��� �./ �� � ��� ��� ��� �� �#���� � � �� ����� �� � ������� � ����� �� � ��#������� ������� � ���� ����� �� + � ���� �� ��� ��� ��������� ����� ����� ��� ����� �� � � � �����#� ����������� ������ ���� ���� 3 ����� � ������� ���#�� � � ����� �� � ������ ���� ��� �� ��� ������ ��� ���# �� � �� ���#�� �� �� ���� �� ������ ���� ��� �� ��� �� �� �� �� � � ����� ��� �� �� ���#�� �� � �#�� �� ���� �� ������ ����#� � ��� �� ���� �� � �� � �#�� ����� ������ ����� �� #���� ��� �#����� �� � ������� � ���8#� � ������� ��� ���� ��� �� � ������ �� ��#������ ���� � � � ������� ���� ��� +�� ��� �� � ��������������>����� � � ���� �#� �>� ��� ��������� � ������ ������� �� ���� � ��� �� � ������� ����>������ �� ���� �� �������� �1� ���� �* �������� ��� ���� � ��� �� �������� ����#�� ���� �� �������� ���� � ��������� �� � ��� ����� ����� �� ��� �� ���� �� �#�� �� �* �*��� �� ��� ��� ����� �� �� � �� �����*��� ��� '%i ��� �� � � ���� ��������� � ������ � � ���# (=i ��� �� ������� ������� ��� �#���� ��� �� �#� �� �� �� h� �� �� )�� �� ������� .� �� � �� , ������ � �� .������� ��� �#���� �� ��� � �� =�$!! ��#�������� $�'?% ���� �� ���� ��������� � ������ � � ��-�! ���� �� ��� ������� /#� �������� �� ��� ����� ��� �� � ������� �������� !% �� � 3 ���� �� � ���� ��� =�!b' ��������� � � $��=!������� �� � � �� �#����� 4���� ���� ����������� ������ ���� ���� � ��� ��������� � �� )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" =! �#���� � ������� �� ����� � ����� �� � + � ���� ���� �� � + � ���� ���� �� � � ������� ���� ���� �*������ � ��#�� � ����� �� � ���� �� �� ��� � ��1�� �# #�� �# ��� �� ����� ����# �� ���� � �#�� ��� ��������� ���� �� � �� �#�� ��� ���� �� � � ����� ���� ��� � 3 ����� � ��������� � ��� � ����� �� ����� � � �� ���� �� ���� ��� #��� � ���# � �� ����� � �� �� �����+������ �� ����� � �� � � �� �#���� ��� ���# �� +�� ���� �� ����� � ����� ���� ���� � �� ��� � �� �� �� �� �� �� ��� ���# �� �� �� �� �� ����� ���# � �� � ��� � ���� �� +�� �#�������� ������� �� �� ���� �� � ������� ��#�� �� #���� �� ������� �� � � �� �� � �� � $��=!-� �� +�� ���� �� �� #�� #���� �� ����� �� �� =�$!!� � ����� ����� ��� �#����� �� ��� � #�� �� � � ���9� �������� ��� ����� �� � � �� ���# �� �� �� � ��+ �� �� ���# �� ��� � #�� � ����� � � ��� � ������� ��� �� � ���� #��� � � ���9� +�� ���� ����� ��� ����� �� � ����� ���� �� ���� �� ��� �������� ��� ����� �� � ���� �� �� ���# �� ��� � � � �� �*�� � �� �� �������� �� ��� � �� ���# �� � � �#�� ��� �� ��� 2��������0 ������� �������� ����� ��� �� ����� �� ���� �� � �� �������� �� ��� � �� ���# �� ����� ���� ��� 2������� ��� ����� �� � � �� ���# �� � ��� ����� �� ���� �� � �� �*�� � �� �� �������� �� �������� � ���� �� �� ���# �� ��� �� 2�������� ��� �� ���� ��1��� � ��1� ��������� � � ���� �� �� ���# �� ��� � �� ��� ���� �� �� �� �� ��1� � �#���8#� �� �������� ��� ����� � �� �� � ���� ��������� ��� 5�� ����� �� � � �� �������� �� �� ��� ����� �� ���� �� � �� ���������� � �� �������� ����� �� �� ����� ����# �� ��� ���� �� �� �� �� �������� �� ���� � ��1� �� � ��� �� ������ ��� ��� �� ����� ��� ���� �� �� ���� �� ���� �� ��� �� � ��� ����� � � � � �� ���# � � �� ��� � ��� 2������� ��� ����� �� � � �� ���# �� � ��� ����� �� ���� �� � �� �*�� � �� �� � �������� ��� ��� #�� � ���� �� �� ���# �� ��� �� 2�������� ��� �� ���� ��1��� � ��1� ��������� � � ���� �� �� ���# �� ��� � �� ��� 1 �� ��� �� ���� ������ � ���� �# �� � ���� �� �� � ���� �������� �� �� �� � � � ��� � ��� ������� � ��� ��� �� � � �� ���# � � � ��+ �� �� ���# �� �� ����� ��� �� ������� ���� �� ������� � ��� �#�� � �� ���# �� k��������� �� ���� �� �������� � ��� ���� �� ��� 8#��� � ��� ;�� ���# � �� � �� ���# �< � � ��� 8#� � � ��� ;���# �� �� ���# �� <� ��� ������� � ��������� ����� ���� �� �� ��� �� � � �4�� � 4$� � 4(� � 4=� � 4-� � � 5 ;�< ����� � �� �� �� ���# � � ;%� �< ���# �� �� ���# �� � � 4� �� �������� �� ��� �� 4$ �� ���# �� �������� �� ��� 4( �� ����>����� �� ����# �� 4= �� ��� �� 4�� �� �������� ��� ���# �� �� -% )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" 5���� � � ��� ������ ��� #��� � �*��� � �� � �� 8#��� � ��� ;��� ��� �� � � �������� �� ��1� � ���������< � � 8#� � � ��� ;���# �� ���������< � #�� �� �� �� �� 7� ����� ���� � �� �� �� � $% � ���� �� ��������� � � �� #���� �� � ���� �� ��������� ��� ���#��� �#� � ������ ������� �� �� � �#� �>����� ���� �� 4 ����� ����� �� ��� �� �� ��������� � �� � ������� �� ����� � � � � ���#�� �����+������ ��� ����� ���#�� �� ���� �� � # ������ � ����� �� �������� �� ��+ �� �� �� ��������� ���� �� ���� �� � ��4 � 5 ;$< 3 ���� ���� �� �� # ����������� 7�� �� ������� �� � �#��� �������� � ��+ �� �� � � �� �� �� � % � � %� � ������� ;(< 3 ��� ����#�� �� � ��4 �� 2;�� n4 <� /��� �� ���� �� � ;$< � � ;(< �� �� 5����� � �� � 5����5 � ���4 � � �� �����4 � ;=< ����� � �� �� �#�#�� ��� ��� ���# �� �# � �� ��� 5� ��� �# � �� �� ���� ��� � � ;=< ���� ���� � � �� ���#�� �� � ���� ���# 5 � ;$<� 3 �� �� 6 � � ���� �� ���#�� �� 5 ��� 3h;%� �$<� 3 ��� ����� �����4 � � � ���4 �� �� � �$����$ �*� � �$ $ � � �� ;-< �� �� �� ���# �� � �� �� ��� �# � �� �1� ���� �#���� �� �� ���� ��� � �� �� ������� �� � �� �� �� �� ���# � 9���� ��� ��� ����� �� ��+ �� �� ������� � � ��4 � 5 � �� �� � % � � %� � ������� ;b< � �� � � � ��� �� �� # 1 �� ������ ���� 4 � � � ��� �� �� 1 �� �� � � �� 5 ��� �����#��� �� ��� � ���� �� �� � � ������� ��� ���# ��� �� � ��� 9��� � � � ����� ����� �� �$� ���� �� � �� ����� ����� ��������� ������ � � � � �� ��� �� �� ���� �� � �� �� ��� �� �� �� �� ����� � � �# �� �� ����� ��� ���# �� �� 7� #��� �3�f25 ���1��� � �� �#� �#� ��������� �� ��� �� �� � ������ ����� ��� �� ����� � � �� �������� �� ���� �� �������� .)h� 3k� 3 �� +�� ����� � ����� �� � ��� ��1�� �� ��� ��� �#��� �� ��1� � �� ���# �� � � ���� �� ��� ��� ������� � �� ���� � ����� � ����� �� � ��� ��1�� ��� �#�� ��� ��� �� ���# � � � ���� �� ��� ��� � ����� � ��� � �� ���#� � �� �� �� ���� ����� ��������� � ������� .� ���� � �#� ���#�� �� � �� + �� �� � ���� � � ��� �+�� ���� �� ���� �� ��� �� �������� �� ���# �� ����� ��� ;4$< � � � ��������0� ������� � ��1� �� ���# �� � � � ��� �� ��� �� ����� ��� � ����� ���#� �� �������� ��� �� ��� � �� � + �� �� �� �#�1� � ��� )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" -� ;�!!=< � ����� � ���� �� ���� ��� ��# � �� ��� � �������� �� ��� � ��� #�� � � �������� ��� ����� �� � k���� � ���� ���#� �� � � ��� �� ��#�� �� � ��������0� ������� � ��� ����� � � ��� �� ��� ����� � ��� ����� �� ����� ��� �� � ���� �� �� �� �*�� � �� �� � �������� �� ���# �� ��� #��� ��#�� �������� �� ���# �� � �� � ������ � �� � � �� ��� � �� ��� � � ��������� � ���� ���� � ��� ���� �� ����� ��� ����� � ��� ����� ��� �� ��� + �� �� �#���� �� ���� ����� �� � ��� �+�� ���� ��� ���� �� �� ��� �������� ��� ����� �� � � �� ���������� � �� �������� ����� �� �� ����� ����# � ;4(<� ����� ���#� � ���� � � ���� � �� �� � �������� ��� ���� �� ��� ���� �� � ��� �� ������ �� �� ��� �� ���# �� � �� ���� �� ���� ��1��� � ��� ����� � � �� ����� � ��� ����� �� ����� ��� � ��1������ �� �*�� � �� �� � �������� ��� ��� #�� ;4=< ��� � �� ����� ��� � ��� � ��� �+�� ���� ��� ����� � �������� ��� ����� �� � ���������� �� �� �� � �1� �# #�� ��� � � �� ����� ��� �� ���# �� � ���� ������� � �� � �� ��� �� ��� �������� ��� ����� �� � �� ����� ��� �� ���� + �� � �� �� ��� � �� � � ������� � #�� �� d��� ;�!!'<� ��� ���������� � �� �������� �� �������� ;4-< �� � ��� �+�� ������� � �� ����� ��� ��� ����� �� � ��� �������� ����� � ��� � �������� �� ������� ��� �� ������ �#�m�� � � �� �� 3 � ���#�� �� �� ��#�� � �� �#�� ���� ���+�#� ��� � �������� � ������� � �������� �� ������� ���� ��� �� ��� �� ����� ��� �� � ������� � ��� � 3 ���� �� � �������� �� �������� ��� �#�m�� � � ����� �� ������� �� �� � ���� �� 3 �� ���������� �� ���� � �� � ������� �� �� � �������� �� �������� ��� �# ����� �� �� �+ �� 4� � ���#� � ��� ��� #�� ���� � ������ � �� � �� ��� �� � � ��������0� ������� � ��� ����� � � � �������� ��� ����� �� ����� ��� � ����� $ ����� � �� ���#� � �� �� ��� ��������� � ������ ��� �� ���� � ����� � �� �� ���� �� �������� .)h� 3k ����� ��1� ����� �� � ��� �#�� ��� ��� �� ���# � � � ���� �� ��� ��� � ��� + �� �� � ���� � �� �� ���� ��������� ��� ��� �� ���# � �� �� �������� �� ���# �� ����� ��� ;4$< ��� � ��� �+�� ���� ��� ����� � �������� �� ���# �� �� ��� ���������� �� + �� �� �#���� �#� ���� ����� �� � ����� ���� �� ���� �� ��� �� ������� �������� �������� ����� ��� � � �� �*�� � �� �� � �������� �� ��� � �� ���# �� � ����� + �� �� ��� �� ��� � �� � 5��1� ;�!!=<� �������� �#� ����� � 5���� � ������ �� ��� ��� ����� � � ���������0 ������� � �� ���# � � �������� ��� ����� �� ����� ��� � 3 ���� �� ��������� .���+��� ����� �� ����� �>���#� 2������� �� ���# �� � �%��$�=$(=-$$ �%�='?$(?b!($2�%� %��(($ 2������� �� ���# �� ����� ��� %�-=�=b'!$%b2�%= %�$�($b$b?$%2�%= %�=�!$ 2������� ����� �� ����# � %��=$!'�'�?! %�-b(�%?'=b!2�%� %�%%=' 2������� ��� ��� #�� %�('!'%b(!�$ %��=!--�$=?� %�%%($ .#��� ��� � ���� � ����# %�$??%b%?!?= %���(=--'($! %�%%=� �������� �� �������� �%��-$=%'(�$%2�%$ �%�b%%$'�'!b�2�%( %�!?%� 4 ����� ��>� ��� ���� �8#���� ������#�� ��� �������� #�� � �� � �1� ���� �� �!!�./� ��� ����+��� � ����� �� ����� � � � �>���#�� �� ���� �� �� � ���� �� ��������� �� �� � ���� �� � �� ���� �� ��������� �� �������� �� ���# �� ������� � ��� ��1������� � ��--!�b=%� .��>�8#���� � (((�(%-%� ������ ����� � � %�b(� ���� �� �������� �� �������� �#��� �� �� ���# � � � �� ��� ��� o -$ )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" + �� �� ��� � �� ��� � �� � k���� � ��� ;�!!?< ����� ���#� � � ���� � �� �� �� � �� �������� �� ��� � ��� � ����� �� ��� ����� �� �������� ���������� ������ � �� ���������� � �� �������� ����� �� �� ����� ����# � ;4(<� �� + �� �� � ���� � � ��� �+�� ����� � �� ���# �� �������� �� ���# �� � � ���� �� � �� ���������� � �� ��� ��� #��� 4� ��#�� �� �*�� � �� �� � �������� ��� ��� #�� ;4=< ��� � ��� �+�� ���� ��� ����� � �� ������� � �� ���# �� � �� � ��� �+�� ������� � �������� �� ���# �� �� ��� .� �������� �� ���������� � �� �������� �� �������� ;4-< ��� � ��� �+�� ���� ��� ����� � �� ���# �� �� �� �# �� � � ��� �+�� ��� �� � �� ��������0� ������� � �� ���# � � ��� �� ��� ��+ �� �� ���# �� ��� � ��� + �� � �� ��� �+�� ���� ��� ����� � � �� ���# �� �� �� �#� � �� ���������� � �� �������� �� �������� �� �� ��� � �� � + �� �� �� d��� ;�!!'<� ! ���������� �� ������� �� "����� �#"��������� ,�� � �#���� �� � ���� �� �!!�./� ��� � #�� � ���9�� �������� �� ����� ��� ��� ����� �� � � �����#� �����>����������� � � ��� ���� ��� ��� �� ������ ��� �� �������� ��� ����� �� ��� ��� �� ���# �� �� ��� ������� � �� 8#�� �� � ���� #��� � ����#�� �� ��� ��� �� ���# �� � �� 4�� ��# �#��� �� ��1� � �� ���# �� �o $� 7�� �� �� ���# �� ��#� �� ���# �� �o )#� ���#� � � ���� � � ��� �+�� ���� �� ���� �� ��� �������� �� ���# �� � � � � ��������0� ������� � ��� ����� �� �������� �#� ���#� � �� � ���� � ���� � �� � �� � �������� ��� ���� �� ������� � ��� ����� �� �� ����� �� �������� �� ���# �� � ��� � ����� � �� ��� �#�� � �������� �� ����� � � �� ���# � ���� ��� �� ��� ��� � ��� ��� + �� �� ���� ������ ������� �� � ��� �� � ���������� � � �#���� ������ ��1���� ������ � ���������� �� + �� �� � ���� � �� ����>����� �� ����# � � � ��� ��� #��� � �� ����� ��� � ���� � ���� ��� ����� � �� �������� �� ���# �� ������� � ���������� ��������� ���#�� #�� ���� ��� #��� � �� ����� �� �� ��� ��� � �� � ��� � �� ��*���9� � �������� ��� ����� �� � ����#� � � �# #�� ���������� � � �� �� � � �*�� � � ���������� ����� $ ���� � ������ �� ��� ��� ����� � � �������� �� ���# �� �� �� � �������� ��� ����� �� ����� ��� � 3 ���� �� ��������� .���+��� ����� �� ����� �>���#� 2������� �� ���# �� � ='�b$!'(bb( =$�$b?b($!%�('�� 2������� �� ���# �� ����� ��� %��?�-?b$??% %��b��='(�=' %�%%%% 2������� ����� �� ����# � =b�%''=-?(b =%�?!�%'-?? %��%�� 2������� ��� ��� #�� �%$�''-(%=b !��$%'�(�%� %��-!' .#��� ��� � ���� � ����# �$b�-?!!bb? ��$�(=�$b=%�%$?�������� �� �������� b���%$�?(?$ -=�$$=b=!b$ %�%--! 4 ��� ������#�� ��� �������� #�� � �� � �1� ���� �� �!!�./� ��� ����+��� � ����� �� ����� � � � �> ���#�� �� ���� �� �� � ���� �� ��������� �� �� � ���� �� � �� ���� �� ��������� �������� �� ���# �� �� �� ��� ��1������� � �!$!=�b?%� ���� �� �������� �������� �� ���# �� �� �� )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" -( �� �� ����>����� �� ����# � � � ��� ��� #��� ���� ���� ��� ����� � � �������� �� ���# �� ������� �� �#���� ������ ��1��� ���#�� � � � # � �������� �� � ��#���� +��� � � ��#�� ���� ��� #�� � ���� �� ����� ��� �� 3 ���� �� � �� ����� �� �� ����� � ������ �� ���# �� �� �� �� � ������ � �� � #�� �� ���# �� ���� � � ������� � ���� �� �� �* � �� ����� ��$��%�� �#���� 7� �� 1 ���� � � ���#� �������� � � �� ��� �� ��� �����#� �#���� �� � � � ����� �� ��� #�#�� ���������� �������� ��������� 4 ������ 2� �� ;�!!$<� ��� ���� � � � ��� ���# �� �� =%�;1< ��� �� 7����� � ���� ��� 7���� � � � f. ,��� f���� �� �� ������ k���� � 7� /�� /���� �� �� "�� p �����#��� 4� 5� ;�!!?<� ��� ���1��� #�� =%�;1< ��� � �� ��� ����� �� � �� ���# �� � � � �� ������� ������� �� ������� /������ 8��� ��� 9 ��� $�� /�� ��� :#� k#�1 .� �#� � � ;�!!?<� "�!0�1 ����� �5���� ��� �� � ���� ��� &�� h�� "����� ����#���� 2�� � 7� 2�� p ���5����� � f� 4� ;�!!=<� 7�� ��� ���� �� ��� �������� ����� � � �� �!?%�o '��5��� �� ��� ��6 � /���� �� /�� ��� " � =(!6=-(� ����� �� ��� p 5���� � 4� ;�!!?<� ) ���� � � �������� � ��1 � ��+ �� �� ���# �� ��� �� 7��1� � 5����� 3��� ��� � � � , ������ �� �#�1�� 4� ��� 5� ����� "� ��� p 7����*� f� 7� ;�!!=<� 2������� ������� � �� � ������ � =%�;1< ��� � ����� �� ���� � �����#��>����� �� �� 7��1� � 5���� �� =b-(� .��������� �4� h� �� �� k#���# �� 2�� ���� �������� ������� h�*� � p 7������� ;�!!!<� $ 4�� ���5�� "�!0�1 �5����� .������ �� .�������� 5��1�� �� 2� ;�!!=<� f��� =%�;1< � ���#� �� ����� ��+ �� �� �+ ��� �o %� ���� �&� ��� � &���,$ 4�� ���5�� ; � ������ � ��� ��� 7�4 ;��� �$$6�$(<� 5��1�� �� 2� ;�!!-<� 5�� ����� �� � � � �� ���# �� � � =%�;1< �� ��� ��� � ����� �� ���� ��� �� �� < 5���� -5��� /�� 5����� :�;$<� (��6($-� 5��1�� �� 2�� p 5� ����� "� �� ;�!!-<� �#���� ����� �� � �������� �� �� �� �� � � �������� ���� � �������� � =%�;1< ��� �� �� � � �� �������� "(;(<� (�(6(�'� 5��+ ����� �@=%�;1< .�# ��� �� 4������ ;5�.4< ;�!!?<� "�!0�1 ��� �� =� ���� �& ���� �� & 6 � �� �5����� 3�� .������� d���� "� "� ;�!!'<� ���� � �� ���� � � =%�;1< �� ���# �� �� � ���� �� ; 5���� �& ��� %���� �& ��5��� �� b-6'(� e#�� e�� p f�g� ��� �� 4� ;�!!b<� f� ���� � � �� ��#���0� ��+ �� �� ���# �� �� ����� �# ��� � ���� �� ; 5���� �& ��� %���� �&� � (�6(?� -= )�*� -������ $� �������� . � ���� �� $��� ��� /�� �� !� 0���!1 "232" the decision to defer: factors affecting employee deferral incentives introduction evolution of deferral literature and significance of this study data methodology results discussion of results and policy implications acknowledgements references pii: s1057-0810(02)00088-4 ������������ � � � � � �� � � �� � �� �� � ��� � ����� �� �� � ���� ��� �� � � � ��� � ��� ��� �� � ������ � �� �� � � ��� !� " �� # ! ���������� � ������ ������ ���������� ������� ��� ������ � ��� �� �� ��� � �!"��# ��$ %��"��# � $ � % & '� �( � )**!+ � � $ � � $ � � ���� ! ��$ �( � )**!+ ��� '� � !, � �( � )**! �������� .� '�' � � /�� /��� �� � �0���� � ���'� / $ ��� � ����� / � � � / � �� 1'�� � (������� ���� � � � � � �� ��� � �� � � � ��� � ���� ���� �� �� �� � ����� � � �� �� � (2 ����� ��� ������ .� ����� ��( � �� �� ��(� ���'� � � �� � � �� ���� �� � � �� � /� � ��� $���� � �2 �� � $ ���� �� � � ��� � � �0���� �� � � �(�2 ��� $ ��' � �� � ��� �� 31� �� � ��� � �� �� � � � � (����������� � �� � /��2 $���� � � � ' � ���(2�' � �� (�� � �� �� �� � /� $���� � �2 �� � $ ���� ���� � ���( � � �� ����� '������ � � � � �� �� � � �� � $ �� 0��� � � � )**! 3�� $ � &� � 4 �� 5�� � /��� � � �$ �� &'( �"���� ����� �) 6!!+ 6!)+ -7* *��$ �#�) & � �� ��+ ������������ �+ 3�� �2 � ��� � ���� ������ .� $�� � �2 �� �� 8�' ��� 5�� � �� � / 9�� � :85�9; � $ ��' � (2 &���' :!<=7;� # � � :!<=%; � � ����� :!<>); ��� ( �� ��(? �� �� �� � � ��(� � (�� � � ���� � � �� �� :!<<); ���� � ���� ��� $�� �(� �� � � � � (����������� �� ���( �� ��'��� ��� �� ������� �� � �� $�� �� � �� �2 � ��� �� 9�� � � � $ � � �� � � / �� ��(��� �� �� �� � � ����� � �� ( � ����� � � �� � ������2� @�� $ �� � � �� � �� ��2 ���� � � �� / ����2 ��� '� � ���� � � � � � � (������� ���� � �� ��� �� � � �� �� � ���� ��� ( ���� ��� $ � �������� �� ��$ /� / (� 1'�� ����2 '�� � � �� � /��� �� ����� � ��� �� .� � � �� �� � / ����2 ��� '� �� (�� � � � �� & �$ � � $ � !* :)**!; )<!a,*) !*%>�*b!*c*!cd a � ��� � ���� � � )**! 3�� $ � &� � 4 �� 5�� � /��� � � �$ �� �44� & ! * % > � * b ! * : * ) ; * * * b b � 7 � 8��� �'� � / ������� . ��� �=7�<�77,�<><<+ ��1� �=7�<�77!�b!>>� '+���" �##������) ?���' ����e���� 2���� � :���� � ����;� ����� ��� e���� 2���� � :�� � � ��� ;� � � �� e����� ����� :"� # ;� ! . ��� �=7�<�77,�<><<+ ��1� �=7�<�77!�b!>>� �� 1'�� ����2 '�� � �� �� (����������� � �� �� �� ��� ��2 ���'��� � �� � � �� �� � ��� �� �� � �(� � -�$ �� ����� � � �� �� :)***; '��$ � �� ��� �� $ � � � �� ��(��� �� �� �� (����������� � �� ��� �� �� �� 2 ��� � ��� ���$ $ � �� ������ �� �� � �� � ����� �� 5 ������ $ � � ���� ��?�� �� �2 ���� �� � ��� � ���� ���� ������ / � ����� /2 �� $ �� / � /� (����������� � �� ff$��� gg ������ ��$ � / 0�� ��2 � � ��' ����� � �� ���� � ��� / �� ' � �� � � �� � ��!<<*�� @�� $ �� �� � ��2 ( � � ����2 �� �� � (2 �� ���� � 1� �� � �� �� ' � ��� ��2� � � � 3� �����'� :!<<>; � � h�� � � � '' � :!<<>; ���(� �� � �� '� � � �� �� (����������� � �� �� � � � ��� �� ���� ���� � ��� � � '�� � $ � ��� � �� ' ( �� (����������� � � � � ��� �� .� '��'�� �� �� � ����2 � �� 1�� �� '� �� � ��� � �� � �� / ��� � � � ���� � � �� � � �� �� '��(� �� / ��� � (2 �� � � � � $���� � �2 �� �� � �� �� � ���� �� � ���������� ������ ���� � � �� �� :!<<); ��� �� �� 0��� $ � � �� � (����������� � �� ��� �� � :!<b!; ��� ��� /�� �� 0��� �� ����� � �� � � �� �� �� � ��� � ���� ����� � 0��� ��� � /� � � �� ��?��� � � ��� � � �� �i&3� '��$ � / $ � � �� � �� � � �� �� 85�9� j � :!<b,; ���� � �� � � �� �� ��� ���/ �2 ���� (���(� �� �� �� �� �� �� ���2� ���� :!<b,; ��� � ���� �� � �� �� �� � � �� ��� ���� � � �� � /� �� /� �� '� � :'c�; ��� ��� � �$ ��/ � /�$ � /� � � �� ��?��� � � ��� � ��� ���� � �� ��� 'c� ��� ��� �� �� � � .��� � :!<b%; �� � �' � ��� ���� � �� ���� �� � � �� �� �� � ���� � �$ �� ��� � �� � (2 '��� ��� �� ���' ����� '��� � ��� .� '� � / � ���� � ���� (2 �� � � �� �� � � 'c� ��� �� � � �� /� � ����� / �� 8�� � � 8� :!<bb; ���� � ���� (2 ��� � / �� ��2 �� �� ( �� ��� �������� �� 0�� � � ( ��� � '��12 ��� ��� ( ��� � � �� / �� � ����2� 4 8�� � � 8� :!<<!; �� 2 �� � �� ��� � �� � �� �� ���� ���� �� � � �� �� ��� � � ���� �� �� � /� � �� ����� �� � � �� ����� 0���� 4 �' � �� �� � ��� � $ � � � ���� � � �� �� :!<<); '��$ � � $ � � ���� � � ����2 � � �� � �� 85�9� k � � �� � ����2 ��� '�(� �� � �� $ � � ��� �� ���� ����2 � � ��/�� �� � � � � � 'c� � � ���/ �2 � � ���$ �� .� 2 ��� � ���� � � � � (������� ���� � ��'��� � ��� ������� �� � �� $�� �� � � ��� ����� �� � � �� ( ��� � � � � $ ��/ � (����������� � � � 'c� ��� ��� � � ���� ��� � ����� / ��� $�� �� � �� ( ��� ����� �� � � �� 0�� � � � ( �� ��� � 1'�� ����2 '�� �� .��� � �� �� ��2 ���� �� (����������� � �� �� � ������2 ��� '� �� .� �� ���$ ��2 �� �� 1 ��� ������/� �� $ � � � �� ��� ��� ���'� � � .� � /� � ��� � �� ���� � � �� � � /� (����������� � ��� �� �� �� $��� '� � �� �� � ��� ��� �� ( ���� �� ��� ( 1'�� � $�� ��� ��2�� .� 0��� �� �� � �� ���� � ���� �� $��� '� � �� � �� �� ���� � / � � ���� �� ��� � ����� �� � � � �2 �� ( �� � �� ��2 ��� � ���� � �� �� / �� ���'� � .� � $ � � ��$�� � (2 ����� :!<<,;� #� � � 9��j ��2 :!<<*; � � 9��j ��2 :!<<%;� .� � 1'�� �� � � �� ���/ �2 � ? �� � �� � ��( � �� ���� ������'� � ��� ��$ � '����� � �� ��� � � ����� � $�� �2 �� ���� ��� � � � � ' � ���� ��� 1��'� � 8�� � @����� � � #��� ���� :!<<!;� 8�'���� ��� 2� � � &���' :!<<,;� -�$ � :!<<7;� ���� � � �� �� :!<<b;� 4 ��� � � �� ��// �� � � �� ���� � /� � ��� ���� )<) &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� ( ��// �� � ���� ���$ $�� ( �� � �� ���� �� �� (����������� � �� ��� @�� $ �� ���� � ���� �� -�$ � :!<<7; ��$ � ���� � � �� �� 5 ��� � 1'�� �� � � ���� �� $��� '� � �� � �������2 � '� � �� ��� � ��� 8 ��� �2 �� � � �� �� ( / ��� �/� � � ���� �� � ��� '��� �� �� '� � �� � �� �� � ��� �' � ���2 ���� ����� �� � � �� � � � .� � � �� 1'�� �� � ��$�� � (2 ���� � � �� �� :!<<,� !<<=; ��� ��/� ���� �� $��� '� � �� � � ���' ��� � ��� � �� � ���� ������ ��� � ���� �� '��'�� � �� ��( ���/ '� � / �� ��2 �� ��� :!<>=;� ���� � � �� ��g� ��� ������ ��� � ���� � �� 1' �� � � ��� �� � '������ � 1� �� �� �� � ����� ��� � ��� �' �� �� 1� �� � ��� � �� ���� � '������ � '��� ������� � ��� � �� � � � � (����������� �� '�/ � � /� � 4 �'�/�� � /� � � � '�&9�� � � '�@9#� :!; ���� � � �� �� :!<<%; ���� ���� �� ���' �� �� &9� :����� � �� ( /; � � @9# :� /� � �� ���; � � ����� ( '��1 � ��� �� � ��� $ � ��� �� �� 0���� � ��� � �� � ��� ����� '� � � ��� $ �� (��� $��� � � �� ��$ ' ��� ��2 ��� � �� /� � �� �� �''�� � �''�2 / �� ��� (����������� � ������� .� � ��� � �� ' � �� ( �� � � � � � ��� �� ��� � �� �� ��� .� 2 �� ���� ���� �� 2 �� � �(� �� � � �� � � ��2 / �� �� � $�� �(� � ���� '����� $�� �( � �2 �� /� �� �� � ��� � � ��� � �� � � � � (����������� � ��� ��� � �� � �� �� 2 �(� �� � �� � �� � �� �� � $�� �(� � '����� $�� �� � �� ���'� � � � � ���� ��� � ��� ���� ��'��� � (2 � �$ ���� ���� � ������� �� � ���� � � �� �� �� �(� �� ���� ���� �� $��� '� � �� ����� ( � '� � �� ��� � ��� �� 2 �� �� ��$ ��� ��2 �� /� '��$ / ���� �� � � $��� '� � �� � � ��� � �� � �� � � � ��$ � / �� 1��� � ��/ ( �� �� $��� '� � �� � � � ��� 4 � �� -� � � � . ��� :!<<>; ���� ( �� $��� '� � �� �� � ( ��$ ���� '� � � � ��� /����� ������ � � � �$ �� � ��� $��� ������ ���� � ��� � � l �� � �� � ��� ������/� ���� � � �� �� :!<<b; ��/� ���� �� � � ����� �� � �� �� �� ' � �� ���� !<>, �� !<<, �� � �� � �� ����� �� � ������� � �� �� 5 �� �� 1'�� �� � ��� �� $��� '� � �� � ���� � � �� �� �$ �� ��� � (2 $ ����� �� � 0��g� ' ������ � � 4 $ ����� �$ �$��� ������ ���� ��$ 1' � � � � ' � �� �� ���� / /����� �� /�� ( ��� / �$ ���'� � �� � �(��� ����� /����� '���' ���� .� 2 ���� � � �$��� ������ ���� ��$ ' ����� � '����2 ��� / / ���� � ���2 ��� $��� � �� �� �� .� ���� �� � �� �� � �$ �� ��� � � �� �� � �� (����������� � �� ��� .� � �� � '� � �� �� 6����� � � -��� :!<,7; 1'�� � �� $ ��� � �� $ �� / ������ � �� ��� '� � � � ��� $ �� � $ � ��� �� /� � � (��� $��� � �� �� � � .��� � :!<b>;� #��� ����� &�� � �� � � h �� 2 :!<<7;� @��/ :!<<%;� � � #������ :!<<=; ��''��� �� $ � ���� �� $��� '� � �� � �� �� $ ���� �$ �� ��� � � 4� �� � 1'�� �� � � ���� ��� �� ���� � � �0� � � � �� �2 1'�� ��(� � � � �� 1'�� ��� � �� �� � ���� � �0� �2 ������ � ���� �� � � �� � �� � ���� � �� �� $��� '� � ��� 8� ���2 � � ��� '�� �� �' � �� ���� � � �� �� ���� �� $��� '� � �� � � ��� � � � ���2 �� � �� �� �� $��� '� � �� ������ ' �� �� ��� ���� �� �$ � 1� � � ' � ���� ������/� ' � �� � � $ �� � � ����� �������2 ������ .� ��� ����� �''�2 � � ��� '�� �� $ � �� -� � � � . ��� :!<<>; ���� �� $��� '� � �� � � � ���� �� � ( ��$ ���� '� � � � � m �� ��� � �� �� � �� (����������� � �� �� � �� � ���� �� ���� � /� �� �� @��/ :!<<%; ��� �� �� � ���� �� $ ���� �$ �� ��� � � �� �� �� �� ��2 � ��'' �� �� � �� � � � ����� 6 $ �� 1� � $ ��� �� ���'� � �� / ���� ��� ( �� � �� $ � ���� �� $��� '� � �� � �� �� ���� � / � � �0���� �� �'����� .� $ ���� �$ �� ��� � &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� )<, �� ��2 � ��� $ � �� � �� � � �� �� $ � � � � � � �� � � � � �� ���� �� �� (������� ���� � �� ��� �� � � '��� (� ���� $ ������ � � �� 2 ( ��� �� $ � � ���� �� 2 �� �$ �� ��� /� ��2 ��� / �� � ( ��$ ��� 3�����2 � �� (����������� � �� �� � � � l �� � �� � ��� (�� $ ����� ( � $ ���� � �� �� �� � �$ �� ��� � � �� 2 ��2 ��?��� �� � ( ��$ �� �� � ��� � �� ����� ��� � � � � ��� �� $��� '� � �� ������ /�2� .� � ����2 �� 0 � �� �� �� � � � ��� � ����� .� � ����� �� '� $ ��� � � � ��� � ���� � �� ( � �'� � � �� ���� / � � ���� � �� $��� '� � �� � � � ��� � �$ � � � � 2 ���� ��2� � � � 3� �����'� :!<<>;� �� / ���� ���� !<>! �� !<<,� �(�� � ����� �� � �� � � �� ���� � � �� �� :!<<);� .� 2 ���� ���� � ���� ��� � ��� ( ��� �� �� '� � �� ����� � ��� �� � � ���� �� � � � '� � �� ��� ����� 0�� � � � � � '�� � $ � ��� � �� ' ( �� � ��� � � (����������� � ��� ��� .� 2 0 � ���� (����������� � 1'�� � � ��� $�� �� � �� �� ���� ' � �� � �� � � � � (�� ���� �� � ��� � �� ' ��� � /�� $ �� ' � �� ���� !<bb �� !<<,� ���� (�� / �� � �� 0��� � �� ��� � ��� �� ( / ���� � � ����2 ��� �� � (2 �� !<b> ���� ���� � ������ �� �� ��� '��� ������2 � $ � � � � ��� �� .� � � � ��� � ���� ���� � �� �� � � �� �� !<b> ����� ��� ��� �� � (2 $ ��� � ���'� � ���� �� ��� �� � � ���� � ��� � ���'� � � � ����� ���� ���� /� � � ����� ��� � ���'� �� ��� / (����������� � ��� �� � � � � �(� � ���� �� �� ��� (��� $��� �� �� � � ��2 / ��� ��� h�� � � � '' � :!<<>; �� ���� ���� �� ���2 !<<! �� � �( � !<<%� � ? �� / ���� ���� '� �� 2 ��� ��� � �� �$� � �� �� � �� '��� ���� ���� � ����� ' � ��� .� 2 0 � (��� �� � � �� �� � � �� (����������� � �� �� �� '� � � ��� / �� � ' � ��� .� � ��� � �� ' ( �� ����� � � � � � /� � ��� � ���� / � � �� � � ��� � �� ' � �� � �� � � �� � �� �� ��� ���� � $ �� � �� � ��� � �� � �� � � ���� �� � ��� .� ���� 0 � � � ��� '� � �� ��� � /� (����������� � 0��� � � ���� �� � ��� 1'�� ����2 '�� � �$ � � � �(�$ �� � ��� '� � �� ���� (���(� �� �� � � �� ��� @�� $ �� �� 2 0 �� ff.� �9 �� �� � �� �� � '� � � (�� �� �� � ���� �� ����2 � / 0�� �gg� 5 '� � � ��2 ����2 (2 # � � � ���� :)***; ��'� ��� � �� h�� � � � '' � :!<<>; ����2 �� / � � ��� � �������/2 � ���� ���� �� � �� !<<% �����/� �� !<<<� 4� ��� � �� (����������� � �� �� ��� �(� � � � ���� � /� (����������� � 0��� ���� 2 �� � ��� � � ��� � ��� / �� ' � ��� ������/� �� � ����� � � �� ���� �� ����2 � / 0�� �� .� � ����2 ��� �'�� �� �(�� � ��� � 0 � $ � ���� (2 �� / ���� �(�� � ���� � �� / � ' � ��� ���� � ���� � ��!<<, � � � 9���� )**!� � � �� � � � �������/2 ���� � � �� � �� ���$ $�� ( �� ��� � �� ��� � h�� � � � '' � :!<<>; ����2� � ����� ����� .� � �������/2 �� � � (�� � � ���� �'��2 � (2 ���� � � �� �� :!<<); �� � (2 �� ������ ( / � �� � �� � $ � � �� � � � � �� �����2 � /�� � '������ �� (�� � 0����2 � � � � � �� � (����������� � ��� �� &��� � � �� � ���� � ��' ��� ��� � �� � �� ��� �(��� ���� d)* � �� � :k�&� db � �� � ; � � � � � �� ��$ ���0� � ���� / $���� � �c�� ���� � ��' ��� ��� � �� ����� �� � �� ( �'��2 � � � � /��� ���� / ����� /2 � � � � �� � ��� � ? �� �� � � � �� �� ��� � �� ��� / ���'� �� � �� � '��' ��2 ������ � � � �� � � 1 �� �� � � �� ���� � �� ���'� � .� � � �� � ����� �� b* �� �� ),! ���'� � ���� ��2 � �� � � �� � � � ��� � &���� 31��� / � .� ���� � � � � $ � � )<7 &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� �� !* �����2 � /�� � '������ �� �� b ���� � ���� 8��� � / m���( � !� !<<,� � � '������ � ��� ���� � $ �2 , �� ��� � � �� � ��� � � � �� � ��� � � ' � �� �� = �� ���� @ � �� � 2 � � �� � � � ��� � � � �� '������ �� 1 �� � � � �� � ,��� �� � � �$ ���' ( �� � � �� / $ / � ����� �� ,* '������ �� ���� � ��� / �� � �� ' � ��� .� 0 �� � ��� � � �� � � � ��� �� =��� �� ' � ��� �� � �( � ,!� )*** � � 9���� ,!� )**!� &��� '� � �� ���� � ��' ��� ��� � � � ���� ��� �������� / � ��� � � � �(�� � ���� -������ ��� ���� $��� �� � ��� �� ��� �� $��� ���� ����� ��$ ( �$� ��(� �� $ ����� �� �� � � ��� '������ � ��� ���� �� -������ �� ��� ���0� � ���� � � � � ��� � ���� � �� ����� �� ����2 �� 1� � �� �� ' � �� '� �� �� m���( � !<<,� � �� ( �� ��� ���� ���� � �� � $�� �(� �� � /� �� � � � �� ��� �� ��� � �$ � �� =��� �� ' � �� �' �� �� '������ � ������ � ��� � .� � � ���� �� �� ��� / ( �� � ��� ���� � �'� �� �� � � � � ��� ��� �/ � ���� ��� ��'� �� ��� � � ����� �� �������� � '��$ � ��� ������� '� � �� � � �� � ��� �� @�� $ �� �� � ����2 �� � �� ��� �'� �� � �� �� '� � �� $ '�� � �� ( ��� � � � � ���( �� � �� (����������� � � � � � �� � �� ���' ����� �� � � �'� ( �� ( ������ �� � '� � �� $ '�� �� ���� ( �� �� � �� -��� ��� ���� � �� !* /���'� ��� ��� ' � �� ������ / �� � � � (����������� � � � ( ��� ��� � � � �������� � ��� ��� /���' � � �� � � � �� � ��� ��� ��2 � ��� ����� � � /� �� � � ��2� � �� � �� �$ ��/ 1� �� � ��� �� ��� �� �� !* '������ �� ��� �� ,* ' � ��� ��� � /� �� � �/� �� �� '������ � ��( �� 4 ��� � � ���� '� � /� �� � � ��2� � � � ��� ��2 � ��� ����� � � /� �� � � � � ���� � ��� � � �� � � ���� �� / �� = �� ���2 1� �� � ��� � �� � $ ���� 0��� �� �� � ' � � $�� �(� � � � � � � (����������� � � � ( �� �� � ' � � $�� �(� �� .� � ��� � ' �� � �� / �� '������ � ��( � �� �� � ' � � $�� �(� � � ������� .� � /� �� � � ����� '� � � � .�(� � ! � � ) ���� ���� ��� ���� � ���� � �� (���� ������� � �� � � '������ ��� � (����������� � �� �� � � ��� ��/ ��� �� ���� ��� � k�&� ���� � � '� � �� .� ���' �� �� � /� �� � � � ��� � ���� �$ � �� ,* ' � ��� '������ �� � �� �� � /� �� (����������� � '����� � � � ��� %�,n � /� � ��� �� '������ � � �� �� ��� �� ��� �� .� '��(� � � �� ���� �� 1'�� ����2 '�� � �� �� � ���� � �� � ��?� /) �� *�**7� �� '������ � ��( � � � /� �� � �/� �� '������ � 1� �� � ��� � �� ���� � /� )� �� ��?� /) � � � �� *�!!� (�� �� � � �� ���2 �� � /� ��/� �� / $ � 2 �� 0� � �� / (����������� � �� � '������ � � � �� � ����� .� � � �� ��� �� �� ���� � � �� �� :!<<); ��� � �� � � �� '� � �� $ '�� � �� (����������� �� � '� � �� (�� ����� 0��� � �2 '����� � � ��� 1� �� � ��� � .� � /� �� � � ����� ���� ���� ��� / � � �� � $�� �(� �� � �� �'��$ '� � �� $ '�� �� 4� ��2 ( �� � ���� ��� � �� '�� �� �� � /� �� � � � �� /�� $ � .� � � � � l �� � �� �� '��� ' ������ � �� �� ���� � ��� / �� ' � �� �� � (2 � �� � �� ��2 2 �� � � ��� � � �� ���� �� � ����� ��� � 4� � ���� $ � �� �� '������ �� �� �� �� ���� ( �� ��� ��� � � �� 1'�� ����2 '�� � �$ � � � �(�$ ���� �� (����������� � � � � �� �� ��� �� � ��� �� ���� '� � �� $ '�� � �� �� ��� � ��� �������� 5� �� � � � � ��'� �� ��� � � ���� � ( ��� �� '� � �� $ '�� � &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� )<% ����� ( 1' �� � �� ( ���� .� ���� ���� � ���( �� � �� (����������� � � � � � ��� � �� '� � �� $ '�� � � � � ��� � �� �� �( � �2 �� �� ���� � � �� �� ��� � �� '� � �� � ��� � �� � � � ��� � ���� �� � /�� ! � � ) '� � � �� � /� �� � � �� '������ � 1� �� � ��� � �/� �� '������ � ��( �� (�� � � (����������� � � � � � ��� �� ,* ' � ���� 4� ������ ( �� � ���� � � �� '��� (� �� � ����� ����2 �� � ������ �� '������ �� � �� (�� � �� (��� � � � � (����������� � �� � 2 ��� � 0�� ������� � �� � �� �� �� ����� ��( � �� 0��� ���� �2 / ��� � ��2� �� 5 $ ���� ���'� / � ����� /2 �� � � �� / '������ �� � �� (�� � �� (����������� � ����� �� � 1� �� � ��� � ��� $ � � �� � /) �� � �2 *�!! �� ����� /2 ����� ( 1�� � �2 � ��2 �� .�(� ! 8������ �� � �� � /� �� � � �� = �� ���2 '������ � � ��� � � ( ��� � � � � (����������� � 9�� � 4 � �� '� 4 � ' � � $�� �(� � � �� �a9 & � 5�?� /) ! �*�*>%<, :�=�=b<; *�**b,> :7�%>,; *�**b ) *�*=)%) :�%�7<%; *�**%<) :,�),); *�**7 , �*�*)7,7 :�)�!,%; �*�**!*! :�*�%%); *�*** 7 �*�!!>** :�>�,<b; *�**<*% :7�<,%; *�**=b, :,�>)=; *�*!7 % �*�*%)b< :�,�>,7; *�**=,* :,�,b,; �*�**)!, :�!�!7); *�**7 = �*�*>*>7 :�7�=7,; *�**b,= :7�%=>; �*�***<= :�*�%!!; *�**b > �*�!*>** :�=�**); *�**<*> :7�<7%; *�**>)) :,�b>b; *�**))* :�!�!b<; *�*!7 &����� � � ��� / � �� � �� !* ���� � /�� � '������ �� ������ / �� ( ��� � � � � (����������� � � � �� '������ � � ��� � � � �� � ��� ��� ' � �� �� / � ���� � /�� / �� ���� �� .� � ' � � $�� �(� ��� �� � /� �� � � � �� 1� �� � ��� �$ � �� � ����� ��� �� �� � � /� �� � �/� �� '������ � ��( � ��� ��� � ' � � $�� �(� � .� ��� �� �� � �� �� ��� ���( �� � � �� �� � ' � � $�� �(� � ( ��� (����������� � � � � � :���� � ��' ��� ��� � ;� � /�� � (���� �� �� ������ �� ��� .�(� ) 8������ �� � �� � /� �� � � �� = �� ���2 � $ ���� � ��� �2 � ��� � � ( ��� � � � � (����������� � 9�� � 4 � �� '� 4 � ' � � $�� �(� � � �� �a9 & � # � � 5�?� /) ! �*�*,!%= :�%�<,); *�**!,< :)�,*!; *�**) ) �*�*>)7* :�b�*),; *�**%!) :%�>b,; *�*!, , �*�*)b>* :�%�!*%; *�*****) :�*�=)>; *�*** 7 �*�*%),) :�)�=%b; *�**7,, :!�!b!; *�*** % �*�*>%%! :�b�,!*; *�**!=! :)�=><; *�**%)= :%�<7%; *�*!= = �*�*>)>7 :�>�=>,; *�**%!, :%�>7<; *�******, :*�!!b; *�*!, > �*�!!b** :�%�)b%; *�**%7> :=�*b>; *�**b)7 :)�),*; *�*!= b �*�*,*7= :�%�,>); *�**!,b :)�)b,; *�*****! :�*�%=*; *�**) < �*�*%7<) :�)�>b<; *�**!7! :)�,)>; *�**7%) :!�),); *�**) !* �*�*>=!= :�>�<>); *�**!=) :)�=b%; *�**%)< :%�<)!; *�******> :*�))*; *�*!> !! �*�!),** :�%�%**; *�**!== :)�>%=; *�**%=, :=�=)=%; *�**b%> :)�,)); *�*!< .� � ' � � $�� �(� ��� �� � /� �� � � � �� 1� �� � ��� �$ � �� � ����� ��� ��� � $ ���� � ��� � �� /� �� � � �� ���� � ��� � � �� � � ������ ��� �� �� ,* '������ � ������ � ' � ���� .� ��� �� �� � �� �� ��� ���( �� � � �� �� � ' � � $�� �(� � ( ��� (����������� � � � � � :���� � ��' ��� ��� � ;� � /�� � (���� �� �� ������ �� ��� )<= &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� $ � � � (2 �� ��� ���� � $ �� � �� � ��� �� .� � � �� �� � ������ �(� � ��� / �� ' � �� � � ��� � /) �� � �2 *�!)� ���� / ���� $ � �� ���� � ��� ���/ ��/� �� ����� '������ �� �� ( �� ������ � � ����� ����2 � � � � � (����������� �� �� ��� � � �� � � �� � (2 � � ����� ��� � ��� �� �� � ��� � 4� � ' ���'� �� ���'� � / ���� � ��� � �� �� '����2 ���� ��� � �� � � �� �� ��� ���� � $ �� � �� '������ � � ��� � ����� � � � � � � /� !� 5$ ��/ � ��� � �� � � �� � � '������ �� � � ��� ���� � $ �� � � �� � ��� �� � /� )� 5$ ��/ � ��� � �� (����������� � �� � � '������ �� � � ��� ���� � $ �� � � �� � ��� �� &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� )<> � /� ,� . � $�� � � �� � ��� '� � �� �� �� 0�� � � � � /� 7� . � $�� � � �� � ��� '� � �� �� �� (����������� � ��� �� )<b &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� � ��� � �� ' � �� � � � 4� $���� � �2 �� � ��� � � � � ���� �� � ��� �� � � �� � �� ( / ���' ��� � ��� (2 � /� � � ��� �� .� � ��� ��� �� ��� ���� �� ��� � / 0�� � �� (����������� � � � � � �� ��� � ���� � /�� , � � 7� .� � �� $ �2 ���/ $�� �� � � �$ � � � � �� ������ � �� � �� �2 ���� ' � �� �� ' � ��� 4 ��� � � ���� � ��� �� � � � � (����������� � �� ��� ��� / �� � ���� � � ��� � �� 2 �� '����2 ���� ��� � � �� � ���� ��� � �� �0� � �� � �2 *�!> $ �� �� '������ � � $ �� 4� �� � �� �2 ( � ���� �� ������ �� � 2 ����2 �� �� � �� � � �� �� �� �� (����������� � �� �� � �� ( � /��2 � ' � � � �� � � ���� � � �� � ��� �� ����2� � /� 7 �� ���2 ����� ��2 h�� � � � '' � :!<<>;� ��� �� � � ' � �� �� !<<!a !<<%� ��� � � '�� � $ (����������� � �� �� � � # � � � ���� :)***; ��� � � /�� $ (����������� � �� �� �� ' � �� !<<%a!<<<� 5 �' �� � �� �� ��� ���� � �� �2 � $ ��� �� � ��� ��� �� �� ' � � � /��2 $�� �(� ���' � �� �� � /� �� � � �� � ���� �� '������ �� � �2 �� �� b ���� �� ���/ ��$ � �� �� '� � �� � $ ���� � ��� � � �� ������ ����2 ��� �� �� � ������ .� ��� ���� � $ �� � �� � ��� � ��� �� � � ���'� ��� )=n � �� � ��� � $��2 / ���� �<>�<n �� �)==n� 4� ����� ��' ��� ��� � ������ ��� �� ( � � �� � �� $�� �( � �2 �� � $ ���� � ��� � ����� ��$ ( ��� � /� �� &���� ������ � �� � /��2 $���� � '� � � ���� � ����� � �� � ������ ��� 1��'� � 8� 3 �/2 � � �� � � �� � � � ��� � � �� 4�� ( �� � ��� � � /� ' � �� ��� 7b*n� (�� � �� ��� ����� � � ���� � /� � ��� � ���� � � �� � � �� � � �� � � �� �� ������ / ' � �� � ���� � � ���� � �� ( �� 77n � � =*n � �2 �� � �$ �� ���'� � �� $ � ���� � ��' ��� ��� � � �� � �� � � � �� � �� � � � �� � '��(� � � �� ������ 4� ��2 ( �� � ��� ��� �� �(� � �� � � � /��� � � �� ��� ���� h�� � � � '' � :!<<>; � � # � � � ���� :)***;� ���� � � �� � �� � � � � ���� � ��� ���'� 0�� ��$ / � ��� � ��� �� � � ' � ��� ���� � � ���� ��� '�� �� � � � ��� � �� ' ( �� � � � � � ��� � @�� $ �� ��� �� '������ � $ ����� � � �� '��� (� �� �� �� � �� � � � �� �� � � � ��2 �� � �� / ��$� � �� �� ������ � �� �� � � �� ( � �� � ��� � 2 � � �� � $ ��� � ��� �� � 4� ���� ���� ( (�� � � ���� � � �� / ������ ���� �� �� ���$ $ �� � � ' � �� � ��$ � �� /�� $ � ��� � �� �� � / 0��� ���� �� ���'� � �� �� 8�� � � 8� :!<<!; ���� �� ����� 0��� ��$ � � /� � � �� ��� �� � ���/ � ��� � �'���� � � �� ���� ��� � � �2 �� �� �� .� � ���� / ���$ $���� ' ( �� ����� � ���� ����� 0��� �'' �� / �� ��$ � /� � � ��� � ��� � ���� �� ��� � 4 ��� � �� ����� �� � �� � �� '��� �� � � ����� � �� �� �� � 1� �� �� � � � ��� � ���� � ��� ���� �� �� �� '� � � �� .�(� , ����� �� � � �� �� ���� � � ��� �� .�(� , & � �� �� � � � ��� � ���� ���� � i �� � !<<, !<<7 !<<% !<<= !<<> !<<b !<<< )*** 8�' ��� ��� � :���� d�; )<%,, ,*%%) ,==** %!7>) %)b7) 7b*b= 7<,*, 7%7,= 8�' ��� ��� � :k�&� d�; !=))) !<)%% ),==> ,=)*< ,!>7= )%*=< )7bb7 !<!)> .�� �$ � :���� d�; =,)) >!b) >7)) <=7, !7!*% !<)!= )!!*< ))=<! .�� �$ � :k�&� d�; ,7,, 7)<! 7b7< ==%7 <),< !**=* !!*,= !*!7> # �� � ���'� � !b< )*> )*% )*7 ))7 ))< )!b ),! &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� )<< ��' ��� ��� � � � �� ��( � �� 0��� � �� �� � /� % / $ � �� ����� ��2 � ��� � � �� ��&3 ��� ��� �� � /���� � 1 ��� / �� ' � �� �� �� ����2� � ���������� .� '��� ' ������ � �� $��� ������ �� k�&� �$ � � � � 2 ��� � �� � l �� � �� � ����� �� �� � ����2 �� �� ���� � ���� � ����� /2 �� $ �� / '������ �� �� ������ � �� � /� (����������� �� � � � (��� � / ���� '������ �� � � =��� �� (�� �� ����� ��$ �(� � � $ ���� �� ��� $ � � ��� 1� �� �� �� ���� �� @�� $ �� �� � ����� ����� ��� � � ��� �����/ � � �� � 2� �� � �'��� / �� / �� �� � � � ��� � ���� � �� ����� � .� � /� $�� �( � �2 �� � ��� � � � ' � ���(2�' � �� (�� � ��� � � 1�� � �2 � ��2 '��'�� � � �� �� ���� �� � $ �� 0��� � � �� ��� � (2 �� ����� ���� � � � ��� � � ��� � 1�� � �2 � � � $ �� '������ � � � �� � � �� � � ��� / �� ��( � �� '������ �� �� � �� ���� ������ �� � � /�� �� � $�� �� � � ��� �� �� � ���� � ��� � �� '�� � �� ( 0�� �� �� ����� /2� 5 ��� � '��(� � ���� � � �� �� �( � �2 �� �� $ ��� � ����� / � (�� � � (������� ���� � � �� � � �� 0���� �� � �� � � � � ��� ���� d%<, � �� � � �� � � � � � �� 0��� �� ���'� ��� ?��� ���� d!,% � �� � � 4� � �($ ����2 � �0���� �� (�2 � � � �� ���/ $���� � �� ���� � � �� ���� ��� ��' ��� ��� � � � ����� ��$ / � � / 0�� � �'��� � ���� '� � � .� � � �$ � ���� ����� / � ���/ �2 �� � ��� �� � $ ���� $ ������ �� �� � �� �/ �� ����� ( � ����� � �� �� ������ '������ �� � �� ����� ������ ���� ��2 ( � �0���� �� '��� � � �� 2 �� � �� � � �� '������ � � � �� � �� � � �� '��� ������2 � �� ����� /2 ��� �� ( � � �2 ���'� �� 5�/�� �� ���� (����������� � � � � � �� ��� �� 1'�� ��(� $ ���� � � �� ���� � �0� �2 ��2 �''�2 �� ���/ � ���� ��� (�� �� 2 �� �� � $� � � � � ��� �� 4 ����� � � �� �� � ���� �� �� � � � ��2 ���� $ ���� � � �� ���� � �0� �2 �� ( �� �� $ �2 1'�� � �� "� � �'��� ���� �� � �0����2 �� ������ / '������ �� �� � / 0�� � � � �� �� � /� %� ��� � � �� � � � ��� � ��� ��� �� � /���� � 1� ,** &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� �� �� �� � �2 �� ����� ������� �� ����� ��( � �� ���'� � � �� � ��� � � �'��� (� �� �� ������ '������ �� ���� '��$ � � ��� �������2 � /� �� � $ �� 0��� � � .� 1� �� � ��� � / ��� �� $ �$ � � ��� $ �2 �� / $ ��� � ��� �� �� �� �� ���� �� ����2 � / 0�� � ( ���� �� �� 1�� � $�� �� � � ��� $ � ( �� ' � ���� � � �� ��� �� �� (�� � ���� (����������� � �� � � ��� �� �� $�� �� � � ��� � ���� ������ .� � � / �� �� �� ����� /2 � ����� � � � �� � � 0 ���2 � ���� � ���� ��$ � ���/ l� � � �� � ������ � � �� � � � � $ �� 0��� � �� � ��� �� �� �� �� � � �( � �2 �� � ��� �� �� � � ��� �� � � �� � � �� ��� ��� �� / �� ������� �� ��� � �� $ ������ .� � ' ������ � � � ���� � �$ � � ��� $ �2 ����� ' � ��� �� � � � � �� 2 �� �� �� �� $ ���� �� 0� � �� ��� �� � �� �� /� �� � ��� � � ��� $ �� �� ���� �� ���������� �� �� � �� :!<b!;� .� � ��� � �� ' ( �� � ��� � � ���� � $��� �� ����� ������� & ���" -�������" '� � ����� �� ,a!b� ����� &� :!<b,;� .� � ��� � �� ' ( �� �� /�g 2 ��� ���� � $��� � � � ��� ��� �i&3 ����� ������� & ���" -�������" '� � ����� ��� !)<a!%=� ������ �� :!<>);� 8�' ��� ���� � �� � (� �� � �� � ��� �� � (����� /� & ���" � ������� �5� 777a7%%� ������ �� :!<<,;� � �� � � � ��� � & ���" � �� "� ����������� ��� ba!b� ��2� �� �� &�� o 3� �����'�� h� � :!<<>;� 8������ �� � �� � � �� � �� �� � � � ��� � ���� ���� � � ��� �� ��� ����� ��# -������� 36� !b!a)*%� 8�'���� 8�� ��� 2� 4�� o &���' � �� �� :!<<,;� 4 � � �� � �� $��� � � /����� ����� � ��� �� -�������" ���"���� & ���"� )>a,=� 8�� � j� 8�� o 8� � �� :!<bb;� 5 � �� $ � � �� ��� � '� � / � �� � � �� ��� �� 0�� � � �� � ����� ��� $�� �(� ��� � ��� & ���" -������� �3� ,*<a,)%� 8�� � j� 8�� o 8� � �� :!<<!;� &��������� � � � ��� ������� � �� �� �� ����� � � ���/ 0���� & ���" -������� �7� !7=>a!7b7� 8�� � #� j� 8�� @����� i�� o #��� ����� �� :!<<!;� �� ��� ���� � � ����� � ��� � ��'� � & ���" -������� �7� !>,<a!>b<� -� �� j�� o . ��� � &� :!<<>;� 3$ � � � �� ������� � �� �� �� ������� �� � �� $�� �� � �� ����� � ��� �� & ���" -������� 5�� !a,,� -�$ �� �� #� :!<<7;� .� ������� �� � �� � �� � � ����� � ��� �� .� '� 8m9�k&.5. $ � � � & ���" -������� ��� !%<,a!%<>� -�$ �� �� #�� ����� 3� ��� o �� ��� j� �� :)***;� 8������ � �� ��� ��$�� � � � � � �$ ��/ � ��� �� !<)>a!<<>� & ���" -������� 55� ,b<a7*=� �� ��� �� �� 9�� o .��� �� � @� :!<b%;� -� � �� ����� ���� � �$ �� ���p & ���" -������� ��� ><,ab*%� �� ��� �� �� 9�� o .��� �� � @� :!<b>;� ����� � $ � � � $ ���� �$ �� ��� � � � ����� ���� � � ��� �� �2� & ���" -������� ��� %%>a%b!� ����� 3� ��� o �� ��� j� �� :!<<);� .� ������� �� � �� 1' �� � ����� � ��� �� & ���" -������� �6� 7)>a 7=%� ����� 3� ��� o �� ��� j� �� :!<<,;� 8���� � �� ������� �� � ��� � � ������ � � (� ��� & ���" -�������" '� � ����� 33� ,a%=� ����� 3� ��� o �� ��� j� �� :!<<%;� & � � � (����������� � ������� �� /� � � � ��� �� & ���" -������� 5�� !,!a!%%� ����� 3� ��� o �� ��� j� �� :!<<=;� 9��� ������ 1'�� �� � � �� ��� � '� � / � ���� �� & ���" -������� 5�� %%a<7� ����� 3� ��� o �� ��� j� �� :!<<b;� h��� $ ���� /������ .� � � �� � �� $ � � � & ���" -������� 53� !<>%a!<<<� &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� ,*! 6������ ��� o -���� -� #� :!<,7;� ��� ���� ���"����� � � i���� 9�6����@ ��� @��/ � � :!<<%;� 8�� ��$ 9�����) ��� ���� ������� � 9����� ���!���� 3 /� ���� 8� ���� ��� �� � � �@���� j �� -� �� :!<b,;� & � �� ��� � � ���� � � � ����� � ��� � ��� �� �2� ����� � �' � ��� $ � � � & ���" -�������" '� � ����� ��� !,a,)� #��� ����� ��� &�� � �� 5�� o h �� 2� � �� :!<<7;� 8� ���� � $ ��� �� 1���'���� � � � � � ��� & ���" -������� ��� !%7!a!%>b� #������� � :!<<=;� 31' ���� � � � � �� ������� �� � �� ����� � ��� �� & ���" -������� 5�� !>!%a!>7!� # � "�� o � ����� �� �� :)***;� 8�� ����������� ��� 4 !+� +���!�� � ��� �� ��$ %��"��#� & ���� � k&5� 5��� �2 �� � � � �� & �$ � � 8� � � � � # � �� �� :!<=%;� .� $����� � �� � �� ��� �� � � �� � � �� � �� � ��2 $ ��� �� ����� '������ �� � � ��' ��� (��/ ��� /����$ '� � ���� ��# ����������� �6� !,a,>� #�� 5� ��� o 9��j ��2� 5� 8� :!<<*;� -����� ��' / ( �� � � ��� �� 0 � � �� ��� � '� � / ��� ��� /����$ -�������" �� #���� 3� 7,!a7=>� 9��j ��2� 5� 8� :!<<%;� 9��� ������ ��� �� �� �� 1'�� � $ �� � � ���� �� 85�9� & ���" -�������" '� � ����� 3:� ,a)b� ���� &� 5� :!<>=;� .� ��( ���/ �� ��2 �� ��' ��� ��� � '� � /� & ���" '� � ��� 8�� ��� �3� ,7!a,=*� &���' � �� �� :!<=7;� 8�' ��� ��� � '� � �� 5 �� ��2 �� ���� � �� � (� �� � � �� � � � � �� � ��� & ���" -������� ��� 7)%a77)� h��� 3�� o � '' �� �� :!<<>;� .� � � � � (����������� � �� ��� � � � ��� �� ��$ %��"��# ;��������� ���"���� ,%a7%� ,*) &,-, ��� "# �� �", . -�������" �������� /����$ �� 0����1 ���23�� book-to-market and size as determinants of returns in small illiquid markets: the new zealand case introduction literature review methodology results discussion references financial services review volume 33 number 2 (2025) volume 33, no. 2 2025 editor: john e. grable, ph.d. cfp ® university of georgia advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting conrad ciccotello, jd, ph.d., university of denver sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp ® , baylor university (emeritus) martin seay, ph.d., cfp ® , kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp ® , san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama lu fan, ph.d., cfp ® , university of georgia jasmine fang, ph.d., massey university, new zealand mark fedenia, ph.d., university of wisconsin philip gibson, ph.d., cfp ® , winthrop university stu heckman, ph.d., cfp ® , texas tech university william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea izidin el kalak, cardiff university, united kingdom thomas langdon, ph.d., roger william university, bristol, ri mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden lance palmer, ph.d., cfp ® , cpa ® , university of georgia abed rabbani, ph.d., cfp ® , university of missouri chris robinson, ph.d., york university (emeritus), canada elisabeth sinnewe, ph.d., queensland university of technology jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida issn online 1057-0810 print 1873-5673 contents grable, john e., from the editor, i-ii. invited paper hanlon, robert h., leher, paul, cohen, alexander, miller, eric, hancock, monte, & mitchell, robert, psychophysiological finance and intelligent wellness: a new financial planning practice model, 1-14. regular issue papers birkenmaier, julie, & stratman, hope, consumers' basic bank account complaints and their financial hardships: a content analysis of complaints filed with the consumer financial protection bureau (cfpb), 15-35. anderson, jason n., sanchez, donovan, gallardo, juan e., lawson, derek, & ouyang, congrong, exploring the effect of federal student loan payment resumption on borrowers through sentiment and textual analysis using x, 36-54. sanders, kaplan, & olajide, olamide, consumer margin use: understanding the role of peer influence, investment literacy, and age, 55-73. heckman, stuart, letkiewicz, jodi, & lim, hanna, student willingness to borrow for higher education, 74-92. antonoudi, efthymia, kostandini, genti, & lim, hanna, immigration law enforcement and immigrant homeownership, 93-123. smith, david, & curnutt, gary, the association of cryptocurrency and the use of alternative financial services, 124-143. ahmmed, ferdous, kalenkoski, charlene marie, & browning, christopher m., examining the gender gap in participation in employersponsored retirement plans: oaxaca decomposition, 144-164. stapes, tanya, rollins-koons, ashlyn, & mccoy, megan, a structured literature review on equity in the financial services profession: unpacking gender barriers and advancing women's participation globally, 165-188. academy of financial services officers michelle cull western sydney university president kirsten macdonald griffith university executive vp-program thanh ngo east carolina university vice president finance mustafa nourallah mid sweden university vice president international relations thomas korankye university of arizona vice president marketing & public relations elisabeth sinnewe queensland university of technology chair, australia-new zealand chapter shawn brayman financial planning research consultant immediate past president editor, financial services review john e. grable, ph.d., cfp® university of georgia directors norah feng massey university aaron gilbert auckland university of technology matt goren brett danko education center tom idzorek morningstar eun jin kwak university of wisconsin-green bay dan moisand moisand fitzgerald tamayo richard stebbins university of alabama tom warschauer san diego state university dave yeske golden gate university yu (yulia) zhang kansas state university past presidents shawn brayman, 2023-2025 financial planning research consultant inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 33, no. 2, 2025 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • conrad ciccotello, jd., ph.d., university of denver • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • shinae choi, ph.d., university of alabama • lu fan, ph.d., cfp®, university of georgia • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • philip gibson, ph.d., cfp®, winthrop university • stu heckman, ph.d., cfp®, texas tech university • stephen horan, ph.d., cfa®, university of north carolina wilmington • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • izidin el kalak, cardiff university, united kingdom • thomas langdon, ph.d., roger william university, bristol, ri • mustafa nourallah, centre for research on economic relations, mid sweden university, sweden • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., cfpretired™, cpa,ca, york university (emeritus), canada • elisabeth sinnewe, ph.d., queensland university of technology, australia jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • axton betz-hamilton, ph.d., south dakota state university • alona bilokha, ph.d., university of north florida • brian l. boscaljon, ph.d., penn state behrend • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph..d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph..d. stetson university, deland, fl • patti fisher, ph.d., virginia tech • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • david nanigian, ph.d., cfp®, mount ararat financial services llc • john nofsinger, ph.d., university of alaska anchorage • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • congrong ouyang, ph.d., kansas state university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peerreviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(2) 165 a structured literature review on equity in the financial services profession: unpacking gender barriers and advancing women’s participation globally tanya staples,1 ashlyn rollins-koons,2 and megan mccoy3 abstract women’s involvement and influence in the financial landscape have risen markedly, with a growing share of global wealth now under their control. if current projections prove accurate, women are expected to manage about 55% of the world’s wealth by 2030, fundamentally transforming the financial services and advisory industries. alongside this shift, a more holistic, solution-focused, advice-oriented approach is emerging—departing from the historically productcentric, male-dominated financial sales industry of the past. despite this progress, gender equity within financial services remains elusive. women currently comprise only about 17% of financial advising professionals in canada and the united states. research consistently underscores the importance of gender diversity, noting that many female clients prefer advisors who understand their distinct needs. yet systemic, cultural, and societal barriers continue to limit women’s full participation in the profession. this structured literature review examines these barriers, particularly within the realms of financial advising and planning. it also explores the implications for policymakers, practitioners, and researchers, offering strategies for employers and the broader profession to enhance organizational structures. emphasis is placed on transparency and the development of policies and procedures that actively integrate a gendered perspective. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation staples, t., rollins-koons, a., & mccoy, m. (2025). a structured literature review on equity in the financial services profession: unpacking gender barriers and advancing women’s participation globally. financial services review, 33(2), 165-188. introduction in recent years, women have emerged as significant players in the financial landscape, controlling a substantial portion of wealth globally. by 2020, women held approximately 1 corresponding author (tmshall24@ksu.edu). kansas state university, manhattan, ks, usa. 2 kansas state university, manhattan, kansas, usa. 3 kansas state university, manhattan, kansas, usa. 32% of all wealth (gandhi, 2020). it is anticipated that by 2030, women will control approximately 55% of the world’s wealth. this increase in wealth is anticipated to influence and reshape the https://creativecommons.org/licenses/by-nc/4.0/ mailto:tmshall24@ksu.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(2) 166 financial management and advice industries (bloomberg, 2024; ubs, 2021). despite this progress, gender equity within the financial services profession remains elusive. while women — especially younger women — increasingly make household financial and wealth-related decisions, the profession's composition does not reflect this shift. historically, financial planning was predominately about men selling financial products to men (clempner et al., 2020; steed, 2019). the industry focused on sales and technical ability to develop and present productbased financial solutions (mandell, 2008; richards, 2021). although the profession is thoughtfully shifting toward holistic, solutionoriented financial advice—an approach women are especially well suited for (richards, 2021)— the traditional, product-focused sales model remains entrenched. legacy thinking around the profession and larger industry persists worldwide (pasztor et al., 2019). the persistently small number of women certified financial planner® (cfp®) in canada (see table 1) and the united states (u.s.) does not bode well for the profession's future. table 1. gender of cfps® professionals and percentage change of women cfps® (2008 – 2023) date* men women unknown sum % change 2008 11663 5467 52 17182 2009 11737 5466 55 17258 -0.02% 2010 11774 5487 59 17320 0.38% 2011 12325 5738 66 18129 4.57% 2012 12058 5578 65 17701 -2.79% 2013 11942 5559 68 17569 -0.34% 2014 11767 5461 72 17300 -1.76% 2015 11663 5348 74 17085 -2.07% 2016 11581 5253 83 16917 -1.78% 2017 11568 5205 84 16857 -0.91% 2018 11506 5182 90 16778 -0.44% 2019 11485 5135 94 16714 -0.91% 2020 11632 5220 99 16951 -1.66% 2021 11648 5214 102 16964 -0.11% 2022 11772 5375 103 17250 3.09% 2023 11862 5467 104 17533 1.71% *year end (m. geramas, personal communication, february 3, 2023) only 17% of all financial advising professionals (faps) are women (moreau, 2022), and the representation of women among cfp® professionals remains stagnant (31.2% in canada and 23.7% in the united states) (cfp board, 2023; fp canada, 2023). eleanor blayney, cfp board consumer advocate, outlines the dangers of the "feminine famine" in the profession (keller, 2014, n.p.), which leaves many women clients underserved by the profession. companies with the highest representation of women in senior management have demonstrated higher profits and higher returns on invested capital (msci, 2016). diverse, equitable, and supportive workplaces make good business sense. research underscores the importance of gender diversity in the financial planning profession. many women clients prefer working with women financial advisors or someone who understands their needs and can engage in a relatable manner (sjogren & allan, 2020; strategic insights, 2017). moreover, widows, often facing significant life transitions, seek a holistic trust staples et al. 167 based planning approach. if those needs are unmet, 80% of widows move their assets to another advisor within a year (baghai et al., 2020; kurlowicz, 2014; reiter et al., 2021). women cfp® professionals, more than their male counterparts, tend to focus on comprehensive approaches to planning and helping clients throughout the planning lifecycle (sheedy, 2021). conversely, male planners often emphasize investments and wealth accumulation. however, barriers persist. systemic, cultural, and societal impediments hinder women's full participation and advancement in financial services. a profession that is not open to everyone cannot serve its diverse needs (reiter et al., 2021) in 2013, the cfp board commissioned a study to explore the causes of the under-representation of women in the financial planning workforce (blayney, 2016). blayney identified five broad categories of barriers: structural, societal, prejudice-related, absence of women-focused leadership development, and misinformation. these barriers perpetuate inequity in mentorship, sponsorship, networking, compensation models, and gender prejudice, limiting the success of women in the profession. this paper reviews the literature on these global financial services industry barriers. the goal was to focus specifically on financial planning, but the need for more research on women's advancement in financial planning necessitated a wider lens. the research question shaping this review is: what are women's barriers to entry and advancement in the financial services industry? after examining existing literature, this paper discusses implications for policymakers, practitioners, and researchers and makes recommendations for future research. methodology the motivation for the research was to investigate why there has been no increase in women cfp® professionals during the research timeframe. this structured literature review of gender inequity and the financial services profession was limited from 2008 until march 2023 to align with our research period of financial planning in canada for 15 years. the scope of the review was academic journal articles. the following search terms were used in google scholar following keywords: "gender diversity" and "women" and "financial planning" to begin identifying studies that addressed the barriers of interest in this study which yielded 852 results. when these phrases were paired with "canada" in a revised search, the results dropped significantly to 268, highlighting a notable gap in the literature concerning canadian women cfp® professionals and gender equity, the original study topic. the research team also examined the following three databases: business source premier, proquest, and web of science to ensure saturation using the same search terms. table 2. structured literature review: search results description and findings revised search two words and phrases were once again combined, but financial planning was expanded to financial services and searches extended beyond canada. at this point enough articles resulted from the search 89 relevant sources revised search three we reviewed the abstract for each to ensure that the purpose of the study and the question posed by the study were aligned with our goal of examining the barriers to entry and advancement in the financial services industry for women to determine if any additional articles were missed in the keyword search 37 relevant sources the team reviewed the articles based on a predefined set of parameters that included the following: was the research question(s) clearly defined, was there use of empirical evidence, was existing literature cited, was the paper supported by a theoretical framework, were the key constructs defined, and did the publication contribute the existing literature. after removing duplicates and ensuring each article met the aforementioned criteria, we identified 89 relevant resources. of the articles that remained, the research team reviewed the abstract for each to financial services review, 33(2) 168 ensure that the purpose of the study and the question posed by the study were aligned with our goal of examining the barriers to entry and advancement in the financial services industry for women to determine if any additional articles were missed in the keyword search. this paper reports findings from 37 scholarly journal articles. table 3 provides a summary of these articles. table 3. article summary highlighting key barriers identified results of the structured literature review (n = 37) barrie r one barrier two barrier three barrier four barrier five abraham, m. (2017). pay formalization revisited: considering the effects of manager gender and discretion on closing the gender wage gap. academy of management journal, 60(1), 29-54. https://doi-org.er.lib.kstate.edu/10.5465/amj.2013.1060 x x x baghai, p., howard, o., prakash, l., & zucker, j. (2020). women as the next wave of growth in us wealth management. mckinsey & company x bergmann, n., scheele, a., & sorger, c. (2019). variations of the same? a sectoral analysis of the gender pay gap in germany and austria. gender, work & organization, 26(5), 668–687. https://doi.org/10.1111/gwao.12299 x x x bisco, j., gradisher, s., & mulholland, ba. (2019). women and diversity—why the conversation must continue in financial services. journal of financial service professionals, 73(1), 72–84. x blayney, e. (2016). what is the future of women in financial planning? journal of financial planning, 29(9), 32–33. x x x bloomfield, r. j., rennekamp, k., steenhoven, b., & stewart, s. (2021). penalties for unexpected behavior: double standards for women in finance. the accounting review, 96(2), 107–125. https://doi.org/10.2308/tar2018-0715 x x bordalo, p., coffman, k., gennaioli, n., & shleifer, a. (2019). beliefs about gender. american economic review, 109(3), 739– 773. https://doi.org/10.1257/aer.20170007 x x x carter, n., elliott, v., harding, c., hetterich, a., howes, j., jackson, c., jeffords, p., & levine, b. (2016). when women thrive. businesses thrive. mercer llc. x x x x x https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi-org.er.lib.k-state.edu.mcas.ms%2f10.5465%2famj.2013.1060%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi-org.er.lib.k-state.edu.mcas.ms%2f10.5465%2famj.2013.1060%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1111%2fgwao.12299%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1257%2faer.20170007%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d staples et al. 169 https://www.mercer.ca/content/dam/mercer/ attachments/north-america/canada/ca-2016gender-diversity-wwt-financial-servicesperspective-en.pdf clempner, j., daisley, m., & jaekel, a. (2020). women in financial services 2020. oliver wyman. https://www.oliverwyman.com/content/dam /oliverwyman/v2/publications/2019/november/w omen-in-financial-services-2020.pdf x x x x domski, s. (2018). pipeline or waterslide: advancing women into executive positions in business and financial planning. celebration of learning. https://digitalcommons.augustana.edu/celeb rationoflearning/2018/presentations/24 x x x x egan, m., matvos, g., & seru, a. (2022). when harry fired sally: the double standard in punishing misconduct. journal of political economy, 130(5), 1184-1248. x x financial planning association. (2023). peggy ruhlin discusses the state of women financial planners. journal of financial planning, 12, 20–22. x x girardone, c., kokas, s., & wood, g. (2021). diversity and women in finance: challenges and future perspectives. journal of corporate finance, 71, https://doi.org/10.1016/j.jcorpfin.2021.1019 06 x x x hart, s. a., & ganley, c. m. (2019). the nature of math anxiety in adults: prevalence and correlates. journal of numerical cognition, 5(2), 122–139. x x healy, g., & ahamed, m. m. (2019). 12. gender pay gap, voluntary interventions, and recession: the case of the british financial services sector. british journal of industrial relations, 57(2), 302–327. https://doi.org/10.1111/bjir.12448 x x x x house of commons treasury committee. (2018). women in finance. 1-51. x x x x ibarra, h., carter, n. m., & silva, c. (2010). why men still get more promotions than women. harvard business review, 88(9), 80-85. x x x ibarra, h., ely, r., & kolb, d. (2013). women rising: the unseen barriers. harvard business review, 91(9), 60–67. x x x https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://digitalcommons.augustana.edu/celebrationoflearning/2018/presentations/24 https://digitalcommons.augustana.edu/celebrationoflearning/2018/presentations/24 https://doi.org/10.1016/j.jcorpfin.2021.101906 https://doi.org/10.1016/j.jcorpfin.2021.101906 https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1111%2fbjir.12448%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d financial services review, 33(2) 170 jaekel, a., & st. onge, e. (2016). why women aren’t making it to the top of financial services firms. harvard business review. x x x klein, g., shtudiner, z., & zwilling, m. (2021). uncovering gender bias in attitudes towards financial advisors. journal of economic behavior & organization, 189, 257–273. https://doi.org/10.1016/j.jebo.2021.06.040 x x kumar, a. (2010). self-selection and the forecasting abilities of female equity analysts. journal of accounting research, 48(2), 393–435. http://www.jstor.org/stable/40784954 x kurlowicz, a. (2014). women in financial planning. journal of financial service professionals, 68(3), 56. https://er.lib.kstate.edu/login?url=https://www-proquestcom.er.lib.k-state.edu/scholarlyjournals/women-financialplanning/docview/1527455925/se-2 x kurtz, a. (2018). why is the pay gap for women financial advisors so wide? financial planning (online), https://er.lib.kstate.edu/login?url=https://www-proquestcom.er.lib.k-state.edu/trade-journals/whyis-pay-gap-women-financial-advisors-sowide/docview/2022918109/se-2 x x x x limón, a. t. (2020). addressing the realities facing diverse employees in the financial planning profession. journal of financial planning, 36(11), 23–26. x mccarthy, e. (2016). where are the women? can we close the advisor gender gap? retirement advisor, 17(7), 36-39. x x x neck, c. (2015). disappearing women: why do women leave senior roles in finance? australian journal of management, 40(3), 488-510. x x x noback, i., broersma, l., & van dijk, j. (2016). climbing the ladder: genderspecific career advancement in financial services and the influence of flexible worktime arrangements. british journal of industrial relations, 54(1), 114–135. https://doi.org/10.1111/bjir.12048 x x x x o’dwyer, m., & richards, d. w. (2021). occupational boundaries: gender capital and career progression in the financial x x x x https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1016%2fj.jebo.2021.06.040%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2fscholarly-journals%2fwomen-financial-planning%2fdocview%2f1527455925%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2fscholarly-journals%2fwomen-financial-planning%2fdocview%2f1527455925%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2fscholarly-journals%2fwomen-financial-planning%2fdocview%2f1527455925%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2fscholarly-journals%2fwomen-financial-planning%2fdocview%2f1527455925%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2fscholarly-journals%2fwomen-financial-planning%2fdocview%2f1527455925%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fwhy-is-pay-gap-women-financial-advisors-so-wide%2fdocview%2f2022918109%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fwhy-is-pay-gap-women-financial-advisors-so-wide%2fdocview%2f2022918109%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fwhy-is-pay-gap-women-financial-advisors-so-wide%2fdocview%2f2022918109%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fwhy-is-pay-gap-women-financial-advisors-so-wide%2fdocview%2f2022918109%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fwhy-is-pay-gap-women-financial-advisors-so-wide%2fdocview%2f2022918109%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://doi.org/10.1111/bjir.12048 staples et al. 171 planning industry. financial planning review, 4(2). https://doi.org/10.1002/cfp2.1123 ogden, s. m., mctavish, d., & mckean, l. (2006). clearing the way for gender balance in the management of the uk financial services industry: enablers and barriers. women in management review, 21(1), 40– 53. https://doi.org/10.1108/0964942061064340 2 x x x reiter, m., & kiss, d. e. (2021). efforts in diversity and recruiting in financial planning undergraduate programs. family and consumer sciences research journal, 49(3), 238–253. https://doi.org/10.1111/fcsr.12389 x reiter, m., seay, m., & loving, a. (2022). diversity in financial planning: race, gender, and the likelihood to trust a financial planner. financial planning review, 5(1), e1134. https://doi.org/10.1002/cfp2.1134 x reiter, m., seay, m., macdonald, m., lutter, s., & loving, a. (2022). are there racial and gender preferences when hiring a financial planner? an experimental design on diversity in financial planning. journal of financial counseling and planning, 33(3), 344-357. x richards, d. w., roberts, h., & whiting, r. h. (2020). female financial advisers: where art thou? australian journal of management, 45(4), 624–644. https://doi.org/10.1177/0312896219896389 x x x seeber, c. m. (2015). for female financial planners, the glass is half full. journal of financial planning, 28(9), 20–22. x x sheedy, r. l. (2021). the importance of building a larger cadre of female cfp® professionals. journal of financial planning, 34(8), 62–65. x sommer, matthew, lim, h. n., & macdonald, m. (2018). gender bias and practice profiles in the selection of a financial adviser. journal of financial planning, 31(10), 38-47. https://er.lib.kstate.edu/login?url=https://www-proquestcom.er.lib.k-state.edu/tradejournals/gender-bias-practice-profilesx https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1002%2fcfp2.1123%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1108%2f09649420610643402%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1108%2f09649420610643402%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1111%2ffcsr.12389%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1002%2fcfp2.1134%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1177%2f0312896219896389%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fgender-bias-practice-profiles-selection-financial%2fdocview%2f2131783502%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fgender-bias-practice-profiles-selection-financial%2fdocview%2f2131783502%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fgender-bias-practice-profiles-selection-financial%2fdocview%2f2131783502%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fgender-bias-practice-profiles-selection-financial%2fdocview%2f2131783502%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d financial services review, 33(2) 172 results this structured literature review on gender inequity and the financial planning profession found evidence suggesting five barriers to women's entry and advancement in the financial services industry. specifically: 1. conscious and unconscious gender bias and discrimination. 2. absence of targeted leadership development programs for women. 3. organizational supports do not reflect women's needs and experiences. 4. out-of-date compensation models and policies contribute to a gender pay gap. 5. lack of information, misunderstanding, and myths about the profession. each of these barriers is described, as well as the theories that attempt to explain why these barriers exist. barrier one: conscious and unconscious gender bias and discrimination incidences of gender bias & discrimination our review yielded numerous articles on conscious and unconscious gender bias and discrimination against women in financial planning (blayney, 2016; klein et al., 2021; richards et al., 2020). kurtz (2018) detailed a class action lawsuit led by 17 women alleging widespread gender discrimination and abuse, whereby men were awarded lucrative accounts at a disproportionately higher rate than women, regardless of account origin (kurtz, 2018). rachel arthurs left her firm after more than 20 years, as she had witnessed the allocations of significant cases to her male colleagues. they argued that she could afford to make less as a single, childless woman (domski, 2018). in a recent u.s. survey, 66% of men did not believe that the financial services playing field is uneven, while 50% of women did (in research, 2017). men (43%) did not feel diversity is as crucial to company success as women (62%). nor did men (45%), as opposed to women (68%), feel that diversity is a key factor to overall industry success (women in advice, 2017). the cfp board's women's initiative (win) study (2014) found that 91% of the financial advisor community in the united states felt that men had an edge over women in the profession when it comes to the skills and characteristics necessary to be a great financial planner (blayney, 2014). this perception perpetuates gender inequity within the profession. another study examined whether women advisors in the united states were punished for violations relating to customer disputes, regulatory infractions, and criminal offenses at higher rates and more severely than their male counterparts (egan et al., 2021). the study found that women were judged more harshly than their male counterparts for the same infractions or mistakes, and this disparity was particularly apparent when managers and executives were predominately male, even when the violations committed by men were more costly to the firm. these biased judgments had lasting effects; women were less likely to be promoted or find other positions. as a result, they were reducing the likelihood of advancement (egan et al., 2021). selectionfinancial/docview/2131783502/se-2 tharp, d. t., parks-stamm, e. j., lurtz, m., & kitces, m. (2022). exploring gender differences in marital and parental income premiums among financial advisors. journal of family and economic issues, 43(1), 15–35. https://doi.org/10.1007/s10834-021-097664 x https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fgender-bias-practice-profiles-selection-financial%2fdocview%2f2131783502%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fer.lib.k-state.edu.mcas.ms%2flogin%3furl%3dhttps%3a%2f%2fwww-proquest-com.er.lib.k-state.edu%2ftrade-journals%2fgender-bias-practice-profiles-selection-financial%2fdocview%2f2131783502%2fse-2%26mcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1007%2fs10834-021-09766-4%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d https://mcas-proxyweb.mcas.ms/certificate-checker?login=false&originalurl=https%3a%2f%2fdoi.org.mcas.ms%2f10.1007%2fs10834-021-09766-4%3fmcastsid%3d15600&mcascsrf=76087ceb96069e31ed7dcfb30d3094a6c9ff0364b8a6f6d11c39b214941d760d staples et al. 173 perceptions and socialization conflicting views exist on gender diversity’s importance in the financial planning profession. within the cfp® community, 69% of planners felt that men had a tremendous advantage over women in the financial planning profession (blayney, 2014; fondulas et al., 2014; pasztor et al., 2019). however, a cfp board survey showed that 51% believed skills were gender-neutral, while 49% believed that men possess those skills more than women (blayney, 2016). socialization experiences and biases have been found to reinforce stereotypes, perpetuating unconscious bias against women planners (carter et al., 2016; oldford & fiset, 2021; richards et al., 2020; sommer et al., 2018). western society has viewed women to be less skilled in investment management than men, especially if the amount of money invested exceeds $100,000 (klein et al., 2021). however, women portfolio managers and analysts have outperformed their male counterparts during certain market conditions because they tend to adopt a buy-andhold strategy rather than frequent trading, which men prefer (kumar, 2010; ritholtz, 2016). however, entrenched beliefs have favored men while hindering women's recognition and advancement (filion & paradi, 2016; klein et al., 2021; richards et al., 2019). client biases biases have impacted clients, too. a 2008 study asked 500 israeli respondents about characteristics that alter investment planner choice (klein et al., 2021). the study concluded that gender bias is not apparent at investment houses, regardless of the investment amount. however, when clients chose to invest through banks rather than investment houses, there was a noticeable gender bias against women (klein et al., 2021). respondents with more significant sums of money invested in banks perceived women advisors more negatively. corporate culture and recruitment practices often perpetuate unconscious bias. usually, corporations use a homogeneous recruitment process, hiring employees similar to the current ones and stifling diversity (bloomfield et al., 2020; house of commons treasury committee, 2018). fewer women in the financial planning hiring pipeline compound this bias and recruiting them takes longer. companies do not want to overspend, so they have hired the first available candidate (house of commons treasury committee, 2018). the underrepresentation of women in the industry has perpetuated the biases or stereotypes that women do 'not fit in' and are less knowledgeable than men, so they do not join or stay in the profession (richards et al., 2020; women in finance, 2017;). studies have shown that men often apply for partially qualified positions, aggressively promoting their qualifications and past accomplishments. at the same time, women tend to refrain from such practices (mohr, 2014), leading to the development of a masculine and competitive culture in some organizations (richards et al., 2020). women working in the australian financial industry reported that eliminating this environment would increase women's retention as financial advisors (richards et al., 2020). self-advocacy, confidence, and advancement self-advocacy is another barrier to gender equity. women in financial services have rarely engaged in self-advocacy, impacting their visibility and success (benjamin, 2018). social learning theory suggests that differences in such experiences have caused women to question their professional abilities and job-related skills, which could negatively impact the self-confidence needed to advance (bandura, 1977, 1978, 1982; betz & hackett, 1986; manolova et al., 2007). the cfp board noted that the lack of self-advocacy makes women less visible, less successful, and less likely to motivate other women to join the profession, resulting in one of the top reasons women have been underrepresented in financial planning (benjamin, 2018). communicating career aspirations is crucial for career advancement for women (richards, 2021). confidence and the belief in abilities are essential in sales-driven environments (richards, 2021). confidence and masculine gender capital have been associated with financial service success (o'dwyer & richards, 2020). both masculine and feminine gender capital describe a skill set defined by the associated gender but not reliant on biological determination. feminine gender financial services review, 33(2) 174 capital has been associated with providing care, detail-oriented work, and organizational proficiency. as a result, self-advocacy and promotion tended to be associated with men, requiring women to work even harder to demonstrate having these much-needed professional qualities (richards, 2021). however, assertive self-promotion by women often leaves a negative impression (cooper, 2013; erfani et al., 2023; o'dwyer & richards, 2020). the lack of self-advocacy and confidence has contributed to the unconscious barriers and stereotypes in the financial services industry. these perceived barriers keep women from advancing at the rate and to the level of their male counterparts (benjamin, 2018; cfp board, 2015). barrier two: absence of targeted leadership development programs for women lack of targeted leadership programs and examples a literature review on leadership development programs indicated that a lack of targeted leadership development programs for women within the profession was a huge impediment to the advancement of women to more senior and executive positions within the financial services industry (khoury et al., 2023). canadian banks have led the way relative to their counterparts in the united states and europe concerning women's representation at the executive level. in 2021, women's leadership in canadian credit unions and banks lay at 44% and 33%, respectively, surpassing the 26% observed in europe and the united states (catalyst, 2020). by 2022, bank board representation in canada reached 44%, although the ceo position remained principally occupied by men. only 18% of canadian bank boards had women ceos (khoury et al., 2023). some canadian banks, like the bank of montreal (bmo) and royal bank of canada (rbc), have created formal leadership development programs. bmo's enterprise sponsorship program focuses not exclusively on women but also diverse individuals, including women. these protégés partnered with senior leaders benefit from advocacy, exposure, and introductions to develop meaningful connections that may result in career development opportunities made by the sponsor (canadian bankers association, 2023). the program at rbc, women in leadership, focused explicitly on women, providing a 10-month development and networking program designed to support women advancing within the organization. the lack of targeted leadership development programs, the lack of women role models, and the inadequate management pipeline of women have negatively impacted gender equity at higher levels of the organizational chart of financial services firms (domski, 2018). even when innovative programs and structures were put in place to advance women, not addressing the unconscious bias against women, the presence of sexism, and the underlying paternalistic social constructs, career advancement efforts for women were likely to be futile. there had to be a genuine desire to advance women to make efforts effective (jaekel & st. onge, 2016). cultural barriers and limitations according to jayne-anne gadhia, ceo of virgin money, another bias was the gender imbalance and what it perpetuated (house of commons treasury committee, 2018). women did not want to be involved in senior-level financial services because of the culture; it was white, male, and old (house of commons treasury committee, 2018). pricewaterhousecoopers (pwc) stated that "gender imbalances are in themselves a workplace culture that acts as a reinforcing barrier to women" (house of commons treasury committee, 2018, p.10). this reputation discouraged women, as they did not feel it offered the career advancement they desired (klein et al., 2021). in 2016, the brandon hall group found that over 75% of participating organizations had no mentoring program for women to advance to leadership positions (domski, 2018). in research conducted by management consulting firm oliver wyman, respondents repeatedly confirmed, "all our senior leaders are older, white males. they are the ones who set the culture we experience every day" (jaekel & st. onge, 2016, para 5). the financial planning/advising role centered on communication, relationship management, and staples et al. 175 assisting clients with their financial needs (brimble & murphy, 2012), all which women reputedly did better than men (kurlowicz, 2014; richards et al., 2020). there needs to be a redefinition of talent if ready and available continues to define talent. according to jon terry of pwc, if he saw only men, like him in senior positions, then that is what he was likely to select for the available position (house of commons treasury committee, 2018). representation in leadership roles the lack of equitable advancement for women was not the result of poor corporate culture alone. women themselves might have been unknowingly discouraging the likelihood of advancement. social learning theory suggests that the education system contributed to the likelihood that young women and girls would select courses and programs that did not lead them to senior positions within the financial services or the financial services industry (bandura, 1977). post-secondary programs might have been better for promotion to the c-suite level. many women received degrees in human resources, health services, and the arts instead of finance, technology, or the sciences (cappelli & hamori, 2005). the choice of degree impacted the advancement to ceo, as ceos typically have finance and accounting backgrounds (laff, 2007, as cited in domski, 2018). women have had higher representation in leadership roles that did not typically lead to the promotion of ceo, such as heads of talent and marketing/business development (cappelli & hamori, 2005). experience in operations or operations management is another predictor of successful candidates for a ceo position. at least 25% of current ceos held the position of chief operations officer (coo) prior to succeeding in the ceo position. however, only 11% of women have held positions in operations leadership (catalyst, 2020), keeping them out of the ceo pipeline. while gender equity progress has been made in the workplace, significant barriers to women’s advancement into leadership roles persists. thirty percent of organizations surveyed by the brandon hall group noted a "lack of expressed desire/assertion among women to ascend to a top executive level" (brandon hall group, 2016, p. 17). it is critical to professional advancement to make professional goals known, which men tend to do better than women. not only were women less likely to engage in self-advocacy, but they were also less likely to embellish accomplishments and experience. women also tended to apply for positions they were entirely qualified for and avoided positions they were not (brandon hall group, 2016; domski, 2018). barrier three: organizational supports do not reflect women's needs and experiences traditional gender roles and disparity the structured literature review supports the findings that gender was a traditional social construct that defined men as providers and breadwinners whose space was in the public domain and women as caregivers, providing supportive roles and operating in the private space of the home (blair-loy, 2003). for the women who did work in financial services, stereotypes persisted, as did gender discrimination (richards et al., 2020). those who made the decisions define what it took to rise to the top and stay there (richards et al., 2020). on the surface, there has appeared to be relative gender equity at the entry-level of the financial services industry. at the junior advisor level in the united states, relative equity existed between men and women. in addition, within major canadian banks, there appeared to be a firm commitment to diversity and inclusion for women at canada's 'big six' banks (canadian bankers association, 2023). while the literature indicated gender equity at the entry-level, the reality was that many women financial planners worked outside canada's 'big six' banks in credit unions, insurance companies, and private brokerage firms. as a result, these women were not included in the banks' stated commitment to equity and were absent from the data (canadian bankers association, 2023; dbrs morningstar, 2023). gender inequity continues up the organizational chart of many financial services firms, including banks (girardone et al., 2021; win research, 2017). a persistent structural barrier to advancement for women in financial advising financial services review, 33(2) 176 was the need for formal networks, mentors, and sponsorship opportunities. this barrier was constrained by dated gender role expectations and an organizational culture that did not support the dual role of professional and caregiver, often played by women (richards et al., 2020). in the united states, the top barrier to the advancement of women was the desire to balance career and family (women in advice, 2017). of the women surveyed for the women in advice report (2017), 20% agreed that this was a significant barrier. work-life balance has historically been perceived as an issue solely for women, but that was not the case (lewis et al., 2007). men also identified work-life balance as a career barrier, so dated gender role expectations need to be updated. the desire for a career and family had to be achievable for both women and men (women in advice, 2017). work-life balance and women’s careers some organizations in canada provide support for parental leaves in addition to those provided by the government. research indicated that the impact of a leave had adverse effects on advancement when women return. women absent from the workforce missed critical opportunities to advance through networking, mentoring, or sponsorship (women in advice, 2017). a crucial time for advancement to a more senior advisor role happens at the five-to-nineyear employment mark. this time often coincided with when women tended to have their first child (age 28 30), effectively removing women from the promotion to financial advisor pipeline (women in advice, 2017). this experience plagued the financial services industry in north america. in 2017, while men and women applied for promotions at comparable rates, men were promoted more often than women. the absence of women due to parental leave contributed to promotion disparity in favor of men. literature indicated a 24% gap in the rates of first promotions favoring men over women in north america, although asking for promotions was at comparable rates (catalyst, 2020). return-to-work support could cause further counterproductive results as it had tended to provide gradual transitions with modified workloads rather than the structure and process that helped women reintegrate into the advancement pipeline (women in finance, 2017). in australia, non-traditional flexitime, part-time work arrangements, and working from home are often seen as concessions for women, resulting in questioning women's commitment and contribution to the business. women might not receive bonuses, and their professional advancement might be stalled (brown, 2010). challenging career moves that require long hours, travel, and relocation for promotion might not be feasible for women in the role of caregiver. missing these opportunities for advancement could limit the experience and qualifications needed for future career progression (adams, 2014, as cited in domski, 2018). mentoring, sponsorship, and networking challenges organizations that did not make promotions based strictly on merit did not provide equitable access to advancement opportunities. according to khoury et al. (2023), canadian banks still needed to develop formal pipelines to ensure a sufficient supply of qualified women to fill senior positions. men and women might have received formal mentoring, but the quality of the mentoring differs. women mentors tended to have less power and influence, resulting in more promotions to male mentees (ibarra et al., 2010). in a study by domski (2018), more than 75% of participating financial services organizations had no mentoring program designed to advance women's leadership exclusively. in research, on behalf of state street global advisors (2017), reported that u.s. women faps believe that finding or having a good mentor is significant for their career success. additionally, the literature indicates that men are far more likely to have a sponsor, whereas women receive a mentor. mentoring is less formal and only sometimes creates a pipeline for advancement. unlike a sponsor, a mentor did not 'have skin in the game' in the form of influence, reputation, or corporate power (ibarra et al., 2010). alternatively, a sponsor was a person tasked with making introductions to influential persons crucial to advancement and supporting protégés in their career progression (richards et al., 2020). advocacy, advice, and identifying staples et al. 177 opportunities for protégés were invaluable for promotion and advancement (catalyst, 2020). it was also important to note that if the sponsor or mentor was not the direct supervisor of the protégé, the role was ineffective, as they were not directly engaged in the relationship, impairing or limiting effective communication (ibarra et al., 2010). the limited access to support, such as mentors and, more importantly, sponsors for women, created a feeling that women were not welcome at certain events (limón, 2020). the lack of effective networking for women financial advisors/planners looking to advance their practices could be a real obstacle. according to those interviewed by ogden et al. (2006), networking was more accessible for men for three reasons: 1. care-giving responsibilities encumbered men less and caused less disruption to their workday. 2. many networking events focus on sports and drinking, activities that circle stereotypical male interests and could actively exclude women. 3. traditional networking often takes place outside of regular working hours, which impinges on women's personal responsibilities as caregivers. networking has been an integral part of the financial planning/advising profession for a long time, and the barriers women face have limited career advancement (richards et al., 2020). srt helps think about this as traditional gender roles of women did not include a focus on drinking and sports. barrier four: out-of-date compensation models and policies perpetuate gender pay gap. financial services profession gender pay gap our review resulted in multiple articles that addressed the persistent gender-based discrimination and inequities that existed, resulting in significant gender pay gaps (abraham, 2017; catalyst, 2020; kurtz, 2018; tharp et al., 2019). data from the u.s. department of labor showed that women financial advisors had the widest wage gap of all occupations, earning only 59 cents for every dollar male financial advisors earned and approximately $35,000 less than men in total compensation (kurtz, 2018). differences in pay also existed in canada, with women earning $7.50 an hour less than their male counterparts in the finance, insurance, real estate, rental, and leasing industries (statistics canada, 2022). in these sectors, women had an average hourly wage of $35.40 compared to their male counterparts at $42.93 per hour, which equates to women making 81.6% of their male counterparts (catalyst, 2020). one reason for gender pay inequity might have been the small number of women in higherpaying management roles in financial services and the historical overrepresentation of men in those positions (abraham, 2017). in 2019, only 20% of women in financial services were in leadership positions worldwide (catalyst, 2020). this study added to the existing body of knowledge by suggesting that gender at the management level, including those making the decisions in the pay equity process, needed consideration (abraham, 2017). in 2020, even though women made up 55% of entry-level employees in the sector, only 39.5% of senior management positions were held by women (canadian bankers association, 2023). women's attrition as they climb the corporate ladder has had a compounding negative effect, with fewer women across the sector, fewer women making decisions, fewer women to promote, and fewer women mentors/sponsors from which to learn (in research, 2019). although a pay gap still existed in the financial planning profession, it was less dramatic than that in the financial sector. tharp et al. (2019) identified a 19% pay gap between male and female financial planners when they surveyed 710 financial planners. tharp et al. (2019) claimed that 91% of this 19% pay gap could be explained by the job role, experience, hours worked, professional designation status, and other factors, leaving a much smaller percentage that might be attributable to gender. while this study looked at other factors that affect pay, it did not examine them through a gender lens. regardless, the literature confirmed an actual pay gap in the financial planning profession and the financial services review, 33(2) 178 larger financial services industry, despite conflicting positions as to why the gap persists and the size of the gap. traditional compensation models the traditional compensation model in the financial services industry was performancebased, comprising commissions and bonuses. advisors have been rewarded heavily for bringing in new business and less so for maintaining existing clients. while this model was undergoing significant change with the elimination of deferred sales charge fees and removal of other 'loads' in favor of fee for service and compensation based on assets under management (aum), less than 35% of financial services organizations in the united states, reviewed performance ratings by gender and only 26% adjusted their annual compensation review process based on pay equity (benjamin, 2018). the lack of effective pay equity processes and professionally trained teams permitted the variable compensation (sales-based) environment that tended to privilege men while inhibiting women (richards et al., 2020). while most financial firms claimed to be a meritocracy, their internal data showed a gender pay gap due to non-transparent compensation structures (kurtz, 2018). the sale-based compensation structure that dominated the industry might be a barrier for women entering the financial industry as women preferred more salary-based compensation, which provides for less personal risks and greater income stability (blayney, 2014; richards et al., 2020). many financial companies did not offer a base salary but had production-based formulas or performance-based forms of compensation. a financial planner's compensation is comprised of commissions and aum. research indicated that in australia and new zealand, the precarious nature of commission and bonus compensation made women step back from the profession, and there was no reason to believe that the situation differed in the north american context. this system worked well for men, who managed significant books of business, but not for women, as women still managed only single-digit percentages of client assets (deloitte, 2019). these forms of compensation prevented women financial planners from entering this profession as they feared insufficient income when they had yet to have the opportunity to create a sufficient client base (blayney, 2014). these dated compensation models discouraged women and made them feel unwelcome. however, the models might have been an antecedent to the gender pay gap problem and were barriers to advancement for women in the global financial services industry (benjamin, 2018). in australia, commission-based structures linked to competitive sales cultures discouraged some women faps from pursuing advancement (richards et al., 2020). women faps often felt inhibited by the sales-oriented pressure and the system that based its rewards on key performance indicators linked to sales, variable remuneration, and competition. these factors did not create an environment of cooperation and collaboration that would be preferable to women. men and women's horizontal and vertical segregation in the financial services sector impacted all workplace practices (richards et al., 2020). stereotypical gender biases the persistent unconscious bias relating to stereotypical assumptions about gender roles has compounded the problem. in australia, women in the industry often occupied roles as paraplanners. a paraplanner assists the financial planner in their administrative duties, giving the financial planner more time to prospect and advise clients. the paraplanner role might be a financial planning associate, licensed assistant, or junior investment advisor in the united states and canada. while these positions provided flexible work arrangements such as part-time and hybrid work, they have had limited conversion opportunities to move into the financial advising professional role (richards et al., 2020). a pathway for career progression was needed in the united states financial services industry. in places like the united kingdom, the united states, and canada, regulatory and policy changes have formalized economic equity support for women. however, these are relatively new, and in many cases, policy initiatives have yet to trickle down to the employee level (canadian immigrant, 2021). in the united kingdom, union-represented financial services staples et al. 179 professionals advocate for employment equity and transparency legislation for compensation. nevertheless, pay and bonus gaps still existed at banks in the united kingdom. the pay gap was 35%, while the bonus was 52% (healy & ahamed, 2019). in canada, the ministry for women and gender equity (wage) has created a development plan with targeted initiatives to address systemic barriers to women's economic advancement and progress (women and gender equity canada, 2023). however, the financial services industry seemed particularly resistant to voluntary change (healy & ahamed, 2019). diversity recognition at the federal level, the bank of canada is in an enviable position to lead and drive change in diversity and inclusion for banking throughout canada. in 2020, the bank of canada committed to a policy. it developed various strategies to achieve inclusivity and diversity and reduce racism within its employee complement and among its stakeholders in canada and abroad (bank of canada, 2021). the bank sought to achieve these goals through education, employee development, outreach, scholarship, and recruitment programs. as the bank of canada continued to build upon and expand its diversity, inclusion, and anti-racism efforts, it continued to foster grassroots employee resource groups. the goal was to foster dialogue, provide leaders with the necessary tools and training to effectively champion inclusion at all levels, and identify and address biases and barriers that impeded equal opportunities and outcomes at the bank (bank of canada, 2020) although the average canadian resident might bank at various financial institutions across the country, the bank of canada set the stage for policy and strategies for other banks and financial institutions to follow. although policies and programs designed to foster diversity and inclusion exist, programs offering one-time diversity and inclusion training need to be improved, as actual change requires time and the ability to practice (shin, 2021). removing barriers is a multipart process that includes understanding the current framework, heightening awareness of potential biases, and applying principles of fairness across all levels of organizations. hiring people who reflected the population's diversity was needed to achieve statistical equity (coulson-thomas, 2023; limón, 2020). instead, there needed to be a genuine intent to improve inclusivity and diversity through education, engagement, and professional development. bonus culture and presenteeism two additional related issues are the persistent bonus culture and presenteeism. in this context, presenteeism referred to simply being in the office, noticeable, and therefore more easily able to advocate for oneself and be seen by superiors and those who awarded promotions, bonuses, and allocated client cases. women avoided seniorlevel financial services positions because the culture required employees to be present to be rewarded or promoted (women in finance, 2017). similarly, there might have been an interplay between presenteeism and bonus culture, as women were in the office less than men due to family and caregiver commitments and their preference for remote work (richards et al., 2020). the bonus culture requiring advisors to selfpromote to justify their bonuses has exacerbated the issue, as men argued more forcefully and successfully with management for bonuses than women (house of commons treasury committee, 2018). when success relied on only sales performance, and there was no inclusive measure for compensation, the system rewarded men. performance matters, but it could be the only measure, nor could the system of bonuses be a negotiation (women in finance, 2017). these blatant, gendered structures and processes operate as barriers to entry, retention, and promotion should organizations wish to attract and retain women in the financial services industry. barrier five: lack of information, misunderstanding, and myths about the profession misinformation and misunderstanding the structured literature review revealed that misinformation and misunderstanding negatively impacted the entrance of women (particularly younger) into the financial planning profession financial services review, 33(2) 180 (bordalo et al., 2018; reiter & kiss, 2021; seeber, 2015). although more women were enrolling in financial planning programs at postsecondary institutions in the united states (35% of students complement), and awareness of the profession among women continued to grow, women were still enrolling in these programs at lower rates than their male counterparts (seeber, 2015). according to luke dean, program director at utah valley university, industry firms have been looking to hire the best and brightest women students. still, even if the demand was there, the women students were not (seeber, 2015). a survey of u.s. institutions offering financial planning degrees identified a general lack of understanding, misconceptions, inability to recruit, and overall lack of awareness relating to career opportunities within the financial planning profession that were not sales-related (reiter & kiss, 2021). one of the most significant challenges was the limited supply of candidates (women and ethnically diverse individuals) that employers could recruit into the financial advisor pipeline. there are many avenues that candidates can take to enter the profession, but it is becoming increasingly common to take financial planning as a major at the college level. while 53% of undergraduate financial planning programs have actively recruited women students, the results are mixed (reiter & kiss, 2021). a common theme emerged, indicating acceptable levels of diversity and inclusion, although women's representation in these programs is 28% of the total student population (reiter & kiss, 2021). respondents to the reiter and kiss (2021) survey indicated that financial planning has better representation of women than other finance programs, which was considered a win by the study's authors. formal recruiting existed, but most students learned about financial planning through academic advisors, personal finance, business 101 courses, or word of mouth (reiter & kiss, 2021). career awareness is lacking the cfp board of standards also identified these challenges in 2014. while 90% of women cfp® professionals in the united states understood that the profession required effective communication and a comprehensive approach to planning, that percentage dropped dramatically to 60% for women advisors without the designation. further, only 33% of women who believed the profession required strong sales skills wanted to become financial planners. in comparison, 45% of women surveyed in the 2013 win research project considered the profession one of relationship building (blayney, 2014). there was a lack of awareness of the financial planning profession and the cfp® certification process by women in the united states (blayney, 2014). this lack of awareness and understanding diminished the profession's value as a viable career opportunity (blayney, 2014). math myth and confidence a serious misgiving about financial planning as a career that might discourage women from the profession is that financial planning is all about math. financial planning requires various skills, and math proficiency is a single skill needed to succeed. regardless, several studies have indicated that women experienced higher math anxiety than men (bernstein & others, 1992; hart & ganley, 2019; van mier et al., 2019). men also tended to be more confident about their mathematical skills, even if that confidence is misplaced (bordalo et al., 2018). the sources of such overconfidence were only partially understood, but women needed to be more selfconfident in male-dominated professions like financial planning (bordalo et al., 2018). the combination of the math overconfidence of men and women's apprehension about math bolstered the idea that a math anxiety gap existed between men and women. this anxiety made women less likely to pursue the financial planning profession or required educational programs as a result. beliefs held by women about themselves influence important decisions, include college applications, career path selection, and willingness to contribute ideas in the workplace or compete for a promotion (bordalo et al., 2018). if women believed they would not be successful in a career like financial planning, it was unlikely that they would follow that career path. staples et al. 181 theoretical framework(s) uncovered during the literature review, several promising theories and frameworks were identified that attempt to explain or provide context for the predominance and persistence of men in the financial planning industry. these theories do not predominate in the financial planning discipline. however, exploring the theories below provides further understanding of the effects of role models, social learning theory, preference theory, and social role theory on gender and the financial planning profession. role model theory (gibson, 2004) suggests that "[envisioning] oneself in a higher position requires someone with whom you identify having attained that position" (women in advice, 2017, p. 10). role model theory has focused on how role models influence and persuade people to achieve their goals (morgenroth & ryan, 2015; rahman & day, 2012). according to morgenroth et al. (2015), role models serve three roles: 1) to influence and motivate, 2) to act as behavioral models that represent what is achievable, and 3) to inspire. the theoretical framework is concerned with when and how role models can influence and motivate the actions of others (morgenroth et al., 2015). as there are so few women in the financial planning profession, especially in managerial and executive roles, women need to see role models to which they can aspire (ogden et al., 2006). firms can partner junior women with senior female mentors as seeing women successfully lead and excel in the organization can boost self-efficacy and encourage imitation of productive, leadershiporiented behaviors. in addition, whenever possible, showcase the successes and expertise of female professionals—through newsletters, webinars, and conferences—so that positive examples are more visible. social learning theory (bandura, 1977) considers the environment within which a person exists and cognitive dynamics as influences on human behavior and learning. this learning happens through observation and imitation of the behaviors observed. the observed behaviors and actions can be reinforced internally and externally and can be positive and negative (bandura, 1977). as it relates to women in financial services, having their abilities doubted, coupled with a lack of representation of women at the management and c-suite levels, may cause women to develop low opinions of their abilities about job-related behaviors, causing them not to fulfill their potential (bandura, 1977, 1978, 1982; betz & hackett, 1986; manolova et al., 2007; pasztor et al., 2019). women who experience gender bias tend to internalize this bias, and the longer the experience, the more deeply embedded it becomes. based on social learning theory, it would benefit firms to offer targeted professional development and training by providing leadership, negotiation, and public-speaking workshops that directly address common barriers women face, such as being talked over in meetings or having their expertise questioned. incorporate confidence-building sessions that tackle imposter syndrome, counter bias, and reinforce personal strengths, ensuring women have the necessary tools to thrive. additionally, encourage a continuous learning culture by inviting women to share newly acquired skills and strategies in team forums or “lunch and learn” sessions, thereby reinforcing good practices across the organization. according to catherine hakim's (2006) preference theory, the socio-economic environment within which women exist, and conditions women's choices related to paid work and responsibilities to home life and family. preference theory posits that the five preconditions: access to contraceptives, full access to all occupations, the ability to access white-collar jobs, the ability to work remotely and in a part-time capacity, and a general attitude that encourages and supports a woman's right to choose their lifestyle must be present for women to entirely choose how they reconcile paid employment and family caregiving (hakim, 2006). generally, women fall into three categories: 1. those who choose a career (a workcentric life) 2. women who prioritize family (a homecentric life) 3. and women who choose both paid employment and family financial services review, 33(2) 182 most women fall into this last category (marshall, 2009). according to hakim, very few societies within europe, apart from the netherlands and the united kingdom, provide genuine choices for women, resulting in having achieved these five preconditions (hakim, 2006). if women do not exist within an environment that provides for these five criteria, they may not have genuine choices as it relates to careers and families. to help women exercise genuine choice in reconciling work and family, firms can adopt policies and practices that address these five preconditions. first, they can provide robust healthcare benefits that include comprehensive coverage for contraceptives, thereby supporting women’s autonomy over their family-planning decisions. second, organizations should create paths to all occupations—especially roles traditionally dominated by men—through transparent hiring, promotion processes, and targeted recruitment efforts. third, expanding access to white-collar and professional roles can be achieved by offering skill-building programs, mentorship initiatives, and clear career progression frameworks. fourth, firms can institute flexible work options—such as part-time positions, remote work, or job-sharing—that make it easier for women to balance family obligations while continuing to develop their careers. finally, cultivating an inclusive, supportive culture is essential: leaders should visibly endorse women’s rights to structure their own balance of work and home life, ensuring that career breaks, caregiving responsibilities, or alternative schedules do not hinder long-term advancement. social role theory (srt; eagly, 1984) concerns similarities and differences in social behavior based on gender. srt posits that the allocation of men and women into social roles within society results from their gender. gender-determined roles strengthen and reinforce the division of labor for men and women (eagly et al., 2000). socialization, a significant process identified by social role theory, points to the need or expectancies of men and women to adapt (archer, 1996). this adaptation resulted in societal characteristics and behavior endemic to the sex-defined roles. while srt may not help explain all societies worldwide, in industrialized countries, women are more likely to default to primary caregiver, a nurturing position identified as communal by archer (1996). men are the financial providers participating in the paid economy and are characteristic of agentic expectancies, such as assertive and instrumental (archer, 1996; eagly & wood, 2016). these socially specified roles define a person's responsibilities (bosak, 2018). financial planning firms can harness and highlight traditionally “feminine” or communal strengths—such as empathy, relationshipbuilding, and holistic care—to more effectively serve clients. by training all advisors, not just women, to listen actively and understand clients’ emotional as well as financial well-being, firms can elevate the overall client experience. through highlighting the importance of compassion, patience, and attentive communication in firmwide policies and marketing materials can showcase a firm’s commitment to caring for clients’ entire financial journey, ultimately differentiating the business in a competitive marketplace and making it a place at which, women are more likely to want to work. discussion this structured literature review described the five types of barriers to entry and advancement in the profession, including 1) dated gender roles, 2) antiquated compensation models, 3) gender bias and discrimination, 4) absence of targeted leadership, and 5) persistent myths and misunderstanding about the profession. the challenge extends beyond merely encouraging women to enter the financial planning sector; it involves devising strategies to reduce the ongoing underrepresentation of women in positions of power and influence. across all facets of the financial services sector—except at the entrylevel—women remain in the minority. alarmingly, women exit the profession at higher rates than men, exacerbating the scarcity of available women in the already narrow pipeline for promotion (carter et al., 2016; domski, 2018). the persistent scarcity of women in authority and influence to empower needed change in the direction of diversity, equity, and inclusion (dei) needs attention. this scarcity perpetuates an staples et al. 183 insufficient pipeline of women progressing beyond entry-level roles and into the financial planning profession. the predominance of older white men in leadership roles perpetuates a mirror image of gender inequity (domski et al., 2018; sheedy, 2021). women cannot envision their success if they cannot see it, and therefore, gender equity remains elusive (diehl et al., 2020; rogish et al., 2022). research in corporate finance and management literature consistently demonstrates that diversity drives competition, improves collaboration, and enhances productivity (kogel et al., 2023; woolley et al., 2010). elevating women’s participation in the financial planning and services profession is more than equity; it aligns with the goals of the entire finance industry. the financial services profession must recognize the immense opportunity to engage, promote, and serve the financial needs of women. as women gain more control over personal financial assets and increase their involvement in decisionmaking their demand for more holistic planning and their influence will only continue to grow. to attract more women to the profession, organizations can take strategic steps: 1. barrier one: conscious and unconscious gender bias and discrimination. social role theory posits that gender differences stem from societal expectations and gender role assignment. to ensure a more equitable workplace, financial institutions can challenge entrenched traditional gender roles by ensuring advancement opportunities are based on skills and talent. through open employee dialogue and employee engagement groups, perceptions, beliefs, and stereotypes can be reduced. providing a safe and secure environment for reporting harassment, abuse, or aggression without fear of reprisal can create a more equitable and inclusive workplace. 2. barrier two: absence of targeted leadership development programs for women. role model theory suggests that having women in corporate leadership positions, who can guide and encourage women in junior positions can motivate more women to pursue advancement within an organization. financial institutions can create targeted leadership development programs specifically tailored for women that highlight the achievement of women role models to help fill the pipeline through to second promotion to ensure women are capable leaders. women's leadership style differs significantly from men's, and having visible women role models and programs tailored to meet that demand can encourage more women to pursue advancement. 3. barrier three: organizational supports do not reflect women's needs and experiences. hakim’s preference theory suggests that organizations need to acknowledge and address the diverse preferences of women in the workplace. by hiring (or training) qualified professionals versed in gender sensitivity can implement formalized performance reviews based on meritocracy, while recognizing that women have differing work-life preferences, can ensure equitable advancement. a gender lens can help ensure that these diverse preferences are considered. organizations must provide open and transparent pathways for career progression and recognition to empower women to navigate the profession confidently and successfully. 4. barrier four: out-of-date compensation models and policies contribute to a gender pay gap. according to social role theory, compensation inequity results from societal expectations based on gender. financial institutions can challenge traditional thinking and structures around compensation to reduce this inequity and create or promote policies to provide more income stability. organizations may need to re-evaluated existing compensation models or offer a salarybased compensation structure or a hybrid model whereby the salary declines as the financial services review, 33(2) 184 commissions or aum compensations increase to mitigate gender disparities in earnings and help incentivize women to pursue financial planning careers. employers may also need to challenge traditional forms of compensation by examining the work traditionally performed by women and men to ensure all types of work is valued equitably. 5. barrier five: lack of information, misunderstanding, and myths about the profession. social learning theory stresses the importance of learning through observation and modeling. employers can use this theory to guide their programs and policies to provide young women and prospective students with opportunities to engage and interact with successful women in the profession. educating persons of influence, parents, educators, and mentors about the profession through workshops, cooperative opportunities, and engagement events whereby prospective students can ask questions to correct misunderstandings are valuable. showcasing the profession through various forms of media to dispel myths and misconceptions about the profession may provide a more accurate and appealing picture of the profession. by dismantling barriers and fostering an inclusive environment, the financial planning profession can genuinely reflect the needs and aspirations of current and future clients. limitations the initial goal of this paper was to examine the canadian landscape. however, the absence of canadian data expanded the focus to the united states, australia, and the united kingdom, where the financial services industry and financial planning profession are well developed. even with the expanded geographic region, research was so limited that it required the inclusion of the financial services industry. while all attempts to mitigate bias have been employed, articles were excluded that examined the experiences of women in other professions and industry to ensure relevance to the financial planning profession. additionally, while multiple databases were searched, some relevant research studies and articles may have been overlooked because of the search terms and the database coverage. finally, our search was limited to a fifteen-year period between 2008 – 2023. research conducted before that date, or very recently has not been included. conclusion this research sought to understand the state of women in the financial planning profession in canada over the last 15 years. the findings contribute to the existing research in four critical ways. first, it provides a compilation of existing literature on the nature of gender inequity in financial services, drawing on experiences in australia, the united kingdom, the united states, and, to a lesser extent, canada. second, it identifies common barriers across different regions that provide unique insights for consideration in the canadian context. next, it reveals possible theoretical frameworks such as role model theory (gibson, 2004), social learning theory (bandura & walters 1977), preference theory (hakim, 2006), and social role theory (eagly, 1984) that can help provide context to understand gender inequity and why it is so persistent in the financial planning profession. finally, it uncovers an essential future primary research agenda, namely gender inequity and canada's financial planning data gap, to make evidence-based decisions and policy recommendations. these insights can guide policy development designed to remove barriers for women and improve gender equity throughout the profession worldwide. references abraham, m. (2017). pay formalization revisited: considering the effects of manager gender and discretion on closing the gender wage gap. academy of management journal, 60(1), 29-54. https://doi-org.er.lib.kstate.edu/10.5465/amj.2013.1060 baghai, p., howard, o., prakash, l., & zucker, j. (2020). women as the next wave of growth in us wealth management. mckinsey & company https://doi-org.er.lib.k-state.edu/10.5465/amj.2013.1060 https://doi-org.er.lib.k-state.edu/10.5465/amj.2013.1060 staples et al. 185 bandura, a. (1978). reflections on selfefficacy. advances in behaviour research and therapy, 1(4), 237-269. bandura, a. (1982). self-efficacy mechanism in human agency. american psychologist, 37(2), 122. bandura, a., & walters, r. h. (1977). social learning theory. prentice hall. bergmann, n., scheele, a., & sorger, c. (2019). variations of the same? a sectoral analysis of the gender pay gap in germany and austria. gender, work & organization, 26(5), 668–687. https://doi.org/10.1111/gwao.12299 bernstein, j. d., & others, a. (1992). barriers to women entering the workforce: math anxiety. new jersey equity research bulletin, 3, 3-6. https://eric.ed.gov/?id=ed359381 betz, n. e., & hackett, g. (1986). applications of self-efficacy theory to understanding career choice behavior. journal of social and clinical psychology, 4(3), 279-289. bisco, j., gradisher, s., & mulholland, b. (2019). women and diversity--why the conversation must continue in financial services. journal of financial service professionals, 73(1), 72–84. blayney, e. (2014). making more room for women in the financial planning profession. cfp board. https://www.cfp.net/knowledge/reportsand-statistics/diversity-and-womensresearch/making-more-room-forwomen-in-the-financial-planningprofession blayney, e. (2016). what is the future of women in financial planning? journal of financial planning, 29(9), 32–33. balentine, c. (2024, december 9). women will control $30 trillion by 2030 in massive shift for wealth management. bloomberg. https://www.bloomberg.com/news/articl es/2024-12-09/women-will-control-30trillion-by-2030-in-massive-shift-forwealthmanagement?leadsource=uverify%20w all bloomfield, r. j., rennekamp, k., steenhoven, b., & stewart, s. (2021). penalties for unexpected behavior: double standards for women in finance. the accounting review, 96(2), 107–125. https://doi.org/10.2308/tar-2018-0715 bordalo, p., coffman, k., gennaioli, n., & shleifer, a. (2019). beliefs about gender. american economic review, 109(3), 739–773. https://doi.org/10.1257/aer.20170007 canadian bankers association. (2023) focus: representation of women at banks in canada. https://cba.ca/representation-ofwomen-at-banks-in-canada carter, n., elliott, v., harding, c., hetterich, a., howes, j., jackson, c., jeffords, p., & levine, b. (2016). when women thrive. mercer llc. https://www.mercer.ca/content/dam/mer cer/attachments/northamerica/canada/ca-2016-genderdiversity-wwt-financial-servicesperspective-en.pdf clempner, j., daisley, m., & jaekel, a. (2020). women in financial services 2020. oliver wyman. https://www.oliverwyman.com/content/ dam/oliverwyman/v2/publications/2019/november /women-in-financial-services-2020.pdf coulson-thomas, c. (2023). building future boards: women and diversity in leadership roles. effective executive, 26(4), 25–45. domski, s. (2018). pipeline or waterslide: advancing women into executive positions in business and financial planning. celebration of learning. https://digitalcommons.augustana.edu/ce lebrationoflearning/2018/presentations/2 4 eagly, a. h., wood, w., & diekman, a. b. (2000). social role theory of sex https://doi.org/10.1111/gwao.12299 https://eric.ed.gov/?id=ed359381 https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.bloomberg.com/news/articles/2024-12-09/women-will-control-30-trillion-by-2030-in-massive-shift-for-wealth-management?leadsource=uverify%20wall https://www.bloomberg.com/news/articles/2024-12-09/women-will-control-30-trillion-by-2030-in-massive-shift-for-wealth-management?leadsource=uverify%20wall https://www.bloomberg.com/news/articles/2024-12-09/women-will-control-30-trillion-by-2030-in-massive-shift-for-wealth-management?leadsource=uverify%20wall https://www.bloomberg.com/news/articles/2024-12-09/women-will-control-30-trillion-by-2030-in-massive-shift-for-wealth-management?leadsource=uverify%20wall https://www.bloomberg.com/news/articles/2024-12-09/women-will-control-30-trillion-by-2030-in-massive-shift-for-wealth-management?leadsource=uverify%20wall https://www.bloomberg.com/news/articles/2024-12-09/women-will-control-30-trillion-by-2030-in-massive-shift-for-wealth-management?leadsource=uverify%20wall https://doi.org/10.2308/tar-2018-0715 https://doi.org/10.1257/aer.20170007 https://cba.ca/representation-of-women-at-banks-in-canada https://cba.ca/representation-of-women-at-banks-in-canada https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.mercer.ca/content/dam/mercer/attachments/north-america/canada/ca-2016-gender-diversity-wwt-financial-services-perspective-en.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://www.oliverwyman.com/content/dam/oliver-wyman/v2/publications/2019/november/women-in-financial-services-2020.pdf https://digitalcommons.augustana.edu/celebrationoflearning/2018/presentations/24 https://digitalcommons.augustana.edu/celebrationoflearning/2018/presentations/24 https://digitalcommons.augustana.edu/celebrationoflearning/2018/presentations/24 financial services review, 33(2) 186 differences and similarities: a current appraisal. the developmental social psychology of gender, 12(174), 5-12. eastman, m. t., rallis, d., mazzucchelli, g. (2016). the tipping point: women on boards and financial performance. msci. https://www.msci.com/documents/1019 9/fd1f8228-cc07-4789-acee3f9ed97ee8bb egan, m., matvos, g., & seru, a. (2022). when harry fired sally: the double standard in punishing misconduct. journal of political economy, 130(5), 1184-1248. financial planning association. (2023). peggy ruhlin discusses the state of women financial planners. journal of financial planning, 36(12), 20–22. gandhi, j. (2020). how much wealth do women control? forbes middle east. retrieved january 29, 2025, from https://www.forbesmiddleeast.com/lead ership/women-leaders/how-muchwealth-do-women-control gibson, d. e. (2004). role models in career development: new directions for theory and research. journal of vocational behavior, 65(1), 134-156. https://doi.org/https://doi.org/10.1016/s 0001-8791(03)00051-4 gibson, s. k. (2004). social learning (cognitive) theory and implications for human resource development. advances in developing human resources, 6(2), 193–210. https://doi.org/10.1177/1523422 304263429 girardone, c., kokas, s., & wood, g. (2021). diversity and women in finance: challenges and future perspectives. journal of corporate finance, 71, 101906. https://doi.org/10.1016/j.jcorpfin.2021.1 01906 hakim, c. (2006). women, careers, and work-life preferences. british journal of guidance & counselling, 34(3), 279–294. hannon, j. m., & milkovich, g. t. (1996). the effect of human resource reputation signals on share prices: an event study. human resource management, 35(3), 405–424. https://doi.org/10.1002/(sici)1099050x(199623)35:3<405::aidhrm7>3.0.co;2-0 hart, s. a., & ganley, c. m. (2019). the nature of math anxiety in adults: prevalence and correlates. journal of numerical cognition, 5(2), 122–139. https://doi.org/10.5964/jnc.v5i2.195 healy, g., & ahamed, m. m. (2019). gender pay gap, voluntary interventions, and recession: the case of the british financial services sector. british journal of industrial relations, 57(2), 302–327. https://doi.org/10.1111/bjir.12448 house of commons treasury committee (2018). women in finance. 1-51. ibarra, h., carter, n. m., & silva, c. (2010). why men still get more promotions than women? harvard business review, 88(9), 80-85. ibarra, h., ely, r., & kolb, d. (2013). women rising: the unseen barriers. harvard business review, 91(9), 60–67. kögel, c. criscuolo, c. gal, p. & schwellnus, c. (2023), “gender diversity in senior management and firm productivity: evidence from nine oecd countries”, oecd productivity working papers, no. 34, https://doi.org/10.1787/58ad664a-en. khoury, m.-g., al-shaibani, a., & driscoll, m. (2023). women at canadian banks: thick glass ceiling. dbrs morningstar. https://dbrs.morningstar.com/research/4 10635/women-at-canadian-banks-thickglass-ceiling klein, g., shtudiner, z., & zwilling, m. (2021). uncovering gender bias in attitudes towards financial advisors. journal of economic behavior & organization, https://www.msci.com/documents/10199/fd1f8228-cc07-4789-acee-3f9ed97ee8bb https://www.msci.com/documents/10199/fd1f8228-cc07-4789-acee-3f9ed97ee8bb https://www.msci.com/documents/10199/fd1f8228-cc07-4789-acee-3f9ed97ee8bb https://www.forbesmiddleeast.com/leadership/women-leaders/how-much-wealth-do-women-control https://www.forbesmiddleeast.com/leadership/women-leaders/how-much-wealth-do-women-control https://www.forbesmiddleeast.com/leadership/women-leaders/how-much-wealth-do-women-control https://doi.org/https:/doi.org/10.1016/s0001-8791(03)00051-4 https://doi.org/https:/doi.org/10.1016/s0001-8791(03)00051-4 https://doi.org/10.1177/1523422304263429 https://doi.org/10.1177/1523422304263429 https://doi.org/10.1016/j.jcorpfin.2021.101906 https://doi.org/10.1016/j.jcorpfin.2021.101906 https://doi.org/10.1002/(sici)1099-050x(199623)35:3%3c405::aid-hrm7%3e3.0.co;2-0 https://doi.org/10.1002/(sici)1099-050x(199623)35:3%3c405::aid-hrm7%3e3.0.co;2-0 https://doi.org/10.1002/(sici)1099-050x(199623)35:3%3c405::aid-hrm7%3e3.0.co;2-0 https://doi.org/10.5964/jnc.v5i2.195 https://doi.org/10.1111/bjir.12448 https://doi.org/10.1787/58ad664a-en https://dbrs.morningstar.com/research/410635/women-at-canadian-banks-thick-glass-ceiling https://dbrs.morningstar.com/research/410635/women-at-canadian-banks-thick-glass-ceiling https://dbrs.morningstar.com/research/410635/women-at-canadian-banks-thick-glass-ceiling staples et al. 187 189, 257–273. https://doi.org/10.1016/j.jebo.2021.06.0 40 kumar, a. (2010). self-selection and the forecasting abilities of female equity analysts. journal of accounting research, 48(2), 393–435. http://www.jstor.org/stable/40784954 kurlowicz, a. (2014). women in financial planning. journal of financial service professionals, 68(3), 56. https://er.lib.kstate.edu/login?url=https://wwwproquest-com.er.lib.kstate.edu/scholarly-journals/womenfinancialplanning/docview/1527455925/se-2 kurtz, a. (2018). why is the pay gap for women financial advisors so wide? financial planning (online), https://er.lib.kstate.edu/login?url=https://wwwproquest-com.er.lib.k-state.edu/tradejournals/why-is-pay-gap-womenfinancial-advisors-sowide/docview/2022918109/se-2 limón, a. t. (2020). addressing the realities facing diverse employees in the financial planning profession. journal of financial planning, 36(11), 23–26. manolova, t. s., brush, c., & edelman, l. f. (2007). what do women (and men) want? entrepreneurial expectancies of women and men nascent entrepreneurs. frontiers of entrepreneurship research, 2008-01. http://dx.doi.org/10.2139/ssrn.1060241 manolova, t., brush, c., edelman, l., & shaver, k. (2012). one size does not fit all: entrepreneurial expectancies and growth intentions of us women and men nascent entrepreneurs. entrepreneurship & regional development, 24(1/2), 7–27. https://doi-org.er.lib.kstate.edu/10.1080/08985626.2012.63734 4 marshall, s. m. (2009). women higher education administrators with children: negotiating personal and professional lives. naspa journal about women in higher education, 2(1), 190-223. mccarthy, e. (2016). where are the women? can we close the advisor gender gap? retirement advisor, 17(7), 36-39. moreau, kait. (2022). the under-representation of women as independent financial advisors – a 20-year trend. environics research. https://environics.ca/article/the-underrepresentation-of-women-asindependent-financial-advisors/ morgenroth, t., ryan, m. k., & peters, k. (2015). the motivational theory of role modeling: how role models influence role aspirants’ goals. review of general psychology, 19(4), 465–483. https://doi.org/10.1037/gpr0000059 neck, c. (2015). disappearing women: why do women leave senior roles in finance? australian journal of management, 40(3), 488-510. noback, i., broersma, l., & van dijk, j. (2016). climbing the ladder: gender-specific career advancement in financial services and the influence of flexible work-time arrangements. british journal of industrial relations, 54(1), 114–135. https://doi.org/10.1111/bjir.12048 o’dwyer, m., & richards, d. w. (2021). occupational boundaries: gender capital and career progression in the financial planning industry. financial planning review, 4(2). https://doi.org/10.1002/cfp2.1123 ogden, s. m., mctavish, d., & mckean, l. (2006). clearing the way for gender balance in the management of the uk financial services industry: enablers and barriers. women in management review, 21(1), 40–53. https://doi.org/10.1108/0964942061064 3402 pasztor, j., sunder, a., & henderson, r. (2019). success and satisfaction of women in financial planning. financial planning review, 2(2), 1-22. https://doi.org/10.1016/j.jebo.2021.06.040 https://doi.org/10.1016/j.jebo.2021.06.040 http://www.jstor.org/stable/40784954 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/scholarly-journals/women-financial-planning/docview/1527455925/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/scholarly-journals/women-financial-planning/docview/1527455925/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/scholarly-journals/women-financial-planning/docview/1527455925/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/scholarly-journals/women-financial-planning/docview/1527455925/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/scholarly-journals/women-financial-planning/docview/1527455925/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/scholarly-journals/women-financial-planning/docview/1527455925/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/why-is-pay-gap-women-financial-advisors-so-wide/docview/2022918109/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/why-is-pay-gap-women-financial-advisors-so-wide/docview/2022918109/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/why-is-pay-gap-women-financial-advisors-so-wide/docview/2022918109/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/why-is-pay-gap-women-financial-advisors-so-wide/docview/2022918109/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/why-is-pay-gap-women-financial-advisors-so-wide/docview/2022918109/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/why-is-pay-gap-women-financial-advisors-so-wide/docview/2022918109/se-2 https://doi-org.er.lib.k-state.edu/10.1080/08985626.2012.637344 https://doi-org.er.lib.k-state.edu/10.1080/08985626.2012.637344 https://doi-org.er.lib.k-state.edu/10.1080/08985626.2012.637344 https://environics.ca/article/the-under-representation-of-women-as-independent-financial-advisors/ https://environics.ca/article/the-under-representation-of-women-as-independent-financial-advisors/ https://environics.ca/article/the-under-representation-of-women-as-independent-financial-advisors/ https://doi.org/10.1037/gpr0000059 https://doi.org/10.1002/cfp2.1123 https://doi.org/10.1108/09649420610643402 https://doi.org/10.1108/09649420610643402 financial services review, 33(2) 188 rahman, h., & day, j. (2012). who is your role model? the relationship between role models and student entrepreneurial motivation. international council for small business (icsb). 2(1), 0-35. https://er.lib.kstate.edu/login?url=https://www.proques t.com/conference-papersproceedings/who-is-your-role-modelrelationshipbetween/docview/1848669213/se-2 reiter, m., & kiss, d. e. (2021). efforts in diversity and recruiting in financial planning undergraduate programs. family and consumer sciences research journal, 49(3), 238–253. https://doi.org/10.1111/fcsr.12389 reiter, m., seay, m., macdonald, m., lutter, s., & loving, a. (2022). are there racial and gender preferences when hiring a financial planner? an experimental design on diversity in financial planning. journal of financial counseling and planning, 33(3), 344-357. reiter, m., seay, m., & loving, a. (2022). diversity in financial planning: race, gender, and the likelihood to trust a financial planner. financial planning review, 5(1), e1134. https://doi.org/10.1002/cfp2.1134 richards, d. w., roberts, h., & whiting, r. h. (2020). female financial advisers: where art thou? australian journal of management, 45(4), 624–644. https://doi.org/10.1177/0312896219896 389 rogish, a., shemluck, n., sandler, s., danielecki, p., & ramsay, t. (2019). within reach? deloitte insights. https://www2.deloitte.com/us/en/insight s/industry/financial-services/women-infinancial-services-leadership-roles.html seeber, c. m. (2015). for female financial planners, the glass is half full. journal of financial planning, 28(9), 20–22. sheedy, r. l. (2021). the importance of building a larger cadre of female cfp® professionals. journal of financial planning, 34(8), 62–65. smith, j. b. (1998). buyer–seller relationships: similarity, relationship management, and quality. psychology & marketing, 15(1), 3–21 sommer, matthew, lim, h. n., & macdonald, m. (2018). gender bias and practice profiles in the selection of a financial adviser. journal of financial planning, 31(10), 38-47. https://er.lib.kstate.edu/login?url=https://wwwproquest-com.er.lib.k-state.edu/tradejournals/gender-bias-practice-profilesselectionfinancial/docview/2131783502/se-2 tharp, d., lurtz, m., mielitz, k., & kitces, m. (2019). examining the gender pay gap among financial planning professionals: a blinder-oaxaca decomposition. financial planning review, 2(3-4), e1061. tharp, d. t., parks-stamm, e. j., lurtz, m., & kitces, m. (2022). exploring gender differences in marital and parental income premiums among financial advisors. journal of family and economic issues, 43(1), 15–35. https://doi.org/10.1007/s10834-02109766-4 women in financial services. (2023). catalyst. https://www.catalyst.org/research/wome n-in-financial-services/ woolley, a. m. chabris, c. pentland, a., hashmi, n. & malone, t. w. (2010). evidence for a collective intelligence factor in the performance of human groups. science 330, 686-688. doi:10.1126/science.1193147 world economic forum. (2022). global gender gap report. https://www.weforum.org/reports/globalgender-gap-report-2022/in-full/1benchmarking-gender-gaps-2022/ https://er.lib.k-state.edu/login?url=https://www.proquest.com/conference-papers-proceedings/who-is-your-role-model-relationship-between/docview/1848669213/se-2 https://er.lib.k-state.edu/login?url=https://www.proquest.com/conference-papers-proceedings/who-is-your-role-model-relationship-between/docview/1848669213/se-2 https://er.lib.k-state.edu/login?url=https://www.proquest.com/conference-papers-proceedings/who-is-your-role-model-relationship-between/docview/1848669213/se-2 https://er.lib.k-state.edu/login?url=https://www.proquest.com/conference-papers-proceedings/who-is-your-role-model-relationship-between/docview/1848669213/se-2 https://er.lib.k-state.edu/login?url=https://www.proquest.com/conference-papers-proceedings/who-is-your-role-model-relationship-between/docview/1848669213/se-2 https://er.lib.k-state.edu/login?url=https://www.proquest.com/conference-papers-proceedings/who-is-your-role-model-relationship-between/docview/1848669213/se-2 https://doi.org/10.1111/fcsr.12389 https://doi.org/10.1002/cfp2.1134 https://doi.org/10.1177/0312896219896389 https://doi.org/10.1177/0312896219896389 https://www2.deloitte.com/us/en/insights/industry/financial-services/women-in-financial-services-leadership-roles.html https://www2.deloitte.com/us/en/insights/industry/financial-services/women-in-financial-services-leadership-roles.html https://www2.deloitte.com/us/en/insights/industry/financial-services/women-in-financial-services-leadership-roles.html https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/gender-bias-practice-profiles-selection-financial/docview/2131783502/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/gender-bias-practice-profiles-selection-financial/docview/2131783502/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/gender-bias-practice-profiles-selection-financial/docview/2131783502/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/gender-bias-practice-profiles-selection-financial/docview/2131783502/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/gender-bias-practice-profiles-selection-financial/docview/2131783502/se-2 https://er.lib.k-state.edu/login?url=https://www-proquest-com.er.lib.k-state.edu/trade-journals/gender-bias-practice-profiles-selection-financial/docview/2131783502/se-2 https://doi.org/10.1007/s10834-021-09766-4 https://doi.org/10.1007/s10834-021-09766-4 https://www.catalyst.org/research/women-in-financial-services/ https://www.catalyst.org/research/women-in-financial-services/ https://www.weforum.org/reports/global-gender-gap-report-2022/in-full/1-benchmarking-gender-gaps-2022/ https://www.weforum.org/reports/global-gender-gap-report-2022/in-full/1-benchmarking-gender-gaps-2022/ https://www.weforum.org/reports/global-gender-gap-report-2022/in-full/1-benchmarking-gender-gaps-2022/ financial services review, 33(2) 15 consumers’ basic bank account complaints and their financial hardships: a content analysis of complaints filed with the consumer financial protection bureau (cfpb) julie birkenmaier1 and hope stratman2 abstract although savings and/or checking account ownership is widespread, significant account problems occur that carry negative implications for consumer finances. this study aims to profile american consumers’ bank account experiences when they encounter challenges with the use of their basic bank account that are not resolved through initial contact with their financial institution. a systematic sample of consumer saving and checking account complaints submitted to the consumer financial protection bureau in 2022-2023 is used to conduct content analysis to identify prevalent themes. the resulting content analysis categories are used in a predictive model to determine the drivers of financial hardship. results suggest that experiencing fraud issues and automatic teller machine (atm) malfunctions led to increased odds of experiencing financial issues. also problematic were challenges relating to funds withheld by the financial institution, account transaction issues, and problems with account features. customer service issues that led to increased odds of financial hardship were staff’s inability to solve their customer issues, weak engagement with their customers, and lack of or wrong information provided to their customers. financial institutions can use these results to focus on the most critical issues that negatively impact customer finances. policy changes to financial institutions, both internal and external, can focus on decreasing the rate and implications of fraud and atm challenges on consumer finances. internally, improving customer service in several key areas through rigorous training to standards and monitoring as well as enhanced grievance procedures, may also be impactful. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation birkenmaier, j. & stratman, h. (2025). consumers’ basic bank account complaints and their financial hardships: a content analysis of complaints filed with the consumer financial protection bureau (cfpb). financial services review, 33(2), 15-35. 1 corresponding author (julie.birkenmaier@slu.edu). saint louis university school of social work, st. louis. missouri, usa 2 saint louis university school of social work, st. louis. missouri, usa. https://creativecommons.org/licenses/by-nc/4.0/ mailto:julie.birkenmaier@slu.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(2) 16 introduction financial institutions play a large and powerful role in the economy and in the lives of everyday american consumers. however, declining u.s. consumer trust in the financial services industry in recent years has co-occurred with increasing consumer vulnerability (i.e., consumers' ability to effectively engage in the market due to the risk of financial hardship) with mainstream financial institutions for a number of reasons (chawla et al., 2023; edelman, 2022; o’connor et al., 2019). first, consumer vulnerability is influenced by financial literacy, which is generally low (klapper & lusardi, 2019). lower financial literacy leaves consumers in a less-advantaged position to determine the best product fit for them in their circumstances and to understand the various processes in place regarding their transactions and investment opportunities. in addition, the number of financial institutions and geographic distribution of bank branches has decreased dramatically in recent years (national bureau of economic research, 2023), which means that many americans preferring to have the option of in-person transactions and communications (including when needed to rectify issues) have fewer options. consumers who live in rural areas and without reliable, strong internet connectivity have fewer bank options (lee et al., 2022). given the increased difficulty of reaching them, financial institutions may have less incentive to offer products that meet their needs. customer service in retail financial institutions has become increasingly automated, which can make it difficult to gain needed information and participate in dispute resolution (cfpb, 2023b). consumers with lower income may struggle with product cost, such as in maintaining minimum balances or meeting transaction requirements (fdic, 2022). these consumers may also perceive unfairness and discrimination in their treatment when working with financial institution staff (kamran & uusitalo, 2019; lim & letkiewicz, 2023). personal attributes, such as disability, physical and mental health challenges, and experiences of intimate partner violence, can also contribute to consumer vulnerability (mogaji, 2020; scott, 2023). in sum, consumer vulnerability and diminished trust in financial institutions can lead to suboptimal product ownership experience with negative ramifications for personal finances. within this context, millions of american consumers experience challenges with their basic bank account every year. some of these consumers will try to resolve these issues by contacting customer service and using the appeal mechanisms of their financial institution. if unsuccessful at a satisfactory resolution, consumers may contact federal agencies that can assist in mediation. challenges with many financial products, such as credit cards, mortgages, and others have been examined, with results suggesting potential negative ramifications for consumers (dou et al., 2024; estelami & liu, 2023; halvorsen & møkkelgård, 2018; polat et al., 2023). however, the ramifications of bank account challenges is an under-studied, yet important, aspect of the consumer financial experience. background theoretical framework several theoretical frameworks inform this research. first, the integrative consumer vulnerability framework is used to understand consumer vulnerability in the marketplace. this framework recognizes multiple dimensions of vulnerability, such as emotional, practical, and relational factors, as well as their interrelationships. additionally, the framework highlights vulnerability as a dynamic that consumers can experience across various domains and contexts. the framework includes two components: a transient and a systemic, class-based component. the transient component is specific only to the current episode, and the systemic, class-based component is based on demographic or sociocultural factors (commuri & ekici, 2008). in this study, the transient component is used to examine the ways in which current bank account issues (e.g., disputed transactions) relate to customer service and financial hardship. the framework is also used to examine how systemic and class-based components (i.e., consumer vulnerabilities) are associated with account challenges, customer service, and financial hardships. recent research findings support the notion that any consumer can experience vulnerability due to inequities in the birkenmaier & stratman 17 financial marketplace, and that vulnerable consumers are more likely to experience mistreatment (e.g., lack of disability accommodation) (garrett & toumanoff, 2010; lim & letkiewicz, 2023; salisbury et al, 2023). second, a range of theories and factors can explain why consumers might complain about a retail financial institution by appealing to a federal agency in hopes of a satisfactory resolution, rather than just switching to a different financial institution. practical matters may take precedence, such as the inability to access their funds to move them, the inability to open an account at another financial institution, the hope of financial compensation or reimbursement as a result of filing a complaint, or other related switching costs that present a financial or logistical disadvantage to a change. beyond these, affect control theory (act) suggests that emotions, particularly anger at or loyalty to a particular financial institution, play a role in consumer decision-making. while anger might propel consumers to seek a change, consumers may feel a relational bond to a specific financial institution such that their view of themselves would be negatively affected by the perceived disloyalty involved in a change (chebat & benamor, 2010; durkin, 2003; kabadayi, 2016). in sum, a combination of emotional, practical, and relational factors can influence a consumer’s decision to complain and advocate for a positive resolution to a complaint to a federal agency rather than switch financial institutions. bank account challenges despite the fact that 95.5% of americans own a savings or checking account at a bank or credit union, problems persist in access to and longterm use of safe, affordable, and beneficial bank accounts. bank accounts are considered a “gateway” product that offers a pathway to the use of other financial products and services from mainstream financial institutions. however, challenges to ownership of an account remain for the 5.9 million u.s. households without accounts. the challenge to ownership is especially present for those households that are over-represented among those with no account (i.e., the ‘unbanked”), including lower-income, less educated, disabled, and black and hispanic households. these households state that they do not have enough money, do not trust financial institutions, and fear unpredictable fees, among other reasons for avoiding accounts (fdic, 2022). for another 25.7% of the population, account ownership has not led to safe, lower cost products; instead, they persist in using highercost products and services from alternative financial product and service (afps) providers (e.g., pawnshops, rent-to-own, payday loans) (fdic, 2019). consumer complaints about checking or savings accounts have steadily risen in recent years (cfpb, 2020, 2021, 2022a). the most common complaints in 2023 related to accounts that were closed without an explanation, managing an account, closing an account, opening an account; or to problems with a lender or other company charging an account (cfpb, 2023e). consumer protection in the financial marketplace both the federal and state governments provide regulatory structures and efforts to protect consumers in the financial marketplace. consumers who experience issues with their bank accounts can complain and seek assistance from the better business bureau (bbb), the federal trade commission (ftc), the federal reserve bank, the consumer financial protection bureau (cfpb), and their local police. these agencies can seek compensatory actions from the financial institution for specific consumers. more broadly, they can educate the public about their rights and responsibilities, collect information, conduct investigations, and sue companies. the regulators can also develop and enforce rules to maintain a fair marketplace (ftc, n.d.). the role of the consumer financial protection bureau in consumer protection the consumer financial protection bureau (cfpb) is a prominent consumer protection agency. it was established in 2010 as a division of the federal reserve system to advocate for the rights and wellbeing of consumers in the u.s. financial sphere (cfpb, n.d.-a). as one of its many functions, the cfpb monitors consumer protections by maintaining a publicly available database of consumer complaints, which it uses financial services review, 33(2) 18 to follow up with large financial services institutions and ensure resolution of such complaints (cfpb, n.d.-b). using the complaint process, the cfpb has facilitated timely responses to consumer complaints from more than 6,100 financial companies (cfpb, n.d.-d). in 2022-2023, the cfpb facilitated an explanatory response to over half of all complaints filed (61% 2022; 53% in 2023), and approximately one-third of the complaints were closed with non-monetary relief (31% in 2022; 40% in 2023) (cfpb, 2022a, 2023a). even without a specific response, complaints to government entities positively affect bankcustomer relations. hayes et al. (2021) found that the threat of consumer complaints affects how banks treat their customers in communities with low trust of financial institutions. research on bank account challenges previous studies about bank account challenges have reported a wide variety of issues with consumer accounts and with the mechanisms and processes designed to assist consumers when they have issues. complaints about unauthorized transactions from accounts, including fraud, are prevalent (morgan, 2021). customers also report concerns and issues related to their use of online and mobile banking (park, 2016) and of automatic teller machines (atms) (gyamfi et al., 2016; nndwamato, 2018). regarding both, customers complain about lack of security, lack of legal support, technical illiteracy, among others (nndwamato, 2018; park, 2016). consumers also report issues related to financial institutions unilaterally closing consumer bank accounts without revealing a reason and without notice (bank policy institute, 2020; jelisejevs, 2021), then reopening them without a consumer’s consent or knowledge (cfpb, 2023c). through the bank secrecy act and anti-money laundering act of 2020, financial institutions can close accounts to combat money laundering and terrorism. in the process, financial institutions determine “suspicious” activity through a rulesbased algorithm's examination of transactions and, without any specific finding that a crime has been committed, close the account without explanation to the consumer. because financial institutions are at legal risk by mistakenly keeping accounts with suspicious activity open, there is a strong incentive to close accounts. the vast majority of the individuals whose accounts have been closed in this manner are likely innocent of any wrongdoing (bank policy institute, 2018, 2024). financial hardship and bank accounts financial hardship is defined as “a state of distress in which an individual is unable to maintain a standard of living” (o’conner et al., 2019, p. 422). consumers are vulnerable to financial hardships when unable to access their funds in accounts due to consumer protection gaps, fraud procedures, and system failures. this inability can lead to material hardship (i.e., hardship related to food, housing, medical services, and other basic needs), difficulty in making ends meet (i.e., paying bills), having debt in collections (i.e., being contacted by a debt collector), or being unable to absorb a financial shock by accessing cash in a short period of time (warmath et al., 2022). little literature has examined the direct relationship of bank account challenges and financial hardship, but related research informs this study. brenner et al. (2020) found that consumer fraud victimization adversely affects an individual’s financial well-being. they found that fraud (i.e., misrepresentation of information and misuse of money by third parties) is associated with a loss of confidence in financial matters, which negatively affects future financial decision-making. lim and letkiewicz (2023) found several adverse financial results associated with compromised bank accounts due to fraud. the events associated with a compromised account due to fraud included income shocks, a health emergency, having work hours or pay reduced, using payday loans, the use of reloadable cards, and initiating non-bank international transfers. being more likely to stop taking medication due to cost was also associated. several studies examined financial hardship’s effect on account ownership. for instance, goodstein and kutzbach (2024) found that job loss and corresponding income reduction leads to a large decrease in the likelihood of having an account for lower-income, renter households. the fdic (2022) found that financial hardship (e.g., losing or quitting a job, being furloughed, birkenmaier & stratman 19 having reduced work hours, or having a significant loss of income) contributed to closing an account. study justification and research questions little research has been published on the characteristics of complaints from consumers related to their basic bank accounts, or the relationship of bank account challenges to customer service or financial hardship. in addition, little is known about the relationship of bank account challenges to demographic or sociocultural factors that make consumers vulnerable. this study fills a gap in the literature to provide a detailed discussion of the challenges consumers face with their bank accounts from large u.s. financial institutions. the research questions are: 1. what are the characteristics of consumer complaints about their basic bank accounts from large u.s. financial institutions, as related to bank problems, customer service, financial hardship, and emotional hardship; and 2. do consumer problems with basic bank accounts relate to consumer financial hardship? methods consumer research has made extensive use of content analysis (e.g., bartikowski & laroche, 2019; estelami & liu, 2023; kim et al., 2013; lecoeuvre et al., 2021). following these examples, this study used a content analysis method to examine the research questions. content analysis is a research methodology that systematically analyzes and interprets qualitative data to identify patterns, themes, and meanings within the data. the process involves breaking down the content (e.g., narrative consumer complaints), into manageable units (e.g., topics), then categorizing and coding these units based on predetermined criteria. a key element is the development of a coding scheme, or set of categories based on the research objectives, and applying the categories to the data in a systematic way based on their content or characteristics (krippendorff, 2019). in this study, researchers used content analysis to code qualitative consumer complaints into quantitative data, using the following multi-step process to design and implement the coding procedure. two coders first agreed on four metacategories (i.e., financial institution problem(s), unsatisfactory customer service, financial hardship resulting from the financial institution problem(s), and emotional hardship resulting from the financial institution problem(s)). these meta-categories were developed based on previous scholarship using the cfpb complaint database, including research on credit card fraud that found financial and emotional hardship (estelami & liu, 2023), account fraud literature about consumer vulnerability and compromised accounts (lim & letkiewicz, 2023), as well as a review of the first 25 complaints in the dataset. next, the two coders independently read a small number of the complaints and identified relevant “thought categories” expressed by the complaints (e.g., “cannot view account online" and “unauthorized account opening”). these thought categories were compared between the two coders, who used discussion to resolve differences. after several rounds of creating new thought categories based on approximately 10% of the complaints, the thought categories were grouped under thought domains (e.g., “account incentive problem,” “account transaction problem,” and “fee problem”) and metacategories (krippendorff, 2019). after the creation of the initial thought domains, each complaint record was coded at the thought domain level with as many thought domains as needed. the coders occasionally added new thought categories and thought domains, and/or agreed to move a thought category from one thought domain to another (e.g., “lack of access to provisional credit” from the thought domain of “policy or procedure problem” to “lost funds or unable to access funds”). a fifth meta-category, customer vulnerability, was added during the coding process, and thought domains and categories were created for it. previously coded complaints were retroactively re-coded each time a code was changed or added to reflect the current coding scheme, and coders continued to compare coding and resolve discrepancies through discussion throughout the coding. this process allows the transformation of qualitative data (narrative complaint data) into quantitative data (numbered codes) regarding categories and domains that the complainants expressed. using the integrative consumer vulnerability framework, emotional (i.e., emotions arising financial services review, 33(2) 20 from the complaints), practical (i.e., facts of the complaint), and relational factors (i.e., interactions with staff) were coded. discrepancies in the coding of thought domains were tracked. the reliability of the coding was found to be acceptable for each of the metacategories, with weighted cohen’s kappa of 0.60 for financial institution problem, 0.56 for unsatisfactory customer service, 0.41 for financial hardship, 0.39 for emotional hardship, and 0.49 for consumer vulnerability. in this study, weighted cohen’s kappa, which accounts for varying degrees of disagreement, was calculated using a percentage of agreement method. zero represented complete agreement, and partial agreement was assigned a percentage value. the observed agreement (po) was calculated considering these weights, and the expected agreement (pe) was derived from the marginal totals. the final kappa values were then computed using the standard formula (cohen, 1968). data since 2011, the cfpb consumer complaint database has received over four million complaints online or over the telephone (cfpb, n.d.-b). the cfpb first routes consumer complaints to the relevant financial services company for review. the company then has a period of time to respond to the issue, communicating with the complainant as needed, and the cfpb keeps the complainant updated on the company's response. the complainant also has a chance to give feedback on the company's response (cfpb, n.d.-c). with the consumer's permission, the cfpb publishes complaints against large financial services providers (i.e., over $10 billion in assets) in the consumer complaint database, which is publicly available on their website, while complaints about smaller financial institutions are sent directly to their regulators. in addition, only complaints for which the financial institution acknowledges that the complainant is or was a customer are included. complaints do not appear in the dataset until they are resolved in some manner (e.g., communicated with the customer, or provided non-monetary or monetary relief). consumer complaints address a variety of products in the financial services industry, including credit reporting; debt collection; checking or savings accounts (cfpb, n.d.-b). the cfpb categorizes each complaint according to a product, sub-product, issue, and sub-issue, along with the date the complaint was received, the state, and the zip code. in addition, the original brief narrative summary of the complaint (in the complainant’s own words), and whether the company provided a response, is included. the cfpb does not provide any information about customer service, financial hardship, or emotional hardship. while the data is not representative of or generalizable to the entire american population, the bank account challenges reported within are likely an underestimate of the problems, given that individual challenges may be resolved after contact with the financial institution; the individual may give up after contact does not result in a resolution; and not everyone with a challenge will submit a complaint (friedline & pawar, 2023). the complaints can reveal the consumer’s perspective on harmful, unethical, and/or illegal corporate activities, and provide insight into potential thematic or systemic failures, mistreatment, discrimination, and abusive policies and practices. study sample for this study, data filters were applied to narrow the millions of data records to specific financial services categories, specific consumer issues, and specific years. complaint data from the cfpb from august 24, 2022 to august 25, 2023 were downloaded (n=12,468 records). the present study's dataset was created by using the filters of “checking and savings accounts” and “complaints with a narrative and consent to be used in the public dataset.” for the sub-issues, the following filters were used: managing an account; closing an account; problem with lender or another company charging your account; opening an account; problem caused by your funds being low; problem with fraud alerts or security freezes; and incorrect information on your report. the dataset included all public responses from the company. the researchers sought a sample size that balanced efficiency and representativeness with sufficient data to draw birkenmaier & stratman 21 meaningful conclusions. the complaints were not listed in the dataset in any type of order, including by date submitted. therefore, a systematic sample of 20% of the dataset was examined by selecting every 5th complaint, starting with the first listed complaint, for a final analytic dataset of 2,493 records (krippendorff, 2019). complaints that fell into any of the following categories were excluded: no financial institution problem, related to a business or trust account, insufficient information, no current financial institution problem, and related to a settlement fund. the analytic sample size of consumer complaints was n=2,030. complaint demographic information as seen in table 1, the majority of the complainants were not an older adult, veteran, or member of the armed forces. the majority of the complaints were regarding checking, rather than savings, accounts, and were related to national banks and credit unions. fifty-eight different institutions were the subject of complaints. table 1. complaint demographic information category n % of complaints demographic none 1,617 79.65 older american 145 7.15 armed forces member or veteran 268 13.21 total 2030 100 product type checking account 1834 90.39 savings account 195 9.61 total 2030 100 financial institution types national bank 12 89.89 regional bank 21 3.77 credit union 8 4.69 financial services company 12 3.45 credit reporting agency 2 0.10 multi-services 1 4.09 other 2 0.10 total 58 100 financial services review, 33(2) 22 results content analysis results table 2 provides the meta-categories, thought domains and their percentage frequencies, and thought categories resulting from the content analysis. a total of five meta-categories and their thought domains emerged from the data: financial institution problem (eight domains), unsatisfactory customer service (five domains), financial hardship (five domains), emotional hardship (five domains), and customer vulnerability (three domains). three of these meta-categories financial institution problem, unsatisfactory customer service, and customer vulnerability are antecedents that precede the other two and may contribute to financial and emotional hardship. within the financial institution problem metacategory, the themes that emerged regarding the most common challenges consumers encountered in the use of their bank accounts ranged from transaction issues, to access issues, to policy and procedure issues. frequently mentioned issues were fraud or theft of their funds, excessive fees, and the unexpected closure of their account. the unsatisfactory customer service meta-category captured instances of insufficient or undesirable behavior from customer service staff, often weak engagement related to attempts to resolve the problem and poor quality information regarding the issue. the third meta-category, financial hardship, tracked how the financial institution problem may have negatively impacted the customer, either financially or legally. such impact is further explored in the emotional hardship meta-category, which captures the negative emotional experiences e.g., stress or frustration caused by the situation. finally, the customer vulnerability meta-category identifies whether consumers may have been particularly impacted by the problem based on existing vulnerabilities, such as demographic or situational factors. table 2. meta-categories, thought domains, and thought categories identified through content analysis of cfpb complaints related to bank accounts thought domain % thought categories meta-category: financial institution problem incentive 5.9 lack of action on promised account incentive/promotion account transaction 32.4 funds debited in account, product not received wrong account debited or credited transaction dispute account debited for authorized transfer, but lost funds unauthorized automatic debit unauthorized transfer account access and features 40.6 unable to view account online unauthorized account opening bank closed account bank failed to close account in a timely manner account frozen undesired change to account debit card shut off could not link account customer closed or opened an account to address problem, yet problem persisted birkenmaier & stratman 23 funds withheld 27.8 bank took or withheld funds excessive delays in returning funds to customer, including from fraud lacked provisional credit access during fraud investigation, or provisional credit initially given is revoked fees 22.9 charged overdraft fees charged fees with no advanced notice charged excessive fees charged incorrect fees charged unjustified fees overdraft fees charged when funds are available charged fees even when account features disallowed the fee account overdrawn or overdrafted fees – unspecified policy or procedure 41.6 inadequate procedures for preventing fraud or theft policy or procedure was unreasonable or inappropriate inadequate policy or procedure for account handling after owner was incapacitated due to death or disability undesired or unexpected account design feature repeated problem with delayed deposit clearing customer did not trust/feel safe with the financial institution fraud or theft 20.0 fraud or theft claim denied customer held financially liable or responsible for fraud/theft fees resulted from fraud atm 3.0 atm malfunctioned charged fees for multiple atm withdrawals due to limits meta-category: unsatisfactory customer service information 40.1 insufficient information or explanation shared did not provide requested information or explanation inconsistent or conflicting information provided on various contacts shared untrue or misleading information, lied or willfully deceived unfulfilled promise of help shared customer personal information without permission weak engagement or responsiveness 51.7 customer service unavailable insufficient communication responsiveness or delay delay or error in processing unreasonably long telephone hold time unwilling to listen, unempathetic lack of apology hung up on, no call back would not investigate financial services review, 33(2) 24 would not help no after-hours assistance wrongly routed call would not send paperwork, or lost paperwork undesirable engagement 9.4 rude or unprofessional staff behavior used sarcasm or laughed at the problem acted irritated or short with customer threatened retribution or intimidated customer blamed customer for problem used high pressure tactics asked invasive questions wrongfully disclosed sensitive customer information stole funds from the customer, tampered with the customer's account, opened an account in the customer's name, or otherwise used customer's information for personal gain gaslit the customer (‘you should be happy that..”) unable to solve problem 9.4 runaround in the customer service process could not speak directly to person desired customer repeatedly asked to fill out the same paperwork required information that customer cannot access asked customer to do something they cannot do discriminatory practice 3.4 perceived discrimination profiled customer physical accessibility issue to resolve problem (e.g., location) meta-category: financial hardship legal difficulties 0.3 under threat of criminal charges concerned about being taxed or legal issues tax filing challenges funds problem 63.2 unable to access money in account lost money or missing money account drained account overdrawn (re) opening account blocked 3.4 unable to open an account unable to reopen a closed account unable to pay/complete transaction 14.7 unable to pay bills or rent unable to make a desired purchase or payment borrowed money or received donations, favors, help could not complete or must cancel a transaction checks for authorized transactions bounced incurred fees from merchants related to late or absent payment went into debt to a merchant birkenmaier & stratman 25 blocked from credit or consumer reporting issue 5.6 limited or no access to credit concerned about credit record and/or score debt sent to collections or in consumer reporting system (e.g., chexsystems) meta-category: emotional hardship stress 18.1 described stress related to financial hardship emphasized consequences of financial hardship frustration 43.9 used of exclamation marks used all caps used rhetorical questions used sarcasm stated "mad," "angry," or "disgusted" threatened to sue the financial institution accused the financial institution of deception hung up on customer service staff used emphatic language such as “ridiculous,” “crooks,” “unacceptable,” “please help,” "tricked," "cheated," etc. embarrassed 0.4 stated “embarrassed” concern 1.9 stated “concerned,” “worried,” “disturbing,” “bothers me” concerned about professional reputation feared that money was not safe with the financial institution other 5.2 other emotions (e.g., “upset,” “disappointed,” “devastated”) described other emotional experiences (e.g., not being able to eat or sleep, feeling sick or shaking, etc.) meta-category: consumer vulnerability pre-existing financial shock 0.9 unemployed laid off fired divorced demographic factors 9.1 age (i.e., older adult, young adult) single parent racial minority less educated lower income living in rural area disability geographic factors (distance to a branch) citizenship status language previously or currently incarcerated mental illness financial services review, 33(2) 26 chronic medical issues woman transgender or nonbinary gender identity situational factors 1.4 power of attorney or guardianship beneficiary of an estate or account account co-owner with a deceased person intimate partner violence sick/hospitalized in a natural disaster homeless home destroyed recent birth in the family note: percentages add up to more than 100%, as complaints were coded in as many meta-categories and thought domains as needed. regression results to determine the relationship between financial hardship (the dependent variable) and the antecedent meta-categories, an analysis using generalized linear model ordinal regression was conducted. this method is appropriate due to the ordinal nature of the dependent variable. this approach allows for the handling of ordered categories while respecting their natural ranking without assuming equal intervals between them. the framework also provides flexibility in specifying link functions and distributions (agresti, 2010). to transform the meta-categories into measures that are needed to run regression analysis, each thought domain was quantified for each complaint report by summing the number of times a related thought category within the thought domain was mentioned or expressed in the complaint. this approach to coding content analysis output for purposes of subsequent quantitative analysis is consistent with prior consumer research studies (estelami & liu, 2023; pan & zhang, 2011). financial hardship, which served as the dependent variable in the regression analysis, was coded by adding the number of times each of the five underlying thought categories legal hardship (e.g., tax issue), funds problem (e.g., cannot access funds in account), account blockage (e.g., cannot open or reopen an account), inability to pay (e.g., cannot pay a bill), and/or credit blocked (e.g., consumer credit issue) was found in the complaint report. the resulting score ranged from 0 (in which case the complainant’s statement mentioned none of the thought categories) to 4 (after combining one and two categories). the antecedent variables were also quantified and used as independent variables. a dichotomous financial institution problem variable was created by noting whether or not each of the associated thought domains were expressed by the complaint (0 = not expressed, 1 = expressed) (i.e., incentive problem, account transaction problem, account features problem, funds issue, fee issue, policy or procedure issue, fraud issue, or problem with automatic teller machine (atm)). in the same way, a dichotomous unsatisfactory customer service variable was created by noting whether or not each of the associated thought domains were mentioned in the complaint (i.e., issues with provided information, weak engagement, undesirable action, inability to solve problems, and discriminatory behavior). dichotomous variables were also created for consumer vulnerabilities (i.e., pre-existing financial shocks, demographic factors, and situational issues). two demographic variables were identified in the complaint filing process and included in the model as control variables: older adults (0 = non older adult, 1 = older adult), and armed forces member or veteran (0 = nonmember, 1 = member). as can be seen from table 3, the model's log likelihood was -3273.54. the aic and bic were 3.24 and -11931.9, respectively. the model deviance was 3436.77, and the pearson chisquare was 3207.40. the likelihood ratio chisquare test (χ²(11) = 31.81, p < .001) suggests that birkenmaier & stratman 27 the model as a whole is statistically significant compared to an intercept-only model. regarding the thought domains, fraud-related issues were the most strongly associated with increased financial hardship. for each one-unit increase in the fraud score, the odds of being in a higher financial hardship category increased by 120% (or = 2.20, 95% ci [1.97, 2.46], p < .001), holding other factors constant. for each one-unit increase in the atm score, the odds of being in a higher financial hardship category increased by 79% (or = 1.79, 95 ci [1.42, 2.54, p<.001), holding other factors constant. similarly, for each one-unit increase in account transaction score, the odds of being in a higher financial hardship category increased by 71% (or = 1.71, 95% ci [1.55, 1.89], p < .001), holding other factors constant. an increase in the customer vulnerability score, account access and features score, and several customer service scores (i.e., unable to solve problems, information, and weak engagement) also result in higher odds (50% 12%) of being in a higher hardship category. neither of the two demographic variables captured through the cfpb complaint filing system, membership in the u.s. armed forces or veteran, nor being an older adult, has a significant relationship to financial hardship. no other variables were omitted from the regression analysis. financial services review, 33(2) 28 table 3. logistic regression results (n=2030) predictor odds ratio standard error z value significance level 95% confidence interval intercept 0.21 0.01 -28.68 0.00 0.19, 0.23 incentive problem 0.21 0.04 -9.27 0.00 0.15, 0.29 account transaction problem 1.72 0.09 10.82 0.00 1.55, 1.89 account feature problem 1.50 0.07 9.02 0.00 1.37, 1.64 funds withheld problem 1.70 0.08 11.31 0.00 1.55, 1.86 fees problem 0.72 0.04 -5.97 0.00 0.64, 0.80 fraud problem 2.20 0.13 13.86 0.00 2.0, 2.47 atm problem 1.79 0.21 13.77 0.00 1.97, 2.46 information problem 1.14 0.05 2.99 0.00 1.05, 1.24 weak engagement problem 1.12 0.05 2.63 0.00 1.03, 1.22 staff unable to solve problem 1.17 0.08 2.35 0.02 1.03, 1.334 consumer vulnerability 1.50 0.11 6.97 0.00 1.33, 1.68 note: (χ²(11) = 31.81, p < .001) birkenmaier & stratman 29 discussion this study aimed to (1) characterize consumer complaints about their basic bank accounts from large u.s. financial institutions, and (2) examine the relationship between consumer problems with basic bank accounts and financial hardship. consistent with the integrative consumer vulnerability framework and affect control theory, complaints contained emotional elements (emotional hardship meta-category), practical elements (financial institution problem meta-category), and relational elements (unsatisfactory customer service metacategory). based on the content analysis, the largest percent of consumer complaints about financial institution problems center on funds being withheld by the financial institution, policy and procedure problems, and bank account access. these findings are consistent with previous research about highly prevalent complaints (gyamfi et al., 2016; morgan, 2021; nndwamato, 2018; park, 2016). the largest percentage of customer service complaints are due to weak engagement and poor quality information, also consistent with prior research (cfpb, 2023a). the results of this study highlight the negative financial hardships associated with specific types of financial institution account challenges. findings suggest that problems with the largest impact on financial hardship are fraud or theft, the use of atms, and account transactions, which is also consistent with previous research (brenner et al., 2020; lim & letkiewicz, 2023). findings indicate that being a consumer with at least one type of vulnerability, such as experiencing a preexisting financial shock (e.g., unemployment, divorce), embodying or experiencing certain demographic factors (e.g., single parenthood, belonging to a racialized minority group), and/or situational factors also increases the odds of experiencing financial hardship. financial institution actions to mitigate customer financial hardship these results suggest that financial institutions can take actions to prevent and mitigate customer financial hardship related to their bank account. fraud or theft financial institutions freezing or closing accounts due to suspected fraud, then taking such action as reopening them without a consumer’s consent or knowledge (cfpb, 2023c), providing little information or recourse to consumers when limiting access to their funds, and levying fees and fines related to the account status was the topic of extensive comments in the dataset. combined with recent findings that the vast majority of customers whose accounts have been closed due to the suspicion of fraud are likely innocent of any wrongdoing (bank policy institute, 2018, 2024), these results can spur internal examination and discussion about policies, procedures, and software and related technology related to suspected fraud that align with federal law, and are also more responsive to customers. consumers also frequently mentioned dissatisfaction with bank fraud departments such as the lack of opportunity to communicate directly with the department staff, inability to learn about the evidence being examined in the case, or the inability to file a grievance (estelami & liu, 2024). other common fraud complaints related to scams included situations in which customers were scammed by thieves pretending to be bank staff, and lack of or slow bank reimbursement from fraud, which sometimes prompted additional fees from merchants. financial institutions could take actions to further educate customers about common fraud threats, and alter internal policies and procedures to provide more helpful responses to customers who have experienced such fraud (hsu, 2024). atm use although there was a small percentage of the complaints, the problematic use of atms significantly raised the odds of being in a higher financial hardship category. these study results highlight a specific area for which financial institutions may be able to take steps internally and fairly easily to reduce hardship. based on the content analysis, the most frequently mentioned issue was challenges with malfunctioning atms and the difficulty in resolving the issue with staff. specifically, a number of the complaints emerged after attempted atm deposits funds, after which the funds were not reflected in the account total. in these complaints, staff were often unable or unwilling to assist the customer for several financial services review, 33(2) 30 reasons that seemed unreasonable to the customer (e.g., the customer was told that the atm was owned by a separate company, the video of the activity was not available to the financial institution, or similar reasons). changing atm practices to increase accountability to the customer for deposits may reduce financial hardship and these types of complaints. customer service issues three customer service issues poor quality information, weak engagement, and inability to solve customer issues emerged as significant associations with financial hardship. for example, complaints included instances when marketing or account materials, signage, or staff provided wrong or inadequate information to customers; staff did not offer assistance; staff failed to follow through on verbal commitments to contact and/or assist the customer in some manner; or staff used unprofessional (i.e., rude or condescending) language and/or behavior in their interaction. customers also complained about the frustrations of dealing with unhelpful automated systems and/or the inability to speak to any human being. all these challenges point to the need for systemic changes in staffing, staff training, monitoring, as well as accountability for customer service complaints. changes at the institutional level or the regulation level that can significantly improve the (correct) information flow to customers, perceived engagement with their account issues, and their ability to problem solve in real time could decrease the rate of customer service complaints about service issues. policy implications these results provide empirical evidence that can be useful in policy work that impacts issues mentioned previously. for example, these results may assist in efforts by industry and advocates to reshape federal law regarding the ways financial institutions are tasked with monitoring for terrorist financing, money laundering, and tax evasion that can result in unnecessary financial hardship for consumers (the associations, 2024; bank policy institute, 2024; jelisejevs, 2021). among other challenging externalities, compliance with current law can result in incidents of financial institutions closing consumer bank accounts without notice or providing a reason to the consumer, resulting in financial hardship. financial institutions and regulatory agencies could improve fraud protection by requiring real-time fraud notifications, increasing consumer access to fraud departments, supplying guidance on technology challenges, and providing faster provisional credit during investigations (hsu, 2024). results also suggested the need to standardize atm functionality and accountability requirements. if regulators were to enforce standardized protocols across banks for handling atm issues, including prompt reconciliation, consumers could be assured that the problems highlighted in this study would be minimized (gyamfi et al., 2016). these results can also assist in policy efforts to improve customer service at banks. for example, the cfpb (2023d) issued an advisory opinion focused on consumer requests for information concerning their bank accounts. their ongoing efforts may address some of the issues, but more is needed to address other types of customer service challenges. the cfpb’s ongoing regulatory actions to hold banks accountable for opening fake accounts is critical for improving consumer trust (cfpb, 2024), but penalties must be examined to ensure that they are large enough to be a significant disincentive for repeat behavior. one possibility is to create an industrywide customer service standard for financial institutions that would establish and enforce minimum standards for timely responses, clear information on grievance resolution processes, and direct access to knowledgeable staff (lim & letkiewicz, 2023). study limitations and future research this study, while providing valuable insights, has several limitations. first, cfpb complaint data may not be representative of all bank account challenges experienced by consumers. those who file complaints may differ systematically from those who do not, potentially biasing the results. second, the study relies on self-reported data, which may be subject to recall bias or exaggeration. third, the cross-sectional nature of the data limits the ability to establish causal relationships between bank account challenges and financial hardship. birkenmaier & stratman 31 future research could address these limitations by incorporating longitudinal designs to track the progression of bank account issues and their impact on financial well-being over time. additionally, mixed-methods approaches, combining quantitative analysis with in-depth qualitative interviews, could provide a more nuanced understanding of consumers' experiences. studies comparing cfpb complaint data with other sources of consumer feedback (e.g., social media) could validate findings and identify potential reporting biases. finally, experimental studies manipulating different aspects of bank account management and customer service could isolate causal factors contributing to financial hardship, informing more targeted interventions by financial institutions and policymakers. conclusion the cfpb complaint data indicates that consumers experience considerable difficulty with their bank accounts, which is one of the many reasons they seek cfpb intermediation. notable is the fact that 70% of the cfpb filings used in this study experienced at least one financial hardship, yet only 17% of the complaints were closed with any monetary relief. this is despite the fact that a powerful federal entity (i.e., the cfpb) was involved. for less than one percent of the complaints did the financial institutions believe that the complaint presents an opportunity for customer service improvement. consumer complaints represent a wealth of information that can assist financial institutions to focus on opportunities to address concerns toward reducing customer financial hardship. references agresti, a. (2010). analysis of ordinal categorical data. wiley. bank policy institute. (2018, october 29). getting to effectiveness – report on u.s. financial institution resources devoted to bsa/aml & sanctions compliance. https://bpi.com/getting-to-effectivenessreport-on-u-s-financial-institutionresources-devoted-to-bsa-aml-sanctionscompliance/ bank policy institute. (2020). the truth about suspicious activity reports. https://bpi.com/the-truth-aboutsuspicious-activity-reports/ bank policy institute. (2024). statement for the record from the bank policy institute before the u.s. house financial services committee. https://bpi.com/bpistatement-for-the-record-for-houseoversight-hearing-of-fincen-and-theoffice-of-terrorism-and-financialintelligence/ bank policy institute, financial technology association, independent community bankers of america, american gaming association, securities industry and financial markets association. (2024). re: request for comments regarding reports by financial institutions of suspicious transactions and fincen form 111-suspicious activity report (omb control numbers 1506–0001, 1506–0006, 1506–0015, 1506–0019, 1506–0029, 1506–0061, and 15060065). https://bpi.com/wpcontent/uploads/2024/07/bpi-icbafta-sifma-aga-joint-tradecomment-letter-fincen-sar-30-daynotice-7-1-2024.pdf bartikowski, b., laroche, m., & richard, m. o. (2019). a content analysis of fear appeal advertising in canada, china, and france. journal of business research, 103, 232-239. https://doi.org/10.1016/j.jbusres.2019.01 .049 brenner, l., meyll, t., stolper, o., & walter, a. (2020). consumer fraud victimization and financial well-being. journal of economic psychology, 76, 102243. https://doi.org/10.1016/j.joep.2019.1022 43 chawla, i., russell, m., white, k., & devaney, s. (2023). fear and trust in financial institutions: a content analysis. financial services review, 31(2/3), 211–228. https://bpi.com/getting-to-effectiveness-report-on-u-s-financial-institution-resources-devoted-to-bsa-aml-sanctions-compliance/ https://bpi.com/getting-to-effectiveness-report-on-u-s-financial-institution-resources-devoted-to-bsa-aml-sanctions-compliance/ https://bpi.com/getting-to-effectiveness-report-on-u-s-financial-institution-resources-devoted-to-bsa-aml-sanctions-compliance/ https://bpi.com/getting-to-effectiveness-report-on-u-s-financial-institution-resources-devoted-to-bsa-aml-sanctions-compliance/ https://bpi.com/the-truth-about-suspicious-activity-reports/ https://bpi.com/the-truth-about-suspicious-activity-reports/ https://bpi.com/bpi-statement-for-the-record-for-house-oversight-hearing-of-fincen-and-the-office-of-terrorism-and-financial-intelligence/ https://bpi.com/bpi-statement-for-the-record-for-house-oversight-hearing-of-fincen-and-the-office-of-terrorism-and-financial-intelligence/ https://bpi.com/bpi-statement-for-the-record-for-house-oversight-hearing-of-fincen-and-the-office-of-terrorism-and-financial-intelligence/ https://bpi.com/bpi-statement-for-the-record-for-house-oversight-hearing-of-fincen-and-the-office-of-terrorism-and-financial-intelligence/ https://bpi.com/bpi-statement-for-the-record-for-house-oversight-hearing-of-fincen-and-the-office-of-terrorism-and-financial-intelligence/ https://bpi.com/wp-content/uploads/2024/07/bpi-icba-fta-sifma-aga-joint-trade-comment-letter-fincen-sar-30-day-notice-7-1-2024.pdf https://bpi.com/wp-content/uploads/2024/07/bpi-icba-fta-sifma-aga-joint-trade-comment-letter-fincen-sar-30-day-notice-7-1-2024.pdf https://bpi.com/wp-content/uploads/2024/07/bpi-icba-fta-sifma-aga-joint-trade-comment-letter-fincen-sar-30-day-notice-7-1-2024.pdf https://bpi.com/wp-content/uploads/2024/07/bpi-icba-fta-sifma-aga-joint-trade-comment-letter-fincen-sar-30-day-notice-7-1-2024.pdf https://bpi.com/wp-content/uploads/2024/07/bpi-icba-fta-sifma-aga-joint-trade-comment-letter-fincen-sar-30-day-notice-7-1-2024.pdf https://doi.org/10.1016/j.jbusres.2019.01.049 https://doi.org/10.1016/j.jbusres.2019.01.049 https://doi.org/10.1016/j.joep.2019.102243 https://doi.org/10.1016/j.joep.2019.102243 financial services review, 33(2) 32 https://doi.org/10.61190/fsr.v31i2/3.354 1 chebat, j. c., ben-amor, k., & davidow, m. (2010). why bank customers dissatisfied with service recovery remain loyal: an affect control theory approach. journal of marketing trends, 1(1), 77-83. https://citeseerx.ist.psu.edu/document?re pid=rep1&type=pdf&doi=2e7d38a47f92 78568fb7f82cc1b7ce0095d4c193 cohen, j. (1968). weighted kappa: nominal scale agreement provision for scaled disagreement or partial credit. psychological bulletin, 70(4), 213-220. https://doi.org/10.1037/h0026256 commuri s. & ekici, a. (2008). an enlargement of the notion of consumer vulnerability. journal of macromarketing, 28(2), 104200. https://doi.org/10.1177/0276146708316 consumer financial protection bureau (cfpb). (n.d.-a). building the cfpb. https://www.consumerfinance.gov/dataresearch/research-reports/building-thecfpb/ consumer financial protection bureau (cfpb). (n.d.-c). cfpb financial well-being scale: questionnaire https://files.consumerfinance.gov/f/docu ments/bcfp_fin-well-being_fullscorecard.pdf consumer financial protection bureau (cfpb). (n.d.-b). consumer complaint database. https://www.consumerfinance.gov/dataresearch/consumer-complaints/ consumer financial protection bureau (cfpb). (n.d.-c). learn how the complaint process works. https://www.consumerfinance.gov/comp laint/process/ consumer financial protection bureau (cfpb). (n.d.-d). consumer complaint program. https://www.consumerfinance.gov/comp liance/consumer-complaint-program/ consumer financial protection bureau (cfpb). (2020). consumer response annual report 2019. https://files.consumerfinance.gov/f/docu ments/cfpb_consumer-response-annualreport_2019.pdf consumer financial protection bureau (cfpb). (2021). consumer response annual report 2020. https://files.consumerfinance.gov/f/docu ments/cfpb_2021-consumer-responseannual-report_2022-03.pdf consumer financial protection bureau (cfpb). (2022). consumer response annual report 2021. https://files.consumerfinance.gov/f/docu ments/cfpb_2021-consumer-responseannual-report_2022-03.pdf consumer financial protection bureau (cfpb). (2023) consumer response annual report 2022. https://files.consumerfinance.gov/f/docu ments/cfpb_2022-consumer-responseannual-report_2023-03.pdf consumer financial protection bureau (cfpb) (2023, june 6). chatbots in consumer finance. https://www.consumerfinance.gov/dataresearch/research-reports/chatbots-inconsumer-finance/chatbots-inconsumer-finance/ consumer financial protection bureau (cfpb) (2023c). consumer financial protection circular 2023-02: reopening deposit accounts that consumer previously closed. https://www.consumerfinance.gov/comp liance/circulars/consumer-financialprotection-circular-2023-02-reopeningdeposit-accounts-that-consumerspreviously-closed/ consumer financial protection bureau, (cfpb). (2023, october 16). consumer information requests to large banks and credit unions. federal register, 88(198), 71120-71124. https://www.federalregister.gov/docume nts/2023/10/16/2023-22774/consumerinformation-requests-to-large-banksand-credit-unions https://doi.org/10.61190/fsr.v31i2/3.3541 https://doi.org/10.61190/fsr.v31i2/3.3541 https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=2e7d38a47f9278568fb7f82cc1b7ce0095d4c193 https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=2e7d38a47f9278568fb7f82cc1b7ce0095d4c193 https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=2e7d38a47f9278568fb7f82cc1b7ce0095d4c193 https://psycnet.apa.org/doi/10.1037/h0026256 https://doi.org/10.1177/0276146708316049 https://www.consumerfinance.gov/data-research/research-reports/building-the-cfpb/ https://www.consumerfinance.gov/data-research/research-reports/building-the-cfpb/ https://www.consumerfinance.gov/data-research/research-reports/building-the-cfpb/ https://files.consumerfinance.gov/f/documents/bcfp_fin-well-being_full-scorecard.pdf https://files.consumerfinance.gov/f/documents/bcfp_fin-well-being_full-scorecard.pdf https://files.consumerfinance.gov/f/documents/bcfp_fin-well-being_full-scorecard.pdf https://www.consumerfinance.gov/data-research/consumer-complaints/ https://www.consumerfinance.gov/data-research/consumer-complaints/ https://www.consumerfinance.gov/complaint/process/ https://www.consumerfinance.gov/complaint/process/ https://www.consumerfinance.gov/compliance/consumer-complaint-program/ https://www.consumerfinance.gov/compliance/consumer-complaint-program/ https://files.consumerfinance.gov/f/documents/cfpb_consumer-response-annual-report_2019.pdf https://files.consumerfinance.gov/f/documents/cfpb_consumer-response-annual-report_2019.pdf https://files.consumerfinance.gov/f/documents/cfpb_consumer-response-annual-report_2019.pdf https://files.consumerfinance.gov/f/documents/cfpb_2021-consumer-response-annual-report_2022-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2021-consumer-response-annual-report_2022-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2021-consumer-response-annual-report_2022-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2021-consumer-response-annual-report_2022-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2021-consumer-response-annual-report_2022-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2021-consumer-response-annual-report_2022-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2022-consumer-response-annual-report_2023-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2022-consumer-response-annual-report_2023-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2022-consumer-response-annual-report_2023-03.pdf https://www.consumerfinance.gov/data-research/research-reports/chatbots-in-consumer-finance/chatbots-in-consumer-finance/ https://www.consumerfinance.gov/data-research/research-reports/chatbots-in-consumer-finance/chatbots-in-consumer-finance/ https://www.consumerfinance.gov/data-research/research-reports/chatbots-in-consumer-finance/chatbots-in-consumer-finance/ https://www.consumerfinance.gov/data-research/research-reports/chatbots-in-consumer-finance/chatbots-in-consumer-finance/ https://www.consumerfinance.gov/compliance/circulars/consumer-financial-protection-circular-2023-02-reopening-deposit-accounts-that-consumers-previously-closed/ https://www.consumerfinance.gov/compliance/circulars/consumer-financial-protection-circular-2023-02-reopening-deposit-accounts-that-consumers-previously-closed/ https://www.consumerfinance.gov/compliance/circulars/consumer-financial-protection-circular-2023-02-reopening-deposit-accounts-that-consumers-previously-closed/ https://www.consumerfinance.gov/compliance/circulars/consumer-financial-protection-circular-2023-02-reopening-deposit-accounts-that-consumers-previously-closed/ https://www.consumerfinance.gov/compliance/circulars/consumer-financial-protection-circular-2023-02-reopening-deposit-accounts-that-consumers-previously-closed/ https://www.federalregister.gov/documents/2023/10/16/2023-22774/consumer-information-requests-to-large-banks-and-credit-unions https://www.federalregister.gov/documents/2023/10/16/2023-22774/consumer-information-requests-to-large-banks-and-credit-unions https://www.federalregister.gov/documents/2023/10/16/2023-22774/consumer-information-requests-to-large-banks-and-credit-unions https://www.federalregister.gov/documents/2023/10/16/2023-22774/consumer-information-requests-to-large-banks-and-credit-unions birkenmaier & stratman 33 consumer financial protection bureau (cfpb). (2023e). consumer response annual report. https://files.consumerfinance.gov/f/docu ments/cfpb_2022-consumer-responseannual-report_2023-03.pdf consumer financial protection bureau (cfpb). (2024, [july 9]). cfpb takes action against fifth third for wrongfully triggering auto repossessions and opening fake bank accounts. https://www.consumerfinance.gov/about -us/newsroom/cfpb-takes-actionagainst-fifth-third-for-wrongfullytriggering-auto-repossessions-andopening-fake-bank-accounts/ dou, y., hung, m., she, g., & wang, l. l. (2024). learning from peers: evidence from disclosure of consumer complaints. journal of accounting and economics, 77(2-3), 101620. https://doi.org/10.1016/j.jacceco.2023.1 01620 edelman. (2022). how distrust and economic fears are contributing to headwinds in the financial services sector. https://www.edelman.com/trust/2022trust-barometer/trust-financial-servicessector estelami, h. & liu, k. (2023). content analysis of american consumers’ credit card fraud complaints filed with the consumer financial protection bureau. journal of financial crime. https://doi.org/10.1108/jfc-03-20230070 federal deposit insurance corporation (fdic). (2019). how america banks: household use of banking and financial services. https://www.fdic.gov/analysis/househol d-survey/2019/2019report.pdf federal deposit insurance corporation (fdic). (2022). 2021 fdic national survey of unbanked and underbanked households. https://www.fdic.gov/analysis/househol d-survey/2021execsum.pdf federal trade commission (ftc). (n.d.). bureau of consumer protection. https://www.ftc.gov/about-ftc/bureausoffices/bureau-consumer-protection friedline, t. & pawar, a. (2023). can a public bank redress the private banking industry’s harms to chicago’s residents? an analysis of the consumer financial protection bureau’s complaints database [policy brief]. poverty solutions, university of michigan. https://sites.fordschool.umich.edu/povert y2021/files/2023/01/final-terri-pb2023.pdf garrett, d. e., & toumanoff, p. g. (2010). are consumers disadvantaged or vulnerable? an examination of consumer complaints to the better business bureau. journal of consumer affairs, 44(1), 3-23. https://doi.org/10.1111/j.17456606.2010.01155.x goodstein, r. m., & kutzbach, m. j. (2024). the effect of job loss on bank account ownership. journal of money, credit and banking. https://doi.org/10.1111/jmcb.13127 gyamfi, n. k., mohammed, m. a., nuamahgyambra, k., katsriku, f., & abdulah, j. d. (2016). enhancing the security features of automated teller machines (atms): a ghanaian perspective. international journal of applied science and technology, 6(1), 102-112. halvorsen, j. l. & møkkelgård, p. j. (2018). financial literacy and other determinants of consumer complaint behavior: an empirical study utilizing the cfpb’s complaint database. [master's thesis, norwegian school of economics]. nhh brage. hayes, r. m., jiang, f., & pan, y. (2021). voice of the customers: local trust culture and consumer complaints to the cfpb. journal of accounting research, 59(3), 1077-1121. https://doi.org/10.1111/1475679x.12364 https://files.consumerfinance.gov/f/documents/cfpb_2022-consumer-response-annual-report_2023-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2022-consumer-response-annual-report_2023-03.pdf https://files.consumerfinance.gov/f/documents/cfpb_2022-consumer-response-annual-report_2023-03.pdf https://www.consumerfinance.gov/about-us/newsroom/cfpb-takes-action-against-fifth-third-for-wrongfully-triggering-auto-repossessions-and-opening-fake-bank-accounts/ https://www.consumerfinance.gov/about-us/newsroom/cfpb-takes-action-against-fifth-third-for-wrongfully-triggering-auto-repossessions-and-opening-fake-bank-accounts/ https://www.consumerfinance.gov/about-us/newsroom/cfpb-takes-action-against-fifth-third-for-wrongfully-triggering-auto-repossessions-and-opening-fake-bank-accounts/ https://www.consumerfinance.gov/about-us/newsroom/cfpb-takes-action-against-fifth-third-for-wrongfully-triggering-auto-repossessions-and-opening-fake-bank-accounts/ https://www.consumerfinance.gov/about-us/newsroom/cfpb-takes-action-against-fifth-third-for-wrongfully-triggering-auto-repossessions-and-opening-fake-bank-accounts/ https://doi.org/10.1016/j.jacceco.2023.101620 https://doi.org/10.1016/j.jacceco.2023.101620 https://www.edelman.com/trust/2022-trust-barometer/trust-financial-services-sector https://www.edelman.com/trust/2022-trust-barometer/trust-financial-services-sector https://www.edelman.com/trust/2022-trust-barometer/trust-financial-services-sector https://doi.org/10.1108/jfc-03-2023-0070 https://doi.org/10.1108/jfc-03-2023-0070 https://www.fdic.gov/analysis/household-survey/2019/2019report.pdf https://www.fdic.gov/analysis/household-survey/2019/2019report.pdf https://www.fdic.gov/analysis/household-survey/2021execsum.pdf https://www.fdic.gov/analysis/household-survey/2021execsum.pdf https://www.ftc.gov/about-ftc/bureaus-offices/bureau-consumer-protection https://www.ftc.gov/about-ftc/bureaus-offices/bureau-consumer-protection https://sites.fordschool.umich.edu/poverty2021/files/2023/01/final-terri-pb-2023.pdf https://sites.fordschool.umich.edu/poverty2021/files/2023/01/final-terri-pb-2023.pdf https://sites.fordschool.umich.edu/poverty2021/files/2023/01/final-terri-pb-2023.pdf https://doi.org/10.1111/j.1745-6606.2010.01155.x https://doi.org/10.1111/j.1745-6606.2010.01155.x https://doi.org/10.1111/jmcb.13127 https://doi.org/10.1111/1475-679x.12364 https://doi.org/10.1111/1475-679x.12364 financial services review, 33(2) 34 hsu, m. j. (2024). acting comptroller of the currency michael j. hsu remarks for the financial literacy and education commission’s public meeting: banks’ role in addressing fraud against consumers. comptroller of the currency. https://www.occ.gov/newsissuances/speeches/2024/pub-speech2024-75.pdf jelisejevs, a. (2021). crucial issues with legal protection of consumers human rights when banks unilaterally close accounts. athens journal of law, 7, 617-634. https://doi.org/10.30958/ajl.7-4-11 kabadayi, s. (2016). customers’ dissatisfaction with banking channels and their intention to leave banks: the moderating effect of trust and trusting beliefs. journal of financial services marketing, 21, 194– 208. https://doi.org/10.1057/s41264016-0005-2 kamran, s. & uusitalo, o. (2019). banks’ unfairness and the vulnerability of lowincome unbanked consumers. the service industries journal, 39(1), 65-85. https://doi.org/10.1080/02642069.2018. 1436704 kim, y., lee, w. n., & jung, j. h. (2013). changing the stakes: a content analysis of internet gambling advertising in tv poker programs between 2006 and 2010. journal of business research, 66(9), 1644-1650. https://doi.org/10.1016/j.jbusres.2012.12 .010 klapper, l. & lusardi, a. (2019). financial literacy and financial resilience: evidence from around the world. financial management, 49(3), 589-614. https://doi.org/10.1111/fima.12283 krippendorff, k. (2019). content analysis: an introduction to its methodology. sage. lee, n. t., seddon, j., tanner, b. & lai, s. (2022). why the federal government needs to step up efforts to close the rural broadband divide. brookings. https://www.brookings.edu/articles/why -the-federal-government-needs-to-stepup-their-efforts-to-close-the-ruralbroadband-divide/ lecoeuvre, l., turner, r., & kuppelwieser, v. g. (2021). customer experience in the b2b area: the impact of age-related impressions. journal of retailing and consumer services, 58, 102216. https://doi.org/10.1016/j.jretconser.2020 .102216 lim, h., & letkiewicz, j. c. (2023). consumer experience of mistreatment and fraud in financial services: implications from an integrative consumer vulnerability framework. journal of consumer policy, 46(2), 109-135. https://doi.org/10.1007/s10603-02309535-w mogaji, e. (2020). consumers’ financial vulnerability when accessing financial services. research agenda working papers, 3, 27-39. https://ssrn.com/abstract=3562889 morgan, r. e. (2021). financial fraud in the united states, 2017. u.s. department of justice. https://bjs.ojp.gov/content/pub/pdf/ffus1 7.pdf national bureau of economic research. (2023). banks that relied on branches, not remote depositors, fared better in recent turmoil. https://www.nber.org/digest/202310/ban ks-relied-branches-not-remotedepositors-fared-better-recent-turmoil nndwamato, m. (2018). influence of automated teller machine service quality on overall customer satisfaction in retail banking sector (order no. 28284652). [doctoral dissertation, university of johannesburg]. proquest dissertations & theses global. o'connor, g. e., newmeyer, c. e., wong, n. y. c., bayuk, j. b., cook, l. a., komarova, y., loibl, c., ong, l. l., & warmath, d. (2019). conceptualizing the multiple https://www.occ.gov/news-issuances/speeches/2024/pub-speech-2024-75.pdf https://www.occ.gov/news-issuances/speeches/2024/pub-speech-2024-75.pdf https://www.occ.gov/news-issuances/speeches/2024/pub-speech-2024-75.pdf https://doi.org/10.30958/ajl.7-4-11 https://doi.org/10.1057/s41264-016-0005-2 https://doi.org/10.1057/s41264-016-0005-2 https://doi.org/10.1080/02642069.2018.1436704 https://doi.org/10.1080/02642069.2018.1436704 https://doi.org/10.1016/j.jbusres.2012.12.010 https://doi.org/10.1016/j.jbusres.2012.12.010 https://doi.org/10.1111/fima.12283 https://www.brookings.edu/articles/why-the-federal-government-needs-to-step-up-their-efforts-to-close-the-rural-broadband-divide/ https://www.brookings.edu/articles/why-the-federal-government-needs-to-step-up-their-efforts-to-close-the-rural-broadband-divide/ https://www.brookings.edu/articles/why-the-federal-government-needs-to-step-up-their-efforts-to-close-the-rural-broadband-divide/ https://www.brookings.edu/articles/why-the-federal-government-needs-to-step-up-their-efforts-to-close-the-rural-broadband-divide/ https://doi.org/10.1016/j.jretconser.2020.102216 https://doi.org/10.1016/j.jretconser.2020.102216 https://ssrn.com/abstract=3562889 https://bjs.ojp.gov/content/pub/pdf/ffus17.pdf https://bjs.ojp.gov/content/pub/pdf/ffus17.pdf https://www.nber.org/digest/202310/banks-relied-branches-not-remote-depositors-fared-better-recent-turmoil https://www.nber.org/digest/202310/banks-relied-branches-not-remote-depositors-fared-better-recent-turmoil https://www.nber.org/digest/202310/banks-relied-branches-not-remote-depositors-fared-better-recent-turmoil birkenmaier & stratman 35 dimensions of consumer financial vulnerability. journal of business research, 100, 421-430. https://doi.org/10.1016/j.jbusres.2018.12 .033 pan, y., & zhang, j. q. (2011). born unequal: a study of the helpfulness of usergenerated product reviews. journal of retailing, 87(4), 598-612. https://doi.org/10.1016/j.jretai.2011.05.0 02 park, c. (2016). consumers and mobile financial services. board of the federal reserve, division of consumer and community affairs. https://www.federalreserve.gov/econres data/mobile-device-report-201203.pdf polat, a., ajmal, m. m., & rafay, a. (2023). mortgage financing frauds in the us: an analysis of cfpb complaint database. in a. rafay (ed.), concepts, cases, and regulations in financial fraud and corruption (pp. 1-27). igi global. salisbury, l. c., nenkov, g. y., blanchard, s. j., hill, r. p., brown, a. l., & martin, k. d. (2023). beyond income: dynamic consumer financial vulnerability. journal of marketing, 87(5), 657-678. https://doi.org/10.1177/0022242922115 0 scott, a. (2023). financial abuse in a banking context: why and how financial institutions can respond. journal of business ethics, 1-16. https://doi.org/10.1007/s10551-02305460-7 warmath, d., elizabeth o'connor, g., wong, n., & newmeyer, c. (2022). the role of social psychological factors in vulnerability to financial hardship. journal of consumer affairs, 56(3), 1148-1177. https://doi.org/10.1111/joca.12468 https://doi.org/10.1016/j.jbusres.2018.12.033 https://doi.org/10.1016/j.jbusres.2018.12.033 https://doi.org/10.1016/j.jbusres.2018.12.033 https://doi.org/10.1016/j.jretai.2011.05.002 https://doi.org/10.1016/j.jretai.2011.05.002 https://www.federalreserve.gov/econresdata/mobile-device-report-201203.pdf https://www.federalreserve.gov/econresdata/mobile-device-report-201203.pdf https://doi.org/10.1177/00222429221150910 https://doi.org/10.1177/00222429221150910 https://doi.org/10.1007/s10551-023-05460-7 https://doi.org/10.1007/s10551-023-05460-7 https://doi.org/10.1111/joca.12468 financial services review, 32(2) i volume 32 issue 2 from the editor one of the pleasures associated with editing financial services review (fsr) is that i have the privilege of seeing the future of the financial services profession unfold in real-time using a research lens. from a personal point of view, the papers in this issue of fsr have had a profound and positive impact on my thinking and my classroom teaching. consider the paper written by drs. chuck grace, adam metzler, yang miao, longlong feng, and alireza fazeli (unveiling the winning contribution patterns for enhanced financial health). for years, i have wondered about household savings rates. as grace et al. point out in their invited paper, savings rates in canada have been declining for two decades. it turns out that one reason is that investors today may be focusing too much on returns rather than the act of saving. as noted in this paper, the "winning strategy" for accumulating wealth over the lifespan involves engaging in consistent saving behavior. the work of grace et al. offers important implications for investors, policymakers, asset managers, and financial advisors. the work of dr. yu zhang (esg perceptions: investigating investor motivations and characteristics) answers another question i've been asking for several years. while managing portfolios using an environmental, social, and governance (esg) perspective has gained a large following among portfolio managers and some financial advisors, the notion of using esg policies when selecting securities is still controversial, with some asking, “what is the ultimate purpose?” in dr. zhang's paper, readers learn how using an esg perspective can add value to the portfolio management process. dr. zhang documents the investor factors associated with esg adoption. going further, dr. zhang answered my question of why esg matters. esg policy adoption acts as a hedge against governmental and environmental shocks. stated another way, esg investing is essentially a risk reduction strategy. this is certainly something that i will incorporate into my teaching and publications. over the past five decades, researchers have spent a great deal of effort studying retirement planning issues. one might think that all the questions have been answered. well, not quite. as drs. afm jalal ahamed and yam limbu show in their paper (retirement planning: a moderated mediation model of cognitive beliefs, retirement planning attitude, and money availability), much of what the field knows comes from data obtained from consumers living the north america, europe, australia, and new zealand. an important question is, "what about the rest of the world?" drs. ahamed and limbu provide some insight into answering the question. the paper examines the role of cognitive factors associated with retirement planning intentions from a developing country perspective. they found that retirement planning attitudes mediate the relationship between cognitive factors and retirement planning intentions. they also noted a negative association between risk tolerance and retirement planning intentions through attitudes. financial self-efficacy had a positive influence. drs. ahamed and limbu's work also shows how the availability of financial resources moderates these relationships. i hope that this paper will prompt other researchers to expand their perspective when thinking about retirement planning issues. the final paper in this issue was written by drs. john young, crystal hudson, and c. w. copeland. their paper (factors mediating the association between financial socialization and well-being: an african american perspective) answers an important question; namely, how is well-being established, financial services review, 32(2) ii particularly in the african american community? these researchers evaluated the relationship between financial socialization and well-being (financial and subjective) mediated by three motivations: (a) financial knowledge, (b) goal setting, and (c) self-control. using comparisons with european americans, the gudmunson and danes financial socialization framework, and the fisher and fisher information motivation behavior model, they found a difference between african americans and european americans in relation to self-control. specifically, they found that self-control mediates directly through financial behaviors and indirectly through financial skills. in their models, goal setting was also important, as was financial knowledge. as one reviewer noted, more studies of this type are needed to help advance the well-being of all individuals and households. the conclusions from this paper have direct policy and practice implications. i am going to switch gears here and give an update about fsr. fsr was recently added to cabells scholarly analytics list of journals. i am in the process of taking steps to petition other indexing services to include fsr. with a track record going back 32 years, with some of the biggest names in the field (e.g., harry markowitz, william sharpe, sherman hanna, and others) having published in fsr, i certainly hope that fsr can make a compelling argument for expanded indexing. when you get a chance, please visit the fsr website (https://openjournals.libs.uga.edu/fsr/index). you will see announcements for two special issues that will be published in 2025. • the first is being co-edited by drs. peter öhman, mustafa nourallah, and izidin el kalak. if you have a paper that deals with the topic of "navigating contemporary fintech solutions" please consider submitting your paper. • dr. barry mulholland is serving as the editor for the second issue, which deals with the "pedagogy of financial planning." this is your opportunity to share, in a peer-reviewed outlet, your work related to the development and use of case studies, teaching activities, pedagogical models, experiential leaving techniques, the use of ai and other tools in the classroom, curriculum development, cross-disciplinary perspectives on teaching, and addressing diversity, equity, and inclusion in financial planning education. another important announcement is that you can now access all fsr back issues and papers on the journal's website. this means that anyone, anywhere, has complete open access to the fsr archives. this also means that google scholar searches will include more full-text fsr outcomes. for the record, the archiving process took hundreds of volunteer hours. who were these volunteers? incredibly, this was accomplished by shawn brayman and wookjae heo (and dr. heo's graduate students). if there is ever an award for the most underappreciated (and overworked) volunteers in an academic setting, shawn and wookjae are deserving of a prize. my deepest thanks to these two "fsr superheros." until next time, all the best, john e. grable, ph.d., cfp® editor pii: s1057-0810(00)00046-9 effective teaching and use of the constant growth dividend discount model thomas h. paynea, j. howard finchb,* adepartment of finance, the university of tennessee at martin, martin, tn 38238, usa bdepartment of finance, the university of tennessee at chattanooga, 615 mccallie avenue, chattanooga, tn 37403, usa abstract the appropriate application of the constant growth dividend discount model (ddm) requires an understanding of the fundamental nature of the model and its parameters. it is important that students not only be able to mechanically “plug and chug” the formula, but that they also understand the model’s assumptions, inputs, sensitivity to error and practical limitations. this paper demonstrates that the valuation measure derived from using the ddm is very sensitive to the relationship between the required return on investment (ks) and the assumed growth rate (g) in earnings and dividends. examples show that the valuation error increases at an increasing rate when the values of ks and g converge in the formula. classroom experience has indicated that students believe and strive to compute a single “correct” valuation of the share price. they should realize that the goal of valuation analysis is to estimate a reasonable range for the intrinsic value of a share price, rather than a single point estimate as often implied by end-of-chapter and exam-type problems using the ddm. © 1999 elsevier science inc. all rights reserved. jel classification:a200; g120 keywords:dividend discount models; asset pricing; stock valuation; valuation models 1. introduction the theoretical soundness and practical simplicity of the constant growth dividend discount model (ddm) (gordon & shapiro, 1956; gordon, 1962) have led to its extensive * corresponding author. tel.:11-423-755-5250; fax:11-423-785-2329. e-mail address:howard-finch@utc.edu (j.h. finch). financial services review 8 (1999) 283–291 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00046-9 application for common stock valuation. students are usually introduced to the formula and its conceptual framework in their first finance course. in more advanced courses, they apply the model to security analysis, to cases involving security issuance and mergers and acquisitions, and to other valuation related problems. while most students eventually become comfortable with mechanically “plugging and chugging” the formula, they often have little understanding of its practical limitations. this paper provides a method for illustrating the nature of estimation and a means of demonstrating the nonlinear sensitivity of the ddm to variations in required rate of return (ks) and the growth rate (g) estimates. nearly all texts discuss the basic assumption that ks must be greater than g for the model to hold. this requirement is predicated on the practical limitation that a stock’s price must be non-negative. similarly, ks must be greater than g since equivalence would result in an infinite value. together, the economic constraints that stock prices are non-negative and finite make the model’s assumption of ks . g fairly easy for students to grasp. what is not necessarily intuitive is how the relationship between ks and g affects the estimate of the stock’s intrinsic value. once the mathematical and economic constraints of the model have been considered, students should realize that proper implementation of the ddm requires more than the calculation of a single point estimate. students, expecting to arrive at a “right answer,” often fail to recognize the distinction between an estimate and a known price. this is not a trivial point since students often have the misconception than any answer arrived at mathematically represents absolute truth. for all applications of the model, but particularly in case studies, students should be required to provide an in-depth explanation of all assumptions, parameter estimation methods, and conclusions. only in this way will they develop a true grasp of the model and its limitations. 2. background valuation error resulting from implementation of various forms of the ddm has been addressed in numerous studies. jacobs and levy (1988) note that the ddm expected returns are not generally predictive and sometimes negatively correlated with actual returns. hickman and petry (1990) find that dividend discount approaches produce errors averaging 88% of the actual price, and 4.21 times those of price-earnings methods. in addition, a high degree of error in growth and required return estimates is found irrespective of specific modeling assumptions. gehr (1992), noting that price estimation bias in the ddm is the result of required return and growth prediction error, proposes application of a probability weighted range of the parameter estimates. finally, good (1989) points out that, since the next period dividend is largely a known quantity, the reliability of the ddm is primarily dependent on the estimation of required return and growth rates. it is this very point which requires increased emphasis during classroom discussions of the ddm. 3. sensitivity analysis formally, the constant growth dividend discount model is given by: v0 5 d1/(ks 2 g) (1) 284 t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 where v0 is the estimated intrinsic value per share, ks is the required rate of return, g is the forecasted growth rate in earnings and dividends (assuming a constant payout ratio), and d1 is the forecasted next period dividend. the assumptions of this model are that g will be at a constant rate for the foreseeable future, and that ks . g. following good (1989), practical application of this model requires the estimation of two key inputs, ks and g. because many texts first cover the valuation formula and only later the various techniques to determine these inputs, students often miss the key relationship in the denominator of the formula. for illustrative purposes, define this relationship by: z 5 (ks 2 g) (2) consider a hypothetical valuation problem where the expected next period dividend (d1) 5 $1.50, ks 5 22%, and g5 8%. then z5 14%. applying the ddm, students will estimate the share price by: v0 5 ($1.50)/(.222 .08) v0 5 $10.71 many students will instinctively stop here, assuming the problem is “finished” (and often laboring to check their “answer” with that of a classmate or solutions manual). however, it is important to convey that the value of v0 is anestimateof intrinsic value and to make clear that it is not the “price” of the stock. table 1 extends the example problem, allowing for different estimates of ks and g. note that, as the difference z gets smaller, the resulting estimate of share value grows larger. specifically, as ks and g converge, the valuation estimate increasesgeometrically. fig. 1 illustrates this relationship graphically. table 1 the constant growth dividend discount model: the effect of changes in z5 ks 2 g on value estimation (d1 5 $1.50) z 5 ks 2 g value estimate (v0) absolute difference (percentage difference) incremental cumulativea 14% $10.71 2 2 12% $12.50 $1.79 $1.79 (16.7%) (16.7%) 10% $15.00 $2.50 $4.29 (20%) (40%) 8% $18.75 $3.75 $8.04 (25%) (75%) 6% $25.00 $6.25 $14.29 (33%) (133%) 4% $37.50 $12.50 $26.79 (50%) (250%) 2% $75.00 $37.50 $64.29 (100%) (600%) a cumulative differences are computed relative to the initial (z5 14%) value. incremental differences are relative to immediately preceding values. 285t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 suppose the market is in equilibrium, so that the actual value per share (which is unknown) is, in fact, the original estimate of $10.71. fig. 1 shows that as z gets smaller,valuation estimate increases at an increasing rate. thus, students should be aware of how sensitive the model’s estimate of intrinsic value is to the relationship between ks and g. 4. moving from textbook examples to “real-world” applications investment textbooks and other resources illustrating financial analysis techniques often provide the necessary model parameters. in other words, the numbers are “given” and it is left to the “analyst” to simply perform the math. as an example, suppose a class was given the following data to use in the estimation of the intrinsic value for a firm’s share price. 4.1. information for computing the required rate of return on investment the risk-free rate of return in the market is 5%, the expected return for the market is 14%, and this firm’s beta is 1.10. the yield to maturity on its long term outstanding bonds is 8.25%, and management estimates an 8% premium for stocks over bonds. 4.2. information for estimating the growth rate for earnings and dividends in 1996 the firm’s earnings per share (eps) was $2.00, and in 1999 eps was $2.32. the firm maintains a 50% dividend payout ratio, and 1999 return on equity (roe) was 11%. analysts who follow the firm’s stock estimate earnings growth of 6 to 8% in the next few years. fig. 1. valuation estimate as ks and g converge. 286 t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 then students, following textbook methodology, should estimate the ddm parameters as follows: 4.2.1. estimating ks method 1 – the capm: ks 5rf 1 [km 2 rf]ß (3) ks 5 5%1 [14 2 5%]1.10 ks 5 14.90% method 2 – bond yield plus risk premium approach: ks 5 ytm 1 risk premium (4) ks 5 8.25%1 8% ks5 16.25% so, the average estimated required return would be: ks 5 (14.9%1 16.25%)/25 15.58% 4.2.2. estimating g method 1 – retention growth model: g 5 b(roe) (5) g 5 .50(11%) g 5 5.5% method 2 – point to point estimate: fv 5 pv(11 g)n (6) $2.325 $2.00(11 g)3 g 5 5.07% method 3 – analysts’ forecast: low analyst estimate: g5 6% high analyst estimate: g5 8% average analyst estimate: g5 7% giving equal weight to each technique, the average estimated growth rate in earnings and dividends is: g5 (5.5%1 5.07%1 7%)/3 5 5.85% applying the ddm, we know that last year’s dividend d0 was $2.32 (0.50)5 $1.16, giving a share price estimate of: 287t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 v0 5 $1.16(1.0585) .15582 .0585 v0 5 1.2279 .0973 v0 5 $12.62 investment texts and other resources that outline these various methodologies generally do so in a very discrete fashion. they do not adequately address the range of estimates arrived when different methodologies and assumptions are applied to the same problem. the main point which students should be aware of is how sensitive this estimate of intrinsic value is to z, the relationship between the estimates of ks and g. reworking the problem, using both the highest and lowest input estimates, gives the following range of values for v0: v0 5 $1.16(1.08) .16252 .08 5 $15.19 v0 5 $1.16(1.0507) .1492 .0507 5 $12.39 students, working on assigned problems or an exam problem, will likely panic at these divergent results, but a practicing security analyst would be very comfortable with a conclusion such as “we feel the shares to be fairly valued in the $12 to $15 range.” 5. non-constant dividend growth another application of the ddm regularly covered in finance courses is the nonconstant, or two-stage growth model. this method simply incorporates multiple growth rates into the share valuation analysis. the nonconstant growth model involves three consecutive steps: 1) estimate the dividends in the high growth period(s) individually, 2) estimate the share price at the final growth phase using the constant growth ddm, and 3) discount the future cash flow stream back to the present at the required rate of return and sum. changing the previous problem slightly, suppose the firm in the example is expected to grow earnings and dividends at a 15% annual rate over the next three years, with subsequent growth slowing to the previously estimated average rate of 5.85% annually. then the dividends for the next three years are forecasted to be d1 5 $1.16(1.15)5 $1.33 d2 5 $1.33(1.15)5 $1.53 d3 5 $1.53(1.15)5 $1.76 the estimate of share price would then be given by: 288 t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 v0 5 1.33 1.1558 1 1.53 (1.1558)2 1 1.76 (1.1558)3 1 1.76(1.0585)/(.15582 .0585) (1.1558)3 v0 5 1.151 1.151 1.141 12.40 v0 5 $15.84 typically, students spend the majority of their time on the first and third steps in solving the problem. however, instructors should emphasize the contribution of the constant growth estimation of share price at the end of the high growth phase to the overall share estimate. in this problem, this step contributes $12.40/$15.845 78% of the total share value. here again, the emphasis should be on estimating a reasonable range for the price at the end of the high growth period (end of year three), given its significance to the overall share valuation estimate. 6. an applied example from portfolio management class examples from student-produced stock valuation reports highlight these issues from an individual financial management perspective. classes are offered at the authors’ respective institutions in which students manage actual securities portfolios worth over $150,000 each. to make new stock recommendations, students are required to produce research reports that include an individual firm analysis, an industry analysis, and a quantitative valuation of the firm’s share value using the nonconstant dividend growth model and other techniques. the following example illustrates how students in our classes have applied the nonconstant growth model to value a firm in 1998. the first period growth rate, g1, was estimated using a four-year point to point estimate as 13.29%. the students’ assumption was that this growth rate could be maintained for three more years. the estimated dividend stream at this growth rate for the next three years was $0.70 per share in 1998, $0.79 in 1999, and $0.89 in 2000. a scenario analysis was performed using a range of required rates of return (ks) and secondary growth rates (g2). following are the share valuation estimates. scenario 1:ks 5 13.5%, g2 5 12% present value of g1 stage dividends: 0.70/(1.135)1 0.79/(1.135)2 1 0.89/(1.135)3 5 $1.85 estimated price at the end of year 2000: (0.89)(1.12)/(0.1352 0.12)5 $66.45 present value of price 2000: 289t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 66.45/(1.135)3 5 $45.45 estimated current share price: $1.85 1 $45.45 5 $47.30 (3.91%) (96.09%) (100%) scenario 2:ks 5 14%, g2 5 12.25% present value of g1 stage dividends: 0.70/(1.14)1 0.79/(1.14)2 1 0.89/(1.14)3 5 $1.82 estimated share price at end of year 2000: 0.89(1.1225)/(0.142 0.1225)5 $57.09 present value of year 2000 price: $57.09/(1.14)3 5 $38.53 estimated current price: $1.821 $38.535 $40.35 (4.51%) (95.49%) (100%) scenario 3:ks 5 14.5%, g2 5 12.5% present value of g1 stage dividends: 0.70/(1.145)1 0.79/(1.145)2 1 0.89/(1.145)3 5 $1.81 estimated share price at end of year 2000: 0.89(1.125)/(0.1452 0.125)5 $50.06 present value of year 2000 price: $50.06/(1.145)3 5 $33.35 estimated current share price: $1.811 $33.355 $35.16 (5.15%) (94.85%) (100%) a couple of points are worth emphasizing in this actual application of the nonconstant growth dividend discount model. first, assuming a maturing firm, out-year growth rates for each scenario are lower than the 13.29% point-to-point estimate for g1. secondly, as the uncertainty about the future growth rate increases so does the investor’s required rate of return. since the assumed growth rate after year 2000 was less certain, the required rate of 290 t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 return was increased in scenario 2 to reflect this higher risk. finally, following gordon and gould (1978), this uncertainty generally makes ks increase with g. based in part on more optimistic industry and economic assumptions, higher growth rate and required return estimates were used in scenario 3. the student analysts can readily see that iterations have a profound effect on the value of the estimated share price. in this case, the overall price ranges from $35.16 to $47.30. as the proficiency with the models and their underlying assumptions improves, it is very useful to place the staged growth model and parameter estimate formulas into an excel spreadsheet. by doing this, estimates for ks and g derived from the various models can be directly referenced in the constant or staged growth ddm. this eliminates duplicative calculations and allows students to change basic model assumptions and immediately see the impact on the price estimate. students in the portfolio management classes regularly express their surprise (and occasional frustration) at how small changes in their assumptions for ks and g can result in a wide range of current price estimates. 7. summary while a very powerful tool for estimating value, proper application of the constant growth dividend discount model requires an understanding of the fundamental nature of the model and its parameters. the ease of calculation makes the model intuitively appealing for finance students. however, they should be able not only to “plug and chug” the ddm formula, but also to understand the model’s inputs and sensitivity to the relationship between the required rate of return and the growth rate. students should be aware that the sensitivity of the ddm to error increases geometrically when the estimates of ks and g converge. since estimates for ks and g can vary widely depending upon the estimation methods used, it is both insufficient and impractical to merely calculate a value based on a single set of narrow assumptions. rather, students need to recognize that implementation of the model means estimating a set of feasible values to arrive at a range of intrinsic value per share for which the student (or analyst) can feel confident. references black, f. (1986). noise.journal of finance, march,529–544. gehr, a. k. (1992). a bias in dividend discount models.financial analysts journal, january/february, 75–80. good, w. r. (1989). bias in stock market valuation.financial analysts journal, september/october, 6–7. gordon, m. j. (1962).the investment, financing and valuation of the corporation.richard d. irwin. gordon, m. j. & gould, l. i. (1978). the cost of equity capital: a reconsideration.journal of finance, 33, 849–861. gordon, m. j., & shapiro, e. (1956). capital equipment analysis: the required rate of profit.management science, october,102–110. hickman, k. & petry, g. h. (1990). a comparison of stock price predictions using court accepted formulas, dividend discount, and p/e models.financial management, summer, 76–87. jacobs, b. i. & levy, k. n. (1988). on the value of ‘value’.financial analysts journal, july/august, 47–62. 291t.h. payne, j.h. finch / financial services review 8 (1999) 283–291 pii: s1057-0810(02)00098-7 ��� ����� ��� �� �� �� � � �� �������� � ����� �� �� �� ��� ��� ������ ���� �� �� ��� �� � � � ���� ! �������� � ����� ���������� ���� ��� ����� ��������� � � �!�� � "��� ��� #$ %�� ����� #&&'( ���� ��� � ��� ��� ���� ') *������� #&&'( ���� �� '+ �������� #&&' �������� ��� ����� ��� �� �� �� � � ��', ��.� � � � ��� ������� �� �� � � ���, � ������ �� �� � � �� � ����� � ������ �� � ��/� � � � �� ���� �� ���� �� ���� �� ����� � ������� �� � �� � �� � � ��, .��0� � �� ��� ��� ������ �������, �� ����� � �� �� ����� ��� � �� �� � � �� �� ��� ������� �� �1�� �� � � ��� � ������ -� � � �� ����� �� �� �� ������ ��� ��� 2� �� -�� � � ��� � � �� �� ���� �� �1�� �� � �3� � �� ������ ��� ���� � ��� ���� ��� ��� ������� �� � � � ��� � � ���� �� �� ����-�� �� �� ��� ��� ������ ������� ��, � � � ���, ��0 �� �� � �� ��� ��� � � �� ����� 2� �� �� % �� ��� ��� �� �� �� � � ��� � #&&' �-��� �� %� ���� 4��� �-� �� � ��������� "#$ %�����&�%�����' �&&( �5& (��)��*�' ������( 2��.( ��� �� � � �� � ���� ������ ��� ��� ��� ������ ������� � 0 ������, 0�--6� � �� ������� ��� ���� �� � �� � � �� 7��-�� �� 8����, '++9( : �. , '++5;� <�0����, � �� � � �� � � 0�-������ ��� � ��� � �� � ��, ���� ���, �� ������� �� �� �� �� ��� �� ������� ��� �� ����� ����� �� ����� � � ��� ��� �� ��� 2� ��6��� � ���� ����� �� ���� -� � ����� � ���� � ��� � � ���� ������ �� �� �� �� ��� ��� ������ ������� 7���� , #&&';� � ������� ���� � � � ����� �� ��� ��� ������ ������� �� �� ��� �� � �� �� ����� ��, � � � ���, ��0 �� �� � �� ��� ��� � � �� ����� 2� �� �� % �� ��� ��� �� �� �� � � ��� � � �� %��� ��� "�� �0 '& 7#&&'; ')5/'+= '&=56&)'&>&'>? / ��� ���� � �� � #&&' �-��� �� %� ���� 4��� �-� �� � ��������� �44� % ' & = 5 6 & ) ' & 7 & # ; & & & + ) 6 5 ! ��-�� �'6)5&6#@=69@&9( � �� �'6)5&6#@=69)&&� #+,��� �**����' 1� ��� a� �� ����� 7���� ���� ;� ' �����, ��� ��� � � �� �������, �� �� ��� �� �� � �.� �0��� � �� ��� ��� � � �� ����� 2� �� �� % �� ���, 4��� ����� � �.� �� 0 ���� � �� � �� -� 0�� ���������-����-� � �� ��� 2� ��b� � �� ���� �� ��� �� �� ��3� ����� �� %������������-� �� �� �� ����� ��� �� �� �� � � �� 7 �� cc�� �bb;, -��� 0 � ��� � � ��� ��3� ����� �, -� �� � �� ��� ��� �� �� ��� � �� � �� � � � �� � �� ��� �� ��� ��� �� �� � �� ���� ��� .�� ��� �� ���� ������� �� �� �� 0���� �� � ��>� � �. � �-� ������� �� ���� d�� 7����, 5&e �� ���� � � �� ��;� <�0����, � ����� �� �� ����� � � �� � ��, �� ��� 2� �� ���� �� ����� � ����� � ������, �� � ��� � �� � �� �� � � ��� ��� � �� � � � � -� ����� �� � � ��0 0�-7�� ����; �� �� ��� � �( ���, ��0 � � � � �� � �� �� � �. �� � �� �� �� ������ *� ���� � �� � �� �� .��0 �� � �����, �� �� -�� �� �� .��0 �� ��3� ��� � �� �� �����, � �� ��� 2� �� ���� �� ����� �� ������ �� � ����� � ������ �� ��� �� � � �� � �� �� �� ��� ��� 0�� 0��. � �� �� �� ��� �� ���� �� ����� �� ����� �� ���� � � ������ ��, 4 0 - ��� � ���� ���� �� �� �� ��� ��� ������ ������� � � ����� � �� � � ���� �� �� ������ ��� � ��� ����� ������������ ����������� 2 ���� �� � ��, ���� �� �����@=&,&&& �� � �� -� � �� f�%� 0�� � - �����-��� �� �� ������� g� ����, ��@),&&& �� ��� ��� �� � �� g� � � -, � �� � �� ��� ��� ��� �� �� ����-� �� ��� �� �� � � ��� � � ���1�� �� � � ��� � ����� � ����-� � �� ���� �� �������� �� �� ��� �� �� �������� ��� ���1�� �� � 0���� �� �� ��� �� ���� ��� -�( ���� ���( ���� ��� �( ����� � ��( �� ����� ��� �� �� ���-� �� ����� �( �� �� � ��� �� �� � ��� 2�� �� �� � '++#, � �� � �� ���. �� ��� �� �� 0��� ��3� ��� � .� '&6� �� � �� �� ���� � �� ���1�� �� � �� � � ��� �� � � �� ��� ���� � �� ���������� �� �� � � �� 0 � 1�� ��� � ��� �� �� � ��� �� � ��� �� �� �� �� ��� �� �������� ���������� �� �� �� ��� �� ��3� ��� �� ����� � �� -� � �� � � ��� �� � �� � � �� ����� � �� �� ��� � .�� ��������� ���� ��� �� � � �� � � ���� ���� ��� ��� � � �� �� � �� �� � �� % ��� �� � ���, �� �� *�������� ��� �� � � ������ ���� �� � 9 � �� ���� �� 0� ���� ��� �� @ � � �� �� % ��� � � �� �� � ��� �� ����� � �#)= 3��� ��� ��-�� �� � �� ��� 2� ��b� ��������� �� � �� � � � ��, �� �� ���� 8���� 4� ��� �� -� �� � ��-���� � �� 9=/== 3��� ��� ���� � ���� ��� � � � 1�� � ��� 0 � �� � ��� �� �� 3��� ��� �� �� � ��6 -��� �� �-� ��� ��, � �� ��, �� ���� � ���� 3��� �� �� �� �� ������ � �, =,&&& � �� � �� ��� � � .� �� �� � �� ��� �� ��� �� � �� � � ��� ��� � =@e� � ��� ���� �� ����������� ������������ ������� �� �� �� � - �� � � �� � �� � �� � � .��� �� ���� ��( ���, � --�0 ���� 0�� �� ������3� ��� � �� � �� � �� ������� �� �� � .��� �� ���� 0�� �� �� � ��� �� �� � 70� ��, �� ������, �����b �� � ; 0��-� � ��� � � � �� 0��� �� 3� ��� ������ ���3� ��� ������ 0 � ����-� � ���� � � %��� � �� � 0��-� ����� � ������3� ��� � �� � � �� ����� � �� � ��3� ��� � �� � �� ��� �� � � ��� �� ')) "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! �� ���� ���� � �� � ��3� � �� � � ��� -����� ����� ��� � �� � � ��� ������� �� 7%� �����, '++#;� � � � ������� �� �� � � �� �� � �� ��� �� �� � ��-�� � � �� �� � �� �� � � � ��� ������� ��� � �� ��� 2� ��, ��� �� � � ���� �� �������� ����� � � ���� = � �� � ��� �0 �� ��� �� �� � �� �� �� ��� ��� � ��� �� ��� � �� ���, ��-� ��� �� ��� �� ��� � �� ��� �� .��0-���� � ��� �� �� �� �� ���� ��� � �� ���� � � ���� �� 0��. �� �� ��� �� � � ���� ��, 3��� ��� �� � ��� � ',&&& �� �� ��� �� � ����� 0���� �� ���� �� ��� �� 0� � ��-�� � � �� ������� ��� ��� ����� �� �� ��� �� � � ���� �� �� ���� ��� 0 � �� ��� ���� �� ��� �� � ����� � ����-�� �� � � ��� �� � � �� ��������� ��� �� � � , .��0 � 6�7�%� &�� 80� 8����9%����� #��,��������, � ��� ������� � �� ��� 2� ��b� 8� �� � ��� �� �� 7#&&'; ��� �� � � �� � �� � �� � ��� �0� ����� �� � � � - 0� ���� ���� � �� �� � � �� ���� �� �������� ����6���������� �� �� �� �� �� ��6���������� �� �� 7� ����, '++&;� *���6���������� �� �� ��� �� �� ������� ��� �� ���� ��� �� �� � �� ���� ��� �� ��� ���� � � �� ���� ������ �� �� � �� � �� �� �� � �� �� � � � �� ��� ��� �� �� � �� � � � �� �� �� �� �� cc8� � �� �� �� �����bb � 0�--6.��0� �� ��� �� ����6���������� �� ��� ��� � 1�� ����-�� 0 � ����6���������� �� �� � � � �� � �� �� �� ��� ��� 0 ��� ��� �� � ����� ��� -���-�� �� � � � ��� ���� 0 � �� � �6.��� �� ���� �� �� --�0 �� ��3� ��� � �� � �� � �� � �� � � ��� ��-�� 4� ��� � � � ����6���������� �� ��, �� �� ��6���������� �� �� � �� � �� ������ � �� � ��� �� ���� ���-� � -����� .��0-���� � �� ��-�� <�0 � �� � � �������� �� �� � � ��� �� .��� � �� ���� � � " ���, �� ���� � 0�� ��� � � �� � � ��� � � ���� � �� -����� .��0-���� � 0��-� �� �-� �� �� ��� � �� � �� 0 ��� � �� �� �� ����� ��� ���� � ��� �� ����6���������� ������ �� �� ��6���������� �� ��, ���� ��� �� ��� ��� ���� ������ � �� � � �� �� � � ���� � �� ���� � � 4�� � � �� ����� ��� �4��� � � �� -�� �� � �� �� ����� 5=e ��� � �� �� �� ���� ��� ��� �� �� ��� �� �� � �� � �� ���� �� � � -����� ���� 0��� ����������� � � � ��� � � �� �� �� �� ��� �4��� ���� �� ���� �� �� � @&e �� � �� � �� . �� � � ��� ��� �� �� �� �� � ����-� � �� 7 � �� ��-� �� ����6�������� ��;� 4�, ��� �� ��-�, �� #=e � ���� ���� �� ��� ��, �� � �� �1�� ��� �� � � 0��� ���� �� @&e �� �� � �� � �� � � � ��� ��� #=e � � �� � � � � �� ���� �� ���� �� ���� � ���� �� �� g� �� � ��� � ��, ���-� � �� ���� �� ���� �� ����� �� ������ ��� �� 0� . � �� � ��� <�0����, ���� � � �. � �� �� �� � � � --�0 �� ��� =e �� � �� � �� � � ��� %��� �� � � �1�� ��� �� ������ � --�0 � �� � �� � ������ ��� ��� � �� �� ������3� ��� � �� � �� ����� ��� ���� �� � ������ �� �� �� �� � � � ��� �� �� @&e � �� � �, �� �4��� ��� � � �� �� �� � �� �� ��6���������� �� � � ��� � '++5 78 ��� , ����� �� %0���� , '++);� � � ���� ����� � � �� ��-� 7 � � � � ��, ����; �� ��� ��� �� � �� � � � ��� �� � ������� �� 0� �� � � ��� .��0-���� ��3� ����� � �� �� �� �� ���������� �� �� ������ � �� � �� �� ��0 � ���� � ��� �� �� 3� � �� .��0-���� �� ��� � , �� �� ��0 �� � � �� �� � �� � ���� . �� �� �� �� � �� � � �� �� ��6���������� �� �� �������� � �� � '+5&� �� � � �� �� �� �� ��0 ��� � �� ��� ������� �� ������ �� �� � 7��-�� �� 8����, '++9;� "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! ')+ ! ������ ����� ����" �� ��� ����� ���� ��� ������ �� ��� � � ��� ��� h -� ������, ���� ��� �� �� � ���-� �� � ���� �� ��� �� %��� ��� � �� '+$&�� ������ �� ����-�� ���� �� �� ��6���������� � �� ��, ��� �� � �� �� �� �� �� � �� � � �� �� � � ��� � �� �� ������ ��� ������ �� ��� �� �� ��� �� � �� �� ����� � ��� ���� � ��-� �� ��� �� 0�� ��� �� �� ������� �� �� �� �� � � ��� ��-� � � �� � � ��� -����� ����� ��� ������ � � �� � �� � � ��-�� ��� �� ��� �� � � ��, � � ��� -����� ����� ���� �� �� � ������ � �-� � �� � �� ���������� �� �� �� ��� �� 0 ��� ������ � ��� ������ 7'+5'; ������ �� ��� 0 � �� ��� �� �� � � ��� -����� ����� ���� 0 � � �����-� � ���� ����� � � � � � ��� ��-� � �� -� � �� �� �� �� 3��� ���� � � ���� 1����� 0��-� �� -� � �� 3��� ��� � � �� ���� � � � � � - ���� �-� � �� � � 0��-� ��0�� �� 3��� �� ������ �� � ����� .� -�� � ���� �� �� : �. 7'++5; �� �� � �. �� �� 1����� � �� � � ���� ���� � �� � �� � 1�� �� � �� �� �� ������ �������� � � �� ��� �� �� � �����, .� � ��� ������� �� -�, ��� � ���� � � ���� d�� �� �� �� �� � ��� �� ���� � � �� � �0 � � ��� �� � ��� �� 3��� �� � � � � .��0-���� � -�� �� ������ ����-� .��0� <�0����, � ���� ��� ���� � � ���� �� �� � �� � �� �� � ��� �� �� � �� � �0 �� 3��� �� � ���� ��� �-�� ����, �� 0� �� � ����� � �� � �0 �� � �� 3��� �� � -� ������ 0 � ������ � � � ���� -���-� � �� ��� 2� ��, � � � ���� ����� � �� ���� 0 � �� 3��� �� �� -� �� � �������� � � �� � �� %���� ���� ��� ��� � �� %���� ���� �� � � ����� ����� �� ��� �� � ������ ��� � 1�� � � ��� �0 �� ���� �� 0� ��� �� �� �� �� ����� 0��-� ���� � .��0 � � ��� -����� ����� ���� � �� � �� ��� ��� �� ��� ��� � �� %���� ���� �� ��0��� �� 3��� ��� cch� ������ �� ��� �� �� � � � -����� �� ��� �� � ����� 0��-� .��0 �� ��0�� � � � ��� 3��� ��ibb # $��� ���� ��� �� --�� � � �� ������ �������, ���� ��� �� ��--�0 �� 3��� ��� ��-� ��� � �� ��� 2� �� ���� �� �� � �. � � �� � '+++� '+& "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! ��� �� � � �� �� �� ������ 0��-� ������ d� � � � � � � 3��� �� ��� �� �� �� � -�� �� �� � � �� h� -� �� ����-�� ��� �� ���� � � � �� ������ � � ��0�� �� 3��� �� 7� � � � ���;, � � � � �� � �� ���0 �� � -�� �� � ��� �� � � �� �� �� ������ ����-� �� �-� � � .�� ��� ������� �� �� � �� %���� ���� �� � �� ����-��� �� ��� �� � )&e �� � � � -����� �� ��� �� � ����� 0��-� .��0 � � � � �� �� �-� � �� �� ��� � ����� ��� ������ ��, � � 3��� �� � �� �� �� ���� � ������ � �� �� &�)& 0��� � -�� �� � � �� � � ��� � �� �� ������ 7*� � � �� ����� � � ��� � � � �� ��-� �� � �0� �� ��-�� �� �� � ��� �� � � � ��������; *�0 ���� ��� �� ��--�0 �� 3��� �� ��� �� �� �� � ��� ��, -�� ��-� ��� � �� ��� 2� �� � � �� � '+++� �� � 3��� �� ��3� ��� ���� ��� ����� �� -� �� � � �� ���� ��� 3��� ��� %�-�� �� � � � ��� ��� ���0 0 � <�0����, �� � ��� ���� ��������� �� -�� �� ��-� ���� ��� � �� � �� 0��-� ������ d� � �� ��� � � -� �� � � ���-� �� ���� � � � � � ��� ��� 3��� �� �� �� �� � ����� � ��� ��� �� ��� �, � �� �, ��3� � �� �� �� � � �� 4� � ����� � ������ �, 0� � �� ������ ���� ��� ��� 2 -� � � ���i � ������ �� � �� � �� � ������ d� � �� �� �, �� ?$#=,&&& 7 �� ���-�� �� ���� � '++); 0��-� �� � 0 � �� ���� 2 -� b� �� ��� ���� � <�0����, � ��� �� ?',=&&,&&& 0��-� � d� ���� 2 -� b� ���-�� �� ���� �-�� �3� d� �� �� �, � ���� � ���� ��� �� � ��� �� � � � �� ��� 2 -� b� ���� �� ����� �� �� �� �������� �� � ��� �� �� � �-��� �� � 3��� �� ��3� ��� �� � �� � � � �� �� ��� ��� ������ �� � � ��� � �� �� �� ������ ��-� �� �� � � � d �� �� � � �� �� � �� ��� ��� ��--�� �� 1������ �� �� � �� %���� ���� �� � �� �� � ��@&e �� � � � -����� �� �� � � �� ��� ��� �� � � 0��-� ������ �� �� � � ����-�� ��� ����, � � 3��� �� 0��-� �� �� ���� � ������ � �� �� &�@&� % ��� �� ������� �� �� �� � � � � ������ � �� 0��� ���� 0�� ��� �� � � ��� 1�� .��0-���� ��3� ����� � �� ����� � �� ��� 2� �� �� ���� 0�� ���b , �� ���� � � �� �� ����� ���� 3��� ��� �� �� ���� � 3��� �� � � � �� ��0���� � �� � � -���� � �� � � ���� -� � � ������ � �� 0��� 3� ��� �� ��3� ��� � ���� �� *� ��� ���� 3��� �� � � �� ���� � �� ������� � � � � -���� �� �� � �� � �� � � ��0�� � %�, 3��� ��� 0 � �� ����� �� �� -�0 � ���� �� �� ����� - �� ��-���� �� �� �� �� "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! '+' �� � �� ���� ���, � �� �� � ��� �� ����� � �#)= 3��� ��� 0 � 9=/== �� ���� �� �� � �� 3��� ���� j��� ��� �� � ��� �� �� 0� �� � �� �� ���� � � 1�� � �� � ������� � � ��6 -���� % ��6 -��� 3��� ���, ���� � �� 0� �� ��-�� � ��� ����, �� 0� �� �� 0� �� � � � ��� � �� 3��� ��� �� 0� �� �� ���� �� � � � ��� � � ��6 -��� 3��� �� 0 � � ������ � �� �� &�5 0��-� �� � � �� � 0 �� ����� '�9 � �� � � 7&�5� # �� � �;( � �� 3��� �� 0 � �� � �� ���� � 0��-� �� #�' �� � � 7&�5� @ �� � �; � �� �� ������ � � �� �� � �� ��-� 7���� , #&&'; �� � -�� �� �� � � ��� � �� �� ����� ��� � �� � � �� � � ��� � � �-� '� ��� �� � � ��� � � �� 3��� ��� � ��, �� � �� � � 0��-� � �� � ����� � � ��� �� 9&'�#& �� � � � ���� ���� � �� �� � ��6 -��� �� � �� 3��� ���� *� � � �� � ���� �-� �� � � �� �� �� � � � -�� �� � ��--�0�� % ��6 -��� 3��� ��� #@9 0� �� � � � �� 9$) � �� 3��� ��� =' ���� �� � � � �� '=@ �� ���� �-� �� � � $#' � �-� ' ����� ��� �� �� �� � � �� � -�� �� �� � � ��� � �� �� ����� � �� �� � �� ��-� *����� �� 3��� ��� ������ � �� �� 7 0� �� � � � ��; % ��6 -��� 3��� ��� + &�#& @�$& '# &�@& 5�#& '= &�9& '#�&& #' &�9= ')�+& #@ &�=& #@�&& #9 &�$& #)�)& '5 &�$= ##�'& #' &�5& #+�9& #9 &�5= @$�&& #@ &�)& @$�)& #& &�)= @9�&& #= &�+& 9=�&& %�� � #@9 #+$�)& � �� 3��� ��� 7 ���� �� � � � ��; # &�@& '�)& @ &�9& @�$& 9 &�=& $�&& ) &�$& '9�9& $ &�$= ''�5& $ &�5& '#�$& = &�5= ''�#= $ &�)& '9�9& 5 &�)= '5�)= 9 &�+& '&�)& %�� � =' '&9�9& �� #)= 9&'�#& '+# "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! ���� � � �� ��-�, �� � �� � � � ����-��� � �� � �� ����� � ������ �� ���� � $9�$e 79&'�#&>$#';( ���, ����� � �')9 3��� ��� �� �� #)= ���� � �� ��0���� ������ � 4 � ������� � � .� ���� � ��������� ��� � �� � � ���-� �� ��� �� ������ � �� ����� � ��0�� �� � 0���� ���� '5= � #&& 3��� ��� ������ , ������ �� �� �� � � �� � ��6 -��� 3��� ��� ������ � �� 3��� ��� ��0���� ������ � *� � � � � � � �� �� � �� ����� ��� � � � � ��� ����� �� 3��� ���� 4� �� � � �� � �� � ���� � ���� 3��� ���, �� �� ����� 0��-� ���� ��� ��� �� ���� -�0 ������ � ���� ��������, � �� � � �� � �� � � � �� 3��� ��� 7���� � �� ������ � ���;, �� �� ����� 0��-� � ��� ����, � ����0� �� � ��-��� � -. ��� � 0 ������ �� ����� 0��� �� �� ������� �� � ����� 0� � �� � ��b ������ �� �� ����� ��( ���� ����� �� 0��-� ���� ��� ���� �� �� ������ �� -���-� ��� ���� � �� � � � � �� �� �� 0� � � � � -��3� ��� � � � �� ��� ��� �� �� � ���� � �� ��� �� � �� �� ������ % ��� &��' � (������� ���� ��� � � � ��� �� � � � �� �� �� �� � � ��-��� �� ����� �1�� ��� �������� � � 2��. ������� �� �1�� ��� 72�� ��, '++@;� ��� 2��. �1�� ��� ���-��� ���� � �� � � ����� ����� �� ������ ��� � �� %���� ���� �� ������� �� �.�� �� � �� - � � .� ��� �� �� �� � �� �� ������ �� � �� � �� � ����-� � �� �� �� �� 4� ������ �� � � �� ', �� � � �� �3� �� ��� �� �� � ��� �� �� � ����� �� �� �� �� � � -�� �� � �� � � 7������ �� � �� � �� ����� �� � � ��; ����-�� 0 � �� �� � �� '� < �� �� � 2��. �1�� ��� � "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! '+@ ����� �� � � ����� �� � � �-� '� ��� � ��� � ���� ������ �� ��� �� � �-��� � � � �� � �� �� ��� ��� ������ � %-��� �� � ��� � �� �� � � �� � �� �� �� ����� �� � �� � 7'; ����, �� �-��� �������� � �� � � � �� �� 1�����b � �� � � �� � �� � �� �� � �� � � � �� �� 1�����b � �� ����� �� � ��� ��� ������ �-� ���0� �� ������ �� �� � �� � �� � �� �� � �� ������ �� ����� ���� -�0 � � �� 7 ���, �� ��� � � �� �����;� 4� ���� � � ����� �� �� 0��� �� � � ����� �� �� ����� �� �����, �� � �� ����� � �� �1�� �� ��0�0 �� �� ��0 ��� 4� � � �� ��-�, �� � �� ����� � �� �1�� �� ��0�0 �� ���� $9�$e 7 �� � -�� ���� �� �� ��-� � � �-� '; � $#e �� �� ������ �� ��� ��0���� ������ � � ��� ��� ��� � �� � ����� ���� �� � ����� �� �� �� ������ �-� �� �� � ��� �������� � �� 2��. �1�� ��� � ��� �1�� ��� --�0� ��� �� � � � 1�����b �� � �� � ���� � �� �� � ���� � � ����� �� ���� ������� �� � �� ������ � � 4 ����-� �� ���� � d�� � ���� ������ �� �� ��-�� �� �� �� �������� �� �� � � ����� ��� �� �� �� � � �� ����� �� ���� ���� ) ����������� �*������ �� �� ��6���������� �� � � �� ��������� ���� � �� ������ �� ��� --�0 ��� �� ��� ��-�� �� ����� � � ��� � 0��� �� ����� �� ��� ����� � � �� ����6���������� �� ��� � �� , �� -����� � �� � � ���� ������� � � � 0 � ���� ������ �� �� �� �� ���� ��� � � � ����� 4� �� ��, �� � � ���� � � � ���� �� � �� ������ �� �� �� � � �� 3��� ���� ��� ���� -�� � ���� � ��� � � �� � � � �� �� �� �� � � �� �� ��� � �� � �� -����� � � ��� ��3� ��� ����� ��� � �� � �� � � �� # �� = � �� ��� �� � ��� �� � �� ���� � �� � ��� ��-� ��� ���� � � ��� ���� �� � �� � �� � � �� �3� �� ������ � �� ������� 0����� � � 6�� � � �������� 0 � � �� ���� -�� � 72�� ��, '++=; �� � � � 6�� � ��� �� 3��� ��� ���� �� ���� ��� �� ��, -�� �� �� ��� � ��� � �� ������ �� � � � ���� ��� �� ��� ����� � � 6�� � 3��� ��� �� ������ �� ���� ��� � ���� ��� �� ��� % ��� �� 3��� ��� �� ������ � ���� ��� �� ��, � � ��� � �� � � 6�� � ������ �� .���� � �� ��� � � �-�, �� -����� � �� � � ���� �������� ���� ��� ��� �� � �� �� ����� �� �� �������� 3��� ��� � � � 7'; ������� ��� �3� � � �� �� �� �� �� � � 0� �� ���� 3��� ��� ��� ���, �� 7#; ����� �� �� � �� � � ��� �� � ���� � � � �� � � �� � � ���� ��� �3� ������� ��� �� �������� 3��� ��� 7 �� � � 6�� �; �� � �� �� � �� � � ���� �� � �3� � ���� � �� � ���� ��� ������� 2� �� � ����6-���������� ��� �� �� � �� � �� � �� �� ���� -��0 �� � � �� ���� � ����0� � � ��� ����, �� �� ����� 0��-� �� �1�� �� ��0 �� � ����� ��� �� ����� ��� -����� � ���� �� �� ���� -�� �� *�� , ���� ��� �� � �� �� ����� �� � �� 7'; � ���� ������� ��� �� �� �������� 3��� ��� �� 7#; -�0�� ����� �� �� � �� � � �� ���� ��� � � � ������� ����� �� � �� � �� � ���� ��� �� ����� �� � �� � �� � ���� �� ���� ������� �� � ���� -�� � � �� � � ����0� ���� � ����� ��� ���� � � �� �� ����� ��� � � �� '+9 "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! 0��-� �� �1�� �� ��0�0 ��� �� � ����� �3� �� ������� ���� � �� � �� � � ��� .��0-���� ��3� ����� � � �� �� � �� ���� � -� �� � �� ��� ��-��� �� �� � ���� �� �� � � ��� ����� ��� �� �� �� � � �� �� � �� � � 0 � �� � � ����� + $�����" � ������� � � ���� �� �� �� � �� �� ����� ������� ���� � �� ����� ��� �� �� �� � � ��, ���� �� ������� 0 ������ �� �� 0 �������� ��� ���� �� �� ��� �� ���� ���� ��� ��� ��� ������ �� ��� �� �� ��� �� � �� �� ����� �� �� �� ��� �� �� � � �� �� �� ������ �� ������� �� �� ����� �� .��0-���� -����� � � � - 3� ��� � �� � �� � 1����� � ����� � �� � � �� � �� ��-�� ��� 2��. ��������� --�0� ��� � � �� � �1�� ��� �� � ��������� �� 0��� ����� �� �� � � ������ �� � � �� �3� �� ����� �� ���� � ��� ���� ��� �� � � �� ��� , --�0 �� ��� � � ��� � �� � ���� �� � �� � � ���� �������� ����, �� � � .��� �� ���� �� �� --�0 �� �� ���� 3� ��� � �� �� ������� �� � �� �� ������ �� �� ���� �� �� ��� �� ��� <�����-, � � ���� �� �� �� � �� �� ����� �� �� ������� 0 -�� �� ����� � ���� ���-��� � �� �� �� ��� �� ���� �� ������� � 0�- � � �� � �� 0�� .� �� �� � � �� �� ���� ��'��,�� ������ 4 0��-� .� � � �. k �� � 4� �� ��� �� � -��� � �-d��, �� � �� �� ��� 2� �� �� % �� ���( � 2�� �� �� �� �� ���� 8����( 2���� 2 �. � �� �� �4���( �� ���� � ���� �� � � � �� 8 -�� � �� %�� ���� ��. �� � f� ���� ��� �� � � -� �-� �� � ��� � � � ���1�� � -��������� ������, h� <� 7'+5';� %� -��, ����� �� �3� � -�� ������� 4� #*�%������� ,������,��� 7#�� ���;� h �� �� ��, ��� "�l� ������ .�� ���� , �� �� 7#&&';� � . �� 7���� ��; �� � � �� �� �� �� ��� ��� �� �� �� � � ��� "������ �& 0����%��� ��������, �7'';, '&)� 2�� ��, �� �� 7'++@;� ��� 2��. ������� �� �1�� ��� � ���� �-� �� ��� ����-�� �� ����� � �� ������ 8$# � #��, �����), �, #� ��� 2� ��b� 8� �� � ��� �� �� 7#&&';� ��� ��� � � �� ��� �� 2� �� �� % �� ���, ������, �g� 2�� ��, �� �� 7'++=;� �����%���� ��� 7�.��% )��� ��%������ ���,��������' :���*����� ��� %�� �%��� ������� �%��� �;������� h��. �� � ��� ������ �� �� ��� ���������� �� �� * �� ����� �� �� ������� � ���� ��� 8 ��� , *�, �����, "�, m %0���� , k� 7'++);� � ��0 ������ ��� �� ��� �� � � ��� 6�� 8� "������, 7��;� � ����, �� 7'++&;� *���6 ��� �� �� ��6���������� � �� �� ������� ���� ��� ��� ��� � �� �� � ��� 8$# � #��, �����), 7#;, #@� ��-��, ��, m 8����, "� 7'++9;� 4� �� ��� �� �� � �� �� ������ � �� �� �� �� �� ���� 4� �� ������ ��������� �� ��� � �-� ��� ������ ��i 8$# � #��, �����), !7h � ��;, '� %� �����, 2� 7'++#;� ��� � � �� ���� �� �� ��� �� �� �� ��� 8$# � #��, �����), �, #� : �. , �� 7'++5;� 4� �� ������ �� ��� ���� ��� -n 0��i 8$# � #��, �����), 4, #� "�-� ��.� / 0����%��� �����%�� �����) � 12�� 3 4�5 �! '+= the cfp certification examination process: a discussion of the modified angoff scoring method introduction the cfp certification examination the goal of professional examinations angoff methodology and the cfp exam some examples the beuk adjustment procedure examination equating summary acknowledgements references pii: s1057-0810(01)00069-5 retirement planning guidelines: a delphi study of financial planners and educators sue alexander greningera, vickie l. hamptonb,*, karrol a. kitt a, susan jacquetc adepartment of human ecology, university of texas at austin, austin, tx 78712, usa bfamily financial planning program, box 41162, texas tech university, lubbock, tx 79409, usa ccalifornia social work education center, uc berkeley school of social work, 120 haviland hall, berkeley, ca 94720, usa received 22 december 2000; received in revised form 19 february 2001; accepted 26 february 2001 abstract retirement planning guidelines were determined using a delphi research design among 188 financial planners and educators. consensus was found for using a 4% inflation rate, an 8.5% rate of return on investments, and a replacement ratio of 70–89% of current income when making retirement projections. nine-tenths of the experts agreed that families should have achieved 50–60% of their retirement savings goal by age 50 and 85–90% by age 60. regarding asset allocation, over 60% felt it was prudent to start moving toward more conservative investments about 3–5 years before retirement. recommendations were developed on the proportion of growth-oriented equities to hold at various points prior to and after retiring. while the level of consensus was high, occupational and gender differences were noted. © 2001 elsevier science inc. all rights reserved. jel classification:d12 keywords:retirement planning; retirement needs analysis; retirement guidelines; asset allocation * corresponding author. tel.:11-806-742-3050; fax:11-806-742-1639. e-mail addresses:vhampton@hs.ttu.edu (v.l. hampton), sgreninger@mail.utexas.edu (s.a. greninger), kkitt@mail.utexas.edu (k.a. kitt), sjacquet@uclink.berkeley.edu (s. jacquet). financial services review 9 (2000) 231–245 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00069-5 1. introduction the convergence of several demographic and economic trends have created great interest among the financial community and the general population alike in planning for that period of life called retirement. the vast numbers of aging baby boomers and longer retirement periods due to increased longevity have raised questions about financial preparedness for retirement and the survival of the social security system. in addition, employment issues that affect the value of retirement plans such as corporate downsizing and the growing use of defined contribution plans rather than defined benefit plans (employee benefit research institute [ebri], 2000a) have increasingly shifted responsibility for financial well-being in retirement from employers to individuals and families. the past two decades have produced a tremendous proliferation of financial publications, programs, software/web sites, and advisors eager to provide investors with financial information and services. retirement planning has been a key area of emphasis for the relatively new profession of financial planners. according to the certified financial planner board of standards (1999), 85% of those people who engage the services of a financial planner seek professional assistance because they want help with retirement planning. in addition, many institutions of higher education are now offering degree and certificate programs for students interested in pursuing careers in this growing field. at the end of 2000, 122 colleges and universities offer 169 programs that are registered by the cfp board, allowing students of these programs to take the cfp™ certification exam that includes retirement planning as one of the major topics (cfp board of standards, 2000). starting with the november 2001 exam, a target 18% of this exam will test retirement knowledge. other universities and programs offer studies and certifications for financial planners and human resource professionals that are focused even more specifically on retirement issues. the importance of retirement planning for the well-being of families and individuals as well as for the economy and society, coupled with the growth of the financial advisory and educational establishment, provide fertile research questions. interestingly, numerous studies have looked at whether or not individuals are financially prepared for retirement, but the financial community has just started to seriously question the meaning of the term “retirement.” based on census data, (hobbs & damon, 1996), the average retirement age has dropped from 67 to 62 over the last decade and could go as low as age 60. but there is strong survey and antidotal evidence that retirement does not equal “quit working” for a large portion of the population (national center for women and retirement research, 2000). gustman, mitchell and steinmeier (1995) report that there is no consensus in the literature regarding the definition of retirement. if we do not understand what retirement means to individuals, how can we judge whether the population is financially prepared? this one issue underscores the importance of gathering information, both qualitative information such as goals and risk tolerances and quantitative data, before completing a capital needs analysis to determine what clients need to do in order to successfully meet their retirement goals. while there are various methods of computing retirement needs for families and individuals, these models are based on a common set of assumptions: 232 s.a. greninger et al. / financial services review 9 (2000) 231–245 y retirement age y inflation rate y rate of return on investments before and after retirement y tax bracket before and after retirement y life expectancy y level of annual expenditures required during retirement these assumptions are incorporated into models used by financial planners, computer calculators, and software programs to assess the required savings needed to meet one’s retirement goals. the actual assumptions made regarding inflation and investment rates, retirement age, longevity, and the cost of living during retirement make a very large difference in determining how much must be saved to meet a retirement goal. for example, an individual whose projected retirement needs are $50,000 annually in today’s dollars will need to save about $15,800 annually assuming 1) no current retirement assets, 2) 30 years to save until retirement, 3) 20 years of retirement, 4) 9% return on investments before and after retirement, and 5) 4% inflation before and after retirement. the annual savings requirement drops about 50% to approximately $7,950 by just raising the investment return to 10% and dropping the inflation rate to 3%. based on the importance of these assumptions, this research asks several questions previous studies have not addressed. are there basic retirement planning guidelines that financial planners and educators tend to recommend in their work? how much agreement or consensus exists regarding these guidelines? can a consensus be developed by facilitating communication between planners and educators? this research effort was embarked upon in an effort to find answers to these questions. 2. review of the literature there is much research interest in how well americans are preparing for the future. for example, a retirement confidence survey (rcs) designed to monitor savings, investing, and planning behaviors and attitudes among a representative sample of u.s. workers has been conducted annually during the last ten years. even with the robust economy the past few years, the 2000 rcs reported that retirement confidence in the u.s. had remained virtually unchanged. although 26% of the workers were very confident and 47% were somewhat confident about having adequate resources to live comfortably in retirement, the rcs researchers feared that some of this confidence might be overstated or based on false hopes given some of their other research findings. the majority of respondents able to provide a number reported having less than $50,000 accumulated for retirement with almost one-fourth having less than $10,000, and only 16% reported having $100,000 or more accumulated in total retirement funds. results using the survey’s retirement readiness rating instrument indicated that approximately 30% of the respondents scored in the poor or very poor ranges with regard to preparedness (employee benefit research institute, 2000b). other research has documented similar negative results regarding retirement preparation, especially among members of the baby-boom generation. an econometric model using accumulations and allocations constructed by kotlikoff and auerbach (1994) projected that 233s.a. greninger et al. / financial services review 9 (2000) 231–245 baby boomers were not saving at a rate that would allow them to maintain their preretirement level of consumption after retiring, concluding that the baby boomers’ financial preparation was poor. mitchell and moore (1998) reported that the median married couple at age 55 owned financial assets, excluding their pension plan, of only $73,000 and that this level of savings was insufficient to fund 20 years or more of retirement expenses at preretirement living standards. concern over the lack of participation in 401(k) plans was noted by bassett, fleming, and rodrigues (1998) who reported that over one-third of employees eligible to participate declined to do so. not all research has reached such negative conclusions. studies by easterlin, schaeffer and macunovich (1993) and the congressional budget office (1993) concluded that baby boomers would generally be better off in retirement than their parents, and the american association of retired persons (1994) concluded that baby boomers would be better off in retirement than the elderly had been in 1990. each of these studies cautioned that there would be unevenness in the financial well-being of baby boomers during retirement with certain subgroups being less well off than current retirees or their parents. in studies more related to the assumptions used in the capital needs analysis, researchers including england (1988), palmer (1994), and mcgill, brown, haley, and schieber (1996) have focused on replacement rates or the ratio of income required to provide retirement living expenses divided by preretirement income. rates vary depending on the components and sophistication in measurement but range from 55 to 80%. the replacement rates recommended for lower income individuals and families are higher than those who are more affluent. replacement ratios are often used in a retirement needs analysis in lieu of more accurate expenditure data. in a significant evaluation of retirement needs assumptions, tacchino and saltzman (1999) challenged the presumption that expenditures are constant throughout the retirement years. they present evidence that individuals continue to save during the early years of retirement and that spending patterns voluntarily decline over the retirement period. they suggest that retirement needs analyses should incorporate this information into the models to avoid overstating the amount required to meet one’s retirement goals. work has also been done in the area of retirement age and asset allocation before and during retirement. recent work by montalto, yuh, and hanna (2000) indicates that planned retirement age increases as people get older, reinforcing the need to revisit the retirement needs analysis periodically. others have investigated the relative importance of various factors (e.g., social security, pensions, and health status) that might affect retirement age (samwick, 1998; uccello, 1998). a large body of literature exists regarding the relationship between asset allocation and investment time horizon. some of this work supports the strategy generally offered by financial practitioners that longer investment horizons should be associated with more equities while shorter investment horizons require larger portions of fixed income products (bierman, 1997; bodie, merton, & samuelson, 1992; butler & domian, 1991; thaler & williamson, 1994; thorley, 1995). however, several papers by samuelson (1989, 1990, 1994) and kritzman (1994) indicate that asset allocation for an individual should be independent of time horizon. bierman (1998) and olsen and khaki (1998) present evidence that supports either increasing or decreasing equity positions with changing time horizons, 234 s.a. greninger et al. / financial services review 9 (2000) 231–245 while levy and gunthorpe (1993) and hodges, taylor, and yoder (1997) argue that equities should consume a smaller position in one’s portfolio as time horizons increase. while a body of research does exist regarding replacement ratios, retirement age, and asset allocation that can be applied to the analysis of retirement needs, the literature does not address the other assumptions included in the capital needs model. in addition, the existing research on asset allocation leaves much uncertainty. therefore, this paper will explore the guidelines financial planners use and financial educators recommend in addition to other important retirement planning considerations, goals that may conflict with successful retirement planning, guidelines for meeting retirement needs, and asset allocation guidelines. 3. objectives and methodology the major objectives of this research study were: (1) to ascertain retirement planning considerations and guidelines from a panel of financial experts comprised of both planners and educators (2) to determine if a consensus of opinion existed or could be established regarding these considerations and guidelines (3) to determine what differences in opinions might exist between occupational and gender subgroups in the panel of experts although it is important to acknowledge that guidelines should be reviewed in terms of the contextual situation including such factors as economic conditions, personal goals, and the individual’s or family’s financial status and life-cycle stage, this research sought to determine baseline norms for retirement planning issues through a consensus-developing process. this study was part of a broader project whose goal was to identify and refine a set of benchmarks and ratios for assessing financial well-being. funding for the project was provided by a small grant from the cfp board. the study utilized a delphi research methodology whereby a panel of experts was queried in a sequential set of mailed-out questionnaires designed to facilitate consensus-building. the delphi process, as described by linstone and turoff (1975), is a way of focusing and organizing communication among individuals such “that the process is effective in allowing a group of individuals, as a whole, to deal with a complex problem” (p. 3). a conventional delphi study according to linstone and turoff is described as follows: a small monitor team designs a questionnaire which is sent to a larger respondent group. after the questionnaire is returned, the monitor team summarizes the results and, based upon the results, develops a new questionnaire for the responding group. the respondent group is usually given at least one opportunity to reevaluate its original answers based upon examination of the group response (p. 5). the delphi methodology is attributed to the rand corporation where it was developed in the 1950s to study defense-oriented issues. although this method has not been used in financial planning research, variants of the methodology have since been applied to complex problems in a relatively wide array of disciplines. in the present study, a three-person research team of personal finance educators, who are 235s.a. greninger et al. / financial services review 9 (2000) 231–245 also all cfp™ certificants, identified financial guidelines and recommendations from existing educational and research literature. their findings were then reviewed by an advisory committee comprised of five cfp™ practitioners and one additional university educator with an interest in financial planning. the advisory committee also reviewed and had input regarding the sampling technique and research design proposed for the study. invitations were extended to 400 financial planners randomly selected from a national listing of cfp™ professionals and 340 financial educators selected from the membership of the association for financial counseling and planning educators (afcpe) in 1994 asking them to participate in the project. both practitioners and educators were included in the sample because of the importance of industry involvement in professional education. since educators train students to work in industry, it is important that educators and practitioners agree upon the basic concepts that define a profession. while there could be overlap between the cfp™ certificants who are largely practitioners and afcpe members who are primarily university educators, it was quite small. an initial acceptance rate of 38% was achieved with 122 financial planners and 159 educators agreeing to participate. in keeping with the delphi protocol, four separate mailed questionnaires distributed over several months were sent to these participants. round one featured an open-ended questionnaire that asked for input from the expert panel regarding the following 20 financial concepts identified in the literature: y diversification between and within investment types y asset allocation appropriate for life cycle stage and goals y extent and adequacy of regular savings/investing program y cash reserves and liquidity y exposure to insolvency y debt safety level y housing expense relative to income level y tax burden y inflation protection y frequency of financial review y specificity of financial goals y progress toward goal attainment y adequacy of life insurance coverage y adequacy of disability insurance coverage y adequacy of property insurance coverage y protection from liability exposure y adequacy of medical insurance coverage y adequacy of long-term care coverage y adequacy of retirement planning given life cycle stage and goals y adequacy of estate planning participants were asked to list factors they felt were important in assessing the financial well-being of individuals and families for each of these areas. based on the open-ended responses from round one, the research team developed two separate closed-end questionnaires that comprised round two. the decision to split round 236 s.a. greninger et al. / financial services review 9 (2000) 231–245 two into two halves was made in order to keep the questionnaire length short enough to encourage participation. results from round two, where there was not a general consensus, were then summarized by the research team and reported back to the participants on the round three questionnaire. in accordance with the recommended protocol for the delphi method, this allowed the participants an opportunity to reevaluate their answers based on the group results. in round three, the participants were asked to rate their level of agreement with summary statements that had emerged from round two results on a 5-point likert scale where 15 definitely do not agree, 25 do not agree, 35 uncertain, 45 agree, and 55 strongly agree. these scaled responses allowed differences to be tested between the occupational and gender subgroups. in the analyses of the data, we used exploratory factor analysis and anovas for continuous variables and frequencies and chi-square to handle categorical variables. comparisons of means and frequencies were based for the most part on gender and occupations. separate exploratory factor analyses were performed on the responses to two sets of items. the intention was to determine whether there were global concepts (tabachnick & fidell, 1996) that might be embedded within each set of items that might be identified for future research endeavors. in round two, participants rated the “importance of factors to consider when planning for retirement,” which contained 13 items and the “likelihood of goals being perceived as more important than (therefore, in conflict with) retirement planning for the typical family,” which contained six items. in both analyses, we used a principal components extraction method with varimax rotation. these analyses were performed on ratings made by the participants on 5-point likert scales. when frequency information is reported from the ratings on the likert scales, percentages may be reported in a truncated manner in terms of the percentages of participants who agreed, disagreed, and were uncertain. however, the means and medians reported in this paper from the likert scales reflect all five response categories. findings from 188 participants who responded to retirement questions in round two and/or round three of the study are reported in this paper. of the 188 respondents, 113 (60%) were educators and 75 (40%) were planners. there were more males than females, 55% versus 45%, respectively. a significant relationship between gender and occupation existed in the sample with planners predominantly being male (77%) and educators predominantly being female (59%) [x2 (1, n 5 188) 5 25.71,p , .000]. there was also a significant difference in the educational level of the two occupational subgroups. as might be expected, the planners were more likely to have received bachelor degrees whereas the educators were more likely to have received advanced degrees, particularly at the doctoral level [x2 (5, n 5 188) 5 49.74,p , .000]. 4. research findings 4.1. retirement planning considerations in round two, the panel of experts was asked to rate the importance of a number of factors to consider when planning for retirement on a 5-point likert scale where 15 very 237s.a. greninger et al. / financial services review 9 (2000) 231–245 unimportant, 25 unimportant, 35 uncertain, 45 important, and 55 very important. the list of considerations was developed from responses given to an open-ended question asked in round one. the mean, median and modal ratings for these considerations are presented in table 1. in addition to rating each, the participants were asked to identify the three considerations that they believed were most important overall. the frequency with which each consideration was listed in the top three is also included in table 1. the factors that were considered as most important to the panel of experts were available income sources, availability of a vested pension plan, availability of tax-deferred plans, and availability of employer matching for building retirement funds. an exploratory factor analysis was performed on the ratings of the above considerations when planning for retirement to determine whether there were any global concepts among the 13 items. in the initial analysis, three factors were observed with eigen values that exceeded 1.0. factor loadings are presented in table 2. factor 1, which was named “availability of income sources,” had an eigen value of 3.9 and explained 30% of the variance. this factor included the following retirement considerations: availability of profit sharing plan, vested pension plan, tax-deferred plans, employer matching, and income sources. the mean rating for factor 1 was 4.5 on the 5-point likert scale for importance. the mean ratings on this factor differed significantly (p , .01) by gender with higher importance assigned to this component by women than men. the second factor, which was named “needs and changes,” had an eigen value of 1.4 and explained 11% of the variance. factor 2 was comprised of career change needs, past job stability/changes, and conflicting goals. the mean rating of the overall panel was 3.8 on this factor with women again assigning significantly (p , .05) higher importance than men to these retirement considerations. the third factor was a weak one, having an eigen value of 0.7 and explaining only 5% of the table 1 important considerations when planning for retirement consideration top three frequency mean rating* median rating modal rating available income sources 103 4.7 5 5 availability of a vested pension plan 88 4.6 5 5 availability of tax-deferred plans 64 4.4 4 4 availability of employer matching for building retirement funds 46 4.4 5 5 availability of profit-sharing plan 19 4.3 4 4 need to meet other conflicting goals (especially children’s education) 41 4.1 4 4 eligibility for social security 31 4.0 4 4 past job stability and changes 17 3.8 4 4 potential to continue working on a part-time basis 12 3.6 4 4 need to change career 3 3.4 3 3 potential for a life insurance buy-out 0 3.0 3 3 potential for receiving an inheritance 4 3.0 4 3 potential for use of reverse mortgage 1 2.8 3 3 * rating scale: 15 very important, 25 unimportant, 35 uncertain, 45 important, 55 very important. 238 s.a. greninger et al. / financial services review 9 (2000) 231–245 variance. this factor was named “potential resources” and included the following retirement considerations: reverse mortgages, insurance buyouts, part-time employment, social security eligibility, and inheritances. the mean rating on factor 3 on the overall panel was 3.3. in this case, the mean rating was significantly (p , .05) higher for educators than planners. 4.2. conflicting goals in round two, respondents were asked to rate other goals on the likelihood that they might be perceived as more important than retirement planning for the typical family. again a 5-point scale was used with 15 very unlikely, 25 unlikely, 3 5 uncertain, 45 likely, and 55 very likely. the mean ratings of the panel for these goals are presented in table 3. children’s education received the highest mean rating, indicating it was perceived as the most likely goal to conflict with retirement planning. the other goal with a relatively high table 2 component loadings of retirement planning considerations considerations factor 1 availability of income sources factor 2 needs & changes factor 3 potential resources availability of: profit-sharing plan .80 .10 .01 vested pension plan .76 .15 .10 tax-deferred plan .73 .13 .11 employer matching funds .69 .34 .10 income sources .59 .15 .13 need to change careers .11 .88 .29 past job stability/changes .24 .49 .15 need to meet other conflicting goals .32 .46 .14 potential for: reverse mortgage usage .01 2.04 .62 life-insurance buyout 2.05 .23 .62 part-time employment .19 .28 .53 social security eligibility .29 .13 .40 inheritance availability .10 .15 .38 eigen value 3.9 1.4 0.7 % of variance 30.0 11.1 5.0 table 3 other goals that potentially conflict with retirement planning goal mean rating* median mode college education for children 4.5 5 5 maintenance of current standard of living 4.3 4 4 better house or second home 3.7 4 4 vacation and travel 3.6 4 4 resources to change career 3.3 3 3 estate preservation 3.0 3 3 * rating scale: 15 very unlikely, 25 unlikely, 3 5 uncertain, 45 likely, 5 5 very likely. 239s.a. greninger et al. / financial services review 9 (2000) 231–245 probability of negatively impacting retirement planning was maintenance of current living standards. when the ratings for these goals where factor analyzed, two factors emerged. these two factors were named “maintaining the good life” and “future security.” the first one had an eigen value of 1.5 and explained 26% of the variance. this “good life” factor included the following goals: vacations and travel, better house or second home, and maintenance of current living standard. the mean rating on this factor for the overall panel of experts was 3.9. the “future security” factor had an eigen value of only 0.8 and explained 13% of the variance. the goals included in this factor included the need to change careers, preserve one’s estate, and pay for children’s college education. this factor’s mean was 3.6 for the entire panel, with educators rating this factor significantly (p , .05) higher than planners. 4.3. retirement needs assumptions two key planning assumptions used in retirement needs analyses are the expected rates of inflation and investment return. when asked about the appropriate rate of inflation to use for retirement planning, round two respondents reported a range of 2–10% with the mean being 4.3%. the most common response was 4% inflation, indicated by 35% of the respondents. slightly less than one-fourth of the experts said 5%, while 14% said 3% inflation. when asked how they decided on the rate to use, almost seven-tenths of the participants said they used an average rate over time as their assumed inflation rate for retirement planning. one-fourth indicated they used their client’s preference, and 6% said they used the current rate of inflation. in round three, the participants were asked to rate their agreement on a 5-point likert scale with the following two statements: (1) the rate of inflation used for retirement planning should be based on an average rate of inflation over time. (2) four percentage is currently an appropriate rate of inflation to use in retirement planning projections. the mean ratings for these two statements were 4.3 and 3.8, respectively. approximately nine-tenths of the experts agreed with the first statement, while 4% disagreed, and 6% were uncertain. although the mean rating was lower for the second statement, almost three-fourths of the participants agreed at some level with this statement; 12% disagreed and 15% were uncertain. based on long-term inflation data from ibbotson associates (2000), participants were consistent in their responses. over the period from 1926 to 1999, inflation has averaged 3% annually; however, the most recent 20-year period has averaged 4% annually. looking at 20-year rolling periods starting with 1960, there are several that produce average annual rates of inflation of 5–6%. regarding expected investment returns, the recommended rate for retirement planning reported in round two ranged from 2 to 12% with both the median and modal responses being 8%. when asked how the investment rate of return was determined, respondents said that this rate was based on factors such as the historical rate of return; the type of assets in a portfolio; client preferences, risk tolerances, and timeframe; inflation; and taxes. in round 240 s.a. greninger et al. / financial services review 9 (2000) 231–245 three, the panel of experts overwhelmingly agreed (94%) that inflation and taxes should be considered when projecting investment return for retirement planning purposes, resulting in a mean rating of 4.5 on the 5-point scale. there was a similar percentage (95%) who agreed that the investment rate of return used in retirement planning should be based on client preferences, risk tolerance, and time frame as well as on the historical rate of return on assets in the client’s portfolio; the mean rating on this statement was also 4.5. only 2% of the experts disagreed with these two planning assumptions. in round three, the mean annual rate of investment return deemed appropriate for retirement planning purposes before inflation and taxes are taken into account was 8.5% for overall investments. this was a little higher than the comparable mean in round two. for specific asset classes, the means in round three were 4.3% for cash equivalents, 10.4% for stocks, 7.1% for bonds, and 6.9% for real estate. the rate of return thought to be appropriate for cash equivalents differed significantly (p , .05) between the two occupational groups. in this case, the difference between planners and educators was qualified by interaction between gender and occupation with female educators reporting higher return rates than either male or female planners. ibbotson associates (2000) data provide benchmark returns for three of the above asset classes. between 1926 and 1999 large company stocks, long-term corporate bonds, and u.s. treasury bills (cash) averaged annual returns of 13%, 6%, and 4%, respectively. over the most recent 20 years, annual returns for all three asset classes have been higher than over the longer term at 18%, 11%, and 7%, respectively. 4.4. guidelines for meeting retirement income needs in round two, over one-half (56%) of the experts reported that they recommended a percentage of current expenses to estimate postretirement needs. the mean reported for this benchmark was 74% with the median being 75%. because there were a number of low percentages reported, the research team decided to reask this question in round three using a closed-end format. the majority of the experts (81%) felt that amounts ranging from 70 to 89% of current expenses were useful in estimating postretirement needs. a few more of the experts selected 70–79% rather than 80–89% of expenses—42% and 39% of the experts, respectively. according to round two participants, the mean percentages of retirement needs the typical individual/family should have achieved by various ages were as follows: at age 30, 18%; at age 40, 37%; at age 50, 59%; at age 60, 85%; at age 65, 96%; and at age 70, 99%. all of these means were significantly (p , .001) different from each other. the percentage of total needs recommended for retirement by age 70 differed significantly (p , .05) by gender with the men’s responses averaging 100% and the women’s averaging 96%, respectively. based on the round two responses, four benchmarks regarding the percentage of overall retirement savings that should be achieved by certain preretirement age levels were developed. these were included in the round three questionnaire for the experts to rate in terms of their agreement. the mean ratings and distribution of agreement, disagreement, and uncertainty regarding these recommended benchmarks are presented in table 4. the major241s.a. greninger et al. / financial services review 9 (2000) 231–245 ity of the experts agreed with each of these benchmarks. as might be expected, there was less consensus among the experts regarding the percentage of retirement savings that should be achieved at younger ages than at older ages. more than nine-tenths of the experts agreed with the benchmarks for 50 and 60 year old individuals while approximately two-thirds agreed with the benchmark for 30-year old individuals. 4.5. timing and asset allocations guidelines in round two, respondents were asked when they thought it was prudent for the average preretiree to begin moving their investments to more conservative investments as they approached their planned retirement target date. the frequency of the panel’s responses were as follows: never, 21%; one year before retirement, 5%; three years before retirement, 27%; five years before retirement, 34%; seven years before retirement, 7%; and ten years before retirement, 6%. this indicated that over six-tenths of the participants thought it was prudent to move toward more conservative investments about three-to-five years before retirement. however, over one-fifth thought it was never prudent to do this. there was a significant (p , .001) difference between the occupational groups on this matter with educators recommending an earlier move to conservative investments than planners. in round two, the participants were also asked what percentage of assets they thought should be placed in growth-oriented equities based on an individual’s proximity to retirement acknowledging that norms would need to be adjusted based on factors such as a person’s risk tolerance and income needs. the mean percentages of assets in growth-oriented equities recommended by the panel based on proximity to retirement were as follows: 15 years before retirement, 69%; 10 years before, 62%; 5 years before, 51%; at retirement, 40%; 5 years after retirement, 33%; 10 years after, 28%; and 15 years after, 25%. significant differences based on occupation existed for the recommended levels both for 10 years and 15 years after retirement. in both cases, planners recommended higher mean percentages in growthoriented equities than the educators. when the panel of experts was asked in round three to indicate their level of agreement with whether the declining pattern of growth-oriented investments specified in the previous table 4 level of agreement with recommended retirement savings goals by age recommended guideline mean rating* % of experts responding agree uncertain disagree by age 60, individuals/families should have achieved 85–90% of their retirement savings goal 4.4 94 4 2 by age 50, individuals/families should have achieved 50–60% of their retirement savings goal 4.2 90 7 3 by age 40, individuals/families should have achieved 30–40% of their retirement savings goal 3.9 78 18 4 by age 30, individuals/families should have achieved 10–20% of their retirement savings goal 3.7 68 15 17 * rating scale: 15 definitely do not agree, 25 do not agree, 35 uncertain, 45 agree, 55 strongly agree. 242 s.a. greninger et al. / financial services review 9 (2000) 231–245 paragraph was appropriate for preretirees and retirees, the mean response was 3.4 on the 5-point scale. there was a significant (p , .05) difference between the means of the occupational groups on this issue with planners expressing less agreement than educators. when asked what percentage of investments should be kept in growth-oriented investments during retirement, the mean percentage recommended by the panel of experts was 31% with the median being 25%. there was again a significant (p , .01) difference in the means based on occupation with the planners recommending 36% and the educators recommending 26%. the panel did not generally agree with the advice that an investor should maintain stock holdings in his/her portfolio equal to “100 minus his/her age.” the mean rating on this statement was 2.7 with only 29% agreeing with this statement. almost one-half (45%) disagreed with this often-quoted rule of thumb and more than one-fourth (26%) were uncertain. planners rated this piece of advice significantly (p , .05) lower on the 5-point scale than did the educators. 5. summary and conclusions this study determined that there was a relatively high level of consensus among the financial experts regarding general retirement planning guidelines. there was more agreement on the guidelines for planning assumptions and meeting retirement needs than about timing and asset allocation. a strong consensus was found for using a 4% inflation rate, a 8.5% rate of return on investments, and a replacement ratio of 70–89% of current income when making retirement projections. over nine-tenths of the experts agreed that individuals and families should have achieved 50–60% of their retirement savings goal by age 50 and 85–90% by age 60. although the majority (60%) of the experts agreed that it was prudent to start moving portfolio holdings toward more conservative investments about 3–5 years before retirement, the planners recommended that the repositioning occur in closer proximity to retirement than the educators. planners also recommended that a significantly higher proportion of growth-oriented assets be held during retirement than educators. there were few differences in the study’s findings based on the gender of the experts. in conclusion, this project has contributed new knowledge and information that should be incorporated into the curriculum and educational materials designed for retirement education programs and personal finance classes. as with much research, the results of this project raised as many questions for future research as it provided answers. on the guidelines where there was a high level of agreement, it would be useful to know how the advice stacks up against the reality of what families and individuals are actually doing. for example, it would be interesting to determine how many 50-years olds have actually achieved 50–60% of their retirement savings goal or how many retirees currently hold growth-oriented equities in the recommended proportions suggested by the panel. furthermore, it would be of value to find out what guidelines people tend to have in their heads regarding retirement planning and how well these match up with those of the experts. 243s.a. greninger et al. / financial services review 9 (2000) 231–245 acknowledgments preparation of this article was supported in part by a grant from the certified financial planner board of standards (cfp board). references american association of retired persons. (1994).aging baby boomers: how secure is their economic future? washington, dc. bassett, w. f., fleming, m. j., & rodrigues, a. p. (1998). how workers use 401(k) plans: the participation, contribution, and withdrawal decisions.national tax journal, li, 263–289. bierman h., jr. (1997). portfolio allocation and the investment horizon.journal of portfolio management, 23, 51–55. bierman jr., h. (1998). a utility approach to the portfolio allocation decision and the investment horizon.journal of portfolio management, 25,81–87. bodie, z., merton, r. c., & samuelson, w. f. (1992). labor supply flexibility and portfolio choice in a life cycle model.journal of economic dynamics and control, 16,427–449. butler, k. c., & domian, d. l. (1991). risk, diversification and the investment horizon.journal of portfolio management, 17,41–47. certified financial planning board of standards. (1999).1999 cfp board consumer survey.denver: author. retrieved december 21, 2000 from the world wide web: http://www.cfp-board.org/press_1999consumer. html certified financial planning board of standards. (2000). twenty-two educational institutions register programs with the cfp board in 2000.cfp board report, 7(6), 4. congressional budget office. (1993). baby boomers in retirement: an early perspective.washington, dc. easterlin, r. a., schaeffer, c. m., & macunovich, d. j. (1993). will the baby boomers be less well off than their parents? income, wealth, and family circumstances over the life cycle in the united states.population and development review, 19, 497–522. employee benefit research institute. (2000a).personal account retirement plans: an analysis of the survey of consumer finances(ebri issues brief no. 223). washington, dc: author. retrieved december 20, 2000 from the world wide web: http://www.ebri.org/ret_findings.htm employee benefit research institute. (2000b).2000 retirement confidence survey (rcs) summary of findings. washington, dc: author. england, c. b. (1988). how much retirement income do employees need?benefits quarterly, iv(1), 1–5. gustman, a. l., mitchell, o. s., & steinmeier, t. l. (1995). retirement measures in the health and retirement study.journal of human resources, 30(1995 supplement), s57–s83. hobbs, f. b., & damon, b. l. (1996). 651 in the united states (current population reports p23–190). washington, dc: us department of commerce bureau of the census. hodges, c. w., taylor, w. r. l., & yoder, j. a. (1997). stocks, bonds, the sharpe ratio, and the investment horizon.financial analysts journal, 53,74–80. ibbotson associates. (2000). basic series: summary statistics of annual total returns.stocks, bonds, bill, and inflation 2000 yearbook. chicago: author. kotlikoff, l. j., & auerbach, a. j. (1994). u.s. fiscal and savings crises and their impact for baby boomers. retirement in the 21st century. . . ready or not. . . . employee benefit research institute, 85–119. kritzman, m. (1994). what practitioners need to know about time diversification.financial analysts journal, 50, 14–18. levy, h., & gunthorpe, d. (1993). optimal investment proportions in senior securities and equities under alternative holding periods.journal of portfolio management, 19,30–36. linstone, h. a., & turoff, m. (eds.) (1975).the delphi method: techniques and applications.boston: addison-wesley publishing co. 244 s.a. greninger et al. / financial services review 9 (2000) 231–245 mcgill, d. m., brown, k. n., haley, j. j., & schieber, s. j. (1996).fundamentals of private pensions(7th ed.). philadelphia: university of pennsylvania press. mitchell, o. s., & moore, j. f. (1998). can americans afford to retire? new evidence on retirement savings adequacy.the journal of risk and insurance, 65(3), 371–400. montalto, c. p., yuh, y., & hanna, s. (2000). determinants of planned retirement age.financial services review, 9(1), 1–15. national center for women and retirement research. (2000).re-casting retirement planning for the 21st century.southampton, ny: author. olsen, r. a., & khaki, m. (1998). risk, rationality, and time diversification.financial analysts journal, 54, 58–63. palmer, b. a. (1994). retirement income ratios: an update.benefits quarterly, 10(2). samuelson, p. a. (1989). the judgment of economic science on rational portfolio management: indexing, timing, and long-horizon effects.journal of portfolio management, 16,4–12. samuelson, p. a. (1990). asset allocation could be dangerous to your health.journal of portfolio management, 16, 5–8. samuelson, p. a. (1994). the long-term case for equities.journal of portfolio management, 21,15–24. samwick, a. a. (1998). new evidence on pensions, social security, and the timing of retirement.journal of public economics, 70(2), 207–236. schister, s. j. (1996). conceptual and measurement problems in contemporary measures of income needs in retirement.benefits quarterly,(second quarter). tabachnick, b. g., & fidell, l. s. (1996).using multivariate statistics(3rd ed.). new york: harper collins. tacchino, k., & saltzman, c. (1999). do accumulation models overstate what’s needed to retire?journal of financial planning, 12(2), 62–73. thaler, r. h., & williamson, j. p. (1994). college and university endowment funds: why not 100% equities? journal of portfolio management, 21,27–37. thorley, s. r. (1995). the time diversification controversy.financial analysts journal, 51,68–76. uccello, c. e. (1998). factors influencing retirement: their implications for raising retirement age (aarp public policy institute, report no. 9810). washington, dc: american association of retired persons. 245s.a. greninger et al. / financial services review 9 (2000) 231–245 121 the impact of the online marketplace on fraud: evidence from craigslist from its early adoption in 1995 to its wider expansion in 2006 efthymia antonoudi,1 martin seay,2 han na lim,3 and elizabeth kiss4 abstract this research addresses the influence of craigslist’s adoption and presence on fraud arrests within metropolitan statistical areas (msas). utilizing the consumer vulnerability framework (hill & sharma, 2020), the study used diverse data sources, including craigslist entry data, the uniform crime reporting (ucr) dataset, and the us census bureau current population survey (cps) data from 1995-2006. employing differences-in-differences (did) models, this study's primary findings indicate a reduction in fraud arrests, ranging from 11% to 23% following the introduction of craigslist. while these results might appear counterintuitive, our findings suggest that online marketplace design and enforcement capacity may jointly influence fraud patterns. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation antonoudi, e., seay, m., lim, h., & kiss, e. (2025). the impact of the online marketplace on fraud: evidence from craigslist from its early adoption in 1995 to its wider expansion in 2006, financial services review, 33(4), 121-133. background the internet has fundamentally altered the way the marketplace operates. using the internet and advanced technologies as an online marketplace brought countless opportunities for consumers to increase their utility by decreasing the costs of searching for information (kroft & pope, 2014), increasing efficiency by matching consumers to suppliers of goods and services, and reducing the transaction costs associated with buying, selling and giving away used goods (fremstad, 2017). 1 corresponding author (eanton@uga.edu). university of georgia, athens, ga, usa. 2 kansas state university, manhattan, ks, usa. 3 california state university, fullerton, ca, usa. 4 kansas state university, manhattan, ks, usa. there are conflicting hypotheses with respect to whether the introduction of craigslist to the local consumer marketplaces has increased or decreased fraud. in this paper, we provide empirical evidence to test these two hypotheses. craigslist, launched in the san francisco bay area, is a classified advertising website operating since 1995 that enables multiple unrelated matching markets to interact on a single consolidated platform (cunningham et al., 2017). users can post advertisements about jobs, https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review 33(4) 122 housing, services, personal items, and for-sale items. in 2000, craigslist expanded to include other major cities such as boston, new york, and los angeles, and in 2001, it entered the denver, atlanta, and austin markets. gradually, craigslist started operating in other smaller metropolitan statistical areas (msas) and, by 2010, covered most us cities and expanded to include multiple markets outside of the us (cunningham et al., 2017). in 2005, craigslist experienced remarkable growth and received more than 8 billion page views per month by the end of 2006 (kroft & pope, 2014). because of the enormous growth and the widespread usage of the website after 2006, we focused our analysis on the years from craigslist's inception in 1995 to its widespread expansion in 2006. craigslist altered how consumers buy, sell, and dispose of secondhand goods (fremstad, 2017). it created a new marketplace promoting online search and lowering information acquisition costs. internet outlets increase market efficiencies by reducing search costs and lowering price dispersion (brynjolfsson & smith, 2000). craigslist connects consumers in mutually beneficial transactions but the opportunity for fraudulent activity may also increase. for instance, in 2014, a man in des plaines, illinois, was charged with multiple felonies after posting fraudulent apartment listings on craigslist and collecting deposits from unsuspecting renters. victims uncovered the scam when they attempted to move in and found the apartments already occupied or that the keys did not work (knowles, 2014). while such cases illustrate that fraud can and does occur on digital platforms, the structured environment of craigslist may also facilitate detection and prosecution in ways that were not possible with offline or traditional scams. this highlights how the introduction of craigslist may not reduce all fraud but may influence how it is reported, tracked, and resolved. building on this idea, we investigate whether the presence and adoption of craigslist as an online marketplace affected fraud-related arrests in locations that adopted the platform compared to those that did not. this study contributes to the financial services literature by examining how online marketplaces influence consumer risk exposure, fraud-related enforcement, and financial behavior. as more financial transactions occur digitally, understanding the role of marketplace infrastructure—such as craigslist—can inform fraud prevention strategies, regulatory frameworks, and consumer protection practices. review of the relevant literature fraud millions of individuals fall victim to fraud around the world every year. according to the recent united states federal trade commission (ftc) data book, individuals reported losing $8.8 billion to scams in 2022, an increase of $2.6 billion yearly (rayo, 2023). while the dollar amount continues to increase, the number of individuals affected has remained staggering; about 35.6 million were fraud victims in 2004 (anderson, 2004). in 2006, the ftc received over 670,000 consumer sentinel complaints amounting to over $1.1 billion in fraud losses, of which 36% were identity theft complaints, and 64% were other types of fraud (federal trade commission, 2006). fraud can also have enormous nonmonetary costs, such as emotional stress and psychological trauma. these nonmonetary costs are hard to quantify and likely greater than the financial losses (lee & soberon‐ferrer, 1997). financial crimes are perpetrated in various ways, with the most common types of fraud related to false representation and identity theft. the top consumer fraud issues identified by the 2004 ftc survey are advance-fee loan scams, being billed for membership without having agreed to it, credit card insurance and credit repair services, paying money for a purchase without receiving the promised prize, being billed for internet services without agreeing, and purchasing a membership in a pyramid scheme (anderson, 2004). not much has changed 20 years later. in the united states, investment scams were the costliest form of fraud in 2022, with reported losses to american consumers of $3.8 billion, followed closely by impersonator scams with reported losses of $2.6 billion (rayo, 2023). antonoudi et al. 123 technology and fraud current scholarly work on online fraud has examined a variety of types and methods of these crimes. online consumer fraud victimization has concentrated on the sophisticated nature of fraudsters (garg & nilizadeh, 2013; lee, 2021a, 2021b; van wilsem, 2013). park et al. (2014) analyzed the prevalence of advance-fee fraud paying a fee upfront in anticipation of receiving something of greater valueon craigslist and showed that ten groups of scammers were responsible for nearly half of the total scam attempts. other scams thriving on craigslist that have been the subject of scholarly work were automobile-related scams targeting mostly educated white males (garg & nilizadeh, 2013). theory and conceptual framework consumer vulnerability occurs when individuals face a heightened risk of harm due to limited access to resources or reduced control over their use in the marketplace (hill & sharma, 2020). in online marketplaces like craigslist, such vulnerability can be shaped by structural features—such as platform availability—that may alter exposure to fraud. in our framework, the key independent variable is the presence or absence of craigslist in a given metropolitan statistical area (msa). this structural condition creates different contexts in which consumers operate, potentially influencing fraud risk. while individual and demographic factors such as income, employment, education, and race also play a role, our analysis focuses on how platform availability interacts with these factors to affect fraud-related outcomes. according to the consumer vulnerability framework, one might expect craigslist’s entry into a market to heighten exposure to fraud by expanding opportunities for deceptive interactions and reducing face-to-face accountability. this would suggest a positive relationship between platform adoption and fraud arrests. however, several mechanisms may produce the opposite effect. craigslist’s peer-topeer design may increase transparency, foster community moderation through user flagging of suspicious listings, and generate digital trails that deter criminal behavior. additionally, if craigslist displaced aggressive or misleading advertising from traditional media, it may have reduced consumer exposure to certain fraud schemes. these competing mechanisms make the relationship between craigslist and fraud an empirical question. research question did the introduction and presence of the craigslist online marketplace impact fraud arrest occurrences? hypothesis given the competing theoretical arguments, we adopt a two-tailed hypothesis structure: h₀: craigslist’s entry has no effect on fraud arrest rates. h₁: craigslist’s entry is associated with changes in fraud arrest rates. methods data and sample we use a combination of three different data sources to determine craigslist entry, fraud arrests, and county-level characteristics. upon request, the authors of kroft & pope (2014) provided data associated with craigslist entry for every msa from 1995 to 2006. craigslist was first introduced in san francisco (1995), then in other large msas like new york and boston (2000), followed by mid-sized cities in the early 2000s. kroft and pope (2014) noted that early adopters tended to be cities with higher incomes, higher education levels, and larger populations. to conserve space, the timeline and average demographic characteristics of early versus late adopters are available upon request. we used the us census bureau county-level data (u.s. census bureau, 1995-2006) to extract the study's control variables, including median income, unemployment percentage, poverty rate, education, and racial background of each county. msas are defined in terms of entire counties, consisting of at least one urbanized area with a population of 50,000 or more, along with adjacent counties with a high degree of economic and social integration with the urbanized core. we used historical delineation files from the us census bureau webpage to aggregate cps data at financial services review 33(4) 124 the msa level (u.s. census bureau, 2023). we converted the data to the most recent definitions. police agencies report county-level fraud arrest data to the federal bureau of investigation’s uniform crime reporting (ucr) program each year, providing base data for this analysis. we used information from the quarterly census of employment and wages (qcew) to aggregate county-level data at the msa level to match the craigslist entry data. for the study's primary analysis, the period from 1995 to 2006 was investigated using data from 1991 to test the parallel pre-trends assumption. unfortunately, the ucr data aggregates fraud types and does not allow us to isolate fraud categories most directly associated with craigslist transactions. during this analysis period, the sample, on average, included 384 msas. we analyzed msas over the 15 years between 1991 and 2006, except 1993. that year was excluded because the fbi crime data were not available. the available data provided an initial sample size of 5,760 msas, reduced by 61 msas that did not report crime data to a sample of 5,699 msas. as anderson (2014) noted, there are problems with the fraud reports. unusually large fraud reports have arrest counts that are a compilation of police agency reports, and not all agencies submit reports consistently every year (anderson, 2014). following anderson (2014) and fone et al. (2023), we dropped from the primary analysis msas that reported unusually high fraud arrests defined as more than two standard deviations from the mean.5 as such, we removed 377 msas, providing a final analytic sample of 5,322 observations. of this final sample of 5,322 observations, 399 msas reported zero fraud arrests during the analysis period in a given year, with the remaining msas reporting arrests above zero. to further safeguard against this issue, we control for the number of agencies reporting within an msa for any given year. to do that, we added a specification that controls for the number 5 msas with fraud arrest rates more than two standard deviations above the mean were excluded. these msas tended to have smaller populations and more volatile reporting patterns. 6 one possible explanation is that craigslist’s introduction shifted local fraud patterns by of agencies that report arrest data (anderson, 2014) and includes only those msas with 90% coverage or above that provided a sample of 3,624 observations. coverage is the percentage of police and other enforcement agencies that reported arrests. variables dependent variable the study’s dependent variable is fraud arrests per 10,000 people in each msa. ucr data defines fraud as the intentional perversion of the truth to induce another person or other entity in reliance upon it to part with something of value or to surrender a legal right (federal bureau of investigation, 2004). this definition includes fraudulent conversion and obtaining money or property by false pretenses. confidence games and bad checks, leaving a full-service gas station without paying, credit card/automatic teller machine fraud, impersonation, welfare fraud, and wire fraud were all included in this definition of fraud (federal bureau of investigation, 2004). forgery and counterfeiting are excluded from this variable. while this definition includes many types of consumer fraud, such as wire fraud, impersonation and false pretenses, we note that some broader categories are also included. the ucr data does not allow us to isolate only consumer fraud, which we acknowledge as a limitation in interpreting results.6 independent variable the primary predictor variable is a binary variable of whether or not craigslist was available in the metropolitan statistical area (msa) in a given period. the craigslist variable is constructed as a binary variable coded as 0 if craigslist was not available and coded as 1 if craigslist was available in that msa for each year during the analysis period. control variables influencing enforcement focus or offender tactics. for example, fraudsters may have moved away from traditional schemes (e.g., bad checks or benefit fraud) toward online impersonation or listing scams, which could still affect the overall fraud arrest rate. antonoudi et al. 125 after aggregating available individual and household-level data, we measured control variables at the msa level. control variables included household median income, poverty rate, and racial composition at the respondent level. in addition, we aggregated the msa-level employment status and education status collected at the respondents' level; a summary of all variables, their definitions, and data sources is provided in appendix table a. statistical analyses a difference-in-differences (did) approach compares the change in fraud arrests over time between msas that adopted craigslist and those that did not. this statistical method allows us to isolate the effect of craigslist's introduction by controlling for broader time trends and regional fixed characteristics. equation the empirical model is specified as equation (1) below. ym,t = β0 + β1craigslistm * time t + β2xt + ym + nt + ε,m,t where ym,t represents fraud/10,000 people and the subscript m is the msa with t representing the year. the parameter β0 is the intercept and β1 is the coefficient for the craigslist dummy variable. the coefficient y is the msa fixed effects and n is the period fixed effects. finally, ε,m,t is the error term. the key terms in our model are explained in table 1 below. table 1. explanation of terms used in the difference-in-differences (did) model term or symbol description how it is measured or applied y (m,t) dependent variable: fraud arrests per 10,000 people in msa m at time t calculated from ucr data, adjusted by msa population β₀ intercept constant term in the model craigslist (m) × time (t) interaction term: treatment effect (did estimator) binary = 1 if craigslist is present in msa m during year t; 0 otherwise β₁ coefficient of interest (treatment effect) measures craigslist's effect on fraud arrests x (t) control variables income, unemployment, education, race; aggregated to msa level β₂ coefficients on control variables estimated effects of each covariate msa fixed effects (yₘ) controls for time-invariant msa characteristics e.g., local policy or geography that does not vary over time year fixed effects (nₜ) controls for time-specific shocks national events or economic conditions in year t error term (ε m,t) random error unobserved factors not captured in the model note. the did estimate reflects: (post-treatment − pre-treatment in treated msas) − (post-treatment − pre-treatment in control msas). this structure helps isolate the effect of craigslist from confounding time and regional factors. we used differences-in-differences models (did) to examine associations between the ucr and census-based variables over time and to study the differential impact of craigslist’s presence in msas that adopted it versus those financial services review 33(4) 126 that did not. the approach compares the change in outcomes over time between a group exposed to a treatment (msas with craigslist) and a control group (msas without craigslist). this design controls for unobserved, time-invariant characteristics of each msa (via fixed effects) and trends common to all msas over time (yearfixed effects). in this study, the "treatment" is craigslist's entry into the market. the did estimate captures the difference between pre-and post-craigslist fraud arrests in treated areas, net of any changes in untreated areas over the same time. results table 2 provides summary statistics for fraudrelated crime arrests, financial status, and demographics based on ucr and ipums cps data sources. the mean number of fraud arrests per 10,000 people was 10.488, with a standard deviation of 11.228. it should be noted that the total number of observations dropped when looking at aggregated cps data by msa, as not all msas had reported data. table 2. summary statistics for craigslist with ucr and ipums cps data variable mean standard deviation n craigslist 0.071 0.258 5,322 fraud arrests/10kp 10.488 11.228 5,322 median household income 47,363.474 9,715.377 4,292 poverty rate 12.805 4.521 4,292 unemployment rate 0.028 0.011 3,019 income under 50k 0.525 0.132 3,022 50k-100k income 0.191 0.072 3,020 % african/american 0.114 0.107 2,902 less than high-school 0.175 0.085 3,022 high-school 0.234 0.075 3,022 some college 0.209 0.043 3,022 note. summary statistics reflect averages across metropolitan statistical areas (msas) from 1995–2006. craigslist entry data were obtained from kroft & pope (2014). fraud arrest data are from the fbi’s uniform crime reporting (ucr) program, aggregated to the msa level. demographic and socioeconomic variables were derived from ipums cps data. observations vary across variables due to missing data in some msas or years 127 the relatively high standard deviation of fraud arrests per 10,000 residents (11.228) compared to the mean (10.488) reflects the expected rightskewness in crime data. to further address distributional concerns, we conducted additional robustness checks using alternative specifications, including the log transformation and standardization of the dependent variable. these yielded consistent treatment effects, suggesting that the skewness in fraud arrest data does not substantively affect the main results. to conserve space, results from the alternative specifications are available upon request. we also conducted robustness checks by excluding extreme outliers with alternative model specifications. table 3 presents the results of the difference-indifferences analysis estimating the effect of craigslist entry on fraud arrests.7 table 3 presents the results using the full sample (column 1), a coverage of 90% or more (column 2), the full sample using the msa population as weight (column 3), and a coverage of 90% or more using the msa population as the weight to account for the different size of each msa (column 4). some msas may have a population of a few hundred thousand people, while others can have a population in the millions. overall, results in table 3 support the rejection of the null hypothesis of no effect in favor of the alternative hypothesis that fraud was reduced by the introduction of craigslist, based on the significant and negative coefficient on the craigslist variable. given that the mean fraud arrests per 10,000 people was 10.488, the estimate in column (1) indicated that after craigslist entry, the fraud arrests decreased by 0.709 or about 7% compared to the control group. the results in column (2), where only msas with 90% coverage or more were included, painted a similar picture, showing that fraud arrests decreased by about 10% after craigslist entered an msa. the last two columns, (3) and (4), where we used population weights, produced similar results. when accounting for the population weights, the coefficients showed a decrease of 11% and 23% for the full sample and the one with 90% coverage or higher, respectively. the results were generally robust to other specifications using other coverage levels. table 3 differences-in-differences analysis with all years and coverage restrictions and weights all msas msas with coverage>90% all msas, weighted msas with coverage>90%, weighted (1) (2) (3) (4) b se b b se b b se b b se b craigslist 0.709*** 0.294 -0.994*** 1.117 -1.127** 0.586 -2.445*** 0.792 obs. 5,322 3,624 5,322 3,624 e (e2-a) 0.75 0.775 0.789 0.807 e (df-a) 377 370 377 370 note: * p < 0.05; ** p < 0 .01; *** p < 0.001. all specifications include msa-fixed effects and time-fixed effects. robust standard errors (se) are provided in separate columns. to validate the did approach, we tested for parallel pre-treatment trends using event study specifications with leads and lags of craigslist entry (table 4). no significant differences were observed in the pre-entry periods, supporting the validity of the did assumptions. 7 results are robust to the exclusion of san francisco, the first city to adopt craigslist. the estimates remain statistically significant and consistent in magnitude. full results are available upon request. 128 table 4 leads and lags all msas msas with coverage >90% all msas weighted msas with coverage>90%, weighted (1) (2) (3) (4) b se b b se b b se b b se b 5 periods prior -0.799** 0.357 -0.865** 0.386 -0.410 0.722 0.137 1.005 4 periods prior -0.715** 0.35 -0.722** 0.436 -0.274 0.57 -0.099 0.78 3 periods prior 0.561 0.345 -0.203 0.432 0.245 0.891 0.654 1.150 2 periods prior -0.110 0.335 0.168 0.459 0.493 0.57 0.606 0.843 1 period prior -0.442 0.364 -0.507 0.488 0.508 0.68 -0.545 0.54 1 year after -0.732 0.463 -0.805* 0.469 -0.152 0.574 -1.033*** 0.61 2 years after -0.893* 0.503 -1.286* 0.704 -1.925* 1.170 -3.228** 1.517 3 years after -0.731 0.587 -1.060 0.858 -1.851 1.378 -3.187* 1.839 4 years after -1.157 0.794 -1.376 1.014 -2.931* 1.765 -4.263* 2.177 5 years or more after -0.995 0.742 -1.253 0.901 -3.262** 1.502 -4.725** 1.858 obs. 5,322 3,624 5,322 3,624 e (e2-a) 0.749 0.774 0.793 0.812 e (df-a) 377 370 377 370 note: * p < 0.05; ** p < 0.01; *** p < 0.001. all specifications include msa-fixed effects and time-fixed effects. robust standard errors are provided in separate columns. to further support the robustness of our analysis and incorporate additional control variables, we accessed cps data spanning from 1995 to 2006 via ipums cps (flood et al., 2023). we subsequently aggregated the control variables at the msa level by aggregating the data utilizing bureau of labor statistics delineation files (u.s. census bureau, 2023). the empirical results were not affected by this adjustment. these additional results are available upon request. robustness checks in table 5, we conducted a series of robustness checks to examine the lagged effects of craigslist entry on fraud arrests. these checks aimed to investigate whether the introduction of craigslist in prior years had any significant influence on subsequent fraud arrest rates. the table provides results for different time lags, ranging from one to three years, and incorporates controls for lagged outcome variables. more specifically, column (1) used fraud arrests forwarded by one year regressed on craigslist entry using the full sample results. for example, instead of using fraud arrests in 2000, we used fraud arrests in 2001. column (2) used fraud arrests forwarded by 2 years, and column (3) used fraud arrests forwarded by 3 years. the estimates in all three columns were negative and significant, indicating that craigslist entry in prior years had a negative impact on fraud arrests. columns (4), (5), and (6) added controls for the lagged outcome variable. for example, column (4) added an additional control for the fraud arrests in a prior year (i.e., if the dependent variable was fraud arrests in 2000, the lagged variable was fraud arrests in 1999). the estimates remained consistently negative and statistically significant in these columns, except antonoudi et al. 129 for the last column, where statistical significance was not observed. table 5. robustness check with lagged effects post lag1 post lag 2 post lag 3 (1) (2) (3) (4) (5) (6) b se b b se b b se b craigslist (se) -1.858*** (0.707) -2.594*** (0.788) -1.985** (0.788) -1.149** (0.547) -1.292** (0.553) -0.220 (0.519) fraud arrests (se) 0.602* (0.037) post lag 1 (se) 0.701*** (0.046) post lag 2 (se) 0.68*** (0.04) post lag 3 obs. 5,321 5,320 5,319 5,321 5,320 5,319 e (e2-a) 0.665 0.573 0.518 0.737 0.729 0.715 e (df-a) 377 377 377 377 377 377 note: * p <0.05; ** p <0.01; *** p <0.001. all specifications include msa-fixed effects and time-fixed effects. robust standard errors are provided in brackets in separate rows. to further assess the robustness of our findings, we conducted several additional checks. these included models interacting craigslist’s presence with poverty and income levels, specifications using different sample restrictions, and estimators robust to staggered treatment timing, such as the callaway and sant’anna (2021) approach. because this method requires a balanced panel, we restricted the analysis to msas with complete data across all years. across all tests, the results consistently supported the main conclusion: craigslist’s entry was associated with a statistically significant decline in fraud arrests. full results are available upon request. discussion and implications discussion the findings from this study consistently show a negative relationship between craigslist entry and fraud arrest rates. rather than increasing fraud, craigslist’s presence may shape detection processes, reduce certain types of fraud, or shift activity to environments where it is less likely to result in arrest. there are several possible explanations for this outcome. the platform's peer-to-peer model may enhance transparency, limit third-party intermediaries, and increase the chance that users identify and flag suspicious behavior. even in largely anonymous settings, digital trails—such as emails or ip addresses— can deter certain types of fraud by increasing the risk of detection. in addition, craigslist may displace more misleading or aggressive advertising that consumers might otherwise encounter offline. these results are consistent with prior studies suggesting that craigslist and similar platforms can reduce certain types of offline risks. for instance, gurun et al. (2016) found that craigslist’s introduction reduced high-cost mortgage advertising, while cunningham et al. (2024) linked its “erotic services” section to declines in violent crime. at the same time, other work, such as heese et al. (2022), has shown that craigslist may reduce institutional oversight in other domains, such as corporate governance. this suggests that platform effects may vary depending on the type of fraud or enforcement mechanism involved. overall, while our study does not test mechanisms directly, the findings are consistent with the idea that platform design and user interactions can help mitigate consumer vulnerability in digital markets. these dynamics warrant further investigation in future research. financial services review 33(4) 130 conclusions utilizing both traditional did analysis and the staggered did approach along with a wider range of tests for the robustness of the findings, we consistently find a statistically significant negative relationship between craigslist entry and fraud arrests across various msas. this finding supports the view that the introduction of craigslist offers potential benefits in securing consumer transactions in a region. nevertheless, it is crucial to highlight that a decrease in fraud arrests does not automatically imply a decline in the overall instances of fraud. this research illuminates the nuanced impacts of online marketplaces like craigslist on societal metrics such as fraud arrests. the findings underscore the need for continued exploration in this domain, ensuring that the implications of these platforms are measured, examined, and understood. our results align with findings from other empirical studies suggesting that the introduction of craigslist leads to a decrease in certain types of offline crimes by shifting transactions from less public spaces to more public and digital spaces. for example, the introduction of the "erotic services" section on craigslist was linked to a reduction of violent crimes like female homicide rates, suggesting that the platform contributed to making sex work safer by moving it off the streets and allowing workers to better screen clients (cunningham et al., 2024). while small-scale frauds, like false listings, might persist and go unreported (mainly if they have not caused substantial financial or emotional loss for the victims), introducing a public and popular platform might reduce larger-scale fraud schemes due to enhanced public scrutiny. unlike numerous other online endeavors, craigslist has maintained its significant popularity for over twenty years (oravec, 2014). as such, listings receive significant attention from the public. larger-scale fraud schemes may find it challenging to operate under such scrutiny without being flagged or reported by multiple users. in-person interactions, often recommended for exchanges of goods, can reduce the anonymity that some larger fraud schemes rely upon. furthermore, newmark and his colleagues at craigslist have collaborated with law enforcement agencies throughout the united states, limiting postings that might lure people into unsafe or illegal situations (freese, 2011). implications and limitations this study provides empirical evidence of a decrease in fraud arrests following the adoption of craigslist. however, this decrease does not necessarily mean a decrease in fraud occurrence. it could imply a shift in how fraud is detected and reported, with more incidents being resolved through the platform's mechanisms or private settlements rather than official law enforcement channels. online platforms like craigslist can stimulate local economies by offering an avenue for people to sell and buy used and other goods using a free or low-cost medium for trading. as such, economic benefits and opportunities might indirectly contribute to reducing economicdriven types of fraud. the findings of this study can instill more confidence in consumers to transact on online platforms like craigslist. aware of the platform's impact, consumers might become more proactive in educating themselves about safer online transactions, utilizing platform guidelines correctly and best practices to avoid scams. this engagement can lead to economic savings and access to better deals along with a broader range of products and services. consumers should still exercise caution and understand that fraud that goes undetected or unreported on these platforms likely still exists. overall, consumers must understand the differences between online platforms, stay updated on recent developments, and take vigilant and proactive measures to protect themselves from potential risks. financial services professionals need to understand emerging digital behaviors to help clients navigate online marketplaces more securely. by tailoring financial advice to reflect the growing role of peer-to-peer platforms, practitioners can support informed, risk-aware participation in digital economic activity particularly as clients pursue side income or engage in informal trading environments. this study contributes to the fraud and consumer behavior literature and demonstrates how did methods can be applied to evaluate marketplace innovations within financial services. as antonoudi et al. 131 financial transactions increasingly occur online, platforms like craigslist, venmo, and facebook marketplace offer rich opportunities for further did-based evaluations—ranging from fraud to credit access, trust, and digital financial literacy. we hope this study encourages future applications of quasi-experimental designs in digital financial research. this study has several limitations. first, the skewed distribution of fraud arrests is inherent in administrative crime data. while we employed robustness checks and sample restrictions to address this issue, future studies may benefit from alternative fraud measures, including monetary loss or victim-reported data, to capture additional dimensions of fraud. second, the uniform crime reporting (ucr) dataset does not allow isolation of consumer fraud from other fraud-related arrests. while many categories (e.g., impersonation, wire fraud, false pretenses) likely reflect crimes affecting consumers, the aggregate nature of the data introduces some ambiguity. third, the use of arrests is an indirect measure of actual fraud occurrence that depends on several factors related to crime reporting and law enforcement action. fourth, although our difference-in-differences (did) framework adjusts for unobserved time-invariant differences and shared temporal shocks, craigslist adoption may still be correlated with unobserved local characteristics. while we used robustness checks and staggered rollout to mitigate this concern, a formal probit model predicting craigslist entry could further validate our identification strategy and remains an area for future research. finally, we do not test the mechanisms underlying the observed reduction in arrests—such as digital trails, peer flagging, or reduced exposure to deceptive advertising. these remain promising areas for future research on how marketplace design influences consumer safety outcomes. references anderson, d. m. (2014). in school and out of trouble? the minimum dropout age and juvenile crime. review of economics and statistics, 96(2), 318-331. anderson, k. b. (2004). consumer fraud in the united states: an ftc survey. https://www.ftc.gov/sites/default/files/d ocuments/reports/consumer-fraudunited-states-ftcsurvey/040805confraudrpt.pdf brynjolfsson, e., & smith, m. d. (2000). frictionless commerce? a comparison of internet and conventional retailers. management science, 46(4), 563-585. callaway, b., & sant’anna, p. h. (2021). difference-in-differences with multiple time periods. journal of econometrics, 225(2), 200-230. cunningham, s., deangelo, g., & tripp, j. (2017, january). craigslist’s effect on violence against women. paper presented at the american economic association conference, chicago, il. cunningham, s., deangelo, g., & tripp, j. (2024). did craigslist’s erotic services reduce female homicide and rape? journal of human resources, 59(1), 280-315. federal bureau of investigation. (2004). uniform crime reporting handbook. u.s. department of justice. https://ucr.fbi.gov/additional-ucrpublications/ucr_handbook.pdf federal trade commission. (2006). consumer fraud and identity theft complaint data january december 2006. https://www.ftc.gov/sites/default/files/d ocuments/reports_annual/sentinel-cy2006/sentinel-cy2006.pdf flood, s., king, m., rodgers, r., ruggles, s., robert, j. w., backman, d., chen, a., cooper, g., richards, s., schouweiler, m., & westberry, m. (2023). ipums cps: version 11.0 [dataset]. https://www.ipums.org/projects/ipumscps/d030.v11.0 fone, z. s., sabia, j. j., & cesur, r. (2023). the unintended effects of minimum wage increases on crime. journal of public economics, 219, 104780. freese, s. m. (2011). craigslist the company and its founder (r. rowell, ed.). abdo publishing company. fremstad, a. (2017). does craigslist reduce waste? evidence from california and https://www.ftc.gov/sites/default/files/documents/reports/consumer-fraud-united-states-ftc-survey/040805confraudrpt.pdf https://www.ftc.gov/sites/default/files/documents/reports/consumer-fraud-united-states-ftc-survey/040805confraudrpt.pdf https://www.ftc.gov/sites/default/files/documents/reports/consumer-fraud-united-states-ftc-survey/040805confraudrpt.pdf https://www.ftc.gov/sites/default/files/documents/reports/consumer-fraud-united-states-ftc-survey/040805confraudrpt.pdf https://ucr.fbi.gov/additional-ucr-publications/ucr_handbook.pdf https://ucr.fbi.gov/additional-ucr-publications/ucr_handbook.pdf https://www.ftc.gov/sites/default/files/documents/reports_annual/sentinel-cy-2006/sentinel-cy2006.pdf https://www.ftc.gov/sites/default/files/documents/reports_annual/sentinel-cy-2006/sentinel-cy2006.pdf https://www.ftc.gov/sites/default/files/documents/reports_annual/sentinel-cy-2006/sentinel-cy2006.pdf financial services review 33(4) 132 florida. ecological economics, 132, 135-143. garg, v., & nilizadeh, s. (2013, may). craigslist scams and community composition: investigating online fraud victimization. in 2013 ieee symposium on security and privacy workshops (spw) (pp. 123-126). ieee. https://doi.org/10.1109/spw.2013.21 gurun, u. g., matvos, g., & seru, a. (2016). advertising expensive mortgages. the journal of finance, 71(5), 2371–2416. https://doi.org/10.1111/jofi.12423 heese, j., pérez-cavazos, g., & peter, c. d. (2022). when the local newspaper leaves town: the effects of local newspaper closures on corporate misconduct. journal of financial economics, 145(2), 445–463. https://doi.org/10.1016/j.jfineco.2021.08 .015 hill, r. p., & sharma, e. (2020). consumer vulnerability. journal of consumer psychology, 30(3), 551-570. knowles, j. (2014, october 29). additional charges filed in craigslist rental scam. abc7 chicago. https://abc7chicago.com/i-teamcraigslist-apartment-scam-rental-desplaines/372081/ kroft, k., & pope, d. g. (2014). does online search crowd out traditional search and improve matching efficiency? evidence from craigslist. journal of labor economics, 32(2), 259-303. lee, c. s. (2022). how online fraud victims are targeted in china: a crime script analysis of baidu tieba c2c fraud. crime & delinquency, 68(13–14), 2529–2553. https://doi.org/10.1177/0011128721102 9862 lee, c. s. (2021b). online fraud victimization in china: a case study of baidu tieba. victims & offenders, 16(3), 343-362. lee, j., & soberon‐ferrer, h. (1997). consumer vulnerability to fraud: influencing factors. journal of consumer affairs, 31(1), 70-89. oravec, j. a. (2014). craigslist in crisis: issues of censorship and moral panic in the context of online communities. international journal of the academic business world, 8(2), 1-11. park, y., jones, j., mccoy, d., shi, e., & jakobsson, m. (2014). scambaiter: understanding targeted nigerian scams on craigslist. in ndss 2014: proceedings of the network and distributed system security symposium. internet society. https://elaineshi.com/docs/scambaiter.pd f rayo, a. (2023, february). the top scams of 2022. https://consumer.ftc.gov/consumeralerts/2023/02/top-scams-2022 u.s. census bureau. (1995-2006). u.s. census bureau county–level data [data set]. u.s. census bureau. (2023, august 4). delineation files. https://www.census.gov/geographies/ref erence-files/time-series/demo/metromicro/delineation-files.html van wilsem, j. (2013). 'bought it, but never got it' assessing risk factors for online consumer fraud victimization. european sociological review, 29(2), 168-178. https://doi.org/10.1093/esr/jcr053 https://doi.org/10.1111/jofi.12423 https://abc7chicago.com/i-team-craigslist-apartment-scam-rental-des-plaines/372081/ https://abc7chicago.com/i-team-craigslist-apartment-scam-rental-des-plaines/372081/ https://abc7chicago.com/i-team-craigslist-apartment-scam-rental-des-plaines/372081/ https://elaineshi.com/docs/scambaiter.pdf https://elaineshi.com/docs/scambaiter.pdf https://consumer.ftc.gov/consumer-alerts/2023/02/top-scams-2022 https://consumer.ftc.gov/consumer-alerts/2023/02/top-scams-2022 https://www.census.gov/geographies/reference-files/time-series/demo/metro-micro/delineation-files.html https://www.census.gov/geographies/reference-files/time-series/demo/metro-micro/delineation-files.html https://www.census.gov/geographies/reference-files/time-series/demo/metro-micro/delineation-files.html 133 appendix table a. variable definitions and data sources symbol variable description data source y (m,t) fraud arrests fraud arrests per 10,000 people in msa m at time t fbi uniform crime reporting (ucr) program craigslist_mt craigslist entry binary indicator if craigslist is available in msa m at time t kroft & pope (2014) x (t) control variables income, unemployment, education, racial composition ipums cps β1 treatment effect coefficient estimating craigslist's effect on fraud arrests author’s model ym msa fixed effects controls for time-invariant characteristics in each msa author’s calculations nt year fixed effects controls for national shocks or trends at year t author’s calculations εmt error term captures unobserved factors influencing fraud arrests author’s model note. all variables were aggregated or matched to the metropolitan statistical area (msa) level using us census delineation files and the qcew county-to-msa crosswalk. ipums cps refers to data from flood et al. (2023). craigslist entry data were obtained from kroft and pope (2014). pii: s1057-0810(00)00066-4 performance persistence and management skill in nonconventional bond mutual funds james philpota, douglas hearthb,*, james rimbeyb afrank d. hickingbotham school of business, ouachita baptist university, arkadelphia, ar 71998, usa bdepartment of finance, sam m. walton college of business administration, university of arkansas, fayetteville, ar 72701, usa received 18 july 2000; received in revised form 9 october 2000; accepted 21 november 2000 abstract recent empirical research has identified a tendency for equity mutual funds to provide consistent performance relative to other funds over time. studies of bond funds have centered around investment grade, straight bonds and have concluded that fund managers outperform indexes on a gross (although not net) basis, but that performance is hampered by high expense levels. we examine nonconventional bond funds (high-yield bonds, global issues and convertible bonds) and find that short-term performance persistence is present, but limited to the high-yield bond subsample. fund managers are unable to distinguish themselves in the long term, despite the diverse nature of the funds they oversee. © 2001 elsevier science inc. all rights reserved. jel classification:g2/g29 keywords:mutual fund performance; non-conventional bond funds 1. introduction mutual fund relative performance persistence and management effectiveness are controversial and popular topics in the finance literature. several empirical studies identify a tendency for mutual funds to provide consistent returns performance over time relative to * corresponding author. tel.:11-501-575-4505; fax:11-501-575-8407. e-mail addresses:dhearth@walton.uark.edu (d. hearth), jphilpot@alpha.obu.edu (j. philpot), jrimbey@ walton.uark.edu (j. rimbey). financial services review 9 (2000) 247–258 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(00)00066-4 other funds (e.g., grinblatt and titman, 1992, hendricks et al., 1993, brown and goetzmann, 1995). although research (hendricks and patel, 1997, brown and goetzmann, 1997) has yet to establish whether the persistence phenomenon is the result of management skill or biases in the data, gruber (1996) shows that investors who “chase past performance” are rational wealth maximizers. the preponderance of mutual fund studies either sample all mutual fund types or focus solely on equity mutual funds. the study of bond mutual funds as a separate group has been limited, largely because of (until recently) limited sample size. those studies that have examined the performance of investment grade, straight bond mutual funds have found that: (1) these funds in the aggregate outperform appropriate indexes on a gross, but not net, basis (gudikunst and mccarthy, 1992); and (2) bond fund performance is hampered by high expenses and not correlated with prior returns (blake et al., 1993, philpot et al., 1998). the study of nonconventional bond mutual fund performance has been limited, with results to date indicating that high-yield bond mutual funds do not outperform relevant indexes (gudikunst and mccarthy, 1992). we extend current knowledge by examining management skill in managing nonconventional bond mutual funds, defined as speculative grade, global or convertible bond funds, using a performance persistence and performance-related variables approach. we find that relative performance persistence is at best a short-run phenomenon limited to our high-yield bond fund sample. performance persistence is no more likely among funds that had no changes in management than in funds that changed managers during the sample period. further, among high-yield bond funds, risk-adjusted performance is inversely related to portfolio turnover. 2. management effectiveness and determinants of performance the existence of a consistent skill level among conventional bond mutual fund managers has been evaluated in several ways. one approach is to compare the aggregate performance of the mutual fund industry or industry segment to the returns of appropriate market indexes. gudikunst and mccarthy (1992) examine the risk and return characteristics of a sample of 25 bond mutual funds over the time period 1976–1989. they find that, on average, the funds in their sample provided risk-adjusted gross returns that were greater than the returns to a broad bond market index. however, after subtracting fees and expenses, the funds’ returns only matched that of the index. blake et al. (1993) further examine bond fund performance using a survivorship bias-adjusted sample of 41 bond mutual funds and a larger nonadjusted sample of bond funds and find that bond mutual fund average net risk-adjusted returns are lower across samples and subsamples than the returns to the relevant indexes. another common approach is to examine relative fund performance persistence through time. while efficient markets theory predicts that individual mutual fund returns will be uncorrelated over time, a positive serial correlation in individual mutual fund returns may indicate that mutual fund managers are consistent in their ability to generate returns relative to their peers. grinblatt and titman (1992), hendricks et al. (1993) brown and goetzmann (1995) and gruber (1996) all find strong persistence of fund relative performance over varied 248 j. philpot / financial services review 9 (2000) 247–258 time horizons. these findings (which examine either equity funds or all categories of mutual funds together) support the presence and persistence of management skill. however, when bond mutual funds are examined by themselves, blake et al. (1993) find no evidence of interperiod consistency in either the performance ranking or the risk-adjusted returns of the 41 bond mutual funds in their sample. philpot et al. (1998) affirm this result, finding in a sample of 27 investment grade, nonconvertible bond mutual funds, that risk-adjusted performance does not predict future performance. in addition to examining serial correlation in returns, mutual fund management can be evaluated by examining the relation between mutual fund returns and individual fund attributes that represent the manager’s activity, such as portfolio turnover and the level of fund expenses. in an efficient market with costly information, resources spent on security analysis and portfolio management should not increase risk-adjusted net portfolio returns. alternatively, when mutual fund managers are able to increase returns beyond their expenses, fund management activity adds value. carhart (1997) finds that mutual fund expenses decrease fund performance and concludes that there is no benefit from active management. in bond mutual funds specifically, recent studies have found negative relations between bond mutual fund performance and fund expenses and fees (blake et al., 1993; philpot et al., 1998). another measure of management activity is a mutual fund’s portfolio turnover rate. actively managed mutual funds exhibit high turnover rates, while passively managed funds tend to report low portfolio turnover. empirical studies are yet to categorically discern the effects of portfolio turnover on mutual fund returns. grinblatt and titman (1989) examine the risk-adjusted returns to the quarterly-updated portfolios of aggressive growth mutual funds versus the returns to the funds’ initial portfolios. they find that the updated portfolios provided the greater returns, and thus they conclude that active portfolio management generated at least some of the returns to the funds. in a more recent study, carhart (1997) finds a negative relation between portfolio turnover and risk-adjusted returns. when investment grade bond funds are examined separately, this negative relation persists (philpot et al., 1998). a final way to test for management skill is to determine whether managers learn by experience. golec (1996) examines the effects of management characteristics on a sample of equity and balanced mutual fund risk-adjusted returns. he finds that the most significant predictor of mutual fund performance is the amount of time a manager has been with a particular fund. this result may indicate a learning effect, with managers increasing their ability to manage a particular fund over time. several additional mutual fund attributes have been hypothesized to be related to relative returns. the most common include fund size and the presence of sales or distribution fees. there is evidence to suggest that as mutual funds grow in size, their ability to provide commensurate returns becomes handicapped. this is especially true for equity funds. grinblatt and titman (1989) find that in their sample of aggressive growth mutual funds, small funds have higher risk-adjusted gross returns than large funds, and they conclude that mutual funds lose market mobility and the ability to take positions in small-capitalization issues as they increase in size. these empirical findings corroborate an increasing tendency for large equity mutual funds (such as fidelity magellan and janus twenty) to end share sales to new investors. bond mutual funds, however, have been shown to have positive economies of 249j. philpot / financial services review 9 (2000) 247–258 scale as evidenced by a positive relationship between fund size and risk-adjusted returns (gudikunst and mccarthy, 1992, philpot et al., 1998). most studies show that loads and distribution (12b-1) fees have little impact on net mutual fund performance. (see golec, 1996 for recent evidence.) 3. nonconventional bond funds extant studies cast doubt upon the ability of conventional bond mutual fund managers to consistently outperform either a relevant market index or their manager peers. however, there is both conventional wisdom and some empirical evidence to suggest that nonconventional bonds (high-yield, global, and convertibles) as a group have considerably different investment characteristics than conventional bonds. for instance, a convertible bond’s conversion feature essentially transforms the bond into a type of equity call option, with equity-like characteristics. global bonds expose the domestic investor to additional risks, including exchange rate risks and country-specific risks. high-yield bonds in particular have been shown to have returns characteristics more similar to equity than to investment-grade debt (bookstaber and jacob, 1986). blume et al. (1991) examine the performance of high-yield bonds, finding that high-yield bonds have lower risk and higher returns over their sample period than investment grade bonds and that high-yield bonds behave like both bonds and stocks. also, brister et al. (1994) show that the market for high-yield bonds is a distinct debt market segment with its own default structure of interest rates. this study examines management effects in high-yield, convertible and global bond funds. current research has compared the average performance of high yield bond (gudikunst and mccarthy, 1997) and global bond (detzler, 1999) mutual funds to relevant indexes, finding that in the aggregate, these funds do not offer superior returns. however, no study evaluates management skill in nonconventional bonds by examining performance persistence or management activity. as stated earlier, nonconventional bonds have characteristics that make them different from straight domestic bonds. if these characteristics of the nonconventional debt markets create enough diversity among the individual issues in these markets, then professional managers may display consistent relative performance and performance that is commensurate with measures of management activity and tenure. 4. sample we examine a sample of 73 nonconventional bond mutual funds obtained from morningstar ondisc over the period 1988–1997. the sample includes 53 high-yield, 10 convertible and 10 global bond funds. (morningstar ondisc reports that in 1988, there were a total of 90 such funds, including 62 high-yield, 16 convertible and 12 global bond funds. thus our sample includes a substantial number of the then-existing funds.) we calculated the funds’ sharpe performance measure over oneand five-year time horizons using quarterly returns. the sharpe measure uses the standard deviation of returns to adjust for risk and provides a risk-adjusted relative performance measure that best suits this study. 250 j. philpot / financial services review 9 (2000) 247–258 using data from such a long time period may increase the risk of survivorship bias in the sample. the risk of bias is greatest in studies that primarily seek to measure aggregate mutual fund performance against an unmanaged index over time, because the best performing funds are the ones most likely to continue in independent operation. brown and goetzmann (1997) and hendricks and patel (1997) suggest that survivorship may also cause spurious performance persistence estimates. we acknowledge the possibility of survivor bias in our data set. because survival bias in mutual funds is largely a function of returns differences among funds, we compare the mean annual returns of our sampled funds to population means reported by morningstar over the sample period. table 1 shows the mean returns of the sampled funds and the population means. hypothesis tests show no significant differences in the sample and population means for any of the fund types. this observation and the fact that our sample includes a large proportion of the population at the beginning of the sample period suggest that any bias present is likely to be small. 5. analysis we begin by examining the tendency of nonconventional bond mutual fund managers to maintain their risk-adjusted performance ranking from period to period. under the hypothesis of no performance persistence, a mutual fund’s ranking in one period should be independent of its ranking in the prior period. that is, a fund in one quintile rank in one period should be equally likely to be in any quintile in a following period. table 2 shows contingency tables of prior and following one-year quintile ranks of nonconventional mutual fund sharpe measure performance over the sample period. panel a contains results for the full sample of 73 funds. ax2 test rejects the hypothesis of a uniform distribution in the table. inspection of the table values reveals, most notably, that funds performing in the bottom quintile in the first year are likely to remain poor relative performers. there appears to be a very weak tendency for funds in the top or second quintiles to remain good performers. middle-performing funds tend to remain in the middle in the subsequent one-year period. the apparent persistence seems to be driven by the high-yield bond mutual funds. panel b of table 2 shows the contingency table test with one-year sharpe measures for the 53 high-yield funds. the panel b results roughly mirror those of the full sample. panels c and d of table 2 show, respectively, the same contingency table tests for the convertible bond and global bond funds. unlike the high-yield funds, these mutual funds exhibit a uniform table 1 sample and morningstar ondisc population mean annual mutual fund returns over the period 1988–1997, by fund type. t-statistics test the null hypothesis of equal population and sample means. fund type sample mean population mean t-statistic high yield 11.45% 10.63% 0.01 convertible 13.69% 13.24% 0.01 world bond 7.58% 8.13% 20.02 251j. philpot / financial services review 9 (2000) 247–258 distribution. thus, it appears that in the near term (based on one-year returns) there is weak evidence of performance persistence among high-yield bond funds, but not the other fund types. recognizing that mutual funds may change managers and that manager change may affect performance consistency, we tested only the 24 nonconventional bond funds that had the same manager throughout the sample period. panel e of table 2 contains these results.x2 tests are unable to reject the hypothesis of a uniform distribution. thus, even when mantable 2 chi-squared tests of non-conventional bond mutual fund relative performance persistence in one-year increments over the period 1988–1997 using the sharpe measure. the number in each cell indicates the number of funds in each prior year’s quintile grouping and subsequent year quintile rank. panel a: non-conventional bond funds, 73 funds, 9 time periods, 657 observations prior year quintile quintile rank 1 2 3 4 5 total 1 28 28 27 32 20 135 2 34 24 34 27 16 135 3 20 34 37 26 18 135 4 26 27 19 27 27 126 5 27 22 18 14 45 126 total 135 135 135 126 126 657 test for uniform distribution:x2 (16) 5 46.86; p, 0.0001 panel b: high-yield bond funds, 53 funds, 9 time periods, 477 observations. prior year quintile quintile rank 1 2 3 4 5 total 1 23 18 24 20 14 99 2 27 21 18 18 15 99 3 17 21 23 19 19 99 4 13 26 16 19 16 90 5 19 13 18 14 35 90 total 99 99 99 90 90 477 test for uniform distribution:x2 (16) 5 33.16; p5 0.007 panel c: convertible funds, 10 funds, 9 time periods, 90 observations. prior year quintile quintile rank 1 2 3 4 5 total 1 5 3 3 4 3 18 2 3 3 4 6 2 18 3 3 7 5 1 2 18 4 4 2 3 4 5 18 5 3 3 3 3 6 18 total 18 18 18 18 18 90 (continued on next page) 252 j. philpot / financial services review 9 (2000) 247–258 agement is constant, short-term management relative performance is not consistent. it is also well established that high expense ratios diminish bond mutual fund performance (blake et al.., 1993, and philpot et al.., 1998). we separately examined the 35 funds in our sample with the highest average expense ratios to see whether they showed performance persistence. as panel f of table 2 shows, there is weak, but not statistically significant, evidence of performance persistence. most notably, there is some tendency for the worst funds to remain poor performers. table 2 (continued) panel d: global funds, 10 funds, 9 time periods, 90 observations. prior year quintile quintile rank 1 2 3 4 5 total 1 3 6 4 1 4 18 2 5 3 4 3 3 18 3 3 3 1 5 6 18 4 6 3 5 4 0 18 5 1 3 4 5 5 18 total 18 18 18 18 18 90 test for uniform distribution:x2 (16) 5 17.78; p5 0.03369 panel e: non-conventional bond funds having no management change, 24 funds, 9 time periods, 216 observations. prior year quintile quintile rank 1 2 3 4 5 total 1 13 7 11 7 7 45 2 14 9 6 10 6 45 3 7 13 10 7 8 45 4 6 9 13 10 7 45 5 5 7 5 11 8 36 total 45 45 45 45 36 216 test for uniform distribution:x2 (16) 5 13.47; p5 0.6379 panel f: non-conventional bond funds having highest average annual expenses, 35 funds, 9 time periods, 315 observations. prior year quintile quintile rank 1 2 3 4 5 total 1 14 11 7 15 16 63 2 11 19 12 13 8 63 3 12 9 18 14 10 63 4 16 16 14 8 9 63 5 10 8 12 13 20 63 total 63 63 63 63 63 315 test for uniform distribution:x2 (16) 5 24.76; p5 0.0741 253j. philpot / financial services review 9 (2000) 247–258 the tests were repeated using sharpe measures from two five-year periods, 1988–1992 and 1993–1997, to determine whether there is evidence of persistence over longer time horizons. table 3 shows these results. tests conducted with the full sample, the subsample of high-yield bond funds, and the funds with no management change (see panels a, b and c of table 3), showed no significant departure from a uniform distribution. tests conducted with the high-expense funds (panel d of table 3) reject the hypothesis of a uniform distribution. interestingly, the presence of larger numbers in the off-diagonal cells of the table indicates an apparent mean reversion in relative fund performance of this group. these results indicate that over longer time periods, there is no performance persistence among the nonconventional bond funds. thus fund managers do not show consistent relative performance—a finding consistent with the philpot et al. (1998) findings for domestic straight bond funds. also, the discrepancy between the one and five-year results for the full sample and the high-yield bond funds is consistent with the findings of hendricks et al. (1993) that performance persistence is strongest when measured over relatively short periods of time. we next employ a cross-sectional regression model similar to philpot et al. (1998) to analyze relations between a mutual fund’s five-year sharpe measure and five independent table 3 chi-squared tests of non-conventional bond mutual fund relative performance persistence in five-year increments over the period 1988–1997 using the sharpe measure. the number in each cell indicates the number of funds in each prior year’s quintile grouping and subsequent year quintile rank. panel a: non-conventional bond funds, 73 funds. prior five-year quintile five-year quintile rank 1 2 3 4 5 total 1 5 2 1 4 3 15 2 1 4 2 3 5 15 3 3 4 4 2 2 15 4 3 5 1 4 1 14 5 3 0 7 1 3 14 total 15 15 15 14 14 73 test for uniform distribution:x2 (16) 5 22.5; p5 0.1278 panel b: high-yield bond funds, 53 funds. prior five-year quintile five-year quintile rank 1 2 3 4 5 total 1 3 2 2 2 2 11 2 2 2 3 2 2 11 3 3 4 2 0 2 11 4 2 3 1 1 3 10 5 1 0 3 5 1 10 total 11 11 11 10 10 53 (continued on next page) 254 j. philpot / financial services review 9 (2000) 247–258 variables, including the fund’s lagged sharpe measure, its expense ratio, its portfolio turnover rate, and the natural logarithm of its net assets. similar to philpot et al., we estimated a pooled time series-cross sectional regression model using one-year time periods. chow tests indicated that the regression parameters were not stable over time; thus inferences from the pooled data are invalid. this result held for the full sample and all sub samples. in each case, ignoring the chow test results, the pooled regression results failed to indicate any performance persistence. following golec (1996) we also include the mutual fund manager’s tenure. we omit loads and distribution fees because virtually all the funds in our sample charged such fees. the regression equation is estimated as: returni5b01b1(lsharpe)1b2(expensei)1b3(turnoveri)1 b4(assetsi)1b5(tenurei)1ei, (1) where return 5 the fund’s five-year sharpe measure; i 5 1, . . . 73, the number of mutual funds in the regression, and denotes the individual fund; table 3 (continued) panel c: non-conventional bond funds with same manager, 24 funds. prior five-year quintile five-year quintile rank 1 2 3 4 5 total 1 1 1 1 1 1 5 2 0 0 2 1 2 5 3 2 0 2 1 0 5 4 1 3 0 0 1 5 5 1 1 0 2 0 4 total 5 5 5 5 4 24 test for uniform distribution:x2 (16) 5 17.33; p5 0.3645 panel d: non-conventional bond funds with high expenses, 35 funds. prior five-year quintile five-year quintile rank 1 2 3 4 5 total 1 0 1 4 1 1 7 2 2 1 2 2 0 7 3 1 0 0 1 5 7 4 2 2 1 2 0 7 5 2 3 0 1 1 7 total 7 7 7 7 7 24 test for uniform distribution:x2 (16) 5 27.14; p5 0.0399 255j. philpot / financial services review 9 (2000) 247–258 b0 5 an intercept term; lsharpe5 the fund’s prior five-year period sharpe measure; expense5 the fund’s average expense ratio; turnover 5 the fund’s average portfolio turnover rate; assets5 the fund’s total assets at the beginning of the period; tenure 5 the fund manager’s tenure in years; e 5 a residual term. table 4 shows regression results using the full sample. diagnostic tests showed no violations of ols assumptions. the regression results indicate that for nonconventional bond funds over a five-year horizon, risk-adjusted returns are unrelated to any of the independent variables, with the exception of portfolio turnover. the results provide no evidence of consistency in relative performance by nonconventional bond mutual fund managers. the insignificant expense variable (t5 20.5829) and negative turnover variable (t5 22.4148) indicate that resources and effort spent by management are not rewarded with increased net returns. nonconventional bond mutual funds appear to neither benefit from economies of scale nor suffer from scale diseconomies based on large asset size. the management tenure variable, significant at the 0.10 level (t5 1.5442), provides only weak evidence of increased management expertise with increased time managing a particular fund. we repeated the regression model estimation for the high-yield bond fund subsample and the subsample of funds that had no managerial change over the sample period. in both of these cases, the regression model itself was insignificant, meaning that fund returns were not table 4 results from regression of five-year sharpe performance measure on independent variables for nonconventional bond mutual funds, 1988–1997, 73 funds. independent variable parameter estimate t-statistic lsharpe 20.0427 20.3561 expense 20.0360 20.5829 turnover 20.0007 22.4148** assets 20.0001 20.6861 tenure 0.0085 1.5442* model f 5 2.6625, p5 0.0296. model adjusted r2 5 0.1035. * significant at 0.10 level ** significant at 0.01 level lsharpe5 the fund’s prior five-year period sharpe measure. expense5 the fund’s average expense ratio. turnover 5 the fund’s average portfolio turnover rate. assets5 the fund’s total assets at the beginning of the period. tenure 5 the fund manager’s tenure in years. 256 j. philpot / financial services review 9 (2000) 247–258 a function of any of the independent variables. thus there is again no evidence of performance persistence or management effectiveness. 6. conclusion prior work has shown that there is persistence in relative performance in equity mutual funds. this has led to speculation that there are systematic and persistent differences in skill levels among mutual fund managers. these skill differences have not shown up in studies of straight bond funds, perhaps due to the relative homogeneity of investment grade bonds. although managers of high-yield, convertible and global bond funds may face more diverse investment opportunities, we find that there is at best a very modest short-run persistence in relative fund performance of such funds, and this appears limited to the high-yield fund subset. examination of returns over a longer time period shows no evidence of management skill or performance persistence in nonconventional bond mutual funds. given these results and those of other studies, it is apparent that the growth in the bond sector of the mutual fund industry is not the result of expertise in professional management. rather, it appears that rational investors may invest in bonds through mutual funds merely to take advantage of the financial intermediary functions (diversification, liquidity, and so forth) these funds perform. acknowledgment the authors wish to thank the editor and two anonymous reviewers for their helpful comments and suggestions on an earlier version of this paper. references blake, c. r., elton, e. j., & gruber, m. j. (1993). the performance of bond mutual funds.journal of business, 66, 371–403. blume, m. e., keim, d. b., & patel, s. a. (1991). returns and volatility of low-grade bonds 1977–1989.journal of finance, 46,49–74. bookstaber, r., & jacob, d. (1986). the composite hedge: controlling the credit risk of high yield bonds. financial analysts’ journal, 43,12–25. brister, w., r. kennedy, liu, p. (1994). the regulation effect of credit rating on bond interest yield: the case of junk bonds.journal of business finance and accounting, 21,511–531. brown, s. j., & goetzmann, w. n. (1995). performance persistence.journal of finance, 50,679–699. brown, s. j., & goetzmann, w. n. (1997). rejoinder: the j-shape of performance persistence given survivorship bias.review of economics and statistics, 79,167–170. carhart, m. m. (1997). on persistence in mutual fund performance.journal of finance, 52,57–82. detzler, l. m. (1999). the performance of global bond mutual funds.journal of banking and finance, 23, 1195–1217. golec, j. h. (1996). the effects of mutual fund managers’ characteristics on their portfolio performance, risk and fees.financial services review, 5,133–148. 257j. philpot / financial services review 9 (2000) 247–258 grinblatt, m., & titman, s. (1989). mutual fund performance: an analysis of quarterly portfolio holdings.journal of business, 62,393–416. grinblatt, m., & titman, s. (1992). the persistence of mutual fund performance.journal of finance, 47, 1977–1985. gruber, m. (1996). another puzzle: the growth in actively managed mutual funds.journal of finance, 51, 783–810. gudikunst, a., & mccarthy, j. (1992). determinants of bond mutual fund performance.journal of fixed income, 2, 35–46. gudikunst, a., & mccarthy, j. (1997). high-yield bond mutual funds: performance, january effects, and other surprises.journal of fixed income, 7,95–101. hendricks, d., & patel, j. (1997). the j-shape of performance persistence given survivorship bias,review of economics and statistics, 79,161–166. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974–1988.journal of finance, 48,93–130. philpot, j., hearth, d., rimbey, j., & schulman, c. (1998). active management, fund size and bond mutual fund returns.the financial review, 33,115–126. 258 j. philpot / financial services review 9 (2000) 247–258 pii: s1057-0810(99)00026-8 from the editor karen eilers lahey the first issue of the vol. 8 for 1999 brings several new developments to the journal. first, there are changes in the editorial board of financial services review. i very much thank those individuals who have served as associate editors for all of their time and effort. for the newly appointed members we look forward to sending you lots of manuscripts and seeking your advice. secondly, we are moving to a more electronic based system. this new approach will encompass the ability to accept manuscript submission by e-mail attachment from current members, electronic transmission and retrieval of manuscripts and reviews from referees, as well as electronic transfer to the publisher. hopefully this will be more convenient and shorten the review process. third, the journal will no longer be carrying the book, software, and website reviews column that has been so ably edited by dr. douglas r. kahl. all of his help in providing this feature is very much appreciated. finally, as of january 1, 1999 elsevier/north holland has purchased jai press, the journal’s former publisher. you will notice some minor changes in how the journal looks with the new publisher. in addition, elsevier/north holland brings standardization to their journals in such areas as classification by jel codes and keywords to make it easier to electronically search for article via online data bases. look for new submission requirements in the journal as well as on our website at http://www.uakron.edu/cba/fsr/fsr.html. the lead article in this issue is entitled, “gender differences in defined contribution pension decisions” and is written by vickie l. baitelsmit, alexandra bernasek, and nancy a. jianakoplos. they utilize the national survey of consumer finances to examine the question of possible risk aversion differences by gender. their results suggest that there are significant differences in allocation of wealth that may impact retirement consumption well into the next century. kwok ho, moshe arye milevsky, and chris robinson develop a model to determine potential shortfall risk from a desired consumption level for a retired investor. their article entitled, “international equity diversification and shortfall risk” looks at potential benefits of diversification for a 65-year-old canadian woman and a 65-year-old american woman. it very much addresses the issue raised by bajtelsmit, bernasek, and jianakoplos in the first financial services review 8 (1999) v–vi 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00026-8 article about retirement differences by gender and examines a shortfall solution. results indicate that the canadian woman would benefit but the american would not benefit. the third article entitled, “international index funds and the investment portfolio” by scott aiello and natalie chieffe provides a review of recent academic literature on international investing and examines the benefits of using international index funds for portfolio diversification. their results indicate that these funds do not outperform the s & p500, but do provide a reduction in risk. michael hanna, joseph p. mccormack, and grady perdue look at portfolios based on investing in the major market indexes of the the g-7 countries in their article entitled, “a nineties perspective on international diversification”. they conclude that for the period that is tested, the s & p 500 dominates all two-country portfolios that are constructed. the fifth and last article tests for a positive impact for firms that are designated as family friendly firms. dianna c. preece and greg filbeck find that there is no statistically significant difference in raw returns or risk-adjusted returns in their study entitled, “family friendly firms: does it pay to care?”. vi k.e. lahey / financial services review 8 (1999) v–vi financial services review, 33(2) 55 consumer margin use: understanding the role of peer influence, investment literacy, and age kaplan sanders1 and olamide olajide2 abstract very little has been observed regarding household decisions around margin use. using the 2021 wave of the national financial capability study (nfcs), this study investigates margin use from a consumer’s perspective. using probit analysis and observing correlations, relationships between peer influence, investment literacy, age, and margin use are explored. results indicate a positive peer influence on the decision to buy on margin. also, younger individuals and individuals with higher degrees of investment literacy have a higher probability of buying on margin. these findings have implications for policymakers as well as those who provide financial advice. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation sanders, k. & olajide, o. (2025). consumer margin use: understanding the role of peer influence, investment literacy, and age. financial services review, 33(2), 55-73. introduction as of may 2024, debit balances in customers’ margin accounts for financial securities exceeded $809 million dollars (finra, 2024). these debit balances represent an opportunity, at a cost, to increase buying power in an effort to improve portfolio returns. buying on margin allows investors to borrow funds from a broker to increase their purchasing power and opens up a wider range of investment strategies. when investors make the decision to buy on margin, they use their portfolios as collateral, bringing the potential for greater investment volatility, margin calls, and even a possible forced liquidation of the investor’s position. therefore, an investor deciding to borrow on margin must not only consider the terms of the loan, but also the stability of the portfolio, market prospects, and accessibility to other funds in the event of an economic downturn. 1 corresponding author (kaplan.sanders@utahtech.edu). utah tech university, st. george, utah, usa. 2 texas tech university, lubbock, texas, usa. the decision to buy on margin is similar to other consumer decisions. juster and shay (1964) reveal why individuals might make purchases on margin when they compare individual and corporate investments in debt. much like a corporation that weighs alternatives when considering taking on a loan to improve profits, an individual will borrow on margin only when they perceive that the financial and psychological benefits outweigh the financial and psychological costs associated with the debt. in this context, if an individual expects a net increase in utility, considering all costs, they are likely to choose to buy on margin. the majority of the literature on margin has not focused on individual investors but has instead concentrated on the effects that margin requirements and margin purchases have on the market. for example, as margin requirements decrease, the amount of margin available to investors increases, resulting in higher price https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(2) 56 volatility and trading volume (zhang & li, 2013). similar results are found in japanese stock markets (hardouvelis & peristiani, 1992). in addition, buying on margin may lead to speculative stock bubbles when the market prohibits short selling (ackert et al., 2006). observing individual investment behavior using simulated data, ayres and nalebuff (2010) argue that younger investors who face liquidity constraints may only be able to have a diversified portfolio of stocks by borrowing on margin. an investor’s decision to purchase securities on margin may be based on their own human capital or by relying on the decisions and behavior of other investors. the use of an investor’s own human capital to make margin purchases can be assessed by measuring their investment knowledge. as an alternative, investors could choose to rely on the actions of other investors (like their peers or other individuals in their social network), especially when faced with uncertain situations (keynes, 1937), because they believe that these peer investors must be right (xu, 2023). this study seeks to build on existing literature by observing margin use from a household perspective. the 2021 wave of the national financial capability study (nfcs) state-bystate and investor surveys permit a deeper view of margin behavior. this study performs a crosssectional analysis of investors who make purchases on margin. specifically, this study examines the relation of peer or social influence (and possible herding behavior) to the decision to purchase securities on margin while also exploring how age and investment literacy are related. findings from this study would be beneficial to individual or household investors and financial professionals working with clients. for individual and household investors, this study creates and enhances awareness of an inherent peer bias driving investment decisions. this awareness could possibly lead to better investment outcomes and holistic financial health, especially where peer influence is incongruent with an individual’s risk tolerance and investment goals. financial professionals can identify inherent peer biases with their clients, address these biases, and offer financial advice and education tailored to their clients, given these biases. conceptual framework buying on margin involves the purchase of assets using one’s portfolio as collateral. therefore, while the decision to buy on margin is primarily viewed as an investment decision, it includes a debt decision as well. the following review will focus on peer influence, financial literacy, and age from both an investment and a debt perspective. while previous studies have focused primarily on financial literacy, this study will focus on investment literacy as a domain-specific form of financial literacy. buying on margin as an investment decision for many, peer and social influence play a major role in their investment decisions from financial market participation to investment selection and allocation. several studies indicate the peer effect on financial market participation. one such study is that of nguyen and nguyen (2020) which examined how the interaction of financial literacy and peer effect indicators influence participation in financial markets. the study found peer effect and perceived financial literacy to significantly influence the respondents’ participation in financial markets (similar results focusing on stock market participation were found by hong et al., 2004). the peer effect or social influence observed might be explained indirectly through risk tolerance (mylondis & oikonomou, 2021). this influence of risk tolerance on the relationship between peer influence and financial market participation is supported by the findings of frydman (2015) which showed a positive relationship between peer influence and risky asset allocations. focusing on workforce peer effects, gerrans et al. (2018) explored the role of workforce peer influence on investment strategies. gerrans et al. (2018) showed that workforce peers positively influence a change in investment strategies, particularly for peers of the same gender. furthermore, heimer (2016) discovered a link between peer influence and the disposition effect. the disposition effect explains the tendency to hold on to losing assets for too long and sell winning assets quickly. based on the premise of sanders & olajide 57 a positive association between peer influence and increased trading, heimer (2016) investigated the relationship between these two phenomena. the study found a positive association between peer influence and disposition effect. with the growth of social media, peer and social influence is even more evident in consumer behavior and decision making today. social media creates an avenue for individuals across all age groups to engage in learning, collaboration, and the exchange of ideas. they are able to do this on a wide variety of subject areas including financial topics like investing and managing debt which in turn shapes financial decision making (cao et al., 2020). these social media platforms offer information from various sources ranging from peers to financial experts and financial influencers (place, 2022). also, because social media allows its users to publicize their achievements and lifestyles, it magnifies the impact of peer and social influence by creating a perception of reality that individuals may feel pressured to follow (de veirman et al., 2017). on the one hand, the use of social media has been associated with positive financial outcomes and better financial decision-making (cao et al., 2020). however, it could raise concerns about the validity and efficacy of such information, as well as the possibility of widespread misinformation (corbin, 2023; place, 2022). financial literacy has been linked to household investment decisions (lusardi and mitchel, 2014). for example, financially literate individuals invest more in stocks than less financially literate individuals (van rooij et al., 2011). additionally, higher degrees of financial literacy provide opportunities that would have otherwise been unavailable (huston, 2010). furthermore, individuals whose subjective financial literacy is greater than their objective financial literacy invest more in risky assets like stocks (verma, 2017). another factor that influences investment decisions is age. charles et al. (2013) investigated whether age affected investment decisions and the behavior of investors. the study found that age is related to investment behavior and found that compared to older investors, younger investors invest more in the equity market and were more likely to buy on margin. similarly, while examining the investment decisions of older investors, korniotis and kumar (2011) found that older investors' portfolios reflected greater investment knowledge as they diversified more, held less risky portfolios, traded less frequently, and were less susceptible to disposition effects. however, these older investors were found to have worse investment skills due to cognitive decline. this lower investment skill was observed based on the older investors' lower return and portfolio performance on a risk-adjusted basis. shivapour et al. (2012) investigated the investment motivations among older and younger investors and opined that older investors were more worried about monetary loss, while younger investors were more motivated by financial gains. given the review of the literature, the influence of peers and social networks in financial decision-making, including investment decisions, is significant. however, problems might arise when these investors follow the advice of their peers blindly without critically considering the impact of such decisions on their personal finances, either because of a lack of knowledge or an inability to apply such knowledge. this behavior of blindly following peers is known as herding. khalid (2020) examined the mediating role of financial self-efficacy in explaining the relationship between herding and investment behavior and indicated that financial self-efficacy mediated a negative association between herding and investment behavior. buying on margin as a debt decision peer and social influence are also related to choices around debt. research using the financial socialization framework has shown that family, peers, schools, and mass media are important socialization agents that influence how individuals acquire and shape financial knowledge and their attitudes toward their personal finances, including debt (gutter et al., 2010; lebaron-black, et al., 2023; supinah et al., 2016). throughout an individual’s lifecycle, the influence of these socialization agents might change. for instance, churchill and moschis (1979) showed that as adolescents become young adults, communication about consumption with financial services review, 33(2) 58 peers increases, while communication with parents regarding consumption declines, indicating peer influence on financial knowledge, attitudes, and behaviors (including debt decisions) supersedes parental influence (bakir et al., 2006). turning to the relationship between peer and social influence on debt behavior, the literature is mixed. for example, georgarakos et al., (2014) investigated the effect of social influence on the likelihood of holding and taking on various forms of debt, and the size of the loans, based on the perceived income of their peers. the study revealed that individuals who perceived their income to be less than their peers were more likely to borrow both secured and unsecured loans and were more likely to have sizeable loans. similar results were found by berlemann and salland (2016), who also showed that these individuals were susceptible to peer effects and made more use of overdrafts. focusing on repayment behavior, breza (2010) examined the influence of peer repayment on the repayment decision of individuals and found that borrowers were more likely to repay their loans if their peers moved from being in default to making full repayment. however, jamilakhon et al. (2020), while exploring the association between financial education, debt attitude, peer influence, and power prestige, discovered no significant influence of peers on debt behavior. similar results were found by dusia et al. (2023). age and financial literacy have also been found to be associated with debt decisions. for instance, those with lower levels of financial literacy have been found to make costly debt decisions more frequently (chatterjee, 2013) and carry higher debt balances (brown & graf, 2013; gathergood & disney, 2011; lusardi & tufano, 2015). concerning age, del rio and young (2006), while focusing on unsecured debt, discovered that younger individuals aged 20 to 30 years old were the most likely to hold unsecured debt. similar results were found by eberhardt et al. (2018). interestingly, agarwal (2007) found that age had a “u-shape” with the cost of borrowing when analyzing how age influences different debt decisions like auto loans, home equity loans, home equity lines of credit, credit cards, and mortgages. this means that middle-aged adults borrowed at lower costs (fees and interest rates) than younger and older adults. herding herding theory as explained by keynes (1937) forms the theoretical basis for this study. herding, an integral part of behavioral finance, is a phenomenon that explains the tendency of individuals or investors to make decisions based on the actions of others. (keynes, 1937). according to keynes (1937), these individuals or investors mimic the actions and behaviors of others when faced with volatile and uncertain situations, because they believe that these other investors know better. pompian (2012) explains that herding behavior could be caused by regret aversion. according to pompian (2012), regret aversion is the tendency of investors to make decisions for fear of making a mistake or fear of missing out on a great deal or investment. therefore, they follow the crowd and actions of other investors to limit the likelihood of future regret, believing that the majority must be right (xu, 2023). trust heuristic could be another source of herding (trehan & sinha, 2019). when making investment decisions, people tend to rely on the advice and actions of influential figures, family, their social or religious groups, and their peers. as a result of this reliance, the uncertainty they feel about the situation or decision is reduced and decision making is faster. however, problems arise when these decisions made out of herding are suboptimal for these investors, considering their risk portfolio and other circumstances unique to them. buying on margin is a unique position as it involves the purchase of securities by borrowing from brokerage firms, with the hope of making a profit when the returns from the investment exceed the borrowing costs. because of the investment and debt position that buying on margin is, while it has the potential to increase returns, it brings significant risks. therefore, it is important that the investors are fully aware of the risks involved. when individuals herd, the uncertainty investors feel about their investing decision or position is reduced, even though the inherent risk of the position is not reduced (xu, 2023). so, investors who invest because their sanders & olajide 59 peers are investing (i.e. investors susceptible to herding) are more likely to buy on margin, especially if their peers are holding that position, because it gives them a false sense of certainty about buying on margin even though buying on margin increases the potential volatility of their position. younger individuals may be more likely to make purchases on margin. as previously discussed, buying on margin is very risky; therefore, individuals with higher risk tolerance are more likely to engage in margin purchases. based on the life-cycle hypothesis, age is negatively associated with risk tolerance, as younger individuals tend to have higher risk tolerance because of their extended time horizon. having more time to experience investment volatility allows them to recover from possible losses from riskier investment choices (mylondis & oikonomou, 2021; yao et al., 2011). also, compared to older individuals, younger individuals are more susceptible to peer and social influence (carolan, 2018; khan et al., 2016) and peer influence is positively associated with risk tolerance (frydman, 2015; gardner & steinberg, 2005). therefore, it is expected that younger investors will be more likely to buy on margin. given the herding theory and the review of previous literature that has shown the impact of peer influence, investment literacy, and age on decision making, this study hypothesizes the following: h1: investors susceptible to the herding effect (i.e., those who invest because their peers are also investing) are more likely to buy on margin. h2: investment literacy significantly influences margin purchase decisions. h3: age negatively relates to margin purchase. methodology data/sample this study uses combined data from the 2021 wave of the national financial capability study (nfcs) state-by-state and investor surveys. the investor survey is a follow-up survey from the more broad-based nfcs state-by-state survey and was conducted to take a deeper look at factors associated with investor decisions. both surveys have been commissioned by the finra investor education foundation and were conducted by applied research and consulting llc (arc). the investor survey was completed by 2,824 individuals who hold investments outside of retirement accounts. responses from both surveys were combined to allow for a more robust set of variables in the model. survey weights, which are provided based on data from the american community survey, are used to make the sample representative of the united states population. responses such as “don’t know” and “prefer not to say” are omitted from the sample in most instances. variables the dependent variable was measured based on responses to the following questions: (a) “do any of your investment accounts allow you to make purchases on margin?” and (b) “have you made any securities purchases on margin?” responses were organized into two outcomes. where respondents stated that they do not have or do not know if they have investment accounts that allow margin purchases (n = 1,736), or that they did not make or do not know if they made purchases on margin (n = 408), responses were coded as “no, i have not made any securities purchases on margin.” respondents who affirmed that they had made purchases on margin were coded as “yes, i have made securities purchases on margin” (n = 192). the key explanatory variable in this study is the influence of peers on investment behavior. respondents ranked how well the following statement describes why they invest on a scale of 1 (“does not describe at all”) to 3 (“describes very well”): “my peers are doing it/social activity/connecting with others.” peer influence enters the model as a series of dummy variables with 1 as the reference category. financial literacy pertaining to investmentspecific knowledge (hereafter, investment literacy score) is measured by adding correct responses to 11 items (see appendix for complete language). these items cover topics such as attributes of stock/bond ownership, the risk/return relationship, active/passive financial services review, 33(2) 60 management, margin, short selling, and options. financial literacy is included in the model as a continuous variable. for the variable on expected relative portfolio performance, respondents are asked the following question: “over the next 12 months, how well do you expect your portfolio of investments to perform?” responses range from 1 (“worse than the market as a whole”) to 3 (“better than the market as a whole”). expected relative portfolio performance enters the model as a series of dummy variables with 1 as the reference category. to represent the nature of margin being at least partially a debt decision, and to explore how an individual’s perception of their debt might influence ongoing debt decisions, overindebtedness is included in the model for margin use. responses to the statement “i have too much debt right now” range from 1 (“strongly disagree”) to 7 (“strongly agree”) with 1 as the reference category. other explanatory variables include willingness to take risks (on a scale of 1 to 5 with 5/5 being “very willing” and 1/5 as the reference category) and non-retirement account value (including 10 categories ranging from “less than $2,000” to “$1,000,000 or more” with “less than $2,000” as the reference category). gender, age, race/ethnicity, and marital status are also included in the model. gender is stated as a dichotomous variable with male as the reference category. age enters the model as a continuous variable ranging from 18 to 92. race/ethnicity is reported as “white non-hispanic,” “black nonhispanic,” “hispanic (alone or in combination),” “asian/pacific islander non-hispanic,” and “other non-hispanic (american indian, other, 2+ ethnicities)” and “white non-hispanic” is the reference category. finally, marital status is organized into the following four categories: “single,” “married,” “divorced/separated,” and “widowed” with “single” being the reference category. data analysis this paper estimates the following probit model: yi * = β0 + β xi + ε (1) yi = 1 if yi * > 0 (buy on margin) (2) yi = 0 if yi * ≤ (no buy on margin) (3) where yi* is a latent measure of the decision of an individual i to make purchases on margin. yi is the observed dependent variable (the decision to make purchases on margin) of an individual i. 𝛽0 is the intercept, while 𝛽 is a vector of coefficients showing the association of the independent variables with the latent variable. xi is a matrix that consists of predictor variables, including peer influence, investment literacy score, expected relative portfolio performance, having too much debt, willingness to take risk, value of non-retirement accounts, gender, age, race, and marital status. 𝜀 is the error term, which is assumed to follow a normal distribution. sanders & olajide 61 table 1. descriptive statistics buy on margin? overall sample yes no mean / (std. err.) mean / (std. err.) mean / (std. err.) buy on margin yes 0.094 (0.008) no 0.906 (0.008) why invest: peer influence not at all 0.760 0.292 0.809 (0.011) (0.040) (0.010) somewhat 0.161 0.369 0.139 (0.009) (0.043) (0.009) very well 0.079 0.339 0.052 (0.007) (0.039) (0.006) investment literacy score 5.386 4.992 5.375 (0.059) (0.233) (0.065) expected relative portfolio return worse than market 0.045 0.044 0.045 (0.005) (0.015) (0.005) same as market 0.672 0.426 0.698 (0.012) (0.044) (0.012) better than market 0.282 0.530 0.257 (0.011) (0.044) (0.011) i have too much debt right now 1 (strongly disagree) 0.562 0.312 0.588 (0.012) (0.039) (0.013) 2 0.109 0.088 0.111 (0.008) (0.031) (0.008) 3 0.065 0.072 0.064 (0.006) (0.023) (0.007) 4 0.093 0.125 0.090 (0.008) (0.033) (0.008) 5 0.068 0.057 0.069 (0.006) (0.018) (0.007) 6 0.029 0.074 0.024 (0.004) (0.023) (0.004) 7 (strongly agree) 0.074 0.273 0.054 (0.006) (0.037) (0.006) willingness to take risk 1 (not at all willing) 0.070 0.008 0.076 (0.007) (0.008) (0.007) 2 0.156 0.013 0.171 (0.009) (0.009) (0.010) 3 0.277 0.135 0.291 (0.011) (0.033) (0.011) 4 0.362 0.376 0.361 (0.012) (0.042) (0.012) 5 (very willing) 0.135 0.468 0.101 (0.008) (0.043) (0.008) financial services review, 33(2) 62 non-retirement account value < $2,000 0.065 0.014 0.070 (0.007) (0.011) (0.007) $2,000-$5,000 0.052 0.045 0.053 (0.006) (0.021) (0.006) $5,000-$10,000 0.047 0.089 0.042 (0.006) (0.033) (0.005) $10,000-$25,000 0.069 0.053 0.070 (0.006) (0.018) (0.007) $25,000-50,000 0.072 0.062 0.074 (0.006) (0.018) (0.007) $50,000-$100,000 0.132 0.215 0.124 (0.008) (0.033) (0.008) $100,000-$250,000 0.180 0.197 0.179 (0.009) (0.035) (0.010) $250,000-$500,000 0.151 0.115 0.154 (0.009) (0.025) (0.009) $500,000-$1,000,000 0.113 0.114 0.113 (0.007) (0.028) (0.008) $1,000,000 0.119 0.096 0.121 (0.008) (0.025) (0.008) gender male 0.637 0.750 0.626 (0.012) (0.038) (0.012) female 0.363 0.250 0.374 (0.012) (0.038) (0.012) age 56.907 43.084 58.340 (0.406) (1.255) (0.408) race/ethnicity white non-hispanic 0.724 0.615 0.735 (0.012) (0.046) (0.013) black non-hispanic 0.058 0.095 0.054 (0.006) (0.023) (0.006) hispanic (alone or in combination) 0.106 0.177 0.098 (0.010) (0.044) (0.010) asian/pacific islander nonhispanic 0.093 0.100 0.092 (0.008) (0.030) (0.008) other non-hispanic 0.020 0.013 0.021 (0.003) (0.008) (0.003) marital status single 0.195 0.246 0.189 (0.010) (0.039) (0.010) married 0.661 0.708 0.657 (0.012) (0.040) (0.012) divorced/separated 0.096 0.044 0.101 (0.007) (0.015) (0.008) widowed 0.048 0.002 0.053 (0.005) (0.002) (0.006) n 2,336 192 2,144 note: data from the 2021 nfcs investor and state-by-state surveys. financial services review, 33(2) 63 large differences can be seen between those who do and those who do not buy on margin. for example, fewer than 30% of those who have made purchases on margin report that peer influences do “not at all” describe why they invest compared to over 80% of those who have not made purchases on margin. conversely, over a third (33.9%) of respondents who have made margin purchases report that peer influences describe why they invest “very well,” compared to about 5% of those who do not make purchases on margin (see figure 1). figure 1. investment motivation: my peers are doing it/social activity/connecting with others those who make margin purchases appear to have a lower average investment literacy score (5.196/11) than those who do not make margin purchases (5.406/11). the distribution of scores can be seen in figure 2, showing the same trend of lower investment literacy scores among margin purchasers, but also showing that higher scores (10/11 and 11/11) were slightly more common among those who made purchases on margin than those who did not. a much larger percentage of respondents who have made purchases on margin (53.0%) expect their portfolios to outperform the market relative to those who have not made purchases on margin (25.7%). finally, those who buy on margin appear to report having too much debt more commonly than those who do not. figure 2. comparison of investment literacy score among those who did and did not buy on margin 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000 full sample did not buy on margin bought on margin describes me: not at all somewhat very well 0% 5% 10% 15% 20% 25% 0 1 2 3 4 5 6 7 8 9 10 11 bought on margin did not buy on margin full sample financial services review, 33(2) 64 the relationships between age and peer influence and age and buying on margin are found in figures 3 and 4, respectively. both demonstrate similar downward trends, with responses being more disparate from age to age among younger participants and more concentrated among those who are older (finally closing in on 0% or “not at all” by just over the age of 80) figure 3. investing due to peer influence by age figure 4. percentage of respondents who made purchases on margin by age financial services review, 33(2) 65 probit model results marginal effects estimated from a probit regression model can be found in table 2. among the variables included in the model, peer influence appears to have one of the most substantive associations with making margin purchases. relative to those who considered peer influence to “not at all” describe why they invest, those who stated “somewhat” and “very well” had a significantly higher probability (0.096 and 0.163, respectively) of buying on margin. especially for those who selected “7 strongly agree” (0.136), having too much debt was associated positively with margin purchase, as was non-retirement account value. willingness to take risk was positively associated with buying on margin, though only at the highest selfreported levels of risk tolerance (4/5 and 5/5). age was associated negatively with margin purchase (-0.002) at a significant level, as was being widowed (-0.057) when compared with being single. no significant results were found based on gender or race/ethnicity. table 2. results of probit model of margin purchasing (marginal effects) marg. eff./ (std. err.) why invest: peer influence not at all (ref.) somewhat 0.099 *** (0.020) very well 0.167 *** (0.032) investment literacy score 0.007 ** (0.003) expected relative portfolio return worse than market (ref.) same as market 0.000 (0.025) better than market 0.045 (0.026) i have too much debt right now 1 (strongly disagree) (ref.) 2 0.023 (0.025) 3 0.033 (0.027) 4 0.048 (0.025) 5 0.019 (0.023) 6 0.059 (0.037) 7 (strongly agree) 0.138 *** (0.034) willingness to take risk 1 (not at all willing) (ref.) 2 -0.010 (0.022) 3 0.042 financial services review, 33(2) 66 (0.024) 4 0.070 *** (0.023) 5 (very willing) 0.110 *** (0.028) non-retirement account value < $2,000 (ref.) $2,000-$5,000 0.024 (0.016) $5,000-$10,000 0.076 *** (0.028) $10,000-$25,000 0.063 *** (0.023) $25,000-50,000 0.062 *** (0.020) $50,000-$100,000 0.094 *** (0.019) $100,000-$250,000 0.107 *** (0.023) $250,000-$500,000 0.077 *** (0.020) $500,000-$1,000,000 0.119 *** (0.029) $1,000,000 0.128 *** (0.030) gender male (ref.) female -0.005 (0.015) age -0.002 *** (0.001) race/ethnicity white non-hispanic (ref.) black non-hispanic 0.021 (0.027) hispanic (alone or in combination) 0.040 (0.028) asian/pacific islander non-hispanic 0.009 (0.024) other non-hispanic 0.004 (0.041) marital status single (ref.) married 0.020 (0.015) divorced/separated 0.021 (0.028) widowed -0.056 * (0.023) note: data from the 2021 nfcs investor and state-by-state surveys. *** p < 0.001; ** p < 0.01 * p < 0.05 financial services review, 33(2) 67 discussion in several respects, the decision to make purchases on margin is not unlike the decision to make a real estate investment, where one borrows to purchase a property. the appropriateness of such an investment would be determined by the characteristics of the property as well as the terms of the loan. certainly, using debt to buy real estate (as opposed to an outright cash purchase) increases the risk and opportunity of owning the property, but it would be unreasonable to assume that an investor is savvy or ignorant without considering all of the details of an investment. in a similar vein, both a well-informed and an ignorant investor may use margin to pursue their goals, though likely with differing outcomes. making investment or debt decisions based on peer influence is likely to result in biased decisions where risks and opportunities are misunderstood and where the information received from peers is seen to be more reliable or comprehensive than it actually is. in agreement with observed trends in the descriptive statistics and correlation matrix (see appendix b), a strong positive association between peer influence and buying on margin was found when controlling for other key variables. as making purchases on margin is both an investment and a debt decision, each of these factors is likely at play. the substantive, positive association found in this study aligns with previous findings, which suggest that there are ties between reliance on peer influence (i.e., peer effect, social influence, herding) and risky investment behavior (frydman, 2015; mylonidis & oikonomou et al., 2021) as well as increased borrowing (berlemann & salland, 2016; georgarakos et al., 2014). this is concerning, as such reliance may be accompanied by inferior financial outcomes. in addition, confirming the findings xu (2023) which found that investors feel more certainty about their investment decisions when they imitate their peers, it is unsurprising to see that peer influence and confidence in one's ability to outperform the market are positively correlated (appendix b). there is a strong resemblance between figures 3 and 4, which demonstrate the relationship between margin use and age, and peer influence and age, respectively. marginal effects estimated from a probit model indicate a highly significant, negative relationship between margin use and age as well. at least visibly, the strongest correlations found were between peer influence and age, peer influence and buying on margin, and age and buying on margin (appendix b). though no significant results appeared in a separate model when including an interaction term between peer influence and age, these appear to be among the most important predictors of making margin purchases. in light of other findings, it is not surprising that these younger, less experienced, more heavily peer-influenced investors show a higher probability of borrowing on margin (charles et al., 2013). it is interesting to note the pattern of investment literacy among those who buy on margin relative to those who do not (see figure 2). though a small, positive linear association between investment literacy and margin use is found in the marginal effects (note the conflicting small, negative correlation in appendix b), a different pattern emerges in figure 2, showing that respondents who bought on margin had lower scores much more frequently, but also scored highly (10/11 and 11/11) slightly more frequently than those who did not. though further research is warranted, this suggests that this tool might be used most commonly by those with relatively low and high degrees of investment literacy. therefore, any policy enacted to protect investors from taking risks that are beyond their capacity or awareness should be carefully crafted so as not to create unreasonable barriers of access to margin use for those who are highly literate. additionally, this positive association between investment literacy and margin use deviates from the findings by kim et al. (2022), who found these variables to be negatively associated when using the 2018 wave of the nfcs investor survey. it is possible that the association between investment literacy and margin use is not static. for example, interest rates were substantially lower during the collection of the 2021 data than they were when the 2018 data were collected. this difference in interest rates (i.e., the cost to invest using margin) will impact the decision to make purchases on margin. another example of what could drive this difference in outcome is that financial services review, 33(2) 68 the outlook of the financial markets can vary dramatically. whereas the choice to buy on margin may seem appropriate to savvy investors during one period (e.g., high anticipation of market volatility), they may be more reluctant to use margin during other periods. also, the attractiveness of margin when compared with other debt options likely varies based on a host of factors (e.g., relative interest rates). further exploration into this interesting topic would be beneficial. a relatively strong, negative correlation between peer influence and investment literacy was found in this study (appendix b). it is possible that individuals with higher degrees of investment literacy are less likely to feel the need to rely on the investment advice of peers (or that those who are less interested in relying on their peers feel a greater need to develop their financial knowledge). an exploration of whether investment literacy protects against herding/peer influence would be beneficial. experimental studies exploring financial education and its impact on investment decisions in a controlled environment might yield valuable insights. additionally, related to the findings of khalid (2020) who found that financial self-efficacy mediated a negative association between herding and investment behavior, future studies could investigate the mediating role of financial selfefficacy on herding and margin use. a substantive positive association between feeling over-indebted and buying on margin was only found among those who ranked their sense of having too much debt very highly (7/7). this indicates a higher probability of entering into or maintaining margin agreements in environments where individuals feel their resource constraints most heavily, and further confirms the argument that margin purchase involves a debt component. it is worrisome that those who feel so strongly about their over-indebtedness are borrowing using a tool that allows for easy and immediate investment of the proceeds of the loan, as they are most likely to feel a need to make large investment returns, but lack the capacity to bear the losses that are often associated with leverage. the cross-sectional nature of these data represents a limitation, especially as peer influence was only recently added to the nfcs investor survey in 2021. future research could strengthen the present study by observing individual margin behavior across time. to the best of our knowledge, there are no datasets tracking individual margin behavior longitudinally. additionally, the manner in which these questions on margin use were asked may have been confusing to the respondents. for example, margin is typically required to be enabled on an account in order to place certain trades (e.g., selling options without owning the underlying security), but many of these transactions are completed without carrying a margin balance (using the cash balance within the account to complete the transaction). conclusions and implications given the paucity of literature examining the factors that influence individual decisions to buy on margin, this paper examines factors that are associated with individual decisions to buy stocks and other assets on margin – more specifically, peer influence, investment literacy, and age. a key finding in the current study is that those who invest due to peer influence are more likely to buy on margin, which might indicate herding behavior. within the context of previous research and herding theory, this study demonstrates the importance of improving awareness of the risks inherent in herding behavior, particularly when buying on margin. financial professionals are encouraged to consider how they might educate their clients to protect against the adverse effects of peer influence, especially discouraging the practice of doubling down on herd-influenced bets using margin. for example, financial planners and advisors might consider adding questions to gauge susceptibility to peer influence to their intake process, focusing on this issue with those who seem to be most at risk. additionally, for those more likely to display herding behavior, financial professionals might consider more frequent portfolio reviews, orienting clients towards a goal-focused rather than a peer-focused approach to portfolio management. furthermore, the relationship between investment literacy and buying on margin yielded an interesting result. while the probit analysis showed a positive association, the descriptive sanders & olajide 69 statistics showed a negative association. also, a negative association was found between investment literacy and peer influence, indicating that investors who have higher degrees of investment literacy are less likely to be susceptible to herding. this finding highlights the importance of improving investment literacy for investors through various investor education programs. another interesting finding is that respondents who rated themselves as being highly indebted were more likely to buy on margin. this finding further strengthens the argument that buying on margin is both a debt and an investment decision. these individuals could potentially be aided by policies that seek to provide just-in-time education about the risks and alternatives to margin use. also, financial professionals could educate their clients that buying on margin is both a debt and an investment decision, which could be a beneficial framing. other factors positively associated with margin purchase include age, risk tolerance, and nonretirement portfolio values. being widowed was negatively associated with buying on margin. references ackert, l. f., charupat, n., church, b. k., & deaves, r. (2006). margin, short selling, and lotteries in experimental asset markets. southern economic journal, 73(2), 419-436. agarwal, s., driscoll, j. c., gabaix, x., & laibson, d. (2007). the age of reason: financial decisions over the lifecycle with implications for regulation (nber working paper number 13191). http://www.nber.org/papers/w13191. ayres, i., & nalebuff, b. j. (2010). diversification across time. yale law & economics research paper no. 413. bakir, a., rose, g. m., & shoham, a. (2006). family communication patterns: mothers' and fathers' communication style and children's perceived influence in family decision making. journal of international consumer marketing, 19(2), 75-95. berlemann, m., & salland, j. (2016). the joneses’ income and debt market participation: empirical evidence from bank account data. economics letters, 142, 6-9. breza, e. (2010). peer pressure and loan repayment: evidence from a natural experiment. mit paper, 285, 19. brown, m., & graf, r. (2013). financial literacy, household investment and household debt: evidence from switzerland. swiss institute of banking and finance working paper no. 13/1). retrieved from https://uxtauri.unisg.ch/repec/usg/sfwpfi/wpf1301.pdf cao, y., gong, f., & zeng, t. (2020). antecedents and consequences of using social media for personal finance. journal of financial counseling and planning, 31(1), 162-176. carolan, b. v. (2018). friends’ plans for college and adolescents’ educational expectations: an instrumental variable approach. international journal of adolescence and youth, 23(3), 334-346. charles, m. a., & kasilingam, d. r. (2013). does the investor's age influence their investment behaviour? paradigm, 17(12), 11-24. chatterjee, s. (2013). borrowing decisions of credit constrained consumers and the role of financial literacy. economics bulletin, 33(1), 179-191. churchill jr, g. a., & moschis, g. p. (1979). television and interpersonal influences on adolescent consumer learning. journal of consumer research, 6(1), 23-35. corbin, k. (2023, march 2). finra warns brokerage firms on risks related to ‘finfluencers.’barron's. www.barrons.co m/advisor/articles/finra-finfluencersadvisors-33736bae. de veirman, m., cauberghe, v., & hudders, l. (2017). marketing through instagram influencers: the impact of number of followers and product divergence on brand attitude. international journal of advertising, 36(5), 798-828. del rio, a., & young, g. (2006). the determinants of unsecured borrowing: http://www.nber.org/papers/w13191 https://ux-tauri.unisg.ch/repec/usg/sfwpfi/wpf-1301.pdf https://ux-tauri.unisg.ch/repec/usg/sfwpfi/wpf-1301.pdf https://ux-tauri.unisg.ch/repec/usg/sfwpfi/wpf-1301.pdf http://www.barrons.com/advisor/articles/finra-finfluencers-advisors-33736bae http://www.barrons.com/advisor/articles/finra-finfluencers-advisors-33736bae http://www.barrons.com/advisor/articles/finra-finfluencers-advisors-33736bae financial services review, 33(2) 70 evidence from the bhps. applied financial economics, 16(15), 1119-1144. dusia, s. l., setyabudhi, k. n., & malelak, m. i. (2023). the effect of debt attitude and peer influence on debt behavior. international journal of financial, accounting, and management, 4(4), 463-479. eberhardt, w., bruine de bruin, w., & strough, j. (2018). age differences in financial decision making: the benefits of more experience and less negative emotions. journal of behavioral decision making, 32(1), 79-93. finra. (2024). margin statistics. retrieved from https://www.finra.org/rulesguidance/key-topics/marginaccounts/margin-statistics frydman, c. (2015). what drives peer effects in financial decision-making? neural and behavioral evidence. ssrn electronic journal, 10. https://www.aeaweb.org/conference/201 5/retrieve.php?pdfid=2266&tk=kfkr4rf d gardner, m., & steinberg, l. (2005). peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: an experimental study. developmental psychology, 41(4), 625. gathergood, j., & disney, r. f. (2011). financial literacy and indebtedness: new evidence for uk consumers. retrieved from https://papers.ssrn.com/sol3/papers.cfm? abstract_id=1851343. georgarakos, d., haliassos, m., & pasini, g. (2014). household debt and social interactions. the review of financial studies, 27(5), 1404-1433. gerrans, p., moulang, c., feng, j., & strydom, m. (2018). individual and peer effects in retirement savings investment choices. pacific-basin finance journal, 47, 150-165. gutter, m. s., garrison, s., & copur, z. (2010). social learning opportunities and the financial behaviors of college students. family and consumer sciences research journal, 38(4), 387-404. hardouvelis, g. a., & peristiani, s. (1992). margin requirements, speculative trading, and stock price fluctuations: the case of japan. the quarterly journal of economics, 107(4), 1333-1370. heimer, r. z. (2016). peer pressure: social interaction and the disposition effect. the review of financial studies, 29(11), 3177-3209. hong, h., kubik, j. d., & stein, j. c. (2004). social interaction and stock‐ market participation. the journal of finance, 59(1), 137-163. huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44(2), 296-316. jamilakhon, k., singh, j. s. k., subramaniam, a., & suppramaniam, s. (2020). determinants of debt behavior. a quantitative study among generation y individuals in tashkent, uzbekistan. test: eng. manag. j, 82(12), 891-904. juster, f. t., & shay, r. p. (1964). consumer knowledge of finance rates. in consumer sensitivity to finance rates: an empirical and analytical investigation (pp. 47-75). nber. keynes, j. m. (1937). the general theory of employment. the quarterly journal of economics, 51(2), 209-223. khalid, f. (2020). factor affecting investment behavior: mediating role of selfefficacy. journal of finance & economics research, 5(2), 112-125. khan, h. a., kamal, y., & saleem, s. (2016). peer influence on young adults’ products purchase decisions. business & economic review, 8(se), 83-92. kim, h., kim, k. t., & hanna, s. d. (2022). the effect of investment literacy on the likelihood of retail investor margin trading and having a margin call. finance research letters, 45, 102146. korniotis, g. m., & kumar, a. (2011). do older investors make better investment decisions? the review of economics and statistics, 93(1), 244265. https://www.finra.org/rules-guidance/key-topics/margin-accounts/margin-statistics https://www.finra.org/rules-guidance/key-topics/margin-accounts/margin-statistics https://www.finra.org/rules-guidance/key-topics/margin-accounts/margin-statistics https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1851343 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1851343 sanders & olajide 71 kumar, s., & goyal, n. (2015). behavioural biases in investment decision making– a systematic literature review. qualitative research in financial markets, 7(1), 88108. lebaron-black, a. b., kelley, h. h., jeffrey, h. e., jorgensen, b. l., & jensen, j. f. (2023). financial socialization agents and spending behavior of emerging adults: do parents, peers, employment, and media matter? journal of financial counseling and planning, 34(1), 6-19. lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. american economic journal: journal of economic literature, 52(1), 5-44. lusardi, a., & tufano, p. (2015). debt literacy, financial experiences, and over indebtedness. journal of pension economics & finance, 14(4), 332-368. mylonidis, n., & oikonomou, l. g. (2021). tracing the impact of peers on households’ economic behavior. economic change and restructuring, 54, 655-681. nguyen, t. a. n., & nguyen, k. m. (2020). role of financial literacy and peer effect in promotion of financial market participation: empirical evidence in vietnam. the journal of asian finance, econom03ics and business, 7(6), 1-8. place, n. (2022, august 18). social media ‘finfluencers’ are misleading investors: young investors are increasingly turning to tiktokers and youtubers for stock tips, often with bad results. financial planning. retrieved from www.financialplanning.com/news/beware-thefinfluencer-how-social-media-stars-aremisinforming-young-investors. shivapour, s. k., nguyen, c. m., cole, c. a., & denburg, n. l. (2012). effects of age, sex, and neuropsychological performance on financial decisionmaking. frontiers in neuroscience, 6, 82. supinah, r., japang, m., amin, h., & hwa, m. a. c. (2016). the role of financial socialization agents on young adults’ financial behaviours and attitudes. in wei international academic conference proceedings, rome, italy. trehan, b., & sinha, a. (2019). investigating investors’ herd behavior. amity business journal, 8(1), 31-36. van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101(2), 449-472. verma, r. (2017). role of self deception, overconfidence and financial aliteracy in household financial decision making. the journal of applied business and economics, 19(8), 94-109. xu, y. (2023, may). behavioral finance: an introduction of herd effect take the dotcom bubble in 2000s as an example. 8th international conference on financial innovation and economic development (icfied 2023) (pp. 216224). atlantis press. yao, r., sharpe, d. l., & wang, f. (2011). decomposing the age effect on risk tolerance. the journal of socioeconomics, 40(6), 879-887. zhang, t., & li, h. (2013). buying on margin, selling short in an agent-based market model. physica a: statistical mechanics and its applications, 392(18), 4075-4082. http://www.financial-planning.com/news/beware-the-finfluencer-how-social-media-stars-are-misinforming-young-investors http://www.financial-planning.com/news/beware-the-finfluencer-how-social-media-stars-are-misinforming-young-investors http://www.financial-planning.com/news/beware-the-finfluencer-how-social-media-stars-are-misinforming-young-investors http://www.financial-planning.com/news/beware-the-finfluencer-how-social-media-stars-are-misinforming-young-investors financial services review, 33(2) 72 appendix appendix a. investment literacy questions used in analysis (correct answers bolded) 1. if you buy a company’s stock… a. you own a part of the company b. you have lent money to the company c. you are liable for the company’s debts d. the company will return your original investment to you with interest e. don’t know f. prefer not to say 2. if you buy a company’s bond… a. you own a part of the company b. you have lent money to the company c. you are liable for the company’s debts d. you can vote on shareholder resolutions e. don’t know f. prefer not to say 3. if a company files for bankruptcy, which of the following securities is most at risk of becoming virtually worthless? a. the company’s preferred stock b. the company’s common stock c. the company’s bonds d. don’t know e. prefer not to say 4. in general, investments that are riskier tend to provide higher returns over time than investments with less risk. a. true b. false c. don’t know d. prefer not to say 5. the past performance of an investment is a good indicator of future results. a. true b. false c. don’t know d. prefer not to say 6. over the last 20 years in the us, the best average returns have been generated by: a. stocks b. bonds c. cds d. money market accounts e. precious metals f. don’t know g. prefer not to say 7. what is the main advantage that index funds have when compared to actively managed funds? a. index funds are generally less risky in the short b. index funds generally have lower fees and expenses c. index funds are generally less likely to decline in value d. don’t know e. prefer not to say sanders & olajide 73 8. which of the following best explains why many municipal bonds pay lower yields than other government bonds? a. municipal bonds are lower risk b. there is a greater demand for municipal bonds c. municipal bonds can be tax-free d. don’t know e. prefer not to say 9. you invest $500 to buy $1,000 worth of stock on margin. the value of the stock drops by 50%. you sell it. approximately how much of your original $500 investment are you left with in the end? a. $500 b. $250 c. $0 d. don’t know e. prefer not to say 10. which is the best definition of “selling short”? a. selling shares of a stock shortly after buying it b. selling shares of a stock before it has reached its peak c. selling shares of a stock at a loss d. selling borrowed shares of a stock e. don’t know f. prefer not to say 11. if you own a call option with a strike price of $50 on a security that is priced at $40, and the option is expiring today, which of the following is closest to the value of that option? a. $10 b. $0 c. -$10 d. don’t know e. prefer not to say appendix b. correlation matrix of key variables peer influence age expected performance investment literacy bought on margin peer influence 1 age -0.4532 1 expected performance 0.1312 -0.0925 1 investment literacy -0.2186 0.1842 -0.0266 1 bought on margin 0.3798 -0.2592 0.1732 -0.0361 1 note: data from the 2021 nfcs investor and state-by-state surveys. pii: s1057-0810(02)00096-3 ��� �������� ��� �������� ���� ��� �� ��������� �� ��� ������ �� ��� �� � ��� �� �� �� � ! �� ��� "� #����$ ���������� ����� �� ������� �� ��� ���� ��� �������� ����� ����� �� ����� ����������� � ����� �� !"#$� ��% %������� &' ��� �� ( &))$* ������� �� ������ �� � + ,��� &))$* �������� $.������ &))$ �������� ��� �������� ��� �������� ��� /� �0 �� $--�� ���� �����! ���� ����� � ��! ��� ���� ���� ��������� �� ������� ��� ������� ���� ���� ���� �� � �(��� �� ������� �� ������ 1� ��2� � � ��� ����� �� ������� ������������ �(���(���� ���� ���� �� ���� � � ������� ���� ��22���� �����2 � �� �� 33������������ �� ��������44 ���� ������� ���� ��� ����� ��� ������ �������� �� ������������! �� � 33��������� �� ��������44 ���� ������� ���� ��� �� �� ����� �( ������� �� ������ ��� ��������� ��� ���� ���� � ��� ������ �� ��� �� � � ��� ��� 5�6�� � ������� �2( �� ���� ��2 ������� �� ������ � &))$ � ����� #������ 7��� � �2��� ��� ���� &'� ����� � ��� �( �&* 1& )��� ���( �������� �� ����* 7�����������* #� �����* �������� ��� �������� ���* ������� �� ��� �� � ���� ������ 7� $---! 8��2 ��� ������ ��� � ���9"����9: � �( �������� ��� �������� ��� /;��� "�� $)<=$)&! $$> #���� $>>+0! ������� �� ��� �� �� �� ��� 33�������� ��� �������� ���44 � ���� ( 33� �44� ��� � � ���� �� ���� � ��� ���� ��� ����� ������������ �� ������2! ���� �����! � ���� ���� ����������� 7� � ���� ���� �� ������� ���� ���2 � 33�������� ?� ���2 8�������� /�?8�044 ���� ��� � ������� ���� ���� ���� �� � �(��� �� �� ������ ���� ����2� �� ��� ������ ( �� ���� � �� ������� �� ����� ��� ���� �� ��� 33����� 2����44� �� � 2�� ���� ����� 2���� ���� ��� � � �������� � �� �� ��� ������ ���� �� ���� �������2 ��� �6� �����2 ������� �� ������ � ��������� ��� ���� �(� ��� ��� 2��� � �(��� �� ������� �� �����! ��� ���� �6� ���� ��� ������� ������������ ���� ��������( �� � ������2 ���� /����! ��� �������� ������ � �� ��� ������� ����� 0� ���� ����� � � �� �� �������� #� ����� %����� $) /&))$0 $-'9&)+ $)@'=)+$)a)$ab 9 ��� � ��� ����� � &))$ � ����� #������ 7��� � �2��� ��� ���� ;77� # $ ) @ ' = ) + $ ) / ) & 0 ) ) ) < = > 8� ��������2 ����� � �� �� �$=>)-=c>+=@$++* ��6� �$=>)-=c>+=@@$)� '*���� ���������( ��2� ��d� ������� /���� �� �� �0! � �����d� ������� /��"� #����0� $ �� �� �$=>)-=c>+=&+&'* ��6� �$=>)-=c>+=@@$)� ��������� ��� ������ ���� ���� ���� ���� ��� ����� �� ��� 33�����44 �� ������������ �� �� �� �� ���� ����� ��� � ����� ������ ��������� ��� ��������� ��� ���� �� ��� ��� ���� � ����� ��� �� ������� ��������� ���� � ��� ����2� ���� �� �6��� �! � �� ��� ������� ��� �����2 � ������ �� �����! �����2� ��� ����! ���������� �����! ��� ������ ��������� ��( �� ���� ���� � ���2 � �������� ���� ��� � �� �� ��� 33������� ���� ���������44� � ��������� �� �������� � �� ( ��� ��� ����� ��( �� ��� ��� �� ������� �� ���� ��� ������������� ?����� ! �� �� � �� ����� �� ���� ������ ��� ��� ���� ���2����� ( ���� ���� �� ����� ��� ������� ����� � 7� ��� ���� �� �� �� ��� ������! ���� ��� ���� �6� ���� ��� ���=������� ���������� ��� ��� ������� �� ����� �� � �� ��=�����2 � ����� �� ��6����� ���� ��� ��� �� ��� ��� 7� ��� ���� �� �����������! ��� ��� ���� ����� ��� ������� �� ����� � ���� �� �� �( � ����� ���2 �� ������ ���� �e��� ��� � ���� � �� �2��� ������� ������� 7� ��� ��� �� �� ���� ������� � �����2! �� � ���� ���� ��� �� �� �� ������� �� ����� �� ������� � ��� ��� ���� ��� ��( �� ����� �� � ��� ���! �������� � � ����� �� ��� �( ���� �������2 ��� ���� ���� ��������� ��� �� �� �( ���� � ����2 ���� �� ��� ��� � ��� �(! �� ������� �� ���� ( 2 �����2 ��� � ������ �� ���� ���� �� ������ f� ���� �� ���� �� � ��������� ��� ���� ���� ���� ��������� �� �� ���� � ������� � ��� /$--@�!�0 ��� � ��� ��� :���� /$--@0 ���� � ����� � ��� �� � ������2 ��������� ������� �2���� � ���� ! ��6������ �( :���� ��� � ��� /&)))0! .��� ��� /$--<0! g��� /$--c0! ��� #���� /$--$0 ���� ��� �(�� ��������� ��� ������� 1��� ��� ���! ��� ��������� ����2�2( ��� ��� 2����� ������ ��� ����� � ?����� ! ��� �����2� �� ��� � � ��� ���� �� �� � � ����� ���� ���� �! �� �� ��������� �� ��� �������� ���� �� ��� #������ & ��������� ��� � �������� �� ��� � � ��� ���� ���� ��� ������� �� ����� ������ ( ���� � ��� ���� ��� �����2�� #������ > �� ��(� ��� ������ �� ��� � � ��� ��� �� ��� �� �� ������ ( �� 2��������� � ��2 �� ����������� ����� #������ c ���� ��� �6� ���� ��� ������� �� ��� ��� �� ��� � �� �������� �� ��� ����=� � ���� ������� ���� (! #������ @ � ������ ���� ����2 ��������� � ��� ��������� �� ���������� ��� ��� � ���� ����� � �2�� ������ 7� ��� �� ��� ��� �2� ��� ( ���� ����� �� ����� ��� ������2! ���� �����! ��� ���� ���� ���������� ���� ��� �6����� ����� ��� � ��� .�� ������� ��� � � ����� ��� ���=������ ��� �� 33������244! 33���� �����44! ��� 33���� ����44! ��� ���� �������� ���� ���� ��� ���� �������� 33������� �� ����� ������ (44� 1��! � ������� ������������ ��( �2� ( ���� � �(��� �� ����� �������� ������� �� ����� ��� ������� ���� ���� ���� �� ��� ���� ������� �� ; ����� � �� �� �� ��� ��� ���� �� �������� &) ��� >& �� ��� :�����2 ��� �� $->> /h�#�8� >'' ��� '+! ��������� (0! � �� ����� �� ��� � ���9#���2� ���� ��� ������� �2�� ����� � �� ������� ��� :��� ?� ���2 ��� �� $-@< /$& h�#�8� $+c>0� :������ �� ������2 ����� 2���� �� ��� ������� ����� ! ��� � � ��� ���� � �������� ���� �������� ��� � ������� �� ������� �� ������ 7� ��� ����� ��� ����� �2� ���2 ��� ��� ��2 �� ������� ���� ������ ����2 ������� �� ����� � �� �� ��� �����6� �� ������� �4 � ����(� 7� � �� ���� �� � ���� ��� � �������� �5����� ��� ����� ��� ��� ������ �� ��� 8�������( %����������� ��� �� �� �� �� ��� �����2 � �������� $-+ '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� #������� (! ��� � � ������ ���� � ������� �� ���2 ������( � ��� �������� ( ��( ��2�2� �� ��( �������( ���� �� ������� �� ���� �! ��������� � ���� ������ ( �� � ������� �������(! ��� ���� ��� ���� � ���������� ��� �� ��� �����( ��� ��������� �� �������� ( ������������ � ��� ������� �(���� �� 2��� � � ����� ���������� � � � ���� �� ��2 ��� /$0 ��� ���� � %��� �� :�� � ��� ��� #�� ��� ( �� � ���� ( �� ��� ��* /&0 ��� �������� ( ����������� �� �� ������ ����* ��� />0 ��� �������� ( ����������� ���� �� � ���������� ( ��� � �� ������2 ��������( � ���� ����� �� ���� ����� �( ��� 8�������( %����������� ��� �� $-''� �� ��� �� � � � ��� �e�� ����� ��� �� ����� ������� ��� ���� ����� �������� ( ��� ��� ����� ��� ���� ����! ����! � ���=�� �����2 �( ����� ��� �������� �� �� ������� ��� ��� ��2�� ��� ��� �� ���2 ������( ��� �� ���� ���� ��� ���� � %��� �� :�� �! ��� �� ���� � ������� �� ����� ���( ���� � � �2� ���� �( ����� ����� ��� ���� � �������� ( ��� ���� ����� ���������! �� �� ��������� ��� �� ���� �������� ��� � � ����� �?8� �� ���� ��2 �� ( ������� � ������� 1��=������� ! ����� ��� � ������ ��� �� ����� � � ��� ���� ��� 1��=������� ���������� ���� �� � �� ���� ���� � ��� ���4� �����2� � � �� ( 2 �������� �� �� �� ������ $) (�� �! � �� ���� $@ (�� �! �� ���( �� ��� ������ �� � ���� $@i �� ��� ����� ������ ����� 2 ��� ������� �� ��� � �� ��� �6���� ���� �?8� ��( ��2�2� �� �� ���� � ���������� ���� � � ��������� � ���� ������ ( �� ������� ���������� /���� �� �� ������ � ��� �2�0 �� ��� (�� �� ( ������ ��� � ���� ( �� ���� ���� �� ��� ������ � � ��� �� ���� � ( �� ����� ���� � ����� ��� ������ ��� � ������ ����� ��� ������ � ��� ����� ���� � �� � ���� �� ������� ���� ���=���� ������� � �� /���� ����� ��� ���� ���� ���������0! ���� �� � �� ���� ���� � ���=���� � �� �� � ���� �� ������� ���� ���� ���� � ���� �� ��� ������ � �� ��� ���� ���� ����� ������ �������2 ��������! ��� ������� �� ����! ���( ���� ���� ���� �� � �2� ���� ���� ���� ����� ��� ���� ���� ���������� ;�� �� �� ��( ���� �� �� �������� ��� �� ����( �� ��������� ����� �� �� �� ��� ����2 ��( �� ��� ���� � ������ ( �(����� ���� � ����� ����� �� ��� ���� �� ��� �������� � ��� ���� ��(������ ��� �� ��� ����2����� �� ���� ������ � ���� ! ����� ����� �� ��� ���� ��� ���� � ������2 �(���� �� �6�� ����� �����2���� ���� ���! �� ���� �� �� $-&-� 7� ���� �����6�! �� ������� ���( �� ���� ����� ��( �2� ��� � ��� �2�� ��� � �� � � ������ �� ������� ������2 ���� ���� ������� �� ����� ���� ��2�� ��� ����� ��������� ����� ��� � $ /����� �� ��� ��� ���� :���� � ����������� /$---00 ����� ���� ��� ��j� � �������� �� ��� ��� �� �� �� �� ��� ������ ���� ��� ������� ����� � 7� � ��� ���� ��� ����� ��� ������2 ����� �� ��� � � ��� ����� ��� ��� ��� � 2� ( ��� ����� ���� � ����� ���� � ��� ���� �2� ������ #��� ����� ��� ���� ���� ��������� ���� � ���( ��� ������� �� � 5�6��� ��( �� ���� ����������! ����! ��� � ������ ����� �� ��� � � ��� ��������2 ����� ���������� ���� ������2� 1��� ��� ���! ��� ��� ���� ��� ����� �� � ����������� �� ���=���� ������������! ����2 ���� ���� �� �� �� � �� ������ ��� � �� �� �� ��� � $ ��� ����� ��� ������� ������������ ��( ������� �� �� ���2 ������( �� ���� ��! ���� � ��� ���� ��� � �� :��� � ��� ���! ����� ��� :��� ?� ���2 8�����( /:?80 ��� ����� �� � ��� ����� �� 33���������� � ��� �� ������244� :��� ?� ���2 8�������� �� � �2� ���� �( ��� ���� � %��� ��! ��� ���� �������� ��� �� � �2� ���� �( ������2 ����� ����� /��� ���� � %��� ��! ��� 8���� � � �� ��� 8� ���(! ��� ���� � .������ 7��� ���� 8� �� �����! ��� ����� ������2 ����������� �� ��� ���� �� �����=��� �� �� �����0� ���� ��� ���! �?8� ��� � �2� ( �� �� ��� �� ���� ���� ���� ��� ���=���� '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� $-�������� ���� ��� ���� � %��� �� �2� ���� ��� �?8! ��� ��� ���=���� �������� ��� �������� �� �� �2� ���� �( ��� #��� ����� �6����2� 8��������� /���� ����� �������� ���0 ��� ��� ����� ���� ���� ����������� � /���� ���� �������� ���0� ��� ��6� ��� ��� �� ��� � $ ���� ��� ���� ����� ��� ���� ���� ��������� ��( �� �������� ���� �������=������2 ��������� �� � ���� �� � ��2������� ��( ���� ���� � ��� �2�� ������ 7� �� ���� � ! ��� � � � ��� �?8� �� ������� �������=������2 ���� �� ��� � ���� ����� ���� � ����2 ��� ���� ���� ���������� ���� ���� � �� � ����� � �������� � ��2������� ( ������� ��� � � � �� �2� ���� �� ����! ����� � � �������=�����2 ������������� "��� ����� ���� �?8�! �� ���� ���� h���� ( �� ��� ?� ���2 8�������� � � � ���� �� ��� ���� ����� ��� ���� � ����2 �������� ���� ��� � $ 8���� 2���� �� ������2 ���� ���� ������� �� ����� ���� ��� �������� ��� �������� ��� :��� � ��� ��� ���� ��� ��� ?� ���2 ������( �� ���� � ��� �2� ��� ( ����� ��( :��� ?� ���2 8�������� /:?8�0 ��( ���� ������2 �������� ��� ��� ���� ��� ����� ���� ��2 �� ( �� ����� ���� � � 33� ��� �� ������244 �������� ?� ���2 8�������� /�?8�0 ��� :?8� ��( ���� ������2 �������� ��� �� �� �� ���=���� ��� ����� ���� ���� � �(��� �� ������� �� ����� ���� � %��� �� �2� ���� :?8�* ���� �2� ��� � ���� ���� �������= � ��� ���� � %��� �� �2� ���� �?8� �� �� �� :?8� :��� �������� ��� ��������� � �� ��� ���� �2� ��� �* ���� ����� �������� ��� ��������� � �� ��� #��� ����� �6����2� 8���������* ���� ���� �������� ��� ��������� � �� ����� ���� ���� ����������� � ;� ������ � ���� ����� ��������� :��� ?� ���2 8�������� � ���� ��� ����� ����� ��( ��� ��2�2� �� ���� � ����2 �� �� ��� ���� �����* �� ��( ���( ���� ���� ����� ��� ����� ���� �� ��� :?8� ��( �� ( ���� � ��� 2��� ����� ��� �������� ���� ����� �������� ?� ���2 8�������� ��( ���� ������2 �������� ��� �� �� �� ���� ����� �������� ��� ���� ��( ���� � ��� ��� ��� �� � �(��� �� �� �� ��� ���� ����� �� �� �� 2��� ����� ��� �������� ���� �= ���� ������� ���������* ���( ��( � �� ���� ����� ����� ����� ���� $-+-! �� ���� � :?8� � ���� ��� ����� ��( ���� � ��� �� �� ��� ���� ����� �� � ������ �6����! ��������2 ���� ��� ��� �( ��� ���� � %��� �� ;� ������ � ���� ���� ��������� :���� ��( �� ( ���� � ��� � ���� ��� ���� ���� �?8� ��( ���� ���� ��� ����� �� �� �� ���=���� ��� ����� ���� ��2�2� �� ���� ���� �2���( ��� ���� � ����2 ���������� k� ( ����� �� ��� ����������� ��� � ��2�2� �� ���� ���� �2���( ���������� &)) '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� � ����� �� ��� ��������� �� ���������� ���� �� ����� �������������� ����� ��������������� ����� 8���� 2���� ��� � 2� ( � ���� ���� ���� � �� ������ �2� ���� ��� � �� #��� ������� ������������ ��2�� �� � � ����� 2���� �� ��2� ���� ����2 � ����� ������������ �� �6��� �! ��� ���� ����� � � � � "(��� ��2�� ��������2 ���� ����� ���� ��� 8��� ���2����� ������� ���� �������� �����=� ����2! ����( �� ��� ��������2! ��� � ���� �6�������� /��� #����! $--$! �� $<0� k��� � �6� ����� �2� ��� ( ����� �� ��� �� �������= ���� �� ���� � ����2 ��� ��� �������2 ���l��� �� 2� �� :��� �� ��� ��� ��� �������� � ��� 8�� �� #����� �� $-+$ ��� ��� $-+' �������� �( ��� ���� � %��� �� �� � �� ������ ���������� ������2 �( :���� � � ���! 8����� � ��� ,�;� � 2�� /,������� m ��� �! &)))! �� $-0� ��� ���� ������ ������� ������������ �22 ������ ( ������2 ��� ���� � ��� ���� � �� ������ �� ���6 �� $--+ ���� ��� � 2� ����! 8���2 ���! ���2�� �� �� 2� ���� �� ���� ���� ������(! � ��� � � � ���! ��� � ���� ����� ������(! #� ����! #����! :� ��(� ���� � ���=��������� �� 2� �������� ��� �����2� �� ��� � �! ����� ��� ��2� 8���2 ��� ��� � ���� ��� �� ������� �� � ��� � 4� ���� ���� ��� ������ /�� � ���� ��� ���������� �� ���� �� 2� ! ��� 8� ��! &))$0� ��������� �� ��� � � ������ �� ��� �� ��� �������� ��� � ����� �� ����� 2���� �� ��� ������� ������ ��� ���� ��� � ���( ��2��� ��( ������� ������������! ���� � 2� ��� ��� ! ����� ��� �� �?8 ������ ��������� ( ���� ��� �����2�� :������ /&)))0 ��� ��� ���� ������ < ������ ���� ��� ���! � ���� &)) �����! ���� ( ��������( �����! 2��� �� ���� ������ �� :��� ?� ���2 8�������� ��� ������ �?8� �� � �� �� �6���� � ����� ���� ��2�� �� ����2 �����2� �� ��� ���! � ��2������� � �����2 �� ��� ��������� �� ������� �� ����� � �� ��� ���� 1��=���� ���� ����� ��� ��� ������ �� ������� ������ #�����( /&)))�0 ��� �� ���� ����2 ��� 33$) � 2��� h�#� �����44 � � 8�� �� #����� 8� �� ����� ��� � 2�� #��� �( .��� ����� l���� � �������� ( ���� ����� � ��l�� �� �� ; ������� �������� 8� �� ����� ��� 8����� k�� �������� 8� �� �����l� �������� ( � ���� �� � ���� �� ��� �� #� ��� ,�� �� /$& ,���! &)))0 ��� ����� ���� ���� ��j� ��= ��� � ��� �2� � �� ���� ����� ������2 � ���������� �� ���� ����� ���� ������� ���� � ����2 ���� ������ ����� ��� ��� ��� �� ���! #�����! .����! .", .� ���! � � ���! ���� ��( 7����������! � � "(���! � 2�� #��� �( .��� ����� ! 1������ .������� : ��� �! ��� �. ���� ������ k� ��� ���� ����! ����� ���� ����� �� �� ���=������ :��� ?� ���2 8�������� ��� ���( ������� ����� � ����� �������� � ��� �2� ��� ����� ������ ��( � 2� :��� ?� ���2 8�������� ���� ����� ����� � ��e�� �� �������� ��� �� ��2�2� �� ���� ����� � ����������! ���������� ������2! ��� ����� ����2������ �� �6��� �! �� � ����� �� 2� ! ��� ���������� ������2 �������� �� ,�;� � 2�� ��� ���� � ����� ���� ��� ����� ��� ������2 �������� �� 8���� �������� :���� �������! ���2 ����! ,����! ��� �� � /&))&! �� $$>0 ���� �� �6�� ��� ���������� �� ��� ����� ( �� ,�;� � 2�� ��� ��� �� 2� ���� 8���� �������� :���� 7��� ���� ��� � �� ����� 2�� ���� ���� ���� ����� ( ���������� ���� ��� ���� ( ���� �� �( ����� �����2 ���� ��� ���� ���� ��������! ���� ���� �� � ������ #� ���� /&)))0 ��� ��� ���� :��� k�� ��� ���� ���� ���� ���� ��������� ��� �2������ ���� ��2 � ������* :��� �� ��� ��� ������� �� �� ��� �� �����2 �� ����! �������� �! ��� �� � ���� ����! ��� ���� ���� 7��� ��� �� �� �������� ��������� (! #�����( /&)))�0 � ������ �� ��( ���������� �� '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� &)$ ���� ���� ����=2��� ���� ���� ���� � ������ 2 �� �( ��2� ( >@i �� $--+! ��� ���� ����� c>i �� ����� �� ��(�� ���� �� ���� ���� � ������� ����2 ��� ���� �� ������! ���� ���� ���� � �� ���� ����� ���� ������2� #���� �� � 7��� ���� 8�����( ��� ��� ��� ��� #���� �� � #����2� :��� �� �� ������2 � ������ �� ��� �� ���������� ������ �� ��2���� ��� ����� ( ��� �� ��� 7��� ���! ��� � ���� ���� � � ���� ���� ���� �����2� ����� �� ��� �� � ���� �� ��������� �� ��� 1������ h��� � ��� /&)))0 ������ ���� ������� k����� $--' ��� � �� &)))! ��� k���� �� �� ��� #��� ������ /k�#0 2 ����� >c ���� ���� � �� �� ��� ��� �� �� ��� ������ ( ����� 2���� �� ������� �� ��� � � �� � �� ( �6� ����� �( �������� �������� �� �� ���� �������� ��� ����� ���� �� ����������� ��� ������ �� ������� ��� ��� �� �� ���� ��� ��� �� �� ��� ���! �� �! ��� "������� /$-->0 �������2���� ��� ���� � %��� �� � ��2 �� $-+' ���� �6������ ��� ���� ����� ���� � ����2 ������ ��( �� � 2� ������2 � 2����������! 8����� �! ,�;� � 2��! ��� :���� � � ���� ���� ����� ����( �6� � �� ��� ����� � ��� ������� �� ��� �� �� ����� � �� �� ��� ����� �� ��� ���� � %��� �� ��������� ���� ����� =��� (��� ����( ������ ���� ����� �����4 ������ ��j�(�� ��2������� �������� ��� �� ����� �� ��� ���� �� � � ��2� :�� 2��� ��� � ��� /$--+0 � �� ��������� �� ����� ����( ��� ����� �������� �������� �� ��� ���� � %��� ��4� ������ �������� �� � �� :��� ?� ���2 8�������� �� ��2�2� �� ���������� ������2 �� ��2� ���� #������ &) �������� ���� #������ �( 8� �� /&))$0 ��� ,������� ��� ��� � /&)))0 �6������ ��� ��� (��� �� �6� � � ������ � ��� ��� ����� �� ��� 8���2 ��� �� 2� ��2�� �� �������� ������� �� ���� � 2� �����! ��� ���� ���� ���������! ��� � ��� �2� �� ������ :��� ������� ����� � �������� ����� � ��� �������! ���������2 ���� ��� �e���( �� ��� ����� � ���=��������� �� 2� � �� �� �� ��= �����2� � �� � ! ��� �������� �� ��� ��� ���2 ���� ��� � � �� � �������� ������� �� ��� ������� �� ����� ������ (� 8 �� (! ��� ������ ( ������� �� ��� ����� ��� ������� ��������� � � �� ��� � �� ������� ���������! ��� �( ��� �������! �� ������� � �� �� � ��� ��� �! �� ����� ����� �� ������� �����2 ��� ���� ��� ( �� � ��������� ��� ���� �� ����� ����� ��� ������� ����� ! ����� �� ��������� �� ����� �� ��� � ���!�� �� � ���������� ���������"� �� ��� ����#��������� �� ���������� ��� ��"�������� ����� ��� � �! � ��������� �� �������� �� �� �� ���� �� ������ ��� �� �������� �� ���2 � ������ ���� ������ �� ���� � � �6� ����� �� ��� 3+!+ 4� ����� 5��� ����6��7 �5� 8��� ��� �� � � � ��� �5� ����� ��� 9 �����6��� � % � ��2� $ ���� ������� ��������� �� ��� ������� ������������ ��� ����� ��� ����2�� ���� � ��� ���� ��� � �� .���2 ��! �� ����� ��� ��� ������� �� ���� � ��( �� ���� ���� �( ��� ���� �������� �� �� ���� � �(��� �� ��������� ���� � � ��������� ���� �� ���� � �(��� �� ������������� �� �6��������� ���� ����(! ��2� $ ������� �� ( �� �� �(��� �� ������������ ������ �������� /����� ��� �����! ���� ����� � ��! ��� ���� ���� ���������0! ��� �� �� � ��� � ( ��� ���� ( ���2����� � ��������� /��(����� �� �����! ���� ����� � ����������! ��� &)& '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� ��� ���� � ����20� � ����2� ���� ������������ ������ �������� ��� ��������� � � �������! ��� ��� �6 ����� �� ���� ���� ������ ��� ������� ����� ��( �� ������� �( �������2 �� ��� �� ���� /��������� ��� ����0 � ��� ��� /������������ ��� ����0� ;��� � �� ��2� $ � ��� ���� ��� ��2������ ������ ( ����� � �� �� ��� � �� �� 33n44 ����� � ���� � �� ���� � �(�� �� ����������� ��� �� ������ �� ���� � �� ���� � ��������� ��� ���2��� �� n4� ����� ���� �� ���� � �(��� �� � �� ���� �� �� ( ��� ��������! �� ������2 ���� ���� �2� ������ ��������� �� �6��� �! ����� ��� ����� ��� � ���� ��(���� �� ����� ��� ��� ���� ��( � ���������� ��� ���� ����� 7� ���� ���� ������! ��� ������������ ��� ���� ���� ��� ���������! ���� ��� �� ���� �� ��� ��� �6 �� � � � ������ 7����� ( ����� 2���� �� ����2 ��� � � �� �� ��� � ������ �� ��( �� �����2 n4� �� � �� ��� ���� �������� �� �������� ��� �����������! �� �� � ��� ���� �� ;��� : �� ��2� $� h��� ��� � �! � �(��� �� � �� ��( ���� � �(��� �� ���������� 1�� ��� � �� � ������ �� ����2 ����� ��� /������������0 � �� ���� /���������0 �� ��� ���� ��� ��� ������� ����� � � ���� ! �� �� � ���� ( ���� �������� �� ��2�� ��� ������ �� �( ������2 ���������! ����� ���� ��� � � ������������ ���� ������ �� ( ����� ��� ����� ���� �� ��2� ������� ��������� ������ ��������� ���� �� � ���� ����! ���� ��� � � ��� ��� �� ������������ ��2��� �� 2��� ��2� $� ��� �������� #� ���� � �� �6� '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� &)> ��( �� ��� �� ������ �� �?8�! �� ������������ ����� � �� ��� �� ��� ��� �6� k� ( ��� ��������� ����� �� ���� ���� ��� ����� �� ��� ������� ����� /;��� 80� ��� ��� �! �� ������� ���� ��������� �� ���� ��� ��� ������� �� ����� ������ ( �( �������� ���� ���� ������������ 3+�+ 4� �����7 � :�;�-�� ����7 7� � �������7 �5� 8��� ��� ���� ���� �j� ����2�� ���� ���� �� �� ��� ������� �� ���� � ��� ���� �������� ����(! �� ���� ������������ ��� ������ ��� 5�6��� ��( �� ���� ��� ( ���� ��� ��� �6� ��� ��� �� � ��� ��� :���� /$--@0 ���� � 2���! 33��������� � � ���� �� � ���� � ���� ������������44� � ��������� ��� ����! ����! � ������ �� ���� ��2 ����2�2( �� �6� �����2 ��� ������� ����� � 7� ��������� ���a� ������������ ������ ���� ����2�! ��� � �� �� ���� �� ����� ���� �� ���� ���� ���� ��� ������� ����� � ��������� (! ��� ��������� �� �������� � ��� �� ���( ����2 ����� �� ����� �� ������2 � ������� �� ���� ��� � ������� ������������ ��( ���� ���� �� ������ �������� ��� ����� �� ��� ����� ��� ����� ���! ��� ��������� ���� ��� ��� ����� ���� � ��� ���������� �� ����� ����� �� ������ 3+�+ ';�������7 8��� ��� ����� �� �� � �����*����� ���;� ��� ��������� ��� ���� �� � � ( ����� �� �6���� ������� ���� ( ��� ��� ��� � ���� �� � � ��� ���� �������2 �� ��� ������������ � ���� �� �� �� ��2� ���� � ������� �������� ���� ����� ��� ��� �� �� ����� ���( �������� ��� ������ �� �� ��� �e������ �� ��� ������� ���� �����2 ��� ����� ���2 ���� ������� �� ����� �����(� ���( ���� ��� ������ ��( �� 2��� ��� ���� � �� ���� ������ ���� �� ������� ( �� � ����� ���� ����� ���2 �� �����! ��2�! ���� ������ �� ��� �� �� � ��������� ���� �� � �� ����� �� �� �� ��� � �� �����! � ���� ���� ������! ���� ������ �� ������ ��� ������ ���� � �� ��� #������ � /$--+0 ����� ��� ��� ���� ����� ��� ��� � ���� ����� ��� ���� ����� ��� ��� �� �� �( ��� �����2 ���� ������ ��(���� ��� �� ��������� ������ ���( ���� ���� ��� � � ��� ����� �� ��� ���� ������ ������ �2( �� ����� (�� � ��� ������ ���� ������ ��(���� ( ��� ���� ��� �� � ���������� ����� �� ������� �� ������ ��� ��� �! ��� �� ��� �� ������� ���� ��������� ���� � ���� ��� �����! ����� �� ��� �� ��������� ��� � �� � ���� ( ������� ������� � ���������� �� ������ ���� ?����� ! ��� ����� � ���� ���� ��� �� ��� �� ���� ��������� ��� ��� �����! ��� ��� �����! ����� ��� � �� ����� � ��� ���� ����� � 2������ � �� ��� #������ � ���� �� ������ � ���� ������������(! � 2���2 ���� ��� ��������� �� ��� ������� ����� ���� ����� �� ����� �������� � ���������� ������ 7� �������� �� � � ��(��2 ��� 2��� � � � �� ���� ����� ���! ���� �������2 � ����� ��� �������� �� � ���� �� ��� ��������� ����2����� �� ��������� ������� ������������� ���� �� �� ���� � ( � �� �� ����� ��������� �� ����� ��� ���� ������ ������ �2( ��� ������ ���� �� �� �� �� ������ �� �6��� �! ����2 �� ��� ���� ���� ������ ��(���� ���! ����2� � �� � �� ���� � ������ �� ���� ��� ������� ������ � ���� ������� �� ��2 ����������� � �� ����� �� �� ��� ������� ������ �� :� ����2 ��� �����2 ������ �� ���� � � ����� �� � �� ��� ���� ���� ! �� ��� �� �� �� ���� ��� ������ ��� ���� � ����� �� ���� �� �� ��� ������� �������� �� ���� �� ��� ��� ����2 �� ��� ��� � �� ����� �� ��� ����� ���� ��������� � ������ ���� ��� /$-->0 ����� ���� �� $-<c ����� ��� <)i �� h�#� ������� ������! ��� ���� ��� &)c '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� ��� ���� �� ��� ���� >)i �� $--&� "��� ���������� �� �����4 ��� � �� ������� ������ ���� ���� ��� ������ � � ��� ���������� ��� 8����2� � ����� /&)))0 ��� �� ���� ��� � ����� �� � &@�&i �� ������� ������� ������ �� $-'@! ���( �� � �� ( -�-i �� $---� #��� ���� ��� ��� �������� /&)))0 ���� � ��� �� �� �����2 �� ����� ���5��� � �� �������� ( �����������! ��� ���� ���� ��� �������� ��� �� � ������� �����! ����� �����! ��� ����( �� ��� ������ ������ ����2� � �� ������� ������������ ���� �������� �� ��� � ���� � �( ���2�����2 ��� ��� �������2 ��� � ������ ������ ��� �� ��� ����� ���� ��������� ��������� �� �6��� �! ���( ���� ���� ���� ���=�� ���� ����� ���� ��� ���� ������2��� �����2 �6���� �� �� � ������ ( �� � ������� � �� ������� ( �� ���������2 �� ��� �� ��� ������� ������ � /:��� m ����� ! $--$0� 3+3+ ��8���7 ��� �� �� �� �5� 8��� ��� �� � � �� ����������� ��� ����������� ����2�� � �� ��� ������� �� ���� ���! ��� ������ � ����� ��6 �� ��� ������� �� ����� ������ ( �� ��2������� ( �������� k � � ������ ���� �� ���� ����� ��� ��� ���� ��� 2�� ���� ��� ���� �� ���� � ( � �� �� ���� ������� �� ���� ������ ������ �2( ��� � � ��� ����� #���� ������� �� ����� � � � 2� ( ���� ������= �����! ������� ������������ � � �� � �� �������� ����� �( � ������2 ������ ����� � ������ �� �� ���� ���� ��� ������������ ���� � �������� � ������ ��� ��� ��( ������ ���� ���� ������� �� ������� ���� ����� ���4 � ����� ���� �� �� ���� � ( � �� �� ��� ���� �� ��� ��(����� ��������� ;�(����� �� ����� ���� �� � � �������� � ����� ���� �� 2 ��� ( �������� �( ������� � ��� ���� ������ 5��� ��� � �������2 ��� � �� ��2 �� ������ ��� ���� ��� ���� �( ������� ����� ���2 ��� � ��� ���� ���� ������ ������ h�� �� � ��� ���� ��(����� ��� ������ ���e������* ��� ���� � %��� ��4� ��� ����( ����� ���� �� &)))! c)i �� ���=���� ��(����� �� � ���� � ��� ����� ( /���� � %��� �� #(����! �� � ���! �������� #� ����� �������0� �� ��� �� �! ������ �2( ��� � ���� �=������2 �� ��� 2� �� � � ����� ���� � � �������� �������� �� ��� � ��������� �� � �������� /:(� � m "��� � ! &))$0� ���� ���� ������ ��� ���� ��(����� �� ����� ���� � ��������( ��� ��� ���=����� ���� �� ����� �����! ���� ����� � ��! ���� ���� ���������! ��� f7#� ���� ���� �� ���� ����������� ���� ������ ��� ����� ���� ��������� �������� ��� � �� ��� ������� ��� � ���� � �� ���������������� 1�� ������ �2( ��� ��2������� ( ���������� ���� ������ ��(���� ��� �� ���� � �(��� �� ������������ ���� �e�� ������ �� � ���� ���� ������� ��� ��� �! �� ���2 � ����������� � �(�� �� ����������� ��( �� � �6���� ��� �� ����� ���� ���� ������ ��(���� ���� ���� ��� �� �� ��� ��� ��� �� ��� ����� ���� ��������� �� �������� ( ������������� � �� ��� #������ � /&))$0 � 2�� ���� ������� ������������ � � ��� �� � �� ��� �������� �� ���� ������2 ��� � ����� ��� ��������2 ������� �� ���������� �� ��� ������� ����� ! ���� ���� � �������� � � �� ��(����� � ����� � ��� ����� ���� ������� �� %��� � ����� � ������ ��� ��� ��� =���=������ �� �������� ���� �� �� �� � ��� ���� ���� ��� ���� ��� ��� �����! �� �� �� ���� ���������� ��� ����2����� ��� �� ;� ���������� � ������ �� � ��������� ������� � ���� �� ����! ���� �� ����� ����� ���� ������ ��� � ����2! ����� � �������! ��� ��� ����2����� �� � �������� ( ���� ��������� ������� � ��( �� �� �� ������ ���� 7� �� � �� ���� �����2 �� ���� ���� ������� ������������ ���� ��2�� �� ������� ���� ���� ����� �6����2�� �� ���� ��( �� �� ��� � 33 �e�����(44 ��������� ���� ����������� ��� '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� &)@ �� ��� ��� ��� ���� � ������� ( �� � ����� � ���������� �� ���� ������� �� ��� �� ��������� ��( ��������� ���� �6���� ��� k�� =���=������ ��� � �� ���� ���� ���� ���� �� � �� �� ������� ���� ��� � ����2 5�� � �� � 2������ �6����2��� #��� ����� � �� �� �� �� ����� ���� ���� ���� � ��� ���� � ���� �� ���� ���� ��= ���! &c ��� � ����2 �� ���� ����� /8� �(! $-->0� �������� �� ��� 2 ���� �� ��� � ������ �� ��� ��� ���� �� ��� ������ �� ����� � ����� �� �����2 �� ���� �� ���* ��� � 33�� ����44 ������� �� ������ � �� � �������� � ������! 33����44 �� ����� ������ �� ��� � ������� �� � ��� ���! 8���� �������� :��� ��� ��� ���� �� $--��� ������ ��� ����� �( $)ci! ��� � ������ � �� ���� ������2! � ���� �� ��! ��� �� �2�2�� ��� 5�� /1�� o� � �����! &)))0� ��� ��������� ��� ���� �� �� ������� �� ��� ��� � ����� ����2��� .� �������� ��� ����� ����� ������ ��� � �������� ��������� �� � ���� �� �� ����� � �� ��� ������������ ��� ����! � ����2� ����� ���� ����� ������������ ?����� ! ��� ��������� ��� ���� ��� � �� ( ���� �� ��� ����� ��� ��������� �� ��� ���� 3+<+ ������ ���7 �� � �������� �� �� �� ��7 ���� �� 8��� ��� ����� �� � ��������� �� �������� �� ������� ( ����� �� ��������2 ���� � ����2� � �� �� ���������� �?8�� ��� ����2� � �� ����� ��� � 2���������� ���� ���� � � ���� ���� �� � ������ ��� � ����� �� ���� ���� ���� � �������� � �� ��� � � ������ ���� ������� ������������� ���( ���� ���� �� � �������� ������e��� ��� �� �����2 �� ���2��� �� � �� ���( �� ���������! ��� j��� � ���2 � ���� �� ��� �� �! ������2 ��� �������� ������ � �� ������� ������������ � ��� ��������� �� ��� ���� �� ������ ���� � � �e�� �� �� � ����� ��� ��������� �������* ��� ���� �� � ���� ����! ��� ����� ������ ��( �� ������ �� � �������� ����! � ��������� �� �������� ��� � ������� ��� � �������� �� ������� �� ����� �� �� �� ���� ( ������2 ����� �� ������ #������� (! �� ��� � ���2���� ���� � ��� ��������� ������� � ( ���� � ��4 ����2����� �� ���� ������ ��� ��� �6���� ��� ����� ������ � � ������ �� � ��������� � ���� 7��� ������ � �������2 �� ������� ( �� ������2! ���� �����! ��� ���� ���� ��������� ���� � � �� �� ��� �( ������� �� ���2 ������( �������� ���� :�����2 �������� ��� ���� � ����� ���� ������ �� � �� �� � �� ������ ��� ��� �� �� �� ��� � ��� �������� �� #��� ����� �������� ��� ���� � ����� ���� ������ ����� �� ��� � ���� ��� �� ��� � ����! �� �� �� ������� ���� ������ ����� ��� � ���� ���������� ���( ���� � ��� �����2 �� ���������� ����� �� 7��� ���� �������� ��� ���� 2���� ���� ������ ����� � ��� �( ��� ����� �( ��� ���� ��� ����� �� ��� �����! ��� �������� ������� �� � ��� �� ����� �� ��(�� �� � ��� ��(���� �� � ����� 7� �� �� �� ��� ����2�����! ������2 �������� ��� ���� ���� �����! ����� �! ��� ����� �� ����� �� ���� � �e�����( ���! � ���� ���! ��� ���� ��� ��� ���� #��� ����� �������� ��� ���� ����2� ���� �� ��� �6���� �� �� ���������� ������2! �� ���=�����2! ��� ����� ����2�= ����� 7��� ���� �������� ��� ���� �� ����2� ��� �( ����� �� ( ���������2 �� ��(�� �� ����� ��� � ������2 ���e���� ��� ��� �� ���� ����� ������ 8 �� (! ��� ���� ������ ������ �2( ���� ������ �� ��� ��������( �� ����2 ������ ���� � � ������ �� ���( � ����� �������� ���� ��� /$--'0 ����� ���� ���� ������ � �������2 ����� �� � ����� ���� ������ ( ���������� �� ������ �2( �� � ����� �� ��� � ������� �� ������ &)< '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� $ %���������� ��� � � ��� ��� ��� ����� �� ��� ������� ������ ( � ����� ���������� ������� �� ����2��2 ��� ����2�2( �� �������2 ����� ��� ������� ����� � ��� �� ��� � ����� ���� � ������ � ����� �� ��� ��� �� � �� ����� 2���� �� ��� ������� ������ ���� �� �������� � ��������� ! �� ��=����� ��� ���� �� ��� �� ��� ��� ���� �� ��� � ��� ������2 ��� ������� ����� ��� �( ��� � � ����� �� ������������! ��� �������! �( ��� ��������� ���� � ������������ ��( ��� � ������ ���� ��� ���� �� �� �������� ���� ����� ��� ��� ��( ������� � �� 2���� � ����! ��� ��� ��� ������������� ��� ���� ������ ������ �2( � � �������2 � �������� ������e��� ��� ����������� �������2��� ���� (! ��� ��������� a�� ��=����� ��� ���� �� �� �������� ���� ����� ��� �� ���� � ��� �������� ������ � �� ��� ��� ������� �� ���2 ������(� &��������� � ��! ��! m #������ �! �� � /&))$0� ���� �� ������� ���� ����� ��� ��p & ����� =��>��7 ��� ����� �! �<! &'$9&-c� � ��! ��! m #������ �! �� � /$--+0� ��� ���� ( �� ������� ���� ���������� & ����� =��>��7 ��� ����� �! �!! $c<$9$c+@� ��� ���� :���� 4� ����������� /$---0� ?5� 8��� ��� � �����6��� �( ?5� ,����2��� 52=����� % � �������� ������2���! .8� ��� ���� :���� 4� ������������ ��� ���! f� ,�! �� �! ,�! m "�������! "� /$-->0� �6������ ���� ����� ���� � ����2 ��� �������� �� ���� ��� ��� ��� �� & ����� ' � �� � ��� =�������! 3<! $c>9$@+� :������! %� �� /&)))0� ?5� -��� �5� -�77��� /��� &+9>+0� h�#� :���� � :�� 2���! %�! m � ��� ! .� /$--+0� k� ��� ��� �� ��� ��� ������� �� ����� ��� ���� �6������� ���� ���� ����� ���� � ����2� �� ��� (��� �� #������ &) �������� ���� & ����� =��>��7 ��� ����� �! ��! cc'9c<@� :���� q�! m � ���! %� /&)))0� ����� �� h��� #��� � %��� ! 1,� ; ������=?� � :���! �� �� ��! m ����� ! �� f� /$--$0� k��=�� ���� ����� ���� �����! ������� ���� ����! ��� ������ �2� ������ & ����� =��>��7 ��� ����� �! !<! +&@9+c<� :(� �! %� ��! m "��� � ! ;� ,� /&))$0� %���� ���� �� ����� �� ���2(� � ���� �� � �������� ! � ��� ����! ��� ��6�� ���� ������� �������� & ����� 9���7����� �� ����� � �������! !�! $>>9$@<� 8� ��! g� �� /&))$0� 8���2 ���9� ��� � � � ��� �� 2� � 8�� ��2��2 �� �� � ������� ������2 ��� ���� ����� & ����� =��>��7 ��� ����� �! �<! $@@>9$@'$� 8����2� � ����� /&)))0� , � ����� ��� 5 ���� -��>� /�� $0� #����(! k����� $! #������ @� 8� �(! � /$-->0� :�2 � ��� �2� ���� ��� � ��� ������� =��> 9���7�����! >>9>'� .��� ���! ,� g� /$--<0� ����� ��� ����� �� ��� 8��� ��� ��������� ��( @���� ����� � �� �5� �� ��� ��� �� �� h��� #��� � %��� ! 1,� ; ������=?� � �������! �� ,�! ���2 ����! ��! ,����! ��! m �� �! � �� /&))&0� � ������ �� 8��� ��� ���>��� ��� ��������� �� /> � ���0� h��� #��� � %��� ! 1,� ; ������=?� � ���� � %��� �� #(���� /�� � ���0� �������� #� ����� �������� ����� ��� ������� 2� �������� ��� �������� ��� /$---0� #�� ,����2��� 52=����� % �� ;��� "�� $)<=$)&! $$> #���� $>>+� � ���9"����9: � �( ��� /$---0� ;��� "�� $)<=$)&! $$> #���� $>>+� #�� � �� ?� %� 8���� ���� %��� � 1�� c>c! $)<�� 8��2 ���! $�� #������� ,�������! ,�! m ��� �! ,� /&)))0� f� ���2 ��� �������� � ����� ������ �� ����� ���� ������� �� ���� ���2 ��� ����� �������� � �� ��� 8���2 ��� �� 2� � ?5� ����� ��� /�����! �<! $'9><� g���! � /$--c0� ����� ��� ��������� �� ��� ���>���� 1�� o� �� �� ��=?� � ���� ���! ,� ,� /$--'0� 1���� � ������ ��� ��(���� �(������ =������� /�����! ! $@9&@� � ���! %� 8� /$--@�0� � ��������� �� �������� �� ������� ���� ���������� ����� ��� 9���7�����! �3! &+<9 &-<� '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� &)' http://www.frbservices.org � ���! %� 8� /$--@�0� �������� ���������� ��� ��� ����2����� ��� �2� ����� �� ������� ������������� & ����� =��>��7 ��� ����� �! !#! c<$9c+$� � ���! %� 8�! m :����! q� /$--@0� � ��������� � ����� � �� ��� (���2 ��� ������� ���� ������� ?5� 7� -�� 8��� ��� ������( % �� �� ��� ������ ����� 8��� ��2�! �� ?� �� � :������� #���� ; ���� 1������ h��� � ��� /&)))0� #����� ��� � �����=��=���� ����� 4� ����� ��� �������.���> ��� -���8� ����7����� ����� �� 1�� o� � ����� /&)))0� ���� ������ ������ �� 5��� ����� 5���a� �� 8��! ��� ���( �� � &)! &)))! #������ 8� #��� ����! :�! m ��� ��������! .� /&)))0� � � ���e�� �� ��� ���� ( �� ������� ���� ���������� & ����� =��>��7 ��� ����� �! �3! $&c>9$&@$� #����! �� "� /$--$0� b ����� ���� 8��� ��� ���>��� ��� ����� ��� #�� ;�� ! 1� ���� ;�� �����2 8�����(� #� ���� ! :� /&)))0� 7��� ���� ���� ����� %=% =��>��7 & �����! &c9>+� #�����(! ;� /&)))�0� ����� -��>�� -�7 ����� /��� @$9@c0� h�#� :���� � #�����(! ;� /&)))�0� /�� 5��7 ��� 5����5 ��� �� ����� /��� @@9@-0� h�#� :���� � "��! 8� /&)))0� �� #� ��� ,�� �� � b5 � ��� 5 � �� /�� %$)0! ,��� $&! &)))� ���� ���! .� 8� /$-->0� 7� ��� ������2 ������ ( �� ��� ���p %����� � ���� ��� ���� � � ������� � �� � ����� ��� �� ��������� /������ ���� � %��� �� :��� �� #�� "����� &)+ '+�+ ,��-���� �+�+ � �� . ����� ��� ����� �� /����� !$ 0�$$!1 !#"2�$� the financial modernization act: new perspectives for the finance curriculum introduction the financial modernization act impact of the financial modernization act: industry reorganization along multifunctional lines benefits of a functional perspective in the post-financial modernization act environment properly characterizing the financial sector after the financial modernization act providing a flexible pedagogy for studying the financial structure explaining financial services in a value-added context defining new functions of the financial sector identifying common elements in producing various financial services conclusions references pii: s1057-0810(01)00077-4 hedonic investment douglas e. allen, elton g. mcgoun* department of management, bucknell university, lewisburg, pennsylvania, 17837 usa received 10 november 2000; received in revised form 25 july 2001; accepted 15 august 2001 abstract investing and consuming may not be so different as traditional economic theory has understood them. the consumer research literature has begun to view consumption not simply as rational decisionmaking, but as a more multisensory activity in which emotion and fantasy play important, if not essential, roles. this new perspective has been extended by holt (1995) in a matrix of metaphors in which consumption can be viewed as an interaction with objects and/or other persons as an end in itself and/or a means toward toward other ends. this paper theorizes how this matrix might apply to investment and uses a literary analysis of the best-sellingthe motley fool investment guide to examine whether or not our knowledge of consumers might in this way inform our understanding of investors. © 2001 elsevier science inc. all rights reserved. 1. introduction traditional economic analysis tells us that there is “investment” and there is “consumption” and that in the familiar economic identity, output must be either invested or consumed but not both. but are they really that different? consumption is coming more and more to be seen as a pleasureable (or “hedonic”) activity in itself and not simply as a way to decide which goods and services will have the greatest utility for us. is investment an equally enjoyable activity and not simply a way to decide which investment will earn us the highest return? what is investment? the verb “invest” has nine definitions in (the oxford english dictionary, 1989): 1) to clothe, robe, or envelop (a person) in or with a garment or article of clothing; to dress or adorn; 2) to cover or surround as with a garment; 3) to clothe or endue with attributes, * corresponding author. tel.:�1-570-577-3732; fax:�1-570-577-1338. e-mail address: mcgoun@bucknell.edu (e.g. mcgoun). financial services review 9 (2000) 389–403 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00077-4 qualities, or a character; 4) to clothe with or in the insignia of an office; hence, with the dignity itself; to install in an office or rank with the customary rites or ceremonies; 5) to establish (a person) in the possession of any office, position, property, and so forth; to endow or furnish with power, authority, or privilege; 6) to settle, secure, or vest (a right or power) in (a person); 7) to enclose or hem in with a hostile force so as to shut off approach or escape; to lay siege to; to besiege, beleaguer; to attack; 8) to occupy or engage, to absorb; 9) to employ (money) in the purchase of anything from which interest or profit is expected; now especially in the purchase of property, stocks, shares, and so forth, in order to hold these for the sake of the interest, dividends, or profits accruing from them.“ (oed, 1989) the earliest use of “ investment” in 1583 by stubbes carried the first meaning: “he . . . could haue inuested them in silks, veluets [and so forth].” (ibid.) from an economic standpoint, of course, it is the last meaning that is most important, and an economist might add that the purpose of rational investment is to maximize risk-adjusted returns. note that all of the meanings of “ invest” refer to something that is desirable, or at least neutral. this contrasts with the word “consumption,” the alternative of “ investment,” as in the fundamental economic identity that output (y) must equal consumption (c) plus investment (i). all of the meanings of “consumption” in the oxford english dictionary (1989) use more or less negative language: 1) the action or fact of consuming or destroying; destruction; 2) the dissipation of moisture by evaporation; 3) decay, wasting away, or wearing out; waste: 4) wasting of the body by disease; a wasting disease; 5) wasteful expenditure; waste; 6) the using up of material, the use of anything as food, or for the support of any process, 7) the destructive employment or utilization of the products of industry; the amount of industrial products consumed; 8) exhaustion of a right of action; 9) the test of a motor vehicle with regard to its economical consumption of petrol.“ (oed, 1989) the earliest appearance of “consumption” in 1398 carried the fourth meaning: “whan blode is made thynne.. soo folowyth consumpcyon and wastyng.” (ibid.) and from an economic standpoint, it is probably the sixth and seventh meanings that are most important, applying to producers and consumers respectively. in the seventh definition, the word “utilization” is especially appropriate, as the purpose of rational consumption is to maximize “utility.” but are “ investment” and “consumption” really so different as these definitions imply? in this paper we contend that investing can be seen as more than investors pursuing riskadjusted returns, a thesis which roughly parallels recent developments in the consumer research literature. with few exceptions (e.g., gardner & levy, 1955; levy, 1959; dichter, 1960) research into consumer research up until the 1980s characterized consumption as a process whereby consumers logically process information about brands and products in an attempt to make rational decisions to maximize their utility (holbrook & hirschman, 1982; hirschman & holbrook, 1982). furthermore, the information that consumers process was usually assumed to include objective attributes (e.g., price, volume, calories, miles per gallon, etc.), and it was assumed that these attributes were evaluated in light of the functional consequences they could yield for the consumer. similarly, investing has been viewed as a process in which investors logically process information about investments based on objective attributes (e.g., price, cash flow, liquidity, possibility of default, etc.) in light of the functional benefits they yield (e.g., the maximization of risk-adjusted return). 390 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 although several early consumer researchers (gardner & levy, 1955; levy, 1959) pointed out long ago that people consume products not only for their functions, but also for what they mean in people’s lives, it has not been until more recently that consumer research has viewed consuming from perspectives other than that of rational decision making resulting from information processing. holbrook & hirschman (1982) and hirschman & holbrook (1982) emphasize dimensions of consumption they have variously referred to as experiential, hedonic, esthetic, autotelic, and subjective. this line of research emphasized the multisensory, fantasy, and emotive facets of consumption. that is, consumption experience is received using all of the senses-gustatory, olfactory, tactile, auditory and visual. furthermore, these sensory perceptions can trigger imagery ranging from the recollection of autobiographical and historical events to unadulterated fantasy. finally, consuming involves a range of emotions including joy, jealousy, fear, rage, rapture, and so forth the experiential lens for viewing consumption introduced by holbrook and hirschman is just one lens that has been employed in the consumer research literature for investigating consumption. other paradigmatic perspectives have been introduced as well (holt, 1995). in section ii, we introduce two dimensions of consumption (structure and purpose) suggested by holt and in section iii explicate the four metaphors for consumption (experience, integration, play, classification) which he derives from them. of course, our purpose in these two sections is to explore the extent to which this typology of consumption practices might also apply to investment practices. in section iv, we perform a literary analysis of a popular book concerning personal investment in order to discover whether or not its veneer as a manual of objective investment advice does in fact mask a hedonic core. section v is a brief conclusion. our intention in this paper is not to provide a final word on alternative dimensions of investing or to exhaust the possible areas of inquiry that they suggest. rather our goal is to outline a general program of inquiry that may be utilized to study what people are really doing when they are investing. for the purposes of our discussion we will focus on primarily personal investment practices. our conclusions, however, may provide insights into professional imvestment practices as well. we acknowledge that our approach is quite different from what traditionally has been considered “fi nance research.” by recognizing that people’s experiences are characterized by logics other than rational economic pursuit, our approach even goes beyond “behavioral finance” (statman, 1999; shleifer, 1999; shefrin, 2000) that has so far assumed only that human experience deviates from economic logic due to information processing biases. while an approach such as this is now commonplace in the consumer behavior literature, it is also quite different from what traditionally had been considered “consumer research.” by introducing this new method (of literary analysis) and methodology into finance, we hope that it may begin to enrich the discipline in the same way as it has been enriching marketing. 2. a typology of investment practices holt’s (1995) typology actually takes off from a point one step beyond that which currently characterizes investment research. as we noted in the introduction, traditional 391d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 consumer research and traditional investment research both assume that in their roles as consumers and investors, people’s decisions to consume or to invest are based upon the evaluation of information concerning the objective attributes of something. things to be consumed or invested in have values; that is, the benefits which they can yield to a consumer or investor, and these values are themselves real attributes that can be determined by this evaluation of information and are independent of the person doing the evaluating. this is obviously true of the discounted cash flow (dcf) valuation of investments, which is a solution to one of the most important problems with which finance has concerned itself and can be found in every textbook. whether you are valuing a bond, a stock, a capital project, or a business acquisition, the process is essentially the same: forecast the cash flows, determine the discount rate, compare the costs and the benefits (using some method such as net present value (npv) or internal rate of return (irr)), and then make a decision regarding the investment. of the first three steps, the third (compare the costs and the benefits) is clearly a mechanical process that with the appropriate training everyone will be able to perform in the same ways. now if everyone has the same information and the same ways of making use of it, everyone will make the same forecasts of cash flows in the first step. and if there is a market consensus on the appropriate compensation for the deferral of consumption, for inflation, and for a unit of risk and on the appropriate amount of risk of an investment (via the capital asset pricing model for example), everyone will use the same discount rate in the second step. consequently, of course, everyone will make the same investment decisions in the fourth step. if anyone were systematically erroneous in their forecasts of cash flows or estimates of discount rates, this would quickly become obvious in the form of systematically inferior decisions. so while both traditional consumer research and traditional investment research are alike in their assumption that value is inherent in something, they differ in their view of the nature of this value. consumption value can be either functional (e.g., warmth, nutrition, safety, etc.) or symbolic (e.g., status, affiliation, etc.) functional value is more clearly independent of who it is who is doing the consuming, but since cultural meanings are shared among consumers, so too is symbolic value. investment value, on the other hand, is purely a functional monetary return and as such has nothing to do with culture. only a handful of nontraditional articles has considered that finance in general and investment in particular may be social or cultural phenomena and thus have symbolic value. (see, for example, frankfurter & lane, 1992; mcgoun, 1996, 1997, and 1998; and robinson & mcgoun, 1999.) the essence of the more recent consumption theory we referred to in the introduction is that consumption value, whether functional or symbolic, is not independent of the consumer. different consumers can consume things in a variety of ways, and it is this variety of ways that holt (1995) is attempting to systematically describe in his typology. let us introduce it in his words: in terms of structure, consuming consists both of actions in which consumers directly engage consumption objects (object actions) and interaction with other people in which consumption objects serve as focal resources (interpersonal actions). in terms of purpose, consumers’ actions can be both ends in themselves (autotelic actions) and means to some further ends (instrumental actions). (holt, 1995, page 2) 392 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 these two dimensions of consuming, the structure of action and the purpose of action, yield a 2 � 2 matrix within which are four metaphors for consuming as shown in table 1. in the remainder of this section we will discuss these two dimensions, and in the following section we will discuss the four metaphors they yield. the thing to keep in mind with regard to this grid is that it is the action itself, whether consuming or investing, that is important, not any value that lies in the thing being consumed or being invested in. there are five levels of instrumentation involved here, and this makes it confusing to use the term “ instrumental” to describe one of the purposes of action, since this is only one of the five levels. the first level of instrumentation concerns only investing, since the traditional objective of investing is risk-adjusted monetary returns, and the traditional view of money is that is has value only in exchange; that is, as an instrument for the acquisition of things to be consumed. this then brings us to the second level of instrumentation, that the things which are themselves consumed or invested in are only instruments by which we acquire the value, functional or symbolic, inherent in them. this is the point at which traditional consumption and investment theories leave off. the third level of instrumentation, which takes us to the different plane of analysis which is the subject of this paper, is that the things which are themselves consumed or invested in are only instruments by which we are able to engage in the acts of consuming or investing. overall, holt’s matrix deals with this third level of instrumentation and within it refers to the final two levels of instrumentation – one explicitly and one implicitly. the implicit instrumentation, which we will call level 4, is that consuming and investing must consist of interactions with objects, but in some cases only as instruments for interactions with other people. in other words, all consuming or investing is fundamentally structured as an object action, but in some cases they are ultimately structured as interpersonal actions. holt calls a level 4 instrument a “ focal resource.” the explicit instrumentation, which we will call level 5, is that consuming and investing are instrumental actions for purposes of integration or classification. to illustrate these levels, consider a share of microsoft stock. the traditional view of investing is that it is an instrument (level 2) for earning a monetary return in the form of a dividend or capital gain. of course this monetary return is itself an instrument (level 1) with which we can acquire whatever goods and/or services we desire. academic finance concerns itself only with level 2, leaving the details of level 1 to marketing. and as we noted above, academic finance with rare exceptions acknowledges only the functional aspects of both levels of instrumentation; that is, the ability of microsoft stock to earn monetary returns and the ability of money to acquire goods and services. neither the stock nor the money is considered to have any symbolic value. what we want to examine in this paper are the instrumental (level 3) value of microsoft stock as a means by which we can become table 1 metaphors for consuming (holt, 1995, p. 3) autotelic actions instrumental actions object actions consuming as experience consuming as integration interpersonal actions consuming as play consuming as classification 393d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 investors; that is, participants in the activity of investing, the instrumental (level 4) value of investing in microsoft stock for interactions with other people, and the instrumental (level 5) value of investing in microsoft stock for integration and classification in our lives. we might also point out here that although they may be clearly functionally different, it may not be possible to differentiate levels 2, 3, and 4 symbolically. earning money, becoming an investor, and socializing as an investor are different functions, but being a microsoft shareholder is likely to have the same symbolic value regardless of why we are one, whether due to level 2 or level 3 instrumentation, and it is through level 4 instrumentation that we realize the symbolic value. now holt’s matrix specifically considers the activity of consuming or investing. first of all, the point of the activity can be either the value we obtain from our interaction with the object (not from the object itself) or the value we obtain from our interactions with other people that we are able to engage in with the help of the object. this is the structure of action dimension of the grid; that is, consuming or investing is an activity, and this activity is an interaction with an object or with other people. second of all, the point of the activity can be either the value we obtain from the activity itself or some other value that we are able to obtain as a result of engaging in the activity. this is the purpose of action dimension of the grid; that is, consuming or investing is an activity that is an autotelic end in itself or an instrumental (level 5) means to an end. holt devised his matrix to elucidate different metaphors specifically for consuming, but so far from our discussion it appears as if it can apply to investing as well. beyond the traditional perspective of finance, which recall is concerned only with the functional aspects of instrumental levels 1 and 2, we can see investing at instrumental level 3 as an end in itself. this is the starting point of the matrix. within the matrix on the structure of action dimension, we can purchase microsoft stock for what we can do with the shares themselves (object actions) or for what we can do with other people as a microsoft shareholder (interpersonal actions). and on the purpose of action dimension, these actions with the shares or with others can be ends in themselves (autotelic actions) or means toward other ends in our lives (instrumental level 5 actions). of course at this point, the point of this classification may be quite unclear, but hopefully we have established that investing might be a desirable activity itself exclusive of its risk-adjusted monetary returns and what it is we can do with these monetary returns. in the following section we will discuss the four possible metaphors for investing as an end in itself – as experience, integration, play, and classification. 3. metaphors for investing 3.1. investing as experience investing as experience is a metaphor for the subjective, emotional interactions of investors with investments and for the ways in which investors make sense of and derive meaning from investing. according to holt (1995), consumption experience involves actors applying general interpretive frameworks in the context of specific domains that have unique 394 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 logics. the unique logic of a domain defines the rules of the game and helps actors who enter into the domain share in a coherent, recognizable experience. recall that the premise of holt’s matrix is that the activity of consuming and, as we are arguing, also of investing, can be desirable in and of itself. in a sense, consuming and investing are games (domains) that consumers and investors (actors) play according to certain rules (unique logics). from their participation (emotional interaction) in these games of consuming and investing, consumers and investors can derive (apply general interpretive frameworks) meaning and pleasure (coherent, recognizable experiences). this metaphor of investing as a game to be experienced is not such an unusual one, as markets were described as a “ fair games” at least as early as bachelier’s (1964) famous dissertation completed in 1900 (frankfurter & mcgoun, 1999). and as with most games, at least those played on a large scale, investing has an institutional structure (commercial banks, mutual funds, etc.) formal rules (commercial law, sec regulations, etc.), and informal rules (personal investment strategies, rules of thumb, etc.) to structure it. but the purpose of playing a game is to derive meaning and pleasure from doing so, and this requires an interpretive framework to structure, to understand, to value, and to respond to the experience. in other words, it is necessary to comprehend the unique social world or domain of investing. holt proposes three such interpretive frameworks or dimensions to experience: accounting (to make sense), evaluating (to value), and appreciating (to respond). accounting entails making sense of and typifying what is going on in a particular domain. in the case of investing, distinguishing between a bull and a bear market, or between a market that is experiencing a correction and a market that is developing into a long-run bear market are examples of how investors make sense of and typify what is going on in the investment domain. other examples of accounting in the investment domain may include understanding the terminology (“short selling,” “ technical level of support” ) and the actual procedures for making transactions. moreover, accounting entails more than typing, categorizing, and labeling. it also involves understanding these elements in a broader context. for example, in addition to knowing the market declined 200 points in a day, this fact is also viewed in the broader context of market news and conditions (e.g., interest rates rose or the fed chairman made disparaging remarks about the stock market). another dimension of experiencing entails evaluating. that is, when we experience things, we tend to place values on the elements involved. investors may evaluate the market, other investors, their own investment strategies and so on. these evaluations may be made by reference to norms, history, or conventions. one standard baseline that is used to evaluate the performance of investors is the overall market performance. additionally, the overall market may be evaluated with respect to the historical return of the market. promotional communications for particular companies or professional investors are prone to manipulating baselines for comparison to cast their past performance in the best possible light. performance can also be evaluated in light of a person’s particular history—for example, “buffet had a below par year for his standards.” evaluations can be made with reference to conventions as well. for example, two different investors whose portfolios approximated the overall market returns for the year may be evaluated very differently if one of the investors achieved her results via an index fund and the other achieved her results via investing in 395d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 individual foreign stocks. in many domains, quantification is critical for evaluations, and obviously, finance is a domain par excellence where quantification is crucial. the third element associated with experiencing is appreciating. appreciating entails responding emotionally to elements of the domain. in the context of investing investors may respond emotionally to the markets, various investment vehicles, investment-related media and news stories/information, other investors and their actions, their own actions, and so on. these emotions can run the gamut from disgust to elation. one particular study into high-risk leisure consumption (i.e., skydiving) focuses on an interesting bodily response referred to as flow. drawing on csikszentmihalyi (1974), celsi, rose & leigh (1993) define flow experience as a phenomenological state which totally consumes the actor and “where self, self-awareness, behavior, and context form a unitized singular experience” (p. 11). in common parlance, flow experiences are ones in which “people lose themselves”— that is they lose all self-awareness. this experience usually ensues in a context where the person is greatly challenged mentally and/or physically, but not to the extent to which the actor is completely overmatched. thus, one key for flow to ensue is a balancing between the actor’s skill and the level of challenge presented by the context. in the context of skydiving, skydivers exhibited the tendency to engage in ever-more difficult jumps as their skill levels rose. investing is necessarily a risky activity, since future returns are always uncertain. as dispassionate as an investor may appear to be, there must be an emotional aspect to investing, and to at least some extent, it must be a flow experience. for a novice investor, flow may be derived from pouring over information concerning mutual funds and investing in them. as one’s investment expertise increases, perhaps more accomplished investors move onto picking individual stocks and even becoming on-line day-traders. experts may dabble in options and derivatives to elicit flow from their investment activities. at its extreme, appreciating may lead to dysfunctional states such as addiction or compulsive investing. the parallels that can be made between investing and gambling are obvious, and it stands to reason that many of the dynamics that contribute to gambling addictions are present in investing. furthermore, the burgeoning practices of on-line trading and day trading add the additional element of “ internet addiction,” in which the virtual world can supplant the real world (solomon, 1998). 3.2. investing as integration the second major metaphor for viewing consumption practices is integration. integration refers to the process whereby the individual and the consumption object become more closely aligned. this can take on two different directions. that is, a consumer can manipulate objects of consumption to fit their own self concept or personal style, or a consumer can alter their own self identity or self concept to conform to the consumption practices in certain domains. one necessary process in achieving integration is assimilating. assimilating entails becoming part of the game—becoming a natural player. in order to assimilate one needs to know the rules of the game, how it is played, how score is kept and how to practically engage in the given consumption domain. in short, consumers must master the experience-related processes mentioned above. second, consumers engage in producing practices, or practices 396 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 that enhance the extent to which consumers actually feel as though they are participating in the game. finally, in the integration dimension of consumption, consumers engage in personalizing. personalizing involves practices whereby consumers try to take commodified, generic practices and make them seem like they are specific to themselves. with regard to investing, assimilation means the process whereby those engaged in investing come to see themselves as “ investors.” assimilation is not a trivial matter. for years, millions of people throughout the world have been investing through company or government-administered retirement plans, and nowadays millions more are investing through their own self-administered retirement funds. while all of them have been “ investors” in an economic sense, however, only a minority would actually call themselves that, and such a self-identification has much to say about what it is they are really doing and why they are doing it. there are a number of ways investors engage in assimilating, and numerous institutional resources provide opportunities for investors to assimilate the investing domain. students can formally major in finance or take courses in investing. countless books and other publications exist for newcomers to learn the tricks of the trade. furthermore, actual investing for the average investor has become much easier with the advent of discount brokers and internet transactions. an explosion in information sources has occurred, including several financial television networks and numerous financial periodicals. based on these sources, new investors or average investors are exposed to the language, conventions, and micro processes involved in investing. finally, the actual number of investment vehicles has proliferated. a particularly interesting phenomenon related to assimilation of investing would entail the socialization processes that college graduates go through during the early days of their entry into the wall street domain. one would expect that these new entrants change their physical appearances, lifestyles, and value orientations upon entering this new world. a particular instance in which consumption practices can powerfully become integrated with a consumer’s self-concept is when consumers find themselves in liminal states (belk 1988; schouten, 1991; celsi, rose & leigh, 1993). liminal states occur when people are making major transitions in their lives. such instances may include starting college, entering the workforce, moving to a new area, getting divorced, or retiring. in all of the instances, the familiar cultural and institutional settings that crucially impact self-identity change drastically along with personal financial situations. as such, consumption practices such as investing can be used to help construct/change/reinforce one’s identity as well as to cope with the differing financial needs. thus, it is not surprising that newly retired business people become highly involved in micromanaging their retirement portfolios as a way to maintain their existing self-concepts that have developed after years of being involved in businessrelated practices. they haven’ t just become more intent on trying to earn or preserve enough wealth for their golden years. alternatively, investing may be a way to change or supplement an existing self-concept. for example, the popular press highlights investment clubs that are comprised of elderly women. it can be argued that such investment groups are sensational due to the fact that women have been excluded historically from the financial world. in fact, these investment clubs may be seen as attempts by women to alter their self-identities through accreting to their self-concepts an element that has been sociohistorically denied them. 397d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 with all of the resources and opportunities available to the average investor, producing is relatively unproblematic. however, the extent to which one really regards oneself as a “player” and how one defines what a “player” is may vary. for example, a person investing in mutual funds or using a broker to invest in individual stocks may not perceive themselves to be the same kind of player as warren buffett or peter lynch. however, they may attempt to become more like these stars by reading a lot about them, trying to follow similar investment strategies, predicting what the actors may do. at the extreme, investors may imagine what it would be like to be a celebrated investor or may live vicariously through these investment stars, imagining that investing in mutual funds is akin to what the stars do. to a certain extent, investing may be suited optimally for producing given that most of it involves statement of opinions and predictions. furthermore, most of the professional prognosticators are often wrong, and very rarely are people held accountable for their past opinions and predictions. in terms of personalizing, there seem to be many opportunities to personalize investing, transforming it from a mass enacted practice to a practice which is done uniquely by an individual investor. investors can pick and choose from various investing strategies and incorporate them into their own personal style. the plethora of different investment vehicles and individual stocks, and so forth provides for a virtually infinite array of personalized investment portfolios. indeed, with the advent of tremendous amounts of data available to the average investor, increased accessibility to enacting exchanges, and the popularization of investing in the mass media, investing may actually be more well-suited for integration than many other practices such as sports spectating in which the average person’s ability to engage in the sport at the same level as the stars is limited. one need not be 7 feet tall, or 300 pounds, or possess extreme athletic talent. rather, all one needs is a t.v. and maybe a computer and a little spare cash to become involved. 3.3. investing as classification the fact that an important role of consumption entails the classification of consumers was noted long ago by veblen (1973), who coined the term “conspicuous consumption.” veblen argued that the goods people consumed not only fulfilled functional needs, but also symbolically communicated one’s position in society. likewise, investing can be seen as a practice through which people establish ties of affiliation with in-group members and distinguish themselves from out-group members. classification practices can occur consciously and intentionally in groups that have strong solidarity and are explicitly recognized. alternatively, classification can occur in a less-organized and strategic manner as well. whereas veblen posited a more conscious and strategic pursuit of distinction through consumption, bourdieu (1984) chronicles how these acts of classification can occur in a more subtle and less intentional fashion. in the context of investing, perhaps the most basic division among categories is between those who are actively involved in financial markets versus those who have no discretionary funds to invest. among those who have discretionary funds there may be further classifications from those who keep their money under their mattresses, to those who keep their money in savings accounts, and proceeding from there all the way to those who invest in very 398 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 sophisticated instruments such as options and other derivatives. in addition, classification can occur in a more subtle manner than simply by reference to the types of investments held. as pointed out by bourdieu (1984) the style in which goods are consumed can be at least as important for distinguishing categories as the actual goods themselves. thus, two investors possessing derivatives in their portfolios may be distinguished based on the fact that one has had a professional investor make the decisions whereas the other has the expertise herself. in addition, the mere expertise and familiarity that one shows with an investment instrument during conversation may be distinctive as well. the preceding discussion on the classificatory nature of investing may prompt some skepticism among those who point out that investing usually occurs in the privacy of one’s office or home, and hence is not a practice amenable to conferring distinction. in the cultures in which most investment occurs, the details of one’s financial affairs are considered a private matter. however, to appreciate the public nature of investing one need only be at party where one of the guests has been active in investing and is discussing the market in the midst of people who have never been in a financial position to even consider investing money to any extent. indeed, one primary way that investing can become conspicuous is through storytelling. just as holt (1995) points out that a primary impetus for attending baseball games may be to gather grist for future stories, so too one may engage in active investing to acquire future war stories to be recounted on appropriate occasions. following bourdieu (1993), it is probably the case that a researcher may discern a relative status hierarchy of financial instruments that accords roughly with a status hierarchy of investors, both amateurs and professionals. a final method by which classifications may be communicated is via mentoring. one can imagine the scenario where the young working class person consults the middle-aged professional down the street for advice as to how to invest his aging mother’s retirement funds. the hierarchical, power-laden structure of classification is evident here. the young man becomes indebted to his neighbor and is mystified and perhaps even somewhat awed by the special and unique faculties that his neighbor possesses. 3.4. investing as play perhaps the last metaphor used for understanding consuming, consuming as play, will be received with the most anathema when applied to investing. to those who have considered investment as the serious pursuit of monetary returns, to suggest that investing has a playful character may seem utterly objectionable. to be sure, people take money and investing seriously; however, it is not unheard of to hear someone characterize their investing as “dabbling in the market” or of someone setting aside a small sum to “play around with in the market.” one of the play-related forms that investing surely takes on is by providing an opportunity for communing and socializing. just as sports takes on a social role that affords people from very different walks of life and with nothing in common the opportunity to communicate and socialize, so too can investing. in fact as we have noted above, a common point of discussion at many cocktail parties and other social gatherings may be the market’s latest performance or breaking financial news. investing may be particularly amenable to serving this function 399d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 due to its nature as we have also already noted—namely, investing invites differing points of view and debate without being such that any point of view is definitively correct. in fact, many forums in which professional investors offer their prognostications may be infused at points with a subtle “wink and a nod” suggesting that most understand that prediction is an intricate game which must be taken with a grain of salt. investing as a play activity that helps people commune must surely play a role in the investment clubs that have been receiving recent attention. this became most obvious to one of the authors when he was approached by his spouse about joining such a club. after initially offering notions about market efficiency and the random walk thesis, it became much clearer that joining the club was about much more than maximizing return on investment. rather, it could be seen as being more akin to the friday night poker game, a golf outing, or a crafts club. investing simply serves as the focal medium for interacting with others. 4. case analysis in order to elucidate how holt’s matrix might be applied to investment, we have selected a popular book, the motley fool investment guide (gardner & gardner, 1997) to consider how investing might be targeted to appeal less to a potential investor’s desire for financial security and more to meeting those other needs which we have discussed above. the book was chosen for its being the top-selling recommendation from the on-line bookseller amazon.com in the general investing category in august, 1999. in fact, four of the ten top-selling recommendations at that time were associated with aol’s investment discussion and advice sector conducted by the gardners as the “motley fools.” the subtitle of the book is “how the fool beats wall street’s wise men and how you can too.” it does not mention finance at all; rather, it suggests investing as a social arena in which one class (“you” and “ the fool” ) is in contention with (in order to “beat” ) another (“wall street’s wise men” ). this is amplified in the forward to the book in which its objectives are set out. this book will enable even the rankest novice to invest expertly on his or her own, enjoy the heck out of it, and beat the pants off the market averages . . . all things that too many people think takes an expert, a wise man, or a market insider to do–those foolish enough, that is, to believe that the market can be beaten at all. if you harbor the faintest intellectual curiosity, relish–not wilt from–risk and challenge, instinctively enjoy taking responsibility for your own future, and own a modem, today’s investment environment is for you. (ibid., page 10; italics in original) this paragraph succinctly captures the popular american myth of the plucky (“harboring the intellectual curiosity,” “ relishing risk and challenge,” “ instinctively enjoying taking responsibility for their own future” ) underdogs (“ rankest novices” ) taking on the system (“experts,” “ wise men,” “ market insiders”) and triumphing over it (“beating the pants off them”). throughout the book, finance professionals and others directly and indirectly associated with financial markets and institutions (i.e., the media, academics, accountants) are disparaged in colorful terms. in contrast to these finance professionals, the book depicts the common person to whom it is directed in the most complimentary terms, terms in which 400 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 anyone would be pleased to be described. of course the authors sometimes need to reassure readers, who might justifiably feel intellectually inferior, by indulging in some self-deprecation. of course, their description of themselves as “fools” makes self-deprecation an underlying theme of the book. don’ t let those numbers confuse you; this is fifth-grade fare, and fortunately for fools like us, that’s about as tough as the mathematical work gets in this book. (ibid., page 79) in short, at most points this book looks less like an investment guide than an incitement to a populist rebellion, not only against established financial markets and institutions but also against the decadent way of life they represent. even if one were not interested in participating in such a revolution, at the very least the book can lead us effortlessly to an idyllic domestic hyper-reality similar to that of garrison keilor’s mythical hometown community lake wobegon, where among other things, “all the children are above average.” it’s content often has little to do with the economic rationality one would expect to underly an investment guide, with vivid imagery and metaphors that wander far afield. the entire part viii of the book “here be dragons: investment approaches to avoid” containing chapters 21 “the carnival of freak delights” and 22 “the leibniz pre-harmonic oscillator” is exceptionally peculiar. in the terminology of holt’s matrix, there is almost nothing in this book to recommend investing as an autotelic activity. although it encourages investors to account for their portfolios and evaluate their results against market averages, it discourages any mastery of terminology or rules. these sorts of things are the domain of the inferior “wise” and not the superior “fools.” hence investing is not experience. and although it does suggest that you can “enjoy the heck out of it,” investing is largely presented as something that you should get over with as quickly and easily as possible without any emotional involvement. its recommendation of investment clubs has nothing to do with their social value. hence investing is not play. throughout the motley fool investment guide runs the theme of investing as an instrumental action. it is difficult, however, to specify whether the book is more concerned with investing as integration or as classification. this book draws a sharp social distinction between finance professionals and their associates on one hand and “ real people” on the other. when the book mentions any finance professional favorably it does so for the “down-to-earth” qualities the person shares with “ real people” and attributes his success to them. in and through investing, you can classify yourself as either “wise” or “foolish,” both terms being used with irony. implicit in any investment guide is investing as integration; that is, personalizing the process and taking charge of something yourself that had previously been entrusted to someone thought to be superior at it. but unquestionably this book promotes this in a special way to an exceptional degree. in the motley fool investment guide, financial markets and institutions are a metaphor for any system social, economic, or political from which the reader feels excluded. but according to the book, the “wise” by whom and for whom these systems are run are not wise, and the “fools” who are excluded from these systems are not foolish. the act of investing (in accordance with the prescriptions of the book) is integration in that it redefines the financial system as one that not only encompasses the “fools ”along 401d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 with the “wise” but that in fact serves the “fools” better than the “wise,” and in doing so makes the “fools” feel better about themselves as fools. that the motley fool investment guide promotes investing as classification and investing as integration has some confirmation in the comments on the book posted by readers on the amazon.com web site. out of 62 comments, 17 (27%) explicitly expressed satisfaction as outsiders with the book’s debunking of the “wise” (classification) or described the empowerment they felt taking investments into their own hands. interestingly, a couple of the readers indicated that the debunking was a common theme in investment guides, empowerment, as we have pointed out, being their raison d’etre. another observation we might make concerning the reader comments is that very few readers comment on the financial outcomes of following the book’s advice. one reason, of course, is that it takes quite some time to accumulate meaningful results attributable to any investment strategy. but this does suggest that the consumption of investment books is a practice which itself might be amenable to analysis using holt’s matrix. not only is investing itself a form of consumption, as we have argued in this paper, but embedded in investing are numerous more traditional forms of consumption. the motley fool investment guide and similar books may in fact be ends in themselves, read by people who have no intention of investing. or they may represent yet another level (level 6) of instrumentation; that is, consumption undertaken in order to undertake another form of consumption. 5. conclusion it does indeed appear that the practices of consuming and investing are undertaken for many of the same reasons and that holt’s typology of metaphors for consuming is applicable to investing as well. although in an economic sense the two represent the alterative uses of output, in a psychological or sociological sense, they are very similar instrumental means to noneconomic ends. in fact, it is the notion that there are multiple layers of instrumentality involved in consuming and investing that links traditional perspectives on both activities to newer theories that might offer us greater insights into them. investing is more than an instrument for acquiring cash returns, as finance has traditionally viewed it, and the goods and services one can acquire with cash are more than instruments for enjoying the benefits these goods and services can provide, as marketing has traditionally viewed them. at the higher levels of instrumentation, the things which are consumed or invested in are instruments by which we are able to engage in the acts of consuming or investing, the acts of consuming and investing are interactions with objects which are instruments for interactions with other people, and the acts of consuming and investing are also instrumental actions for purposes of integration or classification in our lives. there may even be additional instrumental hierarchies layered on top of this process, since consuming or investing may be undertaken as instruments to facilitate or enable other acts of consuming or investing. the particular phenomenon of investment guides exemplifies holt’s two instrumental metaphors for consuming. by their very nature, investment guides promote investing as integration; that is, personalizing the process and taking charge of something yourself that 402 d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 had previously been entrusted to someone thought to be superior at it. relatedly, the guides portray the act of investing as classification, drawing a sharp social distinction between finance professionals and their associates on one hand and “ real people” on the other. while we believe that holt’s two autotelic metaphors, experience and play, can also apply to investing, we would expect investment guides to downplay the former lest the process appear too intimidating and the latter lest it appear too frivolous. financial services is a vast and expanding industry, whose size and growth seem disproportionate to its economic role. but if its “product” is more than the “allocation of funds from savings surplus units to savings deficit units,” its structure and function make much more sense. references bachelier, l. “theory of speculation,” trans. a. james boness, in the random character of stock market prices, ed. paul j. cootner, (cambridge, massachusetts: the mit press, 1964). belk, r. (1988). “possessions and the extended self.” journal of consumer research, 12, 265–280. bourdieu, p. (1984). distinction. a social critique of the judgement of taste. cambridge, massachusetts: harvard university press. bourdieu, p. (1993). sociology in question. thousand oaks, ca: sage. celsi, r., randall, r., & leigh, t. (1993). “an exploration of high-risk leisure consumption through skydiving.” journal of consumer research, 20, 1–38. csikszentmihalyi, m. (1974). beyond boredom and anxiety. san francisco: josey –bass. dichter, e. (1960). the strategy of desire, new york: doubleday. frankfurter, g. m., & lane, w. r. (1992). “the rationality of dividends.” international review of financial analysis, 1, 115–129. frankfurter, g. m., & mcgoun, e. g. “ ideology and the theory of financial economics,” journal of economic behavior and organization, (forthcoming). gardner, d., & gardner, t. (1997). the motley fool investment guide, new york: simon & schuster. gardner, b., & levy, s. j. (1955). “the product and the brand.” harvard business review, 33, 33–39. hirschman, e. c., & holbrook, m. (1982). “hedonic consumption: emerging concepts, methods, and propositions.” journal of marketing, 46, 92–101. holbrook, m., & hirschman, e. (1982). “the experiential aspects of consumption.” journal of consumer research, 9, 132–140. holt, d. b. (1995). “how consumers consume: a typology of consumption practices.” journal of consumer research, 22, 1–16. levy, s. (1959). harvard business review, 37, 117–124. mcgoun, e. g. (1998). “dividends and society.” slovene economic review, 49, ( 4), 341–353. mcgoun, e. g. (1997). “hyperreal finance.” critical perspectives on accounting, 8, 97–122. mcgoun, e. g. (1996). “fashion and finance.” international review of financial analysis, 5 (1), 65–78. the oxford english dictionary. (1989). second edition, oxford: clarendon press. robinson, c., & mcgoun, e.g. (1999). “the sociology of personal finance.” financial services review, 7 (3). schouten, j. (1991). “selves in transition,” journal of consumer research, 17, 412–425. shefrin, h. (2000). beyond greed and fear: understanding behavioral finance and the psychology of investing. boston, massachusetts: harvard business school press. shleifer, a. (1999). inefficient markets: an introduction to behavioral finance. oxford: oxford university press. solomon, m. r. (1999). consumer behavior, upper saddle river, nj: prentice hall. statman, m. (1999). “behavioral finance: past battles and future engagements.” financial analysts journal, 55, 18–27. veblen, t. (1973). the theory of the leisure class, boston: houghton mifflin company. 403d.e. allen, e.g. mcgoun / financial services review 9 (2000) 389–403 pii: s1057-0810(01)00074-9 the equity index annuity: an examination of performance and regulatory concerns gregory a. kuhlemeyer* carroll college, department of business administration and economics, waukesha, wi, usa received 8 february 2000; revised 26 march 2001; accepted 4 may 2001 abstract the equity index annuity (eia) is a recent product development in the insurance industry. this paper details eia design and interest crediting techniques, examines expected performance, and discusses possible regulatory concerns. results indicate that the eia is generally expected to perform better than a traditional fixed annuity for contract periods of at least five years, but is substantially below that of a similar direct equity purchase. eia contracts are not appropriate for shorter-term investors when factors of risk and efficient markets are considered. the sec is not currently regulating the eia product, which should raise industry concern. © 2001 elsevier science inc. all rights reserved. jel classification: g00; g29 keywords: equity index annuity; eia 1. introduction the equity index annuity (eia) is a relatively new financial alternative available from a growing list of life insurance companies across the united states. jackson national, conseco, and allianz/lifeusa are among the largest issuers of equity index products. eias represent investment alternatives that allow the purchaser to participate in the appreciation of a basket of equity securities. the appreciation in the deferred annuity contract is generally * corresponding author. tel.: �1-262-951-3061; fax: �1-262-524-7397. e-mail address: gkuhleme@cc.edu (g.a. kuhlemeyer). financial services review 9 (2000) 327–342 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00074-9 based on one of several different equity indexes, predominately the s&p 500, and is constructed with a guaranteed minimum value. there has been expansive growth in retailing eias over the first few years of their existence. since being first offered in 1995 (sales of $0.4 billion), the financial press has reported that eia product sales in 1999 grew to $5.1 billion and have a current asset base exceeding $15 billion in force (koco, 2000). this substantial growth has not been appropriately represented by an increase in consumer information related to the usefulness of using these instruments in personal financial planning. three issues are examined in this paper. the first issue is an examination of the design of common eia contracts. the four common interest-crediting techniques examined are simple point-to-point, point-to-point with annual ratcheting, point-to-point with asian-end, and point-to-point with look back. second, a bootstrapping technique is utilized to detail the expected performance (return and risk) to investors for a variety of eia contracts based on the s&p 500 index. actual historical data are used to create a simulation of likely outcomes over a variety of holding periods for each of the different interest crediting techniques. finally, this paper provides a brief discussion of the regulatory status of eias in comparison to similar investment alternatives. the design of the eia looks like an equity or derivative equity security, has a return based on an equity security, but faces risks that are not as great as comparable equity securities. it is the opinion of this author that this instrument should be regulated similarly to other equity-based variable life insurance products or mutual funds. this introduction is followed by a review of the literature in section 2 and a detailed description of the most popular eia designs and interest-crediting techniques in section 3. an analysis of the methodology used to analyze the risk-return relationship is provided in section 4 while section 5 includes a detailed discussion of the performance results. regulatory concerns are discussed in section 6 and concluding remarks are provided in section 7. 2. literature review this unique equity-based investment has yet to generate significant product-specific literature. a basic investor introduction to stock index futures, stock index options exchangetraded index funds was recently provided by zigler (2000) and provides an excellent description of how to hedge and invest as an alternative to index mutual funds. the eia product has an important role for use by planners and individuals alike in regards to asset allocation and investor risk preferences. in addition, the annuity aspect of this product plays a role in after-tax returns investors receive while the complex derivation places a premium on product and investor education. this paper extends the literature by examining the performance of an eia in relation to previously accepted research and raising concerns about the current regulatory environment for this equity-like security. the traditional life cycle theory on investing suggests that investors should move to declining equity positions as they move closer to retirement. edler and rudolph (1999) include a literature review on the life cycle theory of investing. bodie and crane (1997) attribute this reduction in the portion of equity holdings to the increasing risk of human capital as investors approach retirement. they conclude that investors can appropriately 328 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 self-manage retirement funds given sufficient education and investment information. yuh, hanna and montalto (1998) attribute a larger proportion of the population being in retirement and enjoying longer life spans as the reasons our population struggles to maintain their current spending habits into retirement. the authors recommend that “aggressive investment or savings strategies should be encouraged” by retirees. the increasing emphasis on equity investments with concern for risk tolerance is consistent with bodie and crane (1997) and the development of products similar to eia contracts. many researchers have looked at historical performance in equity and financial markets. in 1976 ibbotson-sinquefield (is) published a significant analysis of asset returns starting with 1926. ibbotson associates (1999) continues to update debt and equity asset class returns contained in this database. wilson and jones (1987) further expand the is equity database by examining returns from the end of 1870 through 1925. the authors find that this additional set of nearly 60 years results in a geometric mean return that is 124 basis points below that of the original is data series. this is of serious concern to investors making decisions using equity-based or derivative securities. the authors also point out that debt outperformed bonds in total for the period 1870 to 1914. this is a period that exceeds 40 years and would be defined as “long-term” to most investors and is troublesome given the plethora of information expounding the virtues of equity investing. if investor expectations exceed what broader and objective data will support, then this could lead to overestimation of performance and a failure to meet their financial goals. jones and wilson (1987) also clearly show the relative importance of dividend income in their early data set and capital appreciation in the is-based data set. the authors attribute this to the advent of personal income taxes and the competitive advantage of capital gains taxation occurring near the end of their early time period. the limiting downside risk of eias is a critical aspect because of the risk averse nature of investors. risk preference has long been discussed and measured in many works. more recently, grable and lytton (1999) discuss methods of assessing risk tolerances of financial planners and their clients and describe previous research on investor willingness and desire to take risk. the complicated nature of eia contracts places a premium on investor education and knowledge. chen and volpe (1998) find the weakest area of personal finance understanding among college students relates to the area of investments and is greatest for those students who were younger, nonbusiness majors, female, or had little work experience. in addition, the authors find that students with more financial knowledge make fewer incorrect investment decisions at a rate of 20% versus those with a lack of financial knowledge at 48.4%. one can reasonably interpret that greater exposure by individuals to financial topics through life is likely to increase their understanding of financial topics. edler and rudolph (1999) find that retirement satisfaction is directly related to engaging in planning activities and that educational level does not enhance satisfaction. the essence of an equity-indexed product is the derivative use in its creation. gerber and shiu (1999) examine the pricing of reset guarantees and show how it can be applied to pricing some specific eia contracts that contain guarantee resets. the basis for the development of these models is the black-scholes formula. many others have significantly expanded option research. unfortunately, the black-scholes model is a bit less robust the 329g.a. kuhlemeyer / financial services review 9 (2000) 327–342 longer the option life which boyle and lin (1997) address in their valuation model for look back options. boyle and lin (1997) is an extension of other valuations of complex options where there is essentially a set of two or more assets from which value is derived (boyle and tse, 1990; and nelken, 1996). in many eia products, consumers have the choice of maturities that often extend for five to ten years. a final aspect for most insurance-based assets is tax deferral as it influences investment choice. investors often choose one alternative over another even though the performance attributes would not normally lead to this decision ignoring the tax deferral issue. in addition, equity investments have competitive advantages with regards to capital gains that may influence investor preferences. siegel and montgomery (1995) recommend legally avoiding taxes as they find that common stock investments are advantaged over treasury bond investments after accounting for taxes, transaction costs, and inflation. 3. eia design and interest crediting techniques 3.1. eia design and vulnerability issues eias are deferred annuities designed to allow the investor to participate in the appreciation of a specified equity index. for example, if the s&p 500 index rises from 1200 to 1600 over the life of the contract, the index has risen by 25%. owners of these deferred annuity contracts will receive some proportion of this return based on the specific interest crediting technique applied to the contract. this proportion is commonly referred to as the participation rate. equity index contracts will typically have a contract period that varies in length from 1 to 10 years and may cap (limit) the total appreciation over the contract period. this product design generates several potential risks. eias rely on the uncertain future performance of the index to generate interest credits to each contract over the contract period. the funding of eia instruments is accomplished with a combination of fixed return bond instruments and equity derivatives. these derivatives are usually related to the underlying index and should theoretically be able to provide a perfectly hedged portfolio that does not increase insurer risk. the reality of this situation is that each company will be unlikely to perfectly hedge this portfolio on a daily basis given somewhat random cash inflows and outflows and asset values that may move the portfolio away from a perfectly hedged situation. to complicate matters, a rising yield curve combined with falling equity markets is likely to increase volatility for index-based derivatives. the higher yield curve and the falling eia equity account may entice current eia holders to increase lapse rates. higher lapse rates occur when the cost of exiting a poor performing contract falls and the probability an investor will receive only the guaranteed minimum rate rises. although this is a situation that has not yet occurred in a substantial fashion, it places the insurer in a situation in which cash outflows is likely to exceed cash inflows and increase its business risk. the most astute investor will recognize the vulnerability of investing in the eia product because of a failure to receive dividend income in addition to any appreciation in the index. an investor who is less discerning may not understand the difference between index performance and the total return performance of a basket of individual securities that 330 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 represent the identical index. a mutual fund that represents an index will receive both capital appreciation and dividend income. historically, dividend yields have ranged from 2.11% to 8.77% out of a 12.96% arithmetic mean total return on the s&p 500 based on data provided by the ibbotson sbbi database (ibbotson 1999). the compounding effect of dividends never received on the s&p 500 index represents a significant opportunity loss of over 25% of the possible wealth over a 5-year period and over 57% for a 10-year period. 3.2. interest-crediting techniques a variety of interest crediting techniques are employed in eia contracts and this paper focuses on four primary interest indexing methods: simple point-to-point, point-to-point with annual ratcheting, point-to-point with asian-end, and point-to-point with look back. competition is increasing within the eia market but insurers appear to realize that the uniqueness and complication of the interest crediting method can create barriers to entry and minimize the competition between (or comparison of) near commodity-like products. the interest credited to any specific contract will generally depend on the choice of equity index, contract period, participation rate, indexing method, and contract guarantees. it should be noted that annuity contract features such as issue age, transfers, withdraws, surrender charges, death benefits, and other special features are ignored. these features may cause a consumer to choose one product over another, but have no direct bearing on the interest credited to an annuity held to contract maturity for a consumer that is preretirement. in addition, the products are deferred annuities and provide contract owners with tax deferability not available with many alternative investments. for explanatory purposes during the remainder of this section, let us assume that the s&p 500 index has the closing values listed below over the next five years. the total index return, ignoring dividends, is calculated simply as s&p 500 index(t) /s&p 500 index(t�0) for a contract period of t periods. the simple point-to-point (ptp) contract is generally based on the anniversary date of the contract. assume a consumer at t�0 purchases a five-year eia annuity on july 1st (anniversary/contract date) using a simple ptp technique based on the s&p 500 index. in addition, this contract has a 70% participation rate and a guarantee period of five years during which the entire principal (100%) is guaranteed at 3% annually. the contract interest guarantee provides that, at a minimum, the example eia contract owner will earn 15.93% over the five-year period as defined by eq. (1). time s&p 500 index index total return eia return t�0 1,600 t�1 1,800 t�2 1,680 t�3 1,810 t�4 1,770 t�5 2,300 43.75% 30.63% 331g.a. kuhlemeyer / financial services review 9 (2000) 327–342 ��1 � i)n � 1� � � (1) in eq. (1), i is the annual guaranteed rate of return, n is the number of years of the contract, and � is the proportion of the principal that is guaranteed to earn the rate i. the investor purchases the contract with an expectation that the return will exceed the minimum value with some probability. the actual eia return calculation is described in eq. (2). �it�n/it�0 � 1� � � (2) it�n represents the index value on the nth anniversary date, it�0 represents the index value on the contract date, n is the number of contract years, and � is the participation rate over the contract period. if index returns match those described above, the five-year annuity will be credited with a 30.63% return based on a participation rate of 70% given the total index return of nearly 44%. the variations to this interest crediting technique are numerous and most tend to involve a ratchet or annual reset. the ratchet version uses the same basic methodology, but credits interest on each anniversary date. thus, in a five-year contract the owner may receive the minimum guarantee one-year and a return based on the index the next year. in contrast, the simple ptp considers only the entire contract period. this obviously increases the return possibility for the consumer and is generally associated with a lower participation rate or guaranteed interest rate. an additional variation calculates interest on the calendar year rather than the anniversary date. midyear contract dates, like the above example, will earn a pro-rata share of the annual interest credit for the initial and concluding calendar years. in some eia contracts the minimum rate of return may be additive rather than compounded thereby allowing the issuer to increase participation rates all else equal. assume again that there is an annuitant who purchases a five-year annuity on july 1st using simple ptp with annual ratcheting (ptp-r) on each anniversary date. this contract has a reduced participation rate of 45%, but 100% of the principal at the beginning of each one-year period is guaranteed to earn a 3% return and capped with an annual maximum return of 12% per period. contract returns are compound rather than additive in nature. the annuitant would earn a total of 29.88% over the entire five year with years two and four receiving the minimum return and the last year capped at the maximum annual return. notice that the ptp-r technique is characterized by much lower participation rates than the simple ptp example, but the total contract return is nearly identical. investor concerns occur with the longer-term simple ptp and the ptp-r methodology as time s&p 500 index index return eia return t�0 1,600 t�1 1,800 12.50% 5.63% t�2 1,680 �6.67% 3.00%* t�3 1,810 7.74% 3.48% t�4 1,770 �2.21% 3.00%* t�5 2,300 29.94% 12.00%* 332 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 the interest credited is dependent on only two points in time–the beginning and ending time periods. index values can vary dramatically over a short period of time, although most would argue that the historical trend is generally positive over very long periods of time. to address this concern some insurance companies offer eia contracts that credit interest where the ending value is an average of the closing index values over some specified period of time. this is generally referred to as an asian-end (ptp-ae). for the five-year contract discussed above, the contract might have a 6-month, 12-month or possibly a 365-day asian-end. if this contract uses the 12-month asian-end, then the ending index value is calculated as the average of the final closing value and the 11 preceding monthly closing values for the index. the 365-day method would average the closing index value on every day during the preceding year. this technique helps mitigate the risk concern some consumers have regarding the final index value, but it will provide a lower expected return calculation. as one would likely conclude, these contracts can generally offer a slightly higher participation rate or floor guarantee than the simple ptp contract. in this example, if the last 12 monthly index values average 2,100 and the participation rate is 80%, then the total contract return would be 25% [(2,100/1,600–1)*0.8]. the fourth product category of eias is a point-to-point look back (ptp-lb) contract. this type of contract fits the market niche when a consumer asks, “why won’t you use the highest value of the index during the contract? it is not my fault the market went back down.” a common contract in this category is the “high water anniversary look back.” with this type of contract the issuer looks back over the index values on each anniversary date of the contract and uses the highest index value as the ending value in the ptp calculation. alternative variations include the high water day look back and the high water month look back to find either the highest single closing day or month during the contract period to determine the amount of interest credited to the account. the consumer should generally expect to see higher returns and a shifting of contract risk to the insurance company with look back contracts. the insurer will offset their higher risk with lower participation rates and floor guarantees to compensate the insurer for the additional risk-taking. for the ptp-lb technique, assume that the annuitant purchases a five-year contract using the high water month look back method and the original example data. let us also assume that the highest closing month for the s&p 500 index during the contract period is 2380. the ptp-lb contract will have a lower participation rate of 60% and earn the annuitant 29.25% [(2380/1600–1)*0.6] over the five-year period. each of these four primary interest crediting technique have numerous permutations. differences are usually based on the specific index used or on how the initial index value, terminal index value, minimum guarantee, and maximum credited return (cap) are determined. the complicated nature of eias places a premium on financial education and understanding for both planners and consumers. 4. evaluation methodology the four basic techniques of crediting interest (ptp, ptp-ae, ptp-r, ptp-lb) were chosen through an examination of over 60 individual eia contracts that were found from 333g.a. kuhlemeyer / financial services review 9 (2000) 327–342 information provided primarily through a report by milliman & robertson, inc (1997). the basic assumptions for the four interest crediting techniques are generic composites of specific information provided in the milliman & robertson report and supported by puertz (1997), gregory (1997), and horowitz (1998). this composite method is used for the purpose of analyzing basic interest crediting techniques rather than testing specific products. as discussed earlier, each eia contract is like a snowflake–similar but never identical. over 80% of the eia contracts identified use the s&p 500 as the index of choice in determining the amount of interest to credit to a contract. alternative indexes represent the nasdaq and various international equity indexes. although this method can be easily extended to other equity indices, this paper utilizes only the s&p 500 due to the ease of data availability via the ibbotson sbbi database (ibbotson, 1999). monthly total return and capital appreciation data for the s&p 500 were collected from the ibbotson sbbi database (ibbotson, 1999). these monthly values were collected for the period of december 1925 through june 1998 resulting in 870 monthly returns. this analysis is conducted assuming that the historical monthly return distribution for the s&p 500 is representative of the future monthly return distribution. each monthly value is then given a unique indicator code ranging from 1 to 870. the next step involves generating 100,000 random numbers. each random variable is then uniformly designed to correspond to any one of the 870 monthly s&p 500 return observations. the purpose of this bootstrapping (resampling) technique is to create a larger sample of potential outcomes for each of the contract periods under study and does not force us to consider only the exact set and sequence of economic situations that have historically occurred. instead, it allows us to consider future unknown situations using the historical distribution of monthly returns. for example, a single simulated year could potentially include the twelve best performing months in history. this scenario is unlikely, but this methodology provides for this extreme possibility in the analysis. an alternative methodology is to use only the actual historical holding periods within the ibbotson sbbi database. two problems occur with this method. first, it will be a smaller database if only nonoverlapping periods are used and other possible combinations of returns are not considered. this problem is mitigated by using overlapping periods and creates 763 observations for a nine-year contract. this solution creates a second problem of placing a lower relative weight on the earliest and latest monthly return observations. for example, both the first and last monthly observations will each be included only in one nine-year data point while intermediate monthly observations are included in 108 nine-year data points. this method can bias longer-termed eia simulations. the data set begins with the “roaring 20’s” and ends with one of the greatest bull markets. this methodology will bias downward longer-term expected eia performance and not be representative of what investors might expect. next, the 100,000 observations are broken into nonoverlapping 12-month, 60-month, and 108-month periods. this results in 8,333 1-year periods, 1,666 5-year periods, and 925 9-year periods for analysis. each of the full data sets allows examination of the expected performance for each of the four previously discussed index crediting techniques. each full data set is also segmented into quintiles based on eia return to examine the expected return and risk of the bottom, middle, and top quintiles. 334 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 a final comparison is made of the implied historical eia returns to similar products such as an s&p 500 total index returns, historical fixed annuity returns, and treasury bill returns. this analysis allows only a weak comparison as eia returns must be simulated using assumed participation, floor, and cap rates when relatively “factual returns” can be generated and applied in an equity index or traditional fixed annuity setting. 5. eia performance outcomes 5.1 one-year eia outcomes the mean return of the one-year ptp is in the neighborhood of 3.6% to 4.0%. results are provided in table 1. the one-year product in its simplest terms is simply a bet by the purchaser who believes the market will move significantly upward in the next year, but with a guaranteed minimum. although there are slight variations between products and firms, the basic one-year ptp has low participation rates between 15% and 40% and principal guarantees between 90% and 100%. if an alternative traditional fixed annuity is currently offering 4% over the next year, then the purchaser must believe that the market index is going to move upward by at least 16% given a 25% participation rate (4%/0.25). if an investor truly expects the market index to climb more than 16%, why is she/he investing in this particular product? a high tax bracket individual with a short-term holding horizon will need to earn an even greater rate with a taxable index mutual fund account. even within this category, the various company specific products may not be close to each table 1 equity index annuity risk-return results for one-year product product type principal guarantee (guarantee rate) participation rate ceiling on credited return mean return (min, max) standard deviation (% at min, % at max) s&p 500 total return (min, max) simple point-to-point 100% 15% n/a 3.63% 1.61% 13.7% (base case) (3%) (3.0%, 21.2%) (73.6%, n/a) (�56.9%, 152.2%) point-to-point 100% 25% n/a 4.70% 3.28% 13.7% (simple) (3%) (3.0%, 35.4%) (60.1%, n/a) (�56.9%, 152.2%) point-to-point 90% 40% n/a 3.79% 8.11% 13.7% (simple) (3%) (�7.3%, 56.6%) (8.0%, n/a) (�56.9%, 152.2%) point-to-point 100% 20% 12% 4.07% 2.10% 13.7% (simple) (3%) (3.0%, 15.0%) (65.4%, 2.1%) (�56.9%, 152.2%) point-to-point 100% 25% n/a 3.55% 1.49% 13.7% 12-month asian-end (3%) (3.0%, 23.4%) (75.3%, n/a) (�56.9%, 152.2%) point-to-point look back (highest month in last year) 100% 15% n/a 3.80% 1.75% 13.7% (3%) (3.0%, 23.1%) (65.2%, n/a) (�56.9%, 152.2%) quintile results are available upon request. 335g.a. kuhlemeyer / financial services review 9 (2000) 327–342 other on the efficient frontier with regards to the risk-return trade-off. a similar return is found between the base case and the 90% principal-guarantee contract, but the range of possible eia returns increase to 63.9% from 18.2% for the base case. a second risk measure, standard deviation, is provided for both (1.61% and 8.11%) but has limited value due to the elimination of one or both tails of the return distribution. this result indicates that different products are designed with different consumer preferences in mind. it is also not possible to definitively state that one annuity provider generates a larger profit margin than another, as the underlying portfolio will be different for each product. the simple ptp examples show that the participation rate can be increased if firms place a cap on the maximum rate of return that can be earned during the contract period. in this particular case, a 5% increase in the participation rate associated with a cap of 12% on the one-year eia will cause a slight 44 basis point increase in the expected return. an examination of the quintile subgroups show that the bottom and middle quintiles always receive the minimum guarantee return of the one-year simple ptp contract. the top quintile group provides a solid mean return of 10.04%. throughout the paper quintile results may be discussed but are not presented due to space limitations. results are available upon request. participation rates with the ptp-ae case can increase substantially (by 10%) to generate an expected return that is near that of the base case ptp because the s&p 500 index has historically moved in a positive direction. the mean return at 3.55% is only 8 basis points lower and the standard deviation 0.12% less than the base case one-year product. overall, the risk and return figures are very similar to the base case with the exception of a much higher participation rate that can be quoted to potential customers. an analysis of index return quintiles yields no additional insight. not surprising, the ptp-lb mean return and standard deviation are 17 and 14 basis points greater respectively for an eia with a monthly look back and identical participation rate. an analysis of index return quintiles also yields no additional insight. the issue at hand is determining if a short-term eia contract is appropriate for a client to “gamble” that the index is going to rise in an efficient market by purchasing a one-year eia. in most scenarios it is unlikely unless the client is exchanging one form of annuity contract for another (to postpone taxes) and prefers to be in the equities market for the near term. 5.2 five-year eia outcomes the longer-term data provide a much more fertile analysis beyond the one-year ptp. five-year results are provided in table 2. the mean total return for the s&p 500 index over the set of five-year periods is 91.1% [range from –59% to 784%]. in comparison, the mean eia return is 34.7% [range from 15.9% to 311.5%] or slightly more than one-third that of a direct investment. the owner of the base case five-year eia should expect that nearly 45% of the time the client will earn only the minimum return guarantee of 15.9%. the full data set is again segregated into quintiles to examine return and risk more carefully. the results are as expected with the bottom quintile receiving the total guarantee floor return of 15.9%. the middle quintile performs below the mean response for the full data set since the return distribution is not truncated to the right. the mean rates for the median 336 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 and top quintiles imply annualized returns of 5.08% and 12.8% respectively over the five-year period. these results are again somewhat troubling as consumers truly benefit only when they can correctly anticipate those economic periods in which the index returns are above the norm. again, efficient markets preclude this from occurring. a consumer that has the ability to properly “time the market” would be better served to make a direct investment since it is more profitable. thus, an expected return of 6.14% over a five-year period should not generate significant excitement when risk-free 3-month treasury bills historically earn an annualized rate of 3.8% and a comparable s&p 500 investment expects to earn an annualized total return of 13.8%. an identical ptp contract with a participation rate that is 10% higher generates an expected mean return of 40.3% (annualized rate of 7.01%). for the 90% principal guarantee alternative the participation rate can be increased to 65% (15% greater than the base case) to provide a similar mean return when compared to the base case. this alternative is also associated with an increase in risk as measured by standard deviation, return range, and proportion of the time receiving the minimum return. as one might expect, the participation rate must increase with the ptp-ae interest crediting technique to compensate for an average closing value that, more often than not, is lower than the actual closing index value. an increase from a 50% to a 55% participation rate table 2 equity index annuity risk-return results for five-year product product type principal guarantee (guarantee rate) participation rate cap on credited return mean return (min, max) standard deviation (% at min, % at max) s&p 500 total return (min, max) simple point-topoint 100% 50% n/a 34.7% 30.0% 91.1% (base case) (3%) (15.9%, 311.5%) (44.7%, n/a) (�59.0%, 784.2%) point-to-point 100% 60% n/a 40.3% 36.9% 91.1% (simple) (3%) (15.9%, 373.8%) (41.5%, n/a) (�59.0%, 784.2%) point-to-point 90% 65% n/a 39.2% 43.2% 91.1% (simple) (3% (4.3%, 405%) (26.8%, n/a) (�59.0%, 784.2%) point-to-point 100% 55% n/a 34.6% 29.4% 91.1% (12-month asian-end) (15.9%, 349.7%) (43.7%, n/a) (�59.0%, 784.2%) point-to-point 100% 50% 12% annual 34.9% 11.0% 91.1% (annual ratchet) (3% per year) (per year) (15.9%, 76.1%) (96.3%, 69.0%)* (�59.0%, 784.2%) point-to-point 100% 45% n/a 35.5% 28.7% 91.1% (look back) (3%) (15.9%, 315.6%) (38.11%, n/a) (–59.0%, 784.2%) * the annual ratchet technique allows each year out of the contract period to receive either the guarantee rate, the cap, or the actual return based on the participation rate. out of the five years, this refers to the percent of observations that have at least one period that the minimum or maximum levels impact the overall return of the product. quintile results are available upon request. 337g.a. kuhlemeyer / financial services review 9 (2000) 327–342 combined with the asian-end method of interest crediting will provide nearly identical performance results. the ptp-r product return is also nearly identical to the base case but investor risk is significantly smaller. the ptp-r method provides a preferred alternative to the base case for most risk averse clients although they give up the possibility of the “homerun” as the maximum expected return falls from 311.5% to 76.1%. over the five-year period, 96.3% of observations have at least a single one-year period that earn the minimum return and 69% of observations have at least a single year that is restricted by the contract cap. the results of the ptp-lb show that using a lower participation rate of 45% will provide a slightly higher mean return, slightly less risk as measured by standard deviation, no increase in the downside risk, and earning the guaranteed return on a lower proportion of the observations. 5.3 nine-year eia outcomes the time period over which an eia contract is written does not generally exceed nine years, although a few 10-year products exist. the nine-year product appears to be a much better purchase for the consumer than its shorter-term counterparts. the results, provided in table 3, show that the simple ptp product has a mean annualized eia return of approximately 7.6%. contrasting the 93.2% mean eia holding period return with a mean holding table 3 equity index annuity risk-return results for nine-year product product type principal guarantee (guarantee rate) participation rate cap on credited return mean return (min, max) standard deviation (% at min, % at max) s&p 500 total return (min, max) simple point-topoint 100% 70% n/a 93.2% 94.0% 222.3% (base case) (3%) (30.5%, 789.0%) (34.7%, n/a) (�56.5%, 1717.6%) point-to-point 100% 80% n/a 105.1% 108.4% 222.3% (simple) (3%) (30.5%, 901.7%) (31.8%, n/a) (�56.5%, 1717.6%) point-to-point 90% 75% n/a 95.2% 103.9% 222.3% (simple) (3%) (17.4%, 845.4%) (26.6%, n/a) (�56.5%, 1717.6%) point-to-point 100% 75% n/a 92.7% 92.4% 222.3% (12-month asian-end) (3%) (30.5%, 745.9%) (35.1%, n/a) (�56.5%, 1717.6%) point-to-point 100% 100% 12% annual 90.4% 23.0% 222.3% (annual ratchet) (3% per year) (per year) (30.5%, 177.2%) (99.5%, 98.7%)* (�56.5%, 1717.6%) point-to-point 100% 65% n/a 95.3% 92.2% 222.3% (look back) (3%) (30.5%, 732.7%) (30.7%, n/a) (�56.5%, 1717.6%) * the annual ratchet technique allows each year out of the contract period to receive either the guarantee rate, the cap, or the actual return based on the participation rate. out of the nine years, this refers to the percent of observations that have at least one period that the minimum or maximum levels impact the overall return of the product. quintile results are available upon request. 338 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 period return on the s&p 500 data of 222.3% illustrates that the consumer is receiving about 42% of the total possible return. an examination of the quintile performance provides a nearly identical explanation as the five-year product and adds no additional insight. increasing the participation rate by 10%, ceteris paribus, results in an expected mean holding period return of 105.1% (annualized rate of 8.31%). the second alternative guarantees 90% of the principal allowing the insurer to increase the participation rate to 75% to maintain a similar mean interest credit over the nine-year contract. unlike the previous case, the risk factors do increase with this technique. the standard deviation increases along with the range of possible returns (17.4% guarantee rather than 30.5%) and fully one-fourth of the holding periods will result in the deferred annuity owner receiving only the 17.4% floor guarantee. the three remaining techniques, ptp-ae, ptp-r, and ptp-lb each act similarly to the five-year product. the biggest surprise is with the ptp-r technique. the annual cap combined with a guarantee rate on the entire principal allows the firm to raise its participation rate dramatically to 100%. this results in a similar mean return with approximately onefourth the standard deviation and a tighter range of possible returns of the base case. the ptp-r interest crediting technique on a longer-term eia will fit well with most risk averse investors as the annuity owner is guaranteed a rate of return and receives all of the appreciation in the index up to the maximum cap set by the insurer. in addition, the return is “locked” each year and cannot be reduced or eliminated because of adverse market conditions in later years. 5.4 fixed annuity comparison in total, the expected performance of the five-year eia will be similar to that of a regular fixed annuity. there are a few data sources that provide historical deferred fixed annuity rates, although they are not precisely what is needed for comparison in this analysis. the first data source (jnl, 1999) is a single annuity issuer. the insurer provides information regarding the current contract value and annualized return of annuities purchased from 1975 through 1998. these annuities have holding periods from 1 to 24 years and annualized returns that range from 5.71% to 8.40% with a mean annualized rate of 7.15%. the most recent 10-year and 11-year annualized yields generate rates of 6.68% and 6.84% respectively. the web site for the thrift savings plan for federal employees (1999) provides an annuity history going back to 1989. the average monthly annuity rates range from 5.56% to 8.76% with a 7.06% mean for the years 1989 through 1998. the annuity shopper (1999) also supplied single premium deferred annuity rates going back to 1980. a simple average of the rate set yields a 7.03% average and a 6.73% monthly weighted average rate. although it is important not to generalize recent history as being historically representative, the long-term fixed annuity yields have been in the 6.6% to 7.1% range. as an additional comparison, the performance of intermediate and long-term government bonds is examined using the ibbotson sbbi database. the results are comparable to the previous annuity data sources as the ten-year period from 1988 through 1997 has geometric mean returns of 7.36% and 6.71% respectively for long-term and intermediate-term government bonds. for a broader comparison, the geometric mean over the entire 1926 to 1998 339g.a. kuhlemeyer / financial services review 9 (2000) 327–342 period for the long-term and intermediate government yields 5.24% and 4.75% respectively. if one assumes that the recent relationship between deferred annuity rates and the yield range of intermediate and long-term bonds continues into the future, then it is reasonable to expect that returns on deferred annuities might yield approximately 5% annually. the results seem to indicate the five-year eia product is comparable in expected returns to traditional fixed annuities ignoring all issuance costs. overall, it appears that the best choice for typical investors who are considering eias is to invest for longer-term periods. highly risk-averse investors who have a sufficient investment horizon should consider multiple long-term contracts (assuming eias continue to exist) to reduce the risks inherent in a single long-term (say nine years) period. an annualized return of 7.6% to 8.3% on a tax-deferred basis for two or three decades will provide a solid investment alternative for the risk involved. on the other hand, it is difficult to justify very short-term eias in any portfolio. 6. product regulation equity index annuities are commonly referred to as having 1) the ability to participate in the market with equity appreciation and 2) a limited downside risk because of contract guarantees. this opens up problems on how the eia will be marketed by the insurance company’s agents without crossing into the securities environment that is regulated by the sec. eias are a lot like the creature that looks like a duck, quacks like a duck, and flies like a duck. is it duck? eias have returns dependent on equity securities and risk that is greater than a fixed income security. is it a security? a basic return-risk analysis of eia product has yet to be published. it appears that insurance companies have intentionally steered away from providing this type of information due to the possibility of future sec regulation and a potential threat of other legal actions. the improperly trained or naı̈ve agent selling eias may not properly inform or understand the consequences of the eia product on the future value of a customer account. in fact, there appears to be a potential quandary for agents as a thorough explanation of this issue to a client could raise two difficult questions. the prospective client might ask: 1) should i just purchase a portfolio of index securities directly? or 2) why should i purchase a product where i do not receive all of the return (dividends)? the first question implies that the purchase would have to be facilitated by a securities licensed representative (or directly by the customer) and the traditional, nonsecurities-licensed insurance agent loses the sale. the second question is more easily handled as the eia contract guarantees a minimum percentage return. the overall problem is that a thorough explanation of the eia product crosses the line into selling a securities product and is dangerously forcing many insurance agents into areas that they are not legally equipped to handle. a subtle, but important point should be made regarding the difference between traditional fixed annuity contracts and the eia. the traditional fixed contract has an initial known rate that the purchaser accepts on the date that the contract is purchased, while the eia purchaser only accepts a minimum guarantee that may not even guarantee a complete return of 340 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 principal. it is the opinion of the author that the above arguments should generate sec regulation of eia contract sales in this country similar to other variable securities. eia contracts also raise an insurance regulatory issue as these products have the potential of increasing the business risk of those insurance companies underwriting eia contracts based on their decision on how to properly fund and manage their underlying portfolio. a.m. best has previously made public announcements that downgrades could occur if eias were to become a significant part of the product portfolio. safeco and keyport life have left the index product business in recent years because of what the popular press (greene, 2000) has reported as poor investment decisions with eia premiums. this seems to provide supporting evidence of the potential risk that exists to insurers in this market that must ultimately be borne by the customers of the insurer. 7. conclusion the equity index annuity is a recent development that allows traditional insurers to provide an equity-like product to their product line. the results of this analysis indicate that the product, on its performance basis only, does not match up well with traditional equitybased investments and, in many instances, traditional annuity performance. the biggest advantage that a single premium deferred annuity has is tax deferral. the eia tax advantage disappears as the direct investment has the benefits of sale timing, a lower u.s. capital gains rate, and the inclusion of dividends. the results presented here are highly dependent on the participation rate assumption. market variations in participation rates may cause varying buying opportunities for investors in eia contracts. regardless of the type of financial professional, professional and ethical responsibilities to the client should generally result in the same use of these instruments. thus, a nonsecurities licensed insurance agent should be working with a securities-licensed representative to help clients achieve their goals if the agent does not have the appropriate tools or training available. in fact, a planner may be able to replicate a similar tax-advantaged portfolio through the use of a combination of traditional fixed annuities that provide a guarantee-like floor return and options on spiders (spdrs). yet, there is substantial risk to the financial professional as the ability to manage a portfolio of derivative securities is essential to the success of this technique and may place additional liability risks on the planner. it is important to keep in mind that the future regulation of these instruments is still in question. the eia product is an exceptional addition to the product list of the insurance representative if the client and the nonsecurities licensed insurance agent both self select each other because of risk-aversion, tax-aversion, or both. in this case, the product allows the agent to better fill the needs of those clients who are less risk-averse and desire the tax deferability of this product. as the market size of this product line continues to grow, it will become more difficult to regulate this burgeoning industry. firms should take a proactive approach at properly educating representatives how to merchandise these instruments without concern for the regulatory environment. additionally, insurers should be careful in setting participation rates to minimize future solvency difficulties. 341g.a. kuhlemeyer / financial services review 9 (2000) 327–342 acknowledgments the author would like to thank participants at the 1999 academy of financial services conference, christine a. mcclatchey, ken b. cyree, karen eilers lahey, and two anonymous reviewers for their helpful and thoughtful comments. any remaining errors are those of the author. references annuity shopper. (1999). http://www.annuityshopper.com/. august, 23. bodie, z., & crane, d. (1997). personal investing: advice, theory, and evidence. financial analysts journal, 53, 13–23. boyle, p., & lin, x. (1997). bounds on contingent claims based on several assets. journal of financial economics, 46, 383–400. boyle, p., & tse, y. (1990). an algorithm for computing values of options on the maximum or minimum of several assets. journal of financial and quantitative analysis, 25, 215–227. chen h., & volpe, r. (1998). an analysis of personal financial literacy among college students. financial services review, 7, 107–128. edler, h., & rudolph, p. (1999). does retirement planning affect the level of retirement satisfaction? financial services review, 8, 117–127. gerber, h., & shiu, e. (1999). from ruin theory to pricing reset guarantees and perpetual put options. insurance mathematics and economics, 24, 3–14. grable, j., & lytton, r. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163–181. greene, a. (2000). equity index annuities grow steadily. annuity market news, 7, 6. gregory, c. (1997). should equity index annuities be registered? some should not be. national underwriter (life & health/financial services), 101, 22. horowitz, e. (1998). the sales and marketing guide to equity index annuities. cincinnati, oh: the national underwriter company. ibbotson associates. (1999). stocks, bonds, bills, and inflation database. chicago, il: ibbotson associates. jackson national life. (1999). the interest history of jackson national life’s annuities. http://www.jnl.com/ consumer/products/xadv2159.htm. august, 23. jones, c., & wilson, j. (1987). stocks, bonds, paper, and inflation: 1870–1985. the journal of portfolio management, 14, 20–24. koco, l. (2000). ’99 equity index annuity sales topped $5 billion. national underwriter (life & health/financial services), 104, 1. milliman & robertson, inc.,. (1997). equity index annuity product feature comparison. november, 24, 1–17. nelken, i. (1996). the handbook of exotic options. chicago, il: irwin professional publishers. puertz, j. (1997). should equity index annuities be registered? some should be. national underwriter (life & health/financial services), 101, 22. siegel, l., & montgomery, d. (1995). stocks, bonds, and bills after taxes and inflation. journal of portfolio management, 21, 17–25. thrift savings plan for federal employees. (1999). historical annuity rate index: 1989 to present. http:// www.tsp.gov/curinfo/annuity_history.html. august, 23. wilson, j., & jones, c. (1987). a comparison of annual common stock returns: 1871–1925 with 1926–85. journal of business, 60, 239–258. yuh y., hanna, s., & montalto, c. (1998). mean and pessimistic projections of retirement adequacy. financial services review, 7, 175–193. zigler, b. (2000). toiling in the investment garden: hedging strategies for investors. aaii journal, 22(8), 10–14. 342 g.a. kuhlemeyer / financial services review 9 (2000) 327–342 financial services review, 33(2) 124 the association of cryptocurrency and the use of alternative financial services gary curnutt1 and david smith2 abstract alternative financial services (afs) have been studied in recent years in terms of how these financial markets are utilized. the products and services include check cashing, pawnshop loans, payday advance loans, electronic cash transmissions, tax refund anticipation arrangements, rentto-own contracts, prepaid debit cards, gift cards, and loans collateralized by automobile titles. cryptocurrency has become part of this afs ecology. the 2023 survey of household economics and decisionmaking collected information on afs use, including the use of cryptocurrency as an afs. this research answered the questions: (a) do users of cryptocurrencies for afs also tend to use them for investments; (b) do users of cryptocurrencies to make payments tend to use them for other afs purposes, and (c) do users of cryptocurrencies to send money to friends and family tend to use them for other afs purposes? creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license. recommended citation curnutt, g., & smith, d. (2025). the association of cryptocurrency and the use of alternative financial services. financial services review, 33(2), 124-143. introduction the rate at which households are considered banked, that is, having access to traditional financial services such as savings and checking accounts, has steadily increased in the united states from 2011 to 2021 (federal deposit insurance corporation, 2021). the increase was from 91.8% to 95.5% of households banked. there are still unbanked households. underbanked households also exist, with some access to but not full use of traditional sources of savings, investments, payment methods, and credit (birkenmaier & fu, 2023). the unbanked and underbanked use alternative financial services that are generally less efficient, harder to access, and more expensive than traditional 1 western carolina university, cullowhee,, nc, usa. 2 corresponding author (smithdavid@wcu.edu). western carolina university, cullowhee, nc, usa. services from banks, credit unions, and other financial institutions (cfpb 2016). since almost 5% of u.s. households face these challenges, it is interesting to understand their use of cryptocurrencies as alternative financial instruments in services and as investments. alternative financial services (afs) have been studied in recent years regarding how these financial markets are utilized (barcellos & zamarro, 2021; birkenmaier & fu, 2023; fan et al., 2024). the products and services include check cashing, pawnshop loans, payday advance loans, electronic cash transmissions, tax refund anticipation arrangements, rent-to-own contracts, prepaid debit cards, gift cards, and loans collateralized by automobile titles (cfpb 2016). https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ curnutt and smith 125 cryptocurrency has become part of this afs ecology (board of governors of the federal reserve system, 2022). for background on the alternative financial service providers before cryptocurrency use, see prager’s 2014 article “determinants of the locations of alternative financial service providers.” investing in cryptocurrency has two distinct sets of market participants. investment professionals, both institutional and sophisticated individuals, use cryptocurrency as an alternative asset class as a portfolio enhancement to hedge against currency volatility, stock market risk, geopolitical instability, and global economic policy stratification (almeida & gonçalves, 2023). the other set of participants consists of individual investors acting more as gamblers (or speculators) attempting to anticipate where the other investors will peg their prices either in a rising market or a downward-trending environment, thus potentially profiting from the movement of the cryptocurrency market prices (roza et al., 2023). the cryptocurrency market capitalization exceeds $1.5 trillion, with over 14,000 separate coins tracked that are available through over 1,000 exchanges worldwide, providing ample room for trading (coingecko, 2024). cryptocurrency investing activity exists both in traditional financial service institutions and in an ecosystem that is an alternative to mainstream investment services. important to the growth of cryptocurrency as an investment (as well as other uses) is the access provided to the public with the creation of crypto exchanges. these are businesses that facilitate exchanging fiat currencies for cryptocurrencies, where anyone can open an account, deposit dollars, buy cryptocurrency, and do the reverse. exchanges allow trade against other cryptocurrencies and fiat currencies, such as the dollar or yen. fees range from .1% to .5% of transactions. traditional brokerage firms also now provide these services (giudici et al., 2020). the retail public can also access cryptocurrency through atms and other kiosk-type arrangements, paying either a fee (generally 1%) or purchasing at rates higher than exchange rates to incorporate a fee (brown, 2023). hypotheses using 2022 data from the survey of household economics and decisionmaking we formulate the following hypotheses: h1. holding cryptocurrencies as an investment is positively associated with using other afss. h2. using cryptocurrencies to make payments is positively associated with using other afss. h3. using cryptocurrencies to send money to friends or family is positively associated with using other afss. by exploring these hypotheses, pathways toward increasing access to financial services for banked and underbanked households can be supported through policy decisions and financial institution actions. this understanding will assist in relieving some of the economic burden felt by those having to use afs services and products because of current barriers and behaviors. similarly, greater coherence in the investment use of cryptocurrency could be the result of understanding the potentially suboptimal investment decisions (kim et al., 2023). the use of cryptocurrency as an afs cryptocurrency has several attributes that make it a compelling alternative financial services product. first, cryptocurrency allows for transaction privacy (houy et al., 2024). privacy is possible because the value represented by the cryptocurrency is stored on a decentralized ledger, generally a blockchain, without identifiable ownership data. this storage unit is called a wallet. value is transferred via anonymous instructions to transfer value from the sender’s wallet to the receiver’s. a second reason cryptocurrency is used is its perceived safety, provided the user selects a safe option suitable to their needs (houy et al., 2024). third, transferring value is fast and inexpensive. fourth, cryptocurrency is not part of the traditional banking system and is thus more acceptable to those distrusting banks and bank regulators. for these and other reasons, cryptocurrency may be the financial instrument of choice for both the financial services review, 33(2) 126 sender and receiver, including for illegal transactions such as money laundering (wronka, 2022). besides direct transfers of cryptocurrency between senders and receivers, cryptocurrency can be used via an intermediated transaction, such as buying an amazon gift card with cryptocurrency, which can then be used exclusively at amazon online for purchases. amazon does not accept cryptocurrencies directly, but this method is essentially the same. services such as coinsbee.com provide this service for amazon and many other retailers, as well as prepaid mastercard and visa cards (coinsbee, 2024). coinsbee.com will sell you a hotels.com gift card if you want to book a hotel stay. buy a doordash card to pay for your latenight meal delivery. exchanges are available where people can buy and sell cryptocurrencies using fiat currencies such as the u.s. dollar. behavioral finance aspects of the use of cryptocurrencies as an afs and investment using cryptocurrency as an alternative to traditional financial instruments may be a logical, efficient, and effective transaction method and investment instrument. it also may be used based on sub-optimal decisionmaking due to behavioral finance factors. these factors describe potentially irrational financial decisions. for example, one reason for using cryptocurrency is distrust of banks (board of governors, 2022). if the distrust is without rational support, perhaps this decision to use cryptocurrency is irrational, and thus, the behavior of using cryptocurrency for transactions is suboptimal. since people are generally lossaverse, a behavioral finance concept, this distrust may be based on that behavioral finance factor (sokol-hessner & rutledge, 2018). the lived experiences of the unbanked and underbanked households include high overdraft fees at traditional banks. having had to pay these fees creates avoidance behaviors that result in utilizing these afss even if traditional services are available (dlugosz et al., 2021). yes, the distrust of banks may be rational, perhaps experientially based, so it is essential not to be rash when declaring irrationality. herding behavior is a behavioral finance concept prevalent in investing in and using cryptocurrencies (omane-adjepong et al., 2021). herding is when a person follows the crowd, taking cues from others, with no leader necessary. one reason for herding is a natural tendency to fear missing out, sometimes made into the acronym fomo. media attention on cryptocurrencies is intense and constant, with thirty or more articles published daily being common (lee & jeong, 2023). this attention attracts new participants in this market, both cryptocurrency investors and users of cryptocurrencies as an afs. the entry of new investors, likely less informed, coupled with the information and misinformation in the media reports, contributes to the volatility of cryptocurrency markets (lee & jeong, 2023). in auction markets like cryptocurrency exchanges, good news drives prices higher than fundamentals suggest, while bad news has the opposite effect. this mispricing can be attributed to behavioral finance factors such as overconfidence, recency bias, anchoring, confirmation bias, disposition effect, loss aversion, and risk aversion (thampanya et al., 2020). media attention can trigger another emotional response that may drive irrational behavior: narrative influence. this is a powerful influence associated with storytelling. the stories in the media encompass not only what the market is doing in terms of pricing but also the personal stories of individual winners and losers. this type of information can affect the readers’ emotions and result in sub-optimal decisionmaking (shaffer et al., 2018). confirmation bias, where a person pays attention only to the stories that fit with their current beliefs, can lead them to make financial decisions without considering all the alternatives and consequences. behavioral finance and literacy financial literacy is an important variable regarding behavioral finance results. there is much discussion on how to measure financial literacy, including questions to be asked, the number of questions needed for a valid answer, and the saliency of the questions in context, among others (ouachani et al., 2021). the shed study used what is commonly called the big three questionnaire developed in 2008 (lusardi & mitchell, 2014) and used by many large group curnutt and smith 127 studies (ouachani et al., 2021). financial literacy is positively correlated with financial decisions in that the more literate the individual is, the better their decisions. better decision-making was based on more rational considerations, while worse decisions were based on less rational behaviors (kumari, 2020). in other words, the more literate individuals were less susceptible to the behavioral finance concepts that would lead them to less optimal financial decisions. risk tolerance the volatility of cryptocurrency is high compared to many other investments (baur & dimpfl, 2021), so by definition, it is risky, with risk being deviance from the mean. as an asset, price fluctuations are affected by size premium, attention-driven overreaction momentum, and familiarity (liu et al., 2022). it would seem, then, that risk tolerance, that is, an individual's susceptibility to making decisions during volatility downwards, would play a large role in participation in cryptocurrency as an investment. risk tolerance is measured in many ways, from self-reported (willingness to take the risk) to objective and subjective lengthy questionnaires (omanovic & zaimovic, 2024). risk tolerance measurements are designed to understand how an individual views risk and risk’s potentiality. three categories of measuring instruments are prevalent: a) propensity measures, b) stated preferences, and c) revealed preferences, with all three having validated results. (eun & grable, 2024). however, due to behavioral finance concepts such as loss aversion, emotional responses, overconfidence, confirmation bias, recency bias, and others, what people confirm a priori is not constant with actual decisions when losses mount, or gains are excessive (guillemette & finke, 2014). the shed study used selfreporting for stated preferences. to mitigate the effects of behavioral finance factors on decisionmaking, an individual needs to take proactive steps to understand their risk perception and identify biases they may have, such as overconfidence and anchoring. this can be done through education (financial literacy) and consulting with professional financial advisors (coaching) (almansour et al., 2023). formal coursework has been shown to reduce behaviors such as the disposition effect, which is the tendency to hold losing investments too long and sell winning investments too early (paraboni & da costa, 2021). relationship between afs use and cryptocurrency to discuss the relationship between the use of afs and the use of cryptocurrency in any manner requires a dissection of the reasons people use afs in the first place. then, it is possible to associate the perceived benefits of cryptocurrency with the current uses of afs. traditional financial products and services have several drawbacks for the unbanked and underbanked, making afss more attractive. first, there may not be traditional banks physically or virtually convenient to the person. a prepaid debit card can be purchased and used in many locations at many times of the day or night, if not constantly. second, the person may not be able to carry the minimum balance required at a bank for an atm/debit card. the person can deposit to a prepaid card when possible and carry a zero balance until they have cash. third, the traditional payment systems contain lags that inconvenience those needing immediate access to their paychecks, thus check-cashing services. this is also true for others who need money now but will not be paid for work until next week, thus payday loans. waiting for a tax refund may not be acceptable, so the refund anticipation loan scheme exists. it is likely that except for those transactions where the counterparty requires cryptocurrency, these afs users will not be cryptocurrency users either for transactions or investments. they don’t have money to invest, and their current afs use is working for them. (carmona, 2022). another narrative regarding afs use is not based on inconvenience or lack of wherewithal. instead, there is a cohort that uses afs because, for some, many, or all transactions, they do not want to use traditional providers and instruments. some people do not trust banks for various reasons. others may want to avoid banks, not out of distrust, but perhaps to avoid reporting requirements or paper trails. cryptocurrency may be an added means for transactions and investing in this case due to the privacy available and financial services review, 33(2) 128 perceived efficiency in terms of speed and cost of the transaction. investment may be due to either intellectually sound investment allocation purposes or speculation aspects that appeal to those interested in quick wealth accumulation (houy et al., 2024). combining the elements of a person already disposed to using afs for transactions with the potential investment gains, it is reasonable to consider that a person would hold wealth in cryptocurrency and use it as needed as an afs, similar to a stock portfolio occasionally partially liquidated into their checking account for spending. methods data this paper uses data from 2022 survey of household economics and decisionmaking (shed). shed is a nationally representative, when properly weighted, survey conducted by the federal reserve board which measures the economic well-being of u.s. households and can help identify potential risks to their finances. we used the 2022 wave due to a more robust set of questions surveying cryptocurrencies compared to older waves of the survey. however, this limited the scope of the survey to just a crosssection. future research would benefit from exploring longitudinal data spanning several future waves of shed to observe how the association of cryptocurrency adoption is associated with the use of existing alternative financial services over time. the full sample of the 2022 survey has a total of 11,667 respondents, of which our analysis included 9,326 respondents. table 1 shows the summary statistics for five demographic control variables for both the full sample and our analysis sample. a two-sample ttest does show statistically significant differences between our analysis sample and the full sample. notably, our analysis sample has more respondents who report being: male, married, white, having a four-year degree, and an older age. table 1. summary statistics for 2022 survey of household decisionmaking and analysis samples 2022 shed full sample analysis sample mean se mean se female 0.4892 0.0046 0.4700 0.0051** married 0.5771 0.0045 0.6047 0.0050 *** white 0.6908 0.0042 0.7066 0.0047 ** four-year degree 0.4304 0.0045 0.4529 0.0021*** age 51.9671 0.1626 52.823 0.1772 *** note: n of 11,667 for shed sample & 9,326 for analysis sample. samples are weighted using weights provided by the shed. two-sample t-test: ** denotes statistical difference from the shed sample mean at the 5% level of significance. *** denotes statistical difference from the shed sample mean at the 1% level of significance. “se” denotes standard error. starting with a full sample of 11,667 respondents, observations were dropped for missing or incomplete responses to survey questions that were used in the model. specifically, observations were dropped from such responses to: “what is the approximate total amount of your household's savings and investments?” (1,841 dropped), and “where do you think your credit score falls?” (500 dropped). this left us with an analysis sample of 9,326 observations. table 2 shows descriptive statistics for our analysis sample. in our analysis sample, only about 8.7%, 811 respondents, report holding cryptocurrency as an investment. while about 1.4%, 129 respondents, report using cryptocurrency to make a purchase and about 1.3%, 118 respondents, report using cryptocurrency to send money to friends or family. these numbers illustrate the significant difference in adoption of cryptocurrency as an alternative to conventional financial infrastructure. about 96.6% of respondents report having a checking/savings/ or money market account. however, cryptocurrency adoption seems to be somewhat comparable to other alternative financial services. about 7.7% of respondents in our analysis sample have purchased a money order from a service provided that wasn’t a bank. curnutt and smith 129 about 5% of respondents have cashed a check from a service provider that wasn’t a bank. about 2.2% of respondents have taken out a payday loan, about 1.6% have taken out a pawnshop or auto title loan, about 0.8% have obtained a cash advance on their tax-refund and about 9.6% have paid an overdraft fee on a bank account. appendix a contains a table detailing each variable and how they were constructed. table 2. descriptive statistics of explanatory variables analysis samples mean se bought or held cryptocurrency as an investment 0.08696 0.0029 used cryptocurrency to make a purchase 0.0138 0.0012 used cryptocurrency to send money 0.0127 0.0012 has a checking/savings/money market account 0.9661 0.0019 purchased money order outside of bank 0.0772 0.0028 cashed check outside of bank 0.0507 0.0023 taken out a payday loan 0.0216 0.0015 taken out a pawnshop/title loan 0.0157 0.0013 obtained a tax refund advance 0.0080 0.0009 paid an overdraft fee on a bank account 0.0959 0.0031 willing to take financial risk 0 not at all 0.1629 0.0038 1 0.0587 0.0024 2 0.0833 0.0029 3 0.1092 0.0032 4 0.0937 0.0030 5 0.1855 0.0040 6 0.1047 0.0032 7 0.1079 0.0032 8 0.0579 0.0024 9 0.0123 0.0011 10 – very willing 0.0240 0.0015 race/ethnicity white, non-hispanic 0.7066 0.0047 black, non-hispanic 0.0986 0.0031 other, non-hispanic 0.0455 0.0022 hispanic 0.1189 0.0033 two or more races, non-hispanic 0.0303 0.0018 female 0.4701 0.0052 married 0.6048 0.0051 total savings & investments financial services review, 33(2) 130 under $50,000 0.4175 0.0051 $50,000 $99,999 0.1284 0.0035 $100,000 $249,999 0.1403 0.0036 $250,000 $499,999 0.1050 0.0032 $500,000 $999,999 0.0967 0.0031 $1,000,000 or more 0.1122 0.0033 where do you think your credit score falls? very poor 0.0272 0.0017 poor 0.0454 0.0022 fair 0.1067 0.0032 good 0.2351 0.0042 excellent 0.5856 0.0051 educational attainment less than high school 0.0390 0.0019 high school diploma 0.1885 0.0039 some college 0.3195 0.0047 bachelor's degree 0.4529 0.0051 report at doing least okay financially 0.7596 0.0044 number of financial literacy questions answered correctly 0 0.0714 0.0027 1 0.1327 0.0035 2 0.2630 0.0046 3 0.5328 0.0052 employed 0.6262 0.0050 age 52.8235 0.1771 note: n of 9,326. sample is weighted using weights provided by the 2022 survey of household decision making. “se” denotes standard error. dependent variables the dependent variables used in our model are variables that ask the respondent about their use or ownership of cryptocurrency in the previous twelve months. first, survey question “s16_a” is a dichotomous variable that asks the question, “in the past year, have you done the following with cryptocurrency, such as bitcoin or ethereum? (bought or held as an investment.)” responses of yes are coded as a 1, while responses of “no” are coded as a 0. our second dependent variable is survey question “s16_b” which asks, “in the past year, have you done the following with cryptocurrency, such as bitcoin or ethereum? (used to buy something or make a payment.)” responses of yes are coded as a 1, while responses of “no” are coded as a 0. our final dependent variable is survey question “s16_c” which asks, “in the past year, have you done the following with cryptocurrency, such as bitcoin or ethereum? (used to send money to friends or family.)” likewise, responses of yes are coded as a 1, while responses of “no” are coded as a 0. curnutt and smith 131 explanatory variables the explanatory variables consist of several variables that capture the effects of other alternative financial services as well as several variables that control for other effects that may otherwise impact the association of holding cryptocurrencies and the use of afss. • survey question “bk1” asks the respondent, “do you (and/or your spouse or partner) have a checking, savings, or money market account.)” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “bk2_a” asks the respondent, “in the past 12 months, did you (and or your spouse or partner) purchase a money order from a place other than a bank?” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “bk2_b” asks the respondent, “in the past 12 months, did you (and or your spouse or partner) cash a check at a place other than a bank?” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “bk2_c” asks the respondent, “in the past 12 months, did you (and or your spouse or partner) take out a payday loan or payday advance?” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “bk2_d” asks the respondent, “in the past 12 months, did you (and or your spouse or partner) take out a pawnshop loan or an auto title loan?” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “bk2_e” asks the respondent, “in the past 12 months, did you (and or your spouse or partner) obtain a tax refund advance to receive your refund faster?” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “bk2_f” asks the respondent, “in the past 12 months, did you (and or your spouse or partner) pay an overdraft fee on a bank account?” responses of yes are coded as a 1, while responses of “no” are coded as a 0. • survey question “fl0” is included in the model to proxy for the respondent’s risk tolerance. fl0 is an ordinary variable ranging from 0 to 10. the question asks, “on a scale of zero to ten, where zero is not at all willing to take risk and ten is very willing to take risks, what number would you be on the scale?” • survey question “ppethm” is a categorical variable that asks the respondent to provide their self-reported “race/ethnicity”. we include this in our model as a proxy to control for the cultural effects that differences in race and ethnicity may have on the adoption of cryptocurrency. respondents can answer 1 for “ white, non-hispanic”, 2 for “black, non-hispanic”, 3 for “other, non-hispanic”, 4 for “hispanic”, or 5 for “2+ races, non-hispanic”. • survey question “ppgender” is a dichotomous variable that asks the respondent to self-report their gender. respondents can answer 1 for “male” or 2 for “female”. we recode this variable as 0 for “male” respondents and 1 for “female” respondents. • survey question “ppmarit5” is an ordinal variable ranging from 1 to 5. respondents can answer 1 for “now married”, 2 for “widowed”, 3 for “divorced”, 4 for “separated”, or 5 for “never married”. we recoded this variable as 0 for responses 2, 3, 4, or 5. this resulted in a dichotomous variable where a value of 0 is not married and 1 is married. • survey question “pps0596” is an ordinal variable where respondents are asked to self-report their savings and investments on a range from 1 to 7. respondents can answer 1 for “less than $50,000”, 2 for “$50,000 to $99,999”, 3 for “$100,000 to $249,999” 4 for “$250,000 to $499,999”, 5 for “$500,000 to $999,999”, 6 for “$1,000,000 or more” or 7 for “not sure”. financial services review, 33(2) 132 responses of 7 are dropped from the sample. • survey question “ppfs1482” is an ordinal variable that asks the respondent to selfreport where they think their credit score falls on a range from 1 to 6. respondents can answer 1 for “very poor”, 2 for “poor”, 3 for “fair”, 4 for “good”, 5 for “excellent”, 6 for don’t know”. responses of 6 are dropped from the sample. • survey question “educ_4cat” is an ordinal variable that asks the respondent to self-report their level of educational attainment. respondents can answer 1 for “less than a high school degree”, 2 for “high school degree or ged”, 3 for “some college/technical or associate’s degree”, or 4 for “bachelor’s degree or more”. • survey question “at_least_okay” is a dichotomous variable that asks the respondent to self-report their subjective financial wellbeing. the respondent can answer 1 for “yes” or 0 for "no”. shed includes three questions about basic financial concepts. “fl2” asks the question, “do you think the following statement is true or false? ‘buying a single company’s stock usually provides a safer return than a stock mutual fund.’” the respondent can answer 1 for “true”, 2 for “false”, or -2 for "don’t know.” “fl4” asks the question “imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account?” the respondent can answer 1 for “more than today”, 2 for “exactly the same”, or 3 for “less than today.” “fl5” asks the respondent “suppose you had $100 in a savings account and the interest rate was 2% per year. after 5 years, how much do you think you would have in the account if you left the money to grow?” the respondent can answer 1 for “more than $102”, 2 for “exactly $102”, 3 for “less than $102” or -2 for “don’t know”. using these three questions, we construct an ordinal variable that measures the respondent’s total number of correct answers for all the financial literacy questions above (fl2, fl4, fl5). this variable is 0 if the respondent answered 0 questions correct, 7.14% of respondents, 1 if the respondent answered 1 question correctly, 13.27% of respondents, 2 if the respondent answered 2 questions correctly, 26.3% of respondents, and 3 if the respondent answered all three questions correctly, 53.28% of respondents. survey question “ppemploy” is an ordinal variable that asks the respondent to report their current employment status. respondents can answer 1 for “working full-time”, 2 for “working part-time”, or 3 for “not working”. we create a dichotomous variable that is coded as 1 if the respondent reported 1 on ppemoly and 0 otherwise. survey question “ppage” is a continuous variable that allows the respondent to enter current age. model this paper examines the association between the use of alternative financial services and the intention of holding or using cryptocurrency for: investment purposes, making payments, and for using to send to friends or family. to accomplish this, we use three separate logistical regression analyses; the three resulting models are as follow: where the latent variable, 𝐼𝑖; is the unobserved net benefit of holding cryptocurrency as an investment for individual 𝑖. the variable 𝐼𝑖 is the observed dichotomous decision of individual i to hold cryptocurrency as an investment. likewise, the latent variable, 𝑃𝑖; is the unobserved net benefit of using cryptocurrency to buy something or make a payment for individual 𝑖. the variable 𝑃𝑖 is the observed dichotomous decision of individual i to use cryptocurrency to buy 𝐼𝑖 = 𝛼0 + 𝛼1𝐴𝐹𝑆𝑖 + 𝛼𝑘𝑋𝑖 +𝜀𝑖 [1.1] 𝑃𝑖 = 𝛽0 + 𝛽1𝐴𝐹𝑆𝑖 + 𝛽𝑘𝑋𝑖 +𝜃𝑖 [1.2] 𝑅𝑖 = 𝛾0 + 𝛾1𝐴𝐹𝑆𝑖 + 𝛾𝑘𝑋𝑖 +𝜆𝑖 [1.3] 𝐼𝑖 = { 1 if 𝐼𝑖 1 > 0 0 𝑖f 𝐼 ≤ 0 } 𝑃𝑖 = { 1 if 𝑃𝑖 > 0 0 if 𝑃𝑖 ≤ 0 } 𝑅𝑖 = { 1 if 𝑅𝑖 > 0 0 if 𝑅𝑖 ≤ 0 } curnutt and smith 133 something or make a payment, for individual 𝑖. the latent variable, 𝑅𝑖; is the unobserved net benefit of using cryptocurrency to send money to friends or family for individual 𝑖. the variable 𝑃𝑖 is the observed dichotomous decision of individual i to use cryptocurrency to send money to friends or family, for individual 𝑖. afs is a matrix of the seven variables that indicate the use of alternative financial services. these variables include having a bank account, purchasing money orders from places other than a bank, cash check at places other than a bank, use of a payday loan, use of a pawnshop or title loan, use of a tax refund advance, and pay overdraft fees on a bank account. x is a matrix of other explanatory variables, which include financial literacy, credit score, risk tolerance, subjective financial wellbeing, marital status, employment status, total wealth, education, age, ethnicity, and gender. for model 1.1, 𝛼1 is a vector of parameters to be estimated for key explanatory variables, afs, and 𝛼𝑘 is a vector of parameters to be estimated for the other explanatory variables. the error term 𝜀𝑖 follows a normal distribution. likewise, in model 1.2, 𝛽1 is a vector of parameters to be estimated for key explanatory variables, afs, and 𝛽𝑘 is a vector of parameters to be estimated for the other explanatory variables. the error term 𝜃𝑖 follows a normal distribution. in model 1.3, 𝛾1 is a vector of parameters to be estimated for key explanatory variables, afs, and 𝛾𝑘 is a vector of parameters to be estimated for the other explanatory variables. the error term 𝜆𝑖 follows a normal distribution. for each model, marginal effects are estimated to show associations between the explanatory variables and the observed dependent variables. results this paper provides empirical evidence on how the adoption of cryptocurrency, or investment therein, is associated with using other afss. table 3 summarizes the marginal effects for each logistical regression model. holding cryptocurrencies as an investment our results show that holding cryptocurrencies as an investment is associated with using other afss. individuals who reported holding cryptocurrencies as an investment over the last 12 months had a 2.68 percent higher probability of having purchased a money order from a source other than a bank during the same duration of time than those who did not. individuals who held cryptocurrencies as an investment also had a 3.33 percent higher probability of having taken out a payday loan in the last 12 months. likewise, individuals who invested in cryptocurrencies also had a 4.2 percent higher probability of taking out a pawnshop or auto title loan in the last 12 months than those who did not hold cryptocurrencies as an investment. holding cryptocurrencies as an investment was also associated with a 3.82 percent higher probability of having paid an overdraft fee on a bank account over the last 12 months. our results also show a relationship between risk tolerance as measured by the self-reporting variable fl0 and investing in cryptocurrency. at lower levels of risk tolerance, individuals demonstrate a modestly increased likelihood of cryptocurrency investment. specifically, at level 2, the marginal effect is 0.0436, indicating a noticeable increase compared to those unwilling to take risks (level 0). at levels 3 and 4, marginal effects are 0.0258 and 0.0315, respectively, suggesting a gradual strengthening of the relationship. the association grows stronger as risk tolerance increases further. at level 5, the marginal effect rises to 0.0422. a marked increase occurs at level 6, where the marginal effect jumps to 0.0704. this upward trend continues, with level 7 showing a marginal effect of 0.1016. at level 8, the likelihood substantially increases, reflected by a marginal effect of 0.1244. the highest associations are observed at levels 9 and 10. level 9 exhibits the strongest relationship, with a marginal effect of 0.167. even at the maximum level of risk tolerance (level 10), the marginal effect remains high at 0.1221, highlighting a significant propensity among highly risk-tolerant individuals to hold cryptocurrencies as investments. using cryptocurrencies as a payment the results for model 1.2 show a statistically significant association for several variables measuring the use of afss with using cryptocurrency to make a payment over the last financial services review, 33(2) 134 12 months. using cryptocurrency to make a payment in the last 12 months was associated with a 0.9 percent higher probability of having purchased a money order from a service provider other than a bank, compared to those who didn’t use crypto to make payments in the last 12 months and ceteris paribus. using crypto as a payment was also associated with a 1.21 percent higher probability of cashing a check from a provider outside of a bank. using crypto to make payments was associated with a 1.32 percent higher probability of obtaining an advance on an anticipated tax refund. we also find a positive association, 1.48 percent, with using crypto to make payments and paying an overdraft fee on a bank account. we find an association with risk tolerance, although only at some of the higher levels of selfreported risk tolerance. compared to those who report they are not at all willing to take risk, those who report a risk tolerance of 6, 8 , or 10 (very willing), were all associated with a higher probability of using crypto to make payments in the last 12 months, 1.16 percent, 2.02 percent, and 3.93 percent, respectively. those who identify as hispanic, had a 0.9 percent higher probability of using cryptocurrencies to make payments in the last 12 months compared to those who identified as white. identifying as female was associated with a 1.18 lower probability of using crypto to make payments. those who had investments between $50,000 and $99,000 had a 1.05 percent higher probability of using crypto to make payments than those who had less than $50,000 in savings and investments, but none of the other levels of investments had a significant association. lastly, we find age to have a negative association, 0.01 percent, and employment to have a positive association, 0.81 percent with using crypto to make payments. using cryptocurrencies to send money the results for model 1.3 show several statistically significant associations with the use of afss and using cryptocurrencies to send money to friends or family within the last 12 months. compared to those who didn’t use crypto to send money to friends or family, those who did had 0.62 higher probability of purchasing a money order from a provider other than a bank in the last 12 months. sending crypto was also positively associated with cashing a check outside a bank by 0.86 percent. our results show that using crypto to send money is associated with a 0.92 percent higher probability of taking out a payday loan, compared to those who didn’t use crypto to send money. we find that the use crypto to send money is associated with a 2.24 percent higher probability of obtaining a tax refund advance. further, we find sending money with crypto is associated with a 0.66 percent higher probability of having paid an overdraft fee on a bank account. model 1.3 shows some association with selfreported risk tolerance and sending money with crypto. compared to those who were not at all willing to take risk, those who reported a risk tolerance of 1,5,8, and 10 all had a higher probability of using crypto to send money, 1.19 percent, 0.78 percent, 1.72 percent, and 5.05 percent, respectively. compared to those who identified as white, those who reported being in any other racial or ethnic identification were more likely to use cryptocurrencies to send money to friends or family. reporting being black and non-hispanic was associated with a 1.7 percent higher probability of using crypto to send money. reporting being other non-hispanic was associated with a 1.61 percent higher probability of using crypto to send money. reporting being hispanic was associated with a 0.63 percent higher probability of using crypto to send money, compared to those who reported being white and non-hispanic. we also find associations with some levels of self-reported credit score and sending crypto to friends or family in the last 12 months. compared to those who report that they have a very poor credit score, those who report having a poor credit score had a 2.14 percent lower probability of using crypto to send money to friends or family. while reporting having an excellent credit score was associated with a 1.89 percent lower probability of using crypto to send money, compared to those who report a very poor credit score. lastly, we find that answering all three financial literacy questions correctly was associated with a 0.98 percent lower probability of sending money to friends or family. financial services review, 33(2) 135 table 3. marginal effects of the three models independent variable marginal effect (standard error) crypto as an investment [1.1] (se) crypto as a payment [1.2] (se) crypto to send money [1.3] (se) has a checking/savings/money market account 0.0177 0.0244 -0.0010 0.0065 -0.0077 0.0051 purchased money order outside of bank 0.0268** 0.0123 0.009** 0.0040 0.0062* 0.0036 cashed check outside of bank 0.0182 0.0153 0.0121** 0.0043 0.0086** 0.0041 taken out a payday loan 0.0334* 0.0196 0.0035 0.0059 0.0092* 0.0053 taken out a pawnshop/title loan 0.042* 0.0238 0.0066 0.0062 -0.0032 0.0060 obtained a tax refund advance 0.0035 0.0263 0.0132* 0.0074 0.0224*** 0.0055 paid an overdraft fee on a bank account 0.0382*** 0.0104 0.0148*** 0.0041 0.0066* 0.0039 willing to take financial risk 0 not at all (reference) 1 0.0217 0.0137 0.0037 0.0057 0.0119* 0.0064 2 0.0436*** 0.0120 -0.0018 0.0044 0.0074 0.0053 3 0.0258** 0.0105 0.0033 0.0049 0.0036 0.0042 4 0.0315*** 0.0114 -0.0005 0.0043 0.0002 0.0041 5 0.0422*** 0.0093 0.0050 0.0044 0.0078* 0.0046 6 0.0704*** 0.0119 0.0116* 0.0063 0.0038 0.0051 7 0.1016*** 0.0136 0.0022 0.0046 0.0087 0.0061 8 0.1244*** 0.0168 0.0202** 0.0084 0.0172** 0.0081 9 0.167*** 0.0346 0.0376 0.0232 0.0353 0.0265 10 – very willing 0.1221*** 0.0249 0.0393*** 0.0132 0.0505*** 0.0167 race/ethnicity white, non-hispanic black, non-hispanic 0.0061 0.0110 0.0139 0.0055 0.017*** 0.0056 financial services review, 33(2) 136 other, non-hispanic 0.0236* 0.0138 0.0064 0.0063 0.0161** 0.0080 hispanic 0.0154 0.0103 0.009** 0.0042 0.0063* 0.0037 two or more races, nonhispanic 0.0236 0.0180 -0.0007 0.0051 0.0053 0.0068 female -0.0489*** 0.0067 0.0118*** 0.0033 -0.0044 0.0029 married 0.0006 0.0070 -0.0046 0.0033 -0.0035 0.0033 total savings & investments under $50,000 $50,000 $99,999 0.0064 0.0106 0.0105* 0.0058 0.0078 0.0054 $100,000 $249,999 0.0088 0.0103 0.0018 0.0052 0.0021 0.0054 $250,000 $499,999 -0.0167 0.0110 0.0045 0.0075 0.0013 0.0064 $500,000 $999,999 -0.0158 0.0112 0.0018 0.0070 0.0059 0.0110 $1,000,000 or more -0.0093 0.0131 0.0027 0.0063 -0.0014 0.0052 where do you think your credit score falls? very poor poor 0.0272 0.0209 -0.0148 0.0105 -0.0214** 0.0098 fair 0.0054 0.0169 -0.0094 0.0103 -0.0162 0.0099 good 0.0432** 0.0174 -0.0118 0.0109 -0.0100 0.0105 excellent 0.0448** 0.0176 -0.0161 0.0109 -0.0189* 0.0102 educational attainment less than high school high school diploma -0.0181 0.0211 0.0016 0.0059 0.0021 0.0055 some college 0.0296 0.0211 0.0039 0.0055 0.0035 0.0053 bachelor's degree 0.0307 0.0215 0.0043 0.0055 0.0023 0.0057 report at doing least okay financially -0.0206** 0.0087 0.0018 0.0037 -0.0032 0.0038 number of financial literacy questions answered correctly 0 (reference) 1 0.0014 0.0139 0.0035 0.0057 0.0024 0.0062 2 0.0108 0.0132 -0.0012 0.0054 -0.0034 0.0056 3 0.0458*** 0.0133 -0.0006 0.0053 -0.0098* 0.0055 employed 0.0362*** 0.0098 0.0081* 0.0001 0.0010 0.0036 curnutt and smith 137 age -0.0023*** 0.0003 0.001*** 0.0043 0.0000 0.0001 note: 9,326 observations from the 2022 survey of household decision making. *** denotes statistical significance at the 1% level, ** denotes statistical significance at the 5% level, * denotes statistical significance at the 10% level. discussion and implications of key results the results of this analysis indicate that cryptocurrencies play a role in servicing the financial needs of some consumers, particularly those who face barriers to accessing traditional banking services. the results highlight how cryptocurrencies can function as alternative financial services, particularly for individuals relying on non-traditional banking methods. this demographic often includes the unbanked and underbanked. users of more traditional alternative financial services may also be enticed to use cryptocurrency to fulfill a need that would otherwise be fulfilled by a traditional financial institution. cryptocurrencies have also increasingly become popular as alternative means for financial transactions due to their decentralized nature, potential cost-effectiveness, and faster transaction times compared to traditional financial systems. cryptocurrencies show significant potential as a cost-effective remittance solution. baur and dimpfl (2021) identify lower transaction fees and improved speed as key incentives driving individuals and businesses towards adopting cryptocurrencies for payments. while yermack (2015) highlights cryptocurrencies' potential in improving payment efficiency and transparency, especially across international borders. this study indicates that minority groups, particularly black and hispanic individuals, are more likely to use cryptocurrencies for sending money or making payments, when compared to non-hispanic whites. this finding is consistent in some regards with the findings of chatterjee and yang (2025) who find that ethnic minorities are more inclined to utilize afs. traditional remittance services can be prohibitively expensive and slow, whereas cryptocurrencies can offer faster and cheaper options. the use of cryptocurrencies for remittances can reduce costs for those sending money across borders, thereby supporting the financial wellbeing of these communities. however, unlike chatterjee and yang (2025), who find the likelihood of the use of afs is associated with lower objective financial literacy, we find that holding crypto as an investment is associated positively with answering all three financial literacy questions correctly. zhang et al. (2025) find that other psychological factors may also contribute to crypto ownership as an investment. each of the three models appears to be influenced in varying ways by individuals’ willingness to accept financial risk. those with higher risk tolerance are more inclined to own crypto as an investment, consistent with the volatile nature of the asset class and the high uncertainty associated with potential returns. this group may be motivated by speculative opportunities or by a belief in the long-term value of digital assets. their risk profile aligns with the behaviors of individuals seeking high-reward ventures, often without the safety nets offered by traditional investment vehicles. three dimensions may influence the relationship between risk-taking preference and investing in a highly risky asset class. first is financial literacy in the domain of cryptocurrency investing. a highly cryptocurrency-literate person would need knowledge of the many fundamental aspects of the currency of interest: the potential volatility, the markets, liquidity, etc. absent adequate information and the ability to use it, this investor likely exhibits an overconfidence bias. they are investing while overestimating their knowledge and abilities. this situation begs the question, does anyone have adequate information and ability as a cryptocurrency investor? if a practitioner has clients investing in cryptocurrency, it might be prudent to ask; a discussion of the investment and its market risks is in order. however, it may be that the advisor’s compliance requirements do not allow that financial services review, 33(2) 138 discussion. or, the advisor is not competent to counsel in this area due to their lack of knowledge or ability as a cryptocurrency investor. overconfidence leads to adverse outcomes in athletics, politics, investing, and most human activities. second, investors are susceptible to additional biases that affect decisionmaking. anchoring bias occurs in many human activities where attention is focused on an initial figure. an investor in bitcoin (btc), for example, may have entered the market on 9/10/23 at $24,131, considering this a fair price. another investor may have entered on 3/24/24 at $69,146 and considered that a fair price. if they exhibit anchoring bias, the two investors will likely make very different decisions about the price of btc if today it was at $49,000, with both decisions not based on the reality of the market for btc but based on their belief in a fair price. this anchoring bias has a role in another bias, the disposition effect. this effect is when an investor holds losers too long and sells winners too early compared to what a rational investment plan would dictate. what a loser or winner is in the investor’s mind can be based on that initial anchor rather than the true rational value of the asset. third, investment decisions to buy, sell, or hold are based on the information available to the investor and the ability of the investor to use that information rationally. the investor may have a recency bias and be affected by confirmation bias. with btc, for example, recency bias will cause the investor to pay more attention to what is being reported now versus the history of btc prices and market movements. while it is common to say that the past is not a predictor of the future in investments, a well-informed investor considers history valuable information. confirmation bias is when the investor filters information to pay attention to what confirms their beliefs and avoids contradictory information. if the media mostly says btc is on an upswing, which the investor wants to hear, they will ignore warning signs from other reports. these two biases skew the information being processed towards irrationality and suboptimal decisionmaking. other biases likely play a part in the investment decisions regarding cryptocurrencies. financial planning practitioners should proactively discover which of their clients are cryptocurrency investors. in this way, they can apply their planning knowledge and skills to overcome literacy gaps, counsel on risk-taking behaviors, integrate these investments into an overall plan, and correctly manage their clients’ portfolios given the cryptocurrency investing behavior. practitioners need to become experts in both the investment side and the behavioral aspects of cryptocurrency investing. in contrast, using cryptocurrency for payment or sending money seems less associated with elevated financial risk tolerance. instead, these behaviors may reflect practical concerns, such as access, affordability, and convenience. however, there is still a meaningful link to some levels of risk tolerance, suggesting that even among those using crypto for utility rather than speculation, a certain comfort with novel and less-regulated systems is necessary. these users might be more open to financial innovation and may perceive the trade-offs, such as price volatility, as acceptable in exchange for benefits like lower remittance fees or increased transaction speed. importantly, there is also a notable association between cryptocurrency usage and the use of alternative financial services (afs), such as payday loans, pawnshop loans, and money orders obtained outside of banks. this connection suggests that for many users, cryptocurrency may serve as a complement to or substitute for traditional afs. individuals who rely on afs often do so because of limited access to conventional banking, and their adoption of cryptocurrency may reflect a search for more accessible, affordable, or immediate financial solutions. in this way, cryptocurrencies may function as part of an informal financial ecosystem that services the needs of the unbanked and underbanked. this connection highlights the potential for crypto to address systemic gaps in financial inclusion, while also raising concerns about the vulnerability of these populations to risk and exploitation in unregulated digital markets. the implications are far-reaching. for consumers, the connection between risk preference and crypto behavior underscores the need for curnutt and smith 139 personalized financial education that accounts for individual risk profiles and financial goals. for advisors, it reinforces the importance of discussing crypto investments with clients, not just in terms of returns but also in the context of behavioral and psychological readiness. advisors must be prepared to help clients understand how their risk preferences may influence both their motivations and decision-making processes across different forms of crypto use. for policymakers and regulators, understanding the interplay between risk tolerance and crypto usage supports the design of more nuanced regulations that address the varying needs of different user groups. those engaging in highrisk investment behavior may require stronger consumer protections, while users relying on crypto for payments or remittances might benefit more from improved infrastructure and safeguards against fraud or access issues. references almansour, b. y., elkrghli, s., & al-mansour, a. y. (2023). behavioral finance factors and investment decisions: a mediating role of risk perception. cogent economics & finance, 11(2), 1–20. https://doi.org/10.1080/23322039.2023. 2239032 almeida, j., & gonçalves, t. c. (2023). portfolio diversification, hedge and safe-haven properties in cryptocurrency investments and financial economics: a systematic literature review. journal of risk and financial management, 16(1), 1–25. https://doi.org/10.3390/jrfm16010003 barcellos, s. h., & zamarro, g. (2021). unbanked status and use of alternative financial services. journal of pension economics & finance, 20(4), 468–481. https://doi.org/10.1017/s147474721900 0052 birkenmaier, j., & fu, q. (2023). bank account access methods, household financial wellbeing, and alternative financial services. journal of poverty, 27(1), 33–59. https://doi.org/10.1080/10875549.2021. 2010863 baur, d. g., & dimpfl, t. (2021). the volatility of bitcoin and its role as a medium of exchange and a store of value. empirical economics, 61(5), 2663–2683. https://doi.org/10.1007/s00181-02001990-5 brown, t. (2023). bitcoin atms: what, why, and where. business review at berkeley. https://businessreview.studentorg.berkel ey.edu/bitcoin-atms-what-why-andwhere/ carmona, t. (2022, may 19). debunking the narratives about cryptocurrency and financial inclusion. brookings. https://www.brookings.edu/articles/debunking-the-narratives-about-cryptocurrency-and-financial-inclusion/ cfpb. (2016). supplemental findings on payday, payday installment, and vehicle title loans, and deposit advance products. https://files.consumerfinance.gov/f/documents/supplemental_report_060116.pdf chatterjee, s., & chang, y. (2025). utilization of alternative financial services and the role of financial capability. journal of consumer affairs, 59(1), e12614. coingecko. (2024). 2024 q1 crypto industry report. https://www.coingecko.com/research/publications/2024-q1-crypto-report coinsbee. (2024). buy gift cards with crypto. https://www.coinsbee.com/en/ dlugosz, j., melzer, b., & morgan, d. p. (2021). who pays the price? overdraft fee ceilings and the unbanked. frb of new york staff report 973. federal reserve bank of new york. https://papers.ssrn.com/sol3/papers.cfm? abstract_id=3876646 eun, j. k., & grable, j. e. (2024). a comparison of financial risk-tolerance assessment methods in predicting subsequent risk tolerance and future portfolio choices. risks, 12(11), 170. doi: 10.3390/risks12110170 https://doi.org/10.1080/23322039.2023.2239032 https://doi.org/10.1080/23322039.2023.2239032 https://doi.org/10.3390/jrfm16010003 https://doi.org/10.1080/10875549.2021.2010863 https://doi.org/10.1080/10875549.2021.2010863 https://doi.org/10.1007/s00181-020-01990-5 https://doi.org/10.1007/s00181-020-01990-5 https://businessreview.studentorg.berkeley.edu/bitcoin-atms-what-why-and-where/ https://businessreview.studentorg.berkeley.edu/bitcoin-atms-what-why-and-where/ https://businessreview.studentorg.berkeley.edu/bitcoin-atms-what-why-and-where/ https://www.brookings.edu/articles/debunking-the-narratives-about-cryptocurrency-and-financial-inclusion/ https://www.brookings.edu/articles/debunking-the-narratives-about-cryptocurrency-and-financial-inclusion/ https://www.brookings.edu/articles/debunking-the-narratives-about-cryptocurrency-and-financial-inclusion/ https://www.coingecko.com/research/publications/2024-q1-crypto-report https://www.coingecko.com/research/publications/2024-q1-crypto-report https://www.coingecko.com/research/publications/2024-q1-crypto-report https://www.coinsbee.com/en/ https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3876646 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3876646 financial services review, 33(2) 140 fan, l., green, l., & park, n. (2024). financial stressors and alternative financial service use: extending the abc-x model of family stress. international journal of consumer studies, 48(1), 1–15. https://doi.org/10.1111/ijcs.13002 federal deposit insurance corporation. (2021). fdic national survey of unbanked and underbanked households. https://www.fdic.gov/analysis/household-survey/2021report.pdf board of governors of the federal reserve system. (2022). survey of household economics and decisionmaking. federal reserve. https://www.federalreserve.gov/consum erscommunities/shed.htm giudici, g., milne, a., & vinogradov, d. (2020). cryptocurrencies: market analysis and perspectives. journal of industrial and business economics, 47, 1–18. https://doi.org/10.1007/s40812-01900138-6 guillemette, m., & finke, m. (2014). do large swings in equity values change risk tolerance? journal of financial planning, 27(6), 44. https://www.proquest.com/trade-journals/do-large-swings-equity-valueschange-risk/docview/1535617637/se2?accountid=14968 houy, s., schmid, p., & bartel, a. (2024). security aspects of cryptocurrency wallets: a systematic literature review. acm computing surveys, 56(1), 1–31. https://doi.org/10.1145/3596906 kim, k. t., hanna, s. d., & lee, s. t. (2023). investment literacy, overconfidence, and cryptocurrency investment. financial services review, 31(2/3), 121–132. https://doi.org/10.61190/fsr.v31i2/3.353 0 kumari, d. a. t. (2020). the impact of financial literacy on investment decisions: with special reference to undergraduates in the western province, sri lanka. asian journal of contemporary education, 4(2), 110–126. https://doi.org/10.18488/journal.137.2020.42.110.126 lee, k., & jeong, d. (2023). too much is too bad: the effect of media coverage on the price volatility of cryptocurrencies. journal of international money and finance, 133, n/a. liu, y., tsyvinski, a., & wu, x. (2022). common risk factors in cryptocurrency. the journal of finance, 77(2), 1133– 1177. https://doi.org/10.1111/jofi.13119 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52(1), 5–44. omane-adjepong, m., alagidede, i. p., lyimo, a. g., & tweneboah, g. (2021). herding behaviour in cryptocurrency and emerging financial markets. cogent economics & finance, 9(1), 1–14. https://doi.org/10.1080/23322039.2021. 1933681 omanovic, a., & zaimovic, a. (2024). the determinants of financial risk tolerance and portfolio allocation: have the covid-19 pandemic and the ukraine war affected our risk tolerance? borsa istanbul review, 24(6), 1381-1392. doi: 10.1016/j.bir.2024.11.001 ouachani, s., belhassine, o., & kammoun, a. (2021). measuring financial literacy: a literature review. managerial finance, 47(2), 266-281. doi: 10.1108/mf-042019-0175 paraboni, a. l., & da costa, n. (2021). improving the level of financial literacy and the influence of cognitive ability in this process. journal of behavioral and experimental economics, 90, 101656. https://doi.org/10.1016/j.socec.2020.101 656 prager, r. a. (2014). determinants of the locations of alternative financial service providers. review of industrial organization, 45(1), 21–38. https://doi.org/10.1111/ijcs.13002 https://www.fdic.gov/analysis/household-survey/2021report.pdf https://www.fdic.gov/analysis/household-survey/2021report.pdf https://www.federalreserve.gov/consumerscommunities/shed.htm https://www.federalreserve.gov/consumerscommunities/shed.htm https://doi.org/10.1007/s40812-019-00138-6 https://doi.org/10.1007/s40812-019-00138-6 https://www.proquest.com/trade-journals/do-large-swings-equity-values-change-risk/docview/1535617637/se-2?accountid=14968 https://www.proquest.com/trade-journals/do-large-swings-equity-values-change-risk/docview/1535617637/se-2?accountid=14968 https://www.proquest.com/trade-journals/do-large-swings-equity-values-change-risk/docview/1535617637/se-2?accountid=14968 https://www.proquest.com/trade-journals/do-large-swings-equity-values-change-risk/docview/1535617637/se-2?accountid=14968 https://doi.org/10.1145/3596906 https://doi.org/10.61190/fsr.v31i2/3.3530 https://doi.org/10.61190/fsr.v31i2/3.3530 https://doi.org/10.18488/journal.137.2020.42.110.126 https://doi.org/10.18488/journal.137.2020.42.110.126 https://doi.org/10.1111/jofi.13119 https://doi.org/10.1080/23322039.2021.1933681 https://doi.org/10.1080/23322039.2021.1933681 https://doi.org/10.1016/j.socec.2020.101656 https://doi.org/10.1016/j.socec.2020.101656 curnutt and smith 141 roza, t. h., tavares, h., kessler, f. h. p., & passos, i. c. (2023). problematic trading: gambling-like behavior in day trading and cryptocurrency investing. trends in psychiatry and psychotherapy (pre-publication). https://doi.org/10.47626/22376089-2023-0623 shaffer, v. a., focella, e. s., hathaway, a., scherer, l. d., & zikmund-fisher, b. j. (2018). on the usefulness of narratives: an interdisciplinary review and theoretical model. annals of behavioral medicine, 52(5), 429–442. https://doi.org/10.1093/abm/kax008 sokol-hessner, p., & rutledge, r. (2018). the psychological and neural basis of loss aversion. current directions in psychological science: a journal of the american psychological society, 28(1), 20–27. https://doi.org/10.1177/0963721418806 510 thampanya, n., wu, j., nasir, m. a., & liu, j. (2020). fundamental and behavioural determinants of stock return volatility in asean-5 countries. journal of international financial markets, institutions & money, 65, 101193. https://doi.org/10.1016/j.intfin.2020.101 193 wronka, c. (2022). money laundering through cryptocurrencies: analysis of the phenomenon and appropriate prevention measures. journal of money laundering control, 25(1), 79–94. https://doi.org/10.1108/jmlc-02-20210017 yermack, d. (2015). is bitcoin a real currency? an economic appraisal. in handbook of digital currency (pp. 31–43). https://doi.org/10.1016/b978-0-12802117-0.00002-3 zhang, y., naveed, k., & qi, j. (2025). crypto investment: the role of investment motivations, investment confidence, and risk perceptions. financial services review, 33(1), 120–141. https://doi.org/10.47626/2237-6089-2023-0623 https://doi.org/10.47626/2237-6089-2023-0623 https://doi.org/10.1093/abm/kax008 https://doi.org/10.1177/0963721418806510 https://doi.org/10.1177/0963721418806510 https://doi.org/10.1016/j.intfin.2020.101193 https://doi.org/10.1016/j.intfin.2020.101193 https://doi.org/10.1016/b978-0-12-802117-0.00002-3 https://doi.org/10.1016/b978-0-12-802117-0.00002-3 financial services review, 33(2) 142 appendix a table a1. variable used in the study variable type variable name description dependent variables s16_a in the past year, have you bought or held cryptocurrency as an investment? (yes = 1, no = 0) s16_b in the past year, have you used cryptocurrency to buy something or make a payment? (yes = 1, no = 0) s16_c in the past year, have you used cryptocurrency to send money to friends or family? (yes = 1, no = 0) explanatory variables bk1 do you (and/or your spouse or partner) have a checking, savings, or money market account? (yes = 1, no = 0) bk2_a in the past 12 months, did you purchase a money order from a place other than a bank? (yes = 1, no = 0) bk2_b in the past 12 months, did you cash a check at a place other than a bank? (yes = 1, no = 0) bk2_c in the past 12 months, did you take out a payday loan or payday advance? (yes = 1, no = 0) bk2_d in the past 12 months, did you take out a pawnshop loan or an auto title loan? (yes = 1, no = 0) bk2_e in the past 12 months, did you obtain a tax refund advance to receive your refund faster? (yes = 1, no = 0) bk2_f in the past 12 months, did you pay an overdraft fee on a bank account? (yes = 1, no = 0) fl0 risk tolerance (scale: 0 = not at all willing to take risks, 10 = very willing to take risks) financial literacy variables fl2 "buying a single company’s stock usually provides a safer return than a stock mutual fund." (1 = true, 2 = false, -2 = don’t know) fl4 "if the interest rate on savings is 1% per year and inflation is 2%, how much would you be able to buy after one year?" (1 = more than today, 2 = same, 3 = less than today) fl5 "if you had $100 in a savings account at 2% interest per year, how much would you have after five years?" (1 = more than $102, 2 = exactly $102, 3 = less than $102, -2 = don’t know) fl_score number of correct financial literacy answers (0 = 0 correct, 1 = 1 correct, 2 = 2 correct, 3 = 3 correct) control variables ppethm self-reported race/ethnicity (1 = white, non-hispanic; 2 = black, non-hispanic; 3 = other, non-hispanic; 4 = hispanic; 5 = two or more races, non-hispanic) ppgender self-reported gender (0 = male, 1 = female) curnutt and smith 143 ppage respondent’s age (continuous variable) ppemploy employment status (1 = working full-time, 0 = otherwise) educ_4cat educational attainment (1 = less than high school, 2 = high school degree/ged, 3 = some college/technical/associate’s degree, 4 = bachelor’s degree or higher) ppmarit5 marital status (1 = married, 0 = not married) pps0596 self-reported savings and investments (1 = less than $50,000, 2 = $50,000 to $99,999, 3 = $100,000 to $249,999, 4 = $250,000 to $499,999, 5 = $500,000 to $999,999, 6 = $1,000,000 or more) ppfs1482 self-reported credit score perception (1 = very poor, 2 = poor, 3 = fair, 4 = good, 5 = excellent) atleast_okay self-reported financial well-being (yes = 1, no = 0) financial services review, 33(3) 61 central bank digital currency: perspectives on design choices and implications, with a focus on e-rupee dr. vidhu shekhar1 and sanjogita ramesh2 abstract central bank digital currencies (cbdcs) are rapidly gaining momentum as 134 countries explore digital currency initiatives, yet critical gaps remain regarding design choices that determine implementation success. this study examines how cbdc design decisions influence effectiveness and integration with existing payment systems in emerging economies with established digital infrastructures. through in-depth interviews with 22 experts, we identify essential implementation considerations and their systemic implications. our findings reveal three critical insights for successful cbdc deployment. first, a two-tier, non-interest-bearing distribution model preserves banking stability while enabling innovation, with offline capabilities essential for broad accessibility. second, cbdcs should complement rather than replace existing digital payment platforms, enhancing system efficiency through immediate settlement finality and programmability. third, phased implementation guided by clear metrics and strategic partnerships proves essential for sustainable adoption. this research contributes a novel four-layer design framework demonstrating the interconnected effects of technological, security, financial, and user experience choices on payment system evolution. these findings guide emerging market policymakers in optimizing cbdc implementation through strategic interoperability, infrastructure leverage, and balanced innovation. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation shekhar, v. & ramesh, s. (2025). central bank digital currency: perspectives on design choices and implications, with a focus on e-rupee. financial services review, 33(3), 61-79. introduction the digital development is transforming the financial industry (nourallah et al., 2021), and 134 countries are exploring the implementation of central bank digital currencies (cbdcs). this represents a notable development in the evolution of the monetary and payment system worldwide. cbdcs arguably signify the most transformative shift in money since the abandonment of the gold standard and the subsequent adoption of fiat currencies. this shift fundamentally alters monetary system structures, governance, and technology by 1 s. p. jain institute of management and research (spjimr), mumbai, india. 2 corresponding author (fpm22.sanjogita@spjimr.org). spjimr, mumbai, india. integrating programmability, enhancing financial inclusion, and redefining cross-border payments. according to the bank of international settlements (bis, 2020), cbdcs can play a pivotal role in modernizing payment systems by enhancing efficiency, security, and inclusivity. of the countries currently exploring cbdc implementation, three have fully launched their systems, 44 are in the pilot phase, and the remainder are in various stages of research and development (atlantic council, 2025). https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(3) 62 this growing interest underscores the recognition of cbdcs' potential to enhance payment efficiency, while simultaneously providing central banks with innovative tools for oversight and policy implementation. nevertheless, the successful implementation of cbdcs demands careful consideration of numerous interconnected design elements, particularly within economies already possessing sophisticated digital payment infrastructures (allen et al., 2020; bis, 2020). the literature on cbdcs has predominantly focused on theoretical frameworks and technical aspects of implementations. previous studies, such as bis (2020), herve tourpe et al. (2023) and soderberg et al. (2023) have explored cbdc design processes, noting that design parameters evolve alongside advancements in technology and policy. however, a research gap persists concerning the specific impacts that cbdc design choices have on payment systems. this gap is further relevant for countries already having robust and widely adopted digital payment platforms serving large populations efficiently. india presents an ideal reference case for examining these dynamics, as it possesses one of the world's most successful digital payment ecosystems through the unified payments interface (upi), while simultaneously piloting its e-rupee cbdc, creating a unique laboratory for understanding cbdc-existing system interactions. to address these gaps, this study provides a structured examination of how cbdc design decisions influence existing financial infrastructures, particularly within emerging economies like india, featuring mature digital payment systems. our research focuses on three research questions: rq1: what critical design choices determine cbdc effectiveness? rq2: how do cbdc design choices impact existing payment platforms and their evolution? rq3: what considerations should guide cbdc implementation in emerging economies with established digital payment infrastructure, such as india? methodologically, we adopt a qualitative approach, conducting in-depth expert interviews to examine critical cbdc design considerations and their implications for existing payment systems. the expert panel comprises 22 professionals from diverse sectors, including fintech, banking, payments, and academia, offering insights into the technical and operational dimensions of cbdc implementation. our findings indicate that cbdc design involves multiple integrated layers, each with distinct implications for payment system functionality and evolution. we highlight three key insights. first, a two-tier, non-interestbearing distribution model preserves banking stability while enabling innovation, with offline capabilities essential for broad accessibility. second, cbdcs should complement rather than replace existing digital payment platforms, enhancing system efficiency through immediate settlement finality and programmability. third, phased implementation guided by clear metrics and strategic partnerships proves essential for sustainable adoption. the remainder of this paper is structured as follows: section 2 presents literature review on cbdc design and implications. section 3 outlines the research methodology, including details of the qualitative approach and expert interviews. section 4 discusses our key findings and section 5 concludes the paper. literature review the literature on cbdcs has evolved in response to shifting priorities in monetary policy, technological innovation, and payment system modernization. current research broadly focuses on four interconnected themes (see sub-sections 2.1–2.4). current status and implementation of cbdcs initial cbdc research emerged from central banks' post-2008 monetary policy challenges. early theoretical contributions by agarwal & kimball (2015) and rogoff (2015) conceptualized cbdcs primarily as instruments for implementing negative interest rates and overcoming the zero lower bound constraints. however, bindseil (2019) identifies a disconnect between these theoretical models and contemporary central bank objectives, noting that most central banks currently envision non-interest-bearing cbdcs subject to stringent quantity limits. shekhar & ramesh 63 subsequent research shifted its emphasis toward technological innovations and improving payment system efficiency. for instance, bech & garratt, (2017) developed a foundational taxonomy distinguishing retail and wholesale cbdcs, demonstrating that blockchain technology could provide digital cash with anonymity features while eliminating cryptocurrency volatility, and proving central bank money could transfer on distributed ledgers in real time, though the technology remained immature. auer & böhme (2021) advance this by establishing “minimally invasive” cbdc design requirements, finding that cryptocurrency-inspired approaches were unsuitable and instead identifying hybrid/intermediated architectures as most promising, while discovering a novel trade-off whereby central banks must choose between operating complex technical infrastructure or complex supervisory regimes. more recent literature increasingly examines interactions between cbdcs and established digital payment infrastructures. tercero-lucas (2023), employing a diamond-dybvig framework, explores financial stability implications, while bindseil & senner (2025) model impacts on monetary policy transmission and financial intermediation. they also highlight that many proposed technological benefits face practical limitations within central banks' conservative design parameters, including non-interestbearing structures, strict holding limits, and automated links to commercial bank accounts. design considerations technical architecture and policy implications the technical architecture of cbdcs critically shapes their functionality, security, and economic viability. a foundational design decision involves selecting between a centralized or distributed ledger architecture. allen et al. (2020) extensively analyze this choice, emphasizing its implications for accessibility, privacy, and systemic resilience, ultimately recommending a centralized architecture. another important debate centers on whether cbdc design should complement or disrupt existing payment systems. agur et al. (2022) suggest that while cbdcs must be distinct from conventional payment platforms, their integration should minimize disruptions to commercial banks, especially in economies with well-established digital payment systems. corbet et al. (2024) challenge this viewpoint by presenting empirical evidence suggesting that regulatory frameworks, rather than technological readiness, predominantly drive cbdc initiatives in emerging markets. the pwc india report (2021) further highlights implementation challenges in emerging markets, advocating for a two-tier issuance architecture that preserves commercial bank roles while enabling programmable payments and financial inclusion for unbanked populations. security considerations further complicate cbdc implementation. tian et al. (2023) highlight cbdcs' potential role in mitigating private-sector cyber risks but also caution that systemic cybersecurity vulnerabilities could threaten financial stability if inadequately addressed. beyond security and integration challenges, existing literature also explores theoretical frameworks for understanding the fundamental trade-offs inherent in cbdc design. mishra & prasad (2024) develop a general equilibrium model that examines the coexistence of cash and cbdc, demonstrating how design choices affect their relative holdings. their analysis shows that cbdcs can expand the monetary policy toolkit by enabling negative nominal interest rates and “helicopter drops” of money. the paper provides insights on design that can preserve elements of a cash-based economy while delivering digital currency benefits. regarding the global landscape of cbdc, claessens et al. (2024) highlight that 130 countries, representing 98% of the global gross domestic product (gdp), are investigating cbdcs, with varying degrees of implementation. china has already conducted 1.8 trillion-yuan (approximately $249.9 billion) worth of cbdc transactions in trials, while countries like nigeria and the bahamas have officially launched cbdcs with mixed results. these international experiences underscore the critical importance of meticulous cbdc design, illuminating both opportunities (financial inclusion, payment system innovation) and challenges (regulatory complexities, cybersecurity threats). collectively, these studies underscore the necessity for cbdc designs to balance privacy, financial services review, 33(3) 64 resilience, and regulatory oversight to achieve widespread adoption and effectiveness. cbdcs: banking disintermediation, stability, and payment system implications the potential impacts of cbdcs on banking systems are widely debated, focusing primarily on bank disintermediation, financial stability, and interactions with existing payment infrastructures. in relation to bank disintermediation, the degree to which cbdcs affect bank deposits depends on whether they are remunerative (interestbearing) or non-remunerative. for instance, the reserve bank of india (rbi) has adopted a noninterest-bearing model for its e-rupee to mitigate disintermediation risks. empirical research by chiu et al. (2023) suggests that interest rate calibration between 0.30% and 1.49% could allow cbdcs to coexist with traditional banking systems without causing severe disruptions. son et al. (2023) offer a contrasting perspective and find that remunerative cbdcs may significantly impact customer behavior and financial intermediary profitability, intensifying competition in the deposit market. concerning financial stability, corbet et al. (2024) caution against rapid cbdc deployment, especially in less-prepared economic environments. they warn that abrupt shifts in deposit structures and absent safeguards could significantly heighten financial instability risks. providing an alternative perspective, luu et al. (2023) offer empirical evidence from a large sample of banks across 86 countries, indicating that cbdc adoption contributes to financial stability by reducing leverage and asset risks while expanding lending. the authors contend that retail cbdcs may promote stability whereas wholesale cbdcs may hamper it. beyond stability concerns, the interaction between cbdcs and existing payment systems presents opportunities and challenges. di maggio et al. (2024) argue that cbdcs might reshape payment systems in emerging economies, particularly by potentially displacing private-sector digital payment providers due to differential taxation policies and transaction cost structures. in the indian context, banerjee & sinha (2023) emphasize that adopting cbdc should be complementary rather than competitive to existing systems like the upi. additionally, privacy and regulatory considerations are crucial factors influencing cbdc adoption. wang & gao (2024) stress the importance of balancing security with user anonymity to maintain customer trust while effectively mitigating illicit activities. ren et al. (2024) highlight that the inherent trade-off between preserving user privacy and meeting anti-money laundering (aml) compliance remains unresolved. the unresolved tensions across disintermediation, stability, payment integration, and privacy demonstrate the interconnected nature of cbdc design choices and their far-reaching implications for financial systems. financial inclusion, innovation, and future cbdc applications one of the most widely cited motivations for cbdc adoption is its potential to enhance financial inclusion, particularly in economies with substantial unbanked populations. tan (2024) models a two-tier cbdc system in which commercial banks function as distributors, incentivizing unbanked populations to engage with formal financial systems by opening digital accounts. the potential for cbdcs to enhance financial inclusion has emerged as a key consideration in their development. traditional banking systems often exclude certain populations due to geographic, economic, or regulatory barriers, creating demand for alternative financial solutions. recent research reveals that cryptocurrencies are increasingly used as alternative financial services for payments and money transfers, particularly among populations seeking alternatives to traditional banking (curnutt & smith, 2025). this trend highlights the potential role cbdcs could play in providing regulated digital payment alternatives that address similar needs while maintaining central bank oversight and financial stability. empirical research supports this claim. dunbar & treku (2024) find a statistically significant relationship between cbdc awareness and reductions in unbanked individuals in the u.s., notably among middle-income and shekhar & ramesh 65 underbanked groups. however, their findings emphasize that cbdc adoption alone is insufficient for achieving sustained financial inclusion, suggesting that broader financial literacy programs and digital infrastructure enhancements are necessary to ensure longterm success. beyond financial inclusion, cbdcs have the potential to stimulate technological advancements and innovation in financial services. ahnert et al. (2022) and chen et al. (2022) show that cbdcs can foster increased competition in digital payments, thereby encouraging the development of novel financial products and business models. despite these promising developments, current literature on cbdc design and implementation highlights critical gaps relevant to the objectives of this study. while theoretical frameworks on cbdc designs exist, there is limited practical guidance regarding methods for integrating cbdcs within advanced digital payment systems. this deficiency is particularly evident concerning choices about technical architecture, integration requirements, and the trade-offs between innovation and financial stability, issues central to rq1 on critical cbdc design choices. furthermore, existing studies such as di maggio et al. (2024) focus on the potential displacement of private-sector providers rather than system evolution, while banerjee & sinha (2023) emphasize complementary adoption without detailed analysis of system interactions. analyses focusing on banking systems and financial stability (chiu et al., 2023; luu et al., 2023; son et al., 2023) frequently neglect the complex interactions between existing payment service providers, fintech innovations, and established market structures. this limitation directly informs rq2 regarding how cbdc design choices affect existing payment platforms and their evolution. finally, how specific cbdc design decisions impact financial services in emerging economies with advanced digital financial infrastructures, such as india, remains underexplored. addressing this research gap aligns with rq3, which focuses on practical considerations and guidance for implementing cbdcs in contexts characterized by mature digital payment infrastructures. research methodology research design and sample selection this study uses qualitative research methods to understand cbdc design choices and their effects on payment systems in emerging economies. a qualitative approach was chosen to capture in-depth expert insights into complex, interrelated design considerations and implementation issues. the study employs purposive sampling to ensure comprehensive representation of expert perspectives on cbdc implementation. our final sample comprised 22 participants in india across three main professional groups. the first group comprised eight fintech and blockchain experts, including technology consultants, blockchain specialists, and digital currency experts who provided technical insights into cbdc design and architecture (ft1-8). the second group consisted of seven banking and payment industry professionals who offered practical implementation perspectives (bp1-7). the third group included seven academic and policy experts who contributed with policyrelated insights (ab1-7). detailed participant demographics, professional affiliations, and experience levels are summarized in appendix a. we determined sample size based on the principle of data saturation, conducting interviews until subsequent discussions no longer yielded new thematic insights. the chosen sample size was sufficient to achieve saturation, with later interviews confirming thematic patterns initially identified. data collection each interview, conducted between april and november 2024, lasted approximately 40 to 60 minutes and was audio-recorded with participant consent. the interviews were carried out using a semi-structured format. we used a set of prepared questions but allowed for open discussion to gain additional insights on emergent themes. all interviews were conducted in english. our data collection process followed a consistent pattern throughout all interviews. each session began with an introduction to the study and confirmation of informed consent. the main interview portion explored participants' views on cbdc design choices, financial services review, 33(3) 66 effects on existing payment systems, and implementation considerations. we encouraged participants to provide specific examples and elaborate on their experiences. this structured, yet flexible, approach helped ensure overall coverage of key topics while allowing for the exploration of unexpected but relevant themes that emerged during the discussions. the study adhered strictly to ethical research standards. all participants received comprehensive information on study objectives, procedures, and confidentiality measures prior to providing informed consent. the participants were informed of their right to withdraw at any point, though none chose to exercise this option. data analysis the analysis of interview data followed the thematic analysis methodology established by braun & clarke (2006). we began by creating detailed transcripts of all recorded interviews, with each transcript checked multiple times for accuracy. initial analysis involved careful reading of all transcripts to identify key themes and patterns in the responses. we then developed a coding framework to organize the data into meaningful categories aligned with our three research questions. the coding process involved the two researchers working independently to ensure the reliability of the interpretation. regular meetings allowed us to discuss and resolve any differences in coding decisions until we reached an agreement. we paid particular attention to emerging patterns related to cbdc design choices and their implications for payment systems. this collaborative approach helped us minimize individual researcher bias. several limitations of our research methodology must be acknowledged. firstly, our expert sample predominantly comprised individuals from major financial centers, potentially limiting insights from other regions. secondly, the timing of data collection coincided with the early stages of cbdc implementation, suggesting that certain findings may require revisiting as implementations progress. thirdly, due to the rapidly evolving nature of cbdc technologies and associated regulatory landscapes, some technical insights reported here may need periodic updating. findings our thematic analysis identified four central themes regarding cbdc implementation and design: (1) need and status of cbdc implementation, (2) critical design elements, (3) implications for financial systems, and (4) future cbdc use cases in financial services. these themes provide detailed insights into how cbdc design choices influence payment systems and financial infrastructure in emerging economies (see appendix b) need and status of cbdc implementation the emergence of cbdcs represents a strategic response to evolving financial landscapes. our expert interviews revealed diverse perspectives on the primary drivers of cbdc adoption, with notable differences between technology and banking professionals. fintech experts highlighted technological innovation and efficiency. as noted by a blockchain specialist (ft1): “cbdcs can be introduced in the economy when there is a decline in cash usage. central banks must evaluate if there is a need to modernize public payment infrastructure.” conversely, banking professionals emphasized regulatory considerations and financial stability. a banking executive (bp5) explained: “the primary motivation should be ensuring monetary sovereignty in an increasingly digital world, rather than simply following technological trends.” india's e-rupee implementation represents a carefully considered approach to digital currency deployment. per the reserve bank of india (2022) report, e-rupee pilot was launched in december 2022, adopting a two-tier distribution model that maintains the role of traditional banking intermediaries while introducing innovative digital currency features. an industry expert (bp1) highlighted the following key advantage: “the technology used in e-rupee is blockchain, which is very difficult to break through; if there is any suspicious activity, information will be broadcasted and traced immediately.” a notable aspect of india's approach is the nonremunerative design of the e-rupee, which was chosen to minimize disruption to the banking sector. as a banking expert (bp5) explained: “banks will play a crucial role in cbdc shekhar & ramesh 67 implementation, providing payment and transaction settlement services.” this design choice reflects careful consideration of financial stability while enabling innovation in payment services. the pilot phase has revealed several promising developments. in the retail segment, the erupee demonstrates capabilities for instant settlement, offline transactions, and enhanced privacy features. the wholesale segment shows potential for improving interbank settlement efficiency and reducing operational costs. a fintech research expert (bp6) noted: “cbdc settlements are instantaneous and final, reducing operational risks and settlement time significantly.” critical design elements of cbdc figure 1 illustrates the layered conceptual framework of cbdc design, as derived from expert interviews and thematic analysis. this structure comprises four interconnected layers: technology, security, financial functions, and user experience, each representing critical areas of decision-making in cbdc implementation. these design layers collectively influence the effectiveness, adoption, and integration of cbdcs within existing financial infrastructures, highlighting the multifaceted considerations central banks and stakeholders must navigate. the interconnected nature of these layers directly impacts how cbdcs integrate with existing payment systems (rq2) and shapes implementation strategies for emerging economies (rq3). expert interviews revealed differences between different stakeholder groups regarding optimal cbdc design choices. while banking experts emphasized the importance of financial layer and a two-tier distribution model to maintain financial stability, technical experts highlighted the need for robust security frameworks and offline functionality in the technology layer. figure 1. cbdc design layers source: authors’ illustrations technology layer the technological framework has strong implications for financial markets and services. the success of cbdc implementation depends on aligning technological choices with policy objectives while addressing practical requirements. interviews revealed divergent priorities: fintech experts emphasized innovation, with a consultant (ft4) noting: “the choice of architecture must balance innovation potential with practical implementation constraints.” on the other hand, banking professionals prioritized integration with existing financial systems. as shown in figure 2, cbdc is of two types: wholesale cbdc for interbank settlements and retail cbdc for customer usage. cbdc can be deployed using a one-tier (central bank direct), two-tier (with intermediaries), or a hybrid model. india's two-tier approach contrasts with china's digital yuan implementation, which employs a centralized management model within a twotier operational model, with direct central bank oversight. while china prioritizes monetary control and surveillance capabilities, india's model preserves banking sector roles. nigeria's enaira represents a third approach, utilizing a hybrid model that faced adoption challenges due to limited integration with existing mobile money systems. as one academic expert (ap3) noted: “each country's existing payment infrastructure shapes viable cbdc architectures.” technology layer security layer financial function layer user experience layer financial services review, 33(3) 68 figure 2. technology architecture of cbdc source: authors' illustrations as shown in figure 2, cbdc platforms can be digital ledger technology (dlt)-based, blockchain-based, or application programming interface (api)-based, each with distinct characteristics. dlt refers to a digital system for recording transactions where records are maintained simultaneously across multiple locations, providing greater security and transparency than traditional databases. blockchain is a type of dlt designed to be cryptographically connected in a sequential order ( bureau of engraving and printing, 2025). apis are sets of protocols and tools that allow different software applications to communicate with each other, enabling seamless integration between systems. dlt platforms offer flexibility and privacy controls, suitable for both wholesale and retail applications. blockchain platforms provide enhanced security but may face scalability issues. api-based platforms allow integration with existing banking systems and high scalability. financial experts (bp5, bp6) favored approaches that maintain compatibility with existing infrastructure, while technology specialists (ft1, ft3) emphasized the distributed ledger technologies' transformative potential. platform selection creates critical trade-offs that directly impact payment system integration (rq2). wholesale cbdcs prioritizing security and settlement finality favor dlt or blockchain architectures, but this choice may limit interoperability with existing banking apis. retail cbdcs requiring high transaction throughput often employ api-based approaches for seamless integration with current payment rails yet sacrifice some of the programmability benefits that dlt platforms offer. cross-border applications demand strong interoperability, typically requiring dlt or hybrid solutions that can bridge different national payment systems while maintaining regulatory compliance. a centralized architecture prioritizes efficiency and oversight but introduces potential vulnerabilities. decentralized architectures enhance resilience and market participation, while hybrid approaches balance innovation with regulatory control. as one banking professional (bp1) noted: “settlement efficiency gains in wholesale applications could transform liquidity management for financial institutions.” academic experts (ap3, ap7) emphasized balancing innovation with stability, suggesting hybrid approaches might offer the best compromise. these technological choices collectively determine both operational characteristics and the cbdc's potential to transform financial market structures. technology architechture of cbdc type wholesale cbdc, retail cbdc platform/ database architechture dlt, blockchain, api model of deployment one-tier, two-tier, hybrid shekhar & ramesh 69 security layer security architecture plays a central role in shaping cbdc functionality and integration capabilities. the choice between privacypreserving and transparency-focused security models creates a design tension that affects both user adoption and regulatory compliance. expert interviews revealed varying security priorities across stakeholder groups. an industry expert (bp1) emphasized: “rbi does not know who the end recipient is; it only tracks how much currency is released and utilized; transactions are private end to end, similar to cash transactions.” a technology expert (ft4) noted: “cbdc is a secure method of payment, with no risk of personal information leakage.” cbdcs must be developed with privacypreserving techniques such as homomorphic encryption, incorporate smart contracts for regulatory oversight, and include automated encryption for monitoring. banking professionals (bp3, bp5) emphasized familiar security frameworks, while technology specialists (ft1, ft7) advocated for advanced cryptographic solutions. risk management strategies, including recoverability, technical stability, and a robust governance framework, must be implemented to ensure safe transactions. the e-rupee is designed to include features resembling physical currency, with central bank supervision and options for anonymity in smaller transactions. however, security choices create implementation challenges. enhanced privacy features may conflict with aml or know your customer (kyc) requirements, while excessive transparency could undermine user adoption. as one technology expert (ft1) noted: “the challenge is creating security that satisfies both user privacy expectations and regulatory oversight needs without compromising system performance.” financial function layer the financial function layer equipped with functionalities to conduct financial operations. this layer encompasses a range of features that enable efficient processing of transactions, management of digital assets, and integration with existing financial systems, enhancing the overall effectiveness of the cbdc framework. our expert interviews revealed contrasting perspectives on key financial function aspects. academic experts highlighted the theoretical benefits of interest-bearing cbdcs, while industry professionals emphasized practical stability concerns. as one senior banker (bp2) explained, “an interest-bearing cbdc could fundamentally alter deposit dynamics, potentially disrupting commercial bank funding models.” remunerative cbdcs offer an interest-bearing characteristic and could act as a liquid government debt instrument, serving as a secure asset. however, there is a risk that these types may disrupt traditional banking systems if they become the favored option for deposits over savings accounts due to higher competitive interest rates, potentially leading to bank runs. conversely, non-remunerative cbdcs exist solely as digital currencies that do not accrue interest. this type prevents bank disintermediation and helps maintain stability in the financial system. the pilot phase of the erupee is intended to be non-remunerative and does not accrue any interest on value storage. the tokenization capability enables the digital representation of financial assets or rights within the cbdc framework, fostering efficient payment and settlement systems and encouraging innovation in financial services. technology experts (ft1, ft7) emphasized the transformative potential of tokenization for asset markets. the programmability feature restricts the use of cbdc tokens to specific applications, such as designated cbdc medicine tokens that can only be used for purchasing medicines. as noted by a fintech expert (ft6): “programmable money creates entirely new possibilities for targeted policy implementation.” this characteristic can impact welfare programs and presents opportunities for innovation in financial products and services. the selection between financial functions creates cascading effects on existing payment systems. non-remunerative designs preserve banking sector stability but may limit cbdc adoption incentives, programmability enables innovative financial products but requires new regulatory frameworks that existing payment providers must navigate. a banking expert (bp2) observed: “programmable features could financial services review, 33(3) 70 either complement existing fintech or displace them entirely.” the interoperability feature facilitates seamless interactions among financial systems, promoting quicker and more effective settlements and payments. banking professionals (bp4, bp6) emphasized the importance of interoperability with existing infrastructure to ensure smooth adoption and system efficiency. user experience layer the user experience influences how individuals access and utilize cbdcs. expert interviews revealed diverse perspectives on the balance between innovation and accessibility. while technology experts (ft2, ft4) advocated for feature-rich interfaces, financial inclusion specialists (bp7, ap7) emphasized simplicity and accessibility across diverse user segments. the e-rupee must be built to scale effectively across all regions of india, ensuring accessibility on various devices while serving diverse population segments, including those with limited connectivity or banking services. a financial inclusion expert (bp7) noted: “cbdc should be designed to accommodate diverse segments of the population, including various age groups and economic backgrounds.” cbdcs can follow either an account-based or token-based structure. in an account-based cbdc, access and claims are tied to the user's bank account and are subject to kyc regulations to identify and verify the user's identity. this helps prevent fraudulent accounts, theft, and unauthorized access. if the cbdc is implemented using the one-tier model, compliance and authentication processes must be adhered to by the central bank (auer & böhme, 2021). conversely, in a token-based framework, cbdcs are issued as digital tokens and distributed by collaborating financial institutions through mobile app-based wallets. these partner banks provide applications to their registered customers, who can fund their cbdc wallets using their bank accounts, enabling peer-to-peer or merchant transactions. the e-rupee (retail) operates on a token-based model; the rbi has teamed up with various banks nationwide to distribute cbdc tokens to the public. banking customers can register with a partner bank to access the e-rupee wallets, loading funds into it via their bank accounts or upi, facilitating direct transactions to peers' wallets or payments to merchants using qr codes without intermediaries. this presents a rapid and secure method of transferring digital currencies. a key innovation highlighted by technology specialists (ft6, ft8) is the offline functionality of india's cbdc. the design allows it to function effectively in low or limited network conditions, making it accessible on essential devices like feature phones and catering to users with varying levels of financial or digital literacy. as one expert (bp1) explained: “the offline capability is crucial for adoption in rural areas where connectivity remains challenging.” indian cbdc’s offline design enables transactions on feature phones without needing a banking app or strong network connectivity in a secure environment (rbi, 2022). this innovative strategy positions the e-rupee as a promising solution for delivering digital financial services remotely while improving user experience for various population segments. additionally, the availability of erupee wallets in multiple local languages further boosts accessibility across different regions and demographics. implementation sequencing and design dependencies expert interviews further revealed that cbdc design choices create sequential dependencies that constrain future options. several experts emphasized that infrastructure decisions made early in implementation become difficult to reverse later. phase 1 foundation decisions: platform architecture (blockchain vs. api) and distribution model (one-tier vs. two-tier) must be established first, as these choices determine interoperability possibilities and regulatory frameworks. phase 2 integration features: security protocols and financial functions (programmability, tokenization) build upon architectural foundations but can be refined during pilot phases. phase 3 user experience: interface design and offline capabilities can be iteratively improved shekhar & ramesh 71 but depend on the underlying technical architecture established in phase 1. implications of cbdc design on financial systems cbdc design choices can have large implications for existing financial systems, directly addressing rq2 regarding impact on payment platforms. our analysis reveals three critical areas where design decisions reshape financial infrastructure: payment system architecture, banking intermediation models, and innovation pathways. while banking professionals emphasized the importance of maintaining financial stability through careful design choices, fintech experts highlighted the transformative potential of programmable cbdcs for service innovation. considering the evolution of digital payments in india, it is to be noted that india's digital payment landscape has been transformed by the upi. interestingly, our expert interviews revealed fundamental differences between upi and cbdc architectures that have major implications for the payment system. a blockchain expert (bp4) noted: “upi transactions are processed in phases and may be declined if the server does not confirm the transaction. in contrast, cbdc payments are settled immediately, with zero transaction fees and failures.” this difference creates both opportunities and challenges for system integration. while fintech experts emphasized the complementary nature of these systems, banking professionals expressed concerns about the potential fragmentation of payment infrastructures. academic experts took a middle position, suggesting that the two systems could coexist with different use cases based on their relative strengths: upi for high-volume, low-value retail transactions and cbdc for settlementcritical or offline use cases. the comparison further reveals that while upi relies on existing banking infrastructure and internet connectivity, cbdc utilizes blockchain technology and enables offline capabilities, representing an essential architectural divergence with implications for payment system evolution. the architectural differences between upi and cbdc systems illustrate how design choices cascade through financial infrastructure. while upi's success demonstrates the potential for digital payment adoption, cbdc's blockchain foundation creates fundamentally different settlement mechanisms that could either complement or compete with existing systems, depending on implementation choices. the emergence of cbdcs is reshaping bank roles within the financial system. banks are now adapting to provide cbdc wallets, customer onboarding, transaction monitoring, and support services. this evolution introduces opportunities, but also challenges for traditional financial institutions. our interviews revealed contrasting perspectives between banking and fintech experts regarding the impact on banking business models. banking professionals emphasized the potential for disintermediation if cbdc design fails to incorporate appropriate safeguards, while fintech experts highlighted new service opportunities enabled by programmable digital currencies. the emergence of cbdc is also driving innovation in banking products. examples include programmable payment solutions, cbdc-driven lending services, and integrated treasury solutions. these innovations enhance service delivery and reduce operational expenses while preserving the traditional banking framework. however, as an academic expert (ap6) cautioned: “the pace of innovation must be balanced against systemic stability considerations.” highlighting the ongoing tension between transformation and stability in cbdc implementation. future cbdc use cases in financial services our analysis of future cbdc applications revealed contrasting perspectives between technology visionaries and practical implementers. while blockchain specialists emphasized transformative potential for crossborder payments and programmable finance, banking practitioners focused on incremental improvements to existing services and practical adoption challenges. cross-border cbdc implementation faces significant coordination challenges that temper optimistic predictions. while a fintech expert (ft5) predicted that: “swift will become financial services review, 33(3) 72 ineffective due to global trade divergence towards cbdc.”, banking professionals expressed more measured views, with one senior banker (bp5) noting: “international payment system transformation will be evolutionary rather than revolutionary, with cbdcs gradually integrating with existing frameworks.” the practical obstacles include regulatory harmonization, liquidity management, and the need for gradual transition mechanisms that preserve existing correspondent banking relationships. the potential for cooperative cross-border cbdc initiatives emerged as an opportunity. a blockchain expert (ft3) noted: “there is potential for a cbdc bridge in several countries, including singapore, hong kong, and the uae.” this initiative can facilitate faster and more affordable cross-border payments, especially concerning dedollarization trends. the application of cbdcs for financial inclusion and social benefit distribution emerged as a key theme, but with different emphases across expert groups. technology consultants highlighted sophisticated programmable features, while financial inclusion specialists emphasized practical implementation considerations. the tokenization of social benefits via cbdcs introduces advantages that increase the efficiency and effectiveness of welfare distribution. an academic expert (ap4) emphasized the potential for targeted distribution: “purpose-specific tokens could ensure that benefits reach intended beneficiaries with minimal leakage.” however, financial inclusion practitioners highlighted implementation challenges. a financial inclusion expert (bp7) noted: “lastmile distribution requires not just technology but trusted human intermediaries, particularly in rural areas.” this perspective emphasized that technological innovation alone cannot address financial inclusion challenges without appropriate distribution strategies and supporting infrastructure. synthesis of expert perspectives our analysis revealed important tensions between different stakeholder groups regarding cbdc design and implementation. technology experts favored innovative, feature-rich approaches that maximize the transformative potential of digital currencies. banking professionals emphasized stability, integration with existing systems, and minimizing disruption to established business models. academic and regulatory experts focused on the broader socioeconomic implications, particularly regarding financial inclusion and monetary policy effectiveness. despite these differences, several areas of consensus emerged across expert groups. first, all participants agreed that cbdc should complement rather than replace existing payment systems, particularly in economies with established digital payment infrastructure. second, a two-tier distribution model was widely endorsed as appropriate for preserving financial stability while enabling innovation. third, experts across all groups emphasized the importance of offline functionality for addressing financial inclusion objectives, particularly in regions with limited connectivity. the stakeholder tensions reflect deeper questions about the pace and scope of financial system transformation. the consensus around complementary rather than disruptive implementation suggests that successful cbdc design requires careful calibration between innovation and stability, a finding particularly relevant for emerging economies with established digital payment infrastructures (rq3). conclusion as 134 countries explore cbdc implementation, this study offers actionable insights into the design factors that shape implementation outcomes, particularly in emerging economies with advanced digital financial infrastructures. using india’s e-rupee as a reference case, we examined how specific design decisions affect integration with existing financial systems. drawing on interviews with 22 experts across fintech, banking, and academia, we identified several interdependent design considerations critical to cbdc effectiveness. three core insights emerged from the analysis. first, experts broadly supported a two-tier, noninterest-bearing model that preserves banking sector stability while enabling innovation. emphasis was placed on settlement finality and shekhar & ramesh 73 offline capabilities as prerequisites for broadbased adoption and accessibility. second, cbdcs should be designed to complement existing digital payment platforms, such as india’s upi. when properly integrated, cbdcs can enhance transaction efficiency, security, and programmability, without disrupting well-functioning payment ecosystems. third, a phased implementation strategy, guided by clear performance metrics and public-private collaboration, is essential. early architectural and governance choices have long-term consequences. this reinforces the need for strategic foresight. these findings contribute to a relatively underdeveloped dimension of cbdc research by introducing a generalizable four-layer design framework. this framework comprises four interrelated dimensions: technology, security, financial functionality, and user experience. it provides a structured approach for understanding how design decisions cascade through financial systems, shaping accessibility, stability, and innovation. beyond its analytical utility, the framework serves as a practical tool for policymakers navigating complex implementation environments. we recommend that regulators in emerging economies prioritize interoperability with existing payment platforms, leverage established financial infrastructure, and implement governance models that balance innovation with systemic stability. we acknowledge that our expert sample predominantly included professionals from major financial centers, potentially limiting insights from regional or rural contexts where implementation challenges may differ significantly. broader geographic representation could provide a deeper understanding of local-level complexities. additionally, our data was collected at an early stage of global cbdc implementation, suggesting that certain findings might evolve as practical experiences accumulate. future studies at more advanced implementation stages would validate or refine these insights. future research should quantitatively evaluate cbdc outcomes, undertake comparative analyses across diverse implementation environments, and conduct longitudinal assessments of cbdc impacts on financial inclusion and banking dynamics. as global cbdc adoption progresses, empirical insights will become important to inform responsible monetary innovation. references agarwal, r., & kimball, m. (2015). breaking through the zero lower bound (imf working paper no. 15/224). international monetary fund. https://www.aeaweb.org/conference/2 016/retrieve.php?pdfid=13281&tk=69 t3ahqy agur, i., ari, a., & dell’ariccia, g. (2022). designing central bank digital currencies. journal of monetary economics, 125, 62–79. ahnert, t., assenmacher, k., hoffmann, p., leonello, a., monnet, c., & porcellacchia, d. (2022). the economics of central bank digital currency. european central bank. https://papers.ssrn.com/sol3/papers.cf m?abstract_id=4192178 allen, s., čapkun, s., eyal, i., fanti, g., ford, b. a., grimmelmann, j., juels, a., kostiainen, k., meiklejohn, s., miller, a., prasad, e., wüst, k., & zhang, f. (2020). design choices for central bank digital currency: policy and technical considerations (working paper no. 27634). national bureau of economic research. https://doi.org/10.3386/w27634 atlantic council. (2025). central bank digital currency tracker. central bank digital currency tracker. https://www.atlanticcouncil.org/cbdctr acker/ auer, r., & böhme, r. (2021). central bank digital currency: the quest for minimally invasive technology. bis papers, 948. banerjee, s., & sinha, m. (2023). promoting financial inclusion through central bank digital currency: an evaluation of payment system viability in india. australasian accounting, business and finance journal, 17(1), 176–204. https://www.aeaweb.org/conference/2016/retrieve.php?pdfid=13281&tk=69t3ahqy https://www.aeaweb.org/conference/2016/retrieve.php?pdfid=13281&tk=69t3ahqy https://www.aeaweb.org/conference/2016/retrieve.php?pdfid=13281&tk=69t3ahqy https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4192178 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4192178 financial services review, 33(3) 74 bech, m. l., & garratt, r. (2017). central bank cryptocurrencies (ssrn scholarly paper no. 3041906). social science research network. https://papers.ssrn.com/abstract=3041 906 bindseil, u. (2019). central bank digital currency: financial system implications and control. international journal of political economy, 48(4), 303–335. https://doi.org/10.1080/08911916.201 9.1693160 bindseil, u., & senner, r. (2025). modeling central bank digital currencies. journal of economic surveys. https://doi.org/10.1111/joes.12686 bis. (2020). central bank digital currencies: foundational principles and core features. braun, v., & clarke, v. (2006). using thematic analysis in psychology. qualitative research in psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp 063oa bureau of engraving and printing. (2025, march 28). how a bitcoin is created. https://www.moneyfactory.com/howa-bitcoin-is-created/ chen, s., goel, t., qiu, h., & shim, i. (2022). cbdcs in emerging market economies. bis papers. https://papers.ssrn.com/sol3/papers.cf m?abstract_id=4085690 chiu, j., davoodalhosseini, s. m., jiang, j., & zhu, y. (2023). bank market power and central bank digital currency: theory and quantitative assessment. journal of political economy, 131(5), 1213–1248. https://doi.org/10.1086/722517 claessens, s., cong, l. w., moshirian, f., & park, c.-y. (2024). opportunities and challenges associated with the development of fintech and central bank digital currency. journal of financial stability, 73, 101280. https://doi.org/10.1016/j.jfs.2024.1012 80 corbet, s., cumming, d., glatzer, z., & johan, s. (2024). understanding the rapid development of cbdc in emerging economies. finance research letters, 70, 106226. https://doi.org/10.1016/j.frl.2024.1062 26 curnutt, g., & smith, d. (2025). the association of cryptocurrency and the use of alternative financial services. financial services review, 33(2), article 2. https://doi.org/10.61190/fsr.v33i2.409 4 di maggio, m., ghosh, p., ghosh, s., & wu, a. (2024). impact of retail cbdc on digital payments and bank deposits: evidence from india (ssrn working paper no. 4779520). https://papers.ssrn.com/sol3/papers.cf m?abstract_id=4779520 dunbar, k., & treku, d. n. (2024). examining the impact of a central bank digital currency on the access to banking. international review of financial analysis, 93, 103220. https://doi.org/10.1016/j.irfa.2024.103 220 luu, h. n., nguyen, c. p., & nasir, m. a. (2023). implications of central bank digital currency for financial stability: evidence from the global banking sector. journal of international financial markets, institutions and money, 89, 101864. https://doi.org/10.1016/j.intfin.2023.10 1864 mishra, b., & prasad, e. (2024). a simple model of a central bank digital currency. journal of financial stability, 73, 101282. https://doi.org/10.1016/j.jfs.2024.1012 82 nourallah, m., strandberg, c., & öhman, p. (2021). mobile bank applications: loyalty of young bank customers. financial services review, 29(2), 147–167. https://doi.org/10.61190/fsr.v29i2.345 1 https://www.moneyfactory.com/how-a-bitcoin-is-created/ https://www.moneyfactory.com/how-a-bitcoin-is-created/ https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4779520 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4779520 shekhar & ramesh 75 pwc india. (2021). central bank digital currency in the indian context. https://www.pwc.in/assets/pdfs/consult ing/financial-services/cbdc-in-theindian-context.pdf ren, y.-s., ma, c., & wang, y. (2024). a new financial regulatory framework for digital finance: inspired by cbdc. global finance journal, 62, 101025. https://doi.org/10.1016/j.gfj.2024.1010 25 reserve bank of india. (2022). concept note on central bank digital currency. https://rbidocs.rbi.org.in/rdocs/publica tionreport/pdfs/conceptnoteac b531172e0b4dfc9a6e506c2c24f fb6.pdf rogoff, k. (2015). costs and benefits to phasing out paper currency. nber macroeconomics annual, 29, 445– 456. https://doi.org/10.1086/680657 soderberg, g., bechara, m. m., bossu, w., che, m. n. x., davidovic, s., kiff, m. j., lukonga, m. i., griffoli, m. t. m., sun, t., & yoshinaga, a. (2022). behind the scenes of central bank digital currency: emerging trends, insights, and policy lessons. international monetary fund. https://www.imf.org/en/publications/ wp/issues/2022/11/01/behind-thescenes-of-central-bank-digitalcurrency-525520 son, j., ryu, d., & webb, r. i. (2023). central bank digital currency: payment choices and commercial bank profitability. international review of financial analysis, 90, 102874. tan, b. j. (2024). central bank digital currency and financial inclusion. journal of macroeconomics, 81, 103620. https://doi.org/10.1016/j.jmacro.2024. 103620 tercero-lucas, d. (2023). central bank digital currencies and financial stability in a modern monetary system. journal of financial stability, 69, 101188. https://doi.org/10.1016/j.jfs.2023.1011 88 tian, s., zhao, b., & olivares, r. o. (2023). cybersecurity risks and central banks’ sentiment on central bank digital currency: evidence from global cyberattacks. finance research letters, 53, 103609. https://doi.org/10.1016/j.frl.2022.1036 09 tourpe, h., lannquist, a., & soderberg, g. (2023, september). a guide to central bank digital currency product development. international monetary fund. https://www.imf.org/en/publications/ wp/issues/2023/09/01/guide-tocbdc-product-development wang, h., & gao, s. (2024). the future of the international financial system: the emerging cbdc network and its impact on regulation. regulation & governance, 18(1), 288–306. https://doi.org/10.1111/rego.12520 https://rbidocs.rbi.org.in/rdocs/publicationreport/pdfs/conceptnoteacb531172e0b4dfc9a6e506c2c24ffb6.pdf https://rbidocs.rbi.org.in/rdocs/publicationreport/pdfs/conceptnoteacb531172e0b4dfc9a6e506c2c24ffb6.pdf https://rbidocs.rbi.org.in/rdocs/publicationreport/pdfs/conceptnoteacb531172e0b4dfc9a6e506c2c24ffb6.pdf https://rbidocs.rbi.org.in/rdocs/publicationreport/pdfs/conceptnoteacb531172e0b4dfc9a6e506c2c24ffb6.pdf financial services review, 33(3) 76 appendix a. categorization of expert interview participants group 1. fintech and blockchain experts (n=8) expert id role/position expertise area experience geographic region ft1 swift operations digital currencies, blockchain senior level uk ft2 global lead analystdigital transformations blockchain and fintech mid-level uae and india ft3 digital payment consultant blockchain and fintech senior level usa ft4 senior consultant digital solutions blockchain and fintech mid-level south east asia ft5 senior consultant digital payments and blockchain blockchain and fintech mid-level singapore ft6 senior consultant digital payment project fintech and cbdc mid-level singapore ft7 technology and strategydirector blockchain expert senior-level india ft8 emerging tech evangelist, startup enabler fintech and cbdc senior level india shekhar & ramesh 77 group 2. banking and payment industry professionals (n=8) expert id role/position expertise area experience geographic region bp1 cbdc project headpayments industry banking and fintech senior level india bp2 ceo-fintech payments-based startup banking and fintech senior level india bp3 banker-assistant manager, private bank banking and fintech mid-level india bp4 sme consultant and directionfintech startup banking and fintech senior-level india bp5 fintech startup founder, former vpprivate bank banking and fintech senior level india bp6 consultant and researcherpaytm (payments startup) banking and fintech (blockchain and payments) senior level india bp7 director-fintech organization, financial inclusion of smes and consultant financial inclusion and fintech senior level india financial services review, 33(3) 78 group 3. academic and policy experts (n=7) expert id role/position expertise area experience geographic region ap1 associate professor information technology seniorlevel india ap2 adjunct faculty and consultant finance and fintech mid-level india ap3 associate professor operations and technology senior-level india ap4 economist/resear cher policy analysis senior-level india ap5 lawyer-private firm intellectual property, cyber laws senior-level india ap6 economist and researcher global think tank policy and research senior-level india ap7 adjunct faculty and consultantdigital finance finance and technology seniorlevel india note: experience levels are categorized as: senior level: >15 years of experience, mid-level: 5-15 years of experience shekhar & ramesh 79 appendix b. thematic analysis table initial codes from interviews sub-themes main-themes decline in physical cash usage, growth of digital payments, upi success and adoption, alternative source of digital payments, need for payment infrastructure update changes in payment behaviour, digital payment evolution, infrastructure update need and status of cbdc implementation technology platform choices, security requirements, privacy concerns, programmability, interest bearing vs noninterest bearing, user interface design, blockchain advantages, distribution models technology architecture security framework financial functions user experience critical design elements upi comparison, existing payment systems limitations, role of banks, settlement efficiency, operational costs payment system impacts, banking sector transformation, infrastructure development implications for financial systems cross-border capabilities, cbdc bridge potential, programmability benefits, financial inclusion opportunities, last-mile access international settlements, specialised applications, inclusion initiatives future cbdc use cases in financial services pii: s1057-0810(99)00041-4 financial risk tolerance revisited: the development of a risk assessment instrument< john grable,a,*, ruth h. lytton,b akansas state university, 318 justin hall, manhattan, ks 66506, usa bvirginia tech, 101 wallace hall, blacksburg, va 24061, usa received 8 january 1999; received in revised form 30 august 1999; accepted 20 october 1999 abstract this paper explores conceptual, methodological, and empirical issues related to the development of a financial risk-tolerance assessment instrument. financial risk tolerance is a significant factor in a number of household financial decisions, yet few recognized, valid, and reliable methods of assessment are available for use by financial service providers and educators. empirical results from a multistage development of a 13-item risk assessment instrument are discussed. the multidimensional instrument is presented as the foundation for the development of a more widely used and accepted index. future use by practitioners and researchers is encouraged to further validate the usefulness of the instrument. © 2000 elsevier science inc. all rights reserved. jel classification:d81 keywords:risk tolerance; risk assessment; heuristics 1. introduction whether measured for the purpose of self-assessment or for documentation of investment suitability, financial risk tolerance is assumed to be a fundamental issue underlying a number of financial decisions. for this reason, researchers have long been interested in understanding < support for this research was provided by the certified financial planner board of standards. * corresponding author. tel.:11-785-532-1486; fax:11-785-532-5505. e-mail address:grable@humec.ksu.edu (j. grable) financial services review 8 (1999) 163–181 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00041-4 the relationship between personal financial risk tolerance and factors as diverse as the life cycle and asset allocation choice decisions. unfortunately, according to droms (1988), maccrimmon and wehrung (1986), roszkowski (1995), and roszkowski, snelbecker, and leimberg (1993), there are few, if any, generally recognized measures or instruments designed to ascertain someone’s financial risk tolerance or preference. according to roszkowski et al. “most existing devices appear to have been created by various financial planning concerns for their local ‘in-house’ use or are adaptations of techniques that were meant for use in scientific studies . . . no one measure has yet emerged as the standard by which the others can be evaluated” (roszkowski et al., 1993, p. 230). the need for a widely accepted and commonly used instrument is as great today as any time in the past. without such an instrument financial service providers and researchers have been forced to use other assessment techniques that may not adequately measure the underlying construct of financial risk tolerance. furthermore, according to risk-tolerance researchers (e.g., droms, 1988) and financial planning practitioners (e.g., opiela, 1996), the lack of a widely accepted risk-assessment instrument has been an ongoing problem slowing the pace of research in the area of financial management within the larger context of personal financial planning and investment management. the purpose of this paper is to present a framework for the development of a financial risk tolerance assessment instrument, and, based upon this framework, propose a financial risk-tolerance assessment instrument with corresponding reliability and validity estimates. the concepts presented in this paper are offered with the hope of moving the financial service profession closer to the ultimate development and adoption of a standardized financial risk-tolerance assessment instrument. 2. financial risk-tolerance assessment: a review the study of risk has been of interest to investors and academics for hundreds of years (bernstein, 1996); however, most research attempts to understand financial risk tolerance are relatively recent. over the 75 years of study in the united states, the assessment of financial risk tolerance has tended to revolve around five methodologies: choice dilemmas, utility theory, objective measures, heuristic judgments, and subjective assessment. the following discussion briefly describes these methods. choice dilemmas were a popular method of risk assessment until the mid-1970s. basically, choice dilemmas are scenarios where respondents are asked to make a risk choice for themselves or someone else regarding an everyday life event. after years of use these tests were found to generate little evidence of general risk-taking propensity across situations because the items were one-dimensional. maccrimmon and wehrung (1986) summarized findings related to choice dilemmas by concluding that items that ask someone “how risk tolerant are you?” measure only a small part of the multidimensional nature of risk and that most people misstate their risk tolerance in these situations. utility theory continues to be a popular method of assessing financial risk tolerance; however, recent research challenges the standard utility function assumption by showing that 164 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 most people do not have a constant risk aversion throughout the entire domain of wealth (shefrin and statman, 1993). it has been suggested that utility theory cannot adequately represent risk-taking preferences and tolerances because “the magnitudes of potential loss and gain amounts, their chances of occurrence, and the exposure to potential loss contribute to the degree of threat (versus opportunity) in a risky situation” (kahneman and tversky, 1979, p. 266). in other words, people tend to be consistently more willing to take risks when certain losses are anticipated, and are more willing to settle for a sure gain when absolute gains are anticipated (statman, 1995). the difficulty of measuring and assessing someone’s risk tolerance has prompted some researchers to recommend that “financial planners should focus on measurements of objective risk tolerance” (sung and hanna, 1996, p. 228). objective measure analysis appears to offer great potential in the assessment of financial risk tolerance (schooley and worden, 1996); however, objective risk-tolerance measures that require researchers to deduce someone’s risk tolerance via their asset holdings may also pose serious validity problems. objective measures assume that investors act in a rational way and that a person’s asset allocation is a result of personal choice rather than the advice of a third party. as a result, objective measures 1) tend to be descriptive rather than predictive, 2) do not account for the multidimensional nature of risk, and (c) often fail to explain actual investor behavior (elvekrog, 1996; train, 1995). financial services professionals commonly use heuristic judgments to assess and predict financial risk tolerance (roszkowski et al., 1993). this method assumes strong correlation’s between demographic and socioeconomic characteristics and financial risk tolerance (grable and lytton, 1998). for example, it is commonly assumed that older investors are inherently less risk tolerant than younger investors are. based on this heuristic, older individuals are typically advised to invest less of their assets in equities and more in fixed income securities. although often assumed to be based on empirically tested assumptions, heuristic judgments often fail to adequately explain or predict actual investor behavior. in many cases heuristic judgments are little more than commonly accepted myths (cutler, 1995). as haliassos and bertaut (1995) and yoo (1994) concluded, “the current body of theoretical literature does not adequately describe the behavior of individuals” (yoo, 1994, p.1), leaving many to conclude that past research gives limited insight into the relationship between demographic and socioeconomic characteristics and risk tolerance. research findings related to choice dilemmas, utility analysis, objective functions, and heuristic judgments have led some researchers and practitioners studying risk-tolerance theory to conclude that these methods are not entirely appropriate when attempting to assess a person’s financial risk tolerance (e.g., grable and lytton, 1998; maccrimmon and wehrung, 1986; statman, 1995). instead, it has been argued that the best way to concisely and accurately identify a person’s financial risk tolerance is to use an assessment instrument designed specifically to measure subjective risk tolerance using multidimensional financial scenarios and situations (maccrimmon and wehrung, 1986). however, as noted above, there are few, if any, widely accepted and commonly used measures or instruments designed to ascertain someone’s financial risk tolerance (roszkowski, 1995). maccrimmon and wehrung (1986) recommended the use of a questionnaire type instrument over other types of measures or experiments because a questionnaire does not subject 165j. grable, r.h. lytton / financial services review 8 (1999) 163–181 a respondent’s tolerances to “subtle influences of the decision analyst during the assessment process” (maccrimmon and wehrung, 1986, p. 65). questionnaires also were recommended because they allow large numbers of subjects to participate in assessments, thus eliminating response biases that can arise when multiple analysts are used to assess tolerances on an interactive basis. additionally, instead of relying on a single item, maccrimmon and wehrung recommended that surveys and experiments include situation items where respondents are asked to make financial decisions concerning lotteries, stocks, bonds, mutual funds, real estate, options, commodities, and other types of investments. 2.1. a review of instrument development issues roszkowski (1998) noted that assessing someone’s level of risk tolerance is a difficult process because risk tolerance is an elusive, ambiguous concept. some researchers have suggested that risk taking is constant across situations, but evidence indicates that, for example, a person’s level of risk tolerance for physical activities is not a good gauge of risk taking in financial situations (roszkowski, 1998; rowland, 1996). because many people are unsophisticated about investments, it is essential that assessment instruments consider different classes of assets and situations. without this consideration and the addition of multidimensional questions, research indicates that people will tend to overestimate their actual level of risk tolerance because of a desire to appear socially acceptable. a greater range of financial choices also permits researchers to make more specific distinctions among individuals. when making risky financial choices, the literature suggests that people consider four distinct elements: 1) the probability of gains, 2) the probability of loss, 3) the dollar amount of potential gains, and 4) the potential dollar loss (maccrimmon and wehrung, 1986). to assess risk tolerance accurately, roszkowski et al. (1993) suggest that risk-tolerance assessments include items querying respondents’ tolerances for guaranteed versus probable gambles, minimum probability of success items that require a risky course of action, and items offering minimum returns that require respondents to undertake a risky course of action. others have suggested including financial assessment items that elicit a choice between a sure loss of a definite amount and the probable loss of a larger amount. most people become risk seeking in “the sense that they are more willing to risk a large loss than to accept a small, but certain, loss” (roszkowski, 1995, p. 44). it also has been recommended throughout the literature that multidimensional situations remain within the context of personal finance rather than including situations outside the realm of personal finance (rowland, 1996). roszkowski and snelbecker (1989) found “that to gauge risk-taking propensity, it is necessary to ask many different items and to integrate the answers. diversifying the items used to assess risk tolerance is a sound procedure to follow” (roszkowski and snelbecker, 1989, p. 118); however, it is important to keep in mind that a questionnaire need not be too long. roszkowski and bean (1990), based on the results of a comprehensive review of response biases found in the literature, concluded that questionnaire length is inversely related to response rate, and that shorter questionnaires are almost always better than longer ones. 166 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 2.2. a review of validity and reliability issues the concurrent issues of validity and reliability play an important role in the development of financial risk-tolerance assessment instruments (maccrimmon and wehrung, 1986; roszkowski et al., 1993; roszkowski, 1995). the following discussion provides a brief outline of validity and reliability concepts as they relate to the development of a financial risk-tolerance assessment instrument. validity issues play a critical role in the creation and use of instruments designed specifically to predict and measure behavioral attitudes (babbie, 1983). face validity must be assured by combining, modifying, and integrating successfully used financial risk-tolerance items. these types of items generally emerge from a review of previous research, but they may also be developed from empirical observation. it also is important that an instrument obtain convergent validity by comparing different measures of the same trait in tests to assure that they are correlated significantly and substantially with one another. if an index is created from answers obtained from an instrument, it is crucial to test for internal validation. internal validation assures researchers that a relationship between individual items and the measure itself exists. also of importance is the consistency offered by an instrument. according to pedhazur (1982), in non-experimental research “the reliability of the measure of the independent variable tends to be low to moderate (i.e., ranging from about 0.5 to about 0.8). this is particularly so with some of the attributes used in such research (e.g., cognitive styles, self-concept, ego strength, and attitudes). therefore, the bias in estimating the regression coefficient in non-experimental research may be considerable” (pedhazur, 1982, p. 34). this indicates that for a “test to predict criterion, predictive validity is more important than reliability” (isaac and michael, 1995, p. 131). although reliability plays a secondary role, compared to validity, when developing a financial risk-tolerance assessment instrument, it is still important to judge an instrument by how consistently findings emerge from one measurement to another. future instruments designed to measure risk-tolerance attitudes should show alphas in the range of 0.5 to 0.8, with a correspondingly high criterion-related validity (henerson, morris, and fitz–gibbon, 1987; isaac and michael, 1995; pedhazur, 1982). 2.3. summary the literature suggests that a financial risk-tolerance assessment instrument must include at least five elements: 1) some central concept of risk, 2) allowance for the derivation of a risk measure, 3) relevance to respondents, 4) ease of administration, and 5) adequate validity and reliability (maccrimmon and wehrung, 1986). assessment items used within an instrument must also meet several other requirements as suggested by maccrimmon and wehrung: 1) cover a variety of risky financial situations in a multidimensional manner including standard versus natural occurring risks, behaviors, attitudes, threats, opportunities, and simple versus complex situations; 2) be consistent and non-redundant; 3) be interesting to complete; and 4) take a limited amount of time to complete. of course, an instrument must also show a high degree of validity and reliability. together the issues that guide the 167j. grable, r.h. lytton / financial services review 8 (1999) 163–181 development of any measurement index, as well as those issues unique to measuring the construct of financial risk tolerance, should serve as a starting point in the development of an assessment instrument. 3. the development and testing of an instrument the following process, as originally outlined by babbie (1983), was used as the framework for the development and testing of the instrument presented in this paper. babbie recommended that an instrument be created by 1) selecting items for an instrument, 2) conducting an item analysis, 3) creating index scores, and 4) testing for index and instrument validity and reliability. during the initial development stage of the instrument over 100 assessment items were originally selected from a review of academic and trade publications. the face validity of these initial items was examined by the researchers. items that appeared to measure something other than financial risk tolerance (e.g., preferences for general risk seeking, tolerances for physical pain, etc.) were eliminated; however, special attention was made to assure that subtle differences in financial risk tolerance, such an investment risks and gambles, were represented. using this method the original 100 items were reduced to 50. using a pilot study of undergraduate and graduate students, data were obtained to examine relationships among the remaining 50 assessment items. it was assumed that these items were valid on their face, and that the items were related to one another empirically. this assumption was based on the understanding that the majority of the items were selected from practitioner sources (e.g., financial planning trade publications and in-house brokerage assessments), and that as such, the items were useful in assisting financial planners and their clients in assessing risk-tolerance attitudes. the data were used to conduct bivariate and multivariate item analyses. a bivariate item analysis was conducted to assure that items were empirically related to each other. correlation coefficients were developed for each pair of the 50 items. it was hypothesized that respondents who appeared highly risk tolerant on one item should also appear highly risk tolerant on the other items. items that showed inconsistency in correlation’s (i.e., respondents who were generally highly risk tolerant tended to be less risk tolerant on an item) were removed from the pool of items. to avoid multicollinearity problems as outlined by babbie (1983), items were also eliminated if there was a very strong relationship between two items. based on this bivariate analysis approximately 30 items were chosen for inclusion in the item pool. to further reduce the item pool two tests were conducted on these 30 items. first, each item that offered respondents a risk-free alternative or a non-response choice was eliminated. this was done to conform to research findings that suggest possible skewing of responses towards non-response categories (kahneman and tversky, 1979). second, index scores were developed for each respondent using the remaining items. answer choices for each item were given a weight (maximum range 1 to 4) according to the riskiness of the response. higher weightings indicated a riskier choice, whereas lower weightings indicated a less risky choice. the index was constructed by summing the weights corresponding to 168 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 each response. these items were evaluated using a multivariate item analysis. specifically, financial risk-tolerance composite scores were regressed on each of the remaining items. this test indicated that 20 items had a strong relationship with the final composite index. this evaluation confirmed that respondents who scored low (or high) on one item generally scored similarly on other items. the final 20 items are shown it table 1. a discussion of each item, as a measure of a dimension of financial risk tolerance, follows table 1. 3.1. item justification the 20 items in table 1 were originally selected from published measures as reported by bernstein (1993), epstein and garfield (1992), goldberg (1995), malkiel (1994), mehrabian (1991), mellan (1994), 1995), pring (1993), shefrin and statman (1985), 1993), statman (1995), tobias (1978),pioneer news(1996), yamauchi and templer (1982), and other researchers. in addition to meeting validity and reliability requirements, the items 1) offer a high degree of face validity, 2) allow for the derivation of a risk measure, 3) offer relevance to potential respondents, and 4) offer ease of administration (maccrimmon and wehrung, 1986). the items in table 1, if used individually, will tend to measure a distinct dimension or limited dimensions of financial risk tolerance. table 2 indicates the dimension of risk each item was originally thought to assess. by understanding how each item works alone, it is possible to see how an instrument using the items in a concurrent assessment of risk tolerance will be more likely to accurately measure a person’s overall financial risk tolerance. note that in some cases an item is useful in measuring more than one dimension of risk. these types of assessment items tend to offer a more accurate measure of someone’s risk tolerance, although items that measure only one dimension are useful as well. a brief description of each risk dimension, and how items within each dimension work to assess financial risk tolerance, follows table 2. guaranteed versus probable gambles require a respondent to make risk calculations. for example, in item 2, the probable chance of winning $100,000 is less than any of the other options; however, in terms of a mathematical calculation, the payout of $5,000 ($100,0003 5%) is greater than the mathematical payout offered in the other answers. as a result, a respondent who chooses the most risky choice, on its face, is considered to have a higher risk tolerance compared to someone who chooses another answer. in addition to item two, items 11, 13, 14, 15, and 20 all offer a respondent a guaranteed safe option with a corresponding probable gain. in every case, a respondent who chooses a gamble over the guaranteed return should be considered more risk tolerant. general risk taking propensity, or general risk choice, can be measured by using items four and 13. for example, according to mellan (1995), a respondent who finds it easy to pass up a bargain when shopping is thought to have a relatively high tolerance for financial risk. this correlation between general risk choice and financial risk tolerance is based on the concept that some individuals view money as a source of anxiety, and because anxiety can be a hindrance to making risky financial choices, an inverse relationship exists between anxiety and risk tolerance. similarly, the interpretation of item 13 suggests that the more risk tolerant individuals would “borrow money from friends and relatives. . . [to] qualify for a 169j. grable, r.h. lytton / financial services review 8 (1999) 163–181 table 1 financial risk tolerance assessment items items 1. in general, how would your best friend describe you as a risk taker? a. a real gambler b. willing to take risks after completing adequate research c. cautious d. a real risk avoider 2. you are on a tv game show and can choose one of the following. which would you take? a. $1,000 in cash b. a 50% chance at winning $5,000 c. a 25% chance at winning $10,000 d. a 5% chance at winning $100,000 3. you have just finished saving for a “once-in-a-lifetime” vacation. three weeks before you plan to leave, you lose your job. you would: a. cancel the vacation b. take a much more modest vacation c. go as scheduled, reasoning that you need the time to prepare for a job search d. extend your vacation, because this might be your last chance to go first-class 4. how would you respond to the following statement? “it’s hard for me to pass up a bargain.” a. very true b. sometimes true c. not at all true 5. if you unexpectedly received $20,000 toinvest,what would you do? a. deposit it in a bank account, money market account, or an insured cd b. invest it in safe high quality bonds or bond mutual funds c. invest it in stocks or stock mutual funds 6. in terms of experience, how comfortable are you investing in stocks or stock mutual funds? a. not at all comfortable b. somewhat comfortable c. very comfortable 7. which situation would make you the happiest? a. you win $50,000 in a publisher’s contest b. you inherit $50,000 from a rich relative c. you earn $50,000 by risking $1,000 in the options market d. any of the above—after all, you’re happy with the $50,000 8. when you think of the word “risk” which of the following words comes to mind first? a. loss b. uncertainty c. opportunity d. thrill 9. you inherit a mortgage-free house worth $80,000. the house is in a nice neighborhood, and you believe that it should increase in value faster than inflation. unfortunately, the house needs repairs. if rented today, the house would bring in $600 monthly, but if updates and repairs were made, the house would rent for $800 per month. to finance the repairs you’ll need to take out a mortgage on the property. you would: a. sell the house b. rent the house as is c. remodel and update the house, and then rent it 10. in your opinion, is it more important to be protected from rising consumer prices (inflation) or to maintain the safety of your money from loss or theft? a. much more important to secure the safety of my money b. much more important to be protected from rising prices (inflation) 170 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 table 1(continued) items 11. you’ve just taken a job at a small fast growing company. after your first year you are offered the following bonus choices. which one would you choose? a. a five year employment contract b. a $25,000 bonus c. stock in the company currently worth $25,000 with the hope of selling out later at a large profit 12. some experts are predicting prices of assets such as gold, jewels, collectibles, and real estate (hard assets) to increase in value; bond prices may fall, however, experts tend to agree that government bonds are relatively safe. most of your investment assets are now in high interest government bonds. what would you do? a. hold the bonds b. sell the bonds, put half the proceeds into money market accounts, and the other half into hard assets c. sell the bonds and put the total proceeds into hard assets d. sell the bonds, put all the money into hard assets, and borrow additional money to buy more 13. assume you are going to buy a home in the next few weeks. your strategy would probably be: a. to buy an affordable house where you can make monthly payments comfortably b. to stretch a bit financially to buy the house you really want c. to buy the most expensive house you can qualify for d. to borrow money from friends and relatives so you can qualify for a bigger mortgage 14. given the best and worst case returns of the four investment choices below, which would you prefer? a. $200 gain best case; $0 gain/loss worst case b. $800 gain best case; $200 loss worst case c. $2,600 gain best case; $800 loss worst case d. $4,800 gain best case; $2,400 loss worst case 15. assume that you are applying for a mortgage. interest rates have been coming down over the past few months. there’s the possibility that this trend will continue. but some economists are predicting rates to increase. you have the option of locking in your mortgage interest rate or letting it float. if you lock in, you will get the current rate, even if interest rates go up. if the rates go down, you’ll have to settle for the higher locked in rate. you plan to live in the house for at least three years. what would you do? a. definitely lock in the interest rate b. probably lock in the interest rate c. probably let the interest rate float d. definitely let the interest rate float 16. in addition to whatever you own, you have been given $1,000. you are now asked to choose between: a. a sure gain of $500 b. a 50% chance to gain $1,000 and a 50% chance to gain nothing 17. in addition to whatever you own, you have been given $2,000. you are now asked to choose between: a. a sure loss of $500 b. a 50% chance to lose $1,000 and a 50% chance to lose nothing 18. suppose a relative left you an inheritance of $100,000, stipulating in the will that you invest all the money in one of the following choices. which one would you select? a. a savings account or money market mutual fund b. a mutual fund that owns stocks and bonds c. a portfolio of 15 common stocks d. commodities like gold, silver, and oil 19. if you had to invest $20,000, which of the following investment choices would you find most appealing? a. 60% in low-risk investments 30% in medium-risk investments 10% in high-risk investments b. 30% in low-risk investments 40% in medium-risk investments 30% in high-risk investments c. 10% in low-risk investments 40% in medium-risk investments 50% in high-risk investments 171j. grable, r.h. lytton / financial services review 8 (1999) 163–181 bigger mortgage” as opposed to an individual who would choose the affordable house with a comfortable monthly payment. financial risk tolerance, as a choice between a sure loss and a sure gain, can be measured effectively by framing questions that require respondents to choose among alternatives without complete information. items seven and 14 are based on findings which suggest that risk-tolerant individuals are more likely to feel a sense of satisfaction when they make money by taking some sort of action with incomplete information (malkiel, 1994; pring, 1993; rowland, 1996). on the other hand, a person who is less proactive in earning a gain because of limited information, yet still receives a significant payout (e.g., receiving an inheritance), is generally less risk tolerant. the literature also suggests that respondents who perceive themselves as experienced investors or more knowledgeable about personal finance issues also tend to be more risk tolerant than others (goldberg, 1995; grable and joo, 1997). items one, five, six, eight, nine, 10, 12, 15, and 18 all have aspects of the question that require some degree of expertise or knowledge to answer the item. for instance, several of these items require specific knowledge and expertise about interest rates, mortgage markets, and investing. these type of assessment items are useful because, if it is true that experience and knowledge are positively related to risk tolerance, a respondent who answers aggressively to these items should, on average, be more risk tolerant than others. highly related to expertise and knowledge is a respondent’s overall comfort level when making a risky choice. items one, three, four, six through nine, 11, 13, 15, and 19 all assess attitudinal temperament towards risk taking. these items fit well with the general consensus that certain individuals share psychological traits that allow them to make risky choices (carducci and wong, 1998). for example, someone may be inherently more comfortable table 1(continued) items 20. your trusted friend and neighbor, an experienced geologist, is putting together a group of investors to fund an exploratory gold mining venture. the venture could pay back 50 to 100 times the investment if successful. if the mine is a bust, the entire investment is worthless. your friend estimates the chance of success is only 20%. if you had the money, how much would you invest? a. nothing b. one month’s salary c. three month’s salary d. six month’s salary scoring 1. a5 4; b 5 3; c 5 2; d 5 1 11. a5 1; b 5 2; c 5 3 2. a5 1; b 5 2; c 5 3; d 5 4 12. a5 1; b 5 2; c 5 3; d 5 4 3. a5 1; b 5 2; c 5 3; d 5 4 13. a5 1; b 5 2; c 5 3; d 5 4 4. a5 1; b 5 2; c 5 3 14. a5 1; b 5 2; c 5 3; d 5 4 5. a5 1; b 5 2; c 5 3 15. a5 1; b 5 2; c 5 2; d 5 3 6. a5 1; b 5 2; c 5 3 16. a5 1; b 5 3 7. a5 2; b 5 1; c 5 3; d 5 1 17. a5 1; b 5 3 8. a5 1; b 5 2; c 5 3; d 5 4 18. a5 1; b 5 2; c 5 3; d 5 4 9. a5 1; b 5 2; c 5 3 19. a5 1; b 5 2; c 5 3 10. a5 1; b 5 3 20. a5 1; b 5 2; c 5 3; d 5 4 172 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 investing in hard assets such as real estate compared to equity investments such as stocks. it is important to assess these varying dimensions of financial risk tolerance. one way to do this is to use word associations. for example, as measured by item 8, respondents who perceive risk as synonymous with loss are generally more risk averse than those who perceive risk as an opportunity or thrill (bernstein, 1993; mehrabian, 1991; pring, 1993). in general, risk tolerant respondents are likely to feel a sense of confidence and satisfaction when making a risky choice; less risk tolerant respondents will tend to shy away from taking risks. items that measure speculative risk, as the name implies, assume that respondents who have a higher propensity to make a speculation are more risk tolerant in terms of their finances than others. items two, nine, 12, 14, and 20 combine other aspects of risk taking by forcing a respondent to either take a safe course of action or speculate on the degree of return offered by a situation. generally, respondents who elect to forgo higher rates of return in pursuit of stability or sure gains are considered to be less risk tolerant than others (malkiel, 1994; mehrabian, 1991). items 16 and 17 were adapted from prospect theory. prospect theory states that investors evaluate their choice in terms of potential gains and losses relative to some reference point (shefrin and statman, 1993). item 16 is described in terms of gains, whereas item 17 is table 2 dimensions of risk assessed by each item ite m g ua ra nt ee d vs . pr ob ab le ga m bl es g en er al ris k ch oi ce c ho ic e be tw ee n su re lo ss an d su re ga in r is k as ex pe rie nc e an d kn ow le dg e r is k as a le ve lo f co m fo rt s pe cu la tiv e ris k p ro sp ec t th eo ry in ve st m en t ris k item 1 x x item 2 x x item 3 x item 4 x x item 5 x x item 6 x x x item 7 x x item 8 x x item 9 x x x x item 10 x item 11 x x item 12 x x x item 13 x x x item 14 x x x item 15 x x x item 16 x item 17 x item 18 x x item 19 x x item 20 x x 173j. grable, r.h. lytton / financial services review 8 (1999) 163–181 described in terms of losses. this distinction is subtle but profound. in both cases the mathematical payout (i.e., cash flows) are identical. individuals who choose the sure choice act consistent with risk aversion theory, whereas those who choose the gamble in both cases are most likely risk takers (statman, 1995). when combined together by averaging answer weights into a single score, these two items work well in predicting financial risk tolerance. most individuals, assuming they have no prior knowledge of the items, choose the sure gain in item 16 and the chance in item 17. this would indicate a person with moderate risk tolerance. again, a person who chooses both the sure gain in item 16 and sure loss in item 17 will generally exhibit risk-aversion characteristics, whereas someone who takes a chance in both situations will exhibit risk-taking characteristics. a respondent’s propensity to take direct investment risks is assessed through items five, six, nine, 12, 18, and 19. these items combine the attributes of knowledge and temperament in the assessment of risk tolerance. knowledge and temperament tend to determine a respondent’s ability to deal successfully with emotional investments (mehrabian, 1991). as such, a person who is willing to invest money, either earned or gifted, into equities, real estate, or hard assets, instead of choosing to hold less volatile investments, is considered to be more risk tolerant than others (bernstein, 1993). in summary, the 20 individual items were originally determined to measure either one or several dimension of financial risk tolerance. at a minimum, the 20 items measured at least eight dimensions of risk, including: 1) guaranteed versus probable gambles, 2) general risk choice, 3) choice between sure loss and sure gain, 4) risk as related to experience and knowledge, 5) risk as a level of comfort, 6) speculative risk, 7) prospect theory, and 8) investment risk. although individually no one item was sufficient to accurately assess financial risk tolerance, it was concluded that when combined together these items could provide a useful and accurate measure of a person’s financial risk tolerance. 3.2. initial use of the instrument with a research sample the next phase of this study involved administering the 20-item instrument to a larger group of respondents. a convenience sample of faculty and staff from a southern state university (n 5 1,075) was chosen. a modified dillman (1978) method was used to direct the management of the survey. specifically, one-half of all employees (approximately 2,000) received an instrument. a reminder card was mailed two weeks after the first instrument was sent. a duplicate instrument was then mailed one week later. after adjusting for missing data and unusable responses it was determined that the survey had a useable response rate of 54%. approximately 55% of respondents were female. seventy-two percentage were married. respondent ages ranged from a low of 20 to a high of 75 years, with an average of 43 years. incomes ranged from less than $20,000 to over $90,000. respondents who were employed in staff positions outnumbered members of the faculty (61% and 39%, respectively). the majority of respondents (63%) possessed a four year college degree or higher, whereas the remainder (37%) had an associate degree, high school diploma, or less than a high school education. seventy-seven percentage of the sample indicated that they had a somewhat vague or moderate knowledge of investments and personal finance issues. likewise, 77% of the sample indicated that they expected future economic conditions over the next five years to 174 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 be about the same or worse. in general, this convenience sample represented populations found on most four-year college campuses, namely, a group with slightly higher attained education, income, and socioeconomic levels, on average, than the general population. future research using the items and instrument should incorporate different sample frames and populations to confirm the generalizability of findings. financial risk-tolerance scores were determined by each respondent’s score on the 20items in table 1. an index was constructed by summing the weights corresponding to each response. higher scores represented higher levels of risk tolerance whereas lower scores represented lower levels of risk tolerance. the average risk-tolerance score was 37, with a standard deviation of 6.40, and a range of 20 to 63. the reliability estimate for the 20-item instrument was 0.78, indicating an acceptable level of consistency (henerson et al., 1987; pedhazur and schmelkin, 1991). these data were useful in confirming that, overall, 27% of respondents were classified as having low risk tolerances. the majority of respondents (60%) were classified as having moderate risk tolerances, with 13% being classified as having high-risk tolerances. these results were consistent with distributions of financial risk-tolerance scores found in the literature (maccrimmon and wehrung, 1985). table 3 shows the mean and standard deviation scores from the sample for each of the 20 items, as well as correlation’s between each item and the composite score. the correlation range (0.20–0.67) indicated a weak to moderately strong relationship between the individual items and the index score. these bivariate analyses suggested the need for further analysis to refine the scale. table 3 means, standard deviations, and correlations for the 20 risk assessment items (n 5 1,075) item mean standard deviation item correlation with index score item 1 2.48 0.64 0.54 item 2 1.87 0.94 0.58 item 3 1.85 0.87 0.43 item 4 2.00 0.64 0.20 item 5 2.13 0.84 0.67 item 6 1.83 0.97 0.53 item 7 1.46 0.76 0.28 item 8 2.12 0.57 0.44 item 9 2.47 0.78 0.35 item 10 1.93 1.00 0.43 item 11 2.31 0.65 0.49 item 12 1.55 0.58 0.27 item 13 1.45 0.56 0.42 item 14 2.27 0.94 0.63 item 15 1.78 0.52 0.30 item 16 1.66 0.94 0.47 item 17 2.39 0.92 0.41 item 18 2.02 0.80 0.52 item 19 1.68 0.65 0.62 item 20 1.58 0.71 0.45 index 37.00 6.40 n.a. 175j. grable, r.h. lytton / financial services review 8 (1999) 163–181 as was the case in the original pilot study, a multivariate item analysis was conducted by regressing financial risk-tolerance composite scores on each of the 20 assessment items. this test indicated that indeed each assessment item had a strong relationship with the final composite index. external item analysis was measured by comparing individual assessment item scores to other item scores and the total index score. this test was conducted to confirm that individual index items provided similar scores for both those with high and low levels of risk tolerance. results of these item analyses were consistent with the pilot study findings. in general, it was determined that persons who were categorized as having low risk tolerances tended to be less confident in their investment behaviors, less aggressive in their investing behaviors, and more likely to avoid risky financial situations than those who were categorized into higher risk-tolerance categories. 3.3. factor analysis application to further explore the issue of multidimensionality in the instrument and to ensure a parsimonious measure, principal components factor analysis was performed on the 20 items. the purpose of factor analysis is to reduce and summarize data by identifying the underlying, or common interrelationships, which can then be conceptually explained, or named, as factors. as such, this phase of the analysis served two purposes relative to the risk tolerance assessment recommendations offered by maccrimmon and wehrung (1986). the first purpose of the factor analysis was the identification of the underlying dimensions, or factors; this ensured that within the 20 items the instrument offered a multidimensional approach to financial situations yet focused on the central concept of risk. the second purpose of the factor analysis was the elimination of items that did not significantly contribute to the measurement of the underlying dimensions. this ensured that the instrument was brief, nonredundant, and interesting to complete. four statistical criteria are commonly considered in the analysis and interpretation of principal components factor analysis. the eigenvalue-one criterion, the screen test, the proportion of variance accounted for, and the interpretability of the resulting factors (i.e., composite dimensions) were each considered. the first three criterion are used to determine the optimum number of factors, or underlying dimensions, which can be extracted from the data. these are considered in conjunction with interpretability. rotation, or the turning of the reference axes of the factors about the origin, is often used to effect a factor solution that is simpler or more theoretically meaningful. for this analysis, varimax rotation was used to simplify the factor loading structure and to increase interpretability. with varimax, an orthogonal rotation, the factors remain uncorrelated, and the sum of the variance accounted for by the factors does not change. based on the eigenvalue greater than one criterion, the principal components factor analysis generated a four-factor solution that accounted for 38.6% of the total variance. results of both the unrotated and varimax rotated solutions were considered for interpretability. based on the screen test, additional factor analyses with unrotated and varimax rotated solutions were limited to three factors. the latter results were interpretable, with clean loading across factors. this solution accounted for 33.3% of the variance, or a slight 176 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 reduction from the initial “best linear combination” solution generated solely on the basis of the eigenvalue-one criterion. in interpreting the rotated factor pattern, shown in table 4, an item “loaded” on a factor if the factor loading was equal to or greater than 0.45. these items were used in the labeling of the factors, with the exception of item two that was included although the loading equaled 0.4442. excluding this item significantly reduced the cronbach’s coefficient alpha reliability measure for the factor (from 0.4425–0.3027), as well as for the index (from 0.7507–0.7274). the resulting three-factor solution consisted of 13 items. consequently, seven items (numbers four, seven, nine, 10, 11, 13 and 15) were omitted from the instrument because they lacked sufficient loadings to support the internal consistency of the factors. as noted earlier, other researchers have recommended that a risk-tolerance assessment index must produce a reliability coefficient in the range of 0.5 to 0.8 to insure consistency (e.g., henerson et al., 1987; isaac and michael, 1995; maccrimmon and wehrung, 1985; pedhazur, 1982). the cronbach’s coefficient alpha reliability measure of 0.7507 for the instrument falls within the upper end of this range. reliability estimates for the three factors, as shown in table 4, are not as strong. however, the underlying factors were not intended to be used as distinct measures. this analysis was done to demonstrate the multidimensionality of the instrument. results from the factor analysis suggest that the 13-item instrument measured financial risk tolerance on three constructs: 1) investment risk, 2) risk comfort and experience, and 3) speculative risk. these, in turn, encompassed all of the assumed dimensions as reported in table 2. as such, it was concluded that the 13-item instrument offers a strong degree of multidimensionality in the measurement of financial risk tolerance. specifically, the 13 items, when combined into one instrument work together in the assessment of 1) the probability of gains; 2) the probability of losses; 3) the dollar amount of potential gains; 4) table 4. factor loadings for three-factor principal components solutiona factor 1 factor 2 factor 3 factor 1. investment risk (a 5 .720) item 5 .744 .193 0.162 item 6 .635 .095 0.027 item 19 .604 .268 0.192 item 18 .472 .228 0.141 item 14 .465 .401 0.211 factor 2. risk comfort & experience (a 5 .502) item 1 .177 .590 0.210 item 8 .142 .528 0.096 item 12 .132 .503 20.340 item 20 .151 .492 0.129 item 3 2.101 .459 0.387 factor 3. speculative risk (a 5 .443) item 16 .016 .234 0.587 item 17 .134 .004 0.577 item 2 .316 .237 0.444 a total instrumenta 5 0.7507. 177j. grable, r.h. lytton / financial services review 8 (1999) 163–181 the potential dollar loss through the assessment of guaranteed versus probable gambles; 5) minimum probability of success given a risky course of action; and 6) minimum returns given a risky course of action. these situations are accounted for by the inclusion of assessment items that measure choices and tolerances for guaranteed versus probable gambles, prospect theory dilemmas, and potential versus probable gains and losses. 3.4. validity issues the final step in the development of the resulting 13-item parsimonious risk-tolerance assessment instrument involved a test of validity. this test was conducted by analyzing the instrument’s construct validity, which is defined as the extent to which one can be sure the index represents financial risk tolerance (henerson et al., 1987; litwin, 1995; silva, 1993). “demonstrating concurrent validity of an instrument provides good evidence of its construct validity” (henerson et al., 1987, p. 143). concurrent validity tests a measure against another measure that has proven psychometric properties. concurrent validity is calculated as a correlation coefficient (litwin, 1995). in addition to the 20 items in the data collection instrument, respondents were also asked to answer a risk-assessment item developed by the national opinion research center at the university of chicago under the sponsorship of the federal research board as originally asked in the survey of consumer finances (scf). the scf is used to gather data on assets, liabilities, financial attitudes, and financial behaviors of individuals and families. the scf questions asks: which of the following statements on this page comes closest to the amount of financial risk that you are willing to take when you save or make investments? 1. take substantial financial risk expecting to earn substantial returns 2. take above average financial risks expecting to earn above average returns 3. take average financial risks expecting to earn average returns 4. not willing to take any financial risks the scf risk assessment item has been widely used as a proxy for financial risk tolerance, although no published documentation exists to substantiate the validity of this item. however, based on the use of the item in published research (e.g., chang, 1994; grable and lytton, 1998; sung and hanna, 1996; yuh and devaney, 1996), one can assume at least a moderate degree of item validity. also, scores on the item have been very consistent over time, suggesting a high level of reliability. correlation analysis between the scf item and the index scores derived from the 13-item instrument yielded a coefficient of 0.5383. it was determined that a single item, like that of the scf, is not able to measure wide variations within the overall dimension of financial risk tolerance. thus, it is not surprising that the coefficient was moderate. the positive correlation indicates that although both are measuring risk tolerance similarly, the larger 13-item index appears to be measuring multiple dimensions of financial risk tolerance that are not accounted for in the scf item. in summary, the 13-item financial risk-tolerance assessment instrument was found to meet the requirements for a multidimensional financial risk-tolerance assessment instrument as 178 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 outlined in the literature. furthermore, the assessment of reliability and validity, based on testing with this convenience sample, support the potential usefulness of the instrument. in summary, the variety of items included, the high degree of validity and reliability, and the potential ease of administration make this a viable assessment tool for researchers, academics, and practitioners. 4. conclusions financial service providers and researchers, in their respective roles as managers, consultants, and investors, share the common objective of quickly assessing individual financial risk tolerances and preferences (both their own and their clients). unfortunately, according to snelbecker, roszkowski, and cutler (1990) all too often financial service providers, researchers, and household financial managers have little more than qualitative descriptions and intuitive subjective judgments for use in understanding financial beliefs, feelings, needs, and aspirations that affect risk tolerances. instead of relying on a standardized measure of risk tolerance or empirically tested risk and investment rules, many individuals rely on one-dimensional assessments, objective measures, and other heuristics to gauge their own or someone else’s risk-taking propensities. as indicated above, these methods of risk assessment are less effective than using a multidimensional risk-assessment instrument. the original 20-item, and the factor reduced 13-item, instrument extend previous research into risk assessment whereas offering a solid foundation in the development of a widely accepted instrument. financial service providers, educators, and researchers are encouraged to use the instrument as a tool for quickly and accurately assessing the financial risk tolerances of clients and other respondents. further tests of the instrument, both in random surveys and experimental settings, will lead to improved reliability and validity of the instrument, as well as to the eventual development of a financial risk-tolerance assessment instrument for use in private and public organizations. ultimately the continued use, evaluation, and adaptation of this instrument will have a positive impact on the daily practices of financial service providers, and most importantly, on the lives of financial services clientele and constituencies. references babbie, e. r. (1983).the practice of social research.belmont, ca: wadsworth publishing company. bernstein, j. (1993).the investor’s quotient: the psychology of successful investing in commodities and stocks. new york: john wiley & sons. bernstein, p. l. (1996).against the gods: the remarkable story of risk.new york: john wiley & sons. carducci, b. j. & wong, a. s. (1998). type a and risk taking in everyday money matters.journal of business and psychology, 12,355–359. chang, y.r. (1994). saving behavior of us households in the 1980s: results from the 1983 and 1986 survey of consumer finance.financial counseling and planning, 5,45–64. cutler, n. e. (1995). three myths of risk-tolerance: what clients are not telling you.journal of the american society of clu & chfc, 49,33–37. 179j. grable, r.h. lytton / financial services review 8 (1999) 163–181 dillman, d. a. (1978).mail and telephone surveys.new york: john wiley & sons. droms, w. g. (1988). measuring investor risk capacity with a pass.journal of financial planning, 1, 88–92. elvekrog, m. (1996). psychological ‘unbalance’ in investing. a matter of emotion, not analysis.better investing, october, 9–11. epstein, i., & garfield, d. (1992).the psychology of smart investing.new york: john wiley & sons. goldberg, s. t. (1995). yes, but how do you feel about this investment?kiplinger’s personal finance magazine, august, 68–71. grable, j. e. & joo, s. (1997). determinants of risk preference: implications for family and consumer science professionals.family economics and resource management biennial, 2,19–24. grable, j. e. & lytton, r. h. (1998). investor risk tolerance: testing the efficacy of demographics as differentiating and classifying factors.financial counseling and planning, 9,61–74. haliassos, m. & bertaut, c. c. (1995). why do so few hold stocks?the economic journal, 105,1110–1129. henerson, m. e., morris, l. l. & fitz–gibbon, c. t. (1987).how to measure attitudes.newbury park, ca: sage. isaac, s. & michael, w.b. (1995).handbook in research and evaluation(3rd ed.). san diego, ca: educational and industrial testing services. kahneman, d. & tversky, a. (1979). prospect theory: an analysis of decision under risk.econometrica, 47, 263–291. litwin, m. s. (1995).how to measure survey reliability and validity.thousand oaks, ca: sage. maccrimmon, k. r. & wehrung, d. a. (1985). a portfolio of risk measures.theory and decision, 19,1–29. maccrimmon, k. r. & wehrung, d. a. (1986).risk management.new york: the free press. malkiel, b. g. (1994).a random walk down wall street(6th ed). new york: w.w. norton & company. mehrabian, a. (1991).your inner path to investment success: insight into the psychology of investing.chicago: probus publishing company. mellan, o. (1994).money harmony: resolving money conflicts in your life and relationships.new york: walker and company. mellan, o. (1995). beyond reason: to build good rapport, consider your client’s money personality.fee advisor, november/december, 62–65. opiela, n. (1996). business practice management: charles haines, jr.registered investment adviser, 2,ria1, 6. pedhazur, e. j. (1982).multiple regression in behavioral research(2nd ed). fort worth, tx: harcourt brace javanovich, inc. pedhazur, e. j. & schmelkin, l. p. (1991).measurement, design, and analysis: an integrated approach. hillsdale, nj: lawrence erlbaum associates. pioneer news. (1996). what type of investor are you?author, 10(1), 1–2. pring, m. j. (1993).investment psychology explained: classic strategies to beat the markets.new york: john wiley & sons. roszkowski, m. j. (1995).risk tolerance assessment.bryn mawr, pa: the american college. roszkowski, m. j. (1998). risk tolerance in financial decisions. in: cordell, d. m., editor.readings in financial planning (pp. 281–328). bryn mawr, pa: the american college. roszkowski, m. j. & bean, a.g. (1990). believe it or not! longer questionnaires have lower response rates. journal of business and psychology, 4,495–509. roszkowski, m. j. & snelbecker, g. e. (1989). how much risk can a client stand?best’s review (life/health insurance edition), 90, 44–119 roszkowski, m. j., snelbecker, g. e. & leimberg, s. r. (1993). risk tolerance and risk aversion. in: leimberg, s. r., satinsky, m. j., leclair, r. t. & doyle, r. j. jr., editors.the tools and techniques of financial planning (4th ed., pp. 213–226). cincinnati, oh: national underwriter. rowland, m. (1996). the risk quiz.bloomberg personal, 1,155–158. schooley, d. k. & worden, d. d. (1996). risk aversion measures: comparing attitudes and asset allocation. financial services review, 5,87–99. shefrin, h. & statman, m. (1985). the disposition to sell winners too early and ride losers too long: theory and evidence.the journal of finance, 40,777–792. 180 j. grable, r.h. lytton / financial services review 8 (1999) 163–181 shefrin, h. & statman, m. (1993). behavioral aspects of the design and marketing of financial products.financial management, 22(2), 123–134. silva, f. (1993).psychometric foundations and behavioral assessment.thousand oaks, ca: sage. snelbecker, g. e., roszkowski, m. j., & cutler, n. e. (1990). investors’ risk tolerance and return aspirations, and financial advisors’ interpretations: a conceptual model and exploratory data.the journal of behavioral economics, 19,377–393. statman, m. (1995). a behavioral framework for dollar-cost averaging.the journal of portfolio management, fall, 70–78. sung, j. & hanna, s. (1996). factors related to risk tolerance.financial counseling and planning, 7,11–20. tobias, a. (1978).the only investment guide you’ll ever need. new york: harcourt brace jovanovich. train, j. (1995). buy them high, sell them low.financial times, 12,b14. yamauchi, k. t. & templer, d. i. (1982). the development of a money attitude scale.journal of personality assessment, 46(5), 522–528. yoo, p. s. (1994). age dependent portfolio selection.federal reserve bank of st. louis working paper 94–003a, 22. yuh, y. & devaney, s. a. (1996). determinants of couples’ defined contribution retirement funds.financial counsel and planning, 7,31–38. 181j. grable, r.h. lytton / financial services review 8 (1999) 163–181 pii: s1057-0810(02)00092-6 ��������� �� ���� ����� �� � ������ ����������� � ����� ���� � �� �� � �� ������ ����� ������� ����� � ��� � ������ �� �� �� �� ���� � ��� � �� � ��� �� ��� �� �� ��� ������� � � ��� ��� ���� �� !"#� ��� � � ! " #� �$ � %&&'( � � � �� � �� � ���� ' )� �$ � %&&'( � #� � '* + �$ � %&&' �������� ,��� ����� �##�� � ��� �� #�������� �� ��� �� �� ���� ����� ��� � ���� ����� � ����� �� �� �� ���� �� �� � ����� ���� �� ,� ����� ������ � ���� ���� -������ '..' �� -������ '...� ,��� �/� 0# �� � 1��� ���� ��������� � #� ������ �� ��� ��� � �� ��� � � ���� �� ���� ���1��� ����� #�� �� � ���2� ���� ��� � ����� �� �� 0���� # ������ #�������� ��� � ���� ������ " � �� ��� ��� �� ����� � � ��� �� ������ �� �� #��������� �� ������ �� ,������� � ���� ��� 3� �� �� ��� � ���� ��� ����� ��� '%4����� $��� � #����� � ������� � �� ,������ ������ � ���� ������ ��� 5�"� �������� ,� ������� ���� �� ���� �� ����4� ������� � $��� � #���� �� �� �������� �� ��� �� �� � � � %&&' ��� � � " � � 6� � 3�� ������ � � � �� $%� �� � � � ��& �''( �'! '��(��� & +� ���� �����( ��������� � � ����( �� ����� ���� ��( ,������ ���� � � ���� ������ ,��� ����� 0���� � �� �##�� ����� �� ��� �� #�������� �� ��� �� �� # ������ ��� � ���� ����� � ����� �� �� �� ���� �� �� ,������ ���� ��� ���� ��� ,��� � ��� 0# �� � � ���� ��������� $��� ���� � � ���� � #� ������ �� ��� ��� � �� ��� � � ���� �� ���� ���1��� ������ ,��� � �� � � ���� � � �� �� ������� ��� �� ���� ����� �� ����� 6� ���� � ����� �� 1 0���� �� ��� �� � ������ ����� ���� � � ���� � �� ��� ����� � ��� �� ������ �� �� # ������ #�������� �� � ,������ �� ����� ,� � ��� ����� � �� ��� �� ����� �� �� ,������ ��� � ���� �� �� �� � ���� ��� � ���� �� �� �� 3� �� �� ��� � ���� �� �� ����� ��� '%4����� $��� � #����� � ������� � �� ,������ ������ � ���� ������ ��� 5��� � "��� � �������� 7���� ��� " � � � � 1 '& 8%&&'9 *&*:*'; '&!<4&='&>&'>? : � ����� ���� � � %&&' ��� � � " � � 6� � 3�� ������ � � � �� �66@ " ' & ! < 4 & = ' & 8 & % 9 & & & . % 4 a � b��� �#������ ������� , ��@ �'4%='4%=*4*%'*( ��0@ �'4%='4%=*4*.!'� %�� � ���� � & �����c ����� �� 8���� �����9� ������c ����� �� 8�� ������9� # ��� c ����� �� 8�� � ��� 9� ��� �� #�������� �� ��� ���� � ���� �� ����� �� � �� ���� $� �� ����� �� � #�������� �� ��� ��� 5���� ���� ��� ������� ��� �� � ��# � ���� ��� ���� �� 1��� ���$� ����� ���������� �� ��� $ � 1 �� ���$���� � ���� �� � �� � ���� � �� ���� $ � �� ������� #�������� �� ���� ����� 1��� ����� �� �� � ���� ��#� � �� � ����� ,� #�������� �� ���� ����� ����� �� �� ���� �� ������ � ��� ����������� $� ������ ���� � ���� ��# �� � ��� ������ $ �1 � ���� ��� ���� �� �� ������ � �� �� �� ��1 �� �� �� �� ���� ����������� � ����� ��� ��� � ����������� �� �� ,������ ���� ��� ���� �� �� ������� �� ���� ����� �� ��� ���1 �� ,������ �� ���� �� �d����� ��� ������ ��� �� ���� ����� $ � �� ���� �� ����� ��� ������������ 6� �� ��� � ����� �� ��� ��1 �� ���� ����� �� ���� �� �� �����$� � ����� ���� � � ����� �� 1���� �� ��� �� #�������� �� ��� �� $ � �� ���� �� ���� ����� �� � �� �� ���� � ����$� �� ��� �� ������� ���� 5����$� ����� ��������� �� � �� �� �� ������� ���� ��� � 1��� ������ �� ��� �� ��� ����� �� 3 ����� ����������� �������� �� �� ����� 1���� � ���� ��� �� ��� �� � ������ ���� ������ e���� �� � �� �� ���� ���� �� � #�������� ������� �� e �� �� #���������� �� 1���� �� $ ��$���������� ���� � ��f ,��� � �� � ���� �� ��� ��� ���� ���� �� ������ ,��� � �� � e �� ��4������� $�� � ��#��� ������ �� �� �� �� ��� �� ����# � ,��� � ��� $ � #���� � 1��� ���� �������� ��� ������ ��� ��� � � � ���������� $�� �� #����� ���� � �� ���� ��� � ����� �� �� �� � ���� �� ��� ���$� ������ � �� �� � ����� 7������� ,������ �� ����� �� �� �� �$����� �� �� �� ������ �� ������� �� ��� �� �#���� ����� � ���� ����� ���� �� �� ��� ��� ,��� ����� �����$�� � �� �� ��� ����� $� #�� ����� � $ �� � ��� ��������� �� �� ���� ����� ���� ��� � ����� ������ #�� �� �� � ������ ����������� � ����� ���� � 1� � �� ����� � ����� �� �������� �� � � ���� �� �� ��� �� ��� ����� �� 6� ����� �� �� ����� � �� � �� �� ���� �� ���� ����� ����� �� �� ��� �� #�������� �� ��� �� � ��# �� � �� #���������� � ������ �� ��� ���������� ,� $ � ��� �� #�������� �� ���� ����� �� 1 �� �� �� �� � �� � ��# � ���� ��� ���� ��� �����1��� 8'.!%9 #�� �� � �� $��� ��� �� #�������� �� ��� ���� 1��� ��� ��������� �� ���� ��� � ���� � ���������# �� � #�������� �� ��� �� ��� � ��# � �� � ��4 ����� #�������� �#���������� ��� �� g� � � ���� ���� � �� �� ��� �� �� � � ���� �� � �� ���� # �� ��� #����� �� ��� ��� �� �� ����� �� �������� �� � ���� ��� �� � ���� 1��� �� ��1 ���� �� � � �� ����� �����1���/� �� #� �� ��� �� #�������� �� ��� 1�� ��������� �##�� � �� �� �� ��� �� �� � � � ����$� �� �� 5��� � "��� �� 6� �� ���� � 1 � �� � �� �� ���� ����� ����� �� ��� $ � ���� � ����� � �� �� ��� ��� ��������� �##�������� �� ���$ � 8'.a=9� h � ��� "����� 8'.<&9� ��� ��� ��� ���1 � ���� �� ������ ��� ��� ��� �� �� � #�������� ��#�� � #�������� �� ���� ������ "����� 8'.<;9 ���� � �� ii#������ ���� ����� �� ������� � #�������� �� ��� �� �� �� � �� ����//� ��� � � �������� � ���� ��� ��������� �� ���� ����� �� ��1 � �� ���� ���� ���� �� � #��������� ��� ��� ��� � 8'.==9 0���� � �� � � �� ������ ��� 0 ���� ��� ��������� 1� � ������� �� ��� ��������� �� ������ g�� � ��� h��� ���� 8'..'9 ��� �� � ���� �� �� � �� *&; )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 ������ � �� #��������� 0��$�� � � �� ���� ��� � � � � �� � ���� 1� � ��� ��������� �� ��� ��� 1 � �� ��� �� ,� �� �� ������ 1�� ���� ��� ��������� �� ����� ��� � �� �� � � �� ���� $ ���� �� ����� ��� �� �� � ��� ��� #��������� b���� ��� ,����� 8'...9 ���� �� � ���� � ����4��� ���� ����� �� %&:*&j �� ��� ��� 2���� ��� �� �� ���� �� �� 5��� � "��� �� �## �� � �##��#���� $�� � �� ���� � �������� �� ���� � ��� 8'..a9 �� �� � �� $��� � ���� ��� ���� ��� �� ��� �� ii��� ��������� �� ���� ����� �� � �� � � ���� # � ���� �� ���� � � �//� � ��� �� ���� ����� �� ���� �� ���� �� 1� �� � �� 2���� ���� � �� � � ��#��� ������ �� #�� �� �� ����� 1��� � � �� ������� ��������� 6� �� � � ����� �� � � 8� -��� ��� ���� �� �� '.<a( 7��� ��� " �1 ��� '.<<( 7���� '.='( ����� '..<9 ���� �� �� 5�"�� �� 2���� ���� � ��� � �� � ��� #�� �� �� � � � �� � ��1 �� �� � �� ��� �� � 8���#�� ��� "���� %&&'9 �� �� �������� "���� � �� 0� �� � ���� 2� ����� �� � ��# � ��� ��� ���� �� 8g��� �� '.<;( b��� ��� h ������ '.='9� 1��� �� � �� ��� � ����� 6� �� �� 2� ������$� �� �� 1� �� � �� ��d�� 2���� ���� �� � �� ������� �������� �� � ��# � ������ �� �� �� � � �� ��� �� ���� �������� 1��� $ � �� � �� �� 2���� ���� �� �� � ��#��� ������ �� ��� � � ��� 8%&&&9 ����� ���� �������� ��� � ������ ��� ��#� � �� � ����� ���� � � ����� ��� �� � �� � ������� � �� ���� � #�������� �� �� ������ �� �� �� e��� ���1 ���� �� 2� ����� �� ��1 �������� ��#� �� � ����� ���� � � ����� �� ��� #��� �� ���� ������ �� ��� $ ��� � ���� �� 6����$�� "�� � �0 ���� 86"�9 ��� 0 ���� � �� � � �� ������� �������� �� ,��� � ������ �� # ���� �� ���� ����� $� #���� ��� � � ���� � �� ��� �� � ����� � �������� �� ��� ������� ������� �� ������ ������ 2*.* ) ��������� � ���� � �� ,� ,������ ����� ��� $ � ��� ������ ��d�� ���� ����� ���� � �� �� ���� � 1 � �� �� "��$��������� #������� ��� � �� ��� ��� ����� #��$� �� �����1 � � � �� � �� ����� ��� �� ,� � #��$� �� �� �� ��� � �� ���� � � �� ��������� $��� � � � ���� ��� �� �� ������� ��� � $� �� �� 2� �� ���� �� ��� ���� � ������ ,� ���� ����� ���� ����� ������ �� ����� # ���� �� � #��� � �� ,�$� '� ,� ���� ����$� �� � �� �� ������ �������� ��� ������ �� ����� # ����� ������� ���� *%�' �� ';.�aj� 6� ���� # ���� �� ,������ ����� ���� 0# �� � � � ��� �$� �� � �� �� �� � � �� ��� ��� ��� 0� ���� � $� � ���� �� ����� �������� #���� � 8�)�9� 7�� 0��#� � �� 0� ���� � $� �� � �� � ���� �$��� ?!& $������ �� '..& �� ?'''�% $������ $� '...� ,� ����� �� � � ��� ��� $����1��� � � �)� ��� �� � �� � ���� !j �� '..& �� ';�!j �� '...� "��������� �� ����� �� ��� ��� � $� �� �)� ������ ���$� � �� �� ���� '& � ���� ,� ���� � � �� ��� ��� ��� ��� ��������� $����1��� ��� $ � �� � ���� �� � ���� �� ���� $��� � � � ���� ��� � � �� �� � �� �� ��� ���� $����1��� ��� � #� ��� � �� ��� � � ���� �� ���� ������ �� # ����� ,� ��� �� �� ������ �� �� ��� � ����� ��� ���1� $� ������ *&j �� � �� � ���� ,� �)� ���1�� ��� ��� � ���� � � ���� �� ���� ������ �� # ���� 1��� �� 0 #���� �� '..; ��� '...� 1� � �� ������ 0# �� � � ��d�� ����� ��� �� "��������� # � �#��� �� �� )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 *&! ,�$� ' � ����� ���� ����� k����$� '..' '..% '..* '..; '..! '..a '..< '..= '... �)� 8$������ ?9 '!& '!= '<=�< '*% '<& '=; '.; %&; '=! �)� ���1�� 8j9 &�; ;�* <�a �a�' =�' <�! =�& *�= �a�; 6������� 8j9 !.�% a'�; a&�* ';.�a a;�. ='�. .&�a !'�; aa�! � � �#��� �� �� 8?9 %�a!! %�<;; *�&!a %�'a' %�<== *�&&& *�'&! *�%'* %�=<= +�� ��� � $�>�)� 8j9 '!�; '<�a '<�. %&�a '<�* %' %'�; %'�. %.�* ) � $����1���>�)� 8j9 !�' .&�% =�' ''�; <�% ''�a '&�! '&�% ';�! �0� ���� � $� 8$������ ?9 !&�! !!�a a<�* a;�; <*�* =;�' .'�! '&a '''�% 6�� � �� ��� 8,4$����9 8j9 a= <; <& '&! =! '*% '&= ''a '&; ,h>? 8� �� ��9 !!;�. =*!'�a ';&;<�= *<;;&�* !a<%.�= '&;<&!�< '..!*&�a *&a.&=�. !%<'&!�% ,h>+� 8� �� ��9 *%*'�; !%=&�' =%%;�' %*<==�; *.*a!�. a<;=%�; ''%*%=�' '=aa!%�' %<%!%=�. 6����$�� "�� � �0 ���� 7���� '*; ';! 'a& '<a %&! %%= %!= %<< %=! ���� � �#����������� 8$������ ?9 '!�! .�. *<�= %'�= %&�= *&�= a'�; **�; ''%�% ,��� � ��� 8������� ?9 =�!&% =�!a< %'�<<' %*�%&* !%�*!< *<�<*< !=�'&; <&�*.a =;�&*; "��� @ 5�� �� � ���� �� 7�� ��� ,��� � ,� g���� 3��� ������ �� ,��� �� ��� ,� b ����� g��� �� ,��� �� * & a ) *% * � � � �� �*+, � � � � � ��� �� �� �( . ! /0 ! ! . 1 2 ! 2 3 2 . 4 �� � �� � �##��0���� �� %!j ������ �� ����� # ����� e��� � �# � �� 0 ���� ��� �� �� ,������ ���� ��� ������ � �� � #� ��� ��#���� ������� ��d�� ��� � � � �� � � ���� �� ���� ��������� 6� �������� � � �� #��� � 1 � �� � ,��� � ������� � ����� #��� � � ������ �� 1��� ��#�� � ��#� �� �� � ������� ��� ���$���������� � ������� �� ���� ��#������ ����������� ��� ��� ����� � ��#� ��� 3� �� ��� ��� �� ������ � �� �� ��d�� #��$� �� 1��� �������� ��� �� $��� � � � ���� ��� �� ������� �� ���� ��� ���� �� �� � � ���� ,��� #� � ��� �� ��� � ����� ����� � ����� �� ��� �� ��� �� � ������ $� ���� ����� �� ������ 2*0* ���� � 5� �� � � ���� ,��� � ��� � ���� � '!&4� �� ������� �� �#���� ���� ��� ,� ���� � ����� � ���� � �� �� l������ ��#�� 1�� ���$���� � �� '=aa �����1��� �� b��� �� e�� ��� � �� ��� �� �� ii+ ����� � " ����� � �0 ���� //� ,��� 0 ���� � �� � �� �##�������� ��� ����# �� �� ����� 1�� 1 � � ���� ���� � � ����� �� l������ ���� ��� ,� 0 ���� 1�� ��� � �� � � ���� �� �� ���$� �� �� e���� e�� 6 ��� �� � ������� ������ ������� �� �� l������ ��#�� � 3�� � ���$����� �� �� �� #�$�� �� ,��� � �� '.%*� � � 1 ��1 1�� �� � � �� '.%. �� � ����� �� 0 ���� ��� � �� ��� �� �� ii6����$�� " ����� � ��� 7�� ��� �0 ���� g���� //� ,��� ���� � � ��# � 2�� ��� ��� $ �� �� � �� �� � ���� �� ������� � 2��� � ��� �� � 1 �� �#��� � �� ,��� �� ��1 �� �� �� �� �� �� �� �� + #� ����� ��� �� ��# ����� e���� e�� 66 ��� ���#� � ���� �� ��� 6� �� '.=&� �� ����� ��$ ���������� ��� �� �##�� ����� �� �� ���� � #��� �#� � � 4 � �� �� g��� ���� � �� �� � ������� ��� ������������� ���� 1��� � �# � �� ���$���� � ����� �#���� ���� �� 7����1��� �� ���$����� �� �� �� b�#���� ���� � g���� �� '.=%� �� 6"� �� ���� 0��� � �� ��� #� � �� ����� ,� 6"� �� �� ���� ��� � 0 ���� �� ,��� �� ��� �� $ ��� �# ������� �� '.=a� ,� 6"� ��� ���1� ����� ��$�� ��� ��� ���$����� ��� ,� ���$ � �� ��#��� � ���� � �� �� 0 ���� ���$ � ���� =& �� �� �� �� '.=a �� *'! �� �� �� �� %&&&� ,� � �� ���� ���� � �� 2���� ���� �� 1����� �� 6"� ���� ��� m�� )������� ���� �� �� ������ ���� �� �� ) 1 b��#��� � ���� �� ��� �� e�� �4h��� b��#��� � ���� �� ,� ��d����� �� �� ���� 8.*�;j9 �� ���� � �� �� )������� ���� �� ,� ����� ���� � �#����������� �� �� ���� ���� � ��� �� � �� � ���� ?.*= ������� �� �� �� �� '.=a �� ?''%�% $������ �� �� �� �� '...� ,� ���$ � �� ���� ���� � ��� ������ �� � �� � $� ����4���� � � �� � '; � ���� ��� �� ����� ��� �� ������ ����� ��� ����#�� �� � �� � ���� ���� ?'* ������� �� '.=a �� � � ?=; $������ �� '...� ,� � ��� ��# �� ���� � 1����� �� 6"� �� �� � $� ���� �� ���1� �� �� g���� 8�� ��� ��9 ��� g���� ���� �� ,��� �� � � ��4������� � ���� � ��� $��� �������� #�� ��� � ��� ��� � ��� � #� ��� � �� 4� #� ������ ������ ,� ���� ��# �� ���� � 1����� �� 6"� ���� �m�� 6"� 6�� ��������� ���� �m$ ��� �# ������ �� -��� '..a� ,� �$d �� � �� ���� ���� � �� �� � ����� �� ��1 �� ��� ��������� �#���� �� �� 6"� ��� �� #�� �� � �����#�� �� ��� � �� ������� � ����� �� ��� � ����� � ���� � �� ��� ��������� ���� ��� ,� � 1 � =' $���� �# ������ �� ,������ $������ ���� � $� �� �� �� '...� ,���� ��� �� �� �� � $���� 1 � ������ ?'**�! $������� ,� ���� �� ����� ��� �� � �� $� ���� �1� � ��� � ��� $���� 1�� *!j 1��� �� ���� �� ��� �� � �� $� #�� �� �� �1� � ��� � ��� )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 *&< $���� 1�� ;.j� l�� � $���� 8��� � �� � #���� ������� ����� � ��#� �� ��� �� ��� �� $����� ��� ��� ��� $����9 ���� �� � ������� 'aj �� ��� ��� ,� �# � ��� ��� 0 ���� #������� �� �� ���� � 1�� ?'*�% $������� ,� �� ��� �� ���� � �� � � � ��� �� ��� ��� $����1���� 1��� �� $���� $ ��� ��d�� #�� ��� �� �� �� ��� �� � ����� �� ,� � #�� ��� � �� ������ ���� � ������� ���� ��� � �� #����� � #����� ��� ��� ��������� $����1����� ,���� ����� 1��� �� ���� � � �� ����� ��������� � #� ��� � �� ��� � �� ������ ���� � ��� � ������� ����� ����!����� �� ������ +� �� �� ����� �� �� ,������ ������ ,������ �� �� ��� �� ����� ����� �� �� ��� �� ���� � ���� ���� �� ��#� �� �� ���� �� �� ����� ���� � ��# � e ��� +��� � �� ���� ��� � ���� � ���������#� 0��� �� ,��� � �� �� ����������� �� �� ���� �� ,� #�� ����� �� ��� �� �� � � ��� �� ���� ����� �� ���� �� ��� �� ��� � ���� �� ����� ��� � #���� � ����� � ������� � �� ,������ ������ 5�"� ������� ��� � ���� ����� �� ������ �������� �� "�� �� ��#���� $��� ���� � �� ��� �� 1 �� � ��# �� �� �� ��� � � ������ �� ��� �� �#���� ��� ����� �� ������ ,����� �� 6"� ��� �� � ���� 1�� � �� a!j ������ �� # ���� �� ���� ������ �� 6"� 1�� � #��� �� � ��� �� ����� �� �� ���� �������� ��� �� 0�� � ��������� �� ���� ���� �� ,��� �� #���� ������ ��� ��� ��� � �� ��� �� ��� ����� � ��� ��#���$� � ���� ��� ������ ����� � �� ����� 3� �� ����� �� ���� #�� � �� �� � ��#��� 1����� ���� �� � #�#���� � ������ ��� �� ������ �� � # ������ #��������� ���� �� # � � � �� � ���� � �� ������� $��� �������� ��� � � � ���������� 3 �� ���� ���� �� 0��� �� 6����$�� ��� ���� ��� � ��� � �� ,��� � 1� � ���� ������ #�� ��� ���� �� ���� �� �� �� ��� �� ��� �� 7���� ���� � �� ��� � �� #��� � �� �� � � ������� ���� �������� ��� � ������� ������� ��� � � � ���������� ���� �� �� ��� �� ����� �� ,��� � �� � ���� � �� ������� ��� #������ �� �� �� ��� �� �� ���4���� � ������� � �� ��� �� �� � �� ,� ������� �� 5�"� ������� ��� � ���� ����� 8��� $��� � ����� � ������� � �� �� � ��� � � �9 �� � ����� #�� �� ��� ���� ,������ ���� ��� g��� �� 5�"� ��� � ���� ��� � � � �� ������ ���$� ��� ����� �� ���$� ��� � � � ���� ��d�� ����� �� ,� � ��� � � � �� � � �� �� �� � ��� �� � �� ������� �� ��������� �� �� ����� ��� �� ���� �������� �� �� ������ 7���� ���� � � ���� ����� �� ���� #�#���� $ ��� �� � �� ���� ,������ ���� �� �#��� � �� �� �� 1��� �� �� � ������ 1� � �� ,������ #�#������� ��1 0 �� %�! ������� # ������ " #��� �� ����� ���!� ,��� ����� ����� �� �� ���� ��� � ���� �##�������� � ��� � ��#��� �� �� ,������ �� ���� ����� ���� �� �� �� ��0 �� ��� �� 1����� �� ���������� �� �� ,������ ������ ,� ����� ������ � �� ������� � �� ��� �� ��� �� � �� �� ���� �� ������ � �� � �� ������ �� �� #���������� *&= )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 ,� �� ��� �� �� � � ����� � � �� #����$� �� ��� ��� �� �� 6"� ��� 0� �� 5��� � "��� � "n� !&& ��� 0� �� 7�������� +3o ��� 0 ���� �� � ���� ���� �� ����� ��� '%4����� ��� � #����� � ������� � �� ,������ ����� 8,h9� 5��� � "��� � ������� 8?9 ��� � ���� ����� 8+�9� +� �� �� ����� �� �� ,������ ������ �� ���� � #����� 1��� �������� � $ ���� � '4� �� # ���� �� ��� � ������ ��� ����� �� ������ ,��� ����� ������ � .a ������ �� ������� ���� � ���� ���� -������ '..' �� + �$ � '..=� +��� ��� � ���� ��� ���#� �� $�� � �� �� ���������� '% ������ �� ���� ���� -������ '... �� + �$ � '...� ������� �$� � ������ ��� � ����� �� ����� ���� � �� �� ����� � #�� ��� 0� ��� � �� ��� � �� '%4����� ,h� ? ��� +� ��� � #������ ��� �� ,h>? ��� ,h>+� 0 ���� ��� � 1 � �$���� � ���� �� b ����� g��� �� ,��� �� ,� 6"� #�� �� � �� ���� �� �� ��� ��� ��� � ���� � ��� 0� +��� �� �� "n� !&& ��� �� 7�������� +3o 1 � �$���� � ���� ��� 6 �� ������ $���� �� 7�� ��� 0 ���� ��� � 1 � �� � �� �� �� �� � ���� �� ��� ��� ��� � �4$�� � �� ��� �� �� � � �� ,������ ����4$�� � � ������ ������� � �� � ����� ��� �������� � ������� 1 � �� ���� � ���� �� ������� � ���� ���� ��� � � �� �� ��� ��� ����� ���� ���� �� $ � ��d��� � ��� 0 ���� ��� � 1� � � ������ 3 ����4� ���� ��#������ 1�� ��� $ �1 � �� ������ �� ��� �� �� � �� b��� ������ � �� � ��� 1 � ���� � ���� � �� ������ �� � ���������# $ �1 � �� �� ��� �� �� � �� ,� ����4��d��� � ��� � 1 � �� � �� �� ����� �� �� � �� ������ �� b������� � #� � �##��� � � 1 � �� � �� � �� ������ � �� � ��� �� � #�������� ��� �� ,��� �4$�� � �� ����� 7���� � ������� ����� #�������� 8�k�9 1�� � �� � ����� �� �����1��� ��� � �������� �� ��� �� #�������� �� ���� 3� �� � �� ������ � �� #��������� 1�� � �� �� 3� 2����� 1 ���� � #�������� 8�pe9� ��#��� � �� ��� � � #����$� ��� ��� 1�� � ��# �� 7������� ����� '% ������ �� ���� ���� -������ '... �� + �$ � '...� 1 ��#�� � �� �k� ��� �pe #���������� $ %�� ��!� ,� � �� � ���� ��� �������� � ������ �� � ����� ��� � � �� �� ������ ���� ����� �� ��� �� ��� ����� � �� ���1� �� ,�$� %�� b� ����� �� ����4� ������� � ��� � #���� ��� �� �1� ��� ��� 2���� ��� 0 � ��� � � �� ���� �� ��� � �� � ���� ������ �� ��� �� ���� ������ ,� '%4����� ����4� ������� � � ���� ��� ����� ���� ������� � �� ��� � �� ��� �� ��� ����� �� 1��� �� ���� �� � ����� ����� �� �������� � ������� ��� �� $ �� "���# ������ ,��� #���� �� ����� � 0���� � ��� � �� #� � ������� �� �� �� � �� ������ �� �� � ���� �� � ��� ������ ��� �� �������� � ���� �� # ���� �� ����� 1�� <<�;'j� ��0 �� �� ���� ��� ��� �� �� ���� ����� ��� ������� ��� � �� � ���� � �� ���� �� ��� �� ��������� l��� �� '%4����� ����4� ������� � � ���� ��� � #����� � �� � ����� 7��� � ���� # ��# �� �� 6"� ��� 0 0��$�� � ��� ���� �1� �� ��������� �� �� � ��� ���� ������ ��� � 8�� +3o9� +��� ��� �� ��� � � ��� �� ������ �� # ���� �� -������ '... �� + �$ � '... �� � #��� � �� ,�$� %$� ,�$� * #�� �� � ����������� �� �� ��� ������ $ �1 � � ����� �� � � �� �� �� ��� �� �� � � ���� �� � ����$� �� � ,������ �� ����� ���� ��� ������ � �� � ��� �� �� ��1 8�� � ���� 9� ���� ����� � ��1 8�� �� �� 9 � ���������# $ �1 � �� �� ��� �� ��� ����� �� )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 *&. ,� ���� �� ��� ������ � ��� �� $ $ �1 � �� +� ��� ? ��� � #������ 6�� � �������� �� ���� �� ��� ������ $ �1 � �� 6"� ��� ��� ��� ����� �� ��� �� �� &�'a!;� ���� ����� �� 6"� ����� �� ���� � $���� � #�������� �� �� ��� ���� ��1 �� �� � ��������� �� ��� �� $ � 1��� �� �� � �� ���� � � ��$ � �$� ��� �� � � �� ��� �� �� � �� � �k� ��� �� � ��#������ �� � �� ������ � 1 � � ��# �� ,��� ����������� �� � #��� � �� ,�$� ;� ��� � �� ������ � #��������� 1 � � � ���� � $� �����1��� �� ������� � #� � #�� ��� � �� ��� �� #�������� �� ��� ��� �� � ��#� �� �� � ������� ����� #��������� ,� ������������ �� �� ����� �� �� #�������� �� �� ����� �� ������ � 1�� � ��#���� �� ��$d � �� ����� ����������� ,� ���������� 1 � ���� � �� � �� � �� #�������� ���� ��� � �# �� � ��� �� � ���� 81��� � ��� ��� � �� � ���� �� � �� � ��# �� ������ �9� ��� �� 1 ������� �� �� �� ��� �� �� � � �� � � #�������� ���� ��� �� �� 1��� �� �� � $ ��� ����1 � �� �� � ���� 1 �������� ,�$� % h���4$�� � ��� � �� � ����� �������� � �������� ��� "���# ������� 6� ��� �� �� � � �� ������ � ���� 8j9 "������� � ������ �� � ����� "���# ����� 8�9 '..&:'..= "n� !&& <*�;* %%�=. &�;' +3o <'�.a %<�;& &�%. 6"� a;�a= !a�=' &�&' ���� !*�;= 'a�<< �&�a* ,h '%4����� =*�%* ;�*< ;�;' ? '%4����� a;�=* %&�'% &�&; +� '%4����� a!�*a '.�.& &�&< 8$9 '... "n� !&& a.�;' ';�.< &�*a +3o aa�&= %a�<& &�&= 6"� %'=�=& .&�;& '�<' ���� !*�=< %%�&< �&�;a ,h '%4����� =!�;. *�=< !�!a ? '%4����� a<�.' ;�!= &�=a +� '%4����� !&�=! *�;a �*�<. � 3�� ��� � ���� � �� # � ���� �� ,�$� * b��� ������ � �� � ��� ��� ����4��d��� � � ������ '..&:'..= "n� !&& +3o 6"� ���� ,h '%4����� ? '%4����� +� '%4����� "n� !&& '�&&&& +3o &�!=!% '�&&&& 6"� &�'a!; &�';<! '�&&&& ���� �&�&.;= &�&'&' �&�&!=* '�&&&& ,h '%4����� &�'&%% &�''!& &�'&*' &�'a.< '�&&&& ? '%4����� �&�&;*< &�'*'! �&�&<<< &�a<&' &�%'*& '�&&&& +� '%4����� �&�&a'. �&�&'%a �&�&%<. &�a%%% &�%!&; &�===. '�&&&& *'& )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 ,� �k� ��� � � ���� �� =%�!'j� � �������� � ������ �� ;�**j� ��� � "���# ����� �� ;�%=� ,� ���� � ���� ��� � �# �$��� .=j �� ���� #��������� ,� ���� ��� � ��� � �� ���� #�������� �� ���� 1��� � 1 ���� �� q*j� 3$� ���� #���� �� �� �� � �� ������ � ����� � #�������� 1��� �� =*j � ���� 1�� �� ����� �� $�� �� ��� �� � � � � � 1 ������� �� �� '%4����� � #����� ,� � 1�� ��� � � ���$� �������� ��� � � ���� �� =;j� ������ �� �� � �� ������ � 2��� ������ ,� ���� �� � �� ������ � ���� ���� ���� � � ���� �� =%�!'j 8�� �k�9 �� �� �## � � ���� �� =*�%*j 8�� #�������� 1�� � �� ��#�� � �� ���� �� ���� � ����9� 3� �� � 1 ���� � � ����� � ����$� � �� 1 ������� �� �� ���� � ���� ���� �� � �� �� ����� $ ��1 ���� �� �� �k� �� �� �� �� �� � �� #������ �� �� ������ �� ,��� �� �� � �� ������ � 1���� ���#�� ���� ���# ��1� �� �� ����� ���� �� �k�� ,�$� ; #��������� 1��� � ����� �� =& ��� <!j �� �� ���� ��1�1��� ���#���� �� �� � �� ������ �� 3� �� � 2��� � � ���� �� �� #�������� �� ��� � ��1 �� �� 1 ������� �� �� ���� � ���� � ��� �� $�� ����� � ����� �� � �1� ����� ��d����� �� �� #��������� e��� �� 0 #���� �� ����� �� ��� � ��� �� #��� � ������ ��� ��� �� ��� #��������� ,�$� ! � � ��$ � �� ���� � ����� � �� �� #�� ����� #�������� 1� � �� �� ���� ���� � �� 2����� 1 ���� 8�pe9 �� � � #�� ����� �� ��� �� �� � �� ,��� #�������� ��� � � ���� �� a=�';j� � �������� � ������ �� � ����� �� '!�=%j� ��� � "���# ����� �� &�%a� 6� ��#������ �� �� �k�� �� �pe #�������� �� � ���� �� �� � ��� ,� � ���� ��� ����� �� �� �pe #�������� ��� �� �� $ ��� ���� �� �� #�������� � #��� � �� ,�$� ;� ���#�� ����� �� �pe 1���� #��� $ ��1 ��� �� �� ����� �� ��� #��������� ���1� �� ,�$� ; ���� 1 � � ��# � $� ���������� �� ����� �� �� #�������� ,�$� ; 7����� � #���������� '..&:'..=� ���������� "n� !&& +3o 6"� ���� ,h '%4����� ? '%4����� +� '%4����� 3����� � ���� 8j9 3����� �������� � ������ "���# ����� =;�&$ : : &�& &�& &�& &�& &�& &�& &�& =*�& ;�*; ;�*= &�& &�& &�& &�<< ..�%* &�& &�& =%�!' ;�** ;�%= &�& &�& &�& %�;% .<�!= &�& &�& =&�& ;�!; *�!* &�& &�& &�;< '&�!a ==�.< &�& &�& <!�& !�.* '�=a &�& &�& '�a. %a�a' <'�<& &�& &�& � 3�� ��� � ���� � �� # � ���� �� $ )� � ���$� ��������� ������� ����� #��������� ,�$� ! �2�����41 ���� � #��������� '..&:'..=� ���������� "n� !&& +3o 6"� ���� ,h '%4����� ? '%4����� +� '%4����� 3����� � ���� 8j9 3����� �������� � ������ "���# ����� a=�'; '!�=% &�%a ';�%. ';�%. ';�%. ';�%. ';�%. ';�%. ';�%. � 3�� ��� � ���� � �� # � ���� �� )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 *'' ,� � �� �� #� �� �� #�1 � �� �� #��������� � ��# � ����� �� �k� ��� �pe �##��� � �� � ���� ��� ���#� 1�� ������ �� 3� ���������� � �� �� ������� ���� 1�� �� � �� � 1�� � �##��� � �� $������� � #�������� 1���� �� $ � ��� �� ����� ��� �� ��� �� ��#� �� �� ���� �� ,��� �� ����� �� � #��� � �� ,�$� � a ��� <� 6� ,�$� a �� ���� ���1� ���� �� �k� �##��� � 1���� �� #���� � � � ���� �� =;�<*j� � �������� � ������ �� *�<*j� ��� � "���# ����� �� !�!a� ,�$� < ���� �� � ���� �� �pe � ��������� 1���� �� #���� � � � ���� �� =<�;.� � �������� � ������ �� '*�.<� ��� � "���# ����� �� '�a=� ,� �pe � ��������� #���� � �� ��� ���� #�������� ������ �� '..':'..= # ����� ��� �� �pe�##��� � ����� #���� � � '4� �� #�������� ��� '... ���� 1�� ��� ���� �� �� �k� �� � ����4��d��� � $����� 3 �������� �� �������� 1�� ���� � � ��#����� �� � ����� ���� �� ���� ��� ���#� � ,� �4� �� � ����� ���� �� ���� �� �� '&j � � �� � 1�� �� �������� ���� ������ ��� ���� � � $ �1 � �� � ����� �� �� � �1� #���������� ,��� �� ��� ���#������ ��� 1 1 � ��� ������ �� ��0���� � ����� ��1 �� �� ,4� �� � ����� ��� 2������ �� ����� � �� �� ���� ��� ���#� ���� �� ���� �� �� 'j � � �� � 1�� � �������� ���� ������ ��� ���� � � $ �1 � �� ����� �� & '������ �� ����������� ,��� ����� �##�� � �� ��� �� #�������� �� ��� �� �� # ������ ��� � ���� ����� � ����� �� �� �� ���� �� �� ,������ ���� ��� ���� ��� ,��� � ��� $ � 0# �� � ��� ���� ��������� $��� ���� � � ���� � #� ������ �� ��� ��� � �� ��� � � ���� �� ���� ���1��� ������ ,� 0��� � �� � ���� �� ���� �� ������� �� ���� ����� 1���� � �� �� �� ��� �� ������� �� �� ���� ����� �� ���� ����� ��$��� ,�$� a �k� #��������� '...� ���������� "n� !&& +3o 6"� ���� ,h '%4����� ? '%4����� +� '%4����� 3����� � ���� 8j9 3����� �������� � ������ "���# ����� =;�<* *�<* !�!a &�& &�& &�& %�;* .<�!< &�& &�& � 3�� ��� � ���� � �� # � ���� �� ,�$� < �pe #��������� '...� ���������� "n� !&& +3o 6"� ���� ,h '%4����� ? '%4����� +� '%4����� 3����� � ���� 8j9 3����� �������� � ������ "���# ���� =<�;. '*�.< '�a= ';�%. ';�%. ';�%. ';�%. ';�%. ';�%. ';�%. � 3�� ��� � ���� � �� # � ���� �� *'% )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 ����� �� ��� �������� ���� �� �� ������ �$� � �� ���� ��� �� ����� �$� � �� ��� �� �� ��� 2���� ���� ��� �� ��� ��� ��� � � � ������� � � ����� �� ��� �� � �� #��������� ,� �������� ���� �� � ���� � ���� ���� ��� # ����� �� ���� ��� � ���� ������ �� $ � �� � � ������ ��� 1 ���� ���� ����� 1�� �� ,������ ����4� ������� � ��� � #����� ,��� ��� ��� ������� ��#�� ������ ��� �� # ������ ���� � �� ,������ �� �� ��� �� ������ ,� ���� �$� � ����� ���� $ ���� �� ������� #�� �� � �������� ��� � �������� � ����� ����� �� �� ������� ���� �� ��� ��� ��� � �4� ������� � � ����� 8�� ���#�� ������� ��� ��� ��� � �9 �� � ��� � � �� $ �� ��� � ����� ��� e��� �� � �� $ � �� � ����� 1� � �$��#� ��� � � � ��������� �� ���� �� ����� ������� �� ����4 � ������� � � ����� �� ���� � ���� �� ���� # ���� � ��� �� ���� �� �� ����4��� �� ����4� ������� � � ����� �� $ � �� ��# ���� ���� ��� ��� ��� ,� � ��� �$� � ����� ���� $ ���� �� ������ �# ����� ��� ��� ��� � ���� � �� ��� �� � #��$� �� ��� �� ��� ����� �� � ��� �� #�������� �� ��� ���� 1���� ,� ��� ��� 2���� ���� � �� � ��� $ ���� �� � ������� ����� #�������� �� �� ��� 0�� � ���������� ,� ����� 8��� ���� ���#������9 �$� � ����� �� ���� �������� �� ���� �� �1� ��� ��� ���� �� ������ ���� $ 0 ��� �� ,��� �� ��� $ ��� �� � �� ������� �� �� ��� �� �� � �� $�� $ ��� �� '%4����� ���� � ���� ��� ��#� � �� ������� � ��� ��� � ��� ����� �� 6� �� � ������� �� 1��� � �� �� �� ��� �� �� #� ����� �#���� ����� 1��� �� ���� ��� � ���� �� � � �� �� '%4����� ���� � ������ � ������ �� � �� �� �� ��� �� � ���1� �� � �� �� �� � � #��� �� ,� ���� � ����� ��� ��1 ���������m� ���� �� �� ��� � � ����$� ���� ��� ��� ��m��� �� � ���� � ����� 2��� ����� �� �� ��� ��� ���� ���� � ���� ������� ��� #��������� g ��� �� ���� �##�� �� ���������� �� �� ���� � �� ,������ ����4� ������� � '%4����� � ���� $ �� � �� ���� ���� ��� ��� � ���� ������ � � � ��d�� 1 ���� ���� ����� ���� � ,������ �� ����� ���������� g�� �� 7�� n h��� ����� � 8'..'9� ���$�� #�������� �#����������� , � �� � �� �� � $���� �� 4#� %=:;*� g��� �� g� 8'.<;9� b����� ��� � # ������� � ��� ��������@ 3� ��� ��������� ��#������� $���� � �� �� �� � 4 � ;=:!%� b���� � ���� n ,������ ��� 8'...9� ��1 �� � �� ��� �� 0#���� �� �� ����� ��� �� � ��� ��� #��������f $���� � �� ������� � ) � 7������ 0"� **:;;� b���� �3�� n h ������ +� � 8'.='9� ,� �� � �� �������� �� ��� � #�� �@ 6�� ��������� �� � � $���� � �� , � ���� 28� %<<:%=.� ��� �� ��� ������ +�� n e������� )� 8%&&&9� ���� � ��� � �� � ���� �� � ����� 2���� ���� ��� $���� � �� ������� � ) � 7������ 08� ;':;.� ���� b�"�� n ��� �� g��� 8'.==9� �0 ���� ��� �� �������� ���1��� ����� ��� ��� ��� ��������� #�������� � � ����� $���� � �� , � ���� 42� '.<:%'!� 7���� ��7� 8'.='9� "�� � � ������ � �� � �� ���� �������� ��� ��� �� ���� � � %����� � -�� �(� .� !;!:!a!� 7���� ��7�� n " �1 ��� ��e� 8'.<<9� 3�� � � ����� ��� ��������� $���� � �� , � �� � %����� � � "� ''!:';a� ���$ �� �� 8'.a=9� 6�� ����������� �� ���� � #��������@ e ���� ����� ��� �#���� ��1�� ���� � � %����� � -�� �(� "#� '%..:'*';� -��� � -�7�� n ���� �� �� �� 8'.<a9� ,� 7��� � �� � ��� ����� ��� ��@ 3� �#��� �� ��������� $���� � �� , � ���� 2.� ;;<:;!=� h �� ��� n "������ �� 8'.<&9� 6�� ��������� �� ���� ����� �� �� ��� �� #���������� ���� � � %����� � -�� �(� 8!� aa=:a.%� )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 *'* �����1���� ���� 8'.!%9� ��������� � � ����� $���� � �� , � ���� � <<:.'� �� ����� �l�� g �������� ��h�� 7������ � �+�� n e������� g��� 8'..a9� ,1 ��� � ��� �� ��� ��������� 2���� �� �����@ "���� � ���� �� ���� � � �����@ "���� � ���� �� ���� � � ����� ��� ��1 � �����f $���� � �� ������� � ) � 7������ 02� .:%%� ���#��� 3�� n "���� 3� 8%&&'9� ��#��� �� �� � �� � #����� �������� #� ���� $ ��� �� ��#���� � ���� ��� � #�� �� ��� �� � %����� � $���� �� 09� '<<:'=!� ����� 7��� 8'..<9� ,� ��#� � �� �������� �� l�� ���1�� ��� ��� � #�� �� , � �� � ���� �� -�� �(� 8� ':'<� "������ g��� 8'.<;9� e�� ��� �� ����� ��� ����������� ���� � ���� ��� ��� ����f , � �� � �� �� � $���� �� 2!� ;=:!;� *'; )*%* � �� �� �* + , � �� � ���� �� -�� �( .! /0!!.1 2!232.4 portfolio diversification in a highly inflationary emerging market introduction review of the literature overview of the turkish economy and equity market macroeconomic conditions turkish financial markets personal asset management in turkey data and methodology findings summary and conclusions references pii: s1057-0810(99)00040-2 racial differences in investor decision making michael s. gutter, jonathan j. fox, catherine p. montalto* the ohio state university, 1787 neil avenue, columbus, oh 43210-1295, usa received 14 may 1999; received in revised form 20 september 1999; accepted 18 october 1999 abstract racial differences in investment behavior are investigated using data from the 1995 survey of consumer finances. socioeconomic, financial, and attitudinal variables are incorporated in a life-cycle savings model. the impact of all variables is allowed to differ between black households and white households to understand racial differences in risky asset ownership. we determine that observed racial differences in risky asset ownership are explained by racial differences in the individual determinants of risky asset ownership, not by race in and of itself. specifically, these differences seem to center on the impact of children and household size. © 2000 elsevier science inc. all rights reserved. jel classification:d12; d91; g11; j70 keywords:individual investing; life cycle models; portfolio allocation; racial differences; risky assets 1. introduction increasing volatility and record setting gains in the stock market have heightened media attention on investment planning and investment decision making. recent articles in the wall street journal (wsj) suggest that affluent blacks are less involved in the stock market than whites (boyce, 1998; mabry, 1999), and may not be benefiting from market gains. the wsj report is based on a study sponsored by ariel mutual funds and charles schwab that does not attempt to explain the reasons for observed racial differences in investing. the fact that * corresponding author. tel.:11-614-292-4571; fax:11-614-688-8133. e-mail address:montalto.2@osu.edu (c.p. montalto) financial services review 8 (1999) 149–162 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00040-2 blacks are less invested in stocks than whites may indicate differences in risk tolerance and investment choice, or a cultural difference in investment behavior. a better understanding of racial differences in investor behavior has several implications for individual investors, financial advisors, and retirement policy makers. first, individuals who participate in defined contribution retirement plans are increasingly choosing their own investment allocations. racial differences in investment choices could lead to large differences between blacks and whites in wealth accumulation over the life cycle, possibly leaving black families less prepared for retirement. additionally, lower wealth accumulation reduces the household’s ability to deal with financial shocks such as disability or death of a key income earner. if there are racial differences in investment choices, then future financial education efforts may need to be tailored to the needs of specific racial groups. financial consultants may also need to be aware of differences in risk tolerance that are race specific. if current risk tolerance measures are inaccurate for members of specific racial groups, this could lead to misunderstanding and poor implementation by planners. lack of understanding between financial planners and black clients may deter blacks from ever seeking the advice of financial advisors. for financial advisors this is a significant target market to miss. based on the 1995 survey of consumer finance, the average black household holds nearly $50, 000 in financial assets. a major concern for policy makers is the proposal for private control within social security accounts. if this comes to pass, investment choice will directly impact the cornerstone of american retirement funds. this study outlines the framework for, and empirically tests, a model of investor decision making for the purpose of explaining observed racial differences in risky asset ownership. socieoeconomic, financial, and attitudinal variables are incorporated into the life-cycle model of saving to explain racial differences in the likelihood of ownership of stocks and/or small business. 2. literature review and theory racial differences in economic decision making have been well studied (jackson and lindley, 1989; myers and chung, 1996), but racial differences in individual investment decisions has received less attention. research on investment decisions consistently finds that risk tolerance is a key factor in investor decision making. gender differences in investing have been analyzed (bajtelsmit and bernasek, 1996; bajtelsmit, bernasek, and jiankoplos, 1999; embrey and fox, 1997) as well as relationships between risk tolerance, demographic characteristics, and investment portfolio allocation (myers and chung, 1996; schooley and worden, 1996; zhong and xiao, 1995). much of this research uses ratios of risky assets to net financial assets, or ratios of risky retirement assets to total retirement assets as dependent variables (yuh and hanna, 1997). in addition to racial differences, other factors affecting investment and economic decisions are critical to the present discussion. myers and chung (1996) examine racial differences in home ownership and equity levels, incorporating risk attitudes, investment horizon, and cognitive skills into models predicting expected home equity and home ownership. myers 150 m.s. gutter / financial services review 8 (1999) 149–162 and chung find that the factors affecting home ownership and home equity levels differ between blacks and whites. in particular, black married females with higher incomes and higher cognitive skills show higher levels of home ownership; age, number of years married, and length of investment horizon positively influence home ownership for white married females. schooley and worden (1996) use the 1989 survey of consumer finances to examine factors associated with taking financial risks. measures of individual risk tolerance are calculated from the actual household portfolio allocation. the level of individual risk tolerance is found to relate to socioeconomic factors, attitudes, and expectations of future income like social security or defined benefit pensions. additionally, schooley and worden find that self-reported relative risk aversion accurately reflects the observed riskiness of the household portfolio allocation. embrey and fox (1997) use the 1995 survey of consumer finances to analyze gender differences in investment choice among single-person households. a tobit model is used to analyze probabilities and levels of ownership of stocks versus cds. they find that single women invest more conservatively than single men but that gender is not the significant determinant of investment choice. instead, differences in investment choice between single men and single women are related to differences in financial and demographic characteristics including net worth, income, and age. yuh and hanna (1997) examine the proportion of risky assets in retirement portfolios using data from the 1992 survey of consumer finances. yuh and hanna find that education, age, and risk tolerance predict the proportion of risky assets in retirement portfolios. they discuss the need for educating workers, and make suggestions for benefits coordinators and financial planners. in a similar study, yuh and devaney (1996) find that characteristics such as employer matching, longer employment, home ownership, and higher income are positively associated with amount invested in defined contribution plans. zhong and xiao (1995) use the 1989 survey of consumer finances to analyze factors associated with families holding stocks or bonds. they find that education, age, and race are significant factors in determining stock ownership. households with white household heads are more likely to have stocks than are otherwise similar households with non-white household heads. they also find that households with more educated household heads are more likely to have stocks than otherwise similar households with less educated heads. investment in financial wealth and human wealth is analyzed by shaw (1996) using the 1983 through 1986 panel of the survey of consumer finances. shaw determines that increased risk aversion reduces investment in human wealth, and that risk tolerance and risky investments increase with education. shaw concludes that this increased return on investments is part of the return to education. bajtelsmit, bernasek, and jiankoplos (1999) analyze gender differences in pension decisions. they include socioeconomic variables in an expected utility model of personal investment decisions. they then estimate an interaction model to better explain gender differences in pension decisions. the measure of relative risk aversion of households comes from the proportion of risky assets in the household portfolio. the authors conclude that observed gender differences in pension contributions are attributable to gender differences in 151m.s. gutter / financial services review 8 (1999) 149–162 behavior; for example, the impact of age, education, and marital status on women’s contributions to pensions differ from the impact of these same variables on men’s contributions. a common approach in the previous research is to include an indicator variable for race in the empirical model. this indicator variable allows the intercept to differ between blacks and whites. a statistically significant estimated coefficient on the indicator variable is interpreted as evidence of a racial difference. however, this specification may be too simplistic and restrictive. the individual determinants of the behavior may also differ by race, suggesting that separate equations should be estimated for each racial group. bajtelsmit, bernasek, and jiankoplos (1999), embrey and fox (1997), and jackson and lindley (1989) take this more flexible approach, and we use similar methods in this study. the purpose of this study is to incorporate socioeconomic, financial, and attitudinal variables into the investment decision model. the model allows the impact of these variables to differ between black households and white households to explain racial differences in the likelihood of ownership of stocks and/or small businesses. economic theory provides the framework for the investment decision model. the basic model builds on the notion of lifetime utility maximization from wealth accumulation through periods of borrowing and saving. the allocation of lifetime resources between current and future consumption was first formally introduced in ando and modigliani’s (1963) life cycle hypothesis of savings. they state that the aggregate consumption/saving choice is a function of perception of permanent income, demographics, and interest rates on borrowing and savings. the basic model assumes no bequest motive. expanding on this, deaton (1992) outlines the following two period budget constraint for the representative consumer: at11 5 (1 1 rt11) (at 1 yt-ct) (1) where: at11 5 assets at time t1 1 rt11 5 the real rate of return at 5 assets in period t yt 5 income in period t ct 5 consumption in period t. this function incorporates both current income and factors that may affect the perception of future income flows. deaton describes current utility as the present and future discounted values of the felicity or happiness associated with consumption. the value function equation is given by: vt(at) 5 max n[yt(yt 1 at (ni) 1 etvt11(1 1 rit11((ni))] (2) where: vt(at) 5 the value function for current assets in period t 152 m.s. gutter / financial services review 8 (1999) 149–162 n 5 the number of assets vt 5 the marginal rate of felicity in period t yt 5 income in period t at 5 assets in period t ni 5 real expenditures on assets etvt11(11rit11(sni)) 5 the expected future value of assets. maximizing equation 2 subject to equation 1 yields equation 3 yt(ct)5et[(11rit11)vt11(at11)] (3) according to eq. (3), the marginal felicity of consumption today is a function of the expected value of money at some nonzero interest rate. in other words, the marginal felicity or utility of consumption will be affected by the rate of return on savings. thus, a higher return on savings yields higher possible consumption and hence higher possible utility. the underlying economic theory relates investment choice to various characteristics. in a world of real nonzero interest rates, there is an opportunity cost of not saving. the opportunity cost is the rate of return that could be earned on savings. the higher the rate of return, the more overall consumption can be increased through saving and investing that also increases total lifetime utility. because stocks have the highest real rate of return in the long run (siegel, 1994), it is rational for investors to hold stocks to maximize lifetime returns. in addition, ownership of a business has a high potential return if successful. 3. methodology this study uses data from the 1995 survey of consumer finances. the survey, sponsored by the federal reserve board with cooperation from the department of the treasury, provides detailed information about household asset holdings and debt. the sample for this study is 3,939 households, consisting of households with a black or white household head and where the gender of the household head is reported. the survey of consumer finances consists of five data sets due to multiple imputation of missing responses (kennickell, 1997). montalto and sung (1996) describe the repeated imputation inference technique for analyzing multiply imputed data. repeated imputation inference (rii) procedures use data from all five implicates and incorporate estimates of error due to missing data. rii techniques are used for estimation of the descriptive statistics and the logistic regressions in this study (montalto and yuh, 1998). the scf data set contains a weight variable that can be used to generate estimates representative of the us population (kennickell, mcmanus, and woodburn, 1996). the descriptive statistics reported in this analysis are weighted. the logistic regressions are not weighted. kennickell and mcmanus (1993) and montalto (1998) discuss the drawbacks of using the weight variable in a 153m.s. gutter / financial services review 8 (1999) 149–162 multivariate analysis, as the weights are constructed from variables such as income that are commonly used by researchers in multivariate analyses. a logistic regression is used to determine log likelihood estimates of the probability that a household holds stocks and/or business assets in their portfolio. if we assume that they hold risky assets by choice, then this indicator variable is the likelihood that they choose to hold these assets. the dependent variable in the logistic regression is an indicator variable equal to one if the household owns any stocks or business assets, and zero otherwise. this dependent variable is appropriate for our needs since we are interested in the choice to be invested in risky assets, not the amount invested in risky assets. the explanatory variables used in the model are based on socioeconomic, financial, and attitudinal characteristics (table 1) that are highlighted in the literature as being relevant to risk aversion and risky asset ownership. in addition, indicator variables for receiving or expecting to receive an inheritance in the future are included to control for windfalls. although we do not control for table 1 description of explanatory variables variables description socioeconomic variables age age of reference person age-squared squared value of age of reference person household type (reference category: married couple) single male 51 if householder is unmarried male, 0 otherwise single female 51 if householder is unmarried female, 0 otherwise children present 51 if children are present in the household, 0 otherwise household size number of persons in the household education (reference category: college graduate) less than high school 51 if years of education, 12, 0 otherwise high school 51 if years of education5 12, 0 otherwise some college 51 if years of education. 12 and, 16, 0 otherwise financial variables net worth ($1000) value of total assets minus total liabilities, measured in thousands annual household income (reference category: middle two quartiles) lowest quartile 51 if income# $20,280, 0 otherwise highest quartile 51 if income. $113,000, 0 otherwise received inheritance 51 if household has received an inheritance, 0 otherwise expects future inheritance 51 if household expects future inheritance, 0 otherwise attitudinal variables willingness to take risk (reference category: willing to take average risk) not willing 51 if not willing to take any risk, 0 otherwise above average risk 51 if willing to take above average risk, 0 otherwise substantial risk 51 if willing to take substantial risk, 0 otherwise positive economic outlook 51 if household expects economy to improve over next 5 years, 0 otherwise negative economic outlook 51 if household expects economy to do worse over next 5 years, 0 otherwise felt discouraged with credit51 if household reported previously feeling discouraged when applying for credit, 0 otherwise investment horizon 51 if investment horizon of less than a year 52 if investment horizon of 1–2 years 53 if investment horizon of 2–5 years 54 if investment horizon of 5–10 years 55 if investment horizon of more than 10 years 154 m.s. gutter / financial services review 8 (1999) 149–162 gifts from living family members, the inheritance variables are included to determine whether or not assets are held more by choice than circumstance. we are interested in analyzing racial differences in the likelihood of owning risky assets. jackson and lindley (1989) outline a procedure for testing for statistical differences between two groups by estimating a full interaction model, and decomposing any statistical difference to more accurately understand the nature of the between-group differences. to estimate the interaction model, data from the two groups are pooled. the dependent variable is regressed on an intercept term, the set of independent variables, an indicator variable for race, and a set of interaction variables created by multiplying each independent variable by the race variable. to carry out the decomposition two additional models are estimated—a reduced model that omits the indicator variable and the set of interaction variables, and an intermediate model that includes the indicator variable but omits the interaction variables. the interaction model and the reduced model are compared to determine the joint significance of the indicator variable for race and the set of interaction variables. (in ordinary least squares regression the statistical test is an f-test; a comparable test for maximum likelihood estimation is the likelihood ratio test.) if this joint test is significant, then jackson and lindley (1989) outline a procedure for decomposing the total between-group difference into three components: the constant effect, the endowment effect, and the coefficient or response effect. the endowment effect is the portion of the total difference accounted for by differences between the two groups in the level of the explanatory variables. the coefficient or response effect is a measure of the difference between the two groups in the response of the dependent variable to changes in the independent variables. the interaction model and the intermediate model are compared to determine if there is a significant coefficient effect. the constant effect is the portion of the total difference that cannot be accounted for by differential endowments or differential responses. the estimated parameters on the race indicator variable in the interaction model are used to determine if there is a significant constant effect. in our application, a significant coefficient effect provides evidence of racial differences in the impact of the explanatory variables on risky asset ownership. a significant constant effect provides evidence of racial differences in risky asset ownership beyond those factors controlled for in the model. two hypotheses are tested. the first hypothesis is that the effects of socioeconomic, financial, and attitudinal variables on risky asset ownership differ between black households and white households. a likelihood ratio test is used to test the joint statistical significance of the race indicator variable and the set of interaction variables (comparison of the interaction model and the reduced model). additionally, the significance ofx2 statistics on individual interaction variables is used to identify the specific coefficients that differ between black households and white households. the second hypothesis is that differences in risky asset ownership between black households and white households are due to racial differences in the individual determinants of risky asset ownership (coefficient effect), not to race in and of itself (constant effect). a likelihood ratio test is used to test for the statistical significance of the set of interaction terms (comparison of the interaction model and the 155m.s. gutter / financial services review 8 (1999) 149–162 intermediate model). thex2 statistic is used to test the statistical significance of the race indicator variable in the interaction model. 4. results descriptive statistics for black households and white households are provided in table 2, along with the results of tests for statistical differences between the two groups. the majority of black households are headed by a single female (51%) whereas the majority of white households are headed by a married couple (55%). however, 53% of black households contain children compared to 42% of white households, and 10% of black households contain other adult relatives compared to less than 4% of white households. on average, black households are larger than white households (2.64 and 2.49, respectively). the distribution of asset ownership differs between black households and white households. only 23% of black households hold risky assets compared to 46% of white households. there are also significant differences between black households and white households in the mean value of these assets. black households hold significantly less of each asset type on average. the net worth measure also shows a significant difference between the two racial groups with white households having an average net worth of $246, 287, compared to only $45, 854 for black households. black households report a lower willingness to take financial risks and have a shorter investment horizon compared to white households. a significantly higher proportion of black households (60%) than white households (42%) report they are not willing to take any risk. similar proportions of black and white households are willing to take substantial financial risk. the average investment horizon is approximately three and a half years for black households, compared to over five years for white households. we examine whether the risk tolerance question posed in the scf is perceived similarly by black and white households. to do this, responses to the risk tolerance question are collapsed into an indicator for whether or not the household reports a willingness to take investment risks in expectation of higher returns. this indicator of willingness to take risk is then compared to an objective measure of risk tolerance—whether or not the household owns risky assets. thex2 test for independence reveals that households who report a willingness to take risks are more likely to own risky assets (table 3, panel a). of households reporting that they are willing to take risk, 58% own risky assets, compared to only 24% of households not willing to take risk. we also create an indicator variable for whether households exhibit consistent behavior— indicating they are willing to take risk and actually owning risky assets, or indicating they are not willing to take risk and actually not owning risky assets. this consistency variable is compared with the race variable. the results of thex2 test for independence indicate that white households are more likely than black households to exhibit consistent behavior— being willing (unwilling) to take risk and actually owning (not owning) risky assets (table 3, panel b). these results may suggest that black households differ from white households in the ability to own risky assets given the willingness to take risk. however, this conclusion should be interpreted cautiously since other factors are not controlled. 156 m.s. gutter / financial services review 8 (1999) 149–162 table 2 characteristics of black households and white householdsa variable black households white households test for difference between samples socioeconomic variables age 46.2 49.5 t 5 14.152* household type single male 19.00% 18.88% x2 5 8.077* single female 51.16% 25.80% x2 5 968.277* married 29.90% 55.19% x2 5 805.659* household size 2.64 2.49 t 5 77.835* others living with respondent spouse/partner 36.69% 60.93% x2 5 778.619* children 53.23% 42.00% x2 5 175.724* adult relatives 9.95% 3.70% x2 5 222.252* other adults 2.01% 4.00% x2 5 6.045* education less than high school 28.60% 17.57% x2 5 309.573* high school 35.35% 29.94% x2 5 90.347* some college 23.10% 23.38% x2 5 8.311* college degree 12.98% 29.11% x2 5 532.301* financial variables net worth $45,854 $246,287 t 5 8.37* annual household income $23,826 $52,231 t 5 7.173* own risky assets 22.80% 46.14% x2 5 940.50* net financial assets $49,511 $185,643 t 5 7.809* stocks $377 $17,639 t 5 2.686* proportion of stocks in pensions 12.34% 16.93% x2 5 23.245* business assets $3,355 $53,378 t 5 3.65* stock funds $153 $7,443 t 5 4.452* received inheritance 10.98% 24.05% x2 5 321.16* expects future inheritance 4.34% 15.41% x2 5 191.169* attitudinal variables willingness to take risk not willing 59.72% 42.19% x2 5 511.132* average risk 29.59% 39.72% x2 5 118.695* above average risk 7.03% 14.73% x2 5 156.905* substantial risk 3.65% 3.36% x2 5 2.729 expect united states economy to improve 28.99% 25.22% x2 5 9.180* stay the same 38.54% 51.67% x2 5 136.30* worsen 33.47% 23.11% x2 5 118.714* discouraged when applying for credit 36.10% 12.90% x2 5 1145.795* investment horizon 3.45 5.17 t 5 118.93* reasons for saving financial goals 9.06% 9.07% x2 5 3.023 emergency fund 40.41% 33.51% x2 5 62.252* charity 0.00% 0.10% x2 5 4.274* child’s education 9.28% 8.49% x2 5 4.991 luxuries 1.84% 3.13% x2 5 2.241 durable goods 11.72% 9.31% x2 5 48.399* own education 5.24% 4.67% x2 5 3.011 a source:1995 survey of consumer finances; data from all five implicates; estimates derived using repeatedimputation inference techniques. *p # 0.05. 157m.s. gutter / financial services review 8 (1999) 149–162 for the regression analysis, household income is categorized into three groups defined by the lowest quartile, middle two quartiles, and upper quartile of the sample income distribution. the majority of black households are in the lowest income quartile (57%) compared to less than one third of white households (32%). within each income category, mean income is lower for black households compared to white households (table 3, panel c). the logistic regression results for the interaction model (table 4) indicate that there are racial differences in the determinants of stock and small business ownership. this supports our first hypothesis that the effects of socioeconomic, financial, and attitudinal variables on risky asset ownership differ between black households and white households. the joint test of the race indicator variable and the set of interaction terms is statistically significant (p , 0.001). the decomposition of the between-group difference indicates an insignificant constant effect (p 5 0.956), and a statistically significant coefficient effect (p , 0.001). this supports our second hypothesis that differences in risky asset ownership between black households and white households are due to racial differences in the individual determinants of risky asset ownership, not to race in and of itself. in the intermediate model, the race indicator variable is significant and negative, indicating that black households are less likely than white households to hold stocks and small business investments. however, in the interaction model, race is no longer statistically significant. racial differences in the determinants of stock or small business ownership are due to differences between black households and white households in the impact of explanatory variables on risky asset ownership. specifically, thex2 statistics on the interaction variables for presence of children and household size are statistically significant, indicating that table 3 chi-square tests for independence (column percents reported) panel a: objective and self-reported risk tolerance willing to take risk not willing to take risk total sample test statistic own risky assets 58.0% 23.8% 42.7% 2319.84 do not own risky assets 42.0% 76.2% 57.3% (p 5 0.000) panel b: consistency of risk measures and race black white total sample test statistic consistent 64.3% 66.5% 66.2% 4.84 inconsistent 35.7% 33.5% 33.8% (p 5 0.028) panel c: income distribution and race mean income test statistic black white black white lowest 25% 57.1% 32.3% $8,630 $10,808 988.36 (p 5 0.000) middle 50% 41.9% 61.3% $41,238 $48,262 43.75 (p 5 0.000) highest 25% 1.1% 6.4% $154,998 $293,668 567.56 (p 5 0.000) 158 m.s. gutter / financial services review 8 (1999) 149–162 table 4 logistic analysis of likelihood of investing in risky assetsa variable reduced model intermediate model interaction model intercept 20.0539 20.0842 20.2174 socioeconomic variables age 0.0489** 0.0513*** 0.0506** age-squared 20.0005*** 20.0005*** 20.0005*** household type (reference category: married couple) single male 20.4152*** 20.4028** 20.3611** single female 20.4934*** 20.4442*** 20.4333** children present 0.0633 0.0684 20.0423 household size 20.0516 20.0481 0.0036 education (reference category: college graduate) less than high school 21.1675*** 21.1515*** 21.0561*** high school 20.5532*** 20.5434*** 20.4947*** some college 20.4690*** 20.4616*** 20.4145*** financial variables net worth ($1000) 0.0024*** 0.0024*** 0.0024*** income (reference category: middle two quartiles) lowest quartile 20.9783*** 20.9643*** 20.9282*** highest quartile 0.9673*** 0.9635*** 0.9781*** received inheritance 0.4061*** 0.3900*** 0.4031*** expect to receive inheritance 0.1476 0.1164 0.1310 attitudinal variables willingness to take risk (reference category: willing to take average risk) not willing 20.8074*** 20.7982*** 20.8275*** above average risk 0.3371** 0.3242** 0.3761** substantial risk 0.4046 0.4058 0.4698 positive economic outlook 20.0912 20.0726 20.1350 negative economic outlook 20.1900 20.1775 20.1856 felt discouraged with credit 20.2974* 20.2376 20.2842 investment horizon 0.0217* 0.0201* 0.0209* indicator for black household — 20.4723** 20.0973 interaction variables age — — 20.0471 age-squared — — 20.0006 household type (reference category: married couple) single male — — 20.2438 single female — — 20.2477 children present — — 1.0827* household size — — 20.4227* education (reference category: college graduate) less than high school — — 21.0574 high school — — 20.5167 some college — — 20.5850 financial variables net worth ($1000) — — 20.0003 income (reference category: middle two quartiles) lowest quartile — — 20.2863 highest quartile — — 20.7225 received inheritance — — 20.2140 expect to receive inheritance — — 20.3503 159m.s. gutter / financial services review 8 (1999) 149–162 changes in presence of children in the household and in household size affect risky asset ownership of black households differently than white households (table 4). the presence of children in the household increases the likelihood of risky asset ownership by black households, and increases in household size decrease the likelihood. these two variables do not have statistically significant effects on risky asset ownership of white households. a few determinants of risky asset ownership are common between black households and white households. households with a householder who is a college graduate are more likely to own risky assets compared to otherwise similar households with a less educated householder. households with income in the lowest income quartile are less likely than households with income in the middle two income quartiles to own risky assets. ownership of risky assets is less likely among households indicating that they are not willing to take financial risk compared to households willing to take average financial risk. several other determinants of risky asset ownership are statistically significant for white households but not for black households, although the tests for racial differences in the effects of these variables are not statistically significant. age has a nonlinear effect on the likelihood of risky asset ownership for white households, even when income and net worth are controlled. the likelihood of risky asset ownership increases with age until a maximum, after which the likelihood decreases. white households headed by an unmarried householder are less likely to hold risky assets than otherwise similar households headed by a married couple. net worth and income are positively associated with the likelihood of risky asset ownership for white households. white households with income in the highest income quartile are more likely to own risky assets than otherwise similar households with income in the middle two income quartiles. having received an inheritance increases the likelihood of owning risky assets, controlling for income and net worth. an inheritance may include risky assets in the portfolio resulting in risky asset ownership “by chance” rather than by investment choice. white households that are willing to take above average financial risk or substantial financial risk are more likely to own risky assets than otherwise similar households that are table 4(continued) variable reduced model intermediate model interaction model attitudinal variables willingness to take risk (reference category: willing to take average risk) not willing — — 0.2520 above average risk — — 20.8099 substantial risk — — 20.3540 positive economic outlook — — 0.5663 negative economic outlook — — 0.0543 felt discouraged with credit — — 0.0958 investment horizon — — 20.0033 22 log likelihooda 3642.96*** 3632.23** 3608.84*** a source: 1995 survey of consumer finances; data from all five implicates; estimates derived using repeatedimputation inference techniques b statistic reported is the average of the five22 log likelihood statistics from analysis of the separate implicates. * p-value# 0.05; **p-value# .01; *** p-value# 0.001 160 m.s. gutter / financial services review 8 (1999) 149–162 willing to take only average risk. the investment horizon is positively associated with the likelihood of risky asset ownership for white households. white households that expect the economy to do worse over the next five years, and households that report previously feeling discouraged when applying for credit are less likely to own risky assets. 5. conclusions racial differences in investment decisions related to ownership of risky assets seem to center on the impact of children and household size. the presence of children in the household increases the likelihood of stock or business holdings for black households. household size is negatively related to risky asset ownership for black households. these two variables do not have statistically significant effects for white households. although the household size variable may indicate presence of children, it is also attributable to the presence in black households of other adult relatives who may not be able to contribute to household financial resources, thus reducing the likelihood of risky asset investment. if black households and white households perceive financial risks differently, then racial differences in investment decision making might be due to racial differences in the perception of financial risk. however, we do not find evidence of a racial difference in the impact of reported willingness to take risk through risky asset ownership; none of the interaction terms on the risk variables are significant. however, hanna and gutter (1998) caution that the scf measure of risk tolerance does not strictly fit the economic concept of relative risk aversion. a more theoretically guided measure of risk tolerance may be needed to further explore racial differences in risk tolerance. there are several limitations to this study that are inherent to the data set used. several variables that we would like to include in the multivariate analysis are not available in the survey of consumer finances public use data set. we would like to control for living in a metropolitan versus a non-metropolitan area, because living in a metropolitan area provides a greater pool of community resources and a greater prevalence of financial institutions. a metropolitan indicator variable could serve as a proxy for the amount, but not the actual use, of financial information. we also do not have information on whether investment advice was received from family, peers, or the media. the findings of this study are generally consistent with previous studies that blacks and whites have different investment behaviors (boyce, 1998; mabry, 1999; myers and chung, 1996; zhong and xiao, 1995). our empirical model allows the effect of socioeconomic, financial, and attitudinal variables on risky asset ownership to differ between black households and white households. with this empirical specification we are able to conclude that racial differences in risky asset ownership are due to racial differences in the individual determinants of risky asset ownership, not to race in and of itself. this enhanced understanding of racial differences in investor decision making can help financial practitioners more accurately understand and better serve a racially diverse clientele. improved service should lead to more use of financial advisors by clientele, resulting in improved portfolio allocation. in the long run, strategic portfolio allocation should yield higher net worth, higher lifetime consumption, and higher lifetime utility. 161m.s. gutter / financial services review 8 (1999) 149–162 references ando, a. & modigliani, f. (1963). the life-cycle hypothesis of saving.american economic review, 53(1), 55–74. bajtelsmit, v. & bernasek, a. (1996). why do women invest differently than men?financial counseling and planning, 7,1–9. bajtelsmit, v., bernasek, a. & jiankoplos, n. (1999). gender differences in defined contribution pension decisions.financial services review, 8(1), 1–10. boyce, j. (1998). blacks are bearish on wall street says new survey. the wall street journal, 231(69), c21. deaton, a. (1992).understanding consumption.oxford: clarendon press. embrey, l. & fox, j. (1997). gender differences in the investment decision making process.financial counseling and planning, 8(2), 33–39. hanna, s. & gutter, m. (1998). a theory based measure of risk tolerance.proceedings of the academy of financial services,10–11. jackson, j. & lindley, j. (1989). measuring the extent of wage discrimination: a statistical test and a caveat. applied economics, 21,515–540. kennickell, a. b. (1997).codebook for 1995 survey of consumer finances. washington d.c.: federal reserve system. kennickell, a. b. & mcmanus, d. (1993). sampling for household financial characteristics using frame information on past income.proceedings of the section on survey research methods, 1993 joint statistical meetings of the american statistical association, atlanta, ga. kennickell, a. b., mcmanus, d. & woodburn, r.l. (1996). weighting design for the 1992 survey of consumer finances. working paper. board of governors of the federal reserve system. mabry, t. (1999). black investors stay away from stocks.the wall street journal, 233(94), a2. montalto, c.p. (1998). practical issues and challenges from a users’ perspective.consumer interests annual, 44, 232–233. montalto, c. p. & sung, j. (1996). multiple imputation in the 1992 survey of consumer finances.financial counseling and planning, 7,133–146. montalto, c. p. & yuh, y. (1998). estimating nonlinear models with multiply imputed data.financial counseling and planning, 9(1), 97–101. myers, s. & chung, c. (1996). racial differences in home ownership and home equity among preretirement-aged households.the gerontologist, 36(3), 350–360. schooley, d. k. & worden, d. d. (1996). risk aversion measures: comparing attitudes and asset allocation. financial services review, 5(2), 87–99. shaw, k. (1996). an empirical analysis of risk aversion and income growth.journal of labor economics, 14(4), 626–653. siegel, j. j. (1994).stocks for the long run: a guide to selecting markets for long-term growth. burr ridge, il: irwin professional. yuh, y. & devaney, s. (1996). determinants of couple’s defined contribution retirement funds.financial counseling and planning, 7,31–38. yuh, y. & hanna, s. (1997). the demand for risky assets in retirement portfolios.presented at the annual meeting of the academy of financial services. zhong, l. & xiao, j. (1995). determinants of family bond and stock holdings.financial counseling and planning, 6,107–114. 162 m.s. gutter / financial services review 8 (1999) 149–162 pii: s1057-0810(99)00010-4 credit union industry structure: an examination of potential risks robert j. boldina,*, keith leggettb, and robert strandc adepartment of finance and legal studies, indiana university of pennsylvania, 322a eberly college of business, indiana, pa 15705, usa badjunct faculty member at johns hopkins university, washington dc campus, washington, dc usa, george mason university, fairfax, va, usa cadjunct faculty member at johns hopkins university, washington dc campus, washington, dc, usa abstract the credit union industry has undergone significant changes over the past decade. with more than 11,500 credit unions now serving 74 million members, it continues to attract shareholders/ depositors because of its generally lower cost services and higher returns on savings. the credit union membership access act of august 1998 should help membership continue to grow. at the same time, however, shareholders/depositors should be aware of the unique structure of the credit union industry. this paper examines the inherent risk as a result of the interrelationships among its components. © 1998 elsevier science inc. all rights reserved. 1. introduction on august 7, 1998, the credit union membership access act was signed into law. its passage culminated a 20-month campaign to guarantee the right of consumers to select a financial institution. the more than 11,500 credit unions which serve 74 million consumers, have a decided advantage over commercial banks in that the credit unions enjoy a tax exempt status. moreover, in 1982, federal regulators permitted credit unions to expand from a common membership of a single “like” organization to that of multiple membership groups. because of this recent change, it is important for depositors to understand more fully the structure of the credit union industry. while admittedly small in comparison to the banking * corresponding author. tel.:11-724-357-2465; fax:11-724-357-7520. e-mail address:rboldin@grove.iup.edu (r.j. boldin) financial services review 7 (1998) 207–215 1057-0810/98/$ – see front matter © 1998 elsevier science inc. all rights reserved. pii: s1057-0810(99)00010-4 industry ($361 billion credit union industry versus $6 trillion in assets held by banks), the credit union industry has become more aggressive in its offerings. in addition, the structure of the credit union industry is unique compared to that of the banking industry. “natural person credit unions” (npcus), are tax exempt financial cooperatives, whereby the provider of inputs (saver) is also the user of outputs (borrower). these financial cooperatives that serve depositors/shareholders are, in turn, served by forty corporate credit unions. the “corporates” themselves are also nonprofit, cooperative institutions which are owned by, take deposits from, and provide liquidity, investment and payment services for their member npcus. the corporates, in turn, are tied into the u.s. central credit union, which is itself a corporate but also serves as a liquidity facility for other corporates. the stability of the corporates should be important to anyone who is a member and/or depositor of an npcu. as depositors/shareholders, npcu members need to be aware of the interdependence between npcus and the corporates, and consequently of the soundness of the corporates. thus, the purpose of this research is to examine the level of exposure that corporate credit unions pose due to the structure of the credit union industry, and to examine the implications for corporate credit union capital standards. section 2 provides additional background material and a review of the literature. section 3 presents data regarding corporate credit union capital. section 4 reviews corporate credit union investment practices. section 5 describes the potential exposure. the last section discusses the conclusions. 2. background and literature review the credit union industry is hierarchically structured. at the base of the system is approximately 11,500 npcus. the npcus comprise the membership of the middle tier, that is, forty corporate credit unions which are geographically dispersed throughout the united states. the corporate credit unions define the membership of u.s. central credit union, which sits atop the structure. npcus invest their surplus loanable funds in their regional corporate credit union as either deposits or membership capital share accounts. the regional corporates then invest surplus funds in u.s. central credit union, which invests in liquid financial instruments. each group is heavily invested in the level above it. at the end of 1996, corporates had half of their total investments or $13.6 billion invested in u.s. central. in turn, the npcus had $23.2 billion invested in the corporates. the typical npcu had 42.5 percent of its investment portfolio in corporates, amounting to 75 percent of its capital. the average investment percentage is higher for smaller institutions (national credit union administration, 1996; u.s. department of the treasury, 1997). this type of structure, where one group is heavily invested in the level above it, involves a significant potential exposure for the npcus. as noted by carnell (1997), both capital and liquidity are exposed; if a corporate suffers losses on its investment portfolio, and if these losses exceed the amount of its capital, then member institutions must write-down their investments in the corporate. and, if a corporate’s capital is impaired by investment losses, then it will have more difficulty raising funds to satisfy members’ liquidity needs. in early 1995, events raised concerns about the health of the corporates. when short-term 208 r.j. boldin et al. / financial services review 7 (1998) 207–215 interest rates rose significantly throughout 1994 and early 1995, almost all of the corporates suffered investment losses. the u.s. general accounting office (1995a) estimated that corporates, in aggregate, had $600 million in unrealized investment losses or 62% of the aggregate capital base. according to tolen information services, five corporates became market value insolvent. two failed as a result of these losses: federation de coop credit union, puerto rico, september 1994 and capital corporate federal credit union, lanham, md, january 1995. the corporate credit union network is subjected to increasing levels of competitive stress. in recent years, according to black et al. (1994), a growing number of npcus and even regional corporates have elected to invest directly in the market instead of in their corporate credit unions. this disintermediation particularly affected the largest corporate, u.s. central credit union, which has shrunk at an eight percent annual rate since 1991. the u.s. general accounting office (1994) reported that the u.s. central credit union suffered a 28% run-off of assets between year-end 1993 and june 30, 1994. recent mergers among several corporates came about in reaction to these competitive pressures. it is likely this pattern will continue to follow that of the banking sector as outlined by winton (1997) and houston and ryngaert (1994). as the credit union industry matures, it is expected that more consolidation of their activities for synergistic and cost saving purposes will occur, subject to legal restrictions. due to the structure of the credit union industry, should any instability arise among the corporates, it could translate into more widespread problems for the industry. this risk arises from the fact that a significant amount of npcu funds are held by the corporates. nine percent of the average npcu’s assets are invested in a corporate credit union. these funds are at risk if the corporate suffers a loss. in fact, as outlined by the u.s. general accounting office (1995b), member npcus lost $37 million with the failure of capital corporate federal credit union (capcorp) even after the national credit union share insurance fund (ncusif) guaranteed almost $700 million of npcus’ uninsured deposits in capcorp. previous studies such as those by fried et al. (1993) and stiroh (1994), which focused on u.s. credit unions and thrifts, have not addressed this problem. since 1980, 130 countries around the world have experienced significant banking problems, according to lindgren et al. (1996). in the u.s., the saving and loan crisis cost taxpayers an estimated $150 billion (barth et al., 1995). kane and hendershott (1996) found that the ncusif became economically insolvent in the 1980s. the ncusif, however, was able to rebuild its reserves without losses of taxpayer funds. the ncusif’s problems were limited by cross-industry guarantees and the industry’s mutual structure. credit unions must deposit 1% of their insured deposits with the ncusif and must adjust this deposit as their insured deposit base changes. according to the u.s. general accounting office (1991), if the reserve ratio (equity to insured deposits) for the ncusif falls below 1%, credit unions are required to expense the difference between the reserve ratio and the one percent deposit. also, they must replenish their deposit back to its one percent level. since this deposit is at risk if other credit unions take significant risk, it is imperative that all credit unions monitor this risk. at the same time, however, the mutual structure of credit unions reduces the incentive for management to take on greater risk. thus, agency questions arise where management is more conservative than its owners. while this topic is beyond the 209r.j. boldin et al. / financial services review 7 (1998) 207–215 scope of this paper, it should be noted that management’s ability to benefit from risk taking is constrained by the ability of a credit union to convert from its mutual structure to stock organization. the national credit union administration (ncua) rules require 50% of all eligible members to approve a conversion of a credit union to another charter type. further, it limits the ability of management and directors to benefit from the conversion. theoretical literature has examined several models of npcu behavior, but has devoted very little attention to the corporates. four competing objectives functions have been assigned to credit union behavior as follows: pearce (1984) posits growth maximization; fried and lovell (1993) assume maximization of services with respect to a resource constraint; murray and white (1980) suppose efficiency maximization; and, smith (1984) posits securing net gains to savers and borrowers through better than market rates. research by kane and hendershott (1996), however, suggest that the current weak link in the credit union system is the corporate credit union network. they state that low levels of capitalization at corporate credit unions creates a moral hazard, creating incentives for management to take on greater risk. this paper builds upon their research by examining the interrelationships between npcus and the corporates, instead of focusing solely on the behavior of natural person credit unions or corporate credit unions. 3. corporate credit union capital data until recently, corporates have been thinly capitalized. in fact, at year-end 1995, the average leverage capital ratio for the corporate system was only 2.2%. nonetheless, the u.s. general accounting office (1994) reported that all of the corporates met a regulatory risk-based capital standard. in reaction to the problems in 1995, the national credit union administration (the federal regulatory agency) instituted two important changes to strengthen the capital of the corporates to be effective in 1998. first, a risk-based capital standard was replaced (and abolished) with a leverage standard; each corporate must now meet a 4% capital-to-assets test. second, “membership capital share deposits” (mcsd) now count as primary capital toward the leverage standard, provided that there is a contractual requirement of three-years notice prior to withdrawal. the new standard accentuated the interdependence within the pyramid structure. npcu investments in a corporate still face risk of loss should the corporate face significant losses in its investment portfolio. any losses would first be written off against “core capital,” the equity capital not including mcsd. if a corporate’s trouble is severe enough to exhaust its core capital, the mcsd would have to be written down. conceivably, greater losses could even force a write-down of npcu deposits in the corporate. of course, any write-down of deposits or investments in a corporate would be reflected by a drop in the investing npcu’s capital. while deposits are protected up to $100,000, for the majority of the corporates with federal deposit insurance, this amount is small compared to the level of investment of most npcus. corporate core capital, then, is the essential protection for npcu investments in corporates. 210 r.j. boldin et al. / financial services review 7 (1998) 207–215 table 1 shows the funding structure of the corporates. in aggregate, only 2.7% of assets are backed by core capital. some corporates have significantly thinner core capital as shown in tables 2 and 3. 4. corporate credit union investment practices since the financial marketplace is becoming increasingly complex, financial institutions need to have an effective risk-management system in place. for corporate credit unions, having an adequate asset-liability management system in place is particularly important given the potential interest rate risk these institutions face. the ability to manage interest rate risk has important implications for meeting the liquidity needs of natural person credit unions. the available evidence suggests that many corporate credit union portfolios are not adequately hedged for interest rate risk. on average, as of june 1996, corporate credit unions have a negative gap position for repricing intervals up to 3 years. gap analysis examines the difference between the repricing of a financial institution’s assets and liabilities. typically, a financial institution will calculate its gap position for several different maturity periods. a negative gap exists when short-tem liabilities exceed short-term assets. the average one-year gap position for corporate credit unions is23.3% of average earning assets and the gap position is24.5% of average earning assets for a repricing interval of 3 years and less. as table 1 credit union investment in corporate credit unions as of december 31, 1996 median percent of: investments, % assets, % capital, % all corporate credit unions 75.8 68.6 522.8 all natural person credit unions: 42.5 9.0 75.0 assets under $10 million 55.5 11.4 85.6 $10 to $50 million 33.4 7.6 69.5 $50 to $100 million 24.9 6.0 57.8 over $100 million 17.9 4.6 44.4 source: credit union call reports (form 5100) submitted to the national credit union administration. table 2 aggregate corporate credit union balance sheet millions of dollars as of december 31, 1996 assets $48,024 cash and cash equivalents 2,260 loans and investments 45,270 other assets 494 liabilities $48,024 member deposits 37,698 other borrowing 6,760 membership capital share deposits 2,280 other capital 1,286 source: credit union call reports submitted to the national credit union administration. 211r.j. boldin et al. / financial services review 7 (1998) 207–215 an example, southwest corporate federal credit union, the third largest corporate credit union, has a negative one-year gap position of 20.4% of average earning assets. for repricing intervals up to 3 years, the gap position is234.3%. in fact, 12 corporate credit unions have 3-year gap repricing intervals in excess of 10% of average earning assets; 7 corporates had negative one-year gap positions exceeding 20% of average earning assets; and, one corporate had a negative one-year gap of 40%. this negative gap position shows that interest rates paid on the liabilities of these institutions will react more quickly to changing interest rates than those of their assets. thus, corporate credit unions, in aggregate, may be assuming interest rate risk when interest rates rise. unless derivatives are properly employed, this maturity mismatch could lead to liquidity problems. in conjunction with the above, black (1994) reported that corporate credit unions are assuming more risk in their investment practices and their portfolios than in the past. competition for credit unions’ investable funds will only create greater incentives for corporate credit unions to take on more risk, and will increase the use of complex investment instruments and strategies. capcorp, for example, had $63 million in cmos in september 1990; by september 1994, its holdings of cmos exceeded $1 billion. obviously, this raises concerns regarding the ability of corporate credit unions to manage more complex portfolios, and the ability of ncua to evaluate and supervise the corporates’ risk-taking. the black (1994) report also found that only two corporate credit unions were properly using derivative products. additionally, it noted that line managers at corporates agreed that natural person credit unions and even some corporates had little understanding of the risk and pricing of many financial instruments. this lack of management expertise could expose institutions and the ncusif to significant loses. 5. potential exposure sensitivity analysis was performed to examine npcu exposure to the corporates. in separate simulations, each npcu was assumed to suffer a (1) 5%, and (2) 10% loss on its investment in a corporate. the two simulations supposed that each npcu’s capital was reduced by five/ten percent of its investment in a corporate. the calculations used end-of1996 “call report” data on the balance sheets of every npcu (as reported to the ncua table 3 corporate credit union capital-to-assets ratios: distribution as of december 31, 1996 core capital*-toassets ratio number of institutions aggregate assets ($ in billions) 1.5% 1 17.9 2%–3% 11 4.9 3%–4% 20 23.7 4%–5% 9 1.5 * capital does not include “member capital share deposits.” source: credit union call reports submitted to the national credit union administration. 212 r.j. boldin et al. / financial services review 7 (1998) 207–215 semi-annually by the npcus). then, an institution was identified as having significant exposure if the write-down could have reduced its capital to below four percent of assets (based on standards set in the fdic improvement act (fdicia) of 1991 for federallyinsured banks and thrifts). according to fdicia, a bank is undercapitalized if its leverage ratio falls below 4 percent. congress has recently mandated fdicia-like capital standards on npcus beginning in 2000. this analysis does not intend to suggest that all corporate credit unions could suffer material losses at the same time, leading to broad-scale losses for npcus. it ignores the fact that institutions would certainly react to shore up their positions if they suffer investment losses. the point of the simulation is only to consider the number of institutions that have enough exposure to be significantly impacted in case any corporate credit union suffers a major loss. supporting this approach, the u.s. department of the treasury (1997) commented that given the relative homogeneity of the balance sheet of corporate credit unions, concentration risk associated with a single asset class may be exacerbated within the corporate network. the results in table 4 suggest that a large number of npcus might, in fact, have meaningful exposure to corporates. without offsetting adjustments, npcu capital could fall below a 4% standard in 124 institutions with $0.4 billion of assets under a 5% write-down of investments in corporates, and in 300 institutions with $1.5 billion of assets under a 10% write-down. a 5% write-down could exhaust the capital of 22 institutions with $20 million of assets, while a 10% write-down could wipe out the capital of 61 institutions with $90 million of assets. the impact of a loss on a npcu’s investment in a corporate apparently would fall disproportionately on smaller npcus, which are more reliant on the corporates for investment services. funds held by corporates account for most of the investment portfolios of half the npcus with less than $10 million in assets, compared to less than one-fifth for half of the npcus over $50 million in assets. under either write-down, most of the npcus whose capital could become inadequate have less than $10 million of assets. if larger credit unions continue to have less exposure to the corporates, then the widespread consolidation movement within the industry may help to alleviate this exposure. a 1998 decision by the united states supreme court innational credit union administration v. first national bank & trust co. (1998) slowed this consolidation by limiting future mergers of occupation-orassociation-based federal credit unions to only those which result in membership from a table 4 simulated impact of a loss on natural person credit union funds invested in corporate credit unions credit unions whose capital-to-assets ratio could fall to: five percent loss ten percent loss number assets ($ millions) number assets ($ millions) under 0% 22 20 61 90 0% to 2% 37 100 62 270 2% to 4% 65 240 177 1,170 note: simulation based on december 1996 credit union call report data. 213r.j. boldin et al. / financial services review 7 (1998) 207–215 single group. the 1998 supreme court decision was overturned by the credit union membership access act of 1998, which was signed into law in august 1998. 6. conclusions the findings of this analysis suggest that the structure of the credit union industry imposes interlocking risks on the system. the soundness of every level can be affected by instability of a level above it. the forty corporate credit unions depend on the u.s. central credit union, and each of the corporates has an average of 300 credit union members, ranging from 27 to 1,263, which depend on it. the liquidity role of each corporate can be especially critical if a strong economy drives up loan demand and interest rates at the same time, which is not an unlikely circumstance. as npcus draw down their reserves at corporates to meet local loan demand, corporates can be forced to liquidate investments. at the same time, the higher interest rates could undermine the values of corporate longer-term investments unless they are adequately hedged. if corporates suffer losses when they sell investments to satisfy the liquidity demands, their capital could become impaired. in this case, corporates could be unable to meet the npcus’ liquidity needs. if the u.s. central credit union becomes illiquid for the same reason, there could conceivably be much broader liquidity problems in the credit union industry. in many ways, this structure is similar to the pyramiding structure of bank reserves under the old national banking system, which suffered from periodic liquidity crises, financial panics and banks failures. thus, should any single corporate become destabilized, its credit union members likewise could be destabilized. while this paper represents a simulation, some of the losses actually occurred. as noted by johnson (1995), 247 of the 483 credit union members of the larger corporate that failed in 1995 suffered resultant losses. therefore, the existing structure poses potential systemic risk in the credit union system. given the potential for losses, shareholder (depositor) members of any natural person credit union should be aware of the level of investment of their institution in a corporate credit union, and of the credit union’s dependence on the corporate for liquidity. they need to monitor the capital and risk exposure of any corporate with which their institution deals. since every credit union is owned by its members, the membership has the right and responsibility, acting though the board of directors, to control this exposure. the existence of share insurance for credit union members, however, tends to reduce the incentive to monitor the behavior of their credit union. references barth, j., hudson, c., and jahera, j. (1995). s&l closures and survivors: are there systematic differences in behavior. in a. cottrell, m. lawlor & j. wood (eds.),the causes and costs of depository institution failures.boston: kluwer academic publishers. black, h., deprince, a., ford, w., kudlinski, j., & schweitzer, r. (1994).corporate credit union network investments: risks and risk management. report to the national credit union administration. 214 r.j. boldin et al. / financial services review 7 (1998) 207–215 carnell, r. (1997, november 17). speech before the association of corporate credit unions annual meeting, del mar, ca. fried, h. o., lovell, c. a. k., & eeckaut, p. v. (1993). evaluating the performance of u.s. credit unions.journal of banking and finance 17, 251–265. fried, h. o., & lovell, c. (1993).evaluating the performance of credit unions. madison: filene research institute. houston, j. f., & ryngaert, m. d. (1994). the overall gains from large bank mergers.journal of banking and finance 18, 1155–1176. hube, k., & murray, m. (1995). credit unions still top banks on deals for basic service.the wall street journal, march 3. johnson, r. (1995, march 8).impact on the condition of the credit union industry and the national credit union share insurance fund, adequacy of national credit union administration supervision and regulatory program, and actions to prevent failures in the future.hearings on the failure of capital corporate federal credit union. kane, e., & hendershott, r. (1996). the federal deposit insurance fund that didn’t put a bite on u.s. taxpayers. journal of banking and finance 20, 1305–1328. lindgren, c., garcia, g., & saal, m. (1996).bank soundness and macroeconomic policy. washington, dc: international monetary fund. murray, j., & white, r. (1980). economies of scale and deposit taking financial institutions in canada: a study of british columbia credit unions.journal of money, credit and banking 12, 58–70. national credit union administration. (1996).yearend statistics for federally insured credit unions. national credit union administration v. first national bank & trust co.(1998). 118 s. ct. 927. pearce, d. (1984). recent developments in the credit union industry.economic review, federal reserve bank of kansas city 69, 3–19. smith, d. (1984). a theoretical framework of the analysis of credit union decisionmaking.the journal of finance 39, 1155–1168. stiroh, k. j. (1997). the evolution of an industry: u.s. thrifts in the 1990s.journal of banking and finance, 21, 1375–1394. u.s. department of the treasury. (1997).credit unions. december. u.s. general accounting office. (1991). credit unions: reforms for ensuring future soundness.(gao/ ggd-91-85). july. u.s. general accounting office. (1994).corporate credit unions: conditions, issues, and concerns. (gao/ t-ggd-95-15). october 6. u.s. general accounting office. (1995a).the failure of capital corporate federal credit union. (gao/tggd-95-107). february 28. u.s. general accounting office. (1995b).proposed reforms for corporate credit union regulation. (gao/tggd-95-115). march 8. winton, a. (1997). competition among financial intermediaries when diversification matters.journal of financial intermediation 6, 307–346. 215r.j. boldin et al. / financial services review 7 (1998) 207–215 pii: s1057-0810(00)00062-7 stock selection based on morningstar’s ten-year, five-star general equity mutual funds anthony l. lovisceka,*, w. john jordanb,1 adepartment of finance and legal studies, seton hall university, south orange, nj 07079, usa bdepartment of economics, seton hall university, south orange, nj 07079, usa received 14 june 1999; received in revised form 16 august 2000; accepted 16 august 2000 abstract recent research suggests that the individual investor can build stock portfolios that outperform broad market indices. based on this research and on evidence supporting the persistence of mutual fund performance, we test whether or not the individual investor can build market-superior portfolios from stocks selected from the top holdings of morningstar’s ten-year, five-star general equity mutual funds. we use modern portfolio theory to construct the portfolios. although the portfolios tend to outperform the s&p 500 for the 1990s, we conclude that the evidence is not strong enough to recommend this stock selection strategy to the individual investor. © 2000 elsevier science inc. all rights reserved. jel classification:g11 keywords:portfolios; ten-year, five-star funds 1. introduction a successful stock selection strategy is indispensable to active stock investment management. with advances in communications technology and computer software, the individual investor can now implement any one of many strategies, or a combination of them, such as monitoring analysts’ ratings of stocks, tracking financial and economic variables, charting * corresponding author. tel.:11-973-761-9127; fax:11-973-761-9217. e-mail address: loviscto@shu.edu (a.l. loviscek). 1 w. john jordan died unexpectedly during the revision of this paper. this paper is dedicated to him. financial services review 9 (2000) 145–157 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00062-7 stock prices, using asset pricing models, reading investment newsletters, following the activities of portfolio managers, and implementing stock screening programs. which of these strategies helps the individual investor outperform passive investment alternatives (e.g., investing in a stock index mutual fund), however, is certainly a contested issue. from an efficient markets perspective, fama (1995) provides a lucid summary of why stock selection strategies are generally found wanting. as additional support, metrick (1999) finds no evidence of superior stock selection from among 153 investment newsletters. walker and hatfield (1996) conclude that the individual investor is unlikely to profit from analysts’ recommendations published inu.s.a. today.chandy, peavy and reichenstein (1993) demonstrate that the potential to earn excess returns from value line’s “stock highlight” section is short-lived at best. fromme (1993) and michaud (1998) assert that popular stock selection variables, such as the dividend-yield, the price-earnings ratio, and the price-book ratio, may well be too simplistic to be consistently effective, a conclusion that chan (1992) reaches when examining the use of beta. zimmerman and zogg-wetter (1992) present results that question the efficacy of market timing techniques and tests used to detect market timing ability. rohweder (1998) provides evidence that quantitative portfolio construction methods, because they require accurate forecasts of returns and covariances, may deliver disappointing results. an emerging literature, however, suggests that the profession may have, in some cases, underestimated the efficacy of stock selection methods. for example, gold and lebowitz (1999) show that computerized stock screening programs contain stock selection strategies that can help the individual investor significantly exceed the performance of market indices and professional money managers. sahu and kleiman (1998) uncover evidence of superior stock selection ability among portfolio managers at bank trust departments. rich and reichenstein (1993), using value line’s projections, find that market timing based on the expected market risk premium, dividend yield, and the earnings-price ratio may enable the individual investor to beat the s&p 500. mann and solberg (1991) also suggest the potential of using a risk premium to track stocks. in addition, new stock selection strategies are not hard to find. badrinath and kini (1992) show the potential of constructing portfolios from stocks selected on the basis of capm anomaly variables, such as firm size, the earnings-price ratio, and the price-book ratio. dennis, perfect, snow and wiles (1995) uncover similar evidence, while barbee, mukherji and raines (1996) offer support for the sales-price ratio. herzberg (1998) argues that economic value-added may be an accurate measure of a firm’s ability to generate abnormal earnings. wong, wang, goh and quek (1992) demonstrate the potential of fuzzy neural systems for selecting stocks and forecasting market returns. additional support for active management is found in the mutual fund literature. as evidence, using an optimal weighting method based on modern portfolio theory, elton, gruber and blake (1996) find that historical risk-adjusted returns for common stock funds contain predictive power for up to three years. volkman and wohar (1996) also find evidence of persistence in mutual fund performance. khorana and nelling (1998) show that morningstar’s rating system provides the individual investor with valuable information about the future performance of mutual funds. they discover that higher-rated funds, such as those 146 a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 with four and five stars, tend to have a higher risk-adjusted performance, and one that persists for several years. khorana and nelling’s study is instructive in at least one more way. it provides further justification for advertisements and media reports that highlight five-star funds. often included in these advertisements and reports are the respective track records of the funds and their top holdings. oliver (1992), in fact, states that some investors construct portfolios by selecting stocks from the published holdings in these highly rated funds. the appeal of this selection strategy for the individual investor is seen not only in its ease of implementation, but also in the fact that the portfolio manager has actually purchased the stocks (as opposed to the individual investor selecting stocks based only on, for example, an analyst’s recommendation). in summary, although the evidence on the efficacy of stock selection strategies, and in turn active stock portfolio management, has often been found wanting, there is emerging evidence that some stock selection strategies may be able to help the individual investor outperform broad market indices. this study tests one such strategy. guided by the application of modern portfolio theory used by elton, gruber and blake (1996), the findings of khorana and nelling (1998), and the observation by oliver (1992), it tests the efficacy of building stock portfolios from the top holdings in morningstar’s highest-rated general equity mutual funds. the issue is whether or not these portfolios can help the individual investor outperform a broad index of the u.s. stock market. the hypothesis under test is that the portfolios constructed from these highest-rated funds will not systematically offer a risk-adjusted performance that exceeds that of the broad domestic market. as part of the test, and in line with barber and odean (1999) and odean (1999) that the individual investor should avoid frequent trading, we test the risk-adjusted performance of the portfolios using a series ofex anteone-to-five year buy-and-hold strategies. the analysis is conducted as follows: (a) the selection of the top stocks from morningstar’s highest-rated funds; (b) the application of an optimal portfolio construction approach to these selections; and (c) the drawing of inferences from theex antetests. 2. stock selection and data morningstar’s ten-year, five-star, general equity mutual funds serve as the source for our stock selection. this is the company’s highest rating and is one that it has typically awarded to only about 2% of the approximately 2,000 mutual funds that it evaluates. if elton, gruber and blake (1996) and khorana and nelling (1998) are right, then it should be the case that the application of modern portfolio theory to the top holdings of the ten-year, five-star funds stands the best chance of outperforming a broad market index. the index used in this study is the s&p 500, a standard benchmark used to assess portfolio performance. to construct the portfolios, we begin by selecting the top five holdings (i.e., stocks that have the largest percentage of a fund’s total assets) in each ten-year, five-star fund for the years 1989 through 1993. the reasons for using the top five are threefold. first, not only are these holdings easily identified, but they also represent the fund manager’s top selections. 147a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 second, by order of weight, these holdings have the largest impact on a mutual fund’s performance. third, anything other than a small number of stocks selected from each fund makes the analysis unmanageable for the individual investor. the stock selections come from the end-of-year composition of the ten-year, five-star funds provided inmutual funds(morningstar, 1994) andmutual fund sourcebook(morningstar, 1989–1994). for each stock, monthly rates of return, which include dividends and adjustments for stock splits, are computed from six years of monthly closing prices, as taken from dow jones news retrieval, a source (at the time of this study) readily available to the individual investor. we obtain the monthly rates of return on the s&p 500 from ibbotson associates’stocks, bonds, bills, and inflation(2000). following the literature, we express the returns in excess form by subtracting an estimate of a monthly risk-free rate of return, which is calculated from the monthly closing rate on the three-month u.s. treasury bill. the source is thewall street journal. by relying on the fourth-quarter holdings of the funds, we encounter the possibility of a bias. in this study a bias may arise, for example, if a fund manager, late in the fourth quarter, engages in trading activity known as “window dressing.” in this case, a fund manager adds current stock favorites and drops unfavorable ones, presumably to give the appearance that the fund has been invested in successful stocks. although these newly added stocks were not the basis for the fund’s most recent performance, they become candidates for inclusion in the portfolios used in this study. as long as these stocks reflect the fund manager’s expectations for the forthcoming quarters, and are not dropped at the beginning of the next quarter, no significant bias should occur. moreover, the cost of adding to and subtracting from the top holdings of a fund can be significant, especially for large-cap funds. if done frequently, window dressing could lead to a drop in a fund’s performance and the loss of its prestigious five-star rating. thus, all things considered, in this study we judge that the impact of window dressing on the results is unlikely to be material. the number of top five stocks, across the years, is as follows: 95 (1989), 100 (1990), 143 (1991), 109 (1992), 84 (1993). we are able to obtain data on approximately 92% of all the stocks. foreign listings and newly issued stocks, for which the data are not sufficiently available, account for the rest. two reasons explain the difference in the number of stocks per year: (a) the variation in the number of ten-year, five-star funds and (b) the degree of commonality in the top five stocks across the ten-year, five-star funds. for example, there were 28 ten-year, five-star general equity funds each in 1989 and 1990, 39 in 1991, 33 in 1992, and 25 in 1993. there is also a similarity in the investment style across the ten-year, five-star general equity funds. in fact, about 91% of all the stocks come from funds that fit four closely related classifications that morningstar uses: “growth,” “growth and income,” “equity-income,” and “balanced,” with “growth” and “growth and income” accounting for the majority of the stocks. about 6% come from “income” funds and about 3% come from “small company” funds. no stocks come from “aggressive growth.” the similarity in investment objective leads to a commonality in the holdings, especially in large-cap stocks such as coca cola, ford, general electric, intel, motorola, and philip morris. we should also point out that morningstar’s classification system (e.g., “aggressive growth,” “growth,” “growth and income,” etc.) plays no role in this study’s stock selection process. this is because these fund classifications do not explicitly consider a specific stock’s 148 a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 return to risk,as specified in modern portfolio theory, a point that elton, gruber and blake (1996) and bogle (1998) underscore. thus, while one would expect the return on a stock selected from, say, a growth fund to be higher than the return on a stock chosen from, say, an income fund, the same may not be true for the expected return to risk on each stock. 3. portfolio construction and estimation the construction of the portfolios follows the work of elton, gruber, and padberg (egp) (elton and gruber, 1995, chapter 9) and elton, gruber, and blake (1996). three reasons support the use of their method. first, it offers the individual investor an intuitive interpretation for the decision rules. second, the solution is easily obtained from spreadsheet programs because the egp approach relies on betas and simple equations, and avoids the more demanding, iterative, mathematical programming solutions of the markowitz (1991) full-covariance method. these are features that the individual investor should find attractive. third, as shown by nawrocki (1996), the egp approach gives satisfactory results relative to the full-covariance method. the egp approach, which in this study excludes short sales, is based on modern portfolio theory. the approach consists of four steps: (a) the estimation of the single-index model to obtain systematic and unsystematic risks; (b) the ranking of the stocks by excess return to beta; (c) the calculation of a cutoff level to determine the composition of the portfolio; and (d) the determination of the proportion of the portfolio invested in each security. in equations, the steps are as follows: single-index model: (rit 2 rf) 5 ai 1 bi(rmt 2 rf) 1 eit (1) excess return to beta: ret./riski 5 ri 2 rf bi (2) cut-off level: ci 5 sm 2 o j51 i (rj 2 rf)bj sej 2 1 1 sm 2 o j51 i s bj 2 sej 2 d (3) proportions: wi 5 zi / o j51 n zj, where (4) zi 5 bi/sei 2 ((ri2rf)/bi 2 c*) where ri 5 return on stock i; rm 5 return on the market; ri 5 expected return on stock i; rf 5 return on a riskless asset; 149a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 bi 5 beta of stock i; sm 2 5 variance of the market; sei 2 5 unsystematic risk of stock i; wi 5 proportion invested in stock i; and c* 5 maximum value of ci 4. application and results to apply the egp approach, we use four steps. first, following haugen and baker (1991), markowitz and xu (1994), elton, gruber and blake (1996), and thaler (1999), we compute monthly rates of return on each stock from six years of monthly historical stock prices. following this literature, we then subtract the respective monthly treasury bill rates of return from the monthly returns on each stock to express the stock returns in excess form. for each stock, the arithmetic mean of the monthly excess returns serves as the expected return. for example, as part of building the 1989 portfolio, which begins with 95 stocks, we compute the monthly returns on each of these stocks from 1984 through 1989, subtract the respective monthly treasury bill returns, and find the arithmetic mean of the excess returns. for the 1990 portfolio, which begins with 100 stocks, we compute the monthly returns on each of these stocks from 1985 through 1990, followed by subtracting the respective monthly treasury bill returns and finding the arithmetic mean of the excess returns. the same logic applies to the 1991, 1992, and 1993 portfolios. second, by regressing the monthly excess returns on each stock against the monthly excess returns on the s&p 500, we estimate eq. (1), the single-index model. it provides the estimates of each stock’s beta and unsystematic risk. the latter is obtained by squaring the standard error of the regression. to improve the accuracy of these historical betas as estimates of future betas, following kolb and rodriguez (1990) and tucker, becker, isimbabi and ogden (1994), we adjust the historical betas for the possibility of nonstationarity by using the standard formula 0.3431 0.677 x (historical beta). this formula is designed to account for the observed tendency of betas to regress to 1. third, following eq. (2), we rank the stocks, from highest to lowest, in terms of each stock’s excess return to beta. in doing so, we are following the principles of modern portfolio theory in that we are attempting to maximize a portfolio’s return for a given level of risk. for example, for the 1989 portfolio, we rank the 95 stocks in terms of excess return to beta. for the 1990 portfolio, we rank the 100 stocks in terms of excess return to beta, and so on for the 1991, 1992, and 1993 portfolios. for each ranking of stocks, the task now becomes one of determining which of the stocks make up each portfolio. fourth, to determine the composition of each portfolio, we apply eq. (3), the cutoff formula. the rule is that a stock is included in the portfolio as long as its excess return to beta exceeds its cutoff value, ci. an interpretation of this step is that the more a stock is expected to perform relative to that of the portfolio, the more likely the stock will be included in the portfolio. after determining the composition of each portfolio, we determine how much to invest in each stock by following eq. (4). an interpretation of this equation is that the higher a stock’s expected excess return to beta and the lower its unsystematic risk, the greater the 150 a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 percentage that will be invested in the stock. as an additional test of the results, we construct a set of portfolios without using eq. (4). the result is a set of portfolios of stocks that have equal weights. the composition of each of the egp portfolios is shown in table 1. the 1990 portfolio has the largest number of stocks; thirteen have excess returns to beta that exceed the cutoff values. the 1993 portfolio has the smallest number, with only eight stocks exceeding the cutoff levels. regardless of the number, the composition of each portfolio tends to be table 1 composition of the portfolios built from the top five holdings of morningstar’s ten-year, five-star funds using the elton-gruber-padberg approach 1989 portfolio 1992 portfolio ameritech bombay de beers coca cola dillard’s conseco p.h. glatfelter harley davidson gpu intel houston industries international game technology merck microsoft philip morris minorco pacific telesis weatherford international pacificorp waste management 1993 portfolio chevron 1990 portfolio dsc communications atlantic richfield entergy central and south west intel dqe mci communications gpu microsoft mccormick motorola merck scimed life systems minorco pepsico philip morris royal dutch shell schering plough texaco union electric 1991 portfolio amgen biomet coca cola de beers gpu medtronic merck philip morris royal dutch shell the gap unilever 151a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 dominated by large-cap stocks, such as chevron, coca cola, intel, medtronic, merck, pepsico, philip morris, and royal dutch shell. after building a portfolio for a given year (i.e., 1989, 1990, 1991, 1992, and 1993), we track its annualex anteperformance for five years, beginning on january 1. to test the hypothesis, we calculate the geometric means of the risk-adjusted excess returns on each portfolio and the corresponding geometric means of the risk-adjusted excess returns on the s&p 500. we then compare the two. for example, with respect to the 1989 portfolio, we compute the geometric means for the risk-adjusted excess returns for 1990 through 1991, 1990 through 1992, 1990 through 1993, and 1990 through 1994. the geometric mean for each of these holding periods is then compared to the geometric mean for the s&p 500 for each of the same holding periods. in tracking the performance of each portfolio, we find that two companies, dsc communications and scimed life systems, were acquired during the testing period. both companies are part of the 1993 portfolio. we assume liquidation of the stock at the closing price of the month that the acquisition occurred and the earning of the risk-free rate for the remainder of the testing periods. to test for a difference in the performance between the geometric means for the portfolios and those for the s&p 500, following kolodny, laurence and ghosh (1989) and higgins and peterson (1998), we use the probability values associated with the wilcoxon signed-ranks test. we use this test, as opposed to using t values, because, as higgins and peterson point out, the power of a test involving t values is highly questionable with a sample of five observations (i.e., one-year to five-year buy-and-hold periods) and a nonnormal distribution. the probability values associated with the wilcoxon test that are positive and significant at the 5% level support the observation that the performance of the egp portfolios exceeds that of the s&p 500. the negative values that are significant at the 5% level confirm the opposite observation. the returns, which are risk adjusted and annualized, are shown in tables 2 and 3. the reason for the difference in the two sets of results lies in the weights assigned to the stocks that make up the portfolios. the portfolios in table 2 are constructed by using eqs. (1) through (4). thus, the stocks that make up these portfolios have unequal weights. the portfolios in table 3, however, are built by using only eqs. (1) through (3). thus, the stocks that make up the portfolios in table 3 have equal weights. we examine the portfolios of equally weighted stocks to see to what extent the performance of the egp portfolios varies with changes in the weights. because the egp approach inherently uses the treynor measure of portfolio performance (elton and gruber, 1995, p. 645), all numbers in the tables are expressed in terms of this measure. in the upper part of each table, the column on the left represents the years for which the portfolios are built; the top row refers to the years for which the portfolios are tested; and the bottom row represents the excess return divided by beta (51) for the s&p 500. this manner of presentation permits the examination of one-year, two-year, three-year, four-year, and five-year buy-and-hold strategies for both the egp portfolios and the s&p 500. note that the diagonal elements (e.g., in table 2, -0.0039, 0.5740, -0.2618, 0.1023, and 0.0944) illustrate the one-year strategy. in table 2, an examination of the results across the five buy-and-hold categories (one year, 152 a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 two years, etc.) leads to three conclusions. first, based on a comparison of the 25 geometric means for the egp portfolios (i.e., five portfolios and five holding periods) versus the 25 geometric means for the s&p 500, the egp portfolios outperform the s&p 500 in 17 instances. several of the differences are large, in excess of ten percentage points, with the largest being over 33 percentage points (i.e., 1990 portfolio, one year). second, both the 1990 portfolio and the 1993 portfolio outperform the s&p 500 across all five holding periods. at the 5% level of significance, the probability values from the wilcoxon test confirm this observation. third, the s&p 500 outperforms the 1991 portfolio for all five holding periods, as the probability value confirms. a major reason for the 1991 portfolio’s subpar performance relative to that of the s&p table 2 portfolio performances: s&p 500 versus portfolios built from the top holdings of morningstar’s ten-year, five star funds using the elton-gruber-padberg approach, equations 1 through 4 portfolio building year annual risk-adjusted excess returns of the egp portfolios and the s&p 5001 portfolio testing year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1989 20.0039 0.3630 20.0747 20.0847 20.0716 — — — — 1990 — 0.5740 20.0058 0.0830 20.0986 0.6108 — — — 1991 — — 20.2618 20.0159 0.1153 0.5370 0.0394 — — 1992 — — — 0.1023 20.0046 0.1284 0.4867 0.2455 — 1993 — — — — 0.0944 0.3496 0.3544 0.2538 0.4535 s&p 500 20.1072 0.2411 0.0379 0.0690 20.0174 0.3192 0.1818 0.2843 0.2333 geometric means of the annual risk-adjusted excess returns: s&p 500 versus egp portfolios2 1990 1991 1992 1993 1994 p-value 1989 portfolio 20.0039 0.1652 0.0790 0.0355 0.0132 0.1125 s&p 500 20.1072 0.0526 0.0477 0.0530 0.0385 1991 1992 1993 1994 1995 p-value 1990 portfolio 0.5740 0.2509 0.1922 0.1117 0.1973 0.0212* s&p 500 0.2411 0.1350 0.1125 0.0785 0.1229 1992 1993 1994 1995 1996 p-value 1991 portfolio 20.2618 20.1476 20.0648 0.0564 0.0530 20.0212# s&p 500 0.0379 0.0533 0.0292 0.0951 0.1119 1993 1994 1995 1996 1997 p-value 1992 portfolio 0.1023 0.0475 0.0738 0.1648 0.1805 0.2502 s&p 500 0.0690 0.0249 0.1149 0.1312 0.1603 1994 1995 1996 1997 1998 p-value 1993 portfolio 0.0940 0.2154 0.2601 0.2585 0.2953 0.0212* s&p 500 20.0174 0.1385 0.1528 0.1843 0.1940 1,2 the annual risk-adjusted excess returns are the portfolio (and the s&p 500) returns minus the annual three-month u.s. treasury bill rate divided by the portfolio’s beta. to illustrate the geometric means, the 1989 portfolio’s geometric mean through 1991, a two-year holding period, is 0.1652, the geometric mean of20.0039 and 0.3630. the corresponding value for the s&p 500 is 0.0526, the mean of20.1072 and 0.2411. the probability values (i.e., “p-values”) are based on the wilcoxon signed-ranks statistic, where “*” indicates significance at the 5% level in favor of the egp portfolios and “#” indicates significance at the 5% level in favor of the s&p 500. 153a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 500 is its negative performance in 1992. table 2 illustrates that -0.2618 is by far the worst risk-adjusted excess return of any of the portfolios and, as a result, one that is difficult to offset relative to the performance for the s&p 500. for example, while the 1991 portfolio soundly beats the s&p 500 in 1994 and 1995, registering performances of 0.1153 and 0.5370, respectively, these performances are not enough to help the 1991 portfolio outperform the s&p 500 across any of the buy-and-hold strategies. there is some evidence that the 1989 and 1992 portfolios come fairly close to outperforming the s&p 500. the 1989 portfolio actually does it in three instances, but the overall difference, as shown by the probability value, is not significant at the 5% level. a similar conclusion is reached with the 1992 portfolio. it actually beats the s&p 500 in four of the five cases, but the probability value of 0.2502 casts some doubt on the efficacy of this table 3 portfolio performances: s&p 500 versus portfolios built from the top holdings of morningstar’s ten-year, five star funds using the elton-gruber-padberg approach, equations 1 through 3 portfolio building year annual risk-adjusted excess returns of the egp portfolios and the s&p 5003 portfolio testing year 1990 1991 1992 1993 1994 1995 1996 1997 1998 1989 0.0104 0.3217 20.1412 0.0412 20.1014 — — — — 1990 — 0.5921 0.0222 20.0164 20.1297 0.4973 — — — 1991 — — 20.1934 0.0377 0.0712 0.4943 0.1297 — — 1992 — — — 0.1858 0.0343 0.0881 0.3877 0.1792 — 1993 — — — — 0.0708 0.3022 0.2854 0.1462 0.3644 s&p 500 20.1072 0.2411 0.0379 0.0690 20.0174 0.3192 0.1818 0.2843 0.2333 geometric means of the annual risk-adjusted excess returns: s&p 500 versus egp portfolios4 1990 1991 1992 1993 1994 p-value 1989 portfolio 0.0104 0.1556 0.0468 0.0454 0.0142 0.3430 s&p 500 20.1072 0.0526 0.0477 0.0530 0.0385 1991 1992 1993 1994 1995 p-value 1990 portfolio 0.5921 0.2757 0.1698 0.0864 0.1584 0.0212* s&p 500 0.2411 0.1350 0.1125 0.0785 0.1229 1992 1993 1994 1995 1996 p-value 1991 portfolio 20.1934 20.0851 20.0357 0.0759 0.0864 20.0212# s&p 500 0.0379 0.0533 0.0292 0.0951 0.1119 1993 1994 1995 1996 1997 p-value 1992 portfolio 0.1858 0.1074 0.1010 0.1665 0.1690 0.0690 s&p 500 0.0690 0.0249 0.1149 0.1312 0.1603 1994 1995 1996 1997 1998 p-value 1993 portfolio 0.0780 0.1809 0.2147 0.1972 0.2293 0.0212* s&p 500 20.0174 0.1385 0.1528 0.1843 0.1940 3,4 the annual risk-adjusted excess returns are the portfolio (and the s&p 500) returns minus the annual three-month u.s. treasury bill rate divided by the portfolio’s beta. to illustrate the geometric means, the 1989 portfolio’s geometric mean through 1991, a two-year holding period, is 0.1556, the geometric mean of 0.0104 and 0.3217. the corresponding value for the s&p 500 is 0.0526, the mean of20.1072 and 0.2411. the probability values (i.e., “p-values”) are based on the wilcoxon signed-ranks statistic, where “*” indicates significance at the 5% level in favor of the egp portfolios and “#” indicates significance at the 5% level in favor of the s&p 500. 154 a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 portfolio. the reason for this high value is seen in the portfolio’s performance during 1995. while its risk-adjusted excess return is a respectable 0.1284, the s&p 500’s is 0.3192. as a result, by 1995 the geometric mean of this egp portfolio has underperformed the s&p 500 by over four percentage points, the largest difference (in absolute terms) across the five pairwise comparisons. using the wilcoxon test, we find that the size of the (negative) performance relative to that of the s&p 500 makes it impossible to reject the null hypothesis that no significant difference exists in the results. turning to table 3, the results based on equally weighted stocks, we find similar results. the geometric mean performance for the egp portfolios exceeds that of the s&p 500 in 16 of the 25 possible pairwise comparisons. both the 1990 and 1993 portfolios beat the s&p 500 for all holdings periods, as the probability values confirm. once again, the 1991 portfolio underperforms the s&p 500 across all holding periods. as before, the 1989 portfolio does not outperform the s&p 500. the 1992 portfolio outperforms it in four of the five holding periods. unlike in table 2, however, the difference across the periods is significant at the 10% level. this is because the difference in performance by 1995 is smaller than it is in table 1. do the results suggest how long the individual investor should hold a portfolio? this question motivates an examination of the performances by holding period. as a guide to answering the question, elton, gruber and blake (1996), based on evidence of persistence in mutual fund performance, suggest a holding period of up to three years. in both tables 2 and 3, there is some evidence that tends to support this observation. for example, moving down the first column of the geometric means in table 2, we find that during the one-year holding period the egp portfolios outperform the s&p 500 in four of the five instances (i.e., -0.0039 vs. -0.1072; 0.5740 vs. 0.2411; 0.1023 vs. 0.0690; and 0.0940 vs. -0.0174). the same pattern exists when moving down the second column of the geometric means. when moving down the third column, we find that the geometric means exceed those for the s&p 500 in three of the five cases. taken together, the egp portfolios outperform the s&p 500 in 11 of the 15 cases. applying the wilcoxon test to these results, however, reveals a probability value of only about 0.11, which is not significant at the 5% level. this insignificance is traced to the subpar performances of the 1991 portfolio. a similar conclusion is drawn from the wilcoxon test for the one-year and two-year holding periods. overall, the results suggest that selecting stocks from the top holdings of morningstar’s ten-year, five-star funds has some potential to help the individual investor outperform the s&p 500. we conclude, however, that the evidence is not strong enough for us to reject the null hypothesis; namely, that the portfolios will not outperform the s&p 500. at least two reasons support this conclusion. first, according to the wilcoxon tests, only two of the five portfolios clearly beat the s&p 500. moreover, although the egp portfolios outperform the s&p 500 in 17 of the 25 pairwise comparisons, the overall evidence is probably not convincing enough to motivate the individual investor who buys into index funds to adopt the stock selection strategy tested here. this point is reinforced by the composition of the ten-year, five-star funds, which is the second reason supporting our conclusion. these funds are generally large-cap funds (e.g., fidelity magellan) whose top holdings are large-cap stocks, such as coca cola, intel, merck, and philip morris. these stocks are the largest contributors to a broad market index like the s&p 500. consequently, the selection of the top 155a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 holdings in these highest rated funds may not often perform much better than a broad market index. this point is buttressed by bogle (1998), who finds that index funds, for the period of 1992 through 1996, have generally delivered the highest risk-adjusted returns among morningstar’s equity fund categories. 5. conclusion recent research suggests that the individual investor can construct stock portfolios that outperform broad market indices. based on this research and on the evidence that supports the persistence of mutual fund performance, we test whether or not the individual investor can build market-superior portfolios from stocks selected from the top holdings of morningstar’s ten-year, five-star general equity mutual funds. we use the elton-gruber-padberg approach, which is founded on modern portfolio theory, to construct the portfolios for the years 1989 through 1993. we test the performance of the portfolios against that of the s&p 500 by using a series of one-year to five-year buy-and-hold strategies from 1990 through 1998. we find some evidence that the portfolios tend to outperform the s&p 500. we conclude, however, that the evidence is not strong enough to recommend the stock selection strategy to the individual investor. acknowledgments the authors are grateful to two anonymous referees, richard covington, frederick crowley, participants at a research seminar at seton hall university, and especially the editor, karen eilers lahey, for criticisms and suggestions on earlier stages of this research. references badrinath, s. g., & kini, o. (1992). portfolio management using a factor-analytic stock selection strategy. managerial and decision economics, 13,305–314. barbee, w. c. jr., mukherji, s., & raines, g. a. (1996). do sales-price and debt-equity explain stock returns better than book-market and firm size?financial analysts journal, 52,56–60. barber, b., & odean, t. (1999). trading is hazardous to your wealth.journal of finance, 55,773–806. bogle, j. c. (1998). the implication of style analysis for mutual fund performance evaluation.journal of portfolio management, 24,34–42. chan, c. (1992). the beta coefficient: its use in measuring market risk and stock selection.journal of investing, 1, 29–36. chandy, p. r., peavy, j. w., & reichenstein, w. (1993). a note on the value line stock highlight effect.journal of financial research, 16,171–179. dennis, p., perfect, s. b., snow, k. n., & wiles, k. w. (1995). the effects of rebalancing on size and book-to-market ratio portfolio returns.financial analysts journal, 51,47–56. elton, e. j., & gruber, m. j. (1995).modern portfolio theory and investment analysis(5th ed). new york: john wiley. 156 a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 elton, e. j., gruber, m. j., & blake, c. r. (1996). the persistence of risk-adjusted mutual fund performance. journal of business, 69,133–157. fama, e. f. (1995). random walks in stock market prices.financial analysts journal, 51,75–80. fromme, s. (1993). p/e, dividend yield, and beta: how viable are they as stock selection criteria?journal of investing, 2,34–41. gold, s. c., & lebowitz, p. (1999). computerized stock screening rules for portfolio selection.financial services review, 8,61–70. haugen, r., & baker, n. l. (1991). the efficient market inefficiency of capitalization-weighted stock portfolios. journal of portfolio management, 17,35–40. herzberg, m. m. (1998). implementing ebo/eva analysis in stock selection.journal of investing, 7,45–53. higgins, e. j., & peterson, d. r. (1998). the power of one and two-sample t statistics given event-induced variance increases and nonnormal stock returns: a comparative study.quarterly journal of business and economics, 37,27–49. ibbotson associates. (2000).stocks, bonds, bills, and inflation.chicago: ibbotson associates. kolb, r. w., & rodriguez, r. j. (1990). is the distribution of betas stationary?journal of financial research, 13, 279–283. kolodny, r., laurence, m., & ghosh, a. (1989). in search of excellence. . .for whom?journal of portfolio management, 15,56–60. khorana, a., & nelling, e. (1998). the determinants and predictive ability of mutual fund ratings.journal of investing, 7,61–66. mann, s. v., & solberg, d. p. (1991). should investors avoid the stock market outside of january?financial services review, 1,101–108. markowitz, h. m. (1991).portfolio selection: efficient diversification of investments(2nd ed). cambridge, ma: blackwell. markowitz, h. m., & xu, g. l. (1994). data mining corrections.journal of portfolio management, 21,60–69. metrick, a. (1999). performance evaluation with transactions: the stock selection of investment newsletters. journal of finance, 54,1743–1775. michaud, r. o. (1998). is value multidimensional? implications for style management and global stock selection. journal of investing, 7,61–65. morningstar. (1989–1994).mutual fund sourcebook.chicago: morningstar. morningstar. (1994).mutual funds.chicago: morningstar. nawrocki, d. (1996). portfolio analysis with a large universe of assets.applied economics, 28,1191–1198. odean, t. (1999). do investors trade too much?american economic review, 89,1279–1298. oliver, s. (1992). cheap growth.forbes, 149(7), 145. rich, s. p., & reichenstein, w. (1993). market timing for the individual investor: using the predictability of long-horizon stock returns to enhance portfolio performance.financial services review, 3,29–43. rohweder, h. c. (1998). implementing stock selection ideas: does error tracking optimization do any good? journal of portfolio management, 24,49–59. sahu, a. p., & kleiman, r. t. (1998). the timing and stock selection abilities of bank funds: evidence based on meta-analysis.journal of financial services research, 13,137–152. thaler, r. h. (1999). the end of behavioral finance.financial analysts journal, 55,12–17. tucker, a. l., becker, k. g., isimbabi, m. j., & ogden, j. (1994).contemporary portfolio theory and risk management.new york: west. volkman, d. a., & wohar, m. e. (1996). abnormal profits and relative strength in mutual fund returns.review of financial economics, 5,101–116. walker, m. m., & hatfield, g. b. (1996). professional stock analysts’ recommendations: implications for individual investors.financial services review, 5,13–29. wong, f. s., wang, f. z., goh, t. h., & quek, b. k. (1992). fuzzy neural systems for stock selection.financial analysts journal, 48,47–52. zimmerman, h., & zogg-wetter, c. (1992). on detecting selection and timing ability: the case of stock market indexes.financial analysts journal, 48,80–83. 157a.l. loviscek, w.j. jordan / financial services review 9 (2000) 145–157 pii: s1057-0810(01)00086-5 ��������� �� �� ������ � ����� � ��� ������ ��� ��� � � ���� ������� �� �� �� � �� �������� ��� �� ������ �� ��� ��� �������������� ���� �������� ������ �� �� �� � ��� ��� ��� ������ ����� ��� � ��������� ��� ��� � ��� ������ ��� � ��� �� !� "#� #���� �� �� � �� !��� "##�$ �� � �� � ��� �� ���� "% ������ "##�$ � ���� �& '����(�� "##� �������� ������� �� ��� ������ ��� ��� ��� � ��� ���) �� ��� �� �� ��� ���(��� ��� ����� �� ������� ������� ( � �� ��� ��� � �)� ��� ���* +� � ����, ��� ������ ��� ��� � � �� ��� ����� ���� ��� (��� �-���� ���� ������ �� � ��)����� ����� � � ��� ��� � �� . ��� ��� ���� ��� ���(��� (�� ���� ��� ��������� �� ����������* /� � ����� ���� �� ������� �� �� ��� ������ ��� ��� � ������� � �� ������� �� � ������ �� ���(��� �� � 0� ���� �� ����������� �� ��� )�� )��� �� ����� ���� �� ����� (��)��� � �.�� (���.� � ���� �� ��� �� ���� ��������� � ������ �� � �����* 1� �-����� ��� �� ������ �� � ���)�� �� �� ����������� �� �� ����� ������� � ��� �� ��� � � ��* � "##� 2���� �� 3 �� � +� * ��� � ���� ������� * ��$ �� ������ ��� % �"%$ !�4$ !"4 &�������% ��� ������ ��� ���$ ��� ��$ 5�.�� ���� (�� ��$ 5�.�� (���.�$ 3���������� ��$ +������� �� � � ���� ������" /�� ���� � �� �� ��� ������ �� ��� ������ ��� ��� � ������� � �� (� ����� �� � ��� )�� � �� �* ��� � � � ���� ���� (� �� �������� �� ����������� �� ��� ���(��� �� � ��� ��� �� ������* ��� � �� ���� � �� ����� � � �� (��)��� � �.�� (���.� ��� 65789 �� �� ��������� � ������ �� � ���� 6+��9� � �� � �� )���� ����������� �� ��� �� (�� ��� ��� �� �� �� �� � �� �� ��� ����� � �� ��� ���(���, ������� ��* :�� ���� �� 8 ��� �� 3��� �� ��� �) �# 6"##�9 ";%<"&# �#=;�#>�#?#�?@ < ��� ����� ������ � "##� 2���� �� 3 �� � +� * ��� � ���� ������� * �++� 3 � # = ; � # > � # 6 # � 9 # # # > 4 � = � ��������� �� ������* /��*� �4��>�& ##�==4=$ ��-� �4��>�& ##�=4%%* �'� �� ��������% �* ����������a� �*� �*�� 6�* ������������9� �*�������a� �*� �*�� 6�* �������9* � /��*� �4��>�& ##�=4#=$ ��-� �4��>�& ##�=4%%* "# ����� � � ����� ���(�� +��� ��(�� ����� ��� ������ �� (��)��� ����� �� � ��� ���� ������ ��� �� �� � (� ��� ��� ��� ������ ��� ��� ��� ���* /� ��� �� ��� ��� �� � ����� �������� ������� �� ��� �� �� ��� �����* ��� (��� ��� ��� (� ��� ������� �� ���� ����� )� � )��� �(� �� ��������� �� ����� ��������� � �� � �� ��� ��� �� )� � ���� (�� ��� ��� �� �� * /� � ��� ���� �� �� �� �� �� ����� �������� �� �� � (�� ��� (� �� ���� � � � ��� ������* 3 � � �������� ���� � (�� ��� �� ����������� �� ��� ��������� ��� ��� ������ �� �� ���� ������ �� ���������� �� � ��� ���� �� )��� ���� � � �� . ��� ��� � �� ��� ��� �� ������� �� )������ �* � ������ �� �� � � ����� �� � ��� (��� � ���� �� �� � ��� ������� ( � �� ��� ��� ��� ���� ��� � ���� ���������� �� ��� ����������� �� ��� �� �� .����� �� ��� � � ��� � ��� ����� ���� ��� � � � ���) �� ��� ������ ����� �� ��� � ��� ��(�� ��� ������ ��� ���*1������� ��� ������� �� b��� . �� �� ��� �� � �� ��� ��� ������� �� ���� �������� ���� ���� �� ���� ��� � � � ���� (��� � ����� �� ��������� �� ��������� ��� ���� � �� ��� ���(�� ������� �� � � ��� ��������� ��� � ���������� ��� ����� �� ����������� �� ��� � �� ����������* 7��)��� ��� !��� �&&> �� %#�� !��� �&&& ���� =#�### ���(��� �� ��� 3���������� �� 3 ���� ��� ������� �� 0� ���� � �� 633�09� )��� � ��� ��� �� � �� ���� � �� ) �� � 578 �� ���� �� �� ��� �� ���� +���* /�� �� � (��)��� � 578 �� �� +��)�� �������� �� � �� � ���� ����� �� ���(���� �� ��������,� �� � ��� ��� (� ����� � ��� ������* /�� 578 ��� ���� ���(��� �� � �.�� ������� ����� �� (� ��� ���(��,� ������ �� ������ �� ���(���� �* /�� ��������� � �� ��� � ������ (� 33�0 ��� � � � ��� ��� � �� ��� ����� �� (���.� �� � �.�� � �������� ��� � ���� ���� ������� �� ���� � � ��� ����� �� ���� � ��� (���.� ���� � (� �� ��� �� �� � � ��* 7����� ��������� � (� ��� ��� ���(��� �� � ����� � ��� �.�� (���.� ���� ���* 0� �� ��������� �� ������ �� � � ��������� � 578 ��� �� )��� �������� (� �� � �� (� ��� ��������* 0� �� ��� +��� �� � ���(�� ��� �� � � ��� � ���� �� )� � ���� (�� ��� ��� �� � �� ������� ���� ��� � ����� ��������� ��� �� ��� �� �� * 33�0 ���(��� �� ���� ��������� ��� ��� �� ����� ���� �� � �� � ���� �� � ��� ��������� ��� �� �� (� ����� � !��� �� � ����* /�� ��������� � �� ��� ��������� �� ��� ����� ��� ��� ������ ���� ���(��� �� � �� ���� ��� ������* /�� �� � �� (� �� � (� 33�0 ���(��� �� � � �� . ��� ������ �� ��� � �� ��� � � ������� � ��� ������ �� ��� � )�� �����(�� ��� �� ��� ���� �������� .��� �� � � ��� ���� )��� ���� �� � * 5� � ��� �� �� � ���� ��� ���� ������ � � �� � � �) �� ��� ��� �� ��� �(��� ��� � �c�� � �� ������ ��� ������ � ���� � ������� 63� ��� �&&&9 �� ����� �� )���� � ���� ���� 6d��� ����(��� e���� 1������� �&&&9* +� ��� �� ��� � �� �� 33�0 ���(��� )��� �������� ) �� � ����� �� ������� �� ������ ��* 1� � ������ �� )��� ��� �� )�� ������ ������� � ��� �� ��� � � � �� � �� ������� �� � � �� �� ��� � � ������ ��� ���(��� �� ��� � ������ �� ����� ��� �� �� �� �� � �����* ���� ��� . ����� ���)�� �� �� ��� �������� ������ �� ) �� � ������ � ����� �� ����� ������� �� ��� �� � ��� � � �� ��� ��� �� ���������� ��� � � � �������� �� ���������� �� (����� ����� ��������� ��� ������ � � ��� �(��� ��� ������ ��� ���* � ��� ���������� �������������� �� ����� 3���������� �� ��� ���(��� � ������� � ��� (� �� � ��� ������� ������� ( � �� � ��� ���� � �� �� ��� ��������� �������� ��� � ��� � ) �� �����) �� )��� �� � ��� �� � �� ��� �� )� � ���� (�� ��� �� (� �� �� * /�� � � � �� ��� ��� ��� ��)���� (��� � ����� �� "; )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ���� �� ������ � ���� ������ ��� � �c�� � �� ����� � � ��� � �������� ��� ���� ��� ��� ���(��� ���������� (�� ���� ��� ��������� �� ����������* /�� ���� �� ������� �� � ��� ����������� �� ���(��� )�� �� ��� � �� ��� � �� ��� ������� �� �� ���� ���������� � �&&4 �� )�� ��(��c������ �� ���� (� ��� 1��� � �c� �� ��� ��� ������� �� 8 ��� �� 3����� � �&&;* /�� ���� ��� ������� �� � )�� ������ � � ��� �&&;?�&&> 7� ��� 3��� � �� � ����� ��� ������� �� ��������� �� ff���� ��� � ��� �� )��� ��� �� ��� ���� �� ���� ��� ������ �� ����� � ��� � ��������� �,, 6������� �� /�������� �&&>� �* 9* /�� ������ �� ��� ������� �� ����������� �� � ����� ��) ����� @ &= ( �� �� ������ � b��� �� �� "##�### �������� ��� �* /�(�� � �������� ������� ���� �� � (� ��� ����* /���� ���� �� � ���� � � ���.�� �� ��� ����������� �� � ����� � ������� �* /�� ������� ���(�� �� ��� �� ��� �������� ���(�� �� � ����� �� ��� ������ ������� ������ �� ������ ��� ���������� (� ����������� ��� �* /���� ��� � ���� ���� ���� .�� ���(���� �� � ���(�� � � ������� �� ����� � � (�� ���� �� ��� �� ����� ���� � (��)��� ��� ���(���* /�� �� � ������� ���(�� �� ��� �� ��� ������� ���(�� �� � ����� �� ��� ������� ������ �� ������ ��� ��� � ���� � ��� �* /���� ��� � ��� ���� �� ��� �� � (�� ���� (����* +� ����� ��� � )��� ����(� ��� �� � ��� ���� (�� ��� ������ �� ���� ��� ����������� �� ��������� ����� ���� � � � �&&" )� � ���� � (� � � ��� ���� (�� ��� �� (� �� � (� ��������� �� (����� �� ���������* 3��� �� ����� ��� � ��) ���� ��(� ����� ������ ����) �� ���� �� �� � �� ���� (�� ��� ���� ���� � ��� � � ����� 6����� g ������� �&&&� �* 9* ��������� ��� �� )� � ��� ��� � ����(� ��� (� � � ���� ��������� ��� ��� �� �� ������� ��� ���� (� ���(�� �� ��� �� (� ������ ���� ���� �� � �� (� � ����� ���� ���� ������* ��(� �� ��� ��� � ��� ���� � ��� ��������� �� ��� ������ �� ����������* /��� � ���� ��� ��� �������� ���(�� �� ��� � �� ��� � ���(�� � ��� �� �� �� � � ��� ��)����� � ����� �� ������ ���� ��� ��� �� �� ������� ��� ����* 2�� ����� �� ��� ����� � �������� �� ��� ��(� �� ��� ��� � ���� (��� ��� �� @;# ( �� �� ��� ��� �� ���� ���������� ����� 6����� g ������� �&&&� �* 9* +� ��������� � �� � )��� ���� ���� ��� �� ��� ������� ����� (� ������* �� �� �� ��� ��� �� �(� ��� �� �� ��������� �� ����� � �� � �� ��� � �� )� � ���� (�� ��� ��� �� �� � ��� ��� � ���������� ���� (��� � ���� �� ��� �� �� �� ����� �� ��� � ���(��� (� ����� �� � ����� �� �������� �� ��������� �������� ��� ���* /�� ���� �� ��� /�(�� � ������� �� ����������� �� ��� � '��(�� �� ��� � 6����9 ���(�� � ����� 6���###9 6����9 ������ 6@ ( �� ��9 6�����9 ��������� % ## 6"9 �=# 6 9 ;> 6%9 +� ����� � 6 9 4>;= 6"9 " 6=9 ��(� �� ��� & 6=9 ";;4 6%9 �#; 6"9 ���� � "&" 6%9 ���4> 6�9 � & 6�9 3���������� "#=�> 6�9 %% 6=9 ;= 6 9 ���� � ��� � �� ��. � �������� < < % /���� "#&�� "";=4 & 3��� �� ������� �� ��� ��� �� ������� �� ������ �� 6"##�9* � ���� ��� �� ���� � ��� � �� ��. � ��������* )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ";= .����� �� ��� ����� �� �� � ������ (� ��� � ��� ��� �&&4<�&&; ���� ��� +������ � �� 3���������� �� ���� �� �� 6�&&>9* /�� �����) �� ������� � ����� ���� ��� � .� ���* � ����� ���� �� ���� ����� ��� ����� ) �� =;h �� ���� � ��� � �� ���� ��h �� ���� ���� � ��� � ����� �� �� �* ����� � ���� � h �� ����������� ��� � ����� ���� �� �* 3������ ��� � 6(� ������9 ��� ���� � ���� �� ����� �� �� �� ��� � ����� � � � ����� �� ��� ���� �� ��� � � ���� � � ��� �� ����� �� �� ������ ��� �* '�)�� ��� � ��� ���� ���� � ���� �� ����� �� � ����� �� �� �� �� ��� � )���� ������� )��� ����(� ��� �� �� �� �� � (� �� �� �� ����* 1� ��� ��� ������� �� �� ����� ��� � ����� �� �� � � ����� ���� ��)� =�h �� � ����� ���� � ��� �� ��������� �� �* ������� ��� ���� � �� . ��� �� ��� ������� ���� �� � � ���� ; h �� ��������� ���� (�� ��� ��� (� �� �� � �� ��� � ����� �� �� � 6+������ � �� 3���������� �� ���� �� ��� �&&>9* /� � �� ��� ���� )�� �� � � (� �� �� � � ���� ���� ��� ��(��* +� � ����� �� � ���� ��� i�� ��� ���� � ��� ����� ��� �� ��� �� ���� +��� ���� ��� i� ��� ���� � ��� ����� (���.� ���� ����� �������� ff��� ���� �� �.�� (���.� � ����� � ������� � ��� (��� ����)��� �-�������� ,, 6����� g ������� �&&&(� �* � 9* 1� ��� +��� ������ ���� ��� ��b�� ����� �� ������ )��� ������ ) �� ���� 578�� � )�� ���� � �&&> ���� ��� ������ �� +��� ��� � ��� � ) �� ���� 578 ��������* /�(�� " ��� .�� ��� ������ � -���� �� +�� ����� 578 ���� �� ��(� ��� �* 7��� �� ��� ����� �� ��� �, ������� � ������� ������� �� �� ��(� �� ��� � ����� ��� 578 ���� 6&h9 �� ���� ���� 578 �������� 6>;h9 ������ �� 4h �� �� ���� �� ��� ��� � )� � ��� 578 ���� �� "�h )� � ���� ���� 578 �������� 6������� �� ��� ��� �� ������� �� ������ ��� "##�9* 8������ � ��� ����� � � �( � � �� �� ��(� �� ��� ��� � )��� � �� � ��� 578� )��� � ���� ��� � �� � ������ ������� �� �� ����� ������ 6����� g ������� �&&&(9* � ��� ���������� �� �������������� ������ /�� ��� � �� ������ ��� �� �� )� � ����������� �� �� (� �� ��� ��� �� � �� � � ��-�� �� ������� ) �� �� c�� � ��� �� ��������� �� ��c� �������� ��� ����� �� � ���) �� � � �� ������� (�.�� �� �� � �� . �� � (��� �� �� � � ��� �� �� ��� ���� �����* /�� .��� � � �������� ����� �� �� ������� �� ������ ��� � ��� �� ��� �� ����������� �� � �� �� ������ ��������� � 67 � � �� �� g � ����� �&>%9 ��� ���� (�� ���� ���� � � ��������� � 6j��� 7��)�� g ��� �� �� �&&=9 �� ��� ����c��� � �� ��� ��������� � � /�(�� " 7���.� ���� ����� '��(�� �� ��� � ���(��� 6���###9 ������ 6@ ( �� ��9� � ������ ��( ">%" �&"%� "#; 5�.�� (���.� ��� > %"# "# e�(� >" ";;" �=# /���� %;&; ""%"% %;; 3��� �� ������� �� ��� ��� �� ������� �� ������ �� 6"##�9* � 2�� �� ���� � �� �� � �� ��. � ��������* ( 2�� �� ����������� ��� �* ";4 )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ����� �� ��� k�) �� ������� ��� � 63�) � � �&&;9* e�)����� ��� � � �� ��� �� ��� ��� �� � � ���� )��� ����� �� � ������ ��� � ���� �� ������� ���-����� * /���� ��� �� � ���� �� ���(�� �� ��� �� ��� ���� �� �� �* ������� �� 3�� ��� �� 6�&&"9 �-�� �� ��� ���� �)�� ���� �.�� (���.� ����� 657��9 �� �.�� ���� (�� �� ����� 65��9� )� � ��� �c� ������ �� ��� +��� ������� �-�� ��� �� �� ���� �� ���� �� �� �* /��� �������� ��� � �) ���� ��� �� ����� �� �� � ����� ������� ��� ��� ��� �� ���� ������ ��� ��� ���� �� � ������ �� � �����* /��� ����� ���� �� � ���� � �� ���� �� �� ���������� � �� �� ��)�� � ����� �� )�� ���� � ���� ���� �� ��)�� ������� ��� �� 57�� �� �� � ������ � i� �� �� �� � ���* � �� ��� � ����� ������ (� ��� ������� �� 3���������� �� 8�� � ���� �� �� 6����� g ������� �&&&�9 ��.��� ���� ��� ���� ���� 578 �� +�� � ���� ���� ���� � ������� � �������� ���� � ��� ����� �� �� � �� �� ��� ������� ��� �� �* ���� �� ����� ����� �� � ��� 6�&&&9 ���� �-�� �� ��� ���� �� ���� �� � �� ��������� �� '���� ����� �� 3���� 0� ���� �� �� �� � � � ����� �� �� �� � ��� ��� �� �� �� � �� �������� �� �� �� ��������� ������ �� ����� �� �-�� �� ��� � � �� ��� ���* /��� ������� � ����� �� ������ �� ��� �(��� ���� ��� �k��� � �������� � ��� � 57� �� 5��* /�� � ������� ������� � � �� . ��� ��� � �� ����� ���� ����� ���� � (��)��� ��� �� �� � �� ��� 57� �� �� ���( � ��� ��� ��* /��� ���� .� ������� ��� � ���� ��)�� � 5�� ���� � �� �� � ��� �� � ��� ����� )�� )��� ���� � ���� �� ����� ��� ���� �� � ���� �� ����� ��� 57�* 5���(���� ������� g 3�� 6"###9 �� � �� � ��� � �� ���(��� �� � 0*3* ���������� ���� �� ���� ���� �-����� ��� �� ���� �k��� �� ��� �� � (��)��� � 578 �� �� +��* 8 � ��� � ��� ��� �� ���� ��� ���� �� �� � ��� �������� �� ��� ���� ��������* ��� �� ��� �� ���� ��� ��� .� �� ��� ����� ����� �� ����� �������� �� �� � �� � ���� �� (� � � �� . ��� ��� ��� � �� �)� � ���* /���� ���� �� �� +�� ��� ���� �� ����� ��� �������� �� �� ���� �� (��)��� ��������� ��� ��� �� �� ������� �� ��� ����� �� ��� � (��)��� ����� ��� ���* /�� ������� ������� ���(��� ����� (� ������ � �� � ����� ���� �������� ������ ���� � ���� � 578 �� �� +��* 1� �(����� 6�&&&9 ���� ���� ��������� � ��� � )� � ����) ��� ���(�� ���� ��������� �� � )��� ���� � ���� �� ��� ������ ���� � ��� � ���� �� ���� ���� ����� � ��� � ) ����� ��������� �� �* /� � �������� ���� �� � ��� ���� � ���� �� ���� � .��� �� � � �� ��* �������� �������� g 7��)� 6"###9 �� � �� � ������ �� ��� ���� ������� �� �� ��� ������� ��� � (�� � � �� �� � � �� ����� � �)� �� ��� �� � ���� �������� ������� �� �� ���� � ��* /�� � � � )�� �� ������ )������ ������ .��� �� ��� � �� �� �� �k��� � �� ��� �� � (��)��� 578 �� +��* /�� .� ��� ������� ���� ����� � �� � �� �k��� �� ) �� � � �� � � ��� � �� ��� �� � ���� �� �� .��� � ���� � ���� �� ����� ��� +�� ���� ����� � ��� ���� �� � �� ��* /�� ������ ��� ������� ���� ��� � .��� �� ���. �� � ��)�� ���� ��� ������* �� ���� ������� ��� ���� �� ��� +�� ������� �� (� ��� (�� �� ���� (���.�� )��� (� ������� �� � ��� ��� �� �� ��� �� ���� ������ ���� ��� ����� �� (���.��* ������� ��� ���� � �� ) �� ��� 578 )��� �� �� ��� �� ���� ��� �� � �� �� �� ��� � ��* 2��� �� )��� �� ��� ������� ��� ��� 6������� g ������������� "###9 �������� ��� .� �� 6������� �� ��*� "###9 ���� ����� )�� ��� ��� ��� � � �� (��)��� 578 �� +��� � �. ��� ��� �� ���� �� ����� )�� �� � � �� ���)�� �� �� ����������� �� )��� ���� � ���� �� ���� ) �� ��� 578* 8������ ������ � �� ��� ��� � ��� � �� ��� )��� ����� ����� )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ";; ������ ���� ���� �� � ����� ��� � ���)�� ��* ������� �� ������������ 6"###9 ���� ���� ���� ����� ) �� � ������ � �� �� ��� ������ )��� ���� � ���� �� ����� ��� +��� ��)����� �� � ����� )�� ���� ��������������� � ��� �� �� ��� � � �� (��� �. ��� �� �� ��������* � ��� ���������� �� ��� �� �� ���������� �������� ���� � �������� �� � � ���� ���� ���� � ���� �� ����� �� .��� �� ����� �� �� ��� ��� � ��������� �����* ��� �� ������ �� � ��� �� ����� ����� �� � � ��� ���� (��� � ���)�� �� �� ��� ��� � � ���� 6:����� �&&>9 �� � � �� ����� 65)���� � ������� g : ��� "###$ 2� �� g �� ����� "###$ ��� �� �� g ��� �� ��� �&&;$ 3�� �� g 3������� �&&>9* 3�� �� �� 3������ 6�&&>9 ��� 0*3* ��� �(�� �� ���� ��� ���� �� � �� � ��� �� ��� �&&" �� �&&= ������� �� ������� .��� �� �� � �� (� ��� 8� ���� ������� 7��� * /��� ����� ���� ��� ���� � �� � ����� �� ��� �(��� �� ��� ���(�( � �� �� ���� �� � ���� ���� +�� ����� ���� �� ��* /�� � ������� � ��� ���� )� ��� ����� ��� ��� �� ������ ��� � ��( ��� �� �� ��� �� �� ��� ��� ������ � ���� ��������* +� ��� � ������� � ���� )���� �� ���� � ��� ��� ���� � ���� �� ����� ������ ��� ��* ���� � ��� ��� ���� � ���� ���� � ���� )���� �� ����� ������ (�� �* +� ����� �� ������� ���� ��� �� � +��� ��� ������ ������� ���� )����� �� � ���� ���� ���� � )����� ��� ���� � ���� �� ���� ��� ����* ��� �� �� �� ��� �� �� 6�&&;9 ����� ���� ��� �� ������ �� � ��� � �� �� ���� ���)�� �� �� .��� �� ���������� ������� � ������ �� =; �� ���� �� ��� ����* /�� ���� �� ���� ���)�� �� ���� �� �� � � � �� � c���� �� .��� �� ����* ��� ������ ���) �� ���� ���� )���� �� ������� (����� �� ��� .��� �� ����* /�� ������� ��� �� ���� ��� ������� ����� (� ���b� �� �� ������ ������� ���* 8 � ��� �� ��� �� � �� � �� �� ��� �� ���� ��� (��� �� ����* 7�b����� �� 7�������� g ! ���������� 6�&&&9 �� j��5���� �� l���� 6"###9 .� ���� )���� ���) ������� � �� ����� �� � ��� ���� �� �� �� ��� � �� ���� �� ������* /� � � � �� ) �� ������� �� �� �� � 6m� �� ���� �&&;9 �� �������� ���� )���� ��� (� ���� � ��������� ���� ��� )��� ����� �� � .��� �� ������* 5)��� �� ��* 6"###9 .� ���� )���� ��� ���� � �� ������ )��� ����� �� � ������ ��� � (�� ���� ��� ����� �� � �� ����� �� ����� ) �� � ����� .��� �� � � �� ��* e�)�����3 ��(���� 7��)�� ������� g 7�� � ���� 6�&&&9 .� ���� )���� ��� ��� ���� � ��������� ���� ��� )��� .��� �� � � ��� ��� ��� � ����-�* � ��� ���� ������ ���(���� � �� 33�0 � ��������� ��� ��������� �� ����� � ������ �� ������� ���� " ����� ����� � ������� �� ���� ��� ��� ��� ���* 0� 3���� ���������� ��� :� * 60� 3����9 � ��� � � � ������� �� 33�0� )� � � ���� � � ����� �)�� ��(� ��� �� 0� 3���� :� *� ��� ������� �� 33�0* ���(��� �� 33�0 ��� ��c� �� �� ���� (��� ;h �� ������ �� ��� � �������� ���� (���� � h* �� �� �� !��� �&&>� ��� ���(��� �� (���.�� ���� � (� �� �� � 578* 7��)��� !��� �&&> �� !��� �&&&� ��� �� ���(��� )��� � ��� ��� �� � �� ��� �� �� ) �� ��� 578 �� ��� �� �� ��� �� ���� ��������� ������� �� � �� +��* 7��� ";> )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ����� ���� � ����� �� (���.�� � �� ��� ��� ������� ��� ���� ��� ����� ��(������ �� ��������� � ��� ��* /� ����� ���� ��� � � �� ��� ��� �� 33�0 ���(���� � �� ����� c���� ���� �� ������ )�� ������ �� * /�� c���� ���� �� ���� �� ���� �� � ���* /�� .��� �� � �� �-�� ��� )��� � � �� )�� �� �* /�� �� �� �-�� ��� ��� ������ �� ��� (� ���(��� � ��� �� ��� � � � �� �� ������ �� �k��� ��* /�� �� � �-�� ��� ��� ���(��,� ���������� �� ��� � � � �� �� ��� � � �� ��� �� ��� ��� �� ��� .��� �� � �� ����� ���(�� �������� �* 3����� ��� (�� �� )�� � � � ����� (� 0� 3����* � ������ �� �#�### ���(��� )�� ��� ���� �������� ���� ��� ���� ���� >�### ���(��� � ��� ��� * � ���� ������ )�� � ���� � 3�����(�� �&&& �� ��� .��� ������ )�� ���� � l ��(�� �&&&* � ����� �� "� #; ������� )��� ������� )� � ���������� � ���� ��� �������� ���� �� ���� " h* � ����� �� "�%&& ����(�� ������� )��� �����i� * /�� �� � ��� � � �� ���� ���� �� ���(��� (� 33�0 ����� �� � 4>* h �������� ����* /� � ����� �� ���� �� ��� �� �� �� � �� �� ��� c���� ���� ��� � ���� ���� �� ����� ������ �� ������ ����, ���������� �� ��� � �)� ����� �� ���)�� �� �� ����������� �� �������� ��� � ��� �� ������� �� ������ �� �� )�� ���� ���� �� ������ � ��� �� ��� � � � ��* /���� ����� ��� ��� ��� � ��� � ��� �� ��� ��� �������� �� ���� �� � ���(�� �� ������� ��� �� ��* 8 ���� � �������(���� �� ������ � �� ���(�� �������� � �� ������� ��� �� �� ��� ��� ������ � ����� �� ������ ������� � �������� * /� � ��� .�� ���� �� �� �� ����������� �� �� ��� ��� �������� ����� ��� �� ���� �� �� ��c���� ����� �� �-�� �� �����* 3� �� ��� � ��� �� �� ������� ��� � ��� �� ������� �� ������� �-����� ��� .� ����� ���� ��* � ��������� �� ������� �*�* ������ �� �� � ������ ���� )��� ���� �� � ��� )������ ���� ���� ��� +�� �� ��� 578 �� �� �� �� ������ ��� ���� �� ��� � ��� ���� � ���� � ��* 8 ����� ���� � ���� 33�0 ���) ���� ������� %�*4h �� ���(��� ��� ������ ��� ����� (� ������ ���� )��� ���� ��� ��� �� ��� 578� %%*%h ���� �� ���� �� �� +�� �� %=*�h ���� �� ���� � ) �� ��� 578* /� � ����� ����� �� ����� )�� ���� � � � ��� &h ���� �� +�� �� =�h ��� 578* +� ��� ������ ������� �*�h ���� +��� =*"h 578 �� �%*%h �� �� �� ������ ��� ����* /���� )�� ��� ������ ��� ���� ��� �� �� ���������� � ��� ������ ������ ������� �� (�� � � � ��� � ���� ���)� ���� ��� �� � � ���� ��� � �� ������ ����� �(��� ����������� �� �� � ����� ��� ����� �� ��* ����� ����� ���� �� ��� +�� ������ ����� �� �� = �� ����� �-�� � �� ���� ���� �= ����� �� ��� ������ ��� ��������������� ������ �� ��� )���� ������ ������* ����� ����� ���� �� ��� 578 �������� ����� ���� = �� ����� �-�� � �� ���� ���� �# ����� �� ��� ������ ��� ��������������� * ������� �� �������� � ��� ��� � -� �� ��� �-���� �� ���� �� �� �� ����������� �� � �������* �� ��c���� ����� � ��� ���� ����� �� �� �� ����� ����������� ��� ��� � �� . ��� ��� ���� �-�� �� � �� �� ��� ������ �� ��� �� (�� ��� ��� �������� � ���* )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ";& �* * & �����(� �� ����� � ��� ���(��� )��� ���� �� ������ ��� � �)� ����� �� ���)�� �� �� ����������� �� ������� �� ��� � �� ��� � � �� )�� ��� �� ����* 8��� ������ )��� ��� .� � ff����,,� ff� ����,,� ff�������,, �� ff��� ,,* 8����� ������ ���� ��� � �� . ����� ��������������� �� ��� ��)�� ������ �� ff����,, �� ff� ����,, )� �� ����� ��� ��������������� � ��� ff�������,, �� ff��� ,, ������* ������� �� �� ��c���� ����� ������� ��� � �� . �� � �� ����� ������ ��* �� ��c���� ����� ���� � ��� � �� . ��� ������ �� � ����� �� ���)�� �� (� ���* /�� �)� ������� ������ 6�><% �� %=< 9 ��� ��������������� � ��� ff����,, �� ff� ����,, ������ �� �� �� ���������� � ��� ff�������,, �� ff��� ,, ������* /� � � ������� ��� ���(��� � ��� �)� �� �� ��� ������ 6 =<= �� ==<4 9* /���� ������� ��� ��(��� ) �� � ��� �� ��� �� �� � ���� (��� � � �� �� ����� � �� 6�������9 ����� �����* /� �-����� ����� ������ ��� ������� � ���� ��� �� ������� �� �� ����� �� ���)�� �� ��� ��� ��� �������� � �� ������� ��� �� �� � ������� * ��� � �-�� �� �� ���� � ��� � �� ����� ���� � ) �� ���)�� �� �� ��� ��� ��� ������� �� ���)� ����� �� ��� ������* +� ��� ��� (��� � �� � ) �� �� �-�� ��� �� �� � ��� � �� ����� ���� � �� �� �� � ����� �( � �� �� ���� �� ��������� ������� �* 3���� ���� ��� (��� � �� � �� ������ ��� ������ �� (��)��� � � �� �� ����� � �� �����* ��� ��� ������ ��� (��� � �� � �� ���� ��� )��� (� ����� �� � ��� 6�&&&9 ��� .�� �� �k��� � � ��� ������ ����� ��* ���� �������� �� � �� (� �-������ (� ��� ���(�( � ������)�� ���� ���(�( � ������)�� ���� � � � �# � �� ��� ��� � �" ���" � �% �����% � � ���� � � �= � ���= 6�9 )���� ���)�� ��� ���������� ��� ����� ���)�� �� ������� ���)�� ��� ��� ff��� ,, ���)�� �� ����� �� � ��� ���. ��� ��� ���)�� �� ����� � ��� �� � � ���� ��� ���������* /�� ������� ��� ���� �� � /�(�� %* � ��� �� ���� � � ������� � ���� �� � ��)�� ����� �� ���)�� �� )��� ������ ) �� ��� � ����� 6ff��� ,,9 ����� �� ���)�� ��* 8������ ��� � �� . ����� ���� � ���� ���� ����� �� ��� �� ���������� �� ���� � ��)�� ����� �� ���)�� ��* ��� � ���� � � �� . ��� �������� ) �� ��� �)� �������� ������ ���� � ���� ���� ��� �� ��� ����� �� � ��� � ��) ����� �� ���)�� ��* /� � � ���� ��� ��� ������� ��� ������ ��� ��� � ���� ���)�� �� ����� (�� � � -� ��� ��� ������� �����* /�� ��)��� � ��� ����� � � �� . ����� ���� � ���� �� ������� � ��)�� ���)�� �� ����� ���� ��� � ����� � ��� ����� ��� �� � ��)�� ���)�� �� �����* ������� 6����� � �� 9 ����� ���(��� ��� ���� � ���� ���� � � �� ����� �� � ��� ��� �)� ��)��� ���)�� �� ������ ������ ���� ��� � ����� �����* l������ ��� �� ��c���� ����� �� ������� ��� ����( ������� ������� � �� . ��� ������ �� � ������������ ���)�� �� ����� (� ��� �� �� ����� ����* ��� �� � ��� ������ �� ��� ���� � �� . ��� ��� ������� �� ��)�� � ��� ������ ��� ����� � ����* +� (��� ���� ��� � � ��� � ����� ���� ��� ���)�� �� ���������� � �� ��� �� ��� ��)�� ������* /�� ��-� c���� �� ���� � )��� ������ �� ���(��� � ��� �� � ���� ���������� ��� �� � ������ )�� ���(��� ����� �� ������ � �� �� �� ��� �� ���� � ��� �� ��� � � � ��n "># )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� �*2* ,����� �� ������ � ��������� ���� 33�0 �� � � ����� �� ����� ��� ��� ��(�� �� ��� �� ���(��� � ��� �� ��� � � � ��* /���� � �� � � �� ���� �� ������� �� �� �� �� ��� ���(���� � )�(� �� ���� � �� �� � �� ��� ���� ������� �� �� )��� �� � �� �� �� ������� )� � ���(�� ���(��� �� �� �� /�(�� % ���� ��� �� ������� �� �� ����� �� ���)�� �� o��)�� �� �����p(��� � ��� 3*2* 1�� *�* 3 �� . ��� ������ � 2-�6 9 '��� +���� ��� �"*&"% #* = �*="> � #*### 8����� �* ;4 #*"#> =#*%#& � #*### *%;4 �><% ����� �*=#> #*%=" �>*%#4 � #*### *=�> %=< ����� �*="� #*%# "=*#�4 � #*### *=;4 =<= ����� #* > #*"&; "*";" � #*�%" �*=44 7��� q = ����� ������� #*>�# #*"�& �%*4 = � #*### "*" ; ���� � #*##; #*"=% #*##� � #*&;4 �*##> r@4#### � ��� �*� % #*%## � *= % � #*### %*�%; @4####<@�##### #*%�" #*" ; �*=& � #*"#; �*%44 7��� q @�##### : ���� +���� ��� ��*%%; #*%"% �;*�"& � #*### 8����� �*"&& #*�== ;#*">" � #*### %*44 �><% ����� �*"&" #*"4 "%*&>& � #*### %*4 # %=< ����� �* &4 #*"� >*;4= � #*### * 4" =<= ����� #*4& #*�&% �"*&"= � #*### "*##� 7��� q = ����� ������� #*=&; #*�4= �%*#== � #*### �*>�4 ���� � �#*#=" #*�&4 #*#;� � #*;>& #*& & r@4#### � ��� �*#%& #*""; "�*#�& � #*### "*>"= @4####<@�##### #*=%= #*�;= &*%�" � #*##" �*;#; 7��� q @�##### ������� +���� ��� #*%� #*"&= �*�%= � #*">; 8����� #*=4" #*� & � *� " � #*### �*;= �><% ����� #*�&# #*"=% #*=4= � #* =" �*"#& %=< ����� #*4 # #*�& �#*>4 � #*##� �*>&; =<= ����� #*� = #*�4> #*;=" � #*%>4 �*�=4 7��� q = ����� ������� #*"4> #*�=> "*>>4 � #*#>& �*%#> ���� � �#*#>� #*�&# #*�># � #*4;� #*&"" r@4#### � ��� #*=;4 #*"�= ;*"�% � #*##; �*;;& @4####<@�##### #*"�# #*�4" �*44> � #*�&; �*"%% 7��� q @�##### 8 ��� �� ��� " ��� � ��� ��� >%"*;4&� �� ��c���� "=;*;=#� *�* 6" 9� ������� #*###� '��������� /" � #��"�* )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ">� ��� � ���� (�� ��� ������ ��� � �� � ���* 3�� ���� )��� ��� �� ��� �� ���� �� ������� � ���� ������� � �� ���(��� ���� �� � ��� �� ��� � �� �� 33�0 ����������� ��* /�� ��� ���� ���(��� �� � �� ����� ������ �� � ��� �� ��� � � � �� � � ��� � /�(�� * /�� ��b�� �� �� ���(��� �� � ��� �� ��� �� �� ������� �� �� � �� ����������� ���� ����� � ��� ��� ���� �������� �� �����* /�� 33�0 )�(� �� )�� ��� (� ��� � ����� �� ��� ������� �� �� �� ������� ��� . ���� (� ��� � ����* /�� ��� �� ����� ������ �� � �-����� ��� ��� �� ������ �� ) �� �� ��c���� ����� ��� �� ���� ��� � �� . ��� ������ �� � ��� �� ����� �� �� � ��� ��(�� ������ �* ������� ���) ���� )�(� �� ����� � � �� . ����� ������� (� ��� ��� ) �� ����� ��������������� � � ��� �� ��� )�(�(��� ������ ��* /�� c���� �� �� )������ �� � � � ��� �� ���� � ��� �� �� )������ ����� ���������� �������� � ��� ���� (����� ����� (� �-�� �� ������� �� ����� � � � �� . ��� �� �� 6����9 ����������� �� �� ������� 6�����9 � � ���� ����� ������� �� � � �� ����� ��� � ���* 2-�� �� ����� )��� �������� ��� ������ � ����� �� ��� � �� ������ (� ��� ��* 1� �� ������� ��� �� �� ���������� � �� �� ��� )�(� ��� ����� ������� ����� �� � ��� (� � ����� ��� (� ��� � ���� ����������� �� � ��� ��)�� ���������� ������* +� ������� �� �� �� ���� ������� ����� � ��)�� ���������� ������ ��� ���� ���� �� (� �� �� ���������� � )�(� �� �����* /���� � �� ���� ����������� �� � ����� ��� � ���� ���������� ������ ��� (��� ����� �� �������* /���� ��� �� �� � � �� . ��� ������ �� ����� ���������� ������ ) �� � � � �� �����* �� ������ � �� )� � 33�0 ������ � ������ ���� ��� ��� �� (� ��� ���� �������� � � ��� � /�(�� * /�� ������� �� �� � ���� �� ���(��� )�� ��� ���� �������� ��� ��� ������� ���(��� �����)� (� ��� ��� ���� �������� (� 33�0* � �� ��c���� ���� �� ��� ���� �������� ������ � (� ��� �� � ���� � � �� . ��� ������ �* 8������ 6�����9 ��� �� �� 6����9 ���������� � � �� �� ��� ������� �� �� � �� ��� ���� �������� ������ � �� ���� 6�� ��9 ���������� � ���� �� ��� ��� ���� ��� �� ����� ����* +����. ��� ����� ������� �� ��c���� ����� �� � �� . ��� ������ �� ������� ������� �� � � �� ����� ������ �� )� � )�� ��� ���� �������� ������ � ��� * �*5* ,����� �� ������ ������ ���� )��� ���� )�� ���� �� ������� � ��� �� ��� � � � ��* /�(�� = � ���� ���� ��� ������� ���(�� ������� )��� ��������� �� ��� � �������* /�� � �� . �� � /�(�� 33�0 ������ �� ��� � ��� �� ��� � � �� 6h9 ������ � 0�� h 6����9 ���� �������� ��� h 6����9 +������� �� �� � �� �� �� ���(��� >;*> 6�9 =%*# 6�9 1�(� �� %"*# 6%9 %*= 6=9 �� �� �� ������� �� )�(� �� " *4 6 9 ��*� 6%9 3�� ���� ��� (� 33�0 �� ����� ;*# 6"9 ";*# 6"9 :� �� 33�0 ����������� �� ��*% 6=9 *" 6 9 '� 33�0 ������ �� ��� 4*# 649 �*# 649 /���� �##*# ">" )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� �� )��� ��������� � �� � ���� �� � � �� ��� �� ��� ���� (��� ��� .� (� 5�k� �� 3��i 6"###9 )�� ������� ����� � �� ���� �� � (��� ��* l��� "4h ������� � ������� ���� �� � �� � .��� �� ������� �� � �������� �� � � ������� �� �� ���� ���� "�h �� ������ ���� ��� ������ ������ � ��� � �� �� ��� � � � ��* 2� � �������� �� ����� �� �� (� �-�� �� ��� ��� �� ������ ��* 3��� �� ���� � �� . ��� ������ �� ��� ���� ��� ff��(� �,,� ff�������,, �� ff��� �� �� �� �� �,,* ����� ��� ���� � ���� �� ���� ������� ��(� � )� �� ������� ��� ���� � ���� �� ���� ������� ��� � ������� �� ��� ��?�� �� �* ��������� �� �� ff)��� ���������,,� ff.��� �� �������,, �� ff� �������,, ��� ��� � �� . ����� ������� (� ��� ��* /���� ������� ��� ���� �-�� �� (� ����� ������ ��� (��� � � �� �� ������� ������ (�� �� � �� . ��� ������ �� ������* ������ ���� )��� ���� �� � ��� )��� �� ����� ������� � ��� �� ��� � � � ��� )�� ��� ���� ��������* l������ ��� ���� �������� ������ ������� )�� ff)��� ���������,, �� ff�������,,* 8������ ��� � �� . ����� ���� � ���� �� ���� ��� � ������� �� ff��� �� �� �� �� �,, �� ���� �������� )������ ����� ��� � �� . ����� ���� � ���� �� ���� ff)��� ���������,, �� ff.��� �� �������,, �� ���� ��������* /���� ��� �� � �� . ��� ������ �� (��)��� ���������� ���� . �� �� ����� ���* /� ������� ������� �� �-����� ��� ����� ���� �� ��� .� � ��� �������(���� ��� ��� ����� ������� � ��� ��� ���� �� ���� �� � ���(�� �� �������� ��� �� ������� ���� )� � ��� �������� � /�(�� 4* +� � �-�� �� ���� ��� ���� �� ���� ���� �� �-��� � ���)�� �� ������ ) �� � ���� �k��� � ��� � ��� ��� �� ������ �� ������ �� �� �� ���� ���� ������ ��* ���� ��� ��������� � � �-�� �� ���� ���)�� �� ������ ��� ���� ������ �� ��� � ��� ��� �� ������ �� ������* 8 ������ � ��� ���(�� ��� ��� ��� � � �� �. ��� � � �-�� �� �� � ����� ��� � ��� ��� �� ������ �� ������* ���� �������� �� � �� (� �-������ ��� ��� � � ��� ��� �� ������� ��� �� ��� ���(�( � ��� ������ ��(� ���(�( � ������ ������ ��(� � � � �# � �� ��� ��� � �" ���" � �% �����% � � ���� � � �= � ���= 6"9 )���� ( ���������� ��� ������ �� �(�� �� (� ������� * /�(�� = ������ ������� � � � �� ��� ��� 6h9 ������6�9 ������� �������� h 6����9� ���� ��������( h 6����9 '�(� � ������� "�*� 6%9 ������� 4*� 6"9 %#*> 6"9 1��� ��������� >*& 6�9 %�*4 6�9 8�� ��?�� �� � �>*� 6 9 ;*= 6=9 � ������� �#*= 649 &*# 6 9 8 ��� �� ������� �=* 6=9 �;*; 6%9 l���� �*# 6;9 �*� 649 � ���� ��� ���)��� ����)� * ( 2�� �� ff��(� �,,* )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ">% /�(�� 4 :�� �� ������� ��� �� ������ ������� '�(� � 6 � =#49 1��� ��������� 6 � ��;%9 ������� ���� 6 � ==&9 8�� �� �� �� �� � 6 � % 9 ������� 6 � ��#49 3*2* 2-�6 9 3*2* 2-�6 9 3*2* 2-�6 9 3*2* 2-�6 9 3*2* 2-�6 9 o��)�� �� 6(���p��� 9 '��� #*"#" #*""& �*"" #*#&4 #*"#% �*�## ��*#"=� #*"4" #*%=& �#*=&;� #*">> #*==# �#*#;= #*""> #*&"; : ���� �#*%>=� #*�;% #*4>� #*%44� #*� & �* " �#* ;�� #*�;# #*4" �#*#>4 #*�&& #*&�> #*%"=� #*�4% �*%>= ������� �#* =&� #*�4� #*4%" #*% 4� #*� � �* �% �#*"%; #*�=4 #*;>& #*�"; #*�>& �*�%4 #* %4� #*�=" �*= ; 5 �. ���p��� �#* &# #*��= #*4�% #*=�%� #*#& �*4;# #*;%4 #*��; "*#>> #*%;� #*�"4 �* ; #*#44 #*�#4 �*#4> ��� ��p������ �#* %> #*�"# #*4 4 �#*�=&� #*#&; #*>=% #*"4 � #*��4 �*%#% #*� #*�"= �*�= #*4# � #*��� �*>"& ��� 6(��� q = 9 �><% �#*�;# #*"## #*> % #*% � #*�44 �* �# �#*;%4� #*"# #* ;& #*;>"� #*"#& "*�>= #*�"& #*�&" �*�%> %=< �#*�>& #*�4= #*>"> #*%#4� #*�%; �*%=> �#*4"�� #*�=4 #*=%; #*";4 #*�>> �*%�> #*�=� #*�=# �*�4% =<= �#*#4& #*�=; #*&%% #*";>� #*�%" �*%"# �#* �%� #*� 4 #*44" #*#;> #*�>= �*#>� #*##; #*� �*##; 3����p������� #*"�>�� #*�"; �*" % �#* ##� #*�#% #*4;# #*%%� #*�"" �*%&" #*#4# #*�%" �*#4" �#*%�=� #*��; #*;%# ���� � p��� �#*="> #*�%4 #*=&# �#*#%# #*��; #*&;� �#*�"& #*� % #*>;& �#*4;&� #*� � #*=#; "* 4�� #*�;# ��*;"� +� ��� 6(��� q @�##�###9 r@4#### #*;"" #*�> "*#=& �#*� & #*� % #*>4� �#*4>�� #*�;# #*=#4 #*">& #*�& �*%%= ��*#4"� #*�=; #*% 4 @4####<@�##### #* &&� #*�=4 �*4 ; �#*�=% #*��; #*>=> �#*%=>� #*�%# #*4&& #*"&;�� #*�4= �*% = �#* 4� #*�"4 #*4 # �������� �#*4"� #*"; #*=%; �#* =; #*""> #*4%% �#*; #*"4� #* &; ��*>�& #*%# #*�4" �"*�4" #*";" #*��= '��������� /" #*#; #*#; #*#; #*# #*# #*# #*#>% #*#>% #*#>% #*#4 #*#4 #*#4 �� �� �" 6 *�*� �������9 &4*&�% 6�"� #*###9 &4*&�% 6�"� #*###9 &4*&�% 6�"� #*###9 ;#* �% 6�"� #*###9 ;#* �% 6�"� #*###9 ;#* �% 6�"� #*###9 ��&*�>= 6�"� #*###9 ��&*�>= 6�"� #*###9 ��&*�>= 6�"� #*###9 >%*;%= 6�"� #*###9 >%*;%= 6�"� #*###9 >%*;%= 6�"� #*###9 = "*4&� 6�"� #*###9 = "*4&� 6�"� #*###9 = "*4&� 6�"� #*###9 e�����<:������) ������ 6 *�*� �������9 %*�"" 6>� #*&"49 %*�"" 6>� #*&"49 %*�"" 6>� #*&"49 * 4; 6>� #*>�%9 * 4; 6>� #*>�%9 * 4; 6>� #*>�%9 =*>% 6>� #*4449 =*>% 6>� #*4449 =*>% 6>� #*4449 >*;�# 6>� #*%4;9 >*;�# 6>� #*%4;9 >*;�# 6>� #*%4;9 >*= = 6>� #*%>"9 >*= = 6>� #*%>"9 >*= = 6>� #*%>"9 � &=h � �� . �� � �����* �� &#h � �� . �� � �����* " > ) * � � �+') � �� � �� , * ��� � .� � �� � � ������� / ����� � � 0 � � � 1 ! 2 3 4 � �*6* 7 ������ � ��(����� ��(������ � � ����� ���� �*6*�* �� ���� �8��� ������� � ��� � � �� . ��� ��� �� ���� �* 8������ 6�����9 ��� �� � ���� 6����9 � ���� �� ������ ��(� � � ��� �� ��� � � � ��* /�� � � �� ������ �� ��(� � � ����� 6 � �����9 (� 4#h ��� ������� 6�����9* /���� � �� � �� . ��� ������ � (� ��� �� ����� ����* /�� ��� � �� ����� �� ���)�� �� � ���-�� �� �� ����)��� � -� * /�� � � �� ������ �� ��(� � ��� �� � � ��� ����� ) �� � ���� � ���� �� ������� ������ �� ���)�� �� )��� ������ ) �� � ��� ����� �� ���)�� ��* 7� �������� ��� � � �� ������ �� ��(� � ��� � ����� � ��� ���(�� ������� ���������� �� ��� �� �� ���)�� �� �� ����������� ��* �������� �� � ���� � � ��� � �� . ��� � ��� � ��� ���� ������ ) �� ��� ��)��� ����� �� ���)�� �� ��� (� ��(������� (� �� � �� ��� � ���� ���� �� ���� ��� ���� �* l� � ���� ��� � �� �� ������� �� ���� ��� � � �� ������ �� ��(� � ��� � ����� ��� ����� )�� ��� ��� ��� � � �� �. ���* �*6* * �� ���� ���+ ����� (��� ������� ����� )��� ���� � ���� ���� � � �� ����� �� ������ )��� ���������* ��� � ��� ����� �� ���)�� �� ���� � � � -� ) �� �� � �� . ��� ������ � � � � ��� ����� ������ �� �� ���)�� �� ������ ) �� � ��� ����� �� ���)�� ��* e�)����� � ��� ���(�� ������� � ���� �� ������� ���)�� �� ��� � � �� ������ �� ��������� ��� � ����� )��� ������ ) �� ����� ������ �� � ��� ����� �� ���)�� ��* /�� ������� ��� ��� ����� ��� ���� � ���� � � ���� ���� ������� ���������* /�� � � ��� ���� � ����� ��� ����� )�� ��� ��� ��� � � �� �. ���* �*6*2* �� ���� ��������� � l� � ��� � )� .� � � �� . ��� ��� �� ���� � ) �� ������� ���� � ���� ���� ����� �� ������ � .��� � ������� ���� �� � �� �� � ������� �� .��� �� �������* /���� )�� ��� ��� ��� � � �� �. ��� ��� ���� � �� . ����� ���� � ���� �� ���� �� � ���� �� ��� ���� �* /�� � � �� ������ �� � ������� ���� ��� � �� . ����� �� � � � ��)����� ��� ������� ��� ������ )��� ������ �� ��� �� ��� ��� �����* /�� ���� ��)�� � ������ �� � ������� ���� ���)� ��� ������� ) �� ���* ������� ����� ��� � �� . ����� ���� � ���� �� ������ � ������� ���� ���� � � �� �����* /�� ����� �� ���)�� �� ���� � � ��� � ������� ��* /�� � � �� ������ �� � ������� ���� ��� �� � � � ��� ��)�� ��� ����� �� ���)�� ��* /� � ���� � � � �� . ��� ��� ��� ����� ���)�� �� ������ (���) ��� �������� ��� � �� . �� � ����� � ��)�� ��� ff�������,, ���� ��� ��� �)� ��)�� ������* +� ������� ���� ����� )�� ����� � ���� � �������(�� ����� �� ���)�� �� )�� ��� ���� � ���� �� ������ � ������� ����� ������� (� ���� ����� ) �� ���� ���)�� �� ��� ���)��� �� )��� ������� ���� ��� ���� � � ��� ��(�� �� )���� �� .� �* �*6*5* �� ���� � ����.���� �� /�� � � �� ������ �� ��� ��?�� �� � ��� � ����� ) �� ����� �������� �� � � ���� � �� . ��� ��� ��� �������� ��� �����* /�� � � ��� ���� � ����� � ��� � � �� )�� ��� ��� �. ���* /�� � � ��� � ����� ��� ��)�� ��� ����� �� ���)�� ��� �������� ���� )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ">= ��� ff�� ���)�� ��,, ����� ���)� � � �� . ��� ���� �* /� � )��� ������ �� (��� ��� ��� ���� �� ������ ���� ����� ) �� ��� ����� ���)�� �� ��� ���� ���� �� ��� ��� ����* +� ��� ���(�� � ���� � ��� � � ���� ���� )��� ������ ) �� ��� ��?�� �� � ��� � ����� � � ����� (� ���� ���� ff�������,, )�� ������ �� � �������� ��� � ��� c���� ���� ��* /���� � �� � �� . ��� ��� �� ���� � � �� � �������� �� �����* �*6*6* �� ���� � �� �� �� )��� (� �-�� � ��� �������� �� �� ���,� ������� � � �� . ����� ���� � ���� � ��� ���(�� � ���� � * /�� ���� ����� � �� . ��� ��� �(�� � ��� �� �� � �� ) �� ����� ������ ������ ����� ��� ���� � ���� ���� ������� �� ������* �*�* )���� ��� � ��(������ � � ���� ������ � ����� �� ������ ������ ���� )��� ���� )�� ���� ��� ��� ��� ���� �������� �� ����� ���� ������� * � ���� ��� �� ��� �� ������� �� �� (� ��� �� �-�� �� )������ � �� . ��� ������ �� ��� (��)��� ��� �������� � ��� ������ �� � ���� ���� ������ �� ��� ��� � ������� � ������ ����* /�� ������� � ������ ���� ��� � � ������� ���� 6.��� �� ������� �� � �������9 �� �� � �� (� ��� ��� � ���� f��. ��, �������� ��* /�� ���� ��� �� ���������� ���� (��� � �� � �� � �c��� �� �)� �� ��� ���� ��� �� ������� �� �� (� �-������ �� ��� ���(�( � ��� ������ ��(� ���(�( � ��� ������ ���������� ������ � � � �# � �� ��� ��� � �" ���" � �% �����% � � ���� � � �= � ���= 6%9 /�(�� ; �������� ����� ������� ) �� � �� . ��� ������ �� ��� .� * /���� ��� � �� . ��� ��� �� ������ �� (��)��� ��� ������ ����� ������� ������ ) �� � ������� ����* �� ���� ������ ����� ��� �� � ���� � ���� ���� ������� �� ������ ��(� �* e�)����� ��� ����� )�� � ������ ����� ��� ���� � ���� �� ������ )��� ��������� (����� ��� � ������� �� ���� ���� � ������� ����* 8������ �� ��� ����� ��� ��� ���� � ���� �� ������ ��� � ������� �� ���� � ������� ���� (����� ������ �� )��� ��������� �� ��� ��? �� �� �* /�� ������� � ��� � �� �������� �� ) �� � ������� � � ����� � ��� � � �� � ��� ��� �. ��� (�� �� � ��� ��� ���� �� ��� � �� ��� � ��� ��� �� ������ �� )��� ��������� �� ��� ��?�� �� �* ����� � ���� ����� ������� ������� � ���� ��� � �� ������ ������� ����� ���� � ��� � � �� � ��� ��� �. ���* /� � ���� ��� � � ���� ������� ���� � ��� ������ ��� ��� �� ���������� � ��� ff����,, �� ff� ����,, ������ �� ���)�� �� �� ����������� ��* /�� � � �� ��� �� �� �������� �� ) �� � ������� �� )��� ��������� ���� �������� �� ��� ��� � ������� ���� ��� � �� . ����� � ����� � ��� ���(�� ��� ��� ���������� �� ���� �� ���)�� ��* :���� �� ���)�� �� � ��� � � �� . ��� �� ��� � ��� ������� � ���� �� �� �������� �� �� ��� ��?�� �� �* /�� ������� � �� ������ �� � ������� ��� ��� � ������� ���� � ��� � �� . ����� ������� (� � ���* e�)����� ��� ������ �� )��� ��������� �� ��� �� �� �� �� � ��� � � ��� � ����� ��� ��� ��) �� � �� � ��� ������� ������ �� ��� � ����� � ��� �����* 3���� ���� ���� ��� � � �� . ��� ���� � �� ��� ������� � ���� �� �� �������� �� ) �� )��� ��������� �� ��� ��� � ������� ����� ��� � � �� ��� �� �� )��� ��������� �� (� �� ">4 )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� /�(�� ; ���� �k���� �� ������ ������� ���� ��� �� ������� �� ������ ������� 6(��� � ������� ����9 3*2* 1�� � *�* 3 �� . ��� ������ � 2-�6 9 ������� +���� ��� ��*# 4 #*"&% �"*;4# #*### 8����� #*4#; #*�;> ��*4"> #*##� �*>% ���� � "* 4 #*%%= = *�&& #*### ��*;=4 �><% ����� #*44& #*%�% *=;= #*#%" �*&=% %=< ����� #*==� #*" =*#&> #*#" �*;% =<= ����� #*="; #*"%" =*�4" #*#"% �*4& l��� = ����� ������� ����� #*%%% #*�>= %*"%" #*#;" �*%&= r@4#### � ��� �#*# % #*"= #*#"& #*>4 #*&=> @4####<@�##### �#*�= #*�>; #*4;= #* �� #*>=; q@�##### '� ���)�� �� #*>#% #*%># * ;# #*#% "*"%" : ���� ���)�� �� #*%= #*"4% �*>#" #*�;& �* " ������� ���)�� �� #*%=% #*" � "*� ; #*� % �* "% ��� ���)�� �� '�� �. ��� #*=&& #*�;# �"*%;� #*### �*>"� 1��� ��������� +���� ��� ��*%>4 #*%#& "#*�># #*### 8����� �#*;%= #*�;% �>*#;% #*### #* ># ���� � #*#�> #*"#; #*##; #*&%� �*#�> �><% ����� #*;#= #*"& =*;=> #*#�4 "*#"= %=< ����� #*= % #*"%= =*%=& #*#"� �*;"" =<= ����� #* #& #*""; %*" = #*#;" �*=#4 l��� = ����� ������� ����� #* " #*�;& 4*#&" #*#� �*==4 r@4# ### � ��� #*&4; #*"= � * ># #*### "*4"& @4####<@�##### #*=% #*"# 4*> # #*##& �*;#= q@�##### '� ���)�� �� #*&�4 #*%>= =*4= #*#�; "* &> : ���� ���)�� �� #*>&> #*";# ��*#;& #*##� "* = ������� ���)�� �� #*4; #*"=% ;*#;; #*##> �*&4� ��� ���)�� �� '�� �. ��� #*#4" #*�;# #*�%" #*;�; �*#4 8�� ��?�� �� � +���� ��� �%*%4= #*=>; %"*>"" #*### 8����� �#*=�& #*"4 %*>;> #*# & #*=&= ���� � �#*;#� #*"& =*4&% #*#�; #* &4 �><% ����� �*44# #* &4 ��*"�% #*##� =*"=& %=< ����� �*�&= #* => 4*>�= #*##& %*%#% =<= ����� #*>% #* =; %*%%> #*#4> "*%#% l��� = ����� ������� ����� #*%"# #*"; �*%4& #*" " �*%;> r@4### � ��� �*#># #* % =*&= #*#�= "*& = @4####<@�##### #*>;� #*%>> =*#= #*#"= "*%&# q@�##### )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ">; ���� �������� � �� . ����� � ����� � ��� ���(�� � � ��� �������� ������ ���� � � �� �����* ! "��������� ��� �� ��� ����� ������ ) �� � ����� ��� ������ ��� ��� �� � �� �� ������� �� )������* /�� ����������� �� � ����� ��� ��� ������� �-������ ��� � �) ���� ��� �� � ���� ���� (� � ����� � (� � ��b�� � � �� �� ���� �� �� ������� ������ ��� ��� �� �� ���� ���� ���� 65������� �&&&$ ��������� �&&&9* /�� ������� �� �� � ��� � ������� ���� ���� ����� )� � ��c� �� ������ ��* +� � � �� �� �� ����� ������� � �� (� ��� ���� � ���� �������� ��� �� ������ �� � )�� � � ���� ��� � ���� �� ������ �� )��� ������ �� ���� ��� � ���� �� ��� ��� �� (� (���� � � � )��� ����� �� ������� �� ���� ���* ���� ��� ������ � � �� �� ������� � �� ����� �� � ��� ������ ��� ��� � � ��� ��� ������� �� ��������� �� ��� ������ �� .��� �� �� ������� )���� ��� �� � � ��� ������ 67�b����� � �� ��*� �&&&$ 3�� �� g 3������� �&&>$ d��� ����(��� e���� 1������� �&&&9 � ��� ���� ��� ��� ��� � � �� �� �� (� ���� � ���� ������� �� ������� ������� ������* 8������� ��� .� �� ���� � ���(��� ������ ���������� �� ��� �� � ��)�� ����� �� ����������� �� ���)�� �� ���� ��� ���� � ���� �� ������ ������ � ����)��� ����� �� � � � � � ���� ���� ��� � �� � �� ���)�� �� ��� ��� (� � �c������ � ����� * +� ������� � ����� ���� ���� ������� ������ ��� �� � � �� �� ��� ��� � * +� ���� ����� ������� �� ��� � �� (� ������� �� ��� �� ����� ) �� ��)�� ������ �� ���)�� �� �(��� ��� ������ ��� ���* /� � ����� � ���� � �� �� � ��� ��� �� �������� �� � �� �� )���� �� ����� � ������� ��� ������* 8������ ������ � � ��� � �� �-����� ��� ����� �� ������� �� �� � � �� ��� ���� ��� ���� � ���� �� ��� � ����� ������* /�� ��� . �� �� �� ��� �� �� ��(�� ������� ������� �� )��� ������ ���� (���.� (��� ��� � �� ��� ���(��� (� � ���� �� ��� ������� �� �� � � ���� )�� ���� �������� �� ���� ��� ������ ��� ��� � � ��� ������� ��� �� ��� � � ������ ��* ��#��$�� ������ /�� ������� )��� � �� �� ����� 0� 3���� ���������� :� * ��� ��� � ��� ���� � ) �� �� � ������ �* /�(�� ; 6�� �� ��� 9 ������ ������� 6(��� � ������� ����9 3*2* 1�� � *�* 3 �� . ��� ������ � 2-�6 9 '� ���)�� �� #*��& #*4# #*#%& #*> �*�"4 : ���� ���)�� �� #* 4> #* # �*% # #*" ; �*=&; ������� ���)�� �� �#*#=# #*%&; #*#�4 #*&## #*&=� ��� ���)�� �� '�� �. ��� #*�%> #*"4 #*"; #*4## �*� > 8 ��� �� ��� " ��� � ��� ��� �>�#*&%�� �� ��c���� %% *""�� *�* 6%49� ������� #*###� '��������� /" � #�"4 * ">> )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� %�&������� ������� �� ��� ��� �� ������� �� ������ �� 6"##�9* #,/# ����� � ��� 9�� ��:) ���� ;� ����* 3� ���� ������� �� ��� ��� �� ������� �� ������ ��* ��� ��(��� �����??)))*����*���*��* ������� �� /������� 6�&&>9* ��(�� ������ �44!3�44<* ���(����� ���3* ��� ��(��� �����??)))*����*���*��* 7�b����� �� j* :*� 7�������� �*� g ! ���������� '* �* 6�&&&9* ��� �� ������ �� � �.�� ���� (�� �� ���� �� � � ���* �� �� � �������� /������ <6�9� �<�#* 7 � � �*� �� �� e*� g � ����� �* 6�&>%9* /�� ��������� � �� ������� �� 3���������� �� 8�� �* #���� �� ���� � �� ) (��� �� <� &<4&* ������ �*� g ������� 5* 6�&&&�9* 9�� ����� � ��� � ������ � #���� �� * 3� ���� ������� �� 3���������� �� 8�� � ���� �� �� ������ � ������* ������ �*� g ������� 5* 6�&&&(9* 9�� ����� � ��� � �+�� � #���� �� * 3� ���� ������� �� 3���������� �� 8�� � ���� �� �� ������ � ������* ������ �* :*� g � ���� �* �* 6�&&&9* 8� ���� �� � �� � ���� �� ����� �.�� (���.� ������ �.�� ���� (�� ��* 7 ������ � /�� ��� �� 2<6�9� �>< =* 5�k�� 2*� g 3��i� 2* 6"###9* , ������ ��� � � ������ � ������� � � �������� � �� % ��� � =�� �� �� ����� (���> �������* ���(� ��� ��� ����� ������� +��� ���� �� /� ������� 61��� �� ����� ##�#;9* 5���(���� !* e*� ������� 7*� g 3��� �* 6"###9* 3��� � �� ����� ��������<�������� ���� �� ������ ��� ���� �� � � ��� ������ ��* ,���� �� ,�������(�� 62� #=< %"* 5������� 7* 6�&&&9* / �� �� ���� �� ����������� ��� �� � ������* 9�� #���� �� �� �� � /������ '����(��� %=* 5)���� �* 5*� � ������� !* e*� g : ��� !* �* 6"###9* ��� �� ������ �� � � �� ��� ��� �� �� � ���� ������ ��� ��������* +� ,������� (� �� ��� � ��� ����� ��� � ��� �)# � � ����� (* 3������* 2� ��� e*� g �� ����� �* �* 6"###9* ��� )���� ���� � �� ������n 3��� �� �� � ���� ��� e�3* +� ,������� (� �� ��� � ��� ����� ��� � ��� � ���� �� ? �� � �������� � � ����� (* 3������* �������� '*� �������� �*� g 7��)�� o* 6"###9* � � �� �, � � ��� �� �� �� ������* ������� ��� !���� ""<"4* �������� �*� g ������������� �* 6"###9* ��� �� ����������� �� ��������� �� ��� ���� ���� � ���� ������� �� ���(�� � � ���* +� ,������� (� �� ��� � ��� ����� ��� � ,#�#, �� ���� ��* ���(�����* ��� �� ��� 2*� g ��� �� ��� �* 2* 6�&&;9* ��� �� ������ �� � ��� � �� �� ���� ���)�� �� �� .��� �� ����������* ,�������(�� � /������� <�6"9� "%4<"%>* �������� �* :*� g 3�� ��� ��� /* :* 6�&&"9* /�� ������ � ��)�� �.�� ���� (�� �� ���� �� ������ �� � �� . � ��* 7 ������ � /�� ��� �� 2�6"9� %4�<%4&* +������ � �� 3���������� �� ���� �� �� 6�&&>9* ���(�� ��������� �� �p������ � �� ���* 7�� ������ � 2* :����� /* 6�&&>9* 1����,� ) �� � �� ��� � � �* �� �� � ,� � (� l ��(��* ��������� o* 6�&&&9* 3���� � � �� �� ���� �� � �� ������� 3����� ��� ��* 9�� #���� �� �� �� � /������ ��� �� ="* m� �� ���� 7* 6�&&;9* 1���� g ������ � )��� �� ) �* ,���� � 7 ������ �� !���� #<=�* 3�) � � !* 6�&&;9* +��������, �������� �� ��� ��������� � �� ������� ���� ��� ��������� �� �� � ���� ��� ������� �� )�������� ��� � ������* ������ 1�� ������ � ������ � � ���� �� �� 8 ��� �� /�� 0� ���� �� �� 1������ ������� � 61��� �� ����� 3�� �� &;�;%9* 3 ��(���� �*� 7��)�� �*� ������� �*� g 7�� � ����� e* 1* 6�&&&9* 8 ��� �� � � ������ ��� ��� )���� ������ ���� � �� ������n #����� ��� ���� /������ <46"9� %>�<%>=* 3� ��� �* 6�&&&9* 3�������� � ��� ������� �� �� ������ ���(���* 9�� #���� �� �� �� � /������ 5� ��(��� %%* 3�� ��� �* 2*� g 3������� 7* !* 6�&&>9* ��� �� ������ �� � ��� ���� �� �� �� ������ � ��� ������ ��� ��� �����* #����� ��� ���� /������ <<� "#;<"�#* j��5���� � !* :*� g l����� o* �* 6"###9* 3� �� �� �� �� ��������� � ������ ������� ���� ���� ����� �� �� #�6�9 ���* �� �� � �������� /������ 46�9� 4=<;>* j��� 2*� 7��)�� �*� g ��� �� �� 3* 6�&&=9* � ���� �� ���� ���� � � ��� ��������� � �� '�) s����� �� ������� �� �c� �� ������ ��� �* #���� �� ( /��� ��� ���� �� <� �&<% * )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� ">& http://www.apra.gov.au. http://www.agps.gov.au. 1� �(������ 3* 6�&&&9* 5� ���� �� ����� ) �� ���� ���� ��������� �� � ��� � �������� � �� ��� ���� �c� ��n ���� � ���� �� /��� ��� @�����+ � ��� � ) �+��� #8��� ���% #���� ,���� ( � a �� ��� "��+� ( , ��� ������� 26"%9* � ��� ��� !������ �= "##�* ��� ��(��� �����??������*����* ��?�����*���n�(���� �t �"#=#%=* d��� ����(��� e���� 1������ 6�&&&9* 1����,� � ���� ������� ff�c��� )����,,� ������ * 3� ���� ������� �� ������)�����?3���� �� '�) s����� 3��� �� ���� ���� �� � � ���� ��� ��� 3����� �� 1���� 6.��� ������9* '������� ��? ���� �� +�� ����� ����� � ��� % ��� ��� ��� �� � �� �� �� � � ���� )��� �� � �� ) �� � � �) �� ���� �� �� � ��� �� �� ��� ������* ����� � ��� �� �% ��� ��� �� )� � ������ ��� ���� (��� (� �� � � ��� �� ?�� ��� � �������� �� �� ��� � )��� �� � �� � �� �� �� ���� � �� � ��� �� �� ��� ������* +� ��� ���� ������ �� ���� �� ��� � �� � )��� � �� � (��� #�6�9� �� +���* ��? �� 8� �?� �� � �� �� % � ����������� �� ��� ��� ��� ���� (���.�� �� )� � ��� c���� .� ���� ��� �����* 7���.�� ��������� ��� �� �� ���� ���� ��� �� ��� � � ���,� ������ �� �� ��� � .��� ����� �� )���* /�� ���� ������ � (�� ���� ��� ���� �� ���� ��� � 0*3* -� �� � �� ��% ��� ����� �� � �� ���� �� )�� ��� ������� � ����� � �� �����* 7 ������ � ������ ��� ���� �% � ����������� �� ��� ��� ��� ���� (���.�� �� )� � ��� ��� ���� ���� ��� �� ��� ����� �� ��� ���(��,� � ������� ������ � ��� ��� �� ��� � �� �� ��� � ��� ������* /�� (���.� ��� ������ �� � �� � )����� ���� ��� �� ��� ��������� � �� ��� ��� �, ���������� ���� � ��* +� ��� ���� �� � �������� �� � ���� �� ��� � ���c������ ������� �� �� � �.�� ���� (�� �� ����* 7 ������ �� �% � ����������� �� ��� ��� �� �� ���� � )������ � � ���� ���� � ����� �� ��� �� ��� �-����� �� ���� � �� �� ��� �� ���* /�� ���(���� � �� ��� ��� � �����(�� �� ���� �� ���(��� ���� � ) �� � ��� � ����� �� ��� �* 1������ ��������� ���� ���� ����������� �� �� ��� (��� �� �� � � ���� ��� � �� ��� ��� * "&# )* �� �+')������ ,* -��� � . �� �� � �������� /����� �� 0 ���1 !23 4� http://papers.ssrn.com/paper.taf?abstract_id=210205035. http://papers.ssrn.com/paper.taf?abstract_id=210205035. consultation and resource usage in retirement savings decisions: australian evidence of systematic gender differences introduction2 the australian superannuation industry the literature on superannuation choice the literature on gender and investment decisions the ssau survey estimation and results choice of fund knowledge of superannuation people consulted and resources used people consulted individual logistic regressions on person type consult nobody consult work colleagues consult a professional consult family/friends consult partner multinomial regressions on most important person consulted conclusion acknowledgements references definition of key terms pii: s1057-0810(01)00087-7 ������� ��� ����� �������� ����� �� � � ����� ������ �������� ��� � ���� ����� ����� ���������� ������ �� ��� �� � � ��� ������ �������� ��� � ����� ��� ���� ��� � ����� �� ���� ��� ����� � !��� ��� ���� ���� "� #$%&'� �� �������� �� ������� ���� �������� � ������� ���� �! "# � ���� �������� �$ %������� ���� �������� ���� ����� �&��� �� � � ������� � ���� ��� ������'� ������� ���� ���(�� �� ������� ��#� ����� ������ ���#� � � � ������� ��� ����� �������� )�� * ����� � �����+ ,&��� � � -./ 0�� ����� ������ ���#� � ���(�� �1!� � � ����2 ��� ��#�� 3 �� ���� ����� �� � �������� � � ��� �3�� � ���� ��� ������ ���������� (��� �� ����� � �����+ ��� ���� ������ ���� (���� �� ����� � ����� � �� ����# ��� ��� 4+4�5 �� ��� ����� � ����2 �� ����� � ���� ���#� � ����# ��� ��� ���� ��5 �� ��� �#�#������ ���#� � ���� ��� ���� ������+ 6�� � � �#��� �������� ��� �� ����� � �����2 ��������� ���#��� 3 � ���� ��� �� ����� � ����� ���� � ��� �3�� ��� �������� ������ � ������� ������ ���#� �+ � ���� ,������� -��� �� 7 �+ 8�� ������ ��������+ ��� ����� ��� �( 9�'9� ���� �� � ���� ������ )��* ���( �� -���� ������ ���#� � :���� � ������ � ���� ������ ��� �� ����� ������ ��� �� ��� ������� ������� -�����2 ��� ������� ��� ����� �������� )�� *2 ����� �����&������� ����� ! (����+ 8� ����� ����� ��2 ��� �� ������� ����#�� ��� �#��� � � � ���������� ��� ���� �� ����� � � ��� �����#� ����� � �� ��� 6 ���� -�����2 � � ��� �#��� � � � ���������� � ��� ���� �� ��� ���� �� ����� � �� ���� ������ �� ��� 6 ���� -����� ��� ���+ ;���� #�� ���� � �������� 2 ��� ���� � ������� �� ��� �� ������� � � �#��� � �� ����� ������ ����� � � ����#���� ��� �� ���������+ ��� �� ��������� ���3�� ��� ������� ������� -����� ��� ������ ���� � � ��� �#���� �� �� � ���� -������� �����( �� )����* �!�<�$� ��0$=�>��?��?@ < ��� ��� � ������ � ���� ,������� -��� �� 7 �+ 8�� ������ ��������+ /77a � � 0 $ = � > � � ) � � * � � � > $ = $ � ���� ���b��� (�� ��������� �# ��� �� � c������� -���� 6 �������� ���#��� d�������� � � � 7�������� � 9�� �+ � ���+a ��=��1=�40=��4! ��&a ��=��1=�40=��4>+ �+�� � �������( ������e��������+��# ) +�+ ����*+ ��� �#��� � �� ����� ������ ����� � � �������� � � ������� ���#� �������� �#�#�� �� ����� ������ ����� �+ ����� ������ ����� � ������ �����=���� � ������ ����� � � �&������ � f���� �����2 � � #��������� � f#� �� ��� ���� �#��#�2 3�� ���3�� � � ����� ���#��+ ���� ����� �&��� �� ��� ���� ��� � ����� ���������� (��� ��� �� ����� � ����� � � �� ������� ��#� ����� ������ ���#� �+ ��� ����� ��� �� ��� ����� �� ������� � ��#� ������ �+ -����� � �������� � �������#�� �����(+ -����� � ��������� ��� ���� � � �����������+ -����� 4 ����#���� ��� ���#��� � � -����� 0 �������� ��� �� ��#��� � �� ��� �����+ � ���������� ������ 7 ���� � �����2 ��� ������� �� ���� ��� � ������� � ����� ������ ���#� � ���� ������ � ���g#� � ����� � 3 � �� � � ��� ���� �������#���+ � � ���� �� ���� ��� � ����� ���� ��� �������� � ����� ���� �� ���� ��� �� ��� hh(���� � ������ii+ 6�� � -./ 0�� ����2 ��� �� )�1>�* (�� � � �� ��� 3��� �� ���#�� � ��� (���� � ������ �� ���� ��� ��� � ��� ���#� �� ��� �3�� ��� ������� � � ��� �3�� ��� ��(�� ��� ��� ���#� � �� ��� ����� ���� �� ��� (���+ j� �� ��#��� ���� ��� ������� ��� � ��� ���#� (�� ����� �� �� ������ � ��3��� ����+ :���� ����� � ������ � ����� ������ � � ���#������ � ������ � ����� � ��� (��� �� ����� ����� ��� ������ �����2 ��� �� )�1>�* �#������� ������ � ������� ��#����� �� ������ �#������� �� � ��� � � ���� � ������� � ��� ����� �� ������ �� � (�� �� �&����� ��� ���� ��� � ����� � � ����� ��� ������ � ��3��� ���� (����#� �� ����� � ������� �� ��� ������ �����+ k��� � � -�����#�� )�1>4*2 -������� � � -����� )�1>!*2 d�� � � ������ )�1>>* � � ����# � )�111* �� 3�� ��� (���� � ������ ���#���+ "���� � � :�����3��� )�1>0*2 � � �#��� � � j���� )�111* 3 � ���� ��� (���� � ������ �� �� # �g#� �� ��� 6 ���� -����� ����� �������+ "���� � � :�����3��� )�1>0* 3 � � (���� � ������ ��� ����� ������� � ��� 6 ���� k� ����2 "��� 2 � ���2 8#������� � � ��� 6 ���� -�����2 ��� �� (���� ����# ��� ��� >$5 �� ��� (���� ����� ������ ���#� �� ��� ���� �� ��� ��#��+ ���� ���� ��# � ���� ��� (���� � ������ � � =6+-+ ������� �� � ���� �� � �� ��� (���� � ������ � ��� 6 ���� -�����+ �#��� � � j���� )�111* ���#�� � � (���� � ������ ��� ��b�� 6 ���� k� ���� � �����2 �#� �� �� ���������� ������� ��� ������ ����#��� ��� ���� �� ������ ��3��� �� �� ��� �� )�1>�* �#�������+ ��� (���� � ������ �������� ��� ���� ���#��� � �&���� � � (�� ��� (���� � ������ �&����+ 8 (����� ���#�� ��� �&��� ���� ��� ��� (���� � ������ �� ��� �������� ���������� ���(�� ��� ������#� ����� � ��� ���#� � � ��� � ��� ���#� + 8������ � � 7�� ����� )�114* 3 � ���� (�� ������i� ������ ���#� �� �������2 � ���i� ������ ���#� �� ������� ����� >�5 �� ��� ����+ 6�� � �����#� ����� ������ � ����� ��� ��� �1!�<�11� ����=������2 :� �2 l#�� � � � ,������ )�11$* ������ ���� ��� (���� � ������ ��������� ���#�� �#�� � ��� ���� � (���� �� ��� �� �� � � ���� ��� ������� � ��� ���#� � ��� �� ����� ��������� �&���� �� �� ������#� ����� ��� ���#� �+ 6�� � -./ 0�� ����2 ��(2 j���� � � -��� )�11$* ������ ���� ��� ���������� ���(�� ������ � � � ��� ���#� � ��������� ��f���� ��� ���������� ���(�� ������ ����� �� � � � ��� ��� � � ���#� �+ "���� � � :�����3��� )�1>1* 3 � ���� ��� �������� ���������� ���(�� ������ � � � ��� ���#� � ��� �� ������� �� 6 ���� -����� ����� �������2 �#� �&�� � �� ������ ����� ������� �� (���+ �!4 ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ 8 ������� �� �&��� ���� ��� ��� (���� � ������ �� ���� ��� ����� � �������� �� � �����#�� � � � ����#��� �� � ������� �� �� ����� � (��� ��� � �������� =�������� � ����������+ ��� � �������� =�������� � ���������� ������ ���� � ������� �� � �� ���� ���������=������ �� � ������� � ���� ��� (���� � � � ���� ����� ������� � ��� ���� �� � #�� �������� � ������ �� ����� �+ 6�� � ��� %l-, � ��& ���� ���� �1!� �� �11�2 8������ � � 7�� ����� )�114* 3 � ���� � �����#�� � ������� ��� ���� ������ � � ��� ����� � ����2 ���������� � � ���� �����(� � ������� ������ � �������� + -#������ � ������� 3 �� �� �� ����� )�1>>*2 ���� �� ��#�� ���� ��� ������� � ��� ���#� � ��� �� ����� ��������� �&���� �� �� ��� ����� � �������� �� � �����#�� � �������+ ;������ � � ������#� )�11>* �#����� ��� � �������� =�������� � ���������� �� ���� �� ����� � (��� � ����#��� �� ����� �+ 6�� � -./ 0�� ���� ���� �1!� �� �11�2 ���� 3 � � �������� ���������� ���(�� ������ ���#� � � � � ��� ���#� �+ ���� ���#� ���� (���� � �����#�� � ������� ��� ���� ������ �� �� ����� ����� � ������ � ������� � � � �����#�� ������2 � ����#��� �� � ������� �� � �� ����� ������� �� ������ � � ��� ��������� ���� ������ �� �� ����� ����� ������ � ������#� ���#� � �� ����� ���������� �� � �����+ -#������ � ������� 3 �� �� -��� � � -����� )�110*2 ;������ � � ������#� )�11>* �� ��#�� � ����#��� �� � ������� �� ���� � ����#��� �� �( ������ ���� ������� ������� ������� ����� ���#� � � � ��� (���� � ������+ ����������� ���� ���� ��# � ����� �� �� ��������� ���� ��� � ������� � f#� �� � ����� ������ ���#� �+ ����� ��������� ���� ��� � ������� � ��#�� ��� ������� �� � /������ ���� ���� �������� ���� � � ��� � ������� �� ���� ��� � ����� ������ ���#� �+ j����� )�11�* �������� ������ ���#� � �#�� � ��� 3��� � ����� �� � /������ ���� ���� �������� (��� ��� ������ ���#� � �� ��� ���� � ����� �� � ���� �������� ��� 4� ���� �������� � ���� � � � �>�� ����#�� ��� ����� ���� �������� + j� 3 �� ���� ��� ��b����� �� ����� ������ ���#� � ���#� �#�� � ��� ���� � ����� �� � ���� �������� )������� ���� � � ���=������� ����*+ ,&��� � � �� ���� �������� � #� �� ��� �1>> ;#�� ���� �������� 2 ���# � )�11�* 3 �� ������� ���#��� (��� ��� �������� ���#� � ���#��� � �#�� � ��� ���� � ����� �� � ���� �������� + ,&��� � � ��� � ������� �� ���� ��� � �����2 ������� )�11�*2 � � m���2 �2 /���2 � � k� ��� )�11$* 3 � ���� ���� ��� ���� �� �#�� ����� ���#� � ��� ��� �3�� ��� ������� (�� ��� 6 ���� -����� � ����� �� � ������ )������* ��� (�� �� �� � ������ + m��� �� ��+ )�11$* 3 � ���� ���� ������� ����� ���#� � � � �#�#������ ���#� � ��� ������� (�� � ����� �� � ������+ ���� �� ������ ���� ���� � ����� �� � ������ �(�=������ �� ��� ���� � � � ������ � �=����� �� ��� ����+ ������� )�11�* ���� 3 �� ���� ����� ���#� � ��� ������ (�� � d��������� � ����� � � d��������� /������ � ��� � ��(��+ � ����� ������ �������� ��� ���#��� � ��� � f#� �� �� ����� ������ ��� ��� ���� � ����� ������ ���#� �2 �� � ��� � f#� �� ��� ��� � �� ��� �� ����� � ��� � ����� ������ ���#� �+ � ����� ������ �������� 3 �� � �������� ������� ���� ���(�� �� ����� ������ ��� ��� � � ����� ������ ���#� � )���� ���� ���2 ��������� � � 8����2 �11� /������2 �11$ ��������2 �11$ ����� � � -���� 2 �11>*+ ,��� �� �&�� ��� ��� �� ����� ������ �� �� �� � ������ �� ������� ��#� ����� ���#� � � � ����� �� �� ������� ��� �� ����� ������ �� �� �� �������� �� ������� ��#� ����� ���#� �+ �������� � � ������ )�11!* 3 � ���� �� ������� ��� �� ����� ������ ��� ��� ���� �� � ����� � � ������������� ��� �3�� � ����� � � ����� ���#� � ��� ����� � � ����� 3���2 � � ���� �&�� ��� ��� �� ����� ������ ��� ��� � ������ ����� ���#� � �� ����� �#� �� ���������� ����� 3���+ "� �� 2 "�� �� � � ;�#�� )�11$* ������ ���� ��� ��� ��#�� �� ��� ��� �� � ����� ���#� � ������� ������ � �#������ � ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ �!0 ��� �����=�# � � �� �=�# + %�� �#������ �2 ���� 3 � ���� 3��� � � �#������ ���� ��� � ������= ���� �� ������ �&����� �� ��� �������� ����� ���#� ��� ��� (�� �� ����� ������ ��� ���+ 8������ �� �������� 3 �� ���� �� ����� ������ ��������� ��� ��� �3�� � ���������� �� �#�#�� ����� ���#� � � � ���� �� ����� ������ ��� ��� ���� # �g#� ��������� � ��(�� ��� ���� ����� � � ����� 3��� )���� � � ��� ��2 �1>> "� �� 2 �����2 � � "�� �� 2 �11! ;���� � � ;����2 �11$*+ :���� ��� �������� ��� �&��� �� ��� � f#� �� �� ����� ������ ��� ��� ���� � ����� ������ ���#� �2 � ��#���� ���� �&��� �� ��� ���� ��� � ����� ���������� (��� �� ����� � �����+ -� �� ��� �� ���� �� ��� �� �� ����� ������ �� ���� ����� �2 ��� �&���� � �� ����� ������ �������#�� ���� �� ������� ��� ������ ��� ��� � �� � �� ����� � ��� � ����� ���#� �+ ���� ��#�� �&�� �� ��� ���� ��� � ����� �������#�� �� �&��� � � ��� � f#� �� �� ����� � ����� ���� � ����� ������ ���#� �+ ���� ��#�� �&��� �� (������ ��� ����� ������ ���#� � � �� ����� � ����� ��� ��� �3�� ��� ������� � ���� � =�� ����� � ���� ���#� �+ ��� ��#�� ���� �&��� �� ��� ������ �� ����� � ����� ���� � ����� ������ ���#� �+ � ���� �� ����� ����� ��� -�� ���� � � /���i� 0�� )-./* ��� � � ����� � ��� � ���#�� ��� ��������� ���� "� #��� 42 �1!� �� "#�� ��2 ����+ ��� #���� �� ����� � ���� �� ��� ���� ��! �� �04 ���� ��� ����+ ��� ����� ��� � ���#� � � ����� ���#� � ��� ����#����� ��� ���� ����� � ���+ ��� -./ ��� � ���#� � � ����� ���#� � ��� ����#����� #�� � ��� �����=��=����� ��� � ���#��+ ���� �� �������� ����#�� ��� ��� � � ��� � ���#� ���� ���� ������� ���� ��� ����� � ��� � ���#� � ��� ��� ����������� ������� � ��� ���� � � �� ��� �&� ���i� ������ + ��� �� ����� � ����� �� �������� �� ��� ������� ������� -����� )�,d* ��� �������� ��� ��� 4�=���� ������+ ���� ��� ������2 ��� �� ���� ><�� ����� � ���� ��� ���� (��� ��� ��b����� �� ��� ����� � ����� ���#��� � � �#������+ ��� ���� ��� �������� �� �#��� ���������a � ��� ��� � =�� ����� � ���� � � � ��� ��� �� ����� � ����+ ����� (��� 40� �� ����� � ����� ���� ��� ������ �&��� ��+ ��� -./ ��� � ���#�� � � ����� ���#� � ��� ����#�����2 � � ��� �������� �� � �� ����� � ����� ��� ��������+ /=����� ��� ��������� �� �&���� ��� �3�� � ������� ��� � �� ������� ��#� ����� ������ ���#� � ���(�� �� ����� � ����� � � � =�� ����� � ����� � � �� �&��� � ��� � f#� �� �� ����� � ����� ���� � ������� ������ ���#� �+ � �������� ��� � !" �##��� ����� � ����� �� � �#����� �� ��� ���� �&��� ��+ ��� 3��� ���#��� � ����� � �&��� � ��� � ���� ���� ��� ���� � ��#��� ��2��$ ����� � ���� �� (���� �� ����� � ����� ����� 40� ���� �� 4+4�5 �� ��� ����� � ����+ :���� ��� ������� ����� ���#� � ��� � =�� ����� � ����� (�� �+��52 ��� ������� ����� ���#� ��� �� ����� � ����� (�� �+��>0 �� �+05 ������ ��� � =�� ���� ���#� �+ 8����#�� �� ����� � ����� � �� ����# � ��� 4+4�5 �!! ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ ����� � d���������� ���������� �� �� ����� � ����� � � ����� ������ �������� �� �� %� =�� ����� �1!�<���� ����=������ ����� ����� � ���� 40� 12$!0 ��2��$ � ��#� ����� � ���� > ��4 ��! �&��#� ����� � ���� �� �4! �04 8������ ����� ���#� �+��>05 �+���$5 �+��0�5 #�#������ ������� ����� ���#� 41+��5 ��1+005 �0>+!�5 5 �� #�#������ ������� ����� ���#� � ��+!$5 >!+��5 ���5 �� ��+ � =�� ���#� � d������ �� �� ��� � /=���� / � 4�+$�4� ;�������i� ������ ���� �� ����� �� ���� �� � �$+0��� �1!�<�1$1 ��+ �1>�<���� ���#� � d������ �� �� ��� � /=���� / � !+�>!�� �1!�<�1$1 ����=������ ����� ����� � ���� �>0 42$�> 02��� � ��#� ����� � ���� > ��4 ��! �&��#� ����� � ���� �� �4� �0� 8������ ����� ���#� �+���15 �+��4>5 �+��4!5 #�#������ ������� ����� ���#� �+�15 !1+1$5 $�+�!5 5 �� #�#������ ������� ����� ���#� � 4+!5 10+45 ���5 �� ��+ � =�� ���#� � d������ �� �� ��� � /=���� / � �+��!> ;�������i� ������ ���� �� ����� �� ���� �� � �+���� �1>�<���� ����=������ ����� ����� � ���� �!$ 02��! 02��� � ��#� ����� � ���� > �4� �4> �&��#� ����� � ���� > �4! �04 8������ ����� ���#� � �+�$�45 �+�4$$5 �+�00�5 #�#������ ������� ����� ���#� 40+!!5 �4�+��5 �>0+>>5 5 �� #�#������ ������� ����� ���#� � �0+1$5 >4+��5 ���5 �� ��+ � =�� ���#� � d������ �� �� ��� � /=���� / � �40+04� ;�������i� ������ ���� �� ����� �� ���� �� � �+��44� �1>�<���� ����=������ ) � ;���� � ���* ����� ����� � ���� �!$ 02��0 02��� � ��#� ����� � ���� > �4� �4> �&��#� ����� � ���� > �4! �04 8������ ����� ���#� �+�$�45 �+�0�$5 �+�0>>5 #�#������ ������� ����� ���#� �+40+!!5 �!�+��5 ��0+1$5 5 �� #�#������ ������� ����� ���#� � �4+1�5 >0+�>5 ���5 �� ��+ � =�� ���#� � d������ �� �� ��� � /=���� / � �0$+�4� ;�������i� ������ ���� �� ����� �� ���� �� � �+�!�0� � -�� �3�� � �� � � ����+ �� -�� �3�� � �� � � ���0+ ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ �!$ �� ��� ����� � ���� �&��� ��2 �� �#�#������ ������� ����� ���#� � ����# � ��� � ���������� ��+!$5 �� ��� ������� �#�#������ ������� ����� ���#� �+ 8 8%�c8 � ������ (�� #��� �� ���� ��� ��� �3�� �� �� ��� ������� ��� �� ��� ��� ���#� �+ ��� � ������ ������� � /=���� ��������� ���� �&����� ��� �5 �������� ���#� � ������ � ���� ��� ��� �� ����� � ���� ���#� � ��� ��� �3�� ��� ������� ��� ��� ��� � =�� ����� � ���� ���#� �+ ��� � ������ ���� ������� ���� ��� ;�������i� ������ ���� �� ����� �� ���� 3 �� ���� ��� ����� �� �� �� ����� � ���� ���#� � �� ��� �3�� ��� ��(�� ��� ��� ����� �� �� � =�� ����� � ���� ���#� � �� � �5 ����� �� ��� �3�� ��+ -� �� �� �� (����� ����� �n�� ���� ��� �1!�<�1$1 ������ �� ������� � ���� ��� �1>�<���� ������2 ��� ���� �� ������� � �� �(� ���� ���� � � � � ������ �� ��� �(� ���� ���� ��� ����#�����+ -#������ � ���� ����� � ��� ����2 ��� � ������ �� ��� �(� ������� ������� ���� �� � 05 ����� �� ��� �3�� ��2 �1!�<�1$1 ��� ������ ���#� � ��� ��� �3�� ��� ��(�� ��� ��� �1>�<���� ��� ������ ���#� �2 � � �� � �5 ����� �� ��� �3�� ��2 ��� �1!�<�1$1 ����� �� �� ���#� � �� ������� � ���� ��� �1>�<���� ����� �� �� ���#� �+ 7 ����� �� �� ���� ����� ��2 ��� ����� ��� ��=����#����� ��� ��� �(� ���� �#�����+ ��� ���� � ������ � ����� � ������� ��� ���#��� ��� ��� ������ �1!�<�1$1+ ���� ���� ��� � ��#��� 02��� ����� � ���� �� (���� �� ����� � ����� ����� �>0 ���� �� 0+$5 �� ��� ����� � ����+ ��� ������� ����� ���#� ��� � =�� ����� � ����� (�� �+��4>5 � � ��� ������� ����� ���#� ��� �� ����� � ����� (�� �+���15+ 8 � ������ �� ��� ���� ������� ���� ��� � =�� ����� � ����� ��� ���#� � � ����� �� �� ���#� � ��� �� ��� �3�� ��� ������� � ���� ��� �� ����� � ����� ��� ���#� � � ����� ��+ ����� ���#��� ��� g#��� ������� � ���� ��� �1>�<���� ������ ���#���+ ��� ����� ������ � ����� � ������� ��� ���#��� ��� ��� ������ �1>�<����+ ���� ���� ��� � ��#��� 02��� ����� � ���� �� (���� �� ����� � ����� ����� �!$ ����� � ���� �� �+��5 �� ��� ����� � ����+ ��� ������� ����� ���#� ��� � =�� ����� � ����� (�� �+�4$$5 (���� ��� ������� ����� ���#� ��� �� ����� � ����� (�� �+�$�45 �� ���� ��� 0+$5 ����� ������� ��� � =�� ����� � ���� ������� ����� ���#� �+ :���� �� ����� � ����� ����# � ��� � �� �+��5 �� ��� ����� � ����2 �� ����� � ���� ���#� � ����# � ��� ����� �!5 �� ��� �#�#������ ������ ���#� ��� ��� ������+ 8 � ������ �� ��� ���� ������� ���� �� � �5 ����� �� ��� �3�� �� ��� ��� �� ����� � ���� ���#� � ��� ��� �3�� ��� ������� ��� ��� ��� � =�� ����� � ���� ���#� �+ ��� � ������ �#����� 3 �� ���� ��� ����� �� �� �� ����� � ���� ���#� � ��� ��� �3�� ��� ��(�� ��� ��� ����� �� �� � = �� ����� � ����� ���#� �+ 8 �� ��� ��� ��� ������ ��� �� ���� hh;���� � ���ii �> �������2 �1>$2 � � =�� ����� � ����2 ����� ���� ��� ���#���+ d����� � ;���� � ��� ���� ��� ���� ��� � � ��=�������� � ��� �����2 ��� ���#��� ���( ���� �� � �5 ����� �� ��� �3�� ��2 ��� �� ��#��� � �� �� ��� ��+ �� ��� ���� � ����� � �� ��� ������� ���� ���(�� ������ ���#� � � � ��� �� ����� � ����� � ��������� ��������� ��� � f#� �� �� ����� � ����� ���� � �� ������� ��#� ����� ������ ���#� �+ ����� � ������� ��� ���#��� �� ��� ��������� � ��� ��� ����� ���� �������+ ��� 3��� ������ ������� ��� ��������� #�� � ��� �1!�<���� ���� ���+ ��� ���� � ������ � ����� � ������� ��� ��������� #�� � ��� �1!�<�1$1 ���� ���+ ��� ���#��� 3 � ���� �� ����� � ����� ���� � � ��� �3�� � ������ � ����� ������ ���#� �+ ���� ���#�� �� �� �#������ �+ /���� �� �1$02 �,d �� ����� ������ ����� � (��� ����(��� ���������+ -� �� ����� � ������� � � � �� #��� (��� �� �������� ��� ������� ����� �����(� � ��� ����� � �!> ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ �����2 � ������� � �( ������ ���#� �,d ��b������� � � ��( ��� ��b������� ������� �� ��� ��� ���+ ���� ���� �� ��� �� � �1$0 (�� � ����� ������ j�#�� � �#��� � �����#��� ��� )�1$0* ���� ��g#���� ��� �,d �� � �# �� ��� ��b������� ��� ��� ���(�� ����� �� ��� �� �� �#����+ ;� �1$>2 ��� �#�� ,������� � � � ;��� ��� 9��(�� 8�� )�1$>* )���� � �( �� ��� j#������<j�(�� � 8��* (�� ������+ ���� ��� ��g#���� ��� ������� ������� -����� ���� �� ������ ������ � ����� � � �&���� ��( �� ����� ������ ��b������� ��� �� ����� � (��� ��� ��� ���� ���(�� � � � ������ ��� ���+ -�2 ��� ���� �� ��� �1!�<�1$1 ����=������2 �� ��b������� � � ����#���� � �� ��� ��� ��� (��� �� � �( �� ��� �� ���� �#����+ :��� ��� �,d ������� � ������ �� � ������� �� � �������� �� � �������2 �� ����� � ����� ��� � ��� �3�� � ���� ��� ������ � ����� ������ ���#� �+ ��� ����� ������ � ����� � ������� ��� ��������� #�� � ��� �1>�<���� ���� ���+ ��� ���#��� ������ ���� ��� ���� ���� ������2 ����� �� � ���� ��� ������ ���������� (��� �� ����� � �����+ 8� � �5 ����� �� ��� �3�� ��2 �� ����� � ����� ���� � �������� � � ��� �3�� � ������ � ����� ������ ���#� �+ /������� ����� � ��� ���� ���� ��� ������ � ��#�� ��� �,d � ������ � ��� ���# � � � ������ ��� �� � �������� �� ��� �#����2 � � ��� � ������� ������������ �� ��(=���� � �������� + �� � �#�� ��� ���#��� (��� �� ������ �� hh;���� � ���ii2 ��� �1 �������2 �1>$ ������ ���#� (�� ������� ���� ��� ���� ��� � � ��� ��������� � �# ���� + ��� ���#��� �#����� ��� ����� �� ���� �� ����� � ����� ���� � �������� � � ��� �3�� � ������ � ����� ������ ���#� � �� � �5 ����� �� ��� �3�� ��+ ��� ����� �� �#������ ���� ����� �� � ���� ��� � ����� ���������� (��� �� ����� � �����+ ����� � ��������� ���#��� �� ��� ������� ���� ���(�� ����� ������ ���#� � � � �� ����� �� ���� ������ 7 ������� �� / ��������� �1!�<���� �+���$ �+�$!> �+�1$��� -�� ���� ����� )�=���������* �+��1� )�+0��*� �+�4�1 )�+$>>*��� �� �� �� �� ���#� � � a ��+!$4$ �&a �+>>0� %� =�� �� �� �� ���#� � � a ���+4!$ �&a 1+�11� �1!�<�1$1 �+��4> ��+���1 �+��4 -�� ���� ����� )�=���������* �+���� )�+�4�* �+�4!� )��+�!�4* �� �� �� �� ���#� � � a ��+!$4$ �&a 4+!4>! %� =�� �� �� �� ���#� � � a �!+!$0! �&a 0+���4 �1>�<���� �+�4$$ �+��0> >+�4�� -�� ���� ����� )�=���������* �+��4� )�+�!�*� �+�$1� )�+>0�*� �� �� �� �� ���#� � � a ��+40�0 �&a �+!�$� %� =�� �� �� �� ���#� � � a ���+4!$ �&a 1+�11� �1>�<���� ) � ;���� � ���* �+�0�$ �+���$ >+0��� -�� ���� ����� )�=���������* �+���! )�+>�0*� �+�$01 )�+1��*� �� �� �� �� ���#� � � a ��+40�0 �&a �+!�$� %� =�� �� �� �� ���#� � � a �>+�>1� �&a 1+�11� ���� � �� ��� � �2 (���� �� � �2 �� �� �� �� ����� �+ �� � �2 �� �� �� ����� �+ � -�� �3�� � �� � � ����+ ��� -�� �3�� � �� � � �����+ ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ �!1 $ "��������� ���� ����� �&��� �� ��� ���� ��� � ����� ���������� (��� �� ����� � ����� � � �� ������� ��#� ����� ������ ���#� �+ ��� ��������� ���#��� �#����� ���� ����� �� � �������� ���� ��� ������ ���������� (��� �� ����� � �����+ :�� ������� � ��� ������ ���#� � ���(�� �� ����� � ����� � � � =�� ����� � �����2 ��� ���#��� 3 � ���� ��� ��� ���#� �� ����� � ���� � (���� ��� �� ��� (�� ��� �3�� ��� ������� ��� ��� ��� ���#� �� ����� ����� � ���� ���� ��� ����=������ "� #��� �1!� �� "#�� ����+ 8 �#����� �&��� ���� �� ��� ���� 3 �� ���� ��� ���#��� ��� ����� �� ��� ��� ���#� � ���(�� �1>� � � ����+ ��� ��� ������ ���#� � ��� ����� � ���� ��� �� ��� (�� 0+$ ����� ������� ��� ��� ��� ������ ���#� � � =�� ����� � ���� �#�� � ��� ���� ������ �1>�<����+ ��� ���� ���� ������ ���� �� � � ������ (��� �#� �� ��� ������ � �� ����� � �����2 ��� � ������ (�#�� ���� ������ �#� � ����� �!5 �� ��� �#�#������ ���#� � ���� ��� �1>�<���� ����=������+ 6�� � ��������� � ������2 ��� ���#��� ���� 3 � ���� �� ����� � ����� ���� � ��� �3�� ��� �������� ������ � ������� ������ ���#� �+ :���� ���� �������� �� ���� �� �#��� ��� �#�#�� �������� ��2 � ������� �� #�� ��� � �(����� �� ���� ���� ��� � ����� (�� ������ � (�� �� �#� �� ���� ������+ ��� ����� �� �#������ ���� � ������� ���#�� �� � ��� ������ � ���� ��� �� ����� �� ���� ���� � � ������������ ��� ���# � �� ������ ���#� �+ ��#������� 8������2 8+2 . 7�� �����2 d+ )�114*+ ��� � �����#�� � ������ � � ��� (���� � ������+ � ����� / ���� �� ��� 6���� ��� !� ������ �2 2&)�*2 �!�<�$$+ ;����2 "+ �+2 . ;����2 m+ + )�11$*+ ,�� ���� �������2 �� ����� ������ � � �&������ ���#� � � ������ � � �� ��+ /,� ��� /���� �� 7�� � � � ,�! �*2 22 ��<4�+ ;������ 2 /+2 . ������#�2 d+ )�11>*+ 7 �����#�� ����#� � ����#��� �� � ������� � � ��� (���� � ������+ � ����� �� � � �� ��� / �����2 22)-��� � )�**2 $�<>0+ ��(2 ,+ j+2 j����2 /+2 . -���2 +,+ )�11$*+ ����� � ���#� � ��� ��� (���� � ������ #�� � � ������ ����+ � ����� ��� ���� / ����� ��� ��� ��� ��2 28)8���� )�* � � )4**2 4�0<444+ �#���2 8+ "+2 . j����2 /+ 8+ )�111*+ ��� (���� � ������2 ��� ����� �&��� �� ����# � � � ��� 3 � ���� ����� � �#������ ���� ��� ������ � ��&a �1>$<�114+ ���� �� / ���� �� �� � � ��2 &2 !$<$�+ d��2 ,+2 . ������2 ,+ )�1>>*+ 8 �������� �&��� ���� �� ��� (���� � ������+ / ��� �� �������� � ����� ) ��< "# �*2 >�<>4+ ����2 ,+2 . ��� ��2 k+ )�1>>*+ d����� � ������ � � �&������ ����� ���#� �+ � ����� / ����� �� � � ��2 222 �<�0+ ����# �2 /+ )�111*+ 8�� ������ ������� � ���� (���� ��o 8 b#�� ����#��� � ������ �� ��� h(���� � ������i+ /,� 9�* �������7�� � � � ,�! �*2 :)-��������?�������*2 �<�1+ ��� ��2 k+ �+ )�1>�*+ -���� ���#� � � � ��� (���� � ������+ � ����� / ���� �� �� � � ��2 ')�*2 00<$�+ �#�� ,������� � � � ;��� ��� 9��(�� 8��2 �1$>+ /#���� m�( 10=0��2 ������ ������� �$2 �1$>+ j�����2 l+ )�11�*+ -;� �� �� -�����<� ������� )�0�� ��+*+ %�( l���a /� �#� ;����+ j�#�� � �#��� � �����#��� ���2 �1$0+ j+ � + ���+ ���2 ������ ���� �42 �1$0+ "����2 "+2 . :�����3���2 �+ )�1>0*+ ��� (���� � ������ � ����� ����� ���#� �a ��� � ��� ���� �� ����� ��+ -;� � ����� / �����2 =�)"# � )�**2 4��<404+ "����2 "+2 . :�����3���2 �+ )�1>1*+ 7� ����� � �� ���� ������ � ����� ������ ���#� �o ,���� �� ���� ������ ��# �����+ � ����� ���� �� ��� / �����2 $#) �� )�**2 ��$<�44+ �$� ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ "� �� 2 9+ �+2 �����2 "+ +2 . "�� �� 2 �+ �+ )�11!*+ ;#�� ��� �� ����� �2 �� ����� ������2 � � �&������ ���#���� ���#� �+ � ����� / ���� �� �� � � ��2 80)����#��� )�**2 ���<��$+ "� �� 2 9+ �+2 "�� �� 2 �+ �+2 . ;�#�� 2 :+ -+ )�11$*+ ������� ������� �� ����� ������ � � � �#���� ����� ���#� �+ � ����� ��� ���� / ����� ��� ��� ��� ��2 28)"# � )0**2 !�1<!44+ k���2 d+2 . -�����#��2 �+ )�1>4*+ 8 �#����� � ���������� �� ��� (���� � ������ � ����� ���#� �+ � ����� / �����2 #&2 >�1<>�0+ m���2 �+ /+2 �2 k+ +2 /���2 �+ d+2 . k� ���2 :+ �+ )�11$*+ ��� �� ������� �� ���� ��� � � ����� ������ �������� ��+ / ���� �� ���! ��� ,�! �*2 %)�*2 �1<�0+ ���# �2 �+ "+ )�11�*+ � ! � �� �;� � *� ������2 7ma /���#� /#������ �+ ������� 2 -+ )�11�*+ 6�� � �� ������� �� ������ � �� ������� ��� ����� ������+ � ����� ��� ������� �� � � � >������� !��2 &2 >1<11+ �����2 ,+ + )�1>>*+ :�� � (���� � ������o � ����� > �� � ����������2 $8)4*2 4�<41+ /������2 8+ d+ )�11$*+ -���� ���#� �������������� � � ��� ���� �� �� ����� ������+ � ����� / �����2 :2)d������� )0**2 �10�<�1$�+ �����2 c+ c+2 . -���� 2 9+ j+2 "�+ )�11>*+ ����� ������� �� �� ����� ������ � � �# ���� ��+ /,� )����� � ��7,������; ? �� �� >����2 1><�!+ -���2 �+ :+2 . -�����2 m+ �+ )�110*+ ��� ���=��=���=(��� � �����a ��� ���� �� � ����#��� �� � �������+ / ���� �� �������� � ����� ) ��?"# �*2 0><!$+ -�������2 +2 . -�����2 m+ )�1>!*+ d��=��=���=(��� � � � ������ ������� � ����� ���#� �+ � ����� / ���� �� �� � � ��2 $5)�*2 �1$<���+ ���������2 :+2 . 8����2 �+ )�11�*+ ��� ������� �# �� ���� � � ��� ��������� ����� � ������a ����� �� ���� �� ����� ������ �������+ � ����� ���� �� � � ��2 $82 $��<$44+ ��������2 :+2 . ������2 m+ )�11!*+ � ����� ������2 ����� ���#� �2 � � ��� ���� �� ������ � ��� ��� ������� �� �� ����� ������+ � ��;��� �� � � � � �����2 %2)8���� )4**2 1>1<����+ ��������2 :+ )�11$*+ � ����� ������ ���#� � � � �� ����� ������+ � ����� / �����2 :2)"# � )�**2 !�0<!04+ :� �2 k+2 l#�� �2 m+2 . ,������ 2 "+ )�11$*+ 8 �( ���� �� ��� � ��� ������+ � ����� / �����2 :2)0*2 ��$�< ��>!+ ��,� � . / ���� �� ���! ��� ,�! �* $0 1200$3 $%#4$5$ �$� federal open market committee meetings and stock market performance introduction literature review data and methodology results: the fomc effect conclusion references pii: s1057-0810(01)00080-4 from the editor the first article in this issue raises a very interesting question given the intense level of discussion on the future of social security. steven p. fraser, william w. jennings, and david r. king in their article entitled, “strategic asset allocation for individual investors: the impact of the present value of social security benefits” argue that social security benefits should be included in portfolio asset mix decisions. they start out by asking three questions: 1. “should investors consider social security when making asset mix decisions?” 2. “how should an investor determine the value of social security benefits?” 3. “how does proper valuation of social security affect the asset mix of a financial portfolio?” their analysis indicates that when social security wealth is included, there is an incentive for more stocks in the asset mix. gregory a. kuhlemeyer evaluates a recent product that the insurance industry has developed in his article entitled, “the equity index annuity: an examination of performance and regulatory concerns.” he raises issues about the performance of this type of annuity, its appropriateness for different investors, and the lack of regulation by the sec. is there a differences in black versus white households financial asset portfolio holdings? if so, should financial planners structure their product offerings differently? these two questions form the basis of an article by d. anthony plath and thomas h. stevenson entitled, “financial services and the african-american market: what every financial planner should know.” their analysis looks at the statistically significant differences by race and argues that planners need to recognize the difference in risk preferences when making recommendations. ralph r. trecartin jr. tests the book-to-market ratio in his article entitled, “the reliability of the book-to-market ratio as a risk proxy.” he finds that the ratio is a better predictor of return than cash flow, size, and sales growth. however, it is not reliable for periods of time that are less than 10 years and it is not a reliable proxy for risk. the last two articles provide two very different views of finance theory, based on reviewing literature from other fields. in an article entitled, “on time: contributions from the social sciences,” author barbara s. poole examines the anthropology and psychology financial services review 9 (2000) v–vi 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00080-4 literature on time. she discusses time in terms of culture, pace, environmental factors, and temporal orientation. an earlier version of this paper received the 2000 afs paper award made by the american association of individual investors. the last paper by douglas e. allen and elton g. mcgoun is called, “hedonic investment.” they argue that investing and consuming may be more similar than traditional theory would suggest when examined from a psychological or sociological prospective. they draw on marketing literature for their theoretical discussion and use a literary analysis of “the motley fool investment guide” to illustrate the concepts. both the poole and allen and mcgoun papers will hopefully provide the basis for further thought on alternative views of finance. vi from the editor / financial services review 9 (2000) v–vi 12 capstone as project-led problem based learning: theory and application in personal financial planning sarah asebedo1 and bryan gramse2 abstract the personal financial planning capstone course can be a complex and daunting learning experience for both students and instructors. resources exist to guide instructors through the content of capstone (the what); however, more consideration needs to be given to the how of capstone course delivery. this paper explores capstone through the lens of project-led problem based learning (pj-pbl), offering examples of course design, application, and assessment. research and discussion are needed to optimize capstone course delivery as the final class that prepares students for a rigorous profession. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation asebedo, s. & gramse, b. (2025). capstone as project-led problem based learning: theory and application in personal finance planning. financial services review, 33(4), 12-36. introduction completing a capstone class as a student is daunting, given the breadth and depth of the material necessary to facilitate mastery and synthesis of the cfp exam topic areas. teaching a capstone class is similarly formidable for the instructor responsible for guiding students through a comprehensive case in a way that generates deep learning while building students’ confidence in and connection to the profession they are about to join. fortunately, educators have developed textbooks that guide students through constructing a comprehensive financial planning case from start to finish. for example, grable et al. (2022) provided a client case example and illustrated how to work through all financial planning topic areas according to the cfp board’s financial planning process. the resources 1 corresponding author (sarah.asebedo@ttu.edu), texas tech university, lubbock, texas, usa. 2 texas tech university, lubbock, texas, usa. available today are excellent in defining the components of comprehensive financial planning and identifying how to construct this plan to meet the cfp board’s requirements. furthermore, the software, teaching deliverables, video, and excel resources that accompany textbooks significantly equip instructors with the content necessary for success. given these instructional resources, educators no longer need to define the content required for course delivery (the what). instead, educators face a fundamental and lingering challenge: how to deliver capstone within a course structure that achieves the cfp board’s learning objectives in a way that facilitates mastery and professional proficiency while retaining students’ motivation for the personal financial planning profession and confidence in https://creativecommons.org/licenses/by-nc/4.0/ mailto:sarah.asebedo@ttu.edu asebedo & gramse 13 their skills. additionally, a capstone course structure should generate an optimal instructor experience, resulting in a sense of meaning, mentorship success, and accomplishment while mitigating burnout risk. researchers can address the scientific and practical challenge of how to effectively deliver the cfp board’s capstone course through research that identifies evidence-based methods within the context of student and instructor outcomes. before researchers can achieve this result, we must first consider a theoretical framework to guide study design, empirical evaluation, and to synthesize the results within a cohesive theory-based map that facilitates practical implementation and opens future research channels. this paper takes the first step in addressing this gap by exploring project-led problem-based learning (pj-pbl) as a theoretical paradigm for the cfp board’s capstone course. capstone has synergy with pjpbl’s pedagogy, given its project-based case approach situated within a real-world professional context to solve a client’s comprehensive financial plan (“the problem”) using disciplinary knowledge, lending an applicable foundation for exploring a theoretical perspective to guide an instructional approach. therefore, this paper will address this primary research question: what are the potential advantages of using project-led problem based learning (pj-pbl) as a theoretical foundation for cfp board capstone courses? we extend our exploration to research evaluation as a necessary next step to move theory into practical application through this second research question: how can researchers evaluate cfp board capstone course effectiveness in a structured, researchable way? when viewing capstone through the pj-pbl lens, we can begin to identify the components and methods of primary data collection studies that produce the evidence needed to (a) evaluate pj-pbl as a theory applied to capstone, identify theoretical gaps, uncover other potential overarching theories, or develop new theories; (b) identify instructional methodologies and best practices to guide capstone course delivery; and (c) understand how course-design attributes shape student and instructor outcomes. we address these research questions by conducting a literature review, theoretically mapping pj-pbl concepts with capstone, proposing a research study design framework, and providing practice-based application examples with case studies. this paper makes a unique contribution to the financial planning pedagogical literature, as it expands beyond a content or instructional course guide to introduce a new paradigm to explore instructional design evaluation frameworks across various capstone implementation methods. this new paradigm has the potential to strengthen practice and education collaboration with a resource that the financial planning profession currently lacks. this research should be viewed as an early-stage contribution in a space where very little empirical or pedagogical research currently exists. the cfp capstone course is required as part of the cfp board’s education requirement, yet there is no consensus on best practices for structuring or evaluating it. our goal is not to provide definitive answers, but to initiate scholarly discussions and identify directions for future study. by mapping capstone through the pj-pbl theoretical lens, we hope to lay a foundation that invites continued discussion and development across programs. literature review financial planning pedagogy financial planning pedagogy has significant importance as it strives to provide students with the necessary skills and knowledge to become competent and well-rounded financial planners. the existing literature demonstrates the emergence of creative teaching approaches and emphasizes the requirement for enhancing financial education among students. goetz et al. (2005) discussed the inadequacies of the conventional approach to financial planning education, as it results in students facing a prolonged and costly journey toward entering the professional realm. graduating from university programs without prior financial planning experience is common among students, who often lack awareness of the practical aspects of the financial planning business (goetz et al., 2005). goetz et al. (2005) suggested incorporating case studies, simulations, and financial planning software to enhance the financial services review, 33(4) 14 financial planning curriculum. most recently, heymann et al. (2025) provided the first aggregate survey of cfp board-registered program directors, identifying curriculum design, experiential learning, practitioner involvement, and institutional support as critical components of financial planning pedagogy. building on this, zhang (in press) examined cfp board-registered programs within aacsb-accredited business schools and found significant differences in delivery format, program type, and institutional challenges, suggesting that program context strongly shapes pedagogy. subsequent research uncovers the limitations inherent in the traditional method of financial planning education. west et al. (2019) posited that educational institutions should focus on enhancing the development of abilities such as interpersonal communication, negotiation, marketing, and teamwork. the demand for these skills is rising among students and employers, indicating a greater appreciation for their significance (west et al., 2019). moreover, a survey conducted by weisz (2000) of students and employers engaged in an internship program showed that employers prioritized communication, initiative, and teamwork skills. in contrast, students acknowledged communication and initiative as their weakest competencies (weisz, 2000). this shift highlights the importance of individuals possessing both technical expertise and strong interpersonal competencies to succeed in the workplace. the skills employers value in graduates extend beyond mere theoretical knowledge, emphasizing the practical application of theory in real-world work scenarios (teale, 2013). applying theory in practical settings enables graduates to effectively solve problems, innovate, and adapt to dynamic work environments. by showcasing practical skills, students demonstrate their ability to make meaningful contributions to organizational goals and navigate challenges with competence and confidence. it is insufficient to possess knowledge about financial planning; one must actively apply this knowledge in practice (brau et al., 2015; jacob, 2016). including real-life experiences within the classroom setting and providing students with the opportunity to gain firsthand knowledge of the professional world are integral aspects of a personal financial planning course (martin, 2007). capstone course to address these limitations and increase the inclusion of practical financial planning experience, the cfp board introduced a capstone course requirement for all students who enroll in a program after january 1, 2012. the cfp board (n.d.) defined the capstone course as “a comprehensive financial plan development course created to enhance your knowledge, skills, and abilities” (para. 1). typically, students enroll in the capstone course as the final course in their financial planning curriculum or alongside other courses during their last term of study (martin, 2007). the capstone course is designed to integrate and apply the knowledge and skills acquired throughout the program, culminating in a student’s academic journey in financial planning. the capstone course does not fall under the category of an education course in the traditional sense, as it does not provide predefined learning materials. instead, it aims to assess candidates' proficiency in technical skills and ability to analyze and communicate findings across the curriculum through an oral and written report based on a financial planning case study (jackling & sullivan, 2007). the capstone course aims to prepare students with technical financial planning knowledge and the essential skills to integrate, apply, and communicate it effectively to their clients (grable et al., 2022). the cfp board (2019) outlines a 7step financial planning process in the development of a financial plan: 1. understand the client’s personal and financial circumstances 2. identify and select goals 3. analyze the client’s current course of action and potential alternative course(s) of action 4. develop the financial planning recommendation(s) 5. present the financial planning recommendation(s) asebedo & gramse 15 6. implement the financial planning recommendation(s) 7. monitor progress and update the financial planning process detailed in the fundamentals of writing a financial plan textbook by grable et al. (2022) illustrates the essential steps to begin and finalize a financial plan, emphasizing a thorough and comprehensive approach. when developing a financial planning course, it is essential to foster critical thinking and decision-making skills among students about the financial planning process (martin, 2007). in the capstone course, various pedagogical instruments can be employed to facilitate students' learning of the practical aspects of a financial planning practice. in a survey conducted among professors, the question of which pedagogical tool should be incorporated into a finance course yielded a prevailing response using a case study (thapa & chan, 2013). utilizing a case study in the capstone course offers students a valuable opportunity to delve into a scenario that emulates the dynamics of a typical financial planner-client relationship. solis (2018) conducted a study exploring innovative instructional strategies, including collaboration and the integration of multimedia. collaboration is a powerful instructional strategy promoting active engagement, critical thinking, and student social interaction. by working together, learners can develop essential skills, including communication, teamwork, and problem-solving. including multimedia in the classroom can enhance students' learning experiences by making them more captivating, efficient, and tailored to individual needs. another pedagogical approach involves allowing students to engage in peer financial planning. in a study by goetz et al. (2011), researchers followed a program in which students provided financial planning services to fellow students and faculty. implementing this peer financial planning program enabled students to acquire essential practical skills and apply their theoretical knowledge to real-world situations (goetz et al., 2011). maurer and lee (2011) conducted a similar study comparing the efficacy of peer financial counseling with a semester-long course. their findings indicated that peer financial counseling could produce comparable levels of financial literacy improvement in much smaller periods. financial planning pedagogy is evolving to incorporate innovative, hands-on teaching methods that support the needs of various student populations. there is a consensus on the importance of peer-based learning, the use of case studies and mock sessions, and adapting the curriculum to reflect current realities in financial planning. however, gaps remain in the availability of effective pedagogical strategies for capstone classes and the integration of financial planning education into broader academic programs. in conclusion, financial planning capstone courses are crucial for preparing students for the professional world by providing opportunities to apply their knowledge in complex, real-world scenarios. these courses are designed to be integrative, reflecting the student's entire educational financial planning journey. theory project-based learning (pjbl) project-based learning (pjbl) reflects a studentcentered approach to the education environment that fosters learning by doing and applying ideas based on real-world activities similar to those in the professional working environment associated with the educational curriculum (du & han, 2016; krajcik & blumenfeld, 2014). in pjbl, students take an active and engaged approach to learning through problem-solving and constructing their understanding through a project with a meaningful and relevant problem (krajcik & blumenfeld, 2014). according to krajcik and blumenfeld (2014), the concept of pjbl rests upon a body of learning science theory that indicates students learn more deeply with an enhanced personal connection and investment with the material generated through active inquiry, incorporating (a) active construction, (b) situated learning, (c) social interactions, and (d) cognitive tools. guo et al. (2020) articulated the overarching purpose and composition of pjbl well in this summary: “this creation process requires learners to work together to find solutions to authentic problems in the process of financial services review, 33(4) 16 knowledge integration, application, and construction. instructors and community members (e.g., clients), normally as facilitators, provide feedback and support for learners to assist their learning process” (p. 2). guo et al. (2020) conducted a systematic literature review of 76 empirical studies that investigated student measures and outcomes (cognitive, affective, behavioral, and artifact performance) within a pjbl higher education environment. guo et al. (2020) found preliminary evidence for a positive impact of pjbl on students’ content knowledge, learning strategies, skills, motivation, and product quality; however, they noted that more research is needed to assess the effectiveness of the pjbl structure. pjbl is grounded in several learning theories. first, constructivism is central to pjbl as it positions students at the center of the learning process, where they construct learning and meaning through individual or social interactions (narayan et al., 2013). constructivism posits that learners contribute their prior knowledge and experiences to the learning process through an active interaction rather than a passive relay of information from the instructor to the learner. more specifically, pjbl incorporates social constructivism to recognize the influence of society and social interactions on the learning process. second, cognitivism informs pjbl as a project is positioned atop an underlying body of knowledge that the learner has acquired over time and must now retrieve and place into working memory to invoke planning in response to the project as an external stimulus, construct solutions from this knowledge, and expand their existing knowledge framework to generate new strategies. these active internal mental information processing activities, combined with the learner’s experiences and emotions, are key elements of cognitivism (bruning et al., 2011; paciotti, 2013). under the cognitivism umbrella, bandura’s (1986) social cognitive theory reiterates the role of the social environment in influencing cognitive processing. in social cognitive theory, learning extends beyond internal cognitive processes and reactions to external stimuli to include an interactive engagement with the social context that generates the learning environment. furthermore, social cognitive theory emphasizes human agency, where the learner’s internal motivation, decisionmaking, actions, and outcomes are causal inputs to the learning process and environment. selfefficacy is a central factor contributing to this bidirectional system. self-efficacy “refers to the beliefs in one’s capabilities to organize and execute the courses of action required to produce given attainments” (bandura, 1997, p. 3). in other words, self-efficacy embodies the belief that one's actions will lead to success and that failures are temporary obstacles. self-efficacy influences various aspects of personal growth and progress, including goal setting, resilience in the face of adversity, thought patterns, stress levels, and the risk of depression. further, physiological and affective states, enactive mastery experience (past successes), vicarious experiences, and verbal persuasion combine to shape self-efficacy. through the lens of capstone, social cognitive theory and self-efficacy enhance the actionoriented learner, emphasizing the importance of self-efficacy in advancing through a significant project. self-efficacy also provides deeper insight into the elements of the social learning environment that shape self-efficacy and learning outcomes, such as providing opportunities to experience successes (enactive mastery experiences) and demonstrating how others have succeeded in the past (vicarious experiences). third, situated cognition theory brings a real-life environment to pjpl by positing that students develop knowledge within a situated sociocultural context that is inseparable from the learning process (jenlink, 2013). through situated cognition, students set individual goals in relation to themselves and others, learn to think like professionals, and experience how to solve real-world problems through a guided growth trajectory, moving them from novice to independent expert (jenlink, 2013). connection to practice is critical for situated cognition to provide students with opportunities to participate and engage with community experts, observing how they solve daily problems within a situated context and experiencing real-world problemsolving firsthand. the importance of how capstone is facilitated becomes clear through the asebedo & gramse 17 lens of situated cognition theory, as it transforms a comprehensive curriculum academic exercise into a real-world, applied training model for professional practice. fourth, problem-based learning theory has high synergy with pjbl, and experts have proposed integrating them into a hybrid model of projectled problem-based learning (pj-pbl), emphasizing problem-solving within a real-world project to facilitate skill acquisition and produce tangible, work-related artifacts or services (hanney & savin-baden, 2013). problem-based learning places the student at the center of the learning process, with the teacher assuming a facilitator role to solve ill-structured and authentic problems (fredrickson et al., 2013). hanney and savin-baden (2013) noted that while there are various approaches to problem-based learning, a common thread across all models is that the problem reflects professional practice or a real-world situation within the context of disciplinary knowledge. furthermore, problembased learning often incorporates small-group work, thereby also contributing a social element to the learning process. hanney and savin-baden (2013) noted that the amount of guidance and scaffolding within problem-based learning can vary greatly, but that the level of structure needed depends on the learner’s prior experience with problem-based learning. thus, those with minimal previous experience require a high level of support structure, while those with extensive experience require very little. project-led problem-based learning (pjpbl) given the synergy of problemand project-based learning, we employ the theoretical background of project-based learning (krajcik & blumenfeld, 2014) and the integrated approach of project-led problem-based learning (pj-pbl), as proposed by hanney and savin-baden (2013), as the theoretical map guiding capstone. based on this theoretical foundation, pj-pbl typically contains these components: 1. a driving question or problem that students actively engage with to generate (i.e., construct) understanding by producing ideas and solutions to this real-world and nonlinear problem. 2. situated inquiry frames this question or problem within an authentic and real-world and work-related context, reflecting the professional environment and disciplinary knowledge to facilitate the student’s connection with the value and meaning of the tasks and activities asked of them. 3. collaborative activities (students, teachers, professionals, and community members) facilitate social interaction to share, use, and debate ideas within a learning community. 4. scaffolding provides the structure and tools necessary for students to stretch beyond their current capacity. within this scaffolding, cognitive tools and learning technologies (e.g., software, equipment) extend learning to tasks that would not otherwise be possible, while the project and course structure guide the learner through problem solving, depending on their prior experience with problem based learning. 5. learning artifacts represent students' constructed knowledge (e.g., models, reports, recordings, programs). pj-pbl applied to capstone capstone is a quintessential example of pj-bl in higher education. students engage in active problem-solving for a personal financial planning client case where they must use their acquired knowledge across multiple topic areas to generate new ideas and construct solutions that improve the client’s financial health and capability to achieve their goals (a driving question or problem). this problem-solving is situated in an authentic and real-world context, as the student must solve the comprehensive client case by using real-world laws, regulations, and financial resources (situated inquiry). capstone is the keystone class connecting education with the professional environment. problem-solving with a comprehensive client case reflects the core focus of financial planning work, thereby giving value and meaning to the assignments, tasks, and activities asked of students (situated inquiry). the instructor can also customize the various activities within a capstone class to reflect professional activities, such as facilitating a client financial services review, 33(4) 18 meeting and presenting financial planning concepts. the capstone class can easily incorporate collaborative work with students, teachers, and professionals from the financial planning community (collaborative activities). for example, students can work in groups to discuss, collaborate, and negotiate client recommendations, much like a professional environment where a team works together for the benefit of the client. experienced financial planners can serve as a valuable resource for product quotes, provide feedback on planning strategies, or even act as a live client. additionally, by structuring the grading process so that students complete portions of the plan and submit their work periodically, the instructor can offer feedback and promote discussion of ideas along the way, thereby deepening learning and expanding their thought repertoire. this studentcentered course structure facilitates learning and provides essential scaffolding as students navigate problem-solving within a setting likely new to them. the extent of problem-based learning incorporated into the underlying curriculum will be a significant driver of the scaffolding and guidance needed within capstone to make problem-based learning effective. capstone is an excellent class for incorporating cognitive tools and learning technologies to facilitate learning beyond the textbook (scaffolding with cognitive tools and learning technologies). using software enables more complex analysis and provides access to sophisticated statistical tools to assess client outcomes. last, students can present several learning artifacts resulting from a capstone course, including a written comprehensive plan, software reports, presentations, peer feedback, and facilitating client interactions either live or in video form (learning artifacts). as described, capstone naturally lends itself to a pj-pbl approach. this theoretical connection presents several potential advantages (research question 1). framing capstone within a pj-pbl paradigm helps the instructor and student gain a clear picture of the goals, purpose, and process of this class. this perspective can potentially mitigate stress and anxiety often accompanying a large project from both the instructor's and learner's perspectives. pj-pbl creates a theoretical foundation from which to develop a syllabus and schedule, create a supportive and rigorous course structure, choose which technologies to incorporate, consider how to develop collaborative opportunities both internally (teacher and student) and externally (professionals), and determine optimal assessment tools (e.g., written plan, client meeting recording, etc.). evidence-based learning principles with pj-pbl as the foundational theory guiding the teaching pedagogy for capstone, we can then consider additional learning-based principles within the course framework. these principles, as outlined by halpern and hakel (2003), facilitate long-term retention and transfer to the professional environment. they are particularly salient for capstone as a final course at the top of a layered curriculum intended to synthesize a wide array of content: (a) practice at retrieval, (b) varying the conditions under which learning takes place, (c) alternate formats (present and represent), (d) prior knowledge and experience, (e) learning as influenced by the learner (and instructor), (f) experience alone is a poor teacher (the need for cases, even for experienced practitioners), (g) understanding as an interpretive process with students as active participants, (h) the act of remembering influences future recall, (i) less is more, and (j) what learners do determines what and how much is learned, remembered, and recalled. practice at retrieval practice at retrieval involves repeatedly recalling prior learned information under new conditions and circumstances over time to answer new questions and solve new problems, thereby effectively transferring the prior information to a new context (halpern & hakel, 2003). halpern and hakel (2003) noted that practice at retrieval should be facilitated with minimal cues, allowing the learner to recall and retrieve information independently as much as possible, thereby becoming fluent with the acquired knowledge. capstone naturally facilitates practice at retrieval because it provides a new frame (the client case) through which the student must recall and apply the previously learned financial planning asebedo & gramse 19 information. the structure of capstone within the semester can also facilitate retrieval by creating multiple submission opportunities, where learners can submit their work, receive feedback, and revise it for the client’s current situation, and then again for the recommendations. group work and live discussions also facilitate this process. practice at retrieval during the final capstone class is essential for the long-term and fluent transfer of financial planning information and skills into the professional work environment. varying the conditions halpern and hakel (2003) outlined the benefits of varying learning conditions. while more challenging, this approach generates more retrieval cues than a learning environment where the learning conditions are consistent or limited. as a pj-pbl class, capstone can readily vary the learning conditions in various ways, such as constructing the written plan, using software, working in teams, incorporating professionals into the learning environment, developing presentations, and facilitating client meetings. the main idea is to integrate different types of problems and contexts (e.g., writing the plan and then facilitating a client meeting with the plan) within the course, so that the information can be recalled and applied in various ways. using software is an excellent way to vary learning conditions, as software programs require different data entry structures and produce new report layouts that students may not be accustomed to seeing in textbooks. alternate formats incorporating alternate formats entails asking learners to process information from one format to another within the learning environment (halpern & hakel, 2003). for example, the instructor could ask the learner to present written information verbally or visually. capstone can incorporate this feature by asking students to write their financial plan, utilize visual tools within the plan (e.g., graphs and charts generated by software or excel), and present that plan to a client through an interactive client meeting. prior knowledge and experience learning outcomes reflect the learner’s existing knowledge, experience, and understanding (halpern & hakel, 2003). therefore, instructors should assess prior knowledge and understanding at the beginning of class and evaluate progress, as well as any regression to prior knowledge levels. this approach is easily accomplished in a capstone class by having students submit portions of the written plan throughout the semester, allowing the instructor to have a clear picture of their existing knowledge, ability, and learning progression. learner influenced learning similar to prior knowledge and experience, students’ internal beliefs (e.g., self-efficacy) about how they learn and what they are capable of can affect the learning process, especially when the assigned tasks are challenging and require more effort than other learning experiences they might have had. given the multifaceted nature of capstone, it is likely that students will experience doubts about their learning capabilities at some point along the way. instructors teaching capstone have an opportunity to deliver clear feedback and identify any beliefs that are counterproductive to the learning process through interactive class discussions and periodic submissions of sections of the financial plan throughout the semester. experience as a teacher halpern and hakel (2003) highlighted that experience alone is a poor teacher, often misaligning subjective and objective knowledge. learners with extensive experience might have erroneous beliefs about mastering a complex topic. this point is salient for graduate programs where the learner is an experienced financial planning practitioner. professional financial planners often possess specific and in-depth knowledge of issues relevant to their client base; however, they may lack a broad understanding of the topic. thus, incorporating case studies that push the boundaries of student learning beyond their experience, combined with systematic feedback throughout the case development process, is essential, even for experienced professionals. understanding understanding is the outcome of an interpretive process driven by active student participation financial services review, 33(4) 20 where the learner is asked to engage, interpret, and interact with the information (halpern & hakel, 2003). based on this definition, traditional lectures, exams, and multiple-choice questions are not the most effective tools for facilitating deep understanding. while many of the underlying classes in a financial planning curriculum might utilize these more traditional methods, capstone can promote deep learning and understanding as a pj-pbl class that facilitates this active student engagement, interpretation, and interaction with financial planning knowledge, problems, solutions, and the professional financial planning client context. remembering influences future recall it is essential to recognize that what we ask students to retain through learning and assessment will significantly impact what they recall in the future and what they selectively forget (halpern & hakel, 2003). this learning principle is essential to consider when constructing exams and quizzes. for capstone, this applies to the construction of the client case. it might be tempting to build an elaborate case with nuance and unique circumstances; however, this learning principle would suggest a more basic case that emphasizes synthesis and multiple recall points across the central financial planning topics, which would create a more robust, broader, and flexible knowledge foundation that the learner will carry forward into their professional career. less is more building on the prior point, thinking carefully about what students are asked to remember is combined with the notion that less is more in teaching and educating future financial planners. creating a comprehensive client case that encompasses every technical concept in financial planning is impossible. therefore, it is essential to consider the goal of capstone. the cfp board (n.d.) defined the capstone course as “a comprehensive financial plan development course created to enhance your knowledge, skills, and abilities” (para. 1). we posit that this definition could be expanded to include a goal to develop a deep understanding of foundational financial planning principles and to synthesize this understanding across financial planning topic areas, resulting in an integrated assessment of knowledge, skills, and abilities. defining the goal of capstone is essential to determining the content to include when designing the client case and expectations for student recall and assessment. what learners do last, we end with the learner in mind – what they do, they learn. according to halpern and hakel (2003), we must carefully consider the tasks and activities we ask students to engage in, as these will directly influence the learning content and quality, as well as the flexibility of this knowledge in future contexts and situations. halpern and hakel (2003) emphasized that teaching is less about what the professor does; instead, it is far more important to consider what the learner is asked to do. for capstone, this is an easy principle to apply as it is heavily focused on learner application within the context of a case study, consistent with pj-pbl underlying theory. in fact, we posit that there should be very few lectures in a capstone class and suggest that class sessions focus directly on the case and class discussions. we propose that the capstone instructor serves in the capacity of a mentor and guide, as in a pj-pbl course, with a focus on active and student-centered learning. research considerations viewing capstone through a theoretical and practical lens provides insights into how researchers can evaluate course effectiveness in a structured and research-based manner (research question 2). next, we present a preliminary research design framework, along with its key considerations, based on our pj-pbl theoretical exploration. study type a research study assessing capstone outcomes would benefit from a multi-institutional primary data collection effort. cfp board-registered financial planning programs vary significantly across the u.s. and the globe. it is likely fruitful to identify how and in what ways different perspectives across countries, resources, program structure (e.g., degree level, certificate, modality), college home (e.g., human sciences, business, ag econ, etc.), and professional asebedo & gramse 21 partnerships (e.g., employer and alum synergy) relate to outcomes of interest. this collective effort also presents an opportunity to establish a multi-institutional dataset to investigate various capstone approaches and outcomes across time. significant funding would be needed to facilitate such an effort, and initial research will likely require small pilot studies within institutions as a first step. data type capstone generates various learning artifacts that suggest a mixed-methods study type would be applicable and practical, utilizing a combination of quantitative and qualitative data to provide evidence for multiple outcomes of interest, such as knowledge growth by topic area, topic synthesis, self-efficacy, instructor burnout, and professional development. these different data types offer opportunities for various analysis strategies, depending on the sample size (e.g., descriptive statistics, t-tests, multiple regression, structural equation modeling) and time orientation (e.g., cross-sectional, longitudinal). sample characteristics capstone is taught in a wide array of formats, levels, and modalities across institutions where the theoretical concepts would likely necessitate different applications. for example, an in-person, synchronous, undergraduate-level class functions very differently from one that is delivered 100% online, asynchronously, and at the graduate level. therefore, samples would need to be drawn across various course modalities (in-person, synchronous online, asynchronous online) and degree levels (undergraduate and graduate) to account for differences in applied theoretical concepts across various course characteristics. measures outcomes multiple outcomes become relevant when exploring capstone through a theoretical lens. first, the cfp board’s (2021) capstone learning outcomes provide a foundational starting point for assessment, covering comprehensive knowledge and understanding, as well as effective oral and written communication. “upon completion of this course, the student will be able to: 1. demonstrate a comprehensive understanding of the content found within the financial planning curriculum and effectively apply and integrate this information in the formulation of a financial plan. 2. effectively communicate the financial plan, both orally and in writing, including information based on research, peer, colleague or simulated client interaction and/or results emanating from synthesis of material. 3. analyze personal financial situations that includes both qualitative and quantitative information, evaluating clients’ objectives, needs, and values to develop an appropriate strategy within the financial plan. 4. demonstrate logic and reasoning to identify the strengths and weaknesses of various approaches to a specific problem. 5. evaluate the impact of economic, political, and regulatory issues with regard to the financial plan. 6. apply the cfp board code of ethics and standards of conduct to the financial planning process.” beyond course-based knowledge and skill learning outcomes, pj-pbl theory suggests that it is relevant to consider psychological outcomes for both the instructor and the student. for example, this could include student-centered outcomes such as self-efficacy growth (social cognitive theory), professional identity formation (situated cognition theory), confidence in client interactions and effective classroom-to-career transition (situated cognition theory), problemsolving capability (problem-based learning), ability to navigate group-based problem solving and negotiating conflict in professional relationships (social constructivism). from the instructor’s perspective, self-efficacy (social cognitive theory) is an important outcome of interest as the instructor role shifts to that of a facilitator guiding students' problem-solving within the context of ill-structured and authentic problems (problem-based learning), creating intentional ambiguity in the learning environment that activates and develops students’ problemfinancial services review, 33(4) 22 solving activities and cognitive processes. likely, the greater the instructor’s self-efficacy in assuming a facilitator role, the stronger the student’s outcomes and the more satisfied the instructor, which could lead to a stronger sense of meaning and purpose, with less potential for burnout. this example illuminates how theory might introduce more complex research questions and hypotheses. this example illustrates a potential mediation model where a causal path exists from instructor self-efficacy to student outcomes, which in turn influences instructor psychology and retention. key theoretical indicators other potential key theoretical indicators include the quality and impact of group social interactions (social constructivism, social cognitive theory, problem-based learning), quality and type of professional practice synergy (situated cognition theory), learning artifact efficacy (e.g., oral vs. written, client meetings; problem-based learning), and scaffolding for course structure and technology (problem-based learning). baseline variables additionally, theory suggests that baseline factors reflecting prior knowledge and experiences are relevant to student and instructor outcomes. for example, constructivism emphasizes the influence of previous knowledge and experiences on active student learning. similarly, the extent to which instructors incorporate a problem-solving approach into the underlying curriculum will significantly affect the efficacy of the capstone course structure and learning outcomes (problembased learning). furthermore, the instructor's background, qualifications, and practice experience likely contribute to their self-efficacy when engaging in a project-led, problem-based class. best practices to close, we have identified a set of best practices through a convenience sample of faculty interviews with professors at arizona state university and ball state university, each of whom has extensive experience teaching the capstone course, as well as the authors’ own experiences teaching the capstone course at texas tech university. while not intended to be a comprehensive survey, this practice-informed approach highlights strategies that have been applied and deemed anecdotally effective in various financial planning education programs. case study development case studies incorporating a variety of interactive elements encourage students to utilize their critical thinking skills and investigate alternative methods in developing their financial plans. case studies designed to test students' limits should motivate them to tackle problems from multiple angles, applying both analytical and inventive thinking. to aid students in understanding the case study, the instructor can provide standard financial planning materials that a real-world client would typically use. this includes statements of income, expenses, insurance coverages, assets, liabilities, and previous tax returns. an initial case narrative can also be provided to introduce the students to the client's lifestyle, goals, priorities, and ideal retirement scenario. to ensure students are prepared for the range of clients they may serve, capstone courses (in partnership with the underlying curriculum) can expose them to diverse case studies that reflect different lifecycle stages, socioeconomic circumstances, and cultural backgrounds. for example, one case may feature a high-income pre-retired household, another a young professional with student debt, and another a middle-aged family balancing career demands and education funding. by working across multiple client types, students gain experience addressing varied goals, constraints, and planning opportunities. see appendix a for sample onepage case studies that illustrate how instructors might frame cases to capture different lifecycle and socioeconomic contexts. clients often unintentionally leave out material information when working with a financial planner in real-life situations. to simulate realworld scenarios, important details of the case may be omitted. as a result, students need to ask about the data to uncover any additional valuable information related to the client case. similar to real-life situations, students will not initially be aware of what is missing and will discover it as asebedo & gramse 23 they analyze the client's present circumstances. from an instructor's perspective, it is beneficial to convey the learning purpose behind these information gaps so that students can proactively engage in filling them as part of the learning process, rather than viewing the gaps as mistakes needing correction. professional application throughout the course, students can replicate professional practice by applying the same processes they will use in their professional careers. upon the initial introduction of the case study material, students can start the seven-step financial planning process. students will need to allocate sufficient time to comprehend and evaluate the client's present circumstances and objectives. completing the initial analysis separately for each major section of financial planning, including cash flow and net worth, education, retirement, investment, insurance, taxes, and estate planning, provides beneficial scaffolding within the course structure. presenting the case study within the seven-step financial planning process maximizes learning principles within a pj-pbl structure, where students are introduced to a methodical approach to analyzing and formulating a financial plan from the outset. this approach also keeps financial planning simple for students, so they do not feel overwhelmed by the process and can follow it step-by-step. by breaking down the overall project into smaller tasks as a form of scaffolding, students can build their confidence and ultimately complete a comprehensive financial plan. after students have submitted summaries of the client's current situation, they can begin generating recommendations to maximize the client’s likelihood of achieving their goals. as a scaffolding application, students can submit a written summary of their recommendations, along with justification for each central topic area, and receive instructor feedback to facilitate learning and growth before submitting their final plan. for the final part of the course, students will work to synthesize the current situation and make recommendations to ensure that each component of the financial plan works together effectively. in this phase, they can incorporate feedback from previous scaffolding to process the plan at a deeper level and strengthen the final product. students will also create an action plan to guide the clients in implementing the financial plan. by following this structured approach, students will not only gain practical experience in financial planning but also develop the necessary skills to address the needs of their future clients effectively. to enhance professional interactions within a situated context, instructors can incorporate guest speakers, such as retirement specialists, estate planners, or tax experts, to provide students with exposure to practitioner perspectives and specialized areas of financial planning. professionals could also participate as a reviewer or client for mock client meetings. in the case design phase, professionals can lend their expertise in constructing a case that addresses relevant problems experienced in practice, communicate the case characteristics, and facilitate a discussion of recommendations. the goal of these interactions is to help students connect classroom learning with professional financial planning practice through a situated learning environment. collaboration financial planning is commonly carried out within group settings, making it beneficial to expose students to group work dynamics. in the capstone project, forming small groups consisting of two or three individuals facilitates active participation in all aspects of the project while engaging in collaborative efforts. by working together, students can combine their ideas and analyses to develop a comprehensive financial plan. through active discussions and debates, the group can further refine its plan, ensuring that all aspects of the problem are thoroughly examined and addressed. this approach provides students with practical skills in a simulated group financial planning environment, bringing in the social learning components of pj-pbl. it’s important to note that group work might not be feasible in all settings or modalities. for example, incorporating group work in an asynchronous online class of working professionals from various time zones will be more challenging and likely require a different financial services review, 33(4) 24 implementation strategy than in an in-person cohort-based class. in some cases, group work may not be feasible, and social interaction about the case may require a very different format. effective collaboration between the instructor and students also plays a pivotal role in this process. as students submit work periodically throughout the semester, the instructor provides timely feedback to guide them and facilitate their problem solving. before students proceed with crafting recommendations, instructors can provide feedback to promote student learning and growth in their understanding of the client’s current situation. for example, if students have submitted their analysis of the existing retirement planning situation, feedback is provided to help them deepen and strengthen their understanding of the case before developing retirement planning recommendations. this feedback loop ensures that students stay on track and have a solid foundation before progressing to the next stage of the financial planning process, which theoretically builds their self-efficacy as a key outcome. software integrating software and technology plays a crucial role in financial planning. students need to have hands-on experience with a range of technological tools to support and gain proficiency in their financial planning development. this experience can involve foundational tools (e.g., excel) in addition to specialized financial planning software programs, such as emoney, moneyguidepro, and rightcapital. furthermore, incorporating specialized financial planning software strengthens the classroompractice connection, generating a realistic, situated context. students need to familiarize themselves with the functionalities of these programs to navigate through the intricacies of professional financial planning effectively. by leveraging specialized software, students can extend their learning beyond what is possible with basic tools to construct interconnected models reflecting the topic areas (e.g., multi-year cash flow and balance sheet, education funding, estate flowchart, income tax projection, etc.), conduct monte carlo simulations, and assess the impact of various recommendations on the overall financial plan. familiarity with these software tools can enhance students' ability to craft comprehensive and effective financial plans in their future professional endeavors and ease their transition from the classroom to their careers. financial planning tools scaffolding with strategically placed financial planning tools can deepen students' understanding and enhance their ability to demonstrate acquired knowledge. for example, students can derive significant benefits from utilizing templates to initiate their case work, similar to the practice where professionals often employ practical tools to enhance their effectiveness and efficiency, allowing them to begin planning at a much deeper level. for example, a templated excel sheet may be provided, serving as a valuable tool for organizing client financial information, including cash flow statements, net worth statements, financial ratios, tax projections, and other relevant data. additionally, having a written financial plan template available can be advantageous, as it allows students to allocate their time and effort toward the content of their final plan rather than being preoccupied with formatting details. this template can provide a clear structure, indicating the specific sections where different components of the plan should be included. by utilizing these scaffolding techniques and learning artifact outputs, students can enhance their understanding of financial planning concepts and improve their ability to create comprehensive financial plans successfully. it’s important to note that the type and amount of scaffolding provided through templated material may vary based on prior student experiences and the extent to which the underlying curriculum has already offered that scaffolding to prepare students for a project-led, problem based learning environment. role playing and simulation role playing adds a powerful element to the capstone course by immersing students in the client-planner relationship. students can alternate roles as financial planner, client, and observer, practicing essential skills such as communication, asebedo & gramse 25 empathy, and presentation. live or recorded mock client meetings, presentations of financial plan recommendations, and simulated client objections create authentic practice opportunities. role playing can also be extended by assigning students different client personalities, goals, or communication styles. this variation challenges students to adapt their approach, build rapport in diverse situations, and strengthen their ability to respond to different client needs. role playing also highlights the importance of professional demeanor and adaptability, competencies that are difficult to teach through lectures alone but are essential for building client trust and rapport. incorporating professionals in these role-playing activities is another way to enhance the situated context within capstone. example schedule here, we provide an example schedule of how instructors might consider structuring a capstone class that aligns with the purpose and nature of a pj-pbl-type course. the schedule is designed with a three-part structure: (a) part i: current situation, (b) part ii: recommendations, and (c) part iii: the final plan – putting it all together. the overall schedule flow and rationale for the ordering of topic areas are outlined in the following charts. while neither the current situation nor recommendations are solved in a 100% linear fashion, the current situations are more siloed and linear than recommendations; however, there is some dependency with the retirement and risk management sections due to the need to have a working retirement projection (defined by the monte carlo level) before it is clear how the current insurance policies adequately cover the plan. for example, a second home goal modeled under the current situation might be reduced or eliminated in recommendations to achieve a successful outcome based on the client’s priority for that goal, which could have a significant impact on the need for life insurance. creating recommendations involves heavy synthesis and integration across the plan, hence the arrow across all boxes. overall, flexibility and clear communication from the instructor are essential to manage the parts of the planning process that are fluid and dependent on other areas, while providing feedback and awarding points to help students learn and grow throughout the course. in terms of structuring the class, the following schedule (see figure 1 and table 1) provides an example of a 16-week class based on the flow outlined above. the schedule can easily be adjusted to an eight-week class by combining topic areas that have synergy. this schedule does not account for holidays and periodic class cancellations due to conferences and other activities that are typical within academic calendars. furthermore, this example schedule illustrates how instructors might frame periodic submissions as optional check-in points for feedback; however, these optional interim submissions are easily converted to required submissions for points if desired. based on the authors’ experience, optional submissions tend to work well when the scaffolding need is lower, and flexibility needs must be maximized. on the other hand, interim submissions for points are effective when points are needed to strengthen scaffolding (i.e., students are more likely to submit) and flexibility needs are a lesser priority. financial services review, 33(4) 26 figure 1. course flow current situation recommendations cash flow & net worth •understand baseline financial health and financial capability. investments •investment planning is foundational for the retirement projection. education & retirement • retirement is key for risk modeling, and education is a major goal within the context of retirement planning. risk •determining asset protection, risk exposure, and liquidity needs informs income tax and estate. taxes & estate • tax and estate modeling depends on the completion of other areas. investments •great starting point for building confidence and mastery and less dependent on other areas. affects all subsequent modeling. taxes •taxes and investments are highly interrelated; having these sections together or close together is efficient. education, retirement, & risk •connected and efficient to think through close together with retirement/education then risk. if time, pair retirement & education together followed by risk at a later due date. estate •efficient near the end after addressing underlying areas. cash flow & net worth •financial health and impact on cash flow are clear at end. asebedo & gramse 27 table 1. sample schedule dates (mondaymonday) topic what to do what to submit optional what to submit required module 1: current situation 8/25 9/1 course introduction read: syllabus and schedule client case ch. 2 (writing a financial plan) e-money: certification & training videos 9/1 9/8 current situation cash flow & net worth planning read: ch. 6 (cash flow & net worth) submit written plan for preliminary feedback (no points) by 9/8 @ 11:59 pm (late not accepted)  current situation: cash flow & net worth planning submit for course points and grade by 9/8 @ 11:59 pm  personal introduction (discussions) 9/8 9/15 current situation investments ch. 13 (investments) submit written plan for preliminary feedback (no points) by 9/15 @ 11:59 pm (late not accepted)  current situation: investments submit group names no later than 9/15 @ 11:59 pm (no course points, but groups are locked in from this point forward). table 1 continued on next page. financial services review, 33(4) 28 table 1 continued. dates (mondaymonday) topic what to do what to submit optional what to submit required 9/15 9/22 current situation education current situation retirement ch. 14 (education) ch. 15 (retirement) submit written plan for preliminary feedback (no points) by 9/22 @ 11:59 pm (late not accepted)  current situation: education  current situation: retirement n/a 9/22 9/29 current situation risk ch. 8-12 (risk all sections) submit written plan for preliminary feedback (no points) by 9/29 @ 11:59 pm (late not accepted)  current situation: risk 9/29 10/6 current situation income tax ch. 7 (income tax) submit written plan for preliminary feedback (no points) by 10/6 @ 11:59 pm (late not accepted)  current situation: income tax n/a 10/6 10/13 current situation estate ch. 16 (estate) submit written plan for preliminary feedback (no points) by 10/13 @ 11:59 pm (late not accepted)  current situation: estate 10/13 10/20 current situation o client meeting covering the current situation finalize current situation sections and record client meeting. extra credit points! submit by 10/20 @ 11:59 pm  emoney fundamentals certification – early submission for 15 extra credit points! submit for course points and grade by 10/20 @ 11:59 pm  client meeting video covering the client’s current situation table 1 continued on next page. asebedo & gramse 29 table 1 continued. dates (mondaymonday) topic what to do what to submit optional what to submit required module 2: recommendations (extend current situation work to add recommendations) 10/20 10/27 recommendations investments revisit the book as needed. submit written plan for preliminary feedback (no points) by 10/27 @ 11:59 pm (late not accepted)  recommendations: investments n/a 10/27 11/3 recommendations income taxes revisit the book as needed. submit written plan for preliminary feedback (no points) by 11/3 @ 11:59 pm (late not accepted)  recommendations: income taxes n/a 11/3 11/10 recommendations education retirement revisit the book as needed. submit written plan for preliminary feedback (no points) by 11/10 @ 11:59 pm (late not accepted)  recommendations: education  recommendations: retirement 11/10 11/17 recommendations risk submit written plan for preliminary feedback (no points) by 11/17 @ 11:59 pm (late not accepted)  recommendations: risk n/a 11/17 11/24 recommendations estate submit written plan for preliminary feedback (no points) by 11/24 @ 11:59 pm (late not accepted)  recommendations: estate n/a table 1 continued on next page. financial services review, 33(4) 30 table 1 continued. dates (mondaymonday) topic what to do what to submit optional what to submit required module 3: the final plan—putting it all together (current situation + recommendations + front summary pages) 11/24 12/7 (sunday) putting it all together current situation + recommendations add: o front summary pages o cash flow & net worth planning recommendations o client meeting covering recommendations revisit the book as needed. n/a submit for course points and grade by 12/7 @ 11:59 pm  emoney certification certificate of completion  comprehensive written plan o all current situation sections o all recommendations sections (including cash flow & net worth) o all front summary pages 12/8 (monday) submit for course points and grade by 12/8 @ 11:59 pm  the final comprehensive plan recommendations client meeting  group peer evaluations due 12/8 – 12/11 teaching team grading 12/11 final grades due @ noon asebedo & gramse 31 conclusion the capstone course serves as a central training ground in the educational trajectory of future financial planners, offering a comprehensive and practical learning experience that combines theoretical knowledge with real-world applications. utilizing cases, facilitating collaboration and professional interactions, and incorporating financial planning software enhance student preparedness to adapt and thrive in the dynamic and constantly changing financial planning profession. this opportunity allows the capstone class to solidify its position as a foundational bridge, connecting the training gap between theoretical knowledge and practical skills. through this course, students are exposed to the financial planning environment they’ll encounter in their careers, preparing them to confidently transition into their future professional roles, equipped with foundational experiences and skills to excel in the field. in this paper, we explored the delivery of capstone through the lens of a project-led problem based (pj-pbl) approach, considered ways to assess capstone efficacy through a research process, and introduced relevant learning-based principles, while offering examples of course design, schedule, and application considerations. rather than claiming to provide a definitive model, our intention is to frame this work as an early-stage contribution in an underexplored area. the goal of this paper is to spark further conversation and research to increase resources for teaching capstone, specifically addressing the primary gap that needs to be filled to complement the robust resources currently available for the what of capstone. by positioning this paper as a starting point, we aim to encourage ongoing discussion and collaboration among educators, researchers, and professional organizations to strengthen capstone pedagogy. references bandura, a. (1986). social foundations of thought and action: a social cognitive theory. prentice-hall, inc.. bandura, a. (1997). self-efficacy: the exercise of control. w.h. freeman and company. brau, j. c., nielson, j. k., & sudweeks, b. l. (2015). experiential learning in personal finance: a principles and applications based approach. journal of financial education, 41(2), 49-79. bruning, r. h., schraw, g. j., & norby, m. m. (2011). cognitive psychology and instruction, 5th ed. pearson. cfp board. (n.d.). the capstone course. https://www.cfp.net/getcertified/certification-process/educationrequirement/certification-courseworkrequirement/the-capstone-course cfp board. (2019). code of ethics and standards of conduct. https://www.cfp.net//media/files/cfp-board/standards-andethics/cfp-code-andstandards.pdf?la=en&hash=afc0087b5 3ede26b2a478bd84df7dda3 cfp board. (2021). registration criteria for cfp board registered programs. https://www.cfp.net/-/media/files/cfpboard/education-partners/registeredprograms/become-a-registeredprogram/registration_criteria_for_cfp_b oard_registered_programs_updated_202 1.pdf du, x. & han, j. (2016). a literature review on the definition and process of projectbased learning and other relative studies. creative education, 7, 1079-1083. fredrickson, r., mcmahan, s., & dunlap, k. (2013). problem-based learning theory. in b. j. irby, g. brown, r. lara-alecio, & s. jackson (eds.), the handbook of educational theories (pp. 211-217). iap information age publishing. goetz, j. w., durband, d. b., halley, r. e., & davis, k. (2011). a peer-based financial planning & education service program: https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/the-capstone-course https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/the-capstone-course https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/the-capstone-course https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/the-capstone-course financial services review, 33(4) 32 an innovative pedagogic approach. journal of college teaching and learning, 8(4), 7-14. goetz, j. w., tombs, j. w., & hampton, v. l. (2005). easing college students' transition into the financial planning profession. financial services review, 14, 231-251. grable, j. e., kruger, m. e., & ford, m. r. (2022). the fundamentals of writing a financial plan (2nd ed.). national underwriter academic series. the national underwriter company. guo, p., saab, n., post, l. s., admiraal, w. (2020). a review of project-based learning in higher education: student outcomes and measures. international journal of educational research, 102, https://doi.org/10.1016/j.ijer.2020.10158 6. halpern, d. f., & hakel, m. d. (2003). applying the science of learning to the university and beyond. change: the magazine of higher learning, 35(4), 36-41, doi: 10.1080/00091380309604109 hanney, r., & savin-baden, m. (2013). the problem of projects: understanding the theoretical underpinnings of project-led pbl. london review of education, 11(1), 7-19. heymann, r., schnusenberg, o., & timmerman, i. (2025). mapping the terrain: strategies for building effective university financial planning programs. financial planning review, 8(1), e1198. https://doi.org/10.1002/cfp2.1198 jackling, b., & sullivan, c. (2007). financial planners in australia: an evaluation of gaps in technical and behavioral skills. financial services review, 16(3), 211228. jacob, m. b. (2016). pedagogy of financial education among college students. economics and culture, 13(2), 23-34. jenlink, p. m. (2013). situated cognition theory. in b. j. irby, g. brown, r. lara-alecio, & s. jackson (eds.), the handbook of educational theories (pp. 185-198). iap information age publishing. krajcik, j., & blumenfeld, p. (2006). projectbased learning. in r. k. sawyer (ed.), the cambridge handbook of the learning sciences (pp. 317-333). cambridge university press. krajcik, j. s., & shin, n. (2014). project-based learning. in r. k. sawyer (ed.), the cambridge handbook of the learning sciences (2nd ed). (pp. 275-297). martin jr, c. l. (2007). developing a personal financial planning program: more than just courses. journal of college teaching & learning (tlc), 4(11), 79-84. maurer, t. w., & lee, s.-a. (2011). financial education with college students: comparing peer-led and traditional classroom instruction. journal of family and economic issues, 32(4), 680-689. narayan, r., rodriguez, c., araujo, j., shaqlaih, a., & moss, g. (2013). constructivism— constructivist learning theory. in b. j. irby, g. brown, r. lara-alecio, & s. jackson (eds.), the handbook of educational theories (pp. 169–183). iap information age publishing. paciotti, k. d. (2013). cognitivism—ways of knowing. in b. j. irby, g. brown, r. lara-alecio, & s. jackson (eds.), the handbook of educational theories (pp. 105–113). iap information age publishing. solis, o. j. (2018). exploring innovative pedagogy to deliver financial education: a peer-to-peer event with students. american journal of business education, 11(4), 63-70. teale, j. (2013). improving financial planning graduate employability through enterprise education. australasian accounting business & finance journal, 7(3), 91-106. thapa, s., & chan, k. c. (2013). undergraduate entrepreneurial finance course: some https://doi.org/10.1016/j.ijer.2020.101586 https://doi.org/10.1016/j.ijer.2020.101586 asebedo & gramse 33 curriculum and pedagogical perspectives. the journal of learning in higher education, 9(1), 37. weisz, m. (2000). developing a measure of student attributes. the journal of cooperative education, 35(2/3), 33. west, t., johnson, d., & webb, a. (2019). career outcomes of financial planning students. financial planning research journal, 5(1), 82-110. zhang, s. (in press). how to build a cfp board registered program at an aacsb accredited business school? financial services review. financial services review, 33(4) 34 appendix a capstone case example: oscar & sofia client introduction oscar and sofia are a middle-aged, dual-income household with young children, moderate income, and growing complexity in their financial situation. they are balancing career development, business growth, debt management, and long-term wealth-building goals. client background oscar (36) and sofia (36) have been married for ten years and live in lubbock, tx, with their three children: jasmine (8), marco (5), and camila (2). oscar works as an it administrator at texas tech university, earning $65,000 per year with benefits and retirement plans through his employer. sofia owns a small but growing online retail business with a current income of $50,000; her earnings are expected to grow substantially as the business expands. assets and liabilities cash $30,000 in savings and $5,000 in checking home $300,000 value, mortgage balance approx. $170,000 at 4% retirement oscar contributes to ttu orp and 403(b); sofia has an old 401(k) from a prior employer investments small brokerage account with individual stocks vehicles/property two cars financed at 7%; household property ~$50,000 debt $18,000 credit card balance at 23% interest insurance and benefits oscar has life, disability, and health insurance through ttu. sofia currently has no life or disability coverage. the family is covered by ttu’s health plan. property and casualty insurance is in place, but not up to date. primary goals • retire together at age 65–67 while maintaining current lifestyle • fund 100% of tuition, room and board, and fees for all three children at texas tech university • pay for $20,000 weddings for each child and provide $10,000 vehicles at age 16 • upgrade to a $500,000 home within 5–10 years • purchase a $100,000 luxury vehicle at retirement • travel in retirement ($15,000 annually for 20 years) • increase annual charitable giving from $10,000 to $30,000 during retirement • leave a legacy of $500,000 to each child and $800,000 to charity ($2.3 million total) • eliminate debt by retirement appendix a continued on next page. asebedo & gramse 35 appendix a continued. planning considerations • manage high-interest credit card debt and evaluate refinancing options • explore entity choice and retirement plan opportunities for sofia’s business • align investment strategy with differing risk profiles (oscar 40/100, sofia 60/100) • evaluate education funding strategies (529 plans, ttu tuition benefits) • review estate plan; current documents are limited to simple wills capstone mini-case example: alex rivera (retirement stage, upper-class, single individual) alex rivera is 65 and recently retired after successfully running a specialty toy company that he sold five years ago. he is single, has no dependents, and lives in a luxury condominium in a metropolitan area. alex accumulated substantial wealth from the sale of his business and his long-term investments; his current net worth exceeds $8 million. alex’s retirement income is generated through a diversified portfolio that includes taxable accounts, retirement accounts, and real estate investments. his annual income comfortably exceeds his spending needs, allowing him to travel frequently and pursue hobbies such as aviation. while he has medicare with supplemental insurance, his focus is less on affordability of care and more on ensuring comprehensive coverage and planning for potential long-term care. alex’s primary financial concerns now center on managing estate taxes, preserving wealth for future charitable bequests, and ensuring his investment portfolio remains aligned with his long-term goals. he has no direct heirs but would like to establish a charitable foundation and leave funds for nieces and nephews. student assignment: evaluate alex’s wealth management and risk planning strategy in retirement. consider investment allocation, estate planning tools, and philanthropic strategies appropriate for a highnet-worth retiree with no dependents. capstone mini-case example: daniel and emily carter (accumulation stage, lower-class, married couple) daniel (52) and emily (50) carter are married with three children: ava (17), who is preparing to start college; lucas (14), in high school; and noah (10), in elementary school. daniel works full-time as a mechanic, earning about $45,000 annually, while emily works part-time in retail, bringing in an additional $18,000. their combined household income places them in a lower-class status, and they have little savings set aside. the carters rent their home and carry a modest balance of credit card and auto loan debt. they contribute minimally to retirement accounts because most of their income is consumed by daily expenses. with ava starting college soon, they are worried about how to pay for tuition while still covering household needs and saving for retirement. their financial goals include providing some level of support for their children’s education, paying down high-interest debt, and building a small emergency fund. longer-term, they would like to retire in their mid-60s, though they recognize this may be difficult without significant changes to their current financial habits. appendix a continued on next page. financial services review, 33(4) 36 appendix a continued. student assignment: assess the carters’ competing financial priorities given their limited resources. recommend strategies for debt reduction, education planning, and beginning a sustainable retirement savings plan, while recognizing the realities of a lower-income household. pii: s1057-0810(00)00064-0 the information content of closed-end country fund discounts john e. richarda, james b. wigginsb,* adepartment of finance and business economics, wayne state university, detroit, mi, usa bdepartment of finance, michigan state university, east lansing, mi 48824, usa received 20 july 1999; received in revised form 25 july 2000; accepted 22 august 2000 abstract this paper examines whether premiums and discounts on closed-end country mutual funds (cecfs) contain useful information about future returns. we find that higher cecf premiums are associated both with higher future returns on the relevant foreign market index and with higher future nav returns after controlling for the foreign market return. cecfs trading at large discounts are not necessarily bargains, because their future nav performance can be expected to be relatively poor. © 2000 elsevier science inc. all rights reserved. jel classification:g14; g15; g23 keywords:closed-end fund; discount; investor sentiment 1. introduction both open-end and closed-end mutual funds pool shareholders’ money and invest in financial securities. but while open-end mutual funds stand ready to issue or redeem shares at their net asset value (nav) at any time, closed-end funds do not. closed-end fund shares trade on an exchange like an individual stock, and the share price can fluctuate above or below nav. when the share price is higher (lower) than nav, the fund is said to trade at a premium (discount). the “managerial performance” theory and the “investor sentiment” theory provide two * corresponding author. tel.:11-517-353-2256; fax:11-517-432-1080. e-mail address:jwiggins@msu.edu (j.b. wiggins). financial services review 9 (2000) 171–181 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00064-0 rationales for variations in closed-end fund premiums across funds and over time. the managerial performance theory (malkiel, 1977) hypothesizes that premiums vary with the skills of fund managers; funds that charge high expenses or frequently select poorlyperforming stocks will sell at a large discount from nav. the investor sentiment theory (zweig, 1973; lee, shleifer & thaler, 1991) posits that changes in the (possibly irrational) expectations of individual investors cause the premium to fluctuate over time. when individual investors become more optimistic about the future performance of the underlying investments of the fund, the premium increases. one closed-end fund category of particular interest is closed-end country funds (cecfs). cecfs invest exclusively in a single foreign stock market. cecfs are intriguing because fund shares and underlying assets trade in different markets. since few open-end funds specialize in a single foreign market (six fidelity funds are the major exception), investing in a cecf is often the only way for an individual to purchase a well-diversified portfolio of stocks in a specific foreign country. thus, cecf discounts provide a unique gauge of u.s. investors’ valuation of the stock market in the country. this paper empirically examines the information content of cecf discounts for future returns, testing elements of both the managerial performance and investor sentiment theories. first, we test whether the discount forecasts the fund’s future nav performance, controlling for the return on the foreign market and exchange rate risk. in theory, if investors believe that a particular fund manager has superior stock-picking ability, they will pay a high price for that fund relative to nav. discounts should be larger for funds with inferior managers. likewise, if investors believe a fund incurs excessive operating expenses or transaction costs, the discount will be large, because high expenses will likely translate into relatively low nav returns. our managerial performance test updates previous research of hardouvelis, la porta and wizman (1994). we also test whether the cecf discount forecasts the future return on the market index of the foreign country. since cecf shares trade in the u.s. market but the underlying fund assets trade in a much less accessible foreign market, investing in a cecf is usually the most efficient way for u.s. residents to bet on the prospects for a specific foreign market. for example, suppose u.s. investors increase their expectations of future earnings from taiwanese stocks, but taiwanese investors do not. all else equal, the taiwan cecf share price will increase, but the nav will stay the same, narrowing the discount or increasing the premium. if u.s. investor opinion or sentiment, as manifested in the cecf share price, contains useful information about the foreign market not yet fully reflected in nav, an above-average premium should be associated with above-average future foreign market returns. 2. theory and literature review the investor sentiment theory of closed-end fund pricing can be traced to zweig (1973), who hypothesizes that relatively uninformed individuals are the primary investors in closedend fund shares. to zweig, actions of these uninformed investors are contrary indicators of future stock market performance. when uninformed investors become optimistic, closed-end fund premiums increase, and future stock market performance is then expected to be poor. 172 j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 using data from 1965–1971, zweig finds that the frequency of week-to-week increases in premiums for domestic stock funds predicts the future return on the dow jones industrial average. when the number of increases in premiums is abnormally high in a given week, future djia returns are relatively low. lee, shleifer and thaler (1991) further develop the investor sentiment theory. individuals, rather than institutions like pension funds, are the primary investors in both closed-end funds and small-company stocks. according to this theory, when individual investors become more optimistic about the stock market, the prices of both small-company stocks and closed-end funds are driven up relative to the value of large-company stocks, which comprise the majority of closed-end fund portfolios. since individual investor optimism increases the fund’s price, but not its nav, the discount will narrow. using data from 1965–85, they find that the average discount on domestic stock funds is inversely related to the excess return on small-company stocks versus large-company stocks. bodurtha, kim and lee (1995) test the investor sentiment hypothesis using cecf data over the january 1986-december 1990 period. they find that changes in the average premium on cecfs are positively related to the return on the u.s. stock market, controlling for the return on the foreign market and exchange rate movements. presumably, when u.s. investors become more optimistic, they drive up prices of domestic stocks and cecfs at the same time. in a similar vein, bailey and lim (1992) find that u.s. stock indexes exhibit a higher positive correlation with cecf share prices than with their underlying foreign market indexes, suggesting cecf discounts are influenced by u.s. investor sentiment. the managerial performance theory can be traced to boudreaux (1973) and malkiel (1977). if a manager is perceived to be highly skilled at stock selection or market timing, investors will bid up the price of fund shares, and the fund will trade at a premium. similarly, if a fund levies excessive annual management fees, it should sell at a discount. controlling for other variables, malkiel (1977, 1995) finds no relation between the domestic fund discounts and either historical performance or management fees. thus, past performance and management expenses do not appear to explain cross-sectional variation in discounts for domestic stock funds. other authors test the managerial performance theory by examining the relation between premiums and future (rather than past) performance. for domestic stock funds, pontiff (1994) finds no evidence that premiums predict future nav returns. using a sample including both domestic and international funds, chay and trzcinka (1999) uncover evidence supporting the managerial performance hypothesis. using data from 1963–93, and adjusting for risk using several different u.s. market benchmarks, they find that higher premiums forecast superior nav performance. hardouvelis et al. (1994) test the managerial performance theory for cecfs over the january 1985-january 1993 period. when running separate regressions for each fund, they find only a weak relation between premiums and future nav returns. but running separate regressions is not a powerful procedure, because it cannot detect situations where some funds consistently sell at higher premiums than other funds because they have better managers. hardouvelis et al. (1994) then pool data and restrict regression coefficients to be equal across funds, thereby comparing the performance of one fund against another, and uncover a positive and significant relationship between premiums and future nav returns. funds trading at a premium achieve higher future performance than funds trading at a discount. 173j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 the effect of investment restrictions on cecf premiums complicates testing the investor sentiment and managerial performance theories. errunza (1991), building on the work of errunza and losq (1985, 1989), develops a theoretical model of closed-end country fund premiums, showing that premiums depend on the ease of direct investment in the stock market of the country. u.s. investors, including open-end mutual funds, can freely trade individual stocks in some foreign markets (united kingdom) but are subject to ownership limits in others (singapore). american depository receipts (adrs) trade on the new york stock exchange for many companies from some markets (mexico), but are rare or nonexistent for companies in other markets (malaysia). the more difficult investment in a country through open-end mutual funds or adrs, the larger is the expected cecf premium. bonser-neal, brauer, neal and wheatley (1990) find that announcements of liberalizations of investment restrictions in a country generally lead to decreases in cecf premiums. errunza (1991), bodurtha et al. (1995), and hardouvelis et al. (1994) present evidence that premiums are higher for cecfs investing in restricted markets. the next section describes the implications of investment restrictions for our tests. 3. data and methodology each weekend,barron’s reports closed-end fund share price and nav data. we collect data for 38 funds with at least two years of operations over the january 1988-march 1997 period. we exclude four china cecfs because we lack data on chinese market index returns, and two funds specializing in gold stocks. most funds report weekly nav as of friday afternoon in the united states, andbarron’s reports friday’s nyse closing share price along with it. but a few funds report nav as of thursday or wednesday. the india growth fund consistently reported nav from wednesday, andbarron’s reported wednesday’s closing share price for that fund. brazil, brazilian equity, emerging mexico, mexico, mexico equity & income, singapore, and taiwan consistently reported nav from thursday, andbarron’s reported thursday’s closing share price for those funds. thus, even for funds reporting nav as of wednesday or thursday, the cecf share price is measured at the close of trading on the same day. nav data are unavailable for 266 out of 14,414 observations (less than 2%). following bekaert and urias (1996), we use the nav from the previous week as a proxy for these observations. in the few cases that a cecf share price is not available inbarron’s, prices come from thewall street journalor the center for research in security prices (crsp) tapes. we compute weekly foreign market index returns, foreign exchange rates, and world market returns from morgan stanley capital international (msci) daily index data. the msci data come from dri/mcgraw hill. table 1 presents the list of 38 funds, the number of weekly observations available for each, the mean percentage premium (with a negative number indicating a discount), and the mean weekly nav return. the nav return measures the performance of the underlying assets of the fund, defined as 174 j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 nav returnt 5 [nav t 1 dt navt-1]/nav t-1, (1) where dt is the amount of dividend and capital gain distributions paid to shareholders in week t. all returns are adjusted for stock splits. distribution and stock split data come from the crsp tapes. to test the managerial performance theory, we want to compare funds against one another over a common time period. we estimate a seemingly unrelated regression (sur) system of equations for the 30 of 38 funds that have complete return data over january 1991 through march 1997: table 1 descriptive statistics country fund n mean weekly premium mean weekly nav return mean weekly foreign market return argentina argentina 282 5.38% 0.20% 0.19% australia first australia 481 212.83% 0.15% 0.18% austria austria 388 29.74% 0.08% 0.05% brazil brazil 267 23.24% 0.49% 0.65% brazil brazilian equity 258 20.58% 0.45% 0.53% chile chile 387 26.80% 0.49% 0.49% france france growth 347214.25% 0.13% 0.14% germany emerging germany 360215.85% 0.05% 0.16% germany germany 481 22.11% 0.26% 0.26% germany new germany 371214.73% 0.14% 0.15% india india growth 220 10.79% 0.00% 0.12% indonesia indonesia 366 13.13% 20.01% 20.03% indonesia jakarta 361 2.38% 0.02% 20.04% ireland irish investment 362 214.87% 0.19% 0.15% israel first israel 172 21.91% 0.06% 20.13% italy italy 330 28.06% 0.04% 0.10% japan japan equity 240 4.16% 0.19% 0.16% japan japan otc equity 350 4.92% 20.14% 0.00% korea korea 481 34.75% 0.15% 0.06% korea korea equity 172 21.62% 20.33% 20.16% korea korean investment 262 3.53% 20.13% 0.02% malaysia malaysia 481 23.97% 0.35% 0.35% mexico emerging mexico 234 21.36% 0.23% 0.20% mexico mexico 450 210.36% 0.54% 0.63% mexico mexico equity & income 297 24.03% 0.30% 0.20% pakistan pakistan 168212.01% 20.47% 20.30% philippines first philippine 382 217.53% 0.28% 0.29% portugal portugal 383 27.45% 0.11% 0.04% singapore singapore 346 22.00% 0.22% 0.25% spain growth fund of spain 366215.41% 0.17% 0.16% spain spain 453 2.91% 0.16% 0.12% switzerland switzerland 481 26.44% 0.19% 0.30% taiwan roc taiwan 409 20.02% 0.11% 0.10% taiwan taiwan 481 13.76% 0.52% 0.42% thailand thai 463 6.60% 0.28% 0.27% thailand thai capital 354 27.22% 0.07% 0.06% turkey turkey 379 7.77% 0.12% 0.30% united kingdom united kingdom 481 213.74% 0.21% 0.18% 175j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 nav returnt 5 a0 1 a1premt-1 1 a2fmrt 1 a3fxrt 1 et (2) where: nav returnt 5 rate of return on nav for week t premt-1 5 [cecf share price nav]/nav for week t-1, the percentage premium on the fund fmrt 5 foreign market rate of return for week t, measured in u.s. dollars fxrt 5 percentage change in the spot foreign exchange rate over week t, measured as of the close of trading in the foreign country, in units of u.s. dollar per unit of foreign currency the foreign market return, fmrt, is included in the regression to control for the market or systematic component of the nav return on the fund. following bodurtha et al. (1995) and hardouvelis et al. (1994), we also include fxrt, the change in foreign exchange rates over week t, as a control variable. cecfs may use futures or forward contracts to hedge foreign exchange risk. as a result, exchange rate fluctuations could influence the relation between the dollar nav return and the dollar foreign market index return. we restrict the coefficient a1 to be equal across all 30 funds in estimating the sur system. by applying this restriction, a1 can be interpreted as the marginal sensitivity of the week t nav return to the deviation of the week t-1 premium from the mean across all funds and dates. coefficients on fmrt and fxrt are unrestricted, so each fund has its own beta with respect to its foreign market and its own sensitivity to foreign exchange fluctuations. our specification is somewhat more flexible than hardouvelis et al. (1994), who restrict the fmrt and fxrt coefficients to be the same across funds. the null hypothesis is a1 5 0, that the cecf premium contains no information about the future nav return on the fund, controlling for other variables. if investment restrictions (errunza, 1991) cause premiums to vary across funds, our test remains valid but is less powerful than it would be without such confounding factors. weekly returns for all variables in eq. (2) are measured friday-to-friday for funds reporting nav on friday, thursday-to-thursday for funds reporting nav on thursday, and wednesday-to-wednesday for the india growth fund. cecf shares trade on the nyse until 4 p.m., but the underlying foreign market closes earlier. thus, the end-of-week t-1 cecf share price inevitably contains some information about the week t foreign market return not yet incorporated into end-of-week t-1 nav. this nonsynchronicity induces a spurious positive correlation between the week t-1 premium and the week t nav return. however, our multivariate tests should be free of bias because we include the week t foreign market return fmrt as a regressor. fmrt captures any information about the foreign market return between the foreign market and nyse close on the last day of week t-1 that gets reflected in the week t-1 premium. to see if cecf premiums have information content for future foreign market returns, we regress foreign market index returns in week t on the cecf premium at the end of week t-1 and the return on the world market portfolio in week t for each of the 38 individual funds: fmrt 5 b0 1 b1premt-1 1 b2wmrt 1 et (3) where: 176 j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 fmrt 5 foreign market rate of return for week t, measured in u.s. dollars premt-1 5 [cecf share price nav]/nav at the end of week t-1 wmrt 5 msci world market rate of return for week t, measured in u.s. dollars since individual foreign market returns are influenced by the contemporaneous world market return, wmrt is included as a control variable. the null hypothesis of interest is b1 5 0, that the premium does not predict the subsequent week’s foreign market return. since we can test this hypothesis by running individual regressions for each fund, funds from countries with strict investment restrictions stay separate from funds in more open markets. as noted earlier, the nyse closes at 4 p.m. eastern time but most foreign markets close earlier. the time lag is particularly long for asian funds, where markets can close 12 hr or more before the nyse. because of this time lag, the cecf premium could contain information about the “true” (though as yet unobservable) return on the foreign index between the foreign market close and the u.s. market close. to avoid any spurious correlation resulting from the time lag, for funds reporting nav on friday, foreign market and world market returns are measured over the following monday close-to-monday close week. funds reporting nav on thursday and wednesday use friday-to-friday and thursday-to-thursday foreign and world market returns respectively. 4. empirical results table 2 presents regression results from eq. (2). the coefficient on premt-1, restricted to be the same across funds, is 0.0108 with a t-statistic of 6.273, significant at the 1% level using newey and west (1987) standard errors with one lag. for example, if a fund trades at a premium that is 0.10 higher than the overall average premt-1 for the sample, then the nav return on that fund over the next week tends to be abnormally high by 0.10(0.0108)5 0.00108 or 0.108%. the results in table 2 suggest that cecf premiums contain valuable information about future nav performance after controlling for the foreign market return and exchange rate fluctuations. this is consistent with the managerial performance hypothesis. our results update the findings of hardouvelis et al. (1994) over four more recent years of data. to test the robustness of our results, we also estimate a more general specification, using variables from hardouvelis et al. (1994) and bodurtha et al. (1995). we include returns on large company u.s. stocks and the excess return on u.s. small company stocks to control for possible u.s. investor sentiment effects on the nav return, and also the world market return as an explanatory variable. none of these variables add significant explanatory power to the model and there is no qualitative effect of the premt-1 coefficient estimate. table 3 presents the regression results from eq. (3). coefficient estimates on the world market return, representing the “beta” of the foreign market return with respect to the world index, are positive and statistically significant for all but five funds. of the 38 coefficients on the cecf premium, 11 are positive and significant at the 5% or 1% level in a two-tailed test. under the null hypothesis, we would expect to see only about one of 38 funds statistically significant in the upper 2.5% of the distribution. none of the premt-1 coefficients are 177j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 negative and statistically significant. across all 38 funds, the average coefficient on premt-1 is 0.0218, indicating that if the premium for a particular fund exceeds its in-sample average by 0.05, then the foreign market return over the next week tends to be 0.05(0.0218)5 0.00109 or 0.109% higher than expected. the results in table 3 suggest that cecf premiums contain valuable information about future foreign stock market returns. since few open-end funds specialize in a single country, cecfs are often the most efficient way for individuals to focus their investment in a specific foreign market. if u.s. investors become relatively more optimistic about future earnings or dividends than domestic investors in a specific country, the cecf premium will increase. to some extent, u.s. investors’ beliefs or sentiment are confirmed by future foreign market returns. here, sentiment reflects rational beliefs about future returns rather than irrational waves of optimism or pessimism. table 2 seemingly unrelated regression of weekly nav return against the cecf premium, foreign market return, and foreign exchange return, with the cecf premium coefficient restricted across equations, january 1991–march 1997 fund constant prem fmr fxr adj r-squared first australia 0.0012 0.0108** 0.8799** 0.1289 0.6221 austria 0.0018** 0.0108** 0.7466** 20.0242 0.6974 brazil 0.0009 0.0108** 0.6624** 20.0026 0.6625 chile 0.0023* 0.0108** 0.7988** 0.2372* 0.7159 france growth 0.0017** 0.0108** 0.7882** 0.1395** 0.7249 emerging germany 0.0009 0.0108** 0.8689** 20.0151 0.8481 germany 0.0009 0.0108** 0.9077** 0.0461 0.8165 new germany 0.0021** 0.0108** 0.7386** 0.1696** 0.7504 indonesia 20.0012 0.0108** 0.6140** 20.0009 0.5920 jakarta 0.0002 0.0108** 0.5470** 0.0014 0.6313 ireland 0.0025** 0.0108** 0.7064** 0.1586** 0.7660 italy 0.0011 0.0108** 0.7533** 0.1233* 0.8115 japan otc equity 20.0016 0.0108** 0.6215** 0.3478** 0.4259 korea 20.0001 0.0108** 0.6943** 0.1928 0.5580 malaysia 0.0009 0.0108** 0.8902** 0.0516 0.6265 emerging mexico 0.0026 0.0108** 0.6643** 0.4872** 0.5240 mexico 0.0017 0.0108** 0.6781** 0.1123 0.4755 mexico equity & income 0.0041* 0.0108** 0.4708** 0.5730** 0.5164 first philippine 0.0028** 0.0108** 0.5213** 0.2412** 0.5457 portugal 0.0020** 0.0108** 0.6602** 0.1942** 0.7274 singapore 0.0004 0.0108** 0.6026** 0.0876 0.4276 growth fund of spain 0.0024** 0.0108** 0.7782** 0.0203 0.7911 spain 0.0012 0.0108** 0.7333** 0.1098* 0.6616 switzerland 0.0004 0.0108** 0.7699** 0.1221** 0.7429 roc taiwan 0.0004 0.0108** 0.5790** 0.5387* 0.6700 taiwan 0.0006 0.0108** 0.5793** 0.3106 0.4667 thai 0.0010 0.0108** 0.9442** 20.3155 0.8031 thai capital 0.0008 0.0108** 0.8968** 20.8516** 0.8235 turkey 20.0026 0.0108** 0.8264** 20.0847 0.8341 united kingdom 0.0025** 0.0108** 0.7809** 0.1654** 0.5982 * significant at 5%, ** significant at 1% in two-tailed t-test using newey and west (1987) standard errors with one lag. 178 j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 there are important differences in premt-1 coefficients across countries in table 3. for the more developed markets of australia, france, germany, italy, japan, spain, switzerland, and the united kingdom, just two of twelve funds have significant positive coefficients. for the eight funds from latin american markets, including argentina, brazil, chile, and mexico, none of the coefficients are significantly different from zero. for funds in the emerging asian markets of india, indonesia, korea, malaysia, pakistan, the philippines, singapore, taiwan, and thailand, eight of fourteen funds have significant positive coefficients. on the whole, developed and latin american markets are far more open to u.s. table 3 regressions of the weekly foreign market return against the cecf premium and world market return fund n constant prem wmr adj r-squared argentina 282 20.0022 0.0363 1.2288** 0.0964 first australia 481 0.0031 0.0181 0.6497** 0.1798 austria 388 0.0056* 0.0642** 0.8925** 0.2824 brazil 267 0.0049 0.0115 1.0948** 0.0379 brazilian equity 258 0.0027 20.0104 1.2053** 0.0480 chile 387 0.0047* 0.0001 0.1996 0.0061 france growth 347 20.0003 20.0020 0.9743** 0.3908 emerging germany 360 0.0009 0.0054 0.9227** 0.3598 germany 481 0.0012 0.0108 0.9489** 0.3445 new germany 371 0.0019 0.0109 0.8829** 0.3404 india growth 220 20.0101** 0.1014** 0.1458 0.1054 indonesia 366 20.0057* 0.0387** 0.1994 0.0258 jakarta 361 20.0021 0.0569** 0.2591* 0.0359 irish investment 362 0.0056 0.0382 0.9902** 0.3575 first israel 172 20.0036 20.0202 1.0064** 0.0942 italy 330 0.0022 0.0376 1.0014** 0.1774 japan equity 240 20.0028* 0.0096 1.7524** 0.5291 japan otc equity 350 20.0040** 0.0310* 1.7029** 0.6279 korea 481 20.0021 0.0055 0.5078** 0.0513 korea equity 172 20.0025 0.0440 0.8585** 0.1294 korean investment 262 20.0027 0.0516* 0.5650** 0.0672 malaysia 481 0.0028* 0.0147 0.8190** 0.2178 emerging mexico 234 20.0007 20.0172 1.1135** 0.0705 mexico 450 0.0037 20.0114 0.9277** 0.0986 mexico equity & income 297 20.0016 20.0413 1.0688** 0.0949 pakistan 168 0.0114 0.1225** 0.1558 0.0810 first philippine 382 0.0003 20.0120 0.4050** 0.0310 portugal 383 0.0007 0.0132 0.5608** 0.1223 singapore 346 0.0010 0.0054 1.0494** 0.2875 growth fund of spain 366 20.0025 20.0153 1.1533** 0.4163 spain 453 20.0005 0.0007 1.0751** 0.4001 switzerland 481 0.0032** 0.0244* 0.8152** 0.3295 roc taiwan 409 0.0000 0.0698** 0.7154** 0.0724 taiwan 481 20.0030 0.0423** 0.8357** 0.0827 thai 463 0.0006 0.0159* 0.7059** 0.0891 thai capital 354 0.0027 0.0456 0.8577** 0.1040 turkey 379 0.0007 0.0216 0.5207 0.0132 united kingdom 481 0.0016 0.0086 0.8460** 0.3990 * significant at 5%, ** significant at 1% in two-tailed t-test using newey and west (1987) standard errors with one lag. 179j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 investors than emerging asian markets, with a much wider selection of adrs available. the more accessible the market, the faster the information of u.s. investors is reflected in nav, and thus the less informative is the cecf premium in forecasting future returns. 5. conclusions closed-end country funds purchase diversified portfolios of common stocks located in a specific foreign country. this paper examines whether the cecf discount contains valuable information regarding (1) future nav returns, controlling for foreign market returns, and (2) future returns on the relevant foreign market index. because closed-end funds do not redeem or issue new shares to investors on demand at nav like open-end funds, the cecf premium reflects investor perceptions of the expected future performance of fund managers. all else equal, the higher the perceived future performance of the fund, the higher the premium. we find a positive and statistically significant relation between the premium and future nav returns, controlling for the contemporaneous foreign market return and exchange rate fluctuations. since few open-end funds specialize in a single foreign market, investing in a cecf is typically the most efficient way for an individual to purchase a portfolio of stocks in a specific country. thus, cecf premiums and discounts provide a measure of u.s. investors’ valuation of the stock market in the country. we find that relation between the cecf premium at the end of week t-1 and the week t return on the foreign market is positive and statistically significant for 11 of 38 funds, controlling for the week t return on the world market. our results are important for the individual investor in a couple of ways. first, we confirm previous findings that cecf discounts provide information about the future investment performance of the fund, controlling for the foreign market return and exchange rate fluctuations. a fund with a large discount is not necessarily a bargain, because its future nav returns can be expected to lag behind its respective market. second, this is the first paper to show that the cecf discount forecasts the return on the underlying foreign market. large discounts are associated with relatively low returns on the foreign market, controlling for the world market return. this is especially true in asian markets, which are less open to u.s. investors than many other markets. cecf premiums and discounts at least partially reflect rational assessments of the future performance of both the home market and the fund manager, not simply pricing errors caused by irrational investor sentiment. references bailey, w., & lim, j. (1992). evaluating the diversification benefits of the new country funds.journal of portfolio management, 18,74–80. bekaert, g., & urias, m. (1996). diversification, integration and emerging market closed-end funds.journal of finance, 51,835–869. bodurtha, j., kim, d., & lee, c. (1995). closed-end country funds and u.s. market sentiment.review of financial studies, 8,879–918. 180 j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 bonser-neal, c., brauer, g., neal, r., & wheatley, s. (1990). international investment restrictions and closed-end country fund prices.journal of finance, 45,523–548. boudreaux, k. (1973). discounts and premiums on closed-end mutual funds: a study in valuation.journal of finance, 28,515–522. chay, j., & trzcinka, c. (1999). managerial performance and the cross-sectional pricing of closed-end funds. journal of financial economics, 52,379–408. errunza, v. (1991). pricing of national index funds.review of quantitative finance and accounting, 1,91–100. errunza, v., & losq, e. (1985). international asset pricing under mild segmentation: theory and test.journal of finance40, 401–417. errunza, v., & losq, e. (1989). capital flow controls, international asset pricing, and investors’ welfare: a multi-country framework.journal of finance, 44,1025–1038. hardouvelis, g., la porta, r., & wizman, t. (1994). what moves the discount on country equity funds. in: j. frankel, (ed),the internationalization of equity markets,345–397. lee, c., shleifer, a., & thaler, r. (1991). investor sentiment and the closed-end fund puzzle.journal of finance, 46, 75–109. malkiel, b. (1977). the valuation of closed-end investment-company shares.journal of finance, 32,847–859. malkiel, b. (1995). the structure of closed-end fund discounts revisited.journal of portfolio management, 21, 32–38. newey, w., & west, k. (1987). a simple, positive semi-definite heteroskedasticity and autocorrelation consistent covariance matrix.econometrica, 55,703–708. pontiff, j. (1994). closed-end fund premia and returns: implications for financial market equilibrium.journal of financial economics, 37,341–370. zweig, m. (1973). an investor expectations stock price predictive model using closed-end fund premiums. journal of finance, 28,67–87. 181j.e. richard, j.b. wiggins / financial services review 9 (2000) 171–181 financial services review, 33(3) 20 from intention to adequate emergency fund savings through fintech use: evidence from a u.s. survey study ying chen,1 sarah d. asebedo,2 todd d. little, 3 and weihong ning4 abstract this study applied the theory of planned behavior and technology adoption models to investigate consumers' adequate emergency fund savings through the intention to use fintech to save and actual fintech use. a structural equation model with a confirmatory factor analysis was employed to analyze primary data from a sample of 453 u.s. respondents collected in july 2021. the results show that attitudes toward adequate emergency fund savings were negatively associated with the intention to use fintech to save. subjective norms and perceived behavioral control, respectively, were positively associated with the intention to use fintech to save. perceived behavioral control showed a positive direct relationship with adequate emergency fund savings. intention to use fintech to save showed positive relationships with using saving apps and websites, respectively. however, only saving website use was positively associated with adequate emergency fund savings. the results suggest that the intention to use fintech to save and actual fintech use connect attitudes, subjective norms, and perceived behavioral control with adequate emergency fund savings. the findings shed light on empirical evidence in the current literature regarding the importance of intention to use fintech to save. financial institutions, financial advisors, and policymakers should be aware of the significance of intention to use fintech and actual fintech use in households' savings behaviors. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation chen, y., asebedo, s. d., little, t. d., & ning, w. (2025). from intention to adequate emergency fund savings through fintech use: evidence from a u.s. survey study. financial services review, 33(3), 20-47. introduction the unpreparedness for unexpected events before the covid-19 pandemic (finra, 2022) highlighted the need to address u.s. households' financial vulnerability. during the pandemic, researchers, financial professionals, governments, and policymakers underscored the importance of adequate emergency fund savings in dealing with 1 corresponding author (ying.chen@ccsu.edu), central connecticut state university, new britain, ct, usa. 2 sarah.asebedo@ttu.edu, texas tech university, lubbock, tx, usa. 3 todd.d.little@ttu.edu, texas tech university, lubbock, tx, usa. 4 weihong.ning@ccsu.edu, central connecticut state university, new britain, ct, usa. unexpected events, such as layoffs, sicknesses, and medical expenses. even though the importance of maintaining adequate emergency fund savings has been called for attention (gjertson, 2016), u.s. households still face challenges in achieving financial stability (chen et al., 2024) and are advised to reevaluate their portfolios to better prepare for economic fallouts (despard & roll, 2024). despite an increasing https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ chen et al. 21 number of u.s. households reporting adequate emergency fund savings – the 2021 national financial capability survey (nfcs) report shows that more than half of the households reported having at least three months equivalent of living expenses as emergency fund savings (finra, 2022) – many americans still struggle to pay bills and meet basic needs (despard & roll, 2024). when asked about confidence in coming up with $2,000 within the next month in the face of financial hardships, only 43% of u.s. households reported being able to meet this balance in emergency fund savings in 2021 (finra, 2022). from a behavioral perspective, the theory of planned behavior (tpb) emphasizes that behavioral and psychological factors (i.e., attitudes, subjective norms, perceived behavioral control, and intention) might explain savings behavior (ajzen, 1991). attitudes towards savings, subjective norms, and perceived behavioral control are significant predictors of households' intention to save and thus affect their saving behaviors (satsios & hadjidakis, 2018). when it comes to emergency fund savings, financial professionals recommend at least three months' equivalent of living expenses as adequate emergency fund savings (finra, 2022). however, while households often value shortterm savings, they do not always have enough liquid funds to adequately protect themselves from unexpected financial shocks in the form of adequate emergency fund savings (chase et al., 2011). in the short term, households often overlook the importance of adequate emergency fund savings and, thus, lack such savings to cover unexpected life events because the intentionbehavior relationship varies depending on specific behavior, habits, and experiences (conner & armitage, 1998). households with an intention to save might still fail to act to save (gollwitzer, 1999). this intention-to-behavior gap could represent opportunities for financial technology (fintech) to connect the path from the intention to use fintech to save to actual savings behavior of maintaining adequate emergency fund savings. fintech refers to applying technology to financial services and products to facilitate and improve financial activities (bajunaied et al., 2023; feyen et al., 2023; schueffel, 2016; setiawan et al., 2022). it covers various financial domains, including account management, savings, payments, investment, and insurance, depending on the purposes of financial activities (demirgüçkunt et al., 2018; thakor, 2020). fintech has dramatically drawn the attention of financial institutions, researchers, governments, and policymakers over the past decades. behavioral intention of using technology determines actual fintech use, implying that fintech companies should evaluate and predict whether fintech users will accept the new technology (bajunaied et al., 2023). actual fintech adoption has expanded its benefits (i.e., low costs, convenience, easy access) from financial sectors to their clients and provided opportunities to improve financial behaviors and outcomes (demirgüç-kunt et al., 2018; feyen et al., 2023; ouma et al., 2017; thakor, 2020). during the covid-19 pandemic, an emerging trend for fintech adoption in financial services markets has consistently grown for two reasons: first, the financial services industry can widely use fintech to serve their clients better; second, actual fintech use provides opportunities for households' access to services and products such as savings, insurance, and credit management with low cost (feyen et al., 2023). furthermore, promoting banking and online account combinations encourages bank account holders to learn how to use technology to manage finances, such as savings and investments. research shows that bank account holders have more opportunities to access financial products and services through websites and apps (abis et al., 2025; demirgüç-kunt et al., 2022; thakor, 2020). although 96% of u.s. households were banked in 2023, only 48% of the bank account holders used mobile banking (federal deposit insurance corporation [fdic], 2023), indicating the opportunity to promote fintech to customers in the united states. while research has emphasized the vital role of fintech use, it is critical to apply the tpb to predict adequate emergency fund savings through the intention to use fintech to save and actual fintech use. to our knowledge, extant studies have focused on the intention to conduct savings behavior (i.e., lučić et al., 2025; widjaja et al., financial services review, 33(3) 22 2020), credit borrowing behavior (i.e., xiao et al., 2011), or on the intention to use fintech (i.e., bajunaied et al., 2023). there is a dearth of research on the association between the intention to use fintech for emergency fund savings and the actual use of fintech products and services to save for emergencies. establishing and maintaining adequate emergency fund savings reduces financial vulnerability (stavins, 2021) and enhances overall financial well-being (anvariclark & ansong, 2022; nourallah et al., 2025). however, a mechanism is needed to trigger households' action to save for adequate emergencies. fintech can be an effective way to motivate households to save and maintain their savings. therefore, there it can be fruitful to introduce a comprehensive model to explain households' motivation and actual use of fintech products and services for financial behaviors (löwgren, 2023). this study applies the tpb (ajzen, 1991), technology acceptance model (tam) (davis, 1989; davis et al., 1989), and unified theory of acceptance and use of technology (utaut) (venkatesh et al., 2003) to address households' adequate emergency fund savings while accounting for the intention to use fintech to accomplish their savings behavior through actual fintech use. using primary data with a sample size of 453 respondents collected from u.s. adults in july 2021, this study examines the path from attitudes toward adequate emergency fund savings, subjective norms, and perceived behavioral control to adequate emergency fund savings through the intention to use fintech to save and to actually use fintech. our study has two main contributions. first, it advances the literature on savings behavior in personal financial planning by integrating the impact of psychological characteristics and actual fintech use. second, it provides empirical evidence of fintech as a mechanism to facilitate households' adequate emergency fund savings with practical implications. the remainder of this paper is organized as follows. section two reviews previous literature and develops a research model and the hypotheses. section three explains empirical methodology and data analysis. section four presents empirical results. section five discusses our research findings. section six concludes the paper. theoretical framework and hypothesis development adequate emergency fund savings savings behavior exists when people spend less than earnings (heckman & hanna, 2015). households are motivated by reserving the excess current income to smooth consumption over their lifetime (ando & modigliani, 1963). many extant studies have extended topics related to savings behavior or wealth accumulation from a longterm perspective (anderson et al., 2017; asebedo et al., 2022; bi & montalto, 2004; cole et al., 1992; ouma et al., 2017; peiris, 2021; shefrin & thaler, 1988; strömbäck et al., 2017). built upon the economic theory of savings, heckman and hanna (2015) proposed that social impacts and psychological factors combined with financial access, incentives, and facilitations might encourage savings behavior among low-income households. asebedo et al. (2019) found that financial self-efficacy (fse) explained savings behavior among u.s. older adults. in other words, people who reported more control over their financial situation tended to save more. similarly, applying the social cognitive theory of selfregulation, asebedo and seay (2018) found that fse was associated with increased wealth accumulation among u.s. pre-retirees. from a short-term perspective, leland (1968) proposed that individuals were motivated to make financial decisions between savings and consumption based on their current and expected future income strains in accounting for the uncertainty about the future. skinner (1988) proposed that the accumulated precautionary savings to protect against future income shocks accounted for a large portion of accumulated capital. chase et al. (2011) summarized that emergency fund savings differed from precautionary savings. precautionary savings are assets set aside to respond to income shocks, such as unemployment and pay cuts; emergency fund savings are money set aside to respond to medical or dental expenses and household expenses (i.e., auto and home repairs). as mentioned, a stricter definition of adequate emergency fund savings is similar to the precautionary savings chen et al. 23 recommended to save at least three months equivalent of living expenses. emergency fund savings are short-term money reserves earmarked for unexpected events (chase et al., 2011). johnson and widdows (1985) suggested that emergency fund savings be equivalent to two to six months of living expenses to cover income shocks. there are two main emergency fund measures present in the literature: (a) a dichotomous measure indicating whether an adequate emergency fund savings account exists (johnson & widdows, 1985), and (b) a continuous measure of the actual dollar amount of emergency fund savings (anong & devaney, 2010; bi & montalto, 2004; brobeck, 2008; gjertson, 2016). precautionary savings and emergency fund savings have been used interchangeably. the current study followed the definition of adequate emergency fund savings to cover unexpected events, such as layoffs, medical expenses, and repairs for at least three months of living expenses. researchers have been investigating the factors associated with short-term emergency fund savings (chase et al., 2011; despard & roll, 2024). firstly, previous empirical studies have shown that financial factors affect households' adequate emergency fund savings (brobeck, 2008; satsios et al., 2020). for example, income was positively associated with adequate emergency fund savings (babiarz & robb, 2014). bank account ownership strongly predicted short-term savings, such as emergency fund savings with at least three months of living expenses (despard et al., 2020). secondly, researchers have investigated the impact of objective and subjective financial knowledge on short-term financial behaviors and found mixed results (tang & baker, 2016). for example, fan and zhang (2021) found that objective and subjective financial knowledge were positively associated with emergency fund savings. henager and cude (2016) found similar results and indicated that subjective financial knowledge was more strongly associated with emergency fund savings in younger age groups. in comparison, objective financial knowledge substantially impacted emergency fund savings in older age groups. other studies found that only subjective financial knowledge was positively associated with adequate emergency fund savings (babiarz & robb, 2014; chen et al., 2024; despard et al., 2020; ismail et al., 2017). a variety of demographic characteristics are associated with holding adequate emergency fund savings. first, age was positively associated with adequate emergency fund savings (bi & montalto, 2004). old adults were more likely to have emergency fund savings of $500 or higher than young adults (brobeck, 2008). second, some studies found that men were more likely than women to hold adequate emergency fund savings (wagner & walstad, 2019). finally, white households had a higher probability of having adequate emergency fund savings than non-white households (bi & montalto, 2004; chase et al., 2011; despard et al., 2020). fintech use fintech, an application of technology to financial services, has attracted the attention of researchers, policymakers, and industry over the decades (feyen et al., 2023; schueffel, 2016; setiawan et al., 2022; thakor, 2020). fintech was initially intended to facilitate payments in consumption activities and has expanded its services to lending, investing, and savings in financial sectors through digital payments, robo-advisors, online banking, and financial apps (abis et al., 2025; bajunaied et al., 2023; risman et al., 2022). financial services companies have increasingly adopted fintech to offer timely information and enhance client interactions (nicoletti et al., 2017). the ongoing innovation of fintech has created an efficient and effective bridge between fintech use and financial capability (demirgüç-kunt et al., 2022; ouma et al., 2017). for example, yeo and fisher (2017) found that americans who frequently used fintech were more likely to manage their money, make savings decisions, and understand their financial issues. similarly, nourallah et al. (2024) concluded that fintech use was positively associated with financial capability in the european union. depending on the forms and functions, fintech has been used to perform various financial activities and assist households in making sound financial decisions. for example, mobile bank apps have been developed to facilitate households in conducting day-to-day financial activities, while robofinancial services review, 33(3) 24 advisors are widely used in investment activities (nourallah & öhman, 2021). in the current study, we focus on using fintech to save for emergency fund savings; thus, actual fintech use limits its forms of saving apps use and saving websites use. theoretical foundation ajzen (1991) proposed the tpb to address psychological characteristics associated with a specific behavior. it states that an individual's behavior depends on three antecedents (i.e., attitude, subjective norms, and perceived behavioral control) and motivation (i.e., intention). an attitude is an individual's tendency to favor or disfavor a specific behavior (ajzen, 1991). it aggregates one's knowledge, evaluation, and positive or negative prejudices toward the behavior. subjective norms refer to an individual's perception of others' ideas or attitudes about performing or not performing a specific behavior. it can be considered as social pressure or influence from others, such as family and friends, which affects how much the individual values their expectations. perceived behavioral control refers to an individual's selfassessed ability to perform a behavior. such perception depends on the individual's internal and external factors. the internal factors include determination and ability, while the external factors include resources, opportunities, and support. intention refers to an individual's psychological disposition to act on a specific behavior. it indicates how much the individual plans to exert to perform a specific behavior. the tpb has been widely used to explain human behaviors in various domains, such as healthrelated behaviors (conner & sparks, 2005; godin & kok, 1996), public health and political science (bosnjak et al., 2020), energy savings behaviors (cheung et al., 1999; lin & shi, 2022), and consumer behaviors (george, 2004; paramita et al., 2018; sharif & naghavi, 2021; shih & fang, 2004; xiao et al., 2011). although extant literature has applied tpb to explain financial behaviors, the results indicated that the impacts of the three tpb components (attitude, subjective norms, and perceived behavioral control) on the intention varied. sharif and naghavi (2021) found that all tpb components were significantly correlated with college students' intention to use online financial trading. shih and fang (2004) indicated that attitude and perceived behavior control were significantly related to consumers' intention to use online banking, but subject norms were not. paramita et al. (2018) showed that none of the tpb components were significantly associated with college students' intention to make stock investments. furthermore, the technology acceptance model (tam) (davis, 1989; davis et al., 1989), and the unified theory of acceptance and use of technology (utaut) (venkatesh et al., 2003) are the most widely utilized models of acceptance and usage of innovative technology. generally, these models indicate that users' behavioral attributes of a specific technology, including perceived usefulness, perceived ease of use, and social influence, determine their attitude and intention to use the technology. the intention, in turn, predicts their actual technology usage. with the ubiquity of fintech products and services, scholars have explored the tam and utaut in this domain (i.e., hu et al., 2019; shih & fang, 2004; singh et al., 2020; samartha et al., 2022; wang, 2021) with an emphasis on how behavioral attributes influence attitude and intention to use fintech. wang (2021) revealed that face and voice recognition were the most preferred identifications methods in fintech apps, which positively influenced users' perceived trust and perceived privacy. hu et al. (2019) showed that bank users' trust in fintech services indirectly predicted their attitudes, which consequently positively impacted their intention to adopt these services for a better experience. using the utaut model, samartha et al. (2022) found similar results that ease of use and trust positively influenced the intention to adopt mobile banking apps. so far, a few studies have examined the impact of the intention to use fintech on actual fintech use but have not provided conclusive findings. shih and fang (2004) found a positive relationship between the intention to use online banking and the actual fintech use of online banking services. however, singh et al. (2020) showed that intention to use fintech did not determine actual fintech use frequency. thus, our study extends the use of the tpb to the phenomenon of adequate emergency fund savings through the lens of technology adoption. chen et al. 25 based on the tpb and technology adoption models (i.e., tam, utaut), we posit that these three antecedents (i.e., attitude, subjective norms, and perceived behavioral control) will determine an individual's intention to use fintech to save, which will then influence them to actually use saving-related fintech products or services. finally, the actual use of fintech will affect their adequate emergency fund savings. hypotheses as previously mentioned, the three antecedents of tpb influence an individual's intention. in the context of saving behaviors, research has found that attitude, subjective norms, and perceived behavior control toward savings directly affect individuals' intention to save and thus affect actual savings behaviors such as saving for security needs (devaney et al., 2007) and retirement (magwegwe & lim, 2021). regarding the attitude toward savings, magwegwe and lim (2021) showed that attitudes toward savings directly affect individuals' intentions and actual ira ownership. people might be motivated to save for basic or security needs (devaney et al., 2007). for example, young people who need to meet their monthly basic needs might not intend to save. when their basic needs have been achieved, they might move up to the next level of savings motives and thus intend to save for security needs to cover unexpected financial shocks. on the one hand, maintaining adequate emergency fund savings can help people achieve the goal of financial security (despard & roll, 2024). on the other hand, with the advancement of fintech and its prevalence, the benefits of fintech use have been promoted regarding its usefulness (i.e., savings, account monitoring, security) and ease of use (abis et al., 2025; bajunaied et al., 2023; risman et al., 2022). these factors might enhance households' attitudes toward adequate emergency fund savings with fintech. thus, we hypothesized as follows: h1: a favorable attitude toward adequate emergency fund savings is positively associated with the intention to use fintech to save. regarding subjective norms, families’ and friends' perceptions of savings are positively associated with the intention to save (duflo & saez, 2003; lučić et al., 2025; magwegwe & lim, 2021). for example, lučić et al. (2025) found that parents' and peers' norms significantly and positively affected children's intention to save. similarly, magwegwe and lim (2021) revealed that subjective norms positively affected individuals' intention to save for retirement. moreover, peer norms influenced the contribution portion of retirement savings (beshears et al., 2015). venkatesh et al.'s utaut (2003) suggested that subjective norms influenced individuals' intention to use fintech under voluntariness and limited experience. we argue that people might voluntarily save for emergency funds but have limited experience using fintech. families' and friends' perceptions of adequate emergency fund savings with fintech use might affect their intention to use fintech to save. additionally, an individual's intention to use fintech to save can also be influenced by social media influencers (safitri et al., 2021). thus, we hypothesized as follows: h2: the subjective norm regarding emergency fund savings is positively associated with the intention to use fintech to save. ajzen (1991) suggested that perceived behavioral control could be internally evaluated by individuals' confidence or self-efficacy in specific situations. the current study measured perceived behavioral control through an individual's confidence in managing emergency fund savings. researchers have found that perceived behavioral control had a direct positive relationship with the intention to save (lučić et al., 2025; satsio & hadjidakis, 2018). moreover, individuals' perceptions and evaluations depend on external factors, such as resources, opportunities, and support (ajzen, 1991). fintech, serving as a tool, provides opportunities and technology support to motivate households to save. for example, financial institutions encourage clients to activate online banking services, which usually make money management tools available (becker, 2017; löwgren, 2023). given that perceived behavioral control is determined by internal factors (i.e., confidence, self-efficacy) and external factors (i.e., technology support), we argue that financial services review, 33(3) 26 perceived behavioral control will affect intention to use fintech to save and hypothesized as follows: h3: perceived behavioral control is positively associated with the intention to use fintech to save. the uniqueness of the tpb is that perceived behavioral control directly impacts the action. perceived behavioral control serves as a proxy of actual behavioral control and, thus, reflects an actual behavioral control measurement. asebedo and seay (2018) found that an individual's selfperception of control of financial situations was positively associated with savings behavior among u.s. pre-retirees. accordingly, perceived behavioral control over emergency fund savings could directly affect adequate emergency fund savings. thus, we hypothesized as follows: h4: perceived behavioral control over emergency fund savings is positively associated with adequate emergency fund savings. the intention-behavior relation delineated in tam (davis, 1989; davis et al., 1989) suggested that an individual's intention to use fintech should influence their actual fintech usage. various fintech products and services, such as apps and websites, have been developed and used in financial planning to meet households' needs. the effects on financial outcomes depend on the fintech form(s) used (abis et al., 2025). for example, lee (2019) found that a push notification about outstanding balances from the mobile app reduced credit cardholders' consumption. households tended to use fintech if they believed the benefits of using fintech were greater than its costs (al nawayseh, 2020). most young consumers intended to use mobile bank apps and online banking services to enhance their savings behavior and thus used fintech to save (daqar et al., 2020). accordingly, we expect that if an individual has a strong intention to use fintech to save, they are more likely to actually use fintech (i.e., saving apps or saving websites) to manage their savings. thus, we hypothesized as follows: h5: intention to use fintech to save is positively associated with saving apps use. h6: intention to use fintech to save is positively associated with saving websites use. previous research about the impact of fintech use has been mainly focused on households’ spending and investment (lee, 2019; ouma et al., 2017; sharif & naghavi, 2021; thakor, 2020). however, nourallah et al. (2024) implied that fintech use affected households’ financial capability (i.e., savings behaviors). for example, becker (2017) found that first-time savers significantly increase their savings after activating and using fintech. demirgüç-kunt et al. (2018) suggested that fintech promotion might help households increase the probability of regular savings if they have a mobile bank account with low costs. becker (2017) highlighted that fintech, such as savings apps and websites (guittierrez ramirez, 2023), could serve as money management tools to save and manage money. löwgren (2023) and ouma et al. (2017) found that saving apps triggered households’ savings behaviors, encouraged their savings engagements, and increased the dollar amount saved. using the 2018 nfcs data, chen et al. (2024) found that fintech use positively affected households' adequate emergency fund savings. based on the above literature, actual fintech use can enhance households' financial capability to manage their savings accounts and thus increase their adequate emergency fund savings. therefore, we hypothesized as follows: h7: saving app use is positively associated with adequate emergency fund savings. h8: saving website use is positively associated with adequate emergency fund savings. given the above hypotheses, figure 1 was developed to illustrate the research model of the current study. chen et al. 27 figure 1: proposed research model h2(+) intention to use fintech to save subjective norms perceived behavioral control adequate emergency fund savings attitudes toward adequate emergency fund savings saving app use saving website use financial services review, 33(3) 28 methods survey design, data, and sample this study randomly collected primary data from 491 u.s. working adults aged 18 and above from amazon mturk in july 2021 (cf. litman et al., 2017). given the nature of the human subject research, this study was reviewed and approved by the institution human research protection program (irb2021390). a survey with fintech use, psychological factors, and emergency fund savings based on the tpb (ajzen, 1991; 2006) was designed and distributed to mturk users through qualtrics. before the full data project, a pilot study was conducted to determine whether the proposed survey and procedures were appropriate for a more extensive sample. the pilot study consisted of 45 respondents, including 25 unpaid family members, friends, colleagues, professionals, and 20 paid participants from mturk (litman et al., 2017). in the complete data collection project, the targeted sample size of 450 was calculated based on a 5% margin of error and a 95% confidence level. mturk automatically excluded 21 respondents who did not meet the requirements and paid 470 respondents, including 17 who completed less than 40% of the questionnaire. after excluding these 17 observations, the sample size was 453 in the final analysis. note that two respondents with illogical data reported numbers that suggested the survey was completed without attention or artificially (i.e., income reported as $50,006,000, expenses reported as $70,008,000, and emergency fund reported as $6,000,000 while income reported as $50,000). we used the full information maximum likelihood (fiml) estimation method to deal with missing data (cf. little, 2024). furthermore, content validity was applied in the survey by including questions about highly valid demographic characteristics (ruel et al., 2015). for example, respondents can easily answer questions regarding their age, gender, and race. however, some questions include jargon and are more difficult to answer, such as emergency fund savings, saving accounts, income, and expenses. when asked about emergency fund savings in savings accounts, few respondents might know that emergency fund savings are usually located in liquid accounts such as checking, savings, and money market accounts. to account for this potential ambiguity, the survey question defined and listed all possible liquid accounts, including checking accounts, savings accounts, money market accounts, and certificate deposits (cds), that respondents might use for their emergency fund savings. variable measurement the dependent variable is adequate emergency fund savings, calculated by whether respondents had at least three months of living expenses in their liquid savings accounts (i.e., checking and savings accounts, money market accounts, and short-term cds). following this minimum recommended guideline, we calculated the dollar amount needed for adequate emergency fund savings by multiplying the monthly living expenses by three. then, we compared the dollar amount required with the self-reported emergency fund balance. if the self-reported emergency fund balance was equal to or greater than the dollar amount needed, we coded it as 1; otherwise, 0. the primary explanatory concepts of interest are attitudes toward emergency fund savings, subjective norms, perceived behavioral control over emergency fund savings, intention to use fintech to save for emergencies, and actual fintech use. three latent variables, i.e., attitude, subjective norms, and perceived behavioral control, were utilized to measure respondents' motivations and perceptions of emergency fund savings. indicators were constructed following a recommended methodology based on the tpb questionnaire (ajzen, 2006) (see appendix). the seven-point scale was used for respondents to answer the survey questions of the three latent variables. the attitudes, subjective norms, and perceived behavioral control constructs demonstrated a good (0.7 ≤ 𝛼 ≤ 0.8) internal reliability (cf. taber, 2018). intention to use fintech to save is an observed variable operationalized through a 7-point likert-type question. based on the model, intention was a key variable in the path from attitudes, subjective norms, and perceived behavioral control to emergency fund savings through actual fintech use. as mentioned, actual fintech use was categorized into two observed variables: saving apps and saving websites. objective financial knowledge was derived from houts and knoll's (2020) financial knowledge scale (fks) and constructed by ten objective financial knowledge questions (see appendix). respondents' correct responses represent the value of objective financial knowledge ranging from zero to ten. subjective financial knowledge was derived from a chen et al. 29 self-assessment of their overall financial knowledge. the analysis also included covariate variables (i.e., wealth, income, bank account holders, age, gender, race, educational achievement, marital status, employment, and the number of children). data analysis a structural equation model (sem) with a confirmatory factor analysis (cfa) was employed for the psychological constructs in mplus. the model tested the following effects: (a) the direct effects of attitudes, subjective norms, and perceived behavioral control on the intention to use fintech to save; (b) the direct effects of perceived behavioral control on emergency fund savings based on the assumption that behavioral perceptions (i.e., self-reported emergency fund level) are close enough to actual behavior (ajzen, 1991); (c) the direct effect of intention to use fintech to save on actual fintech use, measured by saving app use and saving website use, respectively; (d) the indirect effects of attitudes, subjective norms, and perceived behavioral control on the actual emergency fund savings through intention to use fintech to save and to actually use fintech. the covariate variables are controlled based on a full partial approach. all constructs are regressed on the covariates but only for the significant effects (little, 2024). the significant effects of control variables on the latent construct predictors represent the measurement of noninvariance (brown, 2015). the variance inflation factors (vifs) test was run to diagnose collinearity/multicollinearity issues in the analysis (henager & cude, 2016). the vifs indicated low multicollinearity concerns between independent variables (mean vif = 1.44; lowest vif = 1.05; highest vif = 1.91). results descriptive statistics tables 1 and 2 show the sample characteristics of continuous and categorical variables. the results indicated that approximately 39% of respondents had adequate emergency fund savings. respondents had a strong intention to use fintech to save (5 out of 7). meanwhile, they did not frequently use saving apps (3.3 out of 7) or websites (3.4 out of 7). the average objective financial knowledge was 5.2 out of 10, and the average subjective financial knowledge was 5.3 out of 7. the mean log wealth was 5.5 out of 7.8 with a standard deviation of 0.2. the mean log income was 4.7 out of 6.3 with a standard deviation of 0.3. the average age of the respondents was 37 years old. the majority were bank account holders (98%), men (66%), white (70%), holding a bachelor's degree (62%), married (67%), full-time employed (90%), and having one child or two (58%). financial services review, 33(3) 30 table 1. sample characteristics of continuous variables variables mean (sd) attitudes toward adequate emergency fund savings at1: bad to good (1-7) 5.902 (1.307) at2: unpleasant to pleasant (1-7) 5.907 (1.232) at3: harmful to beneficial (1-7) 6.158 (1.091) subjective norms sn1: family's impact (1-7) 5.951 (1.078) sn2: friend's impact (1-7) 5.715 (1.156) perceived behavioral control pc1: self-confidence (1-7) 5.344 (1.664) pc2: emergency fund savings control (17) 5.409 (1.486) intention to use fintech to save (1-7) 5.000 (1.801) actual fintech use saving app use (1-7) 3.254 (1.908) saving website use (1-7) 3.372 (1.861) objective financial knowledge (0-10) 5.236 (2.493) subjective financial knowledge (1-7) 5.260 (1.346) log wealth (4.819 – 7.780) 5.530 (0.230) log income (3.146 – 6.301) 4.721 (0.311) age (21 – 71) 36.815 (9.412) source: primary data collected through mturk in july 2021. n = 453 chen et al. 31 table 2. sample characteristics of categorical variables variables n (percent) adequate emergency fund savings 175 (38.63%) bank account holders 446 (98.45%) gender female 156 (34.44%) male 297 (65.56%) race white 318 (70.20%) non-white 135 (29.80%) educational attainment high school graduate or less 35 (7.73%) some colleges, no degree 52 (11.48%) bachelor’s degree 279 (61.59%) graduate degree or higher 87 (19.21%) marital status married 305 (67.33%) non-married 148 (32.67%) employment full-time employed 408 (90.07%) others 45 (9.93%) children zero 169 (37.31%) one 149 (32.89%) two 114 (25.17%) three and more 21 (4.64%) source: primary data collected through mturk in july 2021. n = 453 financial services review, 33(3) 32 table 3 displays the reliability and correlation results of variables. the measures of attitude, subjective norms, and perceived control were internally reliable because all the values of composite reliability exceeded the 0.70 threshold (nunnally & bernstein, 1994). table 3. reliability and correlation matrix 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 c.r. 1 1.00 2 0.20 1.00 3 0.04 -0.05 1.00 4 0.03 -0.01 -0.10* 1.00 5 0.17*** 0.03 0.08 -0.11* 1.00 6 -0.21*** 0.05 0.11* 0.10* -0.33*** 1.00 7 -0.09* 0.04 -0.15** -0.01 -0.17*** 0.19*** 1.00 8 0.13** 0.06 0.01 -0.14** 0.14** -0.49*** -0.25*** 1.00 9 0.06 0.08 -0.21*** 0.01 0.03 -0.05 0.13** 0.05 1.00 10 0.50*** 0.02 0.08 0.00 0.17*** -0.22*** -0.11* 0.09 0.01 1.00 11 0.03 0.02 0.06 0.13** 0.10* -0.05 -0.07 -0.07 -0.01 0.20*** 1.00 12 -0.15** 0.11* -0.13** 0.06 -0.20*** 0.39*** 0.27*** -0.34*** 0.19*** 0.03 0.07 1.00 13 0.15** -0.01 0.16** 0.02 0.19*** -0.21*** -0.24*** 0.26*** -0.03 0.15** 0.06 -0.22*** 1.00 14 0.14** 0.02 0.12** -0.12* 0.23*** -0.32*** -0.27*** 0.28*** -0.11* 0.11* 0.10* -0.37*** 0.45*** 1.00 15 0.24*** 0.12** 0.04 -0.08 0.19*** -0.34*** -0.21*** 0.28*** 0.01 0.13** -0.05 -0.33*** 0.30*** 0.45*** 1.00 16 0.16** 0.09* 0.09 -0.13** 0.22*** -0.29*** -0.15** 0.27*** 0.04 0.08 -0.04 -0.22*** 0.23*** 0.27*** 0.43*** 1.00 17 -0.04 0.16** -0.21*** 0.03 -0.08 0.20*** 0.16** -0.08 0.20*** 0.02 -0.08 0.36*** 0.07 -0.03 -0.08 -0.09 1.00 0.88 18 0.08 0.03 -0.05 -0.04 0.10* -0.10* -0.03 0.12** 0.01 0.07 0.05 0.07 0.28*** 0.29*** 0.14** 0.14** 0.36*** 1.00 0.87 19 0.12* 0.03 0.12* -0.07 0.20*** -0.30*** -0.19*** 0.22*** 0.02 0.22*** 0.19*** -0.21*** 0.54*** 0.43*** 0.24*** 0.29*** 0.06 0.35*** 0.88 notes: 1-adequate emergency fund savings; 2-bank account holders; 3-gender; 4-race; 5-educational attainment; 6-marital status; 7-employment; 8-number of children; 9-age; 10-income; 11-wealth; 12-objective financial knowledge; 13-subjective financial knowledge; 14-intention; 15-savings app use; 16-savings website use; 17-attitudes; 18-subjective norms; 19-perceived behavioral control. c.r. – composite reliability *p < 0.05, **p < 0.01, ***p < 0.001 chen et al. 33 model results the study used sem with a cfa to test the theoretical hypotheses. all latent means were fixed to 0; the latent variances were fixed to 1 for twoand threeitem constructs, and the loadings were freely estimated (little, 2024). table 4 displays factor loadings and cross-loadings of the multi-item measures of attitudes, subjective norms, and perceived behavioral control. besides, each of the measurement items loaded significantly higher on their focal construct than on the other construct, which indicated the validity of these measures (chin, 1998). table 4. loadings and cross-loadings construct indicators factor 1 factor 2 factor 3 attitudes at1 0.819 -0.021 0.071 at2 0.828 0.048 -0.035 at3 0.864 -0.027 -0.033 subjective norms sn1 0.114 0.828 0.011 sn2 -0.091 0.925 -0.007 perceived behavioral control pc1 -0.040 0.070 0.863 pc2 0.038 -0.061 0.916 source: primary data collected through mturk in july 2021. n = 453 figure 2 shows the statistical results with maximum likelihood (ml) estimators within the sem framework. specifically, the model results with an acceptable model fit are as follows: χ2(df 132) = 333.421, p = < 0.001; rmsea = 0.044; 90% ci [0.050, 0.066], cfi = 0.900, tli = 0.849; srmr=0.057. financial services review, 33(3) 34 figure 2: research model results (standardized) *p < 0.05, **p < 0.01, ***p < 0.001 attitudes toward adequate emergency fund savings subjective norms perceived behavioral control intention to use fintech to save adequate emergency fund savings saving app use saving website use 0.408** (h2) chen et al. 35 table 5 shows the standardized estimates of direct effects between the variables in the current study. attitudes toward adequate emergency fund savings were negatively associated with intention to use fintech to save (b = -0.352, p < 0.001). subjective norms (b = 0.408, p < 0.01) and perceived behavioral control (b = 0.571, p < 0.001) were positively associated with the intention to use fintech to save, respectively. furthermore, perceived behavioral control (b = 0.079, p < 0.001) was positively associated with adequate emergency fund savings. the intention to use fintech to save was positively and significantly associated with both saving apps (b = 0.568, p < 0.001) and websites (b = 0.353, p < 0.001) use. however, only saving websites use (b = 0.041, p < 0.01) significantly predicted adequate emergency fund savings. thus, our research hypotheses for the direct effect were supported except for h1 (i.e., wrong direction) and h7 (i.e., insignificant). table 5. standardized estimates for the proposed sem model (direct effect) standardized estimate (se) h1: attitudes toward adequate emergency fund savings → intention to use fintech to save -0.352*** (0.098) h2: subjective norm → intention to use fintech to save 0.408** (0.144) h3: perceived behavioral control →intention to use fintech to save 0.571*** (0.078) h4: perceived behavioral control → adequate emergency fund savings 0.079*** (0.021) h5: intention to use fintech to save → saving app use 0.568*** (0.042) h6: intention to use fintech to save → saving website use 0.353*** (0.046) h7: saving app use → adequate emergency fund savings -0.015 (0.014) h8: saving website use → adequate emergency fund savings 0.041** (0.014) source: primary data collected through mturk in july 2021. n = 453 se: standard error covariate variables included in the analysis. *p < 0.05, **p < 0.01, ***p < 0.001 table 6 shows the standardized estimates of the indirect effects with 95% confidence intervals. three indirect effect paths were significant. first, the path from attitudes to adequate emergency fund savings through the intention to use fintech to save and saving websites use was negative and significant (b = -0.005, p < 0.05; 95% ci [-0.009, -0.001]). second, the path from the subjective norms to adequate emergency financial services review, 33(3) 36 fund savings through intention to use fintech to save and saving website use was positive and significant (b = 0.006, p < 0.05; 95% ci [0.001, 0.011]). third, the path from perceived behavioral control to adequate emergency fund savings through intention to use fintech to save and saving website use was positive and significant (b = 0.008, p < 0.01; 95% ci [0.003, 0.014]). the results indicated that the intention to use fintech to save and saving website use mediated the relationships between antecedents (i.e., attitudes toward adequate emergency fund savings, subjective norms, and perceived behavioral control) and adequate emergency fund savings. table 6. standardized estimates for the proposed sem model (indirect effect) standardized estimate (se) 95% ci ll ul attitudes toward adequate emergency fund savings → intention to use fintech to save → saving app use → aadequate emergency fund savings 0.003 (0.003) -0.002 0.008 subjective norm → intention to use fintech to save → saving app use → aadequate emergency fund savings -0.003 (0.003) -0.009 0.002 perceived behavioral control → intention to use fintech to save → saving app use → aadequate emergency fund savings -0.005 (0.005) -0.012 0.003 attitudes toward adequate emergency fund savings → intention to use fintech to save → saving website use → aadequate emergency fund savings -0.005* (0.003) -0.009 -0.001 subjective norm → intention to use fintech to save → saving website use → aadequate emergency fund savings 0.006* (0.003) 0.001 0.011 perceived behavioral control → intention to use fintech to save → saving website use → aadequate emergency fund savings 0.008** (0.003) 0.003 0.014 se: standard error; ci: confidence interval; ll: lower limit; ul: upper limit covariate variables included in the analysis. *p < 0.05, **p < 0.01, ***p < 0.001 discussion drawing on the tpb (ajzen, 1991), tam (davis, 1989), and utaut (venkatesh et al., 2003), this study utilized primary data to investigate the effects of three key antecedents (i.e., attitudes, subjective norms, perceived behavioral control) on adequate emergency fund savings through the intention to use fintech to save and actual fintech use (i.e., saving app and saving website). first, we explored the associations between antecedents and intention to use fintech to save. the findings revealed that all three antecedents – attitudes toward adequate emergency savings, subjective norms, and perceived behavioral control significantly influenced the intention to use fintech to save. moreover, perceived behavioral control directly affected adequate emergency fund savings. second, we also found that the intention to use fintech to save and actual fintech use (i.e., saving website use) are critical mediators of the relationships between antecedents and adequate emergency fund savings. interestingly, we found a negative relationship between attitudes toward adequate emergency fund savings and intention to use fintech to save, which does not support h1. this unexpected result can be supported by previous research (ajzen, 2006; chen et al. 37 bagozzi, 1992; davis, 1989; vermeir & verbeke, 2006), stating that other factors (i.e., desires, personal values, needs, information, knowledge) influence the attitude-intention relationship. therefore, financial goals may moderate the relationship between attitudes toward adequate emergency fund savings and intention to use fintech to save. also, people with a favorable attitude toward adequate emergency fund savings might lack financial knowledge about fintech and trust, preventing them from intending to use fintech (khan et al., 2023). further investigation should be undertaken into other factors. consistent with the h2, the result shows that the subjective norms are positively associated with the intention to use fintech to save. in other words, families’ and friends’ perceptions of the importance of emergency fund savings impact a household's intention to use apps or websites for savings. perceived behavioral control is associated significantly with an intention to use fintech to save and adequate emergency fund savings, respectively. the result supports h3, stating a positive relationship exists between perceived behavioral control and intention. households who feel they have control over their savings for emergencies are more likely to use fintech to save. as ajzen's (1991) tpb indicated, perceived behavioral control is a critical factor affecting the intention to act. consistent with h4 and findings from xiao et al. (2011), results show that perceived behavioral control is a strong indicator of maintaining adequate emergency fund savings, suggesting that people who are capable of managing funds are more likely to maintain adequate emergency fund savings. the results also show that the intention to use fintech to save is positively associated with two forms of actual fintech use: saving apps and websites. these results support h5 and h6 and echo the previous research on technology use: intention to use technology predicts its actual adoption (davis, 1989; davis et al., 1989; venkatesh et al., 2003). it implies that if consumers strongly intend to use fintech to save, they will actually use it for emergency funds regardless of its form. however, the relationship between actual fintech use and adequate emergency fund savings depends on the forms of fintech used. saving website use is positively associated with adequate emergency fund savings, which supports h8. however, saving app use is nonsignificant in predicting adequate emergency fund savings. therefore, h7 is not supported. this discrepancy might be explained by the habits of using a specific type of fintech for savings (conner & armitage, 1998; venkatesh et al., 2023). consumers have been using websites for a long time and thus may prefer to use websites to perform financial activities, including emergency fund savings. another explanation is the difference in screen size between mobile devices and computers. for example, using websites on a laptop or computer with a larger screen enables multitasking and improves the efficiency of the money management experience. thus, using saving websites to perform financial activities may better support savings management than mobile saving apps. although saving apps are innovative and convenient, it may take a while for consumers to adopt and develop the habit of using them. therefore, the benefits of using saving apps to adequate emergency fund savings have yet to be seen. surprisingly, the study reveals that the path from attitudes to adequate emergency fund savings through intention to use fintech to save and saving website use was significant but negative. although intention motivates consumers to use fintech to save and affects adequate emergency fund savings, attitudes toward adequate emergency fund savings are a domain in the path. alternative emergency fund savings mechanisms might affect attitudes toward adequate emergency fund savings outcomes. for example, credit cards are another resource to consider as an alternative form of dealing with financial emergencies (bi & montalto, 2004). without emergency fund savings, households use credit cards as an alternative method (lusardi et al., 2017; stavins, 2021). credit cards are convenient to use, easy to access, and incur lower costs than other credit mechanisms, such as payday loans (agarwal et al., 2009; chase et al., 2011; leclerc, 2012). suppose alternative emergency mechanisms (i.e., the availability of credit cards) provide convenience; consumers might be more likely to have a favorable attitude towards credit cards as emergency fund savings. thus, consumers might have great intentions to use credit cards for emergency fund savings instead of using fintech to facilitate emergency fund savings. khandelwal et al. (2021) found that favorable money attitudes were less likely to get involved in a financial crisis resulting from credit card mismanagement. therefore, future studies might account for consumers' perceptions of the emergency fund financial services review, 33(3) 38 savings mechanisms and measure their attitudes toward alternative mechanisms. moreover, paths from the subjective norms and perceived behavioral control to adequate emergency fund savings through intention to use fintech and saving website use are positive and significant. the results suggest that intention to use fintech to save and saving website use serve as mediators. the results add to empirical evidence that intention to a specific behavior and actual behavior are highly correlated (ajzen, 1991). if consumers strongly intend to use fintech to save, they might use fintech to set up automatic savings for emergency funds. moreover, actual fintech use (i.e., saving website use) has been found to serve as a mechanism to connect the intention to use fintech for savings with adequate emergency fund savings. this result adds empirical evidence to tam (davis, 1989) and utaut (venkatesh et al., 2003) that usefulness, accessibility, and trust are key features of fintech development. thus, financial institutions are advised to collaborate with fintech companies to develop efficient and effective savings websites to motivate households to engage in saving behavior for emergency funds. limitations and future research the limitations of the current study should be noted. the cross-sectional dataset limits its parameters' variability over time and causal inference (bowen & wiersema, 1999). further research might collect panel data for more accurate predictions and causal inferences. second, three-item measures for subjective norm and perceived behavioral control are recommended to be considered for future study. third, most respondents are white married males with bachelor's degrees. therefore, the sample selection of the current study might be biased. future studies should recruit other potential participants to represent u.s. households. as noted in the discussion, other factors might exist in the model that affect individuals' intentions and adequate emergency fund savings. ajzen (2006) presented an extended tpb model that emphasizes the significance of perceived behavioral control within the model. perceived behavioral control might moderate the relationships between attitudes and intention, subjective norm and intention, and intention and behavior. therefore, further studies might incorporate the moderation of perceived behavioral control into the tpb and test the model's moderation and mediation effects. conclusion and implications in conclusion, this study explains that intentionbehavior relationships through actual fintech use positively influence households' adequate emergency fund savings. it represents a combination of psychological factors and fintech use affecting savings behavior. the results provide empirical evidence for the tpb (ajzen, 1991) and the technology adoption models (i.e., tam, utaut) (davis, 1989; venkatesh et al., 2003). the indirect paths from three antecedents to adequate emergency fund savings through intention to use fintech to save and actual fintech use indicate that intention to use fintech to save and actual fintech use (i.e., using saving websites) link attitudes, subjective norms, perceived behavioral control, and adequate emergency fund savings. this study provides three critical implications. first, the intention to use fintech to save is critical to connect antecedents and consequences (i.e., adequate emergency fund savings, actual fintech use). intentions are categorized into present-oriented and future-oriented (bagozzi, 1992). if future-oriented intentions are more dominant than present-oriented ones, people might not immediately act on savings for emergencies. therefore, the results imply the importance of financial planning to align presentoriented needs (emergency fund savings) with futureoriented intentions (long-term savings). clients need to set up a financial goal or plan to activate their present-oriented intention to save for emergencies and leverage the efficiency of fintech to implement their plan. kickstart, a non-profit organization in kenya, conducted experimental research regarding using fintech to help farmers save money for equipment purchases. the results showed that using fintech to save helped them save faster than expected (omwansa et al., 2013). financial professionals might motivate households to establish financial goals and use fintech for financial management, budgeting, regular savings, and account monitoring. second, actual fintech use links the intention to use fintech to save and the savings behavior. however, the forms of fintech use matter (nourallah & öhman, 2021). saving websites help clients transfer their intention to use fintech to achieve adequate emergency fund savings. websites users’ perceived chen et al. 39 usefulness, perceived ease of use, social influence, and trust could contribute to the intention to use websites for savings. in contrast, our finding regarding saving app use calls attention to the practices. fintech companies need to make more efforts to develop saving apps to improve their efficiency and effectiveness. financial institutions might help consumers enhance their digital financial knowledge to improve the acceptance and trust in using saving apps for savings (setiawan et al., 2022). saving apps that allow fintech users to customize and set savings goals (gargano & rossi, 2024) and to use strategies to accommodate their savings capability might motivate their savings behavior (löwgren, 2023). mobile banking apps with specific features (i.e., ease of use, security) might help establish bank account holders’ trust in actually using fintech, thus improving their engagement in savings (samartha et al., 2022; sandrine et al., 2019). last but not least, governments might help households establish adequate emergency fund savings. according to u.s. congress (2021), the refund to rainy day savings act was proposed to defer 20 percent of tax refunds to taxpayers' emergency fund savings accounts to help them respond successfully to unexpected financial difficulties. this opt-in program would help taxpayers convert future-oriented intentions to present-oriented intentions to save for emergency funds when tax refunds are available through fintech. references abis, d., pia, p., & limbu, y. (2025). fintech and consumers: a systematic review and integrative framework. management decision, 63(1), 49-75. https://doi.org/10.1108/md-07-2023-1136 agarwal, s., skiba, p. m., & tobacman, j. (2009). payday loans and credit cards: new liquidity and credit scoring puzzles? american economic review, 99(2), 412–417. https://doi.org/10.1257/aer.99.2.412 ajzen, i. (1991). the theory of planned behavior. organizational behavior and human decision processes, 50(2), 179–211. https://doi.org/10.1016/07495978(91)90020-t ajzen, i. (2006). constructing a theory of planned behavior questionnaire. [unpublished manuscript]. university of massachusetts amherst. https://people.umass.edu/aizen/pdf/tpb.meas urement.pdf ajzen, i. (2020). the theory of planned behavior: frequently asked questions. human behavior and emerging technologies, 2(4), 314–324. https://doi.org/10.1002/hbe2.195 al nawayseh, m. k. (2020). fintech in covid-19 and beyond: what factors are affecting customers’ choice of fintech applications? journal of open innovation: technology, market, and complexity, 6(4), 153. https://doi.org/10.3390/joitmc6040153 anderson, a., baker, f., & robinson, d. t. (2017). precautionary savings, retirement planning, and misperceptions of financial literacy. journal of financial economics, 126(2), 383–398. https://doi.org/10.1016/j.jfineco.2017.07.008 ando, a., & modigliani, f. (1963). the "life cycle" hypothesis of saving: aggregate implications and tests. the american economic review, 53(1), 55–84. anong, s. t., & devaney, s. a. (2010). determinants of adequate emergency funds including the effects of seeking professional advice and industry affiliation. family and consumer sciences research journal, 38(4), 405–419. https://doi.org/10.1111/j.15523934.2010.00035.x anvari-clark, j., & ansong, d. (2022). predicting financial well-being using the financial capability perspective: the roles of financial shocks, income volatility, financial products, and savings behaviors. journal of family and economic issues, 43(4), 730–743. https://doi.org/10.1007/s10834-022-09849w asebedo, s. d., quadria, t. h., chen, y., & montenegro-montenegro, e. (2022). individual differences in personality and positive emotion for wealth creation. personality and individual differences, 199, https://doi.org/10.1016/0749-5978(91)90020-t https://doi.org/10.1016/0749-5978(91)90020-t financial services review, 33(3) 40 111854. https://doi.org/10.1016/j.paid.2022.111854 asebedo, s. d., & seay, m. c. (2018). financial selfefficacy and the saving behavior of older preretirees. journal of financial counseling and planning, 29(2), 357-368. https://doi.org/10.1891/1052-3073.29.2.357 asebedo, s. d., wilmarth, m. j., seay, m. c., archuleta, k., brase, g. l., & macdonald, m. (2019). personality and saving behavior among older adults. journal of consumer affairs, 53(2), 488–519. https://doi.org/10.1111/joca.12199 babiarz, p., & robb, c. a. (2014). financial literacy and emergency saving. journal of family and economic issues, 35(1), 40-50. https://doi.org/10.1007/s10834-013-9369-9 bagozzi, r. p. (1992). the self-regulation of attitudes, intentions, and behavior. social psychology quarterly, 55(2), 178-204. https://doi.org/10.2307/2786945 bajunaied, k., hussin, n., & kamarudin, s. (2023). behavioral intention to adopt fintech services: an extension of unified theory of acceptance and use of technology. journal of open innovation: technology, market, and complexity, 9(1), 100010. https://doi.org/10.1016/j.joitmc.2023.10001 0 becker, g. (2017). does fintech affect household saving behavior? findings from a natural field experiment (working paper). federal reserve bank of philadelphia. https://www.philadelphiafed.org//media/frbp/assets/events/2017/consumerfinance/fintech-2017/day-2/does-fintechaffect-household-saving-behavior.pdf beshears, j., choi, j. j., laibson, d., madrian, b. c., & milkman, k. l. (2015). the effect of providing peer information on retirement savings decisions. the journal of finance, 70(3), 1161-1201. https://doi.org/10.1111/jofi.12258 bi, l., & montalto, c. p. (2004). emergency funds and alternative forms of saving. financial services review, 13(2), 93–110. bosnjak, m., ajzen, i., & schmidt, p. (2020). the theory of planned behavior: selected recent advances and applications. europe’s journal of psychology, 16(3), 352–356. https://doi.org/10.5964/ejop.v16i3.3107 bowen, h. p., & wiersema, m. f. (1999). matching method to paradigm in strategy research: limitations of cross‐sectional analysis and some methodological alternatives. strategic management journal, 20(7), 625-636. https://doi.org/10.1002/(sici)10970266(199907)20:7%3c625::aidsmj45%3e3.0.co;2-v brobeck, s. (2008). understanding the emergency savings needs of lowand moderate-income households: a survey-based analysis of impacts, causes, and remedies (report). consumer federation of america. brown, t. a. (2015). confirmatory factor analysis for applied research (2nd ed.). guilford press. chase, s., gjertson, l., & collins, j. m. (2011). coming up with cash in a pinch: emergency savings and its alternatives (cfs issue brief no. 6). center for financial security. chen, y., asebedo, s. d., & ning, w. (2024). associations between fintech use and financial knowledge and emergency fund savings adequacy: expected and unexpected findings. journal of personal finance, 23(2). 44-65 cheung, s. f., chan, d. k. s., & wong, z. s. y. (1999). reexamining the theory of planned behavior in understanding wastepaper recycling. environment and behavior, 31(5), 587–612. https://doi.org/10.1177/00139169921972254 chin, w. w. (1998). commentary: issues and opinion on structural equation modeling. mis quarterly, vii–xvi. cole, h. l., mailath, g. j., & postlewaite, a. (1992). social norms, savings behavior, and growth. journal of political economy, 100(6), 1092– 1125. conner, m., & armitage, c. j. (1998). extending the theory of planned behavior: a review and avenues for further research. journal of https://doi.org/10.1016/j.paid.2022.111854 https://doi.org/10.1111/joca.12199 https://www.philadelphiafed.org/-/media/frbp/assets/events/2017/consumer-finance/fintech-2017/day-2/does-fintech-affect-household-saving-behavior.pdf https://www.philadelphiafed.org/-/media/frbp/assets/events/2017/consumer-finance/fintech-2017/day-2/does-fintech-affect-household-saving-behavior.pdf https://www.philadelphiafed.org/-/media/frbp/assets/events/2017/consumer-finance/fintech-2017/day-2/does-fintech-affect-household-saving-behavior.pdf https://www.philadelphiafed.org/-/media/frbp/assets/events/2017/consumer-finance/fintech-2017/day-2/does-fintech-affect-household-saving-behavior.pdf https://doi.org/10.5964/ejop.v16i3.3107 chen et al. 41 applied social psychology, 28(15), 1429– 1464. https://doi.org/10.1111/j.15591816.1998.tb01685.x conner, m., & sparks, p. (2005). theory of planned behaviour and health behaviour. in m. conner & p. norman (eds.), predicting health behaviour (2nd ed., pp. 121–162). open university press. daqar, m. a. m. a., arqawi, s., & karsh, s. a. (2020). fintech in the eyes of millennials and generation z (the financial behavior and fintech perception). banks and bank systems, 15(3), 20-28. http://dx.doi.org/10.21511/bbs.15(3).2020.0 3 davis, f. d. (1989). perceived usefulness, perceived ease of use, and user acceptance of information technology. mis quarterly, 13(3), 319–340. https://doi.org/10.2307/249008 davis, f. d., bagozzi, r. p., & warshaw, p. r. (1989). user acceptance of computer technology: a comparison of two theoretical models. management science, 35(8), 982-1003. https://doi.org/10.1287/mnsc.35.8.982 demirgüç-kunt, a., klapper, l., singer, d., ansar, s., & hess, j. (2018). the global findex database 2017: measuring financial inclusion and the fintech revolution. world bank publications demirgüç-kunt, a., klapper, l., singer, d., & ansar, s. (2022). the global findex database 2021: financial inclusion, digital payments, and resilience in the age of covid-19. world bank publications. despard, m. r., friedline, t., & martin-west, s. (2020). why do households lack emergency savings? the role of financial capability. journal of family and economic issues, 41, 542-557. https://doi.org/10.1007/s10834020-09679-8 despard, m., & roll, s. (2024). job loss and financial distress during covid-19: the protective role of emergency savings. journal of financial counseling and planning, 35(2), 157-172. http://dx.doi.org/10.1891/jfcp2022-0048 devaney, s. a., anong, s. t., & whirl, s. e. (2007). household savings motives. journal of consumer affairs, 41(1), 174-186. https://doi.org/10.1111/j.17456606.2006.00073.x duflo, e., & saez, e. (2003). the role of information and social interactions in retirement plan decisions: evidence from a randomized experiment. the quarterly journal of economics, 118(3), 815-842. https://doi.org/10.1162/00335530360698432 fan, l., & zhang, l. (2021). the influence of financial education sources on emergency savings: the role of financial literacy. family and consumer sciences research journal, 49(4), 344-361. https://doi.org/10.1111/fcsr.12400 federal deposit insurance corporation [fdic] (2023). fdic survey finds 96 percent of u.s. households were banked in 2023. https://www.fdic.gov/news/pressreleases/2024/fdic-survey-finds-96-percentus-households-were-banked-2023 feyen, e., natarajan, h., & saal, m. (2023). fintech and the future of finance: market and policy implications. world bank group. http://documents.worldbank.org/curated/en/ 099450005162250110/p17300600228b7007 0914b0b5edf26e2f9f finra (2021). national financial capability study state-by-state survey instrument. https://finrafoundation.org/sites/finrafoundat ion/files/nfcs-2021-state-by-statequestionnaire.pdf finra (2022). financial capability in the united states: highlight from the finra foundation national financial capability study. https://finrafoundation.org/sites/finrafoundat ion/files/nfcs-report-fifth-edition-july2022.pdf gargano, a., & rossi, a. g. (2024). goal setting and saving in the fintech era. the journal of finance, 79(3), 1931-1976. https://doi.org/10.1111/jofi.13339 george, j. f. (2004). the theory of planned behavior and internet purchasing. internet research, http://dx.doi.org/10.21511/bbs.15(3).2020.03 http://dx.doi.org/10.21511/bbs.15(3).2020.03 https://doi.org/10.2307/249008 https://www.fdic.gov/news/press-releases/2024/fdic-survey-finds-96-percent-us-households-were-banked-2023 https://www.fdic.gov/news/press-releases/2024/fdic-survey-finds-96-percent-us-households-were-banked-2023 https://www.fdic.gov/news/press-releases/2024/fdic-survey-finds-96-percent-us-households-were-banked-2023 http://documents.worldbank.org/curated/en/099450005162250110/p17300600228b70070914b0b5edf26e2f9f http://documents.worldbank.org/curated/en/099450005162250110/p17300600228b70070914b0b5edf26e2f9f http://documents.worldbank.org/curated/en/099450005162250110/p17300600228b70070914b0b5edf26e2f9f https://finrafoundation.org/sites/finrafoundation/files/nfcs-2021-state-by-state-questionnaire.pdf https://finrafoundation.org/sites/finrafoundation/files/nfcs-2021-state-by-state-questionnaire.pdf https://finrafoundation.org/sites/finrafoundation/files/nfcs-2021-state-by-state-questionnaire.pdf https://finrafoundation.org/sites/finrafoundation/files/nfcs-report-fifth-edition-july-2022.pdf https://finrafoundation.org/sites/finrafoundation/files/nfcs-report-fifth-edition-july-2022.pdf https://finrafoundation.org/sites/finrafoundation/files/nfcs-report-fifth-edition-july-2022.pdf financial services review, 33(3) 42 14(3), 198–212. https://doi.org/10.1108/10662240410542634 gjertson, l. (2016). emergency saving and household hardship. journal of family and economic issues, 37, 1-17. https://doi.org/10.1007/s10834-014-9434-z godin, g., & kok, g. (1996). the theory of planned behavior: a review of its applications to health-related behaviors. american journal of health promotion, 11(2), 87-98. https://doi.org/10.4278/0890-1171-11.2.87 gollwitzer, p. m. (1999). implementation intentions: strong effects of simple plans. american psychologist, 54(7), 493–503. https://psycnet.apa.org/doi/10.1037/0003066x.54.7.493 guittierrez ramirez, d. l. (2023). investigating the influence of fintech solutions on pension fund savings growth in mexico (doctoral dissertation, national college of ireland, dublin). heckman, s., & hanna, s. d. (2015). individual and institutional factors related to low-income household saving behavior. journal of financial counseling and planning, 26(2). 187-199. https://doi.org/10.1891/10523073.26.2.187 henager, r., & cude, b. j. (2016). financial literacy and long-and short-term financial behavior in different age groups. journal of financial counseling and planning, 27(1), 3-19. http://dx.doi.org/10.1891/1052-3073.27.1.3 houts, c. r., & knoll, m. a. (2020). the financial knowledge scale: new analyses, findings, and development of a short form. journal of consumer affairs, 54(2), 775-800. https://doi.org/10.1111/joca.12288 hu, z., ding, s., li, s., chen, l., & yang, s. (2019). adoption intention of fintech services for bank users: an empirical examination with an extended technology acceptance model. symmetry, 11(3), 340. https://doi.org/10.3390/sym11030340 ismail, s., faique, f. a., bakri, m. h., zain, z. m., idris, n. h., yazid, z. a., ... & taib, n. m. (2017). the role of financial self-efficacy scale in predicting financial behavior. advanced science letters, 23(5), 4635-4639. https://doi.org/10.1166/asl.2017.8992 johnson, d., & widdows, r. (1985, march). emergency fund levels of households. in proceedings of the american council on consumer interests 31st annual conference (vol. 235, p. 241). khan, m. t. i., liew, t. w., & lee, x. y. (2023). fintech literacy among millennials: the roles of financial literacy and education. cogent social sciences, 9(2), 1-16. https://doi.org/10.1080/23311886.2023.2281 046 khandelwal, r., kolte, a., veer, n., & sharma, p. (2022). compulsive buying behaviour of credit card users and affecting factors such as financial knowledge, prestige and retention time: a cross-sectional research. vision, 26(2), 172-180. https://doi.org/10.1177/0972262920981428 leclerc, k. (2012). influential factors contributing to college student spending habits and credit card debt. perspectives, 4(1), 20. lee, s. k. (2019, november 3). fintech nudges: overspending messages and personal finance management (nyu stern school of business research paper). https://doi.org/10.2139/ssrn.3390777 leland, h. e. (1968). saving and uncertainty: the precautionary demand for saving. the quarterly journal of economics, 82(3), 465473. https://doi.org/10.2307/1879518 lin, b., & shi, l. (2022). identify and bridge the intention-behavior gap in new energy vehicles consumption: based on a new measurement method. sustainable production and consumption, 31, 432-447. https://doi.org/10.1016/j.spc.2022.03.015 litman, l., robinson, j., & abberbock, t. (2017). turkprime.com: a versatile crowdsourcing data acquisition platform for the behavioral sciences. behavior research methods, 49(2), 433-442. https://doi.org/10.3758/s13428016-0727-z https://doi.org/10.1891/1052-3073.26.2.187 https://doi.org/10.1891/1052-3073.26.2.187 https://doi.org/10.1177/0972262920981428 https://doi.org/10.2139/ssrn.3390777 https://nam04.safelinks.protection.outlook.com/?url=https%3a%2f%2fwww.cloudresearch.com%2fproducts%2fturkprime-mturk-toolkit%2f&data=04%7c01%7cying.chen%40ttu.edu%7ca410112eeaca452a014608d93807c571%7c178a51bf8b2049ffb65556245d5c173c%7c0%7c1%7c637602428307680992%7cunknown%7ctwfpbgzsb3d8eyjwijoimc4wljawmdailcjqijoiv2lumziilcjbtii6ik1hawwilcjxvci6mn0%3d%7c2000&sdata=9jyixphtbv5jb8yiuqwtbh98qso6efsgh4jbomp0fre%3d&reserved=0 chen et al. 43 little, t. d. (2024). longitudinal structural equation modeling. guilford publications. löwgren, a. (2023). fintech: a more motivating way of saving money?: persuasive design in fintech. diva portal. https://www.divaportal.org/smash/get/diva2:1776121/fullt ext01.pdf lučić, a., erceg, n., & barbić, d. (2025). determinants of tweens’ saving intentions: a cross-sectional study. international journal of bank marketing, 43(2), 239-261. https://doi.org/10.1108/ijbm-08-2023-0468 lusardi, a., michaud, p. c., & mitchell, o. s. (2017). optimal financial knowledge and wealth inequality. journal of political economy, 125(2), 431-477. https://doi.org/10.1086/690950 magwegwe, f. m., & lim, h. (2021). factors associated with the ownership of individual retirement accounts (iras): applying the theory of planned behavior. journal of financial counseling and planning, 32(1), 116-130. https://doi.org/10.1891/jfcp-1900023 nicoletti, b., nicoletti, w., & weis, a. (2017). future of fintech. springer nature. nourallah, m., chan, h. r., chien, c. l., & öhman, p. (2025). financial capability, behavior, well‐being, and stress among financial advisors. financial planning review, 8(2), e70002. https://doi.org/10.1002/cfp2.70002 nourallah, m., & öhman, p. (2021). impact of advanced technologies on consumer finance and retail investment. in r. ho, a. h. h. ng, & m. nourallah (eds.), impact of globalization and advanced technologies on online business models (pp. 1–15). igi global. https://doi.org/10.4018/978-1-79987603-8.ch001 nourallah, m., öhman, p., & hamati, s. (2024). financial technology and financial capability: study of the european union. global finance journal, 62, 101008. https://doi.org/10.1016/j.gfj.2024.101008 nunnally, j. c., & bernstein, i. h. (1994). psychometric theory (3rd ed.). new york, ny: mcgraw-hill. omwansa, t. k., waema, t. m., chen, c., & sullivan, n. p. (2013). the mobile phone as the tool to redefine savings for the poor: evidence from kenya. african journal of science, technology, innovation and development, 5(5), 355-361. ouma, s. a., odongo, t. m., & were, m. (2017). mobile financial services and financial inclusion: is it a boon for savings mobilization?. review of development finance, 7(1), 29-35. https://doi.org/10.1016/j.rdf.2017.01.001 paramita, r. s., isbanah, y., kusumaningrum, t. m., musdholifah, m., & hartono, u. (2018). young investor behavior: implementation theory of planned behavior. international journal of civil engineering and technology, 9(7), 733-746. peiris, t. u. i. (2021). effect of financial literacy on individual savings behavior; the mediation role of intention to saving. european journal of business and management research, 6(5), 94-99. http://dx.doi.org/10.24018/ejbmr.2021.6 risman, a., ali, a. j., soelton, m., & siswanti, i. (2022). the behavioral finance of msmes in the advancement of financial inclusion and financial technology (fintech). the indonesian accounting review, 13(1), 91101. https://doi.org/10.14414/tiar.v13i1.3213 ruel, e., wagner iii, w. e., & gillespie, b. j. (2015). the practice of survey research: theory and applications. sage publications. safitri, y., luthfia, a., & ramadanty, s. (2021). social media influencer credibility and social influence on intention to use fintech in indonesia. proceedings of the 2nd international conference on interdisciplinary arts & humanities (iconarties). 263-269. https://dx.doi.org/10.2139/ssrn.3800662 samartha, v., shenoy basthikar, s., hawaldar, i. t., spulbar, c., birau, r., & filip, r. d. (2022). a study on the acceptance of mobile-banking https://doi.org/10.4018/978-1-7998-7603-8.ch001 https://doi.org/10.4018/978-1-7998-7603-8.ch001 financial services review, 33(3) 44 applications in india—unified theory of acceptance and sustainable use of technology model (utaut). sustainability, 14(21), 14506. https://doi.org/10.3390/su142114506 sandrine, p. t., perrine, r., & alexandra, n. (2019). service: a case study with a bank savings mobile application. intelligent system and computing, 273-281. http://dx.doi.org/10.5772/intechopen.89542 satsios, n., & hadjidakis, s. (2018). applying the theory of planned behaviour (tpb) in saving behaviour of pomak households. international journal of financial research, 9(2), 122-133. https://doi.org/10.5430/ijfr.v9n2p122 satsios, n., karamanis, k., galanou, a., & sotiropoulos, i. (2020). saving motives and intention towards saving of bulgarian pomaks. international journal of economics and financial issues, 10(3), 97. https://doi.org/10.32479/ijefi.9633 schueffel, p. (2016). taming the beast: a scientific definition of fintech. journal of innovation management, 4(4), 32-54. https://doi.org/10.24840/21830606_004.004_0004 setiawan, m., effendi, n., santoso, t., dewi, v. i., & sapulette, m. s. (2022). digital financial literacy, current behavior of saving and spending and its future foresight. economics of innovation and new technology, 31(4), 320-338. https://doi.org/10.1080/10438599.2020.1799 142 sharif, s. p., & naghavi, n. (2021). online financial trading among young adults: integrating the theory of planned behavior, technology acceptance model, and theory of flow. international journal of human-computer interaction, 37(10), 949-962. https://doi.org/10.1080/10447318.2020.1861 761 shefrin, h. m., & thaler, r. h. (1988). the behavioral life‐cycle hypothesis. economic inquiry, 26(4), 609-643. https://doi.org/10.1111/j.14657295.1988.tb01520.x shih, y. y., & fang, k. (2004). the use of a decomposed theory of planned behavior to study internet banking in taiwan. internet research, 14(3), 213-223. http://doi.org/10.1108/10662240410542643 singh, s., sahni, m. m., & kovid, r. k. (2020). what drives fintech adoption? a multi-method evaluation using an adapted technology acceptance model. management decision, 58(8), 1675-1697. https://doi.org/10.1108/md-09-2019-1318 skinner, j. (1988). risky income, life cycle consumption, and precautionary savings. journal of monetary economics, 22(2), 237255. https://doi.org/10.1016/03043932(88)90021-9 stavins, j. (2021). unprepared for financial shocks: emergency savings and credit card debt. contemporary economic policy, 39(1), 5982. https://doi.org/10.1111/coep.12477 strömbäck, c., lind, t., skagerlund, k., västfjäll, d., & tinghög, g. (2017). does self-control predict financial behavior and financial wellbeing?. journal of behavioral and experimental finance, 14, 30-38. https://doi.org/10.1016/j.jbef.2017.04.002 taber, k. s. (2018). the use of cronbach's alpha when developing and reporting research instruments in science education. research in science education, 48(6), 1273-1296. https://doi.org/10.1007/s11165-016-9602-2 tang, n., & baker, a. (2016). self-esteem, financial knowledge, and financial behavior. journal of economic psychology, 54, 164-176. https://doi.org/10.1016/j.joep.2016.04.005 thakor, a. v. (2020). fintech and banking: what do we know? journal of financial intermediation, 41, 100833. https://doi.org/10.1016/j.jfi.2019.100833 u.s. congress. (2021). s. 2600 – restoring access to opportunities for housing act. congress.gov. https://www.congress.gov/bill/117thcongress/senate-bill/2600 venkatesh, v., davis, f. d., & zhu, y. (2023). competing roles of intention and habit in predicting behavior: a comprehensive https://doi.org/10.24840/2183-0606_004.004_0004 https://doi.org/10.24840/2183-0606_004.004_0004 https://www.congress.gov/bill/117th-congress/senate-bill/2600 https://www.congress.gov/bill/117th-congress/senate-bill/2600 chen et al. 45 literature review, synthesis, and longitudinal field study. international journal of information management, 71, 102644. https://doi.org/10.1016/j.ijinfomgt.2023.102 644 venkatesh, v., morris, m. g., davis, g. b., & davis, f. d. (2003). user acceptance of information technology: toward a unified view. mis quarterly, 425-478. https://doi.org/10.2307/30036540 vermeir, i., & verbeke, w. (2006). sustainable food consumption: exploring the consumer "attitude–behavioral intention" gap. journal of agricultural and environmental ethics, 19, 169-194. https://doi.org/10.1007/s10806005-5485-3 wagner, j., & walstad, w. b. (2019). the effects of financial education on short‐term and long‐ term financial behaviors. journal of consumer affairs, 53(1), 234-259. https://doi.org/10.1111/joca.12210 wang, j. s. (2021). exploring biometric identification in fintech applications based on the modified tam. financial innovation, 7(1), 42. https://doi.org/10.1186/s40854021-00260-2 widjaja, i., arifin, a., & setini, m. (2020). the effects of financial literacy and subjective norms on saving behavior. management science letters, 10(15), 3635-3642. http://doi.org/10.5267/j.msl.2020.6.030 xiao, j. j., tang, c., serido, j., & shim, s. (2011). antecedents and consequences of risky credit behavior among college students: application and extension of the theory of planned behavior. journal of public policy & marketing, 30(2), 239-245. https://doi.org/10.1509/jppm.30.2.239 yeo, j. h., & fisher, p. j. (2017). mobile financial technology and consumers' financial capability in the united states. journal of education & social policy, 7(1), 80-93. https://doi.org/10.1509/jppm.30.2.239 financial services review, 33(3) 46 appendix measurements of key variables variables items references attitudes at1 having adequate (at least 3 months of your living expenses) in emergency fund savings would be extremely bad (1) – extremely good (7). ajzen (2006; 2020) at2 having adequate (at least 3 months of your living expenses) in emergency fund savings would be extremely unpleasant (1) – extremely pleasant (7). at3 having adequate (at least 3 months of your living expenses) in emergency fund savings would be extremely harmful (1) – extremely beneficial (7). subjective norms sn1 my family thinks that i should maintain emergency fund savings that cover at least three months of expenses. sn2 my friends who are important to me think that i should maintain emergency savings that cover at least threemonths of expenses. perceived behavioral control pc1 i am confident that i can easily maintain adequate (at least 3 months of living expenses) emergency fund savings. pc2 i have complete control over maintaining emergency savings (which are at least three months equivalent of living expenses). intention to use fintech to save in1 what is the likelihood that you will use financial apps to save regularly for emergency fund savings to cover at least three-month living expenses? fintech use saving app (sa) how often do you use financial applications (bank apps, mint, quicken, quickbooks, etc.) to save money regularly for emergency funds? finra (2021) saving website (sw) how often do you use a laptop/computer to save money regularly for emergency funds? emergency fund savings how much have you set aside for emergency fund savings? johnson and widdows (1985) expenses what are your household's approximate monthly expenses, on average? objective financial knowledge ob1 imagine that the interest rate on your savings account was 1% per year and inflation was 2% per year. after 1 year, how much would you be able to buy with the money in this account? houts and knoll (2020) chen et al. 47 ob2 if interest rates rise, what should happen to bond prices? ob3 considering a long time period (e.g.,10 or 20 years), which asset described below normally gives the highest return? ob4 normally, which asset described below displays the highest fluctuations over time? ob5 when an investor spreads his or her money among different assets, does the risk of losing a lot of money increase, decrease, or stay the same? ob6 do you think the following statement is true or false? "if you were to invest $1000 in a stock mutual fund, it would be possible to have less than $1000 when you withdraw your money." ob7 whole life insurance has a savings feature while term insurance does not. ob8 a 15-year mortgage typically requires higher monthly payments than a 30-year mortgage, but the total interest paid over the life of the loan will be less. ob9 housing prices in the united states can never go down. ob10 suppose you owe $3,000 on your credit card. you pay a minimum payment of $30 each month. at an annual percentage rate of 12% (or 1% per month), how many years would it take to eliminate your credit card debt if you made no additional new charges? subjective financial knowledge sb how would you assess your overall financial knowledge? finra (2021) pii: s1057-0810(02)00100-2 ���������� �� �� � �� � � ����� �� ���� �� � ����� �� � ���� ���� �� � �� ������� ��� �� ���� ��� � � � �� �� ��� � � � �� �� � � ��� ��� ��� ��� ���������� ����� �� � ����� �������� ������������ �� ���������� ��� �� ��� �� ��� ��� � � �� � !" #�� �� � !$$ % � � �� � �� � ��� � ���� !& � � �� � !$$ % �� �� � " ' �� �� !$$! �������� (� ������ �� ���� ����� �� �� ������ � ����� ) ��� ���� �� �� ���� ��� � � � �� �� ��� �� � ��� �� ���*������� � � ����� �� � �� �� ��� ����� � � � � ������ ����� � �� � + (� ��������� � � �� � �� � � � ��� � �� �� ���� ��� � � ���� � ��� � �� �� �� � ��� � �� ��� ������ �� �� �� �� ��� ��������� �� � �� ��+ �� ������ � ���� ����� � � �� � �� � �� ���� � �� �� ���*������� � � ����� �� � ���� ���� �� � �� ������� �� � ��� �� ��� ��� �� � � �� �� ��� �� � �� � �� � � �+ � !$$ ,�� ��� � �� � � �� � ��� �� -��+ �� ����� � ��� �! � % �. % �& % '!/ "��# ���! � ��� � �� � 0� ��% ���� �� � � � ����% # � ����� ������������% ���������� � ����% � ���� ���� �� � �� ������� � ���� ������ -� �� 1��� � �� � �� �� ���� � �� �������� � /" � �� �� � � � � � �� � ��� � ���� � � ���� �� ����� �� ���� �� �� ���� � ������ ���� � �� �� &.$�+ 2�������� �� �� ��� ���� � ������ 2�������� ���� 3!$$ 4 � � � �� � ������ ���� �� �� ���5 ����� �� �� � �� � �� � 6/ �� ��� � �� ���� �� �� � �� �� �� �� � ��� �� !$ $ �� ������ �� � )� .$ � �� �� ���� � �� � � � 2� ��� �� �� ���� ���� + �� ����� �� !$ " �� ���� � ������ 2�������� ���� �� � ��� � ���� ��� �� � � ��� �� � �� � �� �� �� � ) � �� �� !$.6� �� ����� ���� �� � )� ��� �+ (� � �� ������ � �� � ��� ��� �� � ��� � ���� �� �� ���� � ������ ���� �+ 7�� ��� � ���� � � �� � $ 3!$$ 4 ""86. $"6*$9 $:$ :; 8 � ����� � �� � � !$$ ,�� ��� � �� � � �� � ��� �� -��+ ,--< � $ " 6 * $ 9 $ 3 $ ! 4 $ $ $ $ * ! � =��� �������� �����+ ( +< � *..$*&6!*">./% � )< � *..$*&6!*"&6$+ �$���� ���������! ? � �@� ?���+ �� 3�+�+ � � �4� ��� ���@� ?���+ �� 3�+ ���4+ ( +< � *..$*&6!*"./ % � )< � *..$*&6!*"&6$+ -� �������� a���� �� a��� 3 &&94 ��� �� � �� ������ ���� �� � � � ����� � ��� � �� ����� ���� � �� � � � �� � ����� � �� �� � �� � ������������ � �� � � �� ��� �� � �� � ����� �� ���? �� � ��� � �� ���� �� �� �� �� �������� ���� �� ��+ (��� � �� � ��� ���� � �� 0� ������< 3 4 ��� � �� � � � � ���� �� �� �� ���� ���� � ���� ���� � �������� ���� �� 3!4 �� � �� � ?��� ���� � ��� � �� �� � ����� ���� �� ���� �����+ (� # ���� -������� �� 2���� 3#-24 � � ������� � ����� � � �� � �� �� � ��� � �� ����� 3���4 �� � ����� �� � �� �� � � � �� ���� ��� ��� �� ����� � �� ���� ��� ������ ��� ������ ���� �� � ��� � �� � ����� �� ��� ��� ������ �� + �� ���� �� � ��� �� ����� ��� �� � �� � ��� ��� ����� �� �� � ����� �� � ��b� �� �� � ������ �� �� ������ �� �� � ��� � �� � ������ �� �� �� �� �� � �� �� �� � ����� ���� �� ���� ����� + '��� � �� ��c� � 3 &&"4 � ����� �� � � ��� �� �� �� ���+ -� �� �� � �� && ���� ������ �� ���� �� �� � " 8/ � �� � � �� �� )� �� � �� � � � � � � �� ! � �� ��� � �� $*� � � ����+ (��� ����� �� ������ ��������� � ���� �� ����� �� �������� � ��� � � �� �� � ��� � �� �� � ��� � �� �� � ) ��� ���� �� �� � ����� �� �� �� � ������ �� � ��� �� �� �� �� ��� ���������+ (� ���� ��� � � � 3 &&!� &&>� �� &&/4 �� �� ����� � � ��� � � � � � �� �� 1��� ����� �� a����� � ����� � � �� � � �������� ��� ��� ��� ��� � �� ����+ d�� �� � � � � � ��� ��� � ����� � �� � �� �� ���� � � �� � � � � �� ��� � � � � ��� ���� �� ��������� � � �� � �� ��� � � � + (� ������ �� ���� ����� �� �� ������ � ����� ) ��� ���� �� �� � ���� ��� � � � �� �� ��� �� � ��� �� ���*������� � � ����� �� � �� �� ��� ����� � � � � ������ ����� �� � + 2 ���� ����� � � ��� � � � �� �� �� � �� �� �� ��� �� �� ��� � ������ ��� ��� �� � � �� �� � ��� � �� ������ ��� � ���� " 8/ � �� �� � �� && �� "/8// � �� �� � �� &&/+ (� ����� � � �� � � �� ���� � �� �� � ��� � �� ���� /! �� /" � �� �� � + (� � �� �� ���� � � � �� ��� ��c � ���� ���� � ������+ � ����� ! ������� � �� � � �� )������ �� � ��� �� � ��� � �� � ������ � ��� � �� � 0� ��� �� � �� ����� � �� � ���� � �� �� �� � �� � ��� � ��+ (��� �����e� ��� �c ���� �� ��� � � �� � ����� � �� � ����� . ��� ���� � )� � ���� �� �� � ����� ��� �� �� �+ � ����� > ������� � �� � �� �� �� �� � ���� �� � ����� " ������ � ���� ������ �� �� �����*� ����� �� ��������� �������+ � ������ �� ��� ���������� (� � � � � � ��� ��� �� � � ��� �� � ��� � �� � ����� �� ��� � �� ���� �� �� �� �� �� � ��� � � � � � � ���? �� ���� ��� �� ������ �� � � ����� ���� ��������� �� � ���� ���+ f� � �� � ����� �� ���� ���? � � � �� � �� � �� �� ����� �� ���� ����� �� � �� ���� ���� � ��� ��� � �� ����� ��� ����� �� %����# �� �� �� / � ��+ &'(' %��������� �������� �� � �� ��� � ����� � ? ���� � �� �� �� � �� � ��� � �� � ���� �����g � � � �� ���� �� 3 &&&4 ��� �� ���� 0� ����� �� ) ������ �� � ����� � �� ��� 0� ������ �� "/ "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 �� ���� � � �� �� ���+ 7����� � ����� ��� � �? � �� �� � � � ������� ���� � ��� � �� ���� �� ����� � ��� ����� ��� ��� � ��� � �� � �� � � + # )�� � ����� ��� � �? � �� ��� � � � �� �� �� �� ����� � ����� ���� � ��� � �� � �����+ � �� �� �� � � �? � �� �� � � � ���� � ���� �� �� �� �� � ��� � �� �� � ����� � �� � ���� �� ��� � � � � ���� �� �� ��� � � ���� �+ (� � ��� �� � ��� �� �69 �������� � ��� � � ��� �� �� �� � � � � �� � ��� � ��� � �� �� � � ������� �� ������� ��� �� �� �� �� � ���� �����+ d�� ���� � �� � �� ��� � � �� �� �� ������� �� ��� �� ���� �� � �� � �� �������� ���� ���� �� �� ���� ���� � ����� � �� �� � � � ����� � ��� �? � �� � � ���� �+ (��� ��� � ���� � �� � ��� � � �� �� �� �� �� �� �e� � ������ �� � � �� � �� �? � �� � � ���� �+ 1���� �� � � ��� � � � � � � � � �� ���� �� 3!$$$4 � ? �� � � ���� � ��� � �� ��� 0� ������< 3 4 �� �������� � �� )� �� � � ����� � � ��� �� � ��� � �� � ��������� � ��� ����� ���� � �����g� �� 3!4 �� �������� � � � � ����� ���� �� �� �� �� � ��� � �� �� �� �� �� ����� � � � �� �� �� � ����� � � � ������ ���� �� � �����g (� �� � �� �� ����� � �� � �������� � �� �� � )� �� � � ����� � � ��� � ��������� � ��� � �����+ ��� � �� ����� � � � ����� � � �� �� � �� � � � ��� �� �� �� )� �� � �� �� �� �� ����� �� � ��� � �� ��� ��� � ����� � � � ������ ����+ h� ���� �� � ������ ����� �� ' �0� � 3!$$$4 � ���� �� ����� �� �� ��� � �� �� �� ��� ���� ���� ��� ���� �� � ��� � ��� � �� � �����+ (� � ��� ������ ���� ���� �� � ���� �� � ���� � ��� �� � ��������� �� >i ��b ���� � � � 9+"i � ���� �� ��� ��� ���� �� � � � � �� � ��� �� 6$89&i+ (� � �� ��� ��� ���� ���� �� ��� ��� � ��� � ��� � �� �� � �� ���� �� �� � ���� � � � � �� �� ���������� �� 0���� � �� � � � �� � ��� � �������� e� ������ ��+ &'&' %��������� ���2�� � j��� � �� � �� a��� �� 3 &&94 ) ��� �� � ��� � �� � 0� �� �� � �� �� ���� �� �� �� � � ���� � � �� �� 7 � � � � �� k� �� �� ��� ��� ����� ������� � �� &&"+ (� ����� �� ��� �� ���5������� ���� �� � � ��� �� �� �� (� ���� �� �� � � �� &&" ���� � �� =����� � 7�� �� � 3�=74 � � ���� � �� � ����? � �� ��*a�= � �� �� ���� � 3 &&64+ (� ������ �� � �� � � ��� � �� � 0� �� �� �������� �� �� � ����� � � � � �� � ��� � �� �� � )� �� �� � ��� � �����������+ (� � ��� �c � � ������ ������ �� ���5 ������ �� ���� �� ������������ � � � �� ��� �� � "!i �� ���� �� �� � � 0� � � �� � � � ��� � ��� � ��+ -� � � ������� �� � ����� � �� � � � � � �� � � �c �� � � �� �� ��� �� ��� ���*�� ��� �� �� �� � ��� �� ��� 0����� �� �� �� � �� � � �� ����� ���� ���� � ������ �� ��� � � ����� �����+ j�� � + 3 &&94 � �� � � � 0� ������ ���� �� � � ���� ���� �� � ��� � �� � 0� ��+ 7����� �� � ���� � �� � � � � �� � � 3 � �� �� *� ��� � �� ����� � � � �� � ��� � ��4 � g ����� � � ���� ���� �� a���� � �� a��� � 3 &9>4 � �� ���� 6$i �� &$i �� , � � 3 &9&� &&>4 �� � /"89"i+ � ����� ���� � ��� 0���� � ��� �� � �� � ��� � �� � ��g (����� �� ���5 ����� ����� ������ � � � ��� �� ��� ���� � �� �� � � �� � � ������ � � g 7������ ���� � �� � �� ����� ��� ������ ��� � ���� � �� � ��� �� �� � � ��� � ��g 7����� �� � �� �� � � �� �� � ���� � � � � � ��� ������� � � ��� � ��g "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 "6 ��)��� � �� � � �� � ��� � �� �����g � �� �� ���� � � ) � � ������ � �g � � ��� �� ���� � ���� � �� ��� �� �� �� 0� ����� �� � ��� � �� � 0� �� � � � �� �� �� ��� �� �� �� � 0� ������+ a���� �� a��� 3 &&94 � � �� �� ����� �� � ��� � �� � 0� �� �� �� ���� � � �� �� ��� ����� �� � � � � ���+�����+ ��:������� d ���� + (� � ��� �� � �� � �� � 2� ��� � ���� �� � � � ��� ��� ����� � � ��� ������� � � ��� � �� ���� �� � ��� � ���� �� � � ���� �� � ;6.�$$$ �� � � �� ��� �� �� �� � )� �� � ����� � � ���� �+ (� �� ���� � �� �� � ���� � �� � ���� �� � � �� �� � � � ������� !$i �� ����� �� ��� � � � � � �� � � ������ �� �� � ��� � � /! �� !i ��� � ��� � �� � � /"+ 7�� ���� � �� � ������� !&+.i �� � ����� �� � � � ��� � ��� � �� � � /! �� 6+/i � /"+ ���� � � � � � �� � � � ������� . +"i � /! �� 9+9i � /"+ (� � ��� ����� � ��� � �� � �� ����� � �� � �������� ���� �� �� �� ��� � � ������ �� � ��� � �� � �� �� ������ ������� � � ��� � �� ���� �� ���� ���� � ��� ���� ��� ���� � �� � ���� ���� �� �� �� � � � ��� �+ &'1' ����� ��� ��� � � ��� ��� �� 3 &&94 ��� �� � ��� ���� � �� � ��� �� � ��� � �� � ��< 3 4 �� � �� �� �� � �� ��� � ���� � � ��� �� � �� �� ������ ��� �� 3!4 ���� � � ��� *� ) �� �� �*� ) � � � � �� � ��� ������ �� � � ����+ � ��� � �� � �� �*� ) � � � � �� � � � �� �� � � b ��� �� ����� �� � �� � � � � �� ��� ����� �� � ���� �+ (� �� ��� � ���� ��� �� ��� ���� � ����� �� �� �� �� � ���� �� ����� �� �������� ���� �� � � ) � � �� � ��� � ��+ (� �� �*� ) � � ����� ��� � � �� ���� �� � � ������ �� � ��� � �� � �� ���5 ������ �� ���� � �� �� �� � �� �����+ � ) ��� � �� ������� � � ��������� �� ��� ����� � �� � ��� �� ��� ��� ����� �� �� � � �� ���� � ������ �� � �� �*� ) � ���+ (��� ���� ������ �� � � � ��� � ����� � ����� �� � � ������� �� ������� � �� � ��) �� � ) �� �� �� �� � ��� � �� � ��+ -� � � ������� �� �� � �� ���� ��� � �������� ��� �� � ��� �� l�� �� a�h����� 3!$$$4 ��� �� � �� � ��� �� ������� �� � � �� ������ ��� � � � �� ������ �� ��� �� � �� ����?� � �� � �� � �����+ (� � � ���� �� �� � � ��� � ���� � ������� �� � ��� �� � �� ����?� � �� � �� � �����+ 2�� � �� ���� �� -������� � ��� � �� 2������� 3-�2�4 �� ) ��� � �� d �� �� ��� �� 3!$$$4 �� ����� � � � �� �� �� ����� � � ��� -������� 3��-4 ���� � � � � � + , ������ ��� �� �� � �� � � � � � � ��� ����� �� ; $!�$$$� ��� �� �� ��� ��� 3&!i4� � � ��� � 39 i4� � � ���� � � �� � 36 i4� � � � � ����� �� ;6!"�$$$� ����� ��� �� � �� �� ��� �� � � � � � �� " + (� � ����� �� � �� ����� �� ���� � � � ���� � �� ����� ���� �� �� � >"� >"8">� ""8/>� �� /"� �� ��� �� � � � ���� � �� � ��� ������ �� 0����+ (� ������ ��� �� � �� � ���*������ �� �� ��� ����� ��� �� �� ����� ���� �� � � ��� �� �� � ������ �� �� � 0����� ��������� �� ��� ����� ���� � �� �� �� -�2�+ 7���� �� .>i �� �� �� � �� � �� �� !/i �� � ������ ��� � � �� �� �� ����� �� � ��+ (��� �� ����� �� � ���� �� � �� �� � .>i � � �� 0���� �� �� �� -�2� �� ./i � 0����+ d �� �� ��� �� 3!$$$4 ����� �� �� �� � �� �� ���� ����� �� k��� �� =� � 3 &&64 ��� ) ��� � � � �� � �� ���� � ������� �� � ��� � �� ������� �� (-22:=��7 � ������ ���+ k��� �� =� � ��� ��� �� � >$i �� �� �� �� � � �� ����� �� � ��� �� "9 "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 http://www.umich.edu/�hrswww (-22 �� ��+ 7���� � �� ��� ����� � ��� �� � � � � �������� � �� � �� �� �� �� 0���� �� ��� 0���� � �� �� � � � �� � ��� ����� �� ��� 0���� � �� �� � � ���� � � � ����� �������� �+ 7���� � � ��� � �� � ��� �� � �� �� �� �� � �� ���� � �������+ (� ���� � � �� �� ��� �� �� � �� � ��� � �� =� �� �� �� ���� � 3!$$$4 �� � �� ��� �� �������� �� �� = ��� 2�� � ,������ a�� 3=2,a4+ 7����� ���� ��� � � � ���� � �� ���? �� � �� �� � ��� �� �� ���� ����� �� ����� � ���?*�� �� ��+ � ����� �� ������ �� �� ���?� �� �� �� ��� �� �� � � �� �� ���� � ���? �� � �� + (� � ��� ������� ��� �� �� �������� �� �� =2,a+ 7���� �� �� � ��� �� � ���� ���� ��� ��� ���� �� � ��� �� �� �� � �� �� ���?� �� �� � �� d �� �� ��� �� 3!$$$4 �� k��� �� =� � 3 &&64+ -� �������� � � � � ��� �? � �� ��� �� �� ���?*�� � ������ � �� � � �+ 7� � �� ' ������� �� ���� 3!$$$4 ) ��� �� �� � �� ���� � ������ �� ���� �� �� �� �� �� ���� � ������ � � ��� �� � �������� e� � ��� � �� ������ ��+ (� � ��� �� � �� �� � �� � � �� �� ���� � ������ � � ��� ���� � � ��� �� � �� �� � �� �� �� �������� � ���� �� � �� ����� � � � �� �� � �� � �� (� ���� -�b ���� ,��� ����� � ������ � 3(-,�4+ 7���� �� �� � ��� �� � �� �� � � ��� � ��� �� � �� �������� ���� � ��� �� ��� ����? �� �� ������ ��+ &'3' 4����� �� ���� �� (� �� ��� � �� � ����� �� ���? � � � � �� �� � ��� � �� � �� ���� �� �� �� �� �������� ���� �� �� � � � �� ������� �� � �� � � ����� � � ��� �������� �� k 5� ���� �� l �� �� � 3 &&64+ (� ���� �� ������ � �� � ���� � �� � �� �� � � ������� �� � �� � �� k 5� ����� k �� � ?� �� '� � ?�� �� 3 &&&4+ (� � �� � � ���� �� &9& �=7 �� � �� ��� ���� � �� � �� �������������+ (� �� � �� �� ����� � �� � ��� � �� � !6i �� �� �� � �� �� � �� � ������������ � �� � ���� ."i ��� � �+ d�� � �� � �� � � �� � � � ��� ���? � ����� �� � � �� � �� � ������� � �� � � � � �� � ��� � � � + 3!$$$4+ -� � ��� �� � �� ������������� �� � ������� ���� � � ��������� ��� �� � � ��� � � �� ��� ���� ��� � ��������� �� �� � � ��� � ��� �+ -� � � �� � ���� �� �� 1+�+ = ���� k�� �� ����� �� (� ��� � 3 &&&4 ������ � � �� �� ������� ��� �������� � ��� � "" � �� �� � � �+ 2� �� ���� ���� � �� � � �� ��� � ������� � � �� ���� �� � � ��� &! � � ��� � �� $$ ��� � �� �� � �� �? � ""8/>+ ������ �� � )� $ � �� ���� /" �� 6> � �� � �� �� � ��� ����� �� 9! � �� �� �� � �� �� �� �? � 6" �� 9> �� /& � �� �� �� � ��� ���� 9" �� �� � �� � � >& � � ��� � �� $$ ��� �+ a � "" � �� � � �� �� � � ��� �? � �� � � ��� � �� ����� ���� �� �� ����� �� � ��� �+ 7�� ) �� � ��� ���� ""8/>� 66i �� � � � � � ��� � �� ����� ���� �� �� ����� � ���� />i �� ��� �+ (� � �� �� ������� �� ���� � �� � �� � � � �� ��� � � ��� ���� ����� �� �� � � ��� � �� � � � ��� �? � �� � � � �� ��e� � �� + 7�� �� � ����� ""8/>� !6i �� � � � � � �� ��e� � �� �� &i �� ��� �+ 7 �� � ����� �� �� � �� � ��� ���� 6.+9i �� � ��� � ���� � � ���� � ����� �� ;."�$$$ �� ��� ��� �� � ����� ""8/> � �� � �+ (��� ����� �� " +>i �� �� /"86> � ����� �� ./+9i ��� ���� 6" � �� �� �� �+ -� �� ""8"& � ������ 6&+ i �� � � � ���?��� �� �� ���� � � ��� ���� �� /!+>i �� ��� �+ 2� ����� � )� �� �� �� � � �� �� � � �� � ��� � ���� ��� ���� /$8/> � �� � �� ���� "/+ �� � � ���?��� �� "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 "& .&+!i �� ��� �+ (��� ��� � /" � �� �� �� � � � /+!i �� � � ���?��� �� &+6i �� ��� �+ &'.' ��� ���������� 2� �� � � �� �������� � � � �� � ��� g -� �� � �� � � �� � � � �� ���� �� � ��� � �� � ���� �� � �� ���� � ������ �� /! �� /"g a��� ��� j��� �� � �� 3!$$$4 �� � � ���� �� &&" �=7 �� ��� ���� � �� � � ���� ��� �� � �� � � ��� � �� � + (� � ��� �� � " i �� �� � �� � � �� � ��� �� � /! �� 9&i �� � /"� ���� �� � ������� �� � ��� � �� � ��� �� � � �������� ����� ���?��� �� *��� 3." � � ? �� ��� 4+ (��� ��� � � �� � ��� � � � � � ���� � � � �� �� ��� �� ��� ���*�� ��� �� ��� �� ��� � ���� � � ����� ������ ���� � �� � � � �� � ����� ������� �� � � �� ���� �� � �� � � ��� � �� � + f���� ���� �� � ��� �� � �� � � ���� �� � �� � � ��� � �� � � ��� � ��� k �? �� ���� ��� ������ � �� � � ��� � �� � �� � �� ��� � ��+ 2������� �� �� � � � �� �������� � �� � � � �� � �� � � ��� � �� � + (� ��� � �� )� �� � �� � �� � � � �� � �� � � ��� � �� � + d� �� �� 3!$$$4 �� � ��� �� � ������ �� �� � �� � ��� � � � ������� ���� �� � ������� �� �� � � �� ���? � � ������ �� �� � �� ��� �� ���? �� ��� � �+ ��� ����� �� �� �� � �� � ����� �� ���?��� �� ��� � � � �� � �� � ��� � ������ � ����� �� � ��� �� �������� ���� ������ � ��� � � ���� ��b ����� ���� � ������ ���� �� � �� � � ���� ���� � � �� � ��� � �� � ��� �� ���*� ���� �� ������ � �����+ (� � �� �� ����� � �� � �� ��� � �� )� ��� �� �� � �� ���� � �� � � � �� ���� ���?��� �� ��� � �+ ��� � �� � ���� � �� �� � �� ��� �� �� ����� � � � � ����� �� �� �� �� � ������+ (� � �� ���� � �� � �� � ��� � �� � ����� ������ ���� � �������� � ���� � �� � � � �� � � ��� �? � �� � ��� �� � � � � �� � �� �� �� ���� �+ (� � ������ ��� � �������� ���� � �� � ������������ � � �� ��� �? � �� � �� ���? �� ��� � �+ (� ����� � �� � ��� � �� �� � � � �� � � �� ���� � ����� ����� �� �����*� ����� � ���� �� � � � � � � � � ��� ����� �� ���� � � �� �� ���� � � � ���� � �� � �� �� �� �=7� �������� ���� ��� ���� ������ ������ ��� (-22:=��7 � � � �� 1+�+ = ���� k�� � � � + (��� �����e� ������������ �� � ��� ��� ����� �� �� *� ��� � �� �� � ��� � �� �� �� � �� �� � ���� �����*� ����� �� ��������� � � �� ������ ������ �� ���*������� � � ������ �� � �� ����� � ���� ���� ���� � �� � �� � �� � ��� � ��+ -� �� ������ � ������� ��� �� �� �� � �������� �� � �� �� ��� ����� � � � ����� � �� � �� ���*������� � � �����+ 2 � �� � �� � � � ��� � � ��� �� � �� �� ����� � ��� � �� � � ��� �� � ������ �� � � �� �� ����� � � ��b� �� � ������� ���� �� � �� ����� �� ����� ���� �� �� ��+ � ���� �� ����� ����� 1'(' �%� ���� (� � � ��� �� ���� �� ����� �� ��� �� � ���� �� ��� � � � � � ����� ���� �� ��� �� � !�/$$ �������� � �� 6�/$$ ���� �� �� �� � � �� �� �� �� � ����� ��� �� �� � �� /$ "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 � �� "$+ (� � �� � � ��� 6�/$6 ���� �� �� ��� �� ���� � � � 6�!!6 ��� �� � ���� � � � �� /�9 / ��� �� ����� � � + (� ���� � �� ���� �� �� � �� � � �� ���� $i � �� � � � � �� .+ =��� �� ��� � ����� �� �� � �� � 0��� �� � ���� �� �� � ��� �� � �� � �� �� � � �� � �� �� ��� � � � �� � � �� � � � ���� ��������� �� �� ��� � � + ���� �� �� � )� �� � �� �� � ���� � �� ��� �� �� � �� � � ����� � �� � �� � � �� ����� �� � �� � � � ��� �� �� �� �� ��� �� � �� �� ��� �� ��+ (� �� � �� ���� ��� >�&$. ���� �� ��+ (��� � �� �� ����� � ���� ��� � � ���� �< � ��� � �� ���� + (� � ��� � ����� ��� �� � .�/./ ���� �� �� ��� � � � ��� � �� �� � � �� ������� � )� �� �� ���� ����� ��� �� � �� � � ����� �!/6 ���� �� ��+ (� ��� � � ��� �� � �� ���*������� � � ����� �� ����� �� �� ���� �� � �� � ���� ���� ������ ���� � � ���� � �� ��� ���� �� �� ��� � ����� ��+ (� ���� �� � � � ����� �� � ��� � ��� �� � � �� ��� �� � �� �� ����� ���*������� 0����< � ��� 3�4� �0��� �� �� ���� � �� �?���:� ������ ��� � � �? � ������ =��� ��� ��� �� ����� �� (* �� �� -�2 �� ��fh��� ����? �� ���� ������ ����� �� 0����� ������ �� ��� � � � �� ��+ ������� 0���� �� )� �� � � � �� � �� ����� � ����� � �� � �������� � � ��� � ���� ��� �� �� � �� ������ ������� 0���� � � �� �� � ��� � �� � � �� � �� � � �� � � ����� �� �� �� ��� �+ (� && ���� �� � ����� ���� �� ���� � � �� � ���� � �� &&!+ (� ���� �� � ����� ���� ��� �� � �� ����� �� ���� �� � ����� �� �� � ��� � 3�� �� � �� ��4 ���� ����� �+ (��� ��� ��� ����� � � ��� �� ������� �� �� ��� �� �� ���? ��e ���� �� ����� � ������ �� ������ �� ���� � ������� � � � � � ��� ����� � ��� �� ��� ����� � �� ��� � ����� + � ���� ���� ������ ���� ��� �� � � �� �� � �� �� ��� ����� � � � � ������ �� � + 2 � ���� ���� ������ ���� )� �� � �� � �� �� ���� � � �� &&!+ 1'&' 5 ��� ���� �� �������� �)� ���� (� � �c �� ���*������� � � ����� ��������� � �� �����*� ����� �� �� *� � 2#fl2 �� �� �+ (� 2#fl2 � � ���� ���� � �� � � �� � �� ���� � �� � �� ������� � ���������� �� �� ���� � �� � ��+ (� 2#fl2 �� ��� ��� �� ��� )� � ���� � �� � � �� � � �c � ���� � � ��� � �� �� ) ��� �� � �� �� ��� �� �� � �� � �� �� ��� � ����� �� ��� �� ��� �� ��� � �� � �+ 2��� �� �� �� �� ����� 0� ������ � ������ ������� �� *� � 2#fl2 ���� �� � � �� �� � �� � � ��� ���* ������� � � �����< + -� �� � �� ���� � �� �� �� ���*������� � � ����� �� ���� �� �� �� �� � � �� � ��� �� ���� �� � ����� � �� � ���� ���� �� � �� �������g !+ -� �� � �� ���� � �� �� �� ���*������� � � ����� �� ���� �� �� � �� � � � � �� � ��� �� ���� �� � ����� � �� � ���� ���� �� � �� �������g .+ -� �� � �� ���� � �� �� � �� �� � �� � ���� � � ����c � � � ��� � �� ���� ���� �� ��g (� )� � ���� � �� � � ��� �� 2#fl2 ��� ��� �� � �� �� � �� �� ��� ����� � � � � ������ �� � �� �� ���� �� �e� �� ��� � ����� ��� � � � ���� �� � ����� � + "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 / 1'1' ��� ����� � �6������ �� �����7��� (� ��) )� � ���� � �� � � � � �� � � ����� � ���� ��� �� ��� ������ � ����� �� ( � + h �� � � � � � ��� �� ��� � � � � �� �� �� ���*������� � � ����� ����� ���� �� ���� � ���� � ��� � �� �� �� � ��� � �� � ���� �� � �� k 5� ���� �� l �� �� � 3 &&64� k 5� ���� � + 3 &&&4� a���� �� a��� 3 &&94� �� ����� �� (� ��� � 3 &&&4+ (� ��� ���� � �� � �� ����� � ���� ���� ������ �� � �����c �� ���� � �� � �� ��� ���� ��� � � �� � ���� � �� ����� �� (� ��� � 3 &&&4+ (� � �� � �� � ��� ���� ���� �� �� � �� �� � � � ����� � �� � + � � �� ����� � ���� ��� �� ������ � � ���� � �� = �� �� �� 2���� � 2� ��� �� �� ��� ��+ (� �� �� �������� �� �� ��� �� �� 2�� � �� ���� ���:� ����+ (� ����� �� � � � �� �� ������������ �� �� � ����� ��� �� ���� � �� �� �����+ k ��� � �� �� �� � ���� 3 &&&4� h��� �� 7�)� �� a��� �� 3 &&&4� �� , �� �� �� � ���� 3!$$$4 ) ��� ( � � ��������� �� )� � ���� � �� � � l �� � � � ��������� h�# h �� � a 7 � ! ��1 ���� ���� � �� �� � ���� ����� �� � �� �� ��� ���� m ! ���� ����� ! �� � �� �� ��� ���� � ! �� ��� ����� �� � . �� � �� �� ��� ���� n ! �� m / 7����� �� � �� ��� > �� � �� �� ��� ���� n / �2= � � d��� �� = �� �� � k �? �� 2���� � 2� ��� � ! 2 ��� �� . ��� � ����� ,��� �� ��< � ���� ���� �� �� h � � ,��� �� ��< ,� ������ �� 7��� � �� ��� ! = ��� ��� � �� �� f������)� �� ' ���� . #��*(� ������ =������ �� �� #� � ����� > #��*'�� � =������ � " -#= ���� �� � ����� 3 && ���� �� � ����� 4 ��� � !" � �� ��� �� ����� � ;!>�!"9 k �� � !" �� "$ � �� ��� � ! �� ;!>�!"9 m ����� � ;> �">9 k �� � "$ �� 6" � �� ��� � . �� ;> �">9 m ����� � ;/>�/$$ 1�� � !" � �� ��� > �� ����� n ;/>�/$$ 2h� 2� � �� &&! j���� � �� � � � � � �� � � "" f � � �� � � � � � ! �� � n "" /! "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 ���� � �� � �� ��� ��� �� ������ ��� �� � � �� ���� �� �� � 2���� � 2� ��� �� � � � ��� � � � ���?� �� ��� �� �� ���� � ����?� �� ����� ���� � � �� = �� �� ��+ d�� �� � �� ��� �� ���� � ����� �� � ��� � �� � ������ �� �� ���? ����� � � �� � �� � �� ���� � �� �� ���*������� � � ����� �� � � + 2 � ����� � �� � �� � �� ������ � �� ����� ���� ������ � � �� �� � �� �� � � ��� � � � �� � ���� ��� �� � ��b� �� � �� ��� )����+ 7�� ������ � ���� � ��� ���� � �� � � � � �� � � ���� � �� �� ��� � � �� �� ���� �� �� � ����� ��� �� �� � � ����+ h���� �� ��� � ,��� �� ��< � ���� ���� �� ,��� �� ��< h � � + ��� �� �� �� ���� � ������ ����� � ��� �� � � =���� � ���� � ������� � �� �� �� �� � ���� � �� ,� ���� �� �� � ���� �� �� 1��� � =����� �� =�����+ (� � ���� ����� �� � � ,��� �� ��< ,� ������ �� ,��� �� ��< 7��� � �� ���+ �� ���� ����� � �� ���� � ���� ,��� �� �� � � ���� ��� �� 2a� o���� k ������ ������ � �� =������ a �������� a ������ � =����� �� �� k� ��� �� =����� �� h��� 7��� � �� ��� k ������ , �� ���� � � � ���� 2���� � � ��� � � 2�� ������ ����� �� k ������ �� a������� ����� ���� � �+ (� ����� ����� �� ��� � = ��� ��� � �� �� f������)� �� ' ����+ (��� ����� ��� �� � ��� � = ��� ��� h� ? ��� = ��� ��� f������)� �� ' ����+ h���� > �� � � #��*(� ������ =������ � �� #� � �����+ �� ���� ����� � ��� �� =������ � ��� ������ ' ��� �e� d��� �� �� a������� p� ? ��� ������� ����� 1��� �� �� �� #� � ����� �� ��� �+ h���� " �� ��� � #��*'�� � =������ �+ (��� ����� ��� �� � -� �� a�� ��� k�������� ������ �� k � �+ -���� �� �� � ����� � ���� �� ���� 0� ��� � �� �� � � � � � �� �� ���� � � �� &&! �� ����� �� ( � + ��� � �� �� ��� ������ � � �? �� �� ���� � )� �� ���� ��� ����� �� � � �����+ h��� � � + 3 &&&4 ���� �� �� � � � ����� �� ������� � � � � �� ����� �� �� � � ����� � � ���� � � ��� ��� �� ���? �� � �� ���� � �+ k � � �� � � � � �� � ���� ���� �� � �� �� ���? �� ���� � � + 3 &9>4� , � � 3 &9&� &&>4� �� h� ���� � � + 3!$$$4� �� ��� � �� � �� � � ��� � �� � 6" �� &$i �� ����� ��� � � ������� � �� � ��� � ��+ 7�� �� �� �� 0� ��� � �� � ��� � ���� �� � ��� � �� ����� �� � �� � ; 9� &. �� ;! �9.!+ 7�� �� ���� �� 0� ��� � �� � ��� � � ;>9�>"$8;"9�!>$+ -� �� ���� � �� � ���� � ������ � � ������ ��� �� �� � � �� �� ����� � ��� �� ���� ��� 0� ��� � �� ����� �� � �� �� ��� 0� ��� �+ (� �� � �� � � � � �� ����� � � � � �� �� � �� � � ���� ���� "" �� �� �� ���� �� � � �� � "" � �� &&!+ (� �� �� � �� ���� � �� � �� �� �� ��� �� �� � ��� �� �? �� �� �� � �� ����� �� � � ����� ��� � � ������ ����+ -� �������� a��� �� � + 3!$$$4 ���� �� �� � � �� � � �� � ��� � �� �� �� � � � ���� � �+ � ������� 3'(' ��� ������� �������� � ��� # ��) ( � ! ����� � �������� �� ������� ��� ���*������� � � ����� �� �� ��� � � � �� ��� �� � � � �c �+ (� ��� � �� ��� �� � �� ���*������� � � ����� ��� �� � �� � �� � ��� �� � � + # � ����� �� �� ����� ��� �� � ��� � �� ���� � � ���� �+ (� � � ����� ���� � � ���� � � � �� ! �� �� � � �� � � ���� � � � ! �� . ��� ���� � �� � � � ����� ���� �� �� � � ��� � �� ���� + #�� �� � �� � � ��� � �� � "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 /. ( � ! � �������� �� ������� �� � � ����� ���� ��� ��� � � � l �� � � � � ��������� #�+ �� ��� �� ���� d � d � ! d � . d�� a ��� � ���� d�� a ��� � ���� d�� a ��� � ���� d�� a ��� � ���� h�# a !/!& !!>6 .9! .!& !+> >$>&.+/ 99. 9+/ "96&.+9 /& $&+& &9 !+/ .99>"+6 >6"!6+. 96669+6 ! 7 � !!6> .9& 99" &. .>+6 699 + ">!&"+. $>$$6+/ !9 /!+/ //$&/+6 9.&6$+! $.&" +! "!/ $+. ��1 � �� ��� � � $6$ 69/ !9> "$"9!+" "66$ +& .$969+9 >9 /"+ ">9 &+> !&6>9+9 > 6 6+. >&..$+9 !$/>/+ ! 6!9 !9& >.& & ///+! $9 9&+9 >. >&+. $&!/$+. !9!. +& "."""+6 96."6+! $ &. +. >>"/>+" . $!! 6>6 !6" ./!&&+! ">$>6+. 99$9&+$ "!/6.+ 6 969+6 $$"$.+6 !9// +! >6 9+ 69"!"+" > a�� ��� � � $9. 9 > !/& &..>6+$ !!$"/9+$ $&6"+9 !.69/9+9 !6 699+$ ."!!9+> ! .>&+ !>$"//+6 !!&."+& �2= d��� >$&. . "! &> !6 !+6 > &. +" 66>6"+/ >6.>&+! />>. +$ &$ . +/ !/9/ +$ > 9..+6 6/6$9+$ ! k �? /". .6> !6& >/9"9+$ /!!6>+/ !/ &!+ "!!>6+6 /& !6+> !&/!$+> .6!/9+6 >9>9.+9 !!!.>+9 . f�� �� "6 $ >6 /"996+" 6&6$.+! ..""!+& $/66&+> !/$/&+& / /. +. 69&>&+/ &6 6>+9 ./!&>+6 ��� ,��� �� �� &/6 6> !!/ "$//!+. 6$$./+. 96 .&+6 6&"6!+9 !$ $$>+9 $&.$!+. "/6."+9 6/&!$+ &$""/+> ! ,��� �� �� -! // "6/ "&$ &$$9$+. $"$>"+ "$ $/+> $!"66+ &6!9+/ "/6/!+> 9//&!+. $! &.+& >"!9>+/ . = ��� �� .96 $/> .!. !/$6&+/ >>.6$+6 /"9!/+/ >>/ !+9 //"96+ 6!!!/+9 9/"/+ ."6 +> /!>6.+9 > #��*(� ������ =������ � ."& !./ !. !!/"9+" .& >6+9 & $!$+> " 9"&+! 6$!> +/ /"99+& .>"" +/ " .&"+& $!!.!+/ " #��*'�� � =������ � !> & " /!//!+" 6" &.+6 "$>.+9 !6> >+6 ""$$9+6 !!""6+> "/>> +6 9&>9&+" .$9/$+$ -#= ��� � !!" /!& "&/ >9$ /+9 / "9+/ .> >6+> "! "9+ //$&"+ .6>>&+" .9""6+" "$/$&+9 !"9.6+& ! !!6 9/ .// 6.9 /+> 69 9$+ /."" +$ 96>/&+9 & ./$+/ 69. 6+$ 6$$$!+6 6>""6+/ "&!96+" . !!> $ > ! $ &"/.+. !">>&+$ & >>+$ .6!>!+$ >>!"6+ $../9+6 >&! +> !$9>9+/ 9/.$ +/ > ���� � !!6 .! &" ! /./!+ ! 66$!+! !$$.&.+ !"/">.+/ !"6"$"+ !>"$9/+> !!&&69+> !!6.$&+ !/ 69"+9 2h� j���� � !."> 6>" /$& &6"&9+! .9!>+/ " $.+& /$>$+& ."& .+$ "& $$+. $!&>&+> !$6 &+" "!$. +6 ! f � � !">& 9& /"9 .$$.9+9 >9>&.+& 66$$ +. >&>$$+" /&///+/ & "9+9 !.$>$+6 >$!"6+! 6."/!+9 #�� �< (� ��� �� ���� ��� � qq#�+ �� ��� �� �����ee ��� �� �� ���� � �� ��� �� ����� �� �� � ��� ��� � � �� � ��� �� ���� � �� � �� �+ (� �� ��� �� ���� ��� �� �� � � � � � ����� �� �� �� �� �� � ��� ��� � � �� � ��� �� ���� � �� � �� � ��� �� � � + / > " '� ' � ) ��� � ' " �� *� �� � � �� � � ���� �� % ����# ( + ,& + + ( . . / 0 1 � � � ����� ���� ���� � � ��������� ��� ".+/i+ (� � ��� � �� ���� � ��� ��� � � �� ����� � 6< � ��� �� �� !+/< ��� � � �+ (� �� �� ����� �� �� ��� ���� � �� � �� ���� ��� � � ���� ����� ��� �� + (� � ������ ��� ."+!i �� �� ���� ���� �� ���� ���� � �� ��e� � �� �� ��� ������ ��� !!+$&i+ h�� � �� � �� ����� � �� � ��� ���� ���� ���� � �� � ��� ���� �� ����� � � �� ���� � ��� � ������� �� � �� ���� � �� ��� ���� � �� ���� � �� � � �����+ #�� �� � ���� ���� �� �� ��� ����� � � ���� �� � � ����� ���� � � � �� ! �� �� � ����� � ���� �� � � .+ (��� ������ � � � )� �� � �� � � � � � ��� � ��+ 7�� �� � � ����� ��� ��� ���� ������ ���� ���� ����� ��� �� �� �� � � ��� ����� � � ����� ���� � � ���� � � � �� ! �� �� � � �� � � �� � � . � ��� �������� � � ��� + d� � � � ����� �� ) ��� � �� � � ��� �� ��� � �� � k �?� � � ./+9/i �� �� � � ����� �� d��� � �� � � � ."+"i �� � � !� �� !&+>i �� � � .+ (��� � �� �� ���� � �� �� ���*������� � � ����� � �� � k �?� �� d��� � �������� � �� �� �� ����� � �� k �� � + 3 &&&4� h��� � � + 3 &&&4� �� , �� �� �� � ���� 3!$$$4+ -� � �� � � �� � ����� 0� ������ ���� � ��� � �� � 0� �� �� � � �� )� �� �� � ��� ��� �� � �� ��+ (� � �� �� � ���� ����� �� ���� � ��� k �? � ��� � ���� � �� � ��� k �? ���� �+ (� ���� �� � � � �� �� ��� � ����� � 2�� �� �� ���� ���� �� ���� � � ���� � � " +9i �� �� d��� � � ����� �� � � � 6!+"i �� � � !� �� �� � ����� �� /!+!i �� � � .+ (� � � ���� � �� � �� � � ����� �� � �� � �� � �� � �����+ -� � � � ���� ��� ����� � �� �� � ����� � ,��� �� �� � � �� �� �� � � ����� �� ���� ��� ����� � #��*'�� � =������ � � � �� �� ��+ (� ,��� �� �� ����� ������� � �� � � �� �� �� � � ����� �� � � � ! �� .� ��� �� �� �� � � ����� �� �� � �� ,��� �� �� -�� � � � ! �� .+ d�� �� � �� �� �� ����� � � ��� �� ���� �� ��� �� � � ���� � �� � �� � ������ �� * =r � � �� � �� � � � � �� �� ���� ������ � ��� �� �� �� ��� ������� �� �� ���� � ����� �� � � � �� � � �� ���� �� ��� � ������������� �� � ��� � � ��� �� � ����� �� �������� �+ (� ���� ������ ��� �� ����� � �� � � � � � ����� � ��� �� ��� � �� + ��� � �� �� �� �� ��� 0� ��� � � ���� ���� � � �� �� � �� ���� �������� � �� � �� ��� ���� � 0� ��� + -� ��� � ������ ���� ���� �� ���� �� ����� � � � �� �� �� � �� �� � 36+9i4 �� ���� �������� � �� ���� ���� �� �� �� ����� � � �� ���� �� � �� �� � 3>9+6i4 �� ���� �+ d� � �� � � ����� �� ���� � � ���� � ��� � �� ���� � �� ���� � � � ""+9.i �� �� � � ����� �� � ��� � �� � � �� ��� �� �� �� ����� ����� �� &!+$"i �� �� ���� �� ����� �����+ -� � � !� ���� � � � ����� ���� � � �� "/+//i �� �� �� �� ����� ����� �� �� � ����� �� " +$/i �� � � .+ k� ����� ��� ���� � �� �� ���� �� ����� ����� ���� � �� &"+ 9i �� �� � � ����� �� � ��� � �� � � ! �� �� � 5��� �� "+ 6i �� � � .+ (��� � �� � � � � b �� �� ��� �� ����� �� �� ����� � � + -� � � � ���� ���� �� �� �� ����� � � !!+ &i �� �� � � ����� �� ���� �� �� ���� �� ����� ������ �������� �� !$+..i �� � � ! �� /+66i �� � � .+ ���� � �� � �� ����� �� � �� � � 5�� � ���� �� �� � � ����� �� � ������ ���� � � � ! �� .+ (� � � �� � ����� �� � ���� �� � � "" � �� �� ����� � � � �� � � � ����� �� ��� � � � �� � ���� �� � � � � � �� � ""+ 7�� ���� �� � � ����� �� � ��� � �������� � � � ��� �� � ���� �� � � ����� �� ���� �� � � ���� �� � � �� ���� � ��� � �� �� ��� ��� � � � ! �� .+ (� � � � ����� � � � ��� � �������� � ���� � ���� � �� �� � ���� �� � � ����� �� ���� �+ (��� ������ �������� �� � �� �� ����� �� a���� �� a��� 3 &&94+ "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 /" k � � �� �� � ���� �� � � ����� �� �� ��������� �� �� )� � ���� ��) � �� � �� ���� �������� � �� � � �� � � � � � ��� � � ����� ��� ���� �� � � �� � d��� � � � � ������ �� � � �� �� �� ,��� �� �� ������ �� �� ���� �� ����� � � �� � � � �� � � � �� ���� �� � � �����+ -� � � .� �� � � �� �� � �� �� � ����� � �� � �� �� ����� � �� � �� � ��� �� � � ������ ���� �� ���� �� � � ����� ��� ���� �� �� #��*'�� � =������ � � ������ ���� � �� �� �� ���� �� ����� � + (� ���� � �� �� ���� ���� ������ � � �� �� � ��� � �� ���� � 3!>4 �� �� #��*'�� � =������ � � ����� � �� � �� �� ���� 3&"4 � �� � + -� � � � ���� ���� �� �� �� � � ����� � ���� � � � � ����� ��� ��� �� ��� � � ���� ����� ��� �� � k �? � � � � ������ �� � � �� �� �� #��*'�� � =������ � ������ �� �� �� ����� � � �� ����� � �� � + 7���� �� �� �� �� �� � � � ����� ���� � � ���� � � � �� ! �� �� � ����� �� � � . ���� �� )� ����� �� �� � ����� � �� � + 3'&' �� ��$�� �� ��� �������� � �$) ����� ��� # ��) (� �����*� ����� � ���� ����� �� ( � . ) ��� � � �� �� ��� �� ��� � ����� �� ��� ��� ������� �� ���� � �� � �� ���*������� � � ����� � �� ) #��� �� �� � � + f� � � �� � �� �� � �� ������ � ������� �� � �� $+$" � �� � �� � �� �� � � + (� � �� �� � �� �� � �� �� ��� �� �� ���*������� � � ����� �� � �� )� �� � �� ���� �� ��� ���� � �� �� ����� � � 3.>i4 �� �� �� ��� � ��� � �������� � �� �� ���� � � 3 .i4+ (� � �� �� ��� ��� ��) � �� � � � ������� �� ���� �� )� ����� �� ���� ��� � ���� �� �� � ����� � �� � �� � � � �� .+ -� � � � ��� �� ��� � �� � �� �� �� �*�� ������ ������ ��� �� ��� ���� � �� � �� �� � �� �� ��� �� ���� � �� �� ������� � �� � � � ! �� .+ (� �� �� �*�� ������ �� � � �� ��� �� � ����� � �� � + (� � �� ���� � �� �� �� � �� �*�� ������� ��� � �� �� ��� �� � �� � �� �� ����� � ���� � ��� � �� ���� �������� �� ���� �� ���� �� �*�� ������ ��� ��� ���� ��� � ��� � �� ����� ��� ���� + d� � �� ��� � ����� �� ��� � ) ��� � ��� � � ��� �� ��� � �� � �� � � ���� ��� � ������ �� � � �� ������ � ������� ��+ 7�� ��� ��� �� �� ���� ��� � ������� �� ������ � �� � �� � + (��� � � � �� �� �� �� ��� � ��� �� ��� �� ���� � �� ��� ���� �� � �� ������ � �� � �� ����� + (���� ��� � � � � �� �� �� � ������� �� � �� �� ��� �� ��� �� � ��� � � �� �� ����� ��� �� ���� � �� + -� � � ! ��� �� ��� � �� � ��� ��� �� �� ���� ��� � ������ � � �� �� ������ � ������� �� ���� ��� ���� � � ���� � � �� ������ � �� � + -� � � .� ��� �� �� ��� � ������ �� �� ��� ��� ���� � � ��� �� ������ � �� � + (� ��� ��� � ������ �� � � ����� �� ����� ��� � � � � ���� � �� � � �� � �� ��� ���� �� � �� � ����� �� � + (��� � �� � ���� ��� �� � � � � ���� �� ��� ���� � � �� � � � �� ���*������� � � ����� � �� ���� ��� � � �� � ����� � �� � ����� � �� � + (� ��� �� ����� ��� � ��� � �� � 0� �� � ���� � �� � � ���� ���� � ����� ���� � � ������ � � �� � ��� ����� �� 0� ����� �� �� � � �� �� ��� � ��� � �� �� � ��� �� ���� ����� �� ���*������� � � �����+ d� � �� � ��� � � �� �� ) ��� � �� � � � �� � � �� ��� � ������ �� � � ������� �� � ������� �� ����� � � �+ 7�� ���� �� � � ���� �� � � � �� � � ���� ��� � ������ �� � � �� ������ � ������� �� �� �� ������ � �� � �� ����� + -� � � !� // "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 ( � . =����*� ����� � ���� �� � � ����� ���� ��� ��� � � � ����� d � d � ! d � . d�� a ��� � ���� d�� a ��� � ���� d�� a ��� � ���� h�# >/+ "��� 9+9$�� 6+>"��� 9>+ &��� !6+&$��� >+ >�� & +!>��� ..+.&��� !!+>!��� ��1 &!+ &��� 6/+. ��� !!+"/��� >9+$.��� !!+&"��� .!+.$��� .$+$9��� $.+"!��� ./+& ��� �2= .!+.&��� /+$ ��� !+69��� ."+/6��� /+9.��� >+9 ��� ./+9/��� &+"$��� "+& ��� ��� .+>.�� .+!$� $+9/ >+6>��� >+ .�� !+/"� >+ .�� .+"9�� +&! -#= 6/+ ��� ./+&$��� !6+&&��� &6+"!��� >/+!6��� >$+!"��� &/+&&��� >>+$/��� 6>+$.��� 2h� >>+!>��� ..+$>��� /+>$��� >/+!&��� !&+//��� !!+&9��� !$+!&��� +.!��� "+&$��� h�# � ��1 >+$$�� .+ .� $+ 6 .+6 � +&! $+6/ .+$ � +&& +!6 h�# � �2= $+$" $+ / $+.> $+/$ $+ 9 $+>> +>6 $+9$ +./ h�# � ��� +$ +.6 +"" $+"" $+"! +!$ $+>! $+>6 +!6 h�# � -#= $+$. $+$! /+&&��� $+ ! $+ / .+ "� $+!! $+!$ >+&"�� h�# � 2h� $+$6 $+$9 $+$! +9! + $ $+ " $+$/ $+!" $+.. ��1 � �2= +$6 $+9" $+>& !+$! +.. + / !+!>� +"$ +" ��1 � ��� +/$ +/" +/> !+ >� !+!!� +!! +>/ +6> $+&$ ��1 � -#= +!. $+>6 .+!/��� +"& $+/& !+/&�� !+9 �� $+&6 "+..��� ��1 � 2h� .+.6� .+!&� $+$/ >+.$�� !+9$� !+$$ +9> +$" !+/&� �2= � ��� $+!$ $+ / $+6> $+"9 $+"! $+9! $+&$ $+6& $+"& �2= � -#= +!9 $+>9 !+.$� $+9. $+6> +&! +!& $+9> !+6/� �2= � 2h� +>/ $+66 !+!. !+ > $+6" .+&&� $+66 $+>. +$/ ��� � -#= $+>" $+. !+96�� $+9" $+9! !+&>�� +/$ $+9! /+/9��� ��� � 2h� !+6/� !+/6� +&" +9/ +6/ +.& +$6 $+/ +6. -#= � 2h� >+6"�� .+99�� !+>! .+9"�� !+&.� .+"$� !+6$� +>9 /+> ��� %! $+ > $+ . $+!. $+ 9 $+ 6 $+!6 $+ 6 $+ " $+.> a�� � 9+>9��� "+/6��� >+ .��� +6/��� 6+9/��� >+&&��� $+&"��� /+99��� 6+!$��� a � >>/.+/$ . 9""+.$ />"".+.> ...9>+ $ ".>/6+"$ 6"6>&+>6 ..&>+/$ .$99$+/$ /.! .+"6 #�� �< (� ���� �� � �*�� ������� ��� � ����� �� �� � � �� �� ��� �� ��� � ����� �� ��� )� �� ���� � �� � �� �� ����� �� � � �����+ (� � � �� �� � � � � � ����� �� �� ��+ (� �� ���?� 3���4� 3��4� �� 3�4 ����� � �� ������� �� � $+ i� i� �� "i � � � �� ���� �+ " '� ' � ) ��� � ' " �� *� �� � � �� � � ���� �� % ����# ( + ,& + + ( . . / 0 1 / 6 �� � � �� ��� � ������ �� � � ������� ��+ d�� ����� �� � �� �� ���� ������ � �� � � �� ������ �� �*�� ������ ������ ��� �� ��� � ����� �� � � �� � + -� � � .� �� � � ��) ��� � ������ �� � � ������� �� �� ����� �� �� ������ � �� � ��� �� �� �� � ��� � ������+ (��� ���� ��� �� � ���� �������� � � ������� �� � ���� � �� �� �� � �� ���*������� � � ����� �� � ���� ��� � � ��� �+ (� ���� ��c �� �����*� ����� � ����� �� � �� ���� � �� � �� � �� ��� � �� � �� � ��� � � �� � �� �� ���� � �� ��� ���� � �� �� ��� �� ��� � ������+ d� � �� ��� � �� �� �� � ��� � � �� � � �c � �� � �� �� ���� ��� ���� �� �� ���� ������ �� )� � ���� � �� � �� ���*������� � � �����+ 7�� �� ���� � �� � �� ���� ������ �� )� � ���� � �� � �� ����� + d� � ��� � ������ � � �c � ��� �� ��� � �� � �� ���� ������ �� )� � ���� � �� � ��� ���*������� � � ����� �� � �� � � �� ��� ���� �� � � � ! �� .+ 7�� �� � ��� � � �� � �� � �� �� ������� �� )� � ���� � �� � + 7�� �� ���� � �� � �� ���� ������ �� )� � ���� � �� � �� ����� + -� ��� � �� � �� � ����� �� �� ������� ���� �� )� ����� �� � �� ���*������� � � ������ �� �� � �� ��� ���� �� �� � � + 7���� �� � ��� � ��� � ������ � ���� � � �� ���*������� � � �����+ (� �� ������ ������� �� ��� ���� �� � �� �� ��� �� �� ��� � ������ �� �� �� ���� ����� �� � �)��� �� � �� ���� � �� �� ���*������� � � ����� �� ���� �� �� �� �� � � �� � ��� �� ���� �� � ����� � �� � ���� ���� �� � �� �������+ 7���� �� �� �������� ���� ����� �� � �)��� �� � �� ���� � �� � �� �� � �� � ���� � � ����c � � � ��� � �� ���� ���� �� ��+ 3'1' ���������� �������� (� ��������� � ���� � ?� �� ���� � �� ���� �� �� � �$) ����� ��� # ��) 7��#��� #���� ��� � �� �� ��� � ������ �� ��+ -� ( � >� �� ��� ���� ������ � �� ���� � �� � � �� � � � � ! �� � � � � . �� !� �� � � � . �� + (� ��� �� �� ������ � ������� �� ���� �� )� ����� �� �� ���� � �� � �� � � � � . �� ! ��� ���� �� �� � ��� � � � ����+ ��� � �� �� � �� �� � ���� � ? � �� � �� � �� �� ��� �� �����*� ����� � ����+ #�� �� � �� � � � �� �� �� ���*������� � � ����� �� ������� ���� � � � �� !� ��� �� � � ��� ���� � � � ! �� . �� � � �� .+ (��� ��� � �� � �� ����� � �� �� �� �� ���� ��� � � ������ �� �� ����� �� �� � ���� �� �� ����� � � �� ��+ (� � �� �� ��� ��� ���� � �� �� � � ����� � �� � � � � ! �� �� � � ������� �� ��� �� ��� � �� � � �� �� ��� ����� � � � �� ����� � ���� �� ������ �� �*�� ������ ��� ��� ����+ -� ���� � �� � � �� � � � � . �� !� ��� ���� �� � � ������� �� ���� � � ��� �� ������ � � �� ��+ k �� � � � � . �� � � �� �� ��� ����� � ������ �� � � ������� ��� ���� � �� � � ��� �� ������ ��+ (� �� � ������ � �� � �� ��� ���� ������ �� ��� ����+ -� �� � � . ���� ����� ���� � � � ! �� � � �� ������� ��+ d� � �� � ��� � � � ���� �� � �c �� � �� � �� ��� ���� � ������� �� �� �� ���� ����� � �� � � � � ! �� + 7�� �� ���� � �� � �� � � � � . �� !� ��� ���� �� � � ������� �� �� ��� �� ���� ����� � �� � � � � . �� � � �� �� ��� ����� � ������ �� � � ������� ��+ ���� ���� �� �� ������ � �� � ��� ��� ���� ������+ (� ���� ������ ��� ���� � ���� ��� ���� �� ����� ������� �� ��� �� ���� � �� � �� � � � � ! �� �� �� � ����� ������� �� ��� �� ��� � ��� ���� ������+ -���� �� /9 "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 ( � > ���������� � ���� �� � � ����� � �� � ��� ��� � � � ����� d � ! � d � d � . � d � ! d � . � d � d�� a ��� � ���� d�� a ��� � ���� d�� a ��� � ���� h�# >+$9��� !+9.��� $+ $ $+$$ $+!> $+!/ +" ��� .+>/��� $+$. ��1 9+6&��� "+9$��� !+9/� >+!/��� >+.>�� + ! /+$>��� >+ �� !+>> �2= .+! � +&" +.! $+>& $+! $+6/ $+&. + $ $+$6 ��� $+&" $+"9 +!! +6> +&" $+>$ !+&"� !+9.� $+>$ -#= >+ 9�� +/& .+./� $+.9 $+!9 .+/.� +99 $+/! $+ $��� 2h� $+ 9 $+ $ +"" !+6!��� &+&>�� .+!9 9+/!�� $+> �� $+!> h�# � ��1 $+9 +!> !+>& +66 $+>9 !+!/ .+ $� !+//� + $ h�# � �2= +! $+&/ $+>! $+/> $+>. !+.& !+/! !+ . $+// h�# � ��� $+6" +$" $+9& $+>> $+." $+"9 $+6/ $+"$ +/" h�# � -#= $+$9 $+!/ +/$ $+$. $+$> +9! $+ " $+!. +"" h�# � 2h� !+/9 >+$6� $+>9 "+!!� >+&/� $+$. $+>! $+$9 $+/$ ��1 � �2= !+&$�� !+.6� +$$ $+!/ $+>/ $+ . !+$/ +.& + 6 ��1 � ��� +/. !+!>� $+&& $+99 $+9 $+66 +$$ $+&$ + $ ��1 � -#= +!> $+9> + ! +. + & +6 + / $+"& .+>"��� ��1 � 2h� +/& +$$ !+9>� +.$ + $+&& $+9/ +66 !+/9� �2= � ��� $+"$ $+&" $+6& $+.> $+>! +/" $+9> $+& +!! �2= � -#= +!& +>6 +.$ $+!> $+ / $+9 $+// $+6 $+&! �2= � 2h� +.6 $+&> +$6 $+9$ $+ ! !+"9 $+/9 $+66 $+>. ��� � -#= !+>>�� +9.� >+66��� !+&/��� !+$"� >+!9��� !+9$��� +"> "+ "��� ��� � 2h� $+"" $+!& +./ +$/ + ! $+$" +9& +"9 +!$ -#= � 2h� $+.> $+"! $+>6 +!9 $+9& $+.6 !+!$ !+$! +$> %! $+$> $+$" $+ $+$. $+$. $+ $ $+$> $+$> $+ . a�� � !+!!��� +&9��� +/9��� +./� + > +" �� +99��� +">��� !+ ��� a � 9&!$+" ! / !+ " &/+ . � &&9&+"> �!!"9/+9> � !"."+&$ � $/&+$. �&6>+/& � ..&+66 #�� �< (� ���� �� � �*�� ������� ��� � ����� �� �� � � �� �� ��� �� ��� � ����� �� ��� )� �� �� ���� � �� � �� �� �� � �� � � � �� � �� � ��� � � �+ (� � � �� �� � � � �� �� �� � � ����� �� �� �� � �� � ��� � � �+ (� �� ���?� 3���4� 3��4� �� 3�4 ����� � �� ������� �� � $+ i� i� �� "i � � � �� ���� �+ " '� ' � ) ��� � ' " �� *� �� � � �� � � ���� �� % ����# ( + ,& + + ( . . / 0 1 / & � � � �� )� � ���� � �� � ��� ���*������� � � ����� ��� ���� �� ��� ��� ���� �� ��� �� ��� � �� �� ���� �� � � � � ����c � � � ��� �+ (� �� � ������ ��� � ����� ��� �� ��� � �� �� � �� � � ����� �� ����� + 7�� �� � ��� � � � ���� �� � �� �� ������ ��� � ����� � �� � � � �� ��� ��� ���� � �� � �� ������ ��� � ����� �� � ����� �� ����� + (���� ��� �� ��������� ���� ������ �� ��� � ����� � �� �� �� � �� � ���� � �� �� ���� � � ���� ���� �� ������ � �� ��� �� � ����� �� ����� + (� �� ������ � �� �� ��� �� ��������� � ���� �������� �� ���� ����� �� � �� � �� ���� � �� �� �� ���*������� � � ����� �� ���� �� �� � �� � � � � �� � ��� �� ���� �� � ����� �� � ���� ���� �� � �� �������+ 7�� �� ��� � �� �� ���� �� � � � ��� �� ������ ������� �� � �� � �� ��� ���� �� ��� ���� �� � � ���� �� ������ � �� � �� ����� + k��� �� �����*� ����� �� ��������� � ���� ����� � ������� ��� �� ���� ����� �� � �� � �� ���� � �� � �� � ���� �� � � � ��� � �� ���� + 3'3' 8��7�� ������� ) ���) ��� ( � " ����� �� ���� � �� � ��� � ���� �� �� � �� � � � � �� .+ -� �� ����� � � � ���� ��� � ��� � ��� � ������� � ���� ��� � �� ��� �� � > �� � ���� �� � � � � ��� � ��� � �������� � � ��� "&+&+ � � �� � �� � �� �� � � �� �� ���� � " 8/ �� && + k� �� ����� � � � �� � � � ���� � � �� "/8//� ���� ���� ��� � ��� � ��� � ������� � ���� ��� � �� ����)�� � � . �� ! ���� �� � � � � ��� � ��� � �������� � � ��� /!+6 � ��+ 3'.' �)����� �� ����� ��� ��� � (� ������ �� �� � � � � ���� � �� � � �������� � � � � ��� � �� �� � �� � ��� �� �� �� � �� ���� �� 0���� � �� � ��� �� ���? �� ����� � �� �� �� �� � ��� � �� �� ��+ ( � / ����� ��� �� ����� �� �� ��) ��� � �� �� � �� � � �� � �� �� � �� �� ��� �� � � � �� ��� ���*������� � � ����� ��� � � � �� .+ (� �� ���� � �? � � ��� ����?� �� ���� ���� �� �� � �� ���� ��� �� ������ ��� ��� �� �� � �����+ ( � " (� ���� � �� � ��� � ���� �� �� � �� � � � � �� . d � d � . � ����� � 2� � ����� � 2� � ��� � 6!. 3 6+ i4 "&+& 9"" 3> +9i4 /!+6 #�� � ��� � � ." 6 39!+&i4 ">+& !"6& 3"9+!i4 "9+$ (�� >!>$ 3 $$i4 >>.> 3 $$i4 #�� �< (� ���� �� ��� � qq� ����� �ee � �� ���� � �� � ����� �� �� �� ���� �� �� �� � � ��� � � � �� � �� �� � �� � ����� � �� ��� � ����� �+ (� ���� �� ��� � qq2� ee � � � � � � �� �� �� ��� � ����� ���+ (� ����� �� qq� ��� �ee ��� �� � �� ���� �� �� ���� �� ��� � ����� �� ��� � � ��� � ���� � � �� � �� � � ��� � � �� ��� �� ���� �+ (� ����� �� qq#�� � ��� � � ee ��� �� � �� ���� �� �� ���� �� ��� � ����� �� ��� � � �� � � ��� � ��� � � � �� ��� �� ���� �+ (� ���� � �� ��� � ����� � �� ���� � �� � �� � � � � �� . � � �� �� � ����� � � �� ���� � �� ��� � � � �� .+ 6$ "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 -� �� �� �� � �� ��� �� �� ���� � � � �� �� � �� �� ��� ���*������� � � ����� ��� ������ �� �� �� ��) � �� � �+ (� ���� � �� �� � � �� ���� � � �� �� � ��� ����?� ��� � �� ���� ���� �� ���� ����� ��� �� � ���� ���� ��� �� � �� ���� ���� �� � � �� + d��� � �� ��� �� ���� � ���� � �� �� � � �� � ������ ������ )� �� ���� �� �� ,��� �� �� -�����+ (� � �� � �� ����? �� ����� �� �� �� � �� �� � ���� � � �� �� d��� �� ��� �� ���� ����� ������� �� ������� �� k �� � + 3 &&&4� h��� � � + 3 &&&4� �� , �� �� �� � ���� 3!$$$4� ��� ���� �� �� � 2���� � 2� ��� �� � � � ��� � � � ���?� �� ��� �� �� ���� � ����?� �� ����� ���� � � �� = �� �� ��+ 2 ����� ������ )� �� ���� �� �� �� �� ����� ����� ���� � �� ���� �� ���� � ������ ���� � �� �� �� ���� �� ����?�+ d �� �� ��� �� 3!$$$4 �� k��� �� =� � 3 &&64 ��� �� � � � � �������� � �� � �� �� � �� 0���� �� ��� �� �� � � � �� � ��� ����� �� ��� �� �� � � ���� � � � ����� �������� �+ ( � / =� �� � �� ����������� �� �� ��� �� � � �� � � � � �� . l �� � � � � ��������� 7�� ��� 2�� � = �� ����? k��� d � d � . d � d � . d � d � . d � d � . h�# a ./+&. .9+>6 6/+". 6.+. ! +!! !>+$> !+!" !+/" ! 7 � ..+> >.+$6 9.+/$ 9.+&. >+6. >+.> +/6 +6. ��1 � �� ��� � � .>+&" > + & &!+99 &!+> /+6& /+&/ $+.. $+/. ! ."+!> >$+/6 9!+&/ 9!+>! "+// "+". +.9 !+$" . .>+/! .9+$ 6/+>/ 6>+>. ! +6. !.+/6 +9 +& > a�� ��� � � ./+.. >!+>/ /9+./ /"+ ! !6+.9 .$+&> >+!/ .+&> �2= d��� ./+$6 .&+6/ 69+.& 6/+." &+" ! +!/ !+ $ !+.& ! k �? .>+.& >"+9/ 9&+$/ &!+!6 $+$ /+6$ $+&! +$. . f�� �� /+/ >6+$ 9.+>& 9.+/6 .+9/ >+6/ !+// +"6 ��� ,��� �� �� .>+6! >$+6/ 6>+/" 6$+9" !!+/. !"+9. !+6! .+.! ! ,��� �� �� -.!+/> .&+ 9.+&" 9.+&! >+6& >+>9 +!/ +/$ . = ��� �� .&+!/ >!+6 6/+9" 6"+/" !$+"6 !!+ 6 !+"9 !+ 9 > #��*(� ������ =������ � .6+ > .9+"6 9 + " 69+ 6+/6 &+>$ + 9 !+>& " #��*'�� � =������ � .9+$$ /$+!/ 9$+.$ 6 +/! .+$$ !>+>. /+6$ .+&" -#= ��� � .&+&> >/+./ 9&+!. & +> &+&9 6+99 $+6& $+6 ! ..+>> >$+& 9>+!" 9.+ 6 >+>" >+&. +.$ +&$ . ..+6" .9+/ 69+/& 6/+/$ &+"> !$+9! +66 !+"9 > ���� � ."+ .6+6" /&+/" /"+.! !/+/. . +.. .+6 .+." 2h� j���� � ..+&$ .9+&> 9$+/. 6&+!. 6+"& 9+&9 +69 +6& ! f � � ./+"& >!+ 9 6&+$$ 66+.> 9+9> !$+$. !+ 6 !+/. #�� �< (� ��� �� ���� ��� � qq7�� ��� �� �ee ��� �� �� � �� �� � �� �� ��� �� � � � �� ��� ���* ������� 0���� ��� � � � �� .+ (� ��� �� ���� ��� � qq= ��ee ��� �� �� � �� �� � �� �0��� �� � � � �� �� ��� �� � � � ��� � � � �� .+ (� ��� �� ���� ��� � qq����?ee ��� �� �� � �� �� � �� ����? �� � � � �� �� ��� �� � � � ��� � � � �� .+ (� ��� �� ���� ��� � qqk���ee ��� �� �� � �� �� � �� ���� �� � � � �� �� ��� �� � � � ��� � � � �� .+ "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 6 (� � �� � ���� � �� ����? �� ����� � �� ��� �� � �� �� )� �� � �� � �� � �� � �� � + 2 ������ )� � ���� ��� � � �� ������ �� ����? ���� � ������ �� � ����+ ��� � �� �� ���� � � � �� � � �� � �� � �� � �� � �� ���*������� � � ����� ���� � � � �� . �� ���� � ? � �� ���� � �� �� ��� �� �� �� ����?� � � ��� ��cc ���+ 7���� � ���� ��� � ���� �� ����� �� �� � �� ���� � �� ����?� ���� � ������ �� � &&$�+ ,����� )� � ����� � � ��� �� �� � �� � ����� ��e� �� ��� � � � ������� �� � ��� �� �� � �� �� � ��� �� �� � � � ���� � � �� ���+ � ����������� (� ������ �� ���� ����� �� �� ������ � ) ��� ���� �� ���*������� � � ����� ��� �� ���� ��� � � � �� �� ��� �� � ���� ���� �� � �� ������� �� � ��� �� ���+ (� �� ������ � �� �� ����� � �� � �� � �� ���� � �� �� �� ���*������� � � ����� �� ���� �� �� �� �� �� �� ��� � � � �� � ��� �� ���� �� � ����� � �� � ���� ���� �� � �� �������+ 7���� �� �� � �� ���� � �� �� �� ���*������� � � ����� �� ���� �� �� � �� � � � � �� � ��� �� �� �� ����� � �� � ���� ���� �� � �� �������+ � �� �� �� � �� �� ������ � ������� �� ���� � �� � �� � ���� � � ����c � � � ��� � �� ���� ���� �� ��+ ���������� k ��� j+ 2+� � �� �� �+ #+� s � ����� �+ ,+ 3 &&&4+ 2� ������ � ���� �� ���� � �� � �� k �? �� d��� �� � �� � �� ��� ������ �����+ ����� ��� ����� �� %����#� 93.4� !&8 >6+ k 5� ����� l+ �+� s l �� �� �� '+ 2+ 3 &&64+ ���? � ����� �� � ��� � �� ����� � 0� ��+ -� f+ �+ a���� 3��+4� : ���� ���� ����� �� � �)� ;��� &+++ 3��+ >"8//4+ ,�� � ��� < 1��� ����� �� , ���� � �� ,� ��+ k 5� ����� l+ �+� k �� � ?� 2+� s '� � ?�� ��� #+ 2+ 3 &&&4+ h �� � ���� � �� � �� � �� � ������������ � ����� � �������+ ����� ��� ����� �� %����#� 93 4� 8 $+ k��� � o+� s =� � � �+ k+ 3 &&64+ , ���� ��� �����< 2���� � �� ���� �� ��� �� + ����� ��� �������� � ������ .1� .8!.+ ���� �� h+ '+� a���� � f+ �+� s a��� �� '+ #+ 3 &9>4+ 2 �� � ���? ��� � ����� � ��� � �� � ����� �� �+ � ����� � ������ � ����� (93 4� !!8>/+ � � �� �+ d+� s ���� ��� ,+ a+ 3 &&&4+ �� � � ��� � �� � ����� �� �� �� � �� � ��� � �� � ���� �����g ����� ��� ����� �� %����#� 93!4� 68 !6+ � � �� �+ d+� s ���� ��� ,+ a+ 3!$$$4+ k � �� �� ������< �)� �� ����� �� � ����� � ������� �� � � � 2� ��� ��+ ����� ��� ����� �� %����#� <3 4� ..8>"+ 7� � �� �+ ,+� ' ������� d+ d+� s ����� �+ �+ 3!$$$4+ ��� � ��� �� � �� ���� ��� �������� ��� �����< (� ��� �� �� �� �� � �� � � �� ���� � ������ � � ���+ ����� ��� ����� �� %����#� <3>4� !&"8.!/+ h� ���� �� �+ 2+� � ������ l+ �+� ����� �+ 2+� s ' �0� �� �+ 3!$$$4+ � ��� � �� � ����� ���� �� �< 2 � ��� ����� �� �� ��� � �� �� �� ��� ����+ ����� ��� ����� �� %����#� <3.4� !. 8!>6+ h��� �� a+ �+� 7�)� '+ '+� s a��� ��� =+ ,+ 3 &&&4+ � �� ���� � �� � �� ��� ���� � ������ � ?���+ ����� ��� ����� �� %����#� 93.4� >&8 /!+ � ��� � �� 2+ �+� =� �� �� �+ �+� s ���� �� �+ �+ 3!$$$4+ ���? �� � �� �� �� � �� ���� ��� ��� ����� � ���� � ��� � ��+ ����� ��� ����� �� %����#� <3!4� "&8 6$+ '��� �� 7+ (+� s ��c� �� �+ 3 &&"4+ 2� �� ��� � �� �� � �� �� � ��� � �� �����+ =)� � ����� ����� %�� �� ��� 1++ 6! "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 � ����? � 2+ k+� �� ��*a�= � �� a+� s ���� �� 2+ �+ 3 &&64+ 7 �� � �� �� � �� �� 1+�+< � � �� ��� �� ���� �� ���� � �� =����� � 7�� �� �+ ������� %������ ��������� 913 4� 8!>+ a���� � f+ �+� s a��� � '+ 7+ 3 &&94+ = � 2� ��� �� ����� �� � ��� g # � ��� �� �� � ��� � �� � ���� � 0� ��+ =)� � ����� %��� ��� >������ �� ?.3.4� .6 8>$$+ a��� ��� =+ ,+� j��� j+� s � �� � �+ 3!$$$4+ � � ���� ��� �� � �� � � ��� � �� � + ����� ��� ����� �� %����#� <3 4� 8 "+ , � �� k+ 2+ 3 &9&4+ ( ) � ���� �� � ��� � �� ����� � � � � �� � ����+ =)� � ����� %��� ��� >������ �� .?� 6$!86!"+ , � �� k+ 2+ 3 &&>4+ � ��� � �� ����� � � � � �� � ����< 2� ��� � + ����@�� a��������� (+3!4� "&86"+ , ��� �+ 2+� s �� � ����� (+ �+ 3!$$$4+ 7�� ��� � ���� � �� �� 2���� � 2� ��� � � �? �< d� � � �� �� ��� � �� � ���� � ?���+ ����� ��� ����� �� %����#� <3>4� .>.8."&+ � ��� ��� ��� d+ 3 &&94+ = �� ���� � �� �e� �� � ��)+ ����� ��� ����� �� %����#� 03.4� &"8!$/+ ������ �+� s (� ��� �� �+ 3 &&&4+ (� � � � ���� ���� �� �� 1��� � �� � �+ ������� : ������ � %�� ���� 8/+ ���� � ������ 2�������� ���� 3!$$ 4+ ; �� � ��� �� ����� ���������� 8>+ l�� � ,+ ,+� s a�h������ '+ �+ 3!$$$4+ (� �� � �� ���� � ������ �� � ��� � ��< � ����� ���� �� �*���� � � ����+ ����� ��� ����� �� %����#� <3 4� >68&.+ d �� � �+� s ��� ��� k+ 3!$$$4+ 2�� � �� ���� � ������� �� � ��� � �� �������< 2� *��*������� �����������g ����� ��� ����� �� %����#� <3 4� 6&8&!+ d� �� �� � d+ 3!$$$4+ (� ��� �� �� �� � ����� ���� �� �� � ������ �� ���? �� ��� � �+ ����� ��� ����� �� %����#� <3 4� 68. + j��� j+� � �� � �+� s a��� ��� =+ ,+ 3 &&94+ a � �� � ��������� ���5 ������ �� � ��� � �� � 0� ��+ ����� ��� ����� �� %����#� 03.4� 6"8 &.+ "'�' �)��� �' "�� * ����� ��� ����� �� %����# (+ ,&++(../01 6. longitudinal changes in net worth by household income and demographic characteristics for the first three waves of the hrs introduction review of the literature retirement planning retirement adequacy asset allocation gender differences age of retirement data and methodology hrs data model used in analysis of the data description of explanatory variables results descriptive statistics of net worth cross-sectional analysis of non-housing net worth longitudinal analysis number of retired households changes in asset allocation conclusions references financial services review volume 33 number 1 (2025) volume 33, no. 1 2025 editor: john e. grable, ph.d. cfp ® university of georgia advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting conrad ciccotello, jd, ph.d., university of denver sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp ® , baylor university (emeritus) martin seay, ph.d., cfp ® , kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp ® , san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama lu fan, ph.d., cfp ® , university of georgia jasmine fang, ph.d., massey university, new zealand mark fedenia, ph.d., university of wisconsin philip gibson, ph.d., cfp ® , winthrop university stu heckman, ph.d., cfp ® , texas tech university william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea izidin el kalak, cardiff university, united kingdom thomas langdon, ph.d., roger william university, bristol, ri mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden lance palmer, ph.d., cfp ® , cpa ® , university of georgia abed rabbani, ph.d., cfp ® , university of missouri chris robinson, ph.d., york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida issn online 1057-0810 print 1873-5673 contents grable, john e., from the editor, i-ii. congrong, ouyang, crandall, thomas, & chatterjee, swarn. the impact of the covid-19 income shock on debt management: a mediation analysis. 1-27. liu, zhikun, blanchett, david, sun, qi, & fink, naomi. retirement expectations vs. reality: if covid-19 did not impact retirement expectations significantly, what did? 28-49. koochel, emily, mccoy, megan, & lutter, sonya. protecting wellbeing through financial shocks. 50-66. mccoy, megan, machiz, ives, johnson, portia, white, kenneth, watkins, kimberly, & bennetts, chet. resilient personality or financial resilience framework for coping with physical and mental health during the covid-19 pandemic. 67-85. sommer, matthew, mccoy, megan, & lim, hanna. an investigation of the relationship between gender and investor behavior during a market correction. 86-101. rand, christopher, mccrae, melisande, & martin, jason. retail investors and investment fraud victims: is there a connection? 102-119. zhang, yu, naveed, khurram, & qi, jia. crypto investment: the role of investment motivations, investment confidence, and risk perceptions. 120-141. malladi, rama, & stanoyevitch, alexander. assessing the impact of rebalancing on equal-weighted and value-weighted portfolios over five decades. 142-164. kwak, eun jin, & grable, john e. a domain specific measure of investment risk preference, 165-178. lynn, christina, heckman, stuart, kothakota, michael, & lawson, derek. does overspending harm retirement preparation? 179204. academy of financial services officers president shawn brayman smb research consulting executive vice president program michelle cull western sydney university vice president finance thanh ngo east carolina university vice president communications kirsten macdonald griffith university vice president international relations jasmine fang massey university vice president marketing & pr thomas korankye university of arizona vice president membership matt goren danko education immediate past president tom potts baylor university editor, financial services review john e. grable, ph.d., cfp® university of georgia directors jason anderson university of kansas norah feng massey university wookjae heo purdue university thomas korankye the university of arizona barry mulholland university of akron mustafa nourallah mid sweden university richard stebbins university of alabama yu (yulia) chang kansas state university past presidents inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 33, no. 1, 2025 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • conrad ciccotello, jd, ph.d., university of denver • shinae choi, ph.d., university of alabama • lu fan, ph.d., cfp®, university of georgia • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • philip gibson, ph.d., cfp®, winthrop university • stu heckman, ph.d., cfp®, texas tech university • stephen horan, ph.d., cfa®, university of north carolina wilmington • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • izidin el kalak, cardiff university, united kingdom • thomas langdon, ph.d., roger william university, bristol, ri • mustafa nourallah, centre for research on economic relations, mid sweden university, sweden • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., cfpretired™, cpa,ca, york university (emeritus), canada jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • axton betz-hamilton, ph.d., south dakota state university • alona bilokha, ph.d., university of north florida • brian l. boscaljon, ph.d., penn state behrend • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph..d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph..d. stetson university, deland, fl • patti fisher, ph.d., virginia tech • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • david nanigian, ph.d., cfp®, mount ararat financial services llc • john nofsinger, ph.d., university of alaska anchorage • mustafa nourallah, ph.d., centre for research on economic relations • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • congrong ouyang, ph.d., kansas state university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peerreviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ pii: s1057-0810(00)00048-2 the internet in the personal finance course walt woerheide* department of finance, rochester institute of technology, 108 lomb memorial drive, rochester, ny 14623-5608, usa abstract a revolution is occurring as textbook authors struggle with how to effectively incorporate the internet into the personal finance textbooks and course. this paper considers three questions associated with this revolution. what is the current degree of internet integration? where might we be headed with this new technology? will this integration make for a better course? the current degree of internet integration is reported in tables 1 and 2. an argument is made that we may be headed to an eventual elimination of the textbook, as we know it. there is some preliminary evidence that this will make for a better educational experience. © 1999 elsevier science inc. all rights reserved. jel classification:i20, o30, z00 keywords: internet; pedagogy; textbooks; technology 1. introduction there is a significant event that is happening in the field of personal finance, and no one seems to be making an effort to assess that event and to explore the implications of that event. this event is the incorporation of the internet into the personal finance course. in just a period of a few years, the textbooks in this field have gone from not even mentioning the internet, to various degrees of actively incorporating the internet. this raises some interesting questions. what is the current degree of internet integration? where might we be headed * tel.: 11-716-475-5268; fax:11-716-475-6920. e-mail address:wjwbbu@rit.edu (w. woerheide). financial services review 8 (1999) 305–317 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00048-2 with this new technology? will this integration make for a better course in terms of student comprehension and skill acquisition? these questions should be of great interest to faculty who teach in the area of personal finance. they should also be of interest to practitioners in the personal finance area. one question facing all faculty as technology alternatives continue to appear is which ones are of passing interest and which ones will grow in significance. faculty who want to avoid having their courses, their teaching, and their research skills become outdated have to decide in which technologies to invest their time and efforts. this paper should help anyone who is undecided about the role of the internet as to whether learning more about the internet is a worthwhile professional commitment. practitioners should also be interested in these questions as the answers may impact how they relate to future customers and the types of skills that future employees will bring to the job. the basic premise of this paper is that the internet is a significant incorporation of technology into the personal finance course. the first section of this paper provides some informal history of the growth of technological integration into finance courses. next, a proposal as to the ideal integration of the internet into the personal finance course will be presented. we will then consider why that ideal is not currently being achieved. we next turn to an empirical examination of how the internet is currently being incorporated into our textbooks in this field. the last section will consider whether there are any substantive benefits to the students from this integration. 2. a review of the technological integration in the 1960s there was little if any technological integration into finance textbooks. the high-tech method of problem solving was the slide rule. as calculators came into popular usage, some textbooks integrated this technology by providing instructions on what data to enter and which keys to punch to solve particular problems. for some textbooks today, this is still the extent of the incorporation of technology. the next technological development was to provide problem-solving disks in conjunction with a text. some disks had proprietary software, and some worked off an existing spreadsheet application such as lotus 1-2-3. for some textbooks today, this is still the extent of the incorporation of technology. in the late 1980s, email began to grow in popularity and usage. although email was not incorporated into textbooks, it became one of the primary methods of communication among faculty. in the late ’80’s, many professors were proclaiming that 1) they weren’t computer oriented enough to learn it, 2) their institutions were unable to provide it, and/or 3) there would not be enough email being sent to justify a commitment to learn to use it. by the mid ’90’s, a revolution had occurred in that so many faculty had email accounts and communicated with email that the professional associations such as the fma and afs found it worthwhile to include email addresses in their membership books along with all the other contact information. as we approach the end of the decade, some of these associations are beginning to make email the primary form of communication with their memberships. the point here is that within a short period of time, faculty have gone from where only a few used 306 w. woerheide / financial services review 8 (1999) 305–317 email to where most now use it. this change required only that the hardware and software become available and that people started to realize the potential of the medium. as recently as five years ago, there was little or no attempt at integrating the internet into finance courses. i offer two reasons for this. five years ago, most finance faculty did not have access to the internet nor had little or no idea as to what was available on the internet. at this time, most faculty were just coming into active daily usage of email. the internet was the logical extension of email. now, all finance faculty are aware of the internet, most are likely familiar with at least a few sites on the internet, and some are attempting to ascertain how the internet may be effectively integrated into their courses. the personal finance course is the logical place for the initial efforts at integrating the internet. this is because a great quantity of the material on the internet is aimed at the individual. much of the material on the internet is aimed at selling products to the individual, but this is not what we mean by personal finance. a good portion of the material on the internet is aimed at providing information and even solutions to problems and questions that people have in the area of personal finance. some of this information is provided in the hope that the users will be so pleased with the information provided that said users will buy the services or products of the site sponsors. however, some of the information is provided by entities whose goal is simply to be a provider of information. these sites include government sites (.gov), educational institution sites (.edu), and nonprofit and not-for-profit sites (.org). 3. what is the ideal internet-integrated course? the wealth of information and calculators currently available on the internet is so extensive that one could teach a highly effective course in personal finance without resorting to a primary text. the ideal integrated course starts with a home page for the course. on or before the first day of class, the instructor needs only to provide the students with the url of the course’s home page. the first link on the course home page would be the syllabus. as a web page, the syllabus can contain many links of potential use to a student. for example, instead of just listing the prerequisites for a course, the instructor can make the prerequisites into links that tie into the syllabi of the prerequisite courses. thus, if a student has had a similar course elsewhere, that student can self-validate that other course by seeing if the course he or she had matches the syllabus of the prerequisite course. most publishers now have web sites to support each textbook (much more on this topic later). as these web sites sometimes include material found in a study guide, as well as supplemental material for the textbook, a link to the publisher’s site for the selected text would seem to be mandatory. the instructor’s name on the syllabus could also serve as a link to the instructor’s resume. as many faculty members have not yet posted their own resumes, some business colleges provide a college-wide directory of faculty with biographies for most or all faculty members. this way, instructors don’t have to feel any embarrassment about providing an autobiography if asked by the students about their background and experiences. each class meeting date could be listed as a link to a separate page. on this separate page, 307w. woerheide / financial services review 8 (1999) 305–317 the instructor could list basic information such as what homework would be due for that class, what the learning objectives for that class would be, and even a full outline of the topics to be covered in that class. the final addition to the class home page would be a link for each “chapter” in an on-line, virtual textbook. the chapter would then be a series of links covering all of the appropriate material for that topic. some connecting material between urls would likely be appropriate. similarly, some end of chapter material such as a summary, list of key terms, and problems may have to be inserted. 4. why doesn’t the ideal exist? if the ideal is within reach, then why don’t we observe it? the primary reason we do not observe anyone providing such a course boils down to the economics of doing it. first, the preparation of such a course would be a gargantuan effort, perhaps equivalent to writing a book. no one writes his or her own book for just one course. a second reason is the obsolescence factor. the author of this on-line text would have to spend time every month, if not every week, checking to make sure all of the links are active and replacing those links where the urls have been modified or the links have disappeared. the modification of urls and their disappearance is known as link rot. no one can afford this much time to keep one course up-to-date. yet another reason we have not seen the abandonment of a primary textbook is that a well-written text can help motivate students and serves as an alter ego to the professor in the course. it is like a secondary instructor in the course in that it reinforces what the professor is saying, and sometimes gives the professor someone to play off of. if the ideal is not possible due to economic incentives, then let us consider situations that might be just short of the ideal. for example, one person could put together this course and “publish” it for all to use. the word publish is in quotes because in the internet world, publishing only means providing one’s url to others. in fact, because of the ability of the commercial search engines to find and document anything that are uploaded on the internet, publishing almost means simply uploading a web page to the internet. the answer to the question of why we don’t observe this nearly ideal text is, again, the economics of the situation. with free distribution of a textbook, there is little incentive for anyone to undertake an effort equivalent to writing a full-length text. this last drawback can be overcome in a couple of different ways. one is to make the on-line text a pay site. that pay sites can be successful has been demonstrated by the proliferation of adult sites on the internet. (of course, this observation about adult sites is based on what the author has read and heard and not personal observation.) this approach has several problems. one is students could easily share the cost of a text. thus, five people could share the cost of the site registration, and then all share in the password provided. the second way to overcome the economic incentive problem is to publish a booklet that contains all of the introductory and connecting material as well as the end of chapter material. in other words, a typical chapter in a textbook today that has 25 pages or so of text and five 308 w. woerheide / financial services review 8 (1999) 305–317 pages of end of chapter material, could be replaced with a chapter that runs for five pages, and has two pages of end of chapter material. thus, the author would now be able to participate in the royalties from the sale of a book. in fact, such a booklet could contain the urls for at least some of the sites that are key to each chapter, rather than having all of the sites reached through a textbook home page. the obsolescence factor could be dealt with through an erratum on the internet. because the erratum is cumbersome, it would be appropriate to update the booklet on an annual basis. thus, although the booklet would have a smaller royalty than a full-length text, the smaller royalty would be offset by the fact that there would be little in the way of a secondary market for such a booklet, and a new edition could be sold every year. even if a way is found to effectively publish the book described above, there would still be problems in getting people to adopt such a book (whether it is an on-line pay-site book, a booklet that links to a web site, or a booklet with lots of sites listed). the most significant problem is that an adopting instructor would have to check out many of the sites in order to know how to use each one. two other problems are that web addresses have an obsolescence factor (as previously mentioned) and an availability factor. let us consider each of these problems. this author’s own experience in doing web-related writing in recent years is that the annualobsolescence rate of web addresses is about 40%. this 40% rate merits additional comments. first, if one sticks to thehome pageaddress of large, established organizations (e.g., www.ssa.gov, www.schwab.com, and www.aaii.org), then the obsolescence rate is at most a few percentage points. if one uses detailed addresses within established organizations (such as a page requiring four links from the home page to arrive at), then the obsolescence rate may approach ten percent. in many cases the page is still there, it is just that the details of the direct address may change. finally, the sites provided by individuals and small firms (such as a law or an accounting firm) will have an attrition rate of over 50%. it also appears that the attrition rate for successful pages is declining. that is, web pages that have been successful in attracting traffic are more likely to stay. thus, if a person identifies say 1,000 sites that are useful in a personal finance course, then one-year later, maybe 600 of these sites would still be active. however, two years later, a fairly high portion of this remaining number of sites would still be active. the last problem to adoption is the availability factor. servers go down. no one likes to be in front of a class doing a lecture and dial into a web site only to receive a notification that said site is not responding at that time. similarly, it would be frustrating to a student to be reading a text and needing to go to a particular site, only to find that site is not currently available. there are two solutions to the availability problem. one is to have a back-up site available. for many topics and calculators, there is more than one site for each such topic or calculator that would do the trick. second, in terms of class presentations, an instructor might prepare a transparency of key sites in advance in case an unique site is not available, and perhaps even have handouts available for the student. these supply-side and demand-side problems of an internet text will take time to overcome. although an all-internet-based text is the future, it likely will be a long time before we see such a text. 309w. woerheide / financial services review 8 (1999) 305–317 5. what is the current state of internet integration? until we have figured out how to overcome the supply and demand problems of an all-internet based course, we will be in a transition phase of moving from the traditional textbook-based course supported by such things as powerpoint slides, spreadsheet templates, and personal financial planning disks. i am calling the current state of textbook publication the transition phase based on two observations. first, the current edition of almost every personal finance text has internet-based material. second, almost none of the prior editions of each of these books provides any internet based material beyond perhaps a few references in the instructors’ manuals. 6. what does the transition phase look like? to clarify what this transition phase looks like, tables 1 and 2 show most of the personal finance texts now being adopted at colleges and universities. table 1 lists the authors, book titles and publishers, and copyright dates of the specific books reviewed for this study. it also provides a summary of the internet usage and integration. the final column is a rating of these books based on thecomprehensivenessof their internet usage and integration in each text. [this is not, repeat, not a rating of the quality of the texts or an endorsement or criticism of any type. this rating is simply a measure of the degree of usage and integration of the internet.] category i texts are the ones that appear to be at the cutting edge of usage and integration of the internet. generally, they provide lists of links at the end of the chapter, a publisher sponsored web site that includes additional exercises and problems along with web links that will provide the solutions, resources for the professor, and resources for the student. category ii texts provide a subset of what can be found in a category i text. category iii texts provide only an occasional listing of web sites. there is no comprehensive incorporation or utilization of the internet in these texts. in establishing these categories, the use of the internet by the publisher is not considered internet integration into the text material. thus, the parts of the publishers’ web sites that provide the opportunity to order additional material such as study guides and instructor’s manuals are not counted in the categorization process. similarly, the parts of the web sites that are purely promotional are also ignored. promotional items include information on the authors, the text, and other texts by the same publisher. based on these criteria, the winger and frasca, rosefsky, keown, and kapoor, dlabay, and hughes texts are category i texts. category ii texts include ramaglia and macdonald and gitman and joehnk. category iii texts include garman and forgue, boone, kurtz, and hearth, and ho, perdue, and robinson. 7. what will the next round of textbooks look like? we will not jump to the ideal state, described in section 3, any time soon, and primarily for the reasons described in section 4. nonetheless, we are moving in that direction. the 310 w. woerheide / financial services review 8 (1999) 305–317 table 1 current status of the internet in various personal finance texts authora book (publisher) copyright internet content category bernard winger and ralph frasca personal finance, fourth edition (prentice hall) 1997 a short list of web sites is at the end of each chapter. in addition, there is a web site that supports the text. the url is www.udayton.edu/ sba/wf/pf.htm. the key features of this web site are exercises that provide links to online resources. there are also a large number of general links. other support material is also found at this site, including downloadable files of powerpoint presentations. spreadsheet templates, and a what’s new section. i robert rosefsky personal finance, seventh edition (john wiley & sons) 1999 throughout each chapter, icons have been placed next to paragraphs and other material. each icon provides a reference link at the following site: www.wiley.com/college/rosefsky/modl/. each reference link then provides links to sites with information relevant to the topic at hand, or provides a problem with links to one or more calculators that solve that problem. there is a decision maker software available, a business extra site which provides articles from thewall street journal, and rosefsky’s newsletters and updates. i judith ramaglia and diane macdonald personal finance: tools for decision making (southwestern college publishing) 1999 a short list of web sites is at the end of each chapter. in addition, there is a web site that supports the text. the url is www.swcollege.com/bef/ramaglia/ramaglia.html. the first element in this site is titled internet applications. these are the same as the links at the end of each chapter. there was also a section titled author updates, although this section had no content at the time of writing. ii arthur j. keown personal finance: turning money into wealth (prentice hall) 1998 sites are listed in the margin at the end of the chapter. however, the publisher also supports a web site for the book (www.prenhall.com/persfin). there are five activities for each chapter. these are: chapter interactive exercises, web-based chapter resources, hot topics, chapter articles from kiplinger’s, and a fun button. i jack kapoor, les dlabay, and robert j. hughes personal finance, fifth edition (irwin, mcgrawhill) 1999 at the end of each chapter, some of the projects and application exercises direct students to use the internet, but not all such exercises provide a specific address. in addition, a short list of web sites is presented at the end of each chapter. the substantial number of on-line resources is listed in table 2. the publisher also supports a web site for the book: www.mhhe.com/business/finance/kdh/. i (continued on next page) 311w. woerheide / financial services review 8 (1999) 305–317 transition phase will be a series of steps as each new edition of books comes out, and as some of the books seek to jump to the lead in terms of what they do with the internet. so, what is the next step? first, i think most of the books will move to incorporate some or all of the criteria used to define the category i texts. that is, most of the category iii texts will at least move to category ii in terms of providing a more comprehensive set of links. these links could be provided in the chapters, at the end of the chapters, or on a publisher sponsored internet site. the category ii texts will most likely move to category i in that they will expand their use of links, providing both substantive lists of links and exercises that are tied into links. the most interesting development will be what the texts currently in category i will have to do to keep at the front of the pack. i think the next step is to move from providing simple lists of internet addresses (category iii texts), tangentially related problems and exercises (category ii texts), or complex web support sites (category i texts), to actually integrating web sites into the presentation of the text material. two examples are provided to show what this means. 7.1. example i: the tax chapter most personal finance texts provide a figure of form 1040. in fact, by the time most textbooks are published, the form 1040 reproduced in the text is usually one year out of date. this form may be obtained from at least 20 different sites, including the irs site (www. irs.ustreas.gov/prod/forms_pubs/forms.html). if a text simply instructs the student to view form 1040 at any one of several sites, it would not be necessary to print it as a figure in the book. table 1(continued) authora book (publisher) copyright internet content category kwok ho, grady perdue, and chris robinson personal financial planning 1998 this book contains no internet-based material. iii lawrence gitman and michael joehnk personal financial planning, eighth edition (dryden press) 1999 there are some references to internet addresses in some of the chapters. at the end of each chapter are exactly five exercises, each of which provides an address. ii e. thomas garman and raymond forgue personal finance, fifth edition (houghton mifflin) 1997 the preface states that: “every chapter offers internet addresses for businesses, nonprofit organizations, and government agencies useful to the student of personal finance. web sites are indexed for convenience.” however, almost no such addresses were found. iii louis boone, david kurtz, and douglas hearth planning your financial future (dryden press) 1997 the preface states that: “we integrate the personal computer throughout the text. we highlight personal-finance computer sosftware and internet resources.” i found only one page that contained internet addresses, and this had seven addresses. iii a the authors are listed in reverse alphabetical order, a sequence that should be used more often to eliminate discrimination against people whose names start with letters near the end of the alphabet. 312 w. woerheide / financial services review 8 (1999) 305–317 most personal finance texts include a table showing the tax rate schedules. the tax rate schedules show the tax due and the marginal tax rates for each income level and each filing status. as with form 1040, the current version of these tables is available on the internet (www.irs.ustreas.gov/prod/ind_info/tax_tables/tax_sched.html). by directing the student to the web site with these tables, a book immediately eliminates one factor contributing toward obsolescence. 7.2. example ii: the housing chapter a critical decision that many people must make at least once in their lives is whether they should buy or rent. the topic is discussed in most textbooks, but this author is not aware of any books that work out an extensive, time value based problem or exercise for the financial table 2 specific comparisons of internet-material authors end of chapter listings chapter-related exercises on-line links other material on-line winger & frasca 113 101 (on-line) 383 spreadsheet templates, powerpoint presentations, what’s new section rosefsky 2 163 (on-line) 2 decision maker software, business extra site, rosefsky’s newsletter & updates ramaglia & macdonald 75 2 2 powerpoint prersentations keown 112 148 (on-line) 60 hot topics, articles from kiplinger’s, fun button kapoor, dlabay, & hughes 222 2 136 powerpoint presentations, instructor’s manual, personal finance links (which are grouped by topic area), pfp software (which allows for downloading the software planning package that accompanies the text, and s&p personal wealth (which is a stock market based web site), and an on-line learning center. for each chapter, this learning center provides chapter overview, learning objectives, key terms, powerpoint, pretest, post test, supplementary cases, supplementary reading, study questions, related web sites, summary, and spanish english glossary. ho, perdue, & robinson 2 2 2 2 gitman & joehnk (see chapter related exercises) 75 (in book) 2 2 gitman & forgue 3 2 2 2 boone, kurtz, & hearth 7 2 2 2 313w. woerheide / financial services review 8 (1999) 305–317 solution to this problem. this author used to do this in his lecture on housing. the presentation of the necessary assumptions (and a nontrivial exercise requires quite a few assumptions), the step by step process of deriving the numbers one needs to know, and the present value application for the solution would typically take at least 45 min. now, i go to a web calculator that provides this solution (www.calcbuilder.com/cgi-bin/calcs/ hom10.cgi/financenter). this calculator not only provides the financial answer in present value terms, but it also gives several critical graphs that show the importance of key variables to the solution of the problem. the buy versus rent decision can now be thoroughly discussed in about 15 min, and the topic more completely analyzed than it ever has been before, including what-if scenarios. the incorporation of a web site such as this into a text will allow much greater discussion and elaboration on the relevant issues, and much less time spent on the mechanics of setting up the problem. another key topic of the housing chapter is how one locates a house and how one selects a realtor. there are several sites that provide a checklist that would help a person define what he or she is really looking for. once the parameters of the desired property are defined, then one wants to see as many properties as possible that would meet these criteria. the site www.realtor.com has emerged as the national search site for homes in most communities. in my community, i can find more homes listed on the realtor.com site than i can in any of the sites supported by local brokerage firms or our local newspaper. it is inconceivable that the next round of textbooks could present the topic of finding a home without directing the student to this site as part of the chapter text. 8. is this forecast unanimously shared? this forecast of what the next round of texts will look like is not unanimously shared. the major objection to the concept of building a text on internet connections, as discussed earlier, is concern of what happens when link rot occurs. this is a legitimate concern, but it can be dealt with in several ways. one is volume. for many topics, one could provide multiple addresses that would work. with a normal attrition rate, at least some of the sites are likely to still be working when the next edition of the book replaces the current one. also, the text could be accompanied by a publisher supported web page that contains a frequently updated list of corrections, additions, or deletions for web sites provided in the text. a second approach is not to give any web sites in the text. in lieu of each web site, the text could provide a reference number to a publisher supported web page, and have the student find the address there. without the text itself, access to the web page would have little value to others. a second objection is that not enough finance professors know how to work effectively with the web, and this will prevent such a book from being sufficiently widely adopted to be successful. two years ago, this objection might well have been the telling indicator. but as was argued at the start of this paper, it appears that we are at a watershed in terms of the web becoming a standard tool of usage by most faculty. just as we went in a few short years from a few faculty using email to almost all faculty using email, so we will move from a few faculty accessing the internet regularly to most faculty accessing it regularly. a third objection is that too many students would not be able to access the internet and 314 w. woerheide / financial services review 8 (1999) 305–317 thus such a book as described above would not work. three years ago, this author gave as a term paper project in his personal finance course the task of each student locating 15 web sites, and explaining how he or she would use each site to make personal financial decisions. two years ago, the project was changed to having the students create their own web pages in which they defined personal finance questions they would likely be asking themselves in the next few years and they provided links which answered those questions. despite having hundreds of students go through this course in the last few years, not one student ever asked the instructor how to find a site on the internet, or even how to create a web page. most of my students can create vastly better looking web pages than i can! in fact, the most likely source of movement toward the incorporation of more internet material in the textbooks will be the students themselves. as students find these sites on the internet that they deem of value, they will start asking those instructors who are not using the internet why these sites are not being mentioned and utilized. this will lead these instructors to adopt textbooks that provide the strong internet usage and incorporation so that they will not have to spend many, many hours themselves finding these sites. 9. will students benefit from this integration? the most important person in the classroom is, of course, the student. as described in the historical section, there has been an evolution of technology in the classroom. however, there has been (to this author’s knowledge) little effort at ascertaining whether or not the student has benefited. what has happened is that the skills the student has acquired has changed. for example, when financial calculators became universal and were integrated into many textbooks and required by many instructors, no one seemed to ask if the student was better off using a calculator. many students have learned to solve time value of money problems by memorizing which keys to use to enter the data. they sometimes cannot write out the time value of money formulas associated with those problems. the amount of time spent in class and on homework doing calculations is greatly reduced. but, is their knowledge of time value of money greater or less? the enrollment in the personal finance course can be divided into two groups. the largest group is those students for whom this is their only exposure to this subject. the other group is those who wish to major in this subject area (perhaps in preparation for obtaining their cfp designation). for the first group, their motivations in taking the course usually are to learn about the issues in which they will have to make personal financial decisions during their lifetime. because of the way in which technology, laws, and practices change over time, we cannot in one course hope to teach these students everything they will need to know in their lifetimes. we can only hope to give them a framework on what are good personal finance decisions and practices today, and the ability to acquire the necessary knowledge and skills in the future. internet integration should serve both of these objectives well. internet integration should allow the student to access information supplemental to the text, as well as information that is more timely than can sometimes be provided in a text. internet integration will also provide the student with the skills to acquire personal financial information over his or her lifetime. students taking personal finance as a one-time course are 315w. woerheide / financial services review 8 (1999) 305–317 usually more interested in learning how to get answers to questions than in learning how to solve problems themselves. the internet calculators are great in this respect. the internet integration will likely still be helpful for students wanting to major in this area, but not as critical. the access to supplemental and more up-to-date information is just as valuable for these students, as is the knowledge of how to access such information on the internet. however, students majoring in personal finance need to know the mechanics of how to solve the various types of problems in the course. they need to know the theory behind the calculations being performed. internet integration is unlikely to do this. however, such students will also be taking more than just this one course in this area. it is likely they will learn the theory and skills of doing these calculations in their other courses. it would be nice to collect some empirical data on whether internet integration makes the student more knowledgeable and more skilled. an obvious technique is to run multiple sections of the course, and use the internet in some classes and not in others. the problem here is that these students will in fact be learning slightly different information and skills. for example, the student in the non-internet course may learn how to calculate a mortgage payment given a mortgage rate, but the student in the internet course will learn sites that will do this calculation, as well as how to shop for a mortgage on the internet. thus, does the student learn more, learn quicker, or learn better with internet integration is a moot question. the real question is whether the student is better prepared to handle his or her personal finance problems over his or her lifetime. we can only make a professional guess on this question, i do not think we can empirically answer it. having said this is an unanswerable question, there is some anecdotal evidence that internet integration improves the educational experience for the student. session 6 of the 1999 academy of financial services annual meeting was a tutorial entitled “teaching personal finance online.” one of the presenters had heard the presentation of an earlier version of this paper at the 1998 meeting, and took up the challenge of attempting to construct the ideal internet integrated course. the course developed was a distance learning course. the presenters noted that compared to another course taught on-campus in the traditional approach, the students in the internet integrated course appeared to be more satisfied with the course and to have learned more. 10. summary there are four points to this article: 1. the current and forthcoming editions of personal finance textbooks on the market mark a major turning point in terms of going from being virtually unaware of the internet, to incorporating the internet. 2. the most aggressive next set of personal finance textbook editions will incorporate the internet into the chapters as a crucial part of the text, rather than just presenting it as supplemental information and optional activities. 3. the internet will eventually replace our personal finance textbooks as we currently 316 w. woerheide / financial services review 8 (1999) 305–317 think about them. we may use only shorter reference books that provide the transition to move from web site to web site. 4. the internet integration is changing what the student learns in the personal finance course, and what skills the student acquires. we can only surmise that these changes will make the students better prepared to manage their personal finances over their lifetimes. acknowledgments the author wishes to thank the editor and two anonymous referees for their comments. several comments and thoughts of the referees have been added to the revised version of this paper. i also wish to thank bob bohn for organizing the panel session at the 1998 afs meeting at which many of the ideas for this paper were first presented. i thank don holdren for his positive support of the ideas presented herein by accepting the challenge delivered at that session and developing an all internet, textbook free course. 317w. woerheide / financial services review 8 (1999) 305–317 pii: s1057-0810(99)80012-2 i financial services review, 7(i): 45-55 issn: 1057-0810 copyright © 1998 by jai press inc. all rights of reproduction in any form reserved. explaining persistence in mutual fund performance f. larry detzel and robert a. weigand this study investigates the determinants of persistence in mutual fund performance. previous research that uses factor-mimicking portfolios and characteristic benchmarks to model fund performance fails to explain all the persistence in fund returns. this study employs a model that directly relates mutual fund returns to the characteristics of the stocks hem by funds. adjusting fund returns for the size of the stocks in which funds invest and financial ratios intended to capture fund manager investment styles explains all the persistence in mutual fund returns from 1976-1985, the period in which persis tence is most prevalent. "past performance is no guarantee of future results. "--sec i. introduction despite the sec's admonition regarding investing based on past performance, studies into the behavior of mutual fund investors find that prior period returns are the most significant determinant of new money flows into mutual funds (see carhart, 1997; gruber, 1996; ippolito, 1992; lakonishok, shleifer, & vishny, 1992; and patel, zeckhauser, & hen dricks, 1992). the perception that recent performance is an important consideration in mutual fund selection is undoubtedly enhanced by the marketing methods of the funds themselves. mutual funds devote significantly more print space to reporting their past returns than to the sec's required warning regarding persistence-based investment strate gies. this study investigates the source of persistence in mutual fund performance to help investors better understand what information is relevant when choosing a fund. the results indicate that certain characteristics of the stocks held by mutual funds explain all of the per sistence in fund returns. mutual fund investing has enjoyed phenomenal growth in recent years. at the end of 1995, investors held almost $1.3 trillion in assets at over 2,200 domestic stock mutual f. larry detzel and robert a. weigand • college of business and administration, university of colorado at colorado springs, 1420 austin bluffs parkway, colorado springs, co 80933; phone: (719) 262-3676, or (719) 262-3120. 46 financial services review 7(1) 1998 funds. by the end of 1997, investor holdings approached $2.5 trillion (investment com pany institute, 1998). the flow of new money into the best performing funds far surpasses new investment in funds that lag the overall market (see carhart, 1997 and gruber, 1996). this study investigates the underlying factors that explain the apparent "momentum" in mutual fund returns. while most prior studies find evidence of momentum in fund returns, authors disagree regarding the source of persistence and whether persistence-based investing can generate excess returns. brown and goetzmann (1995) conclude that "investors can use historical information to beat the pack" (p. 697), but also find that investing based on persistence exposes investors to greater total risk than other strategies. carhart (1997) finds that almost all the predictability in mutual fund returns is explained by common factors in stock returns and systematic differences in mutual fund expenses and transaction costs. consistent with the findings of carhart (1997), daniel, grinblatt, titman, and wermers (1997) conclude that actively-managed mutual funds beat mechanical trading rules based on persistence- but only by an amount equal to the average management fee. malkiel (1995) reports that survivorship bias accounts for a significant amount of performance persistence, which implies that the returns to persistence-based investment strategies may be overstated. golec (1996) finds a relation between mutual fund performance, risk, and fees and fund manager characteristics such as age, level of education, and length of tenure with the fund. porter and trifts (1998) study the performance of fund managers who manage the same fund for at least ten years. they present evidence that inferior performance is more likely to persist than superior performance. a considerable body of research suggests that the cross-sectional pattern of stock returns can be explained by characteristics such as finn size, past returns, earnings-to-price ratios, and book-to-market ratios. examining the effect of these variables in an integrated framework, fama and french (1992, 1995, 1996) conclude that the cross-sectional varia tion in expected returns can be largely explained by only two of these characteristics, size and book-to-market equity. a mutual fund's investment policy will tend to favor stocks of a particular size and style class (e.g., small-capitalization value stocks or large-capitalization growth stocks). if the cross-sectional returns of individual stocks can be explained by characteristics such as firm size and book-to-market equity, it is reasonable to expect that mutual fund returns can also be modeled in a similar manner. in this study, size and style characteristics are repre sented by market capitalization, the ratio of book-to-market equity, the ratio of earnings-to market equity (earnings yield), and the ratio of cash flow-to-market equity (cash flow yield). research indicates that mutual funds tend to maintain their investment strategies over long periods of time (malkiel, 1995). consequently, mutual fund performance will corre spond to the performance trends of the size and style classes in which funds invest. if there are periods when small stocks tend to outperform large stocks, or value stocks outperform growth stocks, then persistence in stock mutual fund returns could be due to trends in these underlying factors. this is the basis of the hypothesis tested: hi: mutual fund returns that have been adjusted for size and style character istics, as well as market risk and expense ratios, will display no serial correlation. explaining persistence 47 previous studies conclude that persistence in mutual fund returns cannot be fully explained by fund characteristics such as recent relative performance, firm size, and book to-market equity (carhart, 1997, daniel et al., 1997; gruber, 1996; and hendricks, patel, & zeckhauser, 1993). recent research suggests that these findings may be due to the use of factor-mimicking portfolios that are constructed to match the characteristics of the stocks held by a mutual fund. daniel and titman (1997) report that firms' actual size and book to-market equity contain more explanatory power than time-series estimates of loadings on factor-mimicking portfolios. this study employs a model that directly relates mutual fund returns to the characteristics of the stocks held by funds. consistent with the results reported by previous researchers, unadjusted mutual fund returns display significant persistence, as do fund returns that have been adjusted for market risk and expense ratios. however, accounting for firm size and fund manager investment styles explains all the persistence in mutual fund returns from 1976-1985, the period in which persistence is most prevalent. ii. data and methodology the sample consists of 61 open-end equity mutual funds classified in the 1975 wiesen b e r g e r i n v e s t m e n t c o m p a n i e s serv ice as "growth" or "growth and current income." the sample was selected at random from all such funds that have returns data reported in the 1975 wiesenberger . t h e population of"general equity" funds at year-end 1974 totaled 230 (malkiel, 1995). six of the 61 funds did not survive the entire sample period because they either merged with another fund or were liquidated. these funds are included in the sample until the year before they terminate. seventy-two percent of the funds in the sample charge an up-front sales fee or "load". recent research attempts to explain the persistence in mutual fund performance using factor models (see carhart, 1997 and gruber, 1996) or benchmark portfolios designed to mimic characteristics of the component stocks held by funds (see daniel et al., 1997). these studies conclude that differences in beta, firm size, book-to-market equity, interest rates, and prior period performance cannot fully explain the persistence in mutual fund returns. daniel and titman (1997) question whether factor models adequately explain the cross-section of expected stock returns. after controlling for firm characteristics, they find that expected returns are unrelated to the loadings on market, firm size, and book-to-market equity factors. they conclude that it is firm characteristics rather than covariances that determine expected returns. based on these findings, a model is developed that directly relates mutual fund returns to the characteristics of the stocks held by each fund. the model expresses returns in year t as a function of year t 1 mutual fund characteristics. three specifications of the model are estimated: rit = ~lt + (~2tbetai, t 1 + (t3texpi, t 1 + r~ rit = ~ l t + ~j2tbetai, t 1 + ~3texpi, t 1 + ~4tsizei, t 1 + r~/t (1) (2) rit = ~ l t + ~2 tbe ta i , t 1 + ~)3texpi, t 1 + ~4tsizei, t 1 8 + ~ t 1 + ~ 5 t b / m i , t 1 + ~ 6 t e / m i , t 1 + 8 7 t c f / m i , t l + rit (3) 48 financial services review 7(1) 1998 where t r i t = beta/, t 1 = expi, t i = size/, t 1 = bitvli , t _ i = e ~ i , t _ 1 = c f / m i , t 1 = o~ 1 t o ~ 3 , = 131 to 132, 81 toa7 r i t , r . , r i t = each year 1975 through 1995; total return of mutual fund i in year t; market risk of fund i in year t 1; expense ratio of fund i in year t 1; natural logarithm of the median market capitalization of the common stocks held by mutual fund i at the end of year t 1; median ratio of book-to-market equity of the common stocks held by mutual fund i at the end of year t 1; median earnings yield of the common stocks held by mutual fund i at the end of year t 1; calculated as income before extraordinary items less preferred dividends divided by the market value of common stock; median cash flow yield of the common stocks held by mutual fund i at the end of year t 1; calculated as the cash flow available to common stock divided by the market value of common stock; regression parameters to be estimated (estimation methods described below); year t characteristic-adjusted regression residuals estimated from equations 1, 2 and 3 above. the variables included in the above models are motivated by studies of the cross-sec tion of stock and mutual fund returns. each mutual fund's beta is estimated via ols regres sion, using monthly returns for the 36 months preceding year t. the center for research in securities prices value-weighted index of all nyse, amex, and nasdaq stocks is used as the market proxy. malkiel (1995), gruber (1996), and carhart (1997) find that mutual fund expense ratios are significant in explaining fund performance. these expense ratios are therefore obtained from wiesenberger and morningstar mutual funds ondisc and included as an explanatory variable in the models. firm size, measured as the natural log arithm of the median market capitalization of the stocks held by each fund, is included as a control variable. three ratios intended to capture fund manager investment styles are also included: the median ratio of book-to-market equity (bdvl) of the stocks held by each fund; the median earnings yield (f_/m); and the median cash flow yield (cf/m). the stocks held by each mutual fund are identified using investment schedules reported in moody's bank and finance manual, q-data corporation's sec file, or morningstar, inc.'s mutual fund sourcebook for each year t 1, 1974-1994. financial data on the stocks held by each fund are obtained using standard & poor's compustat annual files database. equations 1, 2, and 3 above are estimated using the stacked cross-sectional regression approach of fama and macbeth (1973). during the sample period 1975-1995 one cross sectional regression is estimated for each year t. this results in 21 cross-sectional regres sions, beginning with 61 observations in 1975 and ending with 55 observations in 1995 due to six non-surviving funds. the time-series means of the standardized slope coefficients from these regressions provide a basis for comparing the relative contributions of the explanatory variables in explaining mutual fund performance. summing the squares of the slope-coefficient t-statistics yields x 2 statistics (see bajaj & vijh, 1995) that test the signif icance of mutual fund characteristics in explaining the cross-sectional variation in annual explaining persistence 49 fund returns. a significant )c 2 statistic indicates that mutual fund returns are related to the characteristics of the stocks held by each fund. equations 1, 2 and 3 above model mutual fund returns as a function of one or more characteristics identified in asset-pricing studies as significant in explaining the cross-sec tional variation in stock returns. thus, the residuals from these regressions r~, r~, r/~ may be viewed as characteristic-adjusted mutual fund returns. if the persistence in fund returns is related to these characteristics, then these adjusted returns should display less serial cor relation than unadjusted mutual fund returns. accordingly, the following models test for persistence in adjusted fund returns: rit = c + p ri, t _ l +£it (4) r~t = cfs+ p~lr~t_l + e~t (5) 8 8 8 8 rit = c + p ri, t _ l +£it (6) where rit, r . , tit = year t characteristic-adjusted returns of fund i estimated from equations l, 2 and 3 above; pct, p[~, p8 = first-order serial correlation coefficients; eit, £it, eit = year t regression residuals estimated from equations 4, 5, and 6. if the size and style characteristics in equations l, 2, and 3 explain the persistence in mutual fund performance, there will be no serial correlation between adjusted fund returns in year t and year t 1, and the p-coefficients will be insignificantly different from zero. finding no serial correlation supports the hypothesis that the characteristics of stocks held by funds explain the persistence in mutual fund returns. ill empirical results table 1 reports average descriptive statistics for the funds in the sample over the period 1975-1995. there is considerable variation among fund characteristics. for example, mutual fund market risk (beta) ranges from 0.50 to 1.60. the average expense ratios dis play substantial variation as well, ranging from 0.30 percent to 1.88 percent of net assets. the size and style characteristics of the stocks held by funds in the sample also differ con siderably. the mean market value of the stocks in each fund ranges from $66 million to $8.1 billion, while the mean book-to-market ratio ranges from 0.20 to 0.98. the sample includes funds with preferences for both small and large stocks as well as growth and value investing. the average annual total return of the funds in the sample varies widely, ranging from -2.4 to 41.1 percent. table 2 reports results from the three characteristic-model regressions (equations 1, 2, and 3). the results suggest that the characteristics of the stocks held by mutual funds are useful in explaining annual fund returns. across all three models, mutual fund returns are positively related to beta at the one percent level. mutual fund expense ratios are also sig 50 financial services review 7(1) 1998 table 1 sample descriptive statistics return mkt cap b/m elm cf/m beta exp net assets mean 16.9 2680 0.55 0.08 0 .14 1.04 0.95 453.48 min imum 2 . 4 66 0.25 0.05 0.06 0 .50 0 .30 12.33 i st q u a r t i l e 11.2 1112 0 .40 0.07 0 .10 0 .90 0 .70 56.97 median 16.1 2468 0.52 0.08 0 .14 1.00 0.93 156.11 3 rd q u a r t i l e 21.9 3924 0.67 0.09 0.17 1.20 1.14 378.57 m a x i m u m 41.1 8137 0.98 0.12 0.23 1.60 1.88 4999 .60 standard deviat ion 8.8 1970 0.18 0.02 0.04 0.21 0.32 909.0 notes: this table reports descriptive statistics for the sample of 61 mutual funds. average statistics are reported for each fund using annual data from 1975-1995. variable definitions are given at the bottom of the table. definition of variables: return = total return (percent) during each year t; beta = mutual fund market risk estimated over the thirty-six months ending with each year t 1; exp = expense ratio in each year; mkt cap = median market value (millions of dollars) of the common stocks held at the end of each year; b/m = median book value-to-market value of the common stocks held at the end of each year; elm = median earnings yield of the common stocks held at the end of each year; earnings yiem is income before extraordinary items less preferred dividends divided by common stock market value; cf/m = median cash flow yield of the common stocks held at the end of each year; cash flow yiem is cash flow available to common stock divided by common stock market value. net assets = mutual fund net assets (in millions of dollars) at the end of each year. nificant in explaining fund returns. although the incomplete specification in model 1 yields an unexpected positive coefficient, the more complete specifications of models 2 and 3 produce the expected negative coefficient (significant at the five percent level). model 2 also includes the natural logarithm of the median market capitalization of the stocks in each fund (size) as an explanatory variable. size displays the expected negative coefficient and is significant at the one percent level. the size variable improves the explanatory power of the model, with the average adjusted r 2 increasing from 15 to 30 per cent. model 3 incorporates the median values of three financial statement ratios intended to capture fund manager investment style: book-to-market (b/m), earnings-to-market (e/m), and cash flow-to-market (cf/m). while b/m and e/m display the expected positive coef ficient, cf/m is insignificant in the regression. the e/m and b/m variables are significant at the one percent level. the adjusted r 2 of 42 percent indicates that inclusion of these vari ables substantially improves the fit of the model. these findings provide support for the idea that annual mutual fund returns are related to the size and style characteristics of the stocks held by funds. table 3 reports average serial correlation coefficients between mutual fund returns in years t and t 1. the correlations are estimated using the models shown in equations 4, 5, and 6. results are reported for both raw mutual fund returns and for fund returns that have been adjusted for fund size and style characteristics using the regression models shown in equations 1, 2, and 3. panel a of table 3 reports results for the period 1976-1995, while panels b and c report results for the 1976-1985 and 1986-1995 periods, respectively. the results reported in panel a show that, for the entire 20-year period, the mean annual serial correlation coefficient from year-by-year regressions of unadjusted fund returns in year t on year t 1 returns is 0.12. the related t-statistic is 1.62, which is signif explaining persistence 51 table 2 time-series means of cross-sectional regression standardized coefficients model coefficient intercept beta exp size b / m e l m cf,/m adj. r 2 1 m e a n 7.77 0 .10 0.01 )~2 234.73** 235.02** 62.91"* t-statistic 2.26* 1.27 0.24 2 m e a n 22.21 0 .06 0 . 0 5 ~2 240.01"* 179.58"* 31.69* t-statistic 3.33** 0.83 1 . 4 6 3 m e a n 16.72 0.13 0 . 0 4 ~2 175.80"* 94.04** 32.18" t-statistic 2 .12" 2.80** 1 . 3 2 0 . 1 4 405.65** 1 . 3 4 0 . 1 1 0.03 0 .10 -0 .01 305.04** 56.74** 67.48** 27.98 1 . 1 8 0.33 1.13 0 . 1 4 0.15 0 .30 0 .42 notes: this table reports results from the regression models shown in equations 1, 2 and 3, which model annual mutual fund returns as a function of the characteristics of the stocks held by each fund in a given year: rit = (~lt + o~2tbetai, t 1 + ct3texpi, t 1 + r~/ (1) rit = ~ l t + 132tbetal, t i + [~3texpi. t i + [~'lt size + ~t (2) rit = ~)lt+~2tbetai, t _ l +~)3texpi, t l +~)4tsizei, t _ l +~)i,t_l +~5tb/mi, t _ l +66te/mi, t_ l (3) + ~7tcf/mi, t i + rit twenty-one cross-sectional regressions are estimated (one for each year 1975-1995). the regression coefficients reported are calculated as the time-series means of the cross-sectional regression standardized coefficients. (the regres sion intercepts do not have standardized coefficients, and are therefore reported in unstandardized form.) t h e )~2 statistics are obtained by summing the squares of the regression coefficient t-statistics. a significant )~2 value indicates that the variable is significant in explaining mutual fund returns. the res are calculated as the time-series mean adjusled-r2s from all cross-sectional regressions. the t-statistics are calculated as the slope coefficient time-series means divided by the time-series standard errors. **, * significant at the one and five percent levels, respectively. definition of variables: rit = total return of mutual fund i in year t; beta/, t 1 = expi, t 1 = size/, t 1 = b,qvli. t _ 1 = ~ i , t i = cf/mi, t 1 = market risk of fund i in year t 1; expense ratio of fund i in year t 1; natural logarithm of the median market capitalization of the common stocks held by mutual fund i at the end of year t 1; median ratio of book-to-market equity of the common stocks held by mutual fund i at the end of year t 1; median earnings yield of the common stocks held by mutual fund i at the end of year t 1; calculated as income before extraordinary items less preferred dividends divided by the market value of common stock; median cash flow yield of the common stocks held by mutual fund i at the end of year t 1; calculated as the cash flow available to common stock divided by the market value of common stock. icant at the ten percent level. mutual fund returns display only mild persistence over the entire period 1976--1995. the persistence in fund returns that have been adjusted for market risk and expense ratios (r~t) are examined next. the average first-order serial correlation coefficient for betaand expense-adjusted returns increases slightly, from 0.12 to 0.15. the t-statistic on the mean of the pet-coefficients from equation 4 is 2.63, indicating that these correlations are, on average, significantly greater than zero. consistent with the findings reported by brown and goetzmann (1995, pp. 691-693), adjusting mutual fund returns for market risk and expense ratios does not explain the persistence in fund returns. the next average serial correlation coefficient reported in table 3 is from fund returns that have been adjusted for beta, expenses, and the median size of the stocks held by each fund (r~t). the year-by-year persistence in mutual fund returns is reduced by inclusion of 52 f i n a n c i a l s e r v i c e s r e v i e w 7(1) 1998 table 3 first-order serial correlation coefficients of adjusted mutual fund returns variables in per formance -charac te r i s t i c s m o d e l mode l s 1, 2, and 3 p a r a m e t e r value panel a: 1976-1995 unadjusted mutual fund returns p 0.12 t-statistic 1.62 § 4 beta, exp pet 0.15 t-statistic 2.63** 5 beta, exp, size p13 0.11 t-statistic 1.62 § 6 beta, exp, size, b/m, e/m, cf/m p8 0.06 t-statistic 1.25 panel b: 1976-1985 unadjusted mutual fund returns p 0.24 t-statistic 2.46** 4 beta, exp pa 0.19 t-statistic 2.89** 5 beta, exp, size pl~ 0.10 t-statistic 1.51 6 beta, exp, size, b/m, e/m, cf/m p~ 0.03 t-statistic 1.02 panel c: 1986-1995 unadjusted mutual fund returns p -0.01 t-statistic -0.06 4 beta, exp pa 0.12 t-statistic 1.12 5 beta, exp, size p13 0.12 t-statistic 0.99 6 beta, exp, size, b/m, e/m, cf/m p~ 0.09 t-statistic 0.99 notes: this table reports results from the regression models shown in equations 4, 5, and 6, which regress adjusted annual mutual fund returns on lagged adjusted fund returns (the regression residuals obtained from estimating equations 1, 2. and 3): ix (x ~ ~t ix rit = c + p ri, t 1 + £it ( 4 ) r~ c~ + pflr~ t + ~ ( 5 ) • = ", 1 6 6 5 5 6 rit = c + p ri, t i + ~'it (6) the average first-order serial correlation coefficients from adjusted mutual fund returns are compared to the average serial correlation coefficient from unadjusted fund returns. finding that the adjustment regressions (equations 1, 2. and 3) decrease the serial correlation in fund returns supports the hypothesis that the characteristics of the stocks held by funds explains the persistence in mutual fund returns. the first-order serial correlation coefficients reported are calcu lated as the time-series means from all cross-sectional regressions. the t-statistics are calculated as the slope coefficient time-series means divided by the time-series standard errors. **, * significant at the one and five percent levels, respectively. § significant at the ten percent level. definition of variables: ¢t r[~ 6 = year t characteristic-adjusted returns of fund i estimated from equations 1, 2 and 3 above; rip it' rit pet, pl3 ps = first-order serial correlation coefficients; a 13 6 = year t regression residuals estimated from equations 4, 5, and 6. eit' eit" £it explaining persistence 53 the size variable. the mean correlation between annual fund returns declines from 0.15 in model 4 to 0.11 in model 5. the related t-statistic is 1.62, which indicates that adjusting annual fund returns in this manner explains only a small amount of the persistence in fund returns. the final results reported in panel a of table 3 are from estimation of equation 6, which examines the correlations between annual fund returns adjusted for beta, expenses, firm size, and the three characteristic ratios b/m, e/m, and cf/m. the average of the year by-year correlations between fund returns adjusted as such is 0.06. the t-statistic on the ps coefficients from equation 6 is 1.25, indicating that on average these correlations are insig nificantly different from zero. despite the mild persistence in fund returns over the 1976 1995 period, there is a significant reduction in persistence after adjusting for beta, expense ratios, firm size and investment style. these results provide support for the hypothesis that accounting for the size and style characteristics of the stocks held by funds using the more parsimonious and direct methods suggested by daniel and titman (1997) can fully explain the persistence in mutual fund returns. previous research into mutual fund persistence finds that "the strongest evidence for repeat performance is over the late 1970s and early 1980s" (brown & goetzmann, 1995, p. 689). for this reason persistence coefficients are reported for the 1976-1985 and 1986 1995 subperiods in panels b and c of table 3. consistent with the findings of previous researchers (brown & goetzmann, 1995; hendricks et al., 1993; and malkiel, 1995), unad justed mutual fund returns display strong persistence in the 1976-1985 period (p = 0.238, t = 2.46). the average first-order serial correlation coefficient for betaand expense adjusted returns (model 4) decreases only slightly, from 0.24 to 0.19. the t-statistic on the mean of the pa-coefficients from model 4 is 2.89, indicating that these correlations remain significantly greater than zero. as was the case in panel a, adjusting mutual fund returns for market risk and expense ratios does not explain a significant amount of the persistence in fund returns. moving from model 4 to model 5 in panel b reveals that including the size variable in the model reduces the average fund persistence coefficient from 0.19 to 0.10. inclusion of the investment style variables b/m, e/m, and cf/m further reduces the year by-year persistence coefficients to 0.03 (see model 6). the t-statistic of 1.02 is insignifi cantly different from zero, which indicates that accounting for the size and style character istics of the stocks held by funds using more direct and parsimonious modeling methods explains all the persistence in fund returns for the 1976-1985 period. the results reported in panel c of table 3, covering the period 1986-1995, demon strate why the characteristics model explains more of the persistence in mutual fund returns from 1976-1985 than for the entire 20-year period (panel a). confirming the findings of previous studies, there is virtually no persistence in fund returns after 1985. none of the persistence coefficients reported in panel c are significant at conventional levels. iv. conclusions and implications this study investigates the factors contributing to persistence in mutual fund performance. the particular hypothesis tested is that a greater amount of the persistence in mutual fund returns can be explained than has been found by previous researchers (carhart, 1997; daniel et al., 1997; and gruber, 1996). motivated by recent studies into the cross-section 54 financial services review 7(1) 1998 of expected stock returns (daniel & titman, 1997), a model is developed that avoids the use of factor-mimicking portfolios and characteristic benchmarks and instead directly relates mutual fund returns to the properties of the stocks held by funds. consistent with the results reported by previous studies, market risk and fund expense ratios explain only a small amount of the momentum in mutual fund returns. examining the period in which mutual fund return persistence has been most pronounced (1975-1986), however, the results indicate that accounting for the size of the stocks held by funds and fund manager investment styles (characterized by ratios such as book-to-market, earnings to-market, and cash flow-to-market) explains all of the persistence in mutual fund returns. both firm size and investment style characteristics contribute to explaining persistence. as found by previous studies, there is little evidence of momentum in fund returns during the late 1980s and early 1990s. these findings suggest that investors interested in allocating money among mutual funds would be wise to consider more than recent past performance. investors should also take into account recent trends in the overall stock market, such as whether large company stocks are outperforming small company stocks and whether value stocks are outperform ing growth stocks. the persistence in fund performance appears to be driven almost entirely by trends in these well-known and widely-publicized investment categories. in other words, instead of simply buying the best-performing funds from prior periods, inves tors should identify the size and style characteristics of funds and research current market trends in these factors. during periods when large-capitalization stocks begin outperform ing smaller stocks, buying funds that invest in larger stocks should also produce superior results. similarly, recent trends in value and growth stocks should be reflected in the rela tive performance of funds that invest according to these criteria. acknowledgment: the authors would like to thank charles bailey, thomas evans, karen eilers lahey (the editor), james mcbrayer, paul miller, shawn phelps, and two anonymous reviewers for their helpful comments and suggestions. references bajaj, m., & vijh, a. (1995). trading behavior and the unbiasedness of the market reaction to divi dend announcements. journal of finance, 50, 255-279. brown, s., & goetzmann, w. (1995). performance persistence. journal of finance, 50, 679--698. carhart, m. (1997). on persistence in mutual fund performance. journal of finance, 52, 57-82. daniel, k., & titman, s. (1997). evidence on the characteristics of cross sectional variation in stock returns. journal of finance, 52, 1-33. daniel, k., grinblatt, m., titman, s., & wermers, r. (1997). measuring mutual fund performance with characteristic-based benchmarks. journal of finance, 52, 1035-1058. fama, e., & macbeth, j. (1973). risk, return, and equilibrium: empirical tests. journal of political economy, 81,607--636. fama, e., & french, k. (1992). the cross-section of expected stock returns. journal of finance, 47, 427--465. fama, e., & french, k. (1995). size and book-to-market factors in earnings and returns. journal of finance, 50, 131-156. explaining persistence 55 fama, e., & french, k. (1996). multifactor explanations of asset pricing anomalies. journal of financial economics, 51, 55-84. golec, j. (1996). the effects of mutual fund managers' characteristics on their portfolio performance, risk and fees. financial services review, 5, 133-147. gruber, m. (1996). another puzzle: the growth in actively managed mutual funds. journal of finance, 51,783-810. hendricks, d., patel, j., & zeckhauser, r. (1993). hot hands in mutual funds: short-run persistence of relative performance, 1974-1988. journal of finance, 48, 93-130. investment company institute. (1998). mutual fund fact book. washington, dc: ici. ippolito, r.a. (1992). consumer reaction to measures of poor quality. journal of law and econom ics, 35, 45-70. lakonishok, j., shleifer, a., & vishny, r. (1992). the structure and performance of the money man agement industry. brookings papers on economic activity: microeconomics, 339-391. malkiel, b. (1995). returns from investing in equity mutual funds ! 971 to 1991. journal of finance, 50, 549-572. patel, j., zeckhauser, r., & hendricks, d. (1992). investment flows and performance: evidence from mutual funds, cross-border investments and new issues. in r. sato, r. levitch & r. ram achandran (eds.), japan, europe and the international financial markets: analytical and empirical perspectives. new york: cambridge university press. porter, g., & trifts, j. (1998). performance persistence of experienced mutual fund managers. finan cial services review, 7(1 ), 57--68. 37 elevating professional skills through authentic, scaffolded learning in a financial planning capstone elisabeth sinnewe1 abstract this paper presents an evidence-based redesign of a financial planning capstone unit in an accredited australian financial planning degree. five innovations were introduced to strengthen students’ professional capability development: (1) an authentic, dynamic client case; (2) a practitioner-led workshop on client engagement; (3) layered scaffolding to support digital literacy and professional communication skills; (4) a redesigned assessment structure featuring role-based expert pitches evaluated by industry judges; and (5) structured teamwork supports to develop collaborative capability. the impact of the redesign was evaluated by triangulating student performance, student evaluations, thematic analysis of reflective accounts, and an independent expert peer review. findings demonstrate improved communication, teamwork, and perceived job readiness, alongside stronger alignment between assessment tasks and real-world financial planning practice. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation sinnewe, e. (2025). elevating professional skills through authentic, scaffolded learning in a financial planning capstone. financial services review, 33(4), 37-47. introduction in the wake of financial advice scandals such as storm financial in 2009, the australian government raised the educational entry requirements for financial advisers with the introduction of the corporations amendment (professional standards of financial advisers) act 2017. 2 amongst other requirements, the increased standards require new financial advisers to have an approved degree listed in the corporations (relevant providers degrees, qualifications and courses standard) determination 2021. 1 corresponding author (elisabeth.sinnewe@qut.edu.au). queensland university of technology, brisbane, australia. 2 storm financial was an australian financial planning company which collapsed in 2009 after funnelling clients’ retirement savings into highly leveraged investments, resulting in substantial losses (parliamentary joint committee on corporations and financial services, 2009). as a result of these regulatory changes and increased industry demand (johnson et al., 2016), the australian higher education sector experienced a proliferation of financial planning degrees at the postgraduate and undergraduate levels from 1995, when only two degrees were offered, to 88 approved financial planning degrees at the undergraduate level and 91 at the postgraduate level offered by australian higher education providers across all states and territories (excluding the northern territory) by 2024. although prescribing technical competencies in key areas is an encouraging step toward establishing financial planning as a profession, technical knowledge alone is insufficient to ensure that graduates are https://creativecommons.org/licenses/by-nc/4.0/ mailto:elisabeth.sinnewe@qut.edu.au financial services review, 33(4) 38 “job ready” (goetz et al., 2005; brimble et al., 2012; west et al., 2019). one potential solution to improve graduate job readiness is to develop professional capabilities through real-world learning in collaboration with industry, typically in the form of structured work-integrated learning (wil) placements such as paraplanning traineeships (smith, 2012). however, following the hayne royal commission in 2019, major australian banks withdrew from offering wealth advisory services, causing a contraction in large-scale, structured wil opportunities. 3 with students reliant on small to medium-sized advice practices, many of which lack formal internship structures, universities face increasing pressure to create authentic, practice-oriented learning experiences within the academic setting. guided by research on authentic learning in developing professional capability (see e.g., kaider et al., 2017; stein et al., 2004), this paper presents an evidence-based redesign of the financial planning capstone unit, combining authentic, dynamic client scenarios with industry-led, applied skills workshops to replicate the kinds of professional learning experiences typically gained through workplace settings. the redesign of this unit was approached as a collaborative effort among industry partners, learning advisors, and academic experts to create scaffolds that support the development of professional skills, including digital literacy, teamwork, and communication with clients.4 to evaluate the redesigned learning experience, multiple sources of evidence were triangulated, including student academic performance, student evaluations, reflective accounts, and an independent expert peer review. together, these data provided early evidence of improved student capability development and stronger alignment between assessment activities and professional practice. yet, future improvements are needed to ensure that skills-development scaffolding is built at both the degree and unit levels. the efficient use of financial planning software was identified in the post-redesign student feedback as an ongoing challenge. this feedback indicates that the support designed to develop students’ digital capability requires recalibration. digital capability is a key professional skill that financial planning graduates need to adapt to technological advancements such as digital and mobile advice (power, 2017). another area of future exploration is establishing a community of best practices to ensure that financial planning educators work together to improve future offerings of financial planning degrees and uphold a high-quality standard of educational practice. the paper has practical implications for financial planning educators: academics may find the detailed description of the financial planning capstone redesign and its evaluation helpful when reviewing and updating their learning and assessment activities. it also offers value to practitioners by demonstrating how their expertise can be utilized in more sustained, strategic ways beyond guest lectures to help shape the next generation of financial planners, thereby strengthening the connection between financial planning education and the profession. this paper proceeds as follows: section 2 outlines the teaching approach in the financial planning capstone and introduces the innovations implemented through the redesign; section 3 reports on the impact of the redesign on student performance and student evaluations, including an independent expert peer review. section 4 discusses the educator and student reflections on the redesign experience, and section 5 concludes. 3 the royal commission into misconduct in the banking, superannuation and financial services industry (2017-2019), commonly known as the hayne royal commission, was a major public inquiry uncovering systemic compliance failures in australia. the findings led to substantial remediation costs, reputational damage, and increased regulatory scrutiny, prompting major banks to withdraw from providing financial planning services. 4 scaffolds in a teaching context refer here to instructional and other supports to assist students in completing complex professional tasks. for a discussion of how teaching scaffolds have been defined in the education literature, refer to simons and klein (2007). sinnewe 39 redesigning the teaching approach in the financial planning capstone at this australian university, students complete a financial planning major as part of their undergraduate business degree. the major consists of financial planning units covering various aspects of financial advice, such as retirement strategies, including superannuation5 , insurance and risk planning, and investment management. this paper focuses on the capstone unit, the final unit of the major, which draws together the various areas covered in previous units. the capstone objective is to build students’ capabilities in: i. collecting and analyzing client data to develop innovative and effective financial planning recommendations; ii. professionally communicating their recommendations to clients in writing as well as verbally; iii. providing ethical advice in the best interest of their clients by working individually and in teams; and iv. identifying their professional strengths, abilities, and values by critical reflection. to achieve these objectives, students are tasked to create a statement of advice (soa) in teams of four using financial planning software for a hypothetical client case study. 6 since the preparation of a soa is complex and adheres to strict regulatory requirements (see s947b, s947c corporations act 2001; australian securities and investments commission [asic] 2021), the task has been broken down into the following assessment items (including the assessment weight towards students’ final result): 1. client interview and client data analysis (20%); 2. preparation of soa (50%); and 3. presentation and personal reflection (30%). in the first assessment item, students individually collect and organize the client data in a fact find and report findings from analyzing the client data in a client file note. students need to summarize key insights from their analysis in an email to the client to obtain their agreement on the scope of the advice and the financial objectives. the second assessment item requires students to work in teams to develop a compliant soa using state-of-the-art financial planning software (xplan) based on their findings from the first assessment item. students can compare their notes from their first assessment to find the most appropriate recommendations for the client. the third assessment item consists of two parts. the first part is a 15-minute presentation. the second part of the final assessment consists of reflecting on their learning, critically evaluating the development of their professional skills, and identifying the values that define their future professional identity. given the centrality of the capstone unit for developing students’ professional capabilities, it was important to identify the barriers and shortcomings students encountered in their learning experience. reviewing student evaluations from previous offerings revealed a wide array of responses to whether they agree that this unit provided them with the opportunity to improve their knowledge and skills. ratings ranged from strongly agree (5) to disagree (1). students' comments such as “client information was unrealistic,” “a little more structure regarding the first two assignments,” “there was no guidance about what should be included,” and “even just explaining the basic expectations that would have been helpful” suggest that the assessment needed to be redesigned, and support was needed in completing these tasks. consequently, a redesign was implemented in 2023. the main innovations of the redesign are presented in table 1. 5 superannuation refers to australia’s mandatory retirement savings system, where employers are required to contribute a legislated percentages of an employee’s earnings to a regulated trust fund that invests on the employee’s behalf until retirement. 6 under the australian corporations act 2001 (s 946), financial advisers must provide a soa to retail clients when proposing financial product recommendations. financial services review, 33(4) 40 table 1: summary of key teaching innovations in the financial planning capstone redesign innovation problem or feedback prompting the change description of the innovation how it was implemented 1. authentic, industryvalidated client case with staged ethical complexity prior case lacked realism and depth due to missing practitioner input a comprehensive case based on practitioner feedback, including staged disclosure midsemester to simulate a change in client circumstances to which students have to respond by considering the impact on financial strategies and client relationship developed draft case → reviewed by program lead + two advisers → integrated revisions. mid-semester “curveball” released with ensuing class discussion on compliance and ethics. 2. practitionermodelled client discovery workshop using studentgenerated questions student feedback indicated a lack of confidence in their capability and a desire for more guidance early in the semester practitioner workshop modelling professional communication and rapport-building. live critique of student developed client questions during the workshop each student posts 2–3 questions on a digital forum → instructor forwards thematically collated pool of questions to advisers → advisers pre-screen questions and provide feedback to help students get a better understanding of how to conduct an effective client meeting. 3. layered scaffolding of digital and professional skills support student evaluations indicated gaps in skills development integrated labs and skills workshop to develop digital competency and professional capabilities, such as ethical reasoning, teamwork and communication. xplan labs (weeks 2 & 9); ethics masterclass (week 7); teamwork workshop (week 5); pitching/presentation skills workshop (week 11); and career advancement class (week 13) 4. redesigned assessment structure with role-based expert pitches and industry judging previous assessment (prerecorded team video) rewarded generic content repetition and lacked individual accountability each student assumes an “expert role” (retirement, risk, investment, etc.) and delivers a live pitch to an industry-judging panel that ‘probes’ each student on their contribution to their team’s soa. students present financial strategies developed in their ‘expert role’ in a 3-minute live pitch → industry judging panel asks probing questions and scores pitches using a structured rubric (appendix a). 5. structured teamwork supports to develop collaborative capability research identified teamwork as a challenging graduate skill to develop often affected by social loafing and coordination issues (de prada et al. 2022; chang & brickman, 2018). teamwork support scaffolds include role assignment, meeting documentation, and peer evaluation to promote accountability, clarify responsibilities, and support team coordination. designated role assignment, including team lead, is implemented through the team contract → teamwork is ‘practiced’ through structured meetings, each recorded using a guided meeting template → meeting records and anonymous peer-evaluations inform individual teamwork marks. note: this table summarises the five (5) key innovations of the capstone redesign, including the specific problem or gap addressed, details about the innovation and how they were implemented. sinnewe 41 the following discussion provides additional details on the teaching innovations presented in table 1. innovation #1 one of the main changes was to improve the authenticity of the hypothetical client case study, which all student teams used as the basis for preparing their financial strategies. 7 a draft of the case study was circulated to the financial planning subject area coordinator, a former practicing financial adviser, and two currently practicing financial planners for feedback. based on their suggestions, the case study was enriched with additional client-specific details to more closely resemble authentic client scenarios. another real-world feature was the release of additional client information midway through the semester to test students’ ability to pivot and adapt to new information.8 innovation #2 in addition, to support student skill development, a series of guest speakers were invited to class. the guest speakers were invited to discuss specific aspects of skills development. for instance, in week 2, each student was asked to submit up to three (3) questions they would ask clients in an initial meeting to a digital discussion forum. the submissions were categorized by the lead educator and discussed at a briefing with the invited guest speakers, two experienced financial planners, who picked a sample from the student question pool to discuss during their workshop in week 3. during the workshop session, students received feedback on the effectiveness of their questions in eliciting information from clients from the practitioners, who also shared their experiences with establishing rapport with clients and building client relationships. innovation #3 other skills development support scaffolds were livestream demonstrations of financial modeling in xplan by an experienced paraplanner. in the lead-up to the assessments, skills development workshops on teamwork, pitching, and career advancement were delivered by senior learning educators. finally, a seminar on ethics in financial planning was delivered by a leading academic 7 although all teams worked on the same client case, the open-ended nature of the task and the breadth of strategy options available in financial planning typically result in considerable variation in the recommendations developed by different teams. expert in the field of business ethics. various templates (e.g., client file note, self-reflection, meeting notes, etc.) with guiding prompts and questions were created to aid students with the completion of the assessments.9 innovation # 4 was the redesign of the student presentations. in the previous offering, students pre-recorded a 15-minute video on zoom presenting their team’s soa. when reviewing the unit, it was found that this part of the assessment regime could be streamlined by cutting down the repetitiveness in presentations as each team member essentially presented on the same financial plan. the aim of updating this assessment piece was also to provide students with the opportunity to receive instant feedback on their presentations by changing the format to a live presentation. the redesigned assessment requirements asked students to present their advice as a team, with each team member “pitching” a specific part of the financial plan within 3 minutes. overall, each team had 15 minutes to present, which gave them a bit of a buffer in case one of the team members exceeded their allocated time limit. presentations were all scheduled for week 12, and teams could present in person or online. the benefit of scheduling all presentations in one workshop was that students could receive feedback on their pitches from a panel of judges consisting of financial planning professionals. for evaluating the student presentations, each judge was tasked to focus on one student per team and provided a scoring sheet to assess the appropriateness and effectiveness of the presented strategies, the delivery style, and the engagement with questions from the audience and the judges.10 innovation #5 to better assess students’ teamwork skills, specific scaffolding for teamwork was developed. as suggested by chang and brickman (2018), students were instructed to sign a team contract at the beginning of the semester. in the team contract, each student agreed to take on an “expert” role within the team, e.g., retirement expert, insurance expert, etc. to support coordination, the team contract also required the nomination 8 the full text of the client scenario is available upon request. 9 teaching resources are available upon request. 10 the scoring sheet is available upon request. financial services review, 33(4) 42 of a team lead to monitor team progress and serve as the main point of contact. this role determined which part of the advice they were responsible for developing throughout the semester. the role assignment also meant that each student presented a different segment of the soa. additionally, students were required to meet at least five (5) times during the semester and for each meeting complete a meeting record using a template with prompts to guide agenda setting, capturing main discussion points and contributions, and creating a post-meeting action register. finally, students filled out an anonymous peer-evaluation form to rate each other’s contributions. the marking rubric was then adapted to assess students’ teamwork capability based on the submitted meeting records. individual marks were adjusted based on the anonymous peer evaluation forms. evaluation of the redesign: initial indicators of impact this section presents emerging evidence of the impact of the teaching innovations, drawing on three sources: academic performance; student evaluations; and feedback received from an expert peer reviewer of educational practice. student performance and evaluation prior to the redesign, the average student scored 69.6 out of 100 points in 2022 (n = 35; sd = 12.6). in 2023, the average score was 78.3 out of 100 (n = 27; sd = 5.7), an improvement of 12.5% over the last time the unit was run. this difference is statistically significant (t-stat = 3.559; df = 48; p<0.001). figure 1 compares the preand post-redesign distribution of final grades.11 figure 1: financial planning capstone yoy results comparison figure 1 shows academic performance in a year-on-year (yoy) comparison. academic performance is measured on a scale from 0 to 100. the results are determined at the end of the semester based on the total across the three assessment items: client data analysis (20%); soa (50%); and presentation and self-reflection (30%). the weighting of the assessment items remained constant over the years. 2022 was the year prior to the implementation of the redesign. 2023 was after the redesign. 11 these results should be interpreted cautiously, as student performance can be influenced by cohort characteristics, prior experience, and other contextual factors beyond the control of the study. nevertheless, because the assessment tasks are explicitly mapped to the unit’s professional and technical learning outcomes, improved grades may offer a limited but meaningful indication that students were better able to demonstrate the required capabilities. sinnewe 43 the university runs a standard student survey each semester to gather data on students’ learning experiences at the unit level. the learning experience is assessed through questions about satisfaction, learner engagement, assessment feedback, and skills development. comparing the evaluations prior to the redesign (2022) with the results postimplementation (2023) reveals the following results: ‘unit satisfaction’ increased from 75% to 86% agree; ‘learner engagement’ remained stable at 100% agree; ‘usefulness of assessment feedback received’ improved from 50% to 86% agree; and ‘opportunities for skills development increased’ from 75% to 86%.12 expert peer review an expert peer review of the financial planning capstone teaching approach was conducted after the redesign. the review was undertaken by a senior fellow of the higher education academy (sfhea) from another faculty, following standard institutional processes. key points from the review are summarized below: • authenticity and industry alignment: the assessment tasks were judged to be highly authentic and closely aligned with the skills, technologies, and expectations of contemporary financial planning practice. the inclusion of a dynamic client scenario introducing new information midsemester was viewed as a valuable mechanism for deepening ethical reasoning and adaptive judgement. • integration of industry expertise: the reviewer highlighted the “seamless integration” of practitioner input throughout the unit, which strengthened the relevance and professional orientation of the learning experience. students gained meaningful learning opportunities and developed communication and teamwork skills that will support their transition to professional practice. • evidence-based redesign: the redesign was perceived as a thoughtful response to student feedback from past offerings and universal issues with skills development in 12 the 2022 response rate was 11.8%. in 2023, the response rate was 25.9%. the standard survey provides unit-level feedback but does not collect item-level data on specific components of the redesign. as a result, students’ perceptions cannot higher education and specifically in financial planning. • structured skills development: specialized workshops (e.g., ethical advice, pitching, xplan modelling) were commended as effective scaffolds supporting students’ capability development. the reviewer also identified areas for future improvement. in particular, greater collaboration with unit coordinators of earlier subjects was suggested as a way to strengthen the development of professional skills leading into the capstone. the reviewer emphasized that a whole-of-degree approach would distribute the skills development more evenly across the degree, enabling repeated practice and greater independence by the capstone stage. overall, the peer review offers credible external support for the effectiveness of the redesign while highlighting opportunities for further refinement reflections on the redesign: student and educator insights this section presents a reflective analysis of the redesign, informed by insights from the educator and students. student reflections from their final assessment were thematically analysed to capture how learners engaged with the redesigned activities.13 the discussion links the pedagogical evidence that informed each innovation with students’ reported learning experiences, offering a consolidated view of how the redesign supported the development of professional skills. innovation #1 authenticity in assessment and learning is required based on the need for a better transition of financial planning students into the profession by enhancing their learning experience at university (goetz et al. 2005; brimble et al. 2012; west et al. 2019). authentic assessment has been associated with increased professional competency (thurab-nkhosi et al. 2018; sewagegn et al. 2020). the re-developed assessment design meets the characteristics of an authentic learning experience (ashfordbe disaggregated by individual learning outcomes or activities. 13 the thematic analysis for this paper was conducted in leximancer. the technical details of the thematic analysis are available upon request. financial services review, 33(4) 44 rowe et al. 2014): in collaborative teams, students develop a soa as their “final product”. there is an opportunity to learn from structured assessment feedback (i.e., marking rubrics) and from real-world experts' suggestions. the final assessment encompasses an element of metacognition in the form of a self-reflective account. deliberate reflections allow students to effectively engage with their learning experience (ribeiro et al., 2019). self-reflective practice helps emerging practitioners shape their “professional identity” (leering, 2014). student evaluations, reflections, and the peer review attested to the authenticity of the assessment. the thematic analysis captured students’ appreciation of the real-world aspect of the assessment. 14 for instance, a student commented: “the insights gained from this experience will profoundly inform my future practice as a professional financial adviser. this scenario, where i had the opportunity to combine the knowledge from my degree into a realworld context, has provided valuable lessons that will shape my approach to client service and financial planning.” innovation #2 the teaching approach adopted for the financial planning capstone strongly emphasizes collaboration with industry experts. feedback from practicing financial planners on students’ client interview questions enriched the curriculum by providing real-world insights, ensuring students' learning is relevant and applicable to professional contexts. the additional “fireside chat”-style panel talk is a highly resource-effective experiential learning activity that exposes students to real-world challenges financial planners face and experiencing face-to-face conversations with senior leaders in the field (voss and blackburne, 2019). a student with professional experience noted that these insights gave them a better understanding of managing a client discussion and factors that can impact client behavior: “in the capstone unit itself, weeks 3 and 4 are good examples of this, where a focus is made from presenters on understanding their 14 twelve (12) students independently mentioned the concept “real-world” in connection to “case study”, approach to managing a client discussion and giving consideration to behavioural outcomes that can impact client behaviour. my experience within the industry has shown me that often, the technically best client strategy is not necessarily the right strategy for that client if behavioural traits for the client are not aligned to that outcome.” innovation #3 vygotsky’s (1978) zone of proximal development suggests that learners can accomplish more with guidance and support from others. in the prior offering, skill development scaffolding was lacking, and students were mainly left to their own devices when it came to learning how to work in teams, communicate professionally, and present to clients. to fill this gap, various instructional and resource support scaffolds were implemented (see section 2). in reflection, the scaffolding seems to have been effective. the student evaluations suggest improved skills development (see section 3.1). however, the complexity of the financial planning technology (xplan) remained a challenge. one student reflected: “the main challenge for me was xplan. i was confident in the superannuation strategies that i was recommending within the plan, but i was undermined by my lack of proficiency within xplan to communicate and reinforce my recommendations especially through the utilisation of the tools within xplan”. this reflection was also supported by student comments identifying similar struggles in the student feedback survey. in hindsight, the first workshop could have been more engaging. the first industry-led workshop was fast-paced and very process-oriented, with little insight into how the output from the software would be used in practice. in contrast, the second workshop was modeled around a client case. students were much more engaged as they could see the practical relevance of the xplan tool. based on these observations, the introductory workshop “scenario”, “context”, or “setting” in their reflections. sinnewe 45 may need to be revised to provide a better learning experience. innovation#4 communication is consistently identified as one of the most critical capabilities for graduate financial planners, yet higher education providers could do more to develop this skill (west et al. 2019; grable & goetz 2017). research in higher education shows that structured engagement with industry experts can strengthen student learning and contribute to the formation of their emerging professional identity (ashton 2009; riley et al. 2021). in line with this evidence, the redesigned capstone incorporated structured, practitioner-led feedback through the judged live presentations to strengthen students’ ability to articulate the rationale behind strategies and manage client conversations. as illustrated in the quote below, students found the feedback received from the financial planning judging panel valuable, identifying “communication” as a critical skill that this unit helped them to develop. “understanding the "why" behind financial planning decisions has been a game-changer. it's not just about executing tasks; it's about comprehending the underlying rationale and being able to effectively communicate it to clients, fostering trust and credibility.” innovation #5 teamwork is one of the most sought after graduate capabilities in business education, yet the most challenging for students to develop (jackson 2010). reflections from this unit align with this broader pattern, with many students identifying teamwork as an area of difficulty. the literature consistently reports that these challenges are associated with social loafing and poor coordination (hall & buzwell 2013; chang & brickman 2018). in response, the redesigned capstone incorporated explicit teamwork instructional and resource scaffolds (see table 1) as suggested by simons and klein (2007). these supports were intended to provide structure while enabling students to experience the realities of collaborative professional work. student reflections strongly supported the value of this approach: “the assignment’s predefined roles were a welcome change, and while defined, still called for considerable teamwork given the interconnected nature of an soa. it was evident that some sections required other to be completed, which emphasised the need for seamless team cooperation.” together, student and educator reflections affirm that the redesign meaningfully strengthened the authenticity, relevance, and pedagogical coherence of the capstone experience. they also illuminate areas for further improvement, especially in supporting students’ digital capabilities. these insights provide a valuable roadmap for future iterations of the unit and reinforce the importance of sustained, evidence-based enhancement in preparing students for the realities of professional financial planning practice. conclusion the redesign of the financial planning capstone unit demonstrated how authentic learning and targeted professional skills development can enhance students’ job readiness within an academic setting. by integrating realistic client scenarios, structured collaboration with industry partners, and support scaffolds for key capabilities, the revised unit offered students a more supported and professionally aligned learning experience. evidence from student reflections, an independent expert peer review, and other impact indicators suggests that the redesign improved students’ understanding of the financial planning process and professional capabilities such as communication and teamwork. despite these promising outcomes, ongoing refinement is required. student evaluations highlighted persistent challenges in developing digital capability, particularly in navigating financial planning software. future iterations of the unit will explore additional supports, such as engaging an xplan-proficient co-facilitator and further embedding digital literacy development across the semester. the expert peer review also emphasized the need for a stronger whole-of-degree approach to capability development by mapping where key professional skills are introduced, practiced and assessed across the curriculum. to support broader cross-institutional curriculum alignment, the establishment of a financial planning pedagogy community of practice (cop) is proposed. cops, groups that share expertise through ongoing dialogue and collaborative problem-solving, have been financial services review, 33(4) 46 shown to enhance teaching quality and student outcomes (wenger & synder 2000; vescio et al 2008). nesting such a cop within existing education networks (e.g., the academy of financial services) offers a sustainable mechanism for sharing innovations, coordinating curriculum development, and building collective capability among financial planning educators. beyond its institution-specific context, this redesign offers a transferable model for financial planning educators seeking to integrate authentic client scenarios, practitioner engagement, and structured capability scaffolding within their own curricula. by illustrating how industry expertise can be incorporated in sustained and meaningful ways, the study also highlights opportunities for practitioners to shape the professional readiness of emerging advisers. these insights contribute to ongoing sector-wide conversations about strengthening the quality and relevance of financial planning education. overall, this study provides insights into how universities can design authentic, industryaligned capstone experiences that meaningfully support students’ transition into the financial planning profession. the findings underscore the value of practitioner engagement, targeted scaffolding, and collaborative curriculum design in preparing graduates to meet the evolving expectations of the advice industry. references ashford-rowe, k., herrington, j. & brown, c. (2014). establishing the critical elements that determine authentic assessment. assessment & evaluation in higher education 39(2), 205-222. ashton, d. (2009). making it professionally: student identity and industry professionals in higher education. journal of education and work 22(4), 283-300. australian securities & investments commission [asic] (2021) rg 175 licensing: financial product advisers – conduct and disclosure, https://asic.gov.au/regulatoryresources/find-a-document/regulatoryguides/rg-175-licensing-financialproduct-advisers-conduct-anddisclosure/ viewed 3 december 2025. brimble, m., cameron, c., freudenberg, b., fraser, c. & macdonald, k. (2012). collaborating with industry to enhance financial planning and accounting education. australasian accounting business and finance journal 6(4), 7993. chang, y. & brickman, p. (2018). when group work doesn’t work: insights from students. cbe life sciences education 17(3), ar52, https://doi.org/10.1187/cbe.17-090199. corporations act 2001 (cth). corporations amendment (professional standards of financial advisers) act 2017 (cth). corporations (relevant providers degrees, qualifications and courses standard) determination 2021 (cth). de prada, e., mareque, m. & pino-juste, m. (2022). teamwork skills in higher education: is university training contributing to their mastery? psicologia: reflexao e critica 35, 5. goetz, j.w., tombs, j.w. & hampton, v.l. (2005). easing college students' transition into the financial planning profession. financial services review 14(3), 231. grable, j.e. & goetz, j.w. (2017). communication essentials for financial planners: strategies and techniques. john wiley & sons. hall, d. & buzwell, s. (2013). the problem of free-riding in group projects: looking beyond social loafing as reason for noncontribution. active learning in higher education 14(1), 37-49. jackson, d., 2010. an international profile of industry-relevant competencies and skill gaps in modern graduates. international journal of management education (oxford brookes university 8(3), 29-58. johnson, d., brimble, m. & zanetti, r. (2016). industry demand for financial planning graduates. financial planning research journal 2(2),106-124. https://asic.gov.au/regulatory-resources/find-a-document/regulatory-guides/rg-175-licensing-financial-product-advisers-conduct-and-disclosure/ https://asic.gov.au/regulatory-resources/find-a-document/regulatory-guides/rg-175-licensing-financial-product-advisers-conduct-and-disclosure/ https://asic.gov.au/regulatory-resources/find-a-document/regulatory-guides/rg-175-licensing-financial-product-advisers-conduct-and-disclosure/ https://asic.gov.au/regulatory-resources/find-a-document/regulatory-guides/rg-175-licensing-financial-product-advisers-conduct-and-disclosure/ https://asic.gov.au/regulatory-resources/find-a-document/regulatory-guides/rg-175-licensing-financial-product-advisers-conduct-and-disclosure/ https://doi.org/10.1187/cbe.17-09-0199 https://doi.org/10.1187/cbe.17-09-0199 sinnewe 47 kaider, f., hains-wesson, r. & young, k. (2017). practical typology of authentic work-integrated learning activities and assessments. asia-pacific journal of cooperative education 18(2), 153-165. leering, m. (2014). conceptualizing reflective practice for legal professionals, journal of law and social policy 23, 83-106. parliamentary joint committee on corporations and financial services. (2009). inquiry into financial products and services in australia. report november 2009, parliament of australia, https://www.aph.gov.au/parliamentary _business/committees/joint/corporati ons_and_financial_services/complete d_inquiries/2008-10/fps/report/index. power, b. (2017). best tech, it’s no distraction. professional planner 102, 28-30. ribeiro, l.m.c., mamede, s., de brito, e.m., moura, a.s., de faria, r.m.d. & schmidt, h.g. (2019). effects of deliberate reflection on students’ engagement in learning and learning outcomes. medical education 53(4), 390-397. riley, j., eldridge, s. & venkatesh, r. (2021). engaging the experts to embrace the future: an interactive professional learning experience. issues in accounting education 36(4), 41-58. sewagegn, a.a. & diale, b.m. (2020). authentic assessment as a tool to enhance student learning in a higher education institution: implication for student competency. in: railean, e.a. assessment, testing, and measurement strategies in global higher education, igi global, 256-271. simons, k.d. & klein, j.d. (2007). the impact of scaffolding and student achievement levels in a problembased learning environment. instructional science 35, 41-72. smith, c. (2012). evaluating the quality of work-integrated learning curricula: a comprehensive framework. higher education research & development 31(2), 247-262. stein, s.j., isaacs, g. & andrews, t. (2004). incorporating authentic learning experiences within a university course. studies in higher education 29(2), 239-258. thurab-nkhosi, d., williams, g. & masonroberts, m. (2018). achieving confidence in competencies through authentic assessment. journal of management development 37(8), 652662. vescio, v., ross, d. & adams, a. (2008). a review of research on the impact of professional learning communities on teaching practice and student learning. teaching and teacher education 24(1), 80-91. voss, h. & blackburne, g. (2019). bringing experiential learning into the classroom:‘fireside talks’. in: gonzalez-perez, m.a., lynden, k. & taras, v. the palgrave handbook of learning and teaching international business and management, palgrave macmillan, 459-473. vygotsky, l.s. & cole, m. (1987). mind in society: the development of higher psychological processes, harvard university press. wenger, e.c & snyder, w.m. (2000). communities of practice: the organizational frontier, harvard business review 78(1), 139-146. west, t., johnson, d. & webb, a. (2019). career outcomes of financial planning students. financial planning research journal 5(1), 82-110. https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_and_financial_services/completed_inquiries/2008-10/fps/report/index https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_and_financial_services/completed_inquiries/2008-10/fps/report/index https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_and_financial_services/completed_inquiries/2008-10/fps/report/index https://www.aph.gov.au/parliamentary_business/committees/joint/corporations_and_financial_services/completed_inquiries/2008-10/fps/report/index financial services review, 33(2) i volume 33 issue 2 from the editor it is with great pleasure and enthusiasm that i welcome dr. elisabeth sinnewe as the incoming editor of financial services review (fsr). with a strong background in personal finance and superannuation research, an impressive record of publishing in top-tier journals, and proven leadership within the financial planning academic community, dr. sinnewe brings a wealth of experience and vision that aligns beautifully with fsr’s mission to advance rigorous and practitioner-relevant research. as chair of the academy of financial services australia–new zealand chapter, dr. sinnewe has demonstrated a clear commitment to fostering academic excellence and industry collaboration—organizing impactful symposiums, establishing research excellence awards, and facilitating expert panels on emerging issues in financial planning. dr. sinnewe’s editorial experience and dedication to expanding fsr’s reach and engagement make her uniquely suited to guide fsr into the future. i am particularly excited about dr. sinnewe’s plans to elevate research quality, broaden interdisciplinary connections, and explore innovative dissemination strategies that will enhance the journal’s impact and visibility. i am also pleased to present the papers in this issue of fsr. the first paper (psychophysiological finance and intelligent wellness: a new financial planning practice model) was written by robert hanlon, paul leher, alexander cohen, eric miller, monte hancock, and robert mitchell. in this invited paper, the authors point out that cfp® board requires certificants to recognize and address client attitudes, behaviors, and stress-related factors that affect financial decisions and well-being. using this as a framework, this paper introduces a new advice-delivery model that integrates psychophysiology, mobile health, and psychology with traditional financial planning to help reduce stress and enhance client outcomes. the second paper (consumers' basic bank account complaints and their financial hardships: a content analysis of complaints filed with the consumer financial protection bureau (cfpb)), written by julie birkenmaier and hope stratman, analyzes consumer complaints to the cfpb to identify key banking issues—such as fraud, atm malfunctions, and poor customer service—that increase the likelihood of financial hardship. findings suggest financial institutions can mitigate these issues through targeted policy changes and improved customer service practices. the third paper (exploring the effect of federal student loan payment resumption on borrowers through sentiment and textual analysis using x) was written by jason n. anderson, donovan sanchez, juan e. gallardo, derek lawson, and congrong ouyang. the authors analyzed x (formerly twitter) data. they found negative sentiment around student loans surged during the october 2023 payment restart, with 46% of mentions being negative and only 1% positive. the results suggest financial planners should be prepared to support borrowers emotionally in addition to offering financial guidance. the next paper in this issue (consumer margin use: understanding the role of peer influence, investment literacy, and age), written by kaplan sanders and olamide olajide, used data from the 2021 national financial capability study to examine household margin use, finding that peer influence, youth, and higher investment literacy are associated with a greater likelihood of buying on margin. the results offer insights for policymakers and financial advisors to understand consumer borrowing behaviors. financial services review, 33(2) ii stuart heckman, jodi letkiewicz, and hanna lim wrote the fifth paper (student willingness to borrow for higher education). this research team applied a human capital model to analyze college students' willingness to borrow for education, using data from the 2020 study on collegiate financial wellness. findings show that borrowing decisions align with rational human capital theory, with higher tuition, career goals, and expected earnings increasing borrowing willingness, while higher income and alternative financial support reduce it. efthymia antonoudi, genti kostandini, and hanna lim authored the next paper (immigration law enforcement and immigrant homeownership). they used american community survey data and a difference-in-differences model to determine that 287(g) immigration enforcement agreements significantly reduce homeownership rates, especially among less-educated, u.s.-born hispanics in states without everify laws. while the secure communities program shows mixed effects, the overall results emphasize how immigration policies can unintentionally affect housing stability and economic integration. the sixth paper in this issue (the association of cryptocurrency and the use of alternative financial services), written by david smith and gary curnutt, used data from the 2023 survey of household economics and decisionmaking to explore how cryptocurrency fits into the broader landscape of alternative financial services (afs). the paper examines whether individuals who use cryptocurrency for afs purposes—such as payments or sending money—also use it for investing or other afs-related activities. the next paper (examining the gender gap in participation in employer-sponsored retirement plans: oaxaca decomposition) was written by ferdous ahmmed, charlene marie kalenkoski, and christopher m. browning. these researchers used 2021 nfcs data to analyze gender differences in participation in employer-sponsored retirement plans and the role of financial literacy. findings show that women not only have lower financial literacy on average but also receive less benefit (or "return") from that literacy compared to men, contributing to a persistent participation gap. this issue concludes with a paper (a structured literature review on equity in the financial services profession: unpacking gender barriers and advancing women's participation globally) written by tanya staples, ashlyn rollins-koons, and megan mccoy. as noted by these authors, as women increasingly control a larger share of global wealth—projected to reach 55% by 2030—the financial services profession is shifting toward more holistic, advice-driven approaches. however, persistent barriers limit women's representation in financial advising, prompting a call for structural reforms, gender-conscious policies, and inclusive practices to better serve diverse clients and support female professionals. this paper provides a pathway for policymakers and certification bodies to attract more women into the profession. finally, on a personal note, i want to say thank you to everyone who has supported the academy of financial services and financial services review over the past several years. as my last official editorial act, i would like to encourage everyone to congratulate and support dr. elisabeth sinnewe as she steps into the editor role. i step down knowing that the future of financial services review is in excellent hands. john e. grable, ph.d., cfp® pii: s1057-0810(00)00060-3 from the editor can an individual succeed in timing the market? there has been a significant amount of interest by both academicians and practitioners in the subject of using trading rules to earn an abnormal return. the first article in this issue by richard chung and lawrence kryzanowski is entitled, “market timing using strategists’ and analysts’ forecasts of s&p 500 earnings per share”. they provide an excellent review of the literature on market timing and test the usefulness of earnings forecasts as trading rules. anthony l. loviscek and w. john jordan’s article, “stock selection based on morningstar’s ten-year, five-star general equity mutual funds” continues the theme of timing the market. they select the top stocks from morningstar’s highest rated mutual funds to form portfolios and measure the performance of them against the market as measured by the s&p 500. they find that the large cap stocks that are selected do not outperform the market. very unfortunately, this article is being published posthumously for dr. jordan whose entire family was killed in an helicopter accident in hawaii. risk tolerance is assumed to have an impact on portfolio allocations for investors at any age, based on recent articles published in this journal. govind hariharan, kenneth s. chapman, and dale l. domian’s article called “risk tolerance and asset allocation for investors nearing retirement” explores the impact of age and approaching retirement on individual choices. they suggest that individuals do not alter their portfolios based on risk tolerance and do decrease their dependence on treasury bills. john e. richard and james b. wiggins’ article entitled, “the information content of closed-end country fund discounts” tests two conflicting theories that seek to explain the variations in closed-end fund premiums across funds and time. they use foreign country funds from the prospective of the united states investor for their tests. the last article by john c. bost and tony cherin seeks to change the provisions of an estate in an article they title, “liquidating a remainder interest: simplifying personal finance”. they walk the reader through the steps necessary to dissolving a trust so that the income beneficiaries can manage the remaining assets themselves. all of these articles examine issues of individual financial management, which is the focus of this journal. as vickie l. bajtelsmit’s special issue on retirement income (volume 9, number 1) indicates, individuals are being given more and more responsibility for their financial futures. thefinancial services reviewencourages empirical research that provides information to those who are interested in exploring and testing strategies and tactics that will help the individual prosper in unknown environment that is likely to continuously change. karen eilers lahey financial services review 9 (2000) v 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00060-3 pii: s1057-0810(99)00036-0 does retirement planning affect the level of retirement satisfaction? harold w. elder, patricia m. rudolph* dept. of finance & economics, university of alabama, tuscaloosa, al 35487-0224, usa abstract this paper analyzes the relationship between retirement planning and retirement satisfaction. do individuals think about and plan for retirement? if they do, do they utilize financial planning services? if they plan, are they more satisfied with retirement than those who did not? data for 1,781 retired individuals from the first wave of the health and retirement study (hrs) are analyzed using an ordered probit model. the results indicate that thinking about retirement and attending planning meetings have a significant positive impact on satisfaction even when income, wealth, marital status and health are included as explanatory variables. © 1999 elsevier science inc. all rights reserved. 1. introduction as the united states population continues to age, retirement will come to the forefront as a public policy question as well as emerging as one of the primary financial planning issues. in this paper, the relationship between an individual’s actions to plan for retirement and the level of retirement satisfaction is analyzed. do individuals think about and plan for retirement? if they do, do they utilize financial planning services? are those who plan more satisfied in retirement than those who did not? our paper contributes to the retirement literature by utilizing the economic models of saving and consumption to analyze the factors that affect the level of retirement satisfaction. as will be seen in the literature review, studies dealing specifically with retirement satisfaction have generally been limited to empirical papers by demographers, gerontologists and * corresponding author. tel.:11-205-348-8966; fax:11-205-348-0590. e-mail address:prudolph@cba.ua.edu (p. rudolph) financial services review 8 (1999) 117–127 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(99)00036-0 sociologists that study statistical relationships using specialized, non-representative samples. our paper examines the factors that affect retirement satisfaction within the context of the life-cycle model of consumption behavior and uses a representative sample of older americans drawn from the first wave of the health and retirement study. section 2 examines earlier work related to savings, retirement decisions and retirement satisfaction by researchers in a variety of disciplines. the third section looks at a model of retirement satisfaction, whereas the fourth describes the data used to estimate the model. the results of the statistical analysis are contained in section 5. section 6 summarizes and concludes the study as well as raising some questions for future research. 2. literature review research on retirement has been produced from a variety of perspectives. each discipline provides a different viewpoint to the discussion of the issues surrounding retirement, as well as bringing to bear a different set of tools to analyze retirement. economic research has focused primarily on public policy questions rather than examining these decisions from a personal planning perspective. nevertheless, the economic models of savings and consumption provide a theoretical framework in which to analyze the behavior of individuals. psychological and sociological studies look at the factors that affect retirement satisfaction but do not provide any theoretical structure for these relationships. as will be noted below, most of these papers utilize specialized samples in their empirical tests. in this paper, we will use the theoretical structure provided by the economists to motivate and explain the factors that affect retirement satisfaction. economic analysis has not directly addressed the factors that affect the level of retirement satisfaction; rather, it has considered related issues, including the adequacy of aggregate savings. poterba (1996) and hurd (1990) provide good overviews of the economics literature on savings decisions and retirement. as pointed out by poterba (1996, p. 130), “the life-cycle model/permanent income hypothesis has been the dominant economic model for analyzing saving behavior.” in these models, based on the seminal works by ando and modigilani (1963) and friedman (1957), individuals make decisions over their lifetimes to consume and shift consumption through saving in a way that will maximize lifetime utility. recent work by bernheim, skinner and weinberg (1997) and banks, blundell and turner (1998) has noted that there is often a drop—sometimes substantial—in household consumption after retirement. life-cycle models imply that this observation could emerge from individual preferences: those who save less for retirement (and thus are able to consume less in that phase) simply have a higher rate of time preference than those who choose to save more for their retirement. in other words, differences in savings and retirement consumption are attributable to differences in tastes. an alternative hypothesis is that individuals may misperceive what their retirement period wealth will be, and are “surprised” by its (low) value when they do retire. consequently, surprised households must align their consumption with the real value of their retirement wealth. although these studies do not address the individual’s level of retirement satisfaction, it seems likely that the “surprised” individuals with fewer than expected financial resources are less satisfied with their lives than those whose expectations about their wealth position are more consistent with reality. 118 h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 in one of the few financial planning papers that deal specifically with retirement satisfaction, brunson, snow and gustafson (1998) focus on mid-life career changes in their model of retirement satisfaction of career and non-career military personnel. for this specialized sample, they find that their adequacy of financial planning measure has a significant positive impact on the level of satisfaction in retirement. psychological and sociological studies have directly addressed the question of retirement satisfaction, and indicate that a broad range of factors is associated with retirement. the individual’s financial resources, retirement preparation and planning, perceived health, participation in leisure activities, occupation before retirement, relationships with family and friends, and the reason for retirement are all important factors in this research. sterns and gray (1999) provide a broad view of the gerontology research on retirement issues. focusing specifically on the issue of retirement satisfaction, brunson (1996), cooper (1993) and cope (1990) provide reviews of earlier work. recently, dorfman (1989), knesek (1992), floyd et al. (1992), reis and gold (1993), macewen et al. (1995), and gall, evans and johnson (1997) all find that retirement planning have a positive impact on actual or anticipated retirement satisfaction. the consistency of the findings provides strong evidence of a relationship; however, the samples used limit the ability to generalize based on these results. for example, dorfman’s (1989) sample consists of 252 men and 199 women from two rural counties in iowa. knesek’s (1992) sample contains 198 males and only 20 females that are employees at a large manufacturing facility in central indiana. the sample used by floyd et al. (1992) contains only 126 subjects from the midwest, whereas the sample used in macewen et al. (1995) consists of 216 canadians. part of our contribution to this literature is to use a larger, more representative sample. 3. the model of retirement satisfaction the life-cycle model of consumption and savings behavior provides a useful point of departure to think about retirement satisfaction. in this model, individuals maximize utility or satisfaction over the entire course of their lifetimes. decisions about savings made early in the life-cycle help determine the resources that are available over later, retirement years. the life-cycle model implies that a retired individual’s utility is a function of her ability to consume goods and services reflected in her current income and accumulated net worth. as noted in the literature review, economists tend to study the role of consumption whereas sociologists, gerontologists and psychologists suggest that other personal factors are also likely to be important. marital status, health status, level of education, whether the individual was forced to retire, and pre-retirement occupation as well as the retirement planning should have an impact on the level of retirement satisfaction. if some individuals plan more than others and make conscious decisions concerning their retirement, it is reasonable to expect that these individuals are more likely to achieve a higher level of satisfaction than those who do not plan. another way of expressing this is that those who plan are less likely to be in the “surprise group” who have to make significant (downward) adjustments to their consumption pattern upon retirement. 119h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 to determine, at least tentatively, whether this relationship is present in the hrs data (that is described in more detail below), we provide a cross tabulation of the levels of satisfaction in retirement with some measures of retirement planning. these relationships are shown in table 1. the frequencies shown there reveals that individuals who plan—as indicated by their thinking about retirement and attendance of retirement planning sessions—are more likely to be very satisfied in retirement than those who did not think about retirement and attend planning meetings. this tabulated information is obviously not conclusive; nevertheless, there does seem to be a link between planning and satisfaction. of those who had thought “a lot” about retirement, 69.1% were “very satisfied” compared with only 24.8% “very satisfied” respondents who had thought “hardly at all.” similarly, only 4% of those who thought “a lot” were “not at all satisfied” compared with 37.3% of those who had thought “hardly at all.” of those who attend meetings, 65% are “very satisfied” compared with only 36.7% of those that did not attend meetings. whether the planning-satisfaction relationship changes in a multivariate setting is the subject of the statistical analysis that follows. other factors that may be strongly influencing the level of satisfaction that a retiree experiences include economic factors, such as income and accumulated wealth, as well as non-economic factors such as health status, existing familial relationships and the reasons for retirement. 4. the hrs data and methodology the hrs was specifically designed by an interdisciplinary panel to gather a broad range of information pertinent to the retirement decision. (the national institute on aging (nia) is the sponsoring organization for the hrs and the data are collected by the institute for social research at the university of michigan. juster and suzman (1995) provide a useful introduction to the survey in a supplemental issue of thejournal of human resources devoted to the hrs. current information about the data collection process and the data itself can be downloaded from the institute for social research web site, http://www.umich.edu/ ;hrswww/.) the survey includes questions on retirement planning, net worth, income and employment history as well as health status and familial relationships. alternative data sets such as the panel study of income dynamics and the federal reserve’s surveys of consumer finances were designed to focus on households’ economic decisions, and accordingly, do not include questions about many important non-economic factors. table 1 frequencies for planning and statisfaction variables thought about retirement attended meetings totals a lot some a little hardly at all yes no very satisfied 69.1% 50.0% 41.7% 24.8% 65.0% 36.7% 782 moderately satisfied 26.9% 41.8% 44.8% 37.8% 28.6% 38.4% 640 not at all 4.0% 8.2% 13.5% 37.3% 6.4% 24.9% 359 totals 525 316 163 777 454 1327 1781 120 h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 in this paper, only the first wave or year of the hrs data is used, so the analysis cannot address questions about the long-term impact of retirement planning on individual satisfaction. nevertheless, a significant number of individuals in the survey have retired, and have had time to draw (at least some) judgment about their level of personal satisfaction. the first wave surveys were collected in march 1992. to be age-eligible at that time a person had to be between 51 and 61 years of age. partners were also interviewed so that some of the respondents are not “age eligible.” from the total 12,652 total respondents, 1,804 answered that they were completely retired. of those, 1,781 answered the questions concerning the level of satisfaction with retirement and are included in our analysis. questions concerning the level of retirement satisfaction are answered by everyone and, thus, reflect only the respondent’s perception even if the respondent is one member of a household. the net worth and income variables are collected on a household level, so that the income and wealth of both the respondent and partner (if there is one) is included. income from all sources includes pensions and social security, investment income and welfare payments along with wages and salaries. net worth includes both financial assets and housing equity. the respondent’s race is coded as one if white and zero otherwise. “married” is used to designate those who indicate they have partners of the same or opposite sex. the education level is measured in years of education. health status is a self-reported scale. more complete definitions of these variables are presented in table 2, as well as descriptive statistics for these measures. because the focus of this study is the impact of planning on retirement satisfaction, a closer look at questions on these subjects may be instructive. we can think of a possible sequence of steps that a respondent might follow in attempting to deal with the decisions concerning retirement. first, did this individual anticipate the coming of his or her retirement? these factors are addressed in the survey by a question that asks whether the respondent had thought about retirement, and the response indicates the extent to which the person had thought about this question: a lot, some, a little or hardly at all. next, the survey considers whether the individual had taken steps to prepare for retirement, by some formal planning process. this is considered in the hrs in a question that inquires as to whether the respondent had attended meetings about retirement planning (yes or no). obviously, thinking about retirement is not synonymous with planning for retirement, nor is attending meetings on retirement planning the only way to get professional advice. nevertheless, it is difficult to see how you could plan for retirement without thinking about it and attending meetings is a low cost means to obtain professional advice. the “thought about” and “attended meetings” responses should have a positive impact on retirement satisfaction if these variables are correlated with the process of making reasonable decisions. finally, the respondents who have retired are asked as to their level of satisfaction with their life in retirement: very satisfied, moderately satisfied or not at all satisfied with retirement. 5. an ordered probit model of retirement satisfaction as can be seen in the manner that the question about retirement satisfaction is posed, respondents provide information about their level of satisfaction based upon an ordinal 121h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 ranking of the choices, choosing the selection that most closely corresponds to their true level of satisfaction. obviously there is no (cardinal) measure to gauge actual satisfaction, but a regression approach that can provide useful insights about this type of question is the ordered probit model (see zavoina and mcelvey, 1975). this model is a latent regression procedure table 2 variable definitions and descriptive statistics variable name variable definition descriptive statistics retsatis coded response to the question, all in all, would you say that your retirement has turned out to be: 1.237 not at all satisfied 0 (.764) moderately satisfied 1 very satisfied 2 retthink before you retired, how much had you thought about retirement? hardly at all 0 a little 1 1.331 some 2 (1.300) a lot 3 retplan binary variable that takes on a value of 1 if the respondent attended retirement planning meetings and zero otherwise. .2549 (.436) retfor binary variable that takes on a value of 1 if the respondent replied he/ she was ‘‘forced into’’ or ‘‘part wanted, part forced’’ into retirement and zero otherwise. .4756 (.4995) hlth this is a binary variable based upon the coded response to the question: i’m going to read you a list of reasons why some people retire. please tell me whether, for you, poor health was not at all important 0 .4677 somewhat important 1 (.4991) moderately important 2 very important 3 if the response was 1, 2, or 3 then the variable equals 1; if 0 was the response then the variable is set to 0. hhinc total household income from all sources including earned income, investment income, pensions and transfers payments. $36482 (31653) totnw total household net worth includes both home equity and non-home equity. non-home equity is the sum of the household’s holdings of financial assets such as cash, stocks, bonds, cd’s, ira’s and keogh accounts as well as the value of real estate held for investment purposes. it does not include an imputed value of defined benefit pension funds or the assets in pension accounts such as 401(k)’s. nor does it include an imputed value for the claim on social security benefits. $229815 (426239) white binary variable that takes on a value of one if the respondent is white and zero otherwise. .7198 (.4492) male binary variable that takes on a value of one if the respondent is male and zero otherwise. .6036 (.4892) married binary variable that takes on a value of one if the respondent has a partner and zero otherwise. ‘‘partner’’ refers to a spouse or live-in companion of the same or opposite sex. .8147 (.3886) edyrs number of years of education completed. 11.58 (3.28) 122 h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 that assumes that an underlying measure of satisfaction exists, but that its value cannot be observed. thus, the model that is estimated is y* 5 b9x 1 e, (1) wherey* is an unobserved dependent variable;b9 is a coefficient vector,x is a vector of independent variables ande is a stochastic error term. instead, we observe values ofy, that correspond to this person’s level of satisfaction, with values starting at 0 (not at all satisfied) and increasing by units of one as satisfaction increases. these responses correspond to a set of parameters, usually labeledm’s, that partition the distribution ofy*. the estimation procedure thus determines the probability that the valuey* falls into a range of themi’s as established by the observed values ofy (in this case, the responses to the questions of retirement satisfaction). this model assumes thate is normally distributed and the mean and variance ofe are normalized to zero and one, respectively. 5. statistical findings the estimates of the ordered probit model of retirement satisfaction are found in table 3, with the results of the basic model shown in column (a), and then additional estimates that incorporate the interrelationship between planning and forced retirement are shown in columns (b) and (c). the central finding from these estimates is that planning for retirement, as measured by how much the respondent thought about retirement and whether the person attended retirement planning meetings, is positively related to the level of retirement satisfaction. this is true independent of household characteristics or economic status. looking at the other variables in the equation, the results are generally consistent with the findings of prior studies. individuals in households with higher incomes and larger net worth are, predictably, more likely to be satisfied. respondents who have partners (married or otherwise) are also significantly more likely to be satisfied. other characteristics are unrelated to the level of satisfaction: males are no more likely than females to be satisfied; whites are no more likely than non-whites to be satisfied; and those with more education do not have greater levels of satisfaction than those with less. the lack of significance for the education variable may be due to the correlation between education and other significant factors such as income and health. some factors do reduce the likelihood of satisfaction. in particular, those individuals who were forced to retire either by poor health or by their employer are less likely to be satisfied. in some cases, the lack of satisfaction for those forced to retire may be related to planning—or the inability to plan—and could be connected to the ability to anticipate the timing of retirement, and make adequate preparations for it. to investigate these questions, additional estimates have been produced and can be found in columns (b) and (c) in table 3. these estimates include interaction terms between the planning variables (retthnk and retplan) with the retirement cause variables: health condition (hlth) and forced retirement (retfor) variables. given the results from these estimates, it is clear that individuals who had the opportunity to plan in the face of either poor health or a forced retirement are more likely to be more 123h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 satisfied. it is interesting to note that only the “thinking about” variable is significant (and the “attended meetings” variable is not) in these estimates. the coefficients on both the original planning and retirement cause variables all yield the same implications and statistical significance as when the interaction terms are not included. all this simply underscores the impact of preparation on the retiree’s level of satisfaction. one problem that could arise in this investigation is the impact self-selection with respect table 3 ordered probit estimates of retirement statisfaction (n 5 1781)a basic model (a) impact of involuntary retirement employer forced (b) health forced (c) constant 1.00 1.10 1.13 (7.037) (7.337) (7.534) male 2.025 2.023 2.020 (0.410) (0.062) (0.324) white .043 .0435 .045 (0.625) (0.069) (0.647) married .277 .275 .267 (3.475) (3.447) (3.359) *** *** *** totnw .289xe-06 .280xe-06 .279xe-06 (2.802) (2.700) (2.717) *** *** *** hhinc .274xe-05 .257xe-05 .264xe-05 (2.389) (2.221) (2.305) ** ** ** edyrs .010 .011 .011 (0.977) (1.116) (1.051) retthnk .186 .117 .112 (6.972) (3.207) (3.105) *** *** *** retplan .204 .269 .216 (2.646) (2.785) (2.300) *** *** ** hlth 2.591 2.584 2.786 (7.543) (7.462) (7.961) *** *** *** retfor 2.764 2.920 2.759 (9.729) (9.039) (9.714) *** *** *** retthnk interaction — .146 .156 (2.769) (3.136) *** *** retplan interaction — 2.145 2.019 (0.947) (0.122) m1 1.425 1.432 1.435 (27.710) (27.757) (27.885) log likelihood ratio (x2) 876.2 884.3 886.5 a asymptotic t-statistics in parentheses; ***, **, *, indicate statistical significance at the 1%, 5%, and 10% levels, respectively.) the two interactions (retthnk and retplan) are with employer-forced retirement in column (b) and health-forced retirement in column (c). 124 h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 to planning on the level of retirement satisfaction. in this case, self-selection would imply that the individuals who plan are more likely to be satisfied. this could be a very real problem, and the parameter estimates for the planning variables in satisfaction equation, as found in table 3, would overestimate the effects of planning on satisfaction. investigations of this, including estimates of a selection corrected ordered probit model and a modified version of the satisfaction estimates, indicate little or no evidence of self-selection. in the corrected model, a binary measure of the level of satisfaction and a similar measure of either the “thinking” or the “going to meetings” variables are used. to aid the understanding of these relationships, table 4 presents separate estimates of ordered probit models of the “thinking about retirement” variable and probit estimates of the “attended meetings” variable. 6. summary and conclusions this paper looks at the relationship between planning and the level of retirement satisfaction. in the context of a basic utility maximization model, retirement satisfaction is expected to be related to financial as well as non-financial variables. data from the hrs table 4 estimates of planning action: think about and attend meetings (n 5 1781) ordered probit estimates: think about retirement binary probit estimates: attend meetings constant 2.185 2.537 (1.282) (12.154) male .386 .237 (6.507)*** (3.110)*** white .099 2.088 (1.431) (0.998) married .180 .249 (2.196)** (2.176)** totnw 2.124xe-06 .920xe-07 (1.915)* (1.058) hhinc .267xe-05 .211xe-05 (3.090)*** (1.758)* edyrs .036 .113 (3.656)*** (5.516)*** retthnk — .220 (7.077)*** hlth 2.311 2.406 (4.364)*** (4.332)*** retfor .890 2.073 (12.555)*** (0.764) m1 .291 — (13.422) log likelihood ratio (x2) 552.3 351.1 a asymptotic t-statistics in parentheses: ***, **, *, indicate statistical significance at the 1%, 5%, and 10% levels, respectively. 125h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 provides a large, representative sample to test the hypothesized relationship between retirement planning and retirement satisfaction. the results from the ordered probit model indicate that thinking about retirement and the attendance at planning meetings are positively related to retirement satisfaction; moreover, planning activities imply a higher likelihood of satisfaction even for those whose retirement decisions were not made voluntarily (either through health problems or an employer mandate). in addition to shedding some light on the impact of planning on retirement satisfaction, this paper introduces an important new data set to those studying retirement behavior from a financial planning perspective. those interested in individual financial decisions will find the health and retirement study a rich source of information on all aspects of financial decision making by older americans. references ando, a., & modigliani, f. (1963). the ‘life-cycle’ hypothesis of saving: aggregate implications and tests. american economic review, 53(1), 55–84. banks, j., blundell, r., & tanner, s. (1998). is there a retirement-savings puzzle?american economic review, 88(4), 769–788. bernheim, b. d., skinner, j., & weinberg, s. (1997). what accounts for the variation in retirement wealth among us households? nber working paper no. 6227. brunson, b. h. (1996). life satisfaction and retirement: military mid-life career change. doctoral dissertation, texas tech university. dissertation abstracts international 7–07a, 3266. brunson, b. h., snow, m., & gustafson, a. w. (1998). mid-life career change: career military versus non-career financial well-being and financial satisfaction.financial counseling and planning, 8(2), 1–12. cooper, d. f. (1993). retirement style, post-retirement work pattern and retirement satisfaction of public school administrators. unpublished doctoral dissertation, university of la verne. cope, j. k. (1990). the intrinsic and extrinsic factors of job satisfaction in post-retirement employment. unpublished doctoral dissertation, university of oklahoma. dorfman, l. t. (1989). retirement preparation and retirement satisfaction in the rural elderly.journal of applied gerontology, 8(4), 432–450. floyd, f. j., haynes, s., doll, e. r., winemiller, d., lemsky, c., burgy, t. m., werle, m., & heilman, n. (1992). assessing retirement satisfaction and perceptions of retirement experiences.psychology and aging, 7(4), 609–621. friedman, m. (1957).a theory of the consumption function.princeton, nj: princeton university press. gall, t. l., evans, d. r., & howard, j. (1997). the retirement adjustment process: changes in the well-being of male retirees across time.journal of gerontology: psychological sciences, 528(3), p110–p117. hurd, m. (1990). research on the elderly, economic status, retirement and consumption and saving.journal of economic literature, 28(2), 565–637. juster, f. t., & suzman, e. (1995). an overview of the health and retirement study.journal of human resources 30(suppl.), s7–s58. knesek, g. e. (1992). early versus regular retirement: differences in measures of life satisfaction.journal of gerontological social work 19(1), 3–34. macewen, k. e., barling, j., kelloway, e. k., & higginbottom, s. f. (1995). predicting retirement anxiety: the roles of parental socialization and personal planning.the journal of social psychology, 135(2), 203–214. poterba, j. a. (1996). personal saving behavior and retirement income modeling: a research assessment in national research council.assessing knowledge of retirement behavior(pp. 123–148). washington, dc: national academy press. 126 h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 reis, m., & gold, d. p. (1993). retirement, personality and life satisfaction: a review and two models.journal of applied gerontology, 12(2), 261–282. sterns, h. l., & gray, j. h. (1999). work, leisure and retirement. in j. c. cavanaugh & s. k. whitbourne (eds.), gerontology: an interdisciplinary perspective(pp. 355–390). new york: oxford university press. wynne, r. j., & groves, d. l. (1995). life span approach to understanding coping styles of the elderly. education, 115(3), 448–459. zavoina, r., & mcelvey, w. (1975). a statistical model for the analysis of ordinal level dependent variables. journal of mathematical sociology, 4(1), 103–120. 127h.w. elder, p.m. rudolph / financial services review 8 (1999) 117–127 106 addressing diversity, equity, and inclusion in financial planning education miranda reiter1, katherine mielitz2, and tanaka chimbane3 abstract this article presents the experiences of three professors who each developed and taught a university-level diversity, equity, and inclusion (dei) course within a financial planning or related academic program. these courses, tailored to different educational levels, addressed dei concepts in the context of financial planning and personal finance and aimed to cultivate awareness and equip future professionals with the tools to navigate dei challenges in practice and research. the study offers valuable insights and recommendations for educators interested in integrating dei into their curricula. the paper details the instructors’ motivations behind developing the courses, student learning objectives, course design and content, and the authors' unique teaching philosophies. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation reiter, m., mielitz, k., & chimbane, t. (2025). addressing diversity, equity, and inclusion in financial planning education. financial services review, 33(4), 106-120. introduction according to the 2024 census population estimates data, about 58% of the population is white non-hispanic, 20% hispanic, 14% black, and 6% asian (u.s. census bureau, n.d.). the labor force representation is similar to the overall racial and ethnic composition of the country (u.s. bureau of labor statistics, 2023). female workers represent about 47% of the u.s. workforce, which aligns with their 50% representation in the census data (schaeffer, 2024). conversely, there are around 103,000 certified financial planning™ professionals (cfp®) as of march 2025, with about 82% identifying as white and 76% identifying as men (cfp board, n.d.a). the largely homogenous white and male makeup of 1 corresponding author (mreiter@ttu.edu). texas tech university, lubbock, texas, usa. 2 kansas state university, manhattan, ks, usa. 3 texas tech university, lubbock, texas, usa. financial planners has persisted for decades (cfp board, 2014; cfp board, 2018). however, it has been noted that this demographic does not accurately represent either the american population or the u.s. labor force. in short, the financial planning profession has a diversity problem (iacurci, 2022; lake, 2022). when diversity is an issue, it often follows that equity and inclusion are also areas that must be addressed in the profession. colloquially, diversity can be understood as “counting the people,'' and inclusion is “the people have a voice.” equity varies from equality and can be understood as “the peoples’ voices matter at the organizational level” (bernstein et al., 2020). these three terms, diversity, equity, and https://creativecommons.org/licenses/by-nc/4.0/ mailto:mreiter@ttu.edu reiter et al. 107 inclusion, are often called “dei.” belonging is a more recent addition to the dei acronym, often seen as “deib.” belonging is the idea that the people who have been included also have a voice and feel accepted and respected (kurfist, 2022; powertofly, n.d.). over the past several years, leaders, practitioners, and researchers have come to recognize the increasing importance of diversity, equity, and inclusion in the financial planning profession. financial planning firms, associations, and other key actors have responded to the glaring underrepresentation of women and people of color in the profession through various channels, such as research and white papers, scholarships, targeted internships, diversity conferences, and more. there has been a strong push to bring awareness to these issues and change the face of professional financial advice. there is evidence that these efforts have resulted in some positive changes. for example, in 2022, women cfpⓡ professionals comprised 30% of the new certificant pool, and professionals from nonwhite backgrounds comprised 15% of the new cfp certificants for that year (donachie, 2023). however, the gap persists, and representation in the profession lags significantly (cfp board, n.d.-a). one of the ways to affect positive change and progress in the diversity, equity, and inclusion space is to introduce it to the future leaders of the financial planning profession: students. while the pipeline for financial professionals is filled from many areas, one of the most prominent ways is through academic programs (reiter & kiss, 2021). there is evidence that the lack of diversity seen at the professional level is often present at the academic level (diversitas, 2022; reiter & kiss, 2021; reiter, 2023). however, there is room for change and growth. as of april 2024, there are more than 390 cfp board registered academic programs, including certificate, undergraduate, graduate, and doctoral programs across the u.s. (cfp board, n.d.-b). this large number of programs represents an amazing opportunity to effectively address issues of diversity, equity, and inclusion, not only at the academic level but more broadly in the profession. introducing dei at the academic level is crucial for several reasons. first, students must be educated through a lens of dei so that these future financial professionals are aware of the issues, barriers, challenges, and opportunities and are equipped with tools to address the challenges, fix the problems, and offer solutions. second, many students in these academic programs will be client-facing. in a society that is increasingly becoming more diverse, they will need relevant education in history, cultural competence, and dei to develop skills to work with clients and other constituents from various backgrounds effectively. culture plays a role in how people approach their finances, and students must understand these nuances and how to approach them with their clients. third, financial planning needs more students and researchers who can conduct ethical and culturally competent research related to diversity, equity, and inclusion to offer empirical-based solutions to the profession. one possibility to help introduce dei to students is through a formal course within a given curriculum. university classes provide students with opportunities to grow and be exposed to additional perspectives and discourse on topics (samuel et al., 2024). according to data from a 2015 association of american colleges and universities survey, about 60% of institutions included diversity as a requirement within general education. having students learn about diversity within the u.s. is an important learning objective for 73% of the institutions surveyed (humphreys, 2016). however, when students obtain dei education specifically targeted to their fields of study, this may help them understand how dei applies directly to their work and what they can do to impact change. the purpose of this paper is to share the experiences of three professors who created three original dei courses within financial planning, personal finance, and human sciences curricula at two separate u.s. universities. specifically, each author provides details on (a) the motivation and purpose of their course, (b) the objectives of their course, (c) course design and content, (d) how they discussed sensitive topics, and (e) their teaching philosophy. the first two classes discussed were offered to undergraduates, while the final course was offered to graduate students. financial services review, 33(4) 108 a summary of course details are provided in table 1. in addition, the authors suggest recommendations for others interested in developing and teaching dei courses in financial planning and related college-level academic programs. to the authors’ knowledge, this paper is the first account of professors’ experiences teaching dei in financial planning and related curricula. as such, this paper fills a gap in the existing literature in several ways. first, personal accounts of teaching diversity-related courses in the financial planning and personal finance contexts are provided, along with the basic structure of the courses. this could serve as a foundation for other instructors seeking resources in this field who have minimal or no existing framework. the authors provide recommendations based on lessons they have learned, which may help other instructors build upon existing work, avoid certain mistakes, and overcome barriers. third, the authors offer their perceptions of the student experience after taking their classes. even though each class was developed independently with different student populations in mind, similar student benefits were achieved. it is important to note that the authors of this paper share their experiences from the perspective of being female, pre-tenure, tenuretrack assistant professors at the time of teaching their respective courses. each of the authors has a ph.d. in personal financial planning. two have their accredited financial counselor (afc®) designations, and one has their certified financial planner (cfp®) designation. table 1. course details class 1 class 2 class 3 name of class financial perspectives throughout the united states cultural and gender diversity in personal finance diversity, equity, and inclusion in financial planning research first semester course taught spring 2019 fall 2023 fall 2022 institution type large r1: doctoral university – very high research activity large r1: doctoral university – very high research activity large r1: doctoral university – very high research activity geographic region southwest united states southwest united states southwest united states program’s college college of education & human sciences college of health & human sciences college of health & human sciences required/elective/ general education offered as a general education diversity credit; offered as an honors college general education diversity credit offered as general education offered as a required course in the doctoral program number of semesters taught 5 2 1 number of students 140 60 15 availability of course offered 2 times per year offered 3 times a year offered once every 2 years number of credit hours 3 3 3 duration of class 16 weeks; 1 hour 20 min class; 2x/week 16 weeks, 1 hour 20 min class; 2x/week 16 weeks; 1 hour 20 min class; 2x/week reiter et al. 109 modality face-to-face (except for the last half of spring 2020, which finished online, synchronous) face-to-face, online (hybrid and asynchronous) face-to-face, online (synchronous) academic level undergraduate undergraduate graduate experiences of three professors teaching financial planning and personal finance dei courses class #1: financial perspectives throughout the united states motivation and purpose of the course everyone has a relationship with money, but not everyone considers how others use money—and how race and ethnicity, sex and gender, age, religion, family structure, and experiences with wealth and wealth disparity play a role in the financial relationship. this course offers an introductory look at how history has shaped our financial behaviors of today and how our financial behaviors and those of others can shape our tomorrow. financial perspectives throughout the united states takes an oft-taboo subject and encourages relevant discourse and investigation into how differences in financial habits and behaviors are not always negative but guided by history, informed by the present, and shape our future as individuals and as a country. the development of this undergraduate course was grounded in the need to get a diversityrelated general education course in the department. primary concerns included, but were not limited to, covering all the possible subject matter while providing students with a course that flowed and made sense from one topic area to the next. many topic areas of diversity, such as disability and ableism, were not covered. course objectives upon completion of this course, students were expected to have attained the following competencies: 1. understand and explain the historical and present-day financial implications of race, ethnicity, sex, gender, aging, religion, family structure, and experiences with wealth and wealth disparity when considering personal financial resources. 2. gather and interpret information, respond and adapt to changing situations, make complex decisions, solve problems, and evaluate personal experiences and behaviors. 3. demonstrate the ability to communicate effectively through class discussions and presentation(s). 4. critically reflect on potential solutions to increase awareness and support diversity in the united states. 5. demonstrate ability to write effectively through course assignments. course design & content the class was designed to cover units on race and ethnicity, sex and gender, religion, aging, family structure, and experiences with wealth and wealth disparities. while largely lecturebased, hands-on learning experiences, small group discussions, and student presentations were worked into the design. race and ethnicity and experiences with wealth and wealth disparity took most of the semester, bookending the other topics. the intersectionality of the subject matter allowed for rich wrap-up discussions to close out the semester. discussing sensitive topics race and ethnicity were discussed first, as this professor believes there is a common theme of intersectionality with race and ethnicity and the other topics included in the curriculum. therefore, race became the foundational discussion point throughout the class. discussing race and ethnicity along with the history of the banking system and trying to reach predominantly white college students about privilege was the most difficult part of the course and the semester but also the most rewarding. while it was risky to begin a semester with such a heavy subject matter, it was imperative that the foundation laid was as strong as it could be. multiple semesters of teaching this course bore financial services review, 33(4) 110 out the need for addressing race as part of the initial foundation of the course. it is possible, even likely, that had the course been taught by a black professor, the predominantly white student body would not have been as receptive to race being addressed as the core framework for investigating personal finance. engaging students and building trust with them was predominantly focused on getting to know students. everyone was provided with a name tent, and students were called by name as often as possible and eventually without the name tents, as familiarity increased. students realized that they were recognized as individuals, which helped build trust in discussing a challenging topic such as race and its influence on personal finance. students were particularly engaged in the hands-on activities and small group discussions. the student feedback from prior semesters informed each of the following semesters, making it easier to meet students where they were and to help them engage with the materials through readings, class discussions, and activities. teaching philosophy the fundamental goal of teaching is to foster personal growth and learning in a safe environment. understanding that every student comes to the table with their own experiences, comfort level, and knowledge is instrumental to providing a welcoming and safe-to-share atmosphere. my vision is: all students feel welcome to participate and learn in a safe, topicspecific setting. my calling, both personal and professional, is to educate people about money. money in its numerous forms and uses—cash, credit, investment products, or otherwise—is not only a part of one’s everyday existence but also of one's future. since everyone has their own functional or dysfunctional relationship with money, it is of the utmost importance to provide an environment which encourages students to openly discuss and practically apply the financial concepts they understand and those they find confusing. money is a tool. helping my students realize that they can use their tools in their own way to make successful financial lives for themselves is imperative. to practically apply financial concepts requires continued investigation and interest in the learning styles of my students, be it hands-on, small or large group, or one-on-one types of instruction. having different options for working through the necessary learning objectives is imperative to a welcoming and truly educational environment. assessment of my students’ mastery of the topics at hand is conducted in creative ways which also addresses and embraces different learning styles. students work independently and as part of small groups and present assignments such as essays, presentations (individual and small group), and hands-on activities. these different assessments allow students to express their mastery of the topics in ways that encourage engagement in the topic and the learning process. class #2: cultural and gender diversity in personal finance motivation and purpose of the course recognizing the crucial impact of financial counseling, coaching, and education in today’s economy, this course emphasized the importance of multicultural components. it is designed to equip undergraduate students with the skills and awareness necessary to effectively serve diverse individuals and families. the insights gained from this course are applicable across various industries, highlighting the class's role in a comprehensive personal finance education. course objectives upon completion of this course, students were expected to have attained the following competencies: 1. develop a comprehensive understanding of the core concepts of differences, fairness, and belonging (dfb), equity vs. equality, power and privilege, and their relevance to financial wellness. 2. evaluate the influence of cultural factors on income generation, accessibility to financial services, and financial wellness. 3. assess the influence of gender on income generation and accessibility to financial services and financial wellness. reiter et al. 111 4. understand the history and current state of financial access and develop strategies to promote financial wellness. course design & content the course content and design are structured to build a comprehensive understanding of diversity, equity, and inclusion (dei) in the context of financial wellness over 16 weeks. the first section, spanning the initial weeks, establishes foundational concepts by defining differences, fairness, and belonging (dfb), power and privilege, equity vs. equality, and financial wellness. these core principles set the stage for deeper exploration in subsequent sections. in the second section, these concepts are applied to analyze how culture influences career and education choices, financial decisions, economic outcomes, and overall financial wellness. the third section continues to use these foundational concepts to examine gender dynamics, exploring the intersections of gender with career and education choices, financial decisions, and financial wellness. finally, the course delves into the historical context of financial wellness, identifying barriers and discussing solutions to achieve greater equity and inclusion. discussing sensitive topics race was strategically taught during the final three weeks of the course. the early parts of the course were designed to build trust and introduce fundamental concepts while minimizing any sense of unease. this approach was taken so that students first established a solid understanding of key principles, their significance in society, and their relation to financial stability. as the course progressed to cover cultural perspectives and gender, race was naturally woven into class discussions. as students became more comfortable with the material and with each other, they began to bring up race spontaneously. this allowed for organic and passive discussions, setting the stage for a more in-depth exploration in the final weeks. in the final weeks of class, the focus shifted to examining how systemic barriers and inequalities, such as segregation and other racial movements, have shaped financial systems and impacted marginalized communities. by this point, students had already encountered and discussed race indirectly and naturally during the discussions on cultural perspectives, gender, and other foundational topics. this strategy enabled students to see race as an integral part of broader societal issues rather than an isolated topic, making the discussions more relevant and impactful. when the time came to cover race indepth, students were not taken by surprise and were better prepared to handle the complexities of the subject with a well-rounded perspective. teaching philosophy my teaching philosophy is deeply studentcentered. i focus on sharing the power dynamic, fostering a learning environment where students actively participate, and continuously assessing understanding through various methods. i focus on three core principles: engagement, encouragement, and personalized coaching. every student is on a unique learning journey, and it is my responsibility to tailor my approach to meet their individual needs. to ensure active participation and ignite curiosity, i create dynamic classes incorporating storytelling, personal anecdotes, and relatable narratives to connect students with the material personally. incorporating practical strategies such as case studies, interactive games, and "think-pair-share" activities helps maintain high engagement and connect theory with real-world application. the transformative power of encouragement and positive feedback is also pivotal in enhancing students’ creative and cognitive abilities. recognizing the unique dynamics of each class has taught me the importance of being flexible and continually evolving my teaching methods. this adaptability, coupled with a commitment to ongoing learning and application of new educational techniques, is essential for improving student outcomes and enhancing my effectiveness as an educator. class #3: diversity, equity, and inclusion in financial planning research motivation and purpose of the course unlike law, medicine, and other long-established professions, diversity, equity, and inclusion issues are relatively new points of broad financial services review, 33(4) 112 discussion and concern in financial planning research. the academic side of financial planning is well-established but fairly new compared to more traditional programs. as such, it follows that dei issues have been scantily investigated in the academic body of literature. academic research serves a very important purpose, particularly when associated with practice-based professions, as it can provide insights and empirical evidence of phenomena occurring within a field. until recently, much of the research conducted on dei was based on market research and less on empirical evidence. the course, diversity, equity, and inclusion in financial planning research, had a two-pronged purpose. first, it was developed to introduce students to the contemporary issues of diversity, equity, and inclusion (dei) in personal financial planning-related research. dei training is not required in the cfp board academic curriculum and is often not taught to students in financial planning programs. the course content also addressed the financial realities of diverse consumers through the lens of dei. the second purpose of the course was to help cultivate future researchers who have the skills to answer some of the profession’s most pressing questions around dei and provide research-based solutions. course objectives upon completion of this course, students were expected to have attained the following competencies: 1. understand what diversity, equity, and inclusion mean in the context of financial planning in the united states. 2. develop an understanding of the landscape of dei issues in pfp (personal financial planning). 3. learn about the various initiatives that have been put forth to address dei in the financial planning profession. 4. become aware of research related to diversity, equity, and inclusion in financial planning, as well as personal finance research related to race, culture, gender, ethnicity, and marginalized groups. 5. create and give presentations on dei pfprelated research. 6. develop skills on how to find, summarize, and analyze dei pfp-related research. 7. learn how to provide peer-review feedback and use feedback to improve research ideas. 8. write a research paper and proposal focused on a dei-related topic. course design and content the 16-week course was taught twice a week in 1-hour and 20-minute sessions. in the first few weeks of class, it was taught in a lecture-type style to introduce students to the relevant dei concepts and terms, previously published works, and initiatives. the course content centered on the following topics: introduction to dei, dei financial planning research, financial planning research on diverse populations, evaluating and writing research papers, and dei research proposals. for reading material, we used two books on research methods, the american psychology association (apa) manual, and a variety of other academic and white papers and book chapters. in addition to lectures, students learned from guest speakers and short videos. in the first week of class, students were asked to conduct a literature review on published peerreviewed research papers on dei to help them learn more about the body of work. this assignment was somewhat of a challenge for students as not a lot had been published at the time of the course in the fall of 2022. however, this assignment allowed students to get a sense of what had been done. in the initial weeks of class, students discussed relevant research papers in groups and then presented their perceptions and thoughts to the class. after students had gotten a significant level of exposure to lectures, guest speakers, and group discussions, they were then responsible for developing a dei research idea. students worked on this idea with the help of peer review from their classmates. each week, students presented various components of their research idea and each time, their peers provided them with feedback and as such, each student received feedback evaluations from their peers each time they presented. in subsequent presentations, aspects of the student’s project reiter et al. 113 changed based on feedback from the class as well as the instructor. the instructor provided live inclass and post-class written feedback on projects and presentations. the final product for the class was a research proposal which several students continued to develop after the class was completed. discussing sensitive topics even when the course was merely a thought and not yet developed, it was understood that teaching it all of it could be considered sensitive to the students and also possibly, at times, uncomfortable for all of us. an important decision had to be made. how would i teach this class while protecting my comfort and the students’ while not jeopardizing the integrity of the course content? i decided that the approach that made the most sense was to recognize that we all instructors and students will possibly feel uncomfortable, but we have a common interest: to improve our profession, and dei is a part of that effort. i believe that being transparent about those feelings and sharing them early in the class is important to help students understand where the instructor is coming from. it also helps to convey to students that they are not alone in how they may feel. i also felt it was imperative to share my excitement with students about why dei is critical to our incredible profession and how they could be a part of making it better. from a pedagogical aspect, i began the course by reviewing common terminology related to dei so that the students and i were on the same page. this included discussing terms like dei, race, ethnicity, sex, gender, and intersectionality. however, i did make it clear that we were not there to debate the importance of dei in the field of financial planning. one way that i did this was to allow the published work to speak for me. as a class, we reviewed many of the white papers released by organizational leaders and financial planning firms that explained and confirmed the importance of dei to financial planning. this step, i believe, helped students to understand that solving issues around diversity, equity, and inclusion was not the instructor’s idea but rather an important issue to address as a profession. i believe allowing the students to have creative space to ponder and discuss what they read and develop ideas related to how they could address dei in their own way helped to eliminate some of the potential resistance to the class and the material. in short, my approach to discussing sensitive topics included creating a safe environment for students to learn, expecting and giving respect and kindness, and reiterating, in various ways, that we were all there to learn, create, and advance the dei research agenda for our field. teaching philosophy students in the financial planning program are the future of our profession. it is my duty as their instructor to make sure they are well-equipped to understand the most important topics of financial planning, as well as current events so that they can help their clients, produce meaningful research when applicable, and ultimately push the industry forward. i do this by facilitating the learning process through teaching and providing a variety of opportunities to learn. i have learned that my motivation for teaching rests in seeing students grasp concepts and gain knowledge that they did not have before entering my classroom. i feel a special gratification seeing students develop new ideas, which were inspired by the class or the teaching, but catapulted by their own intellect and imagination. in creating and teaching the dei course, i knew that it was a special opportunity to open a path for students to learn about diversity, equity, and inclusion while having the opportunity to actively participate in creating solutions for our profession. to help me teach, i often use a combination of lectures accompanied by visual aids, guest speakers, real-world financial planning stories and case studies, a variety of readings including white papers, peer-reviewed articles, and textbooks, videos, podcasts/audio clips, in-class group exercises, q&a sessions, in-class discussions, and industry-related software and tools. in any given class, there will be students with different learning styles and abilities. students may have preferences for auditory, visual, kinesthetic, reading/writing, or combination learning. as such, it is critical to use a variety of ways to reach students since all do not learn the same. in teaching the dei course, i found that remembering this fact was even more important than perhaps other courses. financial services review, 33(4) 114 i also have a strong belief that students themselves are an invaluable teaching resource. they have the ability to educate their peers and share knowledge in ways that the instructor cannot. as mentioned before, in my dei course, students were tasked with engaging in the peerreview process by giving constructive feedback to their peers’ in-class presentations. many students referenced this exercise, which involved applied learning and peer feedback, as one of the most important components of learning in the course. students come to the classroom with a myriad of perspectives and experiences that can greatly enhance learning possibilities. it can be a game-changer to incorporate this as a resource while teaching dei-related topics all while ensuring students’ thoughts and viewpoints are heard. best practices for teaching dei in financial planning and related programs the three courses discussed in this paper were created due to needs identified either by the professors themselves or within the respective colleges and departments. based on student enrollment, student feedback, and faculty feedback, these courses have been very successful thus far. the professors involved in the development and teaching of these courses are personally vested in the benefits of discussing dei matters with students. course development research was not used as a foundation, but from these courses, we have identified six (6) recommendations for future course development. 1. set goals and clear ground rules as with any post-secondary education course, it is important to set clear goals and objectives when teaching (albilehi et al., 2013; eberly center, n.d.), but this may be even more crucial for dei classes. it is critical that students understand why the class is being taught. the instructor should be clear on the objectives for themselves as well as for the students. this is critical to obtaining student buy-in and staying the course when challenges arise. setting ground rules will help to cultivate a respectful and inclusive atmosphere that encourages students to express themselves openly (eberly center, n.d.). clear guidelines help manage discussions and ensure that all voices are heard and respected (eberly center, n.d.). 2. create a trusting atmosphere establishing trust is essential, especially when addressing sensitive topics. a trusting environment fosters more engaged and open dialogue, allowing students to share their thoughts and experiences more freely (eberly center, n.d.). students want to be seen and respected as individuals and may struggle to be brave in a classroom where dei topics are being openly discussed. instructor-student engagement is imperative. students may resist that which makes them uncomfortable. provide space for them to contribute their thoughts and perspectives (eberly center, n.d.). this can be especially true when talking about the topic of race (and its intersection with personal finance), diversity, equity, and inclusion. however, it is not the job of the instructor to change minds. the instructor’s job is to educate and plant seeds. students tend to engage with material when facts, not personal feelings, are used to drive the conversation. being effective requires setting aside ego and knowing that some students will not be able to fully comprehend the impact of the lessons in the current course. it is recommended to develop a timeline for introducing topics which will simultaneously allow for trust to be built and while aligning with the instructor’s comfort level. the timeline and comfort levels will vary based on the individual. for example, some may feel comfortable introducing certain sensitive topics earlier in the course, such as race, while others may feel more comfortable introducing this later. overall, the course should have a flow that allows the subsections to tie together and while not avoiding the more difficult discussions or putting them off in a way that makes them appear less important. ensuring the course reading materials overlap in specific areas may make identifying this flow easier (e.g., the intersection of race and gender, the intersection of race and experiences with reiter et al. 115 homelessness or incarceration, the intersection of gender and family structure). 3. be aware of positionality while a professor of any background can teach a dei course, it is important to be aware of the concept of positionality, specifically when teaching classes that incorporate educating students on race, ethnicity, culture, and related topics, as it may have an effect on the success of the course (hearn, 2015). positionality involves recognizing one’s own social position and power, or lack thereof, and its implications in society and the spaces in which one operates (hearn, 2015). understanding one’s positionality as an educator is important because it allows one to understand certain dynamics in the classroom, such as instructor and student biases, preconceived notions of social constructs, and power dynamics (hearn, 2015). there is plenty of research to suggest that those who come from traditionally underrepresented and devalued backgrounds, for example, african american women in the united states, have different and more complex experiences from their white counterparts in the classroom (gutiérrez y muhs et al., 2012; niemann et al., 2020). namely, there are specific challenges around negotiating self, identity, and power dynamics for african american professors which requires “extensive emotion management” (harlow, 2003, p. 348). these experiences may be exacerbated by the type and values of one’s institution, the demographics of the students in a given class, or the classes that one teaches. miller and struve (2020) assert that those with marginalized racial identities and those with less secure job statuses, such as non-tenure track professors, are more vulnerable while teaching diversity courses as they are simultaneously attempting to “uphold their intellectual authority, all while considering students’ social, cultural, and emotional responses to the course content” (miller & struve, 2020, p. 438). 4. understand challenges teaching a course on diversity, equity, & inclusion general education credits in diversity are often required to help students critically think about the world around them (laird & engberg, 2011). still, the subject of dei can be and often is polarizing. it can be perceived as divisive, and it may take some students more time to warm up to the material. what can complicate matters is that dei topics can sometimes veer more into the “gray” zone rather than the “black or white” or cut-anddried, indisputable facts, as one might find in hard sciences. it is important to remember that when teaching these courses, dei is about people, including the emotions and feelings that we have. to the point of harlow (2003) and hearn (2015), professor emotion management, student biases, and power dynamics, among other factors, may influence one’s experience in dei courses including instructor comfort. miller, howell, and struve (2019) reported on the emotional labor and depleting effects of teaching diversity courses after interviewing 38 faculty members teaching such courses. even experiences with imposter syndrome may present while teaching these courses. imposter syndrome can be described as when one feels that they do not deserve to be in the professional status that they are, and this notion will become known to others (chrousos & mentis, 2020). in dei courses, students and instructors will all likely feel some form of discomfort at a given time. these feelings are normal. what is important to remember in these instances is the “what “and the “why”. why is this course being taught? what do students have to gain by taking this course, and what do they have to lose by not taking it? what is the ultimate goal? when the objectives and motivations for teaching and offering such a course are clear, it makes it easier to move past some of the negative emotions or perceived threats real or not that may come with teaching it. support from colleagues, faculty, and college administrators for the establishment of the class is also invaluable. it is worth noting that the three courses reviewed in this paper were offered within colleges of human sciences, a setting that financial services review, 33(4) 116 may influence both instructional focus and student engagement in dei issues. unlike programs housed in business or economics schools, human sciences curricula often adopt a holistic approach to individual and societal well-being, potentially creating an environment conducive to in-depth exploration of dei concepts. this setting may offer distinct advantages, such as incorporating diverse social factors affecting health and human development into financial education. however, it may also restrict the applicability of findings to other academic environments that prioritize different disciplinary perspectives. consequently, future research could explore how the specific academic home of a course (e.g., human sciences, business, or agriculture) shapes dei-focused curricula and outcomes, thus expanding our understanding of effective strategies for integrating diversity, equity, and inclusion across various academic disciplines. 5. strategies for engaging diverse learners when delivering dei classes, it is important to use a mix of strategies and be flexible to accommodate different learning styles and needs, as dei topics can be sensitive and require a nuanced approach. consider incorporating case studies reflecting diverse cultural backgrounds and real-world scenarios related to equity and inclusion, and include hands-on activities, small group discussions, and presentations to foster collaboration and active learning. engage students by integrating themes like money and music from various cultures and providing multiple accessible resources, such as articles, videos, and podcasts from diverse voices. recognize that dei discussions can be intense; give students permission to take short breaks if the material becomes overwhelming. building trust is crucial, and this can be partially achieved by creating a brave space where students feel comfortable sharing their thoughts (arao & clemens, 2023; winks, 2018). an instructor might also consider grading based more on growth rather than exclusively "right and wrong" answers. prioritize core material, knowing it may be impossible to cover everything in class, and assign additional content for students to explore independently. 6. ways to implement dei into financial planning curriculum without developing a class course development often goes through an arduous revise and resubmit process. if a course is stalled or even in development, including facets of dei in your regularly scheduled coursework is highly recommended. there are numerous ways to incorporate dei education into a curriculum. here are a few examples of how dei may be worked into coursework. in an insurance class, one could facilitate discussion or provide a lecture on the history of life insurance. there is a historical connection to enslaved people (ralph, 2012). this additional historical context would be beneficial and would not take too much time out of the schedule. additionally, in a section on health insurance, one could discuss the influence of race on health insurance and perceptions of in-office medical treatment (hill et al., 2024; schumacher et al., 2024). discussions regarding disability and ableism would also be relevant. it might be appropriate in various financial planning or personal finance courses to integrate a discussion around generational wealth and what that might mean for different types of families from diverse backgrounds. discuss the racial wealth gap and what this means for estate planning among families of color. in both estate planning and retirement planning coursework, students can investigate and discuss research papers that examine the associations of gender, race, and ethnicity in contributing to retirement plans or establishing a will or trust and the importance of historical context. this culturally relevant information could be worked into case studies as well. an investing class could discuss the impact of the pandemic to make the wealthy wealthier (kochhar & moslimani, 2023). additionally, coursework that encourages reiter et al. 117 students to investigate the emergence of fractional share programs and their influence on investing behaviors of various racial and age groups could be given. also, there is a great opportunity to discuss the diversity of religion and its influence on investing strategy–particularly where debt products can and cannot be used as part of an investment portfolio. a foundational personal finance class could include sections on how disability influences budgeting practices or on banking history and the inequities in borrowing and credit. a personal finance or financial counseling class could include an assignment that asks students to write or discuss their family culture and family money stories. in a class that covers aspects of the financial planning profession, such as interviewing, internships, hiring, recruiting, and other human resources issues, instructors can incorporate how diversity, equity, and inclusion are relevant. when discussing client issues and cases, instructors can create scenarios that cover client experiences from diverse backgrounds and identities. again, these are only a few of many examples of how to integrate dei into financial planning and related courses. conclusion the purpose of this paper was to share the unique experiences of three instructors who developed individual personal finance and financial planning-related dei courses at two universities. each course was different and approached diversity, equity, and inclusion in a distinct way, with two of the classes tailored to undergraduates and the other developed for graduate students. given the novelty of dei-related courses in financial planning academic curricula, the authors aimed to provide recommendations for those interested in developing and teaching similar courses in financial planning and related college-level academic programs. the current study adds to the body of literature in several ways. first, it provides first-hand accounts of teaching diverse dei courses in the financial planning context and includes the basic structure of the courses. this may help provide a starting point for other instructors looking for resources in this area and starting with little to no framework. to the authors’ knowledge, this paper is the first account of teaching dei in financial planning curricula. second, the authors shared lessons learned via recommendations, which may allow other instructors to build upon what has already been done, avoid some mistakes, and overcome roadblocks. third, the authors found that their students reported overwhelmingly positive experiences in course evaluations. this paper covered three individual classes developed by three professors, and no overlap occurred, but similar student benefits were identified. affinity for the professors aside, overarchingly, students felt that the programming was valuable and helped inform how they looked at the world and considered what was going on around them. we hope this inspires others that possible negative student sentiment should not be a reason to avoid establishing such a course or incorporating the content into existing courses. students are the future of our profession, and dei is an important component of their educational experience and a requisite for positively affecting change in our profession. dei education and programming are needed at the higher education level. this is the first investigation of its kind to explain how dei, financial planning, and personal finance are interwoven in the classroom through dei courses. that stated, there are some limitations that we would be remiss in not discussing. while part of different programs, two of these courses were taught within the same university. all courses were taught at universities with very high research activity (r1) within the southwestern united states. as such, the sentiments, experiences, and recommendations may not be generalizable to other contexts. additionally, these courses were not purposefully reviewed for research purposes after each semester. student feedback and detailed instructor reflection are not included but should be considered for future research. it should also be mentioned that the experiences each instructor had teaching their course are linked to their own identities and perceptions. financial services review, 33(4) 118 references albilehi, r., han, j. y., & desmidt, h. (2013). curriculum development 101: lessons learned from a curriculum-design project. catesol journal, 24(1), 187– 197. https://doi.org/10.5070/b5.36162 arao, b., & clemens, k. (2023). from safe spaces to brave spaces: a new way to frame dialogue around diversity and social justice. in the art of effective facilitation (pp. 135–150). routledge. https://doi.org/10.4324/9781003447580 bernstein, r. s., bulger, m., salipante, p., & weisinger, j. y. (2020). from diversity to inclusion to equity: a theory of generative interactions. journal of business ethics, 167, 395–410. https://doi.org/10.1007/s10551-01904180-1 cfp board. (2014). making more room for women in the financial planning profession [white paper]. https://www.cfp.net/knowledge/reportsand-statistics/diversity-and-womensresearch/making-more-room-forwomen-in-the-financial-planningprofession cfp board. (2018). racial diversity in financial planning: where we are and where we must go [white paper]. https://www.cfp.net/knowledge/reportsand-statistics/diversity-and-womensresearch/racial-diversity-in-financialplanning-where-we-are-and-where-wemust-go cfp board. (n.d.-a). cfp professional demographics. https://www.cfp.net/knowledge/reportsand-statistics/professional-demographics cfp board. (n.d.-b). find an education program. https://www.cfp.net/getcertified/certification-process/educationrequirement/certification-courseworkrequirement/find-an-education-program/ chrousos, g. p., & mentis, a. f. a. (2020). imposter syndrome threatens diversity. science, 367(6479), 749–750. https://doi.org/10.1126/science.aba8039 diversitas. (2022). diversitas partner university demographic survey. https://blogs.uakron.edu/diversitasfp/wp content/uploads/sites/1663/2024/05/ba1123-1279-diversitas-demographicsurvey7153377.pdf donachie, p. (2023). cfp professionals’ diversity grows, but lags u.s. demographics. https://www.wealthmanagement.com/in dustry/cfp-professionals-diversitygrows-lags-us-demographics eberly center. (n.d.). how to center dei in teaching. carnegie mellon university. https://www.cmu.edu/teaching/designtea ch/diversityequityinclusion/index.html gutiérrez y muhs, g. g., niemann, y. f., gonzález, c. g., & harris, a. p. (eds.). (2012). presumed incompetent: the intersections of race and class for women in academia. university press of colorado. harlow, r. (2003). “race doesn’t matter, but…”: the effect of race on professors’ experiences and emotion management in the undergraduate college classroom. social psychology quarterly, 66(4), 348–363. https://doi.org/10.2307/1519834 hearn, m. c. (2015). positionality, intersectionality, and power: socially locating the higher education teacher in multicultural education. multicultural education review, 4(2), 38–59. https://doi.org/10.1080/2005615x.2011. 11102893 hill, l., artiga, s., & damico, a. (2024). health coverage by race and ethnicity, 2010– 2022. kff. https://www.kff.org/racialequity-and-health-policy/issuebrief/health-coverage-by-race-andethnicity/ https://doi.org/10.5070/b5.36162 https://doi.org/10.4324/9781003447580 https://doi.org/10.1007/s10551-019-04180-1 https://doi.org/10.1007/s10551-019-04180-1 https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/making-more-room-for-women-in-the-financial-planning-profession https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/racial-diversity-in-financial-planning-where-we-are-and-where-we-must-go https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/racial-diversity-in-financial-planning-where-we-are-and-where-we-must-go https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/racial-diversity-in-financial-planning-where-we-are-and-where-we-must-go https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/racial-diversity-in-financial-planning-where-we-are-and-where-we-must-go https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/racial-diversity-in-financial-planning-where-we-are-and-where-we-must-go https://www.cfp.net/knowledge/reports-and-statistics/professional-demographics https://www.cfp.net/knowledge/reports-and-statistics/professional-demographics https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/find-an-education-program/ https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/find-an-education-program/ https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/find-an-education-program/ https://www.cfp.net/get-certified/certification-process/education-requirement/certification-coursework-requirement/find-an-education-program/ https://doi.org/10.1126/science.aba8039 https://blogs.uakron.edu/diversitasfp/wp-content/uploads/sites/1663/2024/05/ba-1123-1279-diversitas-demographic-survey7153377.pdf https://blogs.uakron.edu/diversitasfp/wp-content/uploads/sites/1663/2024/05/ba-1123-1279-diversitas-demographic-survey7153377.pdf https://blogs.uakron.edu/diversitasfp/wp-content/uploads/sites/1663/2024/05/ba-1123-1279-diversitas-demographic-survey7153377.pdf https://blogs.uakron.edu/diversitasfp/wp-content/uploads/sites/1663/2024/05/ba-1123-1279-diversitas-demographic-survey7153377.pdf https://blogs.uakron.edu/diversitasfp/wp-content/uploads/sites/1663/2024/05/ba-1123-1279-diversitas-demographic-survey7153377.pdf https://www.wealthmanagement.com/industry/cfp-professionals-diversity-grows-lags-us-demographics https://www.wealthmanagement.com/industry/cfp-professionals-diversity-grows-lags-us-demographics https://www.wealthmanagement.com/industry/cfp-professionals-diversity-grows-lags-us-demographics https://www.cmu.edu/teaching/designteach/diversityequityinclusion/index.html https://www.cmu.edu/teaching/designteach/diversityequityinclusion/index.html https://doi.org/10.2307/1519834 https://doi.org/10.1080/2005615x.2011.11102893 https://doi.org/10.1080/2005615x.2011.11102893 https://www.kff.org/racial-equity-and-health-policy/issue-brief/health-coverage-by-race-and-ethnicity/ https://www.kff.org/racial-equity-and-health-policy/issue-brief/health-coverage-by-race-and-ethnicity/ https://www.kff.org/racial-equity-and-health-policy/issue-brief/health-coverage-by-race-and-ethnicity/ https://www.kff.org/racial-equity-and-health-policy/issue-brief/health-coverage-by-race-and-ethnicity/ reiter et al. 119 humphreys, d. (2016). progress and prospects for the reform of undergraduate education. liberal education, 102(3), 28–35. iacurci, m. (2022). while slowly improving, a lack of diversity in the financial planning industry persists. cnbc. https://www.cnbc.com/2022/02/15/alack-of-diversity-still-persists-in-thefinancial-planning-industry.html kochhar, r., & moslimani, m. (2023). wealth surged in the pandemic, but debt endures for poorer black and hispanic families. pew research center. https://www.pewresearch.org/race-andethnicity/2023/12/04/wealth-surged-inthe-pandemic-but-debt-endures-forpoorer-black-and-hispanic-families/ kurfist, a. (2022). student belonging: the next dei frontier in higher education. hanover research. laird, t. f. n., & engberg, m. e. (2011). establishing differences between diversity requirements and other courses with varying degrees of diversity inclusivity. the journal of general education, 60(2), 117–137. https://doi.org/10.5325/jgeneeduc.60.2.0 117 lake, r. (2022). the financial advisor industry still lacks diversity. investopedia. https://www.investopedia.com/financialadvisor/fa-profession-accountable-lackdiversity/ miller, r. a., & struve, l. e. (2020). “heavy lifters of the university”: non-tenure track faculty teaching required diversity courses. innovative higher education, 45(6), 437–455. https://doi.org/10.1007/s10755-02009517-7 miller, r. a., howell, c. d., & struve, l. (2019). “constantly, excessively, and all the time”: the emotional labor of teaching diversity courses. international journal of teaching and learning in higher education, 31(3), 491–502. niemann, y. f., gutiérrez y muhs, g., gonzález, c. g., & harris, a. p. (2020). presumed incompetent ii: race, class, power, and resistance of women in academia (1st ed.). utah state university press. powertofly. (n.d.). 4 strategies to develop your leadership skills. https://powertofly.com/up/workivadevelop-your-leadership-skills ralph, m. (2012). “life… in the midst of death”: notes on the relationship between slave insurance, life insurance, and disability. disability studies quarterly, 32(3), 1– 27. https://doi.org/10.18061/dsq.v32i3.3267 reiter, m., & kiss, e. (2021). efforts in diversity and recruiting in financial planning undergraduate programs. family and consumer sciences research journal, 49(3), 238–253. https://doi.org/10.1111/fcsr.12389 reiter, m. (2023). successfully maintaining cfp board registered programs & best practices for engaging diverse students [white paper]. cfp board. https://www.cfp.net/knowledge/reportsand-statistics/diversity-and-womensresearch/how-to-engage-diversestudents samuel, h. s., etim, e. e., shinggu, j. p., nwekemaraizu, u., & bako, b. (2024). diversity and inclusion in higher education. in m. kayyali (ed.), building resiliency in higher education: globalization, digital skills, and student wellness (pp. 239–248). igi global. schaeffer, k. (2024). for women’s history month, a look at gender gains – and gaps – in the u.s. pew research center. https://www.pewresearch.org/shortreads/2024/02/27/for-womens-historymonth-a-look-at-gender-gains-and-gapsin-the-us/ https://www.cnbc.com/2022/02/15/a-lack-of-diversity-still-persists-in-the-financial-planning-industry.html https://www.cnbc.com/2022/02/15/a-lack-of-diversity-still-persists-in-the-financial-planning-industry.html https://www.cnbc.com/2022/02/15/a-lack-of-diversity-still-persists-in-the-financial-planning-industry.html https://www.pewresearch.org/race-and-ethnicity/2023/12/04/wealth-surged-in-the-pandemic-but-debt-endures-for-poorer-black-and-hispanic-families/ https://www.pewresearch.org/race-and-ethnicity/2023/12/04/wealth-surged-in-the-pandemic-but-debt-endures-for-poorer-black-and-hispanic-families/ https://www.pewresearch.org/race-and-ethnicity/2023/12/04/wealth-surged-in-the-pandemic-but-debt-endures-for-poorer-black-and-hispanic-families/ https://www.pewresearch.org/race-and-ethnicity/2023/12/04/wealth-surged-in-the-pandemic-but-debt-endures-for-poorer-black-and-hispanic-families/ https://doi.org/10.5325/jgeneeduc.60.2.0117 https://doi.org/10.5325/jgeneeduc.60.2.0117 https://www.investopedia.com/financial-advisor/fa-profession-accountable-lack-diversity/ https://www.investopedia.com/financial-advisor/fa-profession-accountable-lack-diversity/ https://www.investopedia.com/financial-advisor/fa-profession-accountable-lack-diversity/ https://doi.org/10.1007/s10755-020-09517-7 https://doi.org/10.1007/s10755-020-09517-7 https://powertofly.com/up/workiva-develop-your-leadership-skills https://powertofly.com/up/workiva-develop-your-leadership-skills https://doi.org/10.18061/dsq.v32i3.3267 https://doi.org/10.1111/fcsr.12389 https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/how-to-engage-diverse-students https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/how-to-engage-diverse-students https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/how-to-engage-diverse-students https://www.cfp.net/knowledge/reports-and-statistics/diversity-and-womens-research/how-to-engage-diverse-students https://www.pewresearch.org/short-reads/2024/02/27/for-womens-history-month-a-look-at-gender-gains-and-gaps-in-the-us/ https://www.pewresearch.org/short-reads/2024/02/27/for-womens-history-month-a-look-at-gender-gains-and-gaps-in-the-us/ https://www.pewresearch.org/short-reads/2024/02/27/for-womens-history-month-a-look-at-gender-gains-and-gaps-in-the-us/ https://www.pewresearch.org/short-reads/2024/02/27/for-womens-history-month-a-look-at-gender-gains-and-gaps-in-the-us/ financial services review, 33(4) 120 schumacher, s., hill, l., artiga, s., hamel, l., & valdes, i. (2024). five facts about black women’s experiences in health care. kff. https://www.kff.org/racial-equityand-health-policy/issue-brief/five-factsabout-black-womens-experiences-inhealth-care/ u.s. bureau of labor statistics. (n.d.). race and ethnicity in the labor force. https://www.bls.gov/opub/reports/raceand-ethnicity/2022/home.htm u.s. census bureau. (n.d.). quickfacts. https://www.census.gov/quickfacts/ winks, l. (2018). discomfort, challenge, and brave spaces in higher education. in w. leal filho (ed.), implementing sustainability in the curriculum of universities: approaches, methods and projects (pp. 99–111). springer. https://doi.org/10.1007/978-3-31970281-0_7 https://www.kff.org/racial-equity-and-health-policy/issue-brief/five-facts-about-black-womens-experiences-in-health-care/ https://www.kff.org/racial-equity-and-health-policy/issue-brief/five-facts-about-black-womens-experiences-in-health-care/ https://www.kff.org/racial-equity-and-health-policy/issue-brief/five-facts-about-black-womens-experiences-in-health-care/ https://www.kff.org/racial-equity-and-health-policy/issue-brief/five-facts-about-black-womens-experiences-in-health-care/ https://www.bls.gov/opub/reports/race-and-ethnicity/2022/home.htm https://www.bls.gov/opub/reports/race-and-ethnicity/2022/home.htm https://www.census.gov/quickfacts/ https://doi.org/10.1007/978-3-319-70281-0_7 https://doi.org/10.1007/978-3-319-70281-0_7 financial services review, 33(1) 120 crypto investment: the role of investment motivations, investment confidence, and risk perceptions yu zhang,1 khurram naveed,2 and jia qi3 abstract in an era of increasing interest in cryptocurrency, this research delves into the psychological and behavioral factors influencing cryptocurrency investment and future cryptocurrency investment intentions. analyzing data from the 2021 national financial capability study and the investor survey, the study employs two logistic regression models to investigate the effects of investment motivations, risk perceptions, and investing confidence on cryptocurrency investments in a sample of 1,653 american investors. the findings revealed that motivations and investment confidence positively correlate to the choice to engage in cryptocurrency investments and the propensity for future investments. the risk perception of cryptocurrencies acts as a barrier, discouraging current investment behavior and future investment intentions in this class of assets. the integration of psychological factors into the examination of cryptocurrency investment has two important implications: 1) it enhances the accuracy of investor profiling, and 2) it heightens the awareness of motivational factors, enabling financial advisors and planners to provide personalized guidance that addresses the cognitive and emotional motivations associated with investing in a market as volatile as cryptocurrency. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation zhang, y., naveed, k., & qi, j. (2025). crypto investment: the role of investment motivations, investment confidence, and risk perceptions. financial services review, 33(1), 120-141. introduction blockchain is a decentralized digital ledger that records transactions across multiple computers to ensure security and transparency (yli-huumo et al., 2016). the inception of blockchain can be traced back to the introduction of bitcoin in 2009, which was not only the first successful cryptocurrency but also the first application of blockchain technology (nakamoto, 2008). since then, cryptocurrencies have evolved into 1 corresponding author (yuliazhang@ksu.edu). kansas state university, manhattan, ks, usa. 2 kansas state university, manhattan, ks, usa. 3 university of georgia, athens, ga, usa. sophisticated financial instruments that are attracting an increasing number of investors (fang et al., 2022) and may become a key component of financial markets in the future (kyriazis, 2019). particularly, investors can rely on historical information and patterns in the cryptocurrency market to forecast future investment returns due to the relatively inefficient nature of cryptocurrency markets (kyriazis, 2019). investors can also employ trading https://creativecommons.org/licenses/by-nc/4.0/ mailto:yuliazhang@ksu.edu https://creativecommons.org/licenses/by-nc/4.0/ zhang et al. 121 strategies that involve lower levels of risk while still generating profits, in contrast to what would be expected in more efficient markets; however, the market is moving toward greater efficiency (kyriazis, 2019). according to a study conducted by the pew research center, the majority of americans have a basic awareness of cryptocurrency. however, their degree of confidence in investing in these innovations remains relatively low (faverio & sidoti, 2023). most investors who have invested in, traded, or used cryptocurrency report initiating these activities within the past five years (faverio & sidoti, 2023). additionally, nearly half of american cryptocurrency investors have indicated that their investments have underperformed relative to their expectations (faverio & sidoti, 2023). prior research, although limited, has indicated that several factors, including risk tolerance, financial literacy overconfidence, investment experience, socio-demographic characteristics, and public sentiments, have contributed to the actual investment behavior of individuals in cryptocurrency (almeida & gonçalves, 2023; anderson & lawson, 2023; faverio & sidoti, 2023; kim et al., 2023; zhao & zhang, 2021). nonetheless, there is limited knowledge regarding the factors that shape individuals' behaviors in cryptocurrency investments and intentions on future cryptocurrency investments. to address this gap, this study draws on the selfdetermination theory (sdt) and the theory of planned behavior (tpb) as guiding frameworks. these theories provide a structured approach to investigating the determinants of cryptocurrency investments by integrating various factors, such as investment motivations, investment confidence, and risk perceptions. by applying these established psychological theories, this research seeks to examine and build upon the existing knowledge of financial behaviors and psychological motivations and extend these insights to the context of cryptocurrency investments. the results indicate that motivations such as pursuing short-term profits, seeking entertainment, and learning about investing significantly increased the likelihood of investing in cryptocurrency and future intentions to invest. conversely, perceptions of risk associated with cryptocurrency were linked to a decreased likelihood of making such investments and lowered future investment intentions. gaining comprehension of these variables can yield significant insights into the psychological impact on investment decisions. literature review and theoretical background investment motivations the investment motivations explored in this study encompass a variety of factors, including earning short-term profits, securing long-term gains, engaging in investment for entertainment, excitement, or fun, participating due to peer influence or social activities, investing to support personal values or making a societal impact, and investing to learn more about investing itself. self-determination theory, a macro theory developed by deci and ryan (2012), focuses on intrinsic and extrinsic motivations guiding individuals’ behaviors. these motivations, characterized by the psychological needs for autonomy, competence, and relatedness, may potentially influence investors’ financial behaviors, including conducting cryptocurrency investments. according to deci and ryan (2012), extrinsic motivation arises when individuals engage in behavior with the intention of obtaining external rewards or avoiding penalties. investors start investing because of the expectation of earning short-term profits and long-term gains, which are motivated by external financial profits. determining the price and volatility of cryptocurrencies is challenging and differs from traditional financial products (kim et al., 2022b). the apparent independence of cryptocurrency market price fluctuations from other asset classes signifies a good opportunity for portfolio diversification (bouri et al., 2017). investors with extrinsic motivation may use cryptocurrency as a diversifier in their portfolio, which is also in line with traditional economic rationality. this suggests that individuals make investment decisions on cryptocurrency based on the expected utility or outcomes, aiming to maximize benefits while minimizing risks. financial services review, 33(1) 122 in contrast to extrinsic motivation, intrinsic motivation involves engaging in an activity for its inherent interest and enjoyment, where the activity itself is rewarding. the speculative nature of cryptocurrency markets is distinguished by high volatility levels and potentially highly profitable for diversified portfolios (bouri et al., 2017). investors looking for the excitement of high-stakes games, like the wild swings in cryptocurrency values, may be attracted to invest in cryptocurrency. additionally, intrinsic motivation emphasizes the psychological need for competence, which involves effectively interacting with the environment and exhibiting mastery over tasks (deci & ryan, 2012). investing in cryptocurrencies could also be associated with the intrinsic desire to acquire knowledge and develop expertise in this new financial product, thereby bolstering one’s sense of competence. relatedness encompasses the intrinsic need to establish connections with people and to be part of a community (deci & ryan, 2012). investing motivated by peer influence and connecting with others closely satisfies the psychological need for belonging. additionally, peers can significantly influence whether a particular behavior is performed in alignment with the subjective norms described in the theory of planned behavior. recent studies have shown rising evidence that peers could influence investment behaviors (i.e., bursztyn et al., 2014; chen & ma, 2017; delfino et al., 2016; ouimet & tate, 2020). investors in the cryptocurrency market frequently exhibit irrational behavior by uncritically following the decisions of others without depending on their judgment (ballis & drakos, 2020). using an experimental investigation, delfino et al. (2016) found that the investment selections of participants demonstrated significant association with the choices of their peers, primarily driven by the influence of social information, which reflects the behavior of a larger group. taking a more specific angle, ouimet and tate (2020) employed employee stock purchase plans and found that peer networks guide investment behaviors, resulting in improved investment decisions. bursztyn et al. (2014) showed that both social learning (gaining information from peers) and social utility (desire to align with peers) are driving factors behind investment choices. in addition to disseminating information, prior research has also demonstrated that peers can influence various financial behaviors other than selecting risky assets, such as charitable donation decisions (lieber & skimmyhorn, 2018), the adoption of insurance policies (cai et al., 2015), and retirement savings and enrollment decisions (beshears et al., 2015; duflo & saez, 2003). drawing upon established concepts from selfdetermination theory and extant literature on factors that are closely related to investment decision-making, this study contributes to the body of literature by proposing that: h1a: investment motivations are associated with current investment in cryptocurrency. h1b: investment motivations are associated with the intention to invest in cryptocurrency in the future. investment confidence self-efficacy refers to an individual’s belief in their ability to conduct the behaviors necessary to produce specific performance outcomes (bandura, 1977). as suggested in the theory of planned behaviors, individuals’ beliefs about their ability to control and perform a specific behavior can significantly impact their actions (bandura, 1986). this study defines investment confidence as a particular kind of self-efficacy related to the belief that one can invest comfortably, which is consistent with the perceived behavior control component of the theory of planned behavior. investment confidence may serve as a key determinant of an individual’s willingness to engage in investing activities and make well-informed choices. despite the scarcity of studies specifically focused on investment confidence and cryptocurrency investments, literature documents a significant association between the level of financial self-efficacy and investment behaviors, such as personal finance product selections (farrell et al., 2016) and volatile financial asset ownership (chatterjee et al., 2011). financial self-efficacy is positively associated with the level of risk individuals are willing to assume within their investment portfolios (montford & goldsmith, 2016), positively influence mutual fund investment (mishra et al., 2022), and wealth zhang et al. 123 accumulation across time (chatterjee et al., 2011). individuals with higher self-efficacy have a greater propensity for entrepreneurial investment, characterized by being aggressive (cassar & friedman, 2009). a higher level of self-efficacy, which could be reflected in heightened investment confidence, might manifest in allocating funds towards high-risk financial instruments, such as stocks, bonds, and mutual funds (chatterjee et al., 2011), and potentially cryptocurrencies as the trend evolves. however, it is important to recognize the unique aspects of cryptocurrency, such as the volatility and the technological complexity. therefore, this study evaluates whether the established positive relationship between investment confidence in volatile asset ownership extends to cryptocurrency investment and future investment intentions. the following hypotheses are proposed: h2a: investment confidence is positively associated with current investment in cryptocurrency. h2b: investment confidence is positively associated with the intention to invest in cryptocurrency in the future. risk perception risk perception in the investment domain can be defined as a person’s subjective judgment about the potential losses and uncertainty associated with a particular investment (weber & milliman, 1997). this subjective perception is not always aligned with the objective probability of risks involved but is intertwined with the psychological, conative, and emotional factors (slovic, 1987). risk perception also often incorporates the likelihood of the potential loss and its perceived severity (kahneman & tversky, 1979). in this study, the perception of risk associated with cryptocurrencies can be understood as an attitude component of the theory of planned behaviors, which indicates the extent to which individuals evaluate engaging in cryptocurrency investment negatively (ajzen, 1985). risk perception has been recognized as a key determinant in the growing field of cryptocurrency adoption. previous research has documented a negative relationship between risk perception and the actual adoption of cryptocurrency (anser et al., 2020; dabbous et al., 2022; hasan et al., 2022). for instance, dabbous et al. (2022) found that perceived risk is negatively associated with the willingness to adopt cryptocurrency. similarly, anser et al. (2020) provided supportive evidence that perceived risk negatively correlates with cryptocurrency adoption and moderates the relationship between intentions and actual cryptocurrency adoption among individuals. additionally, hasan et al. (2022) found similar results, with perceived risk negatively influencing behavioral intention for cryptocurrency adoption among university students. although the relationship between risk perception and investment in cryptocurrency has yet to be extensively studied, research does indicate a negative relationship between risk perception and investment decisions more broadly (byrne, 2005; nguyen et al., 2019). nguyen et al. (2019) found that client risk perception negatively correlates with risky-asset allocation decisions. similarly, byrne (2005) found a negative association between risk perception and investment decisions. efforts have been made to understand the role of risk perception. yet, the relationship between risk perception and cryptocurrency as an investment and intentions to invest in cryptocurrency in the future remains inadequately explored. given the existing evidence of a generally negative relationship between risk perception and making risky investments, this study proposes that: h3a: the risk perception of cryptocurrency is negatively associated with current investment in cryptocurrency. h3b: the risk perception of cryptocurrency is negatively associated with the intention to invest in cryptocurrency in the future. control variables previous studies found that investment knowledge, particularly subjective financial knowledge, is positively associated with holding cryptocurrency as an investment (zhao & zhang, 2020). socio-demographic variables also play a financial services review, 33(1) 124 crucial role in shaping investment behaviors, especially in the context of risky assets. younger individuals often accept more risk (jianakoplos & bernasek, 1998), whereas women generally prefer safer investments (powell & ansic, 1997). additionally, simms (2014) highlights the differential use of financial advice among female investors, revealing distinct profiles based on varying socio-economic backgrounds and risk perceptions. cultural and ethnic backgrounds impact risk preferences (harrison et al., 2002; gutter et al., 1999), and marital status and children can lead to more conservative choices (joo & grable, 2004; van rooij et al., 2011). financial literacy through education and higher household incomes are linked to riskier portfolios (van rooij et al., 2011; yao et al., 2004), and more investment experience often correlates with a greater likelihood of choosing high-risk assets (corter & chen, 2006). theoretical background behavioral finance challenges the traditional assumptions of core economic models by acknowledging that individuals are not always rational in their decision-making processes. these traditional assumptions, outlined by fama (1970) in his discussion on the efficient market hypothesis, often include the notions of rationality, complete information, and market efficiency. these principles suggest that individuals always make decisions that maximize their utility, operating under the premise that all available information is reflected in market prices. however, behavioral finance highlights the importance of incorporating psychological and sociological factors into the decision-making process, thus recognizing the limitations of these traditional assumptions (bakar & yi, 2016). through the lens of self-determination theory, the decision to invest in cryptocurrencies could be driven by intrinsic and extrinsic factors and is an act of self-determination and motivation. on the other hand, the theory of planned behavior provides a comprehensive psychological framework that is utilized to understand and predict the actions of individuals during the decision-making process (ajzen, 1985). according to the tpb, an individual’s actual behavior and the intention to perform a behavior are influenced by three primary factors: subjective norms, behavioral control, and attitude. the investment motivation resulting from a desire to interact with others and engage in social activity functions as an illustration of subjective norms, which is also connected with the need for relatedness in self-determination theory. perceived behavioral control concerns an individual’s assessment of the ease or difficulty of carrying out a particular behavior in the theory of planned behavior, which can be seen as a form of self-efficacy (bandura, 1977, 1986). investment confidence, defined in this study, refers to the perception of one’s capability to confidently make investment decisions. a higher degree of confidence in making investment decisions and a greater perceived control over investing behaviors increase the probability that an individual will engage in investments such as cryptocurrency. attitude, another important component of the tpb, refers to the degree of favorable or unfavorable evaluation of performing the behavior (ajzen, 1985). when investing in cryptocurrencies, an individual’s intention to invest may be significantly influenced by their perception of the risks involved. if they perceive these risks as very high or extreme, their evaluation of the potential benefits of investing in cryptocurrencies may become less favorable. this heightened awareness of the risk involved may lead to a more cautious stance towards making such investments. both theories collectively aid in understanding how intrinsic and extrinsic motivations, alongside attitudes toward investing behavior, subjective norms, and perceived behavioral control, might influence investment decisions. the conceptual framework employed in the present study is illustrated in figure 1. utilizing this integrated conceptual framework facilitates a comprehensive exploration of financial and nonfinancial factors that could influence individuals’ decisions to invest in cryptocurrencies. zhang et al. 125 figure 1. conceptual framework methodology dataset and sample this study employs the 2021 national financial capability study (nfcs) state-by-state survey combined with the 2021 national financial capability study investor survey. funded by the finra investor education foundation and conducted by fgs global, the 2021 national financial capability study (nfcs), modified from the 2018 version, surveyed 27,118 adults aged 18 and above who have validated and updated demographic characteristics across the united states. the investor survey comprised 2,824 respondents from the state-by-state survey who had investments outside of retirement accounts and were either the primary or joint decision-makers in their households. given that current and future cryptocurrency investments are the main variables we examine in this study, investors who responded "don't know" or "prefer not to say" in response to questions about current and future cryptocurrency investments, as well as investment motivations, were excluded from the analytical sample, leaving a final sample size of 1,653 investors. measurement cryptocurrency investments and future cryptocurrency investment. this study’s two dependent variables on cryptocurrency investment were binary coded based on investors’ responses to the question, "have you invested in cryptocurrencies, either directly or through a fund that invests in cryptocurrencies?" responses were coded as 1 if the respondent confirmed "yes" and 0 if the respondent responded "no." the second dependent variable, future cryptocurrency investment, was also binary coded. investors were asked to indicate whether they are considering investing in cryptocurrencies in the future. those who responded "yes" to the question were considered to have an interest in potential cryptocurrency financial services review, 33(1) 126 investments and were assigned a value of 1. in contrast, those who responded "no" were assigned a value of 0. the two cryptocurrencyrelated questions were asked independently. investment motivations. to measure the various motivations for investment decision-making effectively, we operationalized investment motivations into binary terms based on selfassessed responses to the question: “how well does each of the following describe why you invest?” respondents were presented with six statements in the original questionnaire: “to make money in the short term,” “to make money in the long term,” “for entertainment/excitement/fun/playing a game,” “my peers are doing it/social activity/connecting with others,” “to make a difference in the world/support values i care about/be socially responsible,” and “to learn about investing.” the investment motivation variables were coded as '1' when respondents selected “describes somewhat” or “describes very well,” and as '0' for “does not describe at all.” this binary coding method effectively aligns with self-determination theory, helping to identify the presence of specific extrinsic or intrinsic motivations for investment decisions. it is important to note that 99.03% of the sample in this study hold long-term gains as an investment motivation, which suggests that this specific motivation may lack discriminatory power. consequently, the motivation for longterm gains was excluded from the regression analyses. investment confidence. investment confidence was measured based on a self-assessment in the survey: “how comfortable are you when it comes to making investment decisions?” with a 10-point likert-type scale that ranged from 1 = not at all comfortable to 10 = extremely comfortable. risk perceptions. investors were required to provide their thoughts regarding the level of risk associated with cryptocurrency as an investment. a 5-point likert scale was used with scores of 1 = not at all risky, 2 = slightly risky, 3 = moderately risky, 4 = very risky, and 5 = extremely risky. control variables. this study also used sociodemographic characteristics as control variables, including age, gender, ethnicity, marital status, educational attainment, presence of dependent child(ren), household income level, homeownership, and possession of a nonretirement account with a high investment total value exceeding $100,000. additionally, investors were asked to describe the amount of financial risk they are willing to take when they save or make investments. the variables were reverse coded, with 1 = not willing to take any financial risks, 2 = take average financial risks expecting to earn average returns, 3 = take above average financial risks expecting to earn above average returns, and 4 = take substantial financial risks expecting to earn substantial returns. objective financial knowledge was quantified by the number of correct responses to 11 multiple-choice questions on investing, with a range of 0 to 11. subjective financial knowledge was measured through a self-assessment of overall investment knowledge on a 7-point likert scale, with 1 = very low and 7 = very high. for details on these financial knowledge questions, please refer to appendix a. analyses two logistic regression models were utilized to analyze the outcomes centered on this study: investors’ current investments in cryptocurrency and future intentions to invest in cryptocurrency. the independent variables incorporate measures of investment motivations, as informed by selfdetermination theory, to capture the psychological needs that may influence an individual’s investment behavior. additionally, components from the theory of planned behavior were integrated, with the perceived risk of cryptocurrency representing the attitude component and investment confidence reflecting perceived behavioral control. logit(𝑝) = log ( 𝑝 1 − 𝑝 ) = 𝛽0 + 𝛽1 ∗ 𝑀𝑜𝑣𝑖𝑎𝑡𝑖𝑜𝑛 + 𝛽2 ∗ 𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 + 𝛽3 ∗ 𝑅𝑖𝑠𝑘 𝑃𝑟𝑒𝑐𝑝𝑡𝑖𝑜𝑛 + 𝛽4 ∗ 𝑋 where, 𝑝 is the probability that (1) the investor was invested in cryptocurrency and (2) intended to do so in the future; 𝑝 1−𝑝 is the odds of the event; 𝑋 is the matrix of demographic control variables. zhang et al. 127 additionally, standardized odds ratios were computed for all independent variables to facilitate comparisons between their effects on current cryptocurrency ownership and future intentions to invest in cryptocurrency. results descriptive results an overview of sample descriptive statistics can be found in table 1. almost one out of four respondents reported having invested in cryptocurrencies, either directly or through a fund that invests in cryptocurrencies (22.32%), while three out of ten (30.79%) reported they were considering investing in cryptocurrencies in the future. a notable proportion of the respondents, 68.18% of the sample investors, expressly indicated that making short-term profit was their motivation. nearly all investments (99.03%) were made with long-term profit in mind. additionally, 28.61% of the investments were made for entertainment purposes, and 20.39% of the investors invested for the motives of social activity. a notable 40.17% of investors started investing to support personal values or effect social change, and 60.13% invested to learn about investment. the average score for assessing the risk associated with cryptocurrencies as an investment was 4.05, ranging from 1 to 5. the mean score of investment confidence, as measured on a scale of 1 to 10, was 7.05. for investment risk tolerance, 54.51% sought average financial risks with the expectation of earning average returns, and 9.80% desired substantial financial risks with the expectation of earning substantial returns. more than half of the investors included in the sample were male (64.79%). most of the investors, 83.55%, were whites. slightly more than 73% of investors had completed college and received a bachelor’s degree or above, whereas nearly a quarter (24.62%) had dependent(s), and over half (66.18%) were married. in terms of age, the proportion of those aged 65 and older was the highest (41.56%), while the proportion of those aged 18 to 24 was the lowest (2.78%). as for wealth, six out of ten (59.77%) respondents have non-retirement investment accounts with a total value of $100,000 or more. in this study’s sample of investors, the proportion of households with incomes below $35,000 was 9.98%, while the proportion of households with incomes exceeding $150,000 was 18.39%. the majority (84.21%) were homeowners. logistic regression results tables 2 and 3 present the logistic regression models’ main findings on current investment in cryptocurrency and future cryptocurrency investment intentions. all hypotheses proposed in this study were supported. full results for current investment in cryptocurrency showed that investors motivated by extrinsic motivations, such as making money in the short term (odds = 1.73, p < 0.01), experienced a 73% increase in the odds of investing in cryptocurrency than those not motivated by short-term gains. investors who are primarily motivated by intrinsic motivations for entertainment (odds = 2.32, p < 0.001) or gaining knowledge for investment (odds = 1.71, p < 0.05) had a 132% and 71% increase, respectively, in the odds of investing in cryptocurrency compared to their counterparts not motivated by these specific motivations. among these motivations, entertainment showed the strongest effect on the odds of investing in cryptocurrency, as reflected in the standardized odds ratios. the perceptions of cryptocurrency-specific risks negatively correlated with cryptocurrency investment (odds = .56, p < 0.001). specifically, as the perception of risk increased, the odds of investing in cryptocurrencies decreased by 44%. investment confidence was positively associated with cryptocurrency investment (odds = 1.21, p < 0.01). this means that for a unit increase in investment confidence, the odds of investing in cryptocurrency increase by 21%. among the key independent variables, investment confidence had the most significant impact on the odds of investing in cryptocurrency, as indicated by the standardized odds ratios. financial services review, 33(1) 128 table 1. descriptive statistics (n = 1,653) variable mean/% std. dev. min max cryptocurrency investment current cryptocurrency investment 22.32% future investment intention 30.79% investment motivation short-term gains 68.18% long-term gains 99.03% entertainment 28.61% peers influence 20.39% support values 40.17% learning investing 60.13% crypto risk perceptions 4.05 0.99 1 5 investment confidence 7.05 2.01 1 10 objective financial knowledge 5.65 2.54 0 11 subjective financial knowledge 4.90 1.30 1 7 investment risk tolerance not willing to take any financial risks 7.74% average financial risks average returns 54.51% above average financial risks above average returns 27.95% substantial financial risks substantial returns 9.80% sociodemographic variables male 64.79% whites 83.55% degree holder 73.50% has dependent(s) 24.62% married 66.18% age category age18to24 2.78% age25to34 8.23% age35to44 13.31% age45to54 12.89% age55to64 21.23% age65+ 41.56% wealth factor high investment balance (>$100,000) 59.77% income level less than $35,000 9.98% $35,000-$49,999 9.20% $50,000-$74,999 19.06% $75,000-$99,999 19.78% $100,000-$149,999 23.59% $150,000 and above 18.39% homeownership 84.21% investors willing to take above-average risks (odds = 2.83, p < 0.05) and substantial risks (odds = 4.23, p < 0.01) were more likely to invest in cryptocurrency. males (odds = 1.83, p < 0.05) and those with financial dependents (odds = 1.61, p < 0.05) were more likely to invest in cryptocurrency. additionally, compared to elder age groups (specifically those aged 65 and zhang et al. 129 above), investors are generally more likely to invest in cryptocurrency 4 . compared to those who earned $150,000 and above, investors earning $75,000 to $100,000 were less likely to invest in cryptocurrency (odds = .57, p < 0.05). age and investment risk tolerance emerged as the most influential factors in current cryptocurrency investing, as indicated by the standardized odds ratios. table 2. logistic regression on current cryptocurrency investment current cryptocurrency investment std. or or se z p>z investment motivations short-term gains 1.29 1.73 0.35 2.73 ** entertainment 1.46 2.32 0.41 4.71 *** peers influence 1.02 1.05 0.23 0.21 support values 0.97 0.93 0.17 -0.38 learning investing 1.30 1.71 0.36 2.54 * crypto risk perceptions 0.56 0.56 0.05 -7.08 *** investment confidence 1.47 1.21 0.08 2.94 ** objective financial knowledge 1.14 1.05 0.04 1.38 subjective financial knowledge 1.01 1.00 0.09 0.04 investment risk tolerance (ref: not willing to take any) average financial risks average returns 1.47 2.18 1.00 1.69 above average financial risks above average returns 1.60 2.83 1.32 2.24 * substantial financial risks substantial returns 1.54 4.23 2.08 2.93 ** sociodemographic variables male 1.33 1.83 0.34 3.21 * whites 0.93 0.83 0.17 -0.94 degree holder 0.74 0.51 0.09 -3.70 *** has dependent(s) 1.23 1.60 0.31 2.44 * married 0.85 0.71 0.14 -1.74 age category (ref: age 65+) 18-24 1.32 5.38 2.54 3.57 *** 25-34 1.80 8.54 2.81 6.53 *** 35-44 1.74 5.12 1.52 5.48 *** 45-54 1.73 5.15 1.44 5.85 *** 55-64 1.44 2.42 0.65 3.31 * wealth factor high investment balance 0.92 0.84 0.15 -0.95 income level (ref: $150,000+) less than $35,000 0.98 0.93 0.32 -0.21 $35,000-$49,999 1.17 1.71 0.56 1.64 $50,000-$74,999 0.87 0.70 0.20 -1.25 $75,000-$99,999 0.80 0.57 0.15 -2.13 * $100,000-$149,999 0.90 0.79 0.19 -1.00 homeownership 0.93 0.82 0.18 -0.90 intercept 0.04 0.03 -4.03 *** note: chi2(29) = 680.26***. pseudo r2 = 0.3875. * p < 0.05, ** p < 0.01, *** p < 0.001. 4 the predicted probability of current cryptocurrency investment peaks at approximately 35.81% in the 2534 age group and then decreases as age increases, suggesting a nonlinear relationship in the probability of investing in cryptocurrency across age groups. the full results on predicted probability for each age category are available upon request. financial services review, 33(1) 130 full results for future intentions to invest in cryptocurrency revealed similar patterns for investment motivations. the motivations for short-term profit ambitions (odds = 2.13, p < 0.001), entertainment and excitement (odds = 1.97, p < 0.001), and learning purpose (odds = 2.18, p < 0.001) were each positively and significantly associated with increased odds of considering investing in cryptocurrency in the future. specifically, these motivations were linked to 113%, 97%, and 118% higher odds of planning to invest in cryptocurrency, respectively. the standardized odds ratio also confirmed that the motivation for learning investment showed the strongest effect on future intentions to invest in cryptocurrency. the more perceived risks associated with cryptocurrency, the lower the odds of future cryptocurrency investment intentions (odds = .34, p < 0.001). this indicates that as risk perceptions increase, the odds of planning to invest in cryptocurrency decrease by 66%. investment confidence (odds = 1.17, p < 0.05) and objective financial knowledge (odds = 1.10, p < 0.01) were positively associated with 17% and 10% higher odds, respectively, of future cryptocurrency investment intentions. based on the comparison of the standardized odds ratios, among the key independent variables, risk perceptions were confirmed to have the most significant impact on future cryptocurrency investment. investors with above-average (odds = 3.09, p < 0.01) and substantial investment risk tolerance (odds = 5.86, p < 0.001) were more likely to consider cryptocurrencies as future investment options. investors with financial independence were also more likely to show intentions to invest in cryptocurrency (odds = 1.64, p < 0.05). whites (odds = .56, p < 0.01) and college degree holders (odds = .64, p < 0.05) were less likely to invest in cryptocurrency in the future. a substantial investment balance (odds = .55, p < 0.01) was negatively linked with future investment intentions in cryptocurrency. additionally, the age group under 65 consistently showed a positive association with the intention to invest in cryptocurrencies in the future5 . among all the independent variables, the standardized odds ratios indicated that age and investment risk tolerance had the most significant impact on future cryptocurrency intentions. 5 similarly, the predicted probability of future cryptocurrency investment intention peaks at 44.62% in the 25 to 34 age group, indicating the highest likelihood of planning to invest in cryptocurrencies. beyond this peak, the intention to invest in cryptocurrencies declines progressively with increasing age. zhang et al. 131 table 3. logistic regression on future cryptocurrency investment intention future investment intention std. or or se z p>z investment motivations short-term gains 1.42 2.13 0.41 3.92 *** entertainment 1.36 1.97 0.36 3.70 *** peers influence 1.02 1.05 0.24 0.22 support values 1.12 1.27 0.23 1.31 learning investing 1.47 2.18 0.44 3.92 *** crypto risk perceptions 0.35 0.34 0.03 -11.63 *** investment confidence 1.37 1.17 0.08 2.45 * objective financial knowledge 1.28 1.10 0.04 2.60 ** subjective financial knowledge 0.84 0.87 0.09 -1.39 investment risk tolerance (ref: not willing to take any) average financial risks average returns 1.44 2.09 0.86 1.79 above average financial risks above average returns 1.66 3.09 1.31 2.66 ** substantial financial risks substantial returns 1.69 5.86 2.78 3.73 *** sociodemographic variables male 1.18 1.41 0.26 1.87 whites 0.81 0.56 0.12 -2.82 ** degree holder 0.82 0.64 0.12 -2.33 * has dependent(s) 1.24 1.64 0.33 2.47 * married 0.97 0.93 0.18 -0.34 age category (ref: age 65+) 18-24 1.30 4.90 2.64 2.95 ** 25-34 1.80 8.57 2.85 6.46 *** 35-44 1.83 5.95 1.71 6.21 *** 45-54 1.70 4.90 1.29 6.02 *** 55-64 1.49 2.65 0.63 4.11 *** wealth factor high investment balance 0.75 0.55 0.10 -3.34 ** income level (ref: $150,000+) less than $35,000 1.01 1.02 0.35 0.07 $35,000-$49,999 1.10 1.37 0.47 0.92 $50,000-$74,999 0.90 0.76 0.22 -0.97 $75,000-$99,999 0.92 0.81 0.22 -0.79 $100,000-$149,999 0.85 0.69 0.17 -1.50 homeownership 0.91 0.78 0.18 -1.06 intercept 1.00 0.75 0.01 note: chi2(29) = 989.04***. pseudo r2 = 0.4845. * p < 0.05, ** p < 0.01, *** p < 0.001. table 4 presents a comparison of the impact of variables investigated in this study across cryptocurrency ownership and future investment intentions. a significant relationship was identified between objective financial knowledge and future intentions to invest in cryptocurrency, while no such relationship was found with current cryptocurrency ownership. male investors were more likely to currently own cryptocurrency, whereas future investment intentions did not differ by gender. investors with an income of $75,000-$99,999 were less likely to hold cryptocurrency at present, but no significant differences in future investment intentions were observed across income levels compared to the reference group of income greater than $150,000. conversely, white investors and those with high investment account balances demonstrated a lower likelihood of expressing future interest in financial services review, 33(1) 132 cryptocurrency investment, despite no significant difference in their current ownership status. table 4. comparison of effects on cryptocurrency ownership and intention using standardized odds ratios variables ownership intention investment motivations short-term gains 1.29** 1.42*** entertainment 1.46*** 1.36*** peers influence 1.02 1.02 support values 0.97 1.12 learning investing 1.30* 1.47*** crypto risk perceptions 0.56*** 0.35*** investment confidence 1.47** 1.37* objective financial knowledge 1.14 1.28** subjective financial knowledge 1.01 0.84 investment risk tolerance (ref: not willing to take any) average financial risks average returns 1.47 1.44 above average financial risks above average returns 1.60* 1.66** substantial financial risks substantial returns 1.54** 1.69*** sociodemographic variables male 1.33* 1.18 whites 0.93 0.81** degree holder 0.74*** 0.82* has dependent(s) 1.23* 1.24* married 0.85 0.97 age category (ref: age 65+) 18-24 1.32*** 1.30** 25-34 1.80*** 1.80*** 35-44 1.74*** 1.83*** 45-54 1.73*** 1.70*** 55-64 1.44* 1.49*** wealth factor high investment balance 0.92 0.75** income level (ref: $150,000+) less than $35,000 0.98 1.01 $35,000-$49,999 1.17 1.10 $50,000-$74,999 0.87 0.90 $75,000-$99,999 0.80* 0.92 $100,000-$149,999 0.90 0.85 homeownership 0.93 0.91 * p < 0.05, ** p < 0.01, *** p < 0.001. discussion given that the majority of cryptocurrency investors, traders, and users initiated their activities within the past five years (faverio & sidoti, 2023), coupled with a surge in interest in cryptocurrency investments, it is crucial and significantly impactful to explore the potential characteristics of cryptocurrency investors. to the best of our knowledge, there is a gap in the existing literature identifying these characteristics of american cryptocurrency investors. this study seeks to bridge this gap by offering valuable insights, grounded in theoretical foundations, into the factors influencing actual cryptocurrency investment behaviors and future intentions in cryptocurrency zhang et al. 133 investments. findings in this study reveal that investment motivations, investment confidence, and perceptions of cryptocurrency risk exhibit a statistically significant association with the decision-making process of cryptocurrency investments, each aligning with the tenets of sdt and tpb. the results imply a positive association between investment motivations and cryptocurrency investment. investing in cryptocurrencies or intending to do so could be interpreted as an activity driven by these investment motivations. three types of motivations were positively and significantly correlated with both current cryptocurrency investment and future intentions to invest in cryptocurrency. specifically, as selfdetermination theory suggests, the desire for short-term financial gain is an illustration of extrinsic motivations in this study, which stem from external rewards and consequences (deci & ryan, 2012). on the other hand, individuals who invest for excitement and entertainment or to gain knowledge about investing exemplify intrinsic motivations fueled by the personal enjoyment derived from the investment process itself. the inherent volatility of cryptocurrencies may appeal to investors driven by a desire for excitement, as it enables them to participate in cryptocurrency investments that simultaneously serve as a means of entertainment and investment (bouri et al., 2017). specifically, the motivation of entertainment exerted the strongest effect on current cryptocurrency investment, as evidenced by the highest standardized odds ratios among various motivational factors. the desire for enjoyment also correlates significantly with the decision-making process regarding future investment choices, highlighting that the unpredictable nature of cryptocurrency prices could offer a thrilling experience to these investors. additionally, investors who engage in investment activities to gain knowledge may perceive cryptocurrency investment as an appropriate strategy within the dynamic cryptocurrency market. through active engagement in cryptocurrency investment, these investors have the potential to acquire practical experience and strengthen their perceived capability, thereby satisfying the psychological need for competencies. more importantly, the desire to learn about investing emerged as the most significant motivator, strongly aligning with intrinsic motivation as suggested by sdt, having the most substantial impact on future intentions to invest in cryptocurrency, demonstrated by the highest standardized odds ratios. investment confidence is another key variable that could serve as the perceived behavioral control component in the theory of planned behavior. the findings confirmed the strong associations between individuals’ confidence in investing and their behaviors and future intentions to invest in cryptocurrencies. according to the existing body of literature on self-efficacy, individuals’ confidence level in their capability to engage in certain financial behaviors can substantially influence their subsequent actions (bandura, 1977, 1986). previous research has established the association between confidence and a propensity for riskier asset allocation, such as stock, bond, and mutual funds (chatterjee et al., 2011; mishra et al., 2022). the present study’s findings of positive and statistically significant associations between investment confidence and cryptocurrency investment or the intention to invest in the future represent a significant extension of the investment options. the results offer significant evidence that investment confidence plays a crucial role in influencing current investment ownership decisions, particularly with volatile assets such as cryptocurrencies, which demonstrate the most pronounced effect among the key variables analyzed. when making financial decisions, those with high investment confidence may be aware of the price volatility yet remain optimistic about their investment abilities. they may also possess better risk management skills and have the option to diversify their investment portfolios by including cryptocurrencies. more importantly, not only did the likelihood of current cryptocurrency investments increase, but so did the intention to invest in cryptocurrencies in the future because of increased investment confidence. this might be due to the heightened level of aggressiveness in investing (cassar & friedman, 2009) exhibited by investors who are confident and comfortable with their investment decisions. individuals with greater investment confidence may investigate financial services review, 33(1) 134 volatile and trendy investment options, such as cryptocurrencies, expecting to gain beneficial experiences in the future. the theory of planned behavior underscores the significant influence of attitudes toward engaging in a specific behavior. the process of making investment decisions involves weighing the tradeoff between risk and projected returns. consistent with findings reported in the existing literature, there is a negative relationship between the perceived risk and the allocation of funds by investors toward certain assets (aini & lutfi, 2019). as the perceived level of risk associated with cryptocurrencies appears to be higher, the likelihood of individuals investing in these assets decreases accordingly. it is important to note that the general investment risk tolerance was positively associated with cryptocurrency investment. this inconsistency suggests that even though the large potential profits of cryptocurrency attract investors, the risk assessment might discourage them from making such investing decisions if the risk involved with cryptocurrency investment is excessive. as indicated by the strong negative association between risk perceptions of cryptocurrencies and future investment intentions, the perceptions of the riskiness of cryptocurrencies could serve as a barrier to future investment possibilities in cryptocurrencies. although men demonstrated a higher propensity to invest in cryptocurrencies, this did not translate into significant differences in their future investment intentions. individuals with college degrees were less likely to invest in cryptocurrencies, whereas those with financial dependents were more inclined to invest and demonstrated consistent intentions to do so in the future. younger cohorts, especially those aged 25 to 34, were more likely to invest in cryptocurrencies than elder cohorts (those aged 65 and above). this may be due to the fact that older cohorts adopt technologically driven products more slowly (zhang & fan, 2023). due to their unfamiliarity with cryptocurrencies as new investment vehicles, older generations may be reluctant to engage in cryptocurrencies. interestingly, individuals with high investment value are not more likely to explore cryptocurrencies as an appealing option to their portfolio in the future. the volatility and unregulated nature of the cryptocurrency market may discourage individuals with large investment account balances from investing in cryptocurrencies. the current study has several limitations. first, the analysis performed in this study focused on a particular investor subgroup, thereby restricting the generalizability of this result to the broader population. the investigations also used a crosssectional dataset, hence precluding the ability to establish causal relationships in the empirical findings. future research utilizing longitudinal data is necessary to validate the observed association in this study. second, the key variables investigated in the current study were measured using a single self-assessment question. for future research, it might be beneficial to incorporate multiple questions to provide a more comprehensive evaluation of investment motivations, investment confidence, and risk perception of cryptocurrencies. third, although the sample consists of 1,653 american investors from a nationally conducted survey, its diversity may be partially reflected in the global investing population, particularly considering the heterogeneous demographics participating in cryptocurrency investments. further research could continue this line of research by integrating qualitative methods to understand psychological investment motivations and cryptocurrency investments better. implication the findings in this research demonstrate the significant roles of investment motivations, investment confidence, and cryptocurrency risk perceptions in determining investments and future investment intentions. this investigation holds substantial implications for policymakers, financial advisors, and planners. the evidence in this paper intends to highlight that investment motivations for making shortterm gains, entertainment, and learning purposes play significant roles in cryptocurrency investment decisions. individuals who invest because of these motivations are more likely to consider investing in cryptocurrencies in the future. financial advisors and planners are important in assisting clients interested in the zhang et al. 135 volatile and speculative cryptocurrency market. financial advisors should be aware of the psychological influences on decision-making and communicate clearly with their clients by properly defining the risk-return profile of cryptocurrency investments. it is crucial to ensure clients understand how these investments fit their overall financial goals. open discussions about investment motivations could lead to a deeper understanding of the client’s needs and aspirations, enabling advisors to serve them more effectively. given that the motivation to learn investing increases the likelihood of investing in cryptocurrency, financial advisors and planners should provide clients with opportunities to satisfy these psychological needs and enhance investment competency. financial advisors should simplify information on financial instruments to facilitate learning. additionally, when working with clients, financial advisors should also emphasize to their clients the significance of independent research and critical thinking. this approach helps mitigate the potential risks of making investment decisions that are solely motivated by the entertainment and thrill associated with cryptocurrency investments, promoting making informed investment choices. given the importance of investor confidence in cryptocurrency investments, policymakers need to reinforce the market regulation and establish consumer protections to protect investors. with the rise in the popularity of cryptocurrencies, financial counselors and advisors must maintain updated knowledge of both conventional investments and the cryptocurrency market to offer appropriate recommendations. additionally, financial planners and advisors should educate highly confident clients about the opportunities and risks of investing in cryptocurrencies to ensure they are well-informed and do not underestimate the associated risks. the risk perception of cryptocurrency is another major determinant of investment behavior in the cryptocurrency market. due to the fact that risk perceptions depend on individual psychological judgment, policymakers can implement clear and comprehensive disclosure requirements for cryptocurrency platforms to aid investors in forming their opinions on cryptocurrencies. any policies that can help mitigate a sudden loss of investing in a particular cryptocurrency could be attractive to investors. financial advisors should tailor their recommendations and product offerings on financial products to the risk perceptions of their clients. specifically, they should inform prospective cryptocurrency investors of accurate information about cryptocurrency to mitigate misunderstanding. by improving risk understanding, advisors can guide investors toward strategies that better align with their financial goals. future research might also look at the detailed aspects of cryptocurrency that people find very or extremely risky. understanding the reasons for these concerns could aid in understanding why investors feel resistant to investing in cryptocurrency. conclusion cryptocurrencies have emerged as a new asset class in the contemporary landscape of investments. behavioral finance emphasizes the integration of psychological insights into financial practices, which could be applicable to cryptocurrency investments. given the novelty and associated risks of cryptocurrency, the factors influencing an investor’s decision to invest in cryptocurrency still need investigation. this study validates the feasibility of the selfdetermination theory and theory of planned behavior as the theoretical foundation for analyzing the decision-making behaviors of investing in cryptocurrency. the results highlight that investment motivations as psychological factors (i.e., pursuing short-term profits, investing motivated by entertainment and gaming, and learning about investing) are strongly and positively associated with cryptocurrency investments and future intentions to invest in cryptocurrency. investment confidence is positively related to cryptocurrency investments and intentions. on the contrary, as the perceived risk level linked to cryptocurrency increases, investors are less likely to invest in these assets. the results of this study enhance our understanding of the profiles of cryptocurrency investors and offer significant implications for the field. financial practitioners need to raise their awareness regarding the diverse motivations of their clients in order to deliver tailored guidance efficiently. financial practitioners must financial services review, 33(1) 136 gain insight into clients’ psychological investment motivations, particularly those demonstrating heightened interest in volatile investments like cryptocurrency. this understanding will assist financial advisors and planners develop individualized recommendations that meet clients’ overall financial goals and psychological needs. additionally, advisors can assist clients in enhancing their investment confidence while openly and comprehensibly communicating the potential risks linked to cryptocurrency, thereby empowering clients to make well-informed decisions on cryptocurrency investments. references aini, n.s., & lutfi, l. (2019). the influence of risk perception, risk tolerance, overconfidence, and loss aversion towards investment decision making. journal of economics, business & accountancy ventura, 21(3), 401–413. https://doi.org/10.14414/jebav.v21i3.16 63 ajzen, i. (1985). from intentions to actions: a theory of planned behavior. https://doi.org/10.1007/978-3-64269746-3_2 almeida, j., & gonçalves, t. c. (2023). a systematic literature review of investor behavior in the cryptocurrency markets. journal of behavioral and experimental finance, 37, 100785. https://doi.org/10.1016/j.jbef.2022.1007 85 anderson, j. n., & lawson, d. r. (2023). a study of achievement, risk, and cryptocurrency using learned needs theory. journal of financial planning, 36(8). https://www.financialplanningassociatio n.org/learning/publications/journal/aug 23-study-achievement-risk-andcryptocurrency-using-learned-needstheory-open anser, m. k., zaigham, g. h. k., imran rasheed, m., pitafi, a. h., iqbal, j., & luqman, a. (2020). social media usage and individuals' intentions toward adopting bitcoin: the role of the theory of planned behavior and perceived risk. international journal of communication systems, 33(17), e4590. https://doi.org/10.1002/dac.4590 bakar, s., & yi, a. n. c. (2016). the impact of psychological factors on investors’ decision making in malaysian stock market: a case of klang valley and pahang. procedia economics and finance, 35, 319-328. ballis, a., & drakos, k. (2020). testing for herding in the cryptocurrency market. finance research letters, 33, 101210. https://doi.org/10.1016/j.frl.2019.06.008 bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84(2), 191-215. https://doi.org/10.1037/0033295x.84.2.191 bandura, a. (1986). social foundations of thought and action. englewood cliffs, nj, 1986(23-28). https://doi.org/10.5465/amr.1987.43065 38 beshears, j., choi, j. j., laibson, d., madrian, b. c., & milkman, k. l. (2015). the effect of providing peer information on retirement savings decisions. the journal of finance, 70(3), 1161-1201. https://doi.org/10.1111/jofi.12258 bouri, e., molnár, p., azzi, g., roubaud, d., & hagfors, l. i. (2017). on the hedge and safe haven properties of bitcoin: is it really more than a diversifier? finance research letters, 20, 192-198. https://doi.org/10.14414/jebav.v21i3.1663 https://doi.org/10.14414/jebav.v21i3.1663 https://doi.org/10.1007/978-3-642-69746-3_2 https://doi.org/10.1007/978-3-642-69746-3_2 https://doi.org/10.1016/j.jbef.2022.100785 https://doi.org/10.1016/j.jbef.2022.100785 https://doi.org/10.1002/dac.4590 https://doi.org/10.1016/j.frl.2019.06.008 https://doi.org/10.1037/0033-295x.84.2.191 https://doi.org/10.1037/0033-295x.84.2.191 https://doi.org/10.5465/amr.1987.4306538 https://doi.org/10.5465/amr.1987.4306538 https://doi.org/10.1111/jofi.12258 zhang et al. 137 bursztyn, l., ederer, f., ferman, b., & yuchtman, n. (2014). understanding mechanisms underlying peer effects: evidence from a field experiment on financial decisions. econometrica, 82(4), 1273-1301. https://doi.org/10.3982/ecta11991 byrne, k. (2005). how do consumers evaluate risk in financial products? journal of financial services marketing, 10(1), 2136. https://doi.org/10.1057/palgrave.fsm.47 70171 cai, j., janvry, a. d., & sadoulet, e. (2015). social networks and the decision to insure. american economic journal: applied economics, 7(2), 81-108. https://doi.org/10.1257/app.20130442 cassar, g., & friedman, h. (2009). does self‐ efficacy affect entrepreneurial investment? strategic entrepreneurship journal, 3(3), 241-260. https://doi.org/10.1002/sej.73 chatterjee, s., finke, m., & harness, n. (2011). the impact of self-efficacy on wealth accumulation and portfolio choice. applied economics letters, 18(7), 627-631. https://doi.org/10.1080/1350485100376 1830 chen, s., & ma, h. (2017). peer effects in decision-making: evidence from corporate investment. china journal of accounting research, 10(2), 167-188. https://doi.org/10.1016/j.cjar.2016.11.00 2 cohn, s. (2022, july 18). these 10 states are leading america in creating a crypto economy. cnbc. https://www.cnbc.com/2022/07/18/these -are-the-10-states-leading-americascrypto-industry.html corter, j. e., & chen, y. j. (2006). do investment risk tolerance attitudes predict portfolio risk? journal of business and psychology, 20, 369-381. https://doi.org/10.1007/s10869-0059010-5 dabbous, a., merhej sayegh, m., & aoun barakat, k. (2022). understanding the adoption of cryptocurrencies for financial transactions within a high-risk context. the journal of risk finance, 23(4), 349-367. https://doi.org/10.1108/jrf-10-20210169 deci, e. l., & ryan, r. m. (2012). selfdetermination theory. handbook of theories of social psychology, 1(20), 416-436. delfino, a., marengo, l., & ploner, m. (2016). i did it your way. an experimental investigation of peer effects in investment choices. journal of economic psychology, 54, 113-123. https://doi.org/10.1016/j.joep.2016.03.0 05 duflo, e., & saez, e. (2003). the role of information and social interactions in retirement plan decisions: evidence from a randomized experiment. the quarterly journal of economics, 118(3), 815-842. https://doi.org/10.1162/0033553036069 8432 fama, e. f. (1970). efficient capital markets. journal of finance, 25(2), 383-417. https://doi.org/10.2307/2325486 fang, f., ventre, c., basios, m., kanthan, l., martinez-rego, d., wu, f., & li, l. (2022). cryptocurrency trading: a comprehensive survey. financial innovation, 8, 13. https://doi.org/10.1186/s40854-02100321-6 https://doi.org/10.3982/ecta11991 https://doi.org/10.1057/palgrave.fsm.4770171 https://doi.org/10.1057/palgrave.fsm.4770171 https://doi.org/10.1257/app.20130442 https://doi.org/10.1002/sej.73 https://doi.org/10.1080/13504851003761830 https://doi.org/10.1080/13504851003761830 https://doi.org/10.1016/j.cjar.2016.11.002 https://doi.org/10.1016/j.cjar.2016.11.002 https://www.cnbc.com/2022/07/18/these-are-the-10-states-leading-americas-crypto-industry.html https://www.cnbc.com/2022/07/18/these-are-the-10-states-leading-americas-crypto-industry.html https://www.cnbc.com/2022/07/18/these-are-the-10-states-leading-americas-crypto-industry.html https://doi.org/10.1007/s10869-005-9010-5 https://doi.org/10.1007/s10869-005-9010-5 https://doi.org/10.1108/jrf-10-2021-0169 https://doi.org/10.1108/jrf-10-2021-0169 https://doi.org/10.1016/j.joep.2016.03.005 https://doi.org/10.1016/j.joep.2016.03.005 https://doi.org/10.1162/00335530360698432 https://doi.org/10.1162/00335530360698432 https://doi.org/10.2307/2325486 https://doi.org/10.1186/s40854-021-00321-6 https://doi.org/10.1186/s40854-021-00321-6 financial services review, 33(1) 138 farrell, l., fry, t. r., & risse, l. (2016). the significance of financial self-efficacy in explaining women's personal finance behaviour. journal of economic psychology, 54, 85-99. https://doi.org/10.1016/j.joep.2015.07.0 01 faverio, m., & sidoti, o. (2023, april 10). majority of americans aren't confident in the safety and reliability of cryptocurrency. pew research center. https://www.pewresearch.org/shortreads/2023/04/10/majority-ofamericans-arent-confident-in-the-safetyand-reliability-of-cryptocurrency/ gutter, m. s., fox, j. j., & montalto, c. p. (1999). racial differences in investor decision making. financial services review, 8(3), 149-162. https://doi.org/10.1016/s10570810(99)00040-2 harrison, g. w., lau, m. i., & williams, m. b. (2002). estimating individual discount rates in denmark: a field experiment. american economic review, 92(5), 1606-1617. https://doi.org/10.1257/0002828027620 24674 hasan, s. z., ayub, h., ellahi, a., & saleem, m. (2022). a moderated mediation model of factors influencing intention to adopt cryptocurrency among university students. human behavior and emerging technologies, 2022, 1-14. https://doi.org/10.1155/2022/9718920 jianakoplos, n. a., & bernasek, a. (1998). are women more risk averse? economic inquiry, 36(4), 620-630. https://doi.org/10.1111/j.14657295.1998.tb01740.x joo, s., & grable, j. e. (2004). an exploratory framework of the determinants of financial satisfaction. journal of family and economic issues, 25(1), 25-50. https://doi.org/10.1023/b:jeei.00000167 22.37994.9f kahneman, d., & tversky, a. (1979). prospect theory: an analysis of decision under risk. econometrica, 47(2), 263-291. https://doi.org/10.2307/1914185 kim, k. t., hanna, s. d., & lee, s. t. (2023). investment literacy, overconfidence, and cryptocurrency investment. financial services review, 31(2/3), 121-132. http://dx.doi.org/10.2139/ssrn.3953242 kim, g., shin, d., choi, j., & lim, s. (2022b). a deep learning-based cryptocurrency price prediction model that uses on-chain data. ieee access, 10, 56232-56248. https://doi.org/10.1109/access.2022.317 7888 kyriazis. (2019). a survey on efficiency and profitable trading opportunities in cryptocurrency markets. journal of risk and financial management, 12(2), 67. http://dx.doi.org/10.3390/jrfm12020067 lieber, e. m., & skimmyhorn, w. (2018). peer effects in financial decisionmaking. journal of public economics, 163, 37-59. https://doi.org/10.1016/j.jpubeco.2018.0 5.001 mishra, a. k., bansal, r., maurya, p. k., kar, s. k., & bakshi, p. k. (2022). predicting the antecedents of consumers' intention toward purchase of mutual funds: a hybrid pls‐sem‐neural network approach. international journal of consumer studies, 47(2), 563-587. https://doi.org/10.1111/ijcs.12850 montford, w., & goldsmith, r. e. (2016). how gender and financial self‐efficacy influence investment risk taking. international journal of consumer studies, 40(1), 101-106. https://doi.org/10.1111/ijcs.12219 nakamoto, s. (2008). bitcoin: a peer-to-peer electronic cash system. united states sentencing commission. https://www.ussc.gov/sites/default/files/ pdf/training/annual-national-trainingseminar/2018/emerging_tech_bitcoin_ crypto.pdf https://doi.org/10.1016/j.joep.2015.07.001 https://doi.org/10.1016/j.joep.2015.07.001 https://www.pewresearch.org/short-reads/2023/04/10/majority-of-americans-arent-confident-in-the-safety-and-reliability-of-cryptocurrency/ https://www.pewresearch.org/short-reads/2023/04/10/majority-of-americans-arent-confident-in-the-safety-and-reliability-of-cryptocurrency/ https://www.pewresearch.org/short-reads/2023/04/10/majority-of-americans-arent-confident-in-the-safety-and-reliability-of-cryptocurrency/ https://www.pewresearch.org/short-reads/2023/04/10/majority-of-americans-arent-confident-in-the-safety-and-reliability-of-cryptocurrency/ https://doi.org/10.1016/s1057-0810(99)00040-2 https://doi.org/10.1016/s1057-0810(99)00040-2 https://doi.org/10.1257/000282802762024674 https://doi.org/10.1257/000282802762024674 https://doi.org/10.1155/2022/9718920 https://doi.org/10.1111/j.1465-7295.1998.tb01740.x https://doi.org/10.1111/j.1465-7295.1998.tb01740.x https://doi.org/10.1023/b:jeei.0000016722.37994.9f https://doi.org/10.1023/b:jeei.0000016722.37994.9f https://doi.org/10.2307/1914185 http://dx.doi.org/10.2139/ssrn.3953242 https://doi.org/10.1109/access.2022.3177888 https://doi.org/10.1109/access.2022.3177888 http://dx.doi.org/10.3390/jrfm12020067 https://doi.org/10.1016/j.jpubeco.2018.05.001 https://doi.org/10.1016/j.jpubeco.2018.05.001 https://doi.org/10.1111/ijcs.12850 https://doi.org/10.1111/ijcs.12219 zhang et al. 139 nguyen, l., gallery, g., & newton, c. (2019). the joint influence of financial risk perception and risk tolerance on individual investment decision‐making. accounting & finance, 59, 747-771. https://doi.org/10.1111/acfi.12295 ouimet, p., & tate, g. (2020). learning from coworkers: peer effects on individual investment decisions. the journal of finance, 75(1), 133-172. https://doi.org/10.1111/jofi.12830 powell, m., & ansic, d. (1997). gender differences in risk behaviour in financial decision-making: an experimental analysis. journal of economic psychology, 18(6), 605-628. https://doi.org/10.1016/s01674870(97)00026-3 simms, k. (2014). investor profiles: meaningful differences in women’s use of investment advice? financial services review, 23(3), 273-286. https://doi.org/10.61190/fsr.v23i3.3200 slovic, p. (1987). perception of risk. science, 236(4799), 280-285. https://doi.org/10.1126/science.3563507 van rooij, m., lusardi, a., & alessie, r. (2011). financial literacy and stock market participation. journal of financial economics, 101(2), 449-472. https://doi.org/10.1016/j.jfineco.2011.03 .006 weber, e. u., & milliman, r. a. (1997). perceived risk attitudes: relating risk perception to risky choice. management science, 43(2), 123-144. https://doi.org/10.1287/mnsc.43.2.123 yao, r., gutter, m. s., & hanna, s. d. (2004). the financial risk tolerance of blacks, hispanics and whites. financial counseling and planning, 15(1), 51-62. https://doi.org/10.1891/jfcp-18-00072 yli-huumo, j., ko, d., choi, s., park, s., & smolander, k. (2016). where is current research on blockchain technology?—a systematic review. plosone, 11(10), e0163477. https://doi.org/10.1371/journal.pone.016 3477 zhang, y., & fan, l. (2023). an examination of mobile fintech utilization from a stresscoping perspective. journal of financial counseling and planning, 34(3), 354366. https://doi.org/10.1891/jfcp-20220061 zhao, h., & zhang, l. (2021). financial literacy or investment experience: which is more influential in cryptocurrency investment? international journal of bank marketing, 39(7), 1208-1226. https://doi.org/10.1108/ijbm-11-20200552 https://doi.org/10.1111/acfi.12295 https://doi.org/10.1111/jofi.12830 https://doi.org/10.1016/s0167-4870(97)00026-3 https://doi.org/10.1016/s0167-4870(97)00026-3 https://doi.org/10.61190/fsr.v23i3.3200 https://doi.org/10.1126/science.3563507 https://doi.org/10.1016/j.jfineco.2011.03.006 https://doi.org/10.1016/j.jfineco.2011.03.006 https://doi.org/10.1287/mnsc.43.2.123 https://doi.org/10.1891/jfcp-18-00072 https://doi.org/10.1371/journal.pone.0163477 https://doi.org/10.1371/journal.pone.0163477 https://doi.org/10.1108/ijbm-11-2020-0552 https://doi.org/10.1108/ijbm-11-2020-0552 financial services review, 33(1) 140 appendix a objective financial knowledge 1. if you buy a company’s stock… you own a part of the company you have lent money to the company you are liable for the company’s debts the company will return your original investment to you with interest 2. if you buy a company’s bond… you own a part of the company you have lent money to the company you are liable for the company’s debts you can vote on shareholder resolutions 3. if a company files for bankruptcy, which of the following securities is most at risk of becoming virtually worthless? the company’s preferred stock the company’s common stock the company’s bonds 4. in general, investments that are riskier tend to provide higher returns over time than investments with less risk. true false 5. the past performance of an investment is a good indicator of future results. true false 6. over the last 20 years in the us, the best average returns have been generated by: stocks bonds cds money market accounts precious metals 7. what is the main advantage that index funds have when compared to actively managed funds? index funds are generally less risky in the short term index funds generally have lower fees and expenses index funds are generally less likely to decline in value 8. which of the following best explains why many municipal bonds pay lower yields than other government bonds? municipal bonds are lower risk there is a greater demand for municipal bonds municipal bonds can be tax-free 9. you invest $500 to buy $1,000 worth of stock on margin. the value of the stock drops by 50%. you sell it. approximately how much of your original $500 investment are you left with in the end? zhang et al. 141 $500 $250 $0 10. which is the best definition of “selling short”? selling shares of a stock shortly after buying it selling shares of a stock before it has reached its peak selling shares of a stock at a loss selling borrowed shares of a stock 11. if you own a call option with a strike price of $50 on a security that is priced at $40, and the option is expiring today, which of the following is closest to the value of that option? $10 $0 -$10.00 subjective financial knowledge on a scale from 1 to 7, where 1 means very low and 7 means very high, how would you assess your overall knowledge about investing? very low 1 2 3 4 5 6 very high 7 1 2 3 4 5 6 7 48 utilizing experiential learning techniques in a financial planning program: allowing students to learn from themselves steve p. fraser1 abstract we describe and provide several illustrations of experiential learning activities used in an introductory financial planning class as part of an undergraduate finance program. examples include awareness and interview exercises where students are afforded the opportunity to learn and receive course credit for examining and analyzing financial planning situations directly related to the students. we have found the use of these exercises better prepares students for class, enriches class discussions, and stimulates meaningful conversations between students and family members on the importance of financial planning. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation fraser, s. p., (2025). utilizing experiential learning techniques in a financial planning program: allowing students to learn from themselves. financial services review, 33(4), 48-61. introduction the importance of an individual’s understanding of financial matters has never been more vital. lusardi and streeter (2023) examined financial literacy and financial well-being in the u.s. using the latest national financial capability study.2 the authors found financial literacy is “low,” particularly among the young and uneducated. further, they reported survey respondents were “ill-informed about their true level of financial knowledge” and overconfident in their abilities. a low level of financial literacy can lead to poor financial decisions and outcomes including difficulty in budgeting, managing debt, or planning for retirement. as just one example, the impact of a lack of financial knowledge on 1 corresponding author (sfraser@fgcu.edu). florida golf coast university, fort meyers, florida, usa. 2 this research will not delineate between such terms as financial education, knowledge, or literacy. see huston (2010) as one example of a more robust and nuanced discussion of these terms. retirement readiness has never been more apparent. the u.s. department of health and human services (hhs, 2024) reported more than 10,000 individuals were turning 65 every day in the united states. fichtner (2024) suggested the number was more than 11,200 a day as the country meets what the retirement income institute describes as the “peak 65® zone”—the largest surge of retirement age americans turning 65 in our nation’s history. unfortunately, the data suggest most individuals of retirement age are not prepared. the center for retirement research reports the combined median balances for households in individual retirement accounts (iras) and 401(k) accounts for those aged 55-64 was $204,000 in 2022, an amount many might https://creativecommons.org/licenses/by-nc/4.0/ fraser 49 question is sufficient to support a long retirement life expectancy. providing improved financial literacy has proved to be an ongoing challenge. one approach has been to introduce financial education in high schools. the national conference of state legislatures (ncsl, 2022) reports 37 states, guam, puerto rico, and the district of columbia have addressed financial literacy legislation in 2022. implementing financial education initiatives in secondary schools is certainly admirable. however, these efforts are not without their challenges. identifying the specific content to be taught, who will administer the curriculum, and who will “teach the teachers” are essential to any program’s success. one place where we know there is a commitment to financial education is at undergraduate institutions offering financial planning programs. a subset of these schools is affiliated with the certified financial planner board of standards (cfp board), the organization responsible for awarding the certified financial planner™ certification. a major component of earning this mark is through the successful completion of an education program designated by the cfp board as a “board-registered” program. these programs have specific content required for new financial professionals—the cfp board’s principal knowledge topics (pkts). the purpose of this paper is to share best practices in teaching financial planning topics with other financial planning programs. this paper is structured as follows: section 2 provides the context of the course, section 3 discusses how experiential learning is integrated in the course, section 4 details the specific exercises used in the class, section 5 presents the results, and section 6 concludes. one financial planning program’s approach the financial planning program described here takes the form of a financial services concentration within an undergraduate b.s. finance degree program. a concentration is a specific track within the finance major that shares 3 often undergraduate programs will embed general principles of financial planning domain topics in other program courses (e.g. risk management and one or more courses with other tracks. in this degree program, parallel concentrations include real estate finance and financial analysis and management, which all share a principles of investing course. the financial services concentration is a cfp board-registered program and therefore addresses all the cfp board’s pkts. in contrast to some programs, the concentration described here utilizes a standalone financial planning course.3 this course is required for all students in the concentration but is often taken as an elective by students in other finance concentrations. as such, there is a good mix of students who possess basic finance skills yet may have different professional aspirations. the class is structured as a survey of the financial planning content areas across the cfp board’s pkts (e.g. general planning, risk and insurance, investment planning, tax planning, retirement planning, and estate planning). students in the concentration subsequently take an individual course in each topic followed by a required capstone course. as with many introductory undergraduate courses, much of the course grade comes from traditional course examinations, in this case 85%. it is within the remaining 15% of the course grade where students can earn credit through experiential learning opportunities to better facilitate the understanding of financial planning concepts. the course provides a natural laboratory for trying new approaches to teaching financial planning concepts. integrating experiential learning on the first day of each semester, faculty recommend students identify family members or acquaintances they might treat as “clients” throughout the semester. the suggestion is that student learning will be enhanced if they can discuss or apply course topics in a personal setting—essentially experience the course content at a personal level. for example, students will have a more meaningful discussion about retirement planning if they do so with family members or acquaintances approaching retirement age. insurance). this can allow greater flexibility in curriculum management using a six-course sequence in lieu of seven courses. financial services review, 33(4) 50 fink (2013) identifies two problems often seen in college teaching. the first is that learning goals do not go beyond what is described as “understand and remember” type of learning. second, and perhaps more relevant, is teachers have difficulty finding activities beyond traditional lectures and leading discussions. the author suggests “significant learning”—or that learning experience that results in something that is truly significant in terms of the students’ lives. characteristics include process (engaging and high energy) and results (lasting change and value in life). hawtrey (2007) states experiential learning involves active, participating activities, also called situational learning. eyssell (1999) demonstrated this idea with the development of a financial planning practicum. in each case, the goal is to make learning more relevant and impactful for students. the experiential approaches used in this class come in two distinct types of assignments. the first approach we use is what we refer to as an awareness exercise. chowdhry and dholakia (2019) examined the relationship between one’s financial self-awareness and what they term as downstream financial outcomes. their results suggest higher levels of financial awareness are associated with positive financial decisions. in our awareness assignments, the goal is for students to become more aware of the specifics of a financial planning concept by examining and reviewing an instrument or contract where they may be affected but did not necessarily take an active role in the decision to form that contract. in these exercises we look for fink’s (2013) “significant learning.” perhaps the best example of an awareness exercise is seen in the insurance arena—specifically car insurance. nearly all college students have some experience driving automobiles, and students are aware the vehicle owner is required by law to maintain auto insurance coverage. the first class period in the property and casualty module of the course might start with the following dialogue: instructor: many of you drove to campus today. we know that you cannot drive without insurance. what is your coverage and how much do you pay? student: i have full coverage. instructor: what does that mean? student: i don’t know…….and my parents pay for it. so much for awareness. unsurprisingly, many students have no idea what coverage they have nor how much they (or their parents) pay. more detail on this assignment is provided in the exercise section to follow. other awareness exercises used in the class examine risk management (homeowner and flood insurance), investment planning, and tax planning. each of these awareness exercises can be interpreted as an attempt to improve awareness as defined by chowdhry and dholakia (2019). the second experiential approach used in the course involves the use of interview assignments. specifically, students are required to interview family members, friends, or acquaintances (their clients) as a method to initiate an active learning experience. itin (1999) distinguishes between experiential “learning” and experiential “education,” suggesting the former results from a reflection of a direct experience. experiential learning rests with the student and may not involve the teacher. the latter is a philosophy involving the interaction with an instructor and encompassing the larger context of the educational learning environment. our interview assignments seek to incorporate the active learning component that does not involve the teacher by bringing the interview results to enrich the classroom discussion. often, the best classes start with students sharing stories from their interviews with parents, grandparents, or other people of significance. killian et al. (2012) implement an interview exercise in an introductory accounting class. their goal was to incorporate an active, student-centered learning activity in the course. they found the exercise was highly effective in helping students achieve intentional learning in the introductory accounting courses and prompting respect for the profession. the authors also utilized a feedback survey which was adapted for use here. cornell et al. (2013) utilized a structured interview approach in introductory courses at both the undergraduate and mba levels. the authors purport that data can be collected by an inexperienced interviewer in a formal manner and fraser 51 that lack of topic-specific knowledge is not an obstacle during the interview. the authors had students interview people in positions of financial responsibility in various organizations with the intent of having students learn from these individuals about specific accounting topics (e.g. policies, internal controls, fraud prevention). here, we use the feedback students gain from their interviews with their “clients” to frame the classroom discussions. these interviews are what fink (2013) might suggest is a “doing experience.” we utilize interview assignments in both the risk planning and estate planning domains. in sum, the awareness and interview exercises create learning experiences to support active learning. the exercises here we briefly outline each of the exercises used in this course. the specific prompt for each assignment is provided as an appendix. awareness exercises risk planning – auto insurance policy review the auto insurance policy review exercise is an introductory exercise in the property and casualty block. we use this exercise to identify different components of coverage (e.g. property, liability), coverage levels, and the tradeoff between deductible levels and premium rates. we discuss premium factors (e.g. age, car type) and highlight state-specific differences and requirements. as expected, students who pay for their own car insurance appear to be more aware of premium costs than their colleagues whose premiums are paid by their parents. interestingly, they lack more specific knowledge of coverage types, and the role deductibles might play in premium rates. we also link this discussion to the more general financial planning and budgeting concepts (e.g. emergency fund ratio). higher deductibles might suggest a potential requirement for a larger emergency fund. for those students who are fortunate enough to have their parents currently paying their insurance, we often see a renewed sense of appreciation for their situation. (see appendix 1 for specific assignment prompt.) cash & debt management planning credit card statement review like the auto insurance policy review exercise, the credit card statement review exercise is a great experiential activity to expand the students’ understanding of consumer credit. ackert and church (2015) suggest college students’ knowledge of credit cards is inadequate, specifically when considering costs. we examine convenience users who seek credit card reward programs versus credit users who focus on interest rates. students are asked to explore the different rates that may apply (e.g. teaser, purchase, default) and when it might be advantageous to pay an annual fee. it is interesting to see some students will report a detailed analysis justifying paying a high annual fee by describing how they intend to earn benefits valued more. it is but one example where students appear to demonstrate a greater degree of critical thinking solely because the analysis involves something relevant to their daily life (and course content)—their own financial wellness. (see appendix 1 for specific assignment prompt.) investment planning risk tolerance quiz & asset allocation the risk tolerance quiz & asset allocation exercise takes an alternative approach than either the auto insurance policy or credit card statement review exercises. this activity is used in the investment module to examine the difficulty of assessing a client’s individual risk tolerance and establishing the appropriate asset allocation. we see a wide variety of risk tools in the marketplace that generate seemingly similar asset allocations. we ask students to complete two assessments (one more academic and one of their choosing found online). we then ask students to describe what they think should be their individual asset allocation. next, students are asked to assess whether the tools they used were beneficial when they tried to set their allocation. this exercise leads to a robust discussion on the role of asset allocation in portfolio performance as well as the difficulty a planner may encounter with operationalizing a financial plan. how does a planner turn a client’s goals into an objective that results in an asset allocation? this exercise makes the process more meaningful as students try to attempt it for themselves. (see appendix 1 for specific assignment prompt.) financial services review, 33(4) 52 tax planning – irs 1040 tax form scavenger hunt the tax form scavenger hunt exercise takes yet another approach in creating an experiential learning activity. college students have a varied set of work histories and experience with filing taxes. some students have never worked and have no real insight into how income taxes are calculated. alternatively, some students run “side hustles” or are part of the “gig economy” and are somewhat more aware of the complexities of the income tax code. as a survey course, the objective is for students to appreciate how taxpayers determine taxable income (e.g. gross income, exclusions, adjustments (for and from) adjusted gross income), and how deductions differ from credits. this exercise is done prior to a class activity involving use of form 1040 and various schedules. the exercise requires students to complete an online, open resource quiz, which is essentially a scavenger hunt. students are asked to identify specific lines on various tax forms and schedules. while a simple exercise, it requires students to “touch” many of the forms and schedules that are used in the class activity. with this experience, the class activity is a more meaningful exercise as they were required to see how some of the schedules and forms interrelate, even if they did not fully understand how they did so when completing the quiz. perhaps more importantly, the quiz, and subsequent class activity, generate a myriad of questions. students are more engaged in the learning process. (see appendix 1 for specific assignment prompt.) interview exercises risk planning homeowners’ and flood insurance interview this homeowners’ and flood insurance interview focuses on property and casualty insurance (p&c). the objective of this assignment is for students to recognize and appreciate the potential need for, and elements of, various property insurance types. prior to the class period where p&c insurance is introduced, students interview homeowners to learn about the level of coverage, coverage types, premium amounts, and deducible levels chosen by the homeowner. we ask students to complete two interviews with homeowners at different ages and different stages of life so they can get multiple perspectives. the university is situated in the gulf coast region of the southeastern united states, site of two major hurricane landfalls in recent years. to enhance the learning opportunity, we added a flood insurance component to the interview. students now hear first-hand accounts of the risk of natural disasters from homeowners and get a glimpse of how homeowner’s insurance policies and flood insurance policies work (or not) together. these interviews allow for a robust class conversation on the current state of the insurance market and its impact on the housing market. (see appendix 2 for specific assignment prompt.) estate planning – estate planning interview the estate planning interview serves as an introductory exercise for the estate planning module of the course. we see that estate planning is one of the cfp board domains where students have little (or no) exposure prior to coming to class. similar to the homeowners’ and flood insurance interview, we ask students to interview two individuals at different stages of life. we encourage students to ask their “clients” openended questions about what estate planning might mean to those interviewed. the interview then proceeds with specific questions addressing the existence and status of estate planning documents (e.g. wills) and the subject’s awareness of property interests and beneficiary information. the feedback shared by students in class leads to in-depth discussions on individual estate-related issues but also highlights earlier planning concepts discussed in the course. the interviews reinforce the need to “understand the client” and that each client is unique. (see appendix 2 for specific assignment prompt.) results here we report student feedback on the utility of the experiential learning activities used in this class. adapting an instrument discussed in killian et al. (2012), we captured input from students on their perceived usefulness of the awareness and interview exercises. specifically, we asked students to evaluate each exercise on a five-point usefulness scale (very useful/useful/neutral/not useful/not at all fraser 53 useful). the results suggest the auto insurance review is the most useful exercise for learning content with 93% of students finding the exercise either very useful or useful. table 1 reports students’ ranking on the usefulness of each exercise on the understanding of the respective disciple areas of financial planning. table 1. rankings of interview and awareness exercise usefulness percentage of students reporting usefulness of exercises to understanding respective financial planning area. how useful was the _____ exercise in understanding the elements and importance of financial planning concepts? not at all useful /strongly disagree not useful /disagree neutral useful /agree very useful /strongly agree useful or very useful /agree or strongly agree homeowner and flood insurance interview exercise 0% 2% 11% 39% 48% 88% estate planning interview exercise 2% 4% 11% 37% 47% 84% auto insurance policy review awareness exercise 2% 0% 5% 35% 58% 93% risk tolerance quiz/asset allocation awareness exercise 2% 4% 28% 35% 32% 67% credit card statement review awareness exercise 2% 0% 12% 28% 56% 84% irs 1040 tax form scavenger hunt awareness exercise 2% 7% 18% 26% 47% 74% we also asked students to select the most valuable exercises and those that were the least valuable to their learning experience. table 2 depicts the results. when framed this way, panel a shows the most valuable exercise was again the auto insurance review when combining the responses for both most valuable and 2nd most valuable. this was followed closely by the homeowners’ and flood insurance interview. interestingly, the pattern of responses across the spectrum from the least valuable to most valuable varies across exercises. for example, students ranked the auto insurance policy review exercise as most 4 this result might be due to a series of factors. some students have no experience filing taxes. additionally, the data was obtained over a fall and valuable while the risk tolerance quiz exercise was clearly least valuable. in contrast, the distribution of responses for the irs 1040 tax scavenger hunt exercise was more bimodal, with nearly an equal percentage of students ranking it as either the most, or least valuable exercise.4 hawtry (2007) suggests student attitudes matter and that an important aspect of experiential learning is that the learner must find activities meaningful. we asked students, “which of the exercises prompted the most meaningful conversation about the importance of financial planning among your friends and/or family?” as spring semester. it is possible tax planning might be perceived as more useful in a spring semester due to tax filing dates. financial services review, 33(4) 54 expected, the interview exercises ranked higher. however, the results also suggest more than 25% of students had meaningful conversations around the awareness exercises. while students were not instructed to conduct interviews, clearly many had impactful conversations with their course “clients” during these exercises. panel b of table 2 shows the estate planning interview was clearly the most meaningful, with nearly half of students selecting this activity. for many students and families, the estate planning interview is the first time many families have discussed these issues, let alone across generations. there is no better example than when a student returns to class reporting they learned in their interview they are the executor or personal representative of a parent’s estate. suddenly, they now want their parents to get their documents in order. appendix 3 illustrates a reporting rubric that captures an assessment tool that faculty can use or adapt to gain feedback on these experiences. table 2. rankings of interview and awareness exercise value panel a percentage of students reporting value of exercises to understanding respective financial planning area. panel b percentage of students reporting which exercise prompted the most meaningful conversation about the importance of financial planning with interview subjects. panel a panel b exercise least valuable 2nd least valuable 2nd most valuable most valuable most meaningful homeowner and flood insurance interview exercise 14% 11% 25% 13% 25% estate planning interview exercise 11% 29% 14% 18% 46% auto insurance policy review awareness exercise 13% 9% 25% 23% 14% risk tolerance quiz/asset allocation awareness exercise 27% 23% 4% 11% 2% credit card statement review awareness exercise 13% 18% 18% 14% 9% irs 1040 tax form scavenger hunt awareness exercise 23% 11% 14% 21% 4% 100% 100% 100% 100% 100% conclusion the purpose of this paper is to share one program’s use of experiential learning activities in an introductory financial planning course. we found the exercises described here to be useful tools in highlighting the importance of financial planning through “doing” exercises. we also found an additional benefit from the use of these types of activities. we grade these assignments very liberally, primarily on a pass/fail basis. if students put forth a reasonable effort and complete all aspects of the assignment, they will receive full credit. this allows students to earn most of their course grade that is not earned through examinations by simply participating fraser 55 fully in the course. faculty at other programs might consider transitioning a portion of their course grade to exercises and assignments similar to those described here. we have found the use of these exercises better prepares students for class, enriches class discussions, and stimulates meaningful conversations between students and family members on the importance of financial planning. the early data suggests the use of these activities is improving the financial literacy of our students and those with whom they interact. references ackert, l. f., & church, b. k. (2015). credit cards, financial responsibility, and college students: an experimental study. international journal of behavioural accounting and finance, 5(1), 1–26. center for retirement research. (2023). 401(k)/ira holdings in 2022: an update from the scf. https://crr.bc.edu/401kira-holdings-in-2022-an-update-fromthe-scf/ chowdhry, n., & dholakia, u. m. (2019). know thyself financially: how financial selfawareness can benefit consumers and financial advisors. financial planning review, 3(1), e1069. https://doi.org/10.1002/cfp2.1069 cornell, r. m., johnson, c. b., & schwartz, w. c., jr. (2013). enhancing experiential learning with structured interviews. journal of education for business, 88, 136–146. https://doi.org/10.1080/08832323.2012. 659562 eyssell, t. h. (1999). learning by doing: offering a university practicum in personal financial planning. financial services review, 8(4), 293–303. fichtner, j. j. (2024). the peak 65® zone is here: creating a new framework for america’s retirement security (retirement income institute original research #026-2024). fink, l. d. (2013). creating significant learning experiences: an integrated approach to designing college courses (2nd ed.). jossey-bass. hawtrey, k. (2007). using experiential learning techniques. the journal of economic education, 38(2), 143–152. huston, s. j. (2010). measuring financial literacy. the journal of consumer affairs, 44(2), 296–316. itin, c. m. (1999). reasserting the philosophy of experiential education as a vehicle for change in the 21st century. the journal of experiential education, 22(2), 91–98. killian, l. j., huber, m. m., & brandon, c. d. (2012). the financial statement interview: intentional learning in the first accounting course. issues in accounting education, 27(1), 337–360. lusardi, a., & streeter, j. l. (2023). financial literacy and financial well-being: evidence from the us. the journal of financial literacy and wellbeing, 1, 169–198. national conference of state legislatures. (2022). financial literacy 2022 legislation. https://www.ncsl.org/research/financialservices-and-commerce/financialliteracy-2022-legislation.aspx u.s. department of health and human services. (2024). aging. https://www.hhs.gov/aging/index.html https://crr.bc.edu/401k-ira-holdings-in-2022-an-update-from-the-scf/ https://crr.bc.edu/401k-ira-holdings-in-2022-an-update-from-the-scf/ https://crr.bc.edu/401k-ira-holdings-in-2022-an-update-from-the-scf/ https://www.ncsl.org/research/financial-services-and-commerce/financial-literacy-2022-legislation.aspx https://www.ncsl.org/research/financial-services-and-commerce/financial-literacy-2022-legislation.aspx https://www.ncsl.org/research/financial-services-and-commerce/financial-literacy-2022-legislation.aspx https://www.hhs.gov/aging/index.html financial services review, 33(4) 56 appendix 1. awareness exercises risk planning – auto insurance policy review awareness exercise please obtain and upload the declarations page from your auto insurance policy. the purpose of this assignment is to better understand the nature of auto policy coverages and the associated premiums/deductibles etc. please redact any personal information you do not wish to share. please ensure the pages show coverage amounts, deductibles, and premium amounts. if you do not drive and therefore do not have insurance, or do not want to provide information on your coverage, please upload a sample policy that you might find from your research. please address the following questions as you review your policy: 1. have you ever reviewed your policy before this exercise? 2. do you feel you have adequate coverage levels? how do they compare to required minimums? 3. do you have your desired levels for each deductible? 4. have you investigated potential avenues to reduce your premium (e.g. discounts, deductible adjustments)? briefly summarize what you learned about your review of your policy. fraser 57 cash & debt management planning credit card statement review awareness exercise your assignment is to analyze a credit card. if you have a card, please analyze one that you have. if you do not have a card, use this exercise to help you determine which card might be right for you. you will have to go to the "fine print!" specifically: 1. use a couple of sentences to describe what type of credit card user you are. (credit or convenience? why? do you think this will be the same upon graduation?) 2. what is the specific card? (specify network & bank/sponsor--e.g. chase visa) 3. what is the annual fee? 4. what is the interest rate(s)? (balance transfers, purchases, cash advances...) 5. what is the late payment fee? 6. what other fees and/or charges come with the card? 7. what is the credit limit on the account? 8. what is the daily cash advance limit? 9. have you ever obtained your credit report? conclude your review with a summary. if you analyzed a card you currently have, please describe why you originally selected this card. would you select it again? or are you now going to look for a new card? why? investment planning risk tolerance quiz & asset allocation awareness exercise your task is to describe your optimal asset allocation. please visit the following link to take a risk tolerance instrument designed by grable, j. e., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8, 163-181. https://pfp.missouri.edu/research/investment-risk-tolerance-assessment/links to an external site. next, seek an additional risk tolerance questionnaire you find from a basic internet search. 1. print out the results/output from both instruments. 2. determine the asset allocation you think is appropriate for your risk tolerance. provide a pie chart that represents your target asset allocation. summarize how you came to this target asset allocation and comment on the suitability and usefulness of the instruments you used to gauge your risk tolerance. https://pfp.missouri.edu/research/investment-risk-tolerance-assessment/ financial services review, 33(4) 58 tax planning – irs 1040 tax form scavenger hunt awareness exercise this quiz is "open internet." you should find .pdf copies of the 2022 tax forms from irs.gov. you might find it helpful to print the forms necessary to complete the quiz and keep with your notes. 1. if line 33 is greater than line 24 on the 2022 form 1040, you will _____ a. receive a refund b. pay this amount in penalty c. owe this amount in tax d. not be allowed a deduction 2. which of the following are itemized deductions on the 2022 form 1040 schedule a? a. charitable gifts b. mortgage interest c. state and local taxes d. all are potential deductions 3. tip income would be entered on what line on the 2022 form 1040? a. line 2a b. line 5a c. tips are not taxable d. line 1 4. which of the following schedules is used for reporting profit or loss from a business on the 2022 form 1040? a. schedule a b. schedule b c. schedule c d. schedule d 5. checking the box on the 2022 form 1040 designating money to go to the presidential election campaign does not change your tax or refund. a. true b. false 6. find the 2022 form 1040 line 12. assume you are filing "married filing jointly." you would only "itemize" if your itemized deductions _____. a. were less than $25,900 b. exceeded $12,950 c. were less than $12,950 d. exceeded $25,900 7. interest and ordinary dividends are reported on the 2022 form 1040 _____. a. schedule a b. schedule b c. schedule c d. schedule d 8. 2022 form 1040 schedule d line 7 represents __________ a. qualified dividends b. non-qualified dividends c. net long-term capital gains d. net short-term capital gains 9. see the 2022 schedule se . what is the level where no self-employment tax (social security) is due and what is the maximum amount of combined wages subject to social security tax? a. $400; $12,550 b. $3,000; $12,550 c. $400; $147,000 d. $3,000; $147,000 10. home mortgage interest paid are reported on 2022 form 1040 schedule a _____ and ____ be limited. a. line 8a, may not b. line 8a, may c. line 8c, may d. line 8c, may not fraser 59 appendix 2. interview exercises risk planning homeowners’ and flood insurance interview please interview two individuals who own their homes. please select the individuals with whom you will have a meaningful conversation. please ask the following basic questions in your interview: 1. what are their ho coverage levels? 2. what are the deductibles? 3. what are the premiums? 4. do they have replacement or actual cash value coverage (acv)? 5. have they experienced premium increases in the past year? if so, how much? 6. do they have flood insurance? please write a summary for each interview and conclude with a separate paragraph on what you found most interesting about these discussions. estate planning – estate planning interview please interview two individuals concerning their estate planning situation. please select the individuals with whom you will have a meaningful conversation. additionally, please select two individuals who are at different points in their lives (e.g. working & retired). please ask the following basic questions in your interview: 1. client age and marital status. (please do not provide any specific identifying information.) 2. please ask them what estate planning means to them. annotate with just a couple of sentences. please do not lead them. just capture their initial response. 3. please ask what (if any) estate planning documents do they have in place? are they current & valid? (wills, poas, amds, etc.) 4. please ask if they know how any of their significant assets are specifically titled? (fs, jtwros, tic, tie, etc.) 5. please ask if they know whether they have all their beneficiaries specified appropriately for all their financial accounts/insurance? please write a summary for each interview and conclude with a separate paragraph on what you found most interesting about these discussions. financial services review, 33(4) 60 appendix 3. experiential learning activity feedback question not at all useful/strongly disagree not useful/ disagree neutral useful/ agree very useful/strongly agree how useful was the homeowner and flood insurance interview exercise in understanding the elements and importance of risk planning? how useful was the estate planning interview exercise in understanding the elements and importance of estate planning? how useful was the auto insurance policy review awareness exercise in understanding the elements and importance of estate planning? how useful was the risk tolerance quiz & asset allocation awareness exercise in understanding the elements and importance of investment planning? how useful was the credit card statement review awareness exercise in understanding the elements and importance of cash and debt management planning? how useful was the irs 1040 tax form scavenger hunt awareness exercise in understanding the elements and importance of tax planning? fraser 61 please identify those activities you found _______ valuable. most least homeowner & flood insurance interview estate planning interview how useful was the auto insurance policy review exercise risk tolerance quiz & asset allocation awareness exercise credit card statement review awareness exercise irs 1040 tax form scavenger hunt awareness exercise which of the exercises prompted the most meaningful conversation about the importance of financial planning among your friends and/or family? homeowner & flood insurance interview estate planning interview how useful was the auto insurance policy review exercise risk tolerance quiz & asset allocation awareness exercise credit card statement review awareness exercise irs 1040 tax form scavenger hunt awareness exercise please describe any suggestions for additional interview or awareness exercises you think should be included in the course. please provide any additional feedback about the course here: pii: s1057-0810(97)90001-9 from the editor karen eilers lahey this last issue of volume 6 is devoted to the rigorous analysis of individual investor options in terms of taxable and nontaxable fund allocations. the first two articles examine the new choices that individual investors face with their retirement dollars. these include deductible iras, nondeductible iras, the new roth iras, and mutual funds. terry l. crain and jeffrey r. austin provide mathematical models for after-tax accumulations for each of the investment alternatives that considers return, taxable return, time horizon, and tax rates for ordinary income and capital gains. the focus in their article entitled, “an analysis of the tradeoff between tax deferred earnings in iras and preferential capital gains” is on individuals who are currently in the 3 1 percent or higher tax bracket. stephan m. horan, jeffrey h. peterson, and robert mcleod extend the analysis to those individuals whose tax brackets may be lower after they retire. “an analysis of non deductible ira contributions and roth ira conversions” examines taxable mutual fund investments versus nondeductible iras and the desirability of converting existing iras to roth iras. they find that the mutual fund option is less attractive when withdrawal tax rates decline and that the conversion to roth iras works best for those who will remain in the same tax bracket upon withdrawal. several large mutual funds have recently been closed to new investors, and are the subject of the article entitled, “performance of mutual funds before and after closing to new investors”. herman manakyan and kartono liano test the impact of the closing of 27 mutual funds. they find that those funds that are closed do not perform as well after closing as in prior years, which raises the question as to what individual investors should do if they are in such a fund. a canadian perspective on an alternative to united states’ equity enhanced certifi cates of deposit are explained and analyzed by moshe arye milevsky and sharon kim in an article entitled, “the optimal choice of index-linked gics: some canadian evi dence”. these instruments are known as indexed linked guaranteed investment certificates. the authors use concepts from option pricing theory to value them by decomposing the payoff into a zero-coupon bond and call options on the underlying stock index as suggested in a paper by robert brooks in volume 5, number 2. “a simple and effective trading rule for individual investors” addresses a trading rule for individual investors that outperforms a passive buy and hold strategy. laurie prather and william bertin test the effect of the announcement of discount rate changes in predicting market movements over a period of 61 years. ix financial services review 6(4) 1997 ryan b. lee and j. tim query provide a review of the 7th edition of risk manage ment and insurance by s. travis pritchett, joan t. schmit, helen i. doerpinghaus, and james l. atheam. douglas kahl, who is responsible for this column needs additional vol unteers. if you are interested in reviewing a book or a website, please let him know. financial services review volume 32 number 3 (2024) volume 32, no. 3 2024 editor: john e. grable, ph.d. cfp ® university of georgia advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp®, baylor university (emeritus) martin seay, ph.d., cfp ® , kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp ® , san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama jasmine fang, ph.d., massey university, new zealand mark fedenia, ph.d., university of wisconsin stu heckman, ph.d., cfp ® , texas tech university william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea thomas langdon, ph.d., roger william university, bristol, ri terrance martin, ph.d., winston-salem state university mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc lance palmer, ph.d., cfp ® , cpa®, university of georgia abed rabbani, ph.d., cfp ® , university of missouri chris robinson, ph.d., york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida issn online 1057-0810 print 1873-5673 contents grable, john e., from the editor, i-ii. lei, shan & fan, lu. cognitive ability and stock investment among chinese middle-aged and older population. 1-19. zhang, zezhong e., hanna, sherman, & xu, lei. the effect of financial knowledge on workers’ expectation of never retiring. 20-31. pilote, pierre-etienne, boulianne, emilio, & magnan, michel. impact of the financial advisor on clients’ financial outcomes: an integrative model. 32-67. liu, yi, guo, tao, & cheng, yuanshan. global perspectives on the determinants of older adults’ subjective well-being: a comprehensive longitudinal study. 68-82. academy of financial services officers president shawn brayman smb research consulting executive vice president program michelle cull western sydney university vice president finance thanh ngo east carolina university vice president communications kirsten macdonald griffith university vice president international relations jasmine fang massey university vice president marketing & pr cora pettipas hsbc global wealth vice president membership matt goren dalton education, cerifi immediate past president tom potts baylor university editor, financial services review john e. grable, ph.d., cfp® university of georgia directors jason anderson university of kansas norah feng massey university wookjae heo purdue university thomas korankye the university of arizona barry mulholland university of akron mustafa nourallah mid sweden university richard stebbins university of alabama yu (yulia) chang kansas state university past presidents inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 32, no. 3, 2024 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • shinae l. choi, ph.d. university of alabama • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • stu heckman, ph.d., cfp®, texas tech university • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • thomas langdon, ph.d., roger william university, bristol, ri • terrance martin, ph.d., winston-salem state university • mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., cfp®, cpa, ca, york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • brian boscaljon, penn state behrend • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph.d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • lu fan, ph.d., cfp®, university of georgia • patti fisher, ph.d., virginia tech • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph.d. stetson university, deland, fl • philip gibson, ph.d., cfp®, winthrop university • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • john nofsinger, ph.d., university of alaska anchorage • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peerreviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ pii: s1057-0810(99)80011-0 i financial services review, 7(i): 25-44 issn: 1057-0810 copyright © 1998 by jai press inc. all fights of reproduction in any form reserved. term spreads and predictions of bond and stock excess returns dale l. domian and william reichenstein several studies conclude that a long-short term spread, in conjunction with one or more other variables, jointly predict returns on long-term corporate bonds and stocks. we extend these studies by examining the predictive content of intermediate-short term spreads, and by examining regressions of excess returns on 1.5-year to 20-year trea sury bonds. we show that the bond market prices an intermediate-short term spread, and not a long-short spread. we believe individuals should vary their debt-equity mix with the level of a default risk premium or the stock market's dividend yield, and vary their debt portfolios' maturity with an intermediate-short term spread. several studies conclude that a term spread, in conjunction with one or more other vari ables, jointly predict returns on long-term corporate bonds and stocks. fama (1976), startz (1982), shiller, campbell, and schoenholtz (1983), fama (1984), fama and french (1989), and fraser (1995) conclude bond returns vary with a term spread. campbell (1987), fama and french (1989), fama (1990a), schwert (1990), chen (1991), and fraser (1995) con clude stock returns vary with a term spread. related results are presented by keim and stambaugh (1986), fama (1986), fama and bliss (1987), stambaugh (1988), hardouvelis (1994), elton, gruber, and mei (1996), and jensen, mercer, and johnson (1996). most of these studies use a long-short term spread--the spread between a long-term yield and a short-term yield to predict returns on long-term securities. there is also a widely stated but unproven explanation about why the term spread predicts returns. "[t]he spread tracks a term or maturity risk premium in expected returns that is similar for all long-term assets. a reasonable and old hypothesis is that the premium compensates for exposure to discount-rate shocks that affect all long-term securities (stocks and bonds) in roughly the same way" (fama & french, 1989, p. 24). in short, the term spread tracks embedded term risk premiums, which are investors' rewards to bearing interest rate or duration risk. dale l. domian • faculty of business administration, memorial university of newfoundland, st. john' s, nf, canada, a ib 3x5; e-mail: ddomian@morgan.ucs.mun.ca. william reiehenstein • pat and thomas r. powers chair in investment management, hankamer school of business, baylor university, p. o. box 98004, waco, tx 76798-8004; e-mail: bill reichenstein@bayior.edu. 26 financial services review 7(1) 1998 we have two concerns with this literature, which provide the motivation for this paper. first, since regressions have been tested on returns on long-term assets but not returns on shorter-term bonds, no one has adequately tested the suspicion that the term spread tracks the rewards to bearing duration risk. to verify this hypothesis, one must show that the impact of the term spread varies across bond maturities. to be specific, in regressions of bond returns on forecasting variables, the slopes for the term spread should be approxi mately linearly related to bonds' durations. second and more important, we suspected that an intermediate-short spread would do a better job than a long-short spread at (1) tracking embedded term premiums and, there fore, (2) predicting bond and stock returns. a long-short term spread can be separated into a long-intermediate and an intermediate-short spread. some prior studies support the pre ferred-habitat theory (e.g., domian, maness, & reichenstein, 1998; mccallum, 1975, mcculloch, 1975). this theory recognizes that some investors prefer and strongly influ ence the short end of the bond market, while others prefer and influence the long end. it fol lows that an intermediate-short spread could closely track a term risk premium, while a long-intermediate spread may not. for many investors, a bond's duration is a good measure of its risk. these investors dominate the short end of the yield curve--that is, they are the marginal price setters. so, term premiums should rise with maturity in the short end of the yield curve, and an inter mediate-short term spread should proxy for the reward to beating duration risk. in contrast, life insurance companies and defined-benefit pension plans dominate the long end of the yield curve. they do not view long-term bonds as riskier than intermediate-term bonds. in fact, they often prefer long duration bonds because they better match the durations of their liabilities. the upshot is that the long-intermediate spread may not track term risk premi ums. it follows that an intermediate-short spread could be a better measure than a long-short spread of embedded term premiums and thus (if the suspected explanation is correct) a better predictor of bond returns. we extend fama and french (1989) and the other studies (1) by examining the predic tive content of other term spreads, especially intermediate-short spreads, (2) by examining regressions of excess returns on 1.5-year to 20-year treasury bonds, and (3) by discussing the investment implications of the research for individual investors. i. data and methods we closely follow the design of fama and french (1989)--the most widely quoted study in the area--to assure that our novel results are not due to differences in time period or vari able definitions. the dependent variables are the excess returns on treasury bonds, corpo rate bonds, and common stocks. the independent variables are various term spreads, a default spread, and the market's dividend yield. we examine returns on treasury bonds with maturities of 1.5, 2, 3, 5, 7, 10, 15, and 20 years. the corporate bond returns cover aaa, aa, a, baa, and below baa (lg, low grade) bonds. the 1.5through 20-year treasury bond returns rely on estimates of par-bond yields (yields on coupon-bearing bonds selling at par) across maturities that come from ibbotson associates (see coleman, fisher, & ibbotson, 1993 and updates). the appendix to our paper describes the return calculations. the corporate bond returns come from ibbotson term spreads and predictions 27 associates' corporate bond file. each corporate bond series has an average maturity of about 20 years. our regressions use cumulative excess returns over one-month bills. a possible com plication is the coupon effect in average returns on less than one-year treasury securities, described by brooks, levy, and livingston (1989). for 1977-1985, the average return on one-month, coupon-beating notes exceeded the average return on one-month, zero-coupon bills. however, we define excess returns net of the bill return to maintain consistency with fama and french (1989). we adopt the terms t1.5, t2, t3, t5, t7, t10, t15, and t20 for the treasury excess returns and aaa, aa, a, baa, and lg for the corporate bond excess returns. the stock series are excess returns on value-weighted and equal-weighted portfolios of nyse stocks and are available from the center for research in security prices. the excess returns are denoted vw and ew. for the independent variables, we examine alternative term spreads, a default risk spread, and a dividend yield. we examine a long-short term spread, long-tb, the spread between the longest available treasury yield and the one-month treasury bill rate. we also examine several intermediate-short yield spreads. they include 10-tb, 5-tb, 3-tb, and 2-tb, where 10-tb denotes the spread between the 10-year treasury yield and the one-month bill rate. other definitions are similar. coleman, fisher, and ibbotson (1993 and updates) provide the treasury yields, including the longest available treasury yield. fama and french (1989) conclude that the bond market's default yield spread and the stock market's dividend yield capture the impact of the same underlying economic factors. def denotes the default spread between the yield on the composite corporate bond portfo lio and the aaa yield. the dividend yield, d/p, denotes the dividend yield on the value-weighted nyse portfolio. the tests center on regressions of excess bond and stock returns from t to t + t, r(t, t + t), on two independent variables, x(t), known at t, r(t, t + t) = or(t) + ~(t) x(t) + e(t, t + t). the first independent variable is a term spread known at t. the second is either the default spread, def(t), or the market's dividend yield, d(t)/p(t). the results of regressions on a yield spread and d/p were essentially similar to the results of regressions on the same yield spread and def. to save space, we only report the results of regressions on def. the major time period in this study is 1942-1994. in preliminary work, we began by replicating fama and french's (1989) results for the 1941-1987 period. we then changed from aaa-tb, their term spread, to long-tb. we prefer long to aaa because, unlike treasury bonds, corporate bonds have default risk and are usually callable. we also moved the starting date to 1942, since few fully-taxable treasury bonds existed before year-end 1941. these changes did not appreciably affect the replication results. extending the sam ple period through 1994 slightly weakens the statistical results and largely accounts for dif ferences between our long-short term spread results and those of fama and french (1989). our results are presented for both the full sample period, 1942-1994, and the latter half of this period, 1969-1994. we report regressions of excess returns for quarterly and oneto four-year investment horizons (i.e., t = q, 1, 2, 3, or 4). the quarterly and annual regressions use non-overlapping returns. the twoto four-year returns are overlapping 28 financial services review 7(1) 1998 annual observations; these regressions are estimated with the newey and west (1987) pro cedure to inhibit potential statistical problems associated with overlapping observations. ii. business conditions and behavior of the forecasting variables a. autocorrelations table 1 presents summary statistics on one-year excess returns on bonds and stocks. the autocorrelations of the forecast variables contain information about the components of expected excess returns they track. the autocorrelations of dividend yield and the default spread are large at the first-order annual lag, and decay slowly for longer lags. this sup table 1 summary statistics for annual observations on one-year excess returns on the bond and stock portfolios and independent variables including several term spreads, 1942-1994 n-order autocorrelation coefficients mean s.d. 1 2 3 4 5 6 7 8 ti.5 1.01 2.16 -0.02 0.26 -0.33 -0.01 -0.00 0.04 -0.03 0.05 t2 0.88 2.74 -0.03 0.20 -0.33 0.01 -0.00 0.04 0.01 0.04 t3 1.09 3.79 0.04 0.17 -0.27 -0.01 -0.04 0.00 0.02 0.06 t5 1.11 5.06 0.03 0.11 -0.19 0.01 -0.05 -0.03 0.06 0.07 t7 0.98 6.08 0.03 0.06 -0.12 0.01 --0.06 -0.05 0.09 0.06 t10 0.83 7.16 0.04 0.04 -0.07 0.03 -0.06 -0.07 0.10 0.06 t15 0.55 8.31 0.04 0.03 -0.03 0.04 -0.04 -0.08 0.11 0.06 t20 0.32 9.00 0.04 -0.01 -0.04 0.01 --0.04 -0.09 0.11 0.05 aaa 0.36 6.98 0.19 0.01 -0.08 -0.09 -0.16 0.06 0.05 -0.09 aa 0.50 7.05 0.20 -0.07 -0.08 -0.13 -0.10 0.04 0.08 -0.10 a 0.93 7.27 0.22 -0.07 -0.16 0.08 0.01 0.07 0.06 -0.12 baa 1.71 7.33 0.21 -0.14 -0.14 -0.06 0.03 0.11 0.07 -0.12 lg 3.02 10.70 0.21 -0.04 -0.16 -0.05 0.12 0.17 0.04 -0.01 vw 7.23 15.60 -0.06 -0.23 0.10 0.31 0.07 -0.09 0.13 0.04 ew 9.67 21.08 0.04 -0.22 0.01 0.20 -0.04 -0.16 0.06 -0.02 d/p 4.19 1.16 0.72 0.52 0.42 0.35 0.30 0.27 0.26 0.16 def 0.70 0.41 0.61 0.35 0.19 0.21 0.29 0.31 0.25 0.20 long-tb 1.44 1.40 0.51 0.23 0.12 0.19 0.32 0.34 0.11 0.04 10-tb 1.35 1.28 0.46 0.22 0.13 0.22 0.38 0.38 0.10 0.02 5-tb 1.10 1.13 0.39 0.21 0.14 0.23 0.38 0.40 0.11 -0.01 3-tb 0.92 1.00 0.33 0.24 0.17 0.24 0.38 0.42 0.13 -0.01 2-tb 0.78 0.90 0.26 0.26 0.17 0.18 0.31 0.40 0.14 0.02 long-3 0.52 0.75 0.60 0.28 0.10 -0.03 -0.03 0.06 0.05 0.12 notes." the rows are one-year excess returns on 1.5-year through 20-year treasury securities, aaa through low-grade (lg) cor porate bonds, and valueand equal-weighted nyse stock returns. d/p is the ratio of dividends on the value-weighted nyse for year t to the value of the portfolio at the end of the year. def is the difference between the end-of-year yield on all (the portfolio of the 100 corporate bonds in the sample) and the aaa yield. long-tb denotes the difference between the long-term treasury bond yield (in coleman, fisher, & lbbotson, 1993 and updates) and the one-month trea sury bill rate. 10-tb is the difference between the 10-year treasury yield and the one-month bill. 5-tb, 3-tb, and 2-tb are similarly defined. long-3 is the difference between the long-term and three-year treasury yields. the columns present mean, standard deviation, and firstthrough eighth-order autocorrelation coefficients. term spreads and predictions 29 ports fama and french's (1989) story that d/p and def track a component of expected return that changes slowly and persists beyond the nber-measured business cycles. the first-order autocorrelations on the term spreads are usually large. secondand third-order autocorrelations are usually close to zero. this suggests that term spreads track a component of expected excess return that persists for one year. we will see that this story receives strong support from the regression analysis. b. plots of business conditions and term spreads plots of the forecasting variables picture the components of expected returns they cap ture. fama and french (1989) plot the dividend yield and default spread since 1927. they move closely together. fama and french (1989) conclude that these variables track a com mon "variation in expected bond and stock returns in response to aspects of business con ditions that tend to persist beyond measured business cycles" (p. 29). to save space, we do not repeat the figure here. figure 1 plots long-tb and 3-tb against business conditions. it shows that these term spreads tend to follow the nber business cycle. they tend to be narrow near cyclical peaks and wide near troughs. the economic story is that the observable term spreads vary closely with the unobservable term premiums. term premiums, which are the rewards to 8 ¢ 0 q--2 o3 o i .6 8 q. , . . 2 i° o~-2 42 46 50 i l l l l r i l l 42 46 50 54 1 54 long-tb term spread 58 62 66 70 74 78 82 86 90 94 year 58 3-tb term spread i i i i i i i i i ~1 i i i 1 i [ i i i i i i i j 82 86 70 74 78 82 n 90 94 year figure 1. term spreads 30 financial services review 7(1) 1998 17.5 15.0 12.5 " 10.0 ~_ 7.5 .e_ >.. 5.0 2.5 0.0 ii ii ii i i i i i i i i i i i i i i i i 42 47 52 57 i l l l l l l l l l 62 67 year i j "1-1 i i i i i ; i i i i i i i i i r i i ] 72 77 82 87 92 figure 2. yield on 3-year treasury bonds extending maturity, tend to be large when business conditions are weak and small when business conditions are strong. separate plots of each interest rate in these term spreads reveal that they tend to fall in recessions and rise in expansions, with the short-term bill yield falling and rising more than the other rates. consequently, the term spreads follow the business cycle. figure 2 plots the three-year treasury yield. it comes close to defining the nber busi ness cycle. except for the 1974 recession, it peaks near the business peak and falls until near the business trough. so, it follows the business cycle. fama and french (1989) show the one-month bill yield also comes close to defining the nber business cycle, but only for the period since the treasury-federal reserve accord in 1951. c. average excess returns table 1 presents average annual excess returns on treasury bonds, corporate bonds, and common stocks for 1942-1994. henceforth, return means excess return. average returns are similar on 1.5-year through 7-year treasury bonds and then fall as maturity extends beyond seven years. specifically, the average term premiums appear to peak at about 1 percent on 1.5-year through 7-year bonds, and to fall on longer-term treasury bonds. as expected, average returns rise as we go from 20-year treasury bonds to corporate bonds, from high-grade to low-grade bonds and from bonds to stocks. note, however, that term spreads and predictions 31 the corporate bond and stock returns contain not only a term premium, but also default and equity risk premiums. the message of this study, however, is less about average returns or risk premiums than it is about the variation in expected term premiums. returns appear to be predictable and, as the next section will show, an intermediate-short spread does a better job than a long-short spread of predicting bond returns. iii. regression analysis a. choice of term spreads table 2 presents the adjusted coefficients of determination (also called adjusted r 2) for regressions of one-year bond and stock returns on def and alternative measures of a term spread for 1942-1994. the major lesson from the table is that an intermediate-short term spread does a better job of predicting bond returns than a long-short spread. for exam ple, 3-tb and def explain 8 to 10 percent more of the variance of 1.5-year through 3-year treasury returns than long-tb and def. for all treasury returns, 5-tb and def consis tently explain 7 to 8 percent more of the variance. this intermediate-short predictive advantage remains sizeable at 5 to 6 percent on aaato a-grade corporate bonds. the inter mediate-short spread, 10-tb, and long-tb predict similar percentages of the variances on lg, vw, and ew returns--the securities that contain equity risk. at least for 1942 1994, 5-tb and def usually do the best job of predicting bond returns. table 2 predictive content of alternative term spreads, 1942-1994 (adjusted coefficients of determination from regressions of one-year excess asset returns on def and alternative term spreads) dep. v a r . long-tb io-tb 5-tb 3-tb 2-tb ti.5 0.04 0.08 0.12 0.14 0.14 t2 0.02 0.05 0.09 0.12 0.12 t3 0.06 0.10 0.13 0.14 0.12 t5 0.10 0.14 0.17 0.17 0.14 t7 0.12 0.16 0.20 0.19 0.15 ti0 0.15 0.19 0.23 0.22 0.17 t15 0.17 0.22 0.25 0.24 0.19 t20 0.18 0.22 0.25 0.24 0.19 aaa 0.28 0.32 0.34 0.32 0.25 aa 0.26 0.29 0.31 0.29 0.22 a 0.33 0.36 0.38 0.35 0.28 baa 0.36 0.37 0.38 0.35 0.28 lg 0.40 0.39 0.37 0.33 0.30 vw 0.16 0.15 0.15 0.15 0.14 ew 0.21 0.20 0.20 0.20 0.20 notes: the dependent variables are one-year excess returns on 1.5-year through 20-year treasury securities, aaa through low-grade (lg) corporate bonds, and value-weighted and equal-weighted nyse stocks. the independent variables are def and one of the term spreads. def is the difference between the end-of-year yield on all (the portfolio of 100 cor porate bonds in the sample) and the aan yield. long-tb denotes the difference between the long-term treasury bond yield (in coleman, fisher, & ibbotson, 1993 and updates) and the one-month treasury bill yield. ilytb is the difference between the 10-year treasury yield and the one-month bill yield. 5-tb, 3-tb, and 2-'rb are similarly defined. 32 financial services review 7(1) 1998 table 3 regressions of one-year excess returns on def and two term spreads, 1942-1994 def t-star inttb t-stat long-int t-star r 2 t 1.5 0.50 0.89 0.98 1.90 -0.55 1.21 0.16 t2 0.77 1.08 1.19 1.84 -0.84 1.43 0.15 t3 0.92 0.93 1.55 1.99 1.52 1.37 0.16 t5 1.39 1.06 2.20 2.13 -1.69 -1.14 0.18 t7 1.72 1.11 2.75 2.27 -1.90 -1.12 0.21 ti0 2.02 1.14 3.38 2.46 -1.86 --0.99 0.23 t 15 2.46 1.22 4.02 2.61 1.89 -0.93 0.25 t20 2.84 1.30 4.26 2.54 -1.79 -0.85 0.24 aaa 2.24 1.36 3.66 3.59 -0.46 -0.24 0.33 aa 2.15 1.13 3.56 3.64 -0.45 -0.24 0.30 a 4.18 2.46 3.67 3.33 -0.19 -0.11 0.36 baa 5.90 3.70 3.21 3.23 0.50 0.29 0.37 lg 12.14 5.21 2.51 2.00 3.73 1.81 0.39 vw 13.82 3.00 0.90 0.47 4.62 1.26 0.16 ew 22.06 4.39 0.06 0.03 7.34 1.79 0.22 notes: the dependent variables are one-year excess asset returns on 1.5-year through 20-year treasury securities, aaa through low-grade (lg) corporate bonds, and value-weighted and equal-weighted nyse stocks. the independent variables are the default yield spread, def, and the term spreads between, respectively, the n-year intermediate and one-month trea sury yields and the longest available and n-year intermediate treasury yields. the n is set at 3 for ti.5 and t2, and 5 for t3 through ew. def is the difference between the end-of-year yield on all (the portfolio of the 100 corporate bonds in the sample) and the aaa yield. r 2 denotes the adjusted coefficient of determination and t-stat denotes the t-statistic on the coefficient of the column to its left. table 3 presents regressions of one-year returns on def, a long-intermediate spread, long-int, and an intermediate-short spread, int-tb. the maturity of the intermediate rate is set based on table 2 results. let us first consider the results for the treasury and investment-grade bonds. the intermediate-short spread always proves significant at the 7 percent level or better, while the long-intermediate spread never approaches significance. these results imply that the bond market prices an intermediate-short spread and not a long-intermediate spread. statistically, these results allow us to accept the model with def and int-tb as independent variables and to reject the model with def and long-int as independent variables. they also allow us to reject fama and french's (1989) model with def and a long-short spread as independent variables. we conclude that the bond market prices an intermediate-short term spread and not a long-short or long-intermediate spread. lg, vw, and ew contain equity risk. for these regressions, the long-intermediate spread approaches significance, while the intermediate-short spread is only significant for low-grade bonds. this suggests (but nothing more) that the long end of the yield curve may contain information embedded in the prices of equities. to consider whether the results are affected by multicollinearity, we regressed each independent variable on the others. the largest r 2 is 0.15, which suggests that multicol linearity is not a problem. an additional test is the "condition number" described by bels ley, kuh, and welsch (1980). for the regressions reported in table 3, the largest condition number of 3.71 suggests a lack of multicollinearity. in the spirit of fama (1990b), we next present results from regressions of bond and stock returns on def and a common term spread. an advantage of using a common term spread is that its slope coefficients provide information about variation in expected term term spreads and predictions 33 premiums as a function of maturity. the tests use the 5-tb term spread to track expected term premiums in bond and stock returns. similar results prevail for other spreads. b. regressions on 5-tb and def table 4 presents regressions of bond and stock returns on 5-tb and def for invest ment horizons of one quarter and one year through four years. we look first at the patterns of the slopes for 5-tb. as maturity lengthens (i.e., looking across a row), the variation of expected returns increases. however, slopes remain essentially flat as we go from aaa to baa bonds. surprisingly, slopes fall as we go from bonds to stocks. for a given bond (i.e., looking down a column), the slopes increase with the investment horizon through one year and, thereafter, either increase slowly or essentially flatten. this study is the first to look at returns across bond maturities. the results imply that 5-tb captures a term premium in expected bond return that is closely related to bond dura tion. we calculated the macaulay duration of treasury bonds assuming they have a 6 per cent coupon rate and are selling at par. in the one-year treasury bond regressions, the correlation is 0.99 between slopes on 5-tb and bonds' durations. each corporate bond series has an average maturity of about 20 years. interestingly, their 5-tb slopes are less than the slope on 20-year treasury bonds. recall that corporate bonds are almost always callable while treasuries are seldom callable. thus, the average call-adjusted duration of corporate bonds should be less than the duration of treasury bonds. therefore, the pattern of slopes across maturities confirms prior scholars' suspicion that a term spread, such as 5-tb, captures variation in bond returns that is closely related to duration risk. for a small change in bond yield, the approximate bond return, ap/p, is: ap/p = ai • modd, where ai is the change in yield and modd, modified duration, is duration divided by 1.03 (1 plus the bond's semi-annual yield). based on this relationship we regressed the 5-tb slopes on the bonds' modd. the slope coefficient on modd is 0.39, which has a use ful interpretation. suppose the term spread, 5-tb, is 1 percent above normal. during the next year, the five-year yield is expected to fall by 0.39 percent, which will narrow the spread and raise returns by ai times modd, that is, by the slope on 5-tb. two aspects of the 5-tb slopes are disappointing: the negative slopes in the multi-year stock regressions and the small slopes in the quarterly treasury regressions. fama and french (1989) also find the term coefficients in these stock regressions to be negative. so, this result appears to have nothing to do with the choice of term spreads. the treasury bond returns rely on estimates of par-bond yields (yields on coupon-bearing bonds selling at par) across maturities. coleman, fisher, and ibbotson (1993 and updates) estimate a dif ferent set of par yields each month. we did not calculate one-month returns because we felt the estimation errors of the month t and month t + 1 par yields could unduly influence the corresponding monthly return. the longer the investment horizon, the more confident we are that the pattern of bond returns reflects "true" bond returns and not estimation errors. thus, estimation errors may partially account for the disappointing results in the quarterly treasury regressions. we now turn to the patterns of the slope coefficients for def. looking across the rows, the slopes generally increase as we go from short-term to long-term bonds, from high-grade to low-grade bonds, and from large stocks to small stocks. looking down the columns, the slopes generally increase with the length of the investment horizon. these patterns imply that the variation in expected returns increases as we go from short-term to t a b l e 4 r eg re ss io ns o f e xc es s r et ur ns o n 5t b a nd d e f , 19 42 -1 99 4 4:: ,. t 1. 5 72 t 3 t 5 t 7 1" 10 t 15 72 0 a aa a a a b oa l g v w e w s lo pe s on 5 -t b q -0 .0 3 -0 .0 5 -0 .0 3 -0 .0 4 -0 .0 2 0. 02 0. 04 0. 02 0. 62 0. 62 0. 72 0. 72 0. 91 1. 04 1. 21 1 0. 73 0. 84 1. 34 1. 97 2. 49 3. 12 3. 76 4. 01 3. 60 3. 50 3. 65 3. 28 3. 02 1. 54 1. 07 2 0. 41 0. 43 0. 99 1. 57 2. 12 2. 85 3. 53 3. 61 4. 07 4. 01 3. 87 3. 44 2. 08 -0 .4 6 -3 .3 3 3 0. 78 0. 92 1. 69 2. 42 3. 04 3. 88 4. 65 4. 73 4. 67 4. 19 3. 93 2. 96 0. 10 -2 .7 2 -8 .7 5 4 0. 88 1. 13 2. 06 2. 93 3. 73 4. 86 5. 93 6. 18 5. 28 4. 95 4. 74 3. 86 2. 15 -0 .9 7 -6 .8 9 tst at is tic s on 5 -t b q -0 .2 2 -0 .3 6 -0 .1 6 -0 .1 4 -0 .0 6 0. 04 0. 09 0. 05 1. 56 1. 61 2. 09 2. 34 2. 15 1. 93 1. 73 1 1. 70 1. 55 1. 81 2. 00 2. 16 2. 39 2. 57 2. 54 3. 69 3. 77 3. 52 3. 48 2. 51 0. 83 0. 49 2 0. 92 0. 76 1. 24 1. 45 1. 62 1. 86 2. 02 1. 89 3. 07 3. 13 3. 02 2. 77 0. 94 -0. 25 -1 .1 7 3 1. 26 1. 23 1. 64 1. 81 1. 91 2. 12 2. 26 2. 10 2. 40 2. 33 2. 06 1. 57 0. 03 -1 .2 7 -2 .4 2 4 1. 09 1. 14 1. 51 1. 64 1. 71 1. 90 2. 02 1. 91 2. 24 2. 19 2. 03 1. 81 0. 95 -0 .3 7 -1 .9 5 s lo pe s o f d e f q 0. 11 0. 14 0. 20 0. 35 0. 43 0. 54 0. 72 0. 75 0. 72 0. 90 1. 15 1. 56 3. 53 3. 94 6. 20 1 0. 20 0. 33 0. 44 0. 85 1. 12 1. 43 1. 86 2. 27 2. 09 2. 00 4. 12 6. 06 13 .3 2 15 .2 9 24 .3 9 2 1. 20 1. 63 2. 40 3. 91 4. 79 5. 64 6. 50 7. 41 6. 56 5. 58 10 .0 4 13 .2 5 25 .5 5 29 .2 1 44 .1 9 3 1. 51 2. 13 3. 46 6. 16 8. 00 9. 83 11 .3 9 13 .0 6 9. 60 8. 79 13 .8 7 17 .3 7 33 .4 3 37 .6 8 53 .6 3 4 1. 50 2. 22 3. 91 7. 27 9. 44 11 .7 4 13 .4 3 15 .2 1 12 .9 5 11 .1 3 •5 .8 0 18 .8 8 35 .9 1 43 .1 9 55 .2 1 tst at is tic s on d e f q 0. 72 0. 75 0. 81 1. 08 1. 13 1. 25 1. 43 1. 42 1. 52 1. 90 2. 51 3. 26 4. 13 3. 11 3. 31 i 0. 38 0. 50 0. 50 0. 74 0. 83 0. 93 1. 05 1. 18 1. 48 1. 15 2. 71 4. 20 6. 21 3. 63 4. 83 2 0. 95 1. 05 1. 20 1. 57 1. 64 1. 64 1. 64 1. 73 2. 01 1. 48 3. 04 5. 48 6. 82 4. 32 5. 04 3 0. 85 1. 00 1. 19 1. 72 1. 93 2. 04 2. 03 2. 26 1. 80 1. 49 2. 80 4. 79 6. 11 5. 16 4. 22 4 0. 78 0. 98 1. 19 1. 77 1. 99 2. 15 2. 10 2. 32 1. 83 1. 46 2. 44 3. 69 6. 62 4. 99 3. 80 r 2 q -0 .0 1 -0 .0 1 -0 .0 1 -0 .0 1 -0 .0 1 -0 .0 1 -0 .0 0 -0 .0 0 0. 04 0. 04 0. 06 0. 09 0. 13 0. 05 0. 07 1 0. 12 0. 09 0. 13 0. 17 0. 20 0. 23 0. 25 0. 25 0. 34 0. 31 0. 38 0. 38 0. 37 0. 15 0. 20 2 0. 02 0. 01 0. 04 0. 08 0. 10 0. 13 0. 14 0. 13 0. 24 0. 20 0. 28 0. 35 0. 41 0. 28 0. 33 3 0. 04 0. 04 0. 08 0. 14 0. 17 0. 20 0. 21 0. 21 0. 22 0. 18 0. 26 0. 33 0. 43 0. 40 0. 42 4 0. 03 0. 05 0. 10 0. 17 0. 20 0. 23 0. 25 0. 26 0. 25 0. 20 0. 28 0. 35 0. 48 0. 42 0. 37 z > x > t" r~ < < n ot es : t he re gr es si on s f or o ne q ua rt er (q ) a nd o ne y ea r ( 1) u se n on ov er la pp in g r et ur ns . t he re gr es si on s f or tw o to fo ur -y ea r r et ur ns u se o ve rl ap pi ng an nu al o bs er va tio ns . t he n ew ey a nd w es t ( 19 87 ) m et ho d is u se d to e st im at e s ta nd ar d e rr or s i n th e tw o to fo ur -y ea r r eg re ss io ns . r 2 de no te s t he a dj us te d co ef fi ci en t o f d et er m in at io n. s ee t ab le 1 fo r d ef in iti on s o f v ar ia bl es , oe term spreads and predictions 35 long-term debt, from high-grade to low-grade bonds, and from large stocks to small stocks. the economic story is that the rewards for beating the risk reflected by def (and dividend yield) are generous when business conditions are weak and small when business conditions are strong. moreover, the risks reflected by def (and d/p) are slow to dissipate, so the expected rewards increase as the investment horizon lengthens. mathematically, the first-order autocorrelation coefficients in table 1 on def (and d/p) are large and higher-order autocorrelation coefficients decay slowly. economically, def (and d/p) reflect business conditions that change slowly over time. finally, it is interesting to view the patterns of coefficients of determination. recall that 5-tb (or another term spread) predicts returns out one year, but it has little impact on returns in years two, three and four. and, def's predictive content strengthens as we go from short-term debt through stocks and as we lengthen the investment horizon. the coef ficients of determination reflect these patterns. for shorter treasury securities, the good ness-of-fit is relatively large after one year and decreases sharply as the investment horizon lengthens. for longer-term treasury and corporate bonds, the r 2 decreases as the horizon goes from one to two years before slowly rising to reflect the impact of def. for low-grade bonds and stocks, the goodness-of-fit increases with horizon since these assets benefit from the strengthening impact of def. altogether, the results imply that 9 percent to 38 percent of the variation in one-year excess bond returns can be predicted by 5-tb and def. the predictable variation in one-year stock returns is 15 to 20 percent. c. regression diagnostics a key issue examined in this study is whether a term spread can reliably predict bond returns. this section examines whether the full period results prevail for different sub-peri ods and for different measures of returns. we form three checks. first, following fama and french (1989) we regress real returns on def and 5-tb for the post-1953 period. second, we examine the excess-return results for 1969-1994, the last half of the full period. third, we examine plots of, for example, one-year actual returns on 10-year treasury bonds ver sus the fitted values from the full period excess-return regressions. comparing actual returns to the fitted values illustrates whether the full period predictive content exists over sub-periods. table 5 presents the real return regressions for 1953-1994. they tell much the same story as the 1942-1994 excess return regressions, but there are a few interesting differ ences. the slopes on 5-tb are similar in size and significance in the two sets of regres sions. however, the slopes on def are much larger in the real return regressions. they are about ten times larger in the t1.5 and t2 regressions. they remain two to three times larger in the long-term treasury and high-grade corporate bond regressions, and these t-statistics are, on average, perhaps 20 percent larger. separate regressions (not shown) reveal that about half the absolute differences in def slopes is due to the change in time period from 1942-1994 to 1953-1994 and half to the change in dependent variable. the comparatively strong real-return results suggest that when business conditions are weak and def is wide, the expected risk-free real rate tends to be generous. thus, def appears to track a larger fraction of the variation in assets' real returns than their excess returns. table 6 presents the excess-return regressions for 1969-1994. they tell essentially the same story as the 1942-1994 regressions, but the overall fit of most regressions is some t a b l e 5 r eg re ss io ns o f r ea l r et ur ns o n 5t b a nd d e f , 19 53 -1 99 4 t 1. 5 t 2 t 3 t 5 t 7 t io t 15 t 20 a a a a a a b a a l g v w e w s lo pe s on 5 -t b q 0. 03 0. 00 0. 02 0. 01 0. 03 0. 05 0. 07 0. 04 0. 72 0. 70 0. 80 0. 81 1. 03 1. 21 1. 31 1 1. 30 1. 41 1. 91 2. 61 3. 16 3. 81 4. 45 4. 67 4. 36 4. 19 4. 29 3. 91 3. 84 2. 16 1. 33 2 1. 01 1. 01 1. 51 2. 17 2. 70 3. 43 4. 08 4. 12 4. 65 4. 48 4. 26 3. 92 3. 15 0. 31 -3 .1 4 3 1. 27 1. 37 2. 02 2. 82 3. 38 4. 14 4. 79 4. 84 5. 02 4. 39 4. 09 3. 34 1. 56 -1 .1 1 -7 .6 0 4 1. 36 1. 56 2. 33 3. 28 4. 01 5. 07 6. 02 6. 33 5. 47 5. 03 4. 85 4. 24 3. 96 1. 47 -4 .6 2 tst at is tic s on 5 -t b q 0. 25 0. 02 0. 09 0. 06 0. 09 0. 15 0. 15 0. 09 1. 69 1. 73 2. 20 2. 48 2. 33 2. 16 1. 80 1 1. 94 1. 82 2. 01 2. 24 2. 39 2. 62 2. 80 2. 76 3. 83 3. 73 3. 52 3. 53 3. 05 1. 02 0. 55 2 1. 16 1. 02 1. 26 1. 46 1. 58 1. 77 1. 89 1. 79 2. 68 2. 63 2. 53 2. 48 1. 40 0. 15 -0 .9 7 3 1. 00 0. 98 1. 25 1. 47 1. 56 1. 73 1. 82 1. 73 2. 03 1. 86 1. 65 1. 40 0. 50 -0 .4 2 -1 .9 0 4 0. 88 0. 92 1. 18 1. 35 1. 43 1. 60 1. 70 1. 66 1. 86 1. 80 1. 68 1. 57 1. 46 0. 52 -1 .2 7 s lo pe s o f d e f q 1. 37 1. 44 1. 53 1. 62 1. 72 1. 89 2. 18 2. 27 1. 68 2. 44 2. 52 2. 63 3, 58 4. 96 6, 69 1 3. 70 3. 88 3. 69 3. 16 3. 08 3. 22 3. 67 4. 29 3. 79 5. 40 7. 27 8. 66 1 i .8 2 21 .6 6 29 .4 9 2 11 ,6 2 12 .4 4 13 .4 8 14 ,2 4 15 .0 5 16 .0 2 17 .4 7 18 .4 5 18 .8 4 19 ,9 4 24 .1 0 24 .2 6 25 .8 3 40 ,7 7 50 .2 1 3 15 .7 9 17 .0 9 19 .5 3 22 .1 1 24 .6 2 27 .5 9 31 .2 3 32 .9 0 28 .0 2 29 .7 9 34 .1 4 32 .4 8 32 .8 0 42 .0 2 45 .7 9 4 18 .4 4 19 .9 1 23 .2 7 26 .7 2 29 .7 8 33 .2 8 37 .2 6 38 .2 6 37 .4 4 37 .7 6 41 .2 5 38 .3 5 37 .7 4 48 .3 2 42 .8 9 tst at is tic s on d e f q 3. 14 2. 69 2. 23 1. 77 1. 58 1. 50 1. 52 1. 49 1. 16 1. 89 2. 05 2. 06 2. 64 2. 13 2. 15 1 2. 01 1. 83 1. 42 0. 96 0. 80 0. 74 0. 76 0. 81 1. 02 1. 48 1. 88 2, 38 2. 71 2. 29 2. 69 2 2. 74 2. 60 2. 36 2. 06 1. 90 1. 80 1. 76 1. 77 2. 54 2. 69 3. 32 3. 64 3. 17 2. 81 2. 76 3 2. 70 2. 65 2. 54 2. 43 2. 41 2. 45 2. 54 2. 60 2. 61 2. 77 3. 41 3. 55 2. 81 2. 67 2. 22 4 2. 82 2. 80 2. 79 2. 73 2. 72 2. 78 2. 88 2. 83 2. 95 3. 05 3. 65 3. 59 3. 36 3. 12 2. 12 r 2 q 0. 06 0. 04 0. 02 0. 01 0. 01 0, 01 0. 01 0. 00 0. 05 0. 07 0. 09 0. 11 0. 11 0. 06 0. 05 1 0. 26 0. 22 0. 22 0. 23 0. 24 0. 27 0. 29 0, 28 0. 38 0. 38 0. 42 0, 41 0. 31 0. 15 0. 14 2 0. 29 0. 26 0. 22 0. 20 0. 19 0. 19 0. 20 0. 19 0. 33 0. 33 0. 38 0. 40 0. 26 0. 24 0. 20 3 0. 28 0. 26 0. 26 0, 26 0. 26 0. 28 0. 29 0. 28 0. 32 0. 32 0. 36 0. 36 0. 20 0. 21 0. 18 4 0. 25 0. 24 0. 26 0. 27 0. 28 0. 29 0. 31 0. 31 0. 36 0. 36 0. 39 0. 40 0. 29 0. 28 0, 10 z > z 7, t < n .. .d x. .. n ot es : t he re gr es si on s f or o ne q ua rt er (q ) a nd o ne y ea r ( 1) u se n on ov ed ap pi ng re tu rn s. t he re gr es si on s f or tw o to fo ur -y ea r r et ur ns u se o ve rl ap pi ng an nu al o bs er va tio ns . t he n ew ey a nd w es t ( 19 87 ) m et ho d is u se d to e st im at e s ta nd ar d e rr or s i n th e tw o to fo ur -y ea r r eg re ss io ns . r 2 de no te s t he a dj us te d co ef fi ci en t o f d et er m in at io n. t 1. 5 th ro ug h t 20 d en ot e re al re tu rn s o n t re as ur y se cu ri tie s. a aa th ro ug h l g d en ot e r ea l r et ur ns o n a aa th ro ug h lo w -g ra de (b el ow b aa ) c or po ra te b on ds . v w a nd e w d en ot e va lu ew ei gh te d a nd e qu al -w ei gh te d n y se st oc k re al re tu rn s. r ea l r et ur ns ar e co m pu te d fr om ib bo ts on a ss oc ia te s' c pi s er ie s. ~d o o t a b l e 6 r eg re ss io ns o f e xc es s r et ur ns o n 5t b a nd d e f , 19 69 -1 99 4 t i. 5 t 2 t 3 t 5 t 7 t io t i5 t 20 a a a a a a b a a l g v w e w s lo pe s on 5 -t b q -0. 09 -0 .1 2 -0 .1 1 -0 .1 3 -0 .1 2 -0 .1 0 -0 .1 1 -0 .1 5 0. 67 0. 63 0. 74 0. 69 0. 98 1. 48 1. 79 i 0. 59 0. 71 1. 26 2. 00 2. 58 3. 26 3. 93 4. 11 4. 04 3. 72 3. 90 3. 50 3. 65 3. 77 3. 35 2 -0 .2 9 -0 .3 8 0. 05 0. 56 0. 99 1. 58 2. 09 1. 91 2. 93 2. 68 2. 50 2. 26 1. 59 2. 78 -0 .2 5 3 -0 .1 4 -0 .1 1 0. 37 0. 86 1. 20 1. 70 2. 08 1. 87 2. 29 1. 45 i. 1 5 0. 25 1. 38 0. 77 -6 .4 6 4 0. 04 0. 25 0. 91 1. 62 2. 20 3. 03 3. 78 3. 81 2. 45 1. 90 1. 59 0. 87 0. 91 4. 11 -2 .8 9 tst at is tic s on 5 -t b q -0 .5 7 -0 .6 3 -0 .4 6 -0 .4 1 -0 .3 3 -0 .2 4 -0 .2 1 -0 .2 8 1. 33 1. 33 1. 75 1. 77 1. 86 2. 27 2. 09 i 1. 11 1. 06 1. 43 1. 79 2. 01 2. 25 2. 42 2. 34 3. 49 3. 23 3. 23 2. 95 2. 75 1. 56 1. 19 2 -0 .5 7 -0 .6 0 0. 05 0. 46 0. 67 0. 92 1. 08 0. 91 1. 62 1. 45 1. 48 1. 27 0. 70 1. 11 -0 .0 6 3 -0 .2 1 -0 .1 5 0. 33 0. 60 0. 70 0. 87 0. 97 0. 77 0. 90 0. 59 0. 47 0. 10 -0 .4 5 0. 32 1. 38 4 0. 06 0. 27 0. 72 0. 92 1. 00 1. 17 1. 26 1. 13 0. 97 0. 77 0. 65 0. 35 0. 41 1 .9 9 -0 .6 7 s lo pe s of d e f q 0. 62 0. 70 0. 79 0. 92 1. 03 1. 20 1. 50 1. 61 1. 38 2. 13 2. 16 2. 26 3. 71 4. 99 7. 37 1 1. 38 1. 73 1. 31 0. 86 0. 82 0. 88 1. 29 2. 09 0. 39 2. 30 3. 75 5. 14 9. 83 16 .6 3 27 .4 2 2 5. 83 7. 14 8. 33 9. 71 11 .0 6 12 .2 1 13 .7 2 14 .9 5 14 .0 1 14 .9 1 17 .9 4 17 .5 4 21 .3 4 26 .7 7 40 .6 5 3 7. 56 9. 41 12 .2 3 15 .9 5 19 .4 2 22 .9 8 27 .0 9 28 .7 6 21 .6 1 23 .5 4 26 .6 3 24 .8 6 27 .5 5 27 .3 9 40 .0 2 4 7. 47 9. 26 12 .6 6 16 .8 9 20 .6 1 24 .3 9 28 .5 1 29 .1 8 27 .6 8 27 .5 9 29 .7 8 26 .8 2 28 .6 9 29 .7 2 34 .2 8 tst at is tic s on d e f q 1. 34 1. 17 0. 99 0. 84 0. 78 0. 78 0. 85 0. 86 0. 81 1. 41 1. 51 1. 48 2. 31 1. 83 1. 99 1 0. 93 0. 91 0. 50 0. 24 0. 19 0. 18 0. 23 0. 33 0. 09 0. 51 0. 83 1. 14 2. 03 1. 73 2. 33 2 2. 09 2. 02 1. 65 1. 46 1. 39 1. 30 1. 26 1. 29 1. 84 1. 96 2. 30 2. 23 2. 38 1. 68 1. 79 3 2. 24 2. 25 1. 95 2. 01 2. 10 2. 17 2. 29 2. 31 1. 99 2. 16 2. 61 2. 62 2. 30 1. 57 1. 54 4 2. 04 2. 04 1. 81 1. 89 1. 97 2. 06 2. 18 2. 08 2. 07 2. 06 2. 59 2. 47 2. 64 2. 04 1. 53 r 2 q -0 .0 0 -0 .0 0 -0 .0 1 -0 .0 1 -0 .0 1 -0 .0 1 --0 .0 1 -0 .0 1 0. 04 0. 05 0. 07 0. 07 0. 11 0. 08 0, 08 1 0. 05 0. 04 0. 07 0. 11 0. 14 0. 17 0. 19 0. 19 0. 33 0. 31 0. 36 0. 31 0. 27 0. 22 0. 21 2 0. 13 0. 11 0. 05 0. 04 0. 04 0. 05 0. 06 0. 05 0. 19 0. 19 0. 24 0. 23 0. 14 0. 17 0. 11 3 0. 12 0. 12 0. 11 0. 12 0. 14 0. 16 0. 18 0. 17 0. 18 0. 18 0. 24 0. 21 0. 11 0. 13 0. 10 4 0. 09 0. 10 0. 11 0. 13 0. 15 0. 18 0. 20 0. 18 0. 22 0. 22 0. 27 0. 25 0. 15 0. 29 0. 01 t~ n ot es .. t he re gr es si on s f or o ne q ua rt er (q ) a nd o ne y ea r ( i ) u se n on ov er la pp in g r et ur ns . t he re gr es si on s f or tw o to fo ur -y ea r r et ur ns u se o ve rl ap pi ng an nu al o bs er va tio ns . t he n ew ey an d w es t ( 19 87 ) m et ho d is u se d to e st im at e st an da rd er ro rs in th e tw o to f ou rye ar re gr es si on s. r 2 de no te s t he a dj us te d co ef fi ci en t o f d et er m in at io n. s ee t ab le 1 fo r d ef in iti on s o f v ar ia bl es . 38 financial services review 7(1 ) 1998 0.3 0 , 2 0.1 f .= i~ -0.0 uj -0.1 -0.2 -0.3 a c t u a l r e t u r n s ~at~v, _ ~ s a , ; a / v " a ' ' • ^ / i ', ; / i l l l l i l l l l l [ l l l l l l l l l l l l l l l l l l l l l l l l l [ i f l [ l l l l l l l l l l l 42 47 52 57 62 67 72 77 82 87 92 year figure 3. excess returns on io-year treasury bonds what weaker. as before, the slopes of 5-tb increase with bond maturity and, looking down the columns, they peak after one year. the slopes on def generally increase as we go from short-term bonds to long-term bonds to common stocks and from one-year to four-year investment horizons. figure 3 plots actual excess returns on 10-year treasury bonds and fitted values from the one-year regression on 5-tb and def for 1942-1994. the general impression is that the regression has some ability to track actual returns, but it performed poorly from about 1968-1973 and 1986-1994. even the best of models will prove less powerful in some short periods, and this could explain the poor recent performance. however, another possible explanation of the model's poor recent performance is data mining--researchers' practice of "mining" data until they find something that would have worked in the past, but it often does not work going forward. we believe the economic story that a term spread, especially an intermediate-short term spread, tracks embedded term premiums. we, therefore, believe that a term spread will be able to predict returns going forward. others may view fama and french (1989) and the related research as the product of data mining and point to the dis appointing post-1986 results for support. the key, in our opinion, is whether someone believes the underlying economic story. with an appreciation of the qualifications discussed in the prior paragraph, we can dis cuss the investment implications and applications of the results. term spreads and predictions 39 iv. findings, interpretations, and applications in this section, we note empirical findings from this and related studies and discuss their interpretations. we also present our views of their implications and applications. first, two factors jointly predict returns on bonds. one factor can be measured by def or d/p and the second by an intermediate-short term spread. the first factor also predicts year-ahead and longer returns on stocks, but term spreads do not reliably predict stock returns. second, the impact of a default yield spread (or the market's dividend yield) on secu rity returns increases as we go from short-term to long-term bonds, from high-grade to low-grade bonds, and from bonds to stocks. moreover, its impact increases as the invest ment horizon lengthens through four years. economically, def (or d/p) appears to be associated with a business cycle that is longer than the traditional cycle defined by the nber. third, the empirical evidence implies that an intermediate-short term spread can better predict bond returns than a long-short spread. we also present our interpretation of why this predictive advantage exists. our interpretation is consistent with substantial institu tional and some empirical evidence, but it remains only one possible interpretation. for many investors including most individual investors, duration is a good measure of a bond' s risk. these investors dominate the short to intermediate end of the market. so, an intermediate-short spread should closely track embedded term premiums. in contrast, life insurance companies and defined-benefit pension plans dominate the long end of the bond market. for these institutions, long-term bonds are less risky than short-term bonds because they better match their long-term liabilities. for example, a whole-life insurance policy may promise $1,000,000 at death or age 65, whichever comes first. for a large group of insured, actuarial science allows the insurance firm to closely estimate its future cash payouts, and these payouts look a lot like a long-term bond. buying long-term bonds allows them to lock-in the projected cash flows. in addition, life insurance regulations impose the lowest reserve requirement on long-term bonds, which further strengthens the preference for long-term bonds. the cash flows of the accumulated benefit obligation (abo), a measure of the defined-benefit pension liability, look a lot like a bond. not sur prisingly, many firms try to invest pension assets (or at least the bond portion of pension assets) in a long-term bond portfolio that matches the abo liability. this creates a strong preference for long-term bonds. it follows that an intermediate-short term spread should closely track embedded term premiums. meanwhile, a long-intermediate term spread should vary primarily with factors besides term premiums. we suspect institutions' demand for long-duration bonds is one such factor, but a study of the factors is beyond the scope of this study. the intermedi ate-short spread is a cleaner, less noisy measure of embedded term premiums than a long-short spread precisely because it ignores these other factors that affect the long-inter mediate spread. a more rigorous explanation may help. in equation form, let the intermediate-short spread at time t, is t, be a function of the term premium at time t, t p r that is, is t = f ( t p t ) . let the long-intermediate term spread, l i r be a function of factors x t, yt, and zt, where one of these factors may be t p t. that is, l i t = f ( x t, yt, zt). the long-short spread, lst , thus depends on all factors: l s t = f ( t p t, xt, y r zt). prior studies used l s t to proxy for t p r we 40 financial services review 7(1) 1998 believe i s t is a cleaner, less noisy proxy for t p t precisely because it ignores the influence of factors x t, yt, and zt, and this accounts for its better predictive performance. fourth, an intermediate-short spread seems to be a jack-of-all-trades. we show that it best predicts bond returns. harvey (1989) shows that it predicts the growth in real gross domestic product up to one year ahead. fama (1990b) shows that it predicts (a) changes in inflation rates, (b) changes in real rates on short-term treasury bills, and (c) distant changes in the level of the bill rate. these results are consistent with this paper's story. the inter mediate-short spread is wide at business troughs (see figure 1). a wide term spread pre dicts that investors will be generously rewarded for extending bond duration during these tough economic times. harvey finds that it predicts fast output growth for the next year as the economy strengthens. fama finds that a wide spread also predicts that the inflation rate will increase from its low level and that the real rate will decrease from its high level. it cannot predict changes in the bill rate one or two years ahead since the rise in inflation and fall in the real rate roughly offset each other. however, the wide term spread reliably pre dicts a rise in the bill rate three and four years ahead; the rise in inflation is longer lived than the fall in the real rate. fifth, for 1942-1994, the average term risk premium (over the one-month bill yield) peaks at about 1 percent on 1.5-year through 7-year treasury bonds, and it falls thereafter. these results support the evidence from prior term structure studies that the average term premium peaks on intermediate bonds and then falls as maturity lengthens (see domian, maness, & reichenstein, 1998; ibbotson associates, 1997; mccallum, 1975; and mccul loch, 1975). the maximum average term premium of about 1 percent pales in comparison with the average equity risk premiums for the same period of 7.23 percent on value-weighted stocks and 9.67 percent on equally-weighted stocks. historically, investors have received much larger average rewards for switching from bonds (of any maturity) to equity than from extending bond maturity. this implies that the choice of debt-equity mix is more important than the choice of bond maturity for investors who periodically rebalance their portfolios back to fixed weights. that is, the debt-equity decision is a more important strategic asset allocation decision than the bond-maturity decision. sixth, the empirical evidence implies that expected term premiums vary widely. table 7 shows the expected one-year excess returns on, respectively, 1.5-year and 20-year trea sury bonds. the expectations rely on values from tables 1 and 4. they assume def is at its historic average level, and 5-tb is at its mean (1.10%) or one standard deviation below (-0.03%) or above (2.23%) its mean. when the term spread exceeds its mean, the 20-year bond enjoys a 3.02 percent expected return advantage. when the term spread is below its mean, the 1.5-year bond has a 4.40 percent return advantage. this variation in expected table 7 expected one-year excess returns across levels of the term spread assumed 5-tb --0.03% 1.10% 2.23% projected t1.5 0.19% 1.01% 1.83% projected t20 -4 .21% 0.32% 4.85% notes: the projected excess returns on t1.5 and t20 come from table 4 parameters assuming def is at its historic average and 5-tb is at its mean (1.10%) or one standard deviation below (-0.03%) or above (2.23%) its mean. t1.5 and t20 denote 1.5-year and 20-year treasury bond excess returns. term spreads and predictions 41 returns is large and economically significant. historically, 5-tb and def predict between a fifth and two-fifths of the variance in year-ahead returns on seven-year and longer bonds. finally, we recommend an application of our research and related research for individ uals who believe that def (or d/p) and an intermediate-short spread will be able to pre dict security returns. we call these individuals the believers. the application can be explained using sharpe's (1987) integrated asset allocation framework. consider someone who has a stable risk tolerance. in markowitz' mean-variance framework, he has a stable utility function. so, his optimal asset mix varies with capital market prospects--i.e., the efficient frontier. if security prices follow a random walk then market prospects never vary, returns are not predictable, and the individual should follow a constant-weight portfolio strategy with periodic rebalancing. this individual's optimal mix is called his strategic asset mix. the empirical results in this paper and in related research imply that an intermedi ate-short term spread and def (or d/p) can partially predict security returns. it follows that believers should consider tactical asset allocation, where the cash-bonds-stock asset mix changes with market conditions. the equity portion should vary directly with the level of def or d/p. the allocation of the fixed-income portfolio between short-term and long-term debt should vary primarily with the level of 5-tb or another intermediate-short spread. how far should believers allow their asset mix to vary? if returns were 100 percent predictable, the individual would move everything ahead of time into next period's best performing asset. if returns were not predictable, as the random walk model assumes, he or she would follow a constant-weight strategy. a few years ago, samuelson (1990) looked at the evidence on returns' predictability, including many studies referenced herein, and con cluded that it could support deviations in an asset-class weight (e.g., stocks, bonds, cash) of plus or minus 10 percent. although his is but one opinion, and a case could be made for larger maximum deviations, it provides one estimate of how far an individual should allow his tactical asset allocation weights to vary from his strategic weights. if the 10 percent maximum deviation is applied, the recommended tactical strategy amounts to a modest deviation from a constant-weights strategy, but one that we believe is justified. v. conclusions prior studies conclude that a long-short term spread can predict returns on long-term bonds and stocks. we extend these studies (1) by examining the predictive content of other term spreads, especially intermediate-short spreads and (2) by examining regressions of excess returns on 1.5-year through 20-year treasury bonds. the major contributions of this study are two of its novel results. first, we show that the bond market prices an intermediate-short spread and not a long-short or a long-interme diate spread. we believe that the intermediate-sbort spread better tracks embedded term premiums than a long-short spread. second, this is the first study to examine regressions of bond returns across maturities. in regressions of 1.5-year to 20-year treasury returns on the forecasting variable, the cor relation is 0.99 between the slopes for the intermediate-short term spread and estimates of 42 financial services review 7(1) 1998 bond durations. these results confirm prior scholars' suspicion that a term spread tracks the reward to bearing duration risk. finally, we recommend that individual investors who want to practice tactical asset allocation strategies should vary their debt-equity mix with the level of a default risk pre mium or the stock market's dividend yield, while the maturity mix of the debt portfolio should vary with an intermediate-short term spread. acknowledgment: we thank karen eilers lahey (the editor), jeffrey m. mercer and two anonymous reviewers for valuable comments. appendix coleman, fisher, and ibbotson (1993 and updates) estimate end-of-month spot rates and par bond yields (yields on coupon-bearing bonds selling at par) on fully taxable treasury bonds beginning end-of-december 1940. few fully taxable treasury bonds existed before year-end 1941. our data series begin december 1941 when estimated par yields on 1, 1.5, 2, 3, 4, 5, 7, 10, 15, and 20-year bonds were generally available. we calculated quarterly returns on coupon-bearing bonds with maturities of 1.5, 2, 3, 5, 7, 10, 15, and 20 years. the calculations assume that the bond is bought at par at the beginning of the quarter and sold at the end of the quarter at a price corresponding to the then prevailing par yield. the ending par yield is the weighted average of surrounding yields. for example, suppose the end-of-september par yield on five-year bonds is 8.44% and the end-of-december four-year and five-year par yields are 7.96% and 7.98%. the end-of-december 4.75-year par yield is taken as 7.975%, the weighted average of the four-year and five-year par yields. the end-of-december price, /'4.75, which includes accrued interest, is 10 4.22(1.039875)0"5 100(1.039875)°'5 p4.75 = ]~ + t= 1 (1.039875) t (1.039875) 10 where 4.22 is the semiannual coupon and 1.039875 is 1.00 plus the december semiannual yield. the (1.039875) 0.5 in the numerator serves to move all payments forward three months (half a semiannual period). the quarterly return is (p4.75 100)/100, where 100 is the par price of the 5-year bond. references belsley, d. a., kuh, e., & welsch, r. e. (1980). regression diagnostics: identifying influential data and sources of collinearity. new york: wiley. brooks, r., levy, h., & livingston, m. (1989). the coupon effect on term premiums. journal of financial research, 12, 15-21. term spreads and predictions 43 campbell, j. y. (1987). stock returns and the term structure. journal of financial economics, 18, 373-399. chen, n. (1991). financial investment opportunities and the macroeconomy. journal of finance, 46, 529-554. coleman, t. s., fisher, l., & ibbotson, r. g. (1993). historical u.s. treasury yield curves, 1993 edi tion. chicago: ibbotson associates. domian, d. l., maness, t. s., & reichenstein, w. (1998). rewards to extending maturity: implica tions for investors. journal of portfolio management, 24, 77-92. elton, e. j., gruber, m. j., & mei, j. (1996). return generating process and the determinants of term premiums. journal of banking and finance, 20, 1251-1269. fama, e. f. (1976). forward rates as predictors of future spot rates. journal of financial economics, 3, 361-377. fama, e. f. (1984). the information in the term structure. journal of financial economics, 13, 509 528. fama, e. f. (1986). term premiums and default premiums in money markets. journal of financial economics, 17, 175-196. fama, e. f. (1990a). stock returns, expected returns, and real activity. journal of finance, 45, 1089 1108. fama, e. f. (1990b). term-structure forecasts of interest rates, inflation, and real returns. journal of monetary economics, 25, 59-76. fama, e. f., & bliss, r. r. 0987). the information in long-maturity forward rates. american eco nomic review, 77, 680-692. fama, e. f., & french, k. r. (1989). business conditions and expected returns on stocks and bonds. journal of financial economics, 25, 23-49. fraser, p. (1995). uk stock and government bond markets: predictability and the term structure. applied financial economics, 5, 61-67. hardouvelis, g. a. (1994). the term structure spread and future changes in long and short rates in the g7 countries: is there a puzzle? journal of monetary economics, 33, 255-283. harvey, c. r. (1989). forecasts of economic growth from the bond and stock markets. financial analysts journal, 44, 38-45. ibbotson associates. (1997). stocks, bonds, bills and inflation, 1997 yearbook. chicago: ibbotson associates. jensen, g. r., mercer, j. m., & johnson, r. r. (1996). business conditions, monetary policy, and expected security returns. journal of financial economics, 40, 213-237. keim, d. b., & stambaugh, r. f. (1986). predicting returns in the stock and bond markets. journal of financial economics, 17, 357-390. mccallum, j. s. (1975). the expected holding period return, uncertainty and the term structure of interest rates. journal of finance, 30, 307-323. mcculloch, j. h. (1975). an estimate of the liquidity premium. journal of political economy, 83, 95-119. newey, w. k., & west, k. d. (1987). a simple, positive semi-definite, heteroskedasticity and auto correlation consistent covariance matrix. econometrica, 55, 703-708. samuelson, p. (1990). asset allocation could be dangerous to your health. journal of portfolio man agement, 16, 5-8. schwert, g. w. (1990). stock returns and real activity: a century of evidence. journal of finance, 45, 1237-1257 sharpe, w. f. (1987). integrated asset allocation. financial analysts journal, 43, 25-32. shiller, r. j., campbell, j. y., & schoenholtz, k. l. (1983). forward rates and future policy: inter preting the term structure of interest rates. brookings papers on economic activity, 173-217. stambaugh, r. f. (1988). the information in forward rates: implications for models of the term struc ture. journal of financial economics, 21, 41-70. 44 financial services review 7(1) 1998 startz, r. (1982). do forecast errors or term premia really make a difference between long and short rates? journal of financial economics, 10, 323-329. financial services review, 33(1) 165 a domain specific measure of investment risk preference eun jin kwak1 and john e. grable2 abstract this study introduces and validates a domain-specific investment risk-preference measure that integrates elements of revealed-preference tests, using choice scenario dyads, with statedpreference approaches that leverage individual experiences and perceptions. data from two surveys were analyzed using ols regression and ordered logit models to evaluate the measure’s efficacy. results demonstrate that the proposed measure is positively associated with a modified version of the survey of consumer finances (scf) self-assessed risk-tolerance item and negatively associated with cash-holding behavior. compared to existing risk-tolerance assessments, this measure offers a practical advantage by allowing financial advisors to align investment products more accurately with a test-taker’s risk-taking comfort level. this direct applicability highlights the measure's unique value in enhancing portfolio personalization and advancing the precision of investment risk assessment tools. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation kwak, e-j., & grable, j. e. (2025). a domain specific measure of investment risk tolerance. financial services review, 33(1), 179-204. introduction the degree to which a household financial decision-maker allocates investable resources across risk-free, fixed-income, and growth assets is associated with the investment preference of the financial decision-maker. this is one reason governmental agencies, certification boards, and other regulatory bodies require financial service providers to assess, before making investment recommendations, their clients' preference for 1 corresponding author (kwake@uwgb.edu). university of wisconsin-green bay, green bay, wi, usa 2 university of georgia, athens, ga, usa 3 the securities and exchange commission (sec) best interest regulations require investment advisers to evaluate the risk-taking preference of clients prior to making investment recommendations. specifics about the regulations can be found at https://www.sec.gov/info/smallbus/secg/regulation-best-interest. readers may also find the following sec web link to be helpful: https://www.sec.gov/exams/adviser_compliance_questions. similarly, the financial industry regulatory authority, inc. (finra) has rules and standards related to client in-take assessments. these rules can be found at: https://www.finra.org/rules-guidance/key-topics/suitability/faq. holding risky assets.3 essentially, regulators want to ensure that the risks taken by an investor align with the investor’s preference for and willingness to take a risk. the importance of accurately gauging a financial decision-maker's preference for risk implies the need for reliable and valid measures of risk preference. while various academic traditions exist to estimate a person's risk-taking preference, two approaches dominate: revealed-preference tests and stated-preference measures. revealed-preference tests are designed https://creativecommons.org/licenses/by-nc/4.0/ mailto:kwake@uwgb.edu https://www.sec.gov/info/smallbus/secg/regulation-best-interest https://www.sec.gov/exams/adviser_compliance_questions https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(1) 166 to describe preferences by observing choices made when someone is faced with an incentivized dilemma, whereas stated-preference measures (sometimes referred to as propensity or elicited risk-preference tests) rely on a testtaker’s self-report of their subjective evaluation of the riskiness of choices. regardless of the assessment approach, the intended outcome is to ascertain a financial decision-maker’s stable preference for trading risk for return (davies, 2016). today, not only is there disagreement about the optimal way to measure financial risk preference, but there is also no consensus on the preferred definition describing risk-taking preference (harnum et al., 2010; mata et al., 2018). risk preference is sometimes referred to as risk appetite, risk aversion, or risk tolerance (rabbani & nobre, 2022). for clarification purposes, in this study, risk preference refers to a financial decision-maker's general feeling that one situation or choice is better than another one (nobre & grable, 2015), regardless of whether this feeling is accurate. current methodologies used to evaluate investment risk preference tend to either place a large cognitive load upon test takers, thus reducing statistical reliability, or they skew towards being overly simplistic, which can reduce validity. this paper aims to describe the development and testing of a domain specific investment risk-preference measure that blends aspects from revealed-preference tests (i.e., choice scenario dyads) with aspects of statedpreference measures that allow a test-taker to take advantage of past investment experience and evolving perceptions when making choices. as discussed in the paper, this new measure offers a quick and accurate way to measure the specific domain of investment risk preference. 4 revealed-preference theory can be presented in one of three forms (feng & seasholes, 2005). the weak axiom suggests that consumers are consistent in their choices, given income and price constraints, and that purchases always reflect preferences. the strong axiom is premised on the idea that transitivity cannot literature review and hypothesis development revealed-preference assessments samuelson (1938) is generally credited with introducing the theory of revealed preferences. the theory states that the consumption preference of a consumer can best be described through a consumer’s behavior. samuelson assumed that consumers are rational, and before making a decision, consumers weigh the costs and benefits of differing alternatives. only after making an informed preference analysis does a consumer decide on the option that matches their preference. revealed-preference theory allows for adding constraints, such as the introduction of budgets and supply and price limitations. the theory has been widely studied, and today, revealed-preference theory underlies many models of consumer choice.4 the use of revealed-preference assessments is one of two primary ways financial advisors assess the willingness (or unwillingness) of their clients to take financial risks. revealed-preference tests, either in a clinical or survey setting, require testtakers to choose between controlled monetary lotteries (arslan et al., 2020). the notion underlying such tests is that risk and risky choices can best be proxied by analyzing the variance in potential monetary outcomes associated with dyadic choices (hertwig et al., 2018). an example of a dyadic choice is when a test-taker is asked to select their preference between an investment that provides a guaranteed gain of $500 or an investment with a 50% chance of generating a guaranteed $1,000 or a 50% chance of making nothing. although the expected value is the same with both choices, someone who is risk intolerant will generally choose the first investment. in contrast, someone who prefers more investment risk will opt for the second investment. as noted by hertwig et al. (2018), "the revealed-preference tradition holds that people's utilities and true beliefs are revealed be violated. for example, if x is preferred to y, then y cannot be revealed as preferred to x. the generalized axiom deals with situations where two or more alternatives are equally preferred (richter, 1966). kwak & grable 167 through the (incentivized) choices they make" (p. 2). through the combination of a series of related choice scenarios, it is possible to estimate a generalized preference for risk, which can then be converted to a measure of constant relative risk aversion. while revealed-preference assessment techniques are mathematically eloquent, the assessment process is not without its critics. guiso and sodini (2013) and mudzingiri and koumba (2021) noted that revealed-preference tests place a cognitive burden on test-takers, often taxing a person’s numeracy skills.5 the result can lead to cognitive biases that result in choices that vary from a person’s true preference. arshan et al. (2020) pointed out that this measurement approach relies on often violated assumptions, including the notion that preferences are temporally stable (i.e., akin to a personality trait). in a landmark study, frey et al. (2017) concluded that nearly all revealed-preference tests, along with other behavioral assessment techniques, capture transient states rather than stable preferences. as evidence of this assertion, in tests of reliability, test-retest estimates tend to be low (e.g., .20; see mata et al., 2018). stated-preference assessments stated-preference assessments are the primary alternative to revealed-preference measures. stated-preference scores are generally estimated using a propensity scale/questionnaire or a singleitem question (cardak & martin, 2019; rabbani & nobre, 2022). while psychologists generally prefer self-reports, economists tend to be skeptical of what people say because they believe there is a loose (at best) association between statements of intent and actual behavior (i.e., a self-report is ‘cheap talk’). the psychological case for using stated-preference measures lies in how risk is defined using a psychometric lens. rather than focus on the variance of returns, psychologists define risk as an activity that offers rewards with a corresponding possibility of loss. when viewed this way, rather than being always quantifiable, losses are hypothesized to be less 5 additionally, the notion of knowing probability outcomes prior to making a risky financial decision is removed from the realities faced by most investors. predictable or ascertainable before engagement in a risk-taking activity. stated-preference tests rely on a person’s introspective ability to gauge future behavior rather than current observable behavior (arslan et al., 2020). as opposed to being a weakness (i.e., introspective evaluation), arslan et al. (2020) argued that scores derived from stated-preference tests offer a more reliable and valid insight into a decision-maker’s true preference orientation. when answering a stated-preference item or items in a questionnaire, a test-taker is forced to establish a reference frame in which the person considers their possible action(s), situational constraints, and anticipated experiences (e.g., disappointment, regret, fear; hertwig et al., 2018). as noted by arslan et al., statedpreference questions allow test-takers to refer to memories, experiences, and perceptions in a way that increases the validity of a preference assessment. summary as this review of the literature suggests, the revealed-preference and stated-preference traditions associated with risk-preference assessment provide unique advantages and disadvantages. in terms of revealed-preference assessments, tests offer a direct way to measure choices through the presentation of bivariate alternatives offering varying degrees of risk. appraisal of these choices does provide insight into a decision-maker’s preference orientation. the primary drawback to existing revealedpreference tests is that they generally rely on presenting predetermined probability outcomes. this aspect of revealed-preference testing is divorced from the realities faced by decisionmakers when formulating investment decisions. also, the inclusion of probabilities into scenarios increases test-taker cognitive load, which is known to reduce the validity of outcome assessments. stated-preference assessment techniques provide more valid and reliable insight into a decision-maker’s preferences; however, this assessment approach lacks a direct financial services review, 33(1) 168 way to assess immediate choices. additionally, scores from stated-preference tests can be difficult to evaluate. scores are also economically unsustainable for use in models based on modern portfolio theory (guiso & sodini, 2013; hanna & lindamood, 2004). the advantages and disadvantages associated with these two assessment approaches have led a handful of researchers to consider alternatives to or extensions of revealed-preference and statedpreference assessment methodologies to gain a more robust understanding of investment risk preferences.6 for example, hanna et al. (2001) developed a measure of subjective risk tolerance based on economic theory. their specific aim was to present a more useable and valid way to derive estimates of constant relative risk aversion. a few years later, hanna and lindamood (2004) designed a revealed-preference test based on pension income gambles. they added a visual element to the assessment process to reduce the cognitive load on test-takers. they reported an improvement in risk-aversion score outcomes compared to a traditionally designed revealedpreference test (i.e., risk-aversion scores were greater than what has been reported using other measures). hanna and lindamood also noted that scores on their measure were positively associated with the survey of consumer finances single-item stated-preference risk-aversion question. grable et al. (2020) took another approach to estimating investment risk preference by combining elements from revealed-preference and propensity measurement techniques to document risk aversion. scores from their measure were found to correlate with outcomes from other tests of risk aversion, as well as with indicators of risk-taking. the current study extends the work of hanna and lindamood and grable et al. by providing an investment domain specific assessment alternative that blends the aspects of a revealed-preference test (i.e., choice scenario dyads) with elements from stated-preference measures that allow a test-taker to take advantage of experiences and perceptions when making choices. 6 in addition to the studies described in this paper, readers may be interested in reviewing barsky et al. (1997), bowen et al. (2015), brink and rankin research hypotheses when the development of a risk-assessment test is contemplated, it is important to document the measure’s validity. when using cross-sectional data, this is usually accomplished by correlating scores with factors associated with risk preference and risk aversion (i.e., a form of concurrent validity). for example, the proportion of one’s portfolio held in less risky assets (e.g., cash) should be lower for those with a low investment risk-preference score (kim et al., 2019). additionally, one should expect differences in investment risk preference by gender and income (mudzingiri & koumba, 2021), as well as other factors (ertac, 2020). hartnett et al. (2019), koekemoer (2018), and dickason and ferreira (2018) reported that females generally exhibit greater financial risk aversion (i.e., they prefer less risk). income and a preference for low-risk investments are thought to be negatively associated (grable & joo, 2004; pinjisakikool, 2017; wong, 2011). similarly, educational attainment and risk aversion are believed to be negatively associated (hallahan et al., 2004; larkin et al., 2013), whereas a preference for risk-taking has been reported to be positively associated with satisfaction, which is occasionally used as an indicator of financial knowledge and confidence (atlas et al., 2019; dare et al., 2020; grable, 2000; robb, 2012). informed by a review of the literature, this study was conceived to test the following hypotheses: h1: investment risk preference is positively associated with risk-taking investment behavior, controlling for gender, education, income, and financial satisfaction. h2: investment risk preference is positively associated with a person's self-evaluation of their risk tolerance, controlling for gender, education, income, and financial satisfaction. data, variables, and methodology data from two surveys, one conducted in 2022 and the other conducted in 2024, were used to evaluate the proposed risk-preference scale. the (2013), charness et al. (2013), and hansson and lagerkvist (2011). kwak & grable 169 following discussion describes each survey, the variables used in this study, and the methods utilized to test the robustness of the proposed measure. survey one data initial data for this study were obtained from a survey distributed by precision sample, llc (https://www.precisionsample.com/) to a panel of adults aged 18 years or older living in the united states as of december 2022. the survey was distributed to 600 individuals who received a modest incentive upon survey completion. useable data from 596 respondents was available and used in the analyses. table 1 provides a descriptive overview of the sample. survey two data follow-up data were gathered from another survey distributed by precision sample, llc to a different panel of adults aged 18 years or older who were living in the united states in november 2024. the survey was sent to 500 individuals, who, like those in the first sample, received an incentive after completing the survey. useable data from 458 respondents was obtained. descriptive data for the sample is provided in table 1. investment risk-preference test the following six questions were presented in a skip pattern to estimate investment riskpreference scores. the test was conceptualized similarly to a revealed-preference assessment in that respondents were asked to choose between two distinct choice alternatives; however, rather than use probability estimates of return variation and/or losses/gains, the choice scenarios were written to align with the way a stated-preference item might be presented. in this way, the test blended aspects from revealed-preference and stated-preference traditions. the series of questions was prefaced with the following statement: "financial decision-makers face difficult choices when selecting investments. given the following two options, which would you prefer to own?” scores next to each choice scenario indicate how answers were coded. whenever a respondent received a score of 0, they were skipped to the next choice scenario; otherwise, they exited the skip pattern with the assigned score. the only exception was for those faced with the final choice alternatives, in which case a respondent received a score of six or seven, depending on their choice. as such, the questioning process resulted in scores ranging from 1 to 7. the mean and standard deviation of scores was 1.53 and 1.10, respectively, in the first survey and 1.81 and 1.00, respectively, in the second survey. scenario 1: of the two investment choices shown below, which do you prefer? (a) a 100% guaranteed short-term bank product with a low return and minimal risk, offering stable value and little to now volatility. (1) (b) a 100% guaranteed short-term government bond with a low return and minimal risk, where the value may fluctuate slightly (i.e., go up or down). (0) scenario 2: of the two investment choices shown below, which do you prefer? (a) a 100% guaranteed short-term government bond with a low return and almost no volatility, where the value may fluctuate slightly (i.e., go up or down). (2) (b) a 100% guaranteed long-term bank product with a higher return and low volatility, offering stable and predictable value growth. (0) scenario 3: of the two investment choices shown below, which do you prefer? (a) a 100% guaranteed long-term bank product with a higher return and low volatility, offering stable and predictable value growth. (3) (b) a long-term government bond with a higher return, subject to moderate volatility, causing its value to fluctuate. (0) scenario 4: of the two investment choices shown below, financial services review, 33(1) 170 which do you prefer? (a) a long-term government bond with a higher return, subject to moderate volatility, causing its value to fluctuate. (4) (b) a mix of stocks and bonds with high volatility, leading to significant fluctuations in value. (0) scenario 5: of the two investment choices shown below, which do you prefer? (a) a mix of stocks and bonds with high volatility, leading to significant fluctuations in value. (5) (b) a mix of stocks and bonds with extremely high volatility, resulting in frequent and significant fluctuations in value. (0) scenario 6: of the two investment choices shown below, which do you prefer? (a) a mix of stocks and bonds with extremely high volatility, resulting in frequent and significant fluctuations in value. (6) (b) commodities or cryptocurrencies with extremely high volatility, causing radical and unpredictable fluctuations in value. (7) self-assessed risk tolerance in alignment with a recommendation by grable and lytton (2001), and similar to the approach used by hanna and lindamood (2004), an adapted version of the survey of consumer finances (scf) investment choice question, which was similar to a question in the chinese household financial survey (see hanna et al., 2018), was used to evaluate a respondent’s selfassessed risk tolerance. scores were used in a validity test of the proposed measure. the original question included the following four response categories: (a) substantial, (b) above average, (c) average, and (d) no risk. the question was revised for this study to include a belowaverage category. the question was asked and coded as follows: which of the following statements comes closest to the amount of financial risk that you are willing to take when you save or make investments? (a) take substantial financial risk expecting to earn substantial returns (coded 5— substantial risk) (b) take above-average financial risks expecting to earn above-average returns (coded 4—above average risk) (c) take average financial risks expecting to earn average returns (coded 3—average risk) (d) take below-average financial risks expecting to earn below-average returns (coded 2—below average risk) (e) not willing to take any financial risks (coded 1—no risk) control variables the following control variables were measured and used as validation factors in the analyses. gender was coded 1 = female and 0 = male. although an “other” category was provided (e.g., non-binary), no respondents (in either survey) selected this category. education was assessed with six ordinal categories ranging from some high school or less to graduate or professional degree. income was measured on an ordinal scale ranging from less than $10,000 to $150,000 or more. financial satisfaction was measured by asking respondents to indicate how satisfied they were with their present financial situation. the following response options were provided: (a) extremely negative, (b) somewhat negative, (c) neither positive nor negative, (d) somewhat positive, and (e) extremely positive. descriptive data for the control variables are shown in table 1. methodology in addition to the descriptive statistics presented in table 1, a correlation analysis was conducted to estimate the associations across the variables of interest in both surveys. this was followed by the estimation of regression models with the proportion of cash held by a respondent in their portfolio as the outcome variable. in the first model, investment risk-preference scores and the control variables were used as the independent factors. the model was operationalized as follows: kwak & grable 171 𝑌𝑖𝑒 = 𝛽0 + 𝛽1𝑋𝑖1 + 𝛽2𝑋𝑖2 +⋯+ 𝛽𝑝𝑋𝑖𝑝 + ɛ (1) where, 𝑌𝑖𝑒 = represents the cash holdings for individual i, 𝛽0 is the regression constant, 𝑋𝑖𝑝 denotes individual i values on the pth of predictor variables in the model, and ɛ is the error term. two regressions were estimated. the first regression included the investment risk preference variable, while the second included the revised scf risk tolerance variable that was used as proxy for self-assessed (sa) risk tolerance. this analysis was conducted to validate the findings from the first test. this was followed by the estimation of an ordered logit regression for each survey where selfassessed risk tolerance was the dependent variable, with investment risk preference and the control variables included as predictors. the model assumes the latent variable 𝑌∗exists corresponding to the self-assessed risk tolerance value 𝑌𝑟. the model further assumed 𝑎1 < 𝑎2 < 𝑎3 < 𝑎4 < 𝑎5, which represent the estimated critical values, and the relationship between 𝑌∗and 𝑌𝑟 depends on whether it is greater than or less than the given critical values, which are defined as follows: 𝑌𝑟 = { 1 𝑖𝑓 𝑌∗ ≤ 𝛼1 2 𝑖𝑓 𝛼1 < 𝑌 ∗ ≤ 𝛼2 3 𝑖𝑓 𝛼2 < 𝑌 ∗ ≤ 𝛼3 4 𝑖𝑓 𝛼3 < 𝑌 ∗ ≤ 𝛼4 5 𝑖𝑓 𝛼4 < 𝑌∗ } (2) if 𝑌𝑟 is the ordinal outcome of self-assessed risktolerance with j categories, then the cumulative probability of 𝑌𝑟 less than or equal to a specific category j = 1, …, j-1 is 𝑃(𝑌𝑟 ≤ 𝑗). in the model, 𝑃(𝑌𝑟 ≤ 𝐽) = 1. the log odds of being less than or equal to a particular category can be defined as: 𝑙𝑜𝑔 𝑃(𝑌𝑟≤𝒋) 𝑃(𝑌𝑟>𝑗) = 𝑙𝑜𝑔𝑖𝑡(𝑃(𝑌𝑟 ≤ 𝑗)) = 𝛽0 + 𝛽𝑗1𝑋1 + 𝛽𝑗2𝑋2 +⋯+ 𝛽𝑗𝑝𝑋𝑝 + 𝜀 (3) for p predictors. results figure 1 shows the distribution of self-assessed risk tolerance scores derived from the modified scf item. the distribution of scores generally matched what has generally been reported in the literature, with the majority of respondents falling into the average to above-average categories. figure 1. self-assessed risk-tolerance score distributions 19% 16% 32% 21% 12% 25% 13% 36% 20% 6% 0% 5% 10% 15% 20% 25% 30% 35% 40% no risk below average average above average substantial survey one survey two financial services review, 33(1) 172 table 1. descriptive sample statistics (sample one n = 596; sample two n = 458) survey one survey two variable mean sd % mean sd % cash holdings 50.48% 36.72% 51.25% 37.36% gender males females 47% 53% 50% 50% education some high school or less high school graduate some college/trade/vocation training associate’s degree bachelor’s degree graduate of professional degree 2% 17% 22% 16% 31% 12% 2% 19% 22% 11% 34% 12% income less than $10,000 $10,000 to $19,999 $20,000 to $29,999 $30,000 to $39,999 $40,000 to $49,999 $50,000 to $59,999 $60,000 to $69,999 $70,000 to $79,999 $80,000 to $89,999 $90,000 to $99,999 $100,000 to $149,999 $150,000 or more 4% 6% 7% 11% 10% 9% 7% 6% 5% 6% 14% 15% 6% 8% 13% 10% 7% 11% 7% 8% 5% 3% 14% 8% financial satisfaction 1 extremely negative 2 somewhat negative 3 neither +/ 4 somewhat positive 5 extremely positive 12% 16% 21% 32% 195% 12% 26% 22% 30% 10% self-assessed risk tolerance no risk below average average above average substantial 19% 16% 32% 21% 12% 25% 13% 36% 20% 6% investment risk preference 1.53 1.10 1.81 1.00 in the first sample, more females than males completed the questions. the gender composition of the sample was evenly split in the second survey. most respondents had completed at least some college, trade, or vocational training in both surveys. income in both samples was broadly distributed across the categories. financial satisfaction was similar across the samples. survey one results table 2 shows the correlation coefficient estimates for the variable associations in the first survey. a positive association between investment risk-preference scores and selfassessed risk tolerance was observed. investment risk preference was not statistically significantly related to the gender, income, or the financial satisfaction of respondents. investment risk preference was, however, negatively correlated with cash ownership. financial services review, 33(1) 173 table 2. correlation estimates across the variables of interest in survey one variable 1 2 3 4 5 6 7 1. sa risk tolerance 1.00 2. cash -.41** 1.00 3. gender -.25** .13* 1.00 4. education .31** -.31** -.09* 1.00 5. income .34** -.37** -.15** .47** 1.00 6. fin. satisfaction .34** -.40** -.14** .27** .36** 1.00 7. inv. risk pref. .15** -.13** -.01 .07 .04 .03 1.00 note. *p < .01, **p < .001. a regression was estimated to determine the degree to which investment risk preference was positively associated with cash ownership behavior, controlling for gender, education, income, and financial satisfaction. the model was statistically significant, f5,595 = 37.408, p < .001. the model explained approximately 24% of the variance in cash ownership reported by respondents. as shown in table 3, investment risk preference was negatively associated with cash holding behavior. this finding provides support for the first research hypothesis. education, income, and financial satisfaction were also negatively associated with cash ownership. the last four columns of table 3 show the model that was developed as a validity test. in this model, self-assessed risk tolerance was included as an independent variable. this model was also statistically significant (f5,595 = 45.72, p < .001). the amount of explained variance in cash ownership was similar to the original model (27%). table 3. ols regression: dependent variable cash holdings as percentage of portfolio in survey one model 1 validity test model variable b se β t b se β t constant 105.46** 5.20 20.28 105.18** 4.88 21.55 gender 3.20 2.68 .04 1.20 -.17 2.67 -.00 -.06 education -3.38* 1.09 -.13 -3.09 -2.51* 1.08 -.09 -2.33 income -2.02** .44 -.19 -4.56 -1.61** .44 -.15 -3.68 fin. satisfaction -3.92** .54 -.28 -7.31 -3.21** .54 -.23 -5.99 inv. risk pref. -3.35** 1.20 -.10 -2.79 sa risk tolerance -7.21** 1.14 -.25 -6.30 note. *p < .01, **p < .001. the second research hypothesis was addressed using an ordered logit regression. coefficients were estimated to determine whether investment risk preference was positively associated with a respondent’s self-assessment of their risk tolerance, controlling for gender, education, income, and financial satisfaction. the model (table 4) was statistically significant, χ2 = 163.616, p < .001. based on pseudo r2 estimates, it was determined that the model explained approximately 25% of the variance in selfevaluation scores (cox and snell r2 and nagelkerke r2, respectively). investment risk preference was found to be positively associated with self-assessed risk tolerance. this finding provides support for the second hypothesis. additionally, each control variable was significant in the model, with the direction of coefficients matching what has generally been reported in the literature. financial services review, 33(1) 174 table 4. ordered logit regression: dependent variable self-assessed risk tolerance in survey one variable estimate se wald 95% conf. int. lower bound 95% conf. int. upper bound gender -.82** .15 28.11 -1.12 -.51 education .24** .06 14.93 .12 .36 income .09** .03 13.43 .04 .14 fin. satisfaction .17** .03 29.08 .11 .23 inv. risk pref. .24** .07 11.80 .10 .37 note. *p < .01, **p < .001 survey two results the tests from the first survey were replicated using data from the second survey. table 5 shows the correlation coefficient estimates for the variable associations in the second survey. a positive association between investment riskpreference scores and self-assessed risk tolerance was observed. investment risk preference was not statistically significantly related to gender; however, scores were positively associated with respondent education, income, and financial satisfaction. investment risk preference was also negatively associated with cash ownership. table 5. correlation estimates across the variables of interest in survey two variable 1 2 3 4 5 6 7 1. sa risk tolerance 1.00 2. cash holdings -.36** 1.00 3. gender -.14** .16** 1.00 4. education .23** -.18** .04 1.00 5. income .32** -.35** -.03 .41** 1.00 6. fin. satisfaction .27** -.36** -.13* .18** .39** 1.00 7. inv. risk pref. .18** -.19** -.08 .10* .14** .09* 1.00 note. *p < .01, **p < .001. a regression was used to assess whether investment risk preference was positively associated with cash ownership controlling for gender, education, income, and financial satisfaction. the model shown in table 6 was statistically significant, f5,453 = 25.177, p < .001. the model explained approximately 22% of the variance in cash ownership reported by respondents. as shown in the table, investment risk preference was negatively associated with holding cash. this finding provides further support for the first research hypothesis. gender, income, and financial satisfaction were also associated with cash ownership. the last four columns of table 6 show the model where self-assessed risk tolerance, rather than investment risk preference, was included as an independent variable. the model was statistically significant, f5,453 = 29.056, p < .001, explaining about 25% of the variance in cash ownership. self-assessed risk tolerance was negatively associated with cash holdings. gender, income, and financial satisfaction were also associated with cash ownership. kwak & grable 175 table 6. ols regression: dependent variable cash holdings as percentage of portfolio in survey two model 1 validity test model variable b se β t b se β t constant 87.88** 7.81 11.26 93.60** 7.79 12.01 gender 8.49* 3.17 .11 2.68 7.31* 3.13 .10 2.33 education -.99 1.23 -.04 -.80 -.39 1.22 -.02 -.32 income -2.39** .52 -.22 -4.60 -2.03** .52 -.19 -3.91 fin. satisfaction -7.62** 1.42 -.25 -5.38 -6.69** 1.41 -.22 -4.75 inv. risk pref. -3.92** 1.28 -.13 -3.07 sa risk tolerance -6.88** 1.38 -.22 -4.99 note. *p < .01, **p < .001. an ordered logit regression was used to evaluate the second research hypothesis. the model was used to establish whether investment risk preference was positively associated with a respondent’s self-evaluation of their risk tolerance, controlling for gender, education, income, and financial satisfaction. the model shown in table 7 was statistically significant, χ2 = 72,436, p < .001. the model explained between 16% and 18% of the variance in self-assessment scores. investment risk preference was found to be positively associated with self-assessed risk tolerance. this finding adds additional support for the second hypothesis. additionally, each control variable was significant in the model. table 7. ordered logit regression: dependent variable self-assessed risk tolerance variable estimate se wald 95% conf. int. lower bound 95% conf. int. upper bound gender -.42* .17 5.88 -76 -.08 education .17* .07 6.18 .04 .30 income .11** .03 14.55 .05 .17 fin. satisfaction .26** .08 10.64 .10 .41 inv. risk pref. .20** .07 8.11 .06 .34 note. *p < .01, **p < .001 discussion this paper describes the development and testing of a domain specific investment risk-preference measure that blends aspects from revealedpreference tests with elements from statedpreference assessments. the resulting scale is one that provides insight into a financial decision maker’s investment preference based on their experiences and perceptions when making asset allocation choices. this measure offers a comparatively quick and valid way to assess the specific domain of investment risk preference. when assessing the practicality of this tool, it is worthwhile to distinguish between investment risk preference, risk tolerance, and risk aversion. risk tolerance reflects an individual’s willingness to accept financial risk, while risk aversion indicates a reluctance to take risks. in contrast, investment risk preference pertains to a decision-maker's subjective perception that one investment option is more favorable than another. as noted by nobre and grable (2015), someone’s preference is akin to a feeling. a preference is not a characteristic trait. this means that someone’s preference can change over time. for example, it is reasonable to anticipate that a financial decision-maker’s preference for investments that provide higher returns with corresponding more risk (i.e., variability in returns) will increase with experience, satisfaction, and expectations. when viewed this way, risk preference becomes an important input when describing someone’s risk profile, which is generally defined as a composite measure that portrays a person’s willingness to take risk that accounts for their perceptions, preferences, capacities, composure, and needs (brayman et al., 2017; hubble et al., 2020). financial services review, 33(1) 176 this distinction between investment risk preference, risk tolerance, and risk aversion highlights the unique contribution of the measure introduced in this study, which differs from the scf self-assessed risk item and other risktolerance assessments. although the models incorporating the investment risk-preference measure and the self-assessed stated-preference item produced comparable results, the investment risk-preference measure provides a notable advantage. specifically, it enables financial advisors to align a test-taker more accurately with investment products that align with their risktaking comfort level, offering a direct and practical application for portfolio personalization. this precise alignment underscores the significant value of the investment risk-preference measure introduced in this research. in this study, investment risk preference was found to be negatively associated with cash holding behavior. investment risk preference was also found to be positively correlated with selfassessed risk tolerance, attained education, income, and financial satisfaction. findings from this study suggest that the domain specific assessment tool presented in this paper provides a way to quickly estimate a person’s preference when allocating household investment resources. those with a low investment risk-preference scale score are expected to be more likely to hold a greater proportion of their household wealth in low return/low risk assets. on the other hand, as scale scores increase, it is reasonable to expect someone to hold proportionately more risky assets in their household portfolio(s). it is worth considering potential study limitations when evaluating the results of this research project. to begin with, the samples were not nationally representative. the samples were chosen to be descriptive of adults who are tasked with making household investment asset allocation decisions. this helps explain the relatively high degree of risk aversion exhibited by survey respondents. the average respondent’s preference for low-risk investments may have also been tied to the market environment when the surveys were distributed. the year 2022 marked the worst market for bonds in over 100 years. returns on equity investments were also negative. the markets in 2024, however, were generally positive, but overshadowed by a rigorously contested presidential election. it is possible that after incurring portfolio losses, and being uncertain about election outcomes, respondents in both surveys shifted their preference towards lower risk investments at the time of survey completion. this possibility can be checked by administering the test again during a bond and equity bull market that is unencumbered by a national election. it would also be beneficial for future studies to test scores in predicting risky asset ownership. whereas scores in this study were found to be robust in describing cash holding behavior it is possible that preferences are stronger in describing risk avoiding behavior than risk-taking behavior. this possibility is something that ought to be explored in future studies. nonetheless, findings from this study indicate that the domain specific investment risk-preference measure does appear to offer a unique insight into a financial decisionmaker’s predilection when making asset allocation choices. references arslan, r. c., brümmer, m., dohmen, t., drewelies, j., hertwig, r., & wagner, g. g. (2020). how people know their risk preference. scientific reports, 10, 15365. 10.1038/s41598-020-72077-5 atlas, s. a., lu, j., micu, p. d., & porto, n. (2019). financial knowledge, confidence, credit use, and financial satisfaction. journal of financial counseling and planning, 30(2), 175190. barsky, r. b., juster, f. t., kimball, m. s., & shapiro, m. d. (1997). preference parameters and behavioral heterogeneity: an experimental approach in the health and retirement study. quarterly journal of economics, 112(2), 537-579 bowen, d. h., buck, s., deck, c., mills, j. n., & shuls, j. v. (2015). risky business: an analysis of teacher risk preferences. education economics, 23(4), 470-480. brayman, s., grable, j. e., griffin, p., & finke, m. (2017). assessing a client’s risk https://doi.org/10.1038%2fs41598-020-72077-5 kwak & grable 177 profile: a review of solution providers. journal of financial service professionals, 71(1), 71-81. brink, a. g., & rankin, f. w. (2013). the effects of risk preference and loss aversion on individual behavior under bonus, penalty, and combined contract frames. behavioral research in accounting, 25(2), 145-170. cardak, b. a., & martin, v. l. (2019). once in a lifetime? the effects of the global financial crisis on household willingness to take financial risk. economic record. doi:10.1111/1475-4932.12506 charness, g., gneezy, u., imas, a. (2013). experimental methods: eliciting risk preferences. journal of economic behavior & organization, 87, 43-51. dare, s. e., van dijk, w. w., van dijk, e., van dillen, l. f., gallucci, m., & simonse, o. (2020). the road to financial satisfaction: testing the paths of knowledge, attitudes, sense of control, and positive financial behaviors. journal of financial therapy, 11(2), 1-30. davies, g. (2016). how not to measure risk tolerance. https://www.linkedin.com/pulse/howmeasure-risk-tolerance-greg-b-davies dickason, z., & ferreira, s. j. (2018). gender and behavior: the effect of gender and ethnicity on financial risk tolerance in south africa. gender and behaviour, 16, 10851-10862. ertac, s. (2020). the formation and malleability of preferences and noncognitive skills. springer. feng, l., & seasholes, m. s. (2005). do investor sophistication and trading experience eliminate behavioral biases in financial markets? review of finance, 9(3), 305351. frey, r., pedroni, a., mata, r., & hertwig, r. (2017). risk preference shares the psychometric structure of major psychological traits. science advances, 3(10). doi: 10.1126/sciadv.1701381 grable, j. e. (2000). financial risk tolerance and additional factors that affect risk taking in everyday money matters. journal of business and psychology, 14, 625-630. grable, j. e., & joo, s-h. (2004). environmental and biopsychosocial factors associated with financial risk tolerance. journal of financial counseling and planning, 15(1), 73-82. grable, j. e., & lytton, r. h. (2001). assessing the concurrent validity of the scf risk assessment item. financial counseling and planning, 12(2), 43-52. grable, j. e., kwak, e-j., fulk, m., & routh, a. (2020). a simplified measure of investor risk aversion. journal of interdisciplinary economics, 34(1), 734. doi.org/10.1177/0260107920924518 guiso, l., & sodini, p. (2013). household finance: an emerging field in g. m. constantinides, m. harris, & r. m. stulz (eds.) (pp. 799-1612), handbook of the economics of finance, volume 2 part b. elsevier: amsterdam. hallahan, t. a., faff, r. w., & mckenzie, m d. (2004). an empirical investigation of personal financial risk tolerance. financial services review, 13, 57-78. hanna, s. d. & lindamood, s. (2004). an improved measure of risk aversion. journal of financial counseling and planning, 15(2), 27-45. hanna, s. d., gutter, m. s., & fan, j. x. (2001). a measure of risk tolerance based on economic theory. journal of financial counseling and planning, 12(2), 53-60. hanna, s. d., kim, k. t., & lindamood, s. (2018). factors related to the risk tolerance of households in china and the united states: implications for the future of financial markets in china. financial services review, 27(3), 279-302. hansson, h., & lagerkvist, c. j. (2011). measuring farmers’ preferences for risk: https://doi.org/10.1111/1475-4932.12506 https://doi.org/10.1177%2f0260107920924518 financial services review, 33(1) 178 a domain-specific preference scale. journal of risk research, 15(7), 737753. harnum, k., cooper, t., & faseruk, a. (2010). a review of the psychology of risk-taking behavior for individual investors. journal of finance issues, 8(1), 19-28. hartnett, n., gerrans, p., & faff, r. (2019). trusting clients’ financial risk tolerance survey scores. financial analysts journal, 75(2), 91-104. hertwig, r., wulff, d. u., & mata, r. (2018). three gaps and what they may mean for risk preference. philosophical transactions of the royal society b, 374(1766). doi.org/10.1098/rstb.2018.0140 hubble, a., grable, j. e., & dannhauser, b. (2020). investment risk profiling: a guide for financial advisors. cfa research reports. https://www.cfainstitute.org/en/research/ industry-research/investment-riskprofiling kim, s. d., cotwright, m., & chatterjee, s. (2019). who are robo-advisor users? journal of finance issues, 18(2), 33-50. koekemoer, z. (2018). the influence of demographic factors on risk tolerance for south african investors. proceedings of the international academic conferences, 6408640, international institute of social and economic sciences. larkin, c., lucey, b. m., & mulholland, m. (2013). risk tolerance and demographic characteristics: preliminary irish evidence. financial services review, 22, 77-91. mata, r., frey, r., richter, d., schupp, r., & hertwig, r. (2018). risk preference: a view from psychology. journal of economic perspectives, 32(2), 155-172. mudzingiri, c., & koumba, u. (2021). eliciting risk preferences experimentally versus using a general risk question: does financial literacy bridge the gap? risks, 9, 140. doi.org/10.3390/risks9080140 nobre, l. h. n., & grable, j. e. (2015). the role of risk profiles and risk tolerance in shaping client decisions. journal of financial service professionals, 69(3), 18–21. pinjisakikool, t. (2017). the influence of personality traits on households’ financial risk tolerance and financial behavior. journal of interdisciplinary economics, 30, 32-54. rabbani, a. g., & nobre, l. h. (2022). financial risk tolerance. in de gruyter handbook of personal finance, pp. 137-156. john e. grable and swarn chatterjee (eds.). de gruyter. richter, m. k. (1966). revealed preference theory. econometrica, 34(3), 635-645. robb, c. a., babiarz, p., & woodyard, a. (2012). the demand for financial professionals' advice: the role of financial knowledge, satisfaction, and confidence. financial services review, 21(4), 291-306. samuelson, p. a. (1938). a note on the pure theory of consumers’ behavior. economica. new series, 5(17), 61-71. wong, a. (2011). financial risk tolerance and selected demographic factors: a comparative study in 3 countries. global journal of finance & banking issues, 5(5), 1-12. https://www.cfainstitute.org/en/research/industry-research/investment-risk-profiling https://www.cfainstitute.org/en/research/industry-research/investment-risk-profiling https://www.cfainstitute.org/en/research/industry-research/investment-risk-profiling 1 launching a cfp board registered program at an aacsbaccredited business college: a case study and analysis william b. elliott1 & xianwu zhang2 abstract despite increasing demand for financial planning education, the discipline remains underrepresented within aacsb-accredited business schools—even though such programs align well with business curricula. this study examines the distribution of cfp board-registered programs across u.s. institutions, analyzing 395 programs (certificates, bachelor’s, master’s, and doctoral degrees). our findings reveal that only 39% (n=153) are housed within aacsbaccredited business schools, with a predominant focus on in-person or blended bachelor’s degrees. in contrast, non-business schools more frequently offer online certificates and exclusively host all three doctoral programs. geographic analysis identifies substantial disparities in program availability relative to state populations. we present a case study of successful implementation of a cfp board-registered program at an aacsb-accredited business school. the discussion outlines key strategies for program development, accreditation alignment, and institutional challenges, providing actionable insights for business schools seeking to expand into financial planning education. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation elliott, w. b. & zhang, x. (2025). launching a cfp board registered program at an aacsbaccredited business college: a case study and analysis. financial services review, 33(4), 1-11. introduction and relevant literature we examine the number and location of all “cfp board registered” programs in the united states, analyzing program format, delivery method (online, face-to-face, hybrid), and level (undergraduate or graduate). as of february 2025, 395 programs were recognized by the cfp board as “registered programs”. we also provide a case study of developing and offering a cfp board registered program in an aacsb college of business, discussing both the advantages and challenges encountered during development and delivery. in 2024, approximately 326,000 professionals worked as “personal financial advisors” to “help 1 john carroll university, university heights, ohio, usa. 2 corresponding author (xzhang@jcu.edu). john carroll university, university heights, ohio, usa. individuals manage their money and plan for their financial future” (u.s. bureau of labor statistics, n.d.). less than one-third (around 106,109 by september 1, 2025) of these financial planners hold the certified financial planner™ credential (cfp® certification), which represents the standard of excellence in financial planning. the cfp board oversees this credential and promotes professionalism through rigorous education, training, and ethical standards. post-secondary academic programs serve as crucial resources for developing additional trained financial planners. over the past 20 years the number of university-level programs has https://creativecommons.org/licenses/by-nc/4.0/ mailto:xzhang@jcu.edu financial services review, 34(4) 2 increased significantly. however, many current financial planning programs are not housed in colleges of business but rather in colleges of human sciences or social sciences. the curricula commonly found in colleges of business have greater overlap with cfp board requirements, particularly the set of core business classes common in business programs and advanced courses in tax and investments. the typical bsba degree program includes core classes highly relevant to financial planning. introductory courses in economics, accounting, finance, and statistical methods relate directly to financial planning, while other typical business core courses in marketing and management are indirectly relevant. given the apparent natural synergies between the common body of knowledge for financial planning programs and other programs within business schools, we find the relatively limited number of programs in business schools puzzling, especially considering increasing demand for new and innovative programming in business education. we argue that adding these core classes, typically only available in business college programs – along with coursework specifically in financial planning, would improve the quality of graduates pursuing financial planning careers. existing research has primarily examined various approaches to improving financial planning education quality through curriculum content changes or surveys of cfp program directors. goetz et al. (2005) argue that an effective cfp program brings the profession into the classroom and should shorten the transition period for students moving from academic to professional environments. they provide a comprehensive list of techniques which they argue can help shorten that transition. to keep financial planning education current, salter et al. (2011) assessed the topics and skills offered by the academic curriculums at cfp® degree-granting institutions and surveyed financial planning professionals to determine desired expertise levels and importance of financial planning topics for entrylevel planners who are recent cfp program graduates. they present the foundation for effective financial planning curriculum design in higher education, identifying personal skills and qualities, as well as investment planning, as both important and highly desired competencies. most recently, heymann et al. (2025) explored factors influencing the success of cfp board registered financial planning programs in the higher education through comprehensive surveys of cfp board program directors and in-depth qualitative insights from the authors’ own experience. they discuss key elements of successful programs including curriculum design, faculty recruitment, and institutional support. they identify critical success factors including experiential learning opportunities, outreach efforts for diverse talent, and networking events. in this paper, we examine all cfp board registered financial planning programs in the united states. we find a surprisingly low percentage of programs housed in schools or colleges of business. we also provide guidance and strategies for developing and offering financial planning programs in business schools. specifically, this paper examines the distribution of cfp board registered programs across u.s. institutions, analyzing 395 programs (certificates, bachelor’s, master’s, and doctoral degrees) based on the format, delivery method, geographical location, whether the program is housed in a business school or not, and program density. the latter portion presents a case study of how the authors proposed, developed and successfully launched a cfp board registered program at an aacsb accredited business school – the boler college of business at john carroll university (jcu). jcu is a relatively small jesuit university located in cleveland’s eastern suburbs. we provide practical strategies and discuss the challenges related to offering cfp registered programs at aacsb accredited business schools. data according to the bureau of labor statistics (bls), demand for personal financial advisors is projected to increase by 17% between 2023 and 2033. this growth rate significantly exceeds the average for all occupations (4%), and the average for financial positions (6%). during the next decade, the bls projects approximately 55,000 annual job openings for personal financial advisors. some openings result from workers switching occupations or retiring. elliott & zhang 3 the supply of personal financial planners is not keeping pace with demand, especially for financial planners holding the cfp® certification – the industry’s gold standard. in 2013, there were just under 70,000 cfp ® holders, by 2024, approximately 100,000 held the certificate. financial planning graduates typically work in financial institutions (the 12 tribes of financial planning, n.d.). required coursework for financial planning majors includes various business classes as prerequisites, including business finance, and economics. however, prerequisites may vary depending on whether the program is housed in a business college or nonbusiness college. non-business schools may not require the same variety of business classes as prerequisites for financial planning majors. financial planning major core classes, especially in cfp board registered programs, include foundation of personal financial planning, tax planning, investment planning, estate planning, risk management and insurance, retirement planning, and a capstone course. our data comes from two sources. all academic financial planning program information was hand collected by the authors for cfp board registered programs listed on the cfp board website (cfp board of standards, n.d.). variables of interest include program types (certificate, bachelor’s, graduate, etc.), program format (delivery method such as in-person or online), program location (state), and number of cfp® certification holders in each state. population data comes from the bureau of labor statistics, using the latest 2024 population survey available. methods our analysis is descriptive and comparative in nature. the descriptive analysis provides national level summary statistics of all cfp board registered programs in the united states regarding program type, delivery method, and state-by-state program numbers. we conducted state-level per-capita comparative analysis on each state’s population and cfp programs. additionally, we compare cfp programs offered by business schools and non-business schools to gain a deeper understanding of differences between the two groups. results as of february 2025, 395 cfp board registered financial planning programs existed in the united states. figure 1 shows the total number of each cfp program type (certificate, bachelor’s, minor, graduate, and doctoral). specifically, a certificate is a non-degree certificate, where students complete specified courses to receive the certificate. as shown in figure 1, 154 programs (39%) led to certificates, 164 (41%) were undergraduate programs, 23 (6%) led to undergraduate minors, 51 (13%) were graduate programs, and 3 (1%) were doctoral programs. figure 2 shows the total number of each delivery format, including classroom/blended, instructor led online, and self-study. self-study refers to delivery methods where students watch prerecorded lectures and complete quizzes or tests via online platforms, without direct instructor involvement. regarding delivery format, 265 programs (67%) were offered in a classroom/blended format, 117 (30%) as online instructor-led format, and 13 (3%) in online selfstudy format. 4 figure 1. types of cfp® programs figure 2. delivery method for cfp® programs 154 164 23 51 3 0 20 40 60 80 100 120 140 160 180 certificate bachelor bachelor-minor graduate phd type of cfp programs 265 117 13 0 50 100 150 200 250 300 classroom/blended online instructor-led online self study formats of cfp programs elliott & zhang 5 figure 3 shows the number of cfp board registered programs in each state. geographically, they are unevenly spread across states. the top three states with the most programs are pennsylvania, california, and texas. three states – alaska, connecticut, and south dakota – have no cfp board registered programs. figure 4 maps cfp® certification holders by state. similar to cfp programs, there is also substantial variation in cfp® certificate holders across states, with california, texas, and florida having the highest numbers. figure 3. cfp programs by state figure 4. cfp® certificate holders by state the previous section does not consider overall state populations. figure 5 shows cfp programs per million inhabitants. on a per capita basis, the highest concentration of cfp programs (measured as number of cfp programs divided by state population in millions) were in rhode island, utah, the district of columbia, west virginia, and kansas. rhode island has the highest density of programs, with 3.6 programs per million residents (see figure 5). three states – alaska, connecticut, south dakota – have no cfp programs while louisiana, washington, and nevada have very low concentrations. interestingly, california, texas, and florida also have relatively lower program concentrations despite their large populations. figure 5. number of cfp programs per million by state figure 6 shows cfp® certification holders per million inhabitants. we find the highest concentrations in colorado (538 planners/million), new hampshire, massachusetts, minnesota, and connecticut. mississippi and west virginia have the fewest cfps per capita, with only 87 per million. some variation is likely to result from income and net worth differences across the 50 states; however significant variance would likely remain even after controlling for income and wealth. we do not opine on the optimal number of cfp programs, or cfp® certificate holders, for particular locales. however, given the wide ranges observed, some states may benefit from increasing the number of cfp board registered programs and cfp® certificate holders. financial services review, 34(4) 6 figure 6. number of cfp® certificate holders per million by state figure 7 shows the number of programs in business colleges with and without accreditation as well as programs in non-business colleges. business schools may be accredited by different accrediting bodies indicating varying requirements in research, teaching and service areas. three accreditation agencies include the association to advance collegiate schools of business (aacsb), the accreditation council for business schools and programs (acbsp), and the international accreditation council for business education (iacbe). of 395 total programs, only 153 (39%) are offered within 110 aacsb-accredited business schools. an additional 49 (12%) programs are offered in 28 acbsp-accredited business schools and colleges. there are 19 (5%) programs in 15 iacbe-accredited business schools and colleges, and 37 (9%) programs in 28 non-accredited business schools. the remaining 137 (35%) programs are housed in 83 non-business schools and colleges. for example, financial planning programs at texas tech university are housed in the school of financial planning in the college of human science and financial planning programs at the university of georgia are housed in the college of family and consumer science. figure 8 shows delivery format comparisons between aacsb schools and non-business schools. significant differences exist in both program delivery format and cfp program types between aacsb accredited business schools and non-business schools. aacsb-accredited schools heavily emphasize classroom/blended format (81%), while non-business schools split evenly between classroom/blended (50%) and online (50%) formats. figure 9 shows that aacsb accredited business schools tend to offer more bachelor’s degree programs (54%) compared to certificates (27%), while non-business school programs tend to offer more certificate programs (62%) compared to bachelor’s degrees (18%). notably, all three cfp board registered doctoral programs are housed in non-business schools. figure 7. types of colleges and schools offering cfp® programs 110 28 15 28 83 0 20 40 60 80 100 120 aacsb acbsp iacbe b-school-non-accredited non-business schools that offer cfp board registered programs 7 figure 8. formats of cfp programs between aascb schools and non-business figure 9. types of cfp programs between aascb schools and non-business a case study: building a program in an aacsb college of business in this section, we discuss the process and strategies used to overcome challenges that the authors faced during development of a new financial planning major in an aacsbaccredited college of business. issues discussed include creating an advisory board, staffing the new program, integrating new faculty and discipline within an existing department, initial program roll-out, program leadership and oversight, and marketing the new program. initial planning, curriculum development, and due diligence 81 % 14 % % 5 % 49.6 % 46.7 3.6 % % 0 % 10 % 20 30 % 40 % 50 % 60 % 70 % % 80 % 90 classroom/blended online instructor-led online self study formats of cfp programs between aacsb business schools and non business schools aacsb non-business financial services review, 34(4) 8 john carroll university is a relatively small (approximately 2,500 fte), private, jesuit institution with a college of arts and sciences, the boler college of business (aacsb accredited), and a newly created college of health sciences. about 35% of our student body are business majors. prior to adding the financial planning and wealth management (fpwm) program, our college offered majors in accounting, economics, finance, international business, marketing, management/hr, and supply chain. in the very early planning stages for the new major, we reviewed the employment data of our recent alumni using in-house data from senior exit surveys and linkedin data. because of our location, near east suburbs of cleveland, ohio, more than 50 publicly traded corporations are headquartered or have a major corporate presence within an hour of downtown. consequently, many of our students take positions in corporate finance. however, we found that approximately 30% of students took positions in financial planning or wealth management, employed by both local and regional planners as well as major national firms. our finance major curriculum focuses on developing skills related to corporate finance and investment banking. while some overlap exists between skillsets required for student success in corporate/investment banking careers and those required of financial planners and wealth managers, each field has unique subjects. prior to the new program, our finance major did not meet cfp board criteria to become a cfp board registered program. while neither necessary, nor sufficient, knowing how many current finance majors chose financial planning careers was useful and gave us confidence that additional coursework developed for a new fpwm major would gain reasonable enrollment early on. however, some enrollment would likely come at the expense of current finance major enrollment. our next critically important step involved contacting executives and owners of financial planning/wealth management firms in the region. we provided them with a brief overview of our proposed program and asked them to join an advisory board. everyone who was contacted, agreed to join, and we scheduled an initial meeting. we believed that organizing the advisory board early in the process was important to take full advantage of their recommendations and build a strong affiliation between the business community and our new fpwm program. our advisory board goals were threefold. first, we hoped to use their experience to gain insights into employee needs at their firms (particularly entry-level positions needed over five years) and what skills and knowledge those employees needed for professional success. second, we hoped to use the advisory board as a vehicle for building lasting relationships between our new program and their businesses, providing potential for in-class speakers, adjunct professors, and internship and job opportunities for future graduates. finally, we hoped that engaging the business community early and staying in touch would make them more likely to make financial contributions supporting the fpwm program. despite being less than an hour from an older, well-known, and established financial planning program (at the university of akron), our advisory board still had difficulty finding enough qualified applicants for openings at their firms. as a result, they clearly encouraged us to begin a new program and committed to working with us to provide our students with internships and job opportunities. we quickly realized that while significant overlap exists between the coursework of our corporate finance major and a fpwm major, significant gaps also existed. even factoring in the existing courses in our accounting department, we needed to develop several new courses. furthermore, we lacked internal academic expertise in areas relating to retirement planning, estate planning, and risk management and insurance, among others. although our corporate finance major is the largest program on campus, we deliver that program with a very lean faculty base. since we planned to house the fpwm program within a blended department of economics and finance and we were under-resourced on the finance program, we hoped to find a new faculty member with experience in both corporate finance and financial planning. through a national search, elliott & zhang 9 we found a person with sufficient background in both areas. initially, the new hire was a full-time visiting position, with an option to convert to a tenure-track line if the new program attracted sufficient enrollment. staffing the new program the new full-time visitor would not only develop and teach many of the courses in the fpwm program but also serve as program director. we believed that because this was a new program with content outside of our typical departmental expertise, having a director was important. the director was charged with guiding curriculum development and course staffing, and promoting the program, both within the university and to outside stakeholders. in particular, the director would: 1. set strategic initiatives for the fpwm program. 2. promote the fpwm program and guide marketing material development. 3. guide program review and evaluation for internal constituents as well as the higher learning commission (jcu’s universitylevel accrediting body) and aacsb. 4. maintain cfp board registration of the program. 5. approve petitions for course substitutions. 6. work with appropriate parties to develop internships and potential student placement. the director/faculty member also had to be integrated into the department and college. a nontrivial challenge we encountered in integrating a new discipline into the department related to academic research. since many financial planning programs, particularly at the doctoral level, are not housed in colleges of business, journal lists typically used in business colleges may not recognize all journals valued by financial planning researchers. some journals on typical business school lists occasionally publish financial planning research, but it is very limited. because the field is relatively new, many journals valued in the financial planning discipline are not yet included by indexing services (meaning they lack citation impact factors) nor are they tracked by many ranking lists such as the australian business dean’s council (abdc). our college’s journal list was constructed using externally validated ranking lists. we use several ranking lists collated by jean harzing (e.g. abdc, abs, hec), the financial times 50 (ft50), and the ssci 5-year impact factor. because many financial planning field journals are either relatively new or not on the businessrelated lists curated by harzing, they do not already appear on our college list. to ensure that our new financial planning colleagues had journals in their field represented on our college list, we chose to use an additional external ranking. we used a ranking provided by grable & ruiz-menjivar (2020). they adopt chen and huang’s (2007) ‘author affiliation’ method and apply it to financial planning journals. as a result, we added eight financial planning journals identified by grable and ruiz-menjivar as “core” household and personal finance journals. the most significant opposition to adding these journals came from colleagues in other disciplines where publishing in top-tier journals is extremely challenging. however, we argued that acceptance rates and publishing difficulty in any college or department will vary across disciplines. after lengthy discussion, the addition of the financial planning “core” journals passed with a majority vote of our college research committee. curriculum and cfp board registration when designing a new financial planning program, it makes sense to design the curriculum to meet all cfp board criteria for becoming a cfp board registered program. this would allow us to differentiate our program from other regional institutions with majors, concentrations, tracks, or certificates in financial planning that lacked cfp board registration. not only does it allow graduates to automatically meet cfp board educational requirements but also improves program positioning relative to other institutions and most importantly provides students with a high-quality major. new program approval at many universities is a lengthy and time-consuming process, and our university is no different. however, we were under pressure to launch the program sooner rather than later. to begin offering coursework as quickly as possible, we decided to create some financial services review, 34(4) 10 classes (which required minimal approvals) and make those classes part of a ‘concentration’ in our existing finance major. at jcu, adding a concentration to an existing major was a much more streamlined process. this meant we could better assess demand for the new fpwm major and begin making students aware of the new program sooner. we have already offered several courses containing some of the cfp board’s prescribed common body of knowledge. our introductory finance course, business finance, as well as a course in investments and several economics courses (all part of our corporate finance major and business core requirements), included cfp board required topics. we added a new introductory course in financial planning and classes in tax planning and estate planning. originally, we hoped to use an existing personal tax course offered by our accounting department, however, it was determined that there was insufficient overlap and insufficient room to include required financial planning content. one of our advisory board members employed a highly experienced attorney who was willing to teach the tax and estate planning courses. a potential disadvantage with the new concentration was that it could create confusion for students once the standalone fpwm major was approved. for that reason, we only offered the concentration for one year and then cancelled it as soon as the fpwm major was approved. at that time, because of overlap between our corporate finance and the fpwm majors, we developed a ‘pathway’ for students who wish to double major in both corporate finance as well as fpwm (and still graduate within four years). once the new major was in the university’s bulletin, our primary challenge has been ensuring that students are aware of the significant differences between our corporate finance major and the new fpwm program. jcu’s advising model involves both professional advisors (during their first two years) and faculty advisors (during the last two years of study). consequently, most messaging and outreach to students has fallen to the professional and faculty advisors. we also expect to develop video content that will help students make more informed decisions about which pathway/major is right for them. conclusion as of 2024, 395 cfp board registered financial planning programs exist in the united states. delivery mode and degree type vary widely among these programs. thirty-nine percent were certificate programs (i.e. non-degree programs), 41% were undergraduate majors, 6% led to an undergraduate minor, 13% were graduate programs, and 1% were doctoral programs. twothirds of programs were offered in a classroom/blended format, 30% were online instructor-led format, and only 3% in online selfstudy format. of the 395 programs, only 39% are offered within aacsb colleges of business while 35% of programs are housed in non-business colleges. considerable variation exists in the concentration of financial planning programs and cfp® certificate holders across different states. rhode island, utah, the district of columbia, west virginia, and kansas have the highest concentration of financial planning programs, with up to four programs per million residents. alaska, connecticut, and south dakota have zero cfp board registered programs. similar variation exists across states regarding cfp® certificate holders. colorado, new hampshire, massachusetts, minnesota, and connecticut have the highest concentration of cfp® professionals. colorado has the most, with 538 cfps per million residents. mississippi and west virginia have only 87 cfp® professionals per million residents. nearly 40% of programs are offered through aacsb accredited business schools. business schools accredited by other accrediting bodies account for another 17%, and 9% of programs are offered by business schools with no specific business accreditation. all remaining 35% of cfp registered programs are housed in nonbusiness schools and colleges. we conclude with a case study describing the process the authors used to assess the need for, plan, develop curriculum, hire faculty, and deliver a cfp board registered program at john carroll university. there are advantages as well elliott & zhang 11 as challenges to housing a financial planning program in an aacsb-accredited college of business. in our experience the advantages outweigh the disadvantages. in particular, typical business core course offerings provide a very strong business foundation for future financial planners. furthermore, some courses likely already deliver a significant percentage of content required by the cfp board criteria. the primary challenges are as follows. first, it may be necessary to add additional publications to a business college’s journal list to accommodate the faculty in a new discipline. second, care must be taken to clearly message potential students about the difference between corporate finance/finance majors and the financial planning and wealth management major. finally, to ensure a clear pathway for graduating seniors to transition into the workplace, engaging with regional professionals early in program development (with the aid of an advisory board) will ensure solid relationships with potential employers. references certified financial planner board of standards. (n.d.). registered programs. https://www.cfp.net/for-educationpartners/registered-programs chen, c. r., & huang, y. (2007). author affiliation index, finance journal ranking, and the pattern of authorship. journal of corporate finance, 13(5), 1008–1026. https://doi.org/10.1016/j.jcorpfin.2007.04.0 11 dean, l. (host), harness, n. (host), & lemoine, c. (host). (n.d.). the 12 tribes of financial planning [audio podcast]. utah valley university; texas a&m university; university of illinois. goetz, j. w., tombs, j. w., & hampton, v. l. (2005). easing college students’ transition into the financial planning profession. financial services review, 14, 231–251. grable, j. e., & ruiz-menjivar, j. (2020). household and personal finance journal rankings using patterns of authorship and the author affiliation index. http://ssrn.com/abstract=2570891 heymann, r., schnusenberg, o., & timmerman, i. (2025). mapping the terrain: strategies for building effective university financial planning programs. financial planning review, 8(1). https://doi.org/10.1002/cfp2.1198 salter, j. r., hampton, v. l., winchester, d., katz, d. b., & evensky, h. r. (2011). entry-level financial planning practice analysis: preparing students to hit the ground running. financial services review, 20(3), 195–216. u.s. bureau of labor statistics. (n.d.). personal financial advisors. u.s. department of labor. https://www.bls.gov/ooh/businessand-financial/personal-financialadvisors.htm https://www.cfp.net/for-education-partners/registered-programs https://www.cfp.net/for-education-partners/registered-programs https://doi.org/10.1016/j.jcorpfin.2007.04.011 https://doi.org/10.1016/j.jcorpfin.2007.04.011 http://ssrn.com/abstract=2570891 https://doi.org/10.1002/cfp2.1198 https://www.bls.gov/ooh/business-and-financial/personal-financial-advisors.htm https://www.bls.gov/ooh/business-and-financial/personal-financial-advisors.htm https://www.bls.gov/ooh/business-and-financial/personal-financial-advisors.htm pii: s1057-0810(00)00063-9 risk tolerance and asset allocation for investors nearing retirement govind hariharana, kenneth s. chapmanb, dale l. domianc,* aschool of management, university at buffalo, buffalo, ny 14260, usa bcalifornia state university, northridge, northridge, ca 91330, usa ccollege of commerce, university of saskatchewan, saskatoon, sk s7n 5a7, canada received 17 may 1999; received in revised form 6 march 2000; accepted 6 march 2000 abstract this paper uses a large individual-level data set to isolate the effects of risk tolerance on portfolio composition. we test and confirm two predictions of the capital asset pricing model: (1) increased risk tolerance reduces an individual’s propensity to purchase risk-free assets; and (2) higher risk tolerance does not affect the composition of an individual’s portfolio of risky assets. more specifically, we find that risk tolerant investors nearing retirement do not reduce their bond allocations in order to buy more stock. © 2000 elsevier science inc. all rights reserved. jel classification:g11; d12 keywords:portfolio choice; asset allocation; risk aversion 1. introduction canner, mankiw and weil (1997) suggest that wall street financial planners often recommend a different mix of financial assets for highly risk tolerant clients than for more risk averse individuals. risk tolerant investors should buy more high risk, high expected return assets such as stocks, than low risk, low expected return assets such as bonds. as plausible as this advice may sound, it differs markedly from the behavior of the rational expected-utility-maximizing investors inhabiting the capital asset pricing model (capm) * corresponding author. tel.:11-306-966-8425; fax:11-306-966-2515. e-mail address: domian@commerce.usask.ca (d.l. domian). financial services review 9 (2000) 159–170 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00063-9 used in most finance and economics research. all risky assets–including both stocks and bonds–are part of the “market portfolio” in the capm. increased tolerance for risk causes capm investors to alter the percentage of assets held in the risky portfolio as opposed to the risk-free asset, but the composition of stocks and bonds held in the risky portfolio is not changed. this paper investigates the behavior of investors nearing retirement. more specifically, we focus on two hypotheses implicit in the capm: (1) risk tolerant individuals hold a smaller proportion of risk-free assets, and (2) the composition of an individual’s portfolio of risky assets will not change as he/she becomes more risk tolerant. we examine survey data reported in the 1992health and retirement survey,concerning individuals’ asset allocations and willingness to take risk as they approach retirement. each person’s portfolio is decomposed into assets held as stock, bonds and treasury bills. assuming that treasury bills are “risk-free assets” while stocks and bonds are “risky,” the capm’s predictions become (1) the proportion of all assets allocated to treasury bills should fall as risk tolerance increases, and (2) the proportion of risky assets allocated to stocks is independent of risk tolerance. wall street seems to be comfortable with the first of these predictions, but not the second. we find support for both predictions. 2. literature review in recent years, both individual and institutional investors have become increasingly aware of the importance of asset allocation. brinson, hood and beebower (1986) show that over 90% of the variability in portfolio returns can be explained by asset allocation. other studies focus on decisions during particular stages of investors’ life cycles. butler and domian (1993) present asset returns over long holding periods in a form useful for preretirement planning. ho, milevsky and robinson (1994) examine how to maximize the probability of a secure and sufficient income during postretirement years. friend and blume (1975) observe that an individual’s risk tolerance can be inferred from the asset allocation decision by calculating the percentage of a person’s assets invested in risky securities. this approach was extended by siegel and hoban (1982, 1991), morin and suarez (1983), bellante and saba (1986), riley and chow (1992) and others. in a recent variant of this approach, bajtelsmit, bernasek and jianakoplos (1999) presents a version of the capital asset pricing model that allows individuals to allocate their funds between risky assets, a risk-free asset and human capital. the proportion of assets investori devotes to risky securities,ai is determined by the following equation: ai 5 se~rm 2 r f! sm 2 ds 1 ci ds 1 1 2 hi d . (1) the first bracketed term is the same for all individuals and consists of the expected difference in the rate of return between the market portfolio and the risk-free asset,e(rm 2 r f), divided by the variance of the market portfolio,sm 2 . the second term suggests thatai is larger for risk tolerant investors sinceci is personi’s relative risk aversion. in the final term,hi is the ratio of human wealth to net wealth. consequently investors with high human capital investments 160 g. hariharan et al. / financial services review 9 (2000) 159–170 hold larger fractions of their wealth in risky assets. bajtelsmit et al. use information onai and hi to infer a value for risk tolerance multiplied by a constant that is the same for all investors. while the studies described above provide useful insights, they do not test the capm’s asset allocation predictions because they have already assumed the result. one reason why these capm predictions are seldom tested is that there are very few direct measures of risk tolerance available. a few empirical studies have uncovered more direct information. viscusi (1992), for example, infers risk tolerance from a willingness to undertake risky endeavors in other areas of life. it would seem to follow that an individual choosing to race automobiles for a living is relatively tolerant of risk. if so, his/her portfolio should contain a small proportion of risk-free assets, perhaps even a negative proportion (e.g., buying stocks on margin). if the capm is correct, it is also true that the portion of the racing enthusiast’s portfolio allocated to risky securities will have the same composition as that of your average professor. of course, many characteristics differentiate our putative race car driver from others in the population. perhaps the fact that he/she isn’t likely to live as long makes him/her less likely to value stocks with large future, but uncertain current, expected returns. all of this makes the approach interesting, but not ideal. many things other than financial risk tolerance affect willingness to engage in other sorts of risky behavior. lebaron, farrelly and guha (1989) and schooley and worden (1996) obtain a measure of risk tolerance by survey. while lebaron et al. have only a small sample and little information other than the risk tolerance, the 1989 survey of consumer finance (scf) used by schooley and worden is more complete. the scf asked participants “which of the following statements comes closest to the amount of financial risk that you (and your husband/wife) are willing to take when you save or make investments? 1. take substantial financial risks expecting to earn substantial returns 2. take above average financial risks expecting to earn above average returns 3. take average financial risks expecting to earn average returns 4. not willing to take any financial risks.” schooley and worden regress the share of risky assets on dummy variables for the answers to this scf question. as the capm predicts, risk tolerant investors hold a smaller proportion of risk-free assets and more of the risky portfolio. our paper substantially replicates their result with a new data set and a different measure of risk tolerance. unlike schooley and worden, our paper tests the capm notion that the composition of the risky portfolio does not change as risk tolerance increases. 3. data and methods the data for our study come from the first wave of the health and retirement survey (hrs) conducted by the university of michigan’s survey research center in 1992. more information on this data set is available athttp://www.umich.edu/;hrswww/center/ center.html.the hrs is a nationally representative sample of 15,000 individuals aged 51–61. the primary objective of this data set was to collect longitudinal information on the health, financial well being and labor market decisions of people approaching retirement. to this end, the hrs will track the same individuals every two years over a prolonged period 161g. hariharan et al. / financial services review 9 (2000) 159–170 stretching at least into the first decade of the 21st century. such a longitudinal data set, it is expected, will provide a wealth of information that can track changes in health status, financial status and labor market status of these individuals and the reasons for such a change. the first wave of data, which is the one used in this paper, was collected in 1992. for the first wave, in addition to the core set of questions on health, wealth and employment characteristics, a set of ten experimental modules collected information on key topics of interest including parents’ wealth, and spending and saving preferences. our interest in the hrs stems from three pieces of information contained in it: (1) information on assets broken down into stocks, bonds, and so forth, (2) information on expectations regarding inflation, economic depression and bequest motives, and (3) information on risk tolerance of the respondents. the core data provide information on the amount of the respondents’ nonhousing wealth invested in stocks of publicly held companies, mutual funds or investment trusts, bonds including corporate, government, and foreign bonds, as well as money in certificates of deposit, government savings bonds and treasury bills. one shortcoming of the information on the asset variables is that finer distinctions, for instance the type of bonds, are not available in this data set. the hrs also provides responses to questions about respondents’ expectations on inflation and depression in the economy, the length of their planning horizon, and the strength of their bequest motives. the inclusion of a survey question intended to assess risk tolerance was of particular interest to us in examining the role of risk tolerance on asset allocation decisions. respondents were asked to suppose that they were the only income earner in the family with a good job guaranteed to provide their current family income for life. they were then asked if they would accept an opportunity to take a new and equally good job with a 50–50 chance of doubling family income and a 50–50 chance that family income would be reduced by a third. depending on their response, individuals were next asked about their willingness to take a job that had a 50–50 chance of doubling their income and a 50–50 chance of either halving or reducing their income by 20%. from the two questions, we can obtain an index of risk tolerance with four values that range from zero (least risk-tolerant/most risk-averse) to three (most risk-tolerant/least risk-averse). we label this variable risk. we use linear regression techniques to relate risk tolerance to an individual’s (1) share of risk-free assets among all assets and (2) share of bonds among risky assets. the hrs data allow us to divide a person’s assets into stocks, bonds and treasury bills. we treat stocks and bonds as risky prospects with different properties, and treasury bills as risk-free assets. as mentioned in the previous section, eq. (1) describing the asset allocation rule from the capital asset pricing model is the basis for the regressions we employ in this paper. eq. (1) suggests that the proportion of all assets held in risky securities is a function of a person’s risk tolerance, his/her relative investment human capital and factors that are common to everybody in the market. common factors include the difference in expected return between risky and risk-free assets as well as variance of the risky assets. life cycle concerns may cause these common factors to be perceived differently by individuals. for example, people nearing death may not be concerned by the long-term expected return and variance of their portfolio, and may focus only on the next few years. for this reason, we specify the following regression: 162 g. hariharan et al. / financial services review 9 (2000) 159–170 tbill1 5 b0 1 b1risk1 b2educ 1 o i53 i bixi 1 e. (2) the dependent variable tbill1 is the proportion of all assets invested in risk-free securities. further details on this variable are presented following our discussion of the independent variables. means and standard deviations of all variables are shown in table 1. the right side of eq. (2) includes risk, our risk tolerance variable, and educ which crudely measures human capital accumulation as the number of years of education completed by the individual. through eq. (1) the capm predictsb1 , 0, andb2 , 0. additional independent variables x3, . . . , xi are described below. strictly speaking, eq. (1) predicts that b3, . . . , bi are all zero. we have included these variables to control for idiosyncratic beliefs and lifecycle considerations that might cause people to have different perceptions of e~rm 2 rf! sm 2 . variables focusing on the individual’s expectations are intended to identify reasons why individuals might differ in their personal assessment of the risk-return tradeoff among stocks, bonds and t-bills. exinfl contains an individual’s assessment of the probability that the table 1 variable definitions, means, and standard deviations variable name variable definition mean standard deviation exinfl perceived probability of double-digit inflation during the next ten years 5.644 2.303 exdepr perceived probability of a major depression during the next ten years 5.185 2.477 age respondent’s age 55.903 3.138 male dummy variable, equals one if respondent is male 0.662 0.473 educ years of education completed 13.367 2.386 risk index of risk tolerance with four values that range from zero (least risk-tolerant) to three (most risk-tolerant) .711 1.071 plan10 dummy variable, equals one if the planning period for spending and saving decisions is ten years or more 0.106 0.308 nopldat dummy variable, equals one if respondent did not answer the planning question 0.009 0.092 married dummy variable, equals one if respondent is married 0.829 0.376 inherit expectation of leaving a sizable inheritance to respondent’s heirs, five values ranging from one (“yes, definitely”) to five (“no, definitely”) 3.102 1.402 noinhdat dummy variable, equals one if respondent did not answer the inheritance question 0.007 0.086 networth total household net worth including financial assets and home equity 333998.1 618652.2 tbill1 proportion of financial assets invested in risk-free securities 0.424 0.452 stock1 proportion of financial assets invested in stock 0.522 0.446 bond1 proportion of financial assets invested in bonds 0.054 0.181 bond2 proportion of risky assets invested in bonds 0.093 0.242 163g. hariharan et al. / financial services review 9 (2000) 159–170 u.s. economy will experience double-digit inflation sometime during the next 10 years. while nominal interest rates are likely to be higher when average economy-wide expectations of inflation are higher, an individual whose expectations of inflation are unusually high may be inclined to avoid the fixed nominal returns of bonds. exdepr is the individual’s assessment of the chances that the u.s. economy will experience a major depression during the next ten years. presumably those that think a depression is likely will be less interested in stocks and more in fixed nominal assets. in many of the studies (for instance, riley and chow, 1992), wealth has been argued to be an important determinant of risk aversion and asset allocation. our measure of wealth is networth, which is the sum of the value of housing paid for, other real estate holdings paid for, value of automobile paid for, value of business(es) owned, amount in individual retirement accounts, amount in savings and other financial securities, and net of any debts owed. plan10 is a dummy variable identifying households who believe that the most relevant planning period for spending and saving decisions is 10 years or more. according to butler and domian (1993) and gunthorpe and levy (1994), households with long planning horizons should use greater stock allocations. since many households didn’t answer the planning question, we have also included a dummy variable called nopldat which identifies them. inherit identifies households that plan to leave an inheritance, using the following survey question, “do you [and your (husband/wife/partner)] expect to leave a sizable inheritance to your heirs? 1. yes, definitely, 2. yes, probably, 3. yes, possibly, 4. probably not, 5. no, definitely.” inherit is important because our data focuses on people shortly before retirement. people who don’t plan to leave an inheritance may have a considerably shorter planning horizon making them less likely to purchase stocks. noinhdat is a dummy variable denoting respondents who didn’t answer the inheritance question. noinhdat includes both those with response to the inherit question, who receive a value of zero for the noinhdat, and those who did not respond to it, who receive a value of 1 for noinhdat. the remaining variables in our regressions describe demographic and lifecycle differences among respondents. we have included information on the person’s age (age), a dummy for people who are married with spouse present (married), and a dummy variable identifying the person’s gender (male). since the individuals in our data set are near retirement, older people may prefer assets that are predictable in the short term (such as treasury bills and bonds). the gender and marital status dummies should pick up taste differences (other than tolerance for risk) that affect asset allocation. we now return to our discussion of the dependent variable tbill1. since our primary emphasis is on the relationship between portfolio choice and risk tolerance, we have chosen to focus on purely financial assets with values related to their risk and expected return 164 g. hariharan et al. / financial services review 9 (2000) 159–170 properties. similarly, we have ignored money kept in savings accounts because the demand for savings accounts may be linked to liquidity concerns. this leaves us with three categories of financial assets, which we have labeled tbills, bonds, and stocks. tbills actually includes money in certificates of deposit, government savings bonds and treasury bills, while bonds includes corporate, municipal, government, foreign bonds and bond funds. the stocks category includes all shares of stock in publicly held corporations, mutual funds or investment trusts. from these three categories, we created the dependent variable tbill1, as well as two other variables stock1 and bond1 for additional regressions. each variable is the ratio of the amount an individual investor holds of that asset type to an individual’s total of the three asset groups. more specifically, if ti denotes the wealth person i has invested in treasury bills while si is wealth in stock and bi is wealth in bonds, then tbill15 ti/(si 1 bi 1 ti). similarly, stock15 si/(si 1 bi 1 ti) and bond15 bi/(si 1 bi 1 ti). the dependent variable, tbill1, in eq. (2) can be replaced by either stock1 or bond1 to explore the determinants of these ratios. while the regression described in eq. (2) allows us to test the capm’s predictions concerning allocation of assets between risky and risk-free assets, we also wish to investigate the effect of risk tolerance on the composition of the risky portfolio. broadly, the capm suggests that risk tolerance will not have any effect on the percentage of risky assets allocated to stocks as opposed to bonds. the informal wall street wisdom cited in the paper’s introduction suggests that risk tolerant people should select more stock relative to bonds than their risk averse brethren. to test this, we specify a linear regression similar to eq. (2) with the fraction of risky assets devoted to bonds, bond25 bi/(si 1 bi), as the dependent variable: bond2 5 g0 1 g1risk1 g2educ 1 o i53 i gixi 1 u. (3) wall street predictsg1 , 0 while the capm predictsg1 5 0. if the determinants of high risk versus low risk allocations within the risky portfolio are similar to the determinants of risky versus risk-free assets, then the estimated eq. (3) should look very much like eq. (2). 4. results the regression results reported in table 2 are broadly consistent with the predictions of the capm that (1) increased tolerance for risk causes capm investors to alter the percentage of assets held in the risky portfolio as opposed to the risk-free asset, and (2) the composition of stocks and bonds held in the risky portfolio should not change as risk tolerance changes. regression (i) in table 2 estimates eq. (2), while regressions (ii) and (iii) replace the tbill1 dependent variable with stock1 and bond1, respectively, to show the effect of our regressors on these other assets. the coefficient for risk in regression (i) is negative and highly significant whereas it is positive and highly significant in regression (ii). this result 165g. hariharan et al. / financial services review 9 (2000) 159–170 confirms the work of schooley and worden (1996) that, as the capm predicts, risk-tolerant individuals are more likely to choose a smaller proportion of risk-free assets. the expectation variables perform broadly as anticipated. individuals expecting a major depression are significantly more inclined to invest in treasury bills; the exdepr coeffitable 2 regression results (i) tbill1 (ii) stock1 (iii) bond1 (iv) bond2 risk 20.0202822* 0.0176538* 0.0026284 0.0006723 0.0081 0.0080 0.0033 0.0053 0.012 0.028 0.431 0.899 exinfl 0.0010041 20.0018642 0.0008601 0.0019292 0.0042 0.0042 0.0018 0.0028 0.813 0.659 0.624 0.496 exdepr 0.0101191* 20.0083302* 20.0017888 20.0000649 0.0040 0.0039 0.0016 0.0026 0.011 0.035 0.275 0.980 age 0.0010009 20.0026679 0.001667 0.0044584* 0.0028 0.0027 0.0011 0.0018 0.716 0.331 0.144 0.014 male 20.0642129* 0.0718902* 20.0076773 20.0212092 0.0190 0.0189 0.0079 0.0129 0.001 ,0.001 0.329 0.100 educ 20.033231* 0.0282265* 0.0050045* 0.0022464 0.0037 0.0037 0.0015 0.0026 ,0.001 ,0.001 0.001 0.384 married 20.0225661 0.0204705 0.0020956 20.021743 0.0237 0.0236 0.0098 0.0162 0.342 0.387 0.831 0.181 plan10 20.0151819 0.019999 20.0048171 20.0051516 0.0283 0.0282 0.0117 0.0179 0.591 0.478 0.681 0.774 nopldat 0.0298024 20.0418523 0.0120499 0.0041397 0.0933 0.0929 0.0386 0.0608 0.749 0.652 0.755 0.946 inherit 0.0128339* 20.0118426 20.0009914 20.0001546 0.0066 0.0065 0.0027 0.0044 0.050 0.070 0.715 0.972 noinhdat 20.1567533 0.1056201 0.0511332 0.0328997 0.1022 0.1019 0.0423 0.0604 0.125 0.300 0.227 0.586 networth 28.50e-08* 6.32e-08* 2.17e-08* 1.85e-08* 1.47e-08 1.47e-08 6.10e-09 8.38e-09 ,0.001 ,0.001 ,0.001 0.027 constant 0.8209468* 0.2838376 20.1047844 20.1721026 0.1720 0.1714 0.0712 0.1150 ,0.001 0.098 0.141 0.135 number of observations 2577 2577 2577 1812 r-squared 0.0785 0.0601 0.0141 0.0102 each table entry contains the coefficient, standard error, and p-value for the associated regressor. the top row lists the dependent variable and a regression number. independent variables are listed in the first column. * significant at the 95% level. 166 g. hariharan et al. / financial services review 9 (2000) 159–170 cients are positive in regression (i), and negative in regression (ii). consistent with schooley and worden, individuals who were less likely to leave an inheritance (i.e., higher values of inherit) used less stock and more t-bills. individuals whose planning horizon was 10 years or more did tend to use more stock; the coefficients on the plan10 dummy variable are positive in regression (ii), negative in regression (i), although they are not statistically significant. the probability of high inflation, exinfl, is the least significant among the expectation variables. this implies that people who expected double-digit inflation did not alter their portfolio composition. this last result may merely reflect the fact that when inflationary expectations are high generally nominal interest rates rise leaving individual investors with no strong preference between stocks and bonds. demographic variables other than age are highly significant. educ is highly significant suggesting that educated people allocate more of their wealth to the risky assets, perhaps because they are better at understanding the risks associated with these assets (tversky & kahneman, 1986). married males are much less likely to invest in risk-free treasury bills than are their single-female counterparts. the gender difference is especially interesting in light of the work by bajtelsmit et al. (1999) which suggests women accumulate fewer assets in retirement due to lower risk tolerance than males. according to our regression (i), even after controlling for differences in a person’s tolerance for risk, women are more likely to invest in risk-free securities, such as treasury bills, than males. apparently there are taste and/or opportunity differences between the genders beyond risk tolerance that are important determinants of portfolio composition. our measure of wealth, networth, is highly significant in all the regressions. however, the magnitude of the coefficients on networth is very small. it is possible that much of the effect of networth is actually captured by other variables, particularly risk and educ. when networth is excluded from these regressions none of the other coefficients is changed indicating perhaps that the above statement has some validity. in addition, some of the components of networth such as the value of physical property, cars and business may be corrupting the results, because they are not truly exogenous. our somewhat short list of explanatory variables and our inability to measure accurately taste and time preference differences may explain the small values for r2 in our regressions. the focus for our study, however, is the role of differences in risk tolerance and our results do validate our hypothesis. perhaps utilizing the later waves of data and their longitudinal nature will improve the explanatory power of these regressions and is a topic for future study. regression (iv) in table 2 presents our estimates of eq. (3) and the hypothesis that the composition of stocks and bonds held in the risky portfolio should not change as risk tolerance changes. the pattern of variables determining the portion of risky assets devoted to low-risk low-return vehicles is very different from the determinants of their zero-risk counterparts. risk-tolerant investors are no less likely to choose bonds than stocks. this refutes the wall street intuition and supports the capm. in fact, the only highly significant right side variables in regression (iv) are age and networth. this suggests that casual intuition linking low-risk and no-risk securities may be inappropriate. table 3 presents a correlation matrix in order to address concerns about multicollinearity. we calculated variance inflation factors (vifs) for all regressions reported. a vif is the ratio of the actual variance of a coefficientbi to what the variance would have been if xi was 167g. hariharan et al. / financial services review 9 (2000) 159–170 t ab le 3 c or re la tio n m at rix of rig ht si de va ria bl es r is k e x in f l e x d e p r a g e m a le e d u c m a r r ie d p la n 10 n o p ld a t in h e r it n o in h d a t n e t w o r t h r is k 1. 00 00 e x in f l 0. 01 91 1. 00 00 e x d e p r 2 0. 00 59 0. 52 91 1. 00 00 a g e 2 0. 06 03 2 0. 02 10 2 0. 01 42 1. 00 00 m a le 0. 02 40 2 0. 02 29 2 0. 06 64 2 0. 02 96 1. 00 00 e d u c 0. 01 32 2 0. 05 31 2 0. 10 40 2 0. 05 99 0. 09 53 1. 00 00 m a r r ie d 2 0. 03 06 2 0. 05 74 2 0. 09 01 2 0. 00 23 0. 26 22 2 0. 04 92 1. 00 00 p la n 10 0. 01 23 2 0. 01 89 2 0. 03 18 2 0. 01 08 0. 02 59 0. 06 29 0. 01 62 1. 00 00 n o p ld a t 2 0. 01 51 0. 03 08 0. 01 89 0. 03 78 2 0. 01 07 2 0. 06 10 2 0. 03 75 2 0. 04 12 1. 00 00 in h e r it 0. 00 55 0. 04 99 0. 05 71 0. 00 09 2 0. 13 38 2 0. 13 27 2 0. 11 99 2 0. 06 43 0. 02 29 1. 00 00 n o in h d a t 2 0. 00 12 0. 00 11 0. 00 46 0. 01 38 0. 05 042 0. 02 32 0. 03 18 0. 01 55 0. 02 03 2 0. 31 35 1. 00 00 n e t w o r t h 0. 00 28 2 0. 05 17 2 0. 07 04 0. 03 23 0. 11 36 0. 19 20 0. 13 22 0. 08 362 0. 00 97 2 0. 24 33 2 0. 02 23 1. 00 0 168 g. hariharan et al. / financial services review 9 (2000) 159–170 uncorrelated with the other x’s (judge, hill, griffiths, lutkepohl & lee, 1988). a value of 5.0 is quite often used as an indication of severe multicollinearity (marquardt & snee, 1975). in our case, the vifs ranged from a minimum 1.01 to a maximum of 1.41, indicating that multicollinearity among the right-side variables in table 2 is not a significant problem. 5. conclusions and future research our regression evidence suggests that risk-tolerant individuals invest lesser amounts in treasury bills. to the extent that t-bills are a reasonable approximation to the capital asset pricing model’s risk-free asset, this tends to confirm the capm prediction that risk-tolerant investors will hold a smaller fraction of their investments in the risk-free asset. in addition, we found that the division of individual portfolios between stocks and bonds was not systematically related to our measure of risk tolerance. this is broadly consistent with the capm notion that risk tolerance plays no role in the composition of the risky portfolio. several authors, including bajtelsmit et al. (1999), reichenstein (1999), and yuh, hanna and montalto (1999), have noted that attitude toward risk is an important determinant of asset accumulation for retirement. increasingly, individuals are able to allocate their pension assets between stocks and bonds or are making all of the decisions on their own. if many individuals differ significantly in risk tolerance from the fund managers that traditionally performed this task, retirement preparedness will become much more variable in years to come. our data confirm the work of bajtelsmit et al. (1999), for example, that among people nearing retirement risk tolerance is greater for men than for women. this suggests that women will likely accumulate substantially less wealth due to a taste for low-risk, low-return assets. research linking measures of risk-tolerance available in thehealth and retirement surveyand survey of consumer financesto more generally available data such as race, gender and occupation may help predict future preparedness for retirement. acknowledgments we are grateful for helpful comments from robert hagerman, joseph ogden, david weil, two anonymous referees, and the editor, karen eilers lahey. we also acknowledge summer research support for govind hariharan from the school of management of the university at buffalo. references bellante, d., & saba, r. (1986). human capital and life-cycle effects on risk aversion.journal of financial research, 9,41–51. bajtelsmit, v., bernasek, a., & jianakoplos, n. (1999). gender differences in defined contribution pension decisions.financial services review, 8,1–10. brinson, g., hood, l. r., & beebower, g. (1986). determinants of portfolio performance.financial analysts journal, 42,39–44. 169g. hariharan et al. / financial services review 9 (2000) 159–170 butler, k., & domian, d. (1993). long-run returns on stock and bond portfolios: implications for retirement planning.financial services review, 2,41–50. canner, n., mankiw, n. g., & weil, d. (1997). an asset allocation puzzle.american economic review, 87, 181–191. friend, i., & blume, m. (1975). the demand for risky assets.american economic review, 65,900–922. gunthorpe, d., & levy, h. (1994). portfolio composition and the investment horizon.financial analysts journal, 50, 51–56. ho, k., milevsky, m., & robinson, c. (1994). asset allocation, life expectancy, and shortfall.financial services review, 3,109–126. judge, g., hill, r., griffiths, w., lutkepohl, h., & lee, t. (1988).introduction to the theory and practice of econometrics,second edition. new york: john wiley & sons. lebaron, d., farrelly, g., & guha, s. (1989). facilitating a dialogue on risk: a questionnaire approach.financial analysts journal, 45,19–24. marquardt, d., & snee, r. (1975). ridge regressions in practice.the american statistician, 29,3–19. morin, r., & suarez, f. (1983). risk aversion revisited.journal of finance, 38,1201–1216. reichenstein, w. (1999). calculating a family’s asset mix.financial services review, 7,195–206. riley, w. jr., & chow, k. (1992). asset allocation and individual risk aversion.financial analysts journal, 48, 32–37. schooley, d., & worden, d. (1996). risk aversion measures: comparing attitudes and asset allocation.financial services review, 5,87–99. siegel, f., & hoban, j., jr. (1982). relative risk aversion revisited.review of economics and statistics, 64, 481–487. siegel, f., & hoban, j., jr. (1991). measuring risk aversion: allocation, leverage, and accumulation.journal of financial research, 14,27–35. tversky, a., & kahneman, d. (1986). rational choice and the framing of decisions.journal of business, 59, 251–278. viscusi, w. k. (1992).fatal tradeoffs: public and private responsibilities for risk.new york: oxford university press. yuh, y., hanna, s., & montalto, c. (1999). mean and pessimistic projections of retirement adequacy.financial services review, 7,175–193. 170 g. hariharan et al. / financial services review 9 (2000) 159–170 pii: s1057-0810(01)00073-7 financial services and the african-american market: what every financial planner should know d. anthony platha,*, thomas h. stevensonb adepartment of finance and business law, university of north carolina at charlotte, charlotte, north carolina 28223 bdepartment of marketing, university of north carolina at charlotte, charlotte, north carolina 28223 received 6 october 1999; received in revised form 12 april 2001; accepted 17 april 2001 abstract african-american consumers differ markedly from their caucasian counterparts in terms of financial product preferences, product research, and investment asset portfolio composition. this study examines some of the principal differences between african-american and caucasian households in evaluating and purchasing investment assets and explores differences in asset holdings between the two racial groups. this information can help financial planners seeking to market to the african-american community better understand this community, tailor investment information for the unique needs of this community, and render more effective service to individuals and families that comprise this attractive and growing market segment. © 2001 elsevier science inc. all rights reserved. jel classification: d12; d31; j15; j70 keywords: individual investing; racial difference; portfolio composition; personal finance; investment portfolio the importance of segmentation in the development of an effective marketing strategy is well established in the business literature. there are many bases used for market segmentation, but demographic factors such as age, occupation, income level, educational attainment, and race are frequently used to identify key markets. race has become increasingly important as a segmentation variable because it has been shown that race can influence * corresponding author. tel.: �1-704-687-4413; fax: �1-704-687-6987. e-mail address: daplath@email.uncc.edu (d.a. plath). financial services review 94 (2000) 343–359 1057-0810/00/$ – see front matter © 2001 elsevier science inc. all rights reserved. pii: s1057-0810(01)00073-7 consumption patterns and because racial minorities have come to represent a larger and larger proportion of the u.s. population. the sheer numbers of the largest minority group, african-americans, and the improving economic status of many of its members, make this racial minority especially attractive. nevertheless, relatively few studies in the finance literature have profiled differences in financial asset portfolio holdings of black versus white households, or examined how these differences should affect the manner in which financial planners tailor their marketing efforts to meet the needs of different racial groups. the purpose of this study is twofold: first, we seek to profile racial differences in investment asset ownership patterns between black and white households, evaluating the extent to which this profile is consistent with information reported in other academic work that examines racial differences in wealth accumulation patterns and risk tolerance levels. second, we seek to use the profile of financial asset portfolio holdings reported here to advise financial planners and others in the financial community how to structure their product offerings to more effectively meet the needs of african-american consumers. the balance of this study provides information and insights regarding the financial holdings and investment asset consumption preferences of black and white households. section i profiles the growth of african-american households in recent years and surveys the literature to ascertain extant knowledge about differences in black versus white financial asset portfolio holdings. section ii introduces the scf dataset that represents the source of the statistical information reported in the article, and explains the statistical tests used to compare black versus white households. section iii presents some of the principal differences in asset holding patterns across black and white consumers, controlling for income, age, and educational attainment in presenting household financial information across different racial groups. section iv uses the information introduced in section iii to develop a profile of african-american households’ wealth-building asset portfolios for financial planners seeking to understand and serve the needs of the african-american market segment. finally, section v provides conclusions regarding the differences between black and white financial consumers and summarizes how these differences influence asset holdings patterns across the two racial groups. 1. wealth accumulation and the african-american consumer the number of african-americans in the u.s. has grown rapidly. in the ten-year period from 1985 to 1995 for example, there was an increase of more than 4.5 million africanamericans in the u.s., a nearly 16 percent increase. the corresponding percentage increase for caucasians in the same period was just under 8 percent. during the same ten-year period, the number of black households grew by 23 percent; white households grew by only 11 percent. moreover, the total u.s. population of blacks is predicted to increase by as many as 17 million by the year 2020. this represents a projected 53 percent increase over the 25 years from 1995; the white population is forecast to grow only 28 percent during the corresponding period (u.s. department of commerce, 1998). in spite of this rapid growth, very little academic research examines how financial asset 344 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 consumption preferences differ between african-american and caucasian households, and most prior research examines investment portfolio differences between black and white households only as a secondary issue. in general, past studies focus on (1) differences in wealth accumulation rates between black and white households, invoking asset portfolio differences only to explain divergent rates of wealth accumulation, or (2) differences in risk tolerance across the races, using portfolio composition differences to illustrate differences in exposure to liquidity and default risk across different racial groups. while these studies touch on investment portfolio composition differences across different racial groups, portfolio differences are seldom the focal point of the research effort and they frequently reference a limited spectrum of investment assets in only two or three financial categories. in spite of these limitations, the wealth accumulation literature does provide a useful starting point to characterize differences in the investment portfolio holdings of black and white financial consumers. for example, it is widely recognized that the rate of wealth accumulation across black households is significantly below the rate at which white households build wealth (blau & graham, 1990; wolff, 1994; myers & chung, 1996). wealth accumulation rates are positively related to household income levels (zhong & xiao, 1995; gutter et al., 1999), positively related to educational achievement (zhong & xiao, 1995; gutter et al., 1999), and positively related to consumer age (blau & graham, 1990; zhong & xiao, 1995), yet this does not lead to the corollary conclusion that the wealth gap between black and white households narrows with advancing income, educational attainment, and age across different racial groups. in contrast, evidence suggests that the wealth gap widens with increasing consumer age (wolff, 1994), while the literature is silent about how the wealth gap changes with advancing income and educational attainment levels across different racial groups. concerning investment portfolio composition differences across different racial groups, boyce (1998), gutter et al. (1999) and badu et al. (1999) report that black investors display little preference for risky financial securities such as common stocks, while white investors select these riskier financial investments with greater frequency. this leads to the conclusion that black consumers should display relatively smaller holdings of risky financial securities, such as municipal and corporate bonds, common stocks, equity mutual funds, and brokerage accounts than white consumers; and relatively larger holdings of low-risk financial securities such as bank savings and time deposits, savings bonds, and treasury bonds in their investment portfolios. in addition, blau & graham (1990) and brimmer (1991) point out that black consumers display a preference for holding tangible, nonfinancial assets that yield consumptive services, while white consumers tend to hold financial assets and income-producing nonfinancial assets. given this evidence, black consumers should display relatively larger investments in consumption-oriented real property, such as a personal residence and vacation property, and a relatively smaller investment in income-producing real property, such as rental property and commercial real estate, in their investment portfolios. consistent with their demonstrated preference for lower risk, brimmer (1991) and badu et al. (1999) also report that black consumers show a greater preference for financial assets with a high degree of liquidity; while white investors will more willingly sacrifice financial asset liquidity in order to earn higher risk-adjusted total investment returns. based on this 345d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 finding, african-american consumers should show relatively larger holdings of highly liquid, short-term instruments such as bank accounts, savings bonds, and bond-based mutual funds; and correspondingly smaller holdings of less liquid, longer-term assets such as corporate bonds, common stocks, and brokerage accounts, in their financial asset portfolios. regardless of racial background, zhong and xiao (1995) and gutter et al. (1999) report that investment risk tolerance increases with increasing income levels, leading to the conclusion that risk-related investment portfolio differences between african-american and caucasian consumers should diminish with increasing income. similar to this finding, zhong and xiao (1995) and gutter et al. (1999) report that investment risk tolerance increases with heightened levels of educational attainment, suggesting that risk-related investment portfolio differences between black and white consumers should also diminish with rising educational attainment. in contrast to these results, the literature offers less consensus on the issue of how investment risk tolerance changes with advancing investor age. morin and suarez (1983) find that investment risk tolerance decreases with increasing age, while zhong and ziao (1995) report that investment risk tolerance increases with increasing age, and gutter et al. (1999) suggest that investment risk tolerance first increases with increasing age, reaches a maximum in middle age, and then declines with further increases in age. in spite of the lack of agreement across these studies, virtually all past research indicates that there is a significant relationship between investment risk tolerance and age, leading to the conclusion that investment portfolio composition across both african-american and caucasian households should differ when age is introduced as a control variable in the analysis. virtually all of the research results discussing investment portfolio composition differences across different racial groups may be considered somewhat tentative, because investment portfolio preference differences usually do not represent the primary focal point of the research efforts cited here, and these studies offer their respective research conclusions on the basis of very few categorical investment choices. the remainder of this study seeks to build on past research conclusions, offering an expanded array of investment alternatives to profile portfolio preference differences across black and white households, and reviewing whether the investment portfolio implications discussed above hold true under the wider array of investment alternatives presented below. while the results reported below highlight statistically significant pairwise differences between black and white investment portfolios, these isolated pairwise differences may fail to capture the full range of investment portfolio differences between black and white consumers. a large body of prior research has shown that the rate of wealth accumulation across african-american households significantly lags that of white households, so statistically significant pairwise differences may not capture many of the substantive differences in black versus white portfolio holdings. the magnitude of investment balances in black households almost always lags white investment balances because white consumers possess greater financial wealth. consequently, this research effort also examines the changing relative magnitude of investment portfolio differences across black and white households, noting whether these differences are increasing, decreasing, or remaining the same, as variables known to influence investment portfolio preferences—such as age, income, and educational attainment level—change. 346 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 2. data and methodology this study reports and evaluates data gathered from the 1998 survey of consumer finances (scf) prepared by the board of governors of the federal reserve system in cooperation with the statistics and income division of the internal revenue service. conducted triannually since 1983, the scf provides detailed information on the financial characteristics of u.s. households, including financial asset and liability holding patterns, real estate ownership, and household net worth. also included is a variety of demographic and attitudinal characteristics covering age, sex, race, educational attainment, income, and other classificatory variables useful for characterizing household balance sheet characteristics across different subgroups within the american population. a more complete description of the scf dataset is given by kennickell et al. (2000). the scf dataset uses a dual-frame sampling plan that incorporates both an area-probability sample and a special list sample derived from irs tax records. the area-probability sample provides information on financial variables that are widely distributed in the general u.s. population, such as automobile ownership and home mortgages. the list sample represents an oversample of relatively wealthy families designed to capture financial data items that are highly concentrated within a relatively small proportion of the population, such as commercial real estate holdings and household trust fund ownership. this unique sampling methodology results in the oversampling of households more likely to be wealthy, which requires that descriptive statistical measures derived from the scf sample be weighted to generate sampling estimates that are projectible to the entire u.s. population (board of governors of the federal reserve system, 2000, p. 27). the descriptive statistics reported below are derived from the full, weighted version of the scf sample. the scf handles missing data attributable to item nonresponse using a multiple imputation procedure known as repeated imputation inference (rii). as montalto and sung (1996) discuss, this procedure uses stochastic multivariate methods to replace each missing value with five different imputed values to approximate the sampling distribution of missing values, so that imputed values can be averaged to produce a more accurate estimate of what a given missing value would have been in the absence of item nonresponse. the statistics reported below rely upon the full public version of the 1998 scf dataset, surveying 4,305 households to produce a total of 21,525 potential observations (board of governors of the federal reserve system, 2000, p. 27). using all five implicates in developing descriptive statistical measures introduces imputation error into the research design, however, and this additional source of variability requires the use of a specialized adjustment procedure to calculate accurate standard error terms for each of the statistics reported below. the adjustment procedure used below permits the development of standard error terms that incorporate both sampling error and imputation error for mean and proportion statistics, following the recommend course of action reported in the 1998 codebook for the survey of consumer finances (board of governors of the federal reserve system, 2000, pp. 28–32). in all cases, the statistically significant mean differences and proportion differences reported below have been obtained using a test statistic suitable when target population variances are assumed to be both unequal and unknown. 347d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 3. results: asset holding patterns across black and white consumers the portfolio of wealth-accumulation assets, including both financial and real property assets, varies markedly between black and white households. tables 1 through 3 present proportionate holdings of various financial asset categories, as well as household real estate holdings covering respondents’ primary residence, other vacation property owned, and other nonvacation property. in addition, these tables provide mean dollar values of financial accounts and real property holdings across the full scf dataset. in all cases, the results report mean financial values rather than the median percentile values, because the public version of the scf dataset is adjusted for outliers and other plausible errors in data reporting and coding before it is released to the public, and the mean is better able to convey the wide dispersion in the reported data for some response groups (board of governors of the federal reserve system, 2000, pp. 7–8). in this case, understanding the wide range of item responses across particular groups is important in helping to characterize differences between black and white investment portfolios. in order to control for differences in asset holdings that can be explained by demographic characteristics such as income, age, and education, the reported data stratify household asset holdings by respondent income (table 1), respondent age (table 2), and the highest educational grade-level attained by respondents (table 3). in each of these cases, advances in income, age, and educational attainment do lead to changes in asset holdings consistent with the results of yuh and hanna (1997). while asset portfolios of black and white consumers do become increasingly similar with advances in income and educational attainment, particularly across bank-related financial asset categories, substantial differences across both financial and real property assets held by black and white households persist even in the highest income and education categories. this finding regarding asset holdings tends to support the conclusion of williams and qualls (1989), who note that as african-american consumers make more money and move up in class standing, they do not lose their ethnic orientation and begin to resemble white consumers. the most notable difference between black and white households, however, is unrelated to cultural considerations. the dollar value of virtually all asset holdings is substantially greater across white respondents. this gap reflects the wide net worth disparity between black and white consumers reported by scott (1998), myers and chung (1996), and a host of other researchers. moreover, the gap does not appear to be related to the relative popularity of various asset categories, such as common stock versus real estate, across different racial groups. as scott (1998), lach (1999), and badu et al. (1999) suggest, african-american households are particularly conservative in their investment style, preferring real estate assets and insurance products to stock and bond investments. even within these relatively more popular investment categories, however, the mean values for all categories of real property investments across the african-american sample lie well below their corresponding values in white households. this trend persists across all income, age, and education levels; and unlike other asset holding patterns, the real estate gap grows wider, not more narrow, as income and educational attainment increase. in most cases, this is attributable to the increased valuation dispersion observed across real estate holdings of white households. the wide range of real estate values is particularly evident in the 348 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 table 1 family asset holdings by income and race characteristics asset category household income in 1998 dollars $10,000 to $24,999 $25,000 to $49,999 $50,000 to $100,000 more than $100,000 race race race race black n � 573 white n � 2,795 black n � 452 white n � 3,781 black n � 286 white n � 3,915 black n � 117 white n � 5,770 proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean financial asset ownership transaction accounts 57.8%*** $ 361*** 88.8% $ 2,901 88.0% $ 4,285 95.7% $ 3,539 96.3% $ 4,220 99.4% $ 5,514 99.4% $ 9,282*** 99.7% $ 32,782 certificates of deposit 4.8%*** $ 374*** 21.8% $ 7,089 13.1% $ 728*** 17.7% $ 6,293 7.5%** $ 278*** 17.7% $ 6,241 18.6% $ 2,776* 16.0% $ 14,349 savings deposits 45.5% $ 1,417 48.1% $ 2,907 64.2% 42,401 60.7% 43,689 76.8% 43,858* 72.3% $ 6,521 64.8% $ 4,418** 60.3% $ 14,035 savings bonds 6.4% $ 212 12.3% $ 388 14.2%** $ 238*** 22.5% $ 902 17.0%** $ 300*** 32.2% $ 1,508 4.1%*** $ 1,526 32.5% $ 3,065 bonds: treasury bonds 0.0% $ 0 0.2% $ 50 0.0%** $ 0** 0.9% $ 632 0.0%*** $ 0* 1.2% $ 273 0.4%*** $ 1,463** 5.4% $ 15,430 municipal bonds 0.0%* $ 0 1.1% $ 611 0.0%** $ 0 1.5% $ 753 0.0%*** $ 0 1.7% $ 1,852 10.0% $ 4,619* 8.4% $ 33,583 corporate bonds 0.0% $ 0 0.8% $ 321 0.0%** $ 0* 1.2% $ 181 0.0%*** $ 0* 1.5% $ 941 0.0%*** $ 0** 2.5% $ 9,615 mortgage-backed bonds 0.0% $ 0 0.1% 4127 0.0%* $ 0 0.7% $ 434 0.0%* $ 0 0.7% $ 398 0.0%** $ 0* 1.8% $ 4,459 stocks 0.9%*** $ 175* 9.0% $ 4,503 11.8% $ 1,160*** 19.3% $ 9,009 14.2%*** $ 6,893** 29.3% $ 22,259 61.6% $ 54,735*** 56.3% $282,400 retirement accounts (ira/keogh) 6.1%*** $ 682*** 18.2% 44,900 26.3% $ 3,817* 29.1% $ 7,985 37.1% $ 6,319*** 44.6% $ 18,160 37.9% $ 9,665*** 67.8% $103,649 mutual funds stock funds 1.3%*** $ 46*** 8.1% $ 2,395 12.7% $ 1,139** 12.3% $ 3,824 14.4%* $ 2,111*** 23.7% $ 13,899 29.9% $ 20,619*** 44.0% $ 81,554 government bond funds 0.0%* $ 0* 0.8% $ 158 3.0% $ 87 1.0% $ 223 0.0%*** $ 0* 2.6% $ 914 0.0%*** $ 0*** 5.0% $ 4,126 municipal bond funds 0.0%** $ 0 2.2% $ 1.378 2.9% $ 330 2.3% $ 748 1.5%* $ 2,316 5.9% $ 1,990 0.6%*** $ 30*** 12.0% $ 17,760 corporate bond funds 0.0%* $ 0 1.2% $ 577777 0.0%** $ 0 1.2% $ 359 2.4% $ 29* 3.4% $ 2,490 0.0%*** $ 0* 6.7% $ 9,046 combination funds 0.0% $ 0*** 1.9% $ 761 2.4% $ 320 2.0% $ 645 5.9% $ 3,017 2.7% $ 1,389 6.2% $ 1,994 7.4% $ 6,706 life insurance 65.8%* $24,199 59.9% $ 20,511 83.4% $65,446 73.8% $ 55,636 90.8% $124,238 86.3% $ 117,744 90.7% $413,460 89.5% $315,783 brokerage accounts 0.9%*** $ 0 8.4% $ 0 5.8%* $ 0 13.3% $ 0 11.2%** $ 0 23.3% $ 0 47.9% $ 0 54.9% $ 0 trust accounts and annuities 0.7%*** $ 820 6.1% $ 1,530 3.2% $ 2,056 5.5% $ 1,442 4.5% $ 482** 8.2% $ 3,691 7.8% $ 4,169 13.6% $ 14,879 real property ownership personal residence 35.8%** $60,735 57.9% $ 88,511 59.7% $92,515 71.3% $109,857 66.8% $115,339 87.8% $ 151,004 80.3% $238,449 93.7% $337,352 indebtedness on personal residence 19.7% $31,211 21.2% $ 45,070 45.2% $67,786 44.2% $ 52,596 63.8% $ 66,770 71.5% $ 80,210 60.5% $164,376 70.7% $164,959 other vacation property 8.4% $ 7,327 11.5% $ 12,767 12.6% $11,171 17.4% $ 12,167 32.8% $ 13,655 25.9% $ 13,378 36.5% $ 18,509 46.4% $ 19,826 indebtedness on other vacation property 2.0% $37,919 1.7% $ 182,039 3.7% $20,499* 5.2% $ 52,618 13.7% $ 56,295 9.3% $ 93,890 15.6% $128,694 21.4% $223,360 other nonvacation property 0.0% $ 0 0.2% $2,491,560 0.0% $ 0 0.4% $ 63,302 1.2% $ 25,000* 2.1% $ 13,856 2.2% $ 25,000*** 4.4% $640,681 indebtedness on other property 0.0% $ 0 0.1% $ 718,188 0.0% $ 0 0.2% $ 37,030 0.0%** $ 0 1.0% 4136,181 2.2% $ 23,000** 1.1% $720,251 note: *** denotes significance at the 0.001 level; ** denotes significance at the 0.01 level; * denotes significance at the 0.05 level. 349 d .a . p lath, t .h . stevenson /f inancial services r eview 94 (2000) 343–359 table 2 family asset holdings by age and race characteristics asset category head of household age 35 to 44 45 to 54 55 to 64 65 to 74 race race race race black n � 528 white n � 3698 black n � 379 white n � 3,948 black n � 235 white n � 2,976 black n � 190 white n � 2,346 proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean financial asset ownership transaction accounts 65.0%*** $ 981*** 91.9% $ 4,648 69.0%** $ 4,460 95.0% $ 7,461 75.8% $ 4,279* 94.5% $ 11,657 61.2%*** $ 2,064*** 97.0% $ 9,607 certificates of deposit 8.0% $ 470*** 9.9% $ 1,177 4.1%*** $ 214*** 13.7% $ 5,416 3.4%*** $ 309*** 21.4% $ 9,163 13.4%*** $ 1,280*** 33.1% $ 14,845 savings deposits 59.0% $ 1,384*** 65.6% $ 3,839 44.0%*** $ 1,400*** 65.7% $ 5,815 58.8% $ 2,287*** 56.3% $ 6,299 37.3% $ 2,299** 51.3% $ 8,099 savings bonds 12.9%*** $ 132*** 29.0% $ 858 6.7%*** $ 208*** 24.9% $ 1,249 9.1%** $ 135*** 20.3% $ 1,681 6.1%** $ 423* 18.0% $ 1,105 bonds: treasury bonds 0.0%* $ 0 1.0% $ 957 0.0%** $ 88* 1.2% $ 1,832 0.0%** $ 0 1.4% $ 5,708 0.0%** $ 0*** 2.4% $ 1,707 municipal bonds 0.0%* $ 0** 0.6% $ 1,028 0.0%*** $ 346 1.8% $ 3,906 0.0%*** $ 0*** 2.8% $ 7,669 0.0%*** $ 0** 5.7% $ 10,326 corporate bonds 0.0%* $ 0 0.6% $ 400 0.0%** $ 0** 1.3% $ 1,913 0.0%** $ 0 0.9% $ 834 0.0%* $ 0** 2.0% $ 2,169 mortgage-backed bonds 0.0%* $ 0* 0.2% $ 186 0.0% $ 0 0.5% $ 280 0.0%* $ 0 0.9% $ 1,398 0.0%** $ 0** 2.0% $ 1,252 stocks 8.0%*** $ 1,404*** 21.8% $ 20,048 9.4%*** $ 8,655*** 25.6% $ 33,985 3.5%*** $ 3,751*** 29.2% $ 83,278 0.2%*** $ 46*** 24.1% $ 71,196 retirement accounts (ira/keogh) 24.4% $ 3,309** 30.4% $ 7,718 12.0%*** $ 2,523*** 39.7% $ 18,409 29.3%* $ 5,482*** 46.3% $ 50,486 9.0%*** $ 4,182*** 48.2% $ 38,114 mutual funds stock funds 12.1% $ 2,179*** 15.7% $ 10,888 4.1%*** $ 2,115*** 24.3% $ 19,364 0.1%*** $ 25*** 16.4% $ 22,683 5.5%** $ 969*** 15.9% $ 23,618 government bond funds 1.2% $ 74 1.1% $ 341 0.0%*** $ 0** 2.3% $ 657 0.0%*** $ 0* 2.1% $ 1,423 0.0%** $ 0** 3.3% $ 2,047 municipal bond funds 1.8% $ 1,561 2.4% $ 909 0.0%*** $ 0** 4.2% $ 2,719 0.1%*** $ 5** 4.7% $ 4,589 0.0%*** $ 0*** 7.3% $ 7,115 corporate bond funds 1.3% $ 16 2.2% $ 1,135 0.0%*** $ 0 3.6% $ 3,435 0.0%** $ 0* 1.7% $ 1,084 0.0%** $ 0* 3.2% $ 4,810 combination funds 2.5% $ 698 2.4% $ 1,155 0.0%*** $ 0** 2.9% $ 1,536 0.0%*** $ 0* 2.4% $ 1,609 4.6% $ 3,650 4.2% $ 2,583 life insurance 67.5% 74,388*** 74.7% $128,183 65.6% $ 38,767*** 75.7% $128,395 81.8% $ 32,420** 79.6% $ 66,342 77.2% $ 27,004 75.8% $ 19,893 brokerage accounts 9.4%*** $ 0 17.0% $ 0 4.8%*** $ 0 23.1% $ 0 3.4%*** $ 0 19.3% $ 0 0.2%*** $ 0 23.8% $ 0 trust accounts and annuities 1.5% $ 1,467 3.7% $ 2,068 4.7% $ 1,108 6.8% $ 3,221 0.1%*** $ 0** 6.4% $ 6,888 0.4%*** $ 44*** 15.0% $ 6,373 real property ownership personal residence 51.6%*** $87,323*** 72.1% $144,338 43.7%*** $101,891** 79.4% $170,754 61.6%* $109,084* 84.2% $167,366 61.6%* $ 62,322*** 85.0% $150,524 indebtedness on personal residence 41.9%*** $69,532 63.8% $ 88,423 30.6%*** $ 52,678** 63.5% $ 88,766 39.1% $ 67,528 50.1% $ 72,353 20.3% $ 40,532 24.3% $ 55,655 other vacation property 13.5% $ 9,324 18.8% $ 13,320 12.7%** $ 12,848 26.4% $ 15,563 24.8% $ 16,244 24.4% $ 16,619 13.3%*** $ 7,067* 29.3% $ 14,585 indebtedness on other vacation property 4.5%* $25,904*** 9.1% $106,916 2.3%*** $ 72,180 9.8% $115,733 15.5% $ 59,558* 7.8% $215,155 0.6%*** $329,042 5.9% $155,829 other nonvacation property 0.0%*** $ 0* 0.7% $190,024 0.3%** $ 25,000* 1.9% $197,567 1.4% $ 25,000* 1.9% $974,283 0.0%*** $ 0* 1.8% $394,334 indebtedness on other property 0.0%* $ 0 0.5% $115,106 0.3% $ 23,000 0.&% $121,921 0.0%* $ 0* 0.6% $695,094 0.0% $ 0 0.6% $343,148 note: *** denotes significance at the 0.001 level; ** denotes significance at the 0.01 level; * denotes significance at the 0.05 level. 350 d .a . p lath, t .h . stevenson /f inancial services r eview 94 (2000) 343–359 table 3 family asset holdings by educational attainment and race characteristics asset category head of household highest educational attainment completed 9th through 11th grade high school graduate colege graduate attended or completed graduate school race race race race black n � 393 white n � 1,252tc;;2black n � 669 white n � 4,085 black n � 161 white n � 3,800 black n � 127 white n � 3,756 proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean proportion mean financial asset ownership transaction accounts 29.8%*** $ 226*** 77.1% $ 2,633 65.7%*** $ 998*** 90.5% $ 3,446 84.7% $ 2,848** 98.4% $ 6,977 96.5% $ 2,297*** 99.4% $ 10,680 certificates of deposit 4.0%*** $ 186*** 15.6% $ 3,970 6.9%*** $ 394*** 19.1% $ 7,030 3.4%*** $ 50** 15.8% $ 7,522 8.4%*** $ 464*** 25.6% $ 10,763 savings deposits 30.3%* $ 319*** 42.8% $ 2,131 48.7% $ 1,592*** 58.7% $ 4,188 40.5% $ 834*** 44.8% $ 3,213 59.5% $ 1,591*** 43.1% $ 5,308 savings bonds 3.3%*** $ 21 11.8% $ 604 8.6%*** $ 131*** 20.8% $ 825 21.1% $ 141*** 33.3% $ 1,407 33.0% $ 1,001** 32.0% $ 3,193 bonds: treasury bonds 0.0% $ 0 0.1% $ 154 0.0% $ 0 0.2% $ 280 0.0%*** $ 0*** 2.0% $ 2,220 0.0%*** $ 0** 3.4% $ 6,261 municipal bonds 0.0% $ 0 0.9% $ 1,280 0.0%** $ 0 0.6% $ 1,428 0.0%*** $ 0*** 2.9% $ 7,847 0.0%*** $ 0*** 8.5% $ 14,574 corporate bonds 0.0%* $ 0 0.8% $ 944 0.0%* $ 0 0.7% $ 1,521 0.0%** $ 0* 0.8% $ 1,646 0.0%*** $ 0 2.8% $ 5,839 mortgage-backed bonds 0.0% $ 0 0.1% $ 74 0.0% $ 0 0.2% $ 757 0.0%* $ 0 0.6% $ 603 0.0%** $ 0* 1.0% $ 1,102 stocks 1.3%*** $ 3** 9.0% $ 4,348 4.7%*** $ 1,086*** 15.5% $ 18,783 14.3%** $ 2,417*** 29.0) $ 28,720 10.2%*** $ 446*** 36.2% $ 56,709 retirement accounts (ira/keogh) 3.4%*** $ 361*** 16.8% $ 2,640 7.4%*** $ 1,036*** 25.1% $ 7,862 7.3%*** $ 1,649*** 43.5% $ 18,495 19.0%*** $ 4,510*** 55.9% $ 32,526 mutual funds stock funds 1.1%* $ 33** 4.8% $ 1,268 6.8% $ 368*** 9.7% $ 6,315 6.3%*** $ 461*** 20.7% $ 12,010 10.1%** $ 1,430*** 27.2% $ 23,636 government bond funds 0.0% $ 0 0.3% $ 327 0.0%*** $ 0 0.9% $ 98 0.0%*** $ 0*** 4.4% $ 1,108 0.1%*** $ 304 5.7% $ 2,643 municipal bond funds 0.0%* $ 0* 1.1% $ 621 1.4% $ 24*** 2.7% $ 1,938 0.0%*** $ 0*** 6.5% $ 4,002 5.8% $ 377** 11.1% $ 7,060 corporate bond funds 0.0% $ 0 0.3% $ 431 0.0%** $ 0 1.3% $ 696 0.0%*** $ 0** 2.2% $ 926 p*** $ 0 4.9% $ 4,911 combination funds 0.0% $ 0 0.7% $ 407 2.8% $ 1,158 1.8% $ 989 0.0%*** $ 0* 4.8% $ 3,518 0.0%*** $ 0*** 6.9% $ 4,388 life insurance 60.3% $ 9,454* 62.0* $34,364 69.2% $41,252 70.1% $ 44,693 84.8% $ 88,807 81.1% $104,042 81.1% $ 92,785* 82.8% $149,202 brokerage accounts 0.0%*** $ 0 6.0% $ 0 3.1%*** $ 0 11.4% $ 0 9.5%*** $ 0 26.0% $ 0 7.9%*** $ 0 35.0% $ 0 trust accounts and annuities ‘0.0%** $ 0* 2.6% $ 679 0.7%*** $ 204* 6.1% $ 1,544 0.3%*** $ 16,314 7.1% $ 6,999 5.2% $ 3*** 9.7% $ 17,843 real property ownership personal residence 26.2%*** $59,357 66.7% $80,816 45.7%*** $75,069** 72.6% $112,786 48.0%* $118,946 71.6% $155,272 67.2% $141,450* 74.5% $207,416 indebtedness on personal residence 12.8%*** $10,540*** 31.8% $40,369 27.5%** $52,189 42.6% $ 56,862 34.1%* $ 71,897 53.8% $ 81,999 42.0% $102,560 54.7% $100,679 other vacation property 3.6%** $12,656 11.7% $14,313 9.3%** $12,296 17.6% $ 12,653 11.6% $ 11,643 22.3% $ 14,216 31.5% $ 12,316 30.8% $ 14,843 indebtedness on other vacation property 0.9% $27,352 2.2% $40,725 3.6% $23,947* 5.2% $ 62,238 9.8% $ 23,699** 6.5% $197,702 21.3% $ 56,443 13.6% $588,194 other nonvacation property 0.0% $ 0 0.5% $86,706 0.5% $25,000 0.6% $287,331 0.1%*** $ 601*** 1.7% $419,850 2.4% $244,659 1.7% $389,778 indebtedness on other property 0.0% $ 0 0.4% $35,546 0.0%* $ 0* 0.3% $ 70,054 0.1%* $ 272** 0.&% $336,592 0.5% $ 50,000 0.6% $409,805 note: *** denotes significance at the 0.001 level; ** denotes significance at the 0.01 level; * denotes significance at the 0.05 level. 351 d .a . p lath, t .h . stevenson /f inancial services r eview 94 (2000) 343–359 nonvacation real estate category, where commercial real estate assets are particularly concentrated across a white ownership group. interestingly, black households do appear to display a preference for consumptive-type real estate, a finding that supports the work of terrell (1971), sobol (1979), blau and graham (1990) and brimmer (1991). black consumers show a far greater relative investment in consumption-oriented real property, such as a personal residence or vacation property, and a smaller corresponding investment in income-producing business property, captured in the other nonvacation property classification within the scf dataset. while white households also report a relative preference for residential and vacation-type properties over business properties, the relative strength of this preference is not as great as it is for black households. in contrast to this interpretation of commercial real estate holdings, however, it is also likely that the absence of business property ownership across black households reflects a lower incidence of family business ownership within black households. while lach (1999) suggests that life insurance holdings often parallel real estate holding patterns across black households, this trend is not evident in the 1998 scf dataset. the mean value of life insurance assets across black households modestly exceeds the value of white households’ life insurance assets for every income category in the scf dataset. while these pairwise mean differences are not statistically significant, the life insurance category represents the only financial asset classification for which the dollar value of black household holdings exceeds that of white households. for all income groups, this supports the assertion by scott (1998) and badu et al. (1999) that african-americans prefer relatively conservative financial alternatives in their investment portfolios. reviewing financial asset holdings across the various income categories shown in table 1 reveals a striking difference in portfolio composition for relatively high-risk, high-return financial assets between black and white consumers. the absence of corporate debt and equity securities within black families’ investment portfolios—and the corresponding concentration of wealth in real property and life insurance assets across these households— creates a stark contrast with white households’ portfolio holdings. this absence of financial diversification, coupled with the concentration of wealth in lower-yielding financial assets and real property, signals that african-american households face far greater unsystematic financial risk, lower portfolio returns, and a diminished rate of wealth accumulation over time in their wealth-creating asset portfolios. while this conclusion is troubling, it is consistent with other research results investigating racial differences in investing preferences. a number of studies—including boyce (1998), zhong and xiao (1995), lach (1999), and gutter et al. (1999)—point out the wide disparity in equity ownership between black and white households. past research offers a number of different explanations for this disparity. bajtelsmit and bernasek (1996) attribute it to the greater relative influence of black women in making household investment decisions and the relatively greater risk aversion observed across these investors. lach (1999) and vatter and palm (1977) attribute it to a lack of understanding of corporate equity and debt instruments across black households, and limited access to information regarding these investment alternatives. still other researchers attribute it to differences in risk tolerance related to socioeconomic factors (schooley & worden, 1996), educational attainment (shaw, 1996), and a savings motive driven by near-future needs for cash, such as saving for college 352 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 expenses, rather than distant-future events, like retirement savings (lach, 1999). finally, burlew et al. (1992) attribute the absence of relatively illiquid financial investments—such as stocks, bonds, and 401-k retirement assets—in african-american households to stronger current consumption preferences across black households, while pitts, whalen, o’keefe, and murray (1989) and morrall (1996) suggest that blacks display a higher preference for very liquid investments as well as cash holdings. the scf results reported in table 1 support this final contention. for most financial assets, there is a wide disparity between black and white households’ portfolio holdings. this gap is much smaller in the case of bank transaction accounts, particularly for upper middleand high-income families, reflecting the black families’ greater preference for highly liquid financial assets. within lower-income families, however, the use of basic banking services is less prevalent among black households. the distance between the percentage of black and white families reporting holdings of checking and savings deposits narrows with increasing income, but the valuation gap between the dollar value of mean account balances across black and white households does not diminish in similar fashion. like other wealth-oriented asset categories, there is a persistent difference between bank demand, time, and savings deposit balances reported by black and white households. this difference remains unexplained by income, age, or advancing educational attainment. consistent with the notion that black families focus on near-term savings goals (lach, 1999), the use of tax-advantaged retirement savings vehicles such as ira and keogh accounts is much less pronounced across african-american families. moreover, tables 1 through 3 illustrate that the retirement savings gap does not diminish with increasing income, household age, or increased educational attainment. while blacks close much of the gap between differential residential property values as income, household age, and educational attainment levels rise; black households remain severely underinvested in retirement wealthbuilding categories. a wide disparity between the ira/keogh holdings of black and white households persists among even the most affluent, well-educated black households. as a final point of interest in table 1, it is noteworthy that while real estate ownership is far more common among lowerand middle-income white families, the use of debt financing to obtain personal real estate is quite similar across both races. following the convention suggested by reichenstein (1998), who shows that time-series price changes in real property and the corresponding mortgage instrument used to finance the property do not necessarily move in lockstep, making it inappropriate to measure the equity position in real estate by netting current mortgage balances against the fair market value of real property, table 1 reports gross property values and mortgage balances associated with these investments separately. examining proportionate data describing the prevalence of mortgage debt across racial groups in various income categories reveals that the incidence of mortgage debt and the magnitude of this debt are frequently quite similar for black and white families. what is different, however, is the size of mortgage debt relative to the market value of real property encumbered by this debt. across all income categories, black families are more heavily leveraged that their white counterparts, with reported mortgage balances representing a far larger percentage of the value of real property owned. turning to the results shown in table 2, which reports asset holdings while controlling for 353d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 differences in respondent age, we find results similar to those offered in the income stratification sample. again, common stock and corporate bond holdings are quite rare across african-american families, regardless of respondents’ age. while zhong and xiao (1995) report an increasing likelihood of stock ownership as respondent age increases within the african-american population, table 2 indicates that this increase is really quite modest. incidence of stock ownership within black households is 8.9 percent in the group of respondents below 35 years of age, and this proportion actually falls to 8 percent in the 35-to-44 year old age category before peaking at 9.4 percent among 45-to-54 year old respondents and falling to 3.5 percent among 55-to-64 year old respondents and less than 1 percent for respondents aged 65-to-74 years. a much stronger relationship between respondent age and the likelihood of equity ownership emerges within white families. here, the incidence of stock ownership is 15.2 percent for respondents below age 35, increasing to 21.8 percent for respondents between 35 and 44 years old, 25.6 percent for respondents aged 45 to 54 years, and 29.2 percent for respondents aged 55 to 64 years before it declines to 24.1 percent for respondents aged 65 to 74 years. for both racial groups, an inverted u-shape characterizes the relationship between respondent age and equity holdings, although the trend is much more pronounced across white households. interestingly, a similar inverted u-shape describes the relationship between respondent age and household investment in equity mutual funds, although for most age groups, the incidence of mutual fund ownership is surprisingly below the rate of individual stock ownership. this trend holds for both white and black consumer groups in the scf sample. in the case of individual equity holdings and equity fund holdings, the inverted u-shape profile corroborates the most recent research results reported in gutter et al. (1999), while contradicting the work of zhong and xiao (1995) and morrin and suarez (1983). using equity security holdings as a proxy for investment risk tolerance, it appears that risk tolerance first increases with increasing age, reaches a maximum somewhere between age 45 and 64, and then diminishes with further increases in age. the age at which risk tolerance appears to reach a maximum is lower for black households, at somewhere between 45 and 54 years of age, than it is for white households, at somewhere between 55 and 64 years of age. turning to differences in the dollar value of equity holdings across the two racial groups, it is clear that black households are seriously underinvested in equities, and this difference persists across all age groups observed in the scf dataset. for black households the mean equity investment starts at $490 for respondents below age 35, and rises to only $8,655 for respondents between 45 and 54 years of age before beginning to decline at higher age levels as households seek greater financial liquidity and investment safety with advancing age. in contrast, white households begin with a mean equity investment of $6,797 for respondents below age 35, and this investment grows to a maximum of $83,278 for respondents between the ages of 55 and 64 before it begins to decline. clearly, white households place a larger quantity of financial resources in equity investments, use these investments to realize a larger absolute level of capital appreciation, and begin liquidating these equity holdings in favor of safer, more liquid financial assets later in life than their black counterparts. the limited presence of corporate debt and equity investments in african-american households and the 354 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 limited time period over which these families invest in higher-yielding assets severely limits the ability of these households to build wealth over time. examining real estate holdings, respondent age appears to influence the rate at which families first acquire real property. while black respondents consistently lag their white counterparts in the sample proportions reporting ownership of residential real estate, the gap between the races is widest for younger respondents. in addition, the proportion of black households owning residential property accelerates dramatically between the 45-to-54 year old age category and the 54-to-64 year old group, while corresponding changes in property ownership proportions reported by white families is a much smaller increase across these two age categories. comparing the two racial groups, black households show an investment preference for residential real estate over almost all other types of assets, yet they tend to acquire this residential real estate at a more advanced age than their white counterparts. interestingly, the ownership incidence of bank-type financial assets—including both demand and savings deposits—follows an inverted u-shaped pattern across advancing age categories within african-american households, and a much more linear pattern across white households. ownership of banking products first rises with increasing age in black households, reaches a maximum within the 55-to-64 year old age category, and then falls within older families. in the white sample, virtually all respondent households maintain at least one bank account by the time they reach the 35-to-44 year old age category, and the proportion of households holding one or more banking products remains virtually unchanged across successively older age categories. attitudes toward bank products and banking relationships are a function of age within african-american households, while in white households these attitudes appear to be invariant to changes in age. many researchers—including yuh and hanna (1997), shaw (1996), and lach (1999)— suggest that preferences for holding risky assets in general, and equity securities in particular, rise with increasing educational attainment. the data reported in table 3 support this contention, as stock and bond ownership rates rise with increasing educational attainment across both black and white subgroups of the scf sample. for both races, equity holdings accelerate dramatically among households in which respondents possess a baccalaureate degree. even among respondents with advanced educational attainment levels, however, black equity holdings significantly lag their white counterparts in both the proportion of the sample owning stock and the mean market value of equity holdings. it is noteworthy that increased educational attainment occurring below the collegegraduate category contributes to increased family holdings of risky assets, but only in a modest way. both black and white respondents who report completing some undergraduate college coursework or receiving a high school diploma evidence only slightly more stock and bond holdings than respondents who failed to complete a secondary school education. earning a college degree is a significant determinant of household investment patterns across both black and white households. the impact on the risk-return characteristics of family asset portfolios, however, remains much more pronounced among white households, which evidence greater diversification across financial asset categories and substantially greater investment in stocks, bonds, and mutual fund assets. in virtually all cases, the proportion of black households reporting ownership of a particular financial asset lies well below the percentage of white families reporting owner355d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 ship of the same asset. the exception to this generalization is life insurance holdings across black households. within white households, the percentage of respondents reporting some life insurance ownership begins at 62 percent in families that fail to complete a high school education and rises progressively with increasing educational attainment until reaching 83 percent among respondents who report attending or completing graduate school. black households report remarkably similar rates of life insurance ownership, beginning at 60 percent for black respondents who fail to complete high school, rising to 81 percent across those who have attended or completed graduate school. the data suggest that black households prefer the relative stability and security of life insurance products over riskier and more price volatile investments is stocks, corporate bonds, and other brokerage assets. this preference transcends increasing income levels, respondent age, and educational attainment levels within the scf sample. finally, educational attainment levels represent a much more important determinant of household banking relationships among black consumers than white consumers. the proportion of black households reporting use of these traditional commercial banking products rises dramatically with increasing education, from less than 30 percent for respondents who fail to complete high school to over 96 percent for respondents who have attended or completed graduate school. across white respondents, more than 72 percent of non-high school graduates report demand deposit account ownership, and this number rises to 99 percent across respondents who have attended or completed graduate school. at higher levels of educational attainment, black households’ commercial banking relationships closely resemble their white counterparts. at lower levels of educational attainment, particularly in the case of individuals who fail to complete a secondary school education, black consumers’ banking relationships are far different than those observed within white households, indicating that a disproportionate number of unbanked consumers are concentrated within the african-american community. 4. investment portfolio preferences within african-american households it is clear from the statistical information presented here and from a review of the recent academic literature that differences between african-american and caucasian households in terms of financial services ownership patterns and wealth-accumulating asset portfolios are broad and substantive. in particular, black households control smaller asset portfolios than their white counterparts. this conclusion persists when controlling for income, age, and education attainment differences between the two racial groups. in addition, black financial portfolios reflect a clear preference for near-term savings, such as savings for a planned consumer purchase or savings for a college education, at the expense of distant-term savings, such as retirement saving. given this preference, african-american retirement account balances are significantly smaller than the average retirement savings balance observed across caucasian households. consistent with a preference for near-term savings, black households value liquidity more than their white counterparts, showing a greater relative inclination to hold cash and cash-equivalent assets while foregoing less liquid assets offering higher rates of return. this 356 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 liquidity preference occurs across all but the lowest income strata. as income levels increase, commercial bank patronage patterns across black and white consumer groups become quite similar, while divergent patronage patterns characterize the lowest income category. within low-income households, caucasians are far more likely than african-americans to maintain conventional demand and time deposit account relationships with commercial banks. at all income, education, and age levels, however, african-american households invest a smaller percentage of their portfolios in the form of mutual funds, brokerage accounts, and outright equity purchases than caucasian households. in addition, black households demonstrate a distinct preference for safety and security in their investment preferences, favoring life insurance and real estate assets over corporate debt and equity securities across all levels of household income and educational attainment. in most african-american households, life insurance represents the single-most important financial investment. life insurance participation rates for black families exceed those of white families across all income categories. moreover, for middleand upper middle-class households, the reported value of life insurance holdings across black families exceeds the corresponding value of white households’ holdings. this result is particularly noteworthy in light of the large body of academic literature that reports the absolute value of household wealth across african-american households to be significantly below the corresponding value in caucasian households. in relation to total household wealth, it is clear that life insurance holdings represent a far more dominant financial asset in black households than in white households. turning to non-financial investment assets, the market value of real property owned by black families—including owner-occupied residential real estate, vacation property, and other nonvacation property—lies substantially below the corresponding property valuation levels reported by white households. similar to the pattern observed for financial asset holdings across the two racial groups, this real estate valuation gap persists when controlling for income, age, and educational attainment differences between the two groups. in addition, black households report greater mortgage indebtedness than their white counterparts, particularly with respect to owner-occupied residential property. african-american families also wait until later in life, typically between the ages of 45 and 54, to acquire residential real estate. in contrast, caucasian households display a major surge in residential property ownership rates after age 35. in combination, these findings suggest a troubling investment pattern in which africanamerican households underinvest in financial asset categories offering relatively high returns, such as common stock, mutual funds, corporate bonds, and municipal bonds, because they perceive that these securities carry unacceptable investment risks and offer diminished liquidity. at the same time, african-american households overinvest, in relative terms, in highly liquid financial assets offering heightened liquidity only in exchange for relatively lower rates of return over shorter investment periods. as a consequence, these households forego the benefits of financial compounding over a lengthy investment horizon at the higher annualized rates of return associated with debt and equity investments. finally, africanamerican households emphasize the notion of intergenerational wealth transfer over the lifetime accumulation of wealth as the dominant goal of investment strategy. this is reflected in a preference across black households for insurance-based contractual financial arrange357d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 ments that monetize upon the death of the investor, at the expense of investment-based financial asset purchases that build value over the investor’s life span. collectively, these investment strategies result in a diminished rate of wealth accumulation over time, and a smaller stock of total accumulated wealth at any given point in time, within africanamerican households. 5. summary and recommendations this paper has shown that the african-american market for financial services has grown rapidly in size and importance. this growth represents a major opportunity for the marketers of financial services, because statistical evidence indicates that the financial asset profiles of black households trail those of their white counterparts in terms of breadth and depth of holdings, particularly in the area of relatively risky, high yield financial assets. moreover, the paper has shown that the african-american segment of the market demonstrates some unique attitudes and preferences that impact purchase decision patterns. to respond to the opportunity to serve the needs of this emerging segment of the market, this paper shows that it is important for financial service providers to be aware that investment preferences and asset accumulation patterns differ across different racial groups. examined from this perspective, the paper provides financial planners seeking to serve the african-american community with information to understand the community, tailor investment information to the unique needs of this community, and interact with this community to render effective service to the families and individuals who comprise this attractive and growing segment of the financial services marketplace. references bajtelsmit, v., & bernasek, a. (1996). why do women invest differently than men? financial counseling and planning, 7, 1–9. badu, y. a., daniels, k. n., & salandro, d. p. (1999). am empirical analysis of differences in black and white asset and liability combinations. financial services review, 8 (3), 129–147. blau, f. d., & graham, j. w. (1990). black-white differences in wealth and asset composition. quarterly journal of economics, 105 (2), 321–339. board of governors of the federal reserve system. (2000). codebook for 1998 survey of consumer finances. washington, d.c.: federal reserve system (february 15). boyce, j. (1998). blacks are bearish on wall street says new survey. the wall street journal (april 9), c-21. brimmer, a. (1991). building wealth and assets. black enterprise (july), 32. burlew, a., banks, w., mcadoo, h., & azibo, d. (eds.). (1992). african-american psychology, newbury park, nj: sage publications, inc. gutter, m. s., fox, j. j., & montalto, c. p. (1999). racial differences in investor decision making. financial services review, 8 (3), 149–162. kennickell, a. b., starr-mccluer, m., & surette, b. j. (2000). recent changes in u.s. family finances: results from the 1998 survey of consumer finances. federal reserve bulletin, 86 (january), 1–29. lach, j. (1999). the color of money. american demographics, (february), 59–60. montalto, c. p., & sung, j. multiple imputation in the 1992 survey of consumer finances. financial counseling and planning, 7 (1996), 133–146. 358 d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 morin, r. a., & suarez, a. f. (1983). risk aversion revisited. journal of finance, 38 (4), 1206–1216. morrall, k. (1996). appealing to the african-american market. bank marketing, 28 (5), 18–23. myers, s., & chung, c. (1996). racial differences in home ownership and home equity among preretirement-aged households. the gerontologist, 36 (3), 350–360. reichenstein, w. (1998). calculating a family’s asset mix. financial services review, 7 (3), 195–206. schooley, d. i., & worden, d. d. (1996). risk aversion measures: comparing attitudes and asset allocation. financial services review, 5 (2), 87–99. scott, m. (1998). don’t let the bull market pass you by. black enterprise, 28 (june), 243. shaw, k. (1996). an empirical analysis of risk aversion and income growth. journal of labor economics, 14 (4), 626–653. yuh, y., & hanna, s. (1997). the demand for risky assets in retirement portfolios. presented at the annual meeting of the academy of financial services. u.s. department of commerce, bureau of the census. (1998). statistical abstract of the united states. washington, d.c.: u.s. government printing office. williams, j. d., & qualls, w. j. (1989). middle-class black consumers and intensity of ethnic identification. psychology and marketing, 6 (winter), 263–286. wolff, e. n. (1994). trends in household wealth in the united states: 1962–1983 and 1983–1989. working paper, 94-103. new york university department of economics (february). zhong l., & xiao, j. (1995). determinants of family bond and stock holdings. financial counseling and planning, 6, 107–114. 359d.a. plath, t.h. stevenson / financial services review 94 (2000) 343–359 pii: s1057-0810(00)00061-5 market timing using strategists’ and analysts’ forecasts of s&p 500 earnings per share richard chunga, lawrence kryzanowskib,* afaculty of commerce, concordia university, 1455 de maisonneuve blvd. west, montreal, quebec, canada, h3g 1m8 bned goodman professor of finance, faculty of commerce, concordia university, 1455 de maisonneuve blvd. west, montreal, quebec, canada, h3g 1m8 received 14 june 1999; received in revised form 2 august 2000; accepted 26 august 2000 abstract this paper examines the bias in and usefulness of top-down and bottom-up consensus forecasts of earnings per share for the s&p 500 index provided by market strategists and analysts to i/b/e/s. these forecasts exhibit a significant optimism bias that decreases over the 12 months up to release of actual earnings per share. the bias is significantly more pronounced for the bottom-up forecasts of analysts. unlike the findings for country timing, we demonstrate that a stock market timer using switching rules based on the consensus forecasts of s&p 500 earnings or the directional switch in the consensus or in the number of switchers cannot generate a free lunch. © 2000 elsevier science inc. all rights reserved. jel classification:g10; g11; g14 keywords:market timing; earnings forecasts; performance 1. introduction market timing is a money management style that attempts to add value by switching from equities to cash and vice versa based on signals typically generated by mathematical models. as noted by benjamin graham inthe intelligent investor(1954), the current popularity of * corresponding author. tel.:11-514-848-2782; fax:11-514-848-4500. e-mail address:lkryzan@vax2.concordia.ca (l. kryzanowski). financial services review 9 (2000) 125–144 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00061-5 a money management style depends on the performance of the market over the recent past. the belief in buy-and-hold investing (market timing) increases (decreases) in popularity and press coverage as the market progresses through a sustained bull market, and decreases (increases) in popularity and press coverage as the stock market progresses through a bear phase. to some extent, this also applies to academic research interest in market timing. thus, the unresolved debate over whether or not market timing can generate superior risk-adjusted performance over a buy-and-hold all-equity strategy is likely to intensify once we leave the current bull phase of the equity market. thus, most investment textbooks (e.g., bodie, kane & marcus, 1999; reilly & brown, 2000) that are used to prepare a new generation of money managers devote significant space to active investment management, on how to measure timing performance, and on how successful mutual funds are in timing movements in the market or its volatility. the academic community is divided on whether the investment community places too little or too much reliance on the earnings forecasts of analysts. brown (1996) argues that the reliance is too little, dreman and berry (1995) argue that the reliance is too much. while no one appears to have formulated and tested methods for domestic market timing using consensus forecasts of earnings per share, at least two studies examine methods for domestic or global asset allocation using consensus forecasts of earnings per share by analysts. emanuelli and pearson (1994) find that aggregate forecasts of earnings per share by analysts (a bottom-up approach) improves country selection (a top-down approach), and that the earnings-estimate revision ratio (an aggregate measure of changes in analysts’ forecasts of earnings per share) enhances returns from international equity allocation over their studied 52-month test period. similarly, bercel (1994) finds that the forecast data of u.s. and non-u.s. analysts, as measured by changes in the forecasts of earnings per share by analysts and the number of analysts changing their forecasts, can be used to generate abnormal returns in seven international markets. our contribution is to test whether a market timing strategy using the signals based on the bottom-up consensus forecasts of earnings per share by analysts and the top-down consensus forecasts of earnings per share by market strategists for the s&p 500 generates superior investment performance. we also examine whether better market timing performance is achievable using the direction of these consensus forecast revisions by market strategists and analysts, or the directional net number of such forecasts revised up or down by market strategists and analysts. the latter switching rules are similar to those used by emanuelli and pearson (1994) and bercel (1994) to obtain superior country timing investment performance. our research results in four major findings. first, we find a significant optimism bias in bottom-up and top-down forecasts of earnings per share by analysts for the s&p 500 index for the current fiscal year (fy1) and subsequent fiscal year (fy2). second, we find that the optimism bias is significantly higher in the bottom-up forecasts compared to the top-down forecasts on average, and in each of the months approaching the month during each year in which i/b/e/s updates the actual earnings per share for the s&p 500 index. third, we find that not only do these optimism biases decrease over the year but that the biases exhibit temporary reversal in january prior to the month during each year in which i/b/e/s updates the actual earnings per share for the s&p 500 index. fourth, we demonstrate that a market timer using parsimonious switching rules based on the top-down (bottom-up) consensus 126 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 forecasts of earnings per share by market strategists (financial analysts) or various directional measures of forecast revisions cannot generate a free lunch or positive alpha (i.e., superior investment performance). the next section addresses individual investor usage of market timing. the third section provides a review of the relevant literatures. the fourth section examines the bias in consensus top-down forecasts of earnings per share for the s&p 500 of market strategists and the bottom-up equivalents for financial analysts for each of the next two years. the fifth section discusses the data. the sixth and seventh sections evaluate the market timing performance of switching rules based on consensus forecasts of s&p 500 earnings per share and revisions thereof, respectively. we end with some concluding remarks. 2. market timing and the individual investor market timers are fund managers, registered investment advisors, accounts, agents of record and other qualified persons who make market timing decisions and recommendations on behalf of individual investors, or effect such transactions on the instructions of individual investors. a market timer’s goal is to improve risk-adjusted performance by reducing risk and/or enhancing return. unlike other forms of active asset allocation, market timers are 100% invested in equities or in cash, or are long in one of these asset classes and short in the other. market timers manage considerable capital. to illustrate, the more than 200 members of the society of asset allocators and fund timers, inc. or saafti manage an estimated $14 billion (http://www.saafti.com/advisors/saafti/home.asp, july 24, 2000). various service providers expend considerable resources to provide market-timing recommendations to individual investors, and to evaluate the performance of these recommendations. three examples follow. first,hulbert financial digest(see http://www.hulbertdigest.com/) calculates “timing-only” returns for investment newsletters. about one-half of the newsletters it monitors provide timing signals and have significant individual subscribers (according to an e-mail response from mark hulbert dated 20 july 2000). second,moniresearch newsletter (see http://www.moniresearch.com/) tracks and measures the performance of client accounts of 100 timing-only money managers, including rydex funds, profund timers and mutual funds. third,timer digestmonitors over 100 of the leading market timing models, and provides commentary on the top funds based on their ranking of performance. in practice, timers use a variety of investment products to effect their timing strategies, including mutual funds (e.g., bull and bear funds designed to correlate positively and negatively with the major indices), variable annuities, equity baskets, exchange-traded index-linked securities (e.g., s&p 500 depository receipts or spdrs, worldwide equity benchmark shares or webs or i shares, djia depository receipts or diamonds and nasdaq 100 trust units or qqq), and futures on the national market indexes. a number of individual investors confine their timing decisions to their variable annuity and retirement savings plan accounts (i.e., self-directed ira, 401k or keogh accounts) to minimize the tax impact of effecting timing decisions. the practice of timing necessarily converts an unrealized capital gain into an immediate realized gain and an associated tax liability (jeffrey & arnott, 1993). 127r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 market-timing using variable annuity accounts and mutual funds cause concern among the sponsors and managers of these plans due to their allegedly adverse effect on fund performance (koco, 1995). the actions of market timers (short-term investors) allegedly drives up fund costs that are primarily borne by long-term unitholders. these costs include higher taxes, trade costs incurred to fund redemptions and to invest inflows, and larger cash balances to protect against unexpected large redemptions or as the result of unexpected large cash purchases of fund units. to alleviate these problems, many companies impose redemption fees, minimal holding periods and limits on transaction sizes by their unitholders or shareholders. others (such as rydex series trust, profund advisors and protomac funds) have investment vehicles to cater to this active investment market. in may 2000, the vanguard group announced that it had filed with the sec to offer exchange traded funds (so-called viper shares) on five of its most prominent index funds (see http: www.saafti.com/advisors/ saafti/newsletter.asp?storyid584623). these viper shares, which are to be listed on the amex, offer investors advantages in terms of tax efficiencies, minimal expense ratios, and continuous trading possibilities with prices continuously marked-to-market and minimal trade impediments. 3. literature survey 3.1. bias in the forecasts by analysts and strategists numerous studies examine whether or not analysts produce unbiased forecasts of investment-relevant information such as earning per share (eps), and whether or not analysts systematically underreact or overreact to new information. brous (1992), brous and kini (1993), francis and philbrick (1993), kang, o’brien and sivaramakrishnan (1994) and dreman and berry (1995) find that analysts normally produce upwardly biased forecasts of earnings per share. harris (1999) finds a similar bias in the long-run forecasts of analysts. while debondt and thaler (1990) find that analysts systematically overreact to new information, ali, klein and rosenfeld (1992), elliot, philbrick and wiedman (1995) and teoh and wong (1997) find evidence that suggests that analysts systematically underreact to new information. easterwood and nutt (1999) report that the reaction of analysts depends upon the nature of the information that becomes available. specifically, their evidence indicates that analysts underreact to new negative information, do not react in the absence of new information, and overreact to new positive information. they conclude that their results are consistent with the hypothesis that analysts are systematically optimistic in their interpretation of new information. lin and mcnichols (1998) report that the optimism of earnings forecasts by analysts depends upon their underwriting relationships. lead and co-underwriter forecasts of earnings growth are significantly more favorable than those made by unaffiliated analysts. das, levine and sivaramakrishnan (1998) find that analysts issue more optimistic forecasts for low predictability firms. chopra (1998) finds that the average consensus earnings per share growth forecasts made by analysts for the s&p 500 index over the 1985–1997 time period is almost twice the actual 128 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 growth rate, and is revised downward continuously over the course of the year. chung and kryzanowski (1999) examine the top-down forecast accuracy and divergence of market strategists for quarterly earnings per share forecasts for the s&p400 and s&p 500 indexes. they find that such forecasts are, on average, optimistically biased, and that the bias is positively related with both the number of market strategists reporting their forecasts to i/b/e/s each month, and the coefficient of variation of such forecasts. 3.2. efficacy of market timing strategies and market timers a number of papers calculate the required forecasting ability to successfully time the market or simulate the performance of investors with different forecasting abilities. sharpe (1975) calculates that a market timer needs to be correct over 75% of the time to outperform a passive, all-equity portfolio. clarke, fit, gerald, berent and statman (1989) use simulation to conclude that a market timer with even modest amounts of information who follows optimal decision rules can outperform a buy-and-hold investor. beebower and varikooty (1991) demonstrate that most of the common tests for detecting significant ability to generate excess returns of up to two percentage require assessment time periods well beyond human life expectancy. shilling (1992) shows that being long in stocks during bull markets is not as profitable as being out of the market during bear markets even if an investor is not invested in many major bull markets. reichenstein and rich (1994) propose that some market timing is justified by the empirical evidence on the partial predictability of stock returns in the long-term. bierman (1995) argues that market timing is an art and not a science, since economists need hindsight to identify market bubbles. wagner (1997) finds that a timer could miss as much as 20% of the tops and bottoms of the s&p 500 over the 108 years since 1885 and still match the average performance of a buy-and-hold investor. a number of papers report on the effectiveness of specific market timing strategies or market timers. hardy (1990) concludes that regression forecasts using models with macro variables are so good that tactical asset allocation can improve the gross return/risk trade-off considerably, even for investment portfolios confined solely to domestic assets. wagner, shellans and paul (1992) examine the performance of twenty-five investment advisors who performed market-timing services for clients and were monitored by the newsletter published by moniresearch. for the period of october 1985 through the end of september 1990, they find that these advisors outperformed a buy-and-hold strategy. brocato and chandy (1994) argue that selection bias can easily produce the results reported by wagner et al. (1992). brocato and chandy (1994) find that the average record of their twenty-five random timers is identical to the average record of the twenty-five real timers studied by wagner et al. (1992). larsen and wozniak (1994/95, 1995) find support over the 1977–1992 period for market timing in the real world based on results for a discrete timing regression model for market timing. over the period studied by brocato and chandy (1994), larsen and wozniak (1994/95, 1995) find significantly superior results for their method of market timing compared to the randomization strategy of brocato and chandy for out-of-sample tests. brocato and chandy (1995) question the robustness of the conclusions of larsen and wozniak 129r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 (1994/95, 1995) to the imposition of realistic transaction costs and taxes to the market-timing results. reichenstein and rich (1993) report that timing portfolios based on market risk premium display a stronger ability to time the market than those based on dividend yield and earnings-price ratio. fuller and kling (1994) find that the models studied by fama and french (1989) are not reliable signal generators for market timing when trading costs, subperiods and consistency across models is considered. prather and bertin (1998) find performance that is superior to a passive buy-and-hold strategy for a market timing trading rule that uses the public information contained in discount rate changes to signal entry and exit into the stock market. breen, glosten and jagannathan (1989) find that a portfolio managed by the predictions of a three-year rolling regression of excess stock return on the one-month risk-free rate is worth an annual management fee of 2% of the value of the assets managed. lee (1997) finds that the value added by market timing identified by breen et al. (1989) is completely eroded by april 1989, and becomes negative when the studied time period includes observations after april 1989. copeland and copeland (1999) test changes in the implied volatility of options on stock index futures as market-timing strategies for reallocating assets among portfolios of various sizes and styles. they conclude that market timing may be feasible at least for portfolio yield enhancement. due to the large number of studies (e.g., ferson and schadt, 1996; kryzanowski, lalancette & to 1997) that assess the market-timing ability of mutual fund managers in various countries, we only note herein due to journal space constraints that these studies predominantly report evidence of little or poor market-timing ability. a recent exception is the study by busse (1999) who finds that market volatility timing (as opposed to market timing) has led to higher risk-adjusted returns. graham and harvey (1994, 1996, 1997) examine the performance of the asset-allocation strategies of 326 newsletters drawn from thehulbert financial digestfor the 1983–95 period. they find that the group appears not to possess any special information about the future direction of the market. 4. data we use the annual estimates of earnings per share for the current and subsequent fiscal year (fy1 and fy2) for the s&p 500 index that are available from i/b/e/s on both a top-down and bottom-up basis. our data set consists of 218 months of such annual forecasts over the period from january 1982 through february 2000. each top-down consensus forecast is the cross-sectional average of the individual forecasts of earnings per share for the s&p 500 made by market strategists each month. for the interested reader, i/b/e/s provides both quarterly and annual top-down consensus forecasts of earnings per share for fy1 through fy3 under the ticker symbol sap5 on a regular basis. the bottom-up consensus forecasts of earnings per market-weight share are a weightedaverage of the consensus forecasts of the earnings per share for each firm included in the s&p 500, where each weight is equal to the weight of that firm in calculating the stock price 130 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 index for the s&p 500. in other words, the bottom-up consensus forecast is a market-value weighted average of the consensus forecasts of earnings per share for each of the firms in the s&p 500 index. for ease of exposition, we refer to this value as earnings per share throughout this paper. the bottom-up consensus forecasts of earnings per share are available from i/b/e/s on a regular basis but are not included in the data package commonly made available to researchers. 5. s&p 500 earnings forecasts and their biases 5.1. expectations and methodology the s&p 500 forecasts of earnings per share generated by both strategists and analysts studied herein are expected to display systematic optimism. as predominantly sell-side employees of brokerage and investment banking firms, these sell side professionals have economic incentives to promote stock purchase rather than to produce the most accurate forecasts (womack, 1996; carleton, chen & steiner, 1998). das, levine and sivaramakrishnan (1998) assert that analysts may engage in deliberate optimism to obtain private information from firm management that can produce earnings forecasts that are substantially better than those produced from only using public information. furthermore, the s&p 500 forecasts of analysts are expected to exhibit more systematic optimism than those of strategists because analysts unlike strategists derive part of their comparative advantage (expertise) from their superior access to the top management of the firms that they follow, and analysts do not want to jeopardize investment banking relationships between the firm they work for and the firm for which they are forecasting earnings (womack, 1996). many examples exist where such access has been adversely affected by a less than favorable report by an analyst (pratt, 1993). in addition, the consensus bottom-up forecasts have a selection bias not present in the consensus top-down forecasts caused by analysts discontinuing the production of forecasts for firms that do not generate sufficient commission revenues for their employers or for which they are pessimistic. the decimal forecast error in the average or consensus fy1 and fy2 forecasts of the earnings per share for the s&p 500 index are calculated monthly for the top-down forecasts made by market strategists and the bottom-up forecasts made by financial analysts. the forecast error in month t in the consensus forecast of the earnings per share for the s&p 500 index for fiscal year i (i5 fy1 or fy2) for reporting group j (j5 strategists or analysts) in month t, feijt, is given by: feijt 5 (fijt /aijt ) 2 1 (1) where fijt is the consensus forecast in month t of earnings per share for the s&p 500 index for fiscal year i for reporting group j; and aijt is the actual earnings per share for the s&p 500 index for fiscal year i that corresponds to the forecast made for fiscal year i by reporting group j in forecast month t. 131r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 in the i/b/e/s database, the top-down and bottom-up forecast horizons for fy1 and fy2 are moved forward by one year at the end of march and february, respectively, of each calendar year. as noted in section two, the practice of market timing necessarily converts an unrealized capital gain into an immediate realized gain and an associated tax liability (jeffrey & arnott, 1993) unless the timing transactions are being incurred in a tax-deferred account. in the following tests, we ignore tax considerations. this turns out not to be an important omission since our tested market timing strategies do not outperform a passive portfolio with an equivalent average level of risk. 5.2. the results the mean and median consensus forecasts of earnings per share in dollars for fy1 and fy2 for the top-down and bottom-up approaches are reported in the first two rows of numbers in panel a of table 1. although not shown in the panel, both of the top-down consensus forecasts of the earnings per share are significantly lower than their bottom-up counterparts at the 0.05 level (the implied significance level from this point onwards unless noted otherwise). the mean and median forecast errors are reported in the last two rows, respectively, of panel a of table 1. the bottom-up forecasts of financial analysts exhibit a statistically significant mean optimism bias of 17.5% and 30.5% for fy1 and fy2, respectively. such a bias is identified previously by ali, klein and rosenfeld (1992), among others. in a similar but more muted vein, the top-down forecasts of market strategists exhibit a significant mean bias of 7.7% and 12.0% for fy1 and fy2, respectively. although the median values are consistently lower, tests using the median values yield similar inferences. interestingly, all of these biases are considerably higher than the values reported in a previous version of this paper for the 167-month period from january 1982 through november 1995. specifically, for this shorter time period, the mean optimism biases of analysts (strategists) are a significant 8.0% (insignificant –0.1%) for fy1, and a significant 19.2% (significant 3.5%) for fy2. the quite different results for the two time periods suggest that the optimistic biases of analysts and of strategists may vary across both calendar and relative time, and that this time-variation will not be reflected well in more parsimonious markettiming strategies without data snooping. these differences also suggest that the optimistic biases of analysts and of strategists may be positively related to the relative weight of bull market months in the time period studied. these are topics for future study. the four series of average forecast errors of earnings per share relative to their respective annual switch months are reported in panel b of table 1. the top-down and bottom-up values are relative to the end of march and february, respectively (i.e., their switch month 0) because, as noted above, the year-end being forecasted is updated or advanced by one year on these dates in the i/b/e/s database. the numbers in this panel are all positive, which signifies optimism biases in the top-down and bottom-up fy1 average forecasts. these biases decrease over the year, as the switch month is approached (i.e., actual values of earnings per share become known). the optimism bias also decreases over the eleven (ten) months up to the first month prior to switch for the bottom-up (top-down) fy2 average 132 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 forecasts, and increases in the remaining month (two months) prior to the switch month of february (march). whether or not the temporary increase in the optimism bias in january for both the top-down and bottom-up consensus forecasts of earnings per share for fy2 is related to the well-known january anomaly requires further study. table 1 average consensus forecasts and forecast errors for the s&p 500 index measure for earnings per share statisticc top-down consensus forecast of earnings per share by strategists for fiscal year bottom-up consensus forecast of earnings per share by analysts for fiscal year fy1d fy2d fy1 fy2 panel a: mean and median consensus forecasts and forecast errors of earnings per share for the s&p 500 index mean 25.97 28.38 28.53 33.49 forecasta median 23.03 25.93 25.78 30.31 mean 0.0770** 0.1199** 0.1746** 0.3053** forecast errorb median 0.0672** 0.0953** 0.1441** 0.2301** panel b: mean consensus forecast errors of earnings per share for the s&p 500 index relative to the month that i/b/e/s updates actual earnings per share for the s&p 500 index in its database month relative to i/b/e/s update of actual earnings per share top-down consensus forecast of earnings per share by strategists for fiscal year bottom-up consensus forecast of earnings per share by analysts for fiscal year fy1d fy2d fy1 fy2 212 0.0884 0.1332 0.3020 0.4301 211 0.0847 0.1302 0.2081 0.3508 210 0.0944 0.1278 0.1995 0.3327 29 0.0918 0.1272 0.2013 0.3156 28 0.0915 0.1297 0.1904 0.3096 27 0.0860 0.1309 0.1778 0.3020 26 0.0786 0.1238 0.1670 0.2953 25 0.0671 0.1164 0.1563 0.2863 24 0.0584 0.1085 0.1413 0.2762 23 0.0498 0.1009 0.1271 0.2622 22 0.0454 0.1077 0.1177 0.2503 21 0.1332 0.1052 0.1038 0.2506 mean and median consensus forecasts and forecast errors of earnings per share for the s&p 500 made by strategists and analysts, and tests of their significance, are presented in panel a. mean consensus forecast errors of earnings per share for the s&p 500 for the months prior to the month that i/b/e/s annually updates the actual annual earnings per share for the s&p 500 are reported (without testing for their significance) in panel b. monthly forecasts are examined for the 218 month period from january 1982 through february 2000. a dollars of earnings per share for the s&p 500 index. b forecast error is equal to the [(forecast earnings per share)4 (actual earnings per share)]21. c tests are conducted for forecast errors against zero. d fy1 and fy2 refer to the current and subsequent fiscal years, respectively. * and ** indicate significance at the 0.05 and 0.01 levels, respectively. 133r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 6. tests of classical market timing strategies using consensus s&p 500 earnings forecasts we now formulate and test several possible methods for market timing using strategists’ top-down and/or analysts’ bottom-up forecasts of earnings per share for the s&p 500 index for fy1 and fy2 to generate average one-month expected returns. the market timing is classical because the switching rule is based on a comparison of the expected return on risky assets with that on risk-free assets. we begin with a discussion of our test procedures. 6.1. test procedures unlike common practice in the literature (e.g., lee, 1997; wagner, 1997), we not only examine portfolio performance for investors whose natural or preferred habitats are to remain in stock (stock) or in cash (cash) but also in neither (none). we assume a 1% transaction cost for all switches between all equity as proxied by the s&p 500 index and all cash as proxied by 30-day t-bills, and adjust the switching rules to account for the transaction cost when switches are away from the preferred habitat. our switching cost is higher than the 0.5% rate used by fuller and kling (1994). for those investors without a natural habitat, the switching rules are examined with and without a transaction cost charge for any switch. the switching rules used are summarized in table 2. all switches are based on a comparison of theexpectedone-month-hence return on equity with the current onetable 2 switching rules for classical market timing for various natural habitatsa switch to:c natural habitatb stockd cashd noned nonee cash if predicted s&p 500 returnf is less than t-bill returnf minus 1% less than t-bill return less than t-bill return minus 1% less than t-bill return stock if predicted s&p 500 return is greater than t-bill return greater than t-bill return plus 1% greater than t-bill return plus 1% greater than t-bill return this table provides the switching rules used in market timing using the consensus earnings per share forecasts for the s&p 500 made by market strategists and financial analysts. the market timing is classical since the portfolio is either fully invested in equities or in cash. transaction costs are ignored and are considered at a rate of 1% of asset value for each switch. a classical market timing refers to switching between cash and equities based on market expectations. b the natural habitats are to remain in stock, in cash, and in neither (none). c the switching rules compare the predicted return on the s&p 500 with the current return on t-bills. d switching rule reflects a 1% transaction cost charge when moving away from this natural habitat. e switching rule reflects no transaction cost charge when moving away from this natural habitat. f the return on equities is proxied by the return on the s&p 500 index and the return on cash is proxied by the return on t-bills. 134 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 month t-bill rate, with or without an adjustment for transaction costs. as expected, the mean s&p 500 monthly return of 1.47% is substantially higher than the corresponding return of 0.50% for t-bills over the studied period. we test the predictive value of the top-down forecasts supplied by the strategists, the bottom-up forecasts supplied by the analysts, and both the top-down and bottom-up forecasts (switch if both screens are satisfied separately). for each of these three types of consensus forecasts, we use two different approaches for calculating the expectedaverageone-monthhence return for the s&p, which is needed in the switching rules. thus, our tests of the efficacy of market timing also depend upon the validity of these models in determining what prices will be in the future, and on the validity of using consensus forecasts of earnings as a proxy for the earnings expectations that will be reflected in stock prices at a future point in time. the first orcurrent earnings-to-price (e/p)approach calculates the predicted one-month return as the summation of the two components of expected total return (i.e., dividend yield plus capital gain return), after calculating each component’s average monthly compound rate of return over the given forecast horizon. for example, for a forecast horizon of eight months, the average monthly compound rate of capital gain is equal to the eighth root of the holding period rate of capital gain. the capital gain component of total expected return uses the price predicted for the end of the forecast horizon. for example, using the top-down consensus forecast for fy1, the predicted or expected price of the s&p 500 at the end of the forecast horizon (i.e., the switch month of march) is obtained by multiplying the current month’s consensus top-down forecast of earnings per share for the s&p 500 for fy1 by thecurrent price-to-earnings ratiofor the s&p 500. alternatively, it is obtained by dividing this consensus forecast by thecurrent earningsto-price ratio that proxies for the market’s current rate of capitalization of a dollar of current s&p 500 earnings. the dividend yield component of total expected return uses the average monthly compound rate that yields the current month’s annual dividend yield. more formally, the expected average monthly return for an investment at the beginning of month t in the s&p 500 index based on the expected realization of the consensus forecast of earnings per share for fiscal year i provided to i/b/e/s by reporting group j that is available at the beginning of month t, rijt, is given by: rijt 5 hffijtspt at ds 1 pt dg 1 sij 2 1j 1 @~dt! 1 12 2 1# (2) where fijt is the consensus forecast of earnings per share for the s&p 500 index for fiscal year i (i5 fy1 or fy2) provided to i/b/e/s by reporting group j (j5 strategist or analysts) that is available at the beginning of month t; pt is the actual price or level of the s&p 500 index at the beginning of month t; and sij is the number of months from the beginning of month t 135r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 to the switch month that is relevant for fiscal year i for a forecast to i/b/e/s by reporting group j. cancelling out the pt terms in eq. (2) and rearranging (2) yields: rijt 5 fsfijt at d 1 sij 2 1g 1 @~dt! 1 12 2 1# (3) thus, the return estimated using the first approach is equivalent to that obtained using a gordon valuation model, when the capital gain component of total expected return is estimated as the average monthly compound rate of change in the consensus forecast of earnings per share for fy1 (or fy2) for the current month when benchmarked against the most current actual earnings per share available at the beginning of that month. the equivalence of the approaches assumes that the dividend payout ratio remains constant. the second orpremiumapproach replaces the reciprocal of thecurrent earnings-to-price in eq. (2) with a premium-based earnings-to-price, which is calculated as the current risk-free rate plus the historical average premium of earnings-to-price over the return on one-month t-bills. the historic premium is proxied by the most recent 36-month movingaverage premium. the metric proposed by elton and gruber (1991) is used to measure market timing ability for each portfolio. the proposed metric for month t is equal to the excess return on the portfolio for month tminus the excess return that would have been obtained on the portfolio in month t if the portfolio maintained its average actual beta at all points in time. more formally, the proposed metric, dt, for month t is: dt 5 r t 2 rb* t (4) where rt is the excess return on the portfolio for month t, and rb*t is the excess return that would have been obtained on the portfolio in month t if the portfolio maintained its average actual beta (represented byb*) at all points in time. this average actual beta portfolio represents a buy-and-hold portfolio with a holding period equivalent to the time period studied, namely, 218 months or slightly over 18 years. each excess return for the portfolio is obtained by subtracting the t-bill rate from the portfolio’s return. this metric correctly measures performance given market timing because additional information is used which is not available to an outside performance assessor. namely, in addition to the time-series of returns for the risk-free and risky assets, we know and use the portfolio proportions at each point in time in measuring portfolio performance. for example, if a portfolio is invested in equities for 85% of the months, then we know that its average beta is 0.85. 6.2. test results the mean and median one-month forecast errors for the s&p 500 returns for the two return prediction methods using the two types of s&p 500 forecasts for the two fiscal years are presented in table 3. as expected based on earlier results, the forecast returns are significantly higher for the bottom-up forecasts compared to the top-down forecasts. for the 136 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 bottom-up forecasts, the forecasted returns are significant (and higher than actual) for only fy1 using both thecurrent e/pandpremiumapproaches. for the top-down forecasts, the forecasted returns are significant (and higher than actual) only for fy1 using thecurrent e/p approach. four summary statistics for each set of 24 portfolios grouped by return prediction method are examined next. the portfolios within each set are differentiated by preferred habitat (four possibilities), type of consensus (three possibilities) and forecast horizon (two possibilities). the statistics include the mean and median performance metric, the number of switches, and the average beta over the 218-month period. we also emphasize the results for the lowest hurdle to the achievement of superior performance, that is, the case of an investor who has no preferred habitat and encounters no transaction costs in making switches. the summary results for the portfolios using thecurrent e/p and premium return prediction methods are presented in panels a and b of table 4, respectively. first, all of the mean performance metrics are not statistically significant, and all significant median performance metrics are negative. thus, these test results are consistent with the notion of sharpe (1975) that superior market timing is extremely difficult to achieve using the switching rules and signals tested to this point. second, even for the lowest hurdle to performance represented by the last (right-most) column in both panels of table 4, the mean performance is not significant. thus, even when no transaction costs are incurred for switching between equities and cash or vice versa, the portfolios formed using our market timing strategies exhibit a table 3 average monthly return forecast errors for the s&p 500 index for various return prediction methods, forward earnings forecasts and forecaster typea return prediction methodb statisticc top-down consensus forecasts of earnings per share by strategists for fiscal year bottom-up consensus forecasts of earnings per share by analysts for fiscal year fy1 fy2 fy1 fy2 mean 0.0206** 20.0031 0.0407** 0.0056 current e/p median 0.0144** 20.0023 0.0261** 0.0037 mean 20.0066 20.0054 0.0154* 0.0036 premium median 20.0019 20.0040 0.0127* 0.0028 this table reports the mean and median monthly return forecast errors, and tests of their significance, for various combinations of return prediction method, fiscal year end and type of forecasters. the return prediction methods are the current earnings-to-price (e/p) and premium approaches. the fiscal year forecasts are for the next year (fy1) and the following year (fy2). the forecaster type is top-down for forecasts made by market strategists, and bottom-up for forecasts made by financial analysts. a the forecast errors are equal to the predicted minus the actual return for the s&p 500 index for each of the months from january 1982 through february 2000. b the return prediction methods are thecurrent earnings-to-price or e/papproach, and thepremiumapproach. these two models are discussed at length in the body of this article. c tests are conducted to determine if the forecast errors are different from zero. * and ** indicate significance at the 0.05 and 0.01 levels, respectively. 137r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 table 4 summary statistics for performance, risk and activity for portfolios actively managed using the current e/p return prediction and premium approaches and consensus earnings forecasts panel a: various portfolio statistics for timing portfolios using current e/p return prediction approach type of consensus forecast of earnings per share used statistica natural habitate stockf cashf nonef noneg consensus top-down forecasts of earnings per share for fy1 by strategists meanb 20.0011 20.0020 20.0011 20.0020 medianb 20.0017 20.0041* 20.0024 20.0030 # of switchesc 8 9 6 10 betad 0.940 0.830 0.890 0.890 consensus top-down forecasts of earnings per share for fy2 by strategists mean 20.0001 0.0003 0.0001 0.0011 median 0.0006 20.0048** 0.0004 20.0039* # of switches 2 8 2 18 beta 0.995 0.491 0.973 0.794 consensus bottom-up forecasts of earnings per share for fy1 by analysts mean 20.0006 20.0009 20.0007 20.0006 median 20.0001 20.0007 20.0005 20.0003 # of switches 4 5 4 4 beta 0.973 0.950 0.968 0.954 consensus bottom-up forecasts of earnings per share for fy2 by analysts mean 20.0001 0.0004 20.0001 20.0001 median 0.0006 0.0004 0.0006 0.0006 # of switches 2 3 2 2 beta 0.995 0.977 0.995 0.995 consensus top-down forecasts of strategists and bottomup forecasts of analysts for earnings per share for fy1 (switches occur only if both screens are satisfied) mean 20.0001 20.0001 20.0001 20.0001 median 0.0006 20.0011 0.0006 0.0006 # of switches 2 1 2 2 beta 0.995 0.936 0.995 0.995 consensus top-down forecasts of strategists and bottomup forecasts of analysts for earnings per share for fy2 (switches occur only if both screens are satisfied) mean 20.0001 20.0001 20.0001 20.0001 median 0.0006 20.0011 0.0006 0.0006 # of switches 2 1 2 2 beta 0.995 0.936 0.995 0.995 panel b: various portfolio statistics for timing portfolios using premium return prediction approach type of consensus forecast of earnings per share used statistica natural habitata stockf cashf nonef noneg consensus top-down forecasts of earnings per share for fy1 by strategists meanb 0.0003 0 0.0002 20.0002 medianb 20.0059** 20.0050* 20.0051** 20.0055** # of switchesc 11 6 7 13 betad 0.560 0.467 0.478 0.528 consensus top-down forecasts of earnings per share for fy2 by strategists mean 0.0013 20.0004 0.0011 0.0011 median 20.0016 20.0047** 20.0038 20.0064** # of switches 5 4 3 11 beta 0.830 0.451 0.808 0.632 consensus bottom-up forecasts of earnings per share for fy1 by analysts mean 0.0006 0.0010 0.0008 0.0017 median 20.0074* 20.0060** 20.0064** 20.0067** # of switches 16 9 10 18 beta 0.720 0.577 0.621 0.654 (continued on next page) 138 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 performance that is not significantly different than a passive buy-and-hold strategy. third, both models result in a low rate of switching between an all-equity and an all-t-bill portfolio. specifically, only 2 of the 24 portfolios based on signals from thecurrent e/pmodel have 10 or more switches, and only 6 of the 24 portfolios based on signals from thepremium model have 10 or more switches. the maximum number of switches is 18 for both signal-switching approaches. this still represents an average holding period of slightly less than a year (i.e., 218 months divided by 19). fourth, all of the portfolios are invested in equities for more than 63% of the months, where 63% is the proportion of the months for which the returns on the s&p 500 index exceed the returns on t-bills for the 218 month period we study herein. this is easily seen if one recalls that the proportion of the months each portfolio is invested in equities is equal to its beta. table 4(continued) consensus bottom-up forecasts of earnings per share for fy2 by analysts mean 0.0001 0.0012 0.0001 0.0002 median 0.0025 20.0016 0.0025 0.0008 # of switches 0 4 0 6 beta 1 0.824 1 0.973 consensus top-down forecasts of strategists and bottomup forecasts of analysts for earnings per share for fy1 (switches occur only if both screens are satisfied) mean 0.0001 0.0018 0.0001 0.0010 median 0.0025 20.0015 0.0025 20.0014 # of switches 0 4 0 3 beta 1 0.819 1 0.868 consensus top-down forecasts of strategists and bottomup forecasts of analysts for earnings per share for fy2 (switches occur only if both screens are satisfied) mean 0.0001 20.0001 0.0001 0.0010 median 0.0025 20.0011 0.0025 20.0014 # of switches 0 1 0 3 beta 1 0.936 1 0.868 panel a provides the mean and median differential performance (and tests of their significance), average beta and number of switches for the portfolios that are actively managed using the current earnings-to-price (current e/p) return prediction approach and the consensus forecasts for earnings per share for the s&p 500 index for fiscal year fy1 (fy2) made by either strategists or analysts. panel b provides the same information except that the premium return prediction approach is used. a tests are conducted to determine if the mean and median performance metrics are significantly different from zero. b the mean and median market timing performance metric for month t is equal to the (excess return on the portfolio for month t)-(the excess return that would have been obtained on the portfolio in month t if the portfolio maintained its average actual beta at all points in time). c the number of switches is the number of switches from equity to cash or vice versa. d the average beta is the mean beta over the 218 month period for the current e/p return prediction approach, and over a 182 month period for the premium return prediction approach where the first three years are lost in calculating the return premium. e the preferred habitats are stock, cash and neither (or none). f switching rule reflects a transaction cost charge of 1% when moving away from this preferred habitat. g switching rule reflects no transaction cost charge when moving away from this preferred habitat. “*” and “**” indicate significance at the 0.05 and 0.01 levels, respectively. 139r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 7. tests of market timing strategies using consensus s&p 500 earnings forecast revisions we now test market timing performance using switching rules based on the direction of consensus forecast revisions and in the directional net number of forecasts revised up and down. the first switching rule,directional consensus,favors equity when the consensus forecast of earnings per share is revised up for this month compared to last, and vice versa if it is lowered. thus, for an upward directional consensus signal, the portfolio switches to equity if it is already in cash, and remains in equity if it is already in equity. similarly, for a downward directional consensus signal, the portfolio switches to cash if it is already in equity, and remains in cash if it is already in cash. the second switching rule,directional number,favors equity when the number of strategists revising their top-down forecasts of earnings per share upward this month exceeds the number revising them downwards, and vice versa if the number of revisions upwards is less than the number downwards. thus, for an upward directional number signal, the portfolio switches to equity if it is already in cash, and remains in equity if it is already in equity. similarly, for a downward directional number signal, the portfolio switches to cash if it is already in equity, and remains in cash if it is already in cash. the second switching rule is not implemented using the analyst bottom-up forecasts because the data for doing such are not available. the proportions of the periods classified bydirectional consensusanddirectional number are presented in table 5. given the forecast revision patterns identified in section five, it is not surprising that the proportion of down periods almost always exceeds that for up periods. four summary statistics for the portfolios using thedirectional consensusand directional numberswitching rules are presented in table 6. compared to the findings reported in the table 5 proportion of periods classified by directional consensus and the directional number of signals measure of revisions of earnings per share by strategists or analysts direction of revisions in earnings per share for the month top-down consensus forecasts of earnings per share by strategists for fiscal year bottom-up consensus forecasts of earnings per share by analysts for fiscal year fy1 fy2 fy1 fy2 directional consensus up 0.40 0.48 0.26 0.28 directional consensusa down 0.57 0.48 0.72 0.71 directional numberb up 0.33 0.39 directional numberb down 0.56 0.44 this table reports the proportion of periods classified by directional consensus and by directional number for fiscal years fy1 and fy2 for consensus forecasts by strategists and analysts. the directional consensus signal is based on the directional change (up or down) in the consensus forecast of earnings per share for the s&p 500 index. the directional number signal is based on the number of forecasts increased by strategists minus the number of forecasts decreased by strategists for earnings per share for the s&p 500 index for each month t. a the directional consensus is up (down) when the consensus forecasts of earnings per share are revised up (down) for this month compared to last month by strategists for top-down forecasts or by analysts for bottom-up forecasts. b the directional number is up (down) when the number of strategists revising their top-down forecasts of earnings per share upwards this month exceeds (is less than) the number revising them downwards. 140 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 table 6 summary statistics for performance, risk and activity for portfolios actively managed using the directional consensus and directional number signals for market timing type of forecast of earnings per share used statistica switching rule based on directional consensuse directional numberf top-down forecasts of earnings per share for fy1 by strategists meanb 20.0039 20.0020 medianb 20.0043** 20.0032** # of switchesc 57 31 betad 0.393 0.369 top-down forecasts of earnings per share for fy2 by strategists mean 20.0057** 20.0034 median 20.0055** 20.0048** # of switches 83 47 beta 0.471 0.456 bottom-up forecasts of earnings per share for fy1 by analysts mean 20.0025** median 20.0031** # of switches 56 beta 0.257 bottom-up forecasts for earnings per share for fy2 by analysts mean 20.0042 median 20.0036** # of switches 62 beta 0.282 top-down forecasts of earnings per share by strategists and bottom-up forecasts of earnings per share by analysts for fy1 (switches occur only if both screens are satisfied) mean 20.0007 median 20.0022** # of switches 11 beta 0.214 top-down forecasts of earnings per share by strategists and bottom-up forecasts of earnings per share by analysts for fy2 (switches occur only if both screens are satisfied) mean 20.0007 median 20.0022** # of switches 11 beta 0.214 panel a provides the mean and median differential performance (and tests of their significance), average beta and number of switches for the portfolios that are actively managed using the directional consensus and the directional number signals based on the earnings per share forecast for the s&p 500 index for fiscal years fy1 and fy2 by strategists and analysts. the directional consensus signal is based on the directional change (up or down) in the consensus forecast of earnings per share for the s&p 500 index. the directional number signal is based on the number of forecasts increased by strategists minus the number of forecasts decreased by strategists for earnings per share for the s&p 500 index for each month t. a tests are conducted to determine if the mean and median performance metrics are significantly different from zero. b the mean and median market timing performance metrics for month t is equal to the (excess return on the portfolio for month t) (the excess return that would have been obtained on the portfolio in month t if the portfolio maintained its average actual beta at all points in time). c the number of switches is the number of switches from equity to cash or vice versa. d the average beta is the mean beta over the 218 month period. e the directional consensus switching rule favors equity when the consensus forecasts of earnings per share are revised up for this month compared to last month, and vice versa if it is lowered. f the directional number switching rule favours equity when the number of strategists revising their top-down forecasts of earnings per share upwards this month exceeds the number revising them downwards, and vice versa if the number of revisions upwards is less than the number downwards. this signal is not available for analysts. * and ** indicate significance at the 0.05 and 0.01 levels, respectively. 141r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 previous section, the portfolios exhibit substantially more switches, and are invested in equities in less than 50% of the 218 months. this low proportion on being in equities is probably due to the downward revision in consensus earnings per share in time relative to the switching month. thus, not surprisingly, all of the median performance metrics are negative and significant, and all of the mean performance metrics are negative although only two are significant. once again, superior timing performance is not achieved! as a further test of robustness designed to reduce the influence of bull market months in our sample, we rerun all of the market timing tests reported in this section and the previous section using data for the shorter 167-month time period from january 1982 through november 1995. these results (unreported to conserve valuable journal space) are not materially different than those reported herein. 8. concluding remarks our findings have at least three important implications for individual investors (and other investment professionals). first, they show that individual investors should use the less optimistically biased forecasts of market earnings that are provided by strategists and not those obtained by aggregating the bottom-up forecasts of analysts. second, our findings suggest that individual investors could use the difference between bottom-up and top-down forecasts of market earnings to extract some information about the level of overoptimism in analyst forecasts. third, our findings provide additional support for the warning by sharpe (1975, p. 61) that only investors “with truly superior predictive ability should even attempt to time the market.” although some authors demonstrate that only a modest amount of information can lead to superior investment performance, the required informational advantage does not appear to be obtainable from the switching rules tested herein. these rules are based on consensus forecasts of market earnings supplied by strategists or analysts or both, on the direction of consensus forecast revisions, and on the directional number of such forecasts revised up and down. since we test the joint hypothesis that market timing is valuable and that it can be implemented using our timing signals and data inputs, whether or not the conclusions of this study are robust to the use of more refined switching rules using the same or similar data input remains for future study. acknowledgments financial support from fcar (fonds pour la formation de chercheurs et l’aide a` la recherche) and sshrc (social sciences and humanities research council of canada) are gratefully acknowledged. we appreciate the constructive comments received from the editor and two referees of this review, and the analyst and strategist data supplied by i/b/e/s inc. all remaining errors are the authors’ responsibility. 142 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 references ali, a., klein, a., & rosenfeld, j. (1992). analysts’ use of information about permanent and transitory earnings components in forecasting annual eps.the accounting review, 30(1), 183–198. beebower, g. l., & varikooty, a. p. (1991). measuring market timing strategies.financial analysts journal, 47(6), 78–92. bercel, a. (1994). consensus expectations and international equity returns.financial analysts journal, 50(4), 76–80. bierman, h. jr. (1995). bubbles, theory, and market timing.the journal of portfolio management, 22(1), 54–56. bodie, z., kane, a., & marcus, a. j. (1999).investments(fourth edition). new york: irwin mcgraw hill. breen, w., glosten, l. r., & jagannathan, r. (1989). economic significance of predictable variations in stock index returns.journal of finance, 44,1177–1189. brocato, j., & chandy, p. r. (1994). does market timing really work in the real world?the journal of portfolio management, 20(2), 39–44. brocato, j., & chandy, p. r. (1995). market timing can work in the real world: comment.journal of portfolio management, 21(3), 82–84. brous, p. a. (1992). common stock offerings and earnings expectations: a test of the release of unfavorable information.journal of finance, 47,1517–1536. brous, p. a., & kini, o. (1993). a reexamination of analysts earnings forecasts for takeover targets.journal of financial economics, 33,201–226. brown, l. d. (1996). analyst forecasting errors and their implications for security analysis: an alternative perspective.financial analysts journal, 52,40–47. busse, j. a. (1999). volatility timing in mutual funds: evidence from daily returns.the review of financial studies, 12(5), 1009–1041. carleton, w. t., chen, c. r., & steiner, t. l. (1998). optimism biases among brokerage and non-brokerage firm’s equity recommendations: agency costs in the investment industry.financial management, 27,17–30. chopra, v. k. (1998). why so much error in analysts earning forecasts?financial analysts journal, 54(6), 35–42. chung, r., & kryzanowski, l. (1999). accuracy of consensus expectations for top-down earnings per share forecasts for two s&p indexes.applied financial economics, 9,233–238. clarke, r. g., fitzgerald, m. t., berent, p., & statman, m. (1989). market timing with imperfect information. financial analysts journal, 45(6), 27–36. copeland, m. m., & copeland, t. e. (1999). market timing: style and size rotation using the vix.financial analysts journal, 55,73–81. das, s., levine, c. b., & sivaramakrishnan, k. (1998). earnings predictability and bias in analysts’ earnings forecasts.the accounting review, 73(2), 277–294. debondt, w. f. m., & thaler, r. m. (1990). do security analysts overreact?american economic review, 80, 52–57. dreman, d. n., & berry, m. a. (1995). analyst forecasting errors and their implications for security analysis. financial analysts journal, 51(3), 30–41. easterwood, j. c., & nutt, s. r. (1999). inefficiency in analysts’ earnings forecasts: systematic misreaction or systematic optimism?the journal of finance, 54(5), 1777–1797. elliott, j. a., philbrick, d. r., & wiedman, c. i. (1995). evidence from archival data on the relation between security analysts’ forecast errors and prior forecast revisions.contemporary accounting research, 11, 919–938. elton, e. j., & gruber, m. j. (1991). differential information and timing ability.journal of banking and finance, 15(1), 117–131. emanuelli, j. f., & pearson, r. g. (1994). using earnings estimates for global asset allocation.financial analyst journal, 50(2), 60–72. francis, j., & philbrick, d. r. (1993). analysts’ decisions as products of a multi-task environment.journal of accounting research, 31,216–230. 143r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 fama, e. f., & french, k. r. (1989). business conditions and expected returns on stocks and bonds.journal of financial economics, 25,23–49. ferson, w. e., & schadt, r. w. (1996). measuring fund strategy and performance in changing economic conditions.journal of finance, 51(2), 425–461. fuller, r. j., & kling, j. l. (1994). can regression-based models predict stock and bond returns?the journal of portfolio management, 20(3), 56–63. graham, b. (1954).the intelligent investor.new york: harper. graham, j. r., & harvey, c. r. (1994). market timing ability and volatility implied in investment newsletters’ asset allocation recommendations.working paper, 5597.national bureau of economic research. graham, j. r., & harvey, c. r. (1996). market timing ability and volatility implied in investment newsletters’ asset allocation recommendations.journal of financial economics, 42(3), 397–422. graham, j. r., & harvey, c. r. (1997). grading the performance of market-timing newsletters.financial analysts journal, 53(6), 54–66. hardy, d. c. (1990). market timing and international diversification.the journal of portfolio management, 16(4), 23–27. harris, r. d. f. (1999). the accuracy, bias and efficiency of analysts’ long run earnings growth forecasts.journal of business finance and accounting, 26,5/6, 725–755. jeffrey, r. h., & arnott, r. d. (1993). is your alpha big enough to cover its taxes?the journal of portfolio management, 19(3), 15–25. kang, s-h., o’brien, j., & sivaramakrishnan, k. (1994). analysts’ interim earnings forecasts: evidence on the forecasting process.journal of accounting research, 32,103–112. koco, l. (1995). variable annuity market timing spurs guidelines, national underwriter,january 23, 7 & 13. kryzanowski, l., lalancette, s., & to, m. c. (1997). performance attribution using an apt with prespecified macrofactors and time-varying risk premia and betas.journal of financial and quantitative analysis, 32(2), 183–204. larsen, g. a. jr., & wozniak, g. d. (1994/95). market timing for active asset allocation: a discrete regression approach.journal of applied business research, 11(1), 125–143. larsen, g. a. jr., & wozniak, g. d. (1995). market timing can work in the real world.the journal of portfolio management, 21(3), 74–81. lee, w. (1997). market timing and short-term interest rates.the journal of portfolio management, 23(3), 35–46. lin, h.-w., & mcnichols, m. f. (1998). underwriting relationships, analysts’ earnings forecasts and investment recommendations.journal of accounting and economics, 25(1), 101–127. prather, l., & bertin, w. j. (1998). the implication of discount rate changes for market timing.review of financial economics, 7(1), 21–33. pratt, t. (1993). wall street’s four-letter word.investment dealer’s digest,march, 29, 18–22. reichenstein, w., & rich, s. p. (1993). market timing for the individual investor: using the predictability of long-horizon stock returns to enhance portfolio performance.financial services review, 3(1), 29–43. reichenstein, w., & rich, s. p. (1994). predicting long-horizon stock returns: evidence and implications. financial analysts journal, 50(1), 73–76. reilly, f. r., & brown, k. c. (2000).investment analysis and portfolio management(sixth edition). new york: the dryden press. sharpe, w. f. (1975). likely gains from market timing.financial analysts journal, 31(2), 60–69. shilling, a. g. (1992). market timing: better than a buy-hold strategy.financial analysts journal, 48(2), 46–50. teoh, s. h., & wong, t. (1997). analysts’ credulity about reported earnings as overoptimism in new equity issues,working paper,university of michigan. wagner, j. c., shellans, s., & paul, r. (1992). market timing works where it matters most . . . in the real world. the journal of portfolio management, 18(4), 86–90. wagner, j. c. (1997). why market timing works.journal of investing, 6(2), 78–81. womack, k. l. (1996). do brokerage analysts’ recommendations have investment value?journal of finance, 51, 137–167. 144 r. chung, l. kryzanowski / financial services review 9 (2000) 125–144 pii: s1057-0810(99)00039-6 an empirical analysis of differences in black and white asset and liability combinations< yaw a. badua, kenneth n. danielsb, daniel p. salandrob,* adepartment of economics and finance, virginia state university, petersburg, va 23806, usa bdepartment of finance, virginia commonwealth university, richmond, va 23284-4000, usa received 20 july 1999; received in revised form 15 october 1999; accepted 29 october 1999 abstract this study analyzes data from the 1992 survey of consumer finances and finds significant differences in asset and liability combinations between black and white households. in addition, white households are identified as having significantly greater net worth and financial assets relative to black households. we are unable to show that the net worth of black households is constrained by barriers in financial markets. our study investigates how this difference in net worth could engender different financing decisions. we find that black households are significantly more risk averse in their choice of assets. further, we find that black households typically pay higher rates for several types of credit instruments, even though they self identify as conducting significantly more extensive searches in the financial markets. © 2000 elsevier science inc. all rights reserved. jel classification:d10; d31; j15 keywords:household behavior; wealth distribution; minority 1. introduction there is substantial evidence showing that blacks have made substantial economic gains during the past three decades. however, economic parity seems not to have been achieved < an earlier version of this paper was presented at the 1999 eastern finance association meetings. * corresponding author. tel.:11-804-828-7088; fax:11-804-828-3972. e:mail address:dpsaland@vcu.edu (d.p. salandro) financial services review 8 (1999) 129–147 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00039-6 and gross inequities between the races still exist. the fact is that blacks have made relatively significant economic gains if income is used as the measure of economic well-being. reports of income gains and a narrowing of the gap between blacks and whites seem, however, overshadowed by the deep-seated wealth inequity between the two races. indeed, browne (1993) questions the over-reliance on income as an index of economic well-being of american ethnic minorities. racial differences in net worth not only engender identifiable differences in financing decisions, but have long-term impacts on future wealth accumulation and the relative economic status of blacks and whites. using the 1992 survey of consumer finances data we attempt to reverify that there are significant differences in wealth as measured by net worth between the races. utilizing canonical correlation analysis we also examine the portfolio choice behavior of black and white households. it is hypothesized that households choose assets in an attempt to maximize the value of their portfolio and then attempt to finance these assets. the ability to maximize the value of the portfolio may be constrained for a number of reasons including the inability to finance chosen assets because of capital market imperfections, incomplete information, and inertia on the part of the households. in addition, as alluded to by brimmer (1991) there may exist a significant difference in the level of risk aversion between the two subgroups. taking all of these items into consideration the structure of a household portfolio is then determined by selecting those assets that can be financed in an ascending order of perceived contribution to wealth. using measures of risk aversion and yield differentials in the set of financial liabilities we attempt to determine if financial market constraints and inertia exist. the rest of the paper is organized as follows. section 2 provides a framework for factors that may affect the asset and liability combinations of households. section 3 discusses the data and methodologies. section 4 summarizes the empirical results. section 5 summarizes the main ideas of the paper along with some general remarks on public policy. 2. factors that affect asset and liability composition recent studies such as browne (1993), blau and graham (1990), oliver and shapiro (1995), and wolff (1994), have confirmed the existence of a considerable wealth gap between blacks and whites. according to brimmer (1991), blacks in 1988 represented 11.23% of us households, received 7.25% of money income, but held only 2.89% of household net wealth. oliver and shapiro report that the ratio of black-to-white median household income reached 0.62 in 1988 whereas the median net worth ratio stood at 0.08 and that white households possess nearly 103 as much mean net financial assets as black households. census data released recently also indicate that in 1993 white households had a median net worth of $45, 740 compared to $4, 418 for black households, a black-white ratio of 0.10. wolff’s study further shows that even among nonwhite and white families of the same income level, white families held considerably more wealth. racial differences in net worth not only engender identifiable differences in financing decisions, but have longterm impacts on future wealth accumulation and the relative economic status of blacks and 130 y.a. badu et al. / financial services review 8 (1999) 129–147 whites. brimmer (1991) alluded to these differences when (he observed that black wealth was concentrated more heavily in property than in other households and that black investors stressed safety over more risky types of investments. oliver and shapiro argue that the different decision patterns may explain why the wealth gap between blacks and whites is likely to increase. we attempt to provide a framework for factors that affect the portfolio decisions, and therefore the asset and liability combinations of households. we find in the literature a patchwork of compelling arguments that identify key factors that affect the asset and liability combinations of households. recently, gutter, fox and montalto (1999) incorporated socioeconomic, financial, and attitudinal factors into the life-cycle savings model to determine portfolio choice. their findings are that the socioeconomic factors have a greater effect than race in observed differences in portfolio choices. household portfolio decisions are made such that households attempt to maximize the value of their portfolio subject to market financing constraints. households do not have the flexibility of financing that is afforded the corporate sector, that is, many assets are only capable of being financed by particular types of liabilities. thus, if a particular type of liability is not provided to the household, the asset that it sought to finance cannot be obtained, or it must be financed with the next, probably more costly liability. for example, one would expect that cars would be financed with installment loans, given the relatively favorable rate of the installment loan to comparable alternatives, and the tendency of the installment loan maturity to match the life of the asset, which reduces interest rate risk. however, it is not uncommon in the us to find households that finance cars on credit with interest rates as high as 36%. transouth financial is a us financial corporation that has taken advantage of this market condition (tlpj, 1999). the flexibility of financing afforded the household is only one factor that may be impacting these types of financing decisions. other factors include transaction costs, tax considerations, incomplete information, liquidity constraints, heterogeneous risk aversion, and household inertia. asymmetric information problems play a much larger role in this scenario than in the corporate setting. the availability of financing in the household sector is subject to the provider’s ability to correctly process costly obtained information submitted by the household that most certainly involves the problem of moral hazard and adverse selection. in this context, households applying for funding may overstate their ability to meet the repayment schedule of the liability in their applications for financing (moral hazard). in addition, the problem of adverse selection may also be observed in certain types of liability acquisition processes. rates charged high default risk households may be such that households on the low end of this risk category may opt out of this type of financing causing average risk to increase. from the financing providers point of view it is the least desirable customers who apply for this type of liability. this may cause the availability of particular asset and liability combinations to be correlated with the demographics of the households. for example, housing assets and mortgage financing may only be available to households whose head is relatively older with an established credit history attempting to purchase housing in particular areas. paxson (1990) shows that borrowing constraints in personal loan markets may affect 131y.a. badu et al. / financial services review 8 (1999) 129–147 portfolio choices as well. in this analysis exogenous and endogenous borrowing ceilings exist. exogenous borrowing ceilings are unaffected by individual portfolio choices and unambiguously lead to more liquid portfolios. endogenous borrowing ceilings allow for maximum loanable amounts that are positively correlated to illiquid assets (collateral) held by the household. thus, a household facing stricter borrowing ceilings may be less likely to put assets into such things as housing and other assets that cannot be easily converted to finance consumption in case of an income shortfall. the amount of risk a household is willing to face will affect the choice of assets made by that household. friend and blume (1975) provided an expected utility based model that shows that an investor’s portfolio decision is a function of the dollar amount of wealth, individual risk preferences and the allocation of wealth among risky assets. schooley and worden (1995) use this expected utility based model to show that portfolio allocations are a reliable indicator of a household’s relative level of risk taking and attitudes toward risk. they also find investment in risky assets is significantly related to the household’s demographic profile. in a related study, bajtelsmit, bernasek, and jianakoplos (1999) find that women exhibit a greater relative risk aversion in their asset allocation that affects the allocation of household wealth to defined contribution pensions. if particular population subgroups tend to be more risk averse in their asset and liability choices, their long-term wealth will be expected to be less than households who choose to be less risk averse. for instance, households that preclude stocks from their portfolio will undoubtedly have less wealth over the long run, because the long term return on common stock investments are higher than the long term return on less volatile investment opportunities. cultural influences associated with race, gender or life-cycle stage may also influence portfolio decisions after controlling for economic characteristics (haliassos and bertaut 1995; bajtelsmit, bernasek, and jianakoplos 1999). this market imperfection or friction caused by inertial factors is usually ignored in most portfolio selection models. a cultural fear of being in debt may inhibit certain population groups from seeking and accepting a mortgage contract or undertaking student loans. the acquisition costs of information (in both monetary and temporal terms) for certain types of investment opportunities may preclude portions of the population from considering them. tendencies toward more visceral based investments, that is, property, and avoidance of financial investments are explained by inertia as well. blau and graham (1990) infer that a significant wealth gap remains between households headed by whites and blacks due to constraints placed on the equity accumulation of black households. if differences in net worth and the constraints faced by households exist, blacks and whites will tend to make significantly different financial decisions. in this scenario each household uses different sources of costly information and processes this information against different financial constraints. this may cause the availability of particular assets and liability combinations to be correlated with the race of the household. for example, as found in brimmer (1991), black households tend to prefer liquidity and tend to concentrate in investment property, whereas white households tend to prefer financial assets and concentrate in single family dwellings. 132 y.a. badu et al. / financial services review 8 (1999) 129–147 3. data and methodologies the source of data for the current study is the 1992 survey of consumer finances (scf). the scf database is a comprehensive survey investigating household demographics and financial information. the results of this survey provide a unique opportunity for an in depth analysis of contemporary household portfolio selection practices of a large cross-section of consumers (households). along with the 2, 456 households selected by standard methods to achieve a sample including responses from all 48 contiguous states, 1, 450 households were “oversampled” (added) from tax data to insure a representative income distribution. the resulting sample of 3, 906 respondents thus reduced the skewed distribution of having only a few wealthy families with a ‘disproportionately large share of income and wealth’ in the first group of 2, 456. also, nonresponse errors arising from some interviewees’ failure to answer specific questions were imputed. the imputation process increased the number of observations to 19, 530. we reduced this total amount to 17, 528 observations after eliminating all households whose head is neither white or black. the data set provides a vast array of household balance sheet items for the respondents (survey of consumer finances 1992 manual). descriptive statistics are examined for statistically significant differences between mean and median values across groups to determine if differences exist between the races in net worth and components of net worth. in addition, net worth and income distributions are examined across quartiles for each income group. we incorporate techniques similar to wang (1995) in our investigation of differences in income and net worth across the races. first, the log of actual income is regressed on a set of demographic variables using eq. (1). next, expected income is estimated as an instrumental variable. ln~inct! 5 b1 1 b2~ age! 1 b3~ age!2 1 b4~educ! 1 b5~semploy! 1 b6~race! 1 b1~marital! 1 b8~sex! 1 b9~work! 1 b10~children! 1 b11~health! 1 b12~pension! 1 nt (1) where inct, is actual income, equal to the observed annual income of all household members from all sources. age is the age of the head of households rounded to the nearest year, educ is a dummy variable equal to one if the education level is equal to twelve years or more. semploy is a dummy variable equal to one if the household head is self employed, race is a dummy variable equal to one if the head of household identifies as white and zero otherwise. marital is a dummy variable equal to one if the respondent identified as being married, pension is a dummy variable equal to one if the respondent belonged to a pension plan, sex is a dummy variable equal to one if male. children is the number of dependent children the respondent identified, health is a dummy variable equal to one if the respondent identified themselves as having excellent, good, or fair health and zero otherwise. work is a dummy variable equal to one if the head of household has five or more years of work. the expected income of each household, xinc, is calculated based on their set of demographic variables by eq. (2): 133y.a. badu et al. / financial services review 8 (1999) 129–147 ln~xinct! 5 b1 1 b2~ age! 1 b3~ age!2 1 b4~educ! 1 b5~semploy! 1 b6~race! 1 b1~marital! 1 b8~sex! 1 b9~work! 1 b10~children! 1 b11~health! 1 b12~pension! (2) a statistically significant coefficient on the race variable can be interpreted as evidence of a racial difference in magnitudes of expected incomes. however, as pointed out by gutter, fox, and montalto (1990), this may be too simplistic an approach. we address this problem below by using the jackson and lindley (1989) techniques. economic theory provides a well-developed framework for modeling net worth accumulation based on the life cycle hypothesis (ando and modigliani, 1963; wolff, 1994; wang, 1995). this body of work determines that net worth is a function of expected income, age, and a set of households demographic variables. following wang (1995) we have developed a model to explain household nonhuman wealth, where wt is defined as household nonhuman wealth and it equals the summation of past savings, st. we use the net worth of each household as a proxy for the nonhuman wealth and therefore assume that past savings includes financial and physical assets. based on this theoretical work we model net worth using the following equation: ln~wt! 5 g1 1 g2 ln~xinc! 1 g3ln~xinc!2 1 g4 ln~tinc! 1 g5~ age! 1 g6~ age!2 1 g7~educ! 1 g8~semploy! 1 g9~race! 1 g10~marital! 1 g11~sex! 1 g12~work! 1 g13~children! 1 g14~health! 1 g15~pension! 1 mt (3) where tinc is transitory income, equal to the difference between actual income and expected income. similar to bajtelsmit, bernasek and jianakoplos (1999) we limit our sample to those households with positive net worth. we correct for this possible sample selection bias using heckman’s procedure (heckman, 1979). once again if a significant coefficient for the race variable is found this provides weak support for racial differences in net worth. we attempt to determine the source of these differences in net worth accumulation between the races. one possible reason why these differences are found is that barriers may exists in financial markets. in an effort to determine if the net worth of black households is constrained by barriers in the financial markets, we employ a technique very similar to that developed by jackson and lindley (1989) and recently utilized by gutter, fox, and montalto (1999). we create a pooled data set by combining the black and white samples. also, several interaction variables are created with the race dummy variable by multiplying each explanatory variable by this dummy variable. the pooled data set then includes the k explanatory variables and k dummy interaction variables. this methodology is a behavioral model that estimates the net worth of various groups in the study. a set of coefficients for the white and black populations are estimated. the expected net worth for blacks (wb) is estimated using the ‘black’ coefficients and ‘black’ explanatory variables. then expected net worth in the absence of constraints (wh) is estimated for blacks using ‘white’ coefficients and ‘black’ 134 y.a. badu et al. / financial services review 8 (1999) 129–147 explanatory variables. the extent of the constraints experienced by black households is measured by (wh-wb) which is decomposed into the constant and coefficient effects. the constant effect is that portion of the residual or difference that cannot be accounted for by differential endowments or differential responses. the coefficient effect is a measure of the differential between group response in the dependent variable to unit changes in the independent variables. this testing method is really a joint test of the two components of the residual differences, that is, the constant and the coefficient effects that are easily determined from the pooled model. if the net worth of the black households is constrained due to barriers in the financial markets, the constant and coefficient effects must be statistically significant and have the proper sign. given that differences exist across the races in levels of net worth we should observe different combinations of assets and liabilities used by the races in household portfolio construction. ranking of asset and liability combinations can be done naturally by employing canonical correlation analysis. canonical correlation analysis can find linear combinations of variables from both sides of the balance sheet such that they are maximally correlated. the canonical variates derived from the canonical correlation analysis are ranked according to significance and are linearly independent from each other. the ranking of the canonical variates allows the ordering preference of the household sector to be revealed because the analysis identifies, in an ascending order, the asset and liability combinations that maximize the value of the household portfolio. canonical correlation analysis is not a new statistical technique but it has not received a great deal of attention in the finance literature. there are however, a few studies that employ canonical correlation analysis to analyze balance sheet data across different financial entities. two of the most recent studies in this vein are van auken, doran, and yoon (1993) who compare korean management techniques to their american counterparts for 45 small-tomedium-sized korean firms and adams (1995) that investigates the balance sheet structure and tests the managerial-discretion hypothesis using data on a cross-section of 33 new zealand life insurance companies. canonical correlation analysis is performed on two subsamples consisting of households whose head identified as white and heads who identified as black in an effort to investigate whether particular demographic characteristics have an effect on household portfolios. the number of observations in these samples is 15,738 and 1,790, respectively. for each sample of households, six asset variables, five liability variables and the age of the head of households are used in the canonical correlation analysis. furthermore, to standardize for size differences all of the asset and liability variables are divided by the households net worth in the canonical correlation analysis. different asset and liability combinations may be caused by individual choices. a driving force in an individual’s choice is that individual’s level of risk aversion. classic financial theory predicts that the higher the level of systematic risk accepted the higher the expected return. friend and blume (1975) and more recently schooley and worden (1996) use a measure of risky assets to net worth to obtain a measure of risk aversion for households. schooley and worden (1996) demonstrate that household portfolio allocations are reliable indicators of attitudes toward risk. this implies that household asset allocations can be utilized to reveal the households relative level of risk taking. we employ a similar analysis 135y.a. badu et al. / financial services review 8 (1999) 129–147 to determine the ’risk aversion’ coefficient for the households in our sample. several different risk aversion coefficients are investigated. first, the least restrictive measure uses a household’s assets in common stock, real estate other than primary and secondary residents, business investments, mutual fund investments and quasi-liquid retirement funds as a ratio of net worth. the second measure excludes the quasi-liquid retirement funds from the risky portfolio. the reason this asset is excluded is that with the data set utilized here it is impossible to determine the composition of these funds and their liquidity lowers their relative risk. the third, and most restrictive measure of risk tolerance includes only common stock, real estate and business investments in the risky asset portfolio. these risk tolerance measures are then calculated for each racial group and compared. other possible sources of differences in portfolio formation strategies and net worth accumulation are a household’s self-imposed liquidity constraints (attitudes toward debt), level of inertia in financial market activity, and level and utilization of information concerning the financial markets. the scf’s structure lends it self to several types of investigations into these particular areas. standard nonparametric statistical tests are applied to several different categories of data. we examine whether a yield differential exists between the races for similar liabilities. we compare percentage yields for first mortgages and a list of installment credits available to households in this survey. we attempt to determine the households attitude toward debt by examining household responses to specific questions posed by the survey. these questions ask the head of the households: “how do you feel about credit?;” and “do you think it is a good idea to borrow for vacations, living expenses, fur coats, cars or educational expenses?” we then investigate whether a household has been turned down for a loan in the past five years and whether the household conducted an active search of the financial markets for the best rate available. 4. empirical results definitions of all variables appear in table 1. descriptive statistics of all the portfolio variables for the two race subsamples are presented in table 2. the data show that households headed by whites had significantly larger mean asset and liability values than do households headed by blacks for every category. indeed, these inequalities are much more dramatic in the asset categories, with the mean value for vehicles for blacks is 18.9% of the mean value for whites and home sites for blacks is 14.5% of the mean value for whites. all other assets comparisons show that blacks have less than 3% of the asset totals of whites. the comparisons of mean liabilities show that blacks have 45% of the mean value of credit card balances, 32% of the mean value of installment credit other than mortgages, and 26% of the mean value of mortgages using principal residence as collateral. it is interesting to note credit card balances and installment credit other than mortgages are typically the more expensive methods of incurring liabilities that households face. this may indicate that not only do black households have smaller asset portfolios but have less access to liability markets than households headed by whites. the head of households are significantly older for white households than for the black households surveyed with mean ages of 51.5 years and 45.9 years, respectively. table 3 presents distributions of net 136 y.a. badu et al. / financial services review 8 (1999) 129–147 worth and income by quartiles across racial groups. the results for the net worth distribution is very similar to those of brimmer (1991), wolff (1994), and oliver and shapiro (1995). panel a shows that although blacks make up only eleven percentage of the sample, they are over represented in the lower quartiles of net worth. approximately twenty-three percentage of the lowest quartile of net worth is made up of households whose heads are black. panel b present similar finding for income, although blacks have a slightly higher representation in the top two income quartiles than the group had for net worth. in panel c, the data show that slightly more than 56% of the households headed by blacks are in the lowest quartile of net worth. indeed, over 86% of all black households are found in the two lowest quartiles of net worth. panel d present findings for income earned in the last year by heads of households, a slightly larger percentage of black households are in the top two quartiles of incomes. we investigate further the significant differences across race for income and net worth table 1 variable definitions liq: includes any dollar amount in checking accounts, money markets and, savings accounts and any nonrealized capital loss of stock value. fin: includes any dollar amount in checking accounts, money markets and, savings accounts and any nonrealized capital loss of stock value plus dollar value of certificates of deposit, market value of common stock, quasi-liquid retirement accounts that can be borrowed against, directly owed mutual funds, directly held bonds, managed assets such as trusts, annuities and managed investment accounts, cash value of whole life insurance, savings bonds and other financial assets such as money owed to the family, cash held or deferred compensation. vehic: dollar value of all vehicles, including motor homes, rv’s, boats and airplanes, owned by the household. houses: dollar value of home site only, mobile home, both site and mobile home, home and land, apartment or property, and farm or ranch property unrelated to business. realest: dollar value of real estate sold for which the seller provided the loan, including accepting a note, land contract or mortgage plus the percentage owned of such properties as time share, apartment building any business property. bus: net worth of any privately-held business, farm, professional practice or partnership owned by the household plus the dollar amount owed to the household by the business minus any amount owed to the business by the household plus the amount of the business guaranteed or collateralized. mrthel: dollar value of the amount owed by the households for any mortgage/land contract ect., which uses the principle dwelling as collateral plus any amount borrowed against home equity credit lines. realdbt: dollar amount still owed by the household for real estate sold and financed by the household plus any dollar amount still owed for any real estate such as lots, vacation homes, time shares, apartment buildings, commercial property or other investment property. ccbal: dollar value of the balance owed on credit cards (bank type, store cards, gasoline cards etc.) install: dollar value owed in installments for such items as household appliances, furniture, hobby or recreational equipment, educational expenses, medical expenses, home improvements, and vehicles plus any debts owed to friends or relatives. odebt: dollar value of amount borrowed against financial assets such as life insurance policies and pension plans. age: 1992 minus the year of the head of households birth. networth: (fin1house1vehic1realest1bus1other non-financial assets such as antiques, rare books jewelry, etc.) minus (mrthel1realdbt1ccbal1install1odebt plus amounts owed on credit line accounts) 137y.a. badu et al. / financial services review 8 (1999) 129–147 t ab le 2 d es cr ip tiv e st at is tic s (v al ue s ar e sh ow n in do lla rs , ex ce pt fo r a g e w hi ch is in ye ar s) v ar ia bl es f in li q v e h ic h o u s e s r e a le s t b u s m r t h e l r e a ld b t c c b a l in s t a ll o d e b t a g e p an el a : w hi te on ly (n 5 1 5 7 3 8 ) m ea n 13 64 62 2 12 76 00 .1 29 84 1. 4 26 70 63 .4 12 42 50 5 24 68 92 1 51 67 5. 5 31 36 27 16 37 .5 86 76 .4 10 03 6. 3 51 .5 m ed ia n 48 58 5 63 00 99 00 95 00 0 0 0 0 0 50 0 0 0 50 s ta nd ar d d ev ia tio n 68 39 95 4 66 04 41 .7 18 11 11 .2 68 43 0. 9 26 30 00 00 17 70 00 00 12 65 07 .2 68 51 12 4 38 31 .3 10 08 17 .6 11 75 11 .5 16 .7 p an el b : b la ck on ly (n 5 17 90 ) m ea n 24 99 0. 9 36 14 .3 56 43 .6 38 84 8. 9 33 63 9. 5 21 05 5. 2 13 50 6. 3 80 92 .4 74 1. 9 27 78 .2 44 0. 8 45 .9 m ed ia n 12 35 20 0 25 00 0 0 0 0 0 0 0 0 43 s ta nd ar d de vi at io n 14 40 17 .7 12 91 4. 19 82 60 .9 81 15 0. 6 28 32 33 .1 13 31 47 .4 35 15 3. 6 11 30 94 .3 19 62 .9 56 17 .4 25 02 .9 16 .3 d iff er en ce of m ea ns tte st t 5 2 2 4 .5 2t 5 2 2 3 .5 1t 5 2 1 6 .6 t 5 2 3 9 .5 t 5 2 5 .7 7t 5 2 1 7 .3 t 5 2 2 9 .2 t 5 5 .5 9 t 5 2 1 6 .1 t 5 2 7 .2 4t 5 2 1 0 .2 t 5 2 1 3 .7 4 w ilc ox on ra nk su m te st z 5 2 4 1 .0 7 z 5 2 4 2 .2 z 5 2 3 2 .5 z 5 2 3 2 .5 z 5 2 2 0 .7 z 5 2 1 5 .7 z 5 1 3 .7 z 5 2 1 1 .3 z 5 2 2 2 .9 z 5 0 .6 9 z 5 2 0 .4 3z 5 2 1 3 .6 6 b la ck /w hi te ra tio m ea ns 0. 01 8 0. 02 8 0. 18 9 0. 14 5 0. 02 7 0. 00 9 0. 26 1 0. 02 6 0. 45 3 0. 32 0. 04 4 2 0. 43 138 y.a. badu et al. / financial services review 8 (1999) 129–147 using a two-step procedure. first expected income is estimated and then used as an instrumental variable in the net worth regression. table 4 presents the results of the income and net worth regressions. all significant coefficients have the expected sign. we find that the income is significantly and positively related to age, education, work experience, marital status, pension status, self-employment status, race, and the health condition of the respondent. the race dummy variable impacts income and net worth in a significant and positive manner, which indicates that there is a significant difference between the races. however, this does not provide a formal test for significant differences between the net worth of the two groups. by using the jackson and lindley (1989) approach we estimate five separate net worth regressions using the white sample, the black sample, the pooled data sample with no table 3 net worth and income distribution by race panel a: number of households in each quartile of net worth. numbers in parenthesis indicates the percentage of the net worth quartile made up of that racial group. total white black net worth, $21 290 4 382 3 379 1 003 (0.7711) (0.2289) $21 290# net worth, $133 900 4 377 3 833 544 (0.8757) (0.1243) $133 900# net worth, $944 125 4 387 4 178 209 (0.9545) (0.0478) net worth$ $944 125 4 382 4 348 34 (0.9922) (0.0078) total 17 528 15 738 1 790 panel b: number of households in each quartile of income. numbers in parenthesis indicates the percentage of the income quartile made up of that racial group. income# $19 000 4 288 3 337 951 (0.7782) (0.2218) $19 000# income, $42 000 4 304 3 780 516 (0.8801) (0.1199) $42 000# income, $114 000 4 540 4 259 281 (0.9381) (0.0619) income$ $114 000 4 396 4 354 42 (0.9909) (0.0096) total 17 578 15 738 1 790 panel c: percentage of total population (17 578) or racial group in a particular net worth quartile. net worth, $21 290 0.2500 0.2147 0.5603 $21 290# net worth, $133 900 0.2497 0.2435 0.3039 $133 900# net worth, $944 125 0.2503 0.2655 0.1168 net worth$ $944 125 0.2500 0.2762 0.0190 panel d: percentage of total population (17,578) or racial group in a particular income quartile. income, $19 000 0.2446 0.2120 0.5313 $19 000# income, $42 000 0.2456 0.2407 0.2883 $42 000# income, $114 000 0.2490 0.2706 0.1570 income$ $114 000 0.2508 0.2767 0.0234 139y.a. badu et al. / financial services review 8 (1999) 129–147 dummy or interaction terms, the pooled data set with a race dummy and finally the pooled data set with a race dummy and interaction terms. table 5 presents the result of the jackson and lindley analysis. we have not reported all of the results of the net worth regressions but they are available upon request from the authors. the finding of particular interest is the constant effect, which is equal to25.1142. we interpret the significant negative constant term as an indicator of lower mean net worth for whites due to the effect of the race dummy variable. we cannot confirm that the net worth of black households is constrained because the constant effect has the wrong sign. the coefficient effect of 5.75872 is the primary reason why the significant residual difference exists. however, this result implies that black table 4 income and net worth regression analysis income regression eq. (1) net worth regression eq. (3) independent variables coefficient (t-statistics) independent variables coefficient (t-statistics) intercept 6.12 intercept 253.68 (59.35)a (211.69)a age 0.05 ln(xinc) 10.31 (13.09)a (10.63)a (age)2 20.0003 ln(xinc)t) 2 20.53 (27.02)a (211.29)a educ 0.91 ln(tinct) 0.92 (31.16)a (48.28)a semploy 0.90 age 0.25 (35.63)a (13.46)a race 0.29 (age)2 20.001 (11.48)a (29.66)a marital 0.59 educ 1.73 (21.23)a (20.46)a sex 0.43 semploy 2.34 (13.18)a (11.09)a work 0.31 race 1.30 (11.59)a (12.20)a children 0.001 marital 0.64 (0.18) (6.29)a health 0.56 sex 0.82 (12.00) (9.52)a pension 0.44 work 0.46 (17.92)a (4.07)a children 20.07 (24.27)a health 0.52 (2.70)b pension 1.01 (7.53)a 9r29 5 0.372 9r29 5 0.481 f-value5 1048.95 f-value5 1290.78 a indicates significantly different from zero at the 5% level. b indicates significantly different from zero at the 10% level. 140 y.a. badu et al. / financial services review 8 (1999) 129–147 households have significantly lower net worth due to their response to the independent variables. we believe that the significant differences in net worth across races should engender different portfolio decisions, which we investigate with canonical correlation analysis. the expected different portfolio decisions must be evident in their impacts on the respective balance sheets and composition. we use six possible canonical roots given the representation of the problem presented in this study. in general, subjective judgment must be exercised to determine which of the canonical loadings warrant interpretation. however, in this particular analysis we keep only the roots whose correlation is at least 63% and in most cases above 90%. the second column of table 6 provides the significant roots for each subsample. the first four roots of the white-only samples and the first three roots of the black-only sample provide a correlation of at least 95%, all other correlation’s are well below 50%. thus, the parameters for the first four roots of the white-only sample and the first three roots of the black-only sample will be analyzed and are detailed in panels a, and b of table 6. the results of the first loading set for the white only sample reveal that the strongest and therefore most prominent relationship exists between the asset vehicles and the liability installment credit other then mortgages. in addition, the holdings of financial assets are significant in the first loading and independent of any liabilities indicating that a significant portion of the net worth of the white only sample is concentrated in this asset. the significance of the independence of these financial assets from any liabilities is that households headed by whites have tangible net worth that provides a wealth effect in their use of financial assets and liabilities, as well as their spending on real assets. the most prominent relationship found in the second loading set occurs between liquid assets and credit card balances. this indicates that white households place a high value on liquidity and use that liquidity to support their credit card balances. the next significant asset and liability combinations are home sites and mortgages with primary residences as collateral followed by the dollar value of investment real estate sold and dollars still owed for real estate sold and financed by the households, respectively, for the white households. this indicates that principal residence is seen as the third most important portfolio asset for households headed by whites and that it is financed with mortgages. the fourth loading show the asset and table 5 jackson and lindley analysis white model black model pooled no race dummy pooled race dummy pooled race and interaction r2 0.5298 0.3611 0.6289 0.6303 0.6320 se 1.09 1.5147 0.7264 0.7251 0.7233 sse 18514.73 4077.0396 8699.397 8666.65 8619.53 n 15045 1454 16499 16499 16499 mean log net worth ww 5 5.329295 total effect5 1.54692 constant effect5 25.1142 wh 5 4.426896 endowment effect5 0.90240 coefficient effect5 0.64452 wb 5 3.782373 residual effect5 0.64452 f-value for residual effect5 13.8799 f-value for coefficient effect5 9.00851 141y.a. badu et al. / financial services review 8 (1999) 129–147 table 6 canonical analysis resultsa root (canonical correlation) assets liabilities panel a: 1st vehic fin install (0.999) (0.993) (0.987) (0.993) white only sample 2nd liq ccbal (0.997) (0.706) (0.915) (n 5 15738) 3rd houses mrthel (0.992) (0.843) (0.848) 4th realest realdbt (0.990) (0.838) (0.836) panel b: 1st vehic liq install ccbal black only sample (0.999) (0.921 (0.913) (0.995) (0.918) (n 5 1790) 2nd realest realdbt (0.985) (0.976) (0.975) 3rd houses mrthel (0.956) (0.949) (0.943) panel c: 1st houses mrthel white (0.997) (0.967) (0.965) 35 and under 2nd liq install (n 5 3298) (0.977) (0.823) (0.906) 3rd realest realdbt (0.941) (0.940) (0.939) panel d: 1st vehic liq install ccbal (0.999) (0.925) (0.919) (0.998) (0.942) black 2nd realest realdbt 35 and under (0.996) (0.989) (0.989) (n 5 550) 3rd houses mrthel (0.940) (0.938) (0.936) 4th fin ccbal (0.710) (0.285) (20.227) panel e: 1st vehic fin install (0.999) (0.996) (0.955) (0.996) white 2nd liq ccbal 36 to 55 years old (0.999) (0.714) (0.920) (n 5 6221) 3rd realest realdbt (0.996) (0.960) (0.962) 4th houses mrthel (0.989) (0.919) (0.907) panel f: 1st houses mrthel black (0.993) (0.891) (0.903) 36 to 55 years old 2nd vehic install (n 5 775) (0.923) (0.687) (0.628) 3rd realest realdbt (0.634) (0.626) (0.630) panel g: 1st houses vehic mrthel install (0.999) (0.964) (0.841) (0.911) (0.895) white 2nd realest realdbt 55 or older (0.933) (0.912) (0.916) (n 5 6219) 3rd liq install (0.736) (0.3941) (0.295) 142 y.a. badu et al. / financial services review 8 (1999) 129–147 liability pairs of the dollar value of investment real estate sold and dollars still owed for real estate sold and financed by the households as the next most important portfolio items for the white households. panel b of table 6 presents the results of the black-only households. like the first sample the most prominent asset and liability combination exists among vehicles and installment credit other than mortgages. also significant in the first root is the asset and liability combination between liquid assets and credit card balances. this finding indicates a heavy reliance on the use of credit cards by black households, which denotes the different usage of the financial system when compared to the white households. the second most prominent loading was between the dollar value of investment real estate sold and dollars still owed for real estate sold and financed by the households, this despite the fact that fewer than ten percent of black households possess positive amounts of investment real estate. the third most prominent combination is between home sites and mortgages using primary residences as collateral. in addition, it is important to note that black households, on average, do not have any assets that are independent of any liabilities, which supports the claim that black households have lower net worth when compared to white households. this finding is supported by the work of wolff (1994), who studied the accumulation of household wealth over the life-cycle by race. wolff concluded from his study that the life-cycle hypothesis is relevant for educated, white, urban, middle-class households only. we believe that our results follow this general pattern. in order to infer the role that life-cycle plays in the balance sheet composition of black and white households, we divided each race subsample by three age profiles. the three age profiles are 35 and under, 36 to 55 years, and 55 and over. table 6 panels c through h presents the results of the canonical analysis conducted on the various age and race subsamples. among some of the results, panel e clearly shows that the white-only 36 to 55 years subsample, has accumulated significantly different net worth than the 35 and under age group or the 55 and over age group. this group’s primary asset and liability combination is vehicles and installment credit other than mortgages, respectively. further, this group has financial assets that are unencumbered by liabilities. this pattern differs from the other white age groups in that these other groups exhibit primary asset and liability combinations in primary residence and mortgages. this primacy of the residence and mortgage as asset and liability combination is also found in the black subgroups except for the youngest group. similarly, the black-only households for ages 35 and under clearly table 6(continued) root (canonical correlation) assets liabilities panel h: 1st houses mrthel black (0.983) (0.927) (0.905) 55 or older 2nd realest realdbt (n 5 465) (0.745) (0.683) (0.662) 3rd vehic install (0.652) (0.558) (0.464) a the most prominent asset and liability combination are presented for each significant root. in addition, any loading for asset and/or liability greater than 0.8 is also presented. 143y.a. badu et al. / financial services review 8 (1999) 129–147 display a heavy reliance on credit card debt, whereas the other two black subsamples do not display significant credit card usage at all. clearly age has an impact on the balance sheet composition of the household. this may also demonstrate that liability choices of the youngest black households are constrained to the more expensive alternatives. in addition, when we compare the 36 to 55 year subsample for each race, we identify two significantly different groups. although, the black-only households in the 36 to 55 year subsample have a healthy balance sheet structure, the lack of any independent financial assets, that is, assets unaccompanied by matching liabilities demonstrates a marked difference in asset accumulation between the races. the healthy balance sheet structure coupled with independent financial assets of the white-only households in the 36 to 55 subsample may indicate that white households engage in selection of riskier investment vehicles that over time offer a higher expected return. it is clear that there are significant portfolio differences across the racial groups. we focus our attention on individual choices as the reason for portfolio differences between black and white houesholds. table 7 presents the results for the risk tolerance measure calculations. employing several different measures, which vary in degree of restrictiveness, for risk tolerance we find that significant differences exist between the two groups. in all measures the white group was found to be more risk tolerant than the black group. in addition, as stated before we find the black group to be significantly younger than the white group on average. morin and suarez (1983) find that all other things equal risk tolerance decreases with age. thus, in spite of the fact that the black group is younger it is significantly less risk tolerant in its choice of assets and liabilities. the data indicate that households headed by blacks acquire less risky asset combinations and therefore according to classical financial theory should be expected to have lower net worth in the long term. table 7 descriptive statistics for various risk aversion coefficientsb and age by race for the total sample riskavz1 riskavz2 riskavz3 age white black white black white black white black mean 0.535 0.169 0.267 0.084 0.245 0.081 51.538 45.936 median 0.351 0.000 0.164 0.000 0.011 0.000 50 43 sd 32.01 0.793 21.18 0.605 21.18 0.603 16.68 16.30 number of observations 15738 1790 15738 1790 15738 1790 15738 1790 difference of meant-test t-statistic5 21.43 t-statistic5 21.08 t-statistic5 20.97 t-statistic5 213.74a wilcoxin ranksum testc z-statistic5 228.49a z-statistic5 228.87a z-statistic5 227.75a z-statistic5 213.66a a indicates a level of significance at least at the 1% level. b riskazv1 is equal to (stocks1 realest 1 bus 1 nmmf 1 retqliz) divided by net worth. riskavz2 is equal to (stocks1 realest 1 bus 1 nmmf) divided by net worth. riskavz3 is equal to (stocks1 realest 1 bus) divided by net worth. if net worth is equal to 0 it was set to $1.00 for these calculations. c tests the hypothesis that the two samples are from populations with the same medians. 144 y.a. badu et al. / financial services review 8 (1999) 129–147 so far our investigation of the differences in net worth between households headed by whites and blacks has centered on the asset side of the balance sheet. we would now like to examine the liability side of the balance sheet. these differences if they exist would manifest themselves in the attitudes of the head of the households. data pertaining to these issues appear in table 8. the scf survey asks respondents to indicate the rate on various loans, including mortgages. we examined all of the loan types and find that first mortgage rates and a set of loan rates for appliances, furniture, hobby or recreational equipment are significantly higher for households headed by blacks. these results are found in panel a. all other mortgage and loan rates were not significantly different across households. attitudes of the households toward debt also seem to be different. significantly more households headed table 8 some liability costs and household attitudes toward debt responses panel a: rates charged first mortgage loan 4b loan 5b white black white black white black mean 8.862 9.523 12.618 14.027 13.660 15.322 std dev 1.716 2.669 5.107 4.570 6.364 5.379 observations 6814 514 1380 205 31 56 wilcoxin rank-sum test z-statistic55.53a z-statistic5 4.05a z-statistic52.07a panel b: household attitudes toward debt how do you feel about credit? have you been turned down or not received as much credit as asked for? did you shop around for the best deal? white black white black white black good idea 5066 796 turned down 0.1629 0.3235 (0.3219) (0.4447) a great deal of shopping 0.3772 0.4648 bad idea 4462 522 not as much credit 0.0221 0.0447 (0.285) (0.2916) good/bad idea 6210 472 not turned down 0.8150 0.6319 (0.3914) 0.2637) almost no shopping 0.1524 0.1737 observations 15738 1790 observations 15738 1790 wilcoxin ranksum test z-statistic5 210.01a z-statistic5 212.71a z-statistic5 2.90a a indicates a level of significance at least at the 5% level. b the scf asks households “not counting credit cards or loans you have already told me about do you any money on loans for household appliances, furniture, hobby or recreational equipment,” these are the rates list for the types of loans. there are at least 12 categories for these types of loans, those presented represented only the loans whose rates are significantly different across races. 145y.a. badu et al. / financial services review 8 (1999) 129–147 by blacks seem to think that credit is a good idea, that is, a good way to finance assets. further, significantly more households headed by blacks have been either turned down for a loan or did not receive as much as was asked for in a loan application. finally, black households self-identify as conducting significantly more thorough searches for the best rate available. to summarize, black households have a more favorable attitude toward credit as a means of financing assets. these households’ requests for credit are turned down or not fully met by the market more often than for white households and, in several classes of important loan types, black households pay a significantly higher rate. all of this despite the fact that households headed by blacks say they conduct extensive market searches for the best rates. 5. conclusion the evidence presented indicates that there is a significant difference in the net worth of households headed by those who identify themselves as white and black. we are unable to show that the net worth of black households is constrained by barriers in the financial markets, although we find that black households typically pay higher rates for several important types of credit instruments and these higher rates are obtained even though the black households self-identify that they conduct extensive searches in the markets for the best rate available. further, we find that white and black households hold different asset and liability combinations in their portfolios and determine that black households are significantly more risk averse (less risk tolerant) in their choice of assets. we think there are some alarming potential patterns. one disconcerting finding in our results is the heavy reliance on credit cards by the under-35 black households. the higher cost of credit cards, coupled with the tendency for most credit card balances to linger as revolving balances, almost certainly adversely affects future net worth of these households. on average black and white households have a heavy reliance on credit cards and public policy should persuade credit card companies to reveal the necessary monthly payment needed to pay the average outstanding credit card balance off over a responsible time period. another distressing finding is the insignificant net worth of black households on average. when this outcome is coupled with the conservative risk tolerance behavior of black households and the unhealthy balance sheet structure of younger black households, we question whether black households, on average, will be able to ‘catch up’ to the net worth of white households. it is not our position that public policy should have equivalent net worth between the races as a goal. however, public policy and future research should address the usage of the financial system as a means of wealth creation. additionally, the impact of intergenerational transfers of wealth on the risk tolerance behavior of households is another important area of study. acknowledgment the authors thank the editor and the anonymous referees for their helpful comments. this paper has also benefited from the discussions of colleagues at the finance department seminar 146 y.a. badu et al. / financial services review 8 (1999) 129–147 series at virginia commonwealth university, especially david dubofsky and gabriel rameriz, and a discussion with james t. lindley from the university of southern mississippi. the authors remain solely responsible for all content. references adams, m. (1995). balance sheet structure and the managerial-discretion hypothesis: an exploratory empirical study of new zealand life insurance companies.journal of accounting and finance, 35(1), 21–45. ando, a. & modigliani, f. (1963). the life-cycle hypothesis of saving.american economic review, 53(1), 55–74. bajtelsmit, v. l., bernasek, a., & jianakoplos, n. a. (1999). gender differences in defined contribution pension decisions.financial services review, 8, 1–10. blau, f. d. & graham, j. w. (1990). black-white differences in wealth and asset composition.the quarterly journal of economics, 105(2), 321–339. brimmer, a. (1991). building wealth and assets.black enterprise, 12(12), 31–32. browne, r. s. (1993). wealth distribution and its impact on minorities.the review of black political economy, 21(3), 111–120. friend, i. & blume m. e. (1975). the demand for risky assets.the american economic review, 65(5), 900–922. gutter, m. s., fox, j. j. & montalto c. p. (2000). racial differences in investor decision making.financial services review, 8,149–162. haliassos, m. & bertaut, c. c. (1995). why do so few hold stocks?the economic journal, 105(september), 1110–1129. heckman, j. (1979). sample selection bias as a specification error.econometrica, 47(january), 153–161. jackson, j. & lindley, j. (1989). measuring the extent of wage discrimination: a statistical test and a caveat. applied economics, 21, 515–540. morin, r. a. & suarez, a. f. (1983). risk aversion revisited.journal of finance, 38(4), 1206–1216. oliver, m. l. & shapiro, t. m. (1995).black wealth/white wealth: a new perspective on racial inequality. new york: routledge. paxson, c. (1990). borrowing constraints and portfolio choice.quarterly journal of economics, 105(2), 535–543. schooley, d. k. & worden, d. d. (1996). risk aversion measures: comparing attitudes and asset allocation. financial services review, 5(2), 87–99. survey of consumer finances (1992). federal reserve. trial lawyers for public justice & the tlpj foundation. (1999). (http://www.tlpj.org/tlpjf/pressreleases/ tran2pr.htm). u.s. census bureau. asset ownership of households: 1993. (http://www.census.gov/hhes/www/wealth/ wlth93f.html). van auken, h. e., doran, b. m. & yoon, k. j. (1993). a financial comparison between korean and us firms: a cross balance sheet canonical correlation analysis.journal of small business management, (31)3, 73–83. wang, y. (1995). permanent income and wealth accumulation: a cross-sectional study of chinese urban and rural households.economic development and cultural change, 43(3), 523–550. wolff, e. n. (1994). trends in household wealth in the united states: 1962–1983 and 1983–1989.review of income and wealth, 40(2), 143–174. 147y.a. badu et al. / financial services review 8 (1999) 129–147 financial services review, 33(2) 36 exploring the effect of federal student loan payment resumption on borrowers through sentiment and textual analysis using x jason n. anderson,1 donovan sanchez,2 juan e. gallardo,3 derek lawson,4 and congrong ouyang5 abstract student loans have taken on an increasingly significant role in funding the higher education experience and payments toward student loan debt have become an important part of many borrowers’ overall financial plan. using brandwatch, this study analyzes x data to better understand student loan borrower sentiment during the resumption of federal student loan payments in october 2023. during the period studied, negative references to student loans on the platform overtook positive sentiment overwhelmingly (46% negative versus 1% positive). topics and phrases labeled negative sentiment ranked higher in mentions than those with positive or neutral sentiment in their respective categories. the findings highlight the need for financial planners to provide appropriate mental health resources to help borrowers manage negative feelings surrounding the federal student loan payment restart. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation anderson, j., sanchez, d., gallardo, j., lawson, d., & ouyang, c. (2025). exploring the effect of federal student loan payment resumption on borrowers through sentiment and textual analysis using x. financial services review, 33(2), 36-54. introduction given the increasing costs of college education, many college attendees rely on student loans to complete their degrees. according to the national center for education statistics (2023), the average total cost for a first-time, full-time undergraduate student living on campus at a public 4-year institution during the 2021-2022 1 corresponding author (jasonanderson@ksu.edu). university of kansas, lawrence, kansas, usa. 2 kansas state university, manhattan, kansas, usa. 3 kansas state university, manhattan, kansas, usa. 4 kansas state university, manhattan, kansas, usa. 5 kansas state university, manhattan, kansas, usa. disclosure: we employed ai algorithms, specifically brandwatch, to analyze the data collected for this study. the ai methods were selected based on their suitability for the research objectives and data characteristics. academic year was $26,000. for first-time, fulltime undergraduate students attending private nonprofit 4-year institutions, the price tag was much higher at $55,800. from march 13, 2020, to september 1, 2023, the u.s. department of education paused payments and set interest rates to 0% on eligible federal student loans as part of the nation’s covid-19 https://creativecommons.org/licenses/by-nc/4.0/ mailto:jasonanderson@ksu.edu https://creativecommons.org/licenses/by-nc/4.0/ anderson et al. 37 emergency relief measures. during this same period, the biden administration proposed a plan to forgive up to $20,000 for qualifying borrowers. this plan was ultimately blocked by the u.s. supreme court in june 2023. after this decision, student loan interest was set to begin accruing in september 2023 with payment resumption in october 2023. the breakneck pace of student loan changes over the last four years has undoubtedly affected student loan borrowers, who must now address their student loan debt in addition to cumulative societal transitions brought on by the pandemic. this project seeks to understand student loan borrower sentiment during payment resumption through an analysis of their expressions on social media. an increased understanding of borrower sentiment will help financial professionals better appreciate the mental health ramifications of this financial transition on borrowers. the authors wish to make a point of clarification in light of company name changes that could create confusion for the reader. in july 2023, elon musk rebranded the well-known microblogging platform, twitter, to x. because this change is so recent, however, most previous studies will refer to the platform by its former name, twitter. to avoid confusion, the authors have chosen to refer to the platform as x in this paper, even retroactively, unless twitter is referenced directly in a quote or article title. the purpose of this research is to evaluate the emotional impact of recent student loan changes on borrowers. one way to gauge the emotional change within student loan borrowers is through the proxy of textual sentiment during times of major student loan changes. to proxy student loan borrowers, researchers can measure individuals who discuss student loans on x. as mentioned previously, from early 2020 to late 2023 student loan borrowers made no payments or accrued interest on eligible federal student loans, and the biden administration proposed a generous loan forgiveness program. after the denial of student loan forgiveness, interest accrual resumed in september 2023 and payments in october 2023. this paper seeks to address the following question: how did sentiment change on x during the resumption of federal student loan payments in october 2023? based on the findings of sinha et al. (2023) – which are covered later in the literature review we hypothesize that: h1: negative sentiment will outweigh positive sentiment during the study’s date range. h2: topics and phrases with negative sentiment will rank higher in mentions than those with positive or neutral sentiment during the study’s date range. literature review student loans because student loans are readily accessible to students, generally have lower interest rates than private loans, and offer fixed rates, it is no surprise that their usage has increased (ebrahimian, 2023). despite their popularity, borrowers have polarized views about student loans. as a means of funding their education, some borrowers feel that the cost is worth it, while others feel weighed down (nuckols et al., 2020). such feelings often prompt individuals to share their student debt experiences on social media platforms. student loans have been shown to have numerous effects on borrowers. for example, kim and chatterjee (2019) identified a negative association between student loans and an individual’s life satisfaction and psychological well-being. student loans have also been associated with psychological distress, financial anxiety, and other negative forms of health and psychological well-being (archuleta et al., 2013; kim & chatterjee, 2019; zhang et al., 2019). student loans have also had negative outcomes to the extent that some individuals develop problems with mental health, smoking, and heavy drinking (qian et al., 2021). not all borrowers view student debt the same way. borrowers in the repayment period of their loans reported higher psychological distress than those who were still enrolled in a university (sato et al., 2020). in addition, and when compared to individuals without student loans, those repaying their loans had lower levels of financial satisfaction (kim et al., 2021). reflecting contradictory findings, joseph and macdonald financial services review, 33(2) 38 (2021) concluded that student loans are negatively associated with financial satisfaction, while robb et al. (2018) found no significant effect. the contradicting results from these studies highlight the complexity of the relationship between student loans and the emotional state of borrowers. in addition to the psychological effects associated with student loan debt, asset accumulation has also been highlighted in research. mountain et al. (2020) found that student loans negatively affected millennials’ homeownership rates. also of concern, student debt has been associated with a reduced likelihood to save for retirement (elliott et al., 2013). beyond concerns related to retirement and housing, student loan borrowers are more likely to carry other types of debts, like car and credit card debt (fry, 2012). debt repayment reduces disposable income for borrowers, and the resulting lack of liquidity can make it more challenging to finance other purchases (gicheva & thompson, 2014). furthermore, having student debt hinders borrowers’ ability to spend. in a recent paper, 18% of student loan holders reported difficulty buying daily necessities because of their existing debt (hanson, 2023). social media as a data source in contrast to traditional research sources, social media data is “user-generated, naturalistic, and unstructured,” creating a fascinating opportunity for researchers to explore public opinion on a variety of topics (sinha et al., 2023, p. 736). social media as a tool to express opinions has become more popular as the number of users has increased with the availability of the internet (ortiz-ospina, 2019). while social media may be used to share various types of information, it has also been used to express sentiment on specific topics that affect a wide range of individuals. considering its massive growth, companies and media organizations are interested in exploring x user data as a means of detecting user sentiment on various products and services (kouloumpis et al., 2011). likewise, significant interest has developed among researchers for purposes of better understanding the public’s opinion on important topics. what makes x (and social media in general) advantageous from a research standpoint is that it provides insight into the minds of users by the voluntary posting of their thoughts, expressions, as well as their interactions with other platform users in “a naturalistic setting” (de choudhury et al., 2013, p. 128) and in real-time (edo-osagie et al., 2020). chancellor et al. (2020) note that new computational methods of social media data analysis could make significant differences, such as using social media data to identify and provide interventions for risky behavior. the significant number of users also provides a benefit to researchers in that there are a large sample of tweets to look at. as zimbra et al. (2018, p. 24) notes, “many researchers and firms have recognized that valuable insights on issues related to business and society may be achieved by analyzing the opinions expressed in the abundance of tweets”. the use of social media data in research harvesting social media data is a non-traditional means of obtaining the opinions and sentiment of subjects. traditional forms of gathering opinion and sentiment data include the use of surveys to try to determine how study participants feel about a particular issue. the internet and modern technologies have created new spaces for researchers to gather data (edo-osagie et al., 2020). in their research seeking to identify depression on x, nadeem et al. (2016) found that social media data represents a potential solution to problems that can arise in self-reported depression questionnaires in that postings on social media often provide a window directly into the state of mind of the social media user. researchers can use social media data to “automatically identify self-expressions” in the construction of a given data set (coppersmith et al. 2014, p. 52). limited research has been conducted to date using social media and x for analysis of student loans. one notable exception is a 2023 study by sinha et al. looking at student loans and mental health expressions on reddit and x. sinha et al. (2023) used scarcity theory to explore reddit and x data to improve understanding of the relationship between student loans and mental health. they found that social media posts about anderson et al. 39 student loans were associated with negative sentiment and had a higher likelihood of containing expressions of sadness and fear. a large body of work used x and other social media data to study mental illness and mood disorders. in their review of studies attempting to predict mental illness, guntuku et al. categorized studies into two camps: finding correlates to, or identifying, mental illness (2017, p. 43). preoţiucpietro et al.’s (2015) research focused on predicting “linguistic markers” of mental illness (p. 28) while de choudhury et al. (2013) predicted major depression using x. similarly, chancellor et al. (2020) used social media data to identify mood and psychosocial disorders. reece et al. (2017, p. 8) used x to predict depression and ptsd, noting that their “method identified these mental health conditions earlier and more accurately than the performance of trained health professionals, and was more precise than previous computational approaches”. finally, coppersmith et al. (2018) demonstrated how automatic procedures using social media data and natural language processing (nlp) can detect suicide risk. considering the unique way individuals use social media, researchers are naturally attracted to it as a means of examining a variety of behaviors, moods, opinions, and sentiment. because x allows users to express themselves and interact with others, researchers can monitor sentiment and analyze data with relative ease. social science explores the interaction of individuals in their environments with other human beings, and x documents these interactions in a way that can be informative to the social science researcher. in their research, quantifying mental health signals in twitter, coppersmith et al. noted that “social media is by nature social, which means that social patterns, a critical part of mental health and illness, may be readily observable in raw twitter data.” (2014, p. 51). coppersmith et al. (2014) successfully used x data to perform individual and population-level analyses to identify mental health disorder signals for depression, bipolar disorder, post-traumatic stress disorder, and seasonal affective disorder. in another study, coppersmith et al. were able to “easily and automatically” identify x users with ptsd by “scanning for tweets expressing explicit diagnoses” instead of relying on “traditional ptsd diagnostic tools” (2014, p. 579). tsugawa et al. (2015, p. 9) found that by analyzing x user history data that “depression can be recognized in users with an accuracy of approximately 69%”. while text is typically the object of analysis for researchers, guntuku et al. looked at image postings and profile pictures on x for clues relating to user mental health and found that anxious users tended to post more photos, and both depressed as well as anxious users posted images “dominated by grayscale” (2019, p. 244). edo-osagie et al. (2020) undertook a scoping review of the use of x for public health research and found that studies most often used x data for surveillance, event detection, pharmacovigilance, forecasting, disease tracking, and geographic identification. sentiment analysis the vastness of social media calls for useful summarization tools to gain insights into the underlying data. sentiment analysis can automate this otherwise cumbersome process to pull opinionated data from massive datasets (giachanou et al., 2017). sentiment analysis strives to improve the automatic recognition of sentiment within a text (zimbra et al., 2018) and detects opinions based on features selected by researchers (giachanou et al., 2017). while there are a variety of ways to seek to understand student loan sentiment among borrowers, sentiment analysis offers a unique opportunity. as noted previously, automating the processing of large chunks of data is valuable from an efficiency and effectiveness standpoint. most beneficial, in the opinion of the authors, is the opportunity to understand borrower sentiment as expressed by them without prompting from a researcher. in a way, sentiment analysis allows researchers to observe human behavior in its natural habitat. x sentiment analysis x is a microblogging platform that allows users to publish their opinions on virtually any topic. social science researchers can examine user sentiment with relative ease on a wide range of financial services review, 33(2) 40 issues and across a large population considering x’s substantial user base and significant number of daily messages (giachanou et al., 2017). x sentiment analysis may be viewed as a “specialized area within sentiment analysis” (zimbra et al., 2018, p. 3) where x data is mined for opinionated text on a given topic (giachanou et al., 2017), with sentiment detection occurring through the establishment of word embeddings (carvalho et al., 2021). x sentiment analysis is not without its challenges. some of these challenges include length limitation of messages (giachanou et al., 2017), “novel language” resulting from length of message constraints (zimbra et al., 2018, p. 24), as well as written errors and content that is constantly changing (giachanou et al., 2017). despite its challenges, x sentiment analysis remains an important tool for researchers. it offers direct insights from users, shedding light on public opinions on critical issues (giachanou et al., 2017). methodology on june 23, 2023, x removed academic api access and significantly lowered the caps allowed for data collection by academics. this development made it nearly impossible for academics to gather data in a cost-effective manner for use in open-sourced statistical software or readily available data science models. in light of these challenges during the creation of this study, the author team elected to use brandwatch (https://www.brandwatch.com/) to collect social media data. using this software, data was collected containing the term “student loan” and the hashtag #studentloans from october 1 to october 31, 2023 (the query was exported on november 1, 2023). the query was set to collect all mentions of student loans across various content sources, including x, reddit, tumbler, youtube, news sites, blogs, forums, and review sites. instead of collecting all data across these platforms, brandwatch collected a statistically significant sample size with a calculated rate of 33.36%. the export filtered out pornographic content and profanities. this larger dataset (n = 317,394 with 208,595 unique authors) was filtered to collect only records from x (n = 248,303 with 176,640 unique authors), which represented 78.2% of the comprehensive dataset. the final sample for this study was 248,303. a textual and sentiment analysis of the collected data was conducted within the brandwatch online software. brandwatch defines sentiment analysis as “the process used to determine the attitude, opinion and emotion expressed by a person about a particular topic in an online mention” (brandwatch, 2023). a sentiment score is calculated based on the proximity of sentiment and search terms, within the context of larger blocks of text (brandwatch, 2023). importantly, if a collection of text cannot be accurately categorized as positive or negative, it is classified as neutral (brandwatch, 2023). brandwatch conducts sentiment analysis and calculates sentiment scores using proprietary algorithms, machine learning, and natural language processing (“nlp”) (brandwatch, 2023). specifically, brandwatch uses a pretrained model using transformer-based deep learning technology to assign sentiment scores (brandwatch, n.d.-a; brandwatch, n.d.-b). transformer models are commonly used to analyze sentiment for social media data and have been used to analyze data from x (bokolo & liu, 2024; gong et al., 2022; kokab et al., 2022; padmalal et al., 2024; sharma et al., 2022). the sentiment model was trained on user annotated sentiment material from the brandwatch platform, third party datasets from academic studies, annotation service data, and brandwatch in-house data (brandwatch, n.d.-a). sentiment accuracy across languages is estimated to be between 60-75% with an average f1 predictive performance of 55-65% (brandwatch, n.d.-a). this accuracy range reflects the challenge of conducting a sentiment analysis across a large dataset, specifically, the tradeoff inherent in a data science model between accuracy and bias (brandwatch, n.d.-a). brandwatch’s model has been benchmarked to others such as monkeylearn, aylien, idol, metamind, alchemyapi, and datumbox with comparable performance (brandwatch, n.d.-a). results figure 1 shows the number of mentions per day across the date range. the mean was 8,009 and https://www.brandwatch.com/ anderson et al. 41 the median 7,147. between september 30 and october 23, mentions spiked on four days: october 1 (n = 12,888), october 4 (n = 23,849), october 18 (n = 12,732), and october 31 (n = 14,882). figure 2 provides the overall sentiment for student loan mentions within the study’s data range. most of the mentions were neutral (53%), with 46% negative and 1% positive. this finding supports the first hypothesis, as negative sentiment overwhelmingly outweighs positive sentiment during the date range. figure 1. student loan mentions by date note: n = 248,291. brandwatch excluded twelve records when completing this analysis. figure 2. overall sentiment for student loan mentions note: n = 248,301. brandwatch excluded two records when completing this analysis. 0 5000 10000 15000 20000 25000 30000 1 -o ct 2 -o ct 3 -o ct 4 -o ct 5 -o ct 6 -o ct 7 -o ct 8 -o ct 9 -o ct 1 0 -o ct 1 1 -o ct 1 2 -o ct 1 3 -o ct 1 4 -o ct 1 5 -o ct 1 6 -o ct 1 7 -o ct 1 8 -o ct 1 9 -o ct 2 0 -o ct 2 1 -o ct 2 2 -o ct 2 3 -o ct 2 4 -o ct 2 5 -o ct 2 6 -o ct 2 7 -o ct 2 8 -o ct 2 9 -o ct 3 0 -o ct 3 1 -o ct 1% 53% 46% positive neutral negative financial services review, 33(2) 42 table 1 provides additional detail on the twenty most popular phrases, persons, and keywords with sentiments of neutral, negative, and positive along with the category sentiment score. the topranking keyword was debt, with 106,605 mentions, followed by forgiveness (42,010), billion (32,453), tax (30,804), and money (27,854). the top-ranking person – in fact, the only person listed in the top twenty topics – was biden (45,422), who also ranked as the second most popular topic across types (president is also listed as twentieth on this list with 17,805 mentions). the top-ranking topics for negative sentiment were debt (69,286 negative sentiment versus 722 positive sentiment), money (24,742 negative sentiment versus 104 positive sentiment), tax (22,976 negative sentiment versus 38 positive sentiment), biden (20,935 negative sentiment versus 218 positive sentiment), and genocide (20,539 negative sentiment versus 26 positive sentiment). every topic, keyword, and phrase listed in the top twenty most popular topics had a negative sentiment score, supporting our second hypothesis. to further investigate the findings presented in table 1 and provide contextual meaning, a positive to negative sentiment ratio was created. this ratio was then compared to the number of mentions. the correlation between the positivenegative sentiment ratio and mentions was calculated to be 0.149, indicating a weak, positive relationship. this weak relationship further supports the findings and gives context to any potential bias in interpretation of high frequency words. when table 1 was sorted by highest positive-negative sentiment ratio, the top five results form a cluster around student loans: borrowers (0.028), forgiveness (0.026), student loan payments (0.023), payments (0.020), and president (0.015). however, each of these ratios remains quite low at <0.03. table 2 groups each of the top topics into three broad categories (student loans, government, and other/ambiguous) to better focus on borrower sentiment attached to student loans. the average positive-negative sentiment ratio was calculated for each category. the student loans category had the highest average positive-negative sentiment ratio at 0.021. the next highest categories were government (0.005) and other/ambiguous (0.002). financial services review, 33(2) 43 table 1. top 20 topics by mentions and corresponding sentiment statistics topic name type mentions negative neutral positive sentiment score +/ratio debt keyword 106605 69286 36596 722 -64 0.010 biden person 45422 20935 24268 218 -45 0.010 forgiveness keyword 42010 16435 25150 425 -38 0.026 billion keyword 32453 14024 18377 50 -43 0.004 tax keyword 30804 22976 7788 38 -74 0.002 money keyword 27854 24742 3006 104 -88 0.004 payments keyword 24643 10852 13571 218 -43 0.020 relief keyword 22649 15013 7488 146 -65 0.010 women keyword 20785 17928 2851 5 -86 0.000 genocide keyword 20638 20539 71 26 -99 0.001 americans keyword 20536 9218 11248 68 -44 0.007 borrowers keyword 19924 7300 12420 203 -35 0.028 student loan payments phrase 19906 8613 11095 197 -42 0.023 tax money phrase 19834 19756 77 0 -99 0.000 children keyword 19762 18614 1124 23 -94 0.001 voted keyword 18593 6397 12165 29 -34 0.005 hard keyword 18557 18029 515 11 -97 0.001 universal keyword 18200 17697 494 8 -97 0.000 health care phrase 17883 17754 119 8 -99 0.000 president keyword 17805 5840 11875 89 -32 0.015 financial services review, 33(2) 44 table 2. average topic positive-negative sentiment ratio across categories topic name type mentions negative neutral positive sentiment score +/ratio category: student loans debt keyword 106605 69286 36596 722 -64 0.010 forgiveness keyword 42010 16435 25150 425 -38 0.026 payments keyword 24643 10852 13571 218 -43 0.020 borrowers keyword 19924 7300 12420 203 -35 0.028 student loan payments phrase 19906 8613 11095 197 -42 0.023 average +/score 0.021 category: government biden person 45422 20935 24268 218 -45 0.010 tax keyword 30804 22976 7788 38 -74 0.002 money keyword 27854 24742 3006 104 -88 0.004 americans keyword 20536 9218 11248 68 -44 0.007 tax money phrase 19834 19756 77 0 -99 0.000 voted keyword 18593 6397 12165 29 -34 0.005 health care phrase 17883 17754 119 8 -99 0.000 president keyword 17805 5840 11875 89 -32 0.015 average +/score 0.005 category: other/ambiguous billion keyword 32453 14024 18377 50 -43 0.004 relief keyword 22649 15013 7488 146 -65 0.010 women keyword 20785 17928 2851 5 -86 0.000 genocide keyword 20638 20539 71 26 -99 0.001 children keyword 19762 18614 1124 23 -94 0.001 hard keyword 18557 18029 515 11 -97 0.001 universal keyword 18200 17697 494 8 -97 0.000 average +/score 0.002 anderson et al. 45 table 3 displays the twenty most popular phrases with sentiments of neutral, negative, and positive, along with the category sentiment score. the topranking phrases were student loan payments (19,906 mentions), followed by tax money (19,834), health care (17,883), forgiving student loan debt (17,469), and funding a genocide (17,217). regarding the mix of sentiment, 16/20 (80%) of topics had a negative sentiment score, and 4/20 (20%) had a neutral sentiment score of 0. the top-ranking topics for positive sentiment were student loan payments (197 positive sentiment versus 8,613 negative sentiment), student loan borrowers (38 positive sentiment versus 4,583 negative sentiment), supreme court (26 positive sentiment versus 11,908 negative sentiment), student debt relief (17 positive sentiment versus 8,154 negative sentiment), and voted for student loan forgiveness (17 positive sentiment versus 119 negative sentiment). even though they ranked highest for positive sentiment, each of these phrases had a negative overall sentiment score. no topics listed had a positive overall sentiment score. in fact, four topics on the list tax money, health care, forgiving student loan debt, and funding a genocide had the maximum negative sentiment score of -99. this finding offers additional support for our second hypothesis. a positive-negative sentiment ratio was created for phrases to compare the relationship between positive and negative sentiment. this ratio was then compared to the number of mentions. the correlation between the positive-negative sentiment ratio and mentions was calculated to be -0.078 (when rows with score of 0 for both positive and negative sentiment were assigned a ratio score of 0) or -0.122 (when rows with score of 0 for both positive and negative sentiment were excluded from the analysis). these correlations indicate a weak, negative relationship between the positive-negative sentiment ratio and mentions. when table 3 was sorted by highest positive-negative sentiment ratio, the top five results were voted for student loan forgiveness (0.143), student loan payments (0.023), student loan borrowers (0.008), millions of americans (0.004), and billion in student loan debt (0.003). each of these phrases had a positive-negative sentiment ratio of <0.15. the eleventh most popular phrase #fitness check comments for full video was unrelated to the topic of student loans. similarly, the seventeenth most popular phrase, bringing him home pt3 #fitness, was also unrelated to student loans. the topic of fitness—and why it might have shown up in the results—is expounded in subsequent parts of the paper. table 4 groups the popular phrases into the previously utilized three categories of student loans, government, and other/ambiguous. the average positive-negative sentiment ratio was calculated for each category. the student loans category had the highest average positivenegative sentiment ratio at 0.02 with the other two categories (other/ambiguous and government) tied with 0.001. financial services review, 33(2) 46 table 3. top 20 phrases and corresponding sentiment statistics rank topic name mentions negative neutral positive sentiment score +/ ratio 1 student loan payments 19906 8613 11095 197 -42 0.023 2 tax money 19834 19756 77 0 -99 0.000 3 health care 17883 17754 119 8 -99 0.000 4 forgiving student loan debt 17469 17427 35 5 -99 0.000 5 funding a genocide 17217 17214 2 0 -99 0.000 6 supreme court 13832 11908 1897 26 -85 0.002 7 student loan billing statement 12165 2092 10070 2 -17 0.001 8 helping 18yo 11266 0 11266 0 0 0.000 9 student loan borrowers 8925 4583 4302 38 -50 0.008 10 student debt relief 8877 8154 704 17 -91 0.002 11 #fitness check comments for full video 8811 0 8811 0 0 0.000 12 million americans 8796 6553 2233 8 -74 0.001 13 voted for student loan forgiveness 8607 119 8469 17 -1 0.143 14 billion in student loan debt 8163 2986 5168 8 -36 0.003 15 student with his student loan 8139 0 8139 0 0 0.000 16 voted to restart 8016 4625 3390 0 -57 0.000 17 bringing him home pt3 #fitness 7638 0 7638 0 0 0.000 18 cancel your student debt 7192 3945 3246 0 -54 0.000 19 canceling an additional 6196 2305 3891 0 -37 0.000 20 millions of americans 5597 470 5123 2 -8 0.004 anderson et al. 47 table 4. average phrase positive-negative sentiment ratio across categories topic name mentions negative neutral positive sentiment score +/ratio category: student loans student loan payments 19906 8613 11095 197 -42 0.023 forgiving student loan debt 17469 17427 35 5 -99 0.000 student loan billing statement 12165 2092 10070 2 -17 0.001 student loan borrowers 8925 4583 4302 38 -50 0.008 student debt relief 8877 8154 704 17 -91 0.002 voted for student loan forgiveness 8607 119 8469 17 -1 0.143 billion in student loan debt 8163 2986 5168 8 -36 0.003 student with his student loan 8139 0 8139 0 0 0.000 cancel your student debt 7192 3945 3246 0 -54 0.000 average +/score 0.020 category: government tax money 19834 19756 77 0 -99 0.000 health care 17883 17754 119 8 -99 0.000 supreme court 13832 11908 1897 26 -85 0.002 average +/score 0.001 category: other/ambiguous funding a genocide 17217 17214 2 0 -99 0.000 helping 18yo 11266 0 11266 0 0 0.000 #fitness check comments for full video 8811 0 8811 0 0 0.000 million americans 8796 6553 2233 8 -74 0.001 voted to restart 8016 4625 3390 0 -57 0.000 bringing him home pt3 #fitness 7638 0 7638 0 0 0.000 canceling an additional 6196 2305 3891 0 -37 0.000 millions of americans 5597 470 5123 2 -8 0.004 average +/score 0.001 financial services review, 33(2) 48 table 5 shows the top emojis used across all tweets gathered. although this table is not directly related to a hypothesis – or attached to sentiment scores – it gives helpful insights into the mindset of the authors when writing about student loan topics. across all tweets (tweets and retweets), the five most popular emojis were smiling face with horns (n = 11,704), money bag (n = 1,864), police cars revolving light (n = 1,696), spool of thread (n = 1,202), and white down pointing backhand index (n = 1,067). for tweets, the five most popular emojis were money bag (n = 595), face with tears of joy (n = 490), loudly crying face (n = 310), white down pointing backhand index (n = 94), and police cars revolving light (n = 55). when sorted by the greatest total number of impressions, the five most popular emojis were smiling face with horns (384,182,044), white down pointing backhand index (102,521,059), money bag (59,243,121), splashing sweat symbol (39,698,047), and movie camera (36,811,251). of the ten top emojis, at least three carry a negative connotation (smiling face with horns, police cars revolving light, loudly crying face) with 14,068 total tweets and retweets and 392,030,725 impressions. table 5. top 10 emojis rank emoji label all tweets retweets tweets impressions 1 😈 smiling face with horns 11,704 11,701 2 384,182,044 2 💰 money bag 1,864 1,268 595 59,243,121 3 🚨 police cars revolving light 1,696 1,639 55 6,804,599 4 🧵 spool of thread 1,202 1,154 46 4,542,805 5 👇 white down pointing backhand index 1,067 971 94 102,521,059 6 🎥 movie camera 1,052 1,049 2 36,811,251 7 💦 splashing sweat symbol 1,001 992 8 39,698,047 8 😂 face with tears of joy 743 251 490 11,401,407 9 🤷🏽‍♂️ shrugging 734 719 14 1,468,062 10 😭 loudly crying face 668 356 310 1,044,082 table 6 shows the top hashtags for the data gathered. like emojis, these hashtags are not attached to sentiment scores. the top hashtag is unrelated to student loans: #fitness. interestingly, this hashtag had 12,753 retweets but no tweets. beyond this anomaly (addressed in the limitations section), the second to tenth hashtags were all related to student loans. the second most popular hashtag was #studentloans followed by #cancelstudentdebt. the sixth most popular hashtag was #scotus, referring to the supreme court of the united states. for tweets (versus retweets), the five most popular hashtags were #studentloans (1,984), #cancelstudentdebt (775), #poortax (386), #studentloan (262), and #cancelallstudentdebtnow (235). the top five hashtags had combined impressions of 34,533,799, with the list garnering 537,460,898. anderson et al. 49 table 6. top 10 hashtags rank hashtag all tweets retweets tweets impressions 1 #fitness 12,753 12,753 0 501,162,722 2 #studentloans 3,567 1,582 1,984 23,392,392 3 #cancelstudentdebt 1,933 1,157 775 3,600,657 4 #studentloan 674 410 262 7,267,211 5 #poortax 446 59 386 146,290 6 #scotus 428 395 31 646,113 7 #fixstudentloans 404 368 34 342,014 8 #studentloanforgiveness 305 122 181 650,748 9 #cancelallstudentdebtnow 266 29 235 127,249 10 #cancelallstudentdebt 260 179 79 125,502 discussion the years and months preceding october 2023 were filled with a whirlwind of change in the federal student loan space. on march 13, 2020, the trump administration announced the suspension of federal student loan payments and interest. a few years later, on august 24, 2022, the biden administration announced blanket student loan forgiveness while payments and interest were still suspended. on june 30 of the following year the supreme court rejected that forgiveness proposal. in the wake of this announcement, borrowers were told that payments and interest would resume – without any balance forgiven – after over three years of no payments or interest accrual on federal student loans. this study’s analysis of x data advances a better understanding of student loan borrower sentiment during the resumption of federal student loan payments in october 2023. our study demonstrates that borrowers experienced negative sentiment during this resumption of payments, especially without the help of forgiveness. references to student loans on the platform demonstrated a negative sentiment that greatly outpaced positive (46% negative versus 1% positive). given the timeline outlined above, it is unsurprising that several of the most popular hashtags found in this study pointed toward student loan forgiveness as a major topic, as shown in the third, sixth, eighth, ninth, and tenth most popular hashtags (#cancelstudentdebt, #scotus, #studentloanforgiveness, #cancelallstudentdebtnow, and #cancelallstudentdebt, respectively). in summary, both of our hypotheses were confirmed. although neutral sentiment represented the greatest percentage across all data points (53%), negative sentiment considerably outweighed positive during the resumption of federal student loan payments. furthermore, the topics and phrases with negative sentiment ranked higher in mentions than those with positive or neutral sentiment. all the top-ranking topics had negative sentiment scores, while phrases had higher-ranking negative sentiment scores (with negative sentiment scores taking the top seven slots and 80% of the top twenty list). although student loan topics as a category ranked higher in average positive-negative sentiment ratios, the correlations between positive-negative sentiment ratios and mentions remained weak, pointing to a lack of bias in the findings. implications financial planners interacting with clients after the payment resumption should introduce appropriate resources to help borrowers cope with the negative feelings they may experience surrounding this area of their financial lives. luckily for the profession, this study’s findings align perfectly with the emergence of financial therapy as a bona fide discipline within the field of personal financial planning. many practitioners in this niche, associations such as financial services review, 33(2) 50 the financial therapy association (fta), and personal financial planning academic programs such as the one at kansas state university, offer training and resources for financial planners hoping to support clients in distressing financial situations. similarly, the cfp board has put greater emphasis on teaching the behavioral aspects of financial planning with the addition of the “psychology of financial planning” in the cfp® certification 2021 principal knowledge topics list (cfp® certification 2021 principal knowledge topics 2021). this means the topic is now taught in all registered educational programs and tested on the cfp® exam. whatever the preferred term (i.e., financial therapy or behavioral finance), teaching planners psychological support mechanisms has significant potential to benefit financial planning clients and student loan borrowers. practitioners would be wise to heed the warnings from industry thought-leaders on how student loans can delay a client’s financial growth. previously outlined studies, such as the one from archuleta et al. (2013), demonstrate the negative aspects of carrying student loan debt, including the alarming connection between student loans and financial anxiety. in a recent article for business insights, researcher and financial therapist dr. megan mccoy explained how student loans can harm borrowers in two crucial ways: increasing shame and delaying financial milestones (aguino & richtmyer, 2023). both can be devastating for a client’s progress toward financial independence. the mounting evidence is a call to action for advisors; once and for all, student loans deserve a dedicated space in the financial plan. the costs are too great – and only increasing – for the emerging generation already saddled with student loan debt. limitations this study does not compare sentiment during repayment to a period beforehand. as such, no comparisons can be made between how sentiment changed before the federal student loan interest and payment restart and after. without the ability to compare to previous sentiment, it is impossible to decipher if negative sentiment is reflective of student loans (in general), the resumption of federal student loan payments and interest, or the collapse of biden’s student loan forgiveness initiative. additionally, textual patterns and sentiment captured during the month of october might not be fully reflective of when student loan payments resumed as a whole, as borrowers had different due dates within the month. this research project was largely descriptive in nature, using frequency-based analysis as the main way to uncover patterns in these gathered data. this tool was useful in identifying the overall narrative within the dataset, which can be especially helpful for exploratory research projects. however, relying too much on frequency-based analysis when using nlp can, at times, influence the interpretation of the results, introduce bias, or ignore broader context. as such, future research on this topic should use tools like regression analysis to push further into the investigation of concrete associations and relationships. accordingly, this paper’s conclusions should be interpreted in concert with those future contributions. the results displayed in table 6 showed that #fitness was the top hashtag across our dataset. table 3 also showed that the phrase “#fitness check comments for full video” was the eleventh-ranking phrase in the dataset. unfortunately, social media data sources contain noise which can cloud the signal. given the phrase had a sentiment score of 0, it is unlikely this result affected the testing of the study’s first hypothesis, although it is probable the phrase artificially boosted overall neutral sentiment. to gain a better understanding of the fitness phenomenon, on october 24, 2023, the authors searched the hashtags #studentloans and #fitness directly on x’s website. a review of the results showed that these two hashtags regularly appeared together within large groupings of hashtags (>5). it is possible the hashtag #fitness was used along with #studentloans to boost search ranking, but perhaps more research could uncover the reasoning behind this peculiar finding. as mentioned previously, the brandwatch query was set to filter out pornographic content and profanities. this filtering of profanities may have artificially lowered negative sentiment for this study, even though it was notably higher than anderson et al. 51 positive sentiment already. when the initial query was modified to filter out these categories, the query generator indicated this filter decreased the number of collected responses by as much as 50%.6 if included, this data would have likely pushed negative sentiment even higher. conclusions this study used brandwatch to analyze borrower sentiment on the x platform during the resumption of federal student loan payments starting in the month of october 2023. the emotions captured by this data-gathering and analysis process were quite bleak; many borrowers expressed overwhelmingly negative emotions directed at this change. however, this negative outlook highlights an opportunity for financial planners to proactively provide adequate support for the financial and mental health of federal student loan borrowers. given appropriate and timely action, even negative student loan events can further solidify the financial planners’ positive presence in their clients’ lives. references aguino, l. & richard richtmyer, r. 2023. “student loans aren't just bad for your wallet — they're bad for your mental health, too.” business insider. accessed april 25, 2024, https://www.businessinsider.com/person al-finance/student-loans-negativelyaffect-borrowers-mental-health-2022-9. archuleta, k. l., dale, a., & spann, s. m. (2013). college students and financial distress: exploring debt, financial satisfaction, and financial anxiety. journal of financial counseling and planning, 24(2), 50–62. bokolo, b. g., & liu, q. (2024). advanced comparative analysis of machine learning and transformer models for depression and suicide detection in social media texts. electronics, 13(20), 3980. https://doi.org/10.3390/electronics13203 6 a preliminary review of the data before filtering showed a large number of records containing the fword 980 brandwatch. (n.d.-a). bcr sentiment methodology and evaluation. brandwatch. (n.d.-b). sentiment analysis – consumer research. brandwatch help. retrieved october 30, 2024, from https://consumer-researchhelp.brandwatch.com/hc/enus/restricted?return_to=https%3a%2f% 2fconsumer-researchhelp.brandwatch.com%2fhc%2fenus%2farticles%2f360013739958sentiment-analysis brandwatch (2023, september 20). sentiment analysis. https://www.brandwatch.com/wpcontent/uploads/brandwatch/sentimentanalysis.pdf carvalho, j., & plastino, a. (2021). on the evaluation and combination of state-ofthe-art features in twitter sentiment analysis. artificial intelligence review, 54(3), 1887–1936. https://doi.org/10.1007/s10462-02009895-6 cfp® certification 2021 principal knowledge topics (2021). accessed april 25, 2024. https://www.cfp.net/-/media/files/cfpboard/cfp-certification/2021-practiceanalysis/2021-principal-knowledgetopics.pdf. chancellor, s., & de choudhury, m. (2020). methods in predictive techniques for mental health status on social media: a critical review. npj digital medicine, 3(1), article 1. https://doi.org/10.1038/s41746-0200233-7 financial services review, 33(2) 52 coppersmith, g., dredze, m., & harman, c. (2014). quantifying mental health signals in twitter. proceedings of the workshop on computational linguistics and clinical psychology: from linguistic signal to clinical reality, 51– 60. https://doi.org/10.3115/v1/w143207 coppersmith, g., harman, c., & dredze, m. (2014). measuring post traumatic stress disorder in twitter. proceedings of the international aaai conference on web and social media, 8(1), article 1. https://doi.org/10.1609/icwsm.v8i1.1457 4 coppersmith, g., leary, r., crutchley, p., & fine, a. (2018). natural language processing of social media as screening for suicide risk. biomedical informatics insights, 10, 1178222618792860. https://doi.org/10.1177/11782226187928 60 de choudhury, m., gamon, m., counts, s., and horvitz, e. (2013). predicting depression via social media. proceedings of the international aaai conference on web and social media 7(1): 128–37. https://doi.org/10.1609/icwsm.v7i1.1443 2 ebrahimian, m. (2023). student loans and social mobility (ssrn scholarly paper 3680159). https://doi.org/10.2139/ssrn.3680159 edo-osagie, o., de la iglesia, b., lake, i., & edeghere, o. (2020). a scoping review of the use of twitter for public health research. computers in biology and medicine, 122, 103770. https://doi.org/10.1016/j.compbiomed.2 020.103770 elliott, w., grinstein-weiss, m., & nam, i. (2013). “student debt and declining retirement savings (working paper no. 13‐34).” center for social development. fry, r. (2012). a record one-in-five households now owe student loan debt. pew research center. giachanou, a., & crestani, f. (2017). like it or not: a survey of twitter sentiment analysis methods. acm computing surveys, 49(2), 1–41. https://doi.org/10.1145/2938640 gicheva, d., & thompson, j. (2014). the effects of student loans on long-term household financial stability. department of economics working paper no. 14-02. greensboro: university of north carolina. gong, x., ying, w., zhong, s., & gong, s. (2022). text sentiment analysis based on transformer and augmentation. frontiers in psychology, 13, 906061. https://doi.org/10.3389/fpsyg.2022.9060 61 guntuku, s. c., preotiuc-pietro, d., eichstaedt, j. c., & ungar, l. h. (2019). what twitter profile and posted images reveal about depression and anxiety. proceedings of the international aaai conference on web and social media, 13, 236–246. https://doi.org/10.1609/icwsm.v13i01.32 25 guntuku, s. c., yaden, d. b., kern, m. l., ungar, l. h., & eichstaedt, j. c. (2017). detecting depression and mental illness on social media: an integrative review. current opinion in behavioral sciences, 18, 43–49. https://doi.org/10.1016/j.cobeha.2017.07 .005 hanson, m (2023, september 20). economic effects of student loan debt. educationdata.org. https://educationdata.org/student-loandebt-economic-impact joseph, m., & macdonald, m. (2021). investigating antecedents to financial satisfaction in emerging adults. 2021 academic research colloquium. https://www.researchgate.net/publicatio n/355903050_investigating_antecedent s_to_financial_satisfaction_in_emergin g_adults https://doi.org/10.1609/icwsm.v13i01.3225 https://doi.org/10.1609/icwsm.v13i01.3225 https://doi.org/10.1016/j.cobeha.2017.07.005 https://doi.org/10.1016/j.cobeha.2017.07.005 anderson et al. 53 kim, j., & chatterjee, s. (2019). student loans, health, and life satisfaction of us households: evidence from a panel study. journal of family and economic issues, 40(1), 36–50. https://doi.org/10.1007/s10834-0189594-3 kim, k. t., lee, j. m., & lee, j. (2021). student loans and financial satisfaction: the moderating role of financial education. journal of financial counseling and planning. https://doi.org/10.1891/jfcp19-00002 kokab, s.t., asghar, s., & naz, s. (2022). transformer-based deep learning models for the sentiment analysis of social media data. array, 14, 100157. https://doi.org/10.1016/j.array.2022.100 157 kouloumpis, e., wilson, t., & moore, j. (2011). twitter sentiment analysis: the good the bad and the omg! proceedings of the international aaai conference on web and social media, 5(1), article 1. https://doi.org/10.1609/icwsm.v5i1.1418 5 mountain, t. p., cao, x., kim, n., & gutter, m. s. (2020). millennials’ future homeownership and the role of student loan debt. family and consumer sciences research journal, 49(1), 5–23. https://doi.org/10.1111/fcsr.12374 nadeem, m., horn, m., coppersmith, g., & sen, s. (2016). identifying depression on twitter (arxiv:1607.07384). arxiv. https://doi.org/10.48550/arxiv.1607.073 84 national center for education statistics. (2023, november 11). price of attending an undergraduate institution. u.s. department of education, institute of education sciences. https://nces.ed.gov/programs/coe/indicat or/cua nuckols, w., bullington, k. e., & gregory, d. e. (2020). was it worth it? using student loans to finance a college degree. higher education politics & economics, 6(1), article 1. https://doi.org/10.32674/hepe.v6i1.1358 ortiz-ospina, e. (2019, september 20). the rise of social media. ourworldindata.org. https://ourworldindata.org/rise-of-socialmedia padmalal, s., dayanand, i. e., rao, g. s., & gore, s. (2024). enhancing sentiment analysis in social media texts using transformerbased nlp models. international journal of electrical and electronics engineering, 11(8), 208–216. https://doi.org/10.14445/23488379/ijee e-v11i8p118 preoţiuc-pietro, d., eichstaedt, j., park, g., sap, m., smith, l., tobolsky, v., schwartz, h. a., & ungar, l. (2015). the role of personality, age, and gender in tweeting about mental illness. proceedings of the 2nd workshop on computational linguistics and clinical psychology: from linguistic signal to clinical reality, 21–30. https://doi.org/10.3115/v1/w15-1203 qian, y., & fan, w. (2023). student loans, mental health, and substance use: a gender comparison among us young adults. journal of american college health, 71(3), 930–941. https://doi.org/10.1080/07448481.2021. 1909046 reece, a. g., reagan, a. j., lix, k. l. m., dodds, p. s., danforth, c. m., & langer, e. j. (2017). forecasting the onset and course of mental illness with twitter data. scientific reports, 7(1), 13006. https://doi.org/10.1038/s41598-01712961-9 robb, c. a., chatterjee, s., porto, n., & cude, b. j. (2019). the influence of student loan debt on financial satisfaction. journal of https://doi.org/10.48550/arxiv.1607.07384 https://doi.org/10.48550/arxiv.1607.07384 https://doi.org/10.1038/s41598-017-12961-9 https://doi.org/10.1038/s41598-017-12961-9 financial services review, 33(2) 54 family and economic issues, 40(1), 51– 73. https://doi.org/10.1007/s10834-0189599-y sato, y., watt, r. g., saijo, y., yoshioka, e., & osaka, k. (2020). student loans and psychological distress: a cross-sectional study of young adults in japan. journal of epidemiology, 30(10), 436–441. https://doi.org/10.2188/jea.je20190057 sharma, u., pandey, p., & kumar, s. (2022). a transformer-based model for evaluation of information relevance in online socialmedia: a case study of covid-19 media posts. new generation computing, 40(4), 1029–1052. https://doi.org/10.1007/s00354-02100151-1 sinha, g. r., larrison, c. r., brooks, i., & kursuncu, u. (2023). comparing naturalistic mental health expressions on student loan debts using reddit and twitter. journal of evidence-based social work, 20(5), 727–742. https://doi.org/10.1080/26408066.2023. 2202668 tsugawa, s., kikuchi, y., kishino, f., nakajima, k., itoh, y., & ohsaki, h. (2015). recognizing depression from twitter activity. proceedings of the 33rd annual acm conference on human factors in computing systems, 3187–3196. https://doi.org/10.1145/2702123.270228 0 zhang, q., & kim, h. (2019). american young adults’ debt and psychological distress. journal of family and economic issues, 40(1), 22–35. https://doi.org/10.1007/s10834-0189605-4 zimbra, d., abbasi, a., zeng, d., & chen, h. (2018). the state-of-the-art in twitter sentiment analysis: a review and benchmark evaluation. acm transactions on management information systems, 9(2), 1–29. https://doi.org/10.1145/3185045 pii: s1057-0810(02)00093-8 � ������ ���� � �� � � � ����� ��� ���� ���� � �������� � � ��� �� ���� ��� �� � �� � ����� ��� �� � � ����� �!� ������� �� ��� ��" ���������� ��� ����� ��� �������� ���������� � ��� �� ����� ���� ��� ��� !� " ��� ��#� ������� �$ %&&&�'&���� ��� ���������� (���������� (��#������ ��� $ �������� ���������� � ��� �� ����� ���� ��� ��� !� " ��� ��#� ������� �$ %&&&�'&���� ��� ����������� ������� ��� )�� ������ � ���������� ����������� ! �� �� (� *&++�� ��� #������� "$ %� � �� "&&&' �������� � ������� ���� "( �� � "&&!' ���� �� !) *���� �� "&&! �������� + ����� � ����� �� ���� �������� � � ��� �� ���� ��� ,�+-��. �/ �� �� 0�� ��� �� ������ � 0�� �+-� ����� � � ��� � ��� ����� � �/���� �� � � � �/���� �� �� �� 1� �� �� ���� � ����� � � ��� �� � � � � ���� � �� � � � ����� � � ����� � � � �� ��� �� 1� 2 � � �� ���� �� �� �� � 0�� �� � ���� ������ ��� � �� ������ ��� �2� � �� ���� � � �� �� �� � � � � � �� ��� � ��� � �� �� � � �� �� � �� � 3�� �� �� �� ������ ��� �� � 4�������� �� !& ��� ���������� � � ����� ��� � �� � � � �� ��� ����� �/�� � ��0�� ����2��� � �� � �� �� � ���� �� � � �� 5 6���7� (&& ,�56 (&&. � ��/� � "&&! +������� ���� �� �� ��� ���� � ��������� ,$ ������ ����� �8!!&' 8!$& .��/ ����������� � �� ��' �������� � �� � ������ �' 6�� ����� � ���� ' 9� �� � � � �����' 4 ��� ��2��� �� � ���� ������ �������� � � ��� �� ���� ��� ,�+-��. �� ������� ���� �� � � �� �� �� ��� ���� �� ��� � �� � �� ���� � �� � ��� � �� � ������ �� �� ���� �� -� �� ����� �� ���� ,����. � �� � ��� � ��� ,:�;.� �� �� ��� �� � �� � ���� ������ ,�������.� �� ��� �� �+-�� ��� ��� ���� ������ � ���� :�;� %��� �� � � 0� <� �� � �������� #����0 !& ,"&&!. ")=>"$? !&(=�&?!&@&!@a > ��� ��� � �� � "&&! +������� ���� �� �� ��� ���� � ��������� 6--b � ! & ( = � & ? ! & , & " . & & & c ) � ? � �������� �� � � ���� 9���b �!�c&$�d"&�"d)&' � /b !�c&$�d$d�"(c$� ,'���� ���������� �����e� ����� ,���� � ����� .� f����� e� ����� ,���� ����� .� f� �� e ������ ,���� ��� .� ! 9���b �!�c&$�d"&�"=?&� " 9���b �!�d!=�)=)�(&($� ��� ��� ����� � � ����� � �� � ��� �� �� 0�� ��� ������ � � ���� ���� � �+-�� � � �/���� �� � � ��� ��� � � � � ��� � �f�� �� ���2 �� 9�� � ����� ���0 � ��� � gg ������ ��� 77 � �� � � � ����� � �� �+-�� ������ � � :�; � ���� ��� � �� �� ������ � ���0 �� �� �� �� �� � ��� �� � ���� �� �� ���� �� � ��� � 9���� 2 �� �� �� �� � ���� � � � ��� ����� � ���� � � � ����� � ��� �� ��� �� � ��� ��2��� ��� %�� ���� ��� �/ � �� ��� ��� �� �� �� ��� � � ��� �� � � ���� ���� �/ � ���� �� �� � � ������ � � ���� ������ ��� �� � ,� ����� � ���� ���0 �� �.� )&��� � ��� �� ��� �� ��������� � � � � �� ��� � � ����� �/�� ��� � ������� � � � �� � � ���� � ���� ����� ������ 0� ����� ��0 �+-� � �� � �� �� � �� ���� �� � �� ��� � �� �� ��� � �� ������ ��� �� � � �� ��� �� �� � 0���� �� � � ��� �� �� �� ��� ��� � ��� � � � � �� ��� ������� � ������ �� ��� ���������� %��� �� � � 0� ��� ��� � ����� �� �������� � �� �� �������� ���� �� � �� �� % � ����� �� ������� � ����� ��� � � ���� � � ����� � � ��� ����� �� 0� � �� � � � ��� � � �� ��� � �� ������ �� ��� � ���� 0���� �� 4 ����� ,!cc(.� 0�� 2 �� � � � ��� ��� �� � ��� ����� �� � � ����� �h�� � �� � � � � �� ���� ����� ����� �� 0� � ������ �� ����� i ���������� j 6�� � � 1�h� ,!cc$. � #��� �� � 1���� � ,"&&&. ����� � �� � ��� �� ���� � 0�� �������� � �� ��� :�; �� �� �� � ��� � ���� ������ �� �� � 1��� ��� ,"&&&. ����� � ������ � ����� � � � ��� ����� � �� ������ � ���� � �� �� � ��� �� �� �������� 0�� ��� ������ � �� �� ����� �� �� ��� �� � �� ���� �� ���� � ����� <� �� �� � ��� �� �� � �� 3�� � �/�����2�� � j��� ��������� � 9� ��� ,!cc&� !cc!.� 0�� ���� � ���� � ���� � � �/�� � ������ � � � ���@� �� ��� �� 0���� � �� ��� ,!cc). ����� � � � �� =k �� ������ �� �� � � �� � �� � ���� � ����� �� � � !& �� ��� ���� �� �� ����� �� �������� �� � ��� �� �� �� � ����� ������� � ������ � ����� � ��� �� �� �� � � ��� � ��� � ��� � �� ���� �� � � �� � ��� -����� �� � ���� 0� � ��� �� �� �� �� �� � � � �������2� �� � �2 �� � ��� ����� �� � � � ��� ��2��� ��� �� � � ����� ��� ���:� �� �� ���� :� �� � 1�� ��� ,!cc&.� ����� � � l��� � j�� ,!cc(. � � ���� ����� � � ��� �� ��� ��� �� � �� ���� ����� ����� �� 0� � �� � ��������� +��� h ,!cc!.� � ���� � j�� ,!cc".� #������ ,!cc?. � � ���� ����� � ��� �� �� �� �� �� ���� ���� ���� ��� � ��� � � ���� ��� � ��� �� i�0����� � ����� � ����� � <�������� � � � � �� ,"&&!. 2 � � ��� �� �� � �� �� � �� ���� ������� ����� 0� � ��� ��� �� � ��� �� �� � � � ����� � ����� ����� � ������� ��� � %7�� �� � *�0 � ,"&&&. � � �� ��� � � 0���� ���� � � ��� �� �� ��� � � � ��� �� �� � ����� �� � ���� � �� f�� �� � ��� �� �� �� � �� � �� � � � ����� � ���h� � �+-� �� ���� 9�� � � ����� �������� � ��� �� �� �� ���� � ��� ���� 9�� 2�� �� gg��� ��� 0���� ��77 � � ���� 0����� �+-�� �� � ��� � ��� ��� �� ���� � ������ � � ��� ����� �� ���� �� ��� ��� �� � �+-�� ��� � � �� ��� ���� ��� � ������ � 6� ��� ,!cc(.� � ��� � ��� �� �� �� �� ������ ��� ��0�� �� �� � ������ �� � 0��� � ���� � ����� � ����� ������ ��� �� ���� � ���� � � � ��� � ������ ���� � � ��� ����� ������ �� ��� �� � �+-�� ��� � �� ��� ��� ��h� �� �� �+-��7 ������ �� �� � ,!cc=. �� ���� � ��� ")? ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 ������� � ��� � � ���� m� � ��� ��� �� � ��� � ������ �������� � ��� �� �� ��� 0����� � � ����� � �� ���� � � �� ��� � 0�� � ��� � ��/�� � �� � �� ��� ,��� �� � 5 9 ����� !ccd � .� ��� ���� ���� �� � ����� �� 2�� � ������ � �� �� �� ��� �� ������ � � � � � � ��� �� �� � ���� ����� �� � ����������� �� �� �/���� ���� ���� ����� �� � ��0������� �� ��� %�� ����� ����� �� ��� ��� 0���� �� � �� � � � ����� ������ �� � �� �/ ��� �� ��� �2� �� �/ � �� ��� ���� � �� �� � ���� � ���� %�+� � ��� 9�� � 0� ��� �� 0����� ������� � ��� � ����� � ������ �� ��� ($ ���� ���� � �+-�� ���� � � �� !cd= � *���� �� !ccd ,!�(d$ 0����.� 6����� � � ��� � ���� �� �� ���� �� 6��� ������ ����� � ������ �� ���� ( ��7� �������� "�� ��� � ��� ���� � � � ��� ������ �� � ���� ����� 0� � �+-��' ��0���� ������ � ������ �� �� �f�� �� ������ ���� *����� �� �� ����� ��� � �� �/ � � ��� �� �� ��������� � ��� � � �� �� �+-�� � ������ � �� ����� �� �������� � 9 �� !� �� � 0� � �� � �� � 0���� � �� �+-� �� � ���� �� ���� � �� �� �� �� ������� + �� �� �� ($ �+-�� ����� �� �� �� ���2�� � gg�� �� � ���� � �� �77 �� gg����� ��h�� ���� � �� �77 � 6��� ������ ����� 4�� �� ���� �� �� ����� ���� ���� � ���� ������ ���� � ����� � � ����� ���� �� �/������ � � ��� ��������� ��� � � ��� �� 9�� � ���� ���� � � ����� � 2��� �� ���2�� � 6��� ������ ���� � gg� ���� � ��������� � ��� �� ��77 gg�� ��� � �� ������ ��� �� ��77 �� gg0���� ���� � �� ��77 �� �� ��� �� �� �+-�� �/ �� �� �� � � ���� �� ���� �� � ��� )&��� � ������� � �� 0���7� � ��� �� ������� � �� ����� � �� ���� �+-�� 0� �� �� �� �� ($ ��� 0���� 0����� �� �� � �� � �� ��� % ��� ��� ���� �� !d�c �� �� � �� �� ��� �� ����� �� � �� 0���� % �� (k �� �� 0���� � �� � � �� ��0�� �� �� � �� ��� � �� =&k �� �� 0���� � �� !? �� ��0�� � �� ��� � �� ��� � � � )&k � �� � 0�� !? � $"� %�&� ��������� � 1� � � � ��� � ��� �� � �� ����� �� ���� �� �� � ��� � ���� 0� � ������ � ��� ����� � �� ���� � �� ����� ��� �� � �� �� �b #� �� � � � � �� �4 ��� #� �� � � �� ,!. 0����� #� �� �� �� 0����� �� �� ��� �� � � g4 ��� #� �� 7 �� 0����� �� �� ��� �� � ��� � ��/ ,�56 (&& �� *�0.' � � ����2��� �� � ��� �� �� ��� �� � � �� �� �� ����� ��� ��� �� � � 1� ����� � 0����� � � ��� � �� �� � ��� � � � �� � � �� 5 6���7� (&& ,�56 (&&. � �� *�0 �� �� � ��� �� � � ��/ ,*�0.� 0���� �� �� ���� !� �8��� ,������ �� � ������.� 9 �� ! �� � � �� �� � � � ��� � �� �� � 0�� ��� � �� �56 (&& � �� � ��� ���/�� 9�� 0���� �� ��� �� � ��� ��� 0���� �� ��� �� � � � � ��� � �� �56 (&& �� &�d! � &�(?� ������ ������ 0���� ���� ��� ��� ��� � �� *�0 �� &�(d ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 ")c 9 �� ! ���� ���� � �������� � � ��� �� ���� ��� ���� � ����� �� � �� � � � ����� �������� � � ��� �� ���� � <��� 0��� ��h�� j � 0��� ��h�� 4� � � � �� �� +/����� ���� � !@d@d= !"@"=@cd &�(( ���� �� ����4 ��� ��� �� <� � !!@$@c$ !"@"=@cd &�&c ���� �� 8�� � + ���� �� <� � d@!(@c& !"@"=@cd &�=& �� �� � �� � �� !"@!(@c( !"@"=@cd &�!$ � ��� <� ���� 5 ���� � c@!&@=! !"@"=@cd &�d" ����� ��� � �� � ������ �� !@d@?c !"@"=@cd &�)" ���� ���� ; ��� <� � �� $@"$@?= !"@"=@cd &�d( ���� 5 � ���� #� � � ���� <� � d@!!@c) !"@"=@cd &�)" ���� 5 � ���� 9� � #� �� #� � � <� � !&@!@c) !"@"=@cd &�!& � ������ 5 8� �� � ������ �� !@d@d= $@"!@=? &�$d �� � � <� � �� � � $@!!@?d !"@"=@cd �&�&) �� � � ������ ��� ������ �� =@=@=" !"@"=@cd &�(( �� ���� � �� � ������ �� "@!?@== !"@"?@?$ &�$& *�� 0 �� 8���� 8�� � *����� � 5 ���� <� � $@")@c) !"@"=@cd &�!( *�� 0 �� 8���� *����� � 5 ���� <� � $@?@c$ !"@"=@cd &�)! 9�� *��� ��� <� � �� !@d@d= c@"=@=$ &�d) *��� 5 6����� � ��� ��� ���� <� � !@)&@?= !"@"=@cd &�"? + ��/ �� !@"@?! !"@"?@?$ &�(! +���� -�� !!@(@c) !"@"=@cd &�c= <��� <� �� � <� � (@!d@?d !"@"=@cd &�?) 8 ���� +��� � 9��� c@(@?d !"@"=@cd &�d" 8 ���� 8�� � 4�� ����� 9��� !!@!?@c$ !"@"=@cd &�)$ 8� �� � ������ ��� ��� ���� � !@d@d= !"@"=@cd &�=c i5n i� � �� �� ��� ��� (@!@?= !"@"=@cd &�c& i5n j��� ���� ��� ��� ��� (@?@c" !"@"=@cd !�&? ��2��� 4 ��� � <� � "@!d@c& !"@"=@cd &�() ��� i ���� � � 5 9���� %���� � � � <� � ?@"d@c$ !"@"=@cd &�?( ��� i ���� 6 ��� 8�� � *����� � <� � "@!=@c( !"@"=@cd &�?c ��� i ���� 6 ��� ����� *����� � <� � "@!=@c( !"@"=@cd &�$) j��� ������ �� !@d@d= (@!!@c& &�=) j� �� � ����� � +��� � <� � !!@=@?d !"@"=@cd &�dd j� �� � ����� � 8��0 � <� � !!@!&@c( !"@"=@cd !�)) 4<� ����� � ; ��� !!@!(@cd !"@"=@cd &�(! 4 ���� <� �� !@d@d= =@!@?) &�?) 4��� <� �� ��� �� =@!(@c$ !"@"=@cd &�"c 4��� 8�� ���� �� ���� � <� � (@""@?= !"@"=@cd &�c! :�-� 8��0 � <� � !@=@c$ !"@"=@cd �&�"! : �� � ��� �� 5 9��� ����� <� � =@c@d( $@"&@=c !�!) : �� � � � ��� 9 ��� 4 ��� � <� � =@"c@c$ !"@"=@cd &�d& : � ���� <� � "@")@=c !"@"?@?$ !�&! :�0 ��� 4��� <� � �� !&@"c@c) !"@"=@cd &�?( :�0 ������ i��� ���� <� � ?@)@=) !"@"?@?$ &�)) :� � � �� ��� !@d@d= =@!=@c" &�d) 6� ������ 5 #�������� ������ �� !@d@d= !"@"=@cd &�d" 6������ ������ � � 5 9���� <� � �� !@)!@?d !"@"=@cd &�=d #���� 8�� � 9��� �� !!@"!@?d !"@"=@cd &�($ #���� 4������ � 9��� �� !"@")@c) !"@"=@cd &�d! "$& ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 � &�("� ������ ������ % �� ���� �� �� � �� ��� �� � � � � ��� �� � !�&&� � ���� �� �� � �� ���� � ����h� �� 9�� � � � � � �� ���� �&�"& � !�)) ��� �� �56 (&&� � ���� �&�") � !�)= ��� �� *�0� ������ ������ 1� ��� ����� � ��������� �� � � � ��� � �� �� �� � 2 � �� �� � � � &�d)� 0� � � �� �� &�)&>&�c=� ��� �� �56 (&&� 0���� �� �� � ����� �� ��� � �� *�0 �� &�(c� 0� � � �� �� &�"=>&�c?� 9���� �� � � � ���� � � ���� �� ��� � ��� �� �� �� ����� �� ���� ����� � ���� ��� ���� ��� � ����� ����� <�� � �� 0���� �� ������ � 0�� � ��� ����� � :�; ��� � �� �+-� �� ����� �� � ������ � � ��� ������ :�; :�; ,". 6��� ��� ������ � ���� ������� �+-�� ����� � �������� 9�� � ��� �� �� � ��� �� ��� ����� � �� � ��� �� �� a!&&�&&& �� ��� � � �� �� )&��� � ������� � � � �� "$$ ������� � ��� �� � � ����� �� ��h��� � �� �� 0���� �� � ���� ��� � �� �� �� � ���� ������ ��� �� � � �+-� 0� � ������ � ���� 0��� ��� � � �� ������ � �� �� ��� �� 3 ���� ��� � ��� �� ���� � � 0���� ����� ���� � �+-� 0� � ������ � ���� 0��� ��� � � �� �� ���0�� � �� ���� ��� �� 3 ���� � � ���� ���� � � 0���� 9�� �� ��� �� ������ ������ � �� ���� ��� ��� �� � � ������ 9�� � � ����� � ������ � ��� ���� ��� ������ �� � ���� �� ��� � ���� �"k � �"(k � �� � ���� ���� ��� � ���� &k � �")k� ����� � � � �� ���� �� �� "k ��� � � � /���� �� �� !(k ��� �� 9� ����� � � ��0 ���� � � �� ��� 0���� �� ��� �� �"(k �� ��� � �� 0� � ���� ��� � ���� �!&k � �")k� 0���� � ��� �� ��� ��� � �� �!"k� �� ��� � �� 0� � ���� ��� � ���� &k � �!&k� � ��� ��� 0���� �� � � ��� �� �������� � � ��� � �+-�� 0� � ����� ������ � ��� � � �� ������ � �� ��� � <�� �/ ����� �� ���� �� �� �� ����� �� � �� 2�� 0���� �� a"(�&&& �� � ��� �� � � �� � �� -� � 0��� "� � � �� �� ���� �� �7� ������ ���0� ���2��� �� � ������ ����� �� � ��� �� � �� ����� � �� �� ���� �� � ��� ����� � �� 9 �� ! ,0 ���� �� . �������� � � ��� �� ���� � <��� 0��� ��h�� j � 0��� ��h�� 4� � � � � ���� ��� ���� <� � (@!?@c& !"@"=@cd &�== ��� �� � �� 9���� 5 � � <� � c@c@?c !"@"=@cd &�)? ������ � �� � �� !&@"&@=" !"@"=@cd &�(! � � �� �� ��� !@d@d= (@(@=? &�=& 9����� � � � ������ �� !@d@d= !"@"=@cd &�=) 9�� � � �� ������ �� !@d@d= "@)@=? &�$& o0��� <� � !&@)@?d !"@"=@cd &�$= 9�� �� � � � ����� �� ��� � �� �� �� ($ �� �� ���� �� �56 (&& � �� � ��� ���/�� % ��� ��� ���� �� !d�c �� �� � �� �� ��� �� ����� �� � � �� �� �� !�(d$ 0���� ,)& �� ��. �/ �� ��� ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 "$! �� �� �� �� ��� ���� ��� � �� �+-� � � � ����� �� �� ������ ���� � ������ �� ��� �� 0���� <�� �� �� � �� � ������ ��� � � � �� ��� � �� �� � � �� ���� �� � ��� �� �� �� ��� � �� �� ��� �� �� ��� ��� ������� �� � � �� �+-� �� � ��� 9� � � �� ��� � �� ��� ������� � � � �� ��� �� � �� ��� ����� ��� ��� �� ��� � � � �� ��� -� � 0��� ) ��� �� � �� �7� ������ ������� ��� �� � ��� �� �� ��� ����� �� �� ���� �� � �� ��� �� �� ���� �� ��� ��� �� � �� �� ��� ���� � ��� 0� �� �� �� ��� � ��� �� � �� � ��� � �� ���� �� ������� � � � � �� ��� � �� �� ���� � � � �� ��� �� �� �� �� ��� � � ��� � �� ����� �� � � �+-� ��� 0���� � ����� � �� � �� �� ��� ���� �� ���� ��� 0���� �� � �� �� ��� ������ �� ���� �� ��� ����� �� � � � �� �� ������ � 0��� � 0���� ���� �� � �+-�� � ��� �� ��� ��� � � � ������ � �� ��� ������ �� � ��� ��� ����� �� ������ � � � ��� �� � �� �56 (&& ��� � 0���� 9�� �� �� � �� �56 (&& ��� � 0��� ������� � �� �� �� � �� � ��� �� ��� ������ 9�� ������� �� �� ���� �� � ���� �� �� ����� ��� � ��� � �� �� �� � �� � � ��� ������ �� �� �� � 0���� � �� �� ������� ������ � ��� �� � �� �56 (&&� � �� 0���� �� � ��� � �� �� �� ����� �� � �� ������� �� � 0�� �� �� �� � �� ��� ����� ���� �� �������� 0��� � �� �� �� � �� �56 (&& ,� ����� � ������ ��. � � � �� 0���� 9�� ����� ��� � ��� � �� �� ��� ���� � � ��� �� ����� �� � �� ����� ��� ������ �� �� 0����� � ��� � �� �� � ���� �� (��� � ��� �� ��� � ����� ��� ������ �� !�&& ������� � �� �� � � �� ��� � � �� �� �� � �� �� �56 (&&� � � ������ ��� �� ,����. � !�&& �� � ��� ��� �� � � ��� � ������ ����� � �������� ,� ������. � ���� �� �� �56 (&&� 9���� ������� � � � �� ��� � �� � ������ � �� ����� �� �� ������ � � �� b !k� "k� � )k� 6���� � 4 � !� !c=(� ��������� �/�� ��� � ��� ���� ����/�� ��� !k � ����/�� ��� "k� ���� �� � ��� �� �� �� ����� � �� �� �� �� �� ��� � ���� <����0� � 4 � * �� ������ ������ ,������ � 1�9� � � � ���� �� �� � � ����.� ������ �� ��������� � � � ���� � =(k� � ���� ���� � �� 6��� ������ ����� 9�� �� � �� ����� � � �� ��� ���� � !k� % �� � ��� ��� � � ��@ �� ���� �� �� �� �������� � � � �� ��� � � 0�� !k � )k ���� � � �� ��� �� ����� �� � 9�� ���� � � � �� ��� ������ � �� "$$ ������� � ��� �� � � ����� ������ � =)" � � � � �� � �� �� � ��� �� ��������� ������ ��� ��� �� ���� ��� � ����� �� �+-� ������ � �� � �� �� ���� ��� ,��� *�j� �� ���������� �������� � 1 ��� � !cc!� !cc".� 0� ��� ������ �� � � � ��/ (��� � ������� � � ��� �� � 0�� ��� ������ � �� ��� � ��� � � � � ����� ������� �� � � ����� �������� 9�� 2�� 0��� �� � � �� �� �� �� �� !cd=� !c="� !c==� !c?"� !c?=� � !cc"� �� �� � � �� � �� �� �� ��/ �� �������� 9�� �����0� � ���� �� �� � ��� ��� ��������� ����� �� �� � ��� �� � ��2 � �� �� ���� �� ����� ��� � ��� � �� �� 0� � ������ � �� ��� � �� �� �� ��� ������ � �� �� � 0�� �� ��� � ���� ��� � #���� � ���� �� d �� �� (��� � ��� �� ��� �� ������ � � �� ��� ��������� � 0� � ��� � � ����� �� � ��� �� ��� � �� �� �� ���� ������� � � � �� ��� ������� �4#� � � � d=�=! � � � ="�=d � � � ==�?! � � � ?"�?d � � � ?=�c! � � � c"�cd � ��d=�=!��md=�=! � ��="�=d��m="�=d � ��==�?!��m==�?! � ��?"�?d��m?"�?d � ��?=�c!��m?=�c! � ��c"�cd��mc"�cd � ��d=�=!�6�:d=�=! � ��="�=d�6�:="�=d � ��==�?!�6�:==�?! � �?"�?d�6�:?"�?d � �?=�c!�6�:?=�c! � �c"�cd�6�:c"�cd � �� � "$" ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 0����� �4#� �� �� ����� ��� � ��� � �� �� ��� � � ��� �' � � 99��� �� �� � ����� ��� �� ��� �� �� !c99>!c��� ��99��� �� ����� �� � �� � �� ��� ������ !k ��� ������ ���� !c99 � !c��' ��m99>�� �� �� ��� � �� �� !c99>!c�� ,����� � ���� ��� � ���.' �99��� �� ����� �� !k ��� � ��� �� � �� �� � 0�� �� ��� � ���� ��� ���� !c99 � !c��' �6�:99>�� �� ����� �� ��� � �� �� � 0�� �� ��� � ���� ��� ���� !c99 � !c�� � �� � �� ����� ��� ��� � � ��� �� 9�� ����� � ���� �� ��������� ����� �� ���0 � 9 �� "� 8��� �� 2 �� �� �� �������� �� �� ��� ��� ��� � �� ������� �� ��� � ���� ������ �� �� ����2��� �� �� �� ��� 0 � ���� ����h�� � � ���� ���� � �������� �� ����� �� ��� �2� �� ������ ��� �� �� ��� � �� ��� �� � �� ����� �� ����� �������� �� �� ��� � � � ����� ����� �� � �� ����� �� ��0 9 �� " +� �� �� ��������� ����2��� � ,���� �� � �� �� � ����� ��� � ��� � �� �� . m� �� ����� ��� ��� �� !k 9� � � �� ��� ����� ,"" � &�c?d". !cd=>!c=! !�"!c="�� &�&&=))�� �&�&&$c&� !c=">!c=d !�!!?$d�� &�&&cdc�� �&�&!$$&�� !c==>!c?! !�(?$dd�� &�&!"d=�� �&�&&)&( !c?">!c?d &�c)($(�� &�&)(")�� �&�&!?cd�� !c?=>!cc! &�=d!$c�� &�&$))"�� �&�&$=c&�� !cc">!ccd !�"(&=(�� �&�&!d!= &�&!")c "k 9� � � �� ��� ����� ,"" � &�c?c?. !cd=>!c=! &�=)!d&�� &�&&??$�� &�&"!&? !c=">!c=d &�c=?dd�� &�&&&!$ &�&&!$( !c==>!c?! !�$&(($�� �&�&&!"" &�&"""" !c?">!c?d &�d)$d=�� &�&"?$?�� &�&&d?& !c?=>!cc! &�()($&�� &�&"$&!�� �&�&&?$(�� !cc">!ccd &�c&c""�� �&�&!)c! &�&"cdc )k 9� � � �� ��� ����� ,"" � &�c?=). !cd=>!c=! &�$&=?$�� &�&!&&"�� &�&)$?= !c=">!c=d &�?(=(&�� �&�&&d"( &�&!!c" !c==>!c?! !�"$?d?�� �&�&!&d( &�&)c)) !c?">!c?d &�$!$&=�� &�&")"=�� &�&"$&$ !c?=>!cc! &�)?!!)�� &�&!)()�� &�&!!=! !cc">!ccd &�d(???�� �&�&!!$c &�&)c$( 9�� �� 0�� ���� ��� �� � ����� � � � �� ���� ��� �� �� ��� � �� ����2��� � ,���� ����h�� � � ���� ��� � �� ���� ������ �����.� � ��� � � ����� �� �� � ��� �� ��� � �� �� �� ���� ������� � � � �� ��� ������� 0� � �� ��/ ��� �� ��� ������ � � �� ��� ��������� � 1� � �4#� � �� ����� ��� � ��� � �� �� ��� � � ��� �� �� ��������� ����� �� �4#��� � d=�=!�� � ="�=d � � � ==�?!�� � ?"�?d�� � ?=�c!�� � c"�cd���d=�=!��md=�=! � ��="�=d��m="�=d���==�?!��m==�?!���?"�?d��m?"�?d � ��?=�c!��m?=�c!���c"�cd��mc"�cd���d=�=!�6�:d=�=! � ��="�=d�6�:="�=d � ��==�?!�6�:==�?! � �?"�?d�6�:?"�?d � �?=�c!�6�:?=�c! � �c"�cd�6�:c"�cd � �� �� 0���� ��m99>�� ������� � �� �� ��� ������ ,k. � �� �� !c99>!c�� ,����� � ���� ��� � ���.� � �6�:99>�� ������� � �� ����� �� ��� � �� �� � 0�� �� ��� � ���� ��� ���� !c99 � !c��� 9�� � � �� �� � � 99��� �� �� �� � ����� ��� �� ��� �� �� !c99>!c��� ��99��� �� �� ����� �� � �� � �� �� ��� ������ !k ��� ������ ���� !c99 � !c��� � �99>�� �� �� ����� �� !k ��� � ��� �� � �� �� � 0�� �� ��� � ���� ��� ���� !c99 � !c��� � ��� �2� &�&( ������ �� ��� �2� &�&! ������ ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 "$) �� � � 0�� ������ ��� � ���� � �� � � � ��� �� �� �� � ���� �� �� �� ����� � ���� � ����� � � ���� ,!cc).� 6� ��� ,!cc(. � �� � ,!cc=. ������ � � � � ����� 0� � ��� ����� ���0 �� � ������ ������ �� �� �� 9���� �� ����2��� � �� �� �� � �� ��� �� ���� ����h�� � � �� ���� � �� ����� �� ��� �2� �� ������ �� ��� � �� ����� ��� �2� �� ������ ��� �� � ����� ���� �� 0�� ���� , ���� ���� �� � ������ � �� ����� � �� � ���0 �� �� �� �� ������� � �� ��� � �� ����2��� � ����� ��� �� ���.� ��� �� � �� ��������� ������ � � ����� �� � ��� ���� �� ��� ���2 �� � �� �� � ���2 �� ��� � �� � � � �� ��� ������ 9 ��� ) � $ ���0 �� !& ��� ���2 �� � !& �� � ���2 �� � � ������ ������ ������ ��� � �� �� �� ���� � � � �� ��� � ��� 9 �� ) 9� ��� ���2 �� � � ����� ,�� �� "$$.� ����� ��� ���� )& �� �� ,!�(d$ 0����. # � ��� ��� ������ ���� ��� ������ ����� ��� �� �� � �/���� �� �56 (&& 4� 0����� �� �� � � �� ���� �� �� 0����� �� �� 1���� � � ��� �� � � � � � ��� 1���� 0� � � �� �� � 1���� 0� � p( �� �� !k 9� � � �� ��� ����� ! "!k !?k !d!c�$k &�$$!k "�((ck $"ck "$?k (c$k " "!k !ck !$c"�(k &�$)dk "�(($k $dck ")?k (=!k ) ""k !?k !"c=�?k &�$"?k "�(=$k $?"k "$?k ((?k $ ""k !ck !"d"�?k &�$"dk "�(=$k ("!k ""ck ($$k ( "!k !=k !"(!�=k &�$"?k "�(c)k c!k ""ck d!dk d "&k !dk !"!&�dk &�$"!k "�)c&k )&(k "&&k ="?k = "&k !?k !"&"�&k &�$""k "�(&!k )d$k ""?k d$!k ? "&k !=k !"&&�"k &�$""k "�(&dk )$ck !c=k d?&k c "(k "&k !!("�)k &�$!ck "�$c(k d((k ")dk $)=k !& "$k "&k !&=(�"k &�$!(k "�("dk d)!k ""?k $d?k "k 9� � � �� ��� ����� ! "(k !?k )=(�&k &�)(?k "�("&k (c)k "d"k $d$k " "(k "&k )$d�(k &�)()k "�$c=k d((k ")dk $)=k ) "(k !ck "?"�ck &�)$)k "�$c&k d"ck ")dk $$(k $ !?k !&k "d!�(k &�))(k "�!?dk ddk !(dk !&&ck ( ")k !dk "(d�&k &�)$k "�()!k (&"k "!!k ((dk d !?k ck "$?�(k &�))"k "�!=&k (=k !!ck !&!=k = "(k "!k "$?�$k &�))=k "�(!?k d=)k ""=k $"dk ? !?k !!k "$)�)k &�))!k "�!?(k ?dk !)=k c=dk c !?k ?k "$&�dk &�))k "�!(dk (&k !!=k !&)dk !& "$k !?k ")c�&k &�))=k "�(=!k ((!k "=)k $c(k )k 9� � � �� ��� ����� ! "(k !?k !!)�)k &�)&=k "�()$k (c)k "d"k $d$k " "(k !)k =?�(k &�"c(k "�$$!k )?=k "dck (d=k ) "(k !&k d&�(k &�"?)k "�"!$k )!"k ?=k d&(k $ "(k "&k (?�dk &�"?=k "�(!dk d((k ")dk $)=k ( !?k ?k (=�ck &�"?"k "�!=)k (&k !!=k !&)dk d !?k =k ((�dk &�"?!k "�!=)k ))k !)$k !&=$k = "(k !ck ((�)k &�"?(k "�(&dk d"ck ")dk $$(k ? !?k (k ("�ck &�"=ck "�!d)k ))k !"$k !&cdk c !?k ck (!�$k &�"=ck "�!??k (=k !!ck !&!=k !& "(k !$k $c�dk &�"?=k "�(?)k $&&k )!=k ($?k "$$ ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 � ������� %�� ����� � ������ � ����� 2 �� �� � � �� � � � ����� ������� � ������ � �� �+-�� �� �� � �� �� � �� � 0��� � �/���� �� � ��� �� �� �� ��� ������ �� � � � �� ��� �� 9 �� ) ����� �� �� ��� ���2 �� !& � � ����� ��� � �� � � � �� ��� ������ � 0��� � ���� �� �� �� �� �� �� 0���� � 0���� �� � � ��� � ��� �� � � ����� �� ��� �� �� ��� 9�� ������ ����� ��� � ��� � �� �� � ��� �� !k� "k� � )k ��� ������ �� !�d!ck� )=(k� � !!)k� ������ ������ � �� )k ������ �� � � � �� ��� � ��� �2� �� ����� �� 9 �� $ 9� �� � ���2 �� � � ����� ,�� �� "$$.� ����� ��� ���� )& �� �� ,!�(d$ 0����. # � ��� ��� ������ ���� ��� ������ ����� ��� �� �� � �/���� �� �56 (&& 4� 0����� �� �� � � �� ���� �� �� 0����� �� �� 1���� � � ��� �� � � � � � ��� 1���� 0� � � �� �� � 1���� 0� � p( �� �� !k 9� � � �� ��� ����� ")( (k "k !"=�"k &�"cck !�?&$k !k &k !$??k ")d $k !k !"d�dk &�"cck !�=c"k !k &k !(!&k ")= (k &k !""�"k &�"c=k !�=c$k !k &k !(!!k ")? $k &k !!c�(k &�"c=k !�=c!k !k &k !(!?k ")c dk $k !!$�?k &�"cdk !�?)&k !k "k !$)dk "$& $k "k !!!�?k &�"c$k !�=c?k !k &k !(&"k "$! (k )k !&(�ck &�"c)k !�?""k !k "k !$="k "$" ?k dk !&"�=k &�"c"k !�?(?k !k "=k !)=$k "$) )k &k cd�"k &�"?ck !�=?ck !k &k !(")k "$$ )k !k c"�dk &�"??k !�=c"k !k &k !(!?k "k 9� � � �� ��� ����� ")( !&k ?k �(?�"k &�!c)k !�c))k "k d"k !)""k ")d !!k ck �(c�(k &�!c!k !�cd&k (k ddk !"?!k ")= ck =k �d"�"k &�!?dk !�???k !k $?k !)$dk ")? !ck !=k �d$�&k &�!c"k "�$dck "c)k !?(k =$(k ")c !?k !dk �d$�$k &�!?ck "�)")k ""dk !=$k ?"ck "$& !=k !(k �d(�$k &�!?dk "�"(!k !((k !="k c&!k "$! ?k dk �dd�!k &�!=?k !�?=!k !k "=k !)=$k "$" !$k !"k �d=�?k &�!=?k "�&?(k $(k c$k !!"&k "$) !(k !)k �="�?k &�!d?k "�!!)k dck !&"k !&()k "$$ !dk !$k �=c�dk &�!(!k "�!?"k c!k !$?k c?!k )k 9� � � �� ��� ����� ")( ?k dk �c$�)k &�&d(k !�c&"k !k "=k !)=$k ")d "!k !ck �c$�ck &�&="k "�d$?k $dck ")?k (=!k ")= !)k !!k �c(�$k &�&($k "�&==k "=k =?k !!=?k ")? !$k !"k �c=�$k &�&!ck "�!(dk $(k c$k !!"&k ")c !=k !(k �c?�!k &�&&)k "�)!(k !((k !="k c&!k "$& "&k !?k �c?�!k &�&&=k "�d&?k )d$k ""?k d$!k "$! !?k !dk �c?�"k &�&&&k "�)?ck ""dk !=$k ?"ck "$" !ck !=k �c?�"k &�&&)k "�())k "c)k !?(k =$(k "$) !(k !)k �c?�)k �&�&&(k "�!?=k dck !&"k !&()k "$$ !dk !$k �cc�&k �&�&)?k "�"d!k c!k !$?k c?!k ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 "$( �� � � � ���2 � �� �� �� �� !k � "k ������� � � ��� ���0 ���2 � 0��� � �/���� �� �� �56 (&&� ,9�� �56 (&& �� �� �� !"�ck � ���� �� &�")$k 0������ ���� �� )& �� �� 0� �/ �� ��. 9�� ��� ���2 �� � � ����� ���� ��� � ��� �� � �� �� � ���� � =&k �� �� 0����� � ���� ������ �� ���� 2�� �� ���� �� ��� �� ����� � ����� � ����� ��� � ����� �� �� ����� �� ��� � �� � � � �� ��� �� ��� 2 �� �� ������ ���� �� ������ � � � �� � � � ����� � �� �� � � ��� �� �� �� 9 �� $ ����� � �� !& �� � ���2 �� � � ����� ��� � �� � � � �� ��� ������ +/ �� �� �� �� ����� � ����� � �� �� ����� �� �� � � � �� ��� � � ���2 �� 1���� � �� �� � ���2 �� � � ��� �� !k ����� � ��� �� �� �� �� �� � � � ��� ��� �� ���� �� �� �56 (&&� ������ ��� � � �� f� � �� 0� � ��� � � � ����� � ����� � �� �� � 0��� ���0 �� � ��� � 9 ��� ) � $ ������ � �� �� ��� �� ������ �� ������ 0���� �� ��� �� ����� ���� �� ������ �� � � �� �� �� � �� ��� ������ 9�� ��� ���2 �� � � ����� �� ���� ���� ������ �� ��� ������ �� �� � � � �� ��� �� 9�� ��� �� �� ��� ��� �� !& ��� ���2 �� � � ����� �� ����/�� ��� �""k ��� �� ���� ��� ������� 9��� ����� �� �� ��� � 0� � � ����� � ������ �� �������� ����� � ����� � �� � ���� ��� � ��� ����� ��0 � � � �� ��� �� ���0��� � � ����� �� �� � �� � ���� �� �� �� 9�� !& ��� ���2 �� � � ����� !k ��� ����� ������ ��� �� �� � 0�� �� � ���� ������ ��� � �� )�$k ��� �� ���� ��� � ��� �� �� �� d�c � c�c ��� �� "k � )k ������� ������ ������ i����� ��� � ��� � ��� 0�� ��� � �� 0� � ���� � ��� ����� � � ���� ���0� �� � � ������ ����� � ���� �� ���� ��� ��� � �2 � 9 �� $ �������� ������ � � ����� ��� +�� 0� � �� �� �� �� ��� � ���� ������ �� ���� � � � �� ��� � ��� � �� 0� � ���0��� � � ����� ����� � �� ��� ����� ��� � ��� � �� �� �� ����� �� !& �� �� 0��� � � ����� 0� � � � "k � )k ��� � ��� �� ���0�� �� �� "k ��� �� 9 ��� ) � $ ��� ����� �� �� �� � � �� ���� �� � ��� �� � � ������ ���� � ���� �� �� � � �����7 � � �� ���� �� � ��� � � � ����/�� ��� �� � �� � �� �56 (&&7� �� �� � � �� ���� �� �� "�&"k ���� �� )&��� � ������� ����� � � �� ����� �� ��� ��� ����� ��� �� �� � � ������ 0� ���� �� � � ��� �� �� � � � � �� ���� �� � 0� � ���� �� �� �56 (&& � ����� � � �� ����2��� � �� � �� �� ,�; � �@�. ��� �� ���� ��� ������� 9�� �56 (&&7� �;�� ?�d) �/����� �� �� �;�� �� !& �� � � ����� �� !k ����� ,�; � (�c$.� "k ����� ,�; � =�&!.� � )k ����� ,�; � ?�!c.� ��� ������ � � � ���� � � ����� �� � ���� ����� � �� �56 (&&� 9�� ����� � ���� ��� ��������� ������ ,��� 9 �� ". �� 2�� �� 2 �� �� �� �� ��� � ��� ����� � � � � � ��� ������� �� �� �� � �� �� �� ��� ������ �� �� ��� �� �� �� �� ������ �� ������ 0���� �+-�� �� ����� ���� �� ������ �� � � �� �� �� � �� ��� ������ <�� �� ��0�� � � � �� ��� �� �� ��������� ����2��� �� �� �� ��� � �� �� � ����� (��� � ��� �� �� �/��� !cc">!ccd �� � � �� ���� ���� � ��� � �� �� ��� ��� �2� �� &�&! ������ 9�� � �� �� ��� 0�� � �� � "k �� )k � � � �� ��� � ��� ���� �� �� ��� �� ���� 0� � !cc">!ccd � � ������ � ������ � ����� �� i�0����� �� ������ ��� ������ ������ ������ � � ����� ��� ����� ��0 ������� �� � �� ��� �� �� ���� !c=" � !c?!� 0� � ����2��� � � �� � � ��� � �� �� ��� ��� �2� � �� ������ � , �� ���. ����� �� ���� ����h��� %�� ��������� ����� � ��� �� 2�� � ��� ��0 � � � �� ��� �� �� �� �� � � 0�� �� �� � ���� ������ ��� � �� �� � ������ �� �� �� <�� !k ��� �� �� ����2��� �� �� �� "$d ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 � �� �� �� �� ��� � � �� �� ��� ��� �2� �� &�&( ����� � ���� �� �� ��/ ��� �������� <�� ������ � � � �� ��� �� �� ���� �� ���� � ��� � ����� �� � ������� ���0��� � � ����� �� � ���� �� ���� 9�� ����2��� �� �� �� � �� �� �� � �� �� ��� ��� �2� � �� ������ � ,���� ���. ����� �� ��� �� ��� �� ��� 0� � )k ��� �� � ���� �� �� ��/ ��� �� ��� 0� � "k ��� � ,!c=">!c=d � !c?=>!cc! �� �� �/��� �� �.� �� ����� � � �� �� ��� � �� �� �� � ��� �2� ����� � �� ���� � � � � ��� ������� ��� ����� �� ��� ���� � �� �� � ��� � �� ����� ��� � ��� � �� �� ���� �� ���� � ����� �� ��� ���� � �� �� ��� � �� � f��� � �� � ��� ,!! �� !? ���� ���� �.� �� ���� � � ��� �� �� �� ����2��� �� � ���� � �� ���� � � ��� �� �� ����2��� �� �� �� ��� ���� �� � �� ��� �� �� � � � � �� ��� ������ 9��� 2 �� � � ��������� �� ��� ���� �� �� �� � � � f�� �� �� ��� � 0�� �� �� � � � �� � � � ������ <��� !cc" � !ccd� �� ����� � ����� �� 0� � �� ��� � �� �� �� ��� �� �������� ���� ���� �� �� 2�� "( �� �� ��h��� 9�� ����� �� �� �� � ����� ��� � ��� � �� �� �� ��� ������� �� !k ��� ������ � ����� �� �� ����� �� ����� � �� "k � )k ��� ������� 1���� � ���� ������ ��� � �� �� ��� ������ �� ������ � ��� � �� �� � ���� !cd= � !cc!� ��� � �� ������ ������ � � � ��� ������ �� 0��� ����� � � �� � � ( �� �� � ������ �� ���� ��� ������ , �� ���� �� ��� ����2��� � ��� !cc">!ccd �� � �� �� ��� ��� �2� � �� ������ � ����� �� ���� ����h��.� � �� � � � � � �/�� �� �� ��� �� ���� �� � � ��� ��� �� ��� ���� �� ������ ��0�� �+-�� ����� � ���� ������ �� 9��������� ���� � � ����� � �� �� �� ����� � � � � ���� �� �� �� 1���� ��� ���� ���� � � �� ��� � �� � � �� ��� � 0� � �� ��3�� � � ������ ����� �������� � ��� �� ,!cd).� � ������ �� � � ����� � � ���� ,!cc).� ��� ��7� ����� �������� � ������ ��� � � �� ��� � ���� ,����� :�;�.� � 0��� � �� � �� ��� � ���� ���������� �� �� ���� ���� 9���� ���� � ��� � �������� 0� � �/��� �� ������ ������ � � ����� � ������� ��� ����� ������ �� �� ���� ������ � ���� � �� �������� � �� � � ��� � 1�� ��� �� !cc">!ccd �� ���� 0��� ������ �� ���� �� ��� � ��� ��� ��7� ������ � �� �� ���� �� � ���� ������ � � ����� ����� ���� � � �� �� ���� � �!!��� �� ����������� 1� �� 2�� �� �� �� � �� ��� �� � �� � � � ����� � ����� �� �+-�� ���� ������ �� 1� ��� 2 � � �� !& ��� ���������� � � ����� ��� � �� � � � �� ��� ����� �/�� � ��0�� ����2��� � �� � �� �� � ���� �� �56 (&& � ��/� �� ����� � ����� �� � ����� ���� ���� 0� �� ���� � � �� �� � 0�� �� �� � ���� ������ ��� � �� �� ��� �2� ���� � �� ������� �� � �� � � � ���� � � ��� ���� �� �� �� �� ��� �2� �� � �� � � � ��� �� � � � �� ��� �� #� �� � � � � ����� 0� � ���� �� �� �� � � ���� � ����� �� ��� �� � ��� � ��� � �� ���� ��� � ��� ��� 0� � ��0 � � � �� ��� �� ����� ���� �� �� �� � ��� � �� 0� � ���� ������ �� ��� � �� �� � �� ������ �� �� �� i�0����� ��� � ��� ��� 0� � ������ � ���� � � � �� ��� �� �� �� ��� �� � �� � � � ����� 0� � ���0 �� � ����� �� ����� ��� ����� �� � ��� 9�� ����� � ��� ���� � � �� � � ��� ������� �� �� ���� �� � �� ��� �� �� ��h��� ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 "$= ���������� � ����� � �� ��� ����� � �� ��� <�������� �� ��� 5 � � � ��� �� 1� ,"&&!.� � ��� �� � �� ����� �� ��� �� �� � �� �� �� ���� ��������� "�������� &�,).� ))!>)$d� � ����� � �� ��� 5 � ����� #� ,!cc).� � � ��� ����� ��� �� ��� �f�� �� �� � ��� �� ���� � �� �� ������� ������� ��������� ,� � ����� %� !(c>!d?� �� �� 4�� 5 9 ����� *� ,!ccd .� #��� � �� �� � � �� � �������� � �� ��b � � �� � � ����� � ������ � ����q :�� ; ������� "����/ ,� � ���� ��� ! ������� %�,".� "!c>")!� �� �� 4�� 5 9 ����� *� ,!ccd .� %� �� � � �� � � �� ������ � � ������b �0� ��� � � 0�� ������ � ���� � �������� � ��� �� �� ��� ������� ,� � ����� &5� !&d=>!&=$� � ����� 1�� 5 j��� �� ,!cc".� +� �� � � �� �������2� �� �� �0 ��� �� �� ��� ���� < �� �� (��������� ,���� �.� =$>?&� ����� � � ��� l��� *�� 5 j��� �� ,!cc(.� �������� � ��� �� �� �� � ���� � ��� �� ��� � "����/ ��������� �� ����� 5,).� ?=c>c!c� ��� ���:� �� ��� �� ���� 8� ��� :� �� #�� 5 1�� ���� �� ,!cc&.� �� �� � ��� ��� �� � � �������� � ��� �� �� � ������� ���� �������� ��,".� (")>($=� �� ���� 8� �� ,!cc).� ��� �� �� ��� � �� �������� � �� � ��hh��b � =k ���� �� � ���� ��������� �������� "�������� �,".� !cc>"!d� ������ ��� �� �� ��� 5 1��� ���� *� ,"&&&.� <� � ������ ����� � �� � �������� � �� ��� ���� �������� ��,).� !)?(>!$!$� ��� ��� 6� i� ,!cd).� � ��� ������b # ��� ������ ��� �� �� �� ���� )�� ������ (��������� "����/� %� "$>$(� *�j� �� �� ��� ���������� ��� �������� j� i�� 5 1 ��� � #� �� ,!cc!.� 9�� � ��� � ��� � �� �� !c"cb +���� �� ���� �������� � �� ��� ���� ,� � ��� =��� ��� �+,).� d=(>=&&� *�j� �� �� ��� ���������� ��� �������� j� i�� 5 1 ��� � #� �� ,!cc".� �������� � �� � ������ �b � � ��� ��� �� �� ���� ��� �� �� ��� � ���� < �� �� (���������� +5,".� $d>()� +��� h � ;� ,!cc!.� 6���� � �� �� � � ��/ �� ��� "����/ ; ���������� ������� ��� ��� ������ +� c!>!&&� i ���������� 8�� j 6�� � #�� 5 1�h� � 9� ,!cc$.� 1� ����� �� ������ � ��� �� ���� � �� ��� �� <� ��� ,+��.� :�� ��������� ����>��� � �? ��� ���#��� ,��� )$(>)c=.� j��� �� 4� ��� ���������� ��� 5 9� ���� #� i� ,!cc&.� � �� ����b �������� � �� � �� � � �� ��� ���� ,� � ��� <������������ �,$.� !($>!d$� j��� �� 4� ��� ���������� ��� 5 9� ���� #� i� ,!cc!.� ��� �� �� ��� � �� �������� � �� � ��hh��� ���� �������� ��,!.� =(>!&c� 4 ������ �� 8� ,!cc(.� 9�� � ��� ��� �� �������� � �� � ������ � ������ ��� ���� < �� �� (���������� &+,$.� )">)?� %7�� ��� 4�� 5 *�0 �� +� �� ,"&&&.� 9�� ����� �� ��� ��������2� � ��/ � �� � � �������� � ��� �� �� ��b � ������� � ������ ��������� �������� "����/� @,).� "(c>"=(� 6� ���� �� ,!cc(.� �������� � �� � ����� � �� �� �b -����� �� � ��� 2 �� � � ��� ������ ����� ���� ��������� ,� � ����� %�� )$!>)=&� #��� ��� �� +�� 5 1���� �� �� �� ,"&&&.� 9�� � ���� �� �� � �� �������� � ��� �� �� � ������ �� ��������� �������� "����/� @,".� !=!>!?!� #������� �� 1� ,!cc?.� 9�� � �� �� � �������2� �� � �� �� �� �/�� ������ �� ������ ���� ��������� �������� "����/� �,".� c(>!&d� �� �� #� 1� ,!cc=.� %� ���� � �� � � � ����� ��� �������� � �� ��� ���� )��������� �,!.� ($>d!� "$? ��0� ������ � �� ��� 1 ��������� �������� "����/ +* 2&**+3 &%�4&�5 a closer look at trading strategies for u.s. equity closed-end investment companies introduction review of the literature data source description methodology results summary and conclusions references financial services review, 32(3) i volume 32 issue 3 from the editor if you are reading this introduction to the latest issue of financial services review (fsr), you are likely a member of the academy of financial services—the organization that publishes the journal. if you are not a member, you ought to be. as i am writing this introduction, the annual conference in columbus, ohio is in full swing, and what a conference it is. there truly is no better place to meet with colleagues who are interested in individual financial management issues. the sessions have been quite diverse, ranging from enhancing consumer well-being, implementing complex investing strategies, cross-border estate planning, and exploring the pedagogy of financial planning. your membership in the academy helps in building a strong conference and journal. i am excited to introduce the papers in this issue. the first paper was written by drs. shan lei and lu fan (cognitive ability and stock investment among chinese middle-aged and older population). in this paper, they examine the relationship between cognitive ability and stock investment among older adults living in china. using data released by the harmonized china health and retirement longitudinal study (charls), they observed a positive association between cognitive ability and stock investment decisions. this paper indicates that although it may feel awkward for a financial advisor to do so in practice, testing client math skills may be a way to indirectly describe a client's stock investment decisions. the second paper in this issue, written by drs. pierre-etienne pilote, emilio boulianne, and michel magnan (impact of the financial advisor on clients' financial outcomes: an integrative model), provides direct evidence regarding the impact financial advisors have in shaping client outcomes. using agency theory, trust theory, and the concept of knowledge, this team of researchers presents an integrative multi-theory model that illustrates how financial advisors impact client outcomes at each step of the financial planning advising process. what is of particular value for researchers is the model's causal mechanism that clarifies the complex processes involved when providing financial advice. in the third paper, drs. yi liu, tao guo, and yuanshan cheng (global perspectives on the determinants of older adults' subjective well-being: a comprehensive longitudinal study) provide an interesting perspective on the topic of well-being. rather than focus on a single country or a single sample, these researchers utilized international longitudinal surveys to examine the well-being of older adults. liu et al. took into consideration microand macro-level determinants of retirement well-being. what they found will be of interest to policymakers and financial advisors. they observed consistent variations in happiness levels across countries. for example, in the united states, they documented a positive relationship between age and subjective well-being; however, in europe, they noted the opposite—a negative association between age and well-being. if you have an interest in well-being, whether it is at the micro/household level or as a policymaker, you will want to dive into the implications section of this paper. this issue of fsr concludes with an important paper written by mr. zezhong e. zhang, dr. sherman d. hanna, and ms. lei xu (the effect of financial knowledge on workers' expectation of never retiring). anyone who regularly reads the wall street journal or personal finance blogs will have come across reports of individuals indicating a lack of willingness or ability to retire. this paper financial services review, 32(3) ii provides direct evidence regarding the veracity of statements made by individuals about their intention to retire. zhang et al. found the tendency to report the intention to "never retire" is related to objective financial knowledge, with more knowledgeable (regardless of how knowledge is measured [i.e., objectively or subjectively]) people being less likely to report an intention of never retiring. they also noted that those who exhibit overconfidence are more likely to give a "never retire" response when responding to a survey. anyone who provides retirement planning advice or provides guidance on retirement planning policy will find this paper to be a fascinating read. now, on to another topic. in my introduction to volume 32, issue 1, i discussed the use of google scholar as a tool to help with referencing. i received some criticism (via emails) about this. some readers found the use of this introductory space in the journal to discuss ways to enhance the readability of submissions inappropriate. well, these readers are not going to like what comes next. as fsr's reputation continues to improve, and the number of submissions continues to climb, manuscripts must be well written and presented in american psychological association (apa) 7th edition format (or as close to that as possible) to have a reasonably good reception among reviewers. this can be a challenge for those who are used to using harvard, mla, chicago, or another style manual. let's look at a quick shortcut to make the conversion from one style manual to another quick and easy by focusing on the reference list. here is how a paper from fsr is referenced using mla: grable, john, and ruth h. lytton. "financial risk tolerance revisited: the development of a risk assessment instrument." financial services review 8.3 (1999): 163-181. it could take hours to reformat a reference list from mla to apa, or i should say that at least until recently, it would have taken hours, but today, the conversion should take just a few minutes. here is what you can do. open chatgpt (or a similar ai program) and ask chatgpt (chat.openai.com) to "convert this reference list to apa." here is the exact output using the reference from above: grable, j., & lytton, r. h. (1999). financial risk tolerance revisited: the development of a risk assessment instrument. financial services review, 8(3), 163-181. i've tried this dozens of times with 95% accuracy. today, there is no excuse to submit a paper to this, or any other, journal with a reference list that is inconsistent with fsr's style guidelines. it is very important, however, to provide a disclosure statement (this can be in the form of a footnote) indicating that ai (i.e., in this case, chatgpt) was used to format the reference list. as long as you (the author) wrote the original reference list (i.e., it was not created by ai), it is permissible to use ai to bring your paper (i.e., your original work) into alignment with fsr's formatting style. of course, you should always disclose the use of ai, but i encourage authors to make their lives (and the lives of the editorial team) a bit easier through the appropriate use of ai. until next time, all the best, john e. grable, ph.d., cfp® editor pii: s1057-0810(97)90020-2 financial services review, 6(2): 77-96 copyright 0 1997 by jai press inc. issn: 1057-0810 all tights of reproduction in any form reserved. personal bankruptcy costs: their relevance and some estimates james s. ang ali m. fatemi the paper argues that there is a need for the formal treatment of personal bankruptcy costs in thefinance literature. the need arises out of the relevance of such costs to both corporate and personal financing decisions. we show that (a) personal bankruptcy costs (like personal taxes) are relevant to the corporate capital structure problem and that (b) differential bankruptcy costs across corporations and individuals can result in a clientele model of individual investment-borrowing decision which could lead to insti tutional arrangements designed to minimize combined bankruptcy costs. further, we develop a theory ofpersonal bankruptcy and a set of testable hypothesis with regard to their costs. some preliminary estimates of personal bankruptcy costs are reported which suggest that they are higher than corporate bankruptcy costs. there is also some evidence of economics of scale in personal bankruptcy costs. corporate bankruptcy costs have been the subject of considerable theoretical discussion (e.g., see diamond, 1994; harris & raviv, 1991; haugen, & senbet, 1978; kraus & litzen berger, 1973; morris, 1982; scott, 1976; white, 1989)) and some empirical measurement (e.g., see altman, 1984; ang, chua, & mcconnell, 1982; deis, guffey, & moore, 1995; franks & torous, 1989; guffey & moore, 1991; kalaba, 1984; warner, 1984; white, 1993). however, the literature has paid only scant attention to personal bankruptcies in general and personal bankruptcy costs in particular. credit research center’s consumer bankruptcy study (1982), durkin and elliehausen (1978) and stanley and girth (197 1) are examples of earlier studies which have dealt with personal bankruptcies. the more recent works include those of bhandari and hein (1993), buckley and brinig (1995), rooney (1996), simmons (1989), sofianos (1985) and sullivan, warren, and westerbrook (1994). most of those, however, have dealt with questions pertaining to the u.s. bankruptcy code and its fallouts. personal bankruptcy costs have been dealt with only on a limited basis despite the fact that they are important to all major areas of finance. more specifically: james s. ang l florida state university, tallahassee, fl 32306. ali m. fatemi l kansas state university, manhattan, ks 66506. 78 financial services review 6(2) 1997 1. personal bankruptcy costs, or more precisely, their expected value, affect the pricing of personal loans. the lenders, realizing that upon bankruptcy they would have to share the borrowers’ assets with third parties for the deadweight costs involved, will impose these expected costs (ex-ante) on the borrower. an assessment of the magnitude of personal bankruptcy costs is, thus, important in delineating these expected costs. furthermore, the presence of economies of scale in personal bankruptcy costs can partly explain why larger borrowers can obtain loans at lower rates than smaller borrowers can. indeed, in the absence of other differentials, such as lower investigation, securing and monitoring costs or lower risk of default, economies of scale in bankruptcy costs would be the only justification for such phenomenon. 2. insofar as personal investment behavior is concerned, the magnitude of personal bankruptcy costs relative to corporate bankruptcy costs is an important factor in deciding whether to invest in levered firms or to lever-up the holdings of unlevered firms’ securities. in the absence of tax effects, the margin of choice is between personal and corporate bankruptcy costs. with differential taxation, a clientele model may emerge that is dependent both on the individual’s tax bracket and on relative bankruptcy costs. furthermore, the existence of such a clientele together with the presence of economies of scale in bankruptcy costs may lead to institutional arrangements that are aimed at minimizing combined (business plus personal) bankruptcy costs. 3. personal bankruptcy costs are relevant to the theory of the capital structure of the firm in much the same way as personal taxes are. indeed, the interactions of personal vs. corporate bankruptcy costs and personal vs. corporate taxes may determine the capital structure of the corporate sector as well as that of an indi vidual firm. this study is a first step in addressing the question of personal bankruptcy costs. sec tion i discusses the relevance of personal bankruptcy costs and develops a clientele model of individual investment-borrowing decisions. section ii presents some insights on a the ory of personal bankruptcy and develops some testable hypothesis. section iii describes the data and the characteristics of costs for a sample of personal bankruptcy filings together with an empirical analysis of the nature of bankruptcy costs. section iv summarizes and concludes the paper. i. theoretical considerations modigliani and miller (1958), in their pioneering work on the theory of capital structure, prove that in perfect and complete markets the choice of the capital structure is inconse quential to the value of the firm. the proof rests on the argument that as long as investors can borrow or lend on their own account (on terms identical to those available to the firm), they can undo the effect of any changes in the firm’s capital structure. however, if the terms of borrowing (lending) differ, the financing decision may no longer be irrelevant. several factors can cause the terms to differ across the two groups. one such factor is taxes. modigliani and miller (1963) show that when only corporate taxes are considered, there personal bankruptcy costs 19 would be an advantage to corporate borrowing, and a comer solution of 100% debt is obtained. however, once personal taxes are considered as well (with a provision for a pro gressive income tax system where income from bonds is taxed at a higher rate than income from equity), miller (1977) shows that equilibrium would be achieved when the spendable income to the marginal investor is the same whether a dollar of pre-tax operating profits is distributed in the form of interest or equity income. formally, at the margin (m): (1 tp”) = (1 fc)( 1 tg”> (1) where tp = personal income tax on bonds; tg = the discounted effective capital gains tax; tc = corporate income tax. ’ miller, then, proceeds to show that under such an equilibrium there would exist an optimal capital structure for the corporate sector as a whole but none for individual firms. as such, bondholders in the zero and low income tax brackets are the beneficiaries of the tax deductibility of interest payments at the corporate level. formally, investor i whose mar ginal income tax bracket is such that: (1 tp’) > (1 t,)(l -t;> (2) would buy corporate bonds and earn the surplus f,,” $,‘. on the other hand, investor j whose income tax bracket is such that: (1 -t;> < (1 -t&(1 tgj) (3) would shun corporate bonds and may invest in either equity or tax-free municipal bonds depending on the implicit tax rate on the tax free vis-a-vis ti. a second factor that may cause the terms of borrowing to differ across the corporation investor groups is bankruptcy costs. in evaluating the effect of corporate bankruptcy costs, a consensus emerged that an optimum capital structure is reached where corporate tax sav ing is just offset by the present value of expected corporate bankruptcy costs at the margin (e.g., see taggart, 1982; kraus & litzenberger, 1973; scott, 1976; kim, 1978, etc.). this view is challenged by miller (1977), who based on the empirical evidence of warner (1984), argues that the expected corporate bankruptcy costs are not large enough to warrant such conclusions. further, haugen, and senbet (1978) argue that the only relevant corporate bankruptcy costs are the ones attributable to liquidation. thus, if one merely assumes that investors are rational and that assets are priced competitively, corporate bankruptcy costs are trivial and insignificant to the firm’s capital structure. however, altman (1984) argues that warner’s analysis suffers from a lack of proper measurement of expected corporate bankruptcy costs. he further argues that because warner’s study employs the data for a restricted sample of railroads, its results are nonrepresentative of the bankruptcy costs for other firms. more importantly, he argues that the relevant corporate bankruptcy costs are not limited to liquidation costs but include indirect costs: costs of lost managerial energies, costs of lost sales and profits, etc. based on his findings, that total corporate bankruptcy 80 financial services review 6(2) 1997 costs can exceed 20% of the value of the firm just prior to bankruptcy and from 11% to 17% 3 years prior to bankruptcy, he argues that corporate bankruptcy costs are nontrivial and that the choice of the capital is relevant to the value of the firm. castanias’ results (1983) tend to provide further support for the argument that corporate bankruptcy costs are nontrivial. he finds that firms in high failure lines of business tend to have less debt in their capital structure. this is consistent with the hypothesis that bankruptcy costs are large enough to induce firms to hold an optimum mix of debt and equity. disregarding the arguments about the magnitude of bankruptcy costs, deangelo and masulis (1980) show that starting from a miller equilibrium situation, the introduction of corporate bankruptcy costs (and other unresolved agency costs) causes the supply curve for corporate bonds to no longer be hor izontal and turn downward-sloping instead. the effect, then, is for an optimal capital struc ture to exist for the individual firms as well as for the corporate sector as a whole. diamond (1994) brings into focus the control role of debt and argues that within such a framework, bankruptcy costs become endogenous and sometimes negative. accordingly, capital struc ture would depend on the correlation between cash flows and profitability of new invest ments, as well as on taxes and bankruptcy costs. interestingly, all the controversy has been centered on corporate bankruptcy costs, their magnitude and relevance, with no attention paid to personal bankruptcy costs. how ever, disregarding these costs is tantamount to assuming that either personal bankruptcy costs are irrelevant or that they are greater than corporate bankruptcy costs for all investors. a. the relevance of personal bankruptcy costs to corporate capital structure the strongest arguments on the irrelevance of corporate bankruptcy costs are those forwarded by haugen and senbet (1978). can the same arguments be used to reason that personal bankruptcy costs are irrelevant? the thrust of haugen and senbet’s argument is that liquidation is an investment decision in the sense that liquidation is preferred when: net liquidation value > going concern value, where the former is the total value of assets when liquidated less the liquidation costs involved, including bankruptcy costs if the desired course of action calls for such filing. under such a scenario liquidation may be preferred even in the absence of debt and the event of bankruptcy. thus, bankruptcy costs are the costs associated with liquidation as an investment decision and are not affected by the amount of debt in the capital structure. therefore, the amount of debt will not affect the probability or the costs of bankruptcy. neither will it affect the net cash flows from liquidation. an important feature of this line of reasoning is that when there is a lack of agreement between the shareholders and the creditors, the former can buy out (take over) the latter, or vice versa, in order to liquidate if the net liquidation value exceeds the going concern value. even if one assumes that haugen and senbet are correct to conclude that corporate bankruptcy costs are irrelevant (see titman, 1984, for an argument otherwise), the follow ing factors preclude the possibility of drawing such conclusions for personal bankruptcy costs: i. informational asymmetry. this is a more serious problem in personal lending as personal assets tend to be more difficult to value and monitor. the lender will persod bankruptcy costs 81 have a more difficult time in deciding when to liquidate personal assets that are pledged, that is, there are costs from premature liquidation. 2. moral hazard. when facing imminent b~ptcy, the personal borrower may decide to spend the to-be-allocated share of the lender (in part or wholly) before declaring bankruptcy. 3. the lender’s limited ability to garnish future income of the borrower implies that for debtors who have limited assets-in-place but high expected future income, liquidation is almost never the preferred alternative. this holds because value at liquidation (a function of in-place assets) is always less than the value of the person’s lifetime income (a function of future income). the limited abil ity to garnish future income may indeed create an incentive for these individuals to declare personal bankruptcy, especially if they can utilize the exemptions provided in the bankruptcy code.3 these exemptions provide for the discharge of debts and retention of many assets and can move an individual from negative to a positive net worth position.4 overall, the use of debt may change the proba bility of bankruptcy and, considering the moral hazard problem, value at liqui dation as well. 4. human capital is a non-tradeable asset and as such haugen and senbet’s take over argument can not be evoked. even if it were possible to take over a person due to his diminished incentive to work (e.g., see rea, 1984), the value of his future income will be significantly less than the corresponding value if his future income were 100% self-owned. 5. finally, in contrast to the corporate case, there are no circumstances under which a rational person would self liquidate at zero debt. therefore, the assumption of the irrelevance of personal bankruptcy costs is not a tenable one even if one accepts the argument that corporate bankruptcy costs are irrelevant. figure 1 depicts miller’s model. if we now allow for the presence of corporate bank ruptcy costs, the supply curve for corporate debt will no longer remain horizontal and will become downward sloping. thus, as shown in figure 2, a lower aggregate level of debt is achieved. more irn~~~tly, as long as corporate ba~ptcy cost functions are non-uni form, there would be an interior solution for each firm. if we next introduce the possibility of personal borrowing (home made leverage) and further allow for the presence of personal bankruptcy costs, the demand curve for corporate debt will shift to the right. as a result, a higher level of aggregate debt is achieved. this is shown in figure 3. at the firm and at the individual level the situations are more complicated. since bc is firm-specific and bp is individual-speci~c, there will be matching among firms and individuals such that a clien tele model may emerge. b. a clientele model of investment-borrowing an alternative assumption associated with disregarding the personal bankruptcy costs is that they are greater than corporate bankruptcy costs for all investors at all debt levels. such an assumption embodies implications which run counter to the observable phenomenon of margin borrowing, that is, personal leverage to buy corporate shares. 82 financial services review 6(2) 1997 i i i i i i i i i i dl figure 1. miller’s model. allowing personal bankruptcy costs to be lower than corporate bankruptcy costs for some investors would alleviate this problem. formally, we consider four cases. first, in the absence of tax effects, an investor whose personal bankruptcy costs, b,,, is greater than the entire range of corporate bankruptcy costs (as a function of debt sizes): d2 di figure 2. the combined effects of personal taxes, corporate taxes, and corporate bankruptcy costs. personal bankruptcy costs 83 d2 d3 “i figure 3. the combined effects of personal and corporate taxes, and personal and corporate bankruptcy costs. would invest in levered firms and refrain from personal leverage. one the other hand, if: b,(s) < b,(s), v s (6) the investor would buy the shares of unlevered firms and lever the position up via margin. other possibilities are: b,(s) > b&/h s] c s, (7) where the investor would invest in levered firms of debt size s1 or smaller, and b,(s) < b&z), s2 c s, (8) where he/she would avoid investment in levered firms of debt size s2 or larger and would instead invest in unlevered firms via personal leverage. combining differential taxes and bankruptcy costs we can arrive at a general clientele model of individual investment-borrowing decision. to this end, consider an investor with a universe of four investment vehicles to choose from: 1. a tax-exempt bond; 2. a taxable corporate bond; 3. the equity of an unlevered firm; 4. the equity of a levered firm, 84 financial services review 6(2) 1997 insofar as the tax regime is concerned, assume that corporate profits after interest payments are taxed at the fc rate before any distributions to the shareholders. distributions to the share holders can take the form of dividends or capital gains. the former is taxed as ordinary income at the rate tp which is not lower than the rate tg at which the latter is taxed: (tp 2 t&. the investor may combine his investment in any of the four investment alternatives with borrowing on personal account in which case the interest paid would be deductible from his ordinary income. the rate of interest paid by any borrower, be it the levered firm or the individual investor, is determined by a set of market-wide conditions and a set of borrower specific factors. market-wide conditions translate into the aggregate supply of and the aggregate demand for loanable funds. ignoring bankruptcy costs for the moment and starting with a market-clearing rate of interest for tax-exempt risk-free bonds, the rate on tax-exempt risky bonds will include a default risk premium such that the certainty equivalent rate of interest, ro, is equal across all such bonds. in order to induce the lenders to purchase taxable bonds, the borrowers will have to offer sufficiently higher rates to pro vide the lenders an after-tax rate of return at least equal to r,. the borrowers will do so as long as the tax advantage of borrowing makes the effective cost of loans less than or equal to r,. at equilibrium, then, the marginal lender is in the same tax bracket as the marginal borrower. if we now assume, as miller (1977) does, that the marginal borrower is the levered firm without access to other-than-interest tax shields, the equilibrium rate of inter est will be rj( 1 tj. if the levered firm can utilize other tax-shielding mechanisms like the investment tax credit, depreciation, etc., deangelo and masulis (1980) show that the equi librium rate of interest will be rj( i t”) where t” < tc. allowing now for bankruptcies to be costly, the interest rate that is charged to a bor rower will include a second component, b, to compensate the lender for the expected bankruptcy costs: b, for the corporate borrower and bp for the individual borrower. this premium is a function of borrower-specific factors like the riskiness of the venture and the borrower’s degree of leverage. to leave the lender with an after-tax compensation no less than the zero expected bankruptcy costs situation, this premium too has to be grossed up by the tax rate applicable to the marginal lender. therefore, the rate of interest paid by the borrower is rj( 1 t”) in the absence of bankruptcy costs and (r. + b,)l( 1 t”) for the corporate borrower and (r. + bj(1 t”) for the individual borrower when bankruptcies are costly. returning to the investor’s problem, his choices of after-tax rate of return are: r, for tax-exempt riskless bonds r, + b, for tax-exempt risky bonds (rj( 1 p))( 1 tp) for taxable riskless bonds ((rr, + b,)l( 1 p))( 1 tp) for taxable corporate bonds the investor will be indifferent between the taxables and the tax-exempts if his tax rate is equal to that of the marginal lender. he will prefer taxables if t,, < p and will prefer tax exempts if tp > p. additionally, regardless of the choice, the investment may be combined with personal leverage if the after-tax cost of personal borrowing ((r. + b,)l( 1 t”))( 1 tp) is less than the after-tax return from lending ((r. + b,)l(l t”))(l fp). it is readily apparent that the critical variables are b, and b,,, that is, personal leverage is preferred if bp < b,. thus, it is personal bankruptcy costs 85 table 1 a clientele model of individual investment-borrowing decisions bp > b,(s) tp > 1, + ts typo i tp < tc + t&, type iii type i: buys shares of levered firms; no personal borrowing type ii: buys shares of unlevered firms via personal leverage type iii: buys bonds; no personal borrowing type iv: buys bonds via personal leverage b,, < b,(s) type ii type iv possible for an individual to be a lender and a borrower at the same time if his expected bankruptcy costs are sufficiently smaller than corporate bankruptcy costs. consider now the alternatives of investing in the equity of an unlevered firm and that of a levered firm. without loss of generality, let us assume that the payout is zero or alter natively, that both dividends and capital gains are taxed at the same rate (t,). it can then be shown that a tax-and bankruptcy-induced clientele model emerges. such a model would compare the differential, after tax returns of each of the four alternatives relative to the other three. bypassing the details of the derivations, we offer the summary in table 1. it is clear that the individual’s choice of the investment medium is dependent on a how his/her tax bracket compares to the corporate tax rate and the marginal investor’s tax rate as well as how his/her bankruptcy costs compare to the corporate bankruptcy costs. ii. a primer on the theory of personal bankruptcy the preceding section illustrates the importance of personal bankruptcy costs to both the investment and the financing decision. before proceeding to report some estimates of these costs, we present a discussion of the some of the differences between personal and corporate bankruptcies. we also present a few testable hypothesis about personal bank ruptcy cost. given that corporate bankruptcies and their costs have been the focus of much debate, it would be helpful to review the process of corporate bankruptcy first. the essential fea tures of corporate bankruptcy (harris & raviv, 1993; john, 1993) can be listed as follows: a failure to pay either the interest or the principal of an obligation, or the viola tion of certain loan covenants, can trigger a corporate bankruptcy. upon such an event, all assets of the corporate firm become available for takeover by the cred itor(s). the limited liability feature of the corporate form prevents the creditor(s) from gaining access to the non-corporate assets of the shareholders. no legal exemptions are granted for the benefit of the residual claimant(s). exceptions consist of some deviations from the absolute priority rule that vary from case to case. financial services review 6(2) 1997 4. at bankruptcy, the business may either be liquidated or be allowed to continue to operate. 5. the creditors bear all of the ex-post bankruptcy costs. (these are usually only partially offset by the ex-ante bankruptcy costs assessed.) these features have two important implications for the process of corporate bank ruptcy. one is that the corporate firm (representing the shareholders/owners or the manag ers as their agents) has no incentive to voluntarily file for bankruptcy. this is due to the fact that when the value of the firm is less than the value of the obligations against it, the wealth position of the residual claimants is equivalent to the value of an unexercised out of the money call option. the residual claimant has no reason to exercise such an option given that the exercise will net the holder a current value of zero. keeping the option alive is pref erable, given that the probability of recovery at some future date is always non-negative. indeed, as long as the owners/managers remain the decision makers for the operations of the entity, they can act to increase the value of the option by increasing the riskiness of the assets. alternatively, we can consider that portion of the value of the equity which is attrib utable to the limited liability feature of the corporate form. this is a put option, which can not have a negative value and can only enhance the value of the equity. the other important implication is that the creditors, in their interest to preserve assets, would want to initiate the bankruptcy proceeding as soon as possible. this follows from the observation, above, that allowing the corporate firm to continue operations could amount to a process of wasting assets in place. accordingly, the creditors would want to assume ownership of all assets as soon as possible. they can then decide whether to liquidate or to continue the operations. with ownership, they position themselves to receive all the poten tial gains from access to assets in place. simultaneously, they would prevent the owners/ shareholders from exploiting risk-shifting opportunities which could enhance value only for themselves. therefore, we can summarize that with corporate bankruptcies: (a) the debtors will never initiate a bankruptcy proceeding, and (b) it is almost always the creditors who initiate the proceedings.5 personal bankruptcies differs from the corporate ones in several important ways: 1. in the event of bankruptcy, not all assets of the individual can be taken over by the creditors. some assets in place (e.g., the principal place of residence, vehi cle, tools, etc.) and the individual’s future income stream are exempt from take over. it is possible, therefore, that these exemptions would preserve a significant portion of assets in place. indeed, exemptions granted by the bankruptcy law and those granted by the states could leave the bankrupt individual significant residual assets following a filing. 2. unlike the corporate case, under certain circumstances, it would be optimal for the individual to initiate bankruptcy proceedings. these arise from: l a desire to secure the individual ownership of the exempt assets, given that in the absence of bankruptcy the exempt assets can dissipate as payment to the creditors. following this line of reasoning, it can be hypothesized that the asset level at which bankruptcy declaration is triggered increases with the level of exemptions. personal bankruptcy costs 87 l a need to facilitate additional borrowing at some future date. in such situa tions, an individual’s decision to file for bankruptcy is analogous to a sover eign country’s willingness to work out its loan(s) in default with the lenders involved. the motivation behind such a decision is the debtor’s perceived need for future borrowing. l an effort to take advantage of the presence of a mandatory cooling off period between bankruptcy filings: an early filing increases the present value of the gains from consumption (made possible by new debt) prior to a second bank ruptcy filing. l an attempt to relax restrictions on consumption: the pressure to satisfy debt obligations could compel individuals to limit spending below the subsistence level. settlement through bankruptcy proceedings may ease or remove such restrictions6 3. when filing for personal bankruptcy, the individuals can choose to give up either part of their assets in place (net of exemptions) or part of their future incomes (typically, for no more than 3 years). the former can be done through a chapter 7 filing and the latter through a chapter 13 filing. accordingly, the individual filer can choose the “least cost” alternative among the two: those who are assets-in-place poor but are future-income rich (e.g., medical school students with non-student loans) would prefer chapter 7. those with low expected future incomes but high assets-in-place (e.g., a person with a principal place of residence that has substantially appreciated in value) will choose chap ter 13. personal bankruptcies are, therefore, subject to a potential moral hazard problem which the corporate ones are not subject to.7 4. there are other moral hazard problems as well: individuals with low levels of asset-in-place (i.e., below the exemption level) may actually incur debt to increase current consumption and to increase their holdings of exempt assets. further, given the aggravated level of informational asymmetry with respect to the valuation of personal assets, an individual’s misrepresentations, made to take advantage of the exemption limit, will have a higher probability of not being detected. 5. in personal bankruptcies, when the value of the asset is near or slightly below the exemption limit, the individual (as the debtor) will be responsible for the ex post costs of bankruptcy. thus, the magnitude of bankruptcy costs does matter to the individual and to his or her decision as to whether to declare bankruptcy. it follows, then, that higher bankruptcy costs may delay the bankruptcy filing if such costs exceeds the necessary assets needed to keep the creditors at bay (i.e., partial debt payment). the personal bankruptcy process can, thus, be summarized as one that gives the indi vidual the choice as to the best alternative for retaining the highest level of wealth. within this framework, the individual would compare his/her expected wealth after bankruptcy to the current wealth to decide whether or not to file. the former is measured as the sum of exempt assets, net of ex-post bankruptcy costs, and the present value of all future income. the latter is measured as the current assets in place plus the present value of all future incomes net of a portion to be paid out for a specified number of years. additionally, the 88 financial services review 6(2) 1997 incentive to declare bankruptcy is often with the individual not with the creditors, as it is in the corporate case. these procedural differences can lead to certain testable hypothesis. to arrive at these, we assume that an individual facing financial distress has fewer resources available to him/ her than does a corporation in distress. the differential access to information and legal advice which this assumption embodies leads to the following hypothesis: hl: when compared to corporations, individuals may wait too long to file for bank ruptcy. as a result, assets-in-place can dwindle to a level below the triggering point for the preservation of exempt assets. h2: when compared to corporations, individuals are at a disadvantage in finding ways to minimize their direct bankruptcy costs. as a result, ex-post personal bankruptcy costs are higher than the corporate ones. h3: the lower level of legal sophistication on the part of near-bankrupt individuals tends to prevent them from appropriately taking advantage of the economic incentives built into the bankruptcy code. in what follows we provide some direct evidence in support of these hypothesists. iii. the empirical results and discussion our main objectives in this section are threefold: to provide some preliminary results on the magnitude of personal bankruptcy costs, to empirically determine if these costs are non trivial and to compare them to business bankruptcy costs. the data is comprised of bank ruptcies filed over the six-year period 1977-1982 with the bankruptcy court, u.s. district court, middle district of florida. it consists of 167 personal bankruptcy petitions filed with the court. of these 112 (67%) may be labeled as nominal asset cases because the bank ruptcy costs involved exceeded the value of the assets involved.* accordingly, the other 55 will be labeled as asset cases. the observation that two-thirds of those cases result in the total loss of assets provides direct evidence in support of our hl hypothesis and indirect evidence in support of the h3 hypothesis. however, direct evidence available elsewhere does provide additional strong support for the last hypothesis. consider, for example, bhandari and weiss (1993) results that the bankruptcy code does not appear to have had a significant effect on the rate of personal bankruptcy filings; buckley and brining’s (1995) findings that differences in state filing rates are not attributable to legal or common economic variables; and sullivan, warren and westerbrook’s (1994) finding that debtors in high-exemption and low-exemption states filed for bankruptcy and selected chapter 13 in similar proportions. these findings all support the notion of lack of sophistication on the part of individuals in financial distress. the breakdown of personal bankruptcy cases by the range of the amount of claims is shown in table 2. the greatest concentration of these cases is in the less than $20,000 range of claims:9 39% of all cases and 53% of asset cases fall in this range. also, it appears that the ratio of total administrative costs to total debt declines as the amount of claims increases, but no such appearance is present for the ratio of administrative costs to net proceeds realized. the former may be regarded as an estimate of bp. when all cases are considered the magnitude of these costs is not, in any of the ranges of claims, less than 9 2 t a b l e 2 s c ha ra ct er is tic s of p er so na l b an kr up tc y c as es h b r an ge of th e b * d ol lu r a m ou nt . t ot al a dm in is tr uf iv e n et p ro ce ed s 2 of c la im s pe r c as e n um be r of c as es t ot al d eb t (t d ) c os ts (a c ) r ea li ze d (n p r ) a c /t d a u n p r ‘c r 0” k a : a ll c as es s o -$ 9, 99 9 io 19 ,9 99 20 29 ,9 99 30 -3 9. 99 9 40 .4 9, 99 9 50 .5 9, 99 9 60 69 ,9 99 70 -7 9. 99 9 80 89 ,9 99 90 99 ,9 99 lo o19 9, 99 9 20 0, 00 0 & m or e t ot al a ve ra ge 35 30 14 17 6 10 5 8 3 6 21 12 16 7 $ 20 5, 02 9 $ 13 ,8 92 $ 26 ,4 19 42 5, 67 4 17 ,5 62 31 ,1 25 34 7, 81 4 5, 69 6 6, 45 6 59 0, 57 3 7, 14 8 17 ,8 41 25 8, 05 3 2, 99 2 6, 25 7 55 8, 55 9 9, 22 9 11 ,3 50 32 1, 77 2 4, 93 1 14 ,7 82 59 7, 77 3 8, 85 2 20 ,3 1 0 25 6, 96 0 1, 80 4 3, 19 7 57 3, 81 2 4, 86 2 4, 86 2 2, 82 1 ,4 99 16 ,9 84 26 ,6 11 $ 22 ,1 31 ,4 73 19 ,5 73 51 ,1 73 $ 29 ,0 88 ,8 91 $1 13 ,1 93 $2 20 ,6 43 $ 2, 42 4, 07 4 $9 ,4 33 $ 18 ,3 87 .0 67 7 .5 25 8 .0 41 2 .5 64 2 .0 16 3 .8 82 2 .0 12 1 .4 00 6 .0 11 5 ,4 78 1 .0 16 5 .8 13 1 .0 15 3 .3 33 5 .0 14 8 .4 35 8 .0 07 0 .5 64 2 .0 07 8 .9 31 7 .0 06 0 .6 36 7 .0 00 8 .3 82 4 ,0 18 ,5 79 b : a ss et c as es o nl y o -$ 9, 99 9 lo 19 ,9 99 20 29 ,9 99 30 39 ,9 99 9 $4 4, 99 7 9, 44 2 $ 21 ,9 69 .0 29 8 .4 29 7 10 13 3, 40 8 12 ,3 07 25 ,8 70 .0 92 2 .4 75 7 4 10 6, 73 6 2, 89 1 3, 65 1 .0 27 0 ,7 91 s 6 20 4, 02 9 5, 15 4 15 ,3 47 .0 25 2 .3 25 2 g r 90 hnancial services review 6(2) 1997 personal bankruptcy costs 91 33% and is on average 58%. when only asset cases are considered, the corresponding fig ures are 29% and 45%.” these are all significantly (beyond the .ol level) greater than zero and apparently much higher than the corresponding figures for business bankruptcies reported elsewhere; for example, an average of 5.3% in warner (1984), 7.5% in ang, chua and mcconnell (1982), or 20% in stanley and girth (1971). these results are, there fore, consistent with the predictions of our second hypothesis. notice, however, that due to the presence of exempt assets (see footnote 3) there is an asymmetry of personal bank ruptcy costs between the borrower and the lender. as a result, although the preceding fig ures may be representative of the costs from the point of view of the lender, they overestimate the costs to the borrower because the value of exempt assets are not included in the net proceeds realized. table 3 provides a breakdown of the component costs of bankruptcy and their relative importance. in approximately half (47%) of the cases, the trustee’s attorney fees amounted to over 50% of all costs and as such is the most important of all component costs. the sec ond most important of all component costs is the trustee’s commissions, which in 38% of all cases amounted to over 50% of all costs.tl since some costs were incurred for only a few cases, no concrete statement can be made regarding the least important component costs. however, for the frequently incurred costs, the referees’ salary and expenses ranks as the least important component cost: for 74% of cases this item amounts to less than 10% of all costs. table 4 provides the data on distributions to claimants. claims are broken down into the priority, secured and unsecured categories. among these, priority claimholders are the ones with the highest percentage of claims paid: approximately 20%. interestingly, secured claimholders fare worse than unsecured claimholders: 0.19% of claims paid to the secured and 0.46% paid to the unsecured claimholders. ‘* however, the distribution to the secured claimholders may be understated if some properties are abandoned to their claimants by the trustee. yet, with consideration given to the possibility of such understatements, the per centage of claims satisfied is extremely low which explains why in a few cases some claim ants did not even bother to file claims.13 this is also suggestive of the ineffectiveness of monitoring debt in the case of individuals. finally, in a similar vein to the conclusions of green and shoven (1983, p. 50), the low percentage of claims paid suggests that the pros table 3 the relative importance of component costs number of crises for which the component cost as a percentage of total administrative cost falls in the runge of u-io% i l-20% 2/-300/o 3140% 41-50% >50% the trustee’s commissions 29 28 referees salary and expenses 124 32 auctioneer’s fees 2 3 attorney fees for the trustee 0 9 the trustee’s other expenses 44 7 attorney fees for bank 0 0 receiver’s expenses 0 1 the appraiser’s fees 9 6 rental expenses 2 0 21 2 2 3 0 11 11 64 1 0 0 0 3 1 8 3 1 1 0 0 1 2 0 0 0 0 0 0 2 1 1 0 92 financial services review 6(2) 1997 table 4 distributions to the claimants total amount ofclaims amount paid percentage of claims paid priority secured unsecured $320,756 $17,947 5.59% 17,780,927 34,999* 0.19% 10,987,208 5 1,240 0.46% case secured claim a~unt paid percentage claim paid 1. $297 $297 100 2. 6,056 6,056 100 3. 450 450 100 4. 468 2,468 100 _. 5 113,844 25,728 22.6 total $124,115 $34,999 28.4 pect of bankruptcy may make personal borrowing so expensive that some portion of the economically desirable demand for personal borrowing not be met.r4 to gain insight into the nature of personal bankruptcy costs, we regressed the total administrative costs of bankruptcy (ac) on net proceeds realized (npr). the estimated equation is as follows: ac = 262 + 0.31 (npr) (7.0) (30.67) r2 = 0.85 (9) accordingly, the administrative costs of personal bankruptcies amount to approximately $262 per filing plus 3 1% of net proceeds realized. when nominal cases are excluded, costs amount to $396 per filing plus 29% of net proceeds.t5 to investigate the possible presence of economies of scale in bankrnptcy costs, the fol lowing quadratic equation was estimated: ac = 120.26 + .49(npr) 0.67 x lo-’ (npr)2 (4.02) (29.3) (11.9) i& = .92 (10) accordingly, it can be concluded that personal bankruptcy costs are a quadratic function of the net proceeds realized and that there are economies of scale in such costsi iv. summary and conclusions corporate bankruptcy costs have been the subject of much theoretical debate and empirical measurement. personal bankruptcy costs on the other hand have not received much atten personal bankruptcy costs 93 tion. this paper is designed to show the need for further investigation into both the rele vance and the magnitude of personal bankruptcy costs. it is argued that we can neither safely assume that personal bankruptcy costs are irrelevant nor can we assume that they are insignificant. we show that the interaction of personal vs. business bankruptcy costs and personal vs. corporate taxes determine the capital structure of the corporate sector as well as that of the individual firm. considering bankruptcy costs (personal and corporate) and taxes (both personal and corporate), we develop a general clientele model of individual investment-borrowing decisions. within such a model the choice of the media (equity vs. debt) and the mode (use vs. non-use of leverage) in investments is determined by the inter action of an individual’s tax bracket and his/her bankruptcy costs, furthermore, within such a framework, levered mutual funds may become the providers of low cost leverage to the individuals with relatively high bankruptcy costs and or high monitoring costs. the incentive to create these funds is then dependent on the magnitude of the spread between the average cost of personal debt and the average cost of personal debt. this spread is a function of the difference between the average expected bankruptcy and monitoring costs for corporate and personal debt, which our exploratory estimates of personal bankruptcy costs indicate may be rather large. these estimates indicate that personal bankruptcy costs can consume as much as 45% of the net proceeds realized on the average and as such are higher than corporate bankruptcy costs documented in the literature. the evidence also suggests that personal bankruptcy costs are a quadratic function of the net proceeds real ized and that there are economies of scale in such costs. this should provide further impe tus for institutional arrangements that minimize combined bankruptcy costs. acknowledgment: the authors would like to thank professor karen eilers lahey for her insightful comments on an earlier version of this paper. notes 1. note that miller sets tg = 0. 2. note that at the margin(m) bcm = 4”‘. this is comparable to the case where only taxes are considered and at equilibrium (1 t,,“) = (1 rcm)( 1 tg”). 3. under chapter 7 of the u.s. 1978 bankruptcy code, the first $7,500 of equity in an indi vidual’s home and $3,050 in other specified assets are exempt from bankruptcy proceedings. how ever, the level of exemptions vary from one state to another, and the petitioner has the choice between federal and state exemptions. see frank and torous (1992) for a comparison of uk and u.s. bankruptcy procedures. gropp, scholz and white (1997) discuss how personal bankruptcy exemptions redistribute credit toward borrowers with high assets. 4. it is from this perspective that shepard (1984) refers to bankruptcy as a rational strategy by many debtors to maximize their wealth. 5. an interesting exception to this rule is the case of john manville which declared bank ruptcy in order to limit the asbestos victims’ claims. the strategy, designed to protect the interest of the shareholders, proved disastrous for the firm. 6. note that these are post-bankruptcy benefits and are often ignored in theoretical discussions. 7. business week, in a recent commentary on personal bankruptcies, reports that about 30% of debtors file under chapter 13 (gutner, 1996). 94 financial services review 6(2) 1997 8. by comparison, business bankruptcy filings during this period consist of 34 cases, only 11 (32%) of which were nominal asset cases. in what follows we will report the highlights of the results for personal cases in the body of the paper and report the results for business bankruptcies in footnotes. the detailed results are, however, available from the authors. 9. by contrast the greatest concentration of business bankruptcy cases is in the more than $150,000 range of claims. 10. the corresponding figures for business bankruptcy cases (b,) are 14% (minimum) and 20% (average) for all cases and 12% and 17% for asset cases only. 11. for business bankruptcy cases the trustee’s attorney fees ranked as the most important component cost. the trustee’s commissions and accountants’ fees tied as the second most important component. 12. in business bankruptcy cases secured and unsecured claimholders fare almost as poorly: 4.6% and 4.3% of claims paid, respectively. 13. this may be due to an important difference between corporate and personal bankruptcies: when a firm is in distress, the lender would consider (a) the liquidation value of all assets of the firm and (b) the capitalized value of all future cash flows and choose a course of action. in case of indi viduals, however, the relevant figures are (a) the liquidation value minus exemptions and (b) the cur rent wealth in-place without regard for future wages (i.e., no slavery, no garnishment of wages, etc.). 14. however, the default rate on consumer loans is very low (l-2%). 15. in comparison, the administrative costs of business bankruptcies include larger fixed costs and lower fixed costs as a percentage of net proceeds realized: for all cases the variable costs amount to $1,996 and the variable costs to 20%. the corresponding figures for asset cases are $2,848 and 19%. 16. the same holds true for our sample of business bankruptcy cases, where the equation is: ac = 635.55 + .38(npr) .97 x 10-6(npr)2 ii2 = .8 (1.01) (9.2) (4.6) references altman, e.i. (1984). a further empirical investigation of the bankruptcy cost question. journal of finance, 39, 1067-1089. ang, j.s., chua, j.h., & mcconnell, j.j. (1982). the administrative costs of corporate bankruptcy: a note. journal of finance, 37,219-226. bhandari, j.s., & weiss, l.a. (1993). the increasing bankruptcy tiling rate: an historical analysis. journal of american bankruptcy law journal, 67, 1 12. buckley, f.h., & brinig, m.f. (1995). the bankruptcy puzzle. george mason university school of law and economics working paper no. 95-006. castanias, r. (1983). bankruptcy risk and optimal capital structure. journal of finance, 38, 1617 1635. credit research center. (1982). consumer bankruptcy study, ~01s. l-2 [monographs no. 23 and 241. credit research center, krannert graduate school of management, purdue university. deangelo, h., & masulis, r. (1980) optimal capital structure under corporate and personal taxa tion. journal of financial economics, 8, 3-30. deis, d.r., guffey, d.m., & moore, w.t. (1995). further evidence on the relationship between bankruptcy costs and firm size. journal of business and economics, 34,69-79. diamond, d.w. (1994). corporate capital structure: the control roles of bank and public debt with taxes and costly bankruptcy. federal reserve bank of richmond economic quarterly, 80, 11-37. personal bankruptcy costs 95 durkin, t.a., & elliehausen, g.e. (1978). 1977 consumer credit survey. washington, dc: board of governors of the federal reserve system. franks, j., & torous, w.n. (1989). an empirical investigation of u.s. firms in reorganization. jour nal of finance, 44,147-769. franks, j., & torous, w.n. (1992). lessons from a comparison of u.s. and uk insolvency codes. oxford review of economic policy, 8(3) 70-82. green, j., & shoven, j. (1983). the effects of financing opportunities and bankruptcy on entrepre neurial risk bearing. in j. roden (ed.), entrepreneurship (pp. 41-59). lexington, ma: lex ington books. gropp, r., scholz, j.k., & white, m. (in press). personal bankruptcy and credit supply and demand. quarterly journal of economics. guffey, d.m., & moore, w.t. (1991). direct bankruptcy costs: evidence from the trucking indus try. financial review, 26, 223-235. gutner, t. (1996, september 9). personal bankruptcy: how to slow the stampede. business week, p. 46. haugen, r.a., & senbet, l.w. (1978). the insignificance of bankruptcy costs to the theory of opti mal capital structure. journal of finance, 33, 383-393. harris, m., & raviv, a. (1991). the theory of capital structure. journal of finance, 46, 297-356. harris, m., & raviv, a. (1993). the design ofbankruptcyprocedures. chicago: center for research in security prices, university of chicago. john, k. (1993). managing financing distress and valuing distressed securities: a survey and a research agenda. financial management, 22, 60-78. kabala, robert e., et al. (1984). estimation of implicit bankruptcy costs. journal of finance, 39, 629-642. kim, e.h. (1978). a mean-variance theory of optimal capital structure and corporate debt capacity. journal of finance, 33,45-63. kraus, a., & litzenberger, r.h. (1973) a state preference model of optimal capital structure. jour nal of finance, 28,91 l-922. miller, m. (1977). debt and taxes. journal of finance, 32,261-275. modigliani, f., & miller, m. (1958). the cost of capital, corporation finance and the theory of investment. american economic review, 47,261-267. modigliani, f., & miller, m. (1963). corporate income taxes and the cost of capital: a correction. american economic review, 53,433-463. morris, j. (1982). taxes, bankruptcy costs and the existence of optimal capital structure. journal of financial research, 5,285-300. rea, s.a., jr. (1984). arm-breaking, consumer credit and personal bankruptcy. economic inquiry, 22, 188-208. rooney, j.f. (1996). u.s. bankruptcy filings top million-a-year pace for first time. chicago daily law bulletin, 142, l-2. scott, j. (1976). a theory of capital structures. the bell journal of economics and management sci ence, 7, 33-54. shepard, l. (1984). personal failures and the bankruptcy reform act of 1978. journal of law and economics, 27,419-437. simmons, p. (1989). bad luck and fixed costs in personal bankruptcies. journal of economic, 99(394), 92-107. sofianos, g. (1985). personal borrowing and bankruptcy: the role of the exemption level (working paper 340,50). new york: solomon brothers center, new york university. stanley, d.t., & girth, m. (1971). bankruptcy, problem, process, reform. washington, dc: the brookings institution. sullivan, t.a., warren, e., & westerbrook, w. (1994). consumer debtors ten years later: a tinan cial comparison of consumer bankrupts 1981-1991. journal of american bankruptcy law journal, 68, l-4. 96 financial services review 6(2) 1997 taggart, r.a. (1982). secular patterns in the financing of u.s. corporations. in b. friedman (ed.), corporate capifal structure in the united states (pp. 13-80). chicago: university of chicago press. titman, s. (1984). the effect of capital structure on the firm’s liquidation decision. journal of financial economics, 13, 137152. warner, j. (1984). bankruptcy costs: some evidence. journal offinance, 32, 337-348. white, m. (1989). the corporate bankruptcy decision. journal of economic perspectives, 3, 129 151. white, m. (1993). the costs of corporate bankruptcy: a u.s.-european comparison (research forum on international economics discussion paper 346,40). university of michigan. pii: s1057-0810(99)00038-4 from the editor karen eilers lahey volume 8, number 3, provides several articles that readers will find both interesting and challenging in terms of their content and conclusions. it is hoped that they will provide you with the necessary incentive to supply additional research in these very important areas. the first article by yaw badu, kenneth daniels, and daniel salandro is entitled “an empirical analysis of differences in black and white asset and liability combinations.” the authors utilize data from the 1992 survey of consumer finances, which is produced by the federal reserve board in conjunction with the department of the treasury. they find that there are significant differences in risk aversion between black households and white households in terms of their choice of assets and the cost of liabilities. white households have significantly greater net worth and financial assets than black households. they raise some very disturbing questions in their conclusions concerning the differences in wealth accumulation between the two groups. the second article, “racial differences in investor decision making” by michael gutter, jonathan fox, and catherine montalto uses the 1995 survey of consumer finances (scf) to examine possible racial differences in risky asset ownership. the authors focus is on the effects of socioeconomic, financial, and attitudinal variables on risky asset ownership between black and white households. the two studies both find statistical differences in the descriptive data by race. this study suggests that the difference is due to racial differences in the individual determinants of risky asset ownership, not to race in and of itself. john grable and ruth lytton develop an instrument to measure financial risk-tolerance in their article entitled, “financial risk tolerance revisited: the development of a risk assessment instrument.” in light of the first two articles’ identification of differences in risk tolerance, an effort to develop a testable instrument of individual attitudes towards risk is important. hopefully, others will be willing to administer it to a wide variety of individuals to determine its reliability and validity on multiple randomly selected samples. a relatively recent development in the real estate brokerage industry is the use of affinity programs. danielle lewis, randy anderson, and leonard zumpano examine the efficiency of firms that use these programs in their article, “an analysis of affinity programs: the case of real estate brokerage participation.” their findings will be of interest to those in the industry and individuals who must select a brokerage firm. financial services review 8 (1999) v–vi 1057-0810/00/$ – see front matter © 2000 elsevier science inc. all rights reserved. pii: s1057-0810(00)00038-4 the last article, “international mutual fund returns and federal reserve policy” is written by robert johnson, gerald buetow, and gerald jensen. their findings provide a possible explanation for home country bias of institutions and individuals when selecting financial assets. vi k.e. lahey / financial services review 8 (1999) v–vi financial services review, 33(2) 93 immigration law enforcement and immigrant homeownership efthymia antonoudi,1 genti kostandini,2 and hanna lim3 abstract we use the american community survey microdata and employ difference-in-differences (did) models to examine how local immigration law enforcement, through 287(g) agreements and the secure communities program, impacts homeownership among different demographic groups. the findings indicate that 287(g) agreements significantly reduce the likelihood of homeownership, particularly among hispanics without a college education and u.s. citizenship, with effects most pronounced in states lacking e-verify mandates. the secure communities program exhibits more nuanced effects, initially showing positive impacts for specific hispanic populations; however, these results are not robust to pre-trend analyses. additional factors such as length of u.s. residence, english proficiency, age, and household income strongly influence immigrant homeownership outcomes, underscoring the complex interplay between policy enforcement and socio-economic assimilation. the results highlight unintended economic consequences of immigration enforcement policies, suggesting important considerations for housing stability, financial security, and integration policies aimed at immigrant and broader community well-being. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation antonoudi, e., kostandini, g., & lim, h. (2025). immigration law enforcement and immigrant homeownership. financial services review, 33(2), 93-123. introduction according to the 2022 american community survey (u.s. census bureau, 2024), about fourteen percent of the u.s. population is foreign-born—nearly triple the 1970 percentage. as of 2022, 77% of u.s. immigrants had some form of legal status, such as legally admitted immigrants, refugees, and temporary residents, and nearly half (49%) had become u.s. citizens (institute of migration research, 2019). in addition, as of 2023, about 29.9 million immigrants were employed, a higher workforce participation rate (64.2%) than that of people born in the united states (59.5%) (bureau of labor statistics, 2024). labor shortages in key industries are often 1 corresponding author (eanton@uga.edu), university of georgia, athens, georgia, usa. 2 university of georgia, athens, georgia, usa. 3 california state university, fullerton, fullerton, california, usa. driven by a lack of skilled workers and an insufficient overall labor supply. immigrants play a crucial role in addressing these shortages, particularly in physically demanding and specialized sectors where native-born workers are less likely to participate. by supplementing the workforce, immigrants help sustain economic productivity and mitigate the effects of demographic shifts, such as an aging population and declining labor force participation (sherman et al., 2019). their contributions are especially significant in industries that face persistent labor gaps, ensuring stability in essential economic sectors. furthermore, they contribute to business creation, with about a quarter (24%) of all new https://creativecommons.org/licenses/by-nc/4.0/ mailto:eanton@uga.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 34(1) 122 u.s. businesses founded by immigrants (chodavadia et al., 2024). immigrants also significantly impact u.s. markets, including the housing market. not only did immigrants address the housing shortage due to a sharp decline in homebuilding after the 2008 financial crisis by providing 30% of all construction workers nationwide, but they also fueled housing demand with a desire to transition to homeownership (national immigration forum, 2024). homeownership provides significant financial benefits to families as it helps them build wealth and financial security (goodman & mayer, 2018). also, homeownership has been considered an important indicator of assimilation for immigrants, which shows how successfully immigrants integrate into their host country's economic, social, and political life (sinning, 2010). more importantly, homeownership represents an outcome of the long-term economic progress of immigrant families and a key factor in their long-term financial security (sinning, 2010). however, there may be significant obstacles for many immigrants in buying a home for reasons specific to them, such as cultural differences in attitudes towards mortgage finance, limited access to loan services, and moving intentions (de coulon & wolff, 2010; rodríguez‐planas, 2018; schoenholtz, 2005). there is a significant wealth gap between native and immigrant families (flores morales, 2019), so research on homeownership for immigrants is expected to improve immigrant families' financial security, narrowing the existing wealth gap. researchers have examined how political context affects immigrant assimilation and adaptation to the united states. for example, researchers have investigated how state-level political contexts for immigrants impact educational outcomes for their children (filindra et al., 2011), how political climates affect immigrant naturalization decisions (cort, 2012), and how those decisions are associated with increased fears of deportation for unauthorized immigrants (amuedo-dorantes et al., 2013). in addition, researchers (painter & yu, 2014) examined how the housing bubble 4 e-verify is an internet-based system that employers use to verify their employees' eligibility to work legally in the u.s. the system compares information an employer enters with records from affected the immigrants' homeownership rates in the united states. however, to our knowledge, no study has considered the impact of immigration policies on homeownership, except for fu (2017) who examined the effect of e-verify4 policies on the homeownership of immigrants. this study attempts to fill this gap by examining the impact of local immigration law enforcement on homeownership. it explores the effects of the 287(g) program and the secure communities program on the homeownership of immigrant populations. as part of the illegal immigration reform and immigrant responsibility act of 1996, the 287(g) program allows local police officers to initiate deportation processes for unauthorized immigrants (amuedo-dorantes et al., 2018), and the secure communities program allows local agencies to run the fingerprints of those arrested through immigration databases to check their immigration status and criminal histories (miles & cox, 2014). the 287(g) agreements and secure communities programs differ from e-verify by directly targeting undocumented immigrants instead of employers. immigration enforcement policies have progressively molded the socio-economic outcomes of immigrant populations in the united states. the 287(g) program and the secure communities program are two key policies that have had profound implications for immigrant communities. both are rooted in local enforcement partnerships with federal immigration authorities, and even though they primarily target undocumented immigrants, they can create ripple effects for broader immigrant populations and u.s.-born citizens (kohli et al., 2011). both programs, while aiming at enhancing public safety, have been criticized for their disproportionate impact on immigrants, including increasing deportations and creating a climate of fear that may influence long-term socio-economic decisions such as homeownership (amuedo-dorantes et al., 2018; kohli et al., 2011). the u.s. department of homeland security and the social security administration to confirm employment eligibility. antonoudi et al. 123 homeownership is a vital indicator of longterm financial security and economic assimilation for immigrants and can be directly or indirectly affected by these immigration policies. these policies can impact the immigrants' motivation and ability to invest in a home by shifting their local social and political climate (dewind & kasinitz, 1997; sinning, 2010). researchers have explored the effects of economic, cultural, and individual factors on immigrant homeownership (amuedo-dorantes & mundra, 2013; borjas, 2002; chakrabarty et al., 2019). however, little is known about how immigration policies such as the 287(g) and secure communities programs precisely affect homeownership outcomes. this study examines how the 287(g) and secure communities programs affect homeownership rates among hispanics, non-citizens, and u.s.born citizens, exploring whether these policies create financial instability for immigrant households. this study aims to fill this critical gap by examining the relationship between these programs and homeownership, with a focus on spatial (policy environment), exposure (years in the united states), and behavioral (english proficiency) factors (xie & greenman, 2011). this study is important because homeownership is a key pathway to financial security (goodman & mayer, 2018) and wealth accumulation, particularly for immigrant and minority communities. understanding the effects of immigration enforcement on homeownership provides insights into broader financial stability issues and access to credit. the findings have implications for financial institutions, policymakers, and housing market stakeholders by highlighting potential homeownership and financial inclusion barriers. lenders can use these insights to serve immigrant communities better, while policymakers can assess the economic consequences of immigration policies. this research discusses access to financial services and strategies to support homeownership opportunities. research question how do 287(g) agreements and secure communities programs impact homeownership probabilities across different population groups? literature review determinants of homeownership among immigrants immigrant homeownership is often seen as a significant milestone in the economic integration process, signaling long-term stability and assisting in wealth accumulation. scholars have long studied the individual and household characteristics that drive this outcome. for example, because homeownership rates among immigrant populations in the united states are generally lower than those of native-born americans, researchers examined the factors that may contribute to this gap, including individuallevel characteristics such as national origin (borjas, 2002), immigration status (amuedodorantes & mundra, 2013), and race and ethnicity (chakrabarty et al., 2019; mundra & uwaifo oyelere, 2018). they also considered communityand national-level external factors such as political climate (allen & ishizawa, 2015), economic environments, and housing market conditions (yu & myers, 2010). researchers highlighted that financial risk tolerance, shaped by macroeconomic and demographic factors, is critical in major financial decisions, including homeownership (kuzniak & grable, 2017). additionally, disparities in financial advice-seeking behavior, influenced by income and financial knowledge, may hinder immigrants' access to homeownership resources (qing & reiter, 2024). psychological and demographic variables, such as self-efficacy and financial literacy, further influence immigrants' ability to navigate complex financial decisions like purchasing a home (kehiaian et al., 2021). borjas (2002) noted significant variances in immigrant households' homeownership rates across national origin groups. for example, in 1990, italian immigrants had the highest homeownership rate (78%) in the united states, and immigrants from the dominican republic had the lowest (14.2%). race and ethnicity and birthplace networks (i.e., social networks of immigrants with the same origin) are significant factors in homeownership (chakrabarty et al., 2019; mundra & uwaifo oyelere, 2018), with some groups (e.g., chinese, indian, and korean natives and immigrants) making headway in homeownership relative to non-hispanic white financial services review, 34(1) 122 natives, while other groups, such as black natives, mexican natives, and cuban immigrants, seeing their rates decline (chakrabarty et al., 2019). using data from the current population survey, mundra and oyelere (2018) found that citizenship affected immigrant homeownership more during the recession than before. the authors also found a decreased impact of length of stay in the united states on homeownership probability during 2007-2012 compared to prior years. similarly, sinning (2010) showed a significant gap in homeownership rates between natives and immigrants in germany, and immigrant homeownership rates converge with those of natives over time, contingent on factors like legal status, length of stay, and access to employment and resources. these studies lay the groundwork for understanding the barriers and pathways immigrants face in achieving homeownership. political and policy contexts affecting homeownership much of the literature has overlooked the role of local and federal policy environments, particularly the influence of immigration enforcement measures in shaping homeownership outcomes. political and economic environments can also be critical in immigrants' homeownership decisions. amnesty policies providing temporary legal status to eligible undocumented immigrants, according to the immigration reform and control act (irca) of 1986, increased the homeownership rate of eligible immigrants by around four percentage points compared to ineligible immigrants (sharpe, 2020). immigrant families in states that adopted everify mandates are less likely to own or buy homes in those states (fu, 2017), and unfavorable state-level political climates towards immigrants are negatively associated with the probability of homeownership among asian and latino immigrants who had moved in the past year (allen & ishizawa, 2015). heightened enforcement deters immigrants, particularly those who may lack legal status, from settling in areas with active policy measures (amuedo-dorantes & mundra, 2013). while their study primarily focuses on mobility and avoidance behaviors, it provides important insights into how enforcement policies disrupt immigrants’ ability to invest in stable, long-term housing. rugh and hall (2016) offer a direct examination of the relationship between immigration enforcement and housing stability by examining the effects of 287(g) agreements on hispanic foreclosure rates, demonstrating how deportation removes wage-earning adults from mixed-status households, increasing the likelihood of foreclosure. their analysis leverages county-level data and a quasiexperimental approach, revealing that implementing 287(g) agreements led to significantly higher foreclosure rates in affected counties. this research is particularly relevant to the current study as it identifies a precise mechanism – income loss due to deportation – through which immigration enforcement undermines homeownership. moreover, their findings situate 287(g) agreements as an important policy environment exacerbating racial disparities in housing outcomes. increased immigration enforcement has been linked to a rise in poverty among households with u.s.-born children, which indirectly affects homeownership by reducing economic resources available for home purchases (amuedo-dorantes & arenas-arroyo, 2021). participation in remittance activities negatively impacts homeownership among immigrants, as financial resources are diverted to support families abroad rather than being invested in home purchases (kuuire et al., 2016). this body of research reveals how immigration enforcement policies intersect with economic and social factors to shape immigrant homeownership outcomes. prior studies have focused on individual and household-level determinants, but fewer have examined the broader spatial and policy contexts that influence these decisions. this study addresses this gap by investigating how implementing 287(g) agreements and the secure communities program affects homeownership among immigrants. situating this analysis within the current literature contributes to a deeper understanding of the long-term economic impacts of immigration policies on immigrant communities. immigration enforcement has a significant negative impact on homeownership rates, particularly among latino communities, by exacerbating financial instability and increasing foreclosure rates (rugh & hall, 2016). immigrant status, legal antonoudi et al. 123 policies, regional variations, and economic factors are crucial in shaping homeownership trajectories. conceptual framework immigrant research has used different terminologies to measure how well the immigrant population settles into the host country, such as integration, assimilation, and acculturation. there was an effort to distinguish these terminologies from each other. for example, the model of acculturation was defined as "the process of cultural and psychological change that follows intercultural contact" and viewed integration and assimilation as different sectors depending on how people seek to acculturate (berry et al., 2006). integration indicates that people adopt the host culture and retain the heritage culture. in contrast, assimilation indicates that people weigh more on involvement with the host society and have less interest in maintaining the heritage culture (berry et al., 2006). despite this effort, previous research kept using those terminologies interchangeably. we used "assimilation" in this study but followed the original research's terminologies. the immigrants’ or their children’s educational, health, and financial outcomes can measure the level of the immigrants’ integration, acculturation, or assimilation. owning a home is an important indicator of an immigrant’s assimilation into the united states, not just because americans view homeownership as a way to build wealth but also because it indicates an intention to settle in the community and host country. while earlier theories on immigrants' assimilation assumed that assimilation is an integral part of the pathway to the american middle class for immigrants, recent studies recognized the diverse experiences of assimilation and emphasized the importance of the social context (greenman & xie, 2008; xie & greenman, 2011). segmented assimilation theory presents heterogeneous assimilation patterns depending on the interactions between immigrants and the host society (dewind & kasinitz, 1997). according to the segmented assimilation theory, american society is highly diverse and segmented, and immigrants may take divergent assimilation paths depending on the local social context in which they are embedded (xie & greenman, 2011). the factors related to segmented assimilation are categorized into spatial, exposure, and behavioral factors. spatial factors represent the intensity with which immigrant families are exposed to the host society locally (xie & greenman, 2011). exposure factors represent the length of time spent and exposure in the host society, and behavioral factors explain the individual-level differences in assimilation (xie & greenman, 2011). in the present study, the two measures of the spatial factor—county-level 287(g) implementation and secure communities implementation—represent how favorable the community’s political environment is toward immigrants. the present study also utilizes years of stay as an exposure factor and english proficiency as a behavioral factor to explain how immigrants are assimilated into the host society—buying a house. these factors are assumed to lower barriers to complicated financial transactions. this study examines these spatial, exposure, and behavioral factors associated with immigrant homeownership. specifically, it is hypothesized that whether the 287(g) program or/and the secure communities program were implemented in the county of residence is associated with the probability of immigrants' homeownership. in addition, it is hypothesized that length of stay in the united states and english proficiency are associated with the probability of immigrants’ homeownership. immigration enforcement policies immigration enforcement, in the form of police-based measures implemented by local or state police and employment-based measures that establish additional employer requirements (amuedo-dorantes & arenas-arroyo, 2021), can impact homeownership probabilities. the 287(g) agreements, secure communities program participation, and omnibus immigration enforcement agreements are all police-based measures. in contrast, e-verify is an employment-based measure, and its implementation varies widely by state. the passage of the illegal immigration reform and immigrant responsibility act of 1996 included what is widely known as the 287(g) program. the first 287(g) agreement was signed in 2002, shortly after the 9/11 terrorist attacks. these agreements were implemented based on either the “task force model,” the “jail financial services review, 34(1) 122 model,” or a combination of both, the “hybrid model.” according to a report by the american immigration council (2021), the first agreements adopted the jail model, but more task force model agreements were implemented starting in 2006. under the jail model, police officers could check to see if the person arrested for other law violations has permission to live in the united states and can initiate deportation for those who do not. under the task force model, officers with appropriate training can interrogate unauthorized immigrants, ask for paperwork, arrest without a warrant, and initiate deportation processes (american immigration council, 2021). because the obama administration discontinued the task force model, the hybrid model, the last of the three agreements, expired on december 31, 2012 (kolker, 2021), and the department of homeland security stopped renewing expired agreements (pham, 2018). between 2012 and 2016, only six agreements were renewed— although 287(g) agreements increased from 35 to 150 during the trump administration from january 2017 to september 2020 (kolker, 2021). however, these agreements focused on two new models, the "jail enforcement model" and the "warrant service officer model," which differed from the jail and task force models. in this study, we focus our analysis on the 287(g) agreements from 2005-2012. figure 1 shows the counties that signed 287(g) agreements each year from 2005 until 2012 (charlton & kostandini, 2021). figure 1. implementation of county 287(g) policies over time as of 2012 (source: charlton & kostandini, 2021) note: 287(g) counties are in red. amuedo-dorantes et al. (2018) found that the 287(g) mandates directly affect immigrant populations through increased deportations and indirectly by increasing the fear of being targeted for removal based on race. kostandini et al. (2014) noted that county 287(g) mandates reduced the supply of unauthorized immigrant workers and county aggregate agricultural expenditures, farm incomes, and vegetable production. while several studies have found that immigration policies reduce the number of undocumented immigrants in adopting jurisdictions (bohn et al., 2014; kostandini et al., 2014; luo et al., 2018; watson, 2013), research on their effect on wages and labor participation of undocumented immigrants generally indicates that u.s. immigration laws in the last two decades have not generated the improved labor outcomes for citizen workers that they were intended to deliver. for example, orrenius and zavodny (2015) showed that everify mandates reduce average hourly earnings among likely unauthorized male mexican immigrants but increase labor force participation among likely unauthorized female mexican immigrants. amuedo-dorantes and bansak (2012) showed that e-verify mandates reduce the employment likelihood of unauthorized male and female workers but have mixed effects on wages, and a decade later, east et al. (2022) found that the secure communities program negatively impacts citizens' employment in middle to high-skill occupations. the implementation of the secure communities (sc) program started in march 2008 after 287(g) policies had already been implemented in several counties. the sc program significantly reduced the authority that 287(g) agreements provided to local agencies to perform tasks such as detaining and initiating deportation procedures in place of immigration and customs enforcement (ice) officers. under the sc program, police officers can run the fingerprints of those arrested against the federal bureau of investigation (fbi) database and the department of homeland security (dhs) database to check their immigration status and criminal history. if these fingerprint checks reveal that someone is unlawfully present in the united states or otherwise subject to removal, then ice officers, not local law enforcement agents, take action. because the sc program was completely activated antonoudi et al. 123 nationwide by january 2013 (miles & cox, 2014), the present study focuses on the years before 2013. methodology data and sample we use microdata from the american community survey (acs) by the united states census bureau from the ipums database. acs surveys households and produces nationally representative information on the population’s social, economic, housing, and demographic characteristics yearly (u.s. census bureau, 2021). specifically, we use the 2005-2013 acs data and focus on the head of household. we examine the effect of county-level 287(g) agreements and the secure communities (sc) program on the homeownership of all hispanics, hispanics without a college education and u.s. citizenship (hereinafter "hwcc"), and u.s.-born citizens. we focus on these three groups for two reasons: (1) even though hwcc immigrants are directly affected, all hispanics may also be affected since police checks may target them; (2) u.s.born citizens living in jurisdictions with 287(g) agreements may benefit from hwcc workers leaving these jurisdictions, which may make it easier for them to find a house to purchase and become homeowners. because acs does not provide information on whether an immigrant is undocumented, we rely on previous literature to focus on the immigrant population—hispanics without a college education and u.s. citizenship (hwcc) who are at higher risk of lacking legal status and most impacted by immigrant law enforcement. we use two ways to identify those immigrants from the acs data that are commonly used in the literature (amuedodorantes & arenas-arroyo, 2021; amuedo‐ dorantes & bansak, 2014; bohn et al., 2014; orrenius & zavodny, 2015; passel & cohn, 2010). the first, hispanics from mexico without a college education and u.s. 5 a total of 289 pumas contained counties that passed 287(g) legislation at some point during the period of the analysis and the rest (1,812 pumas) had no counties with 287(g) legislation. from the 289 pumas that had 287(g) legislation, 271 entirely consisted of counties (or part of counties) that passed 287(g) legislation, and 18 pumas contained a mix of counties that passed 287(g) legislation and citizenship (hereinafter "hwcc1"), include individuals who emigrated from mexico with a high school diploma or less and are not naturalized u.s. citizens. previous literature (orrenius & zavodny, 2015; passel & cohn, 2010) focused on this group because, although not all immigrants in this group have undocumented immigration status, a high proportion belongs to this category. the second definition for immigrants used in separate models as a robustness check includes those who are non-citizens, of hispanic origin (which includes immigrants from el salvador, guatemala, honduras, and other latin american countries), and have a high school diploma or less (hereinafter "hwcc2"). we should note that acs did not provide county-level information before 2005. beginning that year, it includes county information if the county’s population exceeds 65,000 and provides public use micro data areas (pumas) for each respondent. pumas are areas with a minimum of 100,000 residents and do not cross state lines. some counties contain several pumas, and some pumas are made of several small counties. however, the 287(g) and sc programs are implemented at the county level. as a result, it is difficult to determine whether individuals in pumas, including those in several counties, are subject to the immigration policy. following kostandini et al. (2014), we exclude pumas from multiple counties with at least one program county.5 empirical approach we use difference-in-differences (did) models to identify associations between the variables selected over time and compare the differential effect of immigration law enforcement in the jurisdictions that implemented the 287(g) agreements with those that did not. the did approach is well-suited for this study because comparing changes in homeownership rates before and after policy implementation between treatment and control groups isolates the effect counties that did not pass 287(g) legislation. the 18 pumas that contained a mix of counties were dropped from the primary analysis. however, we ran robustness checks that included all 289 pumas containing at least one (or part of one) 287(g) county, and the main results did not change. the results are available from the authors upon request. financial services review, 34(1) 122 of 287(g) agreements and secure communities programs from broader economic trends that may influence homeownership. this method is commonly used in policy evaluation and is particularly effective in settings where a policy is implemented in some locations but not others, allowing for a natural comparison. additionally, including time and location fixed effects helps control for baseline differences between areas and broader macroeconomic conditions that could affect housing markets. we estimate the effect of the agreements on immigrant homeownership by comparing the average change over time in the outcome variable for the treatment group and the average change over time for the control group. given that the dependent variable (homeownership) is binary (1 = homeowner, 0 = non-homeowner), we estimate the following logit model to assess the effect of 287(g) and secure communities programs: pr(yi,p,t = 1) = f(β0 + β1immigpolp,t + β2xi,t + β3zp,t + γp + ηt + εi,p,t) this specification uses a logistic transformation, which is more appropriate than ols regression when the outcome is a dummy variable. in addition to the raw logit coefficients, we report marginal effects, allowing for a more straightforward interpretation of how immigration enforcement policies affect the probability of homeownership. we identify the effects of county-level immigration policies (i.e., 287(g) agreements and the sc program) on the probability of homeownership. let yi,p,t be the outcome of interest on the household head i in puma p and year t. we regress the dependent variable, which is a dummy variable equal to one if the household head owns a house and zero otherwise, on an indicator variable immigpolc,t equal to one if the head of the household is located in puma p that had a 287(g) policy in year t. we then run the same model for the sc programs where the immigpolc,t is equal to one if the individual is located in a puma with an sc program in year t. we control for a vector of individual and family characteristics xi,t, which includes exposure and behavioral factors from the segmented assimilation theory and control variables. following previous studies, we used the length of stay in the united states (in years) as an exposure factor and english proficiency (does not speak english, poor english, good english, very good english) as a behavioral factor. we control for age, number of children, level of education (a bachelor’s degree or not), and household income (logged). we also control for a vector of time-variant puma characteristics zp,t, which includes indicator variables for location in an e-verify state and 287(g) state. we control for puma fixed effects γc and year fixed effects ηt. ei,p,t is the error term. while 287(g) agreements and secure communities programs were implemented as separate immigration enforcement policies, there was some overlap in their adoption. some counties implemented only one of the two programs, while others adopted both. secure communities was introduced in march 2008 and expanded nationwide, eventually covering all jurisdictions by 2013, whereas 287(g) agreements at the county level started in 2005 and were implemented selectively based on agreements between local and federal authorities. to account for this overlap, we control for both policies' presence, ensuring that each enforcement measure's estimated effects are not conflated. we include an indicator variable for secure communities when estimating the effects of 287(g) agreements and vice versa, allowing us to isolate the independent effect of each policy on homeownership outcomes. this approach ensures that the estimates reflect the distinct impact of each program while accounting for jurisdictions that may have implemented both policies. other researchers have used 287(g) agreements (e.g., charlton and kostandini, 2021; kostandini et al., 2014) and the sc program (e.g., miles and cox, 2014) in did frameworks and provided additional discussions on policy exogeneity and other robustness checks. results table 1 presents the summary statistics of our sample from 2005-2013 acs data on homeownership and other characteristics included in the empirical model. the homeownership rate in the total sample was 67.1%. about 85% of the sample were u.s.born, 9% were hispanics, and 2% were hwcc1. the homeownership rate of each group was 69.1%, 51.5%, and 39.1%, antonoudi et al. 123 respectively. on average, the study's sample was about 43 years old, had less than one child, and had a household income of about $77,000. about a third of the sample (33.9%) had at least a bachelor's degree. among those who were not born in the united states, their average years in the u.s. was about 20, and the majority speak english well (22.3%) or very well (34.8%). table 1. summary statistics of main variables mean std. dev. n u.s.-born 0.848 0.359 6,841,478 hispanics 0.087 0.282 6,841,478 hwcc1 0.023 0.151 6,841,478 homeownership 0.671 0.470 6,841,478 homeownership among the u.s. born 0.691 0.462 5,803,208 homeownership among hispanics 0.515 0.500 593,781 homeownership among hwcc1 0.391 0.488 160,251 age 42.959 10.416 6,841,478 number of children 0.193 0.502 6,841,478 household income 77,067.374 77,097.301 6,841,478 bachelor’s degree or higher 0.339 0.473 6,841,478 post 287(g) county 0.079 0.269 6,841,478 post-secure communities county 0.178 0.383 6,841,478 post-e-verify state 0.038 0.190 6,841,478 years in the u.s. among the non-u.s.-born 20.464 12.716 1,038,270 no english among the non-u.s.-born 0.056 0.229 1,038,270 poor english among the non-u.s.-born 0.165 0.371 1,038,270 good english among the non-u.s.-born 0.223 0.416 1,038,270 very good english among the non-u.s.-born 0.348 0.476 1,038,270 speaks only english non-u.s.-born 0.209 0.407 1,038,270 note: data are from the 2005-2013 american community survey. age is provided in years, household income is in u.s. dollars, and all other variables are dummy variables. model 1: the effect of 287(g) agreements on homeownership we start by examining homeownership using only residents of states with 287(g) agreements as controls (columns 1–3) and residents in states without e-verify (columns 4–6). these results are presented in table 2, which shows the estimates of the did model in equation (1). as mentioned, the impact of 287(g) agreements on homeownership is examined using logit models within a difference-in-differences framework as specified in equation (1). results in table 2 and the respective marginal effects presented in table 3 indicate a statistically significant negative relationship between 287(g) agreements and homeownership rates for hispanics, hwcc1 (hispanics without college education and u.s. citizenship), and u.s.-born citizens. the estimated effect of 287(g) agreements on homeownership probability for hispanics is a decline of 5.6 percentage points. for hwcc1, the effect is even more significant, with a reduction of 7.4 percentage points, suggesting that these agreements disproportionately affect immigrants with lower educational attainment. the most considerable observed effect is among u.s.-born citizens, where the estimated decline in homeownership probability is 7.9 percentage points. financial services review, 34(1) 122 table 2. logit regression results (the effect of 287(g) agreements on home ownership using only residents of states with 287(g) agreements at the county and state level (columns 1, 2, and 3) as controls and residents in states without e-verify (columns 4, 5 and 6)) hisp hwcc1 u.s.-born hispe hwcc1 e u.s.-borne (1) (2) (3) (4) (5) (6) post 287g county -.287*** -.386*** -.447*** -.183*** -.416*** -.556*** (0.067) (0.085) (0.052) (0.067) (0.084) (0.051) years in the united states .033*** (0.001) .048*** (0.002) .028*** (0.001) .051*** (0.002) poor english .223*** (0.003) .106*** (0.004) .231*** (0.037) 0.111*** (0.037) good english .531*** (0.043) .338*** (0.044) .531*** (0.043) .354*** (0.044) very good english .42*** (0.044) .271*** (0.049) .37*** (0.045) .285*** (0.048) only english .171*** .049 .211*** .062 (0.057) (0.069) (0.058) (0.067) age .056*** .043*** .071*** .054*** .042*** .071*** (0.0008) (0.001) (0.00004) (0.0008) (0.001) (0.0004) number of children .194*** .148*** .355*** .214*** .146*** .398*** (0.011) (0.015) (0.009) (0.01) (0.015) (0.008) bachelor’s degree or higher .272*** (0.021) .278*** (0.016) .218*** (0.02) .184*** (0.019) log of household income .894*** .618*** .964*** .931*** .622*** 1.012*** (0.018) (0.028) (0.008) (0.018) (0.027) (0.01) u.s.-born .762*** (0.032) .668*** (0.035) obs. 412,660 115,833 2,784,784 453,081 121,335 4,348,766 note: data are from the 2005-2013 american community survey (acs). logistic regressions apply individual weights provided by the acs and include year-fixed effects. robust standard errors are clustered at the puma level. the outcomes variable is a dummy variable indicating homeownership. logistic regressions control for sc as well as e-verify and state level 287(g) agreements in columns (1-3) and state level 287(g) agreements in columns 4-6. *, **, *** denote significance levels at the 10, 5, and 1 percent levels, respectively. the results remain consistent when focusing on individuals in states without e-verify as controls. the probability of homeownership for hispanics in these states declines by 3.6 percentage points, while the effect for hwcc1 in these states is a decrease of 7.9 percentage points. for u.s.-born citizens in states without e-verify, the probability of homeownership decreases by 9.5 percentage points. antonoudi et al. 123 table 3. marginal effects of 287(g) agreements on home ownership group dy/dx std. err. z p-value 95% confidence interval hispanic -.0562 .0130 -4.32 < .001 [-0.0817, -0.0307] hwcc1 -.0735 .0159 -4.61 < .001 [-0.1047, -0.0423] u.s.-born -.0785 .0091 -8.64 < .001 [-0.0964, -0.0607] hispanics e -.0359 .0132 -2.72 .007 [-0.0617, -0.0100] hwcc1 e -.0793 .0159 -4.99 < .001 [-0.1105, -0.0481] u.s.-born e -.0952 .0087 -10.92 < .001 [-0.1123, -0.0781] note. dy/dx represents the marginal effects after the implementation of 287(g) agreements. standard errors are based on the delta method. p-values are reported to three decimal places, with < .001 indicating high statistical significance. confidence intervals are reported at the 95% level. a leads-and-lags model tests the assumption of parallel trends, with results presented in figure 2. the reference period is the year of adoption, which is set to 0 in the figure. the findings confirm that homeownership trends were similar between treatment and control groups before implementing 287(g) agreements for hispanics and hwcci. after policy implementation, the adverse effects on homeownership persist and align with the main estimates for these two groups. however, the pre-trends do not support the findings in table 2 (columns 3 and 6), suggesting that 287(g) agreements are associated with a decline in homeownership among u.s.-born citizens and immigrant populations because of the presence of pre-trends. thus, the results for u.s.-born citizens are invalidated by the leads-and-lags model. beyond the direct effects of 287(g) agreements, several exposure and behavioral factors significantly influence homeownership probabilities. more extended residence in the united states is positively associated with homeownership, suggesting that time allows immigrants to accumulate financial resources and establish stability in housing markets. english proficiency also plays a critical role, with individuals who speak english well being more likely to own a home, likely due to improved access to financial services and better employment opportunities. demographic and economic factors further reinforce these trends. age and household income are positively related to homeownership across all groups, reflecting the life-cycle accumulation of assets and the financial capacity needed for home purchases. the number of children is also associated with a greater likelihood of homeownership, potentially due to families seeking stable housing. within the hispanic sample, u.s.-born individuals exhibit higher homeownership rates than foreign-born hispanics, highlighting advantages such as unrestricted access to mortgage markets and greater financial literacy. these findings suggest that while 287(g) agreements negatively impact homeownership, broader structural factors continue to shape housing outcomes, with financial stability and assimilation playing key roles in mitigating policy effects. model 2: the effect of secure communities on homeownership in the second model, we examine the rollout of the secure communities program in the united states, and, as noted earlier, we take advantage of the variation in the timing of adoption among different counties to examine, using the same did model, whether secure communities have affected homeownership in adopting jurisdictions. the logistic regression model examines the impact of the secure communities program on homeownership. as previously noted, this program was implemented gradually across different counties, allowing for a difference-indifferences approach to identify its effects. table 4 presents estimates using two different control groups: residents of states without county-level 287(g) agreements (columns 1-3) and residents of states without both e-verify and county-level 287(g) agreements (columns 4-6). financial services review, 34(1) 122 table 4. the effect of secure communities on home ownership using only residents of states without county level 287(g) agreements (columns 1, 2, and 3) as controls and residents in states without e-verify and county level 287(g) agreements as controls (column 4, 5 and 6) hisp hwcc1 u.s.-born hispe hwcc1 e u.s.-borne (1) (2) (3) (4) (5) (6) post sc county -.028 .051 .042 .321*** .305** .061 (0.058) (0.085) (0.029) (0.085) (0.015) (0.048) years in the united states .017*** (0.002) .057*** (0.003) .015*** (0.002) .058*** (0.003) poor english .294*** (0.053) .149*** (0.057) .03*** (0.002) .156*** (0.061) good english .591*** (0.056) .438*** (0.062) .581*** (0.061) .426*** (0.068) very good english .356*** .356*** .358*** .369*** (0.06) (0.076) (0.065) (0.082) only english .516*** 0.1 .532*** .172 (0.078) (0.122) (0.086) (0.131) age .05*** .04*** .071*** .049*** .041*** .07*** (0.001) (0.003) (0.0006) (0.002) (0.003) (0.0006) number of children .27*** .144*** .42*** .271*** .153*** .424*** (0.016) (0.027) (0.011) (0.018) (0.028) (0.012) bachelor’s degree or higher .104*** (0.033) .081*** (0.029) .097*** (0.036) .062*** (0.032) log of income .987*** .583*** 1.039*** .993*** .582*** 1.045*** (0.027) (0.038) (0.016) (0.03) (0.041) (0.018) u.s.-born .329*** (0.069) .283*** (0.075) obs. 118,336 27,895 2,617,485 102,970 24,086 2,322,113 note: data are from the 2005-2013 american community survey (acs). logistic regressions apply individual weights provided by the acs and include year-fixed effects. robust standard errors are clustered at the puma level. the outcomes variable is a dummy variable indicating homeownership. logistic regressions control for state-level 287(g) agreements and e-verify (columns 1-3) and state-level 287(g) agreements (columns 4-6). *, **, *** denote significance levels at the 10, 5, and 1 percent levels, respectively. table 5. marginal effects of secure communities program on home ownership group dy/dx std. err. z p-value 95% confidence interval hispanic e .0621 .0164 3.80 < .001 [0.0301, 0.0942] hwcc1 e .0579 .0226 2.56 .010 [0.0136, 0.1022] note. dy/dx represents the marginal effects after the implementation of the secure communities program. standard errors are based on the delta method. p-values are reported to three decimal places, with < .001 indicating high statistical significance. confidence intervals are reported at the 95% level. results in table 4 indicate moderate but significant positive effects of secure communities on homeownership among hispanics (column 4) and hwcc1 (column 5), and the rest of the coefficients are not significant at conventional levels. table 5 presents the marginal effects of the significant coefficients. more specifically, the probability of homeownership for hispanics increased by 6.2 percentage points. the probability of homeownership for hwcc1 increased by 5.8 percentage points in jurisdictions with sc after the implementation compared to those in the control group; however, as illustrated in figure antonoudi et al. 123 3, which shows the pre-trend analysis, the parallel pre-trends assumption does not hold for any of the two groups, thus invalidating the results for these groups. robustness checks to ensure the validity and reliability of our main findings, we conducted a series of robustness checks using alternative specifications, broader control groups, and expanded definitions for immigrant populations. the results from the robustness checks reinforce the primary conclusions drawn from our analysis. first, we tested the sensitivity of our findings to alternative control groups. since did results depend on the control group, we provide additional analysis using a larger pool in the control group to examine the effect of 287(g) agreements and the sc program. the results for 287(g) agreements are presented in table 6. while they contain the same outcomes and independent variables as those in table 2, the control group in the first three columns includes all u.s. residents. in addition, the specification in column (4) focuses only on u.s.-born hispanics; the specification in column (5) focuses on hwcc1 living in the united states for more than 10 years, and the last specification (column 6) focuses only on white u.s.-born citizens. these results, displayed in tables 6 and 7 for the 287(g) program and tables 8 and 9 for the secure communities program, remain consistent with the main estimates. for instance, the marginal effects presented in table 7 indicate significant negative impacts of 287(g) agreements on homeownership probabilities. more precisely, the implementation of 287(g) agreements leads to a decline in the probability of homeownership of approximately 3.4 percentage points for hispanics (p<0.001), 6.9 percentage points for hwcc1 (p<0.001), and 9.1 percentage points for u.s.-born citizens (p<0.001). these marginal effects are very close in magnitude to our primary results, providing further evidence that 287(g) policies adversely affect homeownership, especially for hwcc1 and u.s.-born residents in policy-affected areas. financial services review, 34(1) 122 table 6. the effect of 287(g) agreements on home ownership using all residents of the united states as controls hisp hwcc1 u.s.-born hispus hwcc1 10yrs u.s.-born-w (1) (2) (3) (4) (5) (6) post 287g county -.174*** -.361*** -.529*** -.169*** -.370*** -.552*** (0.063) (0.081) (0.05) (0.058) (0.082) (0.053) years in the united states .029*** (0.001) .051*** (0.001) poor english .227*** (0.003) .114*** (0.033) .056*** (0.037) good english .524*** (0.039) .357*** (0.04) .228*** (0.042) very good english .378*** (0.041) .288*** (0.044) .294*** (0.044) only english .221*** .07 .189*** (0.053) (0.062) (0.066) age .054*** .043*** .012∗∗∗ .063*** .056*** .073*** (0.0007) (0.001) (0.00002) (0.0009) (0.001) (0.0004) number of children .213*** .151*** .059∗∗∗ .237*** .057*** .44*** (0.009) (0.013) (0.0005) (0.013) (0.016) (0.008) bachelor’s degree or higher .222*** (0.018) .072∗∗∗ (0.0005) .252*** (0.021) .097*** (0.017) log of income .919*** .606*** .167∗∗∗ 1.019*** .679*** .952*** (0.016) (0.024) (0.0003) (0.017) (0.025) (0.009) u.s.-born .674*** (0.032) obs. 516,994 140,590 5,123,700 237,950 98,411 4,308,027 note: data are from the 2005-2013 american community survey (acs). logistic regressions apply individual weights provided by the acs and include year-fixed effects. robust standard errors are clustered at the puma level. the outcomes variable is a dummy variable indicating homeownership. logistic regressions control for sc, state-level 287(g) agreements, and e-verify. *, **, *** denote significance levels at the 10, 5, and 1 percent levels, respectively. table 7. marginal effects of 287(g) agreements on home ownership using all residents of the united states as controls group dy/dx std. err. z p-value 95% confidence interval hispanic -.0342 .0124 -2.75 < .001 [-0.0585, -0.0098] hwcc1 -.0688 .0153 -4.49 < .001 [-0.0989, -0.0388] u.s.-born -.0911 .0085 -10.67 < .001 [-0.1078, -0.0744] hispanic e -.0327 .0111 -2.94 .003 [-0.0546, -0.0109] hwcc1 e -.0807 .0177 -4.56 < .001 [-0.1153, -0.0460] u.s.-born e -.0857 .0086 -9.95 < .001 [-0.1026, -0.0688] note. dy/dx represents the marginal effects after the implementation of 287(g) agreements. standard errors are based on the delta method. p-values are reported to three decimal places, with < .001 indicating high statistical significance. confidence intervals are reported at the 95% level. as presented in tables 8 and 9, the secure communities program shows significant, though more minor, positive marginal effects on homeownership probability for hispanics (2.8 percentage points, p<0.001) and hwcc1 (3.1 percentage points, p=0.006), consistent antonoudi et al. 123 with our earlier findings. however, interpretation requires caution due to some parallel trend violations highlighted above and in the appendix figures a.1 and a.2. table 8. the effect of sc on home ownership using all residents of the united states as controls hisp hwcc1 u.s.-born hispus hwcc1 10rys u.s.-born-w (1) (2) (3) (4) (5) (6) post sc county .142*** .16*** -.044*** .127*** -.142** -.036* (0.037) (0.058) (0.018) (0.037) (0.06) (0.019) years in the united states .028*** (0.001) .049*** (0.001) poor english .227*** (0.032) .111*** (0.032) .061* (0.035) good english .517*** (0.037) .35*** (0.038) .291*** (0.04) very good english .376*** (0.039) .29*** (0.041) .308*** (0.041) only english .213*** .063 .187*** (0.049) (0.058) (0.062) age .055*** .043*** .071*** .063*** .056*** .073*** (0.0007) (0.001) (0.0003) (0.0008) (0.001) (0.0004) number of children .208*** .141*** .384*** .233*** .053*** .438*** (0.009) (0.012) (0.007) (0.012) (0.015) (0.008) bachelor’s degree or higher .221*** (0.017) .192*** (0.015) .253*** (0.02) .101*** (0.017) log of income .909*** .6*** .987*** 1.007*** .667*** .94*** (0.014) (0.022) (0.008) (0.016) (0.024) (0.009) u.s.-born .672*** (0.03) obs. 585,549 157,949 5,741,823 271,970 112,722 4,823,266 note: data are from the 2005-2013 american community survey (acs). logistic regressions apply individual weights provided by the acs and include year-fixed effects. robust standard errors are clustered at the puma level. the outcomes variable is a dummy variable indicating homeownership. logistic regressions control county-level 287(g) agreements, state-level 287(g) agreements, and e-verify. *, **, *** denote significance levels at the 10, 5, and 1 percent levels, respectively. financial services review, 34(1) 122 table 9. marginal effects of secure communities on home ownership using all residents of the united states as controls group dy/dx std. err. z p-value 95% confidence interval hispanic .0280 .0073 3.82 < .001 [0.0136, 0.0423] hwcc2 .0305 .0112 2.74 .006 [0.0087, 0.0524] u.s.-born .0076 .0032 2.40 .017 [0.0014, 0.0137] hispanic e .0245 .0072 3.41 .001 [0.0104, 0.0386] hwcc2 e .0308 .0130 2.37 .018 [0.0053, 0.0562] u.s.-born e .0059 .0031 1.94 .052 [0.0001, 0.0120] note. dy/dx represents the marginal effects after the implementation of 287(g) agreements. standard errors are based on the delta method. p-values are reported to three decimal places, with < .001 indicating high statistical significance. confidence intervals are reported at the 95% level. another potential concern is whether the observed effects of immigration enforcement policies are confounded by broader housing market trends, particularly in the aftermath of the 2007–2009 great recession. to account for this, we introduce a bartik-style index constructed at the state level. the results are similar to the main findings.6 finally, as the validity of our difference-in-differences estimations critically depends on the parallel trends assumption, we conducted dynamic analyses (leads-and-lags models) to explore pre-existing trends, as shown in figures 2 and 3 and figures a.1 and a.2 in the appendix for the robustness analysis using the expanded control groups. for the 287(g) agreements, the parallel trends assumption holds well for hispanics and hwcc1, supporting the potential causal interpretation of our main results. however, significant pre-existing trends among u.s.-born citizens suggest that caution is needed when interpreting the magnitude of effects for this group. for the secure communities program, the assumption of parallel trends is not supported, particularly for hispanics and hwcc1, suggesting caution in interpreting these results as causal. these robustness checks, complemented by marginal effects analysis, substantiate our main 6 we follow watson (2013) and yasenov (2019) and include in the did model a macroeconomic control variable for the impact of the great recession, namely the bartik-style measure, which could be a proxy for the labor demand shocks and the trend of the unemployment rate over the great recession period. the bartik-style measure is constructed as follows: bartik= 𝐵𝑎𝑟𝑡𝑖𝑘 = ∑ 𝑠ℎ𝑎𝑟𝑒𝑞𝑟 2000 𝑞 × ∆2000𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑞𝑡 where 𝑠ℎ𝑎𝑟𝑒𝑞𝑟 2000 is the conclusion: immigration enforcement policies, particularly 287(g) agreements, exert significant negative impacts on homeownership probabilities across the various demographic groups examined. these findings persist even when subjected to alternative definitions of immigrant status, expanded control groups, and additional housing market controls, underscoring their broader economic significance. 7 pre-trends the most critical assumption of did is the parallel pre-trend assumption that control and treatment groups should exhibit similar trends before policy implementation. for this reason, we employ a dynamic model (with full leads and lags relative to the pre-adoption year) to examine whether pre-existing differential trends may partly explain significant differences between the treatment and control groups after policy implementation. if that is the case, we cannot attribute our findings to policy implementation since they might be due to pre-existing differences and not immigration laws. the dynamic analysis results are presented in figure 2 and figure 3 for each significant finding concerning the specifications in table 2 industry share in state r in the year 2000. ∆2000𝑒𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡𝑞𝑡is the nationwide growth of industry q between 2000 and year 𝑡. industry q includes construction, agriculture, finance, government, information, manufacturing, professional, retail trade, wholesale trade, and transportation. results are very similar to the main findings and are available upon request. 7 all these results are available upon request. antonoudi et al. 123 for 287(g) agreements and specification in table 4 for the sc program, and the reference time period is the year of adoption, which is set to zero across all specifications. each panel in the figure plots the estimates of the dynamic did model and the 95 percent confidence interval. the results in figure 2, which assess the pre-trends for 287(g) agreements, suggest that the parallel trends assumption holds for hispanics and hwcc1. however, there is some evidence of a significant downward trend in homeownership among u.s.-born citizens before the implementation of 287(g) that invalidates the post-treatment results. similarly figure 3, which presents pre-trends for secure communities, shows apparent pre-existing differences across groups that do not support the positive significant findings for hispanics (column 4 of table 4) and hwcc1 (column 5 of table 4). figure 2. pre-trends on the effects of 287(g) agreements among the different groups -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 financial services review, 34(1) 122 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post u.s.-born -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics e -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 e antonoudi et al. 123 figure 3. pre-trends on the effects of secure communities among the different groups with statistically significant results across all robustness checks, our main conclusions remain unchanged. the pre-trends for the additional models used in robustness checks are provided in figures a.1 and a.2 in -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post u.s._born e -1 -0.5 0 0.5 1 1.5 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 financial services review, 34(1) 122 the appendix. the adverse effects of 287(g) agreements on homeownership are pronounced among hispanics and hwcc1. secure communities programs do not appear to affect homeownership. the robustness of these findings across alternative control groups, broader immigrant definitions, and housing market controls strengthens the credibility of our results. it suggests that strong immigration enforcement policies like 287(g) agreements have significant lasting effects on homeownership decisions. discussion the findings of this study indicate that implementing 287(g) agreements is associated with statistically significant changes in homeownership rates across multiple demographic groups. while hwcc1 experience the most pronounced declines in homeownership hispanic subpopulations also exhibit significant changes. these results suggest that immigration enforcement policies have broader housing market implications beyond the intended policy targets, affecting both immigrant and non-immigrant populations. the impact of 287(g) agreements is particularly pronounced among hwcc1 with marginal effects indicating a reduction of approximately 7.5 percentage points. the decline was also prominent for hispanics, where the probability of homeownership decreased by a magnitude of 3.6 to 5.6 percentage points. this effect was consistent across different model specifications and different alternative definitions of hwcc. these findings align with prior research indicating that immigration enforcement policies can disrupt economic stability and reduce long-term investments such as homeownership (east et al., 2022; rugh & hall, 2016). the significant negative effects for hispanics further suggest that immigration enforcement policies create broader economic spillovers, possibly by reducing local economic activity, discouraging home purchases, or increasing housing market uncertainty. the economic effects of secure communities, on the other hand, are different compared to 287(g) agreements and they do not indicate smaller changes in homeownership across all groups, but the pre-trends analysis does not support these findings. this is not surprising as 287(g) agreements were a lot more aggressive compared to secure communities. the study also reinforces the role of exposure and behavioral factors in shaping homeownership outcomes. consistent with prior literature, more extended residence in the united states is positively associated with homeownership, highlighting the importance of financial accumulation and market integration over time (chatterjee & zahirovic-herbert, 2014; kim et al., 2012; mundra & uwaifo oyelere, 2018). english proficiency remains a key determinant of homeownership, as those who speak english well are more likely to secure mortgage loans and navigate real estate transactions. the relationship between age, income, and homeownership is consistent with existing research, as financial stability and family size influence the decision to invest in long-term housing (goodman & mayer, 2018). the differential effects observed across groups suggest that while economic stability plays a crucial role in homeownership, immigration enforcement policies introduce additional barriers that disproportionately impact specific populations. overall, the results suggest that 287(g) agreements have considerable negative effects on homeownership with the most significant declines observed for hwcc1 and hispanics, particularly in states without e-verify. these findings highlight how immigration enforcement interacts with economic and demographic factors to shape homeownership trends. implications immigration policy has been associated with notable declines in homeownership among hwcc1 and hispanics. this trend suggests immigration enforcement policies may have broader economic implications beyond their intended targets. homeownership serves as a fundamental component of economic stability and wealth accumulation. therefore, reductions in homeownership rates can lead to decreased community investment and hinder economic growth. these findings align with research indicating that intensified immigration enforcement can reduce economic activity and consumer spending, adversely affecting local economies. antonoudi et al. 123 moreover, industries heavily relying on immigrant labor, such as construction and service sectors, may experience workforce shortages due to restrictive immigration enforcement. such shortages can increase labor costs and delay housing projects, exacerbating existing housing shortages and affordability issues. these dynamics underscore the need for balanced immigration policies considering labor market demands and the potential economic consequences of a reduced workforce. beyond economic factors, immigration enforcement policies also have social implications. the fear and uncertainty generated by these policies can lead to decreased civic participation and trust in public institutions among immigrant and nonimmigrant populations. the observed declines in homeownership among hwcc1 and hispanics suggest that these policies may create financial instability for a broader segment of the population than initially intended. this erosion of financial security could have long-term consequences, reinforcing disparities in wealth accumulation and economic mobility. policymakers should consider comprehensive immigration reform that provides transparent and fair pathways to legal status for immigrants lacking legal immigration status, thereby reducing the negative impacts of enforcementfocused approaches on housing markets and local economies. implementing initiatives that facilitate the integration of immigrants into the labor force, particularly in sectors experiencing labor shortages, can support economic growth and stability. additionally, investing in programs that promote financial literacy and homeownership support for immigrant and minority populations may help mitigate some of the economic disruptions caused by immigration enforcement policies. by adopting a holistic approach that balances enforcement with integration and support, policymakers can mitigate the adverse effects of immigration policies on homeownership and broader economic indicators, fostering more resilient communities. limitations while this study provides valuable insights into the impact of immigration enforcement policies on homeownership rates, several limitations should be acknowledged. first, the reliance on available data sources may not fully capture the complexities of individual legal statuses, as such information is often underreported or misclassified. second, the study focuses on 287(g) agreements and secure communities. however, other immigration enforcement measures at the federal, state, and local levels may also contribute to changes in homeownership rates, which are not fully captured in this analysis. variations in local enforcement intensity and community cooperation with federal authorities could result in heterogeneous impacts that our analysis might not fully address. third, while our difference-in-differences approach attempts to control for unobserved confounders, there remains the possibility of omitted variable bias. factors such as local economic conditions, housing market dynamics, and social networks could also influence homeownership decisions but are challenging to measure comprehensively. fourth, the cross-sectional nature of the data limits our ability to establish causal relationships definitively. lastly, the generalizability of our findings may be constrained by regional differences in policy implementation and demographic compositions. future research should consider exploring these variations to enhance the external validity of the results.addressing some of these limitations in subsequent studies would contribute to a more nuanced understanding of how immigration enforcement policies affect housing outcomes among diverse populations. conclusion the findings of this study provide important insights into how immigration enforcement policies impact homeownership among different demographic groups. the results indicate that 287(g) agreements strongly negatively affect homeownership, particularly for hwcc1 and hispanics. these findings underscore the broader economic and social consequences of immigration enforcement, which extend beyond the intended targets of these policies. the results suggest that such policies contribute to financial instability, discourage long-term investments, and reshape local housing markets. as homeownership remains a key pathway to wealth accumulation and economic security, the observed effects raise concerns about widening disparities in housing access. financial services review, 34(1) 122 references american immigration council. (2021, july). the 287(g) program: an overview. retrieved january 12, 2024, from https://www.americanimmigrationcou ncil.org/research/287g-programimmigration allen, r., & ishizawa, h. (2015). state-level political context and immigrant homeownership in the usa. journal of international migration and integration, 16, 1081–1097. https://doi.org/10.1007/s12134-0140393-x amuedo-dorantes, c., & arenas-arroyo, e. (2021). immigration policy and fertility: evidence from undocumented migrants in the u.s. journal of economic behavior & organization, 189, 274–297. https://doi.org/10.1016/j.jebo.2021.06. 027 amuedo-dorantes, c., arenas-arroyo, e., & sevilla, a. (2018). immigration enforcement and economic resources of children with likely unauthorized parents. journal of public economics, 158, 63–78. https://doi.org/10.1016/j.jpubeco.2017 .12.004 amuedo-dorantes, c., & bansak, c. (2012). the labor market impact of mandated employment verification systems. american economic review, 102(3), 543–548. https://doi.org/10.1257/aer.102.3.543 amuedo‐dorantes, c., & bansak, c. (2014). employment verification mandates and the labor market outcomes of likely unauthorized and native workers. contemporary economic policy, 32(3), 671–680. https://doi.org/10.1111/coep.12043 amuedo‐dorantes, c., & mundra, k. (2013). immigrant homeownership and immigration status: evidence from spain. review of international economics, 21(2), 204–218. https://doi.org/10.1111/roie.12031 amuedo-dorantes, c., puttitanun, t., & martinez-donate, a. p. (2013). how do tougher immigration measures affect unauthorized immigrants? demography, 50(3), 1067–1091. https://doi.org/10.1007/s13524-0130200-x berry, j. w., phinney, j. s., sam, d. l., & vedder, p. (2006). immigrant youth: acculturation, identity, and adaptation. applied psychology: an international review, 55(3), 303–332. https://doi.org/10.1111/j.14640597.2006.00256.x bohn, s., lofstrom, m., & raphael, s. (2014). did the 2007 legal arizona workers act reduce the state's unauthorized immigrant population? review of economics and statistics, 96(2), 258– 269. https://doi.org/10.1162/rest_a_0042 9 borjas, g. j. (2002). homeownership in the immigrant population. journal of urban economics, 52(3), 448–476. https://doi.org/10.1016/s00941190(02)00529-6 bureau of labor statistics (bls). (2024). foreign-born workers: labor force characteristics—2023. u.s. department of labor. https://www.bls.gov/news.release/pdf/ forbrn.pdf chakrabarty, d., osei, m. j., winters, j. v., & zhao, d. (2019). which immigrant and minority homeownership rates are gaining ground in the us? journal of economics and finance, 43, 273–297. https://doi.org/10.1007/s12197-0189443-0 charlton, d., & kostandini, g. (2021). can technology compensate for a labor shortage? effects of 287(g) immigration policies on the u.s. dairy industry. american journal of agricultural economics, 103(1), 70– 89. https://doi.org/10.1111/ajae.12125 chatterjee, s., & zahirovic-herbert, v. (2014). a road to assimilation: immigrants and financial markets. journal of economics and finance, 38(2), 345– https://www.americanimmigrationcouncil.org/research/287g-program-immigration https://www.americanimmigrationcouncil.org/research/287g-program-immigration https://www.americanimmigrationcouncil.org/research/287g-program-immigration https://doi.org/10.1007/s12134-014-0393-x https://doi.org/10.1007/s12134-014-0393-x https://doi.org/10.1016/j.jebo.2021.06.027 https://doi.org/10.1016/j.jebo.2021.06.027 https://doi.org/10.1016/j.jpubeco.2017.12.004 https://doi.org/10.1016/j.jpubeco.2017.12.004 https://doi.org/10.1257/aer.102.3.543 https://doi.org/10.1111/coep.12043 https://doi.org/10.1111/roie.12031 https://doi.org/10.1007/s13524-013-0200-x https://doi.org/10.1007/s13524-013-0200-x https://doi.org/10.1111/j.1464-0597.2006.00256.x https://doi.org/10.1111/j.1464-0597.2006.00256.x https://doi.org/10.1162/rest_a_00429 https://doi.org/10.1162/rest_a_00429 https://doi.org/10.1016/s0094-1190(02)00529-6 https://doi.org/10.1016/s0094-1190(02)00529-6 https://www.bls.gov/news.release/pdf/forbrn.pdf https://www.bls.gov/news.release/pdf/forbrn.pdf https://doi.org/10.1007/s12197-018-9443-0 https://doi.org/10.1007/s12197-018-9443-0 https://doi.org/10.1111/ajae.12125 antonoudi et al. 123 358. https://doi.org/10.1007/s12197011-9224-5 chodavadia, s. a., kerr, s., kerr, w. r., & maiden, l. j. (2024). immigrant entrepreneurship: new estimates and a research agenda. nber working paper, 32400. https://www.nber.org/papers/w32400 cort, d. a. (2012). spurred to action or retreat? the effects of reception contexts on naturalization decisions in los angeles. international migration review, 46(2), 483–516. https://doi.org/10.1111/j.17477379.2012.00894.x de coulon, a., & wolff, f.-c. (2010). location intentions of immigrants at retirement: stay/return or go 'back and forth'? applied economics, 42(26), 3319– 3333. https://doi.org/10.1080/00036846.201 0.482518 dewind, j., & kasinitz, p. (1997). everything old is new again? processes and theories of immigrant incorporation. international migration review, 31(4), 1096–1111. https://doi.org/10.1177/019791839703 100412 east, c., hines, a. l., luck, p., mansour, h., & velasquez, a. (2022). the labor market effects of immigration enforcement. journal of labor economics, (forthcoming). https://doi.org/10.1086/721152 filindra, a., blanding, d., & garcía coll, c. (2011). the power of context: statelevel policies and politics and the educational performance of the children of immigrants in the united states. harvard educational review, 81(3), 407–438. https://doi.org/10.17763/haer.81.3.n30 6607254h11281 flores morales, j. (2019). financial security and immigrants' legal status: an analysis of net worth in the united states (working paper no. jsit19-04). retirement and disability research center, university of wisconsinmadison. https://rdrc.wisc.edu/publications/wor king-paper/jsit19-04 fu, w. (2017). essays on immigrant women's labor supply, time use and the impact of e-verify policy [doctoral dissertation, university of kansas]. goodman, l. s., & mayer, c. (2018). homeownership and the american dream. journal of economic perspectives, 32(1), 31–58. https://doi.org/10.1257/jep.32.1.31. greenman, e., & xie, y. (2008). is assimilation theory dead? the effect of assimilation on adolescent well-being. social science research, 37(1), 109–137. https://doi.org/10.1016/j.ssresearch.20 07.07.003 institute of migration research. (2019). quick immigration statistics: united states. retrieved december 23, 2019, from https://www.ilctr.org/quick-usimmigration-statistics/ kehiaian, s. e., williams, a. a., & bird, c. l. (2021). financial, demographic, and psychological differences between chapter 13 bankruptcy filers and nonfilers. financial services review, 29(1), 67–84. https://www.proquest.com/scholarlyjournals/financial-demographicpsychologicaldifferences/docview/2637405462/se-2 kim, j., chatterjee, s., & cho, s. h. (2012). asset ownership of new asian immigrants in the united states. journal of family and economic issues, 33(2), 215–226. https://doi.org/10.1007/s10834-0129317-0 kohli, a., markowitz, p. l., & chavez, l. (2011). secure communities by the numbers: an analysis of demographics and due process. chief justice earl warren institute on law and social policy, university of california, berkeley school of law. kolker, a. f. (2021). the 287(g) program: state and local immigration https://doi.org/10.1007/s12197-011-9224-5 https://doi.org/10.1007/s12197-011-9224-5 https://www.nber.org/papers/w32400 https://doi.org/10.1111/j.1747-7379.2012.00894.x https://doi.org/10.1111/j.1747-7379.2012.00894.x https://doi.org/10.1080/00036846.2010.482518 https://doi.org/10.1080/00036846.2010.482518 https://doi.org/10.1177/019791839703100412 https://doi.org/10.1177/019791839703100412 https://doi.org/10.1086/721152 https://doi.org/10.17763/haer.81.3.n306607254h11281 https://doi.org/10.17763/haer.81.3.n306607254h11281 https://rdrc.wisc.edu/publications/working-paper/jsit19-04 https://rdrc.wisc.edu/publications/working-paper/jsit19-04 https://doi.org/10.1257/jep.32.1.31 https://doi.org/10.1016/j.ssresearch.2007.07.003 https://doi.org/10.1016/j.ssresearch.2007.07.003 https://www.ilctr.org/quick-us-immigration-statistics/ https://www.ilctr.org/quick-us-immigration-statistics/ https://www.proquest.com/scholarly-journals/financial-demographic-psychological-differences/docview/2637405462/se-2 https://www.proquest.com/scholarly-journals/financial-demographic-psychological-differences/docview/2637405462/se-2 https://www.proquest.com/scholarly-journals/financial-demographic-psychological-differences/docview/2637405462/se-2 https://www.proquest.com/scholarly-journals/financial-demographic-psychological-differences/docview/2637405462/se-2 https://doi.org/10.1007/s10834-012-9317-0 https://doi.org/10.1007/s10834-012-9317-0 financial services review, 34(1) 122 enforcement (report no. if11898). congressional research service. kostandini, g., mykerezi, e., & escalante, c. (2014). the impact of immigration enforcement on the u.s. farming sector. american journal of agricultural economics, 96(1), 172– 192. https://doi.org/10.1093/ajae/aat081 kuuire, v., arku, g., luginaah, i., abada, t., & buzzelli, m. (2016). impact of remittance behaviour on immigrant homeownership trajectories: an analysis of the longitudinal survey of immigrants in canada from 2001 to 2005. social indicators research, 127, 11351156. https://doi.org/10.1007/s11205015-1011-9. kuzniak, s., & grable, j. e. (2017). does financial risk tolerance change over time? a test of the role macroeconomic, biopsychosocial and environmental, and social support factors play in shaping changes in risk attitudes. financial services review, 26(4), 315–338. https://doi.org/10.61190/fsr.v26i4.337 0 luo, t., kostandini, g., & jordan, j. l. (2018). the impact of lawa on the family labour supply among farm households. european review of agricultural economics, 45(5), 857–878. https://doi.org/10.1093/erae/jby017 miles, t. j., & cox, a. b. (2014). does immigration enforcement reduce crime? evidence from secure communities. the journal of law and economics, 57(4), 937–973. https://doi.org/10.1086/680935 mundra, k., & uwaifo oyelere, r. (2018). determinants of homeownership among immigrants: changes during the great recession and beyond. international migration review, 52(3), 648–694. https://doi.org/10.1111/imre.12311 national immigration forum. (2024). immigrants and housing. https://immigrationforum.org/article/e xplainer-immigrants-and-housing/ orrenius, p. m., & zavodny, m. (2015). the impact of e-verify mandates on labor market outcomes. southern economic journal, 81(4), 947–959. https://doi.org/10.1002/soej.12023 painter, g., & yu, z. (2014). caught in the housing bubble: immigrants' housing outcomes in traditional gateways and newly emerging destinations. urban studies, 51(4), 781–809. https://doi.org/10.1177/004209801349 4425 passel, j. s., & cohn, d. v. (2010). u.s. unauthorized immigration flows are down sharply since mid-decade. pew research center. pham, h. (2018). 287g agreements in the trump era. wash. & lee l. rev., 75, 1253. qing, d., & reiter, m. (2024). racial/ethnic disparities in financial advice seeking: a decomposition analysis. financial services review, 32(4), 27–50. https://doi.org/10.61190/fsr.v32i4.335 5 rodríguez-planas, n. (2018). mortgage finance and culture. journal of regional science, 58(4), 786–821. https://doi.org/10.1111/jors.12385 rugh, jacob s., and matthew hall. 2016. deporting the american dream: immigration enforcement and latino foreclosures. sociological science 3: 1053–76. doi:10.15195/v3.a46. schoenholtz, a. i. (2005). homeownership and the integration of immigrants in the united states. in e. guild & j. van selm (eds.), international migration and security: opportunities and challenges (pp. 191–216). routledge. sharpe, j. (2020). a pathway to homeownership? evidence from the immigration reform and control act of 1986. contemporary economic policy, 38(3), 435-447. https://search.ebscohost.com/login.asp x?direct=true&authtype=ip,shib&db https://doi.org/10.1093/ajae/aat081 https://doi.org/10.1007/s11205-015-1011-9 https://doi.org/10.1007/s11205-015-1011-9 https://doi.org/10.61190/fsr.v26i4.3370 https://doi.org/10.61190/fsr.v26i4.3370 https://doi.org/10.1093/erae/jby017 https://doi.org/10.1086/680935 https://doi.org/10.1111/imre.12311 https://doi.org/10.1002/soej.12023 https://doi.org/10.1177/0042098013494425 https://doi.org/10.1177/0042098013494425 https://doi.org/10.61190/fsr.v32i4.3355 https://doi.org/10.61190/fsr.v32i4.3355 https://doi.org/10.1111/jors.12385 antonoudi et al. 123 =ecn&an=1851246&site=ehostlive&scope=site&custid=uga1 sherman, a., trisi, d., stone, c., gonzales, s., & parrott, s. (2019). immigrants contribute greatly to u.s. economy, despite administration's zpublic chargey rule rationale. jstor. https://www.cbpp.org/sites/default/file s/atoms/files/8-15-19pov.pdf sinning, m. (2010). homeownership and economic performance of immigrants in germany. urban studies, 47(2), 87– 409. https://doi.org/10.1177/004209800934 9021 u.s. census bureau. (2021). current population survey (https://www.census.gov/programssurveys/cps.html watson, t. (2013). enforcement and immigrant location choice (no. w19626). national bureau of economic research. xie, y., & greenman, e. (2011). the social context of assimilation: testing implications of segmented assimilation theory. social science research, 40(3), 965–984. https://doi.org/10.1016/j.ssresearch.20 11.01.004 yasenov, v. i. (2019). immigrants and the u.s. wage distribution. upjohn institute working paper 20-320. kalamazoo, mi: we upjohn institute for employment research. https://doi. org/10.17848/wp20-320. yu, z., & myers, d. (2010). misleading comparisons of homeownership rates when the variable effect of household formation is ignored: explaining rising homeownership and the homeownership gap between blacks and asians in the u.s. urban studies, 47(12), 2615-2640. https://doi.org/https://doi.org/10.1177/ 0042098009359956 https://doi.org/10.1177/0042098009349021 https://doi.org/10.1177/0042098009349021 https://doi.org/10.1016/j.ssresearch.2011.01.004 https://doi.org/10.1016/j.ssresearch.2011.01.004 financial services review, 34(1) 122 appendix figure a.1. pre-trends on the effects of the 287(g) agreements among the different groups with statistically significant results for the model using all residents of the united states as controls presented under tables 6 and 7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post u.s.-born antonoudi et al. 123 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics born in the u.s. -0.6 -0.4 -0.2 0 0.2 0.4 0.6 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 10 or less years in the u.s. -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post u.s.-born, white financial services review, 34(1) 122 figure a.2. pre-trends on the effects of the secure communities program among the different groups with statistically significant results for the model using all residents of the united states as controls presented under tables 8 and 9 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post u.s.-born antonoudi et al. 123 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hispanics born in the u.s. -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post hwcc1 10 or less years in the u.s. -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 4 years prior 3 years prior 2 years prior 1 year prior adoption 1 year post 2 years post 3 years post 4 years post u.s.-born white financial services review, 33(4) iv editorial teaching tomorrow’s financial planners: insights from research into online marketplaces and fraud; behavior of young adults in a post-communist era; and the value of financial advice. elisabeth sinnewe queensland university of technology, australia contributed articles beyond the special issue theme i would like to start by thanking inga timmerman for spearheading the special issue on financial planning pedagogy, as well as john grable and barry mulholland for laying the foundations for this excellent special issue. in addition to the papers featured in the special issue on financial planning pedagogy, this volume includes three original research articles that offer timely insights into the changing environment, in which financial planners operate. although distinct in focus, each study highlights psychological, technological, or sociobehavioral forces that increasingly shape financial decision-making. collectively, they signal new directions for both professional practice and the future design of financial planning education. in the first article, antonoudi et al. (2025) examine the emergence of online marketplaces and their influence on fraud using data from craigslist’s rollout between 1995 and 2006. their difference-in-difference analysis suggests that the transition from manual bulletin boards to digital platforms initially enhanced transparency and reduced opportunities for fraud. while these findings demonstrate the potential of technology to generate digital trails enabling greater transparency, the authors caution that as digital platforms mature, new forms of technology-enabled fraud inevitably emerge. for instance, according to the us financial trade commission, consumers reported losing usd 12.5 billion to fraud in 2024, with the highest overall losses reported from social media scams. these findings also highlight emerging opportunities for practitioners to engage with clients on issues that extend beyond traditional financial planning. as digital fraud becomes more prevalent, discussions around online risk, information security, and consumer protection are likely to become part of the broader client–adviser relationship, reinforcing the educational and trust-building dimensions of financial planning. from an educational standpoint, the study’s implications resonate with emerging research that frames digital and financial literacy as mutually reinforcing competencies. effective financial decision-making increasingly depends not only on traditional financial skills but also on the capacity to navigate digital financial services and evaluate technology-related risks. koskelainen et al. (2023) advance this perspective by introducing digital financial literacy as an integrated skill set that can be deliberately developed through targeted pedagogical design, offering a clear pathway for enriching financial planning curricula. in the second manuscript, macdonald et al. (2025), provide a narrative review of the value of financial advice and develop a conceptual framework synthesizing the insights gained from the review for future research. their review confirms that while quantitative outcomes, such as portfolio performance, remain relevant, many studies attribute the most positive financial services review, 33(4) v outcomes to broader wellbeing aspects, such as mental wellbeing in the form of reduced financial anxiety (see e.g., archuleta et al., 2020). overall, they conclude that financial advice value is a complex and multidimensional construct, which aligns with the view that financial advice is a credence good: its value cannot be fully verified even after service delivery due to inherent information asymmetries (dulleck & kerschbamer, 2006). these insights suggest productive avenues for future practice-relevant research in developing more nuanced, evidence-informed approaches to articulating and assessing the multidimensional value of financial advice. such work could support clearer communication of adviser value and strengthen alignment professional practice with client wellbeing outcomes. for financial planning education, these findings echo longstanding calls within the financial therapy and financial planning literature for stronger development of relational and interpersonal competencies, including communication, empathy, and behavioral coaching (e.g., asebedo, 2019). in the final paper, ahamed et al. (2025) investigate financial management behavior among young adults in poland, where intergenerational financial knowledge transfer remains limited due to the country’s post-communist economic history. in this quasi-laboratory context, parental financial socialization is comparatively weak, allowing the authors to isolate the role of internal psychological factors that are typically difficult to disentangle from intergenerational influences in other settings. using structural equation modelling and fuzzyset qualitative comparative analysis, the authors show that a strong financial attitude, combined with high subjective financial knowledge, can drive desirable behavior despite low socialization and financial self-efficacy. for practice, the study reinforces that effective client engagement requires tools and techniques capable of identifying clients’ psychological drivers, including financial attitudes, knowledge, and self-efficacy, which are often overlooked in conventional client discovery practices. for educators, the study supports the value of teaching methods that deepen students’ understanding of financial decision-making from the client’s perspective. incorporating structured client-discovery exercises, reflective interviewing techniques, or analyses of client stories can help emerging advisers appreciate behavioral drivers of financial behaviour. some of these approaches have been illustrated by several of the pedagogical contributions to this special issue. taken together, these three studies illuminate critical trends that extend beyond their immediate empirical settings. they point to a profession increasingly characterized by digital risk, behavioral complexity, and growing client diversity, suggesting that tomorrow’s financial planners will require broader behavioral and digital competencies. these themes align closely with the pedagogical innovations highlighted in the special issue and offer productive avenues for future scholarship on the evolving foundations of financial planning education. as inga noted, the continued vibrancy of fsr rests on the commitment of our associate editors, editorial board members, and reviewers, and i am deeply grateful for their contributions and for the collaborative spirit that drives the fsr community forward. financial services review, 33(4) vi references ahamed, a. f. m. j., jakubowska, d., pacholek, b., & dziewanowska, k. (2025). exploring factors affecting young adults' financial management behavior: a hybrid pls-sem and fsqca approach. financial services review, 33(5), 164-190. antonoudi, e., seay, m., lim, h., & kiss, e. (2025). the impact of the online marketplace on fraud: evidence from craigslist from its early adoption in 1995 to its wider expansion in 2006, financial services review, 33(4), 121-133. archuleta, k.l., mielitz, k.s., jayne, d. & le, v. (2020). financial goal setting, financial anxiety, and solution-focused financial therapy (sfft): a quasi-experimental outcome study. contemporary family therapy, 42(1), 68-76. asebedo, s.d. (2019). financial planning client interaction theory (fpcit). journal of personal finance, 18(1), 9-23. dulleck, u. & kerschbamer, r. (2006). on doctors, mechanics, and computer specialists: the economics of credence goods. journal of economic literature, 44(1), 5-42. koskelainen, t., kalmi, p., scornavacca, e. & vartiainen, t. (2023). financial literacy in the digital age a research agenda. journal of consumer affairs, 57(1), 507-528. macdonald, k. l., wildman, k. l., loy, e., & brimble, m. (2025). the value of financial advice: a narrative review and conceptual frameworks. financial services review, 33(4), 134-163. financial services review, 33(2) 1 psychophysiological finance and intelligent wellness: a new financial planning practice model robert hanlon,1 paul leher,2 alexander cohen,3 eric miller,4 monte hancock,5 and robert mitchell6 abstract the certified financial planner board of standards, inc. requires cfp® professionals to identify and respond to a client's attitudes, behaviors, and situations that impact decision-making, the client-planner relationship, and a client’s financial well-being. this practice requirement acknowledges the importance of identifying and analyzing psychological reactions, physiological responses, and financial triggers, which interact to influence client intentions, actions, and outcomes. the paper provides an overview of the way financial stressors and acute and chronic stress can impact the well-being of clients. building on this background, the paper describes a vision for a new advice-delivery model based on the emerging fields of psychophysiological finance and intelligent wellness. the practice model described in this paper shows how advances in mobile health, psychophysiology, and psychology can be blended with a traditional financial planning practice approach to provide clients with comprehensive advice and guidance, attempting to reduce the effects of stress and improve client well-being. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation hanlon, r., leher, p., cohen, a., miller, e., hancock, m., & mitchell, r. (2025). psychophysiological finance and intelligent wellness: a new financial planning practice model. financial services review, 33(2), 1-14. introduction the notion that a client's psychological perspective is associated with the way financial planning recommendations should be made and implemented was recently codified into financial advisory practice standards. in 2021, the certified financial planner boards of standards, inc. (cfp board) introduced six learning objectives related 1 corresponding author (bob.hanlon@livingcenterline.com). living centerline, llc. philadelphia, pa, usa. 2 rutgers university, new brunswick, nj, usa. 3 living centerline, llc. philadelphia, pa, usa. 4 living centerline, llc. philadelphia, pa, usa. 5 living centerline, llc. philadelphia, pa, usa. 6 living centerline, llc. philadelphia, pa, usa. to client psychology for inclusion in program training curricula. in the context of these objectives, the psychology of financial planning involves recognizing and addressing attitudes, behaviors, and situations that influence decisionmaking, the client-planner relationship, and a client's financial well-being. although not specifically mentioned in the learning objectives, https://creativecommons.org/licenses/by-nc/4.0/ mailto:bob.hanlon@livingcenterline.com https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(2) 2 there is a growing realization that client behavior can be explained by identifying and analyzing psychological reactions, physiological responses, and financial triggers that interact to influence intentions, actions, and outcomes (grable et al., 2015; njegovanovic, 2018). the study of these interrelationships is referred to in this paper as psychophysiological finance (psyfi).7 psyfi differs from behavioral economics and behavioral finance, which focus on the cognitive reasoning behind economic decisions. psyfi is based on the hypothesis that physiological activities and behaviors generated by the peripheral nervous system—specifically, the autonomic nervous system that regulates bodily processes like heart rate, respiration, digestion, and skin conductance—interact with an individual's psychological outlook and objective measures of financial wellness. this interaction shapes financial intentions and behaviors. this paper provides an overview of the various ways physiological status—particularly stress reactions—can influence a client's psychological, physical, and financial well-being and presents the conceptual foundation for a new approach to providing financial planning advice that aligns with the notion of transitioning the role of a financial planner to what klontz and associates (2016) called a financial health physician. literature review to understand what the term psyfi means, it is important to have a basic understanding of a factor that links psychology, physiology, and financeology (i.e., the analysis of a household’s financial situation) together in an interrelated way: stress. stress is a condition of life (lehrer et al., 2024). while nearly all people intuitively know this, only some understand what stress is or what reflexes control it. stress can be thought of as the process by which people adapt to the environment around them. individuals encounter stressors on a daily basis. stressors can lead to acute stress (i.e., a shortterm stress response brought on by an immediate 7 the term psychophysiological finance refines a psychophysiological economics hypothesis that describes financial well-being as related to threat) and chronic stress (i.e., a long-term form of stress brought on by prolonged exposure to stressors), which occurs when health, functioning, or one’s sense of well-being are impaired. two neuropsychologists, bruce mcewen and john wingfield, introduced the concept of ‘allostasis’ to describe how the body adapts to stress. they defined allostasis as stability through variability (mcewen & wingfield, 2010). allostasis involves maintaining stability through oscillatory variability around a healthy baseline. oscillation is a characteristic of a wellfunctioning physiological or psychological system (lehrer & eddie, 2013). when one’s heart rate increases, it subsequently decreases. similarly, fluctuations occur in blood pressure, mood, body temperature, and all other physical and psychological functions. engineers describe this phenomenon as a control system governed through negative feedback loops (chirumalla, 2017). these control systems are self-regulating, limiting how far a function can deviate. external demands compel systems to adjust in order to adapt, while negative feedback loops prevent mental or physiological extremes, such as manic states or dangerously high blood pressure. the discussion thus far seems removed from the day-to-day practice of financial planning. however, this is not the case. financial planners must manage their own stress response and anticipate and react to their clients' stress reactions. consider a client who has experienced significant stress. maybe they lost their job or a loved one died. the client’s reflexes may become insufficient or fatigued, leading to seemingly unexplainable decisions and awkward clientfinancial planner interactions. this often occurs when demands are either too severe or too prolonged, which overwhelms a person’s ability to effectively deal with stress. chronic stressors, such as caregiving for a disabled family member or prolonged financial strain, can easily overwhelm a client’s emotional and physiological system. mcewen and behavioral, cognitive, and physiological mechanisms (grable 2013). hanlon et al. 3 wingfield (2010) referred to this state as ‘allostatic overload.’ when allostatic overload occurs, the nervous system no longer functions properly. focus, energy, coordination, and judgment may decline, and the ability to return to a normal state diminishes. anxiety, depression, sleep disturbances, irritability, fatigue, or fear may arise (jovanovic & norrholm, 2016), all of which can negatively impact the client-financial planner relationship. additionally, cognitive functions, including decision-making and performance on demanding tasks, may deteriorate (ramakers et al., 2023). physiological consequences include elevated blood pressure and glucose levels, which can increase susceptibility to infections (jiménez et al., 2021). if this state persists, the individual may become vulnerable to more serious chronic diseases (bellingrath & kudielka, 2017; lovallo, 2015; mcgrady & moss, 2013), including heart disease (cundiff & smith, 2017), asthma (lehrer & moritz, 2022; lehrer et al., 2023), gastrointestinal disorders (jepson, 2008; overmier & murison, 2013), diabetes, and hypertension (brügge, 2001). this insight helps to explain why some clients when faced with outwardly simple financial planning decisions, get bogged down, leading to postponed plan implementation and a tendency to make lessthan-optimal risk-taking decisions (porcelli & delgado, 2009). of particular importance to financial planners is the role of financial stressors. in their periodic survey of stress in america, the american psychological association (2022) regularly identifies financial stressors as major sources of stress symptoms. in 2022, 83% of americans identified inflation as a stressor. approximately 57% said that not having enough money was their primary source of stress, while 43% said that saving enough money for future needs was their main source of stress. in a culture where material wealth translates to social status and respect, financial problems can add to perceptions of relative deprivation compared with neighbors, friends, and other reference groups. further, lower perceived social status can lead to stress responses even when actual income or wealth is above average (beshai et al., 2017). negative perceptions of one’s economic condition can increase susceptibility to other social stressors, such as conflict in love or coworker relationships (lucas et al., 2021; wheeler et al., 2019) and destructive personality patterns involving suspiciousness, anger, social withdrawal, or generally awkward or annoying interpersonal patterns that are occasionally reported by financial planners. in short, experiencing severe financial stressors is related to allostatic overload (french, 2023; patel, 2019). a new practice model: psychophysiological finance and intelligent wellness as the preceding discussion highlights, stressors are prevalent in society, with many financial planning clients exhibiting some degree of stress that is either related to money worries or other life factors (american psychological association, 2022). unfortunately, the type of work a financial planner engages in with a client is typically conducted in a siloed manner. this is true of other advice-giving professionals as well. a financial planner rarely assesses a client's health history, stress reactions, or degree of general wellness. this is because medical, psychological, and financial knowledge is generally kept within one’s professional field, made inaccessible, and rarely shared across fields of practice. scope of practice concerns also lead some financial planners to believe that venturing beyond the core aspects of financial advice could lead to liability exposure (chene et al., 2010). still others are unsure that their training and expertise are sufficient to help clients deal with emotional issues and life stressors (grable et al., 2015; gray 2023). lack of information sharing, in particular, creates barriers to deciding upon and committing to long-term health and wellness goals. commonly, this means insubstantial physical, mental, and financial health complications are left untreated until they have worsened or become observable. in this paper, we propose a model of psyfi and intelligent wellness (the model) that aims to help financial planners increase a client’s well-being through the assessment of the three primary wellness domains: psychological, physiological, and financial (see figure 1). the model is based on decades of practice in the mental health and medical fields and financial planning research. financial services review, 33(2) 4 until recently, adoption of the model was impractical due to technological limitations. as will be discussed below, with the advent of smart technologies and artificial intelligence (ai), it is now possible for a financial planner to use realtime data obtained directly from clients to create integrated and comprehensive strategies that help a client align attitudes, reactions, and behaviors into optimized financial planning behavior. figure 1. the three wellness domains of the psychophysiological finance and intelligent wellness model based on trends in digital health, coupled with information that is emerging from the fields of psychology, psychophysiology, and financeology, we believe the next frontier in financial advice will revolve around what some are calling intelligent wellness (miller et al., 2023). intelligent wellness provides a systematic, evidence-based approach to achieving optimized well-being. intelligent wellness is premised on the practice of systems medicine that examines the composite characteristics of a system, utilizing computational and mathematical tools to analyze the full array of internal and complex component interactions (ahn et al., 2006a; ahn et al., 2006b). rather than specializing in the maintenance of a specific area of health, intelligent wellness seeks to holistically manage a person's entire health ecosystem, which can encompass mental, physical, and financial wellbeing, as each of these components impacts one another. the goal of integrating financial planning with psychology and physiology is to help clients deal with maladaptive patterns of financial behavior. as discussed earlier in the paper, this is important because financial stressors are known to influence a person's physical and mental well-being (grable, 2013; hanlon et al., 2020). similarly, psychological and physiological stress has a negative impact on financial planning outcomes (leher et al., 2024). how might the model be adopted in the financial planning profession? finding an answer to this question was unattainable five years ago. it is possible today, using smart technologies, ai, and a network of connected devices, to collect realtime data from clients. given today's computing power, it is possible to organize data from a variety of sources and sensors to obtain accurate estimates of a client's stress level and reactions. the model does not require a financial planner to be an electrical engineer, a medical professional, a programmer, or a licensed mental health professional. the mechanics that drive the data collection and financial planner feedback are, in many respects, no different than the multitude of technology interfaces that financial planners are already using. the difference is in the valuable information available to a financial planner. a psychological status physiological status financial situation hanlon et al. 5 financial planner does not need to know about a client's ongoing blood pressure, their degree of hydration, or a change in the client's credit score. this information can be gathered unobtrusively, synthesized into actionable information, and delivered directly to the financial planner. with supporting content, the rendered data can be used to drive personalized and actionable recommendations and behavioral interventions that maintain a balance across the three wellness domains (i.e., psychological, physiological, and financial). although all of this may seem futuristic and realistically impossible to do in practice, it is worth remembering that not long ago people expressed skepticism about adopting electric vehicles, blockchain technology, cryptocurrencies, chatbots, ai, wearable devices, and automated teller machines. prior to the covid-19 pandemic, few financial planners would have thought meeting with clients virtually could be effective. today, nearly every financial planning professional, like many physicians, utilizes virtual technologies. rather than being an elusive vision of the future, the ability to use mobile applications as a form of personal financial and stress management is, in fact, something that is happening presently (hanlon et al., 2020). according to the world health organization (n.d.), mhealth and related digital technologies are revolutionizing how people interact with health service providers. consider the adoption of mobile health (mhealth) services. a 2015 pew research center study found that 58% of smartphone users had downloaded a health-related application. the accessibility of mhealth tools has reshaped healthcare delivery, allowing for immediate access to one’s primary care physician and prescriptions regardless of patient location. the model described in this paper is an adaptation of existing mhealth technologies applied to the practice of financial planning. practice example while financial planners can request client data from wearable devices and analyze the data independently, this approach is neither practical nor necessary. instead, technological advancements make it possible for clients to seamlessly share their data through an integrated app, streamlining psychological, physiological, and financial stress measures for use in the financial planning process. this capability is neither novel nor particularly complex to implement. figure 2 illustrates the general data gathering process as conceptualized in the model. figure 2. the psychophysiological finance and intelligent wellness data gathering process psychological inputs from smart device physiological inputs from smart device financial inputs from planner or software integration integrated data app financial services review, 33(2) 6 once data have been gathered, the next step in the model is to use a digital twin. digital twin modeling is most closely aligned with work in product lifecycle management, where real-time comprehensive data allows a model to evolve alongside a product’s various lifestyle stages (grieves, 2022). in contrast to standard simulation technologies, which also utilize digital models to replicate a system’s various processes, digital twin modeling generates an entirely virtual environment through continuously linked data. unlike conventional financial modeling techniques such as monte carlo simulation, the digital twin approach leverages real-time data to identify psychological, physiological, and financial interventions that can be used to enhance a client’s overall well-being. the analysis occurs before client meetings, allowing a financial planner to monitor a client before, during, and after a meeting and proactively adjust recommendations. table 1 shows the type of realtime data and information that can be obtained from a network of connected devices. table 1. the identification and labelling of well-being vectors psychological physiological financial positivity weight emergency fund status engagement blood pressure paying bills on time relationships sugar level spending less than income meaning age sufficient long-term savings accomplishment sleep manageable debt load emotional stability diet credit score optimism heart rate variability appropriate insurance resilience body mass index (bmi) expenditure planning self-esteem vitality a financial planner can then use this information to model outcomes using a digital twin, as shown in figure 3. at its core, the model takes representative data points like those shown in table 1 and analyzes them simultaneously. data can then be manipulated using the digital twin to identify the best mixture of client-centered recommendations that work in combination to reduce stress and increase well-being. by monitoring a client's psychological, physiological, and financial situation in real time, it is possible to identify situations where one wellness domain reaches its allostatic load and becomes incapable of selfregulation. this is important because the current siloed approach to providing advice seldom accounts for stressors outside a financial planner’s practice specialty. a siloed planner may provide useful guidance, but without knowledge of the other wellness domains, the advice may negatively affect how the client deals with other stressors. consider the following example: in this case example, a client is experiencing family stress. while their financial planner senses that something is not “quite right,” the planner concludes that exploring this intuitive insight goes beyond their scope of expertise and practice. instead, the financial planner focuses on the outcomes associated with the client’s stress. in this example, the planner may identify increased spending on the client's part. more specifically, they may question the amount of money being spent and the actual items being purchased. hanlon et al. 7 without further exploratory work, there is no way for the financial planner to know that the client uses shopping to deal with family trauma. the more stress, the more spending on luxury goods. in this case, the financial planner focuses on helping the client work through the spending issue, which appears to be a source of financial stress. working from a siloed perspective, the financial planner will likely make specific budgeting recommendations and other financial observations. doing so, however, is likely to backfire and cause the client even more stress and anxiety (britt-lutter et al., 2019), leading to even more spending. figure 3. the psychophysiological finance and intelligent wellness modelling process figure 4 illustrates the situation. when viewed quantitatively, the client is experiencing moderate to high psychological stress, which stems from the client’s family situation. physiologically, they appear to be coping well. financially, however, the client’s spending behavior is causing financial stress and budgetary issues. this is captured in the first three bars in figure 4. the next three bars show the immediate result after the financial planner makes spending and budgeting recommendations. initially, the client stops spending and adheres to their budget. this immediately improves the financial situation. however, because the financial planner was unaware of the underlying cause of the spending, the recommendations worsen the client's psychological stress. if left unexplored, the client is likely to go back to a reckless spending pattern as a coping mechanism. the corresponding impact of this psychological response to heart rate and blood pressure can then imbalance a range of accompanying physical regulatory responses, leading to reduced health in a variety of functions (lehrer, 2021), thus increasing physiological stress. these outcomes are seen in the bars corresponding to longer-term outcomes. data in app synethsized (figure 1) information sent to financial planner digital twin used to model client outcomes integrated recommendations developed recommendations delivered to client ongoing smart device used to monitor outcomes updates to sent to financial planner adjustments made to plans ongoing monitoring financial services review, 33(2) 8 figure 4. the interrelated nature of psychological, physiological, and financial factors the last three bars in figure 4 represent outcomes associated with using the model. had the financial planner, in this case, been able to document the high level of psychological stress being experienced by the client before making recommendations, it may have been possible to develop interventions—possibly in collaboration with another professional—to help the client deal with the stress arising from the client’s family. this, in turn, would have given the financial planner insight into the change in the client’s spending pattern, which almost certainly would have altered the recommendations presented to the client. ultimately, the client's psychological, physiological, and financial status would have been improved. of course, a key assumption underlying the model’s use is that a financial planner desires to help a client reduce stress and increase wellness rather than focusing entirely on aspects of wealth accumulation and asset protection. using the model, a financial planner is much more likely to identify where stress originates from and then build strategies to help the client across the wellness domains. by quantifying well-being, each wellness domain can be algorithmically analyzed against a target value to improve wellbeing and prevent adverse health spirals. ensuring client action identifying sources of stress, financial or otherwise, is just the first step in helping a client improve their well-being (grier & bryant, 2005). cultivating one's physical, mental, and financial health often requires maintaining complex, strenuous, and initially unpleasant activities such as exercise, reduced resource consumption, and cognitive restructuring (hastings, 2007). while changing behaviors has been shown to increase well-being in the long term, the benefits are indirect and often not experienced until the behavior has been maintained for a long time (rothschild, 1999). providing motivation to clients is one of the model’s most appealing aspects. the motivation literature shows that action occurs through the satisfaction of three inherent psychological needs (mitchell et al., 2022): (a) autonomy—the ability to maintain causal agency in decision-making; (b) competence—the satisfaction of overcoming a challenge; and (c) relatedness—a meaningful interconnection with others. the motivational challenges associated with improving a client’s well-being can be understood by example. consider what happens when someone is encouraged to begin exercising. exercising is often performed due to social current situation after recommendation long-term outcomes model outcome psychological -5 -7 -7 1 physiological 1 1 -4 1 financial -6 -1 -5 1 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 +/ -c en te rli ne hanlon et al. 9 pressure, which means the time delay between starting exercise and improving physical health can seem insurmountable. the result is that wellness behaviors are generally not interesting or fun, at least not at first. this is one reason why mhealth interventions, in particular, have been shown to struggle to support relatedness satisfaction (amagai et al., 2022), as without the in-person component, participants often feel as though their actions are not important to others (mitchell et al., 2022). so, how can a financial planner who wants to adopt the model motivate a client to take action? financial planners need to utilize personal relationships with their clients to reinforce how important psychological, physiological, and financial changes can be. this requires a financial planner to draw on skills, experience, and tools to support deeper client engagement and design more motivating behavioral change programs. one such tool is the use of reward elements such as tokens or achievements that can be used to provide clients with ongoing motivation to improve their financial situation. as rewards are presented, clients will be more likely to visualize otherwise unobservable health and financial milestones, retaining a sense of overcoming a challenge in the face of long-term goals. in the context of the model, the digital twin supports the personalization of gamification, utilizing realtime data to modify the type of recommendations a financial planner will develop and present to the client. takeaways for financial planning professionals adopting the model presented in this paper offers several positive outcomes for financial planners. to begin with, the model helps move financial planning away from a traditional siloed advicegiving approach towards one that integrates financial, psychological, and physiological wellness domains to better understand and service clients. this occurs because gaining an understanding of the role of stress in clients' lives can help in the design and delivery of comprehensive and personalized advice. second, the model integrates with the way many financial planners and their clients use mhealth tools. by embracing the power of digital health technologies, such as telehealth, wearable biosensors, and mhealth apps to track and assess client wellness, financial planners can use realtime data about a client's physical and mental health to tailor financial advice that incorporates ways for a client to better manage their mental, physical, and financial situation. this systematic approach ensures that financial planners can identify and address the multiple ways financial stress contributes to overall health and financial outcomes. third, the model provides a mechanism for building deeper, more meaningful client relationships by encouraging preventative wellness strategies. by integrating bio-behavioral data into their practice, financial planners can begin to deliver actionable recommendations that help clients maintain balance across the three wellness domains (i.e., psychological, physiological, and financial). the dynamic feedback loop built into the model ensures financial planners address financial concerns and contribute to each client's broader sense of wellbeing. rather than waiting for financial stress to manifest into serious health and financial issues, financial planners can use the practice approach described in this paper to proactively guide clients in adopting behaviors that prevent stressrelated diseases and household financial management mistakes. finally, a holistic financial planning approach built upon the model can help the financial planning profession bridge the gap between fragmented mhealth applications, the delivery and implementation of financial planning advice, and ongoing client monitoring. by viewing clients as more than their physical, mental, or financial selves, financial planners can begin to focus on managing wellness and the interconnected components of a client's overall well-being, enhancing long-term health and financial outcomes. in this way, the practice of financial planning can move closer to what klontz et al. (2016) called financial healthcare. it is important to acknowledge that adopting the model significantly expands scope of practice constraints while introducing new ethical considerations. financial planners who adopt the model must recognize the interdisciplinary financial services review, 33(2) 10 nature of integrating financial, psychological, and physiological domains and navigate the complexities this presents. for some financial planners, moving beyond their traditional focus on asset management, retirement planning, and risk management may be beyond their trained scope of practice. when this is the case, it is important to collaborate with mental health and medical professionals to ensure that descriptions of stress and the response to stress are appropriate and valid. working with specialists can also jumpstart a financial planner's continuing education in holistic client care. similarly, using mhealth tools requires a financial planner to be technologically adept. without a process for integrating multiple data streams, including realtime biometric data, a financial planner may feel unprepared to begin working with a client holistically when providing advice on stress management, mental health support, healthy lifestyle changes, financial planning strategies, and overall wellness. a financial planner adopting this model must also ensure they have the necessary training and qualifications to interpret and integrate psychological and physiological data responsibly. there are no specific ethical or procedural guidelines describing how to incorporate intelligent wellness tools into practice, which means financial planners must take steps to ensure that they know when to give advice and when to refer a client to another professional. the medical triage process provides a framework financial planners can use to ensure they stay within the boundaries of their scope of practice. for example, financial planners can provide educational resources about stress management and its impact on financial decisionmaking. sharing general wellness tips, such as the importance of exercise and nutrition in relation to financial planning (e.g., how health impacts work productivity and income), is also permissible. utilizing a triage process allows financial planners to ethically and effectively help clients set financial goals that support holistic mental, physical, and financial wellness. the triage process begins by assessing a client's financial health. the financial planner then identifies the complexity of the issues presented. some issues are routine or preventive, which the planner can easily manage. however, if a client presents a moderately complex problem that requires specialized knowledge, the financial planner might need to consult a specialist or conduct in-depth research. in cases where the issue is critical—such as emotional money behaviors, complex estate questions, or advanced intrafamily tax issues that may trigger psychological or physiological responses beyond the planner's expertise—the financial planner should refer the client to a specialist or collaborate with another mental or physical health professional. an essential part of the triage process is prioritizing client issues based on urgency and importance. high-priority issues include immediate threats like foreclosure, excessive debt, apparent adverse health reactions, and suicidal thoughts. medium-priority issues involve stress related to market corrections or concerns about job loss. lower-priority issues are generally more routine. by addressing the most urgent problems first, a financial planner can identify when a referral is necessary and when it is appropriate to work with limited collaboration in helping clients manage their psychological, physiological, and financial well-being. the final two stages of the triage process involve coordinating care and monitoring client outcomes. financial planners are already wellequipped to act as a client's "financial quarterback," coordinating interactions with accountants and attorneys. including professionals like financial therapists, physicians, or psychologists in one's referral network is a natural extension of the services most financial planners provide. another issue of importance for those thinking about the model is the notion of privacy and confidentiality. the collection and sharing of sensitive health data introduces privacy concerns. anyone accessing a client's health data must adhere to stringent privacy laws, including the health insurance portability and accountability act (hippa). the act mandates that client health data be securely handled and shared only with a client's unambiguous consent (moore et al., 2007). financial planners must also be fluent in the language of informed consent, which is the process of obtaining a client's permission to hanlon et al. 11 collect and use psychological and physiological data in a financial planning context. financial planners should clearly explain how data will be used, what technologies will be employed, and the potential risks involved in integrating their health information with financial advice. lastly, the model does introduce potential conflicts of interest. integrating psychological and physiological measures into the financial planning process can be a concern when wellness products, services, or technologies are recommended without fully disclosing real or perceived conflicts of interest. for instance, promoting a particular digital health tool in exchange for compensation could compromise a financial planner's objectivity. transparency regarding such relationships and avoiding biased recommendations is essential. in a similar vein, it is important to note that not all clients have access to digital health tools or the financial means to invest in such tools. this relates to the notion of equity, where a financial planner needs to be mindful of the burdens and constraints placed on clients. in summary, adopting the model may require some financial planners to broaden their scope of practice and carefully consider a range of ethical issues. while this model has the potential to help a financial planner provide more comprehensive advice by addressing factors associated with a client’s financial, psychological, and physiological well-being, planners must practice within their expertise, safeguard client data, and ensure that all services provided are in the client's best interest (chene et al., 2010; smith, 2009). comprehensive training, collaboration with health and medical professionals, and transparent communication are essential to successfully and ethically adopting this practice model. conclusion over the past decade, researchers, regulators, and certification bodies have increasingly recognized the importance of psychological and physiological influences in shaping the financial behaviors of advisory clients (heye, 2020). the ability of a financial planner to identify and respond to a client’s psychological, attitudinal, and behavioral situation when designing plans to improve not only a client’s financial situation but also their overall sense of well-being is generally recognized as a valued skill. as noted by klontz and associates (2016, p. 52), “financial planners are uniquely positioned to offer the knowledge, tools, and processes to help clients decrease their financial stress and thereby improve clients’ physical health, psychological health, occupational functioning, and relationships.” the developments and innovations discussed in this paper suggest that a new, holistic approach to financial planning that goes beyond traditionally defined client psychological models should be considered. stress management is known to improve financial worries and reduce blood pressure and inflammatory cytokines (i.e., health outcomes). instead of siloing treatments and interventions across medical, mental health, and financial domains, the future is likely based on broadly defining wellness. through the application of digital twin technologies, the model can revolutionize diagnoses and treatments through the creation of a holistic health profile. drawing from a client’s physical, mental, and financial information, a digital twin can act as an accurate representation of a client within their environment, providing pre-emptive warnings and recommendations that allow individuals to understand and take control of their well-being. adaptive digital health technology is a way forward for addressing healthcare barriers and increasing health and financial literacy across the population. disclosure statement the authors would like to acknowledge that grammarly ai was utilized to assist in the editing process of this paper. the software was employed to enhance grammar, spelling, and stylistic elements. all substantive content, ideas, and conclusions presented in this work are solely those of the authors. the use of ai editing tools did not influence the modeling outcomes or the integrity of the model. references ahn, a. c., tewari, m., poon, c.-s., & phillips, r. s. (2006). the clinical applications of a systems approach. plos medicine, 3(7), e209. financial services review, 33(2) 12 https://doi.org/10.1371/journal.pmed.00 30209 ahn, a. c., tewari, m., poon, c.-s., & phillips, r. s. (2006). the limits of reductionism in medicine: could systems biology offer an alternative? plos medicine, 3(6), e208. https://doi.org/10.1371/journal.pmed.00 30208 amagai, s., pila, s., kaat, a. j., nowinski, c. j., & gershon, r. c. (2022). challenges in participant engagement and retention using mobile health apps: literature review. journal of medical internet research, 24(4), e35120. https://doi.org/10.2196/35120 american psychological association. (2022). stress in america: concerned for the future, beset by inflation. https://www.apa.org/news/press/releases /stress/2022/concerned-future-inflation bellingrath, s., & kudielka, b. m. (2017). biological pathways to stress-related disease vulnerability in educators. in t. m. mcintyre, s. e. mcintyre, & d. j. francis (eds.), educator stress: an occupational health perspective (pp. 77– 100). springer. https://doi.org/10.1007/978-3-31953053-6_5 beshai, s., mishra, s., mishra, s., & carleton, r. n. (2017). personal relative deprivation associated with functional disorders via stress: an examination of fibromyalgia and gastrointestinal symptoms. plos one, 12(12), e0189666. https://doi.org/10.1371/journal.pone.018 9666 britt-lutter, s., haselwood, c., & koochel, e. (2019). love and money: reducing stress and improving couple happiness. marriage & family review, 55(4), 330– 345. https://doi.org/10.1080/01494929.2018. 1489374 brügge, k. u. (2001). an integrative psychophysiological model of hypertension, adrenergic activation, and memory (doctoral dissertation). university of johannesburg. cfp board. (2023). cfp® certification learning objectives. https://www.cfp.net//media/files/cfp-board/cfpcertification/2021-practiceanalysis/2021-cfp-certification-learningobjectives.pdf chene, d., gold, j., & gramlich, j. (2010). the scope and practice of comprehensive financial planning: survey results, current standards, and engagement letter recommendations. journal of financial service professionals, 64(1), 47–59. chirumalla, k. (2017). clarifying the feedback loop concept for innovation capability: a literature review. in ispim innovation symposium (pp. 1–4). the international society for professional innovation management (ispim). cundiff, j. m., & smith, t. w. (2017). social status, everyday interpersonal processes, and coronary heart disease: a social psychophysiological view. social and personality psychology compass, 11(4), e12310. https://doi.org/10.1111/spc3.12310 french, d. (2023). from financial wealth shocks to ill‐health: allostatic load and overload. health economics, 32(4), 939– 952. https://doi.org/10.1002/hec.4710 gray, j. (2023). mastering emotional client conversations as a financial planner. journal of financial planning, 36(11), 28–33. grable, j. e. (2013). psychophysiological economics: introducing an emerging field of study. journal of financial service professionals, 67(5), 15–18. grable, j., heo, w., & rabbani, a. (2015). financial anxiety, physiological arousal, and planning intention. journal of financial therapy, 5(2), 2–18. https://doi.org/10.4148/1944-9771.1083 https://www.apa.org/news/press/releases/stress/2022/concerned-future-inflation https://www.apa.org/news/press/releases/stress/2022/concerned-future-inflation https://doi.org/10.1007/978-3-319-53053-6_5 https://doi.org/10.1007/978-3-319-53053-6_5 https://www.cfp.net/-/media/files/cfp-board/cfp-certification/2021-practice-analysis/2021-cfp-certification-learning-objectives.pdf https://www.cfp.net/-/media/files/cfp-board/cfp-certification/2021-practice-analysis/2021-cfp-certification-learning-objectives.pdf https://www.cfp.net/-/media/files/cfp-board/cfp-certification/2021-practice-analysis/2021-cfp-certification-learning-objectives.pdf https://www.cfp.net/-/media/files/cfp-board/cfp-certification/2021-practice-analysis/2021-cfp-certification-learning-objectives.pdf https://www.cfp.net/-/media/files/cfp-board/cfp-certification/2021-practice-analysis/2021-cfp-certification-learning-objectives.pdf hanlon et al. 13 grier, s., & bryant, c. a. (2005). social marketing in public health. annual review of public health, 26(1), 319–339. https://doi.org/10.1146/annurev.publheal th.26.021304.144610 grieves, m. (2022). intelligent digital twins and the development and management of complex systems. digital twin, 2, 8. hanlon, b., hancock, m., lo, c., grable, j., archuleta, k., cohen, a., mazdzer, c., babey, s., miller, e., & nunez, a. (2020). feedback control for optimizing human wellness. in j. s. gero & l. s. c. (eds.), augmented cognition. human cognition and behavior: 14th international conference, ac 2020, held as part of the 22nd hci international conference, hcii 2020, copenhagen, denmark, july 19–24, 2020, proceedings, part ii (pp. 171–190). springer international publishing. https://doi.org/10.1007/978-3-03050334-6_13 hastings, g. (2007). social marketing: why should the devil have all the best tunes? elsevier/butterworth-heinemann. heye, c. (2020). introducing a hierarchy of financial wellness. journal of financial planning, 33(9), 38–41. jepson, n. a. (2008). psychophysiology and worry in irritable bowel syndrome (doctoral dissertation). proquest. jiménez, j. p., botto, a., & fonagy, p. (eds.). (2021). etiopathogenic theories and models in depression. springer. https://doi.org/10.1007/978-3-03079771-4 jovanovic, t., & norrholm, s. d. (2016). human psychophysiology and ptsd. in i. liberzon & k. j. ressler (eds.), neurobiology of ptsd: from brain to mind (pp. 292–316). oxford university press. klontz, b. t., van zutphen, n., & fries, k. (2016). financial planner as healer: maximizing the role of financial health physician. journal of financial planning, 29(12), 52–59. lehrer, p. m. (2021). psychophysiological and systems perspectives on stress and stress management. in principles and practice of stress management (pp. 16–38). guilford press. lehrer, p. m., irvin, c. g., lu, s.-e., & wamboldt, f. s. (2022). relationships among pulmonary function, anxiety, and depression in mild asthma: an exploratory study. biological psychology, 168, 108244. https://doi.org/10.1016/j.biopsycho.2022 .108244 lehrer, p. m., & eddie, d. (2013). dynamic processes in regulation and some implications for biofeedback and biobehavioral interventions. applied psychophysiology and biofeedback, 38, 143–155. https://doi.org/10.1007/s10484-0139227-6 lehrer, p., & moritz, g. (2023). biofeedback has therapeutic effects on asthma, although additional research is needed to document specificity. policy insights from the behavioral and brain sciences, 10(1), 83-89. lehrer, p. m., derby, l., smith caswell, j., grable, j., & hanlon, r. (2024). physiological effects of psychological interventions among persons with financial stress: a systematic review, meta-analysis, and introduction to psychophysiological economics. applied psychophysiology and biofeedback, 1– 19. https://doi.org/10.1007/s10484-02409595-7 lovallo, w. r. (2015). stress and health: biological and psychological interactions (3rd ed.). sage publications. lucas, a., hardie, j. h., & yim, s. s. (2021). pushed together or pulled apart? economic stressors and romantic relationship quality. sociological perspectives, 64(4), 563–586. financial services review, 33(2) 14 https://doi.org/10.1177/0731121421995 244 mcewen, b. s., & wingfield, j. c. (2010). what's in a name? integrating homeostasis, allostasis and stress. hormones and behavior, 57(2), 105–111. https://doi.org/10.1016/j.yhbeh.2009.09. 011 mcgrady, a., & moss, d. (2013). pathways to illness, pathways to health. springer. https://doi.org/10.1007/978-1-46145395-6 miller, e., hanlon, r., lehrer, p., mitchell, k., & hancock, m. (2023). intelligent wellness. in international conference on human-computer interaction (pp. 232– 249). springer nature switzerland. https://doi.org/10.1007/978-3-03135762-4_17 mitchell, r., schuster, l., & jin, h. s. (2022). playing alone: can game design elements satisfy user needs in gamified mhealth services? health promotion international, 37(2), daab168. https://doi.org/10.1093/heapro/daab168 moore, i. n., snyder, s. l., miller, c., qui an, a., blackford, j. u., zhou, c., & hickson, g. b. (2007). confidentiality and privacy in health care from the patient's perspective: does hipaa help? health matrix, 17, 215–238. njegovanovic, a. (2018). neurological aspects of finance, transmitters, emotions, mirror neuronal activity in financial decision. маркетинг і менеджмент інновацій [marketing and management of innovations], 3, 186–198. https://doi.org/10.21272/mmi.2018.3-17 overmier, j. b., & murison, r. (2013). restoring psychology's role in peptic ulcer. applied psychology: health and well‐being, 5(1), 5–27. https://doi.org/10.1111/aphw.12000 patel, p. c. (2019). the great recession and allostatic load in the united states. international journal of stress management, 26(4), 411–417. https://doi.org/10.1037/str0000132 pew research center. (2015, april 1). u.s. smartphone use in 2015. https://www.pewresearch.org/internet/2 015/04/01/us-smartphone-use-in-2015/ porcelli, a. j., & delgado, m. r. (2009). acute stress modulates risk-taking in financial decision making. psychological science, 20(3), 278–283. https://doi.org/10.1111/j.14679280.2009.02288.x ramakers, i., van den houte, m., van oudenhove, l., van den bergh, o., & bogaerts, k. (2023). end-tidal co₂ in patients with panic disorder, stressrelated or functional syndromes, versus healthy controls. applied psychophysiology and biofeedback, 48(2), 149–157. https://doi.org/10.1007/s10484-02209540-7 rothschild, m. l. (1999). carrots, sticks, and promises: a conceptual framework for the management of public health and social issue behaviors. journal of marketing, 63(4), 24–37. https://doi.org/10.1177/0022242999063 00404 smith, j. (2009). professionalism and ethics in financial planning (doctoral dissertation). victoria university. wheeler, b. e., kerpelman, j. l., & yorgason, j. b. (2019). economic hardship, financial distress, and marital quality: the role of relational aggression. journal of family and economic issues, 40, 658–672. https://doi.org/10.1007/s10834-01909629-3 world health organization. (2023). mhealth. https://apps.who.int/gb/ebwha/pdf_files/ wha71/a71_20-en.pdf https://www.pewresearch.org/internet/2015/04/01/us-smartphone-use-in-2015/ https://www.pewresearch.org/internet/2015/04/01/us-smartphone-use-in-2015/ https://apps.who.int/gb/ebwha/pdf_files/wha71/a71_20-en.pdf https://apps.who.int/gb/ebwha/pdf_files/wha71/a71_20-en.pdf pii: s1057-0810(00)00049-4 applications of www technology in teaching finance stuart michelson*, stanley d. smith department of finance, university of central florida, orlando, fl 32816-1400, usa abstract in this article we discuss the need for a personal web page in finance education, some of the ways web pages can be used to benefit students, and present a survey of web page use. we follow with a discussion of how to create and modify a web page using commonly available computer programs. this information allows faculty the ability to provide valuable and timely information and materials for students. © 1999 elsevier science inc. all rights reserved. jel classification:i220 keywords:financial education; internet; web page 1. introduction: why start a web page? in the spring of 1998, a class was informed that e-mail would be used in the fall semester to provide better communication between instructor and students. the students encouraged the instructor to use a web page instead because it was much easier for them to access and use. they cited benefits such as: computer availability on campus, not needing access to e-mail accounts or software, ability to access course files as necessary on their own schedule, irregular checking of e-mail, and dislike of university e-mail accounts. with this charge from the students the instructors decided to take action to create a web page during the summer. one coauthor offered to set up a basic web page for each of the finance faculty. with a basic page, each faculty member was able to make changes to meet individual objectives. it was amazing that creating and maintaining a web page was as easy as word processing. * corresponding author. tel.:11-407-823-6550; fax:11-407-823-6676. e-mail address:stuart.michelson@bus.ucf.edu (s. michelson). financial services review 8 (1999) 319–328 1057-0810/99/$ – see front matter © 1999 elsevier science inc. all rights reserved. pii: s1057-0810(00)00049-4 in the next two sections we present a review of the literature and discuss some of the ways we have used the web pages to benefit students and faculty research efforts. in the fourth and fifth sections we present results of a survey and discuss examples and applications for teaching with web pages. in the sixth section of the paper, we discuss creating and modifying a web page using commonly available computer programs. 2. literature review the literature on the internet is fairly recent, but rapidly expanding. in this section we review several current articles published in this area. anthony herbst (1996) presents an interesting discussion on the fundamentals of the internet and describes the opportunities and challenges for financial engineers on the internet. russ ray (1996), in “an introduction to finance on the internet,” presents a background on the history and evolution of the internet, as well as providing a description of financial resources available. in “a guide to locating financial information on the internet,” james pettijohn (1996) provides a guide and extensive list of financial resources available on the internet. brian grinder (1997) discusses recent developments on the internet and provides an extensive list of financial service internet sites. smith (1996) reviews the use of edgar as a resource for sec documents and applications for classroom use. there are several additional articles that provide information about the use of the internet in teaching and education in the accounting and economics areas [for example see: agarwal and day, 1998; debreceny, smith, and white, 1996; and manning, 1996]. most of these articles provide lists of finance-related internet sites, many that can be described as extensive, but none can be totally comprehensive. the web and the internet are changing and expanding so rapidly that most lists of resources become incomplete as soon as they are published. in fact the web becomes a natural resource for publishing lists of relevant links, since a web page can be updated on a continuous basis. the purpose of this article is not to provide a long list of internet sites, but rather we attempt to describe the benefits, uses, and applications available for a finance faculty member’s web page. 3. web page organization and content a typical method for organizing web pages is to develop a simple home or index page that provides basic personal, professional, and course information. then, each major topic area is a separate, linked page off of the main home page. examples of linked pages include: separate pages for each course, pages for relevant links grouped by topic area, pages for class notes and assignments, schedules, resume, and so forth. in teaching, common applications such as syllabi, lecture notes, schedules and class assignments are used. one of the most important elements to incorporate on a web page is a continuously updated schedule. the class schedule links to information on class lectures, examinations, teaching notes, project and term paper assignments, and related items, for example, articles and sources of information on the internet. students have responded positively to these uses. 320 s. michelson, s.d. smith / financial services review 8 (1999) 319–328 there are many other links that could be provided on the web page. there are free databases and links to proprietary databases, for example, moody’s online, and proprietary bibliographic research tools, for example, abi/inform and econlit. the proprietary links will usually require an id and password to access, since most are available by subscription. other links include interesting articles and recent reports on bank profitability, real estate lending, consumer credit, commercial lending, credit reporting and scoring, trusts, international lending, mergers and changes in delivery of products and services, and interest rates, stock prices, and domestic and international market indexes. additional links include banking publications with free online access, for example,american banker, federal and state regulators of depository institutions, and national and state trade associations for depository institutions. for the study of financial markets, links include information on finance companies, life and property and casualty insurance, pension funds, investment companies, mortgage markets and securities, and stock, bond, options and futures markets. most of the fact books and other data related to these areas had previously been obtained in hardcopy. from the web page the latest information can be accessed very quickly at no cost and reduce the space needed for this data in hardcopy, as well as make this information easily available to students. the web page can be an excellent “information organizer.” examples include links to the most recent government statistical releases on interest rates and other market factors. these links often provide easily downloadable current and historical information. many articles and reports by government bodies are now available on the internet. links to these sources of information can be developed easily by highlighting the article or statistical release title and typing in the address (uniform resource locator or url) for the respective data source. links to investments web sites include individual mutual fund pages, financial markets (nyse, amex, nasdaq, cbot, cme, etc.), brokerage pages, sources for quotes on various markets, as well as sources for obtaining firm specific financial information. another common application is to provide current market index information regarding specific stocks or other financial instruments. these links can be organized or grouped as the instructor may desire. there are many relevant links in the area of corporate financial management. the most obvious are links to the individual firm’s web pages, as well as government links, such as edgar, the sec, and proprietary sources, such as hoovers (portions are currently available without a subscription). these links provide the annual and quarterly statements filed by each corporation. in class, the web page and another organization and access management tool (such as webct) can become a site for official course information. students may be required to regularly check the web site for course information or risk missing important information or updates on assignments, meetings, and data sources. within a day of posting new information on the web page, we typically receive feedback from students indicating their response to new information, as well as notification that they have passed the information on to their “technologically-challenged” classmates. for a “media enhanced” class, which might not meet regularly, this is an excellent communication tool. the instructor may also post student grades continually updated throughout the semester 321s. michelson, s.d. smith / financial services review 8 (1999) 319–328 on the access protected web page. the access management tool, webct, can be password protected, with a unique user id and password for each student. this allows each student to login to his/her own secure area and view only his/her grades. selected statistics on the exams, including mean, median, standard deviation, and a histogram can also be included. this facility allows each student to gauge how they performed relative to the rest of the class. another vehicle for posting grades is to use an excel spreadsheet saved as html. saving the spreadsheet using random student identifying numbers as student id’s allows privacy in posting. instructors can also provide graphs and statistics generated by excel on the page. examples of these applications may be accessed at: http://www.bus.ucf.edu/ssmith or http:// www.bus.ucf.edu/michelson. 4. teaching and learning with web pages: a survey after one semester of use our student evaluations contained several comments that the web page and links were among the “most liked parts” of the course. the graduate students were even more emphatic about their approval of the web page. comments indicated that many graduate students planned to use these web pages after graduation and that they would maintain a membership in the alumni association, since membership includes access to some of the proprietary links on the web page. we recently surveyed two classes of about equal size on the use of the web pages and the results were very interesting and positive. the sample consisted of 90 students, with about 94% seniors and 6% juniors. all of the students were finance majors. the students indicate that about 45% of their university classes use web pages. we found that 98.8% of our students use the course web pages. the primary uses of the web pages include: links 71% syllabus 80% download powerpoint slides* 80% project instructions & examples* 98% general course information 56% other 12% *these data are for one class only because this content is just provided in the one class and the statistic is zero for the other class. the majority of our students (89%) found our web page very easy to use and very helpful.# they indicated that the web pages definitely provide additional material or content (81%).# on average these students used the web pages 11–20 times during the semester and they accessed the web pages one-to-two times per week. we found that 81% of the students “definitely like” the web pages.# [#1 & 2 on a five point scale.] several students provided additional comments on their surveys; the following is a brief summary: ● “the web pages are very helpful, especially for a night student that has very little access to campus during the day.” 322 s. michelson, s.d. smith / financial services review 8 (1999) 319–328 ● “the links are great, they helped with job interviews.” ● “the links were extremely helpful in this class and for other classes.” ● “the links to the library resources were helpful as well.” the students also provided suggestions for improvement of the web pages: “add more links to job sites for the banking industry.” “post grades on web page.” (this is currently done in one class, but not in the other.) “provide more job and career links.” “rearrange your web page with the class information at the top. the results of the survey indicate that students are very receptive to web page use in our classes. they are extremely positive on web page use, they use the pages frequently, they feel the pages provide valuable information, and they freely make suggestions regarding additional information to be added to web page content. 5. teaching and learning with web pages: project examples within our finance department, four faculty members teach five different undergraduate and graduate courses in the financial services area. some of these faculty members use the authors’ web pages as a resource for their classes. projects are assigned and students are informed that they should review the authors’ web pages to access data sources that may be useful. students from another class will often inform us that one of the links was not working. the link can usually be updated in a few minutes while the student is in the office. this is an opportunity to get feedback from the student regarding use of the information. changes that will be useful to the students in the other classes are shared with other faculty so their students can be informed. another interesting side effect resulted from mentioning or demonstrating the web page in interacting with area bankers. one bank has indicated that the web page has been recommended as a resource to some of its employees. as a consequence, a bank research analyst advised us of additional potential sources of information. three examples of using the web page in class projects will be described. the first is a term project that requires the student to conduct a financial analysis of one of the top 100 u.s. bank holding companies (bhcs). each student selects a different bhc. the project introduction material is presented with “measuring and evaluating bank performance” in a commercial bank management class. the project provided realistic applications in the course material using real bhcs to lay a solid foundation for more specific topics discussed later in the course. the students present their analyses allowing comparisons among bhcs with respect to lending portfolios, fee income sources, funding sources, capital adequacy, and market measures of risk and relative values over time. the following discussion highlights areas within the project that describe links to other information in the web page and outside resources. the project utilizes different reports from the federal reserve system’s national information center of banking information (nic) at ,http://www.ffiec.gov/nic.. at this location click on the top 100 bank holding companies to pick one of the bank holding companies. to obtain the performance report—summary 323s. michelson, s.d. smith / financial services review 8 (1999) 319–328 ratios for bhcsfor the bhc, under type of information, select performance ratios and the latest end of year date, for example, 3/31/99, then submit. the summary ratios for 3/31/99, 12/31/98, and 3/31/98 should appear. the students are required to attach this information as an appendix to the project. the students also print out and save to disk additional reports for the company that will be used later in the semester for lectures and daily exams. these reports and examples include consolidated balance sheet, consolidated income statement, securities, loans and lease financing receivables, and deposit liabilities for the last four quarters. the students retrieve a company profile from yahoo. using an internet browser, for example, netscape, they enter the address http://biz.yahoo.com/p/first letter of ticker/ ticker.html. the ticker refers to the company’s stock trading symbol, for example, sti for suntrust banks. for example, for suntrust the address would be,http://biz.yahoo.com/ p/s/sti.html.. a two-page company profile should appear. a small stock performance chart appears at the end of the profile. when students click on the chart and 1 year link it produces a large chart of the company versus the s&p 500 index. the profile and the chart are required to be included as an appendix to the project. students are required to deliver a formal presentation and a written report. this project has evolved over time. in 1996 this project was similar to a project for nonfinancial companies that is described in one of the first articles about using the internet to teach finance and business classes (s. smith 1996). the project was limited to producing financial ratios from information in the company’s 10k annual reports or the 10q quarterly reports and analyzing the performance of the company over time. there were no peer comparisons and the relative market performance analysis was less detailed. using the national information center data, the information is much easier to obtain because links are provided for specific reports and a standardized form is used for all the bhcs, unlike the 10k and 10q reports. most of the common ratios that are used in banking analysis are provided for different time periods with peer group percentiles and means relative to specific banks. the yahoo link mentioned earlier, http://biz.yahoo.com/p/first letter of ticker/ ticker.html, allows comparison of betas, price/book and price/earnings ratios, and graphic comparisons of the stock performance against the s&p 500 index over various time periods. as the world wide web evolves, additional sources become available offering the potential for improved data, leading to better project results. a second project example includes using web page links to describe the project and to access available proprietary databases of the library. this project requires the student to prepare an outline or paper on “management of a financial product or service.” each student selects a different product. the major sources for recent articles and reports isfirstsearch, which allows access to different bibliographic search tools such as abi/inform, business & industry, wilson business abstracts, econlit, and so forth. the major type of search is based on key words related to the selected topic. these search tools provide very current information on most of the major practitioner and academic publications in the financial services area. the objectives of the project are to gain a deeper understanding of managing a specific financial product and to learn how to conduct research using more advanced research methods. before this project, most students think that research is using a common search engine like yahoo. they are very impressed with the power of the 324 s. michelson, s.d. smith / financial services review 8 (1999) 319–328 bibliographic search tools. at the graduate level, an interview with an executive related to the product is also required as part of the paper. a third example in the financial services area is to have students select a money market instrument, predict a future growth rate or price based on a financial model developed by the student, and track the performance during the semester. the prediction models are usually based on one or more financial instruments or economic data, employing a regression model using one instrument’s interest rate as the independent variable and the selected instrument’s interest rate as the dependent variable. web pages are used to provide links to up-to-date databases for this information. online federal reserve statistical releases(http://www. federalreserve.gov/releases)are often current within one business day and historical interest rate information can be easily downloaded for research purposes. a question typically asked is, “how much time and effort does this take?” the initial creation of the web page takes very little time and is discussed in the next section. one can add to the web page over time as schedule and opportunities allow. adding one or a few links takes very little time. when a faculty member or student has a question, it is typical that we access a link on our web pages for a source of information. the web pages have become a library and information organizer. as components are added to the web page, the benefits are more than worth the small investment of time. 6. designing and maintaining a web page one intention of this article is to provide information to allow faculty to easily design and maintain a web page. the authors, working together, were able to use these methods to allow a “web novice” to quickly navigate the learning curve and efficiently design his first web page. the first thing to consider when developing a web page is what editor to use to create the page. in other words, what software will actually put the words, pictures, and graphics on the page? there are many choices in software, but the first that come to mind are microsoft word, netscape, and microsoft frontpage. while frontpage is an extremely powerful web page editor, we will not discuss it in this article because it has a steeper learning curve. that leaves ms word and netscape, fierce competitors, but similar in use and application. we suggest the use of ms word for readers who anticipate that their web page will be viewed more frequently by internet explorer. for those who expect netscape to be the primary browser, we suggest that netscape serve as the editor. note that good web page design calls for testing the web page for accurate viewing under both internet explorer and netscape. we begin our description of creating a web page by using netscape as the editor. the application of word will be very similar in principle. several key points or questions to consider when designing a web page layout: ● what is the important, relevant information to include on the web page? ● consider the content carefully to avoid a cluttered page. ● design the page for efficiency and ease of use. 325s. michelson, s.d. smith / financial services review 8 (1999) 319–328 ● be aware of the loading time for modem users. pictures, sound files, videos, and java script may load very slowly over a 28,800 baud modem. with netscape version 4.01 (or higher) running, go tofile, new, blank page (or page from template or page from wizard).[the standard notation for this article will use italics for commands to select from the editor.] this will produce a blank page to start editing. now simply type in the content on the new page. typical pages include name, position, contact information, course information, brief resume, and links to other pages. for now, don’t worry about formatting; simply type in the content. also notice that it is necessary to press enter at the end of each line, since the editor does not “line-wrap” automatically. web page design templates are also available using most editors, including netscape, word, and frontpage. usually they provide a content outline, a graphical background, and preformatted pages. many universities also provide templates for faculty use to allow standard formatting across all university web pages. next, format the page similar to a word processing document, with a few differences. to make the page clear and easy to read, make the font large and provide plenty of white space around the text. first (with the mouse) highlight the text that forms the title of the page, then click on the desired font size. notice that in some programs the selection is for actual point size, although in others the font choices are limited to small, large, extralarge, and so forth while the text is still highlighted, click on the icon for bold text and the icon for centering text (under alignment). if the icons are not apparent, these commands may be found under format on the tool bar. to change the color of the text as desired, click on the color icon. now format (size, typestyle, bold, italic, colors, centered, etc.) the remainder of the text in a similar manner. with this work completed, the next step is to save the page. go tofile, save as, and then name the file. most web file servers have a required default name for individual home pages, such as home.html or index.html. check with the institution’s resident web manager or technology resources personnel for this information. the web manager may also specify a location (or folder) to store (save) files. note that netscape asks for a title the first time the file is saved, and this title will show up in the title bar of many web browsers. netscape has a nasty habit of occasionally changing the file name or location when saving an existing file. for this reason, we recommend always saving web pages usingsave as, rather thansave, to verify that the file name and location are correct. if the web operating system is windows 95 or windows nt, save the file to a folder as described in this article. if the operating system is unix, readers will probably need to use ftp to transfer the file to proper directory for web server access. even though this is an extra step, the windows version of ftp (file transfer protocol) is fairly easy to use. once the web page is closed, to make further modifications click onfile, edit page, while viewing the web page in netscape. note that this method may be used for editing the page any time in the future. now it’s time to dress up the web page. readers may want to link specific words in their web page to another web page. highlight the designated text then right click the mouse. choose the selection “create link using selected,” then fill in the input blank (“link to a page location or local file”) with the web page location, for example,http://www.uofs.edu. 326 s. michelson, s.d. smith / financial services review 8 (1999) 319–328 now the designated text should be a different color and underlined. scroll the mouse over the text to see the link location at the bottom of the screen. alternatively, highlight the text and then click on the link icon and follow the steps above. another method to add a new link requires clicking oninsert, link, and then filling in the input boxes for “enter text to display for new link” (this is the actual text to link on the page) and “link to a page location or local file” (the link location). readers can use these same methods to link to other pages, for example, individual course pages, other web pages, link pages, or links to take browsers back to department or college pages. it’s also easy to add a graphic or picture to a page, perhaps a graphic of the school’s mascot or insignia. first obtain a copy of the graphic or picture, which are normally in gif or jpeg format. these are standard web page graphic formats and are usually not very large files. copies of the gifs or jpegs are typically available from the web manager or by copying. to copy them, go to the web page with the graphic, then right click on the graphic and click on “save image as.” for the file location, specify the same folder that contains the personal web page. now that the graphic is in the correct folder, position the cursor on the desired location of the personal web page to place the graphic. click on the insert icon and either type in the file name of the graphic (under “image location”) or use “choose file” (much easier) and highlight the name of the graphic. after closing the input box, the graphic should appear. note that images may also be linked to another page or image. to do this, click on the “link” tab (while inserting the image) and fill in the “link to page location or local file” input box. when one clicks on the graphic (and the cursor changes to a hand), the graphic can be appropriately placed on the web page. note that when moving the cursor to the edge of the graphic, it changes to a two-sided arrow that allows for resizing the graphic larger or smaller. right click on the graphic, selectimage propertiesand theimagetab, to select choices with respect to text alignment and text wrapping. to further improve the web page, dress up the background. right click a blank area of the page and selectpage propertiesandbackground. from this input box select the background colors and designs or a graphic gif or jpeg if available. one can also access the same background menu throughformatand thenpage colors and properties. remember that the background should not be so cluttered as to interfere with reading the content. as a final reminder, remember to save the page frequently to avoid losing work. to view the completed web page, exit from the netscape editor then go tofile, open page, specify the file, or choose file, check thenavigatorblank and view the new home web page. for an example of a completed page see either http://www.bus.ucf.edu/ssmith or http://www.bus.ucf.edu/ michelson. recall that to edit the web page further, click onfile, edit page, while viewing the page in netscape. 7. conclusions and recommendations this article provides information on uses and applications of web pages for finance education. suggestions are provided for class-related use and three example class projects are described. the results of a survey of students indicates that the majority are receptive to web page use, found web pages helpful and easy to use, and encourage faculty to develop 327s. michelson, s.d. smith / financial services review 8 (1999) 319–328 additional web page applications. our survey results suggest that students want faculty to utilize the web for distribution of information and for providing relevant resources. faculty members who are hesitant to incorporate course web pages and resources in their classes may be in danger of not staying current in their teaching. in this article we also provide the information necessary for someone new to the web to easily develop a basic web page. it is impossible to cover all of the many facets of web page development without unduly complicating this discussion. there are many manuals on web design and development available in the bookstores. therefore we have subscribed to the principle of “keep it simple” to allow the reader to quickly develop his/her own web page. this knowledge will allow the reader to provide valuable and timely information and materials to their students now! acknowledgment the authors wish to thank the special issue editor, jill vihtelic, and two anonymous referees for very helpful comments. any errors are the responsibility of the authors. references agarwal, r. & day, a. e. (1998). the impact of the internet on economic education.journal of economic education, 29, 99–110. debreceny, r., smith, g. s. & white, c. e. (1996). internet methodologies and the accounting curriculum: a first look. accounting perspectives, 2, 107–124. grinder, b. (1997). an overview of financial services resources on the internet.financial services review, 6, 125–140. herbst, a. f. (1996). the ways in which the financial engineer can use the internet.financial practice and education, 6,111–121. manning, l. m. (1996). economics on the internet: electronic mail in the classroom.journal of economic education, 27, 201–204. pettijohn, j. (1996). a guide to locating financial information on the internet.financial practice and education, 6, 102–110. ray, r. (1996). an introduction to finance on the internet.financial practice and education, 6,95–101. smith, s. d. (1996). using edgar on the internet to teach finance and business courses.journal of financial education, 22, 76–78. 328 s. michelson, s.d. smith / financial services review 8 (1999) 319–328 pii: s1057-0810(02)00099-9 � ����� �� �� ���� � ��� ������ ������� ���� ���� �� � ������� ������� ���� �� �� ���� ��� �� ��� �� �� �������! ����������� ����� �� ����� � ����������� ��� ������� ����� ����� �� �� ����������� ��� ���������� ������ � �� � � !� ����� ��� ����� � ����������� ��� ������� ����� ����� �� �� ����������� ��� ����� �� � "�#��� �� !$$�% ����� �� �� �� ���� &��� !' (� �� �� !$$�% ����#��� ' ������� !$$! �������� ���� ������� ����� � ��� ������� � ����������� �� ������ �� ������� ����� �� �� # ������� ����� �������� �� ��� � �� ��#��� �������� )��� ����� � �������*�� ��� ���+��� ������ ��� � ������� �� ���������� ,��� �� ���� � +������ ��&�������� �� ������� ����� )�� ���+�� ��� ���% -���� ��� ���� ������ �� � ���� ������ �& +������ ����� ��������� ��#����� �� ��� .���� +������ ������ ���� ���� ��� ��#�������� ��#��� &�� ��� ����� � /� ������� ��� ������� ���� ���� � �� � �� ��� ��������� ��&�,��� ��� ����,��� �� ������� /� � �� #�� ��� ��#� �� ���� ������� ��, �� ����� ������ ��� #������� ���� �& ��� ����� � !$$� �� ��� "������ 0��� � ������ ����� ��� "#$ ����� � ��� �% �!% �1 &��' �(�% )������ ����% ���#���� ��&�,���% )������� ��� �% 0� ��������% 2� ������ � ���� ������ ���� ������� ����� � �� � ��#������ �� �������� �� �� ��� � 33������� ����44 � �� � � �� ����� ��������� )�� ���� 33������� ����44 �� ����,��� �& � ��������� ����� �� ��� ��� ����� ����� ��� � �������� ������� �& +������ ��� �� ������� ����� ��� �� ��� ������� ����� # ���� )�� ������ �������� �� ����� ����� �� ���� ���� &�� �� ��� �������� ��� ����� � �� � � ��� ��� �������� �& ����� ������� �� ������������ ��� ��#�������� ��#��� � 5&&������� ����� ���+��� ��� ��� ����� �� ��� �� ��� ��� ����� ��������� ���� �� ,� ���� �� �� � ��� ����� )��� ����� � #�� ���� � ���������� �& ��� ���+��� ��� ��� ������ ���������� ,��� �� ���� � ������� ����� 0� � �� ��� ���� �������� ���� ��������� � �� � � &�� ��� ����� 6������� "�� ���� �� ��, �$ 7!$$�8 !$9:!!$ �$;'.$1�$<$�<= : ��� &���� ������ � !$$� �� ��� "������ 0��� � ������ ����� ��� >00? " � $ ; ' . $ 1 � $ 7 $ ! 8 $ $ $ 9 9 . 9 � ������#������ ������� )� �? ��.1@a.@;@.$;a'% &��? ��.1@a.@;@.b'a1� #����� �((������% � ����cd� ��������� 7���� � ������� ���8� �����d� ��������� 7���� ����8� ��� ���d� ��������� 7���� �� �����8� � )� �? ��.1@a.@;@.ba@@% &��? ��.1@a.@;@.b'a1� ! )� �? ��.1@a.@;@.!!a9% &��? ��.1@a.@;@.b'a1� ��� ����� �� � ���������� �� ��� ������ ��� ������������ �& ��� ����� 0� ��������� �� #�� ���� ����� � ��� ������� &���� &�� ��� ����� ��� #������� ���� �& ��� ����� "� ��� ������� ������� ��� ��� �& ���#����� ���<�� � �������� ���� ������� �� +������ ���������� � ��� � ��� � ��� � 7�99;8 ��� ��� ��*�,���� >������ ��� e��� �, 7�99b8 ��� � ��� �� �##����� �� ������� ��� � � �& ���#���� ��� �*����� ������ +����� �������� 0� ��������� >����� 7�99@8 ��� ��� ��� ",����� 7�99;8 ������� ��&&����� ���� �& � �������� ���� ������� �� +������ ���������� 6�, ������� ������� ��� � �� � � ��������� ��� �������������� ����������� ��� ������������ �& � ������� ����� )��� ����� � + � ���� ��#� � ������� 0������ ��� ������ � #��c��� ���� ��-����� ���� �������� �� �� � &��������� ��� ���� �& � ���#���� � &�, ����� ���� � ��c����� �& ��� ���� � ������� �#��� �� ���� � #��c��� ,�� � � �� ���� �� ������ ��� ������ )��� ,�� � �� �� ��� � ����� ��� �# ; ����� �& +������ ��������� ���� �� �����+���� ���� �����&�� ���� ���� �������� � ���� � �#��������� �� ���� ��� � �� ��� +������ ����� ��� ���� ��-������ ��#�&� �� ��� �� � �� � ����� ���� �� ��� �*� ��� ���#���� ��� ��� �� � ����� &�� �� ��� ���� �& �,� ���#����� ,����� ��� ���� ��������� )����� ,��� ������ �� ��� � �������� ���� ���� ��#��� � &���� �� � ������� ����� ��� �������� ��� ������ ; ����� �& +������ ��������� ����� ��� ������ ��� �� � �� �� �� ��� &����� � + � �� ����� ; �������� 0� ��� #������� ���� ��� ����� ��� ������� � � �& ��� � � � +������ �������� ,���� �� ���� #����� �� ��� ���� ���� �� ��� ����� f��#��� ���� �������� �� ����� ��� &���� �*� ��� ���+�� ���� � ������� ���� ��� #�� ��� �� �������� ��� &��� ��� )�� ���+�� ���������� ,��� �� ��� � ������� ���� � �� � � ,����� ���� ������� ����� ��� �� ��� � � � ����� 6����� �� #������ ��� �� ��� ����# �� �� ��� � �������� ������� �& +������ ���� �����+���� � ������� ��� ������ �& ���� ���� �������� ��� &��� �� �#��� �� ������ +������ ���� �� ������ ���#����� �� �������� �� ������ ,�� ��� ���� ��� 0������� ��� ������� ���� � � �� ������ &�� ��������� � ��������� �� ������ ���� � ����� ��� �� � � �� �& ���� ��� �&&���� 6������� ��� � �������� ���� ������� ���� �� ��� ������� ���� ,� �� � � ���������*�� &������ �� ��� ������� �& ,���� �� ��� &�� �#���+� ��&�������� ���� �� � � �� �� ����� ������ ���� &�� ���� ���� ���� ���#���4� +������ ����������� "� ��� ������� �� &��� �� ��� �#��� ���� ���� ���� ����� ����� ���� �� ��� �*�� ������ ���� ����� ����� �� �� ���� ,���� �� �� ���+� �� ������ �������� "������ ��� �������� �& ���� � �� � � ������� ����� ��. ��� +������ ���� ���� ��� � ���� ������ ���� ��� ��� � ��� ������� ���� �� ������� �� &��� ���� ���� ��#��� � ���� �� +������ ��������� ���� �#���+�� �� ��� ��. ��� ���� ���� ���� ,����� � ������� ���� ��� �&��� ��� ���� ���� ������� �������� � ��������� �� ����� �& �������� ��� #������ �� ��� ���� ��� ���� ��� ��������� ��� -���� �� �� �������� ����+�� ��� ���� ���� �& ��� ������� )����� ����� ���� ���� ������� � �� �& ��&�������� ���� &��� �� ��� �� ��������� "� ������ ��� ���+�� �� ��� � �� ��������� ����� ��� � ������� � ���+�� �� ��� � �� ��������� 0� &���� ��. ��� ���� ���� ���� �� � ��� ���� ������ �� ������� ��� �� ���� ����� � ���������� ���#������� �� ���� ���� ���� ��� ��#��� � ���� �� +������ ��������� ���� �� ���#����� ��� ��">� "���� ��� ��. ��� ���� ���� �#���� �� �� �������� ����� ���� ��#��� � �� � � �� ����� �� ������ � �� ���#����� ��� ��">� !�$ ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0��6������ � ,� &����� ������� ���� ��� ���� ��� ���� ��� �& ��� +������ ������� ���� ��� � �������� )�� ����� ������ �& ��� ,�� � ������� ���� �� �������� ������� �� #� #�� ��� � ������ ���� ����� &�� �������� "������� ������� ��� ���� ����� ��� ������� ������� ������ ��� ,�� � �� � ��� .���� ����� "������� ��� ��� ���� ���� #���&� ��� �� ��� ������� g�#������ �� ��� ��&�,���� ��� �������� ��� ����� � #������#��� �� ���� ���� �������� ,���� ���� ��� ���� �� �� �� #�������� �� ���������� �� � � ����� �������� ������� ,����� ��� ����� 6��� �� ��� ������� ���� ��� �� � ��� ������� ����� � ������� ������ �& ��#�������� ��#��� � 0� �������� �� ��������� ��� � ������� ��#�������� #�� ����� ��� �������� ,��� �������� �� ��� ��� ���� �-��#���� ���� ��� �� ������� ��� �� � ��� �������� � ���#����� � �� ������ ,��� �� ����� �� # �������� 6������� ��� �-��#���� ,����� ��� ������� ���� �� �&��� ��� �#��� �� �� ����� �� ����� � ������ �& ������ ����� ��#����������� �� ��� ����� � g�#������ �� ������� ��� ������� ��� ������� ���� ��� �����+���� � ��#��� ��� �� ����� �& ��� ������ ����� � )�� ���� ���� ������ �& ��� ����� ��� � �������� ������ ��� ��� ���������� �& ��� ������� �& ������ � ������ ��� ��� �� ��� ����������� � �� �& ��� #����� � ������ � ��� � ������� ����� � � ��������� ������� ����� ��� �� ��� &��������� �& � ������� ���� ��� ��� ���� ����� 0� ����������� �& ������� ������ ,� ���� ���� ����� �� � ��� ���� ������� ��� �������� )���� ��� ��� 33 �� �����44 �& ��� ����� ������ ���� ��� ��������� ��� ���� ��� �&��� &���� �� ��� ���� �&+��� �& ����� ����&�������� +���� ��� ���������� 0� ����������� �& ����� ���� ���� ��� #������ ����� �& �������� ���% ,��� ���������� ��� �� ����� ��� &��������� ��� #���� ���� � �� � �� ��� ���#����� ��� ��,� ��� ���� ��� ��� ���� �& ��� ���� ���� �� �� �������� ����� ����� ���� ���� ��� ��� ����#���� �� )�� � ��� ��� +������ ��&�������� ��� ��� ��� ���� ���������� ���� &��� � � � � �������� ��� ����� �-������� �������� +��� ������� ����� �������� ����������� ����� &����� ������������ &������ ��������� ��� &������ ��� �#������ ��������� ������ &��� ��� ��-����� �� ������ ������� �������� ���� �� ��� h����� �������� ��� �������� )�� � ��� ��� # ��&��� ,�� ������� � ���� ����� �� � +��� ������ #������� ��� ��� ��#������� �� #������ �� � ������ �� ���� ����� )�� � ��� ��� ��� ��� #�� ���� � �� �$ ����� �& &��������� ����� )�� ���� ��� �&&��� ��#���������� ��� ����� ��� �� ��� ��� �## ������� >���������� 0����&��� 7�>08 � �,� ���� ���� �� �������&� �## �������� ���� �� ��� � >���� ���� �� � �� � � �� � ��������� ���� ����� ��� � �� !$ ������ �� &�� �� ����� ������� �� � ��� .���� ����� )�� ���#���� # ��&��� ��� ���# �� ��&�������� � �� � ������ �& ���� #������� &��� �������� #������ ���� �� ����� �� �������� �� ����� #������ ���� �� ���� �� ������ )�� ���#���� ��&�,��� &������� � ,��� ����� �& �������� ��� �� ��� ����� �� ��� � ������� ���� ������ ������� �� � �#��� ��� ����� i�,� (�,� ����� ��� #�� ���� � � �� !;$ ��,� ��� ���� ,�� �,���� ��� ����� � ��� ��� ��� g�, ������ ��� ��� � �������� ������� �� ����� ������ ������-���� )�� �������� �� ��� ��� � ��� ��� ��� ��� ��� ���� ��� ����,��� ��� ���� ��� ��� �#������ )�� � ��� ��� +������ ��&�������� ��� ��� ��-����� � ������ &�� ���� ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0��!�� ,����������� )�� �������� �������� &�� ��������� ��� � ��� ��� ���� ��� �� ��� � �� ����,��� �� ��� � ��� ��� 6 ��������� )��� �� �� 0������� ���� ������ ���� #������� ��&�������� ������� � �,�.������ ������� ����� � �������*�� ��� ���� �������� �� � ���� �� ���#����� )�� ����� ����,��� ��� ������ &�� �� �##�������� � =1�$$$ #�� ����� )��� �� � ��� �� & ��� ������� �� ����#���� ����� � ��������� ���� ������� �#���� �� � �� � � &�� �� ��������� =@�$$$ ����� �� 0& ��� ���� �& ��� 6 �������� �� ��� ����� ��� ���� ����,��� �#���� ��� ���� c��� ��� �������*�� ��� ����� )�� ���� ��#���� � �#���� ��� ���� �� ����,��� �� � � �� ���# � �������� ��� ���� �&& ���� �������� ���#����� )�� ������ )� ����� > �� ��� ��� �� ��� �-����� �������� +��� ������� ����� �������� ����������� ����� &����� ������������ ������ #�������� &������ ��������� ��� &������ ��� �#������ )�� )� ����� ������ #�� ���� � �� ; ����� �& ��������� &�������� ����� )�� ��&�,��� � �� �&&��� � ���� ����� �& &��������� ��� ����� ��#���������� ��� ������ �������*�� ,����#���� � ������*�� ����� ���� ��� �������� ������� � ����� ��� ����� >���� ���� �� � �� � � �� � �������� ���� �������� ��� � �� !$ ������ ��� ��� .���� ���� ��� � �������� ,��� #��������� � ��� ���� ���� ���������� )�� �������� #������ �& ��� ��&�,���� ,���� �� �� �� 33�������44 ��� # �� ������� ��� ��������� #����� � �� � ������ �& ���� #������� ��� ����� ��������� 0� � �� &������� ���� �������� ��� �� )�� (�,� /���� &������ #�� ���� �� ����� ��,� ������ ��� + ������ �& ������� &��� � �� �1$ ��,� #�� ����� ��� ����� �� ��� ,� .���,� ��,� ��� ����� ���� �� g�, ������ �������� ��� ������ (�,�� ������ )� ����� > �� ��� � �������� ������ ������� ��� 0������� �� � ������ ��� ��� ���� )�� #���� #���� �& ��� 0������� ��� ��� ������ �� =b$ #�� ������ #�� ����� )��� #���� #���� 0������� ��� ��� ���� ��� ������ � ��� ���� �� ��� )� ����� > ��� )� �� ��� ��� ���� �#����� �� ��� g����� "� ���� �� h��� 7g"h8 0������� #������ ���� ��� )� ����� > �� ������� )�� g"h )� ����� ������ �� =;$$ #�� ����� #�� ����� ��,� ��� ��� g"h ��� ��� ��� �� � ��##��� ����� ����� ���� ����� �� ���� ������� �� �� ��� ��� �� �& ����� �� ��� ������� ���� ���,�� ��� ��������� ��� ��� ��� ������ �� �������� � � ������� �� )�� ������ )� ����� ��������� ��� ��� ��� �� #����� ��� ��������� � �� ��� ��������� ,��� � ���� #������ �& ��� ������ �� �� �� ��� ������� ��� ��� �� ������� ���� ���� ������ )�� ������� ���� �� =!�;$$ #�� ����� &�� �# �� ;$ ������ � ����� ��� � �� ��� ������ ���� ������� �� ������ �� ����� ��� ��� ��-����� ������ )� ������ /� ��� ������ ,��� ��#��� ���� ,� �� � �� ����� #������� ��� #������� ������� � �� #�� ���� � � � � +������ ��&�������� ��� ���� �� ,��� � ��� ��� ��� ������� ������� ��� � �� �& ��� � ��� �� ��� �## ��� ������ �� ��� � �� ��� ����� ��� ������� > �� ��� ��� �� ��� ���� �� �� ���� �& ������ �� � ��� ���� )�� ������� > �� ����� ��� ��� �� ��� �-������� �������� ������������ �#������ ��� &������� ��������� ������ &��� ��� ��-����� �� ������ ����������� ����������� +��� ������� ��� ������������ �������� ����� )�� ������� > �� ������ #�� ���� &��������� ���� �������� ��� � �� ; ������ 6�� �� ��������� &��� �� �#����� ��� ��� ��&����� �� �� 6��������� ����&� #�� ���� ��� ���� ,��� � �� �$ ����� �& +������ ��������� ����� )�� 33,�*��� ��� ���44 � �,� ��� ���� �� ������ � �� ������ �& ��&&����� #������ �*�� #���� �� ������ ������� ��� ,����, ������� �� ��� ������ ��� �������� ,��� ��� �������&� �## �������� )�� �������� �& ��� �##��#����� #������ �*�� #���� ��� ���� ��� ������ ��� ���� &�� ��� ������� �� ���� ������� > �� #�� ���� ��������� #���� ���� �������� ��� !$ ������ ��� ��� .���� &�� � &��� )�� �������� ������� ���#����� ��&����� �� �� j ��������� � �,� ��� �� ���#� ��� � ��������� ����� ��� ��� ���� ���� �������� ��� �� 6�� �� ��������� &��� j �������� > �� #�� ���� !�! ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0���������� ����� ��� ��� ���� � ���� �& ����� �������� ��� �� ������� (�,� /����, ��� ���� � ���� �& ��,� ��� ����� ��� ����� ������� ��� g�, ������ ��� �� ��� ����������� ��#���������� ���,��� ���������� )�� ������� > �� ��� ��� �� � /� . ���� ������ ���� ���� �������� �-��#����� #�� ���� �� ����� ���� ������� ��-���������� ������� ��� /� . ���� ������� ��� ��� � �� ��� � ������ &���� �� ���� �� -���� ��#���� �� ��� #�� � � ��� �� �#���� &�� ����� ,��� ��� ���� �; ������ )�� ������� > �� ����� ���� ��� ��� ��� �� ����� ������ &�� �& =b�;a$ #�� ,����������� )���� ��� ��������� ��� ���� � �� � � ������� ������� ���� ��� � �������� ���� ������� )�� ������� ������ &�� �� ��� �� �������� ����� �� ��� ���� ���� ���� ��������� ,�� � #��� ���#����� ��� ���� ��� ���� �� ��� ����� ��� ���� ������ �� � � �� ,� +�� � �� �& � �� �#� 0� ����� ,����� � ������� ��� � ��� � ��� ��&�������� ���� ,�� � ��#��� � ���� �� �������� �-������ &��� ��� ��� �& ����� ���� ����� /� +�� ���� ��� ���� ��� ��� �& ��� � ��� ��� ��� ������ ��� ���� ���� ����� ���� �� +��� ������� ����������� ����������� ��� ������������ �������� ����� ,���� ��� �� � � �� � � &��� ������� &�� �� ��������� &��� )��� ,�� � ���� �� � ����� ���� ������� ,��� ����� ���� ���#�������� 6������� �& ���� ����� ��� � ���� &�� � �#���+� ��#� �& ����� ���� ��� ��i����� ��� ��������� 6�� ����# �� ��� � ��� ��� # ��&��� �� ��� ������ ,��� �� ����� �� ������������ ����� 6��� �� ,��� �� ����� �� �������� ��� ���� � �� � � �� ��� ���� ��� ����� �� �##���� ���� ��� � ��� ��� ��� ��� ����� �&&�� ��� ������ ����� �& ��� �� ��,� ��� ��������� ��� � ��� � #�������� �� ��� ������� ��� ��� &�� � &��� #���� � �� ��� ��� �#������ ������� ��� ��� ��� ������ ����� �& �������� ��� �� )���� ��� ��&&������� �� ��, ��� ������ ���� ���� � �, ��� ������� �� ������ � ����� 6�� ����# �� ��� ������ ������ ����� �� � ������� ��� ��� ,������ ��� ������� ������ �� ���� ��� ��� )�� ������� ������ ��� ��� ,�*��� ��� ���� ,���� �� � ���� &������� )�� � ��� ��� 6 �������� �������� ��&�������� ������� � ��� ���� ���� �� �#���+� �� ����� #������� ������ ���� ���� ���� ��� ��� �� �,� ���-�� ������� �� ���� ������ � � ,���� ��� � �����#����� �� ������ ��������� �� ,��������� ��#������ �� ��� ��#������� ��� #��&������ �& ��� ��� ����� 0& ��� ��� ���� ��� �� ��#������� ,������ ��� ��� ������� ���� ��� �� ������ ��� � #����� �� �� ��� ������� �� ������� ,��� ��� � ��� ��� 6 �������� ������ #�� ����� ��� �������� ��� ���� ��� �� ��� �������*�� ��� ����� f��# �� ����� � ����� ���� ���� �� � �� # ���� ��� ��.����� �������� � �� � �� �)�) �((��� ��� � 1! �� �������� � (���2���� �&��� �� ������ ���� ���� ��� ��� ��� #�� ������ ����&� � �������� ��, ��� ,� ��� ���� &��� ���� ��� ������ � ���������� ������ ���� ��� +�� ��� ���� �� ��� �� � ����� � �& ��� ��&�������� #�� ����� ���� ���� � �� � � �� �� ��# � ,���� � �� ��� 0�������� �� ��������� �� �#���� ��������� "��� ������� � �, &�� ���������� �� ���� ���#�� ���,���� )��� ��� ���� ���#��.,��� ��������� � �� � � �� ��� &��� ���� �� ,� �� ���� ��� ��� ��� � �� � � &�� ��� �� ������ �� #��#�� � �-��##�� � ��������� ��,� ��� ����� ��� ��&&����� ����� ��� ��#� � ����� ���������� ,��� ���� ��#� �& ����������� h��� �&��� ��� ��#��� ��#�� ����� #��� �& ��� #��c��� �� ���# ���� ��� �#�������� ������� ,� �������� ��� ���� �� ,���� ����&� # ������ ��� �� � ��� ������ ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0��!�b 0� �������� �� ��� ����� ���� ���� ��������� � � �� ��� ������� ���� ����� � �� ���� # ��� �� ��� ��� ������ �� ���#����� �� ��">� )���� �,� ���� ���� ������� � ��� ��������� ���� ���� �� ��� ���� �� ��� �& ��� ����� ����� ���� ����� )�� ��,����� �& ��� ���#����� �� ��"> ���� ���� �� ��� � �� ������ ���������� �� ���� ��� ���� �� ��� �#����� �� &��-���� �� ��� �� ��� � ���� ��&+�� � �� ������� ������ ��� �& ��� ���� ��� ������� ��� �� # ����� � �������� �� � � ,��� �#�� ������ � �& ����� "� � #��#���� �� ,���� � �� # ����� 33�� �����44 �� ��## ����� ��� ���� ��&+�� � ����� � #�� ���� ,��� ��� ���� ����� /� �� � ,������ �$:�; #��� ����� � &�� ���� ���� ���� )���� ��� ��� �������� �� ���,�� � ��� -��������� �� ���# � ��� �������� ������� �� ������ ��� ��� ���� ��&��������� )���� ����� � � �� ��� �� � ����� �� ��� ���� ��&+�� � ����� � #�� ���� � ��� ���� ����� �� �������� ����� ���� &��� ��� ,��� ��� ���� ����� ,� �� � ��� ���� ��� �� # ����� � � �� � � �� ��� ����� "� ,�� �� ������� ����� �� ������� &�� �� ��# � ���� ����� �& ��� �& ����� ��� #�� ��� ��� ���� � ��#��� ���������� ��� ��-����k ����� �� ���� ��� �& ��� #��#���� �& ��� ������� ���� �� �� �� � ��� �������� � ���#����� � �� ������ ������� ���, ���� �& ������ ��� ���� �� #������ �� ��� ,���# ���� 6��� ��������� ��� ���� ��� �� ������ �& ���#����� �� ��� ����� ,� &���� ���� ��&&����� ���#����� �� ��� ��&&����� ���� ����� 6��� ���� �����#����� �� �� �� #&� &�� ��� �������� �� ���, ��, �� ������ � ���� &��� ���� ���� ��� �& ����� �������� )��� &��� ������ ���� ���� �� � #���� ����� ����� ���� ���� ��� ��� ����#���� �� 5& ������ �� ��� �������� ����� �� ��� ���� ���� ���� �� � ���� � �� � �� ��� ����� ���� ������ ���� ���� ��� ��� ���� ���� � #��&������� �� �� &�� ����� �������� 5�� ,�� �� ���� � ���� �� �� ���# � �� � ��� ���� ���������� �� ,��� ����� ������� ����� ���� ,� �� � ������� ��� �������� ���� ���� � �� ��� �� ���� ����� ��� �� �� )�� ���� �� �� �� ����������� )�� �� ��� �� �� �� �� ���� �& ��� � ���� ��� ��� ,���� � ����� ��� ����� ��� ����� )��� ��&�������� ���� �� ������ ��� ��� ����� ��� ���� ,� � �#��� �)�) ���1����( ���(��! � �'��� �� ��� � ���� ���� ��� ������ +������ ���� ���� � 33������� ����44 ��� � ����,��� �& � ������ l������� ���� �& ����� ����� ��� ����� � ��� ��� ��� ������ ��,� ��� �� ����� ������� ����� # ���� 0� &���� ����� ������� ,� �� ��� ��� �������� ������ &�� ���� �& ����� ���� ����� "�&�,��� �� ���� ��� � ������� �� �������� �� � �� � �� 5" 6������� )������ "�&�,��� 76)"8 �&&��� � /� . ���� ������ ��� ��� � �������� &��� ��� 0������� ����� )�� /� . ���� ������ ��� ��� �� ������ �& � �,��� �������� �� ������ ��� ��&�,��� &��� ��� 0������� ����� �� ����� ��� �������� )��� � �������� ��� ���� �� �� � � �#���+� ���� �� ���,��� �� ������ ��� ��&�,���� �������� ��� ��&����������� ��-����� �� ��� ��� ��&�,���� 6�� ����� � ,��� ������� ����� �� # ���� ������� ���� ������ ��� � ���������� � ������� ��� ���� #������ )�� ����� ������ &�� �� ='�;$$ #�� ������������ )�� ��� �� �& �������� �� #��&������ ����� ��� ������ �� ��� � &������ )�� 6)" ��&�,��� ��� ����� ���� ���#������� )�� +��� ���#����� �� ��� ���� ���� ������<����� ����� ��� ���#������ )��� ���#����� � �,� ��� �������� �� ���������� ��, � ������ ,����� ��� �� # ��� �� ��� �� ������ ��� ��������� � ��� ��� ��#��� �� � ���� ���� ������� )�� ������ ���#����� � �,� ��� ������� �� ������ � �� ��.�������� #���&� �� �& ������� ����� ���<�� �#����� ����� ����� ������ #������ )��� ���#����� ���� !�a ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0������ � ���� ������ ���� �� ������� �������� �� ��� 0������� &�� ��� #���� ����� )�� ����� ���#����� �� ��� ����� ���� �� ,���� ��� ���� ������ �� ��#��� ���� �� �-���� ��� ��� � ������� )�� ��&�,��� �� ��� ���#������� �� ��� ����� ��� #�� � � ���� ���� ������ � �� � � ���� ��� ��#��� #��&����� ,� ���� � ��� ���� ���� ����� �� ��� ������� �� ��� ������� )� ���� ��� ���� ��� ��� ��� ��#��� +������ ������ �� ������� ����� ��� ��� ,��� �� #������� � �������� ���� ������ ��������� �������� ��� �� � ������� )���� ��� �� � ���� ��� ������� ��� ������ �� � �� ��#��� � ��� ��� ��������� �& ������ ,�� #����� � ��� ����� ),� &������� ����� ��� ,� �#���� ,� ��������� ,��� ��� ��� ��� h g #��� � ��� ��#���� �� ��� ��� ���� �� �� � ���� # ��� �� ����� ����� 6�������� � ��� #��� � ��� ���� �� � �,��� ��� �� ����� �� ���� �� ��� ��� �&&���� �� ���� ��� ��� ��� �& ,� �#���� ,����� �� ����� +���� /��� ��������� ��� ���� ������ ���# �� ���� ���� ��� ��� ,��� #������ � �� ��� ���� h g ���# ���� )��� �� ��� ����,��� ���� ,� ��� ���� &��� ��� �#��� &��� ��� ����� ������� �� ���� ���#��� ��������� �� �� � h g ����� ��� ��� ���� #�#� �� ��� ��� ,��� �� ����� �� �� ���� ��������� )�� )����.h�� g��� /� �� � #�#� �� ����# � �& � ��������� �� ���� �� ���# �� ����� )�� ����� �& ��� g��� /� ��� � ��*�� 6�� ����# �� � g��� /� ���� �� @a ���������� ��� � �! ���� ���� 7�##�������� � @ &���� b &���8 ��� ���� �� ���� �� =b$�$$$� ������ ��� � �� � � ���� � �, ���# �� �& ���� ��� ���.���� ��&��������� ���� �� � ���#� �& ��� g�, ����� 0�������� � ����� � �� � �� ����� #����� �& ����� )���� 33���#���� �����44 ��� ���� ��#���� � ���� ��������� �� ������ � )����.h�� >������ /� #�� ���� ���� ���� �& i��� � ���� ���� ��� g��� /� ��� >������ /� ��� ��&���� ,��� �� ��# � #����� ,��� ���� #��� ��i������ ��&&����� +������ ��&��������� 0& �#��� �� ������ ����� ��� �� ������ ���� ��� # ���� ���# ���� � ������ ��� �� ���� �� ��� ���������� h g ���# ���� ���� ��� � ���.�# �� ���# �� ��� ���� ��&��������� ��,� ��� h g ���# ��� ��� #��&��� � &�� ���� ��� ���#����� "�&�,��� �� � �� � � ,���� � �,� �� � �#���� �& /� . ���� ���#���� �## �������� ���� 33#���44 ��&�������� �� � ���� >�� �� � &����� ���� ����� �� ��� ���# �� �& � ��#��� )����.h�� g��� /� � )��� ��� � � ���&� &�� #�� ����� ��&�������� �� �&+��� �� ����� ����� ,���� �� � �#��� �� �����#���� ����� � �������� � ��������� � �������� ������� �� ��� ��� ���� ����� ���# ��� �,� ���� �& ���� ��� ����� ,��� ��� ��� �� �& ���������� �� ���� ��� ���� ���������� � ��� ����� �& ��� ����� )�� ��������� ������ ��� � #��������� ,��� ��&�������� &��� ��c�� ��������� ��� ������� �������� ��� �##���� ��� ���� ��� ��� ��������� ������� ��� ��� �� ��� ����� �& ��� ������� �� +������ �� � ����� ���,� �� 0������� ������� )�� ��������� ������ �� ���� ��� ��#���� � ���� ��� ���� ����� ,��� ����� ���� �# �� =�;�$$$ ����� �� �� � ��� i��� � �� ���� 5��� ���4 � ������� �� ���� ���# �� ��#� ��� ������ ��� ���� ������� � ��#����� ���#��� �� �� ��� ��&�,��� ���� ��� �� ��� ���# ���� )��� ��&�,��� 33����44 �� ��,��� ���� ��&�������� ���# �� ��� ���� ��&�������� #�� ���� ��� ����� ��� �� ���� ����� &������ ��� ��� ��������� �� �� ���� ���# �� #��� �� )�� ������ &�� ��� ���# ��� ��� ��&�� ��� �� ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0��!�; ��� ��&�,��� #�� ������ )�� ��&�,��� ��#��� � �� � �� � ����� � ����� &��� ,���� ����� ��#������ �� ��� &�������� ��� ��� ����� �� ��# ����� �� ���� ����� 5��� ��� �� � ��� �-��#���� ��� ��&�,��� �� ������ �& ��� �� � � ���� ���� �& # ��� ��� �-��#���� #��# � ��� ��&�������� ������ #��# �� ���� ��� � � � �� ��� � �������� �# ��� ,������ &�� ���� 5����,���� ��� ,� �� � �� ��� ��� ��������� ����� �& ������ ������� &��� ��� �������� ���� ������� ����� � �� �� � �� � ������ ���� ��� ������ ��� +������ ��,� �������� �� ��� � ����� �� ��� +������ ��,� �������� �� � ����� ���� ��#����������� #������������ ��� � � �� ��� � �����+���� ������ �& ��,� ��� ��&�������� &��� ����� ������ �� �� ����� �� � ������ ��� #������ �� ���� ����� ������� ����� �� �������� �� �� ��������� ������ �& ������� ������ � ����� 5 ���� �� �& ��� ,��� �� ������ ��� +������ ��,� �������� ��� ,� �� � �� ��� � �� � ����� ������� ���� ��� ����� 6������� ��#������ �� ��� ��*� �& ��� ����� ��� ��� � #��������� ��� �� �,� �� � ������ � ��� ,��� ������� �� ���� ����� m�� ��� ,��� �� �������� �� � ������ ���� �� � �� � �� ���������� &��������� ,���� � �, ����������� ��� ������ � �� ,���� ��� ��� ��� � ������ ��� ��� ������� � ��� � � ���� ��� � ���� ! "������# ����� �������� � �& ���� ��� �� �� ����,���� ��� ��� ������ � �� � �,� ������� ��� �� ���� ������ ��� ����� �� ��, ������������� 6�������� � ��� ����� �& ��������� ��� �#��� ��� ���� �� ��,��� ��� �,� �#������ � )������ 6 ��� �� �������� � � ���#���� � � ������ �� ���� ��� � � �#���� #��#���� �� ����� ������ #�����# �� ���� &�� ��� �## � �� ��� ������ 6�, �& �� ,� �� � ��� ������� �� ���4 � ����� ��� ��� )�� ���� ��,� �� ��� ��� �#��� � ���#������ � ������ ������� �� ��� ��� �#��� ��� ������ ����� ����� �& ���� ����� ,�� � �������� )�� +��� ����� ���4 ���� �� �� �� ������ �� ��, ��� ���� ,� � ����� /� �� �� � �� � � &�, ����� ��� � � �������� &��� ���� �� ,� �� �� � ���� ����� �� ��� ��� �� &�� ������ ��� �� ��-���� �������� � ���� ��������k 0& ���� ����� �� ����,��� ������� ��� ��� �� � �� ���� ��� ����� �� ��� ���� ��� ��� ,� ��� � � � �� ������� � 33�����.��44 � ���� �� ��� #��#���� ��� �#���� ��������� ,� ���� �� � ���� ��������� ������� ��� ������ � ��� ������������� i��� �� ������� � ������� #�,��� &��������� ,����������� ��� �� &����� 0& ��� ���� �� ����� �� � � ���,���� &��� ��� &�� ���� �� ��� �� ��#�������� ���# ���� �� ���� ,��� ������ ���������� �������� �� ���� ���� �� ��#������ �� ����� ���� ,� �� � �� ���� #��c������ �� #����������� ��� &�� ������������ l����� � ��� ������� ���� � � �������� ��,�� ,��� � ������� �& ,��� �� ��� ,� �� >����#� ���� �&&���� �� ��� ���� ���� �� ��� i�������� +������� m�� ��� ,��� �� �������� � �,������� ����������� � �,��� ��� ����� �� ��� &���� �& ��� ���� �� � �,������ ����#������ � �& ����� �� ��� ���� �& ��� ����� )��� �� #� ������ ���� ���� ���� �� �� ��� #��c������ ������ ��� ,����� ��� ��� ������ )��� ��� � �����# ����� ,��� ��#�� � �� ��&&����� ���� ��� � �� � � �� ��, +������ �� ��� � �����+���� �� �������� ������ 5���� �#����� ��� ��� ������������ �������� �� � ��� ������� �& ��� �,�� )�� #��c������ ������ ���� � � � ����� ,���� ���&���� ��� � ���. ������ 0& ���� � ,� � !�@ ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0������ &�� ��������� �� ,� � ���&� &�� ��� ������� ���� �� �� � � ��� ��� ������� ����<���� #��c����� ������ �� � ,���������� �� ��� ����������4� ������� ����������� �#��� �� ������� � ��� �� � �� #����� � ��� �-��#����� �� �� �������� ��� #��&������� �����#���� ������� �� � ������� ����� 0� ��� ���� �& ���� ����� �������� �#��� ��� ,� ,��� �� ����� ��� ����� ������� � ��� &�� ��� ����� ,��� �������� "�*� ��� ���� ������ ���� ������� ��� �-��#����� ��� �� �& #��# � ������#����� ��� ��� ������� ,����� ��� �� ����� 0� ����� �� ������ ����� ��� #�� ��� � #��&������� �##������� ��� ��� ,��� �� ��� ���#�� �� ���� ������� ������ ������� �& ������� i��� �� ������ ���#�� #������� � ���# � �& ������� 0� ��������� ������� ��� ��������#��� ���� � ���� ��� ��� �� �� � i��� �� ����� �� �� ����� ����� �� � ��� ���� � �� � i���� 0& ��� ������ �� ��� ���#��� ��� ��� ,��� �� ����� �� ����.������ ���#��� �� � ������ �� � �& ���� ��� ��� �� �� �� � ��� ���� #�,��� �� ��� ������ �� ,������������ ���# �� #��� �� ����<���� #��c������ ��� ����� �-��#����� �� � ���� #��c��� ������� ����� ��� ���� #�������� ��� ��������� � ���+�������� ���� �����+�� ���� ����� ��� ����� � ������� ������ 0& ��� ,� � ������� ��� ��� �� ���� ����� �#���&� � ��������� ������� &�� ����� ��� �� ��� ���� ������ �� � ���� ���� �� ,� � ����� ���� ��� #������ �� � �������� #���� ���� � �������� � #�,�� + ��� �� ����� #�������� ������ �� ��� ��� ��� #��� �� #������ ��� ����� ���������� ,� ��� ��� ������� &�� ��� ���� ����������� &�� ��� ����� )�� ���� ����� ���� �� ��������� #�,�� ��-��������� ,� � ��&+����� &�� ����������� ���� ���#� �� ,� � 6��� �� ��� ,� ���� �� #������� � #������ &�� ��� ����� g�#������ �� ��� ��� �� �& ������� � ��� �� � �� ��� ����� ��� ��� ,��� �� ���� ���� � ���� � ��� #������� �� �������� ��#��� � ��� �� #���� ��� � �� �& ��&�������� ,������ ���� ���� �� �� ���� $ ��� �� ����� �� ����� %���# ��� ���� � ������� ���� ��� � �� ��#���� � #��#������� &�� � ����� � �� �& ��� �#��� �� � �� � �� � ���� ,��� �� &�, �� ��� ����� ��� ���� � ���� �# �� =�;$�$$$� ��� ��� ����� ��. ��������� ����� ������ &���� )�� ��� ��� &�� ������� &���� �� �� � � ������� ���� �� �� ��#��� ��� ������� ���� �� �� �������� &��� ���� "������ ���� ��� � ��� ��� ����� �������������� �� ��������� ��� ������� �� &��� ��� ���#����� ���� �� � � � ��� �� ��������# ,��� ���� ����� � �� ����� � ���� �� &��������� �������������� ��� � � &���� � ������� ������ &�� &������� 0& � ���� �� ��� #���� �� ��� ���� � ��� ��� � ���� #������� �& �������� &��� ������ � ��� ,��� &��� �� ��#��� � ���� �� ��� ��� � ���� �� ����� �� �� � �,� � ���� &�� &�� ��� ��� ������ �& +������ ���� ���� ��� � ��������� ��� ��� ��#��� �& ��� #������������ m�� ��� ��� � �� �& ��������� ,��� ��� ������ � ���� ��� ��� ��� ���, � � 5 ��� ,������� � ����� ��#��� �& ����� ���#���� �� ��� ���#�������� ��� �� ��� ���#����� � �������#��� �� �������� )���� ��� -���� � &�, ����� ������ ��� �������� ���� � ������� � � �� �##�������� ��� � ����� ��� �� /� �� � ������������ �� ��� ����� �� ��� ���� � #������� ������� ��������������� ��������� ��� ���#������ ��� � �� ������ &��� ����� ���#����� 0� �� ��� ��#������� ���� ���� ��� �� &��� ��� ,��� ��� #������� �& ���� � ����� ��� &������ ������ ���� ������� �� ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0��!�' & '�������� ���� �( ��� ��� ��) ���% � ������� ���� #�� ���� �� ����� �������� &�� � ��� #��c����� ���� ���� ��������� �����.�� ��������� ��� ������� ���� ������� �� ��� � �� � ��� #��c����� �� ��� ���#����� �������� ��� &��������� � ������ �& #� �� � ������ ���#����� ��� �� ���� ���� ���� ����� � "������� ��� ��� +������ ��������� ����� ��� ������� ��� �����&�� ��� ���� ���� � ���#���� ����� ,��� � &�, ���# � ����������� )��� ��� ���� � ������ ����� ���� �� ��� ,������� � ����� ���� �& ��#��� �& � ���#��� �� � ���#��� ���� )��� �� ������ � �,� ���� �� �#��� ���� ���� ���� �� ��� ���� ��� ���#������� ������ +��� ��� ����������� ��� ���� ��� ���� �� ���� �� ������� 0� ��� �� ������� �������� ��� ���� ��� �� � �� #� �� � ������� �� � �� �� ����� �� ������ #��c���� �� ���� ������� )�� �������� ��� � ���� ��� ���� &��������� �� �������� ���������� ���� � ����� ��� � � ����� )��� ��� � �� ���.�# ������ ������� �� ���� ��� �� ��� ���������� �� ����� &���� #���� �� � ���� �� ��#�� ��� ����� 0� ��� ������������ �������� �������� ������ � ��� ������ �& ������������ �������� �� �������� ����� /��� ���#��� �� ���� ���� ��������� ����� �� � �� �& ���� �� ����� ���� ��� ��� ��� #�� ���� ���� �� ��� � �� � � �� ���#����� �� ��">� ���� �& ��� ���� ���� ,�� � ���� �� � �� ����� &��� #���� ����������� " � + ����� �� #�� ��� �� ����#���� ��� ����� �� � �� � � ������� ����� ���� ��� #�� ������ �� � &����� ���� �� ���������*�� ������ ���#������ �� ��� ��� �� �& ����� �� ��� ������� ���� ��#����� �� ��� �� &��� � #��c��� �� �������� �� ���� ��� � �� ����� � �� ���� � ����� �� ������� ���� ������ �� ��� ����� )�� ������ ���� �� ������� ������� ���������� � ��#��� ��� ���� ��������� � ����� �� � �������� ���� � �������� �� ������� ���� �� ��� � �� ��� ��#��� �� ���� �� ��# ������ 0������ �& �� ��� ��� �������� �� �� ��� �� ���������� ��� ��� �� �� ,��� ����� )��� ��� � �������� ��� ����������� ��� ������� #������� )������ ���� ������ � �, ��� �������� �� ����� ���������� ��� �������� ���������� ,��� ��� ������ # ���� � ������ ��� ���� ����� ��&�,��� ��� ��, � �, ������� ���� ������ &��� ��&&����� �������� �� ���#��� ����� �� �� ����� ��������� �� ��� ���� ����� ,��� � ��� ����� ��� �� ��� ��� �� ��� ���� ����� * +��������� /� ��� #���� �� ��� ���� ,� ,��� ��� �& ��� +��� ������� ����� ���� ����� �� ��� ������ �& ��� �������� ��� �� ��� ��� � ��� #����� � ��#�������� )�� ���+�� ,� ��� ���� �� � � ���� �� &�������� )�� �� ��� � � ���� ������ �� +������ ����� ��� ��� ���� ��� �& ��� +������ �������� ��� ��#�� �� ��� -�� ��� �& ��� ����������� ��� ��� � � �& �������� 6������� ��� ���#������� � ������ �& ��� ����� ��� �� �� ��� &��� �� ���� &���� ����� � ��� ��������� 6��� �� ��� ��������� �& ��� ���� ��� �� �� ��� ����� � ������� ������ �& ��#�������� ��#��� � )�� &��� ��� &��� ��� ������������ ��� ��� ������ 6����� ��� �������� &�� ��� ���, ���� �& ��� ���� ���� � �� � � �����+���� � �������� ��� � ��� ������� ��� ����� ������� ��#�������� ��� �� �� ���� � ��� ���#����� � �� ������ �� ��� ������ # ���� )��� �� � � ����� &�� &�� ��#��� �������� ��� ���� �� � ���, ���� �& �#���+� ��� � ���� �� ��������� "������ ��� ��# ����� �& ��� �������� &�� ��� ���� ,��� ����� ���� ��#��� � ������ ��� �� !�1 ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0������ �& ��� ���� ���� �� ����� �#��������� 6��� �� ��� � ���� ��� -���� #���� �& ��� ����� ��� ���� �� �� � ���� c� �& �� �������� �� &�� ��� 0� &���� �� �� ��� ������ �� �� � � ���� ��� ��# ������ ������ ��� �& �������� �����# ���� ��-���� �� ���� ��� ���� ,�� � ���� ��� ���������� ������� �� ���#��� 0� �������� ���� ,�� �� ��#���� � ������� ��� � �� ,� ,���� ��� ����� �, � � ���� �( ��������� ��� ���� ����� � � � ��� ��� 6������� ������� a99 >��� � ���� (�, m���� (m �$$!!� j"� )� �? ��.!�!.19b.b@�; ���#?<<,,,�� ��� ������� ��� ����� � ����� j�� ������ >������ ����������� ������ 0�&�������� "������ '�' 5&+�� >���,�� "���� h����� �5 @b�a�.'��;� j"� )� �? ��.1$$.b!;.b!1!� ���� 1ba19 ���#?<<�� ��������� 5" 6������� )������ "����� >�5� ��� ��b;@ >���� ����� >� �;!b1� j"� )� �? ��.1$$.9@'.919' +�d+�,� ����� ���#?<<,,,�+�,� ���� ��� ���� l��� 000 �������� ���������� ������ ������� ������� 0��� b�� "���� /����� g�� �� "���� ��$$ �������� 0h @$@$@.@@1@� j"� 71$$8 ������� 75#���� @8 ���#?<<,,,��������# ������ ��� l����� "� h�&��,��� ��������� 2��� >��������� ���#����� "� �� )����.h�� ���#������� ��$ �������� � ���� (��,� �� �)� j"� )� �? ��.1$$.!ab.;;aa ���#?<<,,,������. ������ ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0��!�9 http://www.bloomberg.com http://telerate.com http://www.�sweb.com http://www.reutersplus.com http://www.trans-lux.com -�(������� ��� ��*�,���� g�� >������ )�� n e��� �,� �� 7�99b8� 6������ ��-��������� ��� ���#���� ��� �*����� �� ���"�. ���������� ����� �� ����� ��� 3�� �� � ��( #(1 ��� �� �7!8� �bb:�b9� � ��� � � �� ��� n � ��� � � "� f� 7�99;8� ���#���� ��� �*����� �� +����� �������� ����� ��� 3�� �� � ��( #(1 ��� �� 47�8� �b!:�a!� ���� �� ��� n ",������ �� l� 7�99;8� 0��� ����� �� �������� �� ������� ��� ����? 0���������� ��&�������� ������� ��� �������� � ������� ����� ��� 3�� �� � ��( #(1 ��� �� 47!8� 9$:91� >������ � 6� 7�99@8� j���� ������ ��&�������� ������� �� ��� � �������� " 1���� ����� ��� #(1 ��� �� ��� '9:1b� !!$ ")�) ���*��(�� "�) �� ��) + ����� ��� ����� �� ,����' �.�--�/ �-�0�� a guide to building a university trading room introduction benefits databases service providers additional thoughts relative to databases simulated trading software hardware that provides the "glitz and the glitter" facility considerations how and where to raise money potential uses of the trading room conclusion address of suppliers references financial services review, 32(4) 13 should investors defer long-term gains in taxable stock portfolios? jeff whitworth1 abstract investors with taxable portfolios sometimes delay the sale of appreciated stock to defer capital gains taxes. while this strategy does help to reduce taxes, it can cause the portfolio to become more concentrated over time, leading to higher overall volatility and lower long-term returns. this paper evaluates the tradeoff between tax efficiency and diversification via monte carlo simulation and finds that diversification is far more important for the investor’s terminal wealth, especially over longer time horizons. under a reasonable set of assumptions, investors are better off rebalancing almost completely each year, even though it requires selling some recent winners and paying capital gains taxes. while tax efficiency does become somewhat more important for individuals who are taxed more heavily on gains or who expect an eventual step-up in basis, even these investors should tolerate only a modest increase in portfolio concentration. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation whitworth, j. (2024). should investors defer long-term gains in taxable stock portfolios? financial services review, 32(4), 13-26. introduction how should taxable investors deal with appreciated stock in their portfolios? while owning stock that has increased in value is obviously a good “problem” to have, it does present investors with a dilemma. selling would trigger a capital gains tax liability, so it may be advantageous to defer the realization of gains as long as possible. however, a policy of never selling appreciated shares inevitably leads to a loss of diversification over time, as the portfolio becomes increasingly dominated by a few big winners. the resulting increase in risk – while undesirable in itself – also reduces the portfolio’s long-term expected growth rate through “volatility drag.” therefore, it is not immediately 1 corresponding author (whitworthj@uhcl.edu). university of houston-clear lake, houston, tx, usa clear whether an investor should sell appreciated shares or continue to hold them. in an early study on tax-efficient investing, constantinides (1983) argues that it is best to realize losses as soon as they occur and to defer gains indefinitely (or until exogenous factors force the investor to liquidate shares). however, in a follow-up study, constantinides (1984) demonstrates that when short-term gains are taxed as ordinary income but long-term gains are taxed at a lower rate, it often makes sense to realize some long-term gains so that future losses can be realized short-term and deducted against ordinary income. smith and smith (2008) propose a strategy of realizing all losses but also realizing enough gains to offset any losses in https://creativecommons.org/licenses/by-nc/4.0/ mailto:whitworthj@uhcl.edu https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 32(4) 14 excess of $3,000 (the maximum that can be deducted against ordinary income in any year). they show via monte carlo simulation that this method generates higher terminal wealth than any of the strategies previously proposed by constantinides. in a more recent simulation study of tax-efficient investing, whitworth (2018) considers the effects of loss recognition, smith and smith’s (2008) limited gain recognition, and the use of appreciated stock for charitable donations. while all of these strategies clearly increase the investor’s wealth over time, a more detailed breakdown of his results shows that most of the increase is attributable not to capital gains tax savings, but to improved portfolio diversification. this raises the question: could diversification be so important that investors should actually sacrifice tax efficiency for a more balanced portfolio? if so, might investors be better off completely rebalancing their portfolios each year, even though it would require selling some appreciated shares and incurring capital gains taxes? or could there be an optimal balance between tax efficiency and diversification that calls for less-than-complete annual rebalancing? these questions – which are not addressed in whitworth (2018) – are the central focus of this paper. it is well known that diversification reduces portfolio risk without reducing expected return – a fact sometimes referred to as the “only free lunch in investing.” since the long-term geometric growth rate on an asset is approximated by g =  − 0.52, where  is the mean annual return and  is the annualized return volatility, the reduction in portfolio risk that accompanies diversification should result in improved long-term growth. indeed, several studies have noted the existence of a “diversification return” (cuthbertson et al., 2016; booth and fama, 1992; willenbrock, 2011; erb and harvey, 2006; bouchey et al., 2012), and because of this, feld (1999) suggests that many investors might be better off selling appreciated 2 these market return assumptions are the same as those used in whitworth (2018) and are generally consistent with the 2024 long-term capital market assumptions of j.p. morgan, which are available at shares and taking the capital gains tax hit rather than dealing with the higher volatility inherent in an overly concentrated portfolio. in a simplified analysis of an investor who is fully concentrated in a single appreciated stock, stein et al. (2000) find that “near-complete diversification” into a lower-volatility portfolio is preferable, “despite a high initial tax cost.” this paper considers the case of a taxable investor managing a multi-stock portfolio and simulates the likely long-term outcomes from different strategies. we assume that all losses are realized each year but also test several criteria for deciding when to sell winning stocks. consistent with the studies cited above, our results indicate that the investor’s after-tax terminal wealth is maximized by keeping the portfolio almost fully diversified via annual rebalancing, even though this method results in more capital gains taxes being paid. even an investor with a high capital gains tax rate or one who expects a step-up in basis (which would increase the incentive to defer gains) should be willing to tolerate only a modest increase in portfolio concentration. the next section describes the monte carlo simulation methodology used in this paper. the following section presents and analyzes the results of the study, and the final section concludes. simulation methodology this study builds on the simulation methodology of whitworth (2018), which itself is based on smith and smith (2008). we assume that an investor has an initial portfolio of $250,000, consisting of $10,000 invested in each of 25 nondividend-paying stocks. for each year t, a simulated market return rmt is generated from a normal distribution with a mean of 8% and a standard deviation of 20%.2 a set of idiosyncratic disturbances εit (i = 1, 2, … 25) is also generated via independent draws from a normal distribution with mean zero and (for reasons that will become clear shortly) a standard deviation of 34.641%. the set of realized stock returns rit (i = 1, 2, … https://am.jpmorgan.com/us/en/assetmanagement/institutional/insights/portfolioinsights/ltcma/. https://am.jpmorgan.com/us/en/asset-management/institutional/insights/portfolio-insights/ltcma/ https://am.jpmorgan.com/us/en/asset-management/institutional/insights/portfolio-insights/ltcma/ https://am.jpmorgan.com/us/en/asset-management/institutional/insights/portfolio-insights/ltcma/ whitworth 15 25) in year t is then determined by adding the market return rmt to each of the 25 disturbances εit. given the parameters above, the simulated annual stock returns rit will be distributed normally with a mean of 8% and a standard deviation of 𝜎𝑖 = √𝜎𝑚 2 + 𝜎𝜀 2 = √0.22 + 0.346412 = 0.4 = 40%. 3 the contemporaneous correlation between the returns of any two different stocks i and j in the same year will be 𝜌𝑖𝑗 = 𝜎𝑖𝑗 𝜎𝑖𝜎𝑗 = 𝜎𝑚 2 𝜎𝑖𝜎𝑗 = 0.22 0.4 ∙ 0.4 = 0.25. these values for σi and ρij are the same as those used by smith and smith (2008) and whitworth (2018). returns across different years are generated independently and thus are uncorrelated.4 ordinary income is taxed at 24%, while long-term capital gains are taxed at 15%5. given these assumptions, the investor does not try to time the market. of course, one could adopt a buy-and-hold approach where no stocks are sold until the end of the investment horizon, but we consider several alternative strategies (described below) where trading decisions are made based on taxes and portfolio diversification. at the end of each year, any lot of stock whose value has declined below its original basis is sold to realize capital losses. any capital gains that have accrued are then handled according to one of three policies. the first two policies are considered in smith and smith (2008) and in whitworth (2018); however, the third policy has 3 as noted in whitworth (2018), this value for σi is reasonable given the individual stock volatilities that have been observed historically [e.g., see statman (1987) and campbell et al. (2001)]. to the extent that investors choose lower-volatility stocks, less rebalancing would be needed to maintain diversification, and tax efficiency would likely become more of a consideration. (of course, the reverse is true if investors choose higher-volatility stocks.) 4 the assumption of serially independent returns is largely consistent with the market efficiency literature summarized by fama (1970, 1991) and malkiel (2003). to the extent that stock returns exhibit not been considered in previous simulation studies. the three strategies for dealing with accrued gains that we examine are as follows: 1) defer all gains: appreciated shares are held as long as possible (i.e., until the end of the investment horizon) so long as their value never dips below the original basis. 2) realize gains to offset excess losses: under this policy, the investor follows smith and smith’s (2008) strategy of realizing only enough gains to offset any losses exceeding $3,000 (the maximum which may be deducted against ordinary income in any year). because an important objective of this strategy is to facilitate rebalancing, shares are sold first from the portfolio’s most heavilyweighted stock until its remaining value equals that of the second heaviest-weighted. then shares are sold from the first two stocks until each of their values equal that of the third heaviest-weighted. this continues until the desired amount of capital gains have been realized, or until there are no more unrealized gains left in the portfolio. 3) rebalance the entire portfolio: under this policy, if any stock comprises more than 4% (i.e., 1/25) of the portfolio, enough shares are sold to bring that stock’s weight down to 4%. when selling shares under policy (2) or (3) to reduce a stock’s weight in the portfolio, it is possible that the investor might own multiple lots of that company’s stock. if so, shares are sold first from the lot with the highest basis-to-value ratio. under policies (1) and (2) – and possibly under policy (3) – the investor will realize a net capital momentum, constantinides’ (1983) strategy of recognizing losses and continuing to hold winners would increase in effectiveness. to the extent that returns are mean-reverting, regular portfolio rebalancing (which requires selling recent winners and buying additional shares of recent losers) would become more desirable. 5 according to the most recently available irs statistics of income (for tax year 2021), 67% of returns reporting long-term gains were subject to the 15% tax rate. as of 2024, this rate applies to single filers with taxable income between $47,026 and $518,900 and to married couples filing jointly with income between $94,051 and $583,750. financial services review, 32(4) 16 loss6, up to $3,000 of which will be deducted against ordinary income (resulting in a tax savings of 24% of the deduction). any unused loss deduction is carried forward to the next year. however, policy (3) will often realize more capital gains than losses, in which case the net gain is taxed at the long-term rate of 15%. (this assumes that the appreciated shares were held for at least a year plus one day to qualify for the longterm capital gains rate.) the proceeds from end-of-year stock sales (plus the tax savings on net losses, or minus taxes paid on net gains) are then reallocated to the leastweighted stocks in the portfolio. funds are invested first in the least-weighted stock until its weight equals the second least-weighted; then funds are invested in the two least-weighted stocks until their respective weights match the third; and so on, until all cash has been reinvested. this rebalancing procedure ensures that the portfolio is as diversified as possible.7 this process continues each year until the end of the investment horizon (assumed to be 10, 20, 30, 40, or 50 years), at which point the entire portfolio is liquidated and taxes are paid on the net long-term gain. for each iteration of the simulation, we concurrently track the value of a buy-and-hold portfolio which begins with $250,000 invested equally across the same 25 stocks, but in which no stocks are ever sold until the final liquidation date. from this, we calculate the percentage by which the investor’s after-tax terminal wealth exceeds that of the buy-and-hold portfolio. some of this wealth increase may be due to improved tax efficiency, while some is attributable to the “diversification return” previously noted. to determine how much is due to better diversification, we also track the value of a portfolio which – like the buy-and-hold portfolio above – starts with the same $250,000 and never realizes any gains or losses until the terminal date. in this portfolio, however, funds are costlessly 6 in the unlikely event that none of the 25 stocks in the portfolio experienced a loss, no gains or losses would be realized that year under policies (1) and (2). 7 implementing this strategy usually will require repurchasing shares that have just been sold to recognize capital losses. in practice, this would run redistributed each year (without any immediate tax consequence) so that the individual stock weights match those in the investor’s portfolio. therefore, this portfolio receives none of the tax benefits that the investor does via tax-efficient trading, but it maintains the same level of diversification. by also comparing its performance versus buy-and-hold, we can ascertain how much of the investor’s wealth gain is due to diversification and how much is from tax savings. for each of the three gain realization policies above and for each possible time horizon (ranging from 10 to 50 years), one million iterations of the simulation are run to obtain parameter estimates. results effect of the three gain realization strategies on terminal wealth and risk table 1 shows the mean and median increases in after-tax terminal wealth (relative to a buy-andhold approach) from implementing the three investment strategies described in the previous section. in addition, the table shows how much of the increase is attributable to tax savings and how much is due to improved diversification. it is immediately clear that all three strategies outperform buy-and-hold and that the difference is greater over longer time horizons. a comparison of panels a and b shows that smith and smith’s (2008) strategy of realizing enough gains to offset excess losses does much better than a policy of simply deferring all gains. the middle column of the table shows that very little of the wealth increase is due to additional tax savings. this is not surprising since recognizing gains is generally not advantageous from a tax standpoint, except that effectively resetting the cost basis on part of the portfolio may create a few more opportunities to harvest future losses. however, as seen in the rightmost column of the table, there is a large diversification return. this afoul of wash sale rules. however, an investor can circumvent these restrictions by waiting 31 days to repurchase the shares, or by immediately purchasing shares of a different stock with similar risk and return characteristics. whitworth 17 is also as expected since the main benefit of smith and smith’s approach is that it routinely rebalances the portfolio (albeit to a limited extent) by harvesting gains from the most heavily-weighted stocks and reinvesting the funds into the least-weighted stocks. this in turn reduces the volatility of the portfolio and the consequent “volatility drag” on returns. table 1. mean (median) increase in after-tax terminal wealth from alternative investment strategies vs. buy-and-hold years invested total increase from tax savings from diversification panel a: realize losses and defer all gains 10 1% ( 1%) 1% ( 1%) 0% ( 0%) 20 4% ( 2%) 2% ( 2%) 0% ( 0%) 30 8% ( 3%) 4% ( 3%) 4% ( 0%) 40 13% ( 4%) 4% ( 3%) 9% ( 1%) 50 22% ( 6%) 7% ( 4%) 15% ( 2%) panel b: realize gains to offset excess losses 10 3% ( 2%) 1% ( 1%) 2% ( 1%) 20 11% ( 8%) 2% ( 2%) 9% ( 6%) 30 26% (17%) 4% ( 3%) 22% (14%) 40 48% (27%) 6% ( 4%) 42% (23%) 50 78% (39%) 9% ( 5%) 69% (34%) panel c: rebalance entire portfolio 10 3% ( 3%) 1% ( 1%) 2% ( 2%) 20 12% (10%) 0% ( 0%) 12% (10%) 30 32% (24%) 0% ( -1%) 32% (25%) 40 64% (44%) -3% ( -4%) 67% (48%) 50 116% (75%) -9% ( -9%) 125% (84%) each investment strategy is simulated 1,000,000 times over 10, 20, 30, 40, or 50 years. the simulated portfolio begins with $250,000 invested equally across 25 non-dividend-paying stocks. each year, we generate a set of 25 stock returns with mean 8%, standard deviation 40%, same-year inter-stock correlation 0.25, and no intertemporal correlation. capital losses are always realized annually. the three respective panels report results assuming that (a) no capital gains are ever realized until the end of the timeline, (b) enough gains are realized each year to offset any losses in excess of $3,000, or (c) the portfolio is completely rebalanced each year. net realized losses are deducted (up to a maximum of $3,000 per year) against ordinary income, which is taxed at 24%. net realized gains are taxed at 15%. after-tax proceeds from stock sales are reinvested in the least-weighted stocks. in each iteration the final after-tax wealth is compared to that of (1) a portfolio using a pure buy-and-hold approach, and (2) a control portfolio that remains untaxed until the end of the horizon but is costlessly rebalanced every year so that its individual stock weights match those in the simulated portfolio. the total increase column reports the mean (and median) percentage by which the simulated portfolio’s terminal value exceeds that of the buy-and-hold portfolio. the tax savings component is computed by comparing the simulated portfolio to the control portfolio that is costlessly rebalanced each year. the diversification component is the total increase minus the tax savings component. financial services review, 32(4) 18 since the added diversification of smith and smith’s strategy provides such large benefits, a natural question is whether an investor might want to diversify even further – perhaps even rebalancing the whole portfolio every year – even though doing so will require realizing and paying taxes on some gains. looking at panel c, the answer is a clear “yes.” even though (as expected) the tax impact of this strategy is worse than the others and actually affects wealth negatively at longer horizons, the additional diversification gains more than make up for this. the net result is that fully rebalancing each year outperforms smith and smith’s more limited gain recognition strategy. table 2 shows the effect of the three strategies on risk, as measured by the standard deviation of terminal wealth in our simulation. as seen in the first column of the table, realizing only losses has almost no effect on portfolio risk, as the standard deviation of terminal wealth is very near what it would be under a buy-and-hold strategy. smith and smith’s (2008) strategy does reduce risk modestly. however, a policy of full portfolio rebalancing does much better, especially over longer horizons. over a 50-year time frame, the standard deviation of terminal wealth for someone using the more limited gain harvesting approach would be about 80% of a buy-and-hold investor’s risk; however, for an investor who fully rebalances each year, it would be only half of that (i.e., 40% of the buy-and-hold risk). the difference between full rebalancing and smith and smith’s more limited gain recognition is relatively modest over 10or 20-year horizons, but it becomes much larger over 40-50 years. table 2. ratio of standard deviation of terminal wealth from alternative investment strategies vs. buy-and-hold years invested realize losses, defer all gains realize offsetting gains rebalance entire portfolio 10 1.01 0.98 0.94 20 1.01 0.95 0.83 30 1.01 0.91 0.68 40 1.01 0.87 0.52 50 0.99 0.80 0.40 the simulation procedure is described in table 1. for each investment strategy and time horizon, this table reports the standard deviation of the 1,000,000 simulated portfolio terminal values divided by the standard deviation of the 1,000,000 corresponding buy-and-hold portfolio values. when comparing tables 1 and 2, it is clear that gain harvesting is most effective at the longest horizons, both at reducing risk and at improving mean and median terminal wealth. as table 1 shows, the diversification return (which is a direct consequence of the lower risk) becomes significantly greater over time. this is not surprising because these longer horizons are precisely where portfolios left to themselves could otherwise become extremely unbalanced. to see why these gain harvesting strategies are especially effective in the long run, it is informative to look directly at how they impact the portfolio’s level of diversification. table 3 shows how concentrated a portfolio may become in any one stock over different time frames. concentration is measured here by the percentage of the portfolio’s value that is invested in its heaviest-weighted stock. (this is a reasonable proxy for more sophisticated diversification measures – for example, a herfindahl-style measure equal to the sum of the squares of the individual stock weights.) bessembinder (2018) notes that “the compounding of random returns over multiple periods will typically impart positive skewness to longer horizon returns, even if the distribution of single-period returns is symmetric.” as a consequence of this return skewness, we would whitworth 19 expect a portfolio left to itself to become increasingly unbalanced over time, as the results in the first column of table 3 confirm. under a buy and hold strategy, just one of the 25 stocks will on average make up 20% of the portfolio (which is not insignificant) after only 10 years. after 50 years, a single stock is likely to become more than half (53%) of the portfolio. merely realizing losses helps a little bit, as the second column of the table shows slightly reduced maximum weights versus the first column. smith and smith’s (2008) strategy (corresponding to the third column) does even better, but it still allows for one stock to become about one-third of the portfolio over longer horizons. by design, a strategy of always rebalancing the whole portfolio keeps each of the 25 stock weights at exactly 4%, which is optimal from a risk perspective. as a comparison of the third and fourth columns shows, this is a little better diversified after 10 years than a portfolio managed according to smith and smith’s strategy, but after 40 or 50 years, it is drastically better. this explains why (as previously seen in tables 1-2), the difference between the two strategies with respect to terminal wealth and risk is relatively small for shorter horizons but much greater over longer horizons. table 3. mean portfolio weight of largest stock in final year years invested buy and hold realize losses, defer all gains realize offsetting gains rebalance entire portfolio 10 20% 19% 8% 4% 20 32% 30% 14% 4% 30 41% 38% 22% 4% 40 48% 43% 29% 4% 50 53% 48% 35% 4% the simulation procedure is described in table 1. results in the four respective columns are reported assuming that (a) no stocks are sold until the end of the horizon; (b) losses are realized annually but no gains are realized until the end of the horizon; (c) losses are realized annually, and enough gains are realized to offset any losses in excess of $3,000; or (d) losses are realized annually, and enough gains are realized to fully rebalance the portfolio annually. for each strategy and time horizon considered, this table reports the mean percentage of the portfolio that is invested in its most heavily-weighted stock at the end of the horizon. effects of less than complete diversification in the previous subsection, we saw that the best results (in terms of terminal wealth, risk, and diversification) were obtained by disregarding some of the tax-efficient strategies previously proposed in the literature and simply diversifying the portfolio fully every year instead. however, the wealth-maximizing policy8 still might lie somewhere between full diversification and the smith and smith (2008) strategy (which still allows for considerable portfolio concentration, as seen in table 3). 8 in the analysis that follows, we focus mostly on how different strategies affect median terminal wealth because medians are less influenced by outliers than means. although the graphs are not shown here, we we now consider the effects of modifying policy (3) described earlier by relaxing the full diversification criterion to varying extents. under this modification, the investor sets a maximum portfolio weight which no individual stock may exceed. this “weight limit” (which remains the same for the entire investment period) may be set as low as 4% if full annual rebalancing is desired, or at a higher percentage if the investor wishes to defer more capital gains while still not allowing the portfolio to get too far out of balance. at the end of each year, if any stock exceeds the weight limit, enough shares are sold to bring its weight also find that mean terminal wealth is affected similarly. financial services review, 32(4) 20 in the portfolio down to the limit. all losses are still recognized, just as in the original strategy. we repeat the simulation for weight limits from 4% to 20% (in increments of 1%). figure 1. median increase in after-tax terminal wealth vs. buy-and-hold for alternative portfolio weight limits each simulated portfolio begins with $250,000 invested equally across 25 non-dividend-paying stocks. each year, we generate a set of 25 stock returns with mean 8%, standard deviation 40%, contemporaneous correlation 0.25, and no intertemporal correlation. capital losses are always realized annually. a maximum weight is defined which no stock may exceed in the portfolio; if any stock’s weight is over the limit at the end of a year, enough shares are sold to bring its weight down to the limit. net capital losses (up to $3,000 per year) are deducted against ordinary income (which is taxed at 24%), and net capital gains are taxed at 15%. after-tax proceeds from stock sales are reinvested in the portfolio’s least-weighted stocks. the final after-tax wealth at the end of the horizon is compared to a buy-and-hold portfolio over the same period. the simulation is repeated 1,000,000 times for investment horizons from 10 to 50 years and for maximum stock weights from 4% to 20%. for each horizon and maximum weight, this graph shows the median percentage by which the simulated portfolio’s terminal value exceeds that of the buy-and-hold portfolio. figure 1 shows the median percentage increase in after-tax terminal wealth (again, relative to a buyand-hold strategy) for different portfolio weight limits that an investor may set. for 10and 20year horizons, a 5% weight limit is optimal, although there is less variation between the outcomes of the alternative strategies over these shorter time frames. for 30or 40-year horizons, a 6% weight limit does slightly better. over 50 years, median terminal wealth is maximized by setting a 7% weight limit. beyond that point, terminal wealth steadily declines as the portfolio 0% 10% 20% 30% 40% 50% 60% 70% 80% 4% 8% 12% 16% 20% m e d ia n in cr e as e in a ft e rta x te rm in al w ea lt h maximum portfolio weight allowed for any one stock whitworth 21 weight limit is raised, and the differences in wealth are especially notable over longer horizons. given the assumptions of our simulation, near-complete diversification is best. it should be noted here that “best” refers only to the policy that maximizes median terminal wealth. in practice, investors may be willing to sacrifice some wealth in exchange for lower portfolio volatility. although we do not attempt here to explicitly model the investor’s utility as a function of risk and expected return, we acknowledge that in many cases the portfolio “weight limit” that maximizes utility may be slightly lower than the one that maximizes expected wealth. for example, in the preceding example, instead of setting the weight limit at 7%, an investor with a 50-year horizon would almost certainly prefer to set it at 6% or even 5% because doing so will reduce portfolio risk while sacrificing only a negligible amount of median terminal wealth. (it is less clear whether they would lower it further to 4% since that would reduce median wealth more discernibly.) nevertheless, median terminal wealth is a key measure of the effectiveness of any investment strategy, so it is a reasonable starting point for the discussion. portfolio risk (as measured by the standard deviation of terminal wealth) is an increasing function of the weight limit since a higher weight limit leads to greater portfolio concentration and less diversification over time. however, it is not immediately clear whether the weight limit is always positively related to shortfall risk (i.e., the risk of earning less than a predetermined rate of return). figure 2 shows the probability of ending up with a negative overall return for different strategies. although the differences are relatively modest across alternative weight limits, shortfall risk is still a monotonically increasing function of portfolio concentration. for each horizon length, the risk is lowest with full diversification and steadily increases as higher portfolio weights are allowed. figure 2. risk of negative overall return for alternative portfolio weight limits the simulation procedure is described in figure 1. for each horizon and maximum stock weight, this graph shows the percentage of outcomes for which the simulated portfolio’s terminal value is less than the initial $250,000 investment. 0% 4% 8% 12% 16% 20% 4% 8% 12% 16% 20% ze ro -r e tu rn s h o rt fa ll r is k maximum portfolio weight allowed for any one stock financial services review, 32(4) 22 if capital gains are untaxed on the terminal date so far, we have assumed that all unrealized capital gains are finally taxed when the portfolio is liquidated at the end of the investment horizon. this makes sense if we interpret the terminal date as the point when the individual intends to consume the wealth they have accumulated. however, other interpretations are possible. for example, if the individual does not intend to consume the assets in the portfolio but instead plans to pass them on to heirs, the terminal date might be interpreted as the investor’s death, at which point the portfolio’s cost basis will be stepped up to current market value, erasing all accrued capital gains tax liability for the inheritors. if so, then it may make sense in some situations to continue holding appreciated stocks despite the increase in portfolio concentration. we would not expect the investor to completely ignore diversification concerns, of course, but the anticipation of a step-up in basis may increase the “weight limit” they would allow. in figure 3, we see that this is indeed true. when capital gains are untaxed on the terminal date, the wealth-maximizing weight limit shifts slightly to the right. for all horizons, median terminal wealth is maximized with a 7% or 8% weight limit (and in all cases, the wealth difference at 7% versus 8% is so slight as to be almost indistinguishable). in reality, most investors probably do not have a single “terminal date” marking the end of the investment horizon but will likely consume some of their wealth over time (resulting in some taxable redemptions) before eventually passing the remainder on to heirs (who will enjoy a stepped-up basis on that portion of the portfolio). it is beyond the scope of this study to model all of these possible complexities, but it is reasonable to believe that in practice the optimal “weight limit” may lie somewhere between the values derived from figures 1 and 3. it is also possible that one might use a dynamic policy that changes over time. for example, an individual who is nearing the end of life and expecting a steppedup basis relatively soon might almost always defer gains, whereas someone with many years to go might realize more gains to keep the portfolio better diversified. if capital gains are taxed at a higher rate it is important to note that the conclusions we have drawn so far depend on the particular parameters of this simulation. one important factor that may influence an investor’s willingness to rebalance the portfolio is the rate at which any realized gains would be taxed. under current law, long-term capital gains are generally taxed at substantially lower rates than ordinary income is. however, it is not guaranteed that these preferentially low rates will always exist in the future. from time to time, policymakers have proposed taxing long-term capital gains at the higher ordinary income tax rate. we expect that such a policy (or more generally, any increase in the capital gains tax rate) would increase the cost of portfolio rebalancing and therefore increase the degree of portfolio concentration that an investor would tolerate. this would be especially true if the investor is expecting much or all of the portfolio to be stepped up in basis at the end of the investment horizon. whitworth 23 figure 3. median increase in untaxed terminal wealth vs. buy-and-hold for alternative portfolio weight limits the simulation procedure is as described in figure 1, except that any unrealized capital gains at the end of the investment horizon are not taxed. after each iteration, the final wealth is compared to a buy-and-hold portfolio which is also untaxed at the end of the horizon. for each horizon and maximum weight, this graph shows the median percentage by which the simulated portfolio’s terminal value exceeds that of the buy-and-hold portfolio. to see this, we repeat the previous simulation but assume that capital gains and ordinary income are both taxed at 40%, which is close to the current top individual income tax rate. the resulting median increases in terminal wealth relative to buy-and-hold are shown in figure 4. as expected, the wealth-maximizing weight limit is greater when realized capital gains are taxed more heavily. for the 10or 20-year horizons, median wealth is maximized when the investor sets a weight limit of 9% or 11%, respectively (although at these horizons, the wealth differences at 9% versus 11% are virtually indistinguishable). for horizons of 30 years or longer, the wealth-maximizing weight limit is 12%. from this it is clear that an increase in the capital gains tax rate increases the degree of portfolio concentration an investor should be willing to tolerate since there is a higher tax liability created by the transactions associated with rebalancing. in fact, it is interesting to note that under this scenario, full annual rebalancing can sometimes be slightly worse than buy-andhold, as evidenced by the negative wealth increase when setting a 4% weight limit for some of the shorter time frames. this is consistent with the increased tax cost of rebalancing transactions and the large number of such transactions that would be required to keep the portfolio equally weighted over time. however, even with the 0% 10% 20% 30% 40% 50% 60% 70% 80% 4% 8% 12% 16% 20% m e d ia n in cr e as e in u n ta xe d t er m in al w ea lt h maximum portfolio weight allowed for any one stock 10-year horizon financial services review, 32(4) 24 higher tax cost, the portfolio still should not be allowed to get too far out of balance if the objective is to maximize terminal wealth. transaction costs unfortunately for investors, rebalancing one’s portfolio regularly also means incurring transaction costs regularly. as capital gains taxes do, brokerage commissions and bid-ask spreads add to the cost of rebalancing and should serve to increase the portfolio concentration that an investor is willing to tolerate. but are transaction costs large enough to make a substantial difference in the wealth-maximizing strategy? we investigate this question by repeating the previous simulation (where realized capital gains are taxed at 40%), except that every sale incurs a cost of 25 basis points (i.e., 0.25% of the value of the shares sold). transaction costs obviously can be higher or lower depending on the characteristics of the stock and/or the trading environment, but this is a reasonable approximation of the effective transaction costs actually observed in u.s. stock markets (e.g., hasbrouck, 2009; chen and velikov, 2023). in unreported results, we find that our simulated terminal wealth values are almost identical to those reported in figure 4. in other words, transaction costs in the major u.s. equity markets are low enough that they do not substantially influence the investor’s decision about whether to defer capital gains or recognize some gains to rebalance the portfolio. figure 4. median increase in untaxed terminal wealth vs. buy-and-hold for alternative portfolio weight limits (tax rate = 40%) the simulation procedure is as described in figure 3, except that ordinary income and any net realized capital gains are taxed at 40%. unrealized gains at the end of the investment horizon remained untaxed. for each horizon and maximum weight, this graph shows the median percentage by which the simulated portfolio’s terminal value exceeds that of the buy-and-hold portfolio. -10% 0% 10% 20% 30% 40% 50% 60% 4% 8% 12% 16% 20% m ed ia n in cr ea se in u n ta xe d t e rm in al w e al th maximum portfolio weight allowed for any one stock 50-year horizon 40-year horizon 30-year horizon 20-year horizon 10-year horizon whitworth 25 other considerations while it is beyond the scope of this study to examine every possible scenario, it is appropriate to acknowledge how the study’s design and some of its specific assumptions may impact the results. first, our conclusions are based on a set of simulated stock returns generated using the methodology and parameters previously described. while this is reasonable and consistent with prior literature (e.g., smith and smith, 2008; whitworth, 2018), other studies have employed a historical resampling approach when investigating similar questions, and it is possible that one might obtain different results by using a different methodology. regarding the specifics of our simulation, one simplifying assumption made is that the stocks do not pay dividends. all other things being equal, taxable individual investors generally should avoid dividend-paying stocks since dividends are tax-inefficient. receiving dividends essentially converts capital gains which could be taxed much later (or not at all, if the investor expects a stepup in basis) into current income which is taxed immediately. in addition, the entire dividend distribution is taxed as income, unlike shares which are liquidated for rebalancing purposes (where only the value above the cost basis is taxed). nevertheless, while some investors do own dividend-paying stocks, most dividends are relatively small and unlikely to have a significant impact on the investor’s terminal wealth. another simplifying assumption of the simulation is that the investor makes no new contributions to the portfolio on an annual basis. in cases where one does make regular additional contributions (e.g., from wage income), the newly invested funds can be allocated to the leastweighted stocks, thereby improving portfolio diversification and reducing (but not eliminating) the need to harvest gains for the purpose of rebalancing. it may be of interest for future studies to simulate the outcomes of alternative strategies when investments are made regularly over time, rather than as a lump sum at the beginning. conclusion this paper has considered alternative policies for when individuals with taxable portfolios should realize capital gains on their winning stocks. constantinides (1983) advocates holding winners as long as possible, a strategy which is taxefficient but can allow the portfolio to become very unbalanced over time. smith and smith (2008) propose a limited gain recognition strategy which does not hinder tax efficiency but does allow for greater portfolio diversification. we show that this leads to much higher terminal wealth, and that almost all of this is due to the “diversification return.” we then show that the investor can do better still by rebalancing the portfolio almost completely each year. even though this strategy requires paying some capital gains taxes earlier than if all gains were deferred, the increased diversification return overwhelms the relatively small reduction in tax efficiency. this, combined with the fact that lower risk is desirable in itself, suggests that taxable stock investors might be better off foregoing the usual advice to delay the sale of appreciated stocks as long as possible. individuals who are taxed more heavily on capital gains or who expect an eventual step-up in basis may wish to defer more gains than a typical investor would. however, even in these cases, only a modest increase in portfolio concentration should be considered since diversification has such a strong impact on long-term returns. references bessembinder, h. (2018). do stocks outperform treasury bills? journal of financial economics, 129(3), 440-457. https://doi.org/10.1016/j.jfineco.2018.06 .004 booth, d. g. and fama, e. f. (1992). diversification returns and asset contributions. financial analysts journal, 48(3), 26-32. https://doi.org/10.2469/faj.v48.n3.26 bouchey, p., nemtchinov, v., paulsen, a., and stein, d.m. (2012). volatility harvesting: why does diversifying and rebalancing create portfolio growth? journal of wealth management, 15(2), 26-35. https://doi.org/10.1016/j.jfineco.2018.06.004 https://doi.org/10.1016/j.jfineco.2018.06.004 https://doi.org/10.2469/faj.v48.n3.26 financial services review, 32(4) 26 campbell, j. y., lettau, m., malkiel, b. g., and xu, y. (2001). have individual stocks become more volatile? an empirical exploration of idiosyncratic risk. journal of finance, 56(1), 1-43. https://doi.org/10.1111/00221082.00318 chen, a. y. and velikov, m. (2023). zeroing in on the expected returns of anomalies. journal of financial and quantitative analysis, 58(3), 968-1004. https://doi.org/10.1017/s002210902200 0874 constantinides, g. m. (1983). capital market equilibrium with personal tax. econometrica, 51(3), 611-636. https://doi.org/10.2307/1912150 constantinides, g. m. (1984). optimal stock trading with personal taxes: implication for prices and the abnormal january return. journal of financial economics, 13(1), 65-89. https://doi.org/10.1016/0304405x(84)90032-1 cuthbertson, k., hayley, s., motson, n., and nitzsche, d. (2016). what does rebalancing really achieve? international journal of finance and economics, 21(3), 224-240. https://doi.org/10.1002/ijfe.1545 erb, c. b. and harvey, c. r. (2006). the strategic and tactical value of commodity futures. financial analysts journal, 62(2), 6997. https://doi.org/10.2469/faj.v62.n2.4084 fama, e. f. (1970). efficient capital markets: a review of theory and empirical work. journal of finance, 25(2), 383-417. https://doi.org/10.1111/j.15406261.1970.tb00518.x fama, e. f. (1991). efficient capital markets: ii. journal of finance, 46(5), 1575-1617. https://doi.org/10.1111/j.15406261.1991.tb04636.x feld, a. (1999). high exposure to low-basis stock: too much of a good thing? cpa journal, 69(11), 60-65. hasbrouck, j. (2009). trading costs and returns for u.s. equities: estimating effective costs from daily data. journal of finance, 64(3), 1445-1477. https://doi.org/10.1111/j.15406261.2009.01469.x malkiel, b. g. (2003). the efficient market hypothesis and its critics. journal of economic perspectives, 17(1), 59-82. https://doi.org/10.1257/0895330033211 64958 smith, m. h., and smith, g. (2008). harvesting capital gains and losses. financial services review, 17(4), 309-321. statman, m. (1987). how many stocks make a diversified portfolio? journal of financial and quantitative analysis, 22(3), 353-363. https://doi.org/10.2307/2330969 stein, d. m., siegel, a. f., narasimhan, p., and appeadu, c. e. (2000). diversification in the presence of taxes. journal of portfolio management, 27(1), 61-71. whitworth, j. (2018). improving long-term portfolio risk and return by using appreciated stocks for charitable donations. financial services review, 27(3), 257-277. willenbrock, s. (2011). diversification return, portfolio rebalancing, and the commodity return puzzle. financial analysts journal, 67(4), 42-49. https://doi.org/10.2469/faj.v67.n4.1 https://doi.org/10.1111/0022-1082.00318 https://doi.org/10.1111/0022-1082.00318 https://doi.org/10.1017/s0022109022000874 https://doi.org/10.1017/s0022109022000874 https://doi.org/10.2307/1912150 https://doi.org/10.1016/0304-405x(84)90032-1 https://doi.org/10.1016/0304-405x(84)90032-1 https://doi.org/10.1002/ijfe.1545 https://doi.org/10.2469/faj.v62.n2.4084 https://doi.org/10.1111/j.1540-6261.1970.tb00518.x https://doi.org/10.1111/j.1540-6261.1970.tb00518.x https://doi.org/10.1111/j.1540-6261.1991.tb04636.x https://doi.org/10.1111/j.1540-6261.1991.tb04636.x https://doi.org/10.1111/j.1540-6261.2009.01469.x https://doi.org/10.1111/j.1540-6261.2009.01469.x https://doi.org/10.1257/089533003321164958 https://doi.org/10.1257/089533003321164958 https://doi.org/10.2307/2330969 https://doi.org/10.2469/faj.v67.n4.1 164 exploring factors affecting young adults' financial management behavior: a hybrid pls-sem and fsqca approach afm jalal ahamed1, dominika jakubowska2, bogdan pacholek3, and katarzyna dziewanowska4 abstract this study investigates the factors influencing financial management behaviors (fmb) among young adults in poland, drawing on the theory of planned behavior (tpb). specifically, it examines the roles of financial attitude, family financial socialization, peer influence, financial self-efficacy, and subjective financial knowledge. data were collected from 340 university students using convenience sampling. a hybrid analytical approach combining partial least squares structural equation modeling (pls-sem) and fuzzy-set qualitative comparative analysis (fsqca) was employed. pls-sem results indicate that financial attitude, subjective financial knowledge, and financial self-efficacy are significantly associated with fmb. contrary to expectations, family financial socialization had no significant effect, while peer influence was significant but negative effect. fsqca revealed two distinct pathways leading to positive fmb, highlighting the importance of financial attitude and knowledge, even in the absence of strong peer or family support. the findings underscore the complexity of financial behavior and suggest that educational interventions should focus on improving financial attitudes and self-efficacy, while also acknowledging the important roles of family and peer influences. this study contributes novel insights into financial behavior research in eastern europe and demonstrates the value of combining sem and fsqca in behavioral finance studies. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation ahamed, a. f. m. j., jakubowska, d., pacholek, b., & dziewanowska, k. (2025). exploring factors affecting young adults' financial management behavior: a hybrid pls-sem and fsqca approach. financial services review, 33(5), 164-190. introduction the study of financial management behaviors (fmb hereafter) has emerged as a critical research domain (ahamed & limbu, 2024; deenanath et al., 2019; goyal et al., 2023; she et al., 2024), particularly in an era characterized by economic volatility and heightened geopolitical tensions. the proliferation of credit 1 corresponding author (jalal.ahamed@his.se). university of skövde, skövde, sweden. 2 university of warmia and mazury, olsztyn, poland. 3 poznań university of poland, poznań, poland. 4 university of warsaw, warsaw, poland. access, advancements in digital banking platforms, and diversification of investment vehicles has increasingly complicated the financial ecosystem, attracting significant attention from policymakers, financial institutions, and researchers. young adults represent a pivotal cohort in this context, as they navigate the complexities of early career https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ ahamed et al. 165 trajectories, debt management, and strategic long-term financial planning. this period of life encapsulates significant transitional milestones—such as labor market entry, engagement in higher education, and personal financial stewardship—all of which necessitate strong financial literacy and informed financial decision-making capabilities. understanding fmb within this demographic is imperative, as financial practices established during this formative stage exert profound implications on their sustained financial solvency and overall economic well-being (porto & xiao, 2022; rai et al., 2019; sabri et al., 2023; she et al., 2024; xiao & porto, 2017). despite a number of reviews, meta-analyses, and empirical studies on fmb from regions like the united states, western europe, northern europe, and southeast asia, (goyal et al., 2021; kaur & singh, 2024; lučić et al., 2024; rudi et al., 2020) a substantial research gap persists in eastern european countries such as poland, the czech republic, and slovakia. this gap is particularly notable given the distinct cultural, social, and political dynamics that differentiate these nations (cwynar, 2021). the shift from communist regimes to market economies, alongside extensive financial reforms, has significantly altered their economic landscapes. however, contemporary scholarship inadequately captures the evolving financial attitudes and behaviors shaped by this unique historical trajectory. this transition has profoundly impacted younger generations, many of whom lack the inherited financial knowledge or established financial culture. moreover, financial literacy in eastern europe remains below that of western europe due to divergent post-wwii socio-economic developments. young adults in the eastern european region face challenges stemming from the lack of financial education during communism, the transition to a market economy, and the conflict between traditional values and modern economic demands. psychological and sociocultural factors, such as tensions between collectivism and individualism and distrust of financial institutions, remain underexplored. thus, further empirical research is crucial to understand emerging fmbs and attitudes in these contexts. in fmb research, the theory of planned behavior (tpb) (ajzen, 1991) has been widely used to predict financial intentions and actions by examining attitudes, subjective norms, and perceived behavioral control (xiao & wu, 2008). positive financial attitudes, such as toward saving and budgeting, are consistently linked to better financial behavior (castrogonzález et al., 2020; perry & morris, 2005). subjective norms—social pressures from significant others—also strongly shape financial decisions (georgiou et al., 2023). both descriptive and injunctive norms influence behavior, often stemming from family socialization and peer influence (bicchieri et al., 2018; salmivaara et al., 2021). parents model financial behavior, while peers contribute through social learning and normative pressure (goyal et al., 2023; gudmunson & danes, 2011). financial self-efficacy, or belief in one’s ability to manage finances, is positively linked to better financial practices and outcomes (farrell et al., 2016; lown, 2011). similarly, subjective financial knowledge—confidence in financial understanding—enhances decision-making and positive financial behavior (allgood & walstad, 2016; perry & morris, 2005). prior research (morris et al., 2023) highlights that financial behavior is shaped by a network of factors, including past experiences and attitudes influenced by knowledge. these psychosocial elements interact and shouldn't be viewed in isolation. yet, despite ample research, asymmetric analysis in this field remains underexplored (algarni et al., 2024; ahamed, 2024). to address these gaps in fmb research, this study investigates key factors and their influence on polish young adults’ financial behavior, including how combinations of these factors lead to positive or negative outcomes, using a hybrid pls-sem and fsqca approach (chaouali et al., 2024; pappas & woodside, 2021). employing both symmetric (net effects) and asymmetric (combinatory effects) modeling methodologies provides a more comprehensive and nuanced understanding of the antecedent factor patterns. sole reliance on a single methodology might limit insights into complex phenomena (gil-cordero et al., 2024). furthermore, complexity theory, which utilizes boolean algebra and asymmetric thinking, enables the exploration of non-linear phenomena (ragin, 2009). fsqca integrates qualitative and quantitative outlooks, financial services review, 33(4) 166 addressing the limitations of variable-oriented analysis while offering the precision of caseoriented analysis (mason et al., 2023). this approach allows for the examination of intricate conditions that predict outcomes, particularly concerning the fmb of young polish adults. by applying a hybrid analytical technique, this study is among the few in the domain that advances theoretical understanding by combining granular findings into propositions. these propositions may contribute to further theory development and policy interventions, particularly within the eastern european region. this study seeks to provide a more nuanced understanding of the factors influencing fmb among young adults in poland, an eastern european country, with the potential to generalize findings to other nations in the region. the results will contribute to the broader literature on financial behavior and offer valuable insights for policymakers and financial educators, both regionally and globally. specifically, the research addresses the following research questions: 1. what factors influence the desirable fmb of polish young adults? 2. to what extent do these factors exert their influence on financial behavior outcomes? 3. what combinations of causal factors contribute to positive or negative financial behaviors among young adults in poland? 4. what factors influence the desirable fmb of polish young adults, to what extent do these factors exert their influence, and what combinations of causal factors may lead to positive or negative financial behaviors among young adults in poland? this study makes three precise contributions to the understanding of fmb among young adults. first, the research investigates the effects of money attitudes, social norms (family financial socialization and peer influence), and perceived behavioral control (subjective financial knowledge and financial self-efficacy) within the framework of the tpb. second, the study employs a dual-method approach by combining pls-sem and fsqca to identify both direct effects and combinations of causal factors influencing financial behavior. third, the research focuses specifically on young adults in eastern european countries, where research in this area is limited, providing novel insights into the factors shaping their financial management practices. overall, the study expands the research on young consumers' fmb and highlights the dynamic interplay of various influencing factors. literature review and hypotheses development tpb as an organizing framework for fmb the theory of planned behavior (tpb) is a well-established conceptual framework used to explain human behavior across various domains, including financial behavior (ajzen, 1991; yeo et al., 2024). tpb posits that behavior is primarily driven by three core factors: attitudes, subjective norms, and perceived behavioral control, which together shape behavioral intentions and subsequent actions (ajzen, 1991; armitage & conner, 2001). this model provides a systematic structure for integrating diverse financial constructs, allowing researchers to organize complex influences into a cohesive framework (magwegwea & lim, 2020; yeo et al., 2024). first, attitudes within tpb capture an individual’s overall evaluation of the behavior in question. in the context of financial management, this includes positive or negative evaluations of saving, budgeting, and investing (shih et al., 2022). for instance, financial attitudes often reflect beliefs about the benefits of saving or the risks associated with financial planning, which have been shown to influence savings intentions (magwegwea & lim, 2020). second, subjective norms refer to the perceived social pressure to engage or not engage in a particular behavior (ajzen, 1991). in personal finance, this includes influences from family, peers, and mentors, often captured through concepts like financial socialization, where parents and peers shape attitudes, skills, and habits related to money (yeo et al., 2024). this aligns well with tpb’s emphasis on social influence, as these norms directly affect financial intentions by reinforcing or discouraging certain financial behaviors (magwegwea & lim, 2020). third, perceived behavioral control encompasses both internal self-efficacy (confidence in one’s ability to manage finances) and external constraints (such as financial resources and access to information) (shih et al., 2022). this component is critical for financial decisionmaking, as a person’s belief in their ability to effectively manage money directly impacts ahamed et al. 167 their financial behaviors (shih et al., 2022). for example, low financial self-efficacy or limited financial knowledge can significantly reduce one’s perceived control over financial decisions, thereby reducing the likelihood of engaging in positive financial behaviors (magwegwea & lim, 2020). together, these three components form a coherent causal chain, linking attitudes, social influences, and perceived control to financial intentions and behaviors (ajzen, 1991; magwegwea & lim, 2020). this structure not only clarifies the relationships between diverse financial factors but also makes it easier to compare findings across studies. unlike standalone measures of financial attitudes or self-efficacy, tpb offers a unified framework that captures the complex, multi-dimensional nature of financial decision-making (rutherford & devaney, 2009; yeo et al., 2024). overall, tpb provides a comprehensive lens for understanding financial behavior, integrating various psychological and social factors into a single, cohesive model. this integration enhances clarity, supports systematic analysis, and facilitates meaningful comparisons across different financial contexts, without dismissing earlier research (magwegwea & lim, 2020; rutherford & devaney, 2009; yeo et al., 2024). financial management behavior (fmb) financial behavior encompasses a wide range of activities such as saving, budgeting, investing, and spending (morris et al., 2023; rai et al., 2019; xiao & wu, 2008). essentially, it involves the techniques individuals use to handle their income and financial situations (saurabh & nandan, 2018). positive financial behavior is defined as actions that promote financial well-being and stability, such as budgeting for expenses, saving for both short and long terms, and preparing for emergencies (saurabh & nandan, 2018). in contrast, engaging in negative financial behaviors, such as excessive reliance on credit and loans, can undermine financial well-being (rai et al., 2019). research indicates that financial behavior is a significant determinant of financial satisfaction (joo & grable, 2004; xiao & porto, 2017). understanding the factors associated with financial behavior is important for informing the design of interventions and educational programs aimed at supporting better financial outcomes. the tpb (ajzen, 1991) provides a robust framework for examining the factors associated with financial behavior. according to tpb, three primary components—attitude, subjective norms, and perceived behavioral control—predict behavioral intentions and actual behavior. past research also highlights that financial behaviors and decisions are influenced by external factors (such as culture, social class, and social groups) and internal factors (like self-esteem), which relate to an individual's self-motivation, learning, and personality (morris et al., 2023). for example, goyal et al. (2023) identified financial socialization, psychological traits, and financial literacy as key factors linked with personal financial management behavior among young indians. this study seeks to explore the factors associated with financial behavior among polish young adults using the tpb framework. by examining the effects of financial self-efficacy, financial attitude, family financial socialization, peer influence, and subjective financial knowledge, this study seeks to provide a comprehensive understanding of the factors that promote positive financial behavior. financial attitude and fmb within the tpb, attitudes are central predictors of behavioral intentions, influencing the likelihood of performing specific actions (ajzen, 1991). financial attitude refers to an individual's positive or negative evaluation of financial practices such as saving, budgeting, investing, and spending (castro-gonzález et al., 2020). positive financial attitudes, reflecting favorable evaluations of prudent financial management, increase the likelihood of engaging in responsible behaviors. previous studies underscore the critical role of financial attitude in predicting financial behavior (rai et al., 2019). for instance, perry and morris (2005) found that individuals with positive financial attitudes are more likely to save regularly and budget carefully. similarly, parrotta and johnson (1998) observed that young adults with positive financial attitudes demonstrated better financial management practices, highlighting the direct impact of attitudes on behavior. those with positive financial attitudes are more likely to prioritize long-term financial goals and recognize the benefits of responsible financial practices, such as saving and budgeting. additionally, these attitudes contribute to emotional regulation, financial services review, 33(4) 168 reducing financial stress and fostering rational decision-making. in light of these insights, we propose the following hypothesis: h1: financial attitude has a positive and significant effect on fmb. social norm and fmb norms represent the social pressure individuals experience to perform specific behaviors based on the beliefs and expectations of significant others (ajzen, 1991). social circles, including family, friends, colleagues, and acquaintances, exert considerable influence on an individual's financial decisions, often subconsciously (georgiou et al., 2023). social norms refer to the collective expectations and rules within a group regarding appropriate behavior (bicchieri et al., 2018). these norms are frequently conveyed through social support networks that offer guidance, advice, and encouragement in fostering sound financial practices. a supportive network can facilitate adherence to financial goals and help when challenges arise. social norms can be categorized into two key subdimensions. first, descriptive norms refer to perceptions of what most individuals do in a particular context (salmivaara et al., 2021). when individuals perceive that their peers are engaging in positive financial behaviors, such as saving, budgeting, or investing, they are more likely to adopt these behaviors themselves. observing others' successful financial management creates expectations and motivates similar actions. second, injunctive norms reflect perceptions of which behaviors are socially approved or disapproved (salmivaara et al., 2021). when responsible financial behaviors are seen as valued and endorsed by significant others—such as family, friends, or the broader community—individuals may experience social pressure to conform. positive reinforcement from important others can further motivate the continuation of good financial habits. in this study, social norms are divided into two categories: family financial socialization and peer influence. family financial socialization and fmb family financial socialization is the process through which individuals acquire financial knowledge, skills, attitudes, and behaviors from family interactions (goyal et al., 2023). this involves direct instruction, observational learning, and discussions about financial matters within the family (legenzova & leckė, 2024). early exposure to financial practices, particularly through parents as role models, can have a lasting influence on financial behaviors. children often emulate their parents' practices, such as saving, budgeting, and prudent decision-making. in addition to modeling, families communicate explicitly about financial matters, providing guidance on financial planning, budgeting, and saving. this socialization shapes financial attitudes and beliefs aligned with family values, promoting responsible financial behavior, such as prioritizing saving and avoiding debt. research consistently supports the positive impact of family financial socialization on financial behavior. for example, shim et al. (2010) found that when parents actively discuss financial matters and model responsible financial practices, young adults exhibit improved financial behaviors. similarly, jorgensen and savla (2010) identified parental influence as a significant determinant of financial attitudes and behaviors among college students. gudmunson and danes (2011) further highlighted that effective family financial socialization enhances financial self-efficacy and confidence, both of which are crucial predictors of responsible financial behavior. in the context of eastern europe, it is suggested that families play an especially strong role in shaping the financial attitudes and behaviors of their children. based on these findings, we propose the following hypothesis: family financial socialization has a positive and significant effect on financial behavior. this relationship is reinforced through mechanisms such as modeling, communication, reinforcement of social norms, and the transmission of financial knowledge and skills. h2: family financial socialization has a positive and significant effect on fmb peer influence and fmb peer influence significantly impacts the attitudes, beliefs, and behaviors of individuals within the same social or age group, particularly during young adulthood when financial independence is being established. peers, including friends and colleagues, shape financial behaviors such as saving, spending, and investing (godase et al., 2023) through social learning and norm-setting. observing ahamed et al. 169 peers who engage in responsible financial practices encourages similar behaviors, while positive peer pressure motivates adherence to sound financial management. in addition to norm-setting, peers often share critical financial information, including budgeting techniques, savings recommendations, and strategies for avoiding financial risks. access to this shared knowledge enhances financial literacy and improves decision-making competencies. furthermore, peers provide emotional support, which can be vital in navigating financial challenges and decision-making processes. within the framework of the tpb, subjective norms—beliefs about what others expect one to do—play a key role in shaping intentions and behaviors. peer influence, therefore, significantly contributes to the formation of these subjective norms, particularly in the financial behaviors of young adults. extant research consistently highlights the positive role of peer interactions in promoting desirable financial behaviors. for example, gudmunson et al. (2016) demonstrated that peer discussions on financial topics can enhance financial literacy and foster responsible financial conduct. similarly, hira et al. (2013) found that peer influence is a key determinant of financial practices such as budgeting, investing, and credit management, underscoring the extensive influence of social interactions on financial decision-making. therefore, we hypothesize that peer influence has a positive and significant effect on financial behavior among young adults. this relationship is supported by mechanisms such as social learning, normative influence, information sharing, emotional support, and comparative evaluation. h3: peer influence has a positive and significant effect on fmb. perceived behavioral control and fmb in the tpb, perceived behavioral control (pbc) is a key concept that reflects an individual's perception of how easy or difficult it is to perform a particular behavior (ajzen, 1991; conner and armitage, 1998). this perception is based on control beliefs, which are factors that can either facilitate or hinder the performance of the behavior (conner & armitage, 1998). control beliefs are divided into two categories: self-efficacy (internal control) and perceived controllability (external control) (ajzen, 1991). the concept of perceived behavioral control in the tpb, is consistent with bandura’s concept of “perceived self-efficacy” (georgiou et al., 2023). in this research, we used two categories for perceived behavioral control, derived from past studies: subjective financial knowledge and financial self-efficacy. these categories provide a comprehensive understanding of perceived behavioral control within the context of financial decision-making, as discussed below. subjective financial knowledge and fmb subjective financial knowledge refers to an individual’s self-perceived understanding and competence in financial matters. unlike objective financial knowledge, which pertains to verifiable knowledge of financial concepts and practices, subjective financial knowledge is based on personal perceptions and confidence in one's financial capabilities. this distinction is pivotal, as it highlights the significant role of self-perception in influencing financial decision-making and behavior. subjective financial knowledge is closely associated with financial self-efficacy, or the belief in one’s ability to effectively manage financial tasks. individuals who perceive themselves as financially competent tend to exhibit greater confidence in their financial decisions, which often leads to more favorable financial outcomes. individuals who believe they possess sufficient financial knowledge are more inclined to engage in prudent financial behaviors, such as consistent saving, strategic investing, and careful budgeting. higher perceived financial knowledge reduces financial anxiety (ahamed & limbu, 2024), fosters proactive management, and enhances motivation to engage in financial planning. it also improves risk assessment, enabling more cautious financial behavior. subjective financial knowledge influences financial intentions by increasing confidence in one’s ability to execute financial plans, aligning with perceived behavioral control in the tpb. past research underscores the strong relationship between subjective financial knowledge and positive financial behaviors, such as budgeting, saving, and investing (morris et al., 2023). allgood and walstad (2016) found that individuals with elevated levels of subjective financial knowledge are significantly more likely to engage in beneficial financial behaviors. similarly, perry and morris financial services review, 33(4) 170 (2005) demonstrated that subjective financial knowledge is a key predictor of financial behavior, even after controlling for objective financial knowledge. in light of these findings, we hypothesize that subjective financial knowledge exerts a positive and significant effect on financial behavior. this relationship is mediated through mechanisms such as enhanced confidence and self-efficacy, superior decision-making, increased motivation and engagement, improved risk perception and management, and strengthened behavioral intentions. h4: subjective financial knowledge has a positive and significant effect on fmb. financial self-efficacy and fmb financial self-efficacy, based on bandura’s self-efficacy theory, refers to an individual’s confidence in their ability to manage financial tasks and make informed decisions (farrell et al., 2016; lown, 2011). bandura’s framework asserts that belief in one’s capabilities significantly influences their success in performing actions (bandura, 1977). individuals with high self-efficacy are more likely to set ambitious goals, persist through challenges, and recover from setbacks (bandura, 1997). in financial contexts, this belief shapes how individuals handle financial challenges and opportunities, impacting both their intentions and behaviors. higher financial self-efficacy strengthens perceived behavioral control, a key element of the tpb, encouraging positive behaviors like budgeting and saving. greater confidence leads individuals to engage in better financial management (lone & bhat, 2024), and those with higher self-efficacy are more motivated to overcome financial difficulties (dare et al., 2023). research supports the link between financial self-efficacy and well-being. farrell et al. (2016) found that higher financial self-efficacy reduces financial stress and enhances well-being, while other studies show it prevents poor financial behaviors (hadar et al., 2013) and promotes prudent habits like saving (lown, 2011). robb and woodyard (2011) identified it as a key predictor of financial satisfaction and effective management. therefore, we conceptualize that financial self-efficacy plays a pivotal role in shaping financial behavior by enhancing perceived behavioral control and fostering positive financial practices. individuals who believe in their ability to manage finances are more likely to adopt behaviors conducive to financial well-being. consequently, we propose the following hypothesis: h5: financial self-efficacy has a positive and significant effect on fmb. methodology participants and sample design we employed a homogeneous convenience sampling technique, to specifically target university students in poland under the age of 29. this approach aimed to minimize sociodemographic variability, thereby enhancing the generalizability of the sample and reducing potential sampling bias (mason et al., 2023). an online survey questionnaire was distributed among business and management students across three universities in poland, located in the capital, the northeastern region, and the west-central region. the survey was conducted between may and july 2024. measures to measure fmb, we adapted 15 items from jorgensen (2007), rated on a 5-point likert scale ranging from 1 (not at all true for me) to 5 (very true for me). financial attitude was assessed using 15 perceptual items from jorgensen's "college students financial survey" (2007), also rated on a 5-point likert scale. financial self-efficacy was measured with six items from farrell et al. (2016), using a 5-point likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). family financial socialization was evaluated using a five-item scale adapted from zhao and zhang (2020), rated on a 5-point likert scale from 1 (strongly disagree) to 5 (strongly agree). peer influence was assessed using a self-developed questionnaire covering various aspects of personal finance, including saving habits, investment decisions, spending behaviors, borrowing and lending practices, and attitudes toward financial planning. each of the six items was rated on a 5-point likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). to measure subjective financial knowledge, we developed eight questions based on a review of the financial literacy literature (deenanath et al., 2019). these questions assessed perceptions across budgeting, investing, loans and debt management, financial security, financial communication, risk ahamed et al. 171 management, interest rates, and insurance, using a 5-point likert scale from 1 (strongly disagree) to 5 (strongly agree). the detailed measurement items can be found in appendix 1. analysis and results sample profile after excluding incomplete responses and addressing missing data, the final usable sample comprised 340 responses, with a median age of 22 years. the sample profile consisted of 340 respondents, with a median age of 22 years. as the table-1 below indicates, the majority of participants were female (64.7%), followed by male (34.7%), and a small proportion who preferred not to disclose their gender (0.6%). in terms of academic standing, the sample included first-year (27.9%), second-year (22.6%), third-year (21.2%), and postgraduate students (28.2%). most respondents were single or in a relationship (96.8%), while a small percentage were married (2.9%) or identified as "other" (0.3%). regarding work experience, 17.4% had no prior experience, 31.5% had less than two years, 25.9% had two to less than four years, 15.6% had four to less than six years, and 9.7% had six years or more. housing arrangements varied, with 4.4% living on campus, 47.4% in off-campus rental housing, 14.1% in their own off-campus residence, 30.6% living with parents or relatives, and 3.5% in other types of arrangements. table 1. sample profile (n =340) frequency % frequency % gender academic standing male 118 34.70 first year 95 27.90 female 220 64.70 second year 77 22.60 prefer not to disclose 2 0.60 third year 72 21.20 post graduate 96 28.20 median age 22 years marital status never been married / single / in a relationship 329 96.80 married 10 2.90 other 1 0.30 working experience housing arrangement none 59 17.40 on-campus 15 4.40 less than 2 years 107 31.50 off-campus rent 161 47.40 two to less than 4 years 88 25.90 off-campus own 48 14.10 four to less than 6 years 53 15.60 live with parents / relatives 104 30.60 six years or more 33 9.70 other 12 3.50 financial services review, 33(4) 172 researchers typically employ suitable methods for confirmatory and exploratory analysis (chanda et al., 2023). in this study, we utilized a combined approach of pls-sem and fsqca to verify the theoretical model (chanda et al., 2023; rasoolimanesh et al., 2021). pls-sem highlights general tendencies, while fsqca uncovers multiple realities in achieving desirable fmb. following configuration theory, fsqca examines complex, non-linear interplays between elements. this study selected fsqca along with pls-sem because it employs a fuzzy (continuous) scale for variables rather than a binary one, contributing novel insights to this research stream (chanda et al., 2023; rasoolimanesh et al., 2021). assessment of the model using pls-sem assessment of the measurement model to evaluate the reliability and validity of the measurement model, we employed smartpls (version 4.1.0.2), following established practices (chanda et al., 2023; chaouali et al., 2024). the analysis demonstrated satisfactory internal consistency and reliability, as indicated by cronbach’s alpha and composite reliability, as well as adequate indicator reliability (outer loadings), convergent validity (ave), and discriminant validity (htmt and fornelllarcker criterion) for most latent constructs (hair jr et al., 2014). while cronbach’s alpha for financial attitude (0.64) and aves for financial behavior and financial self-efficacy fell slightly below the conventional thresholds of 0.70 and 0.50, respectively, prior research supports the acceptability of cronbach’s alpha values above 0.60 in social science contexts (lam, 2012; shi et al., 2012). composite reliability ranged from 0.66 to 0.87, surpassing the minimum acceptable level of 0.60 (fornell & larcker, 1981; lam, 2012). thus, composite reliability alone can ensure sufficient convergent validity, even when more than 50% of the variance stems from error (fornell & larcker, 1981), confirming acceptable internal reliability (lam, 2012). to address common method bias, a common issue in survey research, two strategies were applied. harman’s single-factor test revealed that no single factor accounted for more than 50% of the variance (podsakoff et al., 2003). the construct reliability and validity statistics are depicted in table 2, 3 and 4. additionally, vif collinearity values for the factors were within acceptable limits, indicating that common method bias was not a concern. table 2. construct reliability and validity construct name cronbach's alpha composite reliability average variance extracted (ave) rho_a rho_c fmb 0.71 0.72 0.81 0.46 financial attitude 0.64 0.66 0.81 0.59 family financial socialization 0.71 0.76 0.82 0.53 peer influence 0.82 0.86 0.87 0.52 subjective financial knowledge 0.87 0.88 0.90 0.52 financial self-efficacy 0.78 0.80 0.84 0.48 to establish discriminant validity, this study used the htmt ratio, a method recommended in the literature (chanda et al., 2023; rasoolimanesh et al., 2021). for discriminant validity to be confirmed, each construct’s htmt value should be below 0.90. as shown in table 3, all htmt values meet this criterion. ahamed et al. 173 table 3. discriminant validity (heterotrait-monotrait ratio htmt) construct name 1 2 3 4 5 6 fmb (1) financial self-efficacy (2) 0.49 financial attitude (3) 0.84 0.63 family financial socialization (4) 0.17 0.11 0.20 peer influence (5) 0.23 0.19 0.23 0.32 subjective financial knowledge (6) 0.60 0.28 0.50 0.13 0.12 discriminant validity was further assessed using the fornell-larcker criterion, which requires the square root of each construct's ave to exceed its correlations with other constructs (fornell & larcker, 1981). as displayed in table 4, this criterion was satisfied for all constructs, confirming that each construct shares more variance with its own indicators than with other constructs. table 4. discriminant validity (fornell-larcker criterion) construct name 1 2 3 4 5 6 fmb (1) 0.68 financial self-efficacy (2) 0.39 0.69 financial attitude (3) 0.59 0.47 0.77 family financial socialization (4) 0.13 0.06 0.11 0.73 peer influence (5) -0.20 -0.16 -0.18 -0.21 0.72 subjective financial knowledge (6) 0.49 0.25 0.39 0.04 -0.01 0.72 assessment of the structural model in plssem to evaluate the model, we analyzed the variance explained and path coefficients of the endogenous variables using the adjusted r² (gil-cordero et al., 2024). financial services review, 33(4) 174 figure 1. the conceptual model with beta coefficients and the level of significance (source: figure by authors) the adjusted r² for fmb is 0.45 (figure 1), indicating a good model fit. this suggests that the model explains nearly 50% of the variance in fmb, which is a strong result for a parsimonious model (chanda et al., 2023; gilcordero et al., 2024). as shown in table 5, the path coefficients were tested using bootstrapping with 5000 samples (hair jr et al., 2014). financial attitude significantly associated with desirable fmb (β = 0.40, p > 0.05), supporting hypothesis h1. however, family financial socialization had no significant effect on fmb (β = 0.04, p < 0.05), leading to the rejection of h2. interestingly, peer influence exhibited a significant but negative impact on fmb (β = -0.10, p > 0.05), contrary to h3. both hypotheses related to perceived behavioral control were supported: subjective financial knowledge (β = 0.31, p < 0.05) and financial self-efficacy (β = 0.10, p < 0.05) had positive and significant effects on fmb, supporting h4 and h5. the findings reveal two unexpected results: the lack of effect from family financial socialization and the negative impact of peer influence on fmb. these results highlight the unique socio-economic context of eastern europe and prompted further exploration through qca analysis, as discussed in subsequent sections. table 5. path coefficients and p-values for hypothesized relationships relationships path coefficient (β) p values financial attitude -> fmb (h1) 0.40 0.00 family financial socialization -> fmb (h2) 0.04 0.31 peer influence -> fmb (h3) -0.10 0.02 subjective financial knowledge -> fmb (h4) 0.31 0.00 financial self-efficacy -> fmb (h5) 0.10 0.05 plspredict the plspredict method, following the 10-fold approach (shmueli et al., 2019), evaluates predictive relevance by generating case-level predictions at both the item and construct levels. as a holdout sample-based technique, shmueli et al. (2019) suggest that a structural model exhibits strong predictive power when the differences in rmse between pls and linear ahamed et al. 175 models (lm) are minimal. if the rmse values for all pls items surpass those of the lm, the model demonstrates high predictive error and lacks predictive power. conversely, if most lm rmse values exceed those of the pls, the model has moderate predictive power and error. a model demonstrates high predictive accuracy when only a minority of pls rmse values are lower than those of lm. in this study, as shown in table 6, the rmse values for most pls items were higher than those for lm items, indicating low predictive power for the model (chanda et al., 2023). table 6. results of predictive power assessment using plspredict items of fmb q²_predict rmse pls lm i budget and track spending 0.26 1.02 1.03 i read to increase my financial knowledge 0.24 1.16 1.16 i contribute to an investment account 0.22 1.34 1.36 i contribute to a savings account regularly 0.15 1.49 1.53 i find legal ways to lower my taxes 0.06 1.39 1.41 source: (rasoolimanesh et al., 2021) fsqca analysis we acknowledge that our sample is crosssectional and relatively small, but it meets the criteria for conducting pls-sem (partial least squares structural equation modeling). to address the deemed limitations associated with a smaller sample size, this study also employs fuzzy-set qualitative comparative analysis (fsqca) using version 3.0, an increasingly popular asymmetric method for understanding the complex interdependencies among the antecedents of young adults' financial management behavior (gil-cordero et al., 2024). combining pls-sem and fsqca is advantageous because these approaches offer distinct but complementary insights into the data. pls-sem is a variance-based method that estimates the average influence and significance (net effects) of each predictor on an outcome, making it well-suited for smaller samples and complex models without strict distributional assumptions (de andrés-sánchez & puchades, 2023). in contrast, fsqca is a set-theoretic, case-oriented approach (treats the responses as ‘cases’) that focuses on identifying the specific combinations (or "recipes") of conditions that are sufficient (and sometimes necessary) for the observed outcomes (ahamed et al., 2024; chanda et al., 2023; fainshmidt et al., 2020; ragin, 2009; rasoolimanesh et al., 2021). while pls-sem quantifies the strength and direction of each path within a structural model, fsqca captures complex causal relationships by identifying configurations of factors that together produce a given outcome, even when individual factors alone might have weak or inconsistent effects (ragin, 2008, 2009; de andrés-sánchez & puchades, 2023). a key advantage of fsqca is its ability to account for conjunctural causation and equifinality, meaning that multiple, non-exclusive combinations of conditions can lead to the same outcome (ragin, 2008). this approach recognizes that the same factor may have different effects depending on the context or combination in which it appears, capturing causal asymmetry more effectively than traditional symmetric models like pls-sem (de andrés-sánchez & puchades, 2023). both methods are particularly suitable for studies with smaller cross-sectional samples, which are common in behavioral research. plssem, for instance, is known for its flexibility with small sample sizes and does not require normally distributed data (de andrés-sánchez & puchades, 2023). it also relies on bootstrapping for significance testing, making it effective for studies with limited n (hair et al., 2019). similarly, fsqca, originally developed for medium-sized samples (10–50 cases/responses), is not constrained by the large-n assumptions of conventional regression analysis, making it a powerful tool for analyzing smaller data sets (ahamed, 2024; chanda et al., financial services review, 33(4) 176 2023; fainshmidt et al., 2020; ragin, 2009). recent studies confirm that fsqca can still yield robust results even with small samples, as it assesses subset relations rather than relying on probability distributions (de andrés-sánchez & puchades, 2023; fainshmidt et al., 2020). regarding causal inference, fsqca provides a unique advantage by evaluating necessity and sufficiency relationships within the data. it examines the associations between conditions (or combinations) that are always present when the outcome occurs (necessary conditions) or that consistently lead to the outcome (sufficient conditions) (ragin, 2008, 2009). unlike regression-based approaches, fsqca does not rely on p-values but instead assesses consistency and coverage scores to determine the strength of these causal links (vis & dul, 2018). for example, if financial literacy and self-control consistently co-occur with positive financial management behavior, this combination would be considered a sufficient configuration, supporting a causal interpretation under the assumption of causal complexity. in summary, integrating pls-sem and fsqca in a single study provides a more comprehensive analysis by capturing both the net effects of individual predictors (via plssem) and the complex, context-dependent causal pathways that drive the outcome (via fsqca). this hybrid approach is particularly valuable in studies with small, cross-sectional samples where multiple, interacting factors influence the observed behaviors (gil-cordero et al., 2024; ragin, 2009), acknowledging that different combinations of antecedents can lead to the same outcome. the conceptual qca model is depicted in figure-2. figure 2. the conceptual qca model calibration the first step in fsqca analysis is calibration, which assigns fuzzy set values to each data point, ranging from 0 (full non-membership) to 1 (full membership) (mason et al., 2023; ragin, 2009). this study uses the direct calibration method (ragin, 2008). latent variable scores from the pls-sem software are used, with the 95%, 50%, and 5% percentiles selected as anchors to distinguish membership degrees (ragin, 2008). finally, the original data are calibrated using the "calibration" function in the fsqca software. necessary condition analysis after calibration, the next step is identifying necessary conditions before configurational analysis, assessing if each condition consistently aligns with the outcome (ragin, 2009; mason et al., 2023). necessary condition analysis (nca) examines whether a single condition can predict high membership in the outcome (chanda et al., 2023; rasoolimanesh ahamed et al. 177 et al., 2021). a condition is considered necessary if it must be present for the outcome, with a consistency score above 0.9 (chanda et al., 2023). in this study, fmb is the outcome. table 7 shows no single antecedent sufficiently predicts this behavior, consistent with complexity theory (ragin, 2009; ahamed, 2024). table 7. analysis of the necessary conditions condition outcome high fmb (positive) (~) low/medium fmb consistency coverage consistency coverage financial attitude 0.82 0.76 0.56 0.55 ~ financial attitude 0.51 0.52 0.76 0.82 family financial socialization 0.68 0.68 0.60 0.63 ~ family financial socialization 0.63 0.60 0.70 0.70 peer influence 0.60 0.60 0.67 0.72 ~ peer influence 0.72 0.67 0.63 0.63 subjective financial knowledge 0.78 0.75 0.56 0.57 ~ subjective financial knowledge 0.55 0.54 0.75 0.79 financial self-efficacy 0.57 0.56 0.71 0.75 ~ financial self-efficacy 0.75 0.71 0.59 0.59 note: ~ indicates the absence of a condition. 4.3.3 sufficiency analysis the final step is a sufficiency analysis using a truth table to identify causal combinations for the presence or absence of desirable fmb. fsqca accounts for asymmetrical causal relationships, where explanations for the presence may differ from those for the absence (chanda et al., 2023; rasoolimanesh et al., 2021). both outcomes were analyzed with a minimum frequency of six and a consistency threshold of 0.90, following large sample standards (mason et al., 2023; pappas and woodside, 2021). standard analysis and default settings were used for the intermediate solution's counterfactual analysis. table 8 presents the intermediate solution from the fsqca, identifying configurations that are sufficient for achieving either high (desirable) or low/medium (undesirable) financial management behavior (fmb). black circles (●) indicate the presence of a condition, "x" circles indicate its absence, and blank cells represent "don’t care" conditions, meaning the factor is not critical for the specific outcome (fiss, 2011; pappas & woodside, 2021). large circles denote core conditions, which consistently appear in both parsimonious and intermediate solutions, while small circles indicate peripheral conditions, which play a more context-dependent role (fiss, 2011). key metrics include consistency, which measures the reliability of each configuration in predicting the outcome, raw coverage, indicating the proportion of cases explained by each configuration, and unique coverage, reflecting the distinct contribution of a particular configuration without overlap (ragin, 2008). these metrics provide a detailed view of how various conditions interact to produce complex financial behaviors, capturing both central and context-specific factors. financial services review, 33(4) 178 table 8. results of the intermediate solution (algorithm used quine-mccluskey) conditions outcomes high (positive /desirable) fmb (~) low/medium (negative /undesirable) fmb 1a 1b 2a 2b financial attitude ● ● family financial socialization ● peer influence ● subjective financial knowledge ● ● financial self-efficacy ● ● ● consistency (cc) 0.91 0.90 0.93 0.92 raw coverage (rc) 0.50 0.34 0.53 0.41 unique coverage (uc) 0.22 0.07 0.17 0.05 solution coverage 0.57 0.58 solution consistency 0.90 0.91 consistency cutoff 0.90 0.91 note: table notation source: fiss, 2011; pappas and woodside, 2021. as shown in table 8, the analysis identified two causal combinations for high (positive/desirable) fmb: proposition 1a: a high (positive) financial attitude combined with high (selfperceived) subjective financial knowledge, even with low peer influence and low financial self-efficacy, can lead to desirable fmb. a strong financial attitude, combined with high subjective financial knowledge, can drive desirable fmb, even with low peer influence and financial self-efficacy. individuals with positive financial attitudes view financial management as essential, which, alongside perceived financial competence, boosts proactive decision-making (talwar et al., 2021). low peer influence encourages autonomy, reducing negative fmbs (zulfaris et al., 2020), despite low self-efficacy, intrinsic motivation (di domenico et al., 2022) and belief in financial knowledge guide the desired financial behavior (kuhnen and melzer, 2018; weinstein and stone, 2018). furthermore, individuals seek consistency between beliefs and actions, leading to optimistic, proactive behaviors. self-directed learning compensates for low self-efficacy, reinforcing positive attitudes and financial knowledge over time. in post-communist societies like poland, where formal financial education has been limited, subjective financial knowledge is crucial for decision-making. many young adults rely on self-perceived knowledge from other sources or self-learning to compensate for the lack of formal education. positive financial attitudes, rooted in personal values of responsibility, motivate effective saving and budgeting. peer influence is less significant in a culture valuing independence, and strong attitudes and perceived competence may outweigh low selfefficacy. proposition 1b: a high (positive) financial attitude, along with high family financial socialization, high (self-perceived) ahamed et al. 179 subjective financial knowledge, and high financial self-efficacy, can result in desirable fmb. the combination of a positive financial attitude, strong family financial socialization, high subjective financial knowledge, and elevated financial self-efficacy forms a comprehensive framework that significantly enhances desirable fmb. a positive financial attitude drives alignment with long-term financial goals (talwar et al., 2021), while family financial socialization instills sound practices from an early age (ahamed and limbu, 2024). high subjective financial knowledge boosts confidence in decision-making, and financial self-efficacy reinforces belief in one’s ability to manage finances effectively (lind et al., 2020). together, these elements promote informed decision-making, autonomy, and responsible fmb. this proposition extends the pls results (table 5) and aligns with existing research, emphasizing that family financial socialization is an important predictor of fmb (ahamed and limbu, 2024) and it is most effective when combined with financial attitude, knowledge, and self-efficacy. in the polish context, where intergenerational transmission of financial habits is strong, family financial socialization is expected to play a vital role, particularly in postcommunist settings with limited formal financial education. here, high subjective financial knowledge compensates for educational gaps, while financial self-efficacy is critical in navigating economic pressures, such as rising costs and housing challenges. the analysis also identified two combinations for low/medium (negative/undesirable) fmb: proposition 2a: an unfavorable financial attitude and low (self-perceived) subjective financial knowledge, despite high financial self-efficacy, could still lead to selfreported undesirable fmb. an unfavorable financial attitude and low selfperceived financial knowledge can lead to maladaptive fmbs (lim et al., 2018), even when financial self-efficacy is high (park et al., 2024). negative financial attitudes undermine motivation for prudent practices like saving and budgeting, while low perceived knowledge causes poor decision-making despite confidence in one’s abilities. this misalignment fosters risky fmbs, with psychological factors like cognitive dissonance further exacerbating poor outcomes. thus, high self-efficacy alone cannot prevent undesirable fmb without positive attitudes and adequate perceived knowledge. in poland, where objective financial literacy is low, cultural conservatism and historical economic instability have shaped risk-averse financial attitudes. many individuals lack confidence in their financial knowledge, leading to inadequate savings and aversion to investment (cywar, 2020). limited financial education perpetuates this issue, with even those possessing high self-efficacy unable to engage in optimal fmb without positive attitudes and sufficient knowledge (aboagye and jung, 2018). research shows young adults with negative financial attitudes tend to neglect long-term planning and saving, critical in volatile economies (kaiser et al., 2020). low subjective knowledge further compounds this, as many young poles, perceiving themselves as financially uninformed, engage in risky behaviors such as overspending and poor budgeting (bień and gębski, 2024). consequently, financial self-efficacy, when unaccompanied by positive attitudes and sufficient perceived knowledge, does not facilitate sound fmb. the polish cultural paradigm, which places significant emphasis on resilience and self-reliance, especially in light of the country’s economic history, may foster financial self-efficacy but simultaneously lead to overconfidence and increased risk-taking without comprehensive understanding of the consequences. proposition 2b: a low financial attitude and low family financial socialization can still lead to self-reported undesirable fmb, even with high peer influence and financial self-efficacy. despite strong peer influence and high financial self-efficacy, low financial attitude and weak family financial socialization can still lead to undesirable fmbs. a low financial attitude reflects a lack of priority on financial management, resulting in neglect of budgeting and saving, leading to impulsive spending and poor planning. limited family financial socialization denies individuals early financial education, weakening financial habits and decision-making skills. financial services review, 33(4) 180 while peer influence can shape fmbs, these may not always be prudent, leading to risky spending and insufficient saving. high financial self-efficacy can also result in overconfidence, causing individuals to underestimate risks and make uninformed decisions. this misalignment between attitude, knowledge, and behavior means self-efficacy and peer influence are not enough without intrinsic motivation and foundational financial knowledge. in poland, where financial education often starts late, family socialization is crucial. weak family financial socialization leads to poor financial decisions despite peer influence. research shows peer networks influence spending and saving, but this is less effective if individuals do not value financial responsibility (kaiser et al., 2020). high self-efficacy, without a positive financial attitude and adequate family socialization, results in confidence without effective financial practices, leading to overspending or poor planning. discussion the pls-sem analysis revealed that financial attitude and perceived behavioral control— comprising subjective financial knowledge and financial self-efficacy—are positively linked with effective fmb, aligning with the tpb (ajzen, 1991). similar findings were reported by she et al. (2024) and dare et al. (2023), who identified positive effects of financial knowledge and self-efficacy on fmb. however, social norms, such as family financial socialization and peer influence, presented divergent results. while sabri et al. (2023) found financial socialization positively impacts fmb of malaysian respondents, this study found no significant effect of family financial socialization in poland, likely due to differences in family structures and financial literacy (cwynar, 2020; świecka et al., 2019; swiecka et al., 2020). consequently, young people may not receive the necessary skills and knowledge to manage their finances effectively from their parents. the transition from a communist to a market economy in poland has led to significant changes in financial systems and behaviors (cwynar, 2020). many parents may still adjust to these changes and might not be fully equipped to teach their children modern financial management skills, resulting in less impactful family financial socialization. furthermore, as families become smaller and parents have less time, the transmission of financial knowledge may decrease (kumar et al., 2024). additionally, we found that peer influence had a negative and significant impact on fmb, contrary to expectations. this may be due to the low levels of social trust in poland (curtis et al., 2010), where peer influence often encourages risky fmb, especially among young adults (gardner and steinberg, 2005). social pressures, particularly in consumer-driven environments amplified by social media, lead to impulsive spending on items like fashion and technology. peers, who may also lack financial knowledge, reinforce poor financial habits, resulting in impulsive consumption, inadequate saving, and rising debt. additionally, poland's cultural preference for immediate gratification over long-term planning exacerbates these behaviors (gschwandtner et al., 2022; panek, 2012). the unexpected pls findings, which indicate that social norms have a minimal influence on fmb, are consistent with prior research suggesting that personal norms, rather than social norms, play a more critical role in shaping economic behavior (bašić & verrina, 2023). further, this underscores the necessity for more sophisticated statistical methodologies beyond traditional regression models, to account for the diverse contextual variations influencing causal relationships. the fsqca analysis, in contrast, provided more robust and nuanced insights due to its case-oriented approach. while pls-sem evaluates the net effects of antecedents on outcomes, fsqca identifies various configurations of antecedents sufficient to explain outcomes (chanda et al., 2023). consequently, the two methodologies complement each other. through fsqca, two sufficient configurations leading to positive /desirable fmb and two leading to undesirable / negative fmb were identified (table 8). these findings expand upon the pls-sem results, reinforcing the notion that fmb is a complex and multifactorial phenomenon, necessitating both symmetric and asymmetric analysis for a comprehensive understanding. the fsqca results underscore that financial attitude, subjective financial knowledge, and financial self-efficacy are core antecedents of high fmb, whereas low financial attitude is a pivotal driver of undesirable financial outcomes. these insights also reflect the broader social and ahamed et al. 181 cultural characteristics of polish youth and, more broadly, eastern european populations. research implications theoretical implications the findings highlight the need to further refine theoretical models in the domina that integrate the interplay between financial attitudes, social norms (family socialization and peer influence), and perceived behavioral control (self-efficacy and subjective financial knowledge). the findings of those models should be understood within the specific cultural and social contexts. this approach can provide a more nuanced understanding of how internal and external factors combine to shape financial behaviors. in particular, the role of social influences, such as family and peers, warrants further academic exploration. although the pls-sem analysis did not find a significant positive effect of these factors on the desired financial behaviors of young adults, the fsqca results revealed that family financial socialization contributes to positive fmb (proposition 1b). conversely, the absence of family financial socialization is linked to undesirable fmb. these findings add a new dimension to the literature on fmb and offer valuable insights for future research. this study is among the few that explore the antecedents of fmb in young adults using a hybrid approach, combining both symmetric and asymmetric analytical techniques. this approach provides future researchers with the opportunity to analyze the complex phenomenon of personal fmb not only through regression-based symmetric methods but also by using asymmetric methods to further explain critical findings and identify causal combination patterns. managerial implications from a managerial perspective, the findings highlight the importance of designing comprehensive financial education programs that address not only financial knowledge but also attitudes and family influences (lebaron and kelley, 2021; lim et al., 2018; lind et al., 2020; talwar et al., 2021). financial institutions, educators, and policymakers in poland should develop targeted interventions that focus on improving financial attitudes and enhancing family financial socialization, alongside traditional financial literacy components. programs could include workshops and resources aimed at altering negative financial attitudes and fostering positive financial behaviors within families. additionally, given the influence of peer groups, creating peer-led financial education initiatives could leverage social networks to reinforce positive financial practices. understanding that high self-efficacy alone is insufficient without a supportive attitude and family background emphasizes the need for a holistic approach to financial education that addresses all these factors. these insights can help in crafting more effective policies and programs that promote long-term financial well-being among young adults. conclusion, limitations, and future research the study utilizes pls-sem and fsqca as complementary data analysis approaches, providing greater accuracy in the findings compared to previous research (chanda et al., 2023). this research revealed distinct patterns in financial behavior based on the interplay of various factors. the analysis in this research found that financial attitude and perceived behavioral control (subjective financial knowledge and financial self-efficacy) are positively and significantly linked with positive fmb. however, our findings related to social norms (i.e., family financial socialization and peer influence) contradict some previous research in the domain. this allowed to identify two key combinations for high (positive) financial behavior. the findings indicate that a high financial attitude, when combined with high self-perceived financial knowledge, can lead to desirable financial behavior, even if peer influence and financial self-efficacy are low. this suggests that personal belief in the importance of financial management and confidence in one's financial knowledge are crucial drivers of responsible financial behavior. the research also shows that high financial attitudes, coupled with strong family financial socialization, self-perceived financial knowledge and high financial self-efficacy, contribute significantly to positive financial behavior. this underscores the importance of a holistic approach that integrates individual attitudes, family influences and self-confidence in fostering effective financial management. conversely, an unfavorable financial attitude combined with low self-perceived financial knowledge can lead to undesirable financial behavior, despite high financial self-efficacy. financial services review, 33(4) 182 this highlights that confidence alone is insufficient to ensure positive financial outcomes if it is not supported by a positive attitude and sufficient knowledge. the study further reveals that low financial attitudes and limited family financial socialization can result in negative financial behaviors, even when peer influence and financial self-efficacy are high. this finding suggests that family financial socialization plays a critical role in shaping financial behaviors and that reliance on peer influence alone may not be enough to counteract the effects of poor financial attitudes and inadequate family guidance. several limitations must be acknowledged in this study. first, the research utilized a crosssectional design, capturing data at a single point in time. this approach restricts the ability to establish causal relationships between the factors studied and financial behavior. longitudinal studies would be necessary to track changes over time and determine causality more accurately. second, the study relies on self-reported data, which can be subject to biases such as social desirability or recall inaccuracies. future research should incorporate objective measures of financial behavior and outcomes to validate the selfreported data. additionally, this study focuses on polish young adults as a distinct and understudied population for examining financial behavior within the tpb framework. poland's rapidly changing financial landscape, economic transformation, and evolving social norms create a unique context for financial decision-making, offering insights that may differ from those observed in more extensively studied western or east asian populations. understanding how financial attitudes, social influences, and self-efficacy shape financial behavior in this context can provide valuable contributions to the broader literature on financial planning and education. while this context presents unique opportunities for understanding financial behavior, we recognize that the findings may have limited generalizability beyond polish university students. nevertheless, young adults globally face similar financial challenges, such as managing debt, building financial resilience, and making early career financial decisions, making these findings relevant beyond the polish context (xiao & porto, 2017; sabri et al., 2023). additionally, the combined use of plssem and fsqca captures both generalizable net effects and context-specific causal pathways, offering a more nuanced understanding of financial behavior that can inform educational programs and policy development globally (ragin, 2008; pappas & woodside, 2021). future studies could expand on this work by exploring similar factors in different cultural and economic contexts, further enhancing the practical relevance of the findings (gil-cordero et al., 2024). future research should explore several avenues to build on these findings. longitudinal studies are needed to examine how financial attitudes, family socialization, and self-efficacy interact over time to influence financial behaviors. additionally, research should include diverse populations and cultural contexts to assess the generalizability of the findings and identify culturally relevant factors affecting financial behavior. exploring the mechanisms through which peer influence interacts with financial attitudes and socialization can also provide a deeper understanding of its role in shaping financial decisions. finally, incorporating both qualitative and quantitative methods could offer a more comprehensive view of how individual and social factors collectively impact financial behavior. ahamed et al. 183 appendix 1. measurement items after deleting those with lower factor loadings factor loading financial management behavior i budget and track spending 0.66 i read to increase my financial knowledge. 0.70 i contribute to an investment account 0.75 i contribute to a savings account regularly. 0.71 i find legal ways to lower my taxes. 0.55 financial self-efficacy it is hard to stick to my spending when unexpected expenses arise 0.66 it is challenging to make progress towards my financial goals 0.66 when unexpected expenses occur, i usually have to borrow 0.60 when faced with a financial challenge, i have a hard time figuring out a solution 0.77 i lack confidence in my ability to manage my finances 0.80 i worry about running out of money in future (suppose ten years from now) 0.62 financial attitudes i feel in control of my financial situation 0.83 i feel capable of using my future income to achieve my financial goals 0.82 i feel putting away money each month for savings or investments is important 0.62 family financial socialization discussed family financial matters with me 0.57 spoke to me about the importance of saving 0.80 taught me how to be a smart shopper 0.80 taught me that my actions determine my success in life 0.73 peer influence i am more likely to save money if i know my friends are doing the same 0.55 recommendations from friends influence my investment choices 0.72 i feel pressured to maintain a lifestyle similar to my peers, affecting my spending habits 0.66 i consult my friends before making significant financial decisions 0.81 i am influenced by my peers when deciding to purchase high-value items 0.86 peer discussions have influenced my attitude towards taking loans 0.70 appendix 1 continued on next page. financial services review, 33(4) 184 appendix 1 continued. subjective financial knowledge i am confident in my ability to create a personal budget that reflects my financial goals 0.72 i understand how credit cards work 0.71 i understand the basic principles of investing in the stock market 0.76 i am familiar with the terms and conditions of my mortgage (or loan) 0.63 i know how to protect myself against financial fraud 0.71 i am capable of assessing the risks and returns associated with different investment options 0.76 i understand how interest rates affect savings and borrowing 0.76 i can explain the benefits and drawbacks of different types of insurance policies 0.74 ahamed et al. 185 references aboagye, j., & jung, j. y. (2018). debt holding, financial behavior, and financial satisfaction. journal of financial counseling and planning, 29(2), 208– 217. https://doi.org/10.1891/10523073.29.2.208 ahamed, a. f. m. j., jakubowska, d., & sadílek, t. (2024). financial anxiety of university students in poland and czechia: fsqca analysis. international journal of bank marketing. https://doi.org/10.1108/ijbm-04-20240229 ahamed, a. f. m. j. (2024). the pursuit of subjective well-being through financial well-being, relationship quality, and spiritual well-being: a configuration approach with fuzzy-set qualitative comparative analysis (fsqca). journal of family and economic issues, 1–18. https://doi.org/10.1007/s10834-02409904-7 ahamed, a. f. m. j., & limbu, y. b. (2024). financial anxiety: a systematic review. international journal of bank marketing. advance online publication. https://doi.org/10.1108/ijbm-01-20240021 ahamed, a. f. m. j., & limbu, y. b. (2024). role of social comparison orientation on financial management behavior in a developing nation: examining the mediating role of financial self-efficacy and the moderating effect of financial socialization. the bottom line. advance online publication. https://doi.org/10.1108/bl-12-20230083 ajzen, i. (1991). the theory of planned behavior. organizational behavior and human decision processes, 50(2), 179–211. https://doi.org/10.1016/07495978(91)90020-t algarni, m. a., ali, m., & ali, i. (2024). the role of financial parenting, childhood financial socialization and childhood financial experiences in developing financial well-being among adolescents in their later life. journal of economic and administrative sciences. https://doi.org/10.1108/jeas-01-20240005 armitage, c. j., & conner, m. (2001). efficacy of the theory of planned behaviour: a meta‐analytic review. british journal of social psychology, 40(4), 471–499. https://doi.org/10.1348/014466601164 939 bašić, z., & verrina, e. (2023). personal norms—and not only social norms— shape economic behavior. mpi collective goods discussion paper, 2020/25. https://dx.doi.org/10.2139/ssrn.373158 2 bicchieri, c., muldoon, r., & sontuoso, a. (2018). social norms. in e. n. zalta (ed.), the stanford encyclopedia of philosophy (fall 2018 edition). https://plato.stanford.edu/archives/fall2 018/entries/social-norms/ bień, a., & gębski, ł. (2024). consumers’ financial literacy in poland – the research and the resulting conclusions. journal of banking and financial economics, 1(21), 1–11. https://doi.org/10.7172/23536845.jbfe.2024.1.1 castro-gonzález, s., fernández-lópez, s., rey-ares, l., & rodeiro-pazos, d. (2020). the influence of attitude to money on individuals’ financial wellbeing. social indicators research, 148(3), 747–764. https://doi.org/10.1007/s11205-01902229-w chanda, r. c., vafaei-zadeh, a., hanifah, h., & thurasamy, r. (2023). modeling eco-friendly house purchasing intention: a combined study of plssem and fsqca approaches. international journal of housing markets and analysis. https://doi.org/10.1108/ijhma-062022-0086 chaouali, w., souiden, n., aloui, n., ben dahmane mouelhi, n., woodside, a. g., & ben abdelaziz, f. (2024). roles of barriers and gender in explaining consumers' chatbot resistance in banking: a fuzzy approach. international journal of bank https://doi.org/10.1108/ijbm-04-2024-0229 https://doi.org/10.1108/ijbm-04-2024-0229 https://plato.stanford.edu/archives/fall2018/entries/social-norms/ https://plato.stanford.edu/archives/fall2018/entries/social-norms/ financial services review, 33(4) 186 marketing. https://doi.org/10.1108/ijbm-07-20230325 conner, m., & armitage, c. j. (1998). extending the theory of planned behavior: a review and avenues for further research. journal of applied social psychology, 28(15), 1429–1464. https://doi.org/10.1111/j.15591816.1998.tb01685.x curtis, t., herbst, j., & gumkovska, m. (2010). the social economy of trust: social entrepreneurship experiences in poland. social enterprise journal, 6(3), 194–209. https://doi.org/10.1108/175086110110 88836 cwynar, a. (2020). financial literacy, behaviour and well-being of millennials in poland compared to previous generations: the insights from three large-scale surveys. review of economic perspectives, 20(3), 289– 335. https://doi.org/10.2478/revecp2020-0014 cwynar, a. (2021). financial literacy and financial education in eastern europe. in t. aprea et al. (eds.), the routledge handbook of financial literacy (pp. 203–211). routledge. dare, s. e., van dijk, w. w., van dijk, e., van dillen, l. f., gallucci, m., & simonse, o. (2023). how executive functioning and financial self-efficacy predict subjective financial well-being via positive financial behaviors. journal of family and economic issues, 44(2), 232–248. https://doi.org/10.1007/s10834-02209845-z de andrés-sánchez, j., & puchades, l. g.-v. (2023). combining fsqca and plssem to assess policyholders’ attitude towards life settlements. european research on management and business economics, 29(2), 100220. https://doi.org/10.1016/j.iedeen.2023.1 00220 deenanath, v., danes, s. m., & jang, j. (2019). purposive and unintentional family financial socialization, subjective financial knowledge, and financial behavior of high school students. journal of financial counseling and planning, 30(1), 83–96. https://doi.org/10.1891/10523073.30.1.83 di domenico, s. i., ryan, r. m., bradshaw, e. l., & duineveld, j. j. (2022). motivations for personal financial management: a self-determination theory perspective. frontiers in psychology, 13, 977818. https://doi.org/10.3389/fpsyg.2022.977 818 fainshmidt, s., witt, m. a., aguilera, r. v., & verbeke, a. (2020). the contributions of qualitative comparative analysis (qca) to international business research. journal of international business studies, 51, 455–466. https://doi.org/10.1057/s41267-02000305-y farrell, l., fry, t. r., & risse, l. (2016). the significance of financial self-efficacy in explaining women’s personal finance behaviour. journal of economic psychology, 54, 85–99. https://doi.org/10.1016/j.joep.2016.03. 001 fornell, c., & larcker, d. f. (1981). evaluating structural equation models with unobservable variables and measurement error. journal of marketing research, 18(1), 39–50. https://doi.org/10.1177/002224378101 800104 gardner, m., & steinberg, l. (2005). peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: an experimental study. developmental psychology, 41(4), 625–635. https://doi.org/10.1037/00121649.41.4.625 georgiou, m., daskou, s., anastasiou, a., & siakalli, m. (2023). the effects of the theory of planned behaviour on the switching propensity of retail banking customers at different critical switching incidents. international journal of bank marketing, 41(7), 1872–1898. ahamed et al. 187 https://doi.org/10.1108/ijbm-07-20220270 gil-cordero, e., ledesma-chaves, p., sánchez, r. a., & mariano, a. m. (2024). crypto-wallets revolution! key factors driving behavioral intention to adopt the coinbase wallet using mixed pls-sem/fsqca methodology in the spanish environment. international journal of bank marketing. advance online publication. https://doi.org/10.1108/ijbm-09-20230364 godase, r., jyothi, p., & supriya, m. l. (2023). financial planning propensity in working adults: exploring the role of media. managerial finance, 50(2), 313–328. https://doi.org/10.1108/mf11-2022-0552 goyal, k., kumar, s., & hoffmann, a. (2023). the direct and indirect effects of financial socialization and psychological characteristics on young professionals' personal financial management behavior. international journal of bank marketing, 41(7), 1550–1584. https://doi.org/10.1108/ijbm-09-20220300 goyal, k., kumar, s., & xiao, j. j. (2021). antecedents and consequences of personal financial management behavior: a systematic literature review and future research agenda. international journal of bank marketing, 39(7), 1166–1207. https://doi.org/10.1108/ijbm-07-20200378 gschwandtner, a., jewell, s., & kambhampati, u. s. (2022). lifestyle and life satisfaction: the role of delayed gratification. journal of happiness studies, 23(3), 1043–1072. https://doi.org/10.1007/s10902-02100417-1 hadar, l., sood, s., & fox, c. r. (2013). subjective knowledge in consumer financial decisions. journal of marketing research, 50(3), 303–316. https://doi.org/10.1509/jmr.10.0518 hair jr, j. f., sarstedt, m., hopkins, l., & kuppelwieser, v. g. (2019). partial least squares structural equation modeling (pls-sem): an emerging tool in business research. european business review, 26(2), 106–121. https://doi.org/10.1108/ebr-10-20130128 joo, s.-h., & grable, j. e. (2004). an exploratory framework of the determinants of financial satisfaction. journal of family and economic issues, 25, 25–50. https://doi.org/10.1023/b:jeei.000001 6722.37994.9f jorgensen, b. l. (2007). financial literacy of college students: parental and peer influences [doctoral dissertation, virginia tech]. https://vtechworks.lib.vt.edu/handle/10 919/30115 jorgensen, b. l., & savla, j. (2010). financial literacy of young adults: the importance of parental socialization. family relations, 59(4), 465–478. kaiser, t., lusardi, a., menkhoff, l., & urban, c. (2020). financial education affects financial knowledge and downstream behaviors. journal of financial economics, 139(2), 436– 453. https://doi.org/10.1016/j.jfineco.2020. 07.004 kaur, g., & singh, m. (2024). pathways to individual financial well-being: conceptual framework and future research agenda. fiib business review, 13(1), 27–41. https://doi.org/10.1177/231971452312 18563 kuhnen, c. m., & melzer, b. t. (2018). noncognitive abilities and financial delinquency: the role of self‐efficacy in avoiding financial distress. the journal of finance, 73(6), 2837–2869. https://doi.org/10.1111/jofi.12741 kumar, p., ahlawat, p., deveshwar, a., & yadav, m. (2024). do villagers’ financial socialization, financial literacy, financial attitude, and financial behavior predict their financial wellbeing? evidence from an emerging https://vtechworks.lib.vt.edu/handle/10919/30115 https://vtechworks.lib.vt.edu/handle/10919/30115 financial services review, 33(4) 188 india. journal of family and economic issues, 45, 1–19. https://doi.org/xxxxx (insert doi if available) lam, l. w. (2012). impact of competitiveness on salespeople's commitment and performance. journal of business research, 65(9), 1328–1334. https://doi.org/xxxxx lebaron, a. b., & kelley, h. h. (2021). financial socialization: a decade in review. journal of family and economic issues, 42(suppl 1), 195– 206. https://doi.org/xxxxx legenzova, r., & leckė, g. (2024). the link between family financial socialization in adulthood and investment literacy of p2p investors. journal of family and economic issues, 45, 1–20. https://doi.org/xxxxx lim, t. s., mail, r., abd karim, m. r., ulum, z. k. a. b., jaidi, j., & noordin, r. (2018). a serial mediation model of financial knowledge on the intention to invest: the central role of risk perception and attitude. journal of behavioral and experimental finance, 20, 74–79. https://doi.org/xxxxx lind, t., ahmed, a., skagerlund, k., strömbäck, c., västfjäll, d., & tinghög, g. (2020). competence, confidence, and gender: the role of objective and subjective financial knowledge in household finance. journal of family and economic issues, 41, 626–638. https://doi.org/xxxxx lone, u. m., & bhat, s. a. (2024). impact of financial literacy on financial wellbeing: a mediational role of financial self-efficacy. journal of financial services marketing, 29(1), 122–137. https://doi.org/xxxxx lown, j. m. (2011). development and validation of a financial self-efficacy scale. journal of financial counseling and planning, 22(2), 54–63. https://doi.org/xxxxx (if no doi, leave off doi) lučić, a., erceg, n., & barbić, d. (2024). knowledge and skills are not enough— exploring the determinants of financial behavior formation. international journal of consumer studies, 48(6), e13100. https://doi.org/xxxxx magwegwe, f. m., & lim, h. (2020). factors associated with the ownership of individual retirement accounts (iras): applying the theory of planned behavior. journal of financial counseling and planning, 31(2), 255– 269. https://doi.org/xxxxx (issue info filled based on typical jfcp formatting) mason, m. c., zamparo, g., & pauluzzo, r. (2023). amidst technology, environment and human touch: understanding elderly customers in the bank retail sector. international journal of bank marketing, 41(3), 572– 600. https://doi.org/xxxxx morris, t., kamano, l., & maillet, s. (2023). understanding financial professionals' perceptions of their clients' financial behaviors. international journal of bank marketing, 41(7), 1585–1610. https://doi.org/10.1108/ijbm-07-20220298 panek, e. t. (2012). immediate media: how instant gratification, self-control, and the expansion of media choice affect our everyday lives [doctoral dissertation, university of michigan]. https://deepblue.lib.umich.edu/handle/ 2027.42/94069 pappas, i. o., & woodside, a. g. (2021). fuzzy-set qualitative comparative analysis (fsqca): guidelines for research practice in information systems and marketing. international journal of information management, 58, 102310. https://doi.org/10.1016/j.ijinfomgt.202 0.102310 park, l., ward, d. e., naragon-gainey, k., fujita, k., & koefler, n. (2024). i’m still spending: financial contingency of self-worth predicts financial motivational conflict and compulsive buying. personality and social psychology bulletin, 50(2), 232–252. https://doi.org/10.1177/014616722311 54219 ahamed et al. 189 porto, n., & xiao, j. j. (2022). the role of consumer financial confidence on financial well-being. consumer interests annual, 68. https://www.consumerinterests.org/ass ets/docs/cia/cia2022/porto%20xiao %20cia2022.pdf ragin, c. c. (2008). measurement versus calibration: a set‐theoretic approach. the comparative method: moving beyond qualitative and quantitative strategies, 71–89. ragin, c. c. (2009). redesigning social inquiry: fuzzy sets and beyond. university of chicago press. rai, k., dua, s., & yadav, m. (2019). association of financial attitude, financial behaviour and financial knowledge towards financial literacy: a structural equation modeling approach. fiib business review, 8(1), 51–60. https://doi.org/10.1177/245526582199 1674 rasoolimanesh, s. m., ringle, c. m., sarstedt, m., & olya, h. (2021). the combined use of symmetric and asymmetric approaches: partial least squaresstructural equation modeling and fuzzy-set qualitative comparative analysis. international journal of contemporary hospitality management, 33(5), 1571–1592. https://doi.org/10.1108/ijchm-042020-0335 rudi, j. h., serido, j., & shim, s. (2020). unidirectional and bidirectional relationships between financial parenting and financial self-efficacy: does student loan status matter? journal of family psychology, 34(8), 949–959. https://doi.org/10.1037/fam0000660 rutherford, l., & devaney, s. a. (2009). utilizing the theory of planned behavior to understand convenience use of credit cards. journal of financial counseling and planning, 20(2), 48– 63. https://www.afcpe.org/assets/pdf/vol_2 0_issue_2_rutherford_devaney.pdf sabri, m. f., anthony, m., law, s. h., rahim, h. a., burhan, n. a. s., & ithnin, m. (2023). impact of financial behaviour on financial well-being: evidence among young adults in malaysia. journal of financial services marketing, 28(4), 337–356. https://doi.org/10.1057/s41264-02200136-7 salmivaara, l., lombardini, c., & lankoski, l. (2021). examining social norms among other motives for sustainable food choice: the promise of descriptive norms. journal of cleaner production, 311, 127508. https://doi.org/10.1016/j.jclepro.2021. 127508 saurabh, k., & nandan, t. (2018). role of financial risk attitude and financial behavior as mediators in financial satisfaction: empirical evidence from india. south asian journal of business studies, 7(2), 207–224. https://doi.org/10.1108/sajbs-042017-0050 she, l., rasiah, r., weissmann, m. a., & kaur, h. (2024). using the theory of planned behaviour to explore predictors of financial behaviour among working adults in malaysia. fiib business review, 13(1), 118–135. https://doi.org/10.1177/231971452410 59279 shih, h.-m., chen, b. h., chen, m.-h., wang, c.-h., & wang, l.-f. (2022). a study of the financial behavior based on the theory of planned behavior. international journal of marketing studies, 14(2), 1. https://doi.org/10.5539/ijms.v14n2p1 shi, j., mo, x., & sun, z. (2012). content validity index in scale development. journal of central south university. medical sciences, 37(2), 152–155. https://doi.org/10.3969/j.issn.16727347.2012.02.014 shmueli, g., sarstedt, m., hair, j. f., cheah, j.-h., ting, h., vaithilingam, s., & ringle, c. m. (2019). predictive model assessment in pls-sem: guidelines for using plspredict. european journal of marketing, financial services review, 33(4) 190 53(11), 2322–2347. https://doi.org/10.1108/ejm-022019-0189 świecka, b., grzesiuk, a., korczak, d., & wyszkowska-kaniewska, o. (2019). financial literacy and financial education: theory and survey. walter de gruyter gmbh & co. kg. swiecka, b., yeşildağ, e., özen, e., & grima, s. (2020). financial literacy: the case of poland. sustainability, 12(2), 700. https://doi.org/10.3390/su12020700 talwar, m., talwar, s., kaur, p., tripathy, n., & dhir, a. (2021). has financial attitude impacted the trading activity of retail investors during the covid-19 pandemic? journal of retailing and consumer services, 58, 102341. https://doi.org/10.1016/j.jretconser.20 20.102341 vis, b., & dul, j. (2018). analyzing relationships of necessity not just in kind but also in degree: complementing fsqca with nca. sociological methods & research, 47(4), 872–899. https://doi.org/10.1177/00491241177 13895 weinstein, n., & stone, d. n. (2018). need-depriving effects of financial insecurity: implications for well-being and financial behaviors. journal of experimental psychology: general, 147(10), 1503–1514. https://doi.org/10.1037/xge0000462 xiao, j. j., & porto, n. (2017). financial education and financial satisfaction: financial literacy, behavior, and capability as mediators. international journal of bank marketing, 35(5), 805–817. https://doi.org/10.1108/ijbm-012017-0005 xiao, j. j., & wu, g. (2008). completing debt management plans in credit counseling: an application of the theory of planned behavior. journal of financial counseling and planning, 19(2), 3–15. yeo, k. h. k., lim, w. m., & yii, k. j. (2024). financial planning behaviour: a systematic literature review and new theory development. journal of financial services marketing, 29(3), 979–1001. https://doi.org/10.1057/s41264-02300249-1 zhao, h., & zhang, l. (2020). talking money at home: the value of family financial socialization. international journal of bank marketing, 38(7), 1617–1634. https://doi.org/10.1108/ijbm-042019-0144 zulfaris, m. d., mustafa, h., mahussin, n., alam, m. k., & daud, z. m. (2020). students and money management behavior of a malaysian public university. the journal of asian finance, economics and business, 7(3), 245–251. https://doi.org/10.13106/jafeb.2020.v ol7.no3.245 financial services review volume 32 number 2 (2024) volume 32, no. 2 2024 editor: john e. grable, ph.d. cfp ® university of georgia advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp®, baylor university (emeritus) martin seay, ph.d., cfp ® , kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp ® , san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama jasmine fang, ph.d., massey university, new zealand mark fedenia, ph.d., university of wisconsin stu heckman, ph.d., cfp ® , texas tech university william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea thomas langdon, ph.d., roger william university, bristol, ri terrance martin, ph.d., winston-salem state university mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc lance palmer, ph.d., cfp ® , cpa®, university of georgia abed rabbani, ph.d., cfp ® , university of missouri chris robinson, ph.d., york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida issn online 1057-0810 print 1873-5673 contents grable, john e., from the editor, i-ii. grace, chuck, metzler, adam, miao, yang, feng, longlong, & fazeli, alireza. unveiling the winning contribution patterns for enhanced financial health. 1-28. zhang, yu. esg perceptions: investigating investor motivations and characteristics. 29-52. young, john h., hudson, crystal r., & copeland, c. w. factors mediating the association between financial socialization and well-being: an african american perspective. 53-76. ahamed, afm jalal & limbu, yam b. cognitive beliefs, retirement planning attitude, and money availability. 77-93. academy of financial services officers president shawn brayman smb research consulting executive vice president program michelle cull western sydney university vice president finance thanh ngo east carolina university vice president communications kirsten macdonald griffith university vice president international relations jasmine fang massey university vice president marketing & pr cora pettipas hsbc global wealth vice president membership matt goren dalton education, cerifi immediate past president tom potts baylor university editor, financial services review john e. grable, ph.d., cfp® university of georgia directors jason anderson university of kansas norah feng massey university wookjae heo purdue university thomas korankye the university of arizona barry mulholland university of akron mustafa nourallah mid sweden university richard stebbins university of alabama yu (yulia) chang kansas state university past presidents inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 32, no. 2, 2024 editor john e. grable, ph.d., cfp®, university of georgia editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • shinae l. choi, ph.d. university of alabama • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • stu heckman, ph.d., cfp®, texas tech university • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • thomas langdon, ph.d., roger william university, bristol, ri • terrance martin, ph.d., winston-salem state university • mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • chris robinson, ph.d., cfp®, cpa, ca, york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph.d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • lu fan, ph.d., cfp®, university of georgia • patti fisher, ph.d., virginia tech • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph.d. stetson university, deland, fl • philip gibson, ph.d., cfp®, winthrop university • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d. the ohio state university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • barry mulholland, ph.d., cfp®, university of akron • john nofsinger, ph.d., university of alaska anchorage • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peerreviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ pii: s1057-0810(01)00083-x �������� � �� ��� ��� ������� �� ������ ���� � ��� �������� ��� ��������� � �� �� ��� � ������������ ����� ��������� ������� �� �� �� ������� �� ��� �� ������ �� ����� ��� �� �� ����� � ���� �� !��" #$% ��&���' (� �� )��'�� ��� �� ��**% ��� ����� �� �� �� !""�# ����� �� �� �� ���� �� !" $%�� !""�# �������� & '������ !""� �������� (��� ����� � ��%���� ���%������ ���� ��%�� ������ ��� ����%��� )*���+� ���� *���� ���,���%������ *���� ��� ���� ��-���� �� ��� ���� %���. �/%�� ������� � ���,���� ��� ���� �� ��� �� ����� �������� (�� ������� � � ����0,� �� ������ �� ��� �� � � ���������� � %��� �� %��� �� �����1� ���� �� ������ ������� ��� ���%������ *�� ��� ��� ���� *��� (�� ����0,� �� ������ �� ���� �� �� � ���������. %������ � ���� � ���%�� �� ��� �� ��� ��� ��� ��� �� ��� ��� ����1��� ��.�����. ���,���%������ *���� ��� ���� ��-���� �� ��� ����� ��� 2����.� �������� ���� ��� ���,���%������ *�� �� ��� ���� �� ������ �� ���� ��%��� ���� ���. �� ��� ��� ����1���� � !""� 3%������� �� 4��� ��� ������� *��� ������' + 5��%������ *���# ���� *���# ������ ��� �������.# ���. �� ��� ��� ����1��� � ���� ������ 3������� �� ��.� ��� �� ��� ���� ��� ��0��� �� ���� � ��� ��� � ������� ��%��� � ������ ��� ���� � �� ��� � �������� (�� ����������� �����,��..�� ����� � 2������� �%����� �� ������ ��� �� ��.� �� ������. ������ ���%���� ��� ����� � ������ ��� ���� ��������� ��� �������� �� ��.�� �� ����. �������� ���� ������ ��� ����� ����� ����� ������� �� ��� , ���.�� �� ��.� ��� �� ��� ����� (�� �� ��� ��� ������� �� ������ ��� ��� � ����� �� ��� ������� ��� � ��� ���� � .�������� �� ����� �� ��� ����� (�� (������ ��� � �&&� )(��+ ��� ��� ������� � ������� ��� ������� � ��� ��� ��� �������� � � ��%� �� ��� �� � ��� ���� ���� ��%�� ������ ��� ����%�� )*��+� (�� ����� ����.�� ������ �� !""� �� � ��������� ��� �������%���� �� ��� �� ��� ��� *��� �0��. ��� ���� �� ������ � ��. ����� � �� ��� 6�������� ��� ���� �� ��� �" )!""�+ �789: �"7�,"9�";"�;< 8 ��� ���� ����� � !""� 3%������� �� 4��� ��� ������� *��� 3**� � � " 7 � , " 9 � " ) " � + " " " 9 = ,> � ?�����������. �%����� (���� ��,79",79�,!9"7# �-� ��,79",79�,!!7=� ,-�� � �''�� � + �� ��� �@�� �������% )���� ��������+� �������������@��� ������ )�� ��������+� � ������� ���� �� ���� (�� �� ��� ��� ������� ��� � ������� �� ��� ���� ��%�� ����%�� ��� ����������� ��-,���%������ *���� ��� ��� ���� *��� ��� ���,���%������ *��� ��� ��-, ���%������ � ������ ��������� ����� �%�� �� � a"�)0+� a"=)�+� �43� �� �* 3�4 ����� (�� ���%�� ������� � ������� �� �����2� ���� ��%��� ��� ����� ���� �� ���� � �� ��� ��� ��� � ���������� � � ������ ��������� ������� ������ �� ���� ��%�� ��� ���� �� ��� �� ���� ��-���� ����%���� ����� ��� ��� ��-,�� ����.�� �� ��� ����������� ������ �%� ���� ��� ���������� �� ��� �%��� ���� ������� �� ���� �� ������������ �� ��������.� (�%�� ����� ��� �� ���� �� ��� ��� ������� ���� �� ��� ������ ��� ���� �� ������ �� � �������� (�� ���� �� ������ �� ��������� ��������� �� 2������� �������� ��� �� ����� �� � �����,� � �%����� �����%������ �.����� %�%�� ���� ��.� ��� ��� � �-� �1��. ��� %�%�� � �� ��� ��� ������� ���� ��� �� � ���% %�����. � ������� ������ � �%�� �� ����� ��. �� � 2������� ���.��� b��� ��� ��� ������ �� �� �� �� ��� �� �� ��%� ����%�� � ��� �% ��� � ������ �� �� �� �� ��� �����,� ��� ��� %������������ �%���%����. ��� � 6%������ ��� �� ��� �����,� �������� � ��� �.����� ��� � ������ ������� � ������� ���� ��� �� ����� �� ��� ��� �������� (��� ����� �����1�� ��� ���.,��� �� ��� ��� � ��%�� � ������� �� ������ ��� .���� %���. ���������1�� �� �/%�� � ���,���� ��� ���� �� ��� �� ����� �� ��� ��� � ��%�� ��� �%..���� � ��� �������� �� ����� ��� ��� ���� �� ������� �� ��� ��� � <� �� ��� ��-,���%������ *�� �� �� ����� ���� <� �� ����� �� ��� ���� *�� ��� <� �%����� ��� *��� (�� ���� ����.� � ��� ���%������ *�� �� ����%��� �� ���� ��.�� ����� �� � ������� �� �� ��� ��� �%����� ��� *��� (��� �������� �� �� ����� �� ��� ��� %��� �� ?���� ��� �%���� )�&&�+� �� ���� �%� �������� ������ ��� �� ����� �� %�� ��� %�� ��������� � ��� ���� ����.�� � ��� �� ����� �� ��� ��� � ��%��� ?�� ������� ?���� ��� �%����c� )�&&�+ �������� %��� �/%�� �� ���,���� ��� ����� *� �� �%..����� ���� ��� ������ ���%�� �� ��� ����� ��� ���� �� ��� ��-, �� ����.� ��� �����,� �������������� �%� ���%�� ���� ����%�� /%������� � � ��%����� � ��� ������� � ������� ���� ��� �� ����� �������� (�� ����� �� ��.���1�� �� ������� ������� ! ��� ���� � ���� � �� ��� � ��� �-���� �������%�� �������. �� �� ��� ���� ���� �� �������� ������� ��� ��� ������ ��� ����%���� ������� = ��������� ��� �� ����0 ��� ��� ����� %��� �� ��� ��������� (���� ����� ��� ��������� ��� �-�������� � ��� �-���� ���� %��� �� ��� ���� �%� %�� � �� ����� �������� �� ��� �����,� ��.% ���� (�� ������� ���� .� �� ��� ���%��� � �� %������ �%�� %���. ��� ������ *� ��������� ��� ������� ����%���� ��� ������� ��� d�-������� � ������� ���� ��� �� ����� �� ��� ��� � ��%�� ��� �%..���� �-������ ������������� � ����� ������� �� ��� �������. ��� �� ��� ��� � ��%�� ������� a ��� ���� � �% ��� ��� �����%���. �� ����� � �������� �� ���������� *��� ��� ����� ��-,�� ����� �� ��� ���� �� � ��������� ������������ �������� �������� �� ��� ����� � .��� �� ��� �� �%�� �� b����� 3�������� ��� ����� )�&&�+� 4���� ���0� ����%�� e%�.��� ��� ���� )�&9"+ ��� �����1�� ��� � ���� � ��� ��� ��%�� ���������� ������� �� ����� ����������� ��� %�� �� %������ �� ���� � �� ����0,� �� �� ��� ��� ����1��� ��� ���� �� ����������� �� ��� *���� 'cf���� �� ����� ��� 5������ )�&9=+ ��� �� ���� ���0 %���. ��� ����� ����.�� �� �&9�� (��� �%�� ��� � ���� � ��� ��� ��%�� ���������� �: ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 ������� ��� ���,*�� �� ��� ���� ������ �� �����,�� � �� ������� �������� )�&&a+ �� ����� ��� ���,���%������ *�� ���� ���� ��-���� �� ��� ���� ��� 2��� ���� ��� �� �� �� ������ ��� ������� �������� )�&&a+� ���� ��� ��� ��� ����� �� �������� ����� ��-����� � ��� ���� �� ��� ���� ���� ������� .����� ������� ��� g�� ��� )�&&!+ %�� ����� ������� .���� ������� ��� 2�� ���� ���� �� ������ �� � %������ � �� ��� ��� ����1�� ��� ��������.� � ���%�� ���� �� ��-�� �� ��� ����� ����� (�� (�� �-������ ��� ������� � ������� �� ������ ��� �� ��.�� (�� ��� ��� ���� ��� � �% ��� � ����.�� �� �-�����. �� ��.� ����� �%� ��� ���� ������� ��� ��� ���� *��� ����� � ��� ������ %��� ���� ��������� �� ��-, ��� �� ��� ��� .����� (�� ��� ���� ������� � ����� � ��� �������� � ��%����. ����� �������� ��� 2������� ��� ���� 2� � ��� %�%�� %�� �� ������ �� � ������� ���,������ ����� �%����� �� ���� �� ������ �0� ��� ���� �� ������ ������� ��� ����������� ���%������ *�� ��� ��� ��� ���� *��� ������ �-� ���� � ����� �� �������� �� ����. ���� ����� ����%�� ��� ���0 �� ?���� ��� �%���� )�&&�+� b���� �� ��� )�&&�+� �������,������0� )�&&&+� e%����2���� $������ ��� ���0��� )!"""+ ��� �%����� 5����� ��� (����� )!""�+� ?���� ��� �%���� )�&&�+ � ��%��� ���%������ *���� ���� *���� ���,���%������ *���� ��� ���� ��-���� �� ��� ���� ���% ��. ��� ���������� ������� ��� � %�%�� %���� ?���� ��� �%���� )�&&�+ ���� ���� ���%������ *��� ��� ��� ���� *��� ��%�� ����%�� ��������� %�%�� ��%�� � ��� �� �����c� ������ ���� ��� ����.�� (�� �� ����� ��%�� �� ������ � ���� � ���%������ *�� � �� �-������ �� ��� � ����� ������ �� ��� %�%��� ?�� ������� � ���� *�� ��%�� �� ������ � ��� �� ����� �-������ �� �������� �� ��� ������� (�� 0�� �� ?���� ��� �%����c� )�&&�+ ���%��� �� ���� ���� %��� �� �������� �� ���� �/%�� �� ���,���� ��� ���� �� ��� ���%������ ��� ��� ���� *��� *� ����� ������ �� �� ����� �� ��� !9h ��.���� ������ �� ���% �� �� �� ��� <!�""" �� ��� ���%������ *�� ��� ���� <��aa" �� ��� ���� *��� (�%�� ?���� ��� �%���� )�&&�+� � �������� ���% � ���� ��� �� ����� ��� � ���������� �� %��� � ������� �� �� ��� ��� ��� ��� %��� � ������� ��� �� ���� �� <!�""" )��� � �%�� � ���%�� �������%���� �� �� �� ��� *���+ �� ��� ��-� ������������ ����� ���� ����� � ����� ���� � ��� � ���,���� ��� ��� �� ���� �������� ��� ���� ��.� �������� �� ��� ���%������ *�� �� �� ����� �� ��-,���%������ �� ��� ����� * ��� ���% �� ���� ��� �� ����� ��� �� ����� <!�""" )�� ��� � �%�� � ��� ���%�� �������%���� �� ��+ � ���,����� ���� ��� ���� ��.� �� ��� ���%������ *�� �� �� ����� �� ��-���� �� ��� ����� ��� ���� *������ ����%�� � ��.��� %�%�� ��%� � �� ���� ��� ������ �� ��� �� ����� ���� ��� ����.�� (��� �� ��� �������� %��� �� %�� ?���� ��� �%���� )�&&�+ ���� %�� ���%�� %�%�� %�� ���%��� �� �� ���� ��� ���, ���%������ *�� ���� ��-���� �� ��� ����� *� .������� ����� ���%��� ���� ���� ��� ���� *�� ��� ���%������ *�� ���� �� ����� ���,���%������ *�� ��� ��-���� �� ��� ����� (�� ������ ������� ���,���%������ *��� ��� ��-���� �� ��� ���� ���� ������ �� ��� ���� � ���%�� ��� ��� ��������.� � ���� � ��-�� �� � ����� ������� .���� ����� e%����2��� �� ��� )!"""+ %�� �� �������� ��� �� ���� �� ���� ?���� ��� �%���� )�&&�+ ��� �� � %� �� ���� ���%���� �� �������� �� ���� ���� �������� ���� ��� %�� ��� %�� ���� ��. ��������� � ��� ���� *�� ��� �� ��%� ������ ������� ��� ���%������ *��� �%���� �� ��� )!""�+ %�� �/%�� � ���,���� ��� ��� �� �� ���� ��� �� ��� ��� ��� �� ���� � ��� ���%������ *�� ���� ���� � ��� ���� *�� ��� �����%�� ���� ��� �� �����c� ��.���� ������� �� ��� �� � � �� ��� ��� ��� �� ��� �� � � ���������� � ��� ��� ������ � ��� �� ���� � ��� ��� ����� � *���� ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 �� g� �-���� ��� ����� �������� ��� %�� ����� �� ���� �� ���� %��� �� ?���� ��� �%���� )�&&�+ ��� �%���� �� ��� )!""�+� g� ���% �� �� �� �������� �� ���� �� �%���� �� ��� )!""�+� ���� ��� �� ����� ��� �� ��� ��� %�� � �%�� ��� ����� �� ��� ���� *��� 4 ���� ���� �� ��� �� �����. <� �� ����� �� ��� ���� *�� ���� <� �� ����� �� ��� ���%������ *�� � ��� %��� ����%����. �� ��� ���� ��.� �� ��� ������� '%� �������� ������ ��� �� ����� �� %�� ��� %�� ��������� � ��� ���� ����.�� � ��� �� ����� �� ��� ��� � ��%�� %� �� ��� �� �� ��� ����� �� ��� ������� g� ������%�� ��� ������� � � ii����0,� �� ������cc ��2��� �� ��� ������ �� ��� �� � � ���������� � %��� �� ����� ��������� ���� � �� ��� ��� ���� .� � �-����� ��� �� � %�%�� ��%� �� � .� �� �� ��� ��� ����1�� ��� ���� � ���%��� '%� ���� �� �� ������� � ��� ���� *�� ���� ��� ���%������ *�� ��� �� �� ���� ����0,� �� ������� (�%�� ��� ���� ��%��� ��� �������� �� ��� ��� 0�� %�������� ��������� ��� �� ��� ��� ����1��� ��� ���� � ���%�� ��� ��� ������ �� ��� �� � � ���������� � %��� �� ��� *���� (�� ����0,� �� ������ ����� ��� �� ����� ��� ��� ��� ������ ��� �� �� �� ��� �� � � ���������� � %��� �� ����� �� ��� �� ����2� �� �� ��� ��� �� ��� ���%������ *�� �� ��� ������ �� ��� ���� *��� � �������� �� ��� �� ������� ����������� g� %�� � �� ��� ���������� �� ����0 �� � ��%��� ��� ������� � ������� �� ��� ���� ��%�� �� ������ (�� ������� ������� � ������� �� �� ������ ����� ��� ��� �������. 2 � ������ �� 5��%������ *�� �� ���� �� <!�""" ��� ���� )<a�""" �� ������ ��%���� 2���. j������+ ��� � ������� ���� �� ���� ��%��� �����. ������� ��������� (���� �� ��� ���� ����� � � ������ ���������������� �%� ��� ���������� �� �� ����������� �� ��� � ������� ������ ��� �.�� !� 4 ������ ��������� ������ ��� ����� �%�� �� a"�)0+� a"=)�+� �43 ��� �* 3�4 ����� ���� ��� �� ���� �� *��� �� ������� ��� �%� ��� � �%�� � �� ��� ��� ��� ����� �� %�� ���.��� �� ���� ����� ��� ���� � ������� �� ��� ,� ������ )��.�� ���.� ����+� =� f��,���%������ *�� ���� ��� ��� ��-,�� ����� .����� ��� �� ���� �� <!�""" ��� ���� )<a�""" �� ������ ��%���� 2���. j������+� a� (�� ��� ���� *�� ���� ��� ��� ��-, ��� �� ��� ��� ���%��� �%� ���%�� �� ��� ��� �� ���� �� <!�""" )<a�""" �� ������ ��%���� 2���. j������+� 7� '��� ��-���� �� ��� ����� ����� �� ��� ����� � ��� ��� �������� ������� ��� �-���� �� ����� ��-�� �� ���.,��� ������� .���� ��� ��� ���� ��� ����������� �� ��� ���� (���� �� ��� ��� � ��%�� �� �� �� ������� ��� ��� ��� d�-������� �� ���0 � �� ���� ������� �� ��� ���� ��%��c� ������� �� �������� %��� ��� %�� ��� �� ��� ��� ��������� f��� ���� �� ����� �� �������%�� �� �� *��� ��� �� ����� %�� �� � ������ ���� � �� ����� �/%�� �� ��� � �%�� �� ����� �� *��� 6�� ������ ��%����� ������ ���%�� ��� �� � ��� ��/%���� ������ ���� �� (�� ���%� ���0�.� ������ �������� ��������� ��� ���� ��%�� �������%���� �� �� �� *��� �� <=�""" �� ����� !""!8!""a� <a�""" �� !""78!""� ��� <7�""" �� !""9� (�� ��� ���� ��� ���� �� <7"" ����� ���� �� ��d����� ��j%�� ���� � ��� !""9� 6�� ��-,������ �.� 7" ��� � ��� ��� �-� % �� ��� ��� %����� ��������� �� <7"" �� ����� !""!8!""7 ��� �� <��""" �� !"":� (�� ���%�� �� �� ������� �� ��� ����� �������%���� ��� ���%���� �� ��� *���� �9 ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 (�� (�� � ���� ��� �� ���� �� ������ ���� ��� �������. ����� ����� .��%�� ���� ������� �� ����� ���.������� �� �� ����� *��� ��� ����� �� ��� ��� �������� �� ,� 7�� ��� 7��� !1� � (��� .��%� ��%�� ����%�� �� ������ ���� ������ ��� ���� � )���� ���� <7!�""" �� ��%���� )<=!�""" �� ���.�� 2����+ �� !"""+ �� ����� ���� �� � ������ �� ���� ������� ����� �%� %�� ��.��� ���� � �� ��� )<�7"�""" �� ��%���� ��� <&7�""" �� ���.�� 2����+� (��� .��%� ���� ��� �� �������� ������ ������� ��� ��� *���� !� ,� 7�� ��� 1��� !1� ��' ���-'�'��� 7�� ���' � ���� !1�� (��� .��%� ��%�� ����%�� ����� ��� ��� �� ���� �� � ������ ��� ���� ������� ����� ��� ����� ���� � �-���� ��� �� ��� �� ���.������� �� ��-,���%������ *���� �%� ��� ���� �� ��� ���� *��� (��� .��%� ��%�� �� ��%��� �� ������ � ���� ��� ���� *�� �� � ���,�����%������ *�� ���� ��� �� � ��� ���%�� ���� �0�� �� ������ ���� ��� ���� *��� =� ,� 7�� ��� ��� ���-'�'��� 7�� !1� ����� (��� �� ��� .��%� ���� �� ���� � �� �� ��� ��.� �� ���.������� �� ��� ���� *�� )��%���� ���� ���� � � �� <�7"�""" ��� ���.�� 2���� ���� ���� � � <&7�"""+� (��� ��� �� ��� ���� �� ��� ���,���%������ ����������� *��� (�� ���%� ���0�.� � !""� ��� ������ ��� ���� � �� ��� �� ���.������� �� ���%������ ��� ���� *�� �� ��� ���� !""!� ��� .��%�� ��� �� ��� �� ���� ��-���� �� ��� ����� (�� ������ � �������� �� ��� 2��� .��%� �� ������� ��� ���%������ *�� ��� ��� ���� *��� 6�� ��� ������ .��%�� ��� ������ ���� ������ �� ��� ���� *��� *� �� �� ��%� ���� ��� ���� *�� ���� �� ����� ���� ��� ���,���%������ *�� ��� ��� ���� ��-���� �� ��� ����� 6�� ��� ����� .��%�� ��� ������ �� ������� ��� ���,���%������ *�� ��� ���� ��-���� �� ��� ����� g� ��0� %� ��� ������ ������� ��� ���%������ *�� ��� ��� ���� *�� 2���� �� �������� �� ���� �� ���% � �/%�� � ���,���� ��� ���� �� ���� *��� g� %�� ��� � ���,�����% %����� %�%�� ��%� � ��� �� ����� �� ��� ��� ������� �� ��� � ��%����� � ��� ���� �� ������� (�� ���� ��%�� �� ���� �� ��� �� � ���� � �� ��� ���� �� ��� �� ��� ��� ���� ��� ����� ��� �� � �� ���,�����%�� ���� � ���%��� (�� �������. ��������� ��� %���� & ���� � ���%�� )�� ��� ��-+ �� �� ��� ���� ���% �� �� �� �� � �� ��� ������� �" ��.���� ������ � ��� �� ����� �� ��� �� � � �� ��� ��� �� ��.���� ������ �� ��� �� ����� �� ��� ��-���� �� ��� ���� �! ��.���� ������ � ��� �� ����� �� ��� �� � � ���������� � %��� �� ��� *�� �� ��� ����0,� �� ������ �� ��� ������ �� ��� �� � � ���������� � %��� �� ����� ��� ���% %����� � ���,�� ��%� �� �/%�� �� ��� ���%������ *�� ��� ��� ���� *�� � �% ��� � ����� �� ����� ��� %��� �� ��� �� ��� *�� 6��( ���% %����� � ���,�� %�%�� ��%� 8�� '�'��� 7�� !1�� (�� ���%������ *�� ���� %� <�" �� ���� ������ �� ������ g� ���% � ���� ��%�� �� �� ����� �� ��� �� � ���� � ���%�� �� ��� *��� �%� ��� ���%��� ��%�� �� ��-�� ���� ���� �� ��� ��� ��.���� ������ � ��� ���� ��%��� ���% ��. � �������� ������ �� ��� �� ��� ��� � ��� ���� ��.� .�������� ��� ��� %� �� �������� �������� ����� ������� .���� ��-����� �� ���� �� ��� ���� (�� 0�� �� ����.� � ��� ���%������ *�� �� ��� �-��� ���� d�� � �" �� ��� �����%����� .�������� �� ���� ������ �� ����� �� ��� *��� (��� ��� �� �� �� ����� �� ��-���� �� ��� ����� (�%�� <� �� ����� �� ��� ���%������ *�� ��%�� ���%�� �� �� ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 �& � ���,�����% %������ �� ��� ��� � � ����� �� ������� 6��( �� ���%������ *�� � ��� &����� �!� � �"��� &��� ����� )�+ 1��� !1�� ��� �� ��� ��� �� ��� ���� *�� ��%�� .��� ������� ��-, ��� ��� ��� �� ����� ��� �-���� �� ����� � ��� ����� ���% %������ �����%� ��� ������.������� <� �� ����� �� ��� ���� *�� ��%�� ���%�� �� �� ���% %������ �� ��� ��� � � ����� �� ������� 6��( �� ���� *�� � ��� &�� )!+ 8�� ��� ��� (�� ���� �� ������ ��%�� �� � %������ � ��� ����� �������� �"� �� ��� �!� ��� ���� � ���%�� ��� ��� �� ��� ��� ����1��� g� ������%�� � ��� �������� �� ���� �������� �� ����0 �� ��%���. �� ��� ������ �� ��� �� � � ���������� ������ � �� ��� �� �����c� �%����� ������� (�� ��� �������� ����� �� ���� ��� �� ����� �� ������ � ���� ��� ���%������ *�� � ��� �� ��0��� �� ��� ��� �� � �� ����� ������ �� ��� �� � � ����������� '%� ���� �-�������� �� �%��� ���� ������ ��� ���� ����� ���� ��� ������ �� � %������ � ��� ���� � ���%��� ��� �� ��� ��� ����1�� ��� ��� �%����� ������ � ��� �� ������ 4/%����. )�+ ��� )!+� �� ��� ��� � �� �!� ��� ������ �� ���������� �� ����� ��� ���% %����� ��%�� �� ������ *����%�� �� �-����� ��� �� �� g� ���� ���� ��� ii����0,� �� ������cc� �� � ����0,� �� ������ � �"��� &��� ����� ��� &�� )=+ �� �� ����� ��� ����%���� ��� ����0,� �� ������ .� �� ��� �%����� ������� �-������ ���� � ���%�� ��� ������ � �� ��� ���� * ��� �-������ ������ �� ��� �� � � ���������� ���� ����� ���� ��� ����0,� �� ���� ��� ��%�� ����2� �� �� �����. �� ��� ��-,���%������ *��� '��������� ��� ��%�� �� ������ � ���� ��� ���� *��� *� ��� �� ���� ���� ��� ����0,� �� ������ ���� ������ �� ����� ���� �"� ��� �� �����c� �%����� ������� *� ����� ������ ��� �� ����� �� �� ���� � ����� ������ �� ��� �� � � ���������� �� ��� ���%������ *�� �� ��� �� ������ ���� ��� ���� *��� *� ��� ���� �� ���� ���� ��� ����0,� �� ���� ��������� ���� ��.��� ����� � ���%��� ���.�� �� ��� ��� ����1��� ��� ��.��� ��%�� �� ��� ' ��%���� �� ��� �� ����� ���� �"� (��� ������� ���� ��� ��-���� �� ��� ��� �� �� ��� �� � ���.,��� ������� .����� � ������� ���� ������ ���� ��� ������ �� ��� ��� ���%�� �� ��� ��-���� �� ��� ��� �� �� ��� �� � ���.,��� ������� .��� �����1�� ���� �� ��� �� � � ���������� � ��� *�� ����� � *� ���� ���� ��� ���% %����� ��%� �� ���� ������ � ��� ���%������ *�� ���� ��� 6��( � ��� &����� �!� � �"��� &�� � ��"��� &�� � �"��. )a+ ����� �. �� ��� ������� .���� ������� (�� 2��� ��� �� ��� ��.��,���� ���� �� ��� � ���,�����% %������ �� ��� *�� �� ��� ���� (�� ���� ��� ��� � ����%�� ��� � ���,�����% %������ � ��� ���� ��.� �� ����� �� ��� �� � ������ �� ��� *�� ��� ��-�� ���� �� ��� ��� � ��� �� ��� ��� ����1��� (�� �/%����� ��� �� �� ���2�� �� ������ 6��( � ��� &����� �!� � �"��� &����� �.� � �"�. )7+ (�� ����0,� �� ������ �� ���� ��� ������� ���� ��� �� �������� �� �� �����. )7+ ��� )!+ ��� ��� ��. �� �!� �� � �"��� �.� � �"�. ��� &�� ):+ 9" ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 6�� ���. �� ��� ��� ����1���� ���� ��%�� �����-� ��� �� �")�� �.+� (�%�� �� �� �� ����� ���� � �%����� ������ � �7h ��� � ������� .���� ������ � �"h� ��� ����0,� �� ���������� �� �=�7h� k���� ��� ��� ���%� ���0�.� �%���� �� ������ ���� �%����� ��.���� ������� � %� �� ="h ��� �� ���.���� )�%�j��� �� ����� ����������+ �� ���� ��� ���� *�� ��� ��� ���%������ *��� g� ����%���� ��� ����0,� �� ������� �� �� ����� ��%�� � �%����� �������� ����� � ���%�� ��� �� ��� ��� ����1��� 6�� ���� ��� � �� %�������� �� %�� ��� ��%�� �� ��� ��� ��%� � �" ��� ��� ����� ������� .���� ������ ����������� (���� � ����� ��� ���%��� �� ������� ������ � �"h� (�� ���%��� ��� �� �����.� (�� ����0, � �� ������ �� � ���������. %������ � ��� ���� � ���%�� ��� �� ��� ��� ����1��� �� �� ����� ���� �%����� ��.���� ������ � �"h ��� �� ��� ��� ����1�� � !" ����� ��� ������. � ���� � ���%�� � 9h ��%�� �� � �� �� � � ��.���� ������ ����� ���� 9�:h �� ��� �� � � ���������� �� ����2� �� ��� ���%������ *��� 3���� e ����� ��� ���%��� ���� ��� ��-���� �� ��� ��� � ��� ���� ��.� �� ��� ���%������ *�� �� ��-�� �� ��� ����� ������� .���� ����� (�� ����0,� �� ������ �� ����� ��� ��� �� 9�&h� ���%��� �� ������� ������� � �7h� !�h ��� ="h ��� ����� �� (����� !8a� �������� ���� �� �� ����� ���� �%����� ������ � !�h ������. 9h ���%�� ��� �� �� ��� ��� ����1�� � �7 ����� ���� �� � �� �� �� � ������ � ���� ���� �&�&h )!��:h ���� ��� ����� ������� .���� ���� ��-����� � ��-���� �� ��� ����+ �� ����2� �� ��� ���%������ *�� )(���� =+� *� .������� ��� ���� *�� ���� �� � ������ ������ �� ���� ��%��� ���� ���. �� ��� ��� ����1��� � �� ���� ���� ��� �� �%����� ������� � !�h �� ="h� (�� ������ ������� ��� ���%������ *�� ��� ��� ���� *�� ��� �� �� �� �� ��� �������. �����,� � 5��%������ *��l0���� �%����� ���� ��. �� �� �� ����� �� ��-���� �� ��� ����� ���� *��l����� � ������� � ��-�� �� %�%�� �� ��� ��� .����� (�� ���� *������ �� � ���� �� ��� %�%�� �� ��� ��� .���� ��������� (�%�� ��� ��.��� ��� ���� � ���%��� ��� ��0��� ��� �� ����� ���� �� ������ � ���� ��� ���� *��� �� ������� ��� ���.�� ��� �� ��� ��� ����1��� ��� ��0��� ��� ���� *�� �� ��� ������ ������� (�� ���%��� ����� �� (���� � ����%.� �%� ������� �� ��� ���� �����,� � (���� � (�-,���%������ �� ���� *��� ����0,� �� ������ *� ��� ��� ����1�� )�����+ ����� � ���%�� )�� ���� ��-���� �� ��� ��� � �"h+ ����� � ���%�� )e� ���� ��-���� �� ��� ��� � 9h+ :h 9h �"h �!h :h 9h �"h �!h 7 &��h &�:h &�:h &�7h &�9h &��h &�:h &�:h �" &�ah &�=h &��h &�"h &�:h &�ah &�=h &�!h �7 &�!h 9�&h 9��h 9�7h &�=h &��h &�"h 9�9h !" 9�&h 9�:h 9�=h 9��h &��h 9�&h 9�:h 9�ah !7 9��h 9�=h 9�"h ��:h 9�&h 9�:h 9�=h 9��h =" 9�ah 9�"h ��:h ��!h 9��h 9�ah 9�"h ���h =7 9�!h ���h ��=h :�&h 9�7h 9��h ���h ��ah a" 9�"h ��ah :�&h :�7h 9�=h ��&h ��7h ���h ��.���� ������ �� ��� �� � � �� ��� ��� � �"�� (�� �� ����� ���� �� ������ � ���� ��� ��-,���%������ *�� � ��� �-������ ������ �� ���������� � %��� �� ���� ���� ��� ����0,� �� ����� ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 9� '%� ���� ���� ��� ������� ��� %���������� �%���%����. ��� ������ �� ������ ��� �� ���������� � %��� �� ��� *��� f��� ���� ���� ���� �� � %������ � ���� ��� ���� � � ��� �� ����� ��� ��� ��� ����� ����� �� ��� �� � � ����������� ��� ���%����� �� %�%�� ��.���� ������� � ��� ��� ���%������ *��� b��� ��� ��� %���������� ����� ������ �� ���� �� � �� ��� ���%������ *�� ����%�� ������� ��� j%�� �� ��0��� .� %� �� ���� ��� �� .� ����� �"2" �9� �� ��' :�; 7 � �� *� .������� ��� ���� *�� ��� ��� ��� ���%��� � ��� ���%������ *��� �-���� ��� �%����� �����%������ b��� ��� ��� ���� *�� ��� .������ d�-������� ��� ��� ������� ���� ��� ���%������ *��� (�� ���� *�� ��� ��� ��-, ��� ����������� � ��� �������%����� ����� �� (���� = (�-,���%������ �� ���� *��� ����0,� �� ������ *� ��� ��� ����1�� )�����+ ����� � ���%�� )�� ���� ��-���� �� ��� ��� � !�h+ ����� � ���%�� )e� ���� ��-���� �� ��� ��� � !"h+ :h 9h �"h �!h :h 9h �"h �!h 7 !7�"h !a�ah !=�9h !=�=h !7�7h !7��h !a�:h !a�!h �" !=��h !!��h !���h !"��h !a��h !=�=h !!�7h !���h �7 !��ah �&�&h �9�:h ���ah !!�9h !��:h !"�7h �&�7h !" �&�9h �9�"h �:�ah �7�"h !��7h !"�"h �9��h ���7h !7 �9�ah �:�=h �a�7h �=�"h !"�=h �9�:h ����h �7��h =" ���"h �a��h �!�9h ���!h �&�!h ���=h �7�:h �a��h =7 �7��h �=�=h ���=h &��h �9��h �:�"h �a�!h �!��h a" �a�:h �!�"h �"�"h 9�=h ����h �a�&h �=�"h ���ah ��.���� ������ �� ��� �� � � �� ��� ��� � !��� (�� �� ����� ���� �� ������ � ���� ��� ��-,���%������ *�� � ��� �-������ ������ �� ���������� � %��� �� ���� ���� ��� ����0,� �� ����� (���� ! (�-,���%������ �� ���� *��� ����0,� �� ������ *� ��� ��� ����1�� )�����+ ����� � ���%�� )�� ���� ��-���� �� ��� ��� � �7h+ ����� � ���%�� )e� ���� ��-���� �� ��� ��� � �"h+ :h 9h �"h �!h :h 9h �"h �!h 7 �a�ah �a�!h �a�"h �=�9h �a�:h �a�7h �a�=h �a�!h �" �=�9h �=�ah �=��h �!�9h �a�!h �=�&h �=��h �=�7h �7 �=�!h �!��h �!�!h ���9h �=�9h �=�ah �=��h �!�9h !" �!�:h �!�"h ���ah �"�9h �=�ah �!�&h �!�7h �!��h !7 �!��h ���=h �"�:h �"�"h �=�"h �!�7h ���&h ���7h =" ���:h �"��h &�&h &�!h �!��h �!�"h ���ah �"�&h =7 ����h �"��h &�=h 9�7h �!�=h ���:h �"�&h �"�=h a" �"��h &�:h 9��h ��9h �!�"h ����h �"�ah &��h ��.���� ������ �� ��� �� � � �� ��� ��� � �7�� (�� �� ����� ���� �� ������ � ���� ��� ��-,���%������ *�� � ��� �-������ ������ �� ���������� � %��� �� ���� ���� ��� ����0,� �� ����� 9! ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 6����,�� � �� ��%���� ��� ��� ����� ��-, ��� ����������� � ������.� � ��� 7 ������ '���� �� ����.�� ����%�� �� ��/%��� ��� �� �������%���� �� ��� �.�� ��� ������ �� �������%�� �� ��� �.�� ��� �������� �� ����.�� �� � ������ ������ �� ������ ��� ������ �� ��%�.�� ����� )�������0 ��� �%���� )!"""++� ��� ��� )�&&&+ ��� ����� �� ��� ?��� )�&&&+ ����%�� � �% ��� � ���%�� �������. �� ��� %�� � *��� �� ������ �������. ������ (�� ��� ����� ��� �� ���� ��� ��� )�&&&+ ��� ����� �� ��� ?��� )�&&&+ �� ���� ��� ����������� *�� �� � .���� ������,�������. ����� (�� ���� *�� �� �� � �� ������ ����# ����%�� �� ��� �� ������ %�� � �������. ��� ���%������ *�� ���� ��� ���%���� ��j��� ��� ������� d�-������� ��������� (���� ������� ��� d�-������� ���%�� �0� ��� ���� *�� ��� �������� � ���� ��� ���%������ *�� �� ��� �� ������� �"#" <��-'�'��� 7�� !1� ��� � �9�� ��;�7�� ��� ����� (��� ������ ��� ���� ��%���� �� �������� )�&&a+ ��� ������� ��� g�� ��� )�&&!+� '%� ����� ��� ����������� �� �� (�� ���% %����� ��%�� �� ����� ��� ������� ��� �� ������� ��� 6��( �� ���,���%������ *�� � �� ���� &�� � ����� �!� )�+ 6��( �� ����� ��-���� �� ��� ���� � ��� &��� ����� )9+ (�� ���� ��-���� �� ��� ���� ��� �� ������ ���� ��� ���,���%������ *��� � ��� ��%� � �� �� ��.��2������ ����� ���� �!� (��� �� �������� � � ��.��2���� ���� � ��� ��-���� �� ��� ��� ���%��� ��� �� ��� �� � ���.,��� ������� .����� g� �� � ��� � �� %������� ����� ���% � ���� ��� ���� �� ��� ���� ��� ��-�� ���%���� �%� �� ��� ����� ������� .���� ������ � �"h �� ��� �7h ������ �� ����� ��� !"h �� ����� ���� ��.��� �������� ���%��� ��� ����� �� (���� 7� 3����� ��8�a ���� ��� ���% %����� ��%� �� ���� ������ �� ����� �� ��� ���,���%������ *��� (�� ������������. ��%�� �� ��� ���� �� ��� ���� ��� ����� �� 3����� e�8ea� (�� �% ���� �� ���� ��%�� �������� ���� ���� �� ��� ���� ��� ������� (���� a (�-,���%������ �� ���� *��� ����0,� �� ������ *� ��� ��� ����1�� )�����+ ����� � ���%�� )�� ���� ��-���� �� ��� ��� � ="h+ ����� � ���%�� )e� ���� ��-���� �� ��� ��� � !"h+ :h 9h �"h �!h :h 9h �"h �!h 7 !��7h !:�9h !:��h !7�7h !9�=h !��9h !��ah !:�&h �" !7�=h !a�"h !!�9h !��:h !:�9h !7�9h !7�"h !a�!h �7 !=�!h !��ah �&�9h �9�ah !7�=h !a�"h !!�9h !���h !" !��=h �&��h ���=h �7�:h !=�&h !!�=h !"�9h �&�7h !7 �&�:h ����h �7�"h �=�=h !!�:h !"��h �&�"h ���7h =" ���&h �7�=h �=��h ���=h !��=h �&�!h ���=h �7��h =7 �:�7h �=��h ���ah &�:h !"��h ���9h �7�9h �a��h a" �7��h �!�!h &�&h 9��h �&�"h �:�7h �a�ah �!�:h ��.���� ������ �� ��� �� � � �� ��� ��� � �7�� (�� �� ����� ���� �� ������ � ���� ��� ��-,���%������ *�� � ��� �-������ ������ �� ���������� � %��� �� ���� ���� ��� ����0,� �� ����� ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 9= *� .������� ���� �� ��� ���� ��� ������ �� ����� ����� � ���%�� ��� ��.��� ��%�� � �!� �� ��� �� ��� ��� ����1�� .��� ���.��� ��� ���,���%������ *�� .��� ������� 6�� ����� � ���%��� � :h �� ��.���� � ���,���%������ *�� �� ��� ������ ������ � ��� �� ��� ��� ����1�� �� ���.�� ���� !7 ������ � �� � ��� �� ����� ��� �� ��� =7h ���� ��� �� � � ����������� (�� ���%�� ��-����� � ���� �� ��� ���� �� ������ �� ����� (��������� ���� � ���� ����c� ���%��� �� � �� �%����� ���� � ��� ��� ��-�� �� ��� �%����� ������ ��� ��� �� �����. (���� 7 f��,���%������ *�� �� ���� ��-���� �� ��� ��� ���% %����� ��%� � <� �� ����� *� ��� ��� ����1�� )�����+ ����� � ���%�� )��� ���,���%������ *�� � �! � �7h+ ����� � ���%�� )e�� ���� �� ��� ��� ��-�� ���%���� �� ��� ���� � �"h+ :h 9h �"h �!h :h 9h �"h �!h 7 ��!9�7 ��=&9& ��7�9& ��:a9" � ���� � ��� � ���! � !!"" �" ��:�!! ��&97� !�=7a� !��&"" � !"�� � ���� � � !� !��99� �7 !��9�� !�9a:= =��""� a�9"!7 � ���" !�9=�a =�:a!7 a�:7:& !" !�9�:� a����9 7�9:9a 9�=a&= !�9:!& a�"�:& 7�:"aa ����:� ����� � ���%�� )�!� ���,���%������ *�� � �! � !�h+ ����� � ���%�� )e!� ���� �� ��� ��� ��-�� ���%���� �� ��� ���� � !"h+ :h 9h �"h �!h :h 9h �"h �!h 7 ��!a:& ��=a!: ��aa7� ��77:7 � �!�� � �!� � �!"� � ���� �" ��7��= ��9a:" !��:=a !�7=�= � �"�� � ��"! !��79& !�7"�" �7 !�"�&7 !�797� =�=�&a a�!:7� � ���� !�7=7& =���!! =�&77� !" !�:��! =�:�!7 7��9�� ��=��9 !�77a" =�a79� a�::�" :�!7a9 ����� � ���%�� )�=� ���,���%������ *�� � �! � ="h+ ����� � ���%�� )e=� ���� �� ��� ��� ��-�� ���%���� �� ��� ���� � !"h+ :h 9h �"h �!h :h 9h �"h �!h 7 ��!=:9 ��=!97 ��a!�a ��7==: � �!�� � �!� � �!"� � ���� �" ��77=: ��9��! !���7: !�a�a� � �"�� � ��"! � ���" � ���� �7 ��&��: !�7!"7 =�!!a� a��=�7 � ���� � ���" =���!! =�&77� !" !�7a7" =�7:!� 7�""&! ��"7!a � ���� =�a79� a�::�" :�!7a9 !7 =�="a= 7�"&=& ��99a= �!�!""" =�!!9� a���7: :�9a97 &�9&�7 =" a�=!"a ��=a=& �!�7�a: !��!��& a�"9�� :�a=": �"�":!� �7�:a!& ����� � ���%�� )�a� ���,���%������ *�� � �! � =7h+ ����� � ���%�� )ea� ���� �� ��� ��� ��-�� ���%���� �� ��� ���� � !"h+ :h 9h �"h �!h :h 9h �"h �!h 7 ��!�&9 ��="7� ��=&:9 ��a&77 � �!�� � �!� � �!"� � ���� �" ��7�a� ���7== !�"=7& !�=:99 � �"�� � ��"! � ���" � ���� �7 ��&"�9 !�a��& =�":7! =�&"�9 � ���� � ���" � � �� � "��� !" !�a=a: =�=�&: a��!!& :�:!"� � ���� � ���� a�::�" :�!7a9 !7 =��=&� a�9"�7 ��=&!: ���a""" � ��� a���7: :�9a97 &�9&�7 =" a�"9== :�9&"� ���:&!� �&�9!=& a�"9�� :�a=": �"�":!� �7�:a!& '��� ��-���� �� ��� ���� ���% �� �� �� ��-�� ���%���� �� ������� .���� ������� �! �� ��� ������ �� ��� �� � � ���������� � %���� e��� ��%�� �������� ���� ���� �� ��� ���� ��� ������ ���� ���,���%������ *��� 9a ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 ��-�� �� ��� ������ ������� .���� ����� � ��.��2���� ������� � ��� ������� .���� ��� ��� � ��� ��� �� ����� �� � ���. �� � ��� ��%� ��� � ���� � �� ����.�� � ���� �� ��� ���� ��� ������ �� �� %����� �� �-��� � ���� �� ���� ��� � ��� ���%��� ��� �� ������� .����� ����� ��� �� ����� ���� ��� ��� � ��� �� ��� ��� ����1��� 6�� ���� ����� ��� �� ����� ���� �� ������� ������ � ���� ��� ��-���� �� ��� ��� �� ��� ���% %����� ��%� ��%�� ��� 6��( � �� ���� &�� � ����� �.� )&+ ����� �. �� ��� ���.,��� ������� .���� ����� (��� ���� ��%�� �%���� �� ����� ���� �!� �"�" �9� �� ��' :�; 7 � �� (�� %��� ��� �� ����.� � ��� ���� ��-���� �� ��� ���� ������ � �� ��� ���,���%������ *�� ��� ��� ������� ��� d�-������� ��� ���� (���� ����%��� ������� �� �� �� ��-����� ����2������ k�� �� ������ ��� .� � �������. ������� k�� �� ��.�� ����%��� �� ��������. (���� d�-������� �� %�� � %��� *��� ��%�� �� ������ ���� �� ���� �� �������� ����� ������ �� ��� �������. ��� ������ �� ���.,��� �� ��� ���� � #������ �� ���������� (��� ����� � ��%���� ������ ��� ��� ����� ���.,��� �� ��� ��� � ��%�� � ������� �� ���� ��%���� ��� �� ������ ��� ��0��� �� ��� ��� �� 2�%�� ������ ������� ��� ���%������ *�� ��� ��� ���� *��� (�� ���� �� ������ �� � %������ � ��� ���� ��%��c� �� ��� ��� ����1��� ���� � ���%�� �� ��� �� ��� ��� ��� ����� �� ����� ��.���� �������� g� %�� ���������1�� � ���,���� ��� ��� �� ���� �� ���� ��� ������� (�%�� ��� ���� �� ����� <� �� ����� �� ��� ���%������ *�� ���� <� �� ����� �� ��� ���� *��� (�� ���� ��.� �� ��� ���%������ *�� �� ��� ��� �� %��� ����%���� �� �� ����. �� ����� �� ��-���� �� ��� ���� �� ��� �� � ���� � ���%�� ��� �� � �� ��� ��� ����1��� (��� ���� ��� ��� ��� �� ����� �� ��0� �� ����.� � ��� %�� ��������� � ���� *��� g� %�� ��� ������� � ����0,� �� ������ �� ��� �� � � ���������� � %���� (��� �� ��� ������ �� ��� �� � � ���������� ���� ��%�� �0� ��� ���� ��%�� ���� ����� ������� ��� ��� �������� (�� 2����.� �������� ���� ��� ����0,� �� ������ �� ������ ���� ���� ��� �� �����c� ������ �� ��� �� � � �� ��� ���� (��� ���� ���� ��� �� ����� ��� �� �� � � ����� ������ �� ��� �� � � ���������� �� ����� �� ����2� �� ��� ���%������ *��� (�� ����0,� �� ���� �� � ���������. %������ � � )�+ ��� ���� � ���%�� �� ��� �� ��� ���# )��+ ��� �� ��� ��� ����1��� ��� )���+ ��� ���� �� ����� ��� ��-���� �� ��� ��� .�������� �� ��� �� ��.� �� ������� ��-�� �� ��-��� (�%�� ��� ���� ��%��� ���� ���. �� ��� ��� ����1��� ��%�� �������� �� ������ � ���� ��� ���� *��� 6������� ��� ������� ��� d�-������� � ������� ���� ��� ���� *�� �0� �� �� � �� ������ ������� g� ���� � ��%��� ��� ������ ������� ��� ���,���%������ *�� ��� ���� ��-���� �� ��� ����� '%� �������� ��������� ��� ���,���%������ *�� ��%�� �� ��� ����� �� ����� ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 97 ���� ���. �� ��� ��� ����1���� b��� ��� � ��� ���� �� ��� ��� ���%��� �� � �� ��� �� � ���.,��� ������� .����� ����� ��� �� ����� ���� ��� �� � � ����������� ��� ���,���%������ *�� ����� ��� �� ����.�� 6%������ ��� ��� �� �������� ��� ������� ��� d�-������� � ������� ���� ���� �� ��� ����� (���� ��%�� �0� ��� ���� ��-���� �� ��� ����� � �����%� �������.�� �� ��� ���,���%������ *��� ���������� �� ����� ���� ��%��� ���� ������� �� ��� ��� ����1���� (�� 2����.� ���%������ ��� �� ���-��� ���� �� ������ ��� ���� ���� ����� ��.�����. ����� �� ��� ��� �������� (�� ��������� ��� �� �� �����,�%� �%� �� ��� %���������� ���������� ���� ���. �� � ����1���� (���� ���� ��� �������������� ���� ���� ������������� �� ���/%����� �� �� ����� ������� �� �� ���� ������� ��%�� ���� ������ ��� �� ����� � ����� 2������� .����� �����2������ ���� �������. ������� � ���%������ �� ���� *��� ������������� %�� ������ ��� ����� �� ����� ��� �� ��� ��� ����1�� �� �� .���� �� �� ������ ��� ���������� � � ������ �� ��� �� � � ���������� � ���������� (�%�� �� ��%�.�� �� ������� ��� ���� *�� %�� �� ��� � ���� ������ � �� ��� ���%������ *�� ����� ��� %���������� ��.�����. %�%�� ������� �� ��� .����� ���������� e%�.���� �� 5�� m ����� �� �� )�&9"+� � �� %������ ��%�� � ����������� ������ ��� ������ =������ �� ���� ��� 8�;�� �� � �� �� �� 2� =a8a�� e%����2���� �� ��� $������ 6� ��� m ���0���� 4� �� )!"""+� (�� ���� ���%� ��� ����������� *��� � �� ������ � ��������� =������ �� �99� �' �� �� 1� ����� 26)a+� ��=8�!9� ?����� (� ��� m �%����� $� �� )�&&�+� �� �������� � ��� ������ ������� ���� ����� ������.� �� *��� ��� ��� �������� ������� .����� 0 ���� �� ���� �� 1�� �� 6)a+� !!�8!a!� b����� �� �� 3�������� $� b�� m ������ �� )�&&�+� �� �������� � ���,���%������ *�� �������%����� ��� ���� *�� ��� �������� 0 ���� �� ���� �� 1�� �� 6)a+� !a=8!7:� �%����� n� g�� 5����� �� 5�� m (������ $� 3� )!""�+� *� ��� ��� ��� �� ���� �� ������� ������� ���� ��� ����������� ���� ��%�� ������ ��� ����%���� =������ �� �99� �' �� �� 1� ����� 2�)�+� 778:"� ����� ��� 5� 4�� m ?���� � b� )�&&&+� (�0��. ��� ��� �%� � *�� �������.� 8�� � ��' , ���� 2�%)�!+� :�8�=� 'cf���� ?� $�� �� ����� 5� ��� m 5������� � 3� )�&9=+� 3�� ��%�� ����������� �� ���� ��%�� ������ ��� ����%���� =������ �� ���� ��� 8�;�� �� � �� �� �� *� =78a=� ��������� g� �� )�&&a+� (�� � ���� � %�%�� %�� �������%����� �� � ���,�����%���� 0 ���� �� ���� �� 1�� �� �)!+� �!�8�a�� ��� ���� �� )�&&&+� %���.����������� ������ ��� �������.� ��������� ���%��� 8�� � ��' , ���� 2�%)�!+� �!8�a� �������� � ��� m g�� ���� � �� )�&&!+� 8�;� ��' 7� �� ����� �+ � 9���� � �99������ 4�.������ ?�� �� f$� 3�������,b���� �������,������0�� �� )�&&&+� 3������. ���%�� �� %���. ��� ���� *��� =������ �� 0 ���� �� ����� � 2#):+� =!8 ==� �������0� (�� m �%����� $� ?� )!"""+� !1� *�23&4 ��' ����� ��� ������ 9��� + ��& � ���� ����� ���� e��0����� ?�� f���� 9: ."�" �� ���� �" �������� / 0 ���� �� ���� �� 1�� �� 2� 3#��24 ��5%6 analysis of investment choices for retirement: a new approach and perspective introduction overview of literature analysis of iras and taxable investments options and flexibility non-deductible ira versus open taxable investments options and flexibility summary and conclusion references financial services review volume 33 number 3 (2025) volume 33, no. 3 2025 guest editors: peter öhman, mustafa nourallah & izidin el kalak advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting conrad ciccotello, jd. ph.d., university of denver john grable, ph.d., university of georgia sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp®, baylor university (emeritus) martin seay, ph.d., cfp®, kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama lu fan, ph.d., cfp®, university of georgia jasmine fang, ph.d., massey university, nz mark fedenia, ph.d., university of wisconsin philip gibson, ph.d., cfp®, winthrop university stu heckman, ph.d., cfp®, texas tech university stephen horan, ph.d., cfa®, university of north william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea izidin el kalak, cardiff university, uk thomas langdon, ph.d., roger william university, bristol claire matthews, ph.d., massey university, nz mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden lance palmer, ph.d., cfp®, cpa®, university of georgia abed rabbani, ph.d., cfp®, university of missouri dan richards, ph.d., york university, canada chris robinson, ph.d., cfp®, cpa, ca, york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida bomikazi zeka, ph.d., university of canberra, australia issn online 10.61190/fsr.v33i3 contents öhman, peter, nourallah, mustafa & el kalak, izidin, guest editorial, i-v. special issue: financial technology (fintech) brockbank, alex, kalenkoski, charlene, browning, christopher & guillemette, michael, is using a financial advisor related to cryptocurrency investment? 1-19. chen, ying, asebedo, sarah d., little, todd d., & ning, weihong, from intention to adequate emergency fund savings through fintech use: evidence from a u.s. survey study, 20-47. diab, sara, nourallah, mustafa & öhman, peter, borrowing from family and friends: study of the european union, 48-60. shekhar, vidhu & ramesh, sanjogita, central bank digital currency: perspectives on design choices and implications, with a focus on erupee, 61-78. academy of financial services officers shawn brayman financial planning research consultant president michelle cull western sydney university president-elect kirsten macdonald griffith university executive vp-program thanh ngo east carolina university vice president finance mustafa nourallah mid sweden university vice president international relations thomas korankye university of arizona vice president marketing & public relations elisabeth sinnewe queensland university of technology (qut) chair, australia-new zealand chapter editor, financial services review elisabeth sinnewe queensland university of technology (qut) directors norah feng massey university aaron gilbert auckland university of technology matt goren brett danko education center tom idzorek morningstar eun jin kwak university of wisconsin-green bay dan moisand moisand fitzgerald tamayo richard stebbins university of alabama tom warschauer san diego state university dave yeske golden gate university yu (yulia) zhang kansas state university past presidents shawn brayman, 2023-2025 financial planning research consultant inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 33, no. 3, 2025 editor elisabeth sinnewe, ph.d. queensland university of technology (qut) editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • conrad ciccotello, jd., ph.d., university of denver • john e. grable, ph.d., cfp®, university of georgia • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • shinae choi, ph.d., university of alabama • lu fan, ph.d., cfp®, university of georgia • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • philip gibson, ph.d., cfp®, winthrop university • stu heckman, ph.d., cfp®, texas tech university • stephen horan, ph.d., cfa®, university of north carolina wilmington • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • izidin el kalak, cardiff university, united kingdom • thomas langdon, ph.d., roger william university, bristol, ri • claire matthews, ph.d., massey university, new zealand • mustafa nourallah, centre for research on economic relations, mid sweden university, sweden • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • dan richards, ph.d., york university • chris robinson, ph.d., cfpretired™, cpa, ca, york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida • bomikazi zeka, ph.d., university of canberra, australia editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • axton betz-hamilton, ph.d., south dakota state university • alona bilokha, ph.d., university of north florida • brian l. boscaljon, ph.d., penn state behrend • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph..d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph..d. stetson university, deland, fl • patti fisher, ph.d., virginia tech • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d., the ohio state university • jennifer harrison, ph.d., southern cross university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • ronald mciver, university of south australia • barry mulholland, ph.d., cfp®, university of akron • david nanigian, ph.d., cfp®, mount ararat financial services llc • john nofsinger, ph.d., university of alaska anchorage • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • congrong ouyang, ph.d., kansas state university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • steffen westermann, ph.d., griffith university • tansel yilmazer, ph.d., cfp®, the ohio state university the editor of financial services review wishes to thank university of georgia for support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peer-reviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications pii: s1057-0810(97)90022-6 financial services review, 6(2): 109-124 copyright 0 1997 by jai press inc. issn: 1057-0810 all rights of reproduction in any form reserved. the use of professional designations in the real estate industry susan logan nelson and theron r. nelson a series of surveys investigates the recognition by consumers of professional designa tions in a variety offinancial services fields, but with an emphasis on the real estate industry. the results from the 1991, 1994 and 1996 surveys are consistent in revealing a general lack of recognition by consumers ofprofessional designations, with the nota ble exception of the cpa. the cfp and clu also fared somewhat better than the real estate designations examined both in terms of consumer recognition and perceived competence of those holding the designation. consumers indicate, however, a generally strong desire for dealing with professionally designated individuals when making financial decisions. i. introduction the real estate industry, particularly the brokerage and appraisal segments of the industry, has come under increasing scrutiny from both consumers and government in recent years. evidence which indicates a lack of satisfaction with, and confidence in, real estate service providers continues to build. in 1983, a gallup poll found that only 13% of those surveyed believed that real estate salespeople had high or very high ethical standards, while 28% believed that real estate salespeople had low or very low ethical standards. by 1988, those who believed that real estate salespersons had high or very high ethical standards had remained constant at 13%, while those who believed they had low or very low ethical stan dards had risen to 34% (filisko, 1989). a 1989 survey of almost 72,000 readers who had sold property between 1985 and 1989 was conducted by consumer reports magazine (do real estate.. , 1990) on a related topic. although most surveys of service providers conducted by the magazine indicate that about 10% of customers are left dissatisfied, the survey of real estate brokerage service users found that nearly one-third reported trouble of some type and about 20% were left susan logan nelson l professor of marketing, department of marketing, box 8366, college of business and public administration, university of north dakota, grand forks, nd 58202; e-mail: sunelson@badlands.nodak.edu. theron r. nelson l professor of finance, department of finance, box 7096, college of business and public administration, university of north dakota, grand forks, nd 58202, e-mail: thnelson@badlands.nodak.edu. 110 financial services review 6(2) 1997 dissatisfied-double the expected rate. these findings are consistent with a march 1989 survey conducted by the national association of realtors. the much smaller nar survey of 1,000 people who had recently worked with a real estate salesperson found that 82% were satisfied (filisko, 1989). such reports, when combined with research which has indicated quite clearly that training, education and professional competence are important to consumers in selecting brokers and salespersons (chamber, dotson, & dunlap, 1985; little & myers, 1991; nelson & nelson, 199 1; patton & peterson, 198 1) lend credibility to statements like that of a recent home seller that “[tlhere are good and bad real estate salespeople, and you have to choose the right salesperson when you put your house on the market” (filisko, 1989, p. 46). if we accept that the consuming public is wary of the real estate industry to some extent, then the process of communicating to clients that certain practitioners possess superior skills, training, experience, education and competence becomes critical. most professions which deal directly with consumers in providing a service have turned to the use of professional certification and designation programs for this purpose. ii. the certification and designation process the term “professional certification,” as recognized by the awarding of a professional des ignation, is often confused with the terms “accreditation” and “licensing.” although these terms are similar in that they all relate to standards of education, training, experience, and knowledge, there are distinct differences as summarized in table 1 (bratton & hildebrand, 1980, p. 23). in this paper the term designation is used to represent the completion of a profes sional certification program or process which then entitles the individual to advertise this fact through the use of a particular title which is abbreviated in use. the most common example of this is the cpa designation which is simply an abbreviation of certified public accountant. thus the term designation essentially becomes synonymous with the notion of certification. for certifications (or designations) there must be an association or agency which establishes the program, administers or coordinates the educational programs and testing procedures, and markets the designation. there are numerous associations, probably too many to count (davis, 1990, p. 40), which offer one or more designation programs. per haps the leading organization offering real estate designations is the national association of realtors (nar) and its many affiliates. technically, however, nar does not issue any designations. its most popular designation program, the gri is administered separately in each state. other designation programs are handled through the many institutes, societies and councils of nar. table 1 distinctions among professional credentials t.vpe of‘ crrdenricrl recipient of credenticrl credenticding body required or vo1untut-y accreditation programs licensure individuals association/agency political body voluntary required certification individuals association/agency voluntary professional designations 111 although a reported top executive of a national real estate organization describes des ignations of the “mail order” variety that “you can obtain by sending $50 to get a certificate and some initials to put behind your name” (davis, 1989, p. 1.57) there are usually specific requirements for obtaining professional designations. however, there are no specific stan dards which apply to the organizations sponsoring them. while nar and its affiliates, as well as several other real estate organizations, are well known and have solid reputations, the same is not always true for other organizations. the national commission for health certifying agencies (nchca) was created in 1978 at the urging of the u.s. department of health and human services to foster consistent testing guidelines in the health profes sions. the nchca became, in effect, an organization which certifies organizations that offer certifications to members. with the rapidly growing interest in, and popularity of pro fessional designations in fields outside the health area, the nchca created a sister organi zation, the national commission for professional and occupational certifying agencies (ncpoca), and placed it, with nchca, under the umbrella called the national organi zation for competency assurance (noca). thus there now exists an organization which establishes criteria for approving certifying agencies. noca even provides a package to assist organizations in developing certification programs. iii. goals of certification programs the rapid proliferation of certification programs in many fields is testimony to the popular ity of designations. why, one might ask, are designations so popular? what need do these programs satisfy in the market? commentators have suggested numerous criteria to be fol lowed in structuring certification programs. gilly and galbraith, in a survey of professional associations and/or societies which had implemented or substantially revised a certification program within the past 20 years, identify ten qualification criteria for attaining certifica tion (1988, p. 14). these are, in rank order of importance: 1. professional experience; 2. successful completion of a written examination; 3. completion of a program of study or desired number of years of education; 4. currently employed in the respective industry/trade/field; 5. successful completion of a performance examination; 6. membership in [the sponsoring] professional organization; 7. completion of additional training and/or continuing education; 8. evidence of ethical behavior; 9. personal and/or professional references; 10. completion of an oral evaluation and/or interview. in a survey designated to measure the need for, and desired characteristics of, profes sional certification of individuals with a professional specialty in marketing research, mcdaniel and solano-mendez identify a similar list of criteria for structuring a certifica tion program (1993, p. 25). in order of importance, these are: 112 financial services review 6(2) 1997 1. evidence of ethical behavior; 2. professional experience; 3. current employment in the field; 4. examination over knowledge in the field; 5. academic preparation/program of study; 6. personal and professional references; 7. performance examination; 8. membership in that professional organization; 9. oral examination (i.e, demonstration of skills in a simulated exercise). the importance of ethical behavior revealed in the mcdaniel and solano-mendez sur vey is reflected in most certification programs. an often stated belief is that “. . [a] certifi cation program without a strong code of ethics would be a disservice to our profession, our clients, and ourselves” (wukitsch, 1990b, p. 69). some researchers, however, assert that “certification could be misleading to a naive client. also, whereas certification cannot ensure ethical behavior, it could conceivably be used to mask unethical behavior” (ritten burg & mudock, 1994, p. 8). in many cases, however, adherence to a code of ethics is a condition of membership in the sponsoring organization and is not contingent upon completion of a certification pro gram. such is the case with nar. it is not necessary to complete, or even begin, any of the numerous certification programs available to retain membership in nar. thus adherence to nar’s code of ethics is not contingent upon obtaining a designation, nor do any of the designation programs carry more stringent ethical considerations. the fields offering certification programs are generally those which deal directly with the public in some capacity. in most instances the push for certification is strongest when the field is one in which poor performance offers the potential for harm to consumers. although this is most evident in the health professions, the potential for harm is also present in those fields that involve financial decisions. in some instances, the protection of the public is insured through legislative methods. this is certainly true in the field of real estate brokerage, and has become the case for real estate appraisal (lahey, ott, & lahey, 1993). the legal basis for licensing is, in fact, the need to protect the public welfare. licensing statutes are permitted as an exercise of the police power of the states. thus those fields which require licensing already have a mech anism to protect the public. therefore, the existence of certification programs suggests that the minimum standards of practice required under the licensing statutes are not sufficient to adequately protect consumers, or that the licensing statutes are not adequately enforced. iv. additional certification considerations the goals described above are those most often used to explain the need for creating or main taining certification programs. the need for certification programs when any or all of these goals are not being met is clear. in the field of real estate brokerage, and most related real estate fields, these goals are generally met through legislative requirements (licensing and continuing education) or a strong national organization with an appropriate code of ethics. professional designations 113 unless the public demands a standard of performance beyond that required by the combination of legislative and professional organization efforts, certification programs would appear to serve no real need. one of the major concerns voiced in the appraisal industry is “whether professional designations awarded by associations.. .will be as impor tant, because all an appraiser needs to work for a federal-related institution or agency is to hold a state certificate or license, not a designation” (davis, 1989, p. 154). there are, how ever, several additional factors which may explain the continued demand for designations. nancye kirk, staff vice president for the institute of real estate management, lists three “rewards of professional designations” (professional designations.. .1987, p. 65): 1. increased credibility and enhanced job opportunities; 2. professional contacts which accompany association membership; 3. enhanced earnings potential. glenn crellin, formerly vice president for economics and research at nar, points out two additional reasons for certification. first, the mandatory continuing education (mce) requirement found in real estate puts real estate agents back in the classroom. “once in the class, the motivation to stay to receive something tangible for the effort is expanded. in many cases that result becomes a license or a designation” (crellin, 1991). crellin also points to a variation of the “enhanced professional contacts” role mentioned by kirk. approximately 70-80% of nar’s membership operates in the general area of resi dential brokerage. thus, finding a peer who specializes often becomes a difficult process. for many, the most reliable method of locating specialists is to seek out those agents who hold an appropriate designation. galbraith and gilley (1985, p. 13) provide a summary list of “other purposes” for cer tification, including to: l promote professionalism; l enhance the prestige of the profession; l improve academic programs; l be an income producer; l distinguish individuals from peers and colleagues; l encourage individuals to remain in the profession; l avoid external governmental regulations; l stabilize individual’s job security; l protect clients/employees from incompetent practitioners; and l prevent cannibalism (the recruitment of individuals from other professional asso ciations in order to strengthen one’s own). these goals are obviously designed primarily with the welfare of the recipient of the designation in mind, not the user of the service. in similar fashion, gilley (1988) suggests a seven-step plan to “market” a certification program. interestingly, the marketing plan is i‘. . .focused on meeting the needs of certified or potentially certifiable individuals” (gilley, 1988, p. 110). it would appear that in many fields (gilley was speaking specifically about the interior design field), the concerns expressed by organizations do not relate in any 114 financial servkes review 6(2) 1997 meaningful way to the unmet needs of clients. rather they seem to be based, at least in part, on the need to generate revenue and other benefits associated with a successful certification program. because certification “can guarantee only that an individual meets certain knowl edge, education, and/or experience requirements, it cannot guarantee competence” (ritten burg & murdock, 1994, p. 7). some researchers argue that “certification would benefit only those doing the ~e~i~~ation, entrenched practitioners, and academics-not the pub lic...” (rittenburg & murdock, 1994, p. 8). shimberg (1984, p. 14) similarly cites the ability of certification credentials to ‘&. . .pro vide significant economic benefits to those who possess them.” one danger in such pro grams is the temptation by sponsoring organizations “. . .to restrict the number of certified individuals by increasing the difficulty of the tests or increasing other qualifications, thereby giving special benefits to those who are already certified.” the antit~st charge made under the sherman act is that such programs act as a “conspiracy to monopolize” trade in a particular market by creating membership restrictions for organizations or ser vices essential to competing on an equal basis. although frequently raised, this argument appears to have little, if any, validity. webster, citing internal revenue service and federal trade commission rulings and guidelines, finds that “[alntitrust violations need not be engendered if the program is educational in nature, is open to all, and is not designed to fix prices, exclude members, or allocate markets” (1989, p. 132). v. evaluating certifica~on programs the evaluation of certification programs may proceed from the criteria developed by noca involving the structure of both the sponsoring organization and the certification program. there may be compelling reasons for the existence of a particular certification program based solely, or almost so, on the basic goals of certification described earlier. the ce~i~~ation programs in appraisal, for example, have their origins in meeting these needs in a field which until recently lacked legislative efforts. in similar fashion, the accounting industry has remained almost solely self regulated through the use of the cpa program. the focus of this paper is on a field, real estate brokerage, which has long been char acterized by strong legislative efforts in mandatory education, licensing, and continuing education. such a field would not appear to be in need of independent certification pro grams. however, real estate brokerage boasts many certification programs that apply to individuals who must also comply with state licensing regulations and adhere to a compre hensive organizational code of ethics. thus, with basic certification goals already met, there remain only two reasons why such programs exist: 1. legislative efforts to protect the public are insufficient or are not enforced effectively. if, for example, a distinction exists between complaints filed by consumers against certified brokers (or salespersons) as opposed to noncertified brokers, then a case could be made on this basis. although there is very little, if any, research in this area, it would appear that such is not the case. the hawaii real estate commission, in preparing to ini tiate a mce program, did note that in anaiyzing its complaint files “lilt appears that licens professional designations 115 ees with nine or more years of experience had a disproportionately higher number of complaints than did licensees with less than nine years” (ordway & yee, 1989, p. 13). this study, like other similar efforts, was aimed at an evaluation of the need for mce and did not directly address the use of certification programs per se. it is implied, however, that since the certification programs had been available for a considerable time period prior to this finding, either licensees in hawaii did not pursue such programs in sufficient numbers, or the certification programs did not provide the additional public protection desired. unfortunately, since no research efforts have been aimed specifically at this question, only speculation is possible. 2. certification programs provide member and organization benefits. as discussed earlier, this reasoning seems to provide much of the basis for certifica tion programs in certain fields. since basic consumer protection needs have already been met through legislative efforts, real estate brokerage may fall in this category. if so, a cer tification program must meet only two criteria to be successful: l clients must express a desire for credentials and/or protection beyond those pro vided by basic licensure; and l clients must view those holding the designation as being more competent and/or more ethical than those who do not hold the designation. in this context, it is not even essential that members of the general public recognize the designation. recognition by the client group, however, is certainly a desirable feature. clearly, those seeking certification must believe that the program they have selected meets these criteria or that the supplemental benefits described earlier are sufficient. obtaining a designation, however, can be a difficult and costly process. obtaining the ccim designa tion, for example, is estimated to cost candidates a minimum of $10,000 to $15,000 (davis, 1993, p. 78). candidates continue to seek the designation, says steven f. pope, executive vice president of cirei, because they “believe that they will have a better chance to sur vive if they have the ccim designation” (davis 1993, p. 78). vi. real estate brokerage client surveys lack of recognition by the ultimate consumer group for real estate industry designation holders does not, by itself, completely eliminate the desire to obtain such certification. a poor level of recognition would5 however, cast into serious question the value in the mar ketplace of many real estate industry designations. this study undertakes an assessment of the client group for residential real estate brokerage with regard to familiarity with profes sional designations. the assessment consists of direct mail responses to three questions from a sample of potential real estate brokerage clients. the results of three such surveys are reported here. the first survey took place during 199 1 and was sent to a sample of u.s. homeowners. the second survey took place during 1994 and was sent to recent home buy ers. the third survey took place in 1996 and was also sent to recent home buyers. a differ ent source of mailing labels was obtained for the second and third surveys. these national 116 financial services review 6(2) 1997 lists consisted of names and addresses developed from public records of deed recordings for home sales which took place during january 1994 and january 1996, respectively. the purpose of each survey is to investigate three aspects of client familiarity with professional designations. the first probes the level of recognition of a set of common professional designations, including one nonexistent designation. the second concerns the perceived level of competence for individuals with a designation when compared to individuals who do not hold the designation. the third involves a series of events which commonly call for the use of a service provider who may hold a professional designation. respondents are asked to indicate how important it is that the provider hold an appropri ate professional designation (even if they do not know what the appropriate designation might be). a. the survey instrument the survey instrument consists of questions addressing the three elements described above, along with introductory material describing the nature of the survey and a few demographic questions. question number one provides the list of designations included in the study and asks respondents to indicate which type of business activity each is associ ated with. the list of designations, presented in alphabetic order, represents a cross-section of common nonhealth related professional business services. the designations, and the full title for each, are listed in table 2. the “type of business activity” information is based on the choices available to respondents. in addition to the business activities listed in table 2, two other choices are available. the only other business activity type listed is that of “marketing.” the nonexistent cmr designation, which could stand for “certified market researcher” if such a program existed, is provided for two reasons. first, to test the willingness of respondents to attribute a nonexistent designation to a type of business activity. second, to provide some early information on the potential need for a certification program in the field of marketing research, an area which has recently generated interest in such a program (mcdaniel & solano-mendez, 1993; phillips, 1989; rittenburg & murdock, 1994; schlossberg, 1989; wukitsch, 1990a). the other choice is a “do not recognize” option for designations not recognized by respondents. question number two, using the same set of designations from question one, asks respondents to rate the competence of someone who holds each of the designations when compared to an individual who does not have the designation but offers the service. table 2 designations used in surveys lkyi<pltion full title ccim certified commercial, investment member cfp certified financial planner clu certified life underwriter cmr n/a cpa certified public accountant crs certified residential specialist gri graduate, realtors institute ma1 member, appraisal institute type of businrss activity real estate (general) financial planning insurance nonexistent designation accounting real estate (general) real estate (general) real estate (appraisal) professional designations 117 respondents are asked to rate designated individuals as being “much more,“’ “equally,” or “less” competent than nondesignated individuals. the same “do not recognize” option is available for this question as well. the final question provides a list of personal and business situations. respondents are asked to indicate how important it is, on a six point scale from “extremely” (6) to “not at all” (1) important, that the individual providing the service holds an appropriate profes sional designation, even if the appropriate designation is not known to them right now. a “does not apply” field is provided for those instances where the respondent has never been in the described situation and does not expect to be in the future. the situations describe events that may call for professional services. providers of such services could hold one of the designations provided in questions one and two. thus, once again, health related services are omitted. b. the survey samples the client group for real estate brokerage services consists of those individuals who are likely to have used, or are currently using, such services. the client group for the 1991 survey was defined simply as home owners. to tighten the focus of the survey, and provide a more reliable indicator of the recognition level of real estate brokerage designations, the second and third survey client groups consisted of recent home buyers. national samples were acquired from commercial mailing list services for each survey. in each case, the sur vey instrument, along with a postage paid return mail envelope, was sent to individuals from these lists. the 1991 survey was sent to a total of 950 individuals (selected from a larger list to include 475 female and 47.5 male names). a total of 126 questionnaires were returned for a response rate of 13.3%. the second and third surveys were sent to national samples, each comprised of 2,000 individuals who had recently purchased a home. these lists are developed on a continuing basis from deed recording information in public table 3 demographic characteristics of sample respondents sex: male female i496 slilvry i994 sunvy /yy/ sunny 76.7% 64.0% 50.0% 23.3% 36.0% 50.0% age: total 18-24 25-34 35-49 50-64 65 and over 2.8% i .2% 0.8% 35.5% 35.2% 19.3% 43.3% 44.2% 33.1% 12.8% 12.7% 25.0% 5.6% 6.7% 21.8% 100.0% ioo.o% 100.0% total household income: under $10,000 $ i o,ooo-$19,999 $20,000-$34,999 $35,000-$49,999 $50,00@$74,999 $75,00@$99,999 $loo,ooo-$124,999 $125,000 and over total 100.0% 0.0% 1.7% 13.6% 18.2% 27.8% 14.2% 1 i .4% i3 1% i 100.0% 0.0% 1.9% 14.3% 24.8% 27.3% 19.9% 6.8% 5.0% i oo.07r 8.2% 6.6% 20.5% 27.0% 26.2% 6.6% 3.3% 1 6% b 100.0% 118 financial services review 6(2) 1997 records. the samples utilized for these surveys were from those who recorded deeds during january 1994 and january 1996, respectively. a total of 165 completed questionnaires were received from the 1994 survey for a response rate of 8.3%. in an effort to improve the response rate, a small gift was included in the 1996 survey mailing. the “gift” was a wallet card with information of how to calculate the “wind chill factor.” smce the mailing came from the university of north dakota, and was mailed in january, it was hoped the card would gain the recipient’s attention and lead to an increased willingness to complete and return the survey instrument. however, the 1996 survey resulted in the return of 18 1 com pleted questionnaires for a response rate of only 9.1%. table 3 summarizes the demographic characteristics of the three survey respondent groups. due to the selection process, the 1991 group is equally split between males and females. the 1994 and 1996 groups are based upon the names which appear in the public record and tend to be more heavily dominated by males. these 1994 and 1996 survey respondents are also younger and have slightly higher household income levels. the age differences are consistent with the sample characteristics, since younger families are some what more likely to be in the housing market (as opposed to all home owners). c. survey results tables 4 through 7 summarize responses to the survey questions. although it should come as no surprise, the cpa designation dominates in terms of recognition and perceived competence. as seen in table 4, the cpa designation was recognized by all survey respon dents in 1994, and by virtually all respondents in 1991 and 1996. with one exception, all respondents also correctly categorized the cpa as an accounting designation. the next most recognized designations are the cfp and the clu. recognition of the cfp designa tion appears to be increasing, while that of the clu is declining somewhat. with the exception of the ccim designation, the other five designations in the study (including the nonexistent cmr designation) show similar levels of recognition by the respondents throughout the five-year period. the ccim designation, already the least recognized of those included, well below even that of the nonexistent cmr designation, showed a marked decrease in recognition during the study period. a very similar pattern can be seen in table 5 which summarize responses to the ques tion regarding the competence of designation holders as compared to those offering the same service but who do not hold the designation. once again, the cpa designation dom inates both in terms of recognition and perceived competence. as seen in table 6, the cpa ranks first in all surveys by a wide margin in terms of those who perceive holders as being “much more competent” or “somewhat more competent” than those without the designa tion. the cfp designation moved up one slot to second from 1991 to 1994, changing places with the clu designation. these are followed by the gri, ma1 and crs designa tions. the ccim designation finished last in all surveys, slightly below the nonexistent cmr designation, although the cr1 remained in fourth place from 1991 to 1996, the percentage of respondents who rated holders as more competent dropped noticeably from 26.2% to 18.0% to 16.4%. smaller drops can be seen for the mai, crs and ccim designations. the cfp designation has increased noticeably, while the clu has dropped by the same magni tude. these two designations have essentially traded places in term of perceived compe tence since the first survey in 199 1. p 3 1. 8 f t a b l e 4 f % c on su m er r ec og ni tio n of p ro fe ss io na l d es ig na tio ns pe rc en t se le ct in g c at eg or y c u m c f p c l u c m r c p a c r s g r i m a i 19 96 19 94 i9 91 19 96 19 94 19 91 19 96 19 94 19 91 19 96 19 94 19 91 19 96 19 94 i9 91 19 96 19 94 19 91 19 96 19 94 19 91 19 96 19 94 19 91 a cc ou nt in g 0. 5 0. 5 1. 6 99 .5 10 0. 0 99 .2 0. 5 0. 6 fi na nc ia l pl an ni ng 1. 8 1. 6 56 .4 53 .0 52 .5 1. 1 2. 4 1. 6 0. 8 0. 8 0. 5 0. 6 in su ra nc e 1. 7 1. 2 3. 2 0. 5 0. 8 39 .8 40 .6 48 .4 2. 8 3. 2 2. 2 1. 2 1. 6 m ar ke tin g 1. 1 2. 4 6. 3 0. 6 14 .7 14 .6 16 .8 0. 6 1. 6 1. 1 3. 9 4. 8 3. 2 r e a pp ra is al i.7 1. 2 1. 2 0. 8 2. 3 1. 2 2. 4 5. 6 4. 9 4. 7 6. 6 4. 2 7. 1 12 .2 15 .8 14 .3 r e g en er al 2. 2 4. 3 3. 2 2. 2 0. 6 1. 1 1. 8 0. 8 19 .0 18 .4 15 .1 18 .8 24 .2 21 .4 0. 6 0. 8 d o n ot r ec og ni ze 92 .2 89 ,l 85 .7 43 .1 47 .0 52 .4 56 .4 54 .6 47 .6 81 .9 82 .4 80 .0 0. 5 75 .4 76 .1 77 .8 70 .2 70 .9 68 .3 81 .2 77 .0 80 .0 t a b l e 5 pe rc ei ve d c om pe te nc e of p ro fe ss io na t d es ig na tio n h ol de rs pe rc en t se le ct in g c at eg or y c c /m c f f ’ c l 0 c m r c p a c r .5 g r i m a i 19 96 19 94 19 91 19 96 i9 94 19 91 19 96 19 94 19 91 19 96 19 94 19 91 19 96 19 94 19 91 i9 96 19 94 19 91 19 96 19 94 19 91 19 96 19 94 19 91 m uc h m or e 1. 7 1. 8 4. 8 8. 3 11 .8 ii .3 7. 8 9. 9 13 .6 1. 2 1. 2 2. 4 53 .9 56 .1 64 .2 2. 3 3. 1 3. 2 3. 4 2. 5 8. 7 6. 2 6. 8 5. 5 c om pe te nt so m ew ha t m or e 3. 3 4. 3 5. 5 33 .0 28 .6 22 .6 25 .7 25 .3 27 .2 7. 5 8. 0 8. 1 41 .0 37 .2 29 .3 10 .8 11 .2 11 .2 13 .0 15 .5 17 s 7. 3 7. 4 11 .1 c om pe te nt e qu al ly c om pe te nt 1. 7 3. 7 1. 6 14 .0 9. 9 10 .5 8. 9 8. 6 9. 6 7. 5 7. 4 5. 7 3. 4 4. 9 6. 5 9. 6 7. 5 6. 4 11 .3 8. 7 5. 6 2. 8 6. 8 4. 0 l es s c om pe te nt 0. 8 1. 1 1. 9 1. 6 0. 4 0. 6 0. 8 0. 6 1. 1 1. 8 0. 6 1. 2 1. 6 0. 6 0. 6 0. 6 0. 6 d o n ot r ec og ni ze 93 .3 90 .2 87 .3 43 .6 47 .8 57 .0 57 .0 55 .6 48 .8 83 .8 82 .8 83 .8 0. 6 16 .7 77 .0 17 .6 71 .7 72 .1 68 .2 83 .1 78 .4 19 .4 . professional designations 121 table 6 designation holders ranked as “much’ or “somewhat more competent” percent selecting rmk desigwztion 1996 1994 1991 1996 1994 1991 cpa 94.9 93.3 93.5 i i i cfp 41.3 40.4 33.9 2 2 3 clu 33.5 35.2 40.8 3 3 2 gri 16.4 18.0 26.2 4 4 4 mai 13.6 14.2 16.6 5 6 5 crs 13.1 14.3 14.4 6 5 6 cmr 8.7 9.2 10.5 7 7 7 ccim 5.0 6.1 10.3 8 8 8 table 7 summarizes the responses to question three which investigates the perceived importance of utilizing the services of an individual holding a professional designation in a variety of situations. importance is measured on a six point scale with six being “extremely” important and one being “not at all” important. the mean importance rat ings for this table were computed without including the “does not apply” responses. thus the number of responses utilized to compute the mean varies somewhat between situations. situations most likely to involve the use of an accountant or financial planner have the highest importance ratings. several situations related to real estate brokerage, however, exhibit very high importance ratings as well. note the importance placed on the use of a designated professional during the home purchase process, which is ranked almost identi cally to the home selling process. furthermore, obtaining an appraisal for a home purchase table 7 mean responses in 1994 survey results rank order mean rating does not app1.v (peweat selecting) 1996 1994 1991 1996 1994 i991 “how important is it that the person you deal with has ~1 profes.sioncrl designation when you cue: ” preparing a personal 4.90 5.06 4.78 4.4 2.4 7.1 income tax return investing in stocks and 4.85 4.95 4.69 3.3 7.3 7.9 bonds obtaining an appraisal for 4.63 4.82 5.16 2.2 2.4 4.8 a home purchase planning for retirement investing in real estate buying a home selling a home buying life insurance buying home owners’ insurance buying car insurance being asked to reveal information about 4.73 4.74 4.76 3.3 4.2 4.8 4.41 4.58 4.71 10.5 3.6 7.9 4.26 4.40 4.5 i i.1 1.8 4.0 4.23 4.42 4.41 2.8 3.6 4.0 3.90 3.99 4.05 5.0 3.0 6.4 3.70 3.89 4.19 3.9 2.4 4.8 3.31 3.51 3.83 2.2 1.8 3.2 2.97 3.32 2.32 7.7 9.7 8.8 yourself in a survey* nore: *wording changed between 199 i and lyy4. 122 financial services review 6(2) 1997 and investing in real estate both garnered even higher designation importance ratings than did the brokerage aspect of the industry. vii. conclusions and recommendations several conclusions can be drawn from the survey results. survey respondents, who are likely clients of financial professionals, attach reasonably high level of importance to selecting individuals holding designations when engaged in a variety of financial decision making situations. unfortunately, most consumers are not able to recognize common pro fessional designations and do not see holders of these designations as being a great deal more competent than those without the designations. the general conclusion that can be drawn from this pattern is that consumers perceive a need for professional services, partic ularly financial services, and feel that using a service provider with an appropriate designa tion is important, but have very little idea what an appropriate designation may be. the real estate industry offers its members a wide variety of professional designation programs, as does the financial services industry in general. in contrast, the dominant pro fessional designation is the cpa-a single designation covering a wide variety of specialty areas. the accounting industry is now debating the merits of adding specialty designations (elsea & donelan, 199 1; shambo & eveloff, 1993) and has already approved the “personal financial specialist” (pfs) designation. although the circumstances are a bit different in that the accounting specialty designations would be in addition to the cpa (i.e., only cpa holders would be eligible), some parallels do exist. those who favor additional specializa tion in accounting point out that the medical profession has 23 specialties and the legal pro fession has up to 12 specialties in some states (elsea & donelan, 1991, p. 59). opponents warn of potential “mass confusion both inside and outside the profession” and note that such specialty designations would “overlap designations already offered by other recog nized professional bodies” (shambo & eveloff, 1993, p. 43). proponents of accounting specialization note that “cpas naturally should stand above the crowd of financial services providers” but that the “public has challenged this presumption, however, and has sought advice outside the profession” (shambo & eveloff, 1993, p. 42). this latter argument appears to approach the issue from the viewpoint of the needs of the organization and its members rather than the needs of the consumer. as the accounting profession begins to move toward potential specialization designa tions, the real estate industry should move toward a single, highly recognizable, designa tion which can serve as an “umbrella” designation, much as the cpa designation now does. the financial services industry, although offering numerous professional designation pro grams, is in somewhat better shape in terms of consumer recognition and perceived com petence than is the real estate industry due to the cfp and clu programs. the real estate industry has a significant image problem in many areas-particularly brokerage and appraisal. the creation of a single designation program, with an appropriate level of rigor, based on the cpa and/or cfp model could do much to improve the perceived level of com petence and ethical behavior in the real estate industry. this designation must be broad in scope, require extensive education, training, and experience and be heavily promoted to the public. although a new organization could be created to develop and sponsor this designa professional designations 123 tion, a cooperative program by existing organizations (perhaps with the leadership of nar) may be more appropriate. existing real estate designations could, and probably should, be retained in many instances as specialty areas for those holding the new umbrella designation. this would allow the real estate industry to recognize the legitimate specialty areas which exist, while addressing the problems facing the industry. current designation programs are highly inef fective, at least in terms of recognition by consumers. the demand for financial services, and the desire by clients to utilize the services of persons holding appropriate designations, appears to be reasonably high. the financial ser vices industry, and in particular the real estate brokerage industry, should take note of the current public confusion and ignorance regarding certification programs and take affirma tive steps to improve the image of the industry. references bratton, b., & hildebrand, m. (1980). plain talk about professional certification. znstructional znno vutor, december, 22-49. chambers, t., dotson, m.j., & dunlap, b.j. (1985). consumer perceptions and expectations of real estate practices: an empirical investigation. in d.m. klein kc a.e. smith (eds.), pro ceedings (pp. 70-73). the southern marketing association and florida atlantic university. crellin, g.e. (1991, july 15). letter to authors regarding professional designations offered by nar. davis, j.c. (1989). associations equate knowledge with power, profits; training, experience required to earn designations. national real estate investor, 31 (october), 152. davis, j.c. (1993). cirei emphasizes industry education as a means for success in the 1990s. national real estate investor, 35 (february), 78-80. davis, n.m. (1990). who certifies the certifiers? association management, 42 (february), 38-43. do real estate agents help? (1990). consumer reports, july, 460-464. elsea, j.e., & donelan, j.g. (1991). accredited specialties: what now? the practical accountant, july, 54-60. filisko, g. (1989). salespeople work to clear up image misconceptions. real estate today, 22 (august), 42-47. galbraith, m.w., & gilley, j.w. (1985). an examination of professional certification. lifelong learning: an omnibus of practice and research, 9(2), 1215. gilley, j.w. (1988). marketing your professional certification program. association management, 40 (november), 110-113. gilley, j.w., & galbraith, m.w. (1988). commonalities and characteristics of professional certifica tion: implications for adult education. lifelong learning: an omnibus of practice and research, 12 (l), 11-14. lahey, k.e., ott, d.m., & lahey, v.m. (1993). survey of the effects of state certification on appraisers. the appraisal journal, july, 405-413. little, m.w., & myers, t.a. (1991). an assessment of homebuyer and real estute broker attitudes toward service quality. unpublished paper, department of finance and marketing, virginia commonwealth university. mcdaniel, s.w., & solano-mendez, r. (1993). should marketing researchers be certified? journal of advertising research, july/august, 20-3 1. nelson, s. logan, & nelson, t.r. (1991). a model for real estate brokerage firm selection. the real estate educators association journal, spring, 64-71. 124 financial services review 6(2) 1997 ordway, n., & yee, c. (1989). implementing a new continuing education program for real estate licensees: a case study in hawaii. real estate educators association journal, spring, l-36. patton, w.e., iii, & peterson, a. (1981). a comparative study of attribute importance in the retail image of realtors. in r.h. ross, f.b. kraft, & c.h. davis (eds.), 1981 southwestern market ing proceedings (pp. 154-157). wichita, ks: the southwestern marketing association. phillips, c. (1989). states must license researchers. marketing news, 23 (september 1 l), 27. professional designations: a career boost. (1987). buildings, november, 65-67. rittenburg, t.l., & murdock, g.w. (1994). highly sensitive issue still sparks controversy within the industry. marketing research: a magazine of management and applications, 6 (2), s-10. schlossberg, h. (1989). consensus eludes certification issue. marketing news, 23 (september 1 l), 1. shambo, j., & eveloff, s.h. (1993). is it a good idea to accredit specialists? journal of accountancy, april, 41-44. shimberg, b. (1984). the need for continued competence. journal of real estate education, 3 (1) summer, 10-14. webster, g.d. (1989). certification poses few risks. association management, 41 (october), 132. wukitsch, m.r. (1990a). what about certification? is it needed and is it wanted? marketing research: a magazine of management and applications, 2 (march), 73-75. wukitsch, m.r. (1990b). does ethics fit into the certification concept? marketing research: a mag azine of management and applications, 2 (june), 68-69. pii: s1057-0810(02)00101-4 �������� ��� � ��� ���� � �� � ���� � ��� ������� �������� �� �� �� ���� ����� ���� �������������� �� ���� ������� � �� �������� ���� ��� � ������� ����� �� ��������� �� � ����� �� �� ���� ����� �� � �� �� �� �� � !���"� #$% &'$ ���� (�"���"��� ������� ��" (��� ���� �������� )(�(�* !�� �� �� �� �� �� � !���"� #$% &'$ �� ����� !" #�� � ��� $%%!& �� ����� �� ������� ��� !" '� � ��� $%%!& � �� �� ( )�� ��� $%%$ �������� � �� � ��� ��� �������� ��� � ��� ��� �� ���� ��� � �� � �� ������ ����� �� ���� �* ��� �� + ,���� ���� ��� ������ �� �� � �-� �� �� ���� ���� � �� �� ��� � ��+ ,��� ����� �� ��� �� � �� �� ��� �� �� . �� ������ �� � �� �� ����� �� �� �� � �� ��� � /��� � ������ �� /�+ 0� �� � �� �� � ������ �* �� ���� ���� ��� ������� �� �� ����+ 0� �� � �� �� �� �� �� ��/� �� �������� ��� � ��� ��� � ������ � ��/� �� �� � ��� ���� ���� � � ����1���� �� �� � �� �������� �� � ���� !% ����� ��� ��� �� ������ �� ���� ��������+ 2�� ��/����� ��� ������� �� �* �� ����� �� ������� � ���������� ����1�� ����� � " �����+ � $%%! � ������� �� 3������� # ��� � 2� + %+, ���������� ���)$4& 5!! ���.� "�2�� ��� �& 6���� ��� � ��� � �& ��� ������� �� �� ���& ,�*� ���& �� ��� �� ������� � ���� ������ � �� � ��� ��� �������� ��� � ��� ��� ���� �� ���� ������� ��� ����� �� ��� � �� � �� ������ ����� �� ���� �* ��� �� + 0���� �� ��� �� �� ��� �� ���� � � ��� ����� �� ������� �� �� �� ���7/�� � ��� ����� ����� � �� ��� �� �����7 �� �� �� �� �*�� �� ����� �� �������� �� � �� � /� � �*��������� ����� �� �����+ ,�� ���� ��� ������ �� ��� ���� � �� � ���� �� �� � �-� �� �� ���� ���� � �� �� ��� � �� ��� � � � ��� ������� �� � � �� ��� � ���� ��������� ����� �� ����1���� ����� �� �*������ ��� ����� �� �� ���� �� /��� �� ���� ��� ��� ��� ��� �� �* �� ��+ 8�/����� /� ������� �� �� �� �� �*�� ��� ���� � �� /��� �� ��������� ��� �� �� ��� �� ����. ���9�� �������� ��� � ��� ��� �������� 6���� ��� #���� �� �����/ !% :$%%!; !<"=!4! !%"(�%>!%?%!?@ = ��� ���� � �� � $%%! � ������� �� 3������� # ��� � 2� + �22� # ! % " ( � % > ! % : % $ ; % % ! % ! � < � ������������� � ���+ ,��+� �!�<!4�(a4�$!%%& ��*� �!�<!4�(a4�"<>(+ +/���� �"" ����� �������b���� + � :�+�+ ��������;� ��������b���+���� + � :�+ ������� � �;+ �� � ��� � �� �� � � � � �� ��� � /��� ����� ���� � �������� � � ���� �*���� �� � ���+ 0 � �� �� ��� /� �� ���� �� �� � ����� ��� � �� �� ���7 ���� ����� �� � ��� � ��� ���/ � �� �7/� �� �� ���� �� ���� ��� � ������� � �� ��� ���� ���� � �� � ���+ c ��� �� �� �� ���+ ,�� ��������� � �� �� ����� �� ���� ����/� �� ������ �� �� �� � �� /� � � d���� �* �� ���e � ������ ���� ������+ ,�� � �� ��� ����� �� �� ����� ������ � �� ���� ���� ��� � � ��/�� �*������� ��� ��� ��� ���� ������ ������ � ������� �����+ �� ��� �� ��� ��� �� � �* ��� ��� /�� � � ���� �� � �� �� �� ��� �� �� � ��+ 2������ �� �� �� ���� �� ���� � ���� �� $%%%=$%%! ������ ��� ������ �� � �*����� ������� � ��� �+ ,��� ��� �� � �� ��� ���� �� ������ � ������ /� ��� � �������� ��� � � :��;+ #�� �� ����� � �� ��������� �������� �� � � ��� �� �� ������� /�� � �� �� �� �� ������ �������� �� ���� ������ ������� �� �������� �� �� /��� �� ������ ��� � � ������� ������ �� �� ����� �� "=!%f+ 6 � ��� ���� ��� d���� �* �� ���e � �� ������ ������ � ��� ����� /� �� �� ���� ��� ��� ���� /��� � ������ �������� ��� � ��� � � ��� ����� �� ������� �� �� �� ��� �� /�� � ������ ���� ���� �� � ��� � �� �* ������� � �� /������� �� /� � ������ �*���9 ��� ����* � ���+ 2������ � ���� ��� ����� ���� ����� �� �� ��� �� �� �* �� ��� ���� ���� ���� �� �� ����� �� ���/ � �� � �� �� ���� �� ���� ����� � ����+ 0� /��� �� ��� �� � �� � �� � �� �� �1��� /���� �� �*�� �� �� ��� �* �� �� �� � �������� ��� � � �� ������� �� �� ��� �* �� ��� ��� � �������� ��� � � ��� ��� ���� ��� ��� �� �� ��� �* �� ��� ��� � �� � ��+ ,��������� ��� ��� ��� �� ������� �� � �� ������ �*�� �� :�� ��� �*; �� �� ��� �� �������� ��� � � �� ��� �� � �� �� �� ������ ���� ��� �+ g�� ��������� ��. ���� � ���� ��� �� � ����� ��� ������� � ���� �� /� �� ���� ����+ #� ������7�� �� �� /� �� �� � �� � ��� ���� ��7�� �� ������ �� �� ����� �� �� �� �� :����� ; � ��� ��������� � �� ������ � ������ � ������� �� ���� �� ����� �� � ����� � ��� �� ��� �� ����� � � ����1��� ��� �� �������� � � ��� � � � �� � � ����1�� ��� �� ���� �� �� ���� �� � �� � �������+ 2� � ��� /����� ���� �� �� ������ ��� � "���� �" !%����� ����1��� �� ��������� � �� ����� �� ��� ��� �� ��� �� ��� ��� �� ������ �� � � ��� � � �+ ,��� �� ���� ����� � ��� �� ����� �� !%=!"f ��� �� ����� ��� �� �� �� ��� � �� �� � ��� � �� ��� ��� ����� ��� �������� ��� � ���+ ,��������� � � � � ��� ��� ��� � ��� � �� ��� �� /� �� ��� � � �������� �� � ����� + h� ���� /� � ���� ����. ���9�� �������� ��� � ��� � �� � ��� �� � � ��� �� �� ��� � �. ������ �� ��� � :�3i;+ ,��� � �������� /��� �� �*������� �� �� ���� � �� � ���� ��� �� �� � ��� � � �� ��� �� /��� � ������ �� /������ � ��� � ���� �� ����� � ���� �������� ����� �� � �� ����� �� ���� � ������+ ����� ��� ������� :!jj$; ��� � �� ���� � ��� � �����1� �� �� � ��� ����9 � ��� ������� �� ���������+ 2� � � ��� �/���� � �������� ��� � � �� ��������� � � �� � ��7 ���� ����� �� � ��� � ��� ����� �� ����1��7���� �� � � ��� ��� � 3. ������ �� i ��� � �� ���� ��+ ,��� � ��� �� ������� ����� �� ���� ���� �� . � � � �� �� �� 9��� � ��� � ��� � � �� ��� ��� ��� �� �� �� �� � /��� �� ��� �� ��� ,��� :!jj";� ,��� �� :!jja;� k��� :!jj$;� ��� 8���� ��� #�� :!j>>;+ ����� � � ���. �� ������ � � ��� ��� � �� ������� ��� �� �� �� �� �� /� 9�� �� ��/� �� �������� ��� � ��� ��� ������ � ������ � ��/� �� �� � ��� ��� ����� ��� /� � � ����� � � ����1��� �� �� � �� �������� �� � ���� !% ����� ��� ��� �� ������ �� ���� ��������+ !<4 # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& 2�� ��/����� /� ������ �� � ������ ���� �* �� ����� �� ������� � ���������� ����1�� ����� � " �����+ 0� �� � ���1� �� ��� ������ �� ��� ����� /� � ��� ��� �� �+ ��� ���� ��� � �� ����� ��� � � �� ��� ����� �� 1 ���/���� �� �9��+ ,�� �� ������ �� ��� ����� �� ������1�� �� �����/�+ 2� #� ��� $� /� �����/ �� � �� �� �*�� ��� � ��� � ��� ��� � ����� �� ��� �� �� �� ��� �� d���� �* �� ����e ��� �� ��� � �� /��� �������� ��� � ��� ��� �� � ���+ #� ��� a �������� �� �� �� �� �� �3i ��� ��/ � �� �� ��� � � ���� ��� �� ��� ����� �� ���� ����+ #� ��� < ���� �� � � ���� � ��� ������ � � �� �� ���� � ���� �� �� �������� �� /����������� ��� � ��� ��� �� � ���+ ,��� ��� �/��� ������ � ��9� ��� ������ � �*�� ��� ����� �� /��� �� � �� �� �� ���� �� ��� �� � � ��� � �� �� ��� ����� /�� � �� ������ �� �� #� ��� "+ 0� ��� ������ � � � � �� � �� ��� ��� ��� �� �� ��� �* /��� � � �� �� ���� ����1��� ��� ��� �� �� ��� �� � ��� �� �������� ��� � ��� ���� ���� � ������+ ,�� �� � �� ����� �� �� ��� �� � ���1� ��� �*� � ��� � �� ����� � ��� �� ��� ��������+ 0� ��� ���� �� ��� ��� ���7��� �� ��� � � �� ��� ��7�� � � ��� ���� �3i+ ,��� ����/� � � �� �� ��� /�� � �� �� /� � �� ����� �� d�� ��e �� �� �� � �� � �� ��� � � �� ��� �� �+ �� � ������� �� �� �� �� ���� /� �� ���� ������� � ���� ��� �� �� �������� �� /��� �������� ��� � ��� ��� �� � ���+ ��� ����� ��� ��� �� 9��� ��� �� � ���� ��� �� ��� ��������� �� � �� � ��� ����� � ��� � �� �� ���� � �� ���?�� � � � �� �� ���� ��� ��� �� ������� ���� �� ���� �� � ���+ #� ��� 4 �� � ��� �� ����� /� � � � ��� �� �� ��� ������ �+ � ��� �������� ���������� h���. ���9�� �������� ��� � ��� ���� ����� �� �� � � �������������� ���� �� �� ���� �� ������ �� �� �� � ��� ��� � � ������ ��� �����+ ��/��� 8�/���� ��� ��/�� :$%%!; �������� � ��� ����� �� �� ����9 � ��� ������� ���� ���� �+ 8�/����� ��� ���/ � ��� �� ���� /� �� �� �������+ 6���� ��� � �� � ��� ��� ��� � ������ ����� �� � �1�� � � �� �� ��� � 0� ���� �� ������� :�0�� !jj(; �� �� � �� �� �� !jj( ,�*������ ������ � �� �� ���� ��� ������ �� �� �� �� �+ 5��� :!jj>; ��� ����� :!jj>; ������� �� � �� �� �� ��� � ���� �� /�� �� � �� �0� ��������+ 6 � ��� ���� � ����� � �� ���� � ��� ��/� � � ���� ���� 9��� ������ ���� ��� ��� � � ������ �� ���� ��� �� � /� ���� ��� � ���� ��� �*�. ���9�� �����+ 2� ���� ��� � ������ ���� �� ��� � ��� ����������� � ����� ��� � ��� ��� �� �� �� � �� ��� �� �� ���� ���� �� �� ������ ����� �� �� ��� � ��� �*����� :��� ��� �; ����+ ����� :$%%%; ������� � ��� ����� �� �� � ������ ��� ���� � �� ���� ���� ���� � �� ��� ��� �� �� ��� ���+ 2� ��������� �� � ������� ���� �� ��� �� ���� � /� � ��� ��� � ���� �� �� ����� �� !%=a% ����� ���� � ��� ����+ ,��� �� �� �� ��� � �� ��� ��� �� ��� � �� ������� �� !$%=!a% ����� ���� � �� ���� ����� ��� ��� � ����+ 2������ �� �� ���� �� ��� /� � �� �� � � �� �� �� ����� ����� ��� �� ���� �� � � � ������� � /� � �� /��������� ���������� � �� ������� ��/� �� ���� �+ 2� �� � �� �*�� ��� �� ��� ��� ��� ���� ��� :$%%%; ��� ,������ :!jj!; ���� �� � ���� �� ������ /��� �� �����1��� �� ���� �� � � ����1���� ����� �� �� ���� ����� ���� ��� ������� �* �� �� �� �� ���� ��� ������ �� �������� ��� � ��� � ����� � �� � ���+ c� � ������ �� � �� �� � ���� ����� � ��� ��� �� �� /��� �� ��� ��� ��� ������ ��� � � ����1��� ��� ������+ k���/���� �� ���� �� �� ��� �� ������� �* ��� �� �� ���� ��� # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !<( ��� ��� ��� �� � �� ��� ��������� �� �� �� � ��+ ���� ������� ���� �� � ����� ������ � /� � ��� � ����� /���� �� �� � ��+ 8�/����� �� �� � �� ��� � ��� �� ��� �� ��� �� �� ���� /�� � �������� ��� ������ � /��� �� ��� ����� �� �� � ��� /��� � � ����1��� ��� ������+ +��� �� �� ���2/ � � �������9�" � � �� � � � � �� �:�� � �� �;�� �� &5<� �� �� � � �� .���� ���� � � �� � � �������" �� :�� � ���� �� � �� � �; ���� �� ���.+ 6�� � � �������� �� ������ ���� � � /��� �� �� ���� �� ����� � �� ����� ����� �� c ����� ��� ����� :!j>%; �� /�� � �� �� ��� � ���. �� /��� ��� � �����1� �� �� � ��� ����9 � �� ��� � ��� �� ��� �� ������� �����+ ���� �� �� ��� ,�� ������� �� ���� :!jj(; ��� '���� :!jj4; ���� ��� �� �� ���� � �*� ��� �� �� ��� �* ������ ��� ���+ 8�/����� ��� ��� �� ���� ����� � �� ���� ������ � �� ���� ��� � ���� ��� ���� �������� ��� ��� �� ��� �����1� �� � ������ d���� �* �� ���e � �� ���� �� ������ :!j><; �� '� ��� ' �� ��� #�� :!j>j;+ ,��� ��� � ��� �� ��� ����/� �� ����� �� /� � �� �� �� ��� � �����1� �� ��� ������� ��� ������� ���� ������ �� ����� � ��� � ���� ������� �* ������� �+ # � �� �������� ��� /��� ����� ��� �� � ��� �*���� �� � ���� ��� �� ������ ��� ������+ ,��� �� �� �� /� � �������� ��� � ���+ # � :!jj(; ������� �� �� �� �� d���� �* �� ����e � �� ��� � �� ����� <%!:�; �� ��� ����+ k�/����� ��� �� �� :!j>>; ������ �� �*� � ��� �� ���� ���� �� �� /� � �� ����� �� ��� ��� ����� ��� � �* ����� ���� ��� ������ �� ��������� ����� ��� ����� ������+ '� ���� ��� #����� :!jj<; �*� ���� �� ���� �� �* �� ���� �� �� ��� �* �� ��� � ������������ /� ��� �� �. � � �� � ��+ ,��� ����� �*� ���� �� ��������� � �� ������� ����� ��� ��������� # ������ ��� ����e� "%% ����* � ���� /�� � ����� �� ��� �� ����� �����1� ����+ ���� � c������ ��� l� :$%%!; ��� � �� �� �� � �� �� ��� ����9 � ��� �� ����������� ��� � �� ��� /� ��� � � ��� �*���� ��� �����+ ������ ��� �� /��� �� ��� ��� � � �� :!jj(;� � ������� :!jj<; �� /��� �� �� ����� ���� �� # ����� ��� 0������ :!jj$; ��� ���� ��� ������ � �� �� �� ������ �� /��� �*�� ��� ����� �� �� ���+ � ��� ��������� ���������� �� ������� ���� i���� �� ��� �� �� ���� ���� ��� ����� �� ��� �� ����� �������� ��� �� ��� � ��� �� ��� ����� ��� ��/��� ������ ��� /��� � � ����� � ��� � � � ��� �;�� � ��� �� �� ��� � �����1� ����� ���e ��� �� ���� � �� ��� �+ i���� ���� ��� � ����� ����� ��� /��� �*� �1� ���� �*�� �� ��� � ��� /� �� �� �*�� �� ��� � � ���� ����� �� ���� ����+ # � �� �������� ��� �� �� �*�� �� ��� � �� �� ����� � � � = :����� �� ����� �� ��� �* �� �� ��� � �������� ��� � �; �� ���� �� ��� �� �*�� �� ��� � �� �� ����� � � � � :����� �� ����� �� ��� �* �� �� ��� � �� � ��;� ��� = �� ��������� � �+ 2� �������� �� �� /� /�� � �� +���=�� � +������� /��� ����� ��� ������ = � �+ i���� �� ��� � �� �� � /� �� �� ������ ��9�� ���� �� ���� ��� � ��� �� ��� � �. ������ � ��� ��� ����� ��� � �� ���� ������� �� /��� �� ��� � � � ��� ��� ����� ��e� ���� ��������+ ,�� �3i �� ��� ��� �� ������� ��� �� � ��� � �� ��� �� � � � �� ���� +���>�� � ���3i�+ 2� ����� 3������� �3i �������� � �� � � � �� ���� �� � �� ����� ��� �� /� �� �� /������ � ��� ��� � ����� ����� � � �+ 2� ��� �� �� �� �3i �������� �� �� � ����� �� ��� � : � ��� � �� ���� ��� ��� ����� ��; �� �� ����� � � � >+ !<> # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& #�� �+�+� > ��� �������� �� ����� ��� ��� � � � ��� � �� ��� � �� � � � � ��� ����� �� � ���� � �� �����+ ,�� ��������� . �� � � �3i /� �� �������� �� � �� ���� �� � ��� �� /� �� �� /������ � ����� �� �� ����� �� �* ����� ��� >+ 2 �� � � ��� ��� � ������ �� ���� ���� �� ������ �� /� �� �������� ���� � ������ �������� � �� + h� ������ �� �3i �� ��� ����� > �� �������� �� � ������e� ���� �������� �� ��� �� ��+ #� � ������ ��� ��� � � ���� ���� �3i �� � ��� � ��� �� ��� � �� � ��� � ���� ��� ���� � ��� �� ��� � � ���+ ,�� � ��� �� ���� �� �� �� � ��� ����� �� ����� ��� �� ��� � : ����� � ��; �3i ��� �� � ��� ����� ����� 9��� ��� ���� �� ���� ����+ ,�� ���� ��� � � '� �� �� �������� � �� /��� +:>; ��� �3i� /���� +:>; �� �� ����� ��e� �*�� �� /��� �+ 0��� �� ����� �� �� ���� ������ ��� � �� �� ������� ��� /��� � � ����� ��� �� �3i �� ���� ��� �� �*�� �� /��� �& �� � ��� /����� �� ���� ��� � �� ���� ���+ # � �� �������� ��� ������ ���� �� ����� ����� ��� �*�� � ����9 + 2� ��� ������ /� ���� ��� �� �� � � ���� /� ����� �� ���� ���� ����� ��������� = ��� � �� �� �� ��� �� �3i �� ��� ������ ���������� �� /��� �� /� ��� ��+ 5���� /� � �� ����� /� � �� �� � ��� ��� ����� �� � �� �3i �� �� � � � �� ���� �� � �� ��� �� ��� :�� ����� �; ��� �� �� �� ����� � ��� � � �*�� �� ��� � �� �� ��� �* �� ���� /��� � �. �� � �� � ���+ 6�� ��� �� �� �� �� ��� � �. ������ �� ��� �� ���� �� �� �3i:�6���;� �� /��� �6 : �� �� � ��; ��� �� �� �� � � � �� ���� �� � �� ��� �� ��� :�� ����� �; �� ��� ��� ����� �� �� �� �6 � ���� +���>��6� � �3i��6 � ����� � +���>������ :!; ��� � � �� �������� ��� +���>��6� � �3i��6 � ����� � +���>������ � % :$; ,�� �� ��� >:�6; ���� �� �� ����� �� ��� �* ��� � �� �� �� � �� � � � � �� �� ��� >:��; ���� �� �� ����� �� ��� �* ��� � �� �� �������� ��� � � � � � + 2� �3i:�6���; �� ���� ���� � � ����� �� �� �6 �� �� �� ��� �� ��+ #�� �� � �� ��� �� �� � �*�� �� ��� � �� �� ��� �* /��� �� �� �� � �� ��. ���� � ��/�� ��� ��� ����� �� + g� � ������ /� � ���� �� �� �3i �� ��� � �� ��� ���� � ��� ��� �� � �-� ��� �� �� �� ��� � � ������� �� � � ��� �/���� ����� ��� ��������� ��� �� ���� ��� ��� �� �� /���� ���� ��� ��� � ����+ 2� �� �� � ������ �� �3i �� /��� �� �6 /� � ��� /� �� �� ���� �* �� ��� �� �� ��� � ��� � ��� �� ��� �� ����� ���+ ! � ����"������� �������� �� #� ������ $% ,��� �� ��� �������� � �� �� ���� � �*� ��� �� �� ��������+ 0� ��� � �� ��� ��� ����� �� �� @!%�%%%� /�� � �� �� � �� ��� �� �������� ��� � ���� @a�%%% �� ������ �� ���� �� �� �� �� ������ � ��� �����+ 6�� � ���� ���� �� ����� �� �� ��� �� � ��� � � ����� ������� ��� � � � �* �� � �� aj+4f �� �������� �� � � ������ �� ��� �� � ��� � �* �� � �� a!+4f � ���� �� ��� �� + 6 � ��� ���� �� ����� �� ���� �� ����� ���� ��� �� ����� �* �� � �� $a+4f �� � ������ ��� �� �� �� ��� �� + ,���� �* ��� � ���� ��� ���� ���� �� # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !<j ��� ��� ���� ��� :$%%%;� ��� �� �� ��m� �� �� �� �� �������� ������ �� �� �*���/+ �� �� ��� �� i+#+ � ��� �� �� �� ������ �* �� �� �� $% �����+ # �� ��� ��� $%%!� ��� �*���� �� �� �� ����� !"f /��� �� ��� ��+ ,�� � � ��� � � � �� � �� �* �� �� ��� � � ��� ��� �� �� ��� ��� ������ �� � ������ �� /� �� ���� ����� /�� � �� ������ �� � ��� �� ��� �����+ 6�� �� � ����� �� ��� �� �� ���� � ��������� /� ��� � �� �� ��� �� ����� �� �� �� �� �. � � �� � �� �� !%f ����� ����+ 3* �� � � ���� "� /� ��� � �� �� �� � �� ����� <%f �� � � ��� � �� � ������ ���� ���� ���+ ,��� �� �� ��� � ��� �� ��� � ���/ �� ����9 �� �� ����������� ��� � �� ���� �� �� ���� �* �� ���+ � ������ � ���� �* ��/� �� ����� �� �� �� � � @a�%%% �� �� ��� �� ������ � ����� �������� �� � � �� ��� ����+ k����� �� �* ��� @a�%%% �� ! ���� �� �� ������ ���/���+ 6�� �� �� � �� /� � �� ���� �* �� ���� /� ��� � �� � �� ��� �� ���� "� �� ����� �� ����� ��� �� ��� ����� �� � �����1� ������ ��� � �� ��� � �* ��� ��+ �� �� �� ������ �� �* ��� ��� �� ����� �� ������� � �� :�� ����� � ����� /��� ���� ��� ��; �� �� � �� � �� ��������� �� ���� 4+ 0� ��� � �� a%f �� � �� �� �� �� � ���� �� ��� ��� �� � �� ����� �� �� ���� � �� ��� �� ����� ��� ���������� /�� � /��� �� �*�� � �� �������� �� � � �* �� �+ 2� ���� ���� "jf �� ����� �� ��� �� ����� ��� �����1�� ��� ��� ��� �� � �� ����� �� ����� ���� ��� /��� �� �*�� � �� ��� �� ���� �* �� �+ ,���� ��� � ���� ��� ��� �� ����� #����� ��� #��� :$%%%;+ ,��� ��� �� ������� ������ ��� �� ���� � �� ��� ��� ��� ��� �� ������� ������ ��� �� ����� �� ��� �� ����� �����1� ��� �� !$ � ������ ������ �. � � � ��� � ���� !j4$=!jj>+ ��������� /� ��� � �� �� ���� ��� ��� �� ������� �� �������� �� ��� �� ����� ��� ��� ��� �� � �� ����� ��+ ,��� ��� � ��� �� ��� �� ����e� �� ��� � � ��� �� �� �� � �� + #� � /� �� ��� � �� /��� � �� � �� ��� ��� �� ������� � ���� �� � ��� �� �� � � ����� � ���� �� �� � �� ��� ���� �� /���+ ,� �� �� � � ������ �� �� � �� �� � ��� ��� ��������� ���� �� � ���� /���� �� �� � �� ��� ��� �� ������+ ,�� � �� �� ��� �������� ��� �� ��� � �� �� ���������� ����1�� ��� �� �� ���� � �� �6 �� �� �*� ���� �� ��� �����+ 8�/����� ��� ��� � /��� �� ����� �� ��� ���� �� �� �����7 �� ����9 �� �� ����������� ��� � �� ��� �� �� ���� �* �� ��� :�,g;+ 6�� �� ���� ���� /� ��� � �� �� � �� ����� �� �*������ ��� �� �� � �� ��� !+<!f /���� �� � �� �*������ ��� �� �������� ��� � � ��� $+%(f� /� � � �� �� ����� �� a%+ ,���� ��� � ���� ��� ���� ��� ��� ���� ��� :$%%%;+ ,��� ��� ������� � �� �*������ ��� �� �. � � �� � ��� ��� ��� �� �. � � �������� ��� � ���+ h� ������ � ����� ��� �� � � �������� �� ������ �� ��� � ����� ��� ��� ��� . ����� /� � �� � ��� ����� ��� �� �� ������� ����+ 8�/����� � � ��-� ��� �� � �� ��� �� � �� ���� �� �������� �� � �� ��� � �� ������� � �� ���� �� ������� ��e� ��� ��� ��� � �� �+ k� ��� /� ����� �*� ��� �� ��/� �� ��� � � ��� �� � �� ��� ����� � ����� � �� ��� � ���� � �*� ��� �� � �� �� � � ��� � �+ � ������ � ���� ��� � ����� ,���� ! �������� �� ����� ��e� �� ��� �* �� ���� /��� � ��� �� ����� �� �� �� �6 /� �� ��� /� �� � �� ���� �* �� ��� ��� ��� �� ����� �� �� �� �� ��� ����� � ����� �� ����1���+ ,�� � ����� : �� �; �� dd �� � �� /� �� ���� �* �� ���ee ���� �� �� ����� �� ���� �� ������ ������� � �� ��� ������� �� � � ���������� � �� ��� � �� �� ��� �� �� � � ����1��+ !"% # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& ,���� ! ���/� �� � ���� $< ����� ��� �� ����� �� �� �� �� � � ������ �� ����� �� �� �� �6 /� �� �� ���� �* �� ���� ��/����� � ���� a" ����� ��� �� ����� �� �� �� �� � � ������ �� ����� �� �� �� �6 /��� �� ���� �* �� ��� �� ���1��+ ��������� ,���� ! ���/� �� ��� ��� ����� �� ����1���� �� �6 /� � �� ���� �* �� ��� ��/��� ��� � � �� � ������ � � � �� �� ��� �* �� ���� /��� � � ����� � �� �6 /� �� �� ���� �* �� ���+ ,��� ��� � �� ����� � ��� �� ����� �� /� � ���� � ��+ :$%%!;� �� /��� �� '� �� � ��+ :!j>j;� /�� 9�� �� ���� ������ ��� �� ��� � ��� ���� � �� �� ��� ��� � ���� � �+ ,���� $ ���/� �� �� ���� �� ��/ � �� �� � �� �� ��� �* /��� � ��� �� �6 /� � :�6� �,g; ��� /� �� �� ���� �* �� ���+ ��� ��� ��� ��� � �� �� � �� �� ��� �* /��� � ��� �� ����� �� �� ���� ��� ��������� ��� �� � ���� ��� ��. �� + � ������ � � � ��� � ����� �� ��� ��� ����� �� �� @!%�%%%� � �� �� ��������� ���� ��� � �� �� �� � �� �� ���� ! �� @!%�%%%+ #�� � /� ��� � �� �� ��� �� ����� �� �� ��� �� �� � �� �� !%f ��� �� � �� �*����� �� �� �� � �� �� !+<!f� �� �� ������ ��� �� �� �� �� �� !%� !�<! � >�"j�+ ,� �� �� ������ ������� �* ���� ��� � �� �� �� � �� �� ���� ! �� :!+%>"j;:!%�%%%� � !%�>"j+ ,�� � �� �� �� � ���� ! �� !%�>"j � !%�%%% � >"j+ ,���� ! �� ��� �* �� ���� /��� � ��� ����� ��� @!%�%%% ���� ���� �� ���� ��� � �� ������ ��� � �� �� ���� � �� �� ��� 2���� �� ����1�� �� ��� �* �� ���� /��� � ��� ����� ��� �� � �� � �� /� � ���� �* �� ��� � �� � �� /� �� ���� �* �� ��� �������� ��� � � !% !!�((! !!�$%a !%�>%4 !" !"�j%4 !<�>%4 !<�$a> $% $!�<j< !j�4(< !j�$4" $a $"�(<j $a�a>! $a�a<! $< $(�a<4 $<�((a $<�j$! $" $j�%<a $4�$"$ $4�4$4 a% aj�$<4 a"�!<% a(�<%( a< <j�jaa <<�<"! <j�"<% a" "a�%a! <(�!"! "a�!j> <% (!�4"j 4a�$!4 (4�a$< ,��� ���� ����� � �� �� ��� �* �� ���� /��� � ��� ����� ��� @!%�%%% �� �� �� ��� �6� �� � /� � ��� /� �� �� ���� �* �� ��� ��� ����� � � � ����1��� ���� �� �����/��� ��� � ����� �� ��� ��� ����� �� �� ��� �� � �� @!%�%%%+ �� ����� �� �� ��� �� � ��� � � ����� ������� ��� � � � �* �� � �� aj+4f �� �������� �� � � ������ �� ��� �� � ��� � �* �� � �� a!+4f � ���� �� ��� �� + 6 � ��� ���� �� ����� �� ���� �� ����� ���� ��� �� ���� �* �� � �� $a+4f �� � ������ ��� �� �� �� ��� �� + ,�� ��� �� ����� �� �� �� �� �. � � �� � �� �� ��� �� � �� !%f ����� ����+ 3* �� � � ���� "� �� �� � �� �� ��� �� � ���� <%f �� � � ��� � �� � ������ ���� ���� ���+ ,��� �� �� ��� � ��� �� ��� � ���/ �� ����9 �� �� ����������� ��� � �� ���� �� �� ���� �* �� ���+ ,�� a%f �� � �� �� �� �� � ���� �� ��� �� � �� ��� ��� �� � �� ����� �� �� ���� � �� ��� �� ����� ��� ���������+ /�� � /��� �� �*�� � �� �������� �� � � �*+ 2� ���� ���� "jf �� ����� �� ��� �� ����� ��� �����1�� ��� ��� ��� �� � �� ����� �� ����� ���� ��� /��� �� �*�� � �� ��� �� ���� �* �� �+ 0� ��� � �� �� � �� ����� �� �*������ ��� �� �� � �� ��� !+<!f /���� �� � �� �*������ ��� �� �������� ��� � � ��� $+%(f� /� � � �� �� ����� �� @a%+ # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !"! ,���� $ ��� �� ��� �� �� �� ��� �* /��� � ��� ����� ��� @!%�%%% �� � �� � ��� �� � /� � ��� /� �� ���1� ��� �� �� ���� �* �� ��� l��� ��� �� ����� �� �� :f; c�������� ���� ��� � 3����� ������� �* ���� ��� � c�������� �� ���� ,� �� �� �� ���� �� ���� :����; �����1�� ��� �� ���� h� ��� �� ���� :����; '������� ��� ���� � �� ��� �� ���� 3����� �� ���� 3����� ���� ��� � 3����� �� ��� �* ���� ��� � ����� �� /� � ���� �* �� ��� ! !% !%�%%% !%�>"j !%�%%% >"j 4%! a"" 4%! $"> !%�<$( !%�4(a $ !% !%�4(a !!�"j% !%�<$( j!( 4<$ "$< >>> $(" !%�jja !!�a"( a !% !!�a"( !$�aaa !%�jja j(4 4>a 4!> !�%<( $ja !!�4<$ !$�%(! < !% !$�%(! !a�!%> !!�4<$ !�%a( ($4 44$ !�!"" a!! !$�a"! !$�>$< " �<% !$�>$< (�"!< !$�a"! �"�a!! �"�a!! % �<�>a( % !$�a"! >�(%$ 4 !% >�(%$ j�<<j >�(%$ (<( "$a a%j "$a $$< j�%(a j�$>> ( !% j�$>> !%�%>" j�%(a (j> ""4 <"4 ((a $aj j�"44 j�>>a > !% j�>4a !%�(a$ j�"44 ><j "j< "a> j!! $"" !%�!a! !%�"%< j !% !%�"%< !!�<%4 !%�!a! j%$ 4a$ "ja !�%%" $(! !%�(<( !!�!"j !% !% !!�!"j !$�!!> !%�(<( j"j 4(! 4aj !�%>a $>4 !!�<a$ !!�>(4 !!�((! ����� c� /� �� ���� �* �� ��� ! !% !%�%%% !%�>"j !%�%%% >"j 4%! a"" 4%! $"> !%�<$( !%�4(a $ !% !%�4(a !!�"j% !%�<$( j!( 4<$ "$< >>> $(" !%�jja !!�a"( a !% !!�a"( !$�aaa !%�jja j(4 4>a 4!> !�%<( $ja !!�4<$ !$�%(! < !% !$�%(! !a�!%> !!�4<$ !�%a( ($4 4>$ !�!"" a!! !$�a"! !$�>$< " �<% !$�>$< (�"!< !$�a"! �"�a!! �"�a!! % �<�>a( % !$�a"! (�"!< 4 !% (�"!< >�!"j !$�a"! 4<" <"$ $4( <"$ !j< !$�4(! >�%$% ( !% >�%$% >�(%> !$�4(! 4>j <>$ aj< 44( $%( !a�%j( >�"a< > !% >�"a< j�$4( !a�%j( (aa "!a <4< (>( $$% !a�">< j�%(% j !% j�%(% j�><j !a�">< ((j "<" "!$ >4> $a< !<�!!( j�4a4 !% !% j�4a4 !%�<4a !<�!!( >$> "(j ""$ ja" $<> !<�(%> !%�$"" !!�$%a i���� �� �� � ��� � ���� �� ����� �� ,���� !� ����� � �� ��� ���� ���/� �� �� ����� �� �� ��� �� �� ��� �* /��� � ��� �� ����� �� �� �� �6 ��� ��� ���1� ��� �� �� ���� �* �� ��� ��� � !%����� ����� �� � /���� ����� c ���/� �� �� ����� �� �� ��� �� �� ��� �* /��� � ��� �� ����� �� �� �� �6 /� �� �� ���� �* �� ���+ ! " $ # 0 # �������� � ' � � �� 2 � � � � 3� �� � � ��� � � ����� 4 ����. & � )5 � � & * & � 6 7 & 8 & h�/� �� �������� ��� ���� � �� ��� �� ���� ��� ���� ! �� :%+a;�>"j� � $"> ��� �� ��� �� ����� �� ��� �� ���� /�� � �� >"j� $"> � 4%!+ ,��� "jf �� ��� ����� �� ��� �� ���� �� ��� �� � �� ��� ��� �� ��� �*�� � ��� �� ���� �* �� �+ ,� �� �����1�� ����� �� ��� �� ����� ��� ���� ! �� :%+"j;�4%!� � a""+ ,�� �� ��� �� ���� :����; �� ��� ���� �� �. �� � ��� �� ���� :����; �� �� ���� �� � ������1�� ����� �� ��� �� ���� ��� �� ������ � ����+ ,� �� �� �� ��� �� ���� :����; � ���� ! �� �. �� � 4%!+ ,�� ������ ���� ��� � � ���� ! �� �. �� � �� ��������� ���� ��� � � ���� ! �� � �* �� �������� ��� ���� � �� ��� �� ����7 �* �� ����� �� ��� �� ���� /�� � �� :!%�>"j; � :%+aj4;:$">; � :%+$a4;�a""� � !%�4(a+ ,�� ������ �� ���� � ���� ! �� �. �� � �� ��������� �� ���� �� � �������� ��� ���� � �� ��� �� ���� ������� �� �� � ����� �� ��� �� ���� ������� �� /�� � �� !%�%%%� �!� %�aj4�:$">; � :! � %+$a4;�a""� � !%�<$(+ ,���� �� ��������� ���� ��� � � ���� $ �� �. �� � �� ������ ���� ��� � � ���� !� ��� �� ��������� �� ���� � ���� $ �� �. �� � �� ������ �� ���� � ���� !+ 0� �� �� �� �� � ��� �� �� � �� �� � �� ������ �� ��� �* ���� ��� � �� �� �� � �� ����� �� /� �� �� ���� �* �� ��� � ����� $=j+ � �� ��� �� ���� !%� �� ����� �� /��� ���� ��� �� ��� ����� �� + ,� �� �� �� ��� /��� ��� ��� �� ���� �* �� ������1�� ��� �� ���� �� /��� �� ���� �� �* ��� �� �� ���� �� ��� �� ����+ ,��������� ��� � � ���� !%� �� ������ ���� ��� � �� ���� ��� �� ������ �� ����� ���� �� ��� �� ���� � � � ��� � !<�(%> � !%�$$" � <�<>a+ #�� � ��� �� ���� �* ���� a�%%%� ��� ���� a�%%% �� �� ��� � ����� �������� �� � �+ ,� �� �� ������ �� ��� �* ���� ��� � �� �. �� � �� ������ ���� ��� � �� � �� �* �� �� /�� � �� !%�$"" � :a�%%%;�%�a!4� � !!�$%a+ ,�� �� � ��� �� �� ��� �� �� �6 /� �� �� ���� �* �� ��� �� �� ��� � �� �� � �� ������ ���� ��� � ��� �� �6 /� � �� ���� �* �� ��� ��� ��� ���� �� ��� �� ��� �� ����+ 8�/����� ��� � ���� �� ��� �� ��� �� ������� � � �� ���� " �� � � �*� ���� �� �� �6 /� � �� ���� �* �� ���� /� ��� � �� �� ����� �� ����� ��� �� ��� ����� �� � ���� " � �����1� ��� �� ���� ��� �� ��� �� �* ��� ��+ ,� �� �� ������ ���� ��� � � ���� " ��� �� �6 /� � �� ���� �* �� ��� �� �� ������ ������� �* ��� � �� � �� �* ��� ��+ 8�/����� �� � � �*� ���� �� �� ��� �� ���� �� <�>a( /�� � �* ���� a�%%%� � �� ���� a�%%% �� �� ��� � ����� �������� �� � �+ ��� <�>a( � a�%%% � !�>a( /��� �� ������ ���/��� � �� ��* ����+ ,� �� �� ������ ���� ��� � � ���� " �� (�"!< � :a�%%%;:%+aj4; � >�(%$ � ����� � (�"!< �� �� �6 /� �� �� ���� �* �� ���+ ,�� ��������� ���� ��� � ��� �� ���� � ���� 4 ���� ����� � >�(%$+ �� ������ �� ������ �� ��� �* ���� ��� � �� �� �6 /� � ��� /� �� �� ���� �* �� ���� � �� ����� � �� �� ������ �� ��� �* ���� ��� � �� �� �6 /� � �� ���� �* �� ��� �� ���� �� ��� �� �� �� �6 /� �� �� ���� �* �� ���+ h� � �� �� ��� � ��� �� �� ��� ��� ����� �� �� @!%�%%% �� �� � �� ��� ���� �� � ���� � �� ��/� ���� ������+ 5���� �� ��� � ��� �� <%f ���� �� � ������ ���� ���� ��� � ���� " ��� �� ��� ��� ����� �� �� @!%�%%%� ���� �� �� ��� �� ���� �� <�>a( � ���� "+ 8�/����� ���� a�%%% �� �� ��� � ����� �������� �� � � ��� !�>a( /��� �� ������ ���/���+ 2� ��� ���� �� ����9 �� �� ���� �* �� ��� �� �. �� � � �������� � �� /��� �� ������ ���� ��� � �� �� �6 /� � �� ���� �* �� ��� ���/� �� ����� � ��� �� �� �� �6 /� �� �� ���� �* �� ��� ���/� �� ����� c� /�� � �� >�(%$ � (�"!< � !�!>>+ 8�/����� �� /� � ��� ��/� �� ��� ��� ����� �� � "�%%%� ��� �� ��� �� ���� � ���� " �� $�<!> ��� �� ����9 �� �� ���� �* �� ��� ��� �� � j">+ 2� /� � ��� � �� ��� ��� ����� �� � $%�%%%� ��� �� ��� �� ���� �� j�4(<+ ����� ���� a�%%% �� �� ��� ��� 4�4(< /��� �� ������ # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !"a ���/���+ 2� ��� ���� �� ����9 �� �� ���� �* �� ��� �� !%�a!$+ 2� �� � ����� �� ������ �� ��� ��� ����� �� � �� ������ �� ����9 �� �� ���� �* �� ��� ��� �� ������ �� � ��� ������ ������ ���/���+ 2� ���� ���� /� ������ �� ����� ��� � �������� �� ��� ����� �� ����1�� ������ ��� ����� �� �� �� �� � � ������ ����� �� �� �� �6 /� � ��� /� �� �� ���� �* �� ��� ��� ����� � ��� �� ����� �� ��� ��� � ����+ ,�� ��� � � ���/� �� � �� �� �� ������ � ����� ����� �� ��� �� ����� �� ��� �� ��� ����� �� ����1�� ������ ��� ����� �� �� �� �� � � ������ ����� �� �� �� �6 /� � �� ���� �* �� ��� �� ������ ��� �� ������ � � ������ ����� �� �� �� �6 /� �� �� ���� �* �� ���+ ��������� /� ���� 9�� �� �� ������ �� ��� �� ����� �� ��� �� ���� �� �� ����� �� ����1�� ������ ��� ����� �� �� �� �� � � ������ ����� �� �� �� �6+ ,��� ������ ��� �� �� �� �� �� �� �� �� ���/� �*���������+ ,� �� �� ������ �� ��� �� ����� �� ��� �� ������ �� � �� ��� ���/� �*���������+ & ��� ���������� ��"������� 2� �� ��� � �������� ��� ������ :$%%%�� $%%%�; ��� �� �� ��� � ���. �� � �����1� �������� ��� � ���+ 8�/����� ������ � ������ � ��� �� ��� �� �� �� � �� ��� � �� �� 5 ���� ��� ���� '�� � c���9 :5�'c; ��� � ��� �� ����� ��� ��� � ������+ l� � �� �� �� ��� � �������� �� . � � ��� � ��� �� ���� ���/�� � � /���� �� �� . �� ����+ g � �� �� ��� ���� �� �����/��� �� �� ��� � ����+ �� �� �� �� �� ���� � �*� ���� /� ��� � �� �� ��� ��� ����� �� �� @!%�%%%+ 6�� � ���� ���� �� ����� �� ���� � � ����� ������� ��� � � � �* �� � �� aj+4f �� �������� �� � � ������ �� ��� �� � ��� � �* �� � �� a!+4f � ���� �� ��� �� + ,�� ����� ���� ��� �� ����� �* �� � �� ��� �� � �� $a+4f ��� �� � ������ ��� �� �� �� ��� �� + ,�� � �� �*����� �� �� �� � �� �� !+<!f /���� �� � �� �*����� �� �� �������� ��� � � �� $+%(f� ��� �� �� �� ����� �� @a%+ i����� �� �� �� ���� � �*� ���� �� �� ��� ������� �� �� ���� /� ��� � �� �� ��� �� ����� �� �� �� �� �� � �� �� � � ��� � /� � ��� � !$�4>� ��� � ������ ����� ��� � !(�%(�+ 2� �� �� �� ���� �� � �� �� �� �� ���� ���� :��� �� � �� �� � �� �� ��; �� ��� �� � �� ������ �� �� � ������ �� ��� ��� ���+ 2� �� ������ � � �� �� ��� ������ �� ��. �� �� �� �� �� ���� ���� /� � ��� � !�!$4> ��� � ������ ����� ��� � %�!(%(+ ��������� /� ��� � �� �� ������ ��� �� ���� � �� ��� ��� ��� :�������� �� � ���� � �� ��� �� ����; �� ����� �� ��� ��� �� �� /��� !"f ��� <"f� ��� �� ������ ��� �� ����� �� ��� �� ����� �����1� ��� �� ����� �� ��� ��� �� �� /��� a"f ��� >%f+ ��� ���� ��� � ���� ��� ��� �� ���� � ��+ :$%%%;� /�� � ��� ����� �� ����� � ��� �� !$ � ������ ������ �. � � � ��� � ���� !j4$=!jj>+ 6�� �� ���� ���� /� ��� � �� �� ����� ��e� ��� � � � ��� �� > !�� � �!� �� /���� � � $� /�� � �� ����� �� /� � �� �� �� �� � � � �� 6���� ��� ��� ���� ����� :$%%!;+ ,��� � ��� �� ��� �� � ����� � �� � ��� � ���� �� �� � ��� � � �� ��� �� �� d���e �� ���9 ��� �� ���� ��� ���� �������� � �����*� � ��� $+ k� ��� /� /��� �*� ��� �� ����� ��� � �� � � ��� � � � ��� ���� � ��+ k���/���� /� ������� �� � � � �������� �� � �� � �� �� �� �� � �� �� h�/�� �� :$%%!; �� ��� � ����� � ���. �� �� �������� 9��� �� ��� � ��� ���� �� ��� ����� � !"< # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& �*��������� ��� �� �� ��� �* ����� �� /��� �� ��� ���� �� �������� ������ ������ ��� �� ����� �� �� ��� � �� �� ���� ��� ������ �� ��� � �� ��� ���� � �� � ���+ 0� ��/ � � �� � �� �� ��� � � �� �� "�%%% � ��+ 6�� ��� �� �� �� �� ����� ��e� ����� �� ����1�� �� � ������ /� ������ � "�%%% ��. �� �� �� ������ �� ���+ ,���� �� ��� ������ �� "�%%% ������ �� �� ��� �* �� ���� /��� � � �� ��� �� ���� � ��� ����� ��� �� � = ��� � �� � >�:�?=;� /���� � �� �� ����� �� ����1��� � � !� $� � � � � "�%%% ��� = � �6 /��� �� ����� �� ����� � �� �� �� � ��� = � �6� �,g /��� �� ����� �� ����� � �� �� �� � �� ��� ��� �� ����������� ��� � �� ��� :���� �* �� ���;� = � �� /��� �� ����� �� ����� � �� �������� ��� � �+ ,��� �� ��� �* �� ���� /��� � �� �� ��� � �� �� � �� �*�� �� ��� � �� �� ��� �* �� ���� /��� �� +n>:�?=;o ��� �� �*�� �� ��� � ��� ����� ��� �� �� = ��� � �� � +n�:>:�?=;;o �� �����/�� +�>���=�� � �"�%%% ��! >����=� :a; +���>���=��� � �"�%%% ��! �>����=��!�� !� � :<; 0� �� �� � �� �3i ���� � /��� �� ��� ���+ 6��� � /� ����� �1� �� � �� �� ��� � � � � � ��� ��� ��� � � ���� ��� /� ��� �� �� ��� ��� ����� �� �� �� �� � �� � �� �� �*�� ��� � �� �� � ��� �� �� �+ ,�� 9��� ��� ��� ��� � � � �� �� �3i+ 2� � ��� /����� ��� /� ����� �� ��� � �� � /� � �� �� � ��� ��� ����� �� � �� �3i �� �� � � � �� ���� �� � �� ��� �� :�� �����; ��� :�� �; �� ��� ��� ����� �� � ��� �� �*�� �� ��� � �� �� ��� �* �� ���� /��� � �. ��+ 6�� ��� �� �� �� �� ��� � �. ������ �� ��� � :�3i:�6���;; �� /��� �� � �� ��� �������� ��� � � ����� �� � �� �� � � � �� ���� �� � �� ��� �� :�� �����; ��� :�� �; �� ��� ��� ����� �� � ���� +���>����6� � �3i��6 � ����� � +���>������� :"; 2� � ��� /����� +���>����6� � �3i��6 � ����� � +���>������� � % :4; k� '�����6 � ��� � +���>����6� � �3i��6 � ����� � +���>�������� � %� � 9�� �3i:�6���;� /� �� �� �� � �1� ��� �� ��� �� �� �� ���� �� � � � �� ���� :�3i:�6���;; �� � �� ��� �� :�� �����; ��� :�� �; �� ��� ��� ����� �� �� �� �� � �� � ��� '���:�6���; ����� � 1���+ 6�� �� � �1� ��� � ������ ���� ����� /� ����� � � �� �� �� �� ��� /��� �� � ����� �� �� �� ��� � �� '���:�6���; �� �� ��� !%% � ��� ���� �� ���� ��� !f+ g�� � ��� ��� ����� � �� "�%%% �� �� ����+ � ������ � ,���� a� /��� ���������� ���� �� �������� �� ������� �� �� ��� �* �� ���� /��� � ��� �� �� /��� �� ���� �� �� �6 /� �� �� ���� �* �� ��� /��� �� ����� ��e� ����� �� ����1�� �� � ���� !! �����+ 8�/����� �� �� ����� �� ������ �� �6� ��� ��� �� ���� �* �� ���� �� ����� �� ����1�� ������ ��� �� ��� �� �� ��� �* �� ���� /��� � ��� �� �� � �� ���� �� ��� �� �6 /� � �� ���� �* �� ��� �* ���� � � ���� !< �����+ # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !"" ,���� a #� �� ��� �� �������� ��� � � ��� �� � �� ����� �� � �� ����� 2���� �� ����1�� �� ����� 3*�� �� /��� � �� � �� � �,g 3*�� �� /��� � �� � �� 3*�� �� /��� � �������� ��� � � �3i� �� � �� � �,g ��+ �������� ��� � � �3i� �� � �� ��+ �������� ��� � � �3i� �� � �� � �,g ��+ �� � �� " !<�j4( :a�>>j; !<�(>" :a�j(>; !<�<$< :<�%<!; <<" :%+%%4j; $(a :%+%%%j; !(j :%+%%!a; !% $!�>(a :(�j($; $!�$44 :>�!!$; $!�$"% :j�$(>; 444 :%+%%jj; $($ :%+%%4j; aj( :%+%%$"; !! $a�>(% :j�!jj; $a�!<j :j�a4a; $a�a(j :!!�%a!; 4>4 :%+%%"!; $<j :%+%%%4; <<< :%+%%!$; !a $(�>aj :!!�<$4; $4�(j$ :!!�"<4; $(�4"$ :!<�!4!; (%( :%+%%a4; !>% :%+%%"!; "<" :%+%%%"; !< a%�<a" :!a�(><; $j�$4a :!a�j"a; a%�("( :!(�j!j; (!a :%+%%$"; !a% :%+%%a$; ">" :%+%%%>; !4 a"�a>( :!4�<(a; aa�((< :!4�"(j; a4�<4( :$$�!j%; 4jj :%+%%%<; $$ :%+%%%j; 4>$ :%+%%$<; !( a>�"!j :!j�%""; a4�(%( :!j�$%<; <%�<%% :$4�<>>; 4>< :%+%%"<; �<4 :%+%%!!; ($j :%+%%%a; $% <>�<%a :$"�>("; <"�($! :$"�jaa; "a�%%% :a>�aa"; 4$< :%+%%%>; �$4j :%+%%%"; >(( :%+%%%>; $" ($�""a :<<�("j; 4(�(%a :<<�<!j; >(�a4" :("�%(>; a>% :%+%%%"; �((< :%+%%$<; !�%>4 :%+%%">; a! !!<�$4% :(j�ja$; !%<�j(( :(>�<">; !""�%<4 :!"(�$j%; a4 :%+%%%(; �!�"%$ :%+%%<"; !�a>a :%+%%!>; a$ !$a�4<% :>(�%$<; !!a�a<< :>"�!$!; !(!�a!( :!($�j4"; �"( :%+%%%>; �!�""> :%+%%%j; !�<a% :%+%%%!; a" !"<�%4( :!!a�"$4; !<%�%>4 :!!%�4a"; $$4�4%> :$<a�4(!; �a%% :%+%%%a; �$�!$j :%+%%aj; !�"a! :%+%%!<; ,��� ���� �������� �� �*�� �� ��� � ��� � ������ ����� ��� :�� ����� �����; �� �� ��� �* �� ���� /��� � ��� �� � �� ��� �� ����� ��� �� ��� !%�%%% �� �� ��� �� �� �6� �� � /� � ��� /� �� ���1� ��� �� �� ���� �* �� ���+ 2 ���� �������� �3i ��� ���� ��� ��� � �� ��� ��+ �� � ���+ ,� �� �+�+� �� �� �� �� �*�� �� /��� � ��� �� �� � �� ��� �� ���� �� ��� �� �*�� �� /��� � ��� �� �6� �,g � �� ��� �� �� !< ������ �� � ������ ����� ��� :�+�+� ����; �� ���� �� �� /���+ ,� �� �� �3i �� ��� �� � �������� �� � ����� + 6�� �*� ���� �� �� � dd�3i� �� � ��� �,g ��+ �������� ��� � �ee �������� �� �3i �� /��� �� �� � �� /� � �� ���� �* �� ��� ��� �� �������� ��� � �+ ,�� ���� ��� :���� ���; �3i �� � � � �� ���� �� � �� ��� �� ��� :����� �; �� ��� ��� ����� �� �� �� �� � �� /� � �� �* �� ��� �� ����� � ��� �� �������� � �� /��� �*�� �� ��� ��� �� �� ��� �* /��� � ��� /� � �� ����� �. �� � � � ��� �� ����� �����+ ,� �� � �� ���� �� �� a$ ����� �� �� ���� ��� �� �6� �,g � �� ��� �� �� �� ��� �� �� � �� ���+ h� � �� �� �� �� ��� ��� �� �� �� ����� �� ���� � � ����� ������� ��� � � � �* �� � �� aj+4f �� �������� �� � � ������ �� ��� �� :������� �����;� ��� � �* �� � �� a!+4f � ���� �� ��� �� :/� ����/�� �����;+ ,�� ����� ���� ��� �� ����� �* �� � �� ��� �� � �� $a+4f� �� � ������ ��� �� �� �� ��� �� + ,�� � �� �*����� �� �� ��� �� �� � �� �� !+<!f /���� �� � �� �*����� ��� �� �������� ��� � � �� $+%(f : �� 44 ����� ���� � ��m� �� ��� ��� � ��� 3*����� ���� ������;� ��� �� �� �� ����� �� @a%+ ,�� ��� �� ����� �� �� �� �� �� � �� �� � � ��� � /� � ��� � !$�4>� ��� � ������ ����� ��� � !(�%(�+ ��������� /� ��� � �� �� ������ ��� �� ���� � �� ��� ��� ��� :�������� �� � ���� � �� ��� �� ����; �� ����� �� ��� ��� �� �� /��� !"f ��� <"f� ��� �� ������ ��� �� ����� �� ��� �� ����� �����1� ��� �� ����� �� ��� ��� �� �� /��� a"f ��� >%f+ 6������� /� ��� � �� �� ����� �� �*���� � ��� �� ���� ��� ���� �������� �������� �� /� � � ���9 ��� �� $+ ! " 4 # 0 # �������� � ' � � �� 2 � � � � 3� �� � � ��� � � ����� 4 ����. & � )5 � � & * & � 6 7 & 8 & ,���� a ���� �������� �� �� ��� � �. ������ �� ��� �+ 6�� �� ����� ��e� ����� �� ����1�� ��� % � !4 ������ �� ��� � �� �� �6 �� ���� �� ��� �� ��+ ,� �� ���� ���� ����� �� ����1���� �� � ���� � �� "�"�� �" ��� �� �� � ��e� ��� ��� ����� �� � ��� �� ��� � ��� ����� ��� �� �� �6 ��� �� �� �� �� �+ 0��� �� ����� ��e� ����� �� ����1�� �� � ���� !( ������ �� ��� � ��� ����� ��� �� �� �6 �� ���� ��� �� �� �� ��+ ,� �� �� � ���� � �� ����� � �� �� � ��e� ��� ��� ����� �� � ��� �� ��� � ��� �� � ����� �� � �� �� �+ 8�/����� �� �� ����� �� ��� �� ����������� ��� � �� ���� �� ����� �� ����1�� �� ���� �� ��� � ��� �� �� ���� �� ��� �� �6 /��� �* ��� � a$ �����+ 6�� �� ��� �� � �� ,���� a� /� ���� �� �� �� � �� /� � �� ���� �* �� ��� ��/��� � ������ � �� �� � �� /� �� �� ���� �* �� ��� �� ��������� ��� ��� �+ ��������� �� �� ��� � �. ������ �� ��� � ���/� �� �� ��� �� � �� �� �� �� �� �� �� ��� � �� �� ���� �* �� ���+ 2������ �� ���� � � ������ �� ����� �� ������ �� � � � � �� �� �� ������ �� ��� � �� �� �� ��� ��+ 6 � ��� ���� ����� �� ��� � ��� �� �� ����� �� ����� � ��� ����� !%�%%% �� � ��� �� � ��� �� ��� �� � ��� ������ ������ ���/��� �� ��� "�"%% ��� a%����������� �� ����1��+ ,� �� �� ��� � �� �� ���� �* �� ��� /��� �� ���� ������ �� �� ����� �� ��� �� ���� ��� ����� �� � � ��� �� � ��� ������+ ,���� < �������� �� �� ��� � �. ������ �� ��� � ��� �� ��/� �� ���+ 0� ��� � �� �� � �� �*����� �� �� �6 �� %+$f /���� �� � �� �*����� �� �� �� �� %+af+ ��� � ��� ��� � ���� ��� �� �� � �� ���� �� �� ���� ���+ ,�� ��� ��� � ��� �� ���� �� �� ��� �� ��/� �� ���� � ���� " ����� ��� �� �� � � ������ �� �6 /� �� �� ���� �* �� ��� � ����� � !( ����� �� �� ���� ���+ 8�/����� � ���� !% ����� ��� �� �� � � ������ �� �6 /� � �� ���� �* �� ��� � ����� � a$ ����� �� �� ���� ���+ 2� � � ����� �� ��/� �� �� � �� �� ��� �� � �� �* ����9 ��� + 0� � � ��/� �� � �*���9 ��� �� � ��� � � �� �� #p� "%% ����*�� �� � ��� ����� ��� /� �� �� ���� � ����� �*�� ��� � ������ � � ����1�� ��� ��� �� �� � � ��������+ ,��� /��� �� ����� ����� �� ����1�� ������ ��� �� ����� �� �� �� �� � � ������ �� ����� �� �� �� �6+ ,���� < �������� �� �� ��/� �� �������� ��� � � ��+ �� � ��� /� � ��� /� �� �� ���� �* �� ��� 2���� �� ����1�� �� ����� 3*�� �� /��� � �� � �� � �,g 3*�� �� /��� � �� � �� 3*�� �� /��� � �������� ��� � � �3i� �� � �� � �,g ��+ �������� ��� � � �3i� �� � �� ��+ �������� ��� � � �3i� �� � �� � �,g ��+ �� � �� < !<�$$! :a�a%a; !<�!%( :a�a(<; !<�%>a :a�<aj; !a" :%+%%a%; $" :%+%%%4; !!4 :%+%%$!; " !"�"%! :a�j%>; !"�aaj :a�jj<; !"�<%j :<�!>!; !a> :%+%%%4; �!< :%+%%!$; !"" :%+%%%>; j $!�j"$ :(�<<!; $!�<j> :(�">>; $$�4"a :j�%a(; !j :%+%%%j; �$j$ :%+%%!!; a%4 :%+%%(>; !% $a�j(a :>�(>$; $a�<!% :>�j$<; $"�!!! :!!�%4!; �a( :%+%%$a; �aj4 :%+%%!j; a"$ :%+%%%(; 8������ ��� � ��� ��� � ���� �� ,���� a ��� �� � ��� ���� �������� �� ��� ��� � :��� � ������ ����� ���; �� �� ��� �* �� ���� /��� � ��� �� �� ��� � �. ������ � �� ��� � ��� �� ����� �� �� �� ��/� �� �� ��� �� ��/� �� �6 /� � ��� /� �� �� ���� �* �� ���+ 2� �� ��/� �� ���� /� ��� � �� �� � �� �*����� �� �� �6 �� %+$f /���� �� � �� �*����� �� �� �� �� %+af+ ,�� �������� � �� � ��� �� � �� ��� ��� � �*������ ���� ���/� �� �� ��� ��� � ��� 3*����� :�p3; ���� ������+ # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !"( ' ��"(������� ������� ,��� �� ��� �� �� ���� �� ��� ����� �� ����1�� ������ ��� �� �� � ���� �� ������ ��� � �� �� ��� �* �� ���� /��� � � ����� � �� �6 /� � ��� /� �� �� ���� �* �� ���� ��� ����� � ���� � ��� �� � � ����+ ,���� " ���/� �� � ������� � ��� ���� � ��� ��� �� � �� ��/�� �� ��� �� ��� �� ��� �� ������ �� ����1�� ������ ��� �� �� � � ������ �� �6+ ,��� �� ����� �� /� � �� ,���� " #���� ��� � �������� �� �� ��. ���� � � ����1�� ��� :������� �� ; �������� ��� � � �� ���� � ���� �� � ���7/� � ��� /� �� �� ���� �* �� ���7�� � � � ��� �� �� ��� �� ��� � ������ ����� ��� ��� ���9 ��� �� ���� ��� ���� �������� ��� �� ��� �� �� :f; # ������ ����� ��� :f; ���� ����� �� ����1�� ������ ��� ����� �� �� �� � � ������ �6 ��� �6 � �,g ���� � $ ���� � a �6 � �,g �6 �6 � �,g �6 !<+4> !j+%( $< !< q<% $% !(+%( !j !a a$ !( !"+%( !4 !$ $a !" !$+4> !j+%( q<% !> q<% $j !(+%( a$ !( q<% $4 !"+%( $" !4 aj $! !%+4> !j+%( q<% $( q<% q<% !(+%( q<% $< q<% q<% !"+%( q<% $a q<% a" ,��� ����� ����� � �� ����� ��� � �� �� ��� ����� �� ����1�� ������ ��� �� �� � ���� �� ������ ��� � �� �� ��� �* �� ����� ������� � ��� ��� � ���� �� ,���� a ��� �� + h� � �� � ��/�� � �-� ��� :��� �*�����; ����� �� �� ��� ���� �� � ������ ����� ��� ��� ���� �� ���� �������� ��� �� ����� �� ��. ���� � � ����1�� �� ��� �� � �� ��� �+ ,���� 4 #���� ��� � �������� �� �� ��. ���� � � ����1�� ��� :������� �� ; �������� ��� � � �� ���� � ���� �� � ���7/� � ��� /� �� �� ���� �* �� ���7�� � � � ��� �� �� �* �� � ������ ��� �� :�������; �* �� � :f; �� ��� �� :/� ����/��; �* �� � :f; a!+4% $j+4% $(+4% $"+4% $a+4% $!+4% �6 � �,g aj+4 a$ $> $4 $a $% !( �6 aj+4 !( !" !$ !% > 4 �6 � �,g a!+4 a4 aa a% $( $a $% �6 a!+4 $a $! !> !4 !a !$ ,�� ����� ��� � �� �� ��� ����� �� ����1�� ������ ��� �� �� � �� ��� � �� �� ���� ���� �� � � � ��� �� �* �� ��� ������� � ��� ��� � ���� �� ,���� a ��� �� + 6�� �*� ���� �� �� ������� �* �� � ����� � �� ��� �� �� a!+4f ��� �� ������� �* �� � � ���� �� ��� �� 7/��� �� � ��� ��� /� ����/�7�� $!+4f� ��� � �� �� �� �� ��� � �6 � �� ��� �� �� ����1�� �� ���� �� ��� $% �����+ 2�� ��/����� �� ���� �* �� ��� �� �� �� � ���� �� �� �� ������� ��� ��� �� �� �� �� �� !$ �����+ h� �� �� �� �� �* �� � � ���� �� ��� �� �� ������� �� ���������� ����1�� �� ������+ !"> # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& �� �� ���� � ��������+ 5���� � ��� ���� � ���� �� ������ �� � ������ ����� ��� �� �� ����� �� ��� �� ������ �� ����1�� ������ ��� �� �� � � ������ �� �6+ 6������� �� /� ���� ��� �� ��� � ��� ���� � ���� �� ������ �� ���9 ��� �� ���� ��� ���� ��������� �� ������ �� ����1�� ������ ��� ����� �� �� �� �� � � ������ ����� �� �� �� �6+ (� � �� .� "�� �� � ��� ��� �� ������ � � ���� ������ � �� ����� �� �� ��1�� ����� � 1������ �� �� �� 2� �� ��1��� � �� �� � / �; � � ��+ ,/� �� �� �� � �� ���� � ��� �� � � ���� ��� �� ���� ��� ��� � �* �� ��+ �������� �� ��/�� �� �* �� � �� �� ��� �� � � ����� � �� ������ ��� �� �� �� �� ���� �� �� ����9 �� � �������� ��� � �+ ,���� 4 ��9� � �� ��� �� �� ���� �� ��� ������ ��� ���� � ��� ��� �� ��� ��� ���� ���� ��� ��� �� �� ����� ��� �� �* �� �+ 2������ �� ��/�� �� �� �� �� � ����� �� � ��� �� ����� ������ ��� ����������+ ) ���������� �� *��� ��"��+� ,�� ��� ��-� ��� �� ��� ����� /�� � . �� ��� �� � �� �� �� �� � �� ��� � /��� ��� ���� �� ���� ��� ����9 � �� �������� ��� � ��� ���� � �� � ���+ 0� �� ��� �� �� �� �� �� � d���� �* �� ���e � ������ ������ �� �� � ���� �� ���� �� �*�� /� ��� � �������� ��� � � � � ��+ 0� ��� �� �� �� ��� ��� ��� �� : � �� ���; �� ��� �* /��� � /� ��� � �������� ��� � �� � ����� � � �� � ��� �� ������� � ��� ���� ���+ ,��� �� ���� ��� � �������� �� � ����� � �������� � ���� ��� ���� �� /� �� ���� ����+ ,� �� ���� /� ��� �� �� ��� � ���. �� � � � � �� �� ��� � �. ������ � �� ��� � ��� ��� ��� � �� �� ��� ��� �������� ���� �� � �� �����+ 0� �� � ��� �� �� �� �� ��/� �� �������� ��� � ���7/� � ��� ��� � �*������ ��/�� ��� !% ����� ���� �7��� � ������ � ��/� �� �� � ��� ��� ����� ��� /� � � ���� � � ����1��� �� �� � �� �������� �� � ���� !% ����� ��� ��� �� ������ �� ���� ��������+ ���� ��� ���� /�� ��� ��� ����������� ��� �������� �� ���������� ����1�� �� ���� ������+ 6 � ��� ���� �� ��� ������� �� � � ��� � � � ��� �� � ������ ���� �* �� ���� �� ������� � ���������� ����1�� ����� � " �����+ 2� � ��� /����� � ��� �� "� � ����� �� �������� �� �������� ��� � ��� ���� � �� � ���7���������� �� ������� ��� ���� � ���������� ��� � ����7�� � � � � ���� 1����" � ����� �������� ��� � ��� �� � �� �� �� ��� �� ���/ � ��� � ����+ ��������� ��� �� ������� �� �������� ��� � � /� � 44 ����� ���� � �� ��� ��� � �*������� �� �������� �� ���������� ����1�� �� �� �� ���� �� a% �����+ 6 � ��� ������ � /��� ����� �� �� � � ���. � � �� ���� � ��� �� � �������� ��� � �+ 2������ �� �� �� ���� ��� "=!%f �� �������� ��� � ��� ��� ��� � �1��� �� �* ��� �� �� �� �� �� ��� ����� �� ��� ��������� ��� � � ������� ���/��/� � �� ���+ c��/�� �� ����� �� ����� ��� 0�����/��� :!jjj; �*� ���� �* ��� �� �� �� ���� � ���+ 0�� ��������� �� ��� � � � � �� �3i /� � � ������� ����� �� ����� �� � �� �����+ k���/���� �� �� � ���� �� ����� � � ���� ��� ��� ��� � ��� �� ��� �� ����9 � �� �� ��/ ������ ��� �� �������� ��� � ���� ��� ����� � �� ����� �� �� ��� � �. ������ � �� ��� � ��� ���� ��� �� ������ �� ���� ��������+ 6������� ��� ��� �� ���*�� �� � �� �� � ���� �� /� �� ��� � �*� ��� �� ��� � /�� � ����� �� ���� � ������ /� ��� � �������� ��� � � � � ��� ��� /�� � ������ � ���� �� ���� �� ���� ���/� �� �� ���� �� � ���e ������ 7� �� '� ��� #�� ��� r���� :$%%!;+ # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !"j 2� � � ������� ���� ���� /� ����� ���� . �� ���� ��� � �� ������ �+ ��+��,�� �"���� ,�� � ���� ��� ��� �� � � �� ,2�����36 2�� � � �� h�/ l��� �� � ��� ��������� � ��� � � ����� ��� ��� ������ � ���-� + -��������� ��� ��� �+� p ,���� '+ :!jj";+ #� ���������� � � ��� � �� ���� � �� ������ ��� ����� ��� ��� /��� � /� � ����� � ���� � �� � ��� � � �� ��� ��� �+ ��� %�� ��� �� 4��� ��" (��� ����� 85:!;� a%=<j+ ���� � �+ '+� c������ �+ k+� p l�� )+ :$%%!;+ k��� ������ ���� 0�� e� � /�� � � �� �*���� ����� ��s ��� %�� ��� �� >��� � #���2���� � #������ !%=!(+ c��/�� )+� �� ����� g+� �� ����� )+� p 0�����/���� �+ :!jjj;+ ,�*��� �� ��� �� �� � �� h���. ���9�� ��� � ��� ��� ��� ��� ���� ��� . ���9�� � � � �+ ?� ����� ��; %�� ���� &�:a;� $!j=$a<+ c ������ �+ '+� p ������ #+ �+ :!j>%;+ � �� �� ��� � �� �� �* ���� ���� �� ��� �� �����+ %�� ��� �� �� 0�� ���� ��;� ��� 0������ ���� &� a<=<!+ ���� �� ������� 5+ :!j><;+ g� � �� � � � ������ /� � �������� �*��� 2 ��� � ���� ��� ��� �� ��� ����� �� �� ���+ ��� %�� ��� �� ��������� +��������� &$� 4"=4j+ ��/��� �+ �+� 8�/���� )+ �+� p ��/��� �+ k+ :$%%!;+ 0��� � ���� � �� �����e ���� � � ���� ��� ������� � ��� � ���+ ��� @�� � �� 4����. �� +�������� ��" �������� �� ������� ���+ ��� � ,+ k+� p � � ��� )+ �+ :!jj(;+ �� �������� �� �� ������� �� /��� �* �������� �������� �� 2��� ��� �������� ��� ��� �� �����+ ��������� � ����� 4����.� 8:<;� $$(=$<$+ � ������� � 0+ �+ :!jj<;+ ,�� ���� �� �* �� ������ �� �������� ��� � ���+ ��� %�� ��� �� �� 0�� ���� ���� � �� !,� ��" !��!� )�� ���� "%="4+ '����� 5+ #+ :!jj4;+ ,�� ��� �� ��� ������� �� �� ���+ ?0'�0 0"���� � ���� !<=!>+ '� ��� �+ �+� ' ��� �+ c+� p #�� � �+ #+ :!j>j;+ � ���*� ��� ��� �� �� ��� � �� �* �� ����+ ��� 4����. �� ��������� �"���� 5:a;� a<!=a($+ '� ��� �+ �+� #�� � �+ #+� p r����� 8+ 8+ :$%%!;+ g� � �� ���� �� � ��� ��� ���� � ��� /� � �*���� ��� �*� �������� ����� ���+ >� ���2 '��� � �������� ������ i������� �+ '� ����� )+ �+� p #������ )+ c+ :!jj<;+ � � � � ����* �� � �� /� �� �� ��� �� ����� �����1� ���+ ?�+4� �����+ 6���� ���� �+� p ���� ������ 3+ :$%%!;+ 2������ �� ���� �� �� ����� �� ������ �� ��� �� �� � ��� � + 0�� ���� +������� 4����.� a&:<;� !!!4=!!$"+ 5���� �+ ,+ :!jj>;+ ,�� ���� ��� � � �������+ �� 1��� 6��� ���� !%4=!!%+ 8����� )+� p #��� ,+ �+ :!j>>;+ ���� ������ ���� �� �� � �� ��� ������+ ��� 4����. �� +������� �"���� 66:a;� <"j=<4>+ ����� � )+� p �������� k+ :!jj$;+ h����� ����� ��� ������ �� ���������� ����� � ��� � ��� �� ��� � ���� �� ��� � �+ ��������� � ����� 4����.� 5:!;� "!=4!+ k���� 8+ :!jj$;+ # � ��� � �� ���� � ��� �*�� �� ��� �� # ���� ��� ��������+ #���2���� ������� $b:<;� """="ja+ k�/������ 0+ 5+� p �� ��� '+ �+ :!j>>;+ ,�* �� ���� ��� ������ � ��� �� � � ���+ %�� ��� �� ��������� ��" @��� � � ��� 0�������� 5$:<;� a>(=<%%+ ��������� �+ �+� p ������� #+ 3+ :$%%%�;+ ,�� � ��� �� ���� ��� � ��� �� �� � ���� ��� ��� ��� � ����9 �� �������� ��� � ��� ��� �� � ���+ ��� %�� ��� �� 4��� ��" (��� ����� 86:$;� ��� �+ ��������� �+ �+� p ������� �+ :$%%%�;+ ��� � �1�1��� � �� �� ������������ � � �2��� .� � �� 1�� �""�+ h�/ l���� # ����� � �������� ��+ h�/�� ��� '+ :$%%!;+ ,�� ������ � /� � ��� ������� �� �� ���+ %�� ��� �� ��������� '������2� &�:!!;� !a=$(+ !4% # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& ������ �+ :$%%%;+ ��� ��� � ��� ���� � ��s �+ �c 4+d(+>� ) ��� !%a=!%j+ �� ����� )+ �+� #������ )+ c+� p #��� � �+ :$%%%;+ ���� �� � ��� ��� �� ��� �� ������+ ?�+4� h��� ���+ ��� � 0� ���� �� ������� :!jj(;+ ��� ����� �� �� &aa� �;���� ����� �� �� d0 ��� �� �+ ��� ���� ���� 0+ :$%%%;+ �� �������� �� ���. ���9�� �*��������� ��� � ���+ ��� %�� ��� �� (���� ��2� # ��� !=!$+ # ������ �+ #+� p 0������� �+ �+ :!jj$;+ ��;�� ��" 1������� � � �2�0 �������2 ��� ����+ 3����/��� ������� h)� ���� � � 8���+ # � + )+ #+ :!jj(;+ �� � �����2 �� ��9� ��" "�� �1� ��� �� �� ��&)�* �; 1���e �0� ���� ���� ������� ��� �� ��� �� ���� � ����2 ��� �+ i�� ������� ��+'+ '����� � ���� # ������ i������� �+ �ghh2h5 2�� ��� � ������ � ��� � ��� � ���� :!jj(;+ d� ��1�� ����� ���>�c�� ���� f�� 1�2��+ 8�� ����� �,� ������� p �� ����+ ,��� ��� �+ '+ :!jja;+ h��� ��� � �� �� ���� ��������� ��� � � ��� � �� ���� �+ ��� %�� ��� �� ��������� ��" @��� � � ��� 0�������� 5b:$;� a%!=a!!+ ,������� �+ c+ :!jj!;+ 2���� ���� �� ������ �������� ��� � ��� ��� �� � ���+ %�� ��� �� 0����� ����� ���� (!=((+ ������ �+ h+ :!jj>;+ h ���� �� ��+ ��� ��. %���� (���� ��� 0"���� � ��� �� aj=<"+ # 0 #�������� � '����2��� � 3 ��������� � ����� 4����. &� )5��&* &�67&8& !4! variable annuities versus mutual funds: a monte-carlo analysis of the options introduction the existing literature the certainty equivalent of utility (ceu) a deterministic analysis of va versus mf the stochastic simulation comparative statics conclusion and final remarks acknowledgements references pii: s1057-0810(02)00089-6 �������� ��� � �� ��� �� ��������� ���� �� ������������ ����� ������ ������ ���������� ����� �� � ����� �������� ������������ �� ���������� ��� �� ��� �� �� ���� !�"#�� �$� ������� � !����"�� #$$�% �������� �� ������� ��� #$ !����"�� #$$�% ���� ��� �& '����"�� #$$� �������� �������� ��� � �� � (��)��( ��"��� �� ��� *������� �������� +���� ����� ���� "��� ����� �� �� ������� ��� ������,� ��� ��"���� ����� ����� �� ���� ������ �� �����( �������� ��� � � ������ -��� � �� ��������� ��� �����.���� � ���� � ����� ��� ��� ��� ������ � ���������� ������� � �������� ��� �� ��� ��� ��� ���� ��� ��) ��������� ������� ����� ������� ������ �������� �� ������ �� ��� ����� ���� �� ��� �� ���� � ��� � �� ���� �� ��(�������� ���� )���� ������ ���� ������ �� )��� �� �� ��������� �� ������ �������� �������� �� ��������� � #$$� �������� /������ ���� *�� ��(��� ��������� %&' ����� � ��� �( 0#�% 0#1% �#&% �2# )��* ���( �������� ��� �% �������� ���� � ���� ������ -�� ������� ����������� ��) �������� ��� ��"��� ��� ��� ��) ���� � ����� �� �� �� ������� +��� ���� ���3� �� ���� ��� � ���� �� ��) ������ ��� �� ������"�� �� ��������� )�� ���� "��� �������,��� 4�� '��()�� /�� �� )�� �� ��� �� �� ����( �� �������� ��� � ������ ���� ��������� ���� ��� "��� ������ �� � ���� �������� �� ��� �������� �����( �� �������� ��� � ��������3� )������ ��� ������( �� � ����) �������3� ������3� �������� 5� ��� ���� ����������� ���(�� � )��� ���� �� ��� ���� ����� �� ���3 ������ �� �"���� ���� ���� 6��7�$$$ )���� � (���� ��� �������� 89���(���� #$$�:� ����� '�������� ��� ����� ������� (�� ����� �� � �����(��� ��� ���� ���� ��� � ��� "�����3 ���������� �� �"���� � ���� ����� -�� ���� "��� "����� ��������� )��� ��������� ������� ������( ��� � ;�� ��(�����( ���� � ��������� ��� ������ ��������( ����� ��� ������ �������� ���"��� � ��� ;���3 �������� * ��� ������( � '�������<������3� ������� ������� 4�������� /������� ����) �$ 8#$$�: ##�=#&7 �$7>?$1�$<$�<6 = ��� ���� ������ � #$$� �������� /������ ���� *�� ��(��� ��������� @�� / � $ 7 > ? $ 1 � $ 8 $ # : $ $ $ 1 a ? b � 9����� �����( ������� -��� ��?&&$?a>#?1#2&% �c ��?&&$?a>#?7a>$� &!���� ���������( ���d���������� 8���� ����������:� ������d���������� 8���� �����:� � -��� ��?&&$?a>#?72&b� ��)��� �� ����� '�������<������3� �� ������� )��� ���� ��� ���������� ���� ��� ��� ���� ��(�� ������ �������� ���"���� ��� ��� ����� ������ ��� ��� ����������( ������ )������� �� ������� ������ 8���(� �aaa:� ����� 9�������� ��)������� ��� ����� ��� ������ �������� ���"��� � �������� � e�/� �������� � ;���� ��� � +�" ����� ��� ���� � ���� ������ �������� )��� 4���� e�/�*� 5��� f������ ���� ���� ����� �������� �� �"���� � ��c������� 6�b��$$$ )���� � ����������� 89������� 4�������� /�������� #$$$:� /����� /������� 5��� ����� ���� ��� ����"�c��� ��� "�c��� ��� ��� � ;���� /�� ���� ��� �� �������� ��� ��� ������ ���� �� ��� ���� ������ ������� ��������� "�� �� ���� ������ ��) �������� ��� �� ���� ������ ������ 8!��3� *���� *�������� 0������� #$$$:� -�� ��"��� �� �������� ��� �� )���� �� ��;��� �� ��� �"������( ��� ��� � ������� ����������3� ������� ������ ���( �� �������� ���� �� ������ "���� ������ ������ �������� ���"���� ����� �� �"���� ������ �� ���� ����������3� �������� �� ������� �� �� ����"���� ��) �������� �� ������ �� ��� ������3� ����� )������ ��� ������3� ����������� �� ���)���(�� ��� ��� ��� � ����� �������� ���<�� ������ �� �"���� (���� ���<�� ��������� -�� ��"��� � �������� ��� � �� ������ �������� ��� )����� �� ������ +��� �� ��� �� )����� �� ����� �� ��� ���� � ��������� ��� ���.���� ������ ������"�� �� ����� ����������� )�� ��� �������� �������� ��� � ��� "��� ������"�� �� gg-��33 ����� � ��� �� �������� �(�� /��� ���� �c���� �� �������� ���� ��� �� ��� ������ ��� "��� ��� ������ � ���� �� � � �������� ��� �� )��� ��� e�/� @�"��� �������� 0��� ���������( ���� 2$�$$$ �� �� ��� ������� � ���� �������� ���� ����� f�� ��) �� �������� � ;���� ������ gg� � � �������� ����� ���� ��� �������� �� "������( ��� ���� ������� ���?������� ����� � � � ���� *������ ����33 8����� #$$$:� /��� ����� �� ��� >$$�$$$ ����� � �������� ��� � ������ 84����� #$$$:� -�� ���� � ���� � �� �� ��� �������� �� ��� � ��� ��"���� ��� ��� ���� � ��������� ��� ������ �� �������� 5��� �� �� ������������ ������� ��� ������� � ���� ���� �������� ��� � ��� ���� ��� "��� ��� � ��� ���� �� �������� �� �������( ��� ��������� � ���� ��� ��� ��� �������( ���������� ������� ���"�� �������� �� ����� ��������� � ��� ������� � �� � ������ ����� �� ��� ���� �� � ��.���� ��� ���"��� � 9��(����� ��� e����� /����� 0������ *��������( f ;�� 80*f: �� ���� � "���;�( ��� ������� �� ��� �������� gg�������� 4����� �� �������� �� @���������� 9���� ��� �������� �� ��� �� �������33 8�aa1:� -�� 0*f �� ��� �� ������� ��� ���� ����������� �����"�� �� �������� �� ��� �� ���� ������������� � �������� ��� �� -�� 0*f (������� �� �������� ��� ������� 4������ �(������ ��� � ;���� ��������( ��� 4������ 5����� � �������(������ ��� /����� /������� ��� �������� ������ /������� ��� 4������ -���� 9���������� ��� ������� �� ���� �"������ �� �������� ��� ��� ����� ����� ������?�� �����( �(������ 8�.�� �c� �c ������ ��� -����e����:� ����� ������?���� ��� �? ����� ��� ������� ����� ��������������� f ;����� �� h�/* e�/�*�� ���� ��� ������9��� �������������� ���� ��������� ������� ���� ������ �� ���"�� "���� �� ��� �������� � �������� � �������� ������9��� �� ����� ������ � 62$> ������� �� �aa>� )��� abi � ����� ������ �����"���� �� �������� ����� *�������( �� ��� e�/� /����� /������ ������ ������ �� ���������� ������� ��� ������������� ������� ��������� ### %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ��� 622# ������� �� ;���� ���� �aa7� �� 627$ ������� �� �aab� �� 6>27 ������� �� �aa>� *� � ;���� )��� -����e���� 9�� �� ��� � ��� ����� ���� ������ �� �����( �(������� ��������� ���� �)�?������ � ��� �������� ����� �� ��� ��� ���3� 4���� h����� *��������� '� ������� �������� �������� ����� ��� ���� ��� ���"��� � ��.������ ��������� ��� &7�#&7 �� �������� ���� �aa# �� 7##�a## �� �aa>� h�/* e�/�*� ����� ���� )����� ��� ������ ���� �������� ����� )�� �� �������� ��;������ � �������� ����� "�� ������� ���� �� �aa>� ���"�� "����3 ���� ������ ������� 62a$ �������� 4��� ������ � ����� ������ )��� gg ��������� � ���������33 8"���( ���� ������� ��������( �������� ����:� )��� ������� bi � ����� ������ "���( gg������� ���������33 8)���� �������� ���� ��� �:� '�� ��� ��� �������� � ������� ��������� ��� ������ �� �������� $�$a>i � h�/* e�/�*�3� "������� ������� ��� ���� $���i � ������9���3� ������� 5����� ��� ���� ���������� � ����� ��)����� ��� 0*f �� ��� �� ����)����� �� ���� �� ����� ��� ���� � ��� ��"���� �� ��� ���� �������( �������� ��� � ��� ��� ������ "�� �� ����������( )��� ����� ����� ���� �������� ���� ��� "� �� ������� �� � ������� � ;������� ������� !� 4������ �(���� ��� ������� ������ ������������ �� ��� �������(����� � ���� ����� �������� ���� �� �� ;���� �� ����� "������ ����� �� �� ���������,�� ��;������� *���� ��� ��� � �� �� �� � �������� ���� ��� ���(� ��� ���������,�� ��� � � ������ ���� �� � ����� �������� � � �����3� �������� 80*f� �aa1� � #:� f� ��� ���������� ������ ��� gg�����33 ����� � �������� ���� ������ "� �����)���(��� ��������� ����� ��� ���������� )��� ��������? ���� ��������� �� )��� �� ��� ���� ��� � ��� ��.����� �� �� ��� � ��� ������� ������?�������� f�� /����� /������ ;��� �(��� ���� �� ���� ������� � �������� ���� ��� ���� ���� "��� ��������� *�����(� ��� ������ .�����;��� ��� ;������� ���<�� � �������� ����� �� ������� ��� ���� "� ��"��������� 4�� �c�� ��� ��� ������� ��� "� ���"�� �� �"���� � ��"� ������� � ���� �� .���� � �� � ����(�(� 80*f� �aa1� � 2a:� j��� ���� )��� ��� ������3� ����� ��� "� .�����;��� ���� ����� ��� ���� �� � ������ ������� f�� � ��� ��.��������� � ��� �������� -�� � ��� *���� ���� '��������� *��� �aa1 8�1 e�/�9� k�$#1� @�"� ��) �$7?&�1� ��# /���� &$$>:� �� ���� ��� 4������ -���� 9��������� 84-9: ����"���� ��������� �� ��� ����� � �������� ��� ������ � �������� ��� �� ������ ������� )��� �� ���������� ���������� ��� �� �� ��� �������3 ��� ������ �� � �� ����� ��) �� �������� ��� �������� �� �����( �(������� *� � ������� �� !����"�� �� �aaa� ��� 4-9 ����"������ ��� �������� -�� � l������ ��� '��� 9������(������ 5��)��� ��� ���� � ���� ���� � ��� l������ ��� ���� &$� #$$� ��� 4-9 ��((�� ���� a>�$$$ �������� ���������( ��� �� �����(� � 227 ����� �� )��� �� !����"�� �aaa �� ���� �1$$ ����� �� )��� �� ���� #$$�� -�� ���� ������ ���� � �������� ��� � �� ����� �� ��� l������ )��� ������ ���� ���� 82&i:% ���������,�� ���� �� ������� �������� 8#�i:% "��� ���� 8�2i:% ��������� ����� 8>i:% (��������� ��������� �� "���;�� 8>i:% ��� ����� �������� ��� �� -��� ���� ����(��� �������� ������ � ��� ������3� �������� �� (��� �� �������� �"���� ������� ��������� ����� ��) �� ��������� ������� "����� ���� ����� �� ����� ��� � �� �� ������ �������� ��������� -�� �����(� ���� "��)��� ��� �������� ��� � ��� ��� ���� � ��������� )�� �# ������� "�� 27i � ��� ������� ���������� ��� ��� � )����� � ������ -�� �� ��� ����� ���� gg� � � ���� � %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ##& ��� ������� )��� ���)��� � ��� ��� � �� �� ���( �� 7 �����33 84-9� #$$�� � 2:� @���� � �� � ������ � ��� ��������� �"������ ��� �"��� �)������� � ��� ��"���� b2i � ��� ������� ��� ������� ��������� ������ "������� ��� a>i � ����� ��� ����� gg4���� *�����33 �� ����� ;���� f �������� �� ��� ��� ���� )��� � ������ �� ����� ���� ���� ��� ��������� ����� ������ �� ���� >#i � ����� �������� )�� � ����� �� ��� (��������� j��� ��� ��� ��� ����(������ � ��� ������� ��"����� � ������ � �������� ��� � ��� ���������� ����� �� ����� �� ���.���� ������ ������"�� �� ���� �� ���� ��"����� � ��������������� �� � ������ ����� 9������ �� ���� � �������� ��� � ���� ���( "��� �������� ����������� @��� � ��� ��"��� )��� �������( ��� ������� ��������� �� ���� ���� ���� ������� ��� ��� � ���� ����� �������� ��������� -���� ���� ���( "��� �������� ����"����( ���������� ���� )� ������"� ��� ������� ������ �� �������� ��� �� �� "��� ��� 4������ ��� /���� ����� 8�1 e�/�9� k�&22 8"��� ����:� �1 e�/�9� k�&2� 8���� ����::� ��� ����� ��� �������( �� �������� �"��� ���� ��� ��� ��� �����( ��(�������� �� ����� ��� �� ���� � ���� ��� �� ��� � ��� "��� ��"�������� +��� ��� .������� �� ����� gg���3� ������ ���� ���� ������� ����(��m33� ��� ���)�� ���� "� ggj��� "�� � � �33� ������ ���������,��( ������� �� ���� � �������� ��� � ����3� ������� ��������� �� ��� �� ���� � ���� � ��� �� ��� ��� ���������,����� ������� ������� � ��� �������� ��� � )��� ���.���� ��������� � ����� ���������� �+1+ $���� � �� �� ��� ��������� � �� �����( �� ��� ������������ � ��� ��"��� � �������� ��� �� ��� � ��� �� gg ��( �����33 �� ����� �������� ��������� ���� ������ ���� ��� ��� ��)� ���� ���������,� ��� ��� � �������( �� ����( ������� �����3� ��������� +���� �� �c��������� �� ������ � ��� � ��� ������� ����� �������� �� "����� ��� ��� � ��� �� ��� � ���� ����������� ���� �c��������� � ��� ������3 ��������� �� � �� ������ � ���� �� ����������� ���� ���� ���������,����� �� �����.���� �� ������ ��������� )��� ���� ���.���� ���������� 4�� ��� ���� ���� ��� ������3 ��������� ���� "��� �� ��� � �������� ���� �����,� ��� ��������� �� ��� � ������ ���( �� �������� �� � ������� � ������ �� ����� -��� ���(� ��� ��� ���� ��� �� �� � ����� �� ���� ��� ������� ������� ����(��� �� ������ ��������� ���������� ��� ��������� "����� ��� ���� ����"������ ��� ��������� *� �� �c�� �� � � ��� ��� � ������ f��� ������ 9��� k#a�&�2a ����� �� � ;��� ��(��� ����������� �� ������ ggnfo"����� ������ �� ��� ��� ������� ������ ���( �� �������� � ��� �����( �� ���� ���������� )��� ��� ������ �� ����������� �"���� ������� �� ����� �� �������� �� ����� ��� ������ � � ��"� �� ��� ����� ��(�� �"��(�����33� @������� ������ ���( �� �������� �� ���� "������ ��;��� �� ������� ����� �������� ���� ���� ���"��� ������3� ������� ���<�� ������3� ������� ���"��� ������ �������� ���� ��� ���"��� ���� � �� ������� ���<�� �� ����� �������� ���"���� ������3� ������ ���� 8 ��"�"�� ��� ���� ��.������ ���� gg�� �(����33 �� ��������:� ;������� ����������� ������� ���"��� ��� @�!� ��� ���)����� ��� ������ ���� ���"���� �� ���� ����"��� �����( ��� �"�����(� ��� ��� �������� � ���� �� �������� �� ��� �� ��� � �����( ��� �"�����(� � ��� ��������� �� ��� � ##2 %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� � ������ � ������� 8����� ����� ��� ������� ���������� �������� �� ��� �������� �� ������� �� � ������� � �� �� �� �� ����: � ����� �� �������� �� �((��(��� ��� � ��� ������� � ������� �� ���� ��� �������� �"������ �� �� ���� � ���������(� *� � ���� �� "� ��� ���� �� ���� ����� ����������� ��� �� ������ )��� ����� �������� � ������ ��� ����� � �������� �������� ��������� ��������( �� ��� ����� � ������� (����� �� �������� �"������� � ��� ����� � ������� (����� �� �������� �"������ �� 6�$$�$$$ �� ����� ��� �������� � ���� �� � ����� ��(��� ����� 8� ������ ����� � �������� ��������� )���� �������� �������� � ���"�� � ����� � �� ��������� ����( )��� ��"�������� ;���:� /��� ������ ���� �� ���� ��(�������� ���� ��� ���� ���������,�� �� �������� ��� �� "�� ���� ���� ���������� ��������� �� ���������� *� �c�� �� � ���� � ����� �� ����� *������� 0������ *� �����3� ����(� � �� ����� ��)�� )���� ������� ����"������ �(����� ����� ��������( �� �������� ��� �?����� ����"����( ���������� � ������� ������ �� �������� ���� �������� ��� �� �������� ��� �� ��� � ��.�������� ���� ������� ;������� �� �������� ��� "� ���� )������ ��� ��������3� )������ ���� �������,����� 8����� 9��� k#1?7�?�$�:� *� ��� ����� ��� � ��� ���������,�����<��(������� � ������� ��� ����� ���������� �� ��� /���� � 9��� ������ )���� �� �aa> "����� ��� ������ ����� �� ���� �������� ��� � � �������� � ����% � ����������� �� ;���� "�� ��"��.������ ������� �� �������� ��� � ���� �� ���� ��������� �� � ������ )��� ;��� � � �� 6�$�$$$ ���<�� �� ���������� �� ��������� � ������ � �������� ��� � ��� ������� � ����� �� "� ���� �������� �������� � ��� ����� ��������� "� �� �������� ���� ����( ��� ������3� ��������� ��� �� ���� �c ��(� ���� ������ ��� ����������� ���� ��������� 9��� ����� ������� ���������� ��������� ��������� ����(��� �� ������ ���������� ��� ��� ��������� �� ������ �������,�� "� �������� �������� e���� 9��� ����� ��) 8@���� 9��� 7&$�b:� gg* ����� )�� ��� ������� �� �������"�� ��� ���� ���� ��� �� ��� ������� ������ ���( �� �������� ��� "��� ����) ���� ���� "� ������� � � � ��� �������� � ��) �� �������� �������(����� "� ���������( ��� ����� ��) �� �������� �(���� ���� ��� ������������ ���� ��� �� ��� ������ ���������� )���� ����� ���� � �� ��� � ��� ������ � � � ��� "�(�� �� �������(����� � ��� ���� � � �33� e���� ���� ��������� � ������ � �������� ��� � ��� ������� gg� �����33 �� � ;����( � ������ ��������� � ����"�� �� ���� ������� ��� ��� ����� gg����� �����33 �� ����� ����� ���( ���� �������������� � ��� ����� �� �� "� �������� ��������� ��� �������� 9������ � 9��� ����� �� �� ������ � ��� ��� ����� �� ��� ������ �� ��� �� ������( ���� ������� � ���� � ;����( �� ����� ���� �� "� ��� ������ � ��� �������� ����� ����������� �� ����� ��� ����� ��������� ��� �)�� �� ������ ��� �������������� 9��� ����� @���� 9��� 7&$�>8�: ��.����� ��� ����"�������� � � ����"��� �������"�� "� ������� ��� ������ �������,��� ����"������( ��� ����������3� ������ �� � ������ � �������� ��� �� -� "� �������� �� ���� ����"���� � ������ �� ��.����� �� ��"��� ��� � ����� �������� "� ����� � � ��� ��� � ;�(�� ������ ��� ��� ����� �� �������� ������"�� "� ��� �� �������� -�� ������ "����� ����� ������ ���������� �� �� ������ �������� � ������� �������"�� �� ������ ��� �������� �� ����"���� ����� ������� -��� )����� ������"��� ���� �� ������ �� ���� �� ����"���� )�� ���� ��� ��� )��� ���� ����� �� ���� ���� ���� �� ��� (���� ������������� /� ������� ���� )���� ���� � ���� ������ ���� �������( � � ������ ������� �� ������� ��� �������( ��� ���"����� �� �� ����) ���� ����"������ ��"� �� ������� "� ��� � ���� ��(���������� /��������� ������ � ������ � �������� ��� � ���� �� ��������� )��� � ��) �� �������� �(���� ��� �� � ������ � ���� ������3� �������� 8����� ��� ���� �����( �� ����������( )������� �� �� ������ ������ ����( ��� ������3� ��������:� �� ������ "� ������ �� %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ##7 ��� ������ �� ����� ��� ����������� ����� �� ����������� )��� ���� ��) �� �������� �(����� '�� ��� ��� "������ � ��� 9��� ����� ������� ��� ��� ����� ��� ���������� ���(�� �� ������� � �������� ��� �� ���� ������������ ���� ������� )��������� 8/������� #$$�:� �+�+ 5.� ������� 6������� 5.� � ��� �������� � �������� � � �� 177" 21" �+$+�+81#�"� 9�:+ '�* 1# !�1"� 11� $���+ �##;3 �� �aa>� /����� 5��� 7�# )�� ���������� "� /������ ���� ��� 8 ?*��,���:� * ������� ������� )�� ���������� �� l���� "��� l� � 2�7�� ���)� �� ��� �������� -�� � ��� *���� ���� '��������� *��� �aa1� �� )�� ����� "� 9��(���� ��� ��(��� "� @�������� 9������ �� f���"�� &$� �aa1� )��� �� � ������ ���� � ������� �aaa� -�� *�� ������� ������� ������� � ��� ��� 2> � -���� �1 � ��� e����� /����� 9���� ����"������( ��� 4������ ����� � �������� ��� �� ����"������( ��;�������� ��� �������� �� ��������� � ��� *��� ��� ����"������( ������� ���?�������� ��.���������� ���� �� ��� ����"�������� � ��� 4-93� 9������(����� 8�1 e�/�9� k�$#1� @�"� ��) �$7?&�1� ��# /���� &$$>:� +���� ����� ��� ������� ����� ��������� � ��� e�/� 9��� �����,��( ������� ���������� ���� ��� ������� ����� � �������� ��� �� ��� *�� ����� ������ � ���"�� � ����(�� 4����� �� ������;�� ������� � �������� ��� � �� ������� � ���� ��������� ������ )������ ������� ��� ��������� "��� ���)�� ������ �� g���������3 �������� )��� ��� �� ������ ;������� ������������� ����� "���( ���)�� �� ��� ������ ������� � ���� � ��� �� /�������� �� 4�������,�� ��� ����� � �������� ��� �� ����)��( ������� ��� ��� � ��) �� �������� � ;����� )��� ��� �� ������� � ��� ��� � ��� ���� ����� ���� ��� ��� �� ��� ������3� ����� � ���������� -������� ��� ���� � ���� �� ��������� �� ���� ��� ���� ��� � � �������3� �������� � ������ �� ��� � ����� � ���� � �� ����� ��������� � ��� e�/� 9��� )��� ��������� -�� *�� ������ ���� �� �������� ��� � �� ��������� "� ������ )�� gg���)��(�� ����� ��� �� ����� )������ ��) �� ���������� � ����� � ������;������ � ������� ����� )��� ��� ������ �� ������� �� �� ��� �� �"��� ��� ����) �� �������� ���� ����������� � ��������� � 4������ ��)� �� ���� ����������� � ����� ����� ��� � ����"�� ����� �� ����� ��)33� * ����� � ������;������ �������� ���� ����(� �� �� ����������� ����� ������ �������� ���"��� ���� � "����� "�������� ���� ���� �� ;�(�� ������ ��� ���������� ��� ������������������ ������ -�� *�� ������ ���������� ��� e�/� /��������( 9��������� �� ����� ��� e�/� /��������( 0���������� �����( ���� ������������� ���� ����(� �� ��� ���"�� � ������� ��������� ��� ���"�� � ������;������ ��������� ���� "� ��� �� �������� ��� ������ � ���� ��������� ��� �� ;����� ���<�� ����(� ���� �� ��� ������3� �� ������� ��� ����������� � ����� �� �������� ��� �� (�������� ����� ����� ��� ����� � � �������� ��� � ���� �������� ���� ������� 0���������� �� �� (���� � ��� ������ � ���������� �� �� )��� ��� "� ���������� 8#1 e�/�9� kaa2 ����:� 9������� ��������� ������� ;��� ��� �� ���������� )��� ��������� ���(��( ��� ����� ����� � �� #7 ����� � ��� �������� ��� � ���� �������� ���( ��� ;����( �� ��������� �� � ��� �� ������� ��� � ���� ���������� 8�1 e�/�9� k�$#18": 8&: ��� 82::� �� /������ 7 � ��� *��� 9��(���� ���������� ��� 4-9 �� ��( ��� �����)���(� ����� � � ��� ������ ��� ����������� )�� ������ ���� �� �������"�� "������� ���� ���� "��� ������� � �������� ��� �� ������ �� ���������� ��������� �� ����� �������� ��� �� �� ����� ����������� �� gg� �� ����� ��������33 8��������( ��� ����� ���� �������� ������ �� �����( �(������:� -��� �� ��� (������ � ��� 4-9 9������(������ ##b %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� �� ��� ���( ��� *��� 9��(���� ����� ������ �)� ��(��;���� (���� ��� 4�������,����� � � ��� ����(������ � � �������� ��� �� �� � ����� �� ��� � ����� % ��� ��� ����"�������� � � ��������,�� ������ � �� �������� �� ������� ��� ��) �� �������� �������� 8��� ������: �� �"���� ��� ������ �������� ���� ��(�����( �� ���� � ��� ������ l�)����� ��� ��� ��� ����(������ � ������ ��� ��� ������ ��� �������� ������"�� �� ����� � �������� ��� �� ��� *�� ����� ���� ��� ������ �������� ������� ���.���� ��������� ��� ��������� � ��� ��� ������ �� ��������������� *� ���������� ��� ������ 4������ ��� ���� �� ��� ��"��� � �������� ��� � ��� "��� �� ��;�� �� �� � 4������ ����� ��� ������ �������� ��������� �� ��� ���������� ����������(�� ����(�� ��)����� ��� ;��� ������� � ��� *�� � ���;����� �c������ ��������� ��������� �� � �� ���� ������� ���� )���� ���� ����� ������� � ��� ����� "� �������( ����������� � ��� ����� ���������� )��� ��������( ����� ������ ���������� ��� �������( � ��"���� ������ "� �������� ��� �� *� ���(������ ����������� /�5�7�# ��������� ��������� ����)��( 4������ ������ �� �)��� ����������� �� ����������� )�� �������� �c ����� �� � ������ � ����� ���������� "���( ������� -�� l���� 5��� ��� ��� ������� ���� ���������� ��� ���� )��� �"���� ��� ��� ������� ��� ���� f��� gg������33 ������� � ��� ����� 8���� �� ;������� ������������: ��� "� �)����� ���� ����������� �� �����( "��� gg ��c������� ������33 8�1 e�/�9� k&bb&8�: 8#::� -�� ������ �� ������� ���� ��������� ��������� ��� ���� ���� ���� ������� � �������� ��� � ���� � ��� ������������ ���� ���� ��� ��� ��� gg����33 ������� � ��� �����% ���� ��� gg����33 ������� ��� ����� ������ ���� ��� ����� ��� ����� ;������� ������������ ���� ���� �c������ ������� �� ���� ����� �� ��� �� �������� � ��� ��� �� *�� �� ������ "� �������� ���� �� ����� � ����� ������ 8����� ����� ������ ����<������: ;������� ������������ ��� ��� ���� gg�������33 � �������� ��� �% "��� �� ��������� �"���� ���� ������ � ������ �� ������� � ���� ����� �������(� � �� �����������3� "�������� *�� �� ��������� �� ���� ���� �� ���� ��������� �� ���������� ������ ��� ������� ���������� ���� � "� ����� � � ������� ���"������ )��� �� �������� ���� ����� ��� � �� �c�����( �������� -�� "��� �c�� �� ��(�� "� ����� ��� -����?��?������( *��� �� �������� )���� ������ �� ���������� ������� ������3� ���"����� �� 67$ �� ��� � � ���� �� ������� ������ ���� 8�� ,��� ���"������ � � ���� ���� �� �� ����� ���� �� ��3� "���( ����: 8�7 e�/�9� k�b2&:� /��������� ����� ��� 4��� 9����� 5�����( *��� ��� ���������� ��� ��� ��(�� �� ��� ��� ����(�� ���� � ��� �� � ������ ���� ���������� ��� ���� ��� ���������� ������8�:� �������� ����(��� ����� �� ��� ��� �� ���� ��� ���� ����� ���� �� �"������ �����( �������(����� ��� ���������� � ����� ��� ���� 8�7 e�/�9� k�b1�:� -���� �� ����� ������ "� ��(��� ���� �� ���������� ������ � �������� ��� � �� ��� ���� ��� gg����33 ������ � ��� ��� �% ��� ����� ������ � ��� �������� ���� )���� "� "���� "� � ;������� ������������ *� ������ ���� ������� ����� ��� ���� (�� ����� ����� ��� ��) �� �������� �(������ �� ���� ���� � �� ��� � ��� ��� �� j�� ��� ���� ���� �(����� ��� ���� ��������� )���� ������� � �������� ��� � ���� �c ����� ��������� �� ���� ��������� � ������� �� ������ ������� ����� ������ ���������� �� �� ��� ���������� ��� ������� ����� �� ������� ��"�� �������� "� ��? �������� )�� ���� ������ ���<�� ������� ��� ������� ����������� * �����) � ���� � ��� ����� ��������( �������� ��� � ������� ������������ ��� �����.���� � ��� ������� ����� � ��� ��)� %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ##> �+1+ ,����� �+ $������ 6� + 2�###3 217 $+<+�� "7# 25�=+ ���+33 '���� ������ )�� � ������ � �������� ��� � )�� )�� ������ ������ �������,�� "� � ;������� ����������� ���� ��� "��� �������,�� "� ��� �� �������� � ������3� �������� ��� �� f��� �1 ������ ����� ��� "��� � ���� "� ��� �� ��������� ��� ������ ��� ��� � ��� ����(� ���� ��� ������� � ��� �������� ��� � ��� ������� �� �� ��� ��������� ������ �� �����( �(������ �� �� ��� ��� ��� �� ��� ���� � ggp�(33 ����� �� ��� �� ���� -�� ggp�(33 )���� ����� ������ �(������� ��� ������ �c�������� ���� ������ )�� � ������ � �������� ��� �� ��� �� ��) �������� ������ "� � ���� ����� ��� �������� )�� ����;��� * ��� ��� �� ��� ��� "��� ggp�((��33� ��� �� ������ �� ��� ���� �� � gg ��?� �����33 ���� � �� ���� )�� ������ �� ������3� ����� / ��(�� � ���� �� ������� �� ������3� ����� )���� )�� ���� �� ����(� ���� 6#�$$$ )���� � ������������ ������ ���������� ����� �� �������� ��� �������(���� �� ��� ��� ����� ��� ������� )��� / ��(��� )�� ��� ������ ��� ������� �� ���������� )��� 49!5� f� ��� ������ � ��� ��������� ������ ;��� ���� �(����� / ��(�� ��� 49!5� �������( / ��(�� )�� ��(��(��� �� � ����( ��� ��� ���� �������� ��� ��� ���� �������� ��� 4��� 9����� � �����( *��� -�� ���� ����(�� ������ ������� )��� ��� ���������3 ��� ��� �������� �� �������( � ��� ��� ���� �������� ��� �������� �� ���������( ��� ���� �(����� / ��(��� / ��(�� ��� 49!5 ������ ������3� ����(������� ��� ������� ���� ������� �������( ���� ������3� ������ )��� "���(�� �� "�� ����� * ��� ��� ���� ��� "��� �����( �� � ����� ������ ��������� ��� ������� "������ ��� ���� ��� "����� ��� �c ������� �����( ���� ��� ��� �������� ���� 6�2�$$$ �� ���������3 ��� �� ��� ���� � ��� ���������� /�"��.���� �� ������3� ���������� ��� ����� ����� �)����� ��������� �(����� ������ ��� ��� �������� �� ��� ������ � 67�$$$� f� � ���� ��� ���(���� �� ��������� )�� ��������� -�� � ������ ����� ����� ���� ������3� ��������� )�� ��� "������ � ���� "�� ���� �� ������q�� )�� "������ ������ ������3� � ��� �� �������� ��� ����� /��������� ����� ��� ��� ��������� )��� �)� ��������� )�� ��� ��� ��� � ����� ������ ������3� ������� ;���( )�� ��� ���� �� "� � "�� ���� �� ��������� ���� ���� )��� ��� ���� ����(�� )��� ���(�� )�� �������� � ��� �������� ��� ��� ��� �c ����� �� ������ ��� ��� ��� )��� / ��(��� ����� ����� )�� ���� ��(����� ����� -�c�� ��) ���� ��������3� ��� �� ���� ��������� ��(�� "� ��������"�� ����� � ����(��� '���� ������ ��� ��� ���� ��� / ��(��3� �� � ��������� �������� "�� ��� )�� �������,�� �� ��� ���� � ���� 6�2�$$$� ��� �� )�� ������ 6#$�$$$� �+�+ >�?��� �+ 5 � �� , � � ������ � ��+ 217773 2;� $ +�� 11�� 2���+ ���+33 �� �a17� *�������� 0�,��� ���������� ������� ��� ������ ��� ��������� -�� ���� �"������ ������� ����� �� 9����������� ��� ��� ���� �������� � ��) -������ ��������( � ���� )��� -����� ����� 9����� 9�� � 8-�99: �� 6#1�b>#�#�� 0�,��� � ��� � ���� ��� � ��� �������( ��� ������ �������� ��� )�� ������� �� �c�� � �� ��� -�99 ����� ���� � ��� 0�,��� �� ����� ��� ����� -�99 ���������� ��� ������� �� � ��� �� �� ���� ����� ��� �������� ���� �� � ��� ������� �������� ����������� ��� ���� ��� �� ����(� ��� ��� ������� )��� �� �� ����� '�� ��� ��� ��� ���� 0�,��� ��� ����;�� -�99 � ��� ���� �� �aa7� )��� ��� ���� �� ������ �� �aa>� -�99 ����� ��� �� ������� ��� 0�,���� ���� �� ��� ���� � ;���( ����� 4��� ���� "� ��� � ##1 %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ������( �� � /������ ���(���� ������ 8 ������"�� ;��� "� 0�,���:� -�99 ��������� ����� ����� 0�,��� ���� ;��� ���� �(����� -�99 ����(��( ���� ����� ���������� ���� ����������� �� �"��� � ������ ��� � ��������� ����������� +���� 0�,���3� ������� ;���( ���� �������� ������ �(����� -�99 �� �������� � ������ ��� ��������� � 4������3� 9������� 9��������� @�������� *��� ����� ����(������ )��� ��"��.������ ��� ��� -�99 ����� �� /������ ���(����� )���� )�� (������ "� ��� ����� ������ �����( ���� ��� -�993� ������� gg)��� ��� ����� ������� ��� � �� � ������� ��"� ���������� ����( ���� �������� ����� ���������33� gg-� ���� �����)��� )���� �� � � � ��������� �� ���� �� �������� ������������ ����������� ���������� �������� �� ����� � �� ������ ��� ���� � ��� �������� � � ��"���33� -�� 9���� � * ���� �������� ��� ����� �����3� (���� � /������ ���(����� ��� ������� ��� ����� ����� �� ������� )������ �� ���� ��������� -�993� ���� )�� � �������"�� ���������� � ��� )��� �� ������ �������(����� ��� "��� ����� ��� )������ �� ��� gg� � � -�99 ����� �� ��� ����� ��� ��������� ������ �� �� ��� �������� �� � �� ���� ���������� ����33� �� )���� "� ���� �� ��(�� ���� -�99 )�� ��� �������"�� �� ;���( � ��� ������� ���������� ���� �(����� 0�,���� 9���� ��������(� ��� ����(��� �� ������ �c����� ����� �� �� � ��� ���� ��� ������ ��������� ��������� ��� ��� ������ (��������� �� ������ � ���� ���� ��� ���� �������� ���� )��� ��� ������ ���� � ���� ��� ���� "��� ���������� )��� )�� gg)���(33 �"��� -�99 �����,��( ��� ��������� ����"������ �� ��� �������� ������ �� ���� � ���� 0�,��� )��� �� )�� ���� ��� ��"���m /������ -�99 ��� ;��� ��������� �� ���������� � ���������� ������� ���������% ��� �����"����� �� ���� � ����� ������� ��� gg�� ���3� ��33 �� ���� ���� ���� "��� ������� +�� �������3� -�99 ���� ������ �� ��� ������� ��� ��� ����� ���� ������� �� ;�� ����� ��"��� ��� ������� ����� ��"�m +��� �� � ����� 0�,��� )�� ��� gg����� ���33� ���� ���������� *�� 0�,���� �� �������� �� � �����( �����)� ����� �������( ��� ���� ��� ������ �������� )�� ���� �� � )��� � ��(�� "��� �� �� �� ��� �� ���� ��� -��9� '�� ��� ��� ��� ����� ���������� ��� ��) ��� ������� 0�,��� )�� �������,�� �(���� �+�+ 9 �?�� �+ 50< 6� + 2177"3 2@"� a+b+$+�� 17�3 �� ��� ��� �������� @��,�� ��� �� ����"�� ������ ���������� -��� "������ ����� ���������� )��� ������ )��� ��� ��"������ � ���� � �������� �� � ��) ���� �� ���� ���� 2 ������� ��� �������� ���� ���������( �� ����(��( ��� @��,���3 ��������� )��� ����� "���� ��� ��� ��� ������ �� �����( �(������� ��� ��� � ���� 6�$$�$$$ �� ��������� ����(��� +���� ��� @��,��� ������ �������� ��� ���� ����� ��� � �� ��(��� ����� ������ �������� �� ���� � ��"�������� ���� �� ���� �� ���� � ��� ��� ������� 9����� �� �����( �(������ ���� ����� �� ���� �����������q)��� ��� ��������� �������� ��� ��� ���� 3� �������% gg ��?� �����33 ����� )��� ������� 9�������� ������� �� �������� )�� ����������� ��� �� ���� + ���� �$ ������ � ��� ����;������ �� ������ @��,��� ��� ��?� ����� ������ � �� ������ �� ���� ���� �(������ ���� ���� # ����� �� ������ ��������� ������� �� ��������� ��� �� ���� @��,��� ������� �������� j��� ����� ���� ��� ���� ��������� ��� @��,��� � ������� �� ���� �� �������� ������� @��,��� ;��� ���� �(����� � ���"�� � ��������� ��������( "���� 85��� � !�) j���:� ������ �� �����( �(������ 8+� #$$�:� ��� ������� � ��������� ������ ����� 8��"��:� -�� ����� %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ##a ����� (������ ������� �� ������� ��� � ��� �� �������� �� ������ )���� )�� � ;���� "� ��� 9���� � * ����� -�� � ������ ������ �� � ���� �������� � ����� ����� ���� !�) j��� ��� ��� ����(��,� gg��(��(��� ���"������ � �� ����� ����33� ��� ���� ��� ������ "���� �� ���� ��p������ � ��������� �������� ��� �� ���� �� gg����� )�� �� �������� � ��� )���� ������� �� �c����� �����(���� �������33 9����� ����(��( ���������� ����� !�) j���3� '��� ���� *��� r @�������� *�� )��� ���� ���������� �� ����� )�� �� ��� ���� ��� �� ������� ��(�(�� �� ��� ���� ���� )��� ���� ���� �� ���������( �� ������ �� @��,��� � ��� ������� ����� �������( � ����� ������ ��������� �� � ������ � �� �������� ���� % �����"����� �������� � ����� � ���� ��� � ���� �� ��� ������ ��� ��� �� ������ ������"�� �� ����� � ����� �+�+ 9���� � �+ ��� � $���� ���� 2177@3 2@"1 $ +�� 1�@� 2���+33 f ��� ��� ����� ����� ��� ������� ���� ��������( � ����� ������ �� ������� � �������� �������� ���� ��� ������ *��"���� ��� ������� ���� � ������� �� * ��� �aa$� 5���(���� @������ � ���� �� ����) ��� ������3� �������� ��� �������� � ��� ����� ������� ���� ��� �� ����(�� �� /������ )���� ������ @������ ������� ���� ��� ��� ����� ������� ��� � ����� ������ ���� )��� ������(� /�� �� ����� ��� ������( ������ ����� ��� ���������� �������� ��� �������� ����)�� �������� ������ ������ @������ �������� � ������ � �� ����������( ������ )������ �(����� ��� �� ������( "�� ������� @������ � ����� �� ��� ����� �������� ��� � ��� ��� ��(������ ��� ��� ����� ��� ���)��3� ��(������ �� ��� "�� ������ ��� ����(�� )��� ���������� -��� ������������ )�� �� ����� �� ������� ����� @������ "������� ���� ����� ���� "� ������� gg5���(���� @������33� /�� )�� )���(� �� �aa#� � ��� "���( ��� �� "� ����� �� � ��� ;� � ����� @������ � ��� �$ ���� �� ������ �� ��� )�� ��� �� ��� ������ �� ������ �� ��� ��� �� ������ )������� �� ����� ��������� �� "�� ����� ����(��� *�� � ��� ����(�� �(����� @������ �� �� �� ����� ������������� )��� ���������� ���������� *�� � ��� "�� ������ )��� ���)� �� e���� 5���� )���� ��� � ���� ��� ������� )��� �� �� ����� �������� @������3� ������ ��� ����� ������3� ������� �� ������;������� -�� �� ����� ������� ��� "��� ���� ��� ������� �� ��� ������� )���3� ��������� �� ��� )�� �����( �� � "�������?)����3� �������� -�� "��� ��.����� �� ���������� ������;������� ��� ����� )��� ����� ����(��������� �� ��� � ����( � ��� �������� @������ ;��� ���� �(����� e���� 5���� ����(��( ���� �� ��(��(����� ����)�� �� �� ����� �� � �� �� ������� ����( ��� ��������� )������ ��.�����( �� �� ������;������ �� �����)��� ���� ���( ��� �� �������� ������� �� � �� ��� �������� -�� ����� ����� (������ /������ ���(���� �� ��� "���� �����( ���� ������ ����� �� ����"���� ��� ���� �)�� "� ��� "��� �� ������ � �� ��� ��c����� ����� "��)��� ��� "���3� ����(�� ��(��(���� ��� @������3� ��������� -�� *��"��� /� ���� 9���� ����(����� -�� 9���� ;��� ����� ���� �� � (������ ����� ���"����� �� ��� �� ���� �� ��� ���� "���� � �� ��� �������� ������� � ������� ����� l�)����� ��� 9���� ����� ���� ���� ����� �� ����� ��� � ����� ��������( ������� ��������� )���� )�� �"������� ��� �������� �� ���� ����� l���� ��� 9���� ������ ��� "��� "����� �������� �� � ����������� )��� ��� ������ 8@������: "� � ����( �� ������� �� ��� ����� ��� �� )�� ��� � ����( � ��� �������� ��� ���� ��������� �������� ��������� )���� (��� ���� �� ��� ������ 3� ��"����� !����( ���� �� )�� �������"�� ���� � ����( � ��������� ������� ����� ����� �� �������� ������ ��"�������� ��"����� ��� #&$ %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� 9���� ���� ������ ���� ��� "��� )�� �� ��� "��� ������� �� ����� �� ������ ��� ���� )���� ��������� -���� ��� 9���� ����� gg��� �� ������� � ���� ���"����� ����� "� � �� �����33� -���� �� *��"��� �� ������ � ��� ������ ������� � �� �������� ���� �� �����(� ��� � � ����������� � ���� "��� �������� � ������ ��� ���� ���� ����� ������� l�)����� ���� ������( ��� "� ���� ������� ���� )���� � ��� �� ;��� "���� ��� 9���� ����� ���� �� ����� �� ��� "���3� ������ �� ����) ��3� �)� ���������� ��� �� ����,�� ��� ������"����� � g������3 �� ��� ������ 8����� ��� ������� ����: �� �����( "�� ������� l�� ��� �������� ���� ���� ��� ��������� ������� �� ����� �� ����� �� ������ �� � ���� �� ���� ����� �� ��� � �������� ��� �� ���� � ��� ������ ��(�� "� �� ������ *��� � ������� ����� ��� ���� ����� ����� ������"�� �� ����� �� ����������3� ��������� �+ + �����*� �+ 5������� � � ��+ 2�##13/50<+ 2; �+$���+�� 1# @ 2�+�+���� + 177"3� � c� �� ����� ���c� �� ����� �� �+�� 1#@� 27�. ���+ �###3� ���+ ����+� �� �+$+ 7#�3 f� ���� �>� �aa& *������� *����)� ������� � �������(���3� � ;�� �� /���� ������� 9��� ������ /���� �� )�� ��� ;��� ������ ��� )�� ��.����� �� ;�� ��� � @������ �� �������� 4���� ������( ��� ����� �������� ���� � "����� ��� ������3� ������� ��� ������ �������� ���"���� * ��� ��� �����( ��� ���� *������� ������ ��� ��� �� ��� ������3� ���� ������� *����� *����)�� /��� ���� ������ *����� ��������� �� ��� h�(��� ��� "�(�� ����( *�������3� ��������� /�� ������ �� � ������� ��� �"������ ������� �������� ����( ��� ������ ������3� ������� ��� ������ �������� ���"���� ��� ����� ��� �� �"���� ������� ������(� ��� ���� ������ ��� ;��� ���� ������� � �� �� �������� �� *�������3� �� ����� ���������� f�� � ��� ����� �� �� �"���� ������ �������� )��� *����� � ���� �� � '������3� ����(� ����� '������3� ��.������ ������ �� �������� ��� -����e����� )���� ����� ����� ������ ;���� )��� �� gg�����33� "������ � ������ ������ "��)��� ��� �� �������� ������� ��� ��� �� �������� �� ;��� '������3� ����)�� *����� ������ �� *�������3� ����� ��� ��� ������� ����� )��� �����.����� f���� ����� �� "� *����� �� �"���� ������ )��� �������� �����(� +� �� ��� �aa7� *������� ����� ��� �� ��;����� ��� ����(�(�� -�� ;������� ����������� 8l��� /����(�: �"������ � ������ �� ��� ��� 9���� 9����� ������� 8� ���? ���� �� *�������3� ����:� )���� )�� �� ���� "���� � �� �� �������� �������� ��� -����e����� +� ��� �c ����� 8���� ��� ����� �� ��� ����:� -��� �� ��� ������ ��� �����.���� '������3� �������� 5������ *������� "������� ���� ��� )���� ��� "� � ����� �� ��� ��;����� �� l��� /����� ��� ���� ����������� ;������(� ����(���� �� � ��(��� �������� ����� f� ���� 7� �aa7� *������� ��������� -����e����� )���� ����������� ����� � gg4���� *����33 p�( �� ��� ������ �� ���� f� ���� ##� � ��� �������( )��� '������3� ��� ����������( ���� ���� ������� )�� ����������� � ���� ����( *�������3� ������ ������;���� -����e���� �� ������ ��� �� �������� ��(�����( ���� ������� ��� ��� �� ���� -�� �� ��� ��������� �� ���) �� ��.���� ��� '������3� �� ��� ���� *����� � ���� �� ��� �������� ��)����� f� f���"�� #�� �aab *������� ;��� ���� �(����� "��� -����e���� ��� +� ����(��( ���������� � ��� 4��� 9����� � �����( *�� 849 *:� ��� � ��������� � 9��� ����� ��)� *������� ������� ��� �� ������� �������� 49 * "� �� �� ���� ���������( ��� ������ �� ��� �� ��� ���� �� ��� ��.������ ���� �� � ������ � *�����3� ��� � � ��� ������;���% �����( �� �������� gg��c���� ����"�� ��������33 � ��� �� �������� �� ��� �� ���% ���� ���� -����e���� ����� �� �������"�� ���������(��� ��� ��� ��� ������ ����� �������� �� ��� %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� #&� �� ���� @������ ���(�� ��� �������� ��� ������� ����(�� �� ��� �c �������� � ���� ��� � ��� �� ����� ��� ������� ��� �� ���������� �� �������� �������������� ��"����������� ������������ ��� ��������� �������� ��������( ������� ���� ��������� 8���((����( � � �� �� ������:� 5��� �� ������� ����� �� /������ ���(���� �� ������� �� ���� � ������ 3� ������� *�����(� �� )�� ��"��.������ �������� �� ��� ���� � ������ 3� ������ 49 * ��.����� ������ �� �����( �(������ �� ���� � ��� ���������� �� ����������� ��.������( �� ����� ��� ���� ���� �� ���� "� �������� ���� �� �� ��� ����"������� �� ����� 49 * ���� ��� ���� � #?���� ������� � ������������ gg� � � �c�� � ���� )���� � �� ������ ��� ���������� ��� )��� ���� ����� �������� ��� �� �������� � � � ��� ������ ��� "� "���(�� ��� ���� )����� # ����� � ��� ��������� � � � � ��� ����� �����������33� /���� �)� ����������� )��� ���� ���� ���� # ����� ���� �� *�������3� ����� ��� ����� ����� ����� ���� ����� ����������� )��� ��� �������"��� *� �� ��� ����������� ���� )����� ��� #?���� ������ ��� ����� ���� ���� � ����� ����� "� ���� ���� � � ������ ����� �� �� �������"�� �� ���� /���� *�������3� ;��� )��� �������� � �� ��� � ����� � �������� �� ������� ��� ����� ���� ���� ����� ����������� )��� �� ��� -�� ����� ���� ����� ���� ��� ������ �� �����( �(������ ��� ��������� �� �� �������(������� ��� �"������ �����;������� ��� ��� ��.������� ���� �� ��������( ��� �� ��������� +��� ��(��� �� ������ 3� ����� ���� ��� �� ��� ��������� ������������� ��� 9���� ����� ���� ������ ;��� ��� �� ���� ���� ��� �� ��� )�� ����������� ��� ���� ��� �� ���� ���� ��� ������ �� �����( �(����3� ��������� )��� ���������"��% �� �� gg��������33 ������ �� )�� ��� ��������� �� ���� ������ �������� ����� /���� -����e���� �� ���� �)� �� ����� ��� ���)��( ��� �� �����3� '������3� �����.������ "�� ���� ����( *�������3� /����� /������� ���"��� ��� ����� ���� ���� � ���� ����� ;�� �� ��� ������ �� ���� ������ ��� ����)�� �� �� ������ �� ������ /���� @������ 3� gg���������(�����33 ����� )�� "���� � �� ��� ������( � �� ��.���� ��� '������3�� ��� ����� ���� ���� � ���� ������ ���� ��������� � ���� gg���������(�����33 )�� �������"��� 5������ ��� ������ �(����� + )��� "���� �� �� �� �� ����������� ������ )�� ��.����� �� ���� �� ������ �������� ����� l�)����� ��� ����� ����� ���� ������ ����������� ������� ��� �� ������ )��� ��� ��;������( "���� � �� �� ����� ���� ���� ��� )���� ��� "� � ������ -���� ��� 9���� ��������� ���� ��� ����� ��� ���� ��� ��������( ����(� ������ "� ��� ����(�� �� �� �� ����������� ��� (������ +� ������� ������������ ���� + ��� ��� ������ ��� ������ ��� �������� ����(�� -�� ���� �������� �� ������ ��� � ���� �������� � ������� �� ���� � ��� �� ������� �� ��� �)� ������ ����� �� � ��� gg��������33 ������ ��� ��� gg���������(�����33 ������ *����)� � ������ -�� !���� 9������ 9���� � * ����3 �������� )�� ����� ��� �� ��� ����� *��������( ��� ������� � ����������� ������ ��� 9���� ������ gg-�� (������ 4������ ���� �� ���� � 4������ ������� � ����������� "�(��� �� ��� )��� � ���� ���)� �� ��� ������ �� ���) ���� ��� )�� �������33� !����( ��� ������ ������ "��)��� ��� �� �������� ���� �� ����� *����� ������� ��� ��� ;�� �� �������� �"��� *�������� ��� ����� ����� ���� ��� �������"������ � ��� ����������� )�� � �������� "��� �� � �� � ����� ��� ���� ��� ����� �����3� ���������( ����� ������ )�� �� �� ��� 4�����( �� ����� �������"�� ������ ��� 9���� �������� ��� ����� �����3� �����(� �� ��� gg����������33 ������ ��� �������� ��� ���� �� � ��) ������ -�� /� ���� 9���� (������ ����������� ��� �� � ���� ������ ��������� ����� ���� ��� ����� ����� ��� ��������� #&# %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� ����� ����� ��� ������� � ����������� ������ ��� ���� � ������ ��� ���� # ����� ��� ��� ���� � ��� ���������� �� )���� �� ;�� ���� "���� � �� �� ����(�� )���( �� ����������� -�� �� ��� � ���� �����( �� ����� ��������� ��� "��� �"���� �� ��� � ����� ������ �� ���� �� �� ������ "����� �� ���� ��� ���� ��� � ���������� �� ��� ���� � )���( �� ������������ *� ���������� "����� ��� "��� ����� �� ���������� ���������� ������� � ��� �� ����� � ���� �� ����� �� ������,� ����� �c ����� �� �������� ��� �� � ���������� -���� ��� �� ���� ��������� � �������� ��� �� ��� �� ���� )��� �� ��� "� ����� ������� -�� ����� ��� ������ ��� ��(��;���� �� ;������� ������������ ��� ������������ "�� ��� � ��� �� ;���� �� ���������� �� ������� �� ���������� �������� �������� � �� ���������� �� �������� ����(� �� ����� ;������� ���"����� �� ��� ��� �� ���� ����� �� �� ������� ���� � ����� ������ ��� gg������33 ���� ��� � ���� ��� ��� ������ �� .����� �� j�� *�������� 0�,��� ��� '����� ������ �������� ��"�������� �c ����� �� �������� �� ��� ���� ���� � ��� ���������( ��� ��� ��� �� 4�������� ������ ������ "� ��������"�� �� ��� ������� ��� �������� 5�� �c ������� ��� ������ ����� �������� ���� ������� ��� ������ ��������"�� �� �������� ��� ���� )��� ������ �� ��� ����������� �� ��� � � �������� ���������� ���� ����������� �� ������ �� � ������ ������� 9�������� ��������� ������ ���� ��(��;���� ������( �� ������ f���� ������� � ��������� �� ������� ����?"���� �������� � ��������� ��� ���� �������� -�� ��) ��������� ����(��,�� �� ����������� "��)��� � ���� ��� � ������� !� ����� ��� ��� ���� ���� � �����3� �������� �� � ��(���� ����(��,�� �� ���� �������� ���� ��� "� ���������� @���� � � ����� �(����� ��� ���������� �� ������ ���� ���"��� ��� ��� � )���� ���� ���� ����� �� � �� ������ "�� ��� ��) ���� ��� �������� �� ��� ���"����� �� ��� ����� �� ������� �����3� �������� ��������� -�� ��) (�������� ����(��,�� �������� ������� �� �� ����������( �� �� �������( ������ )���� "����� ��� ��������� ����� �� � ��(��(���� ����� 8 ���������� 8/�����: � -���� k22#:� -���� � � ������ ������ � �� ������ )�� ��(��(��� �� ����)��( �� �� ����� ������ �� ��� ������ 3� �������� �� ��������� �� ���� � )�" ���� )�� ��(��(��� �� �����( �������� �� ��������� ���� ��(��(���� �� �� ���(�� ��� ����� � ��� ������ 3� ������� f"�������� ��������( ������ "� ���� �� ������ �� ��������� ��� ��"��� � �������� ��� �% ��� �� )���� ���� �.����� �"����� ���� �� �� ��� �.����"�� �� ����) ����������� �� "� �������,�� "� ���� � ��� �� )������ ����� ���� �� .�����;�"�� �� gg������33 ��� �c �������� � ���� ��� � ��� �� ������� ����(�� ������� -�� ������� ��((��� �� ������ ���� )���� ��� ���� ����(��,� ��� ���������� �������3 ������� .�����;�"�� �� ���� ��� ������ ��� �������� �� ����� ������� "�� ���� )���� ���� ���� ��� ������"�� � ��� � �������( �� ��������� �� ��������( ������� �������� �� �������� ��� ���( ��(�������� ���� �� ���� ������ ���"����� �� ��� ����� �� ������ ���� (������� ��� ����� ���� �� ��� �"��� �� ����? �"��� ������� ��.������ ��������� �� ���� ������� �������� �� ��������� -�� ��(�������� ������ ������ ���� ���"����� ������ "� �� ���� )������ � ������������� � ����% ��� ���� ��.�������� )���� "� ���� ��� ������ � ��� ������ �� �������� "� ������;��� -���� ������ ���� "� ������� ����(�� �)������ �� �� �)��� � .�����;�"�� �������� ������ ��� ������ )���� ������� ������ �������� ������ �� 6#�7$$� )�������� �� (������� �������� ������ )���� ������� %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� #&& ������ ���?� ? ����� ����� ��� �c ������ ��� )���� � ���;����� ������� ����� ����� ��� ���������3 ��� �� ��� ������ �� �������� ������� �� ��� ��� �� �� )��� �� ��� ����� ����� ��� ���������3 ��� �� ������� ����� ��� ��(��������� -���� ����������� �� �������� ���� ��� )��� �������� ���� ������ ��)��� ��� �������� � ������� �������� �� �������� )���� "� �c ���� �� ������� ����(�� �� )���� @������� �������� �� �������� )���� "� ��;��� ��� ���� �� �� ��� �������� -�� � ��� *���� ���� '��������� *��� -�� ��(�������� ������ "� �� ��� ��������< 4������ ������ -���� �� ������� ��������� �� ���� �� � � ���"����� �� �������� ��) ���"����� �� �� ���� �������� �� ���������( ��� ������(� ����� � "��� ��� ��.��� �� )��� �� �������� �� ����� �� �� gg��������� �� �33% �� ����� ������ ���"����� )������ ���� �� �� ���� �� ������� � �� ������ �������� /��������� ����� ��� ���� ��������� )���� �������� �� ��� ��c���� ����(��� ���� �� ��� �"����� � ������� ���������"�� ����� �� ��� �� ���� � gg �� ���33 ��������m f ������ ���% ����� ��� ���� ��������� )���� ��� ���� �� ����� ���(��� ��� ����� ��� ��� gg������33 � ��� ������ �������� )���� ��"�"�� ��� "� ���������"��� +���� ��� gg������33 � ��� ������ �� �������� �� � (����������� �(����� ����� ��� "� �������������� �� �������� ������% ��)����� ��� (���������� �������� ��(�� ����) ���� ������ �� ������ �� 9����� � 9����� ����"������ �� ���� ������� /���� ���� ����� � �������� ��� � )���� � ��� �� ���� "��� "���� � �� �� �������� gg�������33 �� ���? (����������� ��������� �� )���� ���� ���� ����� �������� ��(�� "� �������� -�� ��(�������� )���� ���� ������ �����( ��������� �� ��� �������� �� �������� ������� �� �� ���"� �� ���� ������ )���� ����"���� � )�" ���� ������( ����� ��� ������ �������� ���"��� � ��������( ��� �c ����� ���� � ���� )���� ����"����� 5���� ��� ������ �� �����( �(������ ��(�� "� ���� ���� ���������� �"��� )���� ���� ���� ���<�� �������� �������� �� ��������� '�����3� � ;��� ��� ����� "��������� ��(�� "� ���� ���� ���� �� �"��� )��� �� �������� ���� (����� ���<�� ������ ��� )�� )���� ���� ������ �� ���� �� �������� ��� �� )��� �� ����� *�� � ��� �� �������� ���3� ���.������ ��������� �� ����� ��� ������� )��� ������� ��������( �� ��� �������� ���� �� ��� *� ����� ����� )��� "� ���� ��������� �� ��� ���������� �������� ���������� *����)� �� -����e���� 9�� �� > 4� /� �#� �$7b 89�'� 9��� � �aa1:� � 3� �� ��� ���3� �� ���� ##7 4�&� �$b& 8a�� 9��� #$$$:� ����� (����� 7&# e�/� a$# 8#$$�:� 9��� ����� @���� 9���� 7#1� �� ��.� 9���(���� �� 8#$$�:� @�� *������ � 0��""��( -��� � ���� �� �' /���� ��������� 9���� ������� 4�"� #&� #$$� �� �$*� 4��� 9����� 5�����( *��� �7 e�/�9� k�b1�� 4����� �� 8#$$$:� @�� ���� ��������� -�������� �� ��� 9��� ����� ��(�������� /������ 9���� f� -���� �������� / ����� f�����(�� l�����( �� �������� -�� �� ���� ���� ����+ , � � d�.� ��� 8������"�� �� ��� <<)))� ��������(������(<��<'�h?����� �����:� 0�,��� �� -����� ����� 9����� 9�� �� >27 /��#� ��#& 84��� * � �aaa:� �������� -��� e��� @�"��� �� �������� �� ����"���� 9����� *�������� � ������ ����� ��� $���� �� '�* 0�� ��+ ���� #b� #$$$� � @ @�"���������� �������� -�� � ��� *���� ���� '��������� *��� �aa1� @�"� ��) �$7?&�1� ��# /���� &$$>� �1 e�/�9�k�$#1 8�aa1:� ����� 9��� k#1?7�?�$�� #&2 %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� http://www.privacyrights.org/ar/dmv-idtheft.htm ����� � +� 8#$$$:� @�� ���� ��������� @��������( @������ ��� @��������( ������ � ��� /����� /������� !��"��� �������� � �+0+ �" ; �� �� �.� $�: ��+ �� $ ��� $� ����� �.� � ��� <��� ��� ,���� � ��+� �$b�� 9��(����� ���(� -� 8�aaa:� /����� ����������% ������� ���� *������� ��������( ��� �������� -�� �� 5.� 6������ '�*���� ����� &� �aaa� ������ �� / ��(��� ����� �a /�+�&� 1a$ 8-�c� * �� #$$$:� f��� ��� 9��� k#a�&�2a� @������ �� e���� /���� 5���� b1� /��#� �&b2 8*��� �aab:� @��,�� �� +� ����� b1# !�j�/�#� �a2 8�aa1:� ���������� 8/�����: � -���� k22#� *��?*5*� /������� �� 8#$$�:� ����) � /������� #$$$ 9��� ����� ��(�������� /� �����( h������ �� ��������( 4��� ��� 9���� � ��� �� �������� *(�� , >� ��� '�* 0����*+��0���(� /����� � ��)� e��������� � ��� @���;�� +������ #$$�� -����?��?������( *��� �7 e�/�9� k�b2&� +� ���� �� *����)�� sss e�/� sss� #$$�� e�/� ��c�� �$&$b 8#$$�:� e�/� 4������ -���� 9��������� 8#$$�:� 6������� 5.� � � ������� ����+ ������� ��� 5����� � 6������� 5.� �� a ���:�� 1777 � %��� �##1+ e�/� 0������ *��������( f ;�� 8�aa1:� 6������� �����+ 6� ����� � � 9������� � � �� ��� 6������� 6��� � �� '������� %+&+ ,���-� �� � )+&+ '�.�� / ����� ��� $���� �� 0����* 1# 2�##13 ��14�� #&7 identity theft: no help for consumers introduction the magnitude of identity theft criminalization of identity theft state actions and legislation the federal identity theft and assumption deterrence act, 1998 (18 u.s.c.§1028, pub. law 105-318, 112 stat. 3007) the inadequacy of criminalization mattly v. spiegel inc. (2000) (19 s.w.3d 890 (tex. app.)) guzman v. toyota motor credit corp. (1999) (745 so.2d 1123 (fla. app.)) polzer v. trw inc. (1998) (682 n.y.s.2d 194) patrick v. union state bank (1996) (681 so.2d 1364 (ala.)) andrews v. transunion corp. (2001)/trw. (7 f.supp.2d 1056 (c.d.calif. 1998), aff'd in part, rev'd in part, 225 f.3d 1063 (9th cir. 2000), cert. grtd., 532 u.s. 902) conclusion references financial services review, 33(1) 28 retirement expectations vs. reality: if covid-19 did not impact retirement expectations significantly, what did? zhikun liu,1 david blanchett,2 qi sun,3 and naomi fink4 abstract using two data sets (a prudential financial wellness survey and the health and retirement study), this study demonstrates that although there is generally a natural upward trend for older (age 50+) americans to progressively delay their expected retirement age, this trend has no statistically significant relationship with the covid-19 pandemic. the distribution of older americans’ expected retirement ages is bimodal, often centered around two social security benefit claiming ages – the early retirement age and full retirement age. however, actual retirement ages are more likely to follow a left-skewed distribution, whereby people appear to retire earlier than expected. the most significant factors that influence participants’ retirement decisions relative to expectations are health (+), wealth (-), age (+), change of marital status (+), mortality expectations (+), education levels (+), disability (-), and major illness diagnosis (-). focusing on these factors can help the retirement benefits community explore strategies to mitigate the negative consequences of gaps between retirement expectations and reality. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation liu, z., blanchett, d., sun, q., & fink, n. (2025). retirement expectations vs. reality: if covid19 did not impact retirement expectations significantly, what did? financial services review, 33(1), 28-49. introduction and literature review for most households, planning for retirement involves some degree of uncertainty, whether about post-retirement spending, longevity, health, or other factors. american workers’ retirement timing expectations play important roles in predicting their retirement decisions (haider and stephens 2007), and thus also play important roles in policy proposals (vanderhei 2022) as well as the design of retirement products 1 corresponding author (zliu@missionsq.org). missionsquare retirement, washington, dc, usa 2 pgim dc solutions, newark, nj, usa 3 pacific life, newport beach, ca, usa 4 nikko asset management co, ltd, tokyo, japan and services (blanchett 2015) to help retirees better plan for their future retirement. research interests in people’s retirement expectations have increased significantly in both academia and industry (beehr 2014; haider and stephens 2007; hanspal et al. 2020, 2021). arguably, the most important retirement decision is when to retire, a factor over which workers have varying degrees of control. https://creativecommons.org/licenses/by-nc/4.0/ mailto:zliu@missionsq.org https://creativecommons.org/licenses/by-nc/4.0/ financial services review, 33(1) 29 from stochastic retirement readiness model design to social security benefits projection estimations and from defined contribution (dc) plan savings strategies to monte carlo retirement-income-replacement goals success rate calculations, people’s expected retirement age is a crucial factor for industry advisors and financial planners to design retirement services, projection models, and to offer sound advice to their clients. retirement expectations determine retirement intentions, and retirement intentions are predictors of retirement behavior (blanchett 2018). individuals who want to maintain their desired standard of living when retired will need to accumulate sufficient financial assets with careful preparation and have a more realistic retirement age projection before the actual retirement. conversely, if an individual does not demonstrate sufficient preparedness, her plans to retire at a given age may not prove achievable. a deeper understanding of the significance of retirement age expectations is needed for such preparations. dubina et al. (2020) [the u.s. bureau of labor statistics] predicted that over 1 out of 4 workers in the u.s. labor force will be 65 or older by 2030, contributing to the sustained decline of the labor force participation rate. many workers and retirees concur that saving for retirement is more important than other household demands (ebri retirement confidence survey 2021). knowing that retirement plans are crucial to retirement outcomes, it is important to study american workers’, especially elderly american adults’ retirement expectations. previous studies investigated several socioeconomic and demographic factors that might potentially impact expected retirement ages. for instance, using the 2006 and 2008 waves of the health and retirement study (hrs), szinovacz et al. (2014) found that debt, assets, education, race, gender, marital status, and income are all factors that are associated with participants’ retirement decisions. on the other side of the coin, past studies have found that the age of eligibility for retirement benefits has a significant impact on actual retirement date (coile & gruber, 2007). research has also found evidence that labor market downturns, such as the global financial crisis of 2009-2009 and the covid-19 pandemic, can impact retirement decision-making. for example, coile and zhang (2022) found that there was an increasing trend of earlier retirement during the covid-19 pandemic. similarly, davis (2021) found that an increasing portion of part-time older workers and workers in high-contact occupations retired earlier during the pandemic. however, few studies have been published to examine the impact of the covid-19 pandemic on older american adults’ retirement expectations. more importantly, as cregan et al. (2023) pointed out in their study, retirement is often a process instead of a one-time decision, and people’s retirement intentions change over time. to fill the gaps in the existing literature, we explore the potential influence of the covid-19 pandemic on employees’ expected retirement age. given the richness of data at our disposal, we also examine different socioeconomic and demographic factors (other than covid-19) that could potentially impact expected and actual retirement ages. more generally, this study seeks to answer several important questions regarding the accuracy of retirement expectations among older americans. do retirees’ expectations match the reality of when they eventually retire? how significant are the gaps between the participants’ actual and expected retirement ages? what factors affect people’s retirement expectations in their 50s and impact their actual retirement decisions a decade or two later? using two different data sets, this study first examines whether the covid-19 pandemic significantly impacted american workers’ retirement expectations. then, it explores the gaps between older americans’ expected and actual retirement ages, further examining which factors are related to worker expectations and actual retirement decisions. the first data set used in this study is based on responses to an online financial wellness assessment offered by prudential financial. this dataset is used to better understand the variability of expected retirement age during covid-19. two separate questionnaires are reviewed. the first was offered from april 20, 2017, to june 27, 2020, and the second from june 28, 2020, to december 1, 2021. there are 154,403 responses liu et al. 30 available for the first dataset and 87,571 for the second that meet the required filters (out of a total of 241,974 responses). if an individual took the questionnaires multiple times, only the last response is included. key variables in this data set include age, gender, marital status, household income, and expected retirement age, and date of survey completion. the second data set employed in this study is the health and retirement study (hrs) data, a nationally representative longitudinal survey of more than 37,000 individuals over the age of 50 that is widely used in the retirement literature. the core hrs survey has been conducted biennially since 1992, and the 2020 wave contains a covid-19 section that recorded the participants’ responses to various pandemicrelated questions. analyses from both data sets indicate that older americans’ retirement expectations (including planned retirement age and social security benefit claiming age) remain largely uninterrupted despite enduring the impact of the covid-19 pandemic on their work and financial situations in 2020. among the factors that influence participants’ retirement decisions, health (+)5, wealth (-), age (+), change of marital status (+), mortality expectations (+), education levels (+), and major illness diagnosis (-) are the most significant factors that impact participants’ expected retirement timing. methodology and results correctly anticipating retirement timing is not only crucial for an individual’s own retirement success but also important for financial planners to provide appropriate advice and services to their clients. a growing body of research, though, indicates that individuals’ retirement age projections are inconsistent with their actual retirement ages. meanwhile, incorporating retirement age uncertainty into a financial plan can significantly impact required retirement savings levels. the more accurately participantfacing calculators and financial plans represent 5 in this text, “(-)” means the impact is negative (i.e., retire earlier than expected and “(+)” means retire later than expected). the range of outcomes with associated probabilities, the more efficiently they can help workers achieve a successful retirement. examining the prudential and hrs survey data sets, this study finds that the covid-19 pandemic did not impact older americans’ retirement expectations significantly. however, acknowledging that other factors than covid-19 (or other significant events) do meaningfully influence both expectations and outcomes, the study investigates different socioeconomic and demographic factors that impact older americans’ actual and expected retirement ages, as well as the gaps between the two. both sets of results indicate that more than half of the participants retire earlier than they expected. the most significant factors that influence participants’ retirement decisions relative to expectations are health, wealth, age, change of marital status, mortality expectations, education levels, disability, and major illness diagnosis. understanding and better capturing the influence of these factors can help the retirement benefits community explore strategies to mitigate the negative consequences of individuals’ inaccurate retirement expectations upon their retirement outcomes. covid-19 was a nonevent for retirement age expectations based on the survey data from prudential financial, retirement age expectations are relatively sticky, especially among older households. figure 1 shows the average expected retirement age by month and age group. note that these are independent observations and are highly unlikely to be the same individual over time (although theoretically possible). the fluctuations during and shortly after the covid-19 period are not significantly larger than those before the pandemic period. meanwhile, although different ages responded differently, their responses within age groups did not differ meaningfully during vs. shortly before and after the covid-19 period. financial services review, 33(1) 31 figure 1. average expected retirement age by month and respondent age group (prudential data) prevalent change in expected retirement age during covid-19 is not observed from figure 1 above, especially among the older age cohorts. figure 2 further supports this observation by showing the month coefficient of an ols regression6 where the dependent variable is retirement age expectation. the independent variables include age, gender, marital status, and income, etc. 6 see appendix table 3 for the detailed ols regression results. the average expected retirement age change is captured by the difference between the monthly variable coefficients minus the average value of these months. 62 63 64 65 66 67 68 69 70 jul-18 jan-19 aug-19 feb-20 sep-20 mar-21 oct-21 a ve ra ge r et ir em en t a ge survey month 45-49 50-54 55-59 60-64 65-70 liu et al. 32 figure 2. ols month coefficient, by month and respondent age group (prudential data) based on figure 2, the average expected retirement age increased briefly during the beginning of the covid-19 pandemic. however, for older cohorts, the change is economically insignificant and (observing the timing of the increase followed by its rapid retreat) may have been caused by stock market volatility rather than by covid-19. the next two exhibits show that in general, there is a natural upward trend for participants to expect a later and later retirement age when they are older. however, this natural trend of delaying retirement has no statistically significant relationship7 with the covid-19 pandemic. 7 tested with both ols and did (difference in difference) regressions in the appendix. see appendix table 1 and table 2 for more details. -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 jan-19 apr-19 jul-19 oct-19 jan-20 apr-20 jul-20 oct-20 jan-21 apr-21 jul-21 o ls c o ef fi ci en t: a ve ra ge e xp ec te d r e ti re m n t a ge c h an ge in y e ar s survey month 40-49 50-59 60-70 financial services review, 33(1) 33 figure 3. average expected retirement age by respondent age (prudential data) figure 4: average expected retirement age by respondent age (hrs data) retirement expectations vs. realities: factors that impact retirement decisions if covid-19 did not alter participants’ retirement expectations significantly, what does? also, are there significant gaps between expected and actual retirement ages? next, this study focuses on investigating the different socioeconomic and demographic factors that could potentially impact older american adults’ expected and actual retirement ages, using the longitudinal hrs data. it also examines how the change of these factors shifts the gaps between the participants’ retirement expectations and reality. in the hrs data, participants’ retirement records are tracked from 1992 to 2018, during which period some participants retired and then went back to work one or more times. therefore, we created four possible definitions for “actual” retirement ages from the hrs: 60 61 62 63 64 65 66 67 68 69 70 35 40 45 50 55 60 65 70 a ve ra ge e xp ec te d r et ir em en t a ge respondent age 61.7 62.0 62.9 63.0 63.7 64.3 64.1 65.3 65.8 65.0 65.3 65.9 65.6 66.0 66.463.2 64.0 64.7 65.4 66.4 66.0 67.1 67.6 67.2 67.7 68.2 67.4 67.9 69.2 56 58 60 62 64 66 68 70 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 r e ti re m en t a ge s hrs survey years plan to stop working think will stop working linear (plan to stop working) linear (think will stop working) liu et al. 34 1) first recorded retirement age – the first recorded year in which the participant retired. 2) last recorded retirement age – the last (up until 2018) recorded year in which the participant retired. 3) mode retirement age – the recorded year in which the participant retired for the longest period from 1992 to 2018. 4) first and mode retirement age – the first and only recorded year in which the participant retired and stayed retired (never went back to work). figure 5 shows the different percentages of participants who retire earlier, later, and in the same year as they expected versus each of these definitions when they were in their 50s. while 53.2% of the hrs participants retired earlier than they expected when they originally answered the survey in 1992 using the first recorded retirement age as the definition of the “actual” retirement age, only 42.4% using he last recorded retirement age. regardless of definition there is a clear trend where people retire earlier than expected, consistent with the research of ebri (2017) and the gallup (2018), among others. figure 5. actual vs. expected retirement age (hrs data) note: the analysis sample used above is 3,441 hrs primary respondents who retired between 1992 and 2018. the expected retirement age used is calculated using the respondents’ expected retirement year recorded in the 1992 wave. across the twelve hrs survey waves, there are 17,170 participants who retired between 1992 and 2018. among them, 7,773 answered the question “when do you think you will stop working?” during the 1992 survey wave, and 4,704 of them are primary respondents8after further restricting the age of these primary 8 in the hrs, the primary response is the household member who answered the survey primarily, not including their spouses or children in the household. respondents to be between 50 and 59 (i.e., filtering out those who are not in their 50s) (see reasons in figure 4 in the appendix), the main analysis sample of this study consists of 3,441 respondents, representing 8,364,876 u.s. 53.2% 15.6% 31.2% 48.2% 17.2% 34.6% 44.6% 15.4% 40.0% 42.2% 13.2% 44.6% 0% 10% 20% 30% 40% 50% 60% earlier than expected same year as expected later than expected first recorded retirement age first=mode retirement age mode recorded retirement age last recorded retirement age financial services review, 33(1) 35 national population after applying the 1992 personal-level analysis weight. after using the first recorded year in which the participant retired as the definition of the actual retirement, we can plot the hrs participants’ average expected, actual retirement ages, and the gaps between the two. figure 6 below shows that older americans’ expected retirement ages are bimodal, often centered around two social security (ss) retirement benefit claiming ages – 62 and 65 years old, which are the initial eligible claiming age and full ss benefit claiming age. note that the full retirement age (fra) in social security for those 50-year-olds in 1992 was around 659 instead of 67. in addition, 65 is also the age for medicare eligibility. rutledge et al. (2015) point out that social security fra is strongly correlated with retirement age. they state that a one-year increase in social security fra is associated with a 0.3-year increase in the retirement age, all else equal. figure 7 indicates that although participants plan to retire around those two ages, their actual retirement ages are more spread out, and they are more likely to retire at age 62 or before, with a clear mode of 62. according to figure 6, more than 34% and 31% of older americans plan to retire around age 62 or 65, respectively, consisting of the majority of the population. however, it is clear that actual retirement ages differ from expectations. figure 7 shows that not only do people actually retire well before the medicare eligibility age of 65 but also that the actual retirement age numbers are more spread out, with a peak at age 62 (15%). more than 44% of the respondents retire before age 62, which is concentrated on the left side of the peak. in contrast, only 24% of participants expected to retire before age 62 when they were in their 50s. 9 some of the cohort (who is in their 50s in 1992) have fra slightly above 65. please see the details of social security fra base on birth years from this link: https://www.ssa.gov/pressoffice/incretage.html. liu et al. 36 figure 6. distribution of expected retirement age for all retired participants between 50 and 60 years old in 1992 (hrs data) note: the number of primary respondents in this chart is 3,289 before weights are applied. after applying the 1992 respondent level analysis weight, it represents 8,361,046 older americans in the united states. 7.1% 4.5% 22.3% 12.0% 1.7% 21.0% 10.8% 2.8% 1.3% 0% 5% 10% 15% 20% 25% 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 75 76 77 78 80 81 85 d en si ty ( % ) first recorded retirement age density (%) 1992 respondent level analyses weight applied financial services review, 33(1) 37 figure 7. distribution of actual retirement age for all retired participants between 50 and 60 years old in 1992 (hrs data) note: the number of primary respondents in this chart is 3,289 before weights are applied. after applying the 1992 respondent level analysis weight, it represents 8,361,046 older americans in the united states. 5.0% 6.2% 7.8% 7.4% 15.2% 9.0% 5.6% 7.7% 4.8% 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0% 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 84 85 d en si ty ( % ) first recorded retirement age density (%) 1992 respondent level analyses weight applied liu et al. 38 figure 8. first reported retirement age minus 1992 expected retirement age (hrs data) note: the number of primary respondents in this chart is 3,283 before weights are applied. after applying the 1992 respondent level analysis weight, it represents 8,343,446 older americans in the united states. using the actual retirement age minus the expected retirement age, which is the error in expectations, the distribution of these differences also implies a higher likelihood of retiring before expectations. as figure 8 indicates, only one in six (16%) retired at the age they expected. including those that accurately forecast their own retirement age, 36 percent of the sample retired within (plus or minus) one year of their expected first retirement age, 49 percent retired within two years, and 64 percent retired within three years of their expected first retirement age. just under 80 percent retired within five years of their expected retirement ages, and almost 5 percent retired more than ten years away from the retirement age they had forecasted when they were in their 50s. 10 the skewness of the distribution is 1.73 and the kurtosis is 3.54, which indicate this distribution is a based on exhibits 6-8, while most older americans’ expected retirement ages are centered around two social security benefit claiming ages – 62 and 65 years old, the distribution of their actual retirement ages follows a left-skewed kernel distribution10. the notable gap (see appendix figure 4) between older americans’ actual and expected retirement ages is the motivation to explore the factors that could potentially influence retirement decisions. using the longitudinal hrs data, this study examines the different socioeconomic and demographic factors that could possibly affect older americans’ retirement expectations and the gaps between their actual and expected retirement ages. the factors tested in this study include wealth, health, gender, education level, selfpositive-skewed (left-skewed) leptokurtic distribution. 0.1%0.1%0.1%0.1%0.2%0.2%0.3%0.4% 0.7%0.7% 1.2% 1.9%2.1% 2.9% 3.2% 4.4% 5.9% 9.7% 7.7% 11.0% 16.1% 8.5% 5.4%5.3% 3.0% 2.2% 1.6% 1.0%1.0% 0.5%0.6%0.6%0.4%0.2%0.1%0.2%0.1%0.0%0.1%0.0% 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 fr eq u en cy d en si ty ( % ) actual minus expected age differences frequency density {years earlier than expected} {years later than expected} 1992 respondent level analyses weight applied financial services review, 33(1) 39 perceived life expectancy, age, financial planning horizon, major illness diagnosis, race, number of children, and marital status. table 1 summarizes some of these socioeconomic and demographic variables with significant effects on older americans’ retirement age decisions based on the results of the regression studies that follow. table 1. association or impact of different factors on retirement age decisions (summary) association with expected retirement age association with actual retirement age longitudinal impact on retirement expectations (1992 2018) longitudinal impact on the gaps (expectation accuracy) longitudinal impact on likelihood of retire earlier or later factors earlier (-) later (+) earlier (-) later (+) hasten (-) delay (+) bigger (+) (less accurate) smaller (-) (more accurate) earlier (-) later (+) health improve expect to retire later** retire later** hasten retirement expectations*** increase gaps** likely to retire later*** wealth increase expect to retire earlier*** retire earlier*** delay retirement expectations*** decrease gaps*** likely to retire earlier** self-perceived life expectancy increase expect to retire later*** (not significant) hasten retirement expectations*** increase gaps*** likely to retire later*** getting married (not statistically significant) hasten retirement expectations** increase gaps** likely to retire later* divorce/widowed expect to retire later** (not significant) delay retirement expectations*** decrease gaps*** likely to retire earlier* major illness diagnosis expect to retire earlier** retire earlier*** delay retirement expectations*** decrease gaps*** likely to retire earlier*** note: * p < 0.05, ** p < 0.01, *** p < 0.001. the first two columns select results from cross-sectional ols regressions, and the last three columns select results from longitudinal fixed-effects and ordered probit regressions. the regression sample is based on the 5,912 participants who were in their 50s during the 1992 hrs survey wave and retired after the year 1992. using the cross-sectional ols regression (results in table 2), we first examine the factors that are associated with older americans’ expected retirement ages and their actual retirement ages. the regression results indicate that keeping everything else equal (ceteris paribus), participants with more education, better health, longer self-perceived life expectancy, a larger number of living children, and older age are associated with both later retirement expectations and actual retirement ages in their 50s. on the contrary, having more wealth and major illness diagnoses are factors that are associated with both earlier retirement expectations and actual retirement ages. liu et al. 40 table 2: ols regression on factors associated with average retirement expectations and actual retirement age decision independent variables dependent variable: actual retirement age dependent variable: average expected retirement age age_1992 0.351*** 0.259*** (0.029) (0.034) female 0.245 -0.306 (0.148) (0.173) married1992 0.0470 0.308 (0.193) (0.227) education years 0.0804** 0.221*** (0.029) (0.034) ln(wealth1992) -0.231*** -0.321*** (0.055) (0.067) race black (race white omitted) 0.0477 -0.482 (0.273) (0.316) race other 0.808* 0.0545 (0.408) (0.500) finplanhorizon1992 0.0795 -0.0974 (0.065) (0.078) everhadcancerorheartprob1992 -0.629** -1.041*** (0.239) (0.287) health1992 0.681*** 0.549*** (0.077) (0.093) numberofdivorces1992 0.0667 0.443** (0.115) (0.135) numberoflivingchildren1992 0.0898* 0.0879 (0.040) (0.048) yearstolive1992 0.00111 0.0410*** (0.008) (0.010) n 4,310 3,664 next, we utilize the longitudinal data to investigate how the participants’ actual and expected retirement ages shift when these socioeconomic and demographic factors change throughout the years (from 1992 to 2018). the results of the longitudinal (time) fixed effects ols regression in table 3 tell us that the increase in wealth, widowhood, and major illness diagnosis will positively (retire later) impact participants’ expected retirement age. these life changes predict an individual will think they will retire later. getting married, improvement in selfreported health, and self-perceived life expectancy increase will cause participants to expect earlier retirement ages. these life changes predict an individual will think they will retire earlier. financial services review, 33(1) 41 table 3. longitudinal fixed effects ols regression on retirement expectations independent variables dependent variable: expected retirement age (1992-2018) health -0.555** (0.170) ln(wealth) 1.037*** (0.140) yearstolive -0.216*** (0.019) married -2.051** (0.671) widowed 3.467*** (0.750) numberoflivingchildren 0.203 (0.181) finplanhorizon 0.0602 (0.106) everhadcancerorheartprob 5.244*** (0.422) n 5,726 when it comes to the difference (gaps) between the participant’s actual retirement age and their expected retirement age, the longitudinal (time) fixed effect ols regression in table 4 indicated the following conclusion: increases in participants’ health, self-perceived life expectancy, and being married will increase the gaps between participants’ first recorded retirement age and their expected retirement age. so these types of life changes predict that an individual will be less accurate in retirement age forecasts. meanwhile, the increase in wealth, being widowed, and major illness diagnoses will decrease the gaps between participants’ actual vs. expected retirement ages. after investigating the wealth factor, we find out an interesting phenomenon: although older americans expect to retire later when they experience a wealth increase prior to their retirement, their actual retirement age is actually getting younger and closer to their original expected retirement age (gaps are smaller). on the other hand, an increased life expectancy pushed back the actual retirement age and led to a growing retirement age gap. for example, keeping everything else equal, a one-year increase in life expectancy is on average associated with 0.22 years increase in the gap between actual and expected retirement. liu et al. 42 table 4. longitudinal fixed effects ols regression on actual minus expectations independent variables dependent variable: gaps between actual and expected retirement ages (1992-2018) health 0.563** (0.171) ln(wealth) -1.069*** (0.142) yearstolive 0.218*** (0.019) married 1.933** (0.686) widowed -3.580*** (0.759) numberoflivingchildren -0.149 (0.182) finplanhorizon -0.0461 (0.107) everhadcancerorheartprob -5.240*** (0.426) n 5,668 last but not least, we want to know the magnitude of the likelihood for participants to retire earlier or later (timing flags) caused by each of these factors. table 5 shows the results of the average marginal effects of the longitudinal ordered probit regression on the timing flags regarding participants’ actual retirement age minus their expected retirement age (the gaps). the regression results imply that the positive improvements in participants’ selfreported health, self-perceived life expectancy, income level, and getting married will likely cause participants to retire later than expected. positive changes in participants’ wealth levels, financial planning horizon increases, major illness diagnoses, as well as widowhood are likely to cause participants to retire earlier than expected, keeping everything else equal. conclusion and implications understanding retirement age expectations compared to actual retirement ages can help policymakers, employers, and industry providers improve retirement benefit design across a range of structures, products, and services, such as defaults and catch-up provisions, investment modeling, and participant advice and education. using the prudential financial survey data and the longitudinal health and retirement study (hrs) data, this study finds that although there is generally a natural upward trend for older american adults to progressively delay their expected retirement, this trend has no statistically significant relationship with the covid-19 pandemic. the study then examines socioeconomic and demographic factors that are thought to impact older american adults’ expected and actual retirement ages. in addition, this study also investigated how the changes in these factors shift the gaps between participants’ expected and actual retirement ages. understanding the relationship between these factors, workers’ expected retirement timing and their ultimate retirement choices, employers and retirement industry providers can help employees better prepare for the (often negative) financial situations that arise when retirement expectations and reality do not align. financial services review, 33(1) 43 table 5: longitudinal ordered probit average marginal effects on reality minus expectations retirement timing flags independent variable: flags on actual minus expected retirement ages independent variables retire earlier than expected (-1) same year as expected (0) retire later than expected (+1) health -0.0283*** 0.00434*** 0.0239*** (0.0066) (0.0010) (0.0056) ln(wealth) 0.0146** -0.00224** -0.0123** (0.0045) (0.0007) (0.0038) years to live -0.00344*** 0.000529*** 0.00291*** (0.0007) (0.0001) (0.0006) married -0.0414* 0.00637* 0.0351* (0.0181) (0.0028) (0.0154) widowed 0.0892** -0.0137** -0.0755** (0.0280) (0.0043) (0.0237) number of living children 0.000759 -0.000117 -0.000642 (0.0036) (0.0006) (0.0030) financial plan horizon 0.0140** -0.00215** -0.0118** (0.0051) (0.0008) (0.0043) ever had cancer or heart prob 0.0966*** -0.0148*** -0.0818*** (0.0178) (0.0028) (0.0151) n 5,668 5,668 5,668 retirement age decisions not only affect the economic well-being of individuals and households in the united states., but also impact the financial solvency of the social security system (montalto, yuh, and hannah, 2000). a growing body of research indicates that retirement age projections are inconsistent with decisions on actual retirement age, and incorporating retirement age uncertainty into a financial plan can significantly impact required retirement savings levels (blanchett, 2018). understanding that many retirement service products and financial planning engines integrate a self-reported retirement age into their design, benefits providers and employers may consider incorporating retirement uncertainty (such as the discrepancy between retirement expectations and reality discussed in this study) into the product design and implementation. the more accurately participant-facing calculators and financial plans 11 age 65 is the fra for those 50-year-olds in 1992 when they took the hrs survey. it is also an important represent the range of outcomes with associated probabilities, the better they can help workers achieve a successful retirement. using the longitudinal hrs data, this study examines the different socioeconomic and demographic factors that impact older americans’ actual and expected retirement ages, as well as the gaps between the two. we find that when even among retirees who were only ten years away from their retirement age (i.e., in their 50s), only one in six accurately predicted their first retirement age. in fact, more than half of the participants ended up retiring earlier than their expectations. although participants’ expected retirement ages are usually centered around two social security claiming ages (62 and 6511), their actual retirement ages are more likely to follow a negatively skewed (retire earlier) distribution. the most significant factors that influence participants’ retirement decisions relative to age for medicare. note that the full ssb fra is now around 67 for younger cohorts. liu et al. 44 expectations are health (+), wealth (-), age (+), change of marital status (+), mortality expectations (+), education levels (+), disability (-), and major illness diagnosis (-). focusing on these factors can help the retirement benefits community explore strategies to mitigate the negative consequences of gaps between retirement expectations and reality. this study also introduces several alternative definitions of “actual” retirement. since participants may go back to work after their first retirement, either full-time or part-time, their retirement well-being could be significantly impacted by their decisions to reenter the workforce. policymakers, employers, and retirement service providers may also take the participants’ potential multiple retirement periods into consideration. future studies could focus on the similarities and differences between these retirement definitions and investigate the patterns in which retirees stop working or reenter the workforce one or more times. references beehr, t. a. (2014). to retire or not to retire: that is not the question. journal of organizational behavior, 35(8), 1093– 1107. https://doi.org/10.1002/job.1965 blanchett, d. m. (2018). the impact of retirement age uncertainty on retirement outcomes. journal of financial planning, 31(9), 36–45. blanchett, d. (2015). dynamic choice and optimal annuitization. the journal of retirement, 3(1), 38-49. coile, c., & gruber, j. (2007). future social security entitlements and the retirement decision. the review of economics and statistics, 89(2), 234-246. coile, c., & zhang, h. (2022). recessions and retirement: new evidence from the covid-19 pandemic. wharton pension research council working paper (no. 20). cregan, c., kulik, c. t., perera, s., & sardeshmukh, s. (2023). when time is running out: a growth curve analysis of older workers' retirement intentions. journal of organizational behavior, 44(5), 760–775. https://doi.org/10.1002/job.2692 davis, o. (2021). employment and retirement among older workers during the covid19 pandemic. schwartz center for economic policy analysis (scepa), department of economics, the new school for social research. dubina, k. s., ice, l., kim, j. l., & rieley, m. j. (2021). projections overview and highlights, 2020–30. monthly labor review, 1-38. employee benefit research institute & greenwald research. (2021). 2021 retirement confidence survey. https://www.ebri.org/docs/defaultsource/rcs/2021-rcs/2021-rcs-summaryreport.pdf gallup. (2018). retirement age expectations among u.s. adults: record high, average still 66. retrieved from https://news.gallup.com/poll/210890/reti rement-age-expectations-among-adultsrecord-high.aspx haider, s. j., & stephens, m. (2007). is there a retirement-consumption puzzle? evidence using subjective retirement expectations. the review of economics and statistics, 89(2), 247–264. https://doi.org/10.1162/rest.89.2.247 hanspal, t., weber, a., & wohlfart, j. (2021). exposure to the covid-19 stock market crash and its effect on household expectations. review of economics and statistics, 103(5), 994–1010. https://doi.org/10.1162/rest_a_00936 hanspal, t., weber, a., & wohlfart, j. (2020). income and wealth shocks and expectations during the covid-19 pandemic. cesifo working paper (no. 8244). center for economic studies and ifo institute (cesifo), munich. montalto, c. p., yuh, y., & hanna, s. (2000). determinants of planned retirement age. financial services review, 9(1), 1–15. https://www.ebri.org/docs/default-source/rcs/2021-rcs/2021-rcs-summary-report.pdf https://www.ebri.org/docs/default-source/rcs/2021-rcs/2021-rcs-summary-report.pdf https://www.ebri.org/docs/default-source/rcs/2021-rcs/2021-rcs-summary-report.pdf https://news.gallup.com/poll/210890/retirement-age-expectations-among-adults-record-high.aspx https://news.gallup.com/poll/210890/retirement-age-expectations-among-adults-record-high.aspx https://news.gallup.com/poll/210890/retirement-age-expectations-among-adults-record-high.aspx https://doi.org/10.1162/rest.89.2.247 financial services review, 33(1) 45 rutledge, m., gillis, c., & webb, a. (2015). will the average retirement age continue to increase? center for retirement research working paper (no. 2015-16). https://crr.bc.edu/wpcontent/uploads/2015/07/wp_201516.pdf szinovacz, m. e., martin, l., & davey, a. (2014). recession and expected retirement age: another look at the evidence. the gerontologist, 54(2), 245257. vanderhei, j. (2022). impact of five legislative proposals and industry innovations on retirement income adequacy. employee benefit research institute. appendixes appendix table 1. ols on expected retirement age (hrs data 2020) ordinary least squares (ols) on expected retirement age expected retirement age ols results age in 2020 0.486*** -0.0232 female -0.0799 -0.3335 married 0.38 9 -0.3312 education years in 2020 0.0814 -0.0518 work affected (yes=1, no-0) -0.134 -0.3341 n 1,05 6 standard errors in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001 liu et al. 46 appendix table 2. did on expected retirement age (hrs data 2018 & 2020) difference-in-difference (did) on expected retirement age outcome age of expected variable social security standard error |t| p>|t| income before control 68.776 treated 68.666 diff (t-c) -0.11 0.379 -0.29 0.772 after control 69.015 treated 69.098 diff (t-c) 0.084 0.39 0.21 0.83 diff-in-diff 0.193 0.544 0.36 0.722 number of observations in the diff-in-diff: 2,199 before after control: 442 406 848 treated: 680 671 1,351 1,122 1,077 financial services review, 33(1) 47 appendix table 3. ols expected retirement age (prudential data 2018 -2021) 40 49 50 59 60 69 estimate std. error t value estimate std. error t value estimate std. error t value (intercept) 51.88 6.02 8.61 ** * 102.21 6.22 16.44 ** * 203.29 6.55 31.02 ** * age 1.24 0.27 4.58 ** * -1.16 0.23 -5.10 ** * -5.00 0.20 -24.48 ** * age_sq -0.01 0.00 -4.53 ** * 0.01 0.00 5.57 ** * 0.04 0.00 27.60 ** * ln_income -1.13 0.04 -28.02 ** * -0.74 0.03 -26.15 ** * 0.32 0.02 14.24 ** * married -0.15 0.05 -2.83 ** -0.41 0.04 -11.09 ** * -0.54 0.03 -18.27 ** * male -0.73 0.05 -15.84 ** * -0.43 0.03 -13.43 ** * -0.10 0.03 -3.81 ** * mon_2018_7 -2.50 0.28 -9.07 ** * -0.75 0.20 -3.74 ** * -0.05 0.18 -0.26 mon_2018_8 -2.76 0.37 -7.40 ** * -0.59 0.26 -2.32 * 0.47 0.19 2.44 * mon_2018_9 -2.20 0.31 -6.98 ** * -0.66 0.25 -2.69 ** -0.10 0.21 -0.49 mon_2018_10 -1.87 0.29 -6.51 ** * -0.11 0.22 -0.50 0.35 0.19 1.81 . mon_2018_11 -2.87 0.24 -11.77 ** * -0.84 0.19 -4.36 ** * 0.12 0.17 0.71 mon_2018_12 -2.80 0.28 -9.98 ** * -1.02 0.21 -4.84 ** * -0.20 0.18 -1.09 mon_2019_1 -2.65 0.21 -12.33 ** * -0.43 0.17 -2.49 * 0.41 0.15 2.68 ** mon_2019_2 -2.86 0.22 -13.23 ** * -0.87 0.17 -5.02 ** * -0.06 0.15 -0.38 mon_2019_3 -2.53 0.22 -11.24 ** * -0.47 0.18 -2.70 ** 0.17 0.15 1.11 mon_2019_4 -2.91 0.22 -13.51 ** * -0.93 0.17 -5.52 ** * 0.20 0.15 1.37 mon_2019_5 -2.90 0.20 -14.54 ** * -0.96 0.16 -5.88 ** * 0.06 0.15 0.43 mon_2019_6 -2.98 0.22 -13.43 ** * -0.83 0.17 -4.80 ** * -0.13 0.15 -0.82 mon_2019_7 -3.54 0.19 -18.43 ** * -0.96 0.16 -6.06 ** * 0.02 0.14 0.16 mon_2019_8 -3.18 0.21 -15.33 ** * -0.89 0.16 -5.45 ** * -0.06 0.15 -0.41 mon_2019_9 -2.96 0.21 -14.33 ** * -0.85 0.16 -5.17 ** * 0.06 0.15 0.40 mon_2019_10 -2.87 0.21 -13.94 ** * -0.72 0.16 -4.43 ** * 0.09 0.14 0.59 mon_2019_11 -2.85 0.22 -12.68 ** * -0.69 0.18 -3.91 ** * 0.13 0.16 0.82 mon_2019_12 -2.84 0.24 -12.04 ** * -0.72 0.18 -3.93 ** * -0.06 0.16 -0.35 mon_2020_1 -2.27 0.20 -11.49 ** * -0.63 0.16 -3.87 ** * 0.15 0.15 1.01 mon_2020_2 -2.38 0.20 -11.72 ** * -0.73 0.17 -4.39 ** * 0.15 0.15 0.98 mon_2020_3 -1.99 0.24 -8.27 ** * -0.19 0.19 -1.01 0.28 0.17 1.65 . mon_2020_4 -2.10 0.26 -8.24 ** * -0.34 0.21 -1.65 . 0.41 0.19 2.17 * mon_2020_5 -2.49 0.26 -9.73 ** * -0.77 0.20 -3.75 ** * 0.05 0.18 0.28 mon_2020_6 -3.11 0.23 -13.69 ** * -1.05 0.18 -5.78 ** * -0.08 0.16 -0.47 mon_2020_7 -2.81 0.21 -13.15 ** * -0.80 0.17 -4.66 ** * -0.19 0.15 -1.26 liu et al. 48 mon_2020_8 -2.88 0.20 -14.24 ** * -0.60 0.16 -3.67 ** * 0.04 0.15 0.26 mon_2020_9 -2.41 0.21 -11.64 ** * -0.54 0.17 -3.26 ** 0.01 0.15 0.06 mon_2020_10 -2.99 0.23 -12.86 ** * -0.90 0.18 -4.96 ** * -0.10 0.16 -0.62 mon_2020_11 -2.67 0.24 -11.29 ** * -0.96 0.18 -5.33 ** * -0.19 0.16 -1.23 mon_2020_12 -2.90 0.23 -12.60 ** * -0.90 0.18 -5.04 ** * -0.09 0.15 -0.60 mon_2021_1 -2.43 0.20 -11.93 ** * -0.64 0.17 -3.86 ** * -0.07 0.15 -0.50 mon_2021_2 -2.48 0.22 -11.19 ** * -0.84 0.18 -4.76 ** * -0.17 0.15 -1.09 mon_2021_3 -2.79 0.23 -12.28 ** * -0.76 0.18 -4.26 ** * -0.21 0.16 -1.34 mon_2021_4 -3.17 0.22 -14.13 ** * -1.05 0.17 -6.04 ** * -0.11 0.16 -0.68 mon_2021_5 -2.79 0.26 -10.76 ** * -0.90 0.20 -4.44 ** * -0.31 0.17 -1.78 . mon_2021_6 -2.97 0.27 -10.87 ** * -0.64 0.22 -2.96 ** 0.28 0.20 1.44 mon_2021_7 -2.58 0.26 -9.81 ** * -0.77 0.21 -3.72 ** * 0.11 0.18 0.62 mon_2021_8 -2.92 0.26 -11.09 ** * -0.71 0.21 -3.43 ** * 0.24 0.18 1.29 mon_2021_9 -2.77 0.27 -10.41 ** * -0.85 0.21 -4.04 ** * 0.13 0.19 0.70 mon_2021_10 -3.43 0.26 -13.18 ** * -1.01 0.20 -5.00 ** * 0.06 0.18 0.31 mon_2021_11 -2.53 0.26 -9.72 ** * -1.14 0.20 -5.66 ** * -0.03 0.18 -0.17 mon_2021_12 -2.41 1.05 -2.28 * -0.60 0.71 -0.84 -0.78 0.62 -1.26 financial services review, 33(1) 49 appendix figure 1. average gaps between actual and expected retirement ages across age groups note: after temporarily releasing the age restriction to 70 instead of 60, the number of primary respondents represented in this chart is 4,186 before weights are applied. after applying the 2018 respondent level analysis weight, it represents 7,444,380 older americans in the united states. -8.1 -8.5 -11.5 -10.9 -9.5 -11.5 -8.1 -7.7 -6.6 -6.0 -5.7 -5.2 -5.6 -3.7 -4.5 -3.4 -3.9 -4.0 -14 -12 -10 -8 -6 -4 -2 0 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 m ea n o f r ea lit y m in u s ex p ec ta ti o n participant age actual first retirement age minus average expected retirement age mean of reality minus expectation 2018 respondent level analysis weight applied financial services review, 33(4) i guest editorial teaching financial planning today emerging practices inga timmerman university of north florida, usa editorial introduction: special issue of financial services review it is with pleasure that i introduce this special issue of financial services review, devoted to advancing financial planning pedagogy. i would like to begin by thanking barry mulholland for initiating this project and shaping its vision. it has been my privilege to complete the issue and work with the contributing authors. over the past two decades, financial planning education has expanded significantly as universities respond to rising demand for academically trained professionals and increasing expectations for program quality. this growth has been driven by several macro trends, including heightened financial market complexity, shifting demographic needs, and broader recognition of the profession’s role in supporting individual and societal financial well-being. foundational work by warschauer (2002) underscored the essential role of universities in shaping the development of the financial planning profession, and subsequent research has documented similar expansion across multiple markets, including australia (bruce & gupta, 2011) and the united states (brady & o’neill, 2013). this trajectory has reinforced the need for systematic guidance in program development. recent studies of high-performing programs highlight elements such as evidencebased curriculum design, diversity and inclusion initiatives, and sustained faculty engagement as central to building effective educational pathways (heymann et al., 2025). at the same time, the literature has emphasized persistent pedagogical challenges. researchers have noted ongoing gaps between technical and behavioral skill development among emerging planners (jackling & sullivan, 2007), which is a critical concern given that effective financial planning requires not only quantitative expertise but also strong interpersonal, communication, and ethical reasoning capabilities. this need for flexible pedagogical models is further reinforced by international evidence. skultety et al. (2020) show that financial planning education is shaped by markedly different regulatory and accreditation requirements across australia, canada, the united kingdom and the united states. these distinctions influence curriculum design, program structure and competency expectations, underscoring the importance of educational frameworks that are capable of adjusting to varied professional standards. as regulatory and pedagogical demands continue to shape financial planning education, the articles in this collection show how educators are responding through curriculum development and intentional instructional practice. this collection brings together eight papers that examine how financial planning is taught and learned, spanning program development, course design, inclusive pedagogy and reflective practice. taken together, these contributions highlight the financial services review, 33(4) ii breadth of innovation occurring in classrooms and programs across the globe and illustrate the continuing evolution of financial planning as an academic discipline. the opening paper, “launching a cfp board-registered program at an aacsbaccredited business college,” examines the institutional landscape of financial planning education and offers a roadmap for universities seeking to develop or expand a registered program inside a college of business. by outlining key considerations in program approval, curriculum integration, and faculty engagement, this paper provides a valuable foundation for schools navigating growth within accredited business environments. from the program level, the issue moves into course-level design and application. “capstone as project-based learning: theory and application in personal financial planning” presents a model of problem-based learning in the capstone course, aligning academic preparation with professional expectations through project-led instruction. “elevating professional skills through authentic, scaffolded learning in a financial planning capstone” continues this theme by showing how scaffolded, real-world tasks can strengthen teamwork, communication, and digital competencies within a capstone setting. early-stage instruction and experiential pedagogy are explored in “utilizing experiential learning techniques in a financial planning program: allowing students to learn from themselves,” which highlights how reflective exercises in introductory courses prepare students for deeper engagement with personal finance and family dialogue. “a study of time value of money educational interventions” complements this by focusing on how financial calculators remain a critical instructional tool in teaching core quantitative concepts and reminding us that technology choice directly shapes learning outcomes. the following two papers shift toward personalized and theory-driven teaching approaches. “a personality-based approach to teaching financial planning and financial literacy” introduces a pedagogical framework that begins with personality assessment and selfreflection, encouraging students to understand who they are before defining their financial goals. “effective financial education strategies: empowering students with personal application” builds on this theme by integrating learning theories to promote instruction that is personally relevant, contextually grounded, and adaptable across student populations. the issue concludes with “addressing diversity, equity, and inclusion in financial planning education. drawing on the experiences of three professors who each developed and taught university-level dei courses within financial planning or related programs, the paper examines how diversity concepts can be meaningfully integrated into curricula. through thoughtful course design, learning objectives, and teaching philosophy, the authors demonstrate how inclusive pedagogy equips future professionals to navigate diversity challenges in both practice and research. together, these papers provide a comprehensive view of financial planning education, from institutional strategy to classroom innovation and inclusive practice. they offer evidence-based instructional approaches, creative program models, and pedagogical tools that educators can adopt and adapt to strengthen both student learning and professional readiness. looking ahead, this body of work also highlights several promising directions for future research. as financial planning programs expand and mature, there is a need for longitudinal studies that track how pedagogical financial services review, 33(4) iii approaches influence student competencies, professional readiness, and career outcomes over time. more research is needed on how technology, artificial intelligence, and digital learning tools reshape instructional design and student engagement. another important area is the need to explore how financial planning education can better integrate behavioral, cultural, and interdisciplinary perspectives to prepare students for an evolving profession. collectively, these future pathways will help strengthen the academic foundations of financial planning and support the continued growth of the discipline. my sincere thanks go to each author for their rigorous and thoughtful work, to the peer reviewers for their constructive feedback, and to the broader community of educators advancing this field. i hope this special issue inspires reflection, experimentation, and continued collaboration in the teaching and learning of financial planning. references brady, j.t., & o’neill, b. (2013). financial planning education in the united states: a survey of degree programs and cooperative extension initiatives. financial planning review, 6(1), 99–117. bruce, k., & gupta, r. (2011). the financial planning education and training agenda in australia. financial services review, 20(1), 61–74. heymann, r., schnusenberg, o., & timmerman, i. (2025). mapping the terrain: strategies for building effective university financial planning programs. financial planning review, https://doi.org/10.1002/cfp2.1198. jackling, b., & sullivan, c. (2007). financial planners in australia: an evaluation of gaps in technical and behavioral skills. financial services review, 16(3), 211-228. skultety, c., kavalamthara, p.j., & cull, m. (2020). financial planning education and regulatory requirements: a cross country comparison between australia, canada, united kingdom and united states of america. western sydney university. warschauer, t. (2002). the role of universities in the development of the personal financial planning profession. financial services review, 11(3), 201-216. https://doi.org/10.1002/cfp2.1198 75 a new approach to teaching personal financial education timothy s. griesdorn1 & sharon a. devaney2 abstract this paper outlines a new approach to teaching a financial planning/financial literacy course. it begins with personality assessments and other tools to help students build a solid foundation upon which they can create goals and develop a financial plan. in contrast, most financial planning textbooks outline the financial planning process, provide an overview of the economic system, and then focus on time value of money concepts and goal setting. all the texts start with the basic assumption that each student knows his or her goals. educators need to be sure their students have a solid foundation of self-awareness before they can plan for their future. a self-awareness foundation would include knowledge of core values and key strength competencies, awareness of the importance of strong relationship skills, hope for the future, expressions of gratitude, and meaningful work. without a solid foundation, goals and objectives cannot provide the same level of life satisfaction and happiness desired in life. sample assignments are provided that would help educators introduce these topics in an undergraduate course. implementation of these elements in an undergraduate personal finance course increased student evaluations of the course by 7%. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation griesdorn, t.s. & devaney, s.a. (2025). a new approach to teaching personal financial education. financial services review, 33(4), 75-86. introduction in the late 1990s and early 2000s, the academic community introduced over 90 financial education programs (vitt et al., 2000). even with the increase in financial literacy programs, fox, bartholomae, and lee (2005) still found a growing need to improve financial literacy levels in the u.s. the study on collegiate financial wellness (scfw) indicated that more college students were exposed to financial education in high school than in college (study on collegiate financial wellness, 2017). however, the research also identified the lack of improvement in overall financial literacy among students despite 1 corresponding author (tgriesdorn@gmail.com). university of the incarnate word, san antonio, texas, usa. 2 purdue university, west lafayette, indiana, usa. attending a financial education course (bartholomae & fox, 2021). one explanation for this outcome involves the tendency among individuals to resist engaging in meaningful change when they feel they are being told what to do. instruction that demands that an individual must save more money, avoid credit card balances, refrain from eating out, and so forth, can result in rejection of the principles taught in financial education programs. people seek education when there is a pressing need. using a personalized learning approach that links personal growth and happiness with financial education should provide for increased https://creativecommons.org/licenses/by-nc/4.0/ mailto:tgriesdorn@gmail.com financial services review, 33(4) 76 receptivity to the principles being offered (alamri et al., 2020). student participation in the process of learning is known to improve financial literacy (peng et al., 2007). the complexity and volume of financial education materials continue to grow exponentially, but financial educators have not examined how they teach the subject. there is little research that explores the sequence in which concepts are presented in financial planning education (financial literacy and education commission, 2006). as behavioral economics has demonstrated, how you present information, and what people see first, matters a lot (thaler & sunstein, 2021). literature review how do students learn best? there is a body of research that indicates student-centered and problem-based learning is beneficial for students. educators who study the science of teaching have long known the principles of effective teaching (nilson, 2016). unfortunately, personal financial planning authors are slow to implement these tools in their textbooks. human factors such as creating a welcoming classroom environment and enhancing student motivation are vital, as well as emphasizing the practical importance of the subject matter and delivering it with active learning techniques (nilson, 2016). additionally, the course should utilize the problem-based learning education approach (savery, 2015). students are presented with real-life situations and are tasked with developing a plan that works for them. often, what works best for their situation plays upon a unique strength they have acquired (rath & gallup, 2017). this proposed model strives to integrate the known pedagogical techniques into a financial planning classroom. student-centric learning starts with the first class and continues to be utilized throughout the entire course. this type of learning has been shown to increase engagement and intrinsic motivation (weimer, 2013). therefore, there is a need to develop an approach to teaching personal finance to college students that utilizes these tools of the science of learning to make it personal, relevant, and timely to the student. methodology the practice of financial planning requires good communication skills (grable & goetz, 2017). this course was designed as a writing-intensive course to assist students in developing their interpersonal communication skills. additionally, they view themselves as the financial planning client, in which they must engage in discovery tools and motivational interviewing techniques. the course was redesigned over the summer of 2019 with the assistance of two instructional designers and feedback from other faculty. the course was developed for a face-to-face classroom environment, but the covid-19 pandemic required the materials to be adapted for distance education. the course was developed for traditional undergraduate student population at the junior level. in addition, the course was intentionally designed such that no prerequisites would be required, thus any student could enroll to benefit from the basic financial education. in addition to the traditional coverage of principles of financial planning and the financial planning process, the first few weeks of the semester are devoted to self-discovery and reflection. the assignments are individualized to the student, so the student acts as the client who is seeking financial planning advice. therefore, the goals and recommendations should be meaningful and applicable for the students’ implementation. the course is personal, reflective, and writing intensive, all of which can be done either face-toface or in an online modality. there are a total of 10 writing assignments and two exams during the semester. students are asked to participate in a post-course evaluation survey. average enrollment in the class is approximately 30 students per semester. the goal of the course is to help students gain a better understanding of themselves, their unique strengths, their core values, and how to leverage these things to foster an increased subjective well-being utilizing the tools of the financial planning process to help them accomplish their goals. see table 1 for a summary of the course content. griesdorn & devaney 77 table 1. revised financial planning course overview course sequence question assessments topics covered weeks 1-3 who am i? strengthsfinder core values spirituality behavioral economics growth mindset motivational interviewing subjective well-being financial planning process vulnerability & trust weeks 4-10 why am i here? relationships gratitude vocation goals risk tolerance time value of money budgeting tax planning risk management financial well-being weeks 11-16 how do i want to be remembered? emotional intelligence money personality retirement estate planning legacy planning record keeping theoretical framework in well-being theory as proposed by martin seligman (2011), overall happiness consists of five distinct elements: positive emotion, engagement, relationships, meaning, and accomplishment (perma). the personal financial management course described in this paper seeks to build on these concepts and help students identify how to improve these traits in their lives through a variety of self-reflection exercises. in addition, feedback to students on these assignments is carefully structured to foster the growth mindset as proposed by carol dweck (2016). student feedback is structured using skill language and encouragement that skills are being improved. for example, student feedback could include appreciation for the thoughtful response while reminding them that attention to detail is an essential skill that could be developed further. finally, students are encouraged to examine their habitual behaviors to see if they help them flourish and become the type of person they desire to be. when they find undesirable behaviors, they are encouraged to replace them with more positive habits as recommended by research conducted by james clear (2018). thus, students are given practical tools for incorporating the new skills they have just learned and how to apply them in their lives for immediate reinforcement. revised course framework for teaching financial planning instructors should start with introspection by the students. if students strive to answer the question “who am i?” before they are exposed to the importance of the material, they may be more likely to attempt to apply financial tools in their lives. to engage students in introspection within a class, it is necessary to start with a shift in their perspective. in doing so, the instructor should explain that there are no right or wrong answers to questions about personality, spending habits, and money scripts. an essential step is to create a “brave” space for students to share their thoughts and feelings (brown, 2018). the first class sets the stage for this by stressing that all information shared in the class is confidential and not to be shared outside of the class without permission. students develop norms for classroom behavior and consequences if those norms are violated. as part of creating a “brave” space where students can feel comfortable sharing their thoughts and feelings openly, a rule is established that no zoom meetings will be recorded unless specifically mentioned in advance. therefore, zoom financial services review, 33(4) 78 meetings are only recorded upon request and for content covered from a lecture on a book chapter. students are then sent on a virtual scavenger hunt to find university resources that can help them during the semester. introspection comes with emotional risk, some students may not be comfortable with what they find out about themselves. making sure the students know what resources are available to them is important, particularly information about counseling resources available at their campus. the scavenger hunt also includes finding career services, library resources, financial aid, health, and tutoring services websites. see figure 1. figure 1. scavenger hunt you will want and need help to be successful with class assignments. each week there will be an assignment that will require you to dig deep into a particular topic. we will be starting with ourselves, our personality, hopes, dreams, desires, fears, strengths, and ambitions. fortunately, there are lots of places on campus eager to help you with various things. this assignment is to help you locate some of the various virtual resources and understand what kinds of help each can provide. your submission will include answers to the questions about each resource. use the campus website and other documents to answer the following questions: 1.) career services center a. what application is used to make appointments with the career services center? b. what three resources, tools, or services available from career services center interest you? 2.) writing and learning center a. how do you make an appointment with the writing and learning center? b. in your own words, what is the mission of the writing and learning center? 3.) testing services a. what services offered by testing services interest you? b. name at least one examination available through credit by examination. 4.) library a. what app allows you to reserve a seat or study space at the library? b. who is the business librarian? 5.) office of financial assistance a. what are two of the services provided by the igrad platform that interest you? b. what does bannerweb allow you to access? 6.) health services and counseling services a. what platform allows for virtual medical appointments (health services)? b. what are two apps recommended by counseling services for reducing anxiety and stress? what is the number to call if you are interested in receiving counseling (counseling services)? during the second class, discussion will focus on research by brené brown on vulnerability and her use of permission slips to identify additional norms that will be used during the semester (brown, 2018). a powerful question for students to ponder is “what will get in the way of my showing up and doing the work?” common examples include not enough time, griesdorn & devaney 79 procrastination, working a part-time job, being afraid to speak, being tired, stressed, distracted, and other possible reasons. brené brown (2018) suggests that whatever gets in the way will become the way. for example, when students indicate procrastination or lack of time as potential issues, class time is set aside to get started on the assignment or to complete some of the assigned work. the core values assignment draws upon brené brown’s (2018) work of identifying and living up to an individual’s core values as detailed in her book dare to lead. students are given a list of possible core values during class and are asked to circle all the potential core values that are meaningful to them. some words in the list of over 100 core values are fairness, faith, family, financial stability, freedom, friendship, fun, etc. most students circle 10 or more items and are given the opportunity to share some of the values they have circled. the instructor should emphasize that there are no right or wrong answers to these questions, but the answers should reflect the student when they feel they are being their best self. as part of the discussion, the question is asked, “what happens when these values conflict?” if everything is important, then how do tough decisions get made? now that students have started the assignment and better understand the importance of this task, they are asked to limit their core values to two and complete the reflection assignment. this approach comes with the tradeoff that they will have to do more of the textbook reading on their own so more class time can be used for contemplation and sharing. see figure 2. figure 2. core values assignment living into your values. using the list of values on page 188 of dare to lead, choose one or two values—the beliefs that are most important to you, that help you find your way in the dark, that fill you with a feeling of purpose. when selecting your values, ask yourself the following questions: • does this define me? • is this who i am at my best? • is this a filter that i use to make hard decisions? value 1: ____________ _________value 2: ________ _________ answer the following questions to investigate your values. value #1 _________________ 01. what are three behaviors that support your value? 02. what are three slippery behaviors that are outside your value? 03. what’s an example of a time when you were fully living this value? value #2 _________________ 01. what are three behaviors that support your value? 02. what are three slippery behaviors that are outside your value? 03. what’s an example of a time when you were fully living this value? the third class starts with a discussion on research related to gratitude and how expressing gratitude daily can improve an individual’s overall sense of well-being and lower instances of financial services review, 33(4) 80 depression (seligman et al., 2005). this discussion is combined with journal and paper exercises focused on gratitude. the first gratitude journal entry is completed in class along with identifying someone for whom they are grateful. this in-class use of time to start the assignment helps overcome procrastination and mental blocks associated with a blank page. only the teacher sees and responds to the journal entries. students keep a gratitude journal for seven days and write a gratitude letter to someone who they feel has made a positive difference in their life but they have failed to thank them adequately. the recipients of these letters often include parents, teachers, coaches, partners, and mentors. these letters often give the student the chance to establish a new, positive dialogue that has been correlated with increased subjective well-being (kaplan, 2016; seligman, 2011). research from janice kaplan (2016) and martin seligman et al. (2005) both show an increase in overall wellbeing associated with the expression of gratitude (see figure 3). figure 3. gratitude assignment research has shown that reflecting on positive things that have happened to us improves our overall feelings of happiness (seligman, 2011). take the next five minutes to do the following: in full sentences, write down three things you are grateful for and why. who is someone you are grateful for? who has made a favorable impact or impression on your life? (students hand this in at the end of the class) gratitude is a deeply personal subject that can change our outlook on life and overall sense of happiness or well-being. our brains are not naturally predisposed to express gratitude, but when we experience it, we are more likely to mirror it back. research continues to show that our thoughts, feelings, actions, body, mind, and spirit all influence each other. that indicates that if you change one thought, feeling, or action, you will likely change other things as well (seligman, 2011). this assignment has two parts. 1.) starting a gratitude journal and 2.) writing a reflection paper. for the next seven days your assignment is to find three things you are grateful for that day and record them on an online journal page that only you and the instructor will see. to complete the gratitude letter assignment, think of someone to whom you are very grateful, but whom you have never properly thanked. write a thoughtful, clear, brief (about 300-word) letter to that person. this letter should clearly describe the person’s contribution, how it was impactful at the time, and what it means today. reach out to that person and arrange a time when you can read the letter to them. post a copy of the letter to canvas. the next assignment involves completing the gallup cliftonstrengths assessment by rath and gallup (2017). this assessment helps students understand some of their unique talents and provides a common vocabulary to discuss strengths and positive psychology. results are shared in class, and a johari window (luft & ingham, 1955) exercise is used to help students identify underdeveloped strengths and develop an action plan to better utilize these strengths (see figure 4). figure 4. cliftonstrengths finder assessment in this assessment you will identify your top five signature strengths. things that you do better than other people, often without even realizing it. these strengths can really help you find your passion and the “right” kind of work to apply these strengths. to complete this assignment, you must take the strengthsfinder assessment, submit your top 5 strengths, and write a short narrative about what these strengths say about you. think about how you can share your strengths with the rest of the class. finally, write about what kinds of jobs you think would allow you to use your strengths each day. griesdorn & devaney 81 for this assignment, please complete the strengthsfinder assessment. once completed, analyze the strengths within the context of the johari window to determine those that currently reside in the open area (known to others and to you) and those in the blind spot (known to others and unknown to you). for example, someone with an “arranger” skill can handle complex issues and see how they should be arranged to be accomplished in the most productive way. to them, it is a natural process, but to someone without this skill, their ability to keep track of many things at one time is lacking. often, students think that their skills are nothing special and that everyone must have them because they come so naturally. however, abilities such as learning new things, setting and accomplishing goals, winning others over, positivity, and adaptability, when mastered through years of practice, become formidable personality attributes. the next step in the foundation-building process is to provide a sense of hope by teaching the growth mindset concepts by carol dweck (2016). this segment of the course stresses the importance of becoming a lifelong learner and being open to new experiences. the discussion encourages students to try something new and gives them time and patience to practice new skills. many students believe they have tried budgeting or making a spending plan before and it just didn’t work, or they are just not good at math. this class provides the opportunity to improve these skills. for example, many budgets or spending plans fail because of non-monthly expenses like car repairs, medical bills, or other expenses. course materials include activities to develop a personalized non-monthly expense list and estimate annual amounts spent on items like gifts, insurance, car repairs, medical expenses, clothing, etc. the course materials emphasize that these expenditures will happen, but the timing and amount might be unknown. the instructor can suggest that students could open a separate checking account and transfer a fixed monthly amount to this account to pay for non-monthly expenses like car repairs, gifts, insurance, and medical expenses. the focus is on skill development and making more optimal financial decisions (see figure 5). figure 5. example of non-monthly expenses set-aside accounts – a great way to ensure you don’t spend more than you make is to plan for future expenses and treat them as a monthly fixed expense. then set that money aside in a separate account to pay for these things as they happen. take the next five to ten minutes to do the following: write down all the categories of expenses that were unexpected. identify whether the expense was likely to happen on a monthly or annual basis. when possible, also include the estimated annual amount. 1.) ________________________________ 2.) ________________________________ total estimate of non-monthly expenditures on an annual basis. $____________________ you might use this as a starting point when completing the non-monthly expense form. review your bank and credit card statements to refresh your memory of significant non-monthly expenses. to conclude the introspection portion of the course, students should read chapter 8: purpose, in grit: the power of passion and perseverance by angela duckworth (2016) and complete assignments that require reflection on their purpose in life. in-class discussion can revolve around the difference between a job, a career, and a vocation in life. students investigate their potential career interests and write a one-to-twopage paper about the results of their study. once the purpose assignment is completed, students are ready to think and dream big about their future aspirations and goals (see figure 6). financial services review, 33(4) 82 figure 6. finding your passion, purpose, or vocation in life spend time thinking about how you will determine your purpose in life. for example, how do i want to be remembered at the end of my life? the following story helps to demonstrate how the same activity can be viewed as a job, or a career, or a vocation. parable of the bricklayers: (page 149 of grit by angela duckworth) three bricklayers are asked: “what are you doing?” the first says, “i am laying bricks.” the second says, “i am building a church.” and the third says, “i am building the house of god.” the first bricklayer has a job. the second has a career or profession. the third has a vocation. everyone has the same occupation, but their subjective experience is different. this assignment has several parts, make sure you clearly indicate what part you are responding to. 1.) read the material from the book grit on finding your passion/purpose. how does angela duckworth, the author of grit define purpose? what is the author’s top-level goal? 2.) create your own version of the parable of the bricklayer by selecting a job relevant to you and develop a listing of characteristics, thoughts, and actions of the person who has a job, a career, or a vocation. 3.) write a two-page paper that reflects upon finding your vocation. how will you know when you have found your vocation in life? how will knowing your vocation make a difference in your life? after students have a vision of their future and have identified goals, they are shown how to use the tools of the financial planning process to accomplish those goals. the tools include a variety of topics like budgeting, risk management, investing, estate planning, consumer loans, and ethics. because the students are now more self-aware, the practicality and usefulness of the tools becomes apparent. in previous courses, students have often chosen to work on homeownership as a common goal. they discuss how this goal relates to their core values, what they need to learn about the home buying process (loan qualifications and types, negotiation techniques, etc.), and how to optimally save for the down payment for this purchase. for example, opening a roth ira for this purpose results in significant tax and retirement benefits versus opening a taxable savings account for the same purpose. results the theoretical framework suggests students will need to master new skills to increase their overall subjective well-being. the perma model (seligman, 2011) indicates the skills will need to include relationships, understanding one’s strengths and building on them, setting and working towards goals, and finding meaning in life. utilizing carol dweck’s (2016) research on the growth mindset, students are given activities to develop their skills in each one of these areas. hence, meaningful goals can be established, and students are more receptive to learning how to use the tools of financial management to help them accomplish their goals. at the end of each semester students are asked to rate their class experience on a 5-point likert scale in 9 separate areas. in 2018 the average score for the 9 questions for this class was 4.56, after the changes were introduced, this average score increased to 4.9 and has remained at that level for the past 6 consecutive semesters. in addition, the student comments have been positive (see figure 7). additional data will be collected via student surveys to better understand the significance of taking this course. griesdorn & devaney 83 figure 7. sample principles of financial planning course feedback this figure provides excerpts from student feedback about the principles of financial planning course between 2018 and 2023. the perma model (seligman, 2011) was used to organize the responses. positive emotion: • "i have never had a class where the student was put before the course. it has made a positive impact on my life, and i am forever grateful." • “i would recommend this professor to anyone since he is someone that truly cares about his student’s success and wants all his students to learn. he creates a wonderful environment for students despite it being an online course.” engagement: • "this course allows you to ponder questions that aren't asked often. it opens your eyes and helps you think about what you want and value in life." • “i intend to take the skills and knowledge further by sharing them with others who may benefit.” relationships: • “i will always value the time spent in your office reflecting on some of the issues of life.” • "this course really helped me during a difficult time in my life." meaning: • "this is a course that is dedicated to helping you succeed in life." • “this has truly been a life changing experience for me.” accomplishment: • “the professor provided students with various assignments that not only allowed students to apply lessons learned but also helped students develop better writing skills, created time for introspection, and guided students by providing mentorship on how students can begin to work for personal goals.” limitations and future research the paper has a few limitations. the survey data is from students and is self-reported. future research would include: (a) qualitative research with a small group of students to explore which assignments were the most meaningful to them, and (b) a follow-up survey to see which financial behaviors were changed. also, the sample size is small. at this university, the course is taught once per semester to about 30 students per semester. the majority of students (average around 75%) are finance majors for whom the course is required, so the ability to extrapolate to other student majors is limited. students who enrolled in this course are predominantly male (70%), so the sample doesn’t reflect gender diversity in equal proportions. there was no control group, but this could be remedied with future research that compares students who took the course with those who have not. future research could include a post-course survey, with follow-up surveys 1-2 years later. additional questions could be included such as retention rates, time until graduation, employment status, and life satisfaction levels. this is exploratory work at one university; pilot studies could take place at additional universities around the country. additional longitudinal data should be collected with the results compared to current methods of teaching financial planning. implications previous research has emphasized the importance of financial education in enhancing individuals' ability to make informed financial decisions (qi et al., 2024). early exposure to personal finance equips students with the skills needed to navigate life's financial complexities (tang & peter, 2015). both the timing and frequency of financial education matter when making positive financial behaviors (augustin & martin, 2023). this innovative approach to teaching financial planning to college students aims to engage those financial services review, 33(4) 84 who might otherwise overlook the significance of the subject. a holistic approach is essential to help individuals internalize the value of financial education. additional results from baseline comparison surveys will be added as this research continues. potential implications of the research include increased student retention, decreased student loan debt, improved financial literacy and health insurance literacy, and a greater appreciation and affiliation with the university. one of the biggest reasons that students give for not continuing with a college education is financial concerns (stewart et al., 2015). if students can improve financial management skills, they may be able to graduate and do so with lower student loan balances. for many students, college is the first time they experience the need for financial management. financial aid packages are awarded to students at the beginning of each semester. then the student needs to be certain the money given to them plus any money earned lasts until the end of the semester, including transportation. an increasing number of students face food insecurity because of lack of resources to buy food and pay for their other expenses (ellison et al., 2021; richards et al., 2023). additionally, if a student drops out of school before earning a bachelor’s degree, then the average income for that student is not likely to increase compared to students who earn a degree. also, the student will have the additional burden of student loan debt repayment. the report on the condition of education 2021, by the u. s. department of education, shows average salaries of students graduating with an associate’s degree or some college earn an average of $39,000 per year, a little more than the average high school graduate with average earnings of $35,000 per year. however, the report shows the average person with a bachelor’s degree earns $55,700 per year (u.s. department of commerce, 2020). this worst-case scenario is happening often with the u. s. department of education reporting the average six-year completion rate for a bachelor’s degree to be 64% in 2020, and four-year completion rates much lower (u. s. department of education, 2021). it should come as no surprise that potential students are hesitant about starting college after high school and many question the value of a college degree. universities need to consider making changes in response to these trends. it is possible that one of the best changes that can be made is how we teach financial management skills and teaching these skills to all students. learning these new skills is an essential part of the maturation process and transition into adulthood (chatterjee et al., 2021). references alamri, h., lowell, v., watson, w., & watson, s. l. (2020). using personalized learning as an instructional approach to motivate learners in online higher education: learner self-determination and intrinsic motivation. journal of research on technology in education, 52(3), 322– 352. https://doi.org/10.1080/15391523.2020. 1728449 augustin, l. a. v., & martin, t. k. (2023). timing and frequency of financial education and positive financial behaviors. financial services review, 31(2), 151–167. https://doi.org/10.61190/fsr.v31i2/3.353 2 bartholomae, s., & fox, j. (2021). a decade review of research on college student financial behavior and well-being. journal of family and economic issues, 42(1), 154–177. https://doi.org/10.1007/s10834-02109756-6 brown, b. (2018). dare to lead: brave work, tough conversations, whole hearts. random house. chatterjee, s., kim, j., & chung, s. (2021). emerging adulthood milestones, perceived capability, and psychological well-being while transitioning to adulthood: evidence from a national study. financial planning review, 4(4), e1132. https://doi.org/10.1002/cfp2.1132 clear, j. (2018). atomic habits: an easy & proven way to build good habits & break bad ones. avery. https://doi.org/10.1080/15391523.2020.1728449 https://doi.org/10.1080/15391523.2020.1728449 https://doi.org/10.61190/fsr.v31i2/3.3532 https://doi.org/10.61190/fsr.v31i2/3.3532 https://doi.org/10.1007/s10834-021-09756-6 https://doi.org/10.1007/s10834-021-09756-6 https://doi.org/10.1002/cfp2.1132 griesdorn & devaney 85 duckworth, a. (2016). grit: the power of passion and perseverance. scribner. dweck, c. s. (2016). mindset: the new psychology of success. random house. ellison, b., bruening, m., hruschka, d. j., nikolaus, c. j., woerden, i. v., rabbitt, m. p., & nickols-richardson, s. m. (2021). viewpoint: food insecurity among college students: a case for consistent and comparable measurement. food policy, 101, article 102031. https://doi.org/10.1016/j.foodpol.2021.1 02031 financial literacy and education commission. (2006). taking ownership of the future: the national strategy for financial literacy. u.s. department of the treasury. https://files.eric.ed.gov/fulltext/ed4967 20.pdf fox, j. j., bartholomae, s., & lee, j. (2005). building the case for financial education. journal of consumer affairs, 39(1), 195– 214. https://doi.org/10.1111/j.17456606.2005.00009.x grable, j. e., & goetz, j. w. (2017). communication essentials for financial planners: strategies and techniques. john wiley & sons. kaplan, j. (2016). the gratitude diaries: how a year looking on the bright side can transform your life. dutton. luft, j., & ingham, h. (1955). the johari window: a graphic model of interpersonal awareness. in proceedings of the western training laboratory in group development. university of california, los angeles. nilson, l. b. (2016). teaching at its best: a research-based resource for college instructors (4th ed.). jossey-bass. peng, t. c. m., bartholomae, s., fox, j. j., & cravener, g. (2007). the impact of personal finance education delivered in high school and college courses. journal of family and economic issues, 28, 265– 284. https://doi.org/10.1007/s10834007-9058-7 qi, j., zhang, y., & worthy, s. (2024). social determinants of health and desirable financial behaviors: the mediation effect of financial knowledge. financial services review, 32(4), 51–77. https://doi.org/10.61190/fsr.v32i4.3608 rath, t., & gallup. (2017). strengthsfinder 2.0: discover your cliftonstrengths. gallup press. richards, r., stokes, n., banna, j., cluskey, m., bergen, m., thomas, v., bushnell, m., & christensen, r. (2023). a comparison of experiences with factors related to food insecurity between college students who are food secure and food insecure: a qualitative study. journal of the academy of nutrition and dietetics, 123(3), 438– 453. https://doi.org/10.1016/j.jand.2022.08.0 01 savery, j. r. (2015). overview of problem-based learning: definitions and distinctions. in a. walker, h. leary, c. e. hmelo-silver, & p. a. ertmer (eds.), essential readings in problem-based learning: exploring and extending the legacy of howard s. barrows (pp. 5–15). purdue university press. https://doi.org/10.2307/j.ctt6wq6fh.6 seligman, m. e. (2011). flourish: a visionary new understanding of happiness and well-being. free press. seligman, m. e., steen, t. a., park, n., & peterson, c. (2005). positive psychology progress: empirical validation of interventions. american psychologist, 60(5), 410–421. https://doi.org/10.1037/0003066x.60.5.410 stewart, s., lim, d. h., & kim, j. (2015). factors influencing college persistence for firsttime students. journal of developmental education, 38(3), 12–16, 18–20. study on collegiate financial wellness. (2017). 2017 key findings report. https://doi.org/10.1016/j.foodpol.2021.102031 https://doi.org/10.1016/j.foodpol.2021.102031 https://files.eric.ed.gov/fulltext/ed496720.pdf https://files.eric.ed.gov/fulltext/ed496720.pdf https://doi.org/10.1111/j.1745-6606.2005.00009.x https://doi.org/10.1111/j.1745-6606.2005.00009.x https://doi.org/10.1007/s10834-007-9058-7 https://doi.org/10.1007/s10834-007-9058-7 https://doi.org/10.61190/fsr.v32i4.3608 https://doi.org/10.1016/j.jand.2022.08.001 https://doi.org/10.1016/j.jand.2022.08.001 https://doi.org/10.2307/j.ctt6wq6fh.6 https://doi.org/10.1037/0003-066x.60.5.410 https://doi.org/10.1037/0003-066x.60.5.410 financial services review, 33(4) 86 https://cssl.osu.edu/posts/632320bc704d-4eef-8bcb87c83019f2e9/documents/2017-scfwkey-findings-report.pdf tang, n., & peter, p. c. (2015). financial knowledge acquisition among the young: the role of financial education, financial experience, and parents’ financial experience. financial services review, 24(2), 119–137. https://doi.org/10.61190/fsr.v24i2.3237 thaler, r. h., & sunstein, c. r. (2021). nudge: the final edition. penguin random house. u.s. department of commerce, census bureau, current population survey. (2020). annual social and economic supplement (asec), 2020: digest of education statistics 2020, table 502.30. u.s. department of education, national center for education statistics. (2021). integrated postsecondary education data system (ipeds), winter 2020–21: graduation rates component (digest of education statistics 2021, table 326.10). vitt, l. a., anderson, c., kent, j., lyter, d. m., siegenthaler, j. k., & ward, j. (2000). personal finance and the rush to competence: financial literacy education in the u.s. institute for sociofinancial studies. weimer, m. (2013). learner-centered teaching: five key changes to practice. john wiley & sons. https://cssl.osu.edu/posts/632320bc-704d-4eef-8bcb-87c83019f2e9/documents/2017-scfw-key-findings-report.pdf https://cssl.osu.edu/posts/632320bc-704d-4eef-8bcb-87c83019f2e9/documents/2017-scfw-key-findings-report.pdf https://cssl.osu.edu/posts/632320bc-704d-4eef-8bcb-87c83019f2e9/documents/2017-scfw-key-findings-report.pdf https://cssl.osu.edu/posts/632320bc-704d-4eef-8bcb-87c83019f2e9/documents/2017-scfw-key-findings-report.pdf https://doi.org/10.61190/fsr.v24i2.3237 financial services review, 32(4) 78 improving communication with financial consumers: insights from a study of phone call phobia wookjae heo,1 yi liu,2 and jae min lee3 abstract a phone call phobia is a symptom of avoiding real-time communication that ranges from mild nervousness to a debilitating fear of making or receiving phone calls. the situation of avoidance of real-time communication, including phone call phobia, can potentially influence personal finance, particularly if it limits an individual's ability to communicate effectively with not only family, friends, and employers but also financial service providers. as the covid-19 pandemic has affected ways of socializing and communicating, it is important to understand the situation and related factors of phone call phobia. this study, therefore, examined factors related to three types of phone call phobia (employer, family, and friends) and six types of communication preferences (face-to-face, phone calls, letters, email, text messaging, and online messaging apps). using data from an online survey conducted in 2021, this study considered a list of comprehensive factors, including psychological factors, financial-psychological factors, financial status, job-related factors, health-related behavior, and demographic factors, in ordered logistic regression and seemingly unrelated regression estimation models. the findings provide insights to improve communication between financial consumers and financial services providers. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation heo, w., liu, y., & lee, j. (2024). improving communication with financial consumers: insights from a study of phone call phobia. financial services review, 32(4), 78-102. introduction the covid-19 pandemic has affected ways of socializing and communicating (choi & choung, 2021). the pandemic increased social isolation and a shift towards virtual communication (gonzálezpadilla & tortolero-blanco, 2020), leading to more experiences of social anxiety in communication (caporucio, 2020). this issue can be more serious and have long-term effects on younger generations, 1 corresponding author (heo28@purdue.edu). purdue university, west layfette, in, usa 2 st. john fisher university, rochester, ny, usa 3 minnesota state university, mankato, mankato, mn, usa as they already have lacked face-to-face interactions and phone calls (rousselle, 2022). phone call phobia, a type of social anxiety disorder, can limit people’s exposure to phone calls and faceto-face interactions (liu et al., 2021). phone call phobia refers to avoidance or being worried about answering a phone call (bragazzi & del puente, 2014). phone call phobia can range from mild https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ heo et al. 79 nervousness to a debilitating fear of making or receiving phone calls (liu et al., 2021). younger generations have grown up with digital tools and are more likely to experience social anxiety disorder, with avoidance of real-time communication being a significant symptom (wang & zhang, 2015). for instance, generation z is more likely to use digital communication tools (auxier & anderson, 2021) and struggle with realtime in-person communication (pichler et al., 2021). the pandemic has aggravated phone call phobia in younger generations who have been more dependent on virtual communication in almost every aspect of their life (silveira et al., 2022), including education, work, and social interaction with family, friends, colleagues, and even businesses or organizations. a shift to virtual communication also presents challenges to people who already have a phobia of real-time conversation, as virtual meetings can lead to feelings of fatigue, exhaustion, and discomfort caused by direct eye gaze (karl et al., 2021). phone call phobia can affect daily functioning related to work, school, and healthcare (ozkan & solmaz, 2015), as well as the acquisition of financial skills and engagement in real-time financial activities. within the field of personal finance, the pandemic’s effect has not been evaded. with technological advancement, virtual communication trends, and pandemic isolation and social agility to changing situations, financial consumers often had to process virtual information themselves and make decisions based solely on types of information such as short videos without understanding or being provided with full contextual information (zhang et al., 2023). shortform videos (e.g., tiktok, youtube shorts, facebook reels, etc.) that were edited to meet time limits often omitted important information and highlighted only attractive or plausible messages, which can deliver misinformation to consumers (tam et al., 2022). for example, some short-form video messages on social media can include incorrect financial knowledge and information, but correction or caution is rarely made (o’sullivan et al., 2022). this situation of avoidance of or limited real-time communication can influence many areas of life, including personal finance, particularly if one's ability to communicate effectively with financial educators and financial service providers is limited due to the macro environment or situation such as the pandemic. individuals with phone call phobia may avoid making phone calls to their creditors or banks to address issues related to their finances, such as disputes over charges or changes to account terms (braithwaite, 2019). braithwaite (2019) criticized the situation that more people have begun preferring to use messaging apps instead of phone calls. this can lead to missed payments, late fees, and even damage to their credit score. additionally, phone call phobia can also impact job opportunities and career growth, which can have a direct impact on personal finance. although this can create many issues related to personal finance outcomes, little research exists to describe realtime communication and personal finance outcomes. in this sense, it is important to capture a profile of phone call phobia during the pandemic and identify factors related to phone call phobia in personal finance, which can be later used for future studies and provide implications. this study, therefore, aims to (a) describe a profile of phone call phobia during the pandemic; (b) identify related factors in the realm of personal finance by incorporating financial, psychological, and sociodemographic characteristics; and (c) examine which communication methods would work well for those with phone call phobia when interacting with financial practitioners. the findings of this study provide insights to improve communication between financial consumers and financial services providers. research questions to understand phone call phobia issues during the pandemic and identify related factors in the domain of personal finance, this study takes financial and psychological characteristics, such as financialpsychological factors and financial status, into consideration in addition to psychological, jobrelated, health-related behavioral, and demographic characteristics. this study also examined which communication methods could work well for financial consumers when interacting with financial practitioners. therefore, the research questions in this study are: financial services review, 32(4) 80 rq1. do psychological factors and financialpsychological factors result in people worrying about or avoiding answering phone calls? rq2. does financial status make people result in people worrying about or avoiding answering phone calls? rq3. what is the preferred communication method for those with phone call phobia? literature review phone call phobia often stems from fear or anxiety related to unexpected financial situations. increased control over personal finances can alleviate anxiety surrounding phone calls, as individuals feel better equipped to handle any potential financial matters (kamarudin et al., 2018) that may arise during a phone conversation. in this sense, having an emergency fund that shows a person’s ability to deal with unexpected financial problems (johnson & widdows, 1985) and serves as a safety net in the face of unforeseen events (anong & devaney, 2010) can contribute to a sense of financial security. consequently, it relieves anxiety related to potential financial crises through an increased, perceived control over their financial issues (kamarudin et al., 2018). psychological and financial-psychological factors associated with phone call phobia psychological factors such as self-esteem and locus of control were considered because phone call phobia is a psychological symptom. the literature suggests that there exists a positive correlation between self-esteem and locus of control with selfconfidence (bunker, 1991; jaaffar et al., 2019; owens, 1993; phares, 1962) and assertiveness (sarkova et al., 2013; williams & john, 1985), as well as stress (galanakis et al., 2016; pilisuk et al., 1993). this correlation may enable individuals to approach phone conversations with a more positive and proactive mindset, resulting in a reduced likelihood of experiencing phone call phobia. moreover, those who exhibit high self-esteem are more likely to have proficient communication skills (kang, 2017). they are more likely to engage in communication that is both efficacious and clear, enabling them to convey their thoughts, needs, and concerns confidently (sarkova et al., 2013), which may alleviate phone call phobia. moreover, individuals with high self-esteem have a lower fear of being rejected, are less concerned about others' opinions, and perceive others as accepting (leary et al., 1995). related research has confirmed that individuals who have a strong tendency to anxiously expect, perceive, and overreact to rejection tend to also suffer from low self-esteem (ayduk et al., 2000). individuals who possess an internal locus of control tend to perceive themselves as having a greater degree of personal control over their lives and outcomes (landau, 1995), while an external locus of control indicates a perception of outside factors such as luck and destiny as determinants of one’s fate and outcomes. as a result, those with an internal locus of control tend to exhibit increased confidence in problem-solving abilities (ng et al., 2006). this heightened sense of control may help lessen the anxiety associated with phone interactions and reduce the probability of developing phone call phobia. in general, understanding an individual's financialpsychological factors and psychological factors including their financial stress, financial satisfaction, self-esteem, and locus of control may be crucial in addressing their phone call phobia. in addition, financial-psychological factors such as financial stress and financial satisfaction may serve as indicators of phone call phobia. financial stress can relate to anxiety due to the fear of not meeting financial obligations or expectations (lee et al., 2023). the fear of disappointing others or falling short of people’s perceived financial expectations (even family and friends) may contribute to anxiety-inducing phone avoidance behavior. on the other hand, high financial satisfaction is associated with lower anxiety (archuleta et al., 2013), a higher sense of control over financial life (adiputra, 2021), and confidence (atlas et al., 2019). therefore, higher levels of financial satisfaction may alleviate phone call phobia, which could ease the pressure associated with phone conversations. financial status and phone call phobia with financial discussions involving matters like medical expenses, major repair payments, or creditor interactions often occurring over the phone, having an emergency fund can help individuals gain more confidence in handling their finances when they take phone calls without heo et al. 81 excessive fear or anxiety. similarly, homeownership provides individuals with a sense of belonging (liu et al., 2022) and represents a state of stability and control (rohe & stewart, 1996). an individual may feel more in control of their phone interactions and more empowered in their stable, safe, and private home environment (kleinhans & elsinga, 2010). taking more loans means having more financial responsibilities, which creates a sense of financial stress (archuleta et al., 2013). research has indicated a moderate association between debt and anxiety (archuleta et al., 2013), and financial loans can exacerbate financial stress and worry (french & mckillop, 2017; worthington, 2006). lenders send payment reminders to borrowers via mobile phones (bursztyn et al., 2019; du et al., 2020), so having more loans entails more conversation about payments, negotiating repayment plans, and other financial issues, which can cause borrowers to hesitate to initiate or accept phone calls related to loans. behavioral and demographic factors associated with phone call phobia furthermore, research has indicated that healthrelated behavioral factors, including alcohol consumption (higley et al., 1991; pohorecky, 1981), soda intake (zhang et al., 2019), and smoking (morrell et al., 2006; patton et al., 1996), can be associated with anxiety levels and thus may be relevant to phone call phobia. according to koval and pederson (1999), schuck and widom (2001), and wilsnack and wilsnack (1997), certain behaviors can serve as coping mechanisms or sources of temporary brief for anxiety symptoms. for instance, people who smoke may experience higher levels of anxiety in general (patton et al., 1996), and this worry may extend to their interactions with other people. when people are unable to smoke while talking on the phone, the absence of their coping technique may cause their anxiety levels to increase (jones & heffner, 2022) and lead to phone call phobia. consequently, by controlling for these variables, a more precise comprehension of the distinct influence of financial-psychological factors on phone call phobia can be attained, independent of potential confounding factors. notably, studies have indicated a positive correlation between job insecurity and heightened levels of anxiety and psychological distress (gallie et al., 2017). this association was particularly pronounced during the covid-19 pandemic (ganson et al., 2021; wilson et al., 2020). the phenomenon of phone call phobia may be impacted by job security, as it is posited that anxiety levels during work-related phone calls may be directly influenced by this factor. individuals experiencing job insecurity may perceive uncontrolled, notinitiated telephone conversations as a possible threat to their employment status, resulting in heightened anxiety and avoidance behaviors. demographic variables may be associated with phone call phobia (dienillah et al., 2018; forgays et al., 2014; hudson & o'regan, 1994; porath, 2011). it has been suggested that younger people may exhibit a greater inclination towards digital communication methods as they have grown up in a digital epoch (porath, 2011). conversely, individuals who are accustomed to telephone conversations as their predominant mode of communication are inclined to be more at ease with phone dialogues (forgays et al., 2014). they may have refined their communication skills in navigating phone interactions. the perceived social expectations and financial stability of individuals during phone conversations may be influenced by their marital status, education, and income (dienillah et al., 2018), which may affect their levels of anxiety during phone conversations. the number of children could contribute to additional stressors and obligations (hudson & o'regan, 1994), which may potentially influence their phone call phobia symptoms. communication preferences and phone call phobia previous research on communication preferences has extensively investigated how people choose between various modes of communication, such as face-to-face interactions, phone calls, letters, emails, text messaging, and online messaging apps like whatsapp, snapchat, and facebook messenger (pierce, 2009; robinson & stubberud, 2012; thayer & ray, 2006; yuan et al., 2016). these studies have shed light on the ever-changing nature of communication in modern culture. faceto-face communication, for example, is highly financial services review, 32(4) 82 valued due to its richness and ability to effectively convey nonverbal clues, making it suited for complex topics (meyer, 2006). while phone calls are extensively used, they can be influenced by the phenomena of "phone call phobia," influencing their communication method of choice (bragazzi & del puente, 2014). according to schneider et al. (2002), the utilization of text messaging resulted in shorter responses, which were perceived as lacking in elaboration, as compared to face-to-face contact. historically, letters and emails have served as conventional means for formal and business-oriented correspondence, while text messages and online messaging programs have emerged as convenient and instantaneous modes of communication in response to the demands of the contemporary, rapidly evolving digital landscape (alvermann, 2002). those who are financially stressed may prefer text messaging and online messaging apps, and they are not comfortable talking with others face-to-face (pierce, 2009). when it comes to interacting with employers, various communication channels are effective for job satisfaction and relationships with employers (braun et al., 2015; westerman & westerman, 2010). this implies that various communication channels are potentially associated with employees’ emotional responses such as anxiety, avoidance, and stress. similarly, conversations with friends and family members may also shift towards text messaging or instant messaging apps, which offer the comfort of composing messages at one's own pace (brown & michinov, 2017; manago et al., 2019; putnam, 2000). this switch to text-based communication may help individuals manage their social interactions and maintain connections, even while dealing with phone anxiety posed by phone call phobia. some may even turn to traditional letters, finding comfort in the absence of real-time conversation. methodology data the data used in this study were collected through an online survey conducted from january 12 to january 29, 2021, through an online survey agency. a random sampling method was used, with 5,906 individuals being contacted and 1,453 respondents participating in the survey. an online survey company randomly sends out survey invitations. among those who respond to these invitations, the company detects and excludes responses from bots or insincere participants. additionally, once a sufficient number of respondents from a specific demographic is reached, the survey design prevents further responses from individuals of the same demographic. the descriptive sample statistics are shown in table 1. the average age of the sample was 45.61 (sd = 18.37) and the sample included 42.95% males and 57.05% females. the average number of children in a household was 1.08 (sd = 2.11). about half of the respondents were single (49.62%). in terms of health-related behavior, the majority of respondents reported that they did not drink alcohol. for instance, 631 respondents did not consume beer; 618 respondents reported not consuming wine; 290 respondents did not consume soda; and 814 respondents reported not consuming cigarettes. although certain portion of respondents reported no consumption of beer, wine, soda, or cigarettes, the mean and standard deviation for are high because a portion of respondents reported a high consumption levels. for instance, trimming outliers reduced the means to more realistic values of s 4.86, 4.93, 10.22, and 11.96 for beer, wine, soda, and cigarettes, respectively. however, to avoid data distortion, respondents who consume high amounts of beer, wine, soda, and cigarettes were retained in our analysis. considering the relatively small gender disparity and the respondents’ concentrated age range of mid-40s, the data can be considered representative of the average working middle-aged american populations. in addition, the sample's diverse communication preferences enable a comprehensive exploration of different communication modes' impact on phone call phobia. including psychological factors like selfesteem and financial stress provides a comprehensive view of participants' well-being, aiding in understanding their influence on phone call phobia. behavioral data on alcohol, soda, and smoking add valuable insights. demographic diversity (age, gender, education, income) ensures the findings can be generalized across various population segments. job insecurity and work status data help understand how employment conditions affect phone call phobia. descriptions of the key variables are included in the next section. financial services review, 32(4) 83 table 1. sample characteristics (n = 1,453) sample characteristics mean s.d. freq. % phone call phobia by employer .68 1.04 by family .72 .98 by friends .54 .82 communication preference face-to-face 2.27 1.61 phone 2.56 1.22 letter 4.53 1.37 e-mail 3.80 1.28 text messaging 3.30 1.56 online messaging 4.54 1.67 psychological factors self-esteem 28.78 5.33 loc 17.46 5.69 fin-psycho factors fin stress 63.00 26.64 fin satisfaction 22.02 7.30 financial status emergency (= have) 751 51.69% homeownership (= own) 727 50.03% number of loans 1.12 1.29 job-related factors job insecurity 19.71 3.91 work status (= working) 857 58.98 behavioral factors alcohol (beer) 10.45 20.86 alcohol (wine) 10.63 21.17 soda 15.35 22.40 smoking 18.45 29.93 demographics age 46.51 18.37 female 829 57.05% single 721 49.62% education lower than high school 46 3.17% high school 367 25.26% associate degree 435 29.94% bachelor’s degree 424 29.18% graduate degree 181 12.46% income lower than $15k 246 16.93% $15k-$25k 214 14.73% $25k-$35k 202 13.90% $35k-$50k 203 13.97% $50k-$75k 247 17.00% $75k-$100k 154 10.60% $100k-$150k 132 9.08% over $150k 55 3.79% number of children 1.08 2.11 financial services review, 32(4) 84 measurements table 2 shows our variable measurements. for phone call phobia, three dichotomous items (yes, no) for each caller (employer, family, friends) were used based on previous studies (e.g., howard & sedgewick, 2021): (a) i feel nervous when my phone rings and it is from (employer, family, friends); (b) phone calls from (employer, family, friends) give me anxiety; and (c) i prefer texting instead of calling (employer, family, friends). all three items were asked by each caller (employer, family, and friends, respectively). a sum of the total “yes” answers to questions by each caller (employer, family, friend) was used to measure phone call phobia. there were thus three phone call phobia areas: (a) phone call phobia by employer, (b) phone call phobia by family, and (c) phone call phobia by friends. scores ranged from 0 (no phone call phobia) to 3 (highest phone call phobia). participants were asked to rank the order of their preferred communication methods. the following six communication methods were provided: (a) face-to-face, (b) phone call, (c) letter, (d) email, (e) text messaging, and (f) online messaging apps (e.g., whatsapp, snapchat, facebook messenger). the rank ranged from 1 (most preferred) to 6 (least preferred). for the analysis, the ranks were converted to reverse coding (1 = least preferred; 6 = most preferred) for ease of interpretation (the higher the number, the more preferred communication method). psychological factors were measured with selfesteem and locus of control. self-esteem was measured using 10 items on a 4-point likert scale (total 10–40) from rosenberg (1965), while locus of control was measured with seven items on a 5point likert scale (total 7–35) from perry and morris (2005). for financial-psychological factors, financial stress was measured with 24 items on a 5point likert scale (total 24–120) from heo et al. (2020), while financial satisfaction was measured with seven items on a 5-point likert scale (total 7– 35) from loibl and hira (2005). as the job-related factors, job insecurity was measured with seven items on a 5-point likert scale (total 7–35) from hellgren et al. (1999), while work status was measured as whether respondents were working or not (1, 0). health-related behavioral factors include alcohol consumption of beer (total number of beer bottles to drink per week) and wine (total number of wine glasses to drink per week), soda consumption (total number of soda cans to drink per week), and smoking (average number of cigarettes to smoke per week). demographic characteristics include age, gender (female = 1, male = 0), marital status (single = 1, otherwise = 0), education (less than high school = 1; high school graduate = 2; associate degree = 3; bachelor’s degree = 4; graduate or higher = 5), income (lower than $15k = 1; $15k–$25k = 2; $25k–$35k = 3; $35k–$50k = 4; $50k–$75k = 5; $75k–$100k = 6; $100k–$150k = 7; over $150k = 8), and the number of children. analytics two analyses were conducted to answer the research questions. to test and answer rq1 and rq2, ordered logistic analyses were utilized. the regressions were estimated as (model 1): 𝑙𝑜𝑔𝑖𝑡(𝑃(𝑌 ≤ 𝑖) = 𝑎𝑖 + 𝑏𝑖𝑙 ∑ 𝑃𝑠𝑦𝑖𝑙 + 𝑏𝑖𝑚 ∑ 𝐹𝑖𝑛𝑃𝑠𝑦𝑖𝑚 + 𝑏𝑖𝑛 ∑ 𝐹𝑖𝑛𝑖𝑛 + 𝑏𝑖𝑜 ∑ 𝐽𝑜𝑏𝑠𝑖𝑜 + 𝑏𝑖𝑝 ∑ 𝐵𝑒ℎ𝑖𝑝 + 𝑏𝑖𝑞 ∑ 𝐷𝑒𝑚𝑜𝑖𝑞 + 𝑒𝑖 … model 1 where psyl is psychological factors including selfesteem and locus of control; finpsym is financialpsychological factors including financial stress and financial satisfaction; finn is financial status including emergency fund, homeownership, and number of loans; jobso is job-related factors including job insecurity and work status; behp is behavioral factors including alcohol consumption, soda drinking, and smoking; demoq is demographic factors; and i denotes the ordinal number of phone call phobia level. this study used a subsample analysis for employer-related phone call phobia after excluding those who are not working. heo et al. 85 table 2. variable measurements item # and types minmax interpretation phone call phobia by employer by family by friends 3 binary items; total sum of 3 items 3 binary items; total sum of 3 items 3 binary items; total sum of 3 items 0 – 3 higher score = higher level of phone call phobia 0 – 3 higher score = higher level of phone call phobia 0 – 3 higher score = higher level of phone call phobia communication preference face-to-face 1 item; 6-degree ranking; reversely coded 1 – 6 higher number = more preferred phone 1 item; 6-degree ranking; reversely coded 1 – 6 higher number = more preferred letter 1 item; 6-degree ranking; reversely coded 1 – 6 higher number = more preferred e-mail 1 item; 6-degree ranking; reversely coded 1 – 6 higher number = more preferred text messaging 1 item; 6-degree ranking; reversely coded 1 – 6 higher number = more preferred online messaging 1 item; 6-degree ranking; reversely coded 1 – 6 higher number = more preferred psychological factors self-esteem 10 items; 4 points likert style scale 10 – 40 higher score = higher level of self-esteem locus of control 7 items; 5 points likert style scale 7 – 35 higher score = high level of external locus of control financial-psychological factors financial stress 24 items; 5 points likert style scale 24 – 120 higher score = higher level of financial stress financial satisfaction 7 items; 5 points likert style scale 7 – 35 higher score = higher level of financial satisfaction financial status emergency have = 1; otherwise = 0 homeownership own = 1; otherwise = 0 number of loans total number of loans 0 – 5 higher number = higher number of loans job-related factors job insecurity 7 items; 5 points likert style scale 7 – 35 higher score = higher level of perceived job insecurity work status working = 1; non-working = 0 health-related behavior beer total number of beer bottles to drink; weekly average 0 – 100 wine total number of wine glasses to drink; weekly average 0 – 100 soda total number of soda cans to drink; weekly average 0 – 100 financial services review, 32(4) 86 smoking total number of cigarettes to smoke; weekly average 0 – 100 demographics age actual age 20 – 88 female female = 1; otherwise = 0 single single = 1; otherwise = 0 education less than high school high school graduate associate degree bachelor's degree graduate degree reference category less than high school = 1; high school graduate = 2; associate degree = 3; bachelor's degree = 4; graduate degree or higher = 5 income less than $15k $15k-$25k $25k-$35k $35k-$50k $50k-$75k $75k-$100k $100k-$150k over $150k reference category less than $15k = 1; $15k-$25k = 2; $25k-$35k = 3; $35k-$50k = 4; $50k-$75k = 5; $75k-$100k = 6; $100k-$150k = 7; over $150k =8 number of children actual number of children in a household heo et al. 87 to answer rq3, seemingly unrelated estimation (sue) was utilized (model 2). the dependent variables were different across the estimated model, making the coefficients of each regression not directly comparable. we study used sue estimation to make marginal effects obtained from separate regressions comparable (weesie, 2000). sue also estimates more robust coefficients because all six models (a–f) in model 2 simultaneously estimate their coefficients by accounting for standard errors across six models. sue was introduced to estimate the comparable coefficients across multiple regressions (srivastava & gilles, 1987) as follows: 𝑌𝑗 = 𝑎𝑗 + 𝑏𝑗𝑙 ∑ 𝑃𝑝ℎ𝑜𝑏𝑖𝑎𝑗𝑙 + 𝑏𝑗𝑘 ∑ 𝑃𝑠𝑦𝑗𝑘 + 𝑏𝑗𝑙 ∑ 𝐹𝑖𝑛𝑃𝑠𝑦𝑗𝑙 + 𝑏𝑗𝑚 ∑ 𝐹𝑖𝑛𝑗𝑚 + 𝑏𝑗𝑛 ∑ 𝐽𝑜𝑏𝑠𝑗𝑛 + 𝑏𝑗𝑜 ∑ 𝐵𝑒ℎ𝑗𝑜 + 𝑏𝑗𝑝 ∑ 𝐷𝑒𝑚𝑜𝑗𝑝 + 𝑒𝑗 … models 2a to 2f where, pphobial is phone call phobia including a call from an employer, family, and friends; and j is the communication method including face-to-face (model 2a), phone (model 2b), letter (model 2c), e-mail (model 2d), text messaging (model 2e), and instant messaging apps (model 2f). results correlation analysis before the ordered logistic regression analysis was undertaken, the correlation between each financial stress factors (i.e., affective response to financial stress; relational response to financial stress; physiological response to financial stress) and each type of phone call phobia (i.e., phone call phobia from employer, phone call phobia from family, phone call phobia from friends) was checked. correlation tests were made to confirm the linearity of the relationships. as shown in table 3, the association between financial stress and phone call phobia was significant. figure 1 illustrates the lowess smoothing graphs showing the linear associations between two selected variables (cleveland, 1979; royston & cox, 2005). table 3. correlation between financial stress and phone call phobia fsa fsr fsp fs total emp pb fam pb frd pb fsa 1.00 fsr .74*** 1.00 fsp .69*** .85*** 1.00 fs total .89*** .93*** .92*** 1.00 emp pb .36*** .36*** .35*** .39*** 1.00 fam pb .32*** .36*** .37*** .38*** .43*** 1.00 frd pb .27*** .28*** .29*** .31*** .45*** .62*** 1.00 note. fsa is the affective response to financial stress; fsr is the relational response to financial stress; fsp is the physiological response to financial stress; fs total is the total sum of fsa, fsr, and fsp; emp pb is phone call phobia from employer; fam pb is phone call phobia from family; and frd pb is phone call phobia from friends. financial services review, 32(4) 88 figure 1. phone call phobia by financial stress and relationships financial and psychological factors associated with phone call phobia as shown in table 4, the relationship between financial stress and phone call phobia was significant and positive, as observed across all three types of callers: employers, family, and friends. this implies that individuals experiencing financial stress were more likely to exhibit a higher level of phone call phobia when it comes to interacting with three groups. when it comes to financial satisfaction, there was a partial association with phone call phobia, specifically in the context of phobia by friends. the result suggests that those with higher levels of financial satisfaction had lower odds of having a higher level of phone call phobia when interacting with friends. they were less likely to experience higher levels of phone call phobia. the number of loans a respondent held was positively associated with the level of phone call phobia, particularly about interactions with friends. this means that individuals with more loans were more likely to experience higher levels of phone call phobia when communicating with their friends. however, emergency and homeownership were not significant. job insecurity played a role in phone call phobia as well. those with higher levels of job insecurity were more likely to experience higher levels of phone call phobia when communicating with employers. furthermore, a respondent's work status had varying effects on phone call phobia depending on the caller. while those who were working were more likely to experience higher levels of phone call phobia when interacting with employers, they were less likely to experience higher levels of phone call phobia when interacting with family members and friends. this suggests that being employed contributes to higher levels of phone call phobia from employers, but it may alleviate the phobia from family and friends. self-esteem also played a role in phone call phobia, with a partial association in the context of phobia experienced when interacting with family members. the data indicates that individuals with lower levels of self-esteem were more likely to exhibit higher levels of phone call phobia when communicating with family. however, the external locus of control was not significant across types of phone call phobia models. among health-related behavioral factors, higher levels of smoking were more likely associated with higher levels of phone heo et al. 89 call phobia when communicating with both family and friends. this suggests that individuals who smoke more were more likely to experience higher levels of phone call phobia in their interactions with their family members and friends. however, alcohol and soda consumption were not significant. concerning the demographic factors, age was negatively related to the odds of exhibiting a higher level of phone call phobia. older individuals were less likely to experience higher levels of phone call phobia, suggesting that they could be more familiar with (or comfortable using) phone calls than younger generations. female respondents were more likely to experience higher levels of phone call phobia when interacting with friends. some income categories had positive associations with phone call phobia when interacting with employers and family. high-income individuals may experience phone call phobia when interacting with employers and family due to several reasons. first, they often face high expectations and pressures at work and home, making phone calls stressful because they might feel the need to respond immediately or communicate perfectly. they also value their time and prefer efficient communication methods like emails or scheduled meetings, making unscheduled phone calls more stressful. additionally, they might feel pressure to maintain a certain image or fear being judged based on their phone conversations. similarly, individuals in low to moderate income groups ($15-25k, $25-35k, $35-50k, $50-75k) face unique pressures that could exacerbate phone call phobia. this group is acutely aware of their job security and financial situation, understanding that employer-initiated phone calls often carry negative connotations such as job performance feedback or termination. unlike written correspondence or in-person meetings, phone calls can be a significant source of anxiety. furthermore, these individuals may be overwhelmed by the demands of both their job and family life, making phone calls an additional stressor. this heightened awareness and experience can intensify anxiety around phone interactions in both professional and personal settings. however, marital status, education, and the number of children were not significant. overall, our findings highlight the complex relationships between financial factors, job insecurity, work status, and communication contexts. these associations have implications for those interested in gaining a better understanding of financial stress and satisfaction related to individuals' communication patterns. financial services review, 32(4) 90 table 4. ordered logistic regression results (model 1) emp pb fam pb frd pb n = 857 n = 1,453 n = 1,453 b se b se b se psychological factors self-esteem -.02 .02 -.06*** .02 -.03 .02 locus of control .02 .02 .02 .01 .00 .02 financial-psychological factors financial stress .02*** .00 .01*** .00 .01** .00 financial satisfaction -.01 .01 -.01 .01 -.02* .01 financial status emergency -.12 .16 .09 .13 -.01 .13 homeownership -.02 .17 -.20 .14 -.17 .14 number of loans .11 .06 .08 .05 .14* .05 job-related factors job insecurity .06** .02 .02 .02 .01 .02 work status -.33* .14 -.35* .14 health-related behavior beer .00 .01 .01 .00 .00 .01 wine -.01 .01 .00 .00 .00 .00 soda -.00 .00 .00 .00 .00 .00 smoking .00 .00 .01*** .00 .01* .00 demographics age -.03*** .00 -.03*** .00 -.04*** .00 female -.07 .15 .15 .12 .27* .13 single -.10 .15 .03 .12 -.23 .13 education (less than high school) high school -.30 .47 -.15 .32 .03 .33 associate degree -.06 .48 .14 .32 .10 .33 bachelor's degree -.12 .49 .30 .33 .28 .34 graduate degree .23 .52 .20 .36 .12 .37 income (less than $15k) $15k-$25k .69* .28 .74*** .19 .22 .20 $25k-$35k .27 .28 .76*** .20 .16 .21 $35k-$50k .57* .28 .82*** .21 .33 .22 $50k-$75k .50 .28 .51* .21 .09 .21 $75k-$100k .46 .30 .40 .25 .26 .25 $100k-$150k .57 .34 .31 .28 .48 .28 over $150k .22 .44 .03 .38 .07 .37 number of children -.04 .03 .05 .03 -.01 .03 intercept 1 .95 .98 -.96 .74 -1.33 .77 intercept 2 1.99 .98 .52 .74 .56 .77 intercept 3 2.85 .98 1.54 .74 1.41 .77 chi2 194.96*** 458.83*** 307.90*** pseudo r2 .09 .14 .11 heo et al. 91 preferred communication method considering phone call phobia table 5 presents the coefficients and standard errors for model 2 (2a, 2b, 2c, 2d, 2e, 2f) that examined the associations between phone call phobia by employers and preferred communication methods. mixed results were observed across the models. those who exhibited higher levels of phone call phobia when interacting with their employer preferred phone calls and letters, while they preferred text messaging and instant messaging apps as their communication methods. those with higher levels of phone call phobia when interacting with family preferred text messaging. those with higher levels of phone call phobia when interacting with family were more likely to prefer text messaging. furthermore, no significant associations were observed between preferring other communication methods (phone, letter, email, text messaging, and online messaging) and phone call phobia within the family context. those who exhibited higher levels of phone call phobia when interacting with friends preferred face-toface and phone communications, while they preferred text messaging and instant messaging apps. concerning psychological factors, external locus of control was negatively related to email preference, meaning that those believing that outside factors, luck, or destiny determine and control one’s fate and outcomes rather than individuals having control over their actions and outcomes were less likely to prefer email. for financial-psychological factors, those with higher levels of financial stress were less likely to choose face-to-face as the preferred communication method but more likely to choose email as their preferred communication method. higher levels of financial satisfaction were negatively related to the letter while positively related to text messaging as the preferred communication method. for financial status factors, having an emergency fund was negatively related to face-to-face communication preference. the number of loans was negatively associated with phone calls and letters while positively related to online messaging apps as the preferred communication method. higher job insecurity was positively related to text messaging preference. for health-related behavior, those with higher levels of beer and soda consumption were less likely to choose letters as their preferred method of communication, while higher beer consumption was positively related to having a text messaging preference. higher smoking levels were positively related to letters, emails, and online messaging as the preferred communication method. for demographic factors, older respondents were less likely to choose face-to-face, phone calls, letters, and email as their preferred methods but more likely to choose text messaging and instant messaging apps. females were more likely to report a preference for face-to-face and email communications, but they were less likely to choose text messaging and instant messaging apps as their preferred communication channels. those having a higher education level compared to those with less than a high school, except a high school diploma, were less likely to choose email as their preferred communication method. some income groups were less likely to choose text messaging and instant messaging apps as their preferred methods, while those with a higher number of children were more likely to prefer email communication. however, self-esteem, homeownership, work status, wine consumption, and marital status were not related to preferred communication methods. in summary, the observed associations between phone call phobia and preferred communication methods varied by with whom they communicate and with what type of communication methods were used after controlling for psychological, financial-psychological, financial status, jobrelated factors, health-related behavior, and demographic characteristics. these findings indicate not only the factors related to preferred communication methods, but they also highlight the association between phone call phobia and communication preferences when approaching financial consumers. financial services review, 32(4) 92 table 5. sue results model 2a model 2b model 2c model 2d model 2e model 2f (face-to-face) (phone) (letter) (email) (text messaging) (online messaging) robust b robust se robust b robust se robust b robust se robust b robust se robust b robust se robust b robust se phone call phobia by employer -.06 .05 -.12** .04 -.16*** .04 .02 .04 .16*** .04 .16** .05 by family -.11 .06 -.09 .05 .01 .05 .08 .05 .11* .05 -.01 .06 by friends -.15* .07 -.16** .06 -.07 .05 -.03 .06 .19** .06 .22** .07 psychological factors self-esteem .00 .01 .00 .01 .00 .01 .00 .01 .00 .01 .01 .01 locus of control .01 .01 .01 .01 -.02 .01 -.02* .01 .01 .01 .01 .01 financial-psychological factors financial stress -.01* .00 .00 .00 .00 .00 .00* .00 .00 .00 .00 .00 financial satisfaction .00 .01 .00 .01 -.01* .01 -.01 .01 .03*** .01 .00 .01 financial status emergency -.19* .10 .00 .07 .07 .08 -.07 .08 .08 .09 .12 .09 homeownership -.09 .10 .03 .08 -.05 .09 .05 .08 .13 .09 -.07 .10 number of loans -.01 .04 -.06* .03 -.07* .03 .00 .03 .03 .04 .11** .04 job-related factors job insecurity .00 .01 -.01 .01 -.01 .01 -.01 .01 .03* .01 .01 .01 work status -.08 .10 .02 .08 .11 .08 .09 .08 -.06 .09 -.08 .10 health-related behavior beer .00 .00 .00 .00 -.01*** .00 -.01 .00 .01** .00 .01 .00 wine .00 .00 .00 .00 .00 .00 -.01 .00 .00 .00 .00 .00 soda .00 .00 .00 .00 -.01* .00 .00 .00 .00 .00 .00 .00 smoking .00 .00 .00 .00 .00** .00 .00** .00 .00 .00 .00** .00 demographics age -.01*** .00 -.01*** .00 -.01* .00 -.01*** .00 .01*** .00 .03*** .00 female .45*** .09 .10 .07 .11 .07 .15* .07 -.49*** .08 -.31*** .09 single .04 .09 -.13 .07 -.03 .08 .00 .07 .08 .08 .05 .09 heo et al. 93 education (less than high school) high school .08 .24 -.01 .17 -.13 .19 -.25 .18 -.11 .25 .42 .26 associate degree .03 .24 .08 .17 .03 .20 -.39* .18 -.01 .25 .26 .26 bachelor's degree .12 .25 .07 .18 .16 .20 -.44* .18 -.13 .25 .22 .27 graduate degree .11 .27 .12 .19 .19 .22 -.57** .20 .12 .27 .02 .29 income (less than $15k) $15k-$25k .21 .15 -.07 .12 -.01 .13 .21 .12 -.01 .14 -.32* .14 $25k-$35k .27 .15 .02 .11 .02 .13 .14 .12 -.23 .15 -.23 .15 $35k-$50k .39* .16 .15 .12 .19 .13 .14 .12 -.38** .14 -.49** .15 $50k-$75k .10 .15 .14 .12 .19 .13 .23 .12 -.28* .14 -.38* .15 $75k-$100k .38* .19 .00 .14 .11 .15 .34 .15 -.22 .17 -.61** .18 $100k-$150k .32 .19 .37 .15 .34 .16 -.05 .16 -.30 .17 -.69** .20 over $150k .30 .24 .08 .16 .10 .22 .17 .21 -.46 .24 -.19 .26 number of children .00 .02 -.02 .02 -.02 .02 .04* .02 .02 .02 -.01 .02 intercept 2.50*** .60 2.77*** .45 5.63*** .48 5.11*** .46 1.89** .58 3.10*** .60 f 4.97*** 6.53*** 6.98*** 4.87*** 10.95*** 11.01*** adjusted r2 .08 .11 .11 .08 .18 .18 financial services review, 32(4) 94 discussion and conclusion financial status, job insecurity, and financial effect our findings suggest that financial characteristics (i.e., financial status, job insecurity, and financialpsychological factors) were the major factors contributing to higher levels of phone call phobia. this brings the attention of financial service providers to a potential communication hurdle. significant factors were all associated with personal finances, such that (a) a respondent has a higher level of financial stress and a lower level of financial satisfaction, (b) a respondent has a higher number of loans, and (c) a respondent has a higher level of job insecurity or not working status. first, the current study reveals a relationship between financial stress and phone call phobia, regardless of the caller's identity (rq1). the presence of perceived psycho-physiological symptoms due to one’s own financial situation can relate to a phone call phobia experience. for example, those with higher levels of financial stress were more likely to experience higher levels of phone call phobia when interacting with employers. financial stress may relate to worry or pressure about work performance and create elevated levels of anxiety in interactions over the phone with employers. it appears that overall financial stress interferes with social and interpersonal relationships. those with higher levels of financial stress were also likely to experience higher levels of phone call phobia with family and friends. this aspect may show the interconnectivity between financial stress and other areas of life across types of interpersonal and social relationships. for those who feel higher levels of financial stress, alternative ways to reach out to them should be considered. our findings also show that the presence of phone call phobia resulting from social interactions with friends may be mitigated by higher levels of financial satisfaction, as it can help alleviate anxiety and concerns associated with conversations, including financial aspects, relating to other areas of life. the results may be related to the positive association between life satisfaction and interpersonal relationships. for example, proctor and linley (2014) found that youths with higher levels of life satisfaction can have positive life outcomes, such as adaptive psycho-social functioning, better interpersonal and social relationships, fewer behavioral problems, and better academic achievement. higher financial satisfaction as an important aspect of life can make one feel more competent and confident in social interactions without comparing their financial situation to others’ or feeling envious (issa, 2023). as a result, this reduced pressure can curb the hesitancy to participate in telephone conversations with friends, thereby promoting deeper social interactions and interpersonal connections. second, our findings support the role of financial status in explaining different levels of phone call phobias (rq2). the total number of loans was associated with the probability of experiencing higher levels of phone call phobia, particularly about calls from friends. although this study did not measure the amount of debt directly, the higher number of loans may represent financial situations that can make debtors experience greater levels of financial instability or challenges. as individuals experience higher levels of debt, they may have a diminished willingness to communicate with others outside their immediate familial sphere, which would affect interpersonal relationships with their friends. this phenomenon can be attributed to the likelihood of receiving calls from debt collectors or discussing their financial issues with others, including friends, leading to heightened discomfort when communicating over the phone. this finding indicates that financial status can contribute to phone call phobia levels, potentially attributable to an individual's social interactions and communication patterns in broader domains. third, our findings show that if someone worries about their job security, the person will likely experience higher levels of phone call phobia. receiving a phone call from employers generally occurs in a professional context and may entail a discussion about work-related subjects, such as job responsibilities, deadlines, or performance feedback. for individuals who are already concerned about their employment status, these phone calls can serve as stimuli that increase their anxiety level. the work status was also identified as a factor of phone call phobia by employers despite mixed results. when respondents were heo et al. 95 working, they were likely to experience lower levels of phone call phobia when communicating with family and friends, while they were more likely to experience higher levels of phone call phobia when communicating with their employer. this implies that a working person avoids phone calls from an employer but does not avoid phone calls from family or friends. this may reflect different characteristics and content of phone calls at work for the hierarchical relationship between employer and employees, which can make employees feel a heightened sense of pressure to perform work-related responsibilities during phone calls. however, employed individuals may feel more confident and comfortable communicating with their family and friends. this can be attributed to the emotional support and establishment of a safe environment that family and friends typically offer even if they are stressed out at work from phone calls with employers. however, if they were not working, this may create another pressure or burden of communicating even with family and friends over the phone, which may be related to their lack of confidence or complex situations without employment. psychological, behavioral, and demographic effects findings from the current study suggest that high self-esteem is negatively associated with phone call phobia, particularly about receiving phone calls from family members. this phenomenon could be attributed to individuals with high self-esteem who have a stronger sense of self-assurance and selfacceptance. they often exhibit a more positive outlook on life (caprara & steca, 2005), associated with enhanced emotional stability (crowe et al., 2016), while they are less likely to be overly concerned with external opinions (leary et al., 1995), including prescriptive and prohibitive advice from family. this emotional resilience and reduced concern about judgment or criticism from others enable them to engage in phone conversations, thereby reducing the probability of experiencing phone call phobia symptoms. the finding may also reflect the familiarity and supportive nature inherent in family relationships. the family relationship and established bonds reduce the anxiety associated with phone conversations, as individuals feel more accepted during such interactions over the phone. after all, family members are frequently perceived as a source of support and approval, particularly when confronted with challenges. our health-related behavior findings indicate that individuals who smoked more exhibited a higher level of resistance to receiving phone calls from family and friends. as a coping method, smokers tend to smoke more frequently when anxious (hughes et al., 1990) and believe that smoking can reduce anxiety or distress levels (zvolensky et al., 2001). smokers are more likely to demonstrate higher levels of distress than nonsmokers (parrott, 1999). in this sense, those smoking more frequently may present higher levels of distress and anxiety in general, experiencing phone call phobia. family and friends’ concerns and expectations may result in smokers’ discomfort with phone calls as smokers may feel frustrated with or unable to handle their close ones’ concerns and expectations. furthermore, findings suggest that age may play a role in reports of high phone call phobia. in comparison to younger respondents, older respondents were less likely to experience elevated levels of phone call phobia. aging is generally associated with maturity (camberis et al., 2014) and a broader range of life experiences and responsibilities in social and personal relationships that they cannot avoid. they could have more experience communicating over the phone than younger people with more alternative communication channels, not leading to experiencing phone call phobia in any circumstance. communication method preference regarding rq3, text messaging was the preferred communication method for those with phone call phobia regardless of the type of interaction. this may explain why text messaging is so widely used and a preferred communication method when communicating with employers, family, and friends. online messaging apps were also preferred by those with phone call phobia in employer and friend relationship situations. a letter was not a preferred communication method for those with an employer interaction phone call financial services review, 32(4) 96 phobia; this might be because typically a letter from an employer may contain more formal (and often negative) information. letters, as a legal form of documentation, may include more crucial information or a unique situation (e.g., termination of employment). those with higher levels of employer phone call phobia preferred virtual communication through online messaging apps. this may reflect the widespread use of messaging apps in business settings. as the rise in remote work has led to an increased dependence on mobile devices to stay connected, third-party messaging apps are widely used within organizations that can streamline communication and increase collaboration among team members (bibb, 2023), although this type of communication can create privacy and data breach issues (goldstein, 2023). face-to-face communication was only significant among those with phone call phobia when interacting with friends. having interactions with friends can be optional and involves alternatives that can be more freely chosen; they can use text messaging and online messaging apps with their friends. in general, our findings show that phone calls and letters were less preferred across most determinants including phone call phobia, psychological, financial-psychological, financial status, jobrelated, health-related behavior, and most demographic factors, while text messaging and online messaging apps were preferred across those determinants except for gender and income. as the non-phone call phobia factors showed, findings indicate trends in communication methods based on convenience (e.g., mobile use) and control over conversations (e.g., choosing when to communicate). implications phone call phobia can potentially influence personal finance outcomes, particularly if it limits an individual's ability to communicate effectively with financial service providers. thus, financial services providers should understand financial consumers’ phone avoidance or phone call phobia and implement more effective communication strategies to approach them. to effectively communicate with clients, financial services providers should first build trust and rapport with their clients. financial services providers should demonstrate empathy and create a safe environment for discussions about financial matters by understanding factors related to levels of phone call phobia and a client's preferred communication approach. this approach strengthens relationships and facilitates open dialogue, thereby empowering clients to express their concerns and goals with greater ease. even though regulators, such as the financial industry regulatory authority (finra), the u.s. securities and exchange commission (sec), and the certified financial planner board of standards (cfp board), require some forms of communication to be documented via letters (sometimes in the form of a certified letter), financial service providers should explore other communication channels for day-to-day interactions. financial service firms are expected to establish and implement policies and procedures to manage client interactions across diversified communication channels while protecting informational security. recent enforcement actions by the sec highlight the importance of these policies. the sec has imposed significant penalties on firms for misuse of personal devices and lack of record keeping (sec, 2023). to mitigate risks, the sec suggests that financial firms should communicate digital communications rules including permissible and prohibited behaviors in electronic messages to their employees. additionally, understanding client and business partner communication preference is recommended to assess potential risks and update their policies accordingly (sec, 2018). for example, financial service firms may need to provide official phone numbers for client contact as outlined in agreement letters to facilitate phone and text messaging in their agreement letter while maintaining compliance. these channels may help financial professionals tailor their approach to meeting clients’ needs. financial services providers should also develop proficient communication techniques to address the anxiety caused by phone call phobia based on trust and individualized communication channels. this personalized communication style can foster stronger interpersonal relationships, cultivate trust, and facilitate open and effective communication. heo et al. 97 limitations this study provides insights into phone call phobia and its association with various personal finance factors and outcomes. nevertheless, it is important to acknowledge certain limitations in this research. first, this study employed a cross-sectional design, which limits its ability to establish causality. to strengthen future research, a longitudinal design would be recommended to capture changes over time and to establish more robust causal relationships. second, the study’s reliance on selfreported data for psychological factors, communication preferences, and alcohol, soda, and smoking consumption may introduce potential biases and measurement error. for example, the high variability in alcohol, soda, and smoking consumption complicates analysis and interpretation. moreover, the sample’s demographic imbalances in education and income and lack of detailed contextual information could constrain the generalizability of the results. to mitigate these limitations, future research should consider employing objective measures, expanding sample diversity, and collecting more in-depth contextual data. third, although this study attempted to examine a comprehensive list of factors associated with phone call phobia, there could have been other variables, such as personality traits, prior traumatic experiences, or other psychological disorders that could have changed the direction and magnitude of some of the results reported in this study. future research should take into consideration these and other theoretically associated factors. references adiputra, i. g. (2021). the influence of financial literacy, financial attitude and locus of control on financial satisfaction: evidence from the community in jakarta. kne social sciences, 636-654. alvermann, d. e. (ed.). (2002). adolescents and literacies in a digital world (vol. 7). peter lang. archuleta, k. l., dale, a., & spann, s. m. (2013). college students and financial distress: exploring debt, financial satisfaction, and financial anxiety. journal of financial counseling and planning, 24(2), 50-62. atlas, s. a., lu, j., micu, p. d., & porto, n. (2019). financial knowledge, confidence, credit use, and financial satisfaction. journal of financial counseling and planning, 30(2), 175-190. auxier, b., & anderson, m. (2021, april 7). social media use in 2021. pew research center. https://www.pewresearch.org/internet/2021/0 4/07/social-media-use-in-2021/ ayduk, o., mendoza-denton, r., mischel, w., downey, g., peake, p. k., & rodriguez, m. (2000). regulating the interpersonal self: strategic self-regulation for coping with rejection sensitivity. journal of personality and social psychology, 79(5), 776. bibb, a. (2023, june 13). revitalizing the employee experience: tailoring your internal communications. forbes. https://www.forbes.com/sites/forbesbusiness council/2023/06/13/revitalizing-theemployee-experience-tailoring-your-internalcommunications/?sh=921c6d13305f bragazzi, n. l., & del puente, g. a. (2014). proposal for including nomophobia in the new dsm-v. psychological research and behavior management, 7, 155–60. braun, s. bark, a. h., & van dick, r. (2015). emails from boss – curse or blessing? relations between communication channels, leader evaluation, and employees’ attitudes. international journal of business communication, 56(1), 50-81. brown, g., & michinov, n. (2017). cultural differences in garnering social capital on facebook: french people prefer close ties and americans prefer distant ties. journal of intercultural communication research, 46(6), 579–593. bunker, l. k. (1991). the role of play and motor skill development in building children's selfconfidence and self-esteem. the elementary school journal, 91(5), 467-471. bursztyn, l., fiorin, s., gottlieb, d., & kanz, m. (2019). moral incentives in credit card debt https://www.pewresearch.org/internet/2021/04/07/social-media-use-in-2021/ https://www.pewresearch.org/internet/2021/04/07/social-media-use-in-2021/ financial services review, 32(4) 98 repayment: evidence from a field experiment. journal of political economy, 127(4), 16411683. camberis, a. l., mcmahon, c. a., gibson, f. l., & boivin, j. (2014). age, psychological maturity, and the transition to motherhood among english-speaking australian women in a metropolitan area. developmental psychology, 50(8), 2154. caprara, g. v., & steca, p. (2005). affective and social self-regulatory efficacy beliefs as determinants of positive thinking and happiness. european psychologist, 10(4), 275-286. caporucio, j. (2020, may 13). how the covid-19 pandemic affects people with social anxiety. medical news today. https://www.medicalnewstoday.com/articles/ social-anxiety-and-covid-19 choi, m., & choung, h. (2021). mediated communication matters during the covid-19 pandemic: the use of interpersonal and masspersonal media and psychological wellbeing. journal of social and personal relationships, 38(8), 2397-2418. cleveland, w. s. (1979). robust locally weighted regression and smoothing scatterplots. journal of the american statistical association, 74, 829–836. crowe, m. l., lopilato, a. c., campbell, w. k., & miller, j. d. (2016). identifying two groups of entitled individuals: cluster analysis reveals emotional stability and self-esteem distinction. journal of personality disorders, 30(6), 762775. dienillah, a. a., anggraeni, l., & sahara, s. (2018). impact of financial inclusion on financial stability based on income group countries. buletin ekonomi moneter dan perbankan, 20(4), 429-442. du, n., li, l., lu, t., & lu, x. (2020). prosocial compliance in p2p lending: a natural field experiment. management science, 66(1), 315-333. forgays, d. k., hyman, i., & schreiber, j. (2014). texting everywhere for everything: gender and age differences in cell phone etiquette and use. computers in human behavior, 31, 314-321. french, d., & mckillop, d. (2017). the impact of debt and financial stress on health in northern irish households. journal of european social policy, 27(5), 458-473. galanakis, m. j., palaiologou, a., patsi, g., velegraki, i. m., & darviri, c. (2016). a literature review on the connection between stress and self-esteem. psychology, 7(5), 687-694. gallie, d., felstead, a., green, f., & inanc, h. (2017). the hidden face of job insecurity. work, employment and society, 31(1), 3653. ganson, k. t., tsai, a. c., weiser, s. d., benabou, s. e., & nagata, j. m. (2021). job insecurity and symptoms of anxiety and depression among us young adults during covid-19. journal of adolescent health, 68(1), 53-56. goldstein, m. (2023, august 8). banks fined $549 million over use of whatsapp and other messaging apps. new york times. https://www.nytimes.com/2023/08/08/busine ss/banks-fines-whatsapp-records.html gonzález-padilla, d. a., & tortolero-blanco, l. (2020). social media influence in the covid19 pandemic. international brazilian journal of urology, 46, 120-124. hellgren, j., sverke, m., & isaksson, k. (1999). a two-dimensional approach to job insecurity: consequences for employee attitudes and well-being. european journal of work and organizational psychology, 8(2), 179-195. heo, w., cho, s., & lee, p. (2020). apr financial stress scale: development and validation of a multidimensional measurement. journal of financial therapy, 11(1), 2. higley, j. d., hasert, m. f., suomi, s. j., & linnoila, m. (1991). nonhuman primate model of alcohol abuse: effects of early experience, personality, and stress on alcohol https://www.medicalnewstoday.com/articles/social-anxiety-and-covid-19 https://www.medicalnewstoday.com/articles/social-anxiety-and-covid-19 https://www.nytimes.com/2023/08/08/business/banks-fines-whatsapp-records.html https://www.nytimes.com/2023/08/08/business/banks-fines-whatsapp-records.html heo et al. 99 consumption. proceedings of the national academy of sciences, 88(16), 7261-7265. hudson, s. a., & o'regan, j. (1994). stress and the graduate psychology student. journal of clinical psychology, 50(6), 973-977. hughes, j. r., higgins, s. t., & hatsukami, d. (1990). effects of abstinence from tobacco: a critical review. research advances in alcohol and drug problems, 10, 317-398. issa, e. e. (2023, july 11). survey: 54% who feel money envy say it harms their mental health. nerdwallet. https://www.nerdwallet.com/article/finance/s urvey-money-envy-is-common-and-54-whofeel-it-say-it-harms-their-mental-health jaaffar, a. h., ibrahim, h. i., rajadurai, j., & sohail, m. s. (2019). psychological impact of work-integrated learning programmes in malaysia: the moderating role of self-esteem on relation between self-efficacy and selfconfidence. international journal of educational psychology. johnson, d., & widdows, r. (1985, march). emergency fund levels of households. in the proceedings of the american council on consumer interests 31st annual conference (vol. 235, p. 241). jones, s. m., & heffner, j. l. (2022). financial anxiety: a potential new target to increase smoking cessation. future oncology, 18(9), 1035-1038. kamarudin, n. s., ramli, n. a., & rasedee, a. f. n. (2018). why preparing an emergency fund is matter to young adults? international journal of engineering and technologies, 7(4), 172-175. kang, s. w. (2017). the influence of voluntary service activities on self-esteem and communication skills among korean university students in healthcare and medical fields. journal of nursing & healthcare, 2(1), 1-6. karl, k. a., peluchette, j. v., & aghakhani, n. (2021). virtual work meetings during the covid-19 pandemic: the good, bad, and ugly. social group research, 53(3), 343-365. kleinhans, r., & elsinga, m. (2010). “buy your home and feel in control”: does home ownership achieve the empowerment of former tenants of social housing? international journal of housing policy, 10(1), 41-61. koval, j. j., & pederson, l. l. (1999). stresscoping and other psychosocial risk factors: a model for smoking in grade 6 students. addictive behaviors, 24(2), 207-218. landau, r. (1995). locus of control and socioeconomic status: does internal locus of control reflect real resources and opportunities or personal coping abilities? social science & medicine, 41(11), 14991505. leary, m. r., tambor, e. s., terdal, s. k., & downs, d. l. (1995). self-esteem as an interpersonal monitor: the sociometer hypothesis. journal of personality and social psychology, 68(3), 518. lee, y. g., kelley, h. h., & lee, j. m. (2023). untying financial stress and financial anxiety: implications for research and financial practitioners. journal of financial therapy, 14(1). liu, q. q., yang, x. j., zhu, x. w., & zhang, d. j. (2021). attachment anxiety, loneliness, rumination and mobile phone dependence: a cross-sectional analysis of a moderated mediation model. current psychology, 40, 5134-5144. liu, s., zhang, f., & wu, f. (2022). contrasting migrants' sense of belonging to the city in selected peri-urban neighbourhoods in beijing. cities, 120, 103499. loibl, c., & hira, t. k. (2005). self-directed financial learning and financial satisfaction. journal of financial counseling and planning, 16(1), 11-22. manago, a. m., brown, g., lawley, k. a., & anderson, g. (2019). adolescents' daily faceto-face and computer-mediated communication: associations with autonomy financial services review, 32(4) 100 and closeness to parents and friends. developmental psychology, 56(1), 153–164. meyer, k. a. (2006). when topics are controversial: is it better to discuss them faceto-face or online? innovative higher education, 31, 175-186. morrell, h. e., & cohen, l. m. (2006). cigarette smoking, anxiety, and depression. journal of psychopathology and behavioral assessment, 28, 281-295. o'sullivan, n. j., nason, g., manecksha, r. p., & o'kelly, f. (2022). the unintentional spread of misinformation on “tiktok”: a paediatric urological perspective. journal of pediatric urology, 18(3), 371-375. owens, t. j. (1993). accentuate the positive-and the negative: rethinking the use of selfesteem, self-deprecation, and self-confidence. social psychology quarterly, 288-299. ozkan, m., & solmaz, b. (2015). mobile addiction of generation z and its effects on their social lifes: (an application among university students in the 18-23 age group). procediasocial and behavioral sciences, 205, 92-98. parrott, a. c. (1999). does cigarette smoking cause stress? american psychologist, 54(10), 817. patton, g. c., hibbert, m., rosier, m. j., carlin, j. b., caust, j., & bowes, g. (1996). is smoking associated with depression and anxiety in teenagers? american journal of public health, 86(2), 225-230. perry, v. g., & morris, m. d. (2005). who is in control? the role of self‐perception, knowledge, and income in explaining consumer financial behavior. journal of consumer affairs, 39(2), 299-313. phares, e. j. (1962). perceptual threshold decrements as a function of skill and chance expectancies. the journal of psychology, 53(2), 399-407. pichler, s., kohli, c., & granitz, n. (2021). ditto for gen z: a framework for leveraging the uniqueness of the new generation. business horizons, 64(5), 599-610. pierce, t. (2009). social anxiety and technology: face-to-face communication versus technological communication among teens. computers in human behavior, 25(6), 13671372. pilisuk, m., montgomery, m. b., parks, s. h., & acredolo, c. (1993). locus of control, life stress, and social networks: gender differences in the health status of the elderly. sex roles, 28(3-4), 147-166. pohorecky, l. a. (1981). the interaction of alcohol and stress: a review. neuroscience & biobehavioral reviews, 5(2), 209-229. porath, s. (2011). text messaging and teenagers: a review of the literature. journal of the research center for educational technology, 7(2), 86-99. proctor, c., & linley, p. a. (2014). life satisfaction in youth. in g. a. fava & c. ruini (eds.), increasing psychological well-being in clinical and educational settings: interventions and cultural contexts (pp. 199215). springer. doi 10.1007/978-94-0178669-0_13 putnam, r. d. (2000). bowling alone: the collapse and revival of american community. simon & schuster. robinson, s., & stubberud, h. a. (2012). communication preferences among university students. academy of educational leadership journal, 16(2), 105. rohe, w. m., & stewart, l. s. (1996). homeownership and neighborhood stability. housing policy debate, 7(1), 37-81. rosenberg, m. (1965). society and the adolescent self-image. princeton university press. rousselle, c. (2022, december 8). gen z won’t talk on phone – now, “the phone lady” is on a mission. fox news. heo et al. 101 https://www.foxnews.com/lifestyle/gen-zwont-talk-phone-now-phone-lady-mission royston, p., & cox, n. j. (2005). a multivariable scatterplot smoother. stata journal, 5, 405– 412. sarkova, m., bacikova‐sleskova, m., orosova, o., madarasova geckova, a., katreniakova, z., klein, d., van den heuvel, w., & van dijk, j. p. (2013). associations between assertiveness, psychological well‐being, and self‐esteem in adolescents. journal of applied social psychology, 43(1), 147-154. schneider, s. j., kerwin, j., frechtling, j., & vivari, b. a. (2002). characteristics of the discussion in online and face-to-face focus groups. social science computer review, 20(1), 31-42. schuck, a. m., & widom, c. s. (2001). childhood victimization and alcohol symptoms in females: causal inferences and hypothesized mediators. child abuse & neglect, 25(8), 1069-1092. silveira, p., morais, r., & petrella, s. (2022). a communication study of young adults and online dependency during the covid-19 pandemic. societies, 12(4), 109. srivastava, v. k., & gilles, d. e. a. (1987). seemingly unrelated regression equations models: estimation and inferences. marcel dekker. tam, j., porter, e. k., & lee, u. j. (2022). examination of information and misinformation about urinary tract infections on tiktok and youtube. urology, 168, 35-40. thayer, s. e., & ray, s. (2006). online communication preferences across age, gender, and duration of internet use. cyberpsychology & behavior, 9(4), 432-440. u.s. securities and exchange commission (2018). observations from investment adviser examinations relating to electronic messaging. the office of compliance inspections and examinations https://www.sec.gov/files/ocie%20risk%2 0alert%20-%20electronic%20messaging.pd f u.s. securities and exchange commission (2023). sec announces enforcement results for fiscal year 2023. https://www.sec.gov/newsroom/pressreleases/2023-234 wang, y., & zhang, y. (2015). relation of mobile phone addiction to perceived social support and subjective well-being in college students. chinese mental health journal, 868-873. weesie, j. (2000). seemingly unrelated estimation and the cluster-adjusted sandwich estimator. stata technical bulletin, 52, 34–47. reprinted in stata technical bulletin reprints, 9, 231– 248. stata press. westerman, c. y. k., & westerman, d. (2010). supervisor impression management: message content and channel effects on impressions. communication studies, 61(5), 585-601. wilsnack, r. w., & wilsnack, s. c. (1997). gender and alcohol: individual and social perspectives. rutgers center of alcohol. williams, j. m., & stout, j. k. (1985). the effect of high and low assertiveness on locus of control and health problems. the journal of psychology, 119(2), 169-173. wilson, j. m., lee, j., fitzgerald, h. n., oosterhoff, b., sevi, b., & shook, n. j. (2020). job insecurity and financial concern during the covid-19 pandemic are associated with worse mental health. journal of occupational and environmental medicine, 62(9), 686-691. worthington, a. c. (2006). debt as a source of financial stress in australian households. international journal of consumer studies, 30(1), 2-15. yuan, s., hussain, s. a., hales, k. d., & cotten, s. r. (2016). what do they like? communication preferences and patterns of older adults in the united states: the role of technology. educational gerontology, 42(3), 163-174. zhang, n., hazarika, b., chen, k., & shi, y. (2023). a cross-national study on the excessive use of https://www.foxnews.com/lifestyle/gen-z-wont-talk-phone-now-phone-lady-mission https://www.foxnews.com/lifestyle/gen-z-wont-talk-phone-now-phone-lady-mission https://www.sec.gov/files/ocie%20risk%20alert%20-%20electronic%20messaging.pdf https://www.sec.gov/files/ocie%20risk%20alert%20-%20electronic%20messaging.pdf https://www.sec.gov/files/ocie%20risk%20alert%20-%20electronic%20messaging.pdf https://www.sec.gov/newsroom/press-releases/2023-234 https://www.sec.gov/newsroom/press-releases/2023-234 financial services review, 32(4) 102 short-video applications among college students. computers in human behavior, 107752. zhang, x., huang, x., xiao, y., jing, d., huang, y., chen, l., luo, d., chen, x. & shen, m. (2019). daily intake of soft drinks is associated with symptoms of anxiety and depression in chinese adolescents. public health nutrition, 22(14), 2553-2560. zvolensky, m. j., feldner, m. t., eifert, g. h., & brown, r. a. (2001). affective style among smokers: understanding anxiety sensitivity, emotional reactivity, and distress tolerance using biological challenge. addictive behaviors, 26(6), 901-915. financial services review, 33(2) 144 examining the gender gap in participation in employersponsored retirement plans: oaxaca decomposition ferdous ahmmed,1 charlene marie kalenkoski,2 and christopher m. browning3 abstract using the 2021 national financial capability study (nfcs), this study examines the association between gender-based participation in employer-sponsored retirement plans and financial literacy. it also decomposes the association between gender-based participation in employer-sponsored retirement plans into its explained and unexplained portions using the oaxaca decomposition. the explained portion measures how much of the gender gap in employer-sponsored, retirement-plan participation is due to the differences in the level of financial literacy. the unexplained portion measures how much of the gender gap in employer-sponsored, retirement-plan participation is due to the difference in the return to financial literacy between men and women. the results show that the explained portion of the gap due to financial literacy is -0.02, and the unexplained portion of the gap due to return to financial literacy is -0.03. the negative explained and unexplained gap due to financial literacy suggests that women have a lower average value of financial literacy and a lower return to financial literacy than men. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation ahmmed, f., kalenkoski, c. m., & browning, c. m. (2025). examining the gender gap in participation in employer-sponsored retirement plans: oaxaca decomposition. financial services review, 33(2), 144-164. introduction according to the u.s. bureau of labor statistics, 68 percent of workers in the private sector had access to retirement benefits through their employer in march 2021, and only 51 percent chose to participate. among the state and local government workers, ninety-two percent had 1 corresponding author (ferdousahmmed@missouristate.edu). missouri state university, springfield, mo, usa. 2 james madison university, harrisonburg, va, usa. 3 texas tech university, lubbock, tx, usa. acknowledgments: the authors thank russell james, thomas korankye, and michael guillemette for their valuable comments and suggestions. access to retirement benefits, but only 82 percent participated (u.s. bureau of labor statistics, 2021). the employee retirement income security act (erisa) ensures that all employees have equal access to employer-sponsored retirement plans regardless of gender, race, or ethnicity, as long as https://creativecommons.org/licenses/by-nc/4.0/ mailto:ferdousahmmed@missouristate.edu https://creativecommons.org/licenses/by-nc/4.0/ ahmmed et al. 145 they meet the eligibility requirements. under erisa, if an employer offers a retirement plan, they must follow specific rules to ensure that the plan is offered to all eligible employees in a nondiscriminatory manner. employees are eligible to participate in an employer-sponsored retirement plan once they meet certain age and service requirements. for example, an employer may require that the employees be at least 21 years old and have worked for the company for at least one year before becoming eligible to participate in the retirement plan. employers must also follow nondiscrimination rules to ensure that the retirement plan is not unfairly weighted in favor of highly compensated employees. these rules aim to prevent employers from providing greater retirement benefits to executives or other highly-paid employees while excluding lower-paid workers. this ensures that everyone has equal access to employer-sponsored retirement plans and is subject to the same eligibility rules. if the employees meet those eligibility requirements, then there should not be a barrier to plan participation. however, participation is voluntary, and the statistics show a gender gap in employer-sponsored retirementplan participation. according to the u.s. bureau of labor statistics (bls 2020), women’s participation in employersponsored retirement plans such as 401(k)s or pensions is significantly less than men's participation. in 2019, only 63% of women participated in an employer-sponsored retirement plan compared to 67% of men (bls 2020). the gap is even wider for women of color. compared to 65% of white women, only 54% of black women participated in employer-sponsored retirement plans in 2019 (bls 2020). a possible reason for this gender gap is that men and women face different social and economic conditions, both within and outside the household, which may lead to differing participation in retirement-savings plans (clark & strauss, 2012). one of these is labor force participation. according to 2018 world bank data, gender gaps in education, discrimination, and social norms shape women’s labor force participation decisions. they are less likely to join the labor force and work for pay than men and thus less likely able to access employersponsored plans. they are more likely to work part-time, in the informal sector, or lower-income occupations. the gender gap in earnings directly impacts women's ability to save for retirement as lower earnings result in reduced income available to save. beyond labor market factors, financial knowledge is one of the most important factors that influence how individuals make their financial decisions, especially when it comes to saving for retirement. prior studies find that individuals with a higher level of financial knowledge are more likely to plan for retirement (lusardi & mitchell, 2011). however, research also shows that women tend to have lower levels of financial literacy than men, which may contribute to the gender gap in retirement preparedness (anderson & collins, 2017). therefore, this study places a special emphasis on financial literacy as a key factor when examining gender gaps in retirement plan participation. other reasons women are less prepared for retirement that are suggested by prior research point to their unpaid caregiving, risk tolerance, investment choices, saving behavior, earnings, information sources, and education (bajtelsmit & bernasek, 1996; feng et al., 2019; weller & tolson, 2017; zhao & zhao, 2018). beyond these observable factors, there may also be an unexplained gap in retirement savings decisions that may be due to unmeasured characteristics that are correlated with measured ones. to investigate this, the current study uses the oaxaca decomposition method, a widely accepted approach for identifying separately the explained and unexplained sources of differences in outcomes. this technique has previously been used in many research studies to investigate the gender pay gap and discrimination (anspal, 2015). this study uses this method to examine how measured characteristics such as financial literacy are related to plan participation, that is, how differences in the level of financial literacy between men and women (the explained portion) lead to differences in participation, all else equal, and how differences in the return to financial literacy (the unexplained portion—due to differences in regression coefficients) lead to differences in participation. other demographic financial services review, 33(2) 146 characteristics that are examined include age, level of education, race, home ownership, income, and risk tolerance. thus, the following research questions are addressed: (1) what factors explain the gender gap in retirement plan participation among single, employed individuals? (2) how do financial literacy and other variables contribute to the explained and unexplained portions of this gap? the analysis sample is made up of single, employed individuals, a group that hasn’t been studied much previously. however, it is important to study this group because they alone are responsible for their own retirement savings. studying the behavior of single individuals to learn more about what influences their retirement plan participation will help financial planners understand how to better support single individuals with their retirement planning. the results show that single women’s participation in employer-sponsored retirement plans is slightly higher (1.36 percentage points) than male worker’s participation a surprising result as it is different from what is found with an all-female-worker sample (bls 2020). objective financial knowledge, age, income, education, homeownership, and risk tolerance are all associated positively with employer-sponsored, retirement-plan participation. literature review previous research finds that women are less prepared for retirement than men and finds that financial knowledge, unpaid caregiving, risk tolerance, investment choices, saving behavior, labor force participation, and information sources are potential reasons for women's lack of retirement preparedness (bajtelsmit & bernasek, 1996; bajtelsmit & van derhei, 1997; dietz et al., 2003; huang & curtin, 2019). however, this research has not focused on single individuals who, unlike married individuals, bear sole responsibility for their financial future, and may behave differently from the population as a whole. financial knowledge the terms financial literacy and financial knowledge have often been used interchangeably in the literature (huston, 2010). some researchers examine the role of financial knowledge in retirement-savings decisions. anderson and collins (2017) find that men possess greater financial knowledge than women, which contributes to the observed gender gaps in retirement savings. this increased financial knowledge allows men to better understand the features and benefits of employer-sponsored retirement plans, leading to greater confidence and participation in these plans. similarly, bucher-koenen et al. (2017) find that women are less likely than men to provide correct answers to questions about basic financial concepts and are more likely to say that they do not know the answer. this gender gap in financial literacy is concerning, especially since women tend to live longer and are more likely to experience widowhood in retirement (hsu, 2016). recent studies such as those by harahap et al. (2022), preston and wright (2023), and tomar et al. (2021), confirm that financial literacy is a strong predictor of retirement-planning behavior and highlight the importance of financial education. unpaid caregiving analyzing data from the survey of consumer finances (scf) from 1989 through 2016, weller and tolson (2018) investigate the relationship between unpaid caregiving and labor earnings stability and the link to retirement savings. they find that unpaid caregiving can adversely affect a caregiver’s hours at work, earnings, employment, and income stability, negatively impacting the caregivers’ savings. women are more likely to experience the effects of caregiving than men (lee & tang, 2015; reinhard et al., 2012). these disparities are particularly significant for single women who lack the support of a partner. while caregiving is not directly analyzed in this study due to data limitations, it remains a relevant context for understanding the gender gap. risk tolerance some researchers have examined the relationships between risk tolerance and retirement savings and wealth accumulation. ahmmed et al. 147 previous studies find that, compared to women, men exhibit a higher level of risk tolerance (bollen & posavac, 2018; gibson et al., 2013). spivey (2010) finds a similar result using data from the national longitudinal survey of youth 1979. this increased risk tolerance of men could translate into a more proactive approach to retirement planning, including participation in employer-sponsored retirement plans. dwyer et al. (2002) and kappal and rastogi (2020) confirm that risk aversion leads women to invest less in risky assets. however, dwyer et al. (2002) and kappal and rastogi (2020) also note that improved financial knowledge can help reduce this gap, highlighting that targeted financial education may be an effective solution. savings behavior differences in saving behavior between men and women can contribute to the gender gap in employer-sponsored, retirement-plan participation. various societal, cultural, and individual factors often influence these differences. fisher et al. (2015) investigate the differences in savings behaviors between genders with data from a nationally representative sample of lowto moderate-income households (nc1172) as well as data from the 2010 survey of consumer finances (scf). nc1172 is a multistate research program sponsored by north central (nc). results show that men and women exhibit different savings behaviors. in the scf, having other members in the household affects savings behavior differently for men and women. additionally, they find that education and counseling positively impact savings behavior among both men and women in low-to-moderateincome households. labor force participation differences in labor force participation between men and women can contribute to the gender gap in employer-sponsored, retirement plan participation. various employment and workforce factors shape this gap. men and women often work in different industries or roles, and some sectors are more likely to offer retirement benefits than others. as a result, access to employer-sponsored plans can vary significantly by gender. cordova et al. (2022) and sierminska et al. (2010) find that women’s lower labor force participation leads to lower retirement wealth accumulation. women more frequently work part-time, possess more diversified work histories influenced by childbearing, and experience more frequent job changes (berger & denton, 2004; niessen-ruenzi & schneider, 2019). this study's focus on employed individuals ensures that differences in employersponsored, retirement-plan participation are not influenced by variations in employment status, thereby enhancing the reliability of the findings. information sources access to financial information also differs by gender. studies have found that women report lower levels of confidence and familiarity with financial concepts than men (chen & volpe, 2002; loibl & hira, 2006). graham et al. (2002) also suggest that information-processing styles may lead to differences in financial strategies. while this factor is not directly included in the model, its indirect effects may be captured through variables such as education or financial literacy. previous studies have looked at the effects of differences in the level (explained) of the explanatory variables but not at the effects of differences in the returns of the explanatory variables (unexplained). understanding the factors contributing to the unexplained gap is essential for addressing the gender gap in employer-sponsored retirement plan participation comprehensively. while the explained gap can be attributed to differences in financial literacy, income, employment, or educational attainment, the unexplained gap focuses on other factors contributing to the gender gap in employersponsored retirement plan participation. the unexplained gap could come from factors that are unobserved or not included in the model, such as discrimination and bias, social and cultural factors, work-life balance, and caregiving responsibilities. according to the u.s. erisa act, there should not be any discrimination between men and women in attaining employersponsored retirement plans. however, gaps in employer-sponsored retirement-plan participation can still exist due to other factors, such as social and cultural factors, work-life balance, and caregiving responsibilities. societal financial services review, 33(2) 148 norms and cultural expectations can influence women’s financial behaviors and saving patterns, potentially contributing to the unexplained gap. women’s increased caregiving responsibilities and work-life balance challenges can affect their employment patterns, leading to interruptions, part-time work, or career breaks, affecting their retirement savings. using the nfcs (2021), the current study examines how differences in the means of the explanatory variables (financial literacy and other demographic and economic variables) and the returns to these explanatory variables between single employed men and single employed women are associated with employer-sponsored, retirement-plan participation. theoretical framework this study’s theoretical framework draws from the life cycle hypothesis (ando & modigliani, 1963) and human capital theory (becker, 1994). the life cycle hypothesis suggests that individuals plan their consumption and savings across different stages of life to maintain a consistent living standard. human capital theory suggests that investments in education and training enhance a person’s skills and knowledge, which in turn increases their economic productivity and value. financial literacy, a form of human capital, has been associated with increased retirement preparedness and a more proactive approach to long-term financial planning (mitchell & lusardi, 2022). building on this theoretical foundation, this study hypothesizes that individuals with higher levels of human capital, measured by objective financial literacy, are more likely to participate in employer-sponsored retirement plans, all else equal. however, the returns to financial literacy may differ between men and women due to discrimination or to unmeasured factors correlated with measured ones. the use of oaxaca decomposition allows separation of these differences into two parts: one based on levels of financial literacy (explained), and the other based on the return to financial literacy (unexplained). the association between age and participating in an employer-sponsored, retirement-savings account is expected to be positive. young adults usually have more liquidity constraints and a lower likelihood of saving for retirement than older adults. the respondent’s level of education is expected to be related positively to participating in an employer-sponsored, retirement-savings account. highly educated individuals can make better financial decisions than less educated individuals. as the individual’s level of education increases, the likelihood of participating in an employersponsored, retirement-savings account may increase. white is a proxy for preferences and constraints that cannot be given a sign a priori. owning a home often indicates greater financial stability and long-term planning, which can encourage individuals to be more proactive about retirement planning. therefore, homeownership is expected to have a positive relationship with participation in employer-sponsored retirement-savings accounts. higher-income increases the financial resources available to the respondents to save for retirement. therefore, a household’s annual income is expected to be related positively to participating in an employer-sponsored, retirement-savings account. ownership of financial assets is influenced by risk tolerance, as financial assets are often risky (nguyen, 2015). therefore, financial risk tolerance is expected to have a positive relationship with participating in an employer-sponsored retirement savings account. data this paper uses data from the 2021 national financial capability study (nfcs). the nfcs is a project of the finra investor education foundation. the online state-by-state survey was administered from june through october 2021 to a sample of 27,118 american adults. the survey includes approximately 500 respondents per state, including the district of columbia. weights are provided to make estimates from the data nationally representative. in 2009, the finra investor education foundation commissioned the first nationwide study to assess the financial capability of american adults. the primary goals of the nfcs ahmmed et al. 149 study are to establish benchmark indicators of financial capability and examine the variations of these indicators with underlying demographic, behavioral, attitudinal, and financial literacy characteristics. the analysis in this paper focuses on single, employed individuals and excludes observations from the sample with the responses "don't know" and "prefer not to say" to the financial literacy, risk tolerance, homeownership, and retirementsavings questions. observations with missing or non-informative responses, such as “don’t know” or “prefer not to say” were excluded from the analysis. these responses do not contribute to measuring financial literacy scores or estimating risk tolerance and other categorical variables. this approach is consistent with previous studies using nfcs data (olajide et al., 2024; pandey & guillemette, 2024). missing data were not imputed, as most variables are categorical and imputing them would introduce additional, potentially untenable, assumptions. this study also excludes self-employed individuals as they are less likely to have a retirement plan through their employer. to abstract from the hours of work decision, a sensitivity analysis is performed on the subsample of full-time employed individuals. the analysis sample size is 4,136. table 1 presents comparisons of the means of demographic variables across the full and analysis samples to show the representativeness of the analysis sample. there are some statistically different means between the full and analysis samples. these are for the female and white variable. seventy one percent of the full sample consists of white individuals, whereas 61% of the analysis sample is white. fifty one percent of full sample is female, whereas 49% of the analysis sample is female. because this paper examines how differences between men and women in the explanatory variables (financial literacy, risk tolerance, and other demographic and economic variables) and differences in the returns to each explanatory variable are associated with employer-sponsored, retirement-plan participation, the dependent variable in the analysis is whether the respondents have any retirement accounts through their current or previous employer. the exact nfcs question that asks this is, “do you or your spouse or your partner have any retirement plans through a current or previous employer, like a pension plan, a thrift savings plan (tsp), or a 401(k)?” the value for the dependent variable is 1 if the respondents answer "yes" and 0 if the respondents answer "no." the survey data unfortunately does not include information on whether retirement plans were offered in the first place. this omission could lead to omitted variable bias if the availability of retirement plans is not randomly distributed across demographic groups. for example, if women are offered retirement plans less often than men due to employment in sectors or jobs with lower benefits, then the observed gap in participation may not fully reflect the actual difference in access. to address this, we performed a sensitivity analysis using a sample of full-time employees, who are generally more likely to have retirement plan offers from the employers. this helps somewhat mitigate the issue by focusing on a group where plan availability is more consistent. the key explanatory variable is objective financial knowledge measured by responses to six questions assessing the respondent's understanding of inflation, compound interest, bond price, mortgage interest, risk, and return. respondents received 1 for each correct answer and 0 for an incorrect answer. thus, this variable is just the sum of correct answers and ranges from 0 to 6. this six-question index has been widely used in financial capability research and reflects core concepts essential to effective retirement planning (lusardi & mitchell, 2011). however, it has limitations, as it measures only objective knowledge and does not account for other aspects of financial capability, such as behavioral application, confidence, or self-efficacy. therefore, while valuable, this index may not fully capture the broader financial decisionmaking abilities that influence retirement plan participation. other explanatory variables are age, level of education, white race, home ownership, income, and risk tolerance. five dummy variables represent age: 25-34, 35-44, 45-54, 55-64, and 65+. the reference category is the 18-24 age financial services review, 33(2) 150 group. the respondents' education level is represented by four dummy variables for some college, associate degree, bachelor's degree, and postgraduate degree. the reference category is high school or less. white is a dummy variable that equals 1 if the respondents' race is white and 0 if the respondents' race is non-white. homeownership is also a dummy variable that equals 1 if respondents own a house and 0 otherwise. income is represented by six dummy variables for $50,000 to $75,000, $75,000 to $100,000, $100,000 to $150,000, $150,000 to $200,000, $200,000 to $300,000, and $300,000 or more. the reference category is less than $50,000. the responses to the risk-tolerance question range from 1 (not at all willing) to 10 (very willing). due to the limited number of observations in certain response categories, the responses are recoded into three categories. the first category is low risk tolerance which includes responses ranging from 1 to 3. the second category is medium risk tolerance, which includes 4 to 7. the third category is high risk tolerance, which includes responses ranging from 8 to 10. medium-risk tolerance and high-risk tolerance are included in the regression, with the reference category being low-risk tolerance. table 2 presents descriptive statistics for all variables for the analysis sample and separately for males and females. in the analysis sample, approximately 56% of individuals have an employer-sponsored retirement plan. the percentage of males in the overall sample is 51%. in the male sample, the percentage of individuals with employer-sponsored retirement plans is 55.59%. the percentage of females in the overall sample is 49%. in the female sample, the percentage of individuals with employersponsored retirement plans is 56.95%. the average financial literacy score for the analysis sample is 2.75. this means that, on average, respondents answered 2.75 of the six financial literacy questions correctly. the average financial literacy score is 3.01 among males, but the average financial literacy score is 2.42 among females. males have scored better than females in financial literacy measures. overall, 32% of individuals have a bachelor’s degree or higher. thirty percent of males have a bachelor’s degree or higher, and 35% of females have a bachelor’s degree or higher. thirty-two percent of males said they are high risk tolerant, but only 19% of females said they are high risk tolerant. for the annual income level, 56% of males and 62% of females said they have an annual income of less than $50,000. table 2 also provides the descriptive statistics for the other demographic variables. table 1. mean comparison of variables between the full sample and the analysis sample category full sample mean (std. dev.) analysis sample mean (std. dev.) t pr(|t| > |t|) female 0.5107 (0.4965) 0.4915 (0.4998) 2.3221 0.0101 white 0.7117 (0.4376) 0.6140 (0.4869) 13.1723 0.0000 ahmmed et al. 151 age (25-64) 0.6899 (0.0030) 0.7063 (0.4554) -5.9568 1.0000 income($50k-$75k) 0.1940 (0.3955) 0.2091 (0.4067) -2.2696 0.9884 table 2. summary statistics full analysis sample male female mean std. err. mean std. err. mean std. err. male 0.5146 0.0081 female 0.4854 0.0081 participation in employer-sponsored retirement plan 0.5618 0.0082 0.5559 0.0116 0.5695 0.0115 objective financial knowledge 2.7525 0.0267 3.0125 0.0383 2.4160 0.0349 white 0.5765 0.0084 0.5764 0.0118 0.5767 0.0117 homeownership 0.3695 0.0079 0.3924 0.0112 0.3399 0.0109 risk tolerance level low 0.2304 0.0069 0.1729 0.0089 0.3048 0.0106 medium 0.5061 0.0083 0.5052 0.0117 0.5072 0.0116 high 0.2635 0.0076 0.3219 0.0111 0.1880 0.0092 age 18-24 0.2641 0.0076 0.2491 0.0108 0.2835 0.0106 25-34 0.3559 0.0080 0.3617 0.0113 0.3484 0.0110 35-44 0.1708 0.0060 0.1686 0.0083 0.1738 0.0086 45-54 0.1233 0.0052 0.1345 0.0074 0.1088 0.0070 55-64 0.0704 0.0040 0.0722 0.0055 0.0681 0.0059 65 + 0.0155 0.0019 0.0139 0.0025 0.0175 0.0030 annual income level less than $50,000 0.5870 0.0081 0.5624 0.0115 0.6188 0.0111 $50,000 to $75,000 0.2033 0.0065 0.2159 0.0092 0.1869 0.0090 financial services review, 33(2) 152 $75,000 to $100,000 0.1090 0.0050 0.1161 0.0072 0.0998 0.0067 $100,000 to $150,000 0.0676 0.0040 0.0679 0.0056 0.0673 0.0057 $150,000 to $200,000 0.0197 0.0021 0.0221 0.0030 0.0165 0.0029 $200,000 to $300,000 0.0093 0.0017 0.0108 0.0026 0.0072 0.0018 more than $300,000 0.0042 0.0010 0.0048 0.0014 0.0034 0.0012 education level high school education or less 0.3105 0.0080 0.3420 0.0115 0.2696 0.0106 some college 0.2607 0.0073 0.2545 0.0102 0.2688 0.0105 associate degree 0.1111 0.0053 0.1073 0.0073 0.1160 0.0076 bachelor’s degree 0.2400 0.0066 0.2420 0.0093 0.2373 0.0093 postgraduate degree 0.0777 0.0039 0.0541 0.0047 0.1083 0.0067 number of observations 4136 2128 2008 notes: this analysis uses data from the finra foundation 2021 nfcs state by state dataset. mean values are shown alongside the standard errors. survey weights are applied. *** indicates significance at the 1% level; ** indicates significance at the 5% level; *indicates significance at the 10% level. model a linear probability model is estimated separately for men and women in order to decompose the gap in participation into its explained and unexplained portions. model 1 (male): ersi = β0m + β1m finliti + γm xi + vmi model 2 (female): ersi = β0f + β1f finliti + γf xi + vfi where ersi is a binary dependent variable that takes a value of 1 if a respondent participates in an employer-sponsored, retirement-plan and 0 otherwise. 𝛽0 is the intercept. 𝛽1 is the association between financial knowledge and employersponsored, retirement-plan participation. finliti is the financial literacy score earned by respondent i. the matrix 𝑋i contains all other explanatory variables related to participation in employersponsored retirement plans. these explanatory variables include age, level of education, race, home ownership, income, and risk tolerance. γm and γf are vectors of the slope parameters for age, level of education, race, home ownership, income, and risk tolerance. vi is the error term that is assumed to follow a normal distribution. robust standard errors are used to adjust for heteroskedasticity. this study uses oaxaca decomposition to compute how much of the gender gap in employer-sponsored retirement-plan participation is due to differences in levels of the explanatory variables (financial literacy, risk tolerance, and other demographic and economic variables) and how much is due to the return to each explanatory variable (the regression coefficients). the oaxaca decomposition is a statistical technique commonly used to break down differences in outcomes, such as earnings or participation rates between groups. it separates the total difference into two parts: one that is explained by differences in observable factors (e.g., financial literacy, income, education) and the other unexplained, often attributed to discrimination or unmeasured factors that are correlated with measured ones. this method is appropriate for analyzing gender gaps in ahmmed et al. 153 employer-sponsored retirement-plan participation, as it shows not only whether men and women differ in characteristics that influence retirement plan participation, but also whether they receive different returns to those characteristics. results the coefficients and robust standard errors for the linear probability models are shown in table 3 for both the male and female employed samples. column “a” of table 3 shows results for the male sample. column “b” shows results for the female sample. consistent with the hypothesis, table 3 results show that the relationship between objective financial literacy and participation in employersponsored retirement-savings plans is positive in both the male and female samples. table 3 also shows that the association between age and employer-sponsored, retirement-plan participation is positive for both male and female samples, suggesting that older individuals are more likely to participate. there is also a positive relationship between a respondent’s education level and participation in employer-sponsored retirement plans for both the male and female samples. similarly, income is positively associated with participation in employersponsored retirement plans for both groups. finally, both risk tolerance and homeownership are positively associated with participation in employer-sponsored retirement plans for the male and female samples. table 3. effects of financial literacy and other explanatory variables on employer-sponsored retirement plan participation: linear probability model (main modelemployed sample) a. male b. female coef. robust std. err. p value coef. robust std. err. p value objective financial knowledge 0.0366*** 0.0065 0.0000 0.0232*** 0.0070 0.0010 white -0.0394* 0.0207 0.0570 -0.0183 0.0212 0.3880 homeownership 0.1352*** 0.0216 0.0000 0.1279*** 0.0231 0.0000 versus (risk tolerance – low) risk tolerance – medium 0.0653** 0.0284 0.0210 0.0579** 0.0237 0.0150 risk tolerance – high 0.1080*** 0.0310 0.0000 0.0606* 0.0316 0.0550 versus (age 18-24) 25-34 0.0691** 0.0278 0.0130 0.1248*** 0.0275 0.0000 35-44 0.0749** 0.0323 0.0210 0.1354*** 0.0324 0.0000 45-54 0.0978*** 0.0340 0.0040 0.1346*** 0.0370 0.0000 55-64 0.0673 0.0413 0.1030 0.1842*** 0.0416 0.0000 65 + 0.2151*** 0.0723 0.0030 -0.0143 0.0873 0.8700 versus (income less than $50,000) financial services review, 33(2) 154 $50,000 to $75,000 0.2000*** 0.0266 0.0000 0.1691*** 0.0276 0.0000 $75,000 to $100,000 0.1995*** 0.0318 0.0000 0.1956*** 0.0326 0.0000 $100,000 to $150,000 0.2246*** 0.0374 0.0000 0.1691*** 0.0371 0.0000 $150,000 to $200,000 0.3236*** 0.0478 0.0000 0.2524*** 0.0538 0.0000 $200,000 to $300,000 0.1673* 0.0913 0.0670 0.2794*** 0.0652 0.0000 more than $300,000 0.2987*** 0.0683 0.0000 0.0424 0.1383 0.7590 versus (education level high school or less) some college 0.0535* 0.0283 0.0590 0.0626** 0.0304 0.0400 associate degree 0.0969** 0.0387 0.0120 0.0730* 0.0398 0.0670 bachelor’s degree 0.1199*** 0.0291 0.0000 0.1853*** 0.0314 0.0000 postgraduate degree 0.0995** 0.0443 0.0250 0.2182*** 0.0362 0.0000 constant 0.1314*** 0.0353 0.0000 0.1873*** 0.0343 0.0000 number of observations 2128 2008 notes: this analysis uses data from the finra foundation 2021 nfcs state by state dataset. coefficient values are shown alongside the robust standard errors. survey weights are applied. *** indicates significance at the 1% level; ** indicates significance at the 5% level; *indicates significance at the 10% level. oaxaca decomposition consistent with oaxaca and ransom (1994), the oaxaca decomposition technique is used with the previous regression results to examine the gender gap in employer-sponsored, retirement -savings accounts. in particular this technique examines how differences in level of financial literacy and other explanatory variables can explain a portion of the gap and how differences in the return to financial literacy and other explanatory variables form the unexplained portion of the gap. the gender gap in retirement-savings participation is y̅m y̅f, where y̅m is the average retirement savings participation of males and y̅f is the average retirement savings participation of females. employer-sponsored retirement savings participation depends on financial literacy and other demographic and economic variables such as age, level of education, race, home ownership, income, and risk tolerance. the corresponding regression retirement savings participation equations for men and women are the following. yim = βmxim + μim (1) yif = βfxif + μif (2) where yim is the retirement savings participation of man i, βm is the vector of effects of xim matrix of independent variables on yim, μim is the error term. yif is the retirement savings participation for woman i, βf is the vector of effects of xif matrix of independent variables on yif, μif is the error term. β̂m is the vector of effects of matrix x on y̅. β̂f is the corresponding vector of effects for women. the average values of xm and xf are x̅m and x̅f. according to the arithmetic relationship, we can write y̅ = β̂x̅ and it holds for both men and women. substituting into the expression for the gender gap in retirement savings participation, we have: ahmmed et al. 155 y̅f − y̅m = β̂f x̅f β̂m x̅m (3) this suggests that average retirement savings participation for men and women could differ either because x̅ differs or because β̂ differs in the current study, either because the average level of financial literacy and/or other explanatory variables differ by gender or because the return to financial literacy and other explanatory variables differ. adding and subtracting β̂mx̅f to the right side of the equation and rearranging and combining terms yields the famous oaxaca decomposition of differences in means. y̅f y̅m = ∑ [β̂jm ×k j=1 (x̅jf − x̅jm)] + ∑ [(β̂jf − β̂jm) × x̅jf]k j=1 (4) for each j, % explained = [β̂jm ×(x̅jf − x̅jm)] y̅f − y̅m × 100 (5) % unexplained = [(β̂jf − β̂jm) × x̅jf] y̅f − y̅m × 100 (6) in the equations 4, 5, and 6, j represents each independent variable. the first term in the brackets on the right side of equation (4) is the difference of average financial literacy and other variables between females and males, multiplied by β̂m, the value of a unit of x for males. it represents the gender gap in employer-sponsored retirement plan participation that can be attributed to the differences in financial literacy and other explanatory variables. this portion of the gender gap in employer-sponsored retirement plan participation is the explained portion of the retirement plan participation gap (differences in mean). the second term in brackets on the right side of equation (4) is the difference in the return of financial literacy and other explanatory variables for females and males. this portion of the gender gap in employer-sponsored retirement plan participation is the unexplained portion of the retirement plan participation gap (difference in slope). to get the portion of the gender gap for each explanatory variable, we divide equation (4) by the total gap in employer-sponsored plan participation (y̅f y̅m) to put the two terms in percentage. the explained portion for each explanatory variable is [β̂jm ×(x̅jf − x̅jm)] y̅f − y̅m × 100 and the unexplained portion for each explanatory variable is = [(β̂jf − β̂jm) × x̅jf] y̅f − y̅m × 100. adding explained and unexplained percentage for each explanatory variable add up to 100 percent. oaxaca decomposition results this study examines the oaxaca decomposition of the gender gap in employer-sponsored, retirement-plan participation for single, employed individuals. table 4 shows the decomposition for single, employed individuals. column “a” shows the explained portion for each explanatory variable and the total explained gap. column “b” shows the unexplained portion for each explanatory variable and the total unexplained gap. table 4 also shows the total gender gap (explained and unexplained) in employer-sponsored, retirement-plan participation for single, employed individuals. table 4 shows that females and males have a 0.0136 (1.36 percentage point) gap in participation in employer-sponsored retirement plans. it means that females’ participation in employer-sponsored retirement plans is 1.36 percentage points higher than males'. of the original 0.0136 gaps, -0.0428 is the result of the difference in financial literacy and other explanatory variables between females and males (explained gap). and 0.0565 is the result of the differences in the return to financial literacy and other explanatory variables between females and males (unexplained gap). the negative explained gap means females have a lower average value of explanatory variables than males. the positive unexplained gap means females have a higher return to explanatory variables than males. table 4 also shows which variables are most responsible for the difference in employersponsored retirement plan participation. for the explained gap, these are objective financial literacy (-142.84%), homeownership (-52.44%), and high risk tolerance (-83.02%). for the unexplained gap, variables that contributed significantly are age 25-34 (189.78%), age 35-44 (122.83%), and bachelor’s degree (130.63%). the positive unexplained gap for age and education suggests that women in financial services review, 33(2) 156 these groups benefit more than men with similar characteristics. women with higher education may make better use of employer-sponsored retirement plans, and younger women may be more engaged in planning for their financial future. the portion of the gender gap in participation in employer-sponsored plans explained by differences in financial literacy is 142.85%, and the portion due to differences in return to financial literacy is -236.01%. the negative explained and unexplained gap due to financial literacy suggests that women have a lower average value of financial literacy and a lower return to financial literacy than men. figure 1 illustrates the explained and unexplained contributions of key variables to the gender gap. figure 1. explained vs. unexplained contribution to gender gap (key variables) table 4. results from oaxaca decomposition (main model-employed sample) a b explained % explained unexplained % unexplained objective financial knowledge -0.0195 -142.8483 -0.0321 -236.0146 white 0.0000 -0.0447 0.0005 3.4146 homeownership -0.0071 -52.4402 -0.0021 -15.3123 risk tolerance – medium 0.0001 0.7108 0.0040 29.1971 risk tolerance – high -0.0113 -83.0208 -0.0114 -83.3842 25-34 -0.0011 -8.1430 0.0258 189.7832 35-44 0.0004 3.1459 0.0167 122.8348 45-54 -0.0025 -18.3422 0.0100 73.4972 55-64 -0.0004 -2.7758 0.0096 70.6010 65 + 0.0003 2.3411 -0.0033 -24.3380 $50,000 to $75,000 -0.0053 -38.6795 -0.0050 -36.8995 -400 -300 -200 -100 0 100 200 300 400 500 c o n tr ib u ti o n t o g en d er g ap % % explained % unexplained ahmmed et al. 157 $75,000 to $100,000 -0.0032 -23.3687 -0.0007 -5.1477 $100,000 to $150,000 -0.0001 -0.8254 -0.0047 -34.7075 $150,000 to $200,000 -0.0016 -11.6480 -0.0016 -11.9160 $200,000 to $300,000 -0.0005 -3.7194 0.0017 12.4403 more than $300,000 -0.0003 -2.3359 -0.0010 -7.1113 some college 0.0008 5.6595 0.0031 22.7930 associate degree 0.0008 5.5398 0.0003 2.1766 bachelor’s degree -0.0007 -4.8371 0.0178 130.6298 postgraduate degree 0.0083 61.0642 0.0084 61.9314 constant 0.0204 150.0984 total explained & unexplained -0.0428 -314.5676 0.0565 414.5662 total gap (explained + unexplained) -0.0428 + 0.0565 = 0.0136 % explained (-0.0428/0.0136)*100 = -314.5676 % unexplained (0.0565/0.0136)*100 = 414.5662 -314.5676% + 414.5662 = 100% sensitivity analysis this study also conducts a sensitivity analysis with single, full-time workers to abstract from the decision regarding the number of hours worked. table 5 shows the results for this subsample. column “a” of table 5 shows results for the male subsample. column “b” shows results for the female subsample. the results of the sensitivity models are consistent with the main models. table 5 shows a positive relationship between objective financial literacy scores and participation in employer-sponsored retirement savings plans for both male and female subsamples. the sensitivity analysis results for the control variables are consistent with the main models for both male and female subsamples. table 6 shows the oaxaca decomposition for the full-time employed subsample. the results of the sensitivity analysis are consistent with the main models. table 6 shows that females' participation in employer-sponsored retirement plans is 2.07 percentage points higher than males. the portion of the gender gap in participation in employersponsored plans explained by differences in financial literacy is -101.81% and the portion due to differences in return to financial literacy is 121.46%. financial services review, 33(2) 158 table 5. effects of financial literacy and other explanatory variables on employer-sponsored retirement plan participation: linear probability model (sensitivity modelfull-time employed subsample). a. male b. female coef. robust std. err. p value coef. robust std. err. p value objective financial knowledge 0.0360*** 0.0073 0.0000 0.0248*** 0.0078 0.0010 white -0.0486** 0.0230 0.0350 -0.0275 0.0236 0.2440 homeownership 0.1059*** 0.0237 0.0000 0.1116*** 0.0247 0.0000 versus (risk tolerance – low) risk tolerance – medium 0.0417 0.0329 0.2060 0.0387 0.0269 0.1500 risk tolerance – high 0.0774** 0.0352 0.0280 0.0248 0.0356 0.4860 versus (age 18-24) 25-34 0.0423 0.0329 0.1990 0.1112*** 0.0328 0.0010 35-44 0.0530 0.0371 0.1530 0.1142*** 0.0373 0.0020 45-54 0.0695* 0.0390 0.0750 0.1037** 0.0420 0.0140 55-64 0.0431 0.0473 0.3620 0.1851*** 0.0458 0.0000 65 + 0.2115*** 0.0774 0.0060 -0.2853** 0.1384 0.0390 versus (income less than $50,000) $50,000 to $75,000 0.1696*** 0.0291 0.0000 0.1419*** 0.0299 0.0000 $75,000 to $100,000 0.1868*** 0.0344 0.0000 0.1610*** 0.0353 0.0000 $100,000 to $150,000 0.2103*** 0.0390 0.0000 0.1386*** 0.0385 0.0000 $150,000 to $200,000 0.2752*** 0.0491 0.0000 0.2888*** 0.0435 0.0000 $200,000 to $300,000 0.1330 0.0949 0.1610 0.2273*** 0.0689 0.0010 more than $300,000 0.3248*** 0.0409 0.0000 -0.0014 0.1403 0.9920 versus (education level high school or less) some college 0.0512 0.0331 0.1220 0.1024*** 0.0371 0.0060 associate degree 0.0657 0.0431 0.1280 0.0838* 0.0458 0.0680 ahmmed et al. 159 bachelor’s degree 0.1156*** 0.0326 0.0000 0.1949*** 0.0356 0.0000 postgraduate degree 0.0956** 0.0465 0.0400 0.2238*** 0.0397 0.0000 constant 0.2553*** 0.0420 0.0000 0.2716*** 0.0408 0.0000 number of observations 1694 1516 notes: this analysis uses data from the finra foundation 2021 nfcs state-by-state dataset. coefficient values are shown alongside the robust standard errors. survey weights are applied. *** indicates significance at the 1% level; ** indicates significance at the 5% level; *indicates significance at the 10% level. table 6. results from oaxaca decomposition (sensitivity modelfull-time employed subsample) a b explained % explained unexplained % unexplained objective financial knowledge -0.0211 -101.8102 -0.0252 -121.4590 white 0.0000 0.2254 0.0001 0.6493 homeownership -0.0050 -24.2819 -0.0001 -0.5460 risk tolerance – medium 0.0009 4.2538 0.0073 35.0394 risk tolerance – high -0.0080 -38.6024 -0.0122 -58.8469 25-34 -0.0001 -0.5706 0.0288 139.0441 35-44 0.0009 4.1254 0.0183 88.2032 45-54 -0.0015 -7.1897 0.0089 43.0729 55-64 -0.0004 -1.8590 0.0109 52.6468 65 + -0.0004 -1.9854 -0.0043 -20.8126 $50,000 to $75,000 -0.0049 -23.4613 -0.0021 -9.8992 $75,000 to $100,000 -0.0022 -10.7782 -0.0029 -14.1192 $100,000 to $150,000 0.0008 3.7146 -0.0060 -28.7940 $150,000 to $200,000 -0.0022 -10.7179 0.0004 1.8648 $200,000 to $300,000 -0.0002 -0.9428 0.0023 11.1890 more than $300,000 -0.0002 -1.0939 -0.0015 -7.3186 some college -0.0005 -2.4918 0.0130 62.9283 associate degree 0.0004 2.1143 0.0067 32.3840 bachelor’s degree 0.0007 3.5466 0.0249 120.1817 postgraduate degree 0.0091 43.7280 0.0118 56.8151 constant -0.0245 -118.1456 financial services review, 33(2) 160 total explained & unexplained -0.0340 -164.0772 0.0547 264.0777 total gap -0.0340 + 0.0547 = 0.0207 % explained (-0.0340/ 0.0207)*100 = -164.0772 % unexplained (0.0547/0.0207)*100 = 264.0777 -164.0772 +264.0777 = 100% discussion and implications this study finds that single women have a lower level of financial literacy and lower returns to financial literacy than single men. this indicates that single women are less knowledgeable about financial matters and benefit less from the financial literacy they possess in terms of participation in employer-sponsored retirement plans. while erisa guarantees equal access to employer-sponsored retirement plans, these findings suggest that equal availability does not necessarily result in similar participation rates. differences in financial literacy, along with demographic and economic factors, continue to create gaps in actual participation. to address this, financial planners can collaborate with employers to develop and implement financial literacy programs specifically targeted at single women. these programs should cover important topics such as retirement planning, investment strategies, budgeting, and risk management. the format of these programs should be tailored to suit different audiences. for example, online modules or gamified learning tools may be more effective for younger workers, whereas in-person workshops or seminars may be more suitable for older individuals. developing tailored educational programs, such as workshops, online courses, and seminars specifically designed for women, focusing on fundamental financial concepts, might help increase financial literacy among single women. additionally, financial planners can work with companies to incorporate financial literacy programs into their employee-benefits packages, encouraging participation among employees. employers can incorporate financial literacy into employee wellness programs and provide incentives for participating in educational sessions or increasing retirement contributions. the study also finds that single females have lower risk tolerance than single males. financial planners can educate single females on risk management to improve their participation in retirement plans by offering one-on-one consultations to assess individual risk tolerance and provide tailored advice on suitable investment strategies. arranging workshops that cover the basics of risk management, different types of risks, and how to balance risk and reward in investment portfolios might also be helpful. to help individuals overcome challenges like low income or risk aversion, employers and financial planners can recommend low-cost, diversified investments such as target-date or index funds. they can also break long-term goals into small, actionable steps to build confidence and encourage participation. furthermore, single women aged 25-34 and 3544 have a much higher return on their retirement plan participation compared to single men. this suggests that, within this age group, women benefit more from their participation in employer-sponsored retirement plans, possibly due to better utilization of employer benefits, more favorable employment conditions, or other sociocultural factors influencing their participation. financial planners can offer customized financial planning services based on age groups, addressing the unique financial needs and goals of single males and females. they can work with employers to create programs and workshops specifically designed for young professional women to maximize their retirement savings potential, encouraging early and consistent participation in employer-sponsored retirement plans to build a strong foundation for future financial security. ahmmed et al. 161 these findings also have important policy implications. policymakers can improve access to retirement savings by supporting communitybased financial literacy programs and mandating automatic enrollment in employer-sponsored retirement plans. finally, the results find that single males have significantly lower returns to bachelor's degrees than single females. given the significant unexplained impact of having a bachelor’s degree, financial planners should leverage the educational background of their clients to tailor more effective retirement planning advice. overall, the findings of this study contribute to both public policy and financial-planning practices by offering actionable recommendations to support retirement preparedness among single individuals. conclusion saving for retirement is one of american adults' most important financial decisions. in the current retirement market of the united states, access to employer-sponsored retirement plans is equal for all individuals as long as they meet the eligibility requirements. the current study is conducted on single, employed individuals to see if there is any gender gap in employer-sponsored plans such as 401(k) or pension plans among single individuals. this study's unique approach involves examining whether a gender gap exists in employer-sponsored plan participation and, if so, determining whether it can be attributed to explained or unexplained factors. the current study examines how differences between men and women in the explanatory variables (financial literacy, risk tolerance, and other demographic and economic variables) and differences in the returns to each explanatory variable are associated with employer-sponsored retirement plan participation. the study uses a nationally representative data set from the 2021 national financial capability study and performs the oaxaca decomposition. this study finds that females’ participation in employersponsored retirement plans is 1.36% higher than males'. the explained and unexplained gap in participation in employer-sponsored retirement plans is -0.0428 and 0.0565, respectively. the negative explained gap means females have a lower average value of explanatory variables than males. the positive unexplained gap means females have a higher return to explanatory variables than males. even though previous studies find that women generally participate less than men in employersponsored retirement plans, the current study finds that single, employed women have slightly higher participation than single, employed men. single women may be more likely to have worries about retirement and their financial future (malone et al., 2010). they are solely responsible for their financial well-being. single women may be more financially independent than married women and single men. this independence can motivate single women to prioritize retirement planning and take positive actions to secure their future. single women may be more motivated to engage in long-term financial planning, such as retirement planning, because they do not have a spouse or partner to rely on for financial help. they understand the significance of making a nest egg to provide for themselves in the future. another reason for participation of single women in employer-sponsored retirement plans is their ability to access and use retirement plans offered by their employers. in contrast, some married women may rely on their spouse's retirement plans. moreover, women generally have longer lifespans than men (maklakov and lummaa, 2013). this longer life expectancy may motivate single women to participate in retirement plans to ensure they have sufficient funds to support themselves later. limitations one of the limitations of this study is that the nfcs does not contain information about the dollar amount of retirement savings in employersponsored plans. this limits the ability to draw comprehensive inferences about the financial readiness of single individuals for retirement. another limitation is the use of cross-sectional data which does not allow causal inference. the data provides a snapshot of the participation in employer-sponsored retirement plans at a single point in time, which means that any observed relationships between explanatory variables and retirement plan participation may not necessarily suggest causation. longitudinal data would be financial services review, 33(2) 162 more appropriate for examining causal relationships but is not available for all relevant variables. the decision to use the 2021 nfcs data was driven by its comprehensive coverage of financial literacy, risk tolerance, and demographic and economic variables pertinent to the study. future research could address this by utilizing longitudinal data to track individuals over time, allowing for the identification of causal relationships between financial literacy, other variables, and retirement plan participation. the study focuses on financial literacy and other demographic and economic variables. however, another limitation of this study is that other important factors are omitted that may influence retirement plan participation, such as future time perspective, peer effects, workplace characteristics, employer matching contributions, and job stability. the exclusion of these variables may lead to omitted variable bias. they are not available in the data. in addition, the measure of financial literacy used in the nfcs may not capture all dimensions of financial knowledge and skills relevant to retirement planning. the financial literacy questions are limited in scope and may not fully reflect an individual’s comprehensive understanding of financial concepts. finally, the oaxaca decomposition method assumes that the differences in retirement plan participation can be decomposed into explained and unexplained components based on observable variables. this method relies on the assumption that the model is correctly specified and that all relevant variables are included. one important limitation of this method is its assumption that the estimated coefficients accurately reflect true returns to the explanatory variables without bias. however, if the model is mis-specified or omits important variables, the resulting decomposition may produce biased or misleading estimates. while the method is useful for identifying gaps and their sources, its results should be interpreted with caution. while the study advocates financial literacy programs and tailored retirement planning strategies, the implementation and effectiveness of these recommendations are not measured. future research should assess the impact of specific interventions intended to increase retirement plan participation and reduce the gender gap. references anderson, d., & collins, j. m. (2017). can knowledge empower women to save more for retirement? (ssrn scholarly paper id 3045306). social science research network. https://doi.org/10.2139/ssrn.3045306 ando, a., & modigliani, f. (1963). the “life cycle” hypothesis of saving: aggregate implications and tests. the american economic review, 53(1), 55–84. anspal, s. (2015). the gender wage gap in estonia: a non-parametric decomposition. baltic journal of economics, 15(1), 1–16. https://doi.org/10.1080/1406099x.2015. 1022436 bajtelsmit, v. l., & bernasek, a. (1996). why do women invest differently than men? financial counseling and planning, 7, 1-10. bajtelsmit, v. l., & vanderhei, j. l. (1997). risk aversion and pension investment choices. in m. s. gordon, o. s. mitchell, & m. m. twinney (eds.), positioning pensions for the twenty-first century (pp. 66–78). pension research council. becker, g. s. (1994). human capital revisited. in human capital: a theoretical and empirical analysis with special reference to education (3rd ed., pp. 15–28). the university of chicago press. berger, e. d., & denton, m. a. (2004). the interplay between women's life course work patterns and financial planning for later life. canadian journal on aging / la revue canadienne du vieillissement, 23(s1), s81–s95. bollen, n. p., & posavac, s. (2018). gender, risk tolerance, and false consensus in asset allocation recommendations. journal of banking & finance, 87, 304-317. https://doi.org/10.2139/ssrn.3045306 https://doi.org/10.1080/1406099x.2015.1022436 https://doi.org/10.1080/1406099x.2015.1022436 ahmmed et al. 163 bucher‐koenen, t., lusardi, a., alessie, r., & van rooij, m. (2017). how financially literate are women? an overview and new insights. journal of consumer affairs, 51(2), 255-283. https://doi.org/10.1111/joca.12142 chen, h., & volpe, r. p. (2002). gender differences in personal financial literacy among college students. financial services review, 11(3), 289–307. clark, g. l., & strauss, k. (2012). saving for retirement: intention, context, and behavior. oxford university press.cordova, k., grabka, m. m., & sierminska, e. (2022). pension wealth and the gender wealth gap. european journal of population, 38(4), 755-810. dietz, b. e., carrozza, m., & ritchey, p. n. (2003). does financial self-efficacy explain gender differences in retirement saving strategies? journal of women & aging, 15(4), 83–96. https://doi.org/10.1300/j074v15n04_07 dwyer, p. d., gilkeson, j. h., & list, j. a. (2002). gender differences in revealed risk taking: evidence from mutual fund investors. economics letters, 76(2), 151–158. https://doi.org/10.1016/s01651765(02)00045-9 feng, j., gerrans, p., moulang, c., whiteside, n., & strydom, m. (2019). why women have lower retirement savings: the australian case. feminist economics, 25(1), 145-173. fisher, p., hayhoe, c., & lown, j. (2015). gender differences in saving behaviors among lowto moderate-income households. financial services review, 24(1), 1–13. https://doi.org/10.61190/fsr.v24i1.3232 gibson, r. j., michayluk, d., & van de venter, g. (2013). financial risk tolerance: an analysis of unexplored factors. financial services review. graham, j. f., stendardi, e. j., myers, j. k., & graham, m. j. (2002). gender differences in investment strategies: an information processing perspective. the international journal of bank marketing, 20(1), 17–26. https://doi.org/10.1108/0265232021041 5953 harahap, s., thoyib, a., sumiati, s., & djazuli, a. (2022). the impact of financial literacy on retirement planning with serial mediation of financial risk tolerance and saving behavior: evidence of medium entrepreneurs in indonesia. international journal of financial studies, 10(3), 66. https://doi.org/10.3390/ijfs10030066 hsu, j. w. (2016). aging and strategic learning: the impact of spousal incentives on financial literacy. journal of human resources, 51(4), 1036–1067. https://doi.org/10.3368/jhr.51.4.10146712r huang, y., & curtin, j. (2019). a review of gender differences in retirement income. journal of retirement studies, 23(2), 4560. https://doi.org/xxxx huston, s. j. (2010). measuring financial literacy. journal of consumer affairs, 44(2), 296-316. https://doi.org/10.1111/j.17456606.2010.01170.x kappal, j. m., & rastogi, s. (2020). investment behaviour of women entrepreneurs. qualitative research in financial markets, 12(4), 485–504. https://doi.org/10.1108/qrfm-04-20200053 lee, y., & tang, f. (2015). more caregiving, less working: caregiving roles and gender difference. journal of applied gerontology, 34(4), 465–483. https://doi.org/10.1177/0733464813508 649 loibl, c., & hira, t. k. (2006). a workplace and gender-related perspective on financial planning information sources and knowledge outcomes. financial services review, 15(1), 21–34. lusardi, a., & mitchell, o. s. (2011). financial literacy and retirement planning in the united states. journal of pension https://doi.org/10.1300/j074v15n04_07 https://doi.org/10.1016/s0165-1765(02)00045-9 https://doi.org/10.1016/s0165-1765(02)00045-9 https://doi.org/10.61190/fsr.v24i1.3232 https://doi.org/10.3390/ijfs10030066 https://doi.org/10.3368/jhr.51.4.1014-6712r https://doi.org/10.3368/jhr.51.4.1014-6712r https://doi.org/10.1108/qrfm-04-2020-0053 https://doi.org/10.1108/qrfm-04-2020-0053 https://doi.org/10.1177/0733464813508649 https://doi.org/10.1177/0733464813508649 financial services review, 33(2) 164 economics and finance, 10(4), 509–525. https://doi.org/10.1017/s147474721100 045x maklakov, a. a., & lummaa, v. (2013). evolution of sex differences in lifespan and aging: causes and constraints. bioessays, 35(8), 717-724. https://doi.org/10.1002/bies.201300017 malone, k., stewart, s. d., wilson, j., & korsching, p. f. (2010). perceptions of financial well-being among american women in diverse families. journal of family and economic issues, 31(1), 63– 81. https://doi.org/10.1007/s10834-0099176-5 mitchell, o. s., & lusardi, a. (2022). financial literacy and financial behavior at older ages (wharton pension research council working paper no. 2022-01). https://doi.org/10.2139/ssrn.4006687 nguyen, t. m. l. (2015). the influence of financial risk tolerance and risk perception on individual investment decision-making in a financial advice context (doctoral dissertation, queensland university of technology). queensland university of technology. niessen-ruenzi, a., & schneider, c. (2019). the gender pension gap in germany (ssrn working paper no. 3455189). https://doi.org/10.2139/ssrn.3455189 oaxaca, r. l., & ransom, m. r. (1994). on discrimination and the decomposition of wage differentials. journal of econometrics, 61(1), 5–21. https://doi.org/10.1016/03044076(94)90074-4 olajide, o., pandey, s., & pandey, i. (2024). social media for investment advice and financial satisfaction: does generation matter? journal of risk and financial management, 17(9), 410. https://doi.org/10.3390/jrfm17090410 pandey, i., & guillemette, m. (2024). social media, investment knowledge, and meme stock trading. journal of behavioral finance, 25(1), 1–17. https://doi.org/10.1080/15427560.2024. 2361875 preston, a., & wright, r. e. (2023). gender, financial literacy and pension savings. economic record, 100(325), 58–83. https://doi.org/10.1111/14754932.12708 reinhard, s. c., feinberg, l., & choula, r. (2012). a call to action: what experts say needs to be done to meet the challenges of family caregiving. aarp public policy institute. https://www.aarp.org/content/dam/aarp/r esearch/public_policy_institute/ltc/2012/ spotlight-12-aarp-ppi-ltc.pdf sierminska, e. m., frick, j. r., & grabka, m. m. (2010). examining the gender wealth gap. oxford economic papers, 62(4), 669–690. https://doi.org/10.1093/oep/gpq007 spivey, c. (2010). desperation or desire? the role of risk aversion in marriage. economic inquiry, 48(2), 499-516. https://doi.org/10.1111/j.14657295.2009.00269.x tomar, s., baker, h. k., kumar, s., & hoffmann, a. o. i. (2021). psychological determinants of retirement financial planning behavior. journal of business research, 133, 432-449. https://doi.org/10.1016/j.jbusres.2021.05 .007 weller, c. e., & tolson, m. e. (2017). too little or too much? women’s economic risk exposure. the journal of retirement, 4(4), 69-83. https://doi.org/10.3905/jor.2017.4.4.069 weller, c. e., & tolson, m. e. (2018). do unpaid caregivers save less for retirement? the journal of retirement, 6(2), 61–73. https://doi.org/10.3905/jor.2018.6.2.061 zhao, r., & zhao, y. (2018). the gender pension gap in china. feminist economics, 24(2), 218–239. https://doi.org/10.1080/13545701.2017. 1411601 https://doi.org/10.1017/s147474721100045x https://doi.org/10.1017/s147474721100045x https://doi.org/10.1007/s10834-009-9176-5 https://doi.org/10.1007/s10834-009-9176-5 https://doi.org/10.2139/ssrn.4006687 https://doi.org/10.2139/ssrn.3455189 https://doi.org/10.1016/0304-4076(94)90074-4 https://doi.org/10.1016/0304-4076(94)90074-4 https://doi.org/10.1080/15427560.2024.2361875 https://doi.org/10.1080/15427560.2024.2361875 https://doi.org/10.1111/1475-4932.12708 https://doi.org/10.1111/1475-4932.12708 https://www.aarp.org/content/dam/aarp/research/public_policy_institute/ltc/2012/spotlight-12-aarp-ppi-ltc.pdf https://www.aarp.org/content/dam/aarp/research/public_policy_institute/ltc/2012/spotlight-12-aarp-ppi-ltc.pdf https://www.aarp.org/content/dam/aarp/research/public_policy_institute/ltc/2012/spotlight-12-aarp-ppi-ltc.pdf https://doi.org/10.1093/oep/gpq007 https://doi.org/10.1016/j.jbusres.2021.05.007 https://doi.org/10.1016/j.jbusres.2021.05.007 https://doi.org/10.3905/jor.2018.6.2.061 https://doi.org/10.1080/13545701.2017.1411601 https://doi.org/10.1080/13545701.2017.1411601 pii: s1057-0810(01)00079-8 ��� ������� � � �� �� �� ���� ��� �� � �� � �� �������� �� � ����� � � �� � ����� ������� �� ������ � �� �� � � !��" ������ �����# ���������� �� ��� ������ ����� � ��� �� � ��� � ������ ����� �� �� ��� ��� �� ��� ��� �� ���� � ��� $������� %& '�(����� )**%+ �������� �� ������� � �� %& ���� )**%+ ���� �� %) �� ��(�� )**% �������� ,� ��� � ��" �� �� �(���� ��� �� ����(���� ���� ���� ��� �� � � ������� ����� �� (� � �� ����� � � ������� � ����� ���� (� ��( ����� ����" ����� �� ��� � � �� � ����� ������� �� ���� �� �� �-�� ��" �� ������ � ��� ��� � � ��� ��� � ��� � ���� � �������� �� �� ( � � � �������� (� ������� �� ���� ��� �� � � ���" ��� �� � � � � � �� �� ���� .� /��" /�� " �� �� � ���� �� ����� �� ��� � �� �� �� ���" �� ���� �� �� (���/ � ��� ��" �� ��� �� � �� ������ �(��� � � ������ ���� �� (���/ �+ ��������� ��� ���/�� ���� ��� �� � � � � � � �(��� � ����� �� � �� ����� � �� �� ��� � �� �� ������� �� �� �(���/ � � )**% 0������� ������� ,��� 1�� ���� � ��������� !�" ��� � ����� �# 2%%3� � � �� �� ��� $�%& �� # 1��� + 1�� �� � �+ � �����+ �������+ ������ � �+ 4 � � ���� � ���� ������ ,� ���� ��� ����� ������� � � � ����� ������" ���� � � �� �� ����� ����� ���� ������ �-5����� ���� ����� 1 �� ��� � ��� ������ �" ��� &* � � �� ���� � ��� �� �� � ���� � ������� ����� � � ������� ����" ������ � �� �� ��� � � &)6 ���� 7���� ����� ������������� �� � � � ������ � � � �� ���� �� � � � ����8� ���� � � � � �� ��� �� � � ������� � ����� � ��� ���� �������� � ������ ��������� ������� � � �������� ,���� ���� � ����� ������� �� � �� ���� ������� ��� ������� ���� ���� 7%8 �� � � ����� ������ � � ��(� � �� ��� ����� �(�� �� � � � � �-�(�� ��� �� � � � � � ������� �" 7)8 '�������� �������� $����� %* 7)**%8 %*%9%%: %*&;5*<%*=*%=> 9 ��� �� � �� �� � )**% 0������� ������� ,��� 1�� ���� � ��������� �,,� � % * & ; 5 * < % * 7 * % 8 * * * ; 6 5 < # � ���� ����� �� � �� ����� �%5;%:5?6&5&&%6+ ��-� �%5;%:5?6&5)&@)� �'���� ����� # ������a(� � � ���� 7�� �����8� ���� � �� � �-5��������" ��� 7?8 �� �������� ��� �� ��� ��-�� � �-�� � �� ������� 7 ������(��8 � ��� ��� � ���� �� ������� ��� �� ��� ��� ��� ��(� �� � � b����/�� ����� � ��� ������� 0������ 7%66;8" � ����� � ���(�� � �(��� � � ���� �� ������� � �� ������ �� �� �� � ������ ����� �� � ����� � ��� ' � � � � ��� ��� ������ � � ��� ������ � � � ���� �� � � � ���� ��� �� �� ��� �������� 7)**%8" c �� 7)***8 ��� .��� 7)**%8� ' � �� �-������ � �� � � � ��� � �� ����� � � /������� ������ ����� � � ��� 4�� � ��� $������� ��� 7)***8� 4 � � � � � ����� �� � � � ����� ��� � ��� �������� 7����" �,115�$0'8" �� ���b��� �� � ������� ����� �� � � �� � �� � ����/��� ���� ��� �� � �� 7 � �� ������ 4������ 1�� �� � � �����8 �� ������ /-�� � � � ���(�� � ������ 1� �� ����� ����" �� ����� ��� � ������ ���� �� �� � ���������� � �� � ������ ���" ����/��� ���� ��� �� � ��� �������(�� ����� ������� ���" �� ��� �� � �� ���� ���� ���� �-������ ������ � � ��� ��� ����� � � ������� � � �� � ' � �-�� ��" ����� ����� � � �� � ����� ������� �� ������ 7�����8" � � � ����� ( �� �� %66<" �� � � ��(� � �� ��� ����� �� &&d �� � � � ��� @&d �� ( ��� � � ) �����" �� �� ����� �� � � ��(� � �� ��� �������� � �� ��� ����e�� �� &*d � � � ��� &*d ( ���� ��� ��� �� � � �� ������� ����� ����� ) ����� �� � ���������� � � � �� (���� ��� �� �� ( ���� ��� ���� ��� �� � � � ��� ��� � �� ����� � � %666 �� ����� �� �� ��(�� %� .���� ��� �������� �� � ���/� �� �����" � ���� )* ��� � � �� ���� ������� � ��� ���� ���� ��� �� � � ����� f� ����" � ���� � ���� �� ���� ��� �� � � � ��� ��� ���� ���� ��� � � ��� ��� /������ � �� �������� ��� ����� �� ��� � ����(�� �� ���� � ���� ��� � � ��� ��� ����� ���� � � � � ��(� � � � � >&"*** �� ����� �� ���� � � ��-���� � >%*"*** �� � ���� �� � � �� ��������� /���� e �� ���� �� � �� 7�,115�$0'" %6668� 1 /�� ������" �� ������ ���� ��� � ����� ��� ��������� ���� (� � �� � �(��� � �" ������� ������ ����� ��� ������� �� � �� .���� �� ������ ���� ��� � ������" �� ����� ������� � �� ��� �� ������ �� ����� � �� �� � � �� ��� ����� �� ��� ��� ���� , �� ���(�� �� �� �(���/ � � ���� ���� ���� (� ������� ��� (� �� ������� �� � � �� ��� ����� ��� �� � ����� �� ���� ��� �� � � � ��� ����� ,� ��� � ���" � �� ���(�� � � � �-�(�� ��� 7����� ��� � � ��(� � �� ��� � 5 ���� �������� ��� �-�(��8 ��(�� % 1��� ��� �� � � � ��� ��� � �� ����� 7%6668 ���� � (�� � � ����� g��� ����� ���� 7�� ��� � �)d � (���8 2� � � � � � g ��� 4 ��� ��� � %66< � 66 && @& * %66: � 6; &* &* * %66@ � 6& @& && * %66) � 6? @* :* * %66* � 6% ?& :& * %6<< � <6 ?* ;* * %6<: � <; )& ;* & %6<@ � <& %& ;* %& g�� �� %6<@ %* @* &* � ����� �,115�$0' 7%666" 4���� ?*8" 1�������� %*) !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 ����� ���� �� ���5 �5 ��� � ���� � 7 � �������� � ���� ���� � �� � � ��� � � ��(� � ��8" ��� �� � � � ������� ���" (���� � ����� ��� � � ���b���5 ����� (� ��(e�� � h�-�(�� ���� ��� �� � �� ��0b���5 ����� ������ �� ����� �����b������ � � ����� ��� �� ���� ,� 3 � ����� ��� �����b������ � � ����� ��� 1" (� 37%� �8 ��� �����b������ � � ����� ��� g" ����� �3 � �� �-��" �� �� ����� �� �b��� (������ 3 � ����� ��� (���� �� ������ �� � � ���� � � �� ( � ������ ���� � �� � � ���� �� ����� ��� �� ��5 ����������� � � � ���� ��� �� �� 7 �-5 ��������= ����� ����8 � � � ������� � ����� �� ����� ��� �� ��5 ����������� � � � ���� ��� �� � � �-5 �������� �� ���� ����� '� � ���� ����� �" � �� ���(�� ��� �� �� � �� � ��� � �� ����� ����� � � � � �� ��� ���� ��� ���� ��� ���� ��� �� ��� � ����� � � � �� � ��� ��� � � ��� � ��� (���/����� 4 � � ���� ������ � �� ��� ��� ���� �� ���� ���� ������ �� �� ��� �� ��� ������ � �� ������ 7' � � �������� � � 4 � � ���� �� � �� ��� i����� � j 4��k��� �" %66?" ��� � )%�8 i����� � ��� 4��k��� � 7%66?8" � ;?) ������(� �� 4 � � ���� � ������ �� �� ���� ����� b��� � ��� � �� ���� ��� � � -��� �� (� ������ ��� ���� ���� � �����!� � �� � � �� � ���� �� � ���� ��� ��������� ��� �� � �� ������ g� ���� � �� ���" �� ������ � �� �������� � � � ���� � �� �� ����� � � ���� �� ��� ��" 7��� � ���b��� �� ����� ���� �� ���� ������� ��� /-�� ���� ��� �� � �� �� � ���/�� �� ��(�� %8 ��� � � ��� � �� � �� �������� � � � ���� � �� � � � � ������� ��� ����� ��� � �� ������ .� ������ � � ������ � (����� � �� � �� �������� �� ���� � (���� � ���� ��� ����� �� �� �� � �� ��� � � � �" ( ���" ��� � ��� ��� � ��� ����� �" � � �������� ��� �� ��� � � �������� � ����� ��� ��� ��� � �� � �" ��� ���� �" ��� ��� � � �� ����� ��� ���b����� ��� ��(� � �� � ���� �������� �� �������� ��� �� � ��� � �� � ���� ��� ��������� � ��� � �� �� �����!�� �� � �� ��� ,� ��� � � )" � ������ � �� �� ��� ��� �� � ������ ,� ��� � � ?" �� �� � � � �� ��� � ��� ���������� �� ������ � �� �� ���������" ��� �� ��� � � @ �� ����� � � �� ������ � � ��� � ������ ��� � � & � ������� �� � �� ��� ������ � � �� ����� ��� � �� ������ � ���������� ������ ����� ��� ���� ������ � �� ��� ������� �� ��� ������ � � � � ������� ����� 7%8 ���� ��� �� � �" 7)8 �-��" ��� 7?8 �� ��� ��� ����� ������� �������� �� �(�( � �� ��� ���� � ��� �� � ���� � ���� ��� �� � � ������ �� ��� (��� ������ � ����� �� �� /����� �� ��� ��� 70� � ��� 2��(��" )***+ ���� �" )***8� ����� � �� � � �� �� ��� 7���8 b� �� � ��� � ��� ��(e�� �� �� g���� �" l �" ��� g��( ��� 7%6<:8 � ��� 7��� �� �� g���� �" ������" j g��( ���" %66%" %66&8 � ���� � ���� ����� �� � �� �� ���� ��� �� � � ������ � ��� ����� �� �� 6*d � �� �����(��� � � �� ���� -�� ����� ,(( � � ��� k� ��� 7)***8 ���� �� � ������ �� � g���� � � ��� 7%66%" %66&8 (� �� ���� �� �� ���� ��� �� � � ������ � ������ �� ��� @*d � �� ����� � � � �� ���� ��� ����� ��� %**d � �� ����� � � �� �� ������� �� !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %*? ����� �� �� � �� � ������ ��� �� ��� ���� � ���� (��� ������ ��� g�� � ��� ������� �-�� ��� ����� �� ��� �� ������ � ��� ��� � � ��-� ��� � ������ �� � �� ������ � ��� ����� �� � � � �� � %**9%@& ����� ����� � ���" ����� � ����� �� � � � � (� ��� �� �� �b�� ��� 7g�����" %66:8� � � ���� �� ���!� � � ��� ������� �������� 7������ �� (� k � ����� j i��� �" %66;8" ����� ��� �� � � � �� ��� ��� � � �� ��� �� �� � ����� � � � ���� �������� � ' � ���������� ����� ��" � � /������� ������� ��� ����� ��� ��������� � ���5 �� ���� ��� ����� � ��� ��������� ��� �� ������� ���� ��� �� � �" ������� �� ��� � b��� � ������ ��� � �� ���!� � � � ������� ����� � �����" ���� ���� ������� ���� (��� � � �" ( ��� ��� ����� � �� ���� ����� �� � � �� � �� ������(�� (���� � � � ����� �� ��� � � ��� � �� ��� ������� �� �� ��� � /�� � �� ����������� 1 ��� ����� �� �� ���" � � (� ����� � ���� ���� � ����� ������� ���� ��� �� � ��� ����� �� � ���� ��� � �� �� ���� � ����� � �b�� � ��� ( �� �� ���� 71������ ����" � ����" j ���� �" )**%8� 1������ ���� � ��� 7)**%8 /�� �� �� ���� ��� �� � � � ��� �� ����� �� ��� �� � � �� �� ���� ��� ������� ��� � �� � ���� ����� �� ��� ��� � � ��" � ��� ��� �� /-�� ����� ��� �� � � ����� ���� ����� ����� � ������ /-�� ��� �� � ����� ��� �� , �� � ���/��� /�� � �h�� ��� ������ � ���� ��� �� � � � /�� �� �� ���� �� � � ���� �� ��� �� � � � �� ������ ' � �-�� ��� � ������" 4�� ��" ��� .��� 7%66;8 � ���� �� �� � �� � � ����� � � � ���� ���� ��� �� � � ��� ������ � �� �� �� ��� � � �� ���� ����� �� ��� (� �� ���� � � �� ��� �� ��� ��������� � �� ����� � �� ��� �� ������ �� � � � � � ���� � �� ��� (� ��� �� �� (� ���� �� ���� ��� � � � �� ��� �� ��� ���� ���� � g���� � �� �� ��� � � �� ��� ��� ����� � �� � ���� �������� ���!� � �" ��� � �� ���� ��� �� � ��" ���� �� ������ �� �� ��(�� %" ��� ��( ����� � ������ � ��� 7%66;8 ����� ���� �!!����� �� ��� 5������ ����� �� ��� � �� ������� �� ������� � ���� ���� � �� ��� � �� � ���� �� � � � � � ���� ���� � �� /-�� ��� �� 7� � ��� 8 ������ ���� ��� �� ��� ��� � �� � ������ �� ��b����/�� ��(�� �� ���� ��� �� � �� 4��� � �� � �� ����� � �h�� �� � �� ��� �� �� ���� � � ��� ������ �� (� ��(�� %" ����� ���� � � � ��� ����� ������� � �� ���� ��� �� � � ������ �� ����� �� (� �� � � � ������ ��� ���� (� ����� ��� �� �� �� ������ � � ����� � �� ��� �(�( 4�5� ��-��" �� � �� � � �����/��� � � � � ����� ���" ��� ��� ��� (� � � ����� ��� � �� ����� 7m�� " %6668� g ��� ��� ����� 7%66;8 /�� �� ����� �� � � ������ ����� ��� �� � �������� ��� �-�(�� ��� �� � �� ���� ��-���!� ��������� ,� ��� " 1�� " g�� ��" ��� �� 7)***8 /�� �� �-�� ��� � �����/��� � � � � �� �� � � � � ����� ���� i �� � ��/��� ���� ��� �� � � �� � �-5 � �� �� ��� ������� � �� � (���n ��� ������ � � ���� � ���� ������ �� ��� ��� �� � � � ��� �� ���� � ��( ���� �� � �� ����� � �� ���� ��� �� � � ��� �� � � �� �� ���� �� %*@ !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 �(6( 4��� * ��) � ����� � ���� (� � � ������� � � � �� ���� ��� �� � � �� ��� ������ ��� ���� ' � �-�� ��� � g������ 7%66;8 ��� g ��� 7%66&8 �� � �� � � � �� ���� ������ �� ��� � ��! � ��������� ��� � � 0� � ��� 2��(�� 7)***8 �� � �� � ���� (� �� ���(�� e���� �� �� � ���� � � ��� ���� ��� �� � � ��� ������ � �" �� ������� �� � ��� � ��� � ������ � � � ,� �� � � � � ����� � � �� ��� ������ � ( ���" � ��� ��� .��� � 7%6668 ��� � � ������ ���������� �� � � � �� � /-�� ��� �� ���� � �� �� � � �� �� ��� � � � � ������ �� ����� ���� �� � �� ��� �� � �� ��� �� ���� �� ��� � ( ��� �� � ��� ����� (� ��� ��������� ����� ������� � ��� ������� � ����� ������� ����" �� �� ����� �� ���� ��� �� � �� ��b���� ���������� ��� �� � �� /-�� ��� �������� ��� �� �� ������� �������� ��� ������ � � ���� � ���� ������ �� ��� ��� ����� � ��� � ��� �� � �� ��� � �� (� �� � � � ������ ��� ��� ����� ��� ���������� � � � ����� ������� � � �� ���� ����� ���� ��� �� � �� ��� � ������ ��" �� � ���� ��� � ��! � � ����� �� ����� � � ���� �� � ��� � �� �� ���� ��� �� � ��� �(�( ��7 ��������% ���� �-��� ���� ��� ����� �� � ��� � ��� � ������ � " ����� �� ���� ��� �� � � ������� ��� ���� g���� � �� � � � � � ������ � ���� ��� �� � � �� ��" �� ������� �� � ��� � ��� � ������ � " � �� � ���(�� ��/�� � ���� � �� �� � ���� � � �� ���� ��� �� � �� (���� ���� (� �� ������ l �����" ���� � ��� � �� ������ �� ���(�� ��( ����� � � �� ������ ����� ��� �� � � � ����� %� 1�� ���� ��� �� � �� ��� ���� ���� ����� � �� ��� ������ � 7� ��� � ���� � �8 � �� ����� �� )� ��� /-�� ����� � �� �� ������� ��� �� � � ���� ������ �� (���� ���� ��� ��� �� �(��� � � �� ����� � ��� �� � ��� �� �-�(�� ��� �-5�������� ����� ��� �� ?� ��� ����� � �� ������ � h�-�(��� � �������� �b�� � � ���� ���� ����� �� � �� ���(��� � � � ��� ��� ���� ���5 �5 ��� � ���� �� ,� ������� ����� ���� �� ����" ����� �� �� �� � � � �� � � ������ �� ��" 4 � � ���� � � � �� ������ � �� ��� � ���� ���������� ������ �� ,� �� �� ��� � �" �� (���� �������� � � �� 4 � � ���� ������ � � ��� �� ����� ��� � � �� � � ���� ���� ���� ��� ��� � ����� ����� 6(�( � ����� 0��� ����" ����� � ����� ��� �� ������ �� � � ���� � � ��� � �� �� � � � �� � ������� � ��� � � � � ������� ���� 7���8 ����� ��� ��� ��� ������ ����" ��� � !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %*& � �5� � � ������� ���� 7����8 ����� � �� � � ,� �� ���" ��� � � ����� ��� � �� �� ���" ����� �� �� ���� ��� �7%� � �8 ��� � ��� ��� ��� ����� �� ���� ����� ��� �-5��������" ����� �� ���� ��� �� ( � � � � ��� ������� �-�� � ��������� ����� �� ��� ��� � ��� 7��� � ������ ���8 ���� ��� �� � � �����" � �� b�� � ���(�� �� ��� � � �� ����� � ���������" � ��� ����� ���� ��� �� ����� ����� �� ��� (�� ��� � ���������� � �� � ������ ��� �� �� ����� ����� � ����� ���" 7����� ��(�� %8" � �� b�� � ���� ��(�� ��� �� �� �� �� �� �� �� �� ���� ����" 7 �� ���� �� ����� �� � � �� ��8" �� ���� ����� ��� �� �������� ���� (�� ����� � � ����� ��� �� ������ �� � � ���� � � �� ���� (�� �� �� ��� �� ��� � �� ����" � �� ��� (� ���� ����� ���� ������� � �� �������� �� ���� � � �� ��� ��� �� ���� � � ���� �� ������� �� ' � �-�� ��" (� ������ ��� ��� �� �� ���� (���� � ��� ����� �����!�� �� �� � �� ����" � �� ��� (� ���(�� ������� � �� ������� ���� � �� � ��� �� ��� ��� �� �� ���� � � ������������ � ���� �� ����� 6(�( � ����� � .� �� ������ ��� � ��� � ���� � �������� � �� (���� � �� � �� ���� ����� � ��� � ��� � � � � � �� �� ���� � ��� �-�� �� 7�,115�$0'" %6668� $���� � ��� � ��� � ��������!� ��� � � � ������� ���� �� ������ � � �� ���� ��� �� � � h�-�(��� � ��� ����" ���� ����� ����� �" ��� � � � ��� ��� � ��� ������� (�� � �� � �� �� ��(�� %� g� � �������� �� �� ��(�� % (�� � �� �" ���� ���� (� ���� ����� � � � ������" ��� ��� ������� ������� � �� ' � �-�� ��" ����� ����� � �� ������� %; � (�� ���� ���� � � ��� � � � ���� ���� (�� �� (���� �� ������ � � � ����� ������ � �� ������� %< � (�� ����� � ��� ��� ����� � �� ���� ���� � ��� � � �" ( ���" ��� � ��� ��� � � ������� �� ���� ��� �� ������� �� � � � �� �� �� �� �� ��(�� %� � ,���� ��� � �� �� ���� ��� �� � �� �� ���� � � �� ����� � ��� ��� ����� �� � ���� �� �b�� ���� .���� �� ��e � ���� � �� ����� ��� � � � � �� ����� ������� %**d � � �" �� �� �� ������� ����� � �� � ������ � �� � � � �� � ������ ��� ��� �� � � �������� � 1 � �� �� �� �� � ����� 7>&**8 �� ����� �� ��� ���� ���� � �� ������� (�� � �� � 7����� � �&**8� � .� ��� �� �� � � ��� ��(� � � � ������ �� �� � �� ���" (� � � ��� b��� ����� 6(6( 8����7�� ��� � �� ���� �����(��� ��� ���� �� ��/���� �� � ��� ��� ������(��� �� �� ��� � � � ����� � ��� ����� �� ������ �� � � ���� � � � � ����� ��� �� ��� ���� ����� � � ��� �������� ������� ��� � � �� ���� ������ �� �������� � � ��� � ��� ������ �� ��� ��(� � � ��� �� � � ��� �������� � � � ��� � � �� ���� ������ �� �������� � � ��� � ��� ������ �� ��� ��(� � � ��� �� %*: !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 � � ��� � �(���� ������� ��� � � � ��������� ��� � � �� ���� " ����� �� �b��� � � � � ��%� � ��� ��� ��� �������� ������� ��� � � � �� � ���� ������ ��� ��(� � �� ��� ��� �������� � � � ��� � � � �� � ���� ������ ��� ��(� � �� ��� ��� � �(���� ������� ��� � � � ��������� ��� � � �� � ���� " ����� �� �b��� ��� � ����%� ����� �7�8 1����� �� � � �� ��� � � � �� �7(8 1����� �� � � �� ��� � ( ���� �7�8 1����� �� � � �� ��� � � ��� ��� � ��� �� �� ��7�8" ��7(8" ��7�8 ��� �� � � � � � �� � � �� ������� �� � � � � �" ( ���" ��� � ��� ��� � ���� � �� �� � � ����" ��� �� ������ 0��� � � *" ��� ����� � ���(� � ����� � %� 6(�( 4*� � ��� 1����� �� �� �� � �� ��� ������ b��� ����� '�� ���" �� ��� � %� ������@" �� � � ������ ��� � � � � � � �� �� � � b��� ��� ,� � ������� ����� �" �� ��( ��� ��� �� ����� �� b��� ���� � � ������ ��� �� � � ( ��� ��� � ��� ��� � �� ��� �� �������� �� �� �� �� �� �� /�� ���� �� �� � � � ������� ��� ��� (� �� ����� �� ��� �% � ���%��� ��%����)����?����@�� � �%�(� ��%(���)(���?(���@(� � �%��� ��%����)����?����@�� � 7%8 ��� �� �� � �� ��� � ���� % ������� �� ���� � 7%8 �" �� ����� ��� � � ��(� � �" 7)8 �� b��� ���� �� �� � �� ��� � � ���� ���� � �" ��� 7?8 �� ��" ����� �� ����� �� ���� ��� �� � ��� ��� ��� � ���� ) �� �� �� �� � �� � ��� � 0b� 7%8� 1 �� ��� � ���� )" �� � �� ���� �� ���� (� ������(�� � � �� � � � ���� �) � ��% � ����)��� ��%����)����?����@�� � �)�(� ��%(���)(���?(���@(� � �)��� ��%����)����?����@�� � 7)8 ' � �� ����" �� ��%� � �%� ��%����@��%� � ���" ����� �� �� ����� ��� b��� ���� �� ��5 �� � ������ ��� � � � � �� ��� �� �������� �� �� �� �� �� �� /�� ���� �� �� ���� ��� (� �� ����� �� ��� �% � ��%� � �� ���%�����)�����?�����@�� 7?8 ����� 7%8 �� ����� ��� � � ��(� � � �� ������� (� � � � �-��" 7)8 �� ����� ��� �� � � �� ��� �� ������� (� ������� ��� � � � �-��" ��� 7?8 ��� ���� � ��� �� � � �� ��� �������� � � � ��� � � ����� � � �� ���� �� ����� 1 �� ��� � ���� )" �� � �� ���� ���� ���� (��� �������� ��� �) � ��% � ��%� � ��� ���%�����)�����?�����@�� 7@8 ����� �� ���� ��� �� ��5 �b��� ���� � � ������ ��� �� �� ���� )� ��� ��� ��� b��� � � ����� ��� � �� ��������� ��� � ��� ��� ��5� ��� ���� � �*� � �& �-��� �������� � &��� �*� ��5 ��-������ �*� ��9 ��&���* �*� �� ��-������ �*� ����� � �� ��� ���� ��n ��� ��" ���� �4 7�� �� � � ��� � �-�� ���� (��� ���8 �-���� �4n !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %*; 6(:( ����������� ���� ������ '��� " � �� ������� �� �� ��� �� �� � �� ��� � � � � �" ( ���" ��� � ��� ��� � ��� ����� � � ��� �����(�� ���� ��� ����� ��� ��� ��� ������ � � � ���� � �� �� � �� ��� �� �� ������ � � (����� �� � � �� ��� ��/�� �� ��� � �� �� � ����� �� ���� � �� �� � �� ��� ��� �� ���� o���� � � ��� �� �� � �� ��� �� � %6<* %666 � � � � �" ( ���" ��� ?5� � � o��� �������� (���� 7�5(����8 ������� �� � �� ,(( � � �gg, �� �(��� 7,(( � � 1�� ��� ��" )***8" �� ����� ��� b��� ���� �� �� � �� ��� ��� ������� �� � � ���� � �� ������" � � ���� � �� )* ������ ��� �� ��� � �� 6*5��� �5(���� �� ���� �� � � -� � � � ��� ��� � ��� ����� �� ,� �� ���� � ( ���" � � ����� � � �� ��� ?�; ��i�� �� � �� ���� ���� (��� �-������ �� � �� ������ ��� ��������� f�� � ����" � � �� ��� & ��i�� �� � �� ����" ��� (��� �-������ �� � �� ?5� � � �5(���� �� �� i����� ��� � � �� ��� ��� ( ����� � � �� ���� ������ ��� k �� � � �9����� � �� � � � ����� � ��� ��� ���� � � ���� ������� ,� ���� ����" �� ����� � � � � ����� � ���� (� ��e�� �� 7� � *�%&" ����� �� �� ����� �� �� �� �� ����(�� �� � �8� ����� �� ���� ������ ��� ��� �� ��� ����� ��" ������ � �� �� �� � �� ��� � � ��� ���� ������ ��� ������� (� e �� 5 � ������ ��� ��(� �� �� � ����� � %<�?:d" :�6)d ��� )�6%d" ��� �� ����� � � � � �" ( ���" ��� �5(����� ��� ��i�� � �� �� ���� ��� ��(� � �� ��� ??�<@" )%�6;" ��� )�&? � � � � �" ( ���" ��� �5(����" ��� �� ������ ��� � ����� � � (� ���� � � � ��� ( ��� �� *�?&<" (� ���� � � � ��� �5(���� �� *�%&?" ���" (� ���� ( ��� ��� �5(���� �� *�&%)� $��� � �� �� � �� ��� ��� ����� �� �� � ��� ��� �� ���� ��� ��(� � �� ���� �� ��� ����� �� ���� ��������� � �� ���� ������ ��� �������� �� �� ������ � �� $��� � ����� �� ��� ������ �� ����� � �� ��� � ��� � �� � � 7��� ���� �" g���� ��" g����� " j ���" %6:<8 � �� ��������9� �������� �� ��-" ����� � ������ � ���� ����� �� ������ �� (� �� �� � � � g ��� 4������ 7k������ j 2�� ��" %6<*" � )*)8� ' � � �������� � � ������ ��� �����(��� �� � � �5����� � � ���� ��� ��(� � � ��� k������ ��� 2�� �� 7%6<*8" � ))<9)?%� , ���� (� ���(�� �� �� ���� �� � � �� ��� � � � � ��� �� ��� %* ����� ��� � �� ����� ��� �� ���� ,� ���� �� ��� ����� � �" �� � ��� ��� (��� �� ���� �� ����� ��� ����� �� ���� �� � %6;* �� ��� %6<6� ��� ������� b��� ���� �� ���� ������ ��� )* ���� ��� � �� %%�&;d" ?�6&d" ��� %�@6d � � � � �" ( ���" ��� �5(����" ��� �� ������ .���� �� � ���� �� �� � � �������� �� ���� � � � ( � �� ��� ��� ����� ��� ������� ������� � ���� � ��� �� �� � �� ���" �� ������� ����� � � ��� � �� ��� ��(� �� ����� ������� �� ����� � (���� � �� %6<*9%666 ��� ��� �� ����" �� ���� �� �� � �� ��� � � � � �� �� ��� ������ � � � ��� (��� �� ����� � �� ��� � � �� ������ 6(2( ; & �*� ������� � &� ��� ���� 0��5 �5�������5 ��� �" � 5 ����� � �� � � ����� ������� ��� ��� �� � ����5����� �� ��� �� � � �(���/ � ��� ������� �� �� �� ���� � � ���� �� � 6 ��� ��� � ����� ��� ��� ' � �� ���" �� ������ �� � � �� ��� �� ���� ��� ����� �� �� � �� e �� � (�(��� � ��� ��(� � � � �� �� � �� ���� 0��� �� � � �� ��� �� ���� �� �� ����� ��� b��� ���� �� � � �� ��� ��� �� �� �� � �%" �)" � � � �4 ��� ������� �� � ���� ����� 0b� 7)8" ����� 4 �� ����� � �� ���� � � ��� �� �� ����� �� ��� � �� �� �� ���� )" @" :" � � � � %< ������ %*< !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 ��� �� �� �� ������� � �� ��� � 4 ����� �� �� (����� � � �� �� 4 � � ���� ������ � �� ���� � ���� �� �� �� �� %"*** ����" � �� ���� ��� %"*** �������� ���� � �� �� � �������� �� ��� ���� � �� 6 ��� ��" ���� (����� � � ��(e�� �������� � ���� � ���� ��� �� ��� �� �� � �� ���" (� ���� (����� � � (��� � �� ����� �� ���� ��� �� � �� �� ��(�� %� ��� ���� � � �� ��������� �� �( ��" �� � �� �� � �%" �)" � � � �4 ������� �� � ���� ����� 0b� 7@8� 1����" �� � ���� �� �� �� �� %"*** ����" � � ���� � �� 6 ��� ��" ��� � � ���� � �� � �� �(��� � �� ' � ���� b��� ��" �� ���� �� �� � �� ��� � � � � 7�7�88 ��� ( ��� 7�7(88 ��� ���� � � �� ��� ��� �� ������ 6(<( ,��� � &*��* ��=����� �*� ��� �� ��� �� �� � ��� � ��� �� ( ��� � ����� � � � ����� ����� �� ��� �������� �� ���� �� ����� ��� ������� ��� ��" ���� ���� ��������� �� �� ��� � � )" ����� ����� �� �� �������� � � � ���� � (� �� ��� � � ���� �� � � ����� ����� ��� �� (� ��������� (�� �" ������� 7%8 �� �����(��� � � �� ���� � � � �" ( ���" ��� � ��� ��� � " 7)8 �� ���� ��� �� � �� �����" 7?8 �� � ��������� � ��� ���� �" ��� 7@8 �� � ����� ��� �� 6(<(�( 4*� -����7����% ������ � � �� � 7 �� � ��� � ��% ������ 1� � �� �� �� �������� ��� � �" �� � ��� � �� � ����� �� �� � �� ��� � � � �" ( ���" ��� � ��� ��� � � ����� ������� ������ �� ��� � %6<*9%666� ��� �� ���� ��������� �� �� �� ���� ���� ������� ��� ��� ��������� ��� ���� �� �� � �� ��� ��� ���� � � � � � �� �� ��� ��� �� �� ����� � � ��� ����� ����+ �������� � � ��� (� ���� ��� ��� ����� ��� ��� ��������� � ��(� �(�� �����(��� � �� �� �� � �� ��� � �� �� � �� ��� � � � � (���� ��������� ������ 6(<(�( � �� ��� ���� � ��� ���� ��� �� � � � �� ��� �� ���� �������� ��� � ����� �� �����" �� ������ �� � � � � � ����� �� � � � ��� �� ������� �� ( ��� ��� � ��� ��� � ��� ����� �� ��� �-�� � � � �� ��� � �� �� ��(�� %� ��� ���� ��� �� � � � � �� ����� �� 7��� �����8 � ���� � � �� 6(<(6( 4�5� ��� ��5 ��� ������ ��� ��� ��� ���� � ���������� �� ��� ��������(�� �� ����� ����� ������� 7%8 � � ��(� � �� �� ��� ��� �-�� �� � � � � �-" 7)8 �������� � � � �� ���� � � �-5 �������� (����" ��� 7?8 ��� ��(� � �� ��� �-�� � �� ������� ��� �� ��� ���� ��� � �� � ���� (���/ � ��� �� ����� � �� ������� ��� � ��� ��� �� �� � �� � �� ���� � , ��� (� ������ �������� �� � ����� �� ����� ���� ��� �� �� ����� �� o��� �� 2�� 4�� �� 1� 7o2418 ��� �� � ��� �� � � ����� � �� ������� �� ��5 ��� �� ��� � �� ���" � �����" ���� �� ������� ����������� ��� �� ���� ������� � ���� ��� ��(�� ������� � � ����� �� ���� ����� �� � �� ������������ ' �� ��� ��� �(��� � �" ��� (���� � �� '������ ,� ����� $������ ������� � �(��� � � )***" ��� ���� �� ������� ��� �� �� ��5 ��� �� � � � �� ������ ��m � ��-"�� ��4����� !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %*6 ��-"�� ��l��� ��-"�� ��� ��c��� l��� ��-"�� � ���� �� �� ������� �������� ��� �� � � ������� ���� � /���� e �� ��� ��m � ��-�� � ���� ��� �� � ��� '������ �(��� � " � � �-�(�� ��� �� �� 7)***8 � >@?"<&*" ��4����� ��-�� � ���� ��� �-�(�� ������� ��� �� � >%*&"6&*" ��l��� ��-�� � ���� ��� �-�(�� ������� ��� �� � >%:%"@&* ��� ��c��� l��� ��-�� � ���� ��� �-�(�� ������� ��� �� � >)<<"?&*� 6(<(�( 4*� * ����� ���� � �*� ��-� ����� � � ���� �� ��� ���� �������� �� � ���� (���� �������� ��� � ����� ������� � �� ��� �� ����� ��� ( �� 7%<5���� � ����� ��� � �� � ������� ���8" �" �� ���� ���" (���� �������� ��� � �� ������� )�� (�� ���� 7%: ����� �� � ������� ���8" � � �" ��� ��� ��" (���� �������� ��� � �� ������� %: � (�� ���� 7) ����� �� � ������� ���8� ����" ���� ��� ���� �������� � ����� ��� ��� %<" %:" %@" %)" %*" <" :" @" ��� ) ������ ,� �� �� ��� � �" �� ����� � � �� 4 � � ���� ������ � � ������ �� ����� � � ���� ��� �� � �" � ���������" ��� �� ���� � � �� � ����� ��� � � ���� � ������� �� �� ��� ��� �� ������ � ������� g�� �� ����� � � ���� �� ��� (� ����" �� ��e�� ��� �� ��b������ .� ���� � � ��� �� /�� ���� �� ��5 ���� ��(� � � �� � ���� ���� �� ����� 1������� �� %=@ � �� �� �� ���� (� ��� ��(� �� �� �� /�� ���� 7�� ��(�� %" � �� ������� %< � (�� ����8" �� �� �� ��� ��(� �� ����� �� ��� ��� (� ��e�� �� � � �-�� ��� � �� ��� ��(� � �" (� ��� (� �� ������ ����� ��� ����� ��� �-�(��" �� �� �� ��� ��(� �� ����� �� ��� �� �� /�� ����" �� � �-��" �� ������� �� �� � �� ��� �4 ��� �������� �� �� �� ����� �� � � � ��� �� �� 4 ����� (�� �� ��e�� ��� � � �-��� > 2���� �� � �� ������� ��� � �4 � �4����" ����� 4 �� �� ���(�� � ����� � ���� �� �� ���" ��� � �� �� �� �� � � ��(� �� ���� ����� ���� �� � ��� (� � � � � � ��(� �� ���� (� ��� >���" ����� ��� �� �� � ���� �� �������� ��� �" � ���� ����� �� � ��� ������� ��� � � � �������� ��� �� ��� �� �� �� � ����� ������(�� � � � ����� ����� �� �� /�� ���� �� ��� %=@7�4 � ��-8� ��� ���� �� �� ��� ��� �� %=@ � �� �� �������� � �" ����� ��� �-�� ���� ������� (��� ���� g���� � �� ������ � � ������� " �� ( ��� %"*** (����� � �� � � ���� � �� 6 ��� �� � � 7��� � ���� ����� � ����� ��� ��8 � � �� ��� ��� � � ���� � �� 6 ��� �� � � 7��� � ���� ����� � ����� ��� ��8" �� � �� ��� ��� ��� �� ����" � � �� ����� ��� � �� � � ��� ��� � � �� ����� ��� ��� ��(� � �� � �������� ��� � � (� �����!�� ��� � � ����� i����� ��� � � �� ��� � � ��� 6 � ����� ��� ��" � �� ��� �� ��� �-5�������� ������ �-�(�� ����� ��� ��� ������� ��� ,� �� �� ���� � (���� �" ���� � ����� ��� ��" ������ %<" %*" ��� @ ����� ��� ����� �� 7��(�� )8� ����� �� � ���� ��� � ���" (� �����5 �5 �� ��� � � � ��� (� ��� �" �� �� �� ��� �� � � ������ �� � � �� ��� ���� ��� �� ������ ��� ���� 7m � ��-8 ��� �� � � ������ ������� ��� ������ ������ �� �������� � � � ����� ������ ��� � �� �� ��� � � ��� �� ��� � � � ��� ��� '�� ���� ��" �� ������� ������ �� ������� ������ �� �� �� ���� 7m � ��-8 �� %%* !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 � ��� ��� �� ��� ��� � � �� ��� � � � ��� �� � : � ����� ������ ���� �������� � � ���� ������ � �� � � �������� �� �� ��� ��� ��� �� �� �� ��" ����� ���� � � �� ��� � ����� ������� � ��" �� ��� �� �� ��� (� �� ������ � �� ���� ��� � ��� ����� ������������ � � ��(� �� � %**d �b�� � ���� � � �� � � �� %<5���� � ����� ��� �" �� ���� 7m � ��-8 ��� �� ������� ������ �� �� ������(�� � � � ����� ������ �� �� ?&d ������ ��� �� ������� �� �� ������(�� ����� �� ���� ����� �� ������� �� � � � �(���� �������� �� � � �-� � � �� ������� (� ���� ���� ��� ?&d" �� ����� � � �� ���� �� ��� ��(� � �� �� � �� ��� ��� ��-�� � � �� �������� ��(�� ) i����� ��� � � �� ��� � � ������ �� �� � ������(�� � � � ����� � � �3���� �� � ����� ��� �� ��� ���� 7m � ��-8 ���� 74����� ��-8 ���� 7l��� ��-8 ���� 7c��� l��� ��-8 %< ����� 4��� <)@?�%< %%%@%�)* <?))�*? ;;%<�&& :<<6�&) ��i� ?*))�)% &%?6�?% ?*6)�;@ ):6&�<@ )%;;�@@ ��c� *�?::: *�@:%) *�?;%& *�?@6? *�?%:* p% :*?:�6: ;@@@�?* :*:&�?& &;?&�@? &);?�:% 4����� ;:6:�)@ %*)%;�:* ;<)6�<; ;?*6�&? :&;<�@: p? 6<%)�66 %?:@;�<* 66@&�?: 6%:&�;@ <%%:�)< 4������ )<@*�?6 ?%?@�;< )6&6�<? )6%@�%: )<@@�<: 4�-���� )<@<%�?* &???<�:* ?*:?%�)* ):&*@�<* )%)<*�<* � ������ %�@* %�6) %�&* %�@* %�)& k�� ��� @�*@ <�)6 &�%@ @�&* ?�:: %* ����� 4��� )%:<�?& )<:;�** )@:&�%: )?;*�?@% ))??�&& ��i� @&*�?< <6)�&: :?%�<? &;@�%< @6?�;< ��c� *�)*;; *�?%%? *�)&:? *�)@)? *�%?%; p% %<?<�*: ))))�<% )**:�6% %6&?�)@ %<;&�:% 4����� )%)@�?& );&)�%6 )?6:�:* )?%)�*? )%<<�&< p? )@?@�;% ??@%�@< )<%;�<? ):6<�@6 )&)?�&@ 4������ %*<;�)& %%;?�<: %%;6�6: %%<%�%* %%<)�@& 4�-���� @?;?�%* ;?:?�)& &@:<�): &*&:�66 @@<<�<% � ������ *�;) %�*: *�<< *�<@ *�;; k�� ��� %�*% )�** %�@% %�)< %�*6 @ ����� 4��� &6%�<& :<?�:& :@)�6? :?)�:: :%;�?) ��i� :@�&) %?@�;? %*:�%) 66�%@ <<�6* ��c� *�%*6* *�%6;% *�%:&* *�%&:; *�%@@* p% &@:�;* &6%�&6 &;*�&% &:@�6@ &&:�:? 4����� &<:�)* :;*�?& :?@�?* :)@�6? :%*�<: p? :?*�6? ;:&�&@ ;*6�;? :6&�@) :;@�%; 4������ @%?�<: ?6?�*6 @*&�%@ @*<�)6 @%?�*< 4�-���� <?%�%) %)&)�)? %*;@�?& %*?)�;? 6;%�6; � ������ *�@) *�:; *�&; *�&@ *�&* k�� ��� *�)& *�6% *�:< *�:) *�&& ��c� � � �������� � ����� � �� !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %%% ��� ����� � ���� ������ ������� ������ �� ������� �� �� � ��� ������ ������ ������ �� ������� �� �� � � � ���� �� � �� ��� � � ��� �� ��� � � � ��� �� �� �� m � ��-" 4����� ��-" ��� l��� ��(��� � �� ��� ����� ���� ������ ������� ��� ������ �� �������� ��� �� ��� �� �� c��� l��� ��(��� � � � � ����� ��� �� � ���� ��� %@ ������ ' � � ����� ��� �� � %<" %:" %@" %)" %*" ��� < �����" �� ��� ������ ��� �� � ��� �� ���� ������� �� ����� ��� �� (� �� � �������" ��� ���� ��� ��� �� ���" � �� ���" �� �� ���� 1� ���� (� ��� ��" �� ����� ��� � �������(�� � �� � �� ��� ��� �� ���" �� �������� (� �� ��i� 7��i�� � � �� ����� ��� ������ ��� �� ��� �� � � �� �-�� � ��� � � %<5���� � ����� ��� �� �� �� l��� ��(��� � ��� %<5" %:5" ��� %@5���� ���� � �� �� �� c��� l��� ��(��� � ��8 2���� �� ������ � ������ ����� � �� 7��� 8 �� �� �����" � �� ���� ��(�� ��� �� ���� ��� ��� ���� � ��� ����� �� � ��� ��� � ��� � ���� � �������� � �� � �� ��� ��� � ��� ��� � ���� � �� � �� �� � � �� � �� ���� ���� ��� ��� ������� �� ��� �� � �� ��(� �� ����� ������ � � �� ������ � �� ���� � ��� �� ��� ���� � �� ���� ��" �� � �� � � � �� ��� ��(� � � �� � �������(�� ������ ��� �� ���� 1 �� ��� � ������ � ��/���� ���(�� ���� ������ � ���� 5� ������" ��� �� ������ �� ���(��" �� ���� �� �� � ������� .���� �� ��� ��(� � �� � ��� �� ���� ��� ���� � ����" �� ����� ��� � �� ���� � ���� ��� �� ���� f�� ��� (����� �� �� � � ������ �-�� � � 7%<5���� � ����� ��� �" ���� 7c��� l��� ��-88" �� ����� ��� � ���� �� � ���� �� �� ���� ��� ��������� � � �� ����� ��� (� ������ ��� � � �� ��� ��� ��� �� ��� ���� �� �� ���� � ������� 0������ ����� �� ��� �� ���� �� �� ���� ���� ��� � �� ��� ��(� � �� � �� ����� ����� ( � �� ������� ������ �� �� � ������(�� � � � ����� ������ ��� �� ��������� ��� ��� ��(� � �� � ��� �� ���� ���� � �� ������ � �� ��� �� ��� 7�� �� ��8" ������ ��� � �� �� ������ ��� �� � ���� ��� � �� ���� �� �� ���� 7�� �� ����8 ��� �� � ����� ��� ����� 7����� �� � �� �-�� � �" ���� 7c��� l��� ��-8" %<5���� � ����� ��� �8" �� � � ���� �� ������ � �� ��� ��� �� ��� �� ���� � ��� � �� ' � �� �����" � � ���� �� � ������ � � � � � �� � � ��� ���� �� �� � �� ���� ��� �� ��� �� ��� ��� �� ��� � ������ ��� (� ���� ��� ��(�� ? ����� �� ���� � ��� ������ � � � � � �� ����� ���� ������ �� ���� ,� %"*** ����� � ���� � �� 6 �������� � ����� ��� �� �� �� ���" � ��� �� ���� � �� ���(�� � ��(�� ? ��� ���b����� 7 � � %"***8 �� ���� � �� ����� ( ����� ������ �� � �� �� � ������ ��� �� ���" (� � ����� ��� � �� �� ��� ����� �� ��� m � ��4����� ��l��� ��c��� l��� ��%< 6&: &%@ ??? %*@ %: 6&; :&? @6% )&6 %@ 6&; ;?% :*) ?6? %) 6@; ;;% ::; @6: %* 6); ;<% ;)* &:6 < 6*% ;;* ;%6 :*: : <&6 ;&* :6< :%) @ <%) ;)) ;** :?) ) ;%) :@% :%) &&6 %%) !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 ���� �� �� ���� 7� ���� ������8 ( ����� ������ ������ �� ������� �� �� � � �� ��� � �� ��� �� � � , �� b�� � ����� �� � � ���� � � ���� � ����� ��� �� 7� �� ��� ����� ��� � � ��(� ��� ���������8 �� ����� �� � � � � ���� � �� �� � �-� � � ��� ���� (� �� ��� �� ���� ��� m � ������ ( ���� ������ ������ �� �������� ��� 6*d � �� ��� � � ���� � < � � �� ������ ����� ����� � ����� � � ���� ������ �� �� ��� �� �� ������ � ������� ��� � � ������ �� �� ��� ��� ������" � � ���� �� � �� ����� 7 � � ��� ���� (��� ��� �� �� )* �����8� .� ��� ������ � ��� � ��� � ���� � � � ����� �� � �� <*9)* � � 5( �� ��� �� � � � � ��� � ����� ��� ��� ��� ������ ��� � � �� ��� � � ��� ��� �� � � ��� � � �� %<5���� � ����� ��� � ��� ����� �� �� ��(�� @� , ��� (� (������ �� ���� �� � ��� � � ��� ����� � ������ ��� ���� ��� �� � �" �� ���� �� � �� ��� � �� ���� ��� � � �� m � ��(��� � � ,� �� m � ��(��� � " �� ������� �� ���" �� ������ �� ���" �� /�� b��� ��� �� ���" ��� �� ������� �� ��� ��� ��� ������ ��� � � ���(�� ���������� � � � �� ���� ��� ���� �� ���� � � �� 4����� ����� l��� ��(��� � � ��� � ���" � �������" ��� � � �� ���� ����� ���������� ��� ����� � � ��� �� ������� �� �� ��� ���� (���/ � �� �� ������ �(��� � � �� ��� ��� �� ���� ���� �� ������� ,� ���� � � � 5( �� ��� �� � � �� � � � ���� � �� ���� �(��� � � ��� ��� ������ ����� ��� ��� �b����� �� ���n .� ������� � �� � � � � � � � � ( ��� �� � ��� (� ������ � � �� ���� �� � ����� ��� �� �� �� �� ���5 �5 ��� � ���� � � �� ���" �� ��" �� � �� �� ���� ��� �� � � �� � ����� � ��� � ���� �b��� ���� � �������� � � �� ���� ��(�� & ����� � ������� � �� ���� � � ��� 6 � ����� ��� �� � � �� m �" 4�����" l���" ��� c��� l��� �(��� � �� ' � �� m � ��(��� � " � ;@d=):d � � 5( �� ���� ��� �� � � ����� � �� �� ���� ������� �� ��� � � �� %<5���� ��� � �� � �� �� ���� ' � �� 4����� ��(��� � " � 66d= %d � � 5( �� �� � ��� ��� �� � �b��� ����� ������� �� ��� � � �� %<5���� � ����� ��(�� @ i����� ��� � � �� ��� � � �������� �� �� � ������(�� � � � ����� � � ��� �� �� �� %<5���� � ����� ��� �37����� � �� � � � ��� ( ���8 ��� ���� 7m � ��-8 ���� 74����� ��-8 ���� 7l��� ��-8 ���� 7c��� l��� ��-8 4��� <)@?�%< <<?&�<) :<<@�:@ :@&&�*6 &<&:�:; ��i� ?*))�)% ??&6�*& )%)?�6& %<;@�6& %&@?�@; ��c� *�?::: *�?<*) �?*<? *�)6*? *�):?@ p% :*?:�6: :@@*�:* &?@%�@@ &*6*�6) @;?*�?) 4����� ;:6:�)@ <??<�%< ::)?�); :)%6�*; &:<*�&? p? 6<%)�66 %*:*?�@* <*&;�*@ ;&*)�@% :;)<�%< 4������ )<@*�?6 ?%?<�)< )6@?�&< )<6@�%& )<)*�)% 4�-���� )<@<%�?* ????%%�&* )*;*;�<* %6)6<�&* %&%;%�** � ������ %�@* %�6) %�@6 %�@* %�)& k�� ��� @�*@ <�)6 &�%@ @�&* ?�:: .���� �� ��� � �� �� �� ���� ��� �� � � �� �� �� �� ��(�� %" ���� ����� ���� � � �� � ������ ��� ��� �� � � � <*d � � � ��� )*d ( ���" � � ���� �� ����� �� ��(�� ) ����� ���� %**d �� � � �� 1 ��� ��� ��� ��� ������ ����" ��� � ���� � �� � � !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %%? ��� �� ��� c��� l��� ��� l��� ��(��� � � � � �� %<5���� � ����� ��� � ���� �� �� �� ��� ���� �� � %**d � � �� ��� ����� � ��/�� �� ������� /����� � � � ��� ��� ���� ����� � � � �� �� � ��� �(��� � �� ��� ����� ����� �� ���� ��� ( ��� �� ���� ����� ��� �b������� � �� �� �� ��� (� ����� �� ���� ��� �� � � � �� ��� ����� �� � �� ���������� ��� �� ���" �� � ������ ��� (� � �� � �������� ' � ��� ��� ������" �� � ���� � � �� ��� � � �� (� ��� g����� �� �� ��� ���� �� ����� �� ���� � ����� ��� �� �� �� (� � � � �� ���� ��� ����� � ����� �� �� ��(��� @ ��� & ��� ��� (� ���� �� �� � �� ��� ��� ����� � ��� � ���" �� �� ����� � � ��� ���� � %**d �b�� � ������ �� ����� � �� � ������ ��� ��� �� � �� .�� ��(�� & �������� � �� �� �� �� m � ���� � " � ;&d=)&d �b�� �=( �� �� � � ��� � ��� ��� �� ��� ��� �� %<5���� � ����� ��� � 7;@d=):d �b���� � 8� ' � �� m � ��(��� � " � ;*d=?*d �� � � ��� ��� �� ��� ��� %: ����� 7:6d=?%d �b���� � 8" ��� � :*d=@*d �� � � ��� ��� � ��� %) ������ ������� �� � � ��� ����� ��� � ��� ��� � ��� ��� � � � ����� � �� %**d �b�� � ��� �� � �� ��� ��� ��� ����� � ������ ��� (���� � �� ��� ����� �� �� � �� ��� ��� �� �� )* ������ g���� � �� �� � � ������(�� �� ��� � � ?�&" �� � ��� �� �� ���� �� ����� �� �� � �� ��� �� � %6;* %6<6" ���� �� �� � �� ��� ���� ��� ������� � ���� ��� ���� ������� � ������ �� ��� (� � � �� �( � ��� ��� ����� ������ �� ��� �+ ����� ��������� � ���� ������ �� ��5 ��������� � �� � � �� � ��� ��� � � ��� � �� ���� , � ��� ���� �� �� � �� �� ���� �� � � �������(�� � ���� (� ���� �� �� � �� ��� � � � � ��� �� �� � �� ��� � ( ���" �� ����� ���� � ��� �� �� ���� ' � �� ���" �� (���/ � ��� ��(�� ���� ��� ���� (� �����/���� �� ���� ���� �� � ������ ��� ��� ��h�-�(�� ���� ��� �� � �� � ���������� .� ���� ����� �� �� �� ��� ��� ��� ���� ��� �� � � �� ����� � � ������� ���� ���� (� ��( ����� ��� ��� ��� ��� ���� ��� �� � �� �� ��� (� �� ��� �� � �� � ���� ��(�� & ������ ��� � � � �� ���� ��b����� �b��� ��� ���� � �������� � � m��� � � ����� ��� 7�����8 ������ �� � � � m � ��4����� ��l��� ��c��� l��� ��%< ;@ 66 9 9 %: :6 6? %** 9 %@ :@ <@ 6@ 9 %) &< ;6 <; %** %* &? ;@ <) 6< < &* :< ;& 6% : @& :& ;% <& @ ?6 &< :& ;: ) @* :* :; ;: ��� ��������� ����� ��� �� ����� �� �� ( ���� � ����� ��� �� � ��� ��� � ��� ����� � �� �������� %%@ !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 ���/���� �� (� �� � � ����� � ,� ��� " �� �������� ����� �� � ���� ��� �������� ��� � � �������� �� � ��� �(��� � �� 1� �� ������� � ��" ������ ���� ����� ��� (���/ ������ �� ������ �(��� � � � �� ��� � �� �� � ��� �(��� � �� ��� ��� �� � �� ������ �(��� � �" �� ��� ������� ���� ��(� �� ��� � ���� ��� � ���� ������� �� � �b�� ���" ��� �� � �� � � ��(� � �� ������� �� �� ������� ����" � ��� ������ �� ���� �� (���/ �� � ���� ������� ��� �� � � ����� '��� " �� � ���� �� ����� �� ��� � �� �� �� � � � ������� ���" �� ���� �� �� (���/ � ��� ������������ � ������ ��� ���� ��� �� � � �� �� ����� ���� � ����� ���������� �� �� � �� ���" ��� �� � �� ��� � �� �� ��� ��� ��� ���� ����� ������� �� � �� ��� ��" �� ������ ���� �(��� � " �� ���� �� �� (���/ �� ��� �� �� � �� �����" �� ��� �� � �� �� ��� ��� �(��� � � 7�� ���� ��� �� � �� ���8" ��� �� ��� ����� � �� �� ��5 �������� (� � ��� ��� �� � � � ������� ����� .� (������ ( � � ���� ����� � ��� ��� ������������ � ������ ��� ���� ��� �� � � ��� ��� ��b����� (� �� ���� 4��� ��� � �� �� ���� ����� �� � �� �� ������� � ����� ����� � � ����� ��� ���� �� ��� �� �� ���� ���� ��/�� ��� �� � �� ���� � �� �b�� �(�� ��� �� ��� �� �� � �� ,���� �� �� � ��� ��� �� �� � � �� ��� � ��������� ���� ���� � ��� (� ���� ������� � ������ � � �� ��� � � ��(� � �� ������� �� �� ���� � �� �����" � �-�������� ����� ��� � �� �� ��� �b�� � ����������" ���" � �� ��� �� ������� ���� � �� � � � � ��������� ���� ��� �� � � ��� �� � �� �� � ������� � � ��� ���� �" �� ��� ��� � ����5���� � � ���� �������� � ���� ��� �� � �� � ��� �� � ��� �� �� ��� �� � � �� )*** 1������ � '�������� �������� � �������� �� ��� �� .1" �� 0�� �" ��� ���� �� ��� �� �������� � � ���� ��� ��� � ���� � � �� ������� ����� � � �� � ��� 1�� ��� �� ��� ����� �� ��� ���� ��!������� 1�� " $� i�" g�� ��" 1� m�" j ��" �� 7)***8� l � ���� ���� �-�(�� ����� �� (��� ������ �� �� %6<*� ��� %66*�n ! ����� 9 �� �� ?���������" �27@8" <@96?� 1������ ����" g�" � ����" i� ��" j ���� �" .� 7)**%8� ,� /-�� ����� ���� ��� �� � � ������ (� ��n ! ����� 9 �� �� ?���������" �<7@8" );9?<� g�����" .� �� 7%66:8� 1��� ��� �� � � � � � ���� ���� ! ����� ,�������� 9�������" @7@8" &<9:?� g������" l� 7%66;8� � � � �� ��� �� � � ��� �� ����� ��� � ��! �� ! ����� 9 �� �� ?���������" �67@8" &%9&&� g ���" q� 7%66&8� f� �� ��� � � � � �� �� � �� ���� ,�������� ����% � ! �����" :�7?8" <9%?� g ���" q�" j �����" i� g� 7%66;8� ���� ��� ����� ���� 1�����" �� �� ��� ��������� ,�������� ����% � ! �����" :67:8" %?9)?� g���� �" 2� ��" l �" $� m�" j g��( ���" 2� m� 7%6<:8� i� ������� � � � � �� ��� ������� ,�������� ����% � ! �����" ��7@8" ?69@&� g���� �" 2� ��" ������" g� i�" j g��( ���" 2� m� 7%66%8� i� ������� � � � � �� ��� ������� ,�������� ����% � ! �����" �<7?8" @*9@<� !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 %%& g���� �" 2� ��" ������" g� i�" j g��( ���" 2� m� 7%66&8� i� ������� � � � � �� ��� ������ ,,� 1� � �� �� ,�������� ����% � ! �����" :�7%8" %??9%?<� ������" ��" 4�� ��" 2� ��" j .���" i� �� 7%66;8� 1� ���� ��� �� � � �!!��� 4*� �������� �� � ��� .�-��&" a<7%8" %<%9%6%� �� ��" �� 4�" j ��������" 1� 7)**%" 4���� %)8� � ����� ������� ���� � �� � ���� �� ��� b���" %*)9%*:� i����� �" $�" j 4��k��� �" �� 2� 7%66?8� � ������ � ��� �� ������ �� �� � �������� ��� � � � f-� �� o������� � ������ 0� �" 0� ��" j 2��(��" 4� �� 7)***8� ��� �� � ���� � � ���� ��� �� � � ��� ������ � ��� ! ����� ,�������� ��� 3���������-� ����% � " 6:7%8" );9@)� 0������" �� l� 7%66;8� '�������� ������� ��� � ����� ������� ��� ������ � ��� m� �� �� ����� � ����� � ,�������� ���-��� .�-��&" 27%8" %;9)<� l �==����� ������������� �� ,(( � � 1�� ��� ��� �� �� � 7 �� � 7��� ��� ��=��� � � %���7 �( 8������ � ����� �� ������ " ,m� $ ��� ,(( � � ��� 1�� ��� ��� ,(( � �" $� 2�" j k� ���" �� i� 7)***8� i �� ���� ��� �� � � ���� ����� @*" 6*" � %** ����� � ��� ������n ,�������� ����% � ! �����" :27%8" ):9??� ���� �" .� .� 7)***" � ����8� ��� �� � ���� � ���� ��� �� � �� ! ����� c�-� ����" :%9:@� � ���" �� ��" j .��� �" �� .� 7%6668� 1��� ��� �� � � ������ ��3�� ��� �� �� ��� (� ���� � � � ��� ( ���� ! ����� c�-� ����" a7%8" &%9&:� k������ .� ��" j 2�� ��" �� 0� 7%6<*8� ����� ����� � �������� ��� � � � 4����� i� �� ,��� k � �����" �� 1�" j i��� �" 1� 4� 7%66;8� 1 ������� 5(���� � � ��� �� ��� ���� ��� �� � �� ! ����� 9 �� �� ?���������" �67)8" ;@9<&� m�� " �� 2� 7%6668� ��-5����� ���� ����� ���� ! ����� b����* ?���������" �7%8" %9:� 4�� �" g�" j $������� ���" .� 7)***8� ��� �� ����� � � /������� ������ ����� � �� ! ����� ,�������� 9�������" �67&8" %*@9%%?� ���� �" �� l�" g���� ��" �� m�" g����� " i� ��" j ���" k� 7%6:<8� � ������ ������� � ���*��d�� 7 � 6;9668� ��� � � � .����� �,115�$0'� 7%666" 4���� ?*8� �� ���� (� ������ ��� � � �� � ����� ������� � ����� ,������ �� �� ����� c ��" �� 7)***" '�(�����8� � ����� ������� ����� 1 ����� � � ��� � � n ��cc ! �����" %%9%@� .���" �� 7)**%8� ��� &)6 � �� � �� ? ��%" 6 7&8" 6:9%*?� %%: !(!( ����)��� �( ����* + ,�������� ���-��� .�-��& � /� �0 � �1��2 http://http://www.collegesavings.org the fallacy of cookie cutter asset allocation: some evidence from "new york's college savings program" introduction literature review asset allocation taxes time horizons suboptimality of state savings plans methodology concepts assumptions variables the model determining rates of return how the simulation was performed factors which influence the outcomes the variability of returns on stocks, bonds, and money market asset allocation taxes and tax preference the holding period of the investment results conclusion acknowledgements references 62 a study of time value of money educational interventions jason anderson1 and blake gray2 time value of money (tvm) concepts are foundational to learning in personal financial planning and finance courses. this study applied connectivism learning theory (clt) to create educational interventions meant to increase student understanding of tvm using excel and a financial calculator as learning tools. results show students exposed to the learning intervention using a financial calculator performed better than those exposed to excel alone or a combination of excel and a financial calculator. these findings point to the structural entrenchment of the financial calculator in the teaching of tvm, and the challenges an instructor might face when introducing other methods to supplant it. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation anderson, j. & gray, b. (2025). a study of time value of money educational interventions. financial services review, 33(4), pp. 62-74. introduction time value of money (tvm) concepts are foundational for students studying personal financial planning and finance. tvm helps students and individuals concretize present-future tradeoffs, a key element in understanding benefits related to delayed gratification behaviors and more advanced applications in retirement and investment planning. these advanced subjects include topics such as the calculation of education and retirement needs and investment fund selection (and the related opportunity cost of other prospects). students who do not adequately comprehend tvm concepts could struggle with increasingly complex topics in subsequent coursework. as such, many professors invest significant resources and time in teaching tvm within foundational courses in the major sequence. tvm can be taught via several instructional strategies. common methods of teaching include 1 corresponding author (jason.anderson@ku.edu), university of kansas, lawrence, ks, usa. 2 btgray@ksu.edu, kansas state university, manhattan, ks, usa. in-class lectures and videos (created by the professor or attained through platforms like youtube) and the use of tools such as financial calculators and microsoft corporation’s excel spreadsheet software (hereafter referred to simply as excel). another important variable in lesson planning is how long to expose students to tvm topics before moving on to other lessons. the questions underlying these selections are crucial to student comprehension. subsequently, this research probes best practices in teaching tvm concepts, both in terms of duration of exposure and selection of effective tools. literature review tvm has a long history of use in decisionmaking across industries, making it necessary learning for business, finance, and personal financial planning students. shrieves & wachowicz (2001) noted that using discounting https://creativecommons.org/licenses/by-nc/4.0/ anderson & gray 63 methods found in tvm calculations dates to the old babylonian period of 1800 b.c. to cite only a few examples of how tvm is used in industry, lucko (2013) explored the role of singularity functions for the application of tvm in the budgeting process in the construction industry. baruom & patterson (1996) discussed the use of heuristics and net present value (npv) – a finance concept closely related to the time-based value of money in financing decisions – in improving financial calculations beyond the traditional approaches in operations management. finally, weil (1990) explored the relationship between generally accepted accounting principles (gaap) and the discounting of cash flows within the accounting profession. once students gain a solid understanding of basic tvm concepts, they quickly learn how integral tvm is to all areas of finance. a review of an open-source finance textbook, principles of finance (published by openstax), shows that it introduces the most basic, single-payment tvm problems in chapter 7 after teaching students about accrual accounting, financial statements, ratio analysis, and corporate structure and governance (dahlquist & knight, 2024). once this foundational tvm idea is presented, more complex concepts are introduced, including multiple payments, both equal and unequal. this level unlocks a whole new world of financial learning, including npv analysis and other computational methods for capital budgeting decisions, loan amortization schedules, interest rate comparisons, and stock and bond valuations. however, gaining a basic understanding of tvm, before scaffolding into higher levels of financial analysis, is not without its challenges. learning tvm can be difficult due to the potentially challenging mathematical relationships underpinning the calculations (delaney et al., 2016; jalbert, 2002). additionally, problems quickly build in complexity when graduating from single, lump sum calculations to multiperiod payments (i.e., annuities). consequently, different experiential and simplified approaches to teaching tvm have previously been explored, including using graphical approaches and excel. arellano et al. (2012) demonstrated the use of excel to teach students tvm concepts using a retirement model. martinez (2013) assessed the use of graphics (the “time value of money tree”) within excel to simplify the learning of tvm (p. 107). jalbert (2002) developed a flow-chart system to help students answer tvm problems (see also jalbert et al., 2004, for an updated version of this method). jalbert’s process was later modified by newfeld (2012) for use with students with dyslexia. gardner (2004) introduced several pedagogical changes to simplify tvm calculations, including helpful modifications to the variable inputs. finally, dempsey (2003), based on the encouraging results of an experimental study conducted by the author, encouraged educators to teach tvm problems mathematically for maximum educational benefits. theoretical framework pedagogically, it is certainly beneficial to test the efficacy of various technologies in learning tvm. indeed, this study presents evidence of varying efficacy of using a financial calculator or excel in the learning of tvm concepts. however, another pedagogical question remains: to what extent does supplementary practice, trial and error, and extra exposure to tvm concepts assist students in learning these important concepts? does practice truly make perfect? to answer this question, this research uses a relatively new educational theory formulated by siemens (2005) and downes (2008): connectivism learning theory (clt). clt accepts the role of technology in education and seeks to explain its role in the learning process. it also emphasizes the importance of interaction and building connections (i.e. “emergence”) and active learning in the acquisition of knowledge (downes, 2008; siemens, 2005). a key concept of clt is “network learning” where connections are formed from exposure to new content, associations, and learning interactions (downes, 2008; siemens, 2005). clt posits that learning evolves in a similar way to the training process for machine learning, which develops through a tree-like structure of pathways (“if this, then that; if not that, then this” programmatic logic) using nodes and connections (siemens, 2005). crucially, greater exposure to a concept (such as tvm) helps build one’s learning network by creating new nodes that contain new information, financial services review, 33(4) 64 which in turn builds stability into the knowledge structure (siemens, 2005). simplistically speaking, the clt learning process describes the progression of exposure to new concepts through trial and error to aid in the acquisition of knowledge. indeed, downes (2008) uses the term “artificial” – not in the same usage as the term artificial intelligence, per se – but to explain the interpretive nature of learning connections as filtered through neurons and neural clusters. this research is dual-purpose. first, this study applies clt by investigating supplementary, interactive learning experiences and opportunities for trial and error (beyond lecture and existing course content) to increase student learning of tvm. stated differently, does additional exposure to tvm content help build a more stable knowledge structure for students – one that leads to better assessment results? this underpins the clt premises that “repetition is an excellent way of strengthening connections” (i.e., creating nodes) and engagement with active learning is a positive development in the education process (siemens, 2005). second, this study tests the efficacy of student exposure to different technologies (financial calculator and excel) to aid in the learning of tvm concepts. hypotheses this study introduces three hypotheses for testing: • hypothesis 1: students who practice tvm questions using excel will increase the number of questions answered correctly in the second exam (e2) as compared to the baseline exam (e1). • hypothesis 2: students who practice tvm questions using a financial calculator will increase the number of questions answered correctly in the second exam (e2) as compared to the baseline exam (e1). • hypothesis 3: students who practice tvm questions using excel and a financial calculator will outperform all other groups in 3 this quiz tested tvm as well as tax concepts covered in the course. the second exam (e2) as compared to the baseline exam (e1). methodology this experimental study was administered to students enrolled in an introductory financial planning course at a midwestern university during the fall 2023 semester. figure 1 provides an overview of the research design. before interventions from this study, students were exposed to a tvm lecture on september 5 and a 10-question quiz that included 5 tvm questions due september 13 in the canvas learning management system (notated as q1 in figure 1). this lecture and quiz constituted the previously built-in exposures to tvm concepts for this course, in addition to recorded tvm financial calculator videos housed in canvas. the lecture was a live demonstration of working through tvm problems with the calculator shown on the screen. after the typical teacher-student interaction and quiz, students took a longer exam (exam 1) on september 14 that contained 10 tvm questions (see appendix a for the questions contained in the first exam)3. this first exam in the experimental sequence – notated as e1 in figure 1 – was the baseline knowledge assessment for this study in lieu of a control group. afterward, three tvm educational interventions were randomly assigned to students enrolled in the class using qualtrics between october 16 and october 26. the three treatment groups are notated as treatment groups a, b, and c in figure 1. treatments were administered to students after voluntary enrollment in the experiment and the signing of an informed consent form (see ic box in figure 1). the three educational interventions assigned to the respective groups were: 1. group a: one 28-minute video explaining how to complete various tvm questions using excel. the video was recorded using the screen-sharing feature in zoom. the video’s content starts with one of the authors (anderson) introducing himself at the beginning of the video and then proceeding to share his screen to display the 10 tvm anderson & gray 65 questions in a word document, complete with the answers to each question shown at the end of the page. after these questions were introduced, anderson proceeded to share an excel workbook on the main screen. the excel workbook was set up with 10 worksheets (one for each question), with a prebuilt area containing the tvm variables: n for number of periods, pv for present value, fv for future value, i for interest rate, and pmt for payment. the final, eleventh sheet displayed the answers and relevant formulas for each question. anderson then worked through each worksheet by first reading a question from the list of 10 questions, inputting the variables corresponding to n, pv, fv, i, and pmt in the prebuilt area, filling in the necessary calculations to uncover the correct answer, and checking the answer against the answer worksheet. the full intervention can be viewed on youtube at the following link: https://www.youtube.com/watch?v=zjcvf 6yqw6i. as a part of the intervention, students were encouraged not only to watch the video but also to complete the questions on their own using excel. 2. group b: one 28-minute video explaining how to complete various tvm questions using a financial calculator (specifically, the texas instruments ba ii plus). like the excel intervention, the video was recorded using the screen-sharing feature in zoom. the video’s content starts with anderson introducing himself with a digital version of the ba ii plus projected to the main screen before sharing the word document with the 10 tvm questions and answers. anderson then re-shared the calculator on the main screen and explained the relevant keys present on the calculator, including buttons to enter tvm variables, clear work, change from beg and end mode, and make a number negative. anderson proceeded to read each tvm question and input the appropriate variables into the ba ii plus to achieve the correct answer. the full 4 this exam tested tvm as well as insurance concepts covered in the course. intervention can be viewed on youtube at the following link: https://www.youtube.com/watch?v=mqhkbefeko. again, students were encouraged not only to watch the video but also to complete the questions on their own using their financial calculator. 3. group c: this group was assigned the 28minute video explaining how to complete various tvm questions using excel and the 28-minute video explaining how to complete them using a financial calculator (for approximately 1 hour total). these three interventions were designed to increase the students’ exposure (time) to tvm concepts and opportunity to engage in trial and error with tvm problems, in line with clt. this is why duration was allowed to vary across interventions (i.e., ~30 minutes vs. 1 hour). students were allowed to take a no-credit tvm quiz to practice the concepts during the interventions and is notated as q2 in figure 1. the intervention videos continued to be available to students once released by the qualtrics survey. after the completion of the interventions, students took a longer exam (exam 2) on november 9 in canvas to assess the efficacy of the educational interventions. this exam is notated as e2 in figure 1.4 consistency between surveys was upheld by maintaining a common question structure, wording choice, and number of questions (10) while ensuring at least three numerical values were modified for n, pv, fv, i, or pmt. post-testing showed some differences between the intervention groups. group a had the highest ending grades for the semester (94% received as) while groups b and c had 81% and 88% as, respectively. the average scores on e1 were 9.12, 8.25, and 8.78 for groups a, b, and c. the average scores on e2 were 9.06, 8.91, and 8.44 for groups a, b, and c. https://www.youtube.com/watch?v=zjcvf6yqw6i https://www.youtube.com/watch?v=zjcvf6yqw6i https://www.youtube.com/watch?v=-mqhkbefeko https://www.youtube.com/watch?v=-mqhkbefeko financial services review, 33(4) 66 results and analysis of the 108 students enrolled in the course, 97 students participated in the study’s interventions. of the 97 who participated in the interventions, two students were dropped from the study for not completing e2. of the remaining 95 participants, 31 were randomly assigned to group a (excel intervention), 32 to group b (calculator intervention), and 32 to group c (excel and calculator). timestamps for submissions on the qualtrics survey ranged from the 10-day period between the dates of october 16 to october 26, 2023, with one outlier showing december 5. statistics provided by youtube for the intervention videos on the dates of october 16-26 were exported and shown in table 1, whereas post-intervention viewing data is shown in table 2. during the intervention period, the financial calculator tvm youtube video had 74 views and the excel tvm video 67 views. the average watch time (per view) and total watch time for the excel video exceeded that of the financial calculator. figure 1. tvm experimental research design in other words, the average view duration and total watch time statistics show students gained more exposure to the excel intervention as compared to the calculator intervention. the average view duration for both videos was less than the total video length, indicating the videos were not watched for their entire duration. this data does not show, however, the breakdown of exposure between the various intervention groups, as unique viewer identifying information was not available from youtube. anderson & gray 67 table 1. youtube video statistics for october 16 26 views total video length average view duration total watch time (hours) excel 67 28:51 9:41 10.8 calculator 74 28:28 7:50 9.7 table 3 shows descriptive statistics for the total gains (or losses) from e1 to e2 by intervention (n = 95). results show the financial calculator intervention was the most effective in improving scores between exams, with a score increase of .656. table 2. youtube video statistics for october 27 november 9 views total video length average view duration total watch time (hours) excel 10 28:28 0:33 0.1 calculator 5 28:51 2:51 0.2 on the other hand, the excel intervention was neutral while the combined intervention showed an average score drop of -0.344 points. data collected shows that the reinforcement of the learning of calculators led to the most students increasing scores and the lowest decreasing scores. table 3. descriptive statistics for gains by intervention observations mean std. dev. min max higher score lower score excel 31 0.000 1.414 -4 2 10 7 calculator 32 0.656 1.911 -4 5 12 4 excel + calculator 32 -0.344 1.945 -9 3 6 6 a regression analysis of tvm exam score gains (e1 to e2) on each intervention is shown in table 4. the only significant result in this analysis is, unsurprisingly, for the financial calculator intervention independent variable (p < 0.05). results show that, all other things held equal, the financial calculator intervention, as compared to the lack of an intervention, predicts a score increase of 1.55. this regression model accounts for 6.2% of the variance in tvm exam score gain. given the findings outlined in this section, hypothesis 1 is rejected as students did not display an increase in tvm exam scores after the excel educational intervention. there is support for hypothesis 2, as students demonstrated a significant increase in exam scores due to the financial calculator intervention. hypothesis 3 was rejected, as students did not show an increase in exam scores after being exposed to both interventions. financial services review, 33(4) 68 discussion and limitations teaching tvm to future financial planners is a complex task. this study presented limited evidence of the efficacy of using a financial calculator to teach tvm concepts to undergraduate students in personal financial planning. this finding is limited, however, by the fact that all students were taught how to use a financial calculator within the course before this study’s interventions, and subsequent learning was on that foundation. table 4. regression of intervention on score gain coeff se 95% confidence interval excel 0.889 0.732 -0.563 2.340 calculator 1.545 * 0.729 0.0988 2.991 excel + calculator 0.545 0.729 -0.901 1.991 constant -0.889 0.644 -2.167 0.389 model fit statistics r-squared 0.062 adjusted r-squared 0.034 * significant at p < 0.05 they were also offered support and resources geared toward using the financial calculator instead of excel, including a lecture and recorded videos. past sections of this course only used the financial calculator, so assigned course peer mentors would have been able to assist with the calculator, but not excel. the intervention groups were not pre-tested for more accurate sorting. this resulted in an interesting finding which confirmed the usefulness of continuing historical methods, i.e., using a financial calculator to teach tvm. two factors, scores correlated with higher overall grades and improvements being ascribed to lower initial scores, were counteracted by the slightly unbalanced populations. it could be assumed that the group with the highest grades would perform the best. instead, the calculator-only group had the lowest overall grades in the class and yet improved the most from e1 to e2. further, the calculator-only group erased its gap from e1 to e2 and performed similarly on e2 compared to the group with the highest grades (the excel-only group). to counteract improvements due to reversion to the means, the two groups had similar e2 scores, but only one group had a statistically significant improvement between e1 and e2. it is certainly possible that students gravitated to the financial calculator – and performed well using this tool – simply due to its structural entrenchment in the teaching of tvm. after all, the syllabus encourages students to buy a financial calculator but mentions nothing of excel. even outside of the historical use of the financial calculator to teach tvm, a brief survey of online resources shows financial calculator videos and study materials are more readily available for students studying tvm. prerecorded lectures from previous years are also available and solely demonstrate effective use of a financial calculator. efforts from this study and others trying to introduce new tools (i.e. excel) and methods to teach tvm face significant headwinds going against the traditional teaching methods. given that software such as excel can be intimidating to master, and financial calculator tools are readily available, it is perhaps difficult to have students transition to a new way of dealing with the problems. in other words, our findings might be a nod to entrenched structural support more so than a win for a specific anderson & gray 69 educational intervention, especially considering the youtube average view duration and total watch times for the excel intervention were greater than for the financial calculator. it may simply be more effective to continue teaching tvm using whatever method with which students are most familiar. theoretically, processing information using different methods would expand knowledge and understanding. this may be true once the individuals have somewhat mastered one method before being introduced to another. data collected from youtube showed limited student engagement with the interventions, given that the average viewing time for videos was significantly lower than their full length. in other words, even though the study participants assented to the qualtrics question “i have watched the video and followed along with the exercises,” youtube analytics showed this was not the case for all students. this might help explain the perplexing contrast between score gains for the financial calculator, yet a potential drop when exposed to both the excel and financial calculator interventions. the drop in score might be explained by the longer intervention time (i.e. watching two videos for 1 hour versus one for 30 minutes) or because students became overwhelmed by learning two different methods for tvm calculations. the fact that there was limited interaction with the videos postintervention could indicate that students felt fully prepared for the upcoming examination or, perhaps, they found little value in continuing to view them after multiple exposures to tvm concepts within the course (see table 2). the practice videos provided examples of 10 calculations each, and students may have skipped solutions to check understanding and stopped after a few were understood. the similar question structure between the quizzes and exams could have played a part, too, in less-than-full engagement from students. alternatively, students could have simply skipped to the parts of the video they found most helpful. this study was not structured to answer these questions, although a future study could do so through design improvements. to expand on potential improvements to the study design, the authors offer the following suggestions. first, youtube analytics for video interventions and responses collected in qualtrics for the educational intervention could not be merged at the individual response level, as no unique viewer identifying information was available from youtube. this precluded the ability to connect watch times to exam performance or highlight potential outliers based on engagement, especially considering student engagement with the videos post-intervention was very limited. future research could connect these two data sources using an alternative video delivery method. second, a data collection of student capabilities and experiences interacting with the various interventions might better explain differences in student scores and engagement. third, future research might go beyond tvm questions that require simplistic manipulation of one missing variable to attempt a better assessment of a more holistic understanding of tvm. in other words, do educational interventions – with the tools presented here or others – help students apply tvm concepts to case studies, client problem solving, and real-world scenarios? clt would bolster such a distinction, as both real-life experience and exposure to information are ways to create and strengthen new learning nodes yet remain distinctive paths within the theoretical framework (siemens, 2005). if one tool promotes the learning of tvm concepts at a higher level, its use might be justified beyond a simple increase in assessment scores. conclusion what was learned from this study that can benefit the teaching of tvm concepts? first, this study expounded on the efficacy of using a financial calculator to teach tvm concepts. although many methods exist for teaching tvm, innovative teachers might first acknowledge a potentially uncomfortable reality: the entrenched method of using a financial calculator to teach tvm might be adequate, even preferred. when most resources support one way of teaching, adopting the supported approach might prove to be the most efficient path forward (even if learning a new method brings unique advantages). in fact, this conundrum is analogous to financial professionals who try to introduce a new software, method, or process into their firm: financial services review, 33(4) 70 there can be significant headwinds. when the wind gets too strong, swimming with the current is recommended. a second point of learning is to acknowledge another observation: students are savvy at building their own learning journey. this study’s comparison of video and intervention data shows students tailor learning to fit their specific needs – even if it means dropping a potentially helpful video once it has served its purpose. however, as long as the course’s learning outcome is achieved, this might be perceived as a positive, rather than a negative, outcome. finally, the results of this study – and specifically the finding that a longer intervention (excel and a financial calculator) was less effective than a shorter one (the financial calculator alone) – encourage teachers to adopt a streamlined approach (i.e., “keep it simple”) to teaching tvm. references arellano, f., mulig, l., & rhame, s. (2012). teaching time value of money using an excel retirement model. american journal of business education, 5(6), 663-676. https://doi.org/10.19030/ajbe.v5i6.7389 baroum, s. m., & patterson, j. h. (1996). the development of cash flow weight procedures for maximizing the net present value of a project. journal of operations management, 14(3), 209– 227. https://doi.org/10.1016/02726963(96)00005-8 dahlquist, j. r., & knight, r. (2024). principles of finance. openstax. https://openstax.org/details/books/princi ples-finance delaney, c. j., rich, s. p., & rose, j. t. (2016). a paradox within the time value of money: a critical thinking exercise for finance students. american journal of business education (ajbe), 9(2), 83-86. https://doi.org/10.19030/ajbe.v9i2.9638 dempsey, s. j. (2003). on the benefits of a mathematical solutions approach to time value of money instruction: arguments and evidence. journal of accounting education, 21(3), 239–260. https://doi.org/10.1016/s07485751(03)00023-x downes, s. (2008). an introduction to connective knowledge. in t. hug (ed.), media, knowledge & education: exploring new spaces, relations and dynamics in digital media ecologies, (pp. 77-102). innsbruck university press. gardner, n. d. (2004). the time value of money: a clarifying and simplifying approach. journal of college teaching & learning, 1(7), 25-30 https://doi.org/10.19030/tlc.v1i7.1965 jalbert, t. (2002). a new method for teaching the time value of money. journal of the american academy of business, 2(1), 72–79. jalbert, t., jalbert, m., & chan, w. y. c. (2004). advances in teaching the time value of money. journal of college teaching and learning, 8(1), 7–12. lucko, g. (2013). supporting financial decisionmaking based on time value of money with singularity functions in cash flow models. construction management and economics, 31(3), 238–253. https://doi.org/10.1080/01446193.2012. 750728 martinez, v. (2013). time value of money made simple: a graphic teaching method. journal of financial education, 39(1/2), 96–117. newfeld, d. (2012). teaching time value of money to dyslexic students: a pilot case study. journal of instructional pedagogies, 8. shrieves, r. e., & wachowicz, j. m. (2001). free cash flow (fcf), economic value added (evatm), and net present value (npv): a reconciliation of variations of discounted-cash-flow (dcf) valuation. the engineering economist, 46(1), 33– 52. https://doi.org/10.1080/0013791010896 7561 siemens, g. (2005). connectivism: learning as https://doi.org/10.19030/ajbe.v5i6.7389 https://doi.org/10.1016/0272-6963(96)00005-8 https://doi.org/10.1016/0272-6963(96)00005-8 https://openstax.org/details/books/principles-finance https://openstax.org/details/books/principles-finance https://doi.org/10.19030/ajbe.v9i2.9638 https://doi.org/10.1016/s0748-5751(03)00023-x https://doi.org/10.1016/s0748-5751(03)00023-x https://doi.org/10.19030/tlc.v1i7.1965 https://doi.org/10.1080/01446193.2012.750728 https://doi.org/10.1080/01446193.2012.750728 https://doi.org/10.1080/00137910108967561 https://doi.org/10.1080/00137910108967561 anderson & gray 71 network-creation. http://masters.donntu.ru/2010/fknt/lozov oi/library/article4.htm weil, r. l. (1990). role of the time value of money in financial reporting. accounting horizons, 4(4), 47. http://masters.donntu.ru/2010/fknt/lozovoi/library/article4.htm http://masters.donntu.ru/2010/fknt/lozovoi/library/article4.htm financial services review, 33(4) 72 appendix a – tvm quiz questions exam 1 (e1) 1. how much must you invest today at 8% interest in order to see your investment grow to $15,000 in 10 years (round the answer to the nearest dollar)? a. $6,330 b. $6,948 (correct answer) c. $7,500 d. $7,620 2. at what annual rate would $500 grow to $1,948 in 12 years? a. 10.0% b. 11.0% c. 12.0% (correct answer) d. 12.5% 3. if lucky louie won a lottery and chose to take $1,000,000 in cash (disregarding taxes), how much money would he have in 30 years if he invested at 6% per annum (round the answer to the nearest dollar)? a. $5,743,491 (correct answer) b. $5,804,268 c. $6,050,972 d. $7,890,345 4. sally contributes $3,000 each year to her roth ira (a type of retirement savings account). what is the value of her account after 5 years if her annual return is 4% per year (round the answer to the nearest dollar)? a. $3,650 b. $8,453 c. $12,931 d. $16,249 (correct answer) 5. dan can earn an 8% annual interest rate. if he invests $20,165 today, how many years does he have to wait to accumulate 32,000 (round to the nearest whole year)? a. 2 years b. 4 years c. 6 years (correct answer) d. 8 years 6. a client intends to place a $2,500 lump sum amount into a savings account that earns 5% interest, compounded monthly. if the lump sum is left to accumulate, how much will be in the account in the future, at the end of 4 years (round the answer to the nearest dollar)? a. $2,057 b. $2,542 c. $3,039 d. $3,052 (correct answer) 7. judy would like to have $200,000 saved in her retirement account in 20 years. at an interest rate of 7 percent, how much should she contribute each and every year (round the answer to the nearest dollar)? a. $3,492 b. $3,569 c. $5,684 anderson & gray 73 d. $4,879 (correct answer) 8. every year, mark puts $500 in a bank account that earns an annual rate of return of 6% at the beginning of the interest rate cycle. how much will you have in that bank account after 6 years? (round the answer to the nearest dollar and use begin mode) a. -$3,697 b. -$3,488 c. $3,488 d. $3,697 (correct answer) 9. billy wants to attend mcpherson college when he graduates high school in 7 years. current annual tuition is $19,000. if tuition rates are rising at 6%, what will be the annual tuition cost when billy enters college in 7 years (round the answer to the nearest dollar)? a. $12,636 b. $23,825 c. $28,514 d. $28,569 (correct answer) 10. if jim wants $25,000 in 5 years and can earn an 8.33% interest rate that is compounded quarterly, how much does he need to invest today (round the answer to the nearest dollar)? a. $16,554 (correct answer) b. $16,980 c. $20,410 d. $37,755 exam 2 (e2) 1. how much must you invest today at 15% interest in order to see your investment grow to $20,000 in 15 years (round the answer to the nearest dollar)? a. $1,568 b. $1,786 c. $2,458 (correct answer) d. $2,594 2. at what annual rate would $1,500 grow to $2,221 in 10 years? a. 3.0% b. 3.5% c. 4.0% (correct answer) d. 4.5% 3. if lucky louie won a lottery and chose to take $500,000 in cash (disregarding taxes), how much money would he have in 15 years if he invested at 8% per annum (round the answer to the nearest dollar)? a. $1,586,085 (correct answer) b. $1,698,345 c. $2,467,711 d. $2,679,345 4. sally contributes $5,000 each year to her roth ira (a type of retirement savings account). what is the value of her account after 10 years if her annual return is 8% per year (round the answer to the nearest dollar)? a. $72,433 (correct answer) b. $75,699 financial services review, 33(4) 74 c. $76,606 d. $77,556 5. dan can earn a 4% annual interest rate. if he invests $29,230 today, how many years does he have to wait to accumulate 40,000 (round to the nearest whole year)? a. 2 years b. 4 years c. 6 years d. 8 years (correct answer) 6. a client intends to place a $4,000 lump sum amount into a savings account that earns 6% interest, compounded monthly. if the lump sum is left to accumulate, how much will be in the account in the future, at the end of 5 years (round the answer to the nearest dollar)? a. $3,529 b. $4,557 c. $5,395 (correct answer) d. $6,722 7. judy would like to have $400,000 saved in her retirement account in 15 years. at an interest rate of 9 percent, how much should she contribute each and every year (round the answer to the nearest dollar)? a. $13,624 (correct answer) b. $13,331 c. $14,444 d. $14,561 8. every year, mark puts $1,000 in a bank account that earns an annual rate of return of 8% at the beginning of the interest rate cycle. how much will you have in that bank account after 8 years? (round the answer to the nearest dollar and use begin mode) a. -$11,488 b. -$10,311 c. $10,311 d. $11,488 (correct answer) 9. billy wants to attend mcpherson college when he graduates high school in 5 years. current annual tuition is $12,000. if tuition rates are rising at 5%, what will be the annual tuition cost when billy enters college in 5 years (round the answer to the nearest dollar)? a. $14,101 b. $14,387 c. $15,211 d. $15,315 (correct answer) 10. if jim wants $22,000 in 7 years and can earn a 7.22% interest rate that is compounded quarterly, how much does he need to invest today (round the answer to the nearest dollar)? a. $13,332 (correct answer) b. $15,554 c. $16,707 d. $17,605 pii: s1057-0810(02)00091-4 � ����� ������ � ����� � � ��� ������� ����� ��� ������ �� �� ���������� ��� ��� �� ��� �� ���������� ����� �� � � � ��������� ���� ��� ���� ����������� ����������� �� ������ �� ��� !�� �" ����� #$$�% ���� !�� ��! ��� &��� #$ ��'������ #$$�% ����'��� �" (������� #$$� �������� ) �� � '�'��� *� '���� +��� ��� ��� ������� ����� ��� ����� � � ���� *������ ������ � +��� &��� ����� ����� ����*� !������ �� ���' �� � '���&�� � �� � ��� � * �� ��� � � �� ������ ��� ,� �� +�� ���� ��� � �� �& ����� ����� � �� � ��� � * �� ��� � �� ��'����� � � ' �� ���� �� ��� ������ � ��� ��� � �� �� ��� ''�� ��&� ����� ��� ��*�� � �� ��� ����� ��� � �� �� ��� ��*�� � ��� ������ ,�� *� ���� &�� � ��-��. ���� ��� � � *� �� �� + � ���� !��� � � �� � �� ����� ��� ��� ������� ��� � �� / � ��� ��� � ����� � ���� ��� � � ��� '��� ��� �� !��� � ��*�� � �� !�� � ������ � #$$� ����! �� �� � �� ) �� ��� � ���� �����!��� !�" ����� � ��� �# ��� $��% �&�# 0�� �% ��� �����% ����&�� �% ������% ) !����� � � ���� ������ ,����� � !����� �� ��1 ��� � ������� *� � � ! � ��� �� ��� ���� �� ��� ��� � '���&�� ��� �&��� ���� ����� '��&��� � ��� *��� & �� � �� ��� +��� � � ��� �� ��!� ��' 2 �� �� &����� � � � ����� �� ��� � � �'����� ( ��� � � � � � � ������� ���� !������ ���� �� �� ��� *��� �������!�� � � ��!� � ����� ���� �& �� ��� �&��� ���� ������ � ���� ���'� �� �� � � �� & �� � �� ������ ���� ����� �� � 33�� � �� ����&�� '��'� � �� ��!����� ' �� � �� �� �� ��� & �� � � ���� � � � ) �� � '�'��� *� !��� ���� ��� ! �� � �� �& � � �''����� �� ������ � � ! � �� ��� � � ��� 4�� ��.�� �������� 2 ��� � �& �� � '�'�� � *������ � ! � �� !������ �� ������ � �2 �� '���&�� � ���� ���� �� * �� � ��� ����� !� ������ �& � �� ���� ���� '��&�� � � ���� ���� ��� � ���� � ����'����� � ��-��. ���� ��� � � / � �''����� ��� ��� ���� �� �!� ����� � � ��� �� ������ ��� ���� '��! ��� �� ��� ������# ��� ����� � *��� � �* � � � �� ���! ��� �! �* �$ 5#$$�6 �#78�99 �$:;-$<�$=$�=> 8 ��� &�� � ������ � #$$� ����! �� �� � �� ) �� ��� � ���� �����!��� �))? � � $ : ; $ < � $ 5 $ # 6 $ $ $ 7 � 9 � @�����'� � � � ����� 4���? ��-#:#-"#<-a7<;% &��? ��-#:#-"#<-9$7"� �'���� �&&������# ��������&b�� ���� ��� 5���� ��������6� ��� ��b�� ���� ��� 5���� ��� �6� � 4���? ��-#:#-"#<-aa;$% &��? ��-#:#-"#<-9$7"� #��� ����� ����� �������� �� ���'?==����� ���� ���������=������������������1� �!�������������� � � ���'����� &�� ��� � � ���� � � � ��'��� � �& !����� � ���'� � &����� � � 4� � � � �*������ � � � �� ���� ��� �� � �� ��� ��& � ��� �& ��� ������� � ! � �� ���'� � � ���� � � � ��'��� �� ,������ � !����� � �������� ' ����� � � ! � �� ������ �� ������ � �2 �� !����� � ���'� �� &�� ��� � '���&�� �� ���� ��� �� ��!� � ����� ���� ������ � � �& ��� � !����� � ��.��� !�� /& '��� � ��� �'���� �� � ���� ���� � �� � �* ��� � ������ �� &�� � ��� ��� � ������� ��!��� �� ��!� ��� ��� �� � !������ �& ��� � � � ��� � � � ��* � �� � �� ���� �� ��������� ��!������ �& ��� !��� � '����� � ���� ��� �� �-� ��� �2 �� ������� ��� � ��� ���� � ��� ��� �����-���� !���� � �� � ������ 4�� '������ * �� �� � '� ����'�� � ���� �� � !������ �� �� 33��������� ��� �����-� ������ �& *����� ���� * �� �! ����� �����'� � !����� � ��� �2 �� �������� 4� � ��� �����-� !���� � �� � �� ���� � �� ��� � ���� � ! � ��� ��� ���� �������� ��! � ��� �����-� �� � �'� � � ��! �� � � �� � ��� ���� ���� �� �� ���� � ���� �� � � � � �� ��� ����� ,� ���� �� ��� �� � �� �� ������ � ! � �� ����� ������ ���� ��� �� � ��� � * �� ��� !������� � �� ������ ��� 4�� ������ � '������ � ����� ��� � � �� �� '��&����� �� � !��� ��* ����� , �� ������� ) ��� �� ���� ����� � � �!� ��� � ��� ��� ���� � � �� �� ���'����� ���� �� � � � � ��������� ���! ��� �� � ��� � � ��� ���� � � ����� �� �� ������ ���*�� >�$ � � >"$ '�� ������ 4� � ����� ���� �� '������� �� �� � ����� � ��� � ��� &����� ��� !����� � ��� � � � ���� � # � �� ���� ��� ����!� �� � � �'���� �� �� ��� !����� �& � � �������� ������� ��� � �� �����* � ���� � #� *� '��! �� � �� �& ��! �* �& ����!� � � ����� �� � ��� ��' � �& ��� � �� ������ � � ���� � 9 '���� ��� ��� ����������� �& ��� '�'�� &����*�� ���� � : �� ��� � �� �� � �& ��� ��� 4�� �� �� � � � � ���� c�� ���� � a� � ��� ��� ��� ���� �� ������� ,������ � !����� � �� � ��� � �* �2 �� ��� � �� ������ � �� � � ��� ��'��� �� �& � ��! ��� ����� � � ��� �� �� �� +�� � ���� ��� !����� * �� �* ��� � � �� ���� ��� �� � ��� � * �� ��� � � �� ������ ��� / � �& ��� +��� ���'� ��� !����� � '������ � �� ���� �� ��� �� � �� �& � !����� �� �� ��� ��� ' � � ��* ��!��� �� ����� '� ���� 4� � ���' � �&�� �� ���� ���� �� ��� ��� ���� � �& � � �� 2 ��� � � ��� 0 �� ���� ��� !����� � ��! ���� +��� ��!� '��'� ����� ��� �� �� ���� �����'� ���� � �� � � �� �������� �� )& �� � �������� � � �� � ��'� �� � �& � !����� � ��! ��� ����� ��� *�� � ��� ���� �&&�� 2 ��� � � ��� � � �!�� ��� ��� ��� ��� / � ����'�� �& � ���� � � � � �� ������ �� 2 ��� � � �� �� �� &� � �� ��� @de@ *�� � �� 5���'?==***��� ���� ������� ����= !�����=�����= f� ��2=�� ���'6� / �� � �� ��� ��� ������ �� ��� ��� � � ����� � ��� &���� * �� ' �� � ��� � � �� ���� ��� � * �� ��� � �� '��&��� ��� �& ��� !������ 4� � ������ � '������ �� �� ������� � �'� +�� & � ��� ���� ����� �� �� �� ��� �� �� � &� �2 �� ��� � � �� ���� ����� ��� !������� � �� '��&��� ��� 4�� '�'�� ���� �� � � � �� ��� ���� c�� ���� 2 �� �� � !��� � ������ ���� �� � &� ��� � � �� * �� �'�� +�� ��������� �� ��� � �� �� !�� � �� ���*��� ��!� �� '��� �& ��� ��� � � '������ � � �.��� !� ����� �� e� ��+ � � � � �.��� !� �"$ �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 http://www.moneycentral.msn.com/investor/calcs/n_riskq/main.asp http://www.moneycen-tral.msn.com/investor/calcs/n_riskq/main.asp &������ ��� � &+� ��� �� ����� �� ��� �&�� �� ����� � ! � ���� �!� & � ����� � ����� �� �� ��.��� !��� �''� ��� ��� �� ��� � �&�� �� � �� �& � �� '�� !�� ����� ���� ���� �''� ��� � �& ��� ����� � &+� �� � � �&�� �'���� ���� ��� ����'��� � ��� � 5�77<6 '��! ��� � ���' ��� c�� ����� � ����� ���� �&&��� ��� '��� � �� &�� �� ����� ��� � �� � � �� �''�� ��� ����� ��2 ��� � ������ � � �& � �� � �� � � �� ����� �� d� -'��&��� � �� !������ ��� � ��'� �� � ��� � � � ��� �� ��!� � !����� � ����� ���� �� �� ��� �� �� � �� * �� ��� � '���� �� !����� � '��&��� ���� � �'� &� � ��� !����� � '������ �� ���� � ! � �� !������ �� ���� ���� � �� ����� ������ � � ���� �� ���� ��� '������ �� ��� ���� c�� � � ���� ���� �� �''��� � �.��� !� ������ � *� �� !��� ���� � !��� ��� � !������ � � �� � ����� ��� �� ���� ��� �'� ��� '���&�� ��� ,� ���� � �� � '�'�� *������ ��� �� �� � ��� ����� ����� ������ ' �� ���� �� ��� ����� @��'���� � � �� �� ��� �� �����'� �� ��� ������ � ��.��� !�� ���� ���� �� � ��� ����� ����� � � ��� � �� ��� '��! ��� ���� � 9� � ���������� �� �������� ��������� ����� �� � � �������� � �� � �� �� ��� ��� � &�� �� ��� �� �&+� � � �������� � � ��& � �� ���� �� ���� ���� � ����� �& �������� ������� �� ��� � �� ������ � � � '��&���� ��� / � ' �'��� �& �� � �� �& ��! �* � �� ���� � ����� ���� ��� ������ � ��� � ����� � �� �� ����� �� ��� � ��� �� �+�� �& � � '��� � ��� �''����� �� ������ � �& � ! � �� ��� � � �� &�� � !������� '���&�� �� � ���� � ����� &�� ��! �* � ���� �&+� � �-������ � ����� �� � �� ��� ��� ���� �� �& ����� � ��� � �� �& ��� '���&�� � * �� ���� �& ��� !������ ���� � �������� � ��� ������ � �& � ! � �� ��� � � �� �� �� '����� �� � !������� '���&�� � � '��&����� '� ��� �� � �����'� �� ���� ��� �&+� � �-������ ��'����� �� � ���� � ����'�� �& �� � � �� /��� � d���� � , ��� � � ������ 5�77<6� 4��� �� ��� ��� � � ����� !� ���� ��� ��� �� �& 4����� �� �������� ����� �� � ��� � �& �� � &� � +��� * �� ���!� ����� ��'����� ��� � �� 4��� + � ���� � ''��� &�� � �����& � �� �& ����� ���� ���� 4�� � �''����� � ����'����� �� � !����� * �� � �� �������� ��� � �� �& *��� ��� 2 ��� � ���� ���� � � ������� 4� � ����� � � � ���� !���� �� � � �� �� �� �� � �� �� �& ���� ����' &�� � !����� * �� � ��* ������ �& � �� ������ ��� 0�� � g � ��� ��� � &��2 � � ������ &�� !����� � ��������� ���� ��� ����� �������� �� ��� '���� � ����� � � ''���� ��� �� �� � �� �� &�� ��� �� � �� �& 0�� � g � �� �� � &� ������ ���� * �� '��&��� ������ ��� ����� ��� � � �� � � � ��-��. ���� ��� � �!�� ��� � ���2 � � �# �� ���� � �� �& ��� ��� �� �������� �� ����� � ��� � ��� ��� �� � �� '��! ��� �� 0�� � g �� ��� ���� !��� ������ ��!� ������� �� ����� ����� �& ������� ��� � �� 0�� � g �� ��*� � �* �c � � ������ 5�77�6 ���������� +��� �� ��� 5�� 6���� "��� ��� ��� ������� �� ������ 50g���6� 4��� + � � � +�� � '� �� ���'� ��� �� ��� �� � � �& � +�� ��� 0g��� ' �� ��� � � � � ��� ���'� �� � ��� � ����� @�� ��� ���!�� � � � ��� ��� 5�77"6 ���� ���� �� � �'�� �� ���'�� �& 0�� � g � ������� ��� +���� ���� *��� 0�� � g � '��+��� �*� +��� &��� ��� � � �� �& +��� �� ��� ���� � &�� � ��� ���� �� * �� ��� 0g��� +���� ��� '� �� ���'� �� � ��� ��'����� c�� � ���� � '� ��� ' �� ��� � �& ��� ��'��� � � ��!����� ����& �!�� ��� ��� ��!���� *����� ������� � � ������� 5�77:6 ���� �� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �"� +��� �� ��� ��� 33����� h ��� ������ ���� � �& 5�� 6���� "��� 4��������� ������� 4��� + � � � � +�� � � � '���� � � '� �� ���'� �� �� �� � �� ��� �� � &��� ��� ����� e����� � � /��� 5#$$$6 '��! �� � 2 � � ��� �� ��� '��&���� �� �& � ! � �� !������ *�� '� ��� �� �� ��� ��� � �* '���&�� ��� 4�� � �� �� � � ��� � � ! � �� !����� ����! �� � � � ��� �� ���� &� � �� @������ 5�77;6 ��� � '��&��� � �� '���&�� � �� ������ e��� ��'�� �& !������ � &&�� &��� ��� � ��� �� �& ������ !� ������ 4��� &� � ���� ��� �� ���� �� � �� ��� � '�� �� '��&���� �� �& � ! � �� !����� '���&�� �� *�� � ������ ����� ���� * �� ��� &��2 � �� � � ���� �& ���� �� / � ���� �& �������� ���� �� ��� 2 ��� � � ��� � ��� �&+� � �� �& �2 �� ������� � ��� � �� �� � ���*�� !�� � !��� � ���*�� +���� ���� � � ��� �� 5�77#6 � � g��� ����� ���� &��� � � 0 �� � 5�7796 ���� '��! �� � ''��� &�� ��� � '�� �� � ��-��. ���� '��&���� �� �& +��� ��! � ��������� �� �� �� � &� � ���� �� !�� � +���� 4�� !�� � !��� � ���*�� � �� �� � � ��!� ���� ������ &��2 � � ��' �� &�� ��� '�' ��� '���� 5� ����� �77a% ������� �� �77a% @���� i ���!��� �77<6� /!�� ��� ����� ����� ��!� ��� � !�� ��� �& �� �� � � � ������ � ��� �� � �� �& & � �� ����� �� '��&��� ������ ��� � � �-� �-���� ��������� (���!� � � /��� 5�77<6 �� �� ��� ����� � �� ' ���*�� & � '��&���� �� � � & � &���� 4��� + � ���� ��� ����� �� � � ���� ��� &�!������ &�� ��� !������ h �� � � �!�� �� ��� !� �� ����� � �& & �� ���� �� ������� �� ���� �� �'��!�� � ��-��. ���� ��� � �� 4� � '��&���� �� ������ �� � �� �� �� � ��� 5,����� i h��+���� �77a% ������ i 4� &��� �77<6� h�*�!��� ��� ��.�� �� �& ��� �������� �� �� ��� ���� !��� &�* !������ * �� �� ���� �� �� ������ ��� ��� ������ �� !��� � ���� �� ��� !��� �� ���� & ��� ��� ����� ��� ��� ����� c�� &�� �� � ����� &�� �� � ���� � � � ����� '� !� ���� '��! ��� � !����� � ���'� ��� ����� �� ��� ����� ���� �� ��� � ���� ��� � & � ���� �&&��� '��� �� !� &����� � � ����� � �! �� �� &��� �������� � � � �� & �� � ��� � ���� ��� � &����� � ���� ��� '���� !�� !�� �� j���� � � � d��� � 5�77<6 + � ���� ��� ��� ����� :-���� ��� � ������ &�� � � �� & �� '��! ��� �! �� �� ���� �� � ������ ��� � �� '��&���� �� '��� ��� �� �!�� � "$-�� �� '�� ��� g�! ���� � � k���� 5#$$$6 �� � �� �������� ���� � � * �� ��� �� � �& � � ��������� 4��� �� ��� ������� :-���� ������ �� �� � &� ��� ������ '��&��� � & ��� ��� ���� ������ ������ &��� ��� '���&�� �� �& ����� & �� &�� ����� � �& � ��'����� '���&�� �� 4��� ���� *������ � ! � �� !������ �� ��� �� ����� ��� � � * �� � '���&�� � �& ��� �������� ��� � � ��� 4�� � ��� ��� + � ���� �� � ''��� &�� ��� ��.��� � �& ��� � �� ��'����� �� � � ���� �� �� �� 33���� ��� �! �� �� � �� ���� � � � �� �� ������� � ��� ����� ������ � �������� �� ��� � ! � �� !�������� 4�� ���� ���'���� � !� �� �� �& !����� � ������ ��� �� �� � ,��� ������ '��&��� � ��� *�� ���'����� �� /���� �� ���� 5�77;� �77<6� h � ���� �'���� � �� �� � � � �� ������ � ��������� � � �� �''� �� !����� ��� � ���� � � �� '� �� �� ����� �''����� �� !��� � � � !�� � � � ���*�� ������ ����� !������ �� � �'��&��� ��� ������ �� � � +�� � ���� � � � � ��-��. ���� ��� �� � ���� � '�' ��� '���� ���� �� (���� 5�77<6 ���� ��������� ��� �� � �& ��� !�� � !��� � ���*�� !��� � � � ��� �� ��� �� �& �&+� � � �������� h � �� �� � � � ��� � � ��� �� /���� �� ���� 5�77;� �77<6� e��� � ����� � ����� ���� ��� ��.�� ����� &�� ��� � ����� �& ��� � '���&�� �� � ��� ���� �& ���� � ��� �� �& '��� !��� �� ���� � ��� �������� ������ &�� &��� � !����� � �"# �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 '���&�� ��� ���� /���� � ������ ����� 7��� 7 �*� � 7��� ������� �� ������ �� � �� � � � �� 5'� �9�6? ������ �!�� !���� 4��� ��� ��*��� �� � ��� �� 4��� ��� �!�� ������ �!�� +��� * �� ��� � �'� ��� ��� ����� � � �!�� &��� � ��� � ��� ��* � � � ����� 8�� 9 ��&� 4��� �� �� &�!�� ! ! �� ������ � ���� �� �!�� ����� �& ���� �� ��� ����� 4��� �!�� ���� � ��� � '���� ����� 4��� �� �� ��!� ����� 4������ �� � � �� '��!� � ��� �� )& ���� *��� '��'��� ������ �� ��� ����� �& � � '����� 4� � ���� � �� � �&&���� �� ��� ����� ��� & �� �� �� *���� 4�� '���&�� � �� ����� &����* � ���� � �� �� � ��� � ������ 4���� � �! �� �� ���� ���� � � '��� � ��.�� ���� ��� ��� � � ��� � �& � ! � ���� '� � '���� ��� ���� �& �� � �!���� +�� �� 4� � �!���� +�� �� *�� ���� ��� �� �� �� e����� � � /��� 5#$$$6� ���'�� ��� ���� �� �� �� ���� ��� ��� ����� � ��� � 33&��� ���� ��� � � ����� �� !� ��� ��� � ��� ���� �! �� �� '���� ��� �� ����� ��� �� � ����'��� �& ��* '���� �� � � � ��� ���� ��� �� �& &��� �� �! �� �� ��� ��� � � '������ ��� ��� ���� �� �� ��� 4������8�� � �� ������ 5�77:6� ���� �& ��� � �� *�� ���� � �!��� ��� ��� �!� ����� &���� ��� ��� �� � �� ��� ���� �� � �& ��� ��� � � -������ 4�� �! �� �� � � � �� & ��� � ! � �� ��� � �� ������ � � � � !����� '��������� ��� �� !� � ����� � ���� �& � ! � �� !������ �� ' � ��� � �� �� ��� & �� ���� ��� '��� !��� �� ����� ) ��� & �� ��!� ����� ��!� ����� �!�� �� ���� & ��� 4�� +��� ��!� ���� � ��� � �� � �� �� �� � ��� ��� ���� � �� ��'��� �& � � � � ���� ��� � � � *� �� �� �� ��� �! � ��� ��� ��� ��!�� �& ��� � &�� ��� � �� ������� � � ���� � &����� � ��� � �� '� � �� �� �� * �� � � ��� ��������� �&��� ��� � �� '���&�� � � ������� &�� ��� & � � ���2 � � ����� ������ � ��� � � � ��� '� ��� �� ����� ��� � � �� ������ � � � � ��� ��� ��� ��� ���'� � ��� 4�� �� �� ��!� ���� ����� &��� ��� � � ��� ���� � ��� ! �� �& ��� & � ��� �� � ��*�� ��� ����� � � � ��� ���� � ������ ) �� ������ ����� ��� ����� ��.�� '������� * �� � �� ��� � !������ �� '���� ��� � �� �� ��� & ��� � ���� ���� ��!� � ������ �& � �� ����� ���� * �� ����� � !����� � �� �� ����& � � � ������ � �& � ��� �� �� � �� ���� ��� ��� ��� � ��!�� �& � �� ���� � �� � ��� � * �� ��� � � �� ������ ��� � !����� *���� � �� ������ �� � �� � ��� � * �� ��� � �� '��! ��� ��� �i� :$$ ��� '��&�� �� ��� ��� � �� �� � �� � ����� � '���&�� �� )& ��� !������� � �� '��&��� �� � �� ��� ���� �� ��� �i� :$$ � �� ��!�� ��� ���� ��� �� �� � &� � ��� ���� � �� � ��� � � � '���� ��� � & �� ������ � ���� � � &+� ��� �� ��� &��� ��� &��� ���� ����� ��� �� �� � ��� �� �� � +���� ��� � � � �����'� � � ����� � '���&�� � ���� ������� ��� � �� ��� ���� � + �� '������ � ���� ��� � �� �& ��� ����� � '���&�� � ��� ��� �� �!�� � ��� 4� � � �� ��� �� ��� ��� &��� ��� ������ *� ��� � �& ��� � � ��� )& � � !� ������ ������� !�� � ����� !� �� ����� �������� ��� *� ��� � � �'���� �� �& ���� ������ ��� ��� �������� 4� � '�� ��� � ��� ���� ���� ��� * �� ��� �� &� �& ��� �i� :$$ ��� �� � ���!� *� ��� � � ���� ����� ������� 4� � ��� ������ � �� ���� � �� � ���� ����� & � ��� � � !��� +�� ������ ��� & �� 4�� ����� *� '���� � ��� ��� ���� � �&&��� � ����� �� !� �� ��� & ��� 4� � ����� �� !� �!������� ��� '������� �& �� &� � � �� � � '�������� ��� �!���� +�� ��� *� �� ���� ������ �� � � +��� '���&�� � � �� * �� � � ��� +��� ��������� 4�� '���&�� � * �� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �"" �� ����� � ���� ���� � ���� � �� ��* ��� ���� � � ����� 4�� &����� � ������ � ����� ��� ���� ��* � � ��� �� � +��� � �& ��� � � �� � ����� ��� � � ���� �� &����*� � ����� ����� ��� ����� ��� ����� ��� ��� '��� � � � �� & �� �� � � �& � ����� *��� ���� ���� ��� � ����� ���� ��� ���� ��� ����� ��� ��� �� '��! �� �� � ���� &����� � &�� � ! � �� +���� ) !������ �� +��� ' �� � ������ ��' ��� c�� � '��&��� �� ���� �� ��� ����� ��� � �� +��� ��� ����� +�� ��� ���� � ��*� �� ������ ��� �� �� �����-��' +���� / �� � ! ��� �� ������ ��' ��� c�� � ��� +��� ��� '����� ��� �''��'� ��� ��� � ����� � ��� � ��������� �� �� �� !�� �� ���*��� �� � ��� � �& ��� !�� �=���*�� ��������� �� ��� 4�� &����� �& ��� ����� ��� � ���* � �� �� ,� �� ���� � '��� � ��� +�� &���� �� �� �� ���� �� &��� ��� ������� *�� � �� 5�6 �� c��� ����� � �� ��� ��� ����� ' �� ��� � �& ��� � ����� ��� &�� � ! � �� +��� � � ���� � ��� � � �� ��� ���'� �� *� �� �� �� �� ���� � �� � � �� ��� ��������� �� �� � &� ��� ���� ���� � +�� +� ��� ���� ������ ��� ����� ' �� ���� � ��� � *�� � �� ��� � ���� 2 � &�� ����� &� � +��� �� ��� ����� �& ��� � ����� ���� ,� �� ��� ������� ) � ��� 5�77<6 5 6 �������� � � �''�� ��� ��� ����� ����������� �� ����� &� +��� �!�� � ���'�� '�� �� ���� � � &��� �7<# �� �77;� ,� ��� �� + �� �� ����� �a ����� � �� ���� ��� ��� ����� �!� ����� � ��� ������� ) � ��� 5�77<6 �������� �� ��� � �� �& '��&��� � �� � ��������� ��� ����� ���� �� ����� &� �!��� +�� � � � � �� ��� '�' ��� � �� ��� ������� :�$$$ +��� ����� � ������ ��' ��� c�� � � 4���� +��� ���� � &�� ��� !��� ��.�� �� �& ��� ����� �2 �� ������ !�� � �& ' �� ��� ������ +���� ���� ����� :�$$$� ��� ��' :l ��� ����� +�� �� � �� �� g��� � �& ��� ������� ����� ���� �"9 �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �����-��' +���� ��� ��� �:l �� � �-��' +���� � � ��� ���� � �� �����-��' +���� 4�� ���� ' �& ��� :�$$$ +��� � � ��� ��� �� � � ���� '� * �� !��� ���� '�� �� �� ������ ��' ��� c�� � !�� ��� 4���� � c� ��+ � � � ���� ��� '������ �& +�� � ��� ��*� �& ��� ����� ��� 2 �� ����� 4�� ���� 2 �� &�� �� � &� � +��� �� !�� �� ���*��� �� ��� � ��2 ��� ������� � ���� �� ����� ��� ��� ����� ����� +�� �� �� � �� :�$$$ +���� !������ � � ��� &��� ����� � & �� � �& ��� � *�� � �� �� �� ���� ������ �� ��� & �� :�$$$ +���� � ���� � ��! �* �& ��� � �� &� � � ����� �& �''��� ������ ��:$$ +��� �!� ����� ������ ��� � ������ m� � ��� ���� �''����� �� ��� ����� ��2 ��� ���� �*� ��� �� �� ���� ����� &�� ���� +��� ��� '� ��-��-��� �� ��� � 5)=�6 � � ��� '� ��-��-���� 5)=�6 ��� �� �� ���� ��� ������ *� ����� * �� ��� ������� :�$$$ +��� ����� � ����� ������ ��' ��� c�� � � ,� ���� ��� ������ ��' ��� c�� � &��� ��� ������� ) � ��� 5�77<6 ��������� ,� ���� ���� �& ��� +��� � � �� � ���� �& ���� �������� &�� '��&��� � ��� ��� � ���� ��� � �� ��� ����'��� +��� * �� ���� !� ��� �� ��� �� � d=� &�� ���� ��� � ����� ��� �� � � �� � ��� �� � ���� &��� � � + �� ���'��� /���� +��� ���' � � �& � � ���'�� � � �� �� ��! � ���'���� ���� &�� ���� ��� � �& ��� �������� � �� � � ��� � !�� �� ��2 ��� �� ���� � � ��'����� �� �� � ��� �� �& ����� �� � �� � �� ��� ��� �� ���� �& +��� �� ��� ���� ���� !�� �� � � � � � +�� ��� ���� ��� ��� & �� :�$$$� �� � �� � �& +��� * �� ���� !� ��� �� � � � ���� !� ���� !�� � ���� �� * �� � � � �! !���� ' � ��� �� � ��� ��� *� �&&�� ��� ��!��� ���� ��� �� �� � � � �& � � �� �� �''�� � �� �� +��� ���� ��!� � '�� � !� )=� � � )=� ��� � �� �& ��� ���� ��� � ���� �& ��� ��� ��� 4� � � �� �� ��� �� �� �� ��� ��� ����� � '������ �� � � !����� ��� � � �� �� �� � ������ �&��� �� � �� � �& ��� � -� �! ! � +��� *� �����'� �� ���� ����� ���� �� � ��� � * �� ��� ��� ����� ���� ���� ' �� '��� ���� ,� '���� #:$ +��� 5:l �& :�$$$6 ��� ����� +�� ��* � � :$$ +��� 5�$l �& ��� :�$$$6 ��� � �-��' ��*� 4�� ���� � +��� �� �� ��� ����� ��' ��� c�� � ������ ��� �& ��� ��� �� ��� ���� ����� �� �& ��� ����-� � '����� � ��� ���� ��� +��� ��� '����� �� ��� ����� ���� ��� ����'��� &�� +��� �� �� �� ��� � ����� ��� &�� �7<# *� ��� ��� '� ��� ��� ��� � � ���� !�� � &�� ����-� � �7<�� / �� ����� ��� �� ��� ���� �� &�� ���� +�� ���� ��� � ! ��� �� ��� �!����� )=� � � )=�� ���'��� !���� &�� ��� �& ��� +��� ��� � � c� �������� 5��� �!����� �& ��� ���� � ��*6� 4� � ���� ��� � ��� ��� � ����� !� )=� � � )=�� 4���� �*� ����� !� !�� �� ��� ����� �������� &�� ���� +��� ,� �� ��� ���� �� �� �& !�� �� ���� ��� ��� �� ��� ����� �� '���� +��� ��� �''��'� ��� ��� � � )& ��� � � �& ��� �*� ����� !� ��� �� � ������� ��� #�#:� ��� +�� � �� � ����� �� �� � ���*�� ������ ���*�� ���'� �� ��� ��+ �� �� ����� ��! � � ���� ��� �!����� ���*�� ����� � � ��� ��� 4�� '� ��� �& ����� ������ �� � �� ����� �� ����� '� �� � �� '��� � � �� ��� �������� ��'����� � � ���� �� � ���*�� ���� * �� �� � �� � ��� * �� )=� � � )=� ��� �� ���� ��� � �� ����� !� �� +��� �& � � ��� � c� ����� �� ��� ���*�� ����� +��� � � )& ��� � � � ���� ��� ��;:� ��� +�� &���� �� ��� !�� � �������� � �� !�� � ����� !��� � ����� &�� +��� * �� � ��� � ������ !�� �� � ���� � ���� ����* ��� � �� � � !�� �� � � !�� � ���*�� ��;: � � #�#: � '����� ��� ��� � ��� � � ��� +��� �� �* ��� '����� �� � � �& ��� � ���� ����� � �� 4�� ���� �& +��� &�� ���� ���� ���� ���� � ���* 4���� �� 4�� ����� !�� � ����� ��!� ��� +���� ���� �� � � &� ��� � ����� ���� �& +���� ���� �� � ��� & �� ���'�� �� ��� � ����� ��' ��� c�� � +��� � ��� ���� ����� * �� �� ���� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �": 4���� � d ���� �& +��� ���� ��� ������� ����� ��� � � & �� ���'�� n��� g���� !�� � � �-��' !�� � ����� !�� � g���� ��� � � �-��' ��� � ����� ��� � g���� ���*�� � �-��' ���*�� ����� ���*�� 4���� �& ����� �7<# �9" ""< :<� "7 :a a: a; �$; �;# �:a< �7<" �9� "9$ <#� 9: a# 7$ a" 77 #�# �<;" �7<9 �9: "�7 a<" 9: <� �#$ :7 �$� #$$ �;:" �7<: �<� """ ;a; "9 ;� �#� "9 7; #9� �<;7 �7<a �:: ";a �$$" "; :" 7" :; ;# �aa #$�# �7<; �<: "$9 <7" #< <: �"7 "a ��# #9# #$#9 �7<< �99 "$" �$99 9$ ;; �97 a: �#� #:a #�77 �7<7 �:a "<a �#"$ ": :� �"$ :< a9 #97 #":7 �77$ �7< "�: 79a �7 ;$ �"7 "# ��a "$# #�"; �77� �aa #<" �"$$ #" a< �:7 a$ �:$ "�; #:#a �77# �9: #7# �::� "7 7" �a; a: ��a "�" #;<� �77" �a: "#9 �::� #< aa �7" :a ��� ";# #<aa �779 �"; "$9 �a9a 9< ;9 "$� a9 �#" :$a "#$" �77: �a# ":" �7;: 9# 99 �9$ 9: �$9 "#$ "�<: �77a �a" #<9 �7:: 9$ <$ �:� 9a �"; "�< "�;9 �77; �<9 ":" #$7a ## :$ :7 9" 7< �9; "$:# �!����� �a$ "#: �#:" ": a< �"7 :" �$< #;� #9�# � " a � () ( � � � & ���� � (� ( � � *�� � +� �� � � �� � � ���� �� , ����% . /� . . 0 � 1 2 3 3 !�� � +��� �������� ��� ���*�� +��� /& ��� +��� &��� � �� ��� ���*�� ��� � � ��� ���� * �� ��� �������� +��� �� �� � ��� ������� ���� �& �����!�� � �� 4� � � ����� � �� � ��� � * �� ��� �� ���� �����!�� � ���� ���*�� +��� �� � �� �� ��� � ��� ������� +���� 4�� ����� �& � � �� �� � � ��� +��� ���' �!�� �� � ��� �� �& ��� ������� ����� ��� �''����� � �� ���� & ��� � �� �� � � �& � �������� � '����� �& � �� &��� ��� ����� +��� � �� � '��� � ��� ���� ����� �� �� ��� ����� ��� �� ���� ��� ''�� ��&� ����� 5����� !�� �� ����� ��� �� � � � � !�� �6 ��� ��� ��* � �� ������ *� �� ��� ��*�� � ��� ����� 5� � ���*��� ����� ���*��� � � ����� ��� �6 ��� ��� � ���� � �� ������ 4�� ����� � ��� � ��� �� ��� ��'����� �� &��� ����*���� ���*�� ����� �*� ��� '�� ,� �� ��� ���� ��! �� � �& �� �� ��� � �� � � ���� �� �& � ��� 4�� ����� *� ����� ��� ���� ��! �� � �!�� ���� �� ��� �''��'� ��� ���� �� �& '���&�� � � �� � ����� � ��� + � �� �& ���� � � ��� �� 5�77#6� 4��� '��! �� ���� � � ''��� &�� ��� �� � �� �& ���� �� ���2 ����� ��'� �� ��� � �� ����� ���� * �� � &&��� ��� �����-���� � �� ����� ��� � �� (��* � &��� ���� � �������� � � ��� � �* ��� ��� ���� �&&�� ���� ���� ���� �� ���2 ����� ���� � &�� ��!������ � c�� � � + � � �� � ������� m� � ���� *� �� &� � �� ��'� �� ��� & �� �� �� �& � ��� ����� ���� * �� ��� � �� ��� � �� 4�� ���� � ���' � �� ������ � & ��� ��� � ��� �� � � �!����� �� � ��� � * �� ��� � �� ������� * �� ���� ����� ,� ���� ���� ���� � '�� �� ��� � � 5h� 6 &�� ���� ����� �!�� ��� ���� ��� ����� 4�� ���� ��� � �& � �� �� � � �� � �� � �!�� ��� ���� � �� '�� �� �� ��� ���� ����� ��� � �� ��� ����'��� ���� ��� � ���� h� � � ��� ���� ��! �� � &�� ��� ���� �7<# � &��� ��� +��� ���� � ��� �7<# ���� �� ��� ���� ���� � ��� �& �7<#� 4� � *���� '������ � ��'����� &�� ���� ���� ���� �� �77;� 4� ���� � ��� �'��� �& ���� � �� � � ��� � *� �� �*� ���� ��� � �� � ���� *� � ! �� ��� ��� � &�� ���� ���� �� �� ���'��� !� ��� ���� ��! �� � � 4� � �''����� � � � ��� �� ��� ����'� ��� ����'� ���� *� �� �� � ������ ��� � ��-&��� ���� &��� ��� ���� � '�� �� ��� � ��&��� � ! � � �� ��� ��� ���� ��! �� � � 4� � ������ ��� � ��-&��� ���� *� �� �� ��� �� � � � ����� �� �� �� � � �� ����� ���� � � ��.��� !� � �� �� ���� &�� �� ����� ��� � �� � � � �� �� ���'��� � ��-��. ���� ��� � '��&���� ��� � ������ � ��� � ��-&��� ���� &�� � � '��� � ��� ���� *� �� ��� �� � �� � ����� ��� �� � � *� �� �� ����� ����� !� �� � ��� e��� �� �� ���� ��!� ��� ��� �� ��� �� �� ���� ��� � ��=��� � ���� ��� � �� � �� � * �� ��� 4��� �� ���� ,� � � ���� �� � ���� �& ���� � � ���� 5�77a6 � � &� � ��� ��� ��� ������ �� � ��� � * �� ��� ��� ���� ��! �� � ��� ���� 4�� � �� � &&��� �� �����!�� *�� ���� ���� �''����� ���� �� ��� ������ ��� ������ 4� � ���� �& '����� � �� ��'����� � !� ��� ��� ��� ������� �� � � �� ��� � ��� � � � ��� �� � �� �& ����� e��� �� !������ ��� ���� � � ��� ��� ��� � �& �� �� �!� ����� &�� '���&�� � ����� � � ��� ���� �& ��� � � �� � �� �� ���� ��� ��� ���� ���� �� ���� � � �� & � �� ������ 4� ��� �� ������ � ��� 33������ +��� &��� ���� ���� *� ���� '� ��� '� ��-������� ��� � 5)=�6� 4� � � ��� ��� � ��� ���� !�� * �� ���� � � �� �� � � � � +�� � !�� ���� �� � &� � ��� & � �� '��&���� �� �& � ������ e������ � ����. � � � � �� 5�77a6 � !� ����� ����� � *�� ��� '� ��-��-����� ��� � ��� '��! �� � ���� � ����� �& & � �� ����� '� �� '��&���� ��� � ���� � ���� ��� ����� ��� �� ���� ��� ���� ��� '��+�� � � �� � ��� ��� �� � �� �� �� ' ���� ��� ����� !�� ����� ���� �� ��� �� ��� ���� � ���� �� �� �&&����� �� ��������� '�� ��� 4� ��� '� ��-��-����� ��� �� �� �� �� � ���� !� !�� �� /& ��� &� � !�� ����� ���� ������� e����� �� ��� 5�77a6 + � ��� )=� ��� � 5���� ��� ���� ��� �=)6 �� ��!� ��� � �� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �"; ���� �� ����� � � +�� � ����� � �� '� * �� & � �� ����� ��� � �� ���� � � g���* �c 5�7776 �� ��� )=� ��� � �� � � �& ��� � ����� � ���� � ���' ��� c�� !��� � ����� � ������� � � /���� �� ���� 5�77;� �77<6 ���� + �� �� � !�� ���� �� �� ��� ���� � � +�� � � ��� � ����� �& & � �� ����� '��&���� ��� /& �'���� �� ��� �''� ��� � �& �� � ����� � ��� ���� *� �� !������ �� ��2 �� ��� !�� � �& ��� )=� ��� � &�� ���� +��� ��� ����� * �� � ������ ���' ��� ���������� ����� ��� � ���� �& ���� ��'� � !� � � ��'� � !� @( +��� �!� ����� &��� *� �� �� � ��� � �� �� ���� ������ 4� � &����� � � ���� �!� ����� �� � ���� � ��� ��� ����� *�� � ��� )& ��� !����� � � � ��� *�� � �� �� �� � &� +��� ���� �� �� ��� ��� ����� & � �������� ���� �� � �� ��� +��� � '��� � ��� ����� 4�� � �� �� � �� ��� 33����� ����� ��� ����� ����� � � � ���� c � ��� �� ��� � �� ���* )=� ��� �� ��� !������ �� ���� &�� ��� ��*��� )= � !�� ��� ������ ��� ��� ��� ���� �& +��� * �� ��� ��*��� )=� ��� � ���� ���� �� ������ ��� '���&�� �� ( � �� ��� �''��� � � ����� �& ��� )=� ��� � *� �� �� � !�� ���� �� ��� �� ��� ���� �& +��� * �� ���� ���� �� � 2 � � �� ���� �� ����� ���� �� ' ������� �� �� � !������� ,� ������ '���&�� �� �& ���� �$ � � #$ +��� �� ������ � ��� �$=#$ +��� * �� ��� ��*��� )=� ��� � * �� ���� ����� ,� �� �$ � � #$ ��� ��� ������ ��� ������ �� � 2 � � �� * �� '��! �� ���2 ��� � !��� +��� � �� ���� ��� � �� ��� �� � �� �+�� �& � !��� +��� � *� �� �� �� � ��� ����� �� ��� ����� � �� � � � ���� � '��! ��� � ��'� � !� *�� �� �����'� �� ���� ��� �� � ��� ��� ���� � � ��� ��� �� � ���� � ����� ��� �� �� �� ���� �� ��� �� ����� ��� � �� � ������� ,� ����� �� ���� � �� ��� & �� ���'�� �& +��� ���� ����� 4���� #� '��� �� ���*� ��� �!����� ��� ���� ��! �� � �& ��� � � &�� ���� �& ��� � ������ � � ��� �� � � �& ��� ���� � ��� ����� � ��� �����!�� � � ��'����� '����� � @��'�� � ��� � �� �� � �� *� + � � �� �*� � �� �� � � � &&��� ���� 4�� � �-���*�� ���� � � ��� �����-!�� � ����� ��!� ��� � ��'����� �� �� �& " � � 9 �� �� ��!����� �� 9 � � "� ���'��� !���� � � ��� �� � �� �& ��� �����-��� � � � �����-!�� � ����� ��� ��!������ 4�� ����� *� ��� ���� �� ��� �� * �� ��� ����� *� �� ��� ����� ����� � � ���� ��� � �& � �� �� ��* � ��� ���* '��� # �& 4���� # � ���� ��� ����� � �� ��'����� � �� ����� ��� �� ��� �� #� � � " � � ��� " ��* ��'����� � �� ����� ��� �� ��� ;� <� � � 7� 4���� ��� ��� �� +�� ���� � � ���'�� �& +��� � � ��� � ����� +��� � �''���� �� �� �� � ��� � * �� ��� ������� � �� �� � � ��� �� �� ������ ���� ��� ��� � � �� ��� �� � �& � ��� *� ��� ���� �������� ������ � *������ ��� '���&�� �� �������� �� � ! � �� !������ ��� � ���� � ����'����� � ����. ���� ��� � � 4���� " '��! ��� �! �� �� � �� � ��� � 2 ��� � � ��� ��� �& � � '���&�� �� ���� ��� � ��� � � *� ������ �� ��� '���&�� �� ���� ���� �� ���� ������� * �� ��� ���� ����� ��� ����� �& ��� ��*-� �� ����� ��!� � ���� ��� � � ��� ��� ����� � ��-� �� ������ 4� � ��� �� �''���� �� �� ��� �''�� �� �& ��� ��� ���� � ��=��� � �����-�&& ���� + � �� ������ ����� � *� ��� �� ��'���� 4� � � �� ��� � � '��� �& ��� � ��=��� � �2 �� � � � � *� ��� �� ���� � � ���� ���'�� � � !��� ���� *������ ����� ��� � � ���'� ���� &�� ��� � �� ������� � ,� � ! �� �"< �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 4���� # �!����� ��� ���� ��! �� � �& ��� � � &�� ��� ����� ����� ����� � � & �� ���'�� @��� ��� �!����� ��� ���� ��! �� � �� �� �� � ����� � �� ��*��� � 7 ��� ����� ��'����� � �� �� �� ���� � ����� ���*�� $�:$"< � � ����� ��� � $�9::a # # � � ���*�� $�"<�" 9 " ����� !�� � $�"79< " 9 � � ��� � $�""$< : : g���� ���*�� $�"$a7 a a � � !�� � $�#<�# ; ; g���� ��� � $�#9a; 7 < g���� !�� � $�#;$7 < 7 ���� # g�* � �� ����� g���� !�� � $�#;$7 < 7 � � !�� � $�#<�# ; ; g���� ��� � $�#9a; 7 < h �� � �� ����� � � ���*�� $�"<�" " " ����� ��� � $�9::a # # ����� ���*�� $�:$"< � � � ��� ����� ���� �� ���� ��'� � � ��� �� ��� � ��� � �� ����� �& ��� � ��� �� ������ 4�� +�� ������ ���� ����� �-�� �-#� � � #-� ��� �& ��*��� � ��� *� �� ����� #-"� "-"� � � "-# ��� �& � ����� � ��� 4�� � ��� �� ����� &��� ����*���� ���*�� � 4�� �� � �� �& 9� :� � � a ���* ���� ��� ����� � ��� � ����� �� ���� ����� +��� ��� �� ������ � �� �� ��� ������ +���� � � ���*�� +��� � �� �� ��� !�� � +���� 4���� " �!����� ��� � � � � ��-��. ���� ��� &�� ��� ����� ����� ����� � � & �� ���'�� �!����� ���� � '�� �� ��� � � � �& ���� � '�� �� ��� � �� ��-��. ���� ���� ��-��. ���� ��� �� � g�* � �� ����� g���� !�� � �"�9�l � $�9797 � � � !�� � �"�"7l # $�9;a# # g���� ��� � 7�;�l " $�"7": " h �� � �� ����� � � ���*�� :�9al : $��9"" 9 ����� ��� � :�<<l 9 $��#7� : ����� ���*�� "�9�l a $�$a;; a � 4�� �� � �� &�� ��� ��� � " �� � �& ���� � '�� �� ��� � � � &�� ��� ��� � : � ��-��. ���� ��� �� � � �� �� &��� �� ��� � ����� ��� � �� � �� !�� �� �� a� ��� ��*��� ��� � �� � �� !�� �� � 4�� � ��-��. ���� ��� � ���� �� �� � ! � � ��� �!����� ���� � '�� �� ��� � �� ��� ��� ���� ��! �� � �& ��� � � &�� ���� ����� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �"7 ��� �!����� ��� � �& ���� ���� �� ��� ������ �!����� ��� ���� ��! �� � � 4�� � ��-� �� ����� ��!� ���� ��� ��*��� ��� � � � � ��� ��*��� � ��-��. ���� ��� � �� 4�� ��*-� �� ����� ��!� ��� ����� � ����� ��� � � � ���� � ����� �� � � � ��-��. ���� ��� �� ) !������ ��� �� ��'��� �!����� ��� � � �� ���'� ���� &�� � ���� � ��� ,� �� �� + � ���� +��� ��� � ��-� �� ����� �&&�� � ���� ��� � �� ) ������ ���� �&&�� � ���� � ���� ��*�� ��� � � ��� ����� +��� ��� � �� �� ��� ��*-� �� ������ m� � ��� �$ +��� * �� ��� ��*��� )=� ��� � *� ���� ��� ��� ��� ���* 4���� 9� ,� ��'��� ��� ���� &����� � 4���� 9 ���� *� �� 4���� "� 4�� ��*-� �� ����� ��!� � � ���� � ��-��. ���� ��� !�� � ��� �� ��� � ��-� �� ������ @�� � : ���*� ��� ��� ��� �& ��� � ��� ��� � # ��� � � � � � ������ � ��� ���� ���� ��� � &�� ��� & �� ���'�� ���* ��� � # �& 4���� "� ��� ���� ���� ��� ��� � � � ���� �� ������ � ��� �$ ��* )=� +���� 4�� ������ ��� � �� ��� &��� <�;al &�� ��� ����� ��� � ���� �� ��;#l &�� ��� ����� !�� � ����� g��� � �� ��� � : *� + � � ������ � ��� ��� 4�� ��!� ���� �� � � � �� �$ +��� � �� ������ � ���� &�� ��� � �� � �� ���*�� ����� ��� &�� ��� ��* � �� !�� � ������ ��� ��� � �� � �� ����� ��� '��&���� �� ���*� � ������ �& �!�� :l &�� ��� ����� ������ *� �� &�� ��� ��* � �� ����� � �� � � ���� ��� � ��� � ������ ������� ��� �!� #l 5;�;#l &�� � � !�� � ����6� 4� � �����!�� � � � � �'� �� � �� ��* )=� ��� �� ��� � ���� ���� �& � !�� � � ����� ��� � ���*�� � ������ � � �*� �& ��� � ��-� �� ����� ��� ���*�� ������ ,� � �'��� ���� � � ���� 2 � � ������ � ��� ���� !�� �-�� � ��� +��� ��� � ��� '�' ��� � �& +���� ,� ��'����� ��� ���!� '����� ��� &�� ��� �$=#$ +�� '���&�� �� � � ������ ��� !����� � ��� � �� � ����� ) ����� *����� *� ��� ��� ���� ��� !����� � � �� !��� ��� +��� ���� ���� � � � ���� � ���2 � � �� ��� ����-� � &�� ��� +���� + � � �� ������� ��� 4�� ����� &�� ���� � � � �� � ��� !����� � � ���� ��� ���� *� �� &�� ����� +��� � �& 4���� 9 �!����� ��� � � � � ��-��. ���� ��� &�� ��� ����� ����� ����� � � �$ +��� * �� ��*��� '� ��-��-����� ��� � � �� ������ � � �!����� ���� � '�� �� ��� � ��-��. ���� ���� ��-��. ���� ��� �� �� �$ � �� ��� � ���� ��� & �� ���'�� ��� � � g�* � �� ����� g���� !�� � �9�77l $�a�:� � ��:<l � � !�� � #����l $�:##$ # ;�;#l g���� ��� � ���9"l $�9;7< " ��;#l h �� � �� ����� � � ���*�� ���$"l $�"#"$ 9 :�:;l ����� ��� � �9�a9l $�"##; : <�;al ����� ���*�� <�a#l $��<<9 a :�#�l � 4�� � ��-��. ���� ��� � ���� �� �� � ! � � ��� �!����� ���� � '�� �� ��� � �� ��� ��� ���� ��! �� � �& ��� � � &�� ���� ����� � 4�� � ��-��. ���� ��� �� � � �� ��� &��� �� ��� � ����� � ��-��. ���� ��� � � �� a� ��� ��*��� � ��-��. ���� ��� � � � 4�� !�� � ��� � : � ���� �� �� � ������ � ��� �!����� ���� � '�� �� ��� � &�� ��� & �� ���'�� '���&�� � 4���� " &��� ��� �!����� ���� � '�� �� ��� � ��� � #� �9$ �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 +��� ��� ���� � *� �� �� �� �!� ����� ���� ����� �� ����-� �� *� �� � *��� *� ��� �� * �� � � ����� +��� � �� e� *� � � � � ��� � � �& ��� ����� *� ��� ����* � ��� ���'� � &����� � �� �� ��������� � � ���� �!� ����� �� !������� �� � �� � !������ '��&�� �� �� ��� � '���&�� � ������ � � ��� ��� � � �� �� ��� ��� � �& � ����� 4�� �� � � ��� ��� � � ��� ����� *��� !��� � � ��� �� ����� . �� � �� ����� ����� �� � �� �� &�� �� � �� *��� &� �� ��� ����� ��� ��*-� �� '���&�� �� �� ��� �� #� � � " &�� � ��-��. ���� ��� � � � ��� � ��-� �� '���&�� �� �� �� �� ��� ��*�� " ��� � �� ��� ����� �� � ��-��. ���� ���� � '�� �� ��� � � ��� ���'& � &�� �!�� �� � '���&�� � '��&���� ��� ��� �� � 33 ��� �& ��� ������ &�� !��� � � ��� ���� �& ������� ���� � !����� �� ��� � ���� � ��� � � �� ��� c� � ,� ���� ����� ��� ���'� ��� !�� � �& � � �� !����� � �& >�$�$$$ �� ��� ��� � �& �7<# � � ���� � � ��� � � �& �77;� m� � �� � ������� � �''����� ����� �� �� � ����� � ���� ��� ��� � � � ��� !�� �� � �& ��� ��� � � � � � ��� ���'�� '�� ��� 4�� !�� � �& ���� ������ '���&�� � � ���* 4���� :� ��� ����� ��*-� �� ����� �� ������ ������� *����� ��� � ��� � ��� � � �& ��� ����� � ��� �� ������ 4� � ����� !� '��&���� �� ����� &�� ��� ����� '���&�� � � c��� 4� ���* ���� �� � ��� �� � �� � � �� � �� '�� �& � ����� � ������ *� ���'��� ��� '��&���� �� �& ��� � � '���&�� �� &�� ��� & �� ���'�� � � &�� ��� �$ � � #$ ��� � �� '���&�� ��� � �� ��� ����� ���'�� '��! �� � � ��� �� �� � � � *� '���� � ���� � �� ��� ���'� &�� ��� & �� ���'��� 4�� '��&���� �� �& ��� ��*-� �� '���&�� �� � �� � � �� . �� � &�* ����� � � ��� � �� ��� � � ��� �� � ��� � ������ � ��� 4�� ����� ��*-� �� ����� ����� �� ������ ��� !�� � �& ��� ����� ����� ����� ��� +��� &�* ����� �& ��� ���'�� '�� �� � � �� � � �� � �'��&��� ��� � ��-� �� ������ ��� ����� ��*-� �� ����� ���� �� ��'�� � ��� ���� !� ��� � � � �� �$ �& ��� 9< ����� �& ����� 4�� � ��-� �� ����� ��� �7 � � ��� !� ����� �& ���� !� ��� � �� 4� � �''��� ������ # �� � ��� � �& ���� !� ��� � � ���� ����� �� � &�� ��� �$ � � #$ +�� '���&�� ��� ,� � � �� ���'��� ��� '��&���� �� �& � � ����� * �� � ������ ���� � �� �� ��� �i� :$$� � �� � � ��.��� !� *�� �� �� ���� &�� �� ����� '��&���� �� � � � �� �!����� � ����. ���� '��&���� ��� n� ��� ��� ���� ��� � � ��� !� !�� � &�� ��� �i� :$$ ��� � 4���� : ��� � ����� *����� &�� ��� ����� ����� ����� !�� � �& >�$�$$$ �&��� �a ����� �& ���'� ��� ���*�� �$ g�*��� )=� ��� �� � �� ���'�� g�* � �� ����� g���� !�� � >;:a$$ >a;777 � � !�� � >�<7#<" >a""99 g���� ��� � >97aa" >"<a9; h �� � �� ����� � � ���*�� >9"<#; >�7�a" ����� ��� � >9a<;� >#$;7< ����� ���*�� >#:::" >�9$$# �i� :$$ >;7#�; >;7#�; ��� !�� �� ��� �� ���� >�$�$$$ *�� !����� � �#="�=<�� ��� � ! �� �� ��� �� !������ �� � � ��� ���'� ��� &��� �7<# �� �77;� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �9� ������� ��� ��� �& ��� � � ��� !� !�� �� ����'� &�� ��� �$ +�� '���&�� � ��� � � !�� � ����� ����� �� ��� !�� � � ������� *� �� ���� ��� ���� ��� �!����� ��� ���� ��! �� � �& ��� �i� :$$ �!�� ��� �; ����� �& � � �� �� � � ���� ��� � � �& ��� ���� '���&�� �� �!�� �� � � �� ��� c� � 4�� �i� :$$ �!����� ��� ���� ��! �� � � $�:7#:� 4�� ������� ���� ��� ���� ��! �� � �� ���* 4���� # � $�:$"< &�� ��� ����� ���*�� ����� ��� &����� � � � ������ ' �'���� � �� *� ��!� �� ��� ��� � � ��� !� !�� � �& >�$�$$$ &�� !��� � ��� �i� :$$ �� � ��'����� � � ��� 4���� : � � � �� #� � ��������� 4�� ��� ��� � ���� ���� ��� '����� �& �!����� � �� &�� ��� '���&�� � �& +��� ��� ����� �& ��� ��� ������� ����� ��� ���� � � �� � ��� � * �� ��� ��� � �& ��� ������ m� � ��� ��� ����� ���� ����� +��� � � ����*� ��� !����� �� � �� ���� ���� ��� !��� � � '���&�� � ���� � �� � ��� � * �� ��� � ����� !� � �� ������ ��� 4�� '����� �� � � �'�� �� �� � &� ��� ���� * �� ��� � �� ��!�� '��&����� �� ��� !����� &��� ��� ������ ����� �� ��� '����� �& � �� � �� � ��� � * �� ��'����� � �� ��� �����!�� � ��-��. ���� ��� � &�� ��� � ��-� �� ����� � � �� ��� �� ��� ��� ��� � ��-��� � ������&&� h ��-� �� '���&�� � !������ * �� �� �� ���2 ����� ��*����� &�� ��� � �� ������� � 4�� ��� � &��� ��� ��*-� �� ����� ������� ���� �& ��� � ��-� �� ������ 4�� ���� � ��=��� � ���� �� � ����� &��� ��� �� �& ��� �����-!�� � � � ��� � �-��' !�� � ������ ���� �& *� �� &��� �� ��� � �� #� 0�� � �& >�$�$$$� �9# �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 ��*-� �� ����� +��� � � �� ������� ��� !����� * �� ���� ������� *����� ��� �� �-� �� ������ � ������ &��� � � �& ��� ��*-� �� ������� ��� / � ��� ��� �&&�� � ''��� &�� � � ��� ��� ������� ����� ��� ���'� �-�'�� +� &����� � �� � ������ � ���� &�� ������ � ��*-� �� !�� � ��� � � ��� ,� �� ��� ��� � � �� ��� � �� �� ��� ���*�� ����� ��� &� � �� �� �& � ���� � �� ��� ����� ��� !�� � ������ ���� �� �� �&&�� ���'� ��� � ��� � � �� . �� &� ��� � �� � � � '���&�� �� ) !������ ���� � � ��-��. ���� � ��-'��&���� �� ��� � � �� �� �� � �����& ��� �� ��� ����� ������ � �������� � �� �� �� � '�'��� ��!��"�� �#���� 4�� � ����� * �� �� ��� � ��� � � ��� � ��&����� &�� ��� � !�� ���� �� �� � � � � � � � �'�� �� ��� �� �� �� ��� j��� � ���� g���� &�� ��� � �� � � ����� ��� ���������� e������ ,� @�� k��� � ����. � ��� i � ��� �� �� 5�77a6� (� �����-'� �� � � ����-�2 �� ��'�� ����� ��� � � ������ ��� ����-������ � � +�� � c�1 ����� ��� ������� ! ������ :596� :a8a$� e������ e� ��� i /��� � 4� 5#$$$6� 4��� � � ��c���� � �� �� � *�����? ��� ����� ����� !����� � '��&���� �� �& � ! � �� !������� ! ����� ����� �� ��5#6� ;;"8<$a� @������� �� �� 5�77;6� / '��� ��� �� � � �� & � '��&���� ��� ! ����� ����� �� ��5�6� :;8<#� @�� ��� �� �� ���!�� k� ,�� i � ��� ��� � ,� 5�77"6� � ��� � ��� 0�� � g � ����� � ��� ��� �&&���� ! ����� ����� ��� ,����� �� ;645#6� �;�8�;7� @����� k� ��� i ���!��� �� e� 5�77<6� 4�� *���o�*�� ��������� � �9��� �� ��98��;� (���!�� ,� g�� i /��� � �� 4� 5�77<6� 4�� ����� � �� ' ���*�� � � �� & � &��� � � ��'� ��� � � ��� � �&&���� � '��&���� ��� 5�� ����� ��� ,����%� ::5�6� <:8�$"� (���� � (� 5�77<6� � �������� ���������� ������8���# ��� ��<� 8������� �� � �� i ��� ����� d�* n���� ����� �� ��� i ��� ��� j� � 5�77#6� 4�� �����-���� � �& ��'����� ����� ��� � �� 5�� ! ����� ����� �� 3�5#6� 9#;89a:� � ����� j� g� 5�77a6� �� � ��! � ���� � �9��� 1� ##:� ����� �� @�� i g���* �c� �� 5�7776� @��' ��� c�� ����� ����� � � ��� &�� '���&�� � ������ � � ����� ��� ����� �� ,����%� �5#6� a�8;$� ������ � (� 5�77:6� �� �� ��� �������8�� �# %�� �� �� ������ � �� ���� 4=� e� ��� e����� d�* n���� j���� �� ��� i d��� �� �� 5�77<6� 4�� ������ � �� � � '��� �� !� �� � �� �& � � �� & � ��� ��� 5�� ! ����� 4�������8� ����� a�8aa� g��� ����� k�� ���� &��� ��� i 0 �� �� � ,� 5�7796� @� ���� � !����� �� �����'���� � � � � � ��� 5�� ! ����� ����� �� 315:6� �:9�8�:;<� g�! ����� �� g�� i k���� � ,� k� 5#$$$6� ����� ������ � ����� � ��� ������� �$-����� :-���� �� ���� �2 �� � � �� & ��� ����� ��� ����� �� ,����%� 15#6� �9:8�:;� � ��� �� � /� 5�77<6� )� !�� � � �� � �� � � ��1 )�'� ��� � � &�� ����� �� ����� � � � ������ ����� ������ � � ! ����� 4�������8� �5�6� a98a:� /��� � (� /�� d���� � k� ,�� , ��� @�� i ������ � @� 5�77<6� � ���� �& ��� !������� �� �� ����� ��� � ������� 5�� ����� ��� ,����%� ::5#6� �a�8�;:� /���� �� ����� k� �� 5�77;6� 7��� % �*� � %��� ������� d�* n���? �����* h ��� /���� �� ����� k� �� 5�77<6� > % � ������ �� �� d�* n���? e����*�� e����� ��*� � �* �c� k� ��� i ������� (� @� 5�77�6� 4�� �'��� �& 0�� � g � �'�� �� � � �� � � ������� ��� � � � ����� '� ���? �! �� �� &��� ��� �!��-���-�� ��� ������� ����� ��� ,����%� �?596� :9;8:a<� �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 �9" ������� � (� �� i ������� � �� �� 5�77:6� �� ����� ��� � � � � � ������� ��� �� &������� ��! � � � ����� ���� * �� ��� ' �� ��� � �& 33����� � ��� ������ �� 0�� � g � ) !����� � � �!��� 5�� ! ����� ���� ��� ,����� �� ;6444596� 9a:89;;� ������� �� ��� i 4� &��� k� ,� 5�77<6� ���&���� �� '��� ��� �� �& ��'�� � ��� � � �� & � �� ������ ����� ��� ����� �� ,����%� �5�6� :;8a<� ������� ) � ��� 5�77<6� ��� ���� � � ������ @��'���� � � @( ��������� ������� �� �� 5�77a6� 99 ����� �& � ����� ��� � � ����� ������ �*� @ ���� -�� �"98�"<� ,������ �� ��� i h��+���� �� e� 5�77a6� ���&��� � �� ����� � ������� ������� ��� � �? �'� ��� � � &�� � ! � �� !������� ����� ��� ����� �� ,����%� �5�6� �"8#7� �99 �()( � ��&���� �(�( ��*��� + ����� ��� ����� �� ,����% -. /�..-0 -�12-33 a stock selection model using morningstar's style box introduction the need for standardization literature on security selection methodology results conclusion acknowledgements references 134 the value of financial advice: a narrative review and conceptual frameworks kirsten l. macdonald,1 karen l . wildman,2 ellana loy, 3 and mark brimble4 abstract efforts to increase the low global consumer uptake and recognition of financial advice as a trusted profession have been hampered by low financial capability, distrust of financial advisors and soaring costs. while considerable research has been published on the value of financial advice, a synthesis of scholarship is surprisingly absent, leading to a lack of credible information on the outcomes of professional financial advice. to determine the ways in which value has been contextualized and measured and the extent to which value has been substantiated, this themes-based narrative review comprehensively examines tangible outcomes, such as investment performance, and less tangible components of value relating to consumer wellbeing and the client-advisor relationship. we conclude by proposing a conceptual framework of the value of financial advice to support the development of a more consistent, rigorous and coherent body of literature across jurisdictions, with increased transparency for consumers and other stakeholders of financial advice. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation macdonald, k. l., wildman, k. l., loy, e., and brimble, m. (2025). the value of financial advice: a narrative review and conceptual frameworks. financial services review, 33(4), 134163. introduction with global concern over the cost of living driven by post-pandemic inflation and interest rate rises, professional, affordable, and accessible financial advice may be needed more than ever. improving access to quality financial advice supports consumer financial sustainability (security, safety, and wellbeing) and better navigation of the environment in which individuals make personal financial decisions. increasing individual financial resilience also creates social and economic payoffs for economies by reducing the burden on government support in working life and retirement (macdonald et al., 2023; oecd, 2020). nonetheless, efforts to increase the low 1 k.macdonald@griffith.edu.au, griffith university, brisbane, queensland, australia. 2 corresponding author (k.wildman@griffith.edu.au), griffith university, brisbane, queensland, australia. 3 e.loy@griffith.edu.au, griffith university, brisbane, queensland, australia. 4 m.brimble@griffith.edu.au, griffith university, brisbane, queensland, australia. global uptake of financial advice to address unmet advice needs are challenged by key trends relating to the demand and supply of advice. the number of individuals seeking advice from a financial advisor is declining and the way in which individuals seek and consume financial advice is changing. only one in 10 australians sought advice in 2021, down 1% per year since 2019 (adviser ratings, 2022; australian securities & investments commission [asic], 2019). although one in four americans currently receive advice from a financial advisor, like australia, advised clients are more likely to be male and older (cfp board, 2023). from 2019 to 2021, ongoing advice in australia https://creativecommons.org/licenses/by-nc/4.0/ macdonald et al. 135 trended downwards (85% to 75%) and piecemeal or one-off advice increased (15% to 24%) while advice-seeking via technology and social media increased (up to 5% of individuals) (adviser ratings, 2022). importantly, the intention of those not currently advised to seek advice in the future is strong (e.g., 41% australia; 25% us), particularly in younger age groups (asic, 2019; cfp board, 2023). however, a multitude of barriers are cited for not proceeding with advice, including cost of advice, not getting around to it, size of financial position not worth getting advice, wanting to control own finances, and anxiety about seeking help from and lack of trust in financial advisors (asic, 2019; westermann et al., 2020). the notion of experts as trusted professionals is common across many fields which feature the provision of advice and client-professional relationships. although financial advisors’ expertise and ability to provide access to knowledge and products that clients do not have access to is a core reason for seeking financial advice, there is a lack of confidence in positive long-run outcomes from advice and trust in advisors (asic, 2019). in addition to a mass exodus of financial advisors in some jurisdictions due to regulatory changes (including increased education requirements), service costs have increased significantly (40% overall in australia 2019-2021; 12%-25% depending on fee type in us) (adviser ratings, 2019; tharp, 2021) and are the major barrier to advice (adviser ratings, 2022; asic, 2019). however, attempting to reduce costs alone to address the affordability and accessibility of advice is arguably futile. cost and trust are among the constructs intrinsically linked to value. a client will not be willing to pay even a lower fee for financial advice if the outcomes are not expected to be positive. confidence in positive outcomes is associated with the quality of advice and relational aspects of financial advice, such as trust in the financial advisor, which are built over time. critically, when asked directly about value, 18% of australians surveyed did not see value in consulting a financial advisor (asic, 2019). thus, even addressing quality of advice concerns or the pervasive distrust of financial advisors stemming from product and advice failures and misconduct must be accompanied by more credible information on the outcomes of personal professional financial advice (macdonald et al., 2023). while considerable research has been produced on the value of financial advice, a synthesis of scholarship is surprisingly absent. a scan of the literature reveals two articles that fall within the vicinity; these being a systematic quantitative literature review by macdonald et al. (2023) which quantified key trends within the literature such as authorship, geography and research methods, and a review by westermann et al. (2020) of the barriers and benefits of financial advice-seeking. while these reviews usefully expand scholarship within their methodological and research scopes, they lack a comprehensive descriptive overview of the value of financial advice in its entirety. this themes-based narrative review fills this gap by describing the history and development of value of advice research. we clarify and critique the ways in which value has been conceptualized, measured, and substantiated in the realm of personal professional financial advice. in particular, this paper strives to answer these research questions: 1) how has value been operationalized in financial advice research (i.e., what is ‘value’)? 2) is value realized for clients (i.e., is the advice worth it)? 3) what recommendations can be made to guide and enhance future research on the value of financial advice (i.e., where do we go from here)? overall, we conclude that the literature demonstrates an increasing interdisciplinary focus and deepening of existing models and empirical testing of elements of value. however, there are still limits to the benefits that can be gained from this work with oftentimes separate threads of research emerging. to address these limitations, we present a conceptual framework of the value of financial advice that pulls the various research threads together into one cohesive model. the framework seeks to increase clarity and understanding about the biopsychosocial and economic benefits for consumers and other stakeholders of financial advice. in so doing, it provides a platform for researchers to explore the holistic impact of financial advice on consumers and households and demonstrate the value this creates over time. such work would financial services review, 33(4) 136 be of interest to practitioners, business owners, policy makers and the research community, and be impactful in terms of contributing to the development of a trustworthy profession to the benefit of all. the remainder of this paper is organized as follows. the next section outlines the review methodology followed by the results structured by four key themes and culminating in a holistic value of advice framework. the last section concludes the paper incorporating challenges, limitations, and future research directions. methodology a narrative review is the chosen method for this literature review. in contrast to a systematic review which uses precise parameters and inclusion and exclusion criteria to systematically document key metrics within the literature, a narrative review allows for more flexibility and nuance (bourhis, 2017). peerreviewed and selected high-quality grey literature (i.e., government articles, phd theses, conference, and industry papers) were sourced from multiple disciplines (e.g., financial, psychological, sociological) in three databases: web of science, scopus, and google scholar. title, abstract and keyword searches were performed using the following search terms or their boolean counterparts: value, financial and advice or: 1. financial advice terms such as: help-seeking, plan, expert; 2. personal finance terms such as: capability, inclusion, literacy, behavior; and 3. wellbeing terms such as: difficult, debt, wellbeing, anxiety, peace of mind. a wide range of search terms were purposefully used to capture a comprehensive portrayal of the value of advice. searches were conducted of articles written in english from any era up until february 2023. these searches continued for five pages after the last relevant article was found, after which it was deemed unlikely that further relevant articles would appear. duplicate articles were then removed. members of the research team then manually screened the remaining articles to include only original empirical research pieces investigating the meaning or measurement of the value of personal financial advice delivered by financial advisors to individuals. the selected 332 articles were then entered into a database and thematically analyzed by the research team using key publication metrics, measurement of value (rq1), and outcomes (rq2) by themes and subthemes. a conceptual framework was then developed and refined as a result of this themesbased approach (rq3). results defining value in the context of personal professional financial advice value is a concept that is both familiar yet vague; it has different meanings to different people but also varies across contexts, disciplines, and dimensions. value has been defined as “relative worth, utility, or importance”, “something (such as a principle or quality) intrinsically valuable or desirable” and “a fair return or equivalent in goods, services, or money for something exchanged” (mirriamwebster inc., n.d.). in consumer research, holbrook (2002, p. 5) defines value as an “interactive relativistic preference experience” emphasizing that value emerges not from the product or service itself, but from the consumption experience. this experiential and dynamic nature of value suggests that its components are interdependent, multifaceted, and subject to change over time (holbrook, 2002). similarly, boztepe (2007) frames value as a derivative of the interaction between the consumer and the product or service, mediated by individual goals, needs, expectations, and emotions. this perspective also highlights the temporal dimension of value, distinguishing between pre-purchase and post-purchase evaluations. in the domain of financial advice, this distinction is particularly salient, as the perceived value of financial advice may evolve with changes in financial circumstances, market conditions, or personal priorities. pagliaro and utkus (2019) offer a tripartite model of the value of financial advice, comprising portfolio outcomes, financial outcomes, and emotional outcomes. while each dimension contributes positively to perceived value, the absence of an integrated framework limits the ability to determine which element is most influential. nonetheless, their work underscores the need for comprehensive value measures that account for both objective and subjective outcomes. macdonald et al. 137 further complicating the assessment of value in financial advice is its classification as a credence service, so consumers may lack the expertise to evaluate quality even after consumption (srinivas, 2000). the literature identifies several barriers to the accurate value assessment, including the intangibility of the service, the absence of clear quality signals, and uncertainty regarding evaluative criteria (alford & sherrell, 1996; bloom & dalpe, 1993; srinivas, 2000). these challenges are compounded by the individualized nature of financial advice which is tailored to the client’s specific goals, life stage, and financial position. although definitions of value vary, a common thread is the relationship between resource expenditure and perceived benefit. in financial advice, this relationship is inherently personalized: cost and outcomes differ across clients, for example, due to different goals, levels of wealth, or life stages, and value may be derived from tangible results (e.g., investment returns) or intangible benefits (e.g., peace of mind). thus, value in financial advice is best understood as a dynamic, experiential construct shaped by the interaction between advisor and client, and contingent upon the client’s evolving needs and perceptions. the possible outcomes of financial advice for clients are therefore wide-ranging and without a holistic approach, a complete conception of the value of financial advice for clients may not be realized. to best conceptualize the value of financial advice, it is therefore necessary to explore the multidimensional nature of financial advice and the environment in which it is delivered (e.g., loy et al., 2021; macdonald et al., 2023; marsden et al., 2011). in the next section, we provide a thematic analysis of the literature examining the value of professional financial advice to clients. this culminates in the presentation of a conceptual framework which serves to both summarize the multifaceted and interrelated aspects of financial advice and support the development of a more rigorous, coherent body of future research. thematic analysis articles were coded into themes based on the aspect of value under investigation. those with a narrow focus were coded into a single theme while others were coded into as many themes as considered appropriate to capture the breadth of the results. these themes were then grouped into four main themes (financial wellbeing, other aspects of wellbeing, quality of advice, and moderating factors), which are summarized in table 1 and form the structure for our review. table 1. themes and sub-themes financial wellbeing objective measures debt fees/costs goal achievement income/welfare benefits insurance investment performance portfolio construction risk trading activity tax efficiency wealth other aspects of wellbeing mental wellbeing physical wellbeing social wellbeing quality of advice relationship quality satisfaction with advice moderating factors advisor factors client factors environmental factors financial services review, 33(4) 138 table 1 continues on next page. table 1 continued. subjective measures financial capability (includes financial knowledge/literacy*, attitudes and behaviors) financial satisfaction financial self-efficacy financial situation* * incorporates subjective and objective measures. financial wellbeing as expected, outcomes pertaining to the client’s financial wellbeing were the most frequently identified in the literature. while financial wellbeing lacks a universally accepted definition (brüggen et al., 2017; kempson et al., 2017; nguyen, 2022), a commonly cited definition by the consumer financial protection bureau (cfpb, 2015, p. 18) describes the construct as “a state wherein a person can fully meet current and ongoing financial obligations, can feel secure in their financial future, and is able to make choices that allow enjoyment of life”. this definition acknowledges the subjective elements of financial wellbeing (regarding security and quality of life), implies that personal characteristics (such as financial knowledge and behaviors) are influential in making financial choices, and allows for measurement of objective outcomes. however, elements featuring in other financial wellbeing definitions, such as financial satisfaction (joo, 2018) and the individual’s environment and life stage (salignac et al., 2020), are missing from the cfpb (2015) definition, which adds to the ambiguity surrounding the construct. while objective measures, such as income, wealth, and debt provide tangible evidence of an individual’s financial situation, many studies rely on subjective measures such as individuals’ perceptions which are subject to recall error and social desirability bias (comerton‐forde et al., 2022). further complicating the conceptualization of financial wellbeing are variations on whether indicators of financial behavior, attitudes and knowledge (together referred to as financial capability) are included as components of financial wellbeing (e.g., kempson et al., 2017) or viewed as external influences (e.g., fu, 2020; muir et al., 2017). a lack of clarity and consistency therefore permeates both the definition and conceptualization of financial wellbeing, and few studies attempt to measure the construct holistically. instead, financial wellbeing research has been criticized for relying “on measures of convenience, rather than measures that are carefully conceived or formally developed” (comerton‐forde et al., 2022, p. 134). to illustrate, kim et al. (2003) operationalized financial wellbeing using four subjective measures adapted from porter (1990) and joo (1998): “satisfaction with personal financial situation”; “perceived financial wellness”; “feelings about current financial situation”; and “level of stress about personal finance”. financial behaviors were absent in this definition, but the five variables representing financial behaviors (“developed a plan for my financial future”; “started or increased my savings”; “reduced some of my personal debts”; “followed a budget or spending plan”; and “cut down on living expenses”) were instead found to be an external influence on financial wellbeing (kim et al., 2003). in contrast, based on the work of ladha et al. (2017), fu (2020) includes both subjective and objective measures: “balance income and expenses”; “build and maintain reserves”; “manage existing debts and has access to potential resources”; “plan and prioritize”; and “manage and recover from financial shocks”. nonetheless, ladha et al. (2017) also include a financial inclusion measure which fu (2020) argues is a predictor of financial wellbeing rather than an outcome. these inconsistencies complicate the development of a framework for measuring the value of financial advice for clients, particularly over time and across macdonald et al. 139 jurisdictions, where not only the measures are changing, but the macro-economic and regulatory environments as well. this stymies the development of a coherent and consistent body of knowledge and limits the collective impact that research can have on practice and policy. thus, in our view, the field of study would benefit from a more consistently applied financial wellbeing framework. objective measures historically, research on financial advice has concentrated on its objective financial benefits, particularly its impact on client portfolio construction and investment performance. these outcomes are not only tangible but also align with the expectations of many clients, making them a logical focus for empirical analysis. these measures can also elicit more reliable results than subjective data which is open to bias. however, this narrow emphasis overlooks the broader value clients derive from the advice process, including behavioral, emotional and strategic benefits. moreover, the reliability of findings on financial outcomes is complicated by contextual factors. much of the extant literature originates from jurisdictions such as the us and germany where commission-based remuneration models, shown to induce advisor conflicts of interest and impact advice quality, were prevalent during the study periods. it is perhaps unsurprising then that the literature on portfolio construction and investment performance of advised portfolios has reported more negative than positive results. evidence suggests that advised portfolios are more diversified (e.g., bluethgen, gintschel, et al., 2008; hackethal et al., 2012; kramer, 2012; marsden et al., 2011) but this increased diversification is costly for clients when achieved through investments made into higher cost mutual funds consistent with advisors’ compensation incentives (e.g., hackethal et al., 2012; kramer, 2012; linnainmaa et al., 2021). it is rare to find reports that advised clients allocate significantly more of their portfolio to cheaper index funds (the study by chalmers and reuter, 2012 is one example). while diversification helps clients avoid home bias (e.g., bluethgen, gintschel, et al., 2008), several studies (e.g., gerhardt & hackethal, 2009; kramer, 2012; marsden et al., 2011) suggest this is due to naïve (or simple) forms of diversification, namely, through higher mutual fund allocation. additionally, there is little evidence that more diversified client portfolios improve investment performance. few studies (e.g., grable & chatterjee, 2014) find significantly better returns in advised individuals while more report minimal changes to risk-adjusted returns (e.g., gerhardt & hackethal, 2009; kramer, 2012; marsden et al., 2011). marsden et al. (2011) argue that the lack of strong evidence for advised clients outperforming the non-advised is a function of research typically being conducted during periods of market growth, where value rises across many markets and industries, and maintaining discipline is relatively easy. advised clients are found to trade more which again is consistent with commission-based renumeration models in some jurisdictions (hackethal et al., 2012), however, linnainmaa et al. (2021) find this is how advisors manage their own portfolios. marsden et al. (2011) also emphasizes that advised investors were more likely than nonadvised to increase risk and portfolio size during a market downturn, which is consistent with advisor observations of some clients taking advantage of asset price reductions during the covid-19 pandemic (loy et al., 2021). reforms to transition the advice industry to a profession, including the removal of commissions, increased education requirements, and a focus on putting clients’ interests first, may partly explain this shift to more positive results. further research is needed to confirm this trend. subjective measures while an individual’s objective financial situation is arguably a vital element to capture in the measurement of financial wellbeing, some researchers maintain that a client’s subjective view of their finances is even more important (e.g., marsden et al., 2011; prawitz et al., 2006). subjective outcomes from financial advice have received less attention in the literature but are gaining traction with most articles having been published in the last decade. measurement largely involves surveys conducted in the us, uk, and australia examining client views after receiving advice, rather than the views of advised versus unadvised clients, as is typically the case with objective measures. working with an advisor is financial services review, 33(4) 140 associated with many positive perceptions, including perceived improvement in financial circumstances (theodos et al., 2015), improved financial confidence (coredata research, 2020), preparedness for retirement (kim et al., 2003), a greater sense of control of and less time worrying about one’s financial situation (irving et al., 2011; pleasence et al., 2007), and improved self-efficacy concerning financial decision-making (kim et al., 2003; marsden et al., 2011). these positive perceptions following advice are consistent across sociodemographic groups including low-income earners (brackertz, 2014), university staff (grable & joo, 2003), college students (britt et al., 2015), and widows (rehl et al., 2016). financial advice is positively associated with improving clients’ financial engagement and behavioral biases. advised clients spend more time searching for financial information (lee & cho, 2005), learning about financial topics (marsden et al., 2011), and striving to improve financial planning behaviors (ford et al., 2020). advisors reduce home bias (see objective measures section), trend chasing (d’acunto et al., 2019), and the disposition effect of managed portfolios (hoechle et al., 2018). marsden et al. (2011) suggest that a key role of financial advisors is to help clients ‘stay the course’ during volatile or declining markets. consistent with this argument, grable and chatterjee (2014) found that clients who sought advice from a financial advisor prior to the global financial crisis (gfc) experienced less wealth volatility during the gfc than non-advised investors. similarly, winchester et al. (2011) found that advised investors were more than one-and-a-half times more likely to maintain a long-term investment strategy than non-advised investors during a market downturn. financially advised individuals also exhibit several important financial behaviors pertaining to savings, planning and budgeting, credit, and buying behaviors. advised clients are more likely to save (gerhardt & hackethal, 2009), budget (bagwell, 2000; brackertz, 2014; theodos et al., 2015) and plan for retirement (kim & garman, 2003; marsden et al., 2011; tang & lachance, 2012). advice recipients are also more likely to reduce their use of credit cards (elliehausen et al., 2007; woodhead et al., 2017), pay their bills on time (cfpb, 2016), compare their options before taking out loan products (fan & chatterjee, 2017) and prioritize long-term goals over current consumption (bagwell, 2000; brackertz, 2014; martin jr et al., 2016). the literature examining the relationship between financial advice and subjective attitudes and behaviors is therefore largely positive and clear in terms of the benefits it provides. the emerging literature examining the relationship between financial advice and financial literacy is less distinct. while a comprehensive review is outside the scope of this paper, our review finds most studies report no or limited improvements in objective financial literacy from financial advice (e.g., britt et al., 2015; tang & lachance, 2012; theodos et al, 2015). these null findings are purportedly the result of several factors, including poor attendance at coaching sessions (theodos et al., 2015); lack of participant motivation, engagement, willpower, or patience (hung & yoong, 2013; theodos et al., 2015); and the temporal nature of the advice (with brief ad-hoc education interventions having higher rates of financial literacy decay) (migliavacca, 2020). we also contend that studies using ‘have obtained financial advice’ (e.g., tang & lachance, 2012) are poor proxies for financial advice due to lack of information and context on what the advice entailed and may therefore also contribute to the null findings. although null or limited associations have been reported between financial advice and objective financial literacy, positive effects have been consistently reported on subjective financial literacy, attitudes and behaviors (e.g., brackertz, 2014; britt et al., 2015; kim & garman, 2003; theodos et al., 2015). these mixed results have lead theodos et al. (2015, p. 151) to conclude that advice effects “may derive directly from behavioral change or skill formation, rather than through increases in more abstract knowledge.” individual factors have also been influential and add context, with trust, willingness to learn, and the length of the clientadvisor relationship having significant positive effects on financial literacy (hung & yoong, 2013; linh et al., 2016; migliavacca, 2020). however, more recent research provides evidence for the argument that individual factors alone are insufficient to improve financial capability (financial literacy, attitudes, and behavior) and wellbeing. fu (2020) used an institutionalist approach, emphasizing that individuals need both the ability (financial macdonald et al. 141 knowledge, attitude, behavior, motivation) and the opportunity (through access to financial products and institutions) to act, thus advocating for a wider context to more definitively measure financial capability and predict the factors that enhance or inhibit financial wellbeing. from the research presented, we conclude that financial advice can enhance individuals’ financial knowledge, attitudes, and behaviors, but that unmeasured psychosocial and environmental factors can affect whether and to what extent financial literacy translates into lasting behavioral change and financial wellbeing. this highlights the need for longitudinal research across jurisdictions that is premised on a conceptual framework of the value of advice to rigorously examine its impact. other aspects of wellbeing the second theme identified in this review pertains to other aspects of wellbeing beyond financial wellbeing, including mental, physical, and social wellbeing. the world health organization (who) (2023, para. 3) defines mental wellbeing as “a state in which an individual acknowledges their own abilities, can cope with everyday life stresses, work productively, and contribute to their community”. physical wellbeing encompasses not only the absence of chronic disease, but feeling healthy and energetic, combined with healthy sleep and regular physical activity (zemtsov & osipova, 2016). social wellbeing is described as a degree of involvement in social networks, together with social support that is both received and perceived to be available when called upon (holt-lunstad et al., 2010). all three aspects of wellbeing are considered integral to overall health (who, 2023) and are covered in varying degrees in the financial advice literature, with mental wellbeing having been examined the most. insights into the mental, physical, and social wellbeing outcomes of financial advice have largely come from studies on debt and welfare advice. research into non-financial wellbeing outcomes dates back to 2000, including a doctoral thesis examining work and personal outcomes of credit counselling (bagwell, 2000) and a peer-reviewed paper on health improvements linked to welfare benefit advice (abbott & hobby, 2000). most of this literature has emerged in the last decade, with a relatively high proportion published since 2018. it predominantly appears in industry and government publications rather than peerreviewed journals. for example, a wave of research into the effect of government-funded debt advice programs emerged from the uk in the 2000-2010 decade. value was operationalized through qualitative and quantitative interviews, with reports highlighting mental wellbeing benefits (reduced stress and anxiety) together with financial wellbeing benefits (feeling more relieved, confident and in control of their finances) following advice (day et al., 2008; pleasence et al., 2007; turley & white, 2007). the literature in this domain is primarily descriptive in its analysis, and overwhelmingly positive in reporting improvements in mental wellbeing. one paper detailed a study involving a randomized controlled trial (rct) which experimentally examined the benefit of a national debtline using comparisons with a control group where no such advice was offered. this study was reportedly the first time such a methodology had been used in the debt advice area (pleasence et al., 2007). this experiment was designed to explore the effectiveness of debt advice on a range of life circumstances. while the rct demonstrated statistical improvements in financial circumstances following advice, improvements in general health, parenting, and family relationships could not be established through the rct even though these findings were reported qualitatively in an accompanying study. the researchers blamed a lack of statistical power (due to participant attrition and the small sample size) for these null findings. apart from this study, few industry or government publications have attempted to quantify the value of advice using more rigorous research designs. peer-reviewed studies also largely focus on wellbeing-related outcomes of debt and welfare advice and also primarily report positive results. encouragingly, some well-designed peerreviewed studies exist which provide stronger empirical support for the more descriptive findings outlined in the industry reports. for instance, a uk study examined the effect of welfare advice using patients attending a medical facility and a quasi-experimental controlled design with follow-up data collected at three months (woodhead et al., 2017). the financial services review, 33(4) 142 regression analyses controlled for age, gender, and ethnicity. compared to the control group, women and black recipients reported mental health improvements, and wellbeing scores increased for those who received positive welfare outcomes. similarly, abbott et al. (2006) performed analyses on groups of individuals with different financial outcomes of welfare advice, finding that those with an income increase had significant improvements in mental health and emotional role functioning at 12-month follow-up, compared to those without an income increase. these results provide further evidence of the association between financial gains and increased mental health and wellbeing. while much of the literature focuses on mental wellbeing outcomes, a smaller subset examines physical and social wellbeing and the impact of financial advice on work performance. the results regarding physical wellbeing are generally, but not always, affirmative. for instance, an intervention study assessed health outcomes following active and non-active credit counselling clients and found active members of the program improved in physical health markers over 18 months (kim et al., 2003). a similar study reported improved health for debt management clients over a two-year period (o'neill et al., 2006). however, abbott et al. (2006) only found bodily pain symptoms had improved in the six months following welfare advice, with no significant improvement in three other physical health markers (general health, vitality, and physical role functioning). the authors suggested that a longer follow-up period might be necessary to engender statistically significant results in physical health, “given the cumulative effect of deprivation on health across the life course” (p. 6). encouragingly, studies also tend to examine and find positive links between professional financial advice and multiple non-financial wellbeing outcomes. studies cover a breadth of analytical strategies from basic frequency data (e.g., coredata research, 2020; irving et al., 2011) to more advanced probit models (europe economics, 2018). the research conducted by europe economics (2018) is an example of a considered, large-scale, mixed-methods design which detailed significant effects in mental wellbeing (depression and panic attacks) and physical wellbeing (arthritis, back pain, and chronic obstructive pulmonary disorder), as well as qualitative evidence on employmentrelated outcomes (increased productivity). furthermore, the researchers also sought to quantify the economic impact of financial advice using a combination of estimates pertaining to healthcare systems and productivity costs. the estimated benefit of clients receiving debt advice equated to an aggregated social gain in mental health and work-related outcomes of £154-307 million per annum (europe economics, 2018). the focus on debt and welfare advice persists in more recent literature (pollard et al., 2020; zeamer, 2020). however, some new studies take a different approach to similar issues, by incorporating psychological constructs and theories into the investigation of financial advice and the various non-financial wellbeing outcomes. for example, ford et al. (2020) employed a qualitative observational design to explore the role of financial couples therapy on relationships and help-seeking behavior, with reports that therapy produced reduced anxieties, better communication and deeper relationships between the couples, and a stronger desire to improve financial behaviors. jackson et al. (2022) followed a cohort of men at risk of suicide due to financial distress who participated in a psychosocial intervention. at six-month follow-up, participants recorded a 49% reduction in depression scores, a 55% reduction in suicidal ideation, and a 26% increase in financial self-efficacy scores (a financial wellbeing measure). these findings reflect a broader shift towards understanding financial advice not only as a driver of economic outcomes, but also as a catalyst for psychological, physical, and relational wellbeing. quality of advice the third theme, quality of advice, has been examined least often in the professional financial advice space. it contains three subthemes: relationship quality; satisfaction with advice; and trust. in government literature, quality of advice has been described as a threeway interaction, encapsulating “the provision of high quality, accessible and affordable financial advice for retail clients” (australian government, 2022, p. 1). in the academic literature, the (obvious) assertion has been made that quality financial advice “should aim to macdonald et al. 143 enhance investor utility” (bluethgen, meyer, et al., 2008, p. 3), however, the credence good that is expert financial advice means the typical client lacks the necessary knowledge or information to accurately assess the quality of the advice provided, even ex-post (bluethgen, meyer, et al., 2008; srinivas, 2000). thus, it is ironic that research into the quality of financial advice has often focused on basic subjective measures and reporting of individuals’ perceptions about the service they receive in the short term or at a single point in time, more so than comprehensive or holistic measures and analyses over the medium to long term. these more subjective accounts are primarily found in industry and government publications, often highlighting high satisfaction and trust in advisors, along with a low likelihood of switching advisor (e.g., farr et al., 2018; madamba & utkus, 2017). grable and joo (2003) also report high satisfaction of advised clients with their advisor through recommendations to family and friends. the non-advised are used in some studies for comparative purposes, with these individuals typically expressing concerns over quality, trust, and affordability, and a perception of being capable of making their own financial decisions (asic, 2019; himawan, 2020). this has led some researchers to advocate for lowcost automated online advice options (farr et al., 2018), in line with the rise of robo-advice. other research directions further investigate psychosocial constructs from the financial wellbeing literature. while trust is an important element of the client-advisor relationship, manders (2021) finds that an examination of personality traits can be used to identify selfprotection measures to counteract the information asymmetry in the client-advisor relationship and thus empower advised clients to improve their evaluation of the quality of advice. additional data and modelling are required to empirically test the impact on advice outcomes, including financial wellbeing. government publications are comparatively abundant, particularly in australia, reflecting the country’s history of financial advice scandals and subsequent regulatory reforms (e.g., asic, 2019; australian government, 2022). more recently, however, there has been a noticeable rise in publications that adopt a more holistic approach to evaluating advice quality or appear in peer-reviewed journals, or both. for instance, an australian study examined the components of advisor-client relationship quality and the extent to which these changed over time. all components of relationship quality were measured with scales previously determined to be reliable and valid. the key determinants of relationship quality were reported to be trust, engagement, and commitment, with many client perceptions of relationship quality increasing over time, despite the instability and regulatory changes within the sector (hunt et al., 2022). another study from italy used a structural equation model to test the impact of length of relationship and ability to understand clients’ emotional associations with money on four relationship outcomes: satisfaction, trust, loyalty, and referral propensity. the better the advisor was at understanding their client’s emotional associations with money, the better the relationship and length of relationship also positively impacted relationship outcomes (lozza et al., 2022). these studies signal a growing academic interest in advice quality as a multidimensional construct central to understanding the value of financial advice, paving the way for more comprehensive conceptualizations of value. for instance, madamba and utkus (2017) employed a mixed-methods approach using a sample of advised clients to identify and quantify the components of trust in financial advisors. they identified three core dimensions of trust, functional, emotional, and ethical, with emotional trust having the greatest influence (53%), followed by ethical (30%), and functional trust (17%). acting in a client’s best interest and advocating for the client were key drivers of an advisor’s trustworthiness. in contrast, bluethgen, meyer, et al. (2008) assessed advice quality through advisors’ return predictions of common asset categories (a measure of overconfidence) and portfolio recommendations to fictitious clients. their findings showed that advice quality improved when advisors were more sophisticated (rational) and less reliant on commission-based remuneration, while advisor experience and education had no significant effect. while also not significant in the survey and interview responses of advisors reported in loy et al. (2023), likert-scale results indicated that increases in education levels and continuing education requirements were perceived by financial services review, 33(4) 144 advisors as having a moderately positive impact on their ability to deliver value to clients. participants believed these benefits would become more pronounced over time, as the profession adjusts to regulatory changes. this suggests a divergence between empirical measures of advice quality and advisors’ perceptions, highlighting the need for further research into how professional standards translate into client outcomes. importantly, loy et al. (2023) also examined education and experience as moderating factors in advisor behavior within the client-professional relationship, providing a foundation for exploring how individual advisor characteristics influence the delivery and perception of advice quality and shape client outcomes. moderating factors as outlined in the results section, determining the value of advice is complicated by the heterogeneity of investor circumstances. similarly, advisor factors, such as an advisor’s education, experience, or conflicts of interest, and environmental factors, including the business model (e.g., advice delivery via faceto-face, robo-advice, video conferencing, telephone etc.), financial markets and the regulatory or cultural environment, interact with client factors and the client-advisor relationship, influencing value through the quality of advice and client outcomes. for example, the impact of new products in financial markets and the state of the markets, the impact of regulation on administrative burden, time and cost of advice leading to stakeholders’ concerns about the accessibility of advice and value for money, and government, regulator and professional body interactions over professional and education standards, are all designed to increase capability to deliver quality advice to add value for clients and protect them from harm (e.g., asic, 2019; loy et al., 2021, 2023). consideration of these moderating factors, along with how value is conceptualized, is therefore vital for a holistic understanding of the value of financial advice (macdonald et al., 2023). an important dimension of advisor factors is the set of actions advisors undertake throughout the financial planning process. conceptualizing value as an evaluation of actions rather than of products or “things” (graeber, 2001) is particularly relevant in financial advice, where advisors engage in diagnosing client needs, setting goals, formulating strategies, implementing plans and monitoring and adjusting those plans over time. these actions represent the core of professional service delivery and are distinct from the outcomes clients may experience. it is critical to differentiate between advisor actions and client outcomes, as the latter are often shaped by factors beyond the advisor’s control. for example, wealth accumulation and investment returns depend heavily on the client’s investment horizon, risk tolerance, and external constraints. as such, financial outcomes alone may not consistently reflect the value of advice (marsden et al., 2011). instead, value may more reliably be found in the advisor’s behavioral coaching and expert guidance, which helps clients stay disciplined and emotionally grounded, particularly during market volatility (loy et al., 2023; pagliaro & utkus, 2019). despite this, loy et al. (2023) found that advisors rated behavioral coaching as the least beneficial financial service they provided, even though open-ended responses revealed its importance in boosting client confidence and contributing to both financial and mental wellbeing. client outcomes are also shaped by client characteristics, including psychological and emotional factors. advisors increasingly recognize their dual role as both technical experts and informal counsellors, even though they may lack formal training in the latter (anthes & lee, 2002; loy et al., 2021). financial psychology research suggests that advisors who can adapt their style to accommodate clients’ emotional needs are more likely to build trust and foster meaningful relationships (anthes and lee, 2002). this highlights the social-emotional context of advice, where value is co-created through relational elements such as emotional, social, and informational support, not just financial expertise. these insights highlight the intricate interplay between advisor actions, client characteristics, and relational dynamics in shaping perceived value. this complexity underscores the need to examine moderating factors such as advisor education, experience, and adaptability, that influence how advice is delivered and received macdonald et al. 145 across diverse client contexts. thus, the value of financial advice is multifaceted, contextdependent, and open to various conceptualizations. tables 2 and 3 serve as exemplars summarizing the findings from the results section, addressing respectively: (i) what constitutes value, i.e., how it is operationalized and measured (rq1); and (ii) whether value is realized (rq2). while this narrative review identifies individual factors that moderate the value of advice, quantifying their interaction or cumulative impact on consumer value remains challenging. in response to this gap, the thematic analysis section introduces a conceptual framework to address rq3. this framework synthesizes existing literature to broaden and deepen understanding of the value of financial advice and aims to stimulate interdisciplinary inquiry and empirical validation through longitudinal, multijurisdictional research. financial services review, 33(4) 146 table 2. measures of value: some examples theme sub-theme author (year) measures financial wellbeing objective measures debt elliehausen et al. (2007) credit score: summary measure of credit worthiness as provided by the empirica risk score, involving six measures of credit use (change in revolving debt; change in the number of credit cards with positive balances; change in credit card utilization as a percentage of credit limit; change in consumer debt; change in total debt; change in total number of accounts with positive balances). goal achievement pagliaro and utkus (2019) probabilistic forecast: estimate of the probability of achieving a financial goal or wealth target (secure retirement) at the end of a specified period. income/welfare benefits various measures are generally reported vaguely or not at all. insurance blanchett (2019) survey: one item assessing whether a household had life insurance at least equal to the total wage income of the household (yes/no). investment performance bluethgen, gintschel, et al. (2008) statistic: tracking error with a benchmark portfolio. hackethal et al. (2012) statistics: sharp ratio; monthly log portfolio returns; net of direct costs, as calculated by the method of dietz (1968); jensen’s alpha; alpha four factor model. marsden, zick and mayer (2011) statistics: annual rate of account growth over the past 1, 2 and 3 years. portfolio construction bluethgen, gintschel, et al. (2008) statistics: equity share of portfolio; share of equity held as mutual funds; ratio of international equity to total equity assets. hackethal et al. (2012) statistic: average share of directly held equity. risk bluethgen, gintschel, et al. (2008) statistic: standard deviation of portfolio returns, derived from the variation of monthly gross portfolio returns as calculated by dietz (1968). hackethal et al. (2012) statistic: variance of monthly portfolio returns. marsden, zick and mayer (2011) survey: increased the investment risk involving existing and new money in retirement account(s) (yes/no). table 2 continued on next page. macdonald et al. 147 table 2 continued. theme sub-theme author (year) measures trading activity bluethgen, gintschel, et al. (2008) statistics: transaction costs; annual sales; purchase turnover. hackethal et al. (2012) statistics: number of purchases per month scaled by account volume; monthly account turnover. wealth grable and chatterjee (2014) statistics: average change in wealth; standard deviation of wealth changes; risk adjusted wealth volatility (using a modified modigliani measure (m2)). marsden, zick and mayer (2011) survey: self-reported retirement savings; emergency fund adequacy; total defined contribution account values; total supplemental account values. subjective measures financial situation littlechild (2013) survey: how confident are you that you will reach your primary financial goal? how much clarity do you have around your financial goals? marsden, zick and mayer (2011) survey: retirement confidence: confidence in having enough money to live comfortably throughout retirement years. theodos et al., 2015 survey: level of financial stress (two items: 1-7 scale. item 1: rating own level of financial stress. item 2: how often individuals wanted to go out but could not afford to). satisfaction with present financial situation (1-7 scale). confidence in ability to achieve financial goals (0100% scale). confidence in ability to make ends meet in an emergency (0-100% scale). progress toward improving household’s financial security/ ability to take care of family/ live more comfortably (0-100% scale). qualitative: specific interview questions not provided. financial attitudes marsden, zick and mayer (2011) survey: could offer an estimate of asset allocation (yes/no); increased the amount of time spent learning about financial topics (yes/no). table 2 continued on next page. financial services review, 33(4) 148 table 2 continued. theme sub-theme author (year) measures financial behaviors marsden, zick and mayer (2011) survey: have established long-term goals and are working to fulfill them; has calculated retirement needs; has a supplement retirement account; frequency of reviewing retirement account statements. after receiving financial advice: increased the amount of money saved regularly; increased the amount of money owed; increased the age of expected retirement. behavioral responses to the recent economic crisis: percentage invested internationally in 2009; number of asset classes in 2009; percentage allocation to stocks in 2009; percentage of total holdings in small or mid-cap funds in 2009; percentage in target retirement date funds in 2009; percentage in actively managed funds in 2009. financial knowledge britt et al. (2015) survey: before and after comparison of objective financial knowledge measured by awarding 1 point per correct answer to the following questions: a) you may obtain a free copy of your credit report each year (true); (b) higher insurance deductibles lead to lower insurance premiums (true); (c) an annuity is a contract issued by a financial institution that guarantees a series of payments over a lifetime (true); (d) a mutual fund is an investment company that invests its shareholders’ money in a diversified portfolio of securities (true); (e) social security and company pension plans are sufficient to meet retirement needs (false); and (f) over 20 years, you will earn more money if you invest in bonds compared to stocks (false). other aspects of wellbeing mental wellbeing abbott et al. (2006) survey: sf-36: domains relating to emotional role functioning; and mental health. brackertz (2014) survey: i feel more positive about the future. my mental and/or emotional wellbeing has improved (1-5 agree-disagree scale). day et al. (2008) interviews: topic guide provided: how well clients feel they are coping, compared with previous financial situation. any wider benefits of the advice process – health, employment, family, etc. irving et al. (2011) survey: assess before and after seeking financial advice: mental health and well-being. woodhead et al. (2017) survey: general health questionnaire (ghq-12): assesses symptoms of mental distress. shortened warwick-edinburgh mental well-being scale (swemwbs): assesses positive mental health. table 2 continued on next page. macdonald et al. 149 table 2 continued. theme sub-theme author (year) measures physical wellbeing abbot et al. (2006) survey: sf-36: domains relating to physical functioning; physical role functioning; bodily pain; self-reported general health; and vitality. brackertz (2014) survey: my physical health has improved (1-5 agree-disagree scale). day et al. (2008) interviews: topic guide provided: any wider benefits of the advice process – health, employment, family, etc. irving et al. (2011) survey: assess before and after seeking financial advice: physical health and well-being. kim et al. (2003) survey: 4 items: self-reported health status; experience of health problems; comparison of physical health with people their age; experience of stress (noted this is a mental wellbeing measure, however for the purposes of this study, all four items were summed to make a composite health score). social wellbeing abbot et al. (2006) survey: sf-36: the domain relating to social functioning. brackertz (2014) survey: my relationship with my family and friends have improved. my relationships with my children has improved. (1-5 agree-disagree scale). day et al. (2008) interviews: topic guide provided: any wider benefits of the advice process – health, employment, family, etc. irving et al. (2011) survey: assessed before and after seeking financial advice: social activities; social relationships; family activities; family relationships. loy et al. (2021) interviews: what are the key things that financial advisors do that help clients navigate a crisis situation? what services do clients need from their financial adviser during a crisis? has the covid-19 pandemic had any impact on your ability to support and provide advice to clients? table 2 continued on next page. financial services review, 33(4) 150 table 2 continued. theme sub-theme author (year) measures quality of advice hunt et al. (2022) survey: separate surveys for clients and financial planners. domains include: trust (11 items); engagement (9 items); relationship quality (17 items); commitment (10 items); empowerment (10 items); ownership (10 items) and client activity (9 items). exact items not reported but original scales are referenced. irving et al. (2011) survey: a series of 12 agree-disagree statements on the financial planning process (e.g., provides a professional service; listens to my ideas and concerns; takes control of planning my financial future for me). loy et al. (2021) interviews: are there any aspects of the client-advisor relationship that add value for clients during times of crisis? how do the relationship aspects of providing advice compare to technical advice and services you provide in terms of importance? how have you managed relationships with clients during the covid-19 pandemic? montmarquette & viennot-briot (2015) econometric model: positive relationship between having a fa for at least 4 years and the level of financial assets compared to a non-advised participant. moderating factors advisor factors irving et al. (2011) survey: a series of 13 agree-disagree statements on financial planner characteristics (e.g., demonstrates a high level of knowledge and expertise in financial planning; provides objective advice; is knowledgeable about the performance of investments). client factors loy et al. (2021) interviews: how have your clients been more/less engaged with their finances and the advice process during covid-19? how have your clients reacted to your firm’s approach to technology during the covid-19 pandemic? environmental factors regulation & professional standards loy et al. (2023) survey: a series of 8 questions on the impact of regulatory change on an advisor’s ability to add value for clients in the short-term and long-term concerning education requirements, continuing professional development, introduction of a code of ethics, ban on commissions (0-10 agree-disagree scale). macdonald et al. 151 table 3. the value of financial advice: a brief summary of some outcomes theme nature of the outcomes author (year) results financial wellbeing objective measures results tend to be more negative than positive bluethgen, gintschel, et al. (2008) advised clients have a significantly larger allocation to mutual funds than non-advised investors, but increased diversification is achieved by investing in higher cost mutual funds. britt et al. (2015) limited improvement in clients’ objective financial knowledge. hackethal et al. (2012) advised accounts have on average lower net returns and inferior risk-return trade-offs (sharpe ratios). kramer (2012) the portfolios of advised investors are better diversified and carry significantly less idiosyncratic risk. subjective measures mostly positive results kim and garman (2003) clients reported feeling more confident making investment decisions. lusardi and mitchell (2011) advised clients were more likely to stick to their retirement plan. marsden, zick and mayer (2011) working with an advisor is related to several important financial behaviours, including goal setting, calculation of retirement needs, accumulation of emergency funds, and positive behavioral responses economic crisis. pleasence et al. (2007) the majority of clients felt more knowledgeable about financial matters after receiving debt advice. rehl et al. (2016) working with a financial advisor significantly improved widows’ financial confidence in their current and future financial situations. table 3 continued on next page. financial services review, 33(4) 152 table 3 continued. theme nature of the outcomes author (year) results other aspects of wellbeing mental wellbeing overwhelmingly positive results abbott et al. (2006) clients who had received an increase in welfare benefits had significant improvements in mental health and emotional role functioning at the 12-month follow-up, compared to those who did not receive an income increase. archuleta et al. (2020) brief solution-focused financial therapy helped reduced financial anxiety. coredata (2020) the non-advised are more likely to experience negative impacts on relationships, mental and physical wellbeing. the advised report improvements in many areas of their lives, including their family relationships, social lives, physical health, and work satisfaction. these results are witnessed in greater degrees in men compared to women. day et al. (2008) clients reported feeling less stressed and relieved after receiving advice, and more optimistic about managing their future finances. pleasence et al. (2007) clients felt their circumstances had changed for the better following advice; for example, they reported less time worrying about their finances; less chance of ‘difficulty living normally’; and improved self-reported mental and physical health. physical wellbeing mostly positive results, some inconclusive results abbot et al. (2006) bodily pain symptoms improved following advice however no significant improvements were reported in general health, vitality, and physical role functioning. kim et al. (2003) in a study comparing active and non-active credit counselling clients, active clients improved in self-reported physical health over an 18-month period. credit counselling had an indirect positive effect on health via (reduced) financial stressor events and (improved) perceived financial wellbeing. o’neill et al. (2006) clients report improved health over time and following advice. improved health is associated with lower incidence of negative financial events, lower financial distress, and better financial wellbeing. table 3 continued on next page. macdonald et al. 153 table 3 continued. theme nature of the outcomes author (year) results social wellbeing mostly positive results, some inconclusive results brackertz (2014) nearly half (45%) of advised clients reported improvements in relationships with children, family and friends. irving et al. (2011) while new clients reported more non-tangible benefits of advice, both new and existing clients reported positive benefits to family relationships, family activities, and social relationships. loy et al. (2021) advisors discussed social benefits of the advisor-client relationship, such as offering friendship and listening to their concerns about the changing landscape during covid-19 pleasance et al. (2007) qualitative results indicated positive improvements in parenting and family relationships, however these results could not be established through a randomized controlled trial (rct) due to a lack of statistical power. quality of advice overwhelmingly positive results hunt et al. (2022) relationship quality in financial advice relationship increased between 2009 and 2016, despite turmoil and change within the sector. trust, engagement, and commitment were the key components of the client-advisor relationship according to clients. madamba & utkus (2017) trust is positively associated with investor age and tenure with advisor. emotional trust had the great impact on overall trust. high trust leads to high satisfaction and low likelihood of switching advisors. trust can be undermined by poor investment performance, communication issues, neglect, and advisors’ self-serving behavior. montmarquette & viennot-briot (2015) compared to a similar long-tenured (15 years or more) non-advised participant, the advised participant has 2.73 times more financial assets. the result is too large to be related to stock selection alone and is more plausibly associated with greater savings associated with having a financial advisor. table 3 continued on next page. financial services review, 33(4) 154 table 3 continued. theme nature of the outcomes author (year) results moderating factors advisor factors mixed results bluethgen, meyer, et al. (2008) advisor education was shown to increase adviser overconfidence yet provided better riskadjusted returns for clients. hackethal et al. (2012) higher investment turnover contributes to lower returns, consistent with commissions being the main source of advisor income. hoechle et al. (2018) a generally independent advisor added value, while conflicts of interest negatively impacted value. client factors mixed results bucher-koenen et al. (2021) women (but not men) with higher financial aptitude reject advisor recommendations more frequently. loy et al. (2023) a lack of financial literacy in clients limits value because the consumer is unaware of what an advisor can help with, has a narrow view of what financial advice involves (e.g., investments and retirement saving) and/or are unaware of the benefits of advice. environmental factors mixed results asic (2019) 37% of participants agreed they could do just as well as a financial advisor by managing their own finances. some participants who had a good relationship with their own advisor still expressed cynicism about the broader financial advice industry. loy et al. (2023) financial advisors generally expressed negative sentiments towards regulation and linked it to the increased cost to serve clients. however, regulatory changes aimed at increasing advisers’ skills (education and continuing professional development requirements) were rated as having a positive impact on advisors’ ability to add value for clients, with participants viewing these changes as positive in the long-term. macdonald et al. 155 conceptual framework this review highlights the complexity and multidimensional nature of value in the context of personal professional financial advice. figures 1, 2, and 3 address rq3 by synthesizing the literature and offering conceptual guidance for future research. these figures explore the value of financial advice through three lenses: 1) the financial advice environment; 2) the client-professional relationship; and 3) client wellbeing outcomes, informed by social support theory as a novel conceptualization of value. figure 1 illustrates the financial advice environment placing the client-advisor relationship at its core and situating it within a broader context of key environmental factors and stakeholders. this model is particularly relevant for understanding the opportunities and constraints faced by australian financial advisors, especially in light of ongoing regulatory reforms and the timeframe over which changes have taken place and continue to occur. importantly, this model is adaptable to other jurisdictions, offering a comparative lens for international research. figure 1. the financial advice environment note: reproduced from authors’ own work in loy et al., 2023 figure 2 shifts focus to the client-advisor dyad, presenting financial advice as an interactive, relational, and reflective process. thematic analysis underpinning this review reveals that client outcomes extend beyond financial metrics (subjective and objective measures), to encompass mental, physical, and social wellbeing. the findings of this narrative review suggest that a well-rounded financial advisor often functions similarly to a support provider. accordingly, value is conceptualized as a dynamic and evolving process that, amongst other moderating variables, is shaped by advisor actions throughout the financial planning process. social support theory provides a compelling framework for examining the dual role of financial advisors as both technical experts and emotional supports. it emphasizes the importance of timely, adequate support in stressful situations, the strength of dyadic relationships in fostering resilience and growth, and the interconnectedness of wellbeing domains (e.g., cohen & wills, 1985; feeney & collins, 2015; holt-lunstad, 2018). figure 2 overlays advisor actions onto cohen & wills’ (1985) four functional domains of social support: 1) emotional and esteem support – offering empathy and validation; 2) social support – spending time to alleviate financial stress; 3) informational and appraisal support – providing facts, feedback and educational guidance to enhance financial knowledge and capability; and 4) instrumental support – delivering financial products and services. administrative broader environment constraint s business financial services review, 33(4) 156 figure 2. the dynamics of the client-advisor relationship note: reproduced from authors’ own work in loy et al., 2023 to realize the full potential of the financial advice profession, particularly in addressing unmet needs such as retirement planning, greater research attention must be directed towards client outcomes. while figure 2 presents an interdisciplinary model of wellbeing and support, figure 3 extends this approach by offering a model that encapsulates the multifaceted nature of value in financial advice. it recognizes that a single advisor action may influence multiple wellbeing domains and that these outcomes are interrelated, with direct effects in one area potentially producing indirect effects in others. this narrative review lays the foundation for future research by offering recommendations in the next section. these aim to guide further exploration of existing value constructs using innovative methodologies and to expand the literature through new conceptualizations of which figure 3 serves as a key exemplar. figure 3. the value of financial advice: a conceptual framework note: adapted from authors’ own work in loy et al., 2021, 2023 conclusion in an increasingly complex and informationrich world, individuals face heightened uncertainty and decision fatigue, prompting a growing reliance on expert advice. financial advisors are widely recognized as experts; however, consumer trust in the long-term macdonald et al. 157 outcomes of financial advice remains limited. historically, the value of financial advice has been assessed primarily through objective financial metrics such as investment performance and returns. yet, while extant literature is extensive, it is fragmented and lacks a cohesive synthesis particularly in relation to how value is conceptualized and measured across diverse contexts and over time. this narrative review identifies a critical gap in the literature: the absence of a holistic framework for evaluating the value of professional financial advice. without such a framework, the collective body of research struggles to inform policy and practice meaningfully. moreover, the literature has paid insufficient attention to the advisor’s role in shaping client outcomes through relational and emotional support. emerging research suggests that subjective benefits such as increased confidence, financial capability, and overall wellbeing may be equally, if not more, important than traditional financial metrics. to address these gaps, this review synthesizes peer-reviewed and high-quality grey literature to provide a comprehensive overview of financial advice. it offers a conceptual framework that integrates advisor actions, client characteristics, and environmental factors, with the aim of guiding more rigorous and coherent research to inform stakeholders including financial services professionals, licensees and policymakers, on strategies to substantiate the perceived and actual value of financial advice and thus, more credible information and transparency for consumers. by engaging a broader and more diverse client base, the financial advice sector can unlock significant and social benefits. as with all narrative reviews, this paper is subject to limitations. although article selection and quality assessment were conducted systematically, the breadth of the literature (over 330 articles) necessitated a thematic approach, which may have excluded some nuances. whilst care was taken to ensure objectivity, narrative synthesis inherently involves subjective judgement (bourhis, 2017). nonetheless, this review deepens and extends understanding of scholarship on the value of professional financial advice and the conceptual framework offers a valuable foundation for future empirical work. to advance the research agenda, several key areas warrant attention. (1) longitudinal research designs longitudinal research design is recommended (e.g., macdonald et al., 2023; pleasence et al., 2007) to overcome several limitations of existing work to progress the research agenda on value. the credence nature of financial advice means its value cannot often be assessed until long after it is delivered. longitudinal studies are essential to capture delayed outcomes, particularly those related to client wellbeing. such designs can illuminate both moderating factors (what influences the advice process and outcomes) and mediating factors (how the various aspects of advice lead to wellbeing outcomes), enabling a more nuanced understanding of value. (2) causal inference and methodological rigor much of the existing literature is observational, limiting the ability to draw causal conclusions. future research should incorporate experimental or quasi-experimental designs to better isolate the effects of financial advice and advisor actions. (3) addressing self-selection and endogeneity a persistent challenge in advice research is disentangling whether clients benefit from advice because they seek it, or whether those who seek advice are already better positioned to benefit. future studies should explore counterfactual scenarios and alternative comparison groups to better understand the true impact of advice. (4) contextualizing value value is inherently subjective and varies across clients and life stages. for example, higher investment returns may not be valuable to a risk-averse retiree. research should explore how value is interpreted relative to individual client goals, preferences, and circumstances. (5) testing the conceptual framework while this paper suggests not all types of support are necessary, the framework presented posits that clients derive the greatest value when the advisor actions align across all domains of social support. future research should empirically test this proposition examining how combinations of advisor behaviors influence different dimensions of client wellbeing. financial services review, 33(4) 158 (6) adapting to emerging trends the rise of robo advice, evolving regulatory environments, and shifting client demographics present new challenges and opportunities. these trends should be incorporated into future studies to assess the robustness and adaptability of the conceptual framework. in conclusion, the value of financial advice is multifaceted, dynamic, and deeply personal. a one-size-fits-all approach to measuring value is insufficient. the conceptual framework developed in this review provides a foundation for more rigorous, interdisciplinary, and context-sensitive research. by embracing complexity and focusing on both tangible and intangible outcomes, future scholarship can better capture the true impact of financial advice and contribute to more effective policy, practice, and client engagement. references abbott, s., & hobby, l. (2000). welfare benefits advice in primary care: evidence of improvements in health. public health, 114(5), 324-327. abbott, s., hobby, l., & cotter, s. (2006). what is the impact on individual health of services in general practice settings which offer welfare benefits advice? health and social care in the community, 14(1), 1-8. https://doi.org/10.1111/j.13652524.2005.00582.x adviser ratings. (2019, december 2). financial advice landscape report: early bird snapshot. adviser ratings. https://pro.adviserratings.com.au/docs/ landscape-report-early-bird.pdf adviser ratings. (2022). australian financial adviser landscape 2022. https://www.ardata.com.au/wpcontent/uploads/2022/05/afalandsca pe2022-ab_r2.pdf alford, b. l., & sherrell, d. l. (1996). the role of affect in consumer satisfaction judgments of credence-based-services. journal of business research, 37, 7184. anthes, w. l., & lee, s. a. (2002). the financial psychology of four lifechanging events. journal of financial planning, 15(5), 76-85. asic. (2019). financial advice: what consumers really think (report no. 627). https://asic.gov.au/regulatoryresources/find-a-document/reports/rep627-financial-advice-what-consumersreally-think/ australian government. (2022). quality of advice review: final report. https://treasury.gov.au/sites/default/file s/2023-01/p2023-358632.pdf bagwell, d. c. (2000). work and personal financial outcomes of credit counseling clients [doctoral dissertation, virginia polytechnic institute and state university]. virginia tech electronic theses and dissertations. https://vtechworks.lib.vt.edu/handle/10 919/29222 blanchett, d. m. (2019). financially sound households use financial planners, not transactional advisers. journal of financial planning, 32(4), 30-40. bloom, p. n., & dalpe, m. p. (1993). the proactive professional. marketing management, 2(1), 26-35. bluethgen, r., gintschel, a., hackethal, a., & mueller, a. (2008). financial advice and individual investors' portfolios. ssrn. https://doi.org/10.2139/ssrn.968197 bluethgen, r., meyer, s., & hackethal, a. (2008). high-quality financial advice wanted! citeseer. https://citeseerx.ist.psu.edu/document? repid=rep1&type=pdf&doi=f1a08e41c 681ec8e93579ba480c1b3cf40a74e2b bourhis, j. (2017). narrative literature review. in m. allen (ed.), the sage encyclopedia of communication research methods (3rd ed., pp. 10761077). sage publications. https://doi.org/10.4135/978148338141 1 boztepe, s. (2007). user value: competing theories and models. international journal of design, 1(2), 55-63. brackertz, n. (2014). the impact of financial counselling on alleviating financial stress in low income households: a national australian empirical study. macdonald et al. 159 social policy and society, 13(3), 389407. britt, s. l., canale, a., fernatt, f., stutz, k., & tibbetts, r. (2015). financial stress and financial counseling: helping college students. journal of financial counseling and planning, 26(2), 172186. brüggen, e. c., hogreve, j., holmlund, m., kabadayi, s., & löfgren, m. (2017). financial well-being: a conceptualization and research agenda. journal of business research, 79, 228237. https://doi.org/10.1016/j.jbusres.2017. 03.013 bucher-koenen, t., hackethal, a., koenen, j., & laudenbach, c. (2021). gender differences in financial advice (working paper). safe. http://dx.doi.org/10.2139/ssrn.2572961 cfp board. (2023). 2023 cfp board consumer sentiment survey: cost of living. https://www.cfp.net/-/media/files/cfpboard/news/2023/financial-planningand-consumer-sentimentdeck.pdf?la=en&hash=76b388e7f3f2 3ad8d5ca48c4a54e9fad chalmers, j. m. r., & reuter, j. (2012). what is the impact of financial advisors on retirement portfolio choices and outcomes? ssrn. http://dx.doi.org/10.2139/ssrn.2078536 comerton‐forde, c., de new, j., salamanca, n., ribar, d. c., nicastro, a., & ross, j. (2022). measuring financial wellbeing with self‐reported and bank record data. economic record, 98(321), 133-151. https://doi.org/10.1111/14754932.12664 consumer financial protection bureau. (2015). financial well-being: the goal of financial education. cfpb. https://files.consumerfinance.gov/f/201 501_cfpb_report_financial-wellbeing.pdf consumer financial protection bureau. (2016). financial coaching: a strategy to improve financial well-being. cfpb. https://files.consumerfinance.gov/f/doc uments/102016_cfpb_financial_coach ing_strategy_to_improve_financial_ well-being.pdf coredata research. (2020). trust in advice: research on the value of financial advice. financial advice new zealand. https://financialadvice.nz/wpcontent/uploads/2020/10/financialadvice-nz-trust-in-advice-doublepage.pdf day, l., collard, s. b., & hay, h. (2008). money advice outreach evaluation: qualitative outcomes for clients. legal services commission. http://www.infohub.moneyadvicetrust. org/content_files/files/debtoutreachout comesclients.pdf dietz, p. o. (1968). components of a measurement model: rate of return, risk, and timing. journal of finance 23, 267–275. elliehausen, g., christopher lundquist, e., & staten, m. e. (2007). the impact of credit counseling on subsequent borrower behavior. journal of consumer affairs, 41(1), 1-28. https://doi.org/10.1111/j.17456606.2006.00066.x europe economics. (2018). the economic impact of debt advice: a report for the money advice service. https://moneyandpensionsservice.org.u k/wpcontent/uploads/2021/03/economicimpact-of-debt-advice-main-report.pdf fan, l., & chatterjee, s. (2017). borrowing decision of households: an examination of the information search process. journal of financial counseling and planning, 28(1), 95106. https://doi.org/10.1891/10523073.28.1.95 farr, d., older, c., ripley, e., & weir, j. (2018). the changing shape of the consumer market for advice. ignition house and critical research. https://www.fca.org.uk/publication/res earch/famr-interim-consumerresearch-report-2018.pdf feeney, b. c., & collins, n. l. (2015). a new look at social support: a theoretical https://dx.doi.org/10.2139/ssrn.2572961 https://moneyandpensionsservice.org.uk/wp-content/uploads/2021/03/economic-impact-of-debt-advice-main-report.pdf https://moneyandpensionsservice.org.uk/wp-content/uploads/2021/03/economic-impact-of-debt-advice-main-report.pdf https://moneyandpensionsservice.org.uk/wp-content/uploads/2021/03/economic-impact-of-debt-advice-main-report.pdf https://moneyandpensionsservice.org.uk/wp-content/uploads/2021/03/economic-impact-of-debt-advice-main-report.pdf financial services review, 33(4) 160 perspective on thriving through relationships. personality and social psychology review, 19(2), 113-147. ford, m. r., ross, d. b., grable, j., & degraff, a. (2020). examining the role of financial therapy on relationship outcomes and help-seeking behavior. contemporary family therapy, 42, 5567. https://doi.org/10.1007/s10591019-09511-y fu, j. (2020). ability or opportunity to act: what shapes financial well-being? world development, 128, 104843. https://doi.org/10.1016/j.worlddev.201 9.104843 gerhardt, r., & hackethal, a. (2009). the influence of financial advisors on household portfolios: a study on private investors switching to financial advice. ssrn. http://dx.doi.org/10.2139/ssrn.1343607 grable, j. e., & chatterjee, s. (2014). reducing wealth volatility: the value of financial advice as measured by zeta. journal of financial planning, 27(8), 45-51. grable, j. e., & joo, s.-h. (2003). a snapshot view of the help-seeking market. journal of financial planning, 16(3), 88-94. graeber, d. (2001). toward an anthropolical theory of value: the false coin of our own dreams. palgrave. https://doi.org/10.1057/978-0-31229906-4 hackethal, a., haliassos, m., & jappelli, t. (2012). financial advisors: a case of babysitters? journal of banking & finance, 36(2), 509-524. https://doi.org/10.1016/j.jbankfin.2011 .08.008 himawan, a. (2020). peace of mind: understanding the non-financial benefits of financial advice. international longevity centre–uk. https://ilcuk.org.uk/peace-of-mind/ hoechle, d., ruenzi, s., schaub, n., & schmid, m. (2017). the impact of financial advice on trade performance and behavioral biases. review of finance, 21(2), 871-910. https://doi.org/10.1093/rof/rfw032 holbrook, m. b. (2002). introduction to consumer value. in m. b. holbrook (ed.), consumer value: a framework for analysis and research (pp. 1-28). routledge. holt-lunstad, j. (2018). why social relationships are important for physical health: a systems approach to understanding and modifying risk and protection. annual review of psychology, 69, 437-458. holt-lunstad, j., smith, t., & layton, j. (2010). social relationships and mortality risk: a meta-analytic review. plos medicine, 7(7), 1000316. https://doi.org/10.1371/journal.pmed.1 000316 hung, a. a., & yoong, j. k. (2013). asking for help: survey and experimental evidence on financial advice and behavior change. in o. s. mitchell & k. smetters (eds.) the market for retirement financial advice (pp. 182212). oxford university press. hunt, k. h., brimble, m., & freudenberg, b. (2022). a move in the right direction: client relationships in financial advice. the journal of wealth management, 25(1), 50-81. irving, k., gallery, g., gallery, n., & newton, c. (2011). “i can’t get no satisfaction”… or can i? a study of satisfaction with financial planning and client well-being. jassa: the finsia journal of applied finance, 2, 36–44. jackson, j., farr, m., birnie, k., davies, p., mamluk, l., o’brien, m., spencer, j., morgan, r., costello, c., & smith, j. (2022). preventing male suicide through a psychosocial intervention that provides psychological support and tackles financial difficulties: a mixed method evaluation. bmc psychiatry, 22, 333. https://doi.org/10.1186/s12888-02203973-5 joo, s.-h. (1998). personal financial wellness and worker job productivity (doctoral dissertation, virginia polytechnic institute and state university). virginia tech electronic theses and https://ilcuk.org.uk/peace-of-mind/ macdonald et al. 161 dissertations. http://hdl.handle.net/10919/30519 joo, s.-h. (2008). personal financial wellness. in j. j. xiao (ed.), handbook of consumer finance research (pp. 21-34). springer. https://ahmadladhani.files.wordpress.c om/2009/10/handbook_of_cfr.pdf kempson, e., finney, a., & poppe, c. (2017). financial well-being: a conceptual model and preliminary analysis. sifo consumption research norway. https://hdl.handle.net/20.500.12199/59 68 kim, j., & garman, e. t. (2003). financial education and advice changes worker attitudes and behaviors. journal of compensation and benefits, 19(5), 713. kim, j., garman, e. t., & sorhaindo, b. (2003). relationships among credit counseling clients' financial wellbeing, financial behaviors, financial stressor events, and health. journal of financial counseling and planning, 14(2), 75-87. kramer, m. m. (2012). financial advice and individual investor portfolio performance. financial management, 41(2), 395-428. ladha, t., asrow, k., parker, s., rhyne, e., & kelly, s. (2017). beyond financial inclusion: financial health as a global framework. center for financial services innovation. https://content.centerforfinancialinclus ion.org/wpcontent/uploads/sites/2/2017/03/finhe althglobal-final.2017.04.11.pdf lee, j., & cho, j. (2005). consumers’ use of information intermediaries and the impact on their information search behavior in the financial market. journal of consumer affairs, 39(1), 95120. https://doi.org/10.1111/j.17456606.2005.00005.x linnainmaa, j. t., melzer, b. t., & previtero, a. (2021). the misguided beliefs of financial advisors. the journal of finance, 76(2), 587-621. https://doi.org/10.1111/jofi.12995 littlechild, j. (2013). measuring adviser value in a fee-based world. vanguard. loy, e., macdonald, k. l., brimble, m., & wildman, k. l. (2021). the value of professional financial advice for consumers in a crisis: experiences of financial advisers during the covid19 pandemic. griffith university. https://www.griffith.edu.au/__data/ass ets/pdf_file/0025/1417804/professiona l-advice-in-a-crisis_amp-griffithreport.pdf loy, e., macdonald, k. l., brimble, m., & wildman, k. l. (2023). the value of financial advice: perspectives of australian financial advisers. griffith university. https://www.griffith.edu.au/__data/ass ets/pdf_file/0027/2070594/value-ofadvice-amp-report.pdf lozza, e., castiglioni, c., bonanomi, a., & poli, f. (2022). money as a symbol in the relationship between financial advisors and their clients: a dyadic study. international journal of bank marketing, 40(7), 613-630. https://doi.org/10.1108/ijbm-04-20210124 macdonald, k. l., loy, e., brimble, m., & wildman, k. l. (2023). the value of personal professional financial advice to clients: a systematic quantitative literature review. accounting & finance, 63(4), 4399-4429. https://doi.org/10.1111/acfi.13099 madamba, a., & utkus, s. p. (2017). trust and financial advice. vanguard. https://static.vgcontent.info/crp/intl/au w/docs/resources/adviser/vanguard_re search_trust_and_financial_advice.pd f manders, j. (2021). let the buyer beware: an analysis of self-protection measures and traits in client-advisor relationships (master’s thesis, griffith university). griffith university. marsden, m., zick, c. d., & mayer, r. n. (2011). the value of seeking financial advice. journal of family and economic issues, 32(4), 625-643. https://doi.org/10.1007/s10834-0119258-z http://hdl.handle.net/10919/30519 https://content.centerforfinancialinclusion.org/wp-content/uploads/sites/2/2017/03/finhealthglobal-final.2017.04.11.pdf https://content.centerforfinancialinclusion.org/wp-content/uploads/sites/2/2017/03/finhealthglobal-final.2017.04.11.pdf https://content.centerforfinancialinclusion.org/wp-content/uploads/sites/2/2017/03/finhealthglobal-final.2017.04.11.pdf https://content.centerforfinancialinclusion.org/wp-content/uploads/sites/2/2017/03/finhealthglobal-final.2017.04.11.pdf https://static.vgcontent.info/crp/intl/auw/docs/resources/adviser/vanguard_research_trust_and_financial_advice.pdf https://static.vgcontent.info/crp/intl/auw/docs/resources/adviser/vanguard_research_trust_and_financial_advice.pdf https://static.vgcontent.info/crp/intl/auw/docs/resources/adviser/vanguard_research_trust_and_financial_advice.pdf https://static.vgcontent.info/crp/intl/auw/docs/resources/adviser/vanguard_research_trust_and_financial_advice.pdf financial services review, 33(4) 162 martin jr, t. k., guillemette, m. a., & browning, c. m. (2016). do retirement planning strategies alter the effect of time preference on retirement wealth? applied economics letters, 23(14), 1003-1005. https://doi.org/10.1080/13504851.2015 .1128068 migliavacca, m. (2020). keep your customer knowledgeable: financial advisors as educators. the european journal of finance, 26(4-5), 402-419. https://doi.org/10.1080/1351847x.201 9.1700148 mirriam-webster inc. (n.d.). value. https://www.merriamwebster.com/dictionary/value montmarquette, c., & viennot-briot, n. (2015). the value of financial advice. annals of economics and finance, 16(1), 69-94. muir, k., hamilton, m., noone, j., marjolin, a., salignac, f., & saunders, p. (2017). exploring financial wellbeing in the australian context. centre for social impact & social policy research centre, university of new south wales. https://www.csi.edu.au/media/explorin g_financial_wellbeing_in_the_austra lian_context_final_report.pdf nguyen, s. m. (2022). visualization and bibliometric analysis on the research of financial wellbeing. international journal of advanced and applied sciences, 9(3), 10-18. https://doi.org/10.21833/ijaas.2022.03. 002 oecd. (2020). oecd pensions outlook 2020. oecd publishing. https://doi.org/10.1787/67ede41b-en o'neill, b., prawitz, a., sorhaindo, b., kim, j., & garman, e. t. (2006). changes in health, negative financial events, and financial distress/financial well-being for debt management program clients. journal of financial counseling and planning, 17(2), 46-63. https://ssrn.com/abstract=2232121 pagliaro, c. a., & utkus, s. p. (2019). assessing the value of advice. vanguard. https://institutional.vanguard.com/cont ent/dam/inst/iigtransformation/insights/pdf/assessingthe-value-of-advice.pdf pleasence, p., buck, a., balmer, n. j., & williams, k. (2007). a helping hand: the impact of debt advice on people’s lives. legal services research centre. https://webarchive.nationalarchives.go v.uk/ukgwa/20100210214545/http://lsr c.org.uk/publications/impact.pdf pollard, c. m., booth, s., louth, j., mackenzie, c., & goodwin‐smith, i. (2020). “i'd be sleeping in the park, i reckon”: lived experience of using financial counselling services in south australia. economic papers, 39(4), 353-366. https://doi.org/10.1111/17593441.12298 porter, n. m. (1990). testing a model of financial well-being (publication no. 9028076) [doctoral dissertation, virginia polytechnic institute and state university]. virginia tech electronic theses and dissertations. https://vtechworks.lib.vt.edu/handle/10 919/39899 prawitz, a., garman, e. t., sorhaindo, b., o'neill, b., kim, j., & drentea, p. (2006). incharge financial distress/financial well-being scale: development, administration, and score interpretation. journal of financial counseling and planning, 17(1), 34-50. rehl, k. m., moor, c. c., leitz, l. y., & grable, j. e. (2016). widows' voices: the value of financial planning. journal of financial service professionals, 70(1), 53-60. salignac, f., hamilton, m., noone, j., marjolin, a., & muir, k. 2020. conceptualizing financial wellbeing: an ecological lifecourse approach. journal of happiness studies, 21, 1581–1602. srinivas, v. k. (2000). individual investors and financial advice: a model of adviceseeking behavior in the financial planning context (publication no. 9967099) [doctoral dissertation, rutgers university–newark]. https://doi.org/10.1787/67ede41b-en https://institutional.vanguard.com/content/dam/inst/iig-transformation/insights/pdf/assessing-the-value-of-advice.pdf https://institutional.vanguard.com/content/dam/inst/iig-transformation/insights/pdf/assessing-the-value-of-advice.pdf https://institutional.vanguard.com/content/dam/inst/iig-transformation/insights/pdf/assessing-the-value-of-advice.pdf https://institutional.vanguard.com/content/dam/inst/iig-transformation/insights/pdf/assessing-the-value-of-advice.pdf https://vtechworks.lib.vt.edu/handle/10919/39899 https://vtechworks.lib.vt.edu/handle/10919/39899 macdonald et al. 163 proquest. https://www.proquest.com/docview/30 4658049 tang, n., & lachance, m.-e. (2012). financial advice: what about low-income consumers? journal of personal finance, 11(2), 121-158. tharp, d. (2021). financial advisor fee trends and the fee compression mirage. kitces. https://www.kitces.com/blog/financialadvisor-average-fee-2020-aum-hourlycomprehensive-financial-plan-cost/ theodos, b., simms, m., treskon, m., stacy, c., brash, r., emam, d., daniels, r., & collazos, j. (2015). an evaluation of the impacts and implementation approaches of financial coaching programs. urban institute. https://www.urban.org/research/public ation/evaluation-impacts-andimplementation-approaches-financialcoaching-programs turley, c., & white, c. (2007). assessing the impact of advice for people with debt problems. legal services research centre. http://www.infohub.moneyadvicetrust. org/content_files/files/assessing_the_i mpact_of_advice.pdf westermann, s., niblock, s. j., harrison, j. l., & kortt, m. a. (2020). financial advice seeking: a review of the barriers and benefits. economic papers, 39(4), 367388. https://doi.org/10.1111/17593441.12294 winchester, d. d., huston, s. j., & finke, m. s. (2011). investor prudence and the role of financial advice. journal of financial service professionals, 65(4), 43-51. woodhead, c., khondoker, m., lomas, r., & raine, r. (2017). impact of co-located welfare advice in healthcare settings: prospective quasi-experimental controlled study. the british journal of psychiatry, 211(6), 388-395. https://doi.org/10.1192/bjp.bp.117.202 713 world health organization. (2023). the global health observatory: health and wellbeing. who. https://www.who.int/data/gho/data/maj or-themes/health-and-well-being zeamer, c. a. (2020). using discourse analysis to evaluate the effectiveness of financial counseling. journal of financial counseling and planning, 32(1), 330-341. https://doi.org/10.1891/jfcp-1800081 zemtsov, a. a., & osipova, t. y. (2016). financial wellbeing as a type of human wellbeing: theoretical review. the european proceedings of social & behavioural sciences, 7, 385-392. https://www.proquest.com/docview/304658049 https://www.proquest.com/docview/304658049 https://www.kitces.com/blog/financial-advisor-average-fee-2020-aum-hourly-comprehensive-financial-plan-cost/ https://www.kitces.com/blog/financial-advisor-average-fee-2020-aum-hourly-comprehensive-financial-plan-cost/ https://www.kitces.com/blog/financial-advisor-average-fee-2020-aum-hourly-comprehensive-financial-plan-cost/ https://www.urban.org/research/publication/evaluation-impacts-and-implementation-approaches-financial-coaching-programs https://www.urban.org/research/publication/evaluation-impacts-and-implementation-approaches-financial-coaching-programs https://www.urban.org/research/publication/evaluation-impacts-and-implementation-approaches-financial-coaching-programs https://www.urban.org/research/publication/evaluation-impacts-and-implementation-approaches-financial-coaching-programs 87 effective financial education strategies: empowering students with personal application richard stebbins1 & diana quito2 abstract financial education is most effective when it is both personalized and applicable to learners’ lives. instruction that reflects individual backgrounds, interests, and learning styles fosters engagement, ownership, and improved outcomes. when financial concepts are taught in real-world contexts, learners are more likely to recognize their relevance and apply them meaningfully. this paper integrates five established learning theories: bloom’s taxonomy, behaviorism, constructivism, experiential learning, and social learning theory. these frameworks support differentiated instruction in financial education and inform the design of inclusive, learner-centered pedagogy. the study’s primary contribution is its application of these theories to specific financial topics, including investment planning, retirement strategies, financial counseling, and estate planning. assignments and instructional tools provided in the appendix illustrate how theory-informed strategies can be implemented in practice. these classroom-tested activities promote financial literacy, support behavioral change, and enhance instructional effectiveness across diverse learner populations. by bridging theoretical models with applied learning experiences, the paper offers educators a practical framework for improving financial education. creative commons license this work is licensed under a creative commons attribution-noncommercial 4.0 license recommended citation stebbins, r. & quito, d. (2025). effective financial education strategies: empowering students with personal application. financial services review, 33(4), 87-105. introduction in contemporary educational environments, the importance of differentiated instruction has become increasingly evident. learners enter classrooms with varied backgrounds, learning styles, and cognitive profiles, requiring pedagogical approaches that are responsive to individual needs (eikeland & ohna, 2022; tomlinson, 2017). tomlinson (2001) emphasizes the value of differentiated instruction in mixedability classrooms, advocating for teaching strategies that accommodate differences in student readiness, interests, and learning preferences. desimone (2009) further identifies 1 corresponding author (rstebbins@ches.ua.edu). university of alabama, tuscaloosa, alabama, usa. 2 rollins college, winter park, florida, usa. professional development as a key mechanism through which educators acquire the skills necessary for effective instruction. together, these perspectives underscore the broader educational shift toward learner-centered models that prioritize adapting to the needs of the learner (dole, 2016; mccombs, 2001) this shift is particularly relevant in the context of financial education. the ultimate goal of financial education, according to the consumer financial protection bureau, is financial well-being (cfpb, 2023). however, financial education has been associated with a range of financial behaviors and outcomes https://creativecommons.org/licenses/by-nc/4.0/ mailto:rstebbins@ches.ua.edu financial services review, 33(4) 88 (alexander & mcelreath, 1999; fernandes et al., 2014; urban et al., 2020). numerous studies have linked it to improvements in financial literacy, decision-making, and overall financial wellbeing (hilgert et al., 2003; lusardi & mitchell, 2014; mandell & klein, 2009). at the same time, conflicting findings suggest that its impact on financial literacy varies considerably across different contexts and populations (hensley, 2015; jump$tart coalition for personal financial literacy, 2006). the need for effective financial education has become increasingly salient in light of evolving economic conditions and the growing complexity of personal financial decision-making (choung et al., 2023; giesecke & waschik, 2025; kalaycı & serra-garcia, 2016; ullah et al., 2024). scholars have called for a comprehensive framework and attempted to assess financial education programs (fox et al., 2005; hathaway & khatiwada, 2008; lyons, 2005, lyons et al., 2006), yet such efforts must also account for learner diversity and instructional design. however, there are no accepted standards of excellence for evaluating program effectiveness, and no single curriculum adequately meets the needs of all learners (hensley, 2015; mccormick, 2009). the effectiveness of financial education is influenced by individual differences in learning styles and preferences (cassidy, 2004; ziernwald et al., 2022) along with instructional design (salas-velasco et al., 2021). this paper responds to these challenges by examining how established learning theories can inform the instructional design of differentiated financial education. the integration of behaviorism, constructivism, social learning theory, experiential learning, and bloom’s taxonomy provides a framework for understanding how individuals learn and apply financial concepts. application of these theories can improve how financial education is taught and how learners can engage with financial content, develop skills, and apply knowledge in real-world contexts. the motivation for this study is both practical and pedagogical. it seeks to bridge the gap between theory and practice by demonstrating how learning theories can be operationalized in the context of financial education. through detailed examples and instructional tools, the paper aims to empower educators with strategies that enhance the relevance, accessibility, and effectiveness of financial education across a range of topics, including investment planning, retirement strategies, financial counseling, and estate planning. theories of learning several theories of learning offer insights into how individuals acquire knowledge, skills, and behaviors. some prominent theories include: bloom’s taxonomy bloom's taxonomy, a well-established framework in educational psychology, offers a structured approach to cognitive learning (bloom et al., 1956). it encompasses various levels of thinking skills, from basic to advanced, facilitating a comprehensive understanding of educational objectives and teaching methodologies. a) knowledge: this level involves recalling previously learned material, including facts, terms, basic concepts, and generalities. for instance, in financial planning, knowledge might entail remembering key investment terms such as "dividend yield" or "asset allocation." b) comprehension: at this level, understanding of facts and ideas is demonstrated by explaining their meaning, interpreting them, or describing main ideas. in a financial context, comprehension could involve explaining the significance of economic indicators or interpreting financial statements to assess a company's performance. c) application: application involves utilizing acquired knowledge in new or novel situations to solve problems. for example, in financial planning, applying knowledge might mean using investment principles to construct a diversified portfolio tailored to a client's risk tolerance and financial goals. d) analysis: analysis entails breaking down information and materials into parts to examine details and relationships. in financial planning, this could involve analyzing market trends, dissecting financial statements to identify strengths and stebbins & quito 89 weaknesses, or evaluating the impact of various economic factors on investment decisions. e) synthesis: synthesis involves compiling information in different ways, building structures or patterns from diverse elements, and integrating parts to form a coherent whole. in financial planning, synthesis might involve creating a comprehensive financial plan that integrates various components such as budgeting, retirement strategies, estate planning, and investment strategies. f) evaluation: this level entails presenting opinions by making judgments about the value and merit of ideas and materials. in financial planning, evaluation might involve assessing the effectiveness of investment strategies, evaluating the risk-return tradeoffs of different financial products, or critiquing the ethical implications of certain financial decisions. instructors can use bloom’s taxonomy to scaffold financial education assignments, starting with basic recall of financial terms and progressing to evaluation of complex financial strategies. for example, a budgeting module might begin with identifying expense categories (knowledge), then move to analyzing spending patterns (analysis), and culminate in students designing and defending a personalized budget plan (synthesis and evaluation). this progression helps students build confidence and competence while allowing educators to assess learning at multiple cognitive levels. this scaffolding structure can also be applied to exams. research shows improved performance on exams that begin with lower-level questions (perlini et al., 1998) with further consideration for online exams using question banks to randomize questions in groups based on difficulty level with all students progressing through that scaffolding (becker et al., 2022). extending beyond bloom’s cognitive domain, warmath and zimmerman (2019) argue for the need to apply bloom’s two other domains of knowledge (affective and psychomotor) to conceptualize financial literacy. according to bloom et al. (1956), the cognitive domain incorporates students’ acquisition of knowledge, and the affective domain focuses on the development of students’ interest, attitudes, and values. the psychomotor domain focuses on objectives related to motor skills and physical development. in their work, warmath and zimmerman drew on these established educational domains to develop a formative scale measuring financial literacy: skills needed for a financial decision (psychomotor), self-efficacy to make the financial decision (affective), and the capacity to develop financial knowledge to apply in future situations (cognitive and psychomotor). warmath and zimmerman’s (2019) use of bloom’s cognitive domains for their composite approach to financial literacy can be leveraged to innovate financial education. from a pedagogical standpoint, educators can supplement teaching financial concepts with building skills and improving self-efficacy. scaffolded learning across bloom’s domains of knowledge can begin with teaching foundational content such as budgeting and credit and moving into hands-on practice through real-world financial decisionmaking simulations such as buying a car or a home (lusardi, 2019). for the affective domain, developing individual financial goals can help to increase confidence by building personal connections with the content. behaviorism behaviorism posits that learning occurs through observable behaviors that are shaped by environmental stimuli and reinforcement. central to behaviorism is the concept of conditioning, where behaviors are either reinforced through rewards or diminished through punishment. the work of psychologists such as ivan pavlov, who studied classical conditioning (pavlov, 1927), and b.f. skinner, known for his research on operant conditioning (skinner, 1938), contributed to the development of behaviorist theory (watson, 2017). instructors can apply behaviorist principles by designing structured reinforcement systems that encourage consistent financial habits. for example, a classroom savings challenge can reward students for tracking expenses over time, reinforcing delayed gratification and goal setting. educators might also use token economies, where students earn points for completing financial financial services review, 33(4) 90 tasks such as creating a budget, analyzing a credit report, or negotiating a mock bill. these points can be exchanged for privileges or recognition, reinforcing desired behaviors. additionally, frequent low-stakes assessments with immediate feedback help reinforce foundational knowledge and correct misconceptions before they become ingrained. constructivism constructivism suggests that individuals construct their understanding of the world through active engagement with their environment (bruner, 2009). learning is viewed as a process of constructing meaning to realistic and relevant contexts through the integration of new information with existing knowledge and experiences (honebein, 1996). the importance of social interaction with multiple perspectives, collaboration, and hands-on experiences in facilitating learning within a constructivist framework (bruner, 2009; honebein, 1996). in financial education, learners can collaborate with peers to solve financial problems, share experiences, and construct their understanding of financial concepts. real-world case studies provide authentic contexts for learners to explore and apply their knowledge, fostering collaborative learning environments where individuals can learn from each other's perspectives and experiences. offering opportunities for personalized exploration, such as through senior theses aligned with individual interests, can enhance engagement and learning outcomes (fogarty & mayo, 1999). experiential learning theory experiential learning theory suggests that learning is a cyclical process involving concrete experiences, reflective observation, abstract conceptualization, and active experimentation (kolb, 1975). learners engage in continuous cycles of learning by doing, reflecting, and applying their experiences to new situations. (burke, 2013; kolb et al., 2014). previous research shows that students engaging in handson activities perform significantly better on standardized tests (stohr-hunt, 1996). in financial education, learners actively experiment with financial concepts and reflect on their experiences to enhance their financial decision-making abilities. a real-world example of experiential learning in action comes from credit card fees that consumers learn to avoid (agarwal et al., 2008) and is further supported by respondents reporting learning more from personal financial experiences than formal financial education (hilgert et al., 2003). in institutions of higher education, student managed investments have been used (grinder et al., 1999) as has opportunities to work with real clients (eyssell, 1999). social learning theory social learning theory posits that learning occurs through observation, imitation, and modeling of others' behaviors. developed by albert bandura, social learning theory emphasizes the role of social context, role models, and reinforcement in shaping behavior (bandura, 1977). financial education begins early with financial socialization. here, a child’s financial knowledge base and relationship with money are influenced by the financial behaviors of parents with longterm effects (danes, 1994; gudmunson & danes, 2011; kim & chatterjee, 2013). gutter et al.’s (2010) study found that in additional to parental modeling, peers influence college students’ financial behaviors and attitudes. for instance, peers can model positive money management behaviors and sensible financial decisions. application in the classroom can look like designing group activities that can reinforce good habits such as collaboratively designing budgets. additionally, homework or projects can include parent or family member interviews to discuss financial decisions to reinforce financial concepts learned in class while leveraging the students’ household knowledge (gonzález et al., 1993). from theory to practice the learning theories discussed above provide a general framework for understanding how students engage with financial education in the classroom. these theories offer insight into how learners process information, develop skills, and build financial knowledge. while this theoretical foundation is important, its full value is realized when applied to specific instructional practices. the next section demonstrates how these theories stebbins & quito 91 inform the design of targeted assignments and learning tools across key financial topics. these topics include investments, retirement planning, financial counseling, and estate planning. each one is accompanied by examples that show how theory-driven strategies are used to support student learning. these examples are grounded in actual classroom activities and instructional materials, all of which are included in the appendix. this transition from general theory to applied practice is intended to help educators implement effective, learner-centered financial education. application for specific financial topics investments personal investment planning provides a clear opportunity to apply constructivist, experiential, and social learning theories. these frameworks guide the design of activities that help students build personalized investment strategies, reflect on financial decisions, and learn collaboratively through peer engagement. investment planning is a critical aspect of personal finance that involves strategic decisionmaking to grow wealth and achieve long-term financial goals. learners are introduced to investment basics through interactive lessons that use real-world examples to illustrate key concepts. they engage in hands-on exercises to build investment portfolios tailored to their financial objectives and risk tolerance levels. students determine their individual risk tolerance assessments to determine their comfort levels with risk and discuss personalized diversification recommendations. through personalized investment plan reviews and feedback, individuals receive guidance on optimizing their investment strategies. appendix a outlines an assignment requiring students to develop a comprehensive investment plan for a hypothetical client and a personalized five-year financial strategy. this project incorporates constructivism, experiential, and social learning theory elements. constructivism applied: • active learning and reflection: students construct their understanding by developing investment plans tailored to real-life scenarios. by reflecting on personalized feedback, they integrate new knowledge with prior experiences. • problem-solving in context: creating a plan for a hypothetical client and themselves allows learners to apply financial concepts to practical situations, fostering deeper comprehension. • collaborative learning: engaging with peers and instructors through discussions and feedback sessions enhances their understanding by exchanging diverse perspectives. experiential learning applied: • hands-on experience: students actively create investment plans, which provides concrete experiences that reinforce learning. • real-world application: designing plans for a hypothetical client and themselves encourages learners to apply theoretical concepts in practical, real-life contexts. • reflection: after receiving feedback, students reflect on their strategies, which deepens their understanding and helps refine their investment approaches. • feedback loop: continuous feedback from instructors fosters iterative learning, allowing students to adjust and improve their plans based on real-world scenarios. this hands-on project exemplifies constructivism by emphasizing active engagement and collaboration, leading to meaningful learning experiences. social learning theory applied: • observation and modeling: students learn by observing peer approaches and instructor demonstrations, enhancing their understanding of effective investment strategies. • peer collaboration: group discussions and peer reviews facilitate shared learning experiences, allowing students to learn from each other’s insights and feedback. • feedback and support: constructive feedback from both peers and instructors creates a supportive environment that financial services review, 33(4) 92 encourages learners to refine their strategies based on collective input. • emulation of best practices: by engaging with exemplars students adopt effective behaviors and decision-making processes in their investment planning. understanding the different types of investments is essential, and learners explore this topic through case studies that examine various investment vehicles and their risk-return profiles. group discussions on investment selection strategies foster collaboration and the sharing of insights among peers. professor-led feedback on investment choices ensures that learners gain a deeper understanding of the implications of their investment decisions and can make informed choices aligned with their financial goals and risk tolerance levels. retirement planning retirement planning instruction benefits from integrating constructivist and experiential learning approaches. these theories support the use of realistic scenarios, hands-on simulations, and reflective exercises that help students connect retirement concepts to their personal financial goals. financial education on retirement planning emphasizes the importance of understanding various retirement income sources and optimizing their mix for financial security in retirement. real-world case studies and group discussions reflecting on what retirement means on a personal level provide elements of constructivism and experiential learning. learners explore different approaches to retirement savings through case studies showcasing diverse strategies. they engage in group discussions to identify challenges and develop solutions related to retirement goals. through the analysis of retirement income sources such as social security, employersponsored retirement plans, individual investment options, and annuities students learn strategies to plan for their individual vision of retirement. individualized feedback on optimizing the retirement income mix ensures that learners can tailor their plans to meet their specific needs and goals, empowering them to make informed decisions and achieve financial security in retirement. foundational knowledge on bloom’s taxonomy can be measured through definitional questions, while higher levels require more depth. developing a comprehensive retirement planning project that applies elements of constructivism with overlap into experiential learning, is outlined in appendix b. this project incorporates real-life scenarios of retirement account selection tailored to individual circumstances, interactive exercises simulating retirement savings calculations, and personalized feedback from the professor on retirement goals and strategies. by engaging in hands-on activities and receiving personalized guidance, learners gain practical skills and insights essential for navigating the complexities of retirement planning. constructivism applied: • real-life scenarios: students apply knowledge to realistic situations, tailoring retirement account selections to individual circumstances, which helps them construct understanding based on personal relevance. (experiential learning theory overlap.) • interactive exercises: engaging in simulations for retirement savings calculations allows learners to actively participate, building knowledge through experience. (experiential learning theory overlap.) • personalized feedback: receiving tailored guidance from the professor encourages reflection and adaptation, reinforcing learning by connecting theoretical concepts to practical application. (experiential learning theory overlap.) • hands-on activities: through active involvement, students construct their understanding of retirement planning complexities, enhancing their problemsolving skills. financial counseling financial counseling education draws heavily on behaviorist, experiential, and social learning principles. these theories inform the use of reinforcement strategies, role-play exercises, and stebbins & quito 93 peer feedback to help students develop practical counseling skills and financial self-awareness. students begin this course by learning about their personal relationship with money, something that likely began with financial socialization (danes, 1994), and identifying their own financial attitudes. students share their first memory of money with the class and engage in instructor-led discussions on positive financial behaviors. elements of behaviorism and experiential strategies here include hands-on budgeting exercises tailored to individual income and expenses. debt management strategies are another essential component of financial education, supported by case studies illustrating effective debt repayment plans. through interactive sessions on debt consolidation and negotiation techniques, learners develop practical skills for managing debt effectively. credit is also addressed, with step-by-step guidance on how credit scores are created and calculating the impact of poor credit. experiential learning theory is again implemented as learners obtain and analyze their own credit report (appendix c1) and learn how to improve their credit history. additionally, learners engage in personalized budget reviews and receive recommendations from the professor. group discussions on overcoming budgeting challenges further allow elements of social learning theory to encourage collaboration and development of strategies for effective cash flow management. on-campus resources, such as the food bank, career services, counseling, landlord/tenant disputes, and law clinics are covered along with community resources. as a practical project students create and virtually share a recipe of a cheap yet healthy meal they made (appendix c2). in addition to budgeting, debt management, credit score improvement, and goal setting, financial counseling addresses critical topics such as dealing with mortgages, car repossession, and medical bill negotiation. learners engage in comprehensive discussions and practical exercises aimed at understanding the complexities of mortgage financing, including the implications of default and foreclosure. they explore strategies for preventing car repossession, such as renegotiating loan terms or exploring alternative financing options. additionally, learners learn negotiation techniques for addressing medical bills, including advocating for lower fees, negotiating payment plans, or seeking financial assistance programs. these concepts are applied throughout the semester with in-class client/counselor role-play sessions, combining elements of constructivism, behaviorism, and experiential learning theories students assume the roles of counselor or client, beginning with instructor-defined scenarios and progressing to student-generated cases that reflect greater complexity. during class, feedback is offered by the professor and classmates. at the end of the semester, students create video where they apply the concepts they have learned, demonstrating mastery of the communication and financial counseling skills learned throughout the semester. this project is available in appendix c3. as evidenced, multiple learning theories were thoughtfully applied throughout the course. here is how the learning theories are applied to the coursework described: behaviorism: • budgeting exercises: students engage in hands-on budgeting tailored to their finances, reinforcing positive financial behaviors through practice. • discussions on financial attitudes: sharing first memories of money and discussing positive behaviors encourages habit formation through reinforcement. experiential learning: • debt management and credit analysis: interactive sessions and personal credit report analysis provide practical, real-world experiences that enhance understanding of financial concepts. • role-playing: in-class client/counselor roleplays simulate real financial counseling scenarios, allowing students to practice and reflect on their skills. • practical projects: creating and sharing a budget-friendly meal recipe involves handson learning, making financial management relevant and applicable to daily life. financial services review, 33(4) 94 constructivism: • collaborative learning: group discussions on budgeting challenges and debt management strategies allow students to construct knowledge through shared experiences and problem-solving. • role-play sessions: students create complex client scenarios, building their understanding by actively engaging in realistic financial counseling situations. • video role play project: creating a video to demonstrate financial counseling skills allows students to synthesize their learning and apply it in a practical context. social learning theory: • peer feedback: in role-plays and discussions, students learn from observing and receiving feedback from classmates and the instructor, enhancing their skills through modeling. • group discussions: collaborative sessions on budgeting and resource management foster learning through shared insights and strategies. these projects effectively integrate various learning theories, providing students with comprehensive, applied learning experiences that develop practical financial skills and critical thinking. estate planning estate planning instruction incorporates constructivist, experiential, and social learning theories to help students engage with complex legal and financial concepts. these theories support the use of case studies, document drafting, and collaborative discussions that promote applied learning and critical thinking. best practices in financial education extend to comprehensive estate planning, a crucial aspect of personal finance that ensures the orderly distribution of assets and the fulfillment of one's wishes upon incapacitation or death. learners engage in interactive lessons on the basics of estate planning, exploring fundamental concepts and legal principles. through case studies illustrating estate planning decisions and their consequences, individuals gain insights into the importance of thoughtful planning and decisionmaking. furthermore, learners receive individualized feedback on estate planning documents and strategies, empowering them to tailor their plans to their unique circumstances and objectives. appendix d provides an assignment to create a comprehensive estate plan for the student (d1) along with a dissecting a will assignment (d2), which serves as practical references for learners as they navigate the complexities of estate planning. these assignments incorporate constructivism, experiential, and learning theory elements. wills and trusts are central to estate planning, and learners delve into these topics through guided exercises in drafting legal documents for their own estate plans. professor-led discussions provide guidance on wills and trusts design and execution. understanding the role and responsibilities of a power of attorney is essential in estate planning, and learners participate in role-playing scenarios to grasp the significance of this legal designation. personalized guidance on selecting and appointing power of attorney agents equips individuals with the knowledge to make informed decisions regarding their representatives. feedback sessions on the power of attorney choices ensure that learners understand the implications of their decisions and are prepared to navigate potential challenges effectively. healthcare directives are another critical component of estate planning, and learners engage in interactive discussions to explore various options and considerations. step-by-step guidance on completing healthcare directives empowers individuals to articulate their medical preferences and end-of-life wishes. professor-led feedback on healthcare directive choices ensures that learners' directives align with their values and preferences, providing clarity and peace of mind regarding their healthcare decisions. through comprehensive education and practical guidance, individuals gain the tools and confidence to navigate the complexities of estate planning effectively and safeguard their legacy for future generations. again, multiple learning theories were applied in the course: stebbins & quito 95 constructivism: • case studies: analyzing real-world scenarios helps students construct knowledge about estate planning by connecting legal concepts to practical applications. • creating estate plans: developing personalized plans encourages learners to actively engage with the material, building their understanding through hands-on activities. • role-playing: scenarios involving power of attorney enhance understanding by allowing students to immerse themselves in realistic situations. experiential learning: • drafting legal documents: exercises in creating wills and trusts provide practical experience, reinforcing learning through application. • interactive discussions: engaging in guided discussions on healthcare directives and estate planning components allows students to reflect and apply their knowledge. • individualized feedback: tailored feedback on estate planning documents helps students refine their strategies and understand the realworld implications of their choices. social learning theory: • professor and guest speaker-led discussions: observing expert demonstrations and participating in discussions fosters learning through observation and collaboration. • feedback sessions: peer and professor feedback on decisions regarding power of attorney and healthcare directives helps learners improve their understanding by modeling best practices. • collaborative learning: group discussions on estate planning options enable students to learn from each other’s insights and experiences. these methods collectively empower learners with practical skills and confidence to navigate the complexities of estate planning effectively. discussion this study presents a structured approach to improving financial education through the application of differentiated instructional strategies. each theory contributes to instructional design by offering strategies that support learner engagement, skill development, and application of financial concepts. however, these theories also present limitations that must be addressed to ensure effective implementation. bloom’s taxonomy provides a structured progression of cognitive skills, but its assumption of linear development may not reflect the varied readiness of learners, particularly when foundational knowledge is uneven. behaviorism emphasizes reinforcement of observable behaviors, which can oversimplify financial reasoning and may encourage short-term compliance rather than long-term behavioral change. constructivism depends on learners having sufficient prior knowledge to build new understanding, which can be problematic when students enter with limited financial experience. experiential learning promotes hands-on engagement, yet simulations may not fully capture the emotional and contextual depth of real financial decisions. social learning theory relies on modeling effective behaviors, but peer influence can reinforce poor habits if not carefully guided. to address these challenges, the paper provides practical examples of assignments and instructional tools that educators can implement across a range of financial topics. these include investment planning, retirement strategies, financial counseling, and estate planning. each example is grounded in classroom-tested activities and aligned with specific learning theories, offering instructors a clear path from theory to practice. the inclusion of detailed assignments in the appendix further supports educators in adapting these strategies to their own teaching contexts. by combining theoretical insight with applied pedagogy, the paper contributes to the development of more effective and inclusive financial education. addressing the diverse learning needs of students remains a crucial aspect of effective financial education. while efforts have been made to cater to various learning preferences, some learners financial services review, 33(4) 96 may still find certain topics less engaging or relevant. future research should explore how these instructional strategies influence higher-order learning outcomes. building on bloom’s taxonomy, studies could examine how synthesis and evaluation are fostered through differentiated instruction and whether certain theories are more effective in promoting these advanced cognitive skills. comparative research could also assess which learning theories lead to stronger student outcomes in financial education, including retention, behavioral change, and decisionmaking confidence. classroom-based studies that measure the impact of these approaches across diverse learner populations would offer valuable insights into instructional effectiveness and equity. desimone (2009) suggests the creation of a common framework to study the impact of professional development on both teachers and students, an idea that can be applied to differentiated financial education. looking forward, innovation in educational technologies presents new opportunities to improve the accessibility and delivery of financial education. artificial intelligence can support adaptive learning environments that respond to individual student needs, while online platforms can expand access to high-quality content and interactive tools. these technologies enable personalized learning pathways, real-time feedback, and data-informed instruction. they also offer potential for simulating complex financial scenarios, tracking learner progress, and tailoring support based on individual performance. however, their integration must be guided by pedagogical principles to ensure that technology enhances rather than replaces meaningful learning experiences. future research should examine how ai-driven tools and digital platforms can support differentiated instruction, promote financial literacy, and improve longterm financial outcomes. a collaborative effort among educators, financial professionals, and institutions will be essential to ensure that these tools are used effectively and equitably. references agarwal, s., driscoll, j. c., gabaix, x., & laibson, d. (2008). learning in the credit card market (no. w13822). national bureau of economic research. https://doi.org/10.3386/w13822 alexander, j. c., jr., & mcelreath, r. b. (1999). does education affect how well students forecast the market? financial services review, 8(4), 253–260. https://doi.org/10.1016/s10570810(00)00043-3 bandura, a. (1977). self-efficacy: toward a unifying theory of behavioral change. psychological review, 84(2), 191–215. https://doi.org/10.1037/0033295x.84.2.191 becker, b., van rijn, p., molenaar, d., & debeer, d. (2022). item order and speededness: implications for test fairness in higher educational high-stakes testing. assessment & evaluation in higher education, 47(7), 1030–1042. https://doi.org/10.1080/02602938.2021. 1991273 bloom, b. s., engelhart, m. d., furst, e. j., hill, w. h., & krathwohl, d. r. (1956). taxonomy of educational objectives: the classification of educational goals. handbook i: cognitive domain. david mckay. bruner, j. s. (2009). the process of education. harvard university press. burke, b. m. (2013). experiential professional development: a model for meaningful and long-lasting change in classrooms. journal of experiential education, 36(3), 247–263. https://doi.org/10.1177/1053825913489 103 cassidy, s. (2004). learning styles: an overview of theories, models, and measures. educational psychology, 24(4), 419– 444. https://doi.org/10.1080/0144341042000 228834 choung, y., chatterjee, s., & pak, t. y. (2023). digital financial literacy and financial well-being. finance research letters, 58, 104438. https://doi.org/10.3386/w13822 https://doi.org/10.1037/0033-295x.84.2.191 https://doi.org/10.1037/0033-295x.84.2.191 https://doi.org/10.1080/02602938.2021.1991273 https://doi.org/10.1080/02602938.2021.1991273 https://doi.org/10.1177/1053825913489103 https://doi.org/10.1177/1053825913489103 https://doi.org/10.1080/0144341042000228834 https://doi.org/10.1080/0144341042000228834 stebbins & quito 97 https://doi.org/10.1016/j.frl.2023.10443 8 consumer financial protection bureau. (2023). financial well-being: the goal of financial education. report, iowa city, ia: consumer financial protection bureau. danes, s. m. (1994). parental perceptions of children's financial socialization. journal of financial counseling and planning, 5, 127–149. desimone, l. m. (2009). improving impact studies of teachers’ professional development: toward better conceptualizations and measures. educational researcher, 38(3), 181– 199. https://doi.org/10.3102/0013189x08331 1 dole, s., bloom, l., & kowalske, k. (2016). transforming pedagogy: changing perspectives from teacher-centered to learner-centered. interdisciplinary journal of problem-based learning, 10(1), article 1. https://doi.org/10.7771/1541-5015.1538 eikeland, i., & ohna, s. e. (2022). differentiation in education: a configurative review. nordic journal of studies in educational policy, 8(3), 157– 170. https://doi.org/10.1080/20020317.2022. 2039351 eyssell, t. h. (1999). learning by doing: offering a university practicum in personal financial planning. financial services review, 8(4), 293–303. https://doi.org/10.1016/s10570810(00)00047-0 fernandes, d., lynch, j. g., jr., & netemeyer, r. g. (2014). financial literacy, financial education, and downstream financial behaviors. management science, 60(8), 1861–1883. https://doi.org/10.1287/mnsc.2013.1849 fogarty, e., & mayo, h. (1999). undergraduate research: the senior thesis in finance. financial services review, 8(4), 223– 234. fox, j., bartholomae, s., & lee, j. (2005). building the case for financial education. journal of consumer affairs, 39(1), 195– 214. https://doi.org/10.1111/j.17456606.2005.00009.x giesecke, j., & waschik, r. (2025). economic analysis of us tariffs introduced over march–april 2025 (centre of policy studies working paper no. g-352). victoria university. gonzález, n., moll, l. c., floyd-tenery, m., rivera, a., rendon, p., gonzales, r., & amanti, c. (1993). teacher research on funds of knowledge: learning from households. grinder, b., cooper, d. w., & britt, m. (1999). an integrative approach to using student investment clubs and student investment funds in the finance curriculum. financial services review, 8(4), 211– 221. https://doi.org/10.1016/s10570810(00)00040-8 gudmunson, c. g., & danes, s. m. (2011). family financial socialization: theory and critical review. journal of family and economic issues, 32, 644–667. https://doi.org/10.1007/s10834-0119275-y gutter, m. s., garrison, s., & copur, z. (2010). social learning opportunities and the financial behaviors of college students. family and consumer sciences research journal, 38(4), 387–404. https://doi.org/10.1111/j.15523934.2010.00034.x hathaway, i., & khatiwada, s. (2008). do financial education programs work? (working paper 08-03). federal reserve bank of cleveland. https://doi.org/10.26509/frbc-wp200803 hensley, b. j. (2015). enhancing links between research and practice to improve consumer financial education and wellhttps://doi.org/10.7771/1541-5015.1538 https://doi.org/10.1080/20020317.2022.2039351 https://doi.org/10.1080/20020317.2022.2039351 https://doi.org/10.26509/frbc-wp-200803 https://doi.org/10.26509/frbc-wp-200803 financial services review, 33(4) 98 being. journal of financial counseling and planning, 26(1), 94–101. hilgert, m. a., hogarth, j. m., & beverly, s. g. (2003). household financial management: the connection between knowledge and behavior. federal reserve bulletin, 89, 309–322. honebein, p. c. (1996). seven goals for the design of constructivist learning environments. in b. g. wilson (ed.), constructivist learning environments: case studies in instructional design (pp. 11–24). educational technology publications. jump$tart coalition for personal financial literacy. (2006). financial literacy shows slight improvement among nation’s high school students. author. kalaycı, k., & serra-garcia, m. (2016). complexity and biases. experimental economics, 19(1), 31–50. https://doi.org/10.1007/s10683-0159434-3 kim, j., & chatterjee, s. (2013). childhood financial socialization and young adults’ financial management. journal of financial counseling and planning, 24(1), 61–79. kolb, d. (1975). towards an applied theory of experiential learning. in c. l. cooper (ed.), theories of group process (pp. 33– 56). john wiley & sons. kolb, d. a., boyatzis, r. e., & mainemelis, c. (2014). experiential learning theory: previous research and new directions. in d. a. kolb (ed.), perspectives on thinking, learning, and cognitive styles (pp. 227–247). routledge. lusardi, a. (2019). financial literacy and the need for financial education: evidence and implications. swiss journal of economics and statistics, 155(1), article 1. https://doi.org/10.1186/s41937-0190027-5 lusardi, a., & mitchell, o. s. (2014). the economic importance of financial literacy: theory and evidence. journal of economic literature, 52(1), 5–44. https://doi.org/10.1257/jel.52.1.5 lyons, a. c. (2005). financial education and program evaluation: challenges and potentials for financial professionals. journal of personal finance, 4(4), 56– 68. lyons, a. c., palmer, l., jayaratne, k. s., & scherpf, e. (2006). are we making the grade? a national overview of financial education and program evaluation. journal of consumer affairs, 40(2), 208– 235. https://doi.org/10.1111/j.17456606.2006.00056.x mandell, l., & klein, l. s. (2009). the impact of financial literacy education on subsequent financial behavior. journal of financial counseling and planning, 20(1), 15–24. mccombs, b. l. (2001). what do we know about learners and learning? the learnercentered framework: bringing the educational system into balance. educational horizons, 79(4), 182–193. http://www.jstor.org/stable/42927064 mccormick, m. h. (2009). the effectiveness of youth financial education: a review of the literature. journal of financial counseling and planning, 20(1), 70–83. pavlov, i. p. (1927). conditioned reflexes: an investigation of the physiological activity of the cerebral cortex (g. v. anrep, trans. & ed.). london: oxford university press. perlini, a. h., lind, d. l., & zumbo, b. d. (1998). context effects on examinations: the effects of time, item order, and item difficulty. canadian psychology/psychologie canadienne, 39(4), 299–306. https://doi.org/10.1037/h0086821 salas-velasco, m., moreno-herrero, d., & sánchez-campillo, j. (2021). teaching financial education in schools and students’ financial literacy: a crosshttps://doi.org/10.1186/s41937-019-0027-5 https://doi.org/10.1186/s41937-019-0027-5 http://www.jstor.org/stable/42927064 stebbins & quito 99 country analysis with pisa data. international journal of finance & economics, 26(3), 4077–4103. https://doi.org/10.1002/ijfe.2005 skinner, b. f. (1938). the behavior of organisms: an experimental analysis. new york: appleton-century. stohr-hunt, p. m. (1996). an analysis of frequency of hands-on experience and science achievement. journal of research in science teaching, 33(1), 101–109. https://doi.org/10.1002/(sici)10982736(199601)33:1<101::aidtea6>3.0.co;2-z tomlinson, c. a. (2001). how to differentiate instruction in mixed-ability classrooms. ascd. tomlinson, c. a. (2017). how to differentiate instruction in academically diverse classrooms. ascd. ullah, r., ismail, h. b., khan, m. t. i., & zeb, a. (2024). nexus between chatgpt usage dimensions and investment decision-making in pakistan: moderating role of financial literacy. technology in society, 76, 102454. https://doi.org/10.1016/j.techsoc.2024.1 02454 urban, c., schmeiser, m., collins, j. m., & brown, a. (2020). the effects of high school personal financial education policies on financial behavior. economics of education review, 78, 101786. https://doi.org/10.1016/j.econedurev.201 8.03.006 warmath, d., & zimmerman, d. (2019). financial literacy as more than knowledge: the development of a formative scale through the lens of bloom’s domains of knowledge. journal of consumer affairs, 53, 1602–1629. https://doi.org/10.1111/joca.12286 watson, j. b. (2017). behaviorism (reprint ed.). routledge. ziernwald, l., hillmayr, d., & holzberger, d. (2022). promoting high-achieving students through differentiated instruction in mixed-ability classrooms: a systematic review. journal of advanced academics, 33(4), 540–573. https://doi.org/10.1177/1932202x22111 2986 https://doi.org/10.1016/j.econedurev.2018.03.006 https://doi.org/10.1016/j.econedurev.2018.03.006 financial services review, 33(4) 100 appendix a investment planning a1 investment plan objectives 1. understand client goals. 2. write investment policy statement. 3. create diversified portfolios to meet those goals. 4. explain your reasoning at a high level for the professor. 5. explain/educate your client on what you have done and why. write an investment plan for two clients. • there are three things to turn in for each client: 1. written investment plan for client. make it pretty. 2. diversified portfolio a. use only investments offered by fidelity. b. no target date funds 3. write up of reasoning behind investment choices at a high level for professor stebbins. meet your clients! 1. bobby braxton is a patent attorney living in colorado. a. he is 42 years old and has not begun saving for retirement. he has no idea what risk tolerance is but doesn’t want to lose all of his money. b. he earns approximately $200,000 annually. c. he has 2 young children and a wife. none of them work. d. he wants to pay for both children to go to texas a&m or somewhere with an equivalent cost. 2. you! a. it’s 5 years after you graduate from bama. a. roll tide! b. you can be employed wherever you want, married or not, kids or not. c. basically it’s where you envision yourself in 5 years. d. create your goals (at least 3) and invest for those goals. stebbins & quito 101 appendix b retirement planning comprehensive retirement planning project 1. please read through the file called “completed client lifestyle intake form”. this will provide you with the data you need to create the plan. jack and jill jones are your clients. they have come to you because they are very concerned about their retirement. several of their friends have been retiring as of late and they hadn’t really thought about their ability to retire until this year. they would like to know if they are on track to meet their retirement goals. they have been saving much more lately and want to make sure it is enough. your job is to determine the likelihood of them meeting their goals using moneyguidepro (mgp) and to determine what changes would be necessary in order for them to achieve as many of their goals as possible will staying within a reasonable boundary of success. 2. once you have completed the mgp analysis, you need to write-up your findings as you would present them to your client (see the grading rubric for a more detailed guideline). you may feel free to include (cut and paste) output from mgp analysis and model your discussion after the presentation feature in mgp. the write-up should at minimum include: • client goals • assumptions used • risk tolerance score/portfolio • explanation of “likelihood” of meeting goals • current likelihood of meeting goals • alternative to meet goals (i.e. additional savings and where) • your recommendations to meet goals • likelihood after recommendations there is no page limit or expectation in answering the above, only that each topic from the rubric is adequately addressed. note: in solving the analysis, you should not use “super solver” to solve your answer. you can adjust each goal individually in the software based off of the ideal and acceptable ranges in conjunction with the importance of each goal. financial services review, 33(4) 102 appendix c financial counseling c1 credit report obtain your free credit report and check it for errors https://www.annualcreditreport.com/index.action you can obtain the report from any credit reporting agency you choose. feel free to share the whole report or just a screenshot with enough shown to prove you obtained the report. if you have any questions about your report or want to dispute an error on the report, i am happy to help. c2 eating healthy on a budget if you don't eat, you will die. this assignment is designed to help you stay alive and better understand how to eat healthy on a budget. this assignment is worth 2 assignment grades. the assignment: • make a healthy meal for less than $5 per serving. o include cost per serving (receipt or list cost of each ingredient) • what is healthy? o https://www.dietaryguidelines.gov/sites/default/files/2021-03/dga_20202025_startsimple_withmyplate_english_color.pdf • provide recipe. • take a picture of the meal. • cite the source of the recipe if applicable. • post all of the above to the discussion board. (note: you can claim your meal in advance by posting what meal you're claiming on the discussion board. this way i don't have half the class making the same meal.) this could be a bulk meal prep of a week + worth of food and then divided into individual meals. this could be a meal for multiple people. restrictions: • no frozen/pre-made meals o note: meal prep services like hellofresh might have introductory offers that could be gamed, but for the purposes of this assignment, this is a meal you make yourself. • no sandwiches • no use of food pantry/wic/snap type benefits. these are definitely great resources when in need, but again not in the spirit of this assignment. resources: nutrition on a budget | nutrition.gov find tips for eating healthy on a budget and saving money when food shopping. eating cheap but eating healthy (rutgers njaes) sshw worldwide challenge daily motivational messages from rutgers njaes. https://www.annualcreditreport.com/index.action https://www.dietaryguidelines.gov/sites/default/files/2021-03/dga_2020-2025_startsimple_withmyplate_english_color.pdf https://www.dietaryguidelines.gov/sites/default/files/2021-03/dga_2020-2025_startsimple_withmyplate_english_color.pdf https://www.nutrition.gov/topics/food-security-and-access/nutrition-budget https://www.nutrition.gov/topics/food-security-and-access/nutrition-budget https://njaes.rutgers.edu/sshw/message/message.php?p=health&m=130 https://njaes.rutgers.edu/sshw/message/message.php?p=health&m=130 stebbins & quito 103 c3 counselor/client role play role play a counseling session with a friend! show your ethos with the way you dress, the location (you are encouraged to use the counseling lab on the 3rd floor of adams hall), and anything else that might help. show the communication skills we learned and the surviving debt content covered to help your client with their issue(s). the video should be at least 10 minutes. you can share them with me in ua box or emailing me (rstebbins@ua.edu) when you've completed the video. videos are due xx/xx at 11:59pm. financial services review, 33(4) 104 appendix d estate planning d1 comprehensive estate plan client profile: look in a mirror. it’s you! you must create an estate plan for yourself today and another one at age 50. you can have a family or not, whatever you envision for yourself at age 50. make any reasonable assumptions needed to create the plan, explaining the assumption and why it was made. please write the plan to yourself. so, for billy smith, it would be written to billy from smith financial planning (or whatever firm name). what is needed in your estate plan? 1. cover letter to yourself giving an overview of the plan. 2. client profile should include a detailed net worth statement with a notation on how each asset (including digital assets) will be retitled (beneficiary designation, titling, will, trust) and where the asset will go. note any valuation discounts that might apply. example: asset value recipient contingent recipient retitling mechanism 401k $100,000 spouse children, per stirpes beneficiary designation total sum of assets 3. list and explain the documents you should have in place. 4. write a will for yourself. a. identify the clauses and explain what they are for/why they are needed. see exemplar. 5. write a living will. 6. write a healthcare power of attorney. 7. beneficiary designations for all accounts that allow beneficiary designations. 8. title at least one asset in a way that avoids probate. 9. required for age 50: 1. use at least two trusts to create ‘strings’ and explain: a. type of trust b. beneficiary/beneficiaries c. strings 10. explain your reasoning for all recommendations. c2 dissect a will the probate process is public, meaning you can access a will at your local courthouse. your assignment is to do just that: choose a will with an identical or similar last name to yours. using the provided will checklist, identify the clauses present and type/put them in the checklist document. note: this does not mean you have to put every single bequest, just one and include any contingent beneficiary/executor/guardian if present. this assignment is worth 4 assignments in the grade book. stebbins & quito 105 • why close to my last name? o to help you confront your mortality, potentially motivating you to plan for your own estate. o to better understand and empathize with your clients' when estate planning is brought up. o so you do not all turn in the same will. • how do i find a will? o https://probate.tuscco.com/probateweb/search/searchtype?key=probate_wills o or, visit in person: o main office: tuscaloosa county courthouse 714 greensboro avenue, suite 121 tuscaloosa, al 35401-1891 office hours: 8:30 am to 5:00 pm • what do i turn in? • the checklist • the will • how do i turn it in? ▪ submit it here. ▪ save the checklist as yourlastnamewillchecklists24 ▪ save the will as yourlastnamewilldissecteds24 ▪ scan it, turn it into a pdf, take pictures of each page, something that lets me see the will so i can check your work. ▪ if you're not sure you successfully uploaded it, email me (rstebbins@ua.edu) with the will and checklist attached. https://nam11.safelinks.protection.outlook.com/?url=https%3a%2f%2fprobate.tuscco.com%2fprobateweb%2fsearch%2fsearchtype%3fkey%3dprobate_wills&data=05%7c01%7crstebbins%40ches.ua.edu%7cfbe34428628b4f4e2be708db052a566b%7c2a00728ef0d040b4a4e8ce433f3fbca7%7c0%7c0%7c638109451613444961%7cunknown%7ctwfpbgzsb3d8eyjwijoimc4wljawmdailcjqijoiv2lumziilcjbtii6ik1hawwilcjxvci6mn0%3d%7c3000%7c%7c%7c&sdata=44%2fjai026uggbqojumg6ntahgwcarkg9p0x7lzut7ya%3d&reserved=0 pii: s1057-0810(02)00090-2 ���� ���� ����� � ����� ���������� ������ �� ����� �� ���� ����� �� �� � � ��� ������ ������ �� �� �� �� ���� � ��� � �� � ��� �� ��� �� �� � ��� � � � � ��� � � � ��� �� ���� � ��� ������!� �� �� �� � �� ��� �� �� "�# � � � ����� �� � � ����� �� "$ %&'�() �*�� ��� ���� �� �� ��������� !!!" ���� �� ��� �� #��� ! $% � !!�" ������� �& �������� !!� �������� ���� ��% �� ' �(� ���������� ������ � ������ � �������� �� �� �%����� ��� #��� � ��� ��� � �� ������� %� � �� �# )����( �������� ��� ������ *+(� $�%� �� �# ,���#�� � -� ����� � .���/0 #���1 ��� %��� � ��(�� ���� ��� � ��� #�� �(� ���� ��� �� �# ���������� ������ �� 2� � �������� �(���� �� ������ � ����� �� �� %� � �� #% �� � �(�� ��� �� ��3% �� ������� �� �� ��� ���� � �4 ������� )� ��� � ����� ���� ������ 5� �� 6���%�� #�� #�%� ������ � ������� �� � #�� ����� �� �� ����4 ������ ��� 7� ' ���� ����� � �� �� ��� �� ����� �(� �� ��� �� �# �(� ���������� ������ � ������ � ��������� � !!� 8���� �� �� � �� 9 �� :�� � �(�� ������� � +,�� � � � ��. ;�" ;��" < ���/��� . ����������� ������ �" ���� ����� � " :���� ������� � � ���� ������ ����������� ������ � � � ���%��� ������ � ���(� )(��� � ������ ) �( � �%� �# �� �� �� ���� ��� �� ���� �(� � � �� �%� ��� �����" ���(��� (� ����� � '�� ������� � �# �(� ������ � ������ �� ���%��� ����� �� ������ � ��� +( � ���(� � �(�%�(� �� �%��� ��� �(� ������ ��� �� ���� ( � � � �� �%� �� � ���4�� ( �( � �(%� ������� ( � ������ � �� � � �� ����� � �� =� ��� � �� .��>�0 ��� ������ �(�� ���������� ������ � ��� � ��� �%����� ��� �(����� ������ �� � ��% �� (��� ������� �(� ���������� ������ � ���(� �# ���� � .(����#��� 4 �) �� �=:0 �� � �%�� �%� ���(� �# ���� � .��?�##� ���@" 7 �� ��� a ���� � ���&0� � �%� � (�� �������� .���� � 7 �� ���� a :� �� ���>" � �(� a -� ���� ���&0� �� � ���%� ������ � ���(� �# ����#�� � ������ � � .+(������ ���@0� b � � �� ���� ��� ��� �) �! . !!�0 ��>c� > �!/>�!d�!e!�ef c ��� #�� � ������ � !!� 8���� �� �� � �� 9 �� :�� � �(�� ������� � ,99g � � ! / > � ! d � ! . ! 0 ! ! ! � ! � � =������� � �%�(��� +���g ���d�h� &/��/>&" #��g ���d�h� &/�hh!h� ,)� � ���� � . ���� �4i%�4��� % .���� ���� �0� � � i%����� % .�� � � 0� � +���g ��� �!�@/d�>&� " #��g ��� �!�@/d�/d&>� -��� ������ �� � ' �=: �� �� ���4 � #�� ������� )(� (��� ��� ���� �� ���� %� � �� #% �� � �" (�)����� ���� � �������( ���� �(� ���%� �# �=:)(� �(� ������ � ���%�� �� ��� ( �(�� ����� �� .����4� a ,�� ��� ����0� ������ .���/0 ��� %��� � ��(�� ���� ��� � ��� #�� �(� �� ��� �� �# �=:� 2� � +����4� � ��( ��� 5� �������� �(���� .��� 0� (� ���� �(�� ������� ����%��� �(� ��� ����� �� �� � ��#��� �� �� �� 7(����� ��� �� ������� ��� ����%�� �� ��)��� �� � �4 ������� �������� �(���� ������� (��� ���(��� %� � �� �%����� ����� � � �4 ������ ��(�� �� ���� �� � �%� � �4����4 � ��(�� �� ���� ������� 9 ������� ���� ���4 �� � � ?� �(� ������ ����� ��� ) �( � �� ���� �� ������ � �%������ ������ ��� ���� ������� )(� #����) � �%�� �%�( �� �(�� ����� �� �=: . ���� jj ���� �e� �# ��%� ����� ������ ���( �� �( � ����� ����� ���� �# �(� �����5� �� ���550 #��� ������ #��� ���� �# �(� ����� � � � �� �# � �� ������ � �%����� � ����� � �� � �� %�� � ������� 9 ������ ����� � �� ���� � ��� � �� %�� �%��� �� (��� � � ?� ������ (��� �� �(����� ����� ����� �(� ���� ��� �� �# �(� �=: ���(� �# ���� �� �������(��� ��� ��� � �� ��������� ���� ����� � �� ������ %� � �� #% �� � � �� ������ �3% � �� %� ����� �� � � �� ��� .�(%�)��� ���>0� 9 �( � ����� )� %�� �� �(�� ���%� ��� �� ������� �=: �� ����� �� �� ������ � ������� �� )(� �������� �(���� ����� � ������ %� � �� #% �� � �� 7� ' ������ %� � �� � ��)��� #�� � �=: ������ � �������� #�� ����� ���� ����� �� ������ �%�( �� ����� ��� ����4�� � � #�� �� �� ��� ���� � ������� �� ������ ���%� ������ � #�� ����� ��� ����4�� 9 ������� ) �� ���� ���#�� � �%� � (�� ���� � �������� �� ���������� ������ � #�� ����4�� +(��� ���%��� � �� �%����� ���������� ������ � �� � ������ � ��������� � �������� ���������� 9 ������ � � � ���� �#�� �%�� ����� � ������� �� �� ������� �� �(� � �� � �� �� �� �(� � � � �# ��%�� � ��� ����� ���( �� �%��(��� ������� +(� ��� ��� �(� ��� ��� �%�� �%� ���� �� � �%� � (�� ��������� ���%� ������ �� �� ���������� ������ �� '0�0 -��1 �� ��� � �! �%�� �%� ���� � .��0 ��3% ��� �(� ������ �� %�� �(� � � �� ������ � ������ �� �%��(��� ������ %��#�� �� b�� �������� # � ������ (�� f�!�!!! ��� ����� �� ��� � (� �%��(���� f�!�!!! )���( �# ������ �)� ������ �(� ������ � ������ ������ � ����%��� �(� ���%� ��� � � �( � ������ � ���� � � �� ��� � �# � ��� +(� � �� ���� �� �� ���� � � �(�� � � %�� ����� �� � ��� ��� ����� ������� � � �%��(���� �(� �� �� ������ ��� ������ � ��� � ��� � �3% �� ������ ���%� � � �(� �(��� ����#�� �� : �� �� ���� �# �( � �������� � �(�� �(� ������ ��� � ����� ��� ���� � ��� �# ( � #% � �� � ���4�� ( �(� � ������% � � �� �� ����� '0'0 ��� �� ���� �� ��!� : ����� �� �� ������ � �������� � 4 �) �� �%� � (�� .�k0� 7 �( �k� � � %�� ������ ���� ���� � �# ( � )����( � �4� ������� � �(� ���� � ���� � �# ( � ��d �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% )����( � (�� ���� � �4� ������ �%�( �� �����%�� � ��� .+�� ���0� +(� ������ � ������ ���� �(� �� � �� ��� �%� � #�� ���� ��� � �# � ��" �(� �(� ������� ���%� #�� �(� ���� � �������� � ����%���� .+(������ ���@0� b�� �������� # � ������ (�� f�!�!!! �� ����� (� ��� �(���� �� ����� /!l �# ( � )����( � ���4�� �� #% � (�� �(� ���� � /!l �# ( � )����( +�� ���� k� ��� �� � 6%�� ( � ��� � � � � �� �(� � �# �(� ���� (� ����%����� �(� ���%� � ( � ����#�� �� +( � �������� ��3% ��� � ������ �� ����� � ������ �# ������ %� #�� � � %�� � ���� � �# �(� ��� ����� #% � �� �%��(��� ���� ����� ���� �# � �4� ������" � � � � �� ��� ���� � � ��� ��3% �� %� � �(� ������ � � �� ��� � � : �� �� ���� �# �( � �������� � �(�� �(� � �4 �# �(� ����#�� � ������� ���� � ��� +(� ������ ���%� � ����� ��� ) �( � ������ � � ���4�� #% ���%���� � �������� �(�� �(� ����� ���� �# ����� )����( ����� �(� � �4� ����� ) �� ������ ����� �� �� �(� )����( �������� �� �(� � �4�#��� ����� ���� � ��� 9 ������� ��� ����� ����� � � ��� ��� ����� �# � �4 �� �� ��( ��� ���� � ��� +(�� � �(�� % ��� ���� �(� � �4� ����� �� � ���� 4 �) � �(�� �(� ����� ���� �# ����� )����( � �4� ������ ) �� ���) �(��%�(�%� �(� � #� �# �(� ������ �� �� �(�� ���� � � �(� ������ � ������ �(� ��� ��� � � �# � �4� � � �4�#��� ������ %� #�� � � )���( �(� � �4 ��� �# �(� ����#�� � ������� 8 �(�� �������� ���%��� � �%����� ��� ��� �%� � �# �(� ������5� )����( %� � �(� � #� �# �(� ������ � .� �(� a -� ���� ���&0� '0*0 9 ��� ��� ! �! m��%� ������ � .m:0 ����)� ������� �� ��4� � �� ���� �# �� �� n%��%�� � � � ������ �%��(���� )(� �� ��� ��� ��) � ������� �� ���� �%��(���� )(� ����4 �� ��� ��� ( �(� 9 � %��� %� � m: ��� �� ��� � ) �( ����� � �(� �%�%��� �� ���%� �# �(� � ������ � �� � ��� ���% � ���( ��� � � b�� �������� ���5� ��� � ������ )� �� �� ������ �(� ���%� �# ( � ����#�� � �� f��!!! ���( �� �(� +(� � %�� ����� f��!!! � ��� � l �(� '��� �� �(� +(� ������ � � �) )���( f��! !� b�� �(� ���� �� �(� �(� � %�� )� �� �(� ����� ���%� ����� �� �� f �!!! �� (� � � � � � � �� f�d! .f �!!! � f��! !0 �� �(� ����% �� :���� �� ����� # �(� '��� �� �(5� ������ � �# f��!!! (� ���� /l� �(� ���� �� �( �(� � %�� �� � �� ���� f��!/! .f �!!! � f��!!!� /l ���� � f�/!0� m��%� ������ � � ���� �% �� #�� ����� �� ������ ��� � �� �(� � %�� ���� � ���� jj�%�� ��)� ����� ( �(�55 9� ��3% ��� � ( �( ����� �# �� �� � � �(� ���� �# �(� ������ �� �� �4 ) �( �(� ��������� 9� ���� ��3% ��� �(� ����%���� �� ������ ������ � �� �� �%� � � �#��� � ��� � �# ���� �� ���%� �� '0�0 2��� �)�� � ��� ! �! �=: ��3% ��� � � %�� �� ���� �(� ���� ���% � �# �� �� �� ���%��� �������� �%�( �� ����� )��4� �� �(� �� 3%������ �� #����) � �=:� � ������ � � %� �%��(�� � ���� �(���� )(� �� ��� #��� � #�)�� �(���� )(� �� ��� � ��� �=: � � � ����� #���� ��� �� ��� �(�� ���%��� ��)�� ��� ���� � ����� �(� ) �( � ��� �(�� ��3% ��� #��3%� � ����#�� � ������ � �� 9� ����)� ������� �� (� �� ��� �� ������ �(�� ���%��� #��� ���� � � �%�� �%� %� � � ���4�� ( �( .,��� ��>�0� �=: ��� ��� (�)����� ��� � ?� � ������5� ������� %� � �� �# )����( #�� � ��� ���� �� ���� %� � �� �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% ��� #% �� � �� ���� �� �� -��4�) �? .��/ 0� +( � �(������ �� �# ��%���� ����%��� �(�� �(� ������5� %� � �� �%��� � ��� ���� �� ����� ,������� �(���� 3%��� � � �( � ����%��� � � � ������ �� ���������� �������( � �(� ���������� ������� ���� � ���( 3%� � �� � � �� ���%���� +(� �������� ��� �� %�� ������� �� � �� �� � � �%�( �� �� � %� �� ���� ��� )(� �(� ������ �5� ���� � ( ����� (�� ��� ����4� 9 ���� � � '�� ����� ���% � %� � ����4 �� �� ��� �� ���%��� � ������ �(� %���� �# �(���� � ������ � ���� �� �%��(���� +(� � )(� �(� ����4 �� �� ������ ����� �(� ������ �� �'�� #��� �(� ��� ��� �� � ���%�� � #��� �) ���( � �# �(��� ������� %���� �# �(����� +(� ����� ��% �� �# �=: ��4 �)�� �� �(� �� %�� � �# ���� � � �4 ) �( �( � ��������� 9 ������� ) �( � (�� �� �� �� � �� �� � �%� �# �� �� �� ���� ) �( �� ��� �(� ������ ����� ��� ) �( ���� � �(� � � �� �%� �� )(�� �� ���� )�� � ���4�� ( �(� :� � ���%��� �� � � ?� ������ ������� ' � � �� � � ���� � ������ � ��� %�� ��� �#�� �# �(� ������ � ������ �� �(� � �4� ������ =� ��� � �� .��>�0 ��4 �)�� ��� �=:5� �� � �� �� �� %�� �(� � �4 �# ���� � �%� �� �� ' � �=: �� �(����� ����� �� � �%����� ��� ������ � �������� )(� ������� �� jj6%�� � 55 �� �(� ���4�� � ���� � �(� � � �� ���% � � � �%�� �%�� ������� �������(��� (��� ������� �=: �� ����� �� �� ������ � ���( 3%��� 8 ���� .��dd0 �������� �=: �� m: � ' � ������� ��� ������ �## ) �( � m: ������ � �������� ���(�� �(� ) �( �=: ���� �� k��� ��� . !!�0 ��%��� �(��� ' ��� ���4 � �� �! ����� .���!c !!!0 �# 3%������� ��� #�� ������� )(� �%��(��� �(� �a, /!! ��� k��� ��� ���� �� � ��� �(� ����� �# ��� ���� � ����� � ����� � ���%� �� k� �������� �(� ���� � �������� �# �=: �� m: � �� ���� � � ' � �� ���%��� �%��� �� � %�� ?� ���%� �� ���(�%�( �=: �%����#���� m:� ��?�## .���@0 ���� ' � �� �� � �� �%��� �� � %�� ?� ���%� � �� � �=: ���� � ��������� k� ����� � �(�� � �� �(� ����4 ���4�� (�� ��� � �� ������� � �4 ���� �� �� � �(� �%��� �� ���� � �������� � �� � ��%��� ������ � ������ �� ����� � �� ���� ��� � � ���%� ���� ?�� � � )( �( ������� �(� ������� ���%� � �������� �(� ����� � �� �# �(� ������ �� +(����� .���@0 �������� �(� �=: �������� �� � �k ������ �� +(����� ���� �(�� ������ � � � ���� ���� �(� �=: �������� �� �� � �� � �� �(� ��(� %�� ��� �� ��� (���� � %��� ��� �(� ������� � �� �� �%�� ��� ��� +(� ��% � ' � �(�� �=: ��� � �� � �� %�� � ������� ���%� � � � ������ � �4 )(� ������� �� �(� �k ��������� ������� ��% �� %�� � �%��� � �� ���� �(� �##��� �� ��� �# �=:� :�����4��� � ���� ����� . !!!0 %�� -� �� =���� � �%��� � �� ������ � �� �� �� 3%��� � �(� ��� �# �(�� �=: ���� � ��� �� �%��� �� ���%� �� +(� ��% � �(�)� �(�� �=: ���� � ��� � �� ��)�� ������� ���%� � � ��)�� ����� � �� �# ���%� �� b�� ������ �� ) �( ��) ����� � ��� �� � �(� �%��� �� ������ � ��������� b�� ������ �� ������ ) �( ( �( ����� � ��� �(� ���%��� ��� ���� ����� �%�� �� �%����#���� �=: �%� ������� �(� ������ �� ������� � �4� ��(���� .���d0 ���� ' � �=: % �����#���� � �%��� � �� �� �� m��� � -�; � . !!!0 ��4� � ##��� � �������(� ���(�� �(� ���4 �� � � %��5� ������ � �� � � � �(� �%�(��� �� � �� � ������5� � ������ � �� � � � ��� ���� )(� ������ ������ #��� �(� � ����#�� � � � ���%����� ��(� %�� ��� � ��� �%��� �� ����#�� � � ! �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% ���#���� �� # �(� ����#�� � � ����� ����4� �� ������� �� � ����#�� � �# �� �� +(� �� � �� �# �(� ����4 ���%� � ���� �(� �� ���%� � ������� �� �(� ������ �5� (�� ?� �������� +(��� ���%��� ���� �� 3%��� � �(� �� ���� �(�� �� � ��� � ��#�� ������ � �� �� �%���� ���� � �� ����4� ��� � � �4 �� ������ � ���%�� �� �(��� ���%��� �%����� �(� � �%��� � �� �� �%������" ������ �=: � � #�� �� ������ � ��������� <�� ��� �������( �%������ �(� �%��� �� �� �# ����� �� �� ������ � ���( 3%��� 9������� .����0 ��% �� �(� � %�� (�� � ��� � ���%� � #�� &/ �# �(� ������� �3% �� #% � ���� � �!� ���� ��� � � +(� �=: ���� � �������� ��� �� ( �(�� ���%� � #�� �� �# �(� &/ #% �� +(� ��% � �� ��% �� �(�� �=: � � �%��� �� �������� #�� #% � ) �( ��) ����� � �� )( �� �� � ���� #�� ����� �� #% �� ����4� � ,�� �� .����0 �� ��� �� �(��� ���%��� � ' �=: �� �� �%��� �� �� �� ���� � ����� ���� #�� ����� �� ���%� � ��� b ����� ������ .���/0 ��� %��� � ��(�� ���� ��� � ��� �� �(� ������ k� ���� �(�� �=: � � �%����� ��� �������� �%� ���� ��� %� �� �(� ��(�� ���� �(������� �� �� �# �������� �������( ����� ������� ���#�� ����4� �(�� ��� � � � .�(� jj� � �(� (� 55 ��� � ���0� ��� ���%��� � �� ���� ?� ������� � ��� ��� �(�� ���� � ��� ����� � ����4� ��� � �� ( �(�� ������� ���%� �� �=: ���� �� � �%�� #�� ������� �� #����) �� � � ?� ������� 9� � ) �( �( � ��(�� ���� #����)��4 �(�� )� ������ �� ��� � ����� ����%��� �=:� � ����� ����� �0�0 ��� 1��� ������ ,������� �(���� � � ����� �� �� �� ������� %� � �� �(���� � � %�� �� ����� �� � � ��4 � )(� �%������ ��� % ����� � 8������ %� � �� �(���� � ���� %�� �(��� �� ���g �(� ������� %� � �� �# � �(� �� � �3%�� �� �(� �%� �# �(� ������� %� � � �� �# ��� ���� ��� �%������" � �(� �� � ���������� # � � � ���%� �� �(� �� �� � ����� ����#�� �" � ������� ��� � �4 ������ .-��4�) �?� ��/ 0� +����4� � ��( ��� .��� 0 ��� � �� �� � ���%�� ������ ��(�� �� �(�� � �� ����� � �� �( � ����� �� �� ��� 7(����� ������� %� � �� �(���� (�� ������� ����%��� �� � � � ���� %�� �(� ����� � ������ ����� )����(� �������� �(���� (�� ������� ����%��� �(� ��� ����� �# ���� � �� �� � � ������ ����� �� �� ��#��� �� �� �� .������ � ���/0� ,������� �(���� ����4� �(� ������ �(� �� ������� �� �)� �(����g #��� � � ���%�� � � 9 �(� #��� � �(���� ������� � ���?� �(� �(� ��� � #���� �(� �##��� �� ����� �� � �� � ��� � �%������ .+����4� a ��( ��� � ��dh� ��� 0� +(� #��� � �(��� � ������ �� �(� �� �� )( �( �(� ������� � ����� �� �� )��� �� �� �(� ����� (�� ��� � ������� � �� �# �(� ������� 9 �(� ���%�� � �(���� ������� ����� � �(� ���%� �# ���( ����� �� �� � �(���� �(� ���#���� ������% ��� 8������ %� � �� �(���� ���� ������� ��� � �4 ������ ) �( � �� ���� ���%� .%� � ��0 #% �� � � 7 �( �������� �(����� �%������ ��� �������� �� �� �� � � .���( ��� � �� � ���� ��0 #��� � ��#��� �� �� �" �(� ���%� #% �� � � ���(��� ) �( �(� �%��� �� ���� #�� �� � � �� ��� #�� ������� : ' �� ������ � �������� �# �(� �������� �(���� ���%� #% �� � � �(�� �# ������ ���� ����� � � 9 ������ ����� �� �� � ���� � ���� ������� �(� �(� ����� �� �� � �� � 9 � � � �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% � � �� � ���(��� ���%� #% �� � � �(���#���� �(� ����� �# �(� %� � �� #% �� � � ������� #�� ������ �(� #�� �� � ) �( �������� �(����� �0'0 :��� ��� �� � �� ����� �� 7(����� -��4�) �? ���%��� �� ��� � ����� �� � �4 ����� � � ��( ��� � +����4� .��>�0 ������� �(� #����) � �)������ ���%� #% �� � g ��#� � #�� # # � ! ��#� � ����#��� # # � ! � )(��� # � �(� �%������ ���� �(� ������ ���%� ���� +�� ���� � ����%��� �(� �� � � � �� �� ����� ��� ����� �� � �� ������� � � � �(� ���� ����� � ���#'� � �� ,������� �(���� ' � ������� ��� ���� ������ �� ���� �� � � �� �� )����( ������ �(� �� )����( �� �� +(� ���� ����� � (����(�� � � ���� %�� �� ��� ��% �� �# ������ ��� � �" +����4� � ��( ��� .��� 0 ������� ������� ' � �%����� ���������� # �(� ������ � �� �# � �� � ����� �) �� �(�� �# � ����� 2� � � � �� ��� ������� � ����� %��� �(�� ' �(� ��������� ��� ����� �� �� � / #�� � .���� ����� � 0 � !�dd #�� � .� �4 ����� � 0� �0*0 �11� � � �� b�� �(� �%����� �# �( � ��% �� ������� ��� ����%�� �� (��� �(��� � ����� �# ������ )( �( �� ����� 7� ���%�� �(� ������� (��� � '�� �%� �# �� �� ��� ����� �� ���� �� � �� "!� +(� ������ � � ���%�� �� �� #�� � ������ � �� #����� +( � � �� #���� � ����� ���� ����%�� ������� ��� ����� ����%��� �(� � ����#�� � ���%� � �� �� �� �(� � �# �(� ���� �� �� �� �� '� � ���� ����� ���( ����� 7 �( ��� �(� ������ ���� �� �(� � � �� '�� �%� � � �4� ����� � ����%����� �(� � %�� ?� ���%� �� �(� � �# � ����� 7 �( � �k ������ � ��������� �(� � %�� �(����� �� ���� /!l �# �(� � �� )����( � � �4� ����� � �(� ���� � /!l �# ( � )����( �(� � �4�#��� ����� .+�� ���0� +(� ������ �(� ����%����� �(� � %�� ?� ���%� � ( � ����#�� � �� ���� � � 7 �( � m: ��������� �(� ���� � �� ������ �(� ���%� �# �(� ������ � �� � ��� ���% � ���( ��� � � 7� ���%�� � %��� ����%��� �(� ������ � �������� ���( �� �(� 9# �(� � %�� )� �� �(� ���%� �# �(� ������ � �� �� f� !�!! �� �(� � �# �(� ����� (� ������� �(� ���%� �# �(� ������ � �� f�!�!! ���( �� �(� 9# �(� ������ � ��� � � ��� � �� ���%� �� �( "�� �(� �(� �� �� �%� � �� �(� ����% � �� �( " � �� %�� �� �(� ���% � �# �(� ���%� ��� � "�� b�� �������� # � ������ ���� �� f�!�!! � ��� � �!l �(� '��� �� �(� �(� ���%� �# �(� ������ � (�� ���) �� f���!!� b�� ���� �� �(� �(� � %�� )� �� �(� ����� ������ �� ����� f !�!!� � �� �(� ����� #% � ���) �� f���!!� �(� ������ �%�� � � � � � �� f��!! �� �(� ����% � �� (��� �(� ����� ����� #% � �3%�� f !�!!� +(� ���� � �# �(� ������5� )����( .f� !�!! �( � �������0 �� ��� ����� �(� � �4� ����� � ����� �(� � �4�#��� ������ 7 �( m:� %� � �� �(� �# ��� � �� ���%� �� �(� ������ �� �� �%��� ���� �(� f�!�!! ���( �� �(" �� �������� )(� �(� �� �(�� ���%� � ���� ��� �(� ������ � � ���� �(� f�!�!!� +(� � %�� ?� ���%� � ����%���� ���� %�� �(� ����� � %�� �� �� �%� � ������� �� �(� �(��� ' �� ���%� �# �(� ������ �� � �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% b�� � �=: ���� � ��������� �(� ������ ���� �� �e� �# ( � � �� )����( �(� � �4� ����� �� "! � 4���� �(� ���� � )����( +�� ���� 8��( �� �(� � � � � �� �e� �# �(� ������5� )����( � ��� �#���� #��� +�� ��� �� �(� � �4� ����� �� �� ���� � �(� � � �� � �� )����( � ����� �(� � �4� ������ +(� � %�� ?� ���%� � �(� ����%���� #�� �(� ����#�� �� b�� ��� #�%� ������ � ������� ��� �(� � %�� ?� ������ ���%� � ��� �(� ���� � �� �(� �������� �(���� ���%� #% �� � �� ����� � � #% �� � ���%�� � ���� +(� ��� � ��� ��% � ��� ����� #��� 9������ � � 3%�#�� . !!!0� +(� �� �(�� ���%� � #�� ��>!c���� ��� %�� � +(� � �4� ������ �� � ��� ��� ����� ����� � ����4� .�a, /!! ������ ��0 � 9������ ����� ����� � ����4�� +(� � �4�#��� ����� %�� � 2��� +�� ���� +� ��� �%���������� � ������� �(� � %�� ?� ���%� ����%��� � � ��� �� �������� � ���%� �� +(� ���%� � ��� ����%���� ���� � �� �(�� ���%� � #�� $� %��� �(��%�( �������� �# ��>!� $� %��� �(��%�( �������� �# ��>�� ���� � �� � �����%� � �%�� � � �� �� �� �(�� ���%� � . ���� �(� $� %��� �##���0 � � �� �(� �(��� "! ������ �%������� ���%� � ��� ���� ����%���� #�� � �� �( ��� � � ��� � b���%���� -���(� ���� :��(�%�( �(� �����%�� � ?� �# �(� ��� ���%� � � ���%� #% �� � � �(� �� ) �( ����� �� �� ����� � ��� � �� �(� �� 4 �� �� �# �(� ���� � ������� �� � �� ������ � 7� ' �(� ���%��� �� �� �� � ��� � ����� ���� �# �(� ����� � �� �( �(��� � �(%� ������ �(� ���%��� #�� $� %��� �(��%�( �������� �# ���( ����� ���%��� #�� ����� �� �� ������ � ����� ���� ��� ��� ����� %�� ��3%���� � ������� 8����� ���%� � ��� ����%�� #�� ���( ������ � �������� #�� ���( ���� %� � �(� ������ ��� � � +(� ��� ������ ���%� � �(� ��� �� �� �� � #�� �(� ������ ���%� � #�� ���( �������� ��� ����%�� � �� �� ���%��� �(� ��� � �(� ��� �� �� �� � #��� �(� �k �������� ��� � ��(��# �# �(��� #��� �(� �� ��������� +(� ������ ���%� � ��� �(� ���� � �� �(� ���%� #% �� � �� ����� � �(� ������ � �������� �(�� ��� � ?�� �(� ������5� �������� %� � ��� b�� �%������ �# �� �(���4 �� )� ����%���� �(� ������� ���%� � � ����� � �� #�� � ����#�� � �# +�� ��� %� � �(� ��% � ��� � � +(� � %�� ?� ���%� � /�!@l ) �( � ��� �� �� �� � �# �>�l� 2� � �������� �(����� �(� ���%� #% �� � � ?��� �� �(� +�� �� ���%� � �(� ��#��� �� �� � #�� �(� ������� +(� ���%��� ��� �� � ��� � #�� ��� #�%� ���� � ������� �� � �� �(� ���� � ������� �� ����%�� ������ �� ��� ����� �(� � �4� ����� ��� ����� +�� ���� �0�0 2 � � �� +(� ������ �� � ��� �# &! �������� � � ����� � �(� ����� ��>!c����� +���� � ������� �(� ���%��� #�� �(� #%�� ������� 7( �� �(� �� �(���4 ����#�� � ���%��� ������ ���%� � #�� � �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% � & ����#�� � �# +�� ���� ��� �(� ������� ���%��� #�� ����4� ��� #�� ������ ���%� �� +( � � �� � ��� � ) �( ���(� � ������� �� ���� �%� ��% ��� ����� ���� �# �(� ����� �� � ��� � �� ��)��� � �� � �(� ������� ��� ������ ���%� �� � �� �=: #����� ������� �� 4��� �(� ��6�� �� �# �(� � )����( +�� ��� ��� � � ������ ���%� � %� � �(� '��� (��# �# �(� ����� � � �� �%��� � � �(�� �(� ��� ������ ���%� � #�� �=: ��� ��)��� ����) �(� ��� ������ ���%� � #�� �%�� �%� ���� �� +(��� ���%��� ��� �� � ��� � ) �( ��?�##5� ' �� .���@0� � 4�) �� �k ��)��� � �� � � ( �(�� ��� ���%� �(� �=: #�� ���( ����� � ����� ��� ����4�� �=: � �� � �%��� �� ������ ���%� � �� m: #�� ����� ��� ����4�� � �� �� ���� � ���%��� �(� ������� ������ ���%� �� � � �� �%��� � � �(�� �(� ��� �� � �� �# ������ ���%� � .�� ����%�� �� �(� ��� �� �� �� � 0� � ���� ��)��� �(� ( �(��� #�� �� ���� �� �� ��3% ��� ������� �� �%� ��� )����( �(� � �4� ����� %� #�� �� � � �4 �� ��������� � � ������ � �� �k ��3% ��� � ������ �� �� �� � ��(��# �# ( � )����( +�� ��� �(��%�(�%� �(� � #� �# �(� ������ �� �( � � � ���� � �4� ������ � ��������� m: ���� � ��� � �� ���� ����� � �� �(� �=: #�� ���( ����� � ����� ��� ����4�� +(� #�%� ���� � ������� �� ��� �� ���� ��� ��� � �� �#'� � � ��� 8������ %� � �� �# )����( �(���� ����� %� �(�� � �������� �(�� ��� � � ( �(�� ���%� � ���� � �4� � �(%� (�� � ( �(�� ��� �� �� �� � " �(� ��� � ��� ���%��� ��� �� ��� ������ b�� �������� #�� ����� ��� ����4�� �(� �=: �������� ���%��� � ��� ����#�� � ���%� �# @� @ ) �( � ��� �� �� �� � �# �>�/d� o�� �k ���� � ���%��� � ��� ����#�� � ���%� �# h�dd ) �( � ��� �� �� �� � �# �>�!�� +( � �#'� � �� �%������ =� ��� � �5� �� �� � � �(�� �=: � �� � ��� ��� � �� �#'� � � ���(� �# ���� �� : #���� ������ ) �( -��4�) �?5� ������� %� � �� #% �� � �(%� ) �� ���#�� �(� �k ���� � �������� �� � �=: �������� #�� ����� ��� ����4�� +(� ���� #�� �������� �(����� (�)����� � �� �� �(���� �(� �������� �(�� ��� �'�� -��4�) �?5� �� ��" � � �� �(���� � �������� �(�� ��� � ?�� �(� �������� �(���� ���%� #% �� � � +� ���� # �(� �#'� � �� �# �=: �������� #�� � ������ ) �( ��( ��� � +����4�5� .��>�0 ���(��� %� � �� #% �� � � �(� ������� ���%� #% �� � � ����%�� � � �(� ���%��� ��� ������� +���� �� b�� ����� � ����� ��� ����4�� ���( �� � �k ������� �� ��� �� ( �(�� ���%� #% �� � � �(� ���%�� #��� �(� �=: ��������� 9 #���� #�� ����� ��� ����4� +���� � : %�� ?� ������ ���%� � � ���%� #% �� � �� ��>!c���� �������� -�� ���� �(���� ��� � m��%� #% �� � 1�m��%� :����g ����� ����4� �%�� �%� �� d !�&� !�@/h &��/ !� d �%� � (�� @�h@ �!��� !�@/h ��/ !�d@ ����������� ������� @��> � �dd !�&dh ���! c m��%� ������� @�!� ���/ !�@&d ��@d !�d> :����g ����� ����4� �%�� �%� �&�> &@�!& !�@!& /�!! !�!!@��� �%� � (�� h�dh �>�!� !�@!& �> !�!! ��� ����������� ������� @� @ ���/d !� �h � ��� c m��%� ������� d�>d �/�/ !�/hh >��! !�!!d��� ��� 1 p !�!!�� � @ �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% .�(� ���� � �4� ����� ������ � �� � ��� 0� � �� �=: ��� � �� � ���� �� ���%� #% �� � �%������ ����� ��� ����4� ��� ����� ��� b�� �(� &! ����� �# �( � ��% �� �(� � %�� ?� ���%� � #�� ����� ���� �� �� #��� � ��) �# �&�l �� � ( �( �# /dl� 7 �( �������� �(���� ���%� #% �� � �� � �4 ������ �������5 ����� ��� ��� ���� ������� �� ������� 9 � � � � ������ ��� ��� �#���� �� �(� � �4� ����� ���� � �� � ��� �� ��� �� ����� ����� ��� ����4� %� � ��� � � �# ( �( ���%� �� +(� �=: �������� �(%� �� �� ?�� ������� #�� �(� �) � � � �4 �# ����� ��� ���� � � ��� ��%�� ������� �� ���� �(� ������% �� �� ���'� #��� �(� %�� � ���� � �� �# ����� ��� ���� � �� ���� � �(� ��� �#�� �# ������ #��� +�� ��� �� ����4�� +( � ���%�� �� ��� ��� ����4� a ,�� ��5� .����0 ' �� �(�� �(�) �=: � �(� ���#���� �������� #�� ����� �� ������ ��� +� ���� #�� �(� ##��� ��� ���%� #% �� � �%������ #�� �=: �� ������� �� �(� �(��� ����� �� �� ������ � ������� ��� 1����%�� ��� ������� � b�� ����� ��� ����4�� �(� �=: �������� ��� � �� � ���%� #% �� � �(�� � ���� �� � ��)�� �(� �(� ����%���� ���%� #% �� � � #�� �(� �(��� ����� �� �� ���� � ������� ��� +(��� ##��� ��� ��� ( �(�� � � '�� �� +(� ���%��� �%����� �(� (����(�� � �(�� �=: � � #�� �� �������� #�� ���� � ����� �� ������ .����� ��� ����4�0� 7( �� �(� ##��� ��� ���%� #% �� � � #�� �=: � �(� ����� �� �� ���� � ������� �� � �� ���� �� ����� � � '�� � #�� ����� ����4�� �( � ��� �� %� �� �(� ����� � ��� � �� ���� � (�� ?� � +���� : %�� ?� ������ ���%� � � ���%� #% �� � �� ���� � ������ �������� -�� ���� �(���� ��� �� m��%� #% �� � 1�m��%� :����g ����� ����4�� �%�� �%� ��@� �>� @ !�// /�d@ !�&@ �%� � (�� @�>� d�h !�// &��> !�>@ ����������� ������� @��@ �� h !�/& &�/h c m��%� ������� &�h> ��!& !�@� !�>@ !� � :����g ����� ����4�� �%�� �%� � �@@ /�/& !�@� h� d !�! �� �%� � (�� h� � �>> !�@� &�@� !�!���� ����������� ������� &��/ �&�h !�/� �!�dd c m��%� ������� >�!h � �h !�/h /�! !�!h :����g ����� ����4�� �%�� �%� �� d !�&� !�@/h &��/ !� d �%� � (�� @�h@ �!��� !�@/h ��/ !�d@ ����������� ������� @��> � �dd !�&dh ���! c m��%� ������� @�!� ���/ !�@&d ��@d !�d> :����g ����� ����4�� �%�� �%� �&�> &@�!& !�@!& /�!! !�!!@��� �%� � (�� h�dh �>�!� !�@!& �> !�!! ��� ����������� ������� @� @ ���/d !� �h � ��� c m��%� ������� d�>d �/�/ !�/hh >��! !�!!d��� �� 1 p !�!/� ��� 1 p !�!!�� � ��/!c����� � �� hc����� �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% � / 9 ������� ��� � ?� ���%� ) �( �%�� �%� ���� � #�� ����� ����4� � ) �( m: #�� ����� ����4�" �=: � �� �(� �#'� � � ���� � �������� #�� � �(�� ������ +(� ���%��� �� ��� �� ������ 5� ������ � �(�� �=: � �(� ������� ��� ���� � �������� #�� ���� ������ �������� �0'0 ,#������ �1�� +(� ��� ��� � �(� ���� ���� � &! ������ 7� ���� � ���?� �(� �����7��� 7�� 99 ��� .��/!c����0 � ( ���� � ���� � � �� #����� .�� hc����0� �(� ���� � �� ��� � � �� ��� ������� �(� 7 �� ��� � ���� ��% � .���&0� 7� ' � � � '�� � �(� �� �(� ������� ���%���� +(� ���%��� ��� 3%�� ��� ���� � � ��� �� �(��� �(� � �� ������ � ������ +���� � 7 �( ������� �� �(� ���%� #% �� � #�� ����4�� �� � �%��� �� �� �=: #�� ���( ����� � ����� ��� ����4� #�� ���( ����� �� �� � �� ��� � �� 7 �( ����� ��� ����4�� �=: � �� � � ���� �� ���%� #% �� � %� � ���( �(� ��/!c���� � ( ���� � � �� ��� � � � �� �(� ����� ��� ����4� ��� �(� ���� ����� ��� �(� �=: �������� � �(� ����� �##��� �� �� ��� ��� ? � � ��� � �� ������ ���%� � � � � ? � �(� ����� �# ���� �� ���%� ��� � � #�� �( � ����� ���%� �� +(� ���%��� #�� �(� ����� �� ��� �%����� �(� ���%��� #�� �(� #%�� ������� �=: ��� ������ �� �� � �##��� �� ������ � �������� #��� � �������� �(���� ���%� #% �� � �������� �� #�� ����4�� !��������� -� ��� �� �� � ������� �� � �4 ������ ) �( ��� ���� �� ���� %� � �� #% �� � � � �� ���%������ �� �� � %�� ��(�� ��� :���� �� �� ��(�� ���� %� � �� #% �� � �� ��� �������� � ������ ���� ����� � (��� ��� ������� � +( � ����� ��� � ����� ����� �(� �������� �(���� ������ ���%� #% �� � � 7� ' �=: ���� � � #�� �� ���� � �������� %� � �(� ��������� ���%�� ��� ���� �� +����4� � ��( ��� #�� ����4�� +(� #� �%�� �# �=: �� � ��� ��� ���� � �������� #�� ��� ������ ��� �� ���� �%����� �� �� �� ��������� � ����� '��� � " (�)����� �(� ���� � 4��� ��� �� � � �(�� �=: � � �� ������ �� ���� � �������� ���� �% �� #�� ������� ������� � #���� ��� �� ��� �(�� ��� � �(� �� �%��� � �# ��� ��� �=: � �� � ������� ��� ������ � �������� #�� ����� �� ������ �%�( �� ����� ��� ����4�� �=: ��� #� � �� �� � ��� ��� ���� � �������� #�� � �4 �� ������ ����%�� ���� ����� � � � �4 ����� � ���#'� � �� ��� ������� ���� � k�)����� �(��� ��� ����� �� �� ����� � #�� �(� �=: �%����� ��� ���%��� #�� ����4�� ���� � �������( ����� �(� � �� (�� ?� � �(� #��3%� �� �# ����#�� � ������ � � ��� ����� ���%��� .����4�� :���� a ,�� ��� ����" �(%�)��� ���>" +(������ ���/0� +( � ��% � ���%��� �(� �=: ������ ��� �#��� ������ #��� �(� � �4�#��� ����� �� �(� � �4� ����� � � �� �(�� � �� ��(� %��� 7� ���� ���%�� ������� ��� %� � � � ���� � �� #���� #�� ����%��� � ����#�� � ���%� �� : 3%��� � ���� � �� �� )(��(�� �=: � 6%�� '� % �� ����� �� �� ���� � � �� (�� ?� � �� % �� ����� �� �� ���� � ����� #% ��� �#�� ��� � �� +(� ���%��� �( � ����� #� � �� �%����� �(� �� ��� �� �# �=: �� � ������ � ��������� � h �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% ��"����� �#���� +(� �������( )�� ��� ���� �� � ��%�� � ) �( $����( � � � � �� ��� � �(� #������ �# �(� ���������� ������ � ������ � ��������� +( � ����� �� �'�� #��� (���#%� ����� �� �� : -��� � $��4 k��� ��� � ���� ��(��� �)� � � ���%� ��#������ � ���� � ���� � �� �� �� �(� !!� )��� b � �� :���� �� � ���� �� +(� �%�(��� ��� ������ ����� � ��� #�� � � ���� � ������� ���������� :�����4���� �� ,�� a ���� ������ 8� �� . !!!0� : � �%��� � �� �� ���)�� �%�� �%� � ���������� ������ �� +���� � �� 4 � �� � �� �� �!� �� ���� � ,� 7�� 7 �� ���� �� 8�� a :� �� -� b� .���>0� ���� ���������� ������ � )��4 #�� �� �� +���� � �� 4 � �� � �� �� �!� � .&0� >dcd!� =� ��� � ��� ;� -� .��>�0� : ��� � �(� �%���� ��� �� �# ���������� ������ � �� � ������ � ��� ��� +���� � �� 4 � �� � �� ;� �� � � �� �� �� � ��. 0� @@&c@/!� 8 ���� � -� 8� .��dd0� m��%� ������ �g : �) �������( �� ���%�%��� � � ���� � � � �� � �� �� :�� � �� � :��� ��� +���� �� :%�%��� ��c�@� k��� ��� � $� $� . !!�0� +��� � ���%� ������ � �(� ���!�� 7��4 � ������ 9������ � �� ;�� a � 3%�#�� � �� :� ����< � ���� � � �� �� �= � ��. 9 �� � �� �� � �� ' �� ����<� =( ����g 9������ :���� ����� 9������� � =� �� .����0� �%�� �%�� ��4� �(� � �%���� 4 � �� � �� �� �!� $� %���� ��( ��� � ��� a +����4�� :� .��>�0� ,������� �(����g : � ���� � �# �� � � ��4 � % �� � �4� ,�������� � � �%� h&c ��� � �(�� $� ��� a -� ���� �� .���&0� <��� � �� � #��� ���������� ������ �g � ���� ���� %��� ��� � ��� � ��� ���%���� 4 � �� � ���� �� 5�� �/� '.�0� /�ch�� -��4�) �?� k� .��/ 0� +(� %� � �� �# )����(� +���� � �� ��� � � � ,������� � � �/�c�/d� ����4�� -� :�� a ,�� ��� �� 8� .����0� ����������� ������� ��� � �� ���) � �� ��� � ��(�� ���� ' � ��� o��4 2 ���� �� 7��4 � ,���� ��� ��� ,��� ;� .��>�0� ��� ��� � �� #�� ���� � � ����� � ����� ������ �� � � !����� �� ����� �%.>0� &>�c &�&� ��?�##� -� �� .���@0� �%����%� ���� � ����%� ������������ �� "�� +���� � �� ������� � � � !������ 7 ���� @/c/!� ��(����� �� .���d0� =��� ������ �g b��� �� ���(� +���� � �� �� � �� ������� � � � !������ 7 ���� �(%�)��� +� .���>0� 8���� � ���%� � ) �( ���� ����� � � 7��4 � ������ ������ � -� .���/0� : ��(�� ���� #����)��4 #�� ���������� ������ �� "�� +���� � �� ������� � � � !������ b���� >!c>d� +(������ �� .���@0� +(� #������ �# ���������� ������ �� 4 � �� � �� �� �� �� ,��� � ��� b���� 7 ���� �&dc �@&� +(������ �� .���/0� +(� � ��� ���� '��� � �� ��������� 4 � �� � �� �� � +���� �� >�.&0� hdc>h� +����4�� :�� a ��( ��� � �� .��dh0� ��� � �� �(� �� � �(� #��� � �� � � �� +���� � �� �� �� � >(.@0� � /�c� >d� +����4�� :�� a ��( ��� � �� .��� 0� : �� ��� �������� �(����g =%�%��� �� ������� ��� � �# % ����� ��� +���� � �� 5 < �� ������ ���� >� �>c& &� m���� ,� ,�� a -�; �� $� �� . !!!0� +(� ����� ������� � �� � � ��� ���� �g ����� � #��� ���������� ������ �� 4 � �� � ���� �� 5�� �/� (� @>ch&� 7 �� ���� �� 8�� a ���� � ,� 7� .���&0� �%�� �%� ����� ���������� ������ �� +���� � �� 4 � �� � �� �� �!� �. 0� h@ch>� �0�0 -�!! �� 20 �� 3 4 � �� � ���� �� 5�� �/ � 6' �7 ��%8�'% � > does loss aversion explain dollar-cost averaging? introduction investing strategies lump sum investing buy and hold strategy value averaging dollar-cost averaging review of literature methodology prospect theory investor utility function application data results data analysis extended sample conclusion acknowledgements references pii: s1057-0810(02)00094-x ����������� ��� ������� �� �������� �� ��� ��� � � � ��� ����� ��� ��� � ������ ���� ��� �� ��� � ��� � �� ��� � �� � ��� ���������� ���� ��� ������ �� ����� ��� ������� ���� ���� ������ � �� ���� � �� ��� !� ��� �"��� # �� ����� ��� ������� ��� ��������� ���� ������ $ �%% &� ���� '���� ��� � #�� "� !($)$� ��� �������� � ��������� !!"# �������� �� ������� $��� " %������� !!"# ������� "& '������� !!" �������� ���� ����( �������) ��� ��� ��� ������� �� �������� �� ���� ��� � � ��� � � ��� �� ��� *+�+ ��$���� �����)� �� $��� ��� ��� � ,� ��� ������+ *���) �� ��-��������� ���� �� �� ����.�/"�"0-1 $� ��2�� � 2� �������) �� ��������� ����� ������� �$ ��� .�31 �( ��������) $�� ��) ���� �� �������� � � �� ����� �$ � �� ��� ��� + 4� ,�� ��)��,� �� �������� �� �������� ���2��� ��� ������� �� � �� ��� �� � � ��� �� ���� ��� � 2� � ��)��,� �� ��) ���� �� �������� �� ��2��� ��+ � ��� 2� � �� �� ��� ,����)� �( ���2��) �� � ��� ��� �� ��-��������� � � �� ������ �$ � �� ��� ��� �5� �� ��� ����� ����� � � ���+ 6�� (� 2� �������) �� ��2 ��� ���� � � �� ���� �� ������� ������ ����� �� ������� ��� ������ �$ ������� ��������� � � �� � ���� ���� �� ���� ��� ��)(+ 7�� ���� �� ������� ������� $�� ����� �� � ��� ��)( �� � �������� ��������� � � $� �2+ � !!" � ������ ������� ���+ � ��)��� ��������+ *+, ���������������" # �"� . �/����.�31# .�������� ��� # 1 � �� ����� �$ ��� # 7����� ����� �(��������# ��� � ������ � ���� ������ ��� . ��� ����� 3�����) 1��� �� �5����� ( ���� � ���) �������� ���������+ �� ��� ��� ��� $ ����� ��� ���� �� ����$� �� �� ���� �� �� ������ ��� ��)���+ ��� � �� ���������� � �� ( ��� � �� �$��� ���� � ���5( $�� �� �����) ��� ����������( �$ � ���� �� )���� �$ ����� �� �� �)� �� ��� ����� (��) ������ � + �� 6 � /"88"0 2����� 99� � �� ���$������ � �� � ������ ��� ���� �$ ��� �� ����� �$ ��� ::+ ���� �� ,�� � ��) � ��� ���� �� ����� ��� 2��� ��� ������ � /��� 3�����)� � ���� � � �� 1 ���� /"88�0 �� 6 ������ /"88;� !!!0� 2�� ,�� �������� � � ������� $�� ��� .�31 �� ���� ���� � � ���0+ 6�� ��� �������� �����2 "! / !!"0 <8� ; "!�;-!="!>!">? � ��� $���� � ���� � !!" � ������ ������� ���+ � ��)��� ��������+ 3��@ � " ! � ; ! = " ! / ! 0 ! ! ! 8 < -a � .�����������) �����+ �� +@ �"-=" &;!8&# $ 5@ �"-=" &;-<= !+ +0 ��� ���� ���$)�� �b��$ �+����� ��+��� /�+ ��� ��0+ ��2����� ����)��) � � ��� �� ��� ������ �� ( � �� �����) ����� ���� 2��� 2�� � � � �� ������ � + �� �� � � ��� ��� �� �� ���) ��� �� ��������� �� ��� ��2 �(���� ��2 �� ��� ��� .�31 �� ������ ������ ����� �� $���� �� �������+ ��� ������� �$ ���� ����( �� �� ������� �������� ��������� 2��� ��� �� ����� �� $���� �� �c���( ������� �� ����)��) � ��� � � ���+ �������� (� 2� ,��� ��� � ��� �� � �� �� ��������� ����� �� �������� ���2��� ��� �� ������ �� � � ��� 2��� ��$$����� ��)��� �$ ����)� ���� 2��� ��� 2�� � ����$� ��+ �� � ����� �� 2� �������) �� ��� ��� ������� �� �������� �� *����� �� ��� �� ���� ��� � � ��� � � ��� ���� ��� �� ��� ��� � ������+ �� ��� � ��� �� � �� �� ��������� ����� �� �������� ���2��� ��� �� ������ �� �c���( � � ��� �� ���( �������� � � + *�� (� ���� ������ �� �� ���� ��� ��� $�� � � ���� ��������+ ��2����� ����� �� ����� ���� ��� ����� ��� 2��� ����) ��� � �� ���� �$ ��� + 6���� (� ��� �� � ��� �� ��� ����� ���� 2��� �������� �� ) �� ������ � �# � �� /"8;;0 ������ ���� ���� �$ ����� ������ � � �� �� �-� �� ��� �$,�����+ ������ (� 2�� � �� ������ ( ��)��,� �� ��� � � �� $���� �� ��� ��)� ( ����� ��� ���� ���� � � ���� � �( ����)��) � � ��� � �� d��� �� ��� � ����� ���� 2��� ��� 4�� � ������ � # �� $ ��� �� �� ��� ������ � �� ,�� �� �������� ���2��� � � �� ��� ����� �� 2�� � ��� ����� �� ��)������ � � ���+ ����� (� ��� � �� �5����� ( �� �� � �� � �� ��� � �� �� ��� ���������� �$ $����� ��� �# $����������� ��� ��� ����)��) � � ��� �� ���� � � ( �� �5�������� ������-�� ��� ���� )���� �� �� ��( ���� � �� ����� ���� ���$��� ��� � �� ���� ���� ��������� � �+ e������ ��� � ��� � � �� �� �� ����� ����) � �� ��� �� ��� �� ��� �$ ������� /��� � ���(� "88=� !!!0+ �� � � ���� ����� 2� ��� 1�.� *� �� 1�.� �. 4�� � � � ( ����5 ������ �� )���� �� � � ( � �� ����� ��� �� ���� �� �������� ��� � /��� � ��� �$ ��� �� ��� �� 1�.� �. 4�� � � �� ����0 2��� �� ��� �� �� ���.� ���� # 2� )� �� ��� ,����)� �� 6�)+ "+ %����� �� � ��� ��� ������ ��� �� ���� 2� � ( �� � ��� � �� ��� �� ��� �� ��� ������+ 1�������� �� �� ������� �� � �(��� 2�� � � ���� ������� ��� ����) ��� 99������-��� ���:: � �� ��� �� ��� �� ��� ������ �� ��� ��)� ( ����)� ��� *+�+ � � �� / ��� )� ����� ���� �$ !+;" 2��� ��� 1�.� �. 4�� � $��� "88! �� !!"0+ %�5�� 2� �� � 6�)+ � 2���� ������� � � ( ��� � ��� �� ��� �� � �� ��� ������ �� ���� ��)������ ����)��) � � �� ���� � ��� �� � � ��� � � �� / ��� )� ����� ���� �$ !+"= 2��� ��� 1�.� �. 4�� � $��� "88! �� !!"0+ 4� ������� �� � � � ( ��� �� ��� �� ��� ������ ���� �� ����� d���+ ���� ������ ���� �� ������ �� ��)������ � � ��� �� ����� ��� �� � � �� ��� 2� �� ���$����� �� �� ���� ��� �� ��� ����)� ��� � ��� � � ���+ �� $ ��� 2������ � � ��� �� ����)� ��� �� ��)������� � �� ��� ����� �� �� ������ ( � � )� � � ���5( $�� ��� + �� �� � � ��� � 10 ���� ������� �$ ��� � � ��-� ��� .�31� 2���� �� ��� ��� ����� �� �������� �� � �� ���� ���� � 2�� + %�2� 2� � �� ���� �� ��� ��� 2� ���:� ��2 ��2 �� �� �� ��������� ����� ������� �� �������� �� � �� ���+ �� $ ��� ��������� ����� �� �������� ���2��� ��� ����� �� � �� ��� �� ������ �� $ � + �� � � �� �5� ���� ���( ���� (+ +10��� ��� ����� �� �������� � � �� 10���� ��� ����� �� ������ ( ���� � ���5( $�� 10��� ��� ����� + +10���� ��� ����� � � �� ������ �������� �$ � � �� ������� �� )�� ��� �� � ��� ��� -$��� � ��� �� ��) ���� �$ ��� -$��� � ��� �� )�� ��� �� � � � �� �������+ ����� ��2 �� �� ������ � �� �� �� ��������� ����� ( � �� � � �� � � � ������ �� �������� �� ��) ���� 2��� � �� ���� 2���� �� 2 (� ��������f ��� ��2�� �� ���� �@ ��� � �� 2 ( �� � 2� ��� � � �� �� $������� ���� �! +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( �2� �����-2��� ��� � $��������g�+�+� �������� �� �� �� ��) ���� �� ��+ ������� ������ ������� �)�� �� � ��� .�31 �� �� �$ �� �� ������� �����2��� �� ���� �� �� ��2��� ��+ 6�� �5 �� �� 3�����)� �� + /"88�0 �� 6 ������ /"88;� !!!0 ��� �d� �� ��-��������� .�31 �� � �������� � �5� � ���� �� �����-�������� ���2��� ��� �� ��� ������� �� ���� ���� � � ���+ �� ��� % � �� 3(�� / !!"0 ������� � �(�����( �� ���� ������ ��� ����� �� ���� �� �� ��2��� �� �� ��))��� �� � ��� ���� �5������� �$ ���� �(�����( � � �� h����,�� �( ��� �5������� �$ � ������ ����� ��2 �� ���� �( � �� ��� ���+ 6�)+ "+ ' � ( � �� ���� ��� �� ��� �� ��� ����) 1�.� *� �� 1�.� �. 4�� � �� � ������ ������ /i �� �( "� "88! �� i��� "� !!"0+ +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( �" �� ��� ������ � �� �$ ��� ����(� 2� ���� $�� ������ ����� �� ��� ��� � . ��� 1 � ���+ ������� �$ ���� � ���� �� ��-��������� �� �������� ���2��� ��������� ����� ������� ���� $�� ��� �� $���� ���� � �� ��� �5����� 2� ���� �� ( � �� � ���������� �� ������ ��������� ����� ����� ) �� �� ���g�+�+� ���� ( ���������� �� ����������� ���������: ��� ����� ��2 �� ���� �( ��� ���+ �� ���� �� �$ 2� � � ��������� 2������ ������ �� ��������� ����� ( ��)��,� �� ( ���� �� �� �2 ����� ��������� � ��� ����� ������ � � � �� ��������� $��� � �� � � ���+ 7����� ����� �� ���� � � ( �� �� �������� �� ����)��) � � ��� ����� (���)�� � � ��� �� ���� ����� �� ������-������)��� ���� � �� 6�)+ + ' � ( � �� ���� ��� �� ��� �� ��� ����) 1�.� �� � �� �� 1�.� �. 4�� � �� � ������ ������ /i �� �( "� "88! �� i��� "� !!"0+ � +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( ���� �������� ������-�� ��� ������) �� � ��� ��� �����+ �������� (� �� ���� ��)������ � � ���� ����� �� � ������ �� ����$� �� ����������� ����� �������+ 6����������� �������( ��� � �� �� ����� ����� ��$$������ �� ������� �� ���) 2� ��$�������� ��$$������ �� �� ��$��c���� �� ���) � � �� ���������� �� �� � � �� ���$,�������� �� ��� ���� ���� � ��� � � ���+ 6�� �5 �� �� j�� �� 7� /"88"0� e� ��� �� � ���( /"88�0 �� � ���( /"88�0 ���� ��� �� � ����)��) � � ��� �� ���� ����� ���� � ��������� �( ���-h����, � � ���� ����� �2��)�+ 6�� (� 2� �5 ���� ��2 �� ��-��������� $���� ���� �$ ��� .�31 � � ������� ��� ������ �$ ������� ��������� �����������+ ������� �$ 2� � � �� �� ��� �� ������� ���� ����� 2� ������� ( ������� ����� ��������� 2��� $���� ����) ��� �� ��)��,� �� ( ����� �� ��� ������ �$ ������� ��������� ����������� �� ��$$����� � ���� �c���( � � ���+ �� ���� � ���� 2� �5 ���� �� ��-��������� ���� �� � � ��� ( �� � �� �� � ��) ���� ��� ������� �� � � ( $� �2�� �( ��) ���� ��� �������+ �����$���� ���� � ��� ��)( 2�� � ������� �$ /"0 ��������) �� ��� ���( e� �� �$ ��� ������� � � �� ��� ������� �� ��) ����� �� / 0 ��������) �� �� � � �� �$ ��� ������� � � �� ��� ������� �� ��������+ 4� ������� ��)� ����� �� ��� �$ ���� ����� � ���� ��� ��)( �� ���� � � ( �� ������� ������� � � � ���� ��� �� k�����) ����+ ��� ��� ����� �$ ��� � ��� �� ��) ��d�� � $� �2�+ �� ������� � 2� ������� ��� ������� �)( �� ����������� $� ��2�� + ������� & ��������� ��� � � �� �������� ��� ���� �( ����������� ,����)�+ ������� < ������� ��� ������� ���� ��+ ������� � �������� ���� ���� ����) ��� � �+ � ����� ����� ��� ������� �$ ��� ����( �� �� �������) �� ��������� ����� �� �������� ���2��� ��� ����� �� �������� �� �� �� ��( �� ��� � � ��� � � ���+ ����� 2� ,��� ���� $�� �� �������� ���2��� � �� ���� ����5 ������ ������ �� �������� ��������� � �� ��� 2��� ����.�/"�"0-��-�� � �c� ����# �� �� � ������ 2� ��� �� �� ��� ������� �)( ���� �� �������� �������+ ������ (� 2� �� ��$��� ��� ����.�/"�"0-��-�� � ���� �� ��-��������� ����.�/"�"0-��-�� � ���� �� ����� �� � ����� 99�� ��-���������:: � � �� ������ �$ � �� ��� ��� + ����� (� 2� �������) �� 2������ ���� �� �� ��2��� �� � � �� ����� �$ � �� ��� ��� ��� ��� $��� ��� �� ��-��������� ����.�/"�"0-��-�� � �5� �� ��� ������������� �$ ��� )� �������� �� ��) ���� ��� ����� + 6����� (� 2� ���� $�� ������ ����� �( ������) $�� ��� �������� �$ �(������� ��������) ��� ����� �� � �� � � ��+ 6�$�� (� 2� ����� 2������ ��� �� ��-��������� ���� � � �� ���� �� ������� ��� ���������� �$ ������� ��������� � �+ �� ��������� �� ������� "� � �� ��� ����� �� � ���) ���� �� � �� ��� ��������� ����� ��� ����� �� �������� �� ��� �� ��� �� ��( �� ������� ������ �����g �����-��� ��������� �� � � � �� �5� ����� 2��� ��� ��-� �� ����� �� � ����$� �� ��� ��)���+ � � � �������� ����� � � ��� ��� � ������ �������� � ��� ������) $�� ������) ������ �����+ ��������) (� 2� �������) �� ��� ��� ������� �� �������� �� ��� � � ��� � � ��� �( ������) �2� ���� �� 2���� �� ����������� � �� ��� �$ ��� ���) �-$ ���� .�31+ ��� ,��� ���� �� ��� ��� ��������� ���-$ ���� .�31 �� � ����� � � �� ����� �$ � �� ��� ��� � �� ���� +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( �& $���� ���� � �� ��� �� �5� �� �� �d�� �������+ ����� 2� ���� ���� �� � �� 5�� ��� ��������� ���-$ ���� .�31 �( ��������) $�� �� ��-��������� �����,� �����+ 6����� 2� ��,�� ��� ���) �-$ ���� .�31 �� ���� ( ��)������ � � ��+ ��� � � ��-� ��� .�31 �� 2������ � $� �2�@ +��"��3���� � +��"��3���� � ��� � ��� � ����� ���� �� /"0 2���� +��"��3���� �� ��� �5������ ��� ������� �� � � �� �� � ��� ��� $����� �� � � �� � �� ��� �� +��"��3������ ��� � ���� �� ��� � � �� ����� �$ � �� ��� ��� / �� ��2� � ��2 �� �� �� � �� ���0+ �� ����� �� ���� �c+ /"0� 2� $� �2 6������ ���2��� �� �� �� �)� /"8=;0� � ������ i ) �� �� � �� ��� � /"88&0� �� ����))� /"88=0� �� ���� $�� ��� �� �������� ���2��� �� �d�� �5���� ������ /� � �� ��� �������0 �� � � �� ��������� �� �� ��(+ .�������� �� �� ��( �$ �5���� ������ �� ���� �� ����) ���.� ������� �)(+ �� $ ��� 2� ��� ����.�/"�"0-��-�� � ���� �� ����� �� ������ ������������ �� ��� �(� ���� �$ ��� ��������� ������ �������� � ����� �� �� ��( � �������)� �� �2 $�� ��� 99 ���� )� �$$���::g�+�+� ��� �(������� ��� �� �$ )��� �� � � ��2� �� ��� �(� ��� �$ ������ �������+ %��� �� � 2� � �� � � �� ����� �$ � �� ��� ��� � �����-������ ( ����� �� �� ����� �� � � � �� ����� �$ �������( ��� �� d���+ �� �� 2� ��2� �� ��� ���� ���� �� � � � �� ����� �$ ��� �� ����>�� ��-��������� �� � �( ������� ���������� � �� � �� �� �� ��� �( �������+ 6�� �5 �� �� � ���( /"88"0� '���2��d� � �� �� 1 �� � � /"88=0� '�� � �� �� �� /"88�0� '� � ���� �� ��� �� /"88;0� �� i �� .��� �� ���) / !!!0 ��))��� �� � �������( ��� �� � � ( �� �� � ������ �� ������� $ ���� $�� ������ ���� �����+ �� �� ��� ��� ������ �� ������� ���� �$ �� ��-$ ���� ������� �$ ��� .�31# ��� )� �� ���� ( �� ������� ��������� � �� ��� � �� �� �� �� ���� �$ ��� �� ��������� ��� ����� �� ��)������ �� ����)� ��� � � ���+ �c+ /"0 �� ,��� ���� �� 2��� ����.�/"�"0-1 ���������� 2���� 99/"�"0:: ��$��� �� ��� ������ �$ )� �� ��� ����� ���� �� ��������� � �� ��� ����� ���������� (� 99-1:: �� 99-��-�� �:: ��$��� �� ��� �������� �$ � �� ��� �� ��� �� � �c� ����+ ��� �����,� ����� �$ ��� �� � �� � �� ��� �c� ����� $�� ���� /"0 �� ���2� � $� �2�@ �3��� � �� � ��� ��� � ���� ��� � ������� �� � ��!� " � � ��� � �� � �� ����" � �� ����"�����" � �� ����"� �� / 0 2���� �3��� �� ��� �� �d�� ��� ������� �� � � �� �� � ��� ��� ��������� � �� ���� ��� ���$,����� �� / ����� ������0 �� �5������ �� �� ����)��,� ��# ��� ���$,����� �� �� ��� � � �� ����� �$ � �� ��� ��� # ����" �� ��� ) �$ ��� �c� ��� ������ $��� ��� �� � �c� ���� /��� ��.� ����0 �� �������� ��2� ���� �� �� ��( � �������)# � ����" �� �� ������:� $���� �� � �� ��� /���.� ����0# ���� � "� �$ ��� � !� �� ! �����2���� �� �� � )��� ��2� / ��� � !0 �� � � ��2� / ��� � !0 �� �2�� �� � �� ��$$����� ��� �� �� ��� ��������� � �� ��� /)��� ��2� � � � ��� �� �$ ��� 2�� � � � ��2� � � � ��� �� �$ �� � ��0# �������) (� �$ �� � !� ���� )� �$$��� �5���� /� � ��2� � � )�� ��� ��� �� �� � )��� ��2�0# �$ �� � !� ��� ��2� ��� �� �� �(�������+ �� ������� ������ �� �d�� � � �� ��� ������� �� ���� � � ���� ��� ����� �� �$ ��� �5������ � � �� ��� �������+ .��������� 2��� � ���� �5���� ������ ��� 10��� � � �� �< +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( ����� �$ ��� ���� � 2 (� �� �������� /��� e� ��� ��� ��) � �� 4�� ���)�� "8==# .���� "8==# ����))�� "88= 2�� ������ ��)��,� �� �������� � � �� ����� �$ ��� 0+ ��2����� 10 ����� ��� � � �� ����� �$ ��� � ( �� ��) ����� � ����� � ( �� ��2��� �� � � ���� �� �� � 2�� � ��� ( ��) ���� ��� ������� /��� � ����� �� +� "88&# 3�����)� �� +� "88�# 6 ������� !!!# �� �� �� +� !!" 2�� ������ ��)��,� �� ��) ���� � � �� ����� �$ ��� 0+ �� ���� �� ����� ��)�� �� ���� ���( ���� � �� ��������� �5� � ����� $�� ��� ���k�����) ,����)� �$ �������� �������+ �$ �� �d�� � � �� ��� ����� �� �������� /���� �� � � ��0� � � �� ����� �$ � �� ��� ��� �� ��������+ �$ �� �d�� � � �� ��� ����� �� ��) ���� /��2��� �� � � ��0� � � �� ����� �$ � �� ��� ��� ���� �� ��) ����+ ����� 10���� $ ���� ��) ���� � � �� ����� �$ ��� �� ����� ��� 2��� ��2��� �� � � ��� �� �������� � � �� ����� �$ ��� �� ���������� 2��� ���� �� � � ���+ �� $ ��� 3�����)� �� + /"88�0 ������� ���� ��� �� �� �� �� � 99��� �5������� �$ �)� ������ �$ ��) ���� � � �� �5���� ������ ������� ��))���� �� � �������� ������� �� � ���� $�� ����������� �������� ����� ���� ���2��� ��� �� �� �d�� ������� �� �� ��� ) ���� ,����) �������� �� ��������::+ ��� ������ $������ ��))��� ����������) �� ��-��������� ��� ������� �� ��������� 2���� ������� $�� ��� ��) ���� ������� �$ ��� �� �d�� � � �� ��� ������� ������������+ �� ��� � �� ����� 2� �� ��$��� ���� /"0 ���� �� ��-��������� . ��� ����� 3�����) 1��� �( ������) �� � � ��� � � ��� �� �2� ������ ��� � �� ��� �$ � ����g�+�+� �� �� ��2�+ 1��� / 0 �������� �$ ��� $� �2��) ����.�/"�"0-1 ��������� ����@ �3��� � �� � ����� �� ��� � �����2��"� � � ��� � ���� ��� � ������� �� � ��!� " � � ��� � �� � �� ����" � �� ����"�����" � �� ����"� �� /&0 2���� � �� ����( � �� � � �� � � �� ��� � �� �$ " �� � ���� �� /�������� ��������� ����� ��� �������0 �� ! �� ��2��� �� ���������� /��) ���� ��������� ����� ��� �������0� ����� ��� � � �� ����� �$ � �� ��� ��� �� ���� ��� �����2� ��� � � �� ����� �$ � �� ��� ��� �� ��2��� ��� ��� ����� � �� � �� � �� ��� � �� ��,������� � �� � �� �c+ / 0+ 1��� � /"0 �� / 0 �� ������ �� ��� "! ������( ������� �� ����) $��� i �� �( "� "88! �� �����) �� i��� "� !!"+ 4� ��� e� ��� ���4�� ����)� ������� �� ������(-���������� ��� �� ��� �� ������� ��� c� ��-� 5���� � � ����� /l1j0 ��� �� ���� �� �� �� �� ������ � ��������� �( e� ��� �� �� 4�� ����)� /"88 0+ �� ���� ����� 2� ���� �� � �� �� ���� / 0+ 4� ���� �� ��2 �$ ��� ���� 99� �� �����:: ����� � � ���+ �����$���� 2� �������) �� ��� �����-�������� �$ ��� )� �������� �� ��) ���� ��� ������� 2��� 99��� ���:: �������� /�����0 �� ��) ���� /�����2�0 ����� �$ ��� ��� ���� $��� ���� / 0+ ��������) �� ��� .�31� 2� ���� � ,�� �� � � �� ������( � � ��� �2� ��� )� ��2 �� �� � �� ��� ��� + �������� (� ��� �� �������� ���2��� ��� )� �������� /��) ����0 ������� �� �������� /��) ����0 � � �� ����� �$ � �� ��� ��� ���� � �� ��)��,� ��� ������� �� ��� �# ,� (� ��� � ���� ���� � �� ��� � �� �� ���� �� �� ��2��� ��+ �� �� � � ��� �� 2�� � �� ���� �� �� ����� ���������� ��� )�� ��� ��� ��2 �� �� �� � �� ��� ��� � ��� �� �� ��� ��������� ����� ��c����� � �� �$ ������+ 7��� ��� ��������� ����� �� �������� ���2��� ��� �� ������ � � ���� � �� ��� ����� � � ���� 2� � � ��� ��� ���� �� ������� ������ �����g ��������� �� � � � �� ( �� �������� 10����+ *���� � ���� �5���� ������ ��� �������� �� ��) ���� � � �� ����� �$ ��� ���� � �� �c� �� ��������(+ �$ ���� �� ����� ��� �� � ��� � � �� ������ ��� �� �� �)� �� +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( �� � �� ���+ 6�� �5 �� �� �$ ��� ��) ���� � � �� ����� �$ � �� ��� ��� ����� ��� 2��� �� �d�� ��) ���� ��� ����� �� )�� ��� �� � ��� �������� � � �� ����� �$ � �� ��� ��� ����� ��� 2��� �� �d�� �������� ��� ����� � �� �������� ��� �� � ��� ���� � � �� �5������ ��) ���� ����� ���� ���2��� �� �d�� ������� �� �������� �� ��� + �� ���� � ��� �� �� c���� �)�� �� ���� ��� �� � � � ��� �5����� �������� ��2 �� ���� �( � �� ��� ��� /��� � ����� �� +� "88&# �� �� �� +� !!" $�� ��$������0+ �� ����� 2����� ��) ���� ������� ���� �� ������ $ ���� �� � �������� �������+ 7$ ������� ��� ���� ���� �� �� ����+ �$ �������� ������� ������ $ ���� �� � ��� ��) ���� ����� ����� �� �������� �$ ��������� ��2 �� ���� �( � �� ��� ���g�+�+� ���� �� � � �� ����� �$ � �� ��� ��� �� )�� ��� �� � ��2��� �� � � �� ����� �$ � �� ��� ��� + ����� �� � �� � � ��� 2� ��� 4 � ���� $�� ��� �� �(�������� �$ ����� � �����2� � !+ � ������� ��������� �$ ��� �� ��-��������� ���� �� �� �� � �� � ���� �� ���� �� � � ����������� ��� ��� �� �d�� ���� �� �� ��2��� �� ���������� �� ��� ��2� 10��� + �� �� � ������ ��� �� ( �� ���� ���$� ���� �$ ���� ���� 2�� � �� �� ������� ������ �����g � 10���� ���������+ ���� �� ��� ����+ ��� ���� � � )���� �� �������� � 2� � ��) ���� $����� ��� ����� + ��������) (� 2� �������) �� ��� $���� ����) ��2�� �$ ��� �� ��-��������� ���� �� �2� 2 (�+ 6����� 2� ��� � � ������ �� ��� ��)��-� �� ���� �$ �c+ /&0 � � � " � $� �2�@ �3� � �� � ����� ����"� ��� � �����2��"� ����" � ��� � ���� ��� � ������� �� � ��!� " � � ��� � �� � �� ����" � �� ����"�����" � �� ����"� �� /<0 2���� ����" �� ����( � �� � � �� � � �� ��� � �� �$ " �$ ��� ��� ������� �� �������� �� ��� �������� � (� �� ! �����2���+ ��� ����� � �� � �� � �� ��� � �� ��,������� � �� � �� �c+ /&0+ 4� ��� �c+ /<0 �� � �� � � ��+ ������� 2� �������) �� ���� � ��� ��)( �� � ��k���� ��� ����� (��) ��� �$ ��� ���� / 0+ �� � ��� �������� /��) ����0 ��� ������� �� � � ( $� �2�� �( �������� /��) ����0 ��� �������+ ��������) (� 2� ������ �� ��� � � �� /��� ���( e� 0 � ���� 99�:: �$ ��� ��� ������� �� �������� /��) ����0+ ����� 2� ����� ��� ������ �$ ������� � � � ���� 99� � "::+ � ��)� ���� �� ���$����� �� ����� �� 2������ ��� ������ �$ ������� � � �� )�� ��� �� � �!m+ � ���� ���� ����( ���� � � ( ������� � � � �� ��� $��� ��� ������� �) ������� ��$$������ ���2��� � ����) ������ $��� i �� �( "� "88! �� i��� "� !!" �� 1�.� ������( ����5 ������ / �8=! ������ �����0+ 4� �������) �� ���� ��� � � � ��� /���) n��)� �������� � i � �� n��� � 1 (�� � 3�� �������� ���) ����� � �2 �� �� � ��0 �� ��� *����� �� ���+ ����� �� �������� �� � *+�+ � ��� � � ��� 2��� ��� $$����� �( ��� ��� � ,� ��� ������# $�� �� � �� ���� 2� ��� ��� 1�.� *� ����5 � ������ � �� � �+ 4� ��� � � ( � � �� � ����� ������� ��������� ����� ������ $�� ����� �� ���� ��� ��� �� �� 2� ��2� �� ��� ���� ���� �� � ����) ����� ( � � � ( ��� �� ������� �� �� ���������) ��� �$$��� �$ � ��� �������� / � ��������� ( �����-���� ����������0+ � �� �� �� ��� ( �)��� �� � ����) ��)� $��c����( � � � � �� ���� �� ��� 2��� ��$��c���� �� ���)# �o +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( 2� � ���� �� �� �� �$$� 2� �� 2� ��) �� � � ��� ��� 2� �� 2 �� �� ���$��� ��� ����� ���� ����� �� �� ���-������� �� ��� ����.�/"�"0-1 ���� 2� ��� ������)� 2��� ��� $�2 � � ������+ 1��) � �� � �( . ��� ������ ���� ����5 ������ �� ��� ���� $��� ' � ���� �+ 4� ������ 1�.� ������� ��� ��� ���( � ����� ��� ������ �$ � -�� �� ����5 �( ��� ����) ��� �� � � ������� �$ �� ���+ 6����������� �� ������( ������� � �( �������� � ( �� �� ��� ���� � � � �� ��� ������� 1�.� ������� �� �$��� ��� ��� �� ��k��� ���������� ������� �)( ����� � � ���+ ��� �c���( ������� �� � �� ��� �� *+�+ �� ��# ������( ������) ��� � � �� ����� �$ �������( ��� �c� �� d���+ ���� �� ���� ������� �� �� ��)������ � � ��� ��� ��� ��k ���� ������ �� � �� ���� ������ �����)� 6����� �c� ���� /j��2� "88�0+ � ��� �� �������� ���$�����( �� ��� ���� ����� �$ ��� � � �� �� ������+ 4��� � �� ���) ��� ����� /������ ����� ��� -$��� � ��0� 2� ��� ��� "-����� �-�� � �� � ���5( $�� ��� ��� -$��� � ��+ 7�� ����� �$ ��� ����( �� �� ��� �� ��� ������ ������ $��� ��� ���� �$ ��� � �� �+ ��������) (� 2� � �� ��� ����� ��� � $�� � �� ������( /6�)+ &0+ �� �� �� �)��� �� �) �� p �) / !!!0� 2� ������$( �2� ��� ������� 2���� ��))��� ��� ����( ��)�� �� ��� �� ��2� ���� ����� �������+ 6���� (� 2� ��,�� ���-������ ������ �� ����) i �� �( "� "88! �� �����) �� i� ( "� "88; /"�8�= ������ �����0+ '����) ���� ������� �� ( i � � �5��������� ��� )� ����� /�!+"<m0# n��� /"+=�m0� �� � �� /;+ &m0� ���) ���� /""+;;m0� *+�+ /"&+&om0� 1 (�� /"&+�=m0� � �2 � /"<+""m0� 3�� ������� /"�+&!m0� ���) n��) /";+; m0� �� �������� /"8+�&m0 � ��� � � ��� �5�������� ��� )� �������� �������+ ������ (� 2� ��,�� ������ ������ �� ����) i� ( � "88;� 2��� ��� �� ��� �$ ��� �� �� 6� ������ �� 1 (�� � ���������� �� �����) ��������� &!� "88= /& � ������ �����0+ '����) ���� ������� i � � /�<"+ �m0� � �2 � /�<;+o=m0� ���) n��) /���+"om0� ���) ���� /�o!+�om0� 3�� ������� /�=o+<om0� n��� /�=;+"<m0� �� � �� /�"!!+<�m0� 1 (�� /�"<;+=;m0� �� �������� /� !&+�8m0 �5�������� �����+ ��� *+�+ /""+"om0 � � �� �5��������� ��� )� �������� �������+ ����� (� 2� ��,�� ����-������ ������ �� ����) 7������ "� "88= �� �����) i��� "� !!" �� � �����d�� �( �������( �� � �( �$ ��� � � ��� ������� /o8; ������ �����0+ ������� �� � �� /�+&<m0� *+�+ /�+8 m0� i � � /8+8�m0� ���) n��) /""+ "m0� �������� /"<+8 m0� ���) ���� /"�+� m0� n��� /&!+"om0� �� 1 (�� /&"+8&m0 � �� ��� )� �������� �������+ � �2 � /�<+;8m0 �� 3�� ������� /�&+8&m0 � � ��� �5����� ��� )� ����� �����) ��� � �� ������+ � � � " �������� �� ������� �$ � � ( � � �� ��� ����� $�� ��� ���� ������� ���-������� ������ �� ����-������ �������� ���������� (+ 1� �� �� �� �� ���� ����� � �2���� /��� �� ��� �$ � ���5������ �)� �������� �� ��) ���� �������� �� ��� ����� 0� ������� /��� � � ����� �$ ��) �������g�������� �� ��) ����0� �� ����� ���� �� ��� 2��� ��� 1�.� �. 4�� � �� ���������+ ��������� ���� �$ ������ � �� �c� ��� ������ � �� �� ��������+ 6�� ��� ���� ������� � �� ������(:� � � �� ��� ������� ������ �� �� � �����d�� �( �������� �� ��) ���� � �2����� �5���� �������� ���� ����� ���� /��)��,� �� l" �� l80� ������ �������� ���� /��)��,� �� l���" 0� �� �� �� ��( � �������) /��)��,� �� l��� " 0+ %��� �� � ��� �������� �$ ������ �������� ���� �� �� �� ��( � �������) ��))��� �� � � �� ��� �� ��������� + �� �� � ������ ���.� � � ������d ���� �� ������� �� �� ���� ��� +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( �; 6�)+ &+ j��� ������ � ��� $�� ��� ���� � � ��� �������� ��� *+�+ �� ��� 1�.� �. 4�� � /i �� �( "� "88! �� i��� "� !!"0+ � = + 2 3 ��� �� � � �245 �� � � ��� � � ���� � ' �� / $ 6( $ 7 ( � ! 8 ( � ( � � � " '���������� �� ������� �� � � ( � � �� ��� ����� ���) n��) �������� i � � n��� 1 (�� 3�� ������� ���) ���� � �2 � �� � �� *+�+ 4�� � 1� � 7��� !+!!!"; �!+!!!& �!+!!!& �!+!!! �!+!!!" �!+!!!" !+!!!! !+!!!< �!+!!!!< !+!!!& !+!!!" 3��-������ !+!!!� !+!!!o �!+!!! �!+!!! !+!!!& !+!!!< !+!!!& !+!!!& !+!!!!; !+!!!& !+!!!" .����� �!+!! < �!+!!=& �!+!!"= �!+!!&; �!+!!o" �!+!!&o �!+!! o �!+!! " �!+!!< !+!!! �!+!!! 3���-������ !+!!! !+!!!< !+!!! !+!!"! !+!!"" �!+!!!< !+!!!< �!+!!!< !+!!!!" !+!!!!" �!+!!!!& �+'+ 7��� !+!";8" !+! 8�=" !+!"<";= !+! &;� !+! !� < !+!";=88 !+!"&<;� !+! "o=" !+! "=� !+!!8;=" !+!!;;;; 3��-������ !+!"<="; !+!";�"< !+!"&o=8 !+!"�"" !+!" <�8 !+!"< 8= !+!!8==; !+! "8;� !+!"oo<o !+!!;= ! !+!!o�=8 .����� !+!&!!;o !+!o=" & !+!"=&88 !+!<88<; !+!�!=!= !+! 8 88 !+! <&! !+!"8 ;= !+!<"!�" !+!" <<8 !+!!8;" 3���-������ !+!";8<= !+!& o"8 !+!"��!< !+!&!!�o !+!"=8�! !+!"= =< !+!"o&<& !+! !=8= !+! &�=8 !+!" =&; !+!!8&8" � �2���� 7��� �"+!& !+ !+< !+<< �!+;= !+;� !+! !+!& !+< �!+<< �!+ � 3��-������ �!+;= "+&o !+ = !+& �!+&= �!+" �!+o! �!+!" �!+ �!+�8 �!+!= .����� !+< �!+;8 !+&8 !+�� �!+<& !+&" !+�8 �!+&& !+;= �!+= �!+o� 3���-������ !+"& !+== !+=" !+"o +!8 &+!8 �!+! !+" !+<= !+!" �!+!8 n������� 7��� 8+o= ";+== ;+8 ;+�< "o+=& "o+� " +�< o+!< =+;� =+<8 o+�8 3��-������ "!+ & "=+�8 ;+;< �+8; "o+= " + � ""+" �+=< =+8= 8+ < ;+;< .����� =+!� 8+o= <+�< =+!= "<+= <+8o ;+88 ;+"& �+<; 8+�; �+ 3���-������ �+<" =+= =+<8 <+"; "o+8 "=+ � <+�; <+=; �+o= <+<& <+"< + 2 3 ��� �� � � �245 �� � � ��� � � ���� � ' �� / $ 6( $ 7 ( � ! 8 ( � ( � 8 � � � " /"������ � 0 ���) n��) �������� i � � n��� 1 (�� 3�� ������� ���) ���� � �2 � �� � �� *+�+ 4�� � .���� ���� 7��� !+&" !+!; !+o" !+"� !+ !+" !+&< !+"& !+"= !+;" " 3��-������ !+ = !+! !+;; !+"" !+&! !+!= !+<" !+"& !+"; !+�; " .����� !+<� !+"= !+<� !+"� !+ < !+ o !+& !+"; !+ & !+=� " 3���-������ !+ 8 !+!= !+& !+ o !+!o !+" !+& !+"! !+ ! !+== " 3�� ���� �( ����� /���� ������ �� (0 e�� !+;< !+& "+"o !+�! !+�< !+&! !+o !+&< !+�� !+8" " l" !+! !+"= !+!; !+!8 !+"" !+ � !+"= !+!< !+"= !+! !+!< l8 �!+!& !+!< !+!; !+"! !+!� !+!= �!+!< !+!& !+!= �!+!" !+!� l���" !+!<" �!+!" !+! 8 �!+!&8 !+!8! !+!; !+!&< �!+!&8 � !+!"� !+!"< !+!"= l��� " !+!o" !+!8� !+! " !+!=; !+!=" !+!o= !+!o� !+!"! !+!88 !+! ; !+! = .e o+=8 ;+�" �+=& <+8= o+"; ;+" ;+=< <+;" <+8; !+=< "+!8 � ������ ����� �� � ��� �� *+�+ �� � 1�.� ������+ 4� �� ��������� ie ����� /��� ��������0 �� $���� ��� �� � $�� � �� ������� ��� �� �(�������� �$ ���� ��( 2 � ��h�����+ l" �� l8 �� jh��)-e�5 l-�� ������� $�� ���� ����� ���� /" �� 8 )�0+ 6�� � �� ������(� 2� ��)���� � �� � � �� ������ ������ ) ���� ��� 2�� � � � �� ��� ������� /1�.� �. 4�� � ������g�-�� � ��0g�+�+� '��� � �� ��3 �� � ����+ l���" �� l��� " ����� �� ������ � /����0 �� ���$�����+ 4� �� ������ ��� 6-�� ������ $�� .��2 e�� ���� /.e0 $�� ��� ������ �� ����) �� i� ( � "88; /.������( �� ��0 �� �����) �� ��������� &!� "88= /��)�����) �$ �������( �� ��� � � ��� � � ���0+ ��� ������ � � �� � ������ ��h������ �$ ��� �� �(�������� � "� �� "!m ��)��,� �� ��� �� ���������� (+ o ! + 2 3 ��� �� � � �245 �� � � ��� � � ���� � ' �� / $ 6( $ 7 ( � ! 8 ( � ( ��� ���� �$ � � ( ��� ����� + 6����������� ����� ��� ��� � $��� ��� ��)������� �$ �������� � � �� ������� 2��� ��� 2�� � ��� ����� �� ��)��,� �� $�� ���������� ��))�����) ���� ��� �$ ����� ���� 2��� ��� 2�� � ����$� ��+ 6�� (� .��2 ��� ����� �� �� ��( ���� $������ ����� ��� �� � ��� �� �������� ���2��� � � �� ��� ������� �� $���� ���� �� �� ��( ��)��,� �� ( ��$$��� �� ����� ������� �$ ����(+ e�$��� ��� ������� �� ( i � � �� n��� � �� ��) ���� ��� ����� + ��� *+�+ � � �� �� ��� �� �� � �� � ��� � � � ���� �� � �2 � �� ��� ���� �� �� � � � ��+ ��� ����� �� ��) ���� ( � �2�� /�5���� �� �������� � i � � �� n��� 0 �� � �� �)� �5���� �������+ '����) ��� ������� ��� � � � ��� ���2 ��) ���� ��� ����� 2��� ��)��� �� ����� �� � �� ����� �������� �� ��� ���-������ ������+ � ��� �� �� ��( �� ���� )�� ��� �� � �� ��� ���-������ ������+ ��� ����� ���� 2��� ��� 2�� � � � �� ����� ��� ( ����� ��� /����� ���0 $�� ���) n��)� �������� � n��� � 3�� �������� � �2 �� �� � �� �� ��� *+�+ /i � � �� 1 (�� 0 � ���� ��� �� ��� ���-������ ������+ �$��� ��� ������� ��� )� � � �� ����� �� �������� /�5���� �� ��� 3�� ������� �� � �2 �0# �� �� �� ���� ����� �� �� �� �� � �����) ��� ������ ������� (�� )�� ��� �� � ��$��� ��� ������+ .���� ���� 2��� ��� 2�� � � � �� ����� ��� ( ����� ��� � ���� ��� �� �����)-���-������-������ /�5���� $�� *+�+ �� n��� 0+ p��� � ���� ��� �� ��� ���-������ ������� ����� ���� 2��� ��� 2�� � � � �� ����� ��� /����� ���0 �� �������� � n��� �� *+�+ /i � � �� 1 (�� 0+ 3�� �� ����-������ ����� ����� �� ���� � �� ���) n��)� 3�� �������� ���) ����� � �2 � �� �� � ��+ 6��� ����� ������ ������ �� �� ������� �� � ����� ���� 2��� ��� 2�� � �� �)�� ���� ����+ ������� 2�� � i � � �� 1 (�� ��� � �� � �� ������ ���� ��)������ � ���� � �$ ��� ������� �������� �� n��� � �� ������ ���� ����)� ��� 2��� ��� 4�� � � � ��+ ����������) (� ��� ��� �$ ����)� ���� �� ��� � h����( �$ ��� � � ��� � � ��� � � ��� ���� ���� �)�� � ���� � �$ ��� ��� � ,� ��� ������+ %�5�� 2� ��� ��� ���� �� $��� � � � " �� )�� � ����)�� �� 2���� �� ���� �$ ��� � ����� ����+ ��������) (� 2� �������) �� ��2 ��� �����-�������� �$ ��� ����� �� �5� ���� �( $��� �� ����� �$ 99���� ����:: ������ ( �� ����� �� �� ���������� �� � � ���: ������ )���� ���) �������g�+�+� �� �� �� ���� ����� ��� � � �2���� �� �������+ 4� �� ( ������ ��� ,����)� �� ��� ���� ������ $�� � � �$ ������(# ��� ,����)� �� ���� � $�� ��� ����� ����������+ 6�)+ < ���� ��d�� ����� �����-������� �� ���������+ �� �� 6 � �� 6����� /"88 0� 2� ,�� k � �����-������� �� �������� ���2��� ��� ����� �� ��� + ��� �� �������� ���2��� ������� �� ��� ����� �� �� k �+ ����� ���� �� �� ��dd ��) ��� ��� ����� �� ��������� ���� � �� ��2 ��-� ����) � ��������� � ��� ( ��c���� ��)��� ������ $�� )�� ��� � � ����� �$ ��) �������� �� ��) ���� �������+ ����������) (� �����-�������� �$ ��� ����� �� /"0 �� �� �� ���� ���� �� / 0 � �2���� �� �� ������ ( ��)��,� �� � ��� �m ��� � ��� ��))��� ��) ���� �� ��������+ ���� �� ���� �� ��� ��� ��� �� �� ������������ 2��� ��� ��� ������� ������� � �� ��� ��� �� � ��������� ���$�� �������� ( � �2�� ������������� �� ��� ��) ���� ( � �2�� ����+ �� ���� ������ 2� ��h��� ��� � �� �� �� �� ���� ����� � �2���� �� ������� � �5� � ���( � �� � �� �$ �������: �����-�������� �� � � ���+ ������� ����� �� ��������� � ���� �� �5� ���� �( ��� �����( �� ��)�� ���� $��� �� ������ �$ �����-������� ������ �����+ ����� ,����)� �� ���������� 2��� ��� ��� ������� �� �������� ��dd � �������� �� ��� � ����� ���� ���� /�+)+� ����))�� "88=0 �� � $�� ��-�5 ��� ���� �$ ��� .�31+ +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( o" � ��������� �� ���� 4� ���� ��d� ��� ���� �� �$ ��� �� �������� ���2��� �� �d�� ������ �� ��������� �� �� ��( �� � � ��� $��� i �� �( "88! �� i��� !!" �� � � � + 6�� ���� /"0� � � �� ������ �$ � �� ��� ��� /�0 �� ����� ��)��,� �� /2��� ��� �� � �5������� �$ i � �0+ �� 6�)+ <+ .����-�������� �� ������ �$ � �� � � �� ��� )� ��� ������� 2��� �������� �� �� �� ���� ����� ��� � � �2���� �� ������� /i �� �( "� "88! �� i��� "� !!"0+ o +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( � � � ���� �$ �� ����j�����:� .�31 ����) ���� � /"0 �� / 0 .�����( 1��� /"0@ ����.�/"�"0-1 1��� / 0@ �� ��-��������� ����.�/"�"0-1 1� � �c� ���� q �� ��� �c� ���� ' '4 1� � �c� ���� q �� ��� �c� ���� ' '4 � � � � � � ��� ���2� � � � ���) n��) !+!!!o� �"+"<& !+!!!!!; !+!<" !+"!; !+==! �!+!!" "+8" !+!!!!" &8+!&8 �&<+o& !+!!!!" !+!8� !+"8o !+; � !+� "+;& �0��������� "+;" � �!+o= +8� +;o &+; �!+ < !+! =+&8 �;+o +&= � +8� &+88 &<+;< �������� �!+!!!&o; !+"88 !+!!!!!< !+ o< !+!&o !+;;= !+!!& "+;" !+!!!�< =+8<o �"<+�� !+!!!!" !+<"& !+! !+o=< !+& "+;� �0��������� �"+o8 � !+&= &+�= &+ 8 !+<< 8+&& "+=! � �+== �"�+;< !+=o <+"� !+"o &<+"< i � � �!+!!"&&o <+o;8 !+!!!!!< !+!<� !+!=o !+=88 !+!!" "+=; �!+!!!<o <=+&; �<&+<<< !+!!!!" !+!;; !+!== !+=<8 !+<" "+=; �0��������� �&+"8 +"! � <+8" &+o< &+o< o8+"8 �"+�8 �+o! � &+8< +!� � <+<; +;o <o+;" n��� �!+!!!=<< !+;�" !+!!!!!< !+!o !+!� !+8!8 !+!!; "+=& �!+!!!"o &&+&;< � =+<=8 !+!!!!" !+!;8 !+!<< !+== !+< "+;� �0��������� � + < � !+=& +8� <+!; +&& � ;8+<o �!+�! ;+&; � <+8< +<� � �+!� "+;� � ="+�& 1 (�� !+!!!!== �!+!8& !+!!!!! !+!�< !+!= !+8!o �!+!!" "+;8 �!+!!!o" &;+ <o � 8+ �o !+!!!!" !+"" !+!;< !+=" !+&" +!< �0��������� !+&; �!+"! +<< � <+&; &+o ;�+o" �"+;o � "�+& �" +8" "+o8 � &+&= "+= � �+o< 3�� ������ �!+!!!" " �!+�8� !+!!!!!< !+!o& !+!== !+=8 !+!!< "+�8 �!+!!!!< &=+" �&=+<<& !+!!!!" !+ <" !+!!& !+;"< !+"8 "+== �0��������� �!+&o �!+<; &+ + = � +" � <<+"= �!+"� <+ � "+8& &+o< �+&; "+;< � "=+; ���) ���� �!+!!!";" "+"�; !+!!!!!< !+"!! !+"! !+=&o �!+!!" "+o8 !+!!!�" <�+o; �� +;�� !+!!!!" !+"&; !+"�� !+;&� !+<� +!" �0��������� �!+;; !+o� <+&< <+<= +;8 &=+�8 +< � +;= � "+8� "+ o &+;� +;" �+=o � �2 � �!+!!!��; !+o88 !+!!!!!8 !+!<< !+!o! !+8!& !+!!" "+8 �!+!!!<& &�+<=& � 8+&o< !+!!!!" !+!<; !+!o !+==o !+<8 "+88 �0��������� �"+!8 !+<8 &+<" &+&8 +=& ��+o� �!+88 =+"= � +<& +! � &+=! +; �"+=" �� � �� �!+!!! 8 �!+�;� !+!!!!!; !+!=o !+!o !+=;< �!+!!� "+o= !+!!! o =+"!8 � =+�;" !+!!!!" !+"&" !+!&" !+= & !+ 8 "+;; �0��������� �!+o" �!+�� &+=< <+<; +"" � �!+!& !+=; +=o � �+!8 &+<" <+;& !+=� < +=8 *+�+ �!+!!!!& &+�!" !+!!!!!" !+!" !+!8& !+8 = !+!!8 "+8; �!+!!!"� ;!+�= �o"+ 8o !+!!!!" !+! = !+" !+=<o !+<& "+=o �0��������� �!+"; "+ 8 &+&; "+" <+=& 8 +;� �!+;� �+<o � +<& "+�= "+�" &+88 &o+"o ����� ��� ������ � �� ��)� ( ���� ������ 2� ��� ������� �� ������( ���������� ��� �� ��� �� ������� ��� c� ��-� 5���� � � ����� /l1j� 2��� 1 �c� ��� ������0 ��� �� ���� �� �� �� �� ������ � ��������� �( e� ��� �� �� 4�� ����)� /"88 0+ ' �� ��� �h����� '-�c� ���+ 99'4:: �� ��� '����� 4 ���� �� ������+ ��� ������ � � �� � ������ ��h������ �$ ��� �� �(�������� � "� �� �� "!m ��� �� ���������� (+ + 2 3 ��� �� � � �245 �� � � ��� � � ���� � ' �� / $ 6( $ 7 ( � ! 8 ( � ( o & ���� �� 2� �� ��� ,�� �( ��)��,� �� �� �������� ���2��� � � �� ��� ����� �� ��������� �� �� ��(+ ���� �� $��� ��� � �� ��� �c� ���� ������� ��$��� ���� ���� � �� ��� ����������� �� ���� )� �$$���+ q �� ��� �� ���������� �� ���� � � ��� /�� � "0+ ��2����� 2� ,�� ��� � �� ��� ������� �� �� �5� ����� �� �������� + ��� ���� )� �$$��� /��)��,� �� �������� �(�����( �� ����� 2��� �0 �� ������� �� � � ���� ��� �������� + �� ����� ��� �� � � � ��2� � � )�� ��� ��� �� �� ��������� � �� ��� �� � )��� ��2�+ �� ���� �(� ���� /"0 ���� ���� h�� ���� ��) ��� �� �������� ���2��� ��� ������� �� ����-� �(��) �� �� ��(+ � '-�c� ��� � ��� �� �5����� ( �2 /r!+!"0 �� � � �� ����� �$ � �� ��� ��� /�0 �� ����� ��)��,� ��+ ����� ,����)� �� ���� � �� ����� �$ � �( ������� ���� � �� 6����� �� + /"8=;0 �� e � �� �� '� ���� �� /"88!0+ �� ��� 10��� $���� ��� .�31 ��))���� �� � � � �� ����� �$ ��� ���� � �� ��������+ p��� �� ( ���� ��� $���� ���� �$ �� ����j������:� .�31 � � �� ������+ ����� ��� �� �d�� ��� ������� � � �� �������� �� ��) ����� �� ������ �� �������� �� ��) ���� � � �� ����� �$ ��� � 2���� � � �� �5� ���� � $� �2�+ '����) ������� �$ )��2��� �� �� ��( ���� ��� ��� ��� ��������� 2 ( $��� ��� � � �� ��� ��� ��� ���� ����5 )���� ( ������ ��2 ���+ ��2����� ���� )��2�� $ ����� �� �� ��( ������ ��������� 2 ( $��� �c���( � � ��� �� ����)� ���� ������ �� �������� ���2��� �� �� ��( �� �������+ ����� ��� �$$��� �$ �� �� ��( �� �������� ��� ���� �� �����c���� ��� ������� �� �������� �� �� ��-��������� �� ���� /"0 ���� �� ����,�� �� ��k��� ���� �� ��(+ ���� �� $��� ���� / 0 �� �� ���� ��d�� �� � � � + 6�� � � ���� 2� ,�� ��)��,� �� �������� � � �� ������ �$ � �� ��� ��� �� ���� �� �� ��)��,� �� � � �� ������ �$ � �� ��� ��� �� ��2��� ��+ ���� ���� � ����� ��� �� � � ����� �� �� � � �� ��������� � �� ��� � �� �� � ����� �� /$������ ����� ��0 �� � � �� ��� ������� �� ���� �� /��2��� ��0+ 6����������� ���� / 0 �������� ������ ,� �� � ���� /"0 � '-�c� ��� � ��� � �)� $��� !+"8 �� !+� + ���� �� $��� ��� � �� ��� �c� ���� �� ���� � �� ����� �$ ���� /"0+ ����� ,����)� ������� �� � �� ��� � �5� � ���( � �� � � �$ ��� ������ )���� ���) ������� �� � � ��� �������+ ��� ���� �� $��� � � � �� � ��2�� ��� ������ � � c�������@ ���� ��� � � ��-� ��� .�31 �� �f ��� � � ��-� ��� .�31 ��))��� ������ �� �������� ���2��� ��c����� ������� ���� $�� ��� �� �� � �� ���+ �� �� � � ��� ��� � � �� ����� �$ /�� ��2 �� ��0 � �� ��� ��� �� �����(��� ��� �� � �� � � �� �� �� �� �5������ �� �� ���� ( ����� ��+ � ��� ���� �� ��������� 2���� ��������� ( �� �� 10��� � � � �� ( �� ������� ( ������ 10����+ ��� ��� �$ ��� �� ��-��������� ���� �� �� �� �������� �� � ���� � � �( �2��) $�� ��� ��) ���� ������� �$ ��� � � �� ��� ������� ������������+ e�$��� ��� $ ��� ��������� � �� ���$��� � � �� �����) �� ��( �� ����� � ���� �5���� ����� �� 2� 2 (� ������ ���2��� ��� � � �� ������ �� ��� ��� -$��� � ��� 2�������� �� )�� ���+ �$��� ��� $ ��� ��������� �� ��� � �� ���$��� � � �� �� ��( �� � ( �� �� $���� �� � � �� �� 2���� �� �d�� ������ �� �� �� �� � ��� ��� -$��� � ��+ �$��� ���� ( �h�����) $�� ��� �2� �� ��� �$ ��� 2�� �� 2� ,�� �� � �������� /��) ����0 � � �� ������ �$ � �� ��� ��� �� ����� ��� 2��� �������� /��) ����0 ��� ����� + �����$���� 2� ���� ��� �� � $���� ���� � �� ��� �5� ��� ��������� ����� �������+ �� ����� �� ��$�� �� 2������ ��� � � ��-� ��� .�31 �� �� ����� ���������� 2� �� � ��� $� �2��) ������������ �� ��������� $��� i �� �( "88! �� i��� !!"+ ���� )� �������� �� ��) ���� ��� ������� �� �������� �� � �� � � �� �� � ����� ) ���� ��� �������� 99 ��� )�:: �������� /���0 �� ��) ���� /���2�0 ����� �$ ��� ��� ���� $��� ���� / 0+ ���� �� �� )� ���� �� 6�)+ �+ o< +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( 6�)+ �+ .����-�������� �� ������ �� ������� �$ � �� � � �� ��� )� �������� /"0 �� ��) ���� / 0 ��� ������� 2��� �������� ��� ��� �������� /���0 �� ��) ���� /���2�0 � � �� ����� �$ � �� ��� ��� /i �� �( "� "88! �� i��� "� !!"0+ + 2 3 ��� �� � � �245 �� � � ��� � � ���� � ' �� / $ 6( $ 7 ( � ! 8 ( � ( o � ���� �� $��� ����� �����-�������� �� �������)+ � �� ������( � � ��� �2� ��� )� ��2 �� �� � �� ��� ��� + �������� (� ��� �� �������� ���2��� ��� )� �������� /��) ����0 ������� �� �������� /��) ����0 � � �� ����� �$ � �� ��� ��� �� ��)��,� ��� ������� �� ��� � /��)��,� �� 5-�� ������� � ��� "m ��� 0+ ��� � ���� �� ��� �2� �� ��� �� �� ��)��,� �� �� ���� �+ �� ���� �� ��� )�� ��� ��� ��2 �� �� � �� ��� ��� � ��� �� �� ��� ��������� ����� ��c����� � �� �$ ������+ ���� ,����) $ � � �� ��� �� �� ��� � � ��-� ��� .�31 �� ����$����� ��� �� ��$ �� � ��� ��� ������� )���� ���) ������� �� ��������� ����� ( �� �����������+ %��� �� � ����� ��� � � � � �� � �� $��� ����� ������ �����+ 6����� ��� �2 ������ �$ ����� ������� ������ ����� ��)�� � �� ����� ��� �� ������ ,����)�+ ������� �� ��)�� �� �)��� �� � ����� �� � �� � �� �� ���������� �� ��� ��� ������� )���� ���) �������+ ������ � �( ���� ������ � �� ��))����� �� � ��2 �� �� ��� �� ����-� �(��)� 2���� �� ��)�� ��� � ����( �� ��� �� ���) �$ � 99 ��� )� ��2 �� �� �� � �� ���::+ �� $ ��� 2� ���$����� $ ����� � (��� �� $ � �� ����� ��� �� �(�������� �$ �c� 99� � �� ����� �$ � �� ��� ��� :: �� � �� ������ /���� �� �� � � � � ���� ��c����0+ ���� ,����) ��k���� 2� ��2� ���� ���� �$ ��� �� ����j������ .�31g�+�+� �� ������� ����� �� � � �� ����� �$ ��� + ������� ��� ������� �� �)�� 2��� ��� �� �� �$ ��� ������( �� ���� /��� '� � ���� �� ��� ��� "88;0+ �� ���� ������ 2� $� �2 �� �� �� + / !!"0 �� �������) �� �(������� ��������) ��� ����� ��$���� �����) �� $��� ��� ��� � 6�� ��� .�����+ 4� ��� 4 � ����� �� �� ���� ��� �(�����( ���2��� ���� �� �� ��2��� �� � � �� ����� �$ � �� ��� ��� + 4� �� ��� ,�� �������� �$ �(�����( ��$��� ��� ������+ p��� ���� �� � � �� ����� �$ � �� ��� ��� /���0 �� ��)��,� �� ( )�� ��� �� � ��) ���� � � �� ����� �$ � �� ��� ��� /���2�0 �� ��� � � � ��� �����) ��� ������ �� �� ��� *+�+ $��� ��� ������� ��))�����) �5������� ��������� ��2 �� ���� �( � �� ��� ���+ ��� � �� ����� ���2 �� � ���� �� � � �� ����� �$ � �� ��� ��� /���0 �� ��)��,� �� ( �� �� �� � ��) ���� � � �� ����� �$ � �� ��� ��� /���2�0 �� ��� *+�+ �����) ��� ������ �� �� i � �� n��� � 1 (�� � 3�� �������� �� ���) ���� $��� ��� ������� ��))�����) �5������� �������� ��2 �� ���� �( � �� ��� ���+ .��������� 2��� ��� �(�����( �������� ( ��������� ���������� �� �� ��)��,� �� �� ���� �� � ��� � �� � � ��� �5�������� ��)��,� �� ����� ����� �� ) ��� /�0+ ���� �� �� ���� ��d�� �� � � � &+ ��� �������� �5������� ��������� �� ���� ��� � � � ��� �����) ��� ������ �� ���������� 2��� ��� �� ��$ �� � ������-�� ��� �� �� ��( ���� � ���))�� � ���+ ����������) (� 2� ,�� �� � ������ �$ �5������� �������� /���������0 �� $� �2�� �( ������ �$ �5������� ��������� /��������0 �� ��� *+�+ /i � �� n��� � 1 (�� � 3�� �������� �� ���) ����0+ ���� ������ ���� �������� � ��� $�� ����� �� � ����$� �� ��� ��)���+ ����� ��� ��)��� �� � ��� �� ��� ������� �� � �$ � � �� ���������� ( ��������$���� /������$����0 ����� � � ���� �� 2� ������$��� /��������$���0 ������� ������$�����) /��������$�����)0 � � ��� ���� �����c���� ������� /��� '� e���� �� �� ��� "8=�# 1�� �� +� !!!# �� �� �� +� !!" $�� ��� � �� ����������� �$ 99�(���� ��� ������ �$ $������:: �� �����c���� h����,� ����� $�� ��� ��� �$ .���� �� � ����$� �� ��� ��)���0+ �� �� 2� ��2� �� � ������ ����� �������� �����) ��� ��� � ,� ��� ������ �� ��� $��� �$ � ��� �� ������)# �� �� � ������ ��� ,����)� �� ��� )�������� ��)+ ��2����� ���� / 0 �� � � �� ������� ��� �� ���� �� c���� ����������)+ �� 10���� �� ��-��������� ���� �� ��� $���� ����) ���� + � ����� �� �������� ��������� ����� �� �������� ���2��� ��� �� ������� 2���� � � �� ���� �� ������ oo +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( � � � & ���� �$ ���� / 0 ��$���� �����) �� $��� ��� ��� � ,� ��� ������ .�����( 3��-������ /i �� �( "� "88! �� i� ( "� "88;0 .����� /i� ( � "88; �� ��������� &!� "88=0 3���-������ /7������ "� "88= �� i��� "� !!"0 � ��� ���2� 4� ' '4 � ��� ���2� 4� ' '4 � ��� ���2� 4� ' '4 ���) n��) !+!!!o �&+" "; ��"+o =8 !+== !+&;; "+;� �!+!!&< <+ ;&& �";+8!; &+&< � !+<<8 "+o8 �<��!< &;+<&< �&�+�"& !+" !+&8= "+oo �0��������� !+8 "!+<8 �=+88 � +8< "!+!; �;+�8 �!+ o+8! ��+ & �������� �o��!< �!+ !8= �<o+;8<" +!< !+<�8 "+; �!+!!o" "o+"� �"!+=� &+8; � !+<"� "+o& ! =+o< � �+" !+=� !+<; "+;8 �0��������� �!+=; 8+8< �;+=" � +�; � " +&! �<+8= �!+!& ;+88 �;+8 i � � !+!!! o"+&=& ��8+!&o !+"� !+�"8 "+;8 �!+!!"" o<+o&= �<�+�8; &+�o � !+<&" "+;8 !+!! <&+" �<=+<8< &+"; � !+<�� "+8o �0��������� !+&� "!+ � �"!+ o � +"8 � ;+;o ��+"; "+=& � <+�o ��+8; n��� � ��!< <&+=o�o �< +=8;o !+o� !+� � "+=" �!+!!&; "+�"& �"�+ <= <+"o � !+�&= "+== !+!! & <+ o= � 8+!� &+&< � !+�" "+8< �0��������� �!+ 8 8+== �8+ & �&+!; ""+8! �;+=8 "+o8 � �+"< ��+�8 1 (�� �;��!< o<+<<=o ��8+8<; "+88 !+<& "+;& �!+!!<= �+8;; � "+ !" &+&o � !+&& "+= !+!!"o &�+<8& �&8+8== &+ & � !+<�< "+= �0��������� �"+<; ""+o< �8+o< � +<! � "!+"8 ��+�; "+;! � ;+o; �;+ = 3�� ������ !+!!! �o+!8<" ��o+8!&= !+!= !+<;& "+8< �!+!!<< & +o;8 � !+;; <+ & � !+< � "+=o !+!!!; &=+"o= �<�+;8 &+!8 � !+<<= "+;= �0��������� !+&8 ""+<= �"!+< �&+"o ;+o! �o+!= "+;o � o+=; �=+�" ���) ���� �;��!< ;=+;�8& �;o+<"<o +" !+&=" "+<& �!+!! &"+&!<� � =+!�< <+!< � !+ &" +"= !+!! ; =+;&8 �&<+;8 &+8" � !+<&= "+o= �0��������� �"+<� "!+8o �8+<! � +!; � ;+ � �o+= + � � �+;o �8+!8 � �2 � !+!!" &8+!8& �<"+<oo !+"< !+<== "+8; �!+!! " <�+"=;; �& + ;� &+ = � !+�<& +!; !+!! o &o+o < �&;+ 8 !+<� !+�!< "+o& �0��������� !+;o o+�o �;+ < � +!o � o+"o �<+!; "+< �+&< �o+;� �� � �� �!+!!" <<+�&" �&8+;<o= "+! !+<=& "+=" �!+!"! &!+���< �"!+&�" "&+=" !+<;= "+;= �8��!< &"+=;= � 8+o<8 !+"< !+<" "+;& �0��������� �"+& "!+o! �8+"= �&+=! 8+�! �&+8" �!+< o+&< �o+�; *+�+ !+!!!" =!+ �o �="+8=; !+" !+<&" "+=� !+!!<< �=+&o �;;+ �o <+ � � !+<�" "+=� �!+!!" oo+& � �� +8� �+8" !+�!; "+=8 �0��������� !+ � ""+�o " +�" + o � "!+� �" +8� � +&" � "!+ " �8+�" 6�� � � �$ ������(� 2� �� ( ������ ��� ,����)� $�� ��� �� � �c� ����+ ����� ��� ������ � �� ��)� ( ���� ������ 2� ��� ������� �� ������( ���������� ��� �� ��� �� ������� ��� c� ��-� 5���� � � ����� /l1j� 2��� 1 �c� ��� ������0 ��� �� ���� �� �� �� �� ������ � ��������� �( e� ��� �� �� 4�� ����)� /"88 0+ � �� ��� 4 � ����� /4�0 ������� �$ ������) ��� �� �(�������� �$ ��@ ��� � ���2� � !+ ' �� ��� �h����� '-�c� ���+ 99'4:: �� ��� '����� 4 ���� �� ������+ ��� ������ � � �� � ������ ��h������ �$ ��� �� �(�������� � "� �� �� "!m ��� �� ���������� (+ + 2 3 ��� �� � � �245 �� � � ��� � � ���� � ' �� / $ 6( $ 7 ( � ! 8 ( � ( o ; ������ ������ � ����������) ��������� �� ��� ����� �� � ��������+ ��������) (� 2� �������) �� ��5� ��� $���� ����) ��2�� �$ ��� �� ��-��������� ���� + 6����� 2� ��� � � ������ �� ��� ��)��-� �� ���� �$ ��� �� � �c� ���� �$ ���� / 0 � ���� 99� � "::� �� ��� �c+ /<0 �� � � ��� $��� i �� �( "88! �� i��� !!"+ ���� �� �� ���� ��d�� �� � � � <+ 6���� ������ �� � ��� ���� �� $��� ��� � �� ��� �c� ���� �� ���������� 2��� ����� �$ � � � �� �� � '-�c� ��� � ��� �� �5����� ( �2+ ��2����� ��� �� ���2� �� ��)��,� �� �� � � ��� � � �� � ��� "!m ��� /�5���� $�� � �2 � �� ���� �� �� ��2��� �� �� *+�+ �� ��2��� ��0+ ���� ,����) ����� ��� �� � ��) ���� /�� ��������0 ��� ������� �� � � ( $� �2�� �( ��) ���� /�� ��������0 ��� �������+ ��������) (� 2� $������ �������) �� 2������ ��� ���� � � ����� �� ��� ������ �$ ������� � � �� � ����� � ���� �� ���� ��� ��)(g�+�+� ��� �� ��( �� ������� ��� ��������� �$ ����� ���������+ ������� ����� �� � � �������� �� ��� � ���� ( ���2��� ��� ��� $��� � �� �� ��� � � ��# ��>��� ����� ���� �� � �� ����� ����� ���� $���� ��� �� ����� �� ����2��)�� ������ ��� ���������� �� ��� ���( e� � �� �� �� �c���( � � ��+ ���� �� �� ������ ( ���� ����) �� ��� � ��)������� 2��� � � � < ���� �$ � 10��� ������� �$ ���� / 0 .�����( 1� � �c� ���� q �� ��� �c� ���� ' � ��� ���2� � � � ���) n��) !+!!!<8 &+!<= �&+<8; =+8&��!o !+!�" !+"& !+=�& !+!!; �0��������� "+�" +"o � � + < � &+;< &+<< &+8� <;+8< �������� !+!!! 8 "+<<� �"+;�< ;+ ��!o !+& 8 !+!=& !+; �!+!!& �0��������� "+!< +<8 � � +&o � &+== &+=! !+=& &+8< i � � �!+!!!8 "+8 �&+o; &+�o��!o !+!< !+!8" !+=8= !+!!� �0��������� � +o& "+o= � �&+" �+!� &+o" <+!� � ;�+&! n��� �!+!!!� "+==< �"+&;� &+=o��!o !+!o !+!<8 !+8" �!+!!" �0��������� �"+� "+o8 � �"+;" � &+ " <+�& +&8 � = +=" 1 (�� �!+!!!"= o+�=" �<+& o �+""��!o !+!<; !+!; !+8!" �!+!<� �0��������� �!+<= &+ � � += �+o! +<� � +"" � <&+&= 3�� ������� !+!!! "" �+8o< �;+� 8 �+ ;��!o !+!;" !+!o8 !+==� !+!"8 �0��������� !+o &+�8 �<+=o o+;= +<� � "+o= � �&+8< ���) ���� �"+8=��!� =+&o& �o+ �� o+<=��!o !+!8� !+" !+="o �!+!!& �0��������� �!+!= &+oo � +�o � "+8� � <+&! +"< � "o+o" � �2 � �!+!!!�= +&= �"+<" "+"o��!� !+!<; !+!;& !+=8 !+!!" �0��������� �"+"" "+<& �"+& &+8; &+;8 &+o; � ��+8< �� � �� "+8 ��!� &+�8" �<+&= ;+="��!o !+"!& !+!o� !+=�< !+!!8 �0��������� !+!o +=< �&+&; �+"� <+o! +!= � <;+ 8 *+�+ ��+ "��!� �+o<o "+o"" "+o;��!o !+!!= !+!8 !+8 � ! �0��������� �!+ � "+8; � !+�� +<o � !+;; �+<! "!o+"= ����� ��� ������ � �� ��)� ( ���� ������ 2� ��� ������� �� ������( ���������� ��� �� ��� �� ������� ��� c� ��-� 5���� � � ����� /l1j� 2��� 1 �c� ��� ������0 ��� �� ���� �� �� �� �� ������ � ��������� �( e� ��� �� �� 4�� ����)� /"88 0+ ' �� ��� �h����� '-�c� ���+ ��� ������ � � �� � ������ ��h������ �$ ��� �� �(�������� � "� �� �� "!m ��� �� ���������� (+ o= +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( $��2 ���� ��) � �� � �� �� � �� � � � � �$ � ������) �5���� ���� �� $����� ����� ���������� ���� � �������� � ��� ���� ��� �������� (�� �� &>+ � ���� ���+ �� $ ��� ��� ��2 �� $��� $���� �� ����� �� ��� $��� �$ � ��) ������� � � ��) ����) $����� ������g�+�+� �� ��) ��� $����� �����+ �� ���� �� ��� �� ��( �� � � ������� � � ���� �� 2��� ��� � � � � ������ � ���� �� ���� ��� ��)( 2��� ���� / 0 .�����( �� �� 7��� ������ 3��-������ .����� 3���-������ .������ � � m .������ � � m .������ � � m .������ � � m ���) n��) � "��o8 � � +&! "�! � � + ";< �&+&; &;& � +! '�2��� �� = & �&+�" � o � � +8; 8= � �o+& "88 �&+o< *��� �� ;<o �"+!& <8o �"+<� ;o �!+!! ";< �!+ 8 �������� � "�;"& �;+"! "�"&< � �;+8� "8& �8+ ! &=o � �&+=< '�2��� �� 8o" �8+8" o"; o!+! " � o�+"! "8 �;+!& *��� �� ;� �&+=; �"; ��+o� o= �!+;� "o; �!+"� i � � � "��&8 �"+&! "�!!� �"+&� "; � +;o &o �!+<8 '�2��� �� =o� �<+ & �o! � �<+!� "" � �8+ o "8& � +"o *��� �� o;< <;+8; <<� <=+& o! <&+=! "o8 <=+;! n��� � "�o<� �<+=& "�!o< �<+&; "=8 �;+8= &8 � �<+o; '�2��� �� 8;8 �8+" o<; �8+"8 " ; o<+=! !� � ��+=o *��� �� ooo <8+�� <"; <=+ o o <;+o8 "=; �&+<& 1 (�� � "�;"" �;+!& "�""! �o+; "8� �8+= <!o �o+o '�2��� �� 8"< �=+oo �<= �o+<< "&< o;+!! & �8+8� *��� �� ;8; ��+ ; �o �;+!! o" <=+<" ";< � +;& 3�� ������� � "�o=8 �o+&! "�"!8 �o+o; "8o o!+" &=< �&+�o '�2��� �� 8"� �=+ < �;& �;+<; ""= o<+<= < �;+ 8 *��� �� ;;< �<+"o �&o ��+=& ;= �<+�� "o! <8+!= ���) ���� � "�o<� �<+=& "�!8< ��+8! "=& � �o+"& &o= �"+& '�2��� �� =<o ��+�� ��! �o+!; "" o!+=; "=< �"+<! *��� �� ;88 �<+"! �<< ��+;< ;" �!+!! "=< �"+ � � �2 � � "�� � �!+=& 8;; <8+8 "=! � ��+ " &o= �"+& '�2��� �� 8!" ��+!" ��= � �&+ < " ; o&+�! "o � ��+&= *��� �� o < <�+=" <"8 <o+!8 �& < +!o "� <o+<= �� � �� � "�o�; ��+ & "�!;! �<+o= "8< �8+�" &8& � �<+=" '�2��� �� 8<< �=+< o!" �;+�" "&" oo+�! " � �o+o= *��� �� ;"& �"+� <o8 �"+<& o& <=+=< "=" � +;; *+�+ � "��"" �!+&; 8=& �!+ & ";" � +<� &�; <8+;8 '�2��� �� ;o< �!+o& <8" �!+ ! =" �!+8< "8 �"+o" *��� �� ;<; �!+"! <8 �!+ o 8! �&+=8 "o� <;+=& ���� � ��� ��)( �� � �������� �� ��������) �� ��� � � �� /��� ���( e� 0 � ���� 99�::� �$ ��� ��� ������� �� �������� /��) ����0+ .������ � � �� ������� � ���� 99� � "::+ � ��)� ���� �� ���$����� �� ����� �� 2������ ��� ��� ��)( �� ������ �� � k�����) ����+ ��� ��)� ���� �� � $� �2�@ 9 � �������� � � � �� 2��� � � ��� � �� !�� �� � � ��������������������������������� ��� � �� !� � + ��� ������ � � �� � ������ ��h������ �$ ��� �� �(�������� � ��� "� � �� "!m ��� �� ���������� (+ +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( o8 2� ��)���� �� �� �� �� �(� ��� ( ��)�� � � � �������� �� ���$��� ������� ��� �$,����� 99,5��:: $�������+ ������� ����� ����$� �� �� ���� �� ���� � �� ����������) ��2 ����� �� ��� ����$� ��# �� �������� ��� �������� 2��� ������� �� ���$��� ���� �� ��(��� 2� � �� ������ ( � �� ��� 99,5�� $������� ��� ��)(::+ ��� �� ��-��������� ���� � ��� ( �� � �� �� � ��) ���� /��������0 ��� ������� �� � � ( $� �2�� �( ��) ���� /��������0 ��� �������+ �����$���� ���� � ��� ��)( 2�� � ������� �$ /"0 ��������) �� ��� ���( e� �� �$ ��� ������� � � �� ��� ������� �� ��) ����� �� / 0 ��������) �� �� � � �� �$ ��� ������� � � �� ��� ������� �� ��������+ 4� ������� ��)� ����� �� ��� �$ ���� ��� ��)( �� ���� � � ( �� ������� ������� � � � ���� ��� �� k�����) ����+ ����� 2� ������ �� ��� � � �� /��� ���( e� 0 )���� �������� /��) ����0 ��� ������� � ���� 99�::+ ����� 2� ����� ��� ������ �$ ������� � � � ���� 99� � "::+ � ��)� ���� �� ���$����� �� ����� �� 2������ ��� ������ �$ ������� � � �� )�� ��� �� � �!m+ � ���� �( �$ ��� ���� �� �� ����� �� �� � � � �+ 4� ,��� �� � ��� ���� �� �� �� ��� $��� i �� �( "88! �� i��� !!"+ 4� ���� ��� �� � ��� ��� ��)( 2�� � �5����� ( 2� �� ���� � � ���� �5���� $�� ��� *+�+� i � � �� � �2 � 2���� ��������� �� ������ �$$ k�����) ����+ ����� 2� �5 ���� ��� ������ �$ ������� ��������� � � �� ���� �� �� ��2��� �� �����) ��� ���� ������ �� ���� ��� �� � ��) ���� ��� ����� �� � � ( $� �2�� �( ��) ���� ��� ����� �� � � ��� ��� ��� *+�+ 4� �� ������� �� � �������� ��� ����� �� �� ( � � ( �� �� $� �2�� �( �������� ��� ����� �� �������� � 1 (�� � 3�� ������� �� ���) ����+ %�5�� 2� �������) �� ������� �� ��( ��$���� �����) �� $��� ��� ������+ 4�� � ������ ����� $�� ��� ���-������ ������ �� �5 �� ( ��� � �� � $�� ��� ���� ������� ��� ��� ��)( �� ( 2�� � �� ��2��� �� �� ��������� ��� ��� *+�+ �����) ��� ������+ �$��� ��� ������� �� �� ( 2�� � �� ��2��� �� �� ����)��) � � ���g�+�+� �������� � n��� � 1 (�� � 3�� �������� � �2 � �� �� � ��+ � ���������� 7�� � ��� �������) ��� ��� ��������� ����� �� �������� ���2��� � � �� ��� ������� �� ��������� � �� ��� �� ���� ��� � � � ��� �� ��� *+�+ 4� ,�� �������� �$ ��)��,� �� �� ��-��������� �� �������� ���2��� ��� ������� �� ��������� � �� ��� �� � � ���+ ������� ��� �� ��-��������� ����.�/"�"0-1 ���� ������ ��� �� ��)��,� �� �������� �� ��) ���� � � �� ����� �$ � �� ��� ��� �� ���� �� �� ��2��� ��� 2���� �� ���������� ����� � � ���+ 7�� ,����)� $ � �� ��� �� �� ��� � � ��-� ��� .�31 �� ������� � �� ��� � � ������ �� ���������� � �� � � �� ��� ������ )���� ���) ������� �� � � ���+ �� ��� �)�� �$ 4 � ����� $�� �(�����(� 2� ,�� �� � ���� � � ��� �5����� ������ ����� ��2 �� ���� �( � �� ��� ��� �����) �� $��� ��� ��� � ,� ��� ������+ ���� ���� � �������� )������ $�� ����� �� � ����$� �� ��� ��)���+ 4� $������ �������) �� ��� $���� ����) �� ��( �$ ��� �� ��-��������� ����.�/"�"0-1 ���� �� � ���� ���� �� ���� $� ��2�� + 4� ���� ��� �� � /"0 ��) ���� ��� ����� �� � � ( �� �� $� �2�� �( ��) ���� ��� ����� �� / 0 ����)��) � � ��� ���� � �������� � n��� � 1 (�� � 3�� �������� � �2 � �� �� � �� ���������� ( �������� �� ���� ������� �� ��( �� � ���� ���� ���� � � ���+ � ��� �� �� ������ ��� �$,�����( �$ ����� ��� ��)��� � � �� ������ ( ;! +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( $$����� �( ��� ��)� ���� �$ �� ���) �� ����)��) � ��� � � ��� /��� e� ���� ���� � ���( �� q�� �� � "88=0+ �� ��� c���� �$ ,����) ��� 99����� :: �� ���� ���������� ���2��� � ��� � � ���� ���� ) �� ����$� �� � � )��� � � ����������� �$ ������� �( ����) $��� ���� �+ 4� ��))��� �� � �� ��-��������� .�31 ���� ������ h�� �� ��,���) ��� ������� �� �������� ��� �����$���� �������� ��������� � �� ��� � � ������ �� �5� � ���( � �� � � $�� ��� ������ )���� ���) �������+ 6����������� ���� ���� � � �� ���� �� ������� ������ ����� ��>�� $���� ��+ �� �� � ������ �� �������� ������ ������� ����) �$ ����)��) � � ��� ��� ���� �� � �$ ���� ����� �� �� ������ ���2��� ����� � � ���+ �!������� �� ��� �+� % �� n+� s 3(��� .+ / !!"0+ ��(������� ��������) ��� ���� �$ �����-����d�� ���� �������@ �� �������� �$ ���� � � �� ������ �����+ *������ �� ���:��# ��� 5����� � (%/<0� =!;�= <+ e � ��� �+ �+� s '� ���� ��� �+ 3+ /"88!0+ ���� ������� �� �� �� ��(+ *������ �� 5�������� ��� ;���������� ��������� (%� !&� "�+ e� ���� �+� ���� .+� � ���(� .+� s q�� �� � �+ /"88=0+ '����������� �� � ���������� �$ ����)��) � � �� ������� �� ���� �� ����+ *������ �� &�������� <���# ��� (�� "! �""o+ e� ���� �+� s � ���(� .+ /"88�0+ ����-� �(��) 2�� � � � �� ����)� ����+ *������ �� 5����� � % � <!&�<<<+ e� ��� ��� �+� ��) �� �+ 6+� s 4�� ���)�� i+ 1+ /"8==0+ � . ��� ����� 3�����) 1��� 2��� ���� � �(��) ��� �� ����+ *������ �� &�������� +���� �� !=� ""o�"&"+ e� ��� ��� �+� s 4�� ����)�� i+ 1+ /"88 0+ l� ��-� 5���� � � ����� ����� ���� �� ��$������ �� �(� ��� ���� � 2��� ���� � �(��) ��� �� ����+ +���� ���� ' �� /�� $$� "<&�"; + .���� �+ p+ /"8==0+ q� �� ��( ����������� �� ���� � � �����@ ���� ������� �������� ����) ���.�+ *������ �� ����� � +���� ������ )� ;8� 8<+ '� e����� 4+� s �� ��� �+ /"8=�0+ '��� ��� ���� � � �� ������ ��f *������ �� 5����� � � � ;8&�=!=+ '� � ����� �+� s ��� ��� e+ /"88;0+ ������ ���� ���� ������) �� ����$� �� �������,� ���� 2��� ����-� �(��) ��� + *������ �� 5����� � %(/�0� "=="�"8" + '���2��d� �+� � ��� i+� s 1 �� � �� �+ /"88=0+ .�����( �� �������( ��� ����� �� � ����)��) � � ��+ *������ �� 5�������� ��� ;���������� ��������� ))� "=8� "o+ '�� �� e+� s �� �� � e+ /"88�0+ ��� 2�� � ����� �$ $����)� �5�� �)� ��� + �� *������ �� 5����� � % � <<��<;8+ 6 � � �+ /"88"0+ �$,����� � ��� � � ���@ ��+ *������ �� 5����� � �=� "�;��"o";+ 6 � � �+ 6+� s 6������ n+ �+ /"88 0+ ��� �����-������� �$ �5������ ���� �������+ *������ �� 5����� � ��� < ;� <o�+ 6 ������� i+ /"88;0+ �� �5 ��� ���� �$ ��� �����-������� �� �������� �$ ��� �� ������@ *n ��������+ *������ �� +���� ��� ��� ����� ��� �!� ""� "+ 6 ������� i+ / !!!0+ 7� ��� ��������� �� �������� ���2��� ��� �� ������ �� ������ ���� ���� �������+ ��� ��������� ' �� / �� 5�������� ��������� !� &�� <�+ 6������ n+ �+� ���2���� 4+ �+� s �� �� �)�� �+ 6+ /"8=;0+ �5������ ���� ������� �� �� �� ��(+ *������ �� 5�������� +���� ���� $!/"0� &�&!+ � ������ j+ �+� i ) �� �� �� �+� s ��� �� '+ /"88&0+ 7� ��� �� ���� ���2��� ��� �5������ � �� �� ��� �� �� ��( �$ ��� ����� �5���� ������ �� ���� �+ *������ �� 5����� � ��� ";;8�"=!"+ �� �)��� .+� �� �)� e+� s p �)� .+ / !!!0+ � e�� �� �� � �� ��( ���2��� ���� ������ �� �5�� �)� � ���@ �������� $��� ������ ��� � k�+ �� ;���� ��� ' �� / �� +���� ��� ��� 5����� � � � &&;�&�<+ � ���(� .+ /"88"0+ ��� 2�� � ����� �$ ��� �� ��� ��� + *������ �� 5����� � �=� """�"�;+ � ���(� .+ /"88�0+ ��� �����-������� �$ �� �� ��( �� �������� ���� �� ����)��) � � ���+ 5����9 ��:� ��� &�������� <���# ��� !� " �&<+ � ���(� .+ /"88=0+ ��� $����� �$ ���������� �� ����)��) � � ���+ ��+' ' ���� �� ������� ��=+ +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( ;" � ���(� .+ / !!!0+ ��� ������� �$ �5������ ������� �� ������ ���� � � ���+ + �#��# <��: �� ;���� ���� )� & � <8+ i �� p+� .���� 3+� s ���)� 1+ / !!!0+ 3 ��,� e ��� ���� � � ��� �� ������ ���� � ��� ���� ������)+ 3��>�� 5����� *������� $$� "�"o+ j��� �+ e+� s 7� � n+ p+� /"88"0+ ����-� �(��) �� �� ����� �� ���� � � �� �������@ ������ ���� ��������+ �� �+ ���� s �+ .� �) /���+0� &���?�0����� "������ <��: � ' � ���� /q� + ��0+ ������� �@ � ������+ j��2� i+ /"88�0+ ���� �������� ��k ���� �� ��� �� �� ��( �$ )��2�� �� ��� ����( ���� (@ �������� $��� ����)��) � � ���+ *��������( �$ .��� )�� 4�� ��) 3 ��� ������+ 1��� i+� q ������ ��� �+� s n���� i+ / !!!0+ ��� ����� �� �>������ ����� �(��������@ �� � (��� �$ ��� *+�+ �� . � �� � ���� � � ���+ 3��>�� 5����� *������� $$� �&�; + 3�����)� � �+ %+� ���� � �� �+� s 1 ����� �+ /"88�0+ ��� ��������� �� ���� ���2��� ��� �� �������+ *������ �� 5�������� ��� ;���������� ��������� ) � "!"�""o+ �� � �+ /"8;;0+ � �����c�� �$ ��� ���� ������) �����(:� �����+ 3 �� "+ *������ �� 5�������� +���� ���� $� " 8�";o+ ����))�� i+ �+ /"88=0+ ���� ���) ��� ��dd ��) ����������� �� ���� ���2��� ��� � � �� ��� ������� �� ��� ��������� � �� ���@ � �2�-$ ���� ���� ��+ *������ �� 5����� � %)� �;��o!&+ ; +2 3����� � ��2 4 5�������� � ���� � ' �� / $ 6( $7 (�!8(�( intertemporal risk-return relationship in the asian markets around the asian crisis introduction methodology data empirical findings conclusion references financial services review volume 33 number 4 (2025) volume 33, no. 4 2025 guest editor: inga timmerman, ph.d., university of north florida advisory editors: vickie bajtelsmit, ph.d., colorado state university (emeritus) shawn brayman, m.e.s., sb research consulting conrad ciccotello, jd. ph.d., university of denver john grable, ph.d., university of georgia sherman hanna, ph.d., the ohio state university tom potts, ph.d., cfp®, baylor university (emeritus) martin seay, ph.d., cfp®, kansas state university meir statman, ph.d., santa clara university tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors: swarn chatterjee, ph.d., university of georgia shinae l. choi, ph.d., university of alabama lu fan, ph.d., cfp®, university of georgia jasmine fang, ph.d., massey university, nz mark fedenia, ph.d., university of wisconsin philip gibson, ph.d., cfp®, winthrop university stu heckman, ph.d., cfp®, texas tech university stephen horan, ph.d., cfa®, university of north william w. jennings, ph.d., cfa®, u.s. airforce academy so-hyun joo, ph.d., ewha womans university, south korea izidin el kalak, cardiff university, uk thomas langdon, ph.d., roger william university, bristol claire matthews, ph.d., massey university, nz mustafa nourallah, ph.d., centre for research on economic relations, mid sweden university, sweden lance palmer, ph.d., cfp®, cpa®, university of georgia abed rabbani, ph.d., cfp®, university of missouri dan richards, ph.d., york university, canada chris robinson, ph.d., cfp®, cpa, ca, york university (emeritus), canada jerry stevens, ph.d., university of richmond ning tang, ph.d., san diego state university inga timmerman, ph.d., university of north florida bomikazi zeka, ph.d., university of canberra, australia issn online 10.61190/fsr.v33i4 contents editorials timmerman, inga, guest editorial, i-iii. sinnewe, elisabeth, editorial, iv-vi. special issue: the pedagogy of financial planning elliott, william b. & zhang, xianwu, launching a cfp board registered program at an aacsb-accredited business college: a case study and analysis, 1-11. asebedo, sarah & gramse, bryan, capstone as project-led problem based learning: theory and application in personal financial planning, 1236. sinnewe, elisabeth, elevating professional skills through authentic, scaffolded learning in a financial planning capstone, 37-47. fraser, steven p., utilizing experiential learning techniques in a financial planning program: allowing students to learn from themselves, 48-61. anderson, jason & gray, blake, a study of time value of money educational interventions, 62-74. griesdorn, tim & devaney, sharon, a new approach to teaching personal financial education, 75-86. stebbins, richard & quito, diana, effective financial education strategies: empowering students with personal application, 87-105. reiter, miranda, mielitz, katherine & chimbane, tanaka, addressing diversity, equity, and inclusion in financial planning education, 106120. original submissions antonoudi, efthhymia, seay, martin, lim, hanna & kiss, elizabeth, the impact of the online marketplace on fraud: evidence from craigslist from its early adoption in 1995 to its wider expansion in 2006, 121-133. macdonald, kirsten, wildman, karen, loy, ellana & brimble, mark, the value of financial advice: a narrative review and conceptual frameworks, 134-163. ahamed, afm jalal, jakubowska, dominika, pacholek, bogdan & dziewanowska, katarzyna, exploring factors affecting young adults' financial management behavior: a hybrid pls-sem and fsqca approach, 164-190. academy of financial services officers shawn brayman financial planning research consultant president michelle cull western sydney university president-elect kirsten macdonald griffith university executive vp-program thanh ngo east carolina university vice president finance mustafa nourallah mid sweden university vice president international relations thomas korankye university of arizona vice president marketing & public relations elisabeth sinnewe queensland university of technology (qut) chair, australia-new zealand chapter editor, financial services review elisabeth sinnewe queensland university of technology (qut) directors norah feng massey university aaron gilbert auckland university of technology matt goren brett danko education center tom idzorek morningstar eun jin kwak university of wisconsin-green bay dan moisand moisand fitzgerald tamayo richard stebbins university of alabama tom warschauer san diego state university dave yeske golden gate university yu (yulia) zhang kansas state university past presidents shawn brayman, 2023-2025 financial planning research consultant inga timmerman, 2020-22 university of north florida janine sam, 2019-20 shepherd university swarn chatterjee, 2018-19 university of georgia robert moreschi, 2016-18 virginia military institute thomas coe, 2015-16 quinnipiac university william chittenden, 2014-15 texas state university lance palmer, 2013-14 university of georgia frank laatsch, 2012-13 university of southern mississippi brian boscaljon, 2011-12 penn state university-erie auburn university, montgomery vickie hampton, 2008-09 texas tech university frank laatsch 2007-08 university of southern mississippi daniel walz, 2006-07 trinity university anne gleason, 2005-06 college of charleston stuart michelson, 2004-05 stetson university grady perdue, 2003-04 university of houston-clear lake vickie bajtelsmit, 2002-03 colorado state university karen eilers lahey, 2001-02 university of akron tom eyssell, 2000-01 university of missouri-st. louis jill lynn vihtelic, 1999-00 saint mary’s college terry zivney, 1998-99 ball state university don holdren, 1997-98 marshall university robert mcleod, 1996-97 university of alabama walt woerheide, 1995-96 the american college dixie mills, 1994-95 illinois state university ted veit, 1993-94 rollins college mona gardner, 1992-93 illinois wesleyan university jean l. heck, 1991-92 villanova university frank k. reilly, 1990-91 university of notre dame lawrence j. gitman, 1989-90 san diego state university travis s. pritchett, 1988-89 university of south carolina tom potts, 1987-88 baylor university robert f. bohn, 1986-87 golden gate university tom warschauer, 1985-86 san diego state university financial services review is the journal of the academy of financial services financial services review the journal of individual financial management vol. 33, no. 4, 2025 editor elisabeth sinnewe, ph.d. queensland university of technology (qut) editorial advisory board • vickie bajtelsmit, ph.d., colorado state university (emeritus) • shawn brayman, m.e.s., sb research consulting • conrad ciccotello, jd., ph.d., university of denver • john e. grable, ph.d., cfp®, university of georgia • sherman hanna, ph.d., the ohio state university • tom potts, ph.d., cfp®, baylor university (emeritus) • martin seay, ph.d., cfp®, kansas state university • meir statman, ph.d., santa clara university • tom warschauer, ph.d., cfp®, san diego state university (emeritus) associate editors • swarn chatterjee, ph.d., university of georgia • shinae choi, ph.d., university of alabama • lu fan, ph.d., cfp®, university of georgia • jasmine fang, ph.d., massey university, new zealand • mark fedenia, ph.d., university of wisconsin • philip gibson, ph.d., cfp®, winthrop university • stu heckman, ph.d., cfp®, texas tech university • stephen horan, ph.d., cfa®, university of north carolina wilmington • william w. jennings, ph.d., cfa®, u.s. airforce academy • so-hyun joo, ph.d., ewha womans university, south korea • izidin el kalak, cardiff university, united kingdom • thomas langdon, ph.d., roger william university, bristol, ri • claire matthews, ph.d., massey university, new zealand • mustafa nourallah, centre for research on economic relations, mid sweden university, sweden • lance palmer, ph.d., cfp®, cpa®, university of georgia • abed rabbani, ph.d., cfp®, university of missouri • dan richards, ph.d., york university • chris robinson, ph.d., cfpretired™, cpa, ca, york university (emeritus), canada • jerry stevens, ph.d., university of richmond • ning tang, ph.d., san diego state university • inga timmerman, ph.d., university of north florida • bomikazi zeka, ph.d., university of canberra, australia editorial board • john anderson, ph.d., university of kansas • kristy archuleta, ph.d., university of georgia • axton betz-hamilton, ph.d., south dakota state university • alona bilokha, ph.d., university of north florida • brian l. boscaljon, ph.d., penn state behrend • colleeen tokar asaad, ph.d., baldwin wallace university • rachel bi, ph.d., utah valley university • chris browning, ph.d., cfp®, texas tech university • john clinebell, ph..d., university of northern colorado (emeritus) • michelle cull, ph.d., western sydney university, australia • james delellio, ph.d., pepperdine university • dale domian, ph.d., cfp®, york university, canada • norah feng, ph.d., massey university, new zealand • giovanni fernandez, ph..d. stetson university, deland, fl • patti fisher, ph.d., virginia tech • jim gilkeson, ph.d., cfa, university of central florida • martie gillen, ph.d., university of florida • chuck grace, cfp®, ivy school of business, canada • drew hanks, ph.d., the ohio state university • jennifer harrison, ph.d., southern cross university • wookjae heo, ph.d., purdue university • stephen m. horan, ph.d., certified financial planner board of standards, inc. • russell james, ph.d., cfp®, texas tech university • kyoung tae kim, ph.d., university of alabama • eun jin kwak, ph.d., university of wisconsin, green bay • derek lawson, ph.d., cfp®, kansas state university • sunwoo lee, ph.d., york university, canada • yi liu, ph.d., cfp®, st. john fisher college • caezilia loibl, ph.d., the ohio state university • megan mccoy, ph.d., lmft, cft-i®, kansas state university • ronald mciver, university of south australia • barry mulholland, ph.d., cfp®, university of akron • david nanigian, ph.d., cfp®, mount ararat financial services llc • john nofsinger, ph.d., university of alaska anchorage • olamide olajide (lami), ph.d., cfp®, afc, texas tech university • congrong ouyang, ph.d., kansas state university • wade d. pfau, ph.d., cfa, ricp, retirement income style awareness, llc • miranda reiter, ph.d., cfp®, texas tech university • aman sunder, ph.d., college for financial planning • kimberly watkins, ph.d., university of georgia • anne wenger, ph.d., san diego state university • steffen westermann, ph.d., griffith university • tansel yilmazer, ph.d., cfp®, the ohio state university • yu zhang, ph.d., kansas state university the editor of financial services reviewwishes to thank university of georgia for its support of the journal financial services review (fsr) is the official publication of the academy of financial services. fsr is a diamond open access journal, which means there are no fees or restrictions for access to or submission of research and no article processing fees if published. the purpose of this double-blind peer-reviewed academic journal is to encourage research that examines the impact of financial issues on individuals. in contrast to the many corporate or institutional journals that are available in finance, the focus of this journal is on individual financial management. fsr provides a forum for those who are interested in the individual perspective on issues in the areas of financial planning, financial counseling, financial literacy, banking/banking services, education in financial services, employee benefits, estate and tax planning, insurance planning, investments, mutual funds, non-bank financial institutions, pension and retirement, planning, and real estate. while the annual meeting held each fall provides an opportunity to discuss and present these topics to colleagues, the journal allows a much wider audience of those interested in this subject matter. to encourage the development of curricula in financial services at the university level, appropriate pedagogical papers are accepted for publication. manuscripts are encouraged that present ideas about appropriate content, methods of teaching, and materials. contributions from practitioners who are actively involved in financial planning, financial services, and professional associations are also encouraged. while the primary purpose of this journal is the publication of traditional academic empirical research, the academy believes that it is important to encourage the cross fertilization of ideas and an exchange of information of interest to both academicians and practitioners. thus, the editor seeks manuscripts from practitioners that present innovative ideas and new information in financial planning and services or suggest new avenues of research for academics. this work is licensed under a creative commons attribution-noncommercial 4.0 international license. author(s) retain copyright and grant the journal right of first publication with the work simultaneously licensed under a creative commons attribution-noncommercial 4.0 international license that allows to share the work with an acknowledgment of the work's authorship and initial publication in this journal. this license allows the author to remix, tweak, and build upon the original work non-commercially. the new work(s) must be non-commercial and acknowledge the original work. https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://creativecommons.org/licenses/by-nc/4.0/ https://www.lib.sfu.ca/help/publish/scholarly-publishing/radical-access/open-access-colour-classifications